vendor/llvm-project/llvmorg-14-init-10186-gff7f2cfa959b

332 files changed, 13659 insertions, 8776 deletions

diff --git a/compiler-rt/include/profile/InstrProfData.inc b/compiler-rt/include/profile/InstrProfData.inc
index 08a642469627..008b8dde5820 100644
--- a/compiler-rt/include/profile/InstrProfData.inc
+++ b/compiler-rt/include/profile/InstrProfData.inc
@@ -75,9 +75,7 @@ INSTR_PROF_DATA(const uint64_t, llvm::Type::getInt64Ty(Ctx), NameRef, \
 INSTR_PROF_DATA(const uint64_t, llvm::Type::getInt64Ty(Ctx), FuncHash, \
                 ConstantInt::get(llvm::Type::getInt64Ty(Ctx), \
                 Inc->getHash()->getZExtValue()))
-INSTR_PROF_DATA(const IntPtrT, llvm::Type::getInt64PtrTy(Ctx), CounterPtr, \
-                ConstantExpr::getBitCast(CounterPtr, \
-                llvm::Type::getInt64PtrTy(Ctx)))
+INSTR_PROF_DATA(const IntPtrT, IntPtrTy, CounterPtr, RelativeCounterPtr)
 /* This is used to map function pointers for the indirect call targets to
  * function name hashes during the conversion from raw to merged profile
  * data.
@@ -129,15 +127,16 @@ INSTR_PROF_VALUE_NODE(PtrToNodeT, llvm::Type::getInt8PtrTy(Ctx), Next, \
 #endif
 INSTR_PROF_RAW_HEADER(uint64_t, Magic, __llvm_profile_get_magic())
 INSTR_PROF_RAW_HEADER(uint64_t, Version, __llvm_profile_get_version())
+INSTR_PROF_RAW_HEADER(uint64_t, BinaryIdsSize, __llvm_write_binary_ids(NULL))
 INSTR_PROF_RAW_HEADER(uint64_t, DataSize, DataSize)
 INSTR_PROF_RAW_HEADER(uint64_t, PaddingBytesBeforeCounters, PaddingBytesBeforeCounters)
 INSTR_PROF_RAW_HEADER(uint64_t, CountersSize, CountersSize)
 INSTR_PROF_RAW_HEADER(uint64_t, PaddingBytesAfterCounters, PaddingBytesAfterCounters)
 INSTR_PROF_RAW_HEADER(uint64_t, NamesSize,  NamesSize)
-INSTR_PROF_RAW_HEADER(uint64_t, CountersDelta, (uintptr_t)CountersBegin)
+INSTR_PROF_RAW_HEADER(uint64_t, CountersDelta,
+                      (uintptr_t)CountersBegin - (uintptr_t)DataBegin)
 INSTR_PROF_RAW_HEADER(uint64_t, NamesDelta, (uintptr_t)NamesBegin)
 INSTR_PROF_RAW_HEADER(uint64_t, ValueKindLast, IPVK_Last)
-INSTR_PROF_RAW_HEADER(uint64_t, BinaryIdsSize, __llvm_write_binary_ids(NULL))
 #undef INSTR_PROF_RAW_HEADER
 /* INSTR_PROF_RAW_HEADER  end */
 
@@ -646,7 +645,7 @@ serializeValueProfDataFrom(ValueProfRecordClosure *Closure,
         (uint64_t)'f' << 16 | (uint64_t)'R' << 8 | (uint64_t)129
 
 /* Raw profile format version (start from 1). */
-#define INSTR_PROF_RAW_VERSION 6
+#define INSTR_PROF_RAW_VERSION 8
 /* Indexed profile format version (start from 1). */
 #define INSTR_PROF_INDEX_VERSION 7
 /* Coverage mapping format version (start from 0). */
diff --git a/compiler-rt/include/sanitizer/asan_interface.h b/compiler-rt/include/sanitizer/asan_interface.h
index 792ef9cfaa32..9bff21c117b3 100644
--- a/compiler-rt/include/sanitizer/asan_interface.h
+++ b/compiler-rt/include/sanitizer/asan_interface.h
@@ -316,7 +316,7 @@ void *__asan_addr_is_in_fake_stack(void *fake_stack, void *addr, void **beg,
 void __asan_handle_no_return(void);
 
 /// Update allocation stack trace for the given allocation to the current stack
-/// trace. Returns 1 if successfull, 0 if not.
+/// trace. Returns 1 if successful, 0 if not.
 int __asan_update_allocation_context(void* addr);
 
 #ifdef __cplusplus
diff --git a/compiler-rt/include/sanitizer/common_interface_defs.h b/compiler-rt/include/sanitizer/common_interface_defs.h
index cd69285b8d4a..692b8f70c969 100644
--- a/compiler-rt/include/sanitizer/common_interface_defs.h
+++ b/compiler-rt/include/sanitizer/common_interface_defs.h
@@ -28,7 +28,7 @@ typedef struct {
   // Enable sandbox support in sanitizer coverage.
   int coverage_sandboxed;
   // File descriptor to write coverage data to. If -1 is passed, a file will
-  // be pre-opened by __sanitizer_sandobx_on_notify(). This field has no
+  // be pre-opened by __sanitizer_sandbox_on_notify(). This field has no
   // effect if coverage_sandboxed == 0.
   intptr_t coverage_fd;
   // If non-zero, split the coverage data into well-formed blocks. This is
diff --git a/compiler-rt/include/sanitizer/dfsan_interface.h b/compiler-rt/include/sanitizer/dfsan_interface.h
index cd3b6d6e2b16..d6209a3ea2b2 100644
--- a/compiler-rt/include/sanitizer/dfsan_interface.h
+++ b/compiler-rt/include/sanitizer/dfsan_interface.h
@@ -150,8 +150,7 @@ int dfsan_get_track_origins(void);
 #ifdef __cplusplus
 }  // extern "C"
 
-template <typename T>
-void dfsan_set_label(dfsan_label label, T &data) { // NOLINT
+template <typename T> void dfsan_set_label(dfsan_label label, T &data) {
   dfsan_set_label(label, (void *)&data, sizeof(T));
 }
 
diff --git a/compiler-rt/include/sanitizer/linux_syscall_hooks.h b/compiler-rt/include/sanitizer/linux_syscall_hooks.h
index 56eae3d40f96..3f3f1e78dfb8 100644
--- a/compiler-rt/include/sanitizer/linux_syscall_hooks.h
+++ b/compiler-rt/include/sanitizer/linux_syscall_hooks.h
@@ -20,1493 +20,1502 @@
 #ifndef SANITIZER_LINUX_SYSCALL_HOOKS_H
 #define SANITIZER_LINUX_SYSCALL_HOOKS_H
 
-#define __sanitizer_syscall_pre_time(tloc) \
+#define __sanitizer_syscall_pre_time(tloc)                                     \
   __sanitizer_syscall_pre_impl_time((long)(tloc))
-#define __sanitizer_syscall_post_time(res, tloc) \
+#define __sanitizer_syscall_post_time(res, tloc)                               \
   __sanitizer_syscall_post_impl_time(res, (long)(tloc))
-#define __sanitizer_syscall_pre_stime(tptr) \
+#define __sanitizer_syscall_pre_stime(tptr)                                    \
   __sanitizer_syscall_pre_impl_stime((long)(tptr))
-#define __sanitizer_syscall_post_stime(res, tptr) \
+#define __sanitizer_syscall_post_stime(res, tptr)                              \
   __sanitizer_syscall_post_impl_stime(res, (long)(tptr))
-#define __sanitizer_syscall_pre_gettimeofday(tv, tz) \
+#define __sanitizer_syscall_pre_gettimeofday(tv, tz)                           \
   __sanitizer_syscall_pre_impl_gettimeofday((long)(tv), (long)(tz))
-#define __sanitizer_syscall_post_gettimeofday(res, tv, tz) \
+#define __sanitizer_syscall_post_gettimeofday(res, tv, tz)                     \
   __sanitizer_syscall_post_impl_gettimeofday(res, (long)(tv), (long)(tz))
-#define __sanitizer_syscall_pre_settimeofday(tv, tz) \
+#define __sanitizer_syscall_pre_settimeofday(tv, tz)                           \
   __sanitizer_syscall_pre_impl_settimeofday((long)(tv), (long)(tz))
-#define __sanitizer_syscall_post_settimeofday(res, tv, tz) \
+#define __sanitizer_syscall_post_settimeofday(res, tv, tz)                     \
   __sanitizer_syscall_post_impl_settimeofday(res, (long)(tv), (long)(tz))
-#define __sanitizer_syscall_pre_adjtimex(txc_p) \
+#define __sanitizer_syscall_pre_adjtimex(txc_p)                                \
   __sanitizer_syscall_pre_impl_adjtimex((long)(txc_p))
-#define __sanitizer_syscall_post_adjtimex(res, txc_p) \
+#define __sanitizer_syscall_post_adjtimex(res, txc_p)                          \
   __sanitizer_syscall_post_impl_adjtimex(res, (long)(txc_p))
-#define __sanitizer_syscall_pre_times(tbuf) \
+#define __sanitizer_syscall_pre_times(tbuf)                                    \
   __sanitizer_syscall_pre_impl_times((long)(tbuf))
-#define __sanitizer_syscall_post_times(res, tbuf) \
+#define __sanitizer_syscall_post_times(res, tbuf)                              \
   __sanitizer_syscall_post_impl_times(res, (long)(tbuf))
 #define __sanitizer_syscall_pre_gettid() __sanitizer_syscall_pre_impl_gettid()
-#define __sanitizer_syscall_post_gettid(res) \
+#define __sanitizer_syscall_post_gettid(res)                                   \
   __sanitizer_syscall_post_impl_gettid(res)
-#define __sanitizer_syscall_pre_nanosleep(rqtp, rmtp) \
+#define __sanitizer_syscall_pre_nanosleep(rqtp, rmtp)                          \
   __sanitizer_syscall_pre_impl_nanosleep((long)(rqtp), (long)(rmtp))
-#define __sanitizer_syscall_post_nanosleep(res, rqtp, rmtp) \
+#define __sanitizer_syscall_post_nanosleep(res, rqtp, rmtp)                    \
   __sanitizer_syscall_post_impl_nanosleep(res, (long)(rqtp), (long)(rmtp))
-#define __sanitizer_syscall_pre_alarm(seconds) \
+#define __sanitizer_syscall_pre_alarm(seconds)                                 \
   __sanitizer_syscall_pre_impl_alarm((long)(seconds))
-#define __sanitizer_syscall_post_alarm(res, seconds) \
+#define __sanitizer_syscall_post_alarm(res, seconds)                           \
   __sanitizer_syscall_post_impl_alarm(res, (long)(seconds))
 #define __sanitizer_syscall_pre_getpid() __sanitizer_syscall_pre_impl_getpid()
-#define __sanitizer_syscall_post_getpid(res) \
+#define __sanitizer_syscall_post_getpid(res)                                   \
   __sanitizer_syscall_post_impl_getpid(res)
 #define __sanitizer_syscall_pre_getppid() __sanitizer_syscall_pre_impl_getppid()
-#define __sanitizer_syscall_post_getppid(res) \
+#define __sanitizer_syscall_post_getppid(res)                                  \
   __sanitizer_syscall_post_impl_getppid(res)
 #define __sanitizer_syscall_pre_getuid() __sanitizer_syscall_pre_impl_getuid()
-#define __sanitizer_syscall_post_getuid(res) \
+#define __sanitizer_syscall_post_getuid(res)                                   \
   __sanitizer_syscall_post_impl_getuid(res)
 #define __sanitizer_syscall_pre_geteuid() __sanitizer_syscall_pre_impl_geteuid()
-#define __sanitizer_syscall_post_geteuid(res) \
+#define __sanitizer_syscall_post_geteuid(res)                                  \
   __sanitizer_syscall_post_impl_geteuid(res)
 #define __sanitizer_syscall_pre_getgid() __sanitizer_syscall_pre_impl_getgid()
-#define __sanitizer_syscall_post_getgid(res) \
+#define __sanitizer_syscall_post_getgid(res)                                   \
   __sanitizer_syscall_post_impl_getgid(res)
 #define __sanitizer_syscall_pre_getegid() __sanitizer_syscall_pre_impl_getegid()
-#define __sanitizer_syscall_post_getegid(res) \
+#define __sanitizer_syscall_post_getegid(res)                                  \
   __sanitizer_syscall_post_impl_getegid(res)
-#define __sanitizer_syscall_pre_getresuid(ruid, euid, suid)          \
-  __sanitizer_syscall_pre_impl_getresuid((long)(ruid), (long)(euid), \
+#define __sanitizer_syscall_pre_getresuid(ruid, euid, suid)                    \
+  __sanitizer_syscall_pre_impl_getresuid((long)(ruid), (long)(euid),           \
                                          (long)(suid))
-#define __sanitizer_syscall_post_getresuid(res, ruid, euid, suid)          \
-  __sanitizer_syscall_post_impl_getresuid(res, (long)(ruid), (long)(euid), \
+#define __sanitizer_syscall_post_getresuid(res, ruid, euid, suid)              \
+  __sanitizer_syscall_post_impl_getresuid(res, (long)(ruid), (long)(euid),     \
                                           (long)(suid))
-#define __sanitizer_syscall_pre_getresgid(rgid, egid, sgid)          \
-  __sanitizer_syscall_pre_impl_getresgid((long)(rgid), (long)(egid), \
+#define __sanitizer_syscall_pre_getresgid(rgid, egid, sgid)                    \
+  __sanitizer_syscall_pre_impl_getresgid((long)(rgid), (long)(egid),           \
                                          (long)(sgid))
-#define __sanitizer_syscall_post_getresgid(res, rgid, egid, sgid)          \
-  __sanitizer_syscall_post_impl_getresgid(res, (long)(rgid), (long)(egid), \
+#define __sanitizer_syscall_post_getresgid(res, rgid, egid, sgid)              \
+  __sanitizer_syscall_post_impl_getresgid(res, (long)(rgid), (long)(egid),     \
                                           (long)(sgid))
-#define __sanitizer_syscall_pre_getpgid(pid) \
+#define __sanitizer_syscall_pre_getpgid(pid)                                   \
   __sanitizer_syscall_pre_impl_getpgid((long)(pid))
-#define __sanitizer_syscall_post_getpgid(res, pid) \
+#define __sanitizer_syscall_post_getpgid(res, pid)                             \
   __sanitizer_syscall_post_impl_getpgid(res, (long)(pid))
 #define __sanitizer_syscall_pre_getpgrp() __sanitizer_syscall_pre_impl_getpgrp()
-#define __sanitizer_syscall_post_getpgrp(res) \
+#define __sanitizer_syscall_post_getpgrp(res)                                  \
   __sanitizer_syscall_post_impl_getpgrp(res)
-#define __sanitizer_syscall_pre_getsid(pid) \
+#define __sanitizer_syscall_pre_getsid(pid)                                    \
   __sanitizer_syscall_pre_impl_getsid((long)(pid))
-#define __sanitizer_syscall_post_getsid(res, pid) \
+#define __sanitizer_syscall_post_getsid(res, pid)                              \
   __sanitizer_syscall_post_impl_getsid(res, (long)(pid))
-#define __sanitizer_syscall_pre_getgroups(gidsetsize, grouplist) \
+#define __sanitizer_syscall_pre_getgroups(gidsetsize, grouplist)               \
   __sanitizer_syscall_pre_impl_getgroups((long)(gidsetsize), (long)(grouplist))
-#define __sanitizer_syscall_post_getgroups(res, gidsetsize, grouplist) \
-  __sanitizer_syscall_post_impl_getgroups(res, (long)(gidsetsize),     \
+#define __sanitizer_syscall_post_getgroups(res, gidsetsize, grouplist)         \
+  __sanitizer_syscall_post_impl_getgroups(res, (long)(gidsetsize),             \
                                           (long)(grouplist))
-#define __sanitizer_syscall_pre_setregid(rgid, egid) \
+#define __sanitizer_syscall_pre_setregid(rgid, egid)                           \
   __sanitizer_syscall_pre_impl_setregid((long)(rgid), (long)(egid))
-#define __sanitizer_syscall_post_setregid(res, rgid, egid) \
+#define __sanitizer_syscall_post_setregid(res, rgid, egid)                     \
   __sanitizer_syscall_post_impl_setregid(res, (long)(rgid), (long)(egid))
-#define __sanitizer_syscall_pre_setgid(gid) \
+#define __sanitizer_syscall_pre_setgid(gid)                                    \
   __sanitizer_syscall_pre_impl_setgid((long)(gid))
-#define __sanitizer_syscall_post_setgid(res, gid) \
+#define __sanitizer_syscall_post_setgid(res, gid)                              \
   __sanitizer_syscall_post_impl_setgid(res, (long)(gid))
-#define __sanitizer_syscall_pre_setreuid(ruid, euid) \
+#define __sanitizer_syscall_pre_setreuid(ruid, euid)                           \
   __sanitizer_syscall_pre_impl_setreuid((long)(ruid), (long)(euid))
-#define __sanitizer_syscall_post_setreuid(res, ruid, euid) \
+#define __sanitizer_syscall_post_setreuid(res, ruid, euid)                     \
   __sanitizer_syscall_post_impl_setreuid(res, (long)(ruid), (long)(euid))
-#define __sanitizer_syscall_pre_setuid(uid) \
+#define __sanitizer_syscall_pre_setuid(uid)                                    \
   __sanitizer_syscall_pre_impl_setuid((long)(uid))
-#define __sanitizer_syscall_post_setuid(res, uid) \
+#define __sanitizer_syscall_post_setuid(res, uid)                              \
   __sanitizer_syscall_post_impl_setuid(res, (long)(uid))
-#define __sanitizer_syscall_pre_setresuid(ruid, euid, suid)          \
-  __sanitizer_syscall_pre_impl_setresuid((long)(ruid), (long)(euid), \
+#define __sanitizer_syscall_pre_setresuid(ruid, euid, suid)                    \
+  __sanitizer_syscall_pre_impl_setresuid((long)(ruid), (long)(euid),           \
                                          (long)(suid))
-#define __sanitizer_syscall_post_setresuid(res, ruid, euid, suid)          \
-  __sanitizer_syscall_post_impl_setresuid(res, (long)(ruid), (long)(euid), \
+#define __sanitizer_syscall_post_setresuid(res, ruid, euid, suid)              \
+  __sanitizer_syscall_post_impl_setresuid(res, (long)(ruid), (long)(euid),     \
                                           (long)(suid))
-#define __sanitizer_syscall_pre_setresgid(rgid, egid, sgid)          \
-  __sanitizer_syscall_pre_impl_setresgid((long)(rgid), (long)(egid), \
+#define __sanitizer_syscall_pre_setresgid(rgid, egid, sgid)                    \
+  __sanitizer_syscall_pre_impl_setresgid((long)(rgid), (long)(egid),           \
                                          (long)(sgid))
-#define __sanitizer_syscall_post_setresgid(res, rgid, egid, sgid)          \
-  __sanitizer_syscall_post_impl_setresgid(res, (long)(rgid), (long)(egid), \
+#define __sanitizer_syscall_post_setresgid(res, rgid, egid, sgid)              \
+  __sanitizer_syscall_post_impl_setresgid(res, (long)(rgid), (long)(egid),     \
                                           (long)(sgid))
-#define __sanitizer_syscall_pre_setfsuid(uid) \
+#define __sanitizer_syscall_pre_setfsuid(uid)                                  \
   __sanitizer_syscall_pre_impl_setfsuid((long)(uid))
-#define __sanitizer_syscall_post_setfsuid(res, uid) \
+#define __sanitizer_syscall_post_setfsuid(res, uid)                            \
   __sanitizer_syscall_post_impl_setfsuid(res, (long)(uid))
-#define __sanitizer_syscall_pre_setfsgid(gid) \
+#define __sanitizer_syscall_pre_setfsgid(gid)                                  \
   __sanitizer_syscall_pre_impl_setfsgid((long)(gid))
-#define __sanitizer_syscall_post_setfsgid(res, gid) \
+#define __sanitizer_syscall_post_setfsgid(res, gid)                            \
   __sanitizer_syscall_post_impl_setfsgid(res, (long)(gid))
-#define __sanitizer_syscall_pre_setpgid(pid, pgid) \
+#define __sanitizer_syscall_pre_setpgid(pid, pgid)                             \
   __sanitizer_syscall_pre_impl_setpgid((long)(pid), (long)(pgid))
-#define __sanitizer_syscall_post_setpgid(res, pid, pgid) \
+#define __sanitizer_syscall_post_setpgid(res, pid, pgid)                       \
   __sanitizer_syscall_post_impl_setpgid(res, (long)(pid), (long)(pgid))
 #define __sanitizer_syscall_pre_setsid() __sanitizer_syscall_pre_impl_setsid()
-#define __sanitizer_syscall_post_setsid(res) \
+#define __sanitizer_syscall_post_setsid(res)                                   \
   __sanitizer_syscall_post_impl_setsid(res)
-#define __sanitizer_syscall_pre_setgroups(gidsetsize, grouplist) \
+#define __sanitizer_syscall_pre_setgroups(gidsetsize, grouplist)               \
   __sanitizer_syscall_pre_impl_setgroups((long)(gidsetsize), (long)(grouplist))
-#define __sanitizer_syscall_post_setgroups(res, gidsetsize, grouplist) \
-  __sanitizer_syscall_post_impl_setgroups(res, (long)(gidsetsize),     \
+#define __sanitizer_syscall_post_setgroups(res, gidsetsize, grouplist)         \
+  __sanitizer_syscall_post_impl_setgroups(res, (long)(gidsetsize),             \
                                           (long)(grouplist))
-#define __sanitizer_syscall_pre_acct(name) \
+#define __sanitizer_syscall_pre_acct(name)                                     \
   __sanitizer_syscall_pre_impl_acct((long)(name))
-#define __sanitizer_syscall_post_acct(res, name) \
+#define __sanitizer_syscall_post_acct(res, name)                               \
   __sanitizer_syscall_post_impl_acct(res, (long)(name))
-#define __sanitizer_syscall_pre_capget(header, dataptr) \
+#define __sanitizer_syscall_pre_capget(header, dataptr)                        \
   __sanitizer_syscall_pre_impl_capget((long)(header), (long)(dataptr))
-#define __sanitizer_syscall_post_capget(res, header, dataptr) \
+#define __sanitizer_syscall_post_capget(res, header, dataptr)                  \
   __sanitizer_syscall_post_impl_capget(res, (long)(header), (long)(dataptr))
-#define __sanitizer_syscall_pre_capset(header, data) \
+#define __sanitizer_syscall_pre_capset(header, data)                           \
   __sanitizer_syscall_pre_impl_capset((long)(header), (long)(data))
-#define __sanitizer_syscall_post_capset(res, header, data) \
+#define __sanitizer_syscall_post_capset(res, header, data)                     \
   __sanitizer_syscall_post_impl_capset(res, (long)(header), (long)(data))
-#define __sanitizer_syscall_pre_personality(personality) \
+#define __sanitizer_syscall_pre_personality(personality)                       \
   __sanitizer_syscall_pre_impl_personality((long)(personality))
-#define __sanitizer_syscall_post_personality(res, personality) \
+#define __sanitizer_syscall_post_personality(res, personality)                 \
   __sanitizer_syscall_post_impl_personality(res, (long)(personality))
-#define __sanitizer_syscall_pre_sigpending(set) \
+#define __sanitizer_syscall_pre_sigpending(set)                                \
   __sanitizer_syscall_pre_impl_sigpending((long)(set))
-#define __sanitizer_syscall_post_sigpending(res, set) \
+#define __sanitizer_syscall_post_sigpending(res, set)                          \
   __sanitizer_syscall_post_impl_sigpending(res, (long)(set))
-#define __sanitizer_syscall_pre_sigprocmask(how, set, oset)          \
-  __sanitizer_syscall_pre_impl_sigprocmask((long)(how), (long)(set), \
+#define __sanitizer_syscall_pre_sigprocmask(how, set, oset)                    \
+  __sanitizer_syscall_pre_impl_sigprocmask((long)(how), (long)(set),           \
                                            (long)(oset))
-#define __sanitizer_syscall_post_sigprocmask(res, how, set, oset)          \
-  __sanitizer_syscall_post_impl_sigprocmask(res, (long)(how), (long)(set), \
+#define __sanitizer_syscall_post_sigprocmask(res, how, set, oset)              \
+  __sanitizer_syscall_post_impl_sigprocmask(res, (long)(how), (long)(set),     \
                                             (long)(oset))
-#define __sanitizer_syscall_pre_getitimer(which, value) \
+#define __sanitizer_syscall_pre_getitimer(which, value)                        \
   __sanitizer_syscall_pre_impl_getitimer((long)(which), (long)(value))
-#define __sanitizer_syscall_post_getitimer(res, which, value) \
+#define __sanitizer_syscall_post_getitimer(res, which, value)                  \
   __sanitizer_syscall_post_impl_getitimer(res, (long)(which), (long)(value))
-#define __sanitizer_syscall_pre_setitimer(which, value, ovalue)        \
-  __sanitizer_syscall_pre_impl_setitimer((long)(which), (long)(value), \
+#define __sanitizer_syscall_pre_setitimer(which, value, ovalue)                \
+  __sanitizer_syscall_pre_impl_setitimer((long)(which), (long)(value),         \
                                          (long)(ovalue))
-#define __sanitizer_syscall_post_setitimer(res, which, value, ovalue)        \
-  __sanitizer_syscall_post_impl_setitimer(res, (long)(which), (long)(value), \
+#define __sanitizer_syscall_post_setitimer(res, which, value, ovalue)          \
+  __sanitizer_syscall_post_impl_setitimer(res, (long)(which), (long)(value),   \
                                           (long)(ovalue))
-#define __sanitizer_syscall_pre_timer_create(which_clock, timer_event_spec, \
-                                             created_timer_id)              \
-  __sanitizer_syscall_pre_impl_timer_create(                                \
+#define __sanitizer_syscall_pre_timer_create(which_clock, timer_event_spec,    \
+                                             created_timer_id)                 \
+  __sanitizer_syscall_pre_impl_timer_create(                                   \
       (long)(which_clock), (long)(timer_event_spec), (long)(created_timer_id))
-#define __sanitizer_syscall_post_timer_create(                         \
-    res, which_clock, timer_event_spec, created_timer_id)              \
-  __sanitizer_syscall_post_impl_timer_create(res, (long)(which_clock), \
-                                             (long)(timer_event_spec), \
+#define __sanitizer_syscall_post_timer_create(                                 \
+    res, which_clock, timer_event_spec, created_timer_id)                      \
+  __sanitizer_syscall_post_impl_timer_create(res, (long)(which_clock),         \
+                                             (long)(timer_event_spec),         \
                                              (long)(created_timer_id))
-#define __sanitizer_syscall_pre_timer_gettime(timer_id, setting) \
+#define __sanitizer_syscall_pre_timer_gettime(timer_id, setting)               \
   __sanitizer_syscall_pre_impl_timer_gettime((long)(timer_id), (long)(setting))
-#define __sanitizer_syscall_post_timer_gettime(res, timer_id, setting) \
-  __sanitizer_syscall_post_impl_timer_gettime(res, (long)(timer_id),   \
+#define __sanitizer_syscall_post_timer_gettime(res, timer_id, setting)         \
+  __sanitizer_syscall_post_impl_timer_gettime(res, (long)(timer_id),           \
                                               (long)(setting))
-#define __sanitizer_syscall_pre_timer_getoverrun(timer_id) \
+#define __sanitizer_syscall_pre_timer_getoverrun(timer_id)                     \
   __sanitizer_syscall_pre_impl_timer_getoverrun((long)(timer_id))
-#define __sanitizer_syscall_post_timer_getoverrun(res, timer_id) \
+#define __sanitizer_syscall_post_timer_getoverrun(res, timer_id)               \
   __sanitizer_syscall_post_impl_timer_getoverrun(res, (long)(timer_id))
-#define __sanitizer_syscall_pre_timer_settime(timer_id, flags, new_setting,   \
-                                              old_setting)                    \
-  __sanitizer_syscall_pre_impl_timer_settime((long)(timer_id), (long)(flags), \
-                                             (long)(new_setting),             \
+#define __sanitizer_syscall_pre_timer_settime(timer_id, flags, new_setting,    \
+                                              old_setting)                     \
+  __sanitizer_syscall_pre_impl_timer_settime((long)(timer_id), (long)(flags),  \
+                                             (long)(new_setting),              \
                                              (long)(old_setting))
-#define __sanitizer_syscall_post_timer_settime(res, timer_id, flags,     \
-                                               new_setting, old_setting) \
-  __sanitizer_syscall_post_impl_timer_settime(                           \
-      res, (long)(timer_id), (long)(flags), (long)(new_setting),         \
+#define __sanitizer_syscall_post_timer_settime(res, timer_id, flags,           \
+                                               new_setting, old_setting)       \
+  __sanitizer_syscall_post_impl_timer_settime(                                 \
+      res, (long)(timer_id), (long)(flags), (long)(new_setting),               \
       (long)(old_setting))
-#define __sanitizer_syscall_pre_timer_delete(timer_id) \
+#define __sanitizer_syscall_pre_timer_delete(timer_id)                         \
   __sanitizer_syscall_pre_impl_timer_delete((long)(timer_id))
-#define __sanitizer_syscall_post_timer_delete(res, timer_id) \
+#define __sanitizer_syscall_post_timer_delete(res, timer_id)                   \
   __sanitizer_syscall_post_impl_timer_delete(res, (long)(timer_id))
-#define __sanitizer_syscall_pre_clock_settime(which_clock, tp) \
+#define __sanitizer_syscall_pre_clock_settime(which_clock, tp)                 \
   __sanitizer_syscall_pre_impl_clock_settime((long)(which_clock), (long)(tp))
-#define __sanitizer_syscall_post_clock_settime(res, which_clock, tp)    \
-  __sanitizer_syscall_post_impl_clock_settime(res, (long)(which_clock), \
+#define __sanitizer_syscall_post_clock_settime(res, which_clock, tp)           \
+  __sanitizer_syscall_post_impl_clock_settime(res, (long)(which_clock),        \
                                               (long)(tp))
-#define __sanitizer_syscall_pre_clock_gettime(which_clock, tp) \
+#define __sanitizer_syscall_pre_clock_gettime(which_clock, tp)                 \
   __sanitizer_syscall_pre_impl_clock_gettime((long)(which_clock), (long)(tp))
-#define __sanitizer_syscall_post_clock_gettime(res, which_clock, tp)    \
-  __sanitizer_syscall_post_impl_clock_gettime(res, (long)(which_clock), \
+#define __sanitizer_syscall_post_clock_gettime(res, which_clock, tp)           \
+  __sanitizer_syscall_post_impl_clock_gettime(res, (long)(which_clock),        \
                                               (long)(tp))
-#define __sanitizer_syscall_pre_clock_adjtime(which_clock, tx) \
+#define __sanitizer_syscall_pre_clock_adjtime(which_clock, tx)                 \
   __sanitizer_syscall_pre_impl_clock_adjtime((long)(which_clock), (long)(tx))
-#define __sanitizer_syscall_post_clock_adjtime(res, which_clock, tx)    \
-  __sanitizer_syscall_post_impl_clock_adjtime(res, (long)(which_clock), \
+#define __sanitizer_syscall_post_clock_adjtime(res, which_clock, tx)           \
+  __sanitizer_syscall_post_impl_clock_adjtime(res, (long)(which_clock),        \
                                               (long)(tx))
-#define __sanitizer_syscall_pre_clock_getres(which_clock, tp) \
+#define __sanitizer_syscall_pre_clock_getres(which_clock, tp)                  \
   __sanitizer_syscall_pre_impl_clock_getres((long)(which_clock), (long)(tp))
-#define __sanitizer_syscall_post_clock_getres(res, which_clock, tp)    \
-  __sanitizer_syscall_post_impl_clock_getres(res, (long)(which_clock), \
+#define __sanitizer_syscall_post_clock_getres(res, which_clock, tp)            \
+  __sanitizer_syscall_post_impl_clock_getres(res, (long)(which_clock),         \
                                              (long)(tp))
-#define __sanitizer_syscall_pre_clock_nanosleep(which_clock, flags, rqtp, \
-                                                rmtp)                     \
-  __sanitizer_syscall_pre_impl_clock_nanosleep(                           \
+#define __sanitizer_syscall_pre_clock_nanosleep(which_clock, flags, rqtp,      \
+                                                rmtp)                          \
+  __sanitizer_syscall_pre_impl_clock_nanosleep(                                \
       (long)(which_clock), (long)(flags), (long)(rqtp), (long)(rmtp))
-#define __sanitizer_syscall_post_clock_nanosleep(res, which_clock, flags, \
-                                                 rqtp, rmtp)              \
-  __sanitizer_syscall_post_impl_clock_nanosleep(                          \
+#define __sanitizer_syscall_post_clock_nanosleep(res, which_clock, flags,      \
+                                                 rqtp, rmtp)                   \
+  __sanitizer_syscall_post_impl_clock_nanosleep(                               \
       res, (long)(which_clock), (long)(flags), (long)(rqtp), (long)(rmtp))
-#define __sanitizer_syscall_pre_nice(increment) \
+#define __sanitizer_syscall_pre_nice(increment)                                \
   __sanitizer_syscall_pre_impl_nice((long)(increment))
-#define __sanitizer_syscall_post_nice(res, increment) \
+#define __sanitizer_syscall_post_nice(res, increment)                          \
   __sanitizer_syscall_post_impl_nice(res, (long)(increment))
 #define __sanitizer_syscall_pre_sched_setscheduler(pid, policy, param)         \
   __sanitizer_syscall_pre_impl_sched_setscheduler((long)(pid), (long)(policy), \
                                                   (long)(param))
-#define __sanitizer_syscall_post_sched_setscheduler(res, pid, policy, param) \
-  __sanitizer_syscall_post_impl_sched_setscheduler(                          \
+#define __sanitizer_syscall_post_sched_setscheduler(res, pid, policy, param)   \
+  __sanitizer_syscall_post_impl_sched_setscheduler(                            \
       res, (long)(pid), (long)(policy), (long)(param))
-#define __sanitizer_syscall_pre_sched_setparam(pid, param) \
+#define __sanitizer_syscall_pre_sched_setparam(pid, param)                     \
   __sanitizer_syscall_pre_impl_sched_setparam((long)(pid), (long)(param))
-#define __sanitizer_syscall_post_sched_setparam(res, pid, param) \
+#define __sanitizer_syscall_post_sched_setparam(res, pid, param)               \
   __sanitizer_syscall_post_impl_sched_setparam(res, (long)(pid), (long)(param))
-#define __sanitizer_syscall_pre_sched_getscheduler(pid) \
+#define __sanitizer_syscall_pre_sched_getscheduler(pid)                        \
   __sanitizer_syscall_pre_impl_sched_getscheduler((long)(pid))
-#define __sanitizer_syscall_post_sched_getscheduler(res, pid) \
+#define __sanitizer_syscall_post_sched_getscheduler(res, pid)                  \
   __sanitizer_syscall_post_impl_sched_getscheduler(res, (long)(pid))
-#define __sanitizer_syscall_pre_sched_getparam(pid, param) \
+#define __sanitizer_syscall_pre_sched_getparam(pid, param)                     \
   __sanitizer_syscall_pre_impl_sched_getparam((long)(pid), (long)(param))
-#define __sanitizer_syscall_post_sched_getparam(res, pid, param) \
+#define __sanitizer_syscall_post_sched_getparam(res, pid, param)               \
   __sanitizer_syscall_post_impl_sched_getparam(res, (long)(pid), (long)(param))
-#define __sanitizer_syscall_pre_sched_setaffinity(pid, len, user_mask_ptr) \
-  __sanitizer_syscall_pre_impl_sched_setaffinity((long)(pid), (long)(len), \
+#define __sanitizer_syscall_pre_sched_setaffinity(pid, len, user_mask_ptr)     \
+  __sanitizer_syscall_pre_impl_sched_setaffinity((long)(pid), (long)(len),     \
                                                  (long)(user_mask_ptr))
-#define __sanitizer_syscall_post_sched_setaffinity(res, pid, len, \
-                                                   user_mask_ptr) \
-  __sanitizer_syscall_post_impl_sched_setaffinity(                \
+#define __sanitizer_syscall_post_sched_setaffinity(res, pid, len,              \
+                                                   user_mask_ptr)              \
+  __sanitizer_syscall_post_impl_sched_setaffinity(                             \
       res, (long)(pid), (long)(len), (long)(user_mask_ptr))
-#define __sanitizer_syscall_pre_sched_getaffinity(pid, len, user_mask_ptr) \
-  __sanitizer_syscall_pre_impl_sched_getaffinity((long)(pid), (long)(len), \
+#define __sanitizer_syscall_pre_sched_getaffinity(pid, len, user_mask_ptr)     \
+  __sanitizer_syscall_pre_impl_sched_getaffinity((long)(pid), (long)(len),     \
                                                  (long)(user_mask_ptr))
-#define __sanitizer_syscall_post_sched_getaffinity(res, pid, len, \
-                                                   user_mask_ptr) \
-  __sanitizer_syscall_post_impl_sched_getaffinity(                \
+#define __sanitizer_syscall_post_sched_getaffinity(res, pid, len,              \
+                                                   user_mask_ptr)              \
+  __sanitizer_syscall_post_impl_sched_getaffinity(                             \
       res, (long)(pid), (long)(len), (long)(user_mask_ptr))
-#define __sanitizer_syscall_pre_sched_yield() \
+#define __sanitizer_syscall_pre_sched_yield()                                  \
   __sanitizer_syscall_pre_impl_sched_yield()
-#define __sanitizer_syscall_post_sched_yield(res) \
+#define __sanitizer_syscall_post_sched_yield(res)                              \
   __sanitizer_syscall_post_impl_sched_yield(res)
-#define __sanitizer_syscall_pre_sched_get_priority_max(policy) \
+#define __sanitizer_syscall_pre_sched_get_priority_max(policy)                 \
   __sanitizer_syscall_pre_impl_sched_get_priority_max((long)(policy))
-#define __sanitizer_syscall_post_sched_get_priority_max(res, policy) \
+#define __sanitizer_syscall_post_sched_get_priority_max(res, policy)           \
   __sanitizer_syscall_post_impl_sched_get_priority_max(res, (long)(policy))
-#define __sanitizer_syscall_pre_sched_get_priority_min(policy) \
+#define __sanitizer_syscall_pre_sched_get_priority_min(policy)                 \
   __sanitizer_syscall_pre_impl_sched_get_priority_min((long)(policy))
-#define __sanitizer_syscall_post_sched_get_priority_min(res, policy) \
+#define __sanitizer_syscall_post_sched_get_priority_min(res, policy)           \
   __sanitizer_syscall_post_impl_sched_get_priority_min(res, (long)(policy))
-#define __sanitizer_syscall_pre_sched_rr_get_interval(pid, interval) \
-  __sanitizer_syscall_pre_impl_sched_rr_get_interval((long)(pid),    \
+#define __sanitizer_syscall_pre_sched_rr_get_interval(pid, interval)           \
+  __sanitizer_syscall_pre_impl_sched_rr_get_interval((long)(pid),              \
                                                      (long)(interval))
-#define __sanitizer_syscall_post_sched_rr_get_interval(res, pid, interval) \
-  __sanitizer_syscall_post_impl_sched_rr_get_interval(res, (long)(pid),    \
+#define __sanitizer_syscall_post_sched_rr_get_interval(res, pid, interval)     \
+  __sanitizer_syscall_post_impl_sched_rr_get_interval(res, (long)(pid),        \
                                                       (long)(interval))
-#define __sanitizer_syscall_pre_setpriority(which, who, niceval)       \
-  __sanitizer_syscall_pre_impl_setpriority((long)(which), (long)(who), \
+#define __sanitizer_syscall_pre_setpriority(which, who, niceval)               \
+  __sanitizer_syscall_pre_impl_setpriority((long)(which), (long)(who),         \
                                            (long)(niceval))
-#define __sanitizer_syscall_post_setpriority(res, which, who, niceval)       \
-  __sanitizer_syscall_post_impl_setpriority(res, (long)(which), (long)(who), \
+#define __sanitizer_syscall_post_setpriority(res, which, who, niceval)         \
+  __sanitizer_syscall_post_impl_setpriority(res, (long)(which), (long)(who),   \
                                             (long)(niceval))
-#define __sanitizer_syscall_pre_getpriority(which, who) \
+#define __sanitizer_syscall_pre_getpriority(which, who)                        \
   __sanitizer_syscall_pre_impl_getpriority((long)(which), (long)(who))
-#define __sanitizer_syscall_post_getpriority(res, which, who) \
+#define __sanitizer_syscall_post_getpriority(res, which, who)                  \
   __sanitizer_syscall_post_impl_getpriority(res, (long)(which), (long)(who))
-#define __sanitizer_syscall_pre_shutdown(arg0, arg1) \
+#define __sanitizer_syscall_pre_shutdown(arg0, arg1)                           \
   __sanitizer_syscall_pre_impl_shutdown((long)(arg0), (long)(arg1))
-#define __sanitizer_syscall_post_shutdown(res, arg0, arg1) \
+#define __sanitizer_syscall_post_shutdown(res, arg0, arg1)                     \
   __sanitizer_syscall_post_impl_shutdown(res, (long)(arg0), (long)(arg1))
-#define __sanitizer_syscall_pre_reboot(magic1, magic2, cmd, arg)      \
-  __sanitizer_syscall_pre_impl_reboot((long)(magic1), (long)(magic2), \
+#define __sanitizer_syscall_pre_reboot(magic1, magic2, cmd, arg)               \
+  __sanitizer_syscall_pre_impl_reboot((long)(magic1), (long)(magic2),          \
                                       (long)(cmd), (long)(arg))
-#define __sanitizer_syscall_post_reboot(res, magic1, magic2, cmd, arg)      \
-  __sanitizer_syscall_post_impl_reboot(res, (long)(magic1), (long)(magic2), \
+#define __sanitizer_syscall_post_reboot(res, magic1, magic2, cmd, arg)         \
+  __sanitizer_syscall_post_impl_reboot(res, (long)(magic1), (long)(magic2),    \
                                        (long)(cmd), (long)(arg))
-#define __sanitizer_syscall_pre_restart_syscall() \
+#define __sanitizer_syscall_pre_restart_syscall()                              \
   __sanitizer_syscall_pre_impl_restart_syscall()
-#define __sanitizer_syscall_post_restart_syscall(res) \
+#define __sanitizer_syscall_post_restart_syscall(res)                          \
   __sanitizer_syscall_post_impl_restart_syscall(res)
-#define __sanitizer_syscall_pre_kexec_load(entry, nr_segments, segments,      \
-                                           flags)                             \
-  __sanitizer_syscall_pre_impl_kexec_load((long)(entry), (long)(nr_segments), \
+#define __sanitizer_syscall_pre_kexec_load(entry, nr_segments, segments,       \
+                                           flags)                              \
+  __sanitizer_syscall_pre_impl_kexec_load((long)(entry), (long)(nr_segments),  \
                                           (long)(segments), (long)(flags))
 #define __sanitizer_syscall_post_kexec_load(res, entry, nr_segments, segments, \
                                             flags)                             \
   __sanitizer_syscall_post_impl_kexec_load(res, (long)(entry),                 \
                                            (long)(nr_segments),                \
                                            (long)(segments), (long)(flags))
-#define __sanitizer_syscall_pre_exit(error_code) \
+#define __sanitizer_syscall_pre_exit(error_code)                               \
   __sanitizer_syscall_pre_impl_exit((long)(error_code))
-#define __sanitizer_syscall_post_exit(res, error_code) \
+#define __sanitizer_syscall_post_exit(res, error_code)                         \
   __sanitizer_syscall_post_impl_exit(res, (long)(error_code))
-#define __sanitizer_syscall_pre_exit_group(error_code) \
+#define __sanitizer_syscall_pre_exit_group(error_code)                         \
   __sanitizer_syscall_pre_impl_exit_group((long)(error_code))
-#define __sanitizer_syscall_post_exit_group(res, error_code) \
+#define __sanitizer_syscall_post_exit_group(res, error_code)                   \
   __sanitizer_syscall_post_impl_exit_group(res, (long)(error_code))
-#define __sanitizer_syscall_pre_wait4(pid, stat_addr, options, ru)   \
-  __sanitizer_syscall_pre_impl_wait4((long)(pid), (long)(stat_addr), \
+#define __sanitizer_syscall_pre_wait4(pid, stat_addr, options, ru)             \
+  __sanitizer_syscall_pre_impl_wait4((long)(pid), (long)(stat_addr),           \
                                      (long)(options), (long)(ru))
-#define __sanitizer_syscall_post_wait4(res, pid, stat_addr, options, ru)   \
-  __sanitizer_syscall_post_impl_wait4(res, (long)(pid), (long)(stat_addr), \
+#define __sanitizer_syscall_post_wait4(res, pid, stat_addr, options, ru)       \
+  __sanitizer_syscall_post_impl_wait4(res, (long)(pid), (long)(stat_addr),     \
                                       (long)(options), (long)(ru))
-#define __sanitizer_syscall_pre_waitid(which, pid, infop, options, ru) \
-  __sanitizer_syscall_pre_impl_waitid(                                 \
+#define __sanitizer_syscall_pre_waitid(which, pid, infop, options, ru)         \
+  __sanitizer_syscall_pre_impl_waitid(                                         \
       (long)(which), (long)(pid), (long)(infop), (long)(options), (long)(ru))
-#define __sanitizer_syscall_post_waitid(res, which, pid, infop, options, ru) \
-  __sanitizer_syscall_post_impl_waitid(res, (long)(which), (long)(pid),      \
-                                       (long)(infop), (long)(options),       \
+#define __sanitizer_syscall_post_waitid(res, which, pid, infop, options, ru)   \
+  __sanitizer_syscall_post_impl_waitid(res, (long)(which), (long)(pid),        \
+                                       (long)(infop), (long)(options),         \
                                        (long)(ru))
-#define __sanitizer_syscall_pre_waitpid(pid, stat_addr, options)       \
-  __sanitizer_syscall_pre_impl_waitpid((long)(pid), (long)(stat_addr), \
+#define __sanitizer_syscall_pre_waitpid(pid, stat_addr, options)               \
+  __sanitizer_syscall_pre_impl_waitpid((long)(pid), (long)(stat_addr),         \
                                        (long)(options))
-#define __sanitizer_syscall_post_waitpid(res, pid, stat_addr, options)       \
-  __sanitizer_syscall_post_impl_waitpid(res, (long)(pid), (long)(stat_addr), \
+#define __sanitizer_syscall_post_waitpid(res, pid, stat_addr, options)         \
+  __sanitizer_syscall_post_impl_waitpid(res, (long)(pid), (long)(stat_addr),   \
                                         (long)(options))
-#define __sanitizer_syscall_pre_set_tid_address(tidptr) \
+#define __sanitizer_syscall_pre_set_tid_address(tidptr)                        \
   __sanitizer_syscall_pre_impl_set_tid_address((long)(tidptr))
-#define __sanitizer_syscall_post_set_tid_address(res, tidptr) \
+#define __sanitizer_syscall_post_set_tid_address(res, tidptr)                  \
   __sanitizer_syscall_post_impl_set_tid_address(res, (long)(tidptr))
-#define __sanitizer_syscall_pre_init_module(umod, len, uargs)         \
-  __sanitizer_syscall_pre_impl_init_module((long)(umod), (long)(len), \
+#define __sanitizer_syscall_pre_init_module(umod, len, uargs)                  \
+  __sanitizer_syscall_pre_impl_init_module((long)(umod), (long)(len),          \
                                            (long)(uargs))
-#define __sanitizer_syscall_post_init_module(res, umod, len, uargs)         \
-  __sanitizer_syscall_post_impl_init_module(res, (long)(umod), (long)(len), \
+#define __sanitizer_syscall_post_init_module(res, umod, len, uargs)            \
+  __sanitizer_syscall_post_impl_init_module(res, (long)(umod), (long)(len),    \
                                             (long)(uargs))
-#define __sanitizer_syscall_pre_delete_module(name_user, flags) \
+#define __sanitizer_syscall_pre_delete_module(name_user, flags)                \
   __sanitizer_syscall_pre_impl_delete_module((long)(name_user), (long)(flags))
-#define __sanitizer_syscall_post_delete_module(res, name_user, flags) \
-  __sanitizer_syscall_post_impl_delete_module(res, (long)(name_user), \
+#define __sanitizer_syscall_post_delete_module(res, name_user, flags)          \
+  __sanitizer_syscall_post_impl_delete_module(res, (long)(name_user),          \
                                               (long)(flags))
-#define __sanitizer_syscall_pre_rt_sigprocmask(how, set, oset, sigsetsize) \
-  __sanitizer_syscall_pre_impl_rt_sigprocmask(                             \
+#define __sanitizer_syscall_pre_rt_sigprocmask(how, set, oset, sigsetsize)     \
+  __sanitizer_syscall_pre_impl_rt_sigprocmask(                                 \
       (long)(how), (long)(set), (long)(oset), (long)(sigsetsize))
-#define __sanitizer_syscall_post_rt_sigprocmask(res, how, set, oset, \
-                                                sigsetsize)          \
-  __sanitizer_syscall_post_impl_rt_sigprocmask(                      \
+#define __sanitizer_syscall_post_rt_sigprocmask(res, how, set, oset,           \
+                                                sigsetsize)                    \
+  __sanitizer_syscall_post_impl_rt_sigprocmask(                                \
       res, (long)(how), (long)(set), (long)(oset), (long)(sigsetsize))
-#define __sanitizer_syscall_pre_rt_sigpending(set, sigsetsize) \
+#define __sanitizer_syscall_pre_rt_sigpending(set, sigsetsize)                 \
   __sanitizer_syscall_pre_impl_rt_sigpending((long)(set), (long)(sigsetsize))
-#define __sanitizer_syscall_post_rt_sigpending(res, set, sigsetsize) \
-  __sanitizer_syscall_post_impl_rt_sigpending(res, (long)(set),      \
+#define __sanitizer_syscall_post_rt_sigpending(res, set, sigsetsize)           \
+  __sanitizer_syscall_post_impl_rt_sigpending(res, (long)(set),                \
                                               (long)(sigsetsize))
-#define __sanitizer_syscall_pre_rt_sigtimedwait(uthese, uinfo, uts, \
-                                                sigsetsize)         \
-  __sanitizer_syscall_pre_impl_rt_sigtimedwait(                     \
+#define __sanitizer_syscall_pre_rt_sigtimedwait(uthese, uinfo, uts,            \
+                                                sigsetsize)                    \
+  __sanitizer_syscall_pre_impl_rt_sigtimedwait(                                \
       (long)(uthese), (long)(uinfo), (long)(uts), (long)(sigsetsize))
-#define __sanitizer_syscall_post_rt_sigtimedwait(res, uthese, uinfo, uts, \
-                                                 sigsetsize)              \
-  __sanitizer_syscall_post_impl_rt_sigtimedwait(                          \
+#define __sanitizer_syscall_post_rt_sigtimedwait(res, uthese, uinfo, uts,      \
+                                                 sigsetsize)                   \
+  __sanitizer_syscall_post_impl_rt_sigtimedwait(                               \
       res, (long)(uthese), (long)(uinfo), (long)(uts), (long)(sigsetsize))
-#define __sanitizer_syscall_pre_rt_tgsigqueueinfo(tgid, pid, sig, uinfo)    \
-  __sanitizer_syscall_pre_impl_rt_tgsigqueueinfo((long)(tgid), (long)(pid), \
+#define __sanitizer_syscall_pre_rt_tgsigqueueinfo(tgid, pid, sig, uinfo)       \
+  __sanitizer_syscall_pre_impl_rt_tgsigqueueinfo((long)(tgid), (long)(pid),    \
                                                  (long)(sig), (long)(uinfo))
 #define __sanitizer_syscall_post_rt_tgsigqueueinfo(res, tgid, pid, sig, uinfo) \
   __sanitizer_syscall_post_impl_rt_tgsigqueueinfo(                             \
       res, (long)(tgid), (long)(pid), (long)(sig), (long)(uinfo))
-#define __sanitizer_syscall_pre_kill(pid, sig) \
+#define __sanitizer_syscall_pre_kill(pid, sig)                                 \
   __sanitizer_syscall_pre_impl_kill((long)(pid), (long)(sig))
-#define __sanitizer_syscall_post_kill(res, pid, sig) \
+#define __sanitizer_syscall_post_kill(res, pid, sig)                           \
   __sanitizer_syscall_post_impl_kill(res, (long)(pid), (long)(sig))
-#define __sanitizer_syscall_pre_tgkill(tgid, pid, sig) \
+#define __sanitizer_syscall_pre_tgkill(tgid, pid, sig)                         \
   __sanitizer_syscall_pre_impl_tgkill((long)(tgid), (long)(pid), (long)(sig))
-#define __sanitizer_syscall_post_tgkill(res, tgid, pid, sig)           \
-  __sanitizer_syscall_post_impl_tgkill(res, (long)(tgid), (long)(pid), \
+#define __sanitizer_syscall_post_tgkill(res, tgid, pid, sig)                   \
+  __sanitizer_syscall_post_impl_tgkill(res, (long)(tgid), (long)(pid),         \
                                        (long)(sig))
-#define __sanitizer_syscall_pre_tkill(pid, sig) \
+#define __sanitizer_syscall_pre_tkill(pid, sig)                                \
   __sanitizer_syscall_pre_impl_tkill((long)(pid), (long)(sig))
-#define __sanitizer_syscall_post_tkill(res, pid, sig) \
+#define __sanitizer_syscall_post_tkill(res, pid, sig)                          \
   __sanitizer_syscall_post_impl_tkill(res, (long)(pid), (long)(sig))
-#define __sanitizer_syscall_pre_rt_sigqueueinfo(pid, sig, uinfo)         \
-  __sanitizer_syscall_pre_impl_rt_sigqueueinfo((long)(pid), (long)(sig), \
+#define __sanitizer_syscall_pre_rt_sigqueueinfo(pid, sig, uinfo)               \
+  __sanitizer_syscall_pre_impl_rt_sigqueueinfo((long)(pid), (long)(sig),       \
                                                (long)(uinfo))
 #define __sanitizer_syscall_post_rt_sigqueueinfo(res, pid, sig, uinfo)         \
   __sanitizer_syscall_post_impl_rt_sigqueueinfo(res, (long)(pid), (long)(sig), \
                                                 (long)(uinfo))
-#define __sanitizer_syscall_pre_sgetmask() \
+#define __sanitizer_syscall_pre_sgetmask()                                     \
   __sanitizer_syscall_pre_impl_sgetmask()
-#define __sanitizer_syscall_post_sgetmask(res) \
+#define __sanitizer_syscall_post_sgetmask(res)                                 \
   __sanitizer_syscall_post_impl_sgetmask(res)
-#define __sanitizer_syscall_pre_ssetmask(newmask) \
+#define __sanitizer_syscall_pre_ssetmask(newmask)                              \
   __sanitizer_syscall_pre_impl_ssetmask((long)(newmask))
-#define __sanitizer_syscall_post_ssetmask(res, newmask) \
+#define __sanitizer_syscall_post_ssetmask(res, newmask)                        \
   __sanitizer_syscall_post_impl_ssetmask(res, (long)(newmask))
-#define __sanitizer_syscall_pre_signal(sig, handler) \
+#define __sanitizer_syscall_pre_signal(sig, handler)                           \
   __sanitizer_syscall_pre_impl_signal((long)(sig), (long)(handler))
-#define __sanitizer_syscall_post_signal(res, sig, handler) \
+#define __sanitizer_syscall_post_signal(res, sig, handler)                     \
   __sanitizer_syscall_post_impl_signal(res, (long)(sig), (long)(handler))
 #define __sanitizer_syscall_pre_pause() __sanitizer_syscall_pre_impl_pause()
-#define __sanitizer_syscall_post_pause(res) \
+#define __sanitizer_syscall_post_pause(res)                                    \
   __sanitizer_syscall_post_impl_pause(res)
 #define __sanitizer_syscall_pre_sync() __sanitizer_syscall_pre_impl_sync()
-#define __sanitizer_syscall_post_sync(res) \
+#define __sanitizer_syscall_post_sync(res)                                     \
   __sanitizer_syscall_post_impl_sync(res)
-#define __sanitizer_syscall_pre_fsync(fd) \
+#define __sanitizer_syscall_pre_fsync(fd)                                      \
   __sanitizer_syscall_pre_impl_fsync((long)(fd))
-#define __sanitizer_syscall_post_fsync(res, fd) \
+#define __sanitizer_syscall_post_fsync(res, fd)                                \
   __sanitizer_syscall_post_impl_fsync(res, (long)(fd))
-#define __sanitizer_syscall_pre_fdatasync(fd) \
+#define __sanitizer_syscall_pre_fdatasync(fd)                                  \
   __sanitizer_syscall_pre_impl_fdatasync((long)(fd))
-#define __sanitizer_syscall_post_fdatasync(res, fd) \
+#define __sanitizer_syscall_post_fdatasync(res, fd)                            \
   __sanitizer_syscall_post_impl_fdatasync(res, (long)(fd))
-#define __sanitizer_syscall_pre_bdflush(func, data) \
+#define __sanitizer_syscall_pre_bdflush(func, data)                            \
   __sanitizer_syscall_pre_impl_bdflush((long)(func), (long)(data))
-#define __sanitizer_syscall_post_bdflush(res, func, data) \
+#define __sanitizer_syscall_post_bdflush(res, func, data)                      \
   __sanitizer_syscall_post_impl_bdflush(res, (long)(func), (long)(data))
-#define __sanitizer_syscall_pre_mount(dev_name, dir_name, type, flags, data) \
-  __sanitizer_syscall_pre_impl_mount((long)(dev_name), (long)(dir_name),     \
-                                     (long)(type), (long)(flags),            \
+#define __sanitizer_syscall_pre_mount(dev_name, dir_name, type, flags, data)   \
+  __sanitizer_syscall_pre_impl_mount((long)(dev_name), (long)(dir_name),       \
+                                     (long)(type), (long)(flags),              \
                                      (long)(data))
 #define __sanitizer_syscall_post_mount(res, dev_name, dir_name, type, flags,   \
                                        data)                                   \
   __sanitizer_syscall_post_impl_mount(res, (long)(dev_name), (long)(dir_name), \
                                       (long)(type), (long)(flags),             \
                                       (long)(data))
-#define __sanitizer_syscall_pre_umount(name, flags) \
+#define __sanitizer_syscall_pre_umount(name, flags)                            \
   __sanitizer_syscall_pre_impl_umount((long)(name), (long)(flags))
-#define __sanitizer_syscall_post_umount(res, name, flags) \
+#define __sanitizer_syscall_post_umount(res, name, flags)                      \
   __sanitizer_syscall_post_impl_umount(res, (long)(name), (long)(flags))
-#define __sanitizer_syscall_pre_oldumount(name) \
+#define __sanitizer_syscall_pre_oldumount(name)                                \
   __sanitizer_syscall_pre_impl_oldumount((long)(name))
-#define __sanitizer_syscall_post_oldumount(res, name) \
+#define __sanitizer_syscall_post_oldumount(res, name)                          \
   __sanitizer_syscall_post_impl_oldumount(res, (long)(name))
-#define __sanitizer_syscall_pre_truncate(path, length) \
+#define __sanitizer_syscall_pre_truncate(path, length)                         \
   __sanitizer_syscall_pre_impl_truncate((long)(path), (long)(length))
-#define __sanitizer_syscall_post_truncate(res, path, length) \
+#define __sanitizer_syscall_post_truncate(res, path, length)                   \
   __sanitizer_syscall_post_impl_truncate(res, (long)(path), (long)(length))
-#define __sanitizer_syscall_pre_ftruncate(fd, length) \
+#define __sanitizer_syscall_pre_ftruncate(fd, length)                          \
   __sanitizer_syscall_pre_impl_ftruncate((long)(fd), (long)(length))
-#define __sanitizer_syscall_post_ftruncate(res, fd, length) \
+#define __sanitizer_syscall_post_ftruncate(res, fd, length)                    \
   __sanitizer_syscall_post_impl_ftruncate(res, (long)(fd), (long)(length))
-#define __sanitizer_syscall_pre_stat(filename, statbuf) \
+#define __sanitizer_syscall_pre_stat(filename, statbuf)                        \
   __sanitizer_syscall_pre_impl_stat((long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_post_stat(res, filename, statbuf) \
+#define __sanitizer_syscall_post_stat(res, filename, statbuf)                  \
   __sanitizer_syscall_post_impl_stat(res, (long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_pre_statfs(path, buf) \
+#define __sanitizer_syscall_pre_statfs(path, buf)                              \
   __sanitizer_syscall_pre_impl_statfs((long)(path), (long)(buf))
-#define __sanitizer_syscall_post_statfs(res, path, buf) \
+#define __sanitizer_syscall_post_statfs(res, path, buf)                        \
   __sanitizer_syscall_post_impl_statfs(res, (long)(path), (long)(buf))
-#define __sanitizer_syscall_pre_statfs64(path, sz, buf) \
+#define __sanitizer_syscall_pre_statfs64(path, sz, buf)                        \
   __sanitizer_syscall_pre_impl_statfs64((long)(path), (long)(sz), (long)(buf))
-#define __sanitizer_syscall_post_statfs64(res, path, sz, buf)           \
-  __sanitizer_syscall_post_impl_statfs64(res, (long)(path), (long)(sz), \
+#define __sanitizer_syscall_post_statfs64(res, path, sz, buf)                  \
+  __sanitizer_syscall_post_impl_statfs64(res, (long)(path), (long)(sz),        \
                                          (long)(buf))
-#define __sanitizer_syscall_pre_fstatfs(fd, buf) \
+#define __sanitizer_syscall_pre_fstatfs(fd, buf)                               \
   __sanitizer_syscall_pre_impl_fstatfs((long)(fd), (long)(buf))
-#define __sanitizer_syscall_post_fstatfs(res, fd, buf) \
+#define __sanitizer_syscall_post_fstatfs(res, fd, buf)                         \
   __sanitizer_syscall_post_impl_fstatfs(res, (long)(fd), (long)(buf))
-#define __sanitizer_syscall_pre_fstatfs64(fd, sz, buf) \
+#define __sanitizer_syscall_pre_fstatfs64(fd, sz, buf)                         \
   __sanitizer_syscall_pre_impl_fstatfs64((long)(fd), (long)(sz), (long)(buf))
-#define __sanitizer_syscall_post_fstatfs64(res, fd, sz, buf)           \
-  __sanitizer_syscall_post_impl_fstatfs64(res, (long)(fd), (long)(sz), \
+#define __sanitizer_syscall_post_fstatfs64(res, fd, sz, buf)                   \
+  __sanitizer_syscall_post_impl_fstatfs64(res, (long)(fd), (long)(sz),         \
                                           (long)(buf))
-#define __sanitizer_syscall_pre_lstat(filename, statbuf) \
+#define __sanitizer_syscall_pre_lstat(filename, statbuf)                       \
   __sanitizer_syscall_pre_impl_lstat((long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_post_lstat(res, filename, statbuf) \
+#define __sanitizer_syscall_post_lstat(res, filename, statbuf)                 \
   __sanitizer_syscall_post_impl_lstat(res, (long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_pre_fstat(fd, statbuf) \
+#define __sanitizer_syscall_pre_fstat(fd, statbuf)                             \
   __sanitizer_syscall_pre_impl_fstat((long)(fd), (long)(statbuf))
-#define __sanitizer_syscall_post_fstat(res, fd, statbuf) \
+#define __sanitizer_syscall_post_fstat(res, fd, statbuf)                       \
   __sanitizer_syscall_post_impl_fstat(res, (long)(fd), (long)(statbuf))
-#define __sanitizer_syscall_pre_newstat(filename, statbuf) \
+#define __sanitizer_syscall_pre_newstat(filename, statbuf)                     \
   __sanitizer_syscall_pre_impl_newstat((long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_post_newstat(res, filename, statbuf) \
+#define __sanitizer_syscall_post_newstat(res, filename, statbuf)               \
   __sanitizer_syscall_post_impl_newstat(res, (long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_pre_newlstat(filename, statbuf) \
+#define __sanitizer_syscall_pre_newlstat(filename, statbuf)                    \
   __sanitizer_syscall_pre_impl_newlstat((long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_post_newlstat(res, filename, statbuf) \
+#define __sanitizer_syscall_post_newlstat(res, filename, statbuf)              \
   __sanitizer_syscall_post_impl_newlstat(res, (long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_pre_newfstat(fd, statbuf) \
+#define __sanitizer_syscall_pre_newfstat(fd, statbuf)                          \
   __sanitizer_syscall_pre_impl_newfstat((long)(fd), (long)(statbuf))
-#define __sanitizer_syscall_post_newfstat(res, fd, statbuf) \
+#define __sanitizer_syscall_post_newfstat(res, fd, statbuf)                    \
   __sanitizer_syscall_post_impl_newfstat(res, (long)(fd), (long)(statbuf))
-#define __sanitizer_syscall_pre_ustat(dev, ubuf) \
+#define __sanitizer_syscall_pre_ustat(dev, ubuf)                               \
   __sanitizer_syscall_pre_impl_ustat((long)(dev), (long)(ubuf))
-#define __sanitizer_syscall_post_ustat(res, dev, ubuf) \
+#define __sanitizer_syscall_post_ustat(res, dev, ubuf)                         \
   __sanitizer_syscall_post_impl_ustat(res, (long)(dev), (long)(ubuf))
-#define __sanitizer_syscall_pre_stat64(filename, statbuf) \
+#define __sanitizer_syscall_pre_stat64(filename, statbuf)                      \
   __sanitizer_syscall_pre_impl_stat64((long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_post_stat64(res, filename, statbuf) \
+#define __sanitizer_syscall_post_stat64(res, filename, statbuf)                \
   __sanitizer_syscall_post_impl_stat64(res, (long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_pre_fstat64(fd, statbuf) \
+#define __sanitizer_syscall_pre_fstat64(fd, statbuf)                           \
   __sanitizer_syscall_pre_impl_fstat64((long)(fd), (long)(statbuf))
-#define __sanitizer_syscall_post_fstat64(res, fd, statbuf) \
+#define __sanitizer_syscall_post_fstat64(res, fd, statbuf)                     \
   __sanitizer_syscall_post_impl_fstat64(res, (long)(fd), (long)(statbuf))
-#define __sanitizer_syscall_pre_lstat64(filename, statbuf) \
+#define __sanitizer_syscall_pre_lstat64(filename, statbuf)                     \
   __sanitizer_syscall_pre_impl_lstat64((long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_post_lstat64(res, filename, statbuf) \
+#define __sanitizer_syscall_post_lstat64(res, filename, statbuf)               \
   __sanitizer_syscall_post_impl_lstat64(res, (long)(filename), (long)(statbuf))
-#define __sanitizer_syscall_pre_setxattr(path, name, value, size, flags) \
-  __sanitizer_syscall_pre_impl_setxattr(                                 \
+#define __sanitizer_syscall_pre_setxattr(path, name, value, size, flags)       \
+  __sanitizer_syscall_pre_impl_setxattr(                                       \
       (long)(path), (long)(name), (long)(value), (long)(size), (long)(flags))
 #define __sanitizer_syscall_post_setxattr(res, path, name, value, size, flags) \
   __sanitizer_syscall_post_impl_setxattr(res, (long)(path), (long)(name),      \
                                          (long)(value), (long)(size),          \
                                          (long)(flags))
-#define __sanitizer_syscall_pre_lsetxattr(path, name, value, size, flags) \
-  __sanitizer_syscall_pre_impl_lsetxattr(                                 \
+#define __sanitizer_syscall_pre_lsetxattr(path, name, value, size, flags)      \
+  __sanitizer_syscall_pre_impl_lsetxattr(                                      \
       (long)(path), (long)(name), (long)(value), (long)(size), (long)(flags))
-#define __sanitizer_syscall_post_lsetxattr(res, path, name, value, size,   \
-                                           flags)                          \
-  __sanitizer_syscall_post_impl_lsetxattr(res, (long)(path), (long)(name), \
-                                          (long)(value), (long)(size),     \
+#define __sanitizer_syscall_post_lsetxattr(res, path, name, value, size,       \
+                                           flags)                              \
+  __sanitizer_syscall_post_impl_lsetxattr(res, (long)(path), (long)(name),     \
+                                          (long)(value), (long)(size),         \
                                           (long)(flags))
-#define __sanitizer_syscall_pre_fsetxattr(fd, name, value, size, flags) \
-  __sanitizer_syscall_pre_impl_fsetxattr(                               \
+#define __sanitizer_syscall_pre_fsetxattr(fd, name, value, size, flags)        \
+  __sanitizer_syscall_pre_impl_fsetxattr(                                      \
       (long)(fd), (long)(name), (long)(value), (long)(size), (long)(flags))
-#define __sanitizer_syscall_post_fsetxattr(res, fd, name, value, size, flags) \
-  __sanitizer_syscall_post_impl_fsetxattr(res, (long)(fd), (long)(name),      \
-                                          (long)(value), (long)(size),        \
+#define __sanitizer_syscall_post_fsetxattr(res, fd, name, value, size, flags)  \
+  __sanitizer_syscall_post_impl_fsetxattr(res, (long)(fd), (long)(name),       \
+                                          (long)(value), (long)(size),         \
                                           (long)(flags))
-#define __sanitizer_syscall_pre_getxattr(path, name, value, size)   \
-  __sanitizer_syscall_pre_impl_getxattr((long)(path), (long)(name), \
+#define __sanitizer_syscall_pre_getxattr(path, name, value, size)              \
+  __sanitizer_syscall_pre_impl_getxattr((long)(path), (long)(name),            \
                                         (long)(value), (long)(size))
-#define __sanitizer_syscall_post_getxattr(res, path, name, value, size)   \
-  __sanitizer_syscall_post_impl_getxattr(res, (long)(path), (long)(name), \
+#define __sanitizer_syscall_post_getxattr(res, path, name, value, size)        \
+  __sanitizer_syscall_post_impl_getxattr(res, (long)(path), (long)(name),      \
                                          (long)(value), (long)(size))
-#define __sanitizer_syscall_pre_lgetxattr(path, name, value, size)   \
-  __sanitizer_syscall_pre_impl_lgetxattr((long)(path), (long)(name), \
+#define __sanitizer_syscall_pre_lgetxattr(path, name, value, size)             \
+  __sanitizer_syscall_pre_impl_lgetxattr((long)(path), (long)(name),           \
                                          (long)(value), (long)(size))
-#define __sanitizer_syscall_post_lgetxattr(res, path, name, value, size)   \
-  __sanitizer_syscall_post_impl_lgetxattr(res, (long)(path), (long)(name), \
+#define __sanitizer_syscall_post_lgetxattr(res, path, name, value, size)       \
+  __sanitizer_syscall_post_impl_lgetxattr(res, (long)(path), (long)(name),     \
                                           (long)(value), (long)(size))
-#define __sanitizer_syscall_pre_fgetxattr(fd, name, value, size)   \
-  __sanitizer_syscall_pre_impl_fgetxattr((long)(fd), (long)(name), \
+#define __sanitizer_syscall_pre_fgetxattr(fd, name, value, size)               \
+  __sanitizer_syscall_pre_impl_fgetxattr((long)(fd), (long)(name),             \
                                          (long)(value), (long)(size))
-#define __sanitizer_syscall_post_fgetxattr(res, fd, name, value, size)   \
-  __sanitizer_syscall_post_impl_fgetxattr(res, (long)(fd), (long)(name), \
+#define __sanitizer_syscall_post_fgetxattr(res, fd, name, value, size)         \
+  __sanitizer_syscall_post_impl_fgetxattr(res, (long)(fd), (long)(name),       \
                                           (long)(value), (long)(size))
-#define __sanitizer_syscall_pre_listxattr(path, list, size)          \
-  __sanitizer_syscall_pre_impl_listxattr((long)(path), (long)(list), \
+#define __sanitizer_syscall_pre_listxattr(path, list, size)                    \
+  __sanitizer_syscall_pre_impl_listxattr((long)(path), (long)(list),           \
                                          (long)(size))
-#define __sanitizer_syscall_post_listxattr(res, path, list, size)          \
-  __sanitizer_syscall_post_impl_listxattr(res, (long)(path), (long)(list), \
+#define __sanitizer_syscall_post_listxattr(res, path, list, size)              \
+  __sanitizer_syscall_post_impl_listxattr(res, (long)(path), (long)(list),     \
                                           (long)(size))
-#define __sanitizer_syscall_pre_llistxattr(path, list, size)          \
-  __sanitizer_syscall_pre_impl_llistxattr((long)(path), (long)(list), \
+#define __sanitizer_syscall_pre_llistxattr(path, list, size)                   \
+  __sanitizer_syscall_pre_impl_llistxattr((long)(path), (long)(list),          \
                                           (long)(size))
-#define __sanitizer_syscall_post_llistxattr(res, path, list, size)          \
-  __sanitizer_syscall_post_impl_llistxattr(res, (long)(path), (long)(list), \
+#define __sanitizer_syscall_post_llistxattr(res, path, list, size)             \
+  __sanitizer_syscall_post_impl_llistxattr(res, (long)(path), (long)(list),    \
                                            (long)(size))
-#define __sanitizer_syscall_pre_flistxattr(fd, list, size)          \
-  __sanitizer_syscall_pre_impl_flistxattr((long)(fd), (long)(list), \
+#define __sanitizer_syscall_pre_flistxattr(fd, list, size)                     \
+  __sanitizer_syscall_pre_impl_flistxattr((long)(fd), (long)(list),            \
                                           (long)(size))
-#define __sanitizer_syscall_post_flistxattr(res, fd, list, size)          \
-  __sanitizer_syscall_post_impl_flistxattr(res, (long)(fd), (long)(list), \
+#define __sanitizer_syscall_post_flistxattr(res, fd, list, size)               \
+  __sanitizer_syscall_post_impl_flistxattr(res, (long)(fd), (long)(list),      \
                                            (long)(size))
-#define __sanitizer_syscall_pre_removexattr(path, name) \
+#define __sanitizer_syscall_pre_removexattr(path, name)                        \
   __sanitizer_syscall_pre_impl_removexattr((long)(path), (long)(name))
-#define __sanitizer_syscall_post_removexattr(res, path, name) \
+#define __sanitizer_syscall_post_removexattr(res, path, name)                  \
   __sanitizer_syscall_post_impl_removexattr(res, (long)(path), (long)(name))
-#define __sanitizer_syscall_pre_lremovexattr(path, name) \
+#define __sanitizer_syscall_pre_lremovexattr(path, name)                       \
   __sanitizer_syscall_pre_impl_lremovexattr((long)(path), (long)(name))
-#define __sanitizer_syscall_post_lremovexattr(res, path, name) \
+#define __sanitizer_syscall_post_lremovexattr(res, path, name)                 \
   __sanitizer_syscall_post_impl_lremovexattr(res, (long)(path), (long)(name))
-#define __sanitizer_syscall_pre_fremovexattr(fd, name) \
+#define __sanitizer_syscall_pre_fremovexattr(fd, name)                         \
   __sanitizer_syscall_pre_impl_fremovexattr((long)(fd), (long)(name))
-#define __sanitizer_syscall_post_fremovexattr(res, fd, name) \
+#define __sanitizer_syscall_post_fremovexattr(res, fd, name)                   \
   __sanitizer_syscall_post_impl_fremovexattr(res, (long)(fd), (long)(name))
-#define __sanitizer_syscall_pre_brk(brk) \
+#define __sanitizer_syscall_pre_brk(brk)                                       \
   __sanitizer_syscall_pre_impl_brk((long)(brk))
-#define __sanitizer_syscall_post_brk(res, brk) \
+#define __sanitizer_syscall_post_brk(res, brk)                                 \
   __sanitizer_syscall_post_impl_brk(res, (long)(brk))
-#define __sanitizer_syscall_pre_mprotect(start, len, prot)          \
-  __sanitizer_syscall_pre_impl_mprotect((long)(start), (long)(len), \
+#define __sanitizer_syscall_pre_mprotect(start, len, prot)                     \
+  __sanitizer_syscall_pre_impl_mprotect((long)(start), (long)(len),            \
                                         (long)(prot))
-#define __sanitizer_syscall_post_mprotect(res, start, len, prot)          \
-  __sanitizer_syscall_post_impl_mprotect(res, (long)(start), (long)(len), \
+#define __sanitizer_syscall_post_mprotect(res, start, len, prot)               \
+  __sanitizer_syscall_post_impl_mprotect(res, (long)(start), (long)(len),      \
                                          (long)(prot))
-#define __sanitizer_syscall_pre_mremap(addr, old_len, new_len, flags, \
-                                       new_addr)                      \
-  __sanitizer_syscall_pre_impl_mremap((long)(addr), (long)(old_len),  \
-                                      (long)(new_len), (long)(flags), \
+#define __sanitizer_syscall_pre_mremap(addr, old_len, new_len, flags,          \
+                                       new_addr)                               \
+  __sanitizer_syscall_pre_impl_mremap((long)(addr), (long)(old_len),           \
+                                      (long)(new_len), (long)(flags),          \
                                       (long)(new_addr))
-#define __sanitizer_syscall_post_mremap(res, addr, old_len, new_len, flags, \
-                                        new_addr)                           \
-  __sanitizer_syscall_post_impl_mremap(res, (long)(addr), (long)(old_len),  \
-                                       (long)(new_len), (long)(flags),      \
+#define __sanitizer_syscall_post_mremap(res, addr, old_len, new_len, flags,    \
+                                        new_addr)                              \
+  __sanitizer_syscall_post_impl_mremap(res, (long)(addr), (long)(old_len),     \
+                                       (long)(new_len), (long)(flags),         \
                                        (long)(new_addr))
-#define __sanitizer_syscall_pre_remap_file_pages(start, size, prot, pgoff, \
-                                                 flags)                    \
-  __sanitizer_syscall_pre_impl_remap_file_pages(                           \
+#define __sanitizer_syscall_pre_remap_file_pages(start, size, prot, pgoff,     \
+                                                 flags)                        \
+  __sanitizer_syscall_pre_impl_remap_file_pages(                               \
       (long)(start), (long)(size), (long)(prot), (long)(pgoff), (long)(flags))
-#define __sanitizer_syscall_post_remap_file_pages(res, start, size, prot,    \
-                                                  pgoff, flags)              \
-  __sanitizer_syscall_post_impl_remap_file_pages(res, (long)(start),         \
-                                                 (long)(size), (long)(prot), \
+#define __sanitizer_syscall_post_remap_file_pages(res, start, size, prot,      \
+                                                  pgoff, flags)                \
+  __sanitizer_syscall_post_impl_remap_file_pages(res, (long)(start),           \
+                                                 (long)(size), (long)(prot),   \
                                                  (long)(pgoff), (long)(flags))
-#define __sanitizer_syscall_pre_msync(start, len, flags) \
+#define __sanitizer_syscall_pre_msync(start, len, flags)                       \
   __sanitizer_syscall_pre_impl_msync((long)(start), (long)(len), (long)(flags))
-#define __sanitizer_syscall_post_msync(res, start, len, flags)         \
-  __sanitizer_syscall_post_impl_msync(res, (long)(start), (long)(len), \
+#define __sanitizer_syscall_post_msync(res, start, len, flags)                 \
+  __sanitizer_syscall_post_impl_msync(res, (long)(start), (long)(len),         \
                                       (long)(flags))
-#define __sanitizer_syscall_pre_munmap(addr, len) \
+#define __sanitizer_syscall_pre_munmap(addr, len)                              \
   __sanitizer_syscall_pre_impl_munmap((long)(addr), (long)(len))
-#define __sanitizer_syscall_post_munmap(res, addr, len) \
+#define __sanitizer_syscall_post_munmap(res, addr, len)                        \
   __sanitizer_syscall_post_impl_munmap(res, (long)(addr), (long)(len))
-#define __sanitizer_syscall_pre_mlock(start, len) \
+#define __sanitizer_syscall_pre_mlock(start, len)                              \
   __sanitizer_syscall_pre_impl_mlock((long)(start), (long)(len))
-#define __sanitizer_syscall_post_mlock(res, start, len) \
+#define __sanitizer_syscall_post_mlock(res, start, len)                        \
   __sanitizer_syscall_post_impl_mlock(res, (long)(start), (long)(len))
-#define __sanitizer_syscall_pre_munlock(start, len) \
+#define __sanitizer_syscall_pre_munlock(start, len)                            \
   __sanitizer_syscall_pre_impl_munlock((long)(start), (long)(len))
-#define __sanitizer_syscall_post_munlock(res, start, len) \
+#define __sanitizer_syscall_post_munlock(res, start, len)                      \
   __sanitizer_syscall_post_impl_munlock(res, (long)(start), (long)(len))
-#define __sanitizer_syscall_pre_mlockall(flags) \
+#define __sanitizer_syscall_pre_mlockall(flags)                                \
   __sanitizer_syscall_pre_impl_mlockall((long)(flags))
-#define __sanitizer_syscall_post_mlockall(res, flags) \
+#define __sanitizer_syscall_post_mlockall(res, flags)                          \
   __sanitizer_syscall_post_impl_mlockall(res, (long)(flags))
-#define __sanitizer_syscall_pre_munlockall() \
+#define __sanitizer_syscall_pre_munlockall()                                   \
   __sanitizer_syscall_pre_impl_munlockall()
-#define __sanitizer_syscall_post_munlockall(res) \
+#define __sanitizer_syscall_post_munlockall(res)                               \
   __sanitizer_syscall_post_impl_munlockall(res)
-#define __sanitizer_syscall_pre_madvise(start, len, behavior)      \
-  __sanitizer_syscall_pre_impl_madvise((long)(start), (long)(len), \
+#define __sanitizer_syscall_pre_madvise(start, len, behavior)                  \
+  __sanitizer_syscall_pre_impl_madvise((long)(start), (long)(len),             \
                                        (long)(behavior))
-#define __sanitizer_syscall_post_madvise(res, start, len, behavior)      \
-  __sanitizer_syscall_post_impl_madvise(res, (long)(start), (long)(len), \
+#define __sanitizer_syscall_post_madvise(res, start, len, behavior)            \
+  __sanitizer_syscall_post_impl_madvise(res, (long)(start), (long)(len),       \
                                         (long)(behavior))
-#define __sanitizer_syscall_pre_mincore(start, len, vec) \
+#define __sanitizer_syscall_pre_mincore(start, len, vec)                       \
   __sanitizer_syscall_pre_impl_mincore((long)(start), (long)(len), (long)(vec))
-#define __sanitizer_syscall_post_mincore(res, start, len, vec)           \
-  __sanitizer_syscall_post_impl_mincore(res, (long)(start), (long)(len), \
+#define __sanitizer_syscall_post_mincore(res, start, len, vec)                 \
+  __sanitizer_syscall_post_impl_mincore(res, (long)(start), (long)(len),       \
                                         (long)(vec))
-#define __sanitizer_syscall_pre_pivot_root(new_root, put_old) \
+#define __sanitizer_syscall_pre_pivot_root(new_root, put_old)                  \
   __sanitizer_syscall_pre_impl_pivot_root((long)(new_root), (long)(put_old))
-#define __sanitizer_syscall_post_pivot_root(res, new_root, put_old) \
-  __sanitizer_syscall_post_impl_pivot_root(res, (long)(new_root),   \
+#define __sanitizer_syscall_post_pivot_root(res, new_root, put_old)            \
+  __sanitizer_syscall_post_impl_pivot_root(res, (long)(new_root),              \
                                            (long)(put_old))
-#define __sanitizer_syscall_pre_chroot(filename) \
+#define __sanitizer_syscall_pre_chroot(filename)                               \
   __sanitizer_syscall_pre_impl_chroot((long)(filename))
-#define __sanitizer_syscall_post_chroot(res, filename) \
+#define __sanitizer_syscall_post_chroot(res, filename)                         \
   __sanitizer_syscall_post_impl_chroot(res, (long)(filename))
-#define __sanitizer_syscall_pre_mknod(filename, mode, dev)           \
-  __sanitizer_syscall_pre_impl_mknod((long)(filename), (long)(mode), \
+#define __sanitizer_syscall_pre_mknod(filename, mode, dev)                     \
+  __sanitizer_syscall_pre_impl_mknod((long)(filename), (long)(mode),           \
                                      (long)(dev))
-#define __sanitizer_syscall_post_mknod(res, filename, mode, dev)           \
-  __sanitizer_syscall_post_impl_mknod(res, (long)(filename), (long)(mode), \
+#define __sanitizer_syscall_post_mknod(res, filename, mode, dev)               \
+  __sanitizer_syscall_post_impl_mknod(res, (long)(filename), (long)(mode),     \
                                       (long)(dev))
-#define __sanitizer_syscall_pre_link(oldname, newname) \
+#define __sanitizer_syscall_pre_link(oldname, newname)                         \
   __sanitizer_syscall_pre_impl_link((long)(oldname), (long)(newname))
-#define __sanitizer_syscall_post_link(res, oldname, newname) \
+#define __sanitizer_syscall_post_link(res, oldname, newname)                   \
   __sanitizer_syscall_post_impl_link(res, (long)(oldname), (long)(newname))
-#define __sanitizer_syscall_pre_symlink(old, new_) \
+#define __sanitizer_syscall_pre_symlink(old, new_)                             \
   __sanitizer_syscall_pre_impl_symlink((long)(old), (long)(new_))
-#define __sanitizer_syscall_post_symlink(res, old, new_) \
+#define __sanitizer_syscall_post_symlink(res, old, new_)                       \
   __sanitizer_syscall_post_impl_symlink(res, (long)(old), (long)(new_))
-#define __sanitizer_syscall_pre_unlink(pathname) \
+#define __sanitizer_syscall_pre_unlink(pathname)                               \
   __sanitizer_syscall_pre_impl_unlink((long)(pathname))
-#define __sanitizer_syscall_post_unlink(res, pathname) \
+#define __sanitizer_syscall_post_unlink(res, pathname)                         \
   __sanitizer_syscall_post_impl_unlink(res, (long)(pathname))
-#define __sanitizer_syscall_pre_rename(oldname, newname) \
+#define __sanitizer_syscall_pre_rename(oldname, newname)                       \
   __sanitizer_syscall_pre_impl_rename((long)(oldname), (long)(newname))
-#define __sanitizer_syscall_post_rename(res, oldname, newname) \
+#define __sanitizer_syscall_post_rename(res, oldname, newname)                 \
   __sanitizer_syscall_post_impl_rename(res, (long)(oldname), (long)(newname))
-#define __sanitizer_syscall_pre_chmod(filename, mode) \
+#define __sanitizer_syscall_pre_chmod(filename, mode)                          \
   __sanitizer_syscall_pre_impl_chmod((long)(filename), (long)(mode))
-#define __sanitizer_syscall_post_chmod(res, filename, mode) \
+#define __sanitizer_syscall_post_chmod(res, filename, mode)                    \
   __sanitizer_syscall_post_impl_chmod(res, (long)(filename), (long)(mode))
-#define __sanitizer_syscall_pre_fchmod(fd, mode) \
+#define __sanitizer_syscall_pre_fchmod(fd, mode)                               \
   __sanitizer_syscall_pre_impl_fchmod((long)(fd), (long)(mode))
-#define __sanitizer_syscall_post_fchmod(res, fd, mode) \
+#define __sanitizer_syscall_post_fchmod(res, fd, mode)                         \
   __sanitizer_syscall_post_impl_fchmod(res, (long)(fd), (long)(mode))
-#define __sanitizer_syscall_pre_fcntl(fd, cmd, arg) \
+#define __sanitizer_syscall_pre_fcntl(fd, cmd, arg)                            \
   __sanitizer_syscall_pre_impl_fcntl((long)(fd), (long)(cmd), (long)(arg))
-#define __sanitizer_syscall_post_fcntl(res, fd, cmd, arg) \
+#define __sanitizer_syscall_post_fcntl(res, fd, cmd, arg)                      \
   __sanitizer_syscall_post_impl_fcntl(res, (long)(fd), (long)(cmd), (long)(arg))
-#define __sanitizer_syscall_pre_fcntl64(fd, cmd, arg) \
+#define __sanitizer_syscall_pre_fcntl64(fd, cmd, arg)                          \
   __sanitizer_syscall_pre_impl_fcntl64((long)(fd), (long)(cmd), (long)(arg))
-#define __sanitizer_syscall_post_fcntl64(res, fd, cmd, arg)           \
-  __sanitizer_syscall_post_impl_fcntl64(res, (long)(fd), (long)(cmd), \
+#define __sanitizer_syscall_post_fcntl64(res, fd, cmd, arg)                    \
+  __sanitizer_syscall_post_impl_fcntl64(res, (long)(fd), (long)(cmd),          \
                                         (long)(arg))
-#define __sanitizer_syscall_pre_pipe(fildes) \
+#define __sanitizer_syscall_pre_pipe(fildes)                                   \
   __sanitizer_syscall_pre_impl_pipe((long)(fildes))
-#define __sanitizer_syscall_post_pipe(res, fildes) \
+#define __sanitizer_syscall_post_pipe(res, fildes)                             \
   __sanitizer_syscall_post_impl_pipe(res, (long)(fildes))
-#define __sanitizer_syscall_pre_pipe2(fildes, flags) \
+#define __sanitizer_syscall_pre_pipe2(fildes, flags)                           \
   __sanitizer_syscall_pre_impl_pipe2((long)(fildes), (long)(flags))
-#define __sanitizer_syscall_post_pipe2(res, fildes, flags) \
+#define __sanitizer_syscall_post_pipe2(res, fildes, flags)                     \
   __sanitizer_syscall_post_impl_pipe2(res, (long)(fildes), (long)(flags))
-#define __sanitizer_syscall_pre_dup(fildes) \
+#define __sanitizer_syscall_pre_dup(fildes)                                    \
   __sanitizer_syscall_pre_impl_dup((long)(fildes))
-#define __sanitizer_syscall_post_dup(res, fildes) \
+#define __sanitizer_syscall_post_dup(res, fildes)                              \
   __sanitizer_syscall_post_impl_dup(res, (long)(fildes))
-#define __sanitizer_syscall_pre_dup2(oldfd, newfd) \
+#define __sanitizer_syscall_pre_dup2(oldfd, newfd)                             \
   __sanitizer_syscall_pre_impl_dup2((long)(oldfd), (long)(newfd))
-#define __sanitizer_syscall_post_dup2(res, oldfd, newfd) \
+#define __sanitizer_syscall_post_dup2(res, oldfd, newfd)                       \
   __sanitizer_syscall_post_impl_dup2(res, (long)(oldfd), (long)(newfd))
-#define __sanitizer_syscall_pre_dup3(oldfd, newfd, flags) \
+#define __sanitizer_syscall_pre_dup3(oldfd, newfd, flags)                      \
   __sanitizer_syscall_pre_impl_dup3((long)(oldfd), (long)(newfd), (long)(flags))
-#define __sanitizer_syscall_post_dup3(res, oldfd, newfd, flags)         \
-  __sanitizer_syscall_post_impl_dup3(res, (long)(oldfd), (long)(newfd), \
+#define __sanitizer_syscall_post_dup3(res, oldfd, newfd, flags)                \
+  __sanitizer_syscall_post_impl_dup3(res, (long)(oldfd), (long)(newfd),        \
                                      (long)(flags))
-#define __sanitizer_syscall_pre_ioperm(from, num, on) \
+#define __sanitizer_syscall_pre_ioperm(from, num, on)                          \
   __sanitizer_syscall_pre_impl_ioperm((long)(from), (long)(num), (long)(on))
-#define __sanitizer_syscall_post_ioperm(res, from, num, on)            \
-  __sanitizer_syscall_post_impl_ioperm(res, (long)(from), (long)(num), \
+#define __sanitizer_syscall_post_ioperm(res, from, num, on)                    \
+  __sanitizer_syscall_post_impl_ioperm(res, (long)(from), (long)(num),         \
                                        (long)(on))
-#define __sanitizer_syscall_pre_ioctl(fd, cmd, arg) \
+#define __sanitizer_syscall_pre_ioctl(fd, cmd, arg)                            \
   __sanitizer_syscall_pre_impl_ioctl((long)(fd), (long)(cmd), (long)(arg))
-#define __sanitizer_syscall_post_ioctl(res, fd, cmd, arg) \
+#define __sanitizer_syscall_post_ioctl(res, fd, cmd, arg)                      \
   __sanitizer_syscall_post_impl_ioctl(res, (long)(fd), (long)(cmd), (long)(arg))
-#define __sanitizer_syscall_pre_flock(fd, cmd) \
+#define __sanitizer_syscall_pre_flock(fd, cmd)                                 \
   __sanitizer_syscall_pre_impl_flock((long)(fd), (long)(cmd))
-#define __sanitizer_syscall_post_flock(res, fd, cmd) \
+#define __sanitizer_syscall_post_flock(res, fd, cmd)                           \
   __sanitizer_syscall_post_impl_flock(res, (long)(fd), (long)(cmd))
-#define __sanitizer_syscall_pre_io_setup(nr_reqs, ctx) \
+#define __sanitizer_syscall_pre_io_setup(nr_reqs, ctx)                         \
   __sanitizer_syscall_pre_impl_io_setup((long)(nr_reqs), (long)(ctx))
-#define __sanitizer_syscall_post_io_setup(res, nr_reqs, ctx) \
+#define __sanitizer_syscall_post_io_setup(res, nr_reqs, ctx)                   \
   __sanitizer_syscall_post_impl_io_setup(res, (long)(nr_reqs), (long)(ctx))
-#define __sanitizer_syscall_pre_io_destroy(ctx) \
+#define __sanitizer_syscall_pre_io_destroy(ctx)                                \
   __sanitizer_syscall_pre_impl_io_destroy((long)(ctx))
-#define __sanitizer_syscall_post_io_destroy(res, ctx) \
+#define __sanitizer_syscall_post_io_destroy(res, ctx)                          \
   __sanitizer_syscall_post_impl_io_destroy(res, (long)(ctx))
-#define __sanitizer_syscall_pre_io_getevents(ctx_id, min_nr, nr, events,    \
-                                             timeout)                       \
-  __sanitizer_syscall_pre_impl_io_getevents((long)(ctx_id), (long)(min_nr), \
-                                            (long)(nr), (long)(events),     \
+#define __sanitizer_syscall_pre_io_getevents(ctx_id, min_nr, nr, events,       \
+                                             timeout)                          \
+  __sanitizer_syscall_pre_impl_io_getevents((long)(ctx_id), (long)(min_nr),    \
+                                            (long)(nr), (long)(events),        \
                                             (long)(timeout))
 #define __sanitizer_syscall_post_io_getevents(res, ctx_id, min_nr, nr, events, \
                                               timeout)                         \
   __sanitizer_syscall_post_impl_io_getevents(res, (long)(ctx_id),              \
                                              (long)(min_nr), (long)(nr),       \
                                              (long)(events), (long)(timeout))
-#define __sanitizer_syscall_pre_io_submit(ctx_id, arg1, arg2)          \
-  __sanitizer_syscall_pre_impl_io_submit((long)(ctx_id), (long)(arg1), \
+#define __sanitizer_syscall_pre_io_submit(ctx_id, arg1, arg2)                  \
+  __sanitizer_syscall_pre_impl_io_submit((long)(ctx_id), (long)(arg1),         \
                                          (long)(arg2))
-#define __sanitizer_syscall_post_io_submit(res, ctx_id, arg1, arg2)          \
-  __sanitizer_syscall_post_impl_io_submit(res, (long)(ctx_id), (long)(arg1), \
+#define __sanitizer_syscall_post_io_submit(res, ctx_id, arg1, arg2)            \
+  __sanitizer_syscall_post_impl_io_submit(res, (long)(ctx_id), (long)(arg1),   \
                                           (long)(arg2))
-#define __sanitizer_syscall_pre_io_cancel(ctx_id, iocb, result)        \
-  __sanitizer_syscall_pre_impl_io_cancel((long)(ctx_id), (long)(iocb), \
+#define __sanitizer_syscall_pre_io_cancel(ctx_id, iocb, result)                \
+  __sanitizer_syscall_pre_impl_io_cancel((long)(ctx_id), (long)(iocb),         \
                                          (long)(result))
-#define __sanitizer_syscall_post_io_cancel(res, ctx_id, iocb, result)        \
-  __sanitizer_syscall_post_impl_io_cancel(res, (long)(ctx_id), (long)(iocb), \
+#define __sanitizer_syscall_post_io_cancel(res, ctx_id, iocb, result)          \
+  __sanitizer_syscall_post_impl_io_cancel(res, (long)(ctx_id), (long)(iocb),   \
                                           (long)(result))
-#define __sanitizer_syscall_pre_sendfile(out_fd, in_fd, offset, count) \
-  __sanitizer_syscall_pre_impl_sendfile((long)(out_fd), (long)(in_fd), \
+#define __sanitizer_syscall_pre_sendfile(out_fd, in_fd, offset, count)         \
+  __sanitizer_syscall_pre_impl_sendfile((long)(out_fd), (long)(in_fd),         \
                                         (long)(offset), (long)(count))
-#define __sanitizer_syscall_post_sendfile(res, out_fd, in_fd, offset, count) \
-  __sanitizer_syscall_post_impl_sendfile(res, (long)(out_fd), (long)(in_fd), \
+#define __sanitizer_syscall_post_sendfile(res, out_fd, in_fd, offset, count)   \
+  __sanitizer_syscall_post_impl_sendfile(res, (long)(out_fd), (long)(in_fd),   \
                                          (long)(offset), (long)(count))
-#define __sanitizer_syscall_pre_sendfile64(out_fd, in_fd, offset, count) \
-  __sanitizer_syscall_pre_impl_sendfile64((long)(out_fd), (long)(in_fd), \
+#define __sanitizer_syscall_pre_sendfile64(out_fd, in_fd, offset, count)       \
+  __sanitizer_syscall_pre_impl_sendfile64((long)(out_fd), (long)(in_fd),       \
                                           (long)(offset), (long)(count))
 #define __sanitizer_syscall_post_sendfile64(res, out_fd, in_fd, offset, count) \
   __sanitizer_syscall_post_impl_sendfile64(res, (long)(out_fd), (long)(in_fd), \
                                            (long)(offset), (long)(count))
-#define __sanitizer_syscall_pre_readlink(path, buf, bufsiz)        \
-  __sanitizer_syscall_pre_impl_readlink((long)(path), (long)(buf), \
+#define __sanitizer_syscall_pre_readlink(path, buf, bufsiz)                    \
+  __sanitizer_syscall_pre_impl_readlink((long)(path), (long)(buf),             \
                                         (long)(bufsiz))
-#define __sanitizer_syscall_post_readlink(res, path, buf, bufsiz)        \
-  __sanitizer_syscall_post_impl_readlink(res, (long)(path), (long)(buf), \
+#define __sanitizer_syscall_post_readlink(res, path, buf, bufsiz)              \
+  __sanitizer_syscall_post_impl_readlink(res, (long)(path), (long)(buf),       \
                                          (long)(bufsiz))
-#define __sanitizer_syscall_pre_creat(pathname, mode) \
+#define __sanitizer_syscall_pre_creat(pathname, mode)                          \
   __sanitizer_syscall_pre_impl_creat((long)(pathname), (long)(mode))
-#define __sanitizer_syscall_post_creat(res, pathname, mode) \
+#define __sanitizer_syscall_post_creat(res, pathname, mode)                    \
   __sanitizer_syscall_post_impl_creat(res, (long)(pathname), (long)(mode))
-#define __sanitizer_syscall_pre_open(filename, flags, mode)          \
-  __sanitizer_syscall_pre_impl_open((long)(filename), (long)(flags), \
+#define __sanitizer_syscall_pre_open(filename, flags, mode)                    \
+  __sanitizer_syscall_pre_impl_open((long)(filename), (long)(flags),           \
                                     (long)(mode))
-#define __sanitizer_syscall_post_open(res, filename, flags, mode)          \
-  __sanitizer_syscall_post_impl_open(res, (long)(filename), (long)(flags), \
+#define __sanitizer_syscall_post_open(res, filename, flags, mode)              \
+  __sanitizer_syscall_post_impl_open(res, (long)(filename), (long)(flags),     \
                                      (long)(mode))
-#define __sanitizer_syscall_pre_close(fd) \
+#define __sanitizer_syscall_pre_close(fd)                                      \
   __sanitizer_syscall_pre_impl_close((long)(fd))
-#define __sanitizer_syscall_post_close(res, fd) \
+#define __sanitizer_syscall_post_close(res, fd)                                \
   __sanitizer_syscall_post_impl_close(res, (long)(fd))
-#define __sanitizer_syscall_pre_access(filename, mode) \
+#define __sanitizer_syscall_pre_access(filename, mode)                         \
   __sanitizer_syscall_pre_impl_access((long)(filename), (long)(mode))
-#define __sanitizer_syscall_post_access(res, filename, mode) \
+#define __sanitizer_syscall_post_access(res, filename, mode)                   \
   __sanitizer_syscall_post_impl_access(res, (long)(filename), (long)(mode))
 #define __sanitizer_syscall_pre_vhangup() __sanitizer_syscall_pre_impl_vhangup()
-#define __sanitizer_syscall_post_vhangup(res) \
+#define __sanitizer_syscall_post_vhangup(res)                                  \
   __sanitizer_syscall_post_impl_vhangup(res)
-#define __sanitizer_syscall_pre_chown(filename, user, group)         \
-  __sanitizer_syscall_pre_impl_chown((long)(filename), (long)(user), \
+#define __sanitizer_syscall_pre_chown(filename, user, group)                   \
+  __sanitizer_syscall_pre_impl_chown((long)(filename), (long)(user),           \
                                      (long)(group))
-#define __sanitizer_syscall_post_chown(res, filename, user, group)         \
-  __sanitizer_syscall_post_impl_chown(res, (long)(filename), (long)(user), \
+#define __sanitizer_syscall_post_chown(res, filename, user, group)             \
+  __sanitizer_syscall_post_impl_chown(res, (long)(filename), (long)(user),     \
                                       (long)(group))
-#define __sanitizer_syscall_pre_lchown(filename, user, group)         \
-  __sanitizer_syscall_pre_impl_lchown((long)(filename), (long)(user), \
+#define __sanitizer_syscall_pre_lchown(filename, user, group)                  \
+  __sanitizer_syscall_pre_impl_lchown((long)(filename), (long)(user),          \
                                       (long)(group))
-#define __sanitizer_syscall_post_lchown(res, filename, user, group)         \
-  __sanitizer_syscall_post_impl_lchown(res, (long)(filename), (long)(user), \
+#define __sanitizer_syscall_post_lchown(res, filename, user, group)            \
+  __sanitizer_syscall_post_impl_lchown(res, (long)(filename), (long)(user),    \
                                        (long)(group))
-#define __sanitizer_syscall_pre_fchown(fd, user, group) \
+#define __sanitizer_syscall_pre_fchown(fd, user, group)                        \
   __sanitizer_syscall_pre_impl_fchown((long)(fd), (long)(user), (long)(group))
-#define __sanitizer_syscall_post_fchown(res, fd, user, group)         \
-  __sanitizer_syscall_post_impl_fchown(res, (long)(fd), (long)(user), \
+#define __sanitizer_syscall_post_fchown(res, fd, user, group)                  \
+  __sanitizer_syscall_post_impl_fchown(res, (long)(fd), (long)(user),          \
                                        (long)(group))
-#define __sanitizer_syscall_pre_chown16(filename, user, group)       \
-  __sanitizer_syscall_pre_impl_chown16((long)(filename), (long)user, \
+#define __sanitizer_syscall_pre_chown16(filename, user, group)                 \
+  __sanitizer_syscall_pre_impl_chown16((long)(filename), (long)user,           \
                                        (long)group)
-#define __sanitizer_syscall_post_chown16(res, filename, user, group)       \
-  __sanitizer_syscall_post_impl_chown16(res, (long)(filename), (long)user, \
+#define __sanitizer_syscall_post_chown16(res, filename, user, group)           \
+  __sanitizer_syscall_post_impl_chown16(res, (long)(filename), (long)user,     \
                                         (long)group)
-#define __sanitizer_syscall_pre_lchown16(filename, user, group)       \
-  __sanitizer_syscall_pre_impl_lchown16((long)(filename), (long)user, \
+#define __sanitizer_syscall_pre_lchown16(filename, user, group)                \
+  __sanitizer_syscall_pre_impl_lchown16((long)(filename), (long)user,          \
                                         (long)group)
-#define __sanitizer_syscall_post_lchown16(res, filename, user, group)       \
-  __sanitizer_syscall_post_impl_lchown16(res, (long)(filename), (long)user, \
+#define __sanitizer_syscall_post_lchown16(res, filename, user, group)          \
+  __sanitizer_syscall_post_impl_lchown16(res, (long)(filename), (long)user,    \
                                          (long)group)
-#define __sanitizer_syscall_pre_fchown16(fd, user, group) \
+#define __sanitizer_syscall_pre_fchown16(fd, user, group)                      \
   __sanitizer_syscall_pre_impl_fchown16((long)(fd), (long)user, (long)group)
-#define __sanitizer_syscall_post_fchown16(res, fd, user, group)       \
-  __sanitizer_syscall_post_impl_fchown16(res, (long)(fd), (long)user, \
+#define __sanitizer_syscall_post_fchown16(res, fd, user, group)                \
+  __sanitizer_syscall_post_impl_fchown16(res, (long)(fd), (long)user,          \
                                          (long)group)
-#define __sanitizer_syscall_pre_setregid16(rgid, egid) \
+#define __sanitizer_syscall_pre_setregid16(rgid, egid)                         \
   __sanitizer_syscall_pre_impl_setregid16((long)rgid, (long)egid)
-#define __sanitizer_syscall_post_setregid16(res, rgid, egid) \
+#define __sanitizer_syscall_post_setregid16(res, rgid, egid)                   \
   __sanitizer_syscall_post_impl_setregid16(res, (long)rgid, (long)egid)
-#define __sanitizer_syscall_pre_setgid16(gid) \
+#define __sanitizer_syscall_pre_setgid16(gid)                                  \
   __sanitizer_syscall_pre_impl_setgid16((long)gid)
-#define __sanitizer_syscall_post_setgid16(res, gid) \
+#define __sanitizer_syscall_post_setgid16(res, gid)                            \
   __sanitizer_syscall_post_impl_setgid16(res, (long)gid)
-#define __sanitizer_syscall_pre_setreuid16(ruid, euid) \
+#define __sanitizer_syscall_pre_setreuid16(ruid, euid)                         \
   __sanitizer_syscall_pre_impl_setreuid16((long)ruid, (long)euid)
-#define __sanitizer_syscall_post_setreuid16(res, ruid, euid) \
+#define __sanitizer_syscall_post_setreuid16(res, ruid, euid)                   \
   __sanitizer_syscall_post_impl_setreuid16(res, (long)ruid, (long)euid)
-#define __sanitizer_syscall_pre_setuid16(uid) \
+#define __sanitizer_syscall_pre_setuid16(uid)                                  \
   __sanitizer_syscall_pre_impl_setuid16((long)uid)
-#define __sanitizer_syscall_post_setuid16(res, uid) \
+#define __sanitizer_syscall_post_setuid16(res, uid)                            \
   __sanitizer_syscall_post_impl_setuid16(res, (long)uid)
-#define __sanitizer_syscall_pre_setresuid16(ruid, euid, suid) \
+#define __sanitizer_syscall_pre_setresuid16(ruid, euid, suid)                  \
   __sanitizer_syscall_pre_impl_setresuid16((long)ruid, (long)euid, (long)suid)
-#define __sanitizer_syscall_post_setresuid16(res, ruid, euid, suid)      \
-  __sanitizer_syscall_post_impl_setresuid16(res, (long)ruid, (long)euid, \
+#define __sanitizer_syscall_post_setresuid16(res, ruid, euid, suid)            \
+  __sanitizer_syscall_post_impl_setresuid16(res, (long)ruid, (long)euid,       \
                                             (long)suid)
-#define __sanitizer_syscall_pre_getresuid16(ruid, euid, suid)          \
-  __sanitizer_syscall_pre_impl_getresuid16((long)(ruid), (long)(euid), \
+#define __sanitizer_syscall_pre_getresuid16(ruid, euid, suid)                  \
+  __sanitizer_syscall_pre_impl_getresuid16((long)(ruid), (long)(euid),         \
                                            (long)(suid))
-#define __sanitizer_syscall_post_getresuid16(res, ruid, euid, suid)          \
-  __sanitizer_syscall_post_impl_getresuid16(res, (long)(ruid), (long)(euid), \
+#define __sanitizer_syscall_post_getresuid16(res, ruid, euid, suid)            \
+  __sanitizer_syscall_post_impl_getresuid16(res, (long)(ruid), (long)(euid),   \
                                             (long)(suid))
-#define __sanitizer_syscall_pre_setresgid16(rgid, egid, sgid) \
+#define __sanitizer_syscall_pre_setresgid16(rgid, egid, sgid)                  \
   __sanitizer_syscall_pre_impl_setresgid16((long)rgid, (long)egid, (long)sgid)
-#define __sanitizer_syscall_post_setresgid16(res, rgid, egid, sgid)      \
-  __sanitizer_syscall_post_impl_setresgid16(res, (long)rgid, (long)egid, \
+#define __sanitizer_syscall_post_setresgid16(res, rgid, egid, sgid)            \
+  __sanitizer_syscall_post_impl_setresgid16(res, (long)rgid, (long)egid,       \
                                             (long)sgid)
-#define __sanitizer_syscall_pre_getresgid16(rgid, egid, sgid)          \
-  __sanitizer_syscall_pre_impl_getresgid16((long)(rgid), (long)(egid), \
+#define __sanitizer_syscall_pre_getresgid16(rgid, egid, sgid)                  \
+  __sanitizer_syscall_pre_impl_getresgid16((long)(rgid), (long)(egid),         \
                                            (long)(sgid))
-#define __sanitizer_syscall_post_getresgid16(res, rgid, egid, sgid)          \
-  __sanitizer_syscall_post_impl_getresgid16(res, (long)(rgid), (long)(egid), \
+#define __sanitizer_syscall_post_getresgid16(res, rgid, egid, sgid)            \
+  __sanitizer_syscall_post_impl_getresgid16(res, (long)(rgid), (long)(egid),   \
                                             (long)(sgid))
-#define __sanitizer_syscall_pre_setfsuid16(uid) \
+#define __sanitizer_syscall_pre_setfsuid16(uid)                                \
   __sanitizer_syscall_pre_impl_setfsuid16((long)uid)
-#define __sanitizer_syscall_post_setfsuid16(res, uid) \
+#define __sanitizer_syscall_post_setfsuid16(res, uid)                          \
   __sanitizer_syscall_post_impl_setfsuid16(res, (long)uid)
-#define __sanitizer_syscall_pre_setfsgid16(gid) \
+#define __sanitizer_syscall_pre_setfsgid16(gid)                                \
   __sanitizer_syscall_pre_impl_setfsgid16((long)gid)
-#define __sanitizer_syscall_post_setfsgid16(res, gid) \
+#define __sanitizer_syscall_post_setfsgid16(res, gid)                          \
   __sanitizer_syscall_post_impl_setfsgid16(res, (long)gid)
-#define __sanitizer_syscall_pre_getgroups16(gidsetsize, grouplist) \
-  __sanitizer_syscall_pre_impl_getgroups16((long)(gidsetsize),     \
+#define __sanitizer_syscall_pre_getgroups16(gidsetsize, grouplist)             \
+  __sanitizer_syscall_pre_impl_getgroups16((long)(gidsetsize),                 \
                                            (long)(grouplist))
-#define __sanitizer_syscall_post_getgroups16(res, gidsetsize, grouplist) \
-  __sanitizer_syscall_post_impl_getgroups16(res, (long)(gidsetsize),     \
+#define __sanitizer_syscall_post_getgroups16(res, gidsetsize, grouplist)       \
+  __sanitizer_syscall_post_impl_getgroups16(res, (long)(gidsetsize),           \
                                             (long)(grouplist))
-#define __sanitizer_syscall_pre_setgroups16(gidsetsize, grouplist) \
-  __sanitizer_syscall_pre_impl_setgroups16((long)(gidsetsize),     \
+#define __sanitizer_syscall_pre_setgroups16(gidsetsize, grouplist)             \
+  __sanitizer_syscall_pre_impl_setgroups16((long)(gidsetsize),                 \
                                            (long)(grouplist))
-#define __sanitizer_syscall_post_setgroups16(res, gidsetsize, grouplist) \
-  __sanitizer_syscall_post_impl_setgroups16(res, (long)(gidsetsize),     \
+#define __sanitizer_syscall_post_setgroups16(res, gidsetsize, grouplist)       \
+  __sanitizer_syscall_post_impl_setgroups16(res, (long)(gidsetsize),           \
                                             (long)(grouplist))
-#define __sanitizer_syscall_pre_getuid16() \
+#define __sanitizer_syscall_pre_getuid16()                                     \
   __sanitizer_syscall_pre_impl_getuid16()
-#define __sanitizer_syscall_post_getuid16(res) \
+#define __sanitizer_syscall_post_getuid16(res)                                 \
   __sanitizer_syscall_post_impl_getuid16(res)
-#define __sanitizer_syscall_pre_geteuid16() \
+#define __sanitizer_syscall_pre_geteuid16()                                    \
   __sanitizer_syscall_pre_impl_geteuid16()
-#define __sanitizer_syscall_post_geteuid16(res) \
+#define __sanitizer_syscall_post_geteuid16(res)                                \
   __sanitizer_syscall_post_impl_geteuid16(res)
-#define __sanitizer_syscall_pre_getgid16() \
+#define __sanitizer_syscall_pre_getgid16()                                     \
   __sanitizer_syscall_pre_impl_getgid16()
-#define __sanitizer_syscall_post_getgid16(res) \
+#define __sanitizer_syscall_post_getgid16(res)                                 \
   __sanitizer_syscall_post_impl_getgid16(res)
-#define __sanitizer_syscall_pre_getegid16() \
+#define __sanitizer_syscall_pre_getegid16()                                    \
   __sanitizer_syscall_pre_impl_getegid16()
-#define __sanitizer_syscall_post_getegid16(res) \
+#define __sanitizer_syscall_post_getegid16(res)                                \
   __sanitizer_syscall_post_impl_getegid16(res)
-#define __sanitizer_syscall_pre_utime(filename, times) \
+#define __sanitizer_syscall_pre_utime(filename, times)                         \
   __sanitizer_syscall_pre_impl_utime((long)(filename), (long)(times))
-#define __sanitizer_syscall_post_utime(res, filename, times) \
+#define __sanitizer_syscall_post_utime(res, filename, times)                   \
   __sanitizer_syscall_post_impl_utime(res, (long)(filename), (long)(times))
-#define __sanitizer_syscall_pre_utimes(filename, utimes) \
+#define __sanitizer_syscall_pre_utimes(filename, utimes)                       \
   __sanitizer_syscall_pre_impl_utimes((long)(filename), (long)(utimes))
-#define __sanitizer_syscall_post_utimes(res, filename, utimes) \
+#define __sanitizer_syscall_post_utimes(res, filename, utimes)                 \
   __sanitizer_syscall_post_impl_utimes(res, (long)(filename), (long)(utimes))
-#define __sanitizer_syscall_pre_lseek(fd, offset, origin) \
+#define __sanitizer_syscall_pre_lseek(fd, offset, origin)                      \
   __sanitizer_syscall_pre_impl_lseek((long)(fd), (long)(offset), (long)(origin))
-#define __sanitizer_syscall_post_lseek(res, fd, offset, origin)        \
-  __sanitizer_syscall_post_impl_lseek(res, (long)(fd), (long)(offset), \
+#define __sanitizer_syscall_post_lseek(res, fd, offset, origin)                \
+  __sanitizer_syscall_post_impl_lseek(res, (long)(fd), (long)(offset),         \
                                       (long)(origin))
-#define __sanitizer_syscall_pre_llseek(fd, offset_high, offset_low, result, \
-                                       origin)                              \
-  __sanitizer_syscall_pre_impl_llseek((long)(fd), (long)(offset_high),      \
-                                      (long)(offset_low), (long)(result),   \
+#define __sanitizer_syscall_pre_llseek(fd, offset_high, offset_low, result,    \
+                                       origin)                                 \
+  __sanitizer_syscall_pre_impl_llseek((long)(fd), (long)(offset_high),         \
+                                      (long)(offset_low), (long)(result),      \
                                       (long)(origin))
-#define __sanitizer_syscall_post_llseek(res, fd, offset_high, offset_low,    \
-                                        result, origin)                      \
-  __sanitizer_syscall_post_impl_llseek(res, (long)(fd), (long)(offset_high), \
-                                       (long)(offset_low), (long)(result),   \
+#define __sanitizer_syscall_post_llseek(res, fd, offset_high, offset_low,      \
+                                        result, origin)                        \
+  __sanitizer_syscall_post_impl_llseek(res, (long)(fd), (long)(offset_high),   \
+                                       (long)(offset_low), (long)(result),     \
                                        (long)(origin))
-#define __sanitizer_syscall_pre_read(fd, buf, count) \
+#define __sanitizer_syscall_pre_read(fd, buf, count)                           \
   __sanitizer_syscall_pre_impl_read((long)(fd), (long)(buf), (long)(count))
-#define __sanitizer_syscall_post_read(res, fd, buf, count)         \
-  __sanitizer_syscall_post_impl_read(res, (long)(fd), (long)(buf), \
+#define __sanitizer_syscall_post_read(res, fd, buf, count)                     \
+  __sanitizer_syscall_post_impl_read(res, (long)(fd), (long)(buf),             \
                                      (long)(count))
-#define __sanitizer_syscall_pre_readv(fd, vec, vlen) \
+#define __sanitizer_syscall_pre_readv(fd, vec, vlen)                           \
   __sanitizer_syscall_pre_impl_readv((long)(fd), (long)(vec), (long)(vlen))
-#define __sanitizer_syscall_post_readv(res, fd, vec, vlen)          \
-  __sanitizer_syscall_post_impl_readv(res, (long)(fd), (long)(vec), \
+#define __sanitizer_syscall_post_readv(res, fd, vec, vlen)                     \
+  __sanitizer_syscall_post_impl_readv(res, (long)(fd), (long)(vec),            \
                                       (long)(vlen))
-#define __sanitizer_syscall_pre_write(fd, buf, count) \
+#define __sanitizer_syscall_pre_write(fd, buf, count)                          \
   __sanitizer_syscall_pre_impl_write((long)(fd), (long)(buf), (long)(count))
-#define __sanitizer_syscall_post_write(res, fd, buf, count)         \
-  __sanitizer_syscall_post_impl_write(res, (long)(fd), (long)(buf), \
+#define __sanitizer_syscall_post_write(res, fd, buf, count)                    \
+  __sanitizer_syscall_post_impl_write(res, (long)(fd), (long)(buf),            \
                                       (long)(count))
-#define __sanitizer_syscall_pre_writev(fd, vec, vlen) \
+#define __sanitizer_syscall_pre_writev(fd, vec, vlen)                          \
   __sanitizer_syscall_pre_impl_writev((long)(fd), (long)(vec), (long)(vlen))
-#define __sanitizer_syscall_post_writev(res, fd, vec, vlen)          \
-  __sanitizer_syscall_post_impl_writev(res, (long)(fd), (long)(vec), \
+#define __sanitizer_syscall_post_writev(res, fd, vec, vlen)                    \
+  __sanitizer_syscall_post_impl_writev(res, (long)(fd), (long)(vec),           \
                                        (long)(vlen))
 
 #ifdef _LP64
 #define __sanitizer_syscall_pre_pread64(fd, buf, count, pos)                   \
   __sanitizer_syscall_pre_impl_pread64((long)(fd), (long)(buf), (long)(count), \
                                        (long)(pos))
-#define __sanitizer_syscall_post_pread64(res, fd, buf, count, pos)    \
-  __sanitizer_syscall_post_impl_pread64(res, (long)(fd), (long)(buf), \
+#define __sanitizer_syscall_post_pread64(res, fd, buf, count, pos)             \
+  __sanitizer_syscall_post_impl_pread64(res, (long)(fd), (long)(buf),          \
                                         (long)(count), (long)(pos))
-#define __sanitizer_syscall_pre_pwrite64(fd, buf, count, pos)    \
-  __sanitizer_syscall_pre_impl_pwrite64((long)(fd), (long)(buf), \
+#define __sanitizer_syscall_pre_pwrite64(fd, buf, count, pos)                  \
+  __sanitizer_syscall_pre_impl_pwrite64((long)(fd), (long)(buf),               \
                                         (long)(count), (long)(pos))
-#define __sanitizer_syscall_post_pwrite64(res, fd, buf, count, pos)    \
-  __sanitizer_syscall_post_impl_pwrite64(res, (long)(fd), (long)(buf), \
+#define __sanitizer_syscall_post_pwrite64(res, fd, buf, count, pos)            \
+  __sanitizer_syscall_post_impl_pwrite64(res, (long)(fd), (long)(buf),         \
                                          (long)(count), (long)(pos))
 #else
 #define __sanitizer_syscall_pre_pread64(fd, buf, count, pos0, pos1)            \
   __sanitizer_syscall_pre_impl_pread64((long)(fd), (long)(buf), (long)(count), \
                                        (long)(pos0), (long)(pos1))
-#define __sanitizer_syscall_post_pread64(res, fd, buf, count, pos0, pos1) \
-  __sanitizer_syscall_post_impl_pread64(res, (long)(fd), (long)(buf),     \
-                                        (long)(count), (long)(pos0), \
-                                        (long)(pos1))
-#define __sanitizer_syscall_pre_pwrite64(fd, buf, count, pos0, pos1) \
-  __sanitizer_syscall_pre_impl_pwrite64(                             \
+#define __sanitizer_syscall_post_pread64(res, fd, buf, count, pos0, pos1)      \
+  __sanitizer_syscall_post_impl_pread64(                                       \
+      res, (long)(fd), (long)(buf), (long)(count), (long)(pos0), (long)(pos1))
+#define __sanitizer_syscall_pre_pwrite64(fd, buf, count, pos0, pos1)           \
+  __sanitizer_syscall_pre_impl_pwrite64(                                       \
       (long)(fd), (long)(buf), (long)(count), (long)(pos0), (long)(pos1))
-#define __sanitizer_syscall_post_pwrite64(res, fd, buf, count, pos0, pos1) \
-  __sanitizer_syscall_post_impl_pwrite64(                                  \
+#define __sanitizer_syscall_post_pwrite64(res, fd, buf, count, pos0, pos1)     \
+  __sanitizer_syscall_post_impl_pwrite64(                                      \
       res, (long)(fd), (long)(buf), (long)(count), (long)(pos0), (long)(pos1))
 #endif
 
-#define __sanitizer_syscall_pre_preadv(fd, vec, vlen, pos_l, pos_h)          \
-  __sanitizer_syscall_pre_impl_preadv((long)(fd), (long)(vec), (long)(vlen), \
+#define __sanitizer_syscall_pre_preadv(fd, vec, vlen, pos_l, pos_h)            \
+  __sanitizer_syscall_pre_impl_preadv((long)(fd), (long)(vec), (long)(vlen),   \
                                       (long)(pos_l), (long)(pos_h))
-#define __sanitizer_syscall_post_preadv(res, fd, vec, vlen, pos_l, pos_h) \
-  __sanitizer_syscall_post_impl_preadv(res, (long)(fd), (long)(vec),      \
-                                       (long)(vlen), (long)(pos_l),       \
+#define __sanitizer_syscall_post_preadv(res, fd, vec, vlen, pos_l, pos_h)      \
+  __sanitizer_syscall_post_impl_preadv(res, (long)(fd), (long)(vec),           \
+                                       (long)(vlen), (long)(pos_l),            \
                                        (long)(pos_h))
-#define __sanitizer_syscall_pre_pwritev(fd, vec, vlen, pos_l, pos_h)          \
-  __sanitizer_syscall_pre_impl_pwritev((long)(fd), (long)(vec), (long)(vlen), \
+#define __sanitizer_syscall_pre_pwritev(fd, vec, vlen, pos_l, pos_h)           \
+  __sanitizer_syscall_pre_impl_pwritev((long)(fd), (long)(vec), (long)(vlen),  \
                                        (long)(pos_l), (long)(pos_h))
-#define __sanitizer_syscall_post_pwritev(res, fd, vec, vlen, pos_l, pos_h) \
-  __sanitizer_syscall_post_impl_pwritev(res, (long)(fd), (long)(vec),      \
-                                        (long)(vlen), (long)(pos_l),       \
+#define __sanitizer_syscall_post_pwritev(res, fd, vec, vlen, pos_l, pos_h)     \
+  __sanitizer_syscall_post_impl_pwritev(res, (long)(fd), (long)(vec),          \
+                                        (long)(vlen), (long)(pos_l),           \
                                         (long)(pos_h))
-#define __sanitizer_syscall_pre_getcwd(buf, size) \
+#define __sanitizer_syscall_pre_getcwd(buf, size)                              \
   __sanitizer_syscall_pre_impl_getcwd((long)(buf), (long)(size))
-#define __sanitizer_syscall_post_getcwd(res, buf, size) \
+#define __sanitizer_syscall_post_getcwd(res, buf, size)                        \
   __sanitizer_syscall_post_impl_getcwd(res, (long)(buf), (long)(size))
-#define __sanitizer_syscall_pre_mkdir(pathname, mode) \
+#define __sanitizer_syscall_pre_mkdir(pathname, mode)                          \
   __sanitizer_syscall_pre_impl_mkdir((long)(pathname), (long)(mode))
-#define __sanitizer_syscall_post_mkdir(res, pathname, mode) \
+#define __sanitizer_syscall_post_mkdir(res, pathname, mode)                    \
   __sanitizer_syscall_post_impl_mkdir(res, (long)(pathname), (long)(mode))
-#define __sanitizer_syscall_pre_chdir(filename) \
+#define __sanitizer_syscall_pre_chdir(filename)                                \
   __sanitizer_syscall_pre_impl_chdir((long)(filename))
-#define __sanitizer_syscall_post_chdir(res, filename) \
+#define __sanitizer_syscall_post_chdir(res, filename)                          \
   __sanitizer_syscall_post_impl_chdir(res, (long)(filename))
-#define __sanitizer_syscall_pre_fchdir(fd) \
+#define __sanitizer_syscall_pre_fchdir(fd)                                     \
   __sanitizer_syscall_pre_impl_fchdir((long)(fd))
-#define __sanitizer_syscall_post_fchdir(res, fd) \
+#define __sanitizer_syscall_post_fchdir(res, fd)                               \
   __sanitizer_syscall_post_impl_fchdir(res, (long)(fd))
-#define __sanitizer_syscall_pre_rmdir(pathname) \
+#define __sanitizer_syscall_pre_rmdir(pathname)                                \
   __sanitizer_syscall_pre_impl_rmdir((long)(pathname))
-#define __sanitizer_syscall_post_rmdir(res, pathname) \
+#define __sanitizer_syscall_post_rmdir(res, pathname)                          \
   __sanitizer_syscall_post_impl_rmdir(res, (long)(pathname))
-#define __sanitizer_syscall_pre_lookup_dcookie(cookie64, buf, len)           \
-  __sanitizer_syscall_pre_impl_lookup_dcookie((long)(cookie64), (long)(buf), \
+#define __sanitizer_syscall_pre_lookup_dcookie(cookie64, buf, len)             \
+  __sanitizer_syscall_pre_impl_lookup_dcookie((long)(cookie64), (long)(buf),   \
                                               (long)(len))
-#define __sanitizer_syscall_post_lookup_dcookie(res, cookie64, buf, len) \
-  __sanitizer_syscall_post_impl_lookup_dcookie(res, (long)(cookie64),    \
+#define __sanitizer_syscall_post_lookup_dcookie(res, cookie64, buf, len)       \
+  __sanitizer_syscall_post_impl_lookup_dcookie(res, (long)(cookie64),          \
                                                (long)(buf), (long)(len))
-#define __sanitizer_syscall_pre_quotactl(cmd, special, id, addr)      \
-  __sanitizer_syscall_pre_impl_quotactl((long)(cmd), (long)(special), \
+#define __sanitizer_syscall_pre_quotactl(cmd, special, id, addr)               \
+  __sanitizer_syscall_pre_impl_quotactl((long)(cmd), (long)(special),          \
                                         (long)(id), (long)(addr))
-#define __sanitizer_syscall_post_quotactl(res, cmd, special, id, addr)      \
-  __sanitizer_syscall_post_impl_quotactl(res, (long)(cmd), (long)(special), \
+#define __sanitizer_syscall_post_quotactl(res, cmd, special, id, addr)         \
+  __sanitizer_syscall_post_impl_quotactl(res, (long)(cmd), (long)(special),    \
                                          (long)(id), (long)(addr))
-#define __sanitizer_syscall_pre_getdents(fd, dirent, count)         \
-  __sanitizer_syscall_pre_impl_getdents((long)(fd), (long)(dirent), \
+#define __sanitizer_syscall_pre_getdents(fd, dirent, count)                    \
+  __sanitizer_syscall_pre_impl_getdents((long)(fd), (long)(dirent),            \
                                         (long)(count))
-#define __sanitizer_syscall_post_getdents(res, fd, dirent, count)         \
-  __sanitizer_syscall_post_impl_getdents(res, (long)(fd), (long)(dirent), \
+#define __sanitizer_syscall_post_getdents(res, fd, dirent, count)              \
+  __sanitizer_syscall_post_impl_getdents(res, (long)(fd), (long)(dirent),      \
                                          (long)(count))
-#define __sanitizer_syscall_pre_getdents64(fd, dirent, count)         \
-  __sanitizer_syscall_pre_impl_getdents64((long)(fd), (long)(dirent), \
+#define __sanitizer_syscall_pre_getdents64(fd, dirent, count)                  \
+  __sanitizer_syscall_pre_impl_getdents64((long)(fd), (long)(dirent),          \
                                           (long)(count))
-#define __sanitizer_syscall_post_getdents64(res, fd, dirent, count)         \
-  __sanitizer_syscall_post_impl_getdents64(res, (long)(fd), (long)(dirent), \
+#define __sanitizer_syscall_post_getdents64(res, fd, dirent, count)            \
+  __sanitizer_syscall_post_impl_getdents64(res, (long)(fd), (long)(dirent),    \
                                            (long)(count))
 #define __sanitizer_syscall_pre_setsockopt(fd, level, optname, optval, optlen) \
   __sanitizer_syscall_pre_impl_setsockopt((long)(fd), (long)(level),           \
                                           (long)(optname), (long)(optval),     \
                                           (long)(optlen))
-#define __sanitizer_syscall_post_setsockopt(res, fd, level, optname, optval, \
-                                            optlen)                          \
-  __sanitizer_syscall_post_impl_setsockopt(res, (long)(fd), (long)(level),   \
-                                           (long)(optname), (long)(optval),  \
+#define __sanitizer_syscall_post_setsockopt(res, fd, level, optname, optval,   \
+                                            optlen)                            \
+  __sanitizer_syscall_post_impl_setsockopt(res, (long)(fd), (long)(level),     \
+                                           (long)(optname), (long)(optval),    \
                                            (long)(optlen))
 #define __sanitizer_syscall_pre_getsockopt(fd, level, optname, optval, optlen) \
   __sanitizer_syscall_pre_impl_getsockopt((long)(fd), (long)(level),           \
                                           (long)(optname), (long)(optval),     \
                                           (long)(optlen))
-#define __sanitizer_syscall_post_getsockopt(res, fd, level, optname, optval, \
-                                            optlen)                          \
-  __sanitizer_syscall_post_impl_getsockopt(res, (long)(fd), (long)(level),   \
-                                           (long)(optname), (long)(optval),  \
+#define __sanitizer_syscall_post_getsockopt(res, fd, level, optname, optval,   \
+                                            optlen)                            \
+  __sanitizer_syscall_post_impl_getsockopt(res, (long)(fd), (long)(level),     \
+                                           (long)(optname), (long)(optval),    \
                                            (long)(optlen))
-#define __sanitizer_syscall_pre_bind(arg0, arg1, arg2) \
+#define __sanitizer_syscall_pre_bind(arg0, arg1, arg2)                         \
   __sanitizer_syscall_pre_impl_bind((long)(arg0), (long)(arg1), (long)(arg2))
-#define __sanitizer_syscall_post_bind(res, arg0, arg1, arg2)          \
-  __sanitizer_syscall_post_impl_bind(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_bind(res, arg0, arg1, arg2)                   \
+  __sanitizer_syscall_post_impl_bind(res, (long)(arg0), (long)(arg1),          \
                                      (long)(arg2))
-#define __sanitizer_syscall_pre_connect(arg0, arg1, arg2) \
+#define __sanitizer_syscall_pre_connect(arg0, arg1, arg2)                      \
   __sanitizer_syscall_pre_impl_connect((long)(arg0), (long)(arg1), (long)(arg2))
-#define __sanitizer_syscall_post_connect(res, arg0, arg1, arg2)          \
-  __sanitizer_syscall_post_impl_connect(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_connect(res, arg0, arg1, arg2)                \
+  __sanitizer_syscall_post_impl_connect(res, (long)(arg0), (long)(arg1),       \
                                         (long)(arg2))
-#define __sanitizer_syscall_pre_accept(arg0, arg1, arg2) \
+#define __sanitizer_syscall_pre_accept(arg0, arg1, arg2)                       \
   __sanitizer_syscall_pre_impl_accept((long)(arg0), (long)(arg1), (long)(arg2))
-#define __sanitizer_syscall_post_accept(res, arg0, arg1, arg2)          \
-  __sanitizer_syscall_post_impl_accept(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_accept(res, arg0, arg1, arg2)                 \
+  __sanitizer_syscall_post_impl_accept(res, (long)(arg0), (long)(arg1),        \
                                        (long)(arg2))
-#define __sanitizer_syscall_pre_accept4(arg0, arg1, arg2, arg3)    \
-  __sanitizer_syscall_pre_impl_accept4((long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_pre_accept4(arg0, arg1, arg2, arg3)                \
+  __sanitizer_syscall_pre_impl_accept4((long)(arg0), (long)(arg1),             \
                                        (long)(arg2), (long)(arg3))
-#define __sanitizer_syscall_post_accept4(res, arg0, arg1, arg2, arg3)    \
-  __sanitizer_syscall_post_impl_accept4(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_accept4(res, arg0, arg1, arg2, arg3)          \
+  __sanitizer_syscall_post_impl_accept4(res, (long)(arg0), (long)(arg1),       \
                                         (long)(arg2), (long)(arg3))
-#define __sanitizer_syscall_pre_getsockname(arg0, arg1, arg2)          \
-  __sanitizer_syscall_pre_impl_getsockname((long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_pre_getsockname(arg0, arg1, arg2)                  \
+  __sanitizer_syscall_pre_impl_getsockname((long)(arg0), (long)(arg1),         \
                                            (long)(arg2))
-#define __sanitizer_syscall_post_getsockname(res, arg0, arg1, arg2)          \
-  __sanitizer_syscall_post_impl_getsockname(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_getsockname(res, arg0, arg1, arg2)            \
+  __sanitizer_syscall_post_impl_getsockname(res, (long)(arg0), (long)(arg1),   \
                                             (long)(arg2))
-#define __sanitizer_syscall_pre_getpeername(arg0, arg1, arg2)          \
-  __sanitizer_syscall_pre_impl_getpeername((long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_pre_getpeername(arg0, arg1, arg2)                  \
+  __sanitizer_syscall_pre_impl_getpeername((long)(arg0), (long)(arg1),         \
                                            (long)(arg2))
-#define __sanitizer_syscall_post_getpeername(res, arg0, arg1, arg2)          \
-  __sanitizer_syscall_post_impl_getpeername(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_getpeername(res, arg0, arg1, arg2)            \
+  __sanitizer_syscall_post_impl_getpeername(res, (long)(arg0), (long)(arg1),   \
                                             (long)(arg2))
-#define __sanitizer_syscall_pre_send(arg0, arg1, arg2, arg3)                  \
-  __sanitizer_syscall_pre_impl_send((long)(arg0), (long)(arg1), (long)(arg2), \
+#define __sanitizer_syscall_pre_send(arg0, arg1, arg2, arg3)                   \
+  __sanitizer_syscall_pre_impl_send((long)(arg0), (long)(arg1), (long)(arg2),  \
                                     (long)(arg3))
-#define __sanitizer_syscall_post_send(res, arg0, arg1, arg2, arg3)    \
-  __sanitizer_syscall_post_impl_send(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_send(res, arg0, arg1, arg2, arg3)             \
+  __sanitizer_syscall_post_impl_send(res, (long)(arg0), (long)(arg1),          \
                                      (long)(arg2), (long)(arg3))
-#define __sanitizer_syscall_pre_sendto(arg0, arg1, arg2, arg3, arg4, arg5) \
-  __sanitizer_syscall_pre_impl_sendto((long)(arg0), (long)(arg1),          \
-                                      (long)(arg2), (long)(arg3),          \
+#define __sanitizer_syscall_pre_sendto(arg0, arg1, arg2, arg3, arg4, arg5)     \
+  __sanitizer_syscall_pre_impl_sendto((long)(arg0), (long)(arg1),              \
+                                      (long)(arg2), (long)(arg3),              \
                                       (long)(arg4), (long)(arg5))
-#define __sanitizer_syscall_post_sendto(res, arg0, arg1, arg2, arg3, arg4, \
-                                        arg5)                              \
-  __sanitizer_syscall_post_impl_sendto(res, (long)(arg0), (long)(arg1),    \
-                                       (long)(arg2), (long)(arg3),         \
+#define __sanitizer_syscall_post_sendto(res, arg0, arg1, arg2, arg3, arg4,     \
+                                        arg5)                                  \
+  __sanitizer_syscall_post_impl_sendto(res, (long)(arg0), (long)(arg1),        \
+                                       (long)(arg2), (long)(arg3),             \
                                        (long)(arg4), (long)(arg5))
-#define __sanitizer_syscall_pre_sendmsg(fd, msg, flags) \
+#define __sanitizer_syscall_pre_sendmsg(fd, msg, flags)                        \
   __sanitizer_syscall_pre_impl_sendmsg((long)(fd), (long)(msg), (long)(flags))
-#define __sanitizer_syscall_post_sendmsg(res, fd, msg, flags)         \
-  __sanitizer_syscall_post_impl_sendmsg(res, (long)(fd), (long)(msg), \
+#define __sanitizer_syscall_post_sendmsg(res, fd, msg, flags)                  \
+  __sanitizer_syscall_post_impl_sendmsg(res, (long)(fd), (long)(msg),          \
                                         (long)(flags))
 #define __sanitizer_syscall_pre_sendmmsg(fd, msg, vlen, flags)                 \
   __sanitizer_syscall_pre_impl_sendmmsg((long)(fd), (long)(msg), (long)(vlen), \
                                         (long)(flags))
-#define __sanitizer_syscall_post_sendmmsg(res, fd, msg, vlen, flags)   \
-  __sanitizer_syscall_post_impl_sendmmsg(res, (long)(fd), (long)(msg), \
+#define __sanitizer_syscall_post_sendmmsg(res, fd, msg, vlen, flags)           \
+  __sanitizer_syscall_post_impl_sendmmsg(res, (long)(fd), (long)(msg),         \
                                          (long)(vlen), (long)(flags))
-#define __sanitizer_syscall_pre_recv(arg0, arg1, arg2, arg3)                  \
-  __sanitizer_syscall_pre_impl_recv((long)(arg0), (long)(arg1), (long)(arg2), \
+#define __sanitizer_syscall_pre_recv(arg0, arg1, arg2, arg3)                   \
+  __sanitizer_syscall_pre_impl_recv((long)(arg0), (long)(arg1), (long)(arg2),  \
                                     (long)(arg3))
-#define __sanitizer_syscall_post_recv(res, arg0, arg1, arg2, arg3)    \
-  __sanitizer_syscall_post_impl_recv(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_recv(res, arg0, arg1, arg2, arg3)             \
+  __sanitizer_syscall_post_impl_recv(res, (long)(arg0), (long)(arg1),          \
                                      (long)(arg2), (long)(arg3))
-#define __sanitizer_syscall_pre_recvfrom(arg0, arg1, arg2, arg3, arg4, arg5) \
-  __sanitizer_syscall_pre_impl_recvfrom((long)(arg0), (long)(arg1),          \
-                                        (long)(arg2), (long)(arg3),          \
+#define __sanitizer_syscall_pre_recvfrom(arg0, arg1, arg2, arg3, arg4, arg5)   \
+  __sanitizer_syscall_pre_impl_recvfrom((long)(arg0), (long)(arg1),            \
+                                        (long)(arg2), (long)(arg3),            \
                                         (long)(arg4), (long)(arg5))
-#define __sanitizer_syscall_post_recvfrom(res, arg0, arg1, arg2, arg3, arg4, \
-                                          arg5)                              \
-  __sanitizer_syscall_post_impl_recvfrom(res, (long)(arg0), (long)(arg1),    \
-                                         (long)(arg2), (long)(arg3),         \
+#define __sanitizer_syscall_post_recvfrom(res, arg0, arg1, arg2, arg3, arg4,   \
+                                          arg5)                                \
+  __sanitizer_syscall_post_impl_recvfrom(res, (long)(arg0), (long)(arg1),      \
+                                         (long)(arg2), (long)(arg3),           \
                                          (long)(arg4), (long)(arg5))
-#define __sanitizer_syscall_pre_recvmsg(fd, msg, flags) \
+#define __sanitizer_syscall_pre_recvmsg(fd, msg, flags)                        \
   __sanitizer_syscall_pre_impl_recvmsg((long)(fd), (long)(msg), (long)(flags))
-#define __sanitizer_syscall_post_recvmsg(res, fd, msg, flags)         \
-  __sanitizer_syscall_post_impl_recvmsg(res, (long)(fd), (long)(msg), \
+#define __sanitizer_syscall_post_recvmsg(res, fd, msg, flags)                  \
+  __sanitizer_syscall_post_impl_recvmsg(res, (long)(fd), (long)(msg),          \
                                         (long)(flags))
 #define __sanitizer_syscall_pre_recvmmsg(fd, msg, vlen, flags, timeout)        \
   __sanitizer_syscall_pre_impl_recvmmsg((long)(fd), (long)(msg), (long)(vlen), \
                                         (long)(flags), (long)(timeout))
-#define __sanitizer_syscall_post_recvmmsg(res, fd, msg, vlen, flags, timeout) \
-  __sanitizer_syscall_post_impl_recvmmsg(res, (long)(fd), (long)(msg),        \
-                                         (long)(vlen), (long)(flags),         \
+#define __sanitizer_syscall_post_recvmmsg(res, fd, msg, vlen, flags, timeout)  \
+  __sanitizer_syscall_post_impl_recvmmsg(res, (long)(fd), (long)(msg),         \
+                                         (long)(vlen), (long)(flags),          \
                                          (long)(timeout))
-#define __sanitizer_syscall_pre_socket(arg0, arg1, arg2) \
+#define __sanitizer_syscall_pre_socket(arg0, arg1, arg2)                       \
   __sanitizer_syscall_pre_impl_socket((long)(arg0), (long)(arg1), (long)(arg2))
-#define __sanitizer_syscall_post_socket(res, arg0, arg1, arg2)          \
-  __sanitizer_syscall_post_impl_socket(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_socket(res, arg0, arg1, arg2)                 \
+  __sanitizer_syscall_post_impl_socket(res, (long)(arg0), (long)(arg1),        \
                                        (long)(arg2))
-#define __sanitizer_syscall_pre_socketpair(arg0, arg1, arg2, arg3)    \
-  __sanitizer_syscall_pre_impl_socketpair((long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_pre_socketpair(arg0, arg1, arg2, arg3)             \
+  __sanitizer_syscall_pre_impl_socketpair((long)(arg0), (long)(arg1),          \
                                           (long)(arg2), (long)(arg3))
-#define __sanitizer_syscall_post_socketpair(res, arg0, arg1, arg2, arg3)    \
-  __sanitizer_syscall_post_impl_socketpair(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_socketpair(res, arg0, arg1, arg2, arg3)       \
+  __sanitizer_syscall_post_impl_socketpair(res, (long)(arg0), (long)(arg1),    \
                                            (long)(arg2), (long)(arg3))
-#define __sanitizer_syscall_pre_socketcall(call, args) \
+#define __sanitizer_syscall_pre_socketcall(call, args)                         \
   __sanitizer_syscall_pre_impl_socketcall((long)(call), (long)(args))
-#define __sanitizer_syscall_post_socketcall(res, call, args) \
+#define __sanitizer_syscall_post_socketcall(res, call, args)                   \
   __sanitizer_syscall_post_impl_socketcall(res, (long)(call), (long)(args))
-#define __sanitizer_syscall_pre_listen(arg0, arg1) \
+#define __sanitizer_syscall_pre_listen(arg0, arg1)                             \
   __sanitizer_syscall_pre_impl_listen((long)(arg0), (long)(arg1))
-#define __sanitizer_syscall_post_listen(res, arg0, arg1) \
+#define __sanitizer_syscall_post_listen(res, arg0, arg1)                       \
   __sanitizer_syscall_post_impl_listen(res, (long)(arg0), (long)(arg1))
-#define __sanitizer_syscall_pre_poll(ufds, nfds, timeout) \
+#define __sanitizer_syscall_pre_poll(ufds, nfds, timeout)                      \
   __sanitizer_syscall_pre_impl_poll((long)(ufds), (long)(nfds), (long)(timeout))
-#define __sanitizer_syscall_post_poll(res, ufds, nfds, timeout)       \
-  __sanitizer_syscall_post_impl_poll(res, (long)(ufds), (long)(nfds), \
+#define __sanitizer_syscall_post_poll(res, ufds, nfds, timeout)                \
+  __sanitizer_syscall_post_impl_poll(res, (long)(ufds), (long)(nfds),          \
                                      (long)(timeout))
-#define __sanitizer_syscall_pre_select(n, inp, outp, exp, tvp)              \
-  __sanitizer_syscall_pre_impl_select((long)(n), (long)(inp), (long)(outp), \
+#define __sanitizer_syscall_pre_select(n, inp, outp, exp, tvp)                 \
+  __sanitizer_syscall_pre_impl_select((long)(n), (long)(inp), (long)(outp),    \
                                       (long)(exp), (long)(tvp))
-#define __sanitizer_syscall_post_select(res, n, inp, outp, exp, tvp) \
-  __sanitizer_syscall_post_impl_select(res, (long)(n), (long)(inp),  \
+#define __sanitizer_syscall_post_select(res, n, inp, outp, exp, tvp)           \
+  __sanitizer_syscall_post_impl_select(res, (long)(n), (long)(inp),            \
                                        (long)(outp), (long)(exp), (long)(tvp))
-#define __sanitizer_syscall_pre_old_select(arg) \
+#define __sanitizer_syscall_pre_old_select(arg)                                \
   __sanitizer_syscall_pre_impl_old_select((long)(arg))
-#define __sanitizer_syscall_post_old_select(res, arg) \
+#define __sanitizer_syscall_post_old_select(res, arg)                          \
   __sanitizer_syscall_post_impl_old_select(res, (long)(arg))
-#define __sanitizer_syscall_pre_epoll_create(size) \
+#define __sanitizer_syscall_pre_epoll_create(size)                             \
   __sanitizer_syscall_pre_impl_epoll_create((long)(size))
-#define __sanitizer_syscall_post_epoll_create(res, size) \
+#define __sanitizer_syscall_post_epoll_create(res, size)                       \
   __sanitizer_syscall_post_impl_epoll_create(res, (long)(size))
-#define __sanitizer_syscall_pre_epoll_create1(flags) \
+#define __sanitizer_syscall_pre_epoll_create1(flags)                           \
   __sanitizer_syscall_pre_impl_epoll_create1((long)(flags))
-#define __sanitizer_syscall_post_epoll_create1(res, flags) \
+#define __sanitizer_syscall_post_epoll_create1(res, flags)                     \
   __sanitizer_syscall_post_impl_epoll_create1(res, (long)(flags))
 #define __sanitizer_syscall_pre_epoll_ctl(epfd, op, fd, event)                 \
   __sanitizer_syscall_pre_impl_epoll_ctl((long)(epfd), (long)(op), (long)(fd), \
                                          (long)(event))
-#define __sanitizer_syscall_post_epoll_ctl(res, epfd, op, fd, event)     \
-  __sanitizer_syscall_post_impl_epoll_ctl(res, (long)(epfd), (long)(op), \
+#define __sanitizer_syscall_post_epoll_ctl(res, epfd, op, fd, event)           \
+  __sanitizer_syscall_post_impl_epoll_ctl(res, (long)(epfd), (long)(op),       \
                                           (long)(fd), (long)(event))
-#define __sanitizer_syscall_pre_epoll_wait(epfd, events, maxevents, timeout) \
-  __sanitizer_syscall_pre_impl_epoll_wait((long)(epfd), (long)(events),      \
+#define __sanitizer_syscall_pre_epoll_wait(epfd, events, maxevents, timeout)   \
+  __sanitizer_syscall_pre_impl_epoll_wait((long)(epfd), (long)(events),        \
                                           (long)(maxevents), (long)(timeout))
-#define __sanitizer_syscall_post_epoll_wait(res, epfd, events, maxevents,     \
-                                            timeout)                          \
-  __sanitizer_syscall_post_impl_epoll_wait(res, (long)(epfd), (long)(events), \
+#define __sanitizer_syscall_post_epoll_wait(res, epfd, events, maxevents,      \
+                                            timeout)                           \
+  __sanitizer_syscall_post_impl_epoll_wait(res, (long)(epfd), (long)(events),  \
                                            (long)(maxevents), (long)(timeout))
-#define __sanitizer_syscall_pre_epoll_pwait(epfd, events, maxevents, timeout, \
-                                            sigmask, sigsetsize)              \
-  __sanitizer_syscall_pre_impl_epoll_pwait(                                   \
-      (long)(epfd), (long)(events), (long)(maxevents), (long)(timeout),       \
+#define __sanitizer_syscall_pre_epoll_pwait(epfd, events, maxevents, timeout,  \
+                                            sigmask, sigsetsize)               \
+  __sanitizer_syscall_pre_impl_epoll_pwait(                                    \
+      (long)(epfd), (long)(events), (long)(maxevents), (long)(timeout),        \
+      (long)(sigmask), (long)(sigsetsize))
+#define __sanitizer_syscall_post_epoll_pwait(res, epfd, events, maxevents,     \
+                                             timeout, sigmask, sigsetsize)     \
+  __sanitizer_syscall_post_impl_epoll_pwait(                                   \
+      res, (long)(epfd), (long)(events), (long)(maxevents), (long)(timeout),   \
       (long)(sigmask), (long)(sigsetsize))
-#define __sanitizer_syscall_post_epoll_pwait(res, epfd, events, maxevents,   \
-                                             timeout, sigmask, sigsetsize)   \
-  __sanitizer_syscall_post_impl_epoll_pwait(                                 \
-      res, (long)(epfd), (long)(events), (long)(maxevents), (long)(timeout), \
+#define __sanitizer_syscall_pre_epoll_pwait2(epfd, events, maxevents, timeout, \
+                                             sigmask, sigsetsize)              \
+  __sanitizer_syscall_pre_impl_epoll_pwait2(                                   \
+      (long)(epfd), (long)(events), (long)(maxevents), (long)(timeout),        \
       (long)(sigmask), (long)(sigsetsize))
-#define __sanitizer_syscall_pre_gethostname(name, len) \
+#define __sanitizer_syscall_post_epoll_pwait2(res, epfd, events, maxevents,    \
+                                              timeout, sigmask, sigsetsize)    \
+  __sanitizer_syscall_post_impl_epoll_pwait2(                                  \
+      res, (long)(epfd), (long)(events), (long)(maxevents), (long)(timeout),   \
+      (long)(sigmask), (long)(sigsetsize))
+#define __sanitizer_syscall_pre_gethostname(name, len)                         \
   __sanitizer_syscall_pre_impl_gethostname((long)(name), (long)(len))
-#define __sanitizer_syscall_post_gethostname(res, name, len) \
+#define __sanitizer_syscall_post_gethostname(res, name, len)                   \
   __sanitizer_syscall_post_impl_gethostname(res, (long)(name), (long)(len))
-#define __sanitizer_syscall_pre_sethostname(name, len) \
+#define __sanitizer_syscall_pre_sethostname(name, len)                         \
   __sanitizer_syscall_pre_impl_sethostname((long)(name), (long)(len))
-#define __sanitizer_syscall_post_sethostname(res, name, len) \
+#define __sanitizer_syscall_post_sethostname(res, name, len)                   \
   __sanitizer_syscall_post_impl_sethostname(res, (long)(name), (long)(len))
-#define __sanitizer_syscall_pre_setdomainname(name, len) \
+#define __sanitizer_syscall_pre_setdomainname(name, len)                       \
   __sanitizer_syscall_pre_impl_setdomainname((long)(name), (long)(len))
-#define __sanitizer_syscall_post_setdomainname(res, name, len) \
+#define __sanitizer_syscall_post_setdomainname(res, name, len)                 \
   __sanitizer_syscall_post_impl_setdomainname(res, (long)(name), (long)(len))
-#define __sanitizer_syscall_pre_newuname(name) \
+#define __sanitizer_syscall_pre_newuname(name)                                 \
   __sanitizer_syscall_pre_impl_newuname((long)(name))
-#define __sanitizer_syscall_post_newuname(res, name) \
+#define __sanitizer_syscall_post_newuname(res, name)                           \
   __sanitizer_syscall_post_impl_newuname(res, (long)(name))
-#define __sanitizer_syscall_pre_uname(arg0) \
+#define __sanitizer_syscall_pre_uname(arg0)                                    \
   __sanitizer_syscall_pre_impl_uname((long)(arg0))
-#define __sanitizer_syscall_post_uname(res, arg0) \
+#define __sanitizer_syscall_post_uname(res, arg0)                              \
   __sanitizer_syscall_post_impl_uname(res, (long)(arg0))
-#define __sanitizer_syscall_pre_olduname(arg0) \
+#define __sanitizer_syscall_pre_olduname(arg0)                                 \
   __sanitizer_syscall_pre_impl_olduname((long)(arg0))
-#define __sanitizer_syscall_post_olduname(res, arg0) \
+#define __sanitizer_syscall_post_olduname(res, arg0)                           \
   __sanitizer_syscall_post_impl_olduname(res, (long)(arg0))
-#define __sanitizer_syscall_pre_getrlimit(resource, rlim) \
+#define __sanitizer_syscall_pre_getrlimit(resource, rlim)                      \
   __sanitizer_syscall_pre_impl_getrlimit((long)(resource), (long)(rlim))
-#define __sanitizer_syscall_post_getrlimit(res, resource, rlim) \
+#define __sanitizer_syscall_post_getrlimit(res, resource, rlim)                \
   __sanitizer_syscall_post_impl_getrlimit(res, (long)(resource), (long)(rlim))
-#define __sanitizer_syscall_pre_old_getrlimit(resource, rlim) \
+#define __sanitizer_syscall_pre_old_getrlimit(resource, rlim)                  \
   __sanitizer_syscall_pre_impl_old_getrlimit((long)(resource), (long)(rlim))
-#define __sanitizer_syscall_post_old_getrlimit(res, resource, rlim)  \
-  __sanitizer_syscall_post_impl_old_getrlimit(res, (long)(resource), \
+#define __sanitizer_syscall_post_old_getrlimit(res, resource, rlim)            \
+  __sanitizer_syscall_post_impl_old_getrlimit(res, (long)(resource),           \
                                               (long)(rlim))
-#define __sanitizer_syscall_pre_setrlimit(resource, rlim) \
+#define __sanitizer_syscall_pre_setrlimit(resource, rlim)                      \
   __sanitizer_syscall_pre_impl_setrlimit((long)(resource), (long)(rlim))
-#define __sanitizer_syscall_post_setrlimit(res, resource, rlim) \
+#define __sanitizer_syscall_post_setrlimit(res, resource, rlim)                \
   __sanitizer_syscall_post_impl_setrlimit(res, (long)(resource), (long)(rlim))
-#define __sanitizer_syscall_pre_prlimit64(pid, resource, new_rlim, old_rlim) \
-  __sanitizer_syscall_pre_impl_prlimit64((long)(pid), (long)(resource),      \
+#define __sanitizer_syscall_pre_prlimit64(pid, resource, new_rlim, old_rlim)   \
+  __sanitizer_syscall_pre_impl_prlimit64((long)(pid), (long)(resource),        \
                                          (long)(new_rlim), (long)(old_rlim))
-#define __sanitizer_syscall_post_prlimit64(res, pid, resource, new_rlim,      \
-                                           old_rlim)                          \
-  __sanitizer_syscall_post_impl_prlimit64(res, (long)(pid), (long)(resource), \
+#define __sanitizer_syscall_post_prlimit64(res, pid, resource, new_rlim,       \
+                                           old_rlim)                           \
+  __sanitizer_syscall_post_impl_prlimit64(res, (long)(pid), (long)(resource),  \
                                           (long)(new_rlim), (long)(old_rlim))
-#define __sanitizer_syscall_pre_getrusage(who, ru) \
+#define __sanitizer_syscall_pre_getrusage(who, ru)                             \
   __sanitizer_syscall_pre_impl_getrusage((long)(who), (long)(ru))
-#define __sanitizer_syscall_post_getrusage(res, who, ru) \
+#define __sanitizer_syscall_post_getrusage(res, who, ru)                       \
   __sanitizer_syscall_post_impl_getrusage(res, (long)(who), (long)(ru))
-#define __sanitizer_syscall_pre_umask(mask) \
+#define __sanitizer_syscall_pre_umask(mask)                                    \
   __sanitizer_syscall_pre_impl_umask((long)(mask))
-#define __sanitizer_syscall_post_umask(res, mask) \
+#define __sanitizer_syscall_post_umask(res, mask)                              \
   __sanitizer_syscall_post_impl_umask(res, (long)(mask))
-#define __sanitizer_syscall_pre_msgget(key, msgflg) \
+#define __sanitizer_syscall_pre_msgget(key, msgflg)                            \
   __sanitizer_syscall_pre_impl_msgget((long)(key), (long)(msgflg))
-#define __sanitizer_syscall_post_msgget(res, key, msgflg) \
+#define __sanitizer_syscall_post_msgget(res, key, msgflg)                      \
   __sanitizer_syscall_post_impl_msgget(res, (long)(key), (long)(msgflg))
-#define __sanitizer_syscall_pre_msgsnd(msqid, msgp, msgsz, msgflg) \
-  __sanitizer_syscall_pre_impl_msgsnd((long)(msqid), (long)(msgp), \
+#define __sanitizer_syscall_pre_msgsnd(msqid, msgp, msgsz, msgflg)             \
+  __sanitizer_syscall_pre_impl_msgsnd((long)(msqid), (long)(msgp),             \
                                       (long)(msgsz), (long)(msgflg))
-#define __sanitizer_syscall_post_msgsnd(res, msqid, msgp, msgsz, msgflg) \
-  __sanitizer_syscall_post_impl_msgsnd(res, (long)(msqid), (long)(msgp), \
+#define __sanitizer_syscall_post_msgsnd(res, msqid, msgp, msgsz, msgflg)       \
+  __sanitizer_syscall_post_impl_msgsnd(res, (long)(msqid), (long)(msgp),       \
                                        (long)(msgsz), (long)(msgflg))
-#define __sanitizer_syscall_pre_msgrcv(msqid, msgp, msgsz, msgtyp, msgflg) \
-  __sanitizer_syscall_pre_impl_msgrcv((long)(msqid), (long)(msgp),         \
-                                      (long)(msgsz), (long)(msgtyp),       \
+#define __sanitizer_syscall_pre_msgrcv(msqid, msgp, msgsz, msgtyp, msgflg)     \
+  __sanitizer_syscall_pre_impl_msgrcv((long)(msqid), (long)(msgp),             \
+                                      (long)(msgsz), (long)(msgtyp),           \
                                       (long)(msgflg))
-#define __sanitizer_syscall_post_msgrcv(res, msqid, msgp, msgsz, msgtyp, \
-                                        msgflg)                          \
-  __sanitizer_syscall_post_impl_msgrcv(res, (long)(msqid), (long)(msgp), \
-                                       (long)(msgsz), (long)(msgtyp),    \
+#define __sanitizer_syscall_post_msgrcv(res, msqid, msgp, msgsz, msgtyp,       \
+                                        msgflg)                                \
+  __sanitizer_syscall_post_impl_msgrcv(res, (long)(msqid), (long)(msgp),       \
+                                       (long)(msgsz), (long)(msgtyp),          \
                                        (long)(msgflg))
-#define __sanitizer_syscall_pre_msgctl(msqid, cmd, buf) \
+#define __sanitizer_syscall_pre_msgctl(msqid, cmd, buf)                        \
   __sanitizer_syscall_pre_impl_msgctl((long)(msqid), (long)(cmd), (long)(buf))
-#define __sanitizer_syscall_post_msgctl(res, msqid, cmd, buf)           \
-  __sanitizer_syscall_post_impl_msgctl(res, (long)(msqid), (long)(cmd), \
+#define __sanitizer_syscall_post_msgctl(res, msqid, cmd, buf)                  \
+  __sanitizer_syscall_post_impl_msgctl(res, (long)(msqid), (long)(cmd),        \
                                        (long)(buf))
-#define __sanitizer_syscall_pre_semget(key, nsems, semflg)        \
-  __sanitizer_syscall_pre_impl_semget((long)(key), (long)(nsems), \
+#define __sanitizer_syscall_pre_semget(key, nsems, semflg)                     \
+  __sanitizer_syscall_pre_impl_semget((long)(key), (long)(nsems),              \
                                       (long)(semflg))
-#define __sanitizer_syscall_post_semget(res, key, nsems, semflg)        \
-  __sanitizer_syscall_post_impl_semget(res, (long)(key), (long)(nsems), \
+#define __sanitizer_syscall_post_semget(res, key, nsems, semflg)               \
+  __sanitizer_syscall_post_impl_semget(res, (long)(key), (long)(nsems),        \
                                        (long)(semflg))
-#define __sanitizer_syscall_pre_semop(semid, sops, nsops) \
+#define __sanitizer_syscall_pre_semop(semid, sops, nsops)                      \
   __sanitizer_syscall_pre_impl_semop((long)(semid), (long)(sops), (long)(nsops))
-#define __sanitizer_syscall_post_semop(res, semid, sops, nsops)         \
-  __sanitizer_syscall_post_impl_semop(res, (long)(semid), (long)(sops), \
+#define __sanitizer_syscall_post_semop(res, semid, sops, nsops)                \
+  __sanitizer_syscall_post_impl_semop(res, (long)(semid), (long)(sops),        \
                                       (long)(nsops))
-#define __sanitizer_syscall_pre_semctl(semid, semnum, cmd, arg)      \
-  __sanitizer_syscall_pre_impl_semctl((long)(semid), (long)(semnum), \
+#define __sanitizer_syscall_pre_semctl(semid, semnum, cmd, arg)                \
+  __sanitizer_syscall_pre_impl_semctl((long)(semid), (long)(semnum),           \
                                       (long)(cmd), (long)(arg))
-#define __sanitizer_syscall_post_semctl(res, semid, semnum, cmd, arg)      \
-  __sanitizer_syscall_post_impl_semctl(res, (long)(semid), (long)(semnum), \
+#define __sanitizer_syscall_post_semctl(res, semid, semnum, cmd, arg)          \
+  __sanitizer_syscall_post_impl_semctl(res, (long)(semid), (long)(semnum),     \
                                        (long)(cmd), (long)(arg))
-#define __sanitizer_syscall_pre_semtimedop(semid, sops, nsops, timeout) \
-  __sanitizer_syscall_pre_impl_semtimedop((long)(semid), (long)(sops),  \
+#define __sanitizer_syscall_pre_semtimedop(semid, sops, nsops, timeout)        \
+  __sanitizer_syscall_pre_impl_semtimedop((long)(semid), (long)(sops),         \
                                           (long)(nsops), (long)(timeout))
-#define __sanitizer_syscall_post_semtimedop(res, semid, sops, nsops, timeout) \
-  __sanitizer_syscall_post_impl_semtimedop(res, (long)(semid), (long)(sops),  \
+#define __sanitizer_syscall_post_semtimedop(res, semid, sops, nsops, timeout)  \
+  __sanitizer_syscall_post_impl_semtimedop(res, (long)(semid), (long)(sops),   \
                                            (long)(nsops), (long)(timeout))
-#define __sanitizer_syscall_pre_shmat(shmid, shmaddr, shmflg)        \
-  __sanitizer_syscall_pre_impl_shmat((long)(shmid), (long)(shmaddr), \
+#define __sanitizer_syscall_pre_shmat(shmid, shmaddr, shmflg)                  \
+  __sanitizer_syscall_pre_impl_shmat((long)(shmid), (long)(shmaddr),           \
                                      (long)(shmflg))
-#define __sanitizer_syscall_post_shmat(res, shmid, shmaddr, shmflg)        \
-  __sanitizer_syscall_post_impl_shmat(res, (long)(shmid), (long)(shmaddr), \
+#define __sanitizer_syscall_post_shmat(res, shmid, shmaddr, shmflg)            \
+  __sanitizer_syscall_post_impl_shmat(res, (long)(shmid), (long)(shmaddr),     \
                                       (long)(shmflg))
-#define __sanitizer_syscall_pre_shmget(key, size, flag) \
+#define __sanitizer_syscall_pre_shmget(key, size, flag)                        \
   __sanitizer_syscall_pre_impl_shmget((long)(key), (long)(size), (long)(flag))
-#define __sanitizer_syscall_post_shmget(res, key, size, flag)          \
-  __sanitizer_syscall_post_impl_shmget(res, (long)(key), (long)(size), \
+#define __sanitizer_syscall_post_shmget(res, key, size, flag)                  \
+  __sanitizer_syscall_post_impl_shmget(res, (long)(key), (long)(size),         \
                                        (long)(flag))
-#define __sanitizer_syscall_pre_shmdt(shmaddr) \
+#define __sanitizer_syscall_pre_shmdt(shmaddr)                                 \
   __sanitizer_syscall_pre_impl_shmdt((long)(shmaddr))
-#define __sanitizer_syscall_post_shmdt(res, shmaddr) \
+#define __sanitizer_syscall_post_shmdt(res, shmaddr)                           \
   __sanitizer_syscall_post_impl_shmdt(res, (long)(shmaddr))
-#define __sanitizer_syscall_pre_shmctl(shmid, cmd, buf) \
+#define __sanitizer_syscall_pre_shmctl(shmid, cmd, buf)                        \
   __sanitizer_syscall_pre_impl_shmctl((long)(shmid), (long)(cmd), (long)(buf))
-#define __sanitizer_syscall_post_shmctl(res, shmid, cmd, buf)           \
-  __sanitizer_syscall_post_impl_shmctl(res, (long)(shmid), (long)(cmd), \
+#define __sanitizer_syscall_post_shmctl(res, shmid, cmd, buf)                  \
+  __sanitizer_syscall_post_impl_shmctl(res, (long)(shmid), (long)(cmd),        \
                                        (long)(buf))
 #define __sanitizer_syscall_pre_ipc(call, first, second, third, ptr, fifth)    \
   __sanitizer_syscall_pre_impl_ipc((long)(call), (long)(first),                \
                                    (long)(second), (long)(third), (long)(ptr), \
                                    (long)(fifth))
-#define __sanitizer_syscall_post_ipc(res, call, first, second, third, ptr, \
-                                     fifth)                                \
-  __sanitizer_syscall_post_impl_ipc(res, (long)(call), (long)(first),      \
-                                    (long)(second), (long)(third),         \
+#define __sanitizer_syscall_post_ipc(res, call, first, second, third, ptr,     \
+                                     fifth)                                    \
+  __sanitizer_syscall_post_impl_ipc(res, (long)(call), (long)(first),          \
+                                    (long)(second), (long)(third),             \
                                     (long)(ptr), (long)(fifth))
-#define __sanitizer_syscall_pre_mq_open(name, oflag, mode, attr)    \
-  __sanitizer_syscall_pre_impl_mq_open((long)(name), (long)(oflag), \
+#define __sanitizer_syscall_pre_mq_open(name, oflag, mode, attr)               \
+  __sanitizer_syscall_pre_impl_mq_open((long)(name), (long)(oflag),            \
                                        (long)(mode), (long)(attr))
-#define __sanitizer_syscall_post_mq_open(res, name, oflag, mode, attr)    \
-  __sanitizer_syscall_post_impl_mq_open(res, (long)(name), (long)(oflag), \
+#define __sanitizer_syscall_post_mq_open(res, name, oflag, mode, attr)         \
+  __sanitizer_syscall_post_impl_mq_open(res, (long)(name), (long)(oflag),      \
                                         (long)(mode), (long)(attr))
-#define __sanitizer_syscall_pre_mq_unlink(name) \
+#define __sanitizer_syscall_pre_mq_unlink(name)                                \
   __sanitizer_syscall_pre_impl_mq_unlink((long)(name))
-#define __sanitizer_syscall_post_mq_unlink(res, name) \
+#define __sanitizer_syscall_post_mq_unlink(res, name)                          \
   __sanitizer_syscall_post_impl_mq_unlink(res, (long)(name))
 #define __sanitizer_syscall_pre_mq_timedsend(mqdes, msg_ptr, msg_len,          \
                                              msg_prio, abs_timeout)            \
   __sanitizer_syscall_pre_impl_mq_timedsend((long)(mqdes), (long)(msg_ptr),    \
                                             (long)(msg_len), (long)(msg_prio), \
                                             (long)(abs_timeout))
-#define __sanitizer_syscall_post_mq_timedsend(res, mqdes, msg_ptr, msg_len,   \
-                                              msg_prio, abs_timeout)          \
-  __sanitizer_syscall_post_impl_mq_timedsend(                                 \
-      res, (long)(mqdes), (long)(msg_ptr), (long)(msg_len), (long)(msg_prio), \
+#define __sanitizer_syscall_post_mq_timedsend(res, mqdes, msg_ptr, msg_len,    \
+                                              msg_prio, abs_timeout)           \
+  __sanitizer_syscall_post_impl_mq_timedsend(                                  \
+      res, (long)(mqdes), (long)(msg_ptr), (long)(msg_len), (long)(msg_prio),  \
       (long)(abs_timeout))
-#define __sanitizer_syscall_pre_mq_timedreceive(mqdes, msg_ptr, msg_len, \
-                                                msg_prio, abs_timeout)   \
-  __sanitizer_syscall_pre_impl_mq_timedreceive(                          \
-      (long)(mqdes), (long)(msg_ptr), (long)(msg_len), (long)(msg_prio), \
+#define __sanitizer_syscall_pre_mq_timedreceive(mqdes, msg_ptr, msg_len,       \
+                                                msg_prio, abs_timeout)         \
+  __sanitizer_syscall_pre_impl_mq_timedreceive(                                \
+      (long)(mqdes), (long)(msg_ptr), (long)(msg_len), (long)(msg_prio),       \
       (long)(abs_timeout))
 #define __sanitizer_syscall_post_mq_timedreceive(res, mqdes, msg_ptr, msg_len, \
                                                  msg_prio, abs_timeout)        \
   __sanitizer_syscall_post_impl_mq_timedreceive(                               \
       res, (long)(mqdes), (long)(msg_ptr), (long)(msg_len), (long)(msg_prio),  \
       (long)(abs_timeout))
-#define __sanitizer_syscall_pre_mq_notify(mqdes, notification) \
+#define __sanitizer_syscall_pre_mq_notify(mqdes, notification)                 \
   __sanitizer_syscall_pre_impl_mq_notify((long)(mqdes), (long)(notification))
-#define __sanitizer_syscall_post_mq_notify(res, mqdes, notification) \
-  __sanitizer_syscall_post_impl_mq_notify(res, (long)(mqdes),        \
+#define __sanitizer_syscall_post_mq_notify(res, mqdes, notification)           \
+  __sanitizer_syscall_post_impl_mq_notify(res, (long)(mqdes),                  \
                                           (long)(notification))
-#define __sanitizer_syscall_pre_mq_getsetattr(mqdes, mqstat, omqstat)       \
-  __sanitizer_syscall_pre_impl_mq_getsetattr((long)(mqdes), (long)(mqstat), \
+#define __sanitizer_syscall_pre_mq_getsetattr(mqdes, mqstat, omqstat)          \
+  __sanitizer_syscall_pre_impl_mq_getsetattr((long)(mqdes), (long)(mqstat),    \
                                              (long)(omqstat))
-#define __sanitizer_syscall_post_mq_getsetattr(res, mqdes, mqstat, omqstat) \
-  __sanitizer_syscall_post_impl_mq_getsetattr(res, (long)(mqdes),           \
+#define __sanitizer_syscall_post_mq_getsetattr(res, mqdes, mqstat, omqstat)    \
+  __sanitizer_syscall_post_impl_mq_getsetattr(res, (long)(mqdes),              \
                                               (long)(mqstat), (long)(omqstat))
-#define __sanitizer_syscall_pre_pciconfig_iobase(which, bus, devfn)         \
-  __sanitizer_syscall_pre_impl_pciconfig_iobase((long)(which), (long)(bus), \
+#define __sanitizer_syscall_pre_pciconfig_iobase(which, bus, devfn)            \
+  __sanitizer_syscall_pre_impl_pciconfig_iobase((long)(which), (long)(bus),    \
                                                 (long)(devfn))
-#define __sanitizer_syscall_post_pciconfig_iobase(res, which, bus, devfn) \
-  __sanitizer_syscall_post_impl_pciconfig_iobase(res, (long)(which),      \
+#define __sanitizer_syscall_post_pciconfig_iobase(res, which, bus, devfn)      \
+  __sanitizer_syscall_post_impl_pciconfig_iobase(res, (long)(which),           \
                                                  (long)(bus), (long)(devfn))
-#define __sanitizer_syscall_pre_pciconfig_read(bus, dfn, off, len, buf) \
-  __sanitizer_syscall_pre_impl_pciconfig_read(                          \
+#define __sanitizer_syscall_pre_pciconfig_read(bus, dfn, off, len, buf)        \
+  __sanitizer_syscall_pre_impl_pciconfig_read(                                 \
       (long)(bus), (long)(dfn), (long)(off), (long)(len), (long)(buf))
-#define __sanitizer_syscall_post_pciconfig_read(res, bus, dfn, off, len, buf) \
-  __sanitizer_syscall_post_impl_pciconfig_read(                               \
+#define __sanitizer_syscall_post_pciconfig_read(res, bus, dfn, off, len, buf)  \
+  __sanitizer_syscall_post_impl_pciconfig_read(                                \
       res, (long)(bus), (long)(dfn), (long)(off), (long)(len), (long)(buf))
-#define __sanitizer_syscall_pre_pciconfig_write(bus, dfn, off, len, buf) \
-  __sanitizer_syscall_pre_impl_pciconfig_write(                          \
+#define __sanitizer_syscall_pre_pciconfig_write(bus, dfn, off, len, buf)       \
+  __sanitizer_syscall_pre_impl_pciconfig_write(                                \
       (long)(bus), (long)(dfn), (long)(off), (long)(len), (long)(buf))
 #define __sanitizer_syscall_post_pciconfig_write(res, bus, dfn, off, len, buf) \
   __sanitizer_syscall_post_impl_pciconfig_write(                               \
       res, (long)(bus), (long)(dfn), (long)(off), (long)(len), (long)(buf))
-#define __sanitizer_syscall_pre_swapon(specialfile, swap_flags) \
+#define __sanitizer_syscall_pre_swapon(specialfile, swap_flags)                \
   __sanitizer_syscall_pre_impl_swapon((long)(specialfile), (long)(swap_flags))
-#define __sanitizer_syscall_post_swapon(res, specialfile, swap_flags) \
-  __sanitizer_syscall_post_impl_swapon(res, (long)(specialfile),      \
+#define __sanitizer_syscall_post_swapon(res, specialfile, swap_flags)          \
+  __sanitizer_syscall_post_impl_swapon(res, (long)(specialfile),               \
                                        (long)(swap_flags))
-#define __sanitizer_syscall_pre_swapoff(specialfile) \
+#define __sanitizer_syscall_pre_swapoff(specialfile)                           \
   __sanitizer_syscall_pre_impl_swapoff((long)(specialfile))
-#define __sanitizer_syscall_post_swapoff(res, specialfile) \
+#define __sanitizer_syscall_post_swapoff(res, specialfile)                     \
   __sanitizer_syscall_post_impl_swapoff(res, (long)(specialfile))
-#define __sanitizer_syscall_pre_sysctl(args) \
+#define __sanitizer_syscall_pre_sysctl(args)                                   \
   __sanitizer_syscall_pre_impl_sysctl((long)(args))
-#define __sanitizer_syscall_post_sysctl(res, args) \
+#define __sanitizer_syscall_post_sysctl(res, args)                             \
   __sanitizer_syscall_post_impl_sysctl(res, (long)(args))
-#define __sanitizer_syscall_pre_sysinfo(info) \
+#define __sanitizer_syscall_pre_sysinfo(info)                                  \
   __sanitizer_syscall_pre_impl_sysinfo((long)(info))
-#define __sanitizer_syscall_post_sysinfo(res, info) \
+#define __sanitizer_syscall_post_sysinfo(res, info)                            \
   __sanitizer_syscall_post_impl_sysinfo(res, (long)(info))
-#define __sanitizer_syscall_pre_sysfs(option, arg1, arg2) \
+#define __sanitizer_syscall_pre_sysfs(option, arg1, arg2)                      \
   __sanitizer_syscall_pre_impl_sysfs((long)(option), (long)(arg1), (long)(arg2))
-#define __sanitizer_syscall_post_sysfs(res, option, arg1, arg2)          \
-  __sanitizer_syscall_post_impl_sysfs(res, (long)(option), (long)(arg1), \
+#define __sanitizer_syscall_post_sysfs(res, option, arg1, arg2)                \
+  __sanitizer_syscall_post_impl_sysfs(res, (long)(option), (long)(arg1),       \
                                       (long)(arg2))
-#define __sanitizer_syscall_pre_syslog(type, buf, len) \
+#define __sanitizer_syscall_pre_syslog(type, buf, len)                         \
   __sanitizer_syscall_pre_impl_syslog((long)(type), (long)(buf), (long)(len))
-#define __sanitizer_syscall_post_syslog(res, type, buf, len)           \
-  __sanitizer_syscall_post_impl_syslog(res, (long)(type), (long)(buf), \
+#define __sanitizer_syscall_post_syslog(res, type, buf, len)                   \
+  __sanitizer_syscall_post_impl_syslog(res, (long)(type), (long)(buf),         \
                                        (long)(len))
-#define __sanitizer_syscall_pre_uselib(library) \
+#define __sanitizer_syscall_pre_uselib(library)                                \
   __sanitizer_syscall_pre_impl_uselib((long)(library))
-#define __sanitizer_syscall_post_uselib(res, library) \
+#define __sanitizer_syscall_post_uselib(res, library)                          \
   __sanitizer_syscall_post_impl_uselib(res, (long)(library))
-#define __sanitizer_syscall_pre_ni_syscall() \
+#define __sanitizer_syscall_pre_ni_syscall()                                   \
   __sanitizer_syscall_pre_impl_ni_syscall()
-#define __sanitizer_syscall_post_ni_syscall(res) \
+#define __sanitizer_syscall_post_ni_syscall(res)                               \
   __sanitizer_syscall_post_impl_ni_syscall(res)
-#define __sanitizer_syscall_pre_ptrace(request, pid, addr, data)    \
-  __sanitizer_syscall_pre_impl_ptrace((long)(request), (long)(pid), \
+#define __sanitizer_syscall_pre_ptrace(request, pid, addr, data)               \
+  __sanitizer_syscall_pre_impl_ptrace((long)(request), (long)(pid),            \
                                       (long)(addr), (long)(data))
-#define __sanitizer_syscall_post_ptrace(res, request, pid, addr, data)    \
-  __sanitizer_syscall_post_impl_ptrace(res, (long)(request), (long)(pid), \
+#define __sanitizer_syscall_post_ptrace(res, request, pid, addr, data)         \
+  __sanitizer_syscall_post_impl_ptrace(res, (long)(request), (long)(pid),      \
                                        (long)(addr), (long)(data))
-#define __sanitizer_syscall_pre_add_key(_type, _description, _payload, plen, \
-                                        destringid)                          \
-  __sanitizer_syscall_pre_impl_add_key((long)(_type), (long)(_description),  \
-                                       (long)(_payload), (long)(plen),       \
+#define __sanitizer_syscall_pre_add_key(_type, _description, _payload, plen,   \
+                                        destringid)                            \
+  __sanitizer_syscall_pre_impl_add_key((long)(_type), (long)(_description),    \
+                                       (long)(_payload), (long)(plen),         \
                                        (long)(destringid))
-#define __sanitizer_syscall_post_add_key(res, _type, _description, _payload, \
-                                         plen, destringid)                   \
-  __sanitizer_syscall_post_impl_add_key(                                     \
-      res, (long)(_type), (long)(_description), (long)(_payload),            \
+#define __sanitizer_syscall_post_add_key(res, _type, _description, _payload,   \
+                                         plen, destringid)                     \
+  __sanitizer_syscall_post_impl_add_key(                                       \
+      res, (long)(_type), (long)(_description), (long)(_payload),              \
       (long)(plen), (long)(destringid))
-#define __sanitizer_syscall_pre_request_key(_type, _description,       \
-                                            _callout_info, destringid) \
-  __sanitizer_syscall_pre_impl_request_key(                            \
-      (long)(_type), (long)(_description), (long)(_callout_info),      \
+#define __sanitizer_syscall_pre_request_key(_type, _description,               \
+                                            _callout_info, destringid)         \
+  __sanitizer_syscall_pre_impl_request_key(                                    \
+      (long)(_type), (long)(_description), (long)(_callout_info),              \
       (long)(destringid))
-#define __sanitizer_syscall_post_request_key(res, _type, _description,  \
-                                             _callout_info, destringid) \
-  __sanitizer_syscall_post_impl_request_key(                            \
-      res, (long)(_type), (long)(_description), (long)(_callout_info),  \
+#define __sanitizer_syscall_post_request_key(res, _type, _description,         \
+                                             _callout_info, destringid)        \
+  __sanitizer_syscall_post_impl_request_key(                                   \
+      res, (long)(_type), (long)(_description), (long)(_callout_info),         \
       (long)(destringid))
 #define __sanitizer_syscall_pre_keyctl(cmd, arg2, arg3, arg4, arg5)            \
   __sanitizer_syscall_pre_impl_keyctl((long)(cmd), (long)(arg2), (long)(arg3), \
                                       (long)(arg4), (long)(arg5))
-#define __sanitizer_syscall_post_keyctl(res, cmd, arg2, arg3, arg4, arg5) \
-  __sanitizer_syscall_post_impl_keyctl(res, (long)(cmd), (long)(arg2),    \
-                                       (long)(arg3), (long)(arg4),        \
+#define __sanitizer_syscall_post_keyctl(res, cmd, arg2, arg3, arg4, arg5)      \
+  __sanitizer_syscall_post_impl_keyctl(res, (long)(cmd), (long)(arg2),         \
+                                       (long)(arg3), (long)(arg4),             \
                                        (long)(arg5))
-#define __sanitizer_syscall_pre_ioprio_set(which, who, ioprio)        \
-  __sanitizer_syscall_pre_impl_ioprio_set((long)(which), (long)(who), \
+#define __sanitizer_syscall_pre_ioprio_set(which, who, ioprio)                 \
+  __sanitizer_syscall_pre_impl_ioprio_set((long)(which), (long)(who),          \
                                           (long)(ioprio))
-#define __sanitizer_syscall_post_ioprio_set(res, which, who, ioprio)        \
-  __sanitizer_syscall_post_impl_ioprio_set(res, (long)(which), (long)(who), \
+#define __sanitizer_syscall_post_ioprio_set(res, which, who, ioprio)           \
+  __sanitizer_syscall_post_impl_ioprio_set(res, (long)(which), (long)(who),    \
                                            (long)(ioprio))
-#define __sanitizer_syscall_pre_ioprio_get(which, who) \
+#define __sanitizer_syscall_pre_ioprio_get(which, who)                         \
   __sanitizer_syscall_pre_impl_ioprio_get((long)(which), (long)(who))
-#define __sanitizer_syscall_post_ioprio_get(res, which, who) \
+#define __sanitizer_syscall_post_ioprio_get(res, which, who)                   \
   __sanitizer_syscall_post_impl_ioprio_get(res, (long)(which), (long)(who))
-#define __sanitizer_syscall_pre_set_mempolicy(mode, nmask, maxnode)       \
-  __sanitizer_syscall_pre_impl_set_mempolicy((long)(mode), (long)(nmask), \
+#define __sanitizer_syscall_pre_set_mempolicy(mode, nmask, maxnode)            \
+  __sanitizer_syscall_pre_impl_set_mempolicy((long)(mode), (long)(nmask),      \
                                              (long)(maxnode))
-#define __sanitizer_syscall_post_set_mempolicy(res, mode, nmask, maxnode) \
-  __sanitizer_syscall_post_impl_set_mempolicy(res, (long)(mode),          \
+#define __sanitizer_syscall_post_set_mempolicy(res, mode, nmask, maxnode)      \
+  __sanitizer_syscall_post_impl_set_mempolicy(res, (long)(mode),               \
                                               (long)(nmask), (long)(maxnode))
-#define __sanitizer_syscall_pre_migrate_pages(pid, maxnode, from, to)      \
-  __sanitizer_syscall_pre_impl_migrate_pages((long)(pid), (long)(maxnode), \
+#define __sanitizer_syscall_pre_migrate_pages(pid, maxnode, from, to)          \
+  __sanitizer_syscall_pre_impl_migrate_pages((long)(pid), (long)(maxnode),     \
                                              (long)(from), (long)(to))
-#define __sanitizer_syscall_post_migrate_pages(res, pid, maxnode, from, to) \
-  __sanitizer_syscall_post_impl_migrate_pages(                              \
+#define __sanitizer_syscall_post_migrate_pages(res, pid, maxnode, from, to)    \
+  __sanitizer_syscall_post_impl_migrate_pages(                                 \
       res, (long)(pid), (long)(maxnode), (long)(from), (long)(to))
-#define __sanitizer_syscall_pre_move_pages(pid, nr_pages, pages, nodes,  \
-                                           status, flags)                \
-  __sanitizer_syscall_pre_impl_move_pages((long)(pid), (long)(nr_pages), \
-                                          (long)(pages), (long)(nodes),  \
+#define __sanitizer_syscall_pre_move_pages(pid, nr_pages, pages, nodes,        \
+                                           status, flags)                      \
+  __sanitizer_syscall_pre_impl_move_pages((long)(pid), (long)(nr_pages),       \
+                                          (long)(pages), (long)(nodes),        \
                                           (long)(status), (long)(flags))
 #define __sanitizer_syscall_post_move_pages(res, pid, nr_pages, pages, nodes,  \
                                             status, flags)                     \
@@ -1517,322 +1526,320 @@
   __sanitizer_syscall_pre_impl_mbind((long)(start), (long)(len), (long)(mode), \
                                      (long)(nmask), (long)(maxnode),           \
                                      (long)(flags))
-#define __sanitizer_syscall_post_mbind(res, start, len, mode, nmask, maxnode, \
-                                       flags)                                 \
-  __sanitizer_syscall_post_impl_mbind(res, (long)(start), (long)(len),        \
-                                      (long)(mode), (long)(nmask),            \
+#define __sanitizer_syscall_post_mbind(res, start, len, mode, nmask, maxnode,  \
+                                       flags)                                  \
+  __sanitizer_syscall_post_impl_mbind(res, (long)(start), (long)(len),         \
+                                      (long)(mode), (long)(nmask),             \
                                       (long)(maxnode), (long)(flags))
-#define __sanitizer_syscall_pre_get_mempolicy(policy, nmask, maxnode, addr, \
-                                              flags)                        \
-  __sanitizer_syscall_pre_impl_get_mempolicy((long)(policy), (long)(nmask), \
-                                             (long)(maxnode), (long)(addr), \
+#define __sanitizer_syscall_pre_get_mempolicy(policy, nmask, maxnode, addr,    \
+                                              flags)                           \
+  __sanitizer_syscall_pre_impl_get_mempolicy((long)(policy), (long)(nmask),    \
+                                             (long)(maxnode), (long)(addr),    \
                                              (long)(flags))
-#define __sanitizer_syscall_post_get_mempolicy(res, policy, nmask, maxnode,   \
-                                               addr, flags)                   \
-  __sanitizer_syscall_post_impl_get_mempolicy(res, (long)(policy),            \
-                                              (long)(nmask), (long)(maxnode), \
+#define __sanitizer_syscall_post_get_mempolicy(res, policy, nmask, maxnode,    \
+                                               addr, flags)                    \
+  __sanitizer_syscall_post_impl_get_mempolicy(res, (long)(policy),             \
+                                              (long)(nmask), (long)(maxnode),  \
                                               (long)(addr), (long)(flags))
-#define __sanitizer_syscall_pre_inotify_init() \
+#define __sanitizer_syscall_pre_inotify_init()                                 \
   __sanitizer_syscall_pre_impl_inotify_init()
-#define __sanitizer_syscall_post_inotify_init(res) \
+#define __sanitizer_syscall_post_inotify_init(res)                             \
   __sanitizer_syscall_post_impl_inotify_init(res)
-#define __sanitizer_syscall_pre_inotify_init1(flags) \
+#define __sanitizer_syscall_pre_inotify_init1(flags)                           \
   __sanitizer_syscall_pre_impl_inotify_init1((long)(flags))
-#define __sanitizer_syscall_post_inotify_init1(res, flags) \
+#define __sanitizer_syscall_post_inotify_init1(res, flags)                     \
   __sanitizer_syscall_post_impl_inotify_init1(res, (long)(flags))
-#define __sanitizer_syscall_pre_inotify_add_watch(fd, path, mask)          \
-  __sanitizer_syscall_pre_impl_inotify_add_watch((long)(fd), (long)(path), \
+#define __sanitizer_syscall_pre_inotify_add_watch(fd, path, mask)              \
+  __sanitizer_syscall_pre_impl_inotify_add_watch((long)(fd), (long)(path),     \
                                                  (long)(mask))
-#define __sanitizer_syscall_post_inotify_add_watch(res, fd, path, mask) \
-  __sanitizer_syscall_post_impl_inotify_add_watch(res, (long)(fd),      \
+#define __sanitizer_syscall_post_inotify_add_watch(res, fd, path, mask)        \
+  __sanitizer_syscall_post_impl_inotify_add_watch(res, (long)(fd),             \
                                                   (long)(path), (long)(mask))
-#define __sanitizer_syscall_pre_inotify_rm_watch(fd, wd) \
+#define __sanitizer_syscall_pre_inotify_rm_watch(fd, wd)                       \
   __sanitizer_syscall_pre_impl_inotify_rm_watch((long)(fd), (long)(wd))
-#define __sanitizer_syscall_post_inotify_rm_watch(res, fd, wd) \
+#define __sanitizer_syscall_post_inotify_rm_watch(res, fd, wd)                 \
   __sanitizer_syscall_post_impl_inotify_rm_watch(res, (long)(fd), (long)(wd))
-#define __sanitizer_syscall_pre_spu_run(fd, unpc, ustatus)       \
-  __sanitizer_syscall_pre_impl_spu_run((long)(fd), (long)(unpc), \
+#define __sanitizer_syscall_pre_spu_run(fd, unpc, ustatus)                     \
+  __sanitizer_syscall_pre_impl_spu_run((long)(fd), (long)(unpc),               \
                                        (long)(ustatus))
-#define __sanitizer_syscall_post_spu_run(res, fd, unpc, ustatus)       \
-  __sanitizer_syscall_post_impl_spu_run(res, (long)(fd), (long)(unpc), \
+#define __sanitizer_syscall_post_spu_run(res, fd, unpc, ustatus)               \
+  __sanitizer_syscall_post_impl_spu_run(res, (long)(fd), (long)(unpc),         \
                                         (long)(ustatus))
-#define __sanitizer_syscall_pre_spu_create(name, flags, mode, fd)      \
-  __sanitizer_syscall_pre_impl_spu_create((long)(name), (long)(flags), \
+#define __sanitizer_syscall_pre_spu_create(name, flags, mode, fd)              \
+  __sanitizer_syscall_pre_impl_spu_create((long)(name), (long)(flags),         \
                                           (long)(mode), (long)(fd))
-#define __sanitizer_syscall_post_spu_create(res, name, flags, mode, fd)      \
-  __sanitizer_syscall_post_impl_spu_create(res, (long)(name), (long)(flags), \
+#define __sanitizer_syscall_post_spu_create(res, name, flags, mode, fd)        \
+  __sanitizer_syscall_post_impl_spu_create(res, (long)(name), (long)(flags),   \
                                            (long)(mode), (long)(fd))
-#define __sanitizer_syscall_pre_mknodat(dfd, filename, mode, dev)     \
-  __sanitizer_syscall_pre_impl_mknodat((long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_pre_mknodat(dfd, filename, mode, dev)              \
+  __sanitizer_syscall_pre_impl_mknodat((long)(dfd), (long)(filename),          \
                                        (long)(mode), (long)(dev))
-#define __sanitizer_syscall_post_mknodat(res, dfd, filename, mode, dev)     \
-  __sanitizer_syscall_post_impl_mknodat(res, (long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_post_mknodat(res, dfd, filename, mode, dev)        \
+  __sanitizer_syscall_post_impl_mknodat(res, (long)(dfd), (long)(filename),    \
                                         (long)(mode), (long)(dev))
-#define __sanitizer_syscall_pre_mkdirat(dfd, pathname, mode)          \
-  __sanitizer_syscall_pre_impl_mkdirat((long)(dfd), (long)(pathname), \
+#define __sanitizer_syscall_pre_mkdirat(dfd, pathname, mode)                   \
+  __sanitizer_syscall_pre_impl_mkdirat((long)(dfd), (long)(pathname),          \
                                        (long)(mode))
-#define __sanitizer_syscall_post_mkdirat(res, dfd, pathname, mode)          \
-  __sanitizer_syscall_post_impl_mkdirat(res, (long)(dfd), (long)(pathname), \
+#define __sanitizer_syscall_post_mkdirat(res, dfd, pathname, mode)             \
+  __sanitizer_syscall_post_impl_mkdirat(res, (long)(dfd), (long)(pathname),    \
                                         (long)(mode))
-#define __sanitizer_syscall_pre_unlinkat(dfd, pathname, flag)          \
-  __sanitizer_syscall_pre_impl_unlinkat((long)(dfd), (long)(pathname), \
+#define __sanitizer_syscall_pre_unlinkat(dfd, pathname, flag)                  \
+  __sanitizer_syscall_pre_impl_unlinkat((long)(dfd), (long)(pathname),         \
                                         (long)(flag))
-#define __sanitizer_syscall_post_unlinkat(res, dfd, pathname, flag)          \
-  __sanitizer_syscall_post_impl_unlinkat(res, (long)(dfd), (long)(pathname), \
+#define __sanitizer_syscall_post_unlinkat(res, dfd, pathname, flag)            \
+  __sanitizer_syscall_post_impl_unlinkat(res, (long)(dfd), (long)(pathname),   \
                                          (long)(flag))
-#define __sanitizer_syscall_pre_symlinkat(oldname, newdfd, newname)       \
-  __sanitizer_syscall_pre_impl_symlinkat((long)(oldname), (long)(newdfd), \
+#define __sanitizer_syscall_pre_symlinkat(oldname, newdfd, newname)            \
+  __sanitizer_syscall_pre_impl_symlinkat((long)(oldname), (long)(newdfd),      \
                                          (long)(newname))
-#define __sanitizer_syscall_post_symlinkat(res, oldname, newdfd, newname) \
-  __sanitizer_syscall_post_impl_symlinkat(res, (long)(oldname),           \
+#define __sanitizer_syscall_post_symlinkat(res, oldname, newdfd, newname)      \
+  __sanitizer_syscall_post_impl_symlinkat(res, (long)(oldname),                \
                                           (long)(newdfd), (long)(newname))
-#define __sanitizer_syscall_pre_linkat(olddfd, oldname, newdfd, newname, \
-                                       flags)                            \
-  __sanitizer_syscall_pre_impl_linkat((long)(olddfd), (long)(oldname),   \
-                                      (long)(newdfd), (long)(newname),   \
+#define __sanitizer_syscall_pre_linkat(olddfd, oldname, newdfd, newname,       \
+                                       flags)                                  \
+  __sanitizer_syscall_pre_impl_linkat((long)(olddfd), (long)(oldname),         \
+                                      (long)(newdfd), (long)(newname),         \
                                       (long)(flags))
 #define __sanitizer_syscall_post_linkat(res, olddfd, oldname, newdfd, newname, \
                                         flags)                                 \
   __sanitizer_syscall_post_impl_linkat(res, (long)(olddfd), (long)(oldname),   \
                                        (long)(newdfd), (long)(newname),        \
                                        (long)(flags))
-#define __sanitizer_syscall_pre_renameat(olddfd, oldname, newdfd, newname) \
-  __sanitizer_syscall_pre_impl_renameat((long)(olddfd), (long)(oldname),   \
+#define __sanitizer_syscall_pre_renameat(olddfd, oldname, newdfd, newname)     \
+  __sanitizer_syscall_pre_impl_renameat((long)(olddfd), (long)(oldname),       \
                                         (long)(newdfd), (long)(newname))
 #define __sanitizer_syscall_post_renameat(res, olddfd, oldname, newdfd,        \
                                           newname)                             \
   __sanitizer_syscall_post_impl_renameat(res, (long)(olddfd), (long)(oldname), \
                                          (long)(newdfd), (long)(newname))
-#define __sanitizer_syscall_pre_futimesat(dfd, filename, utimes)        \
-  __sanitizer_syscall_pre_impl_futimesat((long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_pre_futimesat(dfd, filename, utimes)               \
+  __sanitizer_syscall_pre_impl_futimesat((long)(dfd), (long)(filename),        \
                                          (long)(utimes))
-#define __sanitizer_syscall_post_futimesat(res, dfd, filename, utimes)        \
-  __sanitizer_syscall_post_impl_futimesat(res, (long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_post_futimesat(res, dfd, filename, utimes)         \
+  __sanitizer_syscall_post_impl_futimesat(res, (long)(dfd), (long)(filename),  \
                                           (long)(utimes))
-#define __sanitizer_syscall_pre_faccessat(dfd, filename, mode)          \
-  __sanitizer_syscall_pre_impl_faccessat((long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_pre_faccessat(dfd, filename, mode)                 \
+  __sanitizer_syscall_pre_impl_faccessat((long)(dfd), (long)(filename),        \
                                          (long)(mode))
-#define __sanitizer_syscall_post_faccessat(res, dfd, filename, mode)          \
-  __sanitizer_syscall_post_impl_faccessat(res, (long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_post_faccessat(res, dfd, filename, mode)           \
+  __sanitizer_syscall_post_impl_faccessat(res, (long)(dfd), (long)(filename),  \
                                           (long)(mode))
-#define __sanitizer_syscall_pre_fchmodat(dfd, filename, mode)          \
-  __sanitizer_syscall_pre_impl_fchmodat((long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_pre_fchmodat(dfd, filename, mode)                  \
+  __sanitizer_syscall_pre_impl_fchmodat((long)(dfd), (long)(filename),         \
                                         (long)(mode))
-#define __sanitizer_syscall_post_fchmodat(res, dfd, filename, mode)          \
-  __sanitizer_syscall_post_impl_fchmodat(res, (long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_post_fchmodat(res, dfd, filename, mode)            \
+  __sanitizer_syscall_post_impl_fchmodat(res, (long)(dfd), (long)(filename),   \
                                          (long)(mode))
-#define __sanitizer_syscall_pre_fchownat(dfd, filename, user, group, flag) \
-  __sanitizer_syscall_pre_impl_fchownat((long)(dfd), (long)(filename),     \
-                                        (long)(user), (long)(group),       \
+#define __sanitizer_syscall_pre_fchownat(dfd, filename, user, group, flag)     \
+  __sanitizer_syscall_pre_impl_fchownat((long)(dfd), (long)(filename),         \
+                                        (long)(user), (long)(group),           \
                                         (long)(flag))
-#define __sanitizer_syscall_post_fchownat(res, dfd, filename, user, group,   \
-                                          flag)                              \
-  __sanitizer_syscall_post_impl_fchownat(res, (long)(dfd), (long)(filename), \
-                                         (long)(user), (long)(group),        \
+#define __sanitizer_syscall_post_fchownat(res, dfd, filename, user, group,     \
+                                          flag)                                \
+  __sanitizer_syscall_post_impl_fchownat(res, (long)(dfd), (long)(filename),   \
+                                         (long)(user), (long)(group),          \
                                          (long)(flag))
-#define __sanitizer_syscall_pre_openat(dfd, filename, flags, mode)   \
-  __sanitizer_syscall_pre_impl_openat((long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_pre_openat(dfd, filename, flags, mode)             \
+  __sanitizer_syscall_pre_impl_openat((long)(dfd), (long)(filename),           \
                                       (long)(flags), (long)(mode))
-#define __sanitizer_syscall_post_openat(res, dfd, filename, flags, mode)   \
-  __sanitizer_syscall_post_impl_openat(res, (long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_post_openat(res, dfd, filename, flags, mode)       \
+  __sanitizer_syscall_post_impl_openat(res, (long)(dfd), (long)(filename),     \
                                        (long)(flags), (long)(mode))
-#define __sanitizer_syscall_pre_newfstatat(dfd, filename, statbuf, flag) \
-  __sanitizer_syscall_pre_impl_newfstatat((long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_pre_newfstatat(dfd, filename, statbuf, flag)       \
+  __sanitizer_syscall_pre_impl_newfstatat((long)(dfd), (long)(filename),       \
                                           (long)(statbuf), (long)(flag))
 #define __sanitizer_syscall_post_newfstatat(res, dfd, filename, statbuf, flag) \
   __sanitizer_syscall_post_impl_newfstatat(res, (long)(dfd), (long)(filename), \
                                            (long)(statbuf), (long)(flag))
-#define __sanitizer_syscall_pre_fstatat64(dfd, filename, statbuf, flag) \
-  __sanitizer_syscall_pre_impl_fstatat64((long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_pre_fstatat64(dfd, filename, statbuf, flag)        \
+  __sanitizer_syscall_pre_impl_fstatat64((long)(dfd), (long)(filename),        \
                                          (long)(statbuf), (long)(flag))
-#define __sanitizer_syscall_post_fstatat64(res, dfd, filename, statbuf, flag) \
-  __sanitizer_syscall_post_impl_fstatat64(res, (long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_post_fstatat64(res, dfd, filename, statbuf, flag)  \
+  __sanitizer_syscall_post_impl_fstatat64(res, (long)(dfd), (long)(filename),  \
                                           (long)(statbuf), (long)(flag))
-#define __sanitizer_syscall_pre_readlinkat(dfd, path, buf, bufsiz)   \
-  __sanitizer_syscall_pre_impl_readlinkat((long)(dfd), (long)(path), \
+#define __sanitizer_syscall_pre_readlinkat(dfd, path, buf, bufsiz)             \
+  __sanitizer_syscall_pre_impl_readlinkat((long)(dfd), (long)(path),           \
                                           (long)(buf), (long)(bufsiz))
-#define __sanitizer_syscall_post_readlinkat(res, dfd, path, buf, bufsiz)   \
-  __sanitizer_syscall_post_impl_readlinkat(res, (long)(dfd), (long)(path), \
+#define __sanitizer_syscall_post_readlinkat(res, dfd, path, buf, bufsiz)       \
+  __sanitizer_syscall_post_impl_readlinkat(res, (long)(dfd), (long)(path),     \
                                            (long)(buf), (long)(bufsiz))
-#define __sanitizer_syscall_pre_utimensat(dfd, filename, utimes, flags) \
-  __sanitizer_syscall_pre_impl_utimensat((long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_pre_utimensat(dfd, filename, utimes, flags)        \
+  __sanitizer_syscall_pre_impl_utimensat((long)(dfd), (long)(filename),        \
                                          (long)(utimes), (long)(flags))
-#define __sanitizer_syscall_post_utimensat(res, dfd, filename, utimes, flags) \
-  __sanitizer_syscall_post_impl_utimensat(res, (long)(dfd), (long)(filename), \
+#define __sanitizer_syscall_post_utimensat(res, dfd, filename, utimes, flags)  \
+  __sanitizer_syscall_post_impl_utimensat(res, (long)(dfd), (long)(filename),  \
                                           (long)(utimes), (long)(flags))
-#define __sanitizer_syscall_pre_unshare(unshare_flags) \
+#define __sanitizer_syscall_pre_unshare(unshare_flags)                         \
   __sanitizer_syscall_pre_impl_unshare((long)(unshare_flags))
-#define __sanitizer_syscall_post_unshare(res, unshare_flags) \
+#define __sanitizer_syscall_post_unshare(res, unshare_flags)                   \
   __sanitizer_syscall_post_impl_unshare(res, (long)(unshare_flags))
-#define __sanitizer_syscall_pre_splice(fd_in, off_in, fd_out, off_out, len, \
-                                       flags)                               \
-  __sanitizer_syscall_pre_impl_splice((long)(fd_in), (long)(off_in),        \
-                                      (long)(fd_out), (long)(off_out),      \
+#define __sanitizer_syscall_pre_splice(fd_in, off_in, fd_out, off_out, len,    \
+                                       flags)                                  \
+  __sanitizer_syscall_pre_impl_splice((long)(fd_in), (long)(off_in),           \
+                                      (long)(fd_out), (long)(off_out),         \
                                       (long)(len), (long)(flags))
-#define __sanitizer_syscall_post_splice(res, fd_in, off_in, fd_out, off_out, \
-                                        len, flags)                          \
-  __sanitizer_syscall_post_impl_splice(res, (long)(fd_in), (long)(off_in),   \
-                                       (long)(fd_out), (long)(off_out),      \
+#define __sanitizer_syscall_post_splice(res, fd_in, off_in, fd_out, off_out,   \
+                                        len, flags)                            \
+  __sanitizer_syscall_post_impl_splice(res, (long)(fd_in), (long)(off_in),     \
+                                       (long)(fd_out), (long)(off_out),        \
                                        (long)(len), (long)(flags))
-#define __sanitizer_syscall_pre_vmsplice(fd, iov, nr_segs, flags) \
-  __sanitizer_syscall_pre_impl_vmsplice((long)(fd), (long)(iov),  \
+#define __sanitizer_syscall_pre_vmsplice(fd, iov, nr_segs, flags)              \
+  __sanitizer_syscall_pre_impl_vmsplice((long)(fd), (long)(iov),               \
                                         (long)(nr_segs), (long)(flags))
-#define __sanitizer_syscall_post_vmsplice(res, fd, iov, nr_segs, flags) \
-  __sanitizer_syscall_post_impl_vmsplice(res, (long)(fd), (long)(iov),  \
+#define __sanitizer_syscall_post_vmsplice(res, fd, iov, nr_segs, flags)        \
+  __sanitizer_syscall_post_impl_vmsplice(res, (long)(fd), (long)(iov),         \
                                          (long)(nr_segs), (long)(flags))
-#define __sanitizer_syscall_pre_tee(fdin, fdout, len, flags)                 \
-  __sanitizer_syscall_pre_impl_tee((long)(fdin), (long)(fdout), (long)(len), \
+#define __sanitizer_syscall_pre_tee(fdin, fdout, len, flags)                   \
+  __sanitizer_syscall_pre_impl_tee((long)(fdin), (long)(fdout), (long)(len),   \
                                    (long)(flags))
-#define __sanitizer_syscall_post_tee(res, fdin, fdout, len, flags)    \
-  __sanitizer_syscall_post_impl_tee(res, (long)(fdin), (long)(fdout), \
+#define __sanitizer_syscall_post_tee(res, fdin, fdout, len, flags)             \
+  __sanitizer_syscall_post_impl_tee(res, (long)(fdin), (long)(fdout),          \
                                     (long)(len), (long)(flags))
-#define __sanitizer_syscall_pre_get_robust_list(pid, head_ptr, len_ptr)       \
-  __sanitizer_syscall_pre_impl_get_robust_list((long)(pid), (long)(head_ptr), \
+#define __sanitizer_syscall_pre_get_robust_list(pid, head_ptr, len_ptr)        \
+  __sanitizer_syscall_pre_impl_get_robust_list((long)(pid), (long)(head_ptr),  \
                                                (long)(len_ptr))
-#define __sanitizer_syscall_post_get_robust_list(res, pid, head_ptr, len_ptr) \
-  __sanitizer_syscall_post_impl_get_robust_list(                              \
+#define __sanitizer_syscall_post_get_robust_list(res, pid, head_ptr, len_ptr)  \
+  __sanitizer_syscall_post_impl_get_robust_list(                               \
       res, (long)(pid), (long)(head_ptr), (long)(len_ptr))
-#define __sanitizer_syscall_pre_set_robust_list(head, len) \
+#define __sanitizer_syscall_pre_set_robust_list(head, len)                     \
   __sanitizer_syscall_pre_impl_set_robust_list((long)(head), (long)(len))
-#define __sanitizer_syscall_post_set_robust_list(res, head, len) \
+#define __sanitizer_syscall_post_set_robust_list(res, head, len)               \
   __sanitizer_syscall_post_impl_set_robust_list(res, (long)(head), (long)(len))
-#define __sanitizer_syscall_pre_getcpu(cpu, node, cache) \
+#define __sanitizer_syscall_pre_getcpu(cpu, node, cache)                       \
   __sanitizer_syscall_pre_impl_getcpu((long)(cpu), (long)(node), (long)(cache))
-#define __sanitizer_syscall_post_getcpu(res, cpu, node, cache)         \
-  __sanitizer_syscall_post_impl_getcpu(res, (long)(cpu), (long)(node), \
+#define __sanitizer_syscall_post_getcpu(res, cpu, node, cache)                 \
+  __sanitizer_syscall_post_impl_getcpu(res, (long)(cpu), (long)(node),         \
                                        (long)(cache))
-#define __sanitizer_syscall_pre_signalfd(ufd, user_mask, sizemask)      \
-  __sanitizer_syscall_pre_impl_signalfd((long)(ufd), (long)(user_mask), \
+#define __sanitizer_syscall_pre_signalfd(ufd, user_mask, sizemask)             \
+  __sanitizer_syscall_pre_impl_signalfd((long)(ufd), (long)(user_mask),        \
                                         (long)(sizemask))
-#define __sanitizer_syscall_post_signalfd(res, ufd, user_mask, sizemask)      \
-  __sanitizer_syscall_post_impl_signalfd(res, (long)(ufd), (long)(user_mask), \
+#define __sanitizer_syscall_post_signalfd(res, ufd, user_mask, sizemask)       \
+  __sanitizer_syscall_post_impl_signalfd(res, (long)(ufd), (long)(user_mask),  \
                                          (long)(sizemask))
-#define __sanitizer_syscall_pre_signalfd4(ufd, user_mask, sizemask, flags) \
-  __sanitizer_syscall_pre_impl_signalfd4((long)(ufd), (long)(user_mask),   \
+#define __sanitizer_syscall_pre_signalfd4(ufd, user_mask, sizemask, flags)     \
+  __sanitizer_syscall_pre_impl_signalfd4((long)(ufd), (long)(user_mask),       \
                                          (long)(sizemask), (long)(flags))
 #define __sanitizer_syscall_post_signalfd4(res, ufd, user_mask, sizemask,      \
                                            flags)                              \
   __sanitizer_syscall_post_impl_signalfd4(res, (long)(ufd), (long)(user_mask), \
                                           (long)(sizemask), (long)(flags))
-#define __sanitizer_syscall_pre_timerfd_create(clockid, flags) \
+#define __sanitizer_syscall_pre_timerfd_create(clockid, flags)                 \
   __sanitizer_syscall_pre_impl_timerfd_create((long)(clockid), (long)(flags))
-#define __sanitizer_syscall_post_timerfd_create(res, clockid, flags) \
-  __sanitizer_syscall_post_impl_timerfd_create(res, (long)(clockid), \
+#define __sanitizer_syscall_post_timerfd_create(res, clockid, flags)           \
+  __sanitizer_syscall_post_impl_timerfd_create(res, (long)(clockid),           \
                                                (long)(flags))
-#define __sanitizer_syscall_pre_timerfd_settime(ufd, flags, utmr, otmr)    \
-  __sanitizer_syscall_pre_impl_timerfd_settime((long)(ufd), (long)(flags), \
+#define __sanitizer_syscall_pre_timerfd_settime(ufd, flags, utmr, otmr)        \
+  __sanitizer_syscall_pre_impl_timerfd_settime((long)(ufd), (long)(flags),     \
                                                (long)(utmr), (long)(otmr))
-#define __sanitizer_syscall_post_timerfd_settime(res, ufd, flags, utmr, otmr) \
-  __sanitizer_syscall_post_impl_timerfd_settime(                              \
+#define __sanitizer_syscall_post_timerfd_settime(res, ufd, flags, utmr, otmr)  \
+  __sanitizer_syscall_post_impl_timerfd_settime(                               \
       res, (long)(ufd), (long)(flags), (long)(utmr), (long)(otmr))
-#define __sanitizer_syscall_pre_timerfd_gettime(ufd, otmr) \
+#define __sanitizer_syscall_pre_timerfd_gettime(ufd, otmr)                     \
   __sanitizer_syscall_pre_impl_timerfd_gettime((long)(ufd), (long)(otmr))
-#define __sanitizer_syscall_post_timerfd_gettime(res, ufd, otmr) \
+#define __sanitizer_syscall_post_timerfd_gettime(res, ufd, otmr)               \
   __sanitizer_syscall_post_impl_timerfd_gettime(res, (long)(ufd), (long)(otmr))
-#define __sanitizer_syscall_pre_eventfd(count) \
+#define __sanitizer_syscall_pre_eventfd(count)                                 \
   __sanitizer_syscall_pre_impl_eventfd((long)(count))
-#define __sanitizer_syscall_post_eventfd(res, count) \
+#define __sanitizer_syscall_post_eventfd(res, count)                           \
   __sanitizer_syscall_post_impl_eventfd(res, (long)(count))
-#define __sanitizer_syscall_pre_eventfd2(count, flags) \
+#define __sanitizer_syscall_pre_eventfd2(count, flags)                         \
   __sanitizer_syscall_pre_impl_eventfd2((long)(count), (long)(flags))
-#define __sanitizer_syscall_post_eventfd2(res, count, flags) \
+#define __sanitizer_syscall_post_eventfd2(res, count, flags)                   \
   __sanitizer_syscall_post_impl_eventfd2(res, (long)(count), (long)(flags))
-#define __sanitizer_syscall_pre_old_readdir(arg0, arg1, arg2)          \
-  __sanitizer_syscall_pre_impl_old_readdir((long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_pre_old_readdir(arg0, arg1, arg2)                  \
+  __sanitizer_syscall_pre_impl_old_readdir((long)(arg0), (long)(arg1),         \
                                            (long)(arg2))
-#define __sanitizer_syscall_post_old_readdir(res, arg0, arg1, arg2)          \
-  __sanitizer_syscall_post_impl_old_readdir(res, (long)(arg0), (long)(arg1), \
+#define __sanitizer_syscall_post_old_readdir(res, arg0, arg1, arg2)            \
+  __sanitizer_syscall_post_impl_old_readdir(res, (long)(arg0), (long)(arg1),   \
                                             (long)(arg2))
-#define __sanitizer_syscall_pre_pselect6(arg0, arg1, arg2, arg3, arg4, arg5) \
-  __sanitizer_syscall_pre_impl_pselect6((long)(arg0), (long)(arg1),          \
-                                        (long)(arg2), (long)(arg3),          \
+#define __sanitizer_syscall_pre_pselect6(arg0, arg1, arg2, arg3, arg4, arg5)   \
+  __sanitizer_syscall_pre_impl_pselect6((long)(arg0), (long)(arg1),            \
+                                        (long)(arg2), (long)(arg3),            \
                                         (long)(arg4), (long)(arg5))
-#define __sanitizer_syscall_post_pselect6(res, arg0, arg1, arg2, arg3, arg4, \
-                                          arg5)                              \
-  __sanitizer_syscall_post_impl_pselect6(res, (long)(arg0), (long)(arg1),    \
-                                         (long)(arg2), (long)(arg3),         \
+#define __sanitizer_syscall_post_pselect6(res, arg0, arg1, arg2, arg3, arg4,   \
+                                          arg5)                                \
+  __sanitizer_syscall_post_impl_pselect6(res, (long)(arg0), (long)(arg1),      \
+                                         (long)(arg2), (long)(arg3),           \
                                          (long)(arg4), (long)(arg5))
 #define __sanitizer_syscall_pre_ppoll(arg0, arg1, arg2, arg3, arg4)            \
   __sanitizer_syscall_pre_impl_ppoll((long)(arg0), (long)(arg1), (long)(arg2), \
                                      (long)(arg3), (long)(arg4))
-#define __sanitizer_syscall_post_ppoll(res, arg0, arg1, arg2, arg3, arg4) \
-  __sanitizer_syscall_post_impl_ppoll(res, (long)(arg0), (long)(arg1),    \
-                                      (long)(arg2), (long)(arg3),         \
+#define __sanitizer_syscall_post_ppoll(res, arg0, arg1, arg2, arg3, arg4)      \
+  __sanitizer_syscall_post_impl_ppoll(res, (long)(arg0), (long)(arg1),         \
+                                      (long)(arg2), (long)(arg3),              \
                                       (long)(arg4))
-#define __sanitizer_syscall_pre_syncfs(fd) \
+#define __sanitizer_syscall_pre_syncfs(fd)                                     \
   __sanitizer_syscall_pre_impl_syncfs((long)(fd))
-#define __sanitizer_syscall_post_syncfs(res, fd) \
+#define __sanitizer_syscall_post_syncfs(res, fd)                               \
   __sanitizer_syscall_post_impl_syncfs(res, (long)(fd))
 #define __sanitizer_syscall_pre_perf_event_open(attr_uptr, pid, cpu, group_fd, \
                                                 flags)                         \
   __sanitizer_syscall_pre_impl_perf_event_open((long)(attr_uptr), (long)(pid), \
                                                (long)(cpu), (long)(group_fd),  \
                                                (long)(flags))
-#define __sanitizer_syscall_post_perf_event_open(res, attr_uptr, pid, cpu, \
-                                                 group_fd, flags)          \
-  __sanitizer_syscall_post_impl_perf_event_open(                           \
-      res, (long)(attr_uptr), (long)(pid), (long)(cpu), (long)(group_fd),  \
+#define __sanitizer_syscall_post_perf_event_open(res, attr_uptr, pid, cpu,     \
+                                                 group_fd, flags)              \
+  __sanitizer_syscall_post_impl_perf_event_open(                               \
+      res, (long)(attr_uptr), (long)(pid), (long)(cpu), (long)(group_fd),      \
       (long)(flags))
-#define __sanitizer_syscall_pre_mmap_pgoff(addr, len, prot, flags, fd, pgoff) \
-  __sanitizer_syscall_pre_impl_mmap_pgoff((long)(addr), (long)(len),          \
-                                          (long)(prot), (long)(flags),        \
+#define __sanitizer_syscall_pre_mmap_pgoff(addr, len, prot, flags, fd, pgoff)  \
+  __sanitizer_syscall_pre_impl_mmap_pgoff((long)(addr), (long)(len),           \
+                                          (long)(prot), (long)(flags),         \
                                           (long)(fd), (long)(pgoff))
-#define __sanitizer_syscall_post_mmap_pgoff(res, addr, len, prot, flags, fd, \
-                                            pgoff)                           \
-  __sanitizer_syscall_post_impl_mmap_pgoff(res, (long)(addr), (long)(len),   \
-                                           (long)(prot), (long)(flags),      \
+#define __sanitizer_syscall_post_mmap_pgoff(res, addr, len, prot, flags, fd,   \
+                                            pgoff)                             \
+  __sanitizer_syscall_post_impl_mmap_pgoff(res, (long)(addr), (long)(len),     \
+                                           (long)(prot), (long)(flags),        \
                                            (long)(fd), (long)(pgoff))
-#define __sanitizer_syscall_pre_old_mmap(arg) \
+#define __sanitizer_syscall_pre_old_mmap(arg)                                  \
   __sanitizer_syscall_pre_impl_old_mmap((long)(arg))
-#define __sanitizer_syscall_post_old_mmap(res, arg) \
+#define __sanitizer_syscall_post_old_mmap(res, arg)                            \
   __sanitizer_syscall_post_impl_old_mmap(res, (long)(arg))
-#define __sanitizer_syscall_pre_name_to_handle_at(dfd, name, handle, mnt_id, \
-                                                  flag)                      \
-  __sanitizer_syscall_pre_impl_name_to_handle_at(                            \
+#define __sanitizer_syscall_pre_name_to_handle_at(dfd, name, handle, mnt_id,   \
+                                                  flag)                        \
+  __sanitizer_syscall_pre_impl_name_to_handle_at(                              \
       (long)(dfd), (long)(name), (long)(handle), (long)(mnt_id), (long)(flag))
-#define __sanitizer_syscall_post_name_to_handle_at(res, dfd, name, handle, \
-                                                   mnt_id, flag)           \
-  __sanitizer_syscall_post_impl_name_to_handle_at(                         \
-      res, (long)(dfd), (long)(name), (long)(handle), (long)(mnt_id),      \
+#define __sanitizer_syscall_post_name_to_handle_at(res, dfd, name, handle,     \
+                                                   mnt_id, flag)               \
+  __sanitizer_syscall_post_impl_name_to_handle_at(                             \
+      res, (long)(dfd), (long)(name), (long)(handle), (long)(mnt_id),          \
       (long)(flag))
-#define __sanitizer_syscall_pre_open_by_handle_at(mountdirfd, handle, flags) \
-  __sanitizer_syscall_pre_impl_open_by_handle_at(                            \
+#define __sanitizer_syscall_pre_open_by_handle_at(mountdirfd, handle, flags)   \
+  __sanitizer_syscall_pre_impl_open_by_handle_at(                              \
       (long)(mountdirfd), (long)(handle), (long)(flags))
-#define __sanitizer_syscall_post_open_by_handle_at(res, mountdirfd, handle, \
-                                                   flags)                   \
-  __sanitizer_syscall_post_impl_open_by_handle_at(                          \
+#define __sanitizer_syscall_post_open_by_handle_at(res, mountdirfd, handle,    \
+                                                   flags)                      \
+  __sanitizer_syscall_post_impl_open_by_handle_at(                             \
       res, (long)(mountdirfd), (long)(handle), (long)(flags))
-#define __sanitizer_syscall_pre_setns(fd, nstype) \
+#define __sanitizer_syscall_pre_setns(fd, nstype)                              \
   __sanitizer_syscall_pre_impl_setns((long)(fd), (long)(nstype))
-#define __sanitizer_syscall_post_setns(res, fd, nstype) \
+#define __sanitizer_syscall_post_setns(res, fd, nstype)                        \
   __sanitizer_syscall_post_impl_setns(res, (long)(fd), (long)(nstype))
-#define __sanitizer_syscall_pre_process_vm_readv(pid, lvec, liovcnt, rvec, \
-                                                 riovcnt, flags)           \
-  __sanitizer_syscall_pre_impl_process_vm_readv(                           \
-      (long)(pid), (long)(lvec), (long)(liovcnt), (long)(rvec),            \
+#define __sanitizer_syscall_pre_process_vm_readv(pid, lvec, liovcnt, rvec,     \
+                                                 riovcnt, flags)               \
+  __sanitizer_syscall_pre_impl_process_vm_readv(                               \
+      (long)(pid), (long)(lvec), (long)(liovcnt), (long)(rvec),                \
       (long)(riovcnt), (long)(flags))
-#define __sanitizer_syscall_post_process_vm_readv(res, pid, lvec, liovcnt, \
-                                                  rvec, riovcnt, flags)    \
-  __sanitizer_syscall_post_impl_process_vm_readv(                          \
-      res, (long)(pid), (long)(lvec), (long)(liovcnt), (long)(rvec),       \
+#define __sanitizer_syscall_post_process_vm_readv(res, pid, lvec, liovcnt,     \
+                                                  rvec, riovcnt, flags)        \
+  __sanitizer_syscall_post_impl_process_vm_readv(                              \
+      res, (long)(pid), (long)(lvec), (long)(liovcnt), (long)(rvec),           \
       (long)(riovcnt), (long)(flags))
-#define __sanitizer_syscall_pre_process_vm_writev(pid, lvec, liovcnt, rvec, \
-                                                  riovcnt, flags)           \
-  __sanitizer_syscall_pre_impl_process_vm_writev(                           \
-      (long)(pid), (long)(lvec), (long)(liovcnt), (long)(rvec),             \
+#define __sanitizer_syscall_pre_process_vm_writev(pid, lvec, liovcnt, rvec,    \
+                                                  riovcnt, flags)              \
+  __sanitizer_syscall_pre_impl_process_vm_writev(                              \
+      (long)(pid), (long)(lvec), (long)(liovcnt), (long)(rvec),                \
       (long)(riovcnt), (long)(flags))
-#define __sanitizer_syscall_post_process_vm_writev(res, pid, lvec, liovcnt, \
-                                                   rvec, riovcnt, flags)    \
-  __sanitizer_syscall_post_impl_process_vm_writev(                          \
-      res, (long)(pid), (long)(lvec), (long)(liovcnt), (long)(rvec),        \
+#define __sanitizer_syscall_post_process_vm_writev(res, pid, lvec, liovcnt,    \
+                                                   rvec, riovcnt, flags)       \
+  __sanitizer_syscall_post_impl_process_vm_writev(                             \
+      res, (long)(pid), (long)(lvec), (long)(liovcnt), (long)(rvec),           \
       (long)(riovcnt), (long)(flags))
-#define __sanitizer_syscall_pre_fork() \
-  __sanitizer_syscall_pre_impl_fork()
-#define __sanitizer_syscall_post_fork(res) \
+#define __sanitizer_syscall_pre_fork() __sanitizer_syscall_pre_impl_fork()
+#define __sanitizer_syscall_post_fork(res)                                     \
   __sanitizer_syscall_post_impl_fork(res)
-#define __sanitizer_syscall_pre_vfork() \
-  __sanitizer_syscall_pre_impl_vfork()
-#define __sanitizer_syscall_post_vfork(res) \
+#define __sanitizer_syscall_pre_vfork() __sanitizer_syscall_pre_impl_vfork()
+#define __sanitizer_syscall_post_vfork(res)                                    \
   __sanitizer_syscall_post_impl_vfork(res)
 #define __sanitizer_syscall_pre_sigaction(signum, act, oldact)                 \
   __sanitizer_syscall_pre_impl_sigaction((long)signum, (long)act, (long)oldact)
@@ -2699,6 +2706,13 @@ void __sanitizer_syscall_pre_impl_epoll_pwait(long epfd, long events,
 void __sanitizer_syscall_post_impl_epoll_pwait(long res, long epfd, long events,
                                                long maxevents, long timeout,
                                                long sigmask, long sigsetsize);
+void __sanitizer_syscall_pre_impl_epoll_pwait2(long epfd, long events,
+                                               long maxevents, long timeout,
+                                               long sigmask, long sigsetsize);
+void __sanitizer_syscall_post_impl_epoll_pwait2(long res, long epfd,
+                                                long events, long maxevents,
+                                                long timeout, long sigmask,
+                                                long sigsetsize);
 void __sanitizer_syscall_pre_impl_gethostname(long name, long len);
 void __sanitizer_syscall_post_impl_gethostname(long res, long name, long len);
 void __sanitizer_syscall_pre_impl_sethostname(long name, long len);
@@ -3080,7 +3094,7 @@ void __sanitizer_syscall_post_impl_rt_sigaction(long res, long signum, long act,
 void __sanitizer_syscall_pre_impl_sigaltstack(long ss, long oss);
 void __sanitizer_syscall_post_impl_sigaltstack(long res, long ss, long oss);
 #ifdef __cplusplus
-}  // extern "C"
+} // extern "C"
 #endif
 
