341 files changed, 14606 insertions, 4651 deletions

diff --git a/contrib/compiler-rt/include/sanitizer/allocator_interface.h b/contrib/compiler-rt/include/sanitizer/allocator_interface.h
index 5220631619fc..2d35804b3764 100644
--- a/contrib/compiler-rt/include/sanitizer/allocator_interface.h
+++ b/contrib/compiler-rt/include/sanitizer/allocator_interface.h
@@ -76,6 +76,13 @@ extern "C" {
       void (*malloc_hook)(const volatile void *, size_t),
       void (*free_hook)(const volatile void *));
 
+  /* Drains allocator quarantines (calling thread's and global ones), returns
+     freed memory back to OS and releases other non-essential internal allocator
+     resources in attempt to reduce process RSS.
+     Currently available with ASan only.
+  */
+  void __sanitizer_purge_allocator();
+
 #ifdef __cplusplus
 }  // extern "C"
 #endif
diff --git a/contrib/compiler-rt/include/sanitizer/asan_interface.h b/contrib/compiler-rt/include/sanitizer/asan_interface.h
index 97ba0ceb0b23..e689a730e2c3 100644
--- a/contrib/compiler-rt/include/sanitizer/asan_interface.h
+++ b/contrib/compiler-rt/include/sanitizer/asan_interface.h
@@ -144,6 +144,10 @@ extern "C" {
   void *__asan_addr_is_in_fake_stack(void *fake_stack, void *addr, void **beg,
                                      void **end);
 
+  // Performs cleanup before a [[noreturn]] function.  Must be called
+  // before things like _exit and execl to avoid false positives on stack.
+  void __asan_handle_no_return(void);
+
 #ifdef __cplusplus
 }  // extern "C"
 #endif
diff --git a/contrib/compiler-rt/include/sanitizer/coverage_interface.h b/contrib/compiler-rt/include/sanitizer/coverage_interface.h
index 637379d47c41..081033cec8e5 100644
--- a/contrib/compiler-rt/include/sanitizer/coverage_interface.h
+++ b/contrib/compiler-rt/include/sanitizer/coverage_interface.h
@@ -22,8 +22,11 @@ extern "C" {
   // Record and dump coverage info.
   void __sanitizer_cov_dump();
 
-  //  Dump collected coverage info. Sorts pcs by module into individual
-  //  .sancov files.
+  // Clear collected coverage info.
+  void __sanitizer_cov_reset();
+
+  // Dump collected coverage info. Sorts pcs by module into individual .sancov
+  // files.
   void __sanitizer_dump_coverage(const uintptr_t *pcs, uintptr_t len);
 
 #ifdef __cplusplus
diff --git a/contrib/compiler-rt/include/sanitizer/hwasan_interface.h b/contrib/compiler-rt/include/sanitizer/hwasan_interface.h
new file mode 100644
index 000000000000..f70d87875968
--- /dev/null
+++ b/contrib/compiler-rt/include/sanitizer/hwasan_interface.h
@@ -0,0 +1,33 @@
+//===-- sanitizer/asan_interface.h ------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+// Public interface header.
+//===----------------------------------------------------------------------===//
+#ifndef SANITIZER_HWASAN_INTERFACE_H
+#define SANITIZER_HWASAN_INTERFACE_H
+
+#include <sanitizer/common_interface_defs.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+  // This function may be optionally provided by user and should return
+  // a string containing HWASan runtime options. See asan_flags.h for details.
+  const char* __hwasan_default_options();
+
+  void __hwasan_enable_allocator_tagging();
+  void __hwasan_disable_allocator_tagging();
+
+#ifdef __cplusplus
+}  // extern "C"
+#endif
+
+#endif  // SANITIZER_HWASAN_INTERFACE_H
diff --git a/contrib/compiler-rt/include/sanitizer/lsan_interface.h b/contrib/compiler-rt/include/sanitizer/lsan_interface.h
index 8fb8e756da26..e3e509ff7d9f 100644
--- a/contrib/compiler-rt/include/sanitizer/lsan_interface.h
+++ b/contrib/compiler-rt/include/sanitizer/lsan_interface.h
@@ -64,8 +64,14 @@ extern "C" {
   // for the program it is linked into (if the return value is non-zero). This
   // function must be defined as returning a constant value; any behavior beyond
   // that is unsupported.
+  // To avoid dead stripping, you may need to define this function with
+  // __attribute__((used))
   int __lsan_is_turned_off();
 
+  // This function may be optionally provided by user and should return
+  // a string containing LSan runtime options. See lsan_flags.inc for details.
+  const char *__lsan_default_options();
+
   // This function may be optionally provided by the user and should return
   // a string containing LSan suppressions.
   const char *__lsan_default_suppressions();
diff --git a/contrib/compiler-rt/include/sanitizer/scudo_interface.h b/contrib/compiler-rt/include/sanitizer/scudo_interface.h
new file mode 100644
index 000000000000..cab7e0f5f344
--- /dev/null
+++ b/contrib/compiler-rt/include/sanitizer/scudo_interface.h
@@ -0,0 +1,34 @@
+//===-- sanitizer/scudo_interface.h -----------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+/// Public Scudo interface header.
+//
+//===----------------------------------------------------------------------===//
+#ifndef SANITIZER_SCUDO_INTERFACE_H_
+#define SANITIZER_SCUDO_INTERFACE_H_
+
+#include <sanitizer/common_interface_defs.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+  // This function may be optionally provided by a user and should return
+  // a string containing Scudo runtime options. See scudo_flags.h for details.
+  const char* __scudo_default_options();
+
+  // This function allows to set the RSS limit at runtime. This can be either
+  // the hard limit (HardLimit=1) or the soft limit (HardLimit=0). The limit
+  // can be removed by setting LimitMb to 0. This function's parameters should
+  // be fully trusted to avoid security mishaps.
+  void __scudo_set_rss_limit(unsigned long LimitMb, int HardLimit);
+#ifdef __cplusplus
+}  // extern "C"
+#endif
+
+#endif  // SANITIZER_SCUDO_INTERFACE_H_
diff --git a/contrib/compiler-rt/include/sanitizer/tsan_interface.h b/contrib/compiler-rt/include/sanitizer/tsan_interface.h
index 5ea09ab5c1ff..530885014ac3 100644
--- a/contrib/compiler-rt/include/sanitizer/tsan_interface.h
+++ b/contrib/compiler-rt/include/sanitizer/tsan_interface.h
@@ -44,6 +44,11 @@ const unsigned __tsan_mutex_linker_init      = 1 << 0;
 const unsigned __tsan_mutex_write_reentrant  = 1 << 1;
 // Mutex is read reentrant.
 const unsigned __tsan_mutex_read_reentrant   = 1 << 2;
+// Mutex does not have static storage duration, and must not be used after
+// its destructor runs.  The opposite of __tsan_mutex_linker_init.
+// If this flag is passed to __tsan_mutex_destroy, then the destruction
+// is ignored unless this flag was previously set on the mutex.
+const unsigned __tsan_mutex_not_static       = 1 << 8;
 
 // Mutex operation flags:
 
@@ -70,6 +75,7 @@ void __tsan_mutex_create(void *addr, unsigned flags);
 // Annotate destruction of a mutex.
 // Supported flags:
 //   - __tsan_mutex_linker_init
+//   - __tsan_mutex_not_static
 void __tsan_mutex_destroy(void *addr, unsigned flags);
 
 // Annotate start of lock operation.
diff --git a/contrib/compiler-rt/include/xray/xray_interface.h b/contrib/compiler-rt/include/xray/xray_interface.h
index 564613417069..d08039a67fa2 100644
--- a/contrib/compiler-rt/include/xray/xray_interface.h
+++ b/contrib/compiler-rt/include/xray/xray_interface.h
@@ -106,6 +106,14 @@ extern uintptr_t __xray_function_address(int32_t FuncId);
 /// encounter errors (when there are no instrumented functions, etc.).
 extern size_t __xray_max_function_id();
 
+/// Initialize the required XRay data structures. This is useful in cases where
+/// users want to control precisely when the XRay instrumentation data
+/// structures are initialized, for example when the XRay library is built with
+/// the XRAY_NO_PREINIT preprocessor definition.
+///
+/// Calling __xray_init() more than once is safe across multiple threads.
+extern void __xray_init();
+
 } // end extern "C"
 
 #endif // XRAY_XRAY_INTERFACE_H
diff --git a/contrib/compiler-rt/include/xray/xray_log_interface.h b/contrib/compiler-rt/include/xray/xray_log_interface.h
index cdb20094dc26..6c53cdaebfe2 100644
--- a/contrib/compiler-rt/include/xray/xray_log_interface.h
+++ b/contrib/compiler-rt/include/xray/xray_log_interface.h
@@ -128,6 +128,16 @@ enum XRayLogFlushStatus {
   XRAY_LOG_FLUSHED = 2,
 };
 
+/// This enum indicates the installation state of a logging implementation, when
+/// associating a mode to a particular logging implementation through
+/// `__xray_log_register_impl(...)` or through `__xray_log_select_mode(...`.
+enum XRayLogRegisterStatus {
+  XRAY_REGISTRATION_OK = 0,
+  XRAY_DUPLICATE_MODE = 1,
+  XRAY_MODE_NOT_FOUND = 2,
+  XRAY_INCOMPLETE_IMPL = 3,
+};
+
 /// A valid XRay logging implementation MUST provide all of the function
 /// pointers in XRayLogImpl when being installed through `__xray_set_log_impl`.
 /// To be precise, ALL the functions pointers MUST NOT be nullptr.
@@ -159,6 +169,9 @@ struct XRayLogImpl {
   /// always have a handler for function entry and exit events. In case the
   /// implementation wants to support arg1 (or other future extensions to XRay
   /// logging) those MUST be installed by the installed 'log_init' handler.
+  ///
+  /// Because we didn't want to change the ABI of this struct, the arg1 handler
+  /// may be silently overwritten during initialization as well.
   void (*handle_arg0)(int32_t, XRayEntryType);
 
   /// The log implementation provided routine for when __xray_log_flushLog() is
@@ -186,6 +199,34 @@ struct XRayLogImpl {
 /// called while in any other states.
 void __xray_set_log_impl(XRayLogImpl Impl);
 
+/// This function registers a logging implementation against a "mode"
+/// identifier. This allows multiple modes to be registered, and chosen at
+/// runtime using the same mode identifier through
+/// `__xray_log_select_mode(...)`.
+///
+/// We treat the Mode identifier as a null-terminated byte string, as the
+/// identifier used when retrieving the log impl.
+///
+/// Returns:
+///   - XRAY_REGISTRATION_OK on success.
+///   - XRAY_DUPLICATE_MODE when an implementation is already associated with
+///     the provided Mode; does not update the already-registered
+///     implementation.
+XRayLogRegisterStatus __xray_log_register_mode(const char *Mode,
+                                               XRayLogImpl Impl);
+
+/// This function selects the implementation associated with Mode that has been
+/// registered through __xray_log_register_mode(...) and installs that
+/// implementation (as if through calling __xray_set_log_impl(...)). The same
+/// caveats apply to __xray_log_select_mode(...) as with
+/// __xray_log_set_log_impl(...).
+///
+/// Returns:
+///   - XRAY_REGISTRATION_OK on success.
+///   - XRAY_MODE_NOT_FOUND if there is no implementation associated with Mode;
+///     does not update the currently installed implementation.
+XRayLogRegisterStatus __xray_log_select_mode(const char *Mode);
+
 /// This function removes the currently installed implementation. It will also
 /// uninstall any handlers that have been previously installed. It does NOT
 /// unpatch the instrumentation sleds.
@@ -220,12 +261,19 @@ XRayLogFlushStatus __xray_log_flushLog();
 
 namespace __xray {
 
-// Options used by the LLVM XRay FDR implementation.
+/// Options used by the LLVM XRay FDR logging implementation.
 struct FDRLoggingOptions {
   bool ReportErrors = false;
   int Fd = -1;
 };
 
+/// Options used by the LLVM XRay Basic (Naive) logging implementation.
+struct BasicLoggingOptions {
+  int DurationFilterMicros = 0;
+  size_t MaxStackDepth = 0;
+  size_t ThreadBufferSize = 0;
+};
+
 } // namespace __xray
 
 #endif // XRAY_XRAY_LOG_INTERFACE_H
diff --git a/contrib/compiler-rt/include/xray/xray_records.h b/contrib/compiler-rt/include/xray/xray_records.h
index feb8d228b2fd..d4b7b4c31a9a 100644
--- a/contrib/compiler-rt/include/xray/xray_records.h
+++ b/contrib/compiler-rt/include/xray/xray_records.h
@@ -17,6 +17,8 @@
 #ifndef XRAY_XRAY_RECORDS_H
 #define XRAY_XRAY_RECORDS_H
 
+#include <cstdint>
+
 namespace __xray {
 
 enum FileTypes {
@@ -65,18 +67,23 @@ static_assert(sizeof(XRayFileHeader) == 32, "XRayFileHeader != 32 bytes");
 
 enum RecordTypes {
   NORMAL = 0,
+  ARG_PAYLOAD = 1,
 };
 
 struct alignas(32) XRayRecord {
   // This is the type of the record being written. We use 16 bits to allow us to
   // treat this as a discriminant, and so that the first 4 bytes get packed
   // properly. See RecordTypes for more supported types.
-  uint16_t RecordType = 0;
+  uint16_t RecordType = RecordTypes::NORMAL;
 
   // The CPU where the thread is running. We assume number of CPUs <= 256.
   uint8_t CPU = 0;
 
-  // The type of the event. Usually either ENTER = 0 or EXIT = 1.
+  // The type of the event. One of the following:
+  //   ENTER = 0
+  //   EXIT = 1
+  //   TAIL_EXIT = 2
+  //   ENTER_ARG = 3
   uint8_t Type = 0;
 
   // The function ID for the record.
@@ -94,6 +101,32 @@ struct alignas(32) XRayRecord {
 
 static_assert(sizeof(XRayRecord) == 32, "XRayRecord != 32 bytes");
 
+struct alignas(32) XRayArgPayload {
+  // We use the same 16 bits as a discriminant for the records in the log here
+  // too, and so that the first 4 bytes are packed properly.
+  uint16_t RecordType = RecordTypes::ARG_PAYLOAD;
+
+  // Add a few bytes to pad.
+  uint8_t Padding[2] = {};
+
+  // The function ID for the record.
+  int32_t FuncId = 0;
+
+  // The thread ID for the currently running thread.
+  uint32_t TId = 0;
+
+  // Add more padding.
+  uint8_t Padding2[4] = {};
+
+  // The argument payload.
+  uint64_t Arg = 0;
+
+  // The rest of this record ought to be left as padding.
+  uint8_t TailPadding[8] = {};
+} __attribute__((packed));
+
+static_assert(sizeof(XRayArgPayload) == 32, "XRayArgPayload != 32 bytes");
+
 } // namespace __xray
 
 #endif // XRAY_XRAY_RECORDS_H
diff --git a/contrib/compiler-rt/lib/BlocksRuntime/Block.h b/contrib/compiler-rt/lib/BlocksRuntime/Block.h
index 55cdd01a9123..e6cb142fd755 100644
--- a/contrib/compiler-rt/lib/BlocksRuntime/Block.h
+++ b/contrib/compiler-rt/lib/BlocksRuntime/Block.h
@@ -27,7 +27,7 @@
 
 #if !defined(BLOCK_EXPORT)
 #   if defined(__cplusplus)
-#       define BLOCK_EXPORT extern "C" 
+#       define BLOCK_EXPORT extern "C"
 #   else
 #       define BLOCK_EXPORT extern
 #   endif
diff --git a/contrib/compiler-rt/lib/BlocksRuntime/Block_private.h b/contrib/compiler-rt/lib/BlocksRuntime/Block_private.h
index 8ae821815ebe..91d8d8a9850b 100644
--- a/contrib/compiler-rt/lib/BlocksRuntime/Block_private.h
+++ b/contrib/compiler-rt/lib/BlocksRuntime/Block_private.h
@@ -27,7 +27,7 @@
 
 #if !defined(BLOCK_EXPORT)
 #   if defined(__cplusplus)
-#       define BLOCK_EXPORT extern "C" 
+#       define BLOCK_EXPORT extern "C"
 #   else
 #       define BLOCK_EXPORT extern
 #   endif
diff --git a/contrib/compiler-rt/lib/asan/asan_activation.cc b/contrib/compiler-rt/lib/asan/asan_activation.cc
index 66eba9ce2748..d642be93488d 100644
--- a/contrib/compiler-rt/lib/asan/asan_activation.cc
+++ b/contrib/compiler-rt/lib/asan/asan_activation.cc
@@ -16,8 +16,10 @@
 #include "asan_allocator.h"
 #include "asan_flags.h"
 #include "asan_internal.h"
+#include "asan_mapping.h"
 #include "asan_poisoning.h"
 #include "asan_stack.h"
+#include "sanitizer_common/sanitizer_common.h"
 #include "sanitizer_common/sanitizer_flags.h"
 
 namespace __asan {
@@ -110,8 +112,9 @@ void AsanDeactivate() {
   AllocatorOptions disabled = asan_deactivated_flags.allocator_options;
   disabled.quarantine_size_mb = 0;
   disabled.thread_local_quarantine_size_kb = 0;
-  disabled.min_redzone = 16;  // Redzone must be at least 16 bytes long.
-  disabled.max_redzone = 16;
+  // Redzone must be at least Max(16, granularity) bytes long.
+  disabled.min_redzone = Max(16, (int)SHADOW_GRANULARITY);
+  disabled.max_redzone = disabled.min_redzone;
   disabled.alloc_dealloc_mismatch = false;
   disabled.may_return_null = true;
   ReInitializeAllocator(disabled);
diff --git a/contrib/compiler-rt/lib/asan/asan_allocator.cc b/contrib/compiler-rt/lib/asan/asan_allocator.cc
index 92963ddfc4da..a437ae1cd3b1 100644
--- a/contrib/compiler-rt/lib/asan/asan_allocator.cc
+++ b/contrib/compiler-rt/lib/asan/asan_allocator.cc
@@ -84,7 +84,10 @@ struct ChunkHeader {
   // This field is used for small sizes. For large sizes it is equal to
   // SizeClassMap::kMaxSize and the actual size is stored in the
   // SecondaryAllocator's metadata.
-  u32 user_requested_size;
+  u32 user_requested_size : 29;
+  // align < 8 -> 0
+  // else      -> log2(min(align, 512)) - 2
+  u32 user_requested_alignment_log : 3;
   u32 alloc_context_id;
 };
 
@@ -271,9 +274,9 @@ struct Allocator {
     atomic_store(&max_redzone, options.max_redzone, memory_order_release);
   }
 
-  void Initialize(const AllocatorOptions &options) {
+  void InitLinkerInitialized(const AllocatorOptions &options) {
     SetAllocatorMayReturnNull(options.may_return_null);
-    allocator.Init(options.release_to_os_interval_ms);
+    allocator.InitLinkerInitialized(options.release_to_os_interval_ms);
     SharedInitCode(options);
   }
 
@@ -351,6 +354,20 @@ struct Allocator {
     return Min(Max(rz_log, RZSize2Log(min_rz)), RZSize2Log(max_rz));
   }
 
+  static uptr ComputeUserRequestedAlignmentLog(uptr user_requested_alignment) {
+    if (user_requested_alignment < 8)
+      return 0;
+    if (user_requested_alignment > 512)
+      user_requested_alignment = 512;
+    return Log2(user_requested_alignment) - 2;
+  }
+
+  static uptr ComputeUserAlignment(uptr user_requested_alignment_log) {
+    if (user_requested_alignment_log == 0)
+      return 0;
+    return 1LL << (user_requested_alignment_log + 2);
+  }
+
   // We have an address between two chunks, and we want to report just one.
   AsanChunk *ChooseChunk(uptr addr, AsanChunk *left_chunk,
                          AsanChunk *right_chunk) {
@@ -385,6 +402,8 @@ struct Allocator {
     Flags &fl = *flags();
     CHECK(stack);
     const uptr min_alignment = SHADOW_GRANULARITY;
+    const uptr user_requested_alignment_log =
+        ComputeUserRequestedAlignmentLog(alignment);
     if (alignment < min_alignment)
       alignment = min_alignment;
     if (size == 0) {
@@ -472,6 +491,7 @@ struct Allocator {
       meta[0] = size;
       meta[1] = chunk_beg;
     }
+    m->user_requested_alignment_log = user_requested_alignment_log;
 
     m->alloc_context_id = StackDepotPut(*stack);
 
@@ -573,8 +593,8 @@ struct Allocator {
     }
   }
 
-  void Deallocate(void *ptr, uptr delete_size, BufferedStackTrace *stack,
-                  AllocType alloc_type) {
+  void Deallocate(void *ptr, uptr delete_size, uptr delete_alignment,
+                  BufferedStackTrace *stack, AllocType alloc_type) {
     uptr p = reinterpret_cast<uptr>(ptr);
     if (p == 0) return;
 
@@ -601,11 +621,14 @@ struct Allocator {
         ReportAllocTypeMismatch((uptr)ptr, stack, (AllocType)m->alloc_type,
                                 (AllocType)alloc_type);
       }
-    }
-
-    if (delete_size && flags()->new_delete_type_mismatch &&
-        delete_size != m->UsedSize()) {
-      ReportNewDeleteSizeMismatch(p, delete_size, stack);
+    } else {
+      if (flags()->new_delete_type_mismatch &&
+          (alloc_type == FROM_NEW || alloc_type == FROM_NEW_BR) &&
+          ((delete_size && delete_size != m->UsedSize()) ||
+           ComputeUserRequestedAlignmentLog(delete_alignment) !=
+               m->user_requested_alignment_log)) {
+        ReportNewDeleteTypeMismatch(p, delete_size, delete_alignment, stack);
+      }
     }
 
     QuarantineChunk(m, ptr, stack);
@@ -631,7 +654,7 @@ struct Allocator {
       // If realloc() races with free(), we may start copying freed memory.
       // However, we will report racy double-free later anyway.
       REAL(memcpy)(new_ptr, old_ptr, memcpy_size);
-      Deallocate(old_ptr, 0, stack, FROM_MALLOC);
+      Deallocate(old_ptr, 0, 0, stack, FROM_MALLOC);
     }
     return new_ptr;
   }
@@ -716,6 +739,22 @@ struct Allocator {
     return AsanChunkView(m1);
   }
 
+  void Purge() {
+    AsanThread *t = GetCurrentThread();
+    if (t) {
+      AsanThreadLocalMallocStorage *ms = &t->malloc_storage();
+      quarantine.DrainAndRecycle(GetQuarantineCache(ms),
+                                 QuarantineCallback(GetAllocatorCache(ms)));
+    }
+    {
+      SpinMutexLock l(&fallback_mutex);
+      quarantine.DrainAndRecycle(&fallback_quarantine_cache,
+                                 QuarantineCallback(&fallback_allocator_cache));
+    }
+
+    allocator.ForceReleaseToOS();
+  }
+
   void PrintStats() {
     allocator.PrintStats();
     quarantine.PrintStats();
@@ -750,6 +789,9 @@ bool AsanChunkView::IsQuarantined() const {
 uptr AsanChunkView::Beg() const { return chunk_->Beg(); }
 uptr AsanChunkView::End() const { return Beg() + UsedSize(); }
 uptr AsanChunkView::UsedSize() const { return chunk_->UsedSize(); }
+u32 AsanChunkView::UserRequestedAlignment() const {
+  return Allocator::ComputeUserAlignment(chunk_->user_requested_alignment_log);
+}
 uptr AsanChunkView::AllocTid() const { return chunk_->alloc_tid; }
 uptr AsanChunkView::FreeTid() const { return chunk_->free_tid; }
 AllocType AsanChunkView::GetAllocType() const {
@@ -775,7 +817,7 @@ StackTrace AsanChunkView::GetFreeStack() const {
 }
 
 void InitializeAllocator(const AllocatorOptions &options) {
-  instance.Initialize(options);
+  instance.InitLinkerInitialized(options);
 }
 
 void ReInitializeAllocator(const AllocatorOptions &options) {
@@ -802,12 +844,12 @@ void PrintInternalAllocatorStats() {
 }
 
 void asan_free(void *ptr, BufferedStackTrace *stack, AllocType alloc_type) {
-  instance.Deallocate(ptr, 0, stack, alloc_type);
+  instance.Deallocate(ptr, 0, 0, stack, alloc_type);
 }
 
-void asan_sized_free(void *ptr, uptr size, BufferedStackTrace *stack,
-                     AllocType alloc_type) {
-  instance.Deallocate(ptr, size, stack, alloc_type);
+void asan_delete(void *ptr, uptr size, uptr alignment,
+                 BufferedStackTrace *stack, AllocType alloc_type) {
+  instance.Deallocate(ptr, size, alignment, stack, alloc_type);
 }
 
 void *asan_malloc(uptr size, BufferedStackTrace *stack) {
@@ -823,7 +865,7 @@ void *asan_realloc(void *p, uptr size, BufferedStackTrace *stack) {
     return SetErrnoOnNull(instance.Allocate(size, 8, stack, FROM_MALLOC, true));
   if (size == 0) {
     if (flags()->allocator_frees_and_returns_null_on_realloc_zero) {
-      instance.Deallocate(p, 0, stack, FROM_MALLOC);
+      instance.Deallocate(p, 0, 0, stack, FROM_MALLOC);
       return nullptr;
     }
     // Allocate a size of 1 if we shouldn't free() on Realloc to 0
@@ -839,6 +881,10 @@ void *asan_valloc(uptr size, BufferedStackTrace *stack) {
 
 void *asan_pvalloc(uptr size, BufferedStackTrace *stack) {
   uptr PageSize = GetPageSizeCached();
+  if (UNLIKELY(CheckForPvallocOverflow(size, PageSize))) {
+    errno = errno_ENOMEM;
+    return AsanAllocator::FailureHandler::OnBadRequest();
+  }
   // pvalloc(0) should allocate one page.
   size = size ? RoundUpTo(size, PageSize) : PageSize;
   return SetErrnoOnNull(
@@ -1007,6 +1053,10 @@ uptr __sanitizer_get_allocated_size(const void *p) {
   return allocated_size;
 }
 
+void __sanitizer_purge_allocator() {
+  instance.Purge();
+}
+
 #if !SANITIZER_SUPPORTS_WEAK_HOOKS
 // Provide default (no-op) implementation of malloc hooks.
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_malloc_hook,
diff --git a/contrib/compiler-rt/lib/asan/asan_allocator.h b/contrib/compiler-rt/lib/asan/asan_allocator.h
index ad1aeb58a86b..26483db4c60e 100644
--- a/contrib/compiler-rt/lib/asan/asan_allocator.h
+++ b/contrib/compiler-rt/lib/asan/asan_allocator.h
@@ -58,6 +58,7 @@ class AsanChunkView {
   uptr Beg() const;            // First byte of user memory.
   uptr End() const;            // Last byte of user memory.
   uptr UsedSize() const;       // Size requested by the user.
+  u32 UserRequestedAlignment() const;  // Originally requested alignment.
   uptr AllocTid() const;
   uptr FreeTid() const;
   bool Eq(const AsanChunkView &c) const { return chunk_ == c.chunk_; }
@@ -119,7 +120,11 @@ struct AsanMapUnmapCallback {
 };
 
 #if SANITIZER_CAN_USE_ALLOCATOR64
-# if defined(__powerpc64__)
+# if SANITIZER_FUCHSIA
+const uptr kAllocatorSpace = ~(uptr)0;
+const uptr kAllocatorSize  =  0x40000000000ULL;  // 4T.
+typedef DefaultSizeClassMap SizeClassMap;
+# elif defined(__powerpc64__)
 const uptr kAllocatorSpace =  0xa0000000000ULL;
 const uptr kAllocatorSize  =  0x20000000000ULL;  // 2T.
 typedef DefaultSizeClassMap SizeClassMap;
@@ -193,8 +198,8 @@ struct AsanThreadLocalMallocStorage {
 void *asan_memalign(uptr alignment, uptr size, BufferedStackTrace *stack,
                     AllocType alloc_type);
 void asan_free(void *ptr, BufferedStackTrace *stack, AllocType alloc_type);
-void asan_sized_free(void *ptr, uptr size, BufferedStackTrace *stack,
-                     AllocType alloc_type);
+void asan_delete(void *ptr, uptr size, uptr alignment,
+                 BufferedStackTrace *stack, AllocType alloc_type);
 
 void *asan_malloc(uptr size, BufferedStackTrace *stack);
 void *asan_calloc(uptr nmemb, uptr size, BufferedStackTrace *stack);
diff --git a/contrib/compiler-rt/lib/asan/asan_descriptions.cc b/contrib/compiler-rt/lib/asan/asan_descriptions.cc
index 822a6a62d64e..86c6af7d9f75 100644
--- a/contrib/compiler-rt/lib/asan/asan_descriptions.cc
+++ b/contrib/compiler-rt/lib/asan/asan_descriptions.cc
@@ -122,6 +122,7 @@ static void GetAccessToHeapChunkInformation(ChunkAccess *descr,
   }
   descr->chunk_begin = chunk.Beg();
   descr->chunk_size = chunk.UsedSize();
+  descr->user_requested_alignment = chunk.UserRequestedAlignment();
   descr->alloc_type = chunk.GetAllocType();
 }
 
@@ -150,7 +151,7 @@ static void PrintHeapChunkAccess(uptr addr, const ChunkAccess &descr) {
   str.append(" %zu-byte region [%p,%p)\n", descr.chunk_size,
              (void *)descr.chunk_begin,
              (void *)(descr.chunk_begin + descr.chunk_size));
-  str.append("%s", d.EndLocation());
+  str.append("%s", d.Default());
   Printf("%s", str.data());
 }
 
@@ -260,7 +261,7 @@ static void PrintAccessAndVarIntersection(const StackVarDescr &var, uptr addr,
     // FIXME: we may want to also print the size of the access here,
     // but in case of accesses generated by memset it may be confusing.
     str.append("%s <== Memory access at offset %zd %s this variable%s\n",
-               d.Location(), addr, pos_descr, d.EndLocation());
+               d.Location(), addr, pos_descr, d.Default());
   } else {
     str.append("\n");
   }
@@ -295,7 +296,7 @@ static void DescribeAddressRelativeToGlobal(uptr addr, uptr access_size,
              MaybeDemangleGlobalName(g.name));
   PrintGlobalLocation(&str, g);
   str.append("' (0x%zx) of size %zu\n", g.beg, g.size);
-  str.append("%s", d.EndLocation());
+  str.append("%s", d.Default());
   PrintGlobalNameIfASCII(&str, g);
   Printf("%s", str.data());
 }
@@ -335,6 +336,26 @@ void GlobalAddressDescription::Print(const char *bug_type) const {
   }
 }
 
+bool GlobalAddressDescription::PointsInsideTheSameVariable(
+    const GlobalAddressDescription &other) const {
+  if (size == 0 || other.size == 0) return false;
+
+  for (uptr i = 0; i < size; i++) {
+    const __asan_global &a = globals[i];
+    for (uptr j = 0; j < other.size; j++) {
+      const __asan_global &b = other.globals[j];
+      if (a.beg == b.beg &&
+          a.beg <= addr &&
+          b.beg <= other.addr &&
+          (addr + access_size) < (a.beg + a.size) &&
+          (other.addr + other.access_size) < (b.beg + b.size))
+        return true;
+    }
+  }
+
+  return false;
+}
+
 void StackAddressDescription::Print() const {
   Decorator d;
   char tname[128];
@@ -343,10 +364,10 @@ void StackAddressDescription::Print() const {
          ThreadNameWithParenthesis(tid, tname, sizeof(tname)));
 
   if (!frame_descr) {
-    Printf("%s\n", d.EndLocation());
+    Printf("%s\n", d.Default());
     return;
   }
-  Printf(" at offset %zu in frame%s\n", offset, d.EndLocation());
+  Printf(" at offset %zu in frame%s\n", offset, d.Default());
 
   // Now we print the frame where the alloca has happened.
   // We print this frame as a stack trace with one element.
@@ -355,7 +376,7 @@ void StackAddressDescription::Print() const {
   // previously. That's unfortunate, but I have no better solution,
   // especially given that the alloca may be from entirely different place
   // (e.g. use-after-scope, or different thread's stack).
-  Printf("%s", d.EndLocation());
+  Printf("%s", d.Default());
   StackTrace alloca_stack(&frame_pc, 1);
   alloca_stack.Print();
 
@@ -405,18 +426,18 @@ void HeapAddressDescription::Print() const {
     Printf("%sfreed by thread T%d%s here:%s\n", d.Allocation(),
            free_thread->tid,
            ThreadNameWithParenthesis(free_thread, tname, sizeof(tname)),
-           d.EndAllocation());
+           d.Default());
     StackTrace free_stack = GetStackTraceFromId(free_stack_id);
     free_stack.Print();
     Printf("%spreviously allocated by thread T%d%s here:%s\n", d.Allocation(),
            alloc_thread->tid,
            ThreadNameWithParenthesis(alloc_thread, tname, sizeof(tname)),
-           d.EndAllocation());
+           d.Default());
   } else {
     Printf("%sallocated by thread T%d%s here:%s\n", d.Allocation(),
            alloc_thread->tid,
            ThreadNameWithParenthesis(alloc_thread, tname, sizeof(tname)),
-           d.EndAllocation());
+           d.Default());
   }
   alloc_stack.Print();
   DescribeThread(GetCurrentThread());
diff --git a/contrib/compiler-rt/lib/asan/asan_descriptions.h b/contrib/compiler-rt/lib/asan/asan_descriptions.h
index 0ee677eb7d0e..1a1b01cf20cd 100644
--- a/contrib/compiler-rt/lib/asan/asan_descriptions.h
+++ b/contrib/compiler-rt/lib/asan/asan_descriptions.h
@@ -34,11 +34,8 @@ class Decorator : public __sanitizer::SanitizerCommonDecorator {
  public:
   Decorator() : SanitizerCommonDecorator() {}
   const char *Access() { return Blue(); }
-  const char *EndAccess() { return Default(); }
   const char *Location() { return Green(); }
-  const char *EndLocation() { return Default(); }
   const char *Allocation() { return Magenta(); }
-  const char *EndAllocation() { return Default(); }
 
   const char *ShadowByte(u8 byte) {
     switch (byte) {
@@ -72,9 +69,6 @@ class Decorator : public __sanitizer::SanitizerCommonDecorator {
         return Default();
     }
   }
-  const char *EndShadowByte() { return Default(); }
-  const char *MemoryByte() { return Magenta(); }
-  const char *EndMemoryByte() { return Default(); }
 };
 
 enum ShadowKind : u8 {
@@ -108,6 +102,7 @@ struct ChunkAccess {
   sptr offset;
   uptr chunk_begin;
   uptr chunk_size;
+  u32 user_requested_alignment : 12;
   u32 access_type : 2;
   u32 alloc_type : 2;
 };
@@ -151,6 +146,10 @@ struct GlobalAddressDescription {
   u8 size;
 
   void Print(const char *bug_type = "") const;
+
+  // Returns true when this descriptions points inside the same global variable
+  // as other. Descriptions can have different address within the variable
+  bool PointsInsideTheSameVariable(const GlobalAddressDescription &other) const;
 };
 
 bool GetGlobalAddressInformation(uptr addr, uptr access_size,
diff --git a/contrib/compiler-rt/lib/asan/asan_errors.cc b/contrib/compiler-rt/lib/asan/asan_errors.cc
index b7a38eb7cece..0f4a3abff9a7 100644
--- a/contrib/compiler-rt/lib/asan/asan_errors.cc
+++ b/contrib/compiler-rt/lib/asan/asan_errors.cc
@@ -13,7 +13,6 @@
 //===----------------------------------------------------------------------===//
 