-#endif  // SANITIZER_LINUX_SYSCALL_HOOKS_H
+#endif // SANITIZER_LINUX_SYSCALL_HOOKS_H
diff --git a/compiler-rt/include/sanitizer/tsan_interface.h b/compiler-rt/include/sanitizer/tsan_interface.h
index 565aa391a9fa..2782e61fb8c7 100644
--- a/compiler-rt/include/sanitizer/tsan_interface.h
+++ b/compiler-rt/include/sanitizer/tsan_interface.h
@@ -169,6 +169,9 @@ void __tsan_on_initialize();
 // if TSan should exit as if issues were detected.
 int __tsan_on_finalize(int failed);
 
+// Release TSan internal memory in a best-effort manner.
+void __tsan_flush_memory();
+
 #ifdef __cplusplus
 }  // extern "C"
 #endif
diff --git a/compiler-rt/lib/asan/asan_allocator.cpp b/compiler-rt/lib/asan/asan_allocator.cpp
index 414fba3b427d..3fa36742060b 100644
--- a/compiler-rt/lib/asan/asan_allocator.cpp
+++ b/compiler-rt/lib/asan/asan_allocator.cpp
@@ -102,19 +102,18 @@ class ChunkHeader {
 
  public:
   uptr UsedSize() const {
-    uptr R = user_requested_size_lo;
-    if (sizeof(uptr) > sizeof(user_requested_size_lo))
-      R += (uptr)user_requested_size_hi << (8 * sizeof(user_requested_size_lo));
-    return R;
+    static_assert(sizeof(user_requested_size_lo) == 4,
+                  "Expression below requires this");
+    return FIRST_32_SECOND_64(0, ((uptr)user_requested_size_hi << 32)) +
+           user_requested_size_lo;
   }
 
   void SetUsedSize(uptr size) {
     user_requested_size_lo = size;
-    if (sizeof(uptr) > sizeof(user_requested_size_lo)) {
-      size >>= (8 * sizeof(user_requested_size_lo));
-      user_requested_size_hi = size;
-      CHECK_EQ(user_requested_size_hi, size);
-    }
+    static_assert(sizeof(user_requested_size_lo) == 4,
+                  "Expression below requires this");
+    user_requested_size_hi = FIRST_32_SECOND_64(0, size >> 32);
+    CHECK_EQ(UsedSize(), size);
   }
 
   void SetAllocContext(u32 tid, u32 stack) {
@@ -522,7 +521,7 @@ struct Allocator {
         size > max_user_defined_malloc_size) {
       if (AllocatorMayReturnNull()) {
         Report("WARNING: AddressSanitizer failed to allocate 0x%zx bytes\n",
-               (void*)size);
+               size);
         return nullptr;
       }
       uptr malloc_limit =
@@ -908,13 +907,6 @@ AllocType AsanChunkView::GetAllocType() const {
   return (AllocType)chunk_->alloc_type;
 }
 
-static StackTrace GetStackTraceFromId(u32 id) {
-  CHECK(id);
-  StackTrace res = StackDepotGet(id);
-  CHECK(res.trace);
-  return res;
-}
-
 u32 AsanChunkView::GetAllocStackId() const {
   u32 tid = 0;
   u32 stack = 0;
@@ -931,14 +923,6 @@ u32 AsanChunkView::GetFreeStackId() const {
   return stack;
 }
 
-StackTrace AsanChunkView::GetAllocStack() const {
-  return GetStackTraceFromId(GetAllocStackId());
-}
-
-StackTrace AsanChunkView::GetFreeStack() const {
-  return GetStackTraceFromId(GetFreeStackId());
-}
-
 void InitializeAllocator(const AllocatorOptions &options) {
   instance.InitLinkerInitialized(options);
 }
diff --git a/compiler-rt/lib/asan/asan_allocator.h b/compiler-rt/lib/asan/asan_allocator.h
index 2963e979b55c..27d826fb613a 100644
--- a/compiler-rt/lib/asan/asan_allocator.h
+++ b/compiler-rt/lib/asan/asan_allocator.h
@@ -64,8 +64,6 @@ class AsanChunkView {
   bool Eq(const AsanChunkView &c) const { return chunk_ == c.chunk_; }
   u32 GetAllocStackId() const;
   u32 GetFreeStackId() const;
-  StackTrace GetAllocStack() const;
-  StackTrace GetFreeStack() const;
   AllocType GetAllocType() const;
   bool AddrIsInside(uptr addr, uptr access_size, sptr *offset) const {
     if (addr >= Beg() && (addr + access_size) <= End()) {
diff --git a/compiler-rt/lib/asan/asan_debugging.cpp b/compiler-rt/lib/asan/asan_debugging.cpp
index c01360b52fc9..0b4bf52f2490 100644
--- a/compiler-rt/lib/asan/asan_debugging.cpp
+++ b/compiler-rt/lib/asan/asan_debugging.cpp
@@ -19,6 +19,7 @@
 #include "asan_mapping.h"
 #include "asan_report.h"
 #include "asan_thread.h"
+#include "sanitizer_common/sanitizer_stackdepot.h"
 
 namespace {
 using namespace __asan;
@@ -54,11 +55,11 @@ uptr AsanGetStack(uptr addr, uptr *trace, u32 size, u32 *thread_id,
   StackTrace stack(nullptr, 0);
   if (alloc_stack) {
     if (chunk.AllocTid() == kInvalidTid) return 0;
-    stack = chunk.GetAllocStack();
+    stack = StackDepotGet(chunk.GetAllocStackId());
     if (thread_id) *thread_id = chunk.AllocTid();
   } else {
     if (chunk.FreeTid() == kInvalidTid) return 0;
-    stack = chunk.GetFreeStack();
+    stack = StackDepotGet(chunk.GetFreeStackId());
     if (thread_id) *thread_id = chunk.FreeTid();
   }
 
diff --git a/compiler-rt/lib/asan/asan_descriptions.cpp b/compiler-rt/lib/asan/asan_descriptions.cpp
index 2ba8a02f8410..d7d961685793 100644
--- a/compiler-rt/lib/asan/asan_descriptions.cpp
+++ b/compiler-rt/lib/asan/asan_descriptions.cpp
@@ -251,7 +251,7 @@ static void PrintAccessAndVarIntersection(const StackVarDescr &var, uptr addr,
   }
   str.append("'");
   if (var.line > 0) {
-    str.append(" (line %d)", var.line);
+    str.append(" (line %zd)", var.line);
   }
   if (pos_descr) {
     Decorator d;
@@ -318,7 +318,8 @@ bool DescribeAddressIfGlobal(uptr addr, uptr access_size,
 }
 
 void ShadowAddressDescription::Print() const {
-  Printf("Address %p is located in the %s area.\n", addr, ShadowNames[kind]);
+  Printf("Address %p is located in the %s area.\n", (void *)addr,
+         ShadowNames[kind]);
 }
 
 void GlobalAddressDescription::Print(const char *bug_type) const {
@@ -356,7 +357,7 @@ bool GlobalAddressDescription::PointsInsideTheSameVariable(
 void StackAddressDescription::Print() const {
   Decorator d;
   Printf("%s", d.Location());
-  Printf("Address %p is located in stack of thread %s", addr,
+  Printf("Address %p is located in stack of thread %s", (void *)addr,
          AsanThreadIdAndName(tid).c_str());
 
   if (!frame_descr) {
@@ -469,7 +470,7 @@ AddressDescription::AddressDescription(uptr addr, uptr access_size,
 
 void WildAddressDescription::Print() const {
   Printf("Address %p is a wild pointer inside of access range of size %p.\n",
-         addr, access_size);
+         (void *)addr, (void *)access_size);
 }
 
 void PrintAddressDescription(uptr addr, uptr access_size,
diff --git a/compiler-rt/lib/asan/asan_errors.cpp b/compiler-rt/lib/asan/asan_errors.cpp
index 45166c064877..7cd9fe911afa 100644
--- a/compiler-rt/lib/asan/asan_errors.cpp
+++ b/compiler-rt/lib/asan/asan_errors.cpp
@@ -46,10 +46,9 @@ void ErrorDeadlySignal::Print() {
 void ErrorDoubleFree::Print() {
   Decorator d;
   Printf("%s", d.Error());
-  Report(
-      "ERROR: AddressSanitizer: attempting %s on %p in thread %s:\n",
-      scariness.GetDescription(), addr_description.addr,
-      AsanThreadIdAndName(tid).c_str());
+  Report("ERROR: AddressSanitizer: attempting %s on %p in thread %s:\n",
+         scariness.GetDescription(), (void *)addr_description.addr,
+         AsanThreadIdAndName(tid).c_str());
   Printf("%s", d.Default());
   scariness.Print();
   GET_STACK_TRACE_FATAL(second_free_stack->trace[0],
@@ -62,10 +61,9 @@ void ErrorDoubleFree::Print() {
 void ErrorNewDeleteTypeMismatch::Print() {
   Decorator d;
   Printf("%s", d.Error());
-  Report(
-      "ERROR: AddressSanitizer: %s on %p in thread %s:\n",
-      scariness.GetDescription(), addr_description.addr,
-      AsanThreadIdAndName(tid).c_str());
+  Report("ERROR: AddressSanitizer: %s on %p in thread %s:\n",
+         scariness.GetDescription(), (void *)addr_description.addr,
+         AsanThreadIdAndName(tid).c_str());
   Printf("%s  object passed to delete has wrong type:\n", d.Default());
   if (delete_size != 0) {
     Printf(
@@ -106,7 +104,7 @@ void ErrorFreeNotMalloced::Print() {
   Report(
       "ERROR: AddressSanitizer: attempting free on address "
       "which was not malloc()-ed: %p in thread %s\n",
-      addr_description.Address(), AsanThreadIdAndName(tid).c_str());
+      (void *)addr_description.Address(), AsanThreadIdAndName(tid).c_str());
   Printf("%s", d.Default());
   CHECK_GT(free_stack->size, 0);
   scariness.Print();
@@ -126,7 +124,7 @@ void ErrorAllocTypeMismatch::Print() {
   Printf("%s", d.Error());
   Report("ERROR: AddressSanitizer: %s (%s vs %s) on %p\n",
          scariness.GetDescription(), alloc_names[alloc_type],
-         dealloc_names[dealloc_type], addr_description.Address());
+         dealloc_names[dealloc_type], (void *)addr_description.Address());
   Printf("%s", d.Default());
   CHECK_GT(dealloc_stack->size, 0);
   scariness.Print();
@@ -145,7 +143,7 @@ void ErrorMallocUsableSizeNotOwned::Print() {
   Report(
       "ERROR: AddressSanitizer: attempting to call malloc_usable_size() for "
       "pointer which is not owned: %p\n",
-      addr_description.Address());
+      (void *)addr_description.Address());
   Printf("%s", d.Default());
   stack->Print();
   addr_description.Print();
@@ -158,7 +156,7 @@ void ErrorSanitizerGetAllocatedSizeNotOwned::Print() {
   Report(
       "ERROR: AddressSanitizer: attempting to call "
       "__sanitizer_get_allocated_size() for pointer which is not owned: %p\n",
-      addr_description.Address());
+      (void *)addr_description.Address());
   Printf("%s", d.Default());
   stack->Print();
   addr_description.Print();
@@ -298,9 +296,10 @@ void ErrorStringFunctionMemoryRangesOverlap::Print() {
   Report(
       "ERROR: AddressSanitizer: %s: memory ranges [%p,%p) and [%p, %p) "
       "overlap\n",
-      bug_type, addr1_description.Address(),
-      addr1_description.Address() + length1, addr2_description.Address(),
-      addr2_description.Address() + length2);
+      bug_type, (void *)addr1_description.Address(),
+      (void *)(addr1_description.Address() + length1),
+      (void *)addr2_description.Address(),
+      (void *)(addr2_description.Address() + length2));
   Printf("%s", d.Default());
   scariness.Print();
   stack->Print();
@@ -329,10 +328,10 @@ void ErrorBadParamsToAnnotateContiguousContainer::Print() {
       "      end     : %p\n"
       "      old_mid : %p\n"
       "      new_mid : %p\n",
-      beg, end, old_mid, new_mid);
+      (void *)beg, (void *)end, (void *)old_mid, (void *)new_mid);
   uptr granularity = SHADOW_GRANULARITY;
   if (!IsAligned(beg, granularity))
-    Report("ERROR: beg is not aligned by %d\n", granularity);
+    Report("ERROR: beg is not aligned by %zu\n", granularity);
   stack->Print();
   ReportErrorSummary(scariness.GetDescription(), stack);
 }
@@ -341,7 +340,7 @@ void ErrorODRViolation::Print() {
   Decorator d;
   Printf("%s", d.Error());
   Report("ERROR: AddressSanitizer: %s (%p):\n", scariness.GetDescription(),
-         global1.beg);
+         (void *)global1.beg);
   Printf("%s", d.Default());
   InternalScopedString g1_loc;
   InternalScopedString g2_loc;
@@ -371,7 +370,8 @@ void ErrorInvalidPointerPair::Print() {
   Decorator d;
   Printf("%s", d.Error());
   Report("ERROR: AddressSanitizer: %s: %p %p\n", scariness.GetDescription(),
-         addr1_description.Address(), addr2_description.Address());
+         (void *)addr1_description.Address(),
+         (void *)addr2_description.Address());
   Printf("%s", d.Default());
   GET_STACK_TRACE_FATAL(pc, bp);
   stack.Print();
@@ -538,7 +538,8 @@ static void PrintLegend(InternalScopedString *str) {
 static void PrintShadowBytes(InternalScopedString *str, const char *before,
                              u8 *bytes, u8 *guilty, uptr n) {
   Decorator d;
-  if (before) str->append("%s%p:", before, bytes);
+  if (before)
+    str->append("%s%p:", before, (void *)bytes);
   for (uptr i = 0; i < n; i++) {
     u8 *p = bytes + i;
     const char *before =
@@ -575,7 +576,7 @@ void ErrorGeneric::Print() {
   Printf("%s", d.Error());
   uptr addr = addr_description.Address();
   Report("ERROR: AddressSanitizer: %s on address %p at pc %p bp %p sp %p\n",
-         bug_descr, (void *)addr, pc, bp, sp);
+         bug_descr, (void *)addr, (void *)pc, (void *)bp, (void *)sp);
   Printf("%s", d.Default());
 
   Printf("%s%s of size %zu at %p thread %s%s\n", d.Access(),
diff --git a/compiler-rt/lib/asan/asan_fake_stack.cpp b/compiler-rt/lib/asan/asan_fake_stack.cpp
index bf5c342ee59d..07681c10de91 100644
--- a/compiler-rt/lib/asan/asan_fake_stack.cpp
+++ b/compiler-rt/lib/asan/asan_fake_stack.cpp
@@ -54,10 +54,11 @@ FakeStack *FakeStack::Create(uptr stack_size_log) {
                              : MmapOrDie(size, "FakeStack"));
   res->stack_size_log_ = stack_size_log;
   u8 *p = reinterpret_cast<u8 *>(res);
-  VReport(1, "T%d: FakeStack created: %p -- %p stack_size_log: %zd; "
+  VReport(1,
+          "T%d: FakeStack created: %p -- %p stack_size_log: %zd; "
           "mmapped %zdK, noreserve=%d \n",
-          GetCurrentTidOrInvalid(), p,
-          p + FakeStack::RequiredSize(stack_size_log), stack_size_log,
+          GetCurrentTidOrInvalid(), (void *)p,
+          (void *)(p + FakeStack::RequiredSize(stack_size_log)), stack_size_log,
           size >> 10, flags()->uar_noreserve);
   return res;
 }
diff --git a/compiler-rt/lib/asan/asan_fuchsia.cpp b/compiler-rt/lib/asan/asan_fuchsia.cpp
index b0c7255144ac..15381d5bd0e5 100644
--- a/compiler-rt/lib/asan/asan_fuchsia.cpp
+++ b/compiler-rt/lib/asan/asan_fuchsia.cpp
@@ -31,7 +31,8 @@ namespace __asan {
 // AsanInitInternal->InitializeHighMemEnd (asan_rtl.cpp).
 // Just do some additional sanity checks here.
 void InitializeShadowMemory() {
-  if (Verbosity()) PrintAddressSpaceLayout();
+  if (Verbosity())
+    PrintAddressSpaceLayout();
 
   // Make sure SHADOW_OFFSET doesn't use __asan_shadow_memory_dynamic_address.
   __asan_shadow_memory_dynamic_address = kDefaultShadowSentinel;
@@ -62,7 +63,34 @@ void AsanOnDeadlySignal(int signo, void *siginfo, void *context) {
   UNIMPLEMENTED();
 }
 
-bool PlatformUnpoisonStacks() { return false; }
+bool PlatformUnpoisonStacks() {
+  // The current sp might not point to the default stack. This
+  // could be because we are in a crash stack from fuzzing for example.
+  // Unpoison the default stack and the current stack page.
+  AsanThread *curr_thread = GetCurrentThread();
+  CHECK(curr_thread != nullptr);
+  uptr top = curr_thread->stack_top();
+  uptr bottom = curr_thread->stack_bottom();
+  // The default stack grows from top to bottom. (bottom < top).
+
+  uptr local_stack = reinterpret_cast<uptr>(__builtin_frame_address(0));
+  if (local_stack >= bottom && local_stack <= top) {
+    // The current stack is the default stack.
+    // We only need to unpoison from where we are using until the end.
+    bottom = RoundDownTo(local_stack, GetPageSize());
+    UnpoisonStack(bottom, top, "default");
+  } else {
+    // The current stack is not the default stack.
+    // Unpoison the entire default stack and the current stack page.
+    UnpoisonStack(bottom, top, "default");
+    bottom = RoundDownTo(local_stack, GetPageSize());
+    top = bottom + GetPageSize();
+    UnpoisonStack(bottom, top, "unknown");
+    return true;
+  }
+
+  return false;
+}
 
 // We can use a plain thread_local variable for TSD.
 static thread_local void *per_thread;
@@ -90,14 +118,12 @@ struct AsanThread::InitOptions {
 
 // Shared setup between thread creation and startup for the initial thread.
 static AsanThread *CreateAsanThread(StackTrace *stack, u32 parent_tid,
-                                    uptr user_id, bool detached,
-                                    const char *name) {
+                                    bool detached, const char *name) {
   // In lieu of AsanThread::Create.
   AsanThread *thread = (AsanThread *)MmapOrDie(AsanThreadMmapSize(), __func__);
 
   AsanThreadContext::CreateThreadContextArgs args = {thread, stack};
-  u32 tid =
-      asanThreadRegistry().CreateThread(user_id, detached, parent_tid, &args);
+  u32 tid = asanThreadRegistry().CreateThread(0, detached, parent_tid, &args);
   asanThreadRegistry().SetThreadName(tid, name);
 
   return thread;
@@ -124,7 +150,7 @@ AsanThread *CreateMainThread() {
   CHECK_NE(__sanitizer::MainThreadStackBase, 0);
   CHECK_GT(__sanitizer::MainThreadStackSize, 0);
   AsanThread *t = CreateAsanThread(
-      nullptr, 0, reinterpret_cast<uptr>(self), true,
+      nullptr, 0, true,
       _zx_object_get_property(thrd_get_zx_handle(self), ZX_PROP_NAME, name,
                               sizeof(name)) == ZX_OK
           ? name
@@ -148,13 +174,13 @@ static void *BeforeThreadCreateHook(uptr user_id, bool detached,
                                     uptr stack_size) {
   EnsureMainThreadIDIsCorrect();
   // Strict init-order checking is thread-hostile.
-  if (flags()->strict_init_order) StopInitOrderChecking();
+  if (flags()->strict_init_order)
+    StopInitOrderChecking();
 
   GET_STACK_TRACE_THREAD;
   u32 parent_tid = GetCurrentTidOrInvalid();
 
-  AsanThread *thread =
-      CreateAsanThread(&stack, parent_tid, user_id, detached, name);
+  AsanThread *thread = CreateAsanThread(&stack, parent_tid, detached, name);
 
   // On other systems, AsanThread::Init() is called from the new
   // thread itself.  But on Fuchsia we already know the stack address
diff --git a/compiler-rt/lib/asan/asan_globals.cpp b/compiler-rt/lib/asan/asan_globals.cpp
index 9d7dbc6f264c..5f56fe6f457d 100644
--- a/compiler-rt/lib/asan/asan_globals.cpp
+++ b/compiler-rt/lib/asan/asan_globals.cpp
@@ -35,7 +35,7 @@ struct ListOfGlobals {
   ListOfGlobals *next;
 };
 
-static BlockingMutex mu_for_globals(LINKER_INITIALIZED);
+static Mutex mu_for_globals;
 static LowLevelAllocator allocator_for_globals;
 static ListOfGlobals *list_of_all_globals;
 
@@ -85,12 +85,12 @@ static void ReportGlobal(const Global &g, const char *prefix) {
   Report(
       "%s Global[%p]: beg=%p size=%zu/%zu name=%s module=%s dyn_init=%zu "
       "odr_indicator=%p\n",
-      prefix, &g, (void *)g.beg, g.size, g.size_with_redzone, g.name,
+      prefix, (void *)&g, (void *)g.beg, g.size, g.size_with_redzone, g.name,
       g.module_name, g.has_dynamic_init, (void *)g.odr_indicator);
   if (g.location) {
-    Report("  location (%p): name=%s[%p], %d %d\n", g.location,
-           g.location->filename, g.location->filename, g.location->line_no,
-           g.location->column_no);
+    Report("  location (%p): name=%s[%p], %d %d\n", (void *)g.location,
+           g.location->filename, (void *)g.location->filename,
+           g.location->line_no, g.location->column_no);
   }
 }
 
@@ -108,7 +108,7 @@ static u32 FindRegistrationSite(const Global *g) {
 int GetGlobalsForAddress(uptr addr, Global *globals, u32 *reg_sites,
                          int max_globals) {
   if (!flags()->report_globals) return 0;
-  BlockingMutexLock lock(&mu_for_globals);
+  Lock lock(&mu_for_globals);
   int res = 0;
   for (ListOfGlobals *l = list_of_all_globals; l; l = l->next) {
     const Global &g = *l->g;
@@ -257,7 +257,7 @@ static void UnregisterGlobal(const Global *g) {
 }
 
 void StopInitOrderChecking() {
-  BlockingMutexLock lock(&mu_for_globals);
+  Lock lock(&mu_for_globals);
   if (!flags()->check_initialization_order || !dynamic_init_globals)
     return;
   flags()->check_initialization_order = false;
@@ -359,7 +359,7 @@ void __asan_register_globals(__asan_global *globals, uptr n) {
   if (!flags()->report_globals) return;
   GET_STACK_TRACE_MALLOC;
   u32 stack_id = StackDepotPut(stack);
-  BlockingMutexLock lock(&mu_for_globals);
+  Lock lock(&mu_for_globals);
   if (!global_registration_site_vector) {
     global_registration_site_vector =
         new (allocator_for_globals) GlobalRegistrationSiteVector;
@@ -369,7 +369,8 @@ void __asan_register_globals(__asan_global *globals, uptr n) {
   global_registration_site_vector->push_back(site);
   if (flags()->report_globals >= 2) {
     PRINT_CURRENT_STACK();
-    Printf("=== ID %d; %p %p\n", stack_id, &globals[0], &globals[n - 1]);
+    Printf("=== ID %d; %p %p\n", stack_id, (void *)&globals[0],
+           (void *)&globals[n - 1]);
   }
   for (uptr i = 0; i < n; i++) {
     if (SANITIZER_WINDOWS && globals[i].beg == 0) {
@@ -398,7 +399,7 @@ void __asan_register_globals(__asan_global *globals, uptr n) {
 // We must do this when a shared objects gets dlclosed.
 void __asan_unregister_globals(__asan_global *globals, uptr n) {
   if (!flags()->report_globals) return;
-  BlockingMutexLock lock(&mu_for_globals);
+  Lock lock(&mu_for_globals);
   for (uptr i = 0; i < n; i++) {
     if (SANITIZER_WINDOWS && globals[i].beg == 0) {
       // Skip globals that look like padding from the MSVC incremental linker.
@@ -424,7 +425,7 @@ void __asan_before_dynamic_init(const char *module_name) {
   bool strict_init_order = flags()->strict_init_order;
   CHECK(module_name);
   CHECK(asan_inited);
-  BlockingMutexLock lock(&mu_for_globals);
+  Lock lock(&mu_for_globals);
   if (flags()->report_globals >= 3)
     Printf("DynInitPoison module: %s\n", module_name);
   for (uptr i = 0, n = dynamic_init_globals->size(); i < n; ++i) {
@@ -448,7 +449,7 @@ void __asan_after_dynamic_init() {
       !dynamic_init_globals)
     return;
   CHECK(asan_inited);
-  BlockingMutexLock lock(&mu_for_globals);
+  Lock lock(&mu_for_globals);
   // FIXME: Optionally report that we're unpoisoning globals from a module.
   for (uptr i = 0, n = dynamic_init_globals->size(); i < n; ++i) {
     DynInitGlobal &dyn_g = (*dynamic_init_globals)[i];
diff --git a/compiler-rt/lib/asan/asan_interceptors.cpp b/compiler-rt/lib/asan/asan_interceptors.cpp
index d0a6dd48a748..b28909152e20 100644
--- a/compiler-rt/lib/asan/asan_interceptors.cpp
+++ b/compiler-rt/lib/asan/asan_interceptors.cpp
@@ -49,8 +49,8 @@ namespace __asan {
   ASAN_READ_RANGE((ctx), (s),                                   \
     common_flags()->strict_string_checks ? (len) + 1 : (n))
 
-#define ASAN_READ_STRING(ctx, s, n)                             \
-  ASAN_READ_STRING_OF_LEN((ctx), (s), REAL(strlen)(s), (n))
+#  define ASAN_READ_STRING(ctx, s, n) \
+    ASAN_READ_STRING_OF_LEN((ctx), (s), internal_strlen(s), (n))
 
 static inline uptr MaybeRealStrnlen(const char *s, uptr maxlen) {
 #if SANITIZER_INTERCEPT_STRNLEN
@@ -370,9 +370,9 @@ DEFINE_REAL(char*, index, const char *string, int c)
     ASAN_INTERCEPTOR_ENTER(ctx, strcat);
     ENSURE_ASAN_INITED();
     if (flags()->replace_str) {
-      uptr from_length = REAL(strlen)(from);
+      uptr from_length = internal_strlen(from);
       ASAN_READ_RANGE(ctx, from, from_length + 1);
-      uptr to_length = REAL(strlen)(to);
+      uptr to_length = internal_strlen(to);
       ASAN_READ_STRING_OF_LEN(ctx, to, to_length, to_length);
       ASAN_WRITE_RANGE(ctx, to + to_length, from_length + 1);
       // If the copying actually happens, the |from| string should not overlap
@@ -394,7 +394,7 @@ INTERCEPTOR(char*, strncat, char *to, const char *from, uptr size) {
     uptr from_length = MaybeRealStrnlen(from, size);
     uptr copy_length = Min(size, from_length + 1);
     ASAN_READ_RANGE(ctx, from, copy_length);
-    uptr to_length = REAL(strlen)(to);
+    uptr to_length = internal_strlen(to);
     ASAN_READ_STRING_OF_LEN(ctx, to, to_length, to_length);
     ASAN_WRITE_RANGE(ctx, to + to_length, from_length + 1);
     if (from_length > 0) {
@@ -419,7 +419,7 @@ INTERCEPTOR(char *, strcpy, char *to, const char *from) {
   }
   ENSURE_ASAN_INITED();
   if (flags()->replace_str) {
-    uptr from_size = REAL(strlen)(from) + 1;
+    uptr from_size = internal_strlen(from) + 1;
     CHECK_RANGES_OVERLAP("strcpy", to, from_size, from, from_size);
     ASAN_READ_RANGE(ctx, from, from_size);
     ASAN_WRITE_RANGE(ctx, to, from_size);
@@ -432,7 +432,7 @@ INTERCEPTOR(char*, strdup, const char *s) {
   ASAN_INTERCEPTOR_ENTER(ctx, strdup);
   if (UNLIKELY(!asan_inited)) return internal_strdup(s);
   ENSURE_ASAN_INITED();
-  uptr length = REAL(strlen)(s);
+  uptr length = internal_strlen(s);
   if (flags()->replace_str) {
     ASAN_READ_RANGE(ctx, s, length + 1);
   }
@@ -448,7 +448,7 @@ INTERCEPTOR(char*, __strdup, const char *s) {
   ASAN_INTERCEPTOR_ENTER(ctx, strdup);
   if (UNLIKELY(!asan_inited)) return internal_strdup(s);
   ENSURE_ASAN_INITED();
-  uptr length = REAL(strlen)(s);
+  uptr length = internal_strlen(s);
   if (flags()->replace_str) {
     ASAN_READ_RANGE(ctx, s, length + 1);
   }
@@ -581,7 +581,7 @@ INTERCEPTOR(int, atexit, void (*func)()) {
 #if CAN_SANITIZE_LEAKS
   __lsan::ScopedInterceptorDisabler disabler;
 #endif
-  // Avoid calling real atexit as it is unrechable on at least on Linux.
+  // Avoid calling real atexit as it is unreachable on at least on Linux.
   int res = REAL(__cxa_atexit)((void (*)(void *a))func, nullptr, nullptr);
   REAL(__cxa_atexit)(AtCxaAtexit, nullptr, nullptr);
   return res;
diff --git a/compiler-rt/lib/asan/asan_interceptors.h b/compiler-rt/lib/asan/asan_interceptors.h
index a9249dea45b9..047b044c8bf4 100644
--- a/compiler-rt/lib/asan/asan_interceptors.h
+++ b/compiler-rt/lib/asan/asan_interceptors.h
@@ -133,29 +133,30 @@ DECLARE_REAL(char*, strncpy, char *to, const char *from, uptr size)
 DECLARE_REAL(uptr, strnlen, const char *s, uptr maxlen)
 DECLARE_REAL(char*, strstr, const char *s1, const char *s2)
 
-#if !SANITIZER_MAC
-#define ASAN_INTERCEPT_FUNC(name)                                        \
-  do {                                                                   \
-    if (!INTERCEPT_FUNCTION(name))                                       \
-      VReport(1, "AddressSanitizer: failed to intercept '%s'\n", #name); \
-  } while (0)
-#define ASAN_INTERCEPT_FUNC_VER(name, ver)                                  \
-  do {                                                                      \
-    if (!INTERCEPT_FUNCTION_VER(name, ver))                                 \
-      VReport(1, "AddressSanitizer: failed to intercept '%s@@%s'\n", #name, \
-              #ver);                                                        \
-  } while (0)
-#define ASAN_INTERCEPT_FUNC_VER_UNVERSIONED_FALLBACK(name, ver)              \
-  do {                                                                       \
-    if (!INTERCEPT_FUNCTION_VER(name, ver) && !INTERCEPT_FUNCTION(name))     \
-      VReport(1, "AddressSanitizer: failed to intercept '%s@@%s' or '%s'\n", \
-              #name, #ver, #name);                                           \
-  } while (0)
+#  if !SANITIZER_MAC
+#    define ASAN_INTERCEPT_FUNC(name)                                        \
+      do {                                                                   \
+        if (!INTERCEPT_FUNCTION(name))                                       \
+          VReport(1, "AddressSanitizer: failed to intercept '%s'\n", #name); \
+      } while (0)
+#    define ASAN_INTERCEPT_FUNC_VER(name, ver)                           \
+      do {                                                               \
+        if (!INTERCEPT_FUNCTION_VER(name, ver))                          \
+          VReport(1, "AddressSanitizer: failed to intercept '%s@@%s'\n", \
+                  #name, ver);                                           \
+      } while (0)
+#    define ASAN_INTERCEPT_FUNC_VER_UNVERSIONED_FALLBACK(name, ver)           \
+      do {                                                                    \
+        if (!INTERCEPT_FUNCTION_VER(name, ver) && !INTERCEPT_FUNCTION(name))  \
+          VReport(1,                                                          \
+                  "AddressSanitizer: failed to intercept '%s@@%s' or '%s'\n", \
+                  #name, ver, #name);                                         \
+      } while (0)
 
-#else
+#  else
 // OS X interceptors don't need to be initialized with INTERCEPT_FUNCTION.
-#define ASAN_INTERCEPT_FUNC(name)
-#endif  // SANITIZER_MAC
+#    define ASAN_INTERCEPT_FUNC(name)
+#  endif  // SANITIZER_MAC
 
 #endif  // !SANITIZER_FUCHSIA
 
diff --git a/compiler-rt/lib/asan/asan_linux.cpp b/compiler-rt/lib/asan/asan_linux.cpp
index 4bcbe5d02e33..ad3693d5e6a2 100644
--- a/compiler-rt/lib/asan/asan_linux.cpp
+++ b/compiler-rt/lib/asan/asan_linux.cpp
@@ -128,8 +128,8 @@ void AsanCheckIncompatibleRT() {}
 #else
 static int FindFirstDSOCallback(struct dl_phdr_info *info, size_t size,
                                 void *data) {
-  VReport(2, "info->dlpi_name = %s\tinfo->dlpi_addr = %p\n",
-          info->dlpi_name, info->dlpi_addr);
+  VReport(2, "info->dlpi_name = %s\tinfo->dlpi_addr = %p\n", info->dlpi_name,
+          (void *)info->dlpi_addr);
 
   // Continue until the first dynamic library is found
   if (!info->dlpi_name || info->dlpi_name[0] == 0)
diff --git a/compiler-rt/lib/asan/asan_malloc_linux.cpp b/compiler-rt/lib/asan/asan_malloc_linux.cpp
index c6bec8551bc5..bab80b96f584 100644
--- a/compiler-rt/lib/asan/asan_malloc_linux.cpp
+++ b/compiler-rt/lib/asan/asan_malloc_linux.cpp
@@ -21,129 +21,66 @@
 #  include "asan_interceptors.h"
 #  include "asan_internal.h"
 #  include "asan_stack.h"
+#  include "lsan/lsan_common.h"
 #  include "sanitizer_common/sanitizer_allocator_checks.h"
+#  include "sanitizer_common/sanitizer_allocator_dlsym.h"
 #  include "sanitizer_common/sanitizer_errno.h"
 #  include "sanitizer_common/sanitizer_tls_get_addr.h"
 
 // ---------------------- Replacement functions ---------------- {{{1
 using namespace __asan;
 
-static uptr allocated_for_dlsym;
-static uptr last_dlsym_alloc_size_in_words;
-static const uptr kDlsymAllocPoolSize = 1024;
-static uptr alloc_memory_for_dlsym[kDlsymAllocPoolSize];
-
-static inline bool IsInDlsymAllocPool(const void *ptr) {
-  uptr off = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-  return off < allocated_for_dlsym * sizeof(alloc_memory_for_dlsym[0]);
-}
-
-static void *AllocateFromLocalPool(uptr size_in_bytes) {
-  uptr size_in_words = RoundUpTo(size_in_bytes, kWordSize) / kWordSize;
-  void *mem = (void*)&alloc_memory_for_dlsym[allocated_for_dlsym];
-  last_dlsym_alloc_size_in_words = size_in_words;
-  allocated_for_dlsym += size_in_words;
-  CHECK_LT(allocated_for_dlsym, kDlsymAllocPoolSize);
-  return mem;
-}
-
-static void DeallocateFromLocalPool(const void *ptr) {
-  // Hack: since glibc 2.27 dlsym no longer uses stack-allocated memory to store
-  // error messages and instead uses malloc followed by free. To avoid pool
-  // exhaustion due to long object filenames, handle that special case here.
-  uptr prev_offset = allocated_for_dlsym - last_dlsym_alloc_size_in_words;
-  void *prev_mem = (void*)&alloc_memory_for_dlsym[prev_offset];
-  if (prev_mem == ptr) {
-    REAL(memset)(prev_mem, 0, last_dlsym_alloc_size_in_words * kWordSize);
-    allocated_for_dlsym = prev_offset;
-    last_dlsym_alloc_size_in_words = 0;
+struct DlsymAlloc : public DlSymAllocator<DlsymAlloc> {
+  static bool UseImpl() { return asan_init_is_running; }
+  static void OnAllocate(const void *ptr, uptr size) {
+#  if CAN_SANITIZE_LEAKS
+    // Suppress leaks from dlerror(). Previously dlsym hack on global array was
+    // used by leak sanitizer as a root region.
+    __lsan_register_root_region(ptr, size);
+#  endif
   }
-}
-
-static int PosixMemalignFromLocalPool(void **memptr, uptr alignment,
-                                      uptr size_in_bytes) {
-  if (UNLIKELY(!CheckPosixMemalignAlignment(alignment)))
-    return errno_EINVAL;
-
-  CHECK(alignment >= kWordSize);
-
-  uptr addr = (uptr)&alloc_memory_for_dlsym[allocated_for_dlsym];
-  uptr aligned_addr = RoundUpTo(addr, alignment);
-  uptr aligned_size = RoundUpTo(size_in_bytes, kWordSize);
-
-  uptr *end_mem = (uptr*)(aligned_addr + aligned_size);
-  uptr allocated = end_mem - alloc_memory_for_dlsym;
-  if (allocated >= kDlsymAllocPoolSize)
-    return errno_ENOMEM;
-
-  allocated_for_dlsym = allocated;
-  *memptr = (void*)aligned_addr;
-  return 0;
-}
-
-static inline bool MaybeInDlsym() {
-  // Fuchsia doesn't use dlsym-based interceptors.
-  return !SANITIZER_FUCHSIA && asan_init_is_running;
-}
-
-static inline bool UseLocalPool() { return MaybeInDlsym(); }
-
-static void *ReallocFromLocalPool(void *ptr, uptr size) {
-  const uptr offset = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-  const uptr copy_size = Min(size, kDlsymAllocPoolSize - offset);
-  void *new_ptr;
-  if (UNLIKELY(UseLocalPool())) {
-    new_ptr = AllocateFromLocalPool(size);
-  } else {
-    ENSURE_ASAN_INITED();
-    GET_STACK_TRACE_MALLOC;
-    new_ptr = asan_malloc(size, &stack);
+  static void OnFree(const void *ptr, uptr size) {
+#  if CAN_SANITIZE_LEAKS
+    __lsan_unregister_root_region(ptr, size);
+#  endif
   }
-  internal_memcpy(new_ptr, ptr, copy_size);
-  return new_ptr;
-}
+};
 
 INTERCEPTOR(void, free, void *ptr) {
-  if (UNLIKELY(IsInDlsymAllocPool(ptr))) {
-    DeallocateFromLocalPool(ptr);
-    return;
-  }
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
   GET_STACK_TRACE_FREE;
   asan_free(ptr, &stack, FROM_MALLOC);
 }
 
 #if SANITIZER_INTERCEPT_CFREE
 INTERCEPTOR(void, cfree, void *ptr) {
-  if (UNLIKELY(IsInDlsymAllocPool(ptr)))
-    return;
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
   GET_STACK_TRACE_FREE;
   asan_free(ptr, &stack, FROM_MALLOC);
 }
 #endif // SANITIZER_INTERCEPT_CFREE
 
 INTERCEPTOR(void*, malloc, uptr size) {
-  if (UNLIKELY(UseLocalPool()))
-    // Hack: dlsym calls malloc before REAL(malloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(size);
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Allocate(size);
   ENSURE_ASAN_INITED();
   GET_STACK_TRACE_MALLOC;
   return asan_malloc(size, &stack);
 }
 
 INTERCEPTOR(void*, calloc, uptr nmemb, uptr size) {
-  if (UNLIKELY(UseLocalPool()))
-    // Hack: dlsym calls calloc before REAL(calloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(nmemb * size);
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Callocate(nmemb, size);
   ENSURE_ASAN_INITED();
   GET_STACK_TRACE_MALLOC;
   return asan_calloc(nmemb, size, &stack);
 }
 
 INTERCEPTOR(void*, realloc, void *ptr, uptr size) {
-  if (UNLIKELY(IsInDlsymAllocPool(ptr)))
-    return ReallocFromLocalPool(ptr, size);
-  if (UNLIKELY(UseLocalPool()))
-    return AllocateFromLocalPool(size);
+  if (DlsymAlloc::Use() || DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Realloc(ptr, size);
   ENSURE_ASAN_INITED();
   GET_STACK_TRACE_MALLOC;
   return asan_realloc(ptr, size, &stack);
@@ -205,8 +142,6 @@ INTERCEPTOR(int, mallopt, int cmd, int value) {
 #endif // SANITIZER_INTERCEPT_MALLOPT_AND_MALLINFO
 
 INTERCEPTOR(int, posix_memalign, void **memptr, uptr alignment, uptr size) {
-  if (UNLIKELY(UseLocalPool()))
-    return PosixMemalignFromLocalPool(memptr, alignment, size);
   GET_STACK_TRACE_MALLOC;
   return asan_posix_memalign(memptr, alignment, size, &stack);
 }
diff --git a/compiler-rt/lib/asan/asan_poisoning.cpp b/compiler-rt/lib/asan/asan_poisoning.cpp
index 5f215fe0f9bb..d97af91e692d 100644
--- a/compiler-rt/lib/asan/asan_poisoning.cpp
+++ b/compiler-rt/lib/asan/asan_poisoning.cpp
@@ -66,7 +66,7 @@ void AsanPoisonOrUnpoisonIntraObjectRedzone(uptr ptr, uptr size, bool poison) {
   uptr end = ptr + size;
   if (Verbosity()) {
     Printf("__asan_%spoison_intra_object_redzone [%p,%p) %zd\n",
-           poison ? "" : "un", ptr, end, size);
+           poison ? "" : "un", (void *)ptr, (void *)end, size);
     if (Verbosity() >= 2)
       PRINT_CURRENT_STACK();
   }
diff --git a/compiler-rt/lib/asan/asan_report.cpp b/compiler-rt/lib/asan/asan_report.cpp
index 03f1ed2b0186..1f266334b311 100644
--- a/compiler-rt/lib/asan/asan_report.cpp
+++ b/compiler-rt/lib/asan/asan_report.cpp
@@ -32,12 +32,12 @@ namespace __asan {
 static void (*error_report_callback)(const char*);
 static char *error_message_buffer = nullptr;
 static uptr error_message_buffer_pos = 0;
-static BlockingMutex error_message_buf_mutex(LINKER_INITIALIZED);
+static Mutex error_message_buf_mutex;
 static const unsigned kAsanBuggyPcPoolSize = 25;
 static __sanitizer::atomic_uintptr_t AsanBuggyPcPool[kAsanBuggyPcPoolSize];
 
 void AppendToErrorMessageBuffer(const char *buffer) {
-  BlockingMutexLock l(&error_message_buf_mutex);
+  Lock l(&error_message_buf_mutex);
   if (!error_message_buffer) {
     error_message_buffer =
       (char*)MmapOrDieQuietly(kErrorMessageBufferSize, __func__);
@@ -67,14 +67,14 @@ static void PrintZoneForPointer(uptr ptr, uptr zone_ptr,
                                 const char *zone_name) {
   if (zone_ptr) {
     if (zone_name) {
-      Printf("malloc_zone_from_ptr(%p) = %p, which is %s\n",
-                 ptr, zone_ptr, zone_name);
+      Printf("malloc_zone_from_ptr(%p) = %p, which is %s\n", (void *)ptr,
+             (void *)zone_ptr, zone_name);
     } else {
       Printf("malloc_zone_from_ptr(%p) = %p, which doesn't have a name\n",
-                 ptr, zone_ptr);
+             (void *)ptr, (void *)zone_ptr);
     }
   } else {
-    Printf("malloc_zone_from_ptr(%p) = 0\n", ptr);
+    Printf("malloc_zone_from_ptr(%p) = 0\n", (void *)ptr);
   }
 }
 
@@ -155,10 +155,10 @@ class ScopedInErrorReport {
       DumpProcessMap();
 
     // Copy the message buffer so that we could start logging without holding a
-    // lock that gets aquired during printing.
+    // lock that gets acquired during printing.
     InternalMmapVector<char> buffer_copy(kErrorMessageBufferSize);
     {
-      BlockingMutexLock l(&error_message_buf_mutex);
+      Lock l(&error_message_buf_mutex);
       internal_memcpy(buffer_copy.data(),
                       error_message_buffer, kErrorMessageBufferSize);
       // Clear error_message_buffer so that if we find other errors
@@ -435,9 +435,10 @@ static inline void CheckForInvalidPointerPair(void *p1, void *p2) {
 void ReportMacMzReallocUnknown(uptr addr, uptr zone_ptr, const char *zone_name,
                                BufferedStackTrace *stack) {
   ScopedInErrorReport in_report;
-  Printf("mz_realloc(%p) -- attempting to realloc unallocated memory.\n"
-             "This is an unrecoverable problem, exiting now.\n",
-             addr);
+  Printf(
+      "mz_realloc(%p) -- attempting to realloc unallocated memory.\n"
+      "This is an unrecoverable problem, exiting now.\n",
+      (void *)addr);
   PrintZoneForPointer(addr, zone_ptr, zone_name);
   stack->Print();
   DescribeAddressIfHeap(addr);
@@ -490,7 +491,7 @@ void __asan_report_error(uptr pc, uptr bp, uptr sp, uptr addr, int is_write,
 }
 
 void NOINLINE __asan_set_error_report_callback(void (*callback)(const char*)) {
-  BlockingMutexLock l(&error_message_buf_mutex);
+  Lock l(&error_message_buf_mutex);
   error_report_callback = callback;
 }
 
diff --git a/compiler-rt/lib/asan/asan_rtl.cpp b/compiler-rt/lib/asan/asan_rtl.cpp
index bfaa3bc27027..1b150b393cfe 100644
--- a/compiler-rt/lib/asan/asan_rtl.cpp
+++ b/compiler-rt/lib/asan/asan_rtl.cpp
@@ -557,7 +557,8 @@ void UnpoisonStack(uptr bottom, uptr top, const char *type) {
         "False positive error reports may follow\n"
         "For details see "
         "https://github.com/google/sanitizers/issues/189\n",
-        type, top, bottom, top - bottom, top - bottom);
+        type, (void *)top, (void *)bottom, (void *)(top - bottom),
+        top - bottom);
     return;
   }
   PoisonShadow(bottom, RoundUpTo(top - bottom, SHADOW_GRANULARITY), 0);
diff --git a/compiler-rt/lib/asan/asan_shadow_setup.cpp b/compiler-rt/lib/asan/asan_shadow_setup.cpp
index 6e6260d3413f..fc6de39622b5 100644
--- a/compiler-rt/lib/asan/asan_shadow_setup.cpp
+++ b/compiler-rt/lib/asan/asan_shadow_setup.cpp
@@ -33,7 +33,7 @@ static void ProtectGap(uptr addr, uptr size) {
           "protect_shadow_gap=0:"
           " not protecting shadow gap, allocating gap's shadow\n"
           "|| `[%p, %p]` || ShadowGap's shadow ||\n",
-          GapShadowBeg, GapShadowEnd);
+          (void*)GapShadowBeg, (void*)GapShadowEnd);
     ReserveShadowMemoryRange(GapShadowBeg, GapShadowEnd,
                              "unprotected gap shadow");
     return;
@@ -113,7 +113,7 @@ void InitializeShadowMemory() {
         "Shadow memory range interleaves with an existing memory mapping. "
         "ASan cannot proceed correctly. ABORTING.\n");
     Report("ASan shadow was supposed to be located in the [%p-%p] range.\n",
-           shadow_start, kHighShadowEnd);
+           (void*)shadow_start, (void*)kHighShadowEnd);
     MaybeReportLinuxPIEBug();
     DumpProcessMap();
     Die();
diff --git a/compiler-rt/lib/asan/asan_stats.cpp b/compiler-rt/lib/asan/asan_stats.cpp
index 00ded8f5ef50..9a715ea76fee 100644
--- a/compiler-rt/lib/asan/asan_stats.cpp
+++ b/compiler-rt/lib/asan/asan_stats.cpp
@@ -62,11 +62,11 @@ void AsanStats::MergeFrom(const AsanStats *stats) {
     dst_ptr[i] += src_ptr[i];
 }
 
-static BlockingMutex print_lock(LINKER_INITIALIZED);
+static Mutex print_lock;
 
 static AsanStats unknown_thread_stats(LINKER_INITIALIZED);
 static AsanStats dead_threads_stats(LINKER_INITIALIZED);
-static BlockingMutex dead_threads_stats_lock(LINKER_INITIALIZED);
+static Mutex dead_threads_stats_lock;
 // Required for malloc_zone_statistics() on OS X. This can't be stored in
 // per-thread AsanStats.
 static uptr max_malloced_memory;
@@ -87,7 +87,7 @@ static void GetAccumulatedStats(AsanStats *stats) {
   }
   stats->MergeFrom(&unknown_thread_stats);
   {
-    BlockingMutexLock lock(&dead_threads_stats_lock);
+    Lock lock(&dead_threads_stats_lock);
     stats->MergeFrom(&dead_threads_stats);
   }
   // This is not very accurate: we may miss allocation peaks that happen
@@ -99,7 +99,7 @@ static void GetAccumulatedStats(AsanStats *stats) {
 }
 
 void FlushToDeadThreadStats(AsanStats *stats) {
-  BlockingMutexLock lock(&dead_threads_stats_lock);
+  Lock lock(&dead_threads_stats_lock);
   dead_threads_stats.MergeFrom(stats);
   stats->Clear();
 }
@@ -122,11 +122,11 @@ static void PrintAccumulatedStats() {
   AsanStats stats;
   GetAccumulatedStats(&stats);
   // Use lock to keep reports from mixing up.
-  BlockingMutexLock lock(&print_lock);
+  Lock lock(&print_lock);
   stats.Print();
-  StackDepotStats *stack_depot_stats = StackDepotGetStats();
+  StackDepotStats stack_depot_stats = StackDepotGetStats();
   Printf("Stats: StackDepot: %zd ids; %zdM allocated\n",
-         stack_depot_stats->n_uniq_ids, stack_depot_stats->allocated >> 20);
+         stack_depot_stats.n_uniq_ids, stack_depot_stats.allocated >> 20);
   PrintInternalAllocatorStats();
 }
 
diff --git a/compiler-rt/lib/asan/asan_thread.cpp b/compiler-rt/lib/asan/asan_thread.cpp
index 35d4467e7b53..930139968ec3 100644
--- a/compiler-rt/lib/asan/asan_thread.cpp
+++ b/compiler-rt/lib/asan/asan_thread.cpp
@@ -43,11 +43,11 @@ void AsanThreadContext::OnFinished() {
 static ALIGNED(16) char thread_registry_placeholder[sizeof(ThreadRegistry)];
 static ThreadRegistry *asan_thread_registry;
 
-static BlockingMutex mu_for_thread_context(LINKER_INITIALIZED);
+static Mutex mu_for_thread_context;
 static LowLevelAllocator allocator_for_thread_context;
 
 static ThreadContextBase *GetAsanThreadContext(u32 tid) {
-  BlockingMutexLock lock(&mu_for_thread_context);
+  Lock lock(&mu_for_thread_context);
   return new(allocator_for_thread_context) AsanThreadContext(tid);
 }
 
@@ -83,8 +83,7 @@ AsanThread *AsanThread::Create(thread_callback_t start_routine, void *arg,
   thread->start_routine_ = start_routine;
   thread->arg_ = arg;
   AsanThreadContext::CreateThreadContextArgs args = {thread, stack};
-  asanThreadRegistry().CreateThread(*reinterpret_cast<uptr *>(thread), detached,
-                                    parent_tid, &args);
+  asanThreadRegistry().CreateThread(0, detached, parent_tid, &args);
 
   return thread;
 }
@@ -254,7 +253,7 @@ void AsanThread::Init(const InitOptions *options) {
   int local = 0;
   VReport(1, "T%d: stack [%p,%p) size 0x%zx; local=%p\n", tid(),
           (void *)stack_bottom_, (void *)stack_top_, stack_top_ - stack_bottom_,
-          &local);
+          (void *)&local);
 }
 
 // Fuchsia doesn't use ThreadStart.
@@ -443,7 +442,7 @@ AsanThread *GetCurrentThread() {
 
 void SetCurrentThread(AsanThread *t) {
   CHECK(t->context());
-  VReport(2, "SetCurrentThread: %p for thread %p\n", t->context(),
+  VReport(2, "SetCurrentThread: %p for thread %p\n", (void *)t->context(),
           (void *)GetThreadSelf());
   // Make sure we do not reset the current AsanThread.
   CHECK_EQ(0, AsanTSDGet());
diff --git a/compiler-rt/lib/builtins/README.txt b/compiler-rt/lib/builtins/README.txt
index d66d725e7ab5..53d656d5086d 100644
--- a/compiler-rt/lib/builtins/README.txt
+++ b/compiler-rt/lib/builtins/README.txt
@@ -271,8 +271,8 @@ switchu8
 
 // There is no C interface to the *_vfp_d8_d15_regs functions.  There are
 // called in the prolog and epilog of Thumb1 functions.  When the C++ ABI use
-// SJLJ for exceptions, each function with a catch clause or destuctors needs
-// to save and restore all registers in it prolog and epliog.  But there is 
+// SJLJ for exceptions, each function with a catch clause or destructors needs
+// to save and restore all registers in it prolog and epilog.  But there is
 // no way to access vector and high float registers from thumb1 code, so the 
 // compiler must add call outs to these helper functions in the prolog and 
 // epilog.
@@ -311,9 +311,9 @@ double __floatsidfvfp(int a);           // Appears to convert from
 float __floatsisfvfp(int a);            // Appears to convert from
                                         //     int to float.
 double __floatunssidfvfp(unsigned int a); // Appears to convert from
-                                        //     unisgned int to double.
+                                        //     unsigned int to double.
 float __floatunssisfvfp(unsigned int a); // Appears to convert from
-                                        //     unisgned int to float.
+                                        //     unsigned int to float.
 int __gedf2vfp(double a, double b);     // Appears to return __gedf2
                                         //     (a >= b)
 int __gesf2vfp(float a, float b);       // Appears to return __gesf2
diff --git a/compiler-rt/lib/builtins/arm/truncdfsf2vfp.S b/compiler-rt/lib/builtins/arm/truncdfsf2vfp.S
index a3c0a73466e9..e1c171262a78 100644
--- a/compiler-rt/lib/builtins/arm/truncdfsf2vfp.S
+++ b/compiler-rt/lib/builtins/arm/truncdfsf2vfp.S
@@ -11,9 +11,9 @@
 //
 // extern float __truncdfsf2vfp(double a);
 //
-// Converts double precision float to signle precision result.
+// Converts double precision float to single precision result.
 // Uses Darwin calling convention where a double precision parameter is
-// passed in a R0/R1 pair and a signle precision result is returned in R0.
+// passed in a R0/R1 pair and a single precision result is returned in R0.
 //
 	.syntax unified
 	.p2align 2
diff --git a/compiler-rt/lib/builtins/atomic.c b/compiler-rt/lib/builtins/atomic.c
index 64bf72dfa345..4c3ebb99a513 100644
--- a/compiler-rt/lib/builtins/atomic.c
+++ b/compiler-rt/lib/builtins/atomic.c
@@ -336,6 +336,18 @@ OPTIMISED_CASES
     return tmp;                                                                \
   }
 
+#define ATOMIC_RMW_NAND(n, lockfree, type)                                     \
+  type __atomic_fetch_nand_##n(type *ptr, type val, int model) {               \
+    if (lockfree(ptr))                                                         \
+      return __c11_atomic_fetch_nand((_Atomic(type) *)ptr, val, model);        \
+    Lock *l = lock_for_pointer(ptr);                                           \
+    lock(l);                                                                   \
+    type tmp = *ptr;                                                           \
+    *ptr = ~(tmp & val);                                                       \
+    unlock(l);                                                                 \
+    return tmp;                                                                \
+  }
+
 #define OPTIMISED_CASE(n, lockfree, type) ATOMIC_RMW(n, lockfree, type, add, +)
 OPTIMISED_CASES
 #undef OPTIMISED_CASE
@@ -351,3 +363,6 @@ OPTIMISED_CASES
 #define OPTIMISED_CASE(n, lockfree, type) ATOMIC_RMW(n, lockfree, type, xor, ^)
 OPTIMISED_CASES
 #undef OPTIMISED_CASE
+#define OPTIMISED_CASE(n, lockfree, type) ATOMIC_RMW_NAND(n, lockfree, type)
+OPTIMISED_CASES
+#undef OPTIMISED_CASE
diff --git a/compiler-rt/lib/builtins/clear_cache.c b/compiler-rt/lib/builtins/clear_cache.c
index 3c12b74e8fa6..da0715914b41 100644
--- a/compiler-rt/lib/builtins/clear_cache.c
+++ b/compiler-rt/lib/builtins/clear_cache.c
@@ -35,7 +35,7 @@ uintptr_t GetCurrentProcess(void);
 #include <machine/sysarch.h>
 #endif
 
-#if defined(__OpenBSD__) && (defined(__arm__) || defined(__mips__))
+#if defined(__OpenBSD__) && (defined(__arm__) || defined(__mips__) || defined(__riscv))
 // clang-format off
 #include <sys/types.h>
 #include <machine/sysarch.h>
@@ -166,6 +166,13 @@ void __clear_cache(void *start, void *end) {
                    : "=r"(start_reg)
                    : "r"(start_reg), "r"(end_reg), "r"(flags), "r"(syscall_nr));
   assert(start_reg == 0 && "Cache flush syscall failed.");
+#elif defined(__riscv) && defined(__OpenBSD__)
+  struct riscv_sync_icache_args arg;
+
+  arg.addr = (uintptr_t)start;
+  arg.len = (uintptr_t)end - (uintptr_t)start;
+
+  sysarch(RISCV_SYNC_ICACHE, &arg);
 #else
 #if __APPLE__
   // On Darwin, sys_icache_invalidate() provides this functionality
diff --git a/compiler-rt/lib/builtins/cpu_model.c b/compiler-rt/lib/builtins/cpu_model.c
index 6ee42911b204..b8d807ed651c 100644
--- a/compiler-rt/lib/builtins/cpu_model.c
+++ b/compiler-rt/lib/builtins/cpu_model.c
@@ -422,6 +422,22 @@ getIntelProcessorTypeAndSubtype(unsigned Family, unsigned Model,
       *Subtype = INTEL_COREI7_ICELAKE_CLIENT;
       break;
 
+    // Tigerlake:
+    case 0x8c:
+    case 0x8d:
+      CPU = "tigerlake";
+      *Type = INTEL_COREI7;
+      *Subtype = INTEL_COREI7_TIGERLAKE;
+      break;
+
+    // Alderlake:
+    case 0x97:
+    case 0x9a:
+      CPU = "alderlake";
+      *Type = INTEL_COREI7;
+      *Subtype = INTEL_COREI7_ALDERLAKE;
+      break;
+
     // Icelake Xeon:
     case 0x6a:
     case 0x6c:
diff --git a/compiler-rt/lib/builtins/emutls.c b/compiler-rt/lib/builtins/emutls.c
index 98cabd917d6c..e112fdf51440 100644
--- a/compiler-rt/lib/builtins/emutls.c
+++ b/compiler-rt/lib/builtins/emutls.c
@@ -150,7 +150,7 @@ static void win_error(DWORD last_err, const char *hint) {
                      NULL, last_err, 0, (LPSTR)&buffer, 1, NULL)) {
     fprintf(stderr, "Windows error: %s\n", buffer);
   } else {
-    fprintf(stderr, "Unkown Windows error: %s\n", hint);
+    fprintf(stderr, "Unknown Windows error: %s\n", hint);
   }
   LocalFree(buffer);
 }
@@ -374,6 +374,21 @@ emutls_get_address_array(uintptr_t index) {
   return array;
 }
 
+#ifndef _WIN32
+// Our emulated TLS implementation relies on local state (e.g. for the pthread
+// key), and if we duplicate this state across different shared libraries,
+// accesses to the same TLS variable from different shared libraries will yield
+// different results (see https://github.com/android/ndk/issues/1551 for an
+// example). __emutls_get_address is the only external entry point for emulated
+// TLS, and by making it default visibility and weak, we can rely on the dynamic
+// linker to coalesce multiple copies at runtime and ensure a single unique copy
+// of TLS state. This is a best effort; it won't work if the user is linking
+// with -Bsymbolic or -Bsymbolic-functions, and it also won't work on Windows,
+// where the dynamic linker has no notion of coalescing weak symbols at runtime.
+// A more robust solution would be to create a separate shared library for
+// emulated TLS, to ensure a single copy of its state.
+__attribute__((visibility("default"), weak))
+#endif
 void *__emutls_get_address(__emutls_control *control) {
   uintptr_t index = emutls_get_index(control);
   emutls_address_array *array = emutls_get_address_array(index--);
diff --git a/compiler-rt/lib/builtins/fixdfdi.c b/compiler-rt/lib/builtins/fixdfdi.c
index 511568fc12fd..a48facb68598 100644
--- a/compiler-rt/lib/builtins/fixdfdi.c
+++ b/compiler-rt/lib/builtins/fixdfdi.c
@@ -42,3 +42,7 @@ AEABI_RTABI di_int __aeabi_d2lz(fp_t a) { return __fixdfdi(a); }
 COMPILER_RT_ALIAS(__fixdfdi, __aeabi_d2lz)
 #endif
 #endif
+
+#if defined(__MINGW32__) && defined(__arm__)
+COMPILER_RT_ALIAS(__fixdfdi, __dtoi64)
+#endif
diff --git a/compiler-rt/lib/builtins/fixsfdi.c b/compiler-rt/lib/builtins/fixsfdi.c
index 0cf71c30311a..3a66fb9e2f06 100644
--- a/compiler-rt/lib/builtins/fixsfdi.c
+++ b/compiler-rt/lib/builtins/fixsfdi.c
@@ -42,3 +42,7 @@ AEABI_RTABI di_int __aeabi_f2lz(fp_t a) { return __fixsfdi(a); }
 COMPILER_RT_ALIAS(__fixsfdi, __aeabi_f2lz)
 #endif
 #endif
+
+#if defined(__MINGW32__) && defined(__arm__)
+COMPILER_RT_ALIAS(__fixsfdi, __stoi64)
+#endif
diff --git a/compiler-rt/lib/builtins/fixunsdfdi.c b/compiler-rt/lib/builtins/fixunsdfdi.c
index ccb256d2c7e0..f15f86788e85 100644
--- a/compiler-rt/lib/builtins/fixunsdfdi.c
+++ b/compiler-rt/lib/builtins/fixunsdfdi.c
@@ -40,3 +40,7 @@ AEABI_RTABI du_int __aeabi_d2ulz(fp_t a) { return __fixunsdfdi(a); }
 COMPILER_RT_ALIAS(__fixunsdfdi, __aeabi_d2ulz)
 #endif
 #endif
+
+#if defined(__MINGW32__) && defined(__arm__)
+COMPILER_RT_ALIAS(__fixunsdfdi, __dtou64)
+#endif
diff --git a/compiler-rt/lib/builtins/fixunssfdi.c b/compiler-rt/lib/builtins/fixunssfdi.c
index 647185fbabf1..e8f600df9766 100644
--- a/compiler-rt/lib/builtins/fixunssfdi.c
+++ b/compiler-rt/lib/builtins/fixunssfdi.c
@@ -41,3 +41,7 @@ AEABI_RTABI du_int __aeabi_f2ulz(fp_t a) { return __fixunssfdi(a); }
 COMPILER_RT_ALIAS(__fixunssfdi, __aeabi_f2ulz)
 #endif
 #endif
+
+#if defined(__MINGW32__) && defined(__arm__)
+COMPILER_RT_ALIAS(__fixunssfdi, __stou64)
+#endif
diff --git a/compiler-rt/lib/builtins/fixunsxfdi.c b/compiler-rt/lib/builtins/fixunsxfdi.c
index 097a4e55e931..c8a8061b2cf0 100644
--- a/compiler-rt/lib/builtins/fixunsxfdi.c
+++ b/compiler-rt/lib/builtins/fixunsxfdi.c
@@ -26,7 +26,7 @@
 // mmmm mmmm mmmm
 
 #if defined(_MSC_VER) && !defined(__clang__)
-// MSVC throws a warning about 'unitialized variable use' here,
+// MSVC throws a warning about 'uninitialized variable use' here,
 // disable it for builds that warn-as-error
 #pragma warning(push)
 #pragma warning(disable : 4700)
diff --git a/compiler-rt/lib/builtins/fixunsxfsi.c b/compiler-rt/lib/builtins/fixunsxfsi.c
index 3bc1288d38a1..154abcbd35e7 100644
--- a/compiler-rt/lib/builtins/fixunsxfsi.c
+++ b/compiler-rt/lib/builtins/fixunsxfsi.c
@@ -26,7 +26,7 @@
 // mmmm mmmm mmmm
 
 #if defined(_MSC_VER) && !defined(__clang__)
-// MSVC throws a warning about 'unitialized variable use' here,
+// MSVC throws a warning about 'uninitialized variable use' here,
 // disable it for builds that warn-as-error
 #pragma warning(push)
 #pragma warning(disable : 4700)
diff --git a/compiler-rt/lib/builtins/fixxfdi.c b/compiler-rt/lib/builtins/fixxfdi.c
index a7a0464feb9d..86cf3767b75d 100644
--- a/compiler-rt/lib/builtins/fixxfdi.c
+++ b/compiler-rt/lib/builtins/fixxfdi.c
@@ -25,7 +25,7 @@
 // mmmm mmmm mmmm
 
 #if defined(_MSC_VER) && !defined(__clang__)
-// MSVC throws a warning about 'unitialized variable use' here,
+// MSVC throws a warning about 'uninitialized variable use' here,
 // disable it for builds that warn-as-error
 #pragma warning(push)
 #pragma warning(disable : 4700)
diff --git a/compiler-rt/lib/builtins/floatdidf.c b/compiler-rt/lib/builtins/floatdidf.c
index 7ecb30bca71e..d37c43b1f2f9 100644
--- a/compiler-rt/lib/builtins/floatdidf.c
+++ b/compiler-rt/lib/builtins/floatdidf.c
@@ -101,3 +101,7 @@ AEABI_RTABI double __aeabi_l2d(di_int a) { return __floatdidf(a); }
 COMPILER_RT_ALIAS(__floatdidf, __aeabi_l2d)
 #endif
 #endif
+
+#if defined(__MINGW32__) && defined(__arm__)
+COMPILER_RT_ALIAS(__floatdidf, __i64tod)
+#endif
diff --git a/compiler-rt/lib/builtins/floatdisf.c b/compiler-rt/lib/builtins/floatdisf.c
index faaa1bcb3c8e..5c6316431e39 100644
--- a/compiler-rt/lib/builtins/floatdisf.c
+++ b/compiler-rt/lib/builtins/floatdisf.c
@@ -73,3 +73,7 @@ AEABI_RTABI float __aeabi_l2f(di_int a) { return __floatdisf(a); }
 COMPILER_RT_ALIAS(__floatdisf, __aeabi_l2f)
 #endif
 #endif
+
+#if defined(__MINGW32__) && defined(__arm__)
+COMPILER_RT_ALIAS(__floatdisf, __i64tos)
+#endif
diff --git a/compiler-rt/lib/builtins/floatundidf.c b/compiler-rt/lib/builtins/floatundidf.c
index e5e533042a34..2ec802cdc134 100644
--- a/compiler-rt/lib/builtins/floatundidf.c
+++ b/compiler-rt/lib/builtins/floatundidf.c
@@ -104,3 +104,7 @@ AEABI_RTABI double __aeabi_ul2d(du_int a) { return __floatundidf(a); }
 COMPILER_RT_ALIAS(__floatundidf, __aeabi_ul2d)
 #endif
 #endif
+
+#if defined(__MINGW32__) && defined(__arm__)
+COMPILER_RT_ALIAS(__floatundidf, __u64tod)
+#endif
diff --git a/compiler-rt/lib/builtins/floatundisf.c b/compiler-rt/lib/builtins/floatundisf.c
index 00d61b0c6310..2a4157dc5e4b 100644
--- a/compiler-rt/lib/builtins/floatundisf.c
+++ b/compiler-rt/lib/builtins/floatundisf.c
@@ -70,3 +70,7 @@ AEABI_RTABI float __aeabi_ul2f(du_int a) { return __floatundisf(a); }
 COMPILER_RT_ALIAS(__floatundisf, __aeabi_ul2f)
 #endif
 #endif
+
+#if defined(__MINGW32__) && defined(__arm__)
+COMPILER_RT_ALIAS(__floatundisf, __u64tos)
+#endif
diff --git a/compiler-rt/lib/builtins/mingw_fixfloat.c b/compiler-rt/lib/builtins/mingw_fixfloat.c
deleted file mode 100644
index 945be9d4344a..000000000000
--- a/compiler-rt/lib/builtins/mingw_fixfloat.c
+++ /dev/null
@@ -1,34 +0,0 @@
-//===-- mingw_fixfloat.c - Wrap int/float conversions for arm/windows -----===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-
-#include "int_lib.h"
-
-COMPILER_RT_ABI di_int __fixdfdi(double a);
-COMPILER_RT_ABI di_int __fixsfdi(float a);
-COMPILER_RT_ABI du_int __fixunsdfdi(double a);
-COMPILER_RT_ABI du_int __fixunssfdi(float a);
-COMPILER_RT_ABI double __floatdidf(di_int a);
-COMPILER_RT_ABI float __floatdisf(di_int a);
-COMPILER_RT_ABI double __floatundidf(du_int a);
-COMPILER_RT_ABI float __floatundisf(du_int a);
-
-COMPILER_RT_ABI di_int __dtoi64(double a) { return __fixdfdi(a); }
-
-COMPILER_RT_ABI di_int __stoi64(float a) { return __fixsfdi(a); }
-
-COMPILER_RT_ABI du_int __dtou64(double a) { return __fixunsdfdi(a); }
-
-COMPILER_RT_ABI du_int __stou64(float a) { return __fixunssfdi(a); }
-
-COMPILER_RT_ABI double __i64tod(di_int a) { return __floatdidf(a); }
-
-COMPILER_RT_ABI float __i64tos(di_int a) { return __floatdisf(a); }
-
-COMPILER_RT_ABI double __u64tod(du_int a) { return __floatundidf(a); }
-
-COMPILER_RT_ABI float __u64tos(du_int a) { return __floatundisf(a); }
diff --git a/compiler-rt/lib/builtins/riscv/restore.S b/compiler-rt/lib/builtins/riscv/restore.S
index 12f0d3365655..73f64a920d66 100644
--- a/compiler-rt/lib/builtins/riscv/restore.S
+++ b/compiler-rt/lib/builtins/riscv/restore.S
@@ -93,7 +93,7 @@ __riscv_restore_0:
 __riscv_restore_12:
   ld      s11, 8(sp)
   addi    sp, sp, 16
-  // fallthrough into __riscv_restore_11/10/9/8
+  // fallthrough into __riscv_restore_11/10
 
   .globl  __riscv_restore_11
   .type   __riscv_restore_11,@function
@@ -143,10 +143,6 @@ __riscv_restore_4:
   .type   __riscv_restore_3,@function
   .globl  __riscv_restore_2
   .type   __riscv_restore_2,@function
-  .globl  __riscv_restore_1
-  .type   __riscv_restore_1,@function
-  .globl  __riscv_restore_0
-  .type   __riscv_restore_0,@function
 __riscv_restore_3:
 __riscv_restore_2:
   ld      s2,  0(sp)
@@ -154,6 +150,10 @@ __riscv_restore_2:
   addi    sp, sp, 16
   // fallthrough into __riscv_restore_1/0
 
+  .globl  __riscv_restore_1
+  .type   __riscv_restore_1,@function
+  .globl  __riscv_restore_0
+  .type   __riscv_restore_0,@function
 __riscv_restore_1:
 __riscv_restore_0:
   ld      s0,  0(sp)
diff --git a/compiler-rt/lib/builtins/riscv/save.S b/compiler-rt/lib/builtins/riscv/save.S
index d811bf584fc3..85501aeb4c2e 100644
--- a/compiler-rt/lib/builtins/riscv/save.S
+++ b/compiler-rt/lib/builtins/riscv/save.S
@@ -174,6 +174,8 @@ __riscv_save_2:
   .type   __riscv_save_1,@function
   .globl  __riscv_save_0
   .type   __riscv_save_0,@function
+__riscv_save_1:
+__riscv_save_0:
   addi   sp, sp, -16
   sd     s0, 0(sp)
   sd     ra, 8(sp)
diff --git a/compiler-rt/lib/cfi/cfi.cpp b/compiler-rt/lib/cfi/cfi.cpp
index f691cfb94cfc..95853208f951 100644
--- a/compiler-rt/lib/cfi/cfi.cpp
+++ b/compiler-rt/lib/cfi/cfi.cpp
@@ -320,7 +320,7 @@ void InitShadow() {
 }
 
 THREADLOCAL int in_loader;
-BlockingMutex shadow_update_lock(LINKER_INITIALIZED);
+Mutex shadow_update_lock;
 
 void EnterLoader() NO_THREAD_SAFETY_ANALYSIS {
   if (in_loader == 0) {
@@ -359,7 +359,7 @@ ALWAYS_INLINE void CfiSlowPathCommon(u64 CallSiteTypeId, void *Ptr,
     return;
   }
   CFICheckFn cfi_check = sv.get_cfi_check();
-  VReport(2, "__cfi_check at %p\n", cfi_check);
+  VReport(2, "__cfi_check at %p\n", (void *)cfi_check);
   cfi_check(CallSiteTypeId, Ptr, DiagData);
 }
 
@@ -436,11 +436,11 @@ INTERCEPTOR(int, dlclose, void *handle) {
   return res;
 }
 
-static BlockingMutex interceptor_init_lock(LINKER_INITIALIZED);
+static Mutex interceptor_init_lock;
 static bool interceptors_inited = false;
 
 static void EnsureInterceptorsInitialized() {
-  BlockingMutexLock lock(&interceptor_init_lock);
+  Lock lock(&interceptor_init_lock);
   if (interceptors_inited)
     return;
 
diff --git a/compiler-rt/lib/dfsan/dfsan.cpp b/compiler-rt/lib/dfsan/dfsan.cpp
index 6f9ae141d7ab..ce2c04df83a8 100644
--- a/compiler-rt/lib/dfsan/dfsan.cpp
+++ b/compiler-rt/lib/dfsan/dfsan.cpp
@@ -369,37 +369,6 @@ static void SetOrigin(const void *dst, uptr size, u32 origin) {
       *(u32 *)(end - kOriginAlign) = origin;
 }
 
-static void WriteShadowInRange(dfsan_label label, uptr beg_shadow_addr,
-                               uptr end_shadow_addr) {
-  // TODO: After changing dfsan_label to 8bit, use internal_memset when label
-  // is not 0.
-  dfsan_label *labelp = (dfsan_label *)beg_shadow_addr;
-  if (label) {
-    for (; (uptr)labelp < end_shadow_addr; ++labelp) *labelp = label;
-    return;
-  }
-
-  for (; (uptr)labelp < end_shadow_addr; ++labelp) {
-    // Don't write the label if it is already the value we need it to be.
-    // In a program where most addresses are not labeled, it is common that
-    // a page of shadow memory is entirely zeroed.  The Linux copy-on-write
-    // implementation will share all of the zeroed pages, making a copy of a
-    // page when any value is written.  The un-sharing will happen even if
-    // the value written does not change the value in memory.  Avoiding the
-    // write when both |label| and |*labelp| are zero dramatically reduces
-    // the amount of real memory used by large programs.
-    if (!*labelp)
-      continue;
-
-    *labelp = 0;
-  }
-}
-
-static void WriteShadowWithSize(dfsan_label label, uptr shadow_addr,
-                                uptr size) {
-  WriteShadowInRange(label, shadow_addr, shadow_addr + size * sizeof(label));
-}
-
 #define RET_CHAIN_ORIGIN(id)           \
   GET_CALLER_PC_BP_SP;                 \
   (void)sp;                            \
@@ -451,21 +420,6 @@ void dfsan_copy_memory(void *dst, const void *src, uptr size) {
     dfsan_mem_origin_transfer(dst, src, size);
 }
 
-}  // namespace __dfsan
-
-// If the label s is tainted, set the size bytes from the address p to be a new
-// origin chain with the previous ID o and the current stack trace. This is
-// used by instrumentation to reduce code size when too much code is inserted.
-extern "C" SANITIZER_INTERFACE_ATTRIBUTE void __dfsan_maybe_store_origin(
-    dfsan_label s, void *p, uptr size, dfsan_origin o) {
-  if (UNLIKELY(s)) {
-    GET_CALLER_PC_BP_SP;
-    (void)sp;
-    GET_STORE_STACK_TRACE_PC_BP(pc, bp);
-    SetOrigin(p, size, ChainOrigin(o, &stack));
-  }
-}
-
 // Releases the pages within the origin address range.
 static void ReleaseOrigins(void *addr, uptr size) {
   const uptr beg_origin_addr = (uptr)__dfsan::origin_for(addr);
@@ -484,6 +438,19 @@ static void ReleaseOrigins(void *addr, uptr size) {
     Die();
 }
 
+static void WriteZeroShadowInRange(uptr beg, uptr end) {
+  // Don't write the label if it is already the value we need it to be.
+  // In a program where most addresses are not labeled, it is common that
+  // a page of shadow memory is entirely zeroed.  The Linux copy-on-write
+  // implementation will share all of the zeroed pages, making a copy of a
+  // page when any value is written.  The un-sharing will happen even if
+  // the value written does not change the value in memory.  Avoiding the
+  // write when both |label| and |*labelp| are zero dramatically reduces
+  // the amount of real memory used by large programs.
+  if (!mem_is_zero((const char *)beg, end - beg))
+    internal_memset((void *)beg, 0, end - beg);
+}
+
 // Releases the pages within the shadow address range, and sets
 // the shadow addresses not on the pages to be 0.
 static void ReleaseOrClearShadows(void *addr, uptr size) {
@@ -492,20 +459,22 @@ static void ReleaseOrClearShadows(void *addr, uptr size) {
   const uptr end_shadow_addr = (uptr)__dfsan::shadow_for(end_addr);
 
   if (end_shadow_addr - beg_shadow_addr <
-      common_flags()->clear_shadow_mmap_threshold)
-    return WriteShadowWithSize(0, beg_shadow_addr, size);
+      common_flags()->clear_shadow_mmap_threshold) {
+    WriteZeroShadowInRange(beg_shadow_addr, end_shadow_addr);
+    return;
+  }
 
   const uptr page_size = GetPageSizeCached();
   const uptr beg_aligned = RoundUpTo(beg_shadow_addr, page_size);
   const uptr end_aligned = RoundDownTo(end_shadow_addr, page_size);
 
   if (beg_aligned >= end_aligned) {
-    WriteShadowWithSize(0, beg_shadow_addr, size);
+    WriteZeroShadowInRange(beg_shadow_addr, end_shadow_addr);
   } else {
     if (beg_aligned != beg_shadow_addr)
-      WriteShadowInRange(0, beg_shadow_addr, beg_aligned);
+      WriteZeroShadowInRange(beg_shadow_addr, beg_aligned);
     if (end_aligned != end_shadow_addr)
-      WriteShadowInRange(0, end_aligned, end_shadow_addr);
+      WriteZeroShadowInRange(end_aligned, end_shadow_addr);
     if (!MmapFixedSuperNoReserve(beg_aligned, end_aligned - beg_aligned))
       Die();
   }
@@ -514,7 +483,7 @@ static void ReleaseOrClearShadows(void *addr, uptr size) {
 void SetShadow(dfsan_label label, void *addr, uptr size, dfsan_origin origin) {
   if (0 != label) {
     const uptr beg_shadow_addr = (uptr)__dfsan::shadow_for(addr);
-    WriteShadowWithSize(label, beg_shadow_addr, size);
+    internal_memset((void *)beg_shadow_addr, label, size);
     if (dfsan_get_track_origins())
       SetOrigin(addr, size, origin);
     return;
@@ -526,9 +495,24 @@ void SetShadow(dfsan_label label, void *addr, uptr size, dfsan_origin origin) {
   ReleaseOrClearShadows(addr, size);
 }
 
+}  // namespace __dfsan
+
+// If the label s is tainted, set the size bytes from the address p to be a new
+// origin chain with the previous ID o and the current stack trace. This is
+// used by instrumentation to reduce code size when too much code is inserted.
+extern "C" SANITIZER_INTERFACE_ATTRIBUTE void __dfsan_maybe_store_origin(
+    dfsan_label s, void *p, uptr size, dfsan_origin o) {
+  if (UNLIKELY(s)) {
+    GET_CALLER_PC_BP_SP;
+    (void)sp;
+    GET_STORE_STACK_TRACE_PC_BP(pc, bp);
+    SetOrigin(p, size, ChainOrigin(o, &stack));
+  }
+}
+
 extern "C" SANITIZER_INTERFACE_ATTRIBUTE void __dfsan_set_label(
     dfsan_label label, dfsan_origin origin, void *addr, uptr size) {
-  SetShadow(label, addr, size, origin);
+  __dfsan::SetShadow(label, addr, size, origin);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
@@ -539,7 +523,7 @@ void dfsan_set_label(dfsan_label label, void *addr, uptr size) {
     GET_STORE_STACK_TRACE_PC_BP(pc, bp);
     init_origin = ChainOrigin(0, &stack, true);
   }
-  SetShadow(label, addr, size, init_origin);
+  __dfsan::SetShadow(label, addr, size, init_origin);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
@@ -709,9 +693,9 @@ extern "C" SANITIZER_INTERFACE_ATTRIBUTE void dfsan_print_origin_trace(
     PrintInvalidOriginWarning(label, addr);
 }
 
-extern "C" SANITIZER_INTERFACE_ATTRIBUTE size_t
+extern "C" SANITIZER_INTERFACE_ATTRIBUTE uptr
 dfsan_sprint_origin_trace(const void *addr, const char *description,
-                          char *out_buf, size_t out_buf_size) {
+                          char *out_buf, uptr out_buf_size) {
   CHECK(out_buf);
 
   if (!dfsan_get_track_origins()) {
@@ -780,8 +764,8 @@ extern "C" SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_print_stack_trace() {
   stack.Print();
 }
 
-extern "C" SANITIZER_INTERFACE_ATTRIBUTE size_t
-dfsan_sprint_stack_trace(char *out_buf, size_t out_buf_size) {
+extern "C" SANITIZER_INTERFACE_ATTRIBUTE uptr
+dfsan_sprint_stack_trace(char *out_buf, uptr out_buf_size) {
   CHECK(out_buf);
   GET_CALLER_PC_BP;
   GET_STORE_STACK_TRACE_PC_BP(pc, bp);
@@ -932,7 +916,7 @@ static bool ProtectMemoryRange(uptr beg, uptr size, const char *name) {
 // Consider refactoring these into a shared implementation.
 bool InitShadow(bool init_origins) {
   // Let user know mapping parameters first.
-  VPrintf(1, "dfsan_init %p\n", &__dfsan::dfsan_init);
+  VPrintf(1, "dfsan_init %p\n", (void *)&__dfsan::dfsan_init);
   for (unsigned i = 0; i < kMemoryLayoutSize; ++i)
     VPrintf(1, "%s: %zx - %zx\n", kMemoryLayout[i].name, kMemoryLayout[i].start,
             kMemoryLayout[i].end - 1);
@@ -1007,7 +991,7 @@ static void DFsanInit(int argc, char **argv, char **envp) {
 
   DFsanThread *main_thread = DFsanThread::Create(nullptr, nullptr, nullptr);
   SetCurrentThread(main_thread);
-  main_thread->ThreadStart();
+  main_thread->Init();
 
   dfsan_init_is_running = false;
   dfsan_inited = true;
diff --git a/compiler-rt/lib/dfsan/dfsan.h b/compiler-rt/lib/dfsan/dfsan.h
index b212298157eb..b529e008f300 100644
--- a/compiler-rt/lib/dfsan/dfsan.h
+++ b/compiler-rt/lib/dfsan/dfsan.h
@@ -49,7 +49,7 @@ void dfsan_mem_origin_transfer(const void *dst, const void *src, uptr len);
 }  // extern "C"
 
 template <typename T>
-void dfsan_set_label(dfsan_label label, T &data) {  // NOLINT
+void dfsan_set_label(dfsan_label label, T &data) {
   dfsan_set_label(label, (void *)&data, sizeof(T));
 }
 
diff --git a/compiler-rt/lib/dfsan/dfsan_custom.cpp b/compiler-rt/lib/dfsan/dfsan_custom.cpp
index 3185184f29c8..217bd35c1c54 100644
--- a/compiler-rt/lib/dfsan/dfsan_custom.cpp
+++ b/compiler-rt/lib/dfsan/dfsan_custom.cpp
@@ -583,7 +583,7 @@ SANITIZER_INTERFACE_ATTRIBUTE char *__dfsw_strcat(char *dest, const char *src,
                                                   dfsan_label src_label,
                                                   dfsan_label *ret_label) {
   size_t dest_len = strlen(dest);
-  char *ret = strcat(dest, src);  // NOLINT
+  char *ret = strcat(dest, src);
   dfsan_label *sdest = shadow_for(dest + dest_len);
   const dfsan_label *ssrc = shadow_for(src);
   internal_memcpy((void *)sdest, (const void *)ssrc,
@@ -597,7 +597,7 @@ SANITIZER_INTERFACE_ATTRIBUTE char *__dfso_strcat(
     dfsan_label *ret_label, dfsan_origin dest_origin, dfsan_origin src_origin,
     dfsan_origin *ret_origin) {
   size_t dest_len = strlen(dest);
-  char *ret = strcat(dest, src);  // NOLINT
+  char *ret = strcat(dest, src);
   dfsan_label *sdest = shadow_for(dest + dest_len);
   const dfsan_label *ssrc = shadow_for(src);
   size_t src_len = strlen(src);
@@ -755,6 +755,8 @@ SANITIZER_INTERFACE_ATTRIBUTE void *__dfso_dlopen(
 static void *DFsanThreadStartFunc(void *arg) {
   DFsanThread *t = (DFsanThread *)arg;
   SetCurrentThread(t);
+  t->Init();
+  SetSigProcMask(&t->starting_sigset_, nullptr);
   return t->ThreadStart();
 }
 
@@ -775,6 +777,7 @@ static int dfsan_pthread_create(pthread_t *thread, const pthread_attr_t *attr,
   DFsanThread *t =
       DFsanThread::Create(start_routine_trampoline,
                           (thread_callback_t)start_routine, arg, track_origins);
+  ScopedBlockSignals block(&t->starting_sigset_);
   int res = pthread_create(thread, attr, DFsanThreadStartFunc, t);
 
   if (attr == &myattr)
@@ -1026,6 +1029,33 @@ char *__dfso_get_current_dir_name(dfsan_label *ret_label,
   return __dfsw_get_current_dir_name(ret_label);
 }
 
+// This function is only available for glibc 2.25 or newer.  Mark it weak so
+// linking succeeds with older glibcs.
+SANITIZER_WEAK_ATTRIBUTE int getentropy(void *buffer, size_t length);
+
+SANITIZER_INTERFACE_ATTRIBUTE int __dfsw_getentropy(void *buffer, size_t length,
+                                                    dfsan_label buffer_label,
+                                                    dfsan_label length_label,
+                                                    dfsan_label *ret_label) {
+  int ret = getentropy(buffer, length);
+  if (ret == 0) {
+    dfsan_set_label(0, buffer, length);
+  }
+  *ret_label = 0;
+  return ret;
+}
+
+SANITIZER_INTERFACE_ATTRIBUTE int __dfso_getentropy(void *buffer, size_t length,
+                                                    dfsan_label buffer_label,
+                                                    dfsan_label length_label,
+                                                    dfsan_label *ret_label,
+                                                    dfsan_origin buffer_origin,
+                                                    dfsan_origin length_origin,
+                                                    dfsan_origin *ret_origin) {
+  return __dfsw_getentropy(buffer, length, buffer_label, length_label,
+                           ret_label);
+}
+
 SANITIZER_INTERFACE_ATTRIBUTE
 int __dfsw_gethostname(char *name, size_t len, dfsan_label name_label,
                        dfsan_label len_label, dfsan_label *ret_label) {
@@ -1088,7 +1118,7 @@ int __dfso_getrusage(int who, struct rusage *usage, dfsan_label who_label,
 SANITIZER_INTERFACE_ATTRIBUTE
 char *__dfsw_strcpy(char *dest, const char *src, dfsan_label dst_label,
                     dfsan_label src_label, dfsan_label *ret_label) {
-  char *ret = strcpy(dest, src);  // NOLINT
+  char *ret = strcpy(dest, src);
   if (ret) {
     internal_memcpy(shadow_for(dest), shadow_for(src),
                     sizeof(dfsan_label) * (strlen(src) + 1));
@@ -1102,7 +1132,7 @@ char *__dfso_strcpy(char *dest, const char *src, dfsan_label dst_label,
                     dfsan_label src_label, dfsan_label *ret_label,
                     dfsan_origin dst_origin, dfsan_origin src_origin,
                     dfsan_origin *ret_origin) {
-  char *ret = strcpy(dest, src);  // NOLINT
+  char *ret = strcpy(dest, src);
   if (ret) {
     size_t str_len = strlen(src) + 1;
     dfsan_mem_origin_transfer(dest, src, str_len);
@@ -2489,7 +2519,8 @@ pid_t __dfso_fork(dfsan_label *ret_label, dfsan_origin *ret_origin) {
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_pc_guard, u32 *) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_pc_guard_init, u32 *,
                              u32 *) {}
-SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_pcs_init, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_pcs_init, const uptr *beg,
+                             const uptr *end) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_pc_indir, void) {}
 
 SANITIZER_INTERFACE_WEAK_DEF(void, __dfsw___sanitizer_cov_trace_cmp, void) {}
diff --git a/compiler-rt/lib/dfsan/dfsan_interceptors.cpp b/compiler-rt/lib/dfsan/dfsan_interceptors.cpp
index 92be4fc87d49..d8fb9ea86618 100644
--- a/compiler-rt/lib/dfsan/dfsan_interceptors.cpp
+++ b/compiler-rt/lib/dfsan/dfsan_interceptors.cpp
@@ -17,6 +17,7 @@
 #include "dfsan/dfsan.h"
 #include "dfsan/dfsan_thread.h"
 #include "interception/interception.h"
+#include "sanitizer_common/sanitizer_allocator_dlsym.h"
 #include "sanitizer_common/sanitizer_allocator_interface.h"
 #include "sanitizer_common/sanitizer_common.h"
 #include "sanitizer_common/sanitizer_errno.h"
@@ -26,11 +27,11 @@
 
 using namespace __sanitizer;
 
-namespace {
+static bool interceptors_initialized;
 
-bool interceptors_initialized;
-
-}  // namespace
+struct DlsymAlloc : public DlSymAllocator<DlsymAlloc> {
+  static bool UseImpl() { return !__dfsan::dfsan_inited; }
+};
 
 INTERCEPTOR(void *, reallocarray, void *ptr, SIZE_T nmemb, SIZE_T size) {
   return __dfsan::dfsan_reallocarray(ptr, nmemb, size);
@@ -47,63 +48,37 @@ INTERCEPTOR(void *, aligned_alloc, SIZE_T alignment, SIZE_T size) {
   return __dfsan::dfsan_aligned_alloc(alignment, size);
 }
 
-static uptr allocated_for_dlsym;
-static const uptr kDlsymAllocPoolSize = 1024;
-static uptr alloc_memory_for_dlsym[kDlsymAllocPoolSize];
-
-static bool IsInDlsymAllocPool(const void *ptr) {
-  uptr off = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-  return off < sizeof(alloc_memory_for_dlsym);
-}
-
-static void *AllocateFromLocalPool(uptr size_in_bytes) {
-  uptr size_in_words = RoundUpTo(size_in_bytes, kWordSize) / kWordSize;
-  void *mem = (void *)&alloc_memory_for_dlsym[allocated_for_dlsym];
-  allocated_for_dlsym += size_in_words;
-  CHECK_LT(allocated_for_dlsym, kDlsymAllocPoolSize);
-  return mem;
-}
-
 INTERCEPTOR(void *, calloc, SIZE_T nmemb, SIZE_T size) {
-  if (UNLIKELY(!__dfsan::dfsan_inited))
-    // Hack: dlsym calls calloc before REAL(calloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(nmemb * size);
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Callocate(nmemb, size);
   return __dfsan::dfsan_calloc(nmemb, size);
 }
 
 INTERCEPTOR(void *, realloc, void *ptr, SIZE_T size) {
-  if (UNLIKELY(IsInDlsymAllocPool(ptr))) {
-    uptr offset = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-    uptr copy_size = Min(size, kDlsymAllocPoolSize - offset);
-    void *new_ptr;
-    if (UNLIKELY(!__dfsan::dfsan_inited)) {
-      new_ptr = AllocateFromLocalPool(copy_size);
-    } else {
-      copy_size = size;
-      new_ptr = __dfsan::dfsan_malloc(copy_size);
-    }
-    internal_memcpy(new_ptr, ptr, copy_size);
-    return new_ptr;
-  }
+  if (DlsymAlloc::Use() || DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Realloc(ptr, size);
   return __dfsan::dfsan_realloc(ptr, size);
 }
 
 INTERCEPTOR(void *, malloc, SIZE_T size) {
-  if (UNLIKELY(!__dfsan::dfsan_inited))
-    // Hack: dlsym calls malloc before REAL(malloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(size);
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Allocate(size);
   return __dfsan::dfsan_malloc(size);
 }
 
 INTERCEPTOR(void, free, void *ptr) {
-  if (!ptr || UNLIKELY(IsInDlsymAllocPool(ptr)))
+  if (!ptr)
     return;
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
   return __dfsan::dfsan_deallocate(ptr);
 }
 
 INTERCEPTOR(void, cfree, void *ptr) {
-  if (!ptr || UNLIKELY(IsInDlsymAllocPool(ptr)))
+  if (!ptr)
     return;
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
   return __dfsan::dfsan_deallocate(ptr);
 }
 
@@ -152,12 +127,12 @@ INTERCEPTOR(uptr, malloc_usable_size, void *ptr) {
   if (__dfsan::dfsan_init_is_running)       \
     return REAL(func)(__VA_ARGS__);         \
   ENSURE_DFSAN_INITED();                    \
-  dfsan_set_label(0, __errno_location(), sizeof(int)); /* NOLINT */
+  dfsan_set_label(0, __errno_location(), sizeof(int));
 
 INTERCEPTOR(void *, mmap, void *addr, SIZE_T length, int prot, int flags,
             int fd, OFF_T offset) {
   if (common_flags()->detect_write_exec)
-    ReportMmapWriteExec(prot);
+    ReportMmapWriteExec(prot, flags);
   if (!__dfsan::dfsan_inited)
     return (void *)internal_mmap(addr, length, prot, flags, fd, offset);
   COMMON_INTERCEPTOR_ENTER(mmap, addr, length, prot, flags, fd, offset);
@@ -171,7 +146,7 @@ INTERCEPTOR(void *, mmap, void *addr, SIZE_T length, int prot, int flags,
 INTERCEPTOR(void *, mmap64, void *addr, SIZE_T length, int prot, int flags,
             int fd, OFF64_T offset) {
   if (common_flags()->detect_write_exec)
-    ReportMmapWriteExec(prot);
+    ReportMmapWriteExec(prot, flags);
   if (!__dfsan::dfsan_inited)
     return (void *)internal_mmap(addr, length, prot, flags, fd, offset);
   COMMON_INTERCEPTOR_ENTER(mmap64, addr, length, prot, flags, fd, offset);
diff --git a/compiler-rt/lib/dfsan/dfsan_thread.cpp b/compiler-rt/lib/dfsan/dfsan_thread.cpp
index 6869cf231587..df7e4d9b7421 100644
--- a/compiler-rt/lib/dfsan/dfsan_thread.cpp
+++ b/compiler-rt/lib/dfsan/dfsan_thread.cpp
@@ -67,8 +67,6 @@ void DFsanThread::Destroy() {
 }
 
 thread_return_t DFsanThread::ThreadStart() {
-  Init();
-
   if (!start_routine_) {
     // start_routine_ == 0 if we're on the main thread or on one of the
     // OS X libdispatch worker threads. But nobody is supposed to call
diff --git a/compiler-rt/lib/dfsan/dfsan_thread.h b/compiler-rt/lib/dfsan/dfsan_thread.h
index 8dde626f5569..1c33a1854997 100644
--- a/compiler-rt/lib/dfsan/dfsan_thread.h
+++ b/compiler-rt/lib/dfsan/dfsan_thread.h
@@ -1,5 +1,4 @@
-//===-- dfsan_thread.h -------------------------------------------*- C++
-//-*-===//
+//===-- dfsan_thread.h ------------------------------------------*- C++ -*-===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
@@ -16,6 +15,7 @@
 
 #include "dfsan_allocator.h"
 #include "sanitizer_common/sanitizer_common.h"
+#include "sanitizer_common/sanitizer_posix.h"
 
 namespace __dfsan {
 
@@ -46,6 +46,7 @@ class DFsanThread {
   DFsanThreadLocalMallocStorage &malloc_storage() { return malloc_storage_; }
 
   int destructor_iterations_;
+  __sanitizer_sigset_t starting_sigset_;
 
  private:
   void SetThreadStackAndTls();
diff --git a/compiler-rt/lib/dfsan/done_abilist.txt b/compiler-rt/lib/dfsan/done_abilist.txt
index 3c2670e04c29..eef7c48948cc 100644
--- a/compiler-rt/lib/dfsan/done_abilist.txt
+++ b/compiler-rt/lib/dfsan/done_abilist.txt
@@ -218,6 +218,7 @@ fun:fgets=custom
 fun:fstat=custom
 fun:getcwd=custom
 fun:get_current_dir_name=custom
+fun:getentropy=custom
 fun:gethostname=custom
 fun:getpeername=custom
 fun:getrlimit=custom
@@ -268,7 +269,7 @@ fun:strrchr=custom
 fun:strstr=custom
 
 # Functions which take action based on global state, such as running a callback
-# set by a sepperate function.
+# set by a separate function.
 fun:write=custom
 
 # Functions that take a callback (wrap the callback manually).
diff --git a/compiler-rt/lib/dfsan/libc_ubuntu1404_abilist.txt b/compiler-rt/lib/dfsan/libc_ubuntu1404_abilist.txt
index a1ea0a06b537..433092e2b27b 100644
--- a/compiler-rt/lib/dfsan/libc_ubuntu1404_abilist.txt
+++ b/compiler-rt/lib/dfsan/libc_ubuntu1404_abilist.txt
@@ -1852,6 +1852,7 @@ fun:getdirentries64=uninstrumented
 fun:getdomainname=uninstrumented
 fun:getdtablesize=uninstrumented
 fun:getegid=uninstrumented
+fun:getentropy=uninstrumented
 fun:getenv=uninstrumented
 fun:geteuid=uninstrumented
 fun:getfsent=uninstrumented
diff --git a/compiler-rt/lib/fuzzer/FuzzerBuiltinsMsvc.h b/compiler-rt/lib/fuzzer/FuzzerBuiltinsMsvc.h
index ab191b60ef6e..421dee7f6603 100644
--- a/compiler-rt/lib/fuzzer/FuzzerBuiltinsMsvc.h
+++ b/compiler-rt/lib/fuzzer/FuzzerBuiltinsMsvc.h
@@ -41,7 +41,8 @@ inline uint32_t Clzll(uint64_t X) {
 #if !defined(_M_ARM) && !defined(_M_X64)
   // Scan the high 32 bits.
   if (_BitScanReverse(&LeadZeroIdx, static_cast<unsigned long>(X >> 32)))
-    return static_cast<int>(63 - (LeadZeroIdx + 32)); // Create a bit offset from the MSB.
+    return static_cast<int>(
+        63 - (LeadZeroIdx + 32)); // Create a bit offset from the MSB.
   // Scan the low 32 bits.
   if (_BitScanReverse(&LeadZeroIdx, static_cast<unsigned long>(X)))
     return static_cast<int>(63 - LeadZeroIdx);
diff --git a/compiler-rt/lib/fuzzer/FuzzerCommand.h b/compiler-rt/lib/fuzzer/FuzzerCommand.h
index 87308864af53..f653fe358768 100644
--- a/compiler-rt/lib/fuzzer/FuzzerCommand.h
+++ b/compiler-rt/lib/fuzzer/FuzzerCommand.h
@@ -33,7 +33,7 @@ public:
 
   Command() : CombinedOutAndErr(false) {}
 
-  explicit Command(const Vector<std::string> &ArgsToAdd)
+  explicit Command(const std::vector<std::string> &ArgsToAdd)
       : Args(ArgsToAdd), CombinedOutAndErr(false) {}
 
   explicit Command(const Command &Other)
@@ -58,7 +58,7 @@ public:
 
   // Gets all of the current command line arguments, **including** those after
   // "-ignore-remaining-args=1".
-  const Vector<std::string> &getArguments() const { return Args; }
+  const std::vector<std::string> &getArguments() const { return Args; }
 
   // Adds the given argument before "-ignore_remaining_args=1", or at the end
   // if that flag isn't present.
@@ -68,7 +68,7 @@ public:
 
   // Adds all given arguments before "-ignore_remaining_args=1", or at the end
   // if that flag isn't present.
-  void addArguments(const Vector<std::string> &ArgsToAdd) {
+  void addArguments(const std::vector<std::string> &ArgsToAdd) {
     Args.insert(endMutableArgs(), ArgsToAdd.begin(), ArgsToAdd.end());
   }
 
@@ -155,16 +155,16 @@ private:
   Command(Command &&Other) = delete;
   Command &operator=(Command &&Other) = delete;
 
-  Vector<std::string>::iterator endMutableArgs() {
+  std::vector<std::string>::iterator endMutableArgs() {
     return std::find(Args.begin(), Args.end(), ignoreRemainingArgs());
   }
 
-  Vector<std::string>::const_iterator endMutableArgs() const {
+  std::vector<std::string>::const_iterator endMutableArgs() const {
     return std::find(Args.begin(), Args.end(), ignoreRemainingArgs());
   }
 
   // The command arguments.  Args[0] is the command name.
-  Vector<std::string> Args;
+  std::vector<std::string> Args;
 
   // True indicates stderr is redirected to stdout.
   bool CombinedOutAndErr;
diff --git a/compiler-rt/lib/fuzzer/FuzzerCorpus.h b/compiler-rt/lib/fuzzer/FuzzerCorpus.h
index f8c126072c96..e01891e18fe3 100644
--- a/compiler-rt/lib/fuzzer/FuzzerCorpus.h
+++ b/compiler-rt/lib/fuzzer/FuzzerCorpus.h
@@ -39,13 +39,13 @@ struct InputInfo {
   bool MayDeleteFile = false;
   bool Reduced = false;
   bool HasFocusFunction = false;
-  Vector<uint32_t> UniqFeatureSet;
-  Vector<uint8_t> DataFlowTraceForFocusFunction;
+  std::vector<uint32_t> UniqFeatureSet;
+  std::vector<uint8_t> DataFlowTraceForFocusFunction;
   // Power schedule.
   bool NeedsEnergyUpdate = false;
   double Energy = 0.0;
   double SumIncidence = 0.0;
-  Vector<std::pair<uint32_t, uint16_t>> FeatureFreqs;
+  std::vector<std::pair<uint32_t, uint16_t>> FeatureFreqs;
 
   // Delete feature Idx and its frequency from FeatureFreqs.
   bool DeleteFeatureFreq(uint32_t Idx) {
@@ -209,7 +209,7 @@ public:
   InputInfo *AddToCorpus(const Unit &U, size_t NumFeatures, bool MayDeleteFile,
                          bool HasFocusFunction, bool NeverReduce,
                          std::chrono::microseconds TimeOfUnit,
-                         const Vector<uint32_t> &FeatureSet,
+                         const std::vector<uint32_t> &FeatureSet,
                          const DataFlowTrace &DFT, const InputInfo *BaseII) {
     assert(!U.empty());
     if (FeatureDebug)
@@ -258,7 +258,7 @@ public:
   }
 
   // Debug-only
-  void PrintFeatureSet(const Vector<uint32_t> &FeatureSet) {
+  void PrintFeatureSet(const std::vector<uint32_t> &FeatureSet) {
     if (!FeatureDebug) return;
     Printf("{");
     for (uint32_t Feature: FeatureSet)
@@ -284,7 +284,8 @@ public:
     }
   }
 
-  void Replace(InputInfo *II, const Unit &U) {
+  void Replace(InputInfo *II, const Unit &U,
+               std::chrono::microseconds TimeOfUnit) {
     assert(II->U.size() > U.size());
     Hashes.erase(Sha1ToString(II->Sha1));
     DeleteFile(*II);
@@ -292,6 +293,7 @@ public:
     Hashes.insert(Sha1ToString(II->Sha1));
     II->U = U;
     II->Reduced = true;
+    II->TimeOfUnit = TimeOfUnit;
     DistributionNeedsUpdate = true;
   }
 
@@ -325,7 +327,8 @@ public:
       const auto &II = *Inputs[i];
       Printf("  [% 3zd %s] sz: % 5zd runs: % 5zd succ: % 5zd focus: %d\n", i,
              Sha1ToString(II.Sha1).c_str(), II.U.size(),
-             II.NumExecutedMutations, II.NumSuccessfullMutations, II.HasFocusFunction);
+             II.NumExecutedMutations, II.NumSuccessfullMutations,
+             II.HasFocusFunction);
     }
   }
 
@@ -563,11 +566,11 @@ private:
   }
   std::piecewise_constant_distribution<double> CorpusDistribution;
 
-  Vector<double> Intervals;
-  Vector<double> Weights;
+  std::vector<double> Intervals;
+  std::vector<double> Weights;
 
   std::unordered_set<std::string> Hashes;
-  Vector<InputInfo*> Inputs;
+  std::vector<InputInfo *> Inputs;
 
   size_t NumAddedFeatures = 0;
   size_t NumUpdatedFeatures = 0;
@@ -577,7 +580,7 @@ private:
   bool DistributionNeedsUpdate = true;
   uint16_t FreqOfMostAbundantRareFeature = 0;
   uint16_t GlobalFeatureFreqs[kFeatureSetSize] = {};
-  Vector<uint32_t> RareFeatures;
+  std::vector<uint32_t> RareFeatures;
 
   std::string OutputCorpus;
 };
diff --git a/compiler-rt/lib/fuzzer/FuzzerDataFlowTrace.cpp b/compiler-rt/lib/fuzzer/FuzzerDataFlowTrace.cpp
index 23d422590d19..2f9a4d2d7adc 100644
--- a/compiler-rt/lib/fuzzer/FuzzerDataFlowTrace.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerDataFlowTrace.cpp
@@ -37,7 +37,7 @@ bool BlockCoverage::AppendCoverage(const std::string &S) {
 // Coverage lines have this form:
 // CN X Y Z T
 // where N is the number of the function, T is the total number of instrumented
-// BBs, and X,Y,Z, if present, are the indecies of covered BB.
+// BBs, and X,Y,Z, if present, are the indices of covered BB.
 // BB #0, which is the entry block, is not explicitly listed.
 bool BlockCoverage::AppendCoverage(std::istream &IN) {
   std::string L;
@@ -52,7 +52,7 @@ bool BlockCoverage::AppendCoverage(std::istream &IN) {
       continue;
     }
     if (L[0] != 'C') continue;
-    Vector<uint32_t> CoveredBlocks;
+    std::vector<uint32_t> CoveredBlocks;
     while (true) {
       uint32_t BB = 0;
       SS >> BB;
@@ -68,7 +68,7 @@ bool BlockCoverage::AppendCoverage(std::istream &IN) {
     auto It = Functions.find(FunctionId);
     auto &Counters =
         It == Functions.end()
-            ? Functions.insert({FunctionId, Vector<uint32_t>(NumBlocks)})
+            ? Functions.insert({FunctionId, std::vector<uint32_t>(NumBlocks)})
                   .first->second
             : It->second;
 
@@ -86,8 +86,8 @@ bool BlockCoverage::AppendCoverage(std::istream &IN) {
 //   * any uncovered function gets weight 0.
 //   * a function with lots of uncovered blocks gets bigger weight.
 //   * a function with a less frequently executed code gets bigger weight.
-Vector<double> BlockCoverage::FunctionWeights(size_t NumFunctions) const {
-  Vector<double> Res(NumFunctions);
+std::vector<double> BlockCoverage::FunctionWeights(size_t NumFunctions) const {
+  std::vector<double> Res(NumFunctions);
   for (auto It : Functions) {
     auto FunctionID = It.first;
     auto Counters = It.second;
@@ -104,7 +104,7 @@ Vector<double> BlockCoverage::FunctionWeights(size_t NumFunctions) const {
 }
 
 void DataFlowTrace::ReadCoverage(const std::string &DirPath) {
-  Vector<SizedFile> Files;
+  std::vector<SizedFile> Files;
   GetSizedFilesFromDir(DirPath, &Files);
   for (auto &SF : Files) {
     auto Name = Basename(SF.File);
@@ -115,16 +115,16 @@ void DataFlowTrace::ReadCoverage(const std::string &DirPath) {
   }
 }
 
-static void DFTStringAppendToVector(Vector<uint8_t> *DFT,
+static void DFTStringAppendToVector(std::vector<uint8_t> *DFT,
                                     const std::string &DFTString) {
   assert(DFT->size() == DFTString.size());
   for (size_t I = 0, Len = DFT->size(); I < Len; I++)
     (*DFT)[I] = DFTString[I] == '1';
 }
 
-// converts a string of '0' and '1' into a Vector<uint8_t>
-static Vector<uint8_t> DFTStringToVector(const std::string &DFTString) {
-  Vector<uint8_t> DFT(DFTString.size());
+// converts a string of '0' and '1' into a std::vector<uint8_t>
+static std::vector<uint8_t> DFTStringToVector(const std::string &DFTString) {
+  std::vector<uint8_t> DFT(DFTString.size());
   DFTStringAppendToVector(&DFT, DFTString);
   return DFT;
 }
@@ -159,14 +159,14 @@ static bool ParseDFTLine(const std::string &Line, size_t *FunctionNum,
 }
 
 bool DataFlowTrace::Init(const std::string &DirPath, std::string *FocusFunction,
-                         Vector<SizedFile> &CorporaFiles, Random &Rand) {
+                         std::vector<SizedFile> &CorporaFiles, Random &Rand) {
   if (DirPath.empty()) return false;
   Printf("INFO: DataFlowTrace: reading from '%s'\n", DirPath.c_str());
-  Vector<SizedFile> Files;
+  std::vector<SizedFile> Files;
   GetSizedFilesFromDir(DirPath, &Files);
   std::string L;
   size_t FocusFuncIdx = SIZE_MAX;
-  Vector<std::string> FunctionNames;
+  std::vector<std::string> FunctionNames;
 
   // Collect the hashes of the corpus files.
   for (auto &SF : CorporaFiles)
@@ -191,7 +191,7 @@ bool DataFlowTrace::Init(const std::string &DirPath, std::string *FocusFunction,
     // * chooses a random function according to the weights.
     ReadCoverage(DirPath);
     auto Weights = Coverage.FunctionWeights(NumFunctions);
-    Vector<double> Intervals(NumFunctions + 1);
+    std::vector<double> Intervals(NumFunctions + 1);
     std::iota(Intervals.begin(), Intervals.end(), 0);
     auto Distribution = std::piecewise_constant_distribution<double>(
         Intervals.begin(), Intervals.end(), Weights.begin());
@@ -247,7 +247,7 @@ bool DataFlowTrace::Init(const std::string &DirPath, std::string *FocusFunction,
 }
 
 int CollectDataFlow(const std::string &DFTBinary, const std::string &DirPath,
-                    const Vector<SizedFile> &CorporaFiles) {
+                    const std::vector<SizedFile> &CorporaFiles) {
   Printf("INFO: collecting data flow: bin: %s dir: %s files: %zd\n",
          DFTBinary.c_str(), DirPath.c_str(), CorporaFiles.size());
   if (CorporaFiles.empty()) {
@@ -265,7 +265,7 @@ int CollectDataFlow(const std::string &DFTBinary, const std::string &DirPath,
     // we then request tags in [0,Size/2) and [Size/2, Size), and so on.
     // Function number => DFT.
     auto OutPath = DirPlusFile(DirPath, Hash(FileToVector(F.File)));
-    std::unordered_map<size_t, Vector<uint8_t>> DFTMap;
+    std::unordered_map<size_t, std::vector<uint8_t>> DFTMap;
     std::unordered_set<std::string> Cov;
     Command Cmd;
     Cmd.addArgument(DFTBinary);
diff --git a/compiler-rt/lib/fuzzer/FuzzerDataFlowTrace.h b/compiler-rt/lib/fuzzer/FuzzerDataFlowTrace.h
index 07c03bb25651..054dce1bdcb6 100644
--- a/compiler-rt/lib/fuzzer/FuzzerDataFlowTrace.h
+++ b/compiler-rt/lib/fuzzer/FuzzerDataFlowTrace.h
@@ -39,7 +39,7 @@
 namespace fuzzer {
 
 int CollectDataFlow(const std::string &DFTBinary, const std::string &DirPath,
-                    const Vector<SizedFile> &CorporaFiles);
+                    const std::vector<SizedFile> &CorporaFiles);
 
 class BlockCoverage {
 public:
@@ -77,11 +77,11 @@ public:
     return Result;
   }
 
-  Vector<double> FunctionWeights(size_t NumFunctions) const;
+  std::vector<double> FunctionWeights(size_t NumFunctions) const;
   void clear() { Functions.clear(); }
 
 private:
-  typedef Vector<uint32_t> CoverageVector;
+  typedef std::vector<uint32_t> CoverageVector;
 
   uint32_t NumberOfCoveredBlocks(const CoverageVector &Counters) const {
     uint32_t Res = 0;
@@ -117,9 +117,9 @@ class DataFlowTrace {
  public:
   void ReadCoverage(const std::string &DirPath);
   bool Init(const std::string &DirPath, std::string *FocusFunction,
-            Vector<SizedFile> &CorporaFiles, Random &Rand);
+            std::vector<SizedFile> &CorporaFiles, Random &Rand);
   void Clear() { Traces.clear(); }
-  const Vector<uint8_t> *Get(const std::string &InputSha1) const {
+  const std::vector<uint8_t> *Get(const std::string &InputSha1) const {
     auto It = Traces.find(InputSha1);
     if (It != Traces.end())
       return &It->second;
@@ -128,9 +128,9 @@ class DataFlowTrace {
 
  private:
   // Input's sha1 => DFT for the FocusFunction.
-  std::unordered_map<std::string, Vector<uint8_t> > Traces;
-  BlockCoverage Coverage;
-  std::unordered_set<std::string> CorporaHashes;
+   std::unordered_map<std::string, std::vector<uint8_t>> Traces;
+   BlockCoverage Coverage;
+   std::unordered_set<std::string> CorporaHashes;
 };
 }  // namespace fuzzer
 
diff --git a/compiler-rt/lib/fuzzer/FuzzerDefs.h b/compiler-rt/lib/fuzzer/FuzzerDefs.h
index 1a2752af2f4d..db1f74a545e3 100644
--- a/compiler-rt/lib/fuzzer/FuzzerDefs.h
+++ b/compiler-rt/lib/fuzzer/FuzzerDefs.h
@@ -38,28 +38,8 @@ struct ExternalFunctions;
 // Global interface to functions that may or may not be available.
 extern ExternalFunctions *EF;
 
-// We are using a custom allocator to give a different symbol name to STL
-// containers in order to avoid ODR violations.
-template<typename T>
-  class fuzzer_allocator: public std::allocator<T> {
-    public:
-      fuzzer_allocator() = default;
-
-      template<class U>
-      fuzzer_allocator(const fuzzer_allocator<U>&) {}
-
-      template<class Other>
-      struct rebind { typedef fuzzer_allocator<Other> other;  };
-  };
-
-template<typename T>
-using Vector = std::vector<T, fuzzer_allocator<T>>;
-
-template<typename T>
-using Set = std::set<T, std::less<T>, fuzzer_allocator<T>>;
-
-typedef Vector<uint8_t> Unit;
-typedef Vector<Unit> UnitVector;
+typedef std::vector<uint8_t> Unit;
+typedef std::vector<Unit> UnitVector;
 typedef int (*UserCallback)(const uint8_t *Data, size_t Size);
 
 int FuzzerDriver(int *argc, char ***argv, UserCallback Callback);
diff --git a/compiler-rt/lib/fuzzer/FuzzerDictionary.h b/compiler-rt/lib/fuzzer/FuzzerDictionary.h
index db55907d9363..48f063c7ee4e 100644
--- a/compiler-rt/lib/fuzzer/FuzzerDictionary.h
+++ b/compiler-rt/lib/fuzzer/FuzzerDictionary.h
@@ -52,10 +52,13 @@ class DictionaryEntry {
  public:
   DictionaryEntry() {}
   DictionaryEntry(Word W) : W(W) {}
-  DictionaryEntry(Word W, size_t PositionHint) : W(W), PositionHint(PositionHint) {}
+  DictionaryEntry(Word W, size_t PositionHint)
+      : W(W), PositionHint(PositionHint) {}
   const Word &GetW() const { return W; }
 
-  bool HasPositionHint() const { return PositionHint != std::numeric_limits<size_t>::max(); }
+  bool HasPositionHint() const {
+    return PositionHint != std::numeric_limits<size_t>::max();
+  }
   size_t GetPositionHint() const {
     assert(HasPositionHint());
     return PositionHint;
@@ -108,12 +111,12 @@ private:
 };
 
 // Parses one dictionary entry.
-// If successful, write the enty to Unit and returns true,
+// If successful, writes the entry to Unit and returns true,
 // otherwise returns false.
 bool ParseOneDictionaryEntry(const std::string &Str, Unit *U);
 // Parses the dictionary file, fills Units, returns true iff all lines
 // were parsed successfully.
-bool ParseDictionaryFile(const std::string &Text, Vector<Unit> *Units);
+bool ParseDictionaryFile(const std::string &Text, std::vector<Unit> *Units);
 
 }  // namespace fuzzer
 
diff --git a/compiler-rt/lib/fuzzer/FuzzerDriver.cpp b/compiler-rt/lib/fuzzer/FuzzerDriver.cpp
index ceaa9070512f..6b007f2ad45c 100644
--- a/compiler-rt/lib/fuzzer/FuzzerDriver.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerDriver.cpp
@@ -86,7 +86,7 @@ static const FlagDescription FlagDescriptions [] {
 static const size_t kNumFlags =
     sizeof(FlagDescriptions) / sizeof(FlagDescriptions[0]);
 
-static Vector<std::string> *Inputs;
+static std::vector<std::string> *Inputs;
 static std::string *ProgName;
 
 static void PrintHelp() {
@@ -187,7 +187,7 @@ static bool ParseOneFlag(const char *Param) {
 }
 
 // We don't use any library to minimize dependencies.
-static void ParseFlags(const Vector<std::string> &Args,
+static void ParseFlags(const std::vector<std::string> &Args,
                        const ExternalFunctions *EF) {
   for (size_t F = 0; F < kNumFlags; F++) {
     if (FlagDescriptions[F].IntFlag)
@@ -206,7 +206,7 @@ static void ParseFlags(const Vector<std::string> &Args,
            "Disabling -len_control by default.\n", EF->LLVMFuzzerCustomMutator);
   }
 
-  Inputs = new Vector<std::string>;
+  Inputs = new std::vector<std::string>;
   for (size_t A = 1; A < Args.size(); A++) {
     if (ParseOneFlag(Args[A].c_str())) {
       if (Flags.ignore_remaining_args)
@@ -272,7 +272,7 @@ static void ValidateDirectoryExists(const std::string &Path,
   exit(1);
 }
 
-std::string CloneArgsWithoutX(const Vector<std::string> &Args,
+std::string CloneArgsWithoutX(const std::vector<std::string> &Args,
                               const char *X1, const char *X2) {
   std::string Cmd;
   for (auto &S : Args) {
@@ -283,18 +283,19 @@ std::string CloneArgsWithoutX(const Vector<std::string> &Args,
   return Cmd;
 }
 
-static int RunInMultipleProcesses(const Vector<std::string> &Args,
+static int RunInMultipleProcesses(const std::vector<std::string> &Args,
                                   unsigned NumWorkers, unsigned NumJobs) {
   std::atomic<unsigned> Counter(0);
   std::atomic<bool> HasErrors(false);
   Command Cmd(Args);
   Cmd.removeFlag("jobs");
   Cmd.removeFlag("workers");
-  Vector<std::thread> V;
+  std::vector<std::thread> V;
   std::thread Pulse(PulseThread);
   Pulse.detach();
   for (unsigned i = 0; i < NumWorkers; i++)
-    V.push_back(std::thread(WorkerThread, std::ref(Cmd), &Counter, NumJobs, &HasErrors));
+    V.push_back(std::thread(WorkerThread, std::ref(Cmd), &Counter, NumJobs,
+                            &HasErrors));
   for (auto &T : V)
     T.join();
   return HasErrors ? 1 : 0;
@@ -348,8 +349,8 @@ static std::string GetDedupTokenFromCmdOutput(const std::string &S) {
   return S.substr(Beg, End - Beg);
 }
 
-int CleanseCrashInput(const Vector<std::string> &Args,
-                       const FuzzingOptions &Options) {
+int CleanseCrashInput(const std::vector<std::string> &Args,
+                      const FuzzingOptions &Options) {
   if (Inputs->size() != 1 || !Flags.exact_artifact_path) {
     Printf("ERROR: -cleanse_crash should be given one input file and"
           " -exact_artifact_path\n");
@@ -372,7 +373,7 @@ int CleanseCrashInput(const Vector<std::string> &Args,
   auto U = FileToVector(CurrentFilePath);
   size_t Size = U.size();
 
-  const Vector<uint8_t> ReplacementBytes = {' ', 0xff};
+  const std::vector<uint8_t> ReplacementBytes = {' ', 0xff};
   for (int NumAttempts = 0; NumAttempts < 5; NumAttempts++) {
     bool Changed = false;
     for (size_t Idx = 0; Idx < Size; Idx++) {
@@ -403,7 +404,7 @@ int CleanseCrashInput(const Vector<std::string> &Args,
   return 0;
 }
 
-int MinimizeCrashInput(const Vector<std::string> &Args,
+int MinimizeCrashInput(const std::vector<std::string> &Args,
                        const FuzzingOptions &Options) {
   if (Inputs->size() != 1) {
     Printf("ERROR: -minimize_crash should be given one input file\n");
@@ -503,14 +504,15 @@ int MinimizeCrashInputInternalStep(Fuzzer *F, InputCorpus *Corpus) {
   return 0;
 }
 
-void Merge(Fuzzer *F, FuzzingOptions &Options, const Vector<std::string> &Args,
-           const Vector<std::string> &Corpora, const char *CFPathOrNull) {
+void Merge(Fuzzer *F, FuzzingOptions &Options,
+           const std::vector<std::string> &Args,
+           const std::vector<std::string> &Corpora, const char *CFPathOrNull) {
   if (Corpora.size() < 2) {
     Printf("INFO: Merge requires two or more corpus dirs\n");
     exit(0);
   }
 
-  Vector<SizedFile> OldCorpus, NewCorpus;
+  std::vector<SizedFile> OldCorpus, NewCorpus;
   GetSizedFilesFromDir(Corpora[0], &OldCorpus);
   for (size_t i = 1; i < Corpora.size(); i++)
     GetSizedFilesFromDir(Corpora[i], &NewCorpus);
@@ -518,10 +520,10 @@ void Merge(Fuzzer *F, FuzzingOptions &Options, const Vector<std::string> &Args,
   std::sort(NewCorpus.begin(), NewCorpus.end());
 
   std::string CFPath = CFPathOrNull ? CFPathOrNull : TempPath("Merge", ".txt");
-  Vector<std::string> NewFiles;
-  Set<uint32_t> NewFeatures, NewCov;
+  std::vector<std::string> NewFiles;
+  std::set<uint32_t> NewFeatures, NewCov;
   CrashResistantMerge(Args, OldCorpus, NewCorpus, &NewFiles, {}, &NewFeatures,
-                      {}, &NewCov, CFPath, true);
+                      {}, &NewCov, CFPath, true, Flags.set_cover_merge);
   for (auto &Path : NewFiles)
     F->WriteToOutputCorpus(FileToVector(Path, Options.MaxLen));
   // We are done, delete the control file if it was a temporary one.
@@ -531,17 +533,17 @@ void Merge(Fuzzer *F, FuzzingOptions &Options, const Vector<std::string> &Args,
   exit(0);
 }
 
-int AnalyzeDictionary(Fuzzer *F, const Vector<Unit>& Dict,
-                      UnitVector& Corpus) {
+int AnalyzeDictionary(Fuzzer *F, const std::vector<Unit> &Dict,
+                      UnitVector &Corpus) {
   Printf("Started dictionary minimization (up to %d tests)\n",
          Dict.size() * Corpus.size() * 2);
 
   // Scores and usage count for each dictionary unit.
-  Vector<int> Scores(Dict.size());
-  Vector<int> Usages(Dict.size());
+  std::vector<int> Scores(Dict.size());
+  std::vector<int> Usages(Dict.size());
 
-  Vector<size_t> InitialFeatures;
-  Vector<size_t> ModifiedFeatures;
+  std::vector<size_t> InitialFeatures;
+  std::vector<size_t> ModifiedFeatures;
   for (auto &C : Corpus) {
     // Get coverage for the testcase without modifications.
     F->ExecuteCallback(C.data(), C.size());
@@ -551,7 +553,7 @@ int AnalyzeDictionary(Fuzzer *F, const Vector<Unit>& Dict,
     });
 
     for (size_t i = 0; i < Dict.size(); ++i) {
-      Vector<uint8_t> Data = C;
+      std::vector<uint8_t> Data = C;
       auto StartPos = std::search(Data.begin(), Data.end(),
                                   Dict[i].begin(), Dict[i].end());
       // Skip dictionary unit, if the testcase does not contain it.
@@ -597,9 +599,9 @@ int AnalyzeDictionary(Fuzzer *F, const Vector<Unit>& Dict,
   return 0;
 }
 
-Vector<std::string> ParseSeedInuts(const char *seed_inputs) {
+std::vector<std::string> ParseSeedInuts(const char *seed_inputs) {
   // Parse -seed_inputs=file1,file2,... or -seed_inputs=@seed_inputs_file
-  Vector<std::string> Files;
+  std::vector<std::string> Files;
   if (!seed_inputs) return Files;
   std::string SeedInputs;
   if (Flags.seed_inputs[0] == '@')
@@ -620,9 +622,10 @@ Vector<std::string> ParseSeedInuts(const char *seed_inputs) {
   return Files;
 }
 
-static Vector<SizedFile> ReadCorpora(const Vector<std::string> &CorpusDirs,
-    const Vector<std::string> &ExtraSeedFiles) {
-  Vector<SizedFile> SizedFiles;
+static std::vector<SizedFile>
+ReadCorpora(const std::vector<std::string> &CorpusDirs,
+            const std::vector<std::string> &ExtraSeedFiles) {
+  std::vector<SizedFile> SizedFiles;
   size_t LastNumFiles = 0;
   for (auto &Dir : CorpusDirs) {
     GetSizedFilesFromDir(Dir, &SizedFiles);
@@ -645,7 +648,7 @@ int FuzzerDriver(int *argc, char ***argv, UserCallback Callback) {
     EF->LLVMFuzzerInitialize(argc, argv);
   if (EF->__msan_scoped_disable_interceptor_checks)
     EF->__msan_scoped_disable_interceptor_checks();
-  const Vector<std::string> Args(*argv, *argv + *argc);
+  const std::vector<std::string> Args(*argv, *argv + *argc);
   assert(!Args.empty());
   ProgName = new std::string(Args[0]);
   if (Argv0 != *ProgName) {
@@ -734,7 +737,7 @@ int FuzzerDriver(int *argc, char ***argv, UserCallback Callback) {
     ValidateDirectoryExists(DirName(Options.ExactArtifactPath),
                             Flags.create_missing_dirs);
   }
-  Vector<Unit> Dictionary;
+  std::vector<Unit> Dictionary;
   if (Flags.dict)
     if (!ParseDictionaryFile(FileToString(Flags.dict), &Dictionary))
       return 1;
@@ -794,7 +797,8 @@ int FuzzerDriver(int *argc, char ***argv, UserCallback Callback) {
   if (Flags.verbosity)
     Printf("INFO: Seed: %u\n", Seed);
 
-  if (Flags.collect_data_flow && !Flags.fork && !Flags.merge) {
+  if (Flags.collect_data_flow && !Flags.fork &&
+      !(Flags.merge || Flags.set_cover_merge)) {
     if (RunIndividualFiles)
       return CollectDataFlow(Flags.collect_data_flow, Flags.data_flow_trace,
                         ReadCorpora({}, *Inputs));
@@ -866,10 +870,11 @@ int FuzzerDriver(int *argc, char ***argv, UserCallback Callback) {
     exit(0);
   }
 
+  Options.ForkCorpusGroups = Flags.fork_corpus_groups;
   if (Flags.fork)
     FuzzWithFork(F->GetMD().GetRand(), Options, Args, *Inputs, Flags.fork);
 
-  if (Flags.merge)
+  if (Flags.merge || Flags.set_cover_merge)
     Merge(F, Options, Args, *Inputs, Flags.merge_control_file);
 
   if (Flags.merge_inner) {
@@ -877,7 +882,8 @@ int FuzzerDriver(int *argc, char ***argv, UserCallback Callback) {
     if (Options.MaxLen == 0)
       F->SetMaxInputLen(kDefaultMaxMergeLen);
     assert(Flags.merge_control_file);
-    F->CrashResistantMergeInternalStep(Flags.merge_control_file);
+    F->CrashResistantMergeInternalStep(Flags.merge_control_file,
+                                       !strncmp(Flags.merge_inner, "2", 1));
     exit(0);
   }
 
diff --git a/compiler-rt/lib/fuzzer/FuzzerExtraCounters.cpp b/compiler-rt/lib/fuzzer/FuzzerExtraCounters.cpp
index 04f569a1a879..54ecbf7c62f1 100644
--- a/compiler-rt/lib/fuzzer/FuzzerExtraCounters.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerExtraCounters.cpp
@@ -31,12 +31,4 @@ void ClearExtraCounters() {  // hand-written memset, don't asan-ify.
 
 }  // namespace fuzzer
 
-#else
-// TODO: implement for other platforms.
-namespace fuzzer {
-uint8_t *ExtraCountersBegin() { return nullptr; }
-uint8_t *ExtraCountersEnd() { return nullptr; }
-void ClearExtraCounters() {}
-}  // namespace fuzzer
-
 #endif
diff --git a/compiler-rt/lib/fuzzer/FuzzerExtraCountersDarwin.cpp b/compiler-rt/lib/fuzzer/FuzzerExtraCountersDarwin.cpp
new file mode 100644
index 000000000000..2321ba8a3d40
--- /dev/null
+++ b/compiler-rt/lib/fuzzer/FuzzerExtraCountersDarwin.cpp
@@ -0,0 +1,22 @@
+//===- FuzzerExtraCountersDarwin.cpp - Extra coverage counters for Darwin -===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+// Extra coverage counters defined by user code for Darwin.
+//===----------------------------------------------------------------------===//
+
+#include "FuzzerPlatform.h"
+#include <cstdint>
+
+#if LIBFUZZER_APPLE
+
+namespace fuzzer {
+uint8_t *ExtraCountersBegin() { return nullptr; }
+uint8_t *ExtraCountersEnd() { return nullptr; }
+void ClearExtraCounters() {}
+} // namespace fuzzer
+
+#endif
diff --git a/compiler-rt/lib/fuzzer/FuzzerExtraCountersWindows.cpp b/compiler-rt/lib/fuzzer/FuzzerExtraCountersWindows.cpp
new file mode 100644
index 000000000000..102f5febdaec
--- /dev/null
+++ b/compiler-rt/lib/fuzzer/FuzzerExtraCountersWindows.cpp
@@ -0,0 +1,80 @@
+//===- FuzzerExtraCountersWindows.cpp - Extra coverage counters for Win32 -===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+// Extra coverage counters defined by user code for Windows.
+//===----------------------------------------------------------------------===//
+
+#include "FuzzerPlatform.h"
+#include <cstdint>
+
+#if LIBFUZZER_WINDOWS
+#include <windows.h>
+
+namespace fuzzer {
+
+//
+// The __start___libfuzzer_extra_counters variable is align 16, size 16 to
+// ensure the padding between it and the next variable in this section (either
+// __libfuzzer_extra_counters or __stop___libfuzzer_extra_counters) will be
+// located at (__start___libfuzzer_extra_counters +
+// sizeof(__start___libfuzzer_extra_counters)). Otherwise, the calculation of
+// (stop - (start + sizeof(start))) might be skewed.
+//
+// The section name, __libfuzzer_extra_countaaa ends with "aaa", so it sorts
+// before __libfuzzer_extra_counters alphabetically. We want the start symbol to
+// be placed in the section just before the user supplied counters (if present).
+//
+#pragma section(".data$__libfuzzer_extra_countaaa")
+ATTRIBUTE_ALIGNED(16)
+__declspec(allocate(".data$__libfuzzer_extra_countaaa")) uint8_t
+    __start___libfuzzer_extra_counters[16] = {0};
+
+//
+// Example of what the user-supplied counters should look like. First, the
+// pragma to create the section name. It will fall alphabetically between
+// ".data$__libfuzzer_extra_countaaa" and ".data$__libfuzzer_extra_countzzz".
+// Next, the declspec to allocate the variable inside the specified section.
+// Finally, some array, struct, whatever that is used to track the counter data.
+// The size of this variable is computed at runtime by finding the difference of
+// __stop___libfuzzer_extra_counters and __start___libfuzzer_extra_counters +
+// sizeof(__start___libfuzzer_extra_counters).
+//
+
+//
+//     #pragma section(".data$__libfuzzer_extra_counters")
+//     __declspec(allocate(".data$__libfuzzer_extra_counters"))
+//         uint8_t any_name_variable[64 * 1024];
+//
+
+//
+// Here, the section name, __libfuzzer_extra_countzzz ends with "zzz", so it
+// sorts after __libfuzzer_extra_counters alphabetically. We want the stop
+// symbol to be placed in the section just after the user supplied counters (if
+// present). Align to 1 so there isn't any padding placed between this and the
+// previous variable.
+//
+#pragma section(".data$__libfuzzer_extra_countzzz")
+ATTRIBUTE_ALIGNED(1)
+__declspec(allocate(".data$__libfuzzer_extra_countzzz")) uint8_t
+    __stop___libfuzzer_extra_counters = 0;
+
+uint8_t *ExtraCountersBegin() {
+  return __start___libfuzzer_extra_counters +
+         sizeof(__start___libfuzzer_extra_counters);
+}
+
+uint8_t *ExtraCountersEnd() { return &__stop___libfuzzer_extra_counters; }
+
+ATTRIBUTE_NO_SANITIZE_ALL
+void ClearExtraCounters() {
+  uint8_t *Beg = ExtraCountersBegin();
+  SecureZeroMemory(Beg, ExtraCountersEnd() - Beg);
+}
+
+} // namespace fuzzer
+
+#endif
diff --git a/compiler-rt/lib/fuzzer/FuzzerFlags.def b/compiler-rt/lib/fuzzer/FuzzerFlags.def
index ab31da0ae5d6..11815349b014 100644
--- a/compiler-rt/lib/fuzzer/FuzzerFlags.def
+++ b/compiler-rt/lib/fuzzer/FuzzerFlags.def
@@ -58,12 +58,21 @@ FUZZER_FLAG_INT(max_total_time, 0, "If positive, indicates the maximal total "
 FUZZER_FLAG_INT(help, 0, "Print help.")
 FUZZER_FLAG_INT(fork, 0, "Experimental mode where fuzzing happens "
                 "in a subprocess")
+FUZZER_FLAG_INT(fork_corpus_groups, 0, "For fork mode, enable the corpus-group "
+		"strategy, The main corpus will be grouped according to size, "
+		"and each sub-process will randomly select seeds from different "
+		"groups as the sub-corpus.")
 FUZZER_FLAG_INT(ignore_timeouts, 1, "Ignore timeouts in fork mode")
 FUZZER_FLAG_INT(ignore_ooms, 1, "Ignore OOMs in fork mode")
 FUZZER_FLAG_INT(ignore_crashes, 0, "Ignore crashes in fork mode")
 FUZZER_FLAG_INT(merge, 0, "If 1, the 2-nd, 3-rd, etc corpora will be "
   "merged into the 1-st corpus. Only interesting units will be taken. "
   "This flag can be used to minimize a corpus.")
+FUZZER_FLAG_INT(set_cover_merge, 0, "If 1, the 2-nd, 3-rd, etc corpora will be "
+  "merged into the 1-st corpus. Same as the 'merge' flag, but uses the "
+  "standard greedy algorithm for the set cover problem to "
+  "compute an approximation of the minimum set of testcases that "
+  "provide the same coverage as the initial corpora")
 FUZZER_FLAG_STRING(stop_file, "Stop fuzzing ASAP if this file exists")
 FUZZER_FLAG_STRING(merge_inner, "internal flag")
 FUZZER_FLAG_STRING(merge_control_file,
diff --git a/compiler-rt/lib/fuzzer/FuzzerFork.cpp b/compiler-rt/lib/fuzzer/FuzzerFork.cpp
index 5134a5d979e6..d59d51384201 100644
--- a/compiler-rt/lib/fuzzer/FuzzerFork.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerFork.cpp
@@ -86,18 +86,21 @@ struct FuzzJob {
 };
 
 struct GlobalEnv {
-  Vector<std::string> Args;
-  Vector<std::string> CorpusDirs;
+  std::vector<std::string> Args;
+  std::vector<std::string> CorpusDirs;
   std::string MainCorpusDir;
   std::string TempDir;
   std::string DFTDir;
   std::string DataFlowBinary;
-  Set<uint32_t> Features, Cov;
-  Set<std::string> FilesWithDFT;
-  Vector<std::string> Files;
+  std::set<uint32_t> Features, Cov;
+  std::set<std::string> FilesWithDFT;
+  std::vector<std::string> Files;
+  std::vector<std::size_t> FilesSizes;
   Random *Rand;
   std::chrono::system_clock::time_point ProcessStartTime;
   int Verbosity = 0;
+  int Group = 0;
+  int NumCorpuses = 8;
 
   size_t NumTimeouts = 0;
   size_t NumOOMs = 0;
@@ -136,10 +139,24 @@ struct GlobalEnv {
     if (size_t CorpusSubsetSize =
             std::min(Files.size(), (size_t)sqrt(Files.size() + 2))) {
       auto Time1 = std::chrono::system_clock::now();
-      for (size_t i = 0; i < CorpusSubsetSize; i++) {
-        auto &SF = Files[Rand->SkewTowardsLast(Files.size())];
-        Seeds += (Seeds.empty() ? "" : ",") + SF;
-        CollectDFT(SF);
+      if (Group) { // whether to group the corpus.
+        size_t AverageCorpusSize = Files.size() / NumCorpuses + 1;
+        size_t StartIndex = ((JobId - 1) % NumCorpuses) * AverageCorpusSize;
+        for (size_t i = 0; i < CorpusSubsetSize; i++) {
+          size_t RandNum = (*Rand)(AverageCorpusSize);
+          size_t Index = RandNum + StartIndex;
+          Index = Index < Files.size() ? Index
+                                       : Rand->SkewTowardsLast(Files.size());
+          auto &SF = Files[Index];
+          Seeds += (Seeds.empty() ? "" : ",") + SF;
+          CollectDFT(SF);
+        }
+      } else {
+        for (size_t i = 0; i < CorpusSubsetSize; i++) {
+          auto &SF = Files[Rand->SkewTowardsLast(Files.size())];
+          Seeds += (Seeds.empty() ? "" : ",") + SF;
+          CollectDFT(SF);
+        }
       }
       auto Time2 = std::chrono::system_clock::now();
       auto DftTimeInSeconds = duration_cast<seconds>(Time2 - Time1).count();
@@ -183,7 +200,7 @@ struct GlobalEnv {
     auto Stats = ParseFinalStatsFromLog(Job->LogPath);
     NumRuns += Stats.number_of_executed_units;
 
-    Vector<SizedFile> TempFiles, MergeCandidates;
+    std::vector<SizedFile> TempFiles, MergeCandidates;
     // Read all newly created inputs and their feature sets.
     // Choose only those inputs that have new features.
     GetSizedFilesFromDir(Job->CorpusDir, &TempFiles);
@@ -193,7 +210,7 @@ struct GlobalEnv {
       FeatureFile.replace(0, Job->CorpusDir.size(), Job->FeaturesDir);
       auto FeatureBytes = FileToVector(FeatureFile, 0, false);
       assert((FeatureBytes.size() % sizeof(uint32_t)) == 0);
-      Vector<uint32_t> NewFeatures(FeatureBytes.size() / sizeof(uint32_t));
+      std::vector<uint32_t> NewFeatures(FeatureBytes.size() / sizeof(uint32_t));
       memcpy(NewFeatures.data(), FeatureBytes.data(), FeatureBytes.size());
       for (auto Ft : NewFeatures) {
         if (!Features.count(Ft)) {
@@ -211,15 +228,27 @@ struct GlobalEnv {
 
     if (MergeCandidates.empty()) return;
 
-    Vector<std::string> FilesToAdd;
-    Set<uint32_t> NewFeatures, NewCov;
+    std::vector<std::string> FilesToAdd;
+    std::set<uint32_t> NewFeatures, NewCov;
+    bool IsSetCoverMerge =
+        !Job->Cmd.getFlagValue("set_cover_merge").compare("1");
     CrashResistantMerge(Args, {}, MergeCandidates, &FilesToAdd, Features,
-                        &NewFeatures, Cov, &NewCov, Job->CFPath, false);
+                        &NewFeatures, Cov, &NewCov, Job->CFPath, false,
+                        IsSetCoverMerge);
     for (auto &Path : FilesToAdd) {
       auto U = FileToVector(Path);
       auto NewPath = DirPlusFile(MainCorpusDir, Hash(U));
       WriteToFile(U, NewPath);
-      Files.push_back(NewPath);
+      if (Group) { // Insert the queue according to the size of the seed.
+        size_t UnitSize = U.size();
+        auto Idx =
+            std::upper_bound(FilesSizes.begin(), FilesSizes.end(), UnitSize) -
+            FilesSizes.begin();
+        FilesSizes.insert(FilesSizes.begin() + Idx, UnitSize);
+        Files.insert(Files.begin() + Idx, NewPath);
+      } else {
+        Files.push_back(NewPath);
+      }
     }
     Features.insert(NewFeatures.begin(), NewFeatures.end());
     Cov.insert(NewCov.begin(), NewCov.end());
@@ -228,10 +257,8 @@ struct GlobalEnv {
         if (TPC.PcIsFuncEntry(TE))
           PrintPC("  NEW_FUNC: %p %F %L\n", "",
                   TPC.GetNextInstructionPc(TE->PC));
-
   }
 
-
   void CollectDFT(const std::string &InputPath) {
     if (DataFlowBinary.empty()) return;
     if (!FilesWithDFT.insert(InputPath).second) return;
@@ -283,8 +310,8 @@ void WorkerThread(JobQueue *FuzzQ, JobQueue *MergeQ) {
 
 // This is just a skeleton of an experimental -fork=1 feature.
 void FuzzWithFork(Random &Rand, const FuzzingOptions &Options,
-                  const Vector<std::string> &Args,
-                  const Vector<std::string> &CorpusDirs, int NumJobs) {
+                  const std::vector<std::string> &Args,
+                  const std::vector<std::string> &CorpusDirs, int NumJobs) {
   Printf("INFO: -fork=%d: fuzzing in separate process(s)\n", NumJobs);
 
   GlobalEnv Env;
@@ -294,8 +321,9 @@ void FuzzWithFork(Random &Rand, const FuzzingOptions &Options,
   Env.Verbosity = Options.Verbosity;
   Env.ProcessStartTime = std::chrono::system_clock::now();
   Env.DataFlowBinary = Options.CollectDataFlow;
+  Env.Group = Options.ForkCorpusGroups;
 
-  Vector<SizedFile> SeedFiles;
+  std::vector<SizedFile> SeedFiles;
   for (auto &Dir : CorpusDirs)
     GetSizedFilesFromDir(Dir, &SeedFiles);
   std::sort(SeedFiles.begin(), SeedFiles.end());
@@ -316,13 +344,20 @@ void FuzzWithFork(Random &Rand, const FuzzingOptions &Options,
       Env.Files.push_back(File.File);
   } else {
     auto CFPath = DirPlusFile(Env.TempDir, "merge.txt");
-    Set<uint32_t> NewFeatures, NewCov;
+    std::set<uint32_t> NewFeatures, NewCov;
     CrashResistantMerge(Env.Args, {}, SeedFiles, &Env.Files, Env.Features,
-                        &NewFeatures, Env.Cov, &NewCov, CFPath, false);
+                        &NewFeatures, Env.Cov, &NewCov, CFPath,
+                        /*Verbose=*/false, /*IsSetCoverMerge=*/false);
     Env.Features.insert(NewFeatures.begin(), NewFeatures.end());
     Env.Cov.insert(NewFeatures.begin(), NewFeatures.end());
     RemoveFile(CFPath);
   }
+
+  if (Env.Group) {
+    for (auto &path : Env.Files)
+      Env.FilesSizes.push_back(FileSize(path));
+  }
+
   Printf("INFO: -fork=%d: %zd seed inputs, starting to fuzz in %s\n", NumJobs,
          Env.Files.size(), Env.TempDir.c_str());
 
@@ -337,8 +372,10 @@ void FuzzWithFork(Random &Rand, const FuzzingOptions &Options,
     WriteToFile(Unit({1}), Env.StopFile());
   };
 
+  size_t MergeCycle = 20;
+  size_t JobExecuted = 0;
   size_t JobId = 1;
-  Vector<std::thread> Threads;
+  std::vector<std::thread> Threads;
   for (int t = 0; t < NumJobs; t++) {
     Threads.push_back(std::thread(WorkerThread, &FuzzQ, &MergeQ));
     FuzzQ.Push(Env.CreateNewJob(JobId++));
@@ -358,7 +395,46 @@ void FuzzWithFork(Random &Rand, const FuzzingOptions &Options,
 
     Env.RunOneMergeJob(Job.get());
 
-    // Continue if our crash is one of the ignorred ones.
+    // merge the corpus .
+    JobExecuted++;
+    if (Env.Group && JobExecuted >= MergeCycle) {
+      std::vector<SizedFile> CurrentSeedFiles;
+      for (auto &Dir : CorpusDirs)
+        GetSizedFilesFromDir(Dir, &CurrentSeedFiles);
+      std::sort(CurrentSeedFiles.begin(), CurrentSeedFiles.end());
+
+      auto CFPath = DirPlusFile(Env.TempDir, "merge.txt");
+      std::set<uint32_t> TmpNewFeatures, TmpNewCov;
+      std::set<uint32_t> TmpFeatures, TmpCov;
+      Env.Files.clear();
+      Env.FilesSizes.clear();
+      CrashResistantMerge(Env.Args, {}, CurrentSeedFiles, &Env.Files,
+                          TmpFeatures, &TmpNewFeatures, TmpCov, &TmpNewCov,
+                          CFPath, /*Verbose=*/false, /*IsSetCoverMerge=*/false);
+      for (auto &path : Env.Files)
+        Env.FilesSizes.push_back(FileSize(path));
+      RemoveFile(CFPath);
+      JobExecuted = 0;
+      MergeCycle += 5;
+    }
+
+    // Since the number of corpus seeds will gradually increase, in order to
+    // control the number in each group to be about three times the number of
+    // seeds selected each time, the number of groups is dynamically adjusted.
+    if (Env.Files.size() < 2000)
+      Env.NumCorpuses = 12;
+    else if (Env.Files.size() < 6000)
+      Env.NumCorpuses = 20;
+    else if (Env.Files.size() < 12000)
+      Env.NumCorpuses = 32;
+    else if (Env.Files.size() < 16000)
+      Env.NumCorpuses = 40;
+    else if (Env.Files.size() < 24000)
+      Env.NumCorpuses = 60;
+    else
+      Env.NumCorpuses = 80;
+
+    // Continue if our crash is one of the ignored ones.
     if (Options.IgnoreTimeouts && ExitCode == Options.TimeoutExitCode)
       Env.NumTimeouts++;
     else if (Options.IgnoreOOMs && ExitCode == Options.OOMExitCode)
diff --git a/compiler-rt/lib/fuzzer/FuzzerFork.h b/compiler-rt/lib/fuzzer/FuzzerFork.h
index b29a43e13fbc..fc3e9d636cbc 100644
--- a/compiler-rt/lib/fuzzer/FuzzerFork.h
+++ b/compiler-rt/lib/fuzzer/FuzzerFork.h
@@ -17,8 +17,8 @@
 
 namespace fuzzer {
 void FuzzWithFork(Random &Rand, const FuzzingOptions &Options,
-                  const Vector<std::string> &Args,
-                  const Vector<std::string> &CorpusDirs, int NumJobs);
+                  const std::vector<std::string> &Args,
+                  const std::vector<std::string> &CorpusDirs, int NumJobs);
 } // namespace fuzzer
 
 #endif // LLVM_FUZZER_FORK_H
diff --git a/compiler-rt/lib/fuzzer/FuzzerIO.cpp b/compiler-rt/lib/fuzzer/FuzzerIO.cpp
index 7f149ac6c485..0a58c5377b34 100644
--- a/compiler-rt/lib/fuzzer/FuzzerIO.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerIO.cpp
@@ -23,6 +23,14 @@ namespace fuzzer {
 
 static FILE *OutputFile = stderr;
 
+FILE *GetOutputFile() {
+  return OutputFile;
+}
+
+void SetOutputFile(FILE *NewOutputFile) {
+  OutputFile = NewOutputFile;
+}
+
 long GetEpoch(const std::string &Path) {
   struct stat St;
   if (stat(Path.c_str(), &St))
@@ -90,11 +98,11 @@ void AppendToFile(const uint8_t *Data, size_t Size, const std::string &Path) {
   fclose(Out);
 }
 
-void ReadDirToVectorOfUnits(const char *Path, Vector<Unit> *V, long *Epoch,
+void ReadDirToVectorOfUnits(const char *Path, std::vector<Unit> *V, long *Epoch,
                             size_t MaxSize, bool ExitOnError,
-                            Vector<std::string> *VPaths) {
+                            std::vector<std::string> *VPaths) {
   long E = Epoch ? *Epoch : 0;
-  Vector<std::string> Files;
+  std::vector<std::string> Files;
   ListFilesInDirRecursive(Path, Epoch, &Files, /*TopDir*/true);
   size_t NumLoaded = 0;
   for (size_t i = 0; i < Files.size(); i++) {
@@ -112,8 +120,8 @@ void ReadDirToVectorOfUnits(const char *Path, Vector<Unit> *V, long *Epoch,
   }
 }
 
-void GetSizedFilesFromDir(const std::string &Dir, Vector<SizedFile> *V) {
-  Vector<std::string> Files;
+void GetSizedFilesFromDir(const std::string &Dir, std::vector<SizedFile> *V) {
+  std::vector<std::string> Files;
   ListFilesInDirRecursive(Dir, 0, &Files, /*TopDir*/true);
   for (auto &File : Files)
     if (size_t Size = FileSize(File))
diff --git a/compiler-rt/lib/fuzzer/FuzzerIO.h b/compiler-rt/lib/fuzzer/FuzzerIO.h
index bde18267ea36..401afa0b4477 100644
--- a/compiler-rt/lib/fuzzer/FuzzerIO.h
+++ b/compiler-rt/lib/fuzzer/FuzzerIO.h
@@ -32,9 +32,9 @@ void WriteToFile(const Unit &U, const std::string &Path);
 void AppendToFile(const uint8_t *Data, size_t Size, const std::string &Path);
 void AppendToFile(const std::string &Data, const std::string &Path);
 
-void ReadDirToVectorOfUnits(const char *Path, Vector<Unit> *V, long *Epoch,
+void ReadDirToVectorOfUnits(const char *Path, std::vector<Unit> *V, long *Epoch,
                             size_t MaxSize, bool ExitOnError,
-                            Vector<std::string> *VPaths = 0);
+                            std::vector<std::string> *VPaths = 0);
 
 // Returns "Dir/FileName" or equivalent for the current OS.
 std::string DirPlusFile(const std::string &DirPath,
@@ -54,6 +54,10 @@ void DupAndCloseStderr();
 
 void CloseStdout();
 
+// For testing.
+FILE *GetOutputFile();
+void SetOutputFile(FILE *NewOutputFile);
+
 void Printf(const char *Fmt, ...);
 void VPrintf(bool Verbose, const char *Fmt, ...);
 
@@ -66,7 +70,7 @@ bool IsDirectory(const std::string &Path);
 size_t FileSize(const std::string &Path);
 
 void ListFilesInDirRecursive(const std::string &Dir, long *Epoch,
-                             Vector<std::string> *V, bool TopDir);
+                             std::vector<std::string> *V, bool TopDir);
 
 bool MkDirRecursive(const std::string &Dir);
 void RmDirRecursive(const std::string &Dir);
@@ -85,7 +89,7 @@ struct SizedFile {
   bool operator<(const SizedFile &B) const { return Size < B.Size; }
 };
 
-void GetSizedFilesFromDir(const std::string &Dir, Vector<SizedFile> *V);
+void GetSizedFilesFromDir(const std::string &Dir, std::vector<SizedFile> *V);
 
 char GetSeparator();
 bool IsSeparator(char C);
diff --git a/compiler-rt/lib/fuzzer/FuzzerIOPosix.cpp b/compiler-rt/lib/fuzzer/FuzzerIOPosix.cpp
index 4706a40959be..3700fb098e55 100644
--- a/compiler-rt/lib/fuzzer/FuzzerIOPosix.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerIOPosix.cpp
@@ -53,7 +53,7 @@ std::string Basename(const std::string &Path) {
 }
 
 void ListFilesInDirRecursive(const std::string &Dir, long *Epoch,
-                             Vector<std::string> *V, bool TopDir) {
+                             std::vector<std::string> *V, bool TopDir) {
   auto E = GetEpoch(Dir);
   if (Epoch)
     if (E && *Epoch >= E) return;
@@ -78,7 +78,6 @@ void ListFilesInDirRecursive(const std::string &Dir, long *Epoch,
     *Epoch = E;
 }
 
-
 void IterateDirRecursive(const std::string &Dir,
                          void (*DirPreCallback)(const std::string &Dir),
                          void (*DirPostCallback)(const std::string &Dir),
diff --git a/compiler-rt/lib/fuzzer/FuzzerIOWindows.cpp b/compiler-rt/lib/fuzzer/FuzzerIOWindows.cpp
index 61ad35e281f5..6771fc173c91 100644
--- a/compiler-rt/lib/fuzzer/FuzzerIOWindows.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerIOWindows.cpp
@@ -111,7 +111,7 @@ size_t FileSize(const std::string &Path) {
 }
 
 void ListFilesInDirRecursive(const std::string &Dir, long *Epoch,
-                             Vector<std::string> *V, bool TopDir) {
+                             std::vector<std::string> *V, bool TopDir) {
   auto E = GetEpoch(Dir);
   if (Epoch)
     if (E && *Epoch >= E) return;
@@ -159,7 +159,6 @@ void ListFilesInDirRecursive(const std::string &Dir, long *Epoch,
     *Epoch = E;
 }
 
-
 void IterateDirRecursive(const std::string &Dir,
                          void (*DirPreCallback)(const std::string &Dir),
                          void (*DirPostCallback)(const std::string &Dir),
@@ -297,9 +296,8 @@ static size_t ParseServerAndShare(const std::string &FileName,
   return Pos - Offset;
 }
 
-// Parse the given Ref string from the position Offset, to exactly match the given
-// string Patt.
-// Returns number of characters considered if successful.
+// Parse the given Ref string from the position Offset, to exactly match the
+// given string Patt. Returns number of characters considered if successful.
 static size_t ParseCustomString(const std::string &Ref, size_t Offset,
                                 const char *Patt) {
   size_t Len = strlen(Patt);
diff --git a/compiler-rt/lib/fuzzer/FuzzerInternal.h b/compiler-rt/lib/fuzzer/FuzzerInternal.h
index 37c8a01dc3c6..6637b0034e55 100644
--- a/compiler-rt/lib/fuzzer/FuzzerInternal.h
+++ b/compiler-rt/lib/fuzzer/FuzzerInternal.h
@@ -35,8 +35,8 @@ public:
   Fuzzer(UserCallback CB, InputCorpus &Corpus, MutationDispatcher &MD,
          FuzzingOptions Options);
   ~Fuzzer();
-  void Loop(Vector<SizedFile> &CorporaFiles);
-  void ReadAndExecuteSeedCorpora(Vector<SizedFile> &CorporaFiles);
+  void Loop(std::vector<SizedFile> &CorporaFiles);
+  void ReadAndExecuteSeedCorpora(std::vector<SizedFile> &CorporaFiles);
   void MinimizeCrashLoop(const Unit &U);
   void RereadOutputCorpus(size_t MaxSize);
 
@@ -72,8 +72,9 @@ public:
   void TPCUpdateObservedPCs();
 
   // Merge Corpora[1:] into Corpora[0].
-  void Merge(const Vector<std::string> &Corpora);
-  void CrashResistantMergeInternalStep(const std::string &ControlFilePath);
+  void Merge(const std::vector<std::string> &Corpora);
+  void CrashResistantMergeInternalStep(const std::string &ControlFilePath,
+                                       bool IsSetCoverMerge);
   MutationDispatcher &GetMD() { return MD; }
   void PrintFinalStats();
   void SetMaxInputLen(size_t MaxInputLen);
@@ -141,7 +142,7 @@ private:
   size_t MaxMutationLen = 0;
   size_t TmpMaxMutationLen = 0;
 
-  Vector<uint32_t> UniqFeatureSetTmp;
+  std::vector<uint32_t> UniqFeatureSetTmp;
 
   // Need to know our own thread.
   static thread_local bool IsMyThread;
diff --git a/compiler-rt/lib/fuzzer/FuzzerLoop.cpp b/compiler-rt/lib/fuzzer/FuzzerLoop.cpp
index 86a78ab75174..3205942f6d84 100644
--- a/compiler-rt/lib/fuzzer/FuzzerLoop.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerLoop.cpp
@@ -388,7 +388,7 @@ void Fuzzer::SetMaxMutationLen(size_t MaxMutationLen) {
 
 void Fuzzer::CheckExitOnSrcPosOrItem() {
   if (!Options.ExitOnSrcPos.empty()) {
-    static auto *PCsSet = new Set<uintptr_t>;
+    static auto *PCsSet = new std::set<uintptr_t>;
     auto HandlePC = [&](const TracePC::PCTableEntry *TE) {
       if (!PCsSet->insert(TE->PC).second)
         return;
@@ -413,8 +413,8 @@ void Fuzzer::CheckExitOnSrcPosOrItem() {
 void Fuzzer::RereadOutputCorpus(size_t MaxSize) {
   if (Options.OutputCorpus.empty() || !Options.ReloadIntervalSec)
     return;
-  Vector<Unit> AdditionalCorpus;
-  Vector<std::string> AdditionalCorpusPaths;
+  std::vector<Unit> AdditionalCorpus;
+  std::vector<std::string> AdditionalCorpusPaths;
   ReadDirToVectorOfUnits(
       Options.OutputCorpus.c_str(), &AdditionalCorpus,
       &EpochOfLastReadOfOutputCorpus, MaxSize,
@@ -457,7 +457,7 @@ void Fuzzer::PrintPulseAndReportSlowInput(const uint8_t *Data, size_t Size) {
 
 static void WriteFeatureSetToFile(const std::string &FeaturesDir,
                                   const std::string &FileName,
-                                  const Vector<uint32_t> &FeatureSet) {
+                                  const std::vector<uint32_t> &FeatureSet) {
   if (FeaturesDir.empty() || FeatureSet.empty()) return;
   WriteToFile(reinterpret_cast<const uint8_t *>(FeatureSet.data()),
               FeatureSet.size() * sizeof(FeatureSet[0]),
@@ -548,7 +548,7 @@ bool Fuzzer::RunOne(const uint8_t *Data, size_t Size, bool MayDeleteFile,
       FoundUniqFeaturesOfII == II->UniqFeatureSet.size() &&
       II->U.size() > Size) {
     auto OldFeaturesFile = Sha1ToString(II->Sha1);
-    Corpus.Replace(II, {Data, Data + Size});
+    Corpus.Replace(II, {Data, Data + Size}, TimeOfUnit);
     RenameFeatureSetFile(Options.FeaturesDir, OldFeaturesFile,
                          Sha1ToString(II->Sha1));
     return true;
@@ -784,7 +784,7 @@ void Fuzzer::PurgeAllocator() {
   LastAllocatorPurgeAttemptTime = system_clock::now();
 }
 
-void Fuzzer::ReadAndExecuteSeedCorpora(Vector<SizedFile> &CorporaFiles) {
+void Fuzzer::ReadAndExecuteSeedCorpora(std::vector<SizedFile> &CorporaFiles) {
   const size_t kMaxSaneLen = 1 << 20;
   const size_t kMinDefaultLen = 4096;
   size_t MaxSize = 0;
@@ -849,7 +849,7 @@ void Fuzzer::ReadAndExecuteSeedCorpora(Vector<SizedFile> &CorporaFiles) {
   }
 }
 
-void Fuzzer::Loop(Vector<SizedFile> &CorporaFiles) {
+void Fuzzer::Loop(std::vector<SizedFile> &CorporaFiles) {
   auto FocusFunctionOrAuto = Options.FocusFunction;
   DFT.Init(Options.DataFlowTrace, &FocusFunctionOrAuto, CorporaFiles,
            MD.GetRand());
diff --git a/compiler-rt/lib/fuzzer/FuzzerMerge.cpp b/compiler-rt/lib/fuzzer/FuzzerMerge.cpp
index 162453ceae2c..24bd11958e80 100644
--- a/compiler-rt/lib/fuzzer/FuzzerMerge.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerMerge.cpp
@@ -77,8 +77,8 @@ bool Merger::Parse(std::istream &IS, bool ParseCoverage) {
   size_t ExpectedStartMarker = 0;
   const size_t kInvalidStartMarker = -1;
   size_t LastSeenStartMarker = kInvalidStartMarker;
-  Vector<uint32_t> TmpFeatures;
-  Set<uint32_t> PCs;
+  std::vector<uint32_t> TmpFeatures;
+  std::set<uint32_t> PCs;
   while (std::getline(IS, Line, '\n')) {
     std::istringstream ISS1(Line);
     std::string Marker;
@@ -132,15 +132,16 @@ size_t Merger::ApproximateMemoryConsumption() const  {
 
 // Decides which files need to be merged (add those to NewFiles).
 // Returns the number of new features added.
-size_t Merger::Merge(const Set<uint32_t> &InitialFeatures,
-                     Set<uint32_t> *NewFeatures,
-                     const Set<uint32_t> &InitialCov, Set<uint32_t> *NewCov,
-                     Vector<std::string> *NewFiles) {
+size_t Merger::Merge(const std::set<uint32_t> &InitialFeatures,
+                     std::set<uint32_t> *NewFeatures,
+                     const std::set<uint32_t> &InitialCov,
+                     std::set<uint32_t> *NewCov,
+                     std::vector<std::string> *NewFiles) {
   NewFiles->clear();
   NewFeatures->clear();
   NewCov->clear();
   assert(NumFilesInFirstCorpus <= Files.size());
-  Set<uint32_t> AllFeatures = InitialFeatures;
+  std::set<uint32_t> AllFeatures = InitialFeatures;
 
   // What features are in the initial corpus?
   for (size_t i = 0; i < NumFilesInFirstCorpus; i++) {
@@ -150,7 +151,7 @@ size_t Merger::Merge(const Set<uint32_t> &InitialFeatures,
   // Remove all features that we already know from all other inputs.
   for (size_t i = NumFilesInFirstCorpus; i < Files.size(); i++) {
     auto &Cur = Files[i].Features;
-    Vector<uint32_t> Tmp;
+    std::vector<uint32_t> Tmp;
     std::set_difference(Cur.begin(), Cur.end(), AllFeatures.begin(),
                         AllFeatures.end(), std::inserter(Tmp, Tmp.begin()));
     Cur.swap(Tmp);
@@ -188,15 +189,16 @@ size_t Merger::Merge(const Set<uint32_t> &InitialFeatures,
   return NewFeatures->size();
 }
 
-Set<uint32_t> Merger::AllFeatures() const {
-  Set<uint32_t> S;
+std::set<uint32_t> Merger::AllFeatures() const {
+  std::set<uint32_t> S;
   for (auto &File : Files)
     S.insert(File.Features.begin(), File.Features.end());
   return S;
 }
 
 // Inner process. May crash if the target crashes.
-void Fuzzer::CrashResistantMergeInternalStep(const std::string &CFPath) {
+void Fuzzer::CrashResistantMergeInternalStep(const std::string &CFPath,
+                                             bool IsSetCoverMerge) {
   Printf("MERGE-INNER: using the control file '%s'\n", CFPath.c_str());
   Merger M;
   std::ifstream IF(CFPath);
@@ -212,11 +214,11 @@ void Fuzzer::CrashResistantMergeInternalStep(const std::string &CFPath) {
          M.Files.size() - M.FirstNotProcessedFile);
 
   std::ofstream OF(CFPath, std::ofstream::out | std::ofstream::app);
-  Set<size_t> AllFeatures;
+  std::set<size_t> AllFeatures;
   auto PrintStatsWrapper = [this, &AllFeatures](const char* Where) {
     this->PrintStats(Where, "\n", 0, AllFeatures.size());
   };
-  Set<const TracePC::PCTableEntry *> AllPCs;
+  std::set<const TracePC::PCTableEntry *> AllPCs;
   for (size_t i = M.FirstNotProcessedFile; i < M.Files.size(); i++) {
     Fuzzer::MaybeExitGracefully();
     auto U = FileToVector(M.Files[i].Name);
@@ -234,13 +236,14 @@ void Fuzzer::CrashResistantMergeInternalStep(const std::string &CFPath) {
     // Collect coverage. We are iterating over the files in this order:
     // * First, files in the initial corpus ordered by size, smallest first.
     // * Then, all other files, smallest first.
-    // So it makes no sense to record all features for all files, instead we
-    // only record features that were not seen before.
-    Set<size_t> UniqFeatures;
-    TPC.CollectFeatures([&](size_t Feature) {
-      if (AllFeatures.insert(Feature).second)
-        UniqFeatures.insert(Feature);
-    });
+    std::set<size_t> Features;
+    if (IsSetCoverMerge)
+      TPC.CollectFeatures([&](size_t Feature) { Features.insert(Feature); });
+    else
+      TPC.CollectFeatures([&](size_t Feature) {
+        if (AllFeatures.insert(Feature).second)
+          Features.insert(Feature);
+      });
     TPC.UpdateObservedPCs();
     // Show stats.
     if (!(TotalNumberOfRuns & (TotalNumberOfRuns - 1)))
@@ -249,7 +252,7 @@ void Fuzzer::CrashResistantMergeInternalStep(const std::string &CFPath) {
       PrintStatsWrapper("LOADED");
     // Write the post-run marker and the coverage.
     OF << "FT " << i;
-    for (size_t F : UniqFeatures)
+    for (size_t F : Features)
       OF << " " << F;
     OF << "\n";
     OF << "COV " << i;
@@ -263,15 +266,137 @@ void Fuzzer::CrashResistantMergeInternalStep(const std::string &CFPath) {
   PrintStatsWrapper("DONE  ");
 }
 
-static size_t WriteNewControlFile(const std::string &CFPath,
-                                  const Vector<SizedFile> &OldCorpus,
-                                  const Vector<SizedFile> &NewCorpus,
-                                  const Vector<MergeFileInfo> &KnownFiles) {
+// Merges all corpora into the first corpus. A file is added into
+// the first corpus only if it adds new features. Unlike `Merger::Merge`,
+// this implementation calculates an approximation of the minimum set
+// of corpora files, that cover all known features (set cover problem).
+// Generally, this means that files with more features are preferred for
+// merge into the first corpus. When two files have the same number of
+// features, the smaller one is preferred.
+size_t Merger::SetCoverMerge(const std::set<uint32_t> &InitialFeatures,
+                             std::set<uint32_t> *NewFeatures,
+                             const std::set<uint32_t> &InitialCov,
+                             std::set<uint32_t> *NewCov,
+                             std::vector<std::string> *NewFiles) {
+  assert(NumFilesInFirstCorpus <= Files.size());
+  NewFiles->clear();
+  NewFeatures->clear();
+  NewCov->clear();
+  std::set<uint32_t> AllFeatures;
+  // 1 << 21 - 1 is the maximum feature index.
+  // See 'kFeatureSetSize' in 'FuzzerCorpus.h'.
+  const uint32_t kFeatureSetSize = 1 << 21;
+  std::vector<bool> Covered(kFeatureSetSize, false);
+  size_t NumCovered = 0;
+
+  std::set<uint32_t> ExistingFeatures = InitialFeatures;
+  for (size_t i = 0; i < NumFilesInFirstCorpus; ++i)
+    ExistingFeatures.insert(Files[i].Features.begin(), Files[i].Features.end());
+
+  // Mark the existing features as covered.
+  for (const auto &F : ExistingFeatures) {
+    if (!Covered[F % kFeatureSetSize]) {
+      ++NumCovered;
+      Covered[F % kFeatureSetSize] = true;
+    }
+    // Calculate an underestimation of the set of covered features
+    // since the `Covered` bitvector is smaller than the feature range.
+    AllFeatures.insert(F % kFeatureSetSize);
+  }
+
+  std::set<size_t> RemainingFiles;
+  for (size_t i = NumFilesInFirstCorpus; i < Files.size(); ++i) {
+    // Construct an incremental sequence which represent the
+    // indices to all files (excluding those in the initial corpus).
+    // RemainingFiles = range(NumFilesInFirstCorpus..Files.size()).
+    RemainingFiles.insert(i);
+    // Insert this file's unique features to all features.
+    for (const auto &F : Files[i].Features)
+      AllFeatures.insert(F % kFeatureSetSize);
+  }
+
+  // Integrate files into Covered until set is complete.
+  while (NumCovered != AllFeatures.size()) {
+    // Index to file with largest number of unique features.
+    size_t MaxFeaturesIndex = NumFilesInFirstCorpus;
+    // Indices to remove from RemainingFiles.
+    std::set<size_t> RemoveIndices;
+    // Running max unique feature count.
+    // Updated upon finding a file with more features.
+    size_t MaxNumFeatures = 0;
+
+    // Iterate over all files not yet integrated into Covered,
+    // to find the file which has the largest number of
+    // features that are not already in Covered.
+    for (const auto &i : RemainingFiles) {
+      const auto &File = Files[i];
+      size_t CurrentUnique = 0;
+      // Count number of features in this file
+      // which are not yet in Covered.
+      for (const auto &F : File.Features)
+        if (!Covered[F % kFeatureSetSize])
+          ++CurrentUnique;
+
+      if (CurrentUnique == 0) {
+        // All features in this file are already in Covered: skip next time.
+        RemoveIndices.insert(i);
+      } else if (CurrentUnique > MaxNumFeatures ||
+                 (CurrentUnique == MaxNumFeatures &&
+                  File.Size < Files[MaxFeaturesIndex].Size)) {
+        // Update the max features file based on unique features
+        // Break ties by selecting smaller files.
+        MaxNumFeatures = CurrentUnique;
+        MaxFeaturesIndex = i;
+      }
+    }
+    // Must be a valid index/
+    assert(MaxFeaturesIndex < Files.size());
+    // Remove any feature-less files found.
+    for (const auto &i : RemoveIndices)
+      RemainingFiles.erase(i);
+    if (MaxNumFeatures == 0) {
+      // Did not find a file that adds unique features.
+      // This means that we should have no remaining files.
+      assert(RemainingFiles.size() == 0);
+      assert(NumCovered == AllFeatures.size());
+      break;
+    }
+
+    // MaxFeaturesIndex must be an element of Remaining.
+    assert(RemainingFiles.find(MaxFeaturesIndex) != RemainingFiles.end());
+    // Remove the file with the most features from Remaining.
+    RemainingFiles.erase(MaxFeaturesIndex);
+    const auto &MaxFeatureFile = Files[MaxFeaturesIndex];
+    // Add the features of the max feature file to Covered.
+    for (const auto &F : MaxFeatureFile.Features) {
+      if (!Covered[F % kFeatureSetSize]) {
+        ++NumCovered;
+        Covered[F % kFeatureSetSize] = true;
+        NewFeatures->insert(F);
+      }
+    }
+    // Add the index to this file to the result.
+    NewFiles->push_back(MaxFeatureFile.Name);
+    // Update NewCov with the additional coverage
+    // that MaxFeatureFile provides.
+    for (const auto &C : MaxFeatureFile.Cov)
+      if (InitialCov.find(C) == InitialCov.end())
+        NewCov->insert(C);
+  }
+
+  return NewFeatures->size();
+}
+
+static size_t
+WriteNewControlFile(const std::string &CFPath,
+                    const std::vector<SizedFile> &OldCorpus,
+                    const std::vector<SizedFile> &NewCorpus,
+                    const std::vector<MergeFileInfo> &KnownFiles) {
   std::unordered_set<std::string> FilesToSkip;
   for (auto &SF: KnownFiles)
     FilesToSkip.insert(SF.Name);
 
-  Vector<std::string> FilesToUse;
+  std::vector<std::string> FilesToUse;
   auto MaybeUseFile = [=, &FilesToUse](std::string Name) {
     if (FilesToSkip.find(Name) == FilesToSkip.end())
       FilesToUse.push_back(Name);
@@ -299,19 +424,19 @@ static size_t WriteNewControlFile(const std::string &CFPath,
 }
 
 // Outer process. Does not call the target code and thus should not fail.
-void CrashResistantMerge(const Vector<std::string> &Args,
-                         const Vector<SizedFile> &OldCorpus,
-                         const Vector<SizedFile> &NewCorpus,
-                         Vector<std::string> *NewFiles,
-                         const Set<uint32_t> &InitialFeatures,
-                         Set<uint32_t> *NewFeatures,
-                         const Set<uint32_t> &InitialCov,
-                         Set<uint32_t> *NewCov,
-                         const std::string &CFPath,
-                         bool V /*Verbose*/) {
+void CrashResistantMerge(const std::vector<std::string> &Args,
+                         const std::vector<SizedFile> &OldCorpus,
+                         const std::vector<SizedFile> &NewCorpus,
+                         std::vector<std::string> *NewFiles,
+                         const std::set<uint32_t> &InitialFeatures,
+                         std::set<uint32_t> *NewFeatures,
+                         const std::set<uint32_t> &InitialCov,
+                         std::set<uint32_t> *NewCov, const std::string &CFPath,
+                         bool V, /*Verbose*/
+                         bool IsSetCoverMerge) {
   if (NewCorpus.empty() && OldCorpus.empty()) return;  // Nothing to merge.
   size_t NumAttempts = 0;
-  Vector<MergeFileInfo> KnownFiles;
+  std::vector<MergeFileInfo> KnownFiles;
   if (FileSize(CFPath)) {
     VPrintf(V, "MERGE-OUTER: non-empty control file provided: '%s'\n",
            CFPath.c_str());
@@ -363,6 +488,7 @@ void CrashResistantMerge(const Vector<std::string> &Args,
   // Every inner process should execute at least one input.
   Command BaseCmd(Args);
   BaseCmd.removeFlag("merge");
+  BaseCmd.removeFlag("set_cover_merge");
   BaseCmd.removeFlag("fork");
   BaseCmd.removeFlag("collect_data_flow");
   for (size_t Attempt = 1; Attempt <= NumAttempts; Attempt++) {
@@ -370,14 +496,16 @@ void CrashResistantMerge(const Vector<std::string> &Args,
     VPrintf(V, "MERGE-OUTER: attempt %zd\n", Attempt);
     Command Cmd(BaseCmd);
     Cmd.addFlag("merge_control_file", CFPath);
-    Cmd.addFlag("merge_inner", "1");
+    // If we are going to use the set cover implementation for
+    // minimization add the merge_inner=2 internal flag.
+    Cmd.addFlag("merge_inner", IsSetCoverMerge ? "2" : "1");
     if (!V) {
       Cmd.setOutputFile(getDevNull());
       Cmd.combineOutAndErr();
     }
     auto ExitCode = ExecuteCommand(Cmd);
     if (!ExitCode) {
-      VPrintf(V, "MERGE-OUTER: succesfull in %zd attempt(s)\n", Attempt);
+      VPrintf(V, "MERGE-OUTER: successful in %zd attempt(s)\n", Attempt);
       break;
     }
   }
@@ -395,7 +523,10 @@ void CrashResistantMerge(const Vector<std::string> &Args,
           M.ApproximateMemoryConsumption() >> 20, GetPeakRSSMb());
 
   M.Files.insert(M.Files.end(), KnownFiles.begin(), KnownFiles.end());
-  M.Merge(InitialFeatures, NewFeatures, InitialCov, NewCov, NewFiles);
+  if (IsSetCoverMerge)
+    M.SetCoverMerge(InitialFeatures, NewFeatures, InitialCov, NewCov, NewFiles);
+  else
+    M.Merge(InitialFeatures, NewFeatures, InitialCov, NewCov, NewFiles);
   VPrintf(V, "MERGE-OUTER: %zd new files with %zd new features added; "
           "%zd new coverage edges\n",
          NewFiles->size(), NewFeatures->size(), NewCov->size());
diff --git a/compiler-rt/lib/fuzzer/FuzzerMerge.h b/compiler-rt/lib/fuzzer/FuzzerMerge.h
index e0c6bc539bdb..42f798e1da18 100644
--- a/compiler-rt/lib/fuzzer/FuzzerMerge.h
+++ b/compiler-rt/lib/fuzzer/FuzzerMerge.h
@@ -41,6 +41,7 @@
 #define LLVM_FUZZER_MERGE_H
 
 #include "FuzzerDefs.h"
+#include "FuzzerIO.h"
 
 #include <istream>
 #include <ostream>
@@ -52,11 +53,11 @@ namespace fuzzer {
 struct MergeFileInfo {
   std::string Name;
   size_t Size = 0;
-  Vector<uint32_t> Features, Cov;
+  std::vector<uint32_t> Features, Cov;
 };
 
 struct Merger {
-  Vector<MergeFileInfo> Files;
+  std::vector<MergeFileInfo> Files;
   size_t NumFilesInFirstCorpus = 0;
   size_t FirstNotProcessedFile = 0;
   std::string LastFailure;
@@ -64,23 +65,28 @@ struct Merger {
   bool Parse(std::istream &IS, bool ParseCoverage);
   bool Parse(const std::string &Str, bool ParseCoverage);
   void ParseOrExit(std::istream &IS, bool ParseCoverage);
-  size_t Merge(const Set<uint32_t> &InitialFeatures, Set<uint32_t> *NewFeatures,
-               const Set<uint32_t> &InitialCov, Set<uint32_t> *NewCov,
-               Vector<std::string> *NewFiles);
+  size_t Merge(const std::set<uint32_t> &InitialFeatures,
+               std::set<uint32_t> *NewFeatures,
+               const std::set<uint32_t> &InitialCov, std::set<uint32_t> *NewCov,
+               std::vector<std::string> *NewFiles);
+  size_t SetCoverMerge(const std::set<uint32_t> &InitialFeatures,
+                       std::set<uint32_t> *NewFeatures,
+                       const std::set<uint32_t> &InitialCov,
+                       std::set<uint32_t> *NewCov,
+                       std::vector<std::string> *NewFiles);
   size_t ApproximateMemoryConsumption() const;
-  Set<uint32_t> AllFeatures() const;
+  std::set<uint32_t> AllFeatures() const;
 };
 
-void CrashResistantMerge(const Vector<std::string> &Args,
-                         const Vector<SizedFile> &OldCorpus,
-                         const Vector<SizedFile> &NewCorpus,
-                         Vector<std::string> *NewFiles,
-                         const Set<uint32_t> &InitialFeatures,
-                         Set<uint32_t> *NewFeatures,
-                         const Set<uint32_t> &InitialCov,
-                         Set<uint32_t> *NewCov,
-                         const std::string &CFPath,
-                         bool Verbose);
+void CrashResistantMerge(const std::vector<std::string> &Args,
+                         const std::vector<SizedFile> &OldCorpus,
+                         const std::vector<SizedFile> &NewCorpus,
+                         std::vector<std::string> *NewFiles,
+                         const std::set<uint32_t> &InitialFeatures,
+                         std::set<uint32_t> *NewFeatures,
+                         const std::set<uint32_t> &InitialCov,
+                         std::set<uint32_t> *NewCov, const std::string &CFPath,
+                         bool Verbose, bool IsSetCoverMerge);
 
 }  // namespace fuzzer
 
diff --git a/compiler-rt/lib/fuzzer/FuzzerMutate.cpp b/compiler-rt/lib/fuzzer/FuzzerMutate.cpp
index 4650f1beceac..d663900fdc3a 100644
--- a/compiler-rt/lib/fuzzer/FuzzerMutate.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerMutate.cpp
@@ -485,7 +485,7 @@ void MutationDispatcher::RecordSuccessfulMutationSequence() {
 }
 
 void MutationDispatcher::PrintRecommendedDictionary() {
-  Vector<DictionaryEntry> V;
+  std::vector<DictionaryEntry> V;
   for (auto &DE : PersistentAutoDictionary)
     if (!ManualDictionary.ContainsWord(DE.GetW()))
       V.push_back(DE);
@@ -540,7 +540,7 @@ size_t MutationDispatcher::DefaultMutate(uint8_t *Data, size_t Size,
 // Mutates Data in place, returns new size.
 size_t MutationDispatcher::MutateImpl(uint8_t *Data, size_t Size,
                                       size_t MaxSize,
-                                      Vector<Mutator> &Mutators) {
+                                      std::vector<Mutator> &Mutators) {
   assert(MaxSize > 0);
   // Some mutations may fail (e.g. can't insert more bytes if Size == MaxSize),
   // in which case they will return 0.
@@ -562,7 +562,7 @@ size_t MutationDispatcher::MutateImpl(uint8_t *Data, size_t Size,
 // Mask represents the set of Data bytes that are worth mutating.
 size_t MutationDispatcher::MutateWithMask(uint8_t *Data, size_t Size,
                                           size_t MaxSize,
-                                          const Vector<uint8_t> &Mask) {
+                                          const std::vector<uint8_t> &Mask) {
   size_t MaskedSize = std::min(Size, Mask.size());
   // * Copy the worthy bytes into a temporary array T
   // * Mutate T
diff --git a/compiler-rt/lib/fuzzer/FuzzerMutate.h b/compiler-rt/lib/fuzzer/FuzzerMutate.h
index fd37191156d3..97704e2160aa 100644
--- a/compiler-rt/lib/fuzzer/FuzzerMutate.h
+++ b/compiler-rt/lib/fuzzer/FuzzerMutate.h
@@ -77,7 +77,7 @@ public:
   /// that have '1' in Mask.
   /// Mask.size() should be >= Size.
   size_t MutateWithMask(uint8_t *Data, size_t Size, size_t MaxSize,
-                        const Vector<uint8_t> &Mask);
+                        const std::vector<uint8_t> &Mask);
 
   /// Applies one of the default mutations. Provided as a service
   /// to mutation authors.
@@ -104,7 +104,7 @@ public:
   size_t AddWordFromDictionary(Dictionary &D, uint8_t *Data, size_t Size,
                                size_t MaxSize);
   size_t MutateImpl(uint8_t *Data, size_t Size, size_t MaxSize,
-                    Vector<Mutator> &Mutators);
+                    std::vector<Mutator> &Mutators);
 
   size_t InsertPartOf(const uint8_t *From, size_t FromSize, uint8_t *To,
                       size_t ToSize, size_t MaxToSize);
@@ -133,22 +133,22 @@ public:
   // entries that led to successful discoveries in the past mutations.
   Dictionary PersistentAutoDictionary;
 
-  Vector<DictionaryEntry *> CurrentDictionaryEntrySequence;
+  std::vector<DictionaryEntry *> CurrentDictionaryEntrySequence;
 
   static const size_t kCmpDictionaryEntriesDequeSize = 16;
   DictionaryEntry CmpDictionaryEntriesDeque[kCmpDictionaryEntriesDequeSize];
   size_t CmpDictionaryEntriesDequeIdx = 0;
 
   const Unit *CrossOverWith = nullptr;
-  Vector<uint8_t> MutateInPlaceHere;
-  Vector<uint8_t> MutateWithMaskTemp;
+  std::vector<uint8_t> MutateInPlaceHere;
+  std::vector<uint8_t> MutateWithMaskTemp;
   // CustomCrossOver needs its own buffer as a custom implementation may call
   // LLVMFuzzerMutate, which in turn may resize MutateInPlaceHere.
-  Vector<uint8_t> CustomCrossOverInPlaceHere;
+  std::vector<uint8_t> CustomCrossOverInPlaceHere;
 
-  Vector<Mutator> Mutators;
-  Vector<Mutator> DefaultMutators;
-  Vector<Mutator> CurrentMutatorSequence;
+  std::vector<Mutator> Mutators;
+  std::vector<Mutator> DefaultMutators;
+  std::vector<Mutator> CurrentMutatorSequence;
 };
 
 }  // namespace fuzzer
diff --git a/compiler-rt/lib/fuzzer/FuzzerOptions.h b/compiler-rt/lib/fuzzer/FuzzerOptions.h
index d0c285a6821d..72e256106194 100644
--- a/compiler-rt/lib/fuzzer/FuzzerOptions.h
+++ b/compiler-rt/lib/fuzzer/FuzzerOptions.h
@@ -47,6 +47,7 @@ struct FuzzingOptions {
   int ReportSlowUnits = 10;
   bool OnlyASCII = false;
   bool Entropic = true;
+  bool ForkCorpusGroups = false;
   size_t EntropicFeatureFrequencyThreshold = 0xFF;
   size_t EntropicNumberOfRarestFeatures = 100;
   bool EntropicScalePerExecTime = false;
diff --git a/compiler-rt/lib/fuzzer/FuzzerTracePC.cpp b/compiler-rt/lib/fuzzer/FuzzerTracePC.cpp
index d808b9b00fa3..af8d1ce50f3f 100644
--- a/compiler-rt/lib/fuzzer/FuzzerTracePC.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerTracePC.cpp
@@ -157,7 +157,7 @@ ALWAYS_INLINE uintptr_t TracePC::GetNextInstructionPc(uintptr_t PC) {
 }
 
 void TracePC::UpdateObservedPCs() {
-  Vector<uintptr_t> CoveredFuncs;
+  std::vector<uintptr_t> CoveredFuncs;
   auto ObservePC = [&](const PCTableEntry *TE) {
     if (ObservedPCs.insert(TE).second && DoPrintNewPCs) {
       PrintPC("\tNEW_PC: %p %F %L", "\tNEW_PC: %p",
@@ -300,8 +300,8 @@ void TracePC::PrintCoverage(bool PrintAllCounters) {
       FunctionStr = FunctionStr.substr(3);
     std::string LineStr = DescribePC("%l", VisualizePC);
     size_t NumEdges = Last - First;
-    Vector<uintptr_t> UncoveredPCs;
-    Vector<uintptr_t> CoveredPCs;
+    std::vector<uintptr_t> UncoveredPCs;
+    std::vector<uintptr_t> CoveredPCs;
     for (auto TE = First; TE < Last; TE++)
       if (!ObservedPCs.count(TE))
         UncoveredPCs.push_back(TE->PC);
@@ -391,6 +391,7 @@ void TracePC::HandleCmp(uintptr_t PC, T Arg1, T Arg2) {
   ValueProfileMap.AddValue(PC * 128 + 64 + AbsoluteDistance);
 }
 
+ATTRIBUTE_NO_SANITIZE_MEMORY
 static size_t InternalStrnlen(const char *S, size_t MaxLen) {
   size_t Len = 0;
   for (; Len < MaxLen && S[Len]; Len++) {}
@@ -398,7 +399,8 @@ static size_t InternalStrnlen(const char *S, size_t MaxLen) {
 }
 
 // Finds min of (strlen(S1), strlen(S2)).
-// Needed bacause one of these strings may actually be non-zero terminated.
+// Needed because one of these strings may actually be non-zero terminated.
+ATTRIBUTE_NO_SANITIZE_MEMORY
 static size_t InternalStrnlen2(const char *S1, const char *S2) {
   size_t Len = 0;
   for (; S1[Len] && S2[Len]; Len++)  {}
diff --git a/compiler-rt/lib/fuzzer/FuzzerTracePC.h b/compiler-rt/lib/fuzzer/FuzzerTracePC.h
index a93732972f7d..af1f9d81e950 100644
--- a/compiler-rt/lib/fuzzer/FuzzerTracePC.h
+++ b/compiler-rt/lib/fuzzer/FuzzerTracePC.h
@@ -169,7 +169,7 @@ private:
   size_t NumPCTables;
   size_t NumPCsInPCTables;
 
-  Set<const PCTableEntry*> ObservedPCs;
+  std::set<const PCTableEntry *> ObservedPCs;
   std::unordered_map<uintptr_t, uintptr_t> ObservedFuncs;  // PC => Counter.
 
   uint8_t *FocusFunctionCounterPtr = nullptr;
diff --git a/compiler-rt/lib/fuzzer/FuzzerUtil.cpp b/compiler-rt/lib/fuzzer/FuzzerUtil.cpp
index 05185499bdd1..aeab70f20c28 100644
--- a/compiler-rt/lib/fuzzer/FuzzerUtil.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerUtil.cpp
@@ -43,7 +43,7 @@ void PrintASCIIByte(uint8_t Byte) {
   else if (Byte >= 32 && Byte < 127)
     Printf("%c", Byte);
   else
-    Printf("\\x%02x", Byte);
+    Printf("\\%03o", Byte);
 }
 
 void PrintASCII(const uint8_t *Data, size_t Size, const char *PrintAfter) {
@@ -124,7 +124,7 @@ bool ParseOneDictionaryEntry(const std::string &Str, Unit *U) {
   return true;
 }
 
-bool ParseDictionaryFile(const std::string &Text, Vector<Unit> *Units) {
+bool ParseDictionaryFile(const std::string &Text, std::vector<Unit> *Units) {
   if (Text.empty()) {
     Printf("ParseDictionaryFile: file does not exist or is empty\n");
     return false;
diff --git a/compiler-rt/lib/fuzzer/FuzzerUtil.h b/compiler-rt/lib/fuzzer/FuzzerUtil.h
index a188a7be32a5..71d49097e559 100644
--- a/compiler-rt/lib/fuzzer/FuzzerUtil.h
+++ b/compiler-rt/lib/fuzzer/FuzzerUtil.h
@@ -66,10 +66,10 @@ int CloseProcessPipe(FILE *F);
 const void *SearchMemory(const void *haystack, size_t haystacklen,
                          const void *needle, size_t needlelen);
 
-std::string CloneArgsWithoutX(const Vector<std::string> &Args,
+std::string CloneArgsWithoutX(const std::vector<std::string> &Args,
                               const char *X1, const char *X2);
 
-inline std::string CloneArgsWithoutX(const Vector<std::string> &Args,
+inline std::string CloneArgsWithoutX(const std::vector<std::string> &Args,
                                      const char *X) {
   return CloneArgsWithoutX(Args, X, X);
 }
diff --git a/compiler-rt/lib/fuzzer/FuzzerUtilFuchsia.cpp b/compiler-rt/lib/fuzzer/FuzzerUtilFuchsia.cpp
index 5034b4a28d3f..d80b80cccb80 100644
--- a/compiler-rt/lib/fuzzer/FuzzerUtilFuchsia.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerUtilFuchsia.cpp
@@ -52,6 +52,12 @@ void CrashTrampolineAsm() __asm__("CrashTrampolineAsm");
 
 namespace {
 
+// The signal handler thread uses Zircon exceptions to resume crashed threads
+// into libFuzzer's POSIX signal handlers. The associated event is used to
+// signal when the thread is running, and when it should stop.
+std::thread SignalHandler;
+zx_handle_t SignalHandlerEvent = ZX_HANDLE_INVALID;
+
 // Helper function to handle Zircon syscall failures.
 void ExitOnErr(zx_status_t Status, const char *Syscall) {
   if (Status != ZX_OK) {
@@ -68,23 +74,6 @@ void AlarmHandler(int Seconds) {
   }
 }
 
-// CFAOffset is used to reference the stack pointer before entering the
-// trampoline (Stack Pointer + CFAOffset = prev Stack Pointer). Before jumping
-// to the trampoline we copy all the registers onto the stack. We need to make
-// sure that the new stack has enough space to store all the registers.
-//
-// The trampoline holds CFI information regarding the registers stored in the
-// stack, which is then used by the unwinder to restore them.
-#if defined(__x86_64__)
-// In x86_64 the crashing function might also be using the red zone (128 bytes
-// on top of their rsp).
-constexpr size_t CFAOffset = 128 + sizeof(zx_thread_state_general_regs_t);
-#elif defined(__aarch64__)
-// In aarch64 we need to always have the stack pointer aligned to 16 bytes, so we
-// make sure that we are keeping that same alignment.
-constexpr size_t CFAOffset = (sizeof(zx_thread_state_general_regs_t) + 15) & -(uintptr_t)16;
-#endif
-
 // For the crash handler, we need to call Fuzzer::StaticCrashSignalCallback
 // without POSIX signal handlers.  To achieve this, we use an assembly function
 // to add the necessary CFI unwinding information and a C function to bridge
@@ -163,10 +152,10 @@ constexpr size_t CFAOffset = (sizeof(zx_thread_state_general_regs_t) + 15) & -(u
 
 // Produces an assembler immediate operand for the named or numbered register.
 // This operand contains the offset of the register relative to the CFA.
-#define ASM_OPERAND_REG(reg) \
-  [reg] "i"(offsetof(zx_thread_state_general_regs_t, reg) - CFAOffset),
-#define ASM_OPERAND_NUM(num)                                 \
-  [x##num] "i"(offsetof(zx_thread_state_general_regs_t, r[num]) - CFAOffset),
+#define ASM_OPERAND_REG(reg)                                                   \
+  [reg] "i"(offsetof(zx_thread_state_general_regs_t, reg)),
+#define ASM_OPERAND_NUM(num)                                                   \
+  [x##num] "i"(offsetof(zx_thread_state_general_regs_t, r[num])),
 
 // Trampoline to bridge from the assembly below to the static C++ crash
 // callback.
@@ -178,62 +167,57 @@ static void StaticCrashHandler() {
   }
 }
 
-// Creates the trampoline with the necessary CFI information to unwind through
-// to the crashing call stack:
-//  * Defining the CFA so that it points to the stack pointer at the point
-//    of crash.
-//  * Storing all registers at the point of crash in the stack and refer to them
-//    via CFI information (relative to the CFA).
-//  * Setting the return column so the unwinder knows how to continue unwinding.
-//  * (x86_64) making sure rsp is aligned before calling StaticCrashHandler.
-//  * Calling StaticCrashHandler that will trigger the unwinder.
+// This trampoline function has the necessary CFI information to unwind
+// and get a backtrace:
+//  * The stack contains a copy of all the registers at the point of crash,
+//    the code has CFI directives specifying how to restore them.
+//  * A call to StaticCrashHandler, which will print the stacktrace and exit
+//    the fuzzer, generating a crash artifact.
 //
 // The __attribute__((used)) is necessary because the function
 // is never called; it's just a container around the assembly to allow it to
 // use operands for compile-time computed constants.
 __attribute__((used))
 void MakeTrampoline() {
-  __asm__(".cfi_endproc\n"
-    ".pushsection .text.CrashTrampolineAsm\n"
-    ".type CrashTrampolineAsm,STT_FUNC\n"
-"CrashTrampolineAsm:\n"
-    ".cfi_startproc simple\n"
-    ".cfi_signal_frame\n"
+  __asm__(
+      ".cfi_endproc\n"
+      ".pushsection .text.CrashTrampolineAsm\n"
+      ".type CrashTrampolineAsm,STT_FUNC\n"
+      "CrashTrampolineAsm:\n"
+      ".cfi_startproc simple\n"
+      ".cfi_signal_frame\n"
 #if defined(__x86_64__)
-    ".cfi_return_column rip\n"
-    ".cfi_def_cfa rsp, %c[CFAOffset]\n"
-    FOREACH_REGISTER(CFI_OFFSET_REG, CFI_OFFSET_NUM)
-    "mov %%rsp, %%rbp\n"
-    ".cfi_def_cfa_register rbp\n"
-    "andq $-16, %%rsp\n"
-    "call %c[StaticCrashHandler]\n"
-    "ud2\n"
+      ".cfi_return_column rip\n"
+      ".cfi_def_cfa rsp, 0\n"
+      FOREACH_REGISTER(CFI_OFFSET_REG, CFI_OFFSET_NUM)
+      "call %c[StaticCrashHandler]\n"
+      "ud2\n"
 #elif defined(__aarch64__)
-    ".cfi_return_column 33\n"
-    ".cfi_def_cfa sp, %c[CFAOffset]\n"
-    FOREACH_REGISTER(CFI_OFFSET_REG, CFI_OFFSET_NUM)
-    ".cfi_offset 33, %c[pc]\n"
-    ".cfi_offset 30, %c[lr]\n"
-    "bl %c[StaticCrashHandler]\n"
-    "brk 1\n"
+      ".cfi_return_column 33\n"
+      ".cfi_def_cfa sp, 0\n"
+      FOREACH_REGISTER(CFI_OFFSET_REG, CFI_OFFSET_NUM)
+      ".cfi_offset 33, %c[pc]\n"
+      ".cfi_offset 30, %c[lr]\n"
+      "bl %c[StaticCrashHandler]\n"
+      "brk 1\n"
 #else
 #error "Unsupported architecture for fuzzing on Fuchsia"
 #endif
-    ".cfi_endproc\n"
-    ".size CrashTrampolineAsm, . - CrashTrampolineAsm\n"
-    ".popsection\n"
-    ".cfi_startproc\n"
-    : // No outputs
-    : FOREACH_REGISTER(ASM_OPERAND_REG, ASM_OPERAND_NUM)
+     ".cfi_endproc\n"
+     ".size CrashTrampolineAsm, . - CrashTrampolineAsm\n"
+     ".popsection\n"
+     ".cfi_startproc\n"
+      : // No outputs
+      : FOREACH_REGISTER(ASM_OPERAND_REG, ASM_OPERAND_NUM)
 #if defined(__aarch64__)
-      ASM_OPERAND_REG(pc)
-      ASM_OPERAND_REG(lr)
+        ASM_OPERAND_REG(pc) ASM_OPERAND_REG(lr)
 #endif
-      [StaticCrashHandler] "i" (StaticCrashHandler),
-      [CFAOffset] "i" (CFAOffset));
+        [StaticCrashHandler] "i"(StaticCrashHandler));
 }
 
-void CrashHandler(zx_handle_t *Event) {
+void CrashHandler() {
+  assert(SignalHandlerEvent != ZX_HANDLE_INVALID);
+
   // This structure is used to ensure we close handles to objects we create in
   // this handler.
   struct ScopedHandle {
@@ -251,16 +235,30 @@ void CrashHandler(zx_handle_t *Event) {
                 Self, ZX_EXCEPTION_CHANNEL_DEBUGGER, &Channel.Handle),
             "_zx_task_create_exception_channel");
 
-  ExitOnErr(_zx_object_signal(*Event, 0, ZX_USER_SIGNAL_0),
+  ExitOnErr(_zx_object_signal(SignalHandlerEvent, 0, ZX_USER_SIGNAL_0),
             "_zx_object_signal");
 
   // This thread lives as long as the process in order to keep handling
   // crashes.  In practice, the first crashed thread to reach the end of the
   // StaticCrashHandler will end the process.
   while (true) {
-    ExitOnErr(_zx_object_wait_one(Channel.Handle, ZX_CHANNEL_READABLE,
-                                  ZX_TIME_INFINITE, nullptr),
-              "_zx_object_wait_one");
+    zx_wait_item_t WaitItems[] = {
+        {
+            .handle = SignalHandlerEvent,
+            .waitfor = ZX_SIGNAL_HANDLE_CLOSED,
+            .pending = 0,
+        },
+        {
+            .handle = Channel.Handle,
+            .waitfor = ZX_CHANNEL_READABLE | ZX_CHANNEL_PEER_CLOSED,
+            .pending = 0,
+        },
+    };
+    auto Status = _zx_object_wait_many(
+        WaitItems, sizeof(WaitItems) / sizeof(WaitItems[0]), ZX_TIME_INFINITE);
+    if (Status != ZX_OK || (WaitItems[1].pending & ZX_CHANNEL_READABLE) == 0) {
+      break;
+    }
 
     zx_exception_info_t ExceptionInfo;
     ScopedHandle Exception;
@@ -296,14 +294,17 @@ void CrashHandler(zx_handle_t *Event) {
     // onto the stack and jump into a trampoline with CFI instructions on how
     // to restore it.
 #if defined(__x86_64__)
-    uintptr_t StackPtr = GeneralRegisters.rsp - CFAOffset;
+    uintptr_t StackPtr =
+        (GeneralRegisters.rsp - (128 + sizeof(GeneralRegisters))) &
+        -(uintptr_t)16;
     __unsanitized_memcpy(reinterpret_cast<void *>(StackPtr), &GeneralRegisters,
                          sizeof(GeneralRegisters));
     GeneralRegisters.rsp = StackPtr;
     GeneralRegisters.rip = reinterpret_cast<zx_vaddr_t>(CrashTrampolineAsm);
 
 #elif defined(__aarch64__)
-    uintptr_t StackPtr = GeneralRegisters.sp - CFAOffset;
+    uintptr_t StackPtr =
+        (GeneralRegisters.sp - sizeof(GeneralRegisters)) & -(uintptr_t)16;
     __unsanitized_memcpy(reinterpret_cast<void *>(StackPtr), &GeneralRegisters,
                          sizeof(GeneralRegisters));
     GeneralRegisters.sp = StackPtr;
@@ -327,6 +328,13 @@ void CrashHandler(zx_handle_t *Event) {
   }
 }
 
+void StopSignalHandler() {
+  _zx_handle_close(SignalHandlerEvent);
+  if (SignalHandler.joinable()) {
+    SignalHandler.join();
+  }
+}
+
 } // namespace
 
 // Platform specific functions.
@@ -356,16 +364,14 @@ void SetSignalHandler(const FuzzingOptions &Options) {
     return;
 
   // Set up the crash handler and wait until it is ready before proceeding.
-  zx_handle_t Event;
-  ExitOnErr(_zx_event_create(0, &Event), "_zx_event_create");
+  ExitOnErr(_zx_event_create(0, &SignalHandlerEvent), "_zx_event_create");
 
-  std::thread T(CrashHandler, &Event);
-  zx_status_t Status =
-      _zx_object_wait_one(Event, ZX_USER_SIGNAL_0, ZX_TIME_INFINITE, nullptr);
-  _zx_handle_close(Event);
+  SignalHandler = std::thread(CrashHandler);
+  zx_status_t Status = _zx_object_wait_one(SignalHandlerEvent, ZX_USER_SIGNAL_0,
+                                           ZX_TIME_INFINITE, nullptr);
   ExitOnErr(Status, "_zx_object_wait_one");
 
-  T.detach();
+  std::atexit(StopSignalHandler);
 }
 
 void SleepSeconds(int Seconds) {
diff --git a/compiler-rt/lib/fuzzer/FuzzerUtilWindows.cpp b/compiler-rt/lib/fuzzer/FuzzerUtilWindows.cpp
index 1a54bb569eca..3598758dbb4f 100644
--- a/compiler-rt/lib/fuzzer/FuzzerUtilWindows.cpp
+++ b/compiler-rt/lib/fuzzer/FuzzerUtilWindows.cpp
@@ -204,7 +204,7 @@ const void *SearchMemory(const void *Data, size_t DataLen, const void *Patt,
 }
 
 std::string DisassembleCmd(const std::string &FileName) {
-  Vector<std::string> command_vector;
+  std::vector<std::string> command_vector;
   command_vector.push_back("dumpbin /summary > nul");
   if (ExecuteCommand(Command(command_vector)) == 0)
     return "dumpbin /disasm " + FileName;
diff --git a/compiler-rt/lib/gwp_asan/common.h b/compiler-rt/lib/gwp_asan/common.h
index 7ce367e3ffe9..6b238ad9ecbd 100644
--- a/compiler-rt/lib/gwp_asan/common.h
+++ b/compiler-rt/lib/gwp_asan/common.h
@@ -19,7 +19,28 @@
 #include <stdint.h>
 
 namespace gwp_asan {
-enum class Error {
+
+// Magic header that resides in the AllocatorState so that GWP-ASan bugreports
+// can be understood by tools at different versions. Out-of-process crash
+// handlers, like crashpad on Fuchsia, take the raw contents of the
+// AllocationMetatada array and the AllocatorState, and shove them into the
+// minidump. Online unpacking of these structs needs to know from which version
+// of GWP-ASan it's extracting the information, as the structures are not
+// stable.
+struct AllocatorVersionMagic {
+  // The values are copied into the structure at runtime, during
+  // `GuardedPoolAllocator::init()` so that GWP-ASan remains completely in the
+  // `.bss` segment.
+  static constexpr uint8_t kAllocatorVersionMagic[4] = {'A', 'S', 'A', 'N'};
+  uint8_t Magic[4] = {};
+  // Update the version number when the AllocatorState or AllocationMetadata
+  // change.
+  static constexpr uint16_t kAllocatorVersion = 1;
+  uint16_t Version = 0;
+  uint16_t Reserved = 0;
+};
+
+enum class Error : uint8_t {
   UNKNOWN,
   USE_AFTER_FREE,
   DOUBLE_FREE,
@@ -84,6 +105,7 @@ struct AllocationMetadata {
 // set of information required for understanding a GWP-ASan crash.
 struct AllocatorState {
   constexpr AllocatorState() {}
+  AllocatorVersionMagic VersionMagic{};
 
   // Returns whether the provided pointer is a current sampled allocation that
   // is owned by this pool.
@@ -123,5 +145,38 @@ struct AllocatorState {
   uintptr_t FailureAddress = 0;
 };
 
+// Below are various compile-time checks that the layout of the internal
+// GWP-ASan structures are undisturbed. If they are disturbed, the version magic
+// number needs to be increased by one, and the asserts need to be updated.
+// Out-of-process crash handlers, like breakpad/crashpad, may copy the internal
+// GWP-ASan structures into a minidump for offline reconstruction of the crash.
+// In order to accomplish this, the offline reconstructor needs to know the
+// version of GWP-ASan internal structures that it's unpacking (along with the
+// architecture-specific layout info, which is left as an exercise to the crash
+// handler).
+static_assert(offsetof(AllocatorState, VersionMagic) == 0, "");
+static_assert(sizeof(AllocatorVersionMagic) == 8, "");
+#if defined(__x86_64__)
+static_assert(sizeof(AllocatorState) == 56, "");
+static_assert(offsetof(AllocatorState, FailureAddress) == 48, "");
+static_assert(sizeof(AllocationMetadata) == 568, "");
+static_assert(offsetof(AllocationMetadata, IsDeallocated) == 560, "");
+#elif defined(__aarch64__)
+static_assert(sizeof(AllocatorState) == 56, "");
+static_assert(offsetof(AllocatorState, FailureAddress) == 48, "");
+static_assert(sizeof(AllocationMetadata) == 568, "");
+static_assert(offsetof(AllocationMetadata, IsDeallocated) == 560, "");
+#elif defined(__i386__)
+static_assert(sizeof(AllocatorState) == 32, "");
+static_assert(offsetof(AllocatorState, FailureAddress) == 28, "");
+static_assert(sizeof(AllocationMetadata) == 548, "");
+static_assert(offsetof(AllocationMetadata, IsDeallocated) == 544, "");
+#elif defined(__arm__)
+static_assert(sizeof(AllocatorState) == 32, "");
+static_assert(offsetof(AllocatorState, FailureAddress) == 28, "");
+static_assert(sizeof(AllocationMetadata) == 560, "");
+static_assert(offsetof(AllocationMetadata, IsDeallocated) == 552, "");
+#endif // defined($ARCHITECTURE)
+
 } // namespace gwp_asan
 #endif // GWP_ASAN_COMMON_H_
diff --git a/compiler-rt/lib/gwp_asan/guarded_pool_allocator.cpp b/compiler-rt/lib/gwp_asan/guarded_pool_allocator.cpp
index 8ce5fc9c4dfc..7096b428764c 100644
--- a/compiler-rt/lib/gwp_asan/guarded_pool_allocator.cpp
+++ b/compiler-rt/lib/gwp_asan/guarded_pool_allocator.cpp
@@ -59,6 +59,13 @@ void GuardedPoolAllocator::init(const options::Options &Opts) {
   SingletonPtr = this;
   Backtrace = Opts.Backtrace;
 
+  State.VersionMagic = {{AllocatorVersionMagic::kAllocatorVersionMagic[0],
+                         AllocatorVersionMagic::kAllocatorVersionMagic[1],
+                         AllocatorVersionMagic::kAllocatorVersionMagic[2],
+                         AllocatorVersionMagic::kAllocatorVersionMagic[3]},
+                        AllocatorVersionMagic::kAllocatorVersion,
+                        0};
+
   State.MaxSimultaneousAllocations = Opts.MaxSimultaneousAllocations;
 
   const size_t PageSize = getPlatformPageSize();
diff --git a/compiler-rt/lib/hwasan/hwasan.cpp b/compiler-rt/lib/hwasan/hwasan.cpp
index cbe0dee66dcd..6f0ea64472c6 100644
--- a/compiler-rt/lib/hwasan/hwasan.cpp
+++ b/compiler-rt/lib/hwasan/hwasan.cpp
@@ -16,6 +16,7 @@
 #include "hwasan_checks.h"
 #include "hwasan_dynamic_shadow.h"
 #include "hwasan_globals.h"
+#include "hwasan_mapping.h"
 #include "hwasan_poisoning.h"
 #include "hwasan_report.h"
 #include "hwasan_thread.h"
@@ -141,7 +142,7 @@ static void CheckUnwind() {
 static void HwasanFormatMemoryUsage(InternalScopedString &s) {
   HwasanThreadList &thread_list = hwasanThreadList();
   auto thread_stats = thread_list.GetThreadStats();
-  auto *sds = StackDepotGetStats();
+  auto sds = StackDepotGetStats();
   AllocatorStatCounters asc;
   GetAllocatorStats(asc);
   s.append(
@@ -151,7 +152,7 @@ static void HwasanFormatMemoryUsage(InternalScopedString &s) {
       internal_getpid(), GetRSS(), thread_stats.n_live_threads,
       thread_stats.total_stack_size,
       thread_stats.n_live_threads * thread_list.MemoryUsedPerThread(),
-      sds->allocated, sds->n_uniq_ids, asc[AllocatorStatMapped]);
+      sds.allocated, sds.n_uniq_ids, asc[AllocatorStatMapped]);
 }
 
 #if SANITIZER_ANDROID
@@ -319,7 +320,7 @@ void __hwasan_init_static() {
     InitializeSingleGlobal(global);
 }
 
-void __hwasan_init() {
+__attribute__((constructor(0))) void __hwasan_init() {
   CHECK(!hwasan_init_is_running);
   if (hwasan_inited) return;
   hwasan_init_is_running = 1;
@@ -344,7 +345,7 @@ void __hwasan_init() {
 
   // Needs to be called here because flags()->random_tags might not have been
   // initialized when InitInstrumentation() was called.
-  GetCurrentThread()->InitRandomState();
+  GetCurrentThread()->EnsureRandomStateInited();
 
   SetPrintfAndReportCallback(AppendToErrorMessageBuffer);
   // This may call libc -> needs initialized shadow.
@@ -360,6 +361,7 @@ void __hwasan_init() {
   HwasanTSDThreadInit();
 
   HwasanAllocatorInit();
+  HwasanInstallAtForkHandler();
 
 #if HWASAN_CONTAINS_UBSAN
   __ubsan::InitAsPlugin();
@@ -390,8 +392,15 @@ void __hwasan_print_shadow(const void *p, uptr sz) {
   uptr shadow_last = MemToShadow(ptr_raw + sz - 1);
   Printf("HWASan shadow map for %zx .. %zx (pointer tag %x)\n", ptr_raw,
          ptr_raw + sz, GetTagFromPointer((uptr)p));
-  for (uptr s = shadow_first; s <= shadow_last; ++s)
-    Printf("  %zx: %x\n", ShadowToMem(s), *(tag_t *)s);
+  for (uptr s = shadow_first; s <= shadow_last; ++s) {
+    tag_t mem_tag = *reinterpret_cast<tag_t *>(s);
+    uptr granule_addr = ShadowToMem(s);
+    if (mem_tag && mem_tag < kShadowAlignment)
+      Printf("  %zx: %02x(%02x)\n", granule_addr, mem_tag,
+             *reinterpret_cast<tag_t *>(granule_addr + kShadowAlignment - 1));
+    else
+      Printf("  %zx: %02x\n", granule_addr, mem_tag);
+  }
 }
 
 sptr __hwasan_test_shadow(const void *p, uptr sz) {
diff --git a/compiler-rt/lib/hwasan/hwasan.h b/compiler-rt/lib/hwasan/hwasan.h
index 7338b696ad34..371c43f3cbde 100644
--- a/compiler-rt/lib/hwasan/hwasan.h
+++ b/compiler-rt/lib/hwasan/hwasan.h
@@ -107,6 +107,8 @@ void InitThreads();
 void InitializeInterceptors();
 
 void HwasanAllocatorInit();
+void HwasanAllocatorLock();
+void HwasanAllocatorUnlock();
 
 void *hwasan_malloc(uptr size, StackTrace *stack);
 void *hwasan_calloc(uptr nmemb, uptr size, StackTrace *stack);
@@ -140,6 +142,8 @@ void HwasanAtExit();
 
 void HwasanOnDeadlySignal(int signo, void *info, void *context);
 
+void HwasanInstallAtForkHandler();
+
 void UpdateMemoryUsage();
 
 void AppendToErrorMessageBuffer(const char *buffer);
@@ -183,25 +187,34 @@ void HwasanTagMismatch(uptr addr, uptr access_info, uptr *registers_frame,
     RunFreeHooks(ptr);            \
   } while (false)
 
-#if HWASAN_WITH_INTERCEPTORS && defined(__aarch64__)
+#if HWASAN_WITH_INTERCEPTORS
 // For both bionic and glibc __sigset_t is an unsigned long.
 typedef unsigned long __hw_sigset_t;
 // Setjmp and longjmp implementations are platform specific, and hence the
-// interception code is platform specific too.  As yet we've only implemented
-// the interception for AArch64.
-typedef unsigned long long __hw_register_buf[22];
+// interception code is platform specific too.
+#  if defined(__aarch64__)
+constexpr size_t kHwRegisterBufSize = 22;
+#  elif defined(__x86_64__)
+constexpr size_t kHwRegisterBufSize = 8;
+#  endif
+typedef unsigned long long __hw_register_buf[kHwRegisterBufSize];
 struct __hw_jmp_buf_struct {
   // NOTE: The machine-dependent definition of `__sigsetjmp'
   // assume that a `__hw_jmp_buf' begins with a `__hw_register_buf' and that
   // `__mask_was_saved' follows it.  Do not move these members or add others
   // before it.
+  //
+  // We add a __magic field to our struct to catch cases where libc's setjmp
+  // populated the jmp_buf instead of our interceptor.
   __hw_register_buf __jmpbuf; // Calling environment.
-  int __mask_was_saved;       // Saved the signal mask?
+  unsigned __mask_was_saved : 1;  // Saved the signal mask?
+  unsigned __magic : 31;      // Used to distinguish __hw_jmp_buf from jmp_buf.
   __hw_sigset_t __saved_mask; // Saved signal mask.
 };
 typedef struct __hw_jmp_buf_struct __hw_jmp_buf[1];
 typedef struct __hw_jmp_buf_struct __hw_sigjmp_buf[1];
-#endif // HWASAN_WITH_INTERCEPTORS && __aarch64__
+constexpr unsigned kHwJmpBufMagic = 0x248ACE77;
+#endif  // HWASAN_WITH_INTERCEPTORS
 
 #define ENSURE_HWASAN_INITED()      \
   do {                              \
diff --git a/compiler-rt/lib/hwasan/hwasan_allocation_functions.cpp b/compiler-rt/lib/hwasan/hwasan_allocation_functions.cpp
index 6c2a6077866f..9cd82dbabd19 100644
--- a/compiler-rt/lib/hwasan/hwasan_allocation_functions.cpp
+++ b/compiler-rt/lib/hwasan/hwasan_allocation_functions.cpp
@@ -14,28 +14,21 @@
 
 #include "hwasan.h"
 #include "interception/interception.h"
+#include "sanitizer_common/sanitizer_allocator_dlsym.h"
 #include "sanitizer_common/sanitizer_allocator_interface.h"
 #include "sanitizer_common/sanitizer_tls_get_addr.h"
 
-using namespace __hwasan;
+#if !SANITIZER_FUCHSIA
 
-static uptr allocated_for_dlsym;
-static const uptr kDlsymAllocPoolSize = 1024;
-static uptr alloc_memory_for_dlsym[kDlsymAllocPoolSize];
+using namespace __hwasan;
 
-static bool IsInDlsymAllocPool(const void *ptr) {
-  uptr off = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-  return off < sizeof(alloc_memory_for_dlsym);
-}
+struct DlsymAlloc : public DlSymAllocator<DlsymAlloc> {
+  static bool UseImpl() { return !hwasan_inited; }
+};
 
-static void *AllocateFromLocalPool(uptr size_in_bytes) {
-  uptr size_in_words = RoundUpTo(size_in_bytes, kWordSize) / kWordSize;
-  void *mem = (void *)&alloc_memory_for_dlsym[allocated_for_dlsym];
-  allocated_for_dlsym += size_in_words;
-  CHECK_LT(allocated_for_dlsym, kDlsymAllocPoolSize);
-  return mem;
-}
+extern "C" {
 