 #include "asan_errors.h"
-#include <signal.h>
 #include "asan_descriptions.h"
 #include "asan_mapping.h"
 #include "asan_report.h"
@@ -22,82 +21,26 @@
 
 namespace __asan {
 
-void ErrorStackOverflow::Print() {
-  Decorator d;
-  Printf("%s", d.Warning());
-  Report(
-      "ERROR: AddressSanitizer: %s on address %p"
-      " (pc %p bp %p sp %p T%d)\n", scariness.GetDescription(),
-      (void *)addr, (void *)pc, (void *)bp, (void *)sp, tid);
-  Printf("%s", d.EndWarning());
-  scariness.Print();
-  BufferedStackTrace stack;
-  GetStackTraceWithPcBpAndContext(&stack, kStackTraceMax, pc, bp, context,
-                                  common_flags()->fast_unwind_on_fatal);
-  stack.Print();
-  ReportErrorSummary(scariness.GetDescription(), &stack);
-}
-
-static void MaybeDumpInstructionBytes(uptr pc) {
-  if (!flags()->dump_instruction_bytes || (pc < GetPageSizeCached())) return;
-  InternalScopedString str(1024);
-  str.append("First 16 instruction bytes at pc: ");
-  if (IsAccessibleMemoryRange(pc, 16)) {
-    for (int i = 0; i < 16; ++i) {
-      PrintMemoryByte(&str, "", ((u8 *)pc)[i], /*in_shadow*/ false, " ");
-    }
-    str.append("\n");
-  } else {
-    str.append("unaccessible\n");
-  }
-  Report("%s", str.data());
-}
-
-static void MaybeDumpRegisters(void *context) {
-  if (!flags()->dump_registers) return;
-  SignalContext::DumpAllRegisters(context);
-}
-
-static void MaybeReportNonExecRegion(uptr pc) {
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
-  MemoryMappingLayout proc_maps(/*cache_enabled*/ true);
-  MemoryMappedSegment segment;
-  while (proc_maps.Next(&segment)) {
-    if (pc >= segment.start && pc < segment.end && !segment.IsExecutable())
-      Report("Hint: PC is at a non-executable region. Maybe a wild jump?\n");
-  }
+static void OnStackUnwind(const SignalContext &sig,
+                          const void *callback_context,
+                          BufferedStackTrace *stack) {
+  bool fast = common_flags()->fast_unwind_on_fatal;
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
+  // On FreeBSD the slow unwinding that leverages _Unwind_Backtrace()
+  // yields the call stack of the signal's handler and not of the code
+  // that raised the signal (as it does on Linux).
+  fast = true;
 #endif
+  // Tests and maybe some users expect that scariness is going to be printed
+  // just before the stack. As only asan has scariness score we have no
+  // corresponding code in the sanitizer_common and we use this callback to
+  // print it.
+  static_cast<const ScarinessScoreBase *>(callback_context)->Print();
+  GetStackTrace(stack, kStackTraceMax, sig.pc, sig.bp, sig.context, fast);
 }
 
 void ErrorDeadlySignal::Print() {
-  Decorator d;
-  Printf("%s", d.Warning());
-  const char *description = __sanitizer::DescribeSignalOrException(signo);
-  Report(
-      "ERROR: AddressSanitizer: %s on unknown address %p (pc %p bp %p sp %p "
-      "T%d)\n",
-      description, (void *)addr, (void *)pc, (void *)bp, (void *)sp, tid);
-  Printf("%s", d.EndWarning());
-  if (pc < GetPageSizeCached()) Report("Hint: pc points to the zero page.\n");
-  if (is_memory_access) {
-    const char *access_type =
-        write_flag == SignalContext::WRITE
-            ? "WRITE"
-            : (write_flag == SignalContext::READ ? "READ" : "UNKNOWN");
-    Report("The signal is caused by a %s memory access.\n", access_type);
-    if (addr < GetPageSizeCached())
-      Report("Hint: address points to the zero page.\n");
-  }
-  MaybeReportNonExecRegion(pc);
-  scariness.Print();
-  BufferedStackTrace stack;
-  GetStackTraceWithPcBpAndContext(&stack, kStackTraceMax, pc, bp, context,
-                                  common_flags()->fast_unwind_on_fatal);
-  stack.Print();
-  MaybeDumpInstructionBytes(pc);
-  MaybeDumpRegisters(context);
-  Printf("AddressSanitizer can not provide additional info.\n");
-  ReportErrorSummary(description, &stack);
+  ReportDeadlySignal(signal, tid, &OnStackUnwind, &scariness);
 }
 
 void ErrorDoubleFree::Print() {
@@ -109,7 +52,7 @@ void ErrorDoubleFree::Print() {
       "thread T%d%s:\n",
       scariness.GetDescription(), addr_description.addr, tid,
       ThreadNameWithParenthesis(tid, tname, sizeof(tname)));
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
   scariness.Print();
   GET_STACK_TRACE_FATAL(second_free_stack->trace[0],
                         second_free_stack->top_frame_bp);
@@ -118,7 +61,7 @@ void ErrorDoubleFree::Print() {
   ReportErrorSummary(scariness.GetDescription(), &stack);
 }
 
-void ErrorNewDeleteSizeMismatch::Print() {
+void ErrorNewDeleteTypeMismatch::Print() {
   Decorator d;
   Printf("%s", d.Warning());
   char tname[128];
@@ -127,11 +70,29 @@ void ErrorNewDeleteSizeMismatch::Print() {
       "T%d%s:\n",
       scariness.GetDescription(), addr_description.addr, tid,
       ThreadNameWithParenthesis(tid, tname, sizeof(tname)));
-  Printf("%s  object passed to delete has wrong type:\n", d.EndWarning());
-  Printf(
-      "  size of the allocated type:   %zd bytes;\n"
-      "  size of the deallocated type: %zd bytes.\n",
-      addr_description.chunk_access.chunk_size, delete_size);
+  Printf("%s  object passed to delete has wrong type:\n", d.Default());
+  if (delete_size != 0) {
+    Printf(
+        "  size of the allocated type:   %zd bytes;\n"
+        "  size of the deallocated type: %zd bytes.\n",
+        addr_description.chunk_access.chunk_size, delete_size);
+  }
+  const uptr user_alignment =
+      addr_description.chunk_access.user_requested_alignment;
+  if (delete_alignment != user_alignment) {
+    char user_alignment_str[32];
+    char delete_alignment_str[32];
+    internal_snprintf(user_alignment_str, sizeof(user_alignment_str),
+                      "%zd bytes", user_alignment);
+    internal_snprintf(delete_alignment_str, sizeof(delete_alignment_str),
+                      "%zd bytes", delete_alignment);
+    static const char *kDefaultAlignment = "default-aligned";
+    Printf(
+        "  alignment of the allocated type:   %s;\n"
+        "  alignment of the deallocated type: %s.\n",
+        user_alignment > 0 ? user_alignment_str : kDefaultAlignment,
+        delete_alignment > 0 ? delete_alignment_str : kDefaultAlignment);
+  }
   CHECK_GT(free_stack->size, 0);
   scariness.Print();
   GET_STACK_TRACE_FATAL(free_stack->trace[0], free_stack->top_frame_bp);
@@ -152,7 +113,7 @@ void ErrorFreeNotMalloced::Print() {
       "which was not malloc()-ed: %p in thread T%d%s\n",
       addr_description.Address(), tid,
       ThreadNameWithParenthesis(tid, tname, sizeof(tname)));
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
   CHECK_GT(free_stack->size, 0);
   scariness.Print();
   GET_STACK_TRACE_FATAL(free_stack->trace[0], free_stack->top_frame_bp);
@@ -173,7 +134,7 @@ void ErrorAllocTypeMismatch::Print() {
          scariness.GetDescription(),
          alloc_names[alloc_type], dealloc_names[dealloc_type],
          addr_description.addr);
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
   CHECK_GT(dealloc_stack->size, 0);
   scariness.Print();
   GET_STACK_TRACE_FATAL(dealloc_stack->trace[0], dealloc_stack->top_frame_bp);
@@ -192,7 +153,7 @@ void ErrorMallocUsableSizeNotOwned::Print() {
       "ERROR: AddressSanitizer: attempting to call malloc_usable_size() for "
       "pointer which is not owned: %p\n",
       addr_description.Address());
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
   stack->Print();
   addr_description.Print();
   ReportErrorSummary(scariness.GetDescription(), stack);
@@ -205,7 +166,7 @@ void ErrorSanitizerGetAllocatedSizeNotOwned::Print() {
       "ERROR: AddressSanitizer: attempting to call "
       "__sanitizer_get_allocated_size() for pointer which is not owned: %p\n",
       addr_description.Address());
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
   stack->Print();
   addr_description.Print();
   ReportErrorSummary(scariness.GetDescription(), stack);
@@ -222,7 +183,7 @@ void ErrorStringFunctionMemoryRangesOverlap::Print() {
       bug_type, addr1_description.Address(),
       addr1_description.Address() + length1, addr2_description.Address(),
       addr2_description.Address() + length2);
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
   scariness.Print();
   stack->Print();
   addr1_description.Print();
@@ -235,7 +196,7 @@ void ErrorStringFunctionSizeOverflow::Print() {
   Printf("%s", d.Warning());
   Report("ERROR: AddressSanitizer: %s: (size=%zd)\n",
          scariness.GetDescription(), size);
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
   scariness.Print();
   stack->Print();
   addr_description.Print();
@@ -263,7 +224,7 @@ void ErrorODRViolation::Print() {
   Printf("%s", d.Warning());
   Report("ERROR: AddressSanitizer: %s (%p):\n", scariness.GetDescription(),
          global1.beg);
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
   InternalScopedString g1_loc(256), g2_loc(256);
   PrintGlobalLocation(&g1_loc, global1);
   PrintGlobalLocation(&g2_loc, global2);
@@ -292,7 +253,7 @@ void ErrorInvalidPointerPair::Print() {
   Printf("%s", d.Warning());
   Report("ERROR: AddressSanitizer: %s: %p %p\n", scariness.GetDescription(),
          addr1_description.Address(), addr2_description.Address());
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
   GET_STACK_TRACE_FATAL(pc, bp);
   stack.Print();
   addr1_description.Print();
@@ -477,9 +438,14 @@ static void PrintShadowMemoryForAddress(uptr addr) {
   InternalScopedString str(4096 * 8);
   str.append("Shadow bytes around the buggy address:\n");
   for (int i = -5; i <= 5; i++) {
+    uptr row_shadow_addr = aligned_shadow + i * n_bytes_per_row;
+    // Skip rows that would be outside the shadow range. This can happen when
+    // the user address is near the bottom, top, or shadow gap of the address
+    // space.
+    if (!AddrIsInShadow(row_shadow_addr)) continue;
     const char *prefix = (i == 0) ? "=>" : "  ";
-    PrintShadowBytes(&str, prefix, (u8 *)(aligned_shadow + i * n_bytes_per_row),
-                     (u8 *)shadow_addr, n_bytes_per_row);
+    PrintShadowBytes(&str, prefix, (u8 *)row_shadow_addr, (u8 *)shadow_addr,
+                     n_bytes_per_row);
   }
   if (flags()->print_legend) PrintLegend(&str);
   Printf("%s", str.data());
@@ -491,13 +457,13 @@ void ErrorGeneric::Print() {
   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);
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
 
   char tname[128];
   Printf("%s%s of size %zu at %p thread T%d%s%s\n", d.Access(),
          access_size ? (is_write ? "WRITE" : "READ") : "ACCESS", access_size,
          (void *)addr, tid,
-         ThreadNameWithParenthesis(tid, tname, sizeof(tname)), d.EndAccess());
+         ThreadNameWithParenthesis(tid, tname, sizeof(tname)), d.Default());
 
   scariness.Print();
   GET_STACK_TRACE_FATAL(pc, bp);
diff --git a/contrib/compiler-rt/lib/asan/asan_errors.h b/contrib/compiler-rt/lib/asan/asan_errors.h
index 9a124924470e..518ba0c6945e 100644
--- a/contrib/compiler-rt/lib/asan/asan_errors.h
+++ b/contrib/compiler-rt/lib/asan/asan_errors.h
@@ -27,61 +27,28 @@ struct ErrorBase {
   u32 tid;
 };
 
-struct ErrorStackOverflow : ErrorBase {
-  uptr addr, pc, bp, sp;
-  // ErrorStackOverflow never owns the context.
-  void *context;
-  // VS2013 doesn't implement unrestricted unions, so we need a trivial default
-  // constructor
-  ErrorStackOverflow() = default;
-  ErrorStackOverflow(u32 tid, const SignalContext &sig)
-      : ErrorBase(tid),
-        addr(sig.addr),
-        pc(sig.pc),
-        bp(sig.bp),
-        sp(sig.sp),
-        context(sig.context) {
-    scariness.Clear();
-    scariness.Scare(10, "stack-overflow");
-  }
-  void Print();
-};
-
 struct ErrorDeadlySignal : ErrorBase {
-  uptr addr, pc, bp, sp;
-  // ErrorDeadlySignal never owns the context.
-  void *context;
-  int signo;
-  SignalContext::WriteFlag write_flag;
-  bool is_memory_access;
+  SignalContext signal;
   // VS2013 doesn't implement unrestricted unions, so we need a trivial default
   // constructor
   ErrorDeadlySignal() = default;
-  ErrorDeadlySignal(u32 tid, const SignalContext &sig, int signo_)
-      : ErrorBase(tid),
-        addr(sig.addr),
-        pc(sig.pc),
-        bp(sig.bp),
-        sp(sig.sp),
-        context(sig.context),
-        signo(signo_),
-        write_flag(sig.write_flag),
-        is_memory_access(sig.is_memory_access) {
+  ErrorDeadlySignal(u32 tid, const SignalContext &sig)
+      : ErrorBase(tid), signal(sig) {
     scariness.Clear();
-    if (is_memory_access) {
-      if (addr < GetPageSizeCached()) {
-        scariness.Scare(10, "null-deref");
-      } else if (addr == pc) {
-        scariness.Scare(60, "wild-jump");
-      } else if (write_flag == SignalContext::WRITE) {
-        scariness.Scare(30, "wild-addr-write");
-      } else if (write_flag == SignalContext::READ) {
-        scariness.Scare(20, "wild-addr-read");
-      } else {
-        scariness.Scare(25, "wild-addr");
-      }
-    } else {
+    if (signal.IsStackOverflow()) {
+      scariness.Scare(10, "stack-overflow");
+    } else if (!signal.is_memory_access) {
       scariness.Scare(10, "signal");
+    } else if (signal.addr < GetPageSizeCached()) {
+      scariness.Scare(10, "null-deref");
+    } else if (signal.addr == signal.pc) {
+      scariness.Scare(60, "wild-jump");
+    } else if (signal.write_flag == SignalContext::WRITE) {
+      scariness.Scare(30, "wild-addr-write");
+    } else if (signal.write_flag == SignalContext::READ) {
+      scariness.Scare(20, "wild-addr-read");
+    } else {
+      scariness.Scare(25, "wild-addr");
     }
   }
   void Print();
@@ -104,17 +71,19 @@ struct ErrorDoubleFree : ErrorBase {
   void Print();
 };
 
-struct ErrorNewDeleteSizeMismatch : ErrorBase {
-  // ErrorNewDeleteSizeMismatch doesn't own the stack trace.
+struct ErrorNewDeleteTypeMismatch : ErrorBase {
+  // ErrorNewDeleteTypeMismatch doesn't own the stack trace.
   const BufferedStackTrace *free_stack;
   HeapAddressDescription addr_description;
   uptr delete_size;
+  uptr delete_alignment;
   // VS2013 doesn't implement unrestricted unions, so we need a trivial default
   // constructor
-  ErrorNewDeleteSizeMismatch() = default;
-  ErrorNewDeleteSizeMismatch(u32 tid, BufferedStackTrace *stack, uptr addr,
-                             uptr delete_size_)
-      : ErrorBase(tid), free_stack(stack), delete_size(delete_size_) {
+  ErrorNewDeleteTypeMismatch() = default;
+  ErrorNewDeleteTypeMismatch(u32 tid, BufferedStackTrace *stack, uptr addr,
+                             uptr delete_size_, uptr delete_alignment_)
+      : ErrorBase(tid), free_stack(stack), delete_size(delete_size_),
+        delete_alignment(delete_alignment_) {
     GetHeapAddressInformation(addr, 1, &addr_description);
     scariness.Clear();
     scariness.Scare(10, "new-delete-type-mismatch");
@@ -324,10 +293,9 @@ struct ErrorGeneric : ErrorBase {
 
 // clang-format off
 #define ASAN_FOR_EACH_ERROR_KIND(macro)         \
-  macro(StackOverflow)                          \
   macro(DeadlySignal)                           \
   macro(DoubleFree)                             \
-  macro(NewDeleteSizeMismatch)                  \
+  macro(NewDeleteTypeMismatch)                  \
   macro(FreeNotMalloced)                        \
   macro(AllocTypeMismatch)                      \
   macro(MallocUsableSizeNotOwned)               \
diff --git a/contrib/compiler-rt/lib/asan/asan_fake_stack.cc b/contrib/compiler-rt/lib/asan/asan_fake_stack.cc
index 017b7d2af129..1c6184e3c7fc 100644
--- a/contrib/compiler-rt/lib/asan/asan_fake_stack.cc
+++ b/contrib/compiler-rt/lib/asan/asan_fake_stack.cc
@@ -28,9 +28,9 @@ static const u64 kAllocaRedzoneMask = 31UL;
 
 // For small size classes inline PoisonShadow for better performance.
 ALWAYS_INLINE void SetShadow(uptr ptr, uptr size, uptr class_id, u64 magic) {
-  CHECK_EQ(SHADOW_SCALE, 3);  // This code expects SHADOW_SCALE=3.
   u64 *shadow = reinterpret_cast<u64*>(MemToShadow(ptr));
-  if (class_id <= 6) {
+  if (SHADOW_SCALE == 3 && class_id <= 6) {
+    // This code expects SHADOW_SCALE=3.
     for (uptr i = 0; i < (((uptr)1) << class_id); i++) {
       shadow[i] = magic;
       // Make sure this does not become memset.
@@ -171,7 +171,7 @@ void FakeStack::ForEachFakeFrame(RangeIteratorCallback callback, void *arg) {
   }
 }
 
-#if SANITIZER_LINUX && !SANITIZER_ANDROID
+#if (SANITIZER_LINUX && !SANITIZER_ANDROID) || SANITIZER_FUCHSIA
 static THREADLOCAL FakeStack *fake_stack_tls;
 
 FakeStack *GetTLSFakeStack() {
@@ -183,7 +183,7 @@ void SetTLSFakeStack(FakeStack *fs) {
 #else
 FakeStack *GetTLSFakeStack() { return 0; }
 void SetTLSFakeStack(FakeStack *fs) { }
-#endif  // SANITIZER_LINUX && !SANITIZER_ANDROID
+#endif  // (SANITIZER_LINUX && !SANITIZER_ANDROID) || SANITIZER_FUCHSIA
 
 static FakeStack *GetFakeStack() {
   AsanThread *t = GetCurrentThread();
diff --git a/contrib/compiler-rt/lib/asan/asan_flags.cc b/contrib/compiler-rt/lib/asan/asan_flags.cc
index 6be0d6e94b9a..562168e6d428 100644
--- a/contrib/compiler-rt/lib/asan/asan_flags.cc
+++ b/contrib/compiler-rt/lib/asan/asan_flags.cc
@@ -118,6 +118,10 @@ void InitializeFlags() {
   const char *ubsan_default_options = __ubsan::MaybeCallUbsanDefaultOptions();
   ubsan_parser.ParseString(ubsan_default_options);
 #endif
+#if CAN_SANITIZE_LEAKS
+  const char *lsan_default_options = __lsan::MaybeCallLsanDefaultOptions();
+  lsan_parser.ParseString(lsan_default_options);
+#endif
 
   // Override from command line.
   asan_parser.ParseString(GetEnv("ASAN_OPTIONS"));
@@ -144,6 +148,9 @@ void InitializeFlags() {
            SanitizerToolName);
     Die();
   }
+  // Ensure that redzone is at least SHADOW_GRANULARITY.
+  if (f->redzone < (int)SHADOW_GRANULARITY)
+    f->redzone = SHADOW_GRANULARITY;
   // Make "strict_init_order" imply "check_initialization_order".
   // TODO(samsonov): Use a single runtime flag for an init-order checker.
   if (f->strict_init_order) {
diff --git a/contrib/compiler-rt/lib/asan/asan_flags.inc b/contrib/compiler-rt/lib/asan/asan_flags.inc
index f2216c2e9b3b..00071d39f041 100644
--- a/contrib/compiler-rt/lib/asan/asan_flags.inc
+++ b/contrib/compiler-rt/lib/asan/asan_flags.inc
@@ -79,6 +79,10 @@ ASAN_FLAG(
     "Number of seconds to sleep between printing an error report and "
     "terminating the program. Useful for debugging purposes (e.g. when one "
     "needs to attach gdb).")
+ASAN_FLAG(
+    int, sleep_after_init, 0,
+    "Number of seconds to sleep after AddressSanitizer is initialized. "
+    "Useful for debugging purposes (e.g. when one needs to attach gdb).")
 ASAN_FLAG(bool, check_malloc_usable_size, true,
           "Allows the users to work around the bug in Nvidia drivers prior to "
           "295.*.")
@@ -143,11 +147,6 @@ ASAN_FLAG(int, detect_odr_violation, 2,
           "If >=2, detect violation of One-Definition-Rule (ODR); "
           "If ==1, detect ODR-violation only if the two variables "
           "have different sizes")
-ASAN_FLAG(bool, dump_instruction_bytes, false,
-          "If true, dump 16 bytes starting at the instruction that caused SEGV")
-ASAN_FLAG(bool, dump_registers, true,
-          "If true, dump values of CPU registers when SEGV happens. Only "
-          "available on OS X for now.")
 ASAN_FLAG(const char *, suppressions, "", "Suppressions file name.")
 ASAN_FLAG(bool, halt_on_error, true,
           "Crash the program after printing the first error report "
diff --git a/contrib/compiler-rt/lib/asan/asan_fuchsia.cc b/contrib/compiler-rt/lib/asan/asan_fuchsia.cc
new file mode 100644
index 000000000000..0b5bff4f565e
--- /dev/null
+++ b/contrib/compiler-rt/lib/asan/asan_fuchsia.cc
@@ -0,0 +1,218 @@
+//===-- asan_fuchsia.cc --------------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// This file is a part of AddressSanitizer, an address sanity checker.
+//
+// Fuchsia-specific details.
+//===---------------------------------------------------------------------===//
+
+#include "sanitizer_common/sanitizer_fuchsia.h"
+#if SANITIZER_FUCHSIA
+
+#include "asan_interceptors.h"
+#include "asan_internal.h"
+#include "asan_stack.h"
+#include "asan_thread.h"
+
+#include <limits.h>
+#include <zircon/sanitizer.h>
+#include <zircon/syscalls.h>
+#include <zircon/threads.h>
+
+namespace __asan {
+
+// The system already set up the shadow memory for us.
+// __sanitizer::GetMaxUserVirtualAddress has already been called by
+// AsanInitInternal->InitializeHighMemEnd (asan_rtl.cc).
+// Just do some additional sanity checks here.
+void InitializeShadowMemory() {
+  if (Verbosity()) PrintAddressSpaceLayout();
+
+  // Make sure SHADOW_OFFSET doesn't use __asan_shadow_memory_dynamic_address.
+  __asan_shadow_memory_dynamic_address = kDefaultShadowSentinel;
+  DCHECK(kLowShadowBeg != kDefaultShadowSentinel);
+  __asan_shadow_memory_dynamic_address = kLowShadowBeg;
+
+  CHECK_EQ(kShadowGapEnd, kHighShadowBeg - 1);
+  CHECK_EQ(kHighMemEnd, __sanitizer::ShadowBounds.memory_limit - 1);
+  CHECK_EQ(kHighMemBeg, __sanitizer::ShadowBounds.shadow_limit);
+  CHECK_EQ(kHighShadowBeg, __sanitizer::ShadowBounds.shadow_base);
+  CHECK_EQ(kShadowGapEnd, __sanitizer::ShadowBounds.shadow_base - 1);
+  CHECK_EQ(kLowShadowEnd, 0);
+  CHECK_EQ(kLowShadowBeg, 0);
+}
+
+void AsanApplyToGlobals(globals_op_fptr op, const void *needle) {
+  UNIMPLEMENTED();
+}
+
+void AsanCheckDynamicRTPrereqs() {}
+void AsanCheckIncompatibleRT() {}
+void InitializeAsanInterceptors() {}
+
+void *AsanDoesNotSupportStaticLinkage() { return nullptr; }
+
+void InitializePlatformExceptionHandlers() {}
+void AsanOnDeadlySignal(int signo, void *siginfo, void *context) {
+  UNIMPLEMENTED();
+}
+
+// We can use a plain thread_local variable for TSD.
+static thread_local void *per_thread;
+
+void *AsanTSDGet() { return per_thread; }
+
+void AsanTSDSet(void *tsd) { per_thread = tsd; }
+
+// There's no initialization needed, and the passed-in destructor
+// will never be called.  Instead, our own thread destruction hook
+// (below) will call AsanThread::TSDDtor directly.
+void AsanTSDInit(void (*destructor)(void *tsd)) {
+  DCHECK(destructor == &PlatformTSDDtor);
+}
+
+void PlatformTSDDtor(void *tsd) { UNREACHABLE(__func__); }
+
+static inline size_t AsanThreadMmapSize() {
+  return RoundUpTo(sizeof(AsanThread), PAGE_SIZE);
+}
+
+struct AsanThread::InitOptions {
+  uptr stack_bottom, stack_size;
+};
+
+// 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, uptr stack_bottom,
+                                    uptr stack_size) {
+  // 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);
+  asanThreadRegistry().SetThreadName(tid, name);
+
+  // On other systems, AsanThread::Init() is called from the new
+  // thread itself.  But on Fuchsia we already know the stack address
+  // range beforehand, so we can do most of the setup right now.
+  const AsanThread::InitOptions options = {stack_bottom, stack_size};
+  thread->Init(&options);
+
+  return thread;
+}
+
+// This gets the same arguments passed to Init by CreateAsanThread, above.
+// We're in the creator thread before the new thread is actually started,
+// but its stack address range is already known.  We don't bother tracking
+// the static TLS address range because the system itself already uses an
+// ASan-aware allocator for that.
+void AsanThread::SetThreadStackAndTls(const AsanThread::InitOptions *options) {
+  DCHECK_NE(GetCurrentThread(), this);
+  DCHECK_NE(GetCurrentThread(), nullptr);
+  CHECK_NE(options->stack_bottom, 0);
+  CHECK_NE(options->stack_size, 0);
+  stack_bottom_ = options->stack_bottom;
+  stack_top_ = options->stack_bottom + options->stack_size;
+}
+
+// Called by __asan::AsanInitInternal (asan_rtl.c).
+AsanThread *CreateMainThread() {
+  thrd_t self = thrd_current();
+  char name[ZX_MAX_NAME_LEN];
+  CHECK_NE(__sanitizer::MainThreadStackBase, 0);
+  CHECK_GT(__sanitizer::MainThreadStackSize, 0);
+  AsanThread *t = CreateAsanThread(
+      nullptr, 0, reinterpret_cast<uptr>(self), true,
+      _zx_object_get_property(thrd_get_zx_handle(self), ZX_PROP_NAME, name,
+                              sizeof(name)) == ZX_OK
+          ? name
+          : nullptr,
+      __sanitizer::MainThreadStackBase, __sanitizer::MainThreadStackSize);
+  SetCurrentThread(t);
+  return t;
+}
+
+// This is called before each thread creation is attempted.  So, in
+// its first call, the calling thread is the initial and sole thread.
+static void *BeforeThreadCreateHook(uptr user_id, bool detached,
+                                    const char *name, uptr stack_bottom,
+                                    uptr stack_size) {
+  EnsureMainThreadIDIsCorrect();
+  // Strict init-order checking is thread-hostile.
+  if (flags()->strict_init_order) StopInitOrderChecking();
+
+  GET_STACK_TRACE_THREAD;
+  u32 parent_tid = GetCurrentTidOrInvalid();
+
+  return CreateAsanThread(&stack, parent_tid, user_id, detached, name,
+                          stack_bottom, stack_size);
+}
+
+// This is called after creating a new thread (in the creating thread),
+// with the pointer returned by BeforeThreadCreateHook (above).
+static void ThreadCreateHook(void *hook, bool aborted) {
+  AsanThread *thread = static_cast<AsanThread *>(hook);
+  if (!aborted) {
+    // The thread was created successfully.
+    // ThreadStartHook is already running in the new thread.
+  } else {
+    // The thread wasn't created after all.
+    // Clean up everything we set up in BeforeThreadCreateHook.
+    asanThreadRegistry().FinishThread(thread->tid());
+    UnmapOrDie(thread, AsanThreadMmapSize());
+  }
+}
+
+// This is called in the newly-created thread before it runs anything else,
+// with the pointer returned by BeforeThreadCreateHook (above).
+// cf. asan_interceptors.cc:asan_thread_start
+static void ThreadStartHook(void *hook, uptr os_id) {
+  AsanThread *thread = static_cast<AsanThread *>(hook);
+  SetCurrentThread(thread);
+
+  // In lieu of AsanThread::ThreadStart.
+  asanThreadRegistry().StartThread(thread->tid(), os_id, /*workerthread*/ false,
+                                   nullptr);
+}
+
+// Each thread runs this just before it exits,
+// with the pointer returned by BeforeThreadCreateHook (above).
+// All per-thread destructors have already been called.
+static void ThreadExitHook(void *hook, uptr os_id) {
+  AsanThread::TSDDtor(per_thread);
+}
+
+}  // namespace __asan
+
+// These are declared (in extern "C") by <zircon/sanitizer.h>.
+// The system runtime will call our definitions directly.
+
+void *__sanitizer_before_thread_create_hook(thrd_t thread, bool detached,
+                                            const char *name, void *stack_base,
+                                            size_t stack_size) {
+  return __asan::BeforeThreadCreateHook(
+      reinterpret_cast<uptr>(thread), detached, name,
+      reinterpret_cast<uptr>(stack_base), stack_size);
+}
+
+void __sanitizer_thread_create_hook(void *hook, thrd_t thread, int error) {
+  __asan::ThreadCreateHook(hook, error != thrd_success);
+}
+
+void __sanitizer_thread_start_hook(void *hook, thrd_t self) {
+  __asan::ThreadStartHook(hook, reinterpret_cast<uptr>(self));
+}
+
+void __sanitizer_thread_exit_hook(void *hook, thrd_t self) {
+  __asan::ThreadExitHook(hook, reinterpret_cast<uptr>(self));
+}
+
+#endif  // SANITIZER_FUCHSIA
diff --git a/contrib/compiler-rt/lib/asan/asan_globals.cc b/contrib/compiler-rt/lib/asan/asan_globals.cc
index eebada804f07..0db65d010349 100644
--- a/contrib/compiler-rt/lib/asan/asan_globals.cc
+++ b/contrib/compiler-rt/lib/asan/asan_globals.cc
@@ -384,6 +384,10 @@ void __asan_register_globals(__asan_global *globals, uptr n) {
     }
     RegisterGlobal(&globals[i]);
   }
+
+  // Poison the metadata. It should not be accessible to user code.
+  PoisonShadow(reinterpret_cast<uptr>(globals), n * sizeof(__asan_global),
+               kAsanGlobalRedzoneMagic);
 }
 
 // Unregister an array of globals.
@@ -399,6 +403,9 @@ void __asan_unregister_globals(__asan_global *globals, uptr n) {
     }
     UnregisterGlobal(&globals[i]);
   }
+
+  // Unpoison the metadata.
+  PoisonShadow(reinterpret_cast<uptr>(globals), n * sizeof(__asan_global), 0);
 }
 
 // This method runs immediately prior to dynamic initialization in each TU,
diff --git a/contrib/compiler-rt/lib/asan/asan_init_version.h b/contrib/compiler-rt/lib/asan/asan_init_version.h
index f48cc19cc515..c49fcd740248 100644
--- a/contrib/compiler-rt/lib/asan/asan_init_version.h
+++ b/contrib/compiler-rt/lib/asan/asan_init_version.h
@@ -15,6 +15,8 @@
 #ifndef ASAN_INIT_VERSION_H
 #define ASAN_INIT_VERSION_H
 
+#include "sanitizer_common/sanitizer_platform.h"
+
 extern "C" {
   // Every time the ASan ABI changes we also change the version number in the
   // __asan_init function name.  Objects built with incompatible ASan ABI
@@ -32,7 +34,12 @@ extern "C" {
   // v6=>v7: added 'odr_indicator' to __asan_global
   // v7=>v8: added '__asan_(un)register_image_globals' functions for dead
   //         stripping support on Mach-O platforms
+#if SANITIZER_WORDSIZE == 32 && SANITIZER_ANDROID
+  // v8=>v9: 32-bit Android switched to dynamic shadow
+  #define __asan_version_mismatch_check __asan_version_mismatch_check_v9
+#else
   #define __asan_version_mismatch_check __asan_version_mismatch_check_v8
+#endif
 }
 
 #endif  // ASAN_INIT_VERSION_H
diff --git a/contrib/compiler-rt/lib/asan/asan_interceptors.cc b/contrib/compiler-rt/lib/asan/asan_interceptors.cc
index 34ca22b8616e..cb7dcb3b1ca8 100644
--- a/contrib/compiler-rt/lib/asan/asan_interceptors.cc
+++ b/contrib/compiler-rt/lib/asan/asan_interceptors.cc
@@ -24,6 +24,11 @@
 #include "lsan/lsan_common.h"
 #include "sanitizer_common/sanitizer_libc.h"
 