+SANITIZER_INTERFACE_ATTRIBUTE
 int __sanitizer_posix_memalign(void **memptr, uptr alignment, uptr size) {
   GET_MALLOC_STACK_TRACE;
   CHECK_NE(memptr, 0);
@@ -43,16 +36,19 @@ int __sanitizer_posix_memalign(void **memptr, uptr alignment, uptr size) {
   return res;
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer_memalign(uptr alignment, uptr size) {
   GET_MALLOC_STACK_TRACE;
   return hwasan_memalign(alignment, size, &stack);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer_aligned_alloc(uptr alignment, uptr size) {
   GET_MALLOC_STACK_TRACE;
   return hwasan_aligned_alloc(alignment, size, &stack);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer___libc_memalign(uptr alignment, uptr size) {
   GET_MALLOC_STACK_TRACE;
   void *ptr = hwasan_memalign(alignment, size, &stack);
@@ -61,87 +57,92 @@ void *__sanitizer___libc_memalign(uptr alignment, uptr size) {
   return ptr;
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer_valloc(uptr size) {
   GET_MALLOC_STACK_TRACE;
   return hwasan_valloc(size, &stack);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer_pvalloc(uptr size) {
   GET_MALLOC_STACK_TRACE;
   return hwasan_pvalloc(size, &stack);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void __sanitizer_free(void *ptr) {
-  GET_MALLOC_STACK_TRACE;
-  if (!ptr || UNLIKELY(IsInDlsymAllocPool(ptr)))
+  if (!ptr)
     return;
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
+  GET_MALLOC_STACK_TRACE;
   hwasan_free(ptr, &stack);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void __sanitizer_cfree(void *ptr) {
-  GET_MALLOC_STACK_TRACE;
-  if (!ptr || UNLIKELY(IsInDlsymAllocPool(ptr)))
+  if (!ptr)
     return;
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
+  GET_MALLOC_STACK_TRACE;
   hwasan_free(ptr, &stack);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 uptr __sanitizer_malloc_usable_size(const void *ptr) {
   return __sanitizer_get_allocated_size(ptr);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 struct __sanitizer_struct_mallinfo __sanitizer_mallinfo() {
   __sanitizer_struct_mallinfo sret;
   internal_memset(&sret, 0, sizeof(sret));
   return sret;
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 int __sanitizer_mallopt(int cmd, int value) { return 0; }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void __sanitizer_malloc_stats(void) {
   // FIXME: implement, but don't call REAL(malloc_stats)!
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer_calloc(uptr nmemb, uptr size) {
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Callocate(nmemb, size);
   GET_MALLOC_STACK_TRACE;
-  if (UNLIKELY(!hwasan_inited))
-    // Hack: dlsym calls calloc before REAL(calloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(nmemb * size);
   return hwasan_calloc(nmemb, size, &stack);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer_realloc(void *ptr, uptr size) {
+  if (DlsymAlloc::Use() || DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Realloc(ptr, size);
   GET_MALLOC_STACK_TRACE;
-  if (UNLIKELY(IsInDlsymAllocPool(ptr))) {
-    uptr offset = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-    uptr copy_size = Min(size, kDlsymAllocPoolSize - offset);
-    void *new_ptr;
-    if (UNLIKELY(!hwasan_inited)) {
-      new_ptr = AllocateFromLocalPool(copy_size);
-    } else {
-      copy_size = size;
-      new_ptr = hwasan_malloc(copy_size, &stack);
-    }
-    internal_memcpy(new_ptr, ptr, copy_size);
-    return new_ptr;
-  }
   return hwasan_realloc(ptr, size, &stack);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer_reallocarray(void *ptr, uptr nmemb, uptr size) {
   GET_MALLOC_STACK_TRACE;
   return hwasan_reallocarray(ptr, nmemb, size, &stack);
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer_malloc(uptr size) {
-  GET_MALLOC_STACK_TRACE;
   if (UNLIKELY(!hwasan_init_is_running))
     ENSURE_HWASAN_INITED();
-  if (UNLIKELY(!hwasan_inited))
-    // Hack: dlsym calls malloc before REAL(malloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(size);
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Allocate(size);
+  GET_MALLOC_STACK_TRACE;
   return hwasan_malloc(size, &stack);
 }
 
+}  // extern "C"
+
 #if HWASAN_WITH_INTERCEPTORS
 #  define INTERCEPTOR_ALIAS(RET, FN, ARGS...)                                 \
     extern "C" SANITIZER_INTERFACE_ATTRIBUTE RET WRAP(FN)(ARGS)               \
@@ -170,3 +171,5 @@ INTERCEPTOR_ALIAS(int, mallopt, int cmd, int value);
 INTERCEPTOR_ALIAS(void, malloc_stats, void);
 #  endif
 #endif  // #if HWASAN_WITH_INTERCEPTORS
+
+#endif  // SANITIZER_FUCHSIA
diff --git a/compiler-rt/lib/hwasan/hwasan_allocator.cpp b/compiler-rt/lib/hwasan/hwasan_allocator.cpp
index ef6d4d6c7678..9e1729964e27 100644
--- a/compiler-rt/lib/hwasan/hwasan_allocator.cpp
+++ b/compiler-rt/lib/hwasan/hwasan_allocator.cpp
@@ -107,6 +107,10 @@ void HwasanAllocatorInit() {
     tail_magic[i] = GetCurrentThread()->GenerateRandomTag();
 }
 
+void HwasanAllocatorLock() { allocator.ForceLock(); }
+
+void HwasanAllocatorUnlock() { allocator.ForceUnlock(); }
+
 void AllocatorSwallowThreadLocalCache(AllocatorCache *cache) {
   allocator.SwallowCache(cache);
 }
@@ -158,8 +162,11 @@ static void *HwasanAllocate(StackTrace *stack, uptr orig_size, uptr alignment,
     internal_memset(allocated, flags()->malloc_fill_byte, fill_size);
   }
   if (size != orig_size) {
-    internal_memcpy(reinterpret_cast<u8 *>(allocated) + orig_size, tail_magic,
-                    size - orig_size - 1);
+    u8 *tail = reinterpret_cast<u8 *>(allocated) + orig_size;
+    uptr tail_length = size - orig_size;
+    internal_memcpy(tail, tail_magic, tail_length - 1);
+    // Short granule is excluded from magic tail, so we explicitly untag.
+    tail[tail_length - 1] = 0;
   }
 
   void *user_ptr = allocated;
@@ -201,21 +208,37 @@ static bool PointerAndMemoryTagsMatch(void *tagged_ptr) {
   return PossiblyShortTagMatches(mem_tag, tagged_uptr, 1);
 }
 
+static bool CheckInvalidFree(StackTrace *stack, void *untagged_ptr,
+                             void *tagged_ptr) {
+  // This function can return true if halt_on_error is false.
+  if (!MemIsApp(reinterpret_cast<uptr>(untagged_ptr)) ||
+      !PointerAndMemoryTagsMatch(tagged_ptr)) {
+    ReportInvalidFree(stack, reinterpret_cast<uptr>(tagged_ptr));
+    return true;
+  }
+  return false;
+}
+
 static void HwasanDeallocate(StackTrace *stack, void *tagged_ptr) {
   CHECK(tagged_ptr);
   HWASAN_FREE_HOOK(tagged_ptr);
 
-  if (!PointerAndMemoryTagsMatch(tagged_ptr))
-    ReportInvalidFree(stack, reinterpret_cast<uptr>(tagged_ptr));
+  bool in_taggable_region =
+      InTaggableRegion(reinterpret_cast<uptr>(tagged_ptr));
+  void *untagged_ptr = in_taggable_region ? UntagPtr(tagged_ptr) : tagged_ptr;
+
+  if (CheckInvalidFree(stack, untagged_ptr, tagged_ptr))
+    return;
 
-  void *untagged_ptr = InTaggableRegion(reinterpret_cast<uptr>(tagged_ptr))
-                           ? UntagPtr(tagged_ptr)
-                           : tagged_ptr;
   void *aligned_ptr = reinterpret_cast<void *>(
       RoundDownTo(reinterpret_cast<uptr>(untagged_ptr), kShadowAlignment));
   tag_t pointer_tag = GetTagFromPointer(reinterpret_cast<uptr>(tagged_ptr));
   Metadata *meta =
       reinterpret_cast<Metadata *>(allocator.GetMetaData(aligned_ptr));
+  if (!meta) {
+    ReportInvalidFree(stack, reinterpret_cast<uptr>(tagged_ptr));
+    return;
+  }
   uptr orig_size = meta->get_requested_size();
   u32 free_context_id = StackDepotPut(*stack);
   u32 alloc_context_id = meta->alloc_context_id;
@@ -228,7 +251,11 @@ static void HwasanDeallocate(StackTrace *stack, void *tagged_ptr) {
     CHECK_LT(tail_size, kShadowAlignment);
     void *tail_beg = reinterpret_cast<void *>(
         reinterpret_cast<uptr>(aligned_ptr) + orig_size);
-    if (tail_size && internal_memcmp(tail_beg, tail_magic, tail_size))
+    tag_t short_granule_memtag = *(reinterpret_cast<tag_t *>(
+        reinterpret_cast<uptr>(tail_beg) + tail_size));
+    if (tail_size &&
+        (internal_memcmp(tail_beg, tail_magic, tail_size) ||
+         (in_taggable_region && pointer_tag != short_granule_memtag)))
       ReportTailOverwritten(stack, reinterpret_cast<uptr>(tagged_ptr),
                             orig_size, tail_magic);
   }
@@ -243,8 +270,7 @@ static void HwasanDeallocate(StackTrace *stack, void *tagged_ptr) {
         Min(TaggedSize(orig_size), (uptr)flags()->max_free_fill_size);
     internal_memset(aligned_ptr, flags()->free_fill_byte, fill_size);
   }
-  if (InTaggableRegion(reinterpret_cast<uptr>(tagged_ptr)) &&
-      flags()->tag_in_free && malloc_bisect(stack, 0) &&
+  if (in_taggable_region && flags()->tag_in_free && malloc_bisect(stack, 0) &&
       atomic_load_relaxed(&hwasan_allocator_tagging_enabled)) {
     // Always store full 8-bit tags on free to maximize UAF detection.
     tag_t tag;
@@ -278,13 +304,15 @@ static void HwasanDeallocate(StackTrace *stack, void *tagged_ptr) {
 
 static void *HwasanReallocate(StackTrace *stack, void *tagged_ptr_old,
                               uptr new_size, uptr alignment) {
-  if (!PointerAndMemoryTagsMatch(tagged_ptr_old))
-    ReportInvalidFree(stack, reinterpret_cast<uptr>(tagged_ptr_old));
-
+  void *untagged_ptr_old =
+      InTaggableRegion(reinterpret_cast<uptr>(tagged_ptr_old))
+          ? UntagPtr(tagged_ptr_old)
+          : tagged_ptr_old;
+  if (CheckInvalidFree(stack, untagged_ptr_old, tagged_ptr_old))
+    return nullptr;
   void *tagged_ptr_new =
       HwasanAllocate(stack, new_size, alignment, false /*zeroise*/);
   if (tagged_ptr_old && tagged_ptr_new) {
-    void *untagged_ptr_old =  UntagPtr(tagged_ptr_old);
     Metadata *meta =
         reinterpret_cast<Metadata *>(allocator.GetMetaData(untagged_ptr_old));
     internal_memcpy(
@@ -305,6 +333,8 @@ static void *HwasanCalloc(StackTrace *stack, uptr nmemb, uptr size) {
 }
 
 HwasanChunkView FindHeapChunkByAddress(uptr address) {
+  if (!allocator.PointerIsMine(reinterpret_cast<void *>(address)))
+    return HwasanChunkView();
   void *block = allocator.GetBlockBegin(reinterpret_cast<void*>(address));
   if (!block)
     return HwasanChunkView();
diff --git a/compiler-rt/lib/hwasan/hwasan_exceptions.cpp b/compiler-rt/lib/hwasan/hwasan_exceptions.cpp
index 169e7876cb58..6ed1da335428 100644
--- a/compiler-rt/lib/hwasan/hwasan_exceptions.cpp
+++ b/compiler-rt/lib/hwasan/hwasan_exceptions.cpp
@@ -29,8 +29,8 @@ typedef _Unwind_Reason_Code PersonalityFn(int version, _Unwind_Action actions,
 // is statically linked and the sanitizer runtime and the program are linked
 // against different unwinders. The _Unwind_Context data structure is opaque so
 // it may be incompatible between unwinders.
-typedef _Unwind_Word GetGRFn(_Unwind_Context* context, int index);
-typedef _Unwind_Word GetCFAFn(_Unwind_Context* context);
+typedef uintptr_t GetGRFn(_Unwind_Context* context, int index);
+typedef uintptr_t GetCFAFn(_Unwind_Context* context);
 
 extern "C" SANITIZER_INTERFACE_ATTRIBUTE _Unwind_Reason_Code
 __hwasan_personality_wrapper(int version, _Unwind_Action actions,
diff --git a/compiler-rt/lib/hwasan/hwasan_fuchsia.cpp b/compiler-rt/lib/hwasan/hwasan_fuchsia.cpp
index e299a7e862eb..94e5c5fb69c7 100644
--- a/compiler-rt/lib/hwasan/hwasan_fuchsia.cpp
+++ b/compiler-rt/lib/hwasan/hwasan_fuchsia.cpp
@@ -130,7 +130,7 @@ static void ThreadCreateHook(void *hook, bool aborted) {
 static void ThreadStartHook(void *hook, thrd_t self) {
   Thread *thread = static_cast<Thread *>(hook);
   FinishThreadInitialization(thread);
-  thread->InitRandomState();
+  thread->EnsureRandomStateInited();
 }
 
 // This is the function that sets up the stack ring buffer and enables us to use
@@ -180,6 +180,8 @@ void HwasanTSDThreadInit() {}
 // function is unneeded.
 void InstallAtExitHandler() {}
 
+void HwasanInstallAtForkHandler() {}
+
 // TODO(fxbug.dev/81499): Once we finalize the tagged pointer ABI in zircon, we should come back
 // here and implement the appropriate check that TBI is enabled.
 void InitializeOsSupport() {}
diff --git a/compiler-rt/lib/hwasan/hwasan_interceptors.cpp b/compiler-rt/lib/hwasan/hwasan_interceptors.cpp
index 68f8adec0776..f96ed8804102 100644
--- a/compiler-rt/lib/hwasan/hwasan_interceptors.cpp
+++ b/compiler-rt/lib/hwasan/hwasan_interceptors.cpp
@@ -49,15 +49,14 @@ INTERCEPTOR(int, pthread_create, void *th, void *attr, void *(*callback)(void*),
 
 DEFINE_REAL(int, vfork)
 DECLARE_EXTERN_INTERCEPTOR_AND_WRAPPER(int, vfork)
-#endif // HWASAN_WITH_INTERCEPTORS
 
-#if HWASAN_WITH_INTERCEPTORS && defined(__aarch64__)
 // Get and/or change the set of blocked signals.
 extern "C" int sigprocmask(int __how, const __hw_sigset_t *__restrict __set,
                            __hw_sigset_t *__restrict __oset);
 #define SIG_BLOCK 0
 #define SIG_SETMASK 2
 extern "C" int __sigjmp_save(__hw_sigjmp_buf env, int savemask) {
+  env[0].__magic = kHwJmpBufMagic;
   env[0].__mask_was_saved =
       (savemask && sigprocmask(SIG_BLOCK, (__hw_sigset_t *)0,
                                &env[0].__saved_mask) == 0);
@@ -66,8 +65,14 @@ extern "C" int __sigjmp_save(__hw_sigjmp_buf env, int savemask) {
 
 static void __attribute__((always_inline))
 InternalLongjmp(__hw_register_buf env, int retval) {
+#    if defined(__aarch64__)
+  constexpr size_t kSpIndex = 13;
+#    elif defined(__x86_64__)
+  constexpr size_t kSpIndex = 6;
+#    endif
+
   // Clear all memory tags on the stack between here and where we're going.
-  unsigned long long stack_pointer = env[13];
+  unsigned long long stack_pointer = env[kSpIndex];
   // The stack pointer should never be tagged, so we don't need to clear the
   // tag for this function call.
   __hwasan_handle_longjmp((void *)stack_pointer);
@@ -78,6 +83,7 @@ InternalLongjmp(__hw_register_buf env, int retval) {
   // Must implement this ourselves, since we don't know the order of registers
   // in different libc implementations and many implementations mangle the
   // stack pointer so we can't use it without knowing the demangling scheme.
+#    if defined(__aarch64__)
   register long int retval_tmp asm("x1") = retval;
   register void *env_address asm("x0") = &env[0];
   asm volatile("ldp	x19, x20, [%0, #0<<3];"
@@ -100,9 +106,36 @@ InternalLongjmp(__hw_register_buf env, int retval) {
                "br	x30;"
                : "+r"(env_address)
                : "r"(retval_tmp));
+#    elif defined(__x86_64__)
+  register long int retval_tmp asm("%rsi") = retval;
+  register void *env_address asm("%rdi") = &env[0];
+  asm volatile(
+      // Restore registers.
+      "mov (0*8)(%0),%%rbx;"
+      "mov (1*8)(%0),%%rbp;"
+      "mov (2*8)(%0),%%r12;"
+      "mov (3*8)(%0),%%r13;"
+      "mov (4*8)(%0),%%r14;"
+      "mov (5*8)(%0),%%r15;"
+      "mov (6*8)(%0),%%rsp;"
+      "mov (7*8)(%0),%%rdx;"
+      // Return 1 if retval is 0.
+      "mov $1,%%rax;"
+      "test %1,%1;"
+      "cmovnz %1,%%rax;"
+      "jmp *%%rdx;" ::"r"(env_address),
+      "r"(retval_tmp));
+#    endif
 }
 
 INTERCEPTOR(void, siglongjmp, __hw_sigjmp_buf env, int val) {
+  if (env[0].__magic != kHwJmpBufMagic) {
+    Printf(
+        "WARNING: Unexpected bad jmp_buf. Either setjmp was not called or "
+        "there is a bug in HWASan.\n");
+    return REAL(siglongjmp)(env, val);
+  }
+
   if (env[0].__mask_was_saved)
     // Restore the saved signal mask.
     (void)sigprocmask(SIG_SETMASK, &env[0].__saved_mask,
@@ -114,32 +147,24 @@ INTERCEPTOR(void, siglongjmp, __hw_sigjmp_buf env, int val) {
 // _setjmp on start_thread.  Hence we have to intercept the longjmp on
 // pthread_exit so the __hw_jmp_buf order matches.
 INTERCEPTOR(void, __libc_longjmp, __hw_jmp_buf env, int val) {
+  if (env[0].__magic != kHwJmpBufMagic)
+    return REAL(__libc_longjmp)(env, val);
   InternalLongjmp(env[0].__jmpbuf, val);
 }
 
 INTERCEPTOR(void, longjmp, __hw_jmp_buf env, int val) {
+  if (env[0].__magic != kHwJmpBufMagic) {
+    Printf(
+        "WARNING: Unexpected bad jmp_buf. Either setjmp was not called or "
+        "there is a bug in HWASan.\n");
+    return REAL(longjmp)(env, val);
+  }
   InternalLongjmp(env[0].__jmpbuf, val);
 }
 #undef SIG_BLOCK
 #undef SIG_SETMASK
 
-#endif // HWASAN_WITH_INTERCEPTORS && __aarch64__
-
-static void BeforeFork() {
-  StackDepotLockAll();
-}
-
-static void AfterFork() {
-  StackDepotUnlockAll();
-}
-
-INTERCEPTOR(int, fork, void) {
-  ENSURE_HWASAN_INITED();
-  BeforeFork();
-  int pid = REAL(fork)();
-  AfterFork();
-  return pid;
-}
+#  endif  // HWASAN_WITH_INTERCEPTORS
 
 namespace __hwasan {
 
@@ -156,10 +181,11 @@ void InitializeInterceptors() {
   static int inited = 0;
   CHECK_EQ(inited, 0);
 
-  INTERCEPT_FUNCTION(fork);
-
 #if HWASAN_WITH_INTERCEPTORS
 #if defined(__linux__)
+  INTERCEPT_FUNCTION(__libc_longjmp);
+  INTERCEPT_FUNCTION(longjmp);
+  INTERCEPT_FUNCTION(siglongjmp);
   INTERCEPT_FUNCTION(vfork);
 #endif  // __linux__
   INTERCEPT_FUNCTION(pthread_create);
diff --git a/compiler-rt/lib/hwasan/hwasan_interface_internal.h b/compiler-rt/lib/hwasan/hwasan_interface_internal.h
index 25c0f94fe51f..ef771add411c 100644
--- a/compiler-rt/lib/hwasan/hwasan_interface_internal.h
+++ b/compiler-rt/lib/hwasan/hwasan_interface_internal.h
@@ -169,54 +169,6 @@ SANITIZER_INTERFACE_ATTRIBUTE
 void __hwasan_print_memory_usage();
 
 SANITIZER_INTERFACE_ATTRIBUTE
-int __sanitizer_posix_memalign(void **memptr, uptr alignment, uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void * __sanitizer_memalign(uptr alignment, uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void * __sanitizer_aligned_alloc(uptr alignment, uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void * __sanitizer___libc_memalign(uptr alignment, uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void * __sanitizer_valloc(uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void * __sanitizer_pvalloc(uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void __sanitizer_free(void *ptr);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void __sanitizer_cfree(void *ptr);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-uptr __sanitizer_malloc_usable_size(const void *ptr);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-__hwasan::__sanitizer_struct_mallinfo __sanitizer_mallinfo();
-
-SANITIZER_INTERFACE_ATTRIBUTE
-int __sanitizer_mallopt(int cmd, int value);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void __sanitizer_malloc_stats(void);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void * __sanitizer_calloc(uptr nmemb, uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void * __sanitizer_realloc(void *ptr, uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void * __sanitizer_reallocarray(void *ptr, uptr nmemb, uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void * __sanitizer_malloc(uptr size);
-
-SANITIZER_INTERFACE_ATTRIBUTE
 void *__hwasan_memcpy(void *dst, const void *src, uptr size);
 SANITIZER_INTERFACE_ATTRIBUTE
 void *__hwasan_memset(void *s, int c, uptr n);
diff --git a/compiler-rt/lib/hwasan/hwasan_linux.cpp b/compiler-rt/lib/hwasan/hwasan_linux.cpp
index e22723529f44..ba9e23621cc2 100644
--- a/compiler-rt/lib/hwasan/hwasan_linux.cpp
+++ b/compiler-rt/lib/hwasan/hwasan_linux.cpp
@@ -15,30 +15,30 @@
 #include "sanitizer_common/sanitizer_platform.h"
 #if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD
 
-#include "hwasan.h"
-#include "hwasan_dynamic_shadow.h"
-#include "hwasan_interface_internal.h"
-#include "hwasan_mapping.h"
-#include "hwasan_report.h"
-#include "hwasan_thread.h"
-#include "hwasan_thread_list.h"
+#  include <dlfcn.h>
+#  include <elf.h>
+#  include <errno.h>
+#  include <link.h>
+#  include <pthread.h>
+#  include <signal.h>
+#  include <stdio.h>
+#  include <stdlib.h>
+#  include <sys/prctl.h>
+#  include <sys/resource.h>
+#  include <sys/time.h>
+#  include <unistd.h>
+#  include <unwind.h>
 
-#include <dlfcn.h>
-#include <elf.h>
-#include <link.h>
-#include <pthread.h>
-#include <signal.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <sys/resource.h>
-#include <sys/time.h>
-#include <unistd.h>
-#include <unwind.h>
-#include <sys/prctl.h>
-#include <errno.h>
-
-#include "sanitizer_common/sanitizer_common.h"
-#include "sanitizer_common/sanitizer_procmaps.h"
+#  include "hwasan.h"
+#  include "hwasan_dynamic_shadow.h"
+#  include "hwasan_interface_internal.h"
+#  include "hwasan_mapping.h"
+#  include "hwasan_report.h"
+#  include "hwasan_thread.h"
+#  include "hwasan_thread_list.h"
+#  include "sanitizer_common/sanitizer_common.h"
+#  include "sanitizer_common/sanitizer_procmaps.h"
+#  include "sanitizer_common/sanitizer_stackdepot.h"
 
 // Configurations of HWASAN_WITH_INTERCEPTORS and SANITIZER_ANDROID.
 //
@@ -50,10 +50,10 @@
 //    Tested with check-hwasan on x86_64-linux.
 // HWASAN_WITH_INTERCEPTORS=ON, SANITIZER_ANDROID=ON
 //    Tested with check-hwasan on aarch64-linux-android.
-#if !SANITIZER_ANDROID
+#  if !SANITIZER_ANDROID
 SANITIZER_INTERFACE_ATTRIBUTE
 THREADLOCAL uptr __hwasan_tls;
-#endif
+#  endif
 
 namespace __hwasan {
 
@@ -111,9 +111,9 @@ static void InitializeShadowBaseAddress(uptr shadow_size_bytes) {
 }
 
 void InitializeOsSupport() {
-#define PR_SET_TAGGED_ADDR_CTRL 55
-#define PR_GET_TAGGED_ADDR_CTRL 56
-#define PR_TAGGED_ADDR_ENABLE (1UL << 0)
+#  define PR_SET_TAGGED_ADDR_CTRL 55
+#  define PR_GET_TAGGED_ADDR_CTRL 56
+#  define PR_TAGGED_ADDR_ENABLE (1UL << 0)
   // Check we're running on a kernel that can use the tagged address ABI.
   int local_errno = 0;
   if (internal_iserror(internal_prctl(PR_GET_TAGGED_ADDR_CTRL, 0, 0, 0, 0),
@@ -164,9 +164,9 @@ void InitializeOsSupport() {
       Die();
     }
   }
-#undef PR_SET_TAGGED_ADDR_CTRL
-#undef PR_GET_TAGGED_ADDR_CTRL
-#undef PR_TAGGED_ADDR_ENABLE
+#  undef PR_SET_TAGGED_ADDR_CTRL
+#  undef PR_GET_TAGGED_ADDR_CTRL
+#  undef PR_TAGGED_ADDR_ENABLE
 }
 
 bool InitShadow() {
@@ -241,17 +241,16 @@ bool MemIsApp(uptr p) {
   CHECK(GetTagFromPointer(p) == 0);
 #  endif
 
-  return p >= kHighMemStart || (p >= kLowMemStart && p <= kLowMemEnd);
+  return (p >= kHighMemStart && p <= kHighMemEnd) ||
+         (p >= kLowMemStart && p <= kLowMemEnd);
 }
 
-void InstallAtExitHandler() {
-  atexit(HwasanAtExit);
-}
+void InstallAtExitHandler() { atexit(HwasanAtExit); }
 
 // ---------------------- TSD ---------------- {{{1
 
 extern "C" void __hwasan_thread_enter() {
-  hwasanThreadList().CreateCurrentThread()->InitRandomState();
+  hwasanThreadList().CreateCurrentThread()->EnsureRandomStateInited();
 }
 
 extern "C" void __hwasan_thread_exit() {
@@ -262,7 +261,7 @@ extern "C" void __hwasan_thread_exit() {
     hwasanThreadList().ReleaseThread(t);
 }
 
-#if HWASAN_WITH_INTERCEPTORS
+#  if HWASAN_WITH_INTERCEPTORS
 static pthread_key_t tsd_key;
 static bool tsd_key_inited = false;
 
@@ -286,22 +285,18 @@ void HwasanTSDInit() {
   tsd_key_inited = true;
   CHECK_EQ(0, pthread_key_create(&tsd_key, HwasanTSDDtor));
 }
-#else
+#  else
 void HwasanTSDInit() {}
 void HwasanTSDThreadInit() {}
-#endif
+#  endif
 
-#if SANITIZER_ANDROID
-uptr *GetCurrentThreadLongPtr() {
-  return (uptr *)get_android_tls_ptr();
-}
-#else
-uptr *GetCurrentThreadLongPtr() {
-  return &__hwasan_tls;
-}
-#endif
+#  if SANITIZER_ANDROID
+uptr *GetCurrentThreadLongPtr() { return (uptr *)get_android_tls_ptr(); }
+#  else
+uptr *GetCurrentThreadLongPtr() { return &__hwasan_tls; }
+#  endif
 
-#if SANITIZER_ANDROID
+#  if SANITIZER_ANDROID
 void AndroidTestTlsSlot() {
   uptr kMagicValue = 0x010203040A0B0C0D;
   uptr *tls_ptr = GetCurrentThreadLongPtr();
@@ -316,9 +311,9 @@ void AndroidTestTlsSlot() {
   }
   *tls_ptr = old_value;
 }
-#else
+#  else
 void AndroidTestTlsSlot() {}
-#endif
+#  endif
 
 static AccessInfo GetAccessInfo(siginfo_t *info, ucontext_t *uc) {
   // Access type is passed in a platform dependent way (see below) and encoded
@@ -326,32 +321,32 @@ static AccessInfo GetAccessInfo(siginfo_t *info, ucontext_t *uc) {
   // recoverable. Valid values of Y are 0 to 4, which are interpreted as
   // log2(access_size), and 0xF, which means that access size is passed via
   // platform dependent register (see below).
-#if defined(__aarch64__)
+#  if defined(__aarch64__)
   // Access type is encoded in BRK immediate as 0x900 + 0xXY. For Y == 0xF,
   // access size is stored in X1 register. Access address is always in X0
   // register.
   uptr pc = (uptr)info->si_addr;
   const unsigned code = ((*(u32 *)pc) >> 5) & 0xffff;
   if ((code & 0xff00) != 0x900)
-    return AccessInfo{}; // Not ours.
+    return AccessInfo{};  // Not ours.
 
   const bool is_store = code & 0x10;
   const bool recover = code & 0x20;
   const uptr addr = uc->uc_mcontext.regs[0];
   const unsigned size_log = code & 0xf;
   if (size_log > 4 && size_log != 0xf)
-    return AccessInfo{}; // Not ours.
+    return AccessInfo{};  // Not ours.
   const uptr size = size_log == 0xf ? uc->uc_mcontext.regs[1] : 1U << size_log;
 
-#elif defined(__x86_64__)
+#  elif defined(__x86_64__)
   // Access type is encoded in the instruction following INT3 as
   // NOP DWORD ptr [EAX + 0x40 + 0xXY]. For Y == 0xF, access size is stored in
   // RSI register. Access address is always in RDI register.
   uptr pc = (uptr)uc->uc_mcontext.gregs[REG_RIP];
-  uint8_t *nop = (uint8_t*)pc;
-  if (*nop != 0x0f || *(nop + 1) != 0x1f || *(nop + 2) != 0x40  ||
+  uint8_t *nop = (uint8_t *)pc;
+  if (*nop != 0x0f || *(nop + 1) != 0x1f || *(nop + 2) != 0x40 ||
       *(nop + 3) < 0x40)
-    return AccessInfo{}; // Not ours.
+    return AccessInfo{};  // Not ours.
   const unsigned code = *(nop + 3);
 
   const bool is_store = code & 0x10;
@@ -359,13 +354,13 @@ static AccessInfo GetAccessInfo(siginfo_t *info, ucontext_t *uc) {
   const uptr addr = uc->uc_mcontext.gregs[REG_RDI];
   const unsigned size_log = code & 0xf;
   if (size_log > 4 && size_log != 0xf)
-    return AccessInfo{}; // Not ours.
+    return AccessInfo{};  // Not ours.
   const uptr size =
       size_log == 0xf ? uc->uc_mcontext.gregs[REG_RSI] : 1U << size_log;
 
-#else
-# error Unsupported architecture
-#endif
+#  else
+#    error Unsupported architecture
+#  endif
 
   return AccessInfo{addr, size, is_store, !is_store, recover};
 }
@@ -378,12 +373,12 @@ static bool HwasanOnSIGTRAP(int signo, siginfo_t *info, ucontext_t *uc) {
   SignalContext sig{info, uc};
   HandleTagMismatch(ai, StackTrace::GetNextInstructionPc(sig.pc), sig.bp, uc);
 
-#if defined(__aarch64__)
+#  if defined(__aarch64__)
   uc->uc_mcontext.pc += 4;
-#elif defined(__x86_64__)
-#else
-# error Unsupported architecture
-#endif
+#  elif defined(__x86_64__)
+#  else
+#    error Unsupported architecture
+#  endif
   return true;
 }
 
@@ -396,7 +391,7 @@ static void OnStackUnwind(const SignalContext &sig, const void *,
 void HwasanOnDeadlySignal(int signo, void *info, void *context) {
   // Probably a tag mismatch.
   if (signo == SIGTRAP)
-    if (HwasanOnSIGTRAP(signo, (siginfo_t *)info, (ucontext_t*)context))
+    if (HwasanOnSIGTRAP(signo, (siginfo_t *)info, (ucontext_t *)context))
       return;
 
   HandleDeadlySignal(info, context, GetTid(), &OnStackUnwind, nullptr);
@@ -435,6 +430,18 @@ uptr TagMemoryAligned(uptr p, uptr size, tag_t tag) {
   return AddTagToPointer(p, tag);
 }
 
-} // namespace __hwasan
+void HwasanInstallAtForkHandler() {
+  auto before = []() {
+    HwasanAllocatorLock();
+    StackDepotLockAll();
+  };
+  auto after = []() {
+    StackDepotUnlockAll();
+    HwasanAllocatorUnlock();
+  };
+  pthread_atfork(before, after, after);
+}
+
+}  // namespace __hwasan
 
-#endif // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD
+#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD
diff --git a/compiler-rt/lib/hwasan/hwasan_report.cpp b/compiler-rt/lib/hwasan/hwasan_report.cpp
index 44047c9fdaf8..66d3d155d409 100644
--- a/compiler-rt/lib/hwasan/hwasan_report.cpp
+++ b/compiler-rt/lib/hwasan/hwasan_report.cpp
@@ -37,7 +37,7 @@ namespace __hwasan {
 class ScopedReport {
  public:
   ScopedReport(bool fatal = false) : error_message_(1), fatal(fatal) {
-    BlockingMutexLock lock(&error_message_lock_);
+    Lock lock(&error_message_lock_);
     error_message_ptr_ = fatal ? &error_message_ : nullptr;
     ++hwasan_report_count;
   }
@@ -45,7 +45,7 @@ class ScopedReport {
   ~ScopedReport() {
     void (*report_cb)(const char *);
     {
-      BlockingMutexLock lock(&error_message_lock_);
+      Lock lock(&error_message_lock_);
       report_cb = error_report_callback_;
       error_message_ptr_ = nullptr;
     }
@@ -61,7 +61,7 @@ class ScopedReport {
   }
 
   static void MaybeAppendToErrorMessage(const char *msg) {
-    BlockingMutexLock lock(&error_message_lock_);
+    Lock lock(&error_message_lock_);
     if (!error_message_ptr_)
       return;
     uptr len = internal_strlen(msg);
@@ -72,7 +72,7 @@ class ScopedReport {
   }
 
   static void SetErrorReportCallback(void (*callback)(const char *)) {
-    BlockingMutexLock lock(&error_message_lock_);
+    Lock lock(&error_message_lock_);
     error_report_callback_ = callback;
   }
 
@@ -82,12 +82,12 @@ class ScopedReport {
   bool fatal;
 
   static InternalMmapVector<char> *error_message_ptr_;
-  static BlockingMutex error_message_lock_;
+  static Mutex error_message_lock_;
   static void (*error_report_callback_)(const char *);
 };
 
 InternalMmapVector<char> *ScopedReport::error_message_ptr_;
-BlockingMutex ScopedReport::error_message_lock_;
+Mutex ScopedReport::error_message_lock_;
 void (*ScopedReport::error_report_callback_)(const char *);
 
 // If there is an active ScopedReport, append to its error message.
@@ -351,14 +351,16 @@ static void ShowHeapOrGlobalCandidate(uptr untagged_addr, tag_t *candidate,
       uptr size = GetGlobalSizeFromDescriptor(mem);
       if (size == 0)
         // We couldn't find the size of the global from the descriptors.
-        Printf("%p is located to the %s of a global variable in (%s+0x%x)\n",
-               untagged_addr, candidate == left ? "right" : "left", module_name,
-               module_address);
+        Printf(
+            "%p is located to the %s of a global variable in "
+            "\n    #0 0x%x (%s+0x%x)\n",
+            untagged_addr, candidate == left ? "right" : "left", mem,
+            module_name, module_address);
       else
         Printf(
             "%p is located to the %s of a %zd-byte global variable in "
-            "(%s+0x%x)\n",
-            untagged_addr, candidate == left ? "right" : "left", size,
+            "\n    #0 0x%x (%s+0x%x)\n",
+            untagged_addr, candidate == left ? "right" : "left", size, mem,
             module_name, module_address);
     }
     Printf("%s", d.Default());
@@ -372,6 +374,12 @@ void PrintAddressDescription(
   int num_descriptions_printed = 0;
   uptr untagged_addr = UntagAddr(tagged_addr);
 
+  if (MemIsShadow(untagged_addr)) {
+    Printf("%s%p is HWAsan shadow memory.\n%s", d.Location(), untagged_addr,
+           d.Default());
+    return;
+  }
+
   // Print some very basic information about the address, if it's a heap.
   HwasanChunkView chunk = FindHeapChunkByAddress(untagged_addr);
   if (uptr beg = chunk.Beg()) {
@@ -510,7 +518,7 @@ static void PrintTagInfoAroundAddr(tag_t *tag_ptr, uptr num_rows,
   InternalScopedString s;
   for (tag_t *row = beg_row; row < end_row; row += row_len) {
     s.append("%s", row == center_row_beg ? "=>" : "  ");
-    s.append("%p:", row);
+    s.append("%p:", (void *)row);
     for (uptr i = 0; i < row_len; i++) {
       s.append("%s", row + i == tag_ptr ? "[" : " ");
       print_tag(s, &row[i]);
@@ -549,28 +557,48 @@ static void PrintTagsAroundAddr(tag_t *tag_ptr) {
       "description of short granule tags\n");
 }
 
+uptr GetTopPc(StackTrace *stack) {
+  return stack->size ? StackTrace::GetPreviousInstructionPc(stack->trace[0])
+                     : 0;
+}
+
 void ReportInvalidFree(StackTrace *stack, uptr tagged_addr) {
   ScopedReport R(flags()->halt_on_error);
 
   uptr untagged_addr = UntagAddr(tagged_addr);
   tag_t ptr_tag = GetTagFromPointer(tagged_addr);
-  tag_t *tag_ptr = reinterpret_cast<tag_t*>(MemToShadow(untagged_addr));
-  tag_t mem_tag = *tag_ptr;
+  tag_t *tag_ptr = nullptr;
+  tag_t mem_tag = 0;
+  if (MemIsApp(untagged_addr)) {
+    tag_ptr = reinterpret_cast<tag_t *>(MemToShadow(untagged_addr));
+    if (MemIsShadow(reinterpret_cast<uptr>(tag_ptr)))
+      mem_tag = *tag_ptr;
+    else
+      tag_ptr = nullptr;
+  }
   Decorator d;
   Printf("%s", d.Error());
-  uptr pc = stack->size ? stack->trace[0] : 0;
+  uptr pc = GetTopPc(stack);
   const char *bug_type = "invalid-free";
-  Report("ERROR: %s: %s on address %p at pc %p\n", SanitizerToolName, bug_type,
-         untagged_addr, pc);
+  const Thread *thread = GetCurrentThread();
+  if (thread) {
+    Report("ERROR: %s: %s on address %p at pc %p on thread T%zd\n",
+           SanitizerToolName, bug_type, untagged_addr, pc, thread->unique_id());
+  } else {
+    Report("ERROR: %s: %s on address %p at pc %p on unknown thread\n",
+           SanitizerToolName, bug_type, untagged_addr, pc);
+  }
   Printf("%s", d.Access());
-  Printf("tags: %02x/%02x (ptr/mem)\n", ptr_tag, mem_tag);
+  if (tag_ptr)
+    Printf("tags: %02x/%02x (ptr/mem)\n", ptr_tag, mem_tag);
   Printf("%s", d.Default());
 
   stack->Print();
 
   PrintAddressDescription(tagged_addr, 0, nullptr);
 
-  PrintTagsAroundAddr(tag_ptr);
+  if (tag_ptr)
+    PrintTagsAroundAddr(tag_ptr);
 
   ReportErrorSummary(bug_type, stack);
 }
@@ -578,6 +606,15 @@ void ReportInvalidFree(StackTrace *stack, uptr tagged_addr) {
 void ReportTailOverwritten(StackTrace *stack, uptr tagged_addr, uptr orig_size,
                            const u8 *expected) {
   uptr tail_size = kShadowAlignment - (orig_size % kShadowAlignment);
+  u8 actual_expected[kShadowAlignment];
+  internal_memcpy(actual_expected, expected, tail_size);
+  tag_t ptr_tag = GetTagFromPointer(tagged_addr);
+  // Short granule is stashed in the last byte of the magic string. To avoid
+  // confusion, make the expected magic string contain the short granule tag.
+  if (orig_size % kShadowAlignment != 0) {
+    actual_expected[tail_size - 1] = ptr_tag;
+  }
+
   ScopedReport R(flags()->halt_on_error);
   Decorator d;
   uptr untagged_addr = UntagAddr(tagged_addr);
@@ -614,14 +651,13 @@ void ReportTailOverwritten(StackTrace *stack, uptr tagged_addr, uptr orig_size,
   s.append("Expected:      ");
   for (uptr i = 0; i < kShadowAlignment - tail_size; i++)
     s.append(".. ");
-  for (uptr i = 0; i < tail_size; i++)
-    s.append("%02x ", expected[i]);
+  for (uptr i = 0; i < tail_size; i++) s.append("%02x ", actual_expected[i]);
   s.append("\n");
   s.append("               ");
   for (uptr i = 0; i < kShadowAlignment - tail_size; i++)
     s.append("   ");
   for (uptr i = 0; i < tail_size; i++)
-    s.append("%s ", expected[i] != tail[i] ? "^^" : "  ");
+    s.append("%s ", actual_expected[i] != tail[i] ? "^^" : "  ");
 
   s.append("\nThis error occurs when a buffer overflow overwrites memory\n"
     "to the right of a heap object, but within the %zd-byte granule, e.g.\n"
@@ -647,11 +683,11 @@ void ReportTagMismatch(StackTrace *stack, uptr tagged_addr, uptr access_size,
       GetCurrentThread()->stack_allocations());
 
   Decorator d;
-  Printf("%s", d.Error());
   uptr untagged_addr = UntagAddr(tagged_addr);
   // TODO: when possible, try to print heap-use-after-free, etc.
   const char *bug_type = "tag-mismatch";
-  uptr pc = stack->size ? stack->trace[0] : 0;
+  uptr pc = GetTopPc(stack);
+  Printf("%s", d.Error());
   Report("ERROR: %s: %s on address %p at pc %p\n", SanitizerToolName, bug_type,
          untagged_addr, pc);
 
@@ -666,12 +702,33 @@ void ReportTagMismatch(StackTrace *stack, uptr tagged_addr, uptr access_size,
   tag_t mem_tag = *tag_ptr;
 
   Printf("%s", d.Access());
-  Printf("%s of size %zu at %p tags: %02x/%02x (ptr/mem) in thread T%zd\n",
-         is_store ? "WRITE" : "READ", access_size, untagged_addr, ptr_tag,
-         mem_tag, t->unique_id());
+  if (mem_tag && mem_tag < kShadowAlignment) {
+    tag_t *granule_ptr = reinterpret_cast<tag_t *>((untagged_addr + offset) &
+                                                   ~(kShadowAlignment - 1));
+    // If offset is 0, (untagged_addr + offset) is not aligned to granules.
+    // This is the offset of the leftmost accessed byte within the bad granule.
+    u8 in_granule_offset = (untagged_addr + offset) & (kShadowAlignment - 1);
+    tag_t short_tag = granule_ptr[kShadowAlignment - 1];
+    // The first mismatch was a short granule that matched the ptr_tag.
+    if (short_tag == ptr_tag) {
+      // If the access starts after the end of the short granule, then the first
+      // bad byte is the first byte of the access; otherwise it is the first
+      // byte past the end of the short granule
+      if (mem_tag > in_granule_offset) {
+        offset += mem_tag - in_granule_offset;
+      }
+    }
+    Printf(
+        "%s of size %zu at %p tags: %02x/%02x(%02x) (ptr/mem) in thread T%zd\n",
+        is_store ? "WRITE" : "READ", access_size, untagged_addr, ptr_tag,
+        mem_tag, short_tag, t->unique_id());
+  } else {
+    Printf("%s of size %zu at %p tags: %02x/%02x (ptr/mem) in thread T%zd\n",
+           is_store ? "WRITE" : "READ", access_size, untagged_addr, ptr_tag,
+           mem_tag, t->unique_id());
+  }
   if (offset != 0)
-    Printf("Invalid access starting at offset [%zu, %zu)\n", offset,
-           Min(access_size, static_cast<uptr>(offset) + (1 << kShadowScale)));
+    Printf("Invalid access starting at offset %zu\n", offset);
   Printf("%s", d.Default());
 
   stack->Print();
diff --git a/compiler-rt/lib/hwasan/hwasan_setjmp.S b/compiler-rt/lib/hwasan/hwasan_setjmp_aarch64.S
index 381af63363cc..744748a5101f 100644
--- a/compiler-rt/lib/hwasan/hwasan_setjmp.S
+++ b/compiler-rt/lib/hwasan/hwasan_setjmp_aarch64.S
@@ -1,4 +1,4 @@
-//===-- hwasan_setjmp.S --------------------------------------------------------===//
+//===-- hwasan_setjmp_aarch64.S -------------------------------------------===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
@@ -29,7 +29,7 @@
 // Hence we have to write this function in assembly.
 
 .section .text
-.file "hwasan_setjmp.S"
+.file "hwasan_setjmp_aarch64.S"
 
 .global __interceptor_setjmp
 ASM_TYPE_FUNCTION(__interceptor_setjmp)
@@ -80,24 +80,19 @@ __interceptor_sigsetjmp:
 ASM_SIZE(__interceptor_sigsetjmp)
 
 
-.macro ALIAS first second
-  .globl \second
+.macro WEAK_ALIAS first second
+  .weak \second
   .equ \second\(), \first
 .endm
 
 #if SANITIZER_ANDROID
-ALIAS __interceptor_sigsetjmp, sigsetjmp
-.weak sigsetjmp
-
-ALIAS __interceptor_setjmp_bionic, setjmp
-.weak setjmp
+WEAK_ALIAS __interceptor_sigsetjmp, sigsetjmp
+WEAK_ALIAS __interceptor_setjmp_bionic, setjmp
 #else
-ALIAS __interceptor_sigsetjmp, __sigsetjmp
-.weak __sigsetjmp
+WEAK_ALIAS __interceptor_sigsetjmp, __sigsetjmp
 #endif
 
-ALIAS __interceptor_setjmp, _setjmp
-.weak _setjmp
+WEAK_ALIAS __interceptor_setjmp, _setjmp
 #endif
 
 // We do not need executable stack.
diff --git a/compiler-rt/lib/hwasan/hwasan_setjmp_x86_64.S b/compiler-rt/lib/hwasan/hwasan_setjmp_x86_64.S
new file mode 100644
index 000000000000..7566c1ea0a57
--- /dev/null
+++ b/compiler-rt/lib/hwasan/hwasan_setjmp_x86_64.S
@@ -0,0 +1,82 @@
+//===-- hwasan_setjmp_x86_64.S --------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// setjmp interceptor for x86_64.
+//
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_common/sanitizer_asm.h"
+
+#if HWASAN_WITH_INTERCEPTORS && defined(__x86_64__)
+#include "sanitizer_common/sanitizer_platform.h"
+
+// We want to save the context of the calling function.
+// That requires
+// 1) No modification of the return address by this function.
+// 2) No modification of the stack pointer by this function.
+// 3) (no modification of any other saved register, but that's not really going
+// to occur, and hence isn't as much of a worry).
+//
+// There's essentially no way to ensure that the compiler will not modify the
+// stack pointer when compiling a C function.
+// Hence we have to write this function in assembly.
+//
+// TODO: Handle Intel CET.
+
+.section .text
+.file "hwasan_setjmp_x86_64.S"
+
+.global __interceptor_setjmp
+ASM_TYPE_FUNCTION(__interceptor_setjmp)
+__interceptor_setjmp:
+  CFI_STARTPROC
+  _CET_ENDBR
+  xorl %esi, %esi
+  jmp	__interceptor_sigsetjmp
+  CFI_ENDPROC
+ASM_SIZE(__interceptor_setjmp)
+
+.global __interceptor_sigsetjmp
+ASM_TYPE_FUNCTION(__interceptor_sigsetjmp)
+__interceptor_sigsetjmp:
+  CFI_STARTPROC
+  _CET_ENDBR
+
+  // Save callee save registers.
+  mov %rbx, (0*8)(%rdi)
+  mov %rbp, (1*8)(%rdi)
+  mov %r12, (2*8)(%rdi)
+  mov %r13, (3*8)(%rdi)
+  mov %r14, (4*8)(%rdi)
+  mov %r15, (5*8)(%rdi)
+
+  // Save SP as it was in caller's frame.
+  lea 8(%rsp), %rdx
+  mov %rdx, (6*8)(%rdi)
+
+  // Save return address.
+  mov (%rsp), %rax
+  mov %rax, (7*8)(%rdi)
+
+  jmp __sigjmp_save
+
+  CFI_ENDPROC
+ASM_SIZE(__interceptor_sigsetjmp)
+
+
+.macro WEAK_ALIAS first second
+  .weak \second
+  .equ \second\(), \first
+.endm
+
+WEAK_ALIAS __interceptor_sigsetjmp, __sigsetjmp
+WEAK_ALIAS __interceptor_setjmp, _setjmp
+#endif
+
+// We do not need executable stack.
+NO_EXEC_STACK_DIRECTIVE
diff --git a/compiler-rt/lib/hwasan/hwasan_thread.cpp b/compiler-rt/lib/hwasan/hwasan_thread.cpp
index ee747a3beea5..c776ae179cec 100644
--- a/compiler-rt/lib/hwasan/hwasan_thread.cpp
+++ b/compiler-rt/lib/hwasan/hwasan_thread.cpp
@@ -1,15 +1,15 @@
 
+#include "hwasan_thread.h"
+
 #include "hwasan.h"
+#include "hwasan_interface_internal.h"
 #include "hwasan_mapping.h"
-#include "hwasan_thread.h"
 #include "hwasan_poisoning.h"
-#include "hwasan_interface_internal.h"
-
+#include "sanitizer_common/sanitizer_atomic.h"
 #include "sanitizer_common/sanitizer_file.h"
 #include "sanitizer_common/sanitizer_placement_new.h"
 #include "sanitizer_common/sanitizer_tls_get_addr.h"
 
-
 namespace __hwasan {
 
 static u32 RandomSeed() {
@@ -27,6 +27,7 @@ static u32 RandomSeed() {
 
 void Thread::InitRandomState() {
   random_state_ = flags()->random_tags ? RandomSeed() : unique_id_;
+  random_state_inited_ = true;
 
   // Push a random number of zeros onto the ring buffer so that the first stack
   // tag base will be random.
@@ -40,18 +41,19 @@ void Thread::Init(uptr stack_buffer_start, uptr stack_buffer_size,
   CHECK_EQ(0, stack_top_);
   CHECK_EQ(0, stack_bottom_);
 
-  static u64 unique_id;
-  unique_id_ = unique_id++;
+  static atomic_uint64_t unique_id;
+  unique_id_ = atomic_fetch_add(&unique_id, 1, memory_order_relaxed);
+
   if (auto sz = flags()->heap_history_size)
     heap_allocations_ = HeapAllocationsRingBuffer::New(sz);
 
-  InitStackAndTls(state);
 #if !SANITIZER_FUCHSIA
   // Do not initialize the stack ring buffer just yet on Fuchsia. Threads will
   // be initialized before we enter the thread itself, so we will instead call
   // this later.
   InitStackRingBuffer(stack_buffer_start, stack_buffer_size);
 #endif
+  InitStackAndTls(state);
 }
 
 void Thread::InitStackRingBuffer(uptr stack_buffer_start,
@@ -108,10 +110,9 @@ void Thread::Destroy() {
 }
 
 void Thread::Print(const char *Prefix) {
-  Printf("%sT%zd %p stack: [%p,%p) sz: %zd tls: [%p,%p)\n", Prefix,
-         unique_id_, this, stack_bottom(), stack_top(),
-         stack_top() - stack_bottom(),
-         tls_begin(), tls_end());
+  Printf("%sT%zd %p stack: [%p,%p) sz: %zd tls: [%p,%p)\n", Prefix, unique_id_,
+         (void *)this, stack_bottom(), stack_top(),
+         stack_top() - stack_bottom(), tls_begin(), tls_end());
 }
 
 static u32 xorshift(u32 state) {
@@ -124,17 +125,21 @@ static u32 xorshift(u32 state) {
 // Generate a (pseudo-)random non-zero tag.
 tag_t Thread::GenerateRandomTag(uptr num_bits) {
   DCHECK_GT(num_bits, 0);
-  if (tagging_disabled_) return 0;
+  if (tagging_disabled_)
+    return 0;
   tag_t tag;
   const uptr tag_mask = (1ULL << num_bits) - 1;
   do {
     if (flags()->random_tags) {
-      if (!random_buffer_)
+      if (!random_buffer_) {
+        EnsureRandomStateInited();
         random_buffer_ = random_state_ = xorshift(random_state_);
+      }
       CHECK(random_buffer_);
       tag = random_buffer_ & tag_mask;
       random_buffer_ >>= num_bits;
     } else {
+      EnsureRandomStateInited();
       random_state_ += 1;
       tag = random_state_ & tag_mask;
     }
diff --git a/compiler-rt/lib/hwasan/hwasan_thread.h b/compiler-rt/lib/hwasan/hwasan_thread.h
index 9f20afe1dc76..3db7c1a9454f 100644
--- a/compiler-rt/lib/hwasan/hwasan_thread.h
+++ b/compiler-rt/lib/hwasan/hwasan_thread.h
@@ -28,12 +28,17 @@ class Thread {
 
   void Init(uptr stack_buffer_start, uptr stack_buffer_size,
             const InitState *state = nullptr);
-  void InitRandomState();
+
   void InitStackAndTls(const InitState *state = nullptr);
 
   // Must be called from the thread itself.
   void InitStackRingBuffer(uptr stack_buffer_start, uptr stack_buffer_size);
 
+  inline void EnsureRandomStateInited() {
+    if (UNLIKELY(!random_state_inited_))
+      InitRandomState();
+  }
+
   void Destroy();
 
   uptr stack_top() { return stack_top_; }
@@ -70,6 +75,7 @@ class Thread {
   // via mmap() and *must* be valid in zero-initialized state.
   void ClearShadowForThreadStackAndTLS();
   void Print(const char *prefix);
+  void InitRandomState();
   uptr vfork_spill_;
   uptr stack_top_;
   uptr stack_bottom_;
@@ -89,6 +95,8 @@ class Thread {
 
   bool announced_;
 
+  bool random_state_inited_;  // Whether InitRandomState() has been called.
+
   friend struct ThreadListHead;
 };
 
diff --git a/compiler-rt/lib/hwasan/hwasan_type_test.cpp b/compiler-rt/lib/hwasan/hwasan_type_test.cpp
index 8cff495bae15..5307073fb40b 100644
--- a/compiler-rt/lib/hwasan/hwasan_type_test.cpp
+++ b/compiler-rt/lib/hwasan/hwasan_type_test.cpp
@@ -19,7 +19,7 @@
 #define CHECK_TYPE_SIZE_FITS(TYPE) \
   COMPILER_CHECK(sizeof(__hw_##TYPE) <= sizeof(TYPE))
 
-#if HWASAN_WITH_INTERCEPTORS && defined(__aarch64__)
+#if HWASAN_WITH_INTERCEPTORS
 CHECK_TYPE_SIZE_FITS(jmp_buf);
 CHECK_TYPE_SIZE_FITS(sigjmp_buf);
 #endif
diff --git a/compiler-rt/lib/interception/interception_win.cpp b/compiler-rt/lib/interception/interception_win.cpp
index 98bc756ae53a..38b8c058246a 100644
--- a/compiler-rt/lib/interception/interception_win.cpp
+++ b/compiler-rt/lib/interception/interception_win.cpp
@@ -56,7 +56,7 @@
 //                                      tramp:  jmp QWORD [addr]
 //                                       addr:  .bytes <hook>
 //
-//    Note: <real> is equilavent to <label>.
+//    Note: <real> is equivalent to <label>.
 //
 // 3) HotPatch
 //
@@ -398,8 +398,42 @@ static uptr AllocateMemoryForTrampoline(uptr image_address, size_t size) {
   return allocated_space;
 }
 
+// The following prologues cannot be patched because of the short jump
+// jumping to the patching region.
+
+// ntdll!wcslen in Win11
+//   488bc1          mov     rax,rcx
+//   0fb710          movzx   edx,word ptr [rax]
+//   4883c002        add     rax,2
+//   6685d2          test    dx,dx
+//   75f4            jne     -12
+static const u8 kPrologueWithShortJump1[] = {
+    0x48, 0x8b, 0xc1, 0x0f, 0xb7, 0x10, 0x48, 0x83,
+    0xc0, 0x02, 0x66, 0x85, 0xd2, 0x75, 0xf4,
+};
+
+// ntdll!strrchr in Win11
+//   4c8bc1          mov     r8,rcx
+//   8a01            mov     al,byte ptr [rcx]
+//   48ffc1          inc     rcx
+//   84c0            test    al,al
+//   75f7            jne     -9
+static const u8 kPrologueWithShortJump2[] = {
+    0x4c, 0x8b, 0xc1, 0x8a, 0x01, 0x48, 0xff, 0xc1,
+    0x84, 0xc0, 0x75, 0xf7,
+};
+
 // Returns 0 on error.
 static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
+#if SANITIZER_WINDOWS64
+  if (memcmp((u8*)address, kPrologueWithShortJump1,
+             sizeof(kPrologueWithShortJump1)) == 0 ||
+      memcmp((u8*)address, kPrologueWithShortJump2,
+             sizeof(kPrologueWithShortJump2)) == 0) {
+    return 0;
+  }
+#endif
+
   switch (*(u64*)address) {
     case 0x90909090909006EB:  // stub: jmp over 6 x nop.
       return 8;
@@ -477,6 +511,14 @@ static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
     case 0xA1:  // A1 XX XX XX XX XX XX XX XX :
                 //   movabs eax, dword ptr ds:[XXXXXXXX]
       return 9;
+
+    case 0x83:
+      const u8 next_byte = *(u8*)(address + 1);
+      const u8 mod = next_byte >> 6;
+      const u8 rm = next_byte & 7;
+      if (mod == 1 && rm == 4)
+        return 5;  // 83 ModR/M SIB Disp8 Imm8
+                   //   add|or|adc|sbb|and|sub|xor|cmp [r+disp8], imm8
   }
 
   switch (*(u16*)address) {
@@ -493,6 +535,8 @@ static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
     case 0x5641:  // push r14
     case 0x5741:  // push r15
     case 0x9066:  // Two-byte NOP
+    case 0xc084:  // test al, al
+    case 0x018a:  // mov al, byte ptr [rcx]
       return 2;
 
     case 0x058B:  // 8B 05 XX XX XX XX : mov eax, dword ptr [XX XX XX XX]
@@ -509,6 +553,7 @@ static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
     case 0xd12b48:    // 48 2b d1 : sub rdx, rcx
     case 0x07c1f6:    // f6 c1 07 : test cl, 0x7
     case 0xc98548:    // 48 85 C9 : test rcx, rcx
+    case 0xd28548:    // 48 85 d2 : test rdx, rdx
     case 0xc0854d:    // 4d 85 c0 : test r8, r8
     case 0xc2b60f:    // 0f b6 c2 : movzx eax, dl
     case 0xc03345:    // 45 33 c0 : xor r8d, r8d
@@ -522,6 +567,7 @@ static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
     case 0xca2b48:    // 48 2b ca : sub rcx, rdx
     case 0x10b70f:    // 0f b7 10 : movzx edx, WORD PTR [rax]
     case 0xc00b4d:    // 3d 0b c0 : or r8, r8
+    case 0xc08b41:    // 41 8b c0 : mov eax, r8d
     case 0xd18b48:    // 48 8b d1 : mov rdx, rcx
     case 0xdc8b4c:    // 4c 8b dc : mov r11, rsp
     case 0xd18b4c:    // 4c 8b d1 : mov r10, rcx
diff --git a/compiler-rt/lib/lsan/lsan_allocator.h b/compiler-rt/lib/lsan/lsan_allocator.h
index 9d763789154f..45c6ac406f8a 100644
--- a/compiler-rt/lib/lsan/lsan_allocator.h
+++ b/compiler-rt/lib/lsan/lsan_allocator.h
@@ -50,7 +50,7 @@ struct ChunkMetadata {
 };
 
 #if defined(__mips64) || defined(__aarch64__) || defined(__i386__) || \
-    defined(__arm__) || SANITIZER_RISCV64
+    defined(__arm__) || SANITIZER_RISCV64 || defined(__hexagon__)
 template <typename AddressSpaceViewTy>
 struct AP32 {
   static const uptr kSpaceBeg = 0;
diff --git a/compiler-rt/lib/lsan/lsan_common.cpp b/compiler-rt/lib/lsan/lsan_common.cpp
index 74400d2e8426..308dbb3e41da 100644
--- a/compiler-rt/lib/lsan/lsan_common.cpp
+++ b/compiler-rt/lib/lsan/lsan_common.cpp
@@ -30,7 +30,7 @@ namespace __lsan {
 
 // This mutex is used to prevent races between DoLeakCheck and IgnoreObject, and
 // also to protect the global list of root regions.
-BlockingMutex global_mutex(LINKER_INITIALIZED);
+Mutex global_mutex;
 
 Flags lsan_flags;
 
@@ -79,7 +79,8 @@ class LeakSuppressionContext {
                          int suppression_types_num)
       : context(supprression_types, suppression_types_num) {}
 
-  Suppression *GetSuppressionForStack(u32 stack_trace_id);
+  Suppression *GetSuppressionForStack(u32 stack_trace_id,
+                                      const StackTrace &stack);
 
   const InternalMmapVector<u32> &GetSortedSuppressedStacks() {
     if (!suppressed_stacks_sorted) {
@@ -130,18 +131,13 @@ static LeakSuppressionContext *GetSuppressionContext() {
   return suppression_ctx;
 }
 
-static InternalMmapVector<RootRegion> *root_regions;
+static InternalMmapVectorNoCtor<RootRegion> root_regions;
 
-InternalMmapVector<RootRegion> const *GetRootRegions() { return root_regions; }
-
-void InitializeRootRegions() {
-  CHECK(!root_regions);
-  ALIGNED(64) static char placeholder[sizeof(InternalMmapVector<RootRegion>)];
-  root_regions = new (placeholder) InternalMmapVector<RootRegion>();
+InternalMmapVectorNoCtor<RootRegion> const *GetRootRegions() {
+  return &root_regions;
 }
 
 void InitCommonLsan() {
-  InitializeRootRegions();
   if (common_flags()->detect_leaks) {
     // Initialization which can fail or print warnings should only be done if
     // LSan is actually enabled.
@@ -187,7 +183,8 @@ void ScanRangeForPointers(uptr begin, uptr end,
                           const char *region_type, ChunkTag tag) {
   CHECK(tag == kReachable || tag == kIndirectlyLeaked);
   const uptr alignment = flags()->pointer_alignment();
-  LOG_POINTERS("Scanning %s range %p-%p.\n", region_type, begin, end);
+  LOG_POINTERS("Scanning %s range %p-%p.\n", region_type, (void *)begin,
+               (void *)end);
   uptr pp = begin;
   if (pp % alignment)
     pp = pp + alignment - pp % alignment;
@@ -206,13 +203,15 @@ void ScanRangeForPointers(uptr begin, uptr end,
       LOG_POINTERS(
           "%p is poisoned: ignoring %p pointing into chunk %p-%p of size "
           "%zu.\n",
-          pp, p, chunk, chunk + m.requested_size(), m.requested_size());
+          (void *)pp, p, (void *)chunk, (void *)(chunk + m.requested_size()),
+          m.requested_size());
       continue;
     }
 
     m.set_tag(tag);
-    LOG_POINTERS("%p: found %p pointing into chunk %p-%p of size %zu.\n", pp, p,
-                 chunk, chunk + m.requested_size(), m.requested_size());
+    LOG_POINTERS("%p: found %p pointing into chunk %p-%p of size %zu.\n",
+                 (void *)pp, p, (void *)chunk,
+                 (void *)(chunk + m.requested_size()), m.requested_size());
     if (frontier)
       frontier->push_back(chunk);
   }
@@ -280,7 +279,7 @@ static void ProcessThreads(SuspendedThreadsList const &suspended_threads,
   InternalMmapVector<uptr> registers;
   for (uptr i = 0; i < suspended_threads.ThreadCount(); i++) {
     tid_t os_id = static_cast<tid_t>(suspended_threads.GetThreadID(i));
-    LOG_THREADS("Processing thread %d.\n", os_id);
+    LOG_THREADS("Processing thread %llu.\n", os_id);
     uptr stack_begin, stack_end, tls_begin, tls_end, cache_begin, cache_end;
     DTLS *dtls;
     bool thread_found = GetThreadRangesLocked(os_id, &stack_begin, &stack_end,
@@ -289,14 +288,14 @@ static void ProcessThreads(SuspendedThreadsList const &suspended_threads,
     if (!thread_found) {
       // If a thread can't be found in the thread registry, it's probably in the
       // process of destruction. Log this event and move on.
-      LOG_THREADS("Thread %d not found in registry.\n", os_id);
+      LOG_THREADS("Thread %llu not found in registry.\n", os_id);
       continue;
     }
     uptr sp;
     PtraceRegistersStatus have_registers =
         suspended_threads.GetRegistersAndSP(i, &registers, &sp);
     if (have_registers != REGISTERS_AVAILABLE) {
-      Report("Unable to get registers from thread %d.\n", os_id);
+      Report("Unable to get registers from thread %llu.\n", os_id);
       // If unable to get SP, consider the entire stack to be reachable unless
       // GetRegistersAndSP failed with ESRCH.
       if (have_registers == REGISTERS_UNAVAILABLE_FATAL) continue;
@@ -312,7 +311,8 @@ static void ProcessThreads(SuspendedThreadsList const &suspended_threads,
     }
 
     if (flags()->use_stacks) {
-      LOG_THREADS("Stack at %p-%p (SP = %p).\n", stack_begin, stack_end, sp);
+      LOG_THREADS("Stack at %p-%p (SP = %p).\n", (void *)stack_begin,
+                  (void *)stack_end, (void *)sp);
       if (sp < stack_begin || sp >= stack_end) {
         // SP is outside the recorded stack range (e.g. the thread is running a
         // signal handler on alternate stack, or swapcontext was used).
@@ -326,7 +326,7 @@ static void ProcessThreads(SuspendedThreadsList const &suspended_threads,
           stack_begin += page_size;
         }
         LOG_THREADS("Skipped %d guard page(s) to obtain stack %p-%p.\n",
-                    skipped, stack_begin, stack_end);
+                    skipped, (void *)stack_begin, (void *)stack_end);
       } else {
         // Shrink the stack range to ignore out-of-scope values.
         stack_begin = sp;
@@ -338,7 +338,7 @@ static void ProcessThreads(SuspendedThreadsList const &suspended_threads,
 
     if (flags()->use_tls) {
       if (tls_begin) {
-        LOG_THREADS("TLS at %p-%p.\n", tls_begin, tls_end);
+        LOG_THREADS("TLS at %p-%p.\n", (void *)tls_begin, (void *)tls_end);
         // If the tls and cache ranges don't overlap, scan full tls range,
         // otherwise, only scan the non-overlapping portions
         if (cache_begin == cache_end || tls_end < cache_begin ||
@@ -372,7 +372,8 @@ static void ProcessThreads(SuspendedThreadsList const &suspended_threads,
           uptr dtls_beg = dtv.beg;
           uptr dtls_end = dtls_beg + dtv.size;
           if (dtls_beg < dtls_end) {
-            LOG_THREADS("DTLS %zu at %p-%p.\n", id, dtls_beg, dtls_end);
+            LOG_THREADS("DTLS %d at %p-%p.\n", id, (void *)dtls_beg,
+                        (void *)dtls_end);
             ScanRangeForPointers(dtls_beg, dtls_end, frontier, "DTLS",
                                  kReachable);
           }
@@ -380,7 +381,7 @@ static void ProcessThreads(SuspendedThreadsList const &suspended_threads,
       } else {
         // We are handling a thread with DTLS under destruction. Log about
         // this and continue.
-        LOG_THREADS("Thread %d has DTLS under destruction.\n", os_id);
+        LOG_THREADS("Thread %llu has DTLS under destruction.\n", os_id);
       }
 #endif
     }
@@ -398,8 +399,9 @@ void ScanRootRegion(Frontier *frontier, const RootRegion &root_region,
   uptr intersection_end = Min(region_end, root_region.begin + root_region.size);
   if (intersection_begin >= intersection_end) return;
   LOG_POINTERS("Root region %p-%p intersects with mapped region %p-%p (%s)\n",
-               root_region.begin, root_region.begin + root_region.size,
-               region_begin, region_end,
+               (void *)root_region.begin,
+               (void *)(root_region.begin + root_region.size),
+               (void *)region_begin, (void *)region_end,
                is_readable ? "readable" : "unreadable");
   if (is_readable)
     ScanRangeForPointers(intersection_begin, intersection_end, frontier, "ROOT",
@@ -419,10 +421,8 @@ static void ProcessRootRegion(Frontier *frontier,
 // Scans root regions for heap pointers.
 static void ProcessRootRegions(Frontier *frontier) {
   if (!flags()->use_root_regions) return;
-  CHECK(root_regions);
-  for (uptr i = 0; i < root_regions->size(); i++) {
-    ProcessRootRegion(frontier, (*root_regions)[i]);
-  }
+  for (uptr i = 0; i < root_regions.size(); i++)
+    ProcessRootRegion(frontier, root_regions[i]);
 }
 
 static void FloodFillTag(Frontier *frontier, ChunkTag tag) {
@@ -459,8 +459,8 @@ static void IgnoredSuppressedCb(uptr chunk, void *arg) {
   if (idx >= suppressed.size() || m.stack_trace_id() != suppressed[idx])
     return;
 
-  LOG_POINTERS("Suppressed: chunk %p-%p of size %zu.\n", chunk,
-               chunk + m.requested_size(), m.requested_size());
+  LOG_POINTERS("Suppressed: chunk %p-%p of size %zu.\n", (void *)chunk,
+               (void *)(chunk + m.requested_size()), m.requested_size());
   m.set_tag(kIgnored);
 }
 
@@ -471,15 +471,13 @@ static void CollectIgnoredCb(uptr chunk, void *arg) {
   chunk = GetUserBegin(chunk);
   LsanMetadata m(chunk);
   if (m.allocated() && m.tag() == kIgnored) {
-    LOG_POINTERS("Ignored: chunk %p-%p of size %zu.\n",
-                 chunk, chunk + m.requested_size(), m.requested_size());
+    LOG_POINTERS("Ignored: chunk %p-%p of size %zu.\n", (void *)chunk,
+                 (void *)(chunk + m.requested_size()), m.requested_size());
     reinterpret_cast<Frontier *>(arg)->push_back(chunk);
   }
 }
 
-static uptr GetCallerPC(u32 stack_id, StackDepotReverseMap *map) {
-  CHECK(stack_id);
-  StackTrace stack = map->Get(stack_id);
+static uptr GetCallerPC(const StackTrace &stack) {
   // The top frame is our malloc/calloc/etc. The next frame is the caller.
   if (stack.size >= 2)
     return stack.trace[1];
@@ -488,7 +486,6 @@ static uptr GetCallerPC(u32 stack_id, StackDepotReverseMap *map) {
 
 struct InvalidPCParam {
   Frontier *frontier;
-  StackDepotReverseMap *stack_depot_reverse_map;
   bool skip_linker_allocations;
 };
 
@@ -503,7 +500,7 @@ static void MarkInvalidPCCb(uptr chunk, void *arg) {
     u32 stack_id = m.stack_trace_id();
     uptr caller_pc = 0;
     if (stack_id > 0)
-      caller_pc = GetCallerPC(stack_id, param->stack_depot_reverse_map);
+      caller_pc = GetCallerPC(StackDepotGet(stack_id));
     // If caller_pc is unknown, this chunk may be allocated in a coroutine. Mark
     // it as reachable, as we can't properly report its allocation stack anyway.
     if (caller_pc == 0 || (param->skip_linker_allocations &&
@@ -534,11 +531,9 @@ static void MarkInvalidPCCb(uptr chunk, void *arg) {
 // which we don't care about).
 // On all other platforms, this simply checks to ensure that the caller pc is
 // valid before reporting chunks as leaked.
-void ProcessPC(Frontier *frontier) {
-  StackDepotReverseMap stack_depot_reverse_map;
+static void ProcessPC(Frontier *frontier) {
   InvalidPCParam arg;
   arg.frontier = frontier;
-  arg.stack_depot_reverse_map = &stack_depot_reverse_map;
   arg.skip_linker_allocations =
       flags()->use_tls && flags()->use_ld_allocations && GetLinker() != nullptr;
   ForEachChunk(MarkInvalidPCCb, &arg);
@@ -584,11 +579,6 @@ static void ResetTagsCb(uptr chunk, void *arg) {
     m.set_tag(kDirectlyLeaked);
 }
 
-static void PrintStackTraceById(u32 stack_trace_id) {
-  CHECK(stack_trace_id);
-  StackDepotGet(stack_trace_id).Print();
-}
-
 // ForEachChunk callback. Aggregates information about unreachable chunks into
 // a LeakReport.
 static void CollectLeaksCb(uptr chunk, void *arg) {
@@ -598,16 +588,7 @@ static void CollectLeaksCb(uptr chunk, void *arg) {
   LsanMetadata m(chunk);
   if (!m.allocated()) return;
   if (m.tag() == kDirectlyLeaked || m.tag() == kIndirectlyLeaked) {
-    u32 resolution = flags()->resolution;
-    u32 stack_trace_id = 0;
-    if (resolution > 0) {
-      StackTrace stack = StackDepotGet(m.stack_trace_id());
-      stack.size = Min(stack.size, resolution);
-      stack_trace_id = StackDepotPut(stack);
-    } else {
-      stack_trace_id = m.stack_trace_id();
-    }
-    leak_report->AddLeakedChunk(chunk, stack_trace_id, m.requested_size(),
+    leak_report->AddLeakedChunk(chunk, m.stack_trace_id(), m.requested_size(),
                                 m.tag());
   }
 }
@@ -635,8 +616,9 @@ static void ReportIfNotSuspended(ThreadContextBase *tctx, void *arg) {
   if (tctx->status == ThreadStatusRunning) {
     uptr i = InternalLowerBound(suspended_threads, tctx->os_id);
     if (i >= suspended_threads.size() || suspended_threads[i] != tctx->os_id)
-      Report("Running thread %d was not suspended. False leaks are possible.\n",
-             tctx->os_id);
+      Report(
+          "Running thread %llu was not suspended. False leaks are possible.\n",
+          tctx->os_id);
   }
 }
 
@@ -742,7 +724,7 @@ static bool has_reported_leaks = false;
 bool HasReportedLeaks() { return has_reported_leaks; }
 
 void DoLeakCheck() {
-  BlockingMutexLock l(&global_mutex);
+  Lock l(&global_mutex);
   static bool already_done;
   if (already_done) return;
   already_done = true;
@@ -751,7 +733,7 @@ void DoLeakCheck() {
 }
 
 static int DoRecoverableLeakCheck() {
-  BlockingMutexLock l(&global_mutex);
+  Lock l(&global_mutex);
   bool have_leaks = CheckForLeaks();
   return have_leaks ? 1 : 0;
 }
@@ -780,9 +762,8 @@ Suppression *LeakSuppressionContext::GetSuppressionForAddr(uptr addr) {
 }
 
 Suppression *LeakSuppressionContext::GetSuppressionForStack(
-    u32 stack_trace_id) {
+    u32 stack_trace_id, const StackTrace &stack) {
   LazyInit();
-  StackTrace stack = StackDepotGet(stack_trace_id);
   for (uptr i = 0; i < stack.size; i++) {
     Suppression *s = GetSuppressionForAddr(
         StackTrace::GetPreviousInstructionPc(stack.trace[i]));
@@ -807,6 +788,13 @@ const uptr kMaxLeaksConsidered = 5000;
 void LeakReport::AddLeakedChunk(uptr chunk, u32 stack_trace_id,
                                 uptr leaked_size, ChunkTag tag) {
   CHECK(tag == kDirectlyLeaked || tag == kIndirectlyLeaked);
+
+  if (u32 resolution = flags()->resolution) {
+    StackTrace stack = StackDepotGet(stack_trace_id);
+    stack.size = Min(stack.size, resolution);
+    stack_trace_id = StackDepotPut(stack);
+  }
+
   bool is_directly_leaked = (tag == kDirectlyLeaked);
   uptr i;
   for (i = 0; i < leaks_.size(); i++) {
@@ -869,7 +857,8 @@ void LeakReport::PrintReportForLeak(uptr index) {
          leaks_[index].total_size, leaks_[index].hit_count);
   Printf("%s", d.Default());
 
-  PrintStackTraceById(leaks_[index].stack_trace_id);
+  CHECK(leaks_[index].stack_trace_id);
+  StackDepotGet(leaks_[index].stack_trace_id).Print();
 
   if (flags()->report_objects) {
     Printf("Objects leaked above:\n");
@@ -882,7 +871,7 @@ void LeakReport::PrintLeakedObjectsForLeak(uptr index) {
   u32 leak_id = leaks_[index].id;
   for (uptr j = 0; j < leaked_objects_.size(); j++) {
     if (leaked_objects_[j].leak_id == leak_id)
-      Printf("%p (%zu bytes)\n", leaked_objects_[j].addr,
+      Printf("%p (%zu bytes)\n", (void *)leaked_objects_[j].addr,
              leaked_objects_[j].size);
   }
 }
@@ -905,8 +894,8 @@ uptr LeakReport::ApplySuppressions() {
   LeakSuppressionContext *suppressions = GetSuppressionContext();
   uptr new_suppressions = false;
   for (uptr i = 0; i < leaks_.size(); i++) {
-    Suppression *s =
-        suppressions->GetSuppressionForStack(leaks_[i].stack_trace_id);
+    Suppression *s = suppressions->GetSuppressionForStack(
+        leaks_[i].stack_trace_id, StackDepotGet(leaks_[i].stack_trace_id));
     if (s) {
       s->weight += leaks_[i].total_size;
       atomic_store_relaxed(&s->hit_count, atomic_load_relaxed(&s->hit_count) +
@@ -954,7 +943,7 @@ void __lsan_ignore_object(const void *p) {
     return;
   // Cannot use PointsIntoChunk or LsanMetadata here, since the allocator is not
   // locked.
-  BlockingMutexLock l(&global_mutex);
+  Lock l(&global_mutex);
   IgnoreObjectResult res = IgnoreObjectLocked(p);
   if (res == kIgnoreObjectInvalid)
     VReport(1, "__lsan_ignore_object(): no heap object found at %p", p);
@@ -969,34 +958,32 @@ void __lsan_ignore_object(const void *p) {
 SANITIZER_INTERFACE_ATTRIBUTE
 void __lsan_register_root_region(const void *begin, uptr size) {
 #if CAN_SANITIZE_LEAKS
-  BlockingMutexLock l(&global_mutex);
-  CHECK(root_regions);
+  Lock l(&global_mutex);
   RootRegion region = {reinterpret_cast<uptr>(begin), size};
-  root_regions->push_back(region);
-  VReport(1, "Registered root region at %p of size %llu\n", begin, size);
+  root_regions.push_back(region);
+  VReport(1, "Registered root region at %p of size %zu\n", begin, size);
 #endif // CAN_SANITIZE_LEAKS
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 void __lsan_unregister_root_region(const void *begin, uptr size) {
 #if CAN_SANITIZE_LEAKS
-  BlockingMutexLock l(&global_mutex);
-  CHECK(root_regions);
+  Lock l(&global_mutex);
   bool removed = false;
-  for (uptr i = 0; i < root_regions->size(); i++) {
-    RootRegion region = (*root_regions)[i];
+  for (uptr i = 0; i < root_regions.size(); i++) {
+    RootRegion region = root_regions[i];
     if (region.begin == reinterpret_cast<uptr>(begin) && region.size == size) {
       removed = true;
-      uptr last_index = root_regions->size() - 1;
-      (*root_regions)[i] = (*root_regions)[last_index];
-      root_regions->pop_back();
-      VReport(1, "Unregistered root region at %p of size %llu\n", begin, size);
+      uptr last_index = root_regions.size() - 1;
+      root_regions[i] = root_regions[last_index];
+      root_regions.pop_back();
+      VReport(1, "Unregistered root region at %p of size %zu\n", begin, size);
       break;
     }
   }
   if (!removed) {
     Report(
-        "__lsan_unregister_root_region(): region at %p of size %llu has not "
+        "__lsan_unregister_root_region(): region at %p of size %zu has not "
         "been registered.\n",
         begin, size);
     Die();
diff --git a/compiler-rt/lib/lsan/lsan_common.h b/compiler-rt/lib/lsan/lsan_common.h
index 776ca60b1e97..f9b55e4e8006 100644
--- a/compiler-rt/lib/lsan/lsan_common.h
+++ b/compiler-rt/lib/lsan/lsan_common.h
@@ -18,6 +18,7 @@
 #include "sanitizer_common/sanitizer_common.h"
 #include "sanitizer_common/sanitizer_internal_defs.h"
 #include "sanitizer_common/sanitizer_platform.h"
+#include "sanitizer_common/sanitizer_stackdepot.h"
 #include "sanitizer_common/sanitizer_stoptheworld.h"
 #include "sanitizer_common/sanitizer_symbolizer.h"
 
@@ -139,7 +140,7 @@ struct CheckForLeaksParam {
   bool success = false;
 };
 
-InternalMmapVector<RootRegion> const *GetRootRegions();
+InternalMmapVectorNoCtor<RootRegion> const *GetRootRegions();
 void ScanRootRegion(Frontier *frontier, RootRegion const &region,
                     uptr region_begin, uptr region_end, bool is_readable);
 void ForEachExtraStackRangeCb(uptr begin, uptr end, void* arg);
@@ -279,6 +280,13 @@ int __lsan_is_turned_off();
 
 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 const char *__lsan_default_suppressions();
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __lsan_register_root_region(const void *p, __lsan::uptr size);
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __lsan_unregister_root_region(const void *p, __lsan::uptr size);
+
 }  // extern "C"
 
 #endif  // LSAN_COMMON_H
diff --git a/compiler-rt/lib/lsan/lsan_common_mac.cpp b/compiler-rt/lib/lsan/lsan_common_mac.cpp
index 8516a176eb46..4301dcc615d7 100644
--- a/compiler-rt/lib/lsan/lsan_common_mac.cpp
+++ b/compiler-rt/lib/lsan/lsan_common_mac.cpp
@@ -149,7 +149,7 @@ void ProcessPlatformSpecificAllocations(Frontier *frontier) {
   kern_return_t err = KERN_SUCCESS;
   mach_msg_type_number_t count = VM_REGION_SUBMAP_INFO_COUNT_64;
 
-  InternalMmapVector<RootRegion> const *root_regions = GetRootRegions();
+  InternalMmapVectorNoCtor<RootRegion> const *root_regions = GetRootRegions();
 
   while (err == KERN_SUCCESS) {
     struct vm_region_submap_info_64 info;
diff --git a/compiler-rt/lib/lsan/lsan_fuchsia.cpp b/compiler-rt/lib/lsan/lsan_fuchsia.cpp
index 40e65c6fb729..cb9f62a0341c 100644
--- a/compiler-rt/lib/lsan/lsan_fuchsia.cpp
+++ b/compiler-rt/lib/lsan/lsan_fuchsia.cpp
@@ -62,7 +62,7 @@ void InitializeMainThread() {
   OnCreatedArgs args;
   __sanitizer::GetThreadStackTopAndBottom(true, &args.stack_end,
                                           &args.stack_begin);
-  u32 tid = ThreadCreate(0, GetThreadSelf(), true, &args);
+  u32 tid = ThreadCreate(kMainTid, true, &args);
   CHECK_EQ(tid, 0);
   ThreadStart(tid);
 }
@@ -86,14 +86,13 @@ void GetAllThreadAllocatorCachesLocked(InternalMmapVector<uptr> *caches) {
 void *__sanitizer_before_thread_create_hook(thrd_t thread, bool detached,
                                             const char *name, void *stack_base,
                                             size_t stack_size) {
-  uptr user_id = reinterpret_cast<uptr>(thread);
   ENSURE_LSAN_INITED;
   EnsureMainThreadIDIsCorrect();
   OnCreatedArgs args;
   args.stack_begin = reinterpret_cast<uptr>(stack_base);
   args.stack_end = args.stack_begin + stack_size;
   u32 parent_tid = GetCurrentThread();
-  u32 tid = ThreadCreate(parent_tid, user_id, detached, &args);
+  u32 tid = ThreadCreate(parent_tid, detached, &args);
   return reinterpret_cast<void *>(static_cast<uptr>(tid));
 }
 
diff --git a/compiler-rt/lib/lsan/lsan_interceptors.cpp b/compiler-rt/lib/lsan/lsan_interceptors.cpp
index 90a90a56c54c..ee723f210c9d 100644
--- a/compiler-rt/lib/lsan/lsan_interceptors.cpp
+++ b/compiler-rt/lib/lsan/lsan_interceptors.cpp
@@ -13,6 +13,7 @@
 
 #include "interception/interception.h"
 #include "sanitizer_common/sanitizer_allocator.h"
+#include "sanitizer_common/sanitizer_allocator_dlsym.h"
 #include "sanitizer_common/sanitizer_allocator_report.h"
 #include "sanitizer_common/sanitizer_atomic.h"
 #include "sanitizer_common/sanitizer_common.h"
@@ -43,6 +44,22 @@ int pthread_key_create(unsigned *key, void (*destructor)(void* v));
 int pthread_setspecific(unsigned key, const void *v);
 }
 
+struct DlsymAlloc : DlSymAllocator<DlsymAlloc> {
+  static bool UseImpl() { return lsan_init_is_running; }
+  static void OnAllocate(const void *ptr, uptr size) {
+#if CAN_SANITIZE_LEAKS
+    // Suppress leaks from dlerror(). Previously dlsym hack on global array was
+    // used by leak sanitizer as a root region.
+    __lsan_register_root_region(ptr, size);
+#endif
+  }
+  static void OnFree(const void *ptr, uptr size) {
+#if CAN_SANITIZE_LEAKS
+    __lsan_unregister_root_region(ptr, size);
+#endif
+  }
+};
+
 ///// Malloc/free interceptors. /////
 
 namespace std {
@@ -52,41 +69,34 @@ namespace std {
 
 #if !SANITIZER_MAC
 INTERCEPTOR(void*, malloc, uptr size) {
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Allocate(size);
   ENSURE_LSAN_INITED;
   GET_STACK_TRACE_MALLOC;
   return lsan_malloc(size, stack);
 }
 
 INTERCEPTOR(void, free, void *p) {
+  if (DlsymAlloc::PointerIsMine(p))
+    return DlsymAlloc::Free(p);
   ENSURE_LSAN_INITED;
   lsan_free(p);
 }
 
 INTERCEPTOR(void*, calloc, uptr nmemb, uptr size) {
-  // This hack is not required for Fuchsia because there are no dlsym calls
-  // involved in setting up interceptors.
-#if !SANITIZER_FUCHSIA
-  if (lsan_init_is_running) {
-    // Hack: dlsym calls calloc before REAL(calloc) is retrieved from dlsym.
-    const uptr kCallocPoolSize = 1024;
-    static uptr calloc_memory_for_dlsym[kCallocPoolSize];
-    static uptr allocated;
-    uptr size_in_words = ((nmemb * size) + kWordSize - 1) / kWordSize;
-    void *mem = (void*)&calloc_memory_for_dlsym[allocated];
-    allocated += size_in_words;
-    CHECK(allocated < kCallocPoolSize);
-    return mem;
-  }
-#endif  // !SANITIZER_FUCHSIA
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Callocate(nmemb, size);
   ENSURE_LSAN_INITED;
   GET_STACK_TRACE_MALLOC;
   return lsan_calloc(nmemb, size, stack);
 }
 
-INTERCEPTOR(void*, realloc, void *q, uptr size) {
+INTERCEPTOR(void *, realloc, void *ptr, uptr size) {
+  if (DlsymAlloc::Use() || DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Realloc(ptr, size);
   ENSURE_LSAN_INITED;
   GET_STACK_TRACE_MALLOC;
-  return lsan_realloc(q, size, stack);
+  return lsan_realloc(ptr, size, stack);
 }
 
 INTERCEPTOR(void*, reallocarray, void *q, uptr nmemb, uptr size) {
@@ -458,8 +468,7 @@ INTERCEPTOR(int, pthread_create, void *th, void *attr,
     res = REAL(pthread_create)(th, attr, __lsan_thread_start_func, &p);
   }
   if (res == 0) {
-    int tid = ThreadCreate(GetCurrentThread(), *(uptr *)th,
-                           IsStateDetached(detached));
+    int tid = ThreadCreate(GetCurrentThread(), IsStateDetached(detached));
     CHECK_NE(tid, kMainTid);
     atomic_store(&p.tid, tid, memory_order_release);
     while (atomic_load(&p.tid, memory_order_acquire) != 0)
@@ -470,24 +479,6 @@ INTERCEPTOR(int, pthread_create, void *th, void *attr,
   return res;
 }
 
-INTERCEPTOR(int, pthread_join, void *th, void **ret) {
-  ENSURE_LSAN_INITED;
-  int tid = ThreadTid((uptr)th);
-  int res = REAL(pthread_join)(th, ret);
-  if (res == 0)
-    ThreadJoin(tid);
-  return res;
-}
-
-INTERCEPTOR(int, pthread_detach, void *th) {
-  ENSURE_LSAN_INITED;
-  int tid = ThreadTid((uptr)th);
-  int res = REAL(pthread_detach)(th);
-  if (res == 0)
-    ThreadDetach(tid);
-  return res;
-}
-
 INTERCEPTOR(void, _exit, int status) {
   if (status == 0 && HasReportedLeaks()) status = common_flags()->exitcode;
   REAL(_exit)(status);
@@ -520,8 +511,6 @@ void InitializeInterceptors() {
   LSAN_MAYBE_INTERCEPT_MALLINFO;
   LSAN_MAYBE_INTERCEPT_MALLOPT;
   INTERCEPT_FUNCTION(pthread_create);
-  INTERCEPT_FUNCTION(pthread_detach);
-  INTERCEPT_FUNCTION(pthread_join);
   INTERCEPT_FUNCTION(_exit);
 
   LSAN_MAYBE_INTERCEPT__LWP_EXIT;
diff --git a/compiler-rt/lib/lsan/lsan_mac.cpp b/compiler-rt/lib/lsan/lsan_mac.cpp
index b96893e2801b..10a73f8fa93d 100644
--- a/compiler-rt/lib/lsan/lsan_mac.cpp
+++ b/compiler-rt/lib/lsan/lsan_mac.cpp
@@ -68,7 +68,7 @@ typedef struct {
 ALWAYS_INLINE
 void lsan_register_worker_thread(int parent_tid) {
   if (GetCurrentThread() == kInvalidTid) {
-    u32 tid = ThreadCreate(parent_tid, 0, true);
+    u32 tid = ThreadCreate(parent_tid, true);
     ThreadStart(tid, GetTid());
     SetCurrentThread(tid);
   }
diff --git a/compiler-rt/lib/lsan/lsan_posix.cpp b/compiler-rt/lib/lsan/lsan_posix.cpp
index 5d1c3f6260dd..77118a29f2ea 100644
--- a/compiler-rt/lib/lsan/lsan_posix.cpp
+++ b/compiler-rt/lib/lsan/lsan_posix.cpp
@@ -75,7 +75,7 @@ bool GetThreadRangesLocked(tid_t os_id, uptr *stack_begin, uptr *stack_end,
 }
 
 void InitializeMainThread() {
-  u32 tid = ThreadCreate(kMainTid, 0, true);
+  u32 tid = ThreadCreate(kMainTid, true);
   CHECK_EQ(tid, kMainTid);
   ThreadStart(tid, GetTid());
 }
diff --git a/compiler-rt/lib/lsan/lsan_thread.cpp b/compiler-rt/lib/lsan/lsan_thread.cpp
index 1d224ebca693..ca3dfd03b109 100644
--- a/compiler-rt/lib/lsan/lsan_thread.cpp
+++ b/compiler-rt/lib/lsan/lsan_thread.cpp
@@ -44,8 +44,8 @@ void ThreadContextLsanBase::OnFinished() {
   DTLS_Destroy();
 }
 
-u32 ThreadCreate(u32 parent_tid, uptr user_id, bool detached, void *arg) {
-  return thread_registry->CreateThread(user_id, detached, parent_tid, arg);
+u32 ThreadCreate(u32 parent_tid, bool detached, void *arg) {
+  return thread_registry->CreateThread(0, detached, parent_tid, arg);
 }
 
 void ThreadContextLsanBase::ThreadStart(u32 tid, tid_t os_id,
@@ -68,28 +68,6 @@ ThreadContext *CurrentThreadContext() {
   return (ThreadContext *)thread_registry->GetThreadLocked(GetCurrentThread());
 }
 
-static bool FindThreadByUid(ThreadContextBase *tctx, void *arg) {
-  uptr uid = (uptr)arg;
-  if (tctx->user_id == uid && tctx->status != ThreadStatusInvalid) {
-    return true;
-  }
-  return false;
-}
-
-u32 ThreadTid(uptr uid) {
-  return thread_registry->FindThread(FindThreadByUid, (void *)uid);
-}
-
-void ThreadDetach(u32 tid) {
-  CHECK_NE(tid, kInvalidTid);
-  thread_registry->DetachThread(tid, /* arg */ nullptr);
-}
-
-void ThreadJoin(u32 tid) {
-  CHECK_NE(tid, kInvalidTid);
-  thread_registry->JoinThread(tid, /* arg */ nullptr);
-}
-
 void EnsureMainThreadIDIsCorrect() {
   if (GetCurrentThread() == kMainTid)
     CurrentThreadContext()->os_id = GetTid();
diff --git a/compiler-rt/lib/lsan/lsan_thread.h b/compiler-rt/lib/lsan/lsan_thread.h
index 36643753d019..6ab4172092ae 100644
--- a/compiler-rt/lib/lsan/lsan_thread.h
+++ b/compiler-rt/lib/lsan/lsan_thread.h
@@ -45,11 +45,8 @@ class ThreadContext;
 void InitializeThreadRegistry();
 void InitializeMainThread();
 
-u32 ThreadCreate(u32 tid, uptr uid, bool detached, void *arg = nullptr);
+u32 ThreadCreate(u32 tid, bool detached, void *arg = nullptr);
 void ThreadFinish();
-void ThreadDetach(u32 tid);
-void ThreadJoin(u32 tid);
-u32 ThreadTid(uptr uid);
 
 u32 GetCurrentThread();
 void SetCurrentThread(u32 tid);
diff --git a/compiler-rt/lib/memprof/memprof_allocator.cpp b/compiler-rt/lib/memprof/memprof_allocator.cpp
index 6f01d4dfcb84..696f64d8c324 100644
--- a/compiler-rt/lib/memprof/memprof_allocator.cpp
+++ b/compiler-rt/lib/memprof/memprof_allocator.cpp
@@ -15,6 +15,9 @@
 
 #include "memprof_allocator.h"
 #include "memprof_mapping.h"
+#include "memprof_meminfoblock.h"
+#include "memprof_mibmap.h"
+#include "memprof_rawprofile.h"
 #include "memprof_stack.h"
 #include "memprof_thread.h"
 #include "sanitizer_common/sanitizer_allocator_checks.h"
@@ -25,10 +28,11 @@
 #include "sanitizer_common/sanitizer_flags.h"
 #include "sanitizer_common/sanitizer_internal_defs.h"
 #include "sanitizer_common/sanitizer_list.h"
+#include "sanitizer_common/sanitizer_procmaps.h"
 #include "sanitizer_common/sanitizer_stackdepot.h"
+#include "sanitizer_common/sanitizer_vector.h"
 
 #include <sched.h>
-#include <stdlib.h>
 #include <time.h>
 
 namespace __memprof {
@@ -166,244 +170,6 @@ AllocatorCache *GetAllocatorCache(MemprofThreadLocalMallocStorage *ms) {
   return &ms->allocator_cache;
 }
 
-struct MemInfoBlock {
-  u32 alloc_count;
-  u64 total_access_count, min_access_count, max_access_count;
-  u64 total_size;
-  u32 min_size, max_size;
-  u32 alloc_timestamp, dealloc_timestamp;
-  u64 total_lifetime;
-  u32 min_lifetime, max_lifetime;
-  u32 alloc_cpu_id, dealloc_cpu_id;
-  u32 num_migrated_cpu;
-
-  // Only compared to prior deallocated object currently.
-  u32 num_lifetime_overlaps;
-  u32 num_same_alloc_cpu;
-  u32 num_same_dealloc_cpu;
-
-  u64 data_type_id; // TODO: hash of type name
-
-  MemInfoBlock() : alloc_count(0) {}
-
-  MemInfoBlock(u32 size, u64 access_count, u32 alloc_timestamp,
-               u32 dealloc_timestamp, u32 alloc_cpu, u32 dealloc_cpu)
-      : alloc_count(1), total_access_count(access_count),
-        min_access_count(access_count), max_access_count(access_count),
-        total_size(size), min_size(size), max_size(size),
-        alloc_timestamp(alloc_timestamp), dealloc_timestamp(dealloc_timestamp),
-        total_lifetime(dealloc_timestamp - alloc_timestamp),
-        min_lifetime(total_lifetime), max_lifetime(total_lifetime),
-        alloc_cpu_id(alloc_cpu), dealloc_cpu_id(dealloc_cpu),
-        num_lifetime_overlaps(0), num_same_alloc_cpu(0),
-        num_same_dealloc_cpu(0) {
-    num_migrated_cpu = alloc_cpu_id != dealloc_cpu_id;
-  }
-
-  void Print(u64 id) {
-    u64 p;
-    if (flags()->print_terse) {
-      p = total_size * 100 / alloc_count;
-      Printf("MIB:%llu/%u/%d.%02d/%u/%u/", id, alloc_count, p / 100, p % 100,
-             min_size, max_size);
-      p = total_access_count * 100 / alloc_count;
-      Printf("%d.%02d/%u/%u/", p / 100, p % 100, min_access_count,
-             max_access_count);
-      p = total_lifetime * 100 / alloc_count;
-      Printf("%d.%02d/%u/%u/", p / 100, p % 100, min_lifetime, max_lifetime);
-      Printf("%u/%u/%u/%u\n", num_migrated_cpu, num_lifetime_overlaps,
-             num_same_alloc_cpu, num_same_dealloc_cpu);
-    } else {
-      p = total_size * 100 / alloc_count;
-      Printf("Memory allocation stack id = %llu\n", id);
-      Printf("\talloc_count %u, size (ave/min/max) %d.%02d / %u / %u\n",
-             alloc_count, p / 100, p % 100, min_size, max_size);
-      p = total_access_count * 100 / alloc_count;
-      Printf("\taccess_count (ave/min/max): %d.%02d / %u / %u\n", p / 100,
-             p % 100, min_access_count, max_access_count);
-      p = total_lifetime * 100 / alloc_count;
-      Printf("\tlifetime (ave/min/max): %d.%02d / %u / %u\n", p / 100, p % 100,
-             min_lifetime, max_lifetime);
-      Printf("\tnum migrated: %u, num lifetime overlaps: %u, num same alloc "
-             "cpu: %u, num same dealloc_cpu: %u\n",
-             num_migrated_cpu, num_lifetime_overlaps, num_same_alloc_cpu,
-             num_same_dealloc_cpu);
-    }
-  }
-
-  static void printHeader() {
-    CHECK(flags()->print_terse);
-    Printf("MIB:StackID/AllocCount/AveSize/MinSize/MaxSize/AveAccessCount/"
-           "MinAccessCount/MaxAccessCount/AveLifetime/MinLifetime/MaxLifetime/"
-           "NumMigratedCpu/NumLifetimeOverlaps/NumSameAllocCpu/"
-           "NumSameDeallocCpu\n");
-  }
-
-  void Merge(MemInfoBlock &newMIB) {
-    alloc_count += newMIB.alloc_count;
-
-    total_access_count += newMIB.total_access_count;
-    min_access_count = Min(min_access_count, newMIB.min_access_count);
-    max_access_count = Max(max_access_count, newMIB.max_access_count);
-
-    total_size += newMIB.total_size;
-    min_size = Min(min_size, newMIB.min_size);
-    max_size = Max(max_size, newMIB.max_size);
-
-    total_lifetime += newMIB.total_lifetime;
-    min_lifetime = Min(min_lifetime, newMIB.min_lifetime);
-    max_lifetime = Max(max_lifetime, newMIB.max_lifetime);
-
-    // We know newMIB was deallocated later, so just need to check if it was
-    // allocated before last one deallocated.
-    num_lifetime_overlaps += newMIB.alloc_timestamp < dealloc_timestamp;
-    alloc_timestamp = newMIB.alloc_timestamp;
-    dealloc_timestamp = newMIB.dealloc_timestamp;
-
-    num_same_alloc_cpu += alloc_cpu_id == newMIB.alloc_cpu_id;
-    num_same_dealloc_cpu += dealloc_cpu_id == newMIB.dealloc_cpu_id;
-    alloc_cpu_id = newMIB.alloc_cpu_id;
-    dealloc_cpu_id = newMIB.dealloc_cpu_id;
-  }
-};
-
-static u32 AccessCount = 0;
-static u32 MissCount = 0;
-
-struct SetEntry {
-  SetEntry() : id(0), MIB() {}
-  bool Empty() { return id == 0; }
-  void Print() {
-    CHECK(!Empty());
-    MIB.Print(id);
-  }
-  // The stack id
-  u64 id;
-  MemInfoBlock MIB;
-};
-
-struct CacheSet {
-  enum { kSetSize = 4 };
-
-  void PrintAll() {
-    for (int i = 0; i < kSetSize; i++) {
-      if (Entries[i].Empty())
-        continue;
-      Entries[i].Print();
-    }
-  }
-  void insertOrMerge(u64 new_id, MemInfoBlock &newMIB) {
-    AccessCount++;
-    SetAccessCount++;
-
-    for (int i = 0; i < kSetSize; i++) {
-      auto id = Entries[i].id;
-      // Check if this is a hit or an empty entry. Since we always move any
-      // filled locations to the front of the array (see below), we don't need
-      // to look after finding the first empty entry.
-      if (id == new_id || !id) {
-        if (id == 0) {
-          Entries[i].id = new_id;
-          Entries[i].MIB = newMIB;
-        } else {
-          Entries[i].MIB.Merge(newMIB);
-        }
-        // Assuming some id locality, we try to swap the matching entry
-        // into the first set position.
-        if (i != 0) {
-          auto tmp = Entries[0];
-          Entries[0] = Entries[i];
-          Entries[i] = tmp;
-        }
-        return;
-      }
-    }
-
-    // Miss
-    MissCount++;
-    SetMissCount++;
-
-    // We try to find the entries with the lowest alloc count to be evicted:
-    int min_idx = 0;
-    u64 min_count = Entries[0].MIB.alloc_count;
-    for (int i = 1; i < kSetSize; i++) {
-      CHECK(!Entries[i].Empty());
-      if (Entries[i].MIB.alloc_count < min_count) {
-        min_idx = i;
-        min_count = Entries[i].MIB.alloc_count;
-      }
-    }
-
-    // Print the evicted entry profile information
-    if (!flags()->print_terse)
-      Printf("Evicted:\n");
-    Entries[min_idx].Print();
-
-    // Similar to the hit case, put new MIB in first set position.
-    if (min_idx != 0)
-      Entries[min_idx] = Entries[0];
-    Entries[0].id = new_id;
-    Entries[0].MIB = newMIB;
-  }
-
-  void PrintMissRate(int i) {
-    u64 p = SetAccessCount ? SetMissCount * 10000ULL / SetAccessCount : 0;
-    Printf("Set %d miss rate: %d / %d = %5d.%02d%%\n", i, SetMissCount,
-           SetAccessCount, p / 100, p % 100);
-  }
-
-  SetEntry Entries[kSetSize];
-  u32 SetAccessCount = 0;
-  u32 SetMissCount = 0;
-};
-
-struct MemInfoBlockCache {
-  MemInfoBlockCache() {
-    if (common_flags()->print_module_map)
-      DumpProcessMap();
-    if (flags()->print_terse)
-      MemInfoBlock::printHeader();
-    Sets =
-        (CacheSet *)malloc(sizeof(CacheSet) * flags()->mem_info_cache_entries);
-    Constructed = true;
-  }
-
-  ~MemInfoBlockCache() { free(Sets); }
-
-  void insertOrMerge(u64 new_id, MemInfoBlock &newMIB) {
-    u64 hv = new_id;
-
-    // Use mod method where number of entries should be a prime close to power
-    // of 2.
-    hv %= flags()->mem_info_cache_entries;
-
-    return Sets[hv].insertOrMerge(new_id, newMIB);
-  }
-
-  void PrintAll() {
-    for (int i = 0; i < flags()->mem_info_cache_entries; i++) {
-      Sets[i].PrintAll();
-    }
-  }
-
-  void PrintMissRate() {
-    if (!flags()->print_mem_info_cache_miss_rate)
-      return;
-    u64 p = AccessCount ? MissCount * 10000ULL / AccessCount : 0;
-    Printf("Overall miss rate: %d / %d = %5d.%02d%%\n", MissCount, AccessCount,
-           p / 100, p % 100);
-    if (flags()->print_mem_info_cache_miss_rate_details)
-      for (int i = 0; i < flags()->mem_info_cache_entries; i++)
-        Sets[i].PrintMissRate(i);
-  }
-
-  CacheSet *Sets;
-  // Flag when the Sets have been allocated, in case a deallocation is called
-  // very early before the static init of the Allocator and therefore this table
-  // have completed.
-  bool Constructed = false;
-};
-
 // Accumulates the access count from the shadow for the given pointer and size.
 u64 GetShadowCount(uptr p, u32 size) {
   u64 *shadow = (u64 *)MEM_TO_SHADOW(p);
@@ -454,24 +220,66 @@ struct Allocator {
   uptr max_user_defined_malloc_size;
   atomic_uint8_t rss_limit_exceeded;
 
-  MemInfoBlockCache MemInfoBlockTable;
-  bool destructing;
+  // Holds the mapping of stack ids to MemInfoBlocks.
+  MIBMapTy MIBMap;
+
+  atomic_uint8_t destructing;
+  atomic_uint8_t constructed;
+  bool print_text;
 
   // ------------------- Initialization ------------------------
-  explicit Allocator(LinkerInitialized) : destructing(false) {}
+  explicit Allocator(LinkerInitialized) : print_text(flags()->print_text) {
+    atomic_store_relaxed(&destructing, 0);
+    atomic_store_relaxed(&constructed, 1);
+  }
 
-  ~Allocator() { FinishAndPrint(); }
+  ~Allocator() {
+    atomic_store_relaxed(&destructing, 1);
+    FinishAndWrite();
+  }
+
+  static void PrintCallback(const uptr Key, LockedMemInfoBlock *const &Value,
+                            void *Arg) {
+    SpinMutexLock(&Value->mutex);
+    Value->mib.Print(Key, bool(Arg));
+  }
+
+  void FinishAndWrite() {
+    if (print_text && common_flags()->print_module_map)
+      DumpProcessMap();
 
-  void FinishAndPrint() {
-    if (!flags()->print_terse)
-      Printf("Live on exit:\n");
     allocator.ForceLock();
+
+    InsertLiveBlocks();
+    if (print_text) {
+      MIBMap.ForEach(PrintCallback,
+                     reinterpret_cast<void *>(flags()->print_terse));
+      StackDepotPrintAll();
+    } else {
+      // Serialize the contents to a raw profile. Format documented in
+      // memprof_rawprofile.h.
+      char *Buffer = nullptr;
+
+      MemoryMappingLayout Layout(/*cache_enabled=*/true);
+      u64 BytesSerialized = SerializeToRawProfile(MIBMap, Layout, Buffer);
+      CHECK(Buffer && BytesSerialized && "could not serialize to buffer");
+      report_file.Write(Buffer, BytesSerialized);
+    }
+
+    allocator.ForceUnlock();
+  }
+
+  // Inserts any blocks which have been allocated but not yet deallocated.
+  void InsertLiveBlocks() {
+    if (print_text && !flags()->print_terse)
+      Printf("Live on exit:\n");
+
     allocator.ForEachChunk(
         [](uptr chunk, void *alloc) {
           u64 user_requested_size;
+          Allocator *A = (Allocator *)alloc;
           MemprofChunk *m =
-              ((Allocator *)alloc)
-                  ->GetMemprofChunk((void *)chunk, user_requested_size);
+              A->GetMemprofChunk((void *)chunk, user_requested_size);
           if (!m)
             return;
           uptr user_beg = ((uptr)m) + kChunkHeaderSize;
@@ -479,16 +287,9 @@ struct Allocator {
           long curtime = GetTimestamp();
           MemInfoBlock newMIB(user_requested_size, c, m->timestamp_ms, curtime,
                               m->cpu_id, GetCpuId());
-          ((Allocator *)alloc)
-              ->MemInfoBlockTable.insertOrMerge(m->alloc_context_id, newMIB);
+          InsertOrMerge(m->alloc_context_id, newMIB, A->MIBMap);
         },
         this);
-    allocator.ForceUnlock();
-
-    destructing = true;
-    MemInfoBlockTable.PrintMissRate();
-    MemInfoBlockTable.PrintAll();
-    StackDepotPrintAll();
   }
 
   void InitLinkerInitialized() {
@@ -541,8 +342,7 @@ struct Allocator {
     if (size > kMaxAllowedMallocSize || needed_size > kMaxAllowedMallocSize ||
         size > max_user_defined_malloc_size) {
       if (AllocatorMayReturnNull()) {
-        Report("WARNING: MemProfiler failed to allocate 0x%zx bytes\n",
-               (void *)size);
+        Report("WARNING: MemProfiler failed to allocate 0x%zx bytes\n", size);
         return nullptr;
       }
       uptr malloc_limit =
@@ -621,17 +421,15 @@ struct Allocator {
 
     u64 user_requested_size =
         atomic_exchange(&m->user_requested_size, 0, memory_order_acquire);
-    if (memprof_inited && memprof_init_done && !destructing &&
-        MemInfoBlockTable.Constructed) {
+    if (memprof_inited && memprof_init_done &&
+        atomic_load_relaxed(&constructed) &&
+        !atomic_load_relaxed(&destructing)) {
       u64 c = GetShadowCount(p, user_requested_size);
       long curtime = GetTimestamp();
 
       MemInfoBlock newMIB(user_requested_size, c, m->timestamp_ms, curtime,
                           m->cpu_id, GetCpuId());
-      {
-        SpinMutexLock l(&fallback_mutex);
-        MemInfoBlockTable.insertOrMerge(m->alloc_context_id, newMIB);
-      }
+      InsertOrMerge(m->alloc_context_id, newMIB, MIBMap);
     }
 
     MemprofStats &thread_stats = GetCurrentThreadStats();
@@ -898,7 +696,7 @@ uptr __sanitizer_get_allocated_size(const void *p) {
 }
 
 int __memprof_profile_dump() {
-  instance.FinishAndPrint();
+  instance.FinishAndWrite();
   // In the future we may want to return non-zero if there are any errors
   // detected during the dumping process.
   return 0;
diff --git a/compiler-rt/lib/memprof/memprof_flags.inc b/compiler-rt/lib/memprof/memprof_flags.inc
index 035fd15b9288..ee0760ddc302 100644
--- a/compiler-rt/lib/memprof/memprof_flags.inc
+++ b/compiler-rt/lib/memprof/memprof_flags.inc
@@ -35,15 +35,7 @@ MEMPROF_FLAG(bool, allocator_frees_and_returns_null_on_realloc_zero, true,
              "realloc(p, 0) is equivalent to free(p) by default (Same as the "
              "POSIX standard). If set to false, realloc(p, 0) will return a "
              "pointer to an allocated space which can not be used.")
+MEMPROF_FLAG(bool, print_text, false,
+  "If set, prints the heap profile in text format. Else use the raw binary serialization format.")
 MEMPROF_FLAG(bool, print_terse, false,
-             "If set, prints memory profile in a terse format.")
-
-MEMPROF_FLAG(
-    int, mem_info_cache_entries, 16381,
-    "Size in entries of the mem info block cache, should be closest prime"
-    " number to a power of two for best hashing.")
-MEMPROF_FLAG(bool, print_mem_info_cache_miss_rate, false,
-             "If set, prints the miss rate of the mem info block cache.")
-MEMPROF_FLAG(
-    bool, print_mem_info_cache_miss_rate_details, false,
-    "If set, prints detailed miss rates of the mem info block cache sets.")
+             "If set, prints memory profile in a terse format. Only applicable if print_text = true.")
diff --git a/compiler-rt/lib/memprof/memprof_interceptors.cpp b/compiler-rt/lib/memprof/memprof_interceptors.cpp
index e22768061e70..5575ae2fe444 100644
--- a/compiler-rt/lib/memprof/memprof_interceptors.cpp
+++ b/compiler-rt/lib/memprof/memprof_interceptors.cpp
@@ -204,9 +204,9 @@ INTERCEPTOR(char *, strcat, char *to, const char *from) {
   void *ctx;
   MEMPROF_INTERCEPTOR_ENTER(ctx, strcat);
   ENSURE_MEMPROF_INITED();
-  uptr from_length = REAL(strlen)(from);
+  uptr from_length = internal_strlen(from);
   MEMPROF_READ_RANGE(from, from_length + 1);
-  uptr to_length = REAL(strlen)(to);
+  uptr to_length = internal_strlen(to);
   MEMPROF_READ_STRING(to, to_length);
   MEMPROF_WRITE_RANGE(to + to_length, from_length + 1);
   return REAL(strcat)(to, from);
@@ -219,7 +219,7 @@ INTERCEPTOR(char *, strncat, char *to, const char *from, uptr size) {
   uptr from_length = MaybeRealStrnlen(from, size);
   uptr copy_length = Min(size, from_length + 1);
   MEMPROF_READ_RANGE(from, copy_length);
-  uptr to_length = REAL(strlen)(to);
+  uptr to_length = internal_strlen(to);
   MEMPROF_READ_STRING(to, to_length);
   MEMPROF_WRITE_RANGE(to + to_length, from_length + 1);
   return REAL(strncat)(to, from, size);
@@ -232,7 +232,7 @@ INTERCEPTOR(char *, strcpy, char *to, const char *from) {
     return REAL(strcpy)(to, from);
   }
   ENSURE_MEMPROF_INITED();
-  uptr from_size = REAL(strlen)(from) + 1;
+  uptr from_size = internal_strlen(from) + 1;
   MEMPROF_READ_RANGE(from, from_size);
   MEMPROF_WRITE_RANGE(to, from_size);
   return REAL(strcpy)(to, from);
@@ -244,7 +244,7 @@ INTERCEPTOR(char *, strdup, const char *s) {
   if (UNLIKELY(!memprof_inited))
     return internal_strdup(s);
   ENSURE_MEMPROF_INITED();
-  uptr length = REAL(strlen)(s);
+  uptr length = internal_strlen(s);
   MEMPROF_READ_RANGE(s, length + 1);
   GET_STACK_TRACE_MALLOC;
   void *new_mem = memprof_malloc(length + 1, &stack);
@@ -258,7 +258,7 @@ INTERCEPTOR(char *, __strdup, const char *s) {
   if (UNLIKELY(!memprof_inited))
     return internal_strdup(s);
   ENSURE_MEMPROF_INITED();
-  uptr length = REAL(strlen)(s);
+  uptr length = internal_strlen(s);
   MEMPROF_READ_RANGE(s, length + 1);
   GET_STACK_TRACE_MALLOC;
   void *new_mem = memprof_malloc(length + 1, &stack);
diff --git a/compiler-rt/lib/memprof/memprof_interceptors.h b/compiler-rt/lib/memprof/memprof_interceptors.h
index ca5f3690430a..879a1e1061e5 100644
--- a/compiler-rt/lib/memprof/memprof_interceptors.h
+++ b/compiler-rt/lib/memprof/memprof_interceptors.h
@@ -48,13 +48,13 @@ DECLARE_REAL(char *, strstr, const char *s1, const char *s2)
 #define MEMPROF_INTERCEPT_FUNC_VER(name, ver)                                  \
   do {                                                                         \
     if (!INTERCEPT_FUNCTION_VER(name, ver))                                    \
-      VReport(1, "MemProfiler: failed to intercept '%s@@%s'\n", #name, #ver);  \
+      VReport(1, "MemProfiler: failed to intercept '%s@@%s'\n", #name, ver);   \
   } while (0)
 #define MEMPROF_INTERCEPT_FUNC_VER_UNVERSIONED_FALLBACK(name, ver)             \
   do {                                                                         \
     if (!INTERCEPT_FUNCTION_VER(name, ver) && !INTERCEPT_FUNCTION(name))       \
       VReport(1, "MemProfiler: failed to intercept '%s@@%s' or '%s'\n", #name, \
-              #ver, #name);                                                    \
+              ver, #name);                                                     \
   } while (0)
 
 #endif // MEMPROF_INTERCEPTORS_H
diff --git a/compiler-rt/lib/memprof/memprof_malloc_linux.cpp b/compiler-rt/lib/memprof/memprof_malloc_linux.cpp
index c7330f4619a1..ef753fcaa4ad 100644
--- a/compiler-rt/lib/memprof/memprof_malloc_linux.cpp
+++ b/compiler-rt/lib/memprof/memprof_malloc_linux.cpp
@@ -23,125 +23,52 @@
 #include "memprof_internal.h"
 #include "memprof_stack.h"
 #include "sanitizer_common/sanitizer_allocator_checks.h"
+#include "sanitizer_common/sanitizer_allocator_dlsym.h"
 #include "sanitizer_common/sanitizer_errno.h"
 #include "sanitizer_common/sanitizer_tls_get_addr.h"
 
 // ---------------------- Replacement functions ---------------- {{{1
 using namespace __memprof;
 
-static uptr allocated_for_dlsym;
-static uptr last_dlsym_alloc_size_in_words;
-static const uptr kDlsymAllocPoolSize = 1024;
-static uptr alloc_memory_for_dlsym[kDlsymAllocPoolSize];
-
-static inline bool IsInDlsymAllocPool(const void *ptr) {
-  uptr off = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-  return off < allocated_for_dlsym * sizeof(alloc_memory_for_dlsym[0]);
-}
-
-static void *AllocateFromLocalPool(uptr size_in_bytes) {
-  uptr size_in_words = RoundUpTo(size_in_bytes, kWordSize) / kWordSize;
-  void *mem = (void *)&alloc_memory_for_dlsym[allocated_for_dlsym];
-  last_dlsym_alloc_size_in_words = size_in_words;
-  allocated_for_dlsym += size_in_words;
-  CHECK_LT(allocated_for_dlsym, kDlsymAllocPoolSize);
-  return mem;
-}
-
-static void DeallocateFromLocalPool(const void *ptr) {
-  // Hack: since glibc 2.27 dlsym no longer uses stack-allocated memory to store
-  // error messages and instead uses malloc followed by free. To avoid pool
-  // exhaustion due to long object filenames, handle that special case here.
-  uptr prev_offset = allocated_for_dlsym - last_dlsym_alloc_size_in_words;
-  void *prev_mem = (void *)&alloc_memory_for_dlsym[prev_offset];
-  if (prev_mem == ptr) {
-    REAL(memset)(prev_mem, 0, last_dlsym_alloc_size_in_words * kWordSize);
-    allocated_for_dlsym = prev_offset;
-    last_dlsym_alloc_size_in_words = 0;
-  }
-}
-
-static int PosixMemalignFromLocalPool(void **memptr, uptr alignment,
-                                      uptr size_in_bytes) {
-  if (UNLIKELY(!CheckPosixMemalignAlignment(alignment)))
-    return errno_EINVAL;
-
-  CHECK(alignment >= kWordSize);
-
-  uptr addr = (uptr)&alloc_memory_for_dlsym[allocated_for_dlsym];
-  uptr aligned_addr = RoundUpTo(addr, alignment);
-  uptr aligned_size = RoundUpTo(size_in_bytes, kWordSize);
-
-  uptr *end_mem = (uptr *)(aligned_addr + aligned_size);
-  uptr allocated = end_mem - alloc_memory_for_dlsym;
-  if (allocated >= kDlsymAllocPoolSize)
-    return errno_ENOMEM;
-
-  allocated_for_dlsym = allocated;
-  *memptr = (void *)aligned_addr;
-  return 0;
-}
-
-static inline bool MaybeInDlsym() { return memprof_init_is_running; }
-
-static inline bool UseLocalPool() { return MaybeInDlsym(); }
-
-static void *ReallocFromLocalPool(void *ptr, uptr size) {
-  const uptr offset = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-  const uptr copy_size = Min(size, kDlsymAllocPoolSize - offset);
-  void *new_ptr;
-  if (UNLIKELY(UseLocalPool())) {
-    new_ptr = AllocateFromLocalPool(size);
-  } else {
-    ENSURE_MEMPROF_INITED();
-    GET_STACK_TRACE_MALLOC;
-    new_ptr = memprof_malloc(size, &stack);
-  }
-  internal_memcpy(new_ptr, ptr, copy_size);
-  return new_ptr;
-}
+struct DlsymAlloc : public DlSymAllocator<DlsymAlloc> {
+  static bool UseImpl() { return memprof_init_is_running; }
+};
 
 INTERCEPTOR(void, free, void *ptr) {
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
   GET_STACK_TRACE_FREE;
-  if (UNLIKELY(IsInDlsymAllocPool(ptr))) {
-    DeallocateFromLocalPool(ptr);
-    return;
-  }
   memprof_free(ptr, &stack, FROM_MALLOC);
 }
 
 #if SANITIZER_INTERCEPT_CFREE
 INTERCEPTOR(void, cfree, void *ptr) {
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
   GET_STACK_TRACE_FREE;
-  if (UNLIKELY(IsInDlsymAllocPool(ptr)))
-    return;
   memprof_free(ptr, &stack, FROM_MALLOC);
 }
 #endif // SANITIZER_INTERCEPT_CFREE
 
 INTERCEPTOR(void *, malloc, uptr size) {
-  if (UNLIKELY(UseLocalPool()))
-    // Hack: dlsym calls malloc before REAL(malloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(size);
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Allocate(size);
   ENSURE_MEMPROF_INITED();
   GET_STACK_TRACE_MALLOC;
   return memprof_malloc(size, &stack);
 }
 
 INTERCEPTOR(void *, calloc, uptr nmemb, uptr size) {
-  if (UNLIKELY(UseLocalPool()))
-    // Hack: dlsym calls calloc before REAL(calloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(nmemb * size);
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Callocate(nmemb, size);
   ENSURE_MEMPROF_INITED();
   GET_STACK_TRACE_MALLOC;
   return memprof_calloc(nmemb, size, &stack);
 }
 
 INTERCEPTOR(void *, realloc, void *ptr, uptr size) {
-  if (UNLIKELY(IsInDlsymAllocPool(ptr)))
-    return ReallocFromLocalPool(ptr, size);
-  if (UNLIKELY(UseLocalPool()))
-    return AllocateFromLocalPool(size);
+  if (DlsymAlloc::Use() || DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Realloc(ptr, size);
   ENSURE_MEMPROF_INITED();
   GET_STACK_TRACE_MALLOC;
   return memprof_realloc(ptr, size, &stack);
@@ -201,8 +128,6 @@ INTERCEPTOR(int, mallopt, int cmd, int value) { return 0; }
 #endif // SANITIZER_INTERCEPT_MALLOPT_AND_MALLINFO
 
 INTERCEPTOR(int, posix_memalign, void **memptr, uptr alignment, uptr size) {
-  if (UNLIKELY(UseLocalPool()))
-    return PosixMemalignFromLocalPool(memptr, alignment, size);
   GET_STACK_TRACE_MALLOC;
   return memprof_posix_memalign(memptr, alignment, size, &stack);
 }
diff --git a/compiler-rt/lib/memprof/memprof_meminfoblock.h b/compiler-rt/lib/memprof/memprof_meminfoblock.h
new file mode 100644
index 000000000000..19e424435e79
--- /dev/null
+++ b/compiler-rt/lib/memprof/memprof_meminfoblock.h
@@ -0,0 +1,116 @@
+#ifndef MEMPROF_MEMINFOBLOCK_H_
+#define MEMPROF_MEMINFOBLOCK_H_
+
+#include "memprof_interface_internal.h" // For u32, u64 TODO: Move these out of the internal header.
+#include "sanitizer_common/sanitizer_common.h"
+
+namespace __memprof {
+
+using __sanitizer::Printf;
+
+struct MemInfoBlock {
+  u32 alloc_count;
+  u64 total_access_count, min_access_count, max_access_count;
+  u64 total_size;
+  u32 min_size, max_size;
+  u32 alloc_timestamp, dealloc_timestamp;
+  u64 total_lifetime;
+  u32 min_lifetime, max_lifetime;
+  u32 alloc_cpu_id, dealloc_cpu_id;
+  u32 num_migrated_cpu;
+
+  // Only compared to prior deallocated object currently.
+  u32 num_lifetime_overlaps;
+  u32 num_same_alloc_cpu;
+  u32 num_same_dealloc_cpu;
+
+  u64 data_type_id; // TODO: hash of type name
+
+  MemInfoBlock() : alloc_count(0) {}
+
+  MemInfoBlock(u32 size, u64 access_count, u32 alloc_timestamp,
+               u32 dealloc_timestamp, u32 alloc_cpu, u32 dealloc_cpu)
+      : alloc_count(1), total_access_count(access_count),
+        min_access_count(access_count), max_access_count(access_count),
+        total_size(size), min_size(size), max_size(size),
+        alloc_timestamp(alloc_timestamp), dealloc_timestamp(dealloc_timestamp),
+        total_lifetime(dealloc_timestamp - alloc_timestamp),
+        min_lifetime(total_lifetime), max_lifetime(total_lifetime),
+        alloc_cpu_id(alloc_cpu), dealloc_cpu_id(dealloc_cpu),
+        num_lifetime_overlaps(0), num_same_alloc_cpu(0),
+        num_same_dealloc_cpu(0) {
+    num_migrated_cpu = alloc_cpu_id != dealloc_cpu_id;
+  }
+
+  void Print(u64 id, bool print_terse) const {
+    u64 p;
+
+    if (print_terse) {
+      p = total_size * 100 / alloc_count;
+      Printf("MIB:%llu/%u/%llu.%02llu/%u/%u/", id, alloc_count, p / 100,
+             p % 100, min_size, max_size);
+      p = total_access_count * 100 / alloc_count;
+      Printf("%llu.%02llu/%llu/%llu/", p / 100, p % 100, min_access_count,
+             max_access_count);
+      p = total_lifetime * 100 / alloc_count;
+      Printf("%llu.%02llu/%u/%u/", p / 100, p % 100, min_lifetime,
+             max_lifetime);
+      Printf("%u/%u/%u/%u\n", num_migrated_cpu, num_lifetime_overlaps,
+             num_same_alloc_cpu, num_same_dealloc_cpu);
+    } else {
+      p = total_size * 100 / alloc_count;
+      Printf("Memory allocation stack id = %llu\n", id);
+      Printf("\talloc_count %u, size (ave/min/max) %llu.%02llu / %u / %u\n",
+             alloc_count, p / 100, p % 100, min_size, max_size);
+      p = total_access_count * 100 / alloc_count;
+      Printf("\taccess_count (ave/min/max): %llu.%02llu / %llu / %llu\n",
+             p / 100, p % 100, min_access_count, max_access_count);
+      p = total_lifetime * 100 / alloc_count;
+      Printf("\tlifetime (ave/min/max): %llu.%02llu / %u / %u\n", p / 100,
+             p % 100, min_lifetime, max_lifetime);
+      Printf("\tnum migrated: %u, num lifetime overlaps: %u, num same alloc "
+             "cpu: %u, num same dealloc_cpu: %u\n",
+             num_migrated_cpu, num_lifetime_overlaps, num_same_alloc_cpu,
+             num_same_dealloc_cpu);
+    }
+  }
+
+  static void printHeader() {
+    Printf("MIB:StackID/AllocCount/AveSize/MinSize/MaxSize/AveAccessCount/"
+           "MinAccessCount/MaxAccessCount/AveLifetime/MinLifetime/MaxLifetime/"
+           "NumMigratedCpu/NumLifetimeOverlaps/NumSameAllocCpu/"
+           "NumSameDeallocCpu\n");
+  }
+
+  void Merge(const MemInfoBlock &newMIB) {
+    alloc_count += newMIB.alloc_count;
+
+    total_access_count += newMIB.total_access_count;
+    min_access_count = Min(min_access_count, newMIB.min_access_count);
+    max_access_count = Max(max_access_count, newMIB.max_access_count);
+
+    total_size += newMIB.total_size;
+    min_size = Min(min_size, newMIB.min_size);
+    max_size = Max(max_size, newMIB.max_size);
+
+    total_lifetime += newMIB.total_lifetime;
+    min_lifetime = Min(min_lifetime, newMIB.min_lifetime);
+    max_lifetime = Max(max_lifetime, newMIB.max_lifetime);
+
+    // We know newMIB was deallocated later, so just need to check if it was
+    // allocated before last one deallocated.
+    num_lifetime_overlaps += newMIB.alloc_timestamp < dealloc_timestamp;
+    alloc_timestamp = newMIB.alloc_timestamp;
+    dealloc_timestamp = newMIB.dealloc_timestamp;
+
+    num_same_alloc_cpu += alloc_cpu_id == newMIB.alloc_cpu_id;
+    num_same_dealloc_cpu += dealloc_cpu_id == newMIB.dealloc_cpu_id;
+    alloc_cpu_id = newMIB.alloc_cpu_id;
+    dealloc_cpu_id = newMIB.dealloc_cpu_id;
+  }
+
+} __attribute__((packed));
+
+} // namespace __memprof
+
+#endif // MEMPROF_MEMINFOBLOCK_H_
diff --git a/compiler-rt/lib/memprof/memprof_mibmap.cpp b/compiler-rt/lib/memprof/memprof_mibmap.cpp
new file mode 100644
index 000000000000..47449cf9612b
--- /dev/null
+++ b/compiler-rt/lib/memprof/memprof_mibmap.cpp
@@ -0,0 +1,35 @@
+//===-- memprof_mibmap.cpp -----------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of MemProfiler, a memory profiler.
+//
+//===----------------------------------------------------------------------===//
+
+#include "memprof_mibmap.h"
+#include "sanitizer_common/sanitizer_allocator_internal.h"
+#include "sanitizer_common/sanitizer_mutex.h"
+
+namespace __memprof {
+
+void InsertOrMerge(const uptr Id, const MemInfoBlock &Block, MIBMapTy &Map) {
+  MIBMapTy::Handle h(&Map, static_cast<uptr>(Id), /*remove=*/false,
+                     /*create=*/true);
+  if (h.created()) {
+    LockedMemInfoBlock *lmib =
+        (LockedMemInfoBlock *)InternalAlloc(sizeof(LockedMemInfoBlock));
+    lmib->mutex.Init();
+    lmib->mib = Block;
+    *h = lmib;
+  } else {
+    LockedMemInfoBlock *lmib = *h;
+    SpinMutexLock lock(&lmib->mutex);
+    lmib->mib.Merge(Block);
+  }
+}
+
+} // namespace __memprof
diff --git a/compiler-rt/lib/memprof/memprof_mibmap.h b/compiler-rt/lib/memprof/memprof_mibmap.h
new file mode 100644
index 000000000000..ed5dda174fe5
--- /dev/null
+++ b/compiler-rt/lib/memprof/memprof_mibmap.h
@@ -0,0 +1,24 @@
+#ifndef MEMPROF_MIBMAP_H_
+#define MEMPROF_MIBMAP_H_
+
+#include "memprof_meminfoblock.h"
+#include "sanitizer_common/sanitizer_addrhashmap.h"
+#include "sanitizer_common/sanitizer_mutex.h"
+
+namespace __memprof {
+
+struct LockedMemInfoBlock {
+  __sanitizer::StaticSpinMutex mutex;
+  MemInfoBlock mib;
+};
+
+// The MIB map stores a mapping from stack ids to MemInfoBlocks.
+typedef __sanitizer::AddrHashMap<LockedMemInfoBlock *, 200003> MIBMapTy;
+
+// Insert a new MemInfoBlock or merge with an existing block identified by the
+// stack id.
+void InsertOrMerge(const uptr Id, const MemInfoBlock &Block, MIBMapTy &Map);
+
+} // namespace __memprof
+
+#endif // MEMPROF_MIBMAP_H_
diff --git a/compiler-rt/lib/memprof/memprof_rawprofile.cpp b/compiler-rt/lib/memprof/memprof_rawprofile.cpp
new file mode 100644
index 000000000000..96f315f95b24
--- /dev/null
+++ b/compiler-rt/lib/memprof/memprof_rawprofile.cpp
@@ -0,0 +1,250 @@
+#include "memprof_rawprofile.h"
+#include "memprof_meminfoblock.h"
+#include "sanitizer_common/sanitizer_allocator_internal.h"
+#include "sanitizer_common/sanitizer_linux.h"
+#include "sanitizer_common/sanitizer_procmaps.h"
+#include "sanitizer_common/sanitizer_stackdepot.h"
+#include "sanitizer_common/sanitizer_stackdepotbase.h"
+#include "sanitizer_common/sanitizer_stacktrace.h"
+#include "sanitizer_common/sanitizer_vector.h"
+
+#include <stdlib.h>
+#include <string.h>
+
+namespace __memprof {
+using ::__sanitizer::Vector;
+
+namespace {
+typedef struct __attribute__((__packed__)) {
+  u64 start;
+  u64 end;
+  u64 offset;
+  u8 buildId[32];
+} SegmentEntry;
+
+typedef struct __attribute__((__packed__)) {
+  u64 magic;
+  u64 version;
+  u64 total_size;
+  u64 segment_offset;
+  u64 mib_offset;
+  u64 stack_offset;
+} Header;
+
+template <class T> char *WriteBytes(T Pod, char *&Buffer) {
+  *(T *)Buffer = Pod;
+  return Buffer + sizeof(T);
+}
+
+void RecordStackId(const uptr Key, UNUSED LockedMemInfoBlock *const &MIB,
+                   void *Arg) {
+  // No need to touch the MIB value here since we are only recording the key.
+  auto *StackIds = reinterpret_cast<Vector<u64> *>(Arg);
+  StackIds->PushBack(Key);
+}
+} // namespace
+
+u64 SegmentSizeBytes(MemoryMappingLayoutBase &Layout) {
+  u64 NumSegmentsToRecord = 0;
+  MemoryMappedSegment segment;
+  for (Layout.Reset(); Layout.Next(&segment);)
+    if (segment.IsReadable() && segment.IsExecutable())
+      NumSegmentsToRecord++;
+
+  return sizeof(u64) // A header which stores the number of records.
+         + sizeof(SegmentEntry) * NumSegmentsToRecord;
+}
+
+// The segment section uses the following format:
+// ---------- Segment Info
+// Num Entries
+// ---------- Segment Entry
+// Start
+// End
+// Offset
+// BuildID 32B
+// ----------
+// ...
+void SerializeSegmentsToBuffer(MemoryMappingLayoutBase &Layout,
+                               const u64 ExpectedNumBytes, char *&Buffer) {
+  char *Ptr = Buffer;
+  // Reserve space for the final count.
+  Ptr += sizeof(u64);
+
+  u64 NumSegmentsRecorded = 0;
+  MemoryMappedSegment segment;
+
+  for (Layout.Reset(); Layout.Next(&segment);) {
+    if (segment.IsReadable() && segment.IsExecutable()) {
+      SegmentEntry entry{};
+      entry.start = segment.start;
+      entry.end = segment.end;
+      entry.offset = segment.offset;
+      memcpy(entry.buildId, segment.uuid, sizeof(segment.uuid));
+      memcpy(Ptr, &entry, sizeof(SegmentEntry));
+      Ptr += sizeof(SegmentEntry);
+      NumSegmentsRecorded++;
+    }
+  }
+
+  // Store the number of segments we recorded in the space we reserved.
+  *((u64 *)Buffer) = NumSegmentsRecorded;
+  CHECK(ExpectedNumBytes == static_cast<u64>(Ptr - Buffer) &&
+        "Expected num bytes != actual bytes written");
+}
+
+u64 StackSizeBytes(const Vector<u64> &StackIds) {
+  u64 NumBytesToWrite = sizeof(u64);
+
+  const u64 NumIds = StackIds.Size();
+  for (unsigned k = 0; k < NumIds; ++k) {
+    const u64 Id = StackIds[k];
+    // One entry for the id and then one more for the number of stack pcs.
+    NumBytesToWrite += 2 * sizeof(u64);
+    const StackTrace St = StackDepotGet(Id);
+
+    CHECK(St.trace != nullptr && St.size > 0 && "Empty stack trace");
+    for (uptr i = 0; i < St.size && St.trace[i] != 0; i++) {
+      NumBytesToWrite += sizeof(u64);
+    }
+  }
+  return NumBytesToWrite;
+}
+
+// The stack info section uses the following format:
+//
+// ---------- Stack Info
+// Num Entries
+// ---------- Stack Entry
+// Num Stacks
+// PC1
+// PC2
+// ...
+// ----------
+void SerializeStackToBuffer(const Vector<u64> &StackIds,
+                            const u64 ExpectedNumBytes, char *&Buffer) {
+  const u64 NumIds = StackIds.Size();
+  char *Ptr = Buffer;
+  Ptr = WriteBytes(static_cast<u64>(NumIds), Ptr);
+
+  for (unsigned k = 0; k < NumIds; ++k) {
+    const u64 Id = StackIds[k];
+    Ptr = WriteBytes(Id, Ptr);
+    Ptr += sizeof(u64); // Bump it by u64, we will fill this in later.
+    u64 Count = 0;
+    const StackTrace St = StackDepotGet(Id);
+    for (uptr i = 0; i < St.size && St.trace[i] != 0; i++) {
+      // PCs in stack traces are actually the return addresses, that is,
+      // addresses of the next instructions after the call.
+      uptr pc = StackTrace::GetPreviousInstructionPc(St.trace[i]);
+      Ptr = WriteBytes(static_cast<u64>(pc), Ptr);
+      ++Count;
+    }
+    // Store the count in the space we reserved earlier.
+    *(u64 *)(Ptr - (Count + 1) * sizeof(u64)) = Count;
+  }
+
+  CHECK(ExpectedNumBytes == static_cast<u64>(Ptr - Buffer) &&
+        "Expected num bytes != actual bytes written");
+}
+
+// The MIB section has the following format:
+// ---------- MIB Info
+// Num Entries
+// ---------- MIB Entry 0
+// Alloc Count
+// ...
+// ---------- MIB Entry 1
+// Alloc Count
+// ...
+// ----------
+void SerializeMIBInfoToBuffer(MIBMapTy &MIBMap, const Vector<u64> &StackIds,
+                              const u64 ExpectedNumBytes, char *&Buffer) {
+  char *Ptr = Buffer;
+  const u64 NumEntries = StackIds.Size();
+  Ptr = WriteBytes(NumEntries, Ptr);
+
+  for (u64 i = 0; i < NumEntries; i++) {
+    const u64 Key = StackIds[i];
+    MIBMapTy::Handle h(&MIBMap, Key, /*remove=*/true, /*create=*/false);
+    CHECK(h.exists());
+    Ptr = WriteBytes(Key, Ptr);
+    Ptr = WriteBytes((*h)->mib, Ptr);
+  }
+
+  CHECK(ExpectedNumBytes == static_cast<u64>(Ptr - Buffer) &&
+        "Expected num bytes != actual bytes written");
+}
+
+// Format
+// ---------- Header
+// Magic
+// Version
+// Total Size
+// Segment Offset
+// MIB Info Offset
+// Stack Offset
+// ---------- Segment Info
+// Num Entries
+// ---------- Segment Entry
+// Start
+// End
+// Offset
+// BuildID 32B
+// ----------
+// ...
+// ---------- MIB Info
+// Num Entries
+// ---------- MIB Entry
+// Alloc Count
+// ...
+// ---------- Stack Info
+// Num Entries
+// ---------- Stack Entry
+// Num Stacks
+// PC1
+// PC2
+// ...
+// ----------
+// ...
+u64 SerializeToRawProfile(MIBMapTy &MIBMap, MemoryMappingLayoutBase &Layout,
+                          char *&Buffer) {
+  const u64 NumSegmentBytes = SegmentSizeBytes(Layout);
+
+  Vector<u64> StackIds;
+  MIBMap.ForEach(RecordStackId, reinterpret_cast<void *>(&StackIds));
+  // The first 8b are for the total number of MIB records. Each MIB record is
+  // preceded by a 8b stack id which is associated with stack frames in the next
+  // section.
+  const u64 NumMIBInfoBytes =
+      sizeof(u64) + StackIds.Size() * (sizeof(u64) + sizeof(MemInfoBlock));
+
+  const u64 NumStackBytes = StackSizeBytes(StackIds);
+
+  const u64 TotalSizeBytes =
+      sizeof(Header) + NumSegmentBytes + NumStackBytes + NumMIBInfoBytes;
+
+  // Allocate the memory for the entire buffer incl. info blocks.
+  Buffer = (char *)InternalAlloc(TotalSizeBytes);
+  char *Ptr = Buffer;
+
+  Header header{MEMPROF_RAW_MAGIC_64,
+                MEMPROF_RAW_VERSION,
+                static_cast<u64>(TotalSizeBytes),
+                sizeof(Header),
+                sizeof(Header) + NumSegmentBytes,
+                sizeof(Header) + NumSegmentBytes + NumMIBInfoBytes};
+  Ptr = WriteBytes(header, Ptr);
+
+  SerializeSegmentsToBuffer(Layout, NumSegmentBytes, Ptr);
+  Ptr += NumSegmentBytes;
+
+  SerializeMIBInfoToBuffer(MIBMap, StackIds, NumMIBInfoBytes, Ptr);
+  Ptr += NumMIBInfoBytes;
+
+  SerializeStackToBuffer(StackIds, NumStackBytes, Ptr);
+
+  return TotalSizeBytes;
+}
+
+} // namespace __memprof
diff --git a/compiler-rt/lib/memprof/memprof_rawprofile.h b/compiler-rt/lib/memprof/memprof_rawprofile.h
new file mode 100644
index 000000000000..052bac3267f1
--- /dev/null
+++ b/compiler-rt/lib/memprof/memprof_rawprofile.h
@@ -0,0 +1,21 @@
+#ifndef MEMPROF_RAWPROFILE_H_
+#define MEMPROF_RAWPROFILE_H_
+
+#include "memprof_mibmap.h"
+#include "sanitizer_common/sanitizer_procmaps.h"
+
+namespace __memprof {
+
+// TODO: pull these in from MemProfData.inc
+#define MEMPROF_RAW_MAGIC_64                                                   \
+  (u64)255 << 56 | (u64)'m' << 48 | (u64)'p' << 40 | (u64)'r' << 32 |          \
+      (u64)'o' << 24 | (u64)'f' << 16 | (u64)'r' << 8 | (u64)129
+
+#define MEMPROF_RAW_VERSION 1ULL
+
+u64 SerializeToRawProfile(MIBMapTy &BlockCache, MemoryMappingLayoutBase &Layout,
+                          char *&Buffer);
+
+} // namespace __memprof
+
+#endif // MEMPROF_RAWPROFILE_H_
diff --git a/compiler-rt/lib/memprof/memprof_rtl.cpp b/compiler-rt/lib/memprof/memprof_rtl.cpp
index fee2912d64d4..fb2ef37e51a2 100644
--- a/compiler-rt/lib/memprof/memprof_rtl.cpp
+++ b/compiler-rt/lib/memprof/memprof_rtl.cpp
@@ -264,14 +264,9 @@ void __memprof_record_access(void const volatile *addr) {
   __memprof::RecordAccess((uptr)addr);
 }
 
-// We only record the access on the first location in the range,
-// since we will later accumulate the access counts across the
-// full allocation, and we don't want to inflate the hotness from
-// a memory intrinsic on a large range of memory.
-// TODO: Should we do something else so we can better track utilization?
-void __memprof_record_access_range(void const volatile *addr,
-                                   UNUSED uptr size) {
-  __memprof::RecordAccess((uptr)addr);
+void __memprof_record_access_range(void const volatile *addr, uptr size) {
+  for (uptr a = (uptr)addr; a < (uptr)addr + size; a += kWordSize)
+    __memprof::RecordAccess(a);
 }
 
 extern "C" SANITIZER_INTERFACE_ATTRIBUTE u16
diff --git a/compiler-rt/lib/memprof/memprof_stats.cpp b/compiler-rt/lib/memprof/memprof_stats.cpp
index 8a50d270dc6a..c8faebfa12de 100644
--- a/compiler-rt/lib/memprof/memprof_stats.cpp
+++ b/compiler-rt/lib/memprof/memprof_stats.cpp
@@ -62,11 +62,11 @@ void MemprofStats::MergeFrom(const MemprofStats *stats) {
     dst_ptr[i] += src_ptr[i];
 }
 
-static BlockingMutex print_lock(LINKER_INITIALIZED);
+static Mutex print_lock;
 
 static MemprofStats unknown_thread_stats(LINKER_INITIALIZED);
 static MemprofStats dead_threads_stats(LINKER_INITIALIZED);
-static BlockingMutex dead_threads_stats_lock(LINKER_INITIALIZED);
+static Mutex dead_threads_stats_lock;
 // Required for malloc_zone_statistics() on OS X. This can't be stored in
 // per-thread MemprofStats.
 static uptr max_malloced_memory;
@@ -87,7 +87,7 @@ static void GetAccumulatedStats(MemprofStats *stats) {
   }
   stats->MergeFrom(&unknown_thread_stats);
   {
-    BlockingMutexLock lock(&dead_threads_stats_lock);
+    Lock lock(&dead_threads_stats_lock);
     stats->MergeFrom(&dead_threads_stats);
   }
   // This is not very accurate: we may miss allocation peaks that happen
@@ -99,7 +99,7 @@ static void GetAccumulatedStats(MemprofStats *stats) {
 }
 
 void FlushToDeadThreadStats(MemprofStats *stats) {
-  BlockingMutexLock lock(&dead_threads_stats_lock);
+  Lock lock(&dead_threads_stats_lock);
   dead_threads_stats.MergeFrom(stats);
   stats->Clear();
 }
@@ -113,11 +113,11 @@ static void PrintAccumulatedStats() {
   MemprofStats stats;
   GetAccumulatedStats(&stats);
   // Use lock to keep reports from mixing up.
-  BlockingMutexLock lock(&print_lock);
+  Lock lock(&print_lock);
   stats.Print();
-  StackDepotStats *stack_depot_stats = StackDepotGetStats();
+  StackDepotStats stack_depot_stats = StackDepotGetStats();
   Printf("Stats: StackDepot: %zd ids; %zdM allocated\n",
-         stack_depot_stats->n_uniq_ids, stack_depot_stats->allocated >> 20);
+         stack_depot_stats.n_uniq_ids, stack_depot_stats.allocated >> 20);
   PrintInternalAllocatorStats();
 }
 
diff --git a/compiler-rt/lib/memprof/memprof_thread.cpp b/compiler-rt/lib/memprof/memprof_thread.cpp
index 5ae7a2ee85b9..9512a87cf98e 100644
--- a/compiler-rt/lib/memprof/memprof_thread.cpp
+++ b/compiler-rt/lib/memprof/memprof_thread.cpp
@@ -40,11 +40,11 @@ void MemprofThreadContext::OnFinished() {
 static ALIGNED(16) char thread_registry_placeholder[sizeof(ThreadRegistry)];
 static ThreadRegistry *memprof_thread_registry;
 
-static BlockingMutex mu_for_thread_context(LINKER_INITIALIZED);
+static Mutex mu_for_thread_context;
 static LowLevelAllocator allocator_for_thread_context;
 
 static ThreadContextBase *GetMemprofThreadContext(u32 tid) {
-  BlockingMutexLock lock(&mu_for_thread_context);
+  Lock lock(&mu_for_thread_context);
   return new (allocator_for_thread_context) MemprofThreadContext(tid);
 }
 
@@ -80,8 +80,7 @@ MemprofThread *MemprofThread::Create(thread_callback_t start_routine, void *arg,
   thread->start_routine_ = start_routine;
   thread->arg_ = arg;
   MemprofThreadContext::CreateThreadContextArgs args = {thread, stack};
-  memprofThreadRegistry().CreateThread(*reinterpret_cast<uptr *>(thread),
-                                       detached, parent_tid, &args);
+  memprofThreadRegistry().CreateThread(0, detached, parent_tid, &args);
 
   return thread;
 }
@@ -131,7 +130,7 @@ void MemprofThread::Init(const InitOptions *options) {
   int local = 0;
   VReport(1, "T%d: stack [%p,%p) size 0x%zx; local=%p\n", tid(),
           (void *)stack_bottom_, (void *)stack_top_, stack_top_ - stack_bottom_,
-          &local);
+          (void *)&local);
 }
 
 thread_return_t
@@ -198,7 +197,7 @@ MemprofThread *GetCurrentThread() {
 
 void SetCurrentThread(MemprofThread *t) {
   CHECK(t->context());
-  VReport(2, "SetCurrentThread: %p for thread %p\n", t->context(),
+  VReport(2, "SetCurrentThread: %p for thread %p\n", (void *)t->context(),
           (void *)GetThreadSelf());
   // Make sure we do not reset the current MemprofThread.
   CHECK_EQ(0, TSDGet());
diff --git a/compiler-rt/lib/memprof/tests/driver.cpp b/compiler-rt/lib/memprof/tests/driver.cpp
new file mode 100644
index 000000000000..b402cec1126b
--- /dev/null
+++ b/compiler-rt/lib/memprof/tests/driver.cpp
@@ -0,0 +1,14 @@
+//===-- driver.cpp ----------------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "gtest/gtest.h"
+
+int main(int argc, char **argv) {
+  testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}
diff --git a/compiler-rt/lib/memprof/tests/rawprofile.cpp b/compiler-rt/lib/memprof/tests/rawprofile.cpp
new file mode 100644
index 000000000000..4404ab86092e
--- /dev/null
+++ b/compiler-rt/lib/memprof/tests/rawprofile.cpp
@@ -0,0 +1,188 @@
+#include "memprof/memprof_rawprofile.h"
+
+#include "memprof/memprof_meminfoblock.h"
+#include "sanitizer_common/sanitizer_common.h"
+#include "sanitizer_common/sanitizer_procmaps.h"
+#include "sanitizer_common/sanitizer_stackdepot.h"
+#include "sanitizer_common/sanitizer_stacktrace.h"
+#include "gmock/gmock.h"
+#include "gtest/gtest.h"
+
+#include <memory>
+
+namespace {
+
+using ::__memprof::MemInfoBlock;
+using ::__memprof::MIBMapTy;
+using ::__memprof::SerializeToRawProfile;
+using ::__sanitizer::MemoryMappedSegment;
+using ::__sanitizer::MemoryMappingLayoutBase;
+using ::__sanitizer::StackDepotPut;
+using ::__sanitizer::StackTrace;
+using ::testing::_;
+using ::testing::Action;
+using ::testing::DoAll;
+using ::testing::Return;
+using ::testing::SetArgPointee;
+
+class MockMemoryMappingLayout final : public MemoryMappingLayoutBase {
+public:
+  MOCK_METHOD(bool, Next, (MemoryMappedSegment *), (override));
+  MOCK_METHOD(void, Reset, (), (override));
+};
+
+u64 PopulateFakeMap(const MemInfoBlock &FakeMIB, uptr StackPCBegin,
+                    MIBMapTy &FakeMap) {
+  constexpr int kSize = 5;
+  uptr array[kSize];
+  for (int i = 0; i < kSize; i++) {
+    array[i] = StackPCBegin + i;
+  }
+  StackTrace St(array, kSize);
+  u32 Id = StackDepotPut(St);
+
+  InsertOrMerge(Id, FakeMIB, FakeMap);
+  return Id;
+}
+
+template <class T = u64> T Read(char *&Buffer) {
+  static_assert(std::is_pod<T>::value, "Must be a POD type.");
+  T t = *reinterpret_cast<T *>(Buffer);
+  Buffer += sizeof(T);
+  return t;
+}
+
+TEST(MemProf, Basic) {
+  MockMemoryMappingLayout Layout;
+  MemoryMappedSegment FakeSegment;
+  memset(&FakeSegment, 0, sizeof(FakeSegment));
+  FakeSegment.start = 0x10;
+  FakeSegment.end = 0x20;
+  FakeSegment.offset = 0x10;
+  uint8_t uuid[__sanitizer::kModuleUUIDSize] = {0xC, 0x0, 0xF, 0xF, 0xE, 0xE};
+  memcpy(FakeSegment.uuid, uuid, __sanitizer::kModuleUUIDSize);
+  FakeSegment.protection =
+      __sanitizer::kProtectionExecute | __sanitizer::kProtectionRead;
+
+  const Action<bool(MemoryMappedSegment *)> SetSegment =
+      DoAll(SetArgPointee<0>(FakeSegment), Return(true));
+  EXPECT_CALL(Layout, Next(_))
+      .WillOnce(SetSegment)
+      .WillOnce(Return(false))
+      .WillOnce(SetSegment)
+      .WillRepeatedly(Return(false));
+
+  EXPECT_CALL(Layout, Reset).Times(2);
+
+  MIBMapTy FakeMap;
+  MemInfoBlock FakeMIB;
+  // Since we want to override the constructor set vals to make it easier to
+  // test.
+  memset(&FakeMIB, 0, sizeof(MemInfoBlock));
+  FakeMIB.alloc_count = 0x1;
+  FakeMIB.total_access_count = 0x2;
+
+  u64 FakeIds[2];
+  FakeIds[0] = PopulateFakeMap(FakeMIB, /*StackPCBegin=*/2, FakeMap);
+  FakeIds[1] = PopulateFakeMap(FakeMIB, /*StackPCBegin=*/3, FakeMap);
+
+  char *Ptr = nullptr;
+  u64 NumBytes = SerializeToRawProfile(FakeMap, Layout, Ptr);
+  const char *Buffer = Ptr;
+
+  ASSERT_GT(NumBytes, 0ULL);
+  ASSERT_TRUE(Ptr);
+
+  // Check the header.
+  EXPECT_THAT(Read(Ptr), MEMPROF_RAW_MAGIC_64);
+  EXPECT_THAT(Read(Ptr), MEMPROF_RAW_VERSION);
+  const u64 TotalSize = Read(Ptr);
+  const u64 SegmentOffset = Read(Ptr);
+  const u64 MIBOffset = Read(Ptr);
+  const u64 StackOffset = Read(Ptr);
+
+  // ============= Check sizes.
+  EXPECT_EQ(TotalSize, NumBytes);
+
+  // Should be equal to the size of the raw profile header.
+  EXPECT_EQ(SegmentOffset, 48ULL);
+
+  // We expect only 1 segment entry, 8b for the count and 56b for SegmentEntry
+  // in memprof_rawprofile.cpp.
+  EXPECT_EQ(MIBOffset - SegmentOffset, 64ULL);
+
+  EXPECT_EQ(MIBOffset, 112ULL);
+  // We expect 2 mib entry, 8b for the count and sizeof(u64) +
+  // sizeof(MemInfoBlock) contains stack id + MeminfoBlock.
+  EXPECT_EQ(StackOffset - MIBOffset, 8 + 2 * (8 + sizeof(MemInfoBlock)));
+
+  EXPECT_EQ(StackOffset, 336ULL);
+  // We expect 2 stack entries, with 5 frames - 8b for total count,
+  // 2 * (8b for id, 8b for frame count and 5*8b for fake frames)
+  EXPECT_EQ(TotalSize - StackOffset, 8ULL + 2 * (8 + 8 + 5 * 8));
+
+  // ============= Check contents.
+  unsigned char ExpectedSegmentBytes[64] = {
+      0x01, 0,   0,   0,   0,   0,   0, 0, // Number of entries
+      0x10, 0,   0,   0,   0,   0,   0, 0, // Start
+      0x20, 0,   0,   0,   0,   0,   0, 0, // End
+      0x10, 0,   0,   0,   0,   0,   0, 0, // Offset
+      0x0C, 0x0, 0xF, 0xF, 0xE, 0xE,       // Uuid
+  };
+  EXPECT_EQ(memcmp(Buffer + SegmentOffset, ExpectedSegmentBytes, 64), 0);
+
+  // Check that the number of entries is 2.
+  EXPECT_EQ(*reinterpret_cast<const u64 *>(Buffer + MIBOffset), 2ULL);
+  // Check that stack id is set.
+  EXPECT_EQ(*reinterpret_cast<const u64 *>(Buffer + MIBOffset + 8), FakeIds[0]);
+
+  // Only check a few fields of the first MemInfoBlock.
+  unsigned char ExpectedMIBBytes[sizeof(MemInfoBlock)] = {
+      0x01, 0, 0, 0, // Alloc count
+      0x02, 0, 0, 0, // Total access count
+  };
+  // Compare contents of 1st MIB after skipping count and stack id.
+  EXPECT_EQ(
+      memcmp(Buffer + MIBOffset + 16, ExpectedMIBBytes, sizeof(MemInfoBlock)),
+      0);
+  // Compare contents of 2nd MIB after skipping count and stack id for the first
+  // and only the id for the second.
+  EXPECT_EQ(memcmp(Buffer + MIBOffset + 16 + sizeof(MemInfoBlock) + 8,
+                   ExpectedMIBBytes, sizeof(MemInfoBlock)),
+            0);
+
+  // Check that the number of entries is 2.
+  EXPECT_EQ(*reinterpret_cast<const u64 *>(Buffer + StackOffset), 2ULL);
+  // Check that the 1st stack id is set.
+  EXPECT_EQ(*reinterpret_cast<const u64 *>(Buffer + StackOffset + 8),
+            FakeIds[0]);
+  // Contents are num pcs, value of each pc - 1.
+  unsigned char ExpectedStackBytes[2][6 * 8] = {
+      {
+          0x5, 0, 0, 0, 0, 0, 0, 0, // Number of PCs
+          0x1, 0, 0, 0, 0, 0, 0, 0, // PC ...
+          0x2, 0, 0, 0, 0, 0, 0, 0, 0x3, 0, 0, 0, 0, 0, 0, 0,
+          0x4, 0, 0, 0, 0, 0, 0, 0, 0x5, 0, 0, 0, 0, 0, 0, 0,
+      },
+      {
+          0x5, 0, 0, 0, 0, 0, 0, 0, // Number of PCs
+          0x2, 0, 0, 0, 0, 0, 0, 0, // PC ...
+          0x3, 0, 0, 0, 0, 0, 0, 0, 0x4, 0, 0, 0, 0, 0, 0, 0,
+          0x5, 0, 0, 0, 0, 0, 0, 0, 0x6, 0, 0, 0, 0, 0, 0, 0,
+      },
+  };
+  EXPECT_EQ(memcmp(Buffer + StackOffset + 16, ExpectedStackBytes[0],
+                   sizeof(ExpectedStackBytes[0])),
+            0);
+
+  // Check that the 2nd stack id is set.
+  EXPECT_EQ(
+      *reinterpret_cast<const u64 *>(Buffer + StackOffset + 8 + 6 * 8 + 8),
+      FakeIds[1]);
+
+  EXPECT_EQ(memcmp(Buffer + StackOffset + 16 + 6 * 8 + 8, ExpectedStackBytes[1],
+                   sizeof(ExpectedStackBytes[1])),
+            0);
+}
+
+} // namespace
diff --git a/compiler-rt/lib/msan/msan.cpp b/compiler-rt/lib/msan/msan.cpp
index 4fa772fdcb6e..c554a830e755 100644
--- a/compiler-rt/lib/msan/msan.cpp
+++ b/compiler-rt/lib/msan/msan.cpp
@@ -470,7 +470,7 @@ void __msan_init() {
 
   MsanThread *main_thread = MsanThread::Create(nullptr, nullptr);
   SetCurrentThread(main_thread);
-  main_thread->ThreadStart();
+  main_thread->Init();
 
 #if MSAN_CONTAINS_UBSAN
   __ubsan::InitAsPlugin();
@@ -515,6 +515,7 @@ void __msan_dump_shadow(const void *x, uptr size) {
   }
 
   unsigned char *s = (unsigned char*)MEM_TO_SHADOW(x);
+  Printf("%p[%p]  ", (void *)s, x);
   for (uptr i = 0; i < size; i++)
     Printf("%x%x ", s[i] >> 4, s[i] & 0xf);
   Printf("\n");
@@ -604,7 +605,7 @@ void __msan_set_alloca_origin4(void *a, uptr size, char *descr, uptr pc) {
     id = Origin::CreateStackOrigin(idx).raw_id();
     *id_ptr = id;
     if (print)
-      Printf("First time: idx=%d id=%d %s %p \n", idx, id, descr + 4, pc);
+      Printf("First time: idx=%d id=%d %s 0x%zx \n", idx, id, descr + 4, pc);
   }
   if (print)
     Printf("__msan_set_alloca_origin: descr=%s id=%x\n", descr + 4, id);
diff --git a/compiler-rt/lib/msan/msan.h b/compiler-rt/lib/msan/msan.h
index 963b94a54087..4b2cec31756a 100644
--- a/compiler-rt/lib/msan/msan.h
+++ b/compiler-rt/lib/msan/msan.h
@@ -121,7 +121,7 @@ const MappingDesc kMemoryLayout[] = {
     // The mappings below are used only for 48-bits VMA.
     // TODO(unknown): 48-bit mapping ony covers the usual PIE, non-PIE
     // segments and some more segments totalizing 262144GB of VMA (which cover
-    // only 0.32% of all 48-bit VMA). Memory avaliability can be increase by
+    // only 0.32% of all 48-bit VMA). Memory availability can be increase by
     // adding multiple application segments like 39 and 42 mapping.
     {0x0040000000000ULL, 0x0041000000000ULL, MappingDesc::INVALID, "invalid"},
     {0x0041000000000ULL, 0x0042000000000ULL, MappingDesc::APP, "app-10"},
@@ -219,7 +219,7 @@ const MappingDesc kMemoryLayout[] = {
 #elif SANITIZER_NETBSD || (SANITIZER_LINUX && SANITIZER_WORDSIZE == 64)
 
 #ifdef MSAN_LINUX_X86_64_OLD_MAPPING
-// Requries PIE binary and ASLR enabled.
+// Requires PIE binary and ASLR enabled.
 // Main thread stack and DSOs at 0x7f0000000000 (sometimes 0x7e0000000000).
 // Heap at 0x600000000000.
 const MappingDesc kMemoryLayout[] = {
diff --git a/compiler-rt/lib/msan/msan_chained_origin_depot.cpp b/compiler-rt/lib/msan/msan_chained_origin_depot.cpp
index 5dee80fd4692..49b14131a89b 100644
--- a/compiler-rt/lib/msan/msan_chained_origin_depot.cpp
+++ b/compiler-rt/lib/msan/msan_chained_origin_depot.cpp
@@ -19,7 +19,7 @@ namespace __msan {
 
 static ChainedOriginDepot chainedOriginDepot;
 
-StackDepotStats *ChainedOriginDepotGetStats() {
+StackDepotStats ChainedOriginDepotGetStats() {
   return chainedOriginDepot.GetStats();
 }
 
diff --git a/compiler-rt/lib/msan/msan_chained_origin_depot.h b/compiler-rt/lib/msan/msan_chained_origin_depot.h
index 60ab182fa4c8..ea51c77a905b 100644
--- a/compiler-rt/lib/msan/msan_chained_origin_depot.h
+++ b/compiler-rt/lib/msan/msan_chained_origin_depot.h
@@ -19,7 +19,7 @@
 namespace __msan {
 
 // Gets the statistic of the origin chain storage.
-StackDepotStats *ChainedOriginDepotGetStats();
+StackDepotStats ChainedOriginDepotGetStats();
 
 // Stores a chain with StackDepot ID here_id and previous chain ID prev_id.
 // If successful, returns true and the new chain id new_id.
diff --git a/compiler-rt/lib/msan/msan_interceptors.cpp b/compiler-rt/lib/msan/msan_interceptors.cpp
index 760f74e927d0..eaa3b3ae9404 100644
--- a/compiler-rt/lib/msan/msan_interceptors.cpp
+++ b/compiler-rt/lib/msan/msan_interceptors.cpp
@@ -18,21 +18,22 @@
 #include "msan.h"
 #include "msan_chained_origin_depot.h"
 #include "msan_origin.h"
+#include "msan_poisoning.h"
 #include "msan_report.h"
 #include "msan_thread.h"
-#include "msan_poisoning.h"
-#include "sanitizer_common/sanitizer_errno_codes.h"
-#include "sanitizer_common/sanitizer_platform_limits_posix.h"
-#include "sanitizer_common/sanitizer_platform_limits_netbsd.h"
 #include "sanitizer_common/sanitizer_allocator.h"
+#include "sanitizer_common/sanitizer_allocator_dlsym.h"
 #include "sanitizer_common/sanitizer_allocator_interface.h"
-#include "sanitizer_common/sanitizer_allocator_internal.h"
 #include "sanitizer_common/sanitizer_atomic.h"
 #include "sanitizer_common/sanitizer_common.h"
 #include "sanitizer_common/sanitizer_errno.h"
-#include "sanitizer_common/sanitizer_stackdepot.h"
+#include "sanitizer_common/sanitizer_errno_codes.h"
+#include "sanitizer_common/sanitizer_glibc_version.h"
 #include "sanitizer_common/sanitizer_libc.h"
 #include "sanitizer_common/sanitizer_linux.h"
+#include "sanitizer_common/sanitizer_platform_limits_netbsd.h"
+#include "sanitizer_common/sanitizer_platform_limits_posix.h"
+#include "sanitizer_common/sanitizer_stackdepot.h"
 #include "sanitizer_common/sanitizer_tls_get_addr.h"
 #include "sanitizer_common/sanitizer_vector.h"
 
@@ -74,22 +75,9 @@ bool IsInInterceptorScope() {
   return in_interceptor_scope;
 }
 
-static uptr allocated_for_dlsym;
-static const uptr kDlsymAllocPoolSize = 1024;
-static uptr alloc_memory_for_dlsym[kDlsymAllocPoolSize];
-
-static bool IsInDlsymAllocPool(const void *ptr) {
-  uptr off = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-  return off < sizeof(alloc_memory_for_dlsym);
-}
-
-static void *AllocateFromLocalPool(uptr size_in_bytes) {
-  uptr size_in_words = RoundUpTo(size_in_bytes, kWordSize) / kWordSize;
-  void *mem = (void *)&alloc_memory_for_dlsym[allocated_for_dlsym];
-  allocated_for_dlsym += size_in_words;
-  CHECK_LT(allocated_for_dlsym, kDlsymAllocPoolSize);
-  return mem;
-}
+struct DlsymAlloc : public DlSymAllocator<DlsymAlloc> {
+  static bool UseImpl() { return !msan_inited; }
+};
 
 #define ENSURE_MSAN_INITED() do { \
   CHECK(!msan_init_is_running); \
@@ -220,18 +208,24 @@ INTERCEPTOR(void *, pvalloc, SIZE_T size) {
 #endif
 
 INTERCEPTOR(void, free, void *ptr) {
+  if (UNLIKELY(!ptr))
+    return;
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
   GET_MALLOC_STACK_TRACE;
-  if (!ptr || UNLIKELY(IsInDlsymAllocPool(ptr))) return;
   MsanDeallocate(&stack, ptr);
 }
 
 #if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(void, cfree, void *ptr) {
+  if (UNLIKELY(!ptr))
+    return;
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
   GET_MALLOC_STACK_TRACE;
-  if (!ptr || UNLIKELY(IsInDlsymAllocPool(ptr))) return;
   MsanDeallocate(&stack, ptr);
 }
-#define MSAN_MAYBE_INTERCEPT_CFREE INTERCEPT_FUNCTION(cfree)
+#  define MSAN_MAYBE_INTERCEPT_CFREE INTERCEPT_FUNCTION(cfree)
 #else
 #define MSAN_MAYBE_INTERCEPT_CFREE
 #endif
@@ -286,7 +280,7 @@ INTERCEPTOR(void, malloc_stats, void) {
 INTERCEPTOR(char *, strcpy, char *dest, const char *src) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
-  SIZE_T n = REAL(strlen)(src);
+  SIZE_T n = internal_strlen(src);
   CHECK_UNPOISONED_STRING(src + n, 0);
   char *res = REAL(strcpy)(dest, src);
   CopyShadowAndOrigin(dest, src, n + 1, &stack);
@@ -296,7 +290,7 @@ INTERCEPTOR(char *, strcpy, char *dest, const char *src) {
 INTERCEPTOR(char *, strncpy, char *dest, const char *src, SIZE_T n) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
-  SIZE_T copy_size = REAL(strnlen)(src, n);
+  SIZE_T copy_size = internal_strnlen(src, n);
   if (copy_size < n)
     copy_size++;  // trailing \0
   char *res = REAL(strncpy)(dest, src, n);
@@ -309,7 +303,7 @@ INTERCEPTOR(char *, strncpy, char *dest, const char *src, SIZE_T n) {
 INTERCEPTOR(char *, stpcpy, char *dest, const char *src) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
-  SIZE_T n = REAL(strlen)(src);
+  SIZE_T n = internal_strlen(src);
   CHECK_UNPOISONED_STRING(src + n, 0);
   char *res = REAL(stpcpy)(dest, src);
   CopyShadowAndOrigin(dest, src, n + 1, &stack);
@@ -325,7 +319,7 @@ INTERCEPTOR(char *, strdup, char *src) {
   GET_STORE_STACK_TRACE;
   // On FreeBSD strdup() leverages strlen().
   InterceptorScope interceptor_scope;
-  SIZE_T n = REAL(strlen)(src);
+  SIZE_T n = internal_strlen(src);
   CHECK_UNPOISONED_STRING(src + n, 0);
   char *res = REAL(strdup)(src);
   CopyShadowAndOrigin(res, src, n + 1, &stack);
@@ -336,7 +330,7 @@ INTERCEPTOR(char *, strdup, char *src) {
 INTERCEPTOR(char *, __strdup, char *src) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
-  SIZE_T n = REAL(strlen)(src);
+  SIZE_T n = internal_strlen(src);
   CHECK_UNPOISONED_STRING(src + n, 0);
   char *res = REAL(__strdup)(src);
   CopyShadowAndOrigin(res, src, n + 1, &stack);
@@ -351,7 +345,7 @@ INTERCEPTOR(char *, __strdup, char *src) {
 INTERCEPTOR(char *, gcvt, double number, SIZE_T ndigit, char *buf) {
   ENSURE_MSAN_INITED();
   char *res = REAL(gcvt)(number, ndigit, buf);
-  SIZE_T n = REAL(strlen)(buf);
+  SIZE_T n = internal_strlen(buf);
   __msan_unpoison(buf, n + 1);
   return res;
 }
@@ -363,8 +357,8 @@ INTERCEPTOR(char *, gcvt, double number, SIZE_T ndigit, char *buf) {
 INTERCEPTOR(char *, strcat, char *dest, const char *src) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
-  SIZE_T src_size = REAL(strlen)(src);
-  SIZE_T dest_size = REAL(strlen)(dest);
+  SIZE_T src_size = internal_strlen(src);
+  SIZE_T dest_size = internal_strlen(dest);
   CHECK_UNPOISONED_STRING(src + src_size, 0);
   CHECK_UNPOISONED_STRING(dest + dest_size, 0);
   char *res = REAL(strcat)(dest, src);
@@ -375,8 +369,8 @@ INTERCEPTOR(char *, strcat, char *dest, const char *src) {
 INTERCEPTOR(char *, strncat, char *dest, const char *src, SIZE_T n) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
-  SIZE_T dest_size = REAL(strlen)(dest);
-  SIZE_T copy_size = REAL(strnlen)(src, n);
+  SIZE_T dest_size = internal_strlen(dest);
+  SIZE_T copy_size = internal_strnlen(src, n);
   CHECK_UNPOISONED_STRING(dest + dest_size, 0);
   char *res = REAL(strncat)(dest, src, n);
   CopyShadowAndOrigin(dest + dest_size, src, copy_size, &stack);
@@ -612,7 +606,8 @@ INTERCEPTOR(char *, fcvt, double x, int a, int *b, int *c) {
   char *res = REAL(fcvt)(x, a, b, c);
   __msan_unpoison(b, sizeof(*b));
   __msan_unpoison(c, sizeof(*c));
-  if (res) __msan_unpoison(res, REAL(strlen)(res) + 1);
+  if (res)
+    __msan_unpoison(res, internal_strlen(res) + 1);
   return res;
 }
 #define MSAN_MAYBE_INTERCEPT_FCVT INTERCEPT_FUNCTION(fcvt)
@@ -625,7 +620,8 @@ INTERCEPTOR(char *, getenv, char *name) {
     return REAL(getenv)(name);
   ENSURE_MSAN_INITED();
   char *res = REAL(getenv)(name);
-  if (res) __msan_unpoison(res, REAL(strlen)(res) + 1);
+  if (res)
+    __msan_unpoison(res, internal_strlen(res) + 1);
   return res;
 }
 
@@ -635,7 +631,7 @@ static void UnpoisonEnviron() {
   char **envp = environ;
   for (; *envp; ++envp) {
     __msan_unpoison(envp, sizeof(*envp));
-    __msan_unpoison(*envp, REAL(strlen)(*envp) + 1);
+    __msan_unpoison(*envp, internal_strlen(*envp) + 1);
   }
   // Trailing NULL pointer.
   __msan_unpoison(envp, sizeof(*envp));
@@ -656,7 +652,8 @@ INTERCEPTOR(int, putenv, char *string) {
   return res;
 }
 
-#if SANITIZER_FREEBSD || SANITIZER_NETBSD
+#define SANITIZER_STAT_LINUX (SANITIZER_LINUX && __GLIBC_PREREQ(2, 33))
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD || SANITIZER_STAT_LINUX
 INTERCEPTOR(int, fstat, int fd, void *buf) {
   ENSURE_MSAN_INITED();
   int res = REAL(fstat)(fd, buf);
@@ -664,7 +661,7 @@ INTERCEPTOR(int, fstat, int fd, void *buf) {
     __msan_unpoison(buf, __sanitizer::struct_stat_sz);
   return res;
 }
-#define MSAN_MAYBE_INTERCEPT_FSTAT INTERCEPT_FUNCTION(fstat)
+#  define MSAN_MAYBE_INTERCEPT_FSTAT MSAN_INTERCEPT_FUNC(fstat)
 #else
 #define MSAN_MAYBE_INTERCEPT_FSTAT
 #endif
@@ -677,7 +674,7 @@ INTERCEPTOR(int, __fxstat, int magic, int fd, void *buf) {
     __msan_unpoison(buf, __sanitizer::struct_stat_sz);
   return res;
 }
-#define MSAN_MAYBE_INTERCEPT___FXSTAT INTERCEPT_FUNCTION(__fxstat)
+#  define MSAN_MAYBE_INTERCEPT___FXSTAT MSAN_INTERCEPT_FUNC(__fxstat)
 #else
 #define MSAN_MAYBE_INTERCEPT___FXSTAT
 #endif
@@ -690,20 +687,24 @@ INTERCEPTOR(int, __fxstat64, int magic, int fd, void *buf) {
     __msan_unpoison(buf, __sanitizer::struct_stat64_sz);
   return res;
 }
-#define MSAN_MAYBE_INTERCEPT___FXSTAT64 INTERCEPT_FUNCTION(__fxstat64)
+#  define MSAN_MAYBE_INTERCEPT___FXSTAT64 MSAN_INTERCEPT_FUNC(__fxstat64)
 #else
-#define MSAN_MAYBE_INTERCEPT___FXSTAT64
+#  define MSAN_MAYBE_INTERCEPT___FXSTAT64
 #endif
 
-#if SANITIZER_FREEBSD || SANITIZER_NETBSD
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD || SANITIZER_STAT_LINUX
 INTERCEPTOR(int, fstatat, int fd, char *pathname, void *buf, int flags) {
   ENSURE_MSAN_INITED();
   int res = REAL(fstatat)(fd, pathname, buf, flags);
   if (!res) __msan_unpoison(buf, __sanitizer::struct_stat_sz);
   return res;
 }
-# define MSAN_INTERCEPT_FSTATAT INTERCEPT_FUNCTION(fstatat)
+#  define MSAN_MAYBE_INTERCEPT_FSTATAT MSAN_INTERCEPT_FUNC(fstatat)
 #else
+#  define MSAN_MAYBE_INTERCEPT_FSTATAT
+#endif
+
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(int, __fxstatat, int magic, int fd, char *pathname, void *buf,
             int flags) {
   ENSURE_MSAN_INITED();
@@ -711,7 +712,9 @@ INTERCEPTOR(int, __fxstatat, int magic, int fd, char *pathname, void *buf,
   if (!res) __msan_unpoison(buf, __sanitizer::struct_stat_sz);
   return res;
 }
-# define MSAN_INTERCEPT_FSTATAT INTERCEPT_FUNCTION(__fxstatat)
+#  define MSAN_MAYBE_INTERCEPT___FXSTATAT MSAN_INTERCEPT_FUNC(__fxstatat)
+#else
+#  define MSAN_MAYBE_INTERCEPT___FXSTATAT
 #endif
 
 #if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
@@ -722,9 +725,9 @@ INTERCEPTOR(int, __fxstatat64, int magic, int fd, char *pathname, void *buf,
   if (!res) __msan_unpoison(buf, __sanitizer::struct_stat64_sz);
   return res;
 }
-#define MSAN_MAYBE_INTERCEPT___FXSTATAT64 INTERCEPT_FUNCTION(__fxstatat64)
+#  define MSAN_MAYBE_INTERCEPT___FXSTATAT64 MSAN_INTERCEPT_FUNC(__fxstatat64)
 #else
-#define MSAN_MAYBE_INTERCEPT___FXSTATAT64
+#  define MSAN_MAYBE_INTERCEPT___FXSTATAT64
 #endif
 
 INTERCEPTOR(int, pipe, int pipefd[2]) {
@@ -758,7 +761,7 @@ INTERCEPTOR(char *, fgets_unlocked, char *s, int size, void *stream) {
   ENSURE_MSAN_INITED();
   char *res = REAL(fgets_unlocked)(s, size, stream);
   if (res)
-    __msan_unpoison(s, REAL(strlen)(s) + 1);
+    __msan_unpoison(s, internal_strlen(s) + 1);
   return res;
 }
 #define MSAN_MAYBE_INTERCEPT_FGETS_UNLOCKED INTERCEPT_FUNCTION(fgets_unlocked)
@@ -829,7 +832,7 @@ INTERCEPTOR(int, gethostname, char *name, SIZE_T len) {
   ENSURE_MSAN_INITED();
   int res = REAL(gethostname)(name, len);
   if (!res || (res == -1 && errno == errno_ENAMETOOLONG)) {
-    SIZE_T real_len = REAL(strnlen)(name, len);
+    SIZE_T real_len = internal_strnlen(name, len);
     if (real_len < len)
       ++real_len;
     __msan_unpoison(name, real_len);
@@ -869,27 +872,15 @@ INTERCEPTOR(int, epoll_pwait, int epfd, void *events, int maxevents,
 
 INTERCEPTOR(void *, calloc, SIZE_T nmemb, SIZE_T size) {
   GET_MALLOC_STACK_TRACE;
-  if (UNLIKELY(!msan_inited))
-    // Hack: dlsym calls calloc before REAL(calloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(nmemb * size);
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Callocate(nmemb, size);
   return msan_calloc(nmemb, size, &stack);
 }
 
 INTERCEPTOR(void *, realloc, void *ptr, SIZE_T size) {
+  if (DlsymAlloc::Use() || DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Realloc(ptr, size);
   GET_MALLOC_STACK_TRACE;
-  if (UNLIKELY(IsInDlsymAllocPool(ptr))) {
-    uptr offset = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-    uptr copy_size = Min(size, kDlsymAllocPoolSize - offset);
-    void *new_ptr;
-    if (UNLIKELY(!msan_inited)) {
-      new_ptr = AllocateFromLocalPool(copy_size);
-    } else {
-      copy_size = size;
-      new_ptr = msan_malloc(copy_size, &stack);
-    }
-    internal_memcpy(new_ptr, ptr, copy_size);
-    return new_ptr;
-  }
   return msan_realloc(ptr, size, &stack);
 }
 
@@ -899,16 +890,15 @@ INTERCEPTOR(void *, reallocarray, void *ptr, SIZE_T nmemb, SIZE_T size) {
 }
 
 INTERCEPTOR(void *, malloc, SIZE_T size) {
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Allocate(size);
   GET_MALLOC_STACK_TRACE;
-  if (UNLIKELY(!msan_inited))
-    // Hack: dlsym calls malloc before REAL(malloc) is retrieved from dlsym.
-    return AllocateFromLocalPool(size);
   return msan_malloc(size, &stack);
 }
 
 void __msan_allocated_memory(const void *data, uptr size) {
-  GET_MALLOC_STACK_TRACE;
   if (flags()->poison_in_malloc) {
+    GET_MALLOC_STACK_TRACE;
     stack.tag = STACK_TRACE_TAG_POISON;
     PoisonMemory(data, size, &stack);
   }
@@ -920,8 +910,8 @@ void __msan_copy_shadow(void *dest, const void *src, uptr n) {
 }
 
 void __sanitizer_dtor_callback(const void *data, uptr size) {
-  GET_MALLOC_STACK_TRACE;
   if (flags()->poison_in_dtor) {
+    GET_MALLOC_STACK_TRACE;
     stack.tag = STACK_TRACE_TAG_POISON;
     PoisonMemory(data, size, &stack);
   }
@@ -1023,6 +1013,8 @@ extern "C" int pthread_attr_destroy(void *attr);
 static void *MsanThreadStartFunc(void *arg) {
   MsanThread *t = (MsanThread *)arg;
   SetCurrentThread(t);
+  t->Init();
+  SetSigProcMask(&t->starting_sigset_, nullptr);
   return t->ThreadStart();
 }
 
@@ -1038,7 +1030,7 @@ INTERCEPTOR(int, pthread_create, void *th, void *attr, void *(*callback)(void*),
   AdjustStackSize(attr);
 
   MsanThread *t = MsanThread::Create(callback, param);
-
+  ScopedBlockSignals block(&t->starting_sigset_);
   int res = REAL(pthread_create)(th, attr, MsanThreadStartFunc, t);
 
   if (attr == &myattr)
@@ -1080,9 +1072,9 @@ INTERCEPTOR(void, tzset, int fake) {
   InterceptorScope interceptor_scope;
   REAL(tzset)(fake);
   if (tzname[0])
-    __msan_unpoison(tzname[0], REAL(strlen)(tzname[0]) + 1);
+    __msan_unpoison(tzname[0], internal_strlen(tzname[0]) + 1);
   if (tzname[1])
-    __msan_unpoison(tzname[1], REAL(strlen)(tzname[1]) + 1);
+    __msan_unpoison(tzname[1], internal_strlen(tzname[1]) + 1);
   return;
 }
 
@@ -1092,7 +1084,7 @@ struct MSanAtExitRecord {
 };
 
 struct InterceptorContext {
-  BlockingMutex atexit_mu;
+  Mutex atexit_mu;
   Vector<struct MSanAtExitRecord *> AtExitStack;
 
   InterceptorContext()
@@ -1108,7 +1100,7 @@ InterceptorContext *interceptor_ctx() {
 void MSanAtExitWrapper() {
   MSanAtExitRecord *r;
   {
-    BlockingMutexLock l(&interceptor_ctx()->atexit_mu);
+    Lock l(&interceptor_ctx()->atexit_mu);
 
     uptr element = interceptor_ctx()->AtExitStack.Size() - 1;
     r = interceptor_ctx()->AtExitStack[element];
@@ -1142,7 +1134,7 @@ INTERCEPTOR(int, __cxa_atexit, void (*func)(void *), void *arg,
 
 // Unpoison argument shadow for C++ module destructors.
 INTERCEPTOR(int, atexit, void (*func)()) {
-  // Avoid calling real atexit as it is unrechable on at least on Linux.
+  // Avoid calling real atexit as it is unreachable on at least on Linux.
   if (msan_init_is_running)
     return REAL(__cxa_atexit)((void (*)(void *a))func, 0, 0);
   return setup_at_exit_wrapper((void(*)())func, 0, 0);
@@ -1159,7 +1151,7 @@ static int setup_at_exit_wrapper(void(*f)(), void *arg, void *dso) {
     // NetBSD does not preserve the 2nd argument if dso is equal to 0
     // Store ctx in a local stack-like structure
 
-    BlockingMutexLock l(&interceptor_ctx()->atexit_mu);
+    Lock l(&interceptor_ctx()->atexit_mu);
 
     res = REAL(__cxa_atexit)((void (*)(void *a))MSanAtExitWrapper, 0, 0);
     if (!res) {
@@ -1256,13 +1248,13 @@ int OnExit() {
   do {                                                                     \
     if (!INTERCEPT_FUNCTION_VER(name, ver))                                \
       VReport(1, "MemorySanitizer: failed to intercept '%s@@%s'\n", #name, \
-              #ver);                                                       \
+              ver);                                                        \
   } while (0)
 #define MSAN_INTERCEPT_FUNC_VER_UNVERSIONED_FALLBACK(name, ver)             \
   do {                                                                      \
     if (!INTERCEPT_FUNCTION_VER(name, ver) && !INTERCEPT_FUNCTION(name))    \
       VReport(1, "MemorySanitizer: failed to intercept '%s@@%s' or '%s'\n", \
-              #name, #ver, #name);                                          \
+              #name, ver, #name);                                           \
   } while (0)
 
 #define COMMON_INTERCEPT_FUNCTION(name) MSAN_INTERCEPT_FUNC(name)
@@ -1278,14 +1270,15 @@ int OnExit() {
   CHECK_UNPOISONED_CTX(ctx, ptr, size)
 #define COMMON_INTERCEPTOR_INITIALIZE_RANGE(ptr, size) \
   __msan_unpoison(ptr, size)
-#define COMMON_INTERCEPTOR_ENTER(ctx, func, ...)                  \
-  if (msan_init_is_running) return REAL(func)(__VA_ARGS__);       \
-  ENSURE_MSAN_INITED();                                           \
-  MSanInterceptorContext msan_ctx = {IsInInterceptorScope()};     \
-  ctx = (void *)&msan_ctx;                                        \
-  (void)ctx;                                                      \
-  InterceptorScope interceptor_scope;                             \
-  __msan_unpoison(__errno_location(), sizeof(int)); /* NOLINT */
+#define COMMON_INTERCEPTOR_ENTER(ctx, func, ...)              \
+  if (msan_init_is_running)                                   \
+    return REAL(func)(__VA_ARGS__);                           \
+  ENSURE_MSAN_INITED();                                       \
+  MSanInterceptorContext msan_ctx = {IsInInterceptorScope()}; \
+  ctx = (void *)&msan_ctx;                                    \
+  (void)ctx;                                                  \
+  InterceptorScope interceptor_scope;                         \
+  __msan_unpoison(__errno_location(), sizeof(int));
 #define COMMON_INTERCEPTOR_DIR_ACQUIRE(ctx, path) \
   do {                                            \
   } while (false)
@@ -1454,9 +1447,9 @@ INTERCEPTOR(int, dladdr, void *addr, dlinfo *info) {
   if (res != 0) {
     __msan_unpoison(info, sizeof(*info));
     if (info->dli_fname)
-      __msan_unpoison(info->dli_fname, REAL(strlen)(info->dli_fname) + 1);
+      __msan_unpoison(info->dli_fname, internal_strlen(info->dli_fname) + 1);
     if (info->dli_sname)
-      __msan_unpoison(info->dli_sname, REAL(strlen)(info->dli_sname) + 1);
+      __msan_unpoison(info->dli_sname, internal_strlen(info->dli_sname) + 1);
   }
   return res;
 }
@@ -1465,7 +1458,8 @@ INTERCEPTOR(char *, dlerror, int fake) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, dlerror, fake);
   char *res = REAL(dlerror)(fake);
-  if (res) __msan_unpoison(res, REAL(strlen)(res) + 1);
+  if (res)
+    __msan_unpoison(res, internal_strlen(res) + 1);
   return res;
 }
 
@@ -1483,7 +1477,7 @@ static int msan_dl_iterate_phdr_cb(__sanitizer_dl_phdr_info *info, SIZE_T size,
     if (info->dlpi_phdr && info->dlpi_phnum)
       __msan_unpoison(info->dlpi_phdr, struct_ElfW_Phdr_sz * info->dlpi_phnum);
     if (info->dlpi_name)
-      __msan_unpoison(info->dlpi_name, REAL(strlen)(info->dlpi_name) + 1);
+      __msan_unpoison(info->dlpi_name, internal_strlen(info->dlpi_name) + 1);
   }
   dl_iterate_phdr_data *cbdata = (dl_iterate_phdr_data *)data;
   UnpoisonParam(3);
@@ -1525,7 +1519,7 @@ INTERCEPTOR(wchar_t *, wcscpy, wchar_t *dest, const wchar_t *src) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
   wchar_t *res = REAL(wcscpy)(dest, src);
-  CopyShadowAndOrigin(dest, src, sizeof(wchar_t) * (REAL(wcslen)(src) + 1),
+  CopyShadowAndOrigin(dest, src, sizeof(wchar_t) * (internal_wcslen(src) + 1),
                       &stack);
   return res;
 }
@@ -1533,7 +1527,7 @@ INTERCEPTOR(wchar_t *, wcscpy, wchar_t *dest, const wchar_t *src) {
 INTERCEPTOR(wchar_t *, wcsncpy, wchar_t *dest, const wchar_t *src, SIZE_T n) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
-  SIZE_T copy_size = REAL(wcsnlen)(src, n);
+  SIZE_T copy_size = internal_wcsnlen(src, n);
   if (copy_size < n) copy_size++;           // trailing \0
   wchar_t *res = REAL(wcsncpy)(dest, src, n);
   CopyShadowAndOrigin(dest, src, copy_size * sizeof(wchar_t), &stack);
@@ -1597,7 +1591,7 @@ void *__msan_memmove(void *dest, const void *src, SIZE_T n) {
 
 void __msan_unpoison_string(const char* s) {
   if (!MEM_IS_APP(s)) return;
-  __msan_unpoison(s, REAL(strlen)(s) + 1);
+  __msan_unpoison(s, internal_strlen(s) + 1);
 }
 
 namespace __msan {
@@ -1686,7 +1680,8 @@ void InitializeInterceptors() {
   MSAN_MAYBE_INTERCEPT_FCVT;
   MSAN_MAYBE_INTERCEPT_FSTAT;
   MSAN_MAYBE_INTERCEPT___FXSTAT;
-  MSAN_INTERCEPT_FSTATAT;
+  MSAN_MAYBE_INTERCEPT_FSTATAT;
+  MSAN_MAYBE_INTERCEPT___FXSTATAT;
   MSAN_MAYBE_INTERCEPT___FXSTAT64;
   MSAN_MAYBE_INTERCEPT___FXSTATAT64;
   INTERCEPT_FUNCTION(pipe);
diff --git a/compiler-rt/lib/msan/msan_interface_internal.h b/compiler-rt/lib/msan/msan_interface_internal.h
index 1edacbc7504f..c72c91c3c160 100644
--- a/compiler-rt/lib/msan/msan_interface_internal.h
+++ b/compiler-rt/lib/msan/msan_interface_internal.h
@@ -31,7 +31,7 @@ SANITIZER_INTERFACE_ATTRIBUTE
 void __msan_warning();
 
 // Print a warning and die.
-// Intrumentation inserts calls to this function when building in "fast" mode
+// Instrumentation inserts calls to this function when building in "fast" mode
 // (i.e. -mllvm -msan-keep-going)
 SANITIZER_INTERFACE_ATTRIBUTE __attribute__((noreturn))
 void __msan_warning_noreturn();
diff --git a/compiler-rt/lib/msan/msan_linux.cpp b/compiler-rt/lib/msan/msan_linux.cpp
index d5baee38e710..bced00ba2428 100644
--- a/compiler-rt/lib/msan/msan_linux.cpp
+++ b/compiler-rt/lib/msan/msan_linux.cpp
@@ -37,7 +37,7 @@ namespace __msan {
 void ReportMapRange(const char *descr, uptr beg, uptr size) {
   if (size > 0) {
     uptr end = beg + size - 1;
-    VPrintf(1, "%s : %p - %p\n", descr, beg, end);
+    VPrintf(1, "%s : 0x%zx - 0x%zx\n", descr, beg, end);
   }
 }
 
@@ -45,7 +45,7 @@ static bool CheckMemoryRangeAvailability(uptr beg, uptr size) {
   if (size > 0) {
     uptr end = beg + size - 1;
     if (!MemoryRangeIsAvailable(beg, end)) {
-      Printf("FATAL: Memory range %p - %p is not available.\n", beg, end);
+      Printf("FATAL: Memory range 0x%zx - 0x%zx is not available.\n", beg, end);
       return false;
     }
   }
@@ -65,8 +65,8 @@ static bool ProtectMemoryRange(uptr beg, uptr size, const char *name) {
     }
     if ((uptr)addr != beg) {
       uptr end = beg + size - 1;
-      Printf("FATAL: Cannot protect memory range %p - %p (%s).\n", beg, end,
-             name);
+      Printf("FATAL: Cannot protect memory range 0x%zx - 0x%zx (%s).\n", beg,
+             end, name);
       return false;
     }
   }
@@ -106,7 +106,7 @@ static void CheckMemoryLayoutSanity() {
 
 bool InitShadow(bool init_origins) {
   // Let user know mapping parameters first.
-  VPrintf(1, "__msan_init %p\n", &__msan_init);
+  VPrintf(1, "__msan_init %p\n", reinterpret_cast<void *>(&__msan_init));
   for (unsigned i = 0; i < kMemoryLayoutSize; ++i)
     VPrintf(1, "%s: %zx - %zx\n", kMemoryLayout[i].name, kMemoryLayout[i].start,
             kMemoryLayout[i].end - 1);
@@ -115,7 +115,7 @@ bool InitShadow(bool init_origins) {
 
   if (!MEM_IS_APP(&__msan_init)) {
     Printf("FATAL: Code %p is out of application range. Non-PIE build?\n",
-           (uptr)&__msan_init);
+           reinterpret_cast<void *>(&__msan_init));
     return false;
   }
 
diff --git a/compiler-rt/lib/msan/msan_poisoning.cpp b/compiler-rt/lib/msan/msan_poisoning.cpp
index 15892392f74a..af01aa69f78f 100644
--- a/compiler-rt/lib/msan/msan_poisoning.cpp
+++ b/compiler-rt/lib/msan/msan_poisoning.cpp
@@ -14,6 +14,7 @@
 
 #include "interception/interception.h"
 #include "msan_origin.h"
+#include "msan_thread.h"
 #include "sanitizer_common/sanitizer_common.h"
 
 DECLARE_REAL(void *, memset, void *dest, int c, uptr n)
@@ -241,6 +242,9 @@ void PoisonMemory(const void *dst, uptr size, StackTrace *stack) {
   SetShadow(dst, size, (u8)-1);
 
   if (__msan_get_track_origins()) {
+    MsanThread *t = GetCurrentThread();
+    if (t && t->InSignalHandler())
+      return;
     Origin o = Origin::CreateHeapOrigin(stack);
     SetOrigin(dst, size, o.raw_id());
   }
diff --git a/compiler-rt/lib/msan/msan_report.cpp b/compiler-rt/lib/msan/msan_report.cpp
index e10d9eb62231..ff3e38c7db9e 100644
--- a/compiler-rt/lib/msan/msan_report.cpp
+++ b/compiler-rt/lib/msan/msan_report.cpp
@@ -122,17 +122,17 @@ void ReportStats() {
   ScopedErrorReportLock l;
 
   if (__msan_get_track_origins() > 0) {
-    StackDepotStats *stack_depot_stats = StackDepotGetStats();
+    StackDepotStats stack_depot_stats = StackDepotGetStats();
     // FIXME: we want this at normal exit, too!
     // FIXME: but only with verbosity=1 or something
-    Printf("Unique heap origins: %zu\n", stack_depot_stats->n_uniq_ids);
-    Printf("Stack depot allocated bytes: %zu\n", stack_depot_stats->allocated);
+    Printf("Unique heap origins: %zu\n", stack_depot_stats.n_uniq_ids);
+    Printf("Stack depot allocated bytes: %zu\n", stack_depot_stats.allocated);
 
-    StackDepotStats *chained_origin_depot_stats = ChainedOriginDepotGetStats();
+    StackDepotStats chained_origin_depot_stats = ChainedOriginDepotGetStats();
     Printf("Unique origin histories: %zu\n",
-           chained_origin_depot_stats->n_uniq_ids);
+           chained_origin_depot_stats.n_uniq_ids);
     Printf("History depot allocated bytes: %zu\n",
-           chained_origin_depot_stats->allocated);
+           chained_origin_depot_stats.allocated);
   }
 }
 
@@ -201,13 +201,18 @@ void DescribeMemoryRange(const void *x, uptr size) {
 
   Decorator d;
   Printf("%s", d.Warning());
-  Printf("Shadow map of [%p, %p), %zu bytes:\n", start, end, end - start);
+  uptr start_x = reinterpret_cast<uptr>(x);
+  Printf("Shadow map [%p, %p) of [%p, %p), %zu bytes:\n",
+         reinterpret_cast<void *>(start), reinterpret_cast<void *>(end),
+         reinterpret_cast<void *>(start_x),
+         reinterpret_cast<void *>(start_x + end - start), end - start);
   Printf("%s", d.Default());
   while (s < e) {
     // Line start.
     if (pos % 16 == 0) {
       for (int i = 0; i < 4; ++i) origin_ids[i] = -1;
-      Printf("%p:", s);
+      Printf("%p[%p]:", reinterpret_cast<void *>(s),
+             reinterpret_cast<void *>(start_x - start + s));
     }
     // Group start.
     if (pos % 4 == 0) {
diff --git a/compiler-rt/lib/msan/msan_thread.cpp b/compiler-rt/lib/msan/msan_thread.cpp
index 6ae012acd9a2..40ad6a5019c4 100644
--- a/compiler-rt/lib/msan/msan_thread.cpp
+++ b/compiler-rt/lib/msan/msan_thread.cpp
@@ -66,8 +66,6 @@ void MsanThread::Destroy() {
 }
 
 thread_return_t MsanThread::ThreadStart() {
-  Init();
-
   if (!start_routine_) {
     // start_routine_ == 0 if we're on the main thread or on one of the
     // OS X libdispatch worker threads. But nobody is supposed to call
diff --git a/compiler-rt/lib/msan/msan_thread.h b/compiler-rt/lib/msan/msan_thread.h
index fe795e3a547a..f6ed1534cccd 100644
--- a/compiler-rt/lib/msan/msan_thread.h
+++ b/compiler-rt/lib/msan/msan_thread.h
@@ -15,7 +15,7 @@
 
 #include "msan_allocator.h"
 #include "sanitizer_common/sanitizer_common.h"
-
+#include "sanitizer_common/sanitizer_posix.h"
 namespace __msan {
 
 class MsanThread {
@@ -45,6 +45,7 @@ class MsanThread {
   MsanThreadLocalMallocStorage &malloc_storage() { return malloc_storage_; }
 
   int destructor_iterations_;
+  __sanitizer_sigset_t starting_sigset_;
 
  private:
   // NOTE: There is no MsanThread constructor. It is allocated
diff --git a/compiler-rt/lib/orc/c_api.h b/compiler-rt/lib/orc/c_api.h
index 6677da06ede5..47f46b891d96 100644
--- a/compiler-rt/lib/orc/c_api.h
+++ b/compiler-rt/lib/orc/c_api.h
@@ -50,7 +50,7 @@ ORC_RT_C_EXTERN_C_BEGIN
 
 typedef union {
   char *ValuePtr;
-  char Value[sizeof(ValuePtr)];
+  char Value[sizeof(char *)];
 } __orc_rt_CWrapperFunctionResultDataUnion;
 
 /**
@@ -91,15 +91,15 @@ __orc_rt_CWrapperFunctionResultInit(__orc_rt_CWrapperFunctionResult *R) {
  * Create an __orc_rt_CWrapperFunctionResult with an uninitialized buffer of
  * size Size. The buffer is returned via the DataPtr argument.
  */
-static inline char *
-__orc_rt_CWrapperFunctionResultAllocate(__orc_rt_CWrapperFunctionResult *R,
-                                        size_t Size) {
-  R->Size = Size;
-  if (Size <= sizeof(R->Data.Value))
-    return R->Data.Value;
-
-  R->Data.ValuePtr = (char *)malloc(Size);
-  return R->Data.ValuePtr;
+static inline __orc_rt_CWrapperFunctionResult
+__orc_rt_CWrapperFunctionResultAllocate(size_t Size) {
+  __orc_rt_CWrapperFunctionResult R;
+  R.Size = Size;
+  // If Size is 0 ValuePtr must be 0 or it is considered an out-of-band error.
+  R.Data.ValuePtr = 0;
+  if (Size > sizeof(R.Data.Value))
+    R.Data.ValuePtr = (char *)malloc(Size);
+  return R;
 }
 
 /**
@@ -135,8 +135,8 @@ __orc_rt_CreateCWrapperFunctionResultFromString(const char *Source) {
  * Create an __orc_rt_CWrapperFunctionResult representing an out-of-band
  * error.
  *
- * This function takes ownership of the string argument which must have been
- * allocated with malloc.
+ * This function copies the input string. The client is responsible for freeing
+ * the ErrMsg arg.
  */
 static inline __orc_rt_CWrapperFunctionResult
 __orc_rt_CreateCWrapperFunctionResultFromOutOfBandError(const char *ErrMsg) {
@@ -163,9 +163,9 @@ __orc_rt_DisposeCWrapperFunctionResult(__orc_rt_CWrapperFunctionResult *R) {
  * Get a pointer to the data contained in the given
  * __orc_rt_CWrapperFunctionResult.
  */
-static inline const char *
-__orc_rt_CWrapperFunctionResultData(const __orc_rt_CWrapperFunctionResult *R) {
-  assert((R->Size != 0 || R->Data.ValuePtr == nullptr) &&
+static inline char *
+__orc_rt_CWrapperFunctionResultData(__orc_rt_CWrapperFunctionResult *R) {
+  assert((R->Size != 0 || R->Data.ValuePtr == NULL) &&
          "Cannot get data for out-of-band error value");
   return R->Size > sizeof(R->Data.Value) ? R->Data.ValuePtr : R->Data.Value;
 }
@@ -177,7 +177,7 @@ __orc_rt_CWrapperFunctionResultData(const __orc_rt_CWrapperFunctionResult *R) {
  */
 static inline size_t
 __orc_rt_CWrapperFunctionResultSize(const __orc_rt_CWrapperFunctionResult *R) {
-  assert((R->Size != 0 || R->Data.ValuePtr == nullptr) &&
+  assert((R->Size != 0 || R->Data.ValuePtr == NULL) &&
          "Cannot get size for out-of-band error value");
   return R->Size;
 }
diff --git a/compiler-rt/lib/orc/elfnix_platform.cpp b/compiler-rt/lib/orc/elfnix_platform.cpp
new file mode 100644
index 000000000000..0352f6c4e853
--- /dev/null
+++ b/compiler-rt/lib/orc/elfnix_platform.cpp
@@ -0,0 +1,584 @@
+//===- elfnix_platform.cpp ------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file contains code required to load the rest of the ELF-on-*IX runtime.
+//
+//===----------------------------------------------------------------------===//
+
+#include "elfnix_platform.h"
+#include "common.h"
+#include "error.h"
+#include "wrapper_function_utils.h"
+
+#include <map>
+#include <mutex>
+#include <sstream>
+#include <unordered_map>
+#include <vector>
+
+using namespace __orc_rt;
+using namespace __orc_rt::elfnix;
+
+// Declare function tags for functions in the JIT process.
+ORC_RT_JIT_DISPATCH_TAG(__orc_rt_elfnix_get_initializers_tag)
+ORC_RT_JIT_DISPATCH_TAG(__orc_rt_elfnix_get_deinitializers_tag)
+ORC_RT_JIT_DISPATCH_TAG(__orc_rt_elfnix_symbol_lookup_tag)
+
+// eh-frame registration functions.
+// We expect these to be available for all processes.
+extern "C" void __register_frame(const void *);
+extern "C" void __deregister_frame(const void *);
+
+namespace {
+
+Error validatePointerSectionExtent(const char *SectionName,
+                                   const ExecutorAddrRange &SE) {
+  if (SE.size().getValue() % sizeof(uintptr_t)) {
+    std::ostringstream ErrMsg;
+    ErrMsg << std::hex << "Size of " << SectionName << " 0x"
+           << SE.Start.getValue() << " -- 0x" << SE.End.getValue()
+           << " is not a pointer multiple";
+    return make_error<StringError>(ErrMsg.str());
+  }
+  return Error::success();
+}
+
+Error runInitArray(const std::vector<ExecutorAddrRange> &InitArraySections,
+                   const ELFNixJITDylibInitializers &MOJDIs) {
+
+  for (const auto &ModInits : InitArraySections) {
+    if (auto Err = validatePointerSectionExtent(".init_array", ModInits))
+      return Err;
+
+    using InitFunc = void (*)();
+    for (auto *Init : ModInits.toSpan<InitFunc>())
+      (*Init)();
+  }
+
+  return Error::success();
+}
+struct TLSInfoEntry {
+  unsigned long Key = 0;
+  unsigned long DataAddress = 0;
+};
+
+class ELFNixPlatformRuntimeState {
+private:
+  struct AtExitEntry {
+    void (*Func)(void *);
+    void *Arg;
+  };
+
+  using AtExitsVector = std::vector<AtExitEntry>;
+
+  struct PerJITDylibState {
+    void *Header = nullptr;
+    size_t RefCount = 0;
+    bool AllowReinitialization = false;
+    AtExitsVector AtExits;
+  };
+
+public:
+  static void initialize(void *DSOHandle);
+  static ELFNixPlatformRuntimeState &get();
+  static void destroy();
+
+  ELFNixPlatformRuntimeState(void *DSOHandle)
+      : PlatformJDDSOHandle(DSOHandle) {}
+
+  // Delete copy and move constructors.
+  ELFNixPlatformRuntimeState(const ELFNixPlatformRuntimeState &) = delete;
+  ELFNixPlatformRuntimeState &
+  operator=(const ELFNixPlatformRuntimeState &) = delete;
+  ELFNixPlatformRuntimeState(ELFNixPlatformRuntimeState &&) = delete;
+  ELFNixPlatformRuntimeState &operator=(ELFNixPlatformRuntimeState &&) = delete;
+
+  Error registerObjectSections(ELFNixPerObjectSectionsToRegister POSR);
+  Error deregisterObjectSections(ELFNixPerObjectSectionsToRegister POSR);
+
+  const char *dlerror();
+  void *dlopen(string_view Name, int Mode);
+  int dlclose(void *DSOHandle);
+  void *dlsym(void *DSOHandle, string_view Symbol);
+
+  int registerAtExit(void (*F)(void *), void *Arg, void *DSOHandle);
+  void runAtExits(void *DSOHandle);
+
+  /// Returns the base address of the section containing ThreadData.
+  Expected<std::pair<const char *, size_t>>
+  getThreadDataSectionFor(const char *ThreadData);
+
+  void *getPlatformJDDSOHandle() { return PlatformJDDSOHandle; }
+
+private:
+  PerJITDylibState *getJITDylibStateByHeaderAddr(void *DSOHandle);
+  PerJITDylibState *getJITDylibStateByName(string_view Path);
+  PerJITDylibState &
+  getOrCreateJITDylibState(ELFNixJITDylibInitializers &MOJDIs);
+
+  Error registerThreadDataSection(span<const char> ThreadDataSection);
+
+  Expected<ExecutorAddr> lookupSymbolInJITDylib(void *DSOHandle,
+                                                string_view Symbol);
+
+  Expected<ELFNixJITDylibInitializerSequence>
+  getJITDylibInitializersByName(string_view Path);
+  Expected<void *> dlopenInitialize(string_view Path, int Mode);
+  Error initializeJITDylib(ELFNixJITDylibInitializers &MOJDIs);
+
+  static ELFNixPlatformRuntimeState *MOPS;
+
+  using InitSectionHandler =
+      Error (*)(const std::vector<ExecutorAddrRange> &Sections,
+                const ELFNixJITDylibInitializers &MOJDIs);
+  const std::vector<std::pair<const char *, InitSectionHandler>> InitSections =
+      {{".init_array", runInitArray}};
+
+  void *PlatformJDDSOHandle;
+
+  // FIXME: Move to thread-state.
+  std::string DLFcnError;
+
+  std::recursive_mutex JDStatesMutex;
+  std::unordered_map<void *, PerJITDylibState> JDStates;
+  std::unordered_map<std::string, void *> JDNameToHeader;
+
+  std::mutex ThreadDataSectionsMutex;
+  std::map<const char *, size_t> ThreadDataSections;
+};
+
+ELFNixPlatformRuntimeState *ELFNixPlatformRuntimeState::MOPS = nullptr;
+
+void ELFNixPlatformRuntimeState::initialize(void *DSOHandle) {
+  assert(!MOPS && "ELFNixPlatformRuntimeState should be null");
+  MOPS = new ELFNixPlatformRuntimeState(DSOHandle);
+}
+
+ELFNixPlatformRuntimeState &ELFNixPlatformRuntimeState::get() {
+  assert(MOPS && "ELFNixPlatformRuntimeState not initialized");
+  return *MOPS;
+}
+
+void ELFNixPlatformRuntimeState::destroy() {
+  assert(MOPS && "ELFNixPlatformRuntimeState not initialized");
+  delete MOPS;
+}
+
+Error ELFNixPlatformRuntimeState::registerObjectSections(
+    ELFNixPerObjectSectionsToRegister POSR) {
+  if (POSR.EHFrameSection.Start)
+    __register_frame(POSR.EHFrameSection.Start.toPtr<const char *>());
+
+  if (POSR.ThreadDataSection.Start) {
+    if (auto Err = registerThreadDataSection(
+            POSR.ThreadDataSection.toSpan<const char>()))
+      return Err;
+  }
+
+  return Error::success();
+}
+
+Error ELFNixPlatformRuntimeState::deregisterObjectSections(
+    ELFNixPerObjectSectionsToRegister POSR) {
+  if (POSR.EHFrameSection.Start)
+    __deregister_frame(POSR.EHFrameSection.Start.toPtr<const char *>());
+
+  return Error::success();
+}
+
+const char *ELFNixPlatformRuntimeState::dlerror() { return DLFcnError.c_str(); }
+
+void *ELFNixPlatformRuntimeState::dlopen(string_view Path, int Mode) {
+  std::lock_guard<std::recursive_mutex> Lock(JDStatesMutex);
+
+  // Use fast path if all JITDylibs are already loaded and don't require
+  // re-running initializers.
+  if (auto *JDS = getJITDylibStateByName(Path)) {
+    if (!JDS->AllowReinitialization) {
+      ++JDS->RefCount;
+      return JDS->Header;
+    }
+  }
+
+  auto H = dlopenInitialize(Path, Mode);
+  if (!H) {
+    DLFcnError = toString(H.takeError());
+    return nullptr;
+  }
+
+  return *H;
+}
+
+int ELFNixPlatformRuntimeState::dlclose(void *DSOHandle) {
+  runAtExits(DSOHandle);
+  return 0;
+}
+
+void *ELFNixPlatformRuntimeState::dlsym(void *DSOHandle, string_view Symbol) {
+  auto Addr = lookupSymbolInJITDylib(DSOHandle, Symbol);
+  if (!Addr) {
+    DLFcnError = toString(Addr.takeError());
+    return 0;
+  }
+
+  return Addr->toPtr<void *>();
+}
+
+int ELFNixPlatformRuntimeState::registerAtExit(void (*F)(void *), void *Arg,
+                                               void *DSOHandle) {
+  // FIXME: Handle out-of-memory errors, returning -1 if OOM.
+  std::lock_guard<std::recursive_mutex> Lock(JDStatesMutex);
+  auto *JDS = getJITDylibStateByHeaderAddr(DSOHandle);
+  assert(JDS && "JITDylib state not initialized");
+  JDS->AtExits.push_back({F, Arg});
+  return 0;
+}
+
+void ELFNixPlatformRuntimeState::runAtExits(void *DSOHandle) {
+  // FIXME: Should atexits be allowed to run concurrently with access to
+  // JDState?
+  AtExitsVector V;
+  {
+    std::lock_guard<std::recursive_mutex> Lock(JDStatesMutex);
+    auto *JDS = getJITDylibStateByHeaderAddr(DSOHandle);
+    assert(JDS && "JITDlybi state not initialized");
+    std::swap(V, JDS->AtExits);
+  }
+
+  while (!V.empty()) {
+    auto &AE = V.back();
+    AE.Func(AE.Arg);
+    V.pop_back();
+  }
+}
+
+Expected<std::pair<const char *, size_t>>
+ELFNixPlatformRuntimeState::getThreadDataSectionFor(const char *ThreadData) {
+  std::lock_guard<std::mutex> Lock(ThreadDataSectionsMutex);
+  auto I = ThreadDataSections.upper_bound(ThreadData);
+  // Check that we have a valid entry conovering this address.
+  if (I == ThreadDataSections.begin())
+    return make_error<StringError>("No thread local data section for key");
+  I = std::prev(I);
+  if (ThreadData >= I->first + I->second)
+    return make_error<StringError>("No thread local data section for key");
+  return *I;
+}
+
+ELFNixPlatformRuntimeState::PerJITDylibState *
+ELFNixPlatformRuntimeState::getJITDylibStateByHeaderAddr(void *DSOHandle) {
+  auto I = JDStates.find(DSOHandle);
+  if (I == JDStates.end())
+    return nullptr;
+  return &I->second;
+}
+
+ELFNixPlatformRuntimeState::PerJITDylibState *
+ELFNixPlatformRuntimeState::getJITDylibStateByName(string_view Name) {
+  // FIXME: Avoid creating string copy here.
+  auto I = JDNameToHeader.find(std::string(Name.data(), Name.size()));
+  if (I == JDNameToHeader.end())
+    return nullptr;
+  void *H = I->second;
+  auto J = JDStates.find(H);
+  assert(J != JDStates.end() &&
+         "JITDylib has name map entry but no header map entry");
+  return &J->second;
+}
+
+ELFNixPlatformRuntimeState::PerJITDylibState &
+ELFNixPlatformRuntimeState::getOrCreateJITDylibState(
+    ELFNixJITDylibInitializers &MOJDIs) {
+  void *Header = MOJDIs.DSOHandleAddress.toPtr<void *>();
+
+  auto &JDS = JDStates[Header];
+
+  // If this entry hasn't been created yet.
+  if (!JDS.Header) {
+    assert(!JDNameToHeader.count(MOJDIs.Name) &&
+           "JITDylib has header map entry but no name map entry");
+    JDNameToHeader[MOJDIs.Name] = Header;
+    JDS.Header = Header;
+  }
+
+  return JDS;
+}
+
+Error ELFNixPlatformRuntimeState::registerThreadDataSection(
+    span<const char> ThreadDataSection) {
+  std::lock_guard<std::mutex> Lock(ThreadDataSectionsMutex);
+  auto I = ThreadDataSections.upper_bound(ThreadDataSection.data());
+  if (I != ThreadDataSections.begin()) {
+    auto J = std::prev(I);
+    if (J->first + J->second > ThreadDataSection.data())
+      return make_error<StringError>("Overlapping .tdata sections");
+  }
+  ThreadDataSections.insert(
+      I, std::make_pair(ThreadDataSection.data(), ThreadDataSection.size()));
+  return Error::success();
+}
+
+Expected<ExecutorAddr>
+ELFNixPlatformRuntimeState::lookupSymbolInJITDylib(void *DSOHandle,
+                                                   string_view Sym) {
+  Expected<ExecutorAddr> Result((ExecutorAddr()));
+  if (auto Err = WrapperFunction<SPSExpected<SPSExecutorAddr>(
+          SPSExecutorAddr, SPSString)>::call(&__orc_rt_elfnix_symbol_lookup_tag,
+                                             Result,
+                                             ExecutorAddr::fromPtr(DSOHandle),
+                                             Sym))
+    return std::move(Err);
+  return Result;
+}
+
+Expected<ELFNixJITDylibInitializerSequence>
+ELFNixPlatformRuntimeState::getJITDylibInitializersByName(string_view Path) {
+  Expected<ELFNixJITDylibInitializerSequence> Result(
+      (ELFNixJITDylibInitializerSequence()));
+  std::string PathStr(Path.data(), Path.size());
+  if (auto Err =
+          WrapperFunction<SPSExpected<SPSELFNixJITDylibInitializerSequence>(
+              SPSString)>::call(&__orc_rt_elfnix_get_initializers_tag, Result,
+                                Path))
+    return std::move(Err);
+  return Result;
+}
+
+Expected<void *> ELFNixPlatformRuntimeState::dlopenInitialize(string_view Path,
+                                                              int Mode) {
+  // Either our JITDylib wasn't loaded, or it or one of its dependencies allows
+  // reinitialization. We need to call in to the JIT to see if there's any new
+  // work pending.
+  auto InitSeq = getJITDylibInitializersByName(Path);
+  if (!InitSeq)
+    return InitSeq.takeError();
+
+  // Init sequences should be non-empty.
+  if (InitSeq->empty())
+    return make_error<StringError>(
+        "__orc_rt_elfnix_get_initializers returned an "
+        "empty init sequence");
+
+  // Otherwise register and run initializers for each JITDylib.
+  for (auto &MOJDIs : *InitSeq)
+    if (auto Err = initializeJITDylib(MOJDIs))
+      return std::move(Err);
+
+  // Return the header for the last item in the list.
+  auto *JDS = getJITDylibStateByHeaderAddr(
+      InitSeq->back().DSOHandleAddress.toPtr<void *>());
+  assert(JDS && "Missing state entry for JD");
+  return JDS->Header;
+}
+
+Error ELFNixPlatformRuntimeState::initializeJITDylib(
+    ELFNixJITDylibInitializers &MOJDIs) {
+
+  auto &JDS = getOrCreateJITDylibState(MOJDIs);
+  ++JDS.RefCount;
+
+  for (auto &KV : InitSections) {
+    const auto &Name = KV.first;
+    const auto &Handler = KV.second;
+    auto I = MOJDIs.InitSections.find(Name);
+    if (I != MOJDIs.InitSections.end()) {
+      if (auto Err = Handler(I->second, MOJDIs))
+        return Err;
+    }
+  }
+
+  return Error::success();
+}
+class ELFNixPlatformRuntimeTLVManager {
+public:
+  void *getInstance(const char *ThreadData);
+
+private:
+  std::unordered_map<const char *, char *> Instances;
+  std::unordered_map<const char *, std::unique_ptr<char[]>> AllocatedSections;
+};
+
+void *ELFNixPlatformRuntimeTLVManager::getInstance(const char *ThreadData) {
+  auto I = Instances.find(ThreadData);
+  if (I != Instances.end())
+    return I->second;
+  auto TDS =
+      ELFNixPlatformRuntimeState::get().getThreadDataSectionFor(ThreadData);
+  if (!TDS) {
+    __orc_rt_log_error(toString(TDS.takeError()).c_str());
+    return nullptr;
+  }
+
+  auto &Allocated = AllocatedSections[TDS->first];
+  if (!Allocated) {
+    Allocated = std::make_unique<char[]>(TDS->second);
+    memcpy(Allocated.get(), TDS->first, TDS->second);
+  }
+  size_t ThreadDataDelta = ThreadData - TDS->first;
+  assert(ThreadDataDelta <= TDS->second && "ThreadData outside section bounds");
+
+  char *Instance = Allocated.get() + ThreadDataDelta;
+  Instances[ThreadData] = Instance;
+  return Instance;
+}
+
+void destroyELFNixTLVMgr(void *ELFNixTLVMgr) {
+  delete static_cast<ELFNixPlatformRuntimeTLVManager *>(ELFNixTLVMgr);
+}
+
+} // end anonymous namespace
+
+//------------------------------------------------------------------------------
+//                             JIT entry points
+//------------------------------------------------------------------------------
+
+ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
+__orc_rt_elfnix_platform_bootstrap(char *ArgData, size_t ArgSize) {
+  return WrapperFunction<void(uint64_t)>::handle(
+             ArgData, ArgSize,
+             [](uint64_t &DSOHandle) {
+               ELFNixPlatformRuntimeState::initialize(
+                   reinterpret_cast<void *>(DSOHandle));
+             })
+      .release();
+}
+
+ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
+__orc_rt_elfnix_platform_shutdown(char *ArgData, size_t ArgSize) {
+  ELFNixPlatformRuntimeState::destroy();
+  return WrapperFunctionResult().release();
+}
+
+/// Wrapper function for registering metadata on a per-object basis.
+ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
+__orc_rt_elfnix_register_object_sections(char *ArgData, size_t ArgSize) {
+  return WrapperFunction<SPSError(SPSELFNixPerObjectSectionsToRegister)>::
+      handle(ArgData, ArgSize,
+             [](ELFNixPerObjectSectionsToRegister &POSR) {
+               return ELFNixPlatformRuntimeState::get().registerObjectSections(
+                   std::move(POSR));
+             })
+          .release();
+}
+
+/// Wrapper for releasing per-object metadat.
+ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
+__orc_rt_elfnix_deregister_object_sections(char *ArgData, size_t ArgSize) {
+  return WrapperFunction<SPSError(SPSELFNixPerObjectSectionsToRegister)>::
+      handle(ArgData, ArgSize,
+             [](ELFNixPerObjectSectionsToRegister &POSR) {
+               return ELFNixPlatformRuntimeState::get()
+                   .deregisterObjectSections(std::move(POSR));
+             })
+          .release();
+}
+
+//------------------------------------------------------------------------------
+//                           TLV support
+//------------------------------------------------------------------------------
+
+ORC_RT_INTERFACE void *__orc_rt_elfnix_tls_get_addr_impl(TLSInfoEntry *D) {
+  auto *TLVMgr = static_cast<ELFNixPlatformRuntimeTLVManager *>(
+      pthread_getspecific(D->Key));
+  if (!TLVMgr)
+    TLVMgr = new ELFNixPlatformRuntimeTLVManager();
+  if (pthread_setspecific(D->Key, TLVMgr)) {
+    __orc_rt_log_error("Call to pthread_setspecific failed");
+    return nullptr;
+  }
+
+  return TLVMgr->getInstance(
+      reinterpret_cast<char *>(static_cast<uintptr_t>(D->DataAddress)));
+}
+
+ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
+__orc_rt_elfnix_create_pthread_key(char *ArgData, size_t ArgSize) {
+  return WrapperFunction<SPSExpected<uint64_t>(void)>::handle(
+             ArgData, ArgSize,
+             []() -> Expected<uint64_t> {
+               pthread_key_t Key;
+               if (int Err = pthread_key_create(&Key, destroyELFNixTLVMgr)) {
+                 __orc_rt_log_error("Call to pthread_key_create failed");
+                 return make_error<StringError>(strerror(Err));
+               }
+               return static_cast<uint64_t>(Key);
+             })
+      .release();
+}
+
+//------------------------------------------------------------------------------
+//                           cxa_atexit support
+//------------------------------------------------------------------------------
+
+int __orc_rt_elfnix_cxa_atexit(void (*func)(void *), void *arg,
+                               void *dso_handle) {
+  return ELFNixPlatformRuntimeState::get().registerAtExit(func, arg,
+                                                          dso_handle);
+}
+
+int __orc_rt_elfnix_atexit(void (*func)(void *)) {
+  auto &PlatformRTState = ELFNixPlatformRuntimeState::get();
+  return ELFNixPlatformRuntimeState::get().registerAtExit(
+      func, NULL, PlatformRTState.getPlatformJDDSOHandle());
+}
+
+void __orc_rt_elfnix_cxa_finalize(void *dso_handle) {
+  ELFNixPlatformRuntimeState::get().runAtExits(dso_handle);
+}
+
+//------------------------------------------------------------------------------
+//                        JIT'd dlfcn alternatives.
+//------------------------------------------------------------------------------
+
+const char *__orc_rt_elfnix_jit_dlerror() {
+  return ELFNixPlatformRuntimeState::get().dlerror();
+}
+
+void *__orc_rt_elfnix_jit_dlopen(const char *path, int mode) {
+  return ELFNixPlatformRuntimeState::get().dlopen(path, mode);
+}
+
+int __orc_rt_elfnix_jit_dlclose(void *dso_handle) {
+  return ELFNixPlatformRuntimeState::get().dlclose(dso_handle);
+}
+
+void *__orc_rt_elfnix_jit_dlsym(void *dso_handle, const char *symbol) {
+  return ELFNixPlatformRuntimeState::get().dlsym(dso_handle, symbol);
+}
+
+//------------------------------------------------------------------------------
+//                             ELFNix Run Program
+//------------------------------------------------------------------------------
+
+ORC_RT_INTERFACE int64_t __orc_rt_elfnix_run_program(
+    const char *JITDylibName, const char *EntrySymbolName, int argc,
+    char *argv[]) {
+  using MainTy = int (*)(int, char *[]);
+
+  void *H = __orc_rt_elfnix_jit_dlopen(JITDylibName,
+                                       __orc_rt::elfnix::ORC_RT_RTLD_LAZY);
+  if (!H) {
+    __orc_rt_log_error(__orc_rt_elfnix_jit_dlerror());
+    return -1;
+  }
+
+  auto *Main =
+      reinterpret_cast<MainTy>(__orc_rt_elfnix_jit_dlsym(H, EntrySymbolName));
+
+  if (!Main) {
+    __orc_rt_log_error(__orc_rt_elfnix_jit_dlerror());
+    return -1;
+  }
+
+  int Result = Main(argc, argv);
+
+  if (__orc_rt_elfnix_jit_dlclose(H) == -1)
+    __orc_rt_log_error(__orc_rt_elfnix_jit_dlerror());
+
+  return Result;
+}
diff --git a/compiler-rt/lib/orc/elfnix_platform.h b/compiler-rt/lib/orc/elfnix_platform.h
new file mode 100644
index 000000000000..12b9591979b7
--- /dev/null
+++ b/compiler-rt/lib/orc/elfnix_platform.h
@@ -0,0 +1,131 @@
+//===- elfnix_platform.h ----------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// ORC Runtime support for dynamic loading features on ELF-based platforms.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef ORC_RT_ELFNIX_PLATFORM_H
+#define ORC_RT_ELFNIX_PLATFORM_H
+
+#include "common.h"
+#include "executor_address.h"
+
+// Atexit functions.
+ORC_RT_INTERFACE int __orc_rt_elfnix_cxa_atexit(void (*func)(void *), void *arg,
+                                                void *dso_handle);
+ORC_RT_INTERFACE int __orc_rt_elfnix_atexit(void (*func)(void *));
+ORC_RT_INTERFACE void __orc_rt_elfnix_cxa_finalize(void *dso_handle);
+
+// dlfcn functions.
+ORC_RT_INTERFACE const char *__orc_rt_elfnix_jit_dlerror();
+ORC_RT_INTERFACE void *__orc_rt_elfnix_jit_dlopen(const char *path, int mode);
+ORC_RT_INTERFACE int __orc_rt_elfnix_jit_dlclose(void *dso_handle);
+ORC_RT_INTERFACE void *__orc_rt_elfnix_jit_dlsym(void *dso_handle,
+                                                 const char *symbol);
+
+namespace __orc_rt {
+namespace elfnix {
+
+struct ELFNixPerObjectSectionsToRegister {
+  ExecutorAddrRange EHFrameSection;
+  ExecutorAddrRange ThreadDataSection;
+};
+
+struct ELFNixJITDylibInitializers {
+  using SectionList = std::vector<ExecutorAddrRange>;
+
+  ELFNixJITDylibInitializers() = default;
+  ELFNixJITDylibInitializers(std::string Name, ExecutorAddr DSOHandleAddress)
+      : Name(std::move(Name)), DSOHandleAddress(std::move(DSOHandleAddress)) {}
+
+  std::string Name;
+  ExecutorAddr DSOHandleAddress;
+
+  std::unordered_map<std::string, SectionList> InitSections;
+};
+
+class ELFNixJITDylibDeinitializers {};
+
+using ELFNixJITDylibInitializerSequence =
+    std::vector<ELFNixJITDylibInitializers>;
+
+using ELFNixJITDylibDeinitializerSequence =
+    std::vector<ELFNixJITDylibDeinitializers>;
+
+enum dlopen_mode : int {
+  ORC_RT_RTLD_LAZY = 0x1,
+  ORC_RT_RTLD_NOW = 0x2,
+  ORC_RT_RTLD_LOCAL = 0x4,
+  ORC_RT_RTLD_GLOBAL = 0x8
+};
+
+} // end namespace elfnix
+
+using SPSELFNixPerObjectSectionsToRegister =
+    SPSTuple<SPSExecutorAddrRange, SPSExecutorAddrRange>;
+
+template <>
+class SPSSerializationTraits<SPSELFNixPerObjectSectionsToRegister,
+                             elfnix::ELFNixPerObjectSectionsToRegister> {
+
+public:
+  static size_t size(const elfnix::ELFNixPerObjectSectionsToRegister &MOPOSR) {
+    return SPSELFNixPerObjectSectionsToRegister::AsArgList::size(
+        MOPOSR.EHFrameSection, MOPOSR.ThreadDataSection);
+  }
+
+  static bool
+  serialize(SPSOutputBuffer &OB,
+            const elfnix::ELFNixPerObjectSectionsToRegister &MOPOSR) {
+    return SPSELFNixPerObjectSectionsToRegister::AsArgList::serialize(
+        OB, MOPOSR.EHFrameSection, MOPOSR.ThreadDataSection);
+  }
+
+  static bool deserialize(SPSInputBuffer &IB,
+                          elfnix::ELFNixPerObjectSectionsToRegister &MOPOSR) {
+    return SPSELFNixPerObjectSectionsToRegister::AsArgList::deserialize(
+        IB, MOPOSR.EHFrameSection, MOPOSR.ThreadDataSection);
+  }
+};
+
+using SPSNamedExecutorAddrRangeSequenceMap =
+    SPSSequence<SPSTuple<SPSString, SPSExecutorAddrRangeSequence>>;
+
+using SPSELFNixJITDylibInitializers =
+    SPSTuple<SPSString, SPSExecutorAddr, SPSNamedExecutorAddrRangeSequenceMap>;
+
+using SPSELFNixJITDylibInitializerSequence =
+    SPSSequence<SPSELFNixJITDylibInitializers>;
+
+/// Serialization traits for ELFNixJITDylibInitializers.
+template <>
+class SPSSerializationTraits<SPSELFNixJITDylibInitializers,
+                             elfnix::ELFNixJITDylibInitializers> {
+public:
+  static size_t size(const elfnix::ELFNixJITDylibInitializers &MOJDIs) {
+    return SPSELFNixJITDylibInitializers::AsArgList::size(
+        MOJDIs.Name, MOJDIs.DSOHandleAddress, MOJDIs.InitSections);
+  }
+
+  static bool serialize(SPSOutputBuffer &OB,
+                        const elfnix::ELFNixJITDylibInitializers &MOJDIs) {
+    return SPSELFNixJITDylibInitializers::AsArgList::serialize(
+        OB, MOJDIs.Name, MOJDIs.DSOHandleAddress, MOJDIs.InitSections);
+  }
+
+  static bool deserialize(SPSInputBuffer &IB,
+                          elfnix::ELFNixJITDylibInitializers &MOJDIs) {
+    return SPSELFNixJITDylibInitializers::AsArgList::deserialize(
+        IB, MOJDIs.Name, MOJDIs.DSOHandleAddress, MOJDIs.InitSections);
+  }
+};
+
+} // end namespace __orc_rt
+
+#endif // ORC_RT_ELFNIX_PLATFORM_H
diff --git a/compiler-rt/lib/orc/elfnix_tls.x86-64.S b/compiler-rt/lib/orc/elfnix_tls.x86-64.S
new file mode 100644
index 000000000000..b3e0bef00867
--- /dev/null
+++ b/compiler-rt/lib/orc/elfnix_tls.x86-64.S
@@ -0,0 +1,64 @@
+
+//===-- orc_rt_elfnix_tls_x86-64.s -------------------------------*- ASM -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of the ORC runtime support library.
+//
+//===----------------------------------------------------------------------===//
+
+// The content of this file is x86_64-only
+#if defined(__x86_64__)
+
+#define REGISTER_SAVE_SPACE_SIZE        512
+
+        .text
+
+	// returns address of TLV in %rax, all other registers preserved
+	.globl ___orc_rt_elfnix_tls_get_addr
+___orc_rt_elfnix_tls_get_addr:
+        pushq           %rbp
+        movq            %rsp,        %rbp
+        subq            $REGISTER_SAVE_SPACE_SIZE, %rsp
+        movq            %rcx,    -16(%rbp)
+        movq            %rdx,    -24(%rbp)
+        movq            %rsi,    -32(%rbp)
+        movq            %rdi,    -40(%rbp)
+        movq            %r8,     -48(%rbp)
+        movq            %r9,     -56(%rbp)
+        movq            %r10,    -64(%rbp)
+        movq            %r11,    -72(%rbp)
+	movdqa          %xmm0,  -128(%rbp)
+	movdqa          %xmm1,  -144(%rbp)
+	movdqa          %xmm2,  -160(%rbp)
+	movdqa          %xmm3,  -176(%rbp)
+	movdqa          %xmm4,  -192(%rbp)
+	movdqa          %xmm5,  -208(%rbp)
+	movdqa          %xmm6,  -224(%rbp)
+	movdqa          %xmm7,  -240(%rbp)
+        call           __orc_rt_elfnix_tls_get_addr_impl
+        movq            -16(%rbp),      %rcx
+        movq            -24(%rbp),      %rdx
+        movq            -32(%rbp),      %rsi
+        movq            -40(%rbp),      %rdi
+        movq            -48(%rbp),      %r8
+        movq            -56(%rbp),      %r9
+        movq            -64(%rbp),      %r10
+        movq            -72(%rbp),      %r11
+  movdqa          -128(%rbp),     %xmm0
+	movdqa          -144(%rbp),     %xmm1
+	movdqa          -160(%rbp),     %xmm2
+	movdqa          -176(%rbp),     %xmm3
+	movdqa          -192(%rbp),     %xmm4
+	movdqa          -208(%rbp),     %xmm5
+	movdqa          -224(%rbp),     %xmm6
+	movdqa          -240(%rbp),     %xmm7
+        addq            $REGISTER_SAVE_SPACE_SIZE, %rsp
+        popq            %rbp
+        ret
+
+#endif // defined(__x86_64__)
diff --git a/compiler-rt/lib/orc/executor_address.h b/compiler-rt/lib/orc/executor_address.h
index cfe985bdb60f..79ad9b7f1409 100644
--- a/compiler-rt/lib/orc/executor_address.h
+++ b/compiler-rt/lib/orc/executor_address.h
@@ -37,19 +37,19 @@ private:
 };
 
 /// Represents an address in the executor process.
-class ExecutorAddress {
+class ExecutorAddr {
 public:
-  ExecutorAddress() = default;
-  explicit ExecutorAddress(uint64_t Addr) : Addr(Addr) {}
+  ExecutorAddr() = default;
+  explicit ExecutorAddr(uint64_t Addr) : Addr(Addr) {}
 
-  /// Create an ExecutorAddress from the given pointer.
+  /// Create an ExecutorAddr from the given pointer.
   /// Warning: This should only be used when JITing in-process.
-  template <typename T> static ExecutorAddress fromPtr(T *Value) {
-    return ExecutorAddress(
+  template <typename T> static ExecutorAddr fromPtr(T *Value) {
+    return ExecutorAddr(
         static_cast<uint64_t>(reinterpret_cast<uintptr_t>(Value)));
   }
 
-  /// Cast this ExecutorAddress to a pointer of the given type.
+  /// Cast this ExecutorAddr to a pointer of the given type.
   /// Warning: This should only be esude when JITing in-process.
   template <typename T> T toPtr() const {
     static_assert(std::is_pointer<T>::value, "T must be a pointer type");
@@ -65,53 +65,47 @@ public:
 
   explicit operator bool() const { return Addr != 0; }
 
-  friend bool operator==(const ExecutorAddress &LHS,
-                         const ExecutorAddress &RHS) {
+  friend bool operator==(const ExecutorAddr &LHS, const ExecutorAddr &RHS) {
     return LHS.Addr == RHS.Addr;
   }
 
-  friend bool operator!=(const ExecutorAddress &LHS,
-                         const ExecutorAddress &RHS) {
+  friend bool operator!=(const ExecutorAddr &LHS, const ExecutorAddr &RHS) {
     return LHS.Addr != RHS.Addr;
   }
 
-  friend bool operator<(const ExecutorAddress &LHS,
-                        const ExecutorAddress &RHS) {
+  friend bool operator<(const ExecutorAddr &LHS, const ExecutorAddr &RHS) {
     return LHS.Addr < RHS.Addr;
   }
 
-  friend bool operator<=(const ExecutorAddress &LHS,
-                         const ExecutorAddress &RHS) {
+  friend bool operator<=(const ExecutorAddr &LHS, const ExecutorAddr &RHS) {
     return LHS.Addr <= RHS.Addr;
   }
 
-  friend bool operator>(const ExecutorAddress &LHS,
-                        const ExecutorAddress &RHS) {
+  friend bool operator>(const ExecutorAddr &LHS, const ExecutorAddr &RHS) {
     return LHS.Addr > RHS.Addr;
   }
 
-  friend bool operator>=(const ExecutorAddress &LHS,
-                         const ExecutorAddress &RHS) {
+  friend bool operator>=(const ExecutorAddr &LHS, const ExecutorAddr &RHS) {
     return LHS.Addr >= RHS.Addr;
   }
 
-  ExecutorAddress &operator++() {
+  ExecutorAddr &operator++() {
     ++Addr;
     return *this;
   }
-  ExecutorAddress &operator--() {
+  ExecutorAddr &operator--() {
     --Addr;
     return *this;
   }
-  ExecutorAddress operator++(int) { return ExecutorAddress(Addr++); }
-  ExecutorAddress operator--(int) { return ExecutorAddress(Addr++); }
+  ExecutorAddr operator++(int) { return ExecutorAddr(Addr++); }
+  ExecutorAddr operator--(int) { return ExecutorAddr(Addr++); }
 
-  ExecutorAddress &operator+=(const ExecutorAddrDiff Delta) {
+  ExecutorAddr &operator+=(const ExecutorAddrDiff Delta) {
     Addr += Delta.getValue();
     return *this;
   }
 
-  ExecutorAddress &operator-=(const ExecutorAddrDiff Delta) {
+  ExecutorAddr &operator-=(const ExecutorAddrDiff Delta) {
     Addr -= Delta.getValue();
     return *this;
   }
@@ -121,87 +115,100 @@ private:
 };
 
 /// Subtracting two addresses yields an offset.
-inline ExecutorAddrDiff operator-(const ExecutorAddress &LHS,
-                                  const ExecutorAddress &RHS) {
+inline ExecutorAddrDiff operator-(const ExecutorAddr &LHS,
+                                  const ExecutorAddr &RHS) {
   return ExecutorAddrDiff(LHS.getValue() - RHS.getValue());
 }
 
 /// Adding an offset and an address yields an address.
-inline ExecutorAddress operator+(const ExecutorAddress &LHS,
-                                 const ExecutorAddrDiff &RHS) {
-  return ExecutorAddress(LHS.getValue() + RHS.getValue());
+inline ExecutorAddr operator+(const ExecutorAddr &LHS,
+                              const ExecutorAddrDiff &RHS) {
+  return ExecutorAddr(LHS.getValue() + RHS.getValue());
 }
 
 /// Adding an address and an offset yields an address.
-inline ExecutorAddress operator+(const ExecutorAddrDiff &LHS,
-                                 const ExecutorAddress &RHS) {
-  return ExecutorAddress(LHS.getValue() + RHS.getValue());
+inline ExecutorAddr operator+(const ExecutorAddrDiff &LHS,
+                              const ExecutorAddr &RHS) {
+  return ExecutorAddr(LHS.getValue() + RHS.getValue());
 }
 
 /// Represents an address range in the exceutor process.
-struct ExecutorAddressRange {
-  ExecutorAddressRange() = default;
-  ExecutorAddressRange(ExecutorAddress StartAddress, ExecutorAddress EndAddress)
-      : StartAddress(StartAddress), EndAddress(EndAddress) {}
+struct ExecutorAddrRange {
+  ExecutorAddrRange() = default;
+  ExecutorAddrRange(ExecutorAddr Start, ExecutorAddr End)
+      : Start(Start), End(End) {}
+  ExecutorAddrRange(ExecutorAddr Start, ExecutorAddrDiff Size)
+      : Start(Start), End(Start + Size) {}
 
-  bool empty() const { return StartAddress == EndAddress; }
-  ExecutorAddrDiff size() const { return EndAddress - StartAddress; }
+  bool empty() const { return Start == End; }
+  ExecutorAddrDiff size() const { return End - Start; }
+
+  friend bool operator==(const ExecutorAddrRange &LHS,
+                         const ExecutorAddrRange &RHS) {
+    return LHS.Start == RHS.Start && LHS.End == RHS.End;
+  }
+  friend bool operator!=(const ExecutorAddrRange &LHS,
+                         const ExecutorAddrRange &RHS) {
+    return !(LHS == RHS);
+  }
+  bool contains(ExecutorAddr Addr) const { return Start <= Addr && Addr < End; }
+  bool overlaps(const ExecutorAddrRange &Other) {
+    return !(Other.End <= Start || End <= Other.Start);
+  }
 
   template <typename T> span<T> toSpan() const {
     assert(size().getValue() % sizeof(T) == 0 &&
            "AddressRange is not a multiple of sizeof(T)");
-    return span<T>(StartAddress.toPtr<T *>(), size().getValue() / sizeof(T));
+    return span<T>(Start.toPtr<T *>(), size().getValue() / sizeof(T));
   }
 
-  ExecutorAddress StartAddress;
-  ExecutorAddress EndAddress;
+  ExecutorAddr Start;
+  ExecutorAddr End;
 };
 
-/// SPS serializatior for ExecutorAddress.
-template <> class SPSSerializationTraits<SPSExecutorAddress, ExecutorAddress> {
+/// SPS serializatior for ExecutorAddr.
+template <> class SPSSerializationTraits<SPSExecutorAddr, ExecutorAddr> {
 public:
-  static size_t size(const ExecutorAddress &EA) {
+  static size_t size(const ExecutorAddr &EA) {
     return SPSArgList<uint64_t>::size(EA.getValue());
   }
 
-  static bool serialize(SPSOutputBuffer &BOB, const ExecutorAddress &EA) {
+  static bool serialize(SPSOutputBuffer &BOB, const ExecutorAddr &EA) {
     return SPSArgList<uint64_t>::serialize(BOB, EA.getValue());
   }
 
-  static bool deserialize(SPSInputBuffer &BIB, ExecutorAddress &EA) {
+  static bool deserialize(SPSInputBuffer &BIB, ExecutorAddr &EA) {
     uint64_t Tmp;
     if (!SPSArgList<uint64_t>::deserialize(BIB, Tmp))
       return false;
-    EA = ExecutorAddress(Tmp);
+    EA = ExecutorAddr(Tmp);
     return true;
   }
 };
 
-using SPSExecutorAddressRange =
-    SPSTuple<SPSExecutorAddress, SPSExecutorAddress>;
+using SPSExecutorAddrRange = SPSTuple<SPSExecutorAddr, SPSExecutorAddr>;
 
 /// Serialization traits for address ranges.
 template <>
-class SPSSerializationTraits<SPSExecutorAddressRange, ExecutorAddressRange> {
+class SPSSerializationTraits<SPSExecutorAddrRange, ExecutorAddrRange> {
 public:
-  static size_t size(const ExecutorAddressRange &Value) {
-    return SPSArgList<SPSExecutorAddress, SPSExecutorAddress>::size(
-        Value.StartAddress, Value.EndAddress);
+  static size_t size(const ExecutorAddrRange &Value) {
+    return SPSArgList<SPSExecutorAddr, SPSExecutorAddr>::size(Value.Start,
+                                                              Value.End);
   }
 
-  static bool serialize(SPSOutputBuffer &BOB,
-                        const ExecutorAddressRange &Value) {
-    return SPSArgList<SPSExecutorAddress, SPSExecutorAddress>::serialize(
-        BOB, Value.StartAddress, Value.EndAddress);
+  static bool serialize(SPSOutputBuffer &BOB, const ExecutorAddrRange &Value) {
+    return SPSArgList<SPSExecutorAddr, SPSExecutorAddr>::serialize(
+        BOB, Value.Start, Value.End);
   }
 
-  static bool deserialize(SPSInputBuffer &BIB, ExecutorAddressRange &Value) {
-    return SPSArgList<SPSExecutorAddress, SPSExecutorAddress>::deserialize(
-        BIB, Value.StartAddress, Value.EndAddress);
+  static bool deserialize(SPSInputBuffer &BIB, ExecutorAddrRange &Value) {
+    return SPSArgList<SPSExecutorAddr, SPSExecutorAddr>::deserialize(
+        BIB, Value.Start, Value.End);
   }
 };
 
-using SPSExecutorAddressRangeSequence = SPSSequence<SPSExecutorAddressRange>;
+using SPSExecutorAddrRangeSequence = SPSSequence<SPSExecutorAddrRange>;
 
 } // End namespace __orc_rt
 
diff --git a/compiler-rt/lib/orc/macho_ehframe_registration.cpp b/compiler-rt/lib/orc/macho_ehframe_registration.cpp
new file mode 100644
index 000000000000..d0ea7e70201c
--- /dev/null
+++ b/compiler-rt/lib/orc/macho_ehframe_registration.cpp
@@ -0,0 +1,68 @@
+//===- ehframe_registration.cpp -------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file contains code required to load the rest of the MachO runtime.
+//
+//===----------------------------------------------------------------------===//
+
+#include "adt.h"
+#include "c_api.h"
+#include "common.h"
+
+using namespace __orc_rt;
+
+// eh-frame registration functions.
+// We expect these to be available for all processes.
+extern "C" void __register_frame(const void *);
+extern "C" void __deregister_frame(const void *);
+
+namespace {
+
+template <typename HandleFDEFn>
+void walkEHFrameSection(span<const char> EHFrameSection,
+                        HandleFDEFn HandleFDE) {
+  const char *CurCFIRecord = EHFrameSection.data();
+  uint64_t Size = *reinterpret_cast<const uint32_t *>(CurCFIRecord);
+
+  while (CurCFIRecord != EHFrameSection.end() && Size != 0) {
+    const char *OffsetField = CurCFIRecord + (Size == 0xffffffff ? 12 : 4);
+    if (Size == 0xffffffff)
+      Size = *reinterpret_cast<const uint64_t *>(CurCFIRecord + 4) + 12;
+    else
+      Size += 4;
+    uint32_t Offset = *reinterpret_cast<const uint32_t *>(OffsetField);
+
+    if (Offset != 0)
+      HandleFDE(CurCFIRecord);
+
+    CurCFIRecord += Size;
+    Size = *reinterpret_cast<const uint32_t *>(CurCFIRecord);
+  }
+}
+
+} // end anonymous namespace
+
+ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
+__orc_rt_macho_register_ehframe_section(char *ArgData, size_t ArgSize) {
+  // NOTE: Does not use SPS to deserialize arg buffer, instead the arg buffer
+  // is taken to be the range of the eh-frame section.
+  bool HasError = false;
+  walkEHFrameSection(span<const char>(ArgData, ArgSize), __register_frame);
+  return __orc_rt_CreateCWrapperFunctionResultFromRange((char*)&HasError,
+                                                        sizeof(HasError));
+}
+
+ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
+__orc_rt_macho_deregister_ehframe_section(char *ArgData, size_t ArgSize) {
+  // NOTE: Does not use SPS to deserialize arg buffer, instead the arg buffer
+  // is taken to be the range of the eh-frame section.
+  bool HasError = false;
+  walkEHFrameSection(span<const char>(ArgData, ArgSize), __deregister_frame);
+  return __orc_rt_CreateCWrapperFunctionResultFromRange((char*)&HasError,
+                                                        sizeof(HasError));
+}
diff --git a/compiler-rt/lib/orc/macho_platform.cpp b/compiler-rt/lib/orc/macho_platform.cpp
index 2a960fb548fa..32770219e65c 100644
--- a/compiler-rt/lib/orc/macho_platform.cpp
+++ b/compiler-rt/lib/orc/macho_platform.cpp
@@ -29,11 +29,6 @@ ORC_RT_JIT_DISPATCH_TAG(__orc_rt_macho_get_initializers_tag)
 ORC_RT_JIT_DISPATCH_TAG(__orc_rt_macho_get_deinitializers_tag)
 ORC_RT_JIT_DISPATCH_TAG(__orc_rt_macho_symbol_lookup_tag)
 
-// eh-frame registration functions.
-// We expect these to be available for all processes.
-extern "C" void __register_frame(const void *);
-extern "C" void __deregister_frame(const void *);
-
 // Objective-C types.
 struct objc_class;
 struct objc_image_info;
@@ -55,6 +50,7 @@ extern "C" SEL sel_registerName(const char *) ORC_RT_WEAK_IMPORT;
 // Swift types.
 class ProtocolRecord;
 class ProtocolConformanceRecord;
+class TypeMetadataRecord;
 
 extern "C" void
 swift_registerProtocols(const ProtocolRecord *begin,
@@ -64,36 +60,18 @@ extern "C" void swift_registerProtocolConformances(
     const ProtocolConformanceRecord *begin,
     const ProtocolConformanceRecord *end) ORC_RT_WEAK_IMPORT;
 
-namespace {
-
-template <typename HandleFDEFn>
-void walkEHFrameSection(span<const char> EHFrameSection,
-                        HandleFDEFn HandleFDE) {
-  const char *CurCFIRecord = EHFrameSection.data();
-  uint64_t Size = *reinterpret_cast<const uint32_t *>(CurCFIRecord);
-
-  while (CurCFIRecord != EHFrameSection.end() && Size != 0) {
-    const char *OffsetField = CurCFIRecord + (Size == 0xffffffff ? 12 : 4);
-    if (Size == 0xffffffff)
-      Size = *reinterpret_cast<const uint64_t *>(CurCFIRecord + 4) + 12;
-    else
-      Size += 4;
-    uint32_t Offset = *reinterpret_cast<const uint32_t *>(OffsetField);
+extern "C" void swift_registerTypeMetadataRecords(
+    const TypeMetadataRecord *begin,
+    const TypeMetadataRecord *end) ORC_RT_WEAK_IMPORT;
 
-    if (Offset != 0)
-      HandleFDE(CurCFIRecord);
-
-    CurCFIRecord += Size;
-    Size = *reinterpret_cast<const uint32_t *>(CurCFIRecord);
-  }
-}
+namespace {
 
 Error validatePointerSectionExtent(const char *SectionName,
-                                   const ExecutorAddressRange &SE) {
+                                   const ExecutorAddrRange &SE) {
   if (SE.size().getValue() % sizeof(uintptr_t)) {
     std::ostringstream ErrMsg;
     ErrMsg << std::hex << "Size of " << SectionName << " 0x"
-           << SE.StartAddress.getValue() << " -- 0x" << SE.EndAddress.getValue()
+           << SE.Start.getValue() << " -- 0x" << SE.End.getValue()
            << " is not a pointer multiple";
     return make_error<StringError>(ErrMsg.str());
   }
@@ -101,7 +79,7 @@ Error validatePointerSectionExtent(const char *SectionName,
 }
 
 Error registerObjCSelectors(
-    const std::vector<ExecutorAddressRange> &ObjCSelRefsSections,
+    const std::vector<ExecutorAddrRange> &ObjCSelRefsSections,
     const MachOJITDylibInitializers &MOJDIs) {
 
   if (ORC_RT_UNLIKELY(!sel_registerName))
@@ -112,13 +90,10 @@ Error registerObjCSelectors(
     if (auto Err = validatePointerSectionExtent("__objc_selrefs", ObjCSelRefs))
       return Err;
 
-    fprintf(stderr, "Processing selrefs section at 0x%llx\n",
-            ObjCSelRefs.StartAddress.getValue());
-    for (uintptr_t SelEntry : ObjCSelRefs.toSpan<uintptr_t>()) {
+    for (uintptr_t &SelEntry : ObjCSelRefs.toSpan<uintptr_t>()) {
       const char *SelName = reinterpret_cast<const char *>(SelEntry);
-      fprintf(stderr, "Registering selector \"%s\"\n", SelName);
       auto Sel = sel_registerName(SelName);
-      *reinterpret_cast<SEL *>(SelEntry) = Sel;
+      *reinterpret_cast<SEL *>(&SelEntry) = Sel;
     }
   }
 
@@ -126,7 +101,7 @@ Error registerObjCSelectors(
 }
 
 Error registerObjCClasses(
-    const std::vector<ExecutorAddressRange> &ObjCClassListSections,
+    const std::vector<ExecutorAddrRange> &ObjCClassListSections,
     const MachOJITDylibInitializers &MOJDIs) {
 
   if (ObjCClassListSections.empty())
@@ -170,7 +145,7 @@ Error registerObjCClasses(
 }
 
 Error registerSwift5Protocols(
-    const std::vector<ExecutorAddressRange> &Swift5ProtocolSections,
+    const std::vector<ExecutorAddrRange> &Swift5ProtocolSections,
     const MachOJITDylibInitializers &MOJDIs) {
 
   if (ORC_RT_UNLIKELY(!Swift5ProtocolSections.empty() &&
@@ -179,14 +154,14 @@ Error registerSwift5Protocols(
 
   for (const auto &Swift5Protocols : Swift5ProtocolSections)
     swift_registerProtocols(
-        Swift5Protocols.StartAddress.toPtr<const ProtocolRecord *>(),
-        Swift5Protocols.EndAddress.toPtr<const ProtocolRecord *>());
+        Swift5Protocols.Start.toPtr<const ProtocolRecord *>(),
+        Swift5Protocols.End.toPtr<const ProtocolRecord *>());
 
   return Error::success();
 }
 
 Error registerSwift5ProtocolConformances(
-    const std::vector<ExecutorAddressRange> &Swift5ProtocolConformanceSections,
+    const std::vector<ExecutorAddrRange> &Swift5ProtocolConformanceSections,
     const MachOJITDylibInitializers &MOJDIs) {
 
   if (ORC_RT_UNLIKELY(!Swift5ProtocolConformanceSections.empty() &&
@@ -196,13 +171,28 @@ Error registerSwift5ProtocolConformances(
 
   for (const auto &ProtoConfSec : Swift5ProtocolConformanceSections)
     swift_registerProtocolConformances(
-        ProtoConfSec.StartAddress.toPtr<const ProtocolConformanceRecord *>(),
-        ProtoConfSec.EndAddress.toPtr<const ProtocolConformanceRecord *>());
+        ProtoConfSec.Start.toPtr<const ProtocolConformanceRecord *>(),
+        ProtoConfSec.End.toPtr<const ProtocolConformanceRecord *>());
+
+  return Error::success();
+}
+
+Error registerSwift5Types(const std::vector<ExecutorAddrRange> &Sections,
+                          const MachOJITDylibInitializers &MOJDIs) {
+
+  if (ORC_RT_UNLIKELY(!Sections.empty() && !swift_registerTypeMetadataRecords))
+    return make_error<StringError>(
+        "swift_registerTypeMetadataRecords is not available");
+
+  for (const auto &Section : Sections)
+    swift_registerTypeMetadataRecords(
+        Section.Start.toPtr<const TypeMetadataRecord *>(),
+        Section.End.toPtr<const TypeMetadataRecord *>());
 
   return Error::success();
 }
 
-Error runModInits(const std::vector<ExecutorAddressRange> &ModInitsSections,
+Error runModInits(const std::vector<ExecutorAddrRange> &ModInitsSections,
                   const MachOJITDylibInitializers &MOJDIs) {
 
   for (const auto &ModInits : ModInitsSections) {
@@ -253,8 +243,8 @@ public:
   MachOPlatformRuntimeState(MachOPlatformRuntimeState &&) = delete;
   MachOPlatformRuntimeState &operator=(MachOPlatformRuntimeState &&) = delete;
 
-  Error registerObjectSections(MachOPerObjectSectionsToRegister POSR);
-  Error deregisterObjectSections(MachOPerObjectSectionsToRegister POSR);
+  Error registerThreadDataSection(span<const char> ThreadDataSec);
+  Error deregisterThreadDataSection(span<const char> ThreadDataSec);
 
   const char *dlerror();
   void *dlopen(string_view Name, int Mode);
@@ -273,10 +263,8 @@ private:
   PerJITDylibState *getJITDylibStateByName(string_view Path);
   PerJITDylibState &getOrCreateJITDylibState(MachOJITDylibInitializers &MOJDIs);
 
-  Error registerThreadDataSection(span<const char> ThreadDataSec);
-
-  Expected<ExecutorAddress> lookupSymbolInJITDylib(void *DSOHandle,
-                                                   string_view Symbol);
+  Expected<ExecutorAddr> lookupSymbolInJITDylib(void *DSOHandle,
+                                                string_view Symbol);
 
   Expected<MachOJITDylibInitializerSequence>
   getJITDylibInitializersByName(string_view Path);
@@ -286,13 +274,14 @@ private:
   static MachOPlatformRuntimeState *MOPS;
 
   using InitSectionHandler =
-      Error (*)(const std::vector<ExecutorAddressRange> &Sections,
+      Error (*)(const std::vector<ExecutorAddrRange> &Sections,
                 const MachOJITDylibInitializers &MOJDIs);
   const std::vector<std::pair<const char *, InitSectionHandler>> InitSections =
       {{"__DATA,__objc_selrefs", registerObjCSelectors},
        {"__DATA,__objc_classlist", registerObjCClasses},
        {"__TEXT,__swift5_protos", registerSwift5Protocols},
        {"__TEXT,__swift5_proto", registerSwift5ProtocolConformances},
+       {"__TEXT,__swift5_types", registerSwift5Types},
        {"__DATA,__mod_init_func", runModInits}};
 
   // FIXME: Move to thread-state.
@@ -323,27 +312,28 @@ void MachOPlatformRuntimeState::destroy() {
   delete MOPS;
 }
 
-Error MachOPlatformRuntimeState::registerObjectSections(
-    MachOPerObjectSectionsToRegister POSR) {
-  if (POSR.EHFrameSection.StartAddress)
-    walkEHFrameSection(POSR.EHFrameSection.toSpan<const char>(),
-                       __register_frame);
-
-  if (POSR.ThreadDataSection.StartAddress) {
-    if (auto Err = registerThreadDataSection(
-            POSR.ThreadDataSection.toSpan<const char>()))
-      return Err;
+Error MachOPlatformRuntimeState::registerThreadDataSection(
+    span<const char> ThreadDataSection) {
+  std::lock_guard<std::mutex> Lock(ThreadDataSectionsMutex);
+  auto I = ThreadDataSections.upper_bound(ThreadDataSection.data());
+  if (I != ThreadDataSections.begin()) {
+    auto J = std::prev(I);
+    if (J->first + J->second > ThreadDataSection.data())
+      return make_error<StringError>("Overlapping __thread_data sections");
   }
-
+  ThreadDataSections.insert(
+      I, std::make_pair(ThreadDataSection.data(), ThreadDataSection.size()));
   return Error::success();
 }
 
-Error MachOPlatformRuntimeState::deregisterObjectSections(
-    MachOPerObjectSectionsToRegister POSR) {
-  if (POSR.EHFrameSection.StartAddress)
-    walkEHFrameSection(POSR.EHFrameSection.toSpan<const char>(),
-                       __deregister_frame);
-
+Error MachOPlatformRuntimeState::deregisterThreadDataSection(
+    span<const char> ThreadDataSection) {
+  std::lock_guard<std::mutex> Lock(ThreadDataSectionsMutex);
+  auto I = ThreadDataSections.find(ThreadDataSection.data());
+  if (I == ThreadDataSections.end())
+    return make_error<StringError>("Attempt to deregister unknown thread data "
+                                   "section");
+  ThreadDataSections.erase(I);
   return Error::success();
 }
 
@@ -465,28 +455,15 @@ MachOPlatformRuntimeState::getOrCreateJITDylibState(
   return JDS;
 }
 
-Error MachOPlatformRuntimeState::registerThreadDataSection(
-    span<const char> ThreadDataSection) {
-  std::lock_guard<std::mutex> Lock(ThreadDataSectionsMutex);
-  auto I = ThreadDataSections.upper_bound(ThreadDataSection.data());
-  if (I != ThreadDataSections.begin()) {
-    auto J = std::prev(I);
-    if (J->first + J->second > ThreadDataSection.data())
-      return make_error<StringError>("Overlapping __thread_data sections");
-  }
-  ThreadDataSections.insert(
-      I, std::make_pair(ThreadDataSection.data(), ThreadDataSection.size()));
-  return Error::success();
-}
-
-Expected<ExecutorAddress>
+Expected<ExecutorAddr>
 MachOPlatformRuntimeState::lookupSymbolInJITDylib(void *DSOHandle,
                                                   string_view Sym) {
-  Expected<ExecutorAddress> Result((ExecutorAddress()));
-  if (auto Err = WrapperFunction<SPSExpected<SPSExecutorAddress>(
-          SPSExecutorAddress,
-          SPSString)>::call(&__orc_rt_macho_symbol_lookup_tag, Result,
-                            ExecutorAddress::fromPtr(DSOHandle), Sym))
+  Expected<ExecutorAddr> Result((ExecutorAddr()));
+  if (auto Err = WrapperFunction<SPSExpected<SPSExecutorAddr>(
+          SPSExecutorAddr, SPSString)>::call(&__orc_rt_macho_symbol_lookup_tag,
+                                             Result,
+                                             ExecutorAddr::fromPtr(DSOHandle),
+                                             Sym))
     return std::move(Err);
   return Result;
 }
@@ -589,6 +566,13 @@ void destroyMachOTLVMgr(void *MachOTLVMgr) {
   delete static_cast<MachOPlatformRuntimeTLVManager *>(MachOTLVMgr);
 }
 
+Error runWrapperFunctionCalls(std::vector<WrapperFunctionCall> WFCs) {
+  for (auto &WFC : WFCs)
+    if (auto Err = WFC.runWithSPSRet())
+      return Err;
+  return Error::success();
+}
+
 } // end anonymous namespace
 
 //------------------------------------------------------------------------------
@@ -607,30 +591,41 @@ __orc_rt_macho_platform_shutdown(char *ArgData, size_t ArgSize) {
   return WrapperFunctionResult().release();
 }
 
-/// Wrapper function for registering metadata on a per-object basis.
 ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
-__orc_rt_macho_register_object_sections(char *ArgData, size_t ArgSize) {
-  return WrapperFunction<SPSError(SPSMachOPerObjectSectionsToRegister)>::handle(
-             ArgData, ArgSize,
-             [](MachOPerObjectSectionsToRegister &POSR) {
-               return MachOPlatformRuntimeState::get().registerObjectSections(
-                   std::move(POSR));
+__orc_rt_macho_register_thread_data_section(char *ArgData, size_t ArgSize) {
+  // NOTE: Does not use SPS to deserialize arg buffer, instead the arg buffer
+  // is taken to be the range of the thread data section.
+  return WrapperFunction<SPSError()>::handle(
+             nullptr, 0,
+             [&]() {
+               return MachOPlatformRuntimeState::get()
+                   .registerThreadDataSection(
+                       span<const char>(ArgData, ArgSize));
              })
       .release();
 }
 
-/// Wrapper for releasing per-object metadat.
 ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
-__orc_rt_macho_deregister_object_sections(char *ArgData, size_t ArgSize) {
-  return WrapperFunction<SPSError(SPSMachOPerObjectSectionsToRegister)>::handle(
-             ArgData, ArgSize,
-             [](MachOPerObjectSectionsToRegister &POSR) {
-               return MachOPlatformRuntimeState::get().deregisterObjectSections(
-                   std::move(POSR));
+__orc_rt_macho_deregister_thread_data_section(char *ArgData, size_t ArgSize) {
+  // NOTE: Does not use SPS to deserialize arg buffer, instead the arg buffer
+  // is taken to be the range of the thread data section.
+  return WrapperFunction<SPSError()>::handle(
+             nullptr, 0,
+             [&]() {
+               return MachOPlatformRuntimeState::get()
+                   .deregisterThreadDataSection(
+                       span<const char>(ArgData, ArgSize));
              })
       .release();
 }
 
+ORC_RT_INTERFACE __orc_rt_CWrapperFunctionResult
+__orc_rt_macho_run_wrapper_function_calls(char *ArgData, size_t ArgSize) {
+  return WrapperFunction<SPSError(SPSSequence<SPSWrapperFunctionCall>)>::handle(
+             ArgData, ArgSize, runWrapperFunctionCalls)
+      .release();
+}
+
 //------------------------------------------------------------------------------
 //                            TLV support
 //------------------------------------------------------------------------------
diff --git a/compiler-rt/lib/orc/macho_platform.h b/compiler-rt/lib/orc/macho_platform.h
index 6c05e844b0cd..5b9820a0d1f9 100644
--- a/compiler-rt/lib/orc/macho_platform.h
+++ b/compiler-rt/lib/orc/macho_platform.h
@@ -31,23 +31,17 @@ ORC_RT_INTERFACE void *__orc_rt_macho_jit_dlsym(void *dso_handle,
 namespace __orc_rt {
 namespace macho {
 
-struct MachOPerObjectSectionsToRegister {
-  ExecutorAddressRange EHFrameSection;
-  ExecutorAddressRange ThreadDataSection;
-};
-
 struct MachOJITDylibInitializers {
-  using SectionList = std::vector<ExecutorAddressRange>;
+  using SectionList = std::vector<ExecutorAddrRange>;
 
   MachOJITDylibInitializers() = default;
-  MachOJITDylibInitializers(std::string Name,
-                            ExecutorAddress MachOHeaderAddress)
+  MachOJITDylibInitializers(std::string Name, ExecutorAddr MachOHeaderAddress)
       : Name(std::move(Name)),
         MachOHeaderAddress(std::move(MachOHeaderAddress)) {}
 
   std::string Name;
-  ExecutorAddress MachOHeaderAddress;
-  ExecutorAddress ObjCImageInfoAddress;
+  ExecutorAddr MachOHeaderAddress;
+  ExecutorAddr ObjCImageInfoAddress;
 
   std::unordered_map<std::string, SectionList> InitSections;
 };
@@ -68,38 +62,12 @@ enum dlopen_mode : int {
 
 } // end namespace macho
 
-using SPSMachOPerObjectSectionsToRegister =
-    SPSTuple<SPSExecutorAddressRange, SPSExecutorAddressRange>;
-
-template <>
-class SPSSerializationTraits<SPSMachOPerObjectSectionsToRegister,
-                             macho::MachOPerObjectSectionsToRegister> {
-
-public:
-  static size_t size(const macho::MachOPerObjectSectionsToRegister &MOPOSR) {
-    return SPSMachOPerObjectSectionsToRegister::AsArgList::size(
-        MOPOSR.EHFrameSection, MOPOSR.ThreadDataSection);
-  }
-
-  static bool serialize(SPSOutputBuffer &OB,
-                        const macho::MachOPerObjectSectionsToRegister &MOPOSR) {
-    return SPSMachOPerObjectSectionsToRegister::AsArgList::serialize(
-        OB, MOPOSR.EHFrameSection, MOPOSR.ThreadDataSection);
-  }
-
-  static bool deserialize(SPSInputBuffer &IB,
-                          macho::MachOPerObjectSectionsToRegister &MOPOSR) {
-    return SPSMachOPerObjectSectionsToRegister::AsArgList::deserialize(
-        IB, MOPOSR.EHFrameSection, MOPOSR.ThreadDataSection);
-  }
-};
-
-using SPSNamedExecutorAddressRangeSequenceMap =
-    SPSSequence<SPSTuple<SPSString, SPSExecutorAddressRangeSequence>>;
+using SPSNamedExecutorAddrRangeSequenceMap =
+    SPSSequence<SPSTuple<SPSString, SPSExecutorAddrRangeSequence>>;
 
 using SPSMachOJITDylibInitializers =
-    SPSTuple<SPSString, SPSExecutorAddress, SPSExecutorAddress,
-             SPSNamedExecutorAddressRangeSequenceMap>;
+    SPSTuple<SPSString, SPSExecutorAddr, SPSExecutorAddr,
+             SPSNamedExecutorAddrRangeSequenceMap>;
 
 using SPSMachOJITDylibInitializerSequence =
     SPSSequence<SPSMachOJITDylibInitializers>;
diff --git a/compiler-rt/lib/orc/macho_tlv.arm64.S b/compiler-rt/lib/orc/macho_tlv.arm64.S
new file mode 100644
index 000000000000..f6eb9fc4da39
--- /dev/null
+++ b/compiler-rt/lib/orc/macho_tlv.arm64.S
@@ -0,0 +1,92 @@
+//===-- macho_tlv.arm64.s ---------------------------------------*- ASM -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of the ORC runtime support library.
+//
+//===----------------------------------------------------------------------===//
+
+// The content of this file is arm64-only
+#if defined(__arm64__) || defined(__aarch64__)
+
+#define REGISTER_SAVE_SPACE_SIZE     32 * 24
+
+        .text
+
+  // returns address of TLV in x0, all other registers preserved
+  .globl ___orc_rt_macho_tlv_get_addr
+___orc_rt_macho_tlv_get_addr:
+        sub  sp,  sp, #REGISTER_SAVE_SPACE_SIZE
+        stp x29, x30, [sp, #16 * 1]
+        stp x27, x28, [sp, #16 * 2]
+        stp x25, x26, [sp, #16 * 3]
+        stp x23, x24, [sp, #16 * 4]
+        stp x21, x22, [sp, #16 * 5]
+        stp x19, x20, [sp, #16 * 6]
+        stp x17, x18, [sp, #16 * 7]
+        stp x15, x16, [sp, #16 * 8]
+        stp x13, x14, [sp, #16 * 9]
+        stp x11, x12, [sp, #16 * 10]
+        stp  x9, x10, [sp, #16 * 11]
+        stp  x7,  x8, [sp, #16 * 12]
+        stp  x5,  x6, [sp, #16 * 13]
+        stp  x3,  x4, [sp, #16 * 14]
+        stp  x1,  x2, [sp, #16 * 15]
+        stp q30, q31, [sp, #32 * 8]
+        stp q28, q29, [sp, #32 * 9]
+        stp q26, q27, [sp, #32 * 10]
+        stp q24, q25, [sp, #32 * 11]
+        stp q22, q23, [sp, #32 * 12]
+        stp q20, q21, [sp, #32 * 13]
+        stp q18, q19, [sp, #32 * 14]
+        stp q16, q17, [sp, #32 * 15]
+        stp q14, q15, [sp, #32 * 16]
+        stp q12, q13, [sp, #32 * 17]
+        stp q10, q11, [sp, #32 * 18]
+        stp  q8,  q9, [sp, #32 * 19]
+        stp  q6,  q7, [sp, #32 * 20]
+        stp  q4,  q5, [sp, #32 * 21]
+        stp  q2,  q3, [sp, #32 * 22]
+        stp  q0,  q1, [sp, #32 * 23]
+
+        bl ___orc_rt_macho_tlv_get_addr_impl
+
+        ldp  q0,  q1, [sp, #32 * 23]
+        ldp  q2,  q3, [sp, #32 * 22]
+        ldp  q4,  q5, [sp, #32 * 21]
+        ldp  q6,  q7, [sp, #32 * 20]
+        ldp  q8,  q9, [sp, #32 * 19]
+        ldp q10, q11, [sp, #32 * 18]
+        ldp q12, q13, [sp, #32 * 17]
+        ldp q14, q15, [sp, #32 * 16]
+        ldp q16, q17, [sp, #32 * 15]
+        ldp q18, q19, [sp, #32 * 14]
+        ldp q20, q21, [sp, #32 * 13]
+        ldp q22, q23, [sp, #32 * 12]
+        ldp q24, q25, [sp, #32 * 11]
+        ldp q26, q27, [sp, #32 * 10]
+        ldp q28, q29, [sp, #32 * 9]
+        ldp q30, q31, [sp, #32 * 8]
+        ldp  x1,  x2, [sp, #16 * 15]
+        ldp  x3,  x4, [sp, #16 * 14]
+        ldp  x5,  x6, [sp, #16 * 13]
+        ldp  x7,  x8, [sp, #16 * 12]
+        ldp  x9, x10, [sp, #16 * 11]
+        ldp x11, x12, [sp, #16 * 10]
+        ldp x13, x14, [sp, #16 * 9]
+        ldp x15, x16, [sp, #16 * 8]
+        ldp x17, x18, [sp, #16 * 7]
+        ldp x19, x20, [sp, #16 * 6]
+        ldp x21, x22, [sp, #16 * 5]
+        ldp x23, x24, [sp, #16 * 4]
+        ldp x25, x26, [sp, #16 * 3]
+        ldp x27, x28, [sp, #16 * 2]
+        ldp x29, x30, [sp, #16 * 1]
+        add  sp,  sp, #REGISTER_SAVE_SPACE_SIZE
+        ret
+
+#endif // defined(__arm64__) || defined(__aarch64__)
diff --git a/compiler-rt/lib/orc/macho_tlv.x86-64.S b/compiler-rt/lib/orc/macho_tlv.x86-64.S
index 0affe403eec2..e3daf23e3029 100644
--- a/compiler-rt/lib/orc/macho_tlv.x86-64.S
+++ b/compiler-rt/lib/orc/macho_tlv.x86-64.S
@@ -10,6 +10,9 @@
 //
 //===----------------------------------------------------------------------===//
 
+// The content of this file is x86_64-only
+#if defined(__x86_64__)
+
 #define REGISTER_SAVE_SPACE_SIZE        512
 
         .text
@@ -66,3 +69,5 @@ ___orc_rt_macho_tlv_get_addr:
         addq            $REGISTER_SAVE_SPACE_SIZE, %rsp
         popq            %rbp
         ret
+
+#endif // defined(__x86_64__)
diff --git a/compiler-rt/lib/orc/simple_packed_serialization.h b/compiler-rt/lib/orc/simple_packed_serialization.h
index b561a19d8f04..ec43130a2ef5 100644
--- a/compiler-rt/lib/orc/simple_packed_serialization.h
+++ b/compiler-rt/lib/orc/simple_packed_serialization.h
@@ -176,7 +176,7 @@ public:
 class SPSEmpty {};
 
 /// Represents an address in the executor.
-class SPSExecutorAddress {};
+class SPSExecutorAddr {};
 
 /// SPS tag type for tuples.
 ///
@@ -354,6 +354,27 @@ public:
   }
 };
 
+/// Trivial serialization / deserialization for span<char>
+template <> class SPSSerializationTraits<SPSSequence<char>, span<const char>> {
+public:
+  static size_t size(const span<const char> &S) {
+    return SPSArgList<uint64_t>::size(static_cast<uint64_t>(S.size())) +
+           S.size();
+  }
+  static bool serialize(SPSOutputBuffer &OB, const span<const char> &S) {
+    if (!SPSArgList<uint64_t>::serialize(OB, static_cast<uint64_t>(S.size())))
+      return false;
+    return OB.write(S.data(), S.size());
+  }
+  static bool deserialize(SPSInputBuffer &IB, span<const char> &S) {
+    uint64_t Size;
+    if (!SPSArgList<uint64_t>::deserialize(IB, Size))
+      return false;
+    S = span<const char>(IB.data(), Size);
+    return IB.skip(Size);
+  }
+};
+
 /// SPSTuple serialization for std::pair.
 template <typename SPSTagT1, typename SPSTagT2, typename T1, typename T2>
 class SPSSerializationTraits<SPSTuple<SPSTagT1, SPSTagT2>, std::pair<T1, T2>> {
@@ -396,10 +417,12 @@ public:
     uint64_t Size;
     if (!SPSArgList<uint64_t>::deserialize(IB, Size))
       return false;
+    if (Size > std::numeric_limits<size_t>::max())
+      return false;
     Data = IB.data();
     if (!IB.skip(Size))
       return false;
-    S = {Data, Size};
+    S = {Data, static_cast<size_t>(Size)};
     return true;
   }
 };
diff --git a/compiler-rt/lib/orc/wrapper_function_utils.h b/compiler-rt/lib/orc/wrapper_function_utils.h
index 49faa03e5eb8..23385e1bd794 100644
--- a/compiler-rt/lib/orc/wrapper_function_utils.h
+++ b/compiler-rt/lib/orc/wrapper_function_utils.h
@@ -16,6 +16,7 @@
 #include "c_api.h"
 #include "common.h"
 #include "error.h"
+#include "executor_address.h"
 #include "simple_packed_serialization.h"
 #include <type_traits>
 
@@ -61,7 +62,7 @@ public:
   }
 
   /// Get a pointer to the data contained in this instance.
-  const char *data() const { return __orc_rt_CWrapperFunctionResultData(&R); }
+  char *data() { return __orc_rt_CWrapperFunctionResultData(&R); }
 
   /// Returns the size of the data contained in this instance.
   size_t size() const { return __orc_rt_CWrapperFunctionResultSize(&R); }
@@ -72,10 +73,10 @@ public:
 
   /// Create a WrapperFunctionResult with the given size and return a pointer
   /// to the underlying memory.
-  static char *allocate(WrapperFunctionResult &R, size_t Size) {
-    __orc_rt_DisposeCWrapperFunctionResult(&R.R);
-    __orc_rt_CWrapperFunctionResultInit(&R.R);
-    return __orc_rt_CWrapperFunctionResultAllocate(&R.R, Size);
+  static WrapperFunctionResult allocate(size_t Size) {
+    WrapperFunctionResult R;
+    R.R = __orc_rt_CWrapperFunctionResultAllocate(Size);
+    return R;
   }
 
   /// Copy from the given char range.
@@ -103,6 +104,16 @@ public:
     return createOutOfBandError(Msg.c_str());
   }
 
+  template <typename SPSArgListT, typename... ArgTs>
+  static WrapperFunctionResult fromSPSArgs(const ArgTs &...Args) {
+    auto Result = allocate(SPSArgListT::size(Args...));
+    SPSOutputBuffer OB(Result.data(), Result.size());
+    if (!SPSArgListT::serialize(OB, Args...))
+      return createOutOfBandError(
+          "Error serializing arguments to blob in call");
+    return Result;
+  }
+
   /// If this value is an out-of-band error then this returns the error message,
   /// otherwise returns nullptr.
   const char *getOutOfBandError() const {
@@ -115,19 +126,6 @@ private:
 
 namespace detail {
 
-template <typename SPSArgListT, typename... ArgTs>
-Expected<WrapperFunctionResult>
-serializeViaSPSToWrapperFunctionResult(const ArgTs &...Args) {
-  WrapperFunctionResult Result;
-  char *DataPtr =
-      WrapperFunctionResult::allocate(Result, SPSArgListT::size(Args...));
-  SPSOutputBuffer OB(DataPtr, Result.size());
-  if (!SPSArgListT::serialize(OB, Args...))
-    return make_error<StringError>(
-        "Error serializing arguments to blob in call");
-  return std::move(Result);
-}
-
 template <typename RetT> class WrapperFunctionHandlerCaller {
 public:
   template <typename HandlerT, typename ArgTupleT, std::size_t... I>
@@ -173,12 +171,8 @@ public:
     auto HandlerResult = WrapperFunctionHandlerCaller<RetT>::call(
         std::forward<HandlerT>(H), Args, ArgIndices{});
 
-    if (auto Result = ResultSerializer<decltype(HandlerResult)>::serialize(
-            std::move(HandlerResult)))
-      return std::move(*Result);
-    else
-      return WrapperFunctionResult::createOutOfBandError(
-          toString(Result.takeError()));
+    return ResultSerializer<decltype(HandlerResult)>::serialize(
+        std::move(HandlerResult));
   }
 
 private:
@@ -188,13 +182,12 @@ private:
     SPSInputBuffer IB(ArgData, ArgSize);
     return SPSArgList<SPSTagTs...>::deserialize(IB, std::get<I>(Args)...);
   }
-
 };
 
-// Map function references to function types.
+// Map function pointers to function types.
 template <typename RetT, typename... ArgTs,
           template <typename> class ResultSerializer, typename... SPSTagTs>
-class WrapperFunctionHandlerHelper<RetT (&)(ArgTs...), ResultSerializer,
+class WrapperFunctionHandlerHelper<RetT (*)(ArgTs...), ResultSerializer,
                                    SPSTagTs...>
     : public WrapperFunctionHandlerHelper<RetT(ArgTs...), ResultSerializer,
                                           SPSTagTs...> {};
@@ -217,16 +210,15 @@ class WrapperFunctionHandlerHelper<RetT (ClassT::*)(ArgTs...) const,
 
 template <typename SPSRetTagT, typename RetT> class ResultSerializer {
 public:
-  static Expected<WrapperFunctionResult> serialize(RetT Result) {
-    return serializeViaSPSToWrapperFunctionResult<SPSArgList<SPSRetTagT>>(
-        Result);
+  static WrapperFunctionResult serialize(RetT Result) {
+    return WrapperFunctionResult::fromSPSArgs<SPSArgList<SPSRetTagT>>(Result);
   }
 };
 
 template <typename SPSRetTagT> class ResultSerializer<SPSRetTagT, Error> {
 public:
-  static Expected<WrapperFunctionResult> serialize(Error Err) {
-    return serializeViaSPSToWrapperFunctionResult<SPSArgList<SPSRetTagT>>(
+  static WrapperFunctionResult serialize(Error Err) {
+    return WrapperFunctionResult::fromSPSArgs<SPSArgList<SPSRetTagT>>(
         toSPSSerializable(std::move(Err)));
   }
 };
@@ -234,8 +226,8 @@ public:
 template <typename SPSRetTagT, typename T>
 class ResultSerializer<SPSRetTagT, Expected<T>> {
 public:
-  static Expected<WrapperFunctionResult> serialize(Expected<T> E) {
-    return serializeViaSPSToWrapperFunctionResult<SPSArgList<SPSRetTagT>>(
+  static WrapperFunctionResult serialize(Expected<T> E) {
+    return WrapperFunctionResult::fromSPSArgs<SPSArgList<SPSRetTagT>>(
         toSPSSerializable(std::move(E)));
   }
 };
@@ -310,14 +302,12 @@ public:
       return make_error<StringError>("__orc_rt_jit_dispatch not set");
 
     auto ArgBuffer =
-        detail::serializeViaSPSToWrapperFunctionResult<SPSArgList<SPSTagTs...>>(
-            Args...);
-    if (!ArgBuffer)
-      return ArgBuffer.takeError();
+        WrapperFunctionResult::fromSPSArgs<SPSArgList<SPSTagTs...>>(Args...);
+    if (const char *ErrMsg = ArgBuffer.getOutOfBandError())
+      return make_error<StringError>(ErrMsg);
 
-    WrapperFunctionResult ResultBuffer =
-        __orc_rt_jit_dispatch(&__orc_rt_jit_dispatch_ctx, FnTag,
-                              ArgBuffer->data(), ArgBuffer->size());
+    WrapperFunctionResult ResultBuffer = __orc_rt_jit_dispatch(
+        &__orc_rt_jit_dispatch_ctx, FnTag, ArgBuffer.data(), ArgBuffer.size());
     if (auto ErrMsg = ResultBuffer.getOutOfBandError())
       return make_error<StringError>(ErrMsg);
 
@@ -329,8 +319,8 @@ public:
   static WrapperFunctionResult handle(const char *ArgData, size_t ArgSize,
                                       HandlerT &&Handler) {
     using WFHH =
-        detail::WrapperFunctionHandlerHelper<HandlerT, ResultSerializer,
-                                             SPSTagTs...>;
+        detail::WrapperFunctionHandlerHelper<std::remove_reference_t<HandlerT>,
+                                             ResultSerializer, SPSTagTs...>;
     return WFHH::apply(std::forward<HandlerT>(Handler), ArgData, ArgSize);
   }
 
@@ -362,6 +352,106 @@ public:
   using WrapperFunction<SPSEmpty(SPSTagTs...)>::handle;
 };
 
+/// A function object that takes an ExecutorAddr as its first argument,
+/// casts that address to a ClassT*, then calls the given method on that
+/// pointer passing in the remaining function arguments. This utility
+/// removes some of the boilerplate from writing wrappers for method calls.
+///
+///   @code{.cpp}
+///   class MyClass {
+///   public:
+///     void myMethod(uint32_t, bool) { ... }
+///   };
+///
+///   // SPS Method signature -- note MyClass object address as first argument.
+///   using SPSMyMethodWrapperSignature =
+///     SPSTuple<SPSExecutorAddr, uint32_t, bool>;
+///
+///   WrapperFunctionResult
+///   myMethodCallWrapper(const char *ArgData, size_t ArgSize) {
+///     return WrapperFunction<SPSMyMethodWrapperSignature>::handle(
+///        ArgData, ArgSize, makeMethodWrapperHandler(&MyClass::myMethod));
+///   }
+///   @endcode
+///
+template <typename RetT, typename ClassT, typename... ArgTs>
+class MethodWrapperHandler {
+public:
+  using MethodT = RetT (ClassT::*)(ArgTs...);
+  MethodWrapperHandler(MethodT M) : M(M) {}
+  RetT operator()(ExecutorAddr ObjAddr, ArgTs &...Args) {
+    return (ObjAddr.toPtr<ClassT *>()->*M)(std::forward<ArgTs>(Args)...);
+  }
+
+private:
+  MethodT M;
+};
+
+/// Create a MethodWrapperHandler object from the given method pointer.
+template <typename RetT, typename ClassT, typename... ArgTs>
+MethodWrapperHandler<RetT, ClassT, ArgTs...>
+makeMethodWrapperHandler(RetT (ClassT::*Method)(ArgTs...)) {
+  return MethodWrapperHandler<RetT, ClassT, ArgTs...>(Method);
+}
+
+/// Represents a call to a wrapper function.
+struct WrapperFunctionCall {
+  ExecutorAddr Func;
+  ExecutorAddrRange ArgData;
+
+  WrapperFunctionCall() = default;
+  WrapperFunctionCall(ExecutorAddr Func, ExecutorAddrRange ArgData)
+      : Func(Func), ArgData(ArgData) {}
+
+  /// Run and return result as WrapperFunctionResult.
+  WrapperFunctionResult run() {
+    WrapperFunctionResult WFR(
+        Func.toPtr<__orc_rt_CWrapperFunctionResult (*)(const char *, size_t)>()(
+            ArgData.Start.toPtr<const char *>(),
+            static_cast<size_t>(ArgData.size().getValue())));
+    return WFR;
+  }
+
+  /// Run call and deserialize result using SPS.
+  template <typename SPSRetT, typename RetT> Error runWithSPSRet(RetT &RetVal) {
+    auto WFR = run();
+    if (const char *ErrMsg = WFR.getOutOfBandError())
+      return make_error<StringError>(ErrMsg);
+    SPSInputBuffer IB(WFR.data(), WFR.size());
+    if (!SPSSerializationTraits<SPSRetT, RetT>::deserialize(IB, RetVal))
+      return make_error<StringError>("Could not deserialize result from "
+                                     "serialized wrapper function call");
+    return Error::success();
+  }
+
+  /// Overload for SPS functions returning void.
+  Error runWithSPSRet() {
+    SPSEmpty E;
+    return runWithSPSRet<SPSEmpty>(E);
+  }
+};
+
+class SPSWrapperFunctionCall {};
+
+template <>
+class SPSSerializationTraits<SPSWrapperFunctionCall, WrapperFunctionCall> {
+public:
+  static size_t size(const WrapperFunctionCall &WFC) {
+    return SPSArgList<SPSExecutorAddr, SPSExecutorAddrRange>::size(WFC.Func,
+                                                                   WFC.ArgData);
+  }
+
+  static bool serialize(SPSOutputBuffer &OB, const WrapperFunctionCall &WFC) {
+    return SPSArgList<SPSExecutorAddr, SPSExecutorAddrRange>::serialize(
+        OB, WFC.Func, WFC.ArgData);
+  }
+
+  static bool deserialize(SPSInputBuffer &IB, WrapperFunctionCall &WFC) {
+    return SPSArgList<SPSExecutorAddr, SPSExecutorAddrRange>::deserialize(
+        IB, WFC.Func, WFC.ArgData);
+  }
+};
+
 } // end namespace __orc_rt
 
 #endif // ORC_RT_WRAPPER_FUNCTION_UTILS_H
diff --git a/compiler-rt/lib/profile/InstrProfiling.h b/compiler-rt/lib/profile/InstrProfiling.h
index 237acb33ffa1..5b88d7178012 100644
--- a/compiler-rt/lib/profile/InstrProfiling.h
+++ b/compiler-rt/lib/profile/InstrProfiling.h
@@ -150,7 +150,7 @@ int __llvm_profile_write_file(void);
 int __llvm_orderfile_write_file(void);
 /*!
  * \brief this is a wrapper interface to \c __llvm_profile_write_file.
- * After this interface is invoked, a arleady dumped flag will be set
+ * After this interface is invoked, an already dumped flag will be set
  * so that profile won't be dumped again during program exit.
  * Invocation of interface __llvm_profile_reset_counters will clear
  * the flag. This interface is designed to be used to collect profile
@@ -194,7 +194,8 @@ int __llvm_orderfile_dump(void);
 void __llvm_profile_set_filename(const char *Name);
 
 /*!
- * \brief Set the FILE object for writing instrumentation data.
+ * \brief Set the FILE object for writing instrumentation data. Return 0 if set
+ * successfully or return 1 if failed.
  *
  * Sets the FILE object to be used for subsequent calls to
  * \a __llvm_profile_write_file(). The profile file name set by environment
@@ -213,13 +214,12 @@ void __llvm_profile_set_filename(const char *Name);
  * instrumented image/DSO). This API only modifies the file object within the
  * copy of the runtime available to the calling image.
  *
- * Warning: This is a no-op if continuous mode (\ref
- * __llvm_profile_is_continuous_mode_enabled) is on. The reason for this is
- * that in continuous mode, profile counters are mmap()'d to the profile at
- * program initialization time. Support for transferring the mmap'd profile
- * counts to a new file has not been implemented.
+ * Warning: This is a no-op if EnableMerge is 0 in continuous mode (\ref
+ * __llvm_profile_is_continuous_mode_enabled), because disable merging requires
+ * copying the old profile file to new profile file and this function is usually
+ * used when the proess doesn't have permission to open file.
  */
-void __llvm_profile_set_file_object(FILE *File, int EnableMerge);
+int __llvm_profile_set_file_object(FILE *File, int EnableMerge);
 
 /*! \brief Register to write instrumentation data to file at exit. */
 int __llvm_profile_register_write_file_atexit(void);
@@ -301,14 +301,12 @@ void __llvm_profile_set_dumped();
 COMPILER_RT_VISIBILITY extern int INSTR_PROF_PROFILE_RUNTIME_VAR;
 
 /*!
- * This variable is defined in InstrProfiling.c. Its main purpose is to
- * encode the raw profile version value and other format related information
- * such as whether the profile is from IR based instrumentation. The variable
- * is defined as weak so that compiler can emit an overriding definition
- * depending on user option.  Since we don't support mixing FE and IR based
- * data in the same raw profile data file (in other words, shared libs and
- * main program are expected to be instrumented in the same way), there is
- * no need for this variable to be hidden.
+ * This variable is defined in InstrProfilingVersionVar.c as a hidden symbol
+ * (except on Apple platforms where this symbol is checked by TAPI).  Its main
+ * purpose is to encode the raw profile version value and other format related
+ * information such as whether the profile is from IR based instrumentation. The
+ * variable is defined as weak so that compiler can emit an overriding
+ * definition depending on user option.
  */
 extern uint64_t INSTR_PROF_RAW_VERSION_VAR; /* __llvm_profile_raw_version */
 
diff --git a/compiler-rt/lib/profile/InstrProfilingBuffer.c b/compiler-rt/lib/profile/InstrProfilingBuffer.c
index 21fa7ba1ddd6..68b4f5cd6f52 100644
--- a/compiler-rt/lib/profile/InstrProfilingBuffer.c
+++ b/compiler-rt/lib/profile/InstrProfilingBuffer.c
@@ -116,7 +116,7 @@ uint64_t __llvm_profile_get_size_for_buffer_internal(
       DataSize, CountersSize, NamesSize, &PaddingBytesBeforeCounters,
       &PaddingBytesAfterCounters, &PaddingBytesAfterNames);
 
-  return sizeof(__llvm_profile_header) +
+  return sizeof(__llvm_profile_header) + __llvm_write_binary_ids(NULL) +
          (DataSize * sizeof(__llvm_profile_data)) + PaddingBytesBeforeCounters +
          (CountersSize * sizeof(uint64_t)) + PaddingBytesAfterCounters +
          NamesSize + PaddingBytesAfterNames;
diff --git a/compiler-rt/lib/profile/InstrProfilingFile.c b/compiler-rt/lib/profile/InstrProfilingFile.c
index 518447e3e422..efd9f06ac6ee 100644
--- a/compiler-rt/lib/profile/InstrProfilingFile.c
+++ b/compiler-rt/lib/profile/InstrProfilingFile.c
@@ -92,6 +92,146 @@ static lprofFilename lprofCurFilename = {0,   0, 0, {0}, NULL,
                                          {0}, 0, 0, 0,   PNS_unknown};
 
 static int ProfileMergeRequested = 0;
+static int getProfileFileSizeForMerging(FILE *ProfileFile,
+                                        uint64_t *ProfileFileSize);
+
+#if defined(__APPLE__)
+static const int ContinuousModeSupported = 1;
+static const int UseBiasVar = 0;
+static const char *FileOpenMode = "a+b";
+static void *BiasAddr = NULL;
+static void *BiasDefaultAddr = NULL;
+static int mmapForContinuousMode(uint64_t CurrentFileOffset, FILE *File) {
+  /* Get the sizes of various profile data sections. Taken from
+   * __llvm_profile_get_size_for_buffer(). */
+  const __llvm_profile_data *DataBegin = __llvm_profile_begin_data();
+  const __llvm_profile_data *DataEnd = __llvm_profile_end_data();
+  const uint64_t *CountersBegin = __llvm_profile_begin_counters();
+  const uint64_t *CountersEnd = __llvm_profile_end_counters();
+  const char *NamesBegin = __llvm_profile_begin_names();
+  const char *NamesEnd = __llvm_profile_end_names();
+  const uint64_t NamesSize = (NamesEnd - NamesBegin) * sizeof(char);
+  uint64_t DataSize = __llvm_profile_get_data_size(DataBegin, DataEnd);
+  uint64_t CountersSize = CountersEnd - CountersBegin;
+
+  /* Check that the counter and data sections in this image are
+   * page-aligned. */
+  unsigned PageSize = getpagesize();
+  if ((intptr_t)CountersBegin % PageSize != 0) {
+    PROF_ERR("Counters section not page-aligned (start = %p, pagesz = %u).\n",
+             CountersBegin, PageSize);
+    return 1;
+  }
+  if ((intptr_t)DataBegin % PageSize != 0) {
+    PROF_ERR("Data section not page-aligned (start = %p, pagesz = %u).\n",
+             DataBegin, PageSize);
+    return 1;
+  }
+  int Fileno = fileno(File);
+  /* Determine how much padding is needed before/after the counters and
+   * after the names. */
+  uint64_t PaddingBytesBeforeCounters, PaddingBytesAfterCounters,
+      PaddingBytesAfterNames;
+  __llvm_profile_get_padding_sizes_for_counters(
+      DataSize, CountersSize, NamesSize, &PaddingBytesBeforeCounters,
+      &PaddingBytesAfterCounters, &PaddingBytesAfterNames);
+
+  uint64_t PageAlignedCountersLength =
+      (CountersSize * sizeof(uint64_t)) + PaddingBytesAfterCounters;
+  uint64_t FileOffsetToCounters =
+      CurrentFileOffset + sizeof(__llvm_profile_header) +
+      (DataSize * sizeof(__llvm_profile_data)) + PaddingBytesBeforeCounters;
+  uint64_t *CounterMmap = (uint64_t *)mmap(
+      (void *)CountersBegin, PageAlignedCountersLength, PROT_READ | PROT_WRITE,
+      MAP_FIXED | MAP_SHARED, Fileno, FileOffsetToCounters);
+  if (CounterMmap != CountersBegin) {
+    PROF_ERR(
+        "Continuous counter sync mode is enabled, but mmap() failed (%s).\n"
+        "  - CountersBegin: %p\n"
+        "  - PageAlignedCountersLength: %" PRIu64 "\n"
+        "  - Fileno: %d\n"
+        "  - FileOffsetToCounters: %" PRIu64 "\n",
+        strerror(errno), CountersBegin, PageAlignedCountersLength, Fileno,
+        FileOffsetToCounters);
+    return 1;
+  }
+  return 0;
+}
+#elif defined(__ELF__) || defined(_WIN32)
+
+#define INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR                            \
+  INSTR_PROF_CONCAT(INSTR_PROF_PROFILE_COUNTER_BIAS_VAR, _default)
+intptr_t INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR = 0;
+
+/* This variable is a weak external reference which could be used to detect
+ * whether or not the compiler defined this symbol. */
+#if defined(_MSC_VER)
+COMPILER_RT_VISIBILITY extern intptr_t INSTR_PROF_PROFILE_COUNTER_BIAS_VAR;
+#if defined(_M_IX86) || defined(__i386__)
+#define WIN_SYM_PREFIX "_"
+#else
+#define WIN_SYM_PREFIX
+#endif
+#pragma comment(                                                               \
+    linker, "/alternatename:" WIN_SYM_PREFIX INSTR_PROF_QUOTE(                 \
+                INSTR_PROF_PROFILE_COUNTER_BIAS_VAR) "=" WIN_SYM_PREFIX        \
+                INSTR_PROF_QUOTE(INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR))
+#else
+COMPILER_RT_VISIBILITY extern intptr_t INSTR_PROF_PROFILE_COUNTER_BIAS_VAR
+    __attribute__((weak, alias(INSTR_PROF_QUOTE(
+                             INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR))));
+#endif
+static const int ContinuousModeSupported = 1;
+static const int UseBiasVar = 1;
+/* TODO: If there are two DSOs, the second DSO initilization will truncate the
+ * first profile file. */
+static const char *FileOpenMode = "w+b";
+/* This symbol is defined by the compiler when runtime counter relocation is
+ * used and runtime provides a weak alias so we can check if it's defined. */
+static void *BiasAddr = &INSTR_PROF_PROFILE_COUNTER_BIAS_VAR;
+static void *BiasDefaultAddr = &INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR;
+static int mmapForContinuousMode(uint64_t CurrentFileOffset, FILE *File) {
+  /* Get the sizes of various profile data sections. Taken from
+   * __llvm_profile_get_size_for_buffer(). */
+  const __llvm_profile_data *DataBegin = __llvm_profile_begin_data();
+  const __llvm_profile_data *DataEnd = __llvm_profile_end_data();
+  const uint64_t *CountersBegin = __llvm_profile_begin_counters();
+  const uint64_t *CountersEnd = __llvm_profile_end_counters();
+  uint64_t DataSize = __llvm_profile_get_data_size(DataBegin, DataEnd);
+  /* Get the file size. */
+  uint64_t FileSize = 0;
+  if (getProfileFileSizeForMerging(File, &FileSize))
+    return 1;
+
+  /* Map the profile. */
+  char *Profile = (char *)mmap(NULL, FileSize, PROT_READ | PROT_WRITE,
+                               MAP_SHARED, fileno(File), 0);
+  if (Profile == MAP_FAILED) {
+    PROF_ERR("Unable to mmap profile: %s\n", strerror(errno));
+    return 1;
+  }
+  const uint64_t CountersOffsetInBiasMode =
+      sizeof(__llvm_profile_header) + __llvm_write_binary_ids(NULL) +
+      (DataSize * sizeof(__llvm_profile_data));
+  /* Update the profile fields based on the current mapping. */
+  INSTR_PROF_PROFILE_COUNTER_BIAS_VAR =
+      (intptr_t)Profile - (uintptr_t)CountersBegin + CountersOffsetInBiasMode;
+
+  /* Return the memory allocated for counters to OS. */
+  lprofReleaseMemoryPagesToOS((uintptr_t)CountersBegin, (uintptr_t)CountersEnd);
+  return 0;
+}
+#else
+static const int ContinuousModeSupported = 0;
+static const int UseBiasVar = 0;
+static const char *FileOpenMode = "a+b";
+static void *BiasAddr = NULL;
+static void *BiasDefaultAddr = NULL;
+static int mmapForContinuousMode(uint64_t CurrentFileOffset, FILE *File) {
+  return 0;
+}
+#endif
+
 static int isProfileMergeRequested() { return ProfileMergeRequested; }
 static void setProfileMergeRequested(int EnableMerge) {
   ProfileMergeRequested = EnableMerge;
@@ -101,18 +241,6 @@ static FILE *ProfileFile = NULL;
 static FILE *getProfileFile() { return ProfileFile; }
 static void setProfileFile(FILE *File) { ProfileFile = File; }
 
-COMPILER_RT_VISIBILITY void __llvm_profile_set_file_object(FILE *File,
-                                                           int EnableMerge) {
-  if (__llvm_profile_is_continuous_mode_enabled()) {
-    PROF_WARN("__llvm_profile_set_file_object(fd=%d) not supported, because "
-              "continuous sync mode (%%c) is enabled",
-              fileno(File));
-    return;
-  }
-  setProfileFile(File);
-  setProfileMergeRequested(EnableMerge);
-}
-
 static int getCurFilenameLength();
 static const char *getCurFilename(char *FilenameBuf, int ForceUseBuf);
 static unsigned doMerging() {
@@ -426,13 +554,6 @@ static void truncateCurrentFile(void) {
   fclose(File);
 }
 
-// TODO: Move these functions into InstrProfilingPlatform* files.
-#if defined(__APPLE__)
-static void assertIsZero(int *i) {
-  if (*i)
-    PROF_WARN("Expected flag to be 0, but got: %d\n", *i);
-}
-
 /* Write a partial profile to \p Filename, which is required to be backed by
  * the open file object \p File. */
 static int writeProfileWithFileObject(const char *Filename, FILE *File) {
@@ -444,215 +565,22 @@ static int writeProfileWithFileObject(const char *Filename, FILE *File) {
   return rc;
 }
 
-/* Unlock the profile \p File and clear the unlock flag. */
-static void unlockProfile(int *ProfileRequiresUnlock, FILE *File) {
-  if (!*ProfileRequiresUnlock) {
-    PROF_WARN("%s", "Expected to require profile unlock\n");
-  }
-
-  lprofUnlockFileHandle(File);
-  *ProfileRequiresUnlock = 0;
-}
-
 static void initializeProfileForContinuousMode(void) {
   if (!__llvm_profile_is_continuous_mode_enabled())
     return;
-
-  /* Get the sizes of various profile data sections. Taken from
-   * __llvm_profile_get_size_for_buffer(). */
-  const __llvm_profile_data *DataBegin = __llvm_profile_begin_data();
-  const __llvm_profile_data *DataEnd = __llvm_profile_end_data();
-  const uint64_t *CountersBegin = __llvm_profile_begin_counters();
-  const uint64_t *CountersEnd = __llvm_profile_end_counters();
-  const char *NamesBegin = __llvm_profile_begin_names();
-  const char *NamesEnd = __llvm_profile_end_names();
-  const uint64_t NamesSize = (NamesEnd - NamesBegin) * sizeof(char);
-  uint64_t DataSize = __llvm_profile_get_data_size(DataBegin, DataEnd);
-  uint64_t CountersSize = CountersEnd - CountersBegin;
-
-  /* Check that the counter and data sections in this image are page-aligned. */
-  unsigned PageSize = getpagesize();
-  if ((intptr_t)CountersBegin % PageSize != 0) {
-    PROF_ERR("Counters section not page-aligned (start = %p, pagesz = %u).\n",
-             CountersBegin, PageSize);
+  if (!ContinuousModeSupported) {
+    PROF_ERR("%s\n", "continuous mode is unsupported on this platform");
     return;
   }
-  if ((intptr_t)DataBegin % PageSize != 0) {
-    PROF_ERR("Data section not page-aligned (start = %p, pagesz = %u).\n",
-             DataBegin, PageSize);
-    return;
-  }
-
-  int Length = getCurFilenameLength();
-  char *FilenameBuf = (char *)COMPILER_RT_ALLOCA(Length + 1);
-  const char *Filename = getCurFilename(FilenameBuf, 0);
-  if (!Filename)
-    return;
-
-  FILE *File = NULL;
-  off_t CurrentFileOffset = 0;
-  off_t OffsetModPage = 0;
-
-  /* Whether an exclusive lock on the profile must be dropped after init.
-   * Use a cleanup to warn if the unlock does not occur. */
-  COMPILER_RT_CLEANUP(assertIsZero) int ProfileRequiresUnlock = 0;
-
-  if (!doMerging()) {
-    /* We are not merging profiles, so open the raw profile in append mode. */
-    File = fopen(Filename, "a+b");
-    if (!File)
-      return;
-
-    /* Check that the offset within the file is page-aligned. */
-    CurrentFileOffset = ftello(File);
-    OffsetModPage = CurrentFileOffset % PageSize;
-    if (OffsetModPage != 0) {
-      PROF_ERR("Continuous counter sync mode is enabled, but raw profile is not"
-               "page-aligned. CurrentFileOffset = %" PRIu64 ", pagesz = %u.\n",
-               (uint64_t)CurrentFileOffset, PageSize);
-      return;
-    }
-
-    /* Grow the profile so that mmap() can succeed.  Leak the file handle, as
-     * the file should stay open. */
-    if (writeProfileWithFileObject(Filename, File) != 0)
-      return;
-  } else {
-    /* We are merging profiles. Map the counter section as shared memory into
-     * the profile, i.e. into each participating process. An increment in one
-     * process should be visible to every other process with the same counter
-     * section mapped. */
-    File = lprofOpenFileEx(Filename);
-    if (!File)
-      return;
-
-    ProfileRequiresUnlock = 1;
-
-    uint64_t ProfileFileSize;
-    if (getProfileFileSizeForMerging(File, &ProfileFileSize) == -1)
-      return unlockProfile(&ProfileRequiresUnlock, File);
-
-    if (ProfileFileSize == 0) {
-      /* Grow the profile so that mmap() can succeed.  Leak the file handle, as
-       * the file should stay open. */
-      if (writeProfileWithFileObject(Filename, File) != 0)
-        return unlockProfile(&ProfileRequiresUnlock, File);
-    } else {
-      /* The merged profile has a non-zero length. Check that it is compatible
-       * with the data in this process. */
-      char *ProfileBuffer;
-      if (mmapProfileForMerging(File, ProfileFileSize, &ProfileBuffer) == -1 ||
-          munmap(ProfileBuffer, ProfileFileSize) == -1)
-        return unlockProfile(&ProfileRequiresUnlock, File);
-    }
-  }
-
-  /* mmap() the profile counters so long as there is at least one counter.
-   * If there aren't any counters, mmap() would fail with EINVAL. */
-  if (CountersSize > 0) {
-    int Fileno = fileno(File);
-
-    /* Determine how much padding is needed before/after the counters and after
-     * the names. */
-    uint64_t PaddingBytesBeforeCounters, PaddingBytesAfterCounters,
-        PaddingBytesAfterNames;
-    __llvm_profile_get_padding_sizes_for_counters(
-        DataSize, CountersSize, NamesSize, &PaddingBytesBeforeCounters,
-        &PaddingBytesAfterCounters, &PaddingBytesAfterNames);
-
-    uint64_t PageAlignedCountersLength =
-        (CountersSize * sizeof(uint64_t)) + PaddingBytesAfterCounters;
-    uint64_t FileOffsetToCounters =
-        CurrentFileOffset + sizeof(__llvm_profile_header) +
-        (DataSize * sizeof(__llvm_profile_data)) + PaddingBytesBeforeCounters;
-
-    uint64_t *CounterMmap = (uint64_t *)mmap(
-        (void *)CountersBegin, PageAlignedCountersLength, PROT_READ | PROT_WRITE,
-        MAP_FIXED | MAP_SHARED, Fileno, FileOffsetToCounters);
-    if (CounterMmap != CountersBegin) {
-      PROF_ERR(
-          "Continuous counter sync mode is enabled, but mmap() failed (%s).\n"
-          "  - CountersBegin: %p\n"
-          "  - PageAlignedCountersLength: %" PRIu64 "\n"
-          "  - Fileno: %d\n"
-          "  - FileOffsetToCounters: %" PRIu64 "\n",
-          strerror(errno), CountersBegin, PageAlignedCountersLength, Fileno,
-          FileOffsetToCounters);
-    }
-  }
-
-  if (ProfileRequiresUnlock)
-    unlockProfile(&ProfileRequiresUnlock, File);
-}
-#elif defined(__ELF__) || defined(_WIN32)
-
-#define INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR                            \
-  INSTR_PROF_CONCAT(INSTR_PROF_PROFILE_COUNTER_BIAS_VAR, _default)
-intptr_t INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR = 0;
-
-/* This variable is a weak external reference which could be used to detect
- * whether or not the compiler defined this symbol. */
-#if defined(_WIN32)
-COMPILER_RT_VISIBILITY extern intptr_t INSTR_PROF_PROFILE_COUNTER_BIAS_VAR;
-#pragma comment(linker, "/alternatename:"                                      \
-      INSTR_PROF_QUOTE(INSTR_PROF_PROFILE_COUNTER_BIAS_VAR) "="                \
-      INSTR_PROF_QUOTE(INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR))
-#else
-COMPILER_RT_VISIBILITY extern intptr_t INSTR_PROF_PROFILE_COUNTER_BIAS_VAR
-    __attribute__((weak, alias(INSTR_PROF_QUOTE(
-                             INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR))));
-#endif
-
-static int writeMMappedFile(FILE *OutputFile, char **Profile) {
-  if (!OutputFile)
-    return -1;
-
-  /* Write the data into a file. */
-  setupIOBuffer();
-  ProfDataWriter fileWriter;
-  initFileWriter(&fileWriter, OutputFile);
-  if (lprofWriteData(&fileWriter, NULL, 0)) {
-    PROF_ERR("Failed to write profile: %s\n", strerror(errno));
-    return -1;
-  }
-  fflush(OutputFile);
-
-  /* Get the file size. */
-  uint64_t FileSize = ftell(OutputFile);
-
-  /* Map the profile. */
-  *Profile = (char *)mmap(
-      NULL, FileSize, PROT_READ | PROT_WRITE, MAP_SHARED, fileno(OutputFile), 0);
-  if (*Profile == MAP_FAILED) {
-    PROF_ERR("Unable to mmap profile: %s\n", strerror(errno));
-    return -1;
-  }
-
-  return 0;
-}
-
-static void initializeProfileForContinuousMode(void) {
-  if (!__llvm_profile_is_continuous_mode_enabled())
-    return;
-
-  /* This symbol is defined by the compiler when runtime counter relocation is
-   * used and runtime provides a weak alias so we can check if it's defined. */
-  void *BiasAddr = &INSTR_PROF_PROFILE_COUNTER_BIAS_VAR;
-  void *BiasDefaultAddr = &INSTR_PROF_PROFILE_COUNTER_BIAS_DEFAULT_VAR;
-  if (BiasAddr == BiasDefaultAddr) {
+  if (UseBiasVar && BiasAddr == BiasDefaultAddr) {
     PROF_ERR("%s\n", "__llvm_profile_counter_bias is undefined");
     return;
   }
 
-  /* Get the sizes of various profile data sections. Taken from
-   * __llvm_profile_get_size_for_buffer(). */
-  const __llvm_profile_data *DataBegin = __llvm_profile_begin_data();
-  const __llvm_profile_data *DataEnd = __llvm_profile_end_data();
+  /* Get the sizes of counter section. */
   const uint64_t *CountersBegin = __llvm_profile_begin_counters();
   const uint64_t *CountersEnd = __llvm_profile_end_counters();
-  uint64_t DataSize = __llvm_profile_get_data_size(DataBegin, DataEnd);
-  const uint64_t CountersOffset =
-      sizeof(__llvm_profile_header) + (DataSize * sizeof(__llvm_profile_data));
+  uint64_t CountersSize = CountersEnd - CountersBegin;
 
   int Length = getCurFilenameLength();
   char *FilenameBuf = (char *)COMPILER_RT_ALLOCA(Length + 1);
@@ -661,18 +589,12 @@ static void initializeProfileForContinuousMode(void) {
     return;
 
   FILE *File = NULL;
-  char *Profile = NULL;
-
-  if (!doMerging()) {
-    File = fopen(Filename, "w+b");
-    if (!File)
-      return;
-
-    if (writeMMappedFile(File, &Profile) == -1) {
-      fclose(File);
-      return;
-    }
-  } else {
+  uint64_t CurrentFileOffset = 0;
+  if (doMerging()) {
+    /* We are merging profiles. Map the counter section as shared memory into
+     * the profile, i.e. into each participating process. An increment in one
+     * process should be visible to every other process with the same counter
+     * section mapped. */
     File = lprofOpenFileEx(Filename);
     if (!File)
       return;
@@ -683,37 +605,54 @@ static void initializeProfileForContinuousMode(void) {
       fclose(File);
       return;
     }
-
-    if (!ProfileFileSize) {
-      if (writeMMappedFile(File, &Profile) == -1) {
+    if (ProfileFileSize == 0) {
+      /* Grow the profile so that mmap() can succeed.  Leak the file handle, as
+       * the file should stay open. */
+      if (writeProfileWithFileObject(Filename, File) != 0) {
+        lprofUnlockFileHandle(File);
         fclose(File);
         return;
       }
     } else {
       /* The merged profile has a non-zero length. Check that it is compatible
        * with the data in this process. */
-      if (mmapProfileForMerging(File, ProfileFileSize, &Profile) == -1) {
+      char *ProfileBuffer;
+      if (mmapProfileForMerging(File, ProfileFileSize, &ProfileBuffer) == -1) {
+        lprofUnlockFileHandle(File);
         fclose(File);
         return;
       }
+      (void)munmap(ProfileBuffer, ProfileFileSize);
+    }
+  } else {
+    File = fopen(Filename, FileOpenMode);
+    if (!File)
+      return;
+    /* Check that the offset within the file is page-aligned. */
+    CurrentFileOffset = ftell(File);
+    unsigned PageSize = getpagesize();
+    if (CurrentFileOffset % PageSize != 0) {
+      PROF_ERR("Continuous counter sync mode is enabled, but raw profile is not"
+               "page-aligned. CurrentFileOffset = %" PRIu64 ", pagesz = %u.\n",
+               (uint64_t)CurrentFileOffset, PageSize);
+      return;
+    }
+    if (writeProfileWithFileObject(Filename, File) != 0) {
+      fclose(File);
+      return;
     }
-
-    lprofUnlockFileHandle(File);
   }
 
-  /* Update the profile fields based on the current mapping. */
-  INSTR_PROF_PROFILE_COUNTER_BIAS_VAR =
-      (intptr_t)Profile - (uintptr_t)CountersBegin +
-      CountersOffset;
+  /* mmap() the profile counters so long as there is at least one counter.
+   * If there aren't any counters, mmap() would fail with EINVAL. */
+  if (CountersSize > 0)
+    mmapForContinuousMode(CurrentFileOffset, File);
 
-  /* Return the memory allocated for counters to OS. */
-  lprofReleaseMemoryPagesToOS((uintptr_t)CountersBegin, (uintptr_t)CountersEnd);
-}
-#else
-static void initializeProfileForContinuousMode(void) {
-  PROF_ERR("%s\n", "continuous mode is unsupported on this platform");
+  if (doMerging()) {
+    lprofUnlockFileHandle(File);
+    fclose(File);
+  }
 }
-#endif
 
 static const char *DefaultProfileName = "default.profraw";
 static void resetFilenameToDefault(void) {
@@ -1205,4 +1144,53 @@ int __llvm_profile_register_write_file_atexit(void) {
   return atexit(writeFileWithoutReturn);
 }
 
+COMPILER_RT_VISIBILITY int __llvm_profile_set_file_object(FILE *File,
+                                                          int EnableMerge) {
+  if (__llvm_profile_is_continuous_mode_enabled()) {
+    if (!EnableMerge) {
+      PROF_WARN("__llvm_profile_set_file_object(fd=%d) not supported in "
+                "continuous sync mode when merging is disabled\n",
+                fileno(File));
+      return 1;
+    }
+    if (lprofLockFileHandle(File) != 0) {
+      PROF_WARN("Data may be corrupted during profile merging : %s\n",
+                "Fail to obtain file lock due to system limit.");
+    }
+    uint64_t ProfileFileSize = 0;
+    if (getProfileFileSizeForMerging(File, &ProfileFileSize) == -1) {
+      lprofUnlockFileHandle(File);
+      return 1;
+    }
+    if (ProfileFileSize == 0) {
+      FreeHook = &free;
+      setupIOBuffer();
+      ProfDataWriter fileWriter;
+      initFileWriter(&fileWriter, File);
+      if (lprofWriteData(&fileWriter, 0, 0)) {
+        lprofUnlockFileHandle(File);
+        PROF_ERR("Failed to write file \"%d\": %s\n", fileno(File),
+                 strerror(errno));
+        return 1;
+      }
+      fflush(File);
+    } else {
+      /* The merged profile has a non-zero length. Check that it is compatible
+       * with the data in this process. */
+      char *ProfileBuffer;
+      if (mmapProfileForMerging(File, ProfileFileSize, &ProfileBuffer) == -1) {
+        lprofUnlockFileHandle(File);
+        return 1;
+      }
+      (void)munmap(ProfileBuffer, ProfileFileSize);
+    }
+    mmapForContinuousMode(0, File);
+    lprofUnlockFileHandle(File);
+  } else {
+    setProfileFile(File);
+    setProfileMergeRequested(EnableMerge);
+  }
+  return 0;
+}
+
 #endif
diff --git a/compiler-rt/lib/profile/InstrProfilingInternal.h b/compiler-rt/lib/profile/InstrProfilingInternal.h
index ffa790a4cb66..1394ea8c42f8 100644
--- a/compiler-rt/lib/profile/InstrProfilingInternal.h
+++ b/compiler-rt/lib/profile/InstrProfilingInternal.h
@@ -145,8 +145,8 @@ typedef struct VPDataReaderType {
                                         uint32_t N);
 } VPDataReaderType;
 
-/* Write profile data to destinitation. If SkipNameDataWrite is set to 1,
-   the name data is already in destintation, we just skip over it. */
+/* Write profile data to destination. If SkipNameDataWrite is set to 1,
+   the name data is already in destination, we just skip over it. */
 int lprofWriteData(ProfDataWriter *Writer, VPDataReaderType *VPDataReader,
                    int SkipNameDataWrite);
 int lprofWriteDataImpl(ProfDataWriter *Writer,
diff --git a/compiler-rt/lib/profile/InstrProfilingMerge.c b/compiler-rt/lib/profile/InstrProfilingMerge.c
index 913228513259..674b1898b046 100644
--- a/compiler-rt/lib/profile/InstrProfilingMerge.c
+++ b/compiler-rt/lib/profile/InstrProfilingMerge.c
@@ -22,6 +22,7 @@ void (*VPMergeHook)(ValueProfData *, __llvm_profile_data *);
 COMPILER_RT_VISIBILITY
 uint64_t lprofGetLoadModuleSignature() {
   /* A very fast way to compute a module signature.  */
+  uint64_t Version = __llvm_profile_get_version();
   uint64_t CounterSize = (uint64_t)(__llvm_profile_end_counters() -
                                     __llvm_profile_begin_counters());
   uint64_t DataSize = __llvm_profile_get_data_size(__llvm_profile_begin_data(),
@@ -33,7 +34,7 @@ uint64_t lprofGetLoadModuleSignature() {
   const __llvm_profile_data *FirstD = __llvm_profile_begin_data();
 
   return (NamesSize << 40) + (CounterSize << 30) + (DataSize << 20) +
-         (NumVnodes << 10) + (DataSize > 0 ? FirstD->NameRef : 0);
+         (NumVnodes << 10) + (DataSize > 0 ? FirstD->NameRef : 0) + Version;
 }
 
 /* Returns 1 if profile is not structurally compatible.  */
@@ -44,7 +45,8 @@ int __llvm_profile_check_compatibility(const char *ProfileData,
   __llvm_profile_header *Header = (__llvm_profile_header *)ProfileData;
   __llvm_profile_data *SrcDataStart, *SrcDataEnd, *SrcData, *DstData;
   SrcDataStart =
-      (__llvm_profile_data *)(ProfileData + sizeof(__llvm_profile_header));
+      (__llvm_profile_data *)(ProfileData + sizeof(__llvm_profile_header) +
+                              Header->BinaryIdsSize);
   SrcDataEnd = SrcDataStart + Header->DataSize;
 
   if (ProfileSize < sizeof(__llvm_profile_header))
@@ -63,7 +65,7 @@ int __llvm_profile_check_compatibility(const char *ProfileData,
       Header->ValueKindLast != IPVK_Last)
     return 1;
 
-  if (ProfileSize < sizeof(__llvm_profile_header) +
+  if (ProfileSize < sizeof(__llvm_profile_header) + Header->BinaryIdsSize +
                         Header->DataSize * sizeof(__llvm_profile_data) +
                         Header->NamesSize + Header->CountersSize)
     return 1;
@@ -81,6 +83,14 @@ int __llvm_profile_check_compatibility(const char *ProfileData,
   return 0;
 }
 
+static uintptr_t signextIfWin64(void *V) {
+#ifdef _WIN64
+  return (uintptr_t)(int32_t)(uintptr_t)V;
+#else
+  return (uintptr_t)V;
+#endif
+}
+
 COMPILER_RT_VISIBILITY
 int __llvm_profile_merge_from_buffer(const char *ProfileData,
                                      uint64_t ProfileSize) {
@@ -89,9 +99,11 @@ int __llvm_profile_merge_from_buffer(const char *ProfileData,
   uint64_t *SrcCountersStart;
   const char *SrcNameStart;
   const char *SrcValueProfDataStart, *SrcValueProfData;
+  uintptr_t CountersDelta = Header->CountersDelta;
 
   SrcDataStart =
-      (__llvm_profile_data *)(ProfileData + sizeof(__llvm_profile_header));
+      (__llvm_profile_data *)(ProfileData + sizeof(__llvm_profile_header) +
+                              Header->BinaryIdsSize);
   SrcDataEnd = SrcDataStart + Header->DataSize;
   SrcCountersStart = (uint64_t *)SrcDataEnd;
   SrcNameStart = (const char *)(SrcCountersStart + Header->CountersSize);
@@ -105,15 +117,30 @@ int __llvm_profile_merge_from_buffer(const char *ProfileData,
       DstData = (__llvm_profile_data *)__llvm_profile_begin_data(),
       SrcValueProfData = SrcValueProfDataStart;
        SrcData < SrcDataEnd; ++SrcData, ++DstData) {
-    uint64_t *DstCounters = (uint64_t *)DstData->CounterPtr;
+    // For the in-memory destination, CounterPtr is the distance from the start
+    // address of the data to the start address of the counter. On WIN64,
+    // CounterPtr is a truncated 32-bit value due to COFF limitation. Sign
+    // extend CounterPtr to get the original value.
+    uint64_t *DstCounters =
+        (uint64_t *)((uintptr_t)DstData + signextIfWin64(DstData->CounterPtr));
     unsigned NVK = 0;
 
+    // SrcData is a serialized representation of the memory image. We need to
+    // compute the in-buffer counter offset from the in-memory address distance.
+    // The initial CountersDelta is the in-memory address difference
+    // start(__llvm_prf_cnts)-start(__llvm_prf_data), so SrcData->CounterPtr -
+    // CountersDelta computes the offset into the in-buffer counter section.
+    //
+    // On WIN64, CountersDelta is truncated as well, so no need for signext.
+    uint64_t *SrcCounters =
+        SrcCountersStart +
+        ((uintptr_t)SrcData->CounterPtr - CountersDelta) / sizeof(uint64_t);
+    // CountersDelta needs to be decreased as we advance to the next data
+    // record.
+    CountersDelta -= sizeof(*SrcData);
     unsigned NC = SrcData->NumCounters;
     if (NC == 0)
       return 1;
-    uint64_t *SrcCounters = SrcCountersStart + ((size_t)SrcData->CounterPtr -
-                                                Header->CountersDelta) /
-                                                   sizeof(uint64_t);
     if (SrcCounters < SrcCountersStart ||
         (const char *)SrcCounters >= SrcNameStart ||
         (const char *)(SrcCounters + NC) > SrcNameStart)
diff --git a/compiler-rt/lib/profile/InstrProfilingPlatformFuchsia.c b/compiler-rt/lib/profile/InstrProfilingPlatformFuchsia.c
index 0146b14c193f..9bea795e8e3a 100644
--- a/compiler-rt/lib/profile/InstrProfilingPlatformFuchsia.c
+++ b/compiler-rt/lib/profile/InstrProfilingPlatformFuchsia.c
@@ -52,7 +52,7 @@ static inline void lprofWrite(const char *fmt, ...) {
   int ret = vsnprintf(s, sizeof(s), fmt, ap);
   va_end(ap);
 
-  __sanitizer_log_write(s, ret + 1);
+  __sanitizer_log_write(s, ret);
 }
 
 struct lprofVMOWriterCtx {
@@ -120,7 +120,8 @@ void __llvm_profile_initialize(void) {
   const uint64_t *CountersEnd = __llvm_profile_end_counters();
   const uint64_t DataSize = __llvm_profile_get_data_size(DataBegin, DataEnd);
   const uint64_t CountersOffset =
-      sizeof(__llvm_profile_header) + (DataSize * sizeof(__llvm_profile_data));
+      sizeof(__llvm_profile_header) + __llvm_write_binary_ids(NULL) +
+      (DataSize * sizeof(__llvm_profile_data));
   uint64_t CountersSize = CountersEnd - CountersBegin;
 
   /* Don't publish a VMO if there are no counters. */
@@ -178,9 +179,6 @@ void __llvm_profile_initialize(void) {
    * also consumes the VMO handle. */
   __sanitizer_publish_data(ProfileSinkName, Vmo);
 
-  /* Use the dumpfile symbolizer markup element to write the name of VMO. */
-  lprofWrite("LLVM Profile: {{{dumpfile:%s:%s}}}\n", ProfileSinkName, VmoName);
-
   /* Update the profile fields based on the current mapping. */
   INSTR_PROF_PROFILE_COUNTER_BIAS_VAR =
       (intptr_t)Mapping - (uintptr_t)CountersBegin + CountersOffset;
diff --git a/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c b/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c
index 508624a80cd6..e61f90b2cef9 100644
--- a/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c
+++ b/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c
@@ -17,6 +17,15 @@
 #include "InstrProfiling.h"
 #include "InstrProfilingInternal.h"
 
+#if defined(__FreeBSD__) && !defined(ElfW)
+/*
+ * FreeBSD's elf.h and link.h headers do not define the ElfW(type) macro yet.
+ * If this is added to all supported FreeBSD versions in the future, this
+ * compatibility macro can be removed.
+ */
+#define ElfW(type) __ElfN(type)
+#endif
+
 #define PROF_DATA_START INSTR_PROF_SECT_START(INSTR_PROF_DATA_COMMON)
 #define PROF_DATA_STOP INSTR_PROF_SECT_STOP(INSTR_PROF_DATA_COMMON)
 #define PROF_NAME_START INSTR_PROF_SECT_START(INSTR_PROF_NAME_COMMON)
@@ -76,6 +85,7 @@ COMPILER_RT_VISIBILITY ValueProfNode *__llvm_profile_end_vnodes(void) {
 COMPILER_RT_VISIBILITY ValueProfNode *CurrentVNode = &PROF_VNODES_START;
 COMPILER_RT_VISIBILITY ValueProfNode *EndVNode = &PROF_VNODES_STOP;
 
+#ifdef NT_GNU_BUILD_ID
 static size_t RoundUp(size_t size, size_t align) {
   return (size + align - 1) & ~(align - 1);
 }
@@ -84,11 +94,14 @@ static size_t RoundUp(size_t size, size_t align) {
  * Write binary id length and then its data, because binary id does not
  * have a fixed length.
  */
-int WriteOneBinaryId(ProfDataWriter *Writer, uint64_t BinaryIdLen,
-                     const uint8_t *BinaryIdData) {
+static int WriteOneBinaryId(ProfDataWriter *Writer, uint64_t BinaryIdLen,
+                            const uint8_t *BinaryIdData,
+                            uint64_t BinaryIdPadding) {
   ProfDataIOVec BinaryIdIOVec[] = {
       {&BinaryIdLen, sizeof(uint64_t), 1, 0},
-      {BinaryIdData, sizeof(uint8_t), BinaryIdLen, 0}};
+      {BinaryIdData, sizeof(uint8_t), BinaryIdLen, 0},
+      {NULL, sizeof(uint8_t), BinaryIdPadding, 1},
+  };
   if (Writer->Write(Writer, BinaryIdIOVec,
                     sizeof(BinaryIdIOVec) / sizeof(*BinaryIdIOVec)))
     return -1;
@@ -109,7 +122,8 @@ int WriteOneBinaryId(ProfDataWriter *Writer, uint64_t BinaryIdLen,
  * Note sections like .note.ABI-tag and .note.gnu.build-id are aligned
  * to 4 bytes, so round n_namesz and n_descsz to the nearest 4 bytes.
  */
-int WriteBinaryIdForNote(ProfDataWriter *Writer, const ElfW(Nhdr) * Note) {
+static int WriteBinaryIdForNote(ProfDataWriter *Writer,
+                                const ElfW(Nhdr) * Note) {
   int BinaryIdSize = 0;
 
   const char *NoteName = (const char *)Note + sizeof(ElfW(Nhdr));
@@ -119,11 +133,12 @@ int WriteBinaryIdForNote(ProfDataWriter *Writer, const ElfW(Nhdr) * Note) {
     uint64_t BinaryIdLen = Note->n_descsz;
     const uint8_t *BinaryIdData =
         (const uint8_t *)(NoteName + RoundUp(Note->n_namesz, 4));
-    if (Writer != NULL &&
-        WriteOneBinaryId(Writer, BinaryIdLen, BinaryIdData) == -1)
+    uint8_t BinaryIdPadding = __llvm_profile_get_num_padding_bytes(BinaryIdLen);
+    if (Writer != NULL && WriteOneBinaryId(Writer, BinaryIdLen, BinaryIdData,
+                                           BinaryIdPadding) == -1)
       return -1;
 
-    BinaryIdSize = sizeof(BinaryIdLen) + BinaryIdLen;
+    BinaryIdSize = sizeof(BinaryIdLen) + BinaryIdLen + BinaryIdPadding;
   }
 
   return BinaryIdSize;
@@ -134,8 +149,8 @@ int WriteBinaryIdForNote(ProfDataWriter *Writer, const ElfW(Nhdr) * Note) {
  * If writer is given, write binary ids into profiles.
  * If an error happens while writing, return -1.
  */
-int WriteBinaryIds(ProfDataWriter *Writer, const ElfW(Nhdr) * Note,
-                   const ElfW(Nhdr) * NotesEnd) {
+static int WriteBinaryIds(ProfDataWriter *Writer, const ElfW(Nhdr) * Note,
+                          const ElfW(Nhdr) * NotesEnd) {
   int TotalBinaryIdsSize = 0;
   while (Note < NotesEnd) {
     int Result = WriteBinaryIdForNote(Writer, Note);
@@ -179,5 +194,14 @@ COMPILER_RT_VISIBILITY int __llvm_write_binary_ids(ProfDataWriter *Writer) {
 
   return 0;
 }
+#else /* !NT_GNU_BUILD_ID */
+/*
+ * Fallback implementation for targets that don't support the GNU
+ * extensions NT_GNU_BUILD_ID and __ehdr_start.
+ */
+COMPILER_RT_VISIBILITY int __llvm_write_binary_ids(ProfDataWriter *Writer) {
+  return 0;
+}
+#endif
 
 #endif
diff --git a/compiler-rt/lib/profile/InstrProfilingPlatformOther.c b/compiler-rt/lib/profile/InstrProfilingPlatformOther.c
index 0e59148e2044..48946ce94253 100644
--- a/compiler-rt/lib/profile/InstrProfilingPlatformOther.c
+++ b/compiler-rt/lib/profile/InstrProfilingPlatformOther.c
@@ -46,17 +46,19 @@ void __llvm_profile_register_function(void *Data_) {
   if (!DataFirst) {
     DataFirst = Data;
     DataLast = Data + 1;
-    CountersFirst = Data->CounterPtr;
-    CountersLast = (uint64_t *)Data->CounterPtr + Data->NumCounters;
+    CountersFirst = (uint64_t *)((uintptr_t)Data_ + Data->CounterPtr);
+    CountersLast = CountersFirst + Data->NumCounters;
     return;
   }
 
   DataFirst = (const __llvm_profile_data *)getMinAddr(DataFirst, Data);
-  CountersFirst = (uint64_t *)getMinAddr(CountersFirst, Data->CounterPtr);
+  CountersFirst = (uint64_t *)getMinAddr(
+      CountersFirst, (uint64_t *)((uintptr_t)Data_ + Data->CounterPtr));
 
   DataLast = (const __llvm_profile_data *)getMaxAddr(DataLast, Data + 1);
   CountersLast = (uint64_t *)getMaxAddr(
-      CountersLast, (uint64_t *)Data->CounterPtr + Data->NumCounters);
+      CountersLast,
+      (uint64_t *)((uintptr_t)Data_ + Data->CounterPtr) + Data->NumCounters);
 }
 
 COMPILER_RT_VISIBILITY
diff --git a/compiler-rt/lib/profile/InstrProfilingUtil.c b/compiler-rt/lib/profile/InstrProfilingUtil.c
index 4fa792b72eac..d563e333aca8 100644
--- a/compiler-rt/lib/profile/InstrProfilingUtil.c
+++ b/compiler-rt/lib/profile/InstrProfilingUtil.c
@@ -34,9 +34,15 @@
 #endif
 
 #if defined(__Fuchsia__)
+#include <zircon/process.h>
 #include <zircon/syscalls.h>
 #endif
 
+#if defined(__FreeBSD__)
+#include <signal.h>
+#include <sys/procctl.h>
+#endif
+
 #include "InstrProfiling.h"
 #include "InstrProfilingUtil.h"
 
@@ -325,6 +331,12 @@ COMPILER_RT_VISIBILITY int lprofSuspendSigKill() {
   if (prctl(PR_GET_PDEATHSIG, &PDeachSig) == 0 && PDeachSig == SIGKILL)
     prctl(PR_SET_PDEATHSIG, 0);
   return (PDeachSig == SIGKILL);
+#elif defined(__FreeBSD__)
+  int PDeachSig = 0, PDisableSig = 0;
+  if (procctl(P_PID, 0, PROC_PDEATHSIG_STATUS, &PDeachSig) == 0 &&
+      PDeachSig == SIGKILL)
+    procctl(P_PID, 0, PROC_PDEATHSIG_CTL, &PDisableSig);
+  return (PDeachSig == SIGKILL);
 #else
   return 0;
 #endif
@@ -333,11 +345,18 @@ COMPILER_RT_VISIBILITY int lprofSuspendSigKill() {
 COMPILER_RT_VISIBILITY void lprofRestoreSigKill() {
 #if defined(__linux__)
   prctl(PR_SET_PDEATHSIG, SIGKILL);
+#elif defined(__FreeBSD__)
+  int PEnableSig = SIGKILL;
+  procctl(P_PID, 0, PROC_PDEATHSIG_CTL, &PEnableSig);
 #endif
 }
 
 COMPILER_RT_VISIBILITY int lprofReleaseMemoryPagesToOS(uintptr_t Begin,
                                                        uintptr_t End) {
+#if defined(__ve__)
+  // VE doesn't support madvise.
+  return 0;
+#else
   size_t PageSize = getpagesize();
   uintptr_t BeginAligned = lprofRoundUpTo((uintptr_t)Begin, PageSize);
   uintptr_t EndAligned = lprofRoundDownTo((uintptr_t)End, PageSize);
@@ -352,4 +371,5 @@ COMPILER_RT_VISIBILITY int lprofReleaseMemoryPagesToOS(uintptr_t Begin,
 #endif
   }
   return 0;
+#endif
 }
diff --git a/compiler-rt/lib/profile/InstrProfilingValue.c b/compiler-rt/lib/profile/InstrProfilingValue.c
index 7f368b9f8d4e..08197fdd9ea2 100644
--- a/compiler-rt/lib/profile/InstrProfilingValue.c
+++ b/compiler-rt/lib/profile/InstrProfilingValue.c
@@ -253,7 +253,7 @@ __llvm_profile_instrument_memop(uint64_t TargetValue, void *Data,
 /*
  * A wrapper struct that represents value profile runtime data.
  * Like InstrProfRecord class which is used by profiling host tools,
- * ValueProfRuntimeRecord also implements the abstract intefaces defined in
+ * ValueProfRuntimeRecord also implements the abstract interfaces defined in
  * ValueProfRecordClosure so that the runtime data can be serialized using
  * shared C implementation.
  */
diff --git a/compiler-rt/lib/profile/InstrProfilingVersionVar.c b/compiler-rt/lib/profile/InstrProfilingVersionVar.c
index a6f222150794..e49d171cce41 100644
--- a/compiler-rt/lib/profile/InstrProfilingVersionVar.c
+++ b/compiler-rt/lib/profile/InstrProfilingVersionVar.c
@@ -13,5 +13,14 @@
  * The runtime should only provide its own definition of this symbol when the
  * user has not specified one. Set this up by moving the runtime's copy of this
  * symbol to an object file within the archive.
+ *
+ * Hide this symbol everywhere except Apple platforms, where its presence is
+ * checked by the TAPI tool.
  */
-COMPILER_RT_WEAK uint64_t INSTR_PROF_RAW_VERSION_VAR = INSTR_PROF_RAW_VERSION;
+#if !defined(__APPLE__)
+#define VERSION_VAR_VISIBILITY COMPILER_RT_VISIBILITY
+#else
+#define VERSION_VAR_VISIBILITY
+#endif
+VERSION_VAR_VISIBILITY COMPILER_RT_WEAK uint64_t INSTR_PROF_RAW_VERSION_VAR =
+    INSTR_PROF_RAW_VERSION;
diff --git a/compiler-rt/lib/profile/InstrProfilingWriter.c b/compiler-rt/lib/profile/InstrProfilingWriter.c
index 25f630293227..9cb05570989d 100644
--- a/compiler-rt/lib/profile/InstrProfilingWriter.c
+++ b/compiler-rt/lib/profile/InstrProfilingWriter.c
@@ -32,7 +32,7 @@ static uint32_t VPDataArraySize = sizeof(VPDataArray) / sizeof(*VPDataArray);
 COMPILER_RT_VISIBILITY uint8_t *DynamicBufferIOBuffer = 0;
 COMPILER_RT_VISIBILITY uint32_t VPBufferSize = 0;
 
-/* The buffer writer is reponsponsible in keeping writer state
+/* The buffer writer is responsible in keeping writer state
  * across the call.
  */
 COMPILER_RT_VISIBILITY uint32_t lprofBufferWriter(ProfDataWriter *This,
@@ -283,6 +283,12 @@ lprofWriteDataImpl(ProfDataWriter *Writer, const __llvm_profile_data *DataBegin,
 #define INSTR_PROF_RAW_HEADER(Type, Name, Init) Header.Name = Init;
 #include "profile/InstrProfData.inc"
 
+  /* On WIN64, label differences are truncated 32-bit values. Truncate
+   * CountersDelta to match. */
+#ifdef _WIN64
+  Header.CountersDelta = (uint32_t)Header.CountersDelta;
+#endif
+
   /* Write the profile header. */
   ProfDataIOVec IOVec[] = {{&Header, sizeof(__llvm_profile_header), 1, 0}};
   if (Writer->Write(Writer, IOVec, sizeof(IOVec) / sizeof(*IOVec)))
diff --git a/compiler-rt/lib/sanitizer_common/sancov_flags.inc b/compiler-rt/lib/sanitizer_common/sancov_flags.inc
index cca33fc359f4..de9ede217fc3 100644
--- a/compiler-rt/lib/sanitizer_common/sancov_flags.inc
+++ b/compiler-rt/lib/sanitizer_common/sancov_flags.inc
@@ -14,7 +14,7 @@
 #endif
 
 SANCOV_FLAG(bool, symbolize, true,
-            "If set, converage information will be symbolized by sancov tool "
+            "If set, coverage information will be symbolized by sancov tool "
             "after dumping.")
 
 SANCOV_FLAG(bool, help, false, "Print flags help.")
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_addrhashmap.h b/compiler-rt/lib/sanitizer_common/sanitizer_addrhashmap.h
index 15f81a04350f..7e2fa91089f1 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_addrhashmap.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_addrhashmap.h
@@ -39,6 +39,11 @@ namespace __sanitizer {
 //   the current thread has exclusive access to the data
 //   if !h.exists() then the element never existed
 // }
+// {
+//   Map::Handle h(&m, addr, false, true);
+//   this will create a new element or return a handle to an existing element
+//   if !h.created() this thread does *not* have exclusive access to the data
+// }
 template<typename T, uptr kSize>
 class AddrHashMap {
  private:
@@ -56,7 +61,7 @@ class AddrHashMap {
   static const uptr kBucketSize = 3;
 
   struct Bucket {
-    RWMutex          mtx;
+    Mutex mtx;
     atomic_uintptr_t add;
     Cell             cells[kBucketSize];
   };
@@ -89,6 +94,12 @@ class AddrHashMap {
     bool                   create_;
   };
 
+  typedef void (*ForEachCallback)(const uptr key, const T &val, void *arg);
+  // ForEach acquires a lock on each bucket while iterating over
+  // elements. Note that this only ensures that the structure of the hashmap is
+  // unchanged, there may be a data race to the element itself.
+  void ForEach(ForEachCallback cb, void *arg);
+
  private:
   friend class Handle;
   Bucket *table_;
@@ -98,6 +109,33 @@ class AddrHashMap {
   uptr calcHash(uptr addr);
 };
 
+template <typename T, uptr kSize>
+void AddrHashMap<T, kSize>::ForEach(ForEachCallback cb, void *arg) {
+  for (uptr n = 0; n < kSize; n++) {
+    Bucket *bucket = &table_[n];
+
+    ReadLock lock(&bucket->mtx);
+
+    for (uptr i = 0; i < kBucketSize; i++) {
+      Cell *c = &bucket->cells[i];
+      uptr addr1 = atomic_load(&c->addr, memory_order_acquire);
+      if (addr1 != 0)
+        cb(addr1, c->val, arg);
+    }
+
+    // Iterate over any additional cells.
+    if (AddBucket *add =
+            (AddBucket *)atomic_load(&bucket->add, memory_order_acquire)) {
+      for (uptr i = 0; i < add->size; i++) {
+        Cell *c = &add->cells[i];
+        uptr addr1 = atomic_load(&c->addr, memory_order_acquire);
+        if (addr1 != 0)
+          cb(addr1, c->val, arg);
+      }
+    }
+  }
+}
+
 template<typename T, uptr kSize>
 AddrHashMap<T, kSize>::Handle::Handle(AddrHashMap<T, kSize> *map, uptr addr) {
   map_ = map;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_allocator.h b/compiler-rt/lib/sanitizer_common/sanitizer_allocator.h
index 5ec47416fe0c..ec23465d9584 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_allocator.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_allocator.h
@@ -14,6 +14,7 @@
 #define SANITIZER_ALLOCATOR_H
 
 #include "sanitizer_common.h"
+#include "sanitizer_flat_map.h"
 #include "sanitizer_internal_defs.h"
 #include "sanitizer_lfstack.h"
 #include "sanitizer_libc.h"
@@ -43,12 +44,6 @@ void SetAllocatorOutOfMemory();
 
 void PrintHintAllocatorCannotReturnNull();
 
-// Allocators call these callbacks on mmap/munmap.
-struct NoOpMapUnmapCallback {
-  void OnMap(uptr p, uptr size) const { }
-  void OnUnmap(uptr p, uptr size) const { }
-};
-
 // Callback type for iterating over chunks.
 typedef void (*ForEachChunkCallback)(uptr chunk, void *arg);
 
@@ -70,7 +65,6 @@ inline void RandomShuffle(T *a, u32 n, u32 *rand_state) {
 #include "sanitizer_allocator_size_class_map.h"
 #include "sanitizer_allocator_stats.h"
 #include "sanitizer_allocator_primary64.h"
-#include "sanitizer_allocator_bytemap.h"
 #include "sanitizer_allocator_primary32.h"
 #include "sanitizer_allocator_local_cache.h"
 #include "sanitizer_allocator_secondary.h"
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_bytemap.h b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_bytemap.h
deleted file mode 100644
index 0084bb62c83c..000000000000
--- a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_bytemap.h
+++ /dev/null
@@ -1,107 +0,0 @@
-//===-- sanitizer_allocator_bytemap.h ---------------------------*- C++ -*-===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-//
-// Part of the Sanitizer Allocator.
-//
-//===----------------------------------------------------------------------===//
-#ifndef SANITIZER_ALLOCATOR_H
-#error This file must be included inside sanitizer_allocator.h
-#endif
-
-// Maps integers in rage [0, kSize) to u8 values.
-template <u64 kSize, typename AddressSpaceViewTy = LocalAddressSpaceView>
-class FlatByteMap {
- public:
-  using AddressSpaceView = AddressSpaceViewTy;
-  void Init() {
-    internal_memset(map_, 0, sizeof(map_));
-  }
-
-  void set(uptr idx, u8 val) {
-    CHECK_LT(idx, kSize);
-    CHECK_EQ(0U, map_[idx]);
-    map_[idx] = val;
-  }
-  u8 operator[] (uptr idx) {
-    CHECK_LT(idx, kSize);
-    // FIXME: CHECK may be too expensive here.
-    return map_[idx];
-  }
- private:
-  u8 map_[kSize];
-};
-
-// TwoLevelByteMap maps integers in range [0, kSize1*kSize2) to u8 values.
-// It is implemented as a two-dimensional array: array of kSize1 pointers
-// to kSize2-byte arrays. The secondary arrays are mmaped on demand.
-// Each value is initially zero and can be set to something else only once.
-// Setting and getting values from multiple threads is safe w/o extra locking.
-template <u64 kSize1, u64 kSize2,
-          typename AddressSpaceViewTy = LocalAddressSpaceView,
-          class MapUnmapCallback = NoOpMapUnmapCallback>
-class TwoLevelByteMap {
- public:
-  using AddressSpaceView = AddressSpaceViewTy;
-  void Init() {
-    internal_memset(map1_, 0, sizeof(map1_));
-    mu_.Init();
-  }
-
-  void TestOnlyUnmap() {
-    for (uptr i = 0; i < kSize1; i++) {
-      u8 *p = Get(i);
-      if (!p) continue;
-      MapUnmapCallback().OnUnmap(reinterpret_cast<uptr>(p), kSize2);
-      UnmapOrDie(p, kSize2);
-    }
-  }
-
-  uptr size() const { return kSize1 * kSize2; }
-  uptr size1() const { return kSize1; }
-  uptr size2() const { return kSize2; }
-
-  void set(uptr idx, u8 val) {
-    CHECK_LT(idx, kSize1 * kSize2);
-    u8 *map2 = GetOrCreate(idx / kSize2);
-    CHECK_EQ(0U, map2[idx % kSize2]);
-    map2[idx % kSize2] = val;
-  }
-
-  u8 operator[] (uptr idx) const {
-    CHECK_LT(idx, kSize1 * kSize2);
-    u8 *map2 = Get(idx / kSize2);
-    if (!map2) return 0;
-    auto value_ptr = AddressSpaceView::Load(&map2[idx % kSize2]);
-    return *value_ptr;
-  }
-
- private:
-  u8 *Get(uptr idx) const {
-    CHECK_LT(idx, kSize1);
-    return reinterpret_cast<u8 *>(
-        atomic_load(&map1_[idx], memory_order_acquire));
-  }
-
-  u8 *GetOrCreate(uptr idx) {
-    u8 *res = Get(idx);
-    if (!res) {
-      SpinMutexLock l(&mu_);
-      if (!(res = Get(idx))) {
-        res = (u8*)MmapOrDie(kSize2, "TwoLevelByteMap");
-        MapUnmapCallback().OnMap(reinterpret_cast<uptr>(res), kSize2);
-        atomic_store(&map1_[idx], reinterpret_cast<uptr>(res),
-                     memory_order_release);
-      }
-    }
-    return res;
-  }
-
-  atomic_uintptr_t map1_[kSize1];
-  StaticSpinMutex mu_;
-};
-
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_combined.h b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_combined.h
index 0e81e6764f9a..9a3602f730b3 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_combined.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_combined.h
@@ -112,15 +112,13 @@ class CombinedAllocator {
     return new_p;
   }
 
-  bool PointerIsMine(void *p) {
+  bool PointerIsMine(const void *p) const {
     if (primary_.PointerIsMine(p))
       return true;
     return secondary_.PointerIsMine(p);
   }
 
-  bool FromPrimary(void *p) {
-    return primary_.PointerIsMine(p);
-  }
+  bool FromPrimary(const void *p) const { return primary_.PointerIsMine(p); }
 
   void *GetMetaData(const void *p) {
     if (primary_.PointerIsMine(p))
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_dlsym.h b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_dlsym.h
new file mode 100644
index 000000000000..92b1373ef84d
--- /dev/null
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_dlsym.h
@@ -0,0 +1,79 @@
+//===-- sanitizer_allocator_dlsym.h -----------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// Hack: Sanitizer initializer calls dlsym which may need to allocate and call
+// back into uninitialized sanitizer.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef SANITIZER_ALLOCATOR_DLSYM_H
+#define SANITIZER_ALLOCATOR_DLSYM_H
+
+#include "sanitizer_allocator_internal.h"
+
+namespace __sanitizer {
+
+template <typename Details>
+struct DlSymAllocator {
+  static bool Use() {
+    // Fuchsia doesn't use dlsym-based interceptors.
+    return !SANITIZER_FUCHSIA && UNLIKELY(Details::UseImpl());
+  }
+
+  static bool PointerIsMine(const void *ptr) {
+    // Fuchsia doesn't use dlsym-based interceptors.
+    return !SANITIZER_FUCHSIA &&
+           UNLIKELY(internal_allocator()->FromPrimary(ptr));
+  }
+
+  static void *Allocate(uptr size_in_bytes) {
+    void *ptr = InternalAlloc(size_in_bytes, nullptr, kWordSize);
+    CHECK(internal_allocator()->FromPrimary(ptr));
+    Details::OnAllocate(ptr,
+                        internal_allocator()->GetActuallyAllocatedSize(ptr));
+    return ptr;
+  }
+
+  static void *Callocate(SIZE_T nmemb, SIZE_T size) {
+    void *ptr = InternalCalloc(nmemb, size);
+    CHECK(internal_allocator()->FromPrimary(ptr));
+    Details::OnAllocate(ptr,
+                        internal_allocator()->GetActuallyAllocatedSize(ptr));
+    return ptr;
+  }
+
+  static void Free(void *ptr) {
+    uptr size = internal_allocator()->GetActuallyAllocatedSize(ptr);
+    Details::OnFree(ptr, size);
+    InternalFree(ptr);
+  }
+
+  static void *Realloc(void *ptr, uptr new_size) {
+    if (!ptr)
+      return Allocate(new_size);
+    CHECK(internal_allocator()->FromPrimary(ptr));
+    if (!new_size) {
+      Free(ptr);
+      return nullptr;
+    }
+    uptr size = internal_allocator()->GetActuallyAllocatedSize(ptr);
+    uptr memcpy_size = Min(new_size, size);
+    void *new_ptr = Allocate(new_size);
+    if (new_ptr)
+      internal_memcpy(new_ptr, ptr, memcpy_size);
+    Free(ptr);
+    return new_ptr;
+  }
+
+  static void OnAllocate(const void *ptr, uptr size) {}
+  static void OnFree(const void *ptr, uptr size) {}
+};
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_ALLOCATOR_DLSYM_H
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary32.h b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary32.h
index 38d2a7d117fb..ae1b7e0d5f1c 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary32.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary32.h
@@ -189,7 +189,7 @@ class SizeClassAllocator32 {
     sci->free_list.push_front(b);
   }
 
-  bool PointerIsMine(const void *p) {
+  bool PointerIsMine(const void *p) const {
     uptr mem = reinterpret_cast<uptr>(p);
     if (SANITIZER_SIGN_EXTENDED_ADDRESSES)
       mem &= (kSpaceSize - 1);
@@ -198,8 +198,9 @@ class SizeClassAllocator32 {
     return GetSizeClass(p) != 0;
   }
 
-  uptr GetSizeClass(const void *p) {
-    return possible_regions[ComputeRegionId(reinterpret_cast<uptr>(p))];
+  uptr GetSizeClass(const void *p) const {
+    uptr id = ComputeRegionId(reinterpret_cast<uptr>(p));
+    return possible_regions.contains(id) ? possible_regions[id] : 0;
   }
 
   void *GetBlockBegin(const void *p) {
@@ -251,9 +252,9 @@ class SizeClassAllocator32 {
 
   // Iterate over all existing chunks.
   // The allocator must be locked when calling this function.
-  void ForEachChunk(ForEachChunkCallback callback, void *arg) {
+  void ForEachChunk(ForEachChunkCallback callback, void *arg) const {
     for (uptr region = 0; region < kNumPossibleRegions; region++)
-      if (possible_regions[region]) {
+      if (possible_regions.contains(region) && possible_regions[region]) {
         uptr chunk_size = ClassIdToSize(possible_regions[region]);
         uptr max_chunks_in_region = kRegionSize / (chunk_size + kMetadataSize);
         uptr region_beg = region * kRegionSize;
@@ -292,9 +293,7 @@ class SizeClassAllocator32 {
     return res;
   }
 
-  uptr ComputeRegionBeg(uptr mem) {
-    return mem & ~(kRegionSize - 1);
-  }
+  uptr ComputeRegionBeg(uptr mem) const { return mem & ~(kRegionSize - 1); }
 
   uptr AllocateRegion(AllocatorStats *stat, uptr class_id) {
     DCHECK_LT(class_id, kNumClasses);
@@ -305,7 +304,7 @@ class SizeClassAllocator32 {
     MapUnmapCallback().OnMap(res, kRegionSize);
     stat->Add(AllocatorStatMapped, kRegionSize);
     CHECK(IsAligned(res, kRegionSize));
-    possible_regions.set(ComputeRegionId(res), static_cast<u8>(class_id));
+    possible_regions[ComputeRegionId(res)] = class_id;
     return res;
   }
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary64.h b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary64.h
index b142ee0131b2..f917310cfebb 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary64.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary64.h
@@ -161,7 +161,7 @@ class SizeClassAllocator64 {
   void ForceReleaseToOS() {
     MemoryMapperT memory_mapper(*this);
     for (uptr class_id = 1; class_id < kNumClasses; class_id++) {
-      BlockingMutexLock l(&GetRegionInfo(class_id)->mutex);
+      Lock l(&GetRegionInfo(class_id)->mutex);
       MaybeReleaseToOS(&memory_mapper, class_id, true /*force*/);
     }
   }
@@ -178,7 +178,7 @@ class SizeClassAllocator64 {
     uptr region_beg = GetRegionBeginBySizeClass(class_id);
     CompactPtrT *free_array = GetFreeArray(region_beg);
 
-    BlockingMutexLock l(&region->mutex);
+    Lock l(&region->mutex);
     uptr old_num_chunks = region->num_freed_chunks;
     uptr new_num_freed_chunks = old_num_chunks + n_chunks;
     // Failure to allocate free array space while releasing memory is non
@@ -204,7 +204,7 @@ class SizeClassAllocator64 {
     uptr region_beg = GetRegionBeginBySizeClass(class_id);
     CompactPtrT *free_array = GetFreeArray(region_beg);
 
-    BlockingMutexLock l(&region->mutex);
+    Lock l(&region->mutex);
 #if SANITIZER_WINDOWS
     /* On Windows unmapping of memory during __sanitizer_purge_allocator is
     explicit and immediate, so unmapped regions must be explicitly mapped back
@@ -282,6 +282,8 @@ class SizeClassAllocator64 {
     CHECK(kMetadataSize);
     uptr class_id = GetSizeClass(p);
     uptr size = ClassIdToSize(class_id);
+    if (!size)
+      return nullptr;
     uptr chunk_idx = GetChunkIdx(reinterpret_cast<uptr>(p), size);
     uptr region_beg = GetRegionBeginBySizeClass(class_id);
     return reinterpret_cast<void *>(GetMetadataEnd(region_beg) -
@@ -300,9 +302,8 @@ class SizeClassAllocator64 {
     UnmapWithCallbackOrDie((uptr)address_range.base(), address_range.size());
   }
 
-  static void FillMemoryProfile(uptr start, uptr rss, bool file, uptr *stats,
-                           uptr stats_size) {
-    for (uptr class_id = 0; class_id < stats_size; class_id++)
+  static void FillMemoryProfile(uptr start, uptr rss, bool file, uptr *stats) {
+    for (uptr class_id = 0; class_id < kNumClasses; class_id++)
       if (stats[class_id] == start)
         stats[class_id] = rss;
   }
@@ -315,7 +316,7 @@ class SizeClassAllocator64 {
     Printf(
         "%s %02zd (%6zd): mapped: %6zdK allocs: %7zd frees: %7zd inuse: %6zd "
         "num_freed_chunks %7zd avail: %6zd rss: %6zdK releases: %6zd "
-        "last released: %6zdK region: 0x%zx\n",
+        "last released: %6lldK region: 0x%zx\n",
         region->exhausted ? "F" : " ", class_id, ClassIdToSize(class_id),
         region->mapped_user >> 10, region->stats.n_allocated,
         region->stats.n_freed, in_use, region->num_freed_chunks, avail_chunks,
@@ -328,7 +329,7 @@ class SizeClassAllocator64 {
     uptr rss_stats[kNumClasses];
     for (uptr class_id = 0; class_id < kNumClasses; class_id++)
       rss_stats[class_id] = SpaceBeg() + kRegionSize * class_id;
-    GetMemoryProfile(FillMemoryProfile, rss_stats, kNumClasses);
+    GetMemoryProfile(FillMemoryProfile, rss_stats);
 
     uptr total_mapped = 0;
     uptr total_rss = 0;
@@ -623,7 +624,7 @@ class SizeClassAllocator64 {
 
   static const uptr kRegionSize = kSpaceSize / kNumClassesRounded;
   // FreeArray is the array of free-d chunks (stored as 4-byte offsets).
-  // In the worst case it may reguire kRegionSize/SizeClassMap::kMinSize
+  // In the worst case it may require kRegionSize/SizeClassMap::kMinSize
   // elements, but in reality this will not happen. For simplicity we
   // dedicate 1/8 of the region's virtual space to FreeArray.
   static const uptr kFreeArraySize = kRegionSize / 8;
@@ -665,7 +666,7 @@ class SizeClassAllocator64 {
   };
 
   struct ALIGNED(SANITIZER_CACHE_LINE_SIZE) RegionInfo {
-    BlockingMutex mutex;
+    Mutex mutex;
     uptr num_freed_chunks;  // Number of elements in the freearray.
     uptr mapped_free_array;  // Bytes mapped for freearray.
     uptr allocated_user;  // Bytes allocated for user memory.
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_secondary.h b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_secondary.h
index dd34fe85cc3a..c24354cb5b2a 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_secondary.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_secondary.h
@@ -161,7 +161,7 @@ class LargeMmapAllocator {
     return res;
   }
 
-  bool PointerIsMine(const void *p) {
+  bool PointerIsMine(const void *p) const {
     return GetBlockBegin(p) != nullptr;
   }
 
@@ -179,7 +179,7 @@ class LargeMmapAllocator {
     return GetHeader(p) + 1;
   }
 
-  void *GetBlockBegin(const void *ptr) {
+  void *GetBlockBegin(const void *ptr) const {
     uptr p = reinterpret_cast<uptr>(ptr);
     SpinMutexLock l(&mutex_);
     uptr nearest_chunk = 0;
@@ -301,7 +301,7 @@ class LargeMmapAllocator {
     return GetHeader(reinterpret_cast<uptr>(p));
   }
 
-  void *GetUser(const Header *h) {
+  void *GetUser(const Header *h) const {
     CHECK(IsAligned((uptr)h, page_size_));
     return reinterpret_cast<void*>(reinterpret_cast<uptr>(h) + page_size_);
   }
@@ -318,5 +318,5 @@ class LargeMmapAllocator {
   struct Stats {
     uptr n_allocs, n_frees, currently_allocated, max_allocated, by_size_log[64];
   } stats;
-  StaticSpinMutex mutex_;
+  mutable StaticSpinMutex mutex_;
 };
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_size_class_map.h b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_size_class_map.h
index c50d13303ede..361793f2490a 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_allocator_size_class_map.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_allocator_size_class_map.h
@@ -193,13 +193,13 @@ class SizeClassMap {
       uptr cached = MaxCachedHint(s) * s;
       if (i == kBatchClassID)
         d = p = l = 0;
-      Printf("c%02zd => s: %zd diff: +%zd %02zd%% l %zd "
-             "cached: %zd %zd; id %zd\n",
-             i, Size(i), d, p, l, MaxCachedHint(s), cached, ClassID(s));
+      Printf(
+          "c%02zu => s: %zu diff: +%zu %02zu%% l %zu cached: %zu %zu; id %zu\n",
+          i, Size(i), d, p, l, MaxCachedHint(s), cached, ClassID(s));
       total_cached += cached;
       prev_s = s;
     }
-    Printf("Total cached: %zd\n", total_cached);
+    Printf("Total cached: %zu\n", total_cached);
   }
 
   static void Validate() {
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_asm.h b/compiler-rt/lib/sanitizer_common/sanitizer_asm.h
index 803af3285e18..9ebba91da73f 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_asm.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_asm.h
@@ -6,7 +6,7 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// Various support for assemebler.
+// Various support for assembler.
 //
 //===----------------------------------------------------------------------===//
 
@@ -61,8 +61,15 @@
 #if defined(__ELF__) && (defined(__GNU__) || defined(__FreeBSD__) || \
                          defined(__Fuchsia__) || defined(__linux__))
 // clang-format off
-#define NO_EXEC_STACK_DIRECTIVE .section .note.GNU-stack,"",%progbits  // NOLINT
+#define NO_EXEC_STACK_DIRECTIVE .section .note.GNU-stack,"",%progbits
 // clang-format on
 #else
 #define NO_EXEC_STACK_DIRECTIVE
 #endif
+
+#if (defined(__x86_64__) || defined(__i386__)) && defined(__has_include) && __has_include(<cet.h>)
+#include <cet.h>
+#endif
+#ifndef _CET_ENDBR
+#define _CET_ENDBR
+#endif
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_atomic_clang_mips.h b/compiler-rt/lib/sanitizer_common/sanitizer_atomic_clang_mips.h
index 2b39097112d4..f3d3052e5b7c 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_atomic_clang_mips.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_atomic_clang_mips.h
@@ -18,7 +18,7 @@ namespace __sanitizer {
 
 // MIPS32 does not support atomics > 4 bytes. To address this lack of
 // functionality, the sanitizer library provides helper methods which use an
-// internal spin lock mechanism to emulate atomic oprations when the size is
+// internal spin lock mechanism to emulate atomic operations when the size is
 // 8 bytes.
 static void __spin_lock(volatile int *lock) {
   while (__sync_lock_test_and_set(lock, 1))
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_chained_origin_depot.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_chained_origin_depot.cpp
index 250ac39e1301..472b83d63a08 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_chained_origin_depot.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_chained_origin_depot.cpp
@@ -11,16 +11,57 @@
 
 #include "sanitizer_chained_origin_depot.h"
 
+#include "sanitizer_stackdepotbase.h"
+
 namespace __sanitizer {
 
-bool ChainedOriginDepot::ChainedOriginDepotNode::eq(
-    u32 hash, const args_type &args) const {
-  return here_id == args.here_id && prev_id == args.prev_id;
-}
+namespace {
+struct ChainedOriginDepotDesc {
+  u32 here_id;
+  u32 prev_id;
+};
 
-uptr ChainedOriginDepot::ChainedOriginDepotNode::storage_size(
-    const args_type &args) {
-  return sizeof(ChainedOriginDepotNode);
+struct ChainedOriginDepotNode {
+  using hash_type = u32;
+  u32 link;
+  u32 here_id;
+  u32 prev_id;
+
+  typedef ChainedOriginDepotDesc args_type;
+
+  bool eq(hash_type hash, const args_type &args) const;
+
+  static uptr allocated() { return 0; }
+
+  static hash_type hash(const args_type &args);
+
+  static bool is_valid(const args_type &args);
+
+  void store(u32 id, const args_type &args, hash_type other_hash);
+
+  args_type load(u32 id) const;
+
+  struct Handle {
+    const ChainedOriginDepotNode *node_ = nullptr;
+    u32 id_ = 0;
+    Handle(const ChainedOriginDepotNode *node, u32 id) : node_(node), id_(id) {}
+    bool valid() const { return node_; }
+    u32 id() const { return id_; }
+    int here_id() const { return node_->here_id; }
+    int prev_id() const { return node_->prev_id; }
+  };
+
+  static Handle get_handle(u32 id);
+
+  typedef Handle handle_type;
+};
+
+}  // namespace
+
+static StackDepotBase<ChainedOriginDepotNode, 4, 20> depot;
+
+bool ChainedOriginDepotNode::eq(hash_type hash, const args_type &args) const {
+  return here_id == args.here_id && prev_id == args.prev_id;
 }
 
 /* This is murmur2 hash for the 64->32 bit case.
@@ -36,7 +77,8 @@ uptr ChainedOriginDepot::ChainedOriginDepotNode::storage_size(
    split, or one of two reserved values (-1) or (-2). Either case can
    dominate depending on the workload.
 */
-u32 ChainedOriginDepot::ChainedOriginDepotNode::hash(const args_type &args) {
+ChainedOriginDepotNode::hash_type ChainedOriginDepotNode::hash(
+    const args_type &args) {
   const u32 m = 0x5bd1e995;
   const u32 seed = 0x9747b28c;
   const u32 r = 24;
@@ -61,37 +103,33 @@ u32 ChainedOriginDepot::ChainedOriginDepotNode::hash(const args_type &args) {
   return h;
 }
 
-bool ChainedOriginDepot::ChainedOriginDepotNode::is_valid(
-    const args_type &args) {
-  return true;
-}
+bool ChainedOriginDepotNode::is_valid(const args_type &args) { return true; }
 
-void ChainedOriginDepot::ChainedOriginDepotNode::store(const args_type &args,
-                                                       u32 other_hash) {
+void ChainedOriginDepotNode::store(u32 id, const args_type &args,
+                                   hash_type other_hash) {
   here_id = args.here_id;
   prev_id = args.prev_id;
 }
 
-ChainedOriginDepot::ChainedOriginDepotNode::args_type
-ChainedOriginDepot::ChainedOriginDepotNode::load() const {
+ChainedOriginDepotNode::args_type ChainedOriginDepotNode::load(u32 id) const {
   args_type ret = {here_id, prev_id};
   return ret;
 }
 
-ChainedOriginDepot::ChainedOriginDepotNode::Handle
-ChainedOriginDepot::ChainedOriginDepotNode::get_handle() {
-  return Handle(this);
+ChainedOriginDepotNode::Handle ChainedOriginDepotNode::get_handle(u32 id) {
+  return Handle(&depot.nodes[id], id);
 }
 
 ChainedOriginDepot::ChainedOriginDepot() {}
 
-StackDepotStats *ChainedOriginDepot::GetStats() { return depot.GetStats(); }
+StackDepotStats ChainedOriginDepot::GetStats() const {
+  return depot.GetStats();
+}
 
 bool ChainedOriginDepot::Put(u32 here_id, u32 prev_id, u32 *new_id) {
   ChainedOriginDepotDesc desc = {here_id, prev_id};
   bool inserted;
-  ChainedOriginDepotNode::Handle h = depot.Put(desc, &inserted);
-  *new_id = h.valid() ? h.id() : 0;
+  *new_id = depot.Put(desc, &inserted);
   return inserted;
 }
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_chained_origin_depot.h b/compiler-rt/lib/sanitizer_common/sanitizer_chained_origin_depot.h
index 453cdf6b5449..2e800964a45d 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_chained_origin_depot.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_chained_origin_depot.h
@@ -13,7 +13,6 @@
 #define SANITIZER_CHAINED_ORIGIN_DEPOT_H
 
 #include "sanitizer_common.h"
-#include "sanitizer_stackdepotbase.h"
 
 namespace __sanitizer {
 
@@ -22,7 +21,7 @@ class ChainedOriginDepot {
   ChainedOriginDepot();
 
   // Gets the statistic of the origin chain storage.
-  StackDepotStats *GetStats();
+  StackDepotStats GetStats() const;
 
   // Stores a chain with StackDepot ID here_id and previous chain ID prev_id.
   // If successful, returns true and the new chain id new_id.
@@ -37,48 +36,6 @@ class ChainedOriginDepot {
   void UnlockAll();
 
  private:
-  struct ChainedOriginDepotDesc {
-    u32 here_id;
-    u32 prev_id;
-  };
-
-  struct ChainedOriginDepotNode {
-    ChainedOriginDepotNode *link;
-    u32 id;
-    u32 here_id;
-    u32 prev_id;
-
-    typedef ChainedOriginDepotDesc args_type;
-
-    bool eq(u32 hash, const args_type &args) const;
-
-    static uptr storage_size(const args_type &args);
-
-    static u32 hash(const args_type &args);
-
-    static bool is_valid(const args_type &args);
-
-    void store(const args_type &args, u32 other_hash);
-
-    args_type load() const;
-
-    struct Handle {
-      ChainedOriginDepotNode *node_;
-      Handle() : node_(nullptr) {}
-      explicit Handle(ChainedOriginDepotNode *node) : node_(node) {}
-      bool valid() { return node_; }
-      u32 id() { return node_->id; }
-      int here_id() { return node_->here_id; }
-      int prev_id() { return node_->prev_id; }
-    };
-
-    Handle get_handle();
-
-    typedef Handle handle_type;
-  };
-
-  StackDepotBase<ChainedOriginDepotNode, 4, 20> depot;
-
   ChainedOriginDepot(const ChainedOriginDepot &) = delete;
   void operator=(const ChainedOriginDepot &) = delete;
 };
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common.h b/compiler-rt/lib/sanitizer_common/sanitizer_common.h
index cbdbb0c4c4bd..065154496eb5 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common.h
@@ -192,12 +192,13 @@ class ReservedAddressRange {
 };
 
 typedef void (*fill_profile_f)(uptr start, uptr rss, bool file,
-                               /*out*/uptr *stats, uptr stats_size);
+                               /*out*/ uptr *stats);
 
 // Parse the contents of /proc/self/smaps and generate a memory profile.
-// |cb| is a tool-specific callback that fills the |stats| array containing
-// |stats_size| elements.
-void GetMemoryProfile(fill_profile_f cb, uptr *stats, uptr stats_size);
+// |cb| is a tool-specific callback that fills the |stats| array.
+void GetMemoryProfile(fill_profile_f cb, uptr *stats);
+void ParseUnixMemoryProfile(fill_profile_f cb, uptr *stats, char *smaps,
+                            uptr smaps_len);
 
 // Simple low-level (mmap-based) allocator for internal use. Doesn't have
 // constructor, so all instances of LowLevelAllocator should be
@@ -222,8 +223,8 @@ void CatastrophicErrorWrite(const char *buffer, uptr length);
 void RawWrite(const char *buffer);
 bool ColorizeReports();
 void RemoveANSIEscapeSequencesFromString(char *buffer);
-void Printf(const char *format, ...);
-void Report(const char *format, ...);
+void Printf(const char *format, ...) FORMAT(1, 2);
+void Report(const char *format, ...) FORMAT(1, 2);
 void SetPrintfAndReportCallback(void (*callback)(const char *));
 #define VReport(level, ...)                                              \
   do {                                                                   \
@@ -371,7 +372,7 @@ void ReportErrorSummary(const char *error_type, const AddressInfo &info,
 void ReportErrorSummary(const char *error_type, const StackTrace *trace,
                         const char *alt_tool_name = nullptr);
 
-void ReportMmapWriteExec(int prot);
+void ReportMmapWriteExec(int prot, int mflags);
 
 // Math
 #if SANITIZER_WINDOWS && !defined(__clang__) && !defined(__GNUC__)
@@ -419,9 +420,7 @@ inline uptr LeastSignificantSetBitIndex(uptr x) {
   return up;
 }
 
-inline bool IsPowerOfTwo(uptr x) {
-  return (x & (x - 1)) == 0;
-}
+inline constexpr bool IsPowerOfTwo(uptr x) { return (x & (x - 1)) == 0; }
 
 inline uptr RoundUpToPowerOfTwo(uptr size) {
   CHECK(size);
@@ -433,16 +432,16 @@ inline uptr RoundUpToPowerOfTwo(uptr size) {
   return 1ULL << (up + 1);
 }
 
-inline uptr RoundUpTo(uptr size, uptr boundary) {
+inline constexpr uptr RoundUpTo(uptr size, uptr boundary) {
   RAW_CHECK(IsPowerOfTwo(boundary));
   return (size + boundary - 1) & ~(boundary - 1);
 }
 
-inline uptr RoundDownTo(uptr x, uptr boundary) {
+inline constexpr uptr RoundDownTo(uptr x, uptr boundary) {
   return x & ~(boundary - 1);
 }
 
-inline bool IsAligned(uptr a, uptr alignment) {
+inline constexpr bool IsAligned(uptr a, uptr alignment) {
   return (a & (alignment - 1)) == 0;
 }
 
@@ -618,7 +617,7 @@ class InternalScopedString {
     buffer_.resize(1);
     buffer_[0] = '\0';
   }
-  void append(const char *format, ...);
+  void append(const char *format, ...) FORMAT(2, 3);
   const char *data() const { return buffer_.data(); }
   char *data() { return buffer_.data(); }
 
@@ -697,7 +696,8 @@ enum ModuleArch {
   kModuleArchARMV7S,
   kModuleArchARMV7K,
   kModuleArchARM64,
-  kModuleArchRISCV64
+  kModuleArchRISCV64,
+  kModuleArchHexagon
 };
 
 // Sorts and removes duplicates from the container.
@@ -721,12 +721,15 @@ void SortAndDedup(Container &v, Compare comp = {}) {
   v.resize(last + 1);
 }
 
+constexpr uptr kDefaultFileMaxSize = FIRST_32_SECOND_64(1 << 26, 1 << 28);
+
 // Opens the file 'file_name" and reads up to 'max_len' bytes.
 // The resulting buffer is mmaped and stored in '*buff'.
 // Returns true if file was successfully opened and read.
 bool ReadFileToVector(const char *file_name,
                       InternalMmapVectorNoCtor<char> *buff,
-                      uptr max_len = 1 << 26, error_t *errno_p = nullptr);
+                      uptr max_len = kDefaultFileMaxSize,
+                      error_t *errno_p = nullptr);
 
 // Opens the file 'file_name" and reads up to 'max_len' bytes.
 // This function is less I/O efficient than ReadFileToVector as it may reread
@@ -737,7 +740,7 @@ bool ReadFileToVector(const char *file_name,
 // The total number of read bytes is stored in '*read_len'.
 // Returns true if file was successfully opened and read.
 bool ReadFileToBuffer(const char *file_name, char **buff, uptr *buff_size,
-                      uptr *read_len, uptr max_len = 1 << 26,
+                      uptr *read_len, uptr max_len = kDefaultFileMaxSize,
                       error_t *errno_p = nullptr);
 
 // When adding a new architecture, don't forget to also update
@@ -764,6 +767,8 @@ inline const char *ModuleArchToString(ModuleArch arch) {
       return "arm64";
     case kModuleArchRISCV64:
       return "riscv64";
+    case kModuleArchHexagon:
+      return "hexagon";
   }
   CHECK(0 && "Invalid module arch");
   return "";
@@ -1063,17 +1068,10 @@ class ArrayRef {
   T *end_ = nullptr;
 };
 
-#define PRINTF_128(v)                                                         \
-  (*((u8 *)&v + 0)), (*((u8 *)&v + 1)), (*((u8 *)&v + 2)), (*((u8 *)&v + 3)), \
-      (*((u8 *)&v + 4)), (*((u8 *)&v + 5)), (*((u8 *)&v + 6)),                \
-      (*((u8 *)&v + 7)), (*((u8 *)&v + 8)), (*((u8 *)&v + 9)),                \
-      (*((u8 *)&v + 10)), (*((u8 *)&v + 11)), (*((u8 *)&v + 12)),             \
-      (*((u8 *)&v + 13)), (*((u8 *)&v + 14)), (*((u8 *)&v + 15))
-
 }  // namespace __sanitizer
 
 inline void *operator new(__sanitizer::operator_new_size_type size,
-                          __sanitizer::LowLevelAllocator &alloc) {  // NOLINT
+                          __sanitizer::LowLevelAllocator &alloc) {
   return alloc.Allocate(size);
 }
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc
index 6205d853a4c9..abb38ccfa15d 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc
@@ -204,7 +204,7 @@ extern const short *_tolower_tab_;
 
 #define COMMON_INTERCEPTOR_READ_STRING(ctx, s, n)                   \
     COMMON_INTERCEPTOR_READ_RANGE((ctx), (s),                       \
-      common_flags()->strict_string_checks ? (REAL(strlen)(s)) + 1 : (n) )
+      common_flags()->strict_string_checks ? (internal_strlen(s)) + 1 : (n) )
 
 #ifndef COMMON_INTERCEPTOR_ON_DLOPEN
 #define COMMON_INTERCEPTOR_ON_DLOPEN(filename, flag) \
@@ -435,7 +435,7 @@ INTERCEPTOR(char*, textdomain, const char *domainname) {
   if (domainname) COMMON_INTERCEPTOR_READ_STRING(ctx, domainname, 0);
   char *domain = REAL(textdomain)(domainname);
   if (domain) {
-    COMMON_INTERCEPTOR_INITIALIZE_RANGE(domain, REAL(strlen)(domain) + 1);
+    COMMON_INTERCEPTOR_INITIALIZE_RANGE(domain, internal_strlen(domain) + 1);
   }
   return domain;
 }
@@ -575,8 +575,8 @@ INTERCEPTOR(int, strncasecmp, const char *s1, const char *s2, SIZE_T size) {
 #if SANITIZER_INTERCEPT_STRSTR || SANITIZER_INTERCEPT_STRCASESTR
 static inline void StrstrCheck(void *ctx, char *r, const char *s1,
                                const char *s2) {
-    uptr len1 = REAL(strlen)(s1);
-    uptr len2 = REAL(strlen)(s2);
+    uptr len1 = internal_strlen(s1);
+    uptr len2 = internal_strlen(s2);
     COMMON_INTERCEPTOR_READ_STRING(ctx, s1, r ? r - s1 + len2 : len1 + 1);
     COMMON_INTERCEPTOR_READ_RANGE(ctx, s2, len2 + 1);
 }
@@ -640,10 +640,10 @@ INTERCEPTOR(char*, strtok, char *str, const char *delimiters) {
     // for subsequent calls). We do not need to check strtok's result.
     // As the delimiters can change, we check them every call.
     if (str != nullptr) {
-      COMMON_INTERCEPTOR_READ_RANGE(ctx, str, REAL(strlen)(str) + 1);
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, str, internal_strlen(str) + 1);
     }
     COMMON_INTERCEPTOR_READ_RANGE(ctx, delimiters,
-                                  REAL(strlen)(delimiters) + 1);
+                                  internal_strlen(delimiters) + 1);
     return REAL(strtok)(str, delimiters);
   } else {
     // However, when strict_string_checks is disabled we cannot check the
@@ -657,11 +657,11 @@ INTERCEPTOR(char*, strtok, char *str, const char *delimiters) {
     COMMON_INTERCEPTOR_READ_RANGE(ctx, delimiters, 1);
     char *result = REAL(strtok)(str, delimiters);
     if (result != nullptr) {
-      COMMON_INTERCEPTOR_READ_RANGE(ctx, result, REAL(strlen)(result) + 1);
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, result, internal_strlen(result) + 1);
     } else if (str != nullptr) {
       // No delimiter were found, it's safe to assume that the entire str was
       // scanned.
-      COMMON_INTERCEPTOR_READ_RANGE(ctx, str, REAL(strlen)(str) + 1);
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, str, internal_strlen(str) + 1);
     }
     return result;
   }
@@ -706,7 +706,7 @@ INTERCEPTOR(char*, strchr, const char *s, int c) {
   if (common_flags()->intercept_strchr) {
     // Keep strlen as macro argument, as macro may ignore it.
     COMMON_INTERCEPTOR_READ_STRING(ctx, s,
-      (result ? result - s : REAL(strlen)(s)) + 1);
+      (result ? result - s : internal_strlen(s)) + 1);
   }
   return result;
 }
@@ -737,7 +737,7 @@ INTERCEPTOR(char*, strrchr, const char *s, int c) {
     return internal_strrchr(s, c);
   COMMON_INTERCEPTOR_ENTER(ctx, strrchr, s, c);
   if (common_flags()->intercept_strchr)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, s, REAL(strlen)(s) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, s, internal_strlen(s) + 1);
   return REAL(strrchr)(s, c);
 }
 #define INIT_STRRCHR COMMON_INTERCEPT_FUNCTION(strrchr)
@@ -751,7 +751,7 @@ INTERCEPTOR(SIZE_T, strspn, const char *s1, const char *s2) {
   COMMON_INTERCEPTOR_ENTER(ctx, strspn, s1, s2);
   SIZE_T r = REAL(strspn)(s1, s2);
   if (common_flags()->intercept_strspn) {
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, s2, REAL(strlen)(s2) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, s2, internal_strlen(s2) + 1);
     COMMON_INTERCEPTOR_READ_STRING(ctx, s1, r + 1);
   }
   return r;
@@ -762,7 +762,7 @@ INTERCEPTOR(SIZE_T, strcspn, const char *s1, const char *s2) {
   COMMON_INTERCEPTOR_ENTER(ctx, strcspn, s1, s2);
   SIZE_T r = REAL(strcspn)(s1, s2);
   if (common_flags()->intercept_strspn) {
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, s2, REAL(strlen)(s2) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, s2, internal_strlen(s2) + 1);
     COMMON_INTERCEPTOR_READ_STRING(ctx, s1, r + 1);
   }
   return r;
@@ -781,9 +781,9 @@ INTERCEPTOR(char *, strpbrk, const char *s1, const char *s2) {
   COMMON_INTERCEPTOR_ENTER(ctx, strpbrk, s1, s2);
   char *r = REAL(strpbrk)(s1, s2);
   if (common_flags()->intercept_strpbrk) {
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, s2, REAL(strlen)(s2) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, s2, internal_strlen(s2) + 1);
     COMMON_INTERCEPTOR_READ_STRING(ctx, s1,
-        r ? r - s1 + 1 : REAL(strlen)(s1) + 1);
+        r ? r - s1 + 1 : internal_strlen(s1) + 1);
   }
   return r;
 }
@@ -1251,7 +1251,7 @@ INTERCEPTOR(char *, fgets, char *s, SIZE_T size, void *file) {
   // https://github.com/google/sanitizers/issues/321.
   char *res = REAL(fgets)(s, size, file);
   if (res)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, s, REAL(strlen)(s) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, s, internal_strlen(s) + 1);
   return res;
 }
 #define INIT_FGETS COMMON_INTERCEPT_FUNCTION(fgets)
@@ -1265,7 +1265,7 @@ INTERCEPTOR_WITH_SUFFIX(int, fputs, char *s, void *file) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, fputs, s, file);
   if (!SANITIZER_MAC || s) {  // `fputs(NULL, file)` is supported on Darwin.
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, s, REAL(strlen)(s) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, s, internal_strlen(s) + 1);
   }
   return REAL(fputs)(s, file);
 }
@@ -1280,7 +1280,7 @@ INTERCEPTOR(int, puts, char *s) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, puts, s);
   if (!SANITIZER_MAC || s) {  // `puts(NULL)` is supported on Darwin.
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, s, REAL(strlen)(s) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, s, internal_strlen(s) + 1);
   }
   return REAL(puts)(s);
 }
@@ -1334,7 +1334,7 @@ static void unpoison_tm(void *ctx, __sanitizer_tm *tm) {
     // Can not use COMMON_INTERCEPTOR_WRITE_RANGE here, because tm->tm_zone
     // can point to shared memory and tsan would report a data race.
     COMMON_INTERCEPTOR_INITIALIZE_RANGE(tm->tm_zone,
-                                        REAL(strlen(tm->tm_zone)) + 1);
+                                        internal_strlen(tm->tm_zone) + 1);
   }
 #endif
 }
@@ -1387,7 +1387,7 @@ INTERCEPTOR(char *, ctime, unsigned long *timep) {
   char *res = REAL(ctime)(timep);
   if (res) {
     COMMON_INTERCEPTOR_READ_RANGE(ctx, timep, sizeof(*timep));
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   }
   return res;
 }
@@ -1400,7 +1400,7 @@ INTERCEPTOR(char *, ctime_r, unsigned long *timep, char *result) {
   char *res = REAL(ctime_r)(timep, result);
   if (res) {
     COMMON_INTERCEPTOR_READ_RANGE(ctx, timep, sizeof(*timep));
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   }
   return res;
 }
@@ -1413,7 +1413,7 @@ INTERCEPTOR(char *, asctime, __sanitizer_tm *tm) {
   char *res = REAL(asctime)(tm);
   if (res) {
     COMMON_INTERCEPTOR_READ_RANGE(ctx, tm, sizeof(*tm));
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   }
   return res;
 }
@@ -1426,7 +1426,7 @@ INTERCEPTOR(char *, asctime_r, __sanitizer_tm *tm, char *result) {
   char *res = REAL(asctime_r)(tm, result);
   if (res) {
     COMMON_INTERCEPTOR_READ_RANGE(ctx, tm, sizeof(*tm));
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   }
   return res;
 }
@@ -1463,7 +1463,7 @@ INTERCEPTOR(char *, strptime, char *s, char *format, __sanitizer_tm *tm) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, strptime, s, format, tm);
   if (format)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, format, REAL(strlen)(format) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, format, internal_strlen(format) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -1843,9 +1843,9 @@ INTERCEPTOR(int, ioctl, int d, unsigned long request, ...) {
   const ioctl_desc *desc = ioctl_lookup(request);
   ioctl_desc decoded_desc;
   if (!desc) {
-    VPrintf(2, "Decoding unknown ioctl 0x%x\n", request);
+    VPrintf(2, "Decoding unknown ioctl 0x%lx\n", request);
     if (!ioctl_decode(request, &decoded_desc))
-      Printf("WARNING: failed decoding unknown ioctl 0x%x\n", request);
+      Printf("WARNING: failed decoding unknown ioctl 0x%lx\n", request);
     else
       desc = &decoded_desc;
   }
@@ -1869,26 +1869,26 @@ UNUSED static void unpoison_passwd(void *ctx, __sanitizer_passwd *pwd) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, pwd, sizeof(*pwd));
     if (pwd->pw_name)
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, pwd->pw_name,
-                                     REAL(strlen)(pwd->pw_name) + 1);
+                                     internal_strlen(pwd->pw_name) + 1);
     if (pwd->pw_passwd)
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, pwd->pw_passwd,
-                                     REAL(strlen)(pwd->pw_passwd) + 1);
+                                     internal_strlen(pwd->pw_passwd) + 1);
 #if !SANITIZER_ANDROID
     if (pwd->pw_gecos)
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, pwd->pw_gecos,
-                                     REAL(strlen)(pwd->pw_gecos) + 1);
+                                     internal_strlen(pwd->pw_gecos) + 1);
 #endif
 #if SANITIZER_MAC || SANITIZER_FREEBSD || SANITIZER_NETBSD
     if (pwd->pw_class)
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, pwd->pw_class,
-                                     REAL(strlen)(pwd->pw_class) + 1);
+                                     internal_strlen(pwd->pw_class) + 1);
 #endif
     if (pwd->pw_dir)
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, pwd->pw_dir,
-                                     REAL(strlen)(pwd->pw_dir) + 1);
+                                     internal_strlen(pwd->pw_dir) + 1);
     if (pwd->pw_shell)
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, pwd->pw_shell,
-                                     REAL(strlen)(pwd->pw_shell) + 1);
+                                     internal_strlen(pwd->pw_shell) + 1);
   }
 }
 
@@ -1897,13 +1897,13 @@ UNUSED static void unpoison_group(void *ctx, __sanitizer_group *grp) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, grp, sizeof(*grp));
     if (grp->gr_name)
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, grp->gr_name,
-                                     REAL(strlen)(grp->gr_name) + 1);
+                                     internal_strlen(grp->gr_name) + 1);
     if (grp->gr_passwd)
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, grp->gr_passwd,
-                                     REAL(strlen)(grp->gr_passwd) + 1);
+                                     internal_strlen(grp->gr_passwd) + 1);
     char **p = grp->gr_mem;
     for (; *p; ++p) {
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *p, REAL(strlen)(*p) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *p, internal_strlen(*p) + 1);
     }
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, grp->gr_mem,
                                    (p - grp->gr_mem + 1) * sizeof(*p));
@@ -1916,7 +1916,7 @@ INTERCEPTOR(__sanitizer_passwd *, getpwnam, const char *name) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getpwnam, name);
   if (name)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   __sanitizer_passwd *res = REAL(getpwnam)(name);
   unpoison_passwd(ctx, res);
   return res;
@@ -1931,7 +1931,7 @@ INTERCEPTOR(__sanitizer_passwd *, getpwuid, u32 uid) {
 INTERCEPTOR(__sanitizer_group *, getgrnam, const char *name) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getgrnam, name);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   __sanitizer_group *res = REAL(getgrnam)(name);
   unpoison_group(ctx, res);
   return res;
@@ -1957,7 +1957,7 @@ INTERCEPTOR(int, getpwnam_r, const char *name, __sanitizer_passwd *pwd,
             char *buf, SIZE_T buflen, __sanitizer_passwd **result) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getpwnam_r, name, pwd, buf, buflen, result);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -1984,7 +1984,7 @@ INTERCEPTOR(int, getgrnam_r, const char *name, __sanitizer_group *grp,
             char *buf, SIZE_T buflen, __sanitizer_group **result) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getgrnam_r, name, grp, buf, buflen, result);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -2229,8 +2229,20 @@ INTERCEPTOR(int, clock_getcpuclockid, pid_t pid,
   return res;
 }
 
-#define INIT_CLOCK_GETCPUCLOCKID                  \
-  COMMON_INTERCEPT_FUNCTION(clock_getcpuclockid);
+INTERCEPTOR(int, pthread_getcpuclockid, uptr thread,
+            __sanitizer_clockid_t *clockid) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, pthread_getcpuclockid, thread, clockid);
+  int res = REAL(pthread_getcpuclockid)(thread, clockid);
+  if (!res && clockid) {
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, clockid, sizeof *clockid);
+  }
+  return res;
+}
+
+#define INIT_CLOCK_GETCPUCLOCKID                   \
+  COMMON_INTERCEPT_FUNCTION(clock_getcpuclockid);  \
+  COMMON_INTERCEPT_FUNCTION(pthread_getcpuclockid);
 #else
 #define INIT_CLOCK_GETCPUCLOCKID
 #endif
@@ -2289,7 +2301,7 @@ static void unpoison_glob_t(void *ctx, __sanitizer_glob_t *pglob) {
         ctx, pglob->gl_pathv, (pglob->gl_pathc + 1) * sizeof(*pglob->gl_pathv));
   for (SIZE_T i = 0; i < pglob->gl_pathc; ++i) {
     char *p = pglob->gl_pathv[i];
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p, REAL(strlen)(p) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p, internal_strlen(p) + 1);
   }
 }
 
@@ -2319,19 +2331,19 @@ static void *wrapped_gl_readdir(void *dir) {
 
 static void *wrapped_gl_opendir(const char *s) {
   COMMON_INTERCEPTOR_UNPOISON_PARAM(1);
-  COMMON_INTERCEPTOR_INITIALIZE_RANGE(s, REAL(strlen)(s) + 1);
+  COMMON_INTERCEPTOR_INITIALIZE_RANGE(s, internal_strlen(s) + 1);
   return pglob_copy->gl_opendir(s);
 }
 
 static int wrapped_gl_lstat(const char *s, void *st) {
   COMMON_INTERCEPTOR_UNPOISON_PARAM(2);
-  COMMON_INTERCEPTOR_INITIALIZE_RANGE(s, REAL(strlen)(s) + 1);
+  COMMON_INTERCEPTOR_INITIALIZE_RANGE(s, internal_strlen(s) + 1);
   return pglob_copy->gl_lstat(s, st);
 }
 
 static int wrapped_gl_stat(const char *s, void *st) {
   COMMON_INTERCEPTOR_UNPOISON_PARAM(2);
-  COMMON_INTERCEPTOR_INITIALIZE_RANGE(s, REAL(strlen)(s) + 1);
+  COMMON_INTERCEPTOR_INITIALIZE_RANGE(s, internal_strlen(s) + 1);
   return pglob_copy->gl_stat(s, st);
 }
 
@@ -2410,6 +2422,60 @@ INTERCEPTOR(int, glob64, const char *pattern, int flags,
 #define INIT_GLOB64
 #endif  // SANITIZER_INTERCEPT_GLOB64
 
+#if SANITIZER_INTERCEPT_POSIX_SPAWN
+
+template <class RealSpawnPtr>
+static int PosixSpawnImpl(void *ctx, RealSpawnPtr *real_posix_spawn, pid_t *pid,
+                          const char *file_or_path, const void *file_actions,
+                          const void *attrp, char *const argv[],
+                          char *const envp[]) {
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, file_or_path,
+                                internal_strlen(file_or_path) + 1);
+  if (argv) {
+    for (char *const *s = argv; ; ++s) {
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, s, sizeof(*s));
+      if (!*s) break;
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, *s, internal_strlen(*s) + 1);
+    }
+  }
+  if (envp) {
+    for (char *const *s = envp; ; ++s) {
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, s, sizeof(*s));
+      if (!*s) break;
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, *s, internal_strlen(*s) + 1);
+    }
+  }
+  int res =
+      real_posix_spawn(pid, file_or_path, file_actions, attrp, argv, envp);
+  if (res == 0)
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, pid, sizeof(*pid));
+  return res;
+}
+INTERCEPTOR(int, posix_spawn, pid_t *pid, const char *path,
+            const void *file_actions, const void *attrp, char *const argv[],
+            char *const envp[]) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, posix_spawn, pid, path, file_actions, attrp,
+                           argv, envp);
+  return PosixSpawnImpl(ctx, REAL(posix_spawn), pid, path, file_actions, attrp,
+                        argv, envp);
+}
+INTERCEPTOR(int, posix_spawnp, pid_t *pid, const char *file,
+            const void *file_actions, const void *attrp, char *const argv[],
+            char *const envp[]) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, posix_spawnp, pid, file, file_actions, attrp,
+                           argv, envp);
+  return PosixSpawnImpl(ctx, REAL(posix_spawnp), pid, file, file_actions, attrp,
+                        argv, envp);
+}
+#  define INIT_POSIX_SPAWN                  \
+    COMMON_INTERCEPT_FUNCTION(posix_spawn); \
+    COMMON_INTERCEPT_FUNCTION(posix_spawnp);
+#else  // SANITIZER_INTERCEPT_POSIX_SPAWN
+#  define INIT_POSIX_SPAWN
+#endif  // SANITIZER_INTERCEPT_POSIX_SPAWN
+
 #if SANITIZER_INTERCEPT_WAIT
 // According to sys/wait.h, wait(), waitid(), waitpid() may have symbol version
 // suffixes on Darwin. See the declaration of INTERCEPTOR_WITH_SUFFIX for
@@ -2519,7 +2585,7 @@ INTERCEPTOR(char *, inet_ntop, int af, const void *src, char *dst, u32 size) {
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
   char *res = REAL(inet_ntop)(af, src, dst, size);
-  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   return res;
 }
 INTERCEPTOR(int, inet_pton, int af, const char *src, void *dst) {
@@ -2548,7 +2614,7 @@ INTERCEPTOR(int, inet_pton, int af, const char *src, void *dst) {
 INTERCEPTOR(int, inet_aton, const char *cp, void *dst) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, inet_aton, cp, dst);
-  if (cp) COMMON_INTERCEPTOR_READ_RANGE(ctx, cp, REAL(strlen)(cp) + 1);
+  if (cp) COMMON_INTERCEPTOR_READ_RANGE(ctx, cp, internal_strlen(cp) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -2590,9 +2656,9 @@ INTERCEPTOR(int, getaddrinfo, char *node, char *service,
             struct __sanitizer_addrinfo **out) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getaddrinfo, node, service, hints, out);
-  if (node) COMMON_INTERCEPTOR_READ_RANGE(ctx, node, REAL(strlen)(node) + 1);
+  if (node) COMMON_INTERCEPTOR_READ_RANGE(ctx, node, internal_strlen(node) + 1);
   if (service)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, service, REAL(strlen)(service) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, service, internal_strlen(service) + 1);
   if (hints)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, hints, sizeof(__sanitizer_addrinfo));
   // FIXME: under ASan the call below may write to freed memory and corrupt
@@ -2608,7 +2674,7 @@ INTERCEPTOR(int, getaddrinfo, char *node, char *service,
         COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p->ai_addr, p->ai_addrlen);
       if (p->ai_canonname)
         COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p->ai_canonname,
-                                       REAL(strlen)(p->ai_canonname) + 1);
+                                       internal_strlen(p->ai_canonname) + 1);
       p = p->ai_next;
     }
   }
@@ -2634,9 +2700,9 @@ INTERCEPTOR(int, getnameinfo, void *sockaddr, unsigned salen, char *host,
       REAL(getnameinfo)(sockaddr, salen, host, hostlen, serv, servlen, flags);
   if (res == 0) {
     if (host && hostlen)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, host, REAL(strlen)(host) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, host, internal_strlen(host) + 1);
     if (serv && servlen)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, serv, REAL(strlen)(serv) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, serv, internal_strlen(serv) + 1);
   }
   return res;
 }
@@ -2646,17 +2712,20 @@ INTERCEPTOR(int, getnameinfo, void *sockaddr, unsigned salen, char *host,
 #endif
 
 #if SANITIZER_INTERCEPT_GETSOCKNAME
-INTERCEPTOR(int, getsockname, int sock_fd, void *addr, int *addrlen) {
+INTERCEPTOR(int, getsockname, int sock_fd, void *addr, unsigned *addrlen) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getsockname, sock_fd, addr, addrlen);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, addrlen, sizeof(*addrlen));
-  int addrlen_in = *addrlen;
+  unsigned addr_sz;
+  if (addrlen) {
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, addrlen, sizeof(*addrlen));
+    addr_sz = *addrlen;
+  }
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
   int res = REAL(getsockname)(sock_fd, addr, addrlen);
-  if (res == 0) {
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, addr, Min(addrlen_in, *addrlen));
+  if (!res && addr && addrlen) {
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, addr, Min(addr_sz, *addrlen));
   }
   return res;
 }
@@ -2669,10 +2738,10 @@ INTERCEPTOR(int, getsockname, int sock_fd, void *addr, int *addrlen) {
 static void write_hostent(void *ctx, struct __sanitizer_hostent *h) {
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, h, sizeof(__sanitizer_hostent));
   if (h->h_name)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, h->h_name, REAL(strlen)(h->h_name) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, h->h_name, internal_strlen(h->h_name) + 1);
   char **p = h->h_aliases;
   while (*p) {
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *p, REAL(strlen)(*p) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *p, internal_strlen(*p) + 1);
     ++p;
   }
   COMMON_INTERCEPTOR_WRITE_RANGE(
@@ -3161,13 +3230,17 @@ INTERCEPTOR(int, getpeername, int sockfd, void *addr, unsigned *addrlen) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getpeername, sockfd, addr, addrlen);
   unsigned addr_sz;
-  if (addrlen) addr_sz = *addrlen;
+  if (addrlen) {
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, addrlen, sizeof(*addrlen));
+    addr_sz = *addrlen;
+  }
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
   int res = REAL(getpeername)(sockfd, addr, addrlen);
-  if (!res && addr && addrlen)
+  if (!res && addr && addrlen) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, addr, Min(addr_sz, *addrlen));
+  }
   return res;
 }
 #define INIT_GETPEERNAME COMMON_INTERCEPT_FUNCTION(getpeername);
@@ -3196,7 +3269,7 @@ INTERCEPTOR(int, sysinfo, void *info) {
 INTERCEPTOR(__sanitizer_dirent *, opendir, const char *path) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, opendir, path);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   __sanitizer_dirent *res = REAL(opendir)(path);
   if (res)
     COMMON_INTERCEPTOR_DIR_ACQUIRE(ctx, path);
@@ -3351,10 +3424,10 @@ INTERCEPTOR(char *, setlocale, int category, char *locale) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, setlocale, category, locale);
   if (locale)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, locale, REAL(strlen)(locale) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, locale, internal_strlen(locale) + 1);
   char *res = REAL(setlocale)(category, locale);
   if (res) {
-    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
     unpoison_ctype_arrays(ctx);
   }
   return res;
@@ -3373,7 +3446,7 @@ INTERCEPTOR(char *, getcwd, char *buf, SIZE_T size) {
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
   char *res = REAL(getcwd)(buf, size);
-  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   return res;
 }
 #define INIT_GETCWD COMMON_INTERCEPT_FUNCTION(getcwd);
@@ -3389,7 +3462,7 @@ INTERCEPTOR(char *, get_current_dir_name, int fake) {
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
   char *res = REAL(get_current_dir_name)(fake);
-  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   return res;
 }
 
@@ -3663,7 +3736,7 @@ INTERCEPTOR(int, tcgetattr, int fd, void *termios_p) {
 INTERCEPTOR(char *, realpath, const char *path, char *resolved_path) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, realpath, path, resolved_path);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
 
   // Workaround a bug in glibc where dlsym(RTLD_NEXT, ...) returns the oldest
   // version of a versioned symbol. For realpath(), this gives us something
@@ -3674,11 +3747,12 @@ INTERCEPTOR(char *, realpath, const char *path, char *resolved_path) {
     allocated_path = resolved_path = (char *)WRAP(malloc)(path_max + 1);
 
   char *res = REAL(realpath)(path, resolved_path);
-  if (allocated_path && !res) WRAP(free)(allocated_path);
-  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+  if (allocated_path && !res)
+    WRAP(free)(allocated_path);
+  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   return res;
 }
-#define INIT_REALPATH COMMON_INTERCEPT_FUNCTION(realpath);
+#  define INIT_REALPATH COMMON_INTERCEPT_FUNCTION(realpath);
 #else
 #define INIT_REALPATH
 #endif
@@ -3687,9 +3761,9 @@ INTERCEPTOR(char *, realpath, const char *path, char *resolved_path) {
 INTERCEPTOR(char *, canonicalize_file_name, const char *path) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, canonicalize_file_name, path);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   char *res = REAL(canonicalize_file_name)(path);
-  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   return res;
 }
 #define INIT_CANONICALIZE_FILE_NAME \
@@ -3750,7 +3824,7 @@ INTERCEPTOR(char *, strerror, int errnum) {
   COMMON_INTERCEPTOR_ENTER(ctx, strerror, errnum);
   COMMON_INTERCEPTOR_STRERROR();
   char *res = REAL(strerror)(errnum);
-  if (res) COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+  if (res) COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
   return res;
 }
 #define INIT_STRERROR COMMON_INTERCEPT_FUNCTION(strerror);
@@ -3792,9 +3866,9 @@ INTERCEPTOR(char *, strerror_r, int errnum, char *buf, SIZE_T buflen) {
   // https://github.com/google/sanitizers/issues/321.
   char *res = REAL(strerror_r)(errnum, buf, buflen);
   if (res == buf)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   else
-    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
   return res;
 }
 #endif //(_POSIX_C_SOURCE >= 200112L || _XOPEN_SOURCE >= 600) && !_GNU_SOURCE ||
@@ -3814,7 +3888,7 @@ INTERCEPTOR(int, __xpg_strerror_r, int errnum, char *buf, SIZE_T buflen) {
   int res = REAL(__xpg_strerror_r)(errnum, buf, buflen);
   // This version always returns a null-terminated string.
   if (buf && buflen)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, REAL(strlen)(buf) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, internal_strlen(buf) + 1);
   return res;
 }
 #define INIT_XPG_STRERROR_R COMMON_INTERCEPT_FUNCTION(__xpg_strerror_r);
@@ -3850,7 +3924,7 @@ INTERCEPTOR(int, scandir, char *dirp, __sanitizer_dirent ***namelist,
             scandir_filter_f filter, scandir_compar_f compar) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, scandir, dirp, namelist, filter, compar);
-  if (dirp) COMMON_INTERCEPTOR_READ_RANGE(ctx, dirp, REAL(strlen)(dirp) + 1);
+  if (dirp) COMMON_INTERCEPTOR_READ_RANGE(ctx, dirp, internal_strlen(dirp) + 1);
   scandir_filter = filter;
   scandir_compar = compar;
   // FIXME: under ASan the call below may write to freed memory and corrupt
@@ -3903,7 +3977,7 @@ INTERCEPTOR(int, scandir64, char *dirp, __sanitizer_dirent64 ***namelist,
             scandir64_filter_f filter, scandir64_compar_f compar) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, scandir64, dirp, namelist, filter, compar);
-  if (dirp) COMMON_INTERCEPTOR_READ_RANGE(ctx, dirp, REAL(strlen)(dirp) + 1);
+  if (dirp) COMMON_INTERCEPTOR_READ_RANGE(ctx, dirp, internal_strlen(dirp) + 1);
   scandir64_filter = filter;
   scandir64_compar = compar;
   // FIXME: under ASan the call below may write to freed memory and corrupt
@@ -3999,19 +4073,20 @@ INTERCEPTOR(int, ppoll, __sanitizer_pollfd *fds, __sanitizer_nfds_t nfds,
 INTERCEPTOR(int, wordexp, char *s, __sanitizer_wordexp_t *p, int flags) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, wordexp, s, p, flags);
-  if (s) COMMON_INTERCEPTOR_READ_RANGE(ctx, s, REAL(strlen)(s) + 1);
+  if (s) COMMON_INTERCEPTOR_READ_RANGE(ctx, s, internal_strlen(s) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
   int res = REAL(wordexp)(s, p, flags);
   if (!res && p) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p, sizeof(*p));
-    if (p->we_wordc)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p->we_wordv,
-                                     sizeof(*p->we_wordv) * p->we_wordc);
-    for (uptr i = 0; i < p->we_wordc; ++i) {
+    uptr we_wordc =
+        ((flags & wordexp_wrde_dooffs) ? p->we_offs : 0) + p->we_wordc;
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p->we_wordv,
+                                   sizeof(*p->we_wordv) * (we_wordc + 1));
+    for (uptr i = 0; i < we_wordc; ++i) {
       char *w = p->we_wordv[i];
-      if (w) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, w, REAL(strlen)(w) + 1);
+      if (w) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, w, internal_strlen(w) + 1);
     }
   }
   return res;
@@ -4217,7 +4292,7 @@ INTERCEPTOR(char **, backtrace_symbols, void **buffer, int size) {
   if (res && size) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, size * sizeof(*res));
     for (int i = 0; i < size; ++i)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res[i], REAL(strlen(res[i])) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res[i], internal_strlen(res[i]) + 1);
   }
   return res;
 }
@@ -4335,16 +4410,16 @@ static void write_mntent(void *ctx, __sanitizer_mntent *mnt) {
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, mnt, sizeof(*mnt));
   if (mnt->mnt_fsname)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, mnt->mnt_fsname,
-                                   REAL(strlen)(mnt->mnt_fsname) + 1);
+                                   internal_strlen(mnt->mnt_fsname) + 1);
   if (mnt->mnt_dir)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, mnt->mnt_dir,
-                                   REAL(strlen)(mnt->mnt_dir) + 1);
+                                   internal_strlen(mnt->mnt_dir) + 1);
   if (mnt->mnt_type)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, mnt->mnt_type,
-                                   REAL(strlen)(mnt->mnt_type) + 1);
+                                   internal_strlen(mnt->mnt_type) + 1);
   if (mnt->mnt_opts)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, mnt->mnt_opts,
-                                   REAL(strlen)(mnt->mnt_opts) + 1);
+                                   internal_strlen(mnt->mnt_opts) + 1);
 }
 #endif
 
@@ -4379,7 +4454,7 @@ INTERCEPTOR(__sanitizer_mntent *, getmntent_r, void *fp,
 INTERCEPTOR(int, statfs, char *path, void *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, statfs, path, buf);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -4408,7 +4483,7 @@ INTERCEPTOR(int, fstatfs, int fd, void *buf) {
 INTERCEPTOR(int, statfs64, char *path, void *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, statfs64, path, buf);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -4437,7 +4512,7 @@ INTERCEPTOR(int, fstatfs64, int fd, void *buf) {
 INTERCEPTOR(int, statvfs, char *path, void *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, statvfs, path, buf);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -4471,7 +4546,7 @@ INTERCEPTOR(int, fstatvfs, int fd, void *buf) {
 INTERCEPTOR(int, statvfs64, char *path, void *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, statvfs64, path, buf);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -4500,7 +4575,7 @@ INTERCEPTOR(int, fstatvfs64, int fd, void *buf) {
 INTERCEPTOR(int, initgroups, char *user, u32 group) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, initgroups, user, group);
-  if (user) COMMON_INTERCEPTOR_READ_RANGE(ctx, user, REAL(strlen)(user) + 1);
+  if (user) COMMON_INTERCEPTOR_READ_RANGE(ctx, user, internal_strlen(user) + 1);
   int res = REAL(initgroups)(user, group);
   return res;
 }
@@ -4515,13 +4590,13 @@ INTERCEPTOR(char *, ether_ntoa, __sanitizer_ether_addr *addr) {
   COMMON_INTERCEPTOR_ENTER(ctx, ether_ntoa, addr);
   if (addr) COMMON_INTERCEPTOR_READ_RANGE(ctx, addr, sizeof(*addr));
   char *res = REAL(ether_ntoa)(addr);
-  if (res) COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+  if (res) COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
   return res;
 }
 INTERCEPTOR(__sanitizer_ether_addr *, ether_aton, char *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, ether_aton, buf);
-  if (buf) COMMON_INTERCEPTOR_READ_RANGE(ctx, buf, REAL(strlen)(buf) + 1);
+  if (buf) COMMON_INTERCEPTOR_READ_RANGE(ctx, buf, internal_strlen(buf) + 1);
   __sanitizer_ether_addr *res = REAL(ether_aton)(buf);
   if (res) COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, sizeof(*res));
   return res;
@@ -4543,14 +4618,14 @@ INTERCEPTOR(int, ether_ntohost, char *hostname, __sanitizer_ether_addr *addr) {
   // https://github.com/google/sanitizers/issues/321.
   int res = REAL(ether_ntohost)(hostname, addr);
   if (!res && hostname)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, hostname, REAL(strlen)(hostname) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, hostname, internal_strlen(hostname) + 1);
   return res;
 }
 INTERCEPTOR(int, ether_hostton, char *hostname, __sanitizer_ether_addr *addr) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, ether_hostton, hostname, addr);
   if (hostname)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, hostname, REAL(strlen)(hostname) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, hostname, internal_strlen(hostname) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -4562,7 +4637,7 @@ INTERCEPTOR(int, ether_line, char *line, __sanitizer_ether_addr *addr,
             char *hostname) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, ether_line, line, addr, hostname);
-  if (line) COMMON_INTERCEPTOR_READ_RANGE(ctx, line, REAL(strlen)(line) + 1);
+  if (line) COMMON_INTERCEPTOR_READ_RANGE(ctx, line, internal_strlen(line) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -4570,7 +4645,7 @@ INTERCEPTOR(int, ether_line, char *line, __sanitizer_ether_addr *addr,
   if (!res) {
     if (addr) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, addr, sizeof(*addr));
     if (hostname)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, hostname, REAL(strlen)(hostname) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, hostname, internal_strlen(hostname) + 1);
   }
   return res;
 }
@@ -4591,14 +4666,14 @@ INTERCEPTOR(char *, ether_ntoa_r, __sanitizer_ether_addr *addr, char *buf) {
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
   char *res = REAL(ether_ntoa_r)(addr, buf);
-  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+  if (res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   return res;
 }
 INTERCEPTOR(__sanitizer_ether_addr *, ether_aton_r, char *buf,
             __sanitizer_ether_addr *addr) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, ether_aton_r, buf, addr);
-  if (buf) COMMON_INTERCEPTOR_READ_RANGE(ctx, buf, REAL(strlen)(buf) + 1);
+  if (buf) COMMON_INTERCEPTOR_READ_RANGE(ctx, buf, internal_strlen(buf) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -4864,9 +4939,9 @@ INTERCEPTOR(char *, tmpnam, char *s) {
       // FIXME: under ASan the call below may write to freed memory and corrupt
       // its metadata. See
       // https://github.com/google/sanitizers/issues/321.
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, s, REAL(strlen)(s) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, s, internal_strlen(s) + 1);
     else
-      COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+      COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
   }
   return res;
 }
@@ -4883,7 +4958,7 @@ INTERCEPTOR(char *, tmpnam_r, char *s) {
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
   char *res = REAL(tmpnam_r)(s);
-  if (res && s) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, s, REAL(strlen)(s) + 1);
+  if (res && s) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, s, internal_strlen(s) + 1);
   return res;
 }
 #define INIT_TMPNAM_R COMMON_INTERCEPT_FUNCTION(tmpnam_r);
@@ -4897,7 +4972,7 @@ INTERCEPTOR(char *, ptsname, int fd) {
   COMMON_INTERCEPTOR_ENTER(ctx, ptsname, fd);
   char *res = REAL(ptsname)(fd);
   if (res != nullptr)
-    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
   return res;
 }
 #define INIT_PTSNAME COMMON_INTERCEPT_FUNCTION(ptsname);
@@ -4911,7 +4986,7 @@ INTERCEPTOR(int, ptsname_r, int fd, char *name, SIZE_T namesize) {
   COMMON_INTERCEPTOR_ENTER(ctx, ptsname_r, fd, name, namesize);
   int res = REAL(ptsname_r)(fd, name, namesize);
   if (res == 0)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, name, internal_strlen(name) + 1);
   return res;
 }
 #define INIT_PTSNAME_R COMMON_INTERCEPT_FUNCTION(ptsname_r);
@@ -4925,7 +5000,7 @@ INTERCEPTOR(char *, ttyname, int fd) {
   COMMON_INTERCEPTOR_ENTER(ctx, ttyname, fd);
   char *res = REAL(ttyname)(fd);
   if (res != nullptr)
-    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
   return res;
 }
 #define INIT_TTYNAME COMMON_INTERCEPT_FUNCTION(ttyname);
@@ -4939,7 +5014,7 @@ INTERCEPTOR(int, ttyname_r, int fd, char *name, SIZE_T namesize) {
   COMMON_INTERCEPTOR_ENTER(ctx, ttyname_r, fd, name, namesize);
   int res = REAL(ttyname_r)(fd, name, namesize);
   if (res == 0)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, name, internal_strlen(name) + 1);
   return res;
 }
 #define INIT_TTYNAME_R COMMON_INTERCEPT_FUNCTION(ttyname_r);
@@ -4951,10 +5026,10 @@ INTERCEPTOR(int, ttyname_r, int fd, char *name, SIZE_T namesize) {
 INTERCEPTOR(char *, tempnam, char *dir, char *pfx) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, tempnam, dir, pfx);
-  if (dir) COMMON_INTERCEPTOR_READ_RANGE(ctx, dir, REAL(strlen)(dir) + 1);
-  if (pfx) COMMON_INTERCEPTOR_READ_RANGE(ctx, pfx, REAL(strlen)(pfx) + 1);
+  if (dir) COMMON_INTERCEPTOR_READ_RANGE(ctx, dir, internal_strlen(dir) + 1);
+  if (pfx) COMMON_INTERCEPTOR_READ_RANGE(ctx, pfx, internal_strlen(pfx) + 1);
   char *res = REAL(tempnam)(dir, pfx);
-  if (res) COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+  if (res) COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
   return res;
 }
 #define INIT_TEMPNAM COMMON_INTERCEPT_FUNCTION(tempnam);
@@ -5414,7 +5489,7 @@ asm(
 INTERCEPTOR(SSIZE_T, listxattr, const char *path, char *list, SIZE_T size) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, listxattr, path, list, size);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -5427,7 +5502,7 @@ INTERCEPTOR(SSIZE_T, listxattr, const char *path, char *list, SIZE_T size) {
 INTERCEPTOR(SSIZE_T, llistxattr, const char *path, char *list, SIZE_T size) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, llistxattr, path, list, size);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -5458,8 +5533,8 @@ INTERCEPTOR(SSIZE_T, getxattr, const char *path, const char *name, char *value,
             SIZE_T size) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getxattr, path, name, value, size);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
-  if (name) COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
+  if (name) COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -5471,8 +5546,8 @@ INTERCEPTOR(SSIZE_T, lgetxattr, const char *path, const char *name, char *value,
             SIZE_T size) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, lgetxattr, path, name, value, size);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
-  if (name) COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
+  if (name) COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -5484,7 +5559,7 @@ INTERCEPTOR(SSIZE_T, fgetxattr, int fd, const char *name, char *value,
             SIZE_T size) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, fgetxattr, fd, name, value, size);
-  if (name) COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+  if (name) COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -5554,7 +5629,7 @@ INTERCEPTOR(int, getifaddrs, __sanitizer_ifaddrs **ifap) {
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p, sizeof(__sanitizer_ifaddrs));
       if (p->ifa_name)
         COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p->ifa_name,
-                                       REAL(strlen)(p->ifa_name) + 1);
+                                       internal_strlen(p->ifa_name) + 1);
       if (p->ifa_addr)
         COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p->ifa_addr, struct_sockaddr_sz);
       if (p->ifa_netmask)
@@ -5584,14 +5659,14 @@ INTERCEPTOR(char *, if_indextoname, unsigned int ifindex, char* ifname) {
   // https://github.com/google/sanitizers/issues/321.
   char *res = REAL(if_indextoname)(ifindex, ifname);
   if (res && ifname)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ifname, REAL(strlen)(ifname) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ifname, internal_strlen(ifname) + 1);
   return res;
 }
 INTERCEPTOR(unsigned int, if_nametoindex, const char* ifname) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, if_nametoindex, ifname);
   if (ifname)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, ifname, REAL(strlen)(ifname) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, ifname, internal_strlen(ifname) + 1);
   return REAL(if_nametoindex)(ifname);
 }
 #define INIT_IF_INDEXTONAME                  \
@@ -5849,7 +5924,7 @@ INTERCEPTOR(int, xdr_string, __sanitizer_XDR *xdrs, char **p,
   COMMON_INTERCEPTOR_ENTER(ctx, xdr_string, xdrs, p, maxsize);
   if (p && xdrs->x_op == __sanitizer_XDR_ENCODE) {
     COMMON_INTERCEPTOR_READ_RANGE(ctx, p, sizeof(*p));
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, *p, REAL(strlen)(*p) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, *p, internal_strlen(*p) + 1);
   }
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
@@ -5858,7 +5933,7 @@ INTERCEPTOR(int, xdr_string, __sanitizer_XDR *xdrs, char **p,
   if (p && xdrs->x_op == __sanitizer_XDR_DECODE) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p, sizeof(*p));
     if (res && *p)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *p, REAL(strlen)(*p) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *p, internal_strlen(*p) + 1);
   }
   return res;
 }
@@ -6069,8 +6144,8 @@ INTERCEPTOR(int, __woverflow, __sanitizer_FILE *fp, int ch) {
 INTERCEPTOR(__sanitizer_FILE *, fopen, const char *path, const char *mode) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, fopen, path, mode);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, REAL(strlen)(mode) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, internal_strlen(mode) + 1);
   __sanitizer_FILE *res = REAL(fopen)(path, mode);
   COMMON_INTERCEPTOR_FILE_OPEN(ctx, res, path);
   if (res) unpoison_file(res);
@@ -6079,7 +6154,7 @@ INTERCEPTOR(__sanitizer_FILE *, fopen, const char *path, const char *mode) {
 INTERCEPTOR(__sanitizer_FILE *, fdopen, int fd, const char *mode) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, fdopen, fd, mode);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, REAL(strlen)(mode) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, internal_strlen(mode) + 1);
   __sanitizer_FILE *res = REAL(fdopen)(fd, mode);
   if (res) unpoison_file(res);
   return res;
@@ -6088,8 +6163,8 @@ INTERCEPTOR(__sanitizer_FILE *, freopen, const char *path, const char *mode,
             __sanitizer_FILE *fp) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, freopen, path, mode, fp);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, REAL(strlen)(mode) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, internal_strlen(mode) + 1);
   COMMON_INTERCEPTOR_FILE_CLOSE(ctx, fp);
   __sanitizer_FILE *res = REAL(freopen)(path, mode, fp);
   COMMON_INTERCEPTOR_FILE_OPEN(ctx, res, path);
@@ -6113,7 +6188,7 @@ INTERCEPTOR(int, flopen, const char *path, int flags, ...) {
   va_end(ap);
   COMMON_INTERCEPTOR_ENTER(ctx, flopen, path, flags, mode);
   if (path) {
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   }
   return REAL(flopen)(path, flags, mode);
 }
@@ -6126,7 +6201,7 @@ INTERCEPTOR(int, flopenat, int dirfd, const char *path, int flags, ...) {
   va_end(ap);
   COMMON_INTERCEPTOR_ENTER(ctx, flopen, path, flags, mode);
   if (path) {
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   }
   return REAL(flopenat)(dirfd, path, flags, mode);
 }
@@ -6142,8 +6217,8 @@ INTERCEPTOR(int, flopenat, int dirfd, const char *path, int flags, ...) {
 INTERCEPTOR(__sanitizer_FILE *, fopen64, const char *path, const char *mode) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, fopen64, path, mode);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, REAL(strlen)(mode) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, internal_strlen(mode) + 1);
   __sanitizer_FILE *res = REAL(fopen64)(path, mode);
   COMMON_INTERCEPTOR_FILE_OPEN(ctx, res, path);
   if (res) unpoison_file(res);
@@ -6153,8 +6228,8 @@ INTERCEPTOR(__sanitizer_FILE *, freopen64, const char *path, const char *mode,
             __sanitizer_FILE *fp) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, freopen64, path, mode, fp);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, REAL(strlen)(mode) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, mode, internal_strlen(mode) + 1);
   COMMON_INTERCEPTOR_FILE_CLOSE(ctx, fp);
   __sanitizer_FILE *res = REAL(freopen64)(path, mode, fp);
   COMMON_INTERCEPTOR_FILE_OPEN(ctx, res, path);
@@ -6332,9 +6407,9 @@ INTERCEPTOR(char *, getpass, const char *prompt) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getpass, prompt);
   if (prompt)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, prompt, REAL(strlen)(prompt)+1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, prompt, internal_strlen(prompt)+1);
   char *res = REAL(getpass)(prompt);
-  if (res) COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res)+1);
+  if (res) COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res)+1);
   return res;
 }
 
@@ -6538,17 +6613,42 @@ INTERCEPTOR(int, sem_getvalue, __sanitizer_sem_t *s, int *sval) {
   }
   return res;
 }
-#define INIT_SEM                                                               \
-  COMMON_INTERCEPT_FUNCTION(sem_init);                                         \
-  COMMON_INTERCEPT_FUNCTION(sem_destroy);                                      \
-  COMMON_INTERCEPT_FUNCTION(sem_wait);                                         \
-  COMMON_INTERCEPT_FUNCTION(sem_trywait);                                      \
-  COMMON_INTERCEPT_FUNCTION(sem_timedwait);                                    \
-  COMMON_INTERCEPT_FUNCTION(sem_post);                                         \
-  COMMON_INTERCEPT_FUNCTION(sem_getvalue);
+
+INTERCEPTOR(__sanitizer_sem_t *, sem_open, const char *name, int oflag, ...) {
+  void *ctx;
+  va_list ap;
+  va_start(ap, oflag);
+  u32 mode = va_arg(ap, u32);
+  u32 value = va_arg(ap, u32);
+  COMMON_INTERCEPTOR_ENTER(ctx, sem_open, name, oflag, mode, value);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
+  __sanitizer_sem_t *s = REAL(sem_open)(name, oflag, mode, value);
+  if (s)
+    COMMON_INTERCEPTOR_INITIALIZE_RANGE(s, sizeof(*s));
+  va_end(ap);
+  return s;
+}
+
+INTERCEPTOR(int, sem_unlink, const char *name) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, sem_unlink, name);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
+  return REAL(sem_unlink)(name);
+}
+
+#  define INIT_SEM                            \
+    COMMON_INTERCEPT_FUNCTION(sem_init);      \
+    COMMON_INTERCEPT_FUNCTION(sem_destroy);   \
+    COMMON_INTERCEPT_FUNCTION(sem_wait);      \
+    COMMON_INTERCEPT_FUNCTION(sem_trywait);   \
+    COMMON_INTERCEPT_FUNCTION(sem_timedwait); \
+    COMMON_INTERCEPT_FUNCTION(sem_post);      \
+    COMMON_INTERCEPT_FUNCTION(sem_getvalue);  \
+    COMMON_INTERCEPT_FUNCTION(sem_open);      \
+    COMMON_INTERCEPT_FUNCTION(sem_unlink);
 #else
-#define INIT_SEM
-#endif // SANITIZER_INTERCEPT_SEM
+#  define INIT_SEM
+#endif  // SANITIZER_INTERCEPT_SEM
 
 #if SANITIZER_INTERCEPT_PTHREAD_SETCANCEL
 INTERCEPTOR(int, pthread_setcancelstate, int state, int *oldstate) {
@@ -6631,7 +6731,7 @@ INTERCEPTOR(char *, ctermid, char *s) {
   COMMON_INTERCEPTOR_ENTER(ctx, ctermid, s);
   char *res = REAL(ctermid)(s);
   if (res) {
-    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
   }
   return res;
 }
@@ -6646,7 +6746,7 @@ INTERCEPTOR(char *, ctermid_r, char *s) {
   COMMON_INTERCEPTOR_ENTER(ctx, ctermid_r, s);
   char *res = REAL(ctermid_r)(s);
   if (res) {
-    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_INITIALIZE_RANGE(res, internal_strlen(res) + 1);
   }
   return res;
 }
@@ -6983,8 +7083,8 @@ INTERCEPTOR(SIZE_T, wcsnlen, const wchar_t *s, SIZE_T n) {
 INTERCEPTOR(wchar_t *, wcscat, wchar_t *dst, const wchar_t *src) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, wcscat, dst, src);
-  SIZE_T src_size = REAL(wcslen)(src);
-  SIZE_T dst_size = REAL(wcslen)(dst);
+  SIZE_T src_size = internal_wcslen(src);
+  SIZE_T dst_size = internal_wcslen(dst);
   COMMON_INTERCEPTOR_READ_RANGE(ctx, src, (src_size + 1) * sizeof(wchar_t));
   COMMON_INTERCEPTOR_READ_RANGE(ctx, dst, (dst_size + 1) * sizeof(wchar_t));
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst + dst_size,
@@ -6995,8 +7095,8 @@ INTERCEPTOR(wchar_t *, wcscat, wchar_t *dst, const wchar_t *src) {
 INTERCEPTOR(wchar_t *, wcsncat, wchar_t *dst, const wchar_t *src, SIZE_T n) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, wcsncat, dst, src, n);
-  SIZE_T src_size = REAL(wcsnlen)(src, n);
-  SIZE_T dst_size = REAL(wcslen)(dst);
+  SIZE_T src_size = internal_wcsnlen(src, n);
+  SIZE_T dst_size = internal_wcslen(dst);
   COMMON_INTERCEPTOR_READ_RANGE(ctx, src,
                                 Min(src_size + 1, n) * sizeof(wchar_t));
   COMMON_INTERCEPTOR_READ_RANGE(ctx, dst, (dst_size + 1) * sizeof(wchar_t));
@@ -7015,7 +7115,7 @@ INTERCEPTOR(wchar_t *, wcsncat, wchar_t *dst, const wchar_t *src, SIZE_T n) {
 INTERCEPTOR(wchar_t *, wcsdup, wchar_t *s) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, wcsdup, s);
-  SIZE_T len = REAL(wcslen)(s);
+  SIZE_T len = internal_wcslen(s);
   COMMON_INTERCEPTOR_READ_RANGE(ctx, s, sizeof(wchar_t) * (len + 1));
   wchar_t *result = REAL(wcsdup)(s);
   if (result)
@@ -7029,9 +7129,9 @@ INTERCEPTOR(wchar_t *, wcsdup, wchar_t *s) {
 #endif
 
 #if SANITIZER_INTERCEPT_STRXFRM
-static SIZE_T RealStrLen(const char *str) { return REAL(strlen)(str); }
+static SIZE_T RealStrLen(const char *str) { return internal_strlen(str); }
 
-static SIZE_T RealStrLen(const wchar_t *str) { return REAL(wcslen)(str); }
+static SIZE_T RealStrLen(const wchar_t *str) { return internal_wcslen(str); }
 
 #define STRXFRM_INTERCEPTOR_IMPL(strxfrm, dest, src, len, ...)             \
   {                                                                        \
@@ -7105,7 +7205,7 @@ INTERCEPTOR(int, acct, const char *file) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, acct, file);
   if (file)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, file, REAL(strlen)(file) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, file, internal_strlen(file) + 1);
   return REAL(acct)(file);
 }
 #define INIT_ACCT COMMON_INTERCEPT_FUNCTION(acct)
@@ -7120,7 +7220,7 @@ INTERCEPTOR(const char *, user_from_uid, u32 uid, int nouser) {
   COMMON_INTERCEPTOR_ENTER(ctx, user_from_uid, uid, nouser);
   user = REAL(user_from_uid)(uid, nouser);
   if (user)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, user, REAL(strlen)(user) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, user, internal_strlen(user) + 1);
   return user;
 }
 #define INIT_USER_FROM_UID COMMON_INTERCEPT_FUNCTION(user_from_uid)
@@ -7134,7 +7234,7 @@ INTERCEPTOR(int, uid_from_user, const char *name, u32 *uid) {
   int res;
   COMMON_INTERCEPTOR_ENTER(ctx, uid_from_user, name, uid);
   if (name)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   res = REAL(uid_from_user)(name, uid);
   if (uid)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, uid, sizeof(*uid));
@@ -7152,7 +7252,7 @@ INTERCEPTOR(const char *, group_from_gid, u32 gid, int nogroup) {
   COMMON_INTERCEPTOR_ENTER(ctx, group_from_gid, gid, nogroup);
   group = REAL(group_from_gid)(gid, nogroup);
   if (group)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, group, REAL(strlen)(group) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, group, internal_strlen(group) + 1);
   return group;
 }
 #define INIT_GROUP_FROM_GID COMMON_INTERCEPT_FUNCTION(group_from_gid)
@@ -7166,7 +7266,7 @@ INTERCEPTOR(int, gid_from_group, const char *group, u32 *gid) {
   int res;
   COMMON_INTERCEPTOR_ENTER(ctx, gid_from_group, group, gid);
   if (group)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, group, REAL(strlen)(group) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, group, internal_strlen(group) + 1);
   res = REAL(gid_from_group)(group, gid);
   if (gid)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, gid, sizeof(*gid));
@@ -7182,7 +7282,7 @@ INTERCEPTOR(int, access, const char *path, int mode) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, access, path, mode);
   if (path)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   return REAL(access)(path, mode);
 }
 #define INIT_ACCESS COMMON_INTERCEPT_FUNCTION(access)
@@ -7195,7 +7295,7 @@ INTERCEPTOR(int, faccessat, int fd, const char *path, int mode, int flags) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, faccessat, fd, path, mode, flags);
   if (path)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   return REAL(faccessat)(fd, path, mode, flags);
 }
 #define INIT_FACCESSAT COMMON_INTERCEPT_FUNCTION(faccessat)
@@ -7210,7 +7310,7 @@ INTERCEPTOR(int, getgrouplist, const char *name, u32 basegid, u32 *groups,
   int res;
   COMMON_INTERCEPTOR_ENTER(ctx, getgrouplist, name, basegid, groups, ngroups);
   if (name)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   if (ngroups)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, ngroups, sizeof(*ngroups));
   res = REAL(getgrouplist)(name, basegid, groups, ngroups);
@@ -7234,7 +7334,7 @@ INTERCEPTOR(int, getgroupmembership, const char *name, u32 basegid, u32 *groups,
   COMMON_INTERCEPTOR_ENTER(ctx, getgroupmembership, name, basegid, groups,
                            maxgrp, ngroups);
   if (name)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   res = REAL(getgroupmembership)(name, basegid, groups, maxgrp, ngroups);
   if (!res && groups && ngroups) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, groups, sizeof(*groups) * (*ngroups));
@@ -7252,7 +7352,7 @@ INTERCEPTOR(int, getgroupmembership, const char *name, u32 basegid, u32 *groups,
 INTERCEPTOR(SSIZE_T, readlink, const char *path, char *buf, SIZE_T bufsiz) {
   void* ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, readlink, path, buf, bufsiz);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   SSIZE_T res = REAL(readlink)(path, buf, bufsiz);
   if (res > 0)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, res);
@@ -7269,7 +7369,7 @@ INTERCEPTOR(SSIZE_T, readlinkat, int dirfd, const char *path, char *buf,
             SIZE_T bufsiz) {
   void* ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, readlinkat, dirfd, path, buf, bufsiz);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   SSIZE_T res = REAL(readlinkat)(dirfd, path, buf, bufsiz);
   if (res > 0)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, res);
@@ -7287,7 +7387,7 @@ INTERCEPTOR(int, name_to_handle_at, int dirfd, const char *pathname,
   void* ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, name_to_handle_at, dirfd, pathname, handle,
                            mount_id, flags);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, pathname, REAL(strlen)(pathname) + 1);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, pathname, internal_strlen(pathname) + 1);
 
   __sanitizer_file_handle *sanitizer_handle =
       reinterpret_cast<__sanitizer_file_handle*>(handle);
@@ -7351,7 +7451,7 @@ INTERCEPTOR(SIZE_T, strlcpy, char *dst, char *src, SIZE_T size) {
         ctx, src, Min(internal_strnlen(src, size), size - 1) + 1);
   }
   res = REAL(strlcpy)(dst, src, size);
-  COMMON_INTERCEPTOR_COPY_STRING(ctx, dst, src, REAL(strlen)(dst) + 1);
+  COMMON_INTERCEPTOR_COPY_STRING(ctx, dst, src, internal_strlen(dst) + 1);
   return res;
 }
 
@@ -7379,7 +7479,7 @@ INTERCEPTOR(void *, mmap, void *addr, SIZE_T sz, int prot, int flags, int fd,
             OFF_T off) {
   void *ctx;
   if (common_flags()->detect_write_exec)
-    ReportMmapWriteExec(prot);
+    ReportMmapWriteExec(prot, flags);
   if (COMMON_INTERCEPTOR_NOTHING_IS_INITIALIZED)
     return (void *)internal_mmap(addr, sz, prot, flags, fd, off);
   COMMON_INTERCEPTOR_ENTER(ctx, mmap, addr, sz, prot, flags, fd, off);
@@ -7389,7 +7489,7 @@ INTERCEPTOR(void *, mmap, void *addr, SIZE_T sz, int prot, int flags, int fd,
 INTERCEPTOR(int, mprotect, void *addr, SIZE_T sz, int prot) {
   void *ctx;
   if (common_flags()->detect_write_exec)
-    ReportMmapWriteExec(prot);
+    ReportMmapWriteExec(prot, 0);
   if (COMMON_INTERCEPTOR_NOTHING_IS_INITIALIZED)
     return (int)internal_mprotect(addr, sz, prot);
   COMMON_INTERCEPTOR_ENTER(ctx, mprotect, addr, sz, prot);
@@ -7408,7 +7508,7 @@ INTERCEPTOR(void *, mmap64, void *addr, SIZE_T sz, int prot, int flags, int fd,
             OFF64_T off) {
   void *ctx;
   if (common_flags()->detect_write_exec)
-    ReportMmapWriteExec(prot);
+    ReportMmapWriteExec(prot, flags);
   if (COMMON_INTERCEPTOR_NOTHING_IS_INITIALIZED)
     return (void *)internal_mmap(addr, sz, prot, flags, fd, off);
   COMMON_INTERCEPTOR_ENTER(ctx, mmap64, addr, sz, prot, flags, fd, off);
@@ -7426,7 +7526,7 @@ INTERCEPTOR(char *, devname, u64 dev, u32 type) {
   COMMON_INTERCEPTOR_ENTER(ctx, devname, dev, type);
   name = REAL(devname)(dev, type);
   if (name)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, name, internal_strlen(name) + 1);
   return name;
 }
 #define INIT_DEVNAME COMMON_INTERCEPT_FUNCTION(devname);
@@ -7448,7 +7548,7 @@ INTERCEPTOR(DEVNAME_R_RETTYPE, devname_r, u64 dev, u32 type, char *path,
   COMMON_INTERCEPTOR_ENTER(ctx, devname_r, dev, type, path, len);
   DEVNAME_R_RETTYPE res = REAL(devname_r)(dev, type, path, len);
   if (DEVNAME_R_SUCCESS(res))
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, path, REAL(strlen)(path) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, path, internal_strlen(path) + 1);
   return res;
 }
 #define INIT_DEVNAME_R COMMON_INTERCEPT_FUNCTION(devname_r);
@@ -7478,7 +7578,7 @@ INTERCEPTOR(void, strmode, u32 mode, char *bp) {
   COMMON_INTERCEPTOR_ENTER(ctx, strmode, mode, bp);
   REAL(strmode)(mode, bp);
   if (bp)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, bp, REAL(strlen)(bp) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, bp, internal_strlen(bp) + 1);
 }
 #define INIT_STRMODE COMMON_INTERCEPT_FUNCTION(strmode)
 #else
@@ -7498,37 +7598,42 @@ INTERCEPTOR(struct __sanitizer_ttyent *, getttynam, char *name) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getttynam, name);
   if (name)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   struct __sanitizer_ttyent *ttyent = REAL(getttynam)(name);
   if (ttyent)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ttyent, struct_ttyent_sz);
   return ttyent;
 }
+#define INIT_TTYENT \
+  COMMON_INTERCEPT_FUNCTION(getttyent); \
+  COMMON_INTERCEPT_FUNCTION(getttynam);
+#else
+#define INIT_TTYENT
+#endif
+
+#if SANITIZER_INTERCEPT_TTYENTPATH
 INTERCEPTOR(int, setttyentpath, char *path) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, setttyentpath, path);
   if (path)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   return REAL(setttyentpath)(path);
 }
-#define INIT_TTYENT \
-  COMMON_INTERCEPT_FUNCTION(getttyent); \
-  COMMON_INTERCEPT_FUNCTION(getttynam); \
-  COMMON_INTERCEPT_FUNCTION(setttyentpath)
+#define INIT_TTYENTPATH COMMON_INTERCEPT_FUNCTION(setttyentpath);
 #else
-#define INIT_TTYENT
+#define INIT_TTYENTPATH
 #endif
 
 #if SANITIZER_INTERCEPT_PROTOENT
 static void write_protoent(void *ctx, struct __sanitizer_protoent *p) {
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p, sizeof(*p));
 
-  COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p->p_name, REAL(strlen)(p->p_name) + 1);
+  COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p->p_name, internal_strlen(p->p_name) + 1);
 
   SIZE_T pp_size = 1; // One handles the trailing \0
 
   for (char **pp = p->p_aliases; *pp; ++pp, ++pp_size)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *pp, REAL(strlen)(*pp) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *pp, internal_strlen(*pp) + 1);
 
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, p->p_aliases,
                                   pp_size * sizeof(char **));
@@ -7547,7 +7652,7 @@ INTERCEPTOR(struct __sanitizer_protoent *, getprotobyname, const char *name) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getprotobyname, name);
   if (name)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   struct __sanitizer_protoent *p = REAL(getprotobyname)(name);
   if (p)
     write_protoent(ctx, p);
@@ -7591,7 +7696,7 @@ INTERCEPTOR(int, getprotobyname_r, const char *name,
   COMMON_INTERCEPTOR_ENTER(ctx, getprotobyname_r, name, result_buf, buf,
                            buflen, result);
   if (name)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   int res = REAL(getprotobyname_r)(name, result_buf, buf, buflen, result);
 
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, result, sizeof *result);
@@ -7630,12 +7735,12 @@ INTERCEPTOR(struct __sanitizer_netent *, getnetent) {
   if (n) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n, sizeof(*n));
 
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n->n_name, REAL(strlen)(n->n_name) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n->n_name, internal_strlen(n->n_name) + 1);
 
     SIZE_T nn_size = 1; // One handles the trailing \0
 
     for (char **nn = n->n_aliases; *nn; ++nn, ++nn_size)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *nn, REAL(strlen)(*nn) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *nn, internal_strlen(*nn) + 1);
 
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n->n_aliases,
                                    nn_size * sizeof(char **));
@@ -7647,17 +7752,17 @@ INTERCEPTOR(struct __sanitizer_netent *, getnetbyname, const char *name) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getnetbyname, name);
   if (name)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
   struct __sanitizer_netent *n = REAL(getnetbyname)(name);
   if (n) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n, sizeof(*n));
 
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n->n_name, REAL(strlen)(n->n_name) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n->n_name, internal_strlen(n->n_name) + 1);
 
     SIZE_T nn_size = 1; // One handles the trailing \0
 
     for (char **nn = n->n_aliases; *nn; ++nn, ++nn_size)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *nn, REAL(strlen)(*nn) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *nn, internal_strlen(*nn) + 1);
 
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n->n_aliases,
                                    nn_size * sizeof(char **));
@@ -7672,12 +7777,12 @@ INTERCEPTOR(struct __sanitizer_netent *, getnetbyaddr, u32 net, int type) {
   if (n) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n, sizeof(*n));
 
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n->n_name, REAL(strlen)(n->n_name) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n->n_name, internal_strlen(n->n_name) + 1);
 
     SIZE_T nn_size = 1; // One handles the trailing \0
 
     for (char **nn = n->n_aliases; *nn; ++nn, ++nn_size)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *nn, REAL(strlen)(*nn) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *nn, internal_strlen(*nn) + 1);
 