+// There is no general interception at all on Fuchsia.
+// Only the functions in asan_interceptors_memintrinsics.cc are
+// really defined to replace libc functions.
+#if !SANITIZER_FUCHSIA
+
 #if SANITIZER_POSIX
 #include "sanitizer_common/sanitizer_posix.h"
 #endif
@@ -36,108 +41,6 @@
 
 namespace __asan {
 
-// Return true if we can quickly decide that the region is unpoisoned.
-// We assume that a redzone is at least 16 bytes.
-static inline bool QuickCheckForUnpoisonedRegion(uptr beg, uptr size) {
-  if (size == 0) return true;
-  if (size <= 32)
-    return !AddressIsPoisoned(beg) &&
-           !AddressIsPoisoned(beg + size - 1) &&
-           !AddressIsPoisoned(beg + size / 2);
-  if (size <= 64)
-    return !AddressIsPoisoned(beg) &&
-           !AddressIsPoisoned(beg + size / 4) &&
-           !AddressIsPoisoned(beg + size - 1) &&
-           !AddressIsPoisoned(beg + 3 * size / 4) &&
-           !AddressIsPoisoned(beg + size / 2);
-  return false;
-}
-
-struct AsanInterceptorContext {
-  const char *interceptor_name;
-};
-
-// We implement ACCESS_MEMORY_RANGE, ASAN_READ_RANGE,
-// and ASAN_WRITE_RANGE as macro instead of function so
-// that no extra frames are created, and stack trace contains
-// relevant information only.
-// We check all shadow bytes.
-#define ACCESS_MEMORY_RANGE(ctx, offset, size, isWrite) do {            \
-    uptr __offset = (uptr)(offset);                                     \
-    uptr __size = (uptr)(size);                                         \
-    uptr __bad = 0;                                                     \
-    if (__offset > __offset + __size) {                                 \
-      GET_STACK_TRACE_FATAL_HERE;                                       \
-      ReportStringFunctionSizeOverflow(__offset, __size, &stack);       \
-    }                                                                   \
-    if (!QuickCheckForUnpoisonedRegion(__offset, __size) &&             \
-        (__bad = __asan_region_is_poisoned(__offset, __size))) {        \
-      AsanInterceptorContext *_ctx = (AsanInterceptorContext *)ctx;     \
-      bool suppressed = false;                                          \
-      if (_ctx) {                                                       \
-        suppressed = IsInterceptorSuppressed(_ctx->interceptor_name);   \
-        if (!suppressed && HaveStackTraceBasedSuppressions()) {         \
-          GET_STACK_TRACE_FATAL_HERE;                                   \
-          suppressed = IsStackTraceSuppressed(&stack);                  \
-        }                                                               \
-      }                                                                 \
-      if (!suppressed) {                                                \
-        GET_CURRENT_PC_BP_SP;                                           \
-        ReportGenericError(pc, bp, sp, __bad, isWrite, __size, 0, false);\
-      }                                                                 \
-    }                                                                   \
-  } while (0)
-
-// memcpy is called during __asan_init() from the internals of printf(...).
-// We do not treat memcpy with to==from as a bug.
-// See http://llvm.org/bugs/show_bug.cgi?id=11763.
-#define ASAN_MEMCPY_IMPL(ctx, to, from, size)                           \
-  do {                                                                  \
-    if (UNLIKELY(!asan_inited)) return internal_memcpy(to, from, size); \
-    if (asan_init_is_running) {                                         \
-      return REAL(memcpy)(to, from, size);                              \
-    }                                                                   \
-    ENSURE_ASAN_INITED();                                               \
-    if (flags()->replace_intrin) {                                      \
-      if (to != from) {                                                 \
-        CHECK_RANGES_OVERLAP("memcpy", to, size, from, size);           \
-      }                                                                 \
-      ASAN_READ_RANGE(ctx, from, size);                                 \
-      ASAN_WRITE_RANGE(ctx, to, size);                                  \
-    }                                                                   \
-    return REAL(memcpy)(to, from, size);                                \
-  } while (0)
-
-// memset is called inside Printf.
-#define ASAN_MEMSET_IMPL(ctx, block, c, size)                           \
-  do {                                                                  \
-    if (UNLIKELY(!asan_inited)) return internal_memset(block, c, size); \
-    if (asan_init_is_running) {                                         \
-      return REAL(memset)(block, c, size);                              \
-    }                                                                   \
-    ENSURE_ASAN_INITED();                                               \
-    if (flags()->replace_intrin) {                                      \
-      ASAN_WRITE_RANGE(ctx, block, size);                               \
-    }                                                                   \
-    return REAL(memset)(block, c, size);                                \
-  } while (0)
-
-#define ASAN_MEMMOVE_IMPL(ctx, to, from, size)                           \
-  do {                                                                   \
-    if (UNLIKELY(!asan_inited)) return internal_memmove(to, from, size); \
-    ENSURE_ASAN_INITED();                                                \
-    if (flags()->replace_intrin) {                                       \
-      ASAN_READ_RANGE(ctx, from, size);                                  \
-      ASAN_WRITE_RANGE(ctx, to, size);                                   \
-    }                                                                    \
-    return internal_memmove(to, from, size);                             \
-  } while (0)
-
-#define ASAN_READ_RANGE(ctx, offset, size) \
-  ACCESS_MEMORY_RANGE(ctx, offset, size, false)
-#define ASAN_WRITE_RANGE(ctx, offset, size) \
-  ACCESS_MEMORY_RANGE(ctx, offset, size, true)
-
 #define ASAN_READ_STRING_OF_LEN(ctx, s, len, n)                 \
   ASAN_READ_RANGE((ctx), (s),                                   \
     common_flags()->strict_string_checks ? (len) + 1 : (n))
@@ -145,23 +48,6 @@ struct AsanInterceptorContext {
 #define ASAN_READ_STRING(ctx, s, n)                             \
   ASAN_READ_STRING_OF_LEN((ctx), (s), REAL(strlen)(s), (n))
 
-// Behavior of functions like "memcpy" or "strcpy" is undefined
-// if memory intervals overlap. We report error in this case.
-// Macro is used to avoid creation of new frames.
-static inline bool RangesOverlap(const char *offset1, uptr length1,
-                                 const char *offset2, uptr length2) {
-  return !((offset1 + length1 <= offset2) || (offset2 + length2 <= offset1));
-}
-#define CHECK_RANGES_OVERLAP(name, _offset1, length1, _offset2, length2) do { \
-  const char *offset1 = (const char*)_offset1; \
-  const char *offset2 = (const char*)_offset2; \
-  if (RangesOverlap(offset1, length1, offset2, length2)) { \
-    GET_STACK_TRACE_FATAL_HERE; \
-    ReportStringFunctionMemoryRangesOverlap(name, offset1, length1, \
-                                            offset2, length2, &stack); \
-  } \
-} while (0)
-
 static inline uptr MaybeRealStrnlen(const char *s, uptr maxlen) {
 #if SANITIZER_INTERCEPT_STRNLEN
   if (REAL(strnlen)) {
@@ -275,6 +161,7 @@ DECLARE_REAL_AND_INTERCEPTOR(void, free, void *)
   } while (false)
 
 #include "sanitizer_common/sanitizer_common_interceptors.inc"
+#include "sanitizer_common/sanitizer_signal_interceptors.inc"
 
 // Syscall interceptors don't have contexts, we don't support suppressions
 // for them.
@@ -356,42 +243,6 @@ INTERCEPTOR(int, pthread_join, void *t, void **arg) {
 DEFINE_REAL_PTHREAD_FUNCTIONS
 #endif  // ASAN_INTERCEPT_PTHREAD_CREATE
 
-#if ASAN_INTERCEPT_SIGNAL_AND_SIGACTION
-
-#if SANITIZER_ANDROID
-INTERCEPTOR(void*, bsd_signal, int signum, void *handler) {
-  if (GetHandleSignalMode(signum) != kHandleSignalExclusive)
-    return REAL(bsd_signal)(signum, handler);
-  return 0;
-}
-#endif
-
-INTERCEPTOR(void*, signal, int signum, void *handler) {
-  if (GetHandleSignalMode(signum) != kHandleSignalExclusive)
-    return REAL(signal)(signum, handler);
-  return nullptr;
-}
-
-INTERCEPTOR(int, sigaction, int signum, const struct sigaction *act,
-                            struct sigaction *oldact) {
-  if (GetHandleSignalMode(signum) != kHandleSignalExclusive)
-    return REAL(sigaction)(signum, act, oldact);
-  return 0;
-}
-
-namespace __sanitizer {
-int real_sigaction(int signum, const void *act, void *oldact) {
-  return REAL(sigaction)(signum, (const struct sigaction *)act,
-                         (struct sigaction *)oldact);
-}
-} // namespace __sanitizer
-
-#elif SANITIZER_POSIX
-// We need to have defined REAL(sigaction) on posix systems.
-DEFINE_REAL(int, sigaction, int signum, const struct sigaction *act,
-    struct sigaction *oldact)
-#endif  // ASAN_INTERCEPT_SIGNAL_AND_SIGACTION
-
 #if ASAN_INTERCEPT_SWAPCONTEXT
 static void ClearShadowMemoryForContextStack(uptr stack, uptr ssize) {
   // Align to page size.
@@ -428,6 +279,11 @@ INTERCEPTOR(int, swapcontext, struct ucontext_t *oucp,
 }
 #endif  // ASAN_INTERCEPT_SWAPCONTEXT
 
+#if SANITIZER_NETBSD
+#define longjmp __longjmp14
+#define siglongjmp __siglongjmp14
+#endif
+
 INTERCEPTOR(void, longjmp, void *env, int val) {
   __asan_handle_no_return();
   REAL(longjmp)(env, val);
@@ -462,18 +318,6 @@ INTERCEPTOR(void, __cxa_throw, void *a, void *b, void *c) {
 }
 #endif
 
-void *__asan_memcpy(void *to, const void *from, uptr size) {
-  ASAN_MEMCPY_IMPL(nullptr, to, from, size);
-}
-
-void *__asan_memset(void *block, int c, uptr size) {
-  ASAN_MEMSET_IMPL(nullptr, block, c, size);
-}
-
-void *__asan_memmove(void *to, const void *from, uptr size) {
-  ASAN_MEMMOVE_IMPL(nullptr, to, from, size);
-}
-
 #if ASAN_INTERCEPT_INDEX
 # if ASAN_USE_ALIAS_ATTRIBUTE_FOR_INDEX
 INTERCEPTOR(char*, index, const char *string, int c)
@@ -711,6 +555,7 @@ void InitializeAsanInterceptors() {
   CHECK(!was_called_once);
   was_called_once = true;
   InitializeCommonInterceptors();
+  InitializeSignalInterceptors();
 
   // Intercept str* functions.
   ASAN_INTERCEPT_FUNC(strcat);  // NOLINT
@@ -733,15 +578,9 @@ void InitializeAsanInterceptors() {
   ASAN_INTERCEPT_FUNC(strtoll);
 #endif
 
-  // Intecept signal- and jump-related functions.
+  // Intecept jump-related functions.
   ASAN_INTERCEPT_FUNC(longjmp);
-#if ASAN_INTERCEPT_SIGNAL_AND_SIGACTION
-  ASAN_INTERCEPT_FUNC(sigaction);
-#if SANITIZER_ANDROID
-  ASAN_INTERCEPT_FUNC(bsd_signal);
-#endif
-  ASAN_INTERCEPT_FUNC(signal);
-#endif
+
 #if ASAN_INTERCEPT_SWAPCONTEXT
   ASAN_INTERCEPT_FUNC(swapcontext);
 #endif
@@ -785,3 +624,5 @@ void InitializeAsanInterceptors() {
 }
 
 } // namespace __asan
+
+#endif  // !SANITIZER_FUCHSIA
diff --git a/contrib/compiler-rt/lib/asan/asan_interceptors.h b/contrib/compiler-rt/lib/asan/asan_interceptors.h
index 93fca4f67366..e13bdecfdede 100644
--- a/contrib/compiler-rt/lib/asan/asan_interceptors.h
+++ b/contrib/compiler-rt/lib/asan/asan_interceptors.h
@@ -15,9 +15,30 @@
 #define ASAN_INTERCEPTORS_H
 
 #include "asan_internal.h"
+#include "asan_interceptors_memintrinsics.h"
 #include "interception/interception.h"
 #include "sanitizer_common/sanitizer_platform_interceptors.h"
 
+namespace __asan {
+
+void InitializeAsanInterceptors();
+void InitializePlatformInterceptors();
+
+#define ENSURE_ASAN_INITED()      \
+  do {                            \
+    CHECK(!asan_init_is_running); \
+    if (UNLIKELY(!asan_inited)) { \
+      AsanInitFromRtl();          \
+    }                             \
+  } while (0)
+
+}  // namespace __asan
+
+// There is no general interception at all on Fuchsia.
+// Only the functions in asan_interceptors_memintrinsics.h are
+// really defined to replace libc functions.
+#if !SANITIZER_FUCHSIA
+
 // Use macro to describe if specific function should be
 // intercepted on a given platform.
 #if !SANITIZER_WINDOWS
@@ -34,25 +55,20 @@
 # define ASAN_INTERCEPT_FORK 0
 #endif
 
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
 # define ASAN_USE_ALIAS_ATTRIBUTE_FOR_INDEX 1
 #else
 # define ASAN_USE_ALIAS_ATTRIBUTE_FOR_INDEX 0
 #endif
 
-#if SANITIZER_LINUX && !SANITIZER_ANDROID
+#if (SANITIZER_LINUX && !SANITIZER_ANDROID) || SANITIZER_SOLARIS
 # define ASAN_INTERCEPT_SWAPCONTEXT 1
 #else
 # define ASAN_INTERCEPT_SWAPCONTEXT 0
 #endif
 
 #if !SANITIZER_WINDOWS
-# define ASAN_INTERCEPT_SIGNAL_AND_SIGACTION 1
-#else
-# define ASAN_INTERCEPT_SIGNAL_AND_SIGACTION 0
-#endif
-
-#if !SANITIZER_WINDOWS
 # define ASAN_INTERCEPT_SIGLONGJMP 1
 #else
 # define ASAN_INTERCEPT_SIGLONGJMP 0
@@ -66,7 +82,8 @@
 
 // Android bug: https://code.google.com/p/android/issues/detail?id=61799
 #if ASAN_HAS_EXCEPTIONS && !SANITIZER_WINDOWS && \
-    !(SANITIZER_ANDROID && defined(__i386))
+    !(SANITIZER_ANDROID && defined(__i386)) && \
+    !SANITIZER_SOLARIS
 # define ASAN_INTERCEPT___CXA_THROW 1
 #else
 # define ASAN_INTERCEPT___CXA_THROW 0
@@ -85,16 +102,11 @@
 #endif
 
 DECLARE_REAL(int, memcmp, const void *a1, const void *a2, uptr size)
-DECLARE_REAL(void*, memcpy, void *to, const void *from, uptr size)
-DECLARE_REAL(void*, memset, void *block, int c, uptr size)
 DECLARE_REAL(char*, strchr, const char *str, int c)
 DECLARE_REAL(SIZE_T, strlen, const char *s)
 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)
-struct sigaction;
-DECLARE_REAL(int, sigaction, int signum, const struct sigaction *act,
-                             struct sigaction *oldact)
 
 #if !SANITIZER_MAC
 #define ASAN_INTERCEPT_FUNC(name)                                        \
@@ -113,18 +125,6 @@ DECLARE_REAL(int, sigaction, int signum, const struct sigaction *act,
 #define ASAN_INTERCEPT_FUNC(name)
 #endif  // SANITIZER_MAC
 
-namespace __asan {
-
-void InitializeAsanInterceptors();
-void InitializePlatformInterceptors();
-
-#define ENSURE_ASAN_INITED() do { \
-  CHECK(!asan_init_is_running); \
-  if (UNLIKELY(!asan_inited)) { \
-    AsanInitFromRtl(); \
-  } \
-} while (0)
-
-}  // namespace __asan
+#endif  // !SANITIZER_FUCHSIA
 
 #endif  // ASAN_INTERCEPTORS_H
diff --git a/contrib/compiler-rt/lib/asan/asan_interceptors_memintrinsics.cc b/contrib/compiler-rt/lib/asan/asan_interceptors_memintrinsics.cc
new file mode 100644
index 000000000000..c89cb011492e
--- /dev/null
+++ b/contrib/compiler-rt/lib/asan/asan_interceptors_memintrinsics.cc
@@ -0,0 +1,44 @@
+//===-- asan_interceptors_memintrinsics.cc --------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// This file is a part of AddressSanitizer, an address sanity checker.
+//
+// ASan versions of memcpy, memmove, and memset.
+//===---------------------------------------------------------------------===//
+
+#include "asan_interceptors_memintrinsics.h"
+#include "asan_report.h"
+#include "asan_stack.h"
+#include "asan_suppressions.h"
+
+using namespace __asan;  // NOLINT
+
+void *__asan_memcpy(void *to, const void *from, uptr size) {
+  ASAN_MEMCPY_IMPL(nullptr, to, from, size);
+}
+
+void *__asan_memset(void *block, int c, uptr size) {
+  ASAN_MEMSET_IMPL(nullptr, block, c, size);
+}
+
+void *__asan_memmove(void *to, const void *from, uptr size) {
+  ASAN_MEMMOVE_IMPL(nullptr, to, from, size);
+}
+
+#if SANITIZER_FUCHSIA
+
+// Fuchsia doesn't use sanitizer_common_interceptors.inc, but the only
+// things there it wants are these three.  Just define them as aliases
+// here rather than repeating the contents.
+
+decltype(memcpy) memcpy[[gnu::alias("__asan_memcpy")]];
+decltype(memmove) memmove[[gnu::alias("__asan_memmove")]];
+decltype(memset) memset[[gnu::alias("__asan_memset")]];
+
+#endif  // SANITIZER_FUCHSIA
diff --git a/contrib/compiler-rt/lib/asan/asan_interceptors_memintrinsics.h b/contrib/compiler-rt/lib/asan/asan_interceptors_memintrinsics.h
new file mode 100644
index 000000000000..5a8339a23e97
--- /dev/null
+++ b/contrib/compiler-rt/lib/asan/asan_interceptors_memintrinsics.h
@@ -0,0 +1,148 @@
+//===-- asan_interceptors_memintrinsics.h -----------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// This file is a part of AddressSanitizer, an address sanity checker.
+//
+// ASan-private header for asan_memintrin.cc
+//===---------------------------------------------------------------------===//
+#ifndef ASAN_MEMINTRIN_H
+#define ASAN_MEMINTRIN_H
+
+#include "asan_interface_internal.h"
+#include "asan_internal.h"
+#include "asan_mapping.h"
+#include "interception/interception.h"
+
+DECLARE_REAL(void*, memcpy, void *to, const void *from, uptr size)
+DECLARE_REAL(void*, memset, void *block, int c, uptr size)
+
+namespace __asan {
+
+// Return true if we can quickly decide that the region is unpoisoned.
+// We assume that a redzone is at least 16 bytes.
+static inline bool QuickCheckForUnpoisonedRegion(uptr beg, uptr size) {
+  if (size == 0) return true;
+  if (size <= 32)
+    return !AddressIsPoisoned(beg) &&
+           !AddressIsPoisoned(beg + size - 1) &&
+           !AddressIsPoisoned(beg + size / 2);
+  if (size <= 64)
+    return !AddressIsPoisoned(beg) &&
+           !AddressIsPoisoned(beg + size / 4) &&
+           !AddressIsPoisoned(beg + size - 1) &&
+           !AddressIsPoisoned(beg + 3 * size / 4) &&
+           !AddressIsPoisoned(beg + size / 2);
+  return false;
+}
+
+struct AsanInterceptorContext {
+  const char *interceptor_name;
+};
+
+// We implement ACCESS_MEMORY_RANGE, ASAN_READ_RANGE,
+// and ASAN_WRITE_RANGE as macro instead of function so
+// that no extra frames are created, and stack trace contains
+// relevant information only.
+// We check all shadow bytes.
+#define ACCESS_MEMORY_RANGE(ctx, offset, size, isWrite) do {            \
+    uptr __offset = (uptr)(offset);                                     \
+    uptr __size = (uptr)(size);                                         \
+    uptr __bad = 0;                                                     \
+    if (__offset > __offset + __size) {                                 \
+      GET_STACK_TRACE_FATAL_HERE;                                       \
+      ReportStringFunctionSizeOverflow(__offset, __size, &stack);       \
+    }                                                                   \
+    if (!QuickCheckForUnpoisonedRegion(__offset, __size) &&             \
+        (__bad = __asan_region_is_poisoned(__offset, __size))) {        \
+      AsanInterceptorContext *_ctx = (AsanInterceptorContext *)ctx;     \
+      bool suppressed = false;                                          \
+      if (_ctx) {                                                       \
+        suppressed = IsInterceptorSuppressed(_ctx->interceptor_name);   \
+        if (!suppressed && HaveStackTraceBasedSuppressions()) {         \
+          GET_STACK_TRACE_FATAL_HERE;                                   \
+          suppressed = IsStackTraceSuppressed(&stack);                  \
+        }                                                               \
+      }                                                                 \
+      if (!suppressed) {                                                \
+        GET_CURRENT_PC_BP_SP;                                           \
+        ReportGenericError(pc, bp, sp, __bad, isWrite, __size, 0, false);\
+      }                                                                 \
+    }                                                                   \
+  } while (0)
+
+// memcpy is called during __asan_init() from the internals of printf(...).
+// We do not treat memcpy with to==from as a bug.
+// See http://llvm.org/bugs/show_bug.cgi?id=11763.
+#define ASAN_MEMCPY_IMPL(ctx, to, from, size)                           \
+  do {                                                                  \
+    if (UNLIKELY(!asan_inited)) return internal_memcpy(to, from, size); \
+    if (asan_init_is_running) {                                         \
+      return REAL(memcpy)(to, from, size);                              \
+    }                                                                   \
+    ENSURE_ASAN_INITED();                                               \
+    if (flags()->replace_intrin) {                                      \
+      if (to != from) {                                                 \
+        CHECK_RANGES_OVERLAP("memcpy", to, size, from, size);           \
+      }                                                                 \
+      ASAN_READ_RANGE(ctx, from, size);                                 \
+      ASAN_WRITE_RANGE(ctx, to, size);                                  \
+    }                                                                   \
+    return REAL(memcpy)(to, from, size);                                \
+  } while (0)
+
+// memset is called inside Printf.
+#define ASAN_MEMSET_IMPL(ctx, block, c, size)                           \
+  do {                                                                  \
+    if (UNLIKELY(!asan_inited)) return internal_memset(block, c, size); \
+    if (asan_init_is_running) {                                         \
+      return REAL(memset)(block, c, size);                              \
+    }                                                                   \
+    ENSURE_ASAN_INITED();                                               \
+    if (flags()->replace_intrin) {                                      \
+      ASAN_WRITE_RANGE(ctx, block, size);                               \
+    }                                                                   \
+    return REAL(memset)(block, c, size);                                \
+  } while (0)
+
+#define ASAN_MEMMOVE_IMPL(ctx, to, from, size)                           \
+  do {                                                                   \
+    if (UNLIKELY(!asan_inited)) return internal_memmove(to, from, size); \
+    ENSURE_ASAN_INITED();                                                \
+    if (flags()->replace_intrin) {                                       \
+      ASAN_READ_RANGE(ctx, from, size);                                  \
+      ASAN_WRITE_RANGE(ctx, to, size);                                   \
+    }                                                                    \
+    return internal_memmove(to, from, size);                             \
+  } while (0)
+
+#define ASAN_READ_RANGE(ctx, offset, size) \
+  ACCESS_MEMORY_RANGE(ctx, offset, size, false)
+#define ASAN_WRITE_RANGE(ctx, offset, size) \
+  ACCESS_MEMORY_RANGE(ctx, offset, size, true)
+
+// Behavior of functions like "memcpy" or "strcpy" is undefined
+// if memory intervals overlap. We report error in this case.
+// Macro is used to avoid creation of new frames.
+static inline bool RangesOverlap(const char *offset1, uptr length1,
+                                 const char *offset2, uptr length2) {
+  return !((offset1 + length1 <= offset2) || (offset2 + length2 <= offset1));
+}
+#define CHECK_RANGES_OVERLAP(name, _offset1, length1, _offset2, length2) do { \
+  const char *offset1 = (const char*)_offset1; \
+  const char *offset2 = (const char*)_offset2; \
+  if (RangesOverlap(offset1, length1, offset2, length2)) { \
+    GET_STACK_TRACE_FATAL_HERE; \
+    ReportStringFunctionMemoryRangesOverlap(name, offset1, length1, \
+                                            offset2, length2, &stack); \
+  } \
+} while (0)
+
+}  // namespace __asan
+
+#endif  // ASAN_MEMINTRIN_H
diff --git a/contrib/compiler-rt/lib/asan/asan_internal.h b/contrib/compiler-rt/lib/asan/asan_internal.h
index f09bbd83af25..19133e5291a9 100644
--- a/contrib/compiler-rt/lib/asan/asan_internal.h
+++ b/contrib/compiler-rt/lib/asan/asan_internal.h
@@ -69,8 +69,12 @@ void InitializePlatformExceptionHandlers();
 bool IsSystemHeapAddress(uptr addr);
 
 // asan_rtl.cc
+void PrintAddressSpaceLayout();
 void NORETURN ShowStatsAndAbort();
 
+// asan_shadow_setup.cc
+void InitializeShadowMemory();
+
 // asan_malloc_linux.cc / asan_malloc_mac.cc
 void ReplaceSystemMalloc();
 
@@ -80,6 +84,9 @@ void *AsanDoesNotSupportStaticLinkage();
 void AsanCheckDynamicRTPrereqs();
 void AsanCheckIncompatibleRT();
 
+// asan_thread.cc
+AsanThread *CreateMainThread();
+
 // Support function for __asan_(un)register_image_globals. Searches for the
 // loaded image containing `needle' and then enumerates all global metadata
 // structures declared in that image, applying `op' (e.g.,
diff --git a/contrib/compiler-rt/lib/asan/asan_linux.cc b/contrib/compiler-rt/lib/asan/asan_linux.cc
index 6d47ba432a61..047e1dbb72fa 100644
--- a/contrib/compiler-rt/lib/asan/asan_linux.cc
+++ b/contrib/compiler-rt/lib/asan/asan_linux.cc
@@ -13,10 +13,12 @@
 //===----------------------------------------------------------------------===//
 
 #include "sanitizer_common/sanitizer_platform.h"
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
 
 #include "asan_interceptors.h"
 #include "asan_internal.h"
+#include "asan_premap_shadow.h"
 #include "asan_thread.h"
 #include "sanitizer_common/sanitizer_flags.h"
 #include "sanitizer_common/sanitizer_freebsd.h"
@@ -39,9 +41,17 @@
 #include <sys/link_elf.h>
 #endif
 
-#if SANITIZER_ANDROID || SANITIZER_FREEBSD
+#if SANITIZER_SOLARIS
+#include <link.h>
+#endif
+
+#if SANITIZER_ANDROID || SANITIZER_FREEBSD || SANITIZER_SOLARIS
 #include <ucontext.h>
 extern "C" void* _DYNAMIC;
+#elif SANITIZER_NETBSD
+#include <link_elf.h>
+#include <ucontext.h>
+extern Elf_Dyn _DYNAMIC;
 #else
 #include <sys/ucontext.h>
 #include <link.h>
@@ -77,9 +87,51 @@ void *AsanDoesNotSupportStaticLinkage() {
   return &_DYNAMIC;  // defined in link.h
 }
 
+static void UnmapFromTo(uptr from, uptr to) {
+  CHECK(to >= from);
+  if (to == from) return;
+  uptr res = internal_munmap(reinterpret_cast<void *>(from), to - from);
+  if (UNLIKELY(internal_iserror(res))) {
+    Report(
+        "ERROR: AddresSanitizer failed to unmap 0x%zx (%zd) bytes at address "
+        "%p\n",
+        to - from, to - from, from);
+    CHECK("unable to unmap" && 0);
+  }
+}
+
+#if ASAN_PREMAP_SHADOW
+uptr FindPremappedShadowStart() {
+  uptr granularity = GetMmapGranularity();
+  uptr shadow_start = reinterpret_cast<uptr>(&__asan_shadow);
+  uptr premap_shadow_size = PremapShadowSize();
+  uptr shadow_size = RoundUpTo(kHighShadowEnd, granularity);
+  // We may have mapped too much. Release extra memory.
+  UnmapFromTo(shadow_start + shadow_size, shadow_start + premap_shadow_size);
+  return shadow_start;
+}
+#endif
+
 uptr FindDynamicShadowStart() {
-  UNREACHABLE("FindDynamicShadowStart is not available");
-  return 0;
+#if ASAN_PREMAP_SHADOW
+  if (!PremapShadowFailed())
+    return FindPremappedShadowStart();
+#endif
+
+  uptr granularity = GetMmapGranularity();
+  uptr alignment = granularity * 8;
+  uptr left_padding = granularity;
+  uptr shadow_size = RoundUpTo(kHighShadowEnd, granularity);
+  uptr map_size = shadow_size + left_padding + alignment;
+
+  uptr map_start = (uptr)MmapNoAccess(map_size);
+  CHECK_NE(map_start, ~(uptr)0);
+
+  uptr shadow_start = RoundUpTo(map_start + left_padding, alignment);
+  UnmapFromTo(map_start, shadow_start - left_padding);
+  UnmapFromTo(shadow_start + shadow_size, map_start + map_size);
+
+  return shadow_start;
 }
 
 void AsanApplyToGlobals(globals_op_fptr op, const void *needle) {
@@ -93,6 +145,9 @@ 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);
+
   // Continue until the first dynamic library is found
   if (!info->dlpi_name || info->dlpi_name[0] == 0)
     return 0;
@@ -101,6 +156,21 @@ static int FindFirstDSOCallback(struct dl_phdr_info *info, size_t size,
   if (internal_strncmp(info->dlpi_name, "linux-", sizeof("linux-") - 1) == 0)
     return 0;
 
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
+  // Ignore first entry (the main program)
+  char **p = (char **)data;
+  if (!(*p)) {
+    *p = (char *)-1;
+    return 0;
+  }
+#endif
+
+#if SANITIZER_SOLARIS
+  // Ignore executable on Solaris
+  if (info->dlpi_addr == 0)
+    return 0;
+#endif
+
   *(const char **)data = info->dlpi_name;
   return 1;
 }
@@ -179,4 +249,5 @@ void *AsanDlSymNext(const char *sym) {
 
 } // namespace __asan
 
-#endif // SANITIZER_FREEBSD || SANITIZER_LINUX
+#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD ||
+        // SANITIZER_SOLARIS
diff --git a/contrib/compiler-rt/lib/asan/asan_malloc_linux.cc b/contrib/compiler-rt/lib/asan/asan_malloc_linux.cc
index fd40f47db1c4..6697ff8764ed 100644
--- a/contrib/compiler-rt/lib/asan/asan_malloc_linux.cc
+++ b/contrib/compiler-rt/lib/asan/asan_malloc_linux.cc
@@ -15,7 +15,8 @@
 //===----------------------------------------------------------------------===//
 
 #include "sanitizer_common/sanitizer_platform.h"
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
+#if SANITIZER_FREEBSD || SANITIZER_FUCHSIA || SANITIZER_LINUX || \
+    SANITIZER_NETBSD || SANITIZER_SOLARIS
 
 #include "sanitizer_common/sanitizer_tls_get_addr.h"
 #include "asan_allocator.h"
@@ -30,9 +31,9 @@ static uptr allocated_for_dlsym;
 static const uptr kDlsymAllocPoolSize = 1024;
 static uptr alloc_memory_for_dlsym[kDlsymAllocPoolSize];
 
-static bool IsInDlsymAllocPool(const void *ptr) {
+static INLINE bool IsInDlsymAllocPool(const void *ptr) {
   uptr off = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-  return off < sizeof(alloc_memory_for_dlsym);
+  return off < allocated_for_dlsym * sizeof(alloc_memory_for_dlsym[0]);
 }
 
 static void *AllocateFromLocalPool(uptr size_in_bytes) {
@@ -43,6 +44,26 @@ static void *AllocateFromLocalPool(uptr size_in_bytes) {
   return mem;
 }
 
+static INLINE bool MaybeInDlsym() {
+  // Fuchsia doesn't use dlsym-based interceptors.
+  return !SANITIZER_FUCHSIA && asan_init_is_running;
+}
+
+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(MaybeInDlsym())) {
+    new_ptr = AllocateFromLocalPool(size);
+  } else {
+    ENSURE_ASAN_INITED();
+    GET_STACK_TRACE_MALLOC;
+    new_ptr = asan_malloc(size, &stack);
+  }
+  internal_memcpy(new_ptr, ptr, copy_size);
+  return new_ptr;
+}
+
 INTERCEPTOR(void, free, void *ptr) {
   GET_STACK_TRACE_FREE;
   if (UNLIKELY(IsInDlsymAllocPool(ptr)))
@@ -60,7 +81,7 @@ INTERCEPTOR(void, cfree, void *ptr) {
 #endif // SANITIZER_INTERCEPT_CFREE
 
 INTERCEPTOR(void*, malloc, uptr size) {
-  if (UNLIKELY(asan_init_is_running))
+  if (UNLIKELY(MaybeInDlsym()))
     // Hack: dlsym calls malloc before REAL(malloc) is retrieved from dlsym.
     return AllocateFromLocalPool(size);
   ENSURE_ASAN_INITED();
@@ -69,7 +90,7 @@ INTERCEPTOR(void*, malloc, uptr size) {
 }
 
 INTERCEPTOR(void*, calloc, uptr nmemb, uptr size) {
-  if (UNLIKELY(asan_init_is_running))
+  if (UNLIKELY(MaybeInDlsym()))
     // Hack: dlsym calls calloc before REAL(calloc) is retrieved from dlsym.
     return AllocateFromLocalPool(nmemb * size);
   ENSURE_ASAN_INITED();
@@ -78,21 +99,9 @@ INTERCEPTOR(void*, calloc, uptr nmemb, uptr size) {
 }
 
 INTERCEPTOR(void*, realloc, void *ptr, uptr size) {
-  if (UNLIKELY(IsInDlsymAllocPool(ptr))) {
-    const uptr offset = (uptr)ptr - (uptr)alloc_memory_for_dlsym;
-    const uptr copy_size = Min(size, kDlsymAllocPoolSize - offset);
-    void *new_ptr;
-    if (UNLIKELY(asan_init_is_running)) {
-      new_ptr = AllocateFromLocalPool(size);
-    } else {
-      ENSURE_ASAN_INITED();
-      GET_STACK_TRACE_MALLOC;
-      new_ptr = asan_malloc(size, &stack);
-    }
-    internal_memcpy(new_ptr, ptr, copy_size);
-    return new_ptr;
-  }
-  if (UNLIKELY(asan_init_is_running))
+  if (UNLIKELY(IsInDlsymAllocPool(ptr)))
+    return ReallocFromLocalPool(ptr, size);
+  if (UNLIKELY(MaybeInDlsym()))
     return AllocateFromLocalPool(size);
   ENSURE_ASAN_INITED();
   GET_STACK_TRACE_MALLOC;
@@ -226,4 +235,5 @@ void ReplaceSystemMalloc() {
 }  // namespace __asan
 #endif  // SANITIZER_ANDROID
 
-#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX
+#endif  // SANITIZER_FREEBSD || SANITIZER_FUCHSIA || SANITIZER_LINUX ||
+        // SANITIZER_NETBSD || SANITIZER_SOLARIS
diff --git a/contrib/compiler-rt/lib/asan/asan_mapping.h b/contrib/compiler-rt/lib/asan/asan_mapping.h
index 695740cd982f..d3f360ffc969 100644
--- a/contrib/compiler-rt/lib/asan/asan_mapping.h
+++ b/contrib/compiler-rt/lib/asan/asan_mapping.h
@@ -115,6 +115,13 @@
 // || `[0x40000000, 0x47ffffff]` || LowShadow  ||
 // || `[0x00000000, 0x3fffffff]` || LowMem     ||
 //
+// Shadow mapping on NetBSD/x86-64 with SHADOW_OFFSET == 0x400000000000:
+// || `[0x4feffffffe01, 0x7f7ffffff000]` || HighMem    ||
+// || `[0x49fdffffffc0, 0x4feffffffe00]` || HighShadow ||
+// || `[0x480000000000, 0x49fdffffffbf]` || ShadowGap  ||
+// || `[0x400000000000, 0x47ffffffffff]` || LowShadow  ||
+// || `[0x000000000000, 0x3fffffffffff]` || LowMem     ||
+//
 // Default Windows/i386 mapping:
 // (the exact location of HighShadow/HighMem may vary depending
 //  on WoW64, /LARGEADDRESSAWARE, etc).
@@ -124,11 +131,16 @@
 // || `[0x30000000, 0x35ffffff]` || LowShadow  ||
 // || `[0x00000000, 0x2fffffff]` || LowMem     ||
 