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, n->n_aliases,
                                    nn_size * sizeof(char **));
@@ -7798,7 +7903,7 @@ INTERCEPTOR(int, regcomp, void *preg, const char *pattern, int cflags) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, regcomp, preg, pattern, cflags);
   if (pattern)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, pattern, REAL(strlen)(pattern) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, pattern, internal_strlen(pattern) + 1);
   int res = REAL(regcomp)(preg, pattern, cflags);
   if (!res)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, preg, struct_regex_sz);
@@ -7811,7 +7916,7 @@ INTERCEPTOR(int, regexec, const void *preg, const char *string, SIZE_T nmatch,
   if (preg)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, preg, struct_regex_sz);
   if (string)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, string, REAL(strlen)(string) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, string, internal_strlen(string) + 1);
   int res = REAL(regexec)(preg, string, nmatch, pmatch, eflags);
   if (!res && pmatch)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, pmatch, nmatch * struct_regmatch_sz);
@@ -7825,7 +7930,7 @@ INTERCEPTOR(SIZE_T, regerror, int errcode, const void *preg, char *errbuf,
     COMMON_INTERCEPTOR_READ_RANGE(ctx, preg, struct_regex_sz);
   SIZE_T res = REAL(regerror)(errcode, preg, errbuf, errbuf_size);
   if (errbuf)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, errbuf, REAL(strlen)(errbuf) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, errbuf, internal_strlen(errbuf) + 1);
   return res;
 }
 INTERCEPTOR(void, regfree, const void *preg) {
@@ -7850,15 +7955,15 @@ INTERCEPTOR(SSIZE_T, regnsub, char *buf, SIZE_T bufsiz, const char *sub,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, regnsub, buf, bufsiz, sub, rm, str);
   if (sub)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, sub, REAL(strlen)(sub) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, sub, internal_strlen(sub) + 1);
   // The implementation demands and hardcodes 10 elements
   if (rm)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, rm, 10 * struct_regmatch_sz);
   if (str)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, str, REAL(strlen)(str) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, str, internal_strlen(str) + 1);
   SSIZE_T res = REAL(regnsub)(buf, bufsiz, sub, rm, str);
   if (res > 0 && buf)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, REAL(strlen)(buf) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, internal_strlen(buf) + 1);
   return res;
 }
 INTERCEPTOR(SSIZE_T, regasub, char **buf, const char *sub,
@@ -7866,16 +7971,16 @@ INTERCEPTOR(SSIZE_T, regasub, char **buf, const char *sub,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, regasub, buf, sub, rm, sstr);
   if (sub)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, sub, REAL(strlen)(sub) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, sub, internal_strlen(sub) + 1);
   // Hardcode 10 elements as this is hardcoded size
   if (rm)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, rm, 10 * struct_regmatch_sz);
   if (sstr)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, sstr, REAL(strlen)(sstr) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, sstr, internal_strlen(sstr) + 1);
   SSIZE_T res = REAL(regasub)(buf, sub, rm, sstr);
   if (res > 0 && buf) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, sizeof(char *));
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *buf, REAL(strlen)(*buf) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *buf, internal_strlen(*buf) + 1);
   }
   return res;
 }
@@ -7897,7 +8002,7 @@ INTERCEPTOR(void *, fts_open, char *const *path_argv, int options,
       COMMON_INTERCEPTOR_READ_RANGE(ctx, pa, sizeof(char **));
       if (!*pa)
         break;
-      COMMON_INTERCEPTOR_READ_RANGE(ctx, *pa, REAL(strlen)(*pa) + 1);
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, *pa, internal_strlen(*pa) + 1);
     }
   }
   // TODO(kamil): handle compar callback
@@ -7989,7 +8094,7 @@ INTERCEPTOR(int, sysctlbyname, char *sname, void *oldp, SIZE_T *oldlenp,
   COMMON_INTERCEPTOR_ENTER(ctx, sysctlbyname, sname, oldp, oldlenp, newp,
                            newlen);
   if (sname)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, sname, REAL(strlen)(sname) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, sname, internal_strlen(sname) + 1);
   if (oldlenp)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, oldlenp, sizeof(*oldlenp));
   if (newp && newlen)
@@ -8010,7 +8115,7 @@ INTERCEPTOR(int, sysctlnametomib, const char *sname, int *name,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, sysctlnametomib, sname, name, namelenp);
   if (sname)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, sname, REAL(strlen)(sname) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, sname, internal_strlen(sname) + 1);
   if (namelenp)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, namelenp, sizeof(*namelenp));
   int res = REAL(sysctlnametomib)(sname, name, namelenp);
@@ -8050,7 +8155,7 @@ INTERCEPTOR(void *, asysctlbyname, const char *sname, SIZE_T *len) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, asysctlbyname, sname, len);
   if (sname)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, sname, REAL(strlen)(sname) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, sname, internal_strlen(sname) + 1);
   void *res = REAL(asysctlbyname)(sname, len);
   if (res && len) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, len, sizeof(*len));
@@ -8073,7 +8178,7 @@ INTERCEPTOR(int, sysctlgetmibinfo, char *sname, int *name,
   COMMON_INTERCEPTOR_ENTER(ctx, sysctlgetmibinfo, sname, name, namelenp, cname,
                            csz, rnode, v);
   if (sname)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, sname, REAL(strlen)(sname) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, sname, internal_strlen(sname) + 1);
   if (namelenp)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, namelenp, sizeof(*namelenp));
   if (csz)
@@ -8107,7 +8212,7 @@ INTERCEPTOR(char *, nl_langinfo, long item) {
   COMMON_INTERCEPTOR_ENTER(ctx, nl_langinfo, item);
   char *ret = REAL(nl_langinfo)(item);
   if (ret)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, REAL(strlen)(ret) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, internal_strlen(ret) + 1);
   return ret;
 }
 #define INIT_NL_LANGINFO COMMON_INTERCEPT_FUNCTION(nl_langinfo)
@@ -8127,7 +8232,7 @@ INTERCEPTOR(int, modctl, int operation, void *argp) {
       COMMON_INTERCEPTOR_READ_RANGE(ctx, ml, sizeof(*ml));
       if (ml->ml_filename)
         COMMON_INTERCEPTOR_READ_RANGE(ctx, ml->ml_filename,
-                                      REAL(strlen)(ml->ml_filename) + 1);
+                                      internal_strlen(ml->ml_filename) + 1);
       if (ml->ml_props)
         COMMON_INTERCEPTOR_READ_RANGE(ctx, ml->ml_props, ml->ml_propslen);
     }
@@ -8135,7 +8240,7 @@ INTERCEPTOR(int, modctl, int operation, void *argp) {
   } else if (operation == modctl_unload) {
     if (argp) {
       const char *name = (const char *)argp;
-      COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, name, internal_strlen(name) + 1);
     }
     ret = REAL(modctl)(operation, argp);
   } else if (operation == modctl_stat) {
@@ -8177,7 +8282,7 @@ INTERCEPTOR(long long, strtonum, const char *nptr, long long minval,
   if (errstr) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, errstr, sizeof(const char *));
      if (*errstr)
-      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *errstr, REAL(strlen)(*errstr) + 1);
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *errstr, internal_strlen(*errstr) + 1);
   }
   return ret;
 }
@@ -8197,7 +8302,7 @@ INTERCEPTOR(char *, fparseln, __sanitizer_FILE *stream, SIZE_T *len,
     COMMON_INTERCEPTOR_READ_RANGE(ctx, delim, sizeof(delim[0]) * 3);
   char *ret = REAL(fparseln)(stream, len, lineno, delim, flags);
   if (ret) {
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, REAL(strlen)(ret) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, internal_strlen(ret) + 1);
     if (len)
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, len, sizeof(*len));
     if (lineno)
@@ -8214,7 +8319,7 @@ INTERCEPTOR(char *, fparseln, __sanitizer_FILE *stream, SIZE_T *len,
 INTERCEPTOR(int, statvfs1, const char *path, void *buf, int flags) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, statvfs1, path, buf, flags);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   int res = REAL(statvfs1)(path, buf, flags);
   if (!res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, struct_statvfs_sz);
   return res;
@@ -8495,7 +8600,7 @@ INTERCEPTOR(char *, SHA1File, char *filename, char *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, SHA1File, filename, buf);
   if (filename)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, REAL(strlen)(filename) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, internal_strlen(filename) + 1);
   char *ret = REAL(SHA1File)(filename, buf);
   if (ret)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, SHA1_return_length);
@@ -8506,7 +8611,7 @@ INTERCEPTOR(char *, SHA1FileChunk, char *filename, char *buf, OFF_T offset,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, SHA1FileChunk, filename, buf, offset, length);
   if (filename)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, REAL(strlen)(filename) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, internal_strlen(filename) + 1);
   char *ret = REAL(SHA1FileChunk)(filename, buf, offset, length);
   if (ret)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, SHA1_return_length);
@@ -8582,7 +8687,7 @@ INTERCEPTOR(char *, MD4File, const char *filename, char *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, MD4File, filename, buf);
   if (filename)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, REAL(strlen)(filename) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, internal_strlen(filename) + 1);
   char *ret = REAL(MD4File)(filename, buf);
   if (ret)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, MD4_return_length);
@@ -8665,7 +8770,7 @@ INTERCEPTOR(char *, RMD160File, char *filename, char *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, RMD160File, filename, buf);
   if (filename)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, REAL(strlen)(filename) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, internal_strlen(filename) + 1);
   char *ret = REAL(RMD160File)(filename, buf);
   if (ret)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, RMD160_return_length);
@@ -8676,7 +8781,7 @@ INTERCEPTOR(char *, RMD160FileChunk, char *filename, char *buf, OFF_T offset,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, RMD160FileChunk, filename, buf, offset, length);
   if (filename)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, REAL(strlen)(filename) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, internal_strlen(filename) + 1);
   char *ret = REAL(RMD160FileChunk)(filename, buf, offset, length);
   if (ret)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, RMD160_return_length);
@@ -8752,7 +8857,7 @@ INTERCEPTOR(char *, MD5File, const char *filename, char *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, MD5File, filename, buf);
   if (filename)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, REAL(strlen)(filename) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, internal_strlen(filename) + 1);
   char *ret = REAL(MD5File)(filename, buf);
   if (ret)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, MD5_return_length);
@@ -8882,7 +8987,7 @@ INTERCEPTOR(char *, MD2File, const char *filename, char *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, MD2File, filename, buf);
   if (filename)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, REAL(strlen)(filename) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, internal_strlen(filename) + 1);
   char *ret = REAL(MD2File)(filename, buf);
   if (ret)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, MD2_return_length);
@@ -8960,7 +9065,7 @@ INTERCEPTOR(char *, MD2Data, const unsigned char *data, unsigned int len,
     void *ctx; \
     COMMON_INTERCEPTOR_ENTER(ctx, SHA##LEN##_File, filename, buf); \
     if (filename) \
-      COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, REAL(strlen)(filename) + 1);\
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, internal_strlen(filename) + 1);\
     char *ret = REAL(SHA##LEN##_File)(filename, buf); \
     if (ret) \
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, SHA##LEN##_return_length); \
@@ -8972,7 +9077,7 @@ INTERCEPTOR(char *, MD2Data, const unsigned char *data, unsigned int len,
     COMMON_INTERCEPTOR_ENTER(ctx, SHA##LEN##_FileChunk, filename, buf, offset, \
   length); \
     if (filename) \
-      COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, REAL(strlen)(filename) + 1);\
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, filename, internal_strlen(filename) + 1);\
     char *ret = REAL(SHA##LEN##_FileChunk)(filename, buf, offset, length); \
     if (ret) \
       COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, SHA##LEN##_return_length); \
@@ -8989,10 +9094,10 @@ INTERCEPTOR(char *, MD2Data, const unsigned char *data, unsigned int len,
     return ret; \
   }
 
-SHA2_INTERCEPTORS(224, u32);
-SHA2_INTERCEPTORS(256, u32);
-SHA2_INTERCEPTORS(384, u64);
-SHA2_INTERCEPTORS(512, u64);
+SHA2_INTERCEPTORS(224, u32)
+SHA2_INTERCEPTORS(256, u32)
+SHA2_INTERCEPTORS(384, u64)
+SHA2_INTERCEPTORS(512, u64)
 
 #define INIT_SHA2_INTECEPTORS(LEN) \
   COMMON_INTERCEPT_FUNCTION(SHA##LEN##_Init); \
@@ -9036,7 +9141,7 @@ INTERCEPTOR(int, strvis, char *dst, const char *src, int flag) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, strvis, dst, src, flag);
   if (src)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, REAL(strlen)(src) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, internal_strlen(src) + 1);
   int len = REAL(strvis)(dst, src, flag);
   if (dst)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, len + 1);
@@ -9046,7 +9151,7 @@ INTERCEPTOR(int, stravis, char **dst, const char *src, int flag) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, stravis, dst, src, flag);
   if (src)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, REAL(strlen)(src) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, internal_strlen(src) + 1);
   int len = REAL(stravis)(dst, src, flag);
   if (dst) {
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, sizeof(char *));
@@ -9059,7 +9164,7 @@ INTERCEPTOR(int, strnvis, char *dst, SIZE_T dlen, const char *src, int flag) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, strnvis, dst, dlen, src, flag);
   if (src)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, REAL(strlen)(src) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, internal_strlen(src) + 1);
   int len = REAL(strnvis)(dst, dlen, src, flag);
   // The interface will be valid even if there is no space for NULL char
   if (dst && len > 0)
@@ -9109,7 +9214,7 @@ INTERCEPTOR(char *, svis, char *dst, int c, int flag, int nextc,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, svis, dst, c, flag, nextc, extra);
   if (extra)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, REAL(strlen)(extra) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, internal_strlen(extra) + 1);
   char *end = REAL(svis)(dst, c, flag, nextc, extra);
   if (dst && end)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, end - dst + 1);
@@ -9120,7 +9225,7 @@ INTERCEPTOR(char *, snvis, char *dst, SIZE_T dlen, int c, int flag, int nextc,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, snvis, dst, dlen, c, flag, nextc, extra);
   if (extra)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, REAL(strlen)(extra) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, internal_strlen(extra) + 1);
   char *end = REAL(snvis)(dst, dlen, c, flag, nextc, extra);
   if (dst && end)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst,
@@ -9132,9 +9237,9 @@ INTERCEPTOR(int, strsvis, char *dst, const char *src, int flag,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, strsvis, dst, src, flag, extra);
   if (src)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, REAL(strlen)(src) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, internal_strlen(src) + 1);
   if (extra)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, REAL(strlen)(extra) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, internal_strlen(extra) + 1);
   int len = REAL(strsvis)(dst, src, flag, extra);
   if (dst)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, len + 1);
@@ -9145,9 +9250,9 @@ INTERCEPTOR(int, strsnvis, char *dst, SIZE_T dlen, const char *src, int flag,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, strsnvis, dst, dlen, src, flag, extra);
   if (src)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, REAL(strlen)(src) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, internal_strlen(src) + 1);
   if (extra)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, REAL(strlen)(extra) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, internal_strlen(extra) + 1);
   int len = REAL(strsnvis)(dst, dlen, src, flag, extra);
   // The interface will be valid even if there is no space for NULL char
   if (dst && len >= 0)
@@ -9161,7 +9266,7 @@ INTERCEPTOR(int, strsvisx, char *dst, const char *src, SIZE_T len, int flag,
   if (src)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, src, len);
   if (extra)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, REAL(strlen)(extra) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, internal_strlen(extra) + 1);
   int ret = REAL(strsvisx)(dst, src, len, flag, extra);
   if (dst)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, ret + 1);
@@ -9174,7 +9279,7 @@ INTERCEPTOR(int, strsnvisx, char *dst, SIZE_T dlen, const char *src, SIZE_T len,
   if (src)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, src, len);
   if (extra)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, REAL(strlen)(extra) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, internal_strlen(extra) + 1);
   int ret = REAL(strsnvisx)(dst, dlen, src, len, flag, extra);
   if (dst && ret >= 0)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, ret + 1);
@@ -9188,7 +9293,7 @@ INTERCEPTOR(int, strsenvisx, char *dst, SIZE_T dlen, const char *src,
   if (src)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, src, len);
   if (extra)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, REAL(strlen)(extra) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, extra, internal_strlen(extra) + 1);
   // FIXME: only need to be checked when "flag | VIS_NOLOCALE" doesn't hold
   // according to the implementation
   if (cerr_ptr)
@@ -9215,7 +9320,7 @@ INTERCEPTOR(int, strunvis, char *dst, const char *src) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, strunvis, dst, src);
   if (src)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, REAL(strlen)(src) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, internal_strlen(src) + 1);
   int ret = REAL(strunvis)(dst, src);
   if (ret != -1)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, ret + 1);
@@ -9225,7 +9330,7 @@ INTERCEPTOR(int, strnunvis, char *dst, SIZE_T dlen, const char *src) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, strnunvis, dst, dlen, src);
   if (src)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, REAL(strlen)(src) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, internal_strlen(src) + 1);
   int ret = REAL(strnunvis)(dst, dlen, src);
   if (ret != -1)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, ret + 1);
@@ -9235,7 +9340,7 @@ INTERCEPTOR(int, strunvisx, char *dst, const char *src, int flag) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, strunvisx, dst, src, flag);
   if (src)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, REAL(strlen)(src) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, internal_strlen(src) + 1);
   int ret = REAL(strunvisx)(dst, src, flag);
   if (ret != -1)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, ret + 1);
@@ -9246,7 +9351,7 @@ INTERCEPTOR(int, strnunvisx, char *dst, SIZE_T dlen, const char *src,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, strnunvisx, dst, dlen, src, flag);
   if (src)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, REAL(strlen)(src) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, src, internal_strlen(src) + 1);
   int ret = REAL(strnunvisx)(dst, dlen, src, flag);
   if (ret != -1)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dst, ret + 1);
@@ -9282,7 +9387,7 @@ INTERCEPTOR(struct __sanitizer_cdbr *, cdbr_open, const char *path, int flags) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, cdbr_open, path, flags);
   if (path)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   struct __sanitizer_cdbr *cdbr = REAL(cdbr_open)(path, flags);
   if (cdbr)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, cdbr, sizeof(*cdbr));
@@ -9474,7 +9579,7 @@ INTERCEPTOR(void *, getfsspec, const char *spec) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getfsspec, spec);
   if (spec)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, spec, REAL(strlen)(spec) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, spec, internal_strlen(spec) + 1);
   void *ret = REAL(getfsspec)(spec);
   if (ret)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, struct_fstab_sz);
@@ -9485,7 +9590,7 @@ INTERCEPTOR(void *, getfsfile, const char *file) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getfsfile, file);
   if (file)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, file, REAL(strlen)(file) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, file, internal_strlen(file) + 1);
   void *ret = REAL(getfsfile)(file);
   if (ret)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ret, struct_fstab_sz);
@@ -9529,9 +9634,9 @@ INTERCEPTOR(__sanitizer_FILE *, popen, const char *command, const char *type) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, popen, command, type);
   if (command)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, command, REAL(strlen)(command) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, command, internal_strlen(command) + 1);
   if (type)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, type, REAL(strlen)(type) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, type, internal_strlen(type) + 1);
   __sanitizer_FILE *res = REAL(popen)(command, type);
   COMMON_INTERCEPTOR_FILE_OPEN(ctx, res, nullptr);
   if (res) unpoison_file(res);
@@ -9548,13 +9653,13 @@ INTERCEPTOR(__sanitizer_FILE *, popenve, const char *path,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, popenve, path, argv, envp, type);
   if (path)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   if (argv) {
     for (char *const *pa = argv; ; ++pa) {
       COMMON_INTERCEPTOR_READ_RANGE(ctx, pa, sizeof(char **));
       if (!*pa)
         break;
-      COMMON_INTERCEPTOR_READ_RANGE(ctx, *pa, REAL(strlen)(*pa) + 1);
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, *pa, internal_strlen(*pa) + 1);
     }
   }
   if (envp) {
@@ -9562,11 +9667,11 @@ INTERCEPTOR(__sanitizer_FILE *, popenve, const char *path,
       COMMON_INTERCEPTOR_READ_RANGE(ctx, pa, sizeof(char **));
       if (!*pa)
         break;
-      COMMON_INTERCEPTOR_READ_RANGE(ctx, *pa, REAL(strlen)(*pa) + 1);
+      COMMON_INTERCEPTOR_READ_RANGE(ctx, *pa, internal_strlen(*pa) + 1);
     }
   }
   if (type)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, type, REAL(strlen)(type) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, type, internal_strlen(type) + 1);
   __sanitizer_FILE *res = REAL(popenve)(path, argv, envp, type);
   COMMON_INTERCEPTOR_FILE_OPEN(ctx, res, nullptr);
   if (res) unpoison_file(res);
@@ -9762,7 +9867,7 @@ INTERCEPTOR(char *, fdevname,  int fd) {
   COMMON_INTERCEPTOR_FD_ACCESS(ctx, fd);
   char *name = REAL(fdevname)(fd);
   if (name) {
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, name, REAL(strlen)(name) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, name, internal_strlen(name) + 1);
     if (fd > 0)
       COMMON_INTERCEPTOR_FD_ACQUIRE(ctx, fd);
   }
@@ -9775,7 +9880,7 @@ INTERCEPTOR(char *, fdevname_r,  int fd, char *buf, SIZE_T len) {
   COMMON_INTERCEPTOR_FD_ACCESS(ctx, fd);
   char *name = REAL(fdevname_r)(fd, buf, len);
   if (name && buf && len > 0) {
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, REAL(strlen)(buf) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, internal_strlen(buf) + 1);
     if (fd > 0)
       COMMON_INTERCEPTOR_FD_ACQUIRE(ctx, fd);
   }
@@ -9795,7 +9900,7 @@ INTERCEPTOR(char *, getusershell) {
   COMMON_INTERCEPTOR_ENTER(ctx, getusershell);
   char *res = REAL(getusershell)();
   if (res)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   return res;
 }
 
@@ -9820,7 +9925,7 @@ INTERCEPTOR(int, sl_add, void *sl, char *item) {
   if (sl)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, sl, __sanitizer::struct_StringList_sz);
   if (item)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, item, REAL(strlen)(item) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, item, internal_strlen(item) + 1);
   int res = REAL(sl_add)(sl, item);
   if (!res)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, sl, __sanitizer::struct_StringList_sz);
@@ -9833,10 +9938,10 @@ INTERCEPTOR(char *, sl_find, void *sl, const char *item) {
   if (sl)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, sl, __sanitizer::struct_StringList_sz);
   if (item)
-    COMMON_INTERCEPTOR_READ_RANGE(ctx, item, REAL(strlen)(item) + 1);
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, item, internal_strlen(item) + 1);
   char *res = REAL(sl_find)(sl, item);
   if (res)
-    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, REAL(strlen)(res) + 1);
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, res, internal_strlen(res) + 1);
   return res;
 }
 
@@ -9922,7 +10027,52 @@ INTERCEPTOR(int, getentropy, void *buf, SIZE_T buflen) {
 #define INIT_GETENTROPY
 #endif
 
-#if SANITIZER_INTERCEPT_QSORT
+#if SANITIZER_INTERCEPT_QSORT_R
+typedef int (*qsort_r_compar_f)(const void *, const void *, void *);
+struct qsort_r_compar_params {
+  SIZE_T size;
+  qsort_r_compar_f compar;
+  void *arg;
+};
+static int wrapped_qsort_r_compar(const void *a, const void *b, void *arg) {
+  qsort_r_compar_params *params = (qsort_r_compar_params *)arg;
+  COMMON_INTERCEPTOR_UNPOISON_PARAM(3);
+  COMMON_INTERCEPTOR_INITIALIZE_RANGE(a, params->size);
+  COMMON_INTERCEPTOR_INITIALIZE_RANGE(b, params->size);
+  return params->compar(a, b, params->arg);
+}
+
+INTERCEPTOR(void, qsort_r, void *base, SIZE_T nmemb, SIZE_T size,
+            qsort_r_compar_f compar, void *arg) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, qsort_r, base, nmemb, size, compar, arg);
+  // Run the comparator over all array elements to detect any memory issues.
+  if (nmemb > 1) {
+    for (SIZE_T i = 0; i < nmemb - 1; ++i) {
+      void *p = (void *)((char *)base + i * size);
+      void *q = (void *)((char *)base + (i + 1) * size);
+      COMMON_INTERCEPTOR_UNPOISON_PARAM(3);
+      compar(p, q, arg);
+    }
+  }
+  qsort_r_compar_params params = {size, compar, arg};
+  REAL(qsort_r)(base, nmemb, size, wrapped_qsort_r_compar, &params);
+  COMMON_INTERCEPTOR_WRITE_RANGE(ctx, base, nmemb * size);
+}
+#  define INIT_QSORT_R COMMON_INTERCEPT_FUNCTION(qsort_r)
+#else
+#  define INIT_QSORT_R
+#endif
+
+#if SANITIZER_INTERCEPT_QSORT && SANITIZER_INTERCEPT_QSORT_R
+INTERCEPTOR(void, qsort, void *base, SIZE_T nmemb, SIZE_T size,
+            qsort_r_compar_f compar) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, qsort, base, nmemb, size, compar);
+  WRAP(qsort_r)(base, nmemb, size, compar, nullptr);
+}
+#  define INIT_QSORT COMMON_INTERCEPT_FUNCTION(qsort)
+#elif SANITIZER_INTERCEPT_QSORT && !SANITIZER_INTERCEPT_QSORT_R
 // Glibc qsort uses a temporary buffer allocated either on stack or on heap.
 // Poisoned memory from there may get copied into the comparator arguments,
 // where it needs to be dealt with. But even that is not enough - the results of
@@ -9937,7 +10087,7 @@ INTERCEPTOR(int, getentropy, void *buf, SIZE_T buflen) {
 typedef int (*qsort_compar_f)(const void *, const void *);
 static THREADLOCAL qsort_compar_f qsort_compar;
 static THREADLOCAL SIZE_T qsort_size;
-int wrapped_qsort_compar(const void *a, const void *b) {
+static int wrapped_qsort_compar(const void *a, const void *b) {
   COMMON_INTERCEPTOR_UNPOISON_PARAM(2);
   COMMON_INTERCEPTOR_INITIALIZE_RANGE(a, qsort_size);
   COMMON_INTERCEPTOR_INITIALIZE_RANGE(b, qsort_size);
@@ -9979,60 +10129,34 @@ INTERCEPTOR(void, qsort, void *base, SIZE_T nmemb, SIZE_T size,
   }
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, base, nmemb * size);
 }
-#define INIT_QSORT COMMON_INTERCEPT_FUNCTION(qsort)
+#  define INIT_QSORT COMMON_INTERCEPT_FUNCTION(qsort)
 #else
-#define INIT_QSORT
+#  define INIT_QSORT
 #endif
 
-#if SANITIZER_INTERCEPT_QSORT_R
-typedef int (*qsort_r_compar_f)(const void *, const void *, void *);
-static THREADLOCAL qsort_r_compar_f qsort_r_compar;
-static THREADLOCAL SIZE_T qsort_r_size;
-int wrapped_qsort_r_compar(const void *a, const void *b, void *arg) {
-  COMMON_INTERCEPTOR_UNPOISON_PARAM(3);
-  COMMON_INTERCEPTOR_INITIALIZE_RANGE(a, qsort_r_size);
-  COMMON_INTERCEPTOR_INITIALIZE_RANGE(b, qsort_r_size);
-  return qsort_r_compar(a, b, arg);
+#if SANITIZER_INTERCEPT_BSEARCH
+typedef int (*bsearch_compar_f)(const void *, const void *);
+struct bsearch_compar_params {
+  const void *key;
+  bsearch_compar_f compar;
+};
+
+static int wrapped_bsearch_compar(const void *key, const void *b) {
+  const bsearch_compar_params *params = (const bsearch_compar_params *)key;
+  COMMON_INTERCEPTOR_UNPOISON_PARAM(2);
+  return params->compar(params->key, b);
 }
 
-INTERCEPTOR(void, qsort_r, void *base, SIZE_T nmemb, SIZE_T size,
-            qsort_r_compar_f compar, void *arg) {
+INTERCEPTOR(void *, bsearch, const void *key, const void *base, SIZE_T nmemb,
+            SIZE_T size, bsearch_compar_f compar) {
   void *ctx;
-  COMMON_INTERCEPTOR_ENTER(ctx, qsort_r, base, nmemb, size, compar, arg);
-  // Run the comparator over all array elements to detect any memory issues.
-  if (nmemb > 1) {
-    for (SIZE_T i = 0; i < nmemb - 1; ++i) {
-      void *p = (void *)((char *)base + i * size);
-      void *q = (void *)((char *)base + (i + 1) * size);
-      COMMON_INTERCEPTOR_UNPOISON_PARAM(3);
-      compar(p, q, arg);
-    }
-  }
-  qsort_r_compar_f old_compar = qsort_r_compar;
-  SIZE_T old_size = qsort_r_size;
-  // Handle qsort_r() implementations that recurse using an
-  // interposable function call:
-  bool already_wrapped = compar == wrapped_qsort_r_compar;
-  if (already_wrapped) {
-    // This case should only happen if the qsort() implementation calls itself
-    // using a preemptible function call (e.g. the FreeBSD libc version).
-    // Check that the size and comparator arguments are as expected.
-    CHECK_NE(compar, qsort_r_compar);
-    CHECK_EQ(qsort_r_size, size);
-  } else {
-    qsort_r_compar = compar;
-    qsort_r_size = size;
-  }
-  REAL(qsort_r)(base, nmemb, size, wrapped_qsort_r_compar, arg);
-  if (!already_wrapped) {
-    qsort_r_compar = old_compar;
-    qsort_r_size = old_size;
-  }
-  COMMON_INTERCEPTOR_WRITE_RANGE(ctx, base, nmemb * size);
+  COMMON_INTERCEPTOR_ENTER(ctx, bsearch, key, base, nmemb, size, compar);
+  bsearch_compar_params params = {key, compar};
+  return REAL(bsearch)(&params, base, nmemb, size, wrapped_bsearch_compar);
 }
-#define INIT_QSORT_R COMMON_INTERCEPT_FUNCTION(qsort_r)
+#  define INIT_BSEARCH COMMON_INTERCEPT_FUNCTION(bsearch)
 #else
-#define INIT_QSORT_R
+#  define INIT_BSEARCH
 #endif
 
 #if SANITIZER_INTERCEPT_SIGALTSTACK
@@ -10166,6 +10290,7 @@ static void InitializeCommonInterceptors() {
   INIT_TIME;
   INIT_GLOB;
   INIT_GLOB64;
+  INIT_POSIX_SPAWN;
   INIT_WAIT;
   INIT_WAIT4;
   INIT_INET;
@@ -10401,6 +10526,7 @@ static void InitializeCommonInterceptors() {
   INIT_GETENTROPY;
   INIT_QSORT;
   INIT_QSORT_R;
+  INIT_BSEARCH;
   INIT_SIGALTSTACK;
   INIT_UNAME;
   INIT___XUNAME;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_format.inc b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_format.inc
index 082398ba960a..220abb89c3be 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_format.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_format.inc
@@ -324,8 +324,8 @@ static void scanf_common(void *ctx, int n_inputs, bool allowGnuMalloc,
       continue;
     int size = scanf_get_value_size(&dir);
     if (size == FSS_INVALID) {
-      Report("%s: WARNING: unexpected format specifier in scanf interceptor: ",
-             SanitizerToolName, "%.*s\n", dir.end - dir.begin, dir.begin);
+      Report("%s: WARNING: unexpected format specifier in scanf interceptor: %.*s\n",
+             SanitizerToolName, static_cast<int>(dir.end - dir.begin), dir.begin);
       break;
     }
     void *argp = va_arg(aq, void *);
@@ -469,7 +469,7 @@ static int printf_get_value_size(PrintfDirective *dir) {
         break;                                                     \
       default:                                                     \
         Report("WARNING: unexpected floating-point arg size"       \
-               " in printf interceptor: %d\n", size);              \
+               " in printf interceptor: %zu\n", static_cast<uptr>(size));             \
         return;                                                    \
       }                                                            \
     } else {                                                       \
@@ -484,7 +484,7 @@ static int printf_get_value_size(PrintfDirective *dir) {
         break;                                                     \
       default:                                                     \
         Report("WARNING: unexpected arg size"                      \
-               " in printf interceptor: %d\n", size);              \
+               " in printf interceptor: %zu\n", static_cast<uptr>(size));             \
         return;                                                    \
       }                                                            \
     }                                                              \
@@ -530,7 +530,7 @@ static void printf_common(void *ctx, const char *format, va_list aq) {
         Report(
             "%s: WARNING: unexpected format specifier in printf "
             "interceptor: %.*s (reported once per process)\n",
-            SanitizerToolName, dir.end - dir.begin, dir.begin);
+            SanitizerToolName, static_cast<int>(dir.end - dir.begin), dir.begin);
       break;
     }
     if (dir.convSpecifier == 'n') {
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_netbsd_compat.inc b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_netbsd_compat.inc
index 6aa73ec8c6a2..f6ac3fa5af18 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_netbsd_compat.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_netbsd_compat.inc
@@ -33,7 +33,7 @@
 INTERCEPTOR(int, statvfs, char *path, void *buf) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, statvfs, path, buf);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -99,7 +99,7 @@ INTERCEPTOR(int, getvfsstat, void *buf, SIZE_T bufsize, int flags) {
 INTERCEPTOR(int, statvfs1, const char *path, void *buf, int flags) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, statvfs1, path, buf, flags);
-  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, REAL(strlen)(path) + 1);
+  if (path) COMMON_INTERCEPTOR_READ_RANGE(ctx, path, internal_strlen(path) + 1);
   int res = REAL(statvfs1)(path, buf, flags);
   if (!res) COMMON_INTERCEPTOR_WRITE_RANGE(ctx, buf, struct_statvfs90_sz);
   return res;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_vfork_i386.inc.S b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_vfork_i386.inc.S
index ed693819c6d4..f60b05d157bb 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_vfork_i386.inc.S
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_vfork_i386.inc.S
@@ -6,6 +6,7 @@
 .globl ASM_WRAPPER_NAME(vfork)
 ASM_TYPE_FUNCTION(ASM_WRAPPER_NAME(vfork))
 ASM_WRAPPER_NAME(vfork):
+        _CET_ENDBR
         // Store return address in the spill area and tear down the stack frame.
         sub     $12, %esp
         call    COMMON_INTERCEPTOR_SPILL_AREA
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_vfork_x86_64.inc.S b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_vfork_x86_64.inc.S
index 8147cdd09247..8fd18ea67ffd 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_vfork_x86_64.inc.S
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_vfork_x86_64.inc.S
@@ -6,6 +6,7 @@
 .globl ASM_WRAPPER_NAME(vfork)
 ASM_TYPE_FUNCTION(ASM_WRAPPER_NAME(vfork))
 ASM_WRAPPER_NAME(vfork):
+        _CET_ENDBR
         // Store return address in the spill area and tear down the stack frame.
         push    %rcx
         call    COMMON_INTERCEPTOR_SPILL_AREA
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_libcdep.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_common_libcdep.cpp
index 01ccacc6f320..bc4b477e350f 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_libcdep.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_libcdep.cpp
@@ -26,9 +26,7 @@ void SetSoftRssLimitExceededCallback(void (*Callback)(bool exceeded)) {
 
 #if (SANITIZER_LINUX || SANITIZER_NETBSD) && !SANITIZER_GO
 // Weak default implementation for when sanitizer_stackdepot is not linked in.
-SANITIZER_WEAK_ATTRIBUTE StackDepotStats *StackDepotGetStats() {
-  return nullptr;
-}
+SANITIZER_WEAK_ATTRIBUTE StackDepotStats StackDepotGetStats() { return {}; }
 
 void *BackgroundThread(void *arg) {
   const uptr hard_rss_limit_mb = common_flags()->hard_rss_limit_mb;
@@ -48,16 +46,12 @@ void *BackgroundThread(void *arg) {
         prev_reported_rss = current_rss_mb;
       }
       // If stack depot has grown 10% since last time, print it too.
-      StackDepotStats *stack_depot_stats = StackDepotGetStats();
-      if (stack_depot_stats) {
-        if (prev_reported_stack_depot_size * 11 / 10 <
-            stack_depot_stats->allocated) {
-          Printf("%s: StackDepot: %zd ids; %zdM allocated\n",
-                 SanitizerToolName,
-                 stack_depot_stats->n_uniq_ids,
-                 stack_depot_stats->allocated >> 20);
-          prev_reported_stack_depot_size = stack_depot_stats->allocated;
-        }
+      StackDepotStats stack_depot_stats = StackDepotGetStats();
+      if (prev_reported_stack_depot_size * 11 / 10 <
+          stack_depot_stats.allocated) {
+        Printf("%s: StackDepot: %zd ids; %zdM allocated\n", SanitizerToolName,
+               stack_depot_stats.n_uniq_ids, stack_depot_stats.allocated >> 20);
+        prev_reported_stack_depot_size = stack_depot_stats.allocated;
       }
     }
     // Check RSS against the limit.
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_nolibc.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_common_nolibc.cpp
index 9a4e5388f24d..a20602d8b95a 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_nolibc.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_nolibc.cpp
@@ -25,6 +25,7 @@ void LogMessageOnPrintf(const char *str) {}
 #endif
 void WriteToSyslog(const char *buffer) {}
 void Abort() { internal__exit(1); }
+bool CreateDir(const char *pathname) { return false; }
 #endif // !SANITIZER_WINDOWS
 
 #if !SANITIZER_WINDOWS && !SANITIZER_MAC
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_syscalls.inc b/compiler-rt/lib/sanitizer_common/sanitizer_common_syscalls.inc
index 1b89d6e17684..a38b134085aa 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_syscalls.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_syscalls.inc
@@ -43,45 +43,47 @@
 #include "sanitizer_platform.h"
 #if SANITIZER_LINUX
 
-#include "sanitizer_libc.h"
+#  include "sanitizer_libc.h"
 
-#define PRE_SYSCALL(name)                                                      \
-  SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_syscall_pre_impl_##name
-#define PRE_READ(p, s) COMMON_SYSCALL_PRE_READ_RANGE(p, s)
-#define PRE_WRITE(p, s) COMMON_SYSCALL_PRE_WRITE_RANGE(p, s)
+#  define PRE_SYSCALL(name) \
+    SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_syscall_pre_impl_##name
+#  define PRE_READ(p, s) COMMON_SYSCALL_PRE_READ_RANGE(p, s)
+#  define PRE_WRITE(p, s) COMMON_SYSCALL_PRE_WRITE_RANGE(p, s)
 
-#define POST_SYSCALL(name)                                                     \
-  SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_syscall_post_impl_##name
-#define POST_READ(p, s) COMMON_SYSCALL_POST_READ_RANGE(p, s)
-#define POST_WRITE(p, s) COMMON_SYSCALL_POST_WRITE_RANGE(p, s)
+#  define POST_SYSCALL(name) \
+    SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_syscall_post_impl_##name
+#  define POST_READ(p, s) COMMON_SYSCALL_POST_READ_RANGE(p, s)
+#  define POST_WRITE(p, s) COMMON_SYSCALL_POST_WRITE_RANGE(p, s)
 
-#ifndef COMMON_SYSCALL_ACQUIRE
-# define COMMON_SYSCALL_ACQUIRE(addr) ((void)(addr))
-#endif
+#  ifndef COMMON_SYSCALL_ACQUIRE
+#    define COMMON_SYSCALL_ACQUIRE(addr) ((void)(addr))
+#  endif
 
-#ifndef COMMON_SYSCALL_RELEASE
-# define COMMON_SYSCALL_RELEASE(addr) ((void)(addr))
-#endif
+#  ifndef COMMON_SYSCALL_RELEASE
+#    define COMMON_SYSCALL_RELEASE(addr) ((void)(addr))
+#  endif
 
-#ifndef COMMON_SYSCALL_FD_CLOSE
-# define COMMON_SYSCALL_FD_CLOSE(fd) ((void)(fd))
-#endif
+#  ifndef COMMON_SYSCALL_FD_CLOSE
+#    define COMMON_SYSCALL_FD_CLOSE(fd) ((void)(fd))
+#  endif
 
-#ifndef COMMON_SYSCALL_FD_ACQUIRE
-# define COMMON_SYSCALL_FD_ACQUIRE(fd) ((void)(fd))
-#endif
+#  ifndef COMMON_SYSCALL_FD_ACQUIRE
+#    define COMMON_SYSCALL_FD_ACQUIRE(fd) ((void)(fd))
+#  endif
 
-#ifndef COMMON_SYSCALL_FD_RELEASE
-# define COMMON_SYSCALL_FD_RELEASE(fd) ((void)(fd))
-#endif
+#  ifndef COMMON_SYSCALL_FD_RELEASE
+#    define COMMON_SYSCALL_FD_RELEASE(fd) ((void)(fd))
+#  endif
 
-#ifndef COMMON_SYSCALL_PRE_FORK
-# define COMMON_SYSCALL_PRE_FORK() {}
-#endif
+#  ifndef COMMON_SYSCALL_PRE_FORK
+#    define COMMON_SYSCALL_PRE_FORK() \
+      {}
+#  endif
 
-#ifndef COMMON_SYSCALL_POST_FORK
-# define COMMON_SYSCALL_POST_FORK(res) {}
-#endif
+#  ifndef COMMON_SYSCALL_POST_FORK
+#    define COMMON_SYSCALL_POST_FORK(res) \
+      {}
+#  endif
 
 // FIXME: do some kind of PRE_READ for all syscall arguments (int(s) and such).
 
@@ -130,8 +132,8 @@ struct sanitizer_kernel_sockaddr {
 // Declare it "void" to catch sizeof(kernel_sigset_t).
 typedef void kernel_sigset_t;
 
-static void kernel_write_iovec(const __sanitizer_iovec *iovec,
-                        SIZE_T iovlen, SIZE_T maxlen) {
+static void kernel_write_iovec(const __sanitizer_iovec *iovec, SIZE_T iovlen,
+                               SIZE_T maxlen) {
   for (SIZE_T i = 0; i < iovlen && maxlen; ++i) {
     SSIZE_T sz = Min(iovec[i].iov_len, maxlen);
     POST_WRITE(iovec[i].iov_base, sz);
@@ -141,8 +143,8 @@ static void kernel_write_iovec(const __sanitizer_iovec *iovec,
 
 // This functions uses POST_READ, because it needs to run after syscall to know
 // the real read range.
-static void kernel_read_iovec(const __sanitizer_iovec *iovec,
-                       SIZE_T iovlen, SIZE_T maxlen) {
+static void kernel_read_iovec(const __sanitizer_iovec *iovec, SIZE_T iovlen,
+                              SIZE_T maxlen) {
   POST_READ(iovec, sizeof(*iovec) * iovlen);
   for (SIZE_T i = 0; i < iovlen && maxlen; ++i) {
     SSIZE_T sz = Min(iovec[i].iov_len, maxlen);
@@ -155,8 +157,8 @@ PRE_SYSCALL(recvmsg)(long sockfd, sanitizer_kernel_msghdr *msg, long flags) {
   PRE_READ(msg, sizeof(*msg));
 }
 
-POST_SYSCALL(recvmsg)(long res, long sockfd, sanitizer_kernel_msghdr *msg,
-                      long flags) {
+POST_SYSCALL(recvmsg)
+(long res, long sockfd, sanitizer_kernel_msghdr *msg, long flags) {
   if (res >= 0) {
     if (msg) {
       for (unsigned long i = 0; i < msg->msg_iovlen; ++i) {
@@ -167,13 +169,14 @@ POST_SYSCALL(recvmsg)(long res, long sockfd, sanitizer_kernel_msghdr *msg,
   }
 }
 
-PRE_SYSCALL(recvmmsg)(long fd, sanitizer_kernel_mmsghdr *msg, long vlen,
-                      long flags, void *timeout) {
+PRE_SYSCALL(recvmmsg)
+(long fd, sanitizer_kernel_mmsghdr *msg, long vlen, long flags, void *timeout) {
   PRE_READ(msg, vlen * sizeof(*msg));
 }
 
-POST_SYSCALL(recvmmsg)(long res, long fd, sanitizer_kernel_mmsghdr *msg,
-                       long vlen, long flags, void *timeout) {
+POST_SYSCALL(recvmmsg)
+(long res, long fd, sanitizer_kernel_mmsghdr *msg, long vlen, long flags,
+ void *timeout) {
   if (res >= 0) {
     if (msg) {
       for (unsigned long i = 0; i < msg->msg_hdr.msg_iovlen; ++i) {
@@ -183,7 +186,8 @@ POST_SYSCALL(recvmmsg)(long res, long fd, sanitizer_kernel_mmsghdr *msg,
       POST_WRITE(msg->msg_hdr.msg_control, msg->msg_hdr.msg_controllen);
       POST_WRITE(&msg->msg_len, sizeof(msg->msg_len));
     }
-    if (timeout) POST_WRITE(timeout, struct_timespec_sz);
+    if (timeout)
+      POST_WRITE(timeout, struct_timespec_sz);
   }
 }
 
@@ -203,7 +207,8 @@ PRE_SYSCALL(time)(void *tloc) {}
 
 POST_SYSCALL(time)(long res, void *tloc) {
   if (res >= 0) {
-    if (tloc) POST_WRITE(tloc, sizeof(long));
+    if (tloc)
+      POST_WRITE(tloc, sizeof(long));
   }
 }
 
@@ -211,7 +216,8 @@ PRE_SYSCALL(stime)(void *tptr) {}
 
 POST_SYSCALL(stime)(long res, void *tptr) {
   if (res >= 0) {
-    if (tptr) POST_WRITE(tptr, sizeof(long));
+    if (tptr)
+      POST_WRITE(tptr, sizeof(long));
   }
 }
 
@@ -219,8 +225,10 @@ PRE_SYSCALL(gettimeofday)(void *tv, void *tz) {}
 
 POST_SYSCALL(gettimeofday)(long res, void *tv, void *tz) {
   if (res >= 0) {
-    if (tv) POST_WRITE(tv, timeval_sz);
-    if (tz) POST_WRITE(tz, struct_timezone_sz);
+    if (tv)
+      POST_WRITE(tv, timeval_sz);
+    if (tz)
+      POST_WRITE(tz, struct_timezone_sz);
   }
 }
 
@@ -228,26 +236,30 @@ PRE_SYSCALL(settimeofday)(void *tv, void *tz) {}
 
 POST_SYSCALL(settimeofday)(long res, void *tv, void *tz) {
   if (res >= 0) {
-    if (tv) POST_WRITE(tv, timeval_sz);
-    if (tz) POST_WRITE(tz, struct_timezone_sz);
+    if (tv)
+      POST_WRITE(tv, timeval_sz);
+    if (tz)
+      POST_WRITE(tz, struct_timezone_sz);
   }
 }
 
-#if !SANITIZER_ANDROID
+#  if !SANITIZER_ANDROID
 PRE_SYSCALL(adjtimex)(void *txc_p) {}
 
 POST_SYSCALL(adjtimex)(long res, void *txc_p) {
   if (res >= 0) {
-    if (txc_p) POST_WRITE(txc_p, struct_timex_sz);
+    if (txc_p)
+      POST_WRITE(txc_p, struct_timex_sz);
   }
 }
-#endif
+#  endif
 
 PRE_SYSCALL(times)(void *tbuf) {}
 
 POST_SYSCALL(times)(long res, void *tbuf) {
   if (res >= 0) {
-    if (tbuf) POST_WRITE(tbuf, struct_tms_sz);
+    if (tbuf)
+      POST_WRITE(tbuf, struct_tms_sz);
   }
 }
 
@@ -259,8 +271,10 @@ PRE_SYSCALL(nanosleep)(void *rqtp, void *rmtp) {}
 
 POST_SYSCALL(nanosleep)(long res, void *rqtp, void *rmtp) {
   if (res >= 0) {
-    if (rqtp) POST_WRITE(rqtp, struct_timespec_sz);
-    if (rmtp) POST_WRITE(rmtp, struct_timespec_sz);
+    if (rqtp)
+      POST_WRITE(rqtp, struct_timespec_sz);
+    if (rmtp)
+      POST_WRITE(rmtp, struct_timespec_sz);
   }
 }
 
@@ -296,9 +310,12 @@ PRE_SYSCALL(getresuid)(void *ruid, void *euid, void *suid) {}
 
 POST_SYSCALL(getresuid)(long res, void *ruid, void *euid, void *suid) {
   if (res >= 0) {
-    if (ruid) POST_WRITE(ruid, sizeof(unsigned));
-    if (euid) POST_WRITE(euid, sizeof(unsigned));
-    if (suid) POST_WRITE(suid, sizeof(unsigned));
+    if (ruid)
+      POST_WRITE(ruid, sizeof(unsigned));
+    if (euid)
+      POST_WRITE(euid, sizeof(unsigned));
+    if (suid)
+      POST_WRITE(suid, sizeof(unsigned));
   }
 }
 
@@ -306,9 +323,12 @@ PRE_SYSCALL(getresgid)(void *rgid, void *egid, void *sgid) {}
 
 POST_SYSCALL(getresgid)(long res, void *rgid, void *egid, void *sgid) {
   if (res >= 0) {
-    if (rgid) POST_WRITE(rgid, sizeof(unsigned));
-    if (egid) POST_WRITE(egid, sizeof(unsigned));
-    if (sgid) POST_WRITE(sgid, sizeof(unsigned));
+    if (rgid)
+      POST_WRITE(rgid, sizeof(unsigned));
+    if (egid)
+      POST_WRITE(egid, sizeof(unsigned));
+    if (sgid)
+      POST_WRITE(sgid, sizeof(unsigned));
   }
 }
 
@@ -326,10 +346,11 @@ POST_SYSCALL(getsid)(long res, long pid) {}
 
 PRE_SYSCALL(getgroups)(long gidsetsize, void *grouplist) {}
 
-POST_SYSCALL(getgroups)(long res, long gidsetsize,
-                        __sanitizer___kernel_gid_t *grouplist) {
+POST_SYSCALL(getgroups)
+(long res, long gidsetsize, __sanitizer___kernel_gid_t *grouplist) {
   if (res >= 0) {
-    if (grouplist) POST_WRITE(grouplist, res * sizeof(*grouplist));
+    if (grouplist)
+      POST_WRITE(grouplist, res * sizeof(*grouplist));
   }
 }
 
@@ -374,11 +395,12 @@ PRE_SYSCALL(setsid)() {}
 POST_SYSCALL(setsid)(long res) {}
 
 PRE_SYSCALL(setgroups)(long gidsetsize, __sanitizer___kernel_gid_t *grouplist) {
-  if (grouplist) POST_WRITE(grouplist, gidsetsize * sizeof(*grouplist));
+  if (grouplist)
+    POST_WRITE(grouplist, gidsetsize * sizeof(*grouplist));
 }
 
-POST_SYSCALL(setgroups)(long res, long gidsetsize,
-                        __sanitizer___kernel_gid_t *grouplist) {}
+POST_SYSCALL(setgroups)
+(long res, long gidsetsize, __sanitizer___kernel_gid_t *grouplist) {}
 
 PRE_SYSCALL(acct)(const void *name) {
   if (name)
@@ -388,17 +410,21 @@ PRE_SYSCALL(acct)(const void *name) {
 POST_SYSCALL(acct)(long res, const void *name) {}
 
 PRE_SYSCALL(capget)(void *header, void *dataptr) {
-  if (header) PRE_READ(header, __user_cap_header_struct_sz);
+  if (header)
+    PRE_READ(header, __user_cap_header_struct_sz);
 }
 
 POST_SYSCALL(capget)(long res, void *header, void *dataptr) {
   if (res >= 0)
-    if (dataptr) POST_WRITE(dataptr, __user_cap_data_struct_sz);
+    if (dataptr)
+      POST_WRITE(dataptr, __user_cap_data_struct_sz);
 }
 
 PRE_SYSCALL(capset)(void *header, const void *data) {
-  if (header) PRE_READ(header, __user_cap_header_struct_sz);
-  if (data) PRE_READ(data, __user_cap_data_struct_sz);
+  if (header)
+    PRE_READ(header, __user_cap_header_struct_sz);
+  if (data)
+    PRE_READ(data, __user_cap_data_struct_sz);
 }
 
 POST_SYSCALL(capset)(long res, void *header, const void *data) {}
@@ -411,7 +437,8 @@ PRE_SYSCALL(sigpending)(void *set) {}
 
 POST_SYSCALL(sigpending)(long res, void *set) {
   if (res >= 0) {
-    if (set) POST_WRITE(set, old_sigset_t_sz);
+    if (set)
+      POST_WRITE(set, old_sigset_t_sz);
   }
 }
 
@@ -419,8 +446,10 @@ PRE_SYSCALL(sigprocmask)(long how, void *set, void *oset) {}
 
 POST_SYSCALL(sigprocmask)(long res, long how, void *set, void *oset) {
   if (res >= 0) {
-    if (set) POST_WRITE(set, old_sigset_t_sz);
-    if (oset) POST_WRITE(oset, old_sigset_t_sz);
+    if (set)
+      POST_WRITE(set, old_sigset_t_sz);
+    if (oset)
+      POST_WRITE(oset, old_sigset_t_sz);
   }
 }
 
@@ -428,7 +457,8 @@ PRE_SYSCALL(getitimer)(long which, void *value) {}
 
 POST_SYSCALL(getitimer)(long res, long which, void *value) {
   if (res >= 0) {
-    if (value) POST_WRITE(value, struct_itimerval_sz);
+    if (value)
+      POST_WRITE(value, struct_itimerval_sz);
   }
 }
 
@@ -436,19 +466,23 @@ PRE_SYSCALL(setitimer)(long which, void *value, void *ovalue) {}
 
 POST_SYSCALL(setitimer)(long res, long which, void *value, void *ovalue) {
   if (res >= 0) {
-    if (value) POST_WRITE(value, struct_itimerval_sz);
-    if (ovalue) POST_WRITE(ovalue, struct_itimerval_sz);
+    if (value)
+      POST_WRITE(value, struct_itimerval_sz);
+    if (ovalue)
+      POST_WRITE(ovalue, struct_itimerval_sz);
   }
 }
 
-PRE_SYSCALL(timer_create)(long which_clock, void *timer_event_spec,
-                          void *created_timer_id) {}
+PRE_SYSCALL(timer_create)
+(long which_clock, void *timer_event_spec, void *created_timer_id) {}
 
-POST_SYSCALL(timer_create)(long res, long which_clock, void *timer_event_spec,
-                           void *created_timer_id) {
+POST_SYSCALL(timer_create)
+(long res, long which_clock, void *timer_event_spec, void *created_timer_id) {
   if (res >= 0) {
-    if (timer_event_spec) POST_WRITE(timer_event_spec, struct_sigevent_sz);
-    if (created_timer_id) POST_WRITE(created_timer_id, sizeof(long));
+    if (timer_event_spec)
+      POST_WRITE(timer_event_spec, struct_sigevent_sz);
+    if (created_timer_id)
+      POST_WRITE(created_timer_id, sizeof(long));
   }
 }
 
@@ -456,7 +490,8 @@ PRE_SYSCALL(timer_gettime)(long timer_id, void *setting) {}
 
 POST_SYSCALL(timer_gettime)(long res, long timer_id, void *setting) {
   if (res >= 0) {
-    if (setting) POST_WRITE(setting, struct_itimerspec_sz);
+    if (setting)
+      POST_WRITE(setting, struct_itimerspec_sz);
   }
 }
 
@@ -464,15 +499,18 @@ PRE_SYSCALL(timer_getoverrun)(long timer_id) {}
 
 POST_SYSCALL(timer_getoverrun)(long res, long timer_id) {}
 
-PRE_SYSCALL(timer_settime)(long timer_id, long flags, const void *new_setting,
-                           void *old_setting) {
-  if (new_setting) PRE_READ(new_setting, struct_itimerspec_sz);
+PRE_SYSCALL(timer_settime)
+(long timer_id, long flags, const void *new_setting, void *old_setting) {
+  if (new_setting)
+    PRE_READ(new_setting, struct_itimerspec_sz);
 }
 
-POST_SYSCALL(timer_settime)(long res, long timer_id, long flags,
-                            const void *new_setting, void *old_setting) {
+POST_SYSCALL(timer_settime)
+(long res, long timer_id, long flags, const void *new_setting,
+ void *old_setting) {
   if (res >= 0) {
-    if (old_setting) POST_WRITE(old_setting, struct_itimerspec_sz);
+    if (old_setting)
+      POST_WRITE(old_setting, struct_itimerspec_sz);
   }
 }
 
@@ -481,7 +519,8 @@ PRE_SYSCALL(timer_delete)(long timer_id) {}
 POST_SYSCALL(timer_delete)(long res, long timer_id) {}
 
 PRE_SYSCALL(clock_settime)(long which_clock, const void *tp) {
-  if (tp) PRE_READ(tp, struct_timespec_sz);
+  if (tp)
+    PRE_READ(tp, struct_timespec_sz);
 }
 
 POST_SYSCALL(clock_settime)(long res, long which_clock, const void *tp) {}
@@ -490,37 +529,42 @@ PRE_SYSCALL(clock_gettime)(long which_clock, void *tp) {}
 
 POST_SYSCALL(clock_gettime)(long res, long which_clock, void *tp) {
   if (res >= 0) {
-    if (tp) POST_WRITE(tp, struct_timespec_sz);
+    if (tp)
+      POST_WRITE(tp, struct_timespec_sz);
   }
 }
 
-#if !SANITIZER_ANDROID
+#  if !SANITIZER_ANDROID
 PRE_SYSCALL(clock_adjtime)(long which_clock, void *tx) {}
 
 POST_SYSCALL(clock_adjtime)(long res, long which_clock, void *tx) {
   if (res >= 0) {
-    if (tx) POST_WRITE(tx, struct_timex_sz);
+    if (tx)
+      POST_WRITE(tx, struct_timex_sz);
   }
 }
-#endif
+#  endif
 
 PRE_SYSCALL(clock_getres)(long which_clock, void *tp) {}
 
 POST_SYSCALL(clock_getres)(long res, long which_clock, void *tp) {
   if (res >= 0) {
-    if (tp) POST_WRITE(tp, struct_timespec_sz);
+    if (tp)
+      POST_WRITE(tp, struct_timespec_sz);
   }
 }
 
-PRE_SYSCALL(clock_nanosleep)(long which_clock, long flags, const void *rqtp,
-                             void *rmtp) {
-  if (rqtp) PRE_READ(rqtp, struct_timespec_sz);
+PRE_SYSCALL(clock_nanosleep)
+(long which_clock, long flags, const void *rqtp, void *rmtp) {
+  if (rqtp)
+    PRE_READ(rqtp, struct_timespec_sz);
 }
 
-POST_SYSCALL(clock_nanosleep)(long res, long which_clock, long flags,
-                              const void *rqtp, void *rmtp) {
+POST_SYSCALL(clock_nanosleep)
+(long res, long which_clock, long flags, const void *rqtp, void *rmtp) {
   if (res >= 0) {
-    if (rmtp) POST_WRITE(rmtp, struct_timespec_sz);
+    if (rmtp)
+      POST_WRITE(rmtp, struct_timespec_sz);
   }
 }
 
@@ -532,12 +576,14 @@ PRE_SYSCALL(sched_setscheduler)(long pid, long policy, void *param) {}
 
 POST_SYSCALL(sched_setscheduler)(long res, long pid, long policy, void *param) {
   if (res >= 0) {
-    if (param) POST_WRITE(param, struct_sched_param_sz);
+    if (param)
+      POST_WRITE(param, struct_sched_param_sz);
   }
 }
 
 PRE_SYSCALL(sched_setparam)(long pid, void *param) {
-  if (param) PRE_READ(param, struct_sched_param_sz);
+  if (param)
+    PRE_READ(param, struct_sched_param_sz);
 }
 
 POST_SYSCALL(sched_setparam)(long res, long pid, void *param) {}
@@ -550,23 +596,26 @@ PRE_SYSCALL(sched_getparam)(long pid, void *param) {}
 
 POST_SYSCALL(sched_getparam)(long res, long pid, void *param) {
   if (res >= 0) {
-    if (param) POST_WRITE(param, struct_sched_param_sz);
+    if (param)
+      POST_WRITE(param, struct_sched_param_sz);
   }
 }
 
 PRE_SYSCALL(sched_setaffinity)(long pid, long len, void *user_mask_ptr) {
-  if (user_mask_ptr) PRE_READ(user_mask_ptr, len);
+  if (user_mask_ptr)
+    PRE_READ(user_mask_ptr, len);
 }
 
-POST_SYSCALL(sched_setaffinity)(long res, long pid, long len,
-                                void *user_mask_ptr) {}
+POST_SYSCALL(sched_setaffinity)
+(long res, long pid, long len, void *user_mask_ptr) {}
 
 PRE_SYSCALL(sched_getaffinity)(long pid, long len, void *user_mask_ptr) {}
 
-POST_SYSCALL(sched_getaffinity)(long res, long pid, long len,
-                                void *user_mask_ptr) {
+POST_SYSCALL(sched_getaffinity)
+(long res, long pid, long len, void *user_mask_ptr) {
   if (res >= 0) {
-    if (user_mask_ptr) POST_WRITE(user_mask_ptr, len);
+    if (user_mask_ptr)
+      POST_WRITE(user_mask_ptr, len);
   }
 }
 
@@ -586,7 +635,8 @@ PRE_SYSCALL(sched_rr_get_interval)(long pid, void *interval) {}
 
 POST_SYSCALL(sched_rr_get_interval)(long res, long pid, void *interval) {
   if (res >= 0) {
-    if (interval) POST_WRITE(interval, struct_timespec_sz);
+    if (interval)
+      POST_WRITE(interval, struct_timespec_sz);
   }
 }
 
@@ -610,13 +660,14 @@ PRE_SYSCALL(restart_syscall)() {}
 
 POST_SYSCALL(restart_syscall)(long res) {}
 
-PRE_SYSCALL(kexec_load)(long entry, long nr_segments, void *segments,
-                        long flags) {}
+PRE_SYSCALL(kexec_load)
+(long entry, long nr_segments, void *segments, long flags) {}
 
-POST_SYSCALL(kexec_load)(long res, long entry, long nr_segments, void *segments,
-                         long flags) {
+POST_SYSCALL(kexec_load)
+(long res, long entry, long nr_segments, void *segments, long flags) {
   if (res >= 0) {
-    if (segments) POST_WRITE(segments, struct_kexec_segment_sz);
+    if (segments)
+      POST_WRITE(segments, struct_kexec_segment_sz);
   }
 }
 
@@ -630,22 +681,26 @@ POST_SYSCALL(exit_group)(long res, long error_code) {}
 
 PRE_SYSCALL(wait4)(long pid, void *stat_addr, long options, void *ru) {}
 
-POST_SYSCALL(wait4)(long res, long pid, void *stat_addr, long options,
-                    void *ru) {
+POST_SYSCALL(wait4)
+(long res, long pid, void *stat_addr, long options, void *ru) {
   if (res >= 0) {
-    if (stat_addr) POST_WRITE(stat_addr, sizeof(int));
-    if (ru) POST_WRITE(ru, struct_rusage_sz);
+    if (stat_addr)
+      POST_WRITE(stat_addr, sizeof(int));
+    if (ru)
+      POST_WRITE(ru, struct_rusage_sz);
   }
 }
 
-PRE_SYSCALL(waitid)(long which, long pid, void *infop, long options, void *ru) {
-}
+PRE_SYSCALL(waitid)
+(long which, long pid, void *infop, long options, void *ru) {}
 
-POST_SYSCALL(waitid)(long res, long which, long pid, void *infop, long options,
-                     void *ru) {
+POST_SYSCALL(waitid)
+(long res, long which, long pid, void *infop, long options, void *ru) {
   if (res >= 0) {
-    if (infop) POST_WRITE(infop, siginfo_t_sz);
-    if (ru) POST_WRITE(ru, struct_rusage_sz);
+    if (infop)
+      POST_WRITE(infop, siginfo_t_sz);
+    if (ru)
+      POST_WRITE(ru, struct_rusage_sz);
   }
 }
 
@@ -653,7 +708,8 @@ PRE_SYSCALL(waitpid)(long pid, void *stat_addr, long options) {}
 
 POST_SYSCALL(waitpid)(long res, long pid, void *stat_addr, long options) {
   if (res >= 0) {
-    if (stat_addr) POST_WRITE(stat_addr, sizeof(int));
+    if (stat_addr)
+      POST_WRITE(stat_addr, sizeof(int));
   }
 }
 
@@ -661,7 +717,8 @@ PRE_SYSCALL(set_tid_address)(void *tidptr) {}
 
 POST_SYSCALL(set_tid_address)(long res, void *tidptr) {
   if (res >= 0) {
-    if (tidptr) POST_WRITE(tidptr, sizeof(int));
+    if (tidptr)
+      POST_WRITE(tidptr, sizeof(int));
   }
 }
 
@@ -682,11 +739,14 @@ POST_SYSCALL(delete_module)(long res, const void *name_user, long flags) {}
 
 PRE_SYSCALL(rt_sigprocmask)(long how, void *set, void *oset, long sigsetsize) {}
 
-POST_SYSCALL(rt_sigprocmask)(long res, long how, kernel_sigset_t *set,
-                             kernel_sigset_t *oset, long sigsetsize) {
+POST_SYSCALL(rt_sigprocmask)
+(long res, long how, kernel_sigset_t *set, kernel_sigset_t *oset,
+ long sigsetsize) {
   if (res >= 0) {
-    if (set) POST_WRITE(set, sigsetsize);
-    if (oset) POST_WRITE(oset, sigsetsize);
+    if (set)
+      POST_WRITE(set, sigsetsize);
+    if (oset)
+      POST_WRITE(oset, sigsetsize);
   }
 }
 
@@ -694,29 +754,34 @@ PRE_SYSCALL(rt_sigpending)(void *set, long sigsetsize) {}
 
 POST_SYSCALL(rt_sigpending)(long res, kernel_sigset_t *set, long sigsetsize) {
   if (res >= 0) {
-    if (set) POST_WRITE(set, sigsetsize);
+    if (set)
+      POST_WRITE(set, sigsetsize);
   }
 }
 
-PRE_SYSCALL(rt_sigtimedwait)(const kernel_sigset_t *uthese, void *uinfo,
-                             const void *uts, long sigsetsize) {
-  if (uthese) PRE_READ(uthese, sigsetsize);
-  if (uts) PRE_READ(uts, struct_timespec_sz);
+PRE_SYSCALL(rt_sigtimedwait)
+(const kernel_sigset_t *uthese, void *uinfo, const void *uts, long sigsetsize) {
+  if (uthese)
+    PRE_READ(uthese, sigsetsize);
+  if (uts)
+    PRE_READ(uts, struct_timespec_sz);
 }
 
-POST_SYSCALL(rt_sigtimedwait)(long res, const void *uthese, void *uinfo,
-                              const void *uts, long sigsetsize) {
+POST_SYSCALL(rt_sigtimedwait)
+(long res, const void *uthese, void *uinfo, const void *uts, long sigsetsize) {
   if (res >= 0) {
-    if (uinfo) POST_WRITE(uinfo, siginfo_t_sz);
+    if (uinfo)
+      POST_WRITE(uinfo, siginfo_t_sz);
   }
 }
 
 PRE_SYSCALL(rt_tgsigqueueinfo)(long tgid, long pid, long sig, void *uinfo) {}
 
-POST_SYSCALL(rt_tgsigqueueinfo)(long res, long tgid, long pid, long sig,
-                                void *uinfo) {
+POST_SYSCALL(rt_tgsigqueueinfo)
+(long res, long tgid, long pid, long sig, void *uinfo) {
   if (res >= 0) {
-    if (uinfo) POST_WRITE(uinfo, siginfo_t_sz);
+    if (uinfo)
+      POST_WRITE(uinfo, siginfo_t_sz);
   }
 }
 
@@ -736,7 +801,8 @@ PRE_SYSCALL(rt_sigqueueinfo)(long pid, long sig, void *uinfo) {}
 
 POST_SYSCALL(rt_sigqueueinfo)(long res, long pid, long sig, void *uinfo) {
   if (res >= 0) {
-    if (uinfo) POST_WRITE(uinfo, siginfo_t_sz);
+    if (uinfo)
+      POST_WRITE(uinfo, siginfo_t_sz);
   }
 }
 
@@ -772,11 +838,11 @@ PRE_SYSCALL(bdflush)(long func, long data) {}
 
 POST_SYSCALL(bdflush)(long res, long func, long data) {}
 
-PRE_SYSCALL(mount)(void *dev_name, void *dir_name, void *type, long flags,
-                   void *data) {}
+PRE_SYSCALL(mount)
+(void *dev_name, void *dir_name, void *type, long flags, void *data) {}
 
-POST_SYSCALL(mount)(long res, void *dev_name, void *dir_name, void *type,
-                    long flags, void *data) {
+POST_SYSCALL(mount)
+(long res, void *dev_name, void *dir_name, void *type, long flags, void *data) {
   if (res >= 0) {
     if (dev_name)
       POST_WRITE(dev_name,
@@ -826,11 +892,12 @@ PRE_SYSCALL(stat)(const void *filename, void *statbuf) {
 
 POST_SYSCALL(stat)(long res, const void *filename, void *statbuf) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct___old_kernel_stat_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct___old_kernel_stat_sz);
   }
 }
 
-#if !SANITIZER_ANDROID
+#  if !SANITIZER_ANDROID
 PRE_SYSCALL(statfs)(const void *path, void *buf) {
   if (path)
     PRE_READ(path, __sanitizer::internal_strlen((const char *)path) + 1);
@@ -838,7 +905,8 @@ PRE_SYSCALL(statfs)(const void *path, void *buf) {
 
 POST_SYSCALL(statfs)(long res, const void *path, void *buf) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, struct_statfs_sz);
+    if (buf)
+      POST_WRITE(buf, struct_statfs_sz);
   }
 }
 
@@ -849,7 +917,8 @@ PRE_SYSCALL(statfs64)(const void *path, long sz, void *buf) {
 
 POST_SYSCALL(statfs64)(long res, const void *path, long sz, void *buf) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, struct_statfs64_sz);
+    if (buf)
+      POST_WRITE(buf, struct_statfs64_sz);
   }
 }
 
@@ -857,7 +926,8 @@ PRE_SYSCALL(fstatfs)(long fd, void *buf) {}
 
 POST_SYSCALL(fstatfs)(long res, long fd, void *buf) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, struct_statfs_sz);
+    if (buf)
+      POST_WRITE(buf, struct_statfs_sz);
   }
 }
 
@@ -865,10 +935,11 @@ PRE_SYSCALL(fstatfs64)(long fd, long sz, void *buf) {}
 
 POST_SYSCALL(fstatfs64)(long res, long fd, long sz, void *buf) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, struct_statfs64_sz);
+    if (buf)
+      POST_WRITE(buf, struct_statfs64_sz);
   }
 }
-#endif // !SANITIZER_ANDROID
+#  endif  // !SANITIZER_ANDROID
 
 PRE_SYSCALL(lstat)(const void *filename, void *statbuf) {
   if (filename)
@@ -878,7 +949,8 @@ PRE_SYSCALL(lstat)(const void *filename, void *statbuf) {
 
 POST_SYSCALL(lstat)(long res, const void *filename, void *statbuf) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct___old_kernel_stat_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct___old_kernel_stat_sz);
   }
 }
 
@@ -886,7 +958,8 @@ PRE_SYSCALL(fstat)(long fd, void *statbuf) {}
 
 POST_SYSCALL(fstat)(long res, long fd, void *statbuf) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct___old_kernel_stat_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct___old_kernel_stat_sz);
   }
 }
 
@@ -898,7 +971,8 @@ PRE_SYSCALL(newstat)(const void *filename, void *statbuf) {
 
 POST_SYSCALL(newstat)(long res, const void *filename, void *statbuf) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct_kernel_stat_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct_kernel_stat_sz);
   }
 }
 
@@ -910,7 +984,8 @@ PRE_SYSCALL(newlstat)(const void *filename, void *statbuf) {
 
 POST_SYSCALL(newlstat)(long res, const void *filename, void *statbuf) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct_kernel_stat_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct_kernel_stat_sz);
   }
 }
 
@@ -918,19 +993,21 @@ PRE_SYSCALL(newfstat)(long fd, void *statbuf) {}
 
 POST_SYSCALL(newfstat)(long res, long fd, void *statbuf) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct_kernel_stat_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct_kernel_stat_sz);
   }
 }
 
-#if !SANITIZER_ANDROID
+#  if !SANITIZER_ANDROID
 PRE_SYSCALL(ustat)(long dev, void *ubuf) {}
 
 POST_SYSCALL(ustat)(long res, long dev, void *ubuf) {
   if (res >= 0) {
-    if (ubuf) POST_WRITE(ubuf, struct_ustat_sz);
+    if (ubuf)
+      POST_WRITE(ubuf, struct_ustat_sz);
   }
 }
-#endif  // !SANITIZER_ANDROID
+#  endif  // !SANITIZER_ANDROID
 
 PRE_SYSCALL(stat64)(const void *filename, void *statbuf) {
   if (filename)
@@ -940,7 +1017,8 @@ PRE_SYSCALL(stat64)(const void *filename, void *statbuf) {
 
 POST_SYSCALL(stat64)(long res, const void *filename, void *statbuf) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct_kernel_stat64_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct_kernel_stat64_sz);
   }
 }
 
@@ -948,7 +1026,8 @@ PRE_SYSCALL(fstat64)(long fd, void *statbuf) {}
 
 POST_SYSCALL(fstat64)(long res, long fd, void *statbuf) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct_kernel_stat64_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct_kernel_stat64_sz);
   }
 }
 
@@ -960,71 +1039,80 @@ PRE_SYSCALL(lstat64)(const void *filename, void *statbuf) {
 
 POST_SYSCALL(lstat64)(long res, const void *filename, void *statbuf) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct_kernel_stat64_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct_kernel_stat64_sz);
   }
 }
 
-PRE_SYSCALL(setxattr)(const void *path, const void *name, const void *value,
-                      long size, long flags) {
+PRE_SYSCALL(setxattr)
+(const void *path, const void *name, const void *value, long size, long flags) {
   if (path)
     PRE_READ(path, __sanitizer::internal_strlen((const char *)path) + 1);
   if (name)
     PRE_READ(name, __sanitizer::internal_strlen((const char *)name) + 1);
-  if (value) PRE_READ(value, size);
+  if (value)
+    PRE_READ(value, size);
 }
 
-POST_SYSCALL(setxattr)(long res, const void *path, const void *name,
-                       const void *value, long size, long flags) {}
+POST_SYSCALL(setxattr)
+(long res, const void *path, const void *name, const void *value, long size,
+ long flags) {}
 
-PRE_SYSCALL(lsetxattr)(const void *path, const void *name, const void *value,
-                       long size, long flags) {
+PRE_SYSCALL(lsetxattr)
+(const void *path, const void *name, const void *value, long size, long flags) {
   if (path)
     PRE_READ(path, __sanitizer::internal_strlen((const char *)path) + 1);
   if (name)
     PRE_READ(name, __sanitizer::internal_strlen((const char *)name) + 1);
-  if (value) PRE_READ(value, size);
+  if (value)
+    PRE_READ(value, size);
 }
 
-POST_SYSCALL(lsetxattr)(long res, const void *path, const void *name,
-                        const void *value, long size, long flags) {}
+POST_SYSCALL(lsetxattr)
+(long res, const void *path, const void *name, const void *value, long size,
+ long flags) {}
 
-PRE_SYSCALL(fsetxattr)(long fd, const void *name, const void *value, long size,
-                       long flags) {
+PRE_SYSCALL(fsetxattr)
+(long fd, const void *name, const void *value, long size, long flags) {
   if (name)
     PRE_READ(name, __sanitizer::internal_strlen((const char *)name) + 1);
-  if (value) PRE_READ(value, size);
+  if (value)
+    PRE_READ(value, size);
 }
 
-POST_SYSCALL(fsetxattr)(long res, long fd, const void *name, const void *value,
-                        long size, long flags) {}
+POST_SYSCALL(fsetxattr)
+(long res, long fd, const void *name, const void *value, long size,
+ long flags) {}
 
-PRE_SYSCALL(getxattr)(const void *path, const void *name, void *value,
-                      long size) {
+PRE_SYSCALL(getxattr)
+(const void *path, const void *name, void *value, long size) {
   if (path)
     PRE_READ(path, __sanitizer::internal_strlen((const char *)path) + 1);
   if (name)
     PRE_READ(name, __sanitizer::internal_strlen((const char *)name) + 1);
 }
 
-POST_SYSCALL(getxattr)(long res, const void *path, const void *name,
-                       void *value, long size) {
+POST_SYSCALL(getxattr)
+(long res, const void *path, const void *name, void *value, long size) {
   if (size && res > 0) {
-    if (value) POST_WRITE(value, res);
+    if (value)
+      POST_WRITE(value, res);
   }
 }
 
-PRE_SYSCALL(lgetxattr)(const void *path, const void *name, void *value,
-                       long size) {
+PRE_SYSCALL(lgetxattr)
+(const void *path, const void *name, void *value, long size) {
   if (path)
     PRE_READ(path, __sanitizer::internal_strlen((const char *)path) + 1);
   if (name)
     PRE_READ(name, __sanitizer::internal_strlen((const char *)name) + 1);
 }
 
-POST_SYSCALL(lgetxattr)(long res, const void *path, const void *name,
-                        void *value, long size) {
+POST_SYSCALL(lgetxattr)
+(long res, const void *path, const void *name, void *value, long size) {
   if (size && res > 0) {
-    if (value) POST_WRITE(value, res);
+    if (value)
+      POST_WRITE(value, res);
   }
 }
 
@@ -1033,10 +1121,11 @@ PRE_SYSCALL(fgetxattr)(long fd, const void *name, void *value, long size) {
     PRE_READ(name, __sanitizer::internal_strlen((const char *)name) + 1);
 }
 
-POST_SYSCALL(fgetxattr)(long res, long fd, const void *name, void *value,
-                        long size) {
+POST_SYSCALL(fgetxattr)
+(long res, long fd, const void *name, void *value, long size) {
   if (size && res > 0) {
-    if (value) POST_WRITE(value, res);
+    if (value)
+      POST_WRITE(value, res);
   }
 }
 
@@ -1047,7 +1136,8 @@ PRE_SYSCALL(listxattr)(const void *path, void *list, long size) {
 
 POST_SYSCALL(listxattr)(long res, const void *path, void *list, long size) {
   if (size && res > 0) {
-    if (list) POST_WRITE(list, res);
+    if (list)
+      POST_WRITE(list, res);
   }
 }
 
@@ -1058,7 +1148,8 @@ PRE_SYSCALL(llistxattr)(const void *path, void *list, long size) {
 
 POST_SYSCALL(llistxattr)(long res, const void *path, void *list, long size) {
   if (size && res > 0) {
-    if (list) POST_WRITE(list, res);
+    if (list)
+      POST_WRITE(list, res);
   }
 }
 
@@ -1066,7 +1157,8 @@ PRE_SYSCALL(flistxattr)(long fd, void *list, long size) {}
 
 POST_SYSCALL(flistxattr)(long res, long fd, void *list, long size) {
   if (size && res > 0) {
-    if (list) POST_WRITE(list, res);
+    if (list)
+      POST_WRITE(list, res);
   }
 }
 
@@ -1103,17 +1195,17 @@ PRE_SYSCALL(mprotect)(long start, long len, long prot) {}
 
 POST_SYSCALL(mprotect)(long res, long start, long len, long prot) {}
 
-PRE_SYSCALL(mremap)(long addr, long old_len, long new_len, long flags,
-                    long new_addr) {}
+PRE_SYSCALL(mremap)
+(long addr, long old_len, long new_len, long flags, long new_addr) {}
 
-POST_SYSCALL(mremap)(long res, long addr, long old_len, long new_len,
-                     long flags, long new_addr) {}
+POST_SYSCALL(mremap)
+(long res, long addr, long old_len, long new_len, long flags, long new_addr) {}
 
-PRE_SYSCALL(remap_file_pages)(long start, long size, long prot, long pgoff,
-                              long flags) {}
+PRE_SYSCALL(remap_file_pages)
+(long start, long size, long prot, long pgoff, long flags) {}
 
-POST_SYSCALL(remap_file_pages)(long res, long start, long size, long prot,
-                               long pgoff, long flags) {}
+POST_SYSCALL(remap_file_pages)
+(long res, long start, long size, long prot, long pgoff, long flags) {}
 
 PRE_SYSCALL(msync)(long start, long len, long flags) {}
 
@@ -1189,7 +1281,8 @@ PRE_SYSCALL(link)(const void *oldname, const void *newname) {
 POST_SYSCALL(link)(long res, const void *oldname, const void *newname) {}
 
 PRE_SYSCALL(symlink)(const void *old, const void *new_) {
-  if (old) PRE_READ(old, __sanitizer::internal_strlen((const char *)old) + 1);
+  if (old)
+    PRE_READ(old, __sanitizer::internal_strlen((const char *)old) + 1);
   if (new_)
     PRE_READ(new_, __sanitizer::internal_strlen((const char *)new_) + 1);
 }
@@ -1237,14 +1330,16 @@ PRE_SYSCALL(pipe)(void *fildes) {}
 
 POST_SYSCALL(pipe)(long res, void *fildes) {
   if (res >= 0)
-    if (fildes) POST_WRITE(fildes, sizeof(int) * 2);
+    if (fildes)
+      POST_WRITE(fildes, sizeof(int) * 2);
 }
 
 PRE_SYSCALL(pipe2)(void *fildes, long flags) {}
 
 POST_SYSCALL(pipe2)(long res, void *fildes, long flags) {
   if (res >= 0)
-    if (fildes) POST_WRITE(fildes, sizeof(int) * 2);
+    if (fildes)
+      POST_WRITE(fildes, sizeof(int) * 2);
 }
 
 PRE_SYSCALL(dup)(long fildes) {}
@@ -1272,16 +1367,19 @@ PRE_SYSCALL(flock)(long fd, long cmd) {}
 POST_SYSCALL(flock)(long res, long fd, long cmd) {}
 
 PRE_SYSCALL(io_setup)(long nr_reqs, void **ctx) {
-  if (ctx) PRE_WRITE(ctx, sizeof(*ctx));
+  if (ctx)
+    PRE_WRITE(ctx, sizeof(*ctx));
 }
 
 POST_SYSCALL(io_setup)(long res, long nr_reqs, void **ctx) {
   if (res >= 0) {
-    if (ctx) POST_WRITE(ctx, sizeof(*ctx));
+    if (ctx)
+      POST_WRITE(ctx, sizeof(*ctx));
     // (*ctx) is actually a pointer to a kernel mapped page, and there are
     // people out there who are crazy enough to peek into that page's 32-byte
     // header.
-    if (*ctx) POST_WRITE(*ctx, 32);
+    if (*ctx)
+      POST_WRITE(*ctx, 32);
   }
 }
 
@@ -1289,16 +1387,21 @@ PRE_SYSCALL(io_destroy)(long ctx) {}
 
 POST_SYSCALL(io_destroy)(long res, long ctx) {}
 
-PRE_SYSCALL(io_getevents)(long ctx_id, long min_nr, long nr,
-                          __sanitizer_io_event *ioevpp, void *timeout) {
-  if (timeout) PRE_READ(timeout, struct_timespec_sz);
+PRE_SYSCALL(io_getevents)
+(long ctx_id, long min_nr, long nr, __sanitizer_io_event *ioevpp,
+ void *timeout) {
+  if (timeout)
+    PRE_READ(timeout, struct_timespec_sz);
 }
 
-POST_SYSCALL(io_getevents)(long res, long ctx_id, long min_nr, long nr,
-                           __sanitizer_io_event *ioevpp, void *timeout) {
+POST_SYSCALL(io_getevents)
+(long res, long ctx_id, long min_nr, long nr, __sanitizer_io_event *ioevpp,
+ void *timeout) {
   if (res >= 0) {
-    if (ioevpp) POST_WRITE(ioevpp, res * sizeof(*ioevpp));
-    if (timeout) POST_WRITE(timeout, struct_timespec_sz);
+    if (ioevpp)
+      POST_WRITE(ioevpp, res * sizeof(*ioevpp));
+    if (timeout)
+      POST_WRITE(timeout, struct_timespec_sz);
   }
   for (long i = 0; i < res; i++) {
     // We synchronize io_submit -> io_getevents/io_cancel using the
@@ -1308,26 +1411,26 @@ POST_SYSCALL(io_getevents)(long res, long ctx_id, long min_nr, long nr,
     // synchronize on 0. But there does not seem to be a better solution
     // (except wrapping all operations in own context, which is unreliable).
     // We can not reliably extract fildes in io_getevents.
-    COMMON_SYSCALL_ACQUIRE((void*)ioevpp[i].data);
+    COMMON_SYSCALL_ACQUIRE((void *)ioevpp[i].data);
   }
 }
 
 PRE_SYSCALL(io_submit)(long ctx_id, long nr, __sanitizer_iocb **iocbpp) {
   for (long i = 0; i < nr; ++i) {
     uptr op = iocbpp[i]->aio_lio_opcode;
-    void *data = (void*)iocbpp[i]->aio_data;
-    void *buf = (void*)iocbpp[i]->aio_buf;
+    void *data = (void *)iocbpp[i]->aio_data;
+    void *buf = (void *)iocbpp[i]->aio_buf;
     uptr len = (uptr)iocbpp[i]->aio_nbytes;
     if (op == iocb_cmd_pwrite && buf && len) {
       PRE_READ(buf, len);
     } else if (op == iocb_cmd_pread && buf && len) {
       POST_WRITE(buf, len);
     } else if (op == iocb_cmd_pwritev) {
-      __sanitizer_iovec *iovec = (__sanitizer_iovec*)buf;
+      __sanitizer_iovec *iovec = (__sanitizer_iovec *)buf;
       for (uptr v = 0; v < len; v++)
         PRE_READ(iovec[v].iov_base, iovec[v].iov_len);
     } else if (op == iocb_cmd_preadv) {
-      __sanitizer_iovec *iovec = (__sanitizer_iovec*)buf;
+      __sanitizer_iovec *iovec = (__sanitizer_iovec *)buf;
       for (uptr v = 0; v < len; v++)
         POST_WRITE(iovec[v].iov_base, iovec[v].iov_len);
     }
@@ -1336,19 +1439,18 @@ PRE_SYSCALL(io_submit)(long ctx_id, long nr, __sanitizer_iocb **iocbpp) {
   }
 }
 
-POST_SYSCALL(io_submit)(long res, long ctx_id, long nr,
-    __sanitizer_iocb **iocbpp) {}
+POST_SYSCALL(io_submit)
+(long res, long ctx_id, long nr, __sanitizer_iocb **iocbpp) {}
 
-PRE_SYSCALL(io_cancel)(long ctx_id, __sanitizer_iocb *iocb,
-    __sanitizer_io_event *result) {
-}
+PRE_SYSCALL(io_cancel)
+(long ctx_id, __sanitizer_iocb *iocb, __sanitizer_io_event *result) {}
 
-POST_SYSCALL(io_cancel)(long res, long ctx_id, __sanitizer_iocb *iocb,
-    __sanitizer_io_event *result) {
+POST_SYSCALL(io_cancel)
+(long res, long ctx_id, __sanitizer_iocb *iocb, __sanitizer_io_event *result) {
   if (res == 0) {
     if (result) {
       // See comment in io_getevents.
-      COMMON_SYSCALL_ACQUIRE((void*)result->data);
+      COMMON_SYSCALL_ACQUIRE((void *)result->data);
       POST_WRITE(result, sizeof(*result));
     }
     if (iocb)
@@ -1358,19 +1460,23 @@ POST_SYSCALL(io_cancel)(long res, long ctx_id, __sanitizer_iocb *iocb,
 
 PRE_SYSCALL(sendfile)(long out_fd, long in_fd, void *offset, long count) {}
 
-POST_SYSCALL(sendfile)(long res, long out_fd, long in_fd,
-                       __sanitizer___kernel_off_t *offset, long count) {
+POST_SYSCALL(sendfile)
+(long res, long out_fd, long in_fd, __sanitizer___kernel_off_t *offset,
+ long count) {
   if (res >= 0) {
-    if (offset) POST_WRITE(offset, sizeof(*offset));
+    if (offset)
+      POST_WRITE(offset, sizeof(*offset));
   }
 }
 
 PRE_SYSCALL(sendfile64)(long out_fd, long in_fd, void *offset, long count) {}
 
-POST_SYSCALL(sendfile64)(long res, long out_fd, long in_fd,
-                         __sanitizer___kernel_loff_t *offset, long count) {
+POST_SYSCALL(sendfile64)
+(long res, long out_fd, long in_fd, __sanitizer___kernel_loff_t *offset,
+ long count) {
   if (res >= 0) {
-    if (offset) POST_WRITE(offset, sizeof(*offset));
+    if (offset)
+      POST_WRITE(offset, sizeof(*offset));
   }
 }
 
@@ -1402,9 +1508,7 @@ PRE_SYSCALL(open)(const void *filename, long flags, long mode) {
 
 POST_SYSCALL(open)(long res, const void *filename, long flags, long mode) {}
 
-PRE_SYSCALL(close)(long fd) {
-  COMMON_SYSCALL_FD_CLOSE((int)fd);
-}
+PRE_SYSCALL(close)(long fd) { COMMON_SYSCALL_FD_CLOSE((int)fd); }
 
 POST_SYSCALL(close)(long res, long fd) {}
 
@@ -1440,7 +1544,7 @@ PRE_SYSCALL(fchown)(long fd, long user, long group) {}
 
 POST_SYSCALL(fchown)(long res, long fd, long user, long group) {}
 
-#if SANITIZER_USES_UID16_SYSCALLS
+#  if SANITIZER_USES_UID16_SYSCALLS
 PRE_SYSCALL(chown16)(const void *filename, long user, long group) {
   if (filename)
     PRE_READ(filename,
@@ -1483,13 +1587,16 @@ POST_SYSCALL(setresuid16)(long res, long ruid, long euid, long suid) {}
 
 PRE_SYSCALL(getresuid16)(void *ruid, void *euid, void *suid) {}
 
-POST_SYSCALL(getresuid16)(long res, __sanitizer___kernel_old_uid_t *ruid,
-                          __sanitizer___kernel_old_uid_t *euid,
-                          __sanitizer___kernel_old_uid_t *suid) {
+POST_SYSCALL(getresuid16)
+(long res, __sanitizer___kernel_old_uid_t *ruid,
+ __sanitizer___kernel_old_uid_t *euid, __sanitizer___kernel_old_uid_t *suid) {
   if (res >= 0) {
-    if (ruid) POST_WRITE(ruid, sizeof(*ruid));
-    if (euid) POST_WRITE(euid, sizeof(*euid));
-    if (suid) POST_WRITE(suid, sizeof(*suid));
+    if (ruid)
+      POST_WRITE(ruid, sizeof(*ruid));
+    if (euid)
+      POST_WRITE(euid, sizeof(*euid));
+    if (suid)
+      POST_WRITE(suid, sizeof(*suid));
   }
 }
 
@@ -1499,13 +1606,16 @@ POST_SYSCALL(setresgid16)(long res, long rgid, long egid, long sgid) {}
 
 PRE_SYSCALL(getresgid16)(void *rgid, void *egid, void *sgid) {}
 
-POST_SYSCALL(getresgid16)(long res, __sanitizer___kernel_old_gid_t *rgid,
-                          __sanitizer___kernel_old_gid_t *egid,
-                          __sanitizer___kernel_old_gid_t *sgid) {
+POST_SYSCALL(getresgid16)
+(long res, __sanitizer___kernel_old_gid_t *rgid,
+ __sanitizer___kernel_old_gid_t *egid, __sanitizer___kernel_old_gid_t *sgid) {
   if (res >= 0) {
-    if (rgid) POST_WRITE(rgid, sizeof(*rgid));
-    if (egid) POST_WRITE(egid, sizeof(*egid));
-    if (sgid) POST_WRITE(sgid, sizeof(*sgid));
+    if (rgid)
+      POST_WRITE(rgid, sizeof(*rgid));
+    if (egid)
+      POST_WRITE(egid, sizeof(*egid));
+    if (sgid)
+      POST_WRITE(sgid, sizeof(*sgid));
   }
 }
 
@@ -1517,23 +1627,25 @@ PRE_SYSCALL(setfsgid16)(long gid) {}
 
 POST_SYSCALL(setfsgid16)(long res, long gid) {}
 
-PRE_SYSCALL(getgroups16)(long gidsetsize,
-                         __sanitizer___kernel_old_gid_t *grouplist) {}
+PRE_SYSCALL(getgroups16)
+(long gidsetsize, __sanitizer___kernel_old_gid_t *grouplist) {}
 
-POST_SYSCALL(getgroups16)(long res, long gidsetsize,
-                          __sanitizer___kernel_old_gid_t *grouplist) {
+POST_SYSCALL(getgroups16)
+(long res, long gidsetsize, __sanitizer___kernel_old_gid_t *grouplist) {
   if (res >= 0) {
-    if (grouplist) POST_WRITE(grouplist, res * sizeof(*grouplist));
+    if (grouplist)
+      POST_WRITE(grouplist, res * sizeof(*grouplist));
   }
 }
 
-PRE_SYSCALL(setgroups16)(long gidsetsize,
-                         __sanitizer___kernel_old_gid_t *grouplist) {
-  if (grouplist) POST_WRITE(grouplist, gidsetsize * sizeof(*grouplist));
+PRE_SYSCALL(setgroups16)
+(long gidsetsize, __sanitizer___kernel_old_gid_t *grouplist) {
+  if (grouplist)
+    POST_WRITE(grouplist, gidsetsize * sizeof(*grouplist));
 }
 
-POST_SYSCALL(setgroups16)(long res, long gidsetsize,
-                          __sanitizer___kernel_old_gid_t *grouplist) {}
+POST_SYSCALL(setgroups16)
+(long res, long gidsetsize, __sanitizer___kernel_old_gid_t *grouplist) {}
 
 PRE_SYSCALL(getuid16)() {}
 
@@ -1550,7 +1662,7 @@ POST_SYSCALL(getgid16)(long res) {}
 PRE_SYSCALL(getegid16)() {}
 
 POST_SYSCALL(getegid16)(long res) {}
-#endif // SANITIZER_USES_UID16_SYSCALLS
+#  endif  // SANITIZER_USES_UID16_SYSCALLS
 
 PRE_SYSCALL(utime)(void *filename, void *times) {}
 
@@ -1559,7 +1671,8 @@ POST_SYSCALL(utime)(long res, void *filename, void *times) {
     if (filename)
       POST_WRITE(filename,
                  __sanitizer::internal_strlen((const char *)filename) + 1);
-    if (times) POST_WRITE(times, struct_utimbuf_sz);
+    if (times)
+      POST_WRITE(times, struct_utimbuf_sz);
   }
 }
 
@@ -1570,7 +1683,8 @@ POST_SYSCALL(utimes)(long res, void *filename, void *utimes) {
     if (filename)
       POST_WRITE(filename,
                  __sanitizer::internal_strlen((const char *)filename) + 1);
-    if (utimes) POST_WRITE(utimes, timeval_sz);
+    if (utimes)
+      POST_WRITE(utimes, timeval_sz);
   }
 }
 
@@ -1578,91 +1692,104 @@ PRE_SYSCALL(lseek)(long fd, long offset, long origin) {}
 
 POST_SYSCALL(lseek)(long res, long fd, long offset, long origin) {}
 
-PRE_SYSCALL(llseek)(long fd, long offset_high, long offset_low, void *result,
-                    long origin) {}
+PRE_SYSCALL(llseek)
+(long fd, long offset_high, long offset_low, void *result, long origin) {}
 
-POST_SYSCALL(llseek)(long res, long fd, long offset_high, long offset_low,
-                     void *result, long origin) {
+POST_SYSCALL(llseek)
+(long res, long fd, long offset_high, long offset_low, void *result,
+ long origin) {
   if (res >= 0) {
-    if (result) POST_WRITE(result, sizeof(long long));
+    if (result)
+      POST_WRITE(result, sizeof(long long));
   }
 }
 
 PRE_SYSCALL(readv)(long fd, const __sanitizer_iovec *vec, long vlen) {}
 
-POST_SYSCALL(readv)(long res, long fd, const __sanitizer_iovec *vec,
-                    long vlen) {
+POST_SYSCALL(readv)
+(long res, long fd, const __sanitizer_iovec *vec, long vlen) {
   if (res >= 0) {
-    if (vec) kernel_write_iovec(vec, vlen, res);
+    if (vec)
+      kernel_write_iovec(vec, vlen, res);
   }
 }
 
 PRE_SYSCALL(write)(long fd, const void *buf, long count) {
-  if (buf) PRE_READ(buf, count);
+  if (buf)
+    PRE_READ(buf, count);
 }
 
 POST_SYSCALL(write)(long res, long fd, const void *buf, long count) {}
 
 PRE_SYSCALL(writev)(long fd, const __sanitizer_iovec *vec, long vlen) {}
 
-POST_SYSCALL(writev)(long res, long fd, const __sanitizer_iovec *vec,
-                     long vlen) {
+POST_SYSCALL(writev)
+(long res, long fd, const __sanitizer_iovec *vec, long vlen) {
   if (res >= 0) {
-    if (vec) kernel_read_iovec(vec, vlen, res);
+    if (vec)
+      kernel_read_iovec(vec, vlen, res);
   }
 }
 
-#ifdef _LP64
+#  ifdef _LP64
 PRE_SYSCALL(pread64)(long fd, void *buf, long count, long pos) {}
 
 POST_SYSCALL(pread64)(long res, long fd, void *buf, long count, long pos) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, res);
+    if (buf)
+      POST_WRITE(buf, res);
   }
 }
 
 PRE_SYSCALL(pwrite64)(long fd, const void *buf, long count, long pos) {
-  if (buf) PRE_READ(buf, count);
+  if (buf)
+    PRE_READ(buf, count);
 }
 
-POST_SYSCALL(pwrite64)(long res, long fd, const void *buf, long count,
-                       long pos) {}
-#else
+POST_SYSCALL(pwrite64)
+(long res, long fd, const void *buf, long count, long pos) {}
+#  else
 PRE_SYSCALL(pread64)(long fd, void *buf, long count, long pos0, long pos1) {}
 
-POST_SYSCALL(pread64)(long res, long fd, void *buf, long count, long pos0,
-                      long pos1) {
+POST_SYSCALL(pread64)
+(long res, long fd, void *buf, long count, long pos0, long pos1) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, res);
+    if (buf)
+      POST_WRITE(buf, res);
   }
 }
 
-PRE_SYSCALL(pwrite64)(long fd, const void *buf, long count, long pos0,
-                      long pos1) {
-  if (buf) PRE_READ(buf, count);
+PRE_SYSCALL(pwrite64)
+(long fd, const void *buf, long count, long pos0, long pos1) {
+  if (buf)
+    PRE_READ(buf, count);
 }
 
-POST_SYSCALL(pwrite64)(long res, long fd, const void *buf, long count,
-                       long pos0, long pos1) {}
-#endif
+POST_SYSCALL(pwrite64)
+(long res, long fd, const void *buf, long count, long pos0, long pos1) {}
+#  endif
 
-PRE_SYSCALL(preadv)(long fd, const __sanitizer_iovec *vec, long vlen,
-                    long pos_l, long pos_h) {}
+PRE_SYSCALL(preadv)
+(long fd, const __sanitizer_iovec *vec, long vlen, long pos_l, long pos_h) {}
 
-POST_SYSCALL(preadv)(long res, long fd, const __sanitizer_iovec *vec, long vlen,
-                     long pos_l, long pos_h) {
+POST_SYSCALL(preadv)
+(long res, long fd, const __sanitizer_iovec *vec, long vlen, long pos_l,
+ long pos_h) {
   if (res >= 0) {
-    if (vec) kernel_write_iovec(vec, vlen, res);
+    if (vec)
+      kernel_write_iovec(vec, vlen, res);
   }
 }
 
-PRE_SYSCALL(pwritev)(long fd, const __sanitizer_iovec *vec, long vlen,
-                     long pos_l, long pos_h) {}
+PRE_SYSCALL(pwritev)
+(long fd, const __sanitizer_iovec *vec, long vlen, long pos_l, long pos_h) {}
 
-POST_SYSCALL(pwritev)(long res, long fd, const __sanitizer_iovec *vec,
-                      long vlen, long pos_l, long pos_h) {
+POST_SYSCALL(pwritev)
+(long res, long fd, const __sanitizer_iovec *vec, long vlen, long pos_l,
+ long pos_h) {
   if (res >= 0) {
-    if (vec) kernel_read_iovec(vec, vlen, res);
+    if (vec)
+      kernel_read_iovec(vec, vlen, res);
   }
 }
 
@@ -1717,14 +1844,15 @@ PRE_SYSCALL(quotactl)(long cmd, const void *special, long id, void *addr) {
     PRE_READ(special, __sanitizer::internal_strlen((const char *)special) + 1);
 }
 
-POST_SYSCALL(quotactl)(long res, long cmd, const void *special, long id,
-                       void *addr) {}
+POST_SYSCALL(quotactl)
+(long res, long cmd, const void *special, long id, void *addr) {}
 
 PRE_SYSCALL(getdents)(long fd, void *dirent, long count) {}
 
 POST_SYSCALL(getdents)(long res, long fd, void *dirent, long count) {
   if (res >= 0) {
-    if (dirent) POST_WRITE(dirent, res);
+    if (dirent)
+      POST_WRITE(dirent, res);
   }
 }
 
@@ -1732,15 +1860,16 @@ PRE_SYSCALL(getdents64)(long fd, void *dirent, long count) {}
 
 POST_SYSCALL(getdents64)(long res, long fd, void *dirent, long count) {
   if (res >= 0) {
-    if (dirent) POST_WRITE(dirent, res);
+    if (dirent)
+      POST_WRITE(dirent, res);
   }
 }
 
-PRE_SYSCALL(setsockopt)(long fd, long level, long optname, void *optval,
-                        long optlen) {}
+PRE_SYSCALL(setsockopt)
+(long fd, long level, long optname, void *optval, long optlen) {}
 
-POST_SYSCALL(setsockopt)(long res, long fd, long level, long optname,
-                         void *optval, long optlen) {
+POST_SYSCALL(setsockopt)
+(long res, long fd, long level, long optname, void *optval, long optlen) {
   if (res >= 0) {
     if (optval)
       POST_WRITE(optval,
@@ -1748,77 +1877,88 @@ POST_SYSCALL(setsockopt)(long res, long fd, long level, long optname,
   }
 }
 
-PRE_SYSCALL(getsockopt)(long fd, long level, long optname, void *optval,
-                        void *optlen) {}
+PRE_SYSCALL(getsockopt)
+(long fd, long level, long optname, void *optval, void *optlen) {}
 
-POST_SYSCALL(getsockopt)(long res, long fd, long level, long optname,
-                         void *optval, void *optlen) {
+POST_SYSCALL(getsockopt)
+(long res, long fd, long level, long optname, void *optval, void *optlen) {
   if (res >= 0) {
     if (optval)
       POST_WRITE(optval,
                  __sanitizer::internal_strlen((const char *)optval) + 1);
-    if (optlen) POST_WRITE(optlen, sizeof(int));
+    if (optlen)
+      POST_WRITE(optlen, sizeof(int));
   }
 }
 
 PRE_SYSCALL(bind)(long arg0, sanitizer_kernel_sockaddr *arg1, long arg2) {}
 
-POST_SYSCALL(bind)(long res, long arg0, sanitizer_kernel_sockaddr *arg1,
-                   long arg2) {
+POST_SYSCALL(bind)
+(long res, long arg0, sanitizer_kernel_sockaddr *arg1, long arg2) {
   if (res >= 0) {
-    if (arg1) POST_WRITE(arg1, sizeof(*arg1));
+    if (arg1)
+      POST_WRITE(arg1, sizeof(*arg1));
   }
 }
 
 PRE_SYSCALL(connect)(long arg0, sanitizer_kernel_sockaddr *arg1, long arg2) {}
 
-POST_SYSCALL(connect)(long res, long arg0, sanitizer_kernel_sockaddr *arg1,
-                      long arg2) {
+POST_SYSCALL(connect)
+(long res, long arg0, sanitizer_kernel_sockaddr *arg1, long arg2) {
   if (res >= 0) {
-    if (arg1) POST_WRITE(arg1, sizeof(*arg1));
+    if (arg1)
+      POST_WRITE(arg1, sizeof(*arg1));
   }
 }
 
 PRE_SYSCALL(accept)(long arg0, sanitizer_kernel_sockaddr *arg1, void *arg2) {}
 
-POST_SYSCALL(accept)(long res, long arg0, sanitizer_kernel_sockaddr *arg1,
-                     void *arg2) {
+POST_SYSCALL(accept)
+(long res, long arg0, sanitizer_kernel_sockaddr *arg1, void *arg2) {
   if (res >= 0) {
-    if (arg1) POST_WRITE(arg1, sizeof(*arg1));
-    if (arg2) POST_WRITE(arg2, sizeof(unsigned));
+    if (arg1)
+      POST_WRITE(arg1, sizeof(*arg1));
+    if (arg2)
+      POST_WRITE(arg2, sizeof(unsigned));
   }
 }
 
-PRE_SYSCALL(accept4)(long arg0, sanitizer_kernel_sockaddr *arg1, void *arg2,
-                     long arg3) {}
+PRE_SYSCALL(accept4)
+(long arg0, sanitizer_kernel_sockaddr *arg1, void *arg2, long arg3) {}
 
-POST_SYSCALL(accept4)(long res, long arg0, sanitizer_kernel_sockaddr *arg1,
-                      void *arg2, long arg3) {
+POST_SYSCALL(accept4)
+(long res, long arg0, sanitizer_kernel_sockaddr *arg1, void *arg2, long arg3) {
   if (res >= 0) {
-    if (arg1) POST_WRITE(arg1, sizeof(*arg1));
-    if (arg2) POST_WRITE(arg2, sizeof(unsigned));
+    if (arg1)
+      POST_WRITE(arg1, sizeof(*arg1));
+    if (arg2)
+      POST_WRITE(arg2, sizeof(unsigned));
   }
 }
 
-PRE_SYSCALL(getsockname)(long arg0, sanitizer_kernel_sockaddr *arg1,
-                         void *arg2) {}
+PRE_SYSCALL(getsockname)
+(long arg0, sanitizer_kernel_sockaddr *arg1, void *arg2) {}
 
-POST_SYSCALL(getsockname)(long res, long arg0, sanitizer_kernel_sockaddr *arg1,
-                          void *arg2) {
+POST_SYSCALL(getsockname)
+(long res, long arg0, sanitizer_kernel_sockaddr *arg1, void *arg2) {
   if (res >= 0) {
-    if (arg1) POST_WRITE(arg1, sizeof(*arg1));
-    if (arg2) POST_WRITE(arg2, sizeof(unsigned));
+    if (arg1)
+      POST_WRITE(arg1, sizeof(*arg1));
+    if (arg2)
+      POST_WRITE(arg2, sizeof(unsigned));
   }
 }
 
-PRE_SYSCALL(getpeername)(long arg0, sanitizer_kernel_sockaddr *arg1,
-                         void *arg2) {}
+PRE_SYSCALL(getpeername)
+(long arg0, sanitizer_kernel_sockaddr *arg1, void *arg2) {}
 
-POST_SYSCALL(getpeername)(long res, long arg0, sanitizer_kernel_sockaddr *arg1,
-                          void *arg2) {
+POST_SYSCALL(getpeername)
+(long res, long arg0, sanitizer_kernel_sockaddr *arg1, void *arg2) {
   if (res >= 0) {
-    if (arg1) POST_WRITE(arg1, sizeof(*arg1));
-    if (arg2) POST_WRITE(arg2, sizeof(unsigned));
+    if (arg1)
+      POST_WRITE(arg1, sizeof(*arg1));
+    if (arg2)
+      POST_WRITE(arg2, sizeof(unsigned));
   }
 }
 
@@ -1826,18 +1966,23 @@ PRE_SYSCALL(send)(long arg0, void *arg1, long arg2, long arg3) {}
 
 POST_SYSCALL(send)(long res, long arg0, void *arg1, long arg2, long arg3) {
   if (res) {
-    if (arg1) POST_READ(arg1, res);
+    if (arg1)
+      POST_READ(arg1, res);
   }
 }
 
-PRE_SYSCALL(sendto)(long arg0, void *arg1, long arg2, long arg3,
-                    sanitizer_kernel_sockaddr *arg4, long arg5) {}
+PRE_SYSCALL(sendto)
+(long arg0, void *arg1, long arg2, long arg3, sanitizer_kernel_sockaddr *arg4,
+ long arg5) {}
 
-POST_SYSCALL(sendto)(long res, long arg0, void *arg1, long arg2, long arg3,
-                     sanitizer_kernel_sockaddr *arg4, long arg5) {
+POST_SYSCALL(sendto)
+(long res, long arg0, void *arg1, long arg2, long arg3,
+ sanitizer_kernel_sockaddr *arg4, long arg5) {
   if (res >= 0) {
-    if (arg1) POST_READ(arg1, res);
-    if (arg4) POST_WRITE(arg4, sizeof(*arg4));
+    if (arg1)
+      POST_READ(arg1, res);
+    if (arg4)
+      POST_WRITE(arg4, sizeof(*arg4));
   }
 }
 
@@ -1857,19 +2002,25 @@ PRE_SYSCALL(recv)(long arg0, void *buf, long len, long flags) {}
 
 POST_SYSCALL(recv)(long res, void *buf, long len, long flags) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, res);
+    if (buf)
+      POST_WRITE(buf, res);
   }
 }
 
-PRE_SYSCALL(recvfrom)(long arg0, void *buf, long len, long flags,
-                      sanitizer_kernel_sockaddr *arg4, void *arg5) {}
+PRE_SYSCALL(recvfrom)
+(long arg0, void *buf, long len, long flags, sanitizer_kernel_sockaddr *arg4,
+ void *arg5) {}
 
-POST_SYSCALL(recvfrom)(long res, long arg0, void *buf, long len, long flags,
-                       sanitizer_kernel_sockaddr *arg4, void *arg5) {
+POST_SYSCALL(recvfrom)
+(long res, long arg0, void *buf, long len, long flags,
+ sanitizer_kernel_sockaddr *arg4, void *arg5) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, res);
-    if (arg4) POST_WRITE(arg4, sizeof(*arg4));
-    if (arg5) POST_WRITE(arg5, sizeof(int));
+    if (buf)
+      POST_WRITE(buf, res);
+    if (arg4)
+      POST_WRITE(arg4, sizeof(*arg4));
+    if (arg5)
+      POST_WRITE(arg5, sizeof(int));
   }
 }
 
@@ -1881,14 +2032,16 @@ PRE_SYSCALL(socketpair)(long arg0, long arg1, long arg2, int *sv) {}
 
 POST_SYSCALL(socketpair)(long res, long arg0, long arg1, long arg2, int *sv) {
   if (res >= 0)
-    if (sv) POST_WRITE(sv, sizeof(int) * 2);
+    if (sv)
+      POST_WRITE(sv, sizeof(int) * 2);
 }
 
 PRE_SYSCALL(socketcall)(long call, void *args) {}
 
 POST_SYSCALL(socketcall)(long res, long call, void *args) {
   if (res >= 0) {
-    if (args) POST_WRITE(args, sizeof(long));
+    if (args)
+      POST_WRITE(args, sizeof(long));
   }
 }
 
@@ -1898,25 +2051,31 @@ POST_SYSCALL(listen)(long res, long arg0, long arg1) {}
 
 PRE_SYSCALL(poll)(void *ufds, long nfds, long timeout) {}
 
-POST_SYSCALL(poll)(long res, __sanitizer_pollfd *ufds, long nfds,
-                   long timeout) {
+POST_SYSCALL(poll)
+(long res, __sanitizer_pollfd *ufds, long nfds, long timeout) {
   if (res >= 0) {
-    if (ufds) POST_WRITE(ufds, nfds * sizeof(*ufds));
+    if (ufds)
+      POST_WRITE(ufds, nfds * sizeof(*ufds));
   }
 }
 
-PRE_SYSCALL(select)(long n, __sanitizer___kernel_fd_set *inp,
-                    __sanitizer___kernel_fd_set *outp,
-                    __sanitizer___kernel_fd_set *exp, void *tvp) {}
+PRE_SYSCALL(select)
+(long n, __sanitizer___kernel_fd_set *inp, __sanitizer___kernel_fd_set *outp,
+ __sanitizer___kernel_fd_set *exp, void *tvp) {}
 
-POST_SYSCALL(select)(long res, long n, __sanitizer___kernel_fd_set *inp,
-                     __sanitizer___kernel_fd_set *outp,
-                     __sanitizer___kernel_fd_set *exp, void *tvp) {
+POST_SYSCALL(select)
+(long res, long n, __sanitizer___kernel_fd_set *inp,
+ __sanitizer___kernel_fd_set *outp, __sanitizer___kernel_fd_set *exp,
+ void *tvp) {
   if (res >= 0) {
-    if (inp) POST_WRITE(inp, sizeof(*inp));
-    if (outp) POST_WRITE(outp, sizeof(*outp));
-    if (exp) POST_WRITE(exp, sizeof(*exp));
-    if (tvp) POST_WRITE(tvp, timeval_sz);
+    if (inp)
+      POST_WRITE(inp, sizeof(*inp));
+    if (outp)
+      POST_WRITE(outp, sizeof(*outp));
+    if (exp)
+      POST_WRITE(exp, sizeof(*exp));
+    if (tvp)
+      POST_WRITE(tvp, timeval_sz);
   }
 }
 
@@ -1936,29 +2095,55 @@ PRE_SYSCALL(epoll_ctl)(long epfd, long op, long fd, void *event) {}
 
 POST_SYSCALL(epoll_ctl)(long res, long epfd, long op, long fd, void *event) {
   if (res >= 0) {
-    if (event) POST_WRITE(event, struct_epoll_event_sz);
+    if (event)
+      POST_WRITE(event, struct_epoll_event_sz);
   }
 }
 
-PRE_SYSCALL(epoll_wait)(long epfd, void *events, long maxevents, long timeout) {
+PRE_SYSCALL(epoll_wait)
+(long epfd, void *events, long maxevents, long timeout) {}
+
+POST_SYSCALL(epoll_wait)
+(long res, long epfd, void *events, long maxevents, long timeout) {
+  if (res >= 0) {
+    if (events)
+      POST_WRITE(events, res * struct_epoll_event_sz);
+  }
+}
+
+PRE_SYSCALL(epoll_pwait)
+(long epfd, void *events, long maxevents, long timeout,
+ const kernel_sigset_t *sigmask, long sigsetsize) {
+  if (sigmask)
+    PRE_READ(sigmask, sigsetsize);
 }
 
-POST_SYSCALL(epoll_wait)(long res, long epfd, void *events, long maxevents,
-                         long timeout) {
+POST_SYSCALL(epoll_pwait)
+(long res, long epfd, void *events, long maxevents, long timeout,
+ const void *sigmask, long sigsetsize) {
   if (res >= 0) {
-    if (events) POST_WRITE(events, struct_epoll_event_sz);
+    if (events)
+      POST_WRITE(events, res * struct_epoll_event_sz);
   }
 }
 
-PRE_SYSCALL(epoll_pwait)(long epfd, void *events, long maxevents, long timeout,
-                         const kernel_sigset_t *sigmask, long sigsetsize) {
-  if (sigmask) PRE_READ(sigmask, sigsetsize);
+PRE_SYSCALL(epoll_pwait2)
+(long epfd, void *events, long maxevents,
+ const sanitizer_kernel_timespec *timeout, const kernel_sigset_t *sigmask,
+ long sigsetsize) {
+  if (timeout)
+    PRE_READ(timeout, sizeof(timeout));
+  if (sigmask)
+    PRE_READ(sigmask, sigsetsize);
 }
 
-POST_SYSCALL(epoll_pwait)(long res, long epfd, void *events, long maxevents,
-                          long timeout, const void *sigmask, long sigsetsize) {
+POST_SYSCALL(epoll_pwait2)
+(long res, long epfd, void *events, long maxevents,
+ const sanitizer_kernel_timespec *timeout, const void *sigmask,
+ long sigsetsize) {
   if (res >= 0) {
-    if (events) POST_WRITE(events, struct_epoll_event_sz);
+    if (events)
+      POST_WRITE(events, res * struct_epoll_event_sz);
   }
 }
 
@@ -1993,7 +2178,8 @@ PRE_SYSCALL(newuname)(void *name) {}
 
 POST_SYSCALL(newuname)(long res, void *name) {
   if (res >= 0) {
-    if (name) POST_WRITE(name, struct_new_utsname_sz);
+    if (name)
+      POST_WRITE(name, struct_new_utsname_sz);
   }
 }
 
@@ -2001,7 +2187,8 @@ PRE_SYSCALL(uname)(void *arg0) {}
 
 POST_SYSCALL(uname)(long res, void *arg0) {
   if (res >= 0) {
-    if (arg0) POST_WRITE(arg0, struct_old_utsname_sz);
+    if (arg0)
+      POST_WRITE(arg0, struct_old_utsname_sz);
   }
 }
 
@@ -2009,7 +2196,8 @@ PRE_SYSCALL(olduname)(void *arg0) {}
 
 POST_SYSCALL(olduname)(long res, void *arg0) {
   if (res >= 0) {
-    if (arg0) POST_WRITE(arg0, struct_oldold_utsname_sz);
+    if (arg0)
+      POST_WRITE(arg0, struct_oldold_utsname_sz);
   }
 }
 
@@ -2017,7 +2205,8 @@ PRE_SYSCALL(getrlimit)(long resource, void *rlim) {}
 
 POST_SYSCALL(getrlimit)(long res, long resource, void *rlim) {
   if (res >= 0) {
-    if (rlim) POST_WRITE(rlim, struct_rlimit_sz);
+    if (rlim)
+      POST_WRITE(rlim, struct_rlimit_sz);
   }
 }
 
@@ -2025,7 +2214,8 @@ PRE_SYSCALL(old_getrlimit)(long resource, void *rlim) {}
 
 POST_SYSCALL(old_getrlimit)(long res, long resource, void *rlim) {
   if (res >= 0) {
-    if (rlim) POST_WRITE(rlim, struct_rlimit_sz);
+    if (rlim)
+      POST_WRITE(rlim, struct_rlimit_sz);
   }
 }
 
@@ -2033,29 +2223,33 @@ PRE_SYSCALL(setrlimit)(long resource, void *rlim) {}
 
 POST_SYSCALL(setrlimit)(long res, long resource, void *rlim) {
   if (res >= 0) {
-    if (rlim) POST_WRITE(rlim, struct_rlimit_sz);
+    if (rlim)
+      POST_WRITE(rlim, struct_rlimit_sz);
   }
 }
 
-#if !SANITIZER_ANDROID
-PRE_SYSCALL(prlimit64)(long pid, long resource, const void *new_rlim,
-                       void *old_rlim) {
-  if (new_rlim) PRE_READ(new_rlim, struct_rlimit64_sz);
+#  if !SANITIZER_ANDROID
+PRE_SYSCALL(prlimit64)
+(long pid, long resource, const void *new_rlim, void *old_rlim) {
+  if (new_rlim)
+    PRE_READ(new_rlim, struct_rlimit64_sz);
 }
 
-POST_SYSCALL(prlimit64)(long res, long pid, long resource, const void *new_rlim,
-                        void *old_rlim) {
+POST_SYSCALL(prlimit64)
+(long res, long pid, long resource, const void *new_rlim, void *old_rlim) {
   if (res >= 0) {
-    if (old_rlim) POST_WRITE(old_rlim, struct_rlimit64_sz);
+    if (old_rlim)
+      POST_WRITE(old_rlim, struct_rlimit64_sz);
   }
 }
-#endif
+#  endif
 
 PRE_SYSCALL(getrusage)(long who, void *ru) {}
 
 POST_SYSCALL(getrusage)(long res, long who, void *ru) {
   if (res >= 0) {
-    if (ru) POST_WRITE(ru, struct_rusage_sz);
+    if (ru)
+      POST_WRITE(ru, struct_rusage_sz);
   }
 }
 
@@ -2068,31 +2262,34 @@ PRE_SYSCALL(msgget)(long key, long msgflg) {}
 POST_SYSCALL(msgget)(long res, long key, long msgflg) {}
 
 PRE_SYSCALL(msgsnd)(long msqid, void *msgp, long msgsz, long msgflg) {
-  if (msgp) PRE_READ(msgp, msgsz);
+  if (msgp)
+    PRE_READ(msgp, msgsz);
 }
 
-POST_SYSCALL(msgsnd)(long res, long msqid, void *msgp, long msgsz,
-                     long msgflg) {}
+POST_SYSCALL(msgsnd)
+(long res, long msqid, void *msgp, long msgsz, long msgflg) {}
 
-PRE_SYSCALL(msgrcv)(long msqid, void *msgp, long msgsz, long msgtyp,
-                    long msgflg) {}
+PRE_SYSCALL(msgrcv)
+(long msqid, void *msgp, long msgsz, long msgtyp, long msgflg) {}
 
-POST_SYSCALL(msgrcv)(long res, long msqid, void *msgp, long msgsz, long msgtyp,
-                     long msgflg) {
+POST_SYSCALL(msgrcv)
+(long res, long msqid, void *msgp, long msgsz, long msgtyp, long msgflg) {
   if (res >= 0) {
-    if (msgp) POST_WRITE(msgp, res);
+    if (msgp)
+      POST_WRITE(msgp, res);
   }
 }
 
-#if !SANITIZER_ANDROID
+#  if !SANITIZER_ANDROID
 PRE_SYSCALL(msgctl)(long msqid, long cmd, void *buf) {}
 
 POST_SYSCALL(msgctl)(long res, long msqid, long cmd, void *buf) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, struct_msqid_ds_sz);
+    if (buf)
+      POST_WRITE(buf, struct_msqid_ds_sz);
   }
 }
-#endif
+#  endif
 
 PRE_SYSCALL(semget)(long key, long nsems, long semflg) {}
 
@@ -2106,13 +2303,14 @@ PRE_SYSCALL(semctl)(long semid, long semnum, long cmd, void *arg) {}
 
 POST_SYSCALL(semctl)(long res, long semid, long semnum, long cmd, void *arg) {}
 
-PRE_SYSCALL(semtimedop)(long semid, void *sops, long nsops,
-                        const void *timeout) {
-  if (timeout) PRE_READ(timeout, struct_timespec_sz);
+PRE_SYSCALL(semtimedop)
+(long semid, void *sops, long nsops, const void *timeout) {
+  if (timeout)
+    PRE_READ(timeout, struct_timespec_sz);
 }
 
-POST_SYSCALL(semtimedop)(long res, long semid, void *sops, long nsops,
-                         const void *timeout) {}
+POST_SYSCALL(semtimedop)
+(long res, long semid, void *sops, long nsops, const void *timeout) {}
 
 PRE_SYSCALL(shmat)(long shmid, void *shmaddr, long shmflg) {}
 
@@ -2138,18 +2336,20 @@ POST_SYSCALL(shmdt)(long res, void *shmaddr) {
   }
 }
 
-PRE_SYSCALL(ipc)(long call, long first, long second, long third, void *ptr,
-                 long fifth) {}
+PRE_SYSCALL(ipc)
+(long call, long first, long second, long third, void *ptr, long fifth) {}
 
-POST_SYSCALL(ipc)(long res, long call, long first, long second, long third,
-                  void *ptr, long fifth) {}
+POST_SYSCALL(ipc)
+(long res, long call, long first, long second, long third, void *ptr,
+ long fifth) {}
 
-#if !SANITIZER_ANDROID
+#  if !SANITIZER_ANDROID
 PRE_SYSCALL(shmctl)(long shmid, long cmd, void *buf) {}
 
 POST_SYSCALL(shmctl)(long res, long shmid, long cmd, void *buf) {
   if (res >= 0) {
-    if (buf) POST_WRITE(buf, sizeof(__sanitizer_shmid_ds));
+    if (buf)
+      POST_WRITE(buf, sizeof(__sanitizer_shmid_ds));
   }
 }
 
@@ -2158,10 +2358,11 @@ PRE_SYSCALL(mq_open)(const void *name, long oflag, long mode, void *attr) {
     PRE_READ(name, __sanitizer::internal_strlen((const char *)name) + 1);
 }
 
-POST_SYSCALL(mq_open)(long res, const void *name, long oflag, long mode,
-                      void *attr) {
+POST_SYSCALL(mq_open)
+(long res, const void *name, long oflag, long mode, void *attr) {
   if (res >= 0) {
-    if (attr) POST_WRITE(attr, struct_mq_attr_sz);
+    if (attr)
+      POST_WRITE(attr, struct_mq_attr_sz);
   }
 }
 
@@ -2172,62 +2373,73 @@ PRE_SYSCALL(mq_unlink)(const void *name) {
 
 POST_SYSCALL(mq_unlink)(long res, const void *name) {}
 
-PRE_SYSCALL(mq_timedsend)(long mqdes, const void *msg_ptr, long msg_len,
-                          long msg_prio, const void *abs_timeout) {
-  if (msg_ptr) PRE_READ(msg_ptr, msg_len);
-  if (abs_timeout) PRE_READ(abs_timeout, struct_timespec_sz);
+PRE_SYSCALL(mq_timedsend)
+(long mqdes, const void *msg_ptr, long msg_len, long msg_prio,
+ const void *abs_timeout) {
+  if (msg_ptr)
+    PRE_READ(msg_ptr, msg_len);
+  if (abs_timeout)
+    PRE_READ(abs_timeout, struct_timespec_sz);
 }
 
-POST_SYSCALL(mq_timedsend)(long res, long mqdes, const void *msg_ptr,
-                           long msg_len, long msg_prio,
-                           const void *abs_timeout) {}
+POST_SYSCALL(mq_timedsend)
+(long res, long mqdes, const void *msg_ptr, long msg_len, long msg_prio,
+ const void *abs_timeout) {}
 
-PRE_SYSCALL(mq_timedreceive)(long mqdes, void *msg_ptr, long msg_len,
-                             void *msg_prio, const void *abs_timeout) {
-  if (abs_timeout) PRE_READ(abs_timeout, struct_timespec_sz);
+PRE_SYSCALL(mq_timedreceive)
+(long mqdes, void *msg_ptr, long msg_len, void *msg_prio,
+ const void *abs_timeout) {
+  if (abs_timeout)
+    PRE_READ(abs_timeout, struct_timespec_sz);
 }
 
-POST_SYSCALL(mq_timedreceive)(long res, long mqdes, void *msg_ptr, long msg_len,
-                              int *msg_prio, const void *abs_timeout) {
+POST_SYSCALL(mq_timedreceive)
+(long res, long mqdes, void *msg_ptr, long msg_len, int *msg_prio,
+ const void *abs_timeout) {
   if (res >= 0) {
-    if (msg_ptr) POST_WRITE(msg_ptr, res);
-    if (msg_prio) POST_WRITE(msg_prio, sizeof(*msg_prio));
+    if (msg_ptr)
+      POST_WRITE(msg_ptr, res);
+    if (msg_prio)
+      POST_WRITE(msg_prio, sizeof(*msg_prio));
   }
 }
 
 PRE_SYSCALL(mq_notify)(long mqdes, const void *notification) {
-  if (notification) PRE_READ(notification, struct_sigevent_sz);
+  if (notification)
+    PRE_READ(notification, struct_sigevent_sz);
 }
 
 POST_SYSCALL(mq_notify)(long res, long mqdes, const void *notification) {}
 
 PRE_SYSCALL(mq_getsetattr)(long mqdes, const void *mqstat, void *omqstat) {
-  if (mqstat) PRE_READ(mqstat, struct_mq_attr_sz);
+  if (mqstat)
+    PRE_READ(mqstat, struct_mq_attr_sz);
 }
 
-POST_SYSCALL(mq_getsetattr)(long res, long mqdes, const void *mqstat,
-                            void *omqstat) {
+POST_SYSCALL(mq_getsetattr)
+(long res, long mqdes, const void *mqstat, void *omqstat) {
   if (res >= 0) {
-    if (omqstat) POST_WRITE(omqstat, struct_mq_attr_sz);
+    if (omqstat)
+      POST_WRITE(omqstat, struct_mq_attr_sz);
   }
 }
-#endif  // SANITIZER_ANDROID
+#  endif  // SANITIZER_ANDROID
 
 PRE_SYSCALL(pciconfig_iobase)(long which, long bus, long devfn) {}
 
 POST_SYSCALL(pciconfig_iobase)(long res, long which, long bus, long devfn) {}
 
-PRE_SYSCALL(pciconfig_read)(long bus, long dfn, long off, long len, void *buf) {
-}
+PRE_SYSCALL(pciconfig_read)
+(long bus, long dfn, long off, long len, void *buf) {}
 
-POST_SYSCALL(pciconfig_read)(long res, long bus, long dfn, long off, long len,
-                             void *buf) {}
+POST_SYSCALL(pciconfig_read)
+(long res, long bus, long dfn, long off, long len, void *buf) {}
 
-PRE_SYSCALL(pciconfig_write)(long bus, long dfn, long off, long len,
-                             void *buf) {}
+PRE_SYSCALL(pciconfig_write)
+(long bus, long dfn, long off, long len, void *buf) {}
 
-POST_SYSCALL(pciconfig_write)(long res, long bus, long dfn, long off, long len,
-                              void *buf) {}
+POST_SYSCALL(pciconfig_write)
+(long res, long bus, long dfn, long off, long len, void *buf) {}
 
 PRE_SYSCALL(swapon)(const void *specialfile, long swap_flags) {
   if (specialfile)
@@ -2247,8 +2459,10 @@ POST_SYSCALL(swapoff)(long res, const void *specialfile) {}
 
 PRE_SYSCALL(sysctl)(__sanitizer___sysctl_args *args) {
   if (args) {
-    if (args->name) PRE_READ(args->name, args->nlen * sizeof(*args->name));
-    if (args->newval) PRE_READ(args->name, args->newlen);
+    if (args->name)
+      PRE_READ(args->name, args->nlen * sizeof(*args->name));
+    if (args->newval)
+      PRE_READ(args->name, args->newlen);
   }
 }
 
@@ -2265,7 +2479,8 @@ PRE_SYSCALL(sysinfo)(void *info) {}
 
 POST_SYSCALL(sysinfo)(long res, void *info) {
   if (res >= 0) {
-    if (info) POST_WRITE(info, struct_sysinfo_sz);
+    if (info)
+      POST_WRITE(info, struct_sysinfo_sz);
   }
 }
 
@@ -2294,10 +2509,10 @@ PRE_SYSCALL(ni_syscall)() {}
 POST_SYSCALL(ni_syscall)(long res) {}
 
 PRE_SYSCALL(ptrace)(long request, long pid, long addr, long data) {
-#if !SANITIZER_ANDROID &&                                                   \
-    (defined(__i386) || defined(__x86_64) || defined(__mips64) ||           \
-     defined(__powerpc64__) || defined(__aarch64__) || defined(__s390__) || \
-     SANITIZER_RISCV64)
+#  if !SANITIZER_ANDROID &&                                                   \
+      (defined(__i386) || defined(__x86_64) || defined(__mips64) ||           \
+       defined(__powerpc64__) || defined(__aarch64__) || defined(__s390__) || \
+       SANITIZER_RISCV64)
   if (data) {
     if (request == ptrace_setregs) {
       PRE_READ((void *)data, struct_user_regs_struct_sz);
@@ -2312,14 +2527,14 @@ PRE_SYSCALL(ptrace)(long request, long pid, long addr, long data) {
       PRE_READ(iov->iov_base, iov->iov_len);
     }
   }
-#endif
+#  endif
 }
 
 POST_SYSCALL(ptrace)(long res, long request, long pid, long addr, long data) {
-#if !SANITIZER_ANDROID &&                                                   \
-    (defined(__i386) || defined(__x86_64) || defined(__mips64) ||           \
-     defined(__powerpc64__) || defined(__aarch64__) || defined(__s390__) || \
-     SANITIZER_RISCV64)
+#  if !SANITIZER_ANDROID &&                                                   \
+      (defined(__i386) || defined(__x86_64) || defined(__mips64) ||           \
+       defined(__powerpc64__) || defined(__aarch64__) || defined(__s390__) || \
+       SANITIZER_RISCV64)
   if (res >= 0 && data) {
     // Note that this is different from the interceptor in
     // sanitizer_common_interceptors.inc.
@@ -2340,11 +2555,12 @@ POST_SYSCALL(ptrace)(long res, long request, long pid, long addr, long data) {
       POST_WRITE((void *)data, sizeof(void *));
     }
   }
-#endif
+#  endif
 }
 
-PRE_SYSCALL(add_key)(const void *_type, const void *_description,
-                     const void *_payload, long plen, long destringid) {
+PRE_SYSCALL(add_key)
+(const void *_type, const void *_description, const void *_payload, long plen,
+ long destringid) {
   if (_type)
     PRE_READ(_type, __sanitizer::internal_strlen((const char *)_type) + 1);
   if (_description)
@@ -2352,11 +2568,13 @@ PRE_SYSCALL(add_key)(const void *_type, const void *_description,
              __sanitizer::internal_strlen((const char *)_description) + 1);
 }
 
-POST_SYSCALL(add_key)(long res, const void *_type, const void *_description,
-                      const void *_payload, long plen, long destringid) {}
+POST_SYSCALL(add_key)
+(long res, const void *_type, const void *_description, const void *_payload,
+ long plen, long destringid) {}
 
-PRE_SYSCALL(request_key)(const void *_type, const void *_description,
-                         const void *_callout_info, long destringid) {
+PRE_SYSCALL(request_key)
+(const void *_type, const void *_description, const void *_callout_info,
+ long destringid) {
   if (_type)
     PRE_READ(_type, __sanitizer::internal_strlen((const char *)_type) + 1);
   if (_description)
@@ -2367,13 +2585,14 @@ PRE_SYSCALL(request_key)(const void *_type, const void *_description,
              __sanitizer::internal_strlen((const char *)_callout_info) + 1);
 }
 
-POST_SYSCALL(request_key)(long res, const void *_type, const void *_description,
-                          const void *_callout_info, long destringid) {}
+POST_SYSCALL(request_key)
+(long res, const void *_type, const void *_description,
+ const void *_callout_info, long destringid) {}
 
 PRE_SYSCALL(keyctl)(long cmd, long arg2, long arg3, long arg4, long arg5) {}
 
-POST_SYSCALL(keyctl)(long res, long cmd, long arg2, long arg3, long arg4,
-                     long arg5) {}
+POST_SYSCALL(keyctl)
+(long res, long cmd, long arg2, long arg3, long arg4, long arg5) {}
 
 PRE_SYSCALL(ioprio_set)(long which, long who, long ioprio) {}
 
@@ -2387,50 +2606,62 @@ PRE_SYSCALL(set_mempolicy)(long mode, void *nmask, long maxnode) {}
 
 POST_SYSCALL(set_mempolicy)(long res, long mode, void *nmask, long maxnode) {
   if (res >= 0) {
-    if (nmask) POST_WRITE(nmask, sizeof(long));
+    if (nmask)
+      POST_WRITE(nmask, sizeof(long));
   }
 }
 
-PRE_SYSCALL(migrate_pages)(long pid, long maxnode, const void *from,
-                           const void *to) {
-  if (from) PRE_READ(from, sizeof(long));
-  if (to) PRE_READ(to, sizeof(long));
+PRE_SYSCALL(migrate_pages)
+(long pid, long maxnode, const void *from, const void *to) {
+  if (from)
+    PRE_READ(from, sizeof(long));
+  if (to)
+    PRE_READ(to, sizeof(long));
 }
 
-POST_SYSCALL(migrate_pages)(long res, long pid, long maxnode, const void *from,
-                            const void *to) {}
+POST_SYSCALL(migrate_pages)
+(long res, long pid, long maxnode, const void *from, const void *to) {}
 
-PRE_SYSCALL(move_pages)(long pid, long nr_pages, const void **pages,
-                        const int *nodes, int *status, long flags) {
-  if (pages) PRE_READ(pages, nr_pages * sizeof(*pages));
-  if (nodes) PRE_READ(nodes, nr_pages * sizeof(*nodes));
+PRE_SYSCALL(move_pages)
+(long pid, long nr_pages, const void **pages, const int *nodes, int *status,
+ long flags) {
+  if (pages)
+    PRE_READ(pages, nr_pages * sizeof(*pages));
+  if (nodes)
+    PRE_READ(nodes, nr_pages * sizeof(*nodes));
 }
 
-POST_SYSCALL(move_pages)(long res, long pid, long nr_pages, const void **pages,
-                         const int *nodes, int *status, long flags) {
+POST_SYSCALL(move_pages)
+(long res, long pid, long nr_pages, const void **pages, const int *nodes,
+ int *status, long flags) {
   if (res >= 0) {
-    if (status) POST_WRITE(status, nr_pages * sizeof(*status));
+    if (status)
+      POST_WRITE(status, nr_pages * sizeof(*status));
   }
 }
 
-PRE_SYSCALL(mbind)(long start, long len, long mode, void *nmask, long maxnode,
-                   long flags) {}
+PRE_SYSCALL(mbind)
+(long start, long len, long mode, void *nmask, long maxnode, long flags) {}
 
-POST_SYSCALL(mbind)(long res, long start, long len, long mode, void *nmask,
-                    long maxnode, long flags) {
+POST_SYSCALL(mbind)
+(long res, long start, long len, long mode, void *nmask, long maxnode,
+ long flags) {
   if (res >= 0) {
-    if (nmask) POST_WRITE(nmask, sizeof(long));
+    if (nmask)
+      POST_WRITE(nmask, sizeof(long));
   }
 }
 
-PRE_SYSCALL(get_mempolicy)(void *policy, void *nmask, long maxnode, long addr,
-                           long flags) {}
+PRE_SYSCALL(get_mempolicy)
+(void *policy, void *nmask, long maxnode, long addr, long flags) {}
 
-POST_SYSCALL(get_mempolicy)(long res, void *policy, void *nmask, long maxnode,
-                            long addr, long flags) {
+POST_SYSCALL(get_mempolicy)
+(long res, void *policy, void *nmask, long maxnode, long addr, long flags) {
   if (res >= 0) {
-    if (policy) POST_WRITE(policy, sizeof(int));
-    if (nmask) POST_WRITE(nmask, sizeof(long));
+    if (policy)
+      POST_WRITE(policy, sizeof(int));
+    if (nmask)
+      POST_WRITE(nmask, sizeof(long));
   }
 }
 
@@ -2447,8 +2678,8 @@ PRE_SYSCALL(inotify_add_watch)(long fd, const void *path, long mask) {
     PRE_READ(path, __sanitizer::internal_strlen((const char *)path) + 1);
 }
 
-POST_SYSCALL(inotify_add_watch)(long res, long fd, const void *path,
-                                long mask) {}
+POST_SYSCALL(inotify_add_watch)
+(long res, long fd, const void *path, long mask) {}
 
 PRE_SYSCALL(inotify_rm_watch)(long fd, long wd) {}
 
@@ -2458,8 +2689,10 @@ PRE_SYSCALL(spu_run)(long fd, void *unpc, void *ustatus) {}
 
 POST_SYSCALL(spu_run)(long res, long fd, unsigned *unpc, unsigned *ustatus) {
   if (res >= 0) {
-    if (unpc) POST_WRITE(unpc, sizeof(*unpc));
-    if (ustatus) POST_WRITE(ustatus, sizeof(*ustatus));
+    if (unpc)
+      POST_WRITE(unpc, sizeof(*unpc));
+    if (ustatus)
+      POST_WRITE(ustatus, sizeof(*ustatus));
   }
 }
 
@@ -2468,8 +2701,8 @@ PRE_SYSCALL(spu_create)(const void *name, long flags, long mode, long fd) {
     PRE_READ(name, __sanitizer::internal_strlen((const char *)name) + 1);
 }
 
-POST_SYSCALL(spu_create)(long res, const void *name, long flags, long mode,
-                         long fd) {}
+POST_SYSCALL(spu_create)
+(long res, const void *name, long flags, long mode, long fd) {}
 
 PRE_SYSCALL(mknodat)(long dfd, const void *filename, long mode, long dev) {
   if (filename)
@@ -2477,8 +2710,8 @@ PRE_SYSCALL(mknodat)(long dfd, const void *filename, long mode, long dev) {
              __sanitizer::internal_strlen((const char *)filename) + 1);
 }
 
-POST_SYSCALL(mknodat)(long res, long dfd, const void *filename, long mode,
-                      long dev) {}
+POST_SYSCALL(mknodat)
+(long res, long dfd, const void *filename, long mode, long dev) {}
 
 PRE_SYSCALL(mkdirat)(long dfd, const void *pathname, long mode) {
   if (pathname)
@@ -2503,30 +2736,33 @@ PRE_SYSCALL(symlinkat)(const void *oldname, long newdfd, const void *newname) {
     PRE_READ(newname, __sanitizer::internal_strlen((const char *)newname) + 1);
 }
 
-POST_SYSCALL(symlinkat)(long res, const void *oldname, long newdfd,
-                        const void *newname) {}
+POST_SYSCALL(symlinkat)
+(long res, const void *oldname, long newdfd, const void *newname) {}
 
-PRE_SYSCALL(linkat)(long olddfd, const void *oldname, long newdfd,
-                    const void *newname, long flags) {
+PRE_SYSCALL(linkat)
+(long olddfd, const void *oldname, long newdfd, const void *newname,
+ long flags) {
   if (oldname)
     PRE_READ(oldname, __sanitizer::internal_strlen((const char *)oldname) + 1);
   if (newname)
     PRE_READ(newname, __sanitizer::internal_strlen((const char *)newname) + 1);
 }
 
-POST_SYSCALL(linkat)(long res, long olddfd, const void *oldname, long newdfd,
-                     const void *newname, long flags) {}
+POST_SYSCALL(linkat)
+(long res, long olddfd, const void *oldname, long newdfd, const void *newname,
+ long flags) {}
 
-PRE_SYSCALL(renameat)(long olddfd, const void *oldname, long newdfd,
-                      const void *newname) {
+PRE_SYSCALL(renameat)
+(long olddfd, const void *oldname, long newdfd, const void *newname) {
   if (oldname)
     PRE_READ(oldname, __sanitizer::internal_strlen((const char *)oldname) + 1);
   if (newname)
     PRE_READ(newname, __sanitizer::internal_strlen((const char *)newname) + 1);
 }
 
-POST_SYSCALL(renameat)(long res, long olddfd, const void *oldname, long newdfd,
-                       const void *newname) {}
+POST_SYSCALL(renameat)
+(long res, long olddfd, const void *oldname, long newdfd, const void *newname) {
+}
 
 PRE_SYSCALL(futimesat)(long dfd, const void *filename, void *utimes) {
   if (filename)
@@ -2534,10 +2770,11 @@ PRE_SYSCALL(futimesat)(long dfd, const void *filename, void *utimes) {
              __sanitizer::internal_strlen((const char *)filename) + 1);
 }
 
-POST_SYSCALL(futimesat)(long res, long dfd, const void *filename,
-                        void *utimes) {
+POST_SYSCALL(futimesat)
+(long res, long dfd, const void *filename, void *utimes) {
   if (res >= 0) {
-    if (utimes) POST_WRITE(utimes, timeval_sz);
+    if (utimes)
+      POST_WRITE(utimes, timeval_sz);
   }
 }
 
@@ -2557,15 +2794,15 @@ PRE_SYSCALL(fchmodat)(long dfd, const void *filename, long mode) {
 
 POST_SYSCALL(fchmodat)(long res, long dfd, const void *filename, long mode) {}
 
-PRE_SYSCALL(fchownat)(long dfd, const void *filename, long user, long group,
-                      long flag) {
+PRE_SYSCALL(fchownat)
+(long dfd, const void *filename, long user, long group, long flag) {
   if (filename)
     PRE_READ(filename,
              __sanitizer::internal_strlen((const char *)filename) + 1);
 }
 
-POST_SYSCALL(fchownat)(long res, long dfd, const void *filename, long user,
-                       long group, long flag) {}
+POST_SYSCALL(fchownat)
+(long res, long dfd, const void *filename, long user, long group, long flag) {}
 
 PRE_SYSCALL(openat)(long dfd, const void *filename, long flags, long mode) {
   if (filename)
@@ -2573,34 +2810,36 @@ PRE_SYSCALL(openat)(long dfd, const void *filename, long flags, long mode) {
              __sanitizer::internal_strlen((const char *)filename) + 1);
 }
 
-POST_SYSCALL(openat)(long res, long dfd, const void *filename, long flags,
-                     long mode) {}
+POST_SYSCALL(openat)
+(long res, long dfd, const void *filename, long flags, long mode) {}
 
-PRE_SYSCALL(newfstatat)(long dfd, const void *filename, void *statbuf,
-                        long flag) {
+PRE_SYSCALL(newfstatat)
+(long dfd, const void *filename, void *statbuf, long flag) {
   if (filename)
     PRE_READ(filename,
              __sanitizer::internal_strlen((const char *)filename) + 1);
 }
 
-POST_SYSCALL(newfstatat)(long res, long dfd, const void *filename,
-                         void *statbuf, long flag) {
+POST_SYSCALL(newfstatat)
+(long res, long dfd, const void *filename, void *statbuf, long flag) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct_kernel_stat_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct_kernel_stat_sz);
   }
 }
 
-PRE_SYSCALL(fstatat64)(long dfd, const void *filename, void *statbuf,
-                       long flag) {
+PRE_SYSCALL(fstatat64)
+(long dfd, const void *filename, void *statbuf, long flag) {
   if (filename)
     PRE_READ(filename,
              __sanitizer::internal_strlen((const char *)filename) + 1);
 }
 
-POST_SYSCALL(fstatat64)(long res, long dfd, const void *filename, void *statbuf,
-                        long flag) {
+POST_SYSCALL(fstatat64)
+(long res, long dfd, const void *filename, void *statbuf, long flag) {
   if (res >= 0) {
-    if (statbuf) POST_WRITE(statbuf, struct_kernel_stat64_sz);
+    if (statbuf)
+      POST_WRITE(statbuf, struct_kernel_stat64_sz);
   }
 }
 
@@ -2609,25 +2848,26 @@ PRE_SYSCALL(readlinkat)(long dfd, const void *path, void *buf, long bufsiz) {
     PRE_READ(path, __sanitizer::internal_strlen((const char *)path) + 1);
 }
 
-POST_SYSCALL(readlinkat)(long res, long dfd, const void *path, void *buf,
-                         long bufsiz) {
+POST_SYSCALL(readlinkat)
+(long res, long dfd, const void *path, void *buf, long bufsiz) {
   if (res >= 0) {
     if (buf)
       POST_WRITE(buf, __sanitizer::internal_strlen((const char *)buf) + 1);
   }
 }
 
-PRE_SYSCALL(utimensat)(long dfd, const void *filename, void *utimes,
-                       long flags) {
+PRE_SYSCALL(utimensat)
+(long dfd, const void *filename, void *utimes, long flags) {
   if (filename)
     PRE_READ(filename,
              __sanitizer::internal_strlen((const char *)filename) + 1);
 }
 
-POST_SYSCALL(utimensat)(long res, long dfd, const void *filename, void *utimes,
-                        long flags) {
+POST_SYSCALL(utimensat)
+(long res, long dfd, const void *filename, void *utimes, long flags) {
   if (res >= 0) {
-    if (utimes) POST_WRITE(utimes, struct_timespec_sz);
+    if (utimes)
+      POST_WRITE(utimes, struct_timespec_sz);
   }
 }
 
@@ -2635,24 +2875,28 @@ PRE_SYSCALL(unshare)(long unshare_flags) {}
 
 POST_SYSCALL(unshare)(long res, long unshare_flags) {}
 
-PRE_SYSCALL(splice)(long fd_in, void *off_in, long fd_out, void *off_out,
-                    long len, long flags) {}
+PRE_SYSCALL(splice)
+(long fd_in, void *off_in, long fd_out, void *off_out, long len, long flags) {}
 
-POST_SYSCALL(splice)(long res, long fd_in, void *off_in, long fd_out,
-                     void *off_out, long len, long flags) {
+POST_SYSCALL(splice)
+(long res, long fd_in, void *off_in, long fd_out, void *off_out, long len,
+ long flags) {
   if (res >= 0) {
-    if (off_in) POST_WRITE(off_in, sizeof(long long));
-    if (off_out) POST_WRITE(off_out, sizeof(long long));
+    if (off_in)
+      POST_WRITE(off_in, sizeof(long long));
+    if (off_out)
+      POST_WRITE(off_out, sizeof(long long));
   }
 }
 
-PRE_SYSCALL(vmsplice)(long fd, const __sanitizer_iovec *iov, long nr_segs,
-                      long flags) {}
+PRE_SYSCALL(vmsplice)
+(long fd, const __sanitizer_iovec *iov, long nr_segs, long flags) {}
 
-POST_SYSCALL(vmsplice)(long res, long fd, const __sanitizer_iovec *iov,
-                       long nr_segs, long flags) {
+POST_SYSCALL(vmsplice)
+(long res, long fd, const __sanitizer_iovec *iov, long nr_segs, long flags) {
   if (res >= 0) {
-    if (iov) kernel_read_iovec(iov, nr_segs, res);
+    if (iov)
+      kernel_read_iovec(iov, nr_segs, res);
   }
 }
 
@@ -2662,8 +2906,8 @@ POST_SYSCALL(tee)(long res, long fdin, long fdout, long len, long flags) {}
 
 PRE_SYSCALL(get_robust_list)(long pid, void *head_ptr, void *len_ptr) {}
 
-POST_SYSCALL(get_robust_list)(long res, long pid, void *head_ptr,
-                              void *len_ptr) {}
+POST_SYSCALL(get_robust_list)
+(long res, long pid, void *head_ptr, void *len_ptr) {}
 
 PRE_SYSCALL(set_robust_list)(void *head, long len) {}
 
@@ -2673,27 +2917,31 @@ PRE_SYSCALL(getcpu)(void *cpu, void *node, void *cache) {}
 
 POST_SYSCALL(getcpu)(long res, void *cpu, void *node, void *cache) {
   if (res >= 0) {
-    if (cpu) POST_WRITE(cpu, sizeof(unsigned));
-    if (node) POST_WRITE(node, sizeof(unsigned));
+    if (cpu)
+      POST_WRITE(cpu, sizeof(unsigned));
+    if (node)
+      POST_WRITE(node, sizeof(unsigned));
     // The third argument to this system call is nowadays unused.
   }
 }
 
 PRE_SYSCALL(signalfd)(long ufd, void *user_mask, long sizemask) {}
 
-POST_SYSCALL(signalfd)(long res, long ufd, kernel_sigset_t *user_mask,
-                       long sizemask) {
+POST_SYSCALL(signalfd)
+(long res, long ufd, kernel_sigset_t *user_mask, long sizemask) {
   if (res >= 0) {
-    if (user_mask) POST_WRITE(user_mask, sizemask);
+    if (user_mask)
+      POST_WRITE(user_mask, sizemask);
   }
 }
 
 PRE_SYSCALL(signalfd4)(long ufd, void *user_mask, long sizemask, long flags) {}
 
-POST_SYSCALL(signalfd4)(long res, long ufd, kernel_sigset_t *user_mask,
-                        long sizemask, long flags) {
+POST_SYSCALL(signalfd4)
+(long res, long ufd, kernel_sigset_t *user_mask, long sizemask, long flags) {
   if (res >= 0) {
-    if (user_mask) POST_WRITE(user_mask, sizemask);
+    if (user_mask)
+      POST_WRITE(user_mask, sizemask);
   }
 }
 
@@ -2701,15 +2949,17 @@ PRE_SYSCALL(timerfd_create)(long clockid, long flags) {}
 
 POST_SYSCALL(timerfd_create)(long res, long clockid, long flags) {}
 
-PRE_SYSCALL(timerfd_settime)(long ufd, long flags, const void *utmr,
-                             void *otmr) {
-  if (utmr) PRE_READ(utmr, struct_itimerspec_sz);
+PRE_SYSCALL(timerfd_settime)
+(long ufd, long flags, const void *utmr, void *otmr) {
+  if (utmr)
+    PRE_READ(utmr, struct_itimerspec_sz);
 }
 
-POST_SYSCALL(timerfd_settime)(long res, long ufd, long flags, const void *utmr,
-                              void *otmr) {
+POST_SYSCALL(timerfd_settime)
+(long res, long ufd, long flags, const void *utmr, void *otmr) {
   if (res >= 0) {
-    if (otmr) POST_WRITE(otmr, struct_itimerspec_sz);
+    if (otmr)
+      POST_WRITE(otmr, struct_itimerspec_sz);
   }
 }
 
@@ -2717,7 +2967,8 @@ PRE_SYSCALL(timerfd_gettime)(long ufd, void *otmr) {}
 
 POST_SYSCALL(timerfd_gettime)(long res, long ufd, void *otmr) {
   if (res >= 0) {
-    if (otmr) POST_WRITE(otmr, struct_itimerspec_sz);
+    if (otmr)
+      POST_WRITE(otmr, struct_itimerspec_sz);
   }
 }
 
@@ -2735,33 +2986,42 @@ POST_SYSCALL(old_readdir)(long res, long arg0, void *arg1, long arg2) {
   // Missing definition of 'struct old_linux_dirent'.
 }
 
-PRE_SYSCALL(pselect6)(long arg0, __sanitizer___kernel_fd_set *arg1,
-                      __sanitizer___kernel_fd_set *arg2,
-                      __sanitizer___kernel_fd_set *arg3, void *arg4,
-                      void *arg5) {}
+PRE_SYSCALL(pselect6)
+(long arg0, __sanitizer___kernel_fd_set *arg1,
+ __sanitizer___kernel_fd_set *arg2, __sanitizer___kernel_fd_set *arg3,
+ void *arg4, void *arg5) {}
 
-POST_SYSCALL(pselect6)(long res, long arg0, __sanitizer___kernel_fd_set *arg1,
-                       __sanitizer___kernel_fd_set *arg2,
-                       __sanitizer___kernel_fd_set *arg3, void *arg4,
-                       void *arg5) {
+POST_SYSCALL(pselect6)
+(long res, long arg0, __sanitizer___kernel_fd_set *arg1,
+ __sanitizer___kernel_fd_set *arg2, __sanitizer___kernel_fd_set *arg3,
+ void *arg4, void *arg5) {
   if (res >= 0) {
-    if (arg1) POST_WRITE(arg1, sizeof(*arg1));
-    if (arg2) POST_WRITE(arg2, sizeof(*arg2));
-    if (arg3) POST_WRITE(arg3, sizeof(*arg3));
-    if (arg4) POST_WRITE(arg4, struct_timespec_sz);
+    if (arg1)
+      POST_WRITE(arg1, sizeof(*arg1));
+    if (arg2)
+      POST_WRITE(arg2, sizeof(*arg2));
+    if (arg3)
+      POST_WRITE(arg3, sizeof(*arg3));
+    if (arg4)
+      POST_WRITE(arg4, struct_timespec_sz);
   }
 }
 
-PRE_SYSCALL(ppoll)(__sanitizer_pollfd *arg0, long arg1, void *arg2,
-                   const kernel_sigset_t *arg3, long arg4) {
-  if (arg3) PRE_READ(arg3, arg4);
+PRE_SYSCALL(ppoll)
+(__sanitizer_pollfd *arg0, long arg1, void *arg2, const kernel_sigset_t *arg3,
+ long arg4) {
+  if (arg3)
+    PRE_READ(arg3, arg4);
 }
 
-POST_SYSCALL(ppoll)(long res, __sanitizer_pollfd *arg0, long arg1, void *arg2,
-                    const void *arg3, long arg4) {
+POST_SYSCALL(ppoll)
+(long res, __sanitizer_pollfd *arg0, long arg1, void *arg2, const void *arg3,
+ long arg4) {
   if (res >= 0) {
-    if (arg0) POST_WRITE(arg0, sizeof(*arg0));
-    if (arg2) POST_WRITE(arg2, struct_timespec_sz);
+    if (arg0)
+      POST_WRITE(arg0, sizeof(*arg0));
+    if (arg2)
+      POST_WRITE(arg2, struct_timespec_sz);
   }
 }
 
@@ -2769,81 +3029,79 @@ PRE_SYSCALL(syncfs)(long fd) {}
 
 POST_SYSCALL(syncfs)(long res, long fd) {}
 
-PRE_SYSCALL(perf_event_open)(__sanitizer_perf_event_attr *attr_uptr, long pid,
-                             long cpu, long group_fd, long flags) {
-  if (attr_uptr) PRE_READ(attr_uptr, attr_uptr->size);
+PRE_SYSCALL(perf_event_open)
+(__sanitizer_perf_event_attr *attr_uptr, long pid, long cpu, long group_fd,
+ long flags) {
+  if (attr_uptr)
+    PRE_READ(attr_uptr, attr_uptr->size);
 }
 
-POST_SYSCALL(perf_event_open)(long res, __sanitizer_perf_event_attr *attr_uptr,
-                              long pid, long cpu, long group_fd, long flags) {}
+POST_SYSCALL(perf_event_open)
+(long res, __sanitizer_perf_event_attr *attr_uptr, long pid, long cpu,
+ long group_fd, long flags) {}
 
-PRE_SYSCALL(mmap_pgoff)(long addr, long len, long prot, long flags, long fd,
-                        long pgoff) {}
+PRE_SYSCALL(mmap_pgoff)
+(long addr, long len, long prot, long flags, long fd, long pgoff) {}
 
-POST_SYSCALL(mmap_pgoff)(long res, long addr, long len, long prot, long flags,
-                         long fd, long pgoff) {}
+POST_SYSCALL(mmap_pgoff)
+(long res, long addr, long len, long prot, long flags, long fd, long pgoff) {}
 
 PRE_SYSCALL(old_mmap)(void *arg) {}
 
 POST_SYSCALL(old_mmap)(long res, void *arg) {}
 
-PRE_SYSCALL(name_to_handle_at)(long dfd, const void *name, void *handle,
-                               void *mnt_id, long flag) {}
+PRE_SYSCALL(name_to_handle_at)
+(long dfd, const void *name, void *handle, void *mnt_id, long flag) {}
 
-POST_SYSCALL(name_to_handle_at)(long res, long dfd, const void *name,
-                                void *handle, void *mnt_id, long flag) {}
+POST_SYSCALL(name_to_handle_at)
+(long res, long dfd, const void *name, void *handle, void *mnt_id, long flag) {}
 
 PRE_SYSCALL(open_by_handle_at)(long mountdirfd, void *handle, long flags) {}
 
-POST_SYSCALL(open_by_handle_at)(long res, long mountdirfd, void *handle,
-                                long flags) {}
+POST_SYSCALL(open_by_handle_at)
+(long res, long mountdirfd, void *handle, long flags) {}
 
 PRE_SYSCALL(setns)(long fd, long nstype) {}
 
 POST_SYSCALL(setns)(long res, long fd, long nstype) {}
 
-PRE_SYSCALL(process_vm_readv)(long pid, const __sanitizer_iovec *lvec,
-                              long liovcnt, const void *rvec, long riovcnt,
-                              long flags) {}
+PRE_SYSCALL(process_vm_readv)
+(long pid, const __sanitizer_iovec *lvec, long liovcnt, const void *rvec,
+ long riovcnt, long flags) {}
 
-POST_SYSCALL(process_vm_readv)(long res, long pid,
-                               const __sanitizer_iovec *lvec, long liovcnt,
-                               const void *rvec, long riovcnt, long flags) {
+POST_SYSCALL(process_vm_readv)
+(long res, long pid, const __sanitizer_iovec *lvec, long liovcnt,
+ const void *rvec, long riovcnt, long flags) {
   if (res >= 0) {
-    if (lvec) kernel_write_iovec(lvec, liovcnt, res);
+    if (lvec)
+      kernel_write_iovec(lvec, liovcnt, res);
   }
 }
 
-PRE_SYSCALL(process_vm_writev)(long pid, const __sanitizer_iovec *lvec,
-                               long liovcnt, const void *rvec, long riovcnt,
-                               long flags) {}
+PRE_SYSCALL(process_vm_writev)
+(long pid, const __sanitizer_iovec *lvec, long liovcnt, const void *rvec,
+ long riovcnt, long flags) {}
 
-POST_SYSCALL(process_vm_writev)(long res, long pid,
-                                const __sanitizer_iovec *lvec, long liovcnt,
-                                const void *rvec, long riovcnt, long flags) {
+POST_SYSCALL(process_vm_writev)
+(long res, long pid, const __sanitizer_iovec *lvec, long liovcnt,
+ const void *rvec, long riovcnt, long flags) {
   if (res >= 0) {
-    if (lvec) kernel_read_iovec(lvec, liovcnt, res);
+    if (lvec)
+      kernel_read_iovec(lvec, liovcnt, res);
   }
 }
 
-PRE_SYSCALL(fork)() {
-  COMMON_SYSCALL_PRE_FORK();
-}
+PRE_SYSCALL(fork)() { COMMON_SYSCALL_PRE_FORK(); }
 
-POST_SYSCALL(fork)(long res) {
-  COMMON_SYSCALL_POST_FORK(res);
-}
+POST_SYSCALL(fork)(long res) { COMMON_SYSCALL_POST_FORK(res); }
 
-PRE_SYSCALL(vfork)() {
-  COMMON_SYSCALL_PRE_FORK();
-}
+PRE_SYSCALL(vfork)() { COMMON_SYSCALL_PRE_FORK(); }
 
-POST_SYSCALL(vfork)(long res) {
-  COMMON_SYSCALL_POST_FORK(res);
-}
+POST_SYSCALL(vfork)(long res) { COMMON_SYSCALL_POST_FORK(res); }
 
-PRE_SYSCALL(sigaction)(long signum, const __sanitizer_kernel_sigaction_t *act,
-                       __sanitizer_kernel_sigaction_t *oldact) {
+PRE_SYSCALL(sigaction)
+(long signum, const __sanitizer_kernel_sigaction_t *act,
+ __sanitizer_kernel_sigaction_t *oldact) {
   if (act) {
     PRE_READ(&act->sigaction, sizeof(act->sigaction));
     PRE_READ(&act->sa_flags, sizeof(act->sa_flags));
@@ -2851,15 +3109,16 @@ PRE_SYSCALL(sigaction)(long signum, const __sanitizer_kernel_sigaction_t *act,
   }
 }
 
-POST_SYSCALL(sigaction)(long res, long signum,
-                        const __sanitizer_kernel_sigaction_t *act,
-                        __sanitizer_kernel_sigaction_t *oldact) {
-  if (res >= 0 && oldact) POST_WRITE(oldact, sizeof(*oldact));
+POST_SYSCALL(sigaction)
+(long res, long signum, const __sanitizer_kernel_sigaction_t *act,
+ __sanitizer_kernel_sigaction_t *oldact) {
+  if (res >= 0 && oldact)
+    POST_WRITE(oldact, sizeof(*oldact));
 }
 
-PRE_SYSCALL(rt_sigaction)(long signum,
-                          const __sanitizer_kernel_sigaction_t *act,
-                          __sanitizer_kernel_sigaction_t *oldact, SIZE_T sz) {
+PRE_SYSCALL(rt_sigaction)
+(long signum, const __sanitizer_kernel_sigaction_t *act,
+ __sanitizer_kernel_sigaction_t *oldact, SIZE_T sz) {
   if (act) {
     PRE_READ(&act->sigaction, sizeof(act->sigaction));
     PRE_READ(&act->sa_flags, sizeof(act->sa_flags));
@@ -2867,9 +3126,9 @@ PRE_SYSCALL(rt_sigaction)(long signum,
   }
 }
 
-POST_SYSCALL(rt_sigaction)(long res, long signum,
-                           const __sanitizer_kernel_sigaction_t *act,
-                           __sanitizer_kernel_sigaction_t *oldact, SIZE_T sz) {
+POST_SYSCALL(rt_sigaction)
+(long res, long signum, const __sanitizer_kernel_sigaction_t *act,
+ __sanitizer_kernel_sigaction_t *oldact, SIZE_T sz) {
   if (res >= 0 && oldact) {
     SIZE_T oldact_sz = ((char *)&oldact->sa_mask) - ((char *)oldact) + sz;
     POST_WRITE(oldact, oldact_sz);
@@ -2906,11 +3165,11 @@ POST_SYSCALL(sigaltstack)(long res, void *ss, void *oss) {
 }
 }  // extern "C"
 
-#undef PRE_SYSCALL
-#undef PRE_READ
-#undef PRE_WRITE
-#undef POST_SYSCALL
-#undef POST_READ
-#undef POST_WRITE
+#  undef PRE_SYSCALL
+#  undef PRE_READ
+#  undef PRE_WRITE
+#  undef POST_SYSCALL
+#  undef POST_READ
+#  undef POST_WRITE
 
 #endif  // SANITIZER_LINUX
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_coverage_fuchsia.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_coverage_fuchsia.cpp
index a52db08433e3..1d0dbe592b93 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_coverage_fuchsia.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_coverage_fuchsia.cpp
@@ -51,6 +51,8 @@ constexpr const char kSancovSinkName[] = "sancov";
 // This class relies on zero-initialization.
 class TracePcGuardController final {
  public:
+  constexpr TracePcGuardController() {}
+
   // For each PC location being tracked, there is a u32 reserved in global
   // data called the "guard".  At startup, we assign each guard slot a
   // unique index into the big results array.  Later during runtime, the
@@ -87,7 +89,7 @@ class TracePcGuardController final {
   }
 
   void Dump() {
-    BlockingMutexLock locked(&setup_lock_);
+    Lock locked(&setup_lock_);
     if (array_) {
       CHECK_NE(vmo_, ZX_HANDLE_INVALID);
 
@@ -114,7 +116,7 @@ class TracePcGuardController final {
   // We can always spare the 32G of address space.
   static constexpr size_t MappingSize = sizeof(uptr) << 32;
 
-  BlockingMutex setup_lock_ = BlockingMutex(LINKER_INITIALIZED);
+  Mutex setup_lock_;
   uptr *array_ = nullptr;
   u32 next_index_ = 0;
   zx_handle_t vmo_ = {};
@@ -123,7 +125,7 @@ class TracePcGuardController final {
   size_t DataSize() const { return next_index_ * sizeof(uintptr_t); }
 
   u32 Setup(u32 num_guards) {
-    BlockingMutexLock locked(&setup_lock_);
+    Lock locked(&setup_lock_);
     DCHECK(common_flags()->coverage);
 
     if (next_index_ == 0) {
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cpp
index 73ebeb5fa14a..56220df2ac18 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cpp
@@ -73,7 +73,7 @@ static void SanitizerDumpCoverage(const uptr* unsorted_pcs, uptr len) {
     if (!pc) continue;
 
     if (!__sanitizer_get_module_and_offset_for_pc(pc, nullptr, 0, &pcs[i])) {
-      Printf("ERROR: unknown pc 0x%x (may happen if dlclose is used)\n", pc);
+      Printf("ERROR: unknown pc 0x%zx (may happen if dlclose is used)\n", pc);
       continue;
     }
     uptr module_base = pc - pcs[i];
@@ -151,6 +151,55 @@ class TracePcGuardController {
 
 static TracePcGuardController pc_guard_controller;
 
+// A basic default implementation of callbacks for
+// -fsanitize-coverage=inline-8bit-counters,pc-table.
+// Use TOOL_OPTIONS (UBSAN_OPTIONS, etc) to dump the coverage data:
+// * cov_8bit_counters_out=PATH to dump the 8bit counters.
+// * cov_pcs_out=PATH to dump the pc table.
+//
+// Most users will still need to define their own callbacks for greater
+// flexibility.
+namespace SingletonCounterCoverage {
+
+static char *counters_beg, *counters_end;
+static const uptr *pcs_beg, *pcs_end;
+
+static void DumpCoverage() {
+  const char* file_path = common_flags()->cov_8bit_counters_out;
+  if (file_path && internal_strlen(file_path)) {
+    fd_t fd = OpenFile(file_path);
+    FileCloser file_closer(fd);
+    uptr size = counters_end - counters_beg;
+    WriteToFile(fd, counters_beg, size);
+    if (common_flags()->verbosity)
+      __sanitizer::Printf("cov_8bit_counters_out: written %zd bytes to %s\n",
+                          size, file_path);
+  }
+  file_path = common_flags()->cov_pcs_out;
+  if (file_path && internal_strlen(file_path)) {
+    fd_t fd = OpenFile(file_path);
+    FileCloser file_closer(fd);
+    uptr size = (pcs_end - pcs_beg) * sizeof(uptr);
+    WriteToFile(fd, pcs_beg, size);
+    if (common_flags()->verbosity)
+      __sanitizer::Printf("cov_pcs_out: written %zd bytes to %s\n", size,
+                          file_path);
+  }
+}
+
+static void Cov8bitCountersInit(char* beg, char* end) {
+  counters_beg = beg;
+  counters_end = end;
+  Atexit(DumpCoverage);
+}
+
+static void CovPcsInit(const uptr* beg, const uptr* end) {
+  pcs_beg = beg;
+  pcs_end = end;
+}
+
+}  // namespace SingletonCounterCoverage
+
 }  // namespace
 }  // namespace __sancov
 
@@ -191,7 +240,9 @@ SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_cov_dump() {
 SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_cov_reset() {
   __sancov::pc_guard_controller.Reset();
 }
-// Default empty implementations (weak). Users should redefine them.
+// Default implementations (weak).
+// Either empty or very simple.
+// Most users should redefine them.
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_cmp, void) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_cmp1, void) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_cmp2, void) {}
@@ -206,9 +257,15 @@ SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_div4, void) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_div8, void) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_gep, void) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_pc_indir, void) {}
-SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_8bit_counters_init, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_8bit_counters_init,
+                             char* start, char* end) {
+  __sancov::SingletonCounterCoverage::Cov8bitCountersInit(start, end);
+}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_bool_flag_init, void) {}
-SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_pcs_init, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_pcs_init, const uptr* beg,
+                             const uptr* end) {
+  __sancov::SingletonCounterCoverage::CovPcsInit(beg, end);
+}
 }  // extern "C"
 // Weak definition for code instrumented with -fsanitize-coverage=stack-depth
 // and later linked with code containing a strong definition.
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_deadlock_detector.h b/compiler-rt/lib/sanitizer_common/sanitizer_deadlock_detector.h
index b80cff460eda..0749f633b4bc 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_deadlock_detector.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_deadlock_detector.h
@@ -293,7 +293,7 @@ class DeadlockDetector {
   }
 
   // Returns true iff dtls is empty (no locks are currently held) and we can
-  // add the node to the currently held locks w/o chanding the global state.
+  // add the node to the currently held locks w/o changing the global state.
   // This operation is thread-safe as it only touches the dtls.
   bool onFirstLock(DeadlockDetectorTLS<BV> *dtls, uptr node, u32 stk = 0) {
     if (!dtls->empty()) return false;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_dense_map.h b/compiler-rt/lib/sanitizer_common/sanitizer_dense_map.h
new file mode 100644
index 000000000000..3fa6af76ce29
--- /dev/null
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_dense_map.h
@@ -0,0 +1,678 @@
+//===- sanitizer_dense_map.h - Dense probed hash table ----------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This is fork of llvm/ADT/DenseMap.h class with the following changes:
+//  * Use mmap to allocate.
+//  * No iterators.
+//  * Does not shrink.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef SANITIZER_DENSE_MAP_H
+#define SANITIZER_DENSE_MAP_H
+
+#include "sanitizer_common.h"
+#include "sanitizer_dense_map_info.h"
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_type_traits.h"
+
+namespace __sanitizer {
+
+template <typename DerivedT, typename KeyT, typename ValueT, typename KeyInfoT,
+          typename BucketT>
+class DenseMapBase {
+ public:
+  using size_type = unsigned;
+  using key_type = KeyT;
+  using mapped_type = ValueT;
+  using value_type = BucketT;
+
+  WARN_UNUSED_RESULT bool empty() const { return getNumEntries() == 0; }
+  unsigned size() const { return getNumEntries(); }
+
+  /// Grow the densemap so that it can contain at least \p NumEntries items
+  /// before resizing again.
+  void reserve(size_type NumEntries) {
+    auto NumBuckets = getMinBucketToReserveForEntries(NumEntries);
+    if (NumBuckets > getNumBuckets())
+      grow(NumBuckets);
+  }
+
+  void clear() {
+    if (getNumEntries() == 0 && getNumTombstones() == 0)
+      return;
+
+    const KeyT EmptyKey = getEmptyKey(), TombstoneKey = getTombstoneKey();
+    if (__sanitizer::is_trivially_destructible<ValueT>::value) {
+      // Use a simpler loop when values don't need destruction.
+      for (BucketT *P = getBuckets(), *E = getBucketsEnd(); P != E; ++P)
+        P->getFirst() = EmptyKey;
+    } else {
+      unsigned NumEntries = getNumEntries();
+      for (BucketT *P = getBuckets(), *E = getBucketsEnd(); P != E; ++P) {
+        if (!KeyInfoT::isEqual(P->getFirst(), EmptyKey)) {
+          if (!KeyInfoT::isEqual(P->getFirst(), TombstoneKey)) {
+            P->getSecond().~ValueT();
+            --NumEntries;
+          }
+          P->getFirst() = EmptyKey;
+        }
+      }
+      CHECK_EQ(NumEntries, 0);
+    }
+    setNumEntries(0);
+    setNumTombstones(0);
+  }
+
+  /// Return 1 if the specified key is in the map, 0 otherwise.
+  size_type count(const KeyT &Key) const {
+    const BucketT *TheBucket;
+    return LookupBucketFor(Key, TheBucket) ? 1 : 0;
+  }
+
+  value_type *find(const KeyT &Key) {
+    BucketT *TheBucket;
+    if (LookupBucketFor(Key, TheBucket))
+      return TheBucket;
+    return nullptr;
+  }
+  const value_type *find(const KeyT &Key) const {
+    const BucketT *TheBucket;
+    if (LookupBucketFor(Key, TheBucket))
+      return TheBucket;
+    return nullptr;
+  }
+
+  /// Alternate version of find() which allows a different, and possibly
+  /// less expensive, key type.
+  /// The DenseMapInfo is responsible for supplying methods
+  /// getHashValue(LookupKeyT) and isEqual(LookupKeyT, KeyT) for each key
+  /// type used.
+  template <class LookupKeyT>
+  value_type *find_as(const LookupKeyT &Key) {
+    BucketT *TheBucket;
+    if (LookupBucketFor(Key, TheBucket))
+      return TheBucket;
+    return nullptr;
+  }
+  template <class LookupKeyT>
+  const value_type *find_as(const LookupKeyT &Key) const {
+    const BucketT *TheBucket;
+    if (LookupBucketFor(Key, TheBucket))
+      return TheBucket;
+    return nullptr;
+  }
+
+  /// lookup - Return the entry for the specified key, or a default
+  /// constructed value if no such entry exists.
+  ValueT lookup(const KeyT &Key) const {
+    const BucketT *TheBucket;
+    if (LookupBucketFor(Key, TheBucket))
+      return TheBucket->getSecond();
+    return ValueT();
+  }
+
+  // Inserts key,value pair into the map if the key isn't already in the map.
+  // If the key is already in the map, it returns false and doesn't update the
+  // value.
+  detail::DenseMapPair<value_type *, bool> insert(const value_type &KV) {
+    return try_emplace(KV.first, KV.second);
+  }
+
+  // Inserts key,value pair into the map if the key isn't already in the map.
+  // If the key is already in the map, it returns false and doesn't update the
+  // value.
+  detail::DenseMapPair<value_type *, bool> insert(value_type &&KV) {
+    return try_emplace(__sanitizer::move(KV.first),
+                       __sanitizer::move(KV.second));
+  }
+
+  // Inserts key,value pair into the map if the key isn't already in the map.
+  // The value is constructed in-place if the key is not in the map, otherwise
+  // it is not moved.
+  template <typename... Ts>
+  detail::DenseMapPair<value_type *, bool> try_emplace(KeyT &&Key,
+                                                       Ts &&...Args) {
+    BucketT *TheBucket;
+    if (LookupBucketFor(Key, TheBucket))
+      return {TheBucket, false};  // Already in map.
+
+    // Otherwise, insert the new element.
+    TheBucket = InsertIntoBucket(TheBucket, __sanitizer::move(Key),
+                                 __sanitizer::forward<Ts>(Args)...);
+    return {TheBucket, true};
+  }
+
+  // Inserts key,value pair into the map if the key isn't already in the map.
+  // The value is constructed in-place if the key is not in the map, otherwise
+  // it is not moved.
+  template <typename... Ts>
+  detail::DenseMapPair<value_type *, bool> try_emplace(const KeyT &Key,
+                                                       Ts &&...Args) {
+    BucketT *TheBucket;
+    if (LookupBucketFor(Key, TheBucket))
+      return {TheBucket, false};  // Already in map.
+
+    // Otherwise, insert the new element.
+    TheBucket =
+        InsertIntoBucket(TheBucket, Key, __sanitizer::forward<Ts>(Args)...);
+    return {TheBucket, true};
+  }
+
+  /// Alternate version of insert() which allows a different, and possibly
+  /// less expensive, key type.
+  /// The DenseMapInfo is responsible for supplying methods
+  /// getHashValue(LookupKeyT) and isEqual(LookupKeyT, KeyT) for each key
+  /// type used.
+  template <typename LookupKeyT>
+  detail::DenseMapPair<value_type *, bool> insert_as(value_type &&KV,
+                                                     const LookupKeyT &Val) {
+    BucketT *TheBucket;
+    if (LookupBucketFor(Val, TheBucket))
+      return {TheBucket, false};  // Already in map.
+
+    // Otherwise, insert the new element.
+    TheBucket =
+        InsertIntoBucketWithLookup(TheBucket, __sanitizer::move(KV.first),
+                                   __sanitizer::move(KV.second), Val);
+    return {TheBucket, true};
+  }
+
+  bool erase(const KeyT &Val) {
+    BucketT *TheBucket;
+    if (!LookupBucketFor(Val, TheBucket))
+      return false;  // not in map.
+
+    TheBucket->getSecond().~ValueT();
+    TheBucket->getFirst() = getTombstoneKey();
+    decrementNumEntries();
+    incrementNumTombstones();
+    return true;
+  }
+
+  void erase(value_type *I) {
+    CHECK_NE(I, nullptr);
+    BucketT *TheBucket = &*I;
+    TheBucket->getSecond().~ValueT();
+    TheBucket->getFirst() = getTombstoneKey();
+    decrementNumEntries();
+    incrementNumTombstones();
+  }
+
+  value_type &FindAndConstruct(const KeyT &Key) {
+    BucketT *TheBucket;
+    if (LookupBucketFor(Key, TheBucket))
+      return *TheBucket;
+
+    return *InsertIntoBucket(TheBucket, Key);
+  }
+
+  ValueT &operator[](const KeyT &Key) { return FindAndConstruct(Key).second; }
+
+  value_type &FindAndConstruct(KeyT &&Key) {
+    BucketT *TheBucket;
+    if (LookupBucketFor(Key, TheBucket))
+      return *TheBucket;
+
+    return *InsertIntoBucket(TheBucket, __sanitizer::move(Key));
+  }
+
+  ValueT &operator[](KeyT &&Key) {
+    return FindAndConstruct(__sanitizer::move(Key)).second;
+  }
+
+  /// Equality comparison for DenseMap.
+  ///
+  /// Iterates over elements of LHS confirming that each (key, value) pair in
+  /// LHS is also in RHS, and that no additional pairs are in RHS. Equivalent to
+  /// N calls to RHS.find and N value comparisons. Amortized complexity is
+  /// linear, worst case is O(N^2) (if every hash collides).
+  bool operator==(const DenseMapBase &RHS) const {
+    if (size() != RHS.size())
+      return false;
+
+    const KeyT EmptyKey = getEmptyKey(), TombstoneKey = getTombstoneKey();
+    for (auto *P = getBuckets(), *E = getBucketsEnd(); P != E; ++P) {
+      const KeyT K = P->getFirst();
+      if (!KeyInfoT::isEqual(K, EmptyKey) &&
+          !KeyInfoT::isEqual(K, TombstoneKey)) {
+        const auto *I = RHS.find(K);
+        if (!I || P->getSecond() != I->getSecond())
+          return false;
+      }
+    }
+
+    return true;
+  }
+
+ protected:
+  DenseMapBase() = default;
+
+  void destroyAll() {
+    if (getNumBuckets() == 0)  // Nothing to do.
+      return;
+
+    const KeyT EmptyKey = getEmptyKey(), TombstoneKey = getTombstoneKey();
+    for (BucketT *P = getBuckets(), *E = getBucketsEnd(); P != E; ++P) {
+      if (!KeyInfoT::isEqual(P->getFirst(), EmptyKey) &&
+          !KeyInfoT::isEqual(P->getFirst(), TombstoneKey))
+        P->getSecond().~ValueT();
+      P->getFirst().~KeyT();
+    }
+  }
+
+  void initEmpty() {
+    setNumEntries(0);
+    setNumTombstones(0);
+
+    CHECK_EQ((getNumBuckets() & (getNumBuckets() - 1)), 0);
+    const KeyT EmptyKey = getEmptyKey();
+    for (BucketT *B = getBuckets(), *E = getBucketsEnd(); B != E; ++B)
+      ::new (&B->getFirst()) KeyT(EmptyKey);
+  }
+
+  /// Returns the number of buckets to allocate to ensure that the DenseMap can
+  /// accommodate \p NumEntries without need to grow().
+  unsigned getMinBucketToReserveForEntries(unsigned NumEntries) {
+    // Ensure that "NumEntries * 4 < NumBuckets * 3"
+    if (NumEntries == 0)
+      return 0;
+    // +1 is required because of the strict equality.
+    // For example if NumEntries is 48, we need to return 401.
+    return RoundUpToPowerOfTwo((NumEntries * 4 / 3 + 1) + /* NextPowerOf2 */ 1);
+  }
+
+  void moveFromOldBuckets(BucketT *OldBucketsBegin, BucketT *OldBucketsEnd) {
+    initEmpty();
+
+    // Insert all the old elements.
+    const KeyT EmptyKey = getEmptyKey();
+    const KeyT TombstoneKey = getTombstoneKey();
+    for (BucketT *B = OldBucketsBegin, *E = OldBucketsEnd; B != E; ++B) {
+      if (!KeyInfoT::isEqual(B->getFirst(), EmptyKey) &&
+          !KeyInfoT::isEqual(B->getFirst(), TombstoneKey)) {
+        // Insert the key/value into the new table.
+        BucketT *DestBucket;
+        bool FoundVal = LookupBucketFor(B->getFirst(), DestBucket);
+        (void)FoundVal;  // silence warning.
+        CHECK(!FoundVal);
+        DestBucket->getFirst() = __sanitizer::move(B->getFirst());
+        ::new (&DestBucket->getSecond())
+            ValueT(__sanitizer::move(B->getSecond()));
+        incrementNumEntries();
+
+        // Free the value.
+        B->getSecond().~ValueT();
+      }
+      B->getFirst().~KeyT();
+    }
+  }
+
+  template <typename OtherBaseT>
+  void copyFrom(
+      const DenseMapBase<OtherBaseT, KeyT, ValueT, KeyInfoT, BucketT> &other) {
+    CHECK_NE(&other, this);
+    CHECK_EQ(getNumBuckets(), other.getNumBuckets());
+
+    setNumEntries(other.getNumEntries());
+    setNumTombstones(other.getNumTombstones());
+
+    if (__sanitizer::is_trivially_copyable<KeyT>::value &&
+        __sanitizer::is_trivially_copyable<ValueT>::value)
+      internal_memcpy(reinterpret_cast<void *>(getBuckets()),
+                      other.getBuckets(), getNumBuckets() * sizeof(BucketT));
+    else
+      for (uptr i = 0; i < getNumBuckets(); ++i) {
+        ::new (&getBuckets()[i].getFirst())
+            KeyT(other.getBuckets()[i].getFirst());
+        if (!KeyInfoT::isEqual(getBuckets()[i].getFirst(), getEmptyKey()) &&
+            !KeyInfoT::isEqual(getBuckets()[i].getFirst(), getTombstoneKey()))
+          ::new (&getBuckets()[i].getSecond())
+              ValueT(other.getBuckets()[i].getSecond());
+      }
+  }
+
+  static unsigned getHashValue(const KeyT &Val) {
+    return KeyInfoT::getHashValue(Val);
+  }
+
+  template <typename LookupKeyT>
+  static unsigned getHashValue(const LookupKeyT &Val) {
+    return KeyInfoT::getHashValue(Val);
+  }
+
+  static const KeyT getEmptyKey() { return KeyInfoT::getEmptyKey(); }
+
+  static const KeyT getTombstoneKey() { return KeyInfoT::getTombstoneKey(); }
+
+ private:
+  unsigned getNumEntries() const {
+    return static_cast<const DerivedT *>(this)->getNumEntries();
+  }
+
+  void setNumEntries(unsigned Num) {
+    static_cast<DerivedT *>(this)->setNumEntries(Num);
+  }
+
+  void incrementNumEntries() { setNumEntries(getNumEntries() + 1); }
+
+  void decrementNumEntries() { setNumEntries(getNumEntries() - 1); }
+
+  unsigned getNumTombstones() const {
+    return static_cast<const DerivedT *>(this)->getNumTombstones();
+  }
+
+  void setNumTombstones(unsigned Num) {
+    static_cast<DerivedT *>(this)->setNumTombstones(Num);
+  }
+
+  void incrementNumTombstones() { setNumTombstones(getNumTombstones() + 1); }
+
+  void decrementNumTombstones() { setNumTombstones(getNumTombstones() - 1); }
+
+  const BucketT *getBuckets() const {
+    return static_cast<const DerivedT *>(this)->getBuckets();
+  }
+
+  BucketT *getBuckets() { return static_cast<DerivedT *>(this)->getBuckets(); }
+
+  unsigned getNumBuckets() const {
+    return static_cast<const DerivedT *>(this)->getNumBuckets();
+  }
+
+  BucketT *getBucketsEnd() { return getBuckets() + getNumBuckets(); }
+
+  const BucketT *getBucketsEnd() const {
+    return getBuckets() + getNumBuckets();
+  }
+
+  void grow(unsigned AtLeast) { static_cast<DerivedT *>(this)->grow(AtLeast); }
+
+  template <typename KeyArg, typename... ValueArgs>
+  BucketT *InsertIntoBucket(BucketT *TheBucket, KeyArg &&Key,
+                            ValueArgs &&...Values) {
+    TheBucket = InsertIntoBucketImpl(Key, Key, TheBucket);
+
+    TheBucket->getFirst() = __sanitizer::forward<KeyArg>(Key);
+    ::new (&TheBucket->getSecond())
+        ValueT(__sanitizer::forward<ValueArgs>(Values)...);
+    return TheBucket;
+  }
+
+  template <typename LookupKeyT>
+  BucketT *InsertIntoBucketWithLookup(BucketT *TheBucket, KeyT &&Key,
+                                      ValueT &&Value, LookupKeyT &Lookup) {
+    TheBucket = InsertIntoBucketImpl(Key, Lookup, TheBucket);
+
+    TheBucket->getFirst() = __sanitizer::move(Key);
+    ::new (&TheBucket->getSecond()) ValueT(__sanitizer::move(Value));
+    return TheBucket;
+  }
+
+  template <typename LookupKeyT>
+  BucketT *InsertIntoBucketImpl(const KeyT &Key, const LookupKeyT &Lookup,
+                                BucketT *TheBucket) {
+    // If the load of the hash table is more than 3/4, or if fewer than 1/8 of
+    // the buckets are empty (meaning that many are filled with tombstones),
+    // grow the table.
+    //
+    // The later case is tricky.  For example, if we had one empty bucket with
+    // tons of tombstones, failing lookups (e.g. for insertion) would have to
+    // probe almost the entire table until it found the empty bucket.  If the
+    // table completely filled with tombstones, no lookup would ever succeed,
+    // causing infinite loops in lookup.
+    unsigned NewNumEntries = getNumEntries() + 1;
+    unsigned NumBuckets = getNumBuckets();
+    if (UNLIKELY(NewNumEntries * 4 >= NumBuckets * 3)) {
+      this->grow(NumBuckets * 2);
+      LookupBucketFor(Lookup, TheBucket);
+      NumBuckets = getNumBuckets();
+    } else if (UNLIKELY(NumBuckets - (NewNumEntries + getNumTombstones()) <=
+                        NumBuckets / 8)) {
+      this->grow(NumBuckets);
+      LookupBucketFor(Lookup, TheBucket);
+    }
+    CHECK(TheBucket);
+
+    // Only update the state after we've grown our bucket space appropriately
+    // so that when growing buckets we have self-consistent entry count.
+    incrementNumEntries();
+
+    // If we are writing over a tombstone, remember this.
+    const KeyT EmptyKey = getEmptyKey();
+    if (!KeyInfoT::isEqual(TheBucket->getFirst(), EmptyKey))
+      decrementNumTombstones();
+
+    return TheBucket;
+  }
+
+  /// LookupBucketFor - Lookup the appropriate bucket for Val, returning it in
+  /// FoundBucket.  If the bucket contains the key and a value, this returns
+  /// true, otherwise it returns a bucket with an empty marker or tombstone and
+  /// returns false.
+  template <typename LookupKeyT>
+  bool LookupBucketFor(const LookupKeyT &Val,
+                       const BucketT *&FoundBucket) const {
+    const BucketT *BucketsPtr = getBuckets();
+    const unsigned NumBuckets = getNumBuckets();
+
+    if (NumBuckets == 0) {
+      FoundBucket = nullptr;
+      return false;
+    }
+
+    // FoundTombstone - Keep track of whether we find a tombstone while probing.
+    const BucketT *FoundTombstone = nullptr;
+    const KeyT EmptyKey = getEmptyKey();
+    const KeyT TombstoneKey = getTombstoneKey();
+    CHECK(!KeyInfoT::isEqual(Val, EmptyKey));
+    CHECK(!KeyInfoT::isEqual(Val, TombstoneKey));
+
+    unsigned BucketNo = getHashValue(Val) & (NumBuckets - 1);
+    unsigned ProbeAmt = 1;
+    while (true) {
+      const BucketT *ThisBucket = BucketsPtr + BucketNo;
+      // Found Val's bucket?  If so, return it.
+      if (LIKELY(KeyInfoT::isEqual(Val, ThisBucket->getFirst()))) {
+        FoundBucket = ThisBucket;
+        return true;
+      }
+
+      // If we found an empty bucket, the key doesn't exist in the set.
+      // Insert it and return the default value.
+      if (LIKELY(KeyInfoT::isEqual(ThisBucket->getFirst(), EmptyKey))) {
+        // If we've already seen a tombstone while probing, fill it in instead
+        // of the empty bucket we eventually probed to.
+        FoundBucket = FoundTombstone ? FoundTombstone : ThisBucket;
+        return false;
+      }
+
+      // If this is a tombstone, remember it.  If Val ends up not in the map, we
+      // prefer to return it than something that would require more probing.
+      if (KeyInfoT::isEqual(ThisBucket->getFirst(), TombstoneKey) &&
+          !FoundTombstone)
+        FoundTombstone = ThisBucket;  // Remember the first tombstone found.
+
+      // Otherwise, it's a hash collision or a tombstone, continue quadratic
+      // probing.
+      BucketNo += ProbeAmt++;
+      BucketNo &= (NumBuckets - 1);
+    }
+  }
+
+  template <typename LookupKeyT>
+  bool LookupBucketFor(const LookupKeyT &Val, BucketT *&FoundBucket) {
+    const BucketT *ConstFoundBucket;
+    bool Result = const_cast<const DenseMapBase *>(this)->LookupBucketFor(
+        Val, ConstFoundBucket);
+    FoundBucket = const_cast<BucketT *>(ConstFoundBucket);
+    return Result;
+  }
+
+ public:
+  /// Return the approximate size (in bytes) of the actual map.
+  /// This is just the raw memory used by DenseMap.
+  /// If entries are pointers to objects, the size of the referenced objects
+  /// are not included.
+  uptr getMemorySize() const {
+    return RoundUpTo(getNumBuckets() * sizeof(BucketT), GetPageSizeCached());
+  }
+};
+
+/// Inequality comparison for DenseMap.
+///
+/// Equivalent to !(LHS == RHS). See operator== for performance notes.
+template <typename DerivedT, typename KeyT, typename ValueT, typename KeyInfoT,
+          typename BucketT>
+bool operator!=(
+    const DenseMapBase<DerivedT, KeyT, ValueT, KeyInfoT, BucketT> &LHS,
+    const DenseMapBase<DerivedT, KeyT, ValueT, KeyInfoT, BucketT> &RHS) {
+  return !(LHS == RHS);
+}
+
+template <typename KeyT, typename ValueT,
+          typename KeyInfoT = DenseMapInfo<KeyT>,
+          typename BucketT = detail::DenseMapPair<KeyT, ValueT>>
+class DenseMap : public DenseMapBase<DenseMap<KeyT, ValueT, KeyInfoT, BucketT>,
+                                     KeyT, ValueT, KeyInfoT, BucketT> {
+  friend class DenseMapBase<DenseMap, KeyT, ValueT, KeyInfoT, BucketT>;
+
+  // Lift some types from the dependent base class into this class for
+  // simplicity of referring to them.
+  using BaseT = DenseMapBase<DenseMap, KeyT, ValueT, KeyInfoT, BucketT>;
+
+  BucketT *Buckets = nullptr;
+  unsigned NumEntries = 0;
+  unsigned NumTombstones = 0;
+  unsigned NumBuckets = 0;
+
+ public:
+  /// Create a DenseMap with an optional \p InitialReserve that guarantee that
+  /// this number of elements can be inserted in the map without grow()
+  explicit DenseMap(unsigned InitialReserve) { init(InitialReserve); }
+  constexpr DenseMap() = default;
+
+  DenseMap(const DenseMap &other) : BaseT() {
+    init(0);
+    copyFrom(other);
+  }
+
+  DenseMap(DenseMap &&other) : BaseT() {
+    init(0);
+    swap(other);
+  }
+
+  ~DenseMap() {
+    this->destroyAll();
+    deallocate_buffer(Buckets, sizeof(BucketT) * NumBuckets);
+  }
+
+  void swap(DenseMap &RHS) {
+    Swap(Buckets, RHS.Buckets);
+    Swap(NumEntries, RHS.NumEntries);
+    Swap(NumTombstones, RHS.NumTombstones);
+    Swap(NumBuckets, RHS.NumBuckets);
+  }
+
+  DenseMap &operator=(const DenseMap &other) {
+    if (&other != this)
+      copyFrom(other);
+    return *this;
+  }
+
+  DenseMap &operator=(DenseMap &&other) {
+    this->destroyAll();
+    deallocate_buffer(Buckets, sizeof(BucketT) * NumBuckets, alignof(BucketT));
+    init(0);
+    swap(other);
+    return *this;
+  }
+
+  void copyFrom(const DenseMap &other) {
+    this->destroyAll();
+    deallocate_buffer(Buckets, sizeof(BucketT) * NumBuckets);
+    if (allocateBuckets(other.NumBuckets)) {
+      this->BaseT::copyFrom(other);
+    } else {
+      NumEntries = 0;
+      NumTombstones = 0;
+    }
+  }
+
+  void init(unsigned InitNumEntries) {
+    auto InitBuckets = BaseT::getMinBucketToReserveForEntries(InitNumEntries);
+    if (allocateBuckets(InitBuckets)) {
+      this->BaseT::initEmpty();
+    } else {
+      NumEntries = 0;
+      NumTombstones = 0;
+    }
+  }
+
+  void grow(unsigned AtLeast) {
+    unsigned OldNumBuckets = NumBuckets;
+    BucketT *OldBuckets = Buckets;
+
+    allocateBuckets(RoundUpToPowerOfTwo(Max<unsigned>(64, AtLeast)));
+    CHECK(Buckets);
+    if (!OldBuckets) {
+      this->BaseT::initEmpty();
+      return;
+    }
+
+    this->moveFromOldBuckets(OldBuckets, OldBuckets + OldNumBuckets);
+
+    // Free the old table.
+    deallocate_buffer(OldBuckets, sizeof(BucketT) * OldNumBuckets);
+  }
+
+ private:
+  unsigned getNumEntries() const { return NumEntries; }
+
+  void setNumEntries(unsigned Num) { NumEntries = Num; }
+
+  unsigned getNumTombstones() const { return NumTombstones; }
+
+  void setNumTombstones(unsigned Num) { NumTombstones = Num; }
+
+  BucketT *getBuckets() const { return Buckets; }
+
+  unsigned getNumBuckets() const { return NumBuckets; }
+
+  bool allocateBuckets(unsigned Num) {
+    NumBuckets = Num;
+    if (NumBuckets == 0) {
+      Buckets = nullptr;
+      return false;
+    }
+
+    uptr Size = sizeof(BucketT) * NumBuckets;
+    if (Size * 2 <= GetPageSizeCached()) {
+      // We always allocate at least a page, so use entire space.
+      unsigned Log2 = MostSignificantSetBitIndex(GetPageSizeCached() / Size);
+      Size <<= Log2;
+      NumBuckets <<= Log2;
+      CHECK_EQ(Size, sizeof(BucketT) * NumBuckets);
+      CHECK_GT(Size * 2, GetPageSizeCached());
+    }
+    Buckets = static_cast<BucketT *>(allocate_buffer(Size));
+    return true;
+  }
+
+  static void *allocate_buffer(uptr Size) {
+    return MmapOrDie(RoundUpTo(Size, GetPageSizeCached()), "DenseMap");
+  }
+
+  static void deallocate_buffer(void *Ptr, uptr Size) {
+    UnmapOrDie(Ptr, RoundUpTo(Size, GetPageSizeCached()));
+  }
+};
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_DENSE_MAP_H
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_dense_map_info.h b/compiler-rt/lib/sanitizer_common/sanitizer_dense_map_info.h
new file mode 100644
index 000000000000..85c6427906c1
--- /dev/null
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_dense_map_info.h
@@ -0,0 +1,260 @@
+//===- sanitizer_dense_map_info.h - Type traits for DenseMap ----*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef SANITIZER_DENSE_MAP_INFO_H
+#define SANITIZER_DENSE_MAP_INFO_H
+
+#include "sanitizer_common.h"
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_type_traits.h"
+
+namespace __sanitizer {
+
+namespace detail {
+
+/// Simplistic combination of 32-bit hash values into 32-bit hash values.
+static inline unsigned combineHashValue(unsigned a, unsigned b) {
+  u64 key = (u64)a << 32 | (u64)b;
+  key += ~(key << 32);
+  key ^= (key >> 22);
+  key += ~(key << 13);
+  key ^= (key >> 8);
+  key += (key << 3);
+  key ^= (key >> 15);
+  key += ~(key << 27);
+  key ^= (key >> 31);
+  return (unsigned)key;
+}
+
+// We extend a pair to allow users to override the bucket type with their own
+// implementation without requiring two members.
+template <typename KeyT, typename ValueT>
+struct DenseMapPair {
+  KeyT first = {};
+  ValueT second = {};
+  DenseMapPair() = default;
+  DenseMapPair(const KeyT &f, const ValueT &s) : first(f), second(s) {}
+
+  template <typename KeyT2, typename ValueT2>
+  DenseMapPair(KeyT2 &&f, ValueT2 &&s)
+      : first(__sanitizer::forward<KeyT2>(f)),
+        second(__sanitizer::forward<ValueT2>(s)) {}
+
+  DenseMapPair(const DenseMapPair &other) = default;
+  DenseMapPair &operator=(const DenseMapPair &other) = default;
+  DenseMapPair(DenseMapPair &&other) = default;
+  DenseMapPair &operator=(DenseMapPair &&other) = default;
+
+  KeyT &getFirst() { return first; }
+  const KeyT &getFirst() const { return first; }
+  ValueT &getSecond() { return second; }
+  const ValueT &getSecond() const { return second; }
+};
+
+}  // end namespace detail
+
+template <typename T>
+struct DenseMapInfo {
+  // static inline T getEmptyKey();
+  // static inline T getTombstoneKey();
+  // static unsigned getHashValue(const T &Val);
+  // static bool isEqual(const T &LHS, const T &RHS);
+};
+
+// Provide DenseMapInfo for all pointers. Come up with sentinel pointer values
+// that are aligned to alignof(T) bytes, but try to avoid requiring T to be
+// complete. This allows clients to instantiate DenseMap<T*, ...> with forward
+// declared key types. Assume that no pointer key type requires more than 4096
+// bytes of alignment.
+template <typename T>
+struct DenseMapInfo<T *> {
+  // The following should hold, but it would require T to be complete:
+  // static_assert(alignof(T) <= (1 << Log2MaxAlign),
+  //               "DenseMap does not support pointer keys requiring more than "
+  //               "Log2MaxAlign bits of alignment");
+  static constexpr uptr Log2MaxAlign = 12;
+
+  static inline T *getEmptyKey() {
+    uptr Val = static_cast<uptr>(-1);
+    Val <<= Log2MaxAlign;
+    return reinterpret_cast<T *>(Val);
+  }
+
+  static inline T *getTombstoneKey() {
+    uptr Val = static_cast<uptr>(-2);
+    Val <<= Log2MaxAlign;
+    return reinterpret_cast<T *>(Val);
+  }
+
+  static unsigned getHashValue(const T *PtrVal) {
+    return (unsigned((uptr)PtrVal) >> 4) ^ (unsigned((uptr)PtrVal) >> 9);
+  }
+
+  static bool isEqual(const T *LHS, const T *RHS) { return LHS == RHS; }
+};
+
+// Provide DenseMapInfo for chars.
+template <>
+struct DenseMapInfo<char> {
+  static inline char getEmptyKey() { return ~0; }
+  static inline char getTombstoneKey() { return ~0 - 1; }
+  static unsigned getHashValue(const char &Val) { return Val * 37U; }
+
+  static bool isEqual(const char &LHS, const char &RHS) { return LHS == RHS; }
+};
+
+// Provide DenseMapInfo for unsigned chars.
+template <>
+struct DenseMapInfo<unsigned char> {
+  static inline unsigned char getEmptyKey() { return ~0; }
+  static inline unsigned char getTombstoneKey() { return ~0 - 1; }
+  static unsigned getHashValue(const unsigned char &Val) { return Val * 37U; }
+
+  static bool isEqual(const unsigned char &LHS, const unsigned char &RHS) {
+    return LHS == RHS;
+  }
+};
+
+// Provide DenseMapInfo for unsigned shorts.
+template <>
+struct DenseMapInfo<unsigned short> {
+  static inline unsigned short getEmptyKey() { return 0xFFFF; }
+  static inline unsigned short getTombstoneKey() { return 0xFFFF - 1; }
+  static unsigned getHashValue(const unsigned short &Val) { return Val * 37U; }
+
+  static bool isEqual(const unsigned short &LHS, const unsigned short &RHS) {
+    return LHS == RHS;
+  }
+};
+
+// Provide DenseMapInfo for unsigned ints.
+template <>
+struct DenseMapInfo<unsigned> {
+  static inline unsigned getEmptyKey() { return ~0U; }
+  static inline unsigned getTombstoneKey() { return ~0U - 1; }
+  static unsigned getHashValue(const unsigned &Val) { return Val * 37U; }
+
+  static bool isEqual(const unsigned &LHS, const unsigned &RHS) {
+    return LHS == RHS;
+  }
+};
+
+// Provide DenseMapInfo for unsigned longs.
+template <>
+struct DenseMapInfo<unsigned long> {
+  static inline unsigned long getEmptyKey() { return ~0UL; }
+  static inline unsigned long getTombstoneKey() { return ~0UL - 1L; }
+
+  static unsigned getHashValue(const unsigned long &Val) {
+    return (unsigned)(Val * 37UL);
+  }
+
+  static bool isEqual(const unsigned long &LHS, const unsigned long &RHS) {
+    return LHS == RHS;
+  }
+};
+
+// Provide DenseMapInfo for unsigned long longs.
+template <>
+struct DenseMapInfo<unsigned long long> {
+  static inline unsigned long long getEmptyKey() { return ~0ULL; }
+  static inline unsigned long long getTombstoneKey() { return ~0ULL - 1ULL; }
+
+  static unsigned getHashValue(const unsigned long long &Val) {
+    return (unsigned)(Val * 37ULL);
+  }
+
+  static bool isEqual(const unsigned long long &LHS,
+                      const unsigned long long &RHS) {
+    return LHS == RHS;
+  }
+};
+
+// Provide DenseMapInfo for shorts.
+template <>
+struct DenseMapInfo<short> {
+  static inline short getEmptyKey() { return 0x7FFF; }
+  static inline short getTombstoneKey() { return -0x7FFF - 1; }
+  static unsigned getHashValue(const short &Val) { return Val * 37U; }
+  static bool isEqual(const short &LHS, const short &RHS) { return LHS == RHS; }
+};
+
+// Provide DenseMapInfo for ints.
+template <>
+struct DenseMapInfo<int> {
+  static inline int getEmptyKey() { return 0x7fffffff; }
+  static inline int getTombstoneKey() { return -0x7fffffff - 1; }
+  static unsigned getHashValue(const int &Val) { return (unsigned)(Val * 37U); }
+
+  static bool isEqual(const int &LHS, const int &RHS) { return LHS == RHS; }
+};
+
+// Provide DenseMapInfo for longs.
+template <>
+struct DenseMapInfo<long> {
+  static inline long getEmptyKey() {
+    return (1UL << (sizeof(long) * 8 - 1)) - 1UL;
+  }
+
+  static inline long getTombstoneKey() { return getEmptyKey() - 1L; }
+
+  static unsigned getHashValue(const long &Val) {
+    return (unsigned)(Val * 37UL);
+  }
+
+  static bool isEqual(const long &LHS, const long &RHS) { return LHS == RHS; }
+};
+
+// Provide DenseMapInfo for long longs.
+template <>
+struct DenseMapInfo<long long> {
+  static inline long long getEmptyKey() { return 0x7fffffffffffffffLL; }
+  static inline long long getTombstoneKey() {
+    return -0x7fffffffffffffffLL - 1;
+  }
+
+  static unsigned getHashValue(const long long &Val) {
+    return (unsigned)(Val * 37ULL);
+  }
+
+  static bool isEqual(const long long &LHS, const long long &RHS) {
+    return LHS == RHS;
+  }
+};
+
+// Provide DenseMapInfo for all pairs whose members have info.
+template <typename T, typename U>
+struct DenseMapInfo<detail::DenseMapPair<T, U>> {
+  using Pair = detail::DenseMapPair<T, U>;
+  using FirstInfo = DenseMapInfo<T>;
+  using SecondInfo = DenseMapInfo<U>;
+
+  static inline Pair getEmptyKey() {
+    return detail::DenseMapPair<T, U>(FirstInfo::getEmptyKey(),
+                                      SecondInfo::getEmptyKey());
+  }
+
+  static inline Pair getTombstoneKey() {
+    return detail::DenseMapPair<T, U>(FirstInfo::getTombstoneKey(),
+                                      SecondInfo::getTombstoneKey());
+  }
+
+  static unsigned getHashValue(const Pair &PairVal) {
+    return detail::combineHashValue(FirstInfo::getHashValue(PairVal.first),
+                                    SecondInfo::getHashValue(PairVal.second));
+  }
+
+  static bool isEqual(const Pair &LHS, const Pair &RHS) {
+    return FirstInfo::isEqual(LHS.first, RHS.first) &&
+           SecondInfo::isEqual(LHS.second, RHS.second);
+  }
+};
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_DENSE_MAP_INFO_H
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_file.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_file.cpp
index 0b92dccde4a1..5492560df914 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_file.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_file.cpp
@@ -75,6 +75,20 @@ void ReportFile::ReopenIfNecessary() {
   fd_pid = pid;
 }
 
+static void RecursiveCreateParentDirs(char *path) {
+  if (path[0] == '\0')
+    return;
+  for (int i = 1; path[i] != '\0'; ++i) {
+    char save = path[i];
+    if (!IsPathSeparator(path[i]))
+      continue;
+    path[i] = '\0';
+    /* Some of these will fail, because the directory exists, ignore it. */
+    CreateDir(path);
+    path[i] = save;
+  }
+}
+
 void ReportFile::SetReportPath(const char *path) {
   if (path) {
     uptr len = internal_strlen(path);
@@ -95,6 +109,7 @@ void ReportFile::SetReportPath(const char *path) {
     fd = kStdoutFd;
   } else {
     internal_snprintf(path_prefix, kMaxPathLength, "%s", path);
+    RecursiveCreateParentDirs(path_prefix);
   }
 }
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_file.h b/compiler-rt/lib/sanitizer_common/sanitizer_file.h
index 08671ab67d0f..3d7916171c1e 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_file.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_file.h
@@ -81,6 +81,8 @@ bool FileExists(const char *filename);
 char *FindPathToBinary(const char *name);
 bool IsPathSeparator(const char c);
 bool IsAbsolutePath(const char *path);
+// Returns true on success, false on failure.
+bool CreateDir(const char *pathname);
 // Starts a subprocess and returs its pid.
 // If *_fd parameters are not kInvalidFd their corresponding input/output
 // streams will be redirect to the file. The files will always be closed
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_flag_parser.h b/compiler-rt/lib/sanitizer_common/sanitizer_flag_parser.h
index acc71ccd89ee..3ccc6a6fa537 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_flag_parser.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_flag_parser.h
@@ -138,7 +138,7 @@ inline bool FlagHandler<uptr>::Parse(const char *value) {
 
 template <>
 inline bool FlagHandler<uptr>::Format(char *buffer, uptr size) {
-  uptr num_symbols_should_write = internal_snprintf(buffer, size, "%p", *t_);
+  uptr num_symbols_should_write = internal_snprintf(buffer, size, "0x%zx", *t_);
   return num_symbols_should_write < size;
 }
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_flags.inc b/compiler-rt/lib/sanitizer_common/sanitizer_flags.inc
index 3bc44c6b1eb1..95da82b1a1da 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_flags.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_flags.inc
@@ -160,6 +160,10 @@ COMMON_FLAG(
 COMMON_FLAG(const char *, coverage_dir, ".",
             "Target directory for coverage dumps. Defaults to the current "
             "directory.")
+COMMON_FLAG(const char *, cov_8bit_counters_out, "",
+    "If non-empty, write 8bit counters to this file. ")
+COMMON_FLAG(const char *, cov_pcs_out, "",
+    "If non-empty, write the coverage pc table to this file. ")
 COMMON_FLAG(bool, full_address_space, false,
             "Sanitize complete address space; "
             "by default kernel area on 32-bit platforms will not be sanitized")
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_flat_map.h b/compiler-rt/lib/sanitizer_common/sanitizer_flat_map.h
new file mode 100644
index 000000000000..05fb554d20c1
--- /dev/null
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_flat_map.h
@@ -0,0 +1,173 @@
+//===-- sanitizer_flat_map.h ------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// Part of the Sanitizer Allocator.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef SANITIZER_FLAT_MAP_H
+#define SANITIZER_FLAT_MAP_H
+
+#include "sanitizer_atomic.h"
+#include "sanitizer_common.h"
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_local_address_space_view.h"
+#include "sanitizer_mutex.h"
+
+namespace __sanitizer {
+
+// Call these callbacks on mmap/munmap.
+struct NoOpMapUnmapCallback {
+  void OnMap(uptr p, uptr size) const {}
+  void OnUnmap(uptr p, uptr size) const {}
+};
+
+// Maps integers in rage [0, kSize) to values.
+template <typename T, u64 kSize,
+          typename AddressSpaceViewTy = LocalAddressSpaceView>
+class FlatMap {
+ public:
+  using AddressSpaceView = AddressSpaceViewTy;
+  void Init() { internal_memset(map_, 0, sizeof(map_)); }
+
+  constexpr uptr size() const { return kSize; }
+
+  bool contains(uptr idx) const {
+    CHECK_LT(idx, kSize);
+    return true;
+  }
+
+  T &operator[](uptr idx) {
+    DCHECK_LT(idx, kSize);
+    return map_[idx];
+  }
+
+  const T &operator[](uptr idx) const {
+    DCHECK_LT(idx, kSize);
+    return map_[idx];
+  }
+
+ private:
+  T map_[kSize];
+};
+
+// TwoLevelMap maps integers in range [0, kSize1*kSize2) to values.
+// It is implemented as a two-dimensional array: array of kSize1 pointers
+// to kSize2-byte arrays. The secondary arrays are mmaped on demand.
+// Each value is initially zero and can be set to something else only once.
+// Setting and getting values from multiple threads is safe w/o extra locking.
+template <typename T, u64 kSize1, u64 kSize2,
+          typename AddressSpaceViewTy = LocalAddressSpaceView,
+          class MapUnmapCallback = NoOpMapUnmapCallback>
+class TwoLevelMap {
+  static_assert(IsPowerOfTwo(kSize2), "Use a power of two for performance.");
+
+ public:
+  using AddressSpaceView = AddressSpaceViewTy;
+  void Init() {
+    mu_.Init();
+    internal_memset(map1_, 0, sizeof(map1_));
+  }
+
+  void TestOnlyUnmap() {
+    for (uptr i = 0; i < kSize1; i++) {
+      T *p = Get(i);
+      if (!p)
+        continue;
+      MapUnmapCallback().OnUnmap(reinterpret_cast<uptr>(p), MmapSize());
+      UnmapOrDie(p, kSize2);
+    }
+    Init();
+  }
+
+  uptr MemoryUsage() const {
+    uptr res = 0;
+    for (uptr i = 0; i < kSize1; i++) {
+      T *p = Get(i);
+      if (!p)
+        continue;
+      res += MmapSize();
+    }
+    return res;
+  }
+
+  constexpr uptr size() const { return kSize1 * kSize2; }
+  constexpr uptr size1() const { return kSize1; }
+  constexpr uptr size2() const { return kSize2; }
+
+  bool contains(uptr idx) const {
+    CHECK_LT(idx, kSize1 * kSize2);
+    return Get(idx / kSize2);
+  }
+
+  const T &operator[](uptr idx) const {
+    DCHECK_LT(idx, kSize1 * kSize2);
+    T *map2 = GetOrCreate(idx / kSize2);
+    return *AddressSpaceView::Load(&map2[idx % kSize2]);
+  }
+
+  T &operator[](uptr idx) {
+    DCHECK_LT(idx, kSize1 * kSize2);
+    T *map2 = GetOrCreate(idx / kSize2);
+    return *AddressSpaceView::LoadWritable(&map2[idx % kSize2]);
+  }
+
+ private:
+  constexpr uptr MmapSize() const {
+    return RoundUpTo(kSize2 * sizeof(T), GetPageSizeCached());
+  }
+
+  T *Get(uptr idx) const {
+    DCHECK_LT(idx, kSize1);
+    return reinterpret_cast<T *>(
+        atomic_load(&map1_[idx], memory_order_acquire));
+  }
+
+  T *GetOrCreate(uptr idx) const {
+    DCHECK_LT(idx, kSize1);
+    // This code needs to use memory_order_acquire/consume, but we use
+    // memory_order_relaxed for performance reasons (matters for arm64). We
+    // expect memory_order_relaxed to be effectively equivalent to
+    // memory_order_consume in this case for all relevant architectures: all
+    // dependent data is reachable only by dereferencing the resulting pointer.
+    // If relaxed load fails to see stored ptr, the code will fall back to
+    // Create() and reload the value again with locked mutex as a memory
+    // barrier.
+    T *res = reinterpret_cast<T *>(atomic_load_relaxed(&map1_[idx]));
+    if (LIKELY(res))
+      return res;
+    return Create(idx);
+  }
+
+  NOINLINE T *Create(uptr idx) const {
+    SpinMutexLock l(&mu_);
+    T *res = Get(idx);
+    if (!res) {
+      res = reinterpret_cast<T *>(MmapOrDie(MmapSize(), "TwoLevelMap"));
+      MapUnmapCallback().OnMap(reinterpret_cast<uptr>(res), kSize2);
+      atomic_store(&map1_[idx], reinterpret_cast<uptr>(res),
+                   memory_order_release);
+    }
+    return res;
+  }
+
+  mutable StaticSpinMutex mu_;
+  mutable atomic_uintptr_t map1_[kSize1];
+};
+
+template <u64 kSize, typename AddressSpaceViewTy = LocalAddressSpaceView>
+using FlatByteMap = FlatMap<u8, kSize, AddressSpaceViewTy>;
+
+template <u64 kSize1, u64 kSize2,
+          typename AddressSpaceViewTy = LocalAddressSpaceView,
+          class MapUnmapCallback = NoOpMapUnmapCallback>
+using TwoLevelByteMap =
+    TwoLevelMap<u8, kSize1, kSize2, AddressSpaceViewTy, MapUnmapCallback>;
+}  // namespace __sanitizer
+
+#endif
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.cpp
index 65bc398656c9..c7b30d988365 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.cpp
@@ -112,47 +112,6 @@ void FutexWake(atomic_uint32_t *p, u32 count) {
   CHECK_EQ(status, ZX_OK);
 }
 
-enum MutexState : int { MtxUnlocked = 0, MtxLocked = 1, MtxSleeping = 2 };
-
-BlockingMutex::BlockingMutex() {
-  // NOTE!  It's important that this use internal_memset, because plain
-  // memset might be intercepted (e.g., actually be __asan_memset).
-  // Defining this so the compiler initializes each field, e.g.:
-  //   BlockingMutex::BlockingMutex() : BlockingMutex(LINKER_INITIALIZED) {}
-  // might result in the compiler generating a call to memset, which would
-  // have the same problem.
-  internal_memset(this, 0, sizeof(*this));
-}
-
-void BlockingMutex::Lock() {
-  CHECK_EQ(owner_, 0);
-  atomic_uint32_t *m = reinterpret_cast<atomic_uint32_t *>(&opaque_storage_);
-  if (atomic_exchange(m, MtxLocked, memory_order_acquire) == MtxUnlocked)
-    return;
-  while (atomic_exchange(m, MtxSleeping, memory_order_acquire) != MtxUnlocked) {
-    zx_status_t status =
-        _zx_futex_wait(reinterpret_cast<zx_futex_t *>(m), MtxSleeping,
-                       ZX_HANDLE_INVALID, ZX_TIME_INFINITE);
-    if (status != ZX_ERR_BAD_STATE)  // Normal race.
-      CHECK_EQ(status, ZX_OK);
-  }
-}
-
-void BlockingMutex::Unlock() {
-  atomic_uint32_t *m = reinterpret_cast<atomic_uint32_t *>(&opaque_storage_);
-  u32 v = atomic_exchange(m, MtxUnlocked, memory_order_release);
-  CHECK_NE(v, MtxUnlocked);
-  if (v == MtxSleeping) {
-    zx_status_t status = _zx_futex_wake(reinterpret_cast<zx_futex_t *>(m), 1);
-    CHECK_EQ(status, ZX_OK);
-  }
-}
-
-void BlockingMutex::CheckLocked() const {
-  auto m = reinterpret_cast<atomic_uint32_t const *>(&opaque_storage_);
-  CHECK_NE(MtxUnlocked, atomic_load(m, memory_order_relaxed));
-}
-
 uptr GetPageSize() { return _zx_system_get_page_size(); }
 
 uptr GetMmapGranularity() { return _zx_system_get_page_size(); }
@@ -413,7 +372,7 @@ bool IsAccessibleMemoryRange(uptr beg, uptr size) {
 }
 
 // FIXME implement on this platform.
-void GetMemoryProfile(fill_profile_f cb, uptr *stats, uptr stats_size) {}
+void GetMemoryProfile(fill_profile_f cb, uptr *stats) {}
 
 bool ReadFileToBuffer(const char *file_name, char **buff, uptr *buff_size,
                       uptr *read_len, uptr max_len, error_t *errno_p) {
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_hash.h b/compiler-rt/lib/sanitizer_common/sanitizer_hash.h
index 3d97dcc5d280..f7cf9f234e6f 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_hash.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_hash.h
@@ -38,6 +38,30 @@ class MurMur2HashBuilder {
     return x;
   }
 };
+
+class MurMur2Hash64Builder {
+  static const u64 m = 0xc6a4a7935bd1e995ull;
+  static const u64 seed = 0x9747b28c9747b28cull;
+  static const u64 r = 47;
+  u64 h;
+
+ public:
+  explicit MurMur2Hash64Builder(u64 init = 0) { h = seed ^ (init * m); }
+  void add(u64 k) {
+    k *= m;
+    k ^= k >> r;
+    k *= m;
+    h ^= k;
+    h *= m;
+  }
+  u64 get() {
+    u64 x = h;
+    x ^= x >> r;
+    x *= m;
+    x ^= x >> r;
+    return x;
+  }
+};
 }  //namespace __sanitizer
 
 #endif  // SANITIZER_HASH_H
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_interceptors_ioctl_netbsd.inc b/compiler-rt/lib/sanitizer_common/sanitizer_interceptors_ioctl_netbsd.inc
index 576807ea3a6a..9683b97ab91d 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_interceptors_ioctl_netbsd.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_interceptors_ioctl_netbsd.inc
@@ -1406,7 +1406,7 @@ static void ioctl_table_fill() {
   _(URIO_SEND_COMMAND, READWRITE, struct_urio_command_sz);
   _(URIO_RECV_COMMAND, READWRITE, struct_urio_command_sz);
 #undef _
-} // NOLINT
+}
 
 static bool ioctl_initialized = false;
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_interface_internal.h b/compiler-rt/lib/sanitizer_common/sanitizer_interface_internal.h
index 0b001c1c4830..1600d31c30c0 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_interface_internal.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_interface_internal.h
@@ -111,12 +111,13 @@ extern "C" {
   SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
   void __sanitizer_cov_trace_pc_guard_init(__sanitizer::u32*,
                                            __sanitizer::u32*);
-  SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
-  void __sanitizer_cov_8bit_counters_init();
+  SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE void
+  __sanitizer_cov_8bit_counters_init(char *, char *);
   SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE void
   __sanitizer_cov_bool_flag_init();
   SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE void
-  __sanitizer_cov_pcs_init();
+  __sanitizer_cov_pcs_init(const __sanitizer::uptr *,
+                           const __sanitizer::uptr *);
 } // extern "C"
 
 #endif  // SANITIZER_INTERFACE_INTERNAL_H
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_internal_defs.h b/compiler-rt/lib/sanitizer_common/sanitizer_internal_defs.h
index 84053fec2649..d0db0129d4af 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_internal_defs.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_internal_defs.h
@@ -125,6 +125,10 @@
 # define __has_attribute(x) 0
 #endif
 
+#if !defined(__has_cpp_attribute)
+#  define __has_cpp_attribute(x) 0
+#endif
+
 // For portability reasons we do not include stddef.h, stdint.h or any other
 // system header, but we do need some basic types that are not defined
 // in a portable way by the language itself.
@@ -135,8 +139,13 @@ namespace __sanitizer {
 typedef unsigned long long uptr;
 typedef signed long long sptr;
 #else
+#  if (SANITIZER_WORDSIZE == 64) || SANITIZER_MAC || SANITIZER_WINDOWS
 typedef unsigned long uptr;
 typedef signed long sptr;
+#  else
+typedef unsigned int uptr;
+typedef signed int sptr;
+#  endif
 #endif  // defined(_WIN64)
 #if defined(__x86_64__)
 // Since x32 uses ILP32 data model in 64-bit hardware mode, we must use
@@ -168,10 +177,9 @@ typedef long pid_t;
 typedef int pid_t;
 #endif
 
-#if SANITIZER_FREEBSD || SANITIZER_NETBSD || \
-    SANITIZER_MAC || \
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD || SANITIZER_MAC ||             \
     (SANITIZER_SOLARIS && (defined(_LP64) || _FILE_OFFSET_BITS == 64)) || \
-    (SANITIZER_LINUX && defined(__x86_64__))
+    (SANITIZER_LINUX && (defined(__x86_64__) || defined(__hexagon__)))
 typedef u64 OFF_T;
 #else
 typedef uptr OFF_T;
@@ -250,6 +258,12 @@ typedef u64 tid_t;
 # define NOEXCEPT throw()
 #endif
 
+#if __has_cpp_attribute(clang::fallthrough)
+#  define FALLTHROUGH [[clang::fallthrough]]
+#else
+#  define FALLTHROUGH
+#endif
+
 // Unaligned versions of basic types.
 typedef ALIGNED(1) u16 uu16;
 typedef ALIGNED(1) u32 uu32;
@@ -277,14 +291,17 @@ void NORETURN CheckFailed(const char *file, int line, const char *cond,
                           u64 v1, u64 v2);
 
 // Check macro
-#define RAW_CHECK_MSG(expr, msg) do { \
-  if (UNLIKELY(!(expr))) { \
-    RawWrite(msg); \
-    Die(); \
-  } \
-} while (0)
+#define RAW_CHECK_MSG(expr, msg, ...)          \
+  do {                                         \
+    if (UNLIKELY(!(expr))) {                   \
+      const char* msgs[] = {msg, __VA_ARGS__}; \
+      for (const char* m : msgs) RawWrite(m);  \
+      Die();                                   \
+    }                                          \
+  } while (0)
 
-#define RAW_CHECK(expr) RAW_CHECK_MSG(expr, #expr)
+#define RAW_CHECK(expr) RAW_CHECK_MSG(expr, #expr "\n", )
+#define RAW_CHECK_VA(expr, ...) RAW_CHECK_MSG(expr, #expr "\n", __VA_ARGS__)
 
 #define CHECK_IMPL(c1, op, c2) \
   do { \
@@ -409,8 +426,14 @@ inline void Trap() {
     (void)enable_fp;                      \
   } while (0)
 
-constexpr u32 kInvalidTid = -1;
-constexpr u32 kMainTid = 0;
+// Internal thread identifier allocated by ThreadRegistry.
+typedef u32 Tid;
+constexpr Tid kInvalidTid = -1;
+constexpr Tid kMainTid = 0;
+
+// Stack depot stack identifier.
+typedef u32 StackID;
+const StackID kInvalidStackID = 0;
 
 }  // namespace __sanitizer
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_libc.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_libc.cpp
index 4bc04b486870..d3076f0da489 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_libc.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_libc.cpp
@@ -258,6 +258,18 @@ s64 internal_simple_strtoll(const char *nptr, const char **endptr, int base) {
   }
 }
 
+uptr internal_wcslen(const wchar_t *s) {
+  uptr i = 0;
+  while (s[i]) i++;
+  return i;
+}
+
+uptr internal_wcsnlen(const wchar_t *s, uptr maxlen) {
+  uptr i = 0;
+  while (i < maxlen && s[i]) i++;
+  return i;
+}
+
 bool mem_is_zero(const char *beg, uptr size) {
   CHECK_LE(size, 1ULL << FIRST_32_SECOND_64(30, 40));  // Sanity check.
   const char *end = beg + size;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_libc.h b/compiler-rt/lib/sanitizer_common/sanitizer_libc.h
index bcb81ebbc803..39a212665d0a 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_libc.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_libc.h
@@ -49,7 +49,10 @@ char *internal_strrchr(const char *s, int c);
 char *internal_strstr(const char *haystack, const char *needle);
 // Works only for base=10 and doesn't set errno.
 s64 internal_simple_strtoll(const char *nptr, const char **endptr, int base);
-int internal_snprintf(char *buffer, uptr length, const char *format, ...);
+int internal_snprintf(char *buffer, uptr length, const char *format, ...)
+    FORMAT(3, 4);
+uptr internal_wcslen(const wchar_t *s);
+uptr internal_wcsnlen(const wchar_t *s, uptr maxlen);
 