+#if defined(ASAN_SHADOW_SCALE)
+static const u64 kDefaultShadowScale = ASAN_SHADOW_SCALE;
+#else
 static const u64 kDefaultShadowScale = 3;
+#endif
 static const u64 kDefaultShadowSentinel = ~(uptr)0;
 static const u64 kDefaultShadowOffset32 = 1ULL << 29;  // 0x20000000
 static const u64 kDefaultShadowOffset64 = 1ULL << 44;
-static const u64 kDefaultShort64bitShadowOffset = 0x7FFF8000;  // < 2G.
+static const u64 kDefaultShort64bitShadowOffset =
+    0x7FFFFFFF & (~0xFFFULL << kDefaultShadowScale);  // < 2G.
 static const u64 kIosShadowOffset32 = 1ULL << 30;  // 0x40000000
 static const u64 kIosShadowOffset64 = 0x120200000;
 static const u64 kIosSimShadowOffset32 = 1ULL << 30;
@@ -136,18 +148,20 @@ static const u64 kIosSimShadowOffset64 = kDefaultShadowOffset64;
 static const u64 kAArch64_ShadowOffset64 = 1ULL << 36;
 static const u64 kMIPS32_ShadowOffset32 = 0x0aaa0000;
 static const u64 kMIPS64_ShadowOffset64 = 1ULL << 37;
-static const u64 kPPC64_ShadowOffset64 = 1ULL << 41;
+static const u64 kPPC64_ShadowOffset64 = 1ULL << 44;
 static const u64 kSystemZ_ShadowOffset64 = 1ULL << 52;
 static const u64 kFreeBSD_ShadowOffset32 = 1ULL << 30;  // 0x40000000
 static const u64 kFreeBSD_ShadowOffset64 = 1ULL << 46;  // 0x400000000000
+static const u64 kNetBSD_ShadowOffset64 = 1ULL << 46;  // 0x400000000000
 static const u64 kWindowsShadowOffset32 = 3ULL << 28;  // 0x30000000
 
 #define SHADOW_SCALE kDefaultShadowScale
 
-
-#if SANITIZER_WORDSIZE == 32
+#if SANITIZER_FUCHSIA
+#  define SHADOW_OFFSET (0)
+#elif SANITIZER_WORDSIZE == 32
 #  if SANITIZER_ANDROID
-#    define SHADOW_OFFSET (0)
+#    define SHADOW_OFFSET __asan_shadow_memory_dynamic_address
 #  elif defined(__mips__)
 #    define SHADOW_OFFSET kMIPS32_ShadowOffset32
 #  elif SANITIZER_FREEBSD
@@ -178,6 +192,8 @@ static const u64 kWindowsShadowOffset32 = 3ULL << 28;  // 0x30000000
 #    define SHADOW_OFFSET kSystemZ_ShadowOffset64
 #  elif SANITIZER_FREEBSD
 #    define SHADOW_OFFSET kFreeBSD_ShadowOffset64
+#  elif SANITIZER_NETBSD
+#    define SHADOW_OFFSET kNetBSD_ShadowOffset64
 #  elif SANITIZER_MAC
 #   define SHADOW_OFFSET kDefaultShadowOffset64
 #  elif defined(__mips64)
@@ -189,6 +205,12 @@ static const u64 kWindowsShadowOffset32 = 3ULL << 28;  // 0x30000000
 #  endif
 #endif
 
+#if SANITIZER_ANDROID && defined(__arm__)
+# define ASAN_PREMAP_SHADOW 1
+#else
+# define ASAN_PREMAP_SHADOW 0
+#endif
+
 #define SHADOW_GRANULARITY (1ULL << SHADOW_SCALE)
 #define MEM_TO_SHADOW(mem) (((mem) >> SHADOW_SCALE) + (SHADOW_OFFSET))
 
diff --git a/contrib/compiler-rt/lib/asan/asan_memory_profile.cc b/contrib/compiler-rt/lib/asan/asan_memory_profile.cc
index 05846c37cb6d..603284c8c688 100644
--- a/contrib/compiler-rt/lib/asan/asan_memory_profile.cc
+++ b/contrib/compiler-rt/lib/asan/asan_memory_profile.cc
@@ -107,6 +107,9 @@ static void MemoryProfileCB(const SuspendedThreadsList &suspended_threads_list,
   __lsan::ForEachChunk(ChunkCallback, &hp);
   uptr *Arg = reinterpret_cast<uptr*>(argument);
   hp.Print(Arg[0], Arg[1]);
+
+  if (Verbosity())
+    __asan_print_accumulated_stats();
 }
 
 }  // namespace __asan
diff --git a/contrib/compiler-rt/lib/asan/asan_new_delete.cc b/contrib/compiler-rt/lib/asan/asan_new_delete.cc
index e68c7f3e2400..072f027addd6 100644
--- a/contrib/compiler-rt/lib/asan/asan_new_delete.cc
+++ b/contrib/compiler-rt/lib/asan/asan_new_delete.cc
@@ -125,77 +125,69 @@ INTERCEPTOR(void *, _ZnwmRKSt9nothrow_t, size_t size, std::nothrow_t const&) {
 INTERCEPTOR(void *, _ZnamRKSt9nothrow_t, size_t size, std::nothrow_t const&) {
   OPERATOR_NEW_BODY(FROM_NEW_BR, true /*nothrow*/);
 }
-#endif
+#endif  // !SANITIZER_MAC
 
 #define OPERATOR_DELETE_BODY(type) \
   GET_STACK_TRACE_FREE;\
-  asan_free(ptr, &stack, type);
+  asan_delete(ptr, 0, 0, &stack, type);
+
+#define OPERATOR_DELETE_BODY_SIZE(type) \
+  GET_STACK_TRACE_FREE;\
+  asan_delete(ptr, size, 0, &stack, type);
+
+#define OPERATOR_DELETE_BODY_ALIGN(type) \
+  GET_STACK_TRACE_FREE;\
+  asan_delete(ptr, 0, static_cast<uptr>(align), &stack, type);
+
+#define OPERATOR_DELETE_BODY_SIZE_ALIGN(type) \
+  GET_STACK_TRACE_FREE;\
+  asan_delete(ptr, size, static_cast<uptr>(align), &stack, type);
 
 #if !SANITIZER_MAC
 CXX_OPERATOR_ATTRIBUTE
-void operator delete(void *ptr) NOEXCEPT {
-  OPERATOR_DELETE_BODY(FROM_NEW);
-}
+void operator delete(void *ptr) NOEXCEPT
+{ OPERATOR_DELETE_BODY(FROM_NEW); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete[](void *ptr) NOEXCEPT {
-  OPERATOR_DELETE_BODY(FROM_NEW_BR);
-}
+void operator delete[](void *ptr) NOEXCEPT
+{ OPERATOR_DELETE_BODY(FROM_NEW_BR); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete(void *ptr, std::nothrow_t const&) {
-  OPERATOR_DELETE_BODY(FROM_NEW);
-}
+void operator delete(void *ptr, std::nothrow_t const&)
+{ OPERATOR_DELETE_BODY(FROM_NEW); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete[](void *ptr, std::nothrow_t const&) {
-  OPERATOR_DELETE_BODY(FROM_NEW_BR);
-}
+void operator delete[](void *ptr, std::nothrow_t const&)
+{ OPERATOR_DELETE_BODY(FROM_NEW_BR); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete(void *ptr, size_t size) NOEXCEPT {
-  GET_STACK_TRACE_FREE;
-  asan_sized_free(ptr, size, &stack, FROM_NEW);
-}
+void operator delete(void *ptr, size_t size) NOEXCEPT
+{ OPERATOR_DELETE_BODY_SIZE(FROM_NEW); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete[](void *ptr, size_t size) NOEXCEPT {
-  GET_STACK_TRACE_FREE;
-  asan_sized_free(ptr, size, &stack, FROM_NEW_BR);
-}
+void operator delete[](void *ptr, size_t size) NOEXCEPT
+{ OPERATOR_DELETE_BODY_SIZE(FROM_NEW_BR); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete(void *ptr, std::align_val_t) NOEXCEPT {
-  OPERATOR_DELETE_BODY(FROM_NEW);
-}
+void operator delete(void *ptr, std::align_val_t align) NOEXCEPT
+{ OPERATOR_DELETE_BODY_ALIGN(FROM_NEW); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete[](void *ptr, std::align_val_t) NOEXCEPT {
-  OPERATOR_DELETE_BODY(FROM_NEW_BR);
-}
+void operator delete[](void *ptr, std::align_val_t align) NOEXCEPT
+{ OPERATOR_DELETE_BODY_ALIGN(FROM_NEW_BR); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete(void *ptr, std::align_val_t, std::nothrow_t const&) {
-  OPERATOR_DELETE_BODY(FROM_NEW);
-}
+void operator delete(void *ptr, std::align_val_t align, std::nothrow_t const&)
+{ OPERATOR_DELETE_BODY_ALIGN(FROM_NEW); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete[](void *ptr, std::align_val_t, std::nothrow_t const&) {
-  OPERATOR_DELETE_BODY(FROM_NEW_BR);
-}
+void operator delete[](void *ptr, std::align_val_t align, std::nothrow_t const&)
+{ OPERATOR_DELETE_BODY_ALIGN(FROM_NEW_BR); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete(void *ptr, size_t size, std::align_val_t) NOEXCEPT {
-  GET_STACK_TRACE_FREE;
-  asan_sized_free(ptr, size, &stack, FROM_NEW);
-}
+void operator delete(void *ptr, size_t size, std::align_val_t align) NOEXCEPT
+{ OPERATOR_DELETE_BODY_SIZE_ALIGN(FROM_NEW); }
 CXX_OPERATOR_ATTRIBUTE
-void operator delete[](void *ptr, size_t size, std::align_val_t) NOEXCEPT {
-  GET_STACK_TRACE_FREE;
-  asan_sized_free(ptr, size, &stack, FROM_NEW_BR);
-}
+void operator delete[](void *ptr, size_t size, std::align_val_t align) NOEXCEPT
+{ OPERATOR_DELETE_BODY_SIZE_ALIGN(FROM_NEW_BR); }
 
 #else  // SANITIZER_MAC
-INTERCEPTOR(void, _ZdlPv, void *ptr) {
-  OPERATOR_DELETE_BODY(FROM_NEW);
-}
-INTERCEPTOR(void, _ZdaPv, void *ptr) {
-  OPERATOR_DELETE_BODY(FROM_NEW_BR);
-}
-INTERCEPTOR(void, _ZdlPvRKSt9nothrow_t, void *ptr, std::nothrow_t const&) {
-  OPERATOR_DELETE_BODY(FROM_NEW);
-}
-INTERCEPTOR(void, _ZdaPvRKSt9nothrow_t, void *ptr, std::nothrow_t const&) {
-  OPERATOR_DELETE_BODY(FROM_NEW_BR);
-}
-#endif
+INTERCEPTOR(void, _ZdlPv, void *ptr)
+{ OPERATOR_DELETE_BODY(FROM_NEW); }
+INTERCEPTOR(void, _ZdaPv, void *ptr)
+{ OPERATOR_DELETE_BODY(FROM_NEW_BR); }
+INTERCEPTOR(void, _ZdlPvRKSt9nothrow_t, void *ptr, std::nothrow_t const&)
+{ OPERATOR_DELETE_BODY(FROM_NEW); }
+INTERCEPTOR(void, _ZdaPvRKSt9nothrow_t, void *ptr, std::nothrow_t const&)
+{ OPERATOR_DELETE_BODY(FROM_NEW_BR); }
+#endif  // !SANITIZER_MAC
diff --git a/contrib/compiler-rt/lib/asan/asan_poisoning.cc b/contrib/compiler-rt/lib/asan/asan_poisoning.cc
index abb75ab3bf92..c3a82aa0049e 100644
--- a/contrib/compiler-rt/lib/asan/asan_poisoning.cc
+++ b/contrib/compiler-rt/lib/asan/asan_poisoning.cc
@@ -217,7 +217,7 @@ uptr __asan_region_is_poisoned(uptr beg, uptr size) {
       uptr __bad = __asan_region_is_poisoned(__p, __size);    \
       __asan_report_error(pc, bp, sp, __bad, isWrite, __size, 0);\
     }                                                         \
-  } while (false);                                            \
+  } while (false)
 
 
 extern "C" SANITIZER_INTERFACE_ATTRIBUTE
diff --git a/contrib/compiler-rt/lib/asan/asan_poisoning.h b/contrib/compiler-rt/lib/asan/asan_poisoning.h
index cc3281e08a71..1e00070bcf63 100644
--- a/contrib/compiler-rt/lib/asan/asan_poisoning.h
+++ b/contrib/compiler-rt/lib/asan/asan_poisoning.h
@@ -46,8 +46,11 @@ ALWAYS_INLINE void FastPoisonShadow(uptr aligned_beg, uptr aligned_size,
   // for mapping shadow and zeroing out pages doesn't "just work", so we should
   // probably provide higher-level interface for these operations.
   // For now, just memset on Windows.
-  if (value ||
-      SANITIZER_WINDOWS == 1 ||
+  if (value || SANITIZER_WINDOWS == 1 ||
+      // TODO(mcgrathr): Fuchsia doesn't allow the shadow mapping to be
+      // changed at all.  It doesn't currently have an efficient means
+      // to zero a bunch of pages, but maybe we should add one.
+      SANITIZER_FUCHSIA == 1 ||
       shadow_end - shadow_beg < common_flags()->clear_shadow_mmap_threshold) {
     REAL(memset)((void*)shadow_beg, value, shadow_end - shadow_beg);
   } else {
diff --git a/contrib/compiler-rt/lib/asan/asan_posix.cc b/contrib/compiler-rt/lib/asan/asan_posix.cc
index added746ace8..17c28b0aea2a 100644
--- a/contrib/compiler-rt/lib/asan/asan_posix.cc
+++ b/contrib/compiler-rt/lib/asan/asan_posix.cc
@@ -25,7 +25,6 @@
 #include "sanitizer_common/sanitizer_procmaps.h"
 
 #include <pthread.h>
-#include <signal.h>
 #include <stdlib.h>
 #include <sys/time.h>
 #include <sys/resource.h>
@@ -34,58 +33,9 @@
 namespace __asan {
 
 void AsanOnDeadlySignal(int signo, void *siginfo, void *context) {
-  ScopedDeadlySignal signal_scope(GetCurrentThread());
-  int code = (int)((siginfo_t*)siginfo)->si_code;
-  // Write the first message using fd=2, just in case.
-  // It may actually fail to write in case stderr is closed.
-  internal_write(2, "ASAN:DEADLYSIGNAL\n", 18);
-  SignalContext sig = SignalContext::Create(siginfo, context);
-
-  // Access at a reasonable offset above SP, or slightly below it (to account
-  // for x86_64 or PowerPC redzone, ARM push of multiple registers, etc) is
-  // probably a stack overflow.
-#ifdef __s390__
-  // On s390, the fault address in siginfo points to start of the page, not
-  // to the precise word that was accessed.  Mask off the low bits of sp to
-  // take it into account.
-  bool IsStackAccess = sig.addr >= (sig.sp & ~0xFFF) &&
-                       sig.addr < sig.sp + 0xFFFF;
-#else
-  bool IsStackAccess = sig.addr + 512 > sig.sp && sig.addr < sig.sp + 0xFFFF;
-#endif
-
-#if __powerpc__
-  // Large stack frames can be allocated with e.g.
-  //   lis r0,-10000
-  //   stdux r1,r1,r0 # store sp to [sp-10000] and update sp by -10000
-  // If the store faults then sp will not have been updated, so test above
-  // will not work, because the fault address will be more than just "slightly"
-  // below sp.
-  if (!IsStackAccess && IsAccessibleMemoryRange(sig.pc, 4)) {
-    u32 inst = *(unsigned *)sig.pc;
-    u32 ra = (inst >> 16) & 0x1F;
-    u32 opcd = inst >> 26;
-    u32 xo = (inst >> 1) & 0x3FF;
-    // Check for store-with-update to sp. The instructions we accept are:
-    //   stbu rs,d(ra)          stbux rs,ra,rb
-    //   sthu rs,d(ra)          sthux rs,ra,rb
-    //   stwu rs,d(ra)          stwux rs,ra,rb
-    //   stdu rs,ds(ra)         stdux rs,ra,rb
-    // where ra is r1 (the stack pointer).
-    if (ra == 1 &&
-        (opcd == 39 || opcd == 45 || opcd == 37 || opcd == 62 ||
-         (opcd == 31 && (xo == 247 || xo == 439 || xo == 183 || xo == 181))))
-      IsStackAccess = true;
-  }
-#endif // __powerpc__
-
-  // We also check si_code to filter out SEGV caused by something else other
-  // then hitting the guard page or unmapped memory, like, for example,
-  // unaligned memory access.
-  if (IsStackAccess && (code == si_SEGV_MAPERR || code == si_SEGV_ACCERR))
-    ReportStackOverflow(sig);
-  else
-    ReportDeadlySignal(signo, sig);
+  StartReportDeadlySignal();
+  SignalContext sig(siginfo, context);
+  ReportDeadlySignal(sig);
 }
 
 // ---------------------- TSD ---------------- {{{1
diff --git a/contrib/compiler-rt/lib/asan/asan_premap_shadow.cc b/contrib/compiler-rt/lib/asan/asan_premap_shadow.cc
new file mode 100644
index 000000000000..229eba99fe0e
--- /dev/null
+++ b/contrib/compiler-rt/lib/asan/asan_premap_shadow.cc
@@ -0,0 +1,79 @@
+//===-- asan_premap_shadow.cc ---------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of AddressSanitizer, an address sanity checker.
+//
+// Reserve shadow memory with an ifunc resolver.
+//===----------------------------------------------------------------------===//
+
+#include "asan_mapping.h"
+
+#if ASAN_PREMAP_SHADOW
+
+#include "asan_premap_shadow.h"
+#include "sanitizer_common/sanitizer_posix.h"
+
+namespace __asan {
+
+// The code in this file needs to run in an unrelocated binary. It may not
+// access any external symbol, including its own non-hidden globals.
+
+// Conservative upper limit.
+uptr PremapShadowSize() {
+  uptr granularity = GetMmapGranularity();
+  return RoundUpTo(GetMaxVirtualAddress() >> SHADOW_SCALE, granularity);
+}
+
+// Returns an address aligned to 8 pages, such that one page on the left and
+// PremapShadowSize() bytes on the right of it are mapped r/o.
+uptr PremapShadow() {
+  uptr granularity = GetMmapGranularity();
+  uptr alignment = granularity * 8;
+  uptr left_padding = granularity;
+  uptr shadow_size = PremapShadowSize();
+  uptr map_size = shadow_size + left_padding + alignment;
+
+  uptr map_start = (uptr)MmapNoAccess(map_size);
+  CHECK_NE(map_start, ~(uptr)0);
+
+  uptr shadow_start = RoundUpTo(map_start + left_padding, alignment);
+  uptr shadow_end = shadow_start + shadow_size;
+  internal_munmap(reinterpret_cast<void *>(map_start),
+                  shadow_start - left_padding - map_start);
+  internal_munmap(reinterpret_cast<void *>(shadow_end),
+                  map_start + map_size - shadow_end);
+  return shadow_start;
+}
+
+bool PremapShadowFailed() {
+  uptr shadow = reinterpret_cast<uptr>(&__asan_shadow);
+  uptr resolver = reinterpret_cast<uptr>(&__asan_premap_shadow);
+  // shadow == resolver is how Android KitKat and older handles ifunc.
+  // shadow == 0 just in case.
+  if (shadow == 0 || shadow == resolver)
+    return true;
+  return false;
+}
+} // namespace __asan
+
+extern "C" {
+decltype(__asan_shadow)* __asan_premap_shadow() {
+  // The resolver may be called multiple times. Map the shadow just once.
+  static uptr premapped_shadow = 0;
+  if (!premapped_shadow) premapped_shadow = __asan::PremapShadow();
+  return reinterpret_cast<decltype(__asan_shadow)*>(premapped_shadow);
+}
+
+// __asan_shadow is a "function" that has the same address as the first byte of
+// the shadow mapping.
+INTERFACE_ATTRIBUTE __attribute__((ifunc("__asan_premap_shadow"))) void
+__asan_shadow();
+}
+
+#endif // ASAN_PREMAP_SHADOW
diff --git a/contrib/compiler-rt/lib/asan/asan_premap_shadow.h b/contrib/compiler-rt/lib/asan/asan_premap_shadow.h
new file mode 100644
index 000000000000..41acbdbbb69f
--- /dev/null
+++ b/contrib/compiler-rt/lib/asan/asan_premap_shadow.h
@@ -0,0 +1,30 @@
+//===-- asan_mapping.h ------------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of AddressSanitizer, an address sanity checker.
+//
+// Premap shadow range with an ifunc resolver.
+//===----------------------------------------------------------------------===//
+
+
+#ifndef ASAN_PREMAP_SHADOW_H
+#define ASAN_PREMAP_SHADOW_H
+
+#if ASAN_PREMAP_SHADOW
+namespace __asan {
+// Conservative upper limit.
+uptr PremapShadowSize();
+bool PremapShadowFailed();
+}
+#endif
+
+extern "C" INTERFACE_ATTRIBUTE void __asan_shadow();
+extern "C" decltype(__asan_shadow)* __asan_premap_shadow();
+
+#endif // ASAN_PREMAP_SHADOW_H
diff --git a/contrib/compiler-rt/lib/asan/asan_report.cc b/contrib/compiler-rt/lib/asan/asan_report.cc
index 2e477f258b8d..e3bc02994efd 100644
--- a/contrib/compiler-rt/lib/asan/asan_report.cc
+++ b/contrib/compiler-rt/lib/asan/asan_report.cc
@@ -60,9 +60,8 @@ void PrintMemoryByte(InternalScopedString *str, const char *before, u8 byte,
                      bool in_shadow, const char *after) {
   Decorator d;
   str->append("%s%s%x%x%s%s", before,
-              in_shadow ? d.ShadowByte(byte) : d.MemoryByte(),
-              byte >> 4, byte & 15,
-              in_shadow ? d.EndShadowByte() : d.EndMemoryByte(), after);
+              in_shadow ? d.ShadowByte(byte) : d.MemoryByte(), byte >> 4,
+              byte & 15, d.Default(), after);
 }
 
 static void PrintZoneForPointer(uptr ptr, uptr zone_ptr,
@@ -123,53 +122,15 @@ bool ParseFrameDescription(const char *frame_descr,
 // immediately after printing error report.
 class ScopedInErrorReport {
  public:
-  explicit ScopedInErrorReport(bool fatal = false) {
-    halt_on_error_ = fatal || flags()->halt_on_error;
-
-    if (lock_.TryLock()) {
-      StartReporting();
-      return;
-    }
-
-    // ASan found two bugs in different threads simultaneously.
-
-    u32 current_tid = GetCurrentTidOrInvalid();
-    if (reporting_thread_tid_ == current_tid ||
-        reporting_thread_tid_ == kInvalidTid) {
-      // This is either asynch signal or nested error during error reporting.
-      // Fail simple to avoid deadlocks in Report().
-
-      // Can't use Report() here because of potential deadlocks
-      // in nested signal handlers.
-      const char msg[] = "AddressSanitizer: nested bug in the same thread, "
-                         "aborting.\n";
-      WriteToFile(kStderrFd, msg, sizeof(msg));
-
-      internal__exit(common_flags()->exitcode);
-    }
-
-    if (halt_on_error_) {
-      // Do not print more than one report, otherwise they will mix up.
-      // Error reporting functions shouldn't return at this situation, as
-      // they are effectively no-returns.
-
-      Report("AddressSanitizer: while reporting a bug found another one. "
-             "Ignoring.\n");
-
-      // Sleep long enough to make sure that the thread which started
-      // to print an error report will finish doing it.
-      SleepForSeconds(Max(100, flags()->sleep_before_dying + 1));
-
-      // If we're still not dead for some reason, use raw _exit() instead of
-      // Die() to bypass any additional checks.
-      internal__exit(common_flags()->exitcode);
-    } else {
-      // The other thread will eventually finish reporting
-      // so it's safe to wait
-      lock_.Lock();
-    }
-
-    StartReporting();
+  explicit ScopedInErrorReport(bool fatal = false)
+      : halt_on_error_(fatal || flags()->halt_on_error) {
+    // Make sure the registry and sanitizer report mutexes are locked while
+    // we're printing an error report.
+    // We can lock them only here to avoid self-deadlock in case of
+    // recursive reports.
+    asanThreadRegistry().Lock();
+    Printf(
+        "=================================================================\n");
   }
 
   ~ScopedInErrorReport() {
@@ -217,9 +178,6 @@ class ScopedInErrorReport {
     if (!halt_on_error_)
       internal_memset(&current_error_, 0, sizeof(current_error_));
 
-    CommonSanitizerReportMutex.Unlock();
-    reporting_thread_tid_ = kInvalidTid;
-    lock_.Unlock();
     if (halt_on_error_) {
       Report("ABORTING\n");
       Die();
@@ -237,39 +195,18 @@ class ScopedInErrorReport {
   }
 
  private:
-  void StartReporting() {
-    // Make sure the registry and sanitizer report mutexes are locked while
-    // we're printing an error report.
-    // We can lock them only here to avoid self-deadlock in case of
-    // recursive reports.
-    asanThreadRegistry().Lock();
-    CommonSanitizerReportMutex.Lock();
-    reporting_thread_tid_ = GetCurrentTidOrInvalid();
-    Printf("===================================================="
-           "=============\n");
-  }
-
-  static StaticSpinMutex lock_;
-  static u32 reporting_thread_tid_;
+  ScopedErrorReportLock error_report_lock_;
   // Error currently being reported. This enables the destructor to interact
   // with the debugger and point it to an error description.
   static ErrorDescription current_error_;
   bool halt_on_error_;
 };
 
-StaticSpinMutex ScopedInErrorReport::lock_;
-u32 ScopedInErrorReport::reporting_thread_tid_ = kInvalidTid;
 ErrorDescription ScopedInErrorReport::current_error_;
 
-void ReportStackOverflow(const SignalContext &sig) {
+void ReportDeadlySignal(const SignalContext &sig) {
   ScopedInErrorReport in_report(/*fatal*/ true);
-  ErrorStackOverflow error(GetCurrentTidOrInvalid(), sig);
-  in_report.ReportError(error);
-}
-
-void ReportDeadlySignal(int signo, const SignalContext &sig) {
-  ScopedInErrorReport in_report(/*fatal*/ true);
-  ErrorDeadlySignal error(GetCurrentTidOrInvalid(), sig, signo);
+  ErrorDeadlySignal error(GetCurrentTidOrInvalid(), sig);
   in_report.ReportError(error);
 }
 
@@ -279,11 +216,12 @@ void ReportDoubleFree(uptr addr, BufferedStackTrace *free_stack) {
   in_report.ReportError(error);
 }
 
-void ReportNewDeleteSizeMismatch(uptr addr, uptr delete_size,
+void ReportNewDeleteTypeMismatch(uptr addr, uptr delete_size,
+                                 uptr delete_alignment,
                                  BufferedStackTrace *free_stack) {
   ScopedInErrorReport in_report;
-  ErrorNewDeleteSizeMismatch error(GetCurrentTidOrInvalid(), free_stack, addr,
-                                   delete_size);
+  ErrorNewDeleteTypeMismatch error(GetCurrentTidOrInvalid(), free_stack, addr,
+                                   delete_size, delete_alignment);
   in_report.ReportError(error);
 }
 
@@ -360,17 +298,58 @@ static NOINLINE void ReportInvalidPointerPair(uptr pc, uptr bp, uptr sp,
   in_report.ReportError(error);
 }
 
+static bool IsInvalidPointerPair(uptr a1, uptr a2) {
+  if (a1 == a2)
+    return false;
+
+  // 256B in shadow memory can be iterated quite fast
+  static const uptr kMaxOffset = 2048;
+
+  uptr left = a1 < a2 ? a1 : a2;
+  uptr right = a1 < a2 ? a2 : a1;
+  uptr offset = right - left;
+  if (offset <= kMaxOffset)
+    return __asan_region_is_poisoned(left, offset);
+
+  AsanThread *t = GetCurrentThread();
+
+  // check whether left is a stack memory pointer
+  if (uptr shadow_offset1 = t->GetStackVariableShadowStart(left)) {
+    uptr shadow_offset2 = t->GetStackVariableShadowStart(right);
+    return shadow_offset2 == 0 || shadow_offset1 != shadow_offset2;
+  }
+
+  // check whether left is a heap memory address
+  HeapAddressDescription hdesc1, hdesc2;
+  if (GetHeapAddressInformation(left, 0, &hdesc1) &&
+      hdesc1.chunk_access.access_type == kAccessTypeInside)
+    return !GetHeapAddressInformation(right, 0, &hdesc2) ||
+        hdesc2.chunk_access.access_type != kAccessTypeInside ||
+        hdesc1.chunk_access.chunk_begin != hdesc2.chunk_access.chunk_begin;
+
+  // check whether left is an address of a global variable
+  GlobalAddressDescription gdesc1, gdesc2;
+  if (GetGlobalAddressInformation(left, 0, &gdesc1))
+    return !GetGlobalAddressInformation(right - 1, 0, &gdesc2) ||
+        !gdesc1.PointsInsideTheSameVariable(gdesc2);
+
+  if (t->GetStackVariableShadowStart(right) ||
+      GetHeapAddressInformation(right, 0, &hdesc2) ||
+      GetGlobalAddressInformation(right - 1, 0, &gdesc2))
+    return true;
+
+  // At this point we know nothing about both a1 and a2 addresses.
+  return false;
+}
+
 static INLINE void CheckForInvalidPointerPair(void *p1, void *p2) {
   if (!flags()->detect_invalid_pointer_pairs) return;
   uptr a1 = reinterpret_cast<uptr>(p1);
   uptr a2 = reinterpret_cast<uptr>(p2);
-  AsanChunkView chunk1 = FindHeapChunkByAddress(a1);
-  AsanChunkView chunk2 = FindHeapChunkByAddress(a2);
-  bool valid1 = chunk1.IsAllocated();
-  bool valid2 = chunk2.IsAllocated();
-  if (!valid1 || !valid2 || !chunk1.Eq(chunk2)) {
+
+  if (IsInvalidPointerPair(a1, a2)) {
     GET_CALLER_PC_BP_SP;
-    return ReportInvalidPointerPair(pc, bp, sp, a1, a2);
+    ReportInvalidPointerPair(pc, bp, sp, a1, a2);
   }
 }
 // ----------------------- Mac-specific reports ----------------- {{{1
diff --git a/contrib/compiler-rt/lib/asan/asan_report.h b/contrib/compiler-rt/lib/asan/asan_report.h
index 5a3533a319af..f2cdad47ee81 100644
--- a/contrib/compiler-rt/lib/asan/asan_report.h
+++ b/contrib/compiler-rt/lib/asan/asan_report.h
@@ -46,9 +46,9 @@ bool ParseFrameDescription(const char *frame_descr,
 // Different kinds of error reports.
 void ReportGenericError(uptr pc, uptr bp, uptr sp, uptr addr, bool is_write,
                         uptr access_size, u32 exp, bool fatal);
-void ReportStackOverflow(const SignalContext &sig);
-void ReportDeadlySignal(int signo, const SignalContext &sig);
-void ReportNewDeleteSizeMismatch(uptr addr, uptr delete_size,
+void ReportDeadlySignal(const SignalContext &sig);
+void ReportNewDeleteTypeMismatch(uptr addr, uptr delete_size,
+                                 uptr delete_alignment,
                                  BufferedStackTrace *free_stack);
 void ReportDoubleFree(uptr addr, BufferedStackTrace *free_stack);
 void ReportFreeNotMalloced(uptr addr, BufferedStackTrace *free_stack);
diff --git a/contrib/compiler-rt/lib/asan/asan_rtl.cc b/contrib/compiler-rt/lib/asan/asan_rtl.cc
index 5ae3568ae04a..21fd0e240708 100644
--- a/contrib/compiler-rt/lib/asan/asan_rtl.cc
+++ b/contrib/compiler-rt/lib/asan/asan_rtl.cc
@@ -84,26 +84,6 @@ void ShowStatsAndAbort() {
   Die();
 }
 
-// ---------------------- mmap -------------------- {{{1
-// Reserve memory range [beg, end].
-// We need to use inclusive range because end+1 may not be representable.
-void ReserveShadowMemoryRange(uptr beg, uptr end, const char *name) {
-  CHECK_EQ((beg % GetMmapGranularity()), 0);
-  CHECK_EQ(((end + 1) % GetMmapGranularity()), 0);
-  uptr size = end - beg + 1;
-  DecreaseTotalMmap(size);  // Don't count the shadow against mmap_limit_mb.
-  void *res = MmapFixedNoReserve(beg, size, name);
-  if (res != (void*)beg) {
-    Report("ReserveShadowMemoryRange failed while trying to map 0x%zx bytes. "
-           "Perhaps you're using ulimit -v\n", size);
-    Abort();
-  }
-  if (common_flags()->no_huge_pages_for_shadow)
-    NoHugePagesInRegion(beg, size);
-  if (common_flags()->use_madv_dontdump)
-    DontDumpShadowMemory(beg, size);
-}
-
 // --------------- LowLevelAllocateCallbac ---------- {{{1
 static void OnLowLevelAllocate(uptr ptr, uptr size) {
   PoisonShadow(ptr, size, kAsanInternalHeapMagic);
@@ -327,7 +307,7 @@ static void asan_atexit() {
 
 static void InitializeHighMemEnd() {
 #if !ASAN_FIXED_MAPPING
-  kHighMemEnd = GetMaxVirtualAddress();
+  kHighMemEnd = GetMaxUserVirtualAddress();
   // Increase kHighMemEnd to make sure it's properly
   // aligned together with kHighMemBeg:
   kHighMemEnd |= SHADOW_GRANULARITY * GetMmapGranularity() - 1;
@@ -335,46 +315,7 @@ static void InitializeHighMemEnd() {
   CHECK_EQ((kHighMemBeg % GetMmapGranularity()), 0);
 }
 
-static void ProtectGap(uptr addr, uptr size) {
-  if (!flags()->protect_shadow_gap) {
-    // The shadow gap is unprotected, so there is a chance that someone
-    // is actually using this memory. Which means it needs a shadow...
-    uptr GapShadowBeg = RoundDownTo(MEM_TO_SHADOW(addr), GetPageSizeCached());
-    uptr GapShadowEnd =
-        RoundUpTo(MEM_TO_SHADOW(addr + size), GetPageSizeCached()) - 1;
-    if (Verbosity())
-      Printf("protect_shadow_gap=0:"
-             " not protecting shadow gap, allocating gap's shadow\n"
-             "|| `[%p, %p]` || ShadowGap's shadow ||\n", GapShadowBeg,
-             GapShadowEnd);
-    ReserveShadowMemoryRange(GapShadowBeg, GapShadowEnd,
-                             "unprotected gap shadow");
-    return;
-  }
-  void *res = MmapFixedNoAccess(addr, size, "shadow gap");
-  if (addr == (uptr)res)
-    return;
-  // A few pages at the start of the address space can not be protected.
-  // But we really want to protect as much as possible, to prevent this memory
-  // being returned as a result of a non-FIXED mmap().
-  if (addr == kZeroBaseShadowStart) {
-    uptr step = GetMmapGranularity();
-    while (size > step && addr < kZeroBaseMaxShadowStart) {
-      addr += step;
-      size -= step;
-      void *res = MmapFixedNoAccess(addr, size, "shadow gap");
-      if (addr == (uptr)res)
-        return;
-    }
-  }
-
-  Report("ERROR: Failed to protect the shadow gap. "
-         "ASan cannot proceed correctly. ABORTING.\n");
-  DumpProcessMap();
-  Die();
-}
-
-static void PrintAddressSpaceLayout() {
+void PrintAddressSpaceLayout() {
   Printf("|| `[%p, %p]` || HighMem    ||\n",
          (void*)kHighMemBeg, (void*)kHighMemEnd);
   Printf("|| `[%p, %p]` || HighShadow ||\n",
@@ -426,71 +367,6 @@ static void PrintAddressSpaceLayout() {
           kHighShadowBeg > kMidMemEnd);
 }
 
-static void InitializeShadowMemory() {
-  // Set the shadow memory address to uninitialized.
-  __asan_shadow_memory_dynamic_address = kDefaultShadowSentinel;
-
-  uptr shadow_start = kLowShadowBeg;
-  // Detect if a dynamic shadow address must used and find a available location
-  // when necessary. When dynamic address is used, the macro |kLowShadowBeg|
-  // expands to |__asan_shadow_memory_dynamic_address| which is
-  // |kDefaultShadowSentinel|.
-  if (shadow_start == kDefaultShadowSentinel) {
-    __asan_shadow_memory_dynamic_address = 0;
-    CHECK_EQ(0, kLowShadowBeg);
-    shadow_start = FindDynamicShadowStart();
-  }
-  // Update the shadow memory address (potentially) used by instrumentation.
-  __asan_shadow_memory_dynamic_address = shadow_start;
-
-  if (kLowShadowBeg)
-    shadow_start -= GetMmapGranularity();
-  bool full_shadow_is_available =
-      MemoryRangeIsAvailable(shadow_start, kHighShadowEnd);
-
-#if SANITIZER_LINUX && defined(__x86_64__) && defined(_LP64) &&                \
-    !ASAN_FIXED_MAPPING
-  if (!full_shadow_is_available) {
-    kMidMemBeg = kLowMemEnd < 0x3000000000ULL ? 0x3000000000ULL : 0;
-    kMidMemEnd = kLowMemEnd < 0x3000000000ULL ? 0x4fffffffffULL : 0;
-  }
-#endif
-
-  if (Verbosity()) PrintAddressSpaceLayout();
-
-  if (full_shadow_is_available) {
-    // mmap the low shadow plus at least one page at the left.
-    if (kLowShadowBeg)
-      ReserveShadowMemoryRange(shadow_start, kLowShadowEnd, "low shadow");
-    // mmap the high shadow.
-    ReserveShadowMemoryRange(kHighShadowBeg, kHighShadowEnd, "high shadow");
-    // protect the gap.
-    ProtectGap(kShadowGapBeg, kShadowGapEnd - kShadowGapBeg + 1);
-    CHECK_EQ(kShadowGapEnd, kHighShadowBeg - 1);
-  } else if (kMidMemBeg &&
-      MemoryRangeIsAvailable(shadow_start, kMidMemBeg - 1) &&
-      MemoryRangeIsAvailable(kMidMemEnd + 1, kHighShadowEnd)) {
-    CHECK(kLowShadowBeg != kLowShadowEnd);
-    // mmap the low shadow plus at least one page at the left.
-    ReserveShadowMemoryRange(shadow_start, kLowShadowEnd, "low shadow");
-    // mmap the mid shadow.
-    ReserveShadowMemoryRange(kMidShadowBeg, kMidShadowEnd, "mid shadow");
-    // mmap the high shadow.
-    ReserveShadowMemoryRange(kHighShadowBeg, kHighShadowEnd, "high shadow");
-    // protect the gaps.
-    ProtectGap(kShadowGapBeg, kShadowGapEnd - kShadowGapBeg + 1);
-    ProtectGap(kShadowGap2Beg, kShadowGap2End - kShadowGap2Beg + 1);
-    ProtectGap(kShadowGap3Beg, kShadowGap3End - kShadowGap3Beg + 1);
-  } else {
-    Report("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);
-    DumpProcessMap();
-    Die();
-  }
-}
-
 static void AsanInitInternal() {
   if (LIKELY(asan_inited)) return;
   SanitizerToolName = "AddressSanitizer";
@@ -531,6 +407,7 @@ static void AsanInitInternal() {
   MaybeReexec();
 
   // Setup internal allocator callback.
+  SetLowLevelAllocateMinAlignment(SHADOW_GRANULARITY);
   SetLowLevelAllocateCallback(OnLowLevelAllocate);
 
   InitializeAsanInterceptors();
@@ -575,20 +452,18 @@ static void AsanInitInternal() {
   InitTlsSize();
 
   // Create main thread.
-  AsanThread *main_thread = AsanThread::Create(
-      /* start_routine */ nullptr, /* arg */ nullptr, /* parent_tid */ 0,
-      /* stack */ nullptr, /* detached */ true);
+  AsanThread *main_thread = CreateMainThread();
   CHECK_EQ(0, main_thread->tid());
-  SetCurrentThread(main_thread);
-  main_thread->ThreadStart(internal_getpid(),
-                           /* signal_thread_is_registered */ nullptr);
   force_interface_symbols();  // no-op.
   SanitizerInitializeUnwinder();
 
   if (CAN_SANITIZE_LEAKS) {
     __lsan::InitCommonLsan();
     if (common_flags()->detect_leaks && common_flags()->leak_check_at_exit) {
-      Atexit(__lsan::DoLeakCheck);
+      if (flags()->halt_on_error)
+        Atexit(__lsan::DoLeakCheck);
+      else
+        Atexit(__lsan::DoRecoverableLeakCheckVoid);
     }
   }
 
@@ -608,6 +483,11 @@ static void AsanInitInternal() {
   }
 
   VReport(1, "AddressSanitizer Init done\n");
+
+  if (flags()->sleep_after_init) {
+    Report("Sleeping for %d second(s)\n", flags()->sleep_after_init);
+    SleepForSeconds(flags()->sleep_after_init);
+  }
 }
 
 // Initialize as requested from some part of ASan runtime library (interceptors,
@@ -647,6 +527,7 @@ void NOINLINE __asan_handle_no_return() {
     top = curr_thread->stack_top();
     bottom = ((uptr)&local_stack - PageSize) & ~(PageSize - 1);
   } else {
+    CHECK(!SANITIZER_FUCHSIA);
     // If we haven't seen this thread, try asking the OS for stack bounds.
     uptr tls_addr, tls_size, stack_size;
     GetThreadStackAndTls(/*main=*/false, &bottom, &stack_size, &tls_addr,
diff --git a/contrib/compiler-rt/lib/asan/asan_scariness_score.h b/contrib/compiler-rt/lib/asan/asan_scariness_score.h
index 7f1571416fd5..7f095dd29f29 100644
--- a/contrib/compiler-rt/lib/asan/asan_scariness_score.h
+++ b/contrib/compiler-rt/lib/asan/asan_scariness_score.h
@@ -47,7 +47,7 @@ struct ScarinessScoreBase {
   };
   int GetScore() const { return score; }
   const char *GetDescription() const { return descr; }
-  void Print() {
+  void Print() const {
     if (score && flags()->print_scariness)
       Printf("SCARINESS: %d (%s)\n", score, descr);
   }
diff --git a/contrib/compiler-rt/lib/asan/asan_shadow_setup.cc b/contrib/compiler-rt/lib/asan/asan_shadow_setup.cc
new file mode 100644
index 000000000000..b3cf0b5c1ce2
--- /dev/null
+++ b/contrib/compiler-rt/lib/asan/asan_shadow_setup.cc
@@ -0,0 +1,165 @@
+//===-- asan_shadow_setup.cc ----------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of AddressSanitizer, an address sanity checker.
+//
+// Set up the shadow memory.
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_common/sanitizer_platform.h"
+
+// asan_fuchsia.cc has its own InitializeShadowMemory implementation.
+#if !SANITIZER_FUCHSIA
+
+#include "asan_internal.h"
+#include "asan_mapping.h"
+
+namespace __asan {
+
+// ---------------------- mmap -------------------- {{{1
+// Reserve memory range [beg, end].
+// We need to use inclusive range because end+1 may not be representable.
+void ReserveShadowMemoryRange(uptr beg, uptr end, const char *name) {
+  CHECK_EQ((beg % GetMmapGranularity()), 0);
+  CHECK_EQ(((end + 1) % GetMmapGranularity()), 0);
+  uptr size = end - beg + 1;
+  DecreaseTotalMmap(size);  // Don't count the shadow against mmap_limit_mb.
+  void *res = MmapFixedNoReserve(beg, size, name);
+  if (res != (void *)beg) {
+    Report(
+        "ReserveShadowMemoryRange failed while trying to map 0x%zx bytes. "
+        "Perhaps you're using ulimit -v\n",
+        size);
+    Abort();
+  }
+  if (common_flags()->no_huge_pages_for_shadow) NoHugePagesInRegion(beg, size);
+  if (common_flags()->use_madv_dontdump) DontDumpShadowMemory(beg, size);
+}
+
+static void ProtectGap(uptr addr, uptr size) {
+  if (!flags()->protect_shadow_gap) {
+    // The shadow gap is unprotected, so there is a chance that someone
+    // is actually using this memory. Which means it needs a shadow...
+    uptr GapShadowBeg = RoundDownTo(MEM_TO_SHADOW(addr), GetPageSizeCached());
+    uptr GapShadowEnd =
+        RoundUpTo(MEM_TO_SHADOW(addr + size), GetPageSizeCached()) - 1;
+    if (Verbosity())
+      Printf(
+          "protect_shadow_gap=0:"
+          " not protecting shadow gap, allocating gap's shadow\n"
+          "|| `[%p, %p]` || ShadowGap's shadow ||\n",
+          GapShadowBeg, GapShadowEnd);
+    ReserveShadowMemoryRange(GapShadowBeg, GapShadowEnd,
+                             "unprotected gap shadow");
+    return;
+  }
+  void *res = MmapFixedNoAccess(addr, size, "shadow gap");
+  if (addr == (uptr)res) return;
+  // A few pages at the start of the address space can not be protected.
+  // But we really want to protect as much as possible, to prevent this memory
+  // being returned as a result of a non-FIXED mmap().
+  if (addr == kZeroBaseShadowStart) {
+    uptr step = GetMmapGranularity();
+    while (size > step && addr < kZeroBaseMaxShadowStart) {
+      addr += step;
+      size -= step;
+      void *res = MmapFixedNoAccess(addr, size, "shadow gap");
+      if (addr == (uptr)res) return;
+    }
+  }
+
+  Report(
+      "ERROR: Failed to protect the shadow gap. "
+      "ASan cannot proceed correctly. ABORTING.\n");
+  DumpProcessMap();
+  Die();
+}
+
+static void MaybeReportLinuxPIEBug() {
+#if SANITIZER_LINUX && (defined(__x86_64__) || defined(__aarch64__))
+  Report("This might be related to ELF_ET_DYN_BASE change in Linux 4.12.\n");
+  Report(
+      "See https://github.com/google/sanitizers/issues/856 for possible "
+      "workarounds.\n");
+#endif
+}
+
+void InitializeShadowMemory() {
+  // Set the shadow memory address to uninitialized.
+  __asan_shadow_memory_dynamic_address = kDefaultShadowSentinel;
+
+  uptr shadow_start = kLowShadowBeg;
+  // Detect if a dynamic shadow address must used and find a available location
+  // when necessary. When dynamic address is used, the macro |kLowShadowBeg|
+  // expands to |__asan_shadow_memory_dynamic_address| which is
+  // |kDefaultShadowSentinel|.
+  bool full_shadow_is_available = false;
+  if (shadow_start == kDefaultShadowSentinel) {
+    __asan_shadow_memory_dynamic_address = 0;
+    CHECK_EQ(0, kLowShadowBeg);
+    shadow_start = FindDynamicShadowStart();
+    if (SANITIZER_LINUX) full_shadow_is_available = true;
+  }
+  // Update the shadow memory address (potentially) used by instrumentation.
+  __asan_shadow_memory_dynamic_address = shadow_start;
+
+  if (kLowShadowBeg) shadow_start -= GetMmapGranularity();
+
+  if (!full_shadow_is_available)
+    full_shadow_is_available =
+        MemoryRangeIsAvailable(shadow_start, kHighShadowEnd);
+
+#if SANITIZER_LINUX && defined(__x86_64__) && defined(_LP64) && \
+    !ASAN_FIXED_MAPPING
+  if (!full_shadow_is_available) {
+    kMidMemBeg = kLowMemEnd < 0x3000000000ULL ? 0x3000000000ULL : 0;
+    kMidMemEnd = kLowMemEnd < 0x3000000000ULL ? 0x4fffffffffULL : 0;
+  }
+#endif
+
+  if (Verbosity()) PrintAddressSpaceLayout();
+
+  if (full_shadow_is_available) {
+    // mmap the low shadow plus at least one page at the left.
+    if (kLowShadowBeg)
+      ReserveShadowMemoryRange(shadow_start, kLowShadowEnd, "low shadow");
+    // mmap the high shadow.
+    ReserveShadowMemoryRange(kHighShadowBeg, kHighShadowEnd, "high shadow");
+    // protect the gap.
+    ProtectGap(kShadowGapBeg, kShadowGapEnd - kShadowGapBeg + 1);
+    CHECK_EQ(kShadowGapEnd, kHighShadowBeg - 1);
+  } else if (kMidMemBeg &&
+             MemoryRangeIsAvailable(shadow_start, kMidMemBeg - 1) &&
+             MemoryRangeIsAvailable(kMidMemEnd + 1, kHighShadowEnd)) {
+    CHECK(kLowShadowBeg != kLowShadowEnd);
+    // mmap the low shadow plus at least one page at the left.
+    ReserveShadowMemoryRange(shadow_start, kLowShadowEnd, "low shadow");
+    // mmap the mid shadow.
+    ReserveShadowMemoryRange(kMidShadowBeg, kMidShadowEnd, "mid shadow");
+    // mmap the high shadow.
+    ReserveShadowMemoryRange(kHighShadowBeg, kHighShadowEnd, "high shadow");
+    // protect the gaps.
+    ProtectGap(kShadowGapBeg, kShadowGapEnd - kShadowGapBeg + 1);
+    ProtectGap(kShadowGap2Beg, kShadowGap2End - kShadowGap2Beg + 1);
+    ProtectGap(kShadowGap3Beg, kShadowGap3End - kShadowGap3Beg + 1);
+  } else {
+    Report(
+        "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);
+    MaybeReportLinuxPIEBug();
+    DumpProcessMap();
+    Die();
+  }
+}
+
+}  // namespace __asan
+
+#endif  // !SANITIZER_FUCHSIA
diff --git a/contrib/compiler-rt/lib/asan/asan_stack.h b/contrib/compiler-rt/lib/asan/asan_stack.h
index cc95e0f30a33..8e9df888d798 100644
--- a/contrib/compiler-rt/lib/asan/asan_stack.h
+++ b/contrib/compiler-rt/lib/asan/asan_stack.h
@@ -31,9 +31,8 @@ u32 GetMallocContextSize();
 // The pc will be in the position 0 of the resulting stack trace.
 // The bp may refer to the current frame or to the caller's frame.
 ALWAYS_INLINE
-void GetStackTraceWithPcBpAndContext(BufferedStackTrace *stack, uptr max_depth,
-                                     uptr pc, uptr bp, void *context,
-                                     bool fast) {
+void GetStackTrace(BufferedStackTrace *stack, uptr max_depth, uptr pc, uptr bp,
+                   void *context, bool fast) {
 #if SANITIZER_WINDOWS
   stack->Unwind(max_depth, pc, bp, context, 0, 0, fast);
 #else
@@ -41,10 +40,6 @@ void GetStackTraceWithPcBpAndContext(BufferedStackTrace *stack, uptr max_depth,
   stack->size = 0;
   if (LIKELY(asan_inited)) {
     if ((t = GetCurrentThread()) && !t->isUnwinding()) {
-      // On FreeBSD the slow unwinding that leverages _Unwind_Backtrace()
-      // yields the call stack of the signal's handler and not of the code
-      // that raised the signal (as it does on Linux).
-      if (SANITIZER_FREEBSD && t->isInDeadlySignal()) fast = true;
       uptr stack_top = t->stack_top();
       uptr stack_bottom = t->stack_bottom();
       ScopedUnwinding unwind_scope(t);
@@ -66,32 +61,29 @@ void GetStackTraceWithPcBpAndContext(BufferedStackTrace *stack, uptr max_depth,
 // as early as possible (in functions exposed to the user), as we generally
 // don't want stack trace to contain functions from ASan internals.
 
-#define GET_STACK_TRACE(max_size, fast)                                        \
-  BufferedStackTrace stack;                                                    \
-  if (max_size <= 2) {                                                         \
-    stack.size = max_size;                                                     \
-    if (max_size > 0) {                                                        \
-      stack.top_frame_bp = GET_CURRENT_FRAME();                                \
-      stack.trace_buffer[0] = StackTrace::GetCurrentPc();                      \
-      if (max_size > 1)                                                        \
-        stack.trace_buffer[1] = GET_CALLER_PC();                               \
-    }                                                                          \
-  } else {                                                                     \
-    GetStackTraceWithPcBpAndContext(&stack, max_size,                          \
-                                    StackTrace::GetCurrentPc(),                \
-                                    GET_CURRENT_FRAME(), 0, fast);             \
+#define GET_STACK_TRACE(max_size, fast)                          \
+  BufferedStackTrace stack;                                      \
+  if (max_size <= 2) {                                           \
+    stack.size = max_size;                                       \
+    if (max_size > 0) {                                          \
+      stack.top_frame_bp = GET_CURRENT_FRAME();                  \
+      stack.trace_buffer[0] = StackTrace::GetCurrentPc();        \
+      if (max_size > 1) stack.trace_buffer[1] = GET_CALLER_PC(); \
+    }                                                            \
+  } else {                                                       \
+    GetStackTrace(&stack, max_size, StackTrace::GetCurrentPc(),  \
+                  GET_CURRENT_FRAME(), 0, fast);                 \
   }
 
-#define GET_STACK_TRACE_FATAL(pc, bp)                                          \
-  BufferedStackTrace stack;                                                    \
-  GetStackTraceWithPcBpAndContext(&stack, kStackTraceMax, pc, bp, 0,           \
-                                  common_flags()->fast_unwind_on_fatal)
+#define GET_STACK_TRACE_FATAL(pc, bp)              \
+  BufferedStackTrace stack;                        \
+  GetStackTrace(&stack, kStackTraceMax, pc, bp, 0, \
+                common_flags()->fast_unwind_on_fatal)
 
-#define GET_STACK_TRACE_SIGNAL(sig)                                            \
-  BufferedStackTrace stack;                                                    \
-  GetStackTraceWithPcBpAndContext(&stack, kStackTraceMax,                      \
-                                  (sig).pc, (sig).bp, (sig).context,           \
-                                  common_flags()->fast_unwind_on_fatal)
+#define GET_STACK_TRACE_SIGNAL(sig)                                        \
+  BufferedStackTrace stack;                                                \
+  GetStackTrace(&stack, kStackTraceMax, (sig).pc, (sig).bp, (sig).context, \
+                common_flags()->fast_unwind_on_fatal)
 
 #define GET_STACK_TRACE_FATAL_HERE                                \
   GET_STACK_TRACE(kStackTraceMax, common_flags()->fast_unwind_on_fatal)
diff --git a/contrib/compiler-rt/lib/asan/asan_thread.cc b/contrib/compiler-rt/lib/asan/asan_thread.cc
index b1a0d9a3b37f..ad81512dff08 100644
--- a/contrib/compiler-rt/lib/asan/asan_thread.cc
+++ b/contrib/compiler-rt/lib/asan/asan_thread.cc
@@ -27,11 +27,6 @@ namespace __asan {
 
 // AsanThreadContext implementation.
 
-struct CreateThreadContextArgs {
-  AsanThread *thread;
-  StackTrace *stack;
-};
-
 void AsanThreadContext::OnCreated(void *arg) {
   CreateThreadContextArgs *args = static_cast<CreateThreadContextArgs*>(arg);
   if (args->stack)
@@ -88,7 +83,7 @@ AsanThread *AsanThread::Create(thread_callback_t start_routine, void *arg,
   AsanThread *thread = (AsanThread*)MmapOrDie(size, __func__);
   thread->start_routine_ = start_routine;
   thread->arg_ = arg;
-  CreateThreadContextArgs args = { thread, stack };
+  AsanThreadContext::CreateThreadContextArgs args = {thread, stack};
   asanThreadRegistry().CreateThread(*reinterpret_cast<uptr *>(thread), detached,
                                     parent_tid, &args);
 
@@ -223,12 +218,12 @@ FakeStack *AsanThread::AsyncSignalSafeLazyInitFakeStack() {
   return nullptr;
 }
 
-void AsanThread::Init() {
+void AsanThread::Init(const InitOptions *options) {
   next_stack_top_ = next_stack_bottom_ = 0;
   atomic_store(&stack_switching_, false, memory_order_release);
   fake_stack_ = nullptr;  // Will be initialized lazily if needed.
   CHECK_EQ(this->stack_size(), 0U);
-  SetThreadStackAndTls();
+  SetThreadStackAndTls(options);
   CHECK_GT(this->stack_size(), 0U);
   CHECK(AddrIsInMem(stack_bottom_));
   CHECK(AddrIsInMem(stack_top_ - 1));
@@ -239,6 +234,10 @@ void AsanThread::Init() {
           &local);
 }
 
+// Fuchsia doesn't use ThreadStart.
+// asan_fuchsia.c defines CreateMainThread and SetThreadStackAndTls.
+#if !SANITIZER_FUCHSIA
+
 thread_return_t AsanThread::ThreadStart(
     tid_t os_id, atomic_uintptr_t *signal_thread_is_registered) {
   Init();
@@ -270,7 +269,21 @@ thread_return_t AsanThread::ThreadStart(
   return res;
 }
 
-void AsanThread::SetThreadStackAndTls() {
+AsanThread *CreateMainThread() {
+  AsanThread *main_thread = AsanThread::Create(
+      /* start_routine */ nullptr, /* arg */ nullptr, /* parent_tid */ 0,
+      /* stack */ nullptr, /* detached */ true);
+  SetCurrentThread(main_thread);
+  main_thread->ThreadStart(internal_getpid(),
+                           /* signal_thread_is_registered */ nullptr);
+  return main_thread;
+}
+
+// This implementation doesn't use the argument, which is just passed down
+// from the caller of Init (which see, above).  It's only there to support
+// OS-specific implementations that need more information passed through.
+void AsanThread::SetThreadStackAndTls(const InitOptions *options) {
+  DCHECK_EQ(options, nullptr);
   uptr tls_size = 0;
   uptr stack_size = 0;
   GetThreadStackAndTls(tid() == 0, const_cast<uptr *>(&stack_bottom_),
@@ -283,6 +296,8 @@ void AsanThread::SetThreadStackAndTls() {
   CHECK(AddrIsInStack((uptr)&local));
 }
 
+#endif  // !SANITIZER_FUCHSIA
+
 void AsanThread::ClearShadowForThreadStackAndTLS() {
   PoisonShadow(stack_bottom_, stack_top_ - stack_bottom_, 0);
   if (tls_begin_ != tls_end_)
@@ -302,7 +317,7 @@ bool AsanThread::GetStackFrameAccessByAddr(uptr addr,
     access->frame_descr = (const char *)((uptr*)bottom)[1];
     return true;
   }
-  uptr aligned_addr = addr & ~(SANITIZER_WORDSIZE/8 - 1);  // align addr.
+  uptr aligned_addr = RoundDownTo(addr, SANITIZER_WORDSIZE / 8);  // align addr.
   uptr mem_ptr = RoundDownTo(aligned_addr, SHADOW_GRANULARITY);
   u8 *shadow_ptr = (u8*)MemToShadow(aligned_addr);
   u8 *shadow_bottom = (u8*)MemToShadow(bottom);
@@ -331,6 +346,29 @@ bool AsanThread::GetStackFrameAccessByAddr(uptr addr,
   return true;
 }
 
+uptr AsanThread::GetStackVariableShadowStart(uptr addr) {
+  uptr bottom = 0;
+  if (AddrIsInStack(addr)) {
+    bottom = stack_bottom();
+  } else if (has_fake_stack()) {
+    bottom = fake_stack()->AddrIsInFakeStack(addr);
+    CHECK(bottom);
+  } else
+    return 0;
+
+  uptr aligned_addr = RoundDownTo(addr, SANITIZER_WORDSIZE / 8);  // align addr.
+  u8 *shadow_ptr = (u8*)MemToShadow(aligned_addr);
+  u8 *shadow_bottom = (u8*)MemToShadow(bottom);
+
+  while (shadow_ptr >= shadow_bottom &&
+         (*shadow_ptr != kAsanStackLeftRedzoneMagic &&
+          *shadow_ptr != kAsanStackMidRedzoneMagic &&
+          *shadow_ptr != kAsanStackRightRedzoneMagic))
+    shadow_ptr--;
+
+  return (uptr)shadow_ptr + 1;
+}
+
 bool AsanThread::AddrIsInStack(uptr addr) {
   const auto bounds = GetStackBounds();
   return addr >= bounds.bottom && addr < bounds.top;
diff --git a/contrib/compiler-rt/lib/asan/asan_thread.h b/contrib/compiler-rt/lib/asan/asan_thread.h
index 424f9e68dfea..66091921101f 100644
--- a/contrib/compiler-rt/lib/asan/asan_thread.h
+++ b/contrib/compiler-rt/lib/asan/asan_thread.h
@@ -49,6 +49,11 @@ class AsanThreadContext : public ThreadContextBase {
 
   void OnCreated(void *arg) override;
   void OnFinished() override;
+
+  struct CreateThreadContextArgs {
+    AsanThread *thread;
+    StackTrace *stack;
+  };
 };
 
 // AsanThreadContext objects are never freed, so we need many of them.
@@ -62,7 +67,9 @@ class AsanThread {
   static void TSDDtor(void *tsd);
   void Destroy();
 
-  void Init();  // Should be called from the thread itself.
+  struct InitOptions;
+  void Init(const InitOptions *options = nullptr);
+
   thread_return_t ThreadStart(tid_t os_id,
                               atomic_uintptr_t *signal_thread_is_registered);
 
@@ -83,6 +90,9 @@ class AsanThread {
   };
   bool GetStackFrameAccessByAddr(uptr addr, StackFrameAccess *access);
 
+  // Returns a pointer to the start of the stack variable's shadow memory.
+  uptr GetStackVariableShadowStart(uptr addr);
+
   bool AddrIsInStack(uptr addr);
 
   void DeleteFakeStack(int tid) {
@@ -118,17 +128,15 @@ class AsanThread {
   bool isUnwinding() const { return unwinding_; }
   void setUnwinding(bool b) { unwinding_ = b; }
 
-  // True if we are in a deadly signal handler.
-  bool isInDeadlySignal() const { return in_deadly_signal_; }
-  void setInDeadlySignal(bool b) { in_deadly_signal_ = b; }
-
   AsanThreadLocalMallocStorage &malloc_storage() { return malloc_storage_; }
   AsanStats &stats() { return stats_; }
 
  private:
   // NOTE: There is no AsanThread constructor. It is allocated
   // via mmap() and *must* be valid in zero-initialized state.
-  void SetThreadStackAndTls();
+
+  void SetThreadStackAndTls(const InitOptions *options);
+
   void ClearShadowForThreadStackAndTLS();
   FakeStack *AsyncSignalSafeLazyInitFakeStack();
 
@@ -158,7 +166,6 @@ class AsanThread {
   AsanThreadLocalMallocStorage malloc_storage_;
   AsanStats stats_;
   bool unwinding_;
-  bool in_deadly_signal_;
 };
 
 // ScopedUnwinding is a scope for stacktracing member of a context
@@ -173,20 +180,6 @@ class ScopedUnwinding {
   AsanThread *thread;
 };
 
-// ScopedDeadlySignal is a scope for handling deadly signals.
-class ScopedDeadlySignal {
- public:
-  explicit ScopedDeadlySignal(AsanThread *t) : thread(t) {
-    if (thread) thread->setInDeadlySignal(true);
-  }
-  ~ScopedDeadlySignal() {
-    if (thread) thread->setInDeadlySignal(false);
-  }
-
- private:
-  AsanThread *thread;
-};
-
 // Returns a single instance of registry.
 ThreadRegistry &asanThreadRegistry();
 
diff --git a/contrib/compiler-rt/lib/asan/asan_win.cc b/contrib/compiler-rt/lib/asan/asan_win.cc
index 8a839d913f95..68eedd1863bc 100644
--- a/contrib/compiler-rt/lib/asan/asan_win.cc
+++ b/contrib/compiler-rt/lib/asan/asan_win.cc
@@ -57,8 +57,8 @@ long __asan_unhandled_exception_filter(EXCEPTION_POINTERS *info) {
 
   // FIXME: Handle EXCEPTION_STACK_OVERFLOW here.
 
-  SignalContext sig = SignalContext::Create(exception_record, context);
-  ReportDeadlySignal(exception_record->ExceptionCode, sig);
+  SignalContext sig(exception_record, context);
+  ReportDeadlySignal(sig);
   UNREACHABLE("returned from reporting deadly signal");
 }
 
diff --git a/contrib/compiler-rt/lib/builtins/adddf3.c b/contrib/compiler-rt/lib/builtins/adddf3.c
index c528e9e21f51..9a3901312e51 100644
--- a/contrib/compiler-rt/lib/builtins/adddf3.c
+++ b/contrib/compiler-rt/lib/builtins/adddf3.c
@@ -20,8 +20,11 @@ COMPILER_RT_ABI double __adddf3(double a, double b){
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI double __aeabi_dadd(double a, double b) {
   return __adddf3(a, b);
 }
+#else
+AEABI_RTABI double __aeabi_dadd(double a, double b) COMPILER_RT_ALIAS(__adddf3);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/addsf3.c b/contrib/compiler-rt/lib/builtins/addsf3.c
index fe570687a25e..c5c1a41c3611 100644
--- a/contrib/compiler-rt/lib/builtins/addsf3.c
+++ b/contrib/compiler-rt/lib/builtins/addsf3.c
@@ -20,8 +20,11 @@ COMPILER_RT_ABI float __addsf3(float a, float b) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI float __aeabi_fadd(float a, float b) {
   return __addsf3(a, b);
 }
+#else
+AEABI_RTABI float __aeabi_fadd(float a, float b) COMPILER_RT_ALIAS(__addsf3);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/arm/aeabi_cdcmp.S b/contrib/compiler-rt/lib/builtins/arm/aeabi_cdcmp.S
index 3e7a8b86b739..87dd03dce94d 100644
--- a/contrib/compiler-rt/lib/builtins/arm/aeabi_cdcmp.S
+++ b/contrib/compiler-rt/lib/builtins/arm/aeabi_cdcmp.S
@@ -30,7 +30,7 @@ DEFINE_COMPILERRT_FUNCTION(__aeabi_cdcmpeq)
         push {r0-r3, lr}
         bl __aeabi_cdcmpeq_check_nan
         cmp r0, #1
-#if __ARM_ARCH_ISA_THUMB == 1
+#if defined(USE_THUMB_1)
         beq 1f
         // NaN has been ruled out, so __aeabi_cdcmple can't trap
         mov r0, sp
@@ -46,9 +46,12 @@ DEFINE_COMPILERRT_FUNCTION(__aeabi_cdcmpeq)
         pop {r0-r3, lr}
 
         // NaN has been ruled out, so __aeabi_cdcmple can't trap
+        // Use "it ne" + unconditional branch to guarantee a supported relocation if
+        // __aeabi_cdcmple is in a different section for some builds.
+        IT(ne)
         bne __aeabi_cdcmple
 
-#if defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7EM__)
+#if defined(USE_THUMB_2)
         mov ip, #APSR_C
         msr APSR_nzcvq, ip
 #else
@@ -78,7 +81,7 @@ DEFINE_COMPILERRT_FUNCTION(__aeabi_cdcmple)
 
         bl __aeabi_dcmplt
         cmp r0, #1
-#if __ARM_ARCH_ISA_THUMB == 1
+#if defined(USE_THUMB_1)
         bne 1f
         // Z = 0, C = 0
         movs r0, #1
diff --git a/contrib/compiler-rt/lib/builtins/arm/aeabi_cfcmp.S b/contrib/compiler-rt/lib/builtins/arm/aeabi_cfcmp.S
index 1f304ffd964c..c5fee6b6a08e 100644
--- a/contrib/compiler-rt/lib/builtins/arm/aeabi_cfcmp.S
+++ b/contrib/compiler-rt/lib/builtins/arm/aeabi_cfcmp.S
@@ -30,7 +30,7 @@ DEFINE_COMPILERRT_FUNCTION(__aeabi_cfcmpeq)
         push {r0-r3, lr}
         bl __aeabi_cfcmpeq_check_nan
         cmp r0, #1
-#if __ARM_ARCH_ISA_THUMB == 1
+#if defined(USE_THUMB_1)
         beq 1f
         // NaN has been ruled out, so __aeabi_cfcmple can't trap
         mov r0, sp
@@ -46,9 +46,12 @@ DEFINE_COMPILERRT_FUNCTION(__aeabi_cfcmpeq)
         pop {r0-r3, lr}
 
         // NaN has been ruled out, so __aeabi_cfcmple can't trap
+        // Use "it ne" + unconditional branch to guarantee a supported relocation if
+        // __aeabi_cfcmple is in a different section for some builds.
+        IT(ne)
         bne __aeabi_cfcmple
 
-#if defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7EM__)
+#if defined(USE_THUMB_2)
         mov ip, #APSR_C
         msr APSR_nzcvq, ip
 #else
@@ -78,7 +81,7 @@ DEFINE_COMPILERRT_FUNCTION(__aeabi_cfcmple)
 
         bl __aeabi_fcmplt
         cmp r0, #1
-#if __ARM_ARCH_ISA_THUMB == 1
+#if defined(USE_THUMB_1)
         bne 1f
         // Z = 0, C = 0
         movs r0, #1
diff --git a/contrib/compiler-rt/lib/builtins/arm/aeabi_idivmod.S b/contrib/compiler-rt/lib/builtins/arm/aeabi_idivmod.S
index 4419929f62ac..9c9c80ab5a7b 100644
--- a/contrib/compiler-rt/lib/builtins/arm/aeabi_idivmod.S
+++ b/contrib/compiler-rt/lib/builtins/arm/aeabi_idivmod.S
@@ -20,17 +20,10 @@
 #endif
 
         .syntax unified
-        .syntax unified
         .text
-#if defined(USE_THUMB_PROLOGUE)
-        .thumb
-#endif
+        DEFINE_CODE_STATE
         .p2align 2
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__aeabi_idivmod)
-#else
 DEFINE_COMPILERRT_FUNCTION(__aeabi_idivmod)
-#endif
 #if defined(USE_THUMB_1)
         push    {r0, r1, lr}
         bl      SYMBOL_NAME(__divsi3)
diff --git a/contrib/compiler-rt/lib/builtins/arm/aeabi_memcmp.S b/contrib/compiler-rt/lib/builtins/arm/aeabi_memcmp.S
index 33ea54848b26..e86d6113760e 100644
--- a/contrib/compiler-rt/lib/builtins/arm/aeabi_memcmp.S
+++ b/contrib/compiler-rt/lib/builtins/arm/aeabi_memcmp.S
@@ -14,7 +14,13 @@
         .syntax unified
         .p2align 2
 DEFINE_COMPILERRT_FUNCTION(__aeabi_memcmp)
+#ifdef USE_THUMB_1
+        push    {r7, lr}
+        bl      memcmp
+        pop     {r7, pc}
+#else
         b       memcmp
+#endif
 END_COMPILERRT_FUNCTION(__aeabi_memcmp)
 