 // Return true if all bytes in [mem, mem+size) are zero.
 // Optimized for the case when the result is true.
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_libignore.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_libignore.cpp
index a65d3d896e33..caaba3155a7b 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_libignore.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_libignore.cpp
@@ -22,9 +22,9 @@ LibIgnore::LibIgnore(LinkerInitialized) {
 }
 
 void LibIgnore::AddIgnoredLibrary(const char *name_templ) {
-  BlockingMutexLock lock(&mutex_);
+  Lock lock(&mutex_);
   if (count_ >= kMaxLibs) {
-    Report("%s: too many ignored libraries (max: %d)\n", SanitizerToolName,
+    Report("%s: too many ignored libraries (max: %zu)\n", SanitizerToolName,
            kMaxLibs);
     Die();
   }
@@ -36,7 +36,7 @@ void LibIgnore::AddIgnoredLibrary(const char *name_templ) {
 }
 
 void LibIgnore::OnLibraryLoaded(const char *name) {
-  BlockingMutexLock lock(&mutex_);
+  Lock lock(&mutex_);
   // Try to match suppressions with symlink target.
   InternalMmapVector<char> buf(kMaxPathLength);
   if (name && internal_readlink(name, buf.data(), buf.size() - 1) > 0 &&
@@ -105,7 +105,7 @@ void LibIgnore::OnLibraryLoaded(const char *name) {
           continue;
         if (IsPcInstrumented(range.beg) && IsPcInstrumented(range.end - 1))
           continue;
-        VReport(1, "Adding instrumented range %p-%p from library '%s'\n",
+        VReport(1, "Adding instrumented range 0x%zx-0x%zx from library '%s'\n",
                 range.beg, range.end, mod.full_name());
         const uptr idx =
             atomic_load(&instrumented_ranges_count_, memory_order_relaxed);
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_libignore.h b/compiler-rt/lib/sanitizer_common/sanitizer_libignore.h
index 256f685979f4..18e4d83ed77f 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_libignore.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_libignore.h
@@ -77,7 +77,7 @@ class LibIgnore {
   LibCodeRange instrumented_code_ranges_[kMaxInstrumentedRanges];
 
   // Cold part:
-  BlockingMutex mutex_;
+  Mutex mutex_;
   uptr count_;
   Lib libs_[kMaxLibs];
   bool track_instrumented_libs_;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_linux.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_linux.cpp
index 9b7d87eb85e1..596037d77222 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_linux.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_linux.cpp
@@ -150,17 +150,39 @@ const int FUTEX_WAKE_PRIVATE = FUTEX_WAKE | FUTEX_PRIVATE_FLAG;
 
 namespace __sanitizer {
 
-#if SANITIZER_LINUX && defined(__x86_64__)
-#include "sanitizer_syscall_linux_x86_64.inc"
-#elif SANITIZER_LINUX && SANITIZER_RISCV64
-#include "sanitizer_syscall_linux_riscv64.inc"
-#elif SANITIZER_LINUX && defined(__aarch64__)
-#include "sanitizer_syscall_linux_aarch64.inc"
-#elif SANITIZER_LINUX && defined(__arm__)
-#include "sanitizer_syscall_linux_arm.inc"
-#else
-#include "sanitizer_syscall_generic.inc"
-#endif
+void SetSigProcMask(__sanitizer_sigset_t *set, __sanitizer_sigset_t *old) {
+  CHECK_EQ(0, internal_sigprocmask(SIG_SETMASK, set, old));
+}
+
+ScopedBlockSignals::ScopedBlockSignals(__sanitizer_sigset_t *copy) {
+  __sanitizer_sigset_t set;
+  internal_sigfillset(&set);
+#  if SANITIZER_LINUX && !SANITIZER_ANDROID
+  // Glibc uses SIGSETXID signal during setuid call. If this signal is blocked
+  // on any thread, setuid call hangs.
+  // See test/sanitizer_common/TestCases/Linux/setuid.c.
+  internal_sigdelset(&set, 33);
+#  endif
+  SetSigProcMask(&set, &saved_);
+  if (copy)
+    internal_memcpy(copy, &saved_, sizeof(saved_));
+}
+
+ScopedBlockSignals::~ScopedBlockSignals() { SetSigProcMask(&saved_, nullptr); }
+
+#  if SANITIZER_LINUX && defined(__x86_64__)
+#    include "sanitizer_syscall_linux_x86_64.inc"
+#  elif SANITIZER_LINUX && SANITIZER_RISCV64
+#    include "sanitizer_syscall_linux_riscv64.inc"
+#  elif SANITIZER_LINUX && defined(__aarch64__)
+#    include "sanitizer_syscall_linux_aarch64.inc"
+#  elif SANITIZER_LINUX && defined(__arm__)
+#    include "sanitizer_syscall_linux_arm.inc"
+#  elif SANITIZER_LINUX && defined(__hexagon__)
+#    include "sanitizer_syscall_linux_hexagon.inc"
+#  else
+#    include "sanitizer_syscall_generic.inc"
+#  endif
 
 // --------------- sanitizer_libc.h
 #if !SANITIZER_SOLARIS && !SANITIZER_NETBSD
@@ -415,7 +437,7 @@ uptr internal_unlink(const char *path) {
 }
 
 uptr internal_rename(const char *oldpath, const char *newpath) {
-#if defined(__riscv)
+#if defined(__riscv) && defined(__linux__)
   return internal_syscall(SYSCALL(renameat2), AT_FDCWD, (uptr)oldpath, AT_FDCWD,
                           (uptr)newpath, 0);
 #elif SANITIZER_USES_CANONICAL_LINUX_SYSCALLS
@@ -659,48 +681,6 @@ void FutexWake(atomic_uint32_t *p, u32 count) {
 #    endif
 }
 
-enum { MtxUnlocked = 0, MtxLocked = 1, MtxSleeping = 2 };
-
-BlockingMutex::BlockingMutex() {
-  internal_memset(this, 0, sizeof(*this));
-}
-
-void BlockingMutex::Lock() {
-  CHECK_EQ(owner_, 0);
-  atomic_uint32_t *m = reinterpret_cast<atomic_uint32_t *>(&opaque_storage_);
-  if (atomic_exchange(m, MtxLocked, memory_order_acquire) == MtxUnlocked)
-    return;
-  while (atomic_exchange(m, MtxSleeping, memory_order_acquire) != MtxUnlocked) {
-#if SANITIZER_FREEBSD
-    _umtx_op(m, UMTX_OP_WAIT_UINT, MtxSleeping, 0, 0);
-#elif SANITIZER_NETBSD
-    sched_yield(); /* No userspace futex-like synchronization */
-#else
-    internal_syscall(SYSCALL(futex), (uptr)m, FUTEX_WAIT_PRIVATE, MtxSleeping,
-                     0, 0, 0);
-#endif
-  }
-}
-
-void BlockingMutex::Unlock() {
-  atomic_uint32_t *m = reinterpret_cast<atomic_uint32_t *>(&opaque_storage_);
-  u32 v = atomic_exchange(m, MtxUnlocked, memory_order_release);
-  CHECK_NE(v, MtxUnlocked);
-  if (v == MtxSleeping) {
-#if SANITIZER_FREEBSD
-    _umtx_op(m, UMTX_OP_WAKE, 1, 0, 0);
-#elif SANITIZER_NETBSD
-                   /* No userspace futex-like synchronization */
-#else
-    internal_syscall(SYSCALL(futex), (uptr)m, FUTEX_WAKE_PRIVATE, 1, 0, 0, 0);
-#endif
-  }
-}
-
-void BlockingMutex::CheckLocked() const {
-  auto m = reinterpret_cast<atomic_uint32_t const *>(&opaque_storage_);
-  CHECK_NE(MtxUnlocked, atomic_load(m, memory_order_relaxed));
-}
 #  endif  // !SANITIZER_SOLARIS
 
 // ----------------- sanitizer_linux.h
@@ -1217,7 +1197,8 @@ void ForEachMappedRegion(link_map *map, void (*cb)(const void *, uptr)) {
 }
 #endif
 
-#if defined(__x86_64__) && SANITIZER_LINUX
+#if SANITIZER_LINUX
+#if defined(__x86_64__)
 // We cannot use glibc's clone wrapper, because it messes with the child
 // task's TLS. It writes the PID and TID of the child task to its thread
 // descriptor, but in our case the child task shares the thread descriptor with
@@ -1399,7 +1380,7 @@ uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
 #elif defined(__aarch64__)
 uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
                     int *parent_tidptr, void *newtls, int *child_tidptr) {
-  long long res;
+  register long long res __asm__("x0");
   if (!fn || !child_stack)
     return -EINVAL;
   CHECK_EQ(0, (uptr)child_stack % 16);
@@ -1556,7 +1537,7 @@ uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
              : "cr0", "cr1", "memory", "ctr", "r0", "r27", "r28", "r29");
   return res;
 }
-#elif defined(__i386__) && SANITIZER_LINUX
+#elif defined(__i386__)
 uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
                     int *parent_tidptr, void *newtls, int *child_tidptr) {
   int res;
@@ -1621,7 +1602,7 @@ uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
                        : "memory");
   return res;
 }
-#elif defined(__arm__) && SANITIZER_LINUX
+#elif defined(__arm__)
 uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
                     int *parent_tidptr, void *newtls, int *child_tidptr) {
   unsigned int res;
@@ -1687,7 +1668,8 @@ uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
                        : "memory");
   return res;
 }
-#endif  // defined(__x86_64__) && SANITIZER_LINUX
+#endif
+#endif  // SANITIZER_LINUX
 
 #if SANITIZER_LINUX
 int internal_uname(struct utsname *buf) {
@@ -1779,17 +1761,9 @@ HandleSignalMode GetHandleSignalMode(int signum) {
 #if !SANITIZER_GO
 void *internal_start_thread(void *(*func)(void *arg), void *arg) {
   // Start the thread with signals blocked, otherwise it can steal user signals.
-  __sanitizer_sigset_t set, old;
-  internal_sigfillset(&set);
-#if SANITIZER_LINUX && !SANITIZER_ANDROID
-  // Glibc uses SIGSETXID signal during setuid call. If this signal is blocked
-  // on any thread, setuid call hangs (see test/tsan/setuid.c).
-  internal_sigdelset(&set, 33);
-#endif
-  internal_sigprocmask(SIG_SETMASK, &set, &old);
+  ScopedBlockSignals block(nullptr);
   void *th;
   real_pthread_create(&th, nullptr, func, arg);
-  internal_sigprocmask(SIG_SETMASK, &old, nullptr);
   return th;
 }
 
@@ -1811,7 +1785,7 @@ struct __sanitizer_esr_context {
 
 static bool Aarch64GetESR(ucontext_t *ucontext, u64 *esr) {
   static const u32 kEsrMagic = 0x45535201;
-  u8 *aux = ucontext->uc_mcontext.__reserved;
+  u8 *aux = reinterpret_cast<u8 *>(ucontext->uc_mcontext.__reserved);
   while (true) {
     _aarch64_ctx *ctx = (_aarch64_ctx *)aux;
     if (ctx->size == 0) break;
@@ -1917,7 +1891,11 @@ SignalContext::WriteFlag SignalContext::GetWriteFlag() const {
   u32 instr = *(u32 *)pc;
   return (instr >> 21) & 1 ? WRITE: READ;
 #elif defined(__riscv)
+#if SANITIZER_FREEBSD
+  unsigned long pc = ucontext->uc_mcontext.mc_gpregs.gp_sepc;
+#else
   unsigned long pc = ucontext->uc_mcontext.__gregs[REG_PC];
+#endif
   unsigned faulty_instruction = *(uint16_t *)pc;
 
 #if defined(__riscv_compressed)
@@ -2136,12 +2114,23 @@ static void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
   *sp = ucontext->uc_mcontext.gregs[15];
 #elif defined(__riscv)
   ucontext_t *ucontext = (ucontext_t*)context;
+#    if SANITIZER_FREEBSD
+  *pc = ucontext->uc_mcontext.mc_gpregs.gp_sepc;
+  *bp = ucontext->uc_mcontext.mc_gpregs.gp_s[0];
+  *sp = ucontext->uc_mcontext.mc_gpregs.gp_sp;
+#    else
   *pc = ucontext->uc_mcontext.__gregs[REG_PC];
   *bp = ucontext->uc_mcontext.__gregs[REG_S0];
   *sp = ucontext->uc_mcontext.__gregs[REG_SP];
-#else
-# error "Unsupported arch"
-#endif
+#    endif
+#  elif defined(__hexagon__)
+  ucontext_t *ucontext = (ucontext_t *)context;
+  *pc = ucontext->uc_mcontext.pc;
+  *bp = ucontext->uc_mcontext.r30;
+  *sp = ucontext->uc_mcontext.r29;
+#  else
+#    error "Unsupported arch"
+#  endif
 }
 
 void SignalContext::InitPcSpBp() { GetPcSpBp(context, &pc, &sp, &bp); }
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_linux.h b/compiler-rt/lib/sanitizer_common/sanitizer_linux.h
index 9a23fcfb3b93..6a235db0ee2e 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_linux.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_linux.h
@@ -49,7 +49,17 @@ uptr internal_getdents(fd_t fd, struct linux_dirent *dirp, unsigned int count);
 uptr internal_sigaltstack(const void* ss, void* oss);
 uptr internal_sigprocmask(int how, __sanitizer_sigset_t *set,
     __sanitizer_sigset_t *oldset);
-#if SANITIZER_GLIBC
+
+void SetSigProcMask(__sanitizer_sigset_t *set, __sanitizer_sigset_t *oldset);
+struct ScopedBlockSignals {
+  explicit ScopedBlockSignals(__sanitizer_sigset_t *copy);
+  ~ScopedBlockSignals();
+
+ private:
+  __sanitizer_sigset_t saved_;
+};
+
+#  if SANITIZER_GLIBC
 uptr internal_clock_gettime(__sanitizer_clockid_t clk_id, void *tp);
 #endif
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_linux_s390.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_linux_s390.cpp
index bb2f5b5f9f7d..74db831b0aad 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_linux_s390.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_linux_s390.cpp
@@ -57,8 +57,10 @@ uptr internal_mmap(void *addr, uptr length, int prot, int flags, int fd,
 
 uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
                     int *parent_tidptr, void *newtls, int *child_tidptr) {
-  if (!fn || !child_stack)
-    return -EINVAL;
+  if (!fn || !child_stack) {
+    errno = EINVAL;
+    return -1;
+  }
   CHECK_EQ(0, (uptr)child_stack % 16);
   // Minimum frame size.
 #ifdef __s390x__
@@ -71,9 +73,9 @@ uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
   // And pass parameters.
   ((unsigned long *)child_stack)[1] = (uptr)fn;
   ((unsigned long *)child_stack)[2] = (uptr)arg;
-  register long res __asm__("r2");
+  register uptr res __asm__("r2");
   register void *__cstack      __asm__("r2") = child_stack;
-  register int __flags         __asm__("r3") = flags;
+  register long __flags        __asm__("r3") = flags;
   register int * __ptidptr     __asm__("r4") = parent_tidptr;
   register int * __ctidptr     __asm__("r5") = child_tidptr;
   register void * __newtls     __asm__("r6") = newtls;
@@ -113,6 +115,10 @@ uptr internal_clone(int (*fn)(void *), void *child_stack, int flags, void *arg,
                          "r"(__ctidptr),
                          "r"(__newtls)
                        : "memory", "cc");
+  if (res >= (uptr)-4095) {
+    errno = -res;
+    return -1;
+  }
   return res;
 }
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_local_address_space_view.h b/compiler-rt/lib/sanitizer_common/sanitizer_local_address_space_view.h
index 0e19c4d4a801..a47cfc945cd8 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_local_address_space_view.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_local_address_space_view.h
@@ -17,7 +17,7 @@
 // instantiated with the `LocalAddressSpaceView` type. This type is used to
 // load any pointers in instance methods. This implementation is effectively
 // a no-op. When an object is to be used in an out-of-process manner it is
-// instansiated with the `RemoteAddressSpaceView` type.
+// instantiated with the `RemoteAddressSpaceView` type.
 //
 // By making `AddressSpaceView` a template parameter of an object, it can
 // change its implementation at compile time which has no run time overhead.
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_mac.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_mac.cpp
index 083595d1505f..b67203d4c10e 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_mac.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_mac.cpp
@@ -516,25 +516,6 @@ void FutexWait(atomic_uint32_t *p, u32 cmp) {
 
 void FutexWake(atomic_uint32_t *p, u32 count) {}
 
-BlockingMutex::BlockingMutex() {
-  internal_memset(this, 0, sizeof(*this));
-}
-
-void BlockingMutex::Lock() {
-  CHECK(sizeof(OSSpinLock) <= sizeof(opaque_storage_));
-  CHECK_EQ(OS_SPINLOCK_INIT, 0);
-  CHECK_EQ(owner_, 0);
-  OSSpinLockLock((OSSpinLock*)&opaque_storage_);
-}
-
-void BlockingMutex::Unlock() {
-  OSSpinLockUnlock((OSSpinLock*)&opaque_storage_);
-}
-
-void BlockingMutex::CheckLocked() const {
-  CHECK_NE(*(const OSSpinLock*)&opaque_storage_, 0);
-}
-
 u64 NanoTime() {
   timeval tv;
   internal_memset(&tv, 0, sizeof(tv));
@@ -562,6 +543,9 @@ uptr TlsBaseAddr() {
   asm("movq %%gs:0,%0" : "=r"(segbase));
 #elif defined(__i386__)
   asm("movl %%gs:0,%0" : "=r"(segbase));
+#elif defined(__aarch64__)
+  asm("mrs %x0, tpidrro_el0" : "=r"(segbase));
+  segbase &= 0x07ul;  // clearing lower bits, cpu id stored there
 #endif
   return segbase;
 }
@@ -784,8 +768,8 @@ void *internal_start_thread(void *(*func)(void *arg), void *arg) {
 void internal_join_thread(void *th) { pthread_join((pthread_t)th, 0); }
 
 #if !SANITIZER_GO
-static BlockingMutex syslog_lock(LINKER_INITIALIZED);
-#endif
+static Mutex syslog_lock;
+#  endif
 
 void WriteOneLineToSyslog(const char *s) {
 #if !SANITIZER_GO
@@ -800,7 +784,7 @@ void WriteOneLineToSyslog(const char *s) {
 
 // buffer to store crash report application information
 static char crashreporter_info_buff[__sanitizer::kErrorMessageBufferSize] = {};
-static BlockingMutex crashreporter_info_mutex(LINKER_INITIALIZED);
+static Mutex crashreporter_info_mutex;
 
 extern "C" {
 // Integrate with crash reporter libraries.
@@ -830,7 +814,7 @@ asm(".desc ___crashreporter_info__, 0x10");
 }  // extern "C"
 
 static void CRAppendCrashLogMessage(const char *msg) {
-  BlockingMutexLock l(&crashreporter_info_mutex);
+  Lock l(&crashreporter_info_mutex);
   internal_strlcat(crashreporter_info_buff, msg,
                    sizeof(crashreporter_info_buff));
 #if HAVE_CRASHREPORTERCLIENT_H
@@ -874,7 +858,7 @@ void LogFullErrorReport(const char *buffer) {
   // the reporting thread holds the thread registry mutex, and asl_log waits
   // for GCD to dispatch a new thread, the process will deadlock, because the
   // pthread_create wrapper needs to acquire the lock as well.
-  BlockingMutexLock l(&syslog_lock);
+  Lock l(&syslog_lock);
   if (common_flags()->log_to_syslog)
     WriteToSyslog(buffer);
 
@@ -1330,7 +1314,7 @@ uptr FindAvailableMemoryRange(uptr size, uptr alignment, uptr left_padding,
 }
 
 // FIXME implement on this platform.
-void GetMemoryProfile(fill_profile_f cb, uptr *stats, uptr stats_size) { }
+void GetMemoryProfile(fill_profile_f cb, uptr *stats) {}
 
 void SignalContext::DumpAllRegisters(void *context) {
   Report("Register values:\n");
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_malloc_mac.inc b/compiler-rt/lib/sanitizer_common/sanitizer_malloc_mac.inc
index e3b664f68b61..764e2cef5e74 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_malloc_mac.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_malloc_mac.inc
@@ -23,6 +23,7 @@
 #include <sys/mman.h>
 
 #include "interception/interception.h"
+#include "sanitizer_common/sanitizer_allocator_dlsym.h"
 #include "sanitizer_common/sanitizer_mac.h"
 
 // Similar code is used in Google Perftools,
@@ -192,20 +193,15 @@ void *__sanitizer_mz_malloc(malloc_zone_t *zone, uptr size) {
   return p;
 }
 
+struct DlsymAlloc : public DlSymAllocator<DlsymAlloc> {
+  static bool UseImpl() { return !COMMON_MALLOC_SANITIZER_INITIALIZED; }
+};
+
 extern "C"
 SANITIZER_INTERFACE_ATTRIBUTE
 void *__sanitizer_mz_calloc(malloc_zone_t *zone, size_t nmemb, size_t size) {
-  if (UNLIKELY(!COMMON_MALLOC_SANITIZER_INITIALIZED)) {
-    // Hack: dlsym calls calloc before REAL(calloc) is retrieved from dlsym.
-    const size_t kCallocPoolSize = 1024;
-    static uptr calloc_memory_for_dlsym[kCallocPoolSize];
-    static size_t allocated;
-    size_t size_in_words = ((nmemb * size) + kWordSize - 1) / kWordSize;
-    void *mem = (void*)&calloc_memory_for_dlsym[allocated];
-    allocated += size_in_words;
-    CHECK(allocated < kCallocPoolSize);
-    return mem;
-  }
+  if (DlsymAlloc::Use())
+    return DlsymAlloc::Callocate(nmemb, size);
   COMMON_MALLOC_CALLOC(nmemb, size);
   return p;
 }
@@ -223,6 +219,8 @@ extern "C"
 SANITIZER_INTERFACE_ATTRIBUTE
 void __sanitizer_mz_free(malloc_zone_t *zone, void *ptr) {
   if (!ptr) return;
+  if (DlsymAlloc::PointerIsMine(ptr))
+    return DlsymAlloc::Free(ptr);
   COMMON_MALLOC_FREE(ptr);
 }
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_mutex.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_mutex.cpp
index 46f1d0279ca1..40fe56661250 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_mutex.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_mutex.cpp
@@ -73,7 +73,7 @@ void DebugMutexInit() {
   // Build adjacency matrix.
   bool leaf[kMutexTypeMax];
   internal_memset(&leaf, 0, sizeof(leaf));
-  int cnt[kMutexTypeMax] = {};
+  int cnt[kMutexTypeMax];
   internal_memset(&cnt, 0, sizeof(cnt));
   for (int t = 0; t < kMutexTypeMax; t++) {
     mutex_type_count = t;
@@ -174,7 +174,7 @@ struct InternalDeadlockDetector {
     if (max_idx != MutexInvalid && !mutex_can_lock[max_idx][type]) {
       Printf("%s: internal deadlock: can't lock %s under %s mutex\n", SanitizerToolName,
              mutex_meta[type].name, mutex_meta[max_idx].name);
-      PrintMutexPC(pc);
+      PrintMutexPC(locked[max_idx].pc);
       CHECK(0);
     }
     locked[type].seq = ++sequence;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_mutex.h b/compiler-rt/lib/sanitizer_common/sanitizer_mutex.h
index cbd1c25eb69f..5ec6efaa6490 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_mutex.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_mutex.h
@@ -95,7 +95,11 @@ enum {
 
 // Go linker does not support THREADLOCAL variables,
 // so we can't use per-thread state.
-#define SANITIZER_CHECK_DEADLOCKS (SANITIZER_DEBUG && !SANITIZER_GO)
+// Disable checked locks on Darwin. Although Darwin platforms support
+// THREADLOCAL variables they are not usable early on during process init when
+// `__sanitizer::Mutex` is used.
+#define SANITIZER_CHECK_DEADLOCKS \
+  (SANITIZER_DEBUG && !SANITIZER_GO && SANITIZER_SUPPORTS_THREADLOCAL && !SANITIZER_MAC)
 
 #if SANITIZER_CHECK_DEADLOCKS
 struct MutexMeta {
@@ -111,7 +115,7 @@ struct MutexMeta {
 
 class CheckedMutex {
  public:
-  constexpr CheckedMutex(MutexType type)
+  explicit constexpr CheckedMutex(MutexType type)
 #if SANITIZER_CHECK_DEADLOCKS
       : type_(type)
 #endif
@@ -154,13 +158,13 @@ class CheckedMutex {
 // but this attribute is not supported by some older compilers.
 class MUTEX Mutex : CheckedMutex {
  public:
-  constexpr Mutex(MutexType type = MutexUnchecked) : CheckedMutex(type) {}
+  explicit constexpr Mutex(MutexType type = MutexUnchecked)
+      : CheckedMutex(type) {}
 
   void Lock() ACQUIRE() {
     CheckedMutex::Lock();
     u64 reset_mask = ~0ull;
     u64 state = atomic_load_relaxed(&state_);
-    const uptr kMaxSpinIters = 1500;
     for (uptr spin_iters = 0;; spin_iters++) {
       u64 new_state;
       bool locked = (state & (kWriterLock | kReaderLockMask)) != 0;
@@ -189,8 +193,6 @@ class MUTEX Mutex : CheckedMutex {
         // We've incremented waiting writers, so now block.
         writers_.Wait();
         spin_iters = 0;
-        state = atomic_load(&state_, memory_order_relaxed);
-        DCHECK_NE(state & kWriterSpinWait, 0);
       } else {
         // We've set kWriterSpinWait, but we are still in active spinning.
       }
@@ -199,6 +201,8 @@ class MUTEX Mutex : CheckedMutex {
       // Either way we need to reset kWriterSpinWait
       // next time we take the lock or block again.
       reset_mask = ~kWriterSpinWait;
+      state = atomic_load(&state_, memory_order_relaxed);
+      DCHECK_NE(state & kWriterSpinWait, 0);
     }
   }
 
@@ -212,17 +216,16 @@ class MUTEX Mutex : CheckedMutex {
       DCHECK_NE(state & kWriterLock, 0);
       DCHECK_EQ(state & kReaderLockMask, 0);
       new_state = state & ~kWriterLock;
-      wake_writer =
-          (state & kWriterSpinWait) == 0 && (state & kWaitingWriterMask) != 0;
+      wake_writer = (state & (kWriterSpinWait | kReaderSpinWait)) == 0 &&
+                    (state & kWaitingWriterMask) != 0;
       if (wake_writer)
         new_state = (new_state - kWaitingWriterInc) | kWriterSpinWait;
       wake_readers =
-          (state & (kWriterSpinWait | kWaitingWriterMask)) != 0
+          wake_writer || (state & kWriterSpinWait) != 0
               ? 0
               : ((state & kWaitingReaderMask) >> kWaitingReaderShift);
       if (wake_readers)
-        new_state = (new_state & ~kWaitingReaderMask) +
-                    (wake_readers << kReaderLockShift);
+        new_state = (new_state & ~kWaitingReaderMask) | kReaderSpinWait;
     } while (UNLIKELY(!atomic_compare_exchange_weak(&state_, &state, new_state,
                                                     memory_order_release)));
     if (UNLIKELY(wake_writer))
@@ -233,23 +236,39 @@ class MUTEX Mutex : CheckedMutex {
 
   void ReadLock() ACQUIRE_SHARED() {
     CheckedMutex::Lock();
-    bool locked;
-    u64 new_state;
+    u64 reset_mask = ~0ull;
     u64 state = atomic_load_relaxed(&state_);
-    do {
-      locked =
-          (state & kReaderLockMask) == 0 &&
-          (state & (kWriterLock | kWriterSpinWait | kWaitingWriterMask)) != 0;
+    for (uptr spin_iters = 0;; spin_iters++) {
+      bool locked = (state & kWriterLock) != 0;
+      u64 new_state;
+      if (LIKELY(!locked)) {
+        new_state = (state + kReaderLockInc) & reset_mask;
+      } else if (spin_iters > kMaxSpinIters) {
+        new_state = (state + kWaitingReaderInc) & reset_mask;
+      } else if ((state & kReaderSpinWait) == 0) {
+        // Active spinning, but denote our presence so that unlocking
+        // thread does not wake up other threads.
+        new_state = state | kReaderSpinWait;
+      } else {
+        // Active spinning.
+        state = atomic_load(&state_, memory_order_relaxed);
+        continue;
+      }
+      if (UNLIKELY(!atomic_compare_exchange_weak(&state_, &state, new_state,
+                                                 memory_order_acquire)))
+        continue;
       if (LIKELY(!locked))
-        new_state = state + kReaderLockInc;
-      else
-        new_state = state + kWaitingReaderInc;
-    } while (UNLIKELY(!atomic_compare_exchange_weak(&state_, &state, new_state,
-                                                    memory_order_acquire)));
-    if (UNLIKELY(locked))
-      readers_.Wait();
-    DCHECK_EQ(atomic_load_relaxed(&state_) & kWriterLock, 0);
-    DCHECK_NE(atomic_load_relaxed(&state_) & kReaderLockMask, 0);
+        return;  // We've locked the mutex.
+      if (spin_iters > kMaxSpinIters) {
+        // We've incremented waiting readers, so now block.
+        readers_.Wait();
+        spin_iters = 0;
+      } else {
+        // We've set kReaderSpinWait, but we are still in active spinning.
+      }
+      reset_mask = ~kReaderSpinWait;
+      state = atomic_load(&state_, memory_order_relaxed);
+    }
   }
 
   void ReadUnlock() RELEASE_SHARED() {
@@ -259,9 +278,10 @@ class MUTEX Mutex : CheckedMutex {
     u64 state = atomic_load_relaxed(&state_);
     do {
       DCHECK_NE(state & kReaderLockMask, 0);
-      DCHECK_EQ(state & (kWaitingReaderMask | kWriterLock), 0);
+      DCHECK_EQ(state & kWriterLock, 0);
       new_state = state - kReaderLockInc;
-      wake = (new_state & (kReaderLockMask | kWriterSpinWait)) == 0 &&
+      wake = (new_state &
+              (kReaderLockMask | kWriterSpinWait | kReaderSpinWait)) == 0 &&
              (new_state & kWaitingWriterMask) != 0;
       if (wake)
         new_state = (new_state - kWaitingWriterInc) | kWriterSpinWait;
@@ -305,16 +325,14 @@ class MUTEX Mutex : CheckedMutex {
   //  - a writer is awake and spin-waiting
   //    the flag is used to prevent thundering herd problem
   //    (new writers are not woken if this flag is set)
+  //  - a reader is awake and spin-waiting
   //
-  // Writer support active spinning, readers does not.
+  // Both writers and readers use active spinning before blocking.
   // But readers are more aggressive and always take the mutex
   // if there are any other readers.
-  // Writers hand off the mutex to readers: after wake up readers
-  // already assume ownership of the mutex (don't need to do any
-  // state updates). But the mutex is not handed off to writers,
-  // after wake up writers compete to lock the mutex again.
-  // This is needed to allow repeated write locks even in presence
-  // of other blocked writers.
+  // After wake up both writers and readers compete to lock the
+  // mutex again. This is needed to allow repeated locks even in presence
+  // of other blocked threads.
   static constexpr u64 kCounterWidth = 20;
   static constexpr u64 kReaderLockShift = 0;
   static constexpr u64 kReaderLockInc = 1ull << kReaderLockShift;
@@ -330,7 +348,11 @@ class MUTEX Mutex : CheckedMutex {
                                             << kWaitingWriterShift;
   static constexpr u64 kWriterLock = 1ull << (3 * kCounterWidth);
   static constexpr u64 kWriterSpinWait = 1ull << (3 * kCounterWidth + 1);
+  static constexpr u64 kReaderSpinWait = 1ull << (3 * kCounterWidth + 2);
+
+  static constexpr uptr kMaxSpinIters = 1500;
 
+  Mutex(LinkerInitialized) = delete;
   Mutex(const Mutex &) = delete;
   void operator=(const Mutex &) = delete;
 };
@@ -338,111 +360,6 @@ class MUTEX Mutex : CheckedMutex {
 void FutexWait(atomic_uint32_t *p, u32 cmp);
 void FutexWake(atomic_uint32_t *p, u32 count);
 
-class MUTEX BlockingMutex {
- public:
-  explicit constexpr BlockingMutex(LinkerInitialized)
-      : opaque_storage_ {0, }, owner_ {0} {}
-  BlockingMutex();
-  void Lock() ACQUIRE();
-  void Unlock() RELEASE();
-
-  // This function does not guarantee an explicit check that the calling thread
-  // is the thread which owns the mutex. This behavior, while more strictly
-  // correct, causes problems in cases like StopTheWorld, where a parent thread
-  // owns the mutex but a child checks that it is locked. Rather than
-  // maintaining complex state to work around those situations, the check only
-  // checks that the mutex is owned, and assumes callers to be generally
-  // well-behaved.
-  void CheckLocked() const CHECK_LOCKED();
-
- private:
-  // Solaris mutex_t has a member that requires 64-bit alignment.
-  ALIGNED(8) uptr opaque_storage_[10];
-  uptr owner_;  // for debugging
-};
-
-// Reader-writer spin mutex.
-class MUTEX RWMutex {
- public:
-  RWMutex() {
-    atomic_store(&state_, kUnlocked, memory_order_relaxed);
-  }
-
-  ~RWMutex() {
-    CHECK_EQ(atomic_load(&state_, memory_order_relaxed), kUnlocked);
-  }
-
-  void Lock() ACQUIRE() {
-    u32 cmp = kUnlocked;
-    if (atomic_compare_exchange_strong(&state_, &cmp, kWriteLock,
-                                       memory_order_acquire))
-      return;
-    LockSlow();
-  }
-
-  void Unlock() RELEASE() {
-    u32 prev = atomic_fetch_sub(&state_, kWriteLock, memory_order_release);
-    DCHECK_NE(prev & kWriteLock, 0);
-    (void)prev;
-  }
-
-  void ReadLock() ACQUIRE_SHARED() {
-    u32 prev = atomic_fetch_add(&state_, kReadLock, memory_order_acquire);
-    if ((prev & kWriteLock) == 0)
-      return;
-    ReadLockSlow();
-  }
-
-  void ReadUnlock() RELEASE_SHARED() {
-    u32 prev = atomic_fetch_sub(&state_, kReadLock, memory_order_release);
-    DCHECK_EQ(prev & kWriteLock, 0);
-    DCHECK_GT(prev & ~kWriteLock, 0);
-    (void)prev;
-  }
-
-  void CheckLocked() const CHECK_LOCKED() {
-    CHECK_NE(atomic_load(&state_, memory_order_relaxed), kUnlocked);
-  }
-
- private:
-  atomic_uint32_t state_;
-
-  enum {
-    kUnlocked = 0,
-    kWriteLock = 1,
-    kReadLock = 2
-  };
-
-  void NOINLINE LockSlow() {
-    for (int i = 0;; i++) {
-      if (i < 10)
-        proc_yield(10);
-      else
-        internal_sched_yield();
-      u32 cmp = atomic_load(&state_, memory_order_relaxed);
-      if (cmp == kUnlocked &&
-          atomic_compare_exchange_weak(&state_, &cmp, kWriteLock,
-                                       memory_order_acquire))
-          return;
-    }
-  }
-
-  void NOINLINE ReadLockSlow() {
-    for (int i = 0;; i++) {
-      if (i < 10)
-        proc_yield(10);
-      else
-        internal_sched_yield();
-      u32 prev = atomic_load(&state_, memory_order_acquire);
-      if ((prev & kWriteLock) == 0)
-        return;
-    }
-  }
-
-  RWMutex(const RWMutex &) = delete;
-  void operator=(const RWMutex &) = delete;
-};
-
 template <typename MutexType>
 class SCOPED_LOCK GenericScopedLock {
  public:
@@ -475,12 +392,37 @@ class SCOPED_LOCK GenericScopedReadLock {
   void operator=(const GenericScopedReadLock &) = delete;
 };
 
+template <typename MutexType>
+class SCOPED_LOCK GenericScopedRWLock {
+ public:
+  ALWAYS_INLINE explicit GenericScopedRWLock(MutexType *mu, bool write)
+      ACQUIRE(mu)
+      : mu_(mu), write_(write) {
+    if (write_)
+      mu_->Lock();
+    else
+      mu_->ReadLock();
+  }
+
+  ALWAYS_INLINE ~GenericScopedRWLock() RELEASE() {
+    if (write_)
+      mu_->Unlock();
+    else
+      mu_->ReadUnlock();
+  }
+
+ private:
+  MutexType *mu_;
+  bool write_;
+
+  GenericScopedRWLock(const GenericScopedRWLock &) = delete;
+  void operator=(const GenericScopedRWLock &) = delete;
+};
+
 typedef GenericScopedLock<StaticSpinMutex> SpinMutexLock;
-typedef GenericScopedLock<BlockingMutex> BlockingMutexLock;
-typedef GenericScopedLock<RWMutex> RWMutexLock;
-typedef GenericScopedReadLock<RWMutex> RWMutexReadLock;
 typedef GenericScopedLock<Mutex> Lock;
 typedef GenericScopedReadLock<Mutex> ReadLock;
+typedef GenericScopedRWLock<Mutex> RWLock;
 
 }  // namespace __sanitizer
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_persistent_allocator.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_persistent_allocator.cpp
deleted file mode 100644
index 1ca0375b8a54..000000000000
--- a/compiler-rt/lib/sanitizer_common/sanitizer_persistent_allocator.cpp
+++ /dev/null
@@ -1,18 +0,0 @@
-//===-- sanitizer_persistent_allocator.cpp ----------------------*- C++ -*-===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-//
-// This file is shared between AddressSanitizer and ThreadSanitizer
-// run-time libraries.
-//===----------------------------------------------------------------------===//
-#include "sanitizer_persistent_allocator.h"
-
-namespace __sanitizer {
-
-PersistentAllocator thePersistentAllocator;
-
-}  // namespace __sanitizer
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_persistent_allocator.h b/compiler-rt/lib/sanitizer_common/sanitizer_persistent_allocator.h
deleted file mode 100644
index de4fb6ebc3cf..000000000000
--- a/compiler-rt/lib/sanitizer_common/sanitizer_persistent_allocator.h
+++ /dev/null
@@ -1,71 +0,0 @@
-//===-- sanitizer_persistent_allocator.h ------------------------*- C++ -*-===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-//
-// A fast memory allocator that does not support free() nor realloc().
-// All allocations are forever.
-//===----------------------------------------------------------------------===//
-
-#ifndef SANITIZER_PERSISTENT_ALLOCATOR_H
-#define SANITIZER_PERSISTENT_ALLOCATOR_H
-
-#include "sanitizer_internal_defs.h"
-#include "sanitizer_mutex.h"
-#include "sanitizer_atomic.h"
-#include "sanitizer_common.h"
-
-namespace __sanitizer {
-
-class PersistentAllocator {
- public:
-  void *alloc(uptr size);
-
- private:
-  void *tryAlloc(uptr size);
-  StaticSpinMutex mtx;  // Protects alloc of new blocks for region allocator.
-  atomic_uintptr_t region_pos;  // Region allocator for Node's.
-  atomic_uintptr_t region_end;
-};
-
-inline void *PersistentAllocator::tryAlloc(uptr size) {
-  // Optimisic lock-free allocation, essentially try to bump the region ptr.
-  for (;;) {
-    uptr cmp = atomic_load(&region_pos, memory_order_acquire);
-    uptr end = atomic_load(&region_end, memory_order_acquire);
-    if (cmp == 0 || cmp + size > end) return nullptr;
-    if (atomic_compare_exchange_weak(&region_pos, &cmp, cmp + size,
-                                     memory_order_acquire))
-      return (void *)cmp;
-  }
-}
-
-inline void *PersistentAllocator::alloc(uptr size) {
-  // First, try to allocate optimisitically.
-  void *s = tryAlloc(size);
-  if (s) return s;
-  // If failed, lock, retry and alloc new superblock.
-  SpinMutexLock l(&mtx);
-  for (;;) {
-    s = tryAlloc(size);
-    if (s) return s;
-    atomic_store(&region_pos, 0, memory_order_relaxed);
-    uptr allocsz = 64 * 1024;
-    if (allocsz < size) allocsz = size;
-    uptr mem = (uptr)MmapOrDie(allocsz, "stack depot");
-    atomic_store(&region_end, mem + allocsz, memory_order_release);
-    atomic_store(&region_pos, mem, memory_order_release);
-  }
-}
-
-extern PersistentAllocator thePersistentAllocator;
-inline void *PersistentAlloc(uptr sz) {
-  return thePersistentAllocator.alloc(sz);
-}
-
-} // namespace __sanitizer
-
-#endif // SANITIZER_PERSISTENT_ALLOCATOR_H
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform.h b/compiler-rt/lib/sanitizer_common/sanitizer_platform.h
index 4d3c08893c11..3153de34e5a3 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform.h
@@ -281,11 +281,12 @@
 // mandated by the upstream linux community for all new ports. Other ports
 // may still use legacy syscalls.
 #ifndef SANITIZER_USES_CANONICAL_LINUX_SYSCALLS
-# if (defined(__aarch64__) || defined(__riscv)) && SANITIZER_LINUX
-# define SANITIZER_USES_CANONICAL_LINUX_SYSCALLS 1
-# else
-# define SANITIZER_USES_CANONICAL_LINUX_SYSCALLS 0
-# endif
+#  if (defined(__aarch64__) || defined(__riscv) || defined(__hexagon__)) && \
+      SANITIZER_LINUX
+#    define SANITIZER_USES_CANONICAL_LINUX_SYSCALLS 1
+#  else
+#    define SANITIZER_USES_CANONICAL_LINUX_SYSCALLS 0
+#  endif
 #endif
 
 // udi16 syscalls can only be used when the following conditions are
@@ -377,4 +378,18 @@
 #define SANITIZER_SUPPORTS_INIT_FOR_DLOPEN 0
 #endif
 
+// SANITIZER_SUPPORTS_THREADLOCAL
+// 1 - THREADLOCAL macro is supported by target
+// 0 - THREADLOCAL macro is not supported by target
+#ifndef __has_feature
+// TODO: Support other compilers here
+#  define SANITIZER_SUPPORTS_THREADLOCAL 1
+#else
+#  if __has_feature(tls)
+#    define SANITIZER_SUPPORTS_THREADLOCAL 1
+#  else
+#    define SANITIZER_SUPPORTS_THREADLOCAL 0
+#  endif
+#endif
+
 #endif // SANITIZER_PLATFORM_H
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h b/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h
index 5b710c23fd00..14610f2df78d 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h
@@ -229,11 +229,13 @@
   (SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_CLOCK_GETTIME \
   (SI_FREEBSD || SI_NETBSD || SI_LINUX || SI_SOLARIS)
-#define SANITIZER_INTERCEPT_CLOCK_GETCPUCLOCKID SI_LINUX
+#define SANITIZER_INTERCEPT_CLOCK_GETCPUCLOCKID \
+  (SI_LINUX || SI_FREEBSD || SI_NETBSD)
 #define SANITIZER_INTERCEPT_GETITIMER SI_POSIX
 #define SANITIZER_INTERCEPT_TIME SI_POSIX
 #define SANITIZER_INTERCEPT_GLOB (SI_GLIBC || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_GLOB64 SI_GLIBC
+#define SANITIZER_INTERCEPT_POSIX_SPAWN SI_POSIX
 #define SANITIZER_INTERCEPT_WAIT SI_POSIX
 #define SANITIZER_INTERCEPT_INET SI_POSIX
 #define SANITIZER_INTERCEPT_PTHREAD_GETSCHEDPARAM SI_POSIX
@@ -251,7 +253,8 @@
 #define SANITIZER_INTERCEPT_GETHOSTENT_R (SI_FREEBSD || SI_GLIBC || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_GETSOCKOPT SI_POSIX
 #define SANITIZER_INTERCEPT_ACCEPT SI_POSIX
-#define SANITIZER_INTERCEPT_ACCEPT4 (SI_LINUX_NOT_ANDROID || SI_NETBSD)
+#define SANITIZER_INTERCEPT_ACCEPT4 \
+  (SI_LINUX_NOT_ANDROID || SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_PACCEPT SI_NETBSD
 #define SANITIZER_INTERCEPT_MODF SI_POSIX
 #define SANITIZER_INTERCEPT_RECVMSG SI_POSIX
@@ -309,7 +312,7 @@
 #define SANITIZER_INTERCEPT_PPOLL SI_LINUX_NOT_ANDROID || SI_SOLARIS
 #define SANITIZER_INTERCEPT_WORDEXP                                          \
   (SI_FREEBSD || SI_NETBSD || (SI_MAC && !SI_IOS) || SI_LINUX_NOT_ANDROID || \
-   SI_SOLARIS)  // NOLINT
+   SI_SOLARIS)
 #define SANITIZER_INTERCEPT_SIGWAIT SI_POSIX
 #define SANITIZER_INTERCEPT_SIGWAITINFO SI_LINUX_NOT_ANDROID || SI_SOLARIS
 #define SANITIZER_INTERCEPT_SIGTIMEDWAIT SI_LINUX_NOT_ANDROID || SI_SOLARIS
@@ -337,7 +340,7 @@
 #define SANITIZER_INTERCEPT_ETHER_R (SI_FREEBSD || SI_LINUX_NOT_ANDROID)
 #define SANITIZER_INTERCEPT_SHMCTL                                       \
   (((SI_FREEBSD || SI_LINUX_NOT_ANDROID) && SANITIZER_WORDSIZE == 64) || \
-   SI_NETBSD || SI_SOLARIS)  // NOLINT
+   SI_NETBSD || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_RANDOM_R SI_GLIBC
 #define SANITIZER_INTERCEPT_PTHREAD_ATTR_GET SI_POSIX
 #define SANITIZER_INTERCEPT_PTHREAD_ATTR_GETINHERITSCHED \
@@ -445,7 +448,8 @@
 #define SANITIZER_INTERCEPT_SEM \
   (SI_LINUX || SI_FREEBSD || SI_NETBSD || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_PTHREAD_SETCANCEL SI_POSIX
-#define SANITIZER_INTERCEPT_MINCORE (SI_LINUX || SI_NETBSD || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_MINCORE \
+  (SI_LINUX || SI_NETBSD || SI_FREEBSD || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_PROCESS_VM_READV SI_LINUX
 #define SANITIZER_INTERCEPT_CTERMID \
   (SI_LINUX || SI_MAC || SI_FREEBSD || SI_NETBSD || SI_SOLARIS)
@@ -457,10 +461,13 @@
 #define SANITIZER_INTERCEPT_SEND_SENDTO SI_POSIX
 #define SANITIZER_INTERCEPT_EVENTFD_READ_WRITE SI_LINUX
 
-#define SANITIZER_INTERCEPT_STAT \
-  (SI_FREEBSD || SI_MAC || SI_ANDROID || SI_NETBSD || SI_SOLARIS)
-#define SANITIZER_INTERCEPT_LSTAT (SI_NETBSD || SI_FREEBSD)
-#define SANITIZER_INTERCEPT___XSTAT (!SANITIZER_INTERCEPT_STAT && SI_POSIX)
+#define SI_STAT_LINUX (SI_LINUX && __GLIBC_PREREQ(2, 33))
+#define SANITIZER_INTERCEPT_STAT                                        \
+  (SI_FREEBSD || SI_MAC || SI_ANDROID || SI_NETBSD || SI_SOLARIS ||     \
+   SI_STAT_LINUX)
+#define SANITIZER_INTERCEPT_LSTAT (SI_NETBSD || SI_FREEBSD || SI_STAT_LINUX)
+#define SANITIZER_INTERCEPT___XSTAT \
+  ((!SANITIZER_INTERCEPT_STAT && SI_POSIX) || SI_STAT_LINUX)
 #define SANITIZER_INTERCEPT___XSTAT64 SI_LINUX_NOT_ANDROID
 #define SANITIZER_INTERCEPT___LXSTAT SANITIZER_INTERCEPT___XSTAT
 #define SANITIZER_INTERCEPT___LXSTAT64 SI_LINUX_NOT_ANDROID
@@ -474,7 +481,7 @@
   (SI_LINUX_NOT_ANDROID || SI_MAC || SI_FREEBSD || SI_NETBSD)
 
 #define SANITIZER_INTERCEPT_MMAP SI_POSIX
-#define SANITIZER_INTERCEPT_MMAP64 SI_LINUX_NOT_ANDROID
+#define SANITIZER_INTERCEPT_MMAP64 SI_LINUX_NOT_ANDROID || SI_SOLARIS
 #define SANITIZER_INTERCEPT_MALLOPT_AND_MALLINFO (SI_GLIBC || SI_ANDROID)
 #define SANITIZER_INTERCEPT_MEMALIGN (!SI_FREEBSD && !SI_MAC && !SI_NETBSD)
 #define SANITIZER_INTERCEPT___LIBC_MEMALIGN SI_GLIBC
@@ -496,7 +503,8 @@
 #define SANITIZER_INTERCEPT_GID_FROM_GROUP SI_NETBSD
 #define SANITIZER_INTERCEPT_ACCESS (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_FACCESSAT (SI_NETBSD || SI_FREEBSD)
-#define SANITIZER_INTERCEPT_GETGROUPLIST SI_NETBSD
+#define SANITIZER_INTERCEPT_GETGROUPLIST \
+  (SI_NETBSD || SI_FREEBSD || SI_LINUX)
 #define SANITIZER_INTERCEPT_STRLCPY \
   (SI_NETBSD || SI_FREEBSD || SI_MAC || SI_ANDROID)
 
@@ -517,10 +525,11 @@
 #define SANITIZER_INTERCEPT_DEVNAME_R (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_FGETLN (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_STRMODE (SI_NETBSD || SI_FREEBSD)
-#define SANITIZER_INTERCEPT_TTYENT SI_NETBSD
-#define SANITIZER_INTERCEPT_PROTOENT (SI_NETBSD || SI_LINUX)
+#define SANITIZER_INTERCEPT_TTYENT (SI_NETBSD || SI_FREEBSD)
+#define SANITIZER_INTERCEPT_TTYENTPATH SI_NETBSD
+#define SANITIZER_INTERCEPT_PROTOENT (SI_LINUX || SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_PROTOENT_R SI_GLIBC
-#define SANITIZER_INTERCEPT_NETENT SI_NETBSD
+#define SANITIZER_INTERCEPT_NETENT (SI_LINUX || SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_SETVBUF \
   (SI_NETBSD || SI_FREEBSD || SI_LINUX || SI_MAC)
 #define SANITIZER_INTERCEPT_GETMNTINFO (SI_NETBSD || SI_FREEBSD || SI_MAC)
@@ -536,17 +545,17 @@
 #define SANITIZER_INTERCEPT_MODCTL SI_NETBSD
 #define SANITIZER_INTERCEPT_CAPSICUM SI_FREEBSD
 #define SANITIZER_INTERCEPT_STRTONUM (SI_NETBSD || SI_FREEBSD)
-#define SANITIZER_INTERCEPT_FPARSELN SI_NETBSD
+#define SANITIZER_INTERCEPT_FPARSELN (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_STATVFS1 SI_NETBSD
 #define SANITIZER_INTERCEPT_STRTOI SI_NETBSD
 #define SANITIZER_INTERCEPT_CAPSICUM SI_FREEBSD
 #define SANITIZER_INTERCEPT_SHA1 SI_NETBSD
 #define SANITIZER_INTERCEPT_MD4 SI_NETBSD
 #define SANITIZER_INTERCEPT_RMD160 SI_NETBSD
-#define SANITIZER_INTERCEPT_MD5 SI_NETBSD
+#define SANITIZER_INTERCEPT_MD5 (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_FSEEK (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_MD2 SI_NETBSD
-#define SANITIZER_INTERCEPT_SHA2 SI_NETBSD
+#define SANITIZER_INTERCEPT_SHA2 (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_CDB SI_NETBSD
 #define SANITIZER_INTERCEPT_VIS (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_POPEN SI_POSIX
@@ -571,6 +580,8 @@
 #define SANITIZER_INTERCEPT_QSORT \
   (SI_POSIX && !SI_IOSSIM && !SI_WATCHOS && !SI_TVOS && !SI_ANDROID)
 #define SANITIZER_INTERCEPT_QSORT_R SI_GLIBC
+#define SANITIZER_INTERCEPT_BSEARCH \
+  (SI_POSIX && !SI_IOSSIM && !SI_WATCHOS && !SI_TVOS && !SI_ANDROID)
 // sigaltstack on i386 macOS cannot be intercepted due to setjmp()
 // calling it and assuming that it does not clobber registers.
 #define SANITIZER_INTERCEPT_SIGALTSTACK \
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_freebsd.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_freebsd.cpp
index b5a45ae72cd9..64535805e40d 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_freebsd.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_freebsd.cpp
@@ -69,11 +69,17 @@
 #include <semaphore.h>
 #include <signal.h>
 #include <stddef.h>
+#include <md5.h>
+#include <sha224.h>
+#include <sha256.h>
+#include <sha384.h>
+#include <sha512.h>
 #include <stdio.h>
 #include <stringlist.h>
 #include <term.h>
 #include <termios.h>
 #include <time.h>
+#include <ttyent.h>
 #include <utime.h>
 #include <utmpx.h>
 #include <vis.h>
@@ -170,9 +176,12 @@ uptr __sanitizer_in_addr_sz(int af) {
 unsigned struct_ElfW_Phdr_sz = sizeof(Elf_Phdr);
 int glob_nomatch = GLOB_NOMATCH;
 int glob_altdirfunc = GLOB_ALTDIRFUNC;
+const int wordexp_wrde_dooffs = WRDE_DOOFFS;
 
 unsigned path_max = PATH_MAX;
 
+int struct_ttyent_sz = sizeof(struct ttyent);
+
 // ioctl arguments
 unsigned struct_ifreq_sz = sizeof(struct ifreq);
 unsigned struct_termios_sz = sizeof(struct termios);
@@ -357,6 +366,22 @@ const int si_SEGV_MAPERR = SEGV_MAPERR;
 const int si_SEGV_ACCERR = SEGV_ACCERR;
 const int unvis_valid = UNVIS_VALID;
 const int unvis_validpush = UNVIS_VALIDPUSH;
+
+const unsigned MD5_CTX_sz = sizeof(MD5_CTX);
+const unsigned MD5_return_length = MD5_DIGEST_STRING_LENGTH;
+
+#define SHA2_CONST(LEN)                                                      \
+  const unsigned SHA##LEN##_CTX_sz = sizeof(SHA##LEN##_CTX);                 \
+  const unsigned SHA##LEN##_return_length = SHA##LEN##_DIGEST_STRING_LENGTH; \
+  const unsigned SHA##LEN##_block_length = SHA##LEN##_BLOCK_LENGTH;          \
+  const unsigned SHA##LEN##_digest_length = SHA##LEN##_DIGEST_LENGTH
+
+SHA2_CONST(224);
+SHA2_CONST(256);
+SHA2_CONST(384);
+SHA2_CONST(512);
+
+#undef SHA2_CONST
 }  // namespace __sanitizer
 
 using namespace __sanitizer;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_freebsd.h b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_freebsd.h
index 5e0ca9c7d782..649e64fd1a32 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_freebsd.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_freebsd.h
@@ -16,26 +16,26 @@
 
 #if SANITIZER_FREEBSD
 
-#include "sanitizer_internal_defs.h"
-#include "sanitizer_platform.h"
-#include "sanitizer_platform_limits_posix.h"
+#  include "sanitizer_internal_defs.h"
+#  include "sanitizer_platform.h"
+#  include "sanitizer_platform_limits_posix.h"
 
 // Get sys/_types.h, because that tells us whether 64-bit inodes are
 // used in struct dirent below.
-#include <sys/_types.h>
+#  include <sys/_types.h>
 
 namespace __sanitizer {
 void *__sanitizer_get_link_map_by_dlopen_handle(void *handle);
-#define GET_LINK_MAP_BY_DLOPEN_HANDLE(handle) \
-  (link_map *)__sanitizer_get_link_map_by_dlopen_handle(handle)
+#  define GET_LINK_MAP_BY_DLOPEN_HANDLE(handle) \
+    (link_map *)__sanitizer_get_link_map_by_dlopen_handle(handle)
 
 extern unsigned struct_utsname_sz;
 extern unsigned struct_stat_sz;
-#if defined(__powerpc64__)
+#  if defined(__powerpc64__)
 const unsigned struct___old_kernel_stat_sz = 0;
-#else
+#  else
 const unsigned struct___old_kernel_stat_sz = 32;
-#endif
+#  endif
 extern unsigned struct_rusage_sz;
 extern unsigned siginfo_t_sz;
 extern unsigned struct_itimerval_sz;
@@ -114,11 +114,24 @@ struct __sanitizer_ipc_perm {
   long key;
 };
 
-#if !defined(__i386__)
+struct __sanitizer_protoent {
+  char *p_name;
+  char **p_aliases;
+  int p_proto;
+};
+
+struct __sanitizer_netent {
+  char *n_name;
+  char **n_aliases;
+  int n_addrtype;
+  u32 n_net;
+};
+
+#  if !defined(__i386__)
 typedef long long __sanitizer_time_t;
-#else
+#  else
 typedef long __sanitizer_time_t;
-#endif
+#  endif
 
 struct __sanitizer_shmid_ds {
   __sanitizer_ipc_perm shm_perm;
@@ -147,7 +160,7 @@ struct __sanitizer_ifaddrs {
   unsigned int ifa_flags;
   void *ifa_addr;     // (struct sockaddr *)
   void *ifa_netmask;  // (struct sockaddr *)
-#undef ifa_dstaddr
+#  undef ifa_dstaddr
   void *ifa_dstaddr;  // (struct sockaddr *)
   void *ifa_data;
 };
@@ -229,12 +242,12 @@ struct __sanitizer_cmsghdr {
 };
 
 struct __sanitizer_dirent {
-#if defined(__INO64)
+#  if defined(__INO64)
   unsigned long long d_fileno;
   unsigned long long d_off;
-#else
+#  else
   unsigned int d_fileno;
-#endif
+#  endif
   unsigned short d_reclen;
   // more fields that we don't care about
 };
@@ -243,23 +256,23 @@ struct __sanitizer_dirent {
 typedef int __sanitizer_clock_t;
 typedef int __sanitizer_clockid_t;
 
-#if defined(_LP64) || defined(__x86_64__) || defined(__powerpc__) || \
-    defined(__mips__)
+#  if defined(_LP64) || defined(__x86_64__) || defined(__powerpc__) || \
+      defined(__mips__)
 typedef unsigned __sanitizer___kernel_uid_t;
 typedef unsigned __sanitizer___kernel_gid_t;
-#else
+#  else
 typedef unsigned short __sanitizer___kernel_uid_t;
 typedef unsigned short __sanitizer___kernel_gid_t;
-#endif
+#  endif
 typedef long long __sanitizer___kernel_off_t;
 
-#if defined(__powerpc__) || defined(__mips__)
+#  if defined(__powerpc__) || defined(__mips__)
 typedef unsigned int __sanitizer___kernel_old_uid_t;
 typedef unsigned int __sanitizer___kernel_old_gid_t;
-#else
+#  else
 typedef unsigned short __sanitizer___kernel_old_uid_t;
 typedef unsigned short __sanitizer___kernel_old_gid_t;
-#endif
+#  endif
 
 typedef long long __sanitizer___kernel_loff_t;
 typedef struct {
@@ -366,9 +379,12 @@ struct __sanitizer_glob_t {
 
 extern int glob_nomatch;
 extern int glob_altdirfunc;
+extern const int wordexp_wrde_dooffs;
 
 extern unsigned path_max;
 
+extern int struct_ttyent_sz;
+
 struct __sanitizer_wordexp_t {
   uptr we_wordc;
   char **we_wordv;
@@ -398,39 +414,49 @@ struct __sanitizer_ifconf {
   } ifc_ifcu;
 };
 
-#define IOC_NRBITS 8
-#define IOC_TYPEBITS 8
-#if defined(__powerpc__) || defined(__powerpc64__) || defined(__mips__)
-#define IOC_SIZEBITS 13
-#define IOC_DIRBITS 3
-#define IOC_NONE 1U
-#define IOC_WRITE 4U
-#define IOC_READ 2U
-#else
-#define IOC_SIZEBITS 14
-#define IOC_DIRBITS 2
-#define IOC_NONE 0U
-#define IOC_WRITE 1U
-#define IOC_READ 2U
-#endif
-#define IOC_NRMASK ((1 << IOC_NRBITS) - 1)
-#define IOC_TYPEMASK ((1 << IOC_TYPEBITS) - 1)
-#define IOC_SIZEMASK ((1 << IOC_SIZEBITS) - 1)
-#if defined(IOC_DIRMASK)
-#undef IOC_DIRMASK
-#endif
-#define IOC_DIRMASK ((1 << IOC_DIRBITS) - 1)
-#define IOC_NRSHIFT 0
-#define IOC_TYPESHIFT (IOC_NRSHIFT + IOC_NRBITS)
-#define IOC_SIZESHIFT (IOC_TYPESHIFT + IOC_TYPEBITS)
-#define IOC_DIRSHIFT (IOC_SIZESHIFT + IOC_SIZEBITS)
-#define EVIOC_EV_MAX 0x1f
-#define EVIOC_ABS_MAX 0x3f
+struct __sanitizer__ttyent {
+  char *ty_name;
+  char *ty_getty;
+  char *ty_type;
+  int ty_status;
+  char *ty_window;
+  char *ty_comment;
+  char *ty_group;
+};
+
+#  define IOC_NRBITS 8
+#  define IOC_TYPEBITS 8
+#  if defined(__powerpc__) || defined(__powerpc64__) || defined(__mips__)
+#    define IOC_SIZEBITS 13
+#    define IOC_DIRBITS 3
+#    define IOC_NONE 1U
+#    define IOC_WRITE 4U
+#    define IOC_READ 2U
+#  else
+#    define IOC_SIZEBITS 14
+#    define IOC_DIRBITS 2
+#    define IOC_NONE 0U
+#    define IOC_WRITE 1U
+#    define IOC_READ 2U
+#  endif
+#  define IOC_NRMASK ((1 << IOC_NRBITS) - 1)
+#  define IOC_TYPEMASK ((1 << IOC_TYPEBITS) - 1)
+#  define IOC_SIZEMASK ((1 << IOC_SIZEBITS) - 1)
+#  if defined(IOC_DIRMASK)
+#    undef IOC_DIRMASK
+#  endif
+#  define IOC_DIRMASK ((1 << IOC_DIRBITS) - 1)
+#  define IOC_NRSHIFT 0
+#  define IOC_TYPESHIFT (IOC_NRSHIFT + IOC_NRBITS)
+#  define IOC_SIZESHIFT (IOC_TYPESHIFT + IOC_TYPEBITS)
+#  define IOC_DIRSHIFT (IOC_SIZESHIFT + IOC_SIZEBITS)
+#  define EVIOC_EV_MAX 0x1f
+#  define EVIOC_ABS_MAX 0x3f
 
-#define IOC_DIR(nr) (((nr) >> IOC_DIRSHIFT) & IOC_DIRMASK)
-#define IOC_TYPE(nr) (((nr) >> IOC_TYPESHIFT) & IOC_TYPEMASK)
-#define IOC_NR(nr) (((nr) >> IOC_NRSHIFT) & IOC_NRMASK)
-#define IOC_SIZE(nr) (((nr) >> IOC_SIZESHIFT) & IOC_SIZEMASK)
+#  define IOC_DIR(nr) (((nr) >> IOC_DIRSHIFT) & IOC_DIRMASK)
+#  define IOC_TYPE(nr) (((nr) >> IOC_TYPESHIFT) & IOC_TYPEMASK)
+#  define IOC_NR(nr) (((nr) >> IOC_NRSHIFT) & IOC_NRMASK)
+#  define IOC_SIZE(nr) (((nr) >> IOC_SIZESHIFT) & IOC_SIZEMASK)
 
 extern unsigned struct_ifreq_sz;
 extern unsigned struct_termios_sz;
@@ -621,6 +647,22 @@ extern unsigned IOCTL_KDSKBMODE;
 extern const int si_SEGV_MAPERR;
 extern const int si_SEGV_ACCERR;
 
+extern const unsigned MD5_CTX_sz;
+extern const unsigned MD5_return_length;
+
+#define SHA2_EXTERN(LEN)                          \
+  extern const unsigned SHA##LEN##_CTX_sz;        \
+  extern const unsigned SHA##LEN##_return_length; \
+  extern const unsigned SHA##LEN##_block_length;  \
+  extern const unsigned SHA##LEN##_digest_length
+
+SHA2_EXTERN(224);
+SHA2_EXTERN(256);
+SHA2_EXTERN(384);
+SHA2_EXTERN(512);
+
+#undef SHA2_EXTERN
+
 struct __sanitizer_cap_rights {
   u64 cr_rights[2];
 };
@@ -632,24 +674,24 @@ extern unsigned struct_fstab_sz;
 extern unsigned struct_StringList_sz;
 }  // namespace __sanitizer
 
-#define CHECK_TYPE_SIZE(TYPE) \
-  COMPILER_CHECK(sizeof(__sanitizer_##TYPE) == sizeof(TYPE))
+#  define CHECK_TYPE_SIZE(TYPE) \
+    COMPILER_CHECK(sizeof(__sanitizer_##TYPE) == sizeof(TYPE))
 
-#define CHECK_SIZE_AND_OFFSET(CLASS, MEMBER)                      \
-  COMPILER_CHECK(sizeof(((__sanitizer_##CLASS *)NULL)->MEMBER) == \
-                 sizeof(((CLASS *)NULL)->MEMBER));                \
-  COMPILER_CHECK(offsetof(__sanitizer_##CLASS, MEMBER) ==         \
-                 offsetof(CLASS, MEMBER))
+#  define CHECK_SIZE_AND_OFFSET(CLASS, MEMBER)                      \
+    COMPILER_CHECK(sizeof(((__sanitizer_##CLASS *)NULL)->MEMBER) == \
+                   sizeof(((CLASS *)NULL)->MEMBER));                \
+    COMPILER_CHECK(offsetof(__sanitizer_##CLASS, MEMBER) ==         \
+                   offsetof(CLASS, MEMBER))
 
 // For sigaction, which is a function and struct at the same time,
 // and thus requires explicit "struct" in sizeof() expression.
-#define CHECK_STRUCT_SIZE_AND_OFFSET(CLASS, MEMBER)                      \
-  COMPILER_CHECK(sizeof(((struct __sanitizer_##CLASS *)NULL)->MEMBER) == \
-                 sizeof(((struct CLASS *)NULL)->MEMBER));                \
-  COMPILER_CHECK(offsetof(struct __sanitizer_##CLASS, MEMBER) ==         \
-                 offsetof(struct CLASS, MEMBER))
+#  define CHECK_STRUCT_SIZE_AND_OFFSET(CLASS, MEMBER)                      \
+    COMPILER_CHECK(sizeof(((struct __sanitizer_##CLASS *)NULL)->MEMBER) == \
+                   sizeof(((struct CLASS *)NULL)->MEMBER));                \
+    COMPILER_CHECK(offsetof(struct __sanitizer_##CLASS, MEMBER) ==         \
+                   offsetof(struct CLASS, MEMBER))
 
-#define SIGACTION_SYMNAME sigaction
+#  define SIGACTION_SYMNAME sigaction
 
 #endif
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_linux.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_linux.cpp
index c51327e1269e..9d577570ea1e 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_linux.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_linux.cpp
@@ -28,32 +28,32 @@
 // are not defined anywhere in userspace headers. Fake them. This seems to work
 // fine with newer headers, too.
 #include <linux/posix_types.h>
-#if defined(__x86_64__) ||  defined(__mips__)
-#include <sys/stat.h>
-#else
-#define ino_t __kernel_ino_t
-#define mode_t __kernel_mode_t
-#define nlink_t __kernel_nlink_t
-#define uid_t __kernel_uid_t
-#define gid_t __kernel_gid_t
-#define off_t __kernel_off_t
-#define time_t __kernel_time_t
+#  if defined(__x86_64__) || defined(__mips__) || defined(__hexagon__)
+#    include <sys/stat.h>
+#  else
+#    define ino_t __kernel_ino_t
+#    define mode_t __kernel_mode_t
+#    define nlink_t __kernel_nlink_t
+#    define uid_t __kernel_uid_t
+#    define gid_t __kernel_gid_t
+#    define off_t __kernel_off_t
+#    define time_t __kernel_time_t
 // This header seems to contain the definitions of _kernel_ stat* structs.
-#include <asm/stat.h>
-#undef ino_t
-#undef mode_t
-#undef nlink_t
-#undef uid_t
-#undef gid_t
-#undef off_t
-#endif
+#    include <asm/stat.h>
+#    undef ino_t
+#    undef mode_t
+#    undef nlink_t
+#    undef uid_t
+#    undef gid_t
+#    undef off_t
+#  endif
 
-#include <linux/aio_abi.h>
+#  include <linux/aio_abi.h>
 
-#if !SANITIZER_ANDROID
-#include <sys/statfs.h>
-#include <linux/perf_event.h>
-#endif
+#  if !SANITIZER_ANDROID
+#    include <sys/statfs.h>
+#    include <linux/perf_event.h>
+#  endif
 
 using namespace __sanitizer;
 
@@ -63,9 +63,9 @@ namespace __sanitizer {
 #endif
 }  // namespace __sanitizer
 
-#if !defined(__powerpc64__) && !defined(__x86_64__) && !defined(__aarch64__)\
-                            && !defined(__mips__) && !defined(__s390__)\
-                            && !defined(__sparc__) && !defined(__riscv)
+#  if !defined(__powerpc64__) && !defined(__x86_64__) &&                   \
+      !defined(__aarch64__) && !defined(__mips__) && !defined(__s390__) && \
+      !defined(__sparc__) && !defined(__riscv) && !defined(__hexagon__)
 COMPILER_CHECK(struct___old_kernel_stat_sz == sizeof(struct __old_kernel_stat));
 #endif
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.cpp
index c8f2aa5dba4a..531e07f2d4c5 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.cpp
@@ -666,6 +666,7 @@ unsigned struct_ElfW_Phdr_sz = sizeof(Elf_Phdr);
 
 int glob_nomatch = GLOB_NOMATCH;
 int glob_altdirfunc = GLOB_ALTDIRFUNC;
+const int wordexp_wrde_dooffs = WRDE_DOOFFS;
 
 unsigned path_max = PATH_MAX;
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.h b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.h
index 9e28dcfef041..9407803fc9c3 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.h
@@ -394,6 +394,7 @@ struct __sanitizer_glob_t {
 
 extern int glob_nomatch;
 extern int glob_altdirfunc;
+extern const int wordexp_wrde_dooffs;
 
 extern unsigned path_max;
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.cpp
index 6e5c330b98ef..a1c452855ae7 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.cpp
@@ -91,10 +91,10 @@
 #if SANITIZER_LINUX
 # include <utime.h>
 # include <sys/ptrace.h>
-#if defined(__mips64) || defined(__aarch64__) || defined(__arm__) || \
-    SANITIZER_RISCV64
-#  include <asm/ptrace.h>
-#  ifdef __arm__
+#    if defined(__mips64) || defined(__aarch64__) || defined(__arm__) || \
+        defined(__hexagon__) || SANITIZER_RISCV64
+#      include <asm/ptrace.h>
+#      ifdef __arm__
 typedef struct user_fpregs elf_fpregset_t;
 #   define ARM_VFPREGS_SIZE_ASAN (32 * 8 /*fpregs*/ + 4 /*fpscr*/)
 #   if !defined(ARM_VFPREGS_SIZE)
@@ -242,12 +242,13 @@ namespace __sanitizer {
     defined(__powerpc64__) || defined(__arch64__) || defined(__sparcv9) || \
     defined(__x86_64__) || SANITIZER_RISCV64
 #define SIZEOF_STRUCT_USTAT 32
-#elif defined(__arm__) || defined(__i386__) || defined(__mips__) \
-  || defined(__powerpc__) || defined(__s390__) || defined(__sparc__)
-#define SIZEOF_STRUCT_USTAT 20
-#else
-#error Unknown size of struct ustat
-#endif
+#    elif defined(__arm__) || defined(__i386__) || defined(__mips__) ||    \
+        defined(__powerpc__) || defined(__s390__) || defined(__sparc__) || \
+        defined(__hexagon__)
+#      define SIZEOF_STRUCT_USTAT 20
+#    else
+#      error Unknown size of struct ustat
+#    endif
   unsigned struct_ustat_sz = SIZEOF_STRUCT_USTAT;
   unsigned struct_rlimit64_sz = sizeof(struct rlimit64);
   unsigned struct_statvfs64_sz = sizeof(struct statvfs64);
@@ -312,6 +313,10 @@ unsigned struct_ElfW_Phdr_sz = sizeof(Elf_Phdr);
   int glob_altdirfunc = GLOB_ALTDIRFUNC;
 #endif
 
+#  if !SANITIZER_ANDROID
+  const int wordexp_wrde_dooffs = WRDE_DOOFFS;
+#  endif  // !SANITIZER_ANDROID
+
 #if SANITIZER_LINUX && !SANITIZER_ANDROID &&                               \
     (defined(__i386) || defined(__x86_64) || defined(__mips64) ||          \
      defined(__powerpc64__) || defined(__aarch64__) || defined(__arm__) || \
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
index 4dd27644ed11..d69b344dd613 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
@@ -102,7 +102,10 @@ const unsigned struct_kernel_stat64_sz = 104;
 #elif SANITIZER_RISCV64
 const unsigned struct_kernel_stat_sz = 128;
 const unsigned struct_kernel_stat64_sz = 0;  // RISCV64 does not use stat64
-#endif
+#    elif defined(__hexagon__)
+const unsigned struct_kernel_stat_sz = 128;
+const unsigned struct_kernel_stat64_sz = 0;
+#    endif
 struct __sanitizer_perf_event_attr {
   unsigned type;
   unsigned size;
@@ -367,7 +370,7 @@ struct __sanitizer_group {
   char **gr_mem;
 };
 
-#if defined(__x86_64__) && !defined(_LP64)
+#  if (defined(__x86_64__) && !defined(_LP64)) || defined(__hexagon__)
 typedef long long __sanitizer_time_t;
 #else
 typedef long __sanitizer_time_t;
@@ -475,23 +478,23 @@ struct __sanitizer_dirent {
   unsigned short d_reclen;
   // more fields that we don't care about
 };
-#elif SANITIZER_ANDROID || defined(__x86_64__)
+#  elif SANITIZER_ANDROID || defined(__x86_64__) || defined(__hexagon__)
 struct __sanitizer_dirent {
   unsigned long long d_ino;
   unsigned long long d_off;
   unsigned short d_reclen;
   // more fields that we don't care about
 };
-#else
+#  else
 struct __sanitizer_dirent {
   uptr d_ino;
   uptr d_off;
   unsigned short d_reclen;
   // more fields that we don't care about
 };
-#endif
+#  endif
 
-#if SANITIZER_LINUX && !SANITIZER_ANDROID
+#  if SANITIZER_LINUX && !SANITIZER_ANDROID
 struct __sanitizer_dirent64 {
   unsigned long long d_ino;
   unsigned long long d_off;
@@ -511,8 +514,8 @@ typedef int __sanitizer_clockid_t;
 #endif
 
 #if SANITIZER_LINUX
-#if defined(_LP64) || defined(__x86_64__) || defined(__powerpc__) || \
-    defined(__mips__)
+#    if defined(_LP64) || defined(__x86_64__) || defined(__powerpc__) || \
+        defined(__mips__) || defined(__hexagon__)
 typedef unsigned __sanitizer___kernel_uid_t;
 typedef unsigned __sanitizer___kernel_gid_t;
 #else
@@ -712,6 +715,13 @@ struct __sanitizer_protoent {
   int p_proto;
 };
 
+struct __sanitizer_netent {
+  char *n_name;
+  char **n_aliases;
+  int n_addrtype;
+  u32 n_net;
+};
+
 struct __sanitizer_addrinfo {
   int ai_flags;
   int ai_family;
@@ -773,6 +783,10 @@ extern int glob_altdirfunc;
 
 extern unsigned path_max;
 
+#  if !SANITIZER_ANDROID
+extern const int wordexp_wrde_dooffs;
+#  endif  // !SANITIZER_ANDROID
+
 struct __sanitizer_wordexp_t {
   uptr we_wordc;
   char **we_wordv;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.cpp
index 565b31f68aae..a113cb0d3490 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.cpp
@@ -123,6 +123,7 @@ namespace __sanitizer {
   unsigned struct_ElfW_Phdr_sz = sizeof(ElfW(Phdr));
 
   int glob_nomatch = GLOB_NOMATCH;
+  const int wordexp_wrde_dooffs = WRDE_DOOFFS;
 
   unsigned path_max = PATH_MAX;
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.h b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.h
index 85995e79792d..cbab577bcf26 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.h
@@ -341,6 +341,7 @@ struct __sanitizer_glob_t {
 
 extern int glob_nomatch;
 extern int glob_altdirfunc;
+extern const int wordexp_wrde_dooffs;
 
 extern unsigned path_max;
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_posix.h b/compiler-rt/lib/sanitizer_common/sanitizer_posix.h
index b65dae644767..f91e26e74b87 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_posix.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_posix.h
@@ -20,10 +20,7 @@
 #include "sanitizer_platform_limits_posix.h"
 #include "sanitizer_platform_limits_solaris.h"
 
-#if !SANITIZER_POSIX
-// Make it hard to accidentally use any of functions declared in this file:
-#error This file should only be included on POSIX
-#endif
+#if SANITIZER_POSIX
 
 namespace __sanitizer {
 
@@ -126,4 +123,6 @@ void DecorateMapping(uptr addr, uptr size, const char *name);
 
 }  // namespace __sanitizer
 
+#endif  // SANITIZER_POSIX
+
 #endif  // SANITIZER_POSIX_H
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_posix_libcdep.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_posix_libcdep.cpp
index ddf6844bed13..eed02ce4f6aa 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_posix_libcdep.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_posix_libcdep.cpp
@@ -151,6 +151,8 @@ int Atexit(void (*function)(void)) {
 #endif
 }
 
+bool CreateDir(const char *pathname) { return mkdir(pathname, 0755) == 0; }
+
 bool SupportsColoredOutput(fd_t fd) {
   return isatty(fd) != 0;
 }
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_printf.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_printf.cpp
index b913c92e16f1..3a9e366d2df9 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_printf.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_printf.cpp
@@ -20,10 +20,6 @@
 #include <stdio.h>
 #include <stdarg.h>
 
-#if defined(__x86_64__)
-#  include <emmintrin.h>
-#endif
-
 #if SANITIZER_WINDOWS && defined(_MSC_VER) && _MSC_VER < 1800 &&               \
       !defined(va_copy)
 # define va_copy(dst, src) ((dst) = (src))
@@ -132,8 +128,8 @@ static int AppendPointer(char **buff, const char *buff_end, u64 ptr_value) {
 int VSNPrintf(char *buff, int buff_length,
               const char *format, va_list args) {
   static const char *kPrintfFormatsHelp =
-      "Supported Printf formats: %([0-9]*)?(z|ll)?{d,u,x,X,V}; %p; "
-      "%[-]([0-9]*)?(\\.\\*)?s; %c\n";
+      "Supported Printf formats: %([0-9]*)?(z|l|ll)?{d,u,x,X}; %p; "
+      "%[-]([0-9]*)?(\\.\\*)?s; %c\nProvided format: ";
   RAW_CHECK(format);
   RAW_CHECK(buff_length > 0);
   const char *buff_end = &buff[buff_length - 1];
@@ -164,9 +160,11 @@ int VSNPrintf(char *buff, int buff_length,
     }
     bool have_z = (*cur == 'z');
     cur += have_z;
-    bool have_ll = !have_z && (cur[0] == 'l' && cur[1] == 'l');
+    bool have_l = cur[0] == 'l' && cur[1] != 'l';
+    cur += have_l;
+    bool have_ll = cur[0] == 'l' && cur[1] == 'l';
     cur += have_ll * 2;
-    const bool have_length = have_z || have_ll;
+    const bool have_length = have_z || have_l || have_ll;
     const bool have_flags = have_width || have_length;
     // At the moment only %s supports precision and left-justification.
     CHECK(!((precision >= 0 || left_justified) && *cur != 's'));
@@ -174,6 +172,7 @@ int VSNPrintf(char *buff, int buff_length,
       case 'd': {
         s64 dval = have_ll  ? va_arg(args, s64)
                    : have_z ? va_arg(args, sptr)
+                   : have_l ? va_arg(args, long)
                             : va_arg(args, int);
         result += AppendSignedDecimal(&buff, buff_end, dval, width,
                                       pad_with_zero);
@@ -184,26 +183,20 @@ int VSNPrintf(char *buff, int buff_length,
       case 'X': {
         u64 uval = have_ll  ? va_arg(args, u64)
                    : have_z ? va_arg(args, uptr)
+                   : have_l ? va_arg(args, unsigned long)
                             : va_arg(args, unsigned);
         bool uppercase = (*cur == 'X');
         result += AppendUnsigned(&buff, buff_end, uval, (*cur == 'u') ? 10 : 16,
                                  width, pad_with_zero, uppercase);
         break;
       }
-      case 'V': {
-        for (uptr i = 0; i < 16; i++) {
-          unsigned x = va_arg(args, unsigned);
-          result += AppendUnsigned(&buff, buff_end, x, 16, 2, true, false);
-        }
-        break;
-      }
       case 'p': {
-        RAW_CHECK_MSG(!have_flags, kPrintfFormatsHelp);
+        RAW_CHECK_VA(!have_flags, kPrintfFormatsHelp, format);
         result += AppendPointer(&buff, buff_end, va_arg(args, uptr));
         break;
       }
       case 's': {
-        RAW_CHECK_MSG(!have_length, kPrintfFormatsHelp);
+        RAW_CHECK_VA(!have_length, kPrintfFormatsHelp, format);
         // Only left-justified width is supported.
         CHECK(!have_width || left_justified);
         result += AppendString(&buff, buff_end, left_justified ? -width : width,
@@ -211,17 +204,17 @@ int VSNPrintf(char *buff, int buff_length,
         break;
       }
       case 'c': {
-        RAW_CHECK_MSG(!have_flags, kPrintfFormatsHelp);
+        RAW_CHECK_VA(!have_flags, kPrintfFormatsHelp, format);
         result += AppendChar(&buff, buff_end, va_arg(args, int));
         break;
       }
       case '%' : {
-        RAW_CHECK_MSG(!have_flags, kPrintfFormatsHelp);
+        RAW_CHECK_VA(!have_flags, kPrintfFormatsHelp, format);
         result += AppendChar(&buff, buff_end, '%');
         break;
       }
       default: {
-        RAW_CHECK_MSG(false, kPrintfFormatsHelp);
+        RAW_CHECK_VA(false, kPrintfFormatsHelp, format);
       }
     }
   }
@@ -317,7 +310,6 @@ static void NOINLINE SharedPrintfCode(bool append_pid, const char *format,
                            format, args);
 }
 
-FORMAT(1, 2)
 void Printf(const char *format, ...) {
   va_list args;
   va_start(args, format);
@@ -326,7 +318,6 @@ void Printf(const char *format, ...) {
 }
 
 // Like Printf, but prints the current PID before the output string.
-FORMAT(1, 2)
 void Report(const char *format, ...) {
   va_list args;
   va_start(args, format);
@@ -338,7 +329,6 @@ void Report(const char *format, ...) {
 // Returns the number of symbols that should have been written to buffer
 // (not including trailing '\0'). Thus, the string is truncated
 // iff return value is not less than "length".
-FORMAT(3, 4)
 int internal_snprintf(char *buffer, uptr length, const char *format, ...) {
   va_list args;
   va_start(args, format);
@@ -347,7 +337,6 @@ int internal_snprintf(char *buffer, uptr length, const char *format, ...) {
   return needed_length;
 }
 
-FORMAT(2, 3)
 void InternalScopedString::append(const char *format, ...) {
   uptr prev_len = length();
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_procmaps.h b/compiler-rt/lib/sanitizer_common/sanitizer_procmaps.h
index a56640db43e8..055af366ef06 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_procmaps.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_procmaps.h
@@ -65,13 +65,23 @@ class MemoryMappedSegment {
   MemoryMappedSegmentData *data_;
 };
 
-class MemoryMappingLayout {
+class MemoryMappingLayoutBase {
+ public:
+  virtual bool Next(MemoryMappedSegment *segment) { UNIMPLEMENTED(); }
+  virtual bool Error() const { UNIMPLEMENTED(); };
+  virtual void Reset() { UNIMPLEMENTED(); }
+
+ protected:
+  ~MemoryMappingLayoutBase() {}
+};
+
+class MemoryMappingLayout final : public MemoryMappingLayoutBase {
  public:
   explicit MemoryMappingLayout(bool cache_enabled);
   ~MemoryMappingLayout();
-  bool Next(MemoryMappedSegment *segment);
-  bool Error() const;
-  void Reset();
+  virtual bool Next(MemoryMappedSegment *segment) override;
+  virtual bool Error() const override;
+  virtual void Reset() override;
   // In some cases, e.g. when running under a sandbox on Linux, ASan is unable
   // to obtain the memory mappings. It should fall back to pre-cached data
   // instead of aborting.
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_common.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_common.cpp
index 1b7dd46d8de4..eb351b0f06fd 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_common.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_common.cpp
@@ -145,28 +145,44 @@ void MemoryMappingLayout::DumpListOfModules(
   }
 }
 
-void GetMemoryProfile(fill_profile_f cb, uptr *stats, uptr stats_size) {
+void GetMemoryProfile(fill_profile_f cb, uptr *stats) {
   char *smaps = nullptr;
   uptr smaps_cap = 0;
   uptr smaps_len = 0;
   if (!ReadFileToBuffer("/proc/self/smaps", &smaps, &smaps_cap, &smaps_len))
     return;
+  ParseUnixMemoryProfile(cb, stats, smaps, smaps_len);
+  UnmapOrDie(smaps, smaps_cap);
+}
+
+void ParseUnixMemoryProfile(fill_profile_f cb, uptr *stats, char *smaps,
+                            uptr smaps_len) {
   uptr start = 0;
   bool file = false;
   const char *pos = smaps;
-  while (pos < smaps + smaps_len) {
+  char *end = smaps + smaps_len;
+  if (smaps_len < 2)
+    return;
+  // The following parsing can crash on almost every line
+  // in the case of malformed/truncated input.
+  // Fixing that is hard b/c e.g. ParseDecimal does not
+  // even accept end of the buffer and assumes well-formed input.
+  // So instead we patch end of the input a bit,
+  // it does not affect well-formed complete inputs.
+  *--end = 0;
+  *--end = '\n';
+  while (pos < end) {
     if (IsHex(pos[0])) {
       start = ParseHex(&pos);
       for (; *pos != '/' && *pos > '\n'; pos++) {}
       file = *pos == '/';
     } else if (internal_strncmp(pos, "Rss:", 4) == 0) {
-      while (!IsDecimal(*pos)) pos++;
+      while (pos < end && !IsDecimal(*pos)) pos++;
       uptr rss = ParseDecimal(&pos) * 1024;
-      cb(start, rss, file, stats, stats_size);
+      cb(start, rss, file, stats);
     }
     while (*pos++ != '\n') {}
   }
-  UnmapOrDie(smaps, smaps_cap);
 }
 
 } // namespace __sanitizer
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_solaris.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_solaris.cpp
index bf813f235bb7..e16c4e938cb2 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_solaris.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_solaris.cpp
@@ -55,7 +55,15 @@ bool MemoryMappingLayout::Next(MemoryMappedSegment *segment) {
 
     internal_snprintf(proc_path, sizeof(proc_path), "/proc/self/path/%s",
                       xmapentry->pr_mapname);
-    internal_readlink(proc_path, segment->filename, segment->filename_size);
+    ssize_t sz = internal_readlink(proc_path, segment->filename,
+                                   segment->filename_size - 1);
+
+    // If readlink failed, the map is anonymous.
+    if (sz == -1) {
+      segment->filename[0] = '\0';
+    } else if ((size_t)sz < segment->filename_size)
+      // readlink doesn't NUL-terminate.
+      segment->filename[sz] = '\0';
   }
 
   data_.current += sizeof(prxmap_t);
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_signal_interceptors.inc b/compiler-rt/lib/sanitizer_common/sanitizer_signal_interceptors.inc
index cefb870f7e25..475e577d9982 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_signal_interceptors.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_signal_interceptors.inc
@@ -29,8 +29,16 @@ using namespace __sanitizer;
 #endif
 
 #ifndef SIGNAL_INTERCEPTOR_SIGACTION_IMPL
-#define SIGNAL_INTERCEPTOR_SIGACTION_IMPL(signum, act, oldact) \
-  { return REAL(sigaction_symname)(signum, act, oldact); }
+#  define SIGNAL_INTERCEPTOR_SIGACTION_IMPL(signum, act, oldact)              \
+    {                                                                         \
+      if (!REAL(sigaction_symname)) {                                         \
+        Printf(                                                               \
+            "Warning: REAL(sigaction_symname) == nullptr. This may happen "   \
+            "if you link with ubsan statically. Sigaction will not work.\n"); \
+        return -1;                                                            \
+      }                                                                       \
+      return REAL(sigaction_symname)(signum, act, oldact);                    \
+    }
 #endif
 
 #if SANITIZER_INTERCEPT_BSD_SIGNAL
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_solaris.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_solaris.cpp
index cb53eab8da15..62c40affc9ac 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_solaris.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_solaris.cpp
@@ -225,28 +225,6 @@ void FutexWait(atomic_uint32_t *p, u32 cmp) {
 
 void FutexWake(atomic_uint32_t *p, u32 count) {}
 
-BlockingMutex::BlockingMutex() {
-  CHECK(sizeof(mutex_t) <= sizeof(opaque_storage_));
-  internal_memset(this, 0, sizeof(*this));
-  CHECK_EQ(mutex_init((mutex_t *)&opaque_storage_, USYNC_THREAD, NULL), 0);
-}
-
-void BlockingMutex::Lock() {
-  CHECK(sizeof(mutex_t) <= sizeof(opaque_storage_));
-  CHECK_NE(owner_, (uptr)thr_self());
-  CHECK_EQ(mutex_lock((mutex_t *)&opaque_storage_), 0);
-  CHECK(!owner_);
-  owner_ = (uptr)thr_self();
-}
-
-void BlockingMutex::Unlock() {
-  CHECK(owner_ == (uptr)thr_self());
-  owner_ = 0;
-  CHECK_EQ(mutex_unlock((mutex_t *)&opaque_storage_), 0);
-}
-
-void BlockingMutex::CheckLocked() const { CHECK_EQ((uptr)thr_self(), owner_); }
-
 }  // namespace __sanitizer
 
 #endif  // SANITIZER_SOLARIS
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stack_store.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_stack_store.cpp
new file mode 100644
index 000000000000..ad88e2bbbefc
--- /dev/null
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stack_store.cpp
@@ -0,0 +1,91 @@
+//===-- sanitizer_stack_store.cpp -------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_stack_store.h"
+
+#include "sanitizer_atomic.h"
+#include "sanitizer_common.h"
+#include "sanitizer_stacktrace.h"
+
+namespace __sanitizer {
+
+static constexpr u32 kStackSizeBits = 16;
+
+StackStore::Id StackStore::Store(const StackTrace &trace) {
+  uptr *stack_trace = Alloc(trace.size + 1);
+  CHECK_LT(trace.size, 1 << kStackSizeBits);
+  *stack_trace = trace.size + (trace.tag << kStackSizeBits);
+  internal_memcpy(stack_trace + 1, trace.trace, trace.size * sizeof(uptr));
+  return reinterpret_cast<StackStore::Id>(stack_trace);
+}
+
+StackTrace StackStore::Load(Id id) {
+  const uptr *stack_trace = reinterpret_cast<const uptr *>(id);
+  uptr size = *stack_trace & ((1 << kStackSizeBits) - 1);
+  uptr tag = *stack_trace >> kStackSizeBits;
+  return StackTrace(stack_trace + 1, size, tag);
+}
+
+uptr *StackStore::TryAlloc(uptr count) {
+  // Optimisic lock-free allocation, essentially try to bump the region ptr.
+  for (;;) {
+    uptr cmp = atomic_load(&region_pos_, memory_order_acquire);
+    uptr end = atomic_load(&region_end_, memory_order_acquire);
+    uptr size = count * sizeof(uptr);
+    if (cmp == 0 || cmp + size > end)
+      return nullptr;
+    if (atomic_compare_exchange_weak(&region_pos_, &cmp, cmp + size,
+                                     memory_order_acquire))
+      return reinterpret_cast<uptr *>(cmp);
+  }
+}
+
+uptr *StackStore::Alloc(uptr count) {
+  // First, try to allocate optimisitically.
+  uptr *s = TryAlloc(count);
+  if (LIKELY(s))
+    return s;
+  return RefillAndAlloc(count);
+}
+
+uptr *StackStore::RefillAndAlloc(uptr count) {
+  // If failed, lock, retry and alloc new superblock.
+  SpinMutexLock l(&mtx_);
+  for (;;) {
+    uptr *s = TryAlloc(count);
+    if (s)
+      return s;
+    atomic_store(&region_pos_, 0, memory_order_relaxed);
+    uptr size = count * sizeof(uptr) + sizeof(BlockInfo);
+    uptr allocsz = RoundUpTo(Max<uptr>(size, 64u * 1024u), GetPageSizeCached());
+    uptr mem = (uptr)MmapOrDie(allocsz, "stack depot");
+    BlockInfo *new_block = (BlockInfo *)(mem + allocsz) - 1;
+    new_block->next = curr_;
+    new_block->ptr = mem;
+    new_block->size = allocsz;
+    curr_ = new_block;
+
+    atomic_fetch_add(&mapped_size_, allocsz, memory_order_relaxed);
+
+    allocsz -= sizeof(BlockInfo);
+    atomic_store(&region_end_, mem + allocsz, memory_order_release);
+    atomic_store(&region_pos_, mem, memory_order_release);
+  }
+}
+
+void StackStore::TestOnlyUnmap() {
+  while (curr_) {
+    uptr mem = curr_->ptr;
+    uptr allocsz = curr_->size;
+    curr_ = curr_->next;
+    UnmapOrDie((void *)mem, allocsz);
+  }
+  internal_memset(this, 0, sizeof(*this));
+}
+
+}  // namespace __sanitizer
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stack_store.h b/compiler-rt/lib/sanitizer_common/sanitizer_stack_store.h
new file mode 100644
index 000000000000..b5bbdccc20b1
--- /dev/null
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stack_store.h
@@ -0,0 +1,50 @@
+//===-- sanitizer_stack_store.h ---------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef SANITIZER_STACK_STORE_H
+#define SANITIZER_STACK_STORE_H
+
+#include "sanitizer_atomic.h"
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_mutex.h"
+#include "sanitizer_stacktrace.h"
+
+namespace __sanitizer {
+
+class StackStore {
+ public:
+  constexpr StackStore() = default;
+
+  using Id = uptr;
+
+  Id Store(const StackTrace &trace);
+  StackTrace Load(Id id);
+  uptr Allocated() const { return atomic_load_relaxed(&mapped_size_); }
+
+  void TestOnlyUnmap();
+
+ private:
+  uptr *Alloc(uptr count = 1);
+  uptr *TryAlloc(uptr count);
+  uptr *RefillAndAlloc(uptr count);
+  mutable StaticSpinMutex mtx_ = {};  // Protects alloc of new blocks.
+  atomic_uintptr_t region_pos_ = {};  // Region allocator for Node's.
+  atomic_uintptr_t region_end_ = {};
+  atomic_uintptr_t mapped_size_ = {};
+
+  struct BlockInfo {
+    const BlockInfo *next;
+    uptr ptr;
+    uptr size;
+  };
+  const BlockInfo *curr_ = nullptr;
+};
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_STACK_STORE_H
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stackdepot.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_stackdepot.cpp
index 44a95214e38b..e203b2cc4c89 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stackdepot.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stackdepot.cpp
@@ -14,93 +14,87 @@
 
 #include "sanitizer_common.h"
 #include "sanitizer_hash.h"
+#include "sanitizer_stack_store.h"
 #include "sanitizer_stackdepotbase.h"
 
 namespace __sanitizer {
 
 struct StackDepotNode {
-  StackDepotNode *link;
-  u32 id;
-  atomic_uint32_t hash_and_use_count; // hash_bits : 12; use_count : 20;
-  u32 size;
-  u32 tag;
-  uptr stack[1];  // [size]
+  using hash_type = u64;
+  hash_type stack_hash;
+  u32 link;
 
   static const u32 kTabSizeLog = SANITIZER_ANDROID ? 16 : 20;
-  // Lower kTabSizeLog bits are equal for all items in one bucket.
-  // We use these bits to store the per-stack use counter.
-  static const u32 kUseCountBits = kTabSizeLog;
-  static const u32 kMaxUseCount = 1 << kUseCountBits;
-  static const u32 kUseCountMask = (1 << kUseCountBits) - 1;
-  static const u32 kHashMask = ~kUseCountMask;
 
   typedef StackTrace args_type;
-  bool eq(u32 hash, const args_type &args) const {
-    u32 hash_bits =
-        atomic_load(&hash_and_use_count, memory_order_relaxed) & kHashMask;
-    if ((hash & kHashMask) != hash_bits || args.size != size || args.tag != tag)
-      return false;
-    uptr i = 0;
-    for (; i < size; i++) {
-      if (stack[i] != args.trace[i]) return false;
-    }
-    return true;
+  bool eq(hash_type hash, const args_type &args) const {
+    return hash == stack_hash;
   }
-  static uptr storage_size(const args_type &args) {
-    return sizeof(StackDepotNode) + (args.size - 1) * sizeof(uptr);
-  }
-  static u32 hash(const args_type &args) {
-    MurMur2HashBuilder H(args.size * sizeof(uptr));
+  static uptr allocated();
+  static hash_type hash(const args_type &args) {
+    MurMur2Hash64Builder H(args.size * sizeof(uptr));
     for (uptr i = 0; i < args.size; i++) H.add(args.trace[i]);
+    H.add(args.tag);
     return H.get();
   }
   static bool is_valid(const args_type &args) {
     return args.size > 0 && args.trace;
   }
-  void store(const args_type &args, u32 hash) {
-    atomic_store(&hash_and_use_count, hash & kHashMask, memory_order_relaxed);
-    size = args.size;
-    tag = args.tag;
-    internal_memcpy(stack, args.trace, size * sizeof(uptr));
-  }
-  args_type load() const {
-    return args_type(&stack[0], size, tag);
-  }
-  StackDepotHandle get_handle() { return StackDepotHandle(this); }
+  void store(u32 id, const args_type &args, hash_type hash);
+  args_type load(u32 id) const;
+  static StackDepotHandle get_handle(u32 id);
 
   typedef StackDepotHandle handle_type;
 };
 
-COMPILER_CHECK(StackDepotNode::kMaxUseCount == (u32)kStackDepotMaxUseCount);
-
-u32 StackDepotHandle::id() { return node_->id; }
-int StackDepotHandle::use_count() {
-  return atomic_load(&node_->hash_and_use_count, memory_order_relaxed) &
-         StackDepotNode::kUseCountMask;
-}
-void StackDepotHandle::inc_use_count_unsafe() {
-  u32 prev =
-      atomic_fetch_add(&node_->hash_and_use_count, 1, memory_order_relaxed) &
-      StackDepotNode::kUseCountMask;
-  CHECK_LT(prev + 1, StackDepotNode::kMaxUseCount);
-}
+static StackStore stackStore;
 
 // FIXME(dvyukov): this single reserved bit is used in TSan.
 typedef StackDepotBase<StackDepotNode, 1, StackDepotNode::kTabSizeLog>
     StackDepot;
 static StackDepot theDepot;
+// Keep rarely accessed stack traces out of frequently access nodes to improve
+// caching efficiency.
+static TwoLevelMap<StackStore::Id, StackDepot::kNodesSize1,
+                   StackDepot::kNodesSize2>
+    storeIds;
+// Keep mutable data out of frequently access nodes to improve caching
+// efficiency.
+static TwoLevelMap<atomic_uint32_t, StackDepot::kNodesSize1,
+                   StackDepot::kNodesSize2>
+    useCounts;
 
-StackDepotStats *StackDepotGetStats() {
-  return theDepot.GetStats();
+int StackDepotHandle::use_count() const {
+  return atomic_load_relaxed(&useCounts[id_]);
+}
+
+void StackDepotHandle::inc_use_count_unsafe() {
+  atomic_fetch_add(&useCounts[id_], 1, memory_order_relaxed);
 }
 
-u32 StackDepotPut(StackTrace stack) {
-  StackDepotHandle h = theDepot.Put(stack);
-  return h.valid() ? h.id() : 0;
+uptr StackDepotNode::allocated() {
+  return stackStore.Allocated() + storeIds.MemoryUsage() +
+         useCounts.MemoryUsage();
 }
 
+void StackDepotNode::store(u32 id, const args_type &args, hash_type hash) {
+  stack_hash = hash;
+  storeIds[id] = stackStore.Store(args);
+}
+
+StackDepotNode::args_type StackDepotNode::load(u32 id) const {
+  StackStore::Id store_id = storeIds[id];
+  if (!store_id)
+    return {};
+  return stackStore.Load(store_id);
+}
+
+StackDepotStats StackDepotGetStats() { return theDepot.GetStats(); }
+
+u32 StackDepotPut(StackTrace stack) { return theDepot.Put(stack); }
+
 StackDepotHandle StackDepotPut_WithHandle(StackTrace stack) {
-  return theDepot.Put(stack);
+  return StackDepotNode::get_handle(theDepot.Put(stack));
 }
 
 StackTrace StackDepotGet(u32 id) {
@@ -121,34 +115,14 @@ void StackDepotPrintAll() {
 #endif
 }
 
-bool StackDepotReverseMap::IdDescPair::IdComparator(
-    const StackDepotReverseMap::IdDescPair &a,
-    const StackDepotReverseMap::IdDescPair &b) {
-  return a.id < b.id;
-}
-
-StackDepotReverseMap::StackDepotReverseMap() {
-  map_.reserve(StackDepotGetStats()->n_uniq_ids + 100);
-  for (int idx = 0; idx < StackDepot::kTabSize; idx++) {
-    atomic_uintptr_t *p = &theDepot.tab[idx];
-    uptr v = atomic_load(p, memory_order_consume);
-    StackDepotNode *s = (StackDepotNode*)(v & ~1);
-    for (; s; s = s->link) {
-      IdDescPair pair = {s->id, s};
-      map_.push_back(pair);
-    }
-  }
-  Sort(map_.data(), map_.size(), &IdDescPair::IdComparator);
+StackDepotHandle StackDepotNode::get_handle(u32 id) {
+  return StackDepotHandle(&theDepot.nodes[id], id);
 }
 
-StackTrace StackDepotReverseMap::Get(u32 id) {
-  if (!map_.size())
-    return StackTrace();
-  IdDescPair pair = {id, nullptr};
-  uptr idx = InternalLowerBound(map_, pair, IdDescPair::IdComparator);
-  if (idx > map_.size() || map_[idx].id != id)
-    return StackTrace();
-  return map_[idx].desc->load();
+void StackDepotTestOnlyUnmap() {
+  theDepot.TestOnlyUnmap();
+  storeIds.TestOnlyUnmap();
+  stackStore.TestOnlyUnmap();
 }
 
 } // namespace __sanitizer
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stackdepot.h b/compiler-rt/lib/sanitizer_common/sanitizer_stackdepot.h
index 0e26c1fc37c4..56d655d9404c 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stackdepot.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stackdepot.h
@@ -22,18 +22,18 @@ namespace __sanitizer {
 // StackDepot efficiently stores huge amounts of stack traces.
 struct StackDepotNode;
 struct StackDepotHandle {
-  StackDepotNode *node_;
-  StackDepotHandle() : node_(nullptr) {}
-  explicit StackDepotHandle(StackDepotNode *node) : node_(node) {}
-  bool valid() { return node_; }
-  u32 id();
-  int use_count();
+  StackDepotNode *node_ = nullptr;
+  u32 id_ = 0;
+  StackDepotHandle(StackDepotNode *node, u32 id) : node_(node), id_(id) {}
+  bool valid() const { return node_; }
+  u32 id() const { return id_; }
+  int use_count() const;
   void inc_use_count_unsafe();
 };
 
 const int kStackDepotMaxUseCount = 1U << (SANITIZER_ANDROID ? 16 : 20);
 
-StackDepotStats *StackDepotGetStats();
+StackDepotStats StackDepotGetStats();
 u32 StackDepotPut(StackTrace stack);
 StackDepotHandle StackDepotPut_WithHandle(StackTrace stack);
 // Retrieves a stored stack trace by the id.
@@ -43,29 +43,7 @@ void StackDepotLockAll();
 void StackDepotUnlockAll();
 void StackDepotPrintAll();
 
-// Instantiating this class creates a snapshot of StackDepot which can be
-// efficiently queried with StackDepotGet(). You can use it concurrently with
-// StackDepot, but the snapshot is only guaranteed to contain those stack traces
-// which were stored before it was instantiated.
-class StackDepotReverseMap {
- public:
-  StackDepotReverseMap();
-  StackTrace Get(u32 id);
-
- private:
-  struct IdDescPair {
-    u32 id;
-    StackDepotNode *desc;
-
-    static bool IdComparator(const IdDescPair &a, const IdDescPair &b);
-  };
-
-  InternalMmapVector<IdDescPair> map_;
-
-  // Disallow evil constructors.
-  StackDepotReverseMap(const StackDepotReverseMap&);
-  void operator=(const StackDepotReverseMap&);
-};
+void StackDepotTestOnlyUnmap();
 
 } // namespace __sanitizer
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stackdepotbase.h b/compiler-rt/lib/sanitizer_common/sanitizer_stackdepotbase.h
index 1af2c1892eff..96d1ddc87fd0 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stackdepotbase.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stackdepotbase.h
@@ -16,71 +16,87 @@
 #include <stdio.h>
 
 #include "sanitizer_atomic.h"
+#include "sanitizer_flat_map.h"
 #include "sanitizer_internal_defs.h"
 #include "sanitizer_mutex.h"
-#include "sanitizer_persistent_allocator.h"
 
 namespace __sanitizer {
 
 template <class Node, int kReservedBits, int kTabSizeLog>
 class StackDepotBase {
+  static constexpr u32 kIdSizeLog =
+      sizeof(u32) * 8 - Max(kReservedBits, 1 /* At least 1 bit for locking. */);
+  static constexpr u32 kNodesSize1Log = kIdSizeLog / 2;
+  static constexpr u32 kNodesSize2Log = kIdSizeLog - kNodesSize1Log;
+  static constexpr int kTabSize = 1 << kTabSizeLog;  // Hash table size.
+  static constexpr u32 kUnlockMask = (1ull << kIdSizeLog) - 1;
+  static constexpr u32 kLockMask = ~kUnlockMask;
+
  public:
   typedef typename Node::args_type args_type;
   typedef typename Node::handle_type handle_type;
+  typedef typename Node::hash_type hash_type;
+
+  static constexpr u64 kNodesSize1 = 1ull << kNodesSize1Log;
+  static constexpr u64 kNodesSize2 = 1ull << kNodesSize2Log;
+
   // Maps stack trace to an unique id.
-  handle_type Put(args_type args, bool *inserted = nullptr);
+  u32 Put(args_type args, bool *inserted = nullptr);
   // Retrieves a stored stack trace by the id.
   args_type Get(u32 id);
 
-  StackDepotStats *GetStats() { return &stats; }
+  StackDepotStats GetStats() const {
+    return {
+        atomic_load_relaxed(&n_uniq_ids),
+        nodes.MemoryUsage() + Node::allocated(),
+    };
+  }
 
   void LockAll();
   void UnlockAll();
   void PrintAll();
 
- private:
-  static Node *find(Node *s, args_type args, u32 hash);
-  static Node *lock(atomic_uintptr_t *p);
-  static void unlock(atomic_uintptr_t *p, Node *s);
+  void TestOnlyUnmap() {
+    nodes.TestOnlyUnmap();
+    internal_memset(this, 0, sizeof(*this));
+  }
 
-  static const int kTabSize = 1 << kTabSizeLog;  // Hash table size.
-  static const int kPartBits = 8;
-  static const int kPartShift = sizeof(u32) * 8 - kPartBits - kReservedBits;
-  static const int kPartCount =
-      1 << kPartBits;  // Number of subparts in the table.
-  static const int kPartSize = kTabSize / kPartCount;
-  static const int kMaxId = 1 << kPartShift;
+ private:
+  friend Node;
+  u32 find(u32 s, args_type args, hash_type hash) const;
+  static u32 lock(atomic_uint32_t *p);
+  static void unlock(atomic_uint32_t *p, u32 s);
+  atomic_uint32_t tab[kTabSize];  // Hash table of Node's.
 
-  atomic_uintptr_t tab[kTabSize];   // Hash table of Node's.
-  atomic_uint32_t seq[kPartCount];  // Unique id generators.
+  atomic_uint32_t n_uniq_ids;
 
-  StackDepotStats stats;
+  TwoLevelMap<Node, kNodesSize1, kNodesSize2> nodes;
 
   friend class StackDepotReverseMap;
 };
 
 template <class Node, int kReservedBits, int kTabSizeLog>
-Node *StackDepotBase<Node, kReservedBits, kTabSizeLog>::find(Node *s,
-                                                             args_type args,
-                                                             u32 hash) {
+u32 StackDepotBase<Node, kReservedBits, kTabSizeLog>::find(
+    u32 s, args_type args, hash_type hash) const {
   // Searches linked list s for the stack, returns its id.
-  for (; s; s = s->link) {
-    if (s->eq(hash, args)) {
+  for (; s;) {
+    const Node &node = nodes[s];
+    if (node.eq(hash, args))
       return s;
-    }
+    s = node.link;
   }
-  return nullptr;
+  return 0;
 }
 
 template <class Node, int kReservedBits, int kTabSizeLog>
-Node *StackDepotBase<Node, kReservedBits, kTabSizeLog>::lock(
-    atomic_uintptr_t *p) {
+u32 StackDepotBase<Node, kReservedBits, kTabSizeLog>::lock(atomic_uint32_t *p) {
   // Uses the pointer lsb as mutex.
   for (int i = 0;; i++) {
-    uptr cmp = atomic_load(p, memory_order_relaxed);
-    if ((cmp & 1) == 0 &&
-        atomic_compare_exchange_weak(p, &cmp, cmp | 1, memory_order_acquire))
-      return (Node *)cmp;
+    u32 cmp = atomic_load(p, memory_order_relaxed);
+    if ((cmp & kLockMask) == 0 &&
+        atomic_compare_exchange_weak(p, &cmp, cmp | kLockMask,
+                                     memory_order_acquire))
+      return cmp;
     if (i < 10)
       proc_yield(10);
     else
@@ -90,73 +106,57 @@ Node *StackDepotBase<Node, kReservedBits, kTabSizeLog>::lock(
 
 template <class Node, int kReservedBits, int kTabSizeLog>
 void StackDepotBase<Node, kReservedBits, kTabSizeLog>::unlock(
-    atomic_uintptr_t *p, Node *s) {
-  DCHECK_EQ((uptr)s & 1, 0);
-  atomic_store(p, (uptr)s, memory_order_release);
+    atomic_uint32_t *p, u32 s) {
+  DCHECK_EQ(s & kLockMask, 0);
+  atomic_store(p, s, memory_order_release);
 }
 
 template <class Node, int kReservedBits, int kTabSizeLog>
-typename StackDepotBase<Node, kReservedBits, kTabSizeLog>::handle_type
-StackDepotBase<Node, kReservedBits, kTabSizeLog>::Put(args_type args,
-                                                      bool *inserted) {
-  if (inserted) *inserted = false;
-  if (!Node::is_valid(args)) return handle_type();
-  uptr h = Node::hash(args);
-  atomic_uintptr_t *p = &tab[h % kTabSize];
-  uptr v = atomic_load(p, memory_order_consume);
-  Node *s = (Node *)(v & ~1);
+u32 StackDepotBase<Node, kReservedBits, kTabSizeLog>::Put(args_type args,
+                                                          bool *inserted) {
+  if (inserted)
+    *inserted = false;
+  if (!LIKELY(Node::is_valid(args)))
+    return 0;
+  hash_type h = Node::hash(args);
+  atomic_uint32_t *p = &tab[h % kTabSize];
+  u32 v = atomic_load(p, memory_order_consume);
+  u32 s = v & kUnlockMask;
   // First, try to find the existing stack.
-  Node *node = find(s, args, h);
-  if (node) return node->get_handle();
+  u32 node = find(s, args, h);
+  if (LIKELY(node))
+    return node;
+
   // If failed, lock, retry and insert new.
-  Node *s2 = lock(p);
+  u32 s2 = lock(p);
   if (s2 != s) {
     node = find(s2, args, h);
     if (node) {
       unlock(p, s2);
-      return node->get_handle();
+      return node;
     }
   }
-  uptr part = (h % kTabSize) / kPartSize;
-  u32 id = atomic_fetch_add(&seq[part], 1, memory_order_relaxed) + 1;
-  stats.n_uniq_ids++;
-  CHECK_LT(id, kMaxId);
-  id |= part << kPartShift;
-  CHECK_NE(id, 0);
-  CHECK_EQ(id & (((u32)-1) >> kReservedBits), id);
-  uptr memsz = Node::storage_size(args);
-  s = (Node *)PersistentAlloc(memsz);
-  stats.allocated += memsz;
-  s->id = id;
-  s->store(args, h);
-  s->link = s2;
+  s = atomic_fetch_add(&n_uniq_ids, 1, memory_order_relaxed) + 1;
+  CHECK_EQ(s & kUnlockMask, s);
+  CHECK_EQ(s & (((u32)-1) >> kReservedBits), s);
+  Node &new_node = nodes[s];
+  new_node.store(s, args, h);
+  new_node.link = s2;
   unlock(p, s);
   if (inserted) *inserted = true;
-  return s->get_handle();
+  return s;
 }
 
 template <class Node, int kReservedBits, int kTabSizeLog>
 typename StackDepotBase<Node, kReservedBits, kTabSizeLog>::args_type
 StackDepotBase<Node, kReservedBits, kTabSizeLog>::Get(u32 id) {
-  if (id == 0) {
+  if (id == 0)
     return args_type();
-  }
   CHECK_EQ(id & (((u32)-1) >> kReservedBits), id);
-  // High kPartBits contain part id, so we need to scan at most kPartSize lists.
-  uptr part = id >> kPartShift;
-  for (int i = 0; i != kPartSize; i++) {
-    uptr idx = part * kPartSize + i;
-    CHECK_LT(idx, kTabSize);
-    atomic_uintptr_t *p = &tab[idx];
-    uptr v = atomic_load(p, memory_order_consume);
-    Node *s = (Node *)(v & ~1);
-    for (; s; s = s->link) {
-      if (s->id == id) {
-        return s->load();
-      }
-    }
-  }
-  return args_type();
+  if (!nodes.contains(id))
+    return args_type();
+  const Node &node = nodes[id];
+  return node.load(id);
 }
 
 template <class Node, int kReservedBits, int kTabSizeLog>
@@ -169,24 +169,23 @@ void StackDepotBase<Node, kReservedBits, kTabSizeLog>::LockAll() {
 template <class Node, int kReservedBits, int kTabSizeLog>
 void StackDepotBase<Node, kReservedBits, kTabSizeLog>::UnlockAll() {
   for (int i = 0; i < kTabSize; ++i) {
-    atomic_uintptr_t *p = &tab[i];
+    atomic_uint32_t *p = &tab[i];
     uptr s = atomic_load(p, memory_order_relaxed);
-    unlock(p, (Node *)(s & ~1UL));
+    unlock(p, s & kUnlockMask);
   }
 }
 
 template <class Node, int kReservedBits, int kTabSizeLog>
 void StackDepotBase<Node, kReservedBits, kTabSizeLog>::PrintAll() {
   for (int i = 0; i < kTabSize; ++i) {
-    atomic_uintptr_t *p = &tab[i];
-    lock(p);
-    uptr v = atomic_load(p, memory_order_relaxed);
-    Node *s = (Node *)(v & ~1UL);
-    for (; s; s = s->link) {
-      Printf("Stack for id %u:\n", s->id);
-      s->load().Print();
+    atomic_uint32_t *p = &tab[i];
+    u32 s = atomic_load(p, memory_order_consume) & kUnlockMask;
+    for (; s;) {
+      const Node &node = nodes[s];
+      Printf("Stack for id %u:\n", s);
+      node.load(s).Print();
+      s = node.link;
     }
-    unlock(p, s);
   }
 }
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.cpp
index 07e4409f4a5d..37e9e6dd08d7 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.cpp
@@ -22,8 +22,9 @@ namespace __sanitizer {
 uptr StackTrace::GetNextInstructionPc(uptr pc) {
 #if defined(__sparc__) || defined(__mips__)
   return pc + 8;
-#elif defined(__powerpc__) || defined(__arm__) || defined(__aarch64__)
-  return pc + 4;
+#elif defined(__powerpc__) || defined(__arm__) || defined(__aarch64__) || \
+    defined(__hexagon__)
+  return STRIP_PAC_PC((void *)pc) + 4;
 #elif SANITIZER_RISCV64
   // Current check order is 4 -> 2 -> 6 -> 8
   u8 InsnByte = *(u8 *)(pc);
@@ -64,7 +65,7 @@ void BufferedStackTrace::Init(const uptr *pcs, uptr cnt, uptr extra_top_pc) {
   top_frame_bp = 0;
 }
 
-// Sparc implemention is in its own file.
+// Sparc implementation is in its own file.
 #if !defined(__sparc__)
 
 // In GCC on ARM bp points to saved lr, not fp, so we should check the next
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.h b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.h
index ea330f36f7d7..11c6154b09ea 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.h
@@ -209,11 +209,11 @@ static inline bool IsValidFrame(uptr frame, uptr stack_top, uptr stack_bottom) {
 // StackTrace::GetCurrentPc() faster.
 #if defined(__x86_64__)
 #  define GET_CURRENT_PC()                \
-    ({                                    \
+    (__extension__({                      \
       uptr pc;                            \
       asm("lea 0(%%rip), %0" : "=r"(pc)); \
       pc;                                 \
-    })
+    }))
 #else
 #  define GET_CURRENT_PC() StackTrace::GetCurrentPc()
 #endif
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_libcdep.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_libcdep.cpp
index f60ea7731748..2d1c03f73221 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_libcdep.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_libcdep.cpp
@@ -64,7 +64,7 @@ class StackTraceTextPrinter {
       if (dedup_token_->length())
         dedup_token_->append("--");
       if (stack->info.function != nullptr)
-        dedup_token_->append(stack->info.function);
+        dedup_token_->append("%s", stack->info.function);
     }
   }
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_printer.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_printer.cpp
index c998322d3944..c6356dae23c1 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_printer.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_printer.cpp
@@ -129,7 +129,7 @@ void RenderFrame(InternalScopedString *buffer, const char *format, int frame_no,
       break;
     // Frame number and all fields of AddressInfo structure.
     case 'n':
-      buffer->append("%zu", frame_no);
+      buffer->append("%u", frame_no);
       break;
     case 'p':
       buffer->append("0x%zx", address);
@@ -198,8 +198,8 @@ void RenderFrame(InternalScopedString *buffer, const char *format, int frame_no,
       }
       break;
     default:
-      Report("Unsupported specifier in stack frame format: %c (0x%zx)!\n", *p,
-             *p);
+      Report("Unsupported specifier in stack frame format: %c (%p)!\n", *p,
+             (void *)p);
       Die();
     }
   }
@@ -244,14 +244,14 @@ void RenderData(InternalScopedString *buffer, const char *format,
         buffer->append("%s", StripPathPrefix(DI->file, strip_path_prefix));
         break;
       case 'l':
-        buffer->append("%d", DI->line);
+        buffer->append("%zu", DI->line);
         break;
       case 'g':
         buffer->append("%s", DI->name);
         break;
       default:
-        Report("Unsupported specifier in stack frame format: %c (0x%zx)!\n", *p,
-               *p);
+        Report("Unsupported specifier in stack frame format: %c (%p)!\n", *p,
+               (void *)p);
         Die();
     }
   }
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_sparc.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_sparc.cpp
index 34190fb1bbb2..1e635a66978f 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_sparc.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_sparc.cpp
@@ -9,7 +9,7 @@
 // This file is shared between AddressSanitizer and ThreadSanitizer
 // run-time libraries.
 //
-// Implemention of fast stack unwinding for Sparc.
+// Implementation of fast stack unwinding for Sparc.
 //===----------------------------------------------------------------------===//
 
 #if defined(__sparc__)
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cpp
index 53cfddcfbe0b..403bda1174cc 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cpp
@@ -108,7 +108,7 @@ struct TracerThreadArgument {
   void *callback_argument;
   // The tracer thread waits on this mutex while the parent finishes its
   // preparations.
-  BlockingMutex mutex;
+  Mutex mutex;
   // Tracer thread signals its completion by setting done.
   atomic_uintptr_t done;
   uptr parent_pid;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_netbsd_libcdep.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_netbsd_libcdep.cpp
index 9c7cd64255e5..701db72619a3 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_netbsd_libcdep.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_netbsd_libcdep.cpp
@@ -68,7 +68,7 @@ class SuspendedThreadsListNetBSD final : public SuspendedThreadsList {
 struct TracerThreadArgument {
   StopTheWorldCallback callback;
   void *callback_argument;
-  BlockingMutex mutex;
+  Mutex mutex;
   atomic_uintptr_t done;
   uptr parent_pid;
 };
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.h b/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.h
index 2476b0ea7bf7..42bd157fa627 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.h
@@ -158,7 +158,7 @@ class Symbolizer final {
   // its method should be protected by |mu_|.
   class ModuleNameOwner {
    public:
-    explicit ModuleNameOwner(BlockingMutex *synchronized_by)
+    explicit ModuleNameOwner(Mutex *synchronized_by)
         : last_match_(nullptr), mu_(synchronized_by) {
       storage_.reserve(kInitialCapacity);
     }
@@ -169,7 +169,7 @@ class Symbolizer final {
     InternalMmapVector<const char*> storage_;
     const char *last_match_;
 
-    BlockingMutex *mu_;
+    Mutex *mu_;
   } module_names_;
 
   /// Platform-specific function for creating a Symbolizer object.
@@ -192,7 +192,7 @@ class Symbolizer final {
   // Mutex locked from public methods of |Symbolizer|, so that the internals
   // (including individual symbolizer tools and platform-specific methods) are
   // always synchronized.
-  BlockingMutex mu_;
+  Mutex mu_;
 
   IntrusiveList<SymbolizerTool> tools_;
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_internal.h b/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_internal.h
index 71de1758b3e9..b8670941a05e 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_internal.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_internal.h
@@ -21,7 +21,7 @@ namespace __sanitizer {
 
 // Parsing helpers, 'str' is searched for delimiter(s) and a string or uptr
 // is extracted. When extracting a string, a newly allocated (using
-// InternalAlloc) and null-terminataed buffer is returned. They return a pointer
+// InternalAlloc) and null-terminated buffer is returned. They return a pointer
 // to the next characted after the found delimiter.
 const char *ExtractToken(const char *str, const char *delims, char **result);
 const char *ExtractInt(const char *str, const char *delims, int *result);
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libcdep.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libcdep.cpp
index 98418b426c37..3fc994fd3deb 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libcdep.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libcdep.cpp
@@ -83,7 +83,7 @@ const char *ExtractTokenUpToDelimiter(const char *str, const char *delimiter,
 }
 
 SymbolizedStack *Symbolizer::SymbolizePC(uptr addr) {
-  BlockingMutexLock l(&mu_);
+  Lock l(&mu_);
   const char *module_name = nullptr;
   uptr module_offset;
   ModuleArch arch;
@@ -103,7 +103,7 @@ SymbolizedStack *Symbolizer::SymbolizePC(uptr addr) {
 }
 
 bool Symbolizer::SymbolizeData(uptr addr, DataInfo *info) {
-  BlockingMutexLock l(&mu_);
+  Lock l(&mu_);
   const char *module_name = nullptr;
   uptr module_offset;
   ModuleArch arch;
@@ -124,7 +124,7 @@ bool Symbolizer::SymbolizeData(uptr addr, DataInfo *info) {
 }
 
 bool Symbolizer::SymbolizeFrame(uptr addr, FrameInfo *info) {
-  BlockingMutexLock l(&mu_);
+  Lock l(&mu_);
   const char *module_name = nullptr;
   if (!FindModuleNameAndOffsetForAddress(
           addr, &module_name, &info->module_offset, &info->module_arch))
@@ -141,7 +141,7 @@ bool Symbolizer::SymbolizeFrame(uptr addr, FrameInfo *info) {
 
 bool Symbolizer::GetModuleNameAndOffsetForPC(uptr pc, const char **module_name,
                                              uptr *module_address) {
-  BlockingMutexLock l(&mu_);
+  Lock l(&mu_);
   const char *internal_module_name = nullptr;
   ModuleArch arch;
   if (!FindModuleNameAndOffsetForAddress(pc, &internal_module_name,
@@ -154,7 +154,7 @@ bool Symbolizer::GetModuleNameAndOffsetForPC(uptr pc, const char **module_name,
 }
 
 void Symbolizer::Flush() {
-  BlockingMutexLock l(&mu_);
+  Lock l(&mu_);
   for (auto &tool : tools_) {
     SymbolizerScope sym_scope(this);
     tool.Flush();
@@ -162,7 +162,7 @@ void Symbolizer::Flush() {
 }
 
 const char *Symbolizer::Demangle(const char *name) {
-  BlockingMutexLock l(&mu_);
+  Lock l(&mu_);
   for (auto &tool : tools_) {
     SymbolizerScope sym_scope(this);
     if (const char *demangled = tool.Demangle(name))
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_report.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_report.cpp
index f330ed36640a..869c8935330d 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_report.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_report.cpp
@@ -88,11 +88,17 @@ void ReportErrorSummary(const char *error_type, const StackTrace *stack,
 #endif
 }
 
-void ReportMmapWriteExec(int prot) {
+void ReportMmapWriteExec(int prot, int flags) {
 #if SANITIZER_POSIX && (!SANITIZER_GO && !SANITIZER_ANDROID)
-  if ((prot & (PROT_WRITE | PROT_EXEC)) != (PROT_WRITE | PROT_EXEC))
+  int pflags = (PROT_WRITE | PROT_EXEC);
+  if ((prot & pflags) != pflags)
     return;
 
+#  if SANITIZER_MAC && defined(MAP_JIT)
+  if ((flags & MAP_JIT) == MAP_JIT)
+    return;
+#  endif
+
   ScopedErrorReportLock l;
   SanitizerCommonDecorator d;
 
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_hexagon.inc b/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_hexagon.inc
new file mode 100644
index 000000000000..553bff7503b4
--- /dev/null
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_hexagon.inc
@@ -0,0 +1,131 @@
+//===-- sanitizer_syscall_linux_hexagon.inc ---------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// Implementations of internal_syscall and internal_iserror for Linux/hexagon.
+//
+//===----------------------------------------------------------------------===//
+
+#define SYSCALL(name) __NR_##name
+
+#define __internal_syscall_LL_E(x) \
+  ((union {                        \
+    long long ll;                  \
+    long l[2];                     \
+  }){.ll = x})                     \
+      .l[0],                       \
+      ((union {                    \
+        long long ll;              \
+        long l[2];                 \
+      }){.ll = x})                 \
+          .l[1]
+#define __internal_syscall_LL_O(x) 0, __SYSCALL_LL_E((x))
+
+#define __asm_syscall(...)                                                 \
+  do {                                                                     \
+    __asm__ __volatile__("trap0(#1)" : "=r"(r0) : __VA_ARGS__ : "memory"); \
+    return r0;                                                             \
+  } while (0)
+
+#define __internal_syscall0(n) (__internal_syscall)(n)
+
+static uptr __internal_syscall(long n) {
+  register u32 r6 __asm__("r6") = n;
+  register u32 r0 __asm__("r0");
+  __asm_syscall("r"(r6));
+}
+
+#define __internal_syscall1(n, a1) (__internal_syscall)(n, (long)(a1))
+
+static uptr __internal_syscall(long n, long a) {
+  register u32 r6 __asm__("r6") = n;
+  register u32 r0 __asm__("r0") = a;
+  __asm_syscall("r"(r6), "0"(r0));
+}
+
+#define __internal_syscall2(n, a1, a2) \
+  (__internal_syscall)(n, (long)(a1), (long)(a2))
+
+static uptr __internal_syscall(long n, long a, long b) {
+  register u32 r6 __asm__("r6") = n;
+  register u32 r0 __asm__("r0") = a;
+  register u32 r1 __asm__("r1") = b;
+  __asm_syscall("r"(r6), "0"(r0), "r"(r1));
+}
+
+#define __internal_syscall3(n, a1, a2, a3) \
+  (__internal_syscall)(n, (long)(a1), (long)(a2), (long)(a3))
+
+static uptr __internal_syscall(long n, long a, long b, long c) {
+  register u32 r6 __asm__("r6") = n;
+  register u32 r0 __asm__("r0") = a;
+  register u32 r1 __asm__("r1") = b;
+  register u32 r2 __asm__("r2") = c;
+  __asm_syscall("r"(r6), "0"(r0), "r"(r1), "r"(r2));
+}
+
+#define __internal_syscall4(n, a1, a2, a3, a4) \
+  (__internal_syscall)(n, (long)(a1), (long)(a2), (long)(a3), (long)(a4))
+
+static uptr __internal_syscall(long n, long a, long b, long c, long d) {
+  register u32 r6 __asm__("r6") = n;
+  register u32 r0 __asm__("r0") = a;
+  register u32 r1 __asm__("r1") = b;
+  register u32 r2 __asm__("r2") = c;
+  register u32 r3 __asm__("r3") = d;
+  __asm_syscall("r"(r6), "0"(r0), "r"(r1), "r"(r2), "r"(r3));
+}
+
+#define __internal_syscall5(n, a1, a2, a3, a4, a5)                        \
+  (__internal_syscall)(n, (long)(a1), (long)(a2), (long)(a3), (long)(a4), \
+                       (long)(a5))
+
+static uptr __internal_syscall(long n, long a, long b, long c, long d, long e) {
+  register u32 r6 __asm__("r6") = n;
+  register u32 r0 __asm__("r0") = a;
+  register u32 r1 __asm__("r1") = b;
+  register u32 r2 __asm__("r2") = c;
+  register u32 r3 __asm__("r3") = d;
+  register u32 r4 __asm__("r4") = e;
+  __asm_syscall("r"(r6), "0"(r0), "r"(r1), "r"(r2), "r"(r3), "r"(r4));
+}
+
+#define __internal_syscall6(n, a1, a2, a3, a4, a5, a6)                    \
+  (__internal_syscall)(n, (long)(a1), (long)(a2), (long)(a3), (long)(a4), \
+                       (long)(a5), (long)(a6))
+
+static uptr __internal_syscall(long n, long a, long b, long c, long d, long e,
+                               long f) {
+  register u32 r6 __asm__("r6") = n;
+  register u32 r0 __asm__("r0") = a;
+  register u32 r1 __asm__("r1") = b;
+  register u32 r2 __asm__("r2") = c;
+  register u32 r3 __asm__("r3") = d;
+  register u32 r4 __asm__("r4") = e;
+  register u32 r5 __asm__("r5") = f;
+  __asm_syscall("r"(r6), "0"(r0), "r"(r1), "r"(r2), "r"(r3), "r"(r4), "r"(r5));
+}
+
+#define __SYSCALL_NARGS_X(a1, a2, a3, a4, a5, a6, a7, a8, n, ...) n
+#define __SYSCALL_NARGS(...) \
+  __SYSCALL_NARGS_X(__VA_ARGS__, 7, 6, 5, 4, 3, 2, 1, 0, )
+#define __SYSCALL_CONCAT_X(a, b) a##b
+#define __SYSCALL_CONCAT(a, b) __SYSCALL_CONCAT_X(a, b)
+#define __SYSCALL_DISP(b, ...) \
+  __SYSCALL_CONCAT(b, __SYSCALL_NARGS(__VA_ARGS__))(__VA_ARGS__)
+
+#define internal_syscall(...) __SYSCALL_DISP(__internal_syscall, __VA_ARGS__)
+
+// Helper function used to avoid clobbering of errno.
+bool internal_iserror(uptr retval, int *rverrno) {
+  if (retval >= (uptr)-4095) {
+    if (rverrno)
+      *rverrno = -retval;
+    return true;
+  }
+  return false;
+}
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.cpp
index 745fbf76b01f..2e1cd0238812 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.cpp
@@ -13,6 +13,8 @@
 
 #include "sanitizer_thread_registry.h"
 
+#include "sanitizer_placement_new.h"
+
 namespace __sanitizer {
 
 ThreadContextBase::ThreadContextBase(u32 tid)
@@ -119,7 +121,7 @@ ThreadRegistry::ThreadRegistry(ThreadContextFactory factory, u32 max_threads,
 
 void ThreadRegistry::GetNumberOfThreads(uptr *total, uptr *running,
                                         uptr *alive) {
-  BlockingMutexLock l(&mtx_);
+  ThreadRegistryLock l(this);
   if (total)
     *total = threads_.size();
   if (running) *running = running_threads_;
@@ -127,13 +129,13 @@ void ThreadRegistry::GetNumberOfThreads(uptr *total, uptr *running,
 }
 
 uptr ThreadRegistry::GetMaxAliveThreads() {
-  BlockingMutexLock l(&mtx_);
+  ThreadRegistryLock l(this);
   return max_alive_threads_;
 }
 
 u32 ThreadRegistry::CreateThread(uptr user_id, bool detached, u32 parent_tid,
                                  void *arg) {
-  BlockingMutexLock l(&mtx_);
+  ThreadRegistryLock l(this);
   u32 tid = kInvalidTid;
   ThreadContextBase *tctx = QuarantinePop();
   if (tctx) {
@@ -162,6 +164,12 @@ u32 ThreadRegistry::CreateThread(uptr user_id, bool detached, u32 parent_tid,
     max_alive_threads_++;
     CHECK_EQ(alive_threads_, max_alive_threads_);
   }
+  if (user_id) {
+    // Ensure that user_id is unique. If it's not the case we are screwed.
+    // Ignoring this situation may lead to very hard to debug false
+    // positives later (e.g. if we join a wrong thread).
+    CHECK(live_.try_emplace(user_id, tid).second);
+  }
   tctx->SetCreated(user_id, total_threads_++, detached,
                    parent_tid, arg);
   return tid;
@@ -179,7 +187,7 @@ void ThreadRegistry::RunCallbackForEachThreadLocked(ThreadCallback cb,
 }
 
 u32 ThreadRegistry::FindThread(FindThreadCallback cb, void *arg) {
-  BlockingMutexLock l(&mtx_);
+  ThreadRegistryLock l(this);
   for (u32 tid = 0; tid < threads_.size(); tid++) {
     ThreadContextBase *tctx = threads_[tid];
     if (tctx != 0 && cb(tctx, arg))
@@ -211,7 +219,7 @@ ThreadContextBase *ThreadRegistry::FindThreadContextByOsIDLocked(tid_t os_id) {
 }
 
 void ThreadRegistry::SetThreadName(u32 tid, const char *name) {
-  BlockingMutexLock l(&mtx_);
+  ThreadRegistryLock l(this);
   ThreadContextBase *tctx = threads_[tid];
   CHECK_NE(tctx, 0);
   CHECK_EQ(SANITIZER_FUCHSIA ? ThreadStatusCreated : ThreadStatusRunning,
@@ -220,19 +228,13 @@ void ThreadRegistry::SetThreadName(u32 tid, const char *name) {
 }
 
 void ThreadRegistry::SetThreadNameByUserId(uptr user_id, const char *name) {
-  BlockingMutexLock l(&mtx_);
-  for (u32 tid = 0; tid < threads_.size(); tid++) {
-    ThreadContextBase *tctx = threads_[tid];
-    if (tctx != 0 && tctx->user_id == user_id &&
-        tctx->status != ThreadStatusInvalid) {
-      tctx->SetName(name);
-      return;
-    }
-  }
+  ThreadRegistryLock l(this);
+  if (const auto *tid = live_.find(user_id))
+    threads_[tid->second]->SetName(name);
 }
 
 void ThreadRegistry::DetachThread(u32 tid, void *arg) {
-  BlockingMutexLock l(&mtx_);
+  ThreadRegistryLock l(this);
   ThreadContextBase *tctx = threads_[tid];
   CHECK_NE(tctx, 0);
   if (tctx->status == ThreadStatusInvalid) {
@@ -241,6 +243,8 @@ void ThreadRegistry::DetachThread(u32 tid, void *arg) {
   }
   tctx->OnDetached(arg);
   if (tctx->status == ThreadStatusFinished) {
+    if (tctx->user_id)
+      live_.erase(tctx->user_id);
     tctx->SetDead();
     QuarantinePush(tctx);
   } else {
@@ -252,7 +256,7 @@ void ThreadRegistry::JoinThread(u32 tid, void *arg) {
   bool destroyed = false;
   do {
     {
-      BlockingMutexLock l(&mtx_);
+      ThreadRegistryLock l(this);
       ThreadContextBase *tctx = threads_[tid];
       CHECK_NE(tctx, 0);
       if (tctx->status == ThreadStatusInvalid) {
@@ -260,6 +264,8 @@ void ThreadRegistry::JoinThread(u32 tid, void *arg) {
         return;
       }
       if ((destroyed = tctx->GetDestroyed())) {
+        if (tctx->user_id)
+          live_.erase(tctx->user_id);
         tctx->SetJoined(arg);
         QuarantinePush(tctx);
       }
@@ -275,7 +281,7 @@ void ThreadRegistry::JoinThread(u32 tid, void *arg) {
 // thread before trying to create it, and then failed to actually
 // create it, and so never called StartThread.
 ThreadStatus ThreadRegistry::FinishThread(u32 tid) {
-  BlockingMutexLock l(&mtx_);
+  ThreadRegistryLock l(this);
   CHECK_GT(alive_threads_, 0);
   alive_threads_--;
   ThreadContextBase *tctx = threads_[tid];
@@ -292,6 +298,8 @@ ThreadStatus ThreadRegistry::FinishThread(u32 tid) {
   }
   tctx->SetFinished();
   if (dead) {
+    if (tctx->user_id)
+      live_.erase(tctx->user_id);
     tctx->SetDead();
     QuarantinePush(tctx);
   }
@@ -301,7 +309,7 @@ ThreadStatus ThreadRegistry::FinishThread(u32 tid) {
 
 void ThreadRegistry::StartThread(u32 tid, tid_t os_id, ThreadType thread_type,
                                  void *arg) {
-  BlockingMutexLock l(&mtx_);
+  ThreadRegistryLock l(this);
   running_threads_++;
   ThreadContextBase *tctx = threads_[tid];
   CHECK_NE(tctx, 0);
@@ -333,14 +341,28 @@ ThreadContextBase *ThreadRegistry::QuarantinePop() {
   return tctx;
 }
 
+u32 ThreadRegistry::ConsumeThreadUserId(uptr user_id) {
+  ThreadRegistryLock l(this);
+  u32 tid;
+  auto *t = live_.find(user_id);
+  CHECK(t);
+  tid = t->second;
+  live_.erase(t);
+  auto *tctx = threads_[tid];
+  CHECK_EQ(tctx->user_id, user_id);
+  tctx->user_id = 0;
+  return tid;
+}
+
 void ThreadRegistry::SetThreadUserId(u32 tid, uptr user_id) {
-  BlockingMutexLock l(&mtx_);
+  ThreadRegistryLock l(this);
   ThreadContextBase *tctx = threads_[tid];
   CHECK_NE(tctx, 0);
   CHECK_NE(tctx->status, ThreadStatusInvalid);
   CHECK_NE(tctx->status, ThreadStatusDead);
   CHECK_EQ(tctx->user_id, 0);
   tctx->user_id = user_id;
+  CHECK(live_.try_emplace(user_id, tctx->tid).second);
 }
 
 }  // namespace __sanitizer
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.h b/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.h
index 0b28bbe6ddf6..a259b324220f 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.h
@@ -15,6 +15,7 @@
 #define SANITIZER_THREAD_REGISTRY_H
 
 #include "sanitizer_common.h"
+#include "sanitizer_dense_map.h"
 #include "sanitizer_list.h"
 #include "sanitizer_mutex.h"
 
@@ -127,6 +128,7 @@ class MUTEX ThreadRegistry {
   // Finishes thread and returns previous status.
   ThreadStatus FinishThread(u32 tid);
   void StartThread(u32 tid, tid_t os_id, ThreadType thread_type, void *arg);
+  u32 ConsumeThreadUserId(uptr user_id);
   void SetThreadUserId(u32 tid, uptr user_id);
 
  private:
@@ -135,7 +137,7 @@ class MUTEX ThreadRegistry {
   const u32 thread_quarantine_size_;
   const u32 max_reuse_;
 
-  BlockingMutex mtx_;
+  Mutex mtx_;
 
   u64 total_threads_;   // Total number of created threads. May be greater than
                         // max_threads_ if contexts were reused.
@@ -146,6 +148,7 @@ class MUTEX ThreadRegistry {
   InternalMmapVector<ThreadContextBase *> threads_;
   IntrusiveList<ThreadContextBase> dead_threads_;
   IntrusiveList<ThreadContextBase> invalid_threads_;
+  DenseMap<uptr, Tid> live_;
 
   void QuarantinePush(ThreadContextBase *tctx);
   ThreadContextBase *QuarantinePop();
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_tls_get_addr.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_tls_get_addr.cpp
index 1f664b6cf5b8..b13e2dc9e332 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_tls_get_addr.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_tls_get_addr.cpp
@@ -44,7 +44,7 @@ static atomic_uintptr_t number_of_live_dtls;
 static const uptr kDestroyedThread = -1;
 
 static void DTLS_Deallocate(DTLS::DTVBlock *block) {
-  VReport(2, "__tls_get_addr: DTLS_Deallocate %p %zd\n", block);
+  VReport(2, "__tls_get_addr: DTLS_Deallocate %p\n", (void *)block);
   UnmapOrDie(block, sizeof(DTLS::DTVBlock));
   atomic_fetch_sub(&number_of_live_dtls, 1, memory_order_relaxed);
 }
@@ -66,12 +66,13 @@ static DTLS::DTVBlock *DTLS_NextBlock(atomic_uintptr_t *cur) {
   }
   uptr num_live_dtls =
       atomic_fetch_add(&number_of_live_dtls, 1, memory_order_relaxed);
-  VReport(2, "__tls_get_addr: DTLS_NextBlock %p %zd\n", &dtls, num_live_dtls);
+  VReport(2, "__tls_get_addr: DTLS_NextBlock %p %zd\n", (void *)&dtls,
+          num_live_dtls);
   return new_dtv;
 }
 
 static DTLS::DTV *DTLS_Find(uptr id) {
-  VReport(2, "__tls_get_addr: DTLS_Find %p %zd\n", &dtls, id);
+  VReport(2, "__tls_get_addr: DTLS_Find %p %zd\n", (void *)&dtls, id);
   static constexpr uptr kPerBlock = ARRAY_SIZE(DTLS::DTVBlock::dtvs);
   DTLS::DTVBlock *cur = DTLS_NextBlock(&dtls.dtv_block);
   if (!cur)
@@ -82,7 +83,7 @@ static DTLS::DTV *DTLS_Find(uptr id) {
 
 void DTLS_Destroy() {
   if (!common_flags()->intercept_tls_get_addr) return;
-  VReport(2, "__tls_get_addr: DTLS_Destroy %p\n", &dtls);
+  VReport(2, "__tls_get_addr: DTLS_Destroy %p\n", (void *)&dtls);
   DTLS::DTVBlock *block = (DTLS::DTVBlock *)atomic_exchange(
       &dtls.dtv_block, kDestroyedThread, memory_order_release);
   while (block) {
@@ -117,26 +118,27 @@ DTLS::DTV *DTLS_on_tls_get_addr(void *arg_void, void *res,
     return 0;
   uptr tls_size = 0;
   uptr tls_beg = reinterpret_cast<uptr>(res) - arg->offset - kDtvOffset;
-  VReport(2, "__tls_get_addr: %p {%p,%p} => %p; tls_beg: %p; sp: %p "
-             "num_live_dtls %zd\n",
-          arg, arg->dso_id, arg->offset, res, tls_beg, &tls_beg,
+  VReport(2,
+          "__tls_get_addr: %p {0x%zx,0x%zx} => %p; tls_beg: 0x%zx; sp: %p "
+          "num_live_dtls %zd\n",
+          (void *)arg, arg->dso_id, arg->offset, res, tls_beg, (void *)&tls_beg,
           atomic_load(&number_of_live_dtls, memory_order_relaxed));
   if (dtls.last_memalign_ptr == tls_beg) {
     tls_size = dtls.last_memalign_size;
-    VReport(2, "__tls_get_addr: glibc <=2.18 suspected; tls={%p,%p}\n",
-        tls_beg, tls_size);
+    VReport(2, "__tls_get_addr: glibc <=2.18 suspected; tls={0x%zx,0x%zx}\n",
+            tls_beg, tls_size);
   } else if (tls_beg >= static_tls_begin && tls_beg < static_tls_end) {
     // This is the static TLS block which was initialized / unpoisoned at thread
     // creation.
-    VReport(2, "__tls_get_addr: static tls: %p\n", tls_beg);
+    VReport(2, "__tls_get_addr: static tls: 0x%zx\n", tls_beg);
     tls_size = 0;
   } else if ((tls_beg % 4096) == sizeof(Glibc_2_19_tls_header)) {
     // We may want to check gnu_get_libc_version().
     Glibc_2_19_tls_header *header = (Glibc_2_19_tls_header *)tls_beg - 1;
     tls_size = header->size;
     tls_beg = header->start;
-    VReport(2, "__tls_get_addr: glibc >=2.19 suspected; tls={%p %p}\n",
-        tls_beg, tls_size);
+    VReport(2, "__tls_get_addr: glibc >=2.19 suspected; tls={0x%zx 0x%zx}\n",
+            tls_beg, tls_size);
   } else {
     VReport(2, "__tls_get_addr: Can't guess glibc version\n");
     // This may happen inside the DTOR of main thread, so just ignore it.
@@ -149,7 +151,7 @@ DTLS::DTV *DTLS_on_tls_get_addr(void *arg_void, void *res,
 
 void DTLS_on_libc_memalign(void *ptr, uptr size) {
   if (!common_flags()->intercept_tls_get_addr) return;
-  VReport(2, "DTLS_on_libc_memalign: %p %p\n", ptr, size);
+  VReport(2, "DTLS_on_libc_memalign: %p 0x%zx\n", ptr, size);
   dtls.last_memalign_ptr = reinterpret_cast<uptr>(ptr);
   dtls.last_memalign_size = size;
 }
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_type_traits.h b/compiler-rt/lib/sanitizer_common/sanitizer_type_traits.h
index 2a58d9874d2c..06a44d1b5c7a 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_type_traits.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_type_traits.h
@@ -13,6 +13,8 @@
 #ifndef SANITIZER_TYPE_TRAITS_H
 #define SANITIZER_TYPE_TRAITS_H
 
+#include "sanitizer_common/sanitizer_internal_defs.h"
+
 namespace __sanitizer {
 
 struct true_type {
@@ -57,6 +59,83 @@ struct conditional<false, T, F> {
   using type = F;
 };
 
+template <class T>
+struct remove_reference {
+  using type = T;
+};
+template <class T>
+struct remove_reference<T&> {
+  using type = T;
+};
+template <class T>
+struct remove_reference<T&&> {
+  using type = T;
+};
+
+template <class T>
+WARN_UNUSED_RESULT inline typename remove_reference<T>::type&& move(T&& t) {
+  return static_cast<typename remove_reference<T>::type&&>(t);
+}
+
+template <class T>
+WARN_UNUSED_RESULT inline constexpr T&& forward(
+    typename remove_reference<T>::type& t) {
+  return static_cast<T&&>(t);
+}
+
+template <class T>
+WARN_UNUSED_RESULT inline constexpr T&& forward(
+    typename remove_reference<T>::type&& t) {
+  return static_cast<T&&>(t);
+}
+
+template <class T, T v>
+struct integral_constant {
+  static constexpr const T value = v;
+  typedef T value_type;
+  typedef integral_constant type;
+  constexpr operator value_type() const { return value; }
+  constexpr value_type operator()() const { return value; }
+};
+
+#ifndef __has_builtin
+#  define __has_builtin(x) 0
+#endif
+
+#if __has_builtin(__is_trivially_destructible)
+
+template <class T>
+struct is_trivially_destructible
+    : public integral_constant<bool, __is_trivially_destructible(T)> {};
+
+#elif __has_builtin(__has_trivial_destructor)
+
+template <class T>
+struct is_trivially_destructible
+    : public integral_constant<bool, __has_trivial_destructor(T)> {};
+
+#else
+
+template <class T>
+struct is_trivially_destructible
+    : public integral_constant<bool, /* less efficient fallback */ false> {};
+
+#endif
+
+#if __has_builtin(__is_trivially_copyable)
+
+template <class T>
+struct is_trivially_copyable
+    : public integral_constant<bool, __is_trivially_copyable(T)> {};
+
+#else
+
+template <class T>
+struct is_trivially_copyable
+    : public integral_constant<bool, /* less efficient fallback */ false> {};
+
+#endif
+
 }  // namespace __sanitizer
 
 #endif
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_win.cpp b/compiler-rt/lib/sanitizer_common/sanitizer_win.cpp
index dddd885a45dd..c3607dbed23e 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_win.cpp
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_win.cpp
@@ -16,6 +16,7 @@
 
 #define WIN32_LEAN_AND_MEAN
 #define NOGDI
+#include <direct.h>
 #include <windows.h>
 #include <io.h>
 #include <psapi.h>
@@ -565,6 +566,8 @@ void Abort() {
   internal__exit(3);
 }
 
+bool CreateDir(const char *pathname) { return _mkdir(pathname) == 0; }
+
 #if !SANITIZER_GO
 // Read the file to extract the ImageBase field from the PE header. If ASLR is
 // disabled and this virtual address is available, the loader will typically
@@ -827,27 +830,6 @@ void FutexWake(atomic_uint32_t *p, u32 count) {
     WakeByAddressAll(p);
 }
 
-// ---------------------- BlockingMutex ---------------- {{{1
-
-BlockingMutex::BlockingMutex() {
-  CHECK(sizeof(SRWLOCK) <= sizeof(opaque_storage_));
-  internal_memset(this, 0, sizeof(*this));
-}
-
-void BlockingMutex::Lock() {
-  AcquireSRWLockExclusive((PSRWLOCK)opaque_storage_);
-  CHECK_EQ(owner_, 0);
-  owner_ = GetThreadSelf();
-}
-
-void BlockingMutex::Unlock() {
-  CheckLocked();
-  owner_ = 0;
-  ReleaseSRWLockExclusive((PSRWLOCK)opaque_storage_);
-}
-
-void BlockingMutex::CheckLocked() const { CHECK_EQ(owner_, GetThreadSelf()); }
-
 uptr GetTlsSize() {
   return 0;
 }
@@ -1131,7 +1113,7 @@ bool IsProcessRunning(pid_t pid) {
 int WaitForProcess(pid_t pid) { return -1; }
 
 // FIXME implement on this platform.
-void GetMemoryProfile(fill_profile_f cb, uptr *stats, uptr stats_size) { }
+void GetMemoryProfile(fill_profile_f cb, uptr *stats) {}
 
 void CheckNoDeepBind(const char *filename, int flag) {
   // Do nothing.
diff --git a/compiler-rt/lib/sanitizer_common/symbolizer/scripts/ar_to_bc.sh b/compiler-rt/lib/sanitizer_common/symbolizer/scripts/ar_to_bc.sh
index 5c77bea83294..fa05d61a7c6d 100755
--- a/compiler-rt/lib/sanitizer_common/symbolizer/scripts/ar_to_bc.sh
+++ b/compiler-rt/lib/sanitizer_common/symbolizer/scripts/ar_to_bc.sh
@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 function usage() {
   echo "Usage: $0 INPUT... OUTPUT"
diff --git a/compiler-rt/lib/sanitizer_common/symbolizer/scripts/build_symbolizer.sh b/compiler-rt/lib/sanitizer_common/symbolizer/scripts/build_symbolizer.sh
index c793875db099..d1d61fb7ab2a 100755
--- a/compiler-rt/lib/sanitizer_common/symbolizer/scripts/build_symbolizer.sh
+++ b/compiler-rt/lib/sanitizer_common/symbolizer/scripts/build_symbolizer.sh
@@ -1,4 +1,4 @@
-#!/bin/bash -eu
+#!/usr/bin/env bash
 #
 # Run as: CLANG=bin/clang ZLIB_SRC=src/zlib \
 #             build_symbolizer.sh runtime_build/lib/clang/4.0.0/lib/linux/
@@ -71,6 +71,7 @@ TBLGEN=$CLANG_DIR/llvm-tblgen
 OPT=$CLANG_DIR/opt
 export AR=$CLANG_DIR/llvm-ar
 export LINK=$CLANG_DIR/llvm-link
+TARGET_TRIPLE=$($CC -print-target-triple)
 
 for F in $CC $CXX $TBLGEN $LINK $OPT $AR; do
   if [[ ! -x "$F" ]]; then
@@ -123,7 +124,7 @@ cd ${LIBCXX_BUILD}
 ninja cxx cxxabi
 
 FLAGS="${FLAGS} -fno-rtti -fno-exceptions"
-LLVM_FLAGS="${FLAGS} -nostdinc++ -I${ZLIB_BUILD} -I${LIBCXX_BUILD}/include/c++/v1 -Wno-error=global-constructors"
+LLVM_FLAGS="${FLAGS} -nostdinc++ -I${ZLIB_BUILD} -isystem ${LIBCXX_BUILD}/include/${TARGET_TRIPLE}/c++/v1 -isystem ${LIBCXX_BUILD}/include/c++/v1 -Wno-error=global-constructors"
 
 # Build LLVM.
 if [[ ! -d ${LLVM_BUILD} ]]; then
@@ -156,9 +157,11 @@ $AR rc symbolizer.a sanitizer_symbolize.o sanitizer_wrappers.o
 
 SYMBOLIZER_API_LIST=__sanitizer_symbolize_code,__sanitizer_symbolize_data,__sanitizer_symbolize_flush,__sanitizer_symbolize_demangle
 
+LIBCXX_ARCHIVE_DIR=$(dirname $(find $LIBCXX_BUILD -name libc++.a | head -n1))
+
 # Merge all the object files together and copy the resulting library back.
-$SCRIPT_DIR/ar_to_bc.sh $LIBCXX_BUILD/lib/libc++.a \
-                        $LIBCXX_BUILD/lib/libc++abi.a \
+$SCRIPT_DIR/ar_to_bc.sh $LIBCXX_ARCHIVE_DIR/libc++.a \
+                        $LIBCXX_ARCHIVE_DIR/libc++abi.a \
                         $LLVM_BUILD/lib/libLLVMSymbolize.a \
                         $LLVM_BUILD/lib/libLLVMObject.a \
                         $LLVM_BUILD/lib/libLLVMBinaryFormat.a \
diff --git a/compiler-rt/lib/scudo/scudo_utils.cpp b/compiler-rt/lib/scudo/scudo_utils.cpp
index b7ce8f915817..b0aef752c679 100644
--- a/compiler-rt/lib/scudo/scudo_utils.cpp
+++ b/compiler-rt/lib/scudo/scudo_utils.cpp
@@ -39,7 +39,7 @@ extern int VSNPrintf(char *buff, int buff_length, const char *format,
 
 namespace __scudo {
 
-FORMAT(1, 2) void NORETURN dieWithMessage(const char *Format, ...) {
+void dieWithMessage(const char *Format, ...) {
   static const char ScudoError[] = "Scudo ERROR: ";
   static constexpr uptr PrefixSize = sizeof(ScudoError) - 1;
   // Our messages are tiny, 256 characters is more than enough.
diff --git a/compiler-rt/lib/scudo/scudo_utils.h b/compiler-rt/lib/scudo/scudo_utils.h
index b657c69d9baf..5a9b32f0b234 100644
--- a/compiler-rt/lib/scudo/scudo_utils.h
+++ b/compiler-rt/lib/scudo/scudo_utils.h
@@ -27,7 +27,7 @@ inline Dest bit_cast(const Source& source) {
   return dest;
 }
 
-void NORETURN dieWithMessage(const char *Format, ...);
+void dieWithMessage(const char *Format, ...) NORETURN FORMAT(1, 2);
 
 bool hasHardwareCRC32();
 
diff --git a/compiler-rt/lib/scudo/standalone/combined.h b/compiler-rt/lib/scudo/standalone/combined.h
index fd5360ce0f55..371fb783a06e 100644
--- a/compiler-rt/lib/scudo/standalone/combined.h
+++ b/compiler-rt/lib/scudo/standalone/combined.h
@@ -205,6 +205,16 @@ public:
 #endif // GWP_ASAN_HOOKS
   }
 
+#ifdef GWP_ASAN_HOOKS
+  const gwp_asan::AllocationMetadata *getGwpAsanAllocationMetadata() {
+    return GuardedAlloc.getMetadataRegion();
+  }
+
+  const gwp_asan::AllocatorState *getGwpAsanAllocatorState() {
+    return GuardedAlloc.getAllocatorState();
+  }
+#endif // GWP_ASAN_HOOKS
+
   ALWAYS_INLINE void initThreadMaybe(bool MinimalInit = false) {
     TSDRegistry.initThreadMaybe(this, MinimalInit);
   }
@@ -910,7 +920,7 @@ public:
     if (!Depot->find(Hash, &RingPos, &Size))
       return;
     for (unsigned I = 0; I != Size && I != MaxTraceSize; ++I)
-      Trace[I] = (*Depot)[RingPos + I];
+      Trace[I] = static_cast<uintptr_t>((*Depot)[RingPos + I]);
   }
 
   static void getErrorInfo(struct scudo_error_info *ErrorInfo,
diff --git a/compiler-rt/lib/scudo/standalone/internal_defs.h b/compiler-rt/lib/scudo/standalone/internal_defs.h
index c9ffad136b78..621fc9c45e95 100644
--- a/compiler-rt/lib/scudo/standalone/internal_defs.h
+++ b/compiler-rt/lib/scudo/standalone/internal_defs.h
@@ -78,16 +78,16 @@
 
 namespace scudo {
 
-typedef unsigned long uptr;
-typedef unsigned char u8;
-typedef unsigned short u16;
-typedef unsigned int u32;
-typedef unsigned long long u64;
-typedef signed long sptr;
-typedef signed char s8;
-typedef signed short s16;
-typedef signed int s32;
-typedef signed long long s64;
+typedef uintptr_t uptr;
+typedef uint8_t u8;
+typedef uint16_t u16;
+typedef uint32_t u32;
+typedef uint64_t u64;
+typedef intptr_t sptr;
+typedef int8_t s8;
+typedef int16_t s16;
+typedef int32_t s32;
+typedef int64_t s64;
 
 // The following two functions have platform specific implementations.
 void outputRaw(const char *Buffer);
diff --git a/compiler-rt/lib/scudo/standalone/memtag.h b/compiler-rt/lib/scudo/standalone/memtag.h
index c48e228fbe44..df346bce1bd4 100644
--- a/compiler-rt/lib/scudo/standalone/memtag.h
+++ b/compiler-rt/lib/scudo/standalone/memtag.h
@@ -23,7 +23,9 @@ namespace scudo {
 // We assume that Top-Byte Ignore is enabled if the architecture supports memory
 // tagging. Not all operating systems enable TBI, so we only claim architectural
 // support for memory tagging if the operating system enables TBI.
-#if SCUDO_LINUX && !defined(SCUDO_DISABLE_TBI)
+// HWASan uses the top byte for its own purpose and Scudo should not touch it.
+#if SCUDO_LINUX && !defined(SCUDO_DISABLE_TBI) &&                              \
+    !__has_feature(hwaddress_sanitizer)
 inline constexpr bool archSupportsMemoryTagging() { return true; }
 #else
 inline constexpr bool archSupportsMemoryTagging() { return false; }
@@ -91,9 +93,10 @@ inline bool systemDetectsMemoryTagFaultsTestOnly() {
 #ifndef PR_MTE_TCF_MASK
 #define PR_MTE_TCF_MASK (3UL << PR_MTE_TCF_SHIFT)
 #endif
-  return (static_cast<unsigned long>(
-              prctl(PR_GET_TAGGED_ADDR_CTRL, 0, 0, 0, 0)) &
-          PR_MTE_TCF_MASK) != PR_MTE_TCF_NONE;
+  int res = prctl(PR_GET_TAGGED_ADDR_CTRL, 0, 0, 0, 0);
+  if (res == -1)
+    return false;
+  return (static_cast<unsigned long>(res) & PR_MTE_TCF_MASK) != PR_MTE_TCF_NONE;
 }
 
 inline void enableSystemMemoryTaggingTestOnly() {
diff --git a/compiler-rt/lib/scudo/standalone/primary64.h b/compiler-rt/lib/scudo/standalone/primary64.h
index 13420bf3d222..6c1785512c65 100644
--- a/compiler-rt/lib/scudo/standalone/primary64.h
+++ b/compiler-rt/lib/scudo/standalone/primary64.h
@@ -164,9 +164,9 @@ public:
       PoppedBlocks += Region->Stats.PoppedBlocks;
       PushedBlocks += Region->Stats.PushedBlocks;
     }
-    Str->append("Stats: SizeClassAllocator64: %zuM mapped (%zuM rss) in %zu "
+    Str->append("Stats: SizeClassAllocator64: %zuM mapped (%uM rss) in %zu "
                 "allocations; remains %zu\n",
-                TotalMapped >> 20, 0, PoppedBlocks,
+                TotalMapped >> 20, 0U, PoppedBlocks,
                 PoppedBlocks - PushedBlocks);
 
     for (uptr I = 0; I < NumClasses; I++)
diff --git a/compiler-rt/lib/scudo/standalone/secondary.h b/compiler-rt/lib/scudo/standalone/secondary.h
index 630e64d46edf..abb58a2882af 100644
--- a/compiler-rt/lib/scudo/standalone/secondary.h
+++ b/compiler-rt/lib/scudo/standalone/secondary.h
@@ -485,7 +485,7 @@ void *MapAllocator<Config>::allocate(Options Options, uptr Size, uptr Alignment,
                                      FillContentsMode FillContents) {
   if (Options.get(OptionBit::AddLargeAllocationSlack))
     Size += 1UL << SCUDO_MIN_ALIGNMENT_LOG;
-  Alignment = Max(Alignment, 1UL << SCUDO_MIN_ALIGNMENT_LOG);
+  Alignment = Max(Alignment, uptr(1U) << SCUDO_MIN_ALIGNMENT_LOG);
   const uptr PageSize = getPageSizeCached();
   uptr RoundedSize =
       roundUpTo(roundUpTo(Size, Alignment) + LargeBlock::getHeaderSize() +
@@ -602,12 +602,11 @@ void MapAllocator<Config>::deallocate(Options Options, void *Ptr) {
 
 template <typename Config>
 void MapAllocator<Config>::getStats(ScopedString *Str) const {
-  Str->append(
-      "Stats: MapAllocator: allocated %zu times (%zuK), freed %zu times "
-      "(%zuK), remains %zu (%zuK) max %zuM\n",
-      NumberOfAllocs, AllocatedBytes >> 10, NumberOfFrees, FreedBytes >> 10,
-      NumberOfAllocs - NumberOfFrees, (AllocatedBytes - FreedBytes) >> 10,
-      LargestSize >> 20);
+  Str->append("Stats: MapAllocator: allocated %u times (%zuK), freed %u times "
+              "(%zuK), remains %u (%zuK) max %zuM\n",
+              NumberOfAllocs, AllocatedBytes >> 10, NumberOfFrees,
+              FreedBytes >> 10, NumberOfAllocs - NumberOfFrees,
+              (AllocatedBytes - FreedBytes) >> 10, LargestSize >> 20);
 }
 
 } // namespace scudo
diff --git a/compiler-rt/lib/scudo/standalone/size_class_map.h b/compiler-rt/lib/scudo/standalone/size_class_map.h
index ba0f78453bcb..28b16d976e5e 100644
--- a/compiler-rt/lib/scudo/standalone/size_class_map.h
+++ b/compiler-rt/lib/scudo/standalone/size_class_map.h
@@ -335,8 +335,8 @@ template <typename SCMap> inline void printMap() {
     const uptr L = S ? getMostSignificantSetBitIndex(S) : 0;
     const uptr Cached = SCMap::getMaxCachedHint(S) * S;
     Buffer.append(
-        "C%02zu => S: %zu diff: +%zu %02zu%% L %zu Cached: %zu %zu; id %zu\n",
-        I, S, D, P, L, SCMap::getMaxCachedHint(S), Cached,
+        "C%02zu => S: %zu diff: +%zu %02zu%% L %zu Cached: %u %zu; id %zu\n", I,
+        S, D, P, L, SCMap::getMaxCachedHint(S), Cached,
         SCMap::getClassIdBySize(S));
     TotalCached += Cached;
     PrevS = S;
diff --git a/compiler-rt/lib/scudo/standalone/string_utils.cpp b/compiler-rt/lib/scudo/standalone/string_utils.cpp
index acf85889fcff..13fdb9c6ca6c 100644
--- a/compiler-rt/lib/scudo/standalone/string_utils.cpp
+++ b/compiler-rt/lib/scudo/standalone/string_utils.cpp
@@ -236,7 +236,6 @@ void ScopedString::append(const char *Format, va_list Args) {
   va_end(ArgsCopy);
 }
 
-FORMAT(2, 3)
 void ScopedString::append(const char *Format, ...) {
   va_list Args;
   va_start(Args, Format);
@@ -244,7 +243,6 @@ void ScopedString::append(const char *Format, ...) {
   va_end(Args);
 }
 
-FORMAT(1, 2)
 void Printf(const char *Format, ...) {
   va_list Args;
   va_start(Args, Format);
diff --git a/compiler-rt/lib/scudo/standalone/string_utils.h b/compiler-rt/lib/scudo/standalone/string_utils.h
index 06d23d42246d..dd6ff7893b83 100644
--- a/compiler-rt/lib/scudo/standalone/string_utils.h
+++ b/compiler-rt/lib/scudo/standalone/string_utils.h
@@ -26,15 +26,16 @@ public:
     String.push_back('\0');
   }
   void append(const char *Format, va_list Args);
-  void append(const char *Format, ...);
+  void append(const char *Format, ...) FORMAT(2, 3);
   void output() const { outputRaw(String.data()); }
 
 private:
   Vector<char> String;
 };
 
-int formatString(char *Buffer, uptr BufferLength, const char *Format, ...);
-void Printf(const char *Format, ...);
+int formatString(char *Buffer, uptr BufferLength, const char *Format, ...)
+    FORMAT(3, 4);
+void Printf(const char *Format, ...) FORMAT(1, 2);
 
 } // namespace scudo
 
diff --git a/compiler-rt/lib/scudo/standalone/vector.h b/compiler-rt/lib/scudo/standalone/vector.h
index 2c9a6e2aa655..eae774b56e28 100644
--- a/compiler-rt/lib/scudo/standalone/vector.h
+++ b/compiler-rt/lib/scudo/standalone/vector.h
@@ -19,13 +19,14 @@ namespace scudo {
 // small vectors. The current implementation supports only POD types.
 template <typename T> class VectorNoCtor {
 public:
-  void init(uptr InitialCapacity = 0) {
-    Data = reinterpret_cast<T *>(&LocalData[0]);
+  constexpr void init(uptr InitialCapacity = 0) {
+    Data = &LocalData[0];
     CapacityBytes = sizeof(LocalData);
-    reserve(InitialCapacity);
+    if (InitialCapacity > capacity())
+      reserve(InitialCapacity);
   }
   void destroy() {
-    if (Data != reinterpret_cast<T *>(&LocalData[0]))
+    if (Data != &LocalData[0])
       unmap(Data, CapacityBytes);
   }
   T &operator[](uptr I) {
@@ -55,7 +56,7 @@ public:
   uptr size() const { return Size; }
   const T *data() const { return Data; }
   T *data() { return Data; }
-  uptr capacity() const { return CapacityBytes / sizeof(T); }
+  constexpr uptr capacity() const { return CapacityBytes / sizeof(T); }
   void reserve(uptr NewSize) {
     // Never downsize internal buffer.
     if (NewSize > capacity())
@@ -91,14 +92,14 @@ private:
   }
 
   T *Data = nullptr;
-  u8 LocalData[256] = {};
+  T LocalData[256 / sizeof(T)] = {};
   uptr CapacityBytes = 0;
   uptr Size = 0;
 };
 
 template <typename T> class Vector : public VectorNoCtor<T> {
 public:
-  Vector() { VectorNoCtor<T>::init(); }
+  constexpr Vector() { VectorNoCtor<T>::init(); }
   explicit Vector(uptr Count) {
     VectorNoCtor<T>::init(Count);
     this->resize(Count);
diff --git a/compiler-rt/lib/scudo/standalone/wrappers_c.h b/compiler-rt/lib/scudo/standalone/wrappers_c.h
index 6d0cecdc4b41..5f7f51f3cedf 100644
--- a/compiler-rt/lib/scudo/standalone/wrappers_c.h
+++ b/compiler-rt/lib/scudo/standalone/wrappers_c.h
@@ -32,6 +32,19 @@ struct __scudo_mallinfo {
   __scudo_mallinfo_data_t keepcost;
 };
 
+struct __scudo_mallinfo2 {
+  size_t arena;
+  size_t ordblks;
+  size_t smblks;
+  size_t hblks;
+  size_t hblkhd;
+  size_t usmblks;
+  size_t fsmblks;
+  size_t uordblks;
+  size_t fordblks;
+  size_t keepcost;
+};
+
 // Android sometimes includes malloc.h no matter what, which yields to
 // conflicting return types for mallinfo() if we use our own structure. So if
 // struct mallinfo is declared (#define courtesy of malloc.h), use it directly.
diff --git a/compiler-rt/lib/scudo/standalone/wrappers_c.inc b/compiler-rt/lib/scudo/standalone/wrappers_c.inc
index 43efb02cb860..bbe3617dd0d6 100644
--- a/compiler-rt/lib/scudo/standalone/wrappers_c.inc
+++ b/compiler-rt/lib/scudo/standalone/wrappers_c.inc
@@ -54,6 +54,23 @@ INTERFACE WEAK struct SCUDO_MALLINFO SCUDO_PREFIX(mallinfo)(void) {
   return Info;
 }
 
+INTERFACE WEAK struct __scudo_mallinfo2 SCUDO_PREFIX(mallinfo2)(void) {
+  struct __scudo_mallinfo2 Info = {};
+  scudo::StatCounters Stats;
+  SCUDO_ALLOCATOR.getStats(Stats);
+  // Space allocated in mmapped regions (bytes)
+  Info.hblkhd = Stats[scudo::StatMapped];
+  // Maximum total allocated space (bytes)
+  Info.usmblks = Info.hblkhd;
+  // Space in freed fastbin blocks (bytes)
+  Info.fsmblks = Stats[scudo::StatFree];
+  // Total allocated space (bytes)
+  Info.uordblks = Stats[scudo::StatAllocated];
+  // Total free space (bytes)
+  Info.fordblks = Info.fsmblks;
+  return Info;
+}
+
 INTERFACE WEAK void *SCUDO_PREFIX(malloc)(size_t size) {
   return scudo::setErrnoOnNull(SCUDO_ALLOCATOR.allocate(
       size, scudo::Chunk::Origin::Malloc, SCUDO_MALLOC_ALIGNMENT));
@@ -226,7 +243,7 @@ INTERFACE WEAK int SCUDO_PREFIX(malloc_info)(UNUSED int options, FILE *stream) {
   fputs("<malloc version=\"scudo-1\">\n", stream);
   for (scudo::uptr i = 0; i != max_size; ++i)
     if (sizes[i])
-      fprintf(stream, "<alloc size=\"%lu\" count=\"%lu\"/>\n", i, sizes[i]);
+      fprintf(stream, "<alloc size=\"%zu\" count=\"%zu\"/>\n", i, sizes[i]);
   fputs("</malloc>\n", stream);
   SCUDO_PREFIX(free)(sizes);
   return 0;
diff --git a/compiler-rt/lib/scudo/standalone/wrappers_c_checks.h b/compiler-rt/lib/scudo/standalone/wrappers_c_checks.h
index 7fc1a9600e53..ec9c1a104e83 100644
--- a/compiler-rt/lib/scudo/standalone/wrappers_c_checks.h
+++ b/compiler-rt/lib/scudo/standalone/wrappers_c_checks.h
@@ -46,8 +46,10 @@ inline bool checkPosixMemalignAlignment(uptr Alignment) {
 // builtin supported by recent clang & GCC if it exists, otherwise fallback to a
 // costly division.
 inline bool checkForCallocOverflow(uptr Size, uptr N, uptr *Product) {
-#if __has_builtin(__builtin_umull_overflow)
+#if __has_builtin(__builtin_umull_overflow) && (SCUDO_WORDSIZE == 64U)
   return __builtin_umull_overflow(Size, N, Product);
+#elif __has_builtin(__builtin_umul_overflow) && (SCUDO_WORDSIZE == 32U)
+  return __builtin_umul_overflow(Size, N, Product);
 #else
   *Product = Size * N;
   if (!Size)
diff --git a/compiler-rt/lib/tsan/dd/dd_interceptors.cpp b/compiler-rt/lib/tsan/dd/dd_interceptors.cpp
index f78ef2d44279..2c36f691ec5b 100644
--- a/compiler-rt/lib/tsan/dd/dd_interceptors.cpp
+++ b/compiler-rt/lib/tsan/dd/dd_interceptors.cpp
@@ -285,7 +285,8 @@ static void InitDataSeg() {
     if (is_bss) g_data_end = segment.end;
     prev_is_data = is_data;
   }
-  VPrintf(1, "guessed data_start=%p data_end=%p\n",  g_data_start, g_data_end);
+  VPrintf(1, "guessed data_start=0x%zx data_end=0x%zx\n", g_data_start,
+          g_data_end);
   CHECK_LT(g_data_start, g_data_end);
   CHECK_GE((uptr)&g_data_start, g_data_start);
   CHECK_LT((uptr)&g_data_start, g_data_end);
diff --git a/compiler-rt/lib/tsan/dd/dd_rtl.cpp b/compiler-rt/lib/tsan/dd/dd_rtl.cpp
index 2095217586a8..35b367c0cecb 100644
--- a/compiler-rt/lib/tsan/dd/dd_rtl.cpp
+++ b/compiler-rt/lib/tsan/dd/dd_rtl.cpp
@@ -38,12 +38,12 @@ static void PrintStackTrace(Thread *thr, u32 stk) {
 static void ReportDeadlock(Thread *thr, DDReport *rep) {
   if (rep == 0)
     return;
-  BlockingMutexLock lock(&ctx->report_mutex);
+  Lock lock(&ctx->report_mutex);
   Printf("==============================\n");
   Printf("WARNING: lock-order-inversion (potential deadlock)\n");
   for (int i = 0; i < rep->n; i++) {
-    Printf("Thread %d locks mutex %llu while holding mutex %llu:\n",
-      rep->loop[i].thr_ctx, rep->loop[i].mtx_ctx1, rep->loop[i].mtx_ctx0);
+    Printf("Thread %lld locks mutex %llu while holding mutex %llu:\n",
+           rep->loop[i].thr_ctx, rep->loop[i].mtx_ctx1, rep->loop[i].mtx_ctx0);
     PrintStackTrace(thr, rep->loop[i].stk[1]);
     if (rep->loop[i].stk[0]) {
       Printf("Mutex %llu was acquired here:\n",
diff --git a/compiler-rt/lib/tsan/dd/dd_rtl.h b/compiler-rt/lib/tsan/dd/dd_rtl.h
index b1e19be57d3f..c812ffbd1393 100644
--- a/compiler-rt/lib/tsan/dd/dd_rtl.h
+++ b/compiler-rt/lib/tsan/dd/dd_rtl.h
@@ -19,7 +19,7 @@ namespace __dsan {
 
 typedef DDFlags Flags;
 
-struct Mutex {
+struct UserMutex {
   DDMutex dd;
 };
 
@@ -37,12 +37,12 @@ struct Callback final : public DDCallback {
   u32 Unwind() override;
 };
 
-typedef AddrHashMap<Mutex, 31051> MutexHashMap;
+typedef AddrHashMap<UserMutex, 31051> MutexHashMap;
 
 struct Context {
   DDetector *dd;
 
-  BlockingMutex report_mutex;
+  Mutex report_mutex;
   MutexHashMap mutex_map;
 };
 
diff --git a/compiler-rt/lib/tsan/go/tsan_go.cpp b/compiler-rt/lib/tsan/go/tsan_go.cpp
index 77987f43bf54..104c5b325aee 100644
--- a/compiler-rt/lib/tsan/go/tsan_go.cpp
+++ b/compiler-rt/lib/tsan/go/tsan_go.cpp
@@ -27,13 +27,9 @@ bool IsExpectedReport(uptr addr, uptr size) {
   return false;
 }
 
-void *internal_alloc(MBlockType typ, uptr sz) {
-  return InternalAlloc(sz);
-}
+void *Alloc(uptr sz) { return InternalAlloc(sz); }
 
-void internal_free(void *p) {
-  InternalFree(p);
-}
+void FreeImpl(void *p) { InternalFree(p); }
 
 // Callback into Go.
 static void (*go_runtime_cb)(uptr cmd, void *ctx);
@@ -103,14 +99,16 @@ ReportLocation *SymbolizeData(uptr addr) {
     MBlock *b = ctx->metamap.GetBlock(cbctx.start);
     if (!b)
       return 0;
-    ReportLocation *loc = ReportLocation::New(ReportLocationHeap);
+    auto *loc = New<ReportLocation>();
+    loc->type = ReportLocationHeap;
     loc->heap_chunk_start = cbctx.start;
     loc->heap_chunk_size = b->siz;
     loc->tid = b->tid;
     loc->stack = SymbolizeStackId(b->stk);
     return loc;
   } else {
-    ReportLocation *loc = ReportLocation::New(ReportLocationGlobal);
+    auto *loc = New<ReportLocation>();
+    loc->type = ReportLocationGlobal;
     loc->global.name = internal_strdup(cbctx.name ? cbctx.name : "??");
     loc->global.file = internal_strdup(cbctx.file ? cbctx.file : "??");
     loc->global.line = cbctx.line;
@@ -142,8 +140,7 @@ Processor *ThreadState::proc() {
 extern "C" {
 
 static ThreadState *AllocGoroutine() {
-  ThreadState *thr = (ThreadState*)internal_alloc(MBlockThreadContex,
-      sizeof(ThreadState));
+  auto *thr = (ThreadState *)Alloc(sizeof(ThreadState));
   internal_memset(thr, 0, sizeof(*thr));
   return thr;
 }
@@ -170,25 +167,25 @@ void __tsan_map_shadow(uptr addr, uptr size) {
 }
 
 void __tsan_read(ThreadState *thr, void *addr, void *pc) {
-  MemoryRead(thr, (uptr)pc, (uptr)addr, kSizeLog1);
+  MemoryAccess(thr, (uptr)pc, (uptr)addr, 1, kAccessRead);
 }
 
 void __tsan_read_pc(ThreadState *thr, void *addr, uptr callpc, uptr pc) {
   if (callpc != 0)
     FuncEntry(thr, callpc);
-  MemoryRead(thr, (uptr)pc, (uptr)addr, kSizeLog1);
+  MemoryAccess(thr, (uptr)pc, (uptr)addr, 1, kAccessRead);
   if (callpc != 0)
     FuncExit(thr);
 }
 
 void __tsan_write(ThreadState *thr, void *addr, void *pc) {
-  MemoryWrite(thr, (uptr)pc, (uptr)addr, kSizeLog1);
+  MemoryAccess(thr, (uptr)pc, (uptr)addr, 1, kAccessWrite);
 }
 
 void __tsan_write_pc(ThreadState *thr, void *addr, uptr callpc, uptr pc) {
   if (callpc != 0)
     FuncEntry(thr, callpc);
-  MemoryWrite(thr, (uptr)pc, (uptr)addr, kSizeLog1);
+  MemoryAccess(thr, (uptr)pc, (uptr)addr, 1, kAccessWrite);
   if (callpc != 0)
     FuncExit(thr);
 }
@@ -213,7 +210,7 @@ void __tsan_malloc(ThreadState *thr, uptr pc, uptr p, uptr sz) {
   CHECK(inited);
   if (thr && pc)
     ctx->metamap.AllocBlock(thr, pc, p, sz);
-  MemoryResetRange(0, 0, (uptr)p, sz);
+  MemoryResetRange(thr, pc, (uptr)p, sz);
 }
 
 void __tsan_free(uptr p, uptr sz) {
@@ -223,13 +220,13 @@ void __tsan_free(uptr p, uptr sz) {
 void __tsan_go_start(ThreadState *parent, ThreadState **pthr, void *pc) {
   ThreadState *thr = AllocGoroutine();
   *pthr = thr;
-  int goid = ThreadCreate(parent, (uptr)pc, 0, true);
+  Tid goid = ThreadCreate(parent, (uptr)pc, 0, true);
   ThreadStart(thr, goid, 0, ThreadType::Regular);
 }
 
 void __tsan_go_end(ThreadState *thr) {
   ThreadFinish(thr);
-  internal_free(thr);
+  Free(thr);
 }
 
 void __tsan_proc_create(Processor **pproc) {
@@ -256,9 +253,7 @@ void __tsan_release_merge(ThreadState *thr, void *addr) {
   Release(thr, 0, (uptr)addr);
 }
 
-void __tsan_finalizer_goroutine(ThreadState *thr) {
-  AcquireGlobal(thr, 0);
-}
+void __tsan_finalizer_goroutine(ThreadState *thr) { AcquireGlobal(thr); }
 
 void __tsan_mutex_before_lock(ThreadState *thr, uptr addr, uptr write) {
   if (write)
@@ -285,9 +280,7 @@ void __tsan_go_ignore_sync_begin(ThreadState *thr) {
   ThreadIgnoreSyncBegin(thr, 0);
 }
 
-void __tsan_go_ignore_sync_end(ThreadState *thr) {
-  ThreadIgnoreSyncEnd(thr, 0);
-}
+void __tsan_go_ignore_sync_end(ThreadState *thr) { ThreadIgnoreSyncEnd(thr); }
 
 void __tsan_report_count(u64 *pn) {
   Lock lock(&ctx->report_mtx);
diff --git a/compiler-rt/lib/tsan/rtl/tsan_clock.cpp b/compiler-rt/lib/tsan/rtl/tsan_clock.cpp
index 61848c21d162..d122b67c0aaa 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_clock.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_clock.cpp
@@ -72,9 +72,9 @@
 // clk_ - variable size vector clock, low kClkBits hold timestamp,
 //   the remaining bits hold "acquired" flag (the actual value is thread's
 //   reused counter);
-//   if acquried == thr->reused_, then the respective thread has already
+//   if acquired == thr->reused_, then the respective thread has already
 //   acquired this clock (except possibly for dirty elements).
-// dirty_ - holds up to two indeces in the vector clock that other threads
+// dirty_ - holds up to two indices in the vector clock that other threads
 //   need to acquire regardless of "acquired" flag value;
 // release_store_tid_ - denotes that the clock state is a result of
 //   release-store operation by the thread with release_store_tid_ index.
@@ -272,7 +272,7 @@ void ThreadClock::ReleaseStore(ClockCache *c, SyncClock *dst) {
     // we could update the existing clock and cache it, or replace it with the
     // currently cached clock and release the old one. And for a shared
     // existing clock, we could replace it with the currently cached;
-    // or unshare, update and cache. But, for simplicity, we currnetly reuse
+    // or unshare, update and cache. But, for simplicity, we currently reuse
     // cached clock only when the target clock is empty.
     dst->tab_ = ctx->clock_alloc.Map(cached_idx_);
     dst->tab_idx_ = cached_idx_;
@@ -285,7 +285,7 @@ void ThreadClock::ReleaseStore(ClockCache *c, SyncClock *dst) {
     dst->dirty_[0].epoch = clk_[tid_];
     dst->release_store_tid_ = tid_;
     dst->release_store_reused_ = reused_;
-    // Rememeber that we don't need to acquire it in future.
+    // Remember that we don't need to acquire it in future.
     dst->elem(tid_).reused = reused_;
     // Grab a reference.
     atomic_fetch_add(ref_ptr(dst->tab_), 1, memory_order_relaxed);
@@ -316,7 +316,7 @@ void ThreadClock::ReleaseStore(ClockCache *c, SyncClock *dst) {
   for (uptr i = 0; i < kDirtyTids; i++) dst->dirty_[i].set_tid(kInvalidTid);
   dst->release_store_tid_ = tid_;
   dst->release_store_reused_ = reused_;
-  // Rememeber that we don't need to acquire it in future.
+  // Remember that we don't need to acquire it in future.
   dst->elem(tid_).reused = reused_;
 
   // If the resulting clock is cachable, cache it for future release operations.
diff --git a/compiler-rt/lib/tsan/rtl/tsan_clock.h b/compiler-rt/lib/tsan/rtl/tsan_clock.h
index 31376a1bc9e2..11cbc0c0b86b 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_clock.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_clock.h
@@ -213,7 +213,7 @@ class ThreadClock {
   // We reuse it for subsequent store-release operations without intervening
   // acquire operations. Since it is shared (and thus constant), clock value
   // for the current thread is then stored in dirty entries in the SyncClock.
-  // We host a refernece to the table while it is cached here.
+  // We host a reference to the table while it is cached here.
   u32 cached_idx_;
   u16 cached_size_;
   u16 cached_blocks_;
diff --git a/compiler-rt/lib/tsan/rtl/tsan_debugging.cpp b/compiler-rt/lib/tsan/rtl/tsan_debugging.cpp
index d3d6255090b7..1d3c3849a446 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_debugging.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_debugging.cpp
@@ -195,9 +195,9 @@ const char *__tsan_locate_address(uptr addr, char *name, uptr name_size,
   const char *region_kind = nullptr;
   if (name && name_size > 0) name[0] = 0;
 
-  if (IsMetaMem(addr)) {
+  if (IsMetaMem(reinterpret_cast<u32 *>(addr))) {
     region_kind = "meta shadow";
-  } else if (IsShadowMem(addr)) {
+  } else if (IsShadowMem(reinterpret_cast<RawShadow *>(addr))) {
     region_kind = "shadow";
   } else {
     bool is_stack = false;
@@ -215,9 +215,9 @@ const char *__tsan_locate_address(uptr addr, char *name, uptr name_size,
     } else {
       // TODO(kuba.brecka): We should not lock. This is supposed to be called
       // from within the debugger when other threads are stopped.
-      ctx->thread_registry->Lock();
+      ctx->thread_registry.Lock();
       ThreadContext *tctx = IsThreadStackOrTls(addr, &is_stack);
-      ctx->thread_registry->Unlock();
+      ctx->thread_registry.Unlock();
       if (tctx) {
         region_kind = is_stack ? "stack" : "tls";
       } else {
@@ -252,7 +252,7 @@ int __tsan_get_alloc_stack(uptr addr, uptr *trace, uptr size, int *thread_id,
   *thread_id = b->tid;
   // No locking.  This is supposed to be called from within the debugger when
   // other threads are stopped.
-  ThreadContextBase *tctx = ctx->thread_registry->GetThreadLocked(b->tid);
+  ThreadContextBase *tctx = ctx->thread_registry.GetThreadLocked(b->tid);
   *os_id = tctx->os_id;
 
   StackTrace stack = StackDepotGet(b->stk);
diff --git a/compiler-rt/lib/tsan/rtl/tsan_defs.h b/compiler-rt/lib/tsan/rtl/tsan_defs.h
index f2fb7b1a213f..fe0c1da31599 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_defs.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_defs.h
@@ -18,6 +18,24 @@
 #include "sanitizer_common/sanitizer_mutex.h"
 #include "ubsan/ubsan_platform.h"
 
+#ifndef TSAN_VECTORIZE
+#  define TSAN_VECTORIZE __SSE4_2__
+#endif
+
+#if TSAN_VECTORIZE
+// <emmintrin.h> transitively includes <stdlib.h>,
+// and it's prohibited to include std headers into tsan runtime.
+// So we do this dirty trick.
+#  define _MM_MALLOC_H_INCLUDED
+#  define __MM_MALLOC_H
+#  include <emmintrin.h>
+#  include <smmintrin.h>
+#  define VECTOR_ALIGNED ALIGNED(16)
+typedef __m128i m128;
+#else
+#  define VECTOR_ALIGNED
+#endif
+
 // Setup defaults for compile definitions.
 #ifndef TSAN_NO_HISTORY
 # define TSAN_NO_HISTORY 0
@@ -33,6 +51,19 @@
 
 namespace __tsan {
 
+constexpr uptr kByteBits = 8;
+
+// Thread slot ID.
+enum class Sid : u8 {};
+constexpr uptr kThreadSlotCount = 256;
+constexpr Sid kFreeSid = static_cast<Sid>(255);
+
+// Abstract time unit, vector clock element.
+enum class Epoch : u16 {};
+constexpr uptr kEpochBits = 14;
+constexpr Epoch kEpochZero = static_cast<Epoch>(0);
+constexpr Epoch kEpochOver = static_cast<Epoch>(1 << kEpochBits);
+
 const int kClkBits = 42;
 const unsigned kMaxTidReuse = (1 << (64 - kClkBits)) - 1;
 
@@ -75,8 +106,9 @@ const uptr kShadowCnt = 4;
 // That many user bytes are mapped onto a single shadow cell.
 const uptr kShadowCell = 8;
 
-// Size of a single shadow value (u64).
-const uptr kShadowSize = 8;
+// Single shadow value.
+typedef u64 RawShadow;
+const uptr kShadowSize = sizeof(RawShadow);
 
 // Shadow memory is kShadowMultiplier times larger than user memory.
 const uptr kShadowMultiplier = kShadowSize * kShadowCnt / kShadowCell;
@@ -88,6 +120,9 @@ const uptr kMetaShadowCell = 8;
 // Size of a single meta shadow value (u32).
 const uptr kMetaShadowSize = 4;
 
+// All addresses and PCs are assumed to be compressable to that many bits.
+const uptr kCompressedAddrBits = 44;
+
 #if TSAN_NO_HISTORY
 const bool kCollectHistory = false;
 #else
@@ -154,12 +189,23 @@ struct ReportStack;
 class ReportDesc;
 class RegionAlloc;
 
+typedef uptr AccessType;
+
+enum : AccessType {
+  kAccessWrite = 0,
+  kAccessRead = 1 << 0,
+  kAccessAtomic = 1 << 1,
+  kAccessVptr = 1 << 2,  // read or write of an object virtual table pointer
+  kAccessFree = 1 << 3,  // synthetic memory access during memory freeing
+  kAccessExternalPC = 1 << 4,  // access PC can have kExternalPCBit set
+};
+
 // Descriptor of user's memory block.
 struct MBlock {
   u64  siz : 48;
   u64  tag : 16;
-  u32  stk;
-  u16  tid;
+  StackID stk;
+  Tid tid;
 };
 
 COMPILER_CHECK(sizeof(MBlock) == 16);
diff --git a/compiler-rt/lib/tsan/rtl/tsan_dense_alloc.h b/compiler-rt/lib/tsan/rtl/tsan_dense_alloc.h
index 68ded43c4f6b..9e15f74a0615 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_dense_alloc.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_dense_alloc.h
@@ -49,11 +49,7 @@ class DenseSlabAlloc {
   static_assert(sizeof(T) > sizeof(IndexT),
                 "it doesn't make sense to use dense alloc");
 
-  explicit DenseSlabAlloc(LinkerInitialized, const char *name) {
-    freelist_ = 0;
-    fillpos_ = 0;
-    name_ = name;
-  }
+  DenseSlabAlloc(LinkerInitialized, const char *name) : name_(name) {}
 
   explicit DenseSlabAlloc(const char *name)
       : DenseSlabAlloc(LINKER_INITIALIZED, name) {
@@ -89,6 +85,8 @@ class DenseSlabAlloc {
   }
 
   void FlushCache(Cache *c) {
+    if (!c->pos)
+      return;
     SpinMutexLock lock(&mtx_);
     while (c->pos) {
       IndexT idx = c->cache[--c->pos];
@@ -102,33 +100,39 @@ class DenseSlabAlloc {
     internal_memset(c->cache, 0, sizeof(c->cache));
   }
 
+  uptr AllocatedMemory() const {
+    return atomic_load_relaxed(&fillpos_) * kL2Size * sizeof(T);
+  }
+
  private:
   T *map_[kL1Size];
   SpinMutex mtx_;
-  IndexT freelist_;
-  uptr fillpos_;
-  const char *name_;
+  IndexT freelist_ = {0};
+  atomic_uintptr_t fillpos_ = {0};
+  const char *const name_;
 
   void Refill(Cache *c) {
     SpinMutexLock lock(&mtx_);
     if (freelist_ == 0) {
-      if (fillpos_ == kL1Size) {
+      uptr fillpos = atomic_load_relaxed(&fillpos_);
+      if (fillpos == kL1Size) {
         Printf("ThreadSanitizer: %s overflow (%zu*%zu). Dying.\n",
             name_, kL1Size, kL2Size);
         Die();
       }
-      VPrintf(2, "ThreadSanitizer: growing %s: %zu out of %zu*%zu\n",
-          name_, fillpos_, kL1Size, kL2Size);
+      VPrintf(2, "ThreadSanitizer: growing %s: %zu out of %zu*%zu\n", name_,
+              fillpos, kL1Size, kL2Size);
       T *batch = (T*)MmapOrDie(kL2Size * sizeof(T), name_);
       // Reserve 0 as invalid index.
-      IndexT start = fillpos_ == 0 ? 1 : 0;
+      IndexT start = fillpos == 0 ? 1 : 0;
       for (IndexT i = start; i < kL2Size; i++) {
         new(batch + i) T;
-        *(IndexT*)(batch + i) = i + 1 + fillpos_ * kL2Size;
+        *(IndexT *)(batch + i) = i + 1 + fillpos * kL2Size;
       }
       *(IndexT*)(batch + kL2Size - 1) = 0;
-      freelist_ = fillpos_ * kL2Size + start;
-      map_[fillpos_++] = batch;
+      freelist_ = fillpos * kL2Size + start;
+      map_[fillpos] = batch;
+      atomic_store_relaxed(&fillpos_, fillpos + 1);
     }
     for (uptr i = 0; i < Cache::kSize / 2 && freelist_ != 0; i++) {
       IndexT idx = freelist_;
diff --git a/compiler-rt/lib/tsan/rtl/tsan_external.cpp b/compiler-rt/lib/tsan/rtl/tsan_external.cpp
index a87e12f2936f..19ae174f20a5 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_external.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_external.cpp
@@ -10,9 +10,12 @@
 //
 //===----------------------------------------------------------------------===//
 #include "tsan_rtl.h"
-#include "tsan_interceptors.h"
 #include "sanitizer_common/sanitizer_ptrauth.h"
 
+#if !SANITIZER_GO
+#  include "tsan_interceptors.h"
+#endif
+
 namespace __tsan {
 
 #define CALLERPC ((uptr)__builtin_return_address(0))
@@ -57,16 +60,14 @@ uptr TagFromShadowStackFrame(uptr pc) {
 
 #if !SANITIZER_GO
 
-typedef void(*AccessFunc)(ThreadState *, uptr, uptr, int);
-void ExternalAccess(void *addr, uptr caller_pc, void *tag, AccessFunc access) {
+void ExternalAccess(void *addr, uptr caller_pc, void *tag, AccessType typ) {
   CHECK_LT(tag, atomic_load(&used_tags, memory_order_relaxed));
   ThreadState *thr = cur_thread();
   if (caller_pc) FuncEntry(thr, caller_pc);
   InsertShadowStackFrameForTag(thr, (uptr)tag);
   bool in_ignored_lib;
-  if (!caller_pc || !libignore()->IsIgnored(caller_pc, &in_ignored_lib)) {
-    access(thr, CALLERPC, (uptr)addr, kSizeLog1);
-  }
+  if (!caller_pc || !libignore()->IsIgnored(caller_pc, &in_ignored_lib))
+    MemoryAccess(thr, CALLERPC, (uptr)addr, 1, typ);
   FuncExit(thr);
   if (caller_pc) FuncExit(thr);
 }
@@ -92,7 +93,7 @@ void __tsan_external_register_header(void *tag, const char *header) {
   header = internal_strdup(header);
   char *old_header =
       (char *)atomic_exchange(header_ptr, (uptr)header, memory_order_seq_cst);
-  if (old_header) internal_free(old_header);
+  Free(old_header);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
@@ -111,12 +112,12 @@ void __tsan_external_assign_tag(void *addr, void *tag) {
 
 SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_external_read(void *addr, void *caller_pc, void *tag) {
-  ExternalAccess(addr, STRIP_PAC_PC(caller_pc), tag, MemoryRead);
+  ExternalAccess(addr, STRIP_PAC_PC(caller_pc), tag, kAccessRead);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_external_write(void *addr, void *caller_pc, void *tag) {
-  ExternalAccess(addr, STRIP_PAC_PC(caller_pc), tag, MemoryWrite);
+  ExternalAccess(addr, STRIP_PAC_PC(caller_pc), tag, kAccessWrite);
 }
 }  // extern "C"
 
diff --git a/compiler-rt/lib/tsan/rtl/tsan_fd.cpp b/compiler-rt/lib/tsan/rtl/tsan_fd.cpp
index 50a6b56916aa..255ffa8daf76 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_fd.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_fd.cpp
@@ -26,8 +26,8 @@ struct FdSync {
 
 struct FdDesc {
   FdSync *sync;
-  int creation_tid;
-  u32 creation_stack;
+  Tid creation_tid;
+  StackID creation_stack;
 };
 
 struct FdContext {
@@ -115,7 +115,7 @@ static void init(ThreadState *thr, uptr pc, int fd, FdSync *s,
     MemoryRangeImitateWrite(thr, pc, (uptr)d, 8);
   } else {
     // See the dup-related comment in FdClose.
-    MemoryRead(thr, pc, (uptr)d, kSizeLog8);
+    MemoryAccess(thr, pc, (uptr)d, 8, kAccessRead);
   }
 }
 
@@ -140,7 +140,7 @@ void FdOnFork(ThreadState *thr, uptr pc) {
   }
 }
 
-bool FdLocation(uptr addr, int *fd, int *tid, u32 *stack) {
+bool FdLocation(uptr addr, int *fd, Tid *tid, StackID *stack) {
   for (int l1 = 0; l1 < kTableSizeL1; l1++) {
     FdDesc *tab = (FdDesc*)atomic_load(&fdctx.tab[l1], memory_order_relaxed);
     if (tab == 0)
@@ -163,7 +163,7 @@ void FdAcquire(ThreadState *thr, uptr pc, int fd) {
   FdDesc *d = fddesc(thr, pc, fd);
   FdSync *s = d->sync;
   DPrintf("#%d: FdAcquire(%d) -> %p\n", thr->tid, fd, s);
-  MemoryRead(thr, pc, (uptr)d, kSizeLog8);
+  MemoryAccess(thr, pc, (uptr)d, 8, kAccessRead);
   if (s)
     Acquire(thr, pc, (uptr)s);
 }
@@ -174,7 +174,7 @@ void FdRelease(ThreadState *thr, uptr pc, int fd) {
   FdDesc *d = fddesc(thr, pc, fd);
   FdSync *s = d->sync;
   DPrintf("#%d: FdRelease(%d) -> %p\n", thr->tid, fd, s);
-  MemoryRead(thr, pc, (uptr)d, kSizeLog8);
+  MemoryAccess(thr, pc, (uptr)d, 8, kAccessRead);
   if (s)
     Release(thr, pc, (uptr)s);
 }
@@ -184,7 +184,7 @@ void FdAccess(ThreadState *thr, uptr pc, int fd) {
   if (bogusfd(fd))
     return;
   FdDesc *d = fddesc(thr, pc, fd);
-  MemoryRead(thr, pc, (uptr)d, kSizeLog8);
+  MemoryAccess(thr, pc, (uptr)d, 8, kAccessRead);
 }
 
 void FdClose(ThreadState *thr, uptr pc, int fd, bool write) {
@@ -194,7 +194,7 @@ void FdClose(ThreadState *thr, uptr pc, int fd, bool write) {
   FdDesc *d = fddesc(thr, pc, fd);
   if (write) {
     // To catch races between fd usage and close.
-    MemoryWrite(thr, pc, (uptr)d, kSizeLog8);
+    MemoryAccess(thr, pc, (uptr)d, 8, kAccessWrite);
   } else {
     // This path is used only by dup2/dup3 calls.
     // We do read instead of write because there is a number of legitimate
@@ -204,15 +204,15 @@ void FdClose(ThreadState *thr, uptr pc, int fd, bool write) {
     // 2. Some daemons dup /dev/null in place of stdin/stdout.
     // On the other hand we have not seen cases when write here catches real
     // bugs.
-    MemoryRead(thr, pc, (uptr)d, kSizeLog8);
+    MemoryAccess(thr, pc, (uptr)d, 8, kAccessRead);
   }
   // We need to clear it, because if we do not intercept any call out there
   // that creates fd, we will hit false postives.
   MemoryResetRange(thr, pc, (uptr)d, 8);
   unref(thr, pc, d->sync);
   d->sync = 0;
-  d->creation_tid = 0;
-  d->creation_stack = 0;
+  d->creation_tid = kInvalidTid;
+  d->creation_stack = kInvalidStackID;
 }
 
 void FdFileCreate(ThreadState *thr, uptr pc, int fd) {
@@ -228,7 +228,7 @@ void FdDup(ThreadState *thr, uptr pc, int oldfd, int newfd, bool write) {
     return;
   // Ignore the case when user dups not yet connected socket.
   FdDesc *od = fddesc(thr, pc, oldfd);
-  MemoryRead(thr, pc, (uptr)od, kSizeLog8);
+  MemoryAccess(thr, pc, (uptr)od, 8, kAccessRead);
   FdClose(thr, pc, newfd, write);
   init(thr, pc, newfd, ref(od->sync), write);
 }
diff --git a/compiler-rt/lib/tsan/rtl/tsan_fd.h b/compiler-rt/lib/tsan/rtl/tsan_fd.h
index ce4f2f73bac6..d9648178481c 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_fd.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_fd.h
@@ -53,7 +53,7 @@ void FdSocketCreate(ThreadState *thr, uptr pc, int fd);
 void FdSocketAccept(ThreadState *thr, uptr pc, int fd, int newfd);
 void FdSocketConnecting(ThreadState *thr, uptr pc, int fd);
 void FdSocketConnect(ThreadState *thr, uptr pc, int fd);
-bool FdLocation(uptr addr, int *fd, int *tid, u32 *stack);
+bool FdLocation(uptr addr, int *fd, Tid *tid, StackID *stack);
 void FdOnFork(ThreadState *thr, uptr pc);
 
 uptr File2addr(const char *path);
diff --git a/compiler-rt/lib/tsan/rtl/tsan_flags.cpp b/compiler-rt/lib/tsan/rtl/tsan_flags.cpp
index 49e4a9c21da9..ee89862d17bd 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_flags.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_flags.cpp
@@ -55,6 +55,7 @@ void InitializeFlags(Flags *f, const char *env, const char *env_option_name) {
     // Override some common flags defaults.
     CommonFlags cf;
     cf.CopyFrom(*common_flags());
+    cf.external_symbolizer_path = GetEnv("TSAN_SYMBOLIZER_PATH");
     cf.allow_addr2line = true;
     if (SANITIZER_GO) {
       // Does not work as expected for Go: runtime handles SIGABRT and crashes.
diff --git a/compiler-rt/lib/tsan/rtl/tsan_flags.inc b/compiler-rt/lib/tsan/rtl/tsan_flags.inc
index 2105c754486f..7954a4307fa1 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_flags.inc
+++ b/compiler-rt/lib/tsan/rtl/tsan_flags.inc
@@ -43,7 +43,6 @@ TSAN_FLAG(
     bool, force_seq_cst_atomics, false,
     "If set, all atomics are effectively sequentially consistent (seq_cst), "
     "regardless of what user actually specified.")
-TSAN_FLAG(bool, print_benign, false, "Print matched \"benign\" races at exit.")
 TSAN_FLAG(bool, halt_on_error, false, "Exit after first reported error.")
 TSAN_FLAG(int, atexit_sleep_ms, 1000,
           "Sleep in main thread before exiting for that many ms "
diff --git a/compiler-rt/lib/tsan/rtl/tsan_ignoreset.cpp b/compiler-rt/lib/tsan/rtl/tsan_ignoreset.cpp
index f6e41f668618..1fca1cf4f9fc 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_ignoreset.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_ignoreset.cpp
@@ -19,7 +19,7 @@ IgnoreSet::IgnoreSet()
     : size_() {
 }
 
-void IgnoreSet::Add(u32 stack_id) {
+void IgnoreSet::Add(StackID stack_id) {
   if (size_ == kMaxSize)
     return;
   for (uptr i = 0; i < size_; i++) {
@@ -29,15 +29,7 @@ void IgnoreSet::Add(u32 stack_id) {
   stacks_[size_++] = stack_id;
 }
 
-void IgnoreSet::Reset() {
-  size_ = 0;
-}
-
-uptr IgnoreSet::Size() const {
-  return size_;
-}
-
-u32 IgnoreSet::At(uptr i) const {
+StackID IgnoreSet::At(uptr i) const {
   CHECK_LT(i, size_);
   CHECK_LE(size_, kMaxSize);
   return stacks_[i];
diff --git a/compiler-rt/lib/tsan/rtl/tsan_ignoreset.h b/compiler-rt/lib/tsan/rtl/tsan_ignoreset.h
index 3e318bd674d9..4e2511291ce4 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_ignoreset.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_ignoreset.h
@@ -19,17 +19,16 @@ namespace __tsan {
 
 class IgnoreSet {
  public:
-  static const uptr kMaxSize = 16;
-
   IgnoreSet();
-  void Add(u32 stack_id);
-  void Reset();
-  uptr Size() const;
-  u32 At(uptr i) const;
+  void Add(StackID stack_id);
+  void Reset() { size_ = 0; }
+  uptr Size() const { return size_; }
+  StackID At(uptr i) const;
 
  private:
+  static constexpr uptr kMaxSize = 16;
   uptr size_;
-  u32 stacks_[kMaxSize];
+  StackID stacks_[kMaxSize];
 };
 
 }  // namespace __tsan
diff --git a/compiler-rt/lib/tsan/rtl/tsan_ilist.h b/compiler-rt/lib/tsan/rtl/tsan_ilist.h
new file mode 100644
index 000000000000..d7d8be219dbe
--- /dev/null
+++ b/compiler-rt/lib/tsan/rtl/tsan_ilist.h
@@ -0,0 +1,189 @@
+//===-- tsan_ilist.h --------------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of ThreadSanitizer (TSan), a race detector.
+//
+//===----------------------------------------------------------------------===//
+#ifndef TSAN_ILIST_H
+#define TSAN_ILIST_H
+
+#include "sanitizer_common/sanitizer_internal_defs.h"
+
+namespace __tsan {
+
+class INode {
+ public:
+  INode() = default;
+
+ private:
+  INode* next_ = nullptr;
+  INode* prev_ = nullptr;
+
+  template <typename Base, INode Base::*Node, typename Elem>
+  friend class IList;
+  INode(const INode&) = delete;
+  void operator=(const INode&) = delete;
+};
+
+// Intrusive doubly-linked list.
+//
+// The node class (MyNode) needs to include "INode foo" field,
+// then the list can be declared as IList<MyNode, &MyNode::foo>.
+// This design allows to link MyNode into multiple lists using
+// different INode fields.
+// The optional Elem template argument allows to specify node MDT
+// (most derived type) if it's different from MyNode.
+template <typename Base, INode Base::*Node, typename Elem = Base>
+class IList {
+ public:
+  IList();
+
+  void PushFront(Elem* e);
+  void PushBack(Elem* e);
+  void Remove(Elem* e);
+
+  Elem* PopFront();
+  Elem* PopBack();
+  Elem* Front();
+  Elem* Back();
+
+  // Prev links point towards front of the queue.
+  Elem* Prev(Elem* e);
+  // Next links point towards back of the queue.
+  Elem* Next(Elem* e);
+
+  uptr Size() const;
+  bool Empty() const;
+  bool Queued(Elem* e) const;
+
+ private:
+  INode node_;
+  uptr size_ = 0;
+
+  void Push(Elem* e, INode* after);
+  static INode* ToNode(Elem* e);
+  static Elem* ToElem(INode* n);
+
+  IList(const IList&) = delete;
+  void operator=(const IList&) = delete;
+};
+
+template <typename Base, INode Base::*Node, typename Elem>
+IList<Base, Node, Elem>::IList() {
+  node_.next_ = node_.prev_ = &node_;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+void IList<Base, Node, Elem>::PushFront(Elem* e) {
+  Push(e, &node_);
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+void IList<Base, Node, Elem>::PushBack(Elem* e) {
+  Push(e, node_.prev_);
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+void IList<Base, Node, Elem>::Push(Elem* e, INode* after) {
+  INode* n = ToNode(e);
+  DCHECK_EQ(n->next_, nullptr);
+  DCHECK_EQ(n->prev_, nullptr);
+  INode* next = after->next_;
+  n->next_ = next;
+  n->prev_ = after;
+  next->prev_ = n;
+  after->next_ = n;
+  size_++;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+void IList<Base, Node, Elem>::Remove(Elem* e) {
+  INode* n = ToNode(e);
+  INode* next = n->next_;
+  INode* prev = n->prev_;
+  DCHECK(next);
+  DCHECK(prev);
+  DCHECK(size_);
+  next->prev_ = prev;
+  prev->next_ = next;
+  n->prev_ = n->next_ = nullptr;
+  size_--;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+Elem* IList<Base, Node, Elem>::PopFront() {
+  Elem* e = Front();
+  if (e)
+    Remove(e);
+  return e;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+Elem* IList<Base, Node, Elem>::PopBack() {
+  Elem* e = Back();
+  if (e)
+    Remove(e);
+  return e;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+Elem* IList<Base, Node, Elem>::Front() {
+  return size_ ? ToElem(node_.next_) : nullptr;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+Elem* IList<Base, Node, Elem>::Back() {
+  return size_ ? ToElem(node_.prev_) : nullptr;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+Elem* IList<Base, Node, Elem>::Prev(Elem* e) {
+  INode* n = ToNode(e);
+  DCHECK(n->prev_);
+  return n->prev_ != &node_ ? ToElem(n->prev_) : nullptr;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+Elem* IList<Base, Node, Elem>::Next(Elem* e) {
+  INode* n = ToNode(e);
+  DCHECK(n->next_);
+  return n->next_ != &node_ ? ToElem(n->next_) : nullptr;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+uptr IList<Base, Node, Elem>::Size() const {
+  return size_;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+bool IList<Base, Node, Elem>::Empty() const {
+  return size_ == 0;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+bool IList<Base, Node, Elem>::Queued(Elem* e) const {
+  INode* n = ToNode(e);
+  DCHECK_EQ(!n->next_, !n->prev_);
+  return n->next_;
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+INode* IList<Base, Node, Elem>::ToNode(Elem* e) {
+  return &(e->*Node);
+}
+
+template <typename Base, INode Base::*Node, typename Elem>
+Elem* IList<Base, Node, Elem>::ToElem(INode* n) {
+  return static_cast<Elem*>(reinterpret_cast<Base*>(
+      reinterpret_cast<uptr>(n) -
+      reinterpret_cast<uptr>(&(reinterpret_cast<Elem*>(0)->*Node))));
+}
+
+}  // namespace __tsan
+
+#endif
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interceptors.h b/compiler-rt/lib/tsan/rtl/tsan_interceptors.h
index c5716f53a323..61dbb81ffec4 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_interceptors.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_interceptors.h
@@ -10,44 +10,61 @@ class ScopedInterceptor {
  public:
   ScopedInterceptor(ThreadState *thr, const char *fname, uptr pc);
   ~ScopedInterceptor();
-  void DisableIgnores();
-  void EnableIgnores();
+  void DisableIgnores() {
+    if (UNLIKELY(ignoring_))
+      DisableIgnoresImpl();
+  }
+  void EnableIgnores() {
+    if (UNLIKELY(ignoring_))
+      EnableIgnoresImpl();
+  }
+
  private:
   ThreadState *const thr_;
-  const uptr pc_;
   bool in_ignored_lib_;
   bool ignoring_;
+
+  void DisableIgnoresImpl();
+  void EnableIgnoresImpl();
 };
 
 LibIgnore *libignore();
 
 #if !SANITIZER_GO
 inline bool in_symbolizer() {
-  cur_thread_init();
-  return UNLIKELY(cur_thread()->in_symbolizer);
+  return UNLIKELY(cur_thread_init()->in_symbolizer);
 }
 #endif
 
 }  // namespace __tsan
 
-#define SCOPED_INTERCEPTOR_RAW(func, ...)      \
-  cur_thread_init();                           \
-  ThreadState *thr = cur_thread();             \
-  const uptr caller_pc = GET_CALLER_PC();      \
-  ScopedInterceptor si(thr, #func, caller_pc); \
-  const uptr pc = GET_CURRENT_PC();            \
-  (void)pc;                                    \
-  /**/
+#define SCOPED_INTERCEPTOR_RAW(func, ...)            \
+  ThreadState *thr = cur_thread_init();              \
+  ScopedInterceptor si(thr, #func, GET_CALLER_PC()); \
+  UNUSED const uptr pc = GET_CURRENT_PC();
 
-#define SCOPED_TSAN_INTERCEPTOR(func, ...) \
-    SCOPED_INTERCEPTOR_RAW(func, __VA_ARGS__); \
-    if (REAL(func) == 0) { \
+#ifdef __powerpc64__
+// Debugging of crashes on powerpc after commit:
+// c80604f7a3 ("tsan: remove real func check from interceptors")
+// Somehow replacing if with DCHECK leads to strange failures in:
+// SanitizerCommon-tsan-powerpc64le-Linux :: Linux/ptrace.cpp
+// https://lab.llvm.org/buildbot/#/builders/105
+// https://lab.llvm.org/buildbot/#/builders/121
+// https://lab.llvm.org/buildbot/#/builders/57
+#  define CHECK_REAL_FUNC(func)                                          \
+    if (REAL(func) == 0) {                                               \
       Report("FATAL: ThreadSanitizer: failed to intercept %s\n", #func); \
-      Die(); \
-    }                                                    \
-    if (!thr->is_inited || thr->ignore_interceptors || thr->in_ignored_lib) \
-      return REAL(func)(__VA_ARGS__); \
-/**/
+      Die();                                                             \
+    }
+#else
+#  define CHECK_REAL_FUNC(func) DCHECK(REAL(func))
+#endif
+
+#define SCOPED_TSAN_INTERCEPTOR(func, ...)                                \
+  SCOPED_INTERCEPTOR_RAW(func, __VA_ARGS__);                              \
+  CHECK_REAL_FUNC(func);                                                  \
+  if (!thr->is_inited || thr->ignore_interceptors || thr->in_ignored_lib) \
+    return REAL(func)(__VA_ARGS__);
 
 #define SCOPED_TSAN_INTERCEPTOR_USER_CALLBACK_START() \
     si.DisableIgnores();
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interceptors_mac.cpp b/compiler-rt/lib/tsan/rtl/tsan_interceptors_mac.cpp
index 2d400c7e7098..ed064150d005 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_interceptors_mac.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_interceptors_mac.cpp
@@ -365,7 +365,7 @@ static uptr GetOrCreateSyncAddress(uptr addr, ThreadState *thr, uptr pc) {
   if (h.created()) {
     ThreadIgnoreBegin(thr, pc);
     *h = (uptr) user_alloc(thr, pc, /*size=*/1);
-    ThreadIgnoreEnd(thr, pc);
+    ThreadIgnoreEnd(thr);
   }
   return *h;
 }
@@ -405,8 +405,8 @@ TSAN_INTERCEPTOR(int, swapcontext, ucontext_t *oucp, const ucontext_t *ucp) {
   {
     SCOPED_INTERCEPTOR_RAW(swapcontext, oucp, ucp);
   }
-  // Bacause of swapcontext() semantics we have no option but to copy its
-  // impementation here
+  // Because of swapcontext() semantics we have no option but to copy its
+  // implementation here
   if (!oucp || !ucp) {
     errno = EINVAL;
     return -1;
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp b/compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp
index dd2442842795..25dbe487b280 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp
@@ -90,15 +90,13 @@ DECLARE_REAL(int, pthread_mutexattr_gettype, void *, void *)
 DECLARE_REAL(int, fflush, __sanitizer_FILE *fp)
 DECLARE_REAL_AND_INTERCEPTOR(void *, malloc, uptr size)
 DECLARE_REAL_AND_INTERCEPTOR(void, free, void *ptr)
+extern "C" int pthread_equal(void *t1, void *t2);
 extern "C" void *pthread_self();
 extern "C" void _exit(int status);
 #if !SANITIZER_NETBSD
 extern "C" int fileno_unlocked(void *stream);
 extern "C" int dirfd(void *dirp);
 #endif
-#if SANITIZER_GLIBC
-extern "C" int mallopt(int param, int value);
-#endif
 #if SANITIZER_NETBSD
 extern __sanitizer_FILE __sF[];
 #else
@@ -156,12 +154,11 @@ const int SIG_SETMASK = 2;
 #endif
 
 #define COMMON_INTERCEPTOR_NOTHING_IS_INITIALIZED \
-  (cur_thread_init(), !cur_thread()->is_inited)
+  (!cur_thread_init()->is_inited)
 
 namespace __tsan {
 struct SignalDesc {
   bool armed;
-  bool sigaction;
   __sanitizer_siginfo siginfo;
   ucontext_t ctx;
 };
@@ -169,7 +166,6 @@ struct SignalDesc {
 struct ThreadSignalContext {
   int int_signal_send;
   atomic_uintptr_t in_blocking_func;
-  atomic_uintptr_t have_pending_signals;
   SignalDesc pending_signals[kSigCount];
   // emptyset and oldset are too big for stack.
   __sanitizer_sigset_t emptyset;
@@ -248,8 +244,8 @@ static ThreadSignalContext *SigCtx(ThreadState *thr) {
 
 ScopedInterceptor::ScopedInterceptor(ThreadState *thr, const char *fname,
                                      uptr pc)
-    : thr_(thr), pc_(pc), in_ignored_lib_(false), ignoring_(false) {
-  Initialize(thr);
+    : thr_(thr), in_ignored_lib_(false), ignoring_(false) {
+  LazyInitialize(thr);
   if (!thr_->is_inited) return;
   if (!thr_->ignore_interceptors) FuncEntry(thr, pc);
   DPrintf("#%d: intercept %s()\n", thr_->tid, fname);
@@ -269,25 +265,25 @@ ScopedInterceptor::~ScopedInterceptor() {
   }
 }
 
-void ScopedInterceptor::EnableIgnores() {
-  if (ignoring_) {
-    ThreadIgnoreBegin(thr_, pc_, /*save_stack=*/false);
-    if (flags()->ignore_noninstrumented_modules) thr_->suppress_reports++;
-    if (in_ignored_lib_) {
-      DCHECK(!thr_->in_ignored_lib);
-      thr_->in_ignored_lib = true;
-    }
+NOINLINE
+void ScopedInterceptor::EnableIgnoresImpl() {
+  ThreadIgnoreBegin(thr_, 0);
+  if (flags()->ignore_noninstrumented_modules)
+    thr_->suppress_reports++;
+  if (in_ignored_lib_) {
+    DCHECK(!thr_->in_ignored_lib);
+    thr_->in_ignored_lib = true;
   }
 }
 
-void ScopedInterceptor::DisableIgnores() {
-  if (ignoring_) {
-    ThreadIgnoreEnd(thr_, pc_);
-    if (flags()->ignore_noninstrumented_modules) thr_->suppress_reports--;
-    if (in_ignored_lib_) {
-      DCHECK(thr_->in_ignored_lib);
-      thr_->in_ignored_lib = false;
-    }
+NOINLINE
+void ScopedInterceptor::DisableIgnoresImpl() {
+  ThreadIgnoreEnd(thr_);
+  if (flags()->ignore_noninstrumented_modules)
+    thr_->suppress_reports--;
+  if (in_ignored_lib_) {
+    DCHECK(thr_->in_ignored_lib);
+    thr_->in_ignored_lib = false;
   }
 }
 
@@ -323,7 +319,7 @@ struct BlockingCall {
       , ctx(SigCtx(thr)) {
     for (;;) {
       atomic_store(&ctx->in_blocking_func, 1, memory_order_relaxed);
-      if (atomic_load(&ctx->have_pending_signals, memory_order_relaxed) == 0)
+      if (atomic_load(&thr->pending_signals, memory_order_relaxed) == 0)
         break;
       atomic_store(&ctx->in_blocking_func, 0, memory_order_relaxed);
       ProcessPendingSignals(thr);
@@ -385,14 +381,14 @@ static void at_exit_wrapper() {
 
   Acquire(cur_thread(), (uptr)0, (uptr)ctx);
   ((void(*)())ctx->f)();
-  InternalFree(ctx);
+  Free(ctx);
 }
 
 static void cxa_at_exit_wrapper(void *arg) {
   Acquire(cur_thread(), 0, (uptr)arg);
   AtExitCtx *ctx = (AtExitCtx*)arg;
   ((void(*)(void *arg))ctx->f)(ctx->arg);
-  InternalFree(ctx);
+  Free(ctx);
 }
 
 static int setup_at_exit_wrapper(ThreadState *thr, uptr pc, void(*f)(),
@@ -418,7 +414,7 @@ TSAN_INTERCEPTOR(int, __cxa_atexit, void (*f)(void *a), void *arg, void *dso) {
 
 static int setup_at_exit_wrapper(ThreadState *thr, uptr pc, void(*f)(),
       void *arg, void *dso) {
-  AtExitCtx *ctx = (AtExitCtx*)InternalAlloc(sizeof(AtExitCtx));
+  auto *ctx = New<AtExitCtx>();
   ctx->f = f;
   ctx->arg = arg;
   Release(thr, pc, (uptr)ctx);
@@ -444,7 +440,7 @@ static int setup_at_exit_wrapper(ThreadState *thr, uptr pc, void(*f)(),
   } else {
     res = REAL(__cxa_atexit)(cxa_at_exit_wrapper, ctx, dso);
   }
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreEnd(thr);
   return res;
 }
 
@@ -455,14 +451,14 @@ static void on_exit_wrapper(int status, void *arg) {
   Acquire(thr, pc, (uptr)arg);
   AtExitCtx *ctx = (AtExitCtx*)arg;
   ((void(*)(int status, void *arg))ctx->f)(status, ctx->arg);
-  InternalFree(ctx);
+  Free(ctx);
 }
 
 TSAN_INTERCEPTOR(int, on_exit, void(*f)(int, void*), void *arg) {
   if (in_symbolizer())
     return 0;
   SCOPED_TSAN_INTERCEPTOR(on_exit, f, arg);
-  AtExitCtx *ctx = (AtExitCtx*)InternalAlloc(sizeof(AtExitCtx));
+  auto *ctx = New<AtExitCtx>();
   ctx->f = (void(*)())f;
   ctx->arg = arg;
   Release(thr, pc, (uptr)ctx);
@@ -470,7 +466,7 @@ TSAN_INTERCEPTOR(int, on_exit, void(*f)(int, void*), void *arg) {
   // because we do not see synchronization around atexit callback list.
   ThreadIgnoreBegin(thr, pc);
   int res = REAL(on_exit)(on_exit_wrapper, ctx);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreEnd(thr);
   return res;
 }
 #define TSAN_MAYBE_INTERCEPT_ON_EXIT TSAN_INTERCEPT(on_exit)
@@ -536,10 +532,7 @@ static void LongJmp(ThreadState *thr, uptr *env) {
 }
 
 // FIXME: put everything below into a common extern "C" block?
-extern "C" void __tsan_setjmp(uptr sp) {
-  cur_thread_init();
-  SetJmp(cur_thread(), sp);
-}
+extern "C" void __tsan_setjmp(uptr sp) { SetJmp(cur_thread_init(), sp); }
 