 DEFINE_AEABI_FUNCTION_ALIAS(__aeabi_memcmp4, __aeabi_memcmp)
diff --git a/contrib/compiler-rt/lib/builtins/arm/aeabi_memcpy.S b/contrib/compiler-rt/lib/builtins/arm/aeabi_memcpy.S
index eabfa490494c..e83c5fd4dbb3 100644
--- a/contrib/compiler-rt/lib/builtins/arm/aeabi_memcpy.S
+++ b/contrib/compiler-rt/lib/builtins/arm/aeabi_memcpy.S
@@ -14,7 +14,13 @@
         .syntax unified
         .p2align 2
 DEFINE_COMPILERRT_FUNCTION(__aeabi_memcpy)
+#ifdef USE_THUMB_1
+        push    {r7, lr}
+        bl      memcpy
+        pop     {r7, pc}
+#else
         b       memcpy
+#endif
 END_COMPILERRT_FUNCTION(__aeabi_memcpy)
 
 DEFINE_AEABI_FUNCTION_ALIAS(__aeabi_memcpy4, __aeabi_memcpy)
diff --git a/contrib/compiler-rt/lib/builtins/arm/aeabi_memmove.S b/contrib/compiler-rt/lib/builtins/arm/aeabi_memmove.S
index 1bf08c0d5b75..ee28300e46f2 100644
--- a/contrib/compiler-rt/lib/builtins/arm/aeabi_memmove.S
+++ b/contrib/compiler-rt/lib/builtins/arm/aeabi_memmove.S
@@ -13,7 +13,13 @@
 
         .p2align 2
 DEFINE_COMPILERRT_FUNCTION(__aeabi_memmove)
+#ifdef USE_THUMB_1
+        push    {r7, lr}
+        bl      memmove
+        pop     {r7, pc}
+#else
         b       memmove
+#endif
 END_COMPILERRT_FUNCTION(__aeabi_memmove)
 
 DEFINE_AEABI_FUNCTION_ALIAS(__aeabi_memmove4, __aeabi_memmove)
diff --git a/contrib/compiler-rt/lib/builtins/arm/aeabi_memset.S b/contrib/compiler-rt/lib/builtins/arm/aeabi_memset.S
index 633f592279b5..0a678d7627e7 100644
--- a/contrib/compiler-rt/lib/builtins/arm/aeabi_memset.S
+++ b/contrib/compiler-rt/lib/builtins/arm/aeabi_memset.S
@@ -18,16 +18,29 @@ DEFINE_COMPILERRT_FUNCTION(__aeabi_memset)
         mov     r3, r1
         mov     r1, r2
         mov     r2, r3
+#ifdef USE_THUMB_1
+        push    {r7, lr}
+        bl      memset
+        pop     {r7, pc}
+#else
         b       memset
+#endif
 END_COMPILERRT_FUNCTION(__aeabi_memset)
 
 DEFINE_AEABI_FUNCTION_ALIAS(__aeabi_memset4, __aeabi_memset)
 DEFINE_AEABI_FUNCTION_ALIAS(__aeabi_memset8, __aeabi_memset)
 
+        .p2align 2
 DEFINE_COMPILERRT_FUNCTION(__aeabi_memclr)
         mov     r2, r1
         movs    r1, #0
+#ifdef USE_THUMB_1
+        push    {r7, lr}
+        bl      memset
+        pop     {r7, pc}
+#else
         b       memset
+#endif
 END_COMPILERRT_FUNCTION(__aeabi_memclr)
 
 DEFINE_AEABI_FUNCTION_ALIAS(__aeabi_memclr4, __aeabi_memclr)
diff --git a/contrib/compiler-rt/lib/builtins/arm/aeabi_uidivmod.S b/contrib/compiler-rt/lib/builtins/arm/aeabi_uidivmod.S
index 37dae4a105cc..88a4a6d8bc12 100644
--- a/contrib/compiler-rt/lib/builtins/arm/aeabi_uidivmod.S
+++ b/contrib/compiler-rt/lib/builtins/arm/aeabi_uidivmod.S
@@ -22,16 +22,10 @@
 
         .syntax unified
         .text
-#if defined(USE_THUMB_PROLOGUE)
-        .thumb
-#endif
+        DEFINE_CODE_STATE
         .p2align 2
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__aeabi_uidivmod)
-#else
 DEFINE_COMPILERRT_FUNCTION(__aeabi_uidivmod)
-#endif
-#if __ARM_ARCH_ISA_THUMB == 1
+#if defined(USE_THUMB_1)
         cmp     r0, r1
         bcc     LOCAL_LABEL(case_denom_larger)
         push    {r0, r1, lr}
@@ -44,7 +38,7 @@ LOCAL_LABEL(case_denom_larger):
         movs    r1, r0
         movs    r0, #0
         JMP     (lr)
-#else
+#else // defined(USE_THUMB_1)
         push    { lr }
         sub     sp, sp, #4
         mov     r2, sp
diff --git a/contrib/compiler-rt/lib/builtins/arm/bswapdi2.S b/contrib/compiler-rt/lib/builtins/arm/bswapdi2.S
index 4e5a579cf16a..e9db8bac7994 100644
--- a/contrib/compiler-rt/lib/builtins/arm/bswapdi2.S
+++ b/contrib/compiler-rt/lib/builtins/arm/bswapdi2.S
@@ -11,9 +11,7 @@
 
 	.syntax unified
 	.text
-#if defined(USE_THUMB_PROLOGUE)
-	.thumb
-#endif
+	DEFINE_CODE_STATE
 
 //
 // extern uint64_t __bswapdi2(uint64_t);
@@ -21,11 +19,7 @@
 // Reverse all the bytes in a 64-bit integer.
 //
 	.p2align 2
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__bswapdi2)
-#else
 DEFINE_COMPILERRT_FUNCTION(__bswapdi2)
-#endif
 #if __ARM_ARCH < 6
     // before armv6 does not have "rev" instruction
     // r2 = rev(r0)
diff --git a/contrib/compiler-rt/lib/builtins/arm/bswapsi2.S b/contrib/compiler-rt/lib/builtins/arm/bswapsi2.S
index 60342ae0f2fb..1f6eed5c1bbf 100644
--- a/contrib/compiler-rt/lib/builtins/arm/bswapsi2.S
+++ b/contrib/compiler-rt/lib/builtins/arm/bswapsi2.S
@@ -11,9 +11,7 @@
 
 	.syntax unified
 	.text
-#if defined(USE_THUMB_PROLOGUE)
-	.thumb
-#endif
+	DEFINE_CODE_STATE
 
 //
 // extern uint32_t __bswapsi2(uint32_t);
@@ -21,11 +19,7 @@
 // Reverse all the bytes in a 32-bit integer.
 //
 	.p2align 2
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__bswapsi2)
-#else
 DEFINE_COMPILERRT_FUNCTION(__bswapsi2)
-#endif
 #if __ARM_ARCH < 6
     // before armv6 does not have "rev" instruction
  	eor	r1, r0, r0, ror #16
diff --git a/contrib/compiler-rt/lib/builtins/arm/clzdi2.S b/contrib/compiler-rt/lib/builtins/arm/clzdi2.S
index fe56a183f1b1..fc03b385cdfa 100644
--- a/contrib/compiler-rt/lib/builtins/arm/clzdi2.S
+++ b/contrib/compiler-rt/lib/builtins/arm/clzdi2.S
@@ -15,17 +15,10 @@
 
 	.syntax unified
 	.text
-#if defined(USE_THUMB_PROLOGUE)
-	.thumb
-#endif
-
+	DEFINE_CODE_STATE
 
 	.p2align	2
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__clzdi2)
-#else
 DEFINE_COMPILERRT_FUNCTION(__clzdi2)
-#endif
 #ifdef __ARM_FEATURE_CLZ
 #ifdef __ARMEB__
 	cmp	r0, 0
diff --git a/contrib/compiler-rt/lib/builtins/arm/clzsi2.S b/contrib/compiler-rt/lib/builtins/arm/clzsi2.S
index 28d43041eaf2..f2ce59c90119 100644
--- a/contrib/compiler-rt/lib/builtins/arm/clzsi2.S
+++ b/contrib/compiler-rt/lib/builtins/arm/clzsi2.S
@@ -15,16 +15,10 @@
 
 	.syntax unified
 	.text
-#if defined(USE_THUMB_PROLOGUE)
-	.thumb
-#endif
+	DEFINE_CODE_STATE
 
 	.p2align	2
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__clzsi2)
-#else
 DEFINE_COMPILERRT_FUNCTION(__clzsi2)
-#endif
 #ifdef __ARM_FEATURE_CLZ
 	clz	r0, r0
 	JMP(lr)
diff --git a/contrib/compiler-rt/lib/builtins/arm/comparesf2.S b/contrib/compiler-rt/lib/builtins/arm/comparesf2.S
index 1f7031cbf665..c6c4cc067f07 100644
--- a/contrib/compiler-rt/lib/builtins/arm/comparesf2.S
+++ b/contrib/compiler-rt/lib/builtins/arm/comparesf2.S
@@ -40,25 +40,19 @@
 #include "../assembly.h"
     .syntax unified
     .text
-#if defined(USE_THUMB_PROLOGUE)
-    .thumb
-#endif
+    DEFINE_CODE_STATE
 
 @ int __eqsf2(float a, float b)
 
     .p2align 2
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__eqsf2)
-#else
 DEFINE_COMPILERRT_FUNCTION(__eqsf2)
-#endif
 #if defined(COMPILER_RT_ARMHF_TARGET)
     vmov r0, s0
     vmov r1, s1
 #endif
     // Make copies of a and b with the sign bit shifted off the top.  These will
     // be used to detect zeros and NaNs.
-#if __ARM_ARCH_ISA_THUMB == 1
+#if defined(USE_THUMB_1)
     push    {r6, lr}
     lsls    r2,         r0, #1
     lsls    r3,         r1, #1
@@ -150,7 +144,7 @@ DEFINE_COMPILERRT_FUNCTION(__eqsf2)
     // If a == b, then the Z flag is set, so we can get the correct final value
     // into r0 by simply or'ing with 1 if Z is clear.
     // For Thumb-1, r0 contains -1 if a < b, 0 if a > b and 0 if a == b.
-#if defined(USE_THUMB_1)
+#if !defined(USE_THUMB_1)
     it ne
     orrne   r0,     r0, #1
 #endif
@@ -185,11 +179,7 @@ DEFINE_COMPILERRT_FUNCTION_ALIAS(__nesf2, __eqsf2)
 @ int __gtsf2(float a, float b)
 
     .p2align 2
-#if defined(USE_THUMB)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__gtsf2)
-#else
 DEFINE_COMPILERRT_FUNCTION(__gtsf2)
-#endif
     // Identical to the preceding except in that we return -1 for NaN values.
     // Given that the two paths share so much code, one might be tempted to
     // unify them; however, the extra code needed to do so makes the code size
@@ -263,11 +253,7 @@ DEFINE_COMPILERRT_FUNCTION_ALIAS(__gesf2, __gtsf2)
 @ int __unordsf2(float a, float b)
 
     .p2align 2
-#if defined(USE_THUMB)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__unordsf2)
-#else
 DEFINE_COMPILERRT_FUNCTION(__unordsf2)
-#endif
 
 #if defined(COMPILER_RT_ARMHF_TARGET)
     vmov    r0,         s0
diff --git a/contrib/compiler-rt/lib/builtins/arm/divmodsi4.S b/contrib/compiler-rt/lib/builtins/arm/divmodsi4.S
index 3c8359852009..8a027b741efe 100644
--- a/contrib/compiler-rt/lib/builtins/arm/divmodsi4.S
+++ b/contrib/compiler-rt/lib/builtins/arm/divmodsi4.S
@@ -23,20 +23,14 @@
 
 	.syntax unified
 	.text
-#if defined(USE_THUMB_PROLOGUE)
-	.thumb
-#endif
+  DEFINE_CODE_STATE
 
 @ int __divmodsi4(int divident, int divisor, int *remainder)
 @   Calculate the quotient and remainder of the (signed) division.  The return
 @   value is the quotient, the remainder is placed in the variable.
 
 	.p2align 3
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__divmodsi4)
-#else
 DEFINE_COMPILERRT_FUNCTION(__divmodsi4)
-#endif
 #if __ARM_ARCH_EXT_IDIV__
 	tst     r1, r1
 	beq     LOCAL_LABEL(divzero)
diff --git a/contrib/compiler-rt/lib/builtins/arm/divsi3.S b/contrib/compiler-rt/lib/builtins/arm/divsi3.S
index ad96c9e5bfa9..19757af177eb 100644
--- a/contrib/compiler-rt/lib/builtins/arm/divsi3.S
+++ b/contrib/compiler-rt/lib/builtins/arm/divsi3.S
@@ -22,9 +22,7 @@
 
    .syntax unified
    .text
-#if defined(USE_THUMB_PROLOGUE)
-   .thumb
-#endif
+   DEFINE_CODE_STATE
 
 	.p2align 3
 // Ok, APCS and AAPCS agree on 32 bit args, so it's safe to use the same routine.
@@ -33,11 +31,7 @@ DEFINE_AEABI_FUNCTION_ALIAS(__aeabi_idiv, __divsi3)
 @ int __divsi3(int divident, int divisor)
 @   Calculate and return the quotient of the (signed) division.
 
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__divsi3)
-#else
 DEFINE_COMPILERRT_FUNCTION(__divsi3)
-#endif
 #if __ARM_ARCH_EXT_IDIV__
    tst     r1,r1
    beq     LOCAL_LABEL(divzero)
@@ -49,14 +43,14 @@ LOCAL_LABEL(divzero):
 #else
 ESTABLISH_FRAME
 //  Set aside the sign of the quotient.
-#  if __ARM_ARCH_ISA_THUMB == 1
+#  if defined(USE_THUMB_1)
     movs    r4,     r0
     eors    r4,     r1
 #  else
     eor     r4,     r0, r1
 #  endif
 //  Take absolute value of a and b via abs(x) = (x^(x >> 31)) - (x >> 31).
-#  if   __ARM_ARCH_ISA_THUMB == 1
+#  if defined(USE_THUMB_1)
     asrs    r2,     r0, #31
     asrs    r3,     r1, #31
     eors    r0,     r2
diff --git a/contrib/compiler-rt/lib/builtins/arm/modsi3.S b/contrib/compiler-rt/lib/builtins/arm/modsi3.S
index 0986d84945a6..be263834d7f1 100644
--- a/contrib/compiler-rt/lib/builtins/arm/modsi3.S
+++ b/contrib/compiler-rt/lib/builtins/arm/modsi3.S
@@ -22,19 +22,13 @@
 
 	.syntax unified
 	.text
-#if defined(USE_THUMB_PROLOGUE)
-	.thumb
-#endif
+	DEFINE_CODE_STATE
 
 @ int __modsi3(int divident, int divisor)
 @   Calculate and return the remainder of the (signed) division.
 
 	.p2align 3
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__modsi3)
-#else
 DEFINE_COMPILERRT_FUNCTION(__modsi3)
-#endif
 #if __ARM_ARCH_EXT_IDIV__
 	tst     r1, r1
 	beq     LOCAL_LABEL(divzero)
diff --git a/contrib/compiler-rt/lib/builtins/arm/udivmodsi4.S b/contrib/compiler-rt/lib/builtins/arm/udivmodsi4.S
index 860fced63085..ee3950c9b0eb 100644
--- a/contrib/compiler-rt/lib/builtins/arm/udivmodsi4.S
+++ b/contrib/compiler-rt/lib/builtins/arm/udivmodsi4.S
@@ -16,9 +16,7 @@
 
 	.syntax unified
 	.text
-#if defined(USE_THUMB_PROLOGUE)
-	.thumb
-#endif
+	DEFINE_CODE_STATE
 
 @ unsigned int __udivmodsi4(unsigned int divident, unsigned int divisor,
 @                           unsigned int *remainder)
@@ -26,11 +24,7 @@
 @   value is the quotient, the remainder is placed in the variable.
 
 	.p2align 2
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__udivmodsi4)
-#else
 DEFINE_COMPILERRT_FUNCTION(__udivmodsi4)
-#endif
 #if __ARM_ARCH_EXT_IDIV__
 	tst     r1, r1
 	beq     LOCAL_LABEL(divby0)
diff --git a/contrib/compiler-rt/lib/builtins/arm/udivsi3.S b/contrib/compiler-rt/lib/builtins/arm/udivsi3.S
index 82db2cbb395f..6dea27d404ff 100644
--- a/contrib/compiler-rt/lib/builtins/arm/udivsi3.S
+++ b/contrib/compiler-rt/lib/builtins/arm/udivsi3.S
@@ -16,9 +16,8 @@
 
 	.syntax unified
 	.text
-#if defined(USE_THUMB_PROLOGUE)
-	.thumb
-#endif
+
+DEFINE_CODE_STATE
 
 	.p2align 2
 DEFINE_AEABI_FUNCTION_ALIAS(__aeabi_uidiv, __udivsi3)
@@ -26,11 +25,7 @@ DEFINE_AEABI_FUNCTION_ALIAS(__aeabi_uidiv, __udivsi3)
 @ unsigned int __udivsi3(unsigned int divident, unsigned int divisor)
 @   Calculate and return the quotient of the (unsigned) division.
 
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__udivsi3)
-#else
 DEFINE_COMPILERRT_FUNCTION(__udivsi3)
-#endif
 #if __ARM_ARCH_EXT_IDIV__
 	tst     r1, r1
 	beq     LOCAL_LABEL(divby0)
diff --git a/contrib/compiler-rt/lib/builtins/arm/umodsi3.S b/contrib/compiler-rt/lib/builtins/arm/umodsi3.S
index 90f3e8575c77..069fad34cb9c 100644
--- a/contrib/compiler-rt/lib/builtins/arm/umodsi3.S
+++ b/contrib/compiler-rt/lib/builtins/arm/umodsi3.S
@@ -16,19 +16,13 @@
 
 	.syntax unified
 	.text
-#if defined(USE_THUMB_PROLOGUE)
-	.thumb
-#endif
+	DEFINE_CODE_STATE
 
 @ unsigned int __umodsi3(unsigned int divident, unsigned int divisor)
 @   Calculate and return the remainder of the (unsigned) division.
 
 	.p2align 2
-#if defined(USE_THUMB_PROLOGUE)
-DEFINE_COMPILERRT_THUMB_FUNCTION(__umodsi3)
-#else
 DEFINE_COMPILERRT_FUNCTION(__umodsi3)
-#endif
 #if __ARM_ARCH_EXT_IDIV__
 	tst     r1, r1
 	beq     LOCAL_LABEL(divby0)
diff --git a/contrib/compiler-rt/lib/builtins/ashldi3.c b/contrib/compiler-rt/lib/builtins/ashldi3.c
index fcb0abdb1fce..a5c1836006b9 100644
--- a/contrib/compiler-rt/lib/builtins/ashldi3.c
+++ b/contrib/compiler-rt/lib/builtins/ashldi3.c
@@ -41,8 +41,5 @@ __ashldi3(di_int a, si_int b)
 }
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI di_int __aeabi_llsl(di_int a, si_int b) {
-  return __ashldi3(a, b);
-}
+AEABI_RTABI di_int __aeabi_llsl(di_int a, si_int b) COMPILER_RT_ALIAS(__ashldi3);
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/ashrdi3.c b/contrib/compiler-rt/lib/builtins/ashrdi3.c
index b4ab4c617ba0..84619965eca0 100644
--- a/contrib/compiler-rt/lib/builtins/ashrdi3.c
+++ b/contrib/compiler-rt/lib/builtins/ashrdi3.c
@@ -42,8 +42,5 @@ __ashrdi3(di_int a, si_int b)
 }
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI di_int __aeabi_lasr(di_int a, si_int b) {
-  return __ashrdi3(a, b);
-}
+AEABI_RTABI di_int __aeabi_lasr(di_int a, si_int b) COMPILER_RT_ALIAS(__ashrdi3);
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/assembly.h b/contrib/compiler-rt/lib/builtins/assembly.h
index 4893d483fd82..3f5e59b25442 100644
--- a/contrib/compiler-rt/lib/builtins/assembly.h
+++ b/contrib/compiler-rt/lib/builtins/assembly.h
@@ -68,10 +68,42 @@
 #endif
 
 #if defined(__arm__)
+
+/*
+ * Determine actual [ARM][THUMB[1][2]] ISA using compiler predefined macros:
+ * - for '-mthumb -march=armv6' compiler defines '__thumb__'
+ * - for '-mthumb -march=armv7' compiler defines '__thumb__' and '__thumb2__'
+ */
+#if defined(__thumb2__) || defined(__thumb__)
+#define DEFINE_CODE_STATE .thumb SEPARATOR
+#define DECLARE_FUNC_ENCODING    .thumb_func SEPARATOR
+#if defined(__thumb2__)
+#define USE_THUMB_2
+#define IT(cond)  it cond
+#define ITT(cond) itt cond
+#define ITE(cond) ite cond
+#else
+#define USE_THUMB_1
+#define IT(cond)
+#define ITT(cond)
+#define ITE(cond)
+#endif // defined(__thumb__2)
+#else // !defined(__thumb2__) && !defined(__thumb__)
+#define DEFINE_CODE_STATE .arm SEPARATOR
+#define DECLARE_FUNC_ENCODING
+#define IT(cond)
+#define ITT(cond)
+#define ITE(cond)
+#endif
+
+#if defined(USE_THUMB_1) && defined(USE_THUMB_2)
+#error "USE_THUMB_1 and USE_THUMB_2 can't be defined together."
+#endif
+
 #if defined(__ARM_ARCH_4T__) || __ARM_ARCH >= 5
 #define ARM_HAS_BX
 #endif
-#if !defined(__ARM_FEATURE_CLZ) && __ARM_ARCH_ISA_THUMB != 1 &&                \
+#if !defined(__ARM_FEATURE_CLZ) && !defined(USE_THUMB_1) &&  \
     (__ARM_ARCH >= 6 || (__ARM_ARCH == 5 && !defined(__ARM_ARCH_5__)))
 #define __ARM_FEATURE_CLZ
 #endif
@@ -93,37 +125,15 @@
   JMP(ip)
 #endif
 
-/*
- * Determine actual [ARM][THUMB[1][2]] ISA using compiler predefined macros:
- * - for '-mthumb -march=armv6' compiler defines '__thumb__'
- * - for '-mthumb -march=armv7' compiler defines '__thumb__' and '__thumb2__'
- */
-#if defined(__thumb2__)
-#define USE_THUMB_2 1
-#elif defined(__thumb__)
-#define USE_THUMB_1 1
-#endif
-
-#if defined(USE_THUMB_1) && defined(USE_THUMB_2)
-#error "USE_THUMB_1 and USE_THUMB_2 can't be defined together."
-#endif
-
-#if defined(USE_THUMB_1) || defined(USE_THUMB_1)
-#define USE_THUMB_PROLOGUE 1
-#endif
-
 #if defined(USE_THUMB_2)
-#define IT(cond)  it cond
-#define ITT(cond) itt cond
-#define ITE(cond) ite cond
 #define WIDE(op) op.w
 #else
-#define IT(cond)
-#define ITT(cond)
-#define ITE(cond)
 #define WIDE(op) op
 #endif
-#endif /* defined(__arm__) */
+#else // !defined(__arm)
+#define DECLARE_FUNC_ENCODING
+#define DEFINE_CODE_STATE
+#endif
 
 #define GLUE2(a, b) a##b
 #define GLUE(a, b) GLUE2(a, b)
@@ -137,13 +147,16 @@
 #endif
 
 #define DEFINE_COMPILERRT_FUNCTION(name)                                       \
+  DEFINE_CODE_STATE                                                            \
   FILE_LEVEL_DIRECTIVE SEPARATOR                                               \
   .globl SYMBOL_NAME(name) SEPARATOR                                           \
   SYMBOL_IS_FUNC(SYMBOL_NAME(name)) SEPARATOR                                  \
   DECLARE_SYMBOL_VISIBILITY(name)                                              \
+  DECLARE_FUNC_ENCODING                                                        \
   SYMBOL_NAME(name):
 
 #define DEFINE_COMPILERRT_THUMB_FUNCTION(name)                                 \
+  DEFINE_CODE_STATE                                                            \
   FILE_LEVEL_DIRECTIVE SEPARATOR                                               \
   .globl SYMBOL_NAME(name) SEPARATOR                                           \
   SYMBOL_IS_FUNC(SYMBOL_NAME(name)) SEPARATOR                                  \
@@ -152,16 +165,20 @@
   SYMBOL_NAME(name):
 
 #define DEFINE_COMPILERRT_PRIVATE_FUNCTION(name)                               \
+  DEFINE_CODE_STATE                                                            \
   FILE_LEVEL_DIRECTIVE SEPARATOR                                               \
   .globl SYMBOL_NAME(name) SEPARATOR                                           \
   SYMBOL_IS_FUNC(SYMBOL_NAME(name)) SEPARATOR                                  \
   HIDDEN(SYMBOL_NAME(name)) SEPARATOR                                          \
+  DECLARE_FUNC_ENCODING                                                        \
   SYMBOL_NAME(name):
 
 #define DEFINE_COMPILERRT_PRIVATE_FUNCTION_UNMANGLED(name)                     \
+  DEFINE_CODE_STATE                                                            \
   .globl name SEPARATOR                                                        \
   SYMBOL_IS_FUNC(name) SEPARATOR                                               \
   HIDDEN(name) SEPARATOR                                                       \
+  DECLARE_FUNC_ENCODING                                                        \
   name:
 
 #define DEFINE_COMPILERRT_FUNCTION_ALIAS(name, target)                         \
diff --git a/contrib/compiler-rt/lib/builtins/clear_cache.c b/contrib/compiler-rt/lib/builtins/clear_cache.c
index 7ccbe019dba9..4a01cb46d4ac 100644
--- a/contrib/compiler-rt/lib/builtins/clear_cache.c
+++ b/contrib/compiler-rt/lib/builtins/clear_cache.c
@@ -9,6 +9,7 @@
  */
 
 #include "int_lib.h"
+#include <assert.h>
 #include <stddef.h>
 
 #if __APPLE__
@@ -23,7 +24,7 @@ uint32_t FlushInstructionCache(uintptr_t hProcess, void *lpBaseAddress,
 uintptr_t GetCurrentProcess(void);
 #endif
 
-#if (defined(__FreeBSD__) || defined(__Bitrig__)) && defined(__arm__)
+#if defined(__FreeBSD__) && defined(__arm__)
   #include <sys/types.h>
   #include <machine/sysarch.h>
 #endif
@@ -32,7 +33,7 @@ uintptr_t GetCurrentProcess(void);
   #include <machine/sysarch.h>
 #endif
 
-#if defined(__mips__) && !defined(__FreeBSD__)
+#if defined(__linux__) && defined(__mips__)
   #include <sys/cachectl.h>
   #include <sys/syscall.h>
   #include <unistd.h>
@@ -41,7 +42,7 @@ uintptr_t GetCurrentProcess(void);
      * clear_mips_cache - Invalidates instruction cache for Mips.
      */
     static void clear_mips_cache(const void* Addr, size_t Size) {
-      asm volatile (
+      __asm__ volatile (
         ".set push\n"
         ".set noreorder\n"
         ".set noat\n"
@@ -96,7 +97,7 @@ void __clear_cache(void *start, void *end) {
  * so there is nothing to do
  */
 #elif defined(__arm__) && !defined(__APPLE__)
-    #if defined(__FreeBSD__) || defined(__NetBSD__) || defined(__Bitrig__)
+    #if defined(__FreeBSD__) || defined(__NetBSD__)
         struct arm_sync_icache_args arg;
 
         arg.addr = (uintptr_t)start;
@@ -121,15 +122,13 @@ void __clear_cache(void *start, void *end) {
                           : "=r"(start_reg)
                           : "r"(syscall_nr), "r"(start_reg), "r"(end_reg),
                             "r"(flags));
-         if (start_reg != 0) {
-             compilerrt_abort();
-         }
+         assert(start_reg == 0 && "Cache flush syscall failed.");
     #elif defined(_WIN32)
         FlushInstructionCache(GetCurrentProcess(), start, end - start);
     #else
         compilerrt_abort();
     #endif
-#elif defined(__mips__) && !defined(__FreeBSD__)
+#elif defined(__linux__) && defined(__mips__)
   const uintptr_t start_int = (uintptr_t) start;
   const uintptr_t end_int = (uintptr_t) end;
     #if defined(__ANDROID__) && defined(__LP64__)
@@ -165,6 +164,21 @@ void __clear_cache(void *start, void *end) {
   for (addr = xstart; addr < xend; addr += icache_line_size)
     __asm __volatile("ic ivau, %0" :: "r"(addr));
   __asm __volatile("isb sy");
+#elif defined (__powerpc64__)
+  const size_t line_size = 32;
+  const size_t len = (uintptr_t)end - (uintptr_t)start;
+
+  const uintptr_t mask = ~(line_size - 1);
+  const uintptr_t start_line = ((uintptr_t)start) & mask;
+  const uintptr_t end_line = ((uintptr_t)start + len + line_size - 1) & mask;
+
+  for (uintptr_t line = start_line; line < end_line; line += line_size)
+    __asm__ volatile("dcbf 0, %0" : : "r"(line));
+  __asm__ volatile("sync");
+
+  for (uintptr_t line = start_line; line < end_line; line += line_size)
+    __asm__ volatile("icbi 0, %0" : : "r"(line));
+  __asm__ volatile("isync");
 #else
     #if __APPLE__
         /* On Darwin, sys_icache_invalidate() provides this functionality */
@@ -174,4 +188,3 @@ void __clear_cache(void *start, void *end) {
     #endif
 #endif
 }
-
diff --git a/contrib/compiler-rt/lib/builtins/comparedf2.c b/contrib/compiler-rt/lib/builtins/comparedf2.c
index c5bb169d0021..44e5d2b288a6 100644
--- a/contrib/compiler-rt/lib/builtins/comparedf2.c
+++ b/contrib/compiler-rt/lib/builtins/comparedf2.c
@@ -143,8 +143,11 @@ __gtdf2(fp_t a, fp_t b) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI int __aeabi_dcmpun(fp_t a, fp_t b) {
   return __unorddf2(a, b);
 }
+#else
+AEABI_RTABI int __aeabi_dcmpun(fp_t a, fp_t b) COMPILER_RT_ALIAS(__unorddf2);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/comparesf2.c b/contrib/compiler-rt/lib/builtins/comparesf2.c
index 4badb5e1b9f7..43cd6a6a7003 100644
--- a/contrib/compiler-rt/lib/builtins/comparesf2.c
+++ b/contrib/compiler-rt/lib/builtins/comparesf2.c
@@ -143,8 +143,11 @@ __gtsf2(fp_t a, fp_t b) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI int __aeabi_fcmpun(fp_t a, fp_t b) {
   return __unordsf2(a, b);
 }
+#else
+AEABI_RTABI int __aeabi_fcmpun(fp_t a, fp_t b) COMPILER_RT_ALIAS(__unordsf2);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/cpu_model.c b/contrib/compiler-rt/lib/builtins/cpu_model.c
index 83ea7a49faf7..4c96e9cd85d5 100644
--- a/contrib/compiler-rt/lib/builtins/cpu_model.c
+++ b/contrib/compiler-rt/lib/builtins/cpu_model.c
@@ -54,6 +54,7 @@ enum ProcessorTypes {
   AMD_BTVER1,
   AMD_BTVER2,
   AMDFAM17H,
+  INTEL_KNM,
   CPU_TYPE_MAX
 };
 
@@ -74,6 +75,7 @@ enum ProcessorSubtypes {
   INTEL_COREI7_BROADWELL,
   INTEL_COREI7_SKYLAKE,
   INTEL_COREI7_SKYLAKE_AVX512,
+  INTEL_COREI7_CANNONLAKE,
   CPU_SUBTYPE_MAX
 };
 
@@ -339,6 +341,12 @@ getIntelProcessorTypeAndSubtype(unsigned Family, unsigned Model,
       *Subtype = INTEL_COREI7_SKYLAKE_AVX512; // "skylake-avx512"
       break;
 
+    // Cannonlake:
+    case 0x66:
+      *Type = INTEL_COREI7;
+      *Subtype = INTEL_COREI7_CANNONLAKE; // "cannonlake"
+      break;
+
     case 0x1c: // Most 45 nm Intel Atom processors
     case 0x26: // 45 nm Atom Lincroft
     case 0x27: // 32 nm Atom Medfield
@@ -361,6 +369,10 @@ getIntelProcessorTypeAndSubtype(unsigned Family, unsigned Model,
       *Type = INTEL_KNL; // knl
       break;
 
+    case 0x85:
+      *Type = INTEL_KNM; // knm
+      break;
+
     default: // Unknown family 6 CPU.
       break;
     break;
diff --git a/contrib/compiler-rt/lib/builtins/divdf3.c b/contrib/compiler-rt/lib/builtins/divdf3.c
index 492e32b851e9..04a4dc5571ca 100644
--- a/contrib/compiler-rt/lib/builtins/divdf3.c
+++ b/contrib/compiler-rt/lib/builtins/divdf3.c
@@ -183,8 +183,11 @@ __divdf3(fp_t a, fp_t b) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_ddiv(fp_t a, fp_t b) {
   return __divdf3(a, b);
 }
+#else
+AEABI_RTABI fp_t __aeabi_ddiv(fp_t a, fp_t b) COMPILER_RT_ALIAS(__divdf3);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/divsf3.c b/contrib/compiler-rt/lib/builtins/divsf3.c
index aa6289a6d70a..65294d70fc61 100644
--- a/contrib/compiler-rt/lib/builtins/divsf3.c
+++ b/contrib/compiler-rt/lib/builtins/divsf3.c
@@ -167,8 +167,11 @@ __divsf3(fp_t a, fp_t b) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_fdiv(fp_t a, fp_t b) {
   return __divsf3(a, b);
 }
+#else
+AEABI_RTABI fp_t __aeabi_fdiv(fp_t a, fp_t b) COMPILER_RT_ALIAS(__divsf3);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/divsi3.c b/contrib/compiler-rt/lib/builtins/divsi3.c
index 3852e3990b5b..75aea008ddc1 100644
--- a/contrib/compiler-rt/lib/builtins/divsi3.c
+++ b/contrib/compiler-rt/lib/builtins/divsi3.c
@@ -35,8 +35,5 @@ __divsi3(si_int a, si_int b)
 }
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI si_int __aeabi_idiv(si_int a, si_int b) {
-  return __divsi3(a, b);
-}
+AEABI_RTABI si_int __aeabi_idiv(si_int a, si_int b) COMPILER_RT_ALIAS(__divsi3);
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/emutls.c b/contrib/compiler-rt/lib/builtins/emutls.c
index 12aad3a42b76..5dd8dd154771 100644
--- a/contrib/compiler-rt/lib/builtins/emutls.c
+++ b/contrib/compiler-rt/lib/builtins/emutls.c
@@ -102,7 +102,6 @@ static __inline emutls_address_array* emutls_getspecific() {
 #include <malloc.h>
 #include <stdio.h>
 #include <assert.h>
-#include <immintrin.h>
 
 static LPCRITICAL_SECTION emutls_mutex;
 static DWORD emutls_tls_index = TLS_OUT_OF_INDEXES;
@@ -203,25 +202,24 @@ static __inline emutls_address_array* emutls_getspecific() {
 /* Provide atomic load/store functions for emutls_get_index if built with MSVC.
  */
 #if !defined(__ATOMIC_RELEASE)
+#include <intrin.h>
 
 enum { __ATOMIC_ACQUIRE = 2, __ATOMIC_RELEASE = 3 };
 
 static __inline uintptr_t __atomic_load_n(void *ptr, unsigned type) {
     assert(type == __ATOMIC_ACQUIRE);
+    // These return the previous value - but since we do an OR with 0,
+    // it's equivalent to a plain load.
 #ifdef _WIN64
-    return (uintptr_t) _load_be_u64(ptr);
+    return InterlockedOr64(ptr, 0);
 #else
-    return (uintptr_t) _load_be_u32(ptr);
+    return InterlockedOr(ptr, 0);
 #endif
 }
 
 static __inline void __atomic_store_n(void *ptr, uintptr_t val, unsigned type) {
     assert(type == __ATOMIC_RELEASE);
-#ifdef _WIN64
-    _store_be_u64(ptr, val);
-#else
-    _store_be_u32(ptr, val);
-#endif
+    InterlockedExchangePointer((void *volatile *)ptr, (void *)val);
 }
 
 #endif
diff --git a/contrib/compiler-rt/lib/builtins/enable_execute_stack.c b/contrib/compiler-rt/lib/builtins/enable_execute_stack.c
index 0dc3482c4467..327d460b4253 100644
--- a/contrib/compiler-rt/lib/builtins/enable_execute_stack.c
+++ b/contrib/compiler-rt/lib/builtins/enable_execute_stack.c
@@ -22,7 +22,7 @@
 
 #ifdef _WIN32
 #define WIN32_LEAN_AND_MEAN
-#include <Windows.h>
+#include <windows.h>
 #else
 #ifndef __APPLE__
 #include <unistd.h>
diff --git a/contrib/compiler-rt/lib/builtins/extendhfsf2.c b/contrib/compiler-rt/lib/builtins/extendhfsf2.c
index e7d9fde8abfc..d9c0db84b0ce 100644
--- a/contrib/compiler-rt/lib/builtins/extendhfsf2.c
+++ b/contrib/compiler-rt/lib/builtins/extendhfsf2.c
@@ -23,8 +23,11 @@ COMPILER_RT_ABI float __gnu_h2f_ieee(uint16_t a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI float __aeabi_h2f(uint16_t a) {
   return __extendhfsf2(a);
 }
+#else
+AEABI_RTABI float __aeabi_h2f(uint16_t a) COMPILER_RT_ALIAS(__extendhfsf2);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/extendsfdf2.c b/contrib/compiler-rt/lib/builtins/extendsfdf2.c
index b9e7a7471a98..3d84529a6c53 100644
--- a/contrib/compiler-rt/lib/builtins/extendsfdf2.c
+++ b/contrib/compiler-rt/lib/builtins/extendsfdf2.c
@@ -17,8 +17,11 @@ COMPILER_RT_ABI double __extendsfdf2(float a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI double __aeabi_f2d(float a) {
   return __extendsfdf2(a);
 }
+#else
+AEABI_RTABI double __aeabi_f2d(float a) COMPILER_RT_ALIAS(__extendsfdf2);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/fixdfdi.c b/contrib/compiler-rt/lib/builtins/fixdfdi.c
index 31d76df28255..54e312d3c8f7 100644
--- a/contrib/compiler-rt/lib/builtins/fixdfdi.c
+++ b/contrib/compiler-rt/lib/builtins/fixdfdi.c
@@ -45,13 +45,11 @@ __fixdfdi(fp_t a) {
 #endif
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI di_int
-#if defined(__SOFT_FP__)
-__aeabi_d2lz(fp_t a) {
-#else
-__aeabi_d2lz(double a) {
-#endif
+#if defined(COMPILER_RT_ARMHF_TARGET)
+AEABI_RTABI di_int __aeabi_d2lz(fp_t a) {
   return __fixdfdi(a);
 }
+#else
+AEABI_RTABI di_int __aeabi_d2lz(fp_t a) COMPILER_RT_ALIAS(__fixdfdi);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/fixdfsi.c b/contrib/compiler-rt/lib/builtins/fixdfsi.c
index fc316dcd0545..5b9588175717 100644
--- a/contrib/compiler-rt/lib/builtins/fixdfsi.c
+++ b/contrib/compiler-rt/lib/builtins/fixdfsi.c
@@ -20,8 +20,11 @@ __fixdfsi(fp_t a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI si_int __aeabi_d2iz(fp_t a) {
   return __fixdfsi(a);
 }
+#else
+AEABI_RTABI si_int __aeabi_d2iz(fp_t a) COMPILER_RT_ALIAS(__fixdfsi);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/fixsfdi.c b/contrib/compiler-rt/lib/builtins/fixsfdi.c
index c43473637d60..32e87c60889f 100644
--- a/contrib/compiler-rt/lib/builtins/fixsfdi.c
+++ b/contrib/compiler-rt/lib/builtins/fixsfdi.c
@@ -45,13 +45,11 @@ __fixsfdi(fp_t a) {
 #endif
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI di_int
-#if defined(__SOFT_FP__)
-__aeabi_f2lz(fp_t a) {
-#else
-__aeabi_f2lz(float a) {
-#endif
+#if defined(COMPILER_RT_ARMHF_TARGET)
+AEABI_RTABI di_int __aeabi_f2lz(fp_t a) {
   return __fixsfdi(a);
 }
+#else
+AEABI_RTABI di_int __aeabi_f2lz(fp_t a) COMPILER_RT_ALIAS(__fixsfdi);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/fixsfsi.c b/contrib/compiler-rt/lib/builtins/fixsfsi.c
index 3276df966460..e94e5f3dcd68 100644
--- a/contrib/compiler-rt/lib/builtins/fixsfsi.c
+++ b/contrib/compiler-rt/lib/builtins/fixsfsi.c
@@ -20,8 +20,11 @@ __fixsfsi(fp_t a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI si_int __aeabi_f2iz(fp_t a) {
   return __fixsfsi(a);
 }
+#else
+AEABI_RTABI si_int __aeabi_f2iz(fp_t a) COMPILER_RT_ALIAS(__fixsfsi);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/fixunsdfdi.c b/contrib/compiler-rt/lib/builtins/fixunsdfdi.c
index b734409709bf..bfe4dbb25656 100644
--- a/contrib/compiler-rt/lib/builtins/fixunsdfdi.c
+++ b/contrib/compiler-rt/lib/builtins/fixunsdfdi.c
@@ -42,13 +42,11 @@ __fixunsdfdi(fp_t a) {
 #endif
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI du_int
-#if defined(__SOFT_FP__)
-__aeabi_d2ulz(fp_t a) {
-#else
-__aeabi_d2ulz(double a) {
-#endif
+#if defined(COMPILER_RT_ARMHF_TARGET)
+AEABI_RTABI du_int __aeabi_d2ulz(fp_t a) {
   return __fixunsdfdi(a);
 }
+#else
+AEABI_RTABI du_int __aeabi_d2ulz(fp_t a) COMPILER_RT_ALIAS(__fixunsdfdi);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/fixunsdfsi.c b/contrib/compiler-rt/lib/builtins/fixunsdfsi.c
index bb3d8e0f831b..3c5355beae1a 100644
--- a/contrib/compiler-rt/lib/builtins/fixunsdfsi.c
+++ b/contrib/compiler-rt/lib/builtins/fixunsdfsi.c
@@ -19,8 +19,11 @@ __fixunsdfsi(fp_t a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI su_int __aeabi_d2uiz(fp_t a) {
   return __fixunsdfsi(a);
 }
+#else
+AEABI_RTABI su_int __aeabi_d2uiz(fp_t a) COMPILER_RT_ALIAS(__fixunsdfsi);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/fixunssfdi.c b/contrib/compiler-rt/lib/builtins/fixunssfdi.c
index 5d92245df0d9..080a25bb1e99 100644
--- a/contrib/compiler-rt/lib/builtins/fixunssfdi.c
+++ b/contrib/compiler-rt/lib/builtins/fixunssfdi.c
@@ -43,13 +43,11 @@ __fixunssfdi(fp_t a) {
 #endif
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI du_int
-#if defined(__SOFT_FP__)
-__aeabi_f2ulz(fp_t a) {
-#else
-__aeabi_f2ulz(float a) {
-#endif
+#if defined(COMPILER_RT_ARMHF_TARGET)
+AEABI_RTABI du_int __aeabi_f2ulz(fp_t a) {
   return __fixunssfdi(a);
 }
+#else
+AEABI_RTABI du_int __aeabi_f2ulz(fp_t a) COMPILER_RT_ALIAS(__fixunssfdi);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/fixunssfsi.c b/contrib/compiler-rt/lib/builtins/fixunssfsi.c
index 91d5e8ae5d7f..eca2916a5c88 100644
--- a/contrib/compiler-rt/lib/builtins/fixunssfsi.c
+++ b/contrib/compiler-rt/lib/builtins/fixunssfsi.c
@@ -23,8 +23,11 @@ __fixunssfsi(fp_t a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI su_int __aeabi_f2uiz(fp_t a) {
   return __fixunssfsi(a);
 }
+#else
+AEABI_RTABI su_int __aeabi_f2uiz(fp_t a) COMPILER_RT_ALIAS(__fixunssfsi);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/floatdidf.c b/contrib/compiler-rt/lib/builtins/floatdidf.c
index 681fecef9682..36b856e078d4 100644
--- a/contrib/compiler-rt/lib/builtins/floatdidf.c
+++ b/contrib/compiler-rt/lib/builtins/floatdidf.c
@@ -105,8 +105,11 @@ __floatdidf(di_int a)
 #endif
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI double __aeabi_l2d(di_int a) {
   return __floatdidf(a);
 }
+#else
+AEABI_RTABI double __aeabi_l2d(di_int a) COMPILER_RT_ALIAS(__floatdidf);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/floatdisf.c b/contrib/compiler-rt/lib/builtins/floatdisf.c
index dd548165c373..a2f09eb2ed2c 100644
--- a/contrib/compiler-rt/lib/builtins/floatdisf.c
+++ b/contrib/compiler-rt/lib/builtins/floatdisf.c
@@ -78,8 +78,11 @@ __floatdisf(di_int a)
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI float __aeabi_l2f(di_int a) {
   return __floatdisf(a);
 }
+#else
+AEABI_RTABI float __aeabi_l2f(di_int a) COMPILER_RT_ALIAS(__floatdisf);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/floatsidf.c b/contrib/compiler-rt/lib/builtins/floatsidf.c
index 2ae395bdc1db..fe051123ce7c 100644
--- a/contrib/compiler-rt/lib/builtins/floatsidf.c
+++ b/contrib/compiler-rt/lib/builtins/floatsidf.c
@@ -51,8 +51,11 @@ __floatsidf(int a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_i2d(int a) {
   return __floatsidf(a);
 }
+#else
+AEABI_RTABI fp_t __aeabi_i2d(int a) COMPILER_RT_ALIAS(__floatsidf);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/floatsisf.c b/contrib/compiler-rt/lib/builtins/floatsisf.c
index 08891fcdf201..bf087ee3c295 100644
--- a/contrib/compiler-rt/lib/builtins/floatsisf.c
+++ b/contrib/compiler-rt/lib/builtins/floatsisf.c
@@ -57,8 +57,11 @@ __floatsisf(int a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_i2f(int a) {
   return __floatsisf(a);
 }
+#else
+AEABI_RTABI fp_t __aeabi_i2f(int a) COMPILER_RT_ALIAS(__floatsisf);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/floatundidf.c b/contrib/compiler-rt/lib/builtins/floatundidf.c
index 6c1a931ef2f3..8bc2a096324f 100644
--- a/contrib/compiler-rt/lib/builtins/floatundidf.c
+++ b/contrib/compiler-rt/lib/builtins/floatundidf.c
@@ -104,8 +104,11 @@ __floatundidf(du_int a)
 #endif
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI double __aeabi_ul2d(du_int a) {
   return __floatundidf(a);
 }
+#else
+AEABI_RTABI double __aeabi_ul2d(du_int a) COMPILER_RT_ALIAS(__floatundidf);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/floatundisf.c b/contrib/compiler-rt/lib/builtins/floatundisf.c
index 86841a75dc66..844786ea7777 100644
--- a/contrib/compiler-rt/lib/builtins/floatundisf.c
+++ b/contrib/compiler-rt/lib/builtins/floatundisf.c
@@ -75,8 +75,11 @@ __floatundisf(du_int a)
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI float __aeabi_ul2f(du_int a) {
   return __floatundisf(a);
 }
+#else
+AEABI_RTABI float __aeabi_ul2f(du_int a) COMPILER_RT_ALIAS(__floatundisf);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/floatunsidf.c b/contrib/compiler-rt/lib/builtins/floatunsidf.c
index 8d4807194f0b..75cf6b9177df 100644
--- a/contrib/compiler-rt/lib/builtins/floatunsidf.c
+++ b/contrib/compiler-rt/lib/builtins/floatunsidf.c
@@ -40,8 +40,11 @@ __floatunsidf(unsigned int a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_ui2d(unsigned int a) {
   return __floatunsidf(a);
 }
+#else
+AEABI_RTABI fp_t __aeabi_ui2d(unsigned int a) COMPILER_RT_ALIAS(__floatunsidf);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/floatunsisf.c b/contrib/compiler-rt/lib/builtins/floatunsisf.c
index f194c046d2fb..29525ccedbbe 100644
--- a/contrib/compiler-rt/lib/builtins/floatunsisf.c
+++ b/contrib/compiler-rt/lib/builtins/floatunsisf.c
@@ -48,8 +48,11 @@ __floatunsisf(unsigned int a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_ui2f(unsigned int a) {
   return __floatunsisf(a);
 }
+#else
+AEABI_RTABI fp_t __aeabi_ui2f(unsigned int a) COMPILER_RT_ALIAS(__floatunsisf);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/int_endianness.h b/contrib/compiler-rt/lib/builtins/int_endianness.h
index 7995ddbb953c..e2586c56bac8 100644
--- a/contrib/compiler-rt/lib/builtins/int_endianness.h
+++ b/contrib/compiler-rt/lib/builtins/int_endianness.h
@@ -61,7 +61,7 @@
 
 #endif /* *BSD */
 
-#if defined(__OpenBSD__) || defined(__Bitrig__)
+#if defined(__OpenBSD__)
 #include <machine/endian.h>
 
 #if _BYTE_ORDER == _BIG_ENDIAN
@@ -72,7 +72,7 @@
 #define _YUGA_BIG_ENDIAN    0
 #endif /* _BYTE_ORDER */
 
-#endif /* OpenBSD and Bitrig. */
+#endif /* OpenBSD */
 
 /* .. */
 
diff --git a/contrib/compiler-rt/lib/builtins/int_lib.h b/contrib/compiler-rt/lib/builtins/int_lib.h
index 53caf3a53e63..93394cebf7ea 100644
--- a/contrib/compiler-rt/lib/builtins/int_lib.h
+++ b/contrib/compiler-rt/lib/builtins/int_lib.h
@@ -22,9 +22,11 @@
 
 #if defined(__ELF__)
 #define FNALIAS(alias_name, original_name) \
-  void alias_name() __attribute__((alias(#original_name)))
+  void alias_name() __attribute__((__alias__(#original_name)))
+#define COMPILER_RT_ALIAS(aliasee) __attribute__((__alias__(#aliasee)))
 #else
 #define FNALIAS(alias, name) _Pragma("GCC error(\"alias unsupported on this file format\")")
+#define COMPILER_RT_ALIAS(aliasee) _Pragma("GCC error(\"alias unsupported on this file format\")")
 #endif
 
 /* ABI macro definitions */
diff --git a/contrib/compiler-rt/lib/builtins/lshrdi3.c b/contrib/compiler-rt/lib/builtins/lshrdi3.c
index becbbef4eb09..67b2a7668345 100644
--- a/contrib/compiler-rt/lib/builtins/lshrdi3.c
+++ b/contrib/compiler-rt/lib/builtins/lshrdi3.c
@@ -41,8 +41,5 @@ __lshrdi3(di_int a, si_int b)
 }
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI di_int __aeabi_llsr(di_int a, si_int b) {
-  return __lshrdi3(a, b);
-}
+AEABI_RTABI di_int __aeabi_llsr(di_int a, si_int b) COMPILER_RT_ALIAS(__lshrdi3);
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/muldf3.c b/contrib/compiler-rt/lib/builtins/muldf3.c
index 59a60190eba3..1bb103e38c13 100644
--- a/contrib/compiler-rt/lib/builtins/muldf3.c
+++ b/contrib/compiler-rt/lib/builtins/muldf3.c
@@ -20,8 +20,11 @@ COMPILER_RT_ABI fp_t __muldf3(fp_t a, fp_t b) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_dmul(fp_t a, fp_t b) {
   return __muldf3(a, b);
 }
+#else
+AEABI_RTABI fp_t __aeabi_dmul(fp_t a, fp_t b) COMPILER_RT_ALIAS(__muldf3);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/muldi3.c b/contrib/compiler-rt/lib/builtins/muldi3.c
index 6818a9e2f722..a187315e9165 100644
--- a/contrib/compiler-rt/lib/builtins/muldi3.c
+++ b/contrib/compiler-rt/lib/builtins/muldi3.c
@@ -54,8 +54,5 @@ __muldi3(di_int a, di_int b)
 }
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI di_int __aeabi_lmul(di_int a, di_int b) {
-  return __muldi3(a, b);
-}
+AEABI_RTABI di_int __aeabi_lmul(di_int a, di_int b) COMPILER_RT_ALIAS(__muldi3);
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/mulsf3.c b/contrib/compiler-rt/lib/builtins/mulsf3.c
index f141af1acc58..1e2cf3e717c9 100644
--- a/contrib/compiler-rt/lib/builtins/mulsf3.c
+++ b/contrib/compiler-rt/lib/builtins/mulsf3.c
@@ -20,8 +20,11 @@ COMPILER_RT_ABI fp_t __mulsf3(fp_t a, fp_t b) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_fmul(fp_t a, fp_t b) {
   return __mulsf3(a, b);
 }
+#else
+AEABI_RTABI fp_t __aeabi_fmul(fp_t a, fp_t b) COMPILER_RT_ALIAS(__mulsf3);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/negdf2.c b/contrib/compiler-rt/lib/builtins/negdf2.c
index 5e2544cdb4be..f0bfaad24743 100644
--- a/contrib/compiler-rt/lib/builtins/negdf2.c
+++ b/contrib/compiler-rt/lib/builtins/negdf2.c
@@ -20,8 +20,11 @@ __negdf2(fp_t a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_dneg(fp_t a) {
   return __negdf2(a);
 }
+#else
+AEABI_RTABI fp_t __aeabi_dneg(fp_t a) COMPILER_RT_ALIAS(__negdf2);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/negsf2.c b/contrib/compiler-rt/lib/builtins/negsf2.c
index f90b34335680..05c97d4d5a11 100644
--- a/contrib/compiler-rt/lib/builtins/negsf2.c
+++ b/contrib/compiler-rt/lib/builtins/negsf2.c
@@ -20,8 +20,11 @@ __negsf2(fp_t a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_fneg(fp_t a) {
   return __negsf2(a);
 }
+#else
+AEABI_RTABI fp_t __aeabi_fneg(fp_t a) COMPILER_RT_ALIAS(__negsf2);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/subdf3.c b/contrib/compiler-rt/lib/builtins/subdf3.c
index 38340dfab1a6..a892fa603cf2 100644
--- a/contrib/compiler-rt/lib/builtins/subdf3.c
+++ b/contrib/compiler-rt/lib/builtins/subdf3.c
@@ -22,8 +22,11 @@ __subdf3(fp_t a, fp_t b) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_dsub(fp_t a, fp_t b) {
   return __subdf3(a, b);
 }
+#else
+AEABI_RTABI fp_t __aeabi_dsub(fp_t a, fp_t b) COMPILER_RT_ALIAS(__subdf3);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/subsf3.c b/contrib/compiler-rt/lib/builtins/subsf3.c
index 34276b1447ba..4b2786177dcc 100644
--- a/contrib/compiler-rt/lib/builtins/subsf3.c
+++ b/contrib/compiler-rt/lib/builtins/subsf3.c
@@ -22,8 +22,11 @@ __subsf3(fp_t a, fp_t b) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI fp_t __aeabi_fsub(fp_t a, fp_t b) {
   return __subsf3(a, b);
 }
+#else
+AEABI_RTABI fp_t __aeabi_fsub(fp_t a, fp_t b) COMPILER_RT_ALIAS(__subsf3);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/truncdfhf2.c b/contrib/compiler-rt/lib/builtins/truncdfhf2.c
index 4bb71aa178a0..8354a41b8b6f 100644
--- a/contrib/compiler-rt/lib/builtins/truncdfhf2.c
+++ b/contrib/compiler-rt/lib/builtins/truncdfhf2.c
@@ -16,8 +16,11 @@ COMPILER_RT_ABI uint16_t __truncdfhf2(double a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI uint16_t __aeabi_d2h(double a) {
   return __truncdfhf2(a);
 }
+#else
+AEABI_RTABI uint16_t __aeabi_d2h(double a) COMPILER_RT_ALIAS(__truncdfhf2);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/truncdfsf2.c b/contrib/compiler-rt/lib/builtins/truncdfsf2.c
index 8bf58bb23a3b..195d3e0656e7 100644
--- a/contrib/compiler-rt/lib/builtins/truncdfsf2.c
+++ b/contrib/compiler-rt/lib/builtins/truncdfsf2.c
@@ -16,8 +16,11 @@ COMPILER_RT_ABI float __truncdfsf2(double a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI float __aeabi_d2f(double a) {
   return __truncdfsf2(a);
 }
+#else
+AEABI_RTABI float __aeabi_d2f(double a) COMPILER_RT_ALIAS(__truncdfsf2);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/truncsfhf2.c b/contrib/compiler-rt/lib/builtins/truncsfhf2.c
index f6ce1fa1de05..9c84ab4f938a 100644
--- a/contrib/compiler-rt/lib/builtins/truncsfhf2.c
+++ b/contrib/compiler-rt/lib/builtins/truncsfhf2.c
@@ -22,8 +22,11 @@ COMPILER_RT_ABI uint16_t __gnu_f2h_ieee(float a) {
 }
 
 #if defined(__ARM_EABI__)
+#if defined(COMPILER_RT_ARMHF_TARGET)
 AEABI_RTABI uint16_t __aeabi_f2h(float a) {
   return __truncsfhf2(a);
 }
+#else
+AEABI_RTABI uint16_t __aeabi_f2h(float a) COMPILER_RT_ALIAS(__truncsfhf2);
+#endif
 #endif
-
diff --git a/contrib/compiler-rt/lib/builtins/udivsi3.c b/contrib/compiler-rt/lib/builtins/udivsi3.c
index 8eccf102cc97..bb720f8c382b 100644
--- a/contrib/compiler-rt/lib/builtins/udivsi3.c
+++ b/contrib/compiler-rt/lib/builtins/udivsi3.c
@@ -64,8 +64,5 @@ __udivsi3(su_int n, su_int d)
 }
 
 #if defined(__ARM_EABI__)
-AEABI_RTABI su_int __aeabi_uidiv(su_int n, su_int d) {
-  return __udivsi3(n, d);
-}
+AEABI_RTABI su_int __aeabi_uidiv(su_int n, su_int d) COMPILER_RT_ALIAS(__udivsi3);
 #endif
-
diff --git a/contrib/compiler-rt/lib/cfi/cfi.cc b/contrib/compiler-rt/lib/cfi/cfi.cc
index f720230a70be..f7693b37d29c 100644
--- a/contrib/compiler-rt/lib/cfi/cfi.cc
+++ b/contrib/compiler-rt/lib/cfi/cfi.cc
@@ -280,7 +280,7 @@ void InitShadow() {
   CHECK_EQ(0, GetShadow());
   CHECK_EQ(0, GetShadowSize());
 
-  uptr vma = GetMaxVirtualAddress();
+  uptr vma = GetMaxUserVirtualAddress();
   // Shadow is 2 -> 2**kShadowGranularity.
   SetShadowSize((vma >> (kShadowGranularity - 1)) + 1);
   VReport(1, "CFI: VMA size %zx, shadow size %zx\n", vma, GetShadowSize());
diff --git a/contrib/compiler-rt/lib/cfi/cfi_blacklist.txt b/contrib/compiler-rt/lib/cfi/cfi_blacklist.txt
index cc111be8120e..d8c9d49e3cdd 100644
--- a/contrib/compiler-rt/lib/cfi/cfi_blacklist.txt
+++ b/contrib/compiler-rt/lib/cfi/cfi_blacklist.txt
@@ -1,18 +1,4 @@
-# Standard library types.
-type:std::*
-
-# The stdext namespace contains Microsoft standard library extensions.
-type:stdext::*
-
-# Types with a uuid attribute, i.e. COM types.
-type:attr:uuid
-
-# STL allocators (T *allocator<T *>::allocate(size_type, const void*)).
-# The type signature mandates a cast from uninitialized void* to T*.
-# size_type can either be unsigned int (j) or unsigned long (m).
-fun:*8allocateEjPKv
-fun:*8allocateEmPKv
-
+[cfi-unrelated-cast]
 # std::get_temporary_buffer, likewise (libstdc++, libc++).
 fun:_ZSt20get_temporary_buffer*
 fun:_ZNSt3__120get_temporary_buffer*
diff --git a/contrib/compiler-rt/lib/dfsan/dfsan.cc b/contrib/compiler-rt/lib/dfsan/dfsan.cc
index 3aa99b7f9c27..9e360c959fa7 100644
--- a/contrib/compiler-rt/lib/dfsan/dfsan.cc
+++ b/contrib/compiler-rt/lib/dfsan/dfsan.cc
@@ -21,6 +21,7 @@
 
 #include "sanitizer_common/sanitizer_atomic.h"
 #include "sanitizer_common/sanitizer_common.h"
+#include "sanitizer_common/sanitizer_file.h"
 #include "sanitizer_common/sanitizer_flags.h"
 #include "sanitizer_common/sanitizer_flag_parser.h"
 #include "sanitizer_common/sanitizer_libc.h"
diff --git a/contrib/compiler-rt/lib/esan/esan_sideline.h b/contrib/compiler-rt/lib/esan/esan_sideline.h
index aa3fae1db318..04aff22f4cf0 100644
--- a/contrib/compiler-rt/lib/esan/esan_sideline.h
+++ b/contrib/compiler-rt/lib/esan/esan_sideline.h
@@ -17,6 +17,7 @@
 
 #include "sanitizer_common/sanitizer_atomic.h"
 #include "sanitizer_common/sanitizer_internal_defs.h"
+#include "sanitizer_common/sanitizer_platform_limits_posix.h"
 
 namespace __esan {
 
@@ -46,7 +47,8 @@ public:
 private:
   static int runSideline(void *Arg);
   static void registerSignal(int SigNum);
-  static void handleSidelineSignal(int SigNum, void *SigInfo, void *Ctx);
+  static void handleSidelineSignal(int SigNum, __sanitizer_siginfo *SigInfo,
+                                   void *Ctx);
 
   char *Stack;
   SidelineFunc sampleFunc;
diff --git a/contrib/compiler-rt/lib/esan/esan_sideline_linux.cpp b/contrib/compiler-rt/lib/esan/esan_sideline_linux.cpp
index bc272dfe49f8..4a96910ece35 100644
--- a/contrib/compiler-rt/lib/esan/esan_sideline_linux.cpp
+++ b/contrib/compiler-rt/lib/esan/esan_sideline_linux.cpp
@@ -40,7 +40,8 @@ static const uptr SidelineIdUninitialized = 1;
 static SidelineThread *TheThread;
 
 // We aren't passing SA_NODEFER so the same signal is blocked while here.
-void SidelineThread::handleSidelineSignal(int SigNum, void *SigInfo,
+void SidelineThread::handleSidelineSignal(int SigNum,
+                                          __sanitizer_siginfo *SigInfo,
                                           void *Ctx) {
   VPrintf(3, "Sideline signal %d\n", SigNum);
   CHECK_EQ(SigNum, SIGALRM);
diff --git a/contrib/compiler-rt/lib/esan/working_set_posix.cpp b/contrib/compiler-rt/lib/esan/working_set_posix.cpp
index fcfa87128857..5ec53b959261 100644
--- a/contrib/compiler-rt/lib/esan/working_set_posix.cpp
+++ b/contrib/compiler-rt/lib/esan/working_set_posix.cpp
@@ -34,7 +34,7 @@ bool processWorkingSetSignal(int SigNum, void (*Handler)(int),
   VPrintf(2, "%s: %d\n", __FUNCTION__, SigNum);
   if (SigNum == SIGSEGV) {
     *Result = AppSigAct.handler;
-    AppSigAct.sigaction = (void (*)(int, void*, void*))Handler;
+    AppSigAct.sigaction = (decltype(AppSigAct.sigaction))Handler;
     return false; // Skip real call.
   }
   return true;
@@ -73,7 +73,7 @@ bool processWorkingSetSigprocmask(int How, void *Set, void *OldSet) {
 static void reinstateDefaultHandler(int SigNum) {
   __sanitizer_sigaction SigAct;
   internal_memset(&SigAct, 0, sizeof(SigAct));
-  SigAct.sigaction = (void (*)(int, void*, void*)) SIG_DFL;
+  SigAct.sigaction = (decltype(SigAct.sigaction))SIG_DFL;
   int Res = internal_sigaction(SigNum, &SigAct, nullptr);
   CHECK(Res == 0);
   VPrintf(1, "Unregistered for %d handler\n", SigNum);
@@ -81,7 +81,8 @@ static void reinstateDefaultHandler(int SigNum) {
 