 #if SANITIZER_MAC
 TSAN_INTERCEPTOR(int, setjmp, void *env);
@@ -849,6 +842,54 @@ TSAN_INTERCEPTOR(int, posix_memalign, void **memptr, uptr align, uptr sz) {
 }
 #endif
 
+// Both __cxa_guard_acquire and pthread_once 0-initialize
+// the object initially. pthread_once does not have any
+// other ABI requirements. __cxa_guard_acquire assumes
+// that any non-0 value in the first byte means that
+// initialization is completed. Contents of the remaining
+// bytes are up to us.
+constexpr u32 kGuardInit = 0;
+constexpr u32 kGuardDone = 1;
+constexpr u32 kGuardRunning = 1 << 16;
+constexpr u32 kGuardWaiter = 1 << 17;
+
+static int guard_acquire(ThreadState *thr, uptr pc, atomic_uint32_t *g,
+                         bool blocking_hooks = true) {
+  if (blocking_hooks)
+    OnPotentiallyBlockingRegionBegin();
+  auto on_exit = at_scope_exit([blocking_hooks] {
+    if (blocking_hooks)
+      OnPotentiallyBlockingRegionEnd();
+  });
+
+  for (;;) {
+    u32 cmp = atomic_load(g, memory_order_acquire);
+    if (cmp == kGuardInit) {
+      if (atomic_compare_exchange_strong(g, &cmp, kGuardRunning,
+                                         memory_order_relaxed))
+        return 1;
+    } else if (cmp == kGuardDone) {
+      if (!thr->in_ignored_lib)
+        Acquire(thr, pc, (uptr)g);
+      return 0;
+    } else {
+      if ((cmp & kGuardWaiter) ||
+          atomic_compare_exchange_strong(g, &cmp, cmp | kGuardWaiter,
+                                         memory_order_relaxed))
+        FutexWait(g, cmp | kGuardWaiter);
+    }
+  }
+}
+
+static void guard_release(ThreadState *thr, uptr pc, atomic_uint32_t *g,
+                          u32 v) {
+  if (!thr->in_ignored_lib)
+    Release(thr, pc, (uptr)g);
+  u32 old = atomic_exchange(g, v, memory_order_release);
+  if (old & kGuardWaiter)
+    FutexWake(g, 1 << 30);
+}
+
 // __cxa_guard_acquire and friends need to be intercepted in a special way -
 // regular interceptors will break statically-linked libstdc++. Linux
 // interceptors are especially defined as weak functions (so that they don't
@@ -869,31 +910,17 @@ TSAN_INTERCEPTOR(int, posix_memalign, void **memptr, uptr align, uptr sz) {
 // Used in thread-safe function static initialization.
 STDCXX_INTERCEPTOR(int, __cxa_guard_acquire, atomic_uint32_t *g) {
   SCOPED_INTERCEPTOR_RAW(__cxa_guard_acquire, g);
-  OnPotentiallyBlockingRegionBegin();
-  auto on_exit = at_scope_exit(&OnPotentiallyBlockingRegionEnd);
-  for (;;) {
-    u32 cmp = atomic_load(g, memory_order_acquire);
-    if (cmp == 0) {
-      if (atomic_compare_exchange_strong(g, &cmp, 1<<16, memory_order_relaxed))
-        return 1;
-    } else if (cmp == 1) {
-      Acquire(thr, pc, (uptr)g);
-      return 0;
-    } else {
-      internal_sched_yield();
-    }
-  }
+  return guard_acquire(thr, pc, g);
 }
 
 STDCXX_INTERCEPTOR(void, __cxa_guard_release, atomic_uint32_t *g) {
   SCOPED_INTERCEPTOR_RAW(__cxa_guard_release, g);
-  Release(thr, pc, (uptr)g);
-  atomic_store(g, 1, memory_order_release);
+  guard_release(thr, pc, g, kGuardDone);
 }
 
 STDCXX_INTERCEPTOR(void, __cxa_guard_abort, atomic_uint32_t *g) {
   SCOPED_INTERCEPTOR_RAW(__cxa_guard_abort, g);
-  atomic_store(g, 0, memory_order_relaxed);
+  guard_release(thr, pc, g, kGuardInit);
 }
 
 namespace __tsan {
@@ -935,17 +962,17 @@ static void thread_finalize(void *v) {
 struct ThreadParam {
   void* (*callback)(void *arg);
   void *param;
-  atomic_uintptr_t tid;
+  Tid tid;
+  Semaphore created;
+  Semaphore started;
 };
 
 extern "C" void *__tsan_thread_start_func(void *arg) {
   ThreadParam *p = (ThreadParam*)arg;
   void* (*callback)(void *arg) = p->callback;
   void *param = p->param;
-  int tid = 0;
   {
-    cur_thread_init();
-    ThreadState *thr = cur_thread();
+    ThreadState *thr = cur_thread_init();
     // Thread-local state is not initialized yet.
     ScopedIgnoreInterceptors ignore;
 #if !SANITIZER_MAC && !SANITIZER_NETBSD && !SANITIZER_FREEBSD
@@ -955,14 +982,13 @@ extern "C" void *__tsan_thread_start_func(void *arg) {
       Printf("ThreadSanitizer: failed to set thread key\n");
       Die();
     }
-    ThreadIgnoreEnd(thr, 0);
+    ThreadIgnoreEnd(thr);
 #endif
-    while ((tid = atomic_load(&p->tid, memory_order_acquire)) == 0)
-      internal_sched_yield();
+    p->created.Wait();
     Processor *proc = ProcCreate();
     ProcWire(proc, thr);
-    ThreadStart(thr, tid, GetTid(), ThreadType::Regular);
-    atomic_store(&p->tid, 0, memory_order_release);
+    ThreadStart(thr, p->tid, GetTid(), ThreadType::Regular);
+    p->started.Post();
   }
   void *res = callback(param);
   // Prevent the callback from being tail called,
@@ -984,9 +1010,11 @@ TSAN_INTERCEPTOR(int, pthread_create,
           "fork is not supported. Dying (set die_after_fork=0 to override)\n");
       Die();
     } else {
-      VPrintf(1, "ThreadSanitizer: starting new threads after multi-threaded "
-          "fork is not supported (pid %d). Continuing because of "
-          "die_after_fork=0, but you are on your own\n", internal_getpid());
+      VPrintf(1,
+              "ThreadSanitizer: starting new threads after multi-threaded "
+              "fork is not supported (pid %lu). Continuing because of "
+              "die_after_fork=0, but you are on your own\n",
+              internal_getpid());
     }
   }
   __sanitizer_pthread_attr_t myattr;
@@ -1001,18 +1029,18 @@ TSAN_INTERCEPTOR(int, pthread_create,
   ThreadParam p;
   p.callback = callback;
   p.param = param;
-  atomic_store(&p.tid, 0, memory_order_relaxed);
+  p.tid = kMainTid;
   int res = -1;
   {
     // Otherwise we see false positives in pthread stack manipulation.
     ScopedIgnoreInterceptors ignore;
     ThreadIgnoreBegin(thr, pc);
     res = REAL(pthread_create)(th, attr, __tsan_thread_start_func, &p);
-    ThreadIgnoreEnd(thr, pc);
+    ThreadIgnoreEnd(thr);
   }
   if (res == 0) {
-    int tid = ThreadCreate(thr, pc, *(uptr*)th, IsStateDetached(detached));
-    CHECK_NE(tid, 0);
+    p.tid = ThreadCreate(thr, pc, *(uptr *)th, IsStateDetached(detached));
+    CHECK_NE(p.tid, kMainTid);
     // Synchronization on p.tid serves two purposes:
     // 1. ThreadCreate must finish before the new thread starts.
     //    Otherwise the new thread can call pthread_detach, but the pthread_t
@@ -1020,9 +1048,8 @@ TSAN_INTERCEPTOR(int, pthread_create,
     // 2. ThreadStart must finish before this thread continues.
     //    Otherwise, this thread can call pthread_detach and reset thr->sync
     //    before the new thread got a chance to acquire from it in ThreadStart.
-    atomic_store(&p.tid, tid, memory_order_release);
-    while (atomic_load(&p.tid, memory_order_acquire) != 0)
-      internal_sched_yield();
+    p.created.Post();
+    p.started.Wait();
   }
   if (attr == &myattr)
     pthread_attr_destroy(&myattr);
@@ -1031,10 +1058,10 @@ TSAN_INTERCEPTOR(int, pthread_create,
 
 TSAN_INTERCEPTOR(int, pthread_join, void *th, void **ret) {
   SCOPED_INTERCEPTOR_RAW(pthread_join, th, ret);
-  int tid = ThreadConsumeTid(thr, pc, (uptr)th);
+  Tid tid = ThreadConsumeTid(thr, pc, (uptr)th);
   ThreadIgnoreBegin(thr, pc);
   int res = BLOCK_REAL(pthread_join)(th, ret);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreEnd(thr);
   if (res == 0) {
     ThreadJoin(thr, pc, tid);
   }
@@ -1045,7 +1072,7 @@ DEFINE_REAL_PTHREAD_FUNCTIONS
 
 TSAN_INTERCEPTOR(int, pthread_detach, void *th) {
   SCOPED_INTERCEPTOR_RAW(pthread_detach, th);
-  int tid = ThreadConsumeTid(thr, pc, (uptr)th);
+  Tid tid = ThreadConsumeTid(thr, pc, (uptr)th);
   int res = REAL(pthread_detach)(th);
   if (res == 0) {
     ThreadDetach(thr, pc, tid);
@@ -1066,10 +1093,10 @@ TSAN_INTERCEPTOR(void, pthread_exit, void *retval) {
 #if SANITIZER_LINUX
 TSAN_INTERCEPTOR(int, pthread_tryjoin_np, void *th, void **ret) {
   SCOPED_INTERCEPTOR_RAW(pthread_tryjoin_np, th, ret);
-  int tid = ThreadConsumeTid(thr, pc, (uptr)th);
+  Tid tid = ThreadConsumeTid(thr, pc, (uptr)th);
   ThreadIgnoreBegin(thr, pc);
   int res = REAL(pthread_tryjoin_np)(th, ret);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreEnd(thr);
   if (res == 0)
     ThreadJoin(thr, pc, tid);
   else
@@ -1080,10 +1107,10 @@ TSAN_INTERCEPTOR(int, pthread_tryjoin_np, void *th, void **ret) {
 TSAN_INTERCEPTOR(int, pthread_timedjoin_np, void *th, void **ret,
                  const struct timespec *abstime) {
   SCOPED_INTERCEPTOR_RAW(pthread_timedjoin_np, th, ret, abstime);
-  int tid = ThreadConsumeTid(thr, pc, (uptr)th);
+  Tid tid = ThreadConsumeTid(thr, pc, (uptr)th);
   ThreadIgnoreBegin(thr, pc);
   int res = BLOCK_REAL(pthread_timedjoin_np)(th, ret, abstime);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreEnd(thr);
   if (res == 0)
     ThreadJoin(thr, pc, tid);
   else
@@ -1447,14 +1474,14 @@ TSAN_INTERCEPTOR(int, pthread_rwlock_unlock, void *m) {
 #if !SANITIZER_MAC
 TSAN_INTERCEPTOR(int, pthread_barrier_init, void *b, void *a, unsigned count) {
   SCOPED_TSAN_INTERCEPTOR(pthread_barrier_init, b, a, count);
-  MemoryWrite(thr, pc, (uptr)b, kSizeLog1);
+  MemoryAccess(thr, pc, (uptr)b, 1, kAccessWrite);
   int res = REAL(pthread_barrier_init)(b, a, count);
   return res;
 }
 
 TSAN_INTERCEPTOR(int, pthread_barrier_destroy, void *b) {
   SCOPED_TSAN_INTERCEPTOR(pthread_barrier_destroy, b);
-  MemoryWrite(thr, pc, (uptr)b, kSizeLog1);
+  MemoryAccess(thr, pc, (uptr)b, 1, kAccessWrite);
   int res = REAL(pthread_barrier_destroy)(b);
   return res;
 }
@@ -1462,9 +1489,9 @@ TSAN_INTERCEPTOR(int, pthread_barrier_destroy, void *b) {
 TSAN_INTERCEPTOR(int, pthread_barrier_wait, void *b) {
   SCOPED_TSAN_INTERCEPTOR(pthread_barrier_wait, b);
   Release(thr, pc, (uptr)b);
-  MemoryRead(thr, pc, (uptr)b, kSizeLog1);
+  MemoryAccess(thr, pc, (uptr)b, 1, kAccessRead);
   int res = REAL(pthread_barrier_wait)(b);
-  MemoryRead(thr, pc, (uptr)b, kSizeLog1);
+  MemoryAccess(thr, pc, (uptr)b, 1, kAccessRead);
   if (res == 0 || res == PTHREAD_BARRIER_SERIAL_THREAD) {
     Acquire(thr, pc, (uptr)b);
   }
@@ -1486,20 +1513,11 @@ TSAN_INTERCEPTOR(int, pthread_once, void *o, void (*f)()) {
   else
     a = static_cast<atomic_uint32_t*>(o);
 
-  u32 v = atomic_load(a, memory_order_acquire);
-  if (v == 0 && atomic_compare_exchange_strong(a, &v, 1,
-                                               memory_order_relaxed)) {
+  // Mac OS X appears to use pthread_once() where calling BlockingRegion hooks
+  // result in crashes due to too little stack space.
+  if (guard_acquire(thr, pc, a, !SANITIZER_MAC)) {
     (*f)();
-    if (!thr->in_ignored_lib)
-      Release(thr, pc, (uptr)o);
-    atomic_store(a, 2, memory_order_release);
-  } else {
-    while (v != 2) {
-      internal_sched_yield();
-      v = atomic_load(a, memory_order_acquire);
-    }
-    if (!thr->in_ignored_lib)
-      Acquire(thr, pc, (uptr)o);
+    guard_release(thr, pc, a, kGuardDone);
   }
   return 0;
 }
@@ -1933,24 +1951,45 @@ TSAN_INTERCEPTOR(int, pthread_sigmask, int how, const __sanitizer_sigset_t *set,
 
 namespace __tsan {
 
+static void ReportErrnoSpoiling(ThreadState *thr, uptr pc) {
+  VarSizeStackTrace stack;
+  // StackTrace::GetNestInstructionPc(pc) is used because return address is
+  // expected, OutputReport() will undo this.
+  ObtainCurrentStack(thr, StackTrace::GetNextInstructionPc(pc), &stack);
+  ThreadRegistryLock l(&ctx->thread_registry);
+  ScopedReport rep(ReportTypeErrnoInSignal);
+  if (!IsFiredSuppression(ctx, ReportTypeErrnoInSignal, stack)) {
+    rep.AddStack(stack, true);
+    OutputReport(thr, rep);
+  }
+}
+
 static void CallUserSignalHandler(ThreadState *thr, bool sync, bool acquire,
-                                  bool sigact, int sig,
-                                  __sanitizer_siginfo *info, void *uctx) {
+                                  int sig, __sanitizer_siginfo *info,
+                                  void *uctx) {
   __sanitizer_sigaction *sigactions = interceptor_ctx()->sigactions;
   if (acquire)
     Acquire(thr, 0, (uptr)&sigactions[sig]);
   // Signals are generally asynchronous, so if we receive a signals when
   // ignores are enabled we should disable ignores. This is critical for sync
-  // and interceptors, because otherwise we can miss syncronization and report
+  // and interceptors, because otherwise we can miss synchronization and report
   // false races.
   int ignore_reads_and_writes = thr->ignore_reads_and_writes;
   int ignore_interceptors = thr->ignore_interceptors;
   int ignore_sync = thr->ignore_sync;
+  // For symbolizer we only process SIGSEGVs synchronously
+  // (bug in symbolizer or in tsan). But we want to reset
+  // in_symbolizer to fail gracefully. Symbolizer and user code
+  // use different memory allocators, so if we don't reset
+  // in_symbolizer we can get memory allocated with one being
+  // feed with another, which can cause more crashes.
+  int in_symbolizer = thr->in_symbolizer;
   if (!ctx->after_multithreaded_fork) {
     thr->ignore_reads_and_writes = 0;
     thr->fast_state.ClearIgnoreBit();
     thr->ignore_interceptors = 0;
     thr->ignore_sync = 0;
+    thr->in_symbolizer = 0;
   }
   // Ensure that the handler does not spoil errno.
   const int saved_errno = errno;
@@ -1958,13 +1997,14 @@ static void CallUserSignalHandler(ThreadState *thr, bool sync, bool acquire,
   // This code races with sigaction. Be careful to not read sa_sigaction twice.
   // Also need to remember pc for reporting before the call,
   // because the handler can reset it.
-  volatile uptr pc =
-      sigact ? (uptr)sigactions[sig].sigaction : (uptr)sigactions[sig].handler;
+  volatile uptr pc = (sigactions[sig].sa_flags & SA_SIGINFO)
+                         ? (uptr)sigactions[sig].sigaction
+                         : (uptr)sigactions[sig].handler;
   if (pc != sig_dfl && pc != sig_ign) {
-    if (sigact)
-      ((__sanitizer_sigactionhandler_ptr)pc)(sig, info, uctx);
-    else
-      ((__sanitizer_sighandler_ptr)pc)(sig);
+    // The callback can be either sa_handler or sa_sigaction.
+    // They have different signatures, but we assume that passing
+    // additional arguments to sa_handler works and is harmless.
+    ((__sanitizer_sigactionhandler_ptr)pc)(sig, info, uctx);
   }
   if (!ctx->after_multithreaded_fork) {
     thr->ignore_reads_and_writes = ignore_reads_and_writes;
@@ -1972,6 +2012,7 @@ static void CallUserSignalHandler(ThreadState *thr, bool sync, bool acquire,
       thr->fast_state.SetIgnoreBit();
     thr->ignore_interceptors = ignore_interceptors;
     thr->ignore_sync = ignore_sync;
+    thr->in_symbolizer = in_symbolizer;
   }
   // We do not detect errno spoiling for SIGTERM,
   // because some SIGTERM handlers do spoil errno but reraise SIGTERM,
@@ -1981,27 +2022,16 @@ static void CallUserSignalHandler(ThreadState *thr, bool sync, bool acquire,
   // from rtl_generic_sighandler) we have not yet received the reraised
   // signal; and it looks too fragile to intercept all ways to reraise a signal.
   if (ShouldReport(thr, ReportTypeErrnoInSignal) && !sync && sig != SIGTERM &&
-      errno != 99) {
-    VarSizeStackTrace stack;
-    // StackTrace::GetNestInstructionPc(pc) is used because return address is
-    // expected, OutputReport() will undo this.
-    ObtainCurrentStack(thr, StackTrace::GetNextInstructionPc(pc), &stack);
-    ThreadRegistryLock l(ctx->thread_registry);
-    ScopedReport rep(ReportTypeErrnoInSignal);
-    if (!IsFiredSuppression(ctx, ReportTypeErrnoInSignal, stack)) {
-      rep.AddStack(stack, true);
-      OutputReport(thr, rep);
-    }
-  }
+      errno != 99)
+    ReportErrnoSpoiling(thr, pc);
   errno = saved_errno;
 }
 
-void ProcessPendingSignals(ThreadState *thr) {
+void ProcessPendingSignalsImpl(ThreadState *thr) {
+  atomic_store(&thr->pending_signals, 0, memory_order_relaxed);
   ThreadSignalContext *sctx = SigCtx(thr);
-  if (sctx == 0 ||
-      atomic_load(&sctx->have_pending_signals, memory_order_relaxed) == 0)
+  if (sctx == 0)
     return;
-  atomic_store(&sctx->have_pending_signals, 0, memory_order_relaxed);
   atomic_fetch_add(&thr->in_signal_handler, 1, memory_order_relaxed);
   internal_sigfillset(&sctx->emptyset);
   int res = REAL(pthread_sigmask)(SIG_SETMASK, &sctx->emptyset, &sctx->oldset);
@@ -2010,8 +2040,8 @@ void ProcessPendingSignals(ThreadState *thr) {
     SignalDesc *signal = &sctx->pending_signals[sig];
     if (signal->armed) {
       signal->armed = false;
-      CallUserSignalHandler(thr, false, true, signal->sigaction, sig,
-          &signal->siginfo, &signal->ctx);
+      CallUserSignalHandler(thr, false, true, sig, &signal->siginfo,
+                            &signal->ctx);
     }
   }
   res = REAL(pthread_sigmask)(SIG_SETMASK, &sctx->oldset, 0);
@@ -2028,11 +2058,8 @@ static bool is_sync_signal(ThreadSignalContext *sctx, int sig) {
          (sctx && sig == sctx->int_signal_send);
 }
 
-void ALWAYS_INLINE rtl_generic_sighandler(bool sigact, int sig,
-                                          __sanitizer_siginfo *info,
-                                          void *ctx) {
-  cur_thread_init();
-  ThreadState *thr = cur_thread();
+void sighandler(int sig, __sanitizer_siginfo *info, void *ctx) {
+  ThreadState *thr = cur_thread_init();
   ThreadSignalContext *sctx = SigCtx(thr);
   if (sig < 0 || sig >= kSigCount) {
     VPrintf(1, "ThreadSanitizer: ignoring signal %d\n", sig);
@@ -2048,7 +2075,7 @@ void ALWAYS_INLINE rtl_generic_sighandler(bool sigact, int sig,
     atomic_fetch_add(&thr->in_signal_handler, 1, memory_order_relaxed);
     if (sctx && atomic_load(&sctx->in_blocking_func, memory_order_relaxed)) {
       atomic_store(&sctx->in_blocking_func, 0, memory_order_relaxed);
-      CallUserSignalHandler(thr, sync, true, sigact, sig, info, ctx);
+      CallUserSignalHandler(thr, sync, true, sig, info, ctx);
       atomic_store(&sctx->in_blocking_func, 1, memory_order_relaxed);
     } else {
       // Be very conservative with when we do acquire in this case.
@@ -2057,7 +2084,7 @@ void ALWAYS_INLINE rtl_generic_sighandler(bool sigact, int sig,
       // SIGSYS looks relatively safe -- it's synchronous and can actually
       // need some global state.
       bool acq = (sig == SIGSYS);
-      CallUserSignalHandler(thr, sync, acq, sigact, sig, info, ctx);
+      CallUserSignalHandler(thr, sync, acq, sig, info, ctx);
     }
     atomic_fetch_add(&thr->in_signal_handler, -1, memory_order_relaxed);
     return;
@@ -2068,23 +2095,12 @@ void ALWAYS_INLINE rtl_generic_sighandler(bool sigact, int sig,
   SignalDesc *signal = &sctx->pending_signals[sig];
   if (signal->armed == false) {
     signal->armed = true;
-    signal->sigaction = sigact;
-    if (info)
-      internal_memcpy(&signal->siginfo, info, sizeof(*info));
-    if (ctx)
-      internal_memcpy(&signal->ctx, ctx, sizeof(signal->ctx));
-    atomic_store(&sctx->have_pending_signals, 1, memory_order_relaxed);
+    internal_memcpy(&signal->siginfo, info, sizeof(*info));
+    internal_memcpy(&signal->ctx, ctx, sizeof(signal->ctx));
+    atomic_store(&thr->pending_signals, 1, memory_order_relaxed);
   }
 }
 
-static void rtl_sighandler(int sig) {
-  rtl_generic_sighandler(false, sig, 0, 0);
-}
-
-static void rtl_sigaction(int sig, __sanitizer_siginfo *info, void *ctx) {
-  rtl_generic_sighandler(true, sig, info, ctx);
-}
-
 TSAN_INTERCEPTOR(int, raise, int sig) {
   SCOPED_TSAN_INTERCEPTOR(raise, sig);
   ThreadSignalContext *sctx = SigCtx(thr);
@@ -2118,11 +2134,11 @@ TSAN_INTERCEPTOR(int, pthread_kill, void *tid, int sig) {
   ThreadSignalContext *sctx = SigCtx(thr);
   CHECK_NE(sctx, 0);
   int prev = sctx->int_signal_send;
-  if (tid == pthread_self()) {
+  bool self = pthread_equal(tid, pthread_self());
+  if (self)
     sctx->int_signal_send = sig;
-  }
   int res = REAL(pthread_kill)(tid, sig);
-  if (tid == pthread_self()) {
+  if (self) {
     CHECK_EQ(sctx->int_signal_send, sig);
     sctx->int_signal_send = prev;
   }
@@ -2143,7 +2159,7 @@ TSAN_INTERCEPTOR(int, getaddrinfo, void *node, void *service,
   // inside of getaddrinfo. So ignore memory accesses.
   ThreadIgnoreBegin(thr, pc);
   int res = REAL(getaddrinfo)(node, service, hints, rv);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreEnd(thr);
   return res;
 }
 
@@ -2175,7 +2191,7 @@ void atfork_child() {
     return;
   ThreadState *thr = cur_thread();
   const uptr pc = StackTrace::GetCurrentPc();
-  ForkChildAfter(thr, pc);
+  ForkChildAfter(thr, pc, true);
   FdOnFork(thr, pc);
 }
 
@@ -2196,6 +2212,37 @@ TSAN_INTERCEPTOR(int, vfork, int fake) {
   return WRAP(fork)(fake);
 }
 
+#if SANITIZER_LINUX
+TSAN_INTERCEPTOR(int, clone, int (*fn)(void *), void *stack, int flags,
+                 void *arg, int *parent_tid, void *tls, pid_t *child_tid) {
+  SCOPED_INTERCEPTOR_RAW(clone, fn, stack, flags, arg, parent_tid, tls,
+                         child_tid);
+  struct Arg {
+    int (*fn)(void *);
+    void *arg;
+  };
+  auto wrapper = +[](void *p) -> int {
+    auto *thr = cur_thread();
+    uptr pc = GET_CURRENT_PC();
+    // Start the background thread for fork, but not for clone.
+    // For fork we did this always and it's known to work (or user code has
+    // adopted). But if we do this for the new clone interceptor some code
+    // (sandbox2) fails. So model we used to do for years and don't start the
+    // background thread after clone.
+    ForkChildAfter(thr, pc, false);
+    FdOnFork(thr, pc);
+    auto *arg = static_cast<Arg *>(p);
+    return arg->fn(arg->arg);
+  };
+  ForkBefore(thr, pc);
+  Arg arg_wrapper = {fn, arg};
+  int pid = REAL(clone)(wrapper, stack, flags, &arg_wrapper, parent_tid, tls,
+                        child_tid);
+  ForkParentAfter(thr, pc);
+  return pid;
+}
+#endif
+
 #if !SANITIZER_MAC && !SANITIZER_ANDROID
 typedef int (*dl_iterate_phdr_cb_t)(__sanitizer_dl_phdr_info *info, SIZE_T size,
                                     void *data);
@@ -2207,7 +2254,7 @@ struct dl_iterate_phdr_data {
 };
 
 static bool IsAppNotRodata(uptr addr) {
-  return IsAppMem(addr) && *(u64*)MemToShadow(addr) != kShadowRodata;
+  return IsAppMem(addr) && *MemToShadow(addr) != kShadowRodata;
 }
 
 static int dl_iterate_phdr_cb(__sanitizer_dl_phdr_info *info, SIZE_T size,
@@ -2250,7 +2297,6 @@ static int OnExit(ThreadState *thr) {
 
 struct TsanInterceptorContext {
   ThreadState *thr;
-  const uptr caller_pc;
   const uptr pc;
 };
 
@@ -2291,17 +2337,17 @@ static void HandleRecvmsg(ThreadState *thr, uptr pc,
                     ((TsanInterceptorContext *) ctx)->pc, (uptr) ptr, size, \
                     false)
 
-#define COMMON_INTERCEPTOR_ENTER(ctx, func, ...)      \
-  SCOPED_TSAN_INTERCEPTOR(func, __VA_ARGS__);         \
-  TsanInterceptorContext _ctx = {thr, caller_pc, pc}; \
-  ctx = (void *)&_ctx;                                \
-  (void) ctx;
+#define COMMON_INTERCEPTOR_ENTER(ctx, func, ...) \
+  SCOPED_TSAN_INTERCEPTOR(func, __VA_ARGS__);    \
+  TsanInterceptorContext _ctx = {thr, pc};       \
+  ctx = (void *)&_ctx;                           \
+  (void)ctx;
 
 #define COMMON_INTERCEPTOR_ENTER_NOIGNORE(ctx, func, ...) \
   SCOPED_INTERCEPTOR_RAW(func, __VA_ARGS__);              \
-  TsanInterceptorContext _ctx = {thr, caller_pc, pc};     \
+  TsanInterceptorContext _ctx = {thr, pc};                \
   ctx = (void *)&_ctx;                                    \
-  (void) ctx;
+  (void)ctx;
 
 #define COMMON_INTERCEPTOR_FILE_OPEN(ctx, file, path) \
   if (path)                                           \
@@ -2347,8 +2393,11 @@ static void HandleRecvmsg(ThreadState *thr, uptr pc,
 #define COMMON_INTERCEPTOR_SET_THREAD_NAME(ctx, name) \
   ThreadSetName(((TsanInterceptorContext *) ctx)->thr, name)
 
-#define COMMON_INTERCEPTOR_SET_PTHREAD_NAME(ctx, thread, name) \
-  __tsan::ctx->thread_registry->SetThreadNameByUserId(thread, name)
+#define COMMON_INTERCEPTOR_SET_PTHREAD_NAME(ctx, thread, name)         \
+  if (pthread_equal(pthread_self(), reinterpret_cast<void *>(thread))) \
+    COMMON_INTERCEPTOR_SET_THREAD_NAME(ctx, name);                     \
+  else                                                                 \
+    __tsan::ctx->thread_registry.SetThreadNameByUserId(thread, name)
 
 #define COMMON_INTERCEPTOR_BLOCK_REAL(name) BLOCK_REAL(name)
 
@@ -2420,7 +2469,7 @@ static __sanitizer_sighandler_ptr signal_impl(int sig,
 int sigaction_impl(int sig, const __sanitizer_sigaction *act,
                    __sanitizer_sigaction *old) {
   // Note: if we call REAL(sigaction) directly for any reason without proxying
-  // the signal handler through rtl_sigaction, very bad things will happen.
+  // the signal handler through sighandler, very bad things will happen.
   // The handler will run synchronously and corrupt tsan per-thread state.
   SCOPED_INTERCEPTOR_RAW(sigaction, sig, act, old);
   if (sig <= 0 || sig >= kSigCount) {
@@ -2448,22 +2497,17 @@ int sigaction_impl(int sig, const __sanitizer_sigaction *act,
 #endif
     internal_memcpy(&newact, act, sizeof(newact));
     internal_sigfillset(&newact.sa_mask);
-    if ((uptr)act->handler != sig_ign && (uptr)act->handler != sig_dfl) {
-      if (newact.sa_flags & SA_SIGINFO)
-        newact.sigaction = rtl_sigaction;
-      else
-        newact.handler = rtl_sighandler;
+    if ((act->sa_flags & SA_SIGINFO) ||
+        ((uptr)act->handler != sig_ign && (uptr)act->handler != sig_dfl)) {
+      newact.sa_flags |= SA_SIGINFO;
+      newact.sigaction = sighandler;
     }
     ReleaseStore(thr, pc, (uptr)&sigactions[sig]);
     act = &newact;
   }
   int res = REAL(sigaction)(sig, act, old);
-  if (res == 0 && old) {
-    uptr cb = (uptr)old->sigaction;
-    if (cb == (uptr)rtl_sigaction || cb == (uptr)rtl_sighandler) {
-      internal_memcpy(old, &old_stored, sizeof(*old));
-    }
-  }
+  if (res == 0 && old && old->sigaction == sighandler)
+    internal_memcpy(old, &old_stored, sizeof(*old));
   return res;
 }
 
@@ -2479,20 +2523,16 @@ static __sanitizer_sighandler_ptr signal_impl(int sig,
   return old.handler;
 }
 
-#define TSAN_SYSCALL() \
+#define TSAN_SYSCALL()             \
   ThreadState *thr = cur_thread(); \
-  if (thr->ignore_interceptors) \
-    return; \
-  ScopedSyscall scoped_syscall(thr) \
-/**/
+  if (thr->ignore_interceptors)    \
+    return;                        \
+  ScopedSyscall scoped_syscall(thr)
 
 struct ScopedSyscall {
   ThreadState *thr;
 
-  explicit ScopedSyscall(ThreadState *thr)
-      : thr(thr) {
-    Initialize(thr);
-  }
+  explicit ScopedSyscall(ThreadState *thr) : thr(thr) { LazyInitialize(thr); }
 
   ~ScopedSyscall() {
     ProcessPendingSignals(thr);
@@ -2508,12 +2548,12 @@ static void syscall_access_range(uptr pc, uptr p, uptr s, bool write) {
 static USED void syscall_acquire(uptr pc, uptr addr) {
   TSAN_SYSCALL();
   Acquire(thr, pc, addr);
-  DPrintf("syscall_acquire(%p)\n", addr);
+  DPrintf("syscall_acquire(0x%zx))\n", addr);
 }
 
 static USED void syscall_release(uptr pc, uptr addr) {
   TSAN_SYSCALL();
-  DPrintf("syscall_release(%p)\n", addr);
+  DPrintf("syscall_release(0x%zx)\n", addr);
   Release(thr, pc, addr);
 }
 
@@ -2525,12 +2565,12 @@ static void syscall_fd_close(uptr pc, int fd) {
 static USED void syscall_fd_acquire(uptr pc, int fd) {
   TSAN_SYSCALL();
   FdAcquire(thr, pc, fd);
-  DPrintf("syscall_fd_acquire(%p)\n", fd);
+  DPrintf("syscall_fd_acquire(%d)\n", fd);
 }
 
 static USED void syscall_fd_release(uptr pc, int fd) {
   TSAN_SYSCALL();
-  DPrintf("syscall_fd_release(%p)\n", fd);
+  DPrintf("syscall_fd_release(%d)\n", fd);
   FdRelease(thr, pc, fd);
 }
 
@@ -2540,7 +2580,7 @@ static void syscall_post_fork(uptr pc, int pid) {
   ThreadState *thr = cur_thread();
   if (pid == 0) {
     // child
-    ForkChildAfter(thr, pc);
+    ForkChildAfter(thr, pc, true);
     FdOnFork(thr, pc);
   } else if (pid > 0) {
     // parent
@@ -2700,12 +2740,6 @@ void InitializeInterceptors() {
   REAL(memcpy) = internal_memcpy;
 #endif
 
-  // Instruct libc malloc to consume less memory.
-#if SANITIZER_GLIBC
-  mallopt(1, 0);  // M_MXFAST
-  mallopt(-3, 32*1024);  // M_MMAP_THRESHOLD
-#endif
-
   new(interceptor_ctx()) InterceptorContext();
 
   InitializeCommonInterceptors();
@@ -2843,6 +2877,9 @@ void InitializeInterceptors() {
 
   TSAN_INTERCEPT(fork);
   TSAN_INTERCEPT(vfork);
+#if SANITIZER_LINUX
+  TSAN_INTERCEPT(clone);
+#endif
 #if !SANITIZER_ANDROID
   TSAN_INTERCEPT(dl_iterate_phdr);
 #endif
@@ -2920,25 +2957,36 @@ void InitializeInterceptors() {
 // Note that no_sanitize_thread attribute does not turn off atomic interception
 // so attaching it to the function defined in user code does not help.
 // That's why we now have what we have.
-extern "C" SANITIZER_INTERFACE_ATTRIBUTE
-void __tsan_testonly_barrier_init(u64 *barrier, u32 count) {
-  if (count >= (1 << 8)) {
-      Printf("barrier_init: count is too large (%d)\n", count);
-      Die();
+constexpr u32 kBarrierThreadBits = 10;
+constexpr u32 kBarrierThreads = 1 << kBarrierThreadBits;
+
+extern "C" SANITIZER_INTERFACE_ATTRIBUTE void __tsan_testonly_barrier_init(
+    atomic_uint32_t *barrier, u32 num_threads) {
+  if (num_threads >= kBarrierThreads) {
+    Printf("barrier_init: count is too large (%d)\n", num_threads);
+    Die();
   }
-  // 8 lsb is thread count, the remaining are count of entered threads.
-  *barrier = count;
+  // kBarrierThreadBits lsb is thread count,
+  // the remaining are count of entered threads.
+  atomic_store(barrier, num_threads, memory_order_relaxed);
 }
 
-extern "C" SANITIZER_INTERFACE_ATTRIBUTE
-void __tsan_testonly_barrier_wait(u64 *barrier) {
-  unsigned old = __atomic_fetch_add(barrier, 1 << 8, __ATOMIC_RELAXED);
-  unsigned old_epoch = (old >> 8) / (old & 0xff);
+static u32 barrier_epoch(u32 value) {
+  return (value >> kBarrierThreadBits) / (value & (kBarrierThreads - 1));
+}
+
+extern "C" SANITIZER_INTERFACE_ATTRIBUTE void __tsan_testonly_barrier_wait(
+    atomic_uint32_t *barrier) {
+  u32 old = atomic_fetch_add(barrier, kBarrierThreads, memory_order_relaxed);
+  u32 old_epoch = barrier_epoch(old);
+  if (barrier_epoch(old + kBarrierThreads) != old_epoch) {
+    FutexWake(barrier, (1 << 30));
+    return;
+  }
   for (;;) {
-    unsigned cur = __atomic_load_n(barrier, __ATOMIC_RELAXED);
-    unsigned cur_epoch = (cur >> 8) / (cur & 0xff);
-    if (cur_epoch != old_epoch)
+    u32 cur = atomic_load(barrier, memory_order_relaxed);
+    if (barrier_epoch(cur) != old_epoch)
       return;
-    internal_sched_yield();
+    FutexWait(barrier, cur);
   }
 }
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interface.cpp b/compiler-rt/lib/tsan/rtl/tsan_interface.cpp
index 9bd0e8580b17..048715185151 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_interface.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_interface.cpp
@@ -20,109 +20,58 @@
 
 using namespace __tsan;
 
-void __tsan_init() {
-  cur_thread_init();
-  Initialize(cur_thread());
-}
+void __tsan_init() { Initialize(cur_thread_init()); }
 
 void __tsan_flush_memory() {
   FlushShadowMemory();
 }
 
 void __tsan_read16(void *addr) {
-  MemoryRead(cur_thread(), CALLERPC, (uptr)addr, kSizeLog8);
-  MemoryRead(cur_thread(), CALLERPC, (uptr)addr + 8, kSizeLog8);
+  uptr pc = CALLERPC;
+  ThreadState *thr = cur_thread();
+  MemoryAccess(thr, pc, (uptr)addr, 8, kAccessRead);
+  MemoryAccess(thr, pc, (uptr)addr + 8, 8, kAccessRead);
 }
 
 void __tsan_write16(void *addr) {
-  MemoryWrite(cur_thread(), CALLERPC, (uptr)addr, kSizeLog8);
-  MemoryWrite(cur_thread(), CALLERPC, (uptr)addr + 8, kSizeLog8);
+  uptr pc = CALLERPC;
+  ThreadState *thr = cur_thread();
+  MemoryAccess(thr, pc, (uptr)addr, 8, kAccessWrite);
+  MemoryAccess(thr, pc, (uptr)addr + 8, 8, kAccessWrite);
 }
 
 void __tsan_read16_pc(void *addr, void *pc) {
-  MemoryRead(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog8);
-  MemoryRead(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr + 8, kSizeLog8);
+  uptr pc_no_pac = STRIP_PAC_PC(pc);
+  ThreadState *thr = cur_thread();
+  MemoryAccess(thr, pc_no_pac, (uptr)addr, 8, kAccessRead);
+  MemoryAccess(thr, pc_no_pac, (uptr)addr + 8, 8, kAccessRead);
 }
 
 void __tsan_write16_pc(void *addr, void *pc) {
-  MemoryWrite(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog8);
-  MemoryWrite(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr + 8, kSizeLog8);
+  uptr pc_no_pac = STRIP_PAC_PC(pc);
+  ThreadState *thr = cur_thread();
+  MemoryAccess(thr, pc_no_pac, (uptr)addr, 8, kAccessWrite);
+  MemoryAccess(thr, pc_no_pac, (uptr)addr + 8, 8, kAccessWrite);
 }
 
 // __tsan_unaligned_read/write calls are emitted by compiler.
 
-void __tsan_unaligned_read2(const void *addr) {
-  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 2, false, false);
-}
-
-void __tsan_unaligned_read4(const void *addr) {
-  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 4, false, false);
-}
-
-void __tsan_unaligned_read8(const void *addr) {
-  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 8, false, false);
-}
-
 void __tsan_unaligned_read16(const void *addr) {
-  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 16, false, false);
-}
-
-void __tsan_unaligned_write2(void *addr) {
-  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 2, true, false);
-}
-
-void __tsan_unaligned_write4(void *addr) {
-  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 4, true, false);
-}
-
-void __tsan_unaligned_write8(void *addr) {
-  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 8, true, false);
+  uptr pc = CALLERPC;
+  ThreadState *thr = cur_thread();
+  UnalignedMemoryAccess(thr, pc, (uptr)addr, 8, kAccessRead);
+  UnalignedMemoryAccess(thr, pc, (uptr)addr + 8, 8, kAccessRead);
 }
 
 void __tsan_unaligned_write16(void *addr) {
-  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 16, true, false);
+  uptr pc = CALLERPC;
+  ThreadState *thr = cur_thread();
+  UnalignedMemoryAccess(thr, pc, (uptr)addr, 8, kAccessWrite);
+  UnalignedMemoryAccess(thr, pc, (uptr)addr + 8, 8, kAccessWrite);
 }
 
-// __sanitizer_unaligned_load/store are for user instrumentation.
-
 extern "C" {
 SANITIZER_INTERFACE_ATTRIBUTE
-u16 __sanitizer_unaligned_load16(const uu16 *addr) {
-  __tsan_unaligned_read2(addr);
-  return *addr;
-}
-
-SANITIZER_INTERFACE_ATTRIBUTE
-u32 __sanitizer_unaligned_load32(const uu32 *addr) {
-  __tsan_unaligned_read4(addr);
-  return *addr;
-}
-
-SANITIZER_INTERFACE_ATTRIBUTE
-u64 __sanitizer_unaligned_load64(const uu64 *addr) {
-  __tsan_unaligned_read8(addr);
-  return *addr;
-}
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void __sanitizer_unaligned_store16(uu16 *addr, u16 v) {
-  __tsan_unaligned_write2(addr);
-  *addr = v;
-}
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void __sanitizer_unaligned_store32(uu32 *addr, u32 v) {
-  __tsan_unaligned_write4(addr);
-  *addr = v;
-}
-
-SANITIZER_INTERFACE_ATTRIBUTE
-void __sanitizer_unaligned_store64(uu64 *addr, u64 v) {
-  __tsan_unaligned_write8(addr);
-  *addr = v;
-}
-
-SANITIZER_INTERFACE_ATTRIBUTE
 void *__tsan_get_current_fiber() {
   return cur_thread();
 }
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interface.h b/compiler-rt/lib/tsan/rtl/tsan_interface.h
index 124aa2fd2143..711f064174c2 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_interface.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_interface.h
@@ -95,9 +95,9 @@ SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_write_range(void *addr, unsigned long size);
 
 SANITIZER_INTERFACE_ATTRIBUTE
-void __tsan_read_range_pc(void *addr, unsigned long size, void *pc);  // NOLINT
+void __tsan_read_range_pc(void *addr, unsigned long size, void *pc);
 SANITIZER_INTERFACE_ATTRIBUTE
-void __tsan_write_range_pc(void *addr, unsigned long size, void *pc);  // NOLINT
+void __tsan_write_range_pc(void *addr, unsigned long size, void *pc);
 
 // User may provide function that would be called right when TSan detects
 // an error. The argument 'report' is an opaque pointer that can be used to
@@ -417,12 +417,6 @@ SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_go_atomic64_compare_exchange(ThreadState *thr, uptr cpc, uptr pc,
                                          u8 *a);
 
-SANITIZER_INTERFACE_ATTRIBUTE
-void __tsan_on_initialize();
-
-SANITIZER_INTERFACE_ATTRIBUTE
-int __tsan_on_finalize(int failed);
-
 }  // extern "C"
 
 }  // namespace __tsan
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interface.inc b/compiler-rt/lib/tsan/rtl/tsan_interface.inc
new file mode 100644
index 000000000000..0031800e851f
--- /dev/null
+++ b/compiler-rt/lib/tsan/rtl/tsan_interface.inc
@@ -0,0 +1,182 @@
+//===-- tsan_interface.inc --------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of ThreadSanitizer (TSan), a race detector.
+//
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_common/sanitizer_ptrauth.h"
+#include "tsan_interface.h"
+#include "tsan_rtl.h"
+
+#define CALLERPC ((uptr)__builtin_return_address(0))
+
+using namespace __tsan;
+
+void __tsan_read1(void *addr) {
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 1, kAccessRead);
+}
+
+void __tsan_read2(void *addr) {
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 2, kAccessRead);
+}
+
+void __tsan_read4(void *addr) {
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 4, kAccessRead);
+}
+
+void __tsan_read8(void *addr) {
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 8, kAccessRead);
+}
+
+void __tsan_write1(void *addr) {
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 1, kAccessWrite);
+}
+
+void __tsan_write2(void *addr) {
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 2, kAccessWrite);
+}
+
+void __tsan_write4(void *addr) {
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 4, kAccessWrite);
+}
+
+void __tsan_write8(void *addr) {
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 8, kAccessWrite);
+}
+
+void __tsan_read1_pc(void *addr, void *pc) {
+  MemoryAccess(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, 1, kAccessRead | kAccessExternalPC);
+}
+
+void __tsan_read2_pc(void *addr, void *pc) {
+  MemoryAccess(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, 2, kAccessRead | kAccessExternalPC);
+}
+
+void __tsan_read4_pc(void *addr, void *pc) {
+  MemoryAccess(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, 4, kAccessRead | kAccessExternalPC);
+}
+
+void __tsan_read8_pc(void *addr, void *pc) {
+  MemoryAccess(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, 8, kAccessRead | kAccessExternalPC);
+}
+
+void __tsan_write1_pc(void *addr, void *pc) {
+  MemoryAccess(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, 1, kAccessWrite | kAccessExternalPC);
+}
+
+void __tsan_write2_pc(void *addr, void *pc) {
+  MemoryAccess(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, 2, kAccessWrite | kAccessExternalPC);
+}
+
+void __tsan_write4_pc(void *addr, void *pc) {
+  MemoryAccess(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, 4, kAccessWrite | kAccessExternalPC);
+}
+
+void __tsan_write8_pc(void *addr, void *pc) {
+  MemoryAccess(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, 8, kAccessWrite | kAccessExternalPC);
+}
+
+ALWAYS_INLINE USED void __tsan_unaligned_read2(const void *addr) {
+  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 2, kAccessRead);
+}
+
+ALWAYS_INLINE USED void __tsan_unaligned_read4(const void *addr) {
+  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 4, kAccessRead);
+}
+
+ALWAYS_INLINE USED void __tsan_unaligned_read8(const void *addr) {
+  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 8, kAccessRead);
+}
+
+ALWAYS_INLINE USED void __tsan_unaligned_write2(void *addr) {
+  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 2, kAccessWrite);
+}
+
+ALWAYS_INLINE USED void __tsan_unaligned_write4(void *addr) {
+  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 4, kAccessWrite);
+}
+
+ALWAYS_INLINE USED void __tsan_unaligned_write8(void *addr) {
+  UnalignedMemoryAccess(cur_thread(), CALLERPC, (uptr)addr, 8, kAccessWrite);
+}
+
+extern "C" {
+// __sanitizer_unaligned_load/store are for user instrumentation.
+SANITIZER_INTERFACE_ATTRIBUTE
+u16 __sanitizer_unaligned_load16(const uu16 *addr) {
+  __tsan_unaligned_read2(addr);
+  return *addr;
+}
+
+SANITIZER_INTERFACE_ATTRIBUTE
+u32 __sanitizer_unaligned_load32(const uu32 *addr) {
+  __tsan_unaligned_read4(addr);
+  return *addr;
+}
+
+SANITIZER_INTERFACE_ATTRIBUTE
+u64 __sanitizer_unaligned_load64(const uu64 *addr) {
+  __tsan_unaligned_read8(addr);
+  return *addr;
+}
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __sanitizer_unaligned_store16(uu16 *addr, u16 v) {
+  *addr = v;
+  __tsan_unaligned_write2(addr);
+}
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __sanitizer_unaligned_store32(uu32 *addr, u32 v) {
+  *addr = v;
+  __tsan_unaligned_write4(addr);
+}
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __sanitizer_unaligned_store64(uu64 *addr, u64 v) {
+  *addr = v;
+  __tsan_unaligned_write8(addr);
+}
+}
+
+void __tsan_vptr_update(void **vptr_p, void *new_val) {
+  if (*vptr_p == new_val)
+    return;
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)vptr_p, sizeof(*vptr_p),
+               kAccessWrite | kAccessVptr);
+}
+
+void __tsan_vptr_read(void **vptr_p) {
+  MemoryAccess(cur_thread(), CALLERPC, (uptr)vptr_p, sizeof(*vptr_p),
+               kAccessRead | kAccessVptr);
+}
+
+void __tsan_func_entry(void *pc) { FuncEntry(cur_thread(), STRIP_PAC_PC(pc)); }
+
+void __tsan_func_exit() { FuncExit(cur_thread()); }
+
+void __tsan_ignore_thread_begin() { ThreadIgnoreBegin(cur_thread(), CALLERPC); }
+
+void __tsan_ignore_thread_end() { ThreadIgnoreEnd(cur_thread()); }
+
+void __tsan_read_range(void *addr, uptr size) {
+  MemoryAccessRange(cur_thread(), CALLERPC, (uptr)addr, size, false);
+}
+
+void __tsan_write_range(void *addr, uptr size) {
+  MemoryAccessRange(cur_thread(), CALLERPC, (uptr)addr, size, true);
+}
+
+void __tsan_read_range_pc(void *addr, uptr size, void *pc) {
+  MemoryAccessRange(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, size, false);
+}
+
+void __tsan_write_range_pc(void *addr, uptr size, void *pc) {
+  MemoryAccessRange(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, size, true);
+}
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interface_ann.cpp b/compiler-rt/lib/tsan/rtl/tsan_interface_ann.cpp
index 47314f5ad812..6bd72e18d942 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_interface_ann.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_interface_ann.cpp
@@ -43,15 +43,14 @@ class ScopedAnnotation {
   ThreadState *const thr_;
 };
 
-#define SCOPED_ANNOTATION_RET(typ, ret) \
-    if (!flags()->enable_annotations) \
-      return ret; \
-    ThreadState *thr = cur_thread(); \
-    const uptr caller_pc = (uptr)__builtin_return_address(0); \
-    ScopedAnnotation sa(thr, __func__, caller_pc); \
-    const uptr pc = StackTrace::GetCurrentPc(); \
-    (void)pc; \
-/**/
+#define SCOPED_ANNOTATION_RET(typ, ret)                     \
+  if (!flags()->enable_annotations)                         \
+    return ret;                                             \
+  ThreadState *thr = cur_thread();                          \
+  const uptr caller_pc = (uptr)__builtin_return_address(0); \
+  ScopedAnnotation sa(thr, __func__, caller_pc);            \
+  const uptr pc = StackTrace::GetCurrentPc();               \
+  (void)pc;
 
 #define SCOPED_ANNOTATION(typ) SCOPED_ANNOTATION_RET(typ, )
 
@@ -71,7 +70,6 @@ struct ExpectRace {
 
 struct DynamicAnnContext {
   Mutex mtx;
-  ExpectRace expect;
   ExpectRace benign;
 
   DynamicAnnContext() : mtx(MutexTypeAnnotations) {}
@@ -90,7 +88,7 @@ static void AddExpectRace(ExpectRace *list,
       return;
     }
   }
-  race = (ExpectRace*)internal_alloc(MBlockExpectRace, sizeof(ExpectRace));
+  race = static_cast<ExpectRace *>(Alloc(sizeof(ExpectRace)));
   race->addr = addr;
   race->size = size;
   race->file = f;
@@ -137,81 +135,12 @@ static void InitList(ExpectRace *list) {
 
 void InitializeDynamicAnnotations() {
   dyn_ann_ctx = new(dyn_ann_ctx_placeholder) DynamicAnnContext;
-  InitList(&dyn_ann_ctx->expect);
   InitList(&dyn_ann_ctx->benign);
 }
 
 bool IsExpectedReport(uptr addr, uptr size) {
   ReadLock lock(&dyn_ann_ctx->mtx);
-  if (CheckContains(&dyn_ann_ctx->expect, addr, size))
-    return true;
-  if (CheckContains(&dyn_ann_ctx->benign, addr, size))
-    return true;
-  return false;
-}
-
-static void CollectMatchedBenignRaces(Vector<ExpectRace> *matched,
-    int *unique_count, int *hit_count, atomic_uintptr_t ExpectRace::*counter) {
-  ExpectRace *list = &dyn_ann_ctx->benign;
-  for (ExpectRace *race = list->next; race != list; race = race->next) {
-    (*unique_count)++;
-    const uptr cnt = atomic_load_relaxed(&(race->*counter));
-    if (cnt == 0)
-      continue;
-    *hit_count += cnt;
-    uptr i = 0;
-    for (; i < matched->Size(); i++) {
-      ExpectRace *race0 = &(*matched)[i];
-      if (race->line == race0->line
-          && internal_strcmp(race->file, race0->file) == 0
-          && internal_strcmp(race->desc, race0->desc) == 0) {
-        atomic_fetch_add(&(race0->*counter), cnt, memory_order_relaxed);
-        break;
-      }
-    }
-    if (i == matched->Size())
-      matched->PushBack(*race);
-  }
-}
-
-void PrintMatchedBenignRaces() {
-  Lock lock(&dyn_ann_ctx->mtx);
-  int unique_count = 0;
-  int hit_count = 0;
-  int add_count = 0;
-  Vector<ExpectRace> hit_matched;
-  CollectMatchedBenignRaces(&hit_matched, &unique_count, &hit_count,
-      &ExpectRace::hitcount);
-  Vector<ExpectRace> add_matched;
-  CollectMatchedBenignRaces(&add_matched, &unique_count, &add_count,
-      &ExpectRace::addcount);
-  if (hit_matched.Size()) {
-    Printf("ThreadSanitizer: Matched %d \"benign\" races (pid=%d):\n",
-        hit_count, (int)internal_getpid());
-    for (uptr i = 0; i < hit_matched.Size(); i++) {
-      Printf("%d %s:%d %s\n",
-          atomic_load_relaxed(&hit_matched[i].hitcount),
-          hit_matched[i].file, hit_matched[i].line, hit_matched[i].desc);
-    }
-  }
-  if (hit_matched.Size()) {
-    Printf("ThreadSanitizer: Annotated %d \"benign\" races, %d unique"
-           " (pid=%d):\n",
-        add_count, unique_count, (int)internal_getpid());
-    for (uptr i = 0; i < add_matched.Size(); i++) {
-      Printf("%d %s:%d %s\n",
-          atomic_load_relaxed(&add_matched[i].addcount),
-          add_matched[i].file, add_matched[i].line, add_matched[i].desc);
-    }
-  }
-}
-
-static void ReportMissedExpectedRace(ExpectRace *race) {
-  Printf("==================\n");
-  Printf("WARNING: ThreadSanitizer: missed expected data race\n");
-  Printf("  %s addr=%zx %s:%d\n",
-      race->desc, race->addr, race->file, race->line);
-  Printf("==================\n");
+  return CheckContains(&dyn_ann_ctx->benign, addr, size);
 }
 }  // namespace __tsan
 
@@ -229,20 +158,16 @@ void INTERFACE_ATTRIBUTE AnnotateHappensAfter(char *f, int l, uptr addr) {
 }
 
 void INTERFACE_ATTRIBUTE AnnotateCondVarSignal(char *f, int l, uptr cv) {
-  SCOPED_ANNOTATION(AnnotateCondVarSignal);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateCondVarSignalAll(char *f, int l, uptr cv) {
-  SCOPED_ANNOTATION(AnnotateCondVarSignalAll);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateMutexIsNotPHB(char *f, int l, uptr mu) {
-  SCOPED_ANNOTATION(AnnotateMutexIsNotPHB);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateCondVarWait(char *f, int l, uptr cv,
                                              uptr lock) {
-  SCOPED_ANNOTATION(AnnotateCondVarWait);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateRWLockCreate(char *f, int l, uptr m) {
@@ -279,86 +204,56 @@ void INTERFACE_ATTRIBUTE AnnotateRWLockReleased(char *f, int l, uptr m,
 }
 
 void INTERFACE_ATTRIBUTE AnnotateTraceMemory(char *f, int l, uptr mem) {
-  SCOPED_ANNOTATION(AnnotateTraceMemory);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateFlushState(char *f, int l) {
-  SCOPED_ANNOTATION(AnnotateFlushState);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateNewMemory(char *f, int l, uptr mem,
                                            uptr size) {
-  SCOPED_ANNOTATION(AnnotateNewMemory);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateNoOp(char *f, int l, uptr mem) {
-  SCOPED_ANNOTATION(AnnotateNoOp);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateFlushExpectedRaces(char *f, int l) {
-  SCOPED_ANNOTATION(AnnotateFlushExpectedRaces);
-  Lock lock(&dyn_ann_ctx->mtx);
-  while (dyn_ann_ctx->expect.next != &dyn_ann_ctx->expect) {
-    ExpectRace *race = dyn_ann_ctx->expect.next;
-    if (atomic_load_relaxed(&race->hitcount) == 0) {
-      ctx->nmissed_expected++;
-      ReportMissedExpectedRace(race);
-    }
-    race->prev->next = race->next;
-    race->next->prev = race->prev;
-    internal_free(race);
-  }
 }
 
 void INTERFACE_ATTRIBUTE AnnotateEnableRaceDetection(
     char *f, int l, int enable) {
-  SCOPED_ANNOTATION(AnnotateEnableRaceDetection);
-  // FIXME: Reconsider this functionality later. It may be irrelevant.
 }
 
 void INTERFACE_ATTRIBUTE AnnotateMutexIsUsedAsCondVar(
     char *f, int l, uptr mu) {
-  SCOPED_ANNOTATION(AnnotateMutexIsUsedAsCondVar);
 }
 
 void INTERFACE_ATTRIBUTE AnnotatePCQGet(
     char *f, int l, uptr pcq) {
-  SCOPED_ANNOTATION(AnnotatePCQGet);
 }
 
 void INTERFACE_ATTRIBUTE AnnotatePCQPut(
     char *f, int l, uptr pcq) {
-  SCOPED_ANNOTATION(AnnotatePCQPut);
 }
 
 void INTERFACE_ATTRIBUTE AnnotatePCQDestroy(
     char *f, int l, uptr pcq) {
-  SCOPED_ANNOTATION(AnnotatePCQDestroy);
 }
 
 void INTERFACE_ATTRIBUTE AnnotatePCQCreate(
     char *f, int l, uptr pcq) {
-  SCOPED_ANNOTATION(AnnotatePCQCreate);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateExpectRace(
     char *f, int l, uptr mem, char *desc) {
-  SCOPED_ANNOTATION(AnnotateExpectRace);
-  Lock lock(&dyn_ann_ctx->mtx);
-  AddExpectRace(&dyn_ann_ctx->expect,
-                f, l, mem, 1, desc);
-  DPrintf("Add expected race: %s addr=%zx %s:%d\n", desc, mem, f, l);
 }
 
-static void BenignRaceImpl(
-    char *f, int l, uptr mem, uptr size, char *desc) {
+static void BenignRaceImpl(char *f, int l, uptr mem, uptr size, char *desc) {
   Lock lock(&dyn_ann_ctx->mtx);
   AddExpectRace(&dyn_ann_ctx->benign,
                 f, l, mem, size, desc);
   DPrintf("Add benign race: %s addr=%zx %s:%d\n", desc, mem, f, l);
 }
 
-// FIXME: Turn it off later. WTF is benign race?1?? Go talk to Hans Boehm.
 void INTERFACE_ATTRIBUTE AnnotateBenignRaceSized(
     char *f, int l, uptr mem, uptr size, char *desc) {
   SCOPED_ANNOTATION(AnnotateBenignRaceSized);
@@ -378,7 +273,7 @@ void INTERFACE_ATTRIBUTE AnnotateIgnoreReadsBegin(char *f, int l) {
 
 void INTERFACE_ATTRIBUTE AnnotateIgnoreReadsEnd(char *f, int l) {
   SCOPED_ANNOTATION(AnnotateIgnoreReadsEnd);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreEnd(thr);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateIgnoreWritesBegin(char *f, int l) {
@@ -388,7 +283,7 @@ void INTERFACE_ATTRIBUTE AnnotateIgnoreWritesBegin(char *f, int l) {
 
 void INTERFACE_ATTRIBUTE AnnotateIgnoreWritesEnd(char *f, int l) {
   SCOPED_ANNOTATION(AnnotateIgnoreWritesEnd);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreEnd(thr);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateIgnoreSyncBegin(char *f, int l) {
@@ -398,17 +293,15 @@ void INTERFACE_ATTRIBUTE AnnotateIgnoreSyncBegin(char *f, int l) {
 
 void INTERFACE_ATTRIBUTE AnnotateIgnoreSyncEnd(char *f, int l) {
   SCOPED_ANNOTATION(AnnotateIgnoreSyncEnd);
-  ThreadIgnoreSyncEnd(thr, pc);
+  ThreadIgnoreSyncEnd(thr);
 }
 
 void INTERFACE_ATTRIBUTE AnnotatePublishMemoryRange(
     char *f, int l, uptr addr, uptr size) {
-  SCOPED_ANNOTATION(AnnotatePublishMemoryRange);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateUnpublishMemoryRange(
     char *f, int l, uptr addr, uptr size) {
-  SCOPED_ANNOTATION(AnnotateUnpublishMemoryRange);
 }
 
 void INTERFACE_ATTRIBUTE AnnotateThreadName(
@@ -421,11 +314,9 @@ void INTERFACE_ATTRIBUTE AnnotateThreadName(
 // WTFAnnotateHappensAfter(). Those are being used by Webkit to annotate
 // atomic operations, which should be handled by ThreadSanitizer correctly.
 void INTERFACE_ATTRIBUTE WTFAnnotateHappensBefore(char *f, int l, uptr addr) {
-  SCOPED_ANNOTATION(AnnotateHappensBefore);
 }
 
 void INTERFACE_ATTRIBUTE WTFAnnotateHappensAfter(char *f, int l, uptr addr) {
-  SCOPED_ANNOTATION(AnnotateHappensAfter);
 }
 
 void INTERFACE_ATTRIBUTE WTFAnnotateBenignRaceSized(
@@ -477,15 +368,15 @@ void __tsan_mutex_pre_lock(void *m, unsigned flagz) {
     else
       MutexPreLock(thr, pc, (uptr)m);
   }
-  ThreadIgnoreBegin(thr, pc, /*save_stack=*/false);
-  ThreadIgnoreSyncBegin(thr, pc, /*save_stack=*/false);
+  ThreadIgnoreBegin(thr, 0);
+  ThreadIgnoreSyncBegin(thr, 0);
 }
 
 INTERFACE_ATTRIBUTE
 void __tsan_mutex_post_lock(void *m, unsigned flagz, int rec) {
   SCOPED_ANNOTATION(__tsan_mutex_post_lock);
-  ThreadIgnoreSyncEnd(thr, pc);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreSyncEnd(thr);
+  ThreadIgnoreEnd(thr);
   if (!(flagz & MutexFlagTryLockFailed)) {
     if (flagz & MutexFlagReadLock)
       MutexPostReadLock(thr, pc, (uptr)m, flagz);
@@ -504,44 +395,44 @@ int __tsan_mutex_pre_unlock(void *m, unsigned flagz) {
   } else {
     ret = MutexUnlock(thr, pc, (uptr)m, flagz);
   }
-  ThreadIgnoreBegin(thr, pc, /*save_stack=*/false);
-  ThreadIgnoreSyncBegin(thr, pc, /*save_stack=*/false);
+  ThreadIgnoreBegin(thr, 0);
+  ThreadIgnoreSyncBegin(thr, 0);
   return ret;
 }
 
 INTERFACE_ATTRIBUTE
 void __tsan_mutex_post_unlock(void *m, unsigned flagz) {
   SCOPED_ANNOTATION(__tsan_mutex_post_unlock);
-  ThreadIgnoreSyncEnd(thr, pc);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreSyncEnd(thr);
+  ThreadIgnoreEnd(thr);
 }
 
 INTERFACE_ATTRIBUTE
 void __tsan_mutex_pre_signal(void *addr, unsigned flagz) {
   SCOPED_ANNOTATION(__tsan_mutex_pre_signal);
-  ThreadIgnoreBegin(thr, pc, /*save_stack=*/false);
-  ThreadIgnoreSyncBegin(thr, pc, /*save_stack=*/false);
+  ThreadIgnoreBegin(thr, 0);
+  ThreadIgnoreSyncBegin(thr, 0);
 }
 
 INTERFACE_ATTRIBUTE
 void __tsan_mutex_post_signal(void *addr, unsigned flagz) {
   SCOPED_ANNOTATION(__tsan_mutex_post_signal);
-  ThreadIgnoreSyncEnd(thr, pc);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreSyncEnd(thr);
+  ThreadIgnoreEnd(thr);
 }
 
 INTERFACE_ATTRIBUTE
 void __tsan_mutex_pre_divert(void *addr, unsigned flagz) {
   SCOPED_ANNOTATION(__tsan_mutex_pre_divert);
   // Exit from ignore region started in __tsan_mutex_pre_lock/unlock/signal.
-  ThreadIgnoreSyncEnd(thr, pc);
-  ThreadIgnoreEnd(thr, pc);
+  ThreadIgnoreSyncEnd(thr);
+  ThreadIgnoreEnd(thr);
 }
 
 INTERFACE_ATTRIBUTE
 void __tsan_mutex_post_divert(void *addr, unsigned flagz) {
   SCOPED_ANNOTATION(__tsan_mutex_post_divert);
-  ThreadIgnoreBegin(thr, pc, /*save_stack=*/false);
-  ThreadIgnoreSyncBegin(thr, pc, /*save_stack=*/false);
+  ThreadIgnoreBegin(thr, 0);
+  ThreadIgnoreSyncBegin(thr, 0);
 }
 }  // extern "C"
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interface_atomic.cpp b/compiler-rt/lib/tsan/rtl/tsan_interface_atomic.cpp
index 89bb75394553..24ba3bb1f65d 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_interface_atomic.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_interface_atomic.cpp
@@ -32,6 +32,7 @@ using namespace __tsan;
 static StaticSpinMutex mutex128;
 #endif
 
+#if SANITIZER_DEBUG
 static bool IsLoadOrder(morder mo) {
   return mo == mo_relaxed || mo == mo_consume
       || mo == mo_acquire || mo == mo_seq_cst;
@@ -40,6 +41,7 @@ static bool IsLoadOrder(morder mo) {
 static bool IsStoreOrder(morder mo) {
   return mo == mo_relaxed || mo == mo_release || mo == mo_seq_cst;
 }
+#endif
 
 static bool IsReleaseOrder(morder mo) {
   return mo == mo_release || mo == mo_acq_rel || mo == mo_seq_cst;
@@ -161,16 +163,16 @@ a128 func_cas(volatile a128 *v, a128 cmp, a128 xch) {
 }
 #endif
 
-template<typename T>
-static int SizeLog() {
+template <typename T>
+static int AccessSize() {
   if (sizeof(T) <= 1)
-    return kSizeLog1;
+    return 1;
   else if (sizeof(T) <= 2)
-    return kSizeLog2;
+    return 2;
   else if (sizeof(T) <= 4)
-    return kSizeLog4;
+    return 4;
   else
-    return kSizeLog8;
+    return 8;
   // For 16-byte atomics we also use 8-byte memory access,
   // this leads to false negatives only in very obscure cases.
 }
@@ -202,7 +204,7 @@ static memory_order to_mo(morder mo) {
   case mo_acq_rel: return memory_order_acq_rel;
   case mo_seq_cst: return memory_order_seq_cst;
   }
-  CHECK(0);
+  DCHECK(0);
   return memory_order_seq_cst;
 }
 
@@ -219,27 +221,27 @@ static a128 NoTsanAtomicLoad(const volatile a128 *a, morder mo) {
 #endif
 
 template <typename T>
-static T AtomicLoad(ThreadState *thr, uptr pc, const volatile T *a,
-                    morder mo) NO_THREAD_SAFETY_ANALYSIS {
-  CHECK(IsLoadOrder(mo));
+static T AtomicLoad(ThreadState *thr, uptr pc, const volatile T *a, morder mo) {
+  DCHECK(IsLoadOrder(mo));
   // This fast-path is critical for performance.
   // Assume the access is atomic.
   if (!IsAcquireOrder(mo)) {
-    MemoryReadAtomic(thr, pc, (uptr)a, SizeLog<T>());
+    MemoryAccess(thr, pc, (uptr)a, AccessSize<T>(),
+                 kAccessRead | kAccessAtomic);
     return NoTsanAtomicLoad(a, mo);
   }
   // Don't create sync object if it does not exist yet. For example, an atomic
   // pointer is initialized to nullptr and then periodically acquire-loaded.
   T v = NoTsanAtomicLoad(a, mo);
-  SyncVar *s = ctx->metamap.GetIfExistsAndLock((uptr)a, false);
+  SyncVar *s = ctx->metamap.GetSyncIfExists((uptr)a);
   if (s) {
+    ReadLock l(&s->mtx);
     AcquireImpl(thr, pc, &s->clock);
     // Re-read under sync mutex because we need a consistent snapshot
     // of the value and the clock we acquire.
     v = NoTsanAtomicLoad(a, mo);
-    s->mtx.ReadUnlock();
   }
-  MemoryReadAtomic(thr, pc, (uptr)a, SizeLog<T>());
+  MemoryAccess(thr, pc, (uptr)a, AccessSize<T>(), kAccessRead | kAccessAtomic);
   return v;
 }
 
@@ -257,9 +259,9 @@ static void NoTsanAtomicStore(volatile a128 *a, a128 v, morder mo) {
 
 template <typename T>
 static void AtomicStore(ThreadState *thr, uptr pc, volatile T *a, T v,
-                        morder mo) NO_THREAD_SAFETY_ANALYSIS {
-  CHECK(IsStoreOrder(mo));
-  MemoryWriteAtomic(thr, pc, (uptr)a, SizeLog<T>());
+                        morder mo) {
+  DCHECK(IsStoreOrder(mo));
+  MemoryAccess(thr, pc, (uptr)a, AccessSize<T>(), kAccessWrite | kAccessAtomic);
   // This fast-path is critical for performance.
   // Assume the access is atomic.
   // Strictly saying even relaxed store cuts off release sequence,
@@ -269,36 +271,32 @@ static void AtomicStore(ThreadState *thr, uptr pc, volatile T *a, T v,
     return;
   }
   __sync_synchronize();
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, (uptr)a, true);
+  SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, (uptr)a, false);
+  Lock l(&s->mtx);
   thr->fast_state.IncrementEpoch();
   // Can't increment epoch w/o writing to the trace as well.
   TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
   ReleaseStoreImpl(thr, pc, &s->clock);
   NoTsanAtomicStore(a, v, mo);
-  s->mtx.Unlock();
 }
 
 template <typename T, T (*F)(volatile T *v, T op)>
-static T AtomicRMW(ThreadState *thr, uptr pc, volatile T *a, T v,
-                   morder mo) NO_THREAD_SAFETY_ANALYSIS {
-  MemoryWriteAtomic(thr, pc, (uptr)a, SizeLog<T>());
-  SyncVar *s = 0;
-  if (mo != mo_relaxed) {
-    s = ctx->metamap.GetOrCreateAndLock(thr, pc, (uptr)a, true);
-    thr->fast_state.IncrementEpoch();
-    // Can't increment epoch w/o writing to the trace as well.
-    TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
-    if (IsAcqRelOrder(mo))
-      AcquireReleaseImpl(thr, pc, &s->clock);
-    else if (IsReleaseOrder(mo))
-      ReleaseImpl(thr, pc, &s->clock);
-    else if (IsAcquireOrder(mo))
-      AcquireImpl(thr, pc, &s->clock);
-  }
-  v = F(a, v);
-  if (s)
-    s->mtx.Unlock();
-  return v;
+static T AtomicRMW(ThreadState *thr, uptr pc, volatile T *a, T v, morder mo) {
+  MemoryAccess(thr, pc, (uptr)a, AccessSize<T>(), kAccessWrite | kAccessAtomic);
+  if (LIKELY(mo == mo_relaxed))
+    return F(a, v);
+  SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, (uptr)a, false);
+  Lock l(&s->mtx);
+  thr->fast_state.IncrementEpoch();
+  // Can't increment epoch w/o writing to the trace as well.
+  TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
+  if (IsAcqRelOrder(mo))
+    AcquireReleaseImpl(thr, pc, &s->clock);
+  else if (IsReleaseOrder(mo))
+    ReleaseImpl(thr, pc, &s->clock);
+  else if (IsAcquireOrder(mo))
+    AcquireImpl(thr, pc, &s->clock);
+  return F(a, v);
 }
 
 template<typename T>
@@ -402,20 +400,26 @@ static T NoTsanAtomicCAS(volatile T *a, T c, T v, morder mo, morder fmo) {
 }
 
 template <typename T>
-static bool AtomicCAS(ThreadState *thr, uptr pc, volatile T *a, T *c, T v, morder mo,
-                      morder fmo) NO_THREAD_SAFETY_ANALYSIS {
+static bool AtomicCAS(ThreadState *thr, uptr pc, volatile T *a, T *c, T v,
+                      morder mo, morder fmo) {
   // 31.7.2.18: "The failure argument shall not be memory_order_release
   // nor memory_order_acq_rel". LLVM (2021-05) fallbacks to Monotonic
   // (mo_relaxed) when those are used.
-  CHECK(IsLoadOrder(fmo));
-
-  MemoryWriteAtomic(thr, pc, (uptr)a, SizeLog<T>());
-  SyncVar *s = 0;
-  bool write_lock = IsReleaseOrder(mo);
+  DCHECK(IsLoadOrder(fmo));
 
-  if (mo != mo_relaxed || fmo != mo_relaxed)
-    s = ctx->metamap.GetOrCreateAndLock(thr, pc, (uptr)a, write_lock);
+  MemoryAccess(thr, pc, (uptr)a, AccessSize<T>(), kAccessWrite | kAccessAtomic);
+  if (LIKELY(mo == mo_relaxed && fmo == mo_relaxed)) {
+    T cc = *c;
+    T pr = func_cas(a, cc, v);
+    if (pr == cc)
+      return true;
+    *c = pr;
+    return false;
+  }
 
+  bool release = IsReleaseOrder(mo);
+  SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, (uptr)a, false);
+  RWLock l(&s->mtx, release);
   T cc = *c;
   T pr = func_cas(a, cc, v);
   bool success = pr == cc;
@@ -423,25 +427,16 @@ static bool AtomicCAS(ThreadState *thr, uptr pc, volatile T *a, T *c, T v, morde
     *c = pr;
     mo = fmo;
   }
+  thr->fast_state.IncrementEpoch();
+  // Can't increment epoch w/o writing to the trace as well.
+  TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
 
-  if (s) {
-    thr->fast_state.IncrementEpoch();
-    // Can't increment epoch w/o writing to the trace as well.
-    TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
-
-    if (success && IsAcqRelOrder(mo))
-      AcquireReleaseImpl(thr, pc, &s->clock);
-    else if (success && IsReleaseOrder(mo))
-      ReleaseImpl(thr, pc, &s->clock);
-    else if (IsAcquireOrder(mo))
-      AcquireImpl(thr, pc, &s->clock);
-
-    if (write_lock)
-      s->mtx.Unlock();
-    else
-      s->mtx.ReadUnlock();
-  }
-
+  if (success && IsAcqRelOrder(mo))
+    AcquireReleaseImpl(thr, pc, &s->clock);
+  else if (success && IsReleaseOrder(mo))
+    ReleaseImpl(thr, pc, &s->clock);
+  else if (IsAcquireOrder(mo))
+    AcquireImpl(thr, pc, &s->clock);
   return success;
 }
 
@@ -485,380 +480,356 @@ static morder convert_morder(morder mo) {
   return (morder)(mo & 0x7fff);
 }
 
-#define SCOPED_ATOMIC(func, ...) \
-    ThreadState *const thr = cur_thread(); \
-    if (UNLIKELY(thr->ignore_sync || thr->ignore_interceptors)) { \
-      ProcessPendingSignals(thr); \
-      return NoTsanAtomic##func(__VA_ARGS__); \
-    } \
-    const uptr callpc = (uptr)__builtin_return_address(0); \
-    uptr pc = StackTrace::GetCurrentPc(); \
-    mo = convert_morder(mo); \
-    ScopedAtomic sa(thr, callpc, a, mo, __func__); \
-    return Atomic##func(thr, pc, __VA_ARGS__); \
-/**/
-
-class ScopedAtomic {
- public:
-  ScopedAtomic(ThreadState *thr, uptr pc, const volatile void *a,
-               morder mo, const char *func)
-      : thr_(thr) {
-    FuncEntry(thr_, pc);
-    DPrintf("#%d: %s(%p, %d)\n", thr_->tid, func, a, mo);
-  }
-  ~ScopedAtomic() {
-    ProcessPendingSignals(thr_);
-    FuncExit(thr_);
-  }
- private:
-  ThreadState *thr_;
-};
+#  define ATOMIC_IMPL(func, ...)                                \
+    ThreadState *const thr = cur_thread();                      \
+    ProcessPendingSignals(thr);                                 \
+    if (UNLIKELY(thr->ignore_sync || thr->ignore_interceptors)) \
+      return NoTsanAtomic##func(__VA_ARGS__);                   \
+    mo = convert_morder(mo);                                    \
+    return Atomic##func(thr, GET_CALLER_PC(), __VA_ARGS__);
 
 extern "C" {
 SANITIZER_INTERFACE_ATTRIBUTE
 a8 __tsan_atomic8_load(const volatile a8 *a, morder mo) {
-  SCOPED_ATOMIC(Load, a, mo);
+  ATOMIC_IMPL(Load, a, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a16 __tsan_atomic16_load(const volatile a16 *a, morder mo) {
-  SCOPED_ATOMIC(Load, a, mo);
+  ATOMIC_IMPL(Load, a, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a32 __tsan_atomic32_load(const volatile a32 *a, morder mo) {
-  SCOPED_ATOMIC(Load, a, mo);
+  ATOMIC_IMPL(Load, a, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a64 __tsan_atomic64_load(const volatile a64 *a, morder mo) {
-  SCOPED_ATOMIC(Load, a, mo);
+  ATOMIC_IMPL(Load, a, mo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 a128 __tsan_atomic128_load(const volatile a128 *a, morder mo) {
-  SCOPED_ATOMIC(Load, a, mo);
+  ATOMIC_IMPL(Load, a, mo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_atomic8_store(volatile a8 *a, a8 v, morder mo) {
-  SCOPED_ATOMIC(Store, a, v, mo);
+  ATOMIC_IMPL(Store, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_atomic16_store(volatile a16 *a, a16 v, morder mo) {
-  SCOPED_ATOMIC(Store, a, v, mo);
+  ATOMIC_IMPL(Store, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_atomic32_store(volatile a32 *a, a32 v, morder mo) {
-  SCOPED_ATOMIC(Store, a, v, mo);
+  ATOMIC_IMPL(Store, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_atomic64_store(volatile a64 *a, a64 v, morder mo) {
-  SCOPED_ATOMIC(Store, a, v, mo);
+  ATOMIC_IMPL(Store, a, v, mo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_atomic128_store(volatile a128 *a, a128 v, morder mo) {
-  SCOPED_ATOMIC(Store, a, v, mo);
+  ATOMIC_IMPL(Store, a, v, mo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a8 __tsan_atomic8_exchange(volatile a8 *a, a8 v, morder mo) {
-  SCOPED_ATOMIC(Exchange, a, v, mo);
+  ATOMIC_IMPL(Exchange, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a16 __tsan_atomic16_exchange(volatile a16 *a, a16 v, morder mo) {
-  SCOPED_ATOMIC(Exchange, a, v, mo);
+  ATOMIC_IMPL(Exchange, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a32 __tsan_atomic32_exchange(volatile a32 *a, a32 v, morder mo) {
-  SCOPED_ATOMIC(Exchange, a, v, mo);
+  ATOMIC_IMPL(Exchange, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a64 __tsan_atomic64_exchange(volatile a64 *a, a64 v, morder mo) {
-  SCOPED_ATOMIC(Exchange, a, v, mo);
+  ATOMIC_IMPL(Exchange, a, v, mo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 a128 __tsan_atomic128_exchange(volatile a128 *a, a128 v, morder mo) {
-  SCOPED_ATOMIC(Exchange, a, v, mo);
+  ATOMIC_IMPL(Exchange, a, v, mo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a8 __tsan_atomic8_fetch_add(volatile a8 *a, a8 v, morder mo) {
-  SCOPED_ATOMIC(FetchAdd, a, v, mo);
+  ATOMIC_IMPL(FetchAdd, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a16 __tsan_atomic16_fetch_add(volatile a16 *a, a16 v, morder mo) {
-  SCOPED_ATOMIC(FetchAdd, a, v, mo);
+  ATOMIC_IMPL(FetchAdd, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a32 __tsan_atomic32_fetch_add(volatile a32 *a, a32 v, morder mo) {
-  SCOPED_ATOMIC(FetchAdd, a, v, mo);
+  ATOMIC_IMPL(FetchAdd, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a64 __tsan_atomic64_fetch_add(volatile a64 *a, a64 v, morder mo) {
-  SCOPED_ATOMIC(FetchAdd, a, v, mo);
+  ATOMIC_IMPL(FetchAdd, a, v, mo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 a128 __tsan_atomic128_fetch_add(volatile a128 *a, a128 v, morder mo) {
-  SCOPED_ATOMIC(FetchAdd, a, v, mo);
+  ATOMIC_IMPL(FetchAdd, a, v, mo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a8 __tsan_atomic8_fetch_sub(volatile a8 *a, a8 v, morder mo) {
-  SCOPED_ATOMIC(FetchSub, a, v, mo);
+  ATOMIC_IMPL(FetchSub, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a16 __tsan_atomic16_fetch_sub(volatile a16 *a, a16 v, morder mo) {
-  SCOPED_ATOMIC(FetchSub, a, v, mo);
+  ATOMIC_IMPL(FetchSub, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a32 __tsan_atomic32_fetch_sub(volatile a32 *a, a32 v, morder mo) {
-  SCOPED_ATOMIC(FetchSub, a, v, mo);
+  ATOMIC_IMPL(FetchSub, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a64 __tsan_atomic64_fetch_sub(volatile a64 *a, a64 v, morder mo) {
-  SCOPED_ATOMIC(FetchSub, a, v, mo);
+  ATOMIC_IMPL(FetchSub, a, v, mo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 a128 __tsan_atomic128_fetch_sub(volatile a128 *a, a128 v, morder mo) {
-  SCOPED_ATOMIC(FetchSub, a, v, mo);
+  ATOMIC_IMPL(FetchSub, a, v, mo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a8 __tsan_atomic8_fetch_and(volatile a8 *a, a8 v, morder mo) {
-  SCOPED_ATOMIC(FetchAnd, a, v, mo);
+  ATOMIC_IMPL(FetchAnd, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a16 __tsan_atomic16_fetch_and(volatile a16 *a, a16 v, morder mo) {
-  SCOPED_ATOMIC(FetchAnd, a, v, mo);
+  ATOMIC_IMPL(FetchAnd, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a32 __tsan_atomic32_fetch_and(volatile a32 *a, a32 v, morder mo) {
-  SCOPED_ATOMIC(FetchAnd, a, v, mo);
+  ATOMIC_IMPL(FetchAnd, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a64 __tsan_atomic64_fetch_and(volatile a64 *a, a64 v, morder mo) {
-  SCOPED_ATOMIC(FetchAnd, a, v, mo);
+  ATOMIC_IMPL(FetchAnd, a, v, mo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 a128 __tsan_atomic128_fetch_and(volatile a128 *a, a128 v, morder mo) {
-  SCOPED_ATOMIC(FetchAnd, a, v, mo);
+  ATOMIC_IMPL(FetchAnd, a, v, mo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a8 __tsan_atomic8_fetch_or(volatile a8 *a, a8 v, morder mo) {
-  SCOPED_ATOMIC(FetchOr, a, v, mo);
+  ATOMIC_IMPL(FetchOr, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a16 __tsan_atomic16_fetch_or(volatile a16 *a, a16 v, morder mo) {
-  SCOPED_ATOMIC(FetchOr, a, v, mo);
+  ATOMIC_IMPL(FetchOr, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a32 __tsan_atomic32_fetch_or(volatile a32 *a, a32 v, morder mo) {
-  SCOPED_ATOMIC(FetchOr, a, v, mo);
+  ATOMIC_IMPL(FetchOr, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a64 __tsan_atomic64_fetch_or(volatile a64 *a, a64 v, morder mo) {
-  SCOPED_ATOMIC(FetchOr, a, v, mo);
+  ATOMIC_IMPL(FetchOr, a, v, mo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 a128 __tsan_atomic128_fetch_or(volatile a128 *a, a128 v, morder mo) {
-  SCOPED_ATOMIC(FetchOr, a, v, mo);
+  ATOMIC_IMPL(FetchOr, a, v, mo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a8 __tsan_atomic8_fetch_xor(volatile a8 *a, a8 v, morder mo) {
-  SCOPED_ATOMIC(FetchXor, a, v, mo);
+  ATOMIC_IMPL(FetchXor, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a16 __tsan_atomic16_fetch_xor(volatile a16 *a, a16 v, morder mo) {
-  SCOPED_ATOMIC(FetchXor, a, v, mo);
+  ATOMIC_IMPL(FetchXor, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a32 __tsan_atomic32_fetch_xor(volatile a32 *a, a32 v, morder mo) {
-  SCOPED_ATOMIC(FetchXor, a, v, mo);
+  ATOMIC_IMPL(FetchXor, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a64 __tsan_atomic64_fetch_xor(volatile a64 *a, a64 v, morder mo) {
-  SCOPED_ATOMIC(FetchXor, a, v, mo);
+  ATOMIC_IMPL(FetchXor, a, v, mo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 a128 __tsan_atomic128_fetch_xor(volatile a128 *a, a128 v, morder mo) {
-  SCOPED_ATOMIC(FetchXor, a, v, mo);
+  ATOMIC_IMPL(FetchXor, a, v, mo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a8 __tsan_atomic8_fetch_nand(volatile a8 *a, a8 v, morder mo) {
-  SCOPED_ATOMIC(FetchNand, a, v, mo);
+  ATOMIC_IMPL(FetchNand, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a16 __tsan_atomic16_fetch_nand(volatile a16 *a, a16 v, morder mo) {
-  SCOPED_ATOMIC(FetchNand, a, v, mo);
+  ATOMIC_IMPL(FetchNand, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a32 __tsan_atomic32_fetch_nand(volatile a32 *a, a32 v, morder mo) {
-  SCOPED_ATOMIC(FetchNand, a, v, mo);
+  ATOMIC_IMPL(FetchNand, a, v, mo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a64 __tsan_atomic64_fetch_nand(volatile a64 *a, a64 v, morder mo) {
-  SCOPED_ATOMIC(FetchNand, a, v, mo);
+  ATOMIC_IMPL(FetchNand, a, v, mo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 a128 __tsan_atomic128_fetch_nand(volatile a128 *a, a128 v, morder mo) {
-  SCOPED_ATOMIC(FetchNand, a, v, mo);
+  ATOMIC_IMPL(FetchNand, a, v, mo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic8_compare_exchange_strong(volatile a8 *a, a8 *c, a8 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic16_compare_exchange_strong(volatile a16 *a, a16 *c, a16 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic32_compare_exchange_strong(volatile a32 *a, a32 *c, a32 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic64_compare_exchange_strong(volatile a64 *a, a64 *c, a64 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic128_compare_exchange_strong(volatile a128 *a, a128 *c, a128 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic8_compare_exchange_weak(volatile a8 *a, a8 *c, a8 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic16_compare_exchange_weak(volatile a16 *a, a16 *c, a16 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic32_compare_exchange_weak(volatile a32 *a, a32 *c, a32 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic64_compare_exchange_weak(volatile a64 *a, a64 *c, a64 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 int __tsan_atomic128_compare_exchange_weak(volatile a128 *a, a128 *c, a128 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a8 __tsan_atomic8_compare_exchange_val(volatile a8 *a, a8 c, a8 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a16 __tsan_atomic16_compare_exchange_val(volatile a16 *a, a16 c, a16 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a32 __tsan_atomic32_compare_exchange_val(volatile a32 *a, a32 c, a32 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 a64 __tsan_atomic64_compare_exchange_val(volatile a64 *a, a64 c, a64 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 
 #if __TSAN_HAS_INT128
 SANITIZER_INTERFACE_ATTRIBUTE
 a128 __tsan_atomic128_compare_exchange_val(volatile a128 *a, a128 c, a128 v,
     morder mo, morder fmo) {
-  SCOPED_ATOMIC(CAS, a, c, v, mo, fmo);
+  ATOMIC_IMPL(CAS, a, c, v, mo, fmo);
 }
 #endif
 
 SANITIZER_INTERFACE_ATTRIBUTE
-void __tsan_atomic_thread_fence(morder mo) {
-  char* a = 0;
-  SCOPED_ATOMIC(Fence, mo);
-}
+void __tsan_atomic_thread_fence(morder mo) { ATOMIC_IMPL(Fence, mo); }
 
 SANITIZER_INTERFACE_ATTRIBUTE
 void __tsan_atomic_signal_fence(morder mo) {
@@ -869,25 +840,23 @@ void __tsan_atomic_signal_fence(morder mo) {
 
 // Go
 
-#define ATOMIC(func, ...) \
-    if (thr->ignore_sync) { \
-      NoTsanAtomic##func(__VA_ARGS__); \
-    } else { \
-      FuncEntry(thr, cpc); \
+#  define ATOMIC(func, ...)               \
+    if (thr->ignore_sync) {               \
+      NoTsanAtomic##func(__VA_ARGS__);    \
+    } else {                              \
+      FuncEntry(thr, cpc);                \
       Atomic##func(thr, pc, __VA_ARGS__); \
-      FuncExit(thr); \
-    } \
-/**/
+      FuncExit(thr);                      \
+    }
 
-#define ATOMIC_RET(func, ret, ...) \
-    if (thr->ignore_sync) { \
-      (ret) = NoTsanAtomic##func(__VA_ARGS__); \
-    } else { \
-      FuncEntry(thr, cpc); \
+#  define ATOMIC_RET(func, ret, ...)              \
+    if (thr->ignore_sync) {                       \
+      (ret) = NoTsanAtomic##func(__VA_ARGS__);    \
+    } else {                                      \
+      FuncEntry(thr, cpc);                        \
       (ret) = Atomic##func(thr, pc, __VA_ARGS__); \
-      FuncExit(thr); \
-    } \
-/**/
+      FuncExit(thr);                              \
+    }
 
 extern "C" {
 SANITIZER_INTERFACE_ATTRIBUTE
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interface_inl.h b/compiler-rt/lib/tsan/rtl/tsan_interface_inl.h
deleted file mode 100644
index 5e77d4d3d288..000000000000
--- a/compiler-rt/lib/tsan/rtl/tsan_interface_inl.h
+++ /dev/null
@@ -1,133 +0,0 @@
-//===-- tsan_interface_inl.h ------------------------------------*- C++ -*-===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-//
-// This file is a part of ThreadSanitizer (TSan), a race detector.
-//
-//===----------------------------------------------------------------------===//
-
-#include "tsan_interface.h"
-#include "tsan_rtl.h"
-#include "sanitizer_common/sanitizer_ptrauth.h"
-
-#define CALLERPC ((uptr)__builtin_return_address(0))
-
-using namespace __tsan;
-
-void __tsan_read1(void *addr) {
-  MemoryRead(cur_thread(), CALLERPC, (uptr)addr, kSizeLog1);
-}
-
-void __tsan_read2(void *addr) {
-  MemoryRead(cur_thread(), CALLERPC, (uptr)addr, kSizeLog2);
-}
-
-void __tsan_read4(void *addr) {
-  MemoryRead(cur_thread(), CALLERPC, (uptr)addr, kSizeLog4);
-}
-
-void __tsan_read8(void *addr) {
-  MemoryRead(cur_thread(), CALLERPC, (uptr)addr, kSizeLog8);
-}
-
-void __tsan_write1(void *addr) {
-  MemoryWrite(cur_thread(), CALLERPC, (uptr)addr, kSizeLog1);
-}
-
-void __tsan_write2(void *addr) {
-  MemoryWrite(cur_thread(), CALLERPC, (uptr)addr, kSizeLog2);
-}
-
-void __tsan_write4(void *addr) {
-  MemoryWrite(cur_thread(), CALLERPC, (uptr)addr, kSizeLog4);
-}
-
-void __tsan_write8(void *addr) {
-  MemoryWrite(cur_thread(), CALLERPC, (uptr)addr, kSizeLog8);
-}
-
-void __tsan_read1_pc(void *addr, void *pc) {
-  MemoryRead(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog1);
-}
-
-void __tsan_read2_pc(void *addr, void *pc) {
-  MemoryRead(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog2);
-}
-
-void __tsan_read4_pc(void *addr, void *pc) {
-  MemoryRead(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog4);
-}
-
-void __tsan_read8_pc(void *addr, void *pc) {
-  MemoryRead(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog8);
-}
-
-void __tsan_write1_pc(void *addr, void *pc) {
-  MemoryWrite(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog1);
-}
-
-void __tsan_write2_pc(void *addr, void *pc) {
-  MemoryWrite(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog2);
-}
-
-void __tsan_write4_pc(void *addr, void *pc) {
-  MemoryWrite(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog4);
-}
-
-void __tsan_write8_pc(void *addr, void *pc) {
-  MemoryWrite(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, kSizeLog8);
-}
-
-void __tsan_vptr_update(void **vptr_p, void *new_val) {
-  CHECK_EQ(sizeof(vptr_p), 8);
-  if (*vptr_p != new_val) {
-    ThreadState *thr = cur_thread();
-    thr->is_vptr_access = true;
-    MemoryWrite(thr, CALLERPC, (uptr)vptr_p, kSizeLog8);
-    thr->is_vptr_access = false;
-  }
-}
-
-void __tsan_vptr_read(void **vptr_p) {
-  CHECK_EQ(sizeof(vptr_p), 8);
-  ThreadState *thr = cur_thread();
-  thr->is_vptr_access = true;
-  MemoryRead(thr, CALLERPC, (uptr)vptr_p, kSizeLog8);
-  thr->is_vptr_access = false;
-}
-
-void __tsan_func_entry(void *pc) {
-  FuncEntry(cur_thread(), STRIP_PAC_PC(pc));
-}
-
-void __tsan_func_exit() {
-  FuncExit(cur_thread());
-}
-
-void __tsan_ignore_thread_begin() {
-  ThreadIgnoreBegin(cur_thread(), CALLERPC);
-}
-
-void __tsan_ignore_thread_end() {
-  ThreadIgnoreEnd(cur_thread(), CALLERPC);
-}
-
-void __tsan_read_range(void *addr, uptr size) {
-  MemoryAccessRange(cur_thread(), CALLERPC, (uptr)addr, size, false);
-}
-
-void __tsan_write_range(void *addr, uptr size) {
-  MemoryAccessRange(cur_thread(), CALLERPC, (uptr)addr, size, true);
-}
-
-void __tsan_read_range_pc(void *addr, uptr size, void *pc) {
-  MemoryAccessRange(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, size, false);
-}
-
-void __tsan_write_range_pc(void *addr, uptr size, void *pc) {
-  MemoryAccessRange(cur_thread(), STRIP_PAC_PC(pc), (uptr)addr, size, true);
-}
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interface_java.cpp b/compiler-rt/lib/tsan/rtl/tsan_interface_java.cpp
index 6aa8a7b1d6a7..c090c1f08cbe 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_interface_java.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_interface_java.cpp
@@ -34,52 +34,49 @@ struct JavaContext {
   }
 };
 
-class ScopedJavaFunc {
- public:
-  ScopedJavaFunc(ThreadState *thr, uptr pc)
-      : thr_(thr) {
-    Initialize(thr_);
-    FuncEntry(thr, pc);
-  }
-
-  ~ScopedJavaFunc() {
-    FuncExit(thr_);
-    // FIXME(dvyukov): process pending signals.
-  }
-
- private:
-  ThreadState *thr_;
-};
-
 static u64 jctx_buf[sizeof(JavaContext) / sizeof(u64) + 1];
 static JavaContext *jctx;
 
+MBlock *JavaHeapBlock(uptr addr, uptr *start) {
+  if (!jctx || addr < jctx->heap_begin ||
+      addr >= jctx->heap_begin + jctx->heap_size)
+    return nullptr;
+  for (uptr p = RoundDown(addr, kMetaShadowCell); p >= jctx->heap_begin;
+       p -= kMetaShadowCell) {
+    MBlock *b = ctx->metamap.GetBlock(p);
+    if (!b)
+      continue;
+    if (p + b->siz <= addr)
+      return nullptr;
+    *start = p;
+    return b;
+  }
+  return nullptr;
+}
+
 }  // namespace __tsan
 
-#define SCOPED_JAVA_FUNC(func) \
+#define JAVA_FUNC_ENTER(func)      \
   ThreadState *thr = cur_thread(); \
-  const uptr caller_pc = GET_CALLER_PC(); \
-  const uptr pc = StackTrace::GetCurrentPc(); \
-  (void)pc; \
-  ScopedJavaFunc scoped(thr, caller_pc); \
-/**/
+  (void)thr;
 
 void __tsan_java_init(jptr heap_begin, jptr heap_size) {
-  SCOPED_JAVA_FUNC(__tsan_java_init);
-  DPrintf("#%d: java_init(%p, %p)\n", thr->tid, heap_begin, heap_size);
-  CHECK_EQ(jctx, 0);
-  CHECK_GT(heap_begin, 0);
-  CHECK_GT(heap_size, 0);
-  CHECK_EQ(heap_begin % kHeapAlignment, 0);
-  CHECK_EQ(heap_size % kHeapAlignment, 0);
-  CHECK_LT(heap_begin, heap_begin + heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_init);
+  Initialize(thr);
+  DPrintf("#%d: java_init(0x%zx, 0x%zx)\n", thr->tid, heap_begin, heap_size);
+  DCHECK_EQ(jctx, 0);
+  DCHECK_GT(heap_begin, 0);
+  DCHECK_GT(heap_size, 0);
+  DCHECK_EQ(heap_begin % kHeapAlignment, 0);
+  DCHECK_EQ(heap_size % kHeapAlignment, 0);
+  DCHECK_LT(heap_begin, heap_begin + heap_size);
   jctx = new(jctx_buf) JavaContext(heap_begin, heap_size);
 }
 
 int  __tsan_java_fini() {
-  SCOPED_JAVA_FUNC(__tsan_java_fini);
+  JAVA_FUNC_ENTER(__tsan_java_fini);
   DPrintf("#%d: java_fini()\n", thr->tid);
-  CHECK_NE(jctx, 0);
+  DCHECK_NE(jctx, 0);
   // FIXME(dvyukov): this does not call atexit() callbacks.
   int status = Finalize(thr);
   DPrintf("#%d: java_fini() = %d\n", thr->tid, status);
@@ -87,74 +84,65 @@ int  __tsan_java_fini() {
 }
 
 void __tsan_java_alloc(jptr ptr, jptr size) {
-  SCOPED_JAVA_FUNC(__tsan_java_alloc);
-  DPrintf("#%d: java_alloc(%p, %p)\n", thr->tid, ptr, size);
-  CHECK_NE(jctx, 0);
-  CHECK_NE(size, 0);
-  CHECK_EQ(ptr % kHeapAlignment, 0);
-  CHECK_EQ(size % kHeapAlignment, 0);
-  CHECK_GE(ptr, jctx->heap_begin);
-  CHECK_LE(ptr + size, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_alloc);
+  DPrintf("#%d: java_alloc(0x%zx, 0x%zx)\n", thr->tid, ptr, size);
+  DCHECK_NE(jctx, 0);
+  DCHECK_NE(size, 0);
+  DCHECK_EQ(ptr % kHeapAlignment, 0);
+  DCHECK_EQ(size % kHeapAlignment, 0);
+  DCHECK_GE(ptr, jctx->heap_begin);
+  DCHECK_LE(ptr + size, jctx->heap_begin + jctx->heap_size);
 
-  OnUserAlloc(thr, pc, ptr, size, false);
+  OnUserAlloc(thr, 0, ptr, size, false);
 }
 
 void __tsan_java_free(jptr ptr, jptr size) {
-  SCOPED_JAVA_FUNC(__tsan_java_free);
-  DPrintf("#%d: java_free(%p, %p)\n", thr->tid, ptr, size);
-  CHECK_NE(jctx, 0);
-  CHECK_NE(size, 0);
-  CHECK_EQ(ptr % kHeapAlignment, 0);
-  CHECK_EQ(size % kHeapAlignment, 0);
-  CHECK_GE(ptr, jctx->heap_begin);
-  CHECK_LE(ptr + size, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_free);
+  DPrintf("#%d: java_free(0x%zx, 0x%zx)\n", thr->tid, ptr, size);
+  DCHECK_NE(jctx, 0);
+  DCHECK_NE(size, 0);
+  DCHECK_EQ(ptr % kHeapAlignment, 0);
+  DCHECK_EQ(size % kHeapAlignment, 0);
+  DCHECK_GE(ptr, jctx->heap_begin);
+  DCHECK_LE(ptr + size, jctx->heap_begin + jctx->heap_size);
 
   ctx->metamap.FreeRange(thr->proc(), ptr, size);
 }
 
 void __tsan_java_move(jptr src, jptr dst, jptr size) {
-  SCOPED_JAVA_FUNC(__tsan_java_move);
-  DPrintf("#%d: java_move(%p, %p, %p)\n", thr->tid, src, dst, size);
-  CHECK_NE(jctx, 0);
-  CHECK_NE(size, 0);
-  CHECK_EQ(src % kHeapAlignment, 0);
-  CHECK_EQ(dst % kHeapAlignment, 0);
-  CHECK_EQ(size % kHeapAlignment, 0);
-  CHECK_GE(src, jctx->heap_begin);
-  CHECK_LE(src + size, jctx->heap_begin + jctx->heap_size);
-  CHECK_GE(dst, jctx->heap_begin);
-  CHECK_LE(dst + size, jctx->heap_begin + jctx->heap_size);
-  CHECK_NE(dst, src);
-  CHECK_NE(size, 0);
+  JAVA_FUNC_ENTER(__tsan_java_move);
+  DPrintf("#%d: java_move(0x%zx, 0x%zx, 0x%zx)\n", thr->tid, src, dst, size);
+  DCHECK_NE(jctx, 0);
+  DCHECK_NE(size, 0);
+  DCHECK_EQ(src % kHeapAlignment, 0);
+  DCHECK_EQ(dst % kHeapAlignment, 0);
+  DCHECK_EQ(size % kHeapAlignment, 0);
+  DCHECK_GE(src, jctx->heap_begin);
+  DCHECK_LE(src + size, jctx->heap_begin + jctx->heap_size);
+  DCHECK_GE(dst, jctx->heap_begin);
+  DCHECK_LE(dst + size, jctx->heap_begin + jctx->heap_size);
+  DCHECK_NE(dst, src);
+  DCHECK_NE(size, 0);
 
   // Assuming it's not running concurrently with threads that do
   // memory accesses and mutex operations (stop-the-world phase).
   ctx->metamap.MoveMemory(src, dst, size);
 
-  // Move shadow.
-  u64 *s = (u64*)MemToShadow(src);
-  u64 *d = (u64*)MemToShadow(dst);
-  u64 *send = (u64*)MemToShadow(src + size);
-  uptr inc = 1;
-  if (dst > src) {
-    s = (u64*)MemToShadow(src + size) - 1;
-    d = (u64*)MemToShadow(dst + size) - 1;
-    send = (u64*)MemToShadow(src) - 1;
-    inc = -1;
-  }
-  for (; s != send; s += inc, d += inc) {
-    *d = *s;
-    *s = 0;
-  }
+  // Clear the destination shadow range.
+  // We used to move shadow from src to dst, but the trace format does not
+  // support that anymore as it contains addresses of accesses.
+  RawShadow *d = MemToShadow(dst);
+  RawShadow *dend = MemToShadow(dst + size);
+  internal_memset(d, 0, (dend - d) * sizeof(*d));
 }
 
 jptr __tsan_java_find(jptr *from_ptr, jptr to) {
-  SCOPED_JAVA_FUNC(__tsan_java_find);
-  DPrintf("#%d: java_find(&%p, %p)\n", *from_ptr, to);
-  CHECK_EQ((*from_ptr) % kHeapAlignment, 0);
-  CHECK_EQ(to % kHeapAlignment, 0);
-  CHECK_GE(*from_ptr, jctx->heap_begin);
-  CHECK_LE(to, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_find);
+  DPrintf("#%d: java_find(&0x%zx, 0x%zx)\n", thr->tid, *from_ptr, to);
+  DCHECK_EQ((*from_ptr) % kHeapAlignment, 0);
+  DCHECK_EQ(to % kHeapAlignment, 0);
+  DCHECK_GE(*from_ptr, jctx->heap_begin);
+  DCHECK_LE(to, jctx->heap_begin + jctx->heap_size);
   for (uptr from = *from_ptr; from < to; from += kHeapAlignment) {
     MBlock *b = ctx->metamap.GetBlock(from);
     if (b) {
@@ -166,101 +154,105 @@ jptr __tsan_java_find(jptr *from_ptr, jptr to) {
 }
 
 void __tsan_java_finalize() {
-  SCOPED_JAVA_FUNC(__tsan_java_finalize);
-  DPrintf("#%d: java_mutex_finalize()\n", thr->tid);
-  AcquireGlobal(thr, 0);
+  JAVA_FUNC_ENTER(__tsan_java_finalize);
+  DPrintf("#%d: java_finalize()\n", thr->tid);
+  AcquireGlobal(thr);
 }
 
 void __tsan_java_mutex_lock(jptr addr) {
-  SCOPED_JAVA_FUNC(__tsan_java_mutex_lock);
-  DPrintf("#%d: java_mutex_lock(%p)\n", thr->tid, addr);
-  CHECK_NE(jctx, 0);
-  CHECK_GE(addr, jctx->heap_begin);
-  CHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_mutex_lock);
+  DPrintf("#%d: java_mutex_lock(0x%zx)\n", thr->tid, addr);
+  DCHECK_NE(jctx, 0);
+  DCHECK_GE(addr, jctx->heap_begin);
+  DCHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
 
-  MutexPostLock(thr, pc, addr, MutexFlagLinkerInit | MutexFlagWriteReentrant |
-      MutexFlagDoPreLockOnPostLock);
+  MutexPostLock(thr, 0, addr,
+                MutexFlagLinkerInit | MutexFlagWriteReentrant |
+                    MutexFlagDoPreLockOnPostLock);
 }
 
 void __tsan_java_mutex_unlock(jptr addr) {
-  SCOPED_JAVA_FUNC(__tsan_java_mutex_unlock);
-  DPrintf("#%d: java_mutex_unlock(%p)\n", thr->tid, addr);
-  CHECK_NE(jctx, 0);
-  CHECK_GE(addr, jctx->heap_begin);
-  CHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_mutex_unlock);
+  DPrintf("#%d: java_mutex_unlock(0x%zx)\n", thr->tid, addr);
+  DCHECK_NE(jctx, 0);
+  DCHECK_GE(addr, jctx->heap_begin);
+  DCHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
 
-  MutexUnlock(thr, pc, addr);
+  MutexUnlock(thr, 0, addr);
 }
 
 void __tsan_java_mutex_read_lock(jptr addr) {
-  SCOPED_JAVA_FUNC(__tsan_java_mutex_read_lock);
-  DPrintf("#%d: java_mutex_read_lock(%p)\n", thr->tid, addr);
-  CHECK_NE(jctx, 0);
-  CHECK_GE(addr, jctx->heap_begin);
-  CHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_mutex_read_lock);
+  DPrintf("#%d: java_mutex_read_lock(0x%zx)\n", thr->tid, addr);
+  DCHECK_NE(jctx, 0);
+  DCHECK_GE(addr, jctx->heap_begin);
+  DCHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
 
-  MutexPostReadLock(thr, pc, addr, MutexFlagLinkerInit |
-      MutexFlagWriteReentrant | MutexFlagDoPreLockOnPostLock);
+  MutexPostReadLock(thr, 0, addr,
+                    MutexFlagLinkerInit | MutexFlagWriteReentrant |
+                        MutexFlagDoPreLockOnPostLock);
 }
 
 void __tsan_java_mutex_read_unlock(jptr addr) {
-  SCOPED_JAVA_FUNC(__tsan_java_mutex_read_unlock);
-  DPrintf("#%d: java_mutex_read_unlock(%p)\n", thr->tid, addr);
-  CHECK_NE(jctx, 0);
-  CHECK_GE(addr, jctx->heap_begin);
-  CHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_mutex_read_unlock);
+  DPrintf("#%d: java_mutex_read_unlock(0x%zx)\n", thr->tid, addr);
+  DCHECK_NE(jctx, 0);
+  DCHECK_GE(addr, jctx->heap_begin);
+  DCHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
 
-  MutexReadUnlock(thr, pc, addr);
+  MutexReadUnlock(thr, 0, addr);
 }
 
 void __tsan_java_mutex_lock_rec(jptr addr, int rec) {
-  SCOPED_JAVA_FUNC(__tsan_java_mutex_lock_rec);
-  DPrintf("#%d: java_mutex_lock_rec(%p, %d)\n", thr->tid, addr, rec);
-  CHECK_NE(jctx, 0);
-  CHECK_GE(addr, jctx->heap_begin);
-  CHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
-  CHECK_GT(rec, 0);
+  JAVA_FUNC_ENTER(__tsan_java_mutex_lock_rec);
+  DPrintf("#%d: java_mutex_lock_rec(0x%zx, %d)\n", thr->tid, addr, rec);
+  DCHECK_NE(jctx, 0);
+  DCHECK_GE(addr, jctx->heap_begin);
+  DCHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
+  DCHECK_GT(rec, 0);
 
-  MutexPostLock(thr, pc, addr, MutexFlagLinkerInit | MutexFlagWriteReentrant |
-      MutexFlagDoPreLockOnPostLock | MutexFlagRecursiveLock, rec);
+  MutexPostLock(thr, 0, addr,
+                MutexFlagLinkerInit | MutexFlagWriteReentrant |
+                    MutexFlagDoPreLockOnPostLock | MutexFlagRecursiveLock,
+                rec);
 }
 
 int __tsan_java_mutex_unlock_rec(jptr addr) {
-  SCOPED_JAVA_FUNC(__tsan_java_mutex_unlock_rec);
-  DPrintf("#%d: java_mutex_unlock_rec(%p)\n", thr->tid, addr);
-  CHECK_NE(jctx, 0);
-  CHECK_GE(addr, jctx->heap_begin);
-  CHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_mutex_unlock_rec);
+  DPrintf("#%d: java_mutex_unlock_rec(0x%zx)\n", thr->tid, addr);
+  DCHECK_NE(jctx, 0);
+  DCHECK_GE(addr, jctx->heap_begin);
+  DCHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
 
-  return MutexUnlock(thr, pc, addr, MutexFlagRecursiveUnlock);
+  return MutexUnlock(thr, 0, addr, MutexFlagRecursiveUnlock);
 }
 
 void __tsan_java_acquire(jptr addr) {
-  SCOPED_JAVA_FUNC(__tsan_java_acquire);
-  DPrintf("#%d: java_acquire(%p)\n", thr->tid, addr);
-  CHECK_NE(jctx, 0);
-  CHECK_GE(addr, jctx->heap_begin);
-  CHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_acquire);
+  DPrintf("#%d: java_acquire(0x%zx)\n", thr->tid, addr);
+  DCHECK_NE(jctx, 0);
+  DCHECK_GE(addr, jctx->heap_begin);
+  DCHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
 
-  Acquire(thr, caller_pc, addr);
+  Acquire(thr, 0, addr);
 }
 
 void __tsan_java_release(jptr addr) {
-  SCOPED_JAVA_FUNC(__tsan_java_release);
-  DPrintf("#%d: java_release(%p)\n", thr->tid, addr);
-  CHECK_NE(jctx, 0);
-  CHECK_GE(addr, jctx->heap_begin);
-  CHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_release);
+  DPrintf("#%d: java_release(0x%zx)\n", thr->tid, addr);
+  DCHECK_NE(jctx, 0);
+  DCHECK_GE(addr, jctx->heap_begin);
+  DCHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
 
-  Release(thr, caller_pc, addr);
+  Release(thr, 0, addr);
 }
 
 void __tsan_java_release_store(jptr addr) {
-  SCOPED_JAVA_FUNC(__tsan_java_release);
-  DPrintf("#%d: java_release_store(%p)\n", thr->tid, addr);
-  CHECK_NE(jctx, 0);
-  CHECK_GE(addr, jctx->heap_begin);
-  CHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
+  JAVA_FUNC_ENTER(__tsan_java_release);
+  DPrintf("#%d: java_release_store(0x%zx)\n", thr->tid, addr);
+  DCHECK_NE(jctx, 0);
+  DCHECK_GE(addr, jctx->heap_begin);
+  DCHECK_LT(addr, jctx->heap_begin + jctx->heap_size);
 
-  ReleaseStore(thr, caller_pc, addr);
+  ReleaseStore(thr, 0, addr);
 }
diff --git a/compiler-rt/lib/tsan/rtl/tsan_mman.cpp b/compiler-rt/lib/tsan/rtl/tsan_mman.cpp
index 7765bc070522..ef97ad0bc94e 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_mman.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_mman.cpp
@@ -148,7 +148,7 @@ static void SignalUnsafeCall(ThreadState *thr, uptr pc) {
   ObtainCurrentStack(thr, pc, &stack);
   if (IsFiredSuppression(ctx, ReportTypeSignalUnsafe, stack))
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
+  ThreadRegistryLock l(&ctx->thread_registry);
   ScopedReport rep(ReportTypeSignalUnsafe);
   rep.AddStack(stack, true);
   OutputReport(thr, rep);
@@ -218,9 +218,9 @@ void *user_reallocarray(ThreadState *thr, uptr pc, void *p, uptr size, uptr n) {
 }
 
 void OnUserAlloc(ThreadState *thr, uptr pc, uptr p, uptr sz, bool write) {
-  DPrintf("#%d: alloc(%zu) = %p\n", thr->tid, sz, p);
+  DPrintf("#%d: alloc(%zu) = 0x%zx\n", thr->tid, sz, p);
   ctx->metamap.AllocBlock(thr, pc, p, sz);
-  if (write && thr->ignore_reads_and_writes == 0)
+  if (write && thr->ignore_reads_and_writes == 0 && thr->is_inited)
     MemoryRangeImitateWrite(thr, pc, (uptr)p, sz);
   else
     MemoryResetRange(thr, pc, (uptr)p, sz);
@@ -229,8 +229,8 @@ void OnUserAlloc(ThreadState *thr, uptr pc, uptr p, uptr sz, bool write) {
 void OnUserFree(ThreadState *thr, uptr pc, uptr p, bool write) {
   CHECK_NE(p, (void*)0);
   uptr sz = ctx->metamap.FreeBlock(thr->proc(), p);
-  DPrintf("#%d: free(%p, %zu)\n", thr->tid, p, sz);
-  if (write && thr->ignore_reads_and_writes == 0)
+  DPrintf("#%d: free(0x%zx, %zu)\n", thr->tid, p, sz);
+  if (write && thr->ignore_reads_and_writes == 0 && thr->is_inited)
     MemoryRangeFreed(thr, pc, (uptr)p, sz);
 }
 
@@ -336,7 +336,7 @@ void invoke_free_hook(void *ptr) {
   RunFreeHooks(ptr);
 }
 
-void *internal_alloc(MBlockType typ, uptr sz) {
+void *Alloc(uptr sz) {
   ThreadState *thr = cur_thread();
   if (thr->nomalloc) {
     thr->nomalloc = 0;  // CHECK calls internal_malloc().
@@ -345,7 +345,7 @@ void *internal_alloc(MBlockType typ, uptr sz) {
   return InternalAlloc(sz, &thr->proc()->internal_alloc_cache);
 }
 
-void internal_free(void *p) {
+void FreeImpl(void *p) {
   ThreadState *thr = cur_thread();
   if (thr->nomalloc) {
     thr->nomalloc = 0;  // CHECK calls internal_malloc().
diff --git a/compiler-rt/lib/tsan/rtl/tsan_mman.h b/compiler-rt/lib/tsan/rtl/tsan_mman.h
index a5280d4472c9..efea5e5abdec 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_mman.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_mman.h
@@ -47,42 +47,29 @@ uptr user_alloc_usable_size(const void *p);
 void invoke_malloc_hook(void *ptr, uptr size);
 void invoke_free_hook(void *ptr);
 
-enum MBlockType {
-  MBlockScopedBuf,
-  MBlockString,
-  MBlockStackTrace,
-  MBlockShadowStack,
-  MBlockSync,
-  MBlockClock,
-  MBlockThreadContex,
-  MBlockDeadInfo,
-  MBlockRacyStacks,
-  MBlockRacyAddresses,
-  MBlockAtExit,
-  MBlockFlag,
-  MBlockReport,
-  MBlockReportMop,
-  MBlockReportThread,
-  MBlockReportMutex,
-  MBlockReportLoc,
-  MBlockReportStack,
-  MBlockSuppression,
-  MBlockExpectRace,
-  MBlockSignal,
-  MBlockJmpBuf,
+// For internal data structures.
+void *Alloc(uptr sz);
+void FreeImpl(void *p);
 
-  // This must be the last.
-  MBlockTypeCount
-};
+template <typename T, typename... Args>
+T *New(Args &&...args) {
+  return new (Alloc(sizeof(T))) T(static_cast<Args &&>(args)...);
+}
 
-// For internal data structures.
-void *internal_alloc(MBlockType typ, uptr sz);
-void internal_free(void *p);
+template <typename T>
+void Free(T *&p) {
+  if (p == nullptr)
+    return;
+  FreeImpl(p);
+  p = nullptr;
+}
 
 template <typename T>
-void DestroyAndFree(T *p) {
+void DestroyAndFree(T *&p) {
+  if (p == nullptr)
+    return;
   p->~T();
-  internal_free(p);
+  Free(p);
 }
 
 }  // namespace __tsan
diff --git a/compiler-rt/lib/tsan/rtl/tsan_mutexset.cpp b/compiler-rt/lib/tsan/rtl/tsan_mutexset.cpp
index 813fa3bca936..735179686ba9 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_mutexset.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_mutexset.cpp
@@ -10,15 +10,13 @@
 //
 //===----------------------------------------------------------------------===//
 #include "tsan_mutexset.h"
+
+#include "sanitizer_common/sanitizer_placement_new.h"
 #include "tsan_rtl.h"
 
 namespace __tsan {
 
-const uptr MutexSet::kMaxSize;
-
 MutexSet::MutexSet() {
-  size_ = 0;
-  internal_memset(&descs_, 0, sizeof(descs_));
 }
 
 void MutexSet::Add(u64 id, bool write, u64 epoch) {
@@ -44,9 +42,12 @@ void MutexSet::Add(u64 id, bool write, u64 epoch) {
     CHECK_EQ(size_, kMaxSize - 1);
   }
   // Add new mutex descriptor.
+  descs_[size_].addr = 0;
+  descs_[size_].stack_id = kInvalidStackID;
   descs_[size_].id = id;
   descs_[size_].write = write;
   descs_[size_].epoch = epoch;
+  descs_[size_].seq = seq_++;
   descs_[size_].count = 1;
   size_++;
 }
@@ -70,6 +71,46 @@ void MutexSet::Remove(u64 id) {
   }
 }
 
+void MutexSet::AddAddr(uptr addr, StackID stack_id, bool write) {
+  // Look up existing mutex with the same id.
+  for (uptr i = 0; i < size_; i++) {
+    if (descs_[i].addr == addr) {
+      descs_[i].count++;
+      descs_[i].seq = seq_++;
+      return;
+    }
+  }
+  // On overflow, find the oldest mutex and drop it.
+  if (size_ == kMaxSize) {
+    uptr min = 0;
+    for (uptr i = 0; i < size_; i++) {
+      if (descs_[i].seq < descs_[min].seq)
+        min = i;
+    }
+    RemovePos(min);
+    CHECK_EQ(size_, kMaxSize - 1);
+  }
+  // Add new mutex descriptor.
+  descs_[size_].addr = addr;
+  descs_[size_].stack_id = stack_id;
+  descs_[size_].id = 0;
+  descs_[size_].write = write;
+  descs_[size_].epoch = 0;
+  descs_[size_].seq = seq_++;
+  descs_[size_].count = 1;
+  size_++;
+}
+
+void MutexSet::DelAddr(uptr addr, bool destroy) {
+  for (uptr i = 0; i < size_; i++) {
+    if (descs_[i].addr == addr) {
+      if (destroy || --descs_[i].count == 0)
+        RemovePos(i);
+      return;
+    }
+  }
+}
+
 void MutexSet::RemovePos(uptr i) {
   CHECK_LT(i, size_);
   descs_[i] = descs_[size_ - 1];
@@ -85,4 +126,7 @@ MutexSet::Desc MutexSet::Get(uptr i) const {
   return descs_[i];
 }
 
+DynamicMutexSet::DynamicMutexSet() : ptr_(New<MutexSet>()) {}
+DynamicMutexSet::~DynamicMutexSet() { DestroyAndFree(ptr_); }
+
 }  // namespace __tsan
diff --git a/compiler-rt/lib/tsan/rtl/tsan_mutexset.h b/compiler-rt/lib/tsan/rtl/tsan_mutexset.h
index d63881f40290..93776a664135 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_mutexset.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_mutexset.h
@@ -21,12 +21,22 @@ class MutexSet {
  public:
   // Holds limited number of mutexes.
   // The oldest mutexes are discarded on overflow.
-  static const uptr kMaxSize = 16;
+  static constexpr uptr kMaxSize = 16;
   struct Desc {
+    uptr addr;
+    StackID stack_id;
     u64 id;
     u64 epoch;
-    int count;
+    u32 seq;
+    u32 count;
     bool write;
+
+    Desc() { internal_memset(this, 0, sizeof(*this)); }
+    Desc(const Desc& other) { *this = other; }
+    Desc& operator=(const MutexSet::Desc& other) {
+      internal_memcpy(this, &other, sizeof(*this));
+      return *this;
+    }
   };
 
   MutexSet();
@@ -34,21 +44,37 @@ class MutexSet {
   void Add(u64 id, bool write, u64 epoch);
   void Del(u64 id, bool write);
   void Remove(u64 id);  // Removes the mutex completely (if it's destroyed).
+  void AddAddr(uptr addr, StackID stack_id, bool write);
+  void DelAddr(uptr addr, bool destroy = false);
   uptr Size() const;
   Desc Get(uptr i) const;
 
-  void operator=(const MutexSet &other) {
-    internal_memcpy(this, &other, sizeof(*this));
-  }
-
  private:
 #if !SANITIZER_GO
-  uptr size_;
+  u32 seq_ = 0;
+  uptr size_ = 0;
   Desc descs_[kMaxSize];
-#endif
 
   void RemovePos(uptr i);
-  MutexSet(const MutexSet&);
+#endif
+};
+
+// MutexSet is too large to live on stack.
+// DynamicMutexSet can be use used to create local MutexSet's.
+class DynamicMutexSet {
+ public:
+  DynamicMutexSet();
+  ~DynamicMutexSet();
+  MutexSet* operator->() { return ptr_; }
+  operator MutexSet*() { return ptr_; }
+  DynamicMutexSet(const DynamicMutexSet&) = delete;
+  DynamicMutexSet& operator=(const DynamicMutexSet&) = delete;
+
+ private:
+  MutexSet* ptr_;
+#if SANITIZER_GO
+  MutexSet set_;
+#endif
 };
 
 // Go does not have mutexes, so do not spend memory and time.
@@ -59,9 +85,12 @@ MutexSet::MutexSet() {}
 void MutexSet::Add(u64 id, bool write, u64 epoch) {}
 void MutexSet::Del(u64 id, bool write) {}
 void MutexSet::Remove(u64 id) {}
-void MutexSet::RemovePos(uptr i) {}
+void MutexSet::AddAddr(uptr addr, StackID stack_id, bool write) {}
+void MutexSet::DelAddr(uptr addr, bool destroy) {}
 uptr MutexSet::Size() const { return 0; }
 MutexSet::Desc MutexSet::Get(uptr i) const { return Desc(); }
+DynamicMutexSet::DynamicMutexSet() : ptr_(&set_) {}
+DynamicMutexSet::~DynamicMutexSet() {}
 #endif
 
 }  // namespace __tsan
diff --git a/compiler-rt/lib/tsan/rtl/tsan_platform.h b/compiler-rt/lib/tsan/rtl/tsan_platform.h
index 8bd218e25fd6..7ff0acace8f6 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_platform.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_platform.h
@@ -23,21 +23,19 @@
 
 namespace __tsan {
 
-#if defined(__x86_64__)
-#define HAS_48_BIT_ADDRESS_SPACE 1
-#elif SANITIZER_IOSSIM // arm64 iOS simulators (order of #if matters)
-#define HAS_48_BIT_ADDRESS_SPACE 1
-#elif SANITIZER_IOS // arm64 iOS devices (order of #if matters)
-#define HAS_48_BIT_ADDRESS_SPACE 0
-#elif SANITIZER_MAC // arm64 macOS (order of #if matters)
-#define HAS_48_BIT_ADDRESS_SPACE 1
-#else
-#define HAS_48_BIT_ADDRESS_SPACE 0
-#endif
-
-#if !SANITIZER_GO
+enum {
+  // App memory is not mapped onto shadow memory range.
+  kBrokenMapping = 1 << 0,
+  // Mapping app memory and back does not produce the same address,
+  // this can lead to wrong addresses in reports and potentially
+  // other bad consequences.
+  kBrokenReverseMapping = 1 << 1,
+  // Mapping is non-linear for linear user range.
+  // This is bad and can lead to unpredictable memory corruptions, etc
+  // because range access functions assume linearity.
+  kBrokenLinearity = 1 << 2,
+};
 
-#if HAS_48_BIT_ADDRESS_SPACE
 /*
 C/C++ on linux/x86_64 and freebsd/x86_64
 0000 0000 1000 - 0080 0000 0000: main binary and/or MAP_32BIT mappings (512GB)
@@ -65,9 +63,8 @@ C/C++ on netbsd/amd64 can reuse the same mapping:
  * Stack on NetBSD/amd64 has prereserved 128MB.
  * Heap grows downwards (top-down).
  * ASLR must be disabled per-process or globally.
-
 */
-struct Mapping {
+struct Mapping48AddressSpace {
   static const uptr kMetaShadowBeg = 0x300000000000ull;
   static const uptr kMetaShadowEnd = 0x340000000000ull;
   static const uptr kTraceMemBeg   = 0x600000000000ull;
@@ -82,13 +79,12 @@ struct Mapping {
   static const uptr kMidAppMemEnd  = 0x568000000000ull;
   static const uptr kHiAppMemBeg   = 0x7e8000000000ull;
   static const uptr kHiAppMemEnd   = 0x800000000000ull;
-  static const uptr kAppMemMsk     = 0x780000000000ull;
-  static const uptr kAppMemXor     = 0x040000000000ull;
+  static const uptr kShadowMsk = 0x780000000000ull;
+  static const uptr kShadowXor = 0x040000000000ull;
+  static const uptr kShadowAdd = 0x000000000000ull;
   static const uptr kVdsoBeg       = 0xf000000000000000ull;
 };
 
-#define TSAN_MID_APP_RANGE 1
-#elif defined(__mips64)
 /*
 C/C++ on linux/mips64 (40-bit VMA)
 0000 0000 00 - 0100 0000 00: -                                           (4 GB)
@@ -105,7 +101,7 @@ fe00 0000 00 - ff00 0000 00: heap                                        (4 GB)
 ff00 0000 00 - ff80 0000 00: -                                           (2 GB)
 ff80 0000 00 - ffff ffff ff: modules and main thread stack              (<2 GB)
 */
-struct Mapping40 {
+struct MappingMips64_40 {
   static const uptr kMetaShadowBeg = 0x4000000000ull;
   static const uptr kMetaShadowEnd = 0x5000000000ull;
   static const uptr kTraceMemBeg   = 0xb000000000ull;
@@ -120,14 +116,12 @@ struct Mapping40 {
   static const uptr kMidAppMemEnd  = 0xab00000000ull;
   static const uptr kHiAppMemBeg   = 0xff80000000ull;
   static const uptr kHiAppMemEnd   = 0xffffffffffull;
-  static const uptr kAppMemMsk     = 0xf800000000ull;
-  static const uptr kAppMemXor     = 0x0800000000ull;
+  static const uptr kShadowMsk = 0xf800000000ull;
+  static const uptr kShadowXor = 0x0800000000ull;
+  static const uptr kShadowAdd = 0x0000000000ull;
   static const uptr kVdsoBeg       = 0xfffff00000ull;
 };
 
-#define TSAN_MID_APP_RANGE 1
-#define TSAN_RUNTIME_VMA 1
-#elif defined(__aarch64__) && defined(__APPLE__)
 /*
 C/C++ on Darwin/iOS/ARM64 (36-bit VMA, 64 GB VM)
 0000 0000 00 - 0100 0000 00: -                                    (4 GB)
@@ -141,7 +135,7 @@ C/C++ on Darwin/iOS/ARM64 (36-bit VMA, 64 GB VM)
 0f00 0000 00 - 0fc0 0000 00: traces                               (3 GB)
 0fc0 0000 00 - 1000 0000 00: -
 */
-struct Mapping {
+struct MappingAppleAarch64 {
   static const uptr kLoAppMemBeg   = 0x0100000000ull;
   static const uptr kLoAppMemEnd   = 0x0200000000ull;
   static const uptr kHeapMemBeg    = 0x0200000000ull;
@@ -154,18 +148,14 @@ struct Mapping {
   static const uptr kTraceMemEnd   = 0x0fc0000000ull;
   static const uptr kHiAppMemBeg   = 0x0fc0000000ull;
   static const uptr kHiAppMemEnd   = 0x0fc0000000ull;
-  static const uptr kAppMemMsk     =          0x0ull;
-  static const uptr kAppMemXor     =          0x0ull;
+  static const uptr kShadowMsk = 0x0ull;
+  static const uptr kShadowXor = 0x0ull;
+  static const uptr kShadowAdd = 0x0ull;
   static const uptr kVdsoBeg       = 0x7000000000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
 };
 
-#elif defined(__aarch64__) && !defined(__APPLE__)
-// AArch64 supports multiple VMA which leads to multiple address transformation
-// functions.  To support these multiple VMAS transformations and mappings TSAN
-// runtime for AArch64 uses an external memory read (vmaSize) to select which
-// mapping to use.  Although slower, it make a same instrumented binary run on
-// multiple kernels.
-
 /*
 C/C++ on linux/aarch64 (39-bit VMA)
 0000 0010 00 - 0100 0000 00: main binary
@@ -181,7 +171,7 @@ C/C++ on linux/aarch64 (39-bit VMA)
 7c00 0000 00 - 7d00 0000 00: heap
 7d00 0000 00 - 7fff ffff ff: modules and main thread stack
 */
-struct Mapping39 {
+struct MappingAarch64_39 {
   static const uptr kLoAppMemBeg   = 0x0000001000ull;
   static const uptr kLoAppMemEnd   = 0x0100000000ull;
   static const uptr kShadowBeg     = 0x0800000000ull;
@@ -196,8 +186,9 @@ struct Mapping39 {
   static const uptr kHeapMemEnd    = 0x7d00000000ull;
   static const uptr kHiAppMemBeg   = 0x7e00000000ull;
   static const uptr kHiAppMemEnd   = 0x7fffffffffull;
-  static const uptr kAppMemMsk     = 0x7800000000ull;
-  static const uptr kAppMemXor     = 0x0200000000ull;
+  static const uptr kShadowMsk = 0x7800000000ull;
+  static const uptr kShadowXor = 0x0200000000ull;
+  static const uptr kShadowAdd = 0x0000000000ull;
   static const uptr kVdsoBeg       = 0x7f00000000ull;
 };
 
@@ -216,7 +207,8 @@ C/C++ on linux/aarch64 (42-bit VMA)
 3e000 0000 00 - 3f000 0000 00: heap
 3f000 0000 00 - 3ffff ffff ff: modules and main thread stack
 */
-struct Mapping42 {
+struct MappingAarch64_42 {
+  static const uptr kBroken = kBrokenReverseMapping;
   static const uptr kLoAppMemBeg   = 0x00000001000ull;
   static const uptr kLoAppMemEnd   = 0x01000000000ull;
   static const uptr kShadowBeg     = 0x10000000000ull;
@@ -231,12 +223,13 @@ struct Mapping42 {
   static const uptr kHeapMemEnd    = 0x3f000000000ull;
   static const uptr kHiAppMemBeg   = 0x3f000000000ull;
   static const uptr kHiAppMemEnd   = 0x3ffffffffffull;
-  static const uptr kAppMemMsk     = 0x3c000000000ull;
-  static const uptr kAppMemXor     = 0x04000000000ull;
+  static const uptr kShadowMsk = 0x3c000000000ull;
+  static const uptr kShadowXor = 0x04000000000ull;
+  static const uptr kShadowAdd = 0x00000000000ull;
   static const uptr kVdsoBeg       = 0x37f00000000ull;
 };
 
-struct Mapping48 {
+struct MappingAarch64_48 {
   static const uptr kLoAppMemBeg   = 0x0000000001000ull;
   static const uptr kLoAppMemEnd   = 0x0000200000000ull;
   static const uptr kShadowBeg     = 0x0002000000000ull;
@@ -251,22 +244,12 @@ struct Mapping48 {
   static const uptr kHeapMemEnd    = 0x0ffff00000000ull;
   static const uptr kHiAppMemBeg   = 0x0ffff00000000ull;
   static const uptr kHiAppMemEnd   = 0x1000000000000ull;
-  static const uptr kAppMemMsk     = 0x0fff800000000ull;
-  static const uptr kAppMemXor     = 0x0000800000000ull;
+  static const uptr kShadowMsk = 0x0fff800000000ull;
+  static const uptr kShadowXor = 0x0000800000000ull;
+  static const uptr kShadowAdd = 0x0000000000000ull;
   static const uptr kVdsoBeg       = 0xffff000000000ull;
 };
 
-// Indicates the runtime will define the memory regions at runtime.
-#define TSAN_RUNTIME_VMA 1
-// Indicates that mapping defines a mid range memory segment.
-#define TSAN_MID_APP_RANGE 1
-#elif defined(__powerpc64__)
-// PPC64 supports multiple VMA which leads to multiple address transformation
-// functions.  To support these multiple VMAS transformations and mappings TSAN
-// runtime for PPC64 uses an external memory read (vmaSize) to select which
-// mapping to use.  Although slower, it make a same instrumented binary run on
-// multiple kernels.
-
 /*
 C/C++ on linux/powerpc64 (44-bit VMA)
 0000 0000 0100 - 0001 0000 0000: main binary
@@ -281,7 +264,9 @@ C/C++ on linux/powerpc64 (44-bit VMA)
 0f50 0000 0000 - 0f60 0000 0000: -
 0f60 0000 0000 - 1000 0000 0000: modules and main thread stack
 */
-struct Mapping44 {
+struct MappingPPC64_44 {
+  static const uptr kBroken =
+      kBrokenMapping | kBrokenReverseMapping | kBrokenLinearity;
   static const uptr kMetaShadowBeg = 0x0b0000000000ull;
   static const uptr kMetaShadowEnd = 0x0d0000000000ull;
   static const uptr kTraceMemBeg   = 0x0d0000000000ull;
@@ -294,9 +279,12 @@ struct Mapping44 {
   static const uptr kHeapMemEnd    = 0x0f5000000000ull;
   static const uptr kHiAppMemBeg   = 0x0f6000000000ull;
   static const uptr kHiAppMemEnd   = 0x100000000000ull; // 44 bits
-  static const uptr kAppMemMsk     = 0x0f0000000000ull;
-  static const uptr kAppMemXor     = 0x002100000000ull;
+  static const uptr kShadowMsk = 0x0f0000000000ull;
+  static const uptr kShadowXor = 0x002100000000ull;
+  static const uptr kShadowAdd = 0x000000000000ull;
   static const uptr kVdsoBeg       = 0x3c0000000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
 };
 
 /*
@@ -313,7 +301,7 @@ C/C++ on linux/powerpc64 (46-bit VMA)
 3e00 0000 0000 - 3e80 0000 0000: -
 3e80 0000 0000 - 4000 0000 0000: modules and main thread stack
 */
-struct Mapping46 {
+struct MappingPPC64_46 {
   static const uptr kMetaShadowBeg = 0x100000000000ull;
   static const uptr kMetaShadowEnd = 0x200000000000ull;
   static const uptr kTraceMemBeg   = 0x200000000000ull;
@@ -326,9 +314,12 @@ struct Mapping46 {
   static const uptr kLoAppMemEnd   = 0x010000000000ull;
   static const uptr kHiAppMemBeg   = 0x3e8000000000ull;
   static const uptr kHiAppMemEnd   = 0x400000000000ull; // 46 bits
-  static const uptr kAppMemMsk     = 0x3c0000000000ull;
-  static const uptr kAppMemXor     = 0x020000000000ull;
+  static const uptr kShadowMsk = 0x3c0000000000ull;
+  static const uptr kShadowXor = 0x020000000000ull;
+  static const uptr kShadowAdd = 0x000000000000ull;
   static const uptr kVdsoBeg       = 0x7800000000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
 };
 
 /*
@@ -345,7 +336,7 @@ C/C++ on linux/powerpc64 (47-bit VMA)
 7e00 0000 0000 - 7e80 0000 0000: -
 7e80 0000 0000 - 8000 0000 0000: modules and main thread stack
 */
-struct Mapping47 {
+struct MappingPPC64_47 {
   static const uptr kMetaShadowBeg = 0x100000000000ull;
   static const uptr kMetaShadowEnd = 0x200000000000ull;
   static const uptr kTraceMemBeg   = 0x200000000000ull;
@@ -358,14 +349,14 @@ struct Mapping47 {
   static const uptr kLoAppMemEnd   = 0x010000000000ull;
   static const uptr kHiAppMemBeg   = 0x7e8000000000ull;
   static const uptr kHiAppMemEnd   = 0x800000000000ull; // 47 bits
-  static const uptr kAppMemMsk     = 0x7c0000000000ull;
-  static const uptr kAppMemXor     = 0x020000000000ull;
+  static const uptr kShadowMsk = 0x7c0000000000ull;
+  static const uptr kShadowXor = 0x020000000000ull;
+  static const uptr kShadowAdd = 0x000000000000ull;
   static const uptr kVdsoBeg       = 0x7800000000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
 };
 
-// Indicates the runtime will define the memory regions at runtime.
-#define TSAN_RUNTIME_VMA 1
-#elif defined(__s390x__)
 /*
 C/C++ on linux/s390x
 While the kernel provides a 64-bit address space, we have to restrict ourselves
@@ -380,7 +371,7 @@ a000 0000 0000 - b000 0000 0000: traces - 16TiB (max history * 128k threads)
 b000 0000 0000 - be00 0000 0000: -
 be00 0000 0000 - c000 0000 0000: heap - 2TiB (max supported by the allocator)
 */
-struct Mapping {
+struct MappingS390x {
   static const uptr kMetaShadowBeg = 0x900000000000ull;
   static const uptr kMetaShadowEnd = 0x980000000000ull;
   static const uptr kTraceMemBeg   = 0xa00000000000ull;
@@ -393,13 +384,13 @@ struct Mapping {
   static const uptr kLoAppMemEnd   = 0x0e0000000000ull;
   static const uptr kHiAppMemBeg   = 0xc00000004000ull;
   static const uptr kHiAppMemEnd   = 0xc00000004000ull;
-  static const uptr kAppMemMsk     = 0xb00000000000ull;
-  static const uptr kAppMemXor     = 0x100000000000ull;
+  static const uptr kShadowMsk = 0xb00000000000ull;
+  static const uptr kShadowXor = 0x100000000000ull;
+  static const uptr kShadowAdd = 0x000000000000ull;
   static const uptr kVdsoBeg       = 0xfffffffff000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
 };
-#endif
-
-#elif SANITIZER_GO && !SANITIZER_WINDOWS && HAS_48_BIT_ADDRESS_SPACE
 
 /* Go on linux, darwin and freebsd on x86_64
 0000 0000 1000 - 0000 1000 0000: executable
@@ -414,46 +405,59 @@ struct Mapping {
 6200 0000 0000 - 8000 0000 0000: -
 */
 
-struct Mapping {
+struct MappingGo48 {
   static const uptr kMetaShadowBeg = 0x300000000000ull;
   static const uptr kMetaShadowEnd = 0x400000000000ull;
   static const uptr kTraceMemBeg   = 0x600000000000ull;
   static const uptr kTraceMemEnd   = 0x620000000000ull;
   static const uptr kShadowBeg     = 0x200000000000ull;
   static const uptr kShadowEnd     = 0x238000000000ull;
-  static const uptr kAppMemBeg     = 0x000000001000ull;
-  static const uptr kAppMemEnd     = 0x00e000000000ull;
+  static const uptr kLoAppMemBeg = 0x000000001000ull;
+  static const uptr kLoAppMemEnd = 0x00e000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
+  static const uptr kHiAppMemBeg = 0;
+  static const uptr kHiAppMemEnd = 0;
+  static const uptr kHeapMemBeg = 0;
+  static const uptr kHeapMemEnd = 0;
+  static const uptr kVdsoBeg = 0;
+  static const uptr kShadowMsk = 0;
+  static const uptr kShadowXor = 0;
+  static const uptr kShadowAdd = 0x200000000000ull;
 };
 
-#elif SANITIZER_GO && SANITIZER_WINDOWS
-
 /* Go on windows
 0000 0000 1000 - 0000 1000 0000: executable
 0000 1000 0000 - 00f8 0000 0000: -
 00c0 0000 0000 - 00e0 0000 0000: heap
 00e0 0000 0000 - 0100 0000 0000: -
 0100 0000 0000 - 0500 0000 0000: shadow
-0500 0000 0000 - 0560 0000 0000: -
-0560 0000 0000 - 0760 0000 0000: traces
-0760 0000 0000 - 07d0 0000 0000: metainfo (memory blocks and sync objects)
+0500 0000 0000 - 0700 0000 0000: traces
+0700 0000 0000 - 0770 0000 0000: metainfo (memory blocks and sync objects)
 07d0 0000 0000 - 8000 0000 0000: -
 */
 
-struct Mapping {
-  static const uptr kMetaShadowBeg = 0x076000000000ull;
-  static const uptr kMetaShadowEnd = 0x07d000000000ull;
-  static const uptr kTraceMemBeg   = 0x056000000000ull;
-  static const uptr kTraceMemEnd   = 0x076000000000ull;
+struct MappingGoWindows {
+  static const uptr kMetaShadowBeg = 0x070000000000ull;
+  static const uptr kMetaShadowEnd = 0x077000000000ull;
+  static const uptr kTraceMemBeg = 0x050000000000ull;
+  static const uptr kTraceMemEnd = 0x070000000000ull;
   static const uptr kShadowBeg     = 0x010000000000ull;
   static const uptr kShadowEnd     = 0x050000000000ull;
-  static const uptr kAppMemBeg     = 0x000000001000ull;
-  static const uptr kAppMemEnd     = 0x00e000000000ull;
+  static const uptr kLoAppMemBeg = 0x000000001000ull;
+  static const uptr kLoAppMemEnd = 0x00e000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
+  static const uptr kHiAppMemBeg = 0;
+  static const uptr kHiAppMemEnd = 0;
+  static const uptr kHeapMemBeg = 0;
+  static const uptr kHeapMemEnd = 0;
+  static const uptr kVdsoBeg = 0;
+  static const uptr kShadowMsk = 0;
+  static const uptr kShadowXor = 0;
+  static const uptr kShadowAdd = 0x010000000000ull;
 };
 
-#elif SANITIZER_GO && defined(__powerpc64__)
-
-/* Only Mapping46 and Mapping47 are currently supported for powercp64 on Go. */
-
 /* Go on linux/powerpc64 (46-bit VMA)
 0000 0000 1000 - 0000 1000 0000: executable
 0000 1000 0000 - 00c0 0000 0000: -
@@ -467,15 +471,25 @@ struct Mapping {
 3800 0000 0000 - 4000 0000 0000: -
 */
 
-struct Mapping46 {
+struct MappingGoPPC64_46 {
   static const uptr kMetaShadowBeg = 0x240000000000ull;
   static const uptr kMetaShadowEnd = 0x340000000000ull;
   static const uptr kTraceMemBeg   = 0x360000000000ull;
   static const uptr kTraceMemEnd   = 0x380000000000ull;
   static const uptr kShadowBeg     = 0x200000000000ull;
   static const uptr kShadowEnd     = 0x238000000000ull;
-  static const uptr kAppMemBeg     = 0x000000001000ull;
-  static const uptr kAppMemEnd     = 0x00e000000000ull;
+  static const uptr kLoAppMemBeg = 0x000000001000ull;
+  static const uptr kLoAppMemEnd = 0x00e000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
+  static const uptr kHiAppMemBeg = 0;
+  static const uptr kHiAppMemEnd = 0;
+  static const uptr kHeapMemBeg = 0;
+  static const uptr kHeapMemEnd = 0;
+  static const uptr kVdsoBeg = 0;
+  static const uptr kShadowMsk = 0;
+  static const uptr kShadowXor = 0;
+  static const uptr kShadowAdd = 0x200000000000ull;
 };
 
 /* Go on linux/powerpc64 (47-bit VMA)
@@ -491,21 +505,27 @@ struct Mapping46 {
 6200 0000 0000 - 8000 0000 0000: -
 */
 
-struct Mapping47 {
+struct MappingGoPPC64_47 {
   static const uptr kMetaShadowBeg = 0x300000000000ull;
   static const uptr kMetaShadowEnd = 0x400000000000ull;
   static const uptr kTraceMemBeg   = 0x600000000000ull;
   static const uptr kTraceMemEnd   = 0x620000000000ull;
   static const uptr kShadowBeg     = 0x200000000000ull;
   static const uptr kShadowEnd     = 0x300000000000ull;
-  static const uptr kAppMemBeg     = 0x000000001000ull;
-  static const uptr kAppMemEnd     = 0x00e000000000ull;
+  static const uptr kLoAppMemBeg = 0x000000001000ull;
+  static const uptr kLoAppMemEnd = 0x00e000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
+  static const uptr kHiAppMemBeg = 0;
+  static const uptr kHiAppMemEnd = 0;
+  static const uptr kHeapMemBeg = 0;
+  static const uptr kHeapMemEnd = 0;
+  static const uptr kVdsoBeg = 0;
+  static const uptr kShadowMsk = 0;
+  static const uptr kShadowXor = 0;
+  static const uptr kShadowAdd = 0x200000000000ull;
 };
 
-#define TSAN_RUNTIME_VMA 1
-
-#elif SANITIZER_GO && defined(__aarch64__)
-
 /* Go on linux/aarch64 (48-bit VMA) and darwin/aarch64 (47-bit VMA)
 0000 0000 1000 - 0000 1000 0000: executable
 0000 1000 0000 - 00c0 0000 0000: -
@@ -518,22 +538,27 @@ struct Mapping47 {
 6000 0000 0000 - 6200 0000 0000: traces
 6200 0000 0000 - 8000 0000 0000: -
 */
-
-struct Mapping {
+struct MappingGoAarch64 {
   static const uptr kMetaShadowBeg = 0x300000000000ull;
   static const uptr kMetaShadowEnd = 0x400000000000ull;
   static const uptr kTraceMemBeg   = 0x600000000000ull;
   static const uptr kTraceMemEnd   = 0x620000000000ull;
   static const uptr kShadowBeg     = 0x200000000000ull;
   static const uptr kShadowEnd     = 0x300000000000ull;
-  static const uptr kAppMemBeg     = 0x000000001000ull;
-  static const uptr kAppMemEnd     = 0x00e000000000ull;
+  static const uptr kLoAppMemBeg = 0x000000001000ull;
+  static const uptr kLoAppMemEnd = 0x00e000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
+  static const uptr kHiAppMemBeg = 0;
+  static const uptr kHiAppMemEnd = 0;
+  static const uptr kHeapMemBeg = 0;
+  static const uptr kHeapMemEnd = 0;
+  static const uptr kVdsoBeg = 0;
+  static const uptr kShadowMsk = 0;
+  static const uptr kShadowXor = 0;
+  static const uptr kShadowAdd = 0x200000000000ull;
 };
 
-// Indicates the runtime will define the memory regions at runtime.
-#define TSAN_RUNTIME_VMA 1
-
-#elif SANITIZER_GO && defined(__mips64)
 /*
 Go on linux/mips64 (47-bit VMA)
 0000 0000 1000 - 0000 1000 0000: executable
@@ -547,20 +572,27 @@ Go on linux/mips64 (47-bit VMA)
 6000 0000 0000 - 6200 0000 0000: traces
 6200 0000 0000 - 8000 0000 0000: -
 */
-struct Mapping47 {
+struct MappingGoMips64_47 {
   static const uptr kMetaShadowBeg = 0x300000000000ull;
   static const uptr kMetaShadowEnd = 0x400000000000ull;
   static const uptr kTraceMemBeg = 0x600000000000ull;
   static const uptr kTraceMemEnd = 0x620000000000ull;
   static const uptr kShadowBeg = 0x200000000000ull;
   static const uptr kShadowEnd = 0x300000000000ull;
-  static const uptr kAppMemBeg = 0x000000001000ull;
-  static const uptr kAppMemEnd = 0x00e000000000ull;
+  static const uptr kLoAppMemBeg = 0x000000001000ull;
+  static const uptr kLoAppMemEnd = 0x00e000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
+  static const uptr kHiAppMemBeg = 0;
+  static const uptr kHiAppMemEnd = 0;
+  static const uptr kHeapMemBeg = 0;
+  static const uptr kHeapMemEnd = 0;
+  static const uptr kVdsoBeg = 0;
+  static const uptr kShadowMsk = 0;
+  static const uptr kShadowXor = 0;
+  static const uptr kShadowAdd = 0x200000000000ull;
 };
 
-#define TSAN_RUNTIME_VMA 1
-
-#elif SANITIZER_GO && defined(__s390x__)
 /*
 Go on linux/s390x
 0000 0000 1000 - 1000 0000 0000: executable and heap - 16 TiB
@@ -571,622 +603,367 @@ Go on linux/s390x
 9800 0000 0000 - a000 0000 0000: -
 a000 0000 0000 - b000 0000 0000: traces - 16TiB (max history * 128k threads)
 */
-struct Mapping {
+struct MappingGoS390x {
   static const uptr kMetaShadowBeg = 0x900000000000ull;
   static const uptr kMetaShadowEnd = 0x980000000000ull;
   static const uptr kTraceMemBeg   = 0xa00000000000ull;
   static const uptr kTraceMemEnd   = 0xb00000000000ull;
   static const uptr kShadowBeg     = 0x400000000000ull;
   static const uptr kShadowEnd     = 0x800000000000ull;
-  static const uptr kAppMemBeg     = 0x000000001000ull;
-  static const uptr kAppMemEnd     = 0x100000000000ull;
+  static const uptr kLoAppMemBeg = 0x000000001000ull;
+  static const uptr kLoAppMemEnd = 0x100000000000ull;
+  static const uptr kMidAppMemBeg = 0;
+  static const uptr kMidAppMemEnd = 0;
+  static const uptr kHiAppMemBeg = 0;
+  static const uptr kHiAppMemEnd = 0;
+  static const uptr kHeapMemBeg = 0;
+  static const uptr kHeapMemEnd = 0;
+  static const uptr kVdsoBeg = 0;
+  static const uptr kShadowMsk = 0;
+  static const uptr kShadowXor = 0;
+  static const uptr kShadowAdd = 0x400000000000ull;
 };
 
-#else
-# error "Unknown platform"
-#endif
-
-
-#ifdef TSAN_RUNTIME_VMA
 extern uptr vmaSize;
-#endif
-
 
-enum MappingType {
-  MAPPING_LO_APP_BEG,
-  MAPPING_LO_APP_END,
-  MAPPING_HI_APP_BEG,
-  MAPPING_HI_APP_END,
-#ifdef TSAN_MID_APP_RANGE
-  MAPPING_MID_APP_BEG,
-  MAPPING_MID_APP_END,
-#endif
-  MAPPING_HEAP_BEG,
-  MAPPING_HEAP_END,
-  MAPPING_APP_BEG,
-  MAPPING_APP_END,
-  MAPPING_SHADOW_BEG,
-  MAPPING_SHADOW_END,
-  MAPPING_META_SHADOW_BEG,
-  MAPPING_META_SHADOW_END,
-  MAPPING_TRACE_BEG,
-  MAPPING_TRACE_END,
-  MAPPING_VDSO_BEG,
-};
-
-template<typename Mapping, int Type>
-uptr MappingImpl(void) {
-  switch (Type) {
-#if !SANITIZER_GO
-    case MAPPING_LO_APP_BEG: return Mapping::kLoAppMemBeg;
-    case MAPPING_LO_APP_END: return Mapping::kLoAppMemEnd;
-# ifdef TSAN_MID_APP_RANGE
-    case MAPPING_MID_APP_BEG: return Mapping::kMidAppMemBeg;
-    case MAPPING_MID_APP_END: return Mapping::kMidAppMemEnd;
-# endif
-    case MAPPING_HI_APP_BEG: return Mapping::kHiAppMemBeg;
-    case MAPPING_HI_APP_END: return Mapping::kHiAppMemEnd;
-    case MAPPING_HEAP_BEG: return Mapping::kHeapMemBeg;
-    case MAPPING_HEAP_END: return Mapping::kHeapMemEnd;
-    case MAPPING_VDSO_BEG: return Mapping::kVdsoBeg;
-#else
-    case MAPPING_APP_BEG: return Mapping::kAppMemBeg;
-    case MAPPING_APP_END: return Mapping::kAppMemEnd;
-#endif
-    case MAPPING_SHADOW_BEG: return Mapping::kShadowBeg;
-    case MAPPING_SHADOW_END: return Mapping::kShadowEnd;
-    case MAPPING_META_SHADOW_BEG: return Mapping::kMetaShadowBeg;
-    case MAPPING_META_SHADOW_END: return Mapping::kMetaShadowEnd;
-    case MAPPING_TRACE_BEG: return Mapping::kTraceMemBeg;
-    case MAPPING_TRACE_END: return Mapping::kTraceMemEnd;
-  }
-}
-
-template<int Type>
-uptr MappingArchImpl(void) {
-#if defined(__aarch64__) && !defined(__APPLE__) && !SANITIZER_GO
+template <typename Func, typename Arg>
+ALWAYS_INLINE auto SelectMapping(Arg arg) {
+#if SANITIZER_GO
+#  if defined(__powerpc64__)
   switch (vmaSize) {
-    case 39: return MappingImpl<Mapping39, Type>();
-    case 42: return MappingImpl<Mapping42, Type>();
-    case 48: return MappingImpl<Mapping48, Type>();
+    case 46:
+      return Func::template Apply<MappingGoPPC64_46>(arg);
+    case 47:
+      return Func::template Apply<MappingGoPPC64_47>(arg);
   }
-  DCHECK(0);
-  return 0;
-#elif defined(__powerpc64__)
+#  elif defined(__mips64)
+  return Func::template Apply<MappingGoMips64_47>(arg);
+#  elif defined(__s390x__)
+  return Func::template Apply<MappingGoS390x>(arg);
+#  elif defined(__aarch64__)
+  return Func::template Apply<MappingGoAarch64>(arg);
+#  elif SANITIZER_WINDOWS
+  return Func::template Apply<MappingGoWindows>(arg);
+#  else
+  return Func::template Apply<MappingGo48>(arg);
+#  endif
+#else  // SANITIZER_GO
+#  if defined(__x86_64__) || SANITIZER_IOSSIM || SANITIZER_MAC && !SANITIZER_IOS
+  return Func::template Apply<Mapping48AddressSpace>(arg);
+#  elif defined(__aarch64__) && defined(__APPLE__)
+  return Func::template Apply<MappingAppleAarch64>(arg);
+#  elif defined(__aarch64__) && !defined(__APPLE__)
   switch (vmaSize) {
-#if !SANITIZER_GO
-    case 44: return MappingImpl<Mapping44, Type>();
-#endif
-    case 46: return MappingImpl<Mapping46, Type>();
-    case 47: return MappingImpl<Mapping47, Type>();
+    case 39:
+      return Func::template Apply<MappingAarch64_39>(arg);
+    case 42:
+      return Func::template Apply<MappingAarch64_42>(arg);
+    case 48:
+      return Func::template Apply<MappingAarch64_48>(arg);
   }
-  DCHECK(0);
-  return 0;
-#elif defined(__mips64)
+#  elif defined(__powerpc64__)
   switch (vmaSize) {
-#if !SANITIZER_GO
-    case 40: return MappingImpl<Mapping40, Type>();
-#else
-    case 47: return MappingImpl<Mapping47, Type>();
-#endif
+    case 44:
+      return Func::template Apply<MappingPPC64_44>(arg);
+    case 46:
+      return Func::template Apply<MappingPPC64_46>(arg);
+    case 47:
+      return Func::template Apply<MappingPPC64_47>(arg);
   }
-  DCHECK(0);
-  return 0;
-#else
-  return MappingImpl<Mapping, Type>();
+#  elif defined(__mips64)
+  return Func::template Apply<MappingMips64_40>(arg);
+#  elif defined(__s390x__)
+  return Func::template Apply<MappingS390x>(arg);
+#  else
+#    error "unsupported platform"
+#  endif
 #endif
+  Die();
 }
 
-#if !SANITIZER_GO
-ALWAYS_INLINE
-uptr LoAppMemBeg(void) {
-  return MappingArchImpl<MAPPING_LO_APP_BEG>();
-}
-ALWAYS_INLINE
-uptr LoAppMemEnd(void) {
-  return MappingArchImpl<MAPPING_LO_APP_END>();
+template <typename Func>
+void ForEachMapping() {
+  Func::template Apply<Mapping48AddressSpace>();
+  Func::template Apply<MappingMips64_40>();
+  Func::template Apply<MappingAppleAarch64>();
+  Func::template Apply<MappingAarch64_39>();
+  Func::template Apply<MappingAarch64_42>();
+  Func::template Apply<MappingAarch64_48>();
+  Func::template Apply<MappingPPC64_44>();
+  Func::template Apply<MappingPPC64_46>();
+  Func::template Apply<MappingPPC64_47>();
+  Func::template Apply<MappingS390x>();
+  Func::template Apply<MappingGo48>();
+  Func::template Apply<MappingGoWindows>();
+  Func::template Apply<MappingGoPPC64_46>();
+  Func::template Apply<MappingGoPPC64_47>();
+  Func::template Apply<MappingGoAarch64>();
+  Func::template Apply<MappingGoMips64_47>();
+  Func::template Apply<MappingGoS390x>();
 }
 
-#ifdef TSAN_MID_APP_RANGE
-ALWAYS_INLINE
-uptr MidAppMemBeg(void) {
-  return MappingArchImpl<MAPPING_MID_APP_BEG>();
-}
-ALWAYS_INLINE
-uptr MidAppMemEnd(void) {
-  return MappingArchImpl<MAPPING_MID_APP_END>();
-}
-#endif
+enum MappingType {
+  kLoAppMemBeg,
+  kLoAppMemEnd,
+  kHiAppMemBeg,
+  kHiAppMemEnd,
+  kMidAppMemBeg,
+  kMidAppMemEnd,
+  kHeapMemBeg,
+  kHeapMemEnd,
+  kShadowBeg,
+  kShadowEnd,
+  kMetaShadowBeg,
+  kMetaShadowEnd,
+  kTraceMemBeg,
+  kTraceMemEnd,
+  kVdsoBeg,
+};
+
+struct MappingField {
+  template <typename Mapping>
+  static uptr Apply(MappingType type) {
+    switch (type) {
+      case kLoAppMemBeg:
+        return Mapping::kLoAppMemBeg;
+      case kLoAppMemEnd:
+        return Mapping::kLoAppMemEnd;
+      case kMidAppMemBeg:
+        return Mapping::kMidAppMemBeg;
+      case kMidAppMemEnd:
+        return Mapping::kMidAppMemEnd;
+      case kHiAppMemBeg:
+        return Mapping::kHiAppMemBeg;
+      case kHiAppMemEnd:
+        return Mapping::kHiAppMemEnd;
+      case kHeapMemBeg:
+        return Mapping::kHeapMemBeg;
+      case kHeapMemEnd:
+        return Mapping::kHeapMemEnd;
+      case kVdsoBeg:
+        return Mapping::kVdsoBeg;
+      case kShadowBeg:
+        return Mapping::kShadowBeg;
+      case kShadowEnd:
+        return Mapping::kShadowEnd;
+      case kMetaShadowBeg:
+        return Mapping::kMetaShadowBeg;
+      case kMetaShadowEnd:
+        return Mapping::kMetaShadowEnd;
+      case kTraceMemBeg:
+        return Mapping::kTraceMemBeg;
+      case kTraceMemEnd:
+        return Mapping::kTraceMemEnd;
+    }
+    Die();
+  }
+};
 
 ALWAYS_INLINE
-uptr HeapMemBeg(void) {
-  return MappingArchImpl<MAPPING_HEAP_BEG>();
-}
+uptr LoAppMemBeg(void) { return SelectMapping<MappingField>(kLoAppMemBeg); }
 ALWAYS_INLINE
-uptr HeapMemEnd(void) {
-  return MappingArchImpl<MAPPING_HEAP_END>();
-}
+uptr LoAppMemEnd(void) { return SelectMapping<MappingField>(kLoAppMemEnd); }
 
 ALWAYS_INLINE
-uptr HiAppMemBeg(void) {
-  return MappingArchImpl<MAPPING_HI_APP_BEG>();
-}
+uptr MidAppMemBeg(void) { return SelectMapping<MappingField>(kMidAppMemBeg); }
 ALWAYS_INLINE
-uptr HiAppMemEnd(void) {
-  return MappingArchImpl<MAPPING_HI_APP_END>();
-}
+uptr MidAppMemEnd(void) { return SelectMapping<MappingField>(kMidAppMemEnd); }
 
 ALWAYS_INLINE
-uptr VdsoBeg(void) {
-  return MappingArchImpl<MAPPING_VDSO_BEG>();
-}
-
-#else
+uptr HeapMemBeg(void) { return SelectMapping<MappingField>(kHeapMemBeg); }
+ALWAYS_INLINE
+uptr HeapMemEnd(void) { return SelectMapping<MappingField>(kHeapMemEnd); }
 
 ALWAYS_INLINE
-uptr AppMemBeg(void) {
-  return MappingArchImpl<MAPPING_APP_BEG>();
-}
+uptr HiAppMemBeg(void) { return SelectMapping<MappingField>(kHiAppMemBeg); }
 ALWAYS_INLINE
-uptr AppMemEnd(void) {
-  return MappingArchImpl<MAPPING_APP_END>();
-}
-
-#endif
+uptr HiAppMemEnd(void) { return SelectMapping<MappingField>(kHiAppMemEnd); }
 
-static inline
-bool GetUserRegion(int i, uptr *start, uptr *end) {
-  switch (i) {
-  default:
-    return false;
-#if !SANITIZER_GO
-  case 0:
-    *start = LoAppMemBeg();
-    *end = LoAppMemEnd();
-    return true;
-  case 1:
-    *start = HiAppMemBeg();
-    *end = HiAppMemEnd();
-    return true;
-  case 2:
-    *start = HeapMemBeg();
-    *end = HeapMemEnd();
-    return true;
-# ifdef TSAN_MID_APP_RANGE
-  case 3:
-    *start = MidAppMemBeg();
-    *end = MidAppMemEnd();
-    return true;
-# endif
-#else
-  case 0:
-    *start = AppMemBeg();
-    *end = AppMemEnd();
-    return true;
-#endif
-  }
-}
+ALWAYS_INLINE
+uptr VdsoBeg(void) { return SelectMapping<MappingField>(kVdsoBeg); }
 
 ALWAYS_INLINE
-uptr ShadowBeg(void) {
-  return MappingArchImpl<MAPPING_SHADOW_BEG>();
-}
+uptr ShadowBeg(void) { return SelectMapping<MappingField>(kShadowBeg); }
 ALWAYS_INLINE
-uptr ShadowEnd(void) {
-  return MappingArchImpl<MAPPING_SHADOW_END>();
-}
+uptr ShadowEnd(void) { return SelectMapping<MappingField>(kShadowEnd); }
 
 ALWAYS_INLINE
-uptr MetaShadowBeg(void) {
-  return MappingArchImpl<MAPPING_META_SHADOW_BEG>();
-}
+uptr MetaShadowBeg(void) { return SelectMapping<MappingField>(kMetaShadowBeg); }
 ALWAYS_INLINE
-uptr MetaShadowEnd(void) {
-  return MappingArchImpl<MAPPING_META_SHADOW_END>();
-}
+uptr MetaShadowEnd(void) { return SelectMapping<MappingField>(kMetaShadowEnd); }
 
 ALWAYS_INLINE
-uptr TraceMemBeg(void) {
-  return MappingArchImpl<MAPPING_TRACE_BEG>();
-}
+uptr TraceMemBeg(void) { return SelectMapping<MappingField>(kTraceMemBeg); }
 ALWAYS_INLINE
-uptr TraceMemEnd(void) {
-  return MappingArchImpl<MAPPING_TRACE_END>();
-}
-
+uptr TraceMemEnd(void) { return SelectMapping<MappingField>(kTraceMemEnd); }
 
-template<typename Mapping>
-bool IsAppMemImpl(uptr mem) {
-#if !SANITIZER_GO
+struct IsAppMemImpl {
+  template <typename Mapping>
+  static bool Apply(uptr mem) {
   return (mem >= Mapping::kHeapMemBeg && mem < Mapping::kHeapMemEnd) ||
-# ifdef TSAN_MID_APP_RANGE
          (mem >= Mapping::kMidAppMemBeg && mem < Mapping::kMidAppMemEnd) ||
-# endif
          (mem >= Mapping::kLoAppMemBeg && mem < Mapping::kLoAppMemEnd) ||
          (mem >= Mapping::kHiAppMemBeg && mem < Mapping::kHiAppMemEnd);
-#else
-  return mem >= Mapping::kAppMemBeg && mem < Mapping::kAppMemEnd;
-#endif
-}
-
-ALWAYS_INLINE
-bool IsAppMem(uptr mem) {
-#if defined(__aarch64__) && !defined(__APPLE__) && !SANITIZER_GO
-  switch (vmaSize) {
-    case 39: return IsAppMemImpl<Mapping39>(mem);
-    case 42: return IsAppMemImpl<Mapping42>(mem);
-    case 48: return IsAppMemImpl<Mapping48>(mem);
-  }
-  DCHECK(0);
-  return false;
-#elif defined(__powerpc64__)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 44: return IsAppMemImpl<Mapping44>(mem);
-#endif
-    case 46: return IsAppMemImpl<Mapping46>(mem);
-    case 47: return IsAppMemImpl<Mapping47>(mem);
   }
-  DCHECK(0);
-  return false;
-#elif defined(__mips64)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 40: return IsAppMemImpl<Mapping40>(mem);
-#else
-    case 47: return IsAppMemImpl<Mapping47>(mem);
-#endif
-  }
-  DCHECK(0);
-  return false;
-#else
-  return IsAppMemImpl<Mapping>(mem);
-#endif
-}
+};
 
+ALWAYS_INLINE
+bool IsAppMem(uptr mem) { return SelectMapping<IsAppMemImpl>(mem); }
 
-template<typename Mapping>
-bool IsShadowMemImpl(uptr mem) {
-  return mem >= Mapping::kShadowBeg && mem <= Mapping::kShadowEnd;
-}
+struct IsShadowMemImpl {
+  template <typename Mapping>
+  static bool Apply(uptr mem) {
+    return mem >= Mapping::kShadowBeg && mem <= Mapping::kShadowEnd;
+  }
+};
 
 ALWAYS_INLINE
-bool IsShadowMem(uptr mem) {
-#if defined(__aarch64__) && !defined(__APPLE__) && !SANITIZER_GO
-  switch (vmaSize) {
-    case 39: return IsShadowMemImpl<Mapping39>(mem);
-    case 42: return IsShadowMemImpl<Mapping42>(mem);
-    case 48: return IsShadowMemImpl<Mapping48>(mem);
-  }
-  DCHECK(0);
-  return false;
-#elif defined(__powerpc64__)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 44: return IsShadowMemImpl<Mapping44>(mem);
-#endif
-    case 46: return IsShadowMemImpl<Mapping46>(mem);
-    case 47: return IsShadowMemImpl<Mapping47>(mem);
-  }
-  DCHECK(0);
-  return false;
-#elif defined(__mips64)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 40: return IsShadowMemImpl<Mapping40>(mem);
-#else
-    case 47: return IsShadowMemImpl<Mapping47>(mem);
-#endif
-  }
-  DCHECK(0);
-  return false;
-#else
-  return IsShadowMemImpl<Mapping>(mem);
-#endif
+bool IsShadowMem(RawShadow *p) {
+  return SelectMapping<IsShadowMemImpl>(reinterpret_cast<uptr>(p));
 }
 
-
-template<typename Mapping>
-bool IsMetaMemImpl(uptr mem) {
-  return mem >= Mapping::kMetaShadowBeg && mem <= Mapping::kMetaShadowEnd;
-}
+struct IsMetaMemImpl {
+  template <typename Mapping>
+  static bool Apply(uptr mem) {
+    return mem >= Mapping::kMetaShadowBeg && mem <= Mapping::kMetaShadowEnd;
+  }
+};
 
 ALWAYS_INLINE
-bool IsMetaMem(uptr mem) {
-#if defined(__aarch64__) && !defined(__APPLE__) && !SANITIZER_GO
-  switch (vmaSize) {
-    case 39: return IsMetaMemImpl<Mapping39>(mem);
-    case 42: return IsMetaMemImpl<Mapping42>(mem);
-    case 48: return IsMetaMemImpl<Mapping48>(mem);
-  }
-  DCHECK(0);
-  return false;
-#elif defined(__powerpc64__)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 44: return IsMetaMemImpl<Mapping44>(mem);
-#endif
-    case 46: return IsMetaMemImpl<Mapping46>(mem);
-    case 47: return IsMetaMemImpl<Mapping47>(mem);
-  }
-  DCHECK(0);
-  return false;
-#elif defined(__mips64)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 40: return IsMetaMemImpl<Mapping40>(mem);
-#else
-    case 47: return IsMetaMemImpl<Mapping47>(mem);
-#endif
-  }
-  DCHECK(0);
-  return false;
-#else
-  return IsMetaMemImpl<Mapping>(mem);
-#endif
+bool IsMetaMem(const u32 *p) {
+  return SelectMapping<IsMetaMemImpl>(reinterpret_cast<uptr>(p));
 }
 
+struct MemToShadowImpl {
+  template <typename Mapping>
+  static uptr Apply(uptr x) {
+    DCHECK(IsAppMemImpl::Apply<Mapping>(x));
+    return (((x) & ~(Mapping::kShadowMsk | (kShadowCell - 1))) ^
+            Mapping::kShadowXor) *
+               kShadowMultiplier +
+           Mapping::kShadowAdd;
+  }
+};
 
-template<typename Mapping>
-uptr MemToShadowImpl(uptr x) {
-  DCHECK(IsAppMem(x));
-#if !SANITIZER_GO
-  return (((x) & ~(Mapping::kAppMemMsk | (kShadowCell - 1)))
-      ^ Mapping::kAppMemXor) * kShadowCnt;
-#else
-# ifndef SANITIZER_WINDOWS
-  return ((x & ~(kShadowCell - 1)) * kShadowCnt) | Mapping::kShadowBeg;
-# else
-  return ((x & ~(kShadowCell - 1)) * kShadowCnt) + Mapping::kShadowBeg;
-# endif
-#endif
+ALWAYS_INLINE
+RawShadow *MemToShadow(uptr x) {
+  return reinterpret_cast<RawShadow *>(SelectMapping<MemToShadowImpl>(x));
 }
 
-ALWAYS_INLINE
-uptr MemToShadow(uptr x) {
-#if defined(__aarch64__) && !defined(__APPLE__) && !SANITIZER_GO
-  switch (vmaSize) {
-    case 39: return MemToShadowImpl<Mapping39>(x);
-    case 42: return MemToShadowImpl<Mapping42>(x);
-    case 48: return MemToShadowImpl<Mapping48>(x);
+struct MemToMetaImpl {
+  template <typename Mapping>
+  static u32 *Apply(uptr x) {
+    DCHECK(IsAppMemImpl::Apply<Mapping>(x));
+    return (u32 *)(((((x) & ~(Mapping::kShadowMsk | (kMetaShadowCell - 1)))) /
+                    kMetaShadowCell * kMetaShadowSize) |
+                   Mapping::kMetaShadowBeg);
   }
-  DCHECK(0);
-  return 0;
-#elif defined(__powerpc64__)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 44: return MemToShadowImpl<Mapping44>(x);
-#endif
-    case 46: return MemToShadowImpl<Mapping46>(x);
-    case 47: return MemToShadowImpl<Mapping47>(x);
-  }
-  DCHECK(0);
-  return 0;
-#elif defined(__mips64)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 40: return MemToShadowImpl<Mapping40>(x);
-#else
-    case 47: return MemToShadowImpl<Mapping47>(x);
-#endif
-  }
-  DCHECK(0);
-  return 0;
-#else
-  return MemToShadowImpl<Mapping>(x);
-#endif
-}
+};
 
+ALWAYS_INLINE
+u32 *MemToMeta(uptr x) { return SelectMapping<MemToMetaImpl>(x); }
 
-template<typename Mapping>
-u32 *MemToMetaImpl(uptr x) {
-  DCHECK(IsAppMem(x));
-#if !SANITIZER_GO
-  return (u32*)(((((x) & ~(Mapping::kAppMemMsk | (kMetaShadowCell - 1)))) /
-      kMetaShadowCell * kMetaShadowSize) | Mapping::kMetaShadowBeg);
-#else
-# ifndef SANITIZER_WINDOWS
-  return (u32*)(((x & ~(kMetaShadowCell - 1)) / \
-      kMetaShadowCell * kMetaShadowSize) | Mapping::kMetaShadowBeg);
-# else
-  return (u32*)(((x & ~(kMetaShadowCell - 1)) / \
-      kMetaShadowCell * kMetaShadowSize) + Mapping::kMetaShadowBeg);
-# endif
-#endif
-}
+struct ShadowToMemImpl {
+  template <typename Mapping>
+  static uptr Apply(uptr sp) {
+    if (!IsShadowMemImpl::Apply<Mapping>(sp))
+      return 0;
+    // The shadow mapping is non-linear and we've lost some bits, so we don't
+    // have an easy way to restore the original app address. But the mapping is
+    // a bijection, so we try to restore the address as belonging to
+    // low/mid/high range consecutively and see if shadow->app->shadow mapping
+    // gives us the same address.
+    uptr p =
+        ((sp - Mapping::kShadowAdd) / kShadowMultiplier) ^ Mapping::kShadowXor;
+    if (p >= Mapping::kLoAppMemBeg && p < Mapping::kLoAppMemEnd &&
+        MemToShadowImpl::Apply<Mapping>(p) == sp)
+      return p;
+    if (Mapping::kMidAppMemBeg) {
+      uptr p_mid = p + (Mapping::kMidAppMemBeg & Mapping::kShadowMsk);
+      if (p_mid >= Mapping::kMidAppMemBeg && p_mid < Mapping::kMidAppMemEnd &&
+          MemToShadowImpl::Apply<Mapping>(p_mid) == sp)
+        return p_mid;
+    }
+    return p | Mapping::kShadowMsk;
+  }
+};
 
 ALWAYS_INLINE
-u32 *MemToMeta(uptr x) {
-#if defined(__aarch64__) && !defined(__APPLE__) && !SANITIZER_GO
-  switch (vmaSize) {
-    case 39: return MemToMetaImpl<Mapping39>(x);
-    case 42: return MemToMetaImpl<Mapping42>(x);
-    case 48: return MemToMetaImpl<Mapping48>(x);
-  }
-  DCHECK(0);
-  return 0;
-#elif defined(__powerpc64__)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 44: return MemToMetaImpl<Mapping44>(x);
-#endif
-    case 46: return MemToMetaImpl<Mapping46>(x);
-    case 47: return MemToMetaImpl<Mapping47>(x);
-  }
-  DCHECK(0);
-  return 0;
-#elif defined(__mips64)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 40: return MemToMetaImpl<Mapping40>(x);
-#else
-    case 47: return MemToMetaImpl<Mapping47>(x);
-#endif
-  }
-  DCHECK(0);
-  return 0;
-#else
-  return MemToMetaImpl<Mapping>(x);
-#endif
+uptr ShadowToMem(RawShadow *s) {
+  return SelectMapping<ShadowToMemImpl>(reinterpret_cast<uptr>(s));
 }
 
-
-template<typename Mapping>
-uptr ShadowToMemImpl(uptr s) {
-  DCHECK(IsShadowMem(s));
-#if !SANITIZER_GO
-  // The shadow mapping is non-linear and we've lost some bits, so we don't have
-  // an easy way to restore the original app address. But the mapping is a
-  // bijection, so we try to restore the address as belonging to low/mid/high
-  // range consecutively and see if shadow->app->shadow mapping gives us the
-  // same address.
-  uptr p = (s / kShadowCnt) ^ Mapping::kAppMemXor;
-  if (p >= Mapping::kLoAppMemBeg && p < Mapping::kLoAppMemEnd &&
-      MemToShadow(p) == s)
-    return p;
-# ifdef TSAN_MID_APP_RANGE
-  p = ((s / kShadowCnt) ^ Mapping::kAppMemXor) +
-      (Mapping::kMidAppMemBeg & Mapping::kAppMemMsk);
-  if (p >= Mapping::kMidAppMemBeg && p < Mapping::kMidAppMemEnd &&
-      MemToShadow(p) == s)
-    return p;
-# endif
-  return ((s / kShadowCnt) ^ Mapping::kAppMemXor) | Mapping::kAppMemMsk;
-#else  // #if !SANITIZER_GO
-# ifndef SANITIZER_WINDOWS
-  return (s & ~Mapping::kShadowBeg) / kShadowCnt;
-# else
-  return (s - Mapping::kShadowBeg) / kShadowCnt;
-# endif // SANITIZER_WINDOWS
-#endif
+// Compresses addr to kCompressedAddrBits stored in least significant bits.
+ALWAYS_INLINE uptr CompressAddr(uptr addr) {
+  return addr & ((1ull << kCompressedAddrBits) - 1);
 }
 
-ALWAYS_INLINE
-uptr ShadowToMem(uptr s) {
-#if defined(__aarch64__) && !defined(__APPLE__) && !SANITIZER_GO
-  switch (vmaSize) {
-    case 39: return ShadowToMemImpl<Mapping39>(s);
-    case 42: return ShadowToMemImpl<Mapping42>(s);
-    case 48: return ShadowToMemImpl<Mapping48>(s);
-  }
-  DCHECK(0);
-  return 0;
-#elif defined(__powerpc64__)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 44: return ShadowToMemImpl<Mapping44>(s);
-#endif
-    case 46: return ShadowToMemImpl<Mapping46>(s);
-    case 47: return ShadowToMemImpl<Mapping47>(s);
-  }
-  DCHECK(0);
-  return 0;
-#elif defined(__mips64)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 40: return ShadowToMemImpl<Mapping40>(s);
-#else
-    case 47: return ShadowToMemImpl<Mapping47>(s);
-#endif
+struct RestoreAddrImpl {
+  typedef uptr Result;
+  template <typename Mapping>
+  static Result Apply(uptr addr) {
+    // To restore the address we go over all app memory ranges and check if top
+    // 3 bits of the compressed addr match that of the app range. If yes, we
+    // assume that the compressed address come from that range and restore the
+    // missing top bits to match the app range address.
+    const uptr ranges[] = {
+        Mapping::kLoAppMemBeg,  Mapping::kLoAppMemEnd, Mapping::kMidAppMemBeg,
+        Mapping::kMidAppMemEnd, Mapping::kHiAppMemBeg, Mapping::kHiAppMemEnd,
+        Mapping::kHeapMemBeg,   Mapping::kHeapMemEnd,
+    };
+    const uptr indicator = 0x0e0000000000ull;
+    const uptr ind_lsb = 1ull << LeastSignificantSetBitIndex(indicator);
+    for (uptr i = 0; i < ARRAY_SIZE(ranges); i += 2) {
+      uptr beg = ranges[i];
+      uptr end = ranges[i + 1];
+      if (beg == end)
+        continue;
+      for (uptr p = beg; p < end; p = RoundDown(p + ind_lsb, ind_lsb)) {
+        if ((addr & indicator) == (p & indicator))
+          return addr | (p & ~(ind_lsb - 1));
+      }
+    }
+    Printf("ThreadSanitizer: failed to restore address 0x%zx\n", addr);
+    Die();
   }
-  DCHECK(0);
-  return 0;
-#else
-  return ShadowToMemImpl<Mapping>(s);
-#endif
-}
-
+};
 
+// Restores compressed addr from kCompressedAddrBits to full representation.
+// This is called only during reporting and is not performance-critical.
+inline uptr RestoreAddr(uptr addr) {
+  return SelectMapping<RestoreAddrImpl>(addr);
+}
 
 // The additional page is to catch shadow stack overflow as paging fault.
 // Windows wants 64K alignment for mmaps.
 const uptr kTotalTraceSize = (kTraceSize * sizeof(Event) + sizeof(Trace)
     + (64 << 10) + (64 << 10) - 1) & ~((64 << 10) - 1);
 
-template<typename Mapping>
-uptr GetThreadTraceImpl(int tid) {
-  uptr p = Mapping::kTraceMemBeg + (uptr)tid * kTotalTraceSize;
-  DCHECK_LT(p, Mapping::kTraceMemEnd);
-  return p;
-}
-
-ALWAYS_INLINE
-uptr GetThreadTrace(int tid) {
-#if defined(__aarch64__) && !defined(__APPLE__) && !SANITIZER_GO
-  switch (vmaSize) {
-    case 39: return GetThreadTraceImpl<Mapping39>(tid);
-    case 42: return GetThreadTraceImpl<Mapping42>(tid);
-    case 48: return GetThreadTraceImpl<Mapping48>(tid);
-  }
-  DCHECK(0);
-  return 0;
-#elif defined(__powerpc64__)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 44: return GetThreadTraceImpl<Mapping44>(tid);
-#endif
-    case 46: return GetThreadTraceImpl<Mapping46>(tid);
-    case 47: return GetThreadTraceImpl<Mapping47>(tid);
-  }
-  DCHECK(0);
-  return 0;
-#elif defined(__mips64)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 40: return GetThreadTraceImpl<Mapping40>(tid);
-#else
-    case 47: return GetThreadTraceImpl<Mapping47>(tid);
-#endif
+struct GetThreadTraceImpl {
+  template <typename Mapping>
+  static uptr Apply(uptr tid) {
+    uptr p = Mapping::kTraceMemBeg + tid * kTotalTraceSize;
+    DCHECK_LT(p, Mapping::kTraceMemEnd);
+    return p;
   }
-  DCHECK(0);
-  return 0;
-#else
-  return GetThreadTraceImpl<Mapping>(tid);
-#endif
-}
+};
 
+ALWAYS_INLINE
+uptr GetThreadTrace(int tid) { return SelectMapping<GetThreadTraceImpl>(tid); }
 
-template<typename Mapping>
-uptr GetThreadTraceHeaderImpl(int tid) {
-  uptr p = Mapping::kTraceMemBeg + (uptr)tid * kTotalTraceSize
-      + kTraceSize * sizeof(Event);
-  DCHECK_LT(p, Mapping::kTraceMemEnd);
-  return p;
-}
+struct GetThreadTraceHeaderImpl {
+  template <typename Mapping>
+  static uptr Apply(uptr tid) {
+    uptr p = Mapping::kTraceMemBeg + tid * kTotalTraceSize +
+             kTraceSize * sizeof(Event);
+    DCHECK_LT(p, Mapping::kTraceMemEnd);
+    return p;
+  }
+};
 
 ALWAYS_INLINE
 uptr GetThreadTraceHeader(int tid) {
-#if defined(__aarch64__) && !defined(__APPLE__) && !SANITIZER_GO
-  switch (vmaSize) {
-    case 39: return GetThreadTraceHeaderImpl<Mapping39>(tid);
-    case 42: return GetThreadTraceHeaderImpl<Mapping42>(tid);
-    case 48: return GetThreadTraceHeaderImpl<Mapping48>(tid);
-  }
-  DCHECK(0);
-  return 0;
-#elif defined(__powerpc64__)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 44: return GetThreadTraceHeaderImpl<Mapping44>(tid);
-#endif
-    case 46: return GetThreadTraceHeaderImpl<Mapping46>(tid);
-    case 47: return GetThreadTraceHeaderImpl<Mapping47>(tid);
-  }
-  DCHECK(0);
-  return 0;
-#elif defined(__mips64)
-  switch (vmaSize) {
-#if !SANITIZER_GO
-    case 40: return GetThreadTraceHeaderImpl<Mapping40>(tid);
-#else
-    case 47: return GetThreadTraceHeaderImpl<Mapping47>(tid);
-#endif
-  }
-  DCHECK(0);
-  return 0;
-#else
-  return GetThreadTraceHeaderImpl<Mapping>(tid);
-#endif
+  return SelectMapping<GetThreadTraceHeaderImpl>(tid);
 }
 
 void InitializePlatform();
@@ -1194,7 +971,7 @@ void InitializePlatformEarly();
 void CheckAndProtect();
 void InitializeShadowMemoryPlatform();
 void FlushShadowMemory();
-void WriteMemoryProfile(char *buf, uptr buf_size, uptr nthread, uptr nlive);
+void WriteMemoryProfile(char *buf, uptr buf_size, u64 uptime_ns);
 int ExtractResolvFDs(void *state, int *fds, int nfd);
 int ExtractRecvmsgFDs(void *msg, int *fds, int nfd);
 uptr ExtractLongJmpSp(uptr *env);
diff --git a/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cpp b/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cpp
index cfe597e5380e..73ec14892d28 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cpp
@@ -85,63 +85,68 @@ static void InitializeLongjmpXorKey();
 static uptr longjmp_xor_key;
 #endif
 
-#ifdef TSAN_RUNTIME_VMA
 // Runtime detected VMA size.
 uptr vmaSize;
-#endif
 
 enum {
-  MemTotal  = 0,
-  MemShadow = 1,
-  MemMeta   = 2,
-  MemFile   = 3,
-  MemMmap   = 4,
-  MemTrace  = 5,
-  MemHeap   = 6,
-  MemOther  = 7,
-  MemCount  = 8,
+  MemTotal,
+  MemShadow,
+  MemMeta,
+  MemFile,
+  MemMmap,
+  MemTrace,
+  MemHeap,
+  MemOther,
+  MemCount,
 };
 
-void FillProfileCallback(uptr p, uptr rss, bool file,
-                         uptr *mem, uptr stats_size) {
+void FillProfileCallback(uptr p, uptr rss, bool file, uptr *mem) {
   mem[MemTotal] += rss;
   if (p >= ShadowBeg() && p < ShadowEnd())
     mem[MemShadow] += rss;
   else if (p >= MetaShadowBeg() && p < MetaShadowEnd())
     mem[MemMeta] += rss;
-#if !SANITIZER_GO
+  else if ((p >= LoAppMemBeg() && p < LoAppMemEnd()) ||
+           (p >= MidAppMemBeg() && p < MidAppMemEnd()) ||
+           (p >= HiAppMemBeg() && p < HiAppMemEnd()))
+    mem[file ? MemFile : MemMmap] += rss;
   else if (p >= HeapMemBeg() && p < HeapMemEnd())
     mem[MemHeap] += rss;
-  else if (p >= LoAppMemBeg() && p < LoAppMemEnd())
-    mem[file ? MemFile : MemMmap] += rss;
-  else if (p >= HiAppMemBeg() && p < HiAppMemEnd())
-    mem[file ? MemFile : MemMmap] += rss;
-#else
-  else if (p >= AppMemBeg() && p < AppMemEnd())
-    mem[file ? MemFile : MemMmap] += rss;
-#endif
   else if (p >= TraceMemBeg() && p < TraceMemEnd())
     mem[MemTrace] += rss;
   else
     mem[MemOther] += rss;
 }
 
-void WriteMemoryProfile(char *buf, uptr buf_size, uptr nthread, uptr nlive) {
+void WriteMemoryProfile(char *buf, uptr buf_size, u64 uptime_ns) {
   uptr mem[MemCount];
-  internal_memset(mem, 0, sizeof(mem[0]) * MemCount);
-  __sanitizer::GetMemoryProfile(FillProfileCallback, mem, 7);
-  StackDepotStats *stacks = StackDepotGetStats();
-  internal_snprintf(buf, buf_size,
-      "RSS %zd MB: shadow:%zd meta:%zd file:%zd mmap:%zd"
-      " trace:%zd heap:%zd other:%zd stacks=%zd[%zd] nthr=%zd/%zd\n",
-      mem[MemTotal] >> 20, mem[MemShadow] >> 20, mem[MemMeta] >> 20,
-      mem[MemFile] >> 20, mem[MemMmap] >> 20, mem[MemTrace] >> 20,
-      mem[MemHeap] >> 20, mem[MemOther] >> 20,
-      stacks->allocated >> 20, stacks->n_uniq_ids,
-      nlive, nthread);
+  internal_memset(mem, 0, sizeof(mem));
+  GetMemoryProfile(FillProfileCallback, mem);
+  auto meta = ctx->metamap.GetMemoryStats();
+  StackDepotStats stacks = StackDepotGetStats();
+  uptr nthread, nlive;
+  ctx->thread_registry.GetNumberOfThreads(&nthread, &nlive);
+  uptr internal_stats[AllocatorStatCount];
+  internal_allocator()->GetStats(internal_stats);
+  // All these are allocated from the common mmap region.
+  mem[MemMmap] -= meta.mem_block + meta.sync_obj + stacks.allocated +
+                  internal_stats[AllocatorStatMapped];
+  if (s64(mem[MemMmap]) < 0)
+    mem[MemMmap] = 0;
+  internal_snprintf(
+      buf, buf_size,
+      "%llus: RSS %zd MB: shadow:%zd meta:%zd file:%zd mmap:%zd"
+      " trace:%zd heap:%zd other:%zd intalloc:%zd memblocks:%zd syncobj:%zu"
+      " stacks=%zd[%zd] nthr=%zd/%zd\n",
+      uptime_ns / (1000 * 1000 * 1000), mem[MemTotal] >> 20,
+      mem[MemShadow] >> 20, mem[MemMeta] >> 20, mem[MemFile] >> 20,
+      mem[MemMmap] >> 20, mem[MemTrace] >> 20, mem[MemHeap] >> 20,
+      mem[MemOther] >> 20, internal_stats[AllocatorStatMapped] >> 20,
+      meta.mem_block >> 20, meta.sync_obj >> 20, stacks.allocated >> 20,
+      stacks.n_uniq_ids, nlive, nthread);
 }
 
-#if SANITIZER_LINUX
+#  if SANITIZER_LINUX
 void FlushShadowMemoryCallback(
     const SuspendedThreadsList &suspended_threads_list,
     void *argument) {
@@ -178,12 +183,13 @@ static void MapRodata() {
   internal_unlink(name);  // Unlink it now, so that we can reuse the buffer.
   fd_t fd = openrv;
   // Fill the file with kShadowRodata.
-  const uptr kMarkerSize = 512 * 1024 / sizeof(u64);
-  InternalMmapVector<u64> marker(kMarkerSize);
+  const uptr kMarkerSize = 512 * 1024 / sizeof(RawShadow);
+  InternalMmapVector<RawShadow> marker(kMarkerSize);
   // volatile to prevent insertion of memset
-  for (volatile u64 *p = marker.data(); p < marker.data() + kMarkerSize; p++)
+  for (volatile RawShadow *p = marker.data(); p < marker.data() + kMarkerSize;
+       p++)
     *p = kShadowRodata;
-  internal_write(fd, marker.data(), marker.size() * sizeof(u64));
+  internal_write(fd, marker.data(), marker.size() * sizeof(RawShadow));
   // Map the file into memory.
   uptr page = internal_mmap(0, GetPageSizeCached(), PROT_READ | PROT_WRITE,
                             MAP_PRIVATE | MAP_ANONYMOUS, fd, 0);
@@ -203,9 +209,10 @@ static void MapRodata() {
       char *shadow_start = (char *)MemToShadow(segment.start);
       char *shadow_end = (char *)MemToShadow(segment.end);
       for (char *p = shadow_start; p < shadow_end;
-           p += marker.size() * sizeof(u64)) {
-        internal_mmap(p, Min<uptr>(marker.size() * sizeof(u64), shadow_end - p),
-                      PROT_READ, MAP_PRIVATE | MAP_FIXED, fd, 0);
+           p += marker.size() * sizeof(RawShadow)) {
+        internal_mmap(
+            p, Min<uptr>(marker.size() * sizeof(RawShadow), shadow_end - p),
+            PROT_READ, MAP_PRIVATE | MAP_FIXED, fd, 0);
       }
     }
   }
@@ -219,7 +226,6 @@ void InitializeShadowMemoryPlatform() {
 #endif  // #if !SANITIZER_GO
 
 void InitializePlatformEarly() {
-#ifdef TSAN_RUNTIME_VMA
   vmaSize =
     (MostSignificantSetBitIndex(GET_CURRENT_FRAME()) + 1);
 #if defined(__aarch64__)
@@ -265,7 +271,6 @@ void InitializePlatformEarly() {
   }
 # endif
 #endif
-#endif
 }
 
 void InitializePlatform() {
@@ -341,7 +346,7 @@ int ExtractResolvFDs(void *state, int *fds, int nfd) {
 }
 
 // Extract file descriptors passed via UNIX domain sockets.
-// This is requried to properly handle "open" of these fds.
+// This is required to properly handle "open" of these fds.
 // see 'man recvmsg' and 'man 3 cmsg'.
 int ExtractRecvmsgFDs(void *msgp, int *fds, int nfd) {
   int res = 0;
@@ -447,6 +452,8 @@ static void InitializeLongjmpXorKey() {
 }
 #endif
 
+extern "C" void __tsan_tls_initialization() {}
+
 void ImitateTlsWrite(ThreadState *thr, uptr tls_addr, uptr tls_size) {
   // Check that the thr object is in tls;
   const uptr thr_beg = (uptr)thr;
@@ -456,9 +463,10 @@ void ImitateTlsWrite(ThreadState *thr, uptr tls_addr, uptr tls_size) {
   CHECK_GE(thr_end, tls_addr);
   CHECK_LE(thr_end, tls_addr + tls_size);
   // Since the thr object is huge, skip it.
-  MemoryRangeImitateWrite(thr, /*pc=*/2, tls_addr, thr_beg - tls_addr);
-  MemoryRangeImitateWrite(thr, /*pc=*/2, thr_end,
-                          tls_addr + tls_size - thr_end);
+  const uptr pc = StackTrace::GetNextInstructionPc(
+      reinterpret_cast<uptr>(__tsan_tls_initialization));
+  MemoryRangeImitateWrite(thr, pc, tls_addr, thr_beg - tls_addr);
+  MemoryRangeImitateWrite(thr, pc, thr_end, tls_addr + tls_size - thr_end);
 }
 
 // Note: this function runs with async signals enabled,
diff --git a/compiler-rt/lib/tsan/rtl/tsan_platform_mac.cpp b/compiler-rt/lib/tsan/rtl/tsan_platform_mac.cpp
index d9719a136b21..3faa2d0c6192 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_platform_mac.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_platform_mac.cpp
@@ -139,7 +139,7 @@ static void RegionMemUsage(uptr start, uptr end, uptr *res, uptr *dirty) {
   *dirty = dirty_pages * GetPageSizeCached();
 }
 
-void WriteMemoryProfile(char *buf, uptr buf_size, uptr nthread, uptr nlive) {
+void WriteMemoryProfile(char *buf, uptr buf_size, u64 uptime_ns) {
   uptr shadow_res, shadow_dirty;
   uptr meta_res, meta_dirty;
   uptr trace_res, trace_dirty;
@@ -156,39 +156,41 @@ void WriteMemoryProfile(char *buf, uptr buf_size, uptr nthread, uptr nlive) {
   RegionMemUsage(HeapMemBeg(), HeapMemEnd(), &heap_res, &heap_dirty);
 #else  // !SANITIZER_GO
   uptr app_res, app_dirty;
-  RegionMemUsage(AppMemBeg(), AppMemEnd(), &app_res, &app_dirty);
+  RegionMemUsage(LoAppMemBeg(), LoAppMemEnd(), &app_res, &app_dirty);
 #endif
 
-  StackDepotStats *stacks = StackDepotGetStats();
-  internal_snprintf(buf, buf_size,
-    "shadow   (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
-    "meta     (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
-    "traces   (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
-#if !SANITIZER_GO
-    "low app  (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
-    "high app (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
-    "heap     (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
-#else  // !SANITIZER_GO
-    "app      (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
-#endif
-    "stacks: %zd unique IDs, %zd kB allocated\n"
-    "threads: %zd total, %zd live\n"
-    "------------------------------\n",
-    ShadowBeg(), ShadowEnd(), shadow_res / 1024, shadow_dirty / 1024,
-    MetaShadowBeg(), MetaShadowEnd(), meta_res / 1024, meta_dirty / 1024,
-    TraceMemBeg(), TraceMemEnd(), trace_res / 1024, trace_dirty / 1024,
-#if !SANITIZER_GO
-    LoAppMemBeg(), LoAppMemEnd(), low_res / 1024, low_dirty / 1024,
-    HiAppMemBeg(), HiAppMemEnd(), high_res / 1024, high_dirty / 1024,
-    HeapMemBeg(), HeapMemEnd(), heap_res / 1024, heap_dirty / 1024,
-#else  // !SANITIZER_GO
-    AppMemBeg(), AppMemEnd(), app_res / 1024, app_dirty / 1024,
-#endif
-    stacks->n_uniq_ids, stacks->allocated / 1024,
-    nthread, nlive);
+  StackDepotStats stacks = StackDepotGetStats();
+  uptr nthread, nlive;
+  ctx->thread_registry.GetNumberOfThreads(&nthread, &nlive);
+  internal_snprintf(
+      buf, buf_size,
+      "shadow   (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
+      "meta     (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
+      "traces   (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
+#  if !SANITIZER_GO
+      "low app  (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
+      "high app (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
+      "heap     (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
+#  else  // !SANITIZER_GO
+      "app      (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
+#  endif
+      "stacks: %zd unique IDs, %zd kB allocated\n"
+      "threads: %zd total, %zd live\n"
+      "------------------------------\n",
+      ShadowBeg(), ShadowEnd(), shadow_res / 1024, shadow_dirty / 1024,
+      MetaShadowBeg(), MetaShadowEnd(), meta_res / 1024, meta_dirty / 1024,
+      TraceMemBeg(), TraceMemEnd(), trace_res / 1024, trace_dirty / 1024,
+#  if !SANITIZER_GO
+      LoAppMemBeg(), LoAppMemEnd(), low_res / 1024, low_dirty / 1024,
+      HiAppMemBeg(), HiAppMemEnd(), high_res / 1024, high_dirty / 1024,
+      HeapMemBeg(), HeapMemEnd(), heap_res / 1024, heap_dirty / 1024,
+#  else  // !SANITIZER_GO
+      LoAppMemBeg(), LoAppMemEnd(), app_res / 1024, app_dirty / 1024,
+#  endif
+      stacks.n_uniq_ids, stacks.allocated / 1024, nthread, nlive);
 }
 
-#if !SANITIZER_GO
+#  if !SANITIZER_GO
 void InitializeShadowMemoryPlatform() { }
 
 // On OS X, GCD worker threads are created without a call to pthread_create. We
@@ -215,8 +217,8 @@ static void my_pthread_introspection_hook(unsigned int event, pthread_t thread,
       Processor *proc = ProcCreate();
       ProcWire(proc, thr);
       ThreadState *parent_thread_state = nullptr;  // No parent.
-      int tid = ThreadCreate(parent_thread_state, 0, (uptr)thread, true);
-      CHECK_NE(tid, 0);
+      Tid tid = ThreadCreate(parent_thread_state, 0, (uptr)thread, true);
+      CHECK_NE(tid, kMainTid);
       ThreadStart(thr, tid, GetTid(), ThreadType::Worker);
     }
   } else if (event == PTHREAD_INTROSPECTION_THREAD_TERMINATE) {
@@ -234,11 +236,11 @@ static void my_pthread_introspection_hook(unsigned int event, pthread_t thread,
 #endif
 
 void InitializePlatformEarly() {
-#if !SANITIZER_GO && !HAS_48_BIT_ADDRESS_SPACE
+#  if !SANITIZER_GO && SANITIZER_IOS
   uptr max_vm = GetMaxUserVirtualAddress() + 1;
-  if (max_vm != Mapping::kHiAppMemEnd) {
+  if (max_vm != HiAppMemEnd()) {
     Printf("ThreadSanitizer: unsupported vm address limit %p, expected %p.\n",
-           max_vm, Mapping::kHiAppMemEnd);
+           (void *)max_vm, (void *)HiAppMemEnd());
     Die();
   }
 #endif
@@ -281,13 +283,17 @@ uptr ExtractLongJmpSp(uptr *env) {
 }
 
 #if !SANITIZER_GO
+extern "C" void __tsan_tls_initialization() {}
+
 void ImitateTlsWrite(ThreadState *thr, uptr tls_addr, uptr tls_size) {
   // The pointer to the ThreadState object is stored in the shadow memory
   // of the tls.
   uptr tls_end = tls_addr + tls_size;
   uptr thread_identity = (uptr)pthread_self();
+  const uptr pc = StackTrace::GetNextInstructionPc(
+      reinterpret_cast<uptr>(__tsan_tls_initialization));
   if (thread_identity == main_thread_identity) {
-    MemoryRangeImitateWrite(thr, /*pc=*/2, tls_addr, tls_size);
+    MemoryRangeImitateWrite(thr, pc, tls_addr, tls_size);
   } else {
     uptr thr_state_start = thread_identity;
     uptr thr_state_end = thr_state_start + sizeof(uptr);
@@ -295,10 +301,8 @@ void ImitateTlsWrite(ThreadState *thr, uptr tls_addr, uptr tls_size) {
     CHECK_LE(thr_state_start, tls_addr + tls_size);
     CHECK_GE(thr_state_end, tls_addr);
     CHECK_LE(thr_state_end, tls_addr + tls_size);
-    MemoryRangeImitateWrite(thr, /*pc=*/2, tls_addr,
-                            thr_state_start - tls_addr);
-    MemoryRangeImitateWrite(thr, /*pc=*/2, thr_state_end,
-                            tls_end - thr_state_end);
+    MemoryRangeImitateWrite(thr, pc, tls_addr, thr_state_start - tls_addr);
+    MemoryRangeImitateWrite(thr, pc, thr_state_end, tls_end - thr_state_end);
   }
 }
 #endif
diff --git a/compiler-rt/lib/tsan/rtl/tsan_platform_posix.cpp b/compiler-rt/lib/tsan/rtl/tsan_platform_posix.cpp
index 1c6198cefcd7..763ac444377e 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_platform_posix.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_platform_posix.cpp
@@ -14,12 +14,14 @@
 #include "sanitizer_common/sanitizer_platform.h"
 #if SANITIZER_POSIX
 
-#include "sanitizer_common/sanitizer_common.h"
-#include "sanitizer_common/sanitizer_errno.h"
-#include "sanitizer_common/sanitizer_libc.h"
-#include "sanitizer_common/sanitizer_procmaps.h"
-#include "tsan_platform.h"
-#include "tsan_rtl.h"
+#  include <dlfcn.h>
+
+#  include "sanitizer_common/sanitizer_common.h"
+#  include "sanitizer_common/sanitizer_errno.h"
+#  include "sanitizer_common/sanitizer_libc.h"
+#  include "sanitizer_common/sanitizer_procmaps.h"
+#  include "tsan_platform.h"
+#  include "tsan_rtl.h"
 
 namespace __tsan {
 
@@ -29,6 +31,7 @@ static const char kShadowMemoryMappingHint[] =
     "HINT: if %s is not supported in your environment, you may set "
     "TSAN_OPTIONS=%s=0\n";
 
+#  if !SANITIZER_GO
 static void DontDumpShadow(uptr addr, uptr size) {
   if (common_flags()->use_madv_dontdump)
     if (!DontDumpShadowMemory(addr, size)) {
@@ -39,7 +42,6 @@ static void DontDumpShadow(uptr addr, uptr size) {
     }
 }
 
-#if !SANITIZER_GO
 void InitializeShadowMemory() {
   // Map memory shadow.
   if (!MmapFixedSuperNoReserve(ShadowBeg(), ShadowEnd() - ShadowBeg(),
@@ -70,6 +72,11 @@ void InitializeShadowMemory() {
       meta, meta + meta_size, meta_size >> 30);
 
   InitializeShadowMemoryPlatform();
+
+  on_initialize = reinterpret_cast<void (*)(void)>(
+      dlsym(RTLD_DEFAULT, "__tsan_on_initialize"));
+  on_finalize =
+      reinterpret_cast<int (*)(int)>(dlsym(RTLD_DEFAULT, "__tsan_on_finalize"));
 }
 
 static bool TryProtectRange(uptr beg, uptr end) {
@@ -98,24 +105,24 @@ void CheckAndProtect() {
       continue;
     if (segment.start >= VdsoBeg())  // vdso
       break;
-    Printf("FATAL: ThreadSanitizer: unexpected memory mapping %p-%p\n",
+    Printf("FATAL: ThreadSanitizer: unexpected memory mapping 0x%zx-0x%zx\n",
            segment.start, segment.end);
     Die();
   }
 
-#if defined(__aarch64__) && defined(__APPLE__) && !HAS_48_BIT_ADDRESS_SPACE
+#    if defined(__aarch64__) && defined(__APPLE__) && SANITIZER_IOS
   ProtectRange(HeapMemEnd(), ShadowBeg());
   ProtectRange(ShadowEnd(), MetaShadowBeg());
   ProtectRange(MetaShadowEnd(), TraceMemBeg());
 #else
   ProtectRange(LoAppMemEnd(), ShadowBeg());
   ProtectRange(ShadowEnd(), MetaShadowBeg());
-#ifdef TSAN_MID_APP_RANGE
-  ProtectRange(MetaShadowEnd(), MidAppMemBeg());
-  ProtectRange(MidAppMemEnd(), TraceMemBeg());
-#else
-  ProtectRange(MetaShadowEnd(), TraceMemBeg());
-#endif
+  if (MidAppMemBeg()) {
+    ProtectRange(MetaShadowEnd(), MidAppMemBeg());
+    ProtectRange(MidAppMemEnd(), TraceMemBeg());
+  } else {
+    ProtectRange(MetaShadowEnd(), TraceMemBeg());
+  }
   // Memory for traces is mapped lazily in MapThreadTrace.
   // Protect the whole range for now, so that user does not map something here.
   ProtectRange(TraceMemBeg(), TraceMemEnd());
diff --git a/compiler-rt/lib/tsan/rtl/tsan_platform_windows.cpp b/compiler-rt/lib/tsan/rtl/tsan_platform_windows.cpp
index 19437879a41c..fea893768c79 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_platform_windows.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_platform_windows.cpp
@@ -23,8 +23,7 @@ namespace __tsan {
 void FlushShadowMemory() {
 }
 
-void WriteMemoryProfile(char *buf, uptr buf_size, uptr nthread, uptr nlive) {
-}
+void WriteMemoryProfile(char *buf, uptr buf_size, u64 uptime_ns) {}
 
 void InitializePlatformEarly() {
 }
diff --git a/compiler-rt/lib/tsan/rtl/tsan_report.cpp b/compiler-rt/lib/tsan/rtl/tsan_report.cpp
index 8ef9f0cd4fe8..a926c3761ccf 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_report.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_report.cpp
@@ -19,22 +19,6 @@
 
 namespace __tsan {
 
-ReportStack::ReportStack() : frames(nullptr), suppressable(false) {}
-
-ReportStack *ReportStack::New() {
-  void *mem = internal_alloc(MBlockReportStack, sizeof(ReportStack));
-  return new(mem) ReportStack();
-}
-
-ReportLocation::ReportLocation(ReportLocationType type)
-    : type(type), global(), heap_chunk_start(0), heap_chunk_size(0), tid(0),
-      fd(0), suppressable(false), stack(nullptr) {}
-
-ReportLocation *ReportLocation::New(ReportLocationType type) {
-  void *mem = internal_alloc(MBlockReportStack, sizeof(ReportLocation));
-  return new(mem) ReportLocation(type);
-}
-
 class Decorator: public __sanitizer::SanitizerCommonDecorator {
  public:
   Decorator() : SanitizerCommonDecorator() { }
@@ -68,7 +52,7 @@ ReportDesc::~ReportDesc() {
 #if !SANITIZER_GO
 
 const int kThreadBufSize = 32;
-const char *thread_name(char *buf, int tid) {
+const char *thread_name(char *buf, Tid tid) {
   if (tid == kMainTid)
     return "main thread";
   internal_snprintf(buf, kThreadBufSize, "thread T%d", tid);
@@ -189,23 +173,25 @@ static void PrintLocation(const ReportLocation *loc) {
   if (loc->type == ReportLocationGlobal) {
     const DataInfo &global = loc->global;
     if (global.size != 0)
-      Printf("  Location is global '%s' of size %zu at %p (%s+%p)\n\n",
-             global.name, global.size, global.start,
+      Printf("  Location is global '%s' of size %zu at %p (%s+0x%zx)\n\n",
+             global.name, global.size, reinterpret_cast<void *>(global.start),
              StripModuleName(global.module), global.module_offset);
     else
-      Printf("  Location is global '%s' at %p (%s+%p)\n\n", global.name,
-             global.start, StripModuleName(global.module),
-             global.module_offset);
+      Printf("  Location is global '%s' at %p (%s+0x%zx)\n\n", global.name,
+             reinterpret_cast<void *>(global.start),
+             StripModuleName(global.module), global.module_offset);
   } else if (loc->type == ReportLocationHeap) {
     char thrbuf[kThreadBufSize];
     const char *object_type = GetObjectTypeFromTag(loc->external_tag);
     if (!object_type) {
       Printf("  Location is heap block of size %zu at %p allocated by %s:\n",
-             loc->heap_chunk_size, loc->heap_chunk_start,
+             loc->heap_chunk_size,
+             reinterpret_cast<void *>(loc->heap_chunk_start),
              thread_name(thrbuf, loc->tid));
     } else {
       Printf("  Location is %s of size %zu at %p allocated by %s:\n",
-             object_type, loc->heap_chunk_size, loc->heap_chunk_start,
+             object_type, loc->heap_chunk_size,
+             reinterpret_cast<void *>(loc->heap_chunk_start),
              thread_name(thrbuf, loc->tid));
     }
     print_stack = true;
@@ -225,13 +211,14 @@ static void PrintLocation(const ReportLocation *loc) {
 
 static void PrintMutexShort(const ReportMutex *rm, const char *after) {
   Decorator d;
-  Printf("%sM%zd%s%s", d.Mutex(), rm->id, d.Default(), after);
+  Printf("%sM%lld%s%s", d.Mutex(), rm->id, d.Default(), after);
 }
 
 static void PrintMutexShortWithAddress(const ReportMutex *rm,
                                        const char *after) {
   Decorator d;
-  Printf("%sM%zd (%p)%s%s", d.Mutex(), rm->id, rm->addr, d.Default(), after);
+  Printf("%sM%lld (%p)%s%s", d.Mutex(), rm->id,
+         reinterpret_cast<void *>(rm->addr), d.Default(), after);
 }
 
 static void PrintMutex(const ReportMutex *rm) {
@@ -242,7 +229,8 @@ static void PrintMutex(const ReportMutex *rm) {
     Printf("%s", d.Default());
   } else {
     Printf("%s", d.Mutex());
-    Printf("  Mutex M%llu (%p) created at:\n", rm->id, rm->addr);
+    Printf("  Mutex M%llu (%p) created at:\n", rm->id,
+           reinterpret_cast<void *>(rm->addr));
     Printf("%s", d.Default());
     PrintStack(rm->stack);
   }
@@ -259,12 +247,13 @@ static void PrintThread(const ReportThread *rt) {
   char thrbuf[kThreadBufSize];
   const char *thread_status = rt->running ? "running" : "finished";
   if (rt->thread_type == ThreadType::Worker) {
-    Printf(" (tid=%zu, %s) is a GCD worker thread\n", rt->os_id, thread_status);
+    Printf(" (tid=%llu, %s) is a GCD worker thread\n", rt->os_id,
+           thread_status);
     Printf("\n");
     Printf("%s", d.Default());
     return;
   }
-  Printf(" (tid=%zu, %s) created by %s", rt->os_id, thread_status,
+  Printf(" (tid=%llu, %s) created by %s", rt->os_id, thread_status,
          thread_name(thrbuf, rt->parent_tid));
   if (rt->stack)
     Printf(" at:");
@@ -394,7 +383,7 @@ void PrintReport(const ReportDesc *rep) {
 
 #else  // #if !SANITIZER_GO
 
-const u32 kMainGoroutineId = 1;
+const Tid kMainGoroutineId = 1;
 
 void PrintStack(const ReportStack *ent) {
   if (ent == 0 || ent->frames == 0) {
@@ -405,16 +394,17 @@ void PrintStack(const ReportStack *ent) {
   for (int i = 0; frame; frame = frame->next, i++) {
     const AddressInfo &info = frame->info;
     Printf("  %s()\n      %s:%d +0x%zx\n", info.function,
-        StripPathPrefix(info.file, common_flags()->strip_path_prefix),
-        info.line, (void *)info.module_offset);
+           StripPathPrefix(info.file, common_flags()->strip_path_prefix),
+           info.line, info.module_offset);
   }
 }
 
 static void PrintMop(const ReportMop *mop, bool first) {
   Printf("\n");
   Printf("%s at %p by ",
-      (first ? (mop->write ? "Write" : "Read")
-             : (mop->write ? "Previous write" : "Previous read")), mop->addr);
+         (first ? (mop->write ? "Write" : "Read")
+                : (mop->write ? "Previous write" : "Previous read")),
+         reinterpret_cast<void *>(mop->addr));
   if (mop->tid == kMainGoroutineId)
     Printf("main goroutine:\n");
   else
@@ -426,8 +416,8 @@ static void PrintLocation(const ReportLocation *loc) {
   switch (loc->type) {
   case ReportLocationHeap: {
     Printf("\n");
-    Printf("Heap block of size %zu at %p allocated by ",
-        loc->heap_chunk_size, loc->heap_chunk_start);
+    Printf("Heap block of size %zu at %p allocated by ", loc->heap_chunk_size,
+           reinterpret_cast<void *>(loc->heap_chunk_start));
     if (loc->tid == kMainGoroutineId)
       Printf("main goroutine:\n");
     else
@@ -438,8 +428,9 @@ static void PrintLocation(const ReportLocation *loc) {
   case ReportLocationGlobal: {
     Printf("\n");
     Printf("Global var %s of size %zu at %p declared at %s:%zu\n",
-        loc->global.name, loc->global.size, loc->global.start,
-        loc->global.file, loc->global.line);
+           loc->global.name, loc->global.size,
+           reinterpret_cast<void *>(loc->global.start), loc->global.file,
+           loc->global.line);
     break;
   }
   default:
@@ -469,13 +460,13 @@ void PrintReport(const ReportDesc *rep) {
   } else if (rep->typ == ReportTypeDeadlock) {
     Printf("WARNING: DEADLOCK\n");
     for (uptr i = 0; i < rep->mutexes.Size(); i++) {
-      Printf("Goroutine %d lock mutex %d while holding mutex %d:\n",
-          999, rep->mutexes[i]->id,
-          rep->mutexes[(i+1) % rep->mutexes.Size()]->id);
+      Printf("Goroutine %d lock mutex %llu while holding mutex %llu:\n", 999,
+             rep->mutexes[i]->id,
+             rep->mutexes[(i + 1) % rep->mutexes.Size()]->id);
       PrintStack(rep->stacks[2*i]);
       Printf("\n");
-      Printf("Mutex %d was previously locked here:\n",
-          rep->mutexes[(i+1) % rep->mutexes.Size()]->id);
+      Printf("Mutex %llu was previously locked here:\n",
+             rep->mutexes[(i + 1) % rep->mutexes.Size()]->id);
       PrintStack(rep->stacks[2*i + 1]);
       Printf("\n");
     }
diff --git a/compiler-rt/lib/tsan/rtl/tsan_report.h b/compiler-rt/lib/tsan/rtl/tsan_report.h
index b4e4d8989379..d68c2db88828 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_report.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_report.h
@@ -38,12 +38,8 @@ enum ReportType {
 };
 
 struct ReportStack {
-  SymbolizedStack *frames;
-  bool suppressable;
-  static ReportStack *New();
-
- private:
-  ReportStack();
+  SymbolizedStack *frames = nullptr;
+  bool suppressable = false;
 };
 
 struct ReportMopMutex {
@@ -73,28 +69,24 @@ enum ReportLocationType {
 };
 
 struct ReportLocation {
-  ReportLocationType type;
-  DataInfo global;
-  uptr heap_chunk_start;
-  uptr heap_chunk_size;
-  uptr external_tag;
-  int tid;
-  int fd;
-  bool suppressable;
-  ReportStack *stack;
-
-  static ReportLocation *New(ReportLocationType type);
- private:
-  explicit ReportLocation(ReportLocationType type);
+  ReportLocationType type = ReportLocationGlobal;
+  DataInfo global = {};
+  uptr heap_chunk_start = 0;
+  uptr heap_chunk_size = 0;
+  uptr external_tag = 0;
+  Tid tid = kInvalidTid;
+  int fd = 0;
+  bool suppressable = false;
+  ReportStack *stack = nullptr;
 };
 
 struct ReportThread {
-  int id;
+  Tid id;
   tid_t os_id;
   bool running;
   ThreadType thread_type;
   char *name;
-  u32 parent_tid;
+  Tid parent_tid;
   ReportStack *stack;
 };
 
@@ -114,7 +106,7 @@ class ReportDesc {
   Vector<ReportLocation*> locs;
   Vector<ReportMutex*> mutexes;
   Vector<ReportThread*> threads;
-  Vector<int> unique_tids;
+  Vector<Tid> unique_tids;
   ReportStack *sleep;
   int count;
 
diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl.cpp b/compiler-rt/lib/tsan/rtl/tsan_rtl.cpp
index a21da9c81c6f..ff7726ef0608 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_rtl.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_rtl.cpp
@@ -28,16 +28,6 @@
 #include "tsan_symbolize.h"
 #include "ubsan/ubsan_init.h"
 
-#ifdef __SSE3__
-// <emmintrin.h> transitively includes <stdlib.h>,
-// and it's prohibited to include std headers into tsan runtime.
-// So we do this dirty trick.
-#define _MM_MALLOC_H_INCLUDED
-#define __MM_MALLOC_H
-#include <emmintrin.h>
-typedef __m128i m128;
-#endif
-
 volatile int __tsan_resumed = 0;
 
 extern "C" void __tsan_resume() {
@@ -46,11 +36,17 @@ extern "C" void __tsan_resume() {
 
 namespace __tsan {
 
+#if !SANITIZER_GO
+void (*on_initialize)(void);
+int (*on_finalize)(int);
+#endif
+
 #if !SANITIZER_GO && !SANITIZER_MAC
 __attribute__((tls_model("initial-exec")))
-THREADLOCAL char cur_thread_placeholder[sizeof(ThreadState)] ALIGNED(64);
+THREADLOCAL char cur_thread_placeholder[sizeof(ThreadState)] ALIGNED(
+    SANITIZER_CACHE_LINE_SIZE);
 #endif
-static char ctx_placeholder[sizeof(Context)] ALIGNED(64);
+static char ctx_placeholder[sizeof(Context)] ALIGNED(SANITIZER_CACHE_LINE_SIZE);
 Context *ctx;
 
 // Can be overriden by a front-end.
@@ -62,24 +58,21 @@ void OnInitialize();
 SANITIZER_WEAK_CXX_DEFAULT_IMPL
 bool OnFinalize(bool failed) {
 #if !SANITIZER_GO
-  if (auto *ptr = dlsym(RTLD_DEFAULT, "__tsan_on_finalize"))
-    return reinterpret_cast<decltype(&__tsan_on_finalize)>(ptr)(failed);
+  if (on_finalize)
+    return on_finalize(failed);
 #endif
   return failed;
 }
 SANITIZER_WEAK_CXX_DEFAULT_IMPL
 void OnInitialize() {
 #if !SANITIZER_GO
-  if (auto *ptr = dlsym(RTLD_DEFAULT, "__tsan_on_initialize")) {
-    return reinterpret_cast<decltype(&__tsan_on_initialize)>(ptr)();
-  }
+  if (on_initialize)
+    on_initialize();
 #endif
 }
 #endif
 
-static ALIGNED(64) char thread_registry_placeholder[sizeof(ThreadRegistry)];
-
-static ThreadContextBase *CreateThreadContext(u32 tid) {
+static ThreadContextBase *CreateThreadContext(Tid tid) {
   // Map thread trace when context is created.
   char name[50];
   internal_snprintf(name, sizeof(name), "trace %u", tid);
@@ -98,13 +91,12 @@ static ThreadContextBase *CreateThreadContext(u32 tid) {
     ReleaseMemoryPagesToOS(hdr_end, hdr + sizeof(Trace));
     uptr unused = hdr + sizeof(Trace) - hdr_end;
     if (hdr_end != (uptr)MmapFixedNoAccess(hdr_end, unused)) {
-      Report("ThreadSanitizer: failed to mprotect(%p, %p)\n",
-          hdr_end, unused);
+      Report("ThreadSanitizer: failed to mprotect [0x%zx-0x%zx) \n", hdr_end,
+             unused);
       CHECK("unable to mprotect" && 0);
     }
   }
-  void *mem = internal_alloc(MBlockThreadContex, sizeof(ThreadContext));
-  return new(mem) ThreadContext(tid);
+  return New<ThreadContext>(tid);
 }
 
 #if !SANITIZER_GO
@@ -117,9 +109,8 @@ Context::Context()
     : initialized(),
       report_mtx(MutexTypeReport),
       nreported(),
-      nmissed_expected(),
-      thread_registry(new (thread_registry_placeholder) ThreadRegistry(
-          CreateThreadContext, kMaxTid, kThreadQuarantineSize, kMaxTidReuse)),
+      thread_registry(CreateThreadContext, kMaxTid, kThreadQuarantineSize,
+                      kMaxTidReuse),
       racy_mtx(MutexTypeRacy),
       racy_stacks(),
       racy_addresses(),
@@ -129,7 +120,7 @@ Context::Context()
 }
 
 // The objects are allocated in TLS, so one may rely on zero-initialization.
-ThreadState::ThreadState(Context *ctx, u32 tid, int unique_id, u64 epoch,
+ThreadState::ThreadState(Context *ctx, Tid tid, int unique_id, u64 epoch,
                          unsigned reuse_count, uptr stk_addr, uptr stk_size,
                          uptr tls_addr, uptr tls_size)
     : fast_state(tid, epoch)
@@ -155,16 +146,53 @@ ThreadState::ThreadState(Context *ctx, u32 tid, int unique_id, u64 epoch,
       last_sleep_clock(tid)
 #endif
 {
+  CHECK_EQ(reinterpret_cast<uptr>(this) % SANITIZER_CACHE_LINE_SIZE, 0);
+#if !SANITIZER_GO
+  // C/C++ uses fixed size shadow stack.
+  const int kInitStackSize = kShadowStackSize;
+  shadow_stack = static_cast<uptr *>(
+      MmapNoReserveOrDie(kInitStackSize * sizeof(uptr), "shadow stack"));
+  SetShadowRegionHugePageMode(reinterpret_cast<uptr>(shadow_stack),
+                              kInitStackSize * sizeof(uptr));
+#else
+  // Go uses malloc-allocated shadow stack with dynamic size.
+  const int kInitStackSize = 8;
+  shadow_stack = static_cast<uptr *>(Alloc(kInitStackSize * sizeof(uptr)));
+#endif
+  shadow_stack_pos = shadow_stack;
+  shadow_stack_end = shadow_stack + kInitStackSize;
 }
 
 #if !SANITIZER_GO
-static void MemoryProfiler(Context *ctx, fd_t fd, int i) {
-  uptr n_threads;
-  uptr n_running_threads;
-  ctx->thread_registry->GetNumberOfThreads(&n_threads, &n_running_threads);
+void MemoryProfiler(u64 uptime) {
+  if (ctx->memprof_fd == kInvalidFd)
+    return;
   InternalMmapVector<char> buf(4096);
-  WriteMemoryProfile(buf.data(), buf.size(), n_threads, n_running_threads);
-  WriteToFile(fd, buf.data(), internal_strlen(buf.data()));
+  WriteMemoryProfile(buf.data(), buf.size(), uptime);
+  WriteToFile(ctx->memprof_fd, buf.data(), internal_strlen(buf.data()));
+}
+
+void InitializeMemoryProfiler() {
+  ctx->memprof_fd = kInvalidFd;
+  const char *fname = flags()->profile_memory;
+  if (!fname || !fname[0])
+    return;
+  if (internal_strcmp(fname, "stdout") == 0) {
+    ctx->memprof_fd = 1;
+  } else if (internal_strcmp(fname, "stderr") == 0) {
+    ctx->memprof_fd = 2;
+  } else {
+    InternalScopedString filename;
+    filename.append("%s.%d", fname, (int)internal_getpid());
+    ctx->memprof_fd = OpenFile(filename.data(), WrOnly);
+    if (ctx->memprof_fd == kInvalidFd) {
+      Printf("ThreadSanitizer: failed to open memory profile file '%s'\n",
+             filename.data());
+      return;
+    }
+  }
+  MemoryProfiler(0);
+  MaybeSpawnBackgroundThread();
 }
 
 static void *BackgroundThread(void *arg) {
@@ -172,28 +200,9 @@ static void *BackgroundThread(void *arg) {
   // We don't use ScopedIgnoreInterceptors, because we want ignores to be
   // enabled even when the thread function exits (e.g. during pthread thread
   // shutdown code).
-  cur_thread_init();
-  cur_thread()->ignore_interceptors++;
+  cur_thread_init()->ignore_interceptors++;
   const u64 kMs2Ns = 1000 * 1000;
-
-  fd_t mprof_fd = kInvalidFd;
-  if (flags()->profile_memory && flags()->profile_memory[0]) {
-    if (internal_strcmp(flags()->profile_memory, "stdout") == 0) {
-      mprof_fd = 1;
-    } else if (internal_strcmp(flags()->profile_memory, "stderr") == 0) {
-      mprof_fd = 2;
-    } else {
-      InternalScopedString filename;
-      filename.append("%s.%d", flags()->profile_memory, (int)internal_getpid());
-      fd_t fd = OpenFile(filename.data(), WrOnly);
-      if (fd == kInvalidFd) {
-        Printf("ThreadSanitizer: failed to open memory profile file '%s'\n",
-               filename.data());
-      } else {
-        mprof_fd = fd;
-      }
-    }
-  }
+  const u64 start = NanoTime();
 
   u64 last_flush = NanoTime();
   uptr last_rss = 0;
@@ -211,7 +220,6 @@ static void *BackgroundThread(void *arg) {
         last_flush = NanoTime();
       }
     }
-    // GetRSS can be expensive on huge programs, so don't do it every 100ms.
     if (flags()->memory_limit_mb > 0) {
       uptr rss = GetRSS();
       uptr limit = uptr(flags()->memory_limit_mb) << 20;
@@ -227,9 +235,7 @@ static void *BackgroundThread(void *arg) {
       last_rss = rss;
     }
 
-    // Write memory profile if requested.
-    if (mprof_fd != kInvalidFd)
-      MemoryProfiler(ctx, mprof_fd, i);
+    MemoryProfiler(now - start);
 
     // Flush symbolizer cache if requested.
     if (flags()->flush_symbolizer_ms > 0) {
@@ -260,7 +266,8 @@ static void StopBackgroundThread() {
 #endif
 
 void DontNeedShadowFor(uptr addr, uptr size) {
-  ReleaseMemoryPagesToOS(MemToShadow(addr), MemToShadow(addr + size));
+  ReleaseMemoryPagesToOS(reinterpret_cast<uptr>(MemToShadow(addr)),
+                         reinterpret_cast<uptr>(MemToShadow(addr + size)));
 }
 
 #if !SANITIZER_GO
@@ -297,7 +304,7 @@ void MapShadow(uptr addr, uptr size) {
                                  "meta shadow"))
       Die();
   } else {
-    // Mapping continous heap.
+    // Mapping continuous heap.
     // Windows wants 64K alignment.
     meta_begin = RoundDownTo(meta_begin, 64 << 10);
     meta_end = RoundUpTo(meta_end, 64 << 10);
@@ -310,58 +317,22 @@ void MapShadow(uptr addr, uptr size) {
       Die();
     mapped_meta_end = meta_end;
   }
-  VPrintf(2, "mapped meta shadow for (%p-%p) at (%p-%p)\n",
-      addr, addr+size, meta_begin, meta_end);
+  VPrintf(2, "mapped meta shadow for (0x%zx-0x%zx) at (0x%zx-0x%zx)\n", addr,
+          addr + size, meta_begin, meta_end);
 }
 
 void MapThreadTrace(uptr addr, uptr size, const char *name) {
-  DPrintf("#0: Mapping trace at %p-%p(0x%zx)\n", addr, addr + size, size);
+  DPrintf("#0: Mapping trace at 0x%zx-0x%zx(0x%zx)\n", addr, addr + size, size);
   CHECK_GE(addr, TraceMemBeg());
   CHECK_LE(addr + size, TraceMemEnd());
   CHECK_EQ(addr, addr & ~((64 << 10) - 1));  // windows wants 64K alignment
   if (!MmapFixedSuperNoReserve(addr, size, name)) {
-    Printf("FATAL: ThreadSanitizer can not mmap thread trace (%p/%p)\n",
-        addr, size);
+    Printf("FATAL: ThreadSanitizer can not mmap thread trace (0x%zx/0x%zx)\n",
+           addr, size);
     Die();
   }
 }
 
-static void CheckShadowMapping() {
-  uptr beg, end;
-  for (int i = 0; GetUserRegion(i, &beg, &end); i++) {
-    // Skip cases for empty regions (heap definition for architectures that
-    // do not use 64-bit allocator).
-    if (beg == end)
-      continue;
-    VPrintf(3, "checking shadow region %p-%p\n", beg, end);
-    uptr prev = 0;
-    for (uptr p0 = beg; p0 <= end; p0 += (end - beg) / 4) {
-      for (int x = -(int)kShadowCell; x <= (int)kShadowCell; x += kShadowCell) {
-        const uptr p = RoundDown(p0 + x, kShadowCell);
-        if (p < beg || p >= end)
-          continue;
-        const uptr s = MemToShadow(p);
-        const uptr m = (uptr)MemToMeta(p);
-        VPrintf(3, "  checking pointer %p: shadow=%p meta=%p\n", p, s, m);
-        CHECK(IsAppMem(p));
-        CHECK(IsShadowMem(s));
-        CHECK_EQ(p, ShadowToMem(s));
-        CHECK(IsMetaMem(m));
-        if (prev) {
-          // Ensure that shadow and meta mappings are linear within a single
-          // user range. Lots of code that processes memory ranges assumes it.
-          const uptr prev_s = MemToShadow(prev);
-          const uptr prev_m = (uptr)MemToMeta(prev);
-          CHECK_EQ(s - prev_s, (p - prev) * kShadowMultiplier);
-          CHECK_EQ((m - prev_m) / kMetaShadowSize,
-                   (p - prev) / kMetaShadowCell);
-        }
-        prev = p;
-      }
-    }
-  }
-}
-
 #if !SANITIZER_GO
 static void OnStackUnwind(const SignalContext &sig, const void *,
                           BufferedStackTrace *stack) {
@@ -386,9 +357,10 @@ void CheckUnwind() {
   PrintCurrentStackSlow(StackTrace::GetCurrentPc());
 }
 
+bool is_initialized;
+
 void Initialize(ThreadState *thr) {
   // Thread safe because done before all threads exist.
-  static bool is_initialized = false;
   if (is_initialized)
     return;
   is_initialized = true;
@@ -420,7 +392,6 @@ void Initialize(ThreadState *thr) {
   Processor *proc = ProcCreate();
   ProcWire(proc, thr);
   InitializeInterceptors();
-  CheckShadowMapping();
   InitializePlatform();
   InitializeDynamicAnnotations();
 #if !SANITIZER_GO
@@ -440,8 +411,8 @@ void Initialize(ThreadState *thr) {
           (int)internal_getpid());
 
   // Initialize thread 0.
-  int tid = ThreadCreate(thr, 0, 0, true);
-  CHECK_EQ(tid, 0);
+  Tid tid = ThreadCreate(thr, 0, 0, true);
+  CHECK_EQ(tid, kMainTid);
   ThreadStart(thr, tid, GetTid(), ThreadType::Regular);
 #if TSAN_CONTAINS_UBSAN
   __ubsan::InitAsPlugin();
@@ -450,6 +421,7 @@ void Initialize(ThreadState *thr) {
 
 #if !SANITIZER_GO
   Symbolizer::LateInitialize();
+  InitializeMemoryProfiler();
 #endif
 
   if (flags()->stop_on_start) {
@@ -506,18 +478,8 @@ int Finalize(ThreadState *thr) {
 #endif
   }
 
-  if (ctx->nmissed_expected) {
-    failed = true;
-    Printf("ThreadSanitizer: missed %d expected races\n",
-        ctx->nmissed_expected);
-  }
-
   if (common_flags()->print_suppressions)
     PrintMatchedSuppressions();
-#if !SANITIZER_GO
-  if (flags()->print_benign)
-    PrintMatchedBenignRaces();
-#endif
 
   failed = OnFinalize(failed);
 
@@ -526,7 +488,7 @@ int Finalize(ThreadState *thr) {
 
 #if !SANITIZER_GO
 void ForkBefore(ThreadState *thr, uptr pc) NO_THREAD_SAFETY_ANALYSIS {
-  ctx->thread_registry->Lock();
+  ctx->thread_registry.Lock();
   ctx->report_mtx.Lock();
   ScopedErrorReportLock::Lock();
   // Suppress all reports in the pthread_atfork callbacks.
@@ -545,22 +507,24 @@ void ForkParentAfter(ThreadState *thr, uptr pc) NO_THREAD_SAFETY_ANALYSIS {
   thr->ignore_interceptors--;
   ScopedErrorReportLock::Unlock();
   ctx->report_mtx.Unlock();
-  ctx->thread_registry->Unlock();
+  ctx->thread_registry.Unlock();
 }
 
-void ForkChildAfter(ThreadState *thr, uptr pc) NO_THREAD_SAFETY_ANALYSIS {
+void ForkChildAfter(ThreadState *thr, uptr pc,
+                    bool start_thread) NO_THREAD_SAFETY_ANALYSIS {
   thr->suppress_reports--;  // Enabled in ForkBefore.
   thr->ignore_interceptors--;
   ScopedErrorReportLock::Unlock();
   ctx->report_mtx.Unlock();
-  ctx->thread_registry->Unlock();
+  ctx->thread_registry.Unlock();
 
   uptr nthread = 0;
-  ctx->thread_registry->GetNumberOfThreads(0, 0, &nthread /* alive threads */);
+  ctx->thread_registry.GetNumberOfThreads(0, 0, &nthread /* alive threads */);
   VPrintf(1, "ThreadSanitizer: forked new process with pid %d,"
       " parent had %d threads\n", (int)internal_getpid(), (int)nthread);
   if (nthread == 1) {
-    StartBackgroundThread();
+    if (start_thread)
+      StartBackgroundThread();
   } else {
     // We've just forked a multi-threaded process. We cannot reasonably function
     // after that (some mutexes may be locked before fork). So just enable
@@ -578,19 +542,18 @@ NOINLINE
 void GrowShadowStack(ThreadState *thr) {
   const int sz = thr->shadow_stack_end - thr->shadow_stack;
   const int newsz = 2 * sz;
-  uptr *newstack = (uptr*)internal_alloc(MBlockShadowStack,
-      newsz * sizeof(uptr));
+  auto *newstack = (uptr *)Alloc(newsz * sizeof(uptr));
   internal_memcpy(newstack, thr->shadow_stack, sz * sizeof(uptr));
-  internal_free(thr->shadow_stack);
+  Free(thr->shadow_stack);
   thr->shadow_stack = newstack;
   thr->shadow_stack_pos = newstack + sz;
   thr->shadow_stack_end = newstack + newsz;
 }
 #endif
 
-u32 CurrentStackId(ThreadState *thr, uptr pc) {
+StackID CurrentStackId(ThreadState *thr, uptr pc) {
   if (!thr->is_inited)  // May happen during bootstrap.
-    return 0;
+    return kInvalidStackID;
   if (pc != 0) {
 #if !SANITIZER_GO
     DCHECK_LT(thr->shadow_stack_pos, thr->shadow_stack_end);
@@ -601,13 +564,78 @@ u32 CurrentStackId(ThreadState *thr, uptr pc) {
     thr->shadow_stack_pos[0] = pc;
     thr->shadow_stack_pos++;
   }
-  u32 id = StackDepotPut(
+  StackID id = StackDepotPut(
       StackTrace(thr->shadow_stack, thr->shadow_stack_pos - thr->shadow_stack));
   if (pc != 0)
     thr->shadow_stack_pos--;
   return id;
 }
 
+namespace v3 {
+
+NOINLINE
+void TraceSwitchPart(ThreadState *thr) {
+  Trace *trace = &thr->tctx->trace;
+  Event *pos = reinterpret_cast<Event *>(atomic_load_relaxed(&thr->trace_pos));
+  DCHECK_EQ(reinterpret_cast<uptr>(pos + 1) & TracePart::kAlignment, 0);
+  auto *part = trace->parts.Back();
+  DPrintf("TraceSwitchPart part=%p pos=%p\n", part, pos);
+  if (part) {
+    // We can get here when we still have space in the current trace part.
+    // The fast-path check in TraceAcquire has false positives in the middle of
+    // the part. Check if we are indeed at the end of the current part or not,
+    // and fill any gaps with NopEvent's.
+    Event *end = &part->events[TracePart::kSize];
+    DCHECK_GE(pos, &part->events[0]);
+    DCHECK_LE(pos, end);
+    if (pos + 1 < end) {
+      if ((reinterpret_cast<uptr>(pos) & TracePart::kAlignment) ==
+          TracePart::kAlignment)
+        *pos++ = NopEvent;
+      *pos++ = NopEvent;
+      DCHECK_LE(pos + 2, end);
+      atomic_store_relaxed(&thr->trace_pos, reinterpret_cast<uptr>(pos));
+      // Ensure we setup trace so that the next TraceAcquire
+      // won't detect trace part end.
+      Event *ev;
+      CHECK(TraceAcquire(thr, &ev));
+      return;
+    }
+    // We are indeed at the end.
+    for (; pos < end; pos++) *pos = NopEvent;
+  }
+#if !SANITIZER_GO
+  if (ctx->after_multithreaded_fork) {
+    // We just need to survive till exec.
+    CHECK(part);
+    atomic_store_relaxed(&thr->trace_pos,
+                         reinterpret_cast<uptr>(&part->events[0]));
+    return;
+  }
+#endif
+  part = new (MmapOrDie(sizeof(TracePart), "TracePart")) TracePart();
+  part->trace = trace;
+  thr->trace_prev_pc = 0;
+  {
+    Lock lock(&trace->mtx);
+    trace->parts.PushBack(part);
+    atomic_store_relaxed(&thr->trace_pos,
+                         reinterpret_cast<uptr>(&part->events[0]));
+  }
+  // Make this part self-sufficient by restoring the current stack
+  // and mutex set in the beginning of the trace.
+  TraceTime(thr);
+  for (uptr *pos = &thr->shadow_stack[0]; pos < thr->shadow_stack_pos; pos++)
+    CHECK(TryTraceFunc(thr, *pos));
+  for (uptr i = 0; i < thr->mset.Size(); i++) {
+    MutexSet::Desc d = thr->mset.Get(i);
+    TraceMutexLock(thr, d.write ? EventType::kLock : EventType::kRLock, 0,
+                   d.addr, d.stack_id);
+  }
+}
+
+}  // namespace v3
+
 void TraceSwitch(ThreadState *thr) {
 #if !SANITIZER_GO
   if (ctx->after_multithreaded_fork)
@@ -624,9 +652,7 @@ void TraceSwitch(ThreadState *thr) {
   thr->nomalloc--;
 }
 
-Trace *ThreadTrace(int tid) {
-  return (Trace*)GetThreadTraceHeader(tid);
-}
+Trace *ThreadTrace(Tid tid) { return (Trace *)GetThreadTraceHeader(tid); }
 
 uptr TraceTopPC(ThreadState *thr) {
   Event *events = (Event*)GetThreadTrace(thr->tid);
@@ -652,435 +678,18 @@ extern "C" void __tsan_report_race() {
 }
 #endif
 
-ALWAYS_INLINE
-Shadow LoadShadow(u64 *p) {
-  u64 raw = atomic_load((atomic_uint64_t*)p, memory_order_relaxed);
-  return Shadow(raw);
-}
-
-ALWAYS_INLINE
-void StoreShadow(u64 *sp, u64 s) {
-  atomic_store((atomic_uint64_t*)sp, s, memory_order_relaxed);
-}
-
-ALWAYS_INLINE
-void StoreIfNotYetStored(u64 *sp, u64 *s) {
-  StoreShadow(sp, *s);
-  *s = 0;
-}
-
-ALWAYS_INLINE
-void HandleRace(ThreadState *thr, u64 *shadow_mem,
-                              Shadow cur, Shadow old) {
-  thr->racy_state[0] = cur.raw();
-  thr->racy_state[1] = old.raw();
-  thr->racy_shadow_addr = shadow_mem;
-#if !SANITIZER_GO
-  HACKY_CALL(__tsan_report_race);
-#else
-  ReportRace(thr);
-#endif
-}
-
-static inline bool HappensBefore(Shadow old, ThreadState *thr) {
-  return thr->clock.get(old.TidWithIgnore()) >= old.epoch();
-}
-
-ALWAYS_INLINE
-void MemoryAccessImpl1(ThreadState *thr, uptr addr,
-    int kAccessSizeLog, bool kAccessIsWrite, bool kIsAtomic,
-    u64 *shadow_mem, Shadow cur) {
-
-  // This potentially can live in an MMX/SSE scratch register.
-  // The required intrinsics are:
-  // __m128i _mm_move_epi64(__m128i*);
-  // _mm_storel_epi64(u64*, __m128i);
-  u64 store_word = cur.raw();
-  bool stored = false;
-
-  // scan all the shadow values and dispatch to 4 categories:
-  // same, replace, candidate and race (see comments below).
-  // we consider only 3 cases regarding access sizes:
-  // equal, intersect and not intersect. initially I considered
-  // larger and smaller as well, it allowed to replace some
-  // 'candidates' with 'same' or 'replace', but I think
-  // it's just not worth it (performance- and complexity-wise).
-
-  Shadow old(0);
-
-  // It release mode we manually unroll the loop,
-  // because empirically gcc generates better code this way.
-  // However, we can't afford unrolling in debug mode, because the function
-  // consumes almost 4K of stack. Gtest gives only 4K of stack to death test
-  // threads, which is not enough for the unrolled loop.
-#if SANITIZER_DEBUG
-  for (int idx = 0; idx < 4; idx++) {
-#include "tsan_update_shadow_word_inl.h"
-  }
-#else
-  int idx = 0;
-#include "tsan_update_shadow_word_inl.h"
-  idx = 1;
-  if (stored) {
-#include "tsan_update_shadow_word_inl.h"
-  } else {
-#include "tsan_update_shadow_word_inl.h"
-  }
-  idx = 2;
-  if (stored) {
-#include "tsan_update_shadow_word_inl.h"
-  } else {
-#include "tsan_update_shadow_word_inl.h"
-  }
-  idx = 3;
-  if (stored) {
-#include "tsan_update_shadow_word_inl.h"
-  } else {
-#include "tsan_update_shadow_word_inl.h"
-  }
-#endif
-
-  // we did not find any races and had already stored
-  // the current access info, so we are done
-  if (LIKELY(stored))
-    return;
-  // choose a random candidate slot and replace it
-  StoreShadow(shadow_mem + (cur.epoch() % kShadowCnt), store_word);
-  return;
- RACE:
-  HandleRace(thr, shadow_mem, cur, old);
-  return;
-}
-
-void UnalignedMemoryAccess(ThreadState *thr, uptr pc, uptr addr,
-    int size, bool kAccessIsWrite, bool kIsAtomic) {
-  while (size) {
-    int size1 = 1;
-    int kAccessSizeLog = kSizeLog1;
-    if (size >= 8 && (addr & ~7) == ((addr + 7) & ~7)) {
-      size1 = 8;
-      kAccessSizeLog = kSizeLog8;
-    } else if (size >= 4 && (addr & ~7) == ((addr + 3) & ~7)) {
-      size1 = 4;
-      kAccessSizeLog = kSizeLog4;
-    } else if (size >= 2 && (addr & ~7) == ((addr + 1) & ~7)) {
-      size1 = 2;
-      kAccessSizeLog = kSizeLog2;
-    }
-    MemoryAccess(thr, pc, addr, kAccessSizeLog, kAccessIsWrite, kIsAtomic);
-    addr += size1;
-    size -= size1;
-  }
-}
-
-ALWAYS_INLINE
-bool ContainsSameAccessSlow(u64 *s, u64 a, u64 sync_epoch, bool is_write) {
-  Shadow cur(a);
-  for (uptr i = 0; i < kShadowCnt; i++) {
-    Shadow old(LoadShadow(&s[i]));
-    if (Shadow::Addr0AndSizeAreEqual(cur, old) &&
-        old.TidWithIgnore() == cur.TidWithIgnore() &&
-        old.epoch() > sync_epoch &&
-        old.IsAtomic() == cur.IsAtomic() &&
-        old.IsRead() <= cur.IsRead())
-      return true;
-  }
-  return false;
-}
-
-#if defined(__SSE3__)
-#define SHUF(v0, v1, i0, i1, i2, i3) _mm_castps_si128(_mm_shuffle_ps( \
-    _mm_castsi128_ps(v0), _mm_castsi128_ps(v1), \
-    (i0)*1 + (i1)*4 + (i2)*16 + (i3)*64))
-ALWAYS_INLINE
-bool ContainsSameAccessFast(u64 *s, u64 a, u64 sync_epoch, bool is_write) {
-  // This is an optimized version of ContainsSameAccessSlow.
-  // load current access into access[0:63]
-  const m128 access     = _mm_cvtsi64_si128(a);
-  // duplicate high part of access in addr0:
-  // addr0[0:31]        = access[32:63]
-  // addr0[32:63]       = access[32:63]
-  // addr0[64:95]       = access[32:63]
-  // addr0[96:127]      = access[32:63]
-  const m128 addr0      = SHUF(access, access, 1, 1, 1, 1);
-  // load 4 shadow slots
-  const m128 shadow0    = _mm_load_si128((__m128i*)s);
-  const m128 shadow1    = _mm_load_si128((__m128i*)s + 1);
-  // load high parts of 4 shadow slots into addr_vect:
-  // addr_vect[0:31]    = shadow0[32:63]
-  // addr_vect[32:63]   = shadow0[96:127]
-  // addr_vect[64:95]   = shadow1[32:63]
-  // addr_vect[96:127]  = shadow1[96:127]
-  m128 addr_vect        = SHUF(shadow0, shadow1, 1, 3, 1, 3);
-  if (!is_write) {
-    // set IsRead bit in addr_vect
-    const m128 rw_mask1 = _mm_cvtsi64_si128(1<<15);
-    const m128 rw_mask  = SHUF(rw_mask1, rw_mask1, 0, 0, 0, 0);
-    addr_vect           = _mm_or_si128(addr_vect, rw_mask);
-  }
-  // addr0 == addr_vect?
-  const m128 addr_res   = _mm_cmpeq_epi32(addr0, addr_vect);
-  // epoch1[0:63]       = sync_epoch
-  const m128 epoch1     = _mm_cvtsi64_si128(sync_epoch);
-  // epoch[0:31]        = sync_epoch[0:31]
-  // epoch[32:63]       = sync_epoch[0:31]
-  // epoch[64:95]       = sync_epoch[0:31]
-  // epoch[96:127]      = sync_epoch[0:31]
-  const m128 epoch      = SHUF(epoch1, epoch1, 0, 0, 0, 0);
-  // load low parts of shadow cell epochs into epoch_vect:
-  // epoch_vect[0:31]   = shadow0[0:31]
-  // epoch_vect[32:63]  = shadow0[64:95]
-  // epoch_vect[64:95]  = shadow1[0:31]
-  // epoch_vect[96:127] = shadow1[64:95]
-  const m128 epoch_vect = SHUF(shadow0, shadow1, 0, 2, 0, 2);
-  // epoch_vect >= sync_epoch?
-  const m128 epoch_res  = _mm_cmpgt_epi32(epoch_vect, epoch);
-  // addr_res & epoch_res
-  const m128 res        = _mm_and_si128(addr_res, epoch_res);
-  // mask[0] = res[7]
-  // mask[1] = res[15]
-  // ...
-  // mask[15] = res[127]
-  const int mask        = _mm_movemask_epi8(res);
-  return mask != 0;
-}
-#endif
-
-ALWAYS_INLINE
-bool ContainsSameAccess(u64 *s, u64 a, u64 sync_epoch, bool is_write) {
-#if defined(__SSE3__)
-  bool res = ContainsSameAccessFast(s, a, sync_epoch, is_write);
-  // NOTE: this check can fail if the shadow is concurrently mutated
-  // by other threads. But it still can be useful if you modify
-  // ContainsSameAccessFast and want to ensure that it's not completely broken.
-  // DCHECK_EQ(res, ContainsSameAccessSlow(s, a, sync_epoch, is_write));
-  return res;
-#else
-  return ContainsSameAccessSlow(s, a, sync_epoch, is_write);
-#endif
-}
-
-ALWAYS_INLINE USED
-void MemoryAccess(ThreadState *thr, uptr pc, uptr addr,
-    int kAccessSizeLog, bool kAccessIsWrite, bool kIsAtomic) {
-  u64 *shadow_mem = (u64*)MemToShadow(addr);
-  DPrintf2("#%d: MemoryAccess: @%p %p size=%d"
-      " is_write=%d shadow_mem=%p {%zx, %zx, %zx, %zx}\n",
-      (int)thr->fast_state.tid(), (void*)pc, (void*)addr,
-      (int)(1 << kAccessSizeLog), kAccessIsWrite, shadow_mem,
-      (uptr)shadow_mem[0], (uptr)shadow_mem[1],
-      (uptr)shadow_mem[2], (uptr)shadow_mem[3]);
-#if SANITIZER_DEBUG
-  if (!IsAppMem(addr)) {
-    Printf("Access to non app mem %zx\n", addr);
-    DCHECK(IsAppMem(addr));
-  }
-  if (!IsShadowMem((uptr)shadow_mem)) {
-    Printf("Bad shadow addr %p (%zx)\n", shadow_mem, addr);
-    DCHECK(IsShadowMem((uptr)shadow_mem));
-  }
-#endif
-
-  if (!SANITIZER_GO && !kAccessIsWrite && *shadow_mem == kShadowRodata) {
-    // Access to .rodata section, no races here.
-    // Measurements show that it can be 10-20% of all memory accesses.
-    return;
-  }
-
-  FastState fast_state = thr->fast_state;
-  if (UNLIKELY(fast_state.GetIgnoreBit())) {
-    return;
-  }
-
-  Shadow cur(fast_state);
-  cur.SetAddr0AndSizeLog(addr & 7, kAccessSizeLog);
-  cur.SetWrite(kAccessIsWrite);
-  cur.SetAtomic(kIsAtomic);
-
-  if (LIKELY(ContainsSameAccess(shadow_mem, cur.raw(),
-      thr->fast_synch_epoch, kAccessIsWrite))) {
-    return;
-  }
-
-  if (kCollectHistory) {
-    fast_state.IncrementEpoch();
-    thr->fast_state = fast_state;
-    TraceAddEvent(thr, fast_state, EventTypeMop, pc);
-    cur.IncrementEpoch();
-  }
-
-  MemoryAccessImpl1(thr, addr, kAccessSizeLog, kAccessIsWrite, kIsAtomic,
-      shadow_mem, cur);
-}
-
-// Called by MemoryAccessRange in tsan_rtl_thread.cpp
-ALWAYS_INLINE USED
-void MemoryAccessImpl(ThreadState *thr, uptr addr,
-    int kAccessSizeLog, bool kAccessIsWrite, bool kIsAtomic,
-    u64 *shadow_mem, Shadow cur) {
-  if (LIKELY(ContainsSameAccess(shadow_mem, cur.raw(),
-      thr->fast_synch_epoch, kAccessIsWrite))) {
-    return;
-  }
-
-  MemoryAccessImpl1(thr, addr, kAccessSizeLog, kAccessIsWrite, kIsAtomic,
-      shadow_mem, cur);
-}
-
-static void MemoryRangeSet(ThreadState *thr, uptr pc, uptr addr, uptr size,
-                           u64 val) {
-  (void)thr;
-  (void)pc;
-  if (size == 0)
-    return;
-  // FIXME: fix me.
-  uptr offset = addr % kShadowCell;
-  if (offset) {
-    offset = kShadowCell - offset;
-    if (size <= offset)
-      return;
-    addr += offset;
-    size -= offset;
-  }
-  DCHECK_EQ(addr % 8, 0);
-  // If a user passes some insane arguments (memset(0)),
-  // let it just crash as usual.
-  if (!IsAppMem(addr) || !IsAppMem(addr + size - 1))
-    return;
-  // Don't want to touch lots of shadow memory.
-  // If a program maps 10MB stack, there is no need reset the whole range.
-  size = (size + (kShadowCell - 1)) & ~(kShadowCell - 1);
-  // UnmapOrDie/MmapFixedNoReserve does not work on Windows.
-  if (SANITIZER_WINDOWS || size < common_flags()->clear_shadow_mmap_threshold) {
-    u64 *p = (u64*)MemToShadow(addr);
-    CHECK(IsShadowMem((uptr)p));
-    CHECK(IsShadowMem((uptr)(p + size * kShadowCnt / kShadowCell - 1)));
-    // FIXME: may overwrite a part outside the region
-    for (uptr i = 0; i < size / kShadowCell * kShadowCnt;) {
-      p[i++] = val;
-      for (uptr j = 1; j < kShadowCnt; j++)
-        p[i++] = 0;
-    }
-  } else {
-    // The region is big, reset only beginning and end.
-    const uptr kPageSize = GetPageSizeCached();
-    u64 *begin = (u64*)MemToShadow(addr);
-    u64 *end = begin + size / kShadowCell * kShadowCnt;
-    u64 *p = begin;
-    // Set at least first kPageSize/2 to page boundary.
-    while ((p < begin + kPageSize / kShadowSize / 2) || ((uptr)p % kPageSize)) {
-      *p++ = val;
-      for (uptr j = 1; j < kShadowCnt; j++)
-        *p++ = 0;
-    }
-    // Reset middle part.
-    u64 *p1 = p;
-    p = RoundDown(end, kPageSize);
-    if (!MmapFixedSuperNoReserve((uptr)p1, (uptr)p - (uptr)p1))
-      Die();
-    // Set the ending.
-    while (p < end) {
-      *p++ = val;
-      for (uptr j = 1; j < kShadowCnt; j++)
-        *p++ = 0;
-    }
-  }
-}
-
-void MemoryResetRange(ThreadState *thr, uptr pc, uptr addr, uptr size) {
-  MemoryRangeSet(thr, pc, addr, size, 0);
-}
-
-void MemoryRangeFreed(ThreadState *thr, uptr pc, uptr addr, uptr size) {
-  // Processing more than 1k (4k of shadow) is expensive,
-  // can cause excessive memory consumption (user does not necessary touch
-  // the whole range) and most likely unnecessary.
-  if (size > 1024)
-    size = 1024;
-  CHECK_EQ(thr->is_freeing, false);
-  thr->is_freeing = true;
-  MemoryAccessRange(thr, pc, addr, size, true);
-  thr->is_freeing = false;
-  if (kCollectHistory) {
-    thr->fast_state.IncrementEpoch();
-    TraceAddEvent(thr, thr->fast_state, EventTypeMop, pc);
-  }
-  Shadow s(thr->fast_state);
-  s.ClearIgnoreBit();
-  s.MarkAsFreed();
-  s.SetWrite(true);
-  s.SetAddr0AndSizeLog(0, 3);
-  MemoryRangeSet(thr, pc, addr, size, s.raw());
-}
-
-void MemoryRangeImitateWrite(ThreadState *thr, uptr pc, uptr addr, uptr size) {
-  if (kCollectHistory) {
-    thr->fast_state.IncrementEpoch();
-    TraceAddEvent(thr, thr->fast_state, EventTypeMop, pc);
-  }
-  Shadow s(thr->fast_state);
-  s.ClearIgnoreBit();
-  s.SetWrite(true);
-  s.SetAddr0AndSizeLog(0, 3);
-  MemoryRangeSet(thr, pc, addr, size, s.raw());
-}
-
-void MemoryRangeImitateWriteOrResetRange(ThreadState *thr, uptr pc, uptr addr,
-                                         uptr size) {
-  if (thr->ignore_reads_and_writes == 0)
-    MemoryRangeImitateWrite(thr, pc, addr, size);
-  else
-    MemoryResetRange(thr, pc, addr, size);
-}
-
-ALWAYS_INLINE USED
-void FuncEntry(ThreadState *thr, uptr pc) {
-  DPrintf2("#%d: FuncEntry %p\n", (int)thr->fast_state.tid(), (void*)pc);
-  if (kCollectHistory) {
-    thr->fast_state.IncrementEpoch();
-    TraceAddEvent(thr, thr->fast_state, EventTypeFuncEnter, pc);
-  }
-
-  // Shadow stack maintenance can be replaced with
-  // stack unwinding during trace switch (which presumably must be faster).
-  DCHECK_GE(thr->shadow_stack_pos, thr->shadow_stack);
-#if !SANITIZER_GO
-  DCHECK_LT(thr->shadow_stack_pos, thr->shadow_stack_end);
-#else
-  if (thr->shadow_stack_pos == thr->shadow_stack_end)
-    GrowShadowStack(thr);
-#endif
-  thr->shadow_stack_pos[0] = pc;
-  thr->shadow_stack_pos++;
-}
-
-ALWAYS_INLINE USED
-void FuncExit(ThreadState *thr) {
-  DPrintf2("#%d: FuncExit\n", (int)thr->fast_state.tid());
-  if (kCollectHistory) {
-    thr->fast_state.IncrementEpoch();
-    TraceAddEvent(thr, thr->fast_state, EventTypeFuncExit, 0);
-  }
-
-  DCHECK_GT(thr->shadow_stack_pos, thr->shadow_stack);
-#if !SANITIZER_GO
-  DCHECK_LT(thr->shadow_stack_pos, thr->shadow_stack_end);
-#endif
-  thr->shadow_stack_pos--;
-}
-
-void ThreadIgnoreBegin(ThreadState *thr, uptr pc, bool save_stack) {
+void ThreadIgnoreBegin(ThreadState *thr, uptr pc) {
   DPrintf("#%d: ThreadIgnoreBegin\n", thr->tid);
   thr->ignore_reads_and_writes++;
   CHECK_GT(thr->ignore_reads_and_writes, 0);
   thr->fast_state.SetIgnoreBit();
 #if !SANITIZER_GO
-  if (save_stack && !ctx->after_multithreaded_fork)
+  if (pc && !ctx->after_multithreaded_fork)
     thr->mop_ignore_set.Add(CurrentStackId(thr, pc));
 #endif
 }
 
-void ThreadIgnoreEnd(ThreadState *thr, uptr pc) {
+void ThreadIgnoreEnd(ThreadState *thr) {
   DPrintf("#%d: ThreadIgnoreEnd\n", thr->tid);
   CHECK_GT(thr->ignore_reads_and_writes, 0);
   thr->ignore_reads_and_writes--;
@@ -1100,17 +709,17 @@ uptr __tsan_testonly_shadow_stack_current_size() {
 }
 #endif
 
-void ThreadIgnoreSyncBegin(ThreadState *thr, uptr pc, bool save_stack) {
+void ThreadIgnoreSyncBegin(ThreadState *thr, uptr pc) {
   DPrintf("#%d: ThreadIgnoreSyncBegin\n", thr->tid);
   thr->ignore_sync++;
   CHECK_GT(thr->ignore_sync, 0);
 #if !SANITIZER_GO
-  if (save_stack && !ctx->after_multithreaded_fork)
+  if (pc && !ctx->after_multithreaded_fork)
     thr->sync_ignore_set.Add(CurrentStackId(thr, pc));
 #endif
 }
 
-void ThreadIgnoreSyncEnd(ThreadState *thr, uptr pc) {
+void ThreadIgnoreSyncEnd(ThreadState *thr) {
   DPrintf("#%d: ThreadIgnoreSyncEnd\n", thr->tid);
   CHECK_GT(thr->ignore_sync, 0);
   thr->ignore_sync--;
@@ -1152,8 +761,3 @@ MutexMeta mutex_meta[] = {
 void PrintMutexPC(uptr pc) { StackTrace(&pc, 1).Print(); }
 }  // namespace __sanitizer
 #endif
-
-#if !SANITIZER_GO
-// Must be included in this file to make sure everything is inlined.
-#  include "tsan_interface_inl.h"
-#endif
diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl.h b/compiler-rt/lib/tsan/rtl/tsan_rtl.h
index 8567d0ade877..c71b27e1cbf5 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_rtl.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_rtl.h
@@ -37,14 +37,15 @@
 #include "tsan_clock.h"
 #include "tsan_defs.h"
 #include "tsan_flags.h"
+#include "tsan_ignoreset.h"
 #include "tsan_mman.h"
-#include "tsan_sync.h"
-#include "tsan_trace.h"
-#include "tsan_report.h"
-#include "tsan_platform.h"
 #include "tsan_mutexset.h"
-#include "tsan_ignoreset.h"
+#include "tsan_platform.h"
+#include "tsan_report.h"
+#include "tsan_shadow.h"
 #include "tsan_stack_trace.h"
+#include "tsan_sync.h"
+#include "tsan_trace.h"
 
 #if SANITIZER_WORDSIZE != 64
 # error "ThreadSanitizer is supported only on 64-bit platforms"
@@ -69,6 +70,11 @@ struct AP32 {
 typedef SizeClassAllocator32<AP32> PrimaryAllocator;
 #else
 struct AP64 {  // Allocator64 parameters. Deliberately using a short name.
+#    if defined(__s390x__)
+  typedef MappingS390x Mapping;
+#    else
+  typedef Mapping48AddressSpace Mapping;
+#    endif
   static const uptr kSpaceBeg = Mapping::kHeapMemBeg;
   static const uptr kSpaceSize = Mapping::kHeapMemEnd - Mapping::kHeapMemBeg;
   static const uptr kMetadataSize = 0;
@@ -84,240 +90,6 @@ typedef Allocator::AllocatorCache AllocatorCache;
 Allocator *allocator();
 #endif
 
-const u64 kShadowRodata = (u64)-1;  // .rodata shadow marker
-
-// FastState (from most significant bit):
-//   ignore          : 1
-//   tid             : kTidBits
-//   unused          : -
-//   history_size    : 3
-//   epoch           : kClkBits
-class FastState {
- public:
-  FastState(u64 tid, u64 epoch) {
-    x_ = tid << kTidShift;
-    x_ |= epoch;
-    DCHECK_EQ(tid, this->tid());
-    DCHECK_EQ(epoch, this->epoch());
-    DCHECK_EQ(GetIgnoreBit(), false);
-  }
-
-  explicit FastState(u64 x)
-      : x_(x) {
-  }
-
-  u64 raw() const {
-    return x_;
-  }
-
-  u64 tid() const {
-    u64 res = (x_ & ~kIgnoreBit) >> kTidShift;
-    return res;
-  }
-
-  u64 TidWithIgnore() const {
-    u64 res = x_ >> kTidShift;
-    return res;
-  }
-
-  u64 epoch() const {
-    u64 res = x_ & ((1ull << kClkBits) - 1);
-    return res;
-  }
-
-  void IncrementEpoch() {
-    u64 old_epoch = epoch();
-    x_ += 1;
-    DCHECK_EQ(old_epoch + 1, epoch());
-    (void)old_epoch;
-  }
-
-  void SetIgnoreBit() { x_ |= kIgnoreBit; }
-  void ClearIgnoreBit() { x_ &= ~kIgnoreBit; }
-  bool GetIgnoreBit() const { return (s64)x_ < 0; }
-
-  void SetHistorySize(int hs) {
-    CHECK_GE(hs, 0);
-    CHECK_LE(hs, 7);
-    x_ = (x_ & ~(kHistoryMask << kHistoryShift)) | (u64(hs) << kHistoryShift);
-  }
-
-  ALWAYS_INLINE
-  int GetHistorySize() const {
-    return (int)((x_ >> kHistoryShift) & kHistoryMask);
-  }
-
-  void ClearHistorySize() {
-    SetHistorySize(0);
-  }
-
-  ALWAYS_INLINE
-  u64 GetTracePos() const {
-    const int hs = GetHistorySize();
-    // When hs == 0, the trace consists of 2 parts.
-    const u64 mask = (1ull << (kTracePartSizeBits + hs + 1)) - 1;
-    return epoch() & mask;
-  }
-
- private:
-  friend class Shadow;
-  static const int kTidShift = 64 - kTidBits - 1;
-  static const u64 kIgnoreBit = 1ull << 63;
-  static const u64 kFreedBit = 1ull << 63;
-  static const u64 kHistoryShift = kClkBits;
-  static const u64 kHistoryMask = 7;
-  u64 x_;
-};
-
-// Shadow (from most significant bit):
-//   freed           : 1
-//   tid             : kTidBits
-//   is_atomic       : 1
-//   is_read         : 1
-//   size_log        : 2
-//   addr0           : 3
-//   epoch           : kClkBits
-class Shadow : public FastState {
- public:
-  explicit Shadow(u64 x)
-      : FastState(x) {
-  }
-
-  explicit Shadow(const FastState &s)
-      : FastState(s.x_) {
-    ClearHistorySize();
-  }
-
-  void SetAddr0AndSizeLog(u64 addr0, unsigned kAccessSizeLog) {
-    DCHECK_EQ((x_ >> kClkBits) & 31, 0);
-    DCHECK_LE(addr0, 7);
-    DCHECK_LE(kAccessSizeLog, 3);
-    x_ |= ((kAccessSizeLog << 3) | addr0) << kClkBits;
-    DCHECK_EQ(kAccessSizeLog, size_log());
-    DCHECK_EQ(addr0, this->addr0());
-  }
-
-  void SetWrite(unsigned kAccessIsWrite) {
-    DCHECK_EQ(x_ & kReadBit, 0);
-    if (!kAccessIsWrite)
-      x_ |= kReadBit;
-    DCHECK_EQ(kAccessIsWrite, IsWrite());
-  }
-
-  void SetAtomic(bool kIsAtomic) {
-    DCHECK(!IsAtomic());
-    if (kIsAtomic)
-      x_ |= kAtomicBit;
-    DCHECK_EQ(IsAtomic(), kIsAtomic);
-  }
-
-  bool IsAtomic() const {
-    return x_ & kAtomicBit;
-  }
-
-  bool IsZero() const {
-    return x_ == 0;
-  }
-
-  static inline bool TidsAreEqual(const Shadow s1, const Shadow s2) {
-    u64 shifted_xor = (s1.x_ ^ s2.x_) >> kTidShift;
-    DCHECK_EQ(shifted_xor == 0, s1.TidWithIgnore() == s2.TidWithIgnore());
-    return shifted_xor == 0;
-  }
-
-  static ALWAYS_INLINE
-  bool Addr0AndSizeAreEqual(const Shadow s1, const Shadow s2) {
-    u64 masked_xor = ((s1.x_ ^ s2.x_) >> kClkBits) & 31;
-    return masked_xor == 0;
-  }
-
-  static ALWAYS_INLINE bool TwoRangesIntersect(Shadow s1, Shadow s2,
-      unsigned kS2AccessSize) {
-    bool res = false;
-    u64 diff = s1.addr0() - s2.addr0();
-    if ((s64)diff < 0) {  // s1.addr0 < s2.addr0
-      // if (s1.addr0() + size1) > s2.addr0()) return true;
-      if (s1.size() > -diff)
-        res = true;
-    } else {
-      // if (s2.addr0() + kS2AccessSize > s1.addr0()) return true;
-      if (kS2AccessSize > diff)
-        res = true;
-    }
-    DCHECK_EQ(res, TwoRangesIntersectSlow(s1, s2));
-    DCHECK_EQ(res, TwoRangesIntersectSlow(s2, s1));
-    return res;
-  }
-
-  u64 ALWAYS_INLINE addr0() const { return (x_ >> kClkBits) & 7; }
-  u64 ALWAYS_INLINE size() const { return 1ull << size_log(); }
-  bool ALWAYS_INLINE IsWrite() const { return !IsRead(); }
-  bool ALWAYS_INLINE IsRead() const { return x_ & kReadBit; }
-
-  // The idea behind the freed bit is as follows.
-  // When the memory is freed (or otherwise unaccessible) we write to the shadow
-  // values with tid/epoch related to the free and the freed bit set.
-  // During memory accesses processing the freed bit is considered
-  // as msb of tid. So any access races with shadow with freed bit set
-  // (it is as if write from a thread with which we never synchronized before).
-  // This allows us to detect accesses to freed memory w/o additional
-  // overheads in memory access processing and at the same time restore
-  // tid/epoch of free.
-  void MarkAsFreed() {
-     x_ |= kFreedBit;
-  }
-
-  bool IsFreed() const {
-    return x_ & kFreedBit;
-  }
-
-  bool GetFreedAndReset() {
-    bool res = x_ & kFreedBit;
-    x_ &= ~kFreedBit;
-    return res;
-  }
-
-  bool ALWAYS_INLINE IsBothReadsOrAtomic(bool kIsWrite, bool kIsAtomic) const {
-    bool v = x_ & ((u64(kIsWrite ^ 1) << kReadShift)
-        | (u64(kIsAtomic) << kAtomicShift));
-    DCHECK_EQ(v, (!IsWrite() && !kIsWrite) || (IsAtomic() && kIsAtomic));
-    return v;
-  }
-
-  bool ALWAYS_INLINE IsRWNotWeaker(bool kIsWrite, bool kIsAtomic) const {
-    bool v = ((x_ >> kReadShift) & 3)
-        <= u64((kIsWrite ^ 1) | (kIsAtomic << 1));
-    DCHECK_EQ(v, (IsAtomic() < kIsAtomic) ||
-        (IsAtomic() == kIsAtomic && !IsWrite() <= !kIsWrite));
-    return v;
-  }
-
-  bool ALWAYS_INLINE IsRWWeakerOrEqual(bool kIsWrite, bool kIsAtomic) const {
-    bool v = ((x_ >> kReadShift) & 3)
-        >= u64((kIsWrite ^ 1) | (kIsAtomic << 1));
-    DCHECK_EQ(v, (IsAtomic() > kIsAtomic) ||
-        (IsAtomic() == kIsAtomic && !IsWrite() >= !kIsWrite));
-    return v;
-  }
-
- private:
-  static const u64 kReadShift   = 5 + kClkBits;
-  static const u64 kReadBit     = 1ull << kReadShift;
-  static const u64 kAtomicShift = 6 + kClkBits;
-  static const u64 kAtomicBit   = 1ull << kAtomicShift;
-
-  u64 size_log() const { return (x_ >> (3 + kClkBits)) & 3; }
-
-  static bool TwoRangesIntersectSlow(const Shadow s1, const Shadow s2) {
-    if (s1.addr0() == s2.addr0()) return true;
-    if (s1.addr0() < s2.addr0() && s1.addr0() + s1.size() > s2.addr0())
-      return true;
-    if (s2.addr0() < s1.addr0() && s2.addr0() + s2.size() > s1.addr0())
-      return true;
-    return false;
-  }
-};
-
 struct ThreadSignalContext;
 
 struct JmpBuf {
@@ -380,6 +152,7 @@ struct ThreadState {
   // We do not distinguish beteween ignoring reads and writes
   // for better performance.
   int ignore_reads_and_writes;
+  atomic_sint32_t pending_signals;
   int ignore_sync;
   int suppress_reports;
   // Go does not support ignores.
@@ -387,20 +160,18 @@ struct ThreadState {
   IgnoreSet mop_ignore_set;
   IgnoreSet sync_ignore_set;
 #endif
-  // C/C++ uses fixed size shadow stack embed into Trace.
-  // Go uses malloc-allocated shadow stack with dynamic size.
   uptr *shadow_stack;
   uptr *shadow_stack_end;
   uptr *shadow_stack_pos;
-  u64 *racy_shadow_addr;
-  u64 racy_state[2];
+  RawShadow *racy_shadow_addr;
+  RawShadow racy_state[2];
   MutexSet mset;
   ThreadClock clock;
 #if !SANITIZER_GO
   Vector<JmpBuf> jmp_bufs;
   int ignore_interceptors;
 #endif
-  const u32 tid;
+  const Tid tid;
   const int unique_id;
   bool in_symbolizer;
   bool in_ignored_lib;
@@ -428,7 +199,7 @@ struct ThreadState {
   ThreadSignalContext *signal_ctx;
 
 #if !SANITIZER_GO
-  u32 last_sleep_stack_id;
+  StackID last_sleep_stack_id;
   ThreadClock last_sleep_clock;
 #endif
 
@@ -438,41 +209,49 @@ struct ThreadState {
 
   const ReportDesc *current_report;
 
-  explicit ThreadState(Context *ctx, u32 tid, int unique_id, u64 epoch,
+  // Current position in tctx->trace.Back()->events (Event*).
+  atomic_uintptr_t trace_pos;
+  // PC of the last memory access, used to compute PC deltas in the trace.
+  uptr trace_prev_pc;
+  Sid sid;
+  Epoch epoch;
+
+  explicit ThreadState(Context *ctx, Tid tid, int unique_id, u64 epoch,
                        unsigned reuse_count, uptr stk_addr, uptr stk_size,
                        uptr tls_addr, uptr tls_size);
-};
+} ALIGNED(SANITIZER_CACHE_LINE_SIZE);
 
 #if !SANITIZER_GO
 #if SANITIZER_MAC || SANITIZER_ANDROID
 ThreadState *cur_thread();
 void set_cur_thread(ThreadState *thr);
 void cur_thread_finalize();
-inline void cur_thread_init() { }
-#else
+inline ThreadState *cur_thread_init() { return cur_thread(); }
+#  else
 __attribute__((tls_model("initial-exec")))
 extern THREADLOCAL char cur_thread_placeholder[];
 inline ThreadState *cur_thread() {
   return reinterpret_cast<ThreadState *>(cur_thread_placeholder)->current;
 }
-inline void cur_thread_init() {
+inline ThreadState *cur_thread_init() {
   ThreadState *thr = reinterpret_cast<ThreadState *>(cur_thread_placeholder);
   if (UNLIKELY(!thr->current))
     thr->current = thr;
+  return thr->current;
 }
 inline void set_cur_thread(ThreadState *thr) {
   reinterpret_cast<ThreadState *>(cur_thread_placeholder)->current = thr;
 }
 inline void cur_thread_finalize() { }
-#endif  // SANITIZER_MAC || SANITIZER_ANDROID
+#  endif  // SANITIZER_MAC || SANITIZER_ANDROID
 #endif  // SANITIZER_GO
 
 class ThreadContext final : public ThreadContextBase {
  public:
-  explicit ThreadContext(int tid);
+  explicit ThreadContext(Tid tid);
   ~ThreadContext();
   ThreadState *thr;
-  u32 creation_stack_id;
+  StackID creation_stack_id;
   SyncClock sync;
   // Epoch at which the thread had started.
   // If we see an event from the thread stamped by an older epoch,
@@ -480,6 +259,8 @@ class ThreadContext final : public ThreadContextBase {
   u64 epoch0;
   u64 epoch1;
 
+  v3::Trace trace;
+
   // Override superclass callbacks.
   void OnDead() override;
   void OnJoined(void *arg) override;
@@ -492,13 +273,7 @@ class ThreadContext final : public ThreadContextBase {
 
 struct RacyStacks {
   MD5Hash hash[2];
-  bool operator==(const RacyStacks &other) const {
-    if (hash[0] == other.hash[0] && hash[1] == other.hash[1])
-      return true;
-    if (hash[0] == other.hash[1] && hash[1] == other.hash[0])
-      return true;
-    return false;
-  }
+  bool operator==(const RacyStacks &other) const;
 };
 
 struct RacyAddress {
@@ -524,13 +299,12 @@ struct Context {
 
   Mutex report_mtx;
   int nreported;
-  int nmissed_expected;
   atomic_uint64_t last_symbolize_time_ns;
 
   void *background_thread;
   atomic_uint32_t stop_background_thread;
 
-  ThreadRegistry *thread_registry;
+  ThreadRegistry thread_registry;
 
   Mutex racy_mtx;
   Vector<RacyStacks> racy_stacks;
@@ -543,9 +317,9 @@ struct Context {
   ClockAlloc clock_alloc;
 
   Flags flags;
+  fd_t memprof_fd;
 
-  u64 int_alloc_cnt[MBlockTypeCount];
-  u64 int_alloc_siz[MBlockTypeCount];
+  Mutex slot_mtx;
 };
 
 extern Context *ctx;  // The one and the only global runtime context.
@@ -578,12 +352,12 @@ class ScopedReportBase {
                        const MutexSet *mset);
   void AddStack(StackTrace stack, bool suppressable = false);
   void AddThread(const ThreadContext *tctx, bool suppressable = false);
-  void AddThread(int unique_tid, bool suppressable = false);
-  void AddUniqueTid(int unique_tid);
+  void AddThread(Tid unique_tid, bool suppressable = false);
+  void AddUniqueTid(Tid unique_tid);
   void AddMutex(const SyncVar *s);
   u64 AddMutex(u64 id);
   void AddLocation(uptr addr, uptr size);
-  void AddSleep(u32 stack_id);
+  void AddSleep(StackID stack_id);
   void SetCount(int count);
 
   const ReportDesc *GetReport() const;
@@ -615,7 +389,7 @@ class ScopedReport : public ScopedReportBase {
 
 bool ShouldReport(ThreadState *thr, ReportType typ);
 ThreadContext *IsThreadStackOrTls(uptr addr, bool *is_stack);
-void RestoreStack(int tid, const u64 epoch, VarSizeStackTrace *stk,
+void RestoreStack(Tid tid, const u64 epoch, VarSizeStackTrace *stk,
                   MutexSet *mset, uptr *tag = nullptr);
 
 // The stack could look like:
@@ -662,13 +436,12 @@ void InitializeDynamicAnnotations();
 
 void ForkBefore(ThreadState *thr, uptr pc);
 void ForkParentAfter(ThreadState *thr, uptr pc);
-void ForkChildAfter(ThreadState *thr, uptr pc);
+void ForkChildAfter(ThreadState *thr, uptr pc, bool start_thread);
 
 void ReportRace(ThreadState *thr);
 bool OutputReport(ThreadState *thr, const ScopedReport &srep);
 bool IsFiredSuppression(Context *ctx, ReportType type, StackTrace trace);
 bool IsExpectedReport(uptr addr, uptr size);
-void PrintMatchedBenignRaces();
 
 #if defined(TSAN_DEBUG_OUTPUT) && TSAN_DEBUG_OUTPUT >= 1
 # define DPrintf Printf
@@ -682,10 +455,11 @@ void PrintMatchedBenignRaces();
 # define DPrintf2(...)
 #endif
 
-u32 CurrentStackId(ThreadState *thr, uptr pc);
-ReportStack *SymbolizeStackId(u32 stack_id);
+StackID CurrentStackId(ThreadState *thr, uptr pc);
+ReportStack *SymbolizeStackId(StackID stack_id);
 void PrintCurrentStack(ThreadState *thr, uptr pc);
 void PrintCurrentStackSlow(uptr pc);  // uses libunwind
+MBlock *JavaHeapBlock(uptr addr, uptr *start);
 
 void Initialize(ThreadState *thr);
 void MaybeSpawnBackgroundThread();
@@ -701,34 +475,44 @@ void MemoryAccessImpl(ThreadState *thr, uptr addr,
     u64 *shadow_mem, Shadow cur);
 void MemoryAccessRange(ThreadState *thr, uptr pc, uptr addr,
     uptr size, bool is_write);
-void MemoryAccessRangeStep(ThreadState *thr, uptr pc, uptr addr,
-    uptr size, uptr step, bool is_write);
-void UnalignedMemoryAccess(ThreadState *thr, uptr pc, uptr addr,
-    int size, bool kAccessIsWrite, bool kIsAtomic);
+void UnalignedMemoryAccess(ThreadState *thr, uptr pc, uptr addr, uptr size,
+                           AccessType typ);
 
 const int kSizeLog1 = 0;
 const int kSizeLog2 = 1;
 const int kSizeLog4 = 2;
 const int kSizeLog8 = 3;
 
-void ALWAYS_INLINE MemoryRead(ThreadState *thr, uptr pc,
-                                     uptr addr, int kAccessSizeLog) {
-  MemoryAccess(thr, pc, addr, kAccessSizeLog, false, false);
-}
-
-void ALWAYS_INLINE MemoryWrite(ThreadState *thr, uptr pc,
-                                      uptr addr, int kAccessSizeLog) {
-  MemoryAccess(thr, pc, addr, kAccessSizeLog, true, false);
-}
-
-void ALWAYS_INLINE MemoryReadAtomic(ThreadState *thr, uptr pc,
-                                           uptr addr, int kAccessSizeLog) {
-  MemoryAccess(thr, pc, addr, kAccessSizeLog, false, true);
-}
-
-void ALWAYS_INLINE MemoryWriteAtomic(ThreadState *thr, uptr pc,
-                                            uptr addr, int kAccessSizeLog) {
-  MemoryAccess(thr, pc, addr, kAccessSizeLog, true, true);
+ALWAYS_INLINE
+void MemoryAccess(ThreadState *thr, uptr pc, uptr addr, uptr size,
+                  AccessType typ) {
+  int size_log;
+  switch (size) {
+    case 1:
+      size_log = kSizeLog1;
+      break;
+    case 2:
+      size_log = kSizeLog2;
+      break;
+    case 4:
+      size_log = kSizeLog4;
+      break;
+    default:
+      DCHECK_EQ(size, 8);
+      size_log = kSizeLog8;
+      break;
+  }
+  bool is_write = !(typ & kAccessRead);
+  bool is_atomic = typ & kAccessAtomic;
+  if (typ & kAccessVptr)
+    thr->is_vptr_access = true;
+  if (typ & kAccessFree)
+    thr->is_freeing = true;
+  MemoryAccess(thr, pc, addr, size_log, is_write, is_atomic);
+  if (typ & kAccessVptr)
+    thr->is_vptr_access = false;
+  if (typ & kAccessFree)
+    thr->is_freeing = false;
 }
 
 void MemoryResetRange(ThreadState *thr, uptr pc, uptr addr, uptr size);
@@ -737,26 +521,26 @@ void MemoryRangeImitateWrite(ThreadState *thr, uptr pc, uptr addr, uptr size);
 void MemoryRangeImitateWriteOrResetRange(ThreadState *thr, uptr pc, uptr addr,
                                          uptr size);
 
-void ThreadIgnoreBegin(ThreadState *thr, uptr pc, bool save_stack = true);
-void ThreadIgnoreEnd(ThreadState *thr, uptr pc);
-void ThreadIgnoreSyncBegin(ThreadState *thr, uptr pc, bool save_stack = true);
-void ThreadIgnoreSyncEnd(ThreadState *thr, uptr pc);
+void ThreadIgnoreBegin(ThreadState *thr, uptr pc);
+void ThreadIgnoreEnd(ThreadState *thr);
+void ThreadIgnoreSyncBegin(ThreadState *thr, uptr pc);
+void ThreadIgnoreSyncEnd(ThreadState *thr);
 
 void FuncEntry(ThreadState *thr, uptr pc);
 void FuncExit(ThreadState *thr);
 
-int ThreadCreate(ThreadState *thr, uptr pc, uptr uid, bool detached);
-void ThreadStart(ThreadState *thr, int tid, tid_t os_id,
+Tid ThreadCreate(ThreadState *thr, uptr pc, uptr uid, bool detached);
+void ThreadStart(ThreadState *thr, Tid tid, tid_t os_id,
                  ThreadType thread_type);
 void ThreadFinish(ThreadState *thr);
-int ThreadConsumeTid(ThreadState *thr, uptr pc, uptr uid);
-void ThreadJoin(ThreadState *thr, uptr pc, int tid);
-void ThreadDetach(ThreadState *thr, uptr pc, int tid);
+Tid ThreadConsumeTid(ThreadState *thr, uptr pc, uptr uid);
+void ThreadJoin(ThreadState *thr, uptr pc, Tid tid);
+void ThreadDetach(ThreadState *thr, uptr pc, Tid tid);
 void ThreadFinalize(ThreadState *thr);
 void ThreadSetName(ThreadState *thr, const char *name);
 int ThreadCount(ThreadState *thr);
-void ProcessPendingSignals(ThreadState *thr);
-void ThreadNotJoined(ThreadState *thr, uptr pc, int tid, uptr uid);
+void ProcessPendingSignalsImpl(ThreadState *thr);
+void ThreadNotJoined(ThreadState *thr, uptr pc, Tid tid, uptr uid);
 
 Processor *ProcCreate();
 void ProcDestroy(Processor *proc);
@@ -785,7 +569,7 @@ void Acquire(ThreadState *thr, uptr pc, uptr addr);
 // handle Go finalizers. Namely, finalizer goroutine executes AcquireGlobal
 // right before executing finalizers. This provides a coarse, but simple
 // approximation of the actual required synchronization.
-void AcquireGlobal(ThreadState *thr, uptr pc);
+void AcquireGlobal(ThreadState *thr);
 void Release(ThreadState *thr, uptr pc, uptr addr);
 void ReleaseStoreAcquire(ThreadState *thr, uptr pc, uptr addr);
 void ReleaseStore(ThreadState *thr, uptr pc, uptr addr);
@@ -821,13 +605,16 @@ void TraceSwitch(ThreadState *thr);
 uptr TraceTopPC(ThreadState *thr);
 uptr TraceSize();
 uptr TraceParts();
-Trace *ThreadTrace(int tid);
+Trace *ThreadTrace(Tid tid);
 
 extern "C" void __tsan_trace_switch();
 void ALWAYS_INLINE TraceAddEvent(ThreadState *thr, FastState fs,
                                         EventType typ, u64 addr) {
   if (!kCollectHistory)
     return;
+  // TraceSwitch accesses shadow_stack, but it's called infrequently,
+  // so we check it here proactively.
+  DCHECK(thr->shadow_stack);
   DCHECK_GE((int)typ, 0);
   DCHECK_LE((int)typ, 7);
   DCHECK_EQ(GetLsb(addr, kEventPCBits), addr);
@@ -861,6 +648,149 @@ enum FiberSwitchFlags {
   FiberSwitchFlagNoSync = 1 << 0, // __tsan_switch_to_fiber_no_sync
 };
 
+ALWAYS_INLINE void ProcessPendingSignals(ThreadState *thr) {
+  if (UNLIKELY(atomic_load_relaxed(&thr->pending_signals)))
+    ProcessPendingSignalsImpl(thr);
+}
+
+extern bool is_initialized;
+
+ALWAYS_INLINE
+void LazyInitialize(ThreadState *thr) {
+  // If we can use .preinit_array, assume that __tsan_init
+  // called from .preinit_array initializes runtime before
+  // any instrumented code.
+#if !SANITIZER_CAN_USE_PREINIT_ARRAY
+  if (UNLIKELY(!is_initialized))
+    Initialize(thr);
+#endif
+}
+
+namespace v3 {
+
+void TraceSwitchPart(ThreadState *thr);
+bool RestoreStack(Tid tid, EventType type, Sid sid, Epoch epoch, uptr addr,
+                  uptr size, AccessType typ, VarSizeStackTrace *pstk,
+                  MutexSet *pmset, uptr *ptag);
+
+template <typename EventT>
+ALWAYS_INLINE WARN_UNUSED_RESULT bool TraceAcquire(ThreadState *thr,
+                                                   EventT **ev) {
+  Event *pos = reinterpret_cast<Event *>(atomic_load_relaxed(&thr->trace_pos));
+#if SANITIZER_DEBUG
+  // TraceSwitch acquires these mutexes,
+  // so we lock them here to detect deadlocks more reliably.
+  { Lock lock(&ctx->slot_mtx); }
+  { Lock lock(&thr->tctx->trace.mtx); }
+  TracePart *current = thr->tctx->trace.parts.Back();
+  if (current) {
+    DCHECK_GE(pos, &current->events[0]);
+    DCHECK_LE(pos, &current->events[TracePart::kSize]);
+  } else {
+    DCHECK_EQ(pos, nullptr);
+  }
+#endif
+  // TracePart is allocated with mmap and is at least 4K aligned.
+  // So the following check is a faster way to check for part end.
+  // It may have false positives in the middle of the trace,
+  // they are filtered out in TraceSwitch.
+  if (UNLIKELY(((uptr)(pos + 1) & TracePart::kAlignment) == 0))
+    return false;
+  *ev = reinterpret_cast<EventT *>(pos);
+  return true;
+}
+
+template <typename EventT>
+ALWAYS_INLINE void TraceRelease(ThreadState *thr, EventT *evp) {
+  DCHECK_LE(evp + 1, &thr->tctx->trace.parts.Back()->events[TracePart::kSize]);
+  atomic_store_relaxed(&thr->trace_pos, (uptr)(evp + 1));
+}
+
+template <typename EventT>
+void TraceEvent(ThreadState *thr, EventT ev) {
+  EventT *evp;
+  if (!TraceAcquire(thr, &evp)) {
+    TraceSwitchPart(thr);
+    UNUSED bool res = TraceAcquire(thr, &evp);
+    DCHECK(res);
+  }
+  *evp = ev;
+  TraceRelease(thr, evp);
+}
+
+ALWAYS_INLINE WARN_UNUSED_RESULT bool TryTraceFunc(ThreadState *thr,
+                                                   uptr pc = 0) {
+  if (!kCollectHistory)
+    return true;
+  EventFunc *ev;
+  if (UNLIKELY(!TraceAcquire(thr, &ev)))
+    return false;
+  ev->is_access = 0;
+  ev->is_func = 1;
+  ev->pc = pc;
+  TraceRelease(thr, ev);
+  return true;
+}
+
+WARN_UNUSED_RESULT
+bool TryTraceMemoryAccess(ThreadState *thr, uptr pc, uptr addr, uptr size,
+                          AccessType typ);
+WARN_UNUSED_RESULT
+bool TryTraceMemoryAccessRange(ThreadState *thr, uptr pc, uptr addr, uptr size,
+                               AccessType typ);
+void TraceMemoryAccessRange(ThreadState *thr, uptr pc, uptr addr, uptr size,
+                            AccessType typ);
+void TraceFunc(ThreadState *thr, uptr pc = 0);
+void TraceMutexLock(ThreadState *thr, EventType type, uptr pc, uptr addr,
+                    StackID stk);
+void TraceMutexUnlock(ThreadState *thr, uptr addr);
+void TraceTime(ThreadState *thr);
+
+}  // namespace v3
+
+void GrowShadowStack(ThreadState *thr);
+
+ALWAYS_INLINE
+void FuncEntry(ThreadState *thr, uptr pc) {
+  DPrintf2("#%d: FuncEntry %p\n", (int)thr->fast_state.tid(), (void *)pc);
+  if (kCollectHistory) {
+    thr->fast_state.IncrementEpoch();
+    TraceAddEvent(thr, thr->fast_state, EventTypeFuncEnter, pc);
+  }
+
+  // Shadow stack maintenance can be replaced with
+  // stack unwinding during trace switch (which presumably must be faster).
+  DCHECK_GE(thr->shadow_stack_pos, thr->shadow_stack);
+#if !SANITIZER_GO
+  DCHECK_LT(thr->shadow_stack_pos, thr->shadow_stack_end);
+#else
+  if (thr->shadow_stack_pos == thr->shadow_stack_end)
+    GrowShadowStack(thr);
+#endif
+  thr->shadow_stack_pos[0] = pc;
+  thr->shadow_stack_pos++;
+}
+
+ALWAYS_INLINE
+void FuncExit(ThreadState *thr) {
+  DPrintf2("#%d: FuncExit\n", (int)thr->fast_state.tid());
+  if (kCollectHistory) {
+    thr->fast_state.IncrementEpoch();
+    TraceAddEvent(thr, thr->fast_state, EventTypeFuncExit, 0);
+  }
+
+  DCHECK_GT(thr->shadow_stack_pos, thr->shadow_stack);
+#if !SANITIZER_GO
+  DCHECK_LT(thr->shadow_stack_pos, thr->shadow_stack_end);
+#endif
+  thr->shadow_stack_pos--;
+}
+
+#if !SANITIZER_GO
+extern void (*on_initialize)(void);
+extern int (*on_finalize)(int);
+#endif
+
 }  // namespace __tsan
 
 #endif  // TSAN_RTL_H
diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl_access.cpp b/compiler-rt/lib/tsan/rtl/tsan_rtl_access.cpp
new file mode 100644
index 000000000000..7365fdaa3038
--- /dev/null
+++ b/compiler-rt/lib/tsan/rtl/tsan_rtl_access.cpp
@@ -0,0 +1,604 @@
+//===-- tsan_rtl_access.cpp -----------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of ThreadSanitizer (TSan), a race detector.
+//
+// Definitions of memory access and function entry/exit entry points.
+//===----------------------------------------------------------------------===//
+
+#include "tsan_rtl.h"
+
+namespace __tsan {
+
+namespace v3 {
+
+ALWAYS_INLINE USED bool TryTraceMemoryAccess(ThreadState *thr, uptr pc,
+                                             uptr addr, uptr size,
+                                             AccessType typ) {
+  DCHECK(size == 1 || size == 2 || size == 4 || size == 8);
+  if (!kCollectHistory)
+    return true;
+  EventAccess *ev;
+  if (UNLIKELY(!TraceAcquire(thr, &ev)))
+    return false;
+  u64 size_log = size == 1 ? 0 : size == 2 ? 1 : size == 4 ? 2 : 3;
+  uptr pc_delta = pc - thr->trace_prev_pc + (1 << (EventAccess::kPCBits - 1));
+  thr->trace_prev_pc = pc;
+  if (LIKELY(pc_delta < (1 << EventAccess::kPCBits))) {
+    ev->is_access = 1;
+    ev->is_read = !!(typ & kAccessRead);
+    ev->is_atomic = !!(typ & kAccessAtomic);
+    ev->size_log = size_log;
+    ev->pc_delta = pc_delta;
+    DCHECK_EQ(ev->pc_delta, pc_delta);
+    ev->addr = CompressAddr(addr);
+    TraceRelease(thr, ev);
+    return true;
+  }
+  auto *evex = reinterpret_cast<EventAccessExt *>(ev);
+  evex->is_access = 0;
+  evex->is_func = 0;
+  evex->type = EventType::kAccessExt;
+  evex->is_read = !!(typ & kAccessRead);
+  evex->is_atomic = !!(typ & kAccessAtomic);
+  evex->size_log = size_log;
+  evex->addr = CompressAddr(addr);
+  evex->pc = pc;
+  TraceRelease(thr, evex);
+  return true;
+}
+
+ALWAYS_INLINE USED bool TryTraceMemoryAccessRange(ThreadState *thr, uptr pc,
+                                                  uptr addr, uptr size,
+                                                  AccessType typ) {
+  if (!kCollectHistory)
+    return true;
+  EventAccessRange *ev;
+  if (UNLIKELY(!TraceAcquire(thr, &ev)))
+    return false;
+  thr->trace_prev_pc = pc;
+  ev->is_access = 0;
+  ev->is_func = 0;
+  ev->type = EventType::kAccessRange;
+  ev->is_read = !!(typ & kAccessRead);
+  ev->is_free = !!(typ & kAccessFree);
+  ev->size_lo = size;
+  ev->pc = CompressAddr(pc);
+  ev->addr = CompressAddr(addr);
+  ev->size_hi = size >> EventAccessRange::kSizeLoBits;
+  TraceRelease(thr, ev);
+  return true;
+}
+
+void TraceMemoryAccessRange(ThreadState *thr, uptr pc, uptr addr, uptr size,
+                            AccessType typ) {
+  if (LIKELY(TryTraceMemoryAccessRange(thr, pc, addr, size, typ)))
+    return;
+  TraceSwitchPart(thr);
+  UNUSED bool res = TryTraceMemoryAccessRange(thr, pc, addr, size, typ);
+  DCHECK(res);
+}
+
+void TraceFunc(ThreadState *thr, uptr pc) {
+  if (LIKELY(TryTraceFunc(thr, pc)))
+    return;
+  TraceSwitchPart(thr);
+  UNUSED bool res = TryTraceFunc(thr, pc);
+  DCHECK(res);
+}
+
+void TraceMutexLock(ThreadState *thr, EventType type, uptr pc, uptr addr,
+                    StackID stk) {
+  DCHECK(type == EventType::kLock || type == EventType::kRLock);
+  if (!kCollectHistory)
+    return;
+  EventLock ev;
+  ev.is_access = 0;
+  ev.is_func = 0;
+  ev.type = type;
+  ev.pc = CompressAddr(pc);
+  ev.stack_lo = stk;
+  ev.stack_hi = stk >> EventLock::kStackIDLoBits;
+  ev._ = 0;
+  ev.addr = CompressAddr(addr);
+  TraceEvent(thr, ev);
+}
+
+void TraceMutexUnlock(ThreadState *thr, uptr addr) {
+  if (!kCollectHistory)
+    return;
+  EventUnlock ev;
+  ev.is_access = 0;
+  ev.is_func = 0;
+  ev.type = EventType::kUnlock;
+  ev._ = 0;
+  ev.addr = CompressAddr(addr);
+  TraceEvent(thr, ev);
+}
+
+void TraceTime(ThreadState *thr) {
+  if (!kCollectHistory)
+    return;
+  EventTime ev;
+  ev.is_access = 0;
+  ev.is_func = 0;
+  ev.type = EventType::kTime;
+  ev.sid = static_cast<u64>(thr->sid);
+  ev.epoch = static_cast<u64>(thr->epoch);
+  ev._ = 0;
+  TraceEvent(thr, ev);
+}
+
+}  // namespace v3
+
+ALWAYS_INLINE
+Shadow LoadShadow(u64 *p) {
+  u64 raw = atomic_load((atomic_uint64_t *)p, memory_order_relaxed);
+  return Shadow(raw);
+}
+
+ALWAYS_INLINE
+void StoreShadow(u64 *sp, u64 s) {
+  atomic_store((atomic_uint64_t *)sp, s, memory_order_relaxed);
+}
+
+ALWAYS_INLINE
+void StoreIfNotYetStored(u64 *sp, u64 *s) {
+  StoreShadow(sp, *s);
+  *s = 0;
+}
+
+extern "C" void __tsan_report_race();
+
+ALWAYS_INLINE
+void HandleRace(ThreadState *thr, u64 *shadow_mem, Shadow cur, Shadow old) {
+  thr->racy_state[0] = cur.raw();
+  thr->racy_state[1] = old.raw();
+  thr->racy_shadow_addr = shadow_mem;
+#if !SANITIZER_GO
+  HACKY_CALL(__tsan_report_race);
+#else
+  ReportRace(thr);
+#endif
+}
+
+static inline bool HappensBefore(Shadow old, ThreadState *thr) {
+  return thr->clock.get(old.TidWithIgnore()) >= old.epoch();
+}
+
+ALWAYS_INLINE
+void MemoryAccessImpl1(ThreadState *thr, uptr addr, int kAccessSizeLog,
+                       bool kAccessIsWrite, bool kIsAtomic, u64 *shadow_mem,
+                       Shadow cur) {
+  // This potentially can live in an MMX/SSE scratch register.
+  // The required intrinsics are:
+  // __m128i _mm_move_epi64(__m128i*);
+  // _mm_storel_epi64(u64*, __m128i);
+  u64 store_word = cur.raw();
+  bool stored = false;
+
+  // scan all the shadow values and dispatch to 4 categories:
+  // same, replace, candidate and race (see comments below).
+  // we consider only 3 cases regarding access sizes:
+  // equal, intersect and not intersect. initially I considered
+  // larger and smaller as well, it allowed to replace some
+  // 'candidates' with 'same' or 'replace', but I think
+  // it's just not worth it (performance- and complexity-wise).
+
+  Shadow old(0);
+
+  // It release mode we manually unroll the loop,
+  // because empirically gcc generates better code this way.
+  // However, we can't afford unrolling in debug mode, because the function
+  // consumes almost 4K of stack. Gtest gives only 4K of stack to death test
+  // threads, which is not enough for the unrolled loop.
+#if SANITIZER_DEBUG
+  for (int idx = 0; idx < 4; idx++) {
+#  include "tsan_update_shadow_word.inc"
+  }
+#else
+  int idx = 0;
+#  include "tsan_update_shadow_word.inc"
+  idx = 1;
+  if (stored) {
+#  include "tsan_update_shadow_word.inc"
+  } else {
+#  include "tsan_update_shadow_word.inc"
+  }
+  idx = 2;
+  if (stored) {
+#  include "tsan_update_shadow_word.inc"
+  } else {
+#  include "tsan_update_shadow_word.inc"
+  }
+  idx = 3;
+  if (stored) {
+#  include "tsan_update_shadow_word.inc"
+  } else {
+#  include "tsan_update_shadow_word.inc"
+  }
+#endif
+
+  // we did not find any races and had already stored
+  // the current access info, so we are done
+  if (LIKELY(stored))
+    return;
+  // choose a random candidate slot and replace it
+  StoreShadow(shadow_mem + (cur.epoch() % kShadowCnt), store_word);
+  return;
+RACE:
+  HandleRace(thr, shadow_mem, cur, old);
+  return;
+}
+
+void UnalignedMemoryAccess(ThreadState *thr, uptr pc, uptr addr, uptr size,
+                           AccessType typ) {
+  DCHECK(!(typ & kAccessAtomic));
+  const bool kAccessIsWrite = !(typ & kAccessRead);
+  const bool kIsAtomic = false;
+  while (size) {
+    int size1 = 1;
+    int kAccessSizeLog = kSizeLog1;
+    if (size >= 8 && (addr & ~7) == ((addr + 7) & ~7)) {
+      size1 = 8;
+      kAccessSizeLog = kSizeLog8;
+    } else if (size >= 4 && (addr & ~7) == ((addr + 3) & ~7)) {
+      size1 = 4;
+      kAccessSizeLog = kSizeLog4;
+    } else if (size >= 2 && (addr & ~7) == ((addr + 1) & ~7)) {
+      size1 = 2;
+      kAccessSizeLog = kSizeLog2;
+    }
+    MemoryAccess(thr, pc, addr, kAccessSizeLog, kAccessIsWrite, kIsAtomic);
+    addr += size1;
+    size -= size1;
+  }
+}
+
+ALWAYS_INLINE
+bool ContainsSameAccessSlow(u64 *s, u64 a, u64 sync_epoch, bool is_write) {
+  Shadow cur(a);
+  for (uptr i = 0; i < kShadowCnt; i++) {
+    Shadow old(LoadShadow(&s[i]));
+    if (Shadow::Addr0AndSizeAreEqual(cur, old) &&
+        old.TidWithIgnore() == cur.TidWithIgnore() &&
+        old.epoch() > sync_epoch && old.IsAtomic() == cur.IsAtomic() &&
+        old.IsRead() <= cur.IsRead())
+      return true;
+  }
+  return false;
+}
+
+#if TSAN_VECTORIZE
+#  define SHUF(v0, v1, i0, i1, i2, i3)                    \
+    _mm_castps_si128(_mm_shuffle_ps(_mm_castsi128_ps(v0), \
+                                    _mm_castsi128_ps(v1), \
+                                    (i0)*1 + (i1)*4 + (i2)*16 + (i3)*64))
+ALWAYS_INLINE
+bool ContainsSameAccessFast(u64 *s, u64 a, u64 sync_epoch, bool is_write) {
+  // This is an optimized version of ContainsSameAccessSlow.
+  // load current access into access[0:63]
+  const m128 access = _mm_cvtsi64_si128(a);
+  // duplicate high part of access in addr0:
+  // addr0[0:31]        = access[32:63]
+  // addr0[32:63]       = access[32:63]
+  // addr0[64:95]       = access[32:63]
+  // addr0[96:127]      = access[32:63]
+  const m128 addr0 = SHUF(access, access, 1, 1, 1, 1);
+  // load 4 shadow slots
+  const m128 shadow0 = _mm_load_si128((__m128i *)s);
+  const m128 shadow1 = _mm_load_si128((__m128i *)s + 1);
+  // load high parts of 4 shadow slots into addr_vect:
+  // addr_vect[0:31]    = shadow0[32:63]
+  // addr_vect[32:63]   = shadow0[96:127]
+  // addr_vect[64:95]   = shadow1[32:63]
+  // addr_vect[96:127]  = shadow1[96:127]
+  m128 addr_vect = SHUF(shadow0, shadow1, 1, 3, 1, 3);
+  if (!is_write) {
+    // set IsRead bit in addr_vect
+    const m128 rw_mask1 = _mm_cvtsi64_si128(1 << 15);
+    const m128 rw_mask = SHUF(rw_mask1, rw_mask1, 0, 0, 0, 0);
+    addr_vect = _mm_or_si128(addr_vect, rw_mask);
+  }
+  // addr0 == addr_vect?
+  const m128 addr_res = _mm_cmpeq_epi32(addr0, addr_vect);
+  // epoch1[0:63]       = sync_epoch
+  const m128 epoch1 = _mm_cvtsi64_si128(sync_epoch);
+  // epoch[0:31]        = sync_epoch[0:31]
+  // epoch[32:63]       = sync_epoch[0:31]
+  // epoch[64:95]       = sync_epoch[0:31]
+  // epoch[96:127]      = sync_epoch[0:31]
+  const m128 epoch = SHUF(epoch1, epoch1, 0, 0, 0, 0);
+  // load low parts of shadow cell epochs into epoch_vect:
+  // epoch_vect[0:31]   = shadow0[0:31]
+  // epoch_vect[32:63]  = shadow0[64:95]
+  // epoch_vect[64:95]  = shadow1[0:31]
+  // epoch_vect[96:127] = shadow1[64:95]
+  const m128 epoch_vect = SHUF(shadow0, shadow1, 0, 2, 0, 2);
+  // epoch_vect >= sync_epoch?
+  const m128 epoch_res = _mm_cmpgt_epi32(epoch_vect, epoch);
+  // addr_res & epoch_res
+  const m128 res = _mm_and_si128(addr_res, epoch_res);
+  // mask[0] = res[7]
+  // mask[1] = res[15]
+  // ...
+  // mask[15] = res[127]
+  const int mask = _mm_movemask_epi8(res);
+  return mask != 0;
+}
+#endif
+
+ALWAYS_INLINE
+bool ContainsSameAccess(u64 *s, u64 a, u64 sync_epoch, bool is_write) {
+#if TSAN_VECTORIZE
+  bool res = ContainsSameAccessFast(s, a, sync_epoch, is_write);
+  // NOTE: this check can fail if the shadow is concurrently mutated
+  // by other threads. But it still can be useful if you modify
+  // ContainsSameAccessFast and want to ensure that it's not completely broken.
+  // DCHECK_EQ(res, ContainsSameAccessSlow(s, a, sync_epoch, is_write));
+  return res;
+#else
+  return ContainsSameAccessSlow(s, a, sync_epoch, is_write);
+#endif
+}
+
+ALWAYS_INLINE USED void MemoryAccess(ThreadState *thr, uptr pc, uptr addr,
+                                     int kAccessSizeLog, bool kAccessIsWrite,
+                                     bool kIsAtomic) {
+  RawShadow *shadow_mem = MemToShadow(addr);
+  DPrintf2(
+      "#%d: MemoryAccess: @%p %p size=%d"
+      " is_write=%d shadow_mem=%p {%zx, %zx, %zx, %zx}\n",
+      (int)thr->fast_state.tid(), (void *)pc, (void *)addr,
+      (int)(1 << kAccessSizeLog), kAccessIsWrite, shadow_mem,
+      (uptr)shadow_mem[0], (uptr)shadow_mem[1], (uptr)shadow_mem[2],
+      (uptr)shadow_mem[3]);
+#if SANITIZER_DEBUG
+  if (!IsAppMem(addr)) {
+    Printf("Access to non app mem %zx\n", addr);
+    DCHECK(IsAppMem(addr));
+  }
+  if (!IsShadowMem(shadow_mem)) {
+    Printf("Bad shadow addr %p (%zx)\n", shadow_mem, addr);
+    DCHECK(IsShadowMem(shadow_mem));
+  }
+#endif
+
+  if (!SANITIZER_GO && !kAccessIsWrite && *shadow_mem == kShadowRodata) {
+    // Access to .rodata section, no races here.
+    // Measurements show that it can be 10-20% of all memory accesses.
+    return;
+  }
+
+  FastState fast_state = thr->fast_state;
+  if (UNLIKELY(fast_state.GetIgnoreBit())) {
+    return;
+  }
+
+  Shadow cur(fast_state);
+  cur.SetAddr0AndSizeLog(addr & 7, kAccessSizeLog);
+  cur.SetWrite(kAccessIsWrite);
+  cur.SetAtomic(kIsAtomic);
+
+  if (LIKELY(ContainsSameAccess(shadow_mem, cur.raw(), thr->fast_synch_epoch,
+                                kAccessIsWrite))) {
+    return;
+  }
+
+  if (kCollectHistory) {
+    fast_state.IncrementEpoch();
+    thr->fast_state = fast_state;
+    TraceAddEvent(thr, fast_state, EventTypeMop, pc);
+    cur.IncrementEpoch();
+  }
+
+  MemoryAccessImpl1(thr, addr, kAccessSizeLog, kAccessIsWrite, kIsAtomic,
+                    shadow_mem, cur);
+}
+
+// Called by MemoryAccessRange in tsan_rtl_thread.cpp
+ALWAYS_INLINE USED void MemoryAccessImpl(ThreadState *thr, uptr addr,
+                                         int kAccessSizeLog,
+                                         bool kAccessIsWrite, bool kIsAtomic,
+                                         u64 *shadow_mem, Shadow cur) {
+  if (LIKELY(ContainsSameAccess(shadow_mem, cur.raw(), thr->fast_synch_epoch,
+                                kAccessIsWrite))) {
+    return;
+  }
+
+  MemoryAccessImpl1(thr, addr, kAccessSizeLog, kAccessIsWrite, kIsAtomic,
+                    shadow_mem, cur);
+}
+
+static void MemoryRangeSet(ThreadState *thr, uptr pc, uptr addr, uptr size,
+                           u64 val) {
+  (void)thr;
+  (void)pc;
+  if (size == 0)
+    return;
+  // FIXME: fix me.
+  uptr offset = addr % kShadowCell;
+  if (offset) {
+    offset = kShadowCell - offset;
+    if (size <= offset)
+      return;
+    addr += offset;
+    size -= offset;
+  }
+  DCHECK_EQ(addr % 8, 0);
+  // If a user passes some insane arguments (memset(0)),
+  // let it just crash as usual.
+  if (!IsAppMem(addr) || !IsAppMem(addr + size - 1))
+    return;
+  // Don't want to touch lots of shadow memory.
+  // If a program maps 10MB stack, there is no need reset the whole range.
+  size = (size + (kShadowCell - 1)) & ~(kShadowCell - 1);
+  // UnmapOrDie/MmapFixedNoReserve does not work on Windows.
+  if (SANITIZER_WINDOWS || size < common_flags()->clear_shadow_mmap_threshold) {
+    RawShadow *p = MemToShadow(addr);
+    CHECK(IsShadowMem(p));
+    CHECK(IsShadowMem(p + size * kShadowCnt / kShadowCell - 1));
+    // FIXME: may overwrite a part outside the region
+    for (uptr i = 0; i < size / kShadowCell * kShadowCnt;) {
+      p[i++] = val;
+      for (uptr j = 1; j < kShadowCnt; j++) p[i++] = 0;
+    }
+  } else {
+    // The region is big, reset only beginning and end.
+    const uptr kPageSize = GetPageSizeCached();
+    RawShadow *begin = MemToShadow(addr);
+    RawShadow *end = begin + size / kShadowCell * kShadowCnt;
+    RawShadow *p = begin;
+    // Set at least first kPageSize/2 to page boundary.
+    while ((p < begin + kPageSize / kShadowSize / 2) || ((uptr)p % kPageSize)) {
+      *p++ = val;
+      for (uptr j = 1; j < kShadowCnt; j++) *p++ = 0;
+    }
+    // Reset middle part.
+    RawShadow *p1 = p;
+    p = RoundDown(end, kPageSize);
+    if (!MmapFixedSuperNoReserve((uptr)p1, (uptr)p - (uptr)p1))
+      Die();
+    // Set the ending.
+    while (p < end) {
+      *p++ = val;
+      for (uptr j = 1; j < kShadowCnt; j++) *p++ = 0;
+    }
+  }
+}
+
+void MemoryResetRange(ThreadState *thr, uptr pc, uptr addr, uptr size) {
+  MemoryRangeSet(thr, pc, addr, size, 0);
+}
+
+void MemoryRangeFreed(ThreadState *thr, uptr pc, uptr addr, uptr size) {
+  // Processing more than 1k (4k of shadow) is expensive,
+  // can cause excessive memory consumption (user does not necessary touch
+  // the whole range) and most likely unnecessary.
+  if (size > 1024)
+    size = 1024;
+  CHECK_EQ(thr->is_freeing, false);
+  thr->is_freeing = true;
+  MemoryAccessRange(thr, pc, addr, size, true);
+  thr->is_freeing = false;
+  if (kCollectHistory) {
+    thr->fast_state.IncrementEpoch();
+    TraceAddEvent(thr, thr->fast_state, EventTypeMop, pc);
+  }
+  Shadow s(thr->fast_state);
+  s.ClearIgnoreBit();
+  s.MarkAsFreed();
+  s.SetWrite(true);
+  s.SetAddr0AndSizeLog(0, 3);
+  MemoryRangeSet(thr, pc, addr, size, s.raw());
+}
+
+void MemoryRangeImitateWrite(ThreadState *thr, uptr pc, uptr addr, uptr size) {
+  if (kCollectHistory) {
+    thr->fast_state.IncrementEpoch();
+    TraceAddEvent(thr, thr->fast_state, EventTypeMop, pc);
+  }
+  Shadow s(thr->fast_state);
+  s.ClearIgnoreBit();
+  s.SetWrite(true);
+  s.SetAddr0AndSizeLog(0, 3);
+  MemoryRangeSet(thr, pc, addr, size, s.raw());
+}
+
+void MemoryRangeImitateWriteOrResetRange(ThreadState *thr, uptr pc, uptr addr,
+                                         uptr size) {
+  if (thr->ignore_reads_and_writes == 0)
+    MemoryRangeImitateWrite(thr, pc, addr, size);
+  else
+    MemoryResetRange(thr, pc, addr, size);
+}
+
+void MemoryAccessRange(ThreadState *thr, uptr pc, uptr addr, uptr size,
+                       bool is_write) {
+  if (size == 0)
+    return;
+
+  RawShadow *shadow_mem = MemToShadow(addr);
+  DPrintf2("#%d: MemoryAccessRange: @%p %p size=%d is_write=%d\n", thr->tid,
+           (void *)pc, (void *)addr, (int)size, is_write);
+
+#if SANITIZER_DEBUG
+  if (!IsAppMem(addr)) {
+    Printf("Access to non app mem %zx\n", addr);
+    DCHECK(IsAppMem(addr));
+  }
+  if (!IsAppMem(addr + size - 1)) {
+    Printf("Access to non app mem %zx\n", addr + size - 1);
+    DCHECK(IsAppMem(addr + size - 1));
+  }
+  if (!IsShadowMem(shadow_mem)) {
+    Printf("Bad shadow addr %p (%zx)\n", shadow_mem, addr);
+    DCHECK(IsShadowMem(shadow_mem));
+  }
+  if (!IsShadowMem(shadow_mem + size * kShadowCnt / 8 - 1)) {
+    Printf("Bad shadow addr %p (%zx)\n", shadow_mem + size * kShadowCnt / 8 - 1,
+           addr + size - 1);
+    DCHECK(IsShadowMem(shadow_mem + size * kShadowCnt / 8 - 1));
+  }
+#endif
+
+  if (*shadow_mem == kShadowRodata) {
+    DCHECK(!is_write);
+    // Access to .rodata section, no races here.
+    // Measurements show that it can be 10-20% of all memory accesses.
+    return;
+  }
+
+  FastState fast_state = thr->fast_state;
+  if (fast_state.GetIgnoreBit())
+    return;
+
+  fast_state.IncrementEpoch();
+  thr->fast_state = fast_state;
+  TraceAddEvent(thr, fast_state, EventTypeMop, pc);
+
+  bool unaligned = (addr % kShadowCell) != 0;
+
+  // Handle unaligned beginning, if any.
+  for (; addr % kShadowCell && size; addr++, size--) {
+    int const kAccessSizeLog = 0;
+    Shadow cur(fast_state);
+    cur.SetWrite(is_write);
+    cur.SetAddr0AndSizeLog(addr & (kShadowCell - 1), kAccessSizeLog);
+    MemoryAccessImpl(thr, addr, kAccessSizeLog, is_write, false, shadow_mem,
+                     cur);
+  }
+  if (unaligned)
+    shadow_mem += kShadowCnt;
+  // Handle middle part, if any.
+  for (; size >= kShadowCell; addr += kShadowCell, size -= kShadowCell) {
+    int const kAccessSizeLog = 3;
+    Shadow cur(fast_state);
+    cur.SetWrite(is_write);
+    cur.SetAddr0AndSizeLog(0, kAccessSizeLog);
+    MemoryAccessImpl(thr, addr, kAccessSizeLog, is_write, false, shadow_mem,
+                     cur);
+    shadow_mem += kShadowCnt;
+  }
+  // Handle ending, if any.
+  for (; size; addr++, size--) {
+    int const kAccessSizeLog = 0;
+    Shadow cur(fast_state);
+    cur.SetWrite(is_write);
+    cur.SetAddr0AndSizeLog(addr & (kShadowCell - 1), kAccessSizeLog);
+    MemoryAccessImpl(thr, addr, kAccessSizeLog, is_write, false, shadow_mem,
+                     cur);
+  }
+}
+
+}  // namespace __tsan
+
+#if !SANITIZER_GO
+// Must be included in this file to make sure everything is inlined.
+#  include "tsan_interface.inc"
+#endif
diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl_amd64.S b/compiler-rt/lib/tsan/rtl/tsan_rtl_amd64.S
index 5913aa360c5d..632b19d18158 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_rtl_amd64.S
+++ b/compiler-rt/lib/tsan/rtl/tsan_rtl_amd64.S
@@ -13,6 +13,7 @@ ASM_HIDDEN(__tsan_trace_switch)
 .globl ASM_SYMBOL(__tsan_trace_switch_thunk)
 ASM_SYMBOL(__tsan_trace_switch_thunk):
   CFI_STARTPROC
+  _CET_ENDBR
   # Save scratch registers.
   push %rax
   CFI_ADJUST_CFA_OFFSET(8)
@@ -41,6 +42,25 @@ ASM_SYMBOL(__tsan_trace_switch_thunk):
   push %r11
   CFI_ADJUST_CFA_OFFSET(8)
   CFI_REL_OFFSET(%r11, 0)
+  # All XMM registers are caller-saved.
+  sub $0x100, %rsp
+  CFI_ADJUST_CFA_OFFSET(0x100)
+  vmovdqu %xmm0, 0x0(%rsp)
+  vmovdqu %xmm1, 0x10(%rsp)
+  vmovdqu %xmm2, 0x20(%rsp)
+  vmovdqu %xmm3, 0x30(%rsp)
+  vmovdqu %xmm4, 0x40(%rsp)
+  vmovdqu %xmm5, 0x50(%rsp)
+  vmovdqu %xmm6, 0x60(%rsp)
+  vmovdqu %xmm7, 0x70(%rsp)
+  vmovdqu %xmm8, 0x80(%rsp)
+  vmovdqu %xmm9, 0x90(%rsp)
+  vmovdqu %xmm10, 0xa0(%rsp)
+  vmovdqu %xmm11, 0xb0(%rsp)
+  vmovdqu %xmm12, 0xc0(%rsp)
+  vmovdqu %xmm13, 0xd0(%rsp)
+  vmovdqu %xmm14, 0xe0(%rsp)
+  vmovdqu %xmm15, 0xf0(%rsp)
   # Align stack frame.
   push %rbx  # non-scratch
   CFI_ADJUST_CFA_OFFSET(8)
@@ -58,6 +78,24 @@ ASM_SYMBOL(__tsan_trace_switch_thunk):
   pop %rbx
   CFI_ADJUST_CFA_OFFSET(-8)
   # Restore scratch registers.
+  vmovdqu 0x0(%rsp), %xmm0
+  vmovdqu 0x10(%rsp), %xmm1
+  vmovdqu 0x20(%rsp), %xmm2
+  vmovdqu 0x30(%rsp), %xmm3
+  vmovdqu 0x40(%rsp), %xmm4
+  vmovdqu 0x50(%rsp), %xmm5
+  vmovdqu 0x60(%rsp), %xmm6
+  vmovdqu 0x70(%rsp), %xmm7
+  vmovdqu 0x80(%rsp), %xmm8
+  vmovdqu 0x90(%rsp), %xmm9
+  vmovdqu 0xa0(%rsp), %xmm10
+  vmovdqu 0xb0(%rsp), %xmm11
+  vmovdqu 0xc0(%rsp), %xmm12
+  vmovdqu 0xd0(%rsp), %xmm13
+  vmovdqu 0xe0(%rsp), %xmm14
+  vmovdqu 0xf0(%rsp), %xmm15
+  add $0x100, %rsp
+  CFI_ADJUST_CFA_OFFSET(-0x100)
   pop %r11
   CFI_ADJUST_CFA_OFFSET(-8)
   pop %r10
@@ -93,6 +131,7 @@ ASM_HIDDEN(__tsan_report_race)
 .globl ASM_SYMBOL(__tsan_report_race_thunk)
 ASM_SYMBOL(__tsan_report_race_thunk):
   CFI_STARTPROC
+  _CET_ENDBR
   # Save scratch registers.
   push %rax
   CFI_ADJUST_CFA_OFFSET(8)
@@ -121,6 +160,25 @@ ASM_SYMBOL(__tsan_report_race_thunk):
   push %r11
   CFI_ADJUST_CFA_OFFSET(8)
   CFI_REL_OFFSET(%r11, 0)
+  # All XMM registers are caller-saved.
+  sub $0x100, %rsp
+  CFI_ADJUST_CFA_OFFSET(0x100)
+  vmovdqu %xmm0, 0x0(%rsp)
+  vmovdqu %xmm1, 0x10(%rsp)
+  vmovdqu %xmm2, 0x20(%rsp)
+  vmovdqu %xmm3, 0x30(%rsp)
+  vmovdqu %xmm4, 0x40(%rsp)
+  vmovdqu %xmm5, 0x50(%rsp)
+  vmovdqu %xmm6, 0x60(%rsp)
+  vmovdqu %xmm7, 0x70(%rsp)
+  vmovdqu %xmm8, 0x80(%rsp)
+  vmovdqu %xmm9, 0x90(%rsp)
+  vmovdqu %xmm10, 0xa0(%rsp)
+  vmovdqu %xmm11, 0xb0(%rsp)
+  vmovdqu %xmm12, 0xc0(%rsp)
+  vmovdqu %xmm13, 0xd0(%rsp)
+  vmovdqu %xmm14, 0xe0(%rsp)
+  vmovdqu %xmm15, 0xf0(%rsp)
   # Align stack frame.
   push %rbx  # non-scratch
   CFI_ADJUST_CFA_OFFSET(8)
@@ -138,6 +196,24 @@ ASM_SYMBOL(__tsan_report_race_thunk):
   pop %rbx
   CFI_ADJUST_CFA_OFFSET(-8)
   # Restore scratch registers.
+  vmovdqu 0x0(%rsp), %xmm0
+  vmovdqu 0x10(%rsp), %xmm1
+  vmovdqu 0x20(%rsp), %xmm2
+  vmovdqu 0x30(%rsp), %xmm3
+  vmovdqu 0x40(%rsp), %xmm4
+  vmovdqu 0x50(%rsp), %xmm5
+  vmovdqu 0x60(%rsp), %xmm6
+  vmovdqu 0x70(%rsp), %xmm7
+  vmovdqu 0x80(%rsp), %xmm8
+  vmovdqu 0x90(%rsp), %xmm9
+  vmovdqu 0xa0(%rsp), %xmm10
+  vmovdqu 0xb0(%rsp), %xmm11
+  vmovdqu 0xc0(%rsp), %xmm12
+  vmovdqu 0xd0(%rsp), %xmm13
+  vmovdqu 0xe0(%rsp), %xmm14
+  vmovdqu 0xf0(%rsp), %xmm15
+  add $0x100, %rsp
+  CFI_ADJUST_CFA_OFFSET(-0x100)
   pop %r11
   CFI_ADJUST_CFA_OFFSET(-8)
   pop %r10
@@ -185,6 +261,7 @@ ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(setjmp))
 ASM_SYMBOL_INTERCEPTOR(setjmp):
 #endif
   CFI_STARTPROC
+  _CET_ENDBR
   // save env parameter
   push %rdi
   CFI_ADJUST_CFA_OFFSET(8)
@@ -226,6 +303,7 @@ ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(setjmp))
 ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(_setjmp))
 ASM_SYMBOL_INTERCEPTOR(_setjmp):
   CFI_STARTPROC
+  _CET_ENDBR
   // save env parameter
   push %rdi
   CFI_ADJUST_CFA_OFFSET(8)
@@ -267,6 +345,7 @@ ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(sigsetjmp))
 ASM_SYMBOL_INTERCEPTOR(sigsetjmp):
 #endif
   CFI_STARTPROC
+  _CET_ENDBR
   // save env parameter
   push %rdi
   CFI_ADJUST_CFA_OFFSET(8)
@@ -323,6 +402,7 @@ ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(sigsetjmp))
 ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(__sigsetjmp))
 ASM_SYMBOL_INTERCEPTOR(__sigsetjmp):
   CFI_STARTPROC
+  _CET_ENDBR
   // save env parameter
   push %rdi
   CFI_ADJUST_CFA_OFFSET(8)
diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cpp b/compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cpp
index 27ae279d6304..7d6b41116aa6 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cpp
@@ -35,7 +35,7 @@ struct Callback final : public DDCallback {
     DDCallback::lt = thr->dd_lt;
   }
 
-  u32 Unwind() override { return CurrentStackId(thr, pc); }
+  StackID Unwind() override { return CurrentStackId(thr, pc); }
   int UniqueTid() override { return thr->unique_id; }
 };
 
@@ -53,7 +53,7 @@ static void ReportMutexMisuse(ThreadState *thr, uptr pc, ReportType typ,
     return;
   if (!ShouldReport(thr, typ))
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
+  ThreadRegistryLock l(&ctx->thread_registry);
   ScopedReport rep(typ);
   rep.AddMutex(mid);
   VarSizeStackTrace trace;
@@ -63,51 +63,54 @@ static void ReportMutexMisuse(ThreadState *thr, uptr pc, ReportType typ,
   OutputReport(thr, rep);
 }
 
-void MutexCreate(ThreadState *thr, uptr pc, uptr addr, u32 flagz) NO_THREAD_SAFETY_ANALYSIS {
+void MutexCreate(ThreadState *thr, uptr pc, uptr addr, u32 flagz) {
   DPrintf("#%d: MutexCreate %zx flagz=0x%x\n", thr->tid, addr, flagz);
   if (!(flagz & MutexFlagLinkerInit) && IsAppMem(addr)) {
     CHECK(!thr->is_freeing);
     thr->is_freeing = true;
-    MemoryWrite(thr, pc, addr, kSizeLog1);
+    MemoryAccess(thr, pc, addr, 1, kAccessWrite);
     thr->is_freeing = false;
   }
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
+  SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+  Lock l(&s->mtx);
   s->SetFlags(flagz & MutexCreationFlagMask);
+  // Save stack in the case the sync object was created before as atomic.
   if (!SANITIZER_GO && s->creation_stack_id == 0)
     s->creation_stack_id = CurrentStackId(thr, pc);
-  s->mtx.Unlock();
 }
 
-void MutexDestroy(ThreadState *thr, uptr pc, uptr addr, u32 flagz) NO_THREAD_SAFETY_ANALYSIS {
+void MutexDestroy(ThreadState *thr, uptr pc, uptr addr, u32 flagz) {
   DPrintf("#%d: MutexDestroy %zx\n", thr->tid, addr);
-  SyncVar *s = ctx->metamap.GetIfExistsAndLock(addr, true);
-  if (s == 0)
-    return;
-  if ((flagz & MutexFlagLinkerInit)
-      || s->IsFlagSet(MutexFlagLinkerInit)
-      || ((flagz & MutexFlagNotStatic) && !s->IsFlagSet(MutexFlagNotStatic))) {
-    // Destroy is no-op for linker-initialized mutexes.
-    s->mtx.Unlock();
-    return;
-  }
-  if (common_flags()->detect_deadlocks) {
-    Callback cb(thr, pc);
-    ctx->dd->MutexDestroy(&cb, &s->dd);
-    ctx->dd->MutexInit(&cb, &s->dd);
-  }
   bool unlock_locked = false;
-  if (flags()->report_destroy_locked && s->owner_tid != kInvalidTid &&
-      !s->IsFlagSet(MutexFlagBroken)) {
-    s->SetFlags(MutexFlagBroken);
-    unlock_locked = true;
+  u64 mid = 0;
+  u64 last_lock = 0;
+  {
+    SyncVar *s = ctx->metamap.GetSyncIfExists(addr);
+    if (s == 0)
+      return;
+    Lock l(&s->mtx);
+    if ((flagz & MutexFlagLinkerInit) || s->IsFlagSet(MutexFlagLinkerInit) ||
+        ((flagz & MutexFlagNotStatic) && !s->IsFlagSet(MutexFlagNotStatic))) {
+      // Destroy is no-op for linker-initialized mutexes.
+      return;
+    }
+    if (common_flags()->detect_deadlocks) {
+      Callback cb(thr, pc);
+      ctx->dd->MutexDestroy(&cb, &s->dd);
+      ctx->dd->MutexInit(&cb, &s->dd);
+    }
+    if (flags()->report_destroy_locked && s->owner_tid != kInvalidTid &&
+        !s->IsFlagSet(MutexFlagBroken)) {
+      s->SetFlags(MutexFlagBroken);
+      unlock_locked = true;
+    }
+    mid = s->GetId();
+    last_lock = s->last_lock;
+    if (!unlock_locked)
+      s->Reset(thr->proc());  // must not reset it before the report is printed
   }
-  u64 mid = s->GetId();
-  u64 last_lock = s->last_lock;
-  if (!unlock_locked)
-    s->Reset(thr->proc());  // must not reset it before the report is printed
-  s->mtx.Unlock();
   if (unlock_locked && ShouldReport(thr, ReportTypeMutexDestroyLocked)) {
-    ThreadRegistryLock l(ctx->thread_registry);
+    ThreadRegistryLock l(&ctx->thread_registry);
     ScopedReport rep(ReportTypeMutexDestroyLocked);
     rep.AddMutex(mid);
     VarSizeStackTrace trace;
@@ -119,43 +122,39 @@ void MutexDestroy(ThreadState *thr, uptr pc, uptr addr, u32 flagz) NO_THREAD_SAF
     rep.AddLocation(addr, 1);
     OutputReport(thr, rep);
 
-    SyncVar *s = ctx->metamap.GetIfExistsAndLock(addr, true);
+    SyncVar *s = ctx->metamap.GetSyncIfExists(addr);
     if (s != 0) {
+      Lock l(&s->mtx);
       s->Reset(thr->proc());
-      s->mtx.Unlock();
     }
   }
   thr->mset.Remove(mid);
   // Imitate a memory write to catch unlock-destroy races.
   // Do this outside of sync mutex, because it can report a race which locks
   // sync mutexes.
-  if (IsAppMem(addr)) {
-    CHECK(!thr->is_freeing);
-    thr->is_freeing = true;
-    MemoryWrite(thr, pc, addr, kSizeLog1);
-    thr->is_freeing = false;
-  }
+  if (IsAppMem(addr))
+    MemoryAccess(thr, pc, addr, 1, kAccessWrite | kAccessFree);
   // s will be destroyed and freed in MetaMap::FreeBlock.
 }
 
-void MutexPreLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) NO_THREAD_SAFETY_ANALYSIS {
+void MutexPreLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) {
   DPrintf("#%d: MutexPreLock %zx flagz=0x%x\n", thr->tid, addr, flagz);
   if (!(flagz & MutexFlagTryLock) && common_flags()->detect_deadlocks) {
-    SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, false);
-    s->UpdateFlags(flagz);
-    if (s->owner_tid != thr->tid) {
-      Callback cb(thr, pc);
-      ctx->dd->MutexBeforeLock(&cb, &s->dd, true);
-      s->mtx.ReadUnlock();
-      ReportDeadlock(thr, pc, ctx->dd->GetReport(&cb));
-    } else {
-      s->mtx.ReadUnlock();
+    SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+    {
+      ReadLock l(&s->mtx);
+      s->UpdateFlags(flagz);
+      if (s->owner_tid != thr->tid) {
+        Callback cb(thr, pc);
+        ctx->dd->MutexBeforeLock(&cb, &s->dd, true);
+      }
     }
+    Callback cb(thr, pc);
+    ReportDeadlock(thr, pc, ctx->dd->GetReport(&cb));
   }
 }
 
-void MutexPostLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz,
-                   int rec) NO_THREAD_SAFETY_ANALYSIS {
+void MutexPostLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz, int rec) {
   DPrintf("#%d: MutexPostLock %zx flag=0x%x rec=%d\n",
       thr->tid, addr, flagz, rec);
   if (flagz & MutexFlagRecursiveLock)
@@ -163,43 +162,45 @@ void MutexPostLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz,
   else
     rec = 1;
   if (IsAppMem(addr))
-    MemoryReadAtomic(thr, pc, addr, kSizeLog1);
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
-  s->UpdateFlags(flagz);
-  thr->fast_state.IncrementEpoch();
-  TraceAddEvent(thr, thr->fast_state, EventTypeLock, s->GetId());
-  bool report_double_lock = false;
-  if (s->owner_tid == kInvalidTid) {
-    CHECK_EQ(s->recursion, 0);
-    s->owner_tid = thr->tid;
-    s->last_lock = thr->fast_state.raw();
-  } else if (s->owner_tid == thr->tid) {
-    CHECK_GT(s->recursion, 0);
-  } else if (flags()->report_mutex_bugs && !s->IsFlagSet(MutexFlagBroken)) {
-    s->SetFlags(MutexFlagBroken);
-    report_double_lock = true;
-  }
-  const bool first = s->recursion == 0;
-  s->recursion += rec;
-  if (first) {
-    AcquireImpl(thr, pc, &s->clock);
-    AcquireImpl(thr, pc, &s->read_clock);
-  } else if (!s->IsFlagSet(MutexFlagWriteReentrant)) {
-  }
-  thr->mset.Add(s->GetId(), true, thr->fast_state.epoch());
+    MemoryAccess(thr, pc, addr, 1, kAccessRead | kAccessAtomic);
+  u64 mid = 0;
   bool pre_lock = false;
-  if (first && common_flags()->detect_deadlocks) {
-    pre_lock = (flagz & MutexFlagDoPreLockOnPostLock) &&
-        !(flagz & MutexFlagTryLock);
-    Callback cb(thr, pc);
-    if (pre_lock)
-      ctx->dd->MutexBeforeLock(&cb, &s->dd, true);
-    ctx->dd->MutexAfterLock(&cb, &s->dd, true, flagz & MutexFlagTryLock);
+  bool first = false;
+  bool report_double_lock = false;
+  {
+    SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+    Lock l(&s->mtx);
+    s->UpdateFlags(flagz);
+    thr->fast_state.IncrementEpoch();
+    TraceAddEvent(thr, thr->fast_state, EventTypeLock, s->GetId());
+    if (s->owner_tid == kInvalidTid) {
+      CHECK_EQ(s->recursion, 0);
+      s->owner_tid = thr->tid;
+      s->last_lock = thr->fast_state.raw();
+    } else if (s->owner_tid == thr->tid) {
+      CHECK_GT(s->recursion, 0);
+    } else if (flags()->report_mutex_bugs && !s->IsFlagSet(MutexFlagBroken)) {
+      s->SetFlags(MutexFlagBroken);
+      report_double_lock = true;
+    }
+    first = s->recursion == 0;
+    s->recursion += rec;
+    if (first) {
+      AcquireImpl(thr, pc, &s->clock);
+      AcquireImpl(thr, pc, &s->read_clock);
+    } else if (!s->IsFlagSet(MutexFlagWriteReentrant)) {
+    }
+    thr->mset.Add(s->GetId(), true, thr->fast_state.epoch());
+    if (first && common_flags()->detect_deadlocks) {
+      pre_lock =
+          (flagz & MutexFlagDoPreLockOnPostLock) && !(flagz & MutexFlagTryLock);
+      Callback cb(thr, pc);
+      if (pre_lock)
+        ctx->dd->MutexBeforeLock(&cb, &s->dd, true);
+      ctx->dd->MutexAfterLock(&cb, &s->dd, true, flagz & MutexFlagTryLock);
+    }
+    mid = s->GetId();
   }
-  u64 mid = s->GetId();
-  s->mtx.Unlock();
-  // Can't touch s after this point.
-  s = 0;
   if (report_double_lock)
     ReportMutexMisuse(thr, pc, ReportTypeMutexDoubleLock, addr, mid);
   if (first && pre_lock && common_flags()->detect_deadlocks) {
@@ -208,38 +209,40 @@ void MutexPostLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz,
   }
 }
 
-int MutexUnlock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) NO_THREAD_SAFETY_ANALYSIS {
+int MutexUnlock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) {
   DPrintf("#%d: MutexUnlock %zx flagz=0x%x\n", thr->tid, addr, flagz);
   if (IsAppMem(addr))
-    MemoryReadAtomic(thr, pc, addr, kSizeLog1);
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
-  thr->fast_state.IncrementEpoch();
-  TraceAddEvent(thr, thr->fast_state, EventTypeUnlock, s->GetId());
-  int rec = 0;
+    MemoryAccess(thr, pc, addr, 1, kAccessRead | kAccessAtomic);
+  u64 mid = 0;
   bool report_bad_unlock = false;
-  if (!SANITIZER_GO && (s->recursion == 0 || s->owner_tid != thr->tid)) {
-    if (flags()->report_mutex_bugs && !s->IsFlagSet(MutexFlagBroken)) {
-      s->SetFlags(MutexFlagBroken);
-      report_bad_unlock = true;
-    }
-  } else {
-    rec = (flagz & MutexFlagRecursiveUnlock) ? s->recursion : 1;
-    s->recursion -= rec;
-    if (s->recursion == 0) {
-      s->owner_tid = kInvalidTid;
-      ReleaseStoreImpl(thr, pc, &s->clock);
+  int rec = 0;
+  {
+    SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+    Lock l(&s->mtx);
+    thr->fast_state.IncrementEpoch();
+    TraceAddEvent(thr, thr->fast_state, EventTypeUnlock, s->GetId());
+    if (!SANITIZER_GO && (s->recursion == 0 || s->owner_tid != thr->tid)) {
+      if (flags()->report_mutex_bugs && !s->IsFlagSet(MutexFlagBroken)) {
+        s->SetFlags(MutexFlagBroken);
+        report_bad_unlock = true;
+      }
     } else {
+      rec = (flagz & MutexFlagRecursiveUnlock) ? s->recursion : 1;
+      s->recursion -= rec;
+      if (s->recursion == 0) {
+        s->owner_tid = kInvalidTid;
+        ReleaseStoreImpl(thr, pc, &s->clock);
+      } else {
+      }
     }
+    thr->mset.Del(s->GetId(), true);
+    if (common_flags()->detect_deadlocks && s->recursion == 0 &&
+        !report_bad_unlock) {
+      Callback cb(thr, pc);
+      ctx->dd->MutexBeforeUnlock(&cb, &s->dd, true);
+    }
+    mid = s->GetId();
   }
-  thr->mset.Del(s->GetId(), true);
-  if (common_flags()->detect_deadlocks && s->recursion == 0 &&
-      !report_bad_unlock) {
-    Callback cb(thr, pc);
-    ctx->dd->MutexBeforeUnlock(&cb, &s->dd, true);
-  }
-  u64 mid = s->GetId();
-  s->mtx.Unlock();
-  // Can't touch s after this point.
   if (report_bad_unlock)
     ReportMutexMisuse(thr, pc, ReportTypeMutexBadUnlock, addr, mid);
   if (common_flags()->detect_deadlocks && !report_bad_unlock) {
@@ -249,49 +252,53 @@ int MutexUnlock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) NO_THREAD_SAFET
   return rec;
 }
 
-void MutexPreReadLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) NO_THREAD_SAFETY_ANALYSIS {
+void MutexPreReadLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) {
   DPrintf("#%d: MutexPreReadLock %zx flagz=0x%x\n", thr->tid, addr, flagz);
   if (!(flagz & MutexFlagTryLock) && common_flags()->detect_deadlocks) {
-    SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, false);
-    s->UpdateFlags(flagz);
+    {
+      SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+      ReadLock l(&s->mtx);
+      s->UpdateFlags(flagz);
+      Callback cb(thr, pc);
+      ctx->dd->MutexBeforeLock(&cb, &s->dd, false);
+    }
     Callback cb(thr, pc);
-    ctx->dd->MutexBeforeLock(&cb, &s->dd, false);
-    s->mtx.ReadUnlock();
     ReportDeadlock(thr, pc, ctx->dd->GetReport(&cb));
   }
 }
 
-void MutexPostReadLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) NO_THREAD_SAFETY_ANALYSIS {
+void MutexPostReadLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) {
   DPrintf("#%d: MutexPostReadLock %zx flagz=0x%x\n", thr->tid, addr, flagz);
   if (IsAppMem(addr))
-    MemoryReadAtomic(thr, pc, addr, kSizeLog1);
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, false);
-  s->UpdateFlags(flagz);
-  thr->fast_state.IncrementEpoch();
-  TraceAddEvent(thr, thr->fast_state, EventTypeRLock, s->GetId());
+    MemoryAccess(thr, pc, addr, 1, kAccessRead | kAccessAtomic);
+  u64 mid = 0;
   bool report_bad_lock = false;
-  if (s->owner_tid != kInvalidTid) {
-    if (flags()->report_mutex_bugs && !s->IsFlagSet(MutexFlagBroken)) {
-      s->SetFlags(MutexFlagBroken);
-      report_bad_lock = true;
-    }
-  }
-  AcquireImpl(thr, pc, &s->clock);
-  s->last_lock = thr->fast_state.raw();
-  thr->mset.Add(s->GetId(), false, thr->fast_state.epoch());
   bool pre_lock = false;
-  if (common_flags()->detect_deadlocks) {
-    pre_lock = (flagz & MutexFlagDoPreLockOnPostLock) &&
-        !(flagz & MutexFlagTryLock);
-    Callback cb(thr, pc);
-    if (pre_lock)
-      ctx->dd->MutexBeforeLock(&cb, &s->dd, false);
-    ctx->dd->MutexAfterLock(&cb, &s->dd, false, flagz & MutexFlagTryLock);
+  {
+    SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+    ReadLock l(&s->mtx);
+    s->UpdateFlags(flagz);
+    thr->fast_state.IncrementEpoch();
+    TraceAddEvent(thr, thr->fast_state, EventTypeRLock, s->GetId());
+    if (s->owner_tid != kInvalidTid) {
+      if (flags()->report_mutex_bugs && !s->IsFlagSet(MutexFlagBroken)) {
+        s->SetFlags(MutexFlagBroken);
+        report_bad_lock = true;
+      }
+    }
+    AcquireImpl(thr, pc, &s->clock);
+    s->last_lock = thr->fast_state.raw();
+    thr->mset.Add(s->GetId(), false, thr->fast_state.epoch());
+    if (common_flags()->detect_deadlocks) {
+      pre_lock =
+          (flagz & MutexFlagDoPreLockOnPostLock) && !(flagz & MutexFlagTryLock);
+      Callback cb(thr, pc);
+      if (pre_lock)
+        ctx->dd->MutexBeforeLock(&cb, &s->dd, false);
+      ctx->dd->MutexAfterLock(&cb, &s->dd, false, flagz & MutexFlagTryLock);
+    }
+    mid = s->GetId();
   }
-  u64 mid = s->GetId();
-  s->mtx.ReadUnlock();
-  // Can't touch s after this point.
-  s = 0;
   if (report_bad_lock)
     ReportMutexMisuse(thr, pc, ReportTypeMutexBadReadLock, addr, mid);
   if (pre_lock  && common_flags()->detect_deadlocks) {
@@ -300,28 +307,30 @@ void MutexPostReadLock(ThreadState *thr, uptr pc, uptr addr, u32 flagz) NO_THREA
   }
 }
 
-void MutexReadUnlock(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANALYSIS {
+void MutexReadUnlock(ThreadState *thr, uptr pc, uptr addr) {
   DPrintf("#%d: MutexReadUnlock %zx\n", thr->tid, addr);
   if (IsAppMem(addr))
-    MemoryReadAtomic(thr, pc, addr, kSizeLog1);
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
-  thr->fast_state.IncrementEpoch();
-  TraceAddEvent(thr, thr->fast_state, EventTypeRUnlock, s->GetId());
+    MemoryAccess(thr, pc, addr, 1, kAccessRead | kAccessAtomic);
+  u64 mid = 0;
   bool report_bad_unlock = false;
-  if (s->owner_tid != kInvalidTid) {
-    if (flags()->report_mutex_bugs && !s->IsFlagSet(MutexFlagBroken)) {
-      s->SetFlags(MutexFlagBroken);
-      report_bad_unlock = true;
+  {
+    SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+    Lock l(&s->mtx);
+    thr->fast_state.IncrementEpoch();
+    TraceAddEvent(thr, thr->fast_state, EventTypeRUnlock, s->GetId());
+    if (s->owner_tid != kInvalidTid) {
+      if (flags()->report_mutex_bugs && !s->IsFlagSet(MutexFlagBroken)) {
+        s->SetFlags(MutexFlagBroken);
+        report_bad_unlock = true;
+      }
     }
+    ReleaseImpl(thr, pc, &s->read_clock);
+    if (common_flags()->detect_deadlocks && s->recursion == 0) {
+      Callback cb(thr, pc);
+      ctx->dd->MutexBeforeUnlock(&cb, &s->dd, false);
+    }
+    mid = s->GetId();
   }
-  ReleaseImpl(thr, pc, &s->read_clock);
-  if (common_flags()->detect_deadlocks && s->recursion == 0) {
-    Callback cb(thr, pc);
-    ctx->dd->MutexBeforeUnlock(&cb, &s->dd, false);
-  }
-  u64 mid = s->GetId();
-  s->mtx.Unlock();
-  // Can't touch s after this point.
   thr->mset.Del(mid, false);
   if (report_bad_unlock)
     ReportMutexMisuse(thr, pc, ReportTypeMutexBadReadUnlock, addr, mid);
@@ -331,42 +340,44 @@ void MutexReadUnlock(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANAL
   }
 }
 
-void MutexReadOrWriteUnlock(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANALYSIS {
+void MutexReadOrWriteUnlock(ThreadState *thr, uptr pc, uptr addr) {
   DPrintf("#%d: MutexReadOrWriteUnlock %zx\n", thr->tid, addr);
   if (IsAppMem(addr))
-    MemoryReadAtomic(thr, pc, addr, kSizeLog1);
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
-  bool write = true;
+    MemoryAccess(thr, pc, addr, 1, kAccessRead | kAccessAtomic);
+  u64 mid = 0;
   bool report_bad_unlock = false;
-  if (s->owner_tid == kInvalidTid) {
-    // Seems to be read unlock.
-    write = false;
-    thr->fast_state.IncrementEpoch();
-    TraceAddEvent(thr, thr->fast_state, EventTypeRUnlock, s->GetId());
-    ReleaseImpl(thr, pc, &s->read_clock);
-  } else if (s->owner_tid == thr->tid) {
-    // Seems to be write unlock.
-    thr->fast_state.IncrementEpoch();
-    TraceAddEvent(thr, thr->fast_state, EventTypeUnlock, s->GetId());
-    CHECK_GT(s->recursion, 0);
-    s->recursion--;
-    if (s->recursion == 0) {
-      s->owner_tid = kInvalidTid;
-      ReleaseStoreImpl(thr, pc, &s->clock);
-    } else {
+  {
+    SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+    Lock l(&s->mtx);
+    bool write = true;
+    if (s->owner_tid == kInvalidTid) {
+      // Seems to be read unlock.
+      write = false;
+      thr->fast_state.IncrementEpoch();
+      TraceAddEvent(thr, thr->fast_state, EventTypeRUnlock, s->GetId());
+      ReleaseImpl(thr, pc, &s->read_clock);
+    } else if (s->owner_tid == thr->tid) {
+      // Seems to be write unlock.
+      thr->fast_state.IncrementEpoch();
+      TraceAddEvent(thr, thr->fast_state, EventTypeUnlock, s->GetId());
+      CHECK_GT(s->recursion, 0);
+      s->recursion--;
+      if (s->recursion == 0) {
+        s->owner_tid = kInvalidTid;
+        ReleaseStoreImpl(thr, pc, &s->clock);
+      } else {
+      }
+    } else if (!s->IsFlagSet(MutexFlagBroken)) {
+      s->SetFlags(MutexFlagBroken);
+      report_bad_unlock = true;
     }
-  } else if (!s->IsFlagSet(MutexFlagBroken)) {
-    s->SetFlags(MutexFlagBroken);
-    report_bad_unlock = true;
-  }
-  thr->mset.Del(s->GetId(), write);
-  if (common_flags()->detect_deadlocks && s->recursion == 0) {
-    Callback cb(thr, pc);
-    ctx->dd->MutexBeforeUnlock(&cb, &s->dd, write);
+    thr->mset.Del(s->GetId(), write);
+    if (common_flags()->detect_deadlocks && s->recursion == 0) {
+      Callback cb(thr, pc);
+      ctx->dd->MutexBeforeUnlock(&cb, &s->dd, write);
+    }
+    mid = s->GetId();
   }
-  u64 mid = s->GetId();
-  s->mtx.Unlock();
-  // Can't touch s after this point.
   if (report_bad_unlock)
     ReportMutexMisuse(thr, pc, ReportTypeMutexBadUnlock, addr, mid);
   if (common_flags()->detect_deadlocks) {
@@ -375,31 +386,29 @@ void MutexReadOrWriteUnlock(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFE
   }
 }
 
-void MutexRepair(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANALYSIS {
+void MutexRepair(ThreadState *thr, uptr pc, uptr addr) {
   DPrintf("#%d: MutexRepair %zx\n", thr->tid, addr);
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
+  SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+  Lock l(&s->mtx);
   s->owner_tid = kInvalidTid;
   s->recursion = 0;
-  s->mtx.Unlock();
 }
 
-void MutexInvalidAccess(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANALYSIS {
+void MutexInvalidAccess(ThreadState *thr, uptr pc, uptr addr) {
   DPrintf("#%d: MutexInvalidAccess %zx\n", thr->tid, addr);
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
-  u64 mid = s->GetId();
-  s->mtx.Unlock();
-  ReportMutexMisuse(thr, pc, ReportTypeMutexInvalidAccess, addr, mid);
+  SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, true);
+  ReportMutexMisuse(thr, pc, ReportTypeMutexInvalidAccess, addr, s->GetId());
 }
 
-void Acquire(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANALYSIS {
+void Acquire(ThreadState *thr, uptr pc, uptr addr) {
   DPrintf("#%d: Acquire %zx\n", thr->tid, addr);
   if (thr->ignore_sync)
     return;
-  SyncVar *s = ctx->metamap.GetIfExistsAndLock(addr, false);
+  SyncVar *s = ctx->metamap.GetSyncIfExists(addr);
   if (!s)
     return;
+  ReadLock l(&s->mtx);
   AcquireImpl(thr, pc, &s->clock);
-  s->mtx.ReadUnlock();
 }
 
 static void UpdateClockCallback(ThreadContextBase *tctx_base, void *arg) {
@@ -413,49 +422,48 @@ static void UpdateClockCallback(ThreadContextBase *tctx_base, void *arg) {
   thr->clock.set(&thr->proc()->clock_cache, tctx->tid, epoch);
 }
 
-void AcquireGlobal(ThreadState *thr, uptr pc) {
+void AcquireGlobal(ThreadState *thr) {
   DPrintf("#%d: AcquireGlobal\n", thr->tid);
   if (thr->ignore_sync)
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
-  ctx->thread_registry->RunCallbackForEachThreadLocked(
-      UpdateClockCallback, thr);
+  ThreadRegistryLock l(&ctx->thread_registry);
+  ctx->thread_registry.RunCallbackForEachThreadLocked(UpdateClockCallback, thr);
 }
 
-void ReleaseStoreAcquire(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANALYSIS {
+void ReleaseStoreAcquire(ThreadState *thr, uptr pc, uptr addr) {
   DPrintf("#%d: ReleaseStoreAcquire %zx\n", thr->tid, addr);
   if (thr->ignore_sync)
     return;
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
+  SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, false);
+  Lock l(&s->mtx);
   thr->fast_state.IncrementEpoch();
   // Can't increment epoch w/o writing to the trace as well.
   TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
   ReleaseStoreAcquireImpl(thr, pc, &s->clock);
-  s->mtx.Unlock();
 }
 
-void Release(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANALYSIS {
+void Release(ThreadState *thr, uptr pc, uptr addr) {
   DPrintf("#%d: Release %zx\n", thr->tid, addr);
   if (thr->ignore_sync)
     return;
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
+  SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, false);
+  Lock l(&s->mtx);
   thr->fast_state.IncrementEpoch();
   // Can't increment epoch w/o writing to the trace as well.
   TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
   ReleaseImpl(thr, pc, &s->clock);
-  s->mtx.Unlock();
 }
 
-void ReleaseStore(ThreadState *thr, uptr pc, uptr addr) NO_THREAD_SAFETY_ANALYSIS {
+void ReleaseStore(ThreadState *thr, uptr pc, uptr addr) {
   DPrintf("#%d: ReleaseStore %zx\n", thr->tid, addr);
   if (thr->ignore_sync)
     return;
-  SyncVar *s = ctx->metamap.GetOrCreateAndLock(thr, pc, addr, true);
+  SyncVar *s = ctx->metamap.GetSyncOrCreate(thr, pc, addr, false);
+  Lock l(&s->mtx);
   thr->fast_state.IncrementEpoch();
   // Can't increment epoch w/o writing to the trace as well.
   TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
   ReleaseStoreImpl(thr, pc, &s->clock);
-  s->mtx.Unlock();
 }
 
 #if !SANITIZER_GO
@@ -469,13 +477,13 @@ static void UpdateSleepClockCallback(ThreadContextBase *tctx_base, void *arg) {
 }
 
 void AfterSleep(ThreadState *thr, uptr pc) {
-  DPrintf("#%d: AfterSleep %zx\n", thr->tid);
+  DPrintf("#%d: AfterSleep\n", thr->tid);
   if (thr->ignore_sync)
     return;
   thr->last_sleep_stack_id = CurrentStackId(thr, pc);
-  ThreadRegistryLock l(ctx->thread_registry);
-  ctx->thread_registry->RunCallbackForEachThreadLocked(
-      UpdateSleepClockCallback, thr);
+  ThreadRegistryLock l(&ctx->thread_registry);
+  ctx->thread_registry.RunCallbackForEachThreadLocked(UpdateSleepClockCallback,
+                                                      thr);
 }
 #endif
 
@@ -521,7 +529,7 @@ void AcquireReleaseImpl(ThreadState *thr, uptr pc, SyncClock *c) {
 void ReportDeadlock(ThreadState *thr, uptr pc, DDReport *r) {
   if (r == 0 || !ShouldReport(thr, ReportTypeDeadlock))
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
+  ThreadRegistryLock l(&ctx->thread_registry);
   ScopedReport rep(ReportTypeDeadlock);
   for (int i = 0; i < r->n; i++) {
     rep.AddMutex(r->loop[i].mtx_ctx0);
diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cpp b/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cpp
index 3e809e653c70..811695d144c5 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cpp
@@ -68,8 +68,10 @@ static void StackStripMain(SymbolizedStack *frames) {
   } else if (last && 0 == internal_strcmp(last, "__tsan_thread_start_func")) {
     last_frame->ClearAll();
     last_frame2->next = nullptr;
-  // Strip global ctors init.
-  } else if (last && 0 == internal_strcmp(last, "__do_global_ctors_aux")) {
+    // Strip global ctors init, .preinit_array and main caller.
+  } else if (last && (0 == internal_strcmp(last, "__do_global_ctors_aux") ||
+                      0 == internal_strcmp(last, "__libc_csu_init") ||
+                      0 == internal_strcmp(last, "__libc_start_main"))) {
     last_frame->ClearAll();
     last_frame2->next = nullptr;
   // If both are 0, then we probably just failed to symbolize.
@@ -120,7 +122,7 @@ static ReportStack *SymbolizeStack(StackTrace trace) {
   }
   StackStripMain(top);
 
-  ReportStack *stack = ReportStack::New();
+  auto *stack = New<ReportStack>();
   stack->frames = top;
   return stack;
 }
@@ -132,7 +134,7 @@ bool ShouldReport(ThreadState *thr, ReportType typ) {
   CheckedMutex::CheckNoLocks();
   // For the same reason check we didn't lock thread_registry yet.
   if (SANITIZER_DEBUG)
-    ThreadRegistryLock l(ctx->thread_registry);
+    ThreadRegistryLock l(&ctx->thread_registry);
   if (!flags()->report_bugs || thr->suppress_reports)
     return false;
   switch (typ) {
@@ -154,9 +156,8 @@ bool ShouldReport(ThreadState *thr, ReportType typ) {
 }
 
 ScopedReportBase::ScopedReportBase(ReportType typ, uptr tag) {
-  ctx->thread_registry->CheckLocked();
-  void *mem = internal_alloc(MBlockReport, sizeof(ReportDesc));
-  rep_ = new(mem) ReportDesc;
+  ctx->thread_registry.CheckLocked();
+  rep_ = New<ReportDesc>();
   rep_->typ = typ;
   rep_->tag = tag;
   ctx->report_mtx.Lock();
@@ -165,7 +166,6 @@ ScopedReportBase::ScopedReportBase(ReportType typ, uptr tag) {
 ScopedReportBase::~ScopedReportBase() {
   ctx->report_mtx.Unlock();
   DestroyAndFree(rep_);
-  rep_ = nullptr;
 }
 
 void ScopedReportBase::AddStack(StackTrace stack, bool suppressable) {
@@ -176,8 +176,7 @@ void ScopedReportBase::AddStack(StackTrace stack, bool suppressable) {
 
 void ScopedReportBase::AddMemoryAccess(uptr addr, uptr external_tag, Shadow s,
                                        StackTrace stack, const MutexSet *mset) {
-  void *mem = internal_alloc(MBlockReportMop, sizeof(ReportMop));
-  ReportMop *mop = new(mem) ReportMop;
+  auto *mop = New<ReportMop>();
   rep_->mops.PushBack(mop);
   mop->tid = s.tid();
   mop->addr = addr + s.addr0();
@@ -196,7 +195,7 @@ void ScopedReportBase::AddMemoryAccess(uptr addr, uptr external_tag, Shadow s,
   }
 }
 
-void ScopedReportBase::AddUniqueTid(int unique_tid) {
+void ScopedReportBase::AddUniqueTid(Tid unique_tid) {
   rep_->unique_tids.PushBack(unique_tid);
 }
 
@@ -205,8 +204,7 @@ void ScopedReportBase::AddThread(const ThreadContext *tctx, bool suppressable) {
     if ((u32)rep_->threads[i]->id == tctx->tid)
       return;
   }
-  void *mem = internal_alloc(MBlockReportThread, sizeof(ReportThread));
-  ReportThread *rt = new(mem) ReportThread;
+  auto *rt = New<ReportThread>();
   rep_->threads.PushBack(rt);
   rt->id = tctx->tid;
   rt->os_id = tctx->os_id;
@@ -226,17 +224,17 @@ static bool FindThreadByUidLockedCallback(ThreadContextBase *tctx, void *arg) {
   return tctx->unique_id == (u32)unique_id;
 }
 
-static ThreadContext *FindThreadByUidLocked(int unique_id) {
-  ctx->thread_registry->CheckLocked();
+static ThreadContext *FindThreadByUidLocked(Tid unique_id) {
+  ctx->thread_registry.CheckLocked();
   return static_cast<ThreadContext *>(
-      ctx->thread_registry->FindThreadContextLocked(
+      ctx->thread_registry.FindThreadContextLocked(
           FindThreadByUidLockedCallback, &unique_id));
 }
 
-static ThreadContext *FindThreadByTidLocked(int tid) {
-  ctx->thread_registry->CheckLocked();
-  return static_cast<ThreadContext*>(
-      ctx->thread_registry->GetThreadLocked(tid));
+static ThreadContext *FindThreadByTidLocked(Tid tid) {
+  ctx->thread_registry.CheckLocked();
+  return static_cast<ThreadContext *>(
+      ctx->thread_registry.GetThreadLocked(tid));
 }
 
 static bool IsInStackOrTls(ThreadContextBase *tctx_base, void *arg) {
@@ -251,10 +249,10 @@ static bool IsInStackOrTls(ThreadContextBase *tctx_base, void *arg) {
 }
 
 ThreadContext *IsThreadStackOrTls(uptr addr, bool *is_stack) {
-  ctx->thread_registry->CheckLocked();
-  ThreadContext *tctx = static_cast<ThreadContext*>(
-      ctx->thread_registry->FindThreadContextLocked(IsInStackOrTls,
-                                                    (void*)addr));
+  ctx->thread_registry.CheckLocked();
+  ThreadContext *tctx =
+      static_cast<ThreadContext *>(ctx->thread_registry.FindThreadContextLocked(
+          IsInStackOrTls, (void *)addr));
   if (!tctx)
     return 0;
   ThreadState *thr = tctx->thr;
@@ -264,7 +262,7 @@ ThreadContext *IsThreadStackOrTls(uptr addr, bool *is_stack) {
 }
 #endif
 
-void ScopedReportBase::AddThread(int unique_tid, bool suppressable) {
+void ScopedReportBase::AddThread(Tid unique_tid, bool suppressable) {
 #if !SANITIZER_GO
   if (const ThreadContext *tctx = FindThreadByUidLocked(unique_tid))
     AddThread(tctx, suppressable);
@@ -276,8 +274,7 @@ void ScopedReportBase::AddMutex(const SyncVar *s) {
     if (rep_->mutexes[i]->id == s->uid)
       return;
   }
-  void *mem = internal_alloc(MBlockReportMutex, sizeof(ReportMutex));
-  ReportMutex *rm = new(mem) ReportMutex;
+  auto *rm = New<ReportMutex>();
   rep_->mutexes.PushBack(rm);
   rm->id = s->uid;
   rm->addr = s->addr;
@@ -285,22 +282,21 @@ void ScopedReportBase::AddMutex(const SyncVar *s) {
   rm->stack = SymbolizeStackId(s->creation_stack_id);
 }
 
-u64 ScopedReportBase::AddMutex(u64 id) NO_THREAD_SAFETY_ANALYSIS {
+u64 ScopedReportBase::AddMutex(u64 id) {
   u64 uid = 0;
   u64 mid = id;
   uptr addr = SyncVar::SplitId(id, &uid);
-  SyncVar *s = ctx->metamap.GetIfExistsAndLock(addr, true);
+  SyncVar *s = ctx->metamap.GetSyncIfExists(addr);
   // Check that the mutex is still alive.
   // Another mutex can be created at the same address,
   // so check uid as well.
   if (s && s->CheckId(uid)) {
+    Lock l(&s->mtx);
     mid = s->uid;
     AddMutex(s);
   } else {
     AddDeadMutex(id);
   }
-  if (s)
-    s->mtx.Unlock();
   return mid;
 }
 
@@ -309,8 +305,7 @@ void ScopedReportBase::AddDeadMutex(u64 id) {
     if (rep_->mutexes[i]->id == id)
       return;
   }
-  void *mem = internal_alloc(MBlockReportMutex, sizeof(ReportMutex));
-  ReportMutex *rm = new(mem) ReportMutex;
+  auto *rm = New<ReportMutex>();
   rep_->mutexes.PushBack(rm);
   rm->id = id;
   rm->addr = 0;
@@ -323,10 +318,11 @@ void ScopedReportBase::AddLocation(uptr addr, uptr size) {
     return;
 #if !SANITIZER_GO
   int fd = -1;
-  int creat_tid = kInvalidTid;
-  u32 creat_stack = 0;
+  Tid creat_tid = kInvalidTid;
+  StackID creat_stack = 0;
   if (FdLocation(addr, &fd, &creat_tid, &creat_stack)) {
-    ReportLocation *loc = ReportLocation::New(ReportLocationFD);
+    auto *loc = New<ReportLocation>();
+    loc->type = ReportLocationFD;
     loc->fd = fd;
     loc->tid = creat_tid;
     loc->stack = SymbolizeStackId(creat_stack);
@@ -337,15 +333,19 @@ void ScopedReportBase::AddLocation(uptr addr, uptr size) {
     return;
   }
   MBlock *b = 0;
+  uptr block_begin = 0;
   Allocator *a = allocator();
   if (a->PointerIsMine((void*)addr)) {
-    void *block_begin = a->GetBlockBegin((void*)addr);
+    block_begin = (uptr)a->GetBlockBegin((void *)addr);
     if (block_begin)
-      b = ctx->metamap.GetBlock((uptr)block_begin);
+      b = ctx->metamap.GetBlock(block_begin);
   }
+  if (!b)
+    b = JavaHeapBlock(addr, &block_begin);
   if (b != 0) {
     ThreadContext *tctx = FindThreadByTidLocked(b->tid);
-    ReportLocation *loc = ReportLocation::New(ReportLocationHeap);
+    auto *loc = New<ReportLocation>();
+    loc->type = ReportLocationHeap;
     loc->heap_chunk_start = (uptr)allocator()->GetBlockBegin((void *)addr);
     loc->heap_chunk_size = b->siz;
     loc->external_tag = b->tag;
@@ -358,8 +358,8 @@ void ScopedReportBase::AddLocation(uptr addr, uptr size) {
   }
   bool is_stack = false;
   if (ThreadContext *tctx = IsThreadStackOrTls(addr, &is_stack)) {
-    ReportLocation *loc =
-        ReportLocation::New(is_stack ? ReportLocationStack : ReportLocationTLS);
+    auto *loc = New<ReportLocation>();
+    loc->type = is_stack ? ReportLocationStack : ReportLocationTLS;
     loc->tid = tctx->tid;
     rep_->locs.PushBack(loc);
     AddThread(tctx);
@@ -373,7 +373,7 @@ void ScopedReportBase::AddLocation(uptr addr, uptr size) {
 }
 
 #if !SANITIZER_GO
-void ScopedReportBase::AddSleep(u32 stack_id) {
+void ScopedReportBase::AddSleep(StackID stack_id) {
   rep_->sleep = SymbolizeStackId(stack_id);
 }
 #endif
@@ -387,7 +387,7 @@ ScopedReport::ScopedReport(ReportType typ, uptr tag)
 
 ScopedReport::~ScopedReport() {}
 
-void RestoreStack(int tid, const u64 epoch, VarSizeStackTrace *stk,
+void RestoreStack(Tid tid, const u64 epoch, VarSizeStackTrace *stk,
                   MutexSet *mset, uptr *tag) {
   // This function restores stack trace and mutex set for the thread/epoch.
   // It does so by getting stack trace and mutex set at the beginning of
@@ -450,6 +450,232 @@ void RestoreStack(int tid, const u64 epoch, VarSizeStackTrace *stk,
   ExtractTagFromStack(stk, tag);
 }
 
+namespace v3 {
+
+// Replays the trace up to last_pos position in the last part
+// or up to the provided epoch/sid (whichever is earlier)
+// and calls the provided function f for each event.
+template <typename Func>
+void TraceReplay(Trace *trace, TracePart *last, Event *last_pos, Sid sid,
+                 Epoch epoch, Func f) {
+  TracePart *part = trace->parts.Front();
+  Sid ev_sid = kFreeSid;
+  Epoch ev_epoch = kEpochOver;
+  for (;;) {
+    DCHECK_EQ(part->trace, trace);
+    // Note: an event can't start in the last element.
+    // Since an event can take up to 2 elements,
+    // we ensure we have at least 2 before adding an event.
+    Event *end = &part->events[TracePart::kSize - 1];
+    if (part == last)
+      end = last_pos;
+    for (Event *evp = &part->events[0]; evp < end; evp++) {
+      Event *evp0 = evp;
+      if (!evp->is_access && !evp->is_func) {
+        switch (evp->type) {
+          case EventType::kTime: {
+            auto *ev = reinterpret_cast<EventTime *>(evp);
+            ev_sid = static_cast<Sid>(ev->sid);
+            ev_epoch = static_cast<Epoch>(ev->epoch);
+            if (ev_sid == sid && ev_epoch > epoch)
+              return;
+            break;
+          }
+          case EventType::kAccessExt:
+            FALLTHROUGH;
+          case EventType::kAccessRange:
+            FALLTHROUGH;
+          case EventType::kLock:
+            FALLTHROUGH;
+          case EventType::kRLock:
+            // These take 2 Event elements.
+            evp++;
+            break;
+          case EventType::kUnlock:
+            // This takes 1 Event element.
+            break;
+        }
+      }
+      CHECK_NE(ev_sid, kFreeSid);
+      CHECK_NE(ev_epoch, kEpochOver);
+      f(ev_sid, ev_epoch, evp0);
+    }
+    if (part == last)
+      return;
+    part = trace->parts.Next(part);
+    CHECK(part);
+  }
+  CHECK(0);
+}
+
+static void RestoreStackMatch(VarSizeStackTrace *pstk, MutexSet *pmset,
+                              Vector<uptr> *stack, MutexSet *mset, uptr pc,
+                              bool *found) {
+  DPrintf2("    MATCHED\n");
+  *pmset = *mset;
+  stack->PushBack(pc);
+  pstk->Init(&(*stack)[0], stack->Size());
+  stack->PopBack();
+  *found = true;
+}
+
+// Checks if addr1|size1 is fully contained in addr2|size2.
+// We check for fully contained instread of just overlapping
+// because a memory access is always traced once, but can be
+// split into multiple accesses in the shadow.
+static constexpr bool IsWithinAccess(uptr addr1, uptr size1, uptr addr2,
+                                     uptr size2) {
+  return addr1 >= addr2 && addr1 + size1 <= addr2 + size2;
+}
+
+// Replays the trace of thread tid up to the target event identified
+// by sid/epoch/addr/size/typ and restores and returns stack, mutex set
+// and tag for that event. If there are multiple such events, it returns
+// the last one. Returns false if the event is not present in the trace.
+bool RestoreStack(Tid tid, EventType type, Sid sid, Epoch epoch, uptr addr,
+                  uptr size, AccessType typ, VarSizeStackTrace *pstk,
+                  MutexSet *pmset, uptr *ptag) {
+  // This function restores stack trace and mutex set for the thread/epoch.
+  // It does so by getting stack trace and mutex set at the beginning of
+  // trace part, and then replaying the trace till the given epoch.
+  DPrintf2("RestoreStack: tid=%u sid=%u@%u addr=0x%zx/%zu typ=%x\n", tid,
+           static_cast<int>(sid), static_cast<int>(epoch), addr, size,
+           static_cast<int>(typ));
+  ctx->slot_mtx.CheckLocked();  // needed to prevent trace part recycling
+  ctx->thread_registry.CheckLocked();
+  ThreadContext *tctx =
+      static_cast<ThreadContext *>(ctx->thread_registry.GetThreadLocked(tid));
+  Trace *trace = &tctx->trace;
+  // Snapshot first/last parts and the current position in the last part.
+  TracePart *first_part;
+  TracePart *last_part;
+  Event *last_pos;
+  {
+    Lock lock(&trace->mtx);
+    first_part = trace->parts.Front();
+    if (!first_part)
+      return false;
+    last_part = trace->parts.Back();
+    last_pos = trace->final_pos;
+    if (tctx->thr)
+      last_pos = (Event *)atomic_load_relaxed(&tctx->thr->trace_pos);
+  }
+  DynamicMutexSet mset;
+  Vector<uptr> stack;
+  uptr prev_pc = 0;
+  bool found = false;
+  bool is_read = typ & kAccessRead;
+  bool is_atomic = typ & kAccessAtomic;
+  bool is_free = typ & kAccessFree;
+  TraceReplay(
+      trace, last_part, last_pos, sid, epoch,
+      [&](Sid ev_sid, Epoch ev_epoch, Event *evp) {
+        bool match = ev_sid == sid && ev_epoch == epoch;
+        if (evp->is_access) {
+          if (evp->is_func == 0 && evp->type == EventType::kAccessExt &&
+              evp->_ == 0)  // NopEvent
+            return;
+          auto *ev = reinterpret_cast<EventAccess *>(evp);
+          uptr ev_addr = RestoreAddr(ev->addr);
+          uptr ev_size = 1 << ev->size_log;
+          uptr ev_pc =
+              prev_pc + ev->pc_delta - (1 << (EventAccess::kPCBits - 1));
+          prev_pc = ev_pc;
+          DPrintf2("  Access: pc=0x%zx addr=0x%zx/%zu type=%u/%u\n", ev_pc,
+                   ev_addr, ev_size, ev->is_read, ev->is_atomic);
+          if (match && type == EventType::kAccessExt &&
+              IsWithinAccess(addr, size, ev_addr, ev_size) &&
+              is_read == ev->is_read && is_atomic == ev->is_atomic && !is_free)
+            RestoreStackMatch(pstk, pmset, &stack, mset, ev_pc, &found);
+          return;
+        }
+        if (evp->is_func) {
+          auto *ev = reinterpret_cast<EventFunc *>(evp);
+          if (ev->pc) {
+            DPrintf2("  FuncEnter: pc=0x%llx\n", ev->pc);
+            stack.PushBack(ev->pc);
+          } else {
+            DPrintf2("  FuncExit\n");
+            CHECK(stack.Size());
+            stack.PopBack();
+          }
+          return;
+        }
+        switch (evp->type) {
+          case EventType::kAccessExt: {
+            auto *ev = reinterpret_cast<EventAccessExt *>(evp);
+            uptr ev_addr = RestoreAddr(ev->addr);
+            uptr ev_size = 1 << ev->size_log;
+            prev_pc = ev->pc;
+            DPrintf2("  AccessExt: pc=0x%llx addr=0x%zx/%zu type=%u/%u\n",
+                     ev->pc, ev_addr, ev_size, ev->is_read, ev->is_atomic);
+            if (match && type == EventType::kAccessExt &&
+                IsWithinAccess(addr, size, ev_addr, ev_size) &&
+                is_read == ev->is_read && is_atomic == ev->is_atomic &&
+                !is_free)
+              RestoreStackMatch(pstk, pmset, &stack, mset, ev->pc, &found);
+            break;
+          }
+          case EventType::kAccessRange: {
+            auto *ev = reinterpret_cast<EventAccessRange *>(evp);
+            uptr ev_addr = RestoreAddr(ev->addr);
+            uptr ev_size =
+                (ev->size_hi << EventAccessRange::kSizeLoBits) + ev->size_lo;
+            uptr ev_pc = RestoreAddr(ev->pc);
+            prev_pc = ev_pc;
+            DPrintf2("  Range: pc=0x%zx addr=0x%zx/%zu type=%u/%u\n", ev_pc,
+                     ev_addr, ev_size, ev->is_read, ev->is_free);
+            if (match && type == EventType::kAccessExt &&
+                IsWithinAccess(addr, size, ev_addr, ev_size) &&
+                is_read == ev->is_read && !is_atomic && is_free == ev->is_free)
+              RestoreStackMatch(pstk, pmset, &stack, mset, ev_pc, &found);
+            break;
+          }
+          case EventType::kLock:
+            FALLTHROUGH;
+          case EventType::kRLock: {
+            auto *ev = reinterpret_cast<EventLock *>(evp);
+            bool is_write = ev->type == EventType::kLock;
+            uptr ev_addr = RestoreAddr(ev->addr);
+            uptr ev_pc = RestoreAddr(ev->pc);
+            StackID stack_id =
+                (ev->stack_hi << EventLock::kStackIDLoBits) + ev->stack_lo;
+            DPrintf2("  Lock: pc=0x%zx addr=0x%zx stack=%u write=%d\n", ev_pc,
+                     ev_addr, stack_id, is_write);
+            mset->AddAddr(ev_addr, stack_id, is_write);
+            // Events with ev_pc == 0 are written to the beginning of trace
+            // part as initial mutex set (are not real).
+            if (match && type == EventType::kLock && addr == ev_addr && ev_pc)
+              RestoreStackMatch(pstk, pmset, &stack, mset, ev_pc, &found);
+            break;
+          }
+          case EventType::kUnlock: {
+            auto *ev = reinterpret_cast<EventUnlock *>(evp);
+            uptr ev_addr = RestoreAddr(ev->addr);
+            DPrintf2("  Unlock: addr=0x%zx\n", ev_addr);
+            mset->DelAddr(ev_addr);
+            break;
+          }
+          case EventType::kTime:
+            // TraceReplay already extracted sid/epoch from it,
+            // nothing else to do here.
+            break;
+        }
+      });
+  ExtractTagFromStack(pstk, ptag);
+  return found;
+}
+
+}  // namespace v3
+
+bool RacyStacks::operator==(const RacyStacks &other) const {
+  if (hash[0] == other.hash[0] && hash[1] == other.hash[1])
+    return true;
+  if (hash[0] == other.hash[1] && hash[1] == other.hash[0])
+    return true;
+  return false;
+}
+
 static bool FindRacyStacks(const RacyStacks &hash) {
   for (uptr i = 0; i < ctx->racy_stacks.Size(); i++) {
     if (hash == ctx->racy_stacks[i]) {
@@ -614,7 +840,7 @@ void ReportRace(ThreadState *thr) {
     thr->racy_state[1] = s.raw();
   }
 
-  uptr addr = ShadowToMem((uptr)thr->racy_shadow_addr);
+  uptr addr = ShadowToMem(thr->racy_shadow_addr);
   uptr addr_min = 0;
   uptr addr_max = 0;
   {
@@ -669,11 +895,7 @@ void ReportRace(ThreadState *thr) {
   if (IsFiredSuppression(ctx, typ, traces[0]))
     return;
 
-  // MutexSet is too large to live on stack.
-  Vector<u64> mset_buffer;
-  mset_buffer.Resize(sizeof(MutexSet) / sizeof(u64) + 1);
-  MutexSet *mset2 = new(&mset_buffer[0]) MutexSet();
-
+  DynamicMutexSet mset2;
   Shadow s2(thr->racy_state[1]);
   RestoreStack(s2.tid(), s2.epoch(), &traces[1], mset2, &tags[1]);
   if (IsFiredSuppression(ctx, typ, traces[1]))
@@ -692,7 +914,7 @@ void ReportRace(ThreadState *thr) {
     }
   }
 
-  ThreadRegistryLock l0(ctx->thread_registry);
+  ThreadRegistryLock l0(&ctx->thread_registry);
   ScopedReport rep(typ, tag);
   for (uptr i = 0; i < kMop; i++) {
     Shadow s(thr->racy_state[i]);
@@ -702,8 +924,8 @@ void ReportRace(ThreadState *thr) {
 
   for (uptr i = 0; i < kMop; i++) {
     FastState s(thr->racy_state[i]);
-    ThreadContext *tctx = static_cast<ThreadContext*>(
-        ctx->thread_registry->GetThreadLocked(s.tid()));
+    ThreadContext *tctx = static_cast<ThreadContext *>(
+        ctx->thread_registry.GetThreadLocked(s.tid()));
     if (s.epoch() < tctx->epoch0 || s.epoch() > tctx->epoch1)
       continue;
     rep.AddThread(tctx);
@@ -738,9 +960,7 @@ void PrintCurrentStack(ThreadState *thr, uptr pc) {
 ALWAYS_INLINE USED void PrintCurrentStackSlow(uptr pc) {
 #if !SANITIZER_GO
   uptr bp = GET_CURRENT_FRAME();
-  BufferedStackTrace *ptrace =
-      new(internal_alloc(MBlockStackTrace, sizeof(BufferedStackTrace)))
-          BufferedStackTrace();
+  auto *ptrace = New<BufferedStackTrace>();
   ptrace->Unwind(pc, bp, nullptr, false);
 
   for (uptr i = 0; i < ptrace->size / 2; i++) {
diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cpp b/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cpp
index cdb6e60ebbd0..c8f7124c009d 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cpp
@@ -21,48 +21,14 @@ namespace __tsan {
 
 // ThreadContext implementation.
 
-ThreadContext::ThreadContext(int tid)
-  : ThreadContextBase(tid)
-  , thr()
-  , sync()
-  , epoch0()
-  , epoch1() {
-}
+ThreadContext::ThreadContext(Tid tid)
+    : ThreadContextBase(tid), thr(), sync(), epoch0(), epoch1() {}
 
 #if !SANITIZER_GO
 ThreadContext::~ThreadContext() {
 }
 #endif
 
-void ThreadContext::OnDead() {
-  CHECK_EQ(sync.size(), 0);
-}
-
-void ThreadContext::OnJoined(void *arg) {
-  ThreadState *caller_thr = static_cast<ThreadState *>(arg);
-  AcquireImpl(caller_thr, 0, &sync);
-  sync.Reset(&caller_thr->proc()->clock_cache);
-}
-
-struct OnCreatedArgs {
-  ThreadState *thr;
-  uptr pc;
-};
-
-void ThreadContext::OnCreated(void *arg) {
-  thr = 0;
-  if (tid == kMainTid)
-    return;
-  OnCreatedArgs *args = static_cast<OnCreatedArgs *>(arg);
-  if (!args->thr)  // GCD workers don't have a parent thread.
-    return;
-  args->thr->fast_state.IncrementEpoch();
-  // Can't increment epoch w/o writing to the trace as well.
-  TraceAddEvent(args->thr, args->thr->fast_state, EventTypeMop, 0);
-  ReleaseImpl(args->thr, 0, &sync);
-  creation_stack_id = CurrentStackId(args->thr, args->pc);
-}
-
 void ThreadContext::OnReset() {
   CHECK_EQ(sync.size(), 0);
   uptr trace_p = GetThreadTrace(tid);
@@ -70,94 +36,15 @@ void ThreadContext::OnReset() {
   //!!! ReleaseMemoryToOS(GetThreadTraceHeader(tid), sizeof(Trace));
 }
 
-void ThreadContext::OnDetached(void *arg) {
-  ThreadState *thr1 = static_cast<ThreadState*>(arg);
-  sync.Reset(&thr1->proc()->clock_cache);
-}
-
-struct OnStartedArgs {
-  ThreadState *thr;
-  uptr stk_addr;
-  uptr stk_size;
-  uptr tls_addr;
-  uptr tls_size;
-};
-
-void ThreadContext::OnStarted(void *arg) {
-  OnStartedArgs *args = static_cast<OnStartedArgs*>(arg);
-  thr = args->thr;
-  // RoundUp so that one trace part does not contain events
-  // from different threads.
-  epoch0 = RoundUp(epoch1 + 1, kTracePartSize);
-  epoch1 = (u64)-1;
-  new(thr) ThreadState(ctx, tid, unique_id, epoch0, reuse_count,
-      args->stk_addr, args->stk_size, args->tls_addr, args->tls_size);
-#if !SANITIZER_GO
-  thr->shadow_stack = &ThreadTrace(thr->tid)->shadow_stack[0];
-  thr->shadow_stack_pos = thr->shadow_stack;
-  thr->shadow_stack_end = thr->shadow_stack + kShadowStackSize;
-#else
-  // Setup dynamic shadow stack.
-  const int kInitStackSize = 8;
-  thr->shadow_stack = (uptr*)internal_alloc(MBlockShadowStack,
-      kInitStackSize * sizeof(uptr));
-  thr->shadow_stack_pos = thr->shadow_stack;
-  thr->shadow_stack_end = thr->shadow_stack + kInitStackSize;
-#endif
-  if (common_flags()->detect_deadlocks)
-    thr->dd_lt = ctx->dd->CreateLogicalThread(unique_id);
-  thr->fast_state.SetHistorySize(flags()->history_size);
-  // Commit switch to the new part of the trace.
-  // TraceAddEvent will reset stack0/mset0 in the new part for us.
-  TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
-
-  thr->fast_synch_epoch = epoch0;
-  AcquireImpl(thr, 0, &sync);
-  sync.Reset(&thr->proc()->clock_cache);
-  thr->is_inited = true;
-  DPrintf("#%d: ThreadStart epoch=%zu stk_addr=%zx stk_size=%zx "
-          "tls_addr=%zx tls_size=%zx\n",
-          tid, (uptr)epoch0, args->stk_addr, args->stk_size,
-          args->tls_addr, args->tls_size);
-}
-
-void ThreadContext::OnFinished() {
-#if SANITIZER_GO
-  internal_free(thr->shadow_stack);
-  thr->shadow_stack = nullptr;
-  thr->shadow_stack_pos = nullptr;
-  thr->shadow_stack_end = nullptr;
-#endif
-  if (!detached) {
-    thr->fast_state.IncrementEpoch();
-    // Can't increment epoch w/o writing to the trace as well.
-    TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
-    ReleaseImpl(thr, 0, &sync);
-  }
-  epoch1 = thr->fast_state.epoch();
-
-  if (common_flags()->detect_deadlocks)
-    ctx->dd->DestroyLogicalThread(thr->dd_lt);
-  thr->clock.ResetCached(&thr->proc()->clock_cache);
-#if !SANITIZER_GO
-  thr->last_sleep_clock.ResetCached(&thr->proc()->clock_cache);
-#endif
-#if !SANITIZER_GO
-  PlatformCleanUpThreadState(thr);
-#endif
-  thr->~ThreadState();
-  thr = 0;
-}
-
 #if !SANITIZER_GO
 struct ThreadLeak {
   ThreadContext *tctx;
   int count;
 };
 
-static void MaybeReportThreadLeak(ThreadContextBase *tctx_base, void *arg) {
-  Vector<ThreadLeak> &leaks = *(Vector<ThreadLeak>*)arg;
-  ThreadContext *tctx = static_cast<ThreadContext*>(tctx_base);
+static void CollectThreadLeaks(ThreadContextBase *tctx_base, void *arg) {
+  auto &leaks = *static_cast<Vector<ThreadLeak> *>(arg);
+  auto *tctx = static_cast<ThreadContext *>(tctx_base);
   if (tctx->detached || tctx->status != ThreadStatusFinished)
     return;
   for (uptr i = 0; i < leaks.Size(); i++) {
@@ -166,8 +53,7 @@ static void MaybeReportThreadLeak(ThreadContextBase *tctx_base, void *arg) {
       return;
     }
   }
-  ThreadLeak leak = {tctx, 1};
-  leaks.PushBack(leak);
+  leaks.PushBack({tctx, 1});
 }
 #endif
 
@@ -206,10 +92,10 @@ void ThreadFinalize(ThreadState *thr) {
 #if !SANITIZER_GO
   if (!ShouldReport(thr, ReportTypeThreadLeak))
     return;
-  ThreadRegistryLock l(ctx->thread_registry);
+  ThreadRegistryLock l(&ctx->thread_registry);
   Vector<ThreadLeak> leaks;
-  ctx->thread_registry->RunCallbackForEachThreadLocked(
-      MaybeReportThreadLeak, &leaks);
+  ctx->thread_registry.RunCallbackForEachThreadLocked(CollectThreadLeaks,
+                                                      &leaks);
   for (uptr i = 0; i < leaks.Size(); i++) {
     ScopedReport rep(ReportTypeThreadLeak);
     rep.AddThread(leaks[i].tctx, true);
@@ -221,20 +107,48 @@ void ThreadFinalize(ThreadState *thr) {
 
 int ThreadCount(ThreadState *thr) {
   uptr result;
-  ctx->thread_registry->GetNumberOfThreads(0, 0, &result);
+  ctx->thread_registry.GetNumberOfThreads(0, 0, &result);
   return (int)result;
 }
 
-int ThreadCreate(ThreadState *thr, uptr pc, uptr uid, bool detached) {
+struct OnCreatedArgs {
+  ThreadState *thr;
+  uptr pc;
+};
+
+Tid ThreadCreate(ThreadState *thr, uptr pc, uptr uid, bool detached) {
   OnCreatedArgs args = { thr, pc };
   u32 parent_tid = thr ? thr->tid : kInvalidTid;  // No parent for GCD workers.
-  int tid =
-      ctx->thread_registry->CreateThread(uid, detached, parent_tid, &args);
+  Tid tid = ctx->thread_registry.CreateThread(uid, detached, parent_tid, &args);
   DPrintf("#%d: ThreadCreate tid=%d uid=%zu\n", parent_tid, tid, uid);
   return tid;
 }
 
-void ThreadStart(ThreadState *thr, int tid, tid_t os_id,
+void ThreadContext::OnCreated(void *arg) {
+  thr = 0;
+  if (tid == kMainTid)
+    return;
+  OnCreatedArgs *args = static_cast<OnCreatedArgs *>(arg);
+  if (!args->thr)  // GCD workers don't have a parent thread.
+    return;
+  args->thr->fast_state.IncrementEpoch();
+  // Can't increment epoch w/o writing to the trace as well.
+  TraceAddEvent(args->thr, args->thr->fast_state, EventTypeMop, 0);
+  ReleaseImpl(args->thr, 0, &sync);
+  creation_stack_id = CurrentStackId(args->thr, args->pc);
+}
+
+extern "C" void __tsan_stack_initialization() {}
+
+struct OnStartedArgs {
+  ThreadState *thr;
+  uptr stk_addr;
+  uptr stk_size;
+  uptr tls_addr;
+  uptr tls_size;
+};
+
+void ThreadStart(ThreadState *thr, Tid tid, tid_t os_id,
                  ThreadType thread_type) {
   uptr stk_addr = 0;
   uptr stk_size = 0;
@@ -244,22 +158,13 @@ void ThreadStart(ThreadState *thr, int tid, tid_t os_id,
   if (thread_type != ThreadType::Fiber)
     GetThreadStackAndTls(tid == kMainTid, &stk_addr, &stk_size, &tls_addr,
                          &tls_size);
-
-  if (tid != kMainTid) {
-    if (stk_addr && stk_size)
-      MemoryRangeImitateWrite(thr, /*pc=*/ 1, stk_addr, stk_size);
-
-    if (tls_addr && tls_size) ImitateTlsWrite(thr, tls_addr, tls_size);
-  }
 #endif
 
-  ThreadRegistry *tr = ctx->thread_registry;
+  ThreadRegistry *tr = &ctx->thread_registry;
   OnStartedArgs args = { thr, stk_addr, stk_size, tls_addr, tls_size };
   tr->StartThread(tid, os_id, thread_type, &args);
 
-  tr->Lock();
-  thr->tctx = (ThreadContext*)tr->GetThreadLocked(tid);
-  tr->Unlock();
+  while (!thr->tctx->trace.parts.Empty()) thr->tctx->trace.parts.PopBack();
 
 #if !SANITIZER_GO
   if (ctx->after_multithreaded_fork) {
@@ -268,6 +173,51 @@ void ThreadStart(ThreadState *thr, int tid, tid_t os_id,
     ThreadIgnoreSyncBegin(thr, 0);
   }
 #endif
+
+#if !SANITIZER_GO
+  // Don't imitate stack/TLS writes for the main thread,
+  // because its initialization is synchronized with all
+  // subsequent threads anyway.
+  if (tid != kMainTid) {
+    if (stk_addr && stk_size) {
+      const uptr pc = StackTrace::GetNextInstructionPc(
+          reinterpret_cast<uptr>(__tsan_stack_initialization));
+      MemoryRangeImitateWrite(thr, pc, stk_addr, stk_size);
+    }
+
+    if (tls_addr && tls_size)
+      ImitateTlsWrite(thr, tls_addr, tls_size);
+  }
+#endif
+}
+
+void ThreadContext::OnStarted(void *arg) {
+  OnStartedArgs *args = static_cast<OnStartedArgs *>(arg);
+  thr = args->thr;
+  // RoundUp so that one trace part does not contain events
+  // from different threads.
+  epoch0 = RoundUp(epoch1 + 1, kTracePartSize);
+  epoch1 = (u64)-1;
+  new (thr)
+      ThreadState(ctx, tid, unique_id, epoch0, reuse_count, args->stk_addr,
+                  args->stk_size, args->tls_addr, args->tls_size);
+  if (common_flags()->detect_deadlocks)
+    thr->dd_lt = ctx->dd->CreateLogicalThread(unique_id);
+  thr->fast_state.SetHistorySize(flags()->history_size);
+  // Commit switch to the new part of the trace.
+  // TraceAddEvent will reset stack0/mset0 in the new part for us.
+  TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
+
+  thr->fast_synch_epoch = epoch0;
+  AcquireImpl(thr, 0, &sync);
+  sync.Reset(&thr->proc()->clock_cache);
+  thr->tctx = this;
+  thr->is_inited = true;
+  DPrintf(
+      "#%d: ThreadStart epoch=%zu stk_addr=%zx stk_size=%zx "
+      "tls_addr=%zx tls_size=%zx\n",
+      tid, (uptr)epoch0, args->stk_addr, args->stk_size, args->tls_addr,
+      args->tls_size);
 }
 
 void ThreadFinish(ThreadState *thr) {
@@ -277,7 +227,42 @@ void ThreadFinish(ThreadState *thr) {
   if (thr->tls_addr && thr->tls_size)
     DontNeedShadowFor(thr->tls_addr, thr->tls_size);
   thr->is_dead = true;
-  ctx->thread_registry->FinishThread(thr->tid);
+  thr->is_inited = false;
+#if !SANITIZER_GO
+  thr->ignore_interceptors++;
+#endif
+  ctx->thread_registry.FinishThread(thr->tid);
+}
+
+void ThreadContext::OnFinished() {
+  if (!detached) {
+    thr->fast_state.IncrementEpoch();
+    // Can't increment epoch w/o writing to the trace as well.
+    TraceAddEvent(thr, thr->fast_state, EventTypeMop, 0);
+    ReleaseImpl(thr, 0, &sync);
+  }
+  epoch1 = thr->fast_state.epoch();
+
+#if !SANITIZER_GO
+  UnmapOrDie(thr->shadow_stack, kShadowStackSize * sizeof(uptr));
+#else
+  Free(thr->shadow_stack);
+#endif
+  thr->shadow_stack = nullptr;
+  thr->shadow_stack_pos = nullptr;
+  thr->shadow_stack_end = nullptr;
+
+  if (common_flags()->detect_deadlocks)
+    ctx->dd->DestroyLogicalThread(thr->dd_lt);
+  thr->clock.ResetCached(&thr->proc()->clock_cache);
+#if !SANITIZER_GO
+  thr->last_sleep_clock.ResetCached(&thr->proc()->clock_cache);
+#endif
+#if !SANITIZER_GO
+  PlatformCleanUpThreadState(thr);
+#endif
+  thr->~ThreadState();
+  thr = 0;
 }
 
 struct ConsumeThreadContext {
@@ -285,131 +270,44 @@ struct ConsumeThreadContext {
   ThreadContextBase *tctx;
 };
 
-static bool ConsumeThreadByUid(ThreadContextBase *tctx, void *arg) {
-  ConsumeThreadContext *findCtx = (ConsumeThreadContext *)arg;
-  if (tctx->user_id == findCtx->uid && tctx->status != ThreadStatusInvalid) {
-    if (findCtx->tctx) {
-      // Ensure that user_id is unique. If it's not the case we are screwed.
-      // Something went wrong before, but now there is no way to recover.
-      // Returning a wrong thread is not an option, it may lead to very hard
-      // to debug false positives (e.g. if we join a wrong thread).
-      Report("ThreadSanitizer: dup thread with used id 0x%zx\n", findCtx->uid);
-      Die();
-    }
-    findCtx->tctx = tctx;
-    tctx->user_id = 0;
-  }
-  return false;
-}
-
-int ThreadConsumeTid(ThreadState *thr, uptr pc, uptr uid) {
-  ConsumeThreadContext findCtx = {uid, nullptr};
-  ctx->thread_registry->FindThread(ConsumeThreadByUid, &findCtx);
-  int tid = findCtx.tctx ? findCtx.tctx->tid : kInvalidTid;
-  DPrintf("#%d: ThreadTid uid=%zu tid=%d\n", thr->tid, uid, tid);
-  return tid;
+Tid ThreadConsumeTid(ThreadState *thr, uptr pc, uptr uid) {
+  return ctx->thread_registry.ConsumeThreadUserId(uid);
 }
 
-void ThreadJoin(ThreadState *thr, uptr pc, int tid) {
+void ThreadJoin(ThreadState *thr, uptr pc, Tid tid) {
   CHECK_GT(tid, 0);
   CHECK_LT(tid, kMaxTid);
   DPrintf("#%d: ThreadJoin tid=%d\n", thr->tid, tid);
-  ctx->thread_registry->JoinThread(tid, thr);
+  ctx->thread_registry.JoinThread(tid, thr);
 }
 
-void ThreadDetach(ThreadState *thr, uptr pc, int tid) {
-  CHECK_GT(tid, 0);
-  CHECK_LT(tid, kMaxTid);
-  ctx->thread_registry->DetachThread(tid, thr);
+void ThreadContext::OnJoined(void *arg) {
+  ThreadState *caller_thr = static_cast<ThreadState *>(arg);
+  AcquireImpl(caller_thr, 0, &sync);
+  sync.Reset(&caller_thr->proc()->clock_cache);
 }
 
-void ThreadNotJoined(ThreadState *thr, uptr pc, int tid, uptr uid) {
+void ThreadContext::OnDead() { CHECK_EQ(sync.size(), 0); }
+
+void ThreadDetach(ThreadState *thr, uptr pc, Tid tid) {
   CHECK_GT(tid, 0);
   CHECK_LT(tid, kMaxTid);
-  ctx->thread_registry->SetThreadUserId(tid, uid);
+  ctx->thread_registry.DetachThread(tid, thr);
 }
 
-void ThreadSetName(ThreadState *thr, const char *name) {
-  ctx->thread_registry->SetThreadName(thr->tid, name);
+void ThreadContext::OnDetached(void *arg) {
+  ThreadState *thr1 = static_cast<ThreadState *>(arg);
+  sync.Reset(&thr1->proc()->clock_cache);
 }
 
-void MemoryAccessRange(ThreadState *thr, uptr pc, uptr addr,
-                       uptr size, bool is_write) {
-  if (size == 0)
-    return;
-
-  u64 *shadow_mem = (u64*)MemToShadow(addr);
-  DPrintf2("#%d: MemoryAccessRange: @%p %p size=%d is_write=%d\n",
-      thr->tid, (void*)pc, (void*)addr,
-      (int)size, is_write);
-
-#if SANITIZER_DEBUG
-  if (!IsAppMem(addr)) {
-    Printf("Access to non app mem %zx\n", addr);
-    DCHECK(IsAppMem(addr));
-  }
-  if (!IsAppMem(addr + size - 1)) {
-    Printf("Access to non app mem %zx\n", addr + size - 1);
-    DCHECK(IsAppMem(addr + size - 1));
-  }
-  if (!IsShadowMem((uptr)shadow_mem)) {
-    Printf("Bad shadow addr %p (%zx)\n", shadow_mem, addr);
-    DCHECK(IsShadowMem((uptr)shadow_mem));
-  }
-  if (!IsShadowMem((uptr)(shadow_mem + size * kShadowCnt / 8 - 1))) {
-    Printf("Bad shadow addr %p (%zx)\n",
-               shadow_mem + size * kShadowCnt / 8 - 1, addr + size - 1);
-    DCHECK(IsShadowMem((uptr)(shadow_mem + size * kShadowCnt / 8 - 1)));
-  }
-#endif
-
-  if (*shadow_mem == kShadowRodata) {
-    DCHECK(!is_write);
-    // Access to .rodata section, no races here.
-    // Measurements show that it can be 10-20% of all memory accesses.
-    return;
-  }
-
-  FastState fast_state = thr->fast_state;
-  if (fast_state.GetIgnoreBit())
-    return;
-
-  fast_state.IncrementEpoch();
-  thr->fast_state = fast_state;
-  TraceAddEvent(thr, fast_state, EventTypeMop, pc);
-
-  bool unaligned = (addr % kShadowCell) != 0;
+void ThreadNotJoined(ThreadState *thr, uptr pc, Tid tid, uptr uid) {
+  CHECK_GT(tid, 0);
+  CHECK_LT(tid, kMaxTid);
+  ctx->thread_registry.SetThreadUserId(tid, uid);
+}
 
-  // Handle unaligned beginning, if any.
-  for (; addr % kShadowCell && size; addr++, size--) {
-    int const kAccessSizeLog = 0;
-    Shadow cur(fast_state);
-    cur.SetWrite(is_write);
-    cur.SetAddr0AndSizeLog(addr & (kShadowCell - 1), kAccessSizeLog);
-    MemoryAccessImpl(thr, addr, kAccessSizeLog, is_write, false,
-        shadow_mem, cur);
-  }
-  if (unaligned)
-    shadow_mem += kShadowCnt;
-  // Handle middle part, if any.
-  for (; size >= kShadowCell; addr += kShadowCell, size -= kShadowCell) {
-    int const kAccessSizeLog = 3;
-    Shadow cur(fast_state);
-    cur.SetWrite(is_write);
-    cur.SetAddr0AndSizeLog(0, kAccessSizeLog);
-    MemoryAccessImpl(thr, addr, kAccessSizeLog, is_write, false,
-        shadow_mem, cur);
-    shadow_mem += kShadowCnt;
-  }
-  // Handle ending, if any.
-  for (; size; addr++, size--) {
-    int const kAccessSizeLog = 0;
-    Shadow cur(fast_state);
-    cur.SetWrite(is_write);
-    cur.SetAddr0AndSizeLog(addr & (kShadowCell - 1), kAccessSizeLog);
-    MemoryAccessImpl(thr, addr, kAccessSizeLog, is_write, false,
-        shadow_mem, cur);
-  }
+void ThreadSetName(ThreadState *thr, const char *name) {
+  ctx->thread_registry.SetThreadName(thr->tid, name);
 }
 
 #if !SANITIZER_GO
@@ -421,10 +319,10 @@ void FiberSwitchImpl(ThreadState *from, ThreadState *to) {
 }
 
 ThreadState *FiberCreate(ThreadState *thr, uptr pc, unsigned flags) {
-  void *mem = internal_alloc(MBlockThreadContex, sizeof(ThreadState));
+  void *mem = Alloc(sizeof(ThreadState));
   ThreadState *fiber = static_cast<ThreadState *>(mem);
   internal_memset(fiber, 0, sizeof(*fiber));
-  int tid = ThreadCreate(thr, pc, 0, true);
+  Tid tid = ThreadCreate(thr, pc, 0, true);
   FiberSwitchImpl(thr, fiber);
   ThreadStart(fiber, tid, 0, ThreadType::Fiber);
   FiberSwitchImpl(fiber, thr);
@@ -435,7 +333,7 @@ void FiberDestroy(ThreadState *thr, uptr pc, ThreadState *fiber) {
   FiberSwitchImpl(thr, fiber);
   ThreadFinish(fiber);
   FiberSwitchImpl(fiber, thr);
-  internal_free(fiber);
+  Free(fiber);
 }
 
 void FiberSwitch(ThreadState *thr, uptr pc,
diff --git a/compiler-rt/lib/tsan/rtl/tsan_shadow.h b/compiler-rt/lib/tsan/rtl/tsan_shadow.h
new file mode 100644
index 000000000000..8b7bc341713e
--- /dev/null
+++ b/compiler-rt/lib/tsan/rtl/tsan_shadow.h
@@ -0,0 +1,233 @@
+//===-- tsan_shadow.h -------------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef TSAN_SHADOW_H
+#define TSAN_SHADOW_H
+
+#include "tsan_defs.h"
+#include "tsan_trace.h"
+
+namespace __tsan {
+
+// FastState (from most significant bit):
+//   ignore          : 1
+//   tid             : kTidBits
+//   unused          : -
+//   history_size    : 3
+//   epoch           : kClkBits
+class FastState {
+ public:
+  FastState(u64 tid, u64 epoch) {
+    x_ = tid << kTidShift;
+    x_ |= epoch;
+    DCHECK_EQ(tid, this->tid());
+    DCHECK_EQ(epoch, this->epoch());
+    DCHECK_EQ(GetIgnoreBit(), false);
+  }
+
+  explicit FastState(u64 x) : x_(x) {}
+
+  u64 raw() const { return x_; }
+
+  u64 tid() const {
+    u64 res = (x_ & ~kIgnoreBit) >> kTidShift;
+    return res;
+  }
+
+  u64 TidWithIgnore() const {
+    u64 res = x_ >> kTidShift;
+    return res;
+  }
+
+  u64 epoch() const {
+    u64 res = x_ & ((1ull << kClkBits) - 1);
+    return res;
+  }
+
+  void IncrementEpoch() {
+    u64 old_epoch = epoch();
+    x_ += 1;
+    DCHECK_EQ(old_epoch + 1, epoch());
+    (void)old_epoch;
+  }
+
+  void SetIgnoreBit() { x_ |= kIgnoreBit; }
+  void ClearIgnoreBit() { x_ &= ~kIgnoreBit; }
+  bool GetIgnoreBit() const { return (s64)x_ < 0; }
+
+  void SetHistorySize(int hs) {
+    CHECK_GE(hs, 0);
+    CHECK_LE(hs, 7);
+    x_ = (x_ & ~(kHistoryMask << kHistoryShift)) | (u64(hs) << kHistoryShift);
+  }
+
+  ALWAYS_INLINE
+  int GetHistorySize() const {
+    return (int)((x_ >> kHistoryShift) & kHistoryMask);
+  }
+
+  void ClearHistorySize() { SetHistorySize(0); }
+
+  ALWAYS_INLINE
+  u64 GetTracePos() const {
+    const int hs = GetHistorySize();
+    // When hs == 0, the trace consists of 2 parts.
+    const u64 mask = (1ull << (kTracePartSizeBits + hs + 1)) - 1;
+    return epoch() & mask;
+  }
+
+ private:
+  friend class Shadow;
+  static const int kTidShift = 64 - kTidBits - 1;
+  static const u64 kIgnoreBit = 1ull << 63;
+  static const u64 kFreedBit = 1ull << 63;
+  static const u64 kHistoryShift = kClkBits;
+  static const u64 kHistoryMask = 7;
+  u64 x_;
+};
+
+// Shadow (from most significant bit):
+//   freed           : 1
+//   tid             : kTidBits
+//   is_atomic       : 1
+//   is_read         : 1
+//   size_log        : 2
+//   addr0           : 3
+//   epoch           : kClkBits
+class Shadow : public FastState {
+ public:
+  explicit Shadow(u64 x) : FastState(x) {}
+
+  explicit Shadow(const FastState &s) : FastState(s.x_) { ClearHistorySize(); }
+
+  void SetAddr0AndSizeLog(u64 addr0, unsigned kAccessSizeLog) {
+    DCHECK_EQ((x_ >> kClkBits) & 31, 0);
+    DCHECK_LE(addr0, 7);
+    DCHECK_LE(kAccessSizeLog, 3);
+    x_ |= ((kAccessSizeLog << 3) | addr0) << kClkBits;
+    DCHECK_EQ(kAccessSizeLog, size_log());
+    DCHECK_EQ(addr0, this->addr0());
+  }
+
+  void SetWrite(unsigned kAccessIsWrite) {
+    DCHECK_EQ(x_ & kReadBit, 0);
+    if (!kAccessIsWrite)
+      x_ |= kReadBit;
+    DCHECK_EQ(kAccessIsWrite, IsWrite());
+  }
+
+  void SetAtomic(bool kIsAtomic) {
+    DCHECK(!IsAtomic());
+    if (kIsAtomic)
+      x_ |= kAtomicBit;
+    DCHECK_EQ(IsAtomic(), kIsAtomic);
+  }
+
+  bool IsAtomic() const { return x_ & kAtomicBit; }
+
+  bool IsZero() const { return x_ == 0; }
+
+  static inline bool TidsAreEqual(const Shadow s1, const Shadow s2) {
+    u64 shifted_xor = (s1.x_ ^ s2.x_) >> kTidShift;
+    DCHECK_EQ(shifted_xor == 0, s1.TidWithIgnore() == s2.TidWithIgnore());
+    return shifted_xor == 0;
+  }
+
+  static ALWAYS_INLINE bool Addr0AndSizeAreEqual(const Shadow s1,
+                                                 const Shadow s2) {
+    u64 masked_xor = ((s1.x_ ^ s2.x_) >> kClkBits) & 31;
+    return masked_xor == 0;
+  }
+
+  static ALWAYS_INLINE bool TwoRangesIntersect(Shadow s1, Shadow s2,
+                                               unsigned kS2AccessSize) {
+    bool res = false;
+    u64 diff = s1.addr0() - s2.addr0();
+    if ((s64)diff < 0) {  // s1.addr0 < s2.addr0
+      // if (s1.addr0() + size1) > s2.addr0()) return true;
+      if (s1.size() > -diff)
+        res = true;
+    } else {
+      // if (s2.addr0() + kS2AccessSize > s1.addr0()) return true;
+      if (kS2AccessSize > diff)
+        res = true;
+    }
+    DCHECK_EQ(res, TwoRangesIntersectSlow(s1, s2));
+    DCHECK_EQ(res, TwoRangesIntersectSlow(s2, s1));
+    return res;
+  }
+
+  u64 ALWAYS_INLINE addr0() const { return (x_ >> kClkBits) & 7; }
+  u64 ALWAYS_INLINE size() const { return 1ull << size_log(); }
+  bool ALWAYS_INLINE IsWrite() const { return !IsRead(); }
+  bool ALWAYS_INLINE IsRead() const { return x_ & kReadBit; }
+
+  // The idea behind the freed bit is as follows.
+  // When the memory is freed (or otherwise unaccessible) we write to the shadow
+  // values with tid/epoch related to the free and the freed bit set.
+  // During memory accesses processing the freed bit is considered
+  // as msb of tid. So any access races with shadow with freed bit set
+  // (it is as if write from a thread with which we never synchronized before).
+  // This allows us to detect accesses to freed memory w/o additional
+  // overheads in memory access processing and at the same time restore
+  // tid/epoch of free.
+  void MarkAsFreed() { x_ |= kFreedBit; }
+
+  bool IsFreed() const { return x_ & kFreedBit; }
+
+  bool GetFreedAndReset() {
+    bool res = x_ & kFreedBit;
+    x_ &= ~kFreedBit;
+    return res;
+  }
+
+  bool ALWAYS_INLINE IsBothReadsOrAtomic(bool kIsWrite, bool kIsAtomic) const {
+    bool v = x_ & ((u64(kIsWrite ^ 1) << kReadShift) |
+                   (u64(kIsAtomic) << kAtomicShift));
+    DCHECK_EQ(v, (!IsWrite() && !kIsWrite) || (IsAtomic() && kIsAtomic));
+    return v;
+  }
+
+  bool ALWAYS_INLINE IsRWNotWeaker(bool kIsWrite, bool kIsAtomic) const {
+    bool v = ((x_ >> kReadShift) & 3) <= u64((kIsWrite ^ 1) | (kIsAtomic << 1));
+    DCHECK_EQ(v, (IsAtomic() < kIsAtomic) ||
+                     (IsAtomic() == kIsAtomic && !IsWrite() <= !kIsWrite));
+    return v;
+  }
+
+  bool ALWAYS_INLINE IsRWWeakerOrEqual(bool kIsWrite, bool kIsAtomic) const {
+    bool v = ((x_ >> kReadShift) & 3) >= u64((kIsWrite ^ 1) | (kIsAtomic << 1));
+    DCHECK_EQ(v, (IsAtomic() > kIsAtomic) ||
+                     (IsAtomic() == kIsAtomic && !IsWrite() >= !kIsWrite));
+    return v;
+  }
+
+ private:
+  static const u64 kReadShift = 5 + kClkBits;
+  static const u64 kReadBit = 1ull << kReadShift;
+  static const u64 kAtomicShift = 6 + kClkBits;
+  static const u64 kAtomicBit = 1ull << kAtomicShift;
+
+  u64 size_log() const { return (x_ >> (3 + kClkBits)) & 3; }
+
+  static bool TwoRangesIntersectSlow(const Shadow s1, const Shadow s2) {
+    if (s1.addr0() == s2.addr0())
+      return true;
+    if (s1.addr0() < s2.addr0() && s1.addr0() + s1.size() > s2.addr0())
+      return true;
+    if (s2.addr0() < s1.addr0() && s2.addr0() + s2.size() > s1.addr0())
+      return true;
+    return false;
+  }
+};
+
+const RawShadow kShadowRodata = (RawShadow)-1;  // .rodata shadow marker
+
+}  // namespace __tsan
+
+#endif
diff --git a/compiler-rt/lib/tsan/rtl/tsan_stack_trace.cpp b/compiler-rt/lib/tsan/rtl/tsan_stack_trace.cpp
index 6c703d7f2b10..9bbaafb3a85f 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_stack_trace.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_stack_trace.cpp
@@ -23,14 +23,10 @@ VarSizeStackTrace::~VarSizeStackTrace() {
 }
 
 void VarSizeStackTrace::ResizeBuffer(uptr new_size) {
-  if (trace_buffer) {
-    internal_free(trace_buffer);
-  }
-  trace_buffer =
-      (new_size > 0)
-          ? (uptr *)internal_alloc(MBlockStackTrace,
-                                   new_size * sizeof(trace_buffer[0]))
-          : nullptr;
+  Free(trace_buffer);
+  trace_buffer = (new_size > 0)
+                     ? (uptr *)Alloc(new_size * sizeof(trace_buffer[0]))
+                     : nullptr;
   trace = trace_buffer;
   size = new_size;
 }
diff --git a/compiler-rt/lib/tsan/rtl/tsan_symbolize.cpp b/compiler-rt/lib/tsan/rtl/tsan_symbolize.cpp
index 6478f3a754ac..2e2744d2eae7 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_symbolize.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_symbolize.cpp
@@ -110,7 +110,8 @@ ReportLocation *SymbolizeData(uptr addr) {
   DataInfo info;
   if (!Symbolizer::GetOrInit()->SymbolizeData(addr, &info))
     return 0;
-  ReportLocation *ent = ReportLocation::New(ReportLocationGlobal);
+  auto *ent = New<ReportLocation>();
+  ent->type = ReportLocationGlobal;
   internal_memcpy(&ent->global, &info, sizeof(info));
   return ent;
 }
diff --git a/compiler-rt/lib/tsan/rtl/tsan_sync.cpp b/compiler-rt/lib/tsan/rtl/tsan_sync.cpp
index 5e226b2d12b1..f042abab74e5 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_sync.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_sync.cpp
@@ -20,13 +20,14 @@ void DDMutexInit(ThreadState *thr, uptr pc, SyncVar *s);
 
 SyncVar::SyncVar() : mtx(MutexTypeSyncVar) { Reset(0); }
 
-void SyncVar::Init(ThreadState *thr, uptr pc, uptr addr, u64 uid) {
+void SyncVar::Init(ThreadState *thr, uptr pc, uptr addr, u64 uid,
+                   bool save_stack) {
   this->addr = addr;
   this->uid = uid;
   this->next = 0;
 
-  creation_stack_id = 0;
-  if (!SANITIZER_GO)  // Go does not use them
+  creation_stack_id = kInvalidStackID;
+  if (save_stack && !SANITIZER_GO)  // Go does not use them
     creation_stack_id = CurrentStackId(thr, pc);
   if (common_flags()->detect_deadlocks)
     DDMutexInit(thr, pc, this);
@@ -34,7 +35,7 @@ void SyncVar::Init(ThreadState *thr, uptr pc, uptr addr, u64 uid) {
 
 void SyncVar::Reset(Processor *proc) {
   uid = 0;
-  creation_stack_id = 0;
+  creation_stack_id = kInvalidStackID;
   owner_tid = kInvalidTid;
   last_lock = 0;
   recursion = 0;
@@ -190,63 +191,41 @@ MBlock* MetaMap::GetBlock(uptr p) {
   }
 }
 
-SyncVar* MetaMap::GetOrCreateAndLock(ThreadState *thr, uptr pc,
-                              uptr addr, bool write_lock) {
-  return GetAndLock(thr, pc, addr, write_lock, true);
-}
-
-SyncVar* MetaMap::GetIfExistsAndLock(uptr addr, bool write_lock) {
-  return GetAndLock(0, 0, addr, write_lock, false);
-}
-
-SyncVar *MetaMap::GetAndLock(ThreadState *thr, uptr pc, uptr addr, bool write_lock,
-                             bool create) NO_THREAD_SAFETY_ANALYSIS {
+SyncVar *MetaMap::GetSync(ThreadState *thr, uptr pc, uptr addr, bool create,
+                          bool save_stack) {
   u32 *meta = MemToMeta(addr);
   u32 idx0 = *meta;
   u32 myidx = 0;
-  SyncVar *mys = 0;
+  SyncVar *mys = nullptr;
   for (;;) {
-    u32 idx = idx0;
-    for (;;) {
-      if (idx == 0)
-        break;
-      if (idx & kFlagBlock)
-        break;
+    for (u32 idx = idx0; idx && !(idx & kFlagBlock);) {
       DCHECK(idx & kFlagSync);
       SyncVar * s = sync_alloc_.Map(idx & ~kFlagMask);
-      if (s->addr == addr) {
-        if (myidx != 0) {
+      if (LIKELY(s->addr == addr)) {
+        if (UNLIKELY(myidx != 0)) {
           mys->Reset(thr->proc());
           sync_alloc_.Free(&thr->proc()->sync_cache, myidx);
         }
-        if (write_lock)
-          s->mtx.Lock();
-        else
-          s->mtx.ReadLock();
         return s;
       }
       idx = s->next;
     }
     if (!create)
-      return 0;
-    if (*meta != idx0) {
+      return nullptr;
+    if (UNLIKELY(*meta != idx0)) {
       idx0 = *meta;
       continue;
     }
 
-    if (myidx == 0) {
+    if (LIKELY(myidx == 0)) {
       const u64 uid = atomic_fetch_add(&uid_gen_, 1, memory_order_relaxed);
       myidx = sync_alloc_.Alloc(&thr->proc()->sync_cache);
       mys = sync_alloc_.Map(myidx);
-      mys->Init(thr, pc, addr, uid);
+      mys->Init(thr, pc, addr, uid, save_stack);
     }
     mys->next = idx0;
     if (atomic_compare_exchange_strong((atomic_uint32_t*)meta, &idx0,
         myidx | kFlagSync, memory_order_release)) {
-      if (write_lock)
-        mys->mtx.Lock();
-      else
-        mys->mtx.ReadLock();
       return mys;
     }
   }
@@ -290,4 +269,11 @@ void MetaMap::OnProcIdle(Processor *proc) {
   sync_alloc_.FlushCache(&proc->sync_cache);
 }
 
+MetaMap::MemoryStats MetaMap::GetMemoryStats() const {
+  MemoryStats stats;
+  stats.mem_block = block_alloc_.AllocatedMemory();
+  stats.sync_obj = sync_alloc_.AllocatedMemory();
+  return stats;
+}
+
 }  // namespace __tsan
diff --git a/compiler-rt/lib/tsan/rtl/tsan_sync.h b/compiler-rt/lib/tsan/rtl/tsan_sync.h
index 324aa1b0cea1..fc8fa288a841 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_sync.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_sync.h
@@ -46,14 +46,16 @@ enum MutexFlags {
                                  MutexFlagNotStatic,
 };
 
+// SyncVar is a descriptor of a user synchronization object
+// (mutex or an atomic variable).
 struct SyncVar {
   SyncVar();
 
   uptr addr;  // overwritten by DenseSlabAlloc freelist
   Mutex mtx;
   u64 uid;  // Globally unique id.
-  u32 creation_stack_id;
-  u32 owner_tid;  // Set only by exclusive owners.
+  StackID creation_stack_id;
+  Tid owner_tid;  // Set only by exclusive owners.
   u64 last_lock;
   int recursion;
   atomic_uint32_t flags;
@@ -64,7 +66,7 @@ struct SyncVar {
   // with the mtx. This reduces contention for hot sync objects.
   SyncClock clock;
 
-  void Init(ThreadState *thr, uptr pc, uptr addr, u64 uid);
+  void Init(ThreadState *thr, uptr pc, uptr addr, u64 uid, bool save_stack);
   void Reset(Processor *proc);
 
   u64 GetId() const {
@@ -101,10 +103,8 @@ struct SyncVar {
   }
 };
 
-/* MetaMap allows to map arbitrary user pointers onto various descriptors.
-   Currently it maps pointers to heap block descriptors and sync var descs.
-   It uses 1/2 direct shadow, see tsan_platform.h.
-*/
+// MetaMap maps app addresses to heap block (MBlock) and sync var (SyncVar)
+// descriptors. It uses 1/2 direct shadow, see tsan_platform.h for the mapping.
 class MetaMap {
  public:
   MetaMap();
@@ -115,14 +115,25 @@ class MetaMap {
   void ResetRange(Processor *proc, uptr p, uptr sz);
   MBlock* GetBlock(uptr p);
 
-  SyncVar* GetOrCreateAndLock(ThreadState *thr, uptr pc,
-                              uptr addr, bool write_lock);
-  SyncVar* GetIfExistsAndLock(uptr addr, bool write_lock);
+  SyncVar *GetSyncOrCreate(ThreadState *thr, uptr pc, uptr addr,
+                           bool save_stack) {
+    return GetSync(thr, pc, addr, true, save_stack);
+  }
+  SyncVar *GetSyncIfExists(uptr addr) {
+    return GetSync(nullptr, 0, addr, false, false);
+  }
 
   void MoveMemory(uptr src, uptr dst, uptr sz);
 
   void OnProcIdle(Processor *proc);
 
+  struct MemoryStats {
+    uptr mem_block;
+    uptr sync_obj;
+  };
+
+  MemoryStats GetMemoryStats() const;
+
  private:
   static const u32 kFlagMask  = 3u << 30;
   static const u32 kFlagBlock = 1u << 30;
@@ -133,8 +144,8 @@ class MetaMap {
   SyncAlloc sync_alloc_;
   atomic_uint64_t uid_gen_;
 
-  SyncVar* GetAndLock(ThreadState *thr, uptr pc, uptr addr, bool write_lock,
-                      bool create);
+  SyncVar *GetSync(ThreadState *thr, uptr pc, uptr addr, bool create,
+                   bool save_stack);
 };
 
 }  // namespace __tsan
diff --git a/compiler-rt/lib/tsan/rtl/tsan_trace.h b/compiler-rt/lib/tsan/rtl/tsan_trace.h
index f5e0c407cda8..ffc8c991ece0 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_trace.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_trace.h
@@ -13,8 +13,9 @@
 #define TSAN_TRACE_H
 
 #include "tsan_defs.h"
-#include "tsan_stack_trace.h"
+#include "tsan_ilist.h"
 #include "tsan_mutexset.h"
+#include "tsan_stack_trace.h"
 
 namespace __tsan {
 
@@ -67,6 +68,185 @@ struct Trace {
   Trace() : mtx(MutexTypeTrace) {}
 };
 
+namespace v3 {
+
+enum class EventType : u64 {
+  kAccessExt,
+  kAccessRange,
+  kLock,
+  kRLock,
+  kUnlock,
+  kTime,
+};
+
+// "Base" type for all events for type dispatch.
+struct Event {
+  // We use variable-length type encoding to give more bits to some event
+  // types that need them. If is_access is set, this is EventAccess.
+  // Otherwise, if is_func is set, this is EventFunc.
+  // Otherwise type denotes the type.
+  u64 is_access : 1;
+  u64 is_func : 1;
+  EventType type : 3;
+  u64 _ : 59;
+};
+static_assert(sizeof(Event) == 8, "bad Event size");
+
+// Nop event used as padding and does not affect state during replay.
+static constexpr Event NopEvent = {1, 0, EventType::kAccessExt, 0};
+
+// Compressed memory access can represent only some events with PCs
+// close enough to each other. Otherwise we fall back to EventAccessExt.
+struct EventAccess {
+  static constexpr uptr kPCBits = 15;
+  static_assert(kPCBits + kCompressedAddrBits + 5 == 64,
+                "unused bits in EventAccess");
+
+  u64 is_access : 1;  // = 1
+  u64 is_read : 1;
+  u64 is_atomic : 1;
+  u64 size_log : 2;
+  u64 pc_delta : kPCBits;  // signed delta from the previous memory access PC
+  u64 addr : kCompressedAddrBits;
+};
+static_assert(sizeof(EventAccess) == 8, "bad EventAccess size");
+
+// Function entry (pc != 0) or exit (pc == 0).
+struct EventFunc {
+  u64 is_access : 1;  // = 0
+  u64 is_func : 1;    // = 1
+  u64 pc : 62;
+};
+static_assert(sizeof(EventFunc) == 8, "bad EventFunc size");
+
+// Extended memory access with full PC.
+struct EventAccessExt {
+  // Note: precisely specifying the unused parts of the bitfield is critical for
+  // performance. If we don't specify them, compiler will generate code to load
+  // the old value and shuffle it to extract the unused bits to apply to the new
+  // value. If we specify the unused part and store 0 in there, all that
+  // unnecessary code goes away (store of the 0 const is combined with other
+  // constant parts).
+  static constexpr uptr kUnusedBits = 11;
+  static_assert(kCompressedAddrBits + kUnusedBits + 9 == 64,
+                "unused bits in EventAccessExt");
+
+  u64 is_access : 1;   // = 0
+  u64 is_func : 1;     // = 0
+  EventType type : 3;  // = EventType::kAccessExt
+  u64 is_read : 1;
+  u64 is_atomic : 1;
+  u64 size_log : 2;
+  u64 _ : kUnusedBits;
+  u64 addr : kCompressedAddrBits;
+  u64 pc;
+};
+static_assert(sizeof(EventAccessExt) == 16, "bad EventAccessExt size");
+
+// Access to a memory range.
+struct EventAccessRange {
+  static constexpr uptr kSizeLoBits = 13;
+  static_assert(kCompressedAddrBits + kSizeLoBits + 7 == 64,
+                "unused bits in EventAccessRange");
+
+  u64 is_access : 1;   // = 0
+  u64 is_func : 1;     // = 0
+  EventType type : 3;  // = EventType::kAccessRange
+  u64 is_read : 1;
+  u64 is_free : 1;
+  u64 size_lo : kSizeLoBits;
+  u64 pc : kCompressedAddrBits;
+  u64 addr : kCompressedAddrBits;
+  u64 size_hi : 64 - kCompressedAddrBits;
+};
+static_assert(sizeof(EventAccessRange) == 16, "bad EventAccessRange size");
+
+// Mutex lock.
+struct EventLock {
+  static constexpr uptr kStackIDLoBits = 15;
+  static constexpr uptr kStackIDHiBits =
+      sizeof(StackID) * kByteBits - kStackIDLoBits;
+  static constexpr uptr kUnusedBits = 3;
+  static_assert(kCompressedAddrBits + kStackIDLoBits + 5 == 64,
+                "unused bits in EventLock");
+  static_assert(kCompressedAddrBits + kStackIDHiBits + kUnusedBits == 64,
+                "unused bits in EventLock");
+
+  u64 is_access : 1;   // = 0
+  u64 is_func : 1;     // = 0
+  EventType type : 3;  // = EventType::kLock or EventType::kRLock
+  u64 pc : kCompressedAddrBits;
+  u64 stack_lo : kStackIDLoBits;
+  u64 stack_hi : sizeof(StackID) * kByteBits - kStackIDLoBits;
+  u64 _ : kUnusedBits;
+  u64 addr : kCompressedAddrBits;
+};
+static_assert(sizeof(EventLock) == 16, "bad EventLock size");
+
+// Mutex unlock.
+struct EventUnlock {
+  static constexpr uptr kUnusedBits = 15;
+  static_assert(kCompressedAddrBits + kUnusedBits + 5 == 64,
+                "unused bits in EventUnlock");
+
+  u64 is_access : 1;   // = 0
+  u64 is_func : 1;     // = 0
+  EventType type : 3;  // = EventType::kUnlock
+  u64 _ : kUnusedBits;
+  u64 addr : kCompressedAddrBits;
+};
+static_assert(sizeof(EventUnlock) == 8, "bad EventUnlock size");
+
+// Time change event.
+struct EventTime {
+  static constexpr uptr kUnusedBits = 37;
+  static_assert(kUnusedBits + sizeof(Sid) * kByteBits + kEpochBits + 5 == 64,
+                "unused bits in EventTime");
+
+  u64 is_access : 1;   // = 0
+  u64 is_func : 1;     // = 0
+  EventType type : 3;  // = EventType::kTime
+  u64 sid : sizeof(Sid) * kByteBits;
+  u64 epoch : kEpochBits;
+  u64 _ : kUnusedBits;
+};
+static_assert(sizeof(EventTime) == 8, "bad EventTime size");
+
+struct Trace;
+
+struct TraceHeader {
+  Trace* trace = nullptr;  // back-pointer to Trace containing this part
+  INode trace_parts;       // in Trace::parts
+};
+
+struct TracePart : TraceHeader {
+  // There are a lot of goroutines in Go, so we use smaller parts.
+  static constexpr uptr kByteSize = (SANITIZER_GO ? 128 : 256) << 10;
+  static constexpr uptr kSize =
+      (kByteSize - sizeof(TraceHeader)) / sizeof(Event);
+  // TraceAcquire does a fast event pointer overflow check by comparing
+  // pointer into TracePart::events with kAlignment mask. Since TracePart's
+  // are allocated page-aligned, this check detects end of the array
+  // (it also have false positives in the middle that are filtered separately).
+  // This also requires events to be the last field.
+  static constexpr uptr kAlignment = 0xff0;
+  Event events[kSize];
+
+  TracePart() {}
+};
+static_assert(sizeof(TracePart) == TracePart::kByteSize, "bad TracePart size");
+
+struct Trace {
+  Mutex mtx;
+  IList<TraceHeader, &TraceHeader::trace_parts, TracePart> parts;
+  Event* final_pos =
+      nullptr;  // final position in the last part for finished threads
+
+  Trace() : mtx(MutexTypeTrace) {}
+};
+
+}  // namespace v3
+
 }  // namespace __tsan
 
 #endif  // TSAN_TRACE_H
diff --git a/compiler-rt/lib/tsan/rtl/tsan_update_shadow_word_inl.h b/compiler-rt/lib/tsan/rtl/tsan_update_shadow_word.inc
index d23dfb0ba061..a58ef0f17efa 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_update_shadow_word_inl.h
+++ b/compiler-rt/lib/tsan/rtl/tsan_update_shadow_word.inc
@@ -1,4 +1,4 @@
-//===-- tsan_update_shadow_word_inl.h ---------------------------*- C++ -*-===//
+//===-- tsan_update_shadow_word.inc -----------------------------*- C++ -*-===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
diff --git a/compiler-rt/lib/tsan/rtl/tsan_vector_clock.cpp b/compiler-rt/lib/tsan/rtl/tsan_vector_clock.cpp
new file mode 100644
index 000000000000..278298565d3f
--- /dev/null
+++ b/compiler-rt/lib/tsan/rtl/tsan_vector_clock.cpp
@@ -0,0 +1,126 @@
+//===-- tsan_vector_clock.cpp ---------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of ThreadSanitizer (TSan), a race detector.
+//
+//===----------------------------------------------------------------------===//
+#include "tsan_vector_clock.h"
+
+#include "sanitizer_common/sanitizer_placement_new.h"
+#include "tsan_mman.h"
+
+namespace __tsan {
+
+#if TSAN_VECTORIZE
+const uptr kVectorClockSize = kThreadSlotCount * sizeof(Epoch) / sizeof(m128);
+#endif
+
+VectorClock::VectorClock() { Reset(); }
+
+void VectorClock::Reset() {
+#if !TSAN_VECTORIZE
+  for (uptr i = 0; i < kThreadSlotCount; i++)
+    clk_[i] = kEpochZero;
+#else
+  m128 z = _mm_setzero_si128();
+  m128* vclk = reinterpret_cast<m128*>(clk_);
+  for (uptr i = 0; i < kVectorClockSize; i++) _mm_store_si128(&vclk[i], z);
+#endif
+}
+
+void VectorClock::Acquire(const VectorClock* src) {
+  if (!src)
+    return;
+#if !TSAN_VECTORIZE
+  for (uptr i = 0; i < kThreadSlotCount; i++)
+    clk_[i] = max(clk_[i], src->clk_[i]);
+#else
+  m128* __restrict vdst = reinterpret_cast<m128*>(clk_);
+  m128 const* __restrict vsrc = reinterpret_cast<m128 const*>(src->clk_);
+  for (uptr i = 0; i < kVectorClockSize; i++) {
+    m128 s = _mm_load_si128(&vsrc[i]);
+    m128 d = _mm_load_si128(&vdst[i]);
+    m128 m = _mm_max_epu16(s, d);
+    _mm_store_si128(&vdst[i], m);
+  }
+#endif
+}
+
+static VectorClock* AllocClock(VectorClock** dstp) {
+  if (UNLIKELY(!*dstp))
+    *dstp = New<VectorClock>();
+  return *dstp;
+}
+
+void VectorClock::Release(VectorClock** dstp) const {
+  VectorClock* dst = AllocClock(dstp);
+  dst->Acquire(this);
+}
+
+void VectorClock::ReleaseStore(VectorClock** dstp) const {
+  VectorClock* dst = AllocClock(dstp);
+  *dst = *this;
+}
+
+VectorClock& VectorClock::operator=(const VectorClock& other) {
+#if !TSAN_VECTORIZE
+  for (uptr i = 0; i < kThreadSlotCount; i++)
+    clk_[i] = other.clk_[i];
+#else
+  m128* __restrict vdst = reinterpret_cast<m128*>(clk_);
+  m128 const* __restrict vsrc = reinterpret_cast<m128 const*>(other.clk_);
+  for (uptr i = 0; i < kVectorClockSize; i++) {
+    m128 s = _mm_load_si128(&vsrc[i]);
+    _mm_store_si128(&vdst[i], s);
+  }
+#endif
+  return *this;
+}
+
+void VectorClock::ReleaseStoreAcquire(VectorClock** dstp) {
+  VectorClock* dst = AllocClock(dstp);
+#if !TSAN_VECTORIZE
+  for (uptr i = 0; i < kThreadSlotCount; i++) {
+    Epoch tmp = dst->clk_[i];
+    dst->clk_[i] = clk_[i];
+    clk_[i] = max(clk_[i], tmp);
+  }
+#else
+  m128* __restrict vdst = reinterpret_cast<m128*>(dst->clk_);
+  m128* __restrict vclk = reinterpret_cast<m128*>(clk_);
+  for (uptr i = 0; i < kVectorClockSize; i++) {
+    m128 t = _mm_load_si128(&vdst[i]);
+    m128 c = _mm_load_si128(&vclk[i]);
+    m128 m = _mm_max_epu16(c, t);
+    _mm_store_si128(&vdst[i], c);
+    _mm_store_si128(&vclk[i], m);
+  }
+#endif
+}
+
+void VectorClock::ReleaseAcquire(VectorClock** dstp) {
+  VectorClock* dst = AllocClock(dstp);
+#if !TSAN_VECTORIZE
+  for (uptr i = 0; i < kThreadSlotCount; i++) {
+    dst->clk_[i] = max(dst->clk_[i], clk_[i]);
+    clk_[i] = dst->clk_[i];
+  }
+#else
+  m128* __restrict vdst = reinterpret_cast<m128*>(dst->clk_);
+  m128* __restrict vclk = reinterpret_cast<m128*>(clk_);
+  for (uptr i = 0; i < kVectorClockSize; i++) {
+    m128 c = _mm_load_si128(&vclk[i]);
+    m128 d = _mm_load_si128(&vdst[i]);
+    m128 m = _mm_max_epu16(c, d);
+    _mm_store_si128(&vdst[i], m);
+    _mm_store_si128(&vclk[i], m);
+  }
+#endif
+}
+
+}  // namespace __tsan
diff --git a/compiler-rt/lib/tsan/rtl/tsan_vector_clock.h b/compiler-rt/lib/tsan/rtl/tsan_vector_clock.h
new file mode 100644
index 000000000000..63b206302190
--- /dev/null
+++ b/compiler-rt/lib/tsan/rtl/tsan_vector_clock.h
@@ -0,0 +1,51 @@
+//===-- tsan_vector_clock.h -------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of ThreadSanitizer (TSan), a race detector.
+//
+//===----------------------------------------------------------------------===//
+#ifndef TSAN_VECTOR_CLOCK_H
+#define TSAN_VECTOR_CLOCK_H
+
+#include "tsan_defs.h"
+
+namespace __tsan {
+
+// Fixed-size vector clock, used both for threads and sync objects.
+class VectorClock {
+ public:
+  VectorClock();
+
+  Epoch Get(Sid sid) const;
+  void Set(Sid sid, Epoch v);
+
+  void Reset();
+  void Acquire(const VectorClock* src);
+  void Release(VectorClock** dstp) const;
+  void ReleaseStore(VectorClock** dstp) const;
+  void ReleaseStoreAcquire(VectorClock** dstp);
+  void ReleaseAcquire(VectorClock** dstp);
+
+  VectorClock& operator=(const VectorClock& other);
+
+ private:
+  Epoch clk_[kThreadSlotCount] VECTOR_ALIGNED;
+};
+
+ALWAYS_INLINE Epoch VectorClock::Get(Sid sid) const {
+  return clk_[static_cast<u8>(sid)];
+}
+
+ALWAYS_INLINE void VectorClock::Set(Sid sid, Epoch v) {
+  DCHECK_GE(v, clk_[static_cast<u8>(sid)]);
+  clk_[static_cast<u8>(sid)] = v;
+}
+
+}  // namespace __tsan
+
+#endif  // TSAN_VECTOR_CLOCK_H
diff --git a/compiler-rt/lib/ubsan/ubsan_diag.cpp b/compiler-rt/lib/ubsan/ubsan_diag.cpp
index ef2e495cac8e..8de51bc18770 100644
--- a/compiler-rt/lib/ubsan/ubsan_diag.cpp
+++ b/compiler-rt/lib/ubsan/ubsan_diag.cpp
@@ -157,7 +157,7 @@ static void RenderLocation(InternalScopedString *Buffer, Location Loc) {
     return;
   }
   case Location::LK_Memory:
-    Buffer->append("%p", Loc.getMemoryLocation());
+    Buffer->append("%p", reinterpret_cast<void *>(Loc.getMemoryLocation()));
     return;
   case Location::LK_Symbolized: {
     const AddressInfo &Info = Loc.getSymbolizedStack()->info;
@@ -169,7 +169,7 @@ static void RenderLocation(InternalScopedString *Buffer, Location Loc) {
       RenderModuleLocation(Buffer, Info.module, Info.module_offset,
                            Info.module_arch, common_flags()->strip_path_prefix);
     else
-      Buffer->append("%p", Info.address);
+      Buffer->append("%p", reinterpret_cast<void *>(Info.address));
     return;
   }
   case Location::LK_Null:
@@ -286,7 +286,7 @@ static void PrintMemorySnippet(const Decorator &Decor, MemoryLocation Loc,
   Buffer.append("\n");
 
   // Emit highlights.
-  Buffer.append(Decor.Highlight());
+  Buffer.append("%s", Decor.Highlight());
   Range *InRange = upperBound(Min, Ranges, NumRanges);
   for (uptr P = Min; P != Max; ++P) {
     char Pad = ' ', Byte = ' ';
@@ -355,7 +355,7 @@ Diag::~Diag() {
     Buffer.clear();
   }
 
-  Buffer.append(Decor.Bold());
+  Buffer.append("%s", Decor.Bold());
   RenderLocation(&Buffer, Loc);
   Buffer.append(":");
 
diff --git a/compiler-rt/lib/xray/xray_basic_flags.h b/compiler-rt/lib/xray/xray_basic_flags.h
index 2459effa8bae..b846c1233e8a 100644
--- a/compiler-rt/lib/xray/xray_basic_flags.h
+++ b/compiler-rt/lib/xray/xray_basic_flags.h
@@ -6,7 +6,7 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// This file is a part of XRay, a dynamic runtime instruementation system.
+// This file is a part of XRay, a dynamic runtime instrumentation system.
 //
 // XRay Basic Mode runtime flags.
 //===----------------------------------------------------------------------===//
diff --git a/compiler-rt/lib/xray/xray_buffer_queue.cpp b/compiler-rt/lib/xray/xray_buffer_queue.cpp
index bad91e036cef..748708ccd0f4 100644
--- a/compiler-rt/lib/xray/xray_buffer_queue.cpp
+++ b/compiler-rt/lib/xray/xray_buffer_queue.cpp
@@ -6,7 +6,7 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// This file is a part of XRay, a dynamic runtime instruementation system.
+// This file is a part of XRay, a dynamic runtime instrumentation system.
 //
 // Defines the interface for a buffer queue implementation.
 //
diff --git a/compiler-rt/lib/xray/xray_flags.h b/compiler-rt/lib/xray/xray_flags.h
index edb5a5119f86..cce6fe9d62f9 100644
--- a/compiler-rt/lib/xray/xray_flags.h
+++ b/compiler-rt/lib/xray/xray_flags.h
@@ -6,7 +6,7 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// This file is a part of XRay, a dynamic runtime instruementation system.
+// This file is a part of XRay, a dynamic runtime instrumentation system.
 //
 // XRay runtime flags.
 //===----------------------------------------------------------------------===//
diff --git a/compiler-rt/lib/xray/xray_interface.cpp b/compiler-rt/lib/xray/xray_interface.cpp
index 7669b9ab82be..ddf184c9b857 100644
--- a/compiler-rt/lib/xray/xray_interface.cpp
+++ b/compiler-rt/lib/xray/xray_interface.cpp
@@ -360,7 +360,7 @@ XRayPatchingStatus mprotectAndPatchFunction(int32_t FuncId,
     return XRayPatchingStatus::FAILED;
   }
 
-  // Here we compute the minumum sled and maximum sled associated with a
+  // Here we compute the minimum sled and maximum sled associated with a
   // particular function ID.
   auto SledRange = InstrMap.SledsIndex ? InstrMap.SledsIndex[FuncId - 1]
                                        : findFunctionSleds(FuncId, InstrMap);
diff --git a/compiler-rt/lib/xray/xray_profiling.cpp b/compiler-rt/lib/xray/xray_profiling.cpp
index ef16691562cc..81c33fae88c1 100644
--- a/compiler-rt/lib/xray/xray_profiling.cpp
+++ b/compiler-rt/lib/xray/xray_profiling.cpp
@@ -402,7 +402,7 @@ profilingLoggingInit(size_t, size_t, void *Options,
       return XRayLogInitStatus::XRAY_LOG_UNINITIALIZED;
     }
 
-    // If we've succeded, set the global pointer to the initialised storage.
+    // If we've succeeded, set the global pointer to the initialised storage.
     BQ = reinterpret_cast<BufferQueue *>(&BufferQueueStorage);
   } else {
     BQ->finalize();
diff --git a/compiler-rt/lib/xray/xray_x86_64.cpp b/compiler-rt/lib/xray/xray_x86_64.cpp
index c58584b3a14b..669d2e85bede 100644
--- a/compiler-rt/lib/xray/xray_x86_64.cpp
+++ b/compiler-rt/lib/xray/xray_x86_64.cpp
@@ -148,7 +148,8 @@ bool patchFunctionEntry(const bool Enable, const uint32_t FuncId,
   int64_t TrampolineOffset = reinterpret_cast<int64_t>(Trampoline) -
                              (static_cast<int64_t>(Address) + 11);
   if (TrampolineOffset < MinOffset || TrampolineOffset > MaxOffset) {
-    Report("XRay Entry trampoline (%p) too far from sled (%p)\n", Trampoline,
+    Report("XRay Entry trampoline (%p) too far from sled (%p)\n",
+           reinterpret_cast<void *>(Trampoline),
            reinterpret_cast<void *>(Address));
     return false;
   }
@@ -195,7 +196,8 @@ bool patchFunctionExit(const bool Enable, const uint32_t FuncId,
                              (static_cast<int64_t>(Address) + 11);
   if (TrampolineOffset < MinOffset || TrampolineOffset > MaxOffset) {
     Report("XRay Exit trampoline (%p) too far from sled (%p)\n",
-           __xray_FunctionExit, reinterpret_cast<void *>(Address));
+           reinterpret_cast<void *>(__xray_FunctionExit),
+           reinterpret_cast<void *>(Address));
     return false;
   }
   if (Enable) {
@@ -224,7 +226,8 @@ bool patchFunctionTailExit(const bool Enable, const uint32_t FuncId,
       (static_cast<int64_t>(Address) + 11);
   if (TrampolineOffset < MinOffset || TrampolineOffset > MaxOffset) {
     Report("XRay Tail Exit trampoline (%p) too far from sled (%p)\n",
-           __xray_FunctionTailExit, reinterpret_cast<void *>(Address));
+           reinterpret_cast<void *>(__xray_FunctionTailExit),
+           reinterpret_cast<void *>(Address));
     return false;
   }
   if (Enable) {