 // If this is a shadow fault, we handle it here; otherwise, we pass it to the
 // app to handle it just as the app would do without our tool in place.
-static void handleMemoryFault(int SigNum, void *Info, void *Ctx) {
+static void handleMemoryFault(int SigNum, __sanitizer_siginfo *Info,
+                              void *Ctx) {
   if (SigNum == SIGSEGV) {
     // We rely on si_addr being filled in (thus we do not support old kernels).
     siginfo_t *SigInfo = (siginfo_t *)Info;
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan.cc b/contrib/compiler-rt/lib/hwasan/hwasan.cc
new file mode 100644
index 000000000000..fcc40eb90182
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan.cc
@@ -0,0 +1,303 @@
+//===-- hwasan.cc -----------------------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+// HWAddressSanitizer runtime.
+//===----------------------------------------------------------------------===//
+
+#include "hwasan.h"
+#include "hwasan_thread.h"
+#include "hwasan_poisoning.h"
+#include "sanitizer_common/sanitizer_atomic.h"
+#include "sanitizer_common/sanitizer_common.h"
+#include "sanitizer_common/sanitizer_flags.h"
+#include "sanitizer_common/sanitizer_flag_parser.h"
+#include "sanitizer_common/sanitizer_libc.h"
+#include "sanitizer_common/sanitizer_procmaps.h"
+#include "sanitizer_common/sanitizer_stacktrace.h"
+#include "sanitizer_common/sanitizer_symbolizer.h"
+#include "sanitizer_common/sanitizer_stackdepot.h"
+#include "ubsan/ubsan_flags.h"
+#include "ubsan/ubsan_init.h"
+
+// ACHTUNG! No system header includes in this file.
+
+using namespace __sanitizer;
+
+namespace __hwasan {
+
+void EnterSymbolizer() {
+  HwasanThread *t = GetCurrentThread();
+  CHECK(t);
+  t->EnterSymbolizer();
+}
+void ExitSymbolizer() {
+  HwasanThread *t = GetCurrentThread();
+  CHECK(t);
+  t->LeaveSymbolizer();
+}
+bool IsInSymbolizer() {
+  HwasanThread *t = GetCurrentThread();
+  return t && t->InSymbolizer();
+}
+
+static Flags hwasan_flags;
+
+Flags *flags() {
+  return &hwasan_flags;
+}
+
+int hwasan_inited = 0;
+bool hwasan_init_is_running;
+
+int hwasan_report_count = 0;
+
+void Flags::SetDefaults() {
+#define HWASAN_FLAG(Type, Name, DefaultValue, Description) Name = DefaultValue;
+#include "hwasan_flags.inc"
+#undef HWASAN_FLAG
+}
+
+static void RegisterHwasanFlags(FlagParser *parser, Flags *f) {
+#define HWASAN_FLAG(Type, Name, DefaultValue, Description) \
+  RegisterFlag(parser, #Name, Description, &f->Name);
+#include "hwasan_flags.inc"
+#undef HWASAN_FLAG
+}
+
+static void InitializeFlags() {
+  SetCommonFlagsDefaults();
+  {
+    CommonFlags cf;
+    cf.CopyFrom(*common_flags());
+    cf.external_symbolizer_path = GetEnv("HWASAN_SYMBOLIZER_PATH");
+    cf.malloc_context_size = 20;
+    cf.handle_ioctl = true;
+    // FIXME: test and enable.
+    cf.check_printf = false;
+    cf.intercept_tls_get_addr = true;
+    cf.exitcode = 99;
+    cf.handle_sigill = kHandleSignalExclusive;
+    OverrideCommonFlags(cf);
+  }
+
+  Flags *f = flags();
+  f->SetDefaults();
+
+  FlagParser parser;
+  RegisterHwasanFlags(&parser, f);
+  RegisterCommonFlags(&parser);
+
+#if HWASAN_CONTAINS_UBSAN
+  __ubsan::Flags *uf = __ubsan::flags();
+  uf->SetDefaults();
+
+  FlagParser ubsan_parser;
+  __ubsan::RegisterUbsanFlags(&ubsan_parser, uf);
+  RegisterCommonFlags(&ubsan_parser);
+#endif
+
+  // Override from user-specified string.
+  if (__hwasan_default_options)
+    parser.ParseString(__hwasan_default_options());
+#if HWASAN_CONTAINS_UBSAN
+  const char *ubsan_default_options = __ubsan::MaybeCallUbsanDefaultOptions();
+  ubsan_parser.ParseString(ubsan_default_options);
+#endif
+
+  const char *hwasan_options = GetEnv("HWASAN_OPTIONS");
+  parser.ParseString(hwasan_options);
+#if HWASAN_CONTAINS_UBSAN
+  ubsan_parser.ParseString(GetEnv("UBSAN_OPTIONS"));
+#endif
+  VPrintf(1, "HWASAN_OPTIONS: %s\n", hwasan_options ? hwasan_options : "<empty>");
+
+  InitializeCommonFlags();
+
+  if (Verbosity()) ReportUnrecognizedFlags();
+
+  if (common_flags()->help) parser.PrintFlagDescriptions();
+}
+
+void GetStackTrace(BufferedStackTrace *stack, uptr max_s, uptr pc, uptr bp,
+                   void *context, bool request_fast_unwind) {
+  HwasanThread *t = GetCurrentThread();
+  if (!t || !StackTrace::WillUseFastUnwind(request_fast_unwind)) {
+    // Block reports from our interceptors during _Unwind_Backtrace.
+    SymbolizerScope sym_scope;
+    return stack->Unwind(max_s, pc, bp, context, 0, 0, request_fast_unwind);
+  }
+  stack->Unwind(max_s, pc, bp, context, t->stack_top(), t->stack_bottom(),
+                request_fast_unwind);
+}
+
+void PrintWarning(uptr pc, uptr bp) {
+  GET_FATAL_STACK_TRACE_PC_BP(pc, bp);
+  ReportInvalidAccess(&stack, 0);
+}
+
+} // namespace __hwasan
+
+// Interface.
+
+using namespace __hwasan;
+
+void __hwasan_init() {
+  CHECK(!hwasan_init_is_running);
+  if (hwasan_inited) return;
+  hwasan_init_is_running = 1;
+  SanitizerToolName = "HWAddressSanitizer";
+
+  InitTlsSize();
+
+  CacheBinaryName();
+  InitializeFlags();
+
+  __sanitizer_set_report_path(common_flags()->log_path);
+
+  InitializeInterceptors();
+  InstallDeadlySignalHandlers(HwasanOnDeadlySignal);
+  InstallAtExitHandler(); // Needs __cxa_atexit interceptor.
+
+  DisableCoreDumperIfNecessary();
+  if (!InitShadow()) {
+    Printf("FATAL: HWAddressSanitizer can not mmap the shadow memory.\n");
+    Printf("FATAL: Make sure to compile with -fPIE and to link with -pie.\n");
+    Printf("FATAL: Disabling ASLR is known to cause this error.\n");
+    Printf("FATAL: If running under GDB, try "
+           "'set disable-randomization off'.\n");
+    DumpProcessMap();
+    Die();
+  }
+
+  Symbolizer::GetOrInit()->AddHooks(EnterSymbolizer, ExitSymbolizer);
+
+  InitializeCoverage(common_flags()->coverage, common_flags()->coverage_dir);
+
+  HwasanTSDInit(HwasanTSDDtor);
+
+  HwasanAllocatorInit();
+
+  HwasanThread *main_thread = HwasanThread::Create(nullptr, nullptr);
+  SetCurrentThread(main_thread);
+  main_thread->ThreadStart();
+
+#if HWASAN_CONTAINS_UBSAN
+  __ubsan::InitAsPlugin();
+#endif
+
+  VPrintf(1, "HWAddressSanitizer init done\n");
+
+  hwasan_init_is_running = 0;
+  hwasan_inited = 1;
+}
+
+void __hwasan_print_shadow(const void *x, uptr size) {
+  // FIXME:
+  Printf("FIXME: __hwasan_print_shadow unimplemented\n");
+}
+
+sptr __hwasan_test_shadow(const void *p, uptr sz) {
+  if (sz == 0)
+    return -1;
+  tag_t ptr_tag = GetTagFromPointer((uptr)p);
+  if (ptr_tag == 0)
+    return -1;
+  uptr ptr_raw = GetAddressFromPointer((uptr)p);
+  uptr shadow_first = MEM_TO_SHADOW(ptr_raw);
+  uptr shadow_last = MEM_TO_SHADOW(ptr_raw + sz - 1);
+  for (uptr s = shadow_first; s <= shadow_last; ++s)
+    if (*(tag_t*)s != ptr_tag)
+      return SHADOW_TO_MEM(s) - ptr_raw;
+  return -1;
+}
+
+u16 __sanitizer_unaligned_load16(const uu16 *p) {
+  return *p;
+}
+u32 __sanitizer_unaligned_load32(const uu32 *p) {
+  return *p;
+}
+u64 __sanitizer_unaligned_load64(const uu64 *p) {
+  return *p;
+}
+void __sanitizer_unaligned_store16(uu16 *p, u16 x) {
+  *p = x;
+}
+void __sanitizer_unaligned_store32(uu32 *p, u32 x) {
+  *p = x;
+}
+void __sanitizer_unaligned_store64(uu64 *p, u64 x) {
+  *p = x;
+}
+
+template<unsigned X>
+__attribute__((always_inline))
+static void SigIll() {
+#if defined(__aarch64__)
+  asm("hlt %0\n\t" ::"n"(X));
+#elif defined(__x86_64__) || defined(__i386__)
+  asm("ud2\n\t");
+#else
+  // FIXME: not always sigill.
+  __builtin_trap();
+#endif
+  // __builtin_unreachable();
+}
+
+template<bool IsStore, unsigned LogSize>
+__attribute__((always_inline, nodebug))
+static void CheckAddress(uptr p) {
+  tag_t ptr_tag = GetTagFromPointer(p);
+  uptr ptr_raw = p & ~kAddressTagMask;
+  tag_t mem_tag = *(tag_t *)MEM_TO_SHADOW(ptr_raw);
+  if (UNLIKELY(ptr_tag != mem_tag)) SigIll<0x100 + 0x10 * IsStore + LogSize>();
+}
+
+template<bool IsStore>
+__attribute__((always_inline, nodebug))
+static void CheckAddressSized(uptr p, uptr sz) {
+  CHECK_NE(0, sz);
+  tag_t ptr_tag = GetTagFromPointer(p);
+  uptr ptr_raw = p & ~kAddressTagMask;
+  tag_t *shadow_first = (tag_t *)MEM_TO_SHADOW(ptr_raw);
+  tag_t *shadow_last = (tag_t *)MEM_TO_SHADOW(ptr_raw + sz - 1);
+  for (tag_t *t = shadow_first; t <= shadow_last; ++t)
+    if (UNLIKELY(ptr_tag != *t)) SigIll<0x100 + 0x10 * IsStore + 0xf>();
+}
+
+void __hwasan_load(uptr p, uptr sz) { CheckAddressSized<false>(p, sz); }
+void __hwasan_load1(uptr p) { CheckAddress<false, 0>(p); }
+void __hwasan_load2(uptr p) { CheckAddress<false, 1>(p); }
+void __hwasan_load4(uptr p) { CheckAddress<false, 2>(p); }
+void __hwasan_load8(uptr p) { CheckAddress<false, 3>(p); }
+void __hwasan_load16(uptr p) { CheckAddress<false, 4>(p); }
+
+void __hwasan_store(uptr p, uptr sz) { CheckAddressSized<true>(p, sz); }
+void __hwasan_store1(uptr p) { CheckAddress<true, 0>(p); }
+void __hwasan_store2(uptr p) { CheckAddress<true, 1>(p); }
+void __hwasan_store4(uptr p) { CheckAddress<true, 2>(p); }
+void __hwasan_store8(uptr p) { CheckAddress<true, 3>(p); }
+void __hwasan_store16(uptr p) { CheckAddress<true, 4>(p); }
+
+#if !SANITIZER_SUPPORTS_WEAK_HOOKS
+extern "C" {
+SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
+const char* __hwasan_default_options() { return ""; }
+}  // extern "C"
+#endif
+
+extern "C" {
+SANITIZER_INTERFACE_ATTRIBUTE
+void __sanitizer_print_stack_trace() {
+  GET_FATAL_STACK_TRACE_PC_BP(StackTrace::GetCurrentPc(), GET_CURRENT_FRAME());
+  stack.Print();
+}
+} // extern "C"
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan.h b/contrib/compiler-rt/lib/hwasan/hwasan.h
new file mode 100644
index 000000000000..bcf5282dce7c
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan.h
@@ -0,0 +1,172 @@
+//===-- hwasan.h --------------------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+// Private Hwasan header.
+//===----------------------------------------------------------------------===//
+
+#ifndef HWASAN_H
+#define HWASAN_H
+
+#include "sanitizer_common/sanitizer_flags.h"
+#include "sanitizer_common/sanitizer_internal_defs.h"
+#include "sanitizer_common/sanitizer_stacktrace.h"
+#include "hwasan_interface_internal.h"
+#include "hwasan_flags.h"
+#include "ubsan/ubsan_platform.h"
+
+#ifndef HWASAN_REPLACE_OPERATORS_NEW_AND_DELETE
+# define HWASAN_REPLACE_OPERATORS_NEW_AND_DELETE 1
+#endif
+
+#ifndef HWASAN_CONTAINS_UBSAN
+# define HWASAN_CONTAINS_UBSAN CAN_SANITIZE_UB
+#endif
+
+typedef u8 tag_t;
+
+// Reasonable values are 4 (for 1/16th shadow) and 6 (for 1/64th).
+const uptr kShadowScale = 4;
+const uptr kShadowAlignment = 1UL << kShadowScale;
+
+#define MEM_TO_SHADOW_OFFSET(mem) ((uptr)(mem) >> kShadowScale)
+#define MEM_TO_SHADOW(mem) ((uptr)(mem) >> kShadowScale)
+#define SHADOW_TO_MEM(shadow) ((uptr)(shadow) << kShadowScale)
+
+#define MEM_IS_APP(mem) true
+
+// TBI (Top Byte Ignore) feature of AArch64: bits [63:56] are ignored in address
+// translation and can be used to store a tag.
+const unsigned kAddressTagShift = 56;
+const uptr kAddressTagMask = 0xFFUL << kAddressTagShift;
+
+static inline tag_t GetTagFromPointer(uptr p) {
+  return p >> kAddressTagShift;
+}
+
+static inline uptr GetAddressFromPointer(uptr p) {
+  return p & ~kAddressTagMask;
+}
+
+static inline void * GetAddressFromPointer(const void *p) {
+  return (void *)((uptr)p & ~kAddressTagMask);
+}
+
+static inline uptr AddTagToPointer(uptr p, tag_t tag) {
+  return (p & ~kAddressTagMask) | ((uptr)tag << kAddressTagShift);
+}
+
+namespace __hwasan {
+
+extern int hwasan_inited;
+extern bool hwasan_init_is_running;
+extern int hwasan_report_count;
+
+bool ProtectRange(uptr beg, uptr end);
+bool InitShadow();
+char *GetProcSelfMaps();
+void InitializeInterceptors();
+
+void HwasanAllocatorInit();
+void HwasanAllocatorThreadFinish();
+void HwasanDeallocate(StackTrace *stack, void *ptr);
+
+void *hwasan_malloc(uptr size, StackTrace *stack);
+void *hwasan_calloc(uptr nmemb, uptr size, StackTrace *stack);
+void *hwasan_realloc(void *ptr, uptr size, StackTrace *stack);
+void *hwasan_valloc(uptr size, StackTrace *stack);
+void *hwasan_pvalloc(uptr size, StackTrace *stack);
+void *hwasan_aligned_alloc(uptr alignment, uptr size, StackTrace *stack);
+void *hwasan_memalign(uptr alignment, uptr size, StackTrace *stack);
+int hwasan_posix_memalign(void **memptr, uptr alignment, uptr size,
+                        StackTrace *stack);
+
+void InstallTrapHandler();
+void InstallAtExitHandler();
+
+const char *GetStackOriginDescr(u32 id, uptr *pc);
+
+void EnterSymbolizer();
+void ExitSymbolizer();
+bool IsInSymbolizer();
+
+struct SymbolizerScope {
+  SymbolizerScope() { EnterSymbolizer(); }
+  ~SymbolizerScope() { ExitSymbolizer(); }
+};
+
+void PrintWarning(uptr pc, uptr bp);
+
+void GetStackTrace(BufferedStackTrace *stack, uptr max_s, uptr pc, uptr bp,
+                   void *context, bool request_fast_unwind);
+
+void ReportInvalidAccess(StackTrace *stack, u32 origin);
+void ReportTagMismatch(StackTrace *stack, uptr addr, uptr access_size,
+                       bool is_store);
+void ReportStats();
+void ReportAtExitStatistics();
+void DescribeMemoryRange(const void *x, uptr size);
+void ReportInvalidAccessInsideAddressRange(const char *what, const void *start, uptr size,
+                                 uptr offset);
+
+// Returns a "chained" origin id, pointing to the given stack trace followed by
+// the previous origin id.
+u32 ChainOrigin(u32 id, StackTrace *stack);
+
+const int STACK_TRACE_TAG_POISON = StackTrace::TAG_CUSTOM + 1;
+
+#define GET_MALLOC_STACK_TRACE                                            \
+  BufferedStackTrace stack;                                               \
+  if (hwasan_inited)                                                       \
+  GetStackTrace(&stack, common_flags()->malloc_context_size,              \
+                StackTrace::GetCurrentPc(), GET_CURRENT_FRAME(), nullptr, \
+                common_flags()->fast_unwind_on_malloc)
+
+#define GET_FATAL_STACK_TRACE_PC_BP(pc, bp)              \
+  BufferedStackTrace stack;                              \
+  if (hwasan_inited)                                       \
+  GetStackTrace(&stack, kStackTraceMax, pc, bp, nullptr, \
+                common_flags()->fast_unwind_on_fatal)
+
+class ScopedThreadLocalStateBackup {
+ public:
+  ScopedThreadLocalStateBackup() { Backup(); }
+  ~ScopedThreadLocalStateBackup() { Restore(); }
+  void Backup();
+  void Restore();
+ private:
+  u64 va_arg_overflow_size_tls;
+};
+
+void HwasanTSDInit(void (*destructor)(void *tsd));
+void *HwasanTSDGet();
+void HwasanTSDSet(void *tsd);
+void HwasanTSDDtor(void *tsd);
+
+void HwasanOnDeadlySignal(int signo, void *info, void *context);
+
+}  // namespace __hwasan
+
+#define HWASAN_MALLOC_HOOK(ptr, size)       \
+  do {                                    \
+    if (&__sanitizer_malloc_hook) {       \
+      __sanitizer_malloc_hook(ptr, size); \
+    }                                     \
+    RunMallocHooks(ptr, size);            \
+  } while (false)
+#define HWASAN_FREE_HOOK(ptr)       \
+  do {                            \
+    if (&__sanitizer_free_hook) { \
+      __sanitizer_free_hook(ptr); \
+    }                             \
+    RunFreeHooks(ptr);            \
+  } while (false)
+
+#endif  // HWASAN_H
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan.syms.extra b/contrib/compiler-rt/lib/hwasan/hwasan.syms.extra
new file mode 100644
index 000000000000..8738627e06e8
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan.syms.extra
@@ -0,0 +1,2 @@
+__hwasan_*
+__ubsan_*
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_allocator.cc b/contrib/compiler-rt/lib/hwasan/hwasan_allocator.cc
new file mode 100644
index 000000000000..fbcaf78b88f0
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_allocator.cc
@@ -0,0 +1,330 @@
+//===-- hwasan_allocator.cc --------------------------- ---------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+// HWAddressSanitizer allocator.
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_common/sanitizer_allocator.h"
+#include "sanitizer_common/sanitizer_allocator_checks.h"
+#include "sanitizer_common/sanitizer_allocator_interface.h"
+#include "sanitizer_common/sanitizer_atomic.h"
+#include "sanitizer_common/sanitizer_errno.h"
+#include "sanitizer_common/sanitizer_stackdepot.h"
+#include "hwasan.h"
+#include "hwasan_allocator.h"
+#include "hwasan_thread.h"
+#include "hwasan_poisoning.h"
+
+namespace __hwasan {
+
+enum {
+  CHUNK_INVALID = 0,
+  CHUNK_FREE = 1,
+  CHUNK_ALLOCATED = 2
+};
+
+struct Metadata {
+  u64 state : 2;
+  u64 requested_size : 62;
+  u32 alloc_context_id;
+  u32 free_context_id;
+};
+
+bool HwasanChunkView::IsValid() const {
+  return metadata_ && metadata_->state != CHUNK_INVALID;
+}
+bool HwasanChunkView::IsAllocated() const {
+  return metadata_ && metadata_->state == CHUNK_ALLOCATED;
+}
+uptr HwasanChunkView::Beg() const {
+  return block_;
+}
+uptr HwasanChunkView::End() const {
+  return Beg() + UsedSize();
+}
+uptr HwasanChunkView::UsedSize() const {
+  return metadata_->requested_size;
+}
+u32 HwasanChunkView::GetAllocStackId() const {
+  return metadata_->alloc_context_id;
+}
+u32 HwasanChunkView::GetFreeStackId() const {
+  return metadata_->free_context_id;
+}
+
+struct HwasanMapUnmapCallback {
+  void OnMap(uptr p, uptr size) const {}
+  void OnUnmap(uptr p, uptr size) const {
+    // We are about to unmap a chunk of user memory.
+    // It can return as user-requested mmap() or another thread stack.
+    // Make it accessible with zero-tagged pointer.
+    TagMemory(p, size, 0);
+  }
+};
+
+#if !defined(__aarch64__)
+#error unsupported platform
+#endif
+
+static const uptr kMaxAllowedMallocSize = 2UL << 30;  // 2G
+static const uptr kRegionSizeLog = 20;
+static const uptr kNumRegions = SANITIZER_MMAP_RANGE_SIZE >> kRegionSizeLog;
+typedef TwoLevelByteMap<(kNumRegions >> 12), 1 << 12> ByteMap;
+
+struct AP32 {
+  static const uptr kSpaceBeg = 0;
+  static const u64 kSpaceSize = SANITIZER_MMAP_RANGE_SIZE;
+  static const uptr kMetadataSize = sizeof(Metadata);
+  typedef __sanitizer::CompactSizeClassMap SizeClassMap;
+  static const uptr kRegionSizeLog = __hwasan::kRegionSizeLog;
+  typedef __hwasan::ByteMap ByteMap;
+  typedef HwasanMapUnmapCallback MapUnmapCallback;
+  static const uptr kFlags = 0;
+};
+typedef SizeClassAllocator32<AP32> PrimaryAllocator;
+typedef SizeClassAllocatorLocalCache<PrimaryAllocator> AllocatorCache;
+typedef LargeMmapAllocator<HwasanMapUnmapCallback> SecondaryAllocator;
+typedef CombinedAllocator<PrimaryAllocator, AllocatorCache,
+                          SecondaryAllocator> Allocator;
+
+static Allocator allocator;
+static AllocatorCache fallback_allocator_cache;
+static SpinMutex fallback_mutex;
+static atomic_uint8_t hwasan_allocator_tagging_enabled;
+
+void HwasanAllocatorInit() {
+  atomic_store_relaxed(&hwasan_allocator_tagging_enabled,
+                       !flags()->disable_allocator_tagging);
+  SetAllocatorMayReturnNull(common_flags()->allocator_may_return_null);
+  allocator.Init(common_flags()->allocator_release_to_os_interval_ms);
+}
+
+AllocatorCache *GetAllocatorCache(HwasanThreadLocalMallocStorage *ms) {
+  CHECK(ms);
+  CHECK_LE(sizeof(AllocatorCache), sizeof(ms->allocator_cache));
+  return reinterpret_cast<AllocatorCache *>(ms->allocator_cache);
+}
+
+void HwasanThreadLocalMallocStorage::CommitBack() {
+  allocator.SwallowCache(GetAllocatorCache(this));
+}
+
+static void *HwasanAllocate(StackTrace *stack, uptr size, uptr alignment,
+                          bool zeroise) {
+  alignment = Max(alignment, kShadowAlignment);
+  size = RoundUpTo(size, kShadowAlignment);
+
+  if (size > kMaxAllowedMallocSize) {
+    Report("WARNING: HWAddressSanitizer failed to allocate %p bytes\n",
+           (void *)size);
+    return Allocator::FailureHandler::OnBadRequest();
+  }
+  HwasanThread *t = GetCurrentThread();
+  void *allocated;
+  if (t) {
+    AllocatorCache *cache = GetAllocatorCache(&t->malloc_storage());
+    allocated = allocator.Allocate(cache, size, alignment);
+  } else {
+    SpinMutexLock l(&fallback_mutex);
+    AllocatorCache *cache = &fallback_allocator_cache;
+    allocated = allocator.Allocate(cache, size, alignment);
+  }
+  Metadata *meta =
+      reinterpret_cast<Metadata *>(allocator.GetMetaData(allocated));
+  meta->state = CHUNK_ALLOCATED;
+  meta->requested_size = size;
+  meta->alloc_context_id = StackDepotPut(*stack);
+  if (zeroise)
+    internal_memset(allocated, 0, size);
+
+  void *user_ptr = (flags()->tag_in_malloc &&
+                    atomic_load_relaxed(&hwasan_allocator_tagging_enabled))
+                       ? (void *)TagMemoryAligned((uptr)allocated, size, 0xBB)
+                       : allocated;
+
+  HWASAN_MALLOC_HOOK(user_ptr, size);
+  return user_ptr;
+}
+
+void HwasanDeallocate(StackTrace *stack, void *user_ptr) {
+  CHECK(user_ptr);
+  HWASAN_FREE_HOOK(user_ptr);
+
+  void *p = GetAddressFromPointer(user_ptr);
+  Metadata *meta = reinterpret_cast<Metadata *>(allocator.GetMetaData(p));
+  uptr size = meta->requested_size;
+  meta->state = CHUNK_FREE;
+  meta->requested_size = 0;
+  meta->free_context_id = StackDepotPut(*stack);
+  // This memory will not be reused by anyone else, so we are free to keep it
+  // poisoned.
+  if (flags()->tag_in_free &&
+      atomic_load_relaxed(&hwasan_allocator_tagging_enabled))
+    TagMemoryAligned((uptr)p, size, 0xBC);
+  HwasanThread *t = GetCurrentThread();
+  if (t) {
+    AllocatorCache *cache = GetAllocatorCache(&t->malloc_storage());
+    allocator.Deallocate(cache, p);
+  } else {
+    SpinMutexLock l(&fallback_mutex);
+    AllocatorCache *cache = &fallback_allocator_cache;
+    allocator.Deallocate(cache, p);
+  }
+}
+
+void *HwasanReallocate(StackTrace *stack, void *user_old_p, uptr new_size,
+                     uptr alignment) {
+  alignment = Max(alignment, kShadowAlignment);
+  new_size = RoundUpTo(new_size, kShadowAlignment);
+
+  void *old_p = GetAddressFromPointer(user_old_p);
+  Metadata *meta = reinterpret_cast<Metadata*>(allocator.GetMetaData(old_p));
+  uptr old_size = meta->requested_size;
+  uptr actually_allocated_size = allocator.GetActuallyAllocatedSize(old_p);
+  if (new_size <= actually_allocated_size) {
+    // We are not reallocating here.
+    // FIXME: update stack trace for the allocation?
+    meta->requested_size = new_size;
+    if (!atomic_load_relaxed(&hwasan_allocator_tagging_enabled))
+      return user_old_p;
+    if (flags()->retag_in_realloc)
+      return (void *)TagMemoryAligned((uptr)old_p, new_size, 0xCC);
+    if (new_size > old_size) {
+      tag_t tag = GetTagFromPointer((uptr)user_old_p);
+      TagMemoryAligned((uptr)old_p + old_size, new_size - old_size, tag);
+    }
+    return user_old_p;
+  }
+  uptr memcpy_size = Min(new_size, old_size);
+  void *new_p = HwasanAllocate(stack, new_size, alignment, false /*zeroise*/);
+  if (new_p) {
+    internal_memcpy(new_p, old_p, memcpy_size);
+    HwasanDeallocate(stack, old_p);
+  }
+  return new_p;
+}
+
+HwasanChunkView FindHeapChunkByAddress(uptr address) {
+  void *block = allocator.GetBlockBegin(reinterpret_cast<void*>(address));
+  if (!block)
+    return HwasanChunkView();
+  Metadata *metadata =
+      reinterpret_cast<Metadata*>(allocator.GetMetaData(block));
+  return HwasanChunkView(reinterpret_cast<uptr>(block), metadata);
+}
+
+static uptr AllocationSize(const void *user_ptr) {
+  const void *p = GetAddressFromPointer(user_ptr);
+  if (!p) return 0;
+  const void *beg = allocator.GetBlockBegin(p);
+  if (beg != p) return 0;
+  Metadata *b = (Metadata *)allocator.GetMetaData(p);
+  return b->requested_size;
+}
+
+void *hwasan_malloc(uptr size, StackTrace *stack) {
+  return SetErrnoOnNull(HwasanAllocate(stack, size, sizeof(u64), false));
+}
+
+void *hwasan_calloc(uptr nmemb, uptr size, StackTrace *stack) {
+  if (UNLIKELY(CheckForCallocOverflow(size, nmemb)))
+    return SetErrnoOnNull(Allocator::FailureHandler::OnBadRequest());
+  return SetErrnoOnNull(HwasanAllocate(stack, nmemb * size, sizeof(u64), true));
+}
+
+void *hwasan_realloc(void *ptr, uptr size, StackTrace *stack) {
+  if (!ptr)
+    return SetErrnoOnNull(HwasanAllocate(stack, size, sizeof(u64), false));
+  if (size == 0) {
+    HwasanDeallocate(stack, ptr);
+    return nullptr;
+  }
+  return SetErrnoOnNull(HwasanReallocate(stack, ptr, size, sizeof(u64)));
+}
+
+void *hwasan_valloc(uptr size, StackTrace *stack) {
+  return SetErrnoOnNull(HwasanAllocate(stack, size, GetPageSizeCached(), false));
+}
+
+void *hwasan_pvalloc(uptr size, StackTrace *stack) {
+  uptr PageSize = GetPageSizeCached();
+  if (UNLIKELY(CheckForPvallocOverflow(size, PageSize))) {
+    errno = errno_ENOMEM;
+    return Allocator::FailureHandler::OnBadRequest();
+  }
+  // pvalloc(0) should allocate one page.
+  size = size ? RoundUpTo(size, PageSize) : PageSize;
+  return SetErrnoOnNull(HwasanAllocate(stack, size, PageSize, false));
+}
+
+void *hwasan_aligned_alloc(uptr alignment, uptr size, StackTrace *stack) {
+  if (UNLIKELY(!CheckAlignedAllocAlignmentAndSize(alignment, size))) {
+    errno = errno_EINVAL;
+    return Allocator::FailureHandler::OnBadRequest();
+  }
+  return SetErrnoOnNull(HwasanAllocate(stack, size, alignment, false));
+}
+
+void *hwasan_memalign(uptr alignment, uptr size, StackTrace *stack) {
+  if (UNLIKELY(!IsPowerOfTwo(alignment))) {
+    errno = errno_EINVAL;
+    return Allocator::FailureHandler::OnBadRequest();
+  }
+  return SetErrnoOnNull(HwasanAllocate(stack, size, alignment, false));
+}
+
+int hwasan_posix_memalign(void **memptr, uptr alignment, uptr size,
+                        StackTrace *stack) {
+  if (UNLIKELY(!CheckPosixMemalignAlignment(alignment))) {
+    Allocator::FailureHandler::OnBadRequest();
+    return errno_EINVAL;
+  }
+  void *ptr = HwasanAllocate(stack, size, alignment, false);
+  if (UNLIKELY(!ptr))
+    return errno_ENOMEM;
+  CHECK(IsAligned((uptr)ptr, alignment));
+  *memptr = ptr;
+  return 0;
+}
+
+} // namespace __hwasan
+
+using namespace __hwasan;
+
+void __hwasan_enable_allocator_tagging() {
+  atomic_store_relaxed(&hwasan_allocator_tagging_enabled, 1);
+}
+
+void __hwasan_disable_allocator_tagging() {
+  atomic_store_relaxed(&hwasan_allocator_tagging_enabled, 0);
+}
+
+uptr __sanitizer_get_current_allocated_bytes() {
+  uptr stats[AllocatorStatCount];
+  allocator.GetStats(stats);
+  return stats[AllocatorStatAllocated];
+}
+
+uptr __sanitizer_get_heap_size() {
+  uptr stats[AllocatorStatCount];
+  allocator.GetStats(stats);
+  return stats[AllocatorStatMapped];
+}
+
+uptr __sanitizer_get_free_bytes() { return 1; }
+
+uptr __sanitizer_get_unmapped_bytes() { return 1; }
+
+uptr __sanitizer_get_estimated_allocated_size(uptr size) { return size; }
+
+int __sanitizer_get_ownership(const void *p) { return AllocationSize(p) != 0; }
+
+uptr __sanitizer_get_allocated_size(const void *p) { return AllocationSize(p); }
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_allocator.h b/contrib/compiler-rt/lib/hwasan/hwasan_allocator.h
new file mode 100644
index 000000000000..d025112e9773
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_allocator.h
@@ -0,0 +1,55 @@
+//===-- hwasan_allocator.h ----------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef HWASAN_ALLOCATOR_H
+#define HWASAN_ALLOCATOR_H
+
+#include "sanitizer_common/sanitizer_common.h"
+
+namespace __hwasan {
+
+struct HwasanThreadLocalMallocStorage {
+  uptr quarantine_cache[16];
+  // Allocator cache contains atomic_uint64_t which must be 8-byte aligned.
+  ALIGNED(8) uptr allocator_cache[96 * (512 * 8 + 16)];  // Opaque.
+  void CommitBack();
+
+ private:
+  // These objects are allocated via mmap() and are zero-initialized.
+  HwasanThreadLocalMallocStorage() {}
+};
+
+struct Metadata;
+
+class HwasanChunkView {
+ public:
+  HwasanChunkView() : block_(0), metadata_(nullptr) {}
+  HwasanChunkView(uptr block, Metadata *metadata)
+      : block_(block), metadata_(metadata) {}
+  bool IsValid() const;        // Checks if it points to a valid allocated chunk
+  bool IsAllocated() const;    // Checks if the memory is currently allocated
+  uptr Beg() const;            // First byte of user memory
+  uptr End() const;            // Last byte of user memory
+  uptr UsedSize() const;       // Size requested by the user
+  u32 GetAllocStackId() const;
+  u32 GetFreeStackId() const;
+ private:
+  uptr block_;
+  Metadata *const metadata_;
+};
+
+HwasanChunkView FindHeapChunkByAddress(uptr address);
+
+} // namespace __hwasan
+
+#endif // HWASAN_ALLOCATOR_H
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_blacklist.txt b/contrib/compiler-rt/lib/hwasan/hwasan_blacklist.txt
new file mode 100644
index 000000000000..395ba28f0212
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_blacklist.txt
@@ -0,0 +1,7 @@
+# Blacklist for HWAddressSanitizer. Turns off instrumentation of particular
+# functions or sources. Use with care. You may set location of blacklist
+# at compile-time using -fsanitize-blacklist=<path> flag.
+
+# Example usage:
+# fun:*bad_function_name*
+# src:file_with_tricky_code.cc
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_flags.h b/contrib/compiler-rt/lib/hwasan/hwasan_flags.h
new file mode 100644
index 000000000000..16d60c4d8ba5
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_flags.h
@@ -0,0 +1,30 @@
+//===-- hwasan_flags.h --------------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+//===----------------------------------------------------------------------===//
+#ifndef HWASAN_FLAGS_H
+#define HWASAN_FLAGS_H
+
+namespace __hwasan {
+
+struct Flags {
+#define HWASAN_FLAG(Type, Name, DefaultValue, Description) Type Name;
+#include "hwasan_flags.inc"
+#undef HWASAN_FLAG
+
+  void SetDefaults();
+};
+
+Flags *flags();
+
+}  // namespace __hwasan
+
+#endif  // HWASAN_FLAGS_H
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_flags.inc b/contrib/compiler-rt/lib/hwasan/hwasan_flags.inc
new file mode 100644
index 000000000000..a2cb701ae93d
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_flags.inc
@@ -0,0 +1,29 @@
+//===-- hwasan_flags.inc ------------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Hwasan runtime flags.
+//
+//===----------------------------------------------------------------------===//
+#ifndef HWASAN_FLAG
+# error "Define HWASAN_FLAG prior to including this file!"
+#endif
+
+// HWASAN_FLAG(Type, Name, DefaultValue, Description)
+// See COMMON_FLAG in sanitizer_flags.inc for more details.
+
+HWASAN_FLAG(bool, tag_in_malloc, true, "")
+HWASAN_FLAG(bool, tag_in_free, true, "")
+HWASAN_FLAG(bool, retag_in_realloc, true, "")
+HWASAN_FLAG(bool, print_stats, false, "")
+HWASAN_FLAG(bool, halt_on_error, true, "")
+HWASAN_FLAG(bool, atexit, false, "")
+
+// Test only flag to disable malloc/realloc/free memory tagging on startup.
+// Tagging can be reenabled with __hwasan_enable_allocator_tagging().
+HWASAN_FLAG(bool, disable_allocator_tagging, false, "")
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_interceptors.cc b/contrib/compiler-rt/lib/hwasan/hwasan_interceptors.cc
new file mode 100644
index 000000000000..fb39e9fdacf9
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_interceptors.cc
@@ -0,0 +1,483 @@
+//===-- hwasan_interceptors.cc ----------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+// Interceptors for standard library functions.
+//
+// FIXME: move as many interceptors as possible into
+// sanitizer_common/sanitizer_common_interceptors.h
+//===----------------------------------------------------------------------===//
+
+#include "interception/interception.h"
+#include "hwasan.h"
+#include "hwasan_thread.h"
+#include "hwasan_poisoning.h"
+#include "sanitizer_common/sanitizer_platform_limits_posix.h"
+#include "sanitizer_common/sanitizer_allocator.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_libc.h"
+#include "sanitizer_common/sanitizer_linux.h"
+#include "sanitizer_common/sanitizer_tls_get_addr.h"
+
+#include <stdarg.h>
+// ACHTUNG! No other system header includes in this file.
+// Ideally, we should get rid of stdarg.h as well.
+
+using namespace __hwasan;
+
+using __sanitizer::memory_order;
+using __sanitizer::atomic_load;
+using __sanitizer::atomic_store;
+using __sanitizer::atomic_uintptr_t;
+
+DECLARE_REAL(SIZE_T, strlen, const char *s)
+DECLARE_REAL(SIZE_T, strnlen, const char *s, SIZE_T maxlen)
+DECLARE_REAL(void *, memcpy, void *dest, const void *src, uptr n)
+DECLARE_REAL(void *, memset, void *dest, int c, uptr n)
+
+bool IsInInterceptorScope() {
+  HwasanThread *t = GetCurrentThread();
+  return t && t->InInterceptorScope();
+}
+
+struct InterceptorScope {
+  InterceptorScope() {
+    HwasanThread *t = GetCurrentThread();
+    if (t)
+      t->EnterInterceptorScope();
+  }
+  ~InterceptorScope() {
+    HwasanThread *t = GetCurrentThread();
+    if (t)
+      t->LeaveInterceptorScope();
+  }
+};
+
+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;
+}
+
+#define ENSURE_HWASAN_INITED() do { \
+  CHECK(!hwasan_init_is_running); \
+  if (!hwasan_inited) { \
+    __hwasan_init(); \
+  } \
+} while (0)
+
+
+
+#define HWASAN_READ_RANGE(ctx, offset, size) \
+  CHECK_UNPOISONED(offset, size)
+#define HWASAN_WRITE_RANGE(ctx, offset, size) \
+  CHECK_UNPOISONED(offset, size)
+
+
+
+// Check that [x, x+n) range is unpoisoned.
+#define CHECK_UNPOISONED_0(x, n)                                       \
+  do {                                                                 \
+    sptr __offset = __hwasan_test_shadow(x, n);                         \
+    if (__hwasan::IsInSymbolizer()) break;                              \
+    if (__offset >= 0) {                                               \
+      GET_CALLER_PC_BP_SP;                                             \
+      (void)sp;                                                        \
+      ReportInvalidAccessInsideAddressRange(__func__, x, n, __offset); \
+      __hwasan::PrintWarning(pc, bp);                                   \
+      if (__hwasan::flags()->halt_on_error) {                           \
+        Printf("Exiting\n");                                           \
+        Die();                                                         \
+      }                                                                \
+    }                                                                  \
+  } while (0)
+
+// Check that [x, x+n) range is unpoisoned unless we are in a nested
+// interceptor.
+#define CHECK_UNPOISONED(x, n)                             \
+  do {                                                     \
+    if (!IsInInterceptorScope()) CHECK_UNPOISONED_0(x, n); \
+  } while (0)
+
+#define CHECK_UNPOISONED_STRING_OF_LEN(x, len, n)               \
+  CHECK_UNPOISONED((x),                                         \
+    common_flags()->strict_string_checks ? (len) + 1 : (n) )
+
+
+INTERCEPTOR(int, posix_memalign, void **memptr, SIZE_T alignment, SIZE_T size) {
+  GET_MALLOC_STACK_TRACE;
+  CHECK_NE(memptr, 0);
+  int res = hwasan_posix_memalign(memptr, alignment, size, &stack);
+  return res;
+}
+
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
+INTERCEPTOR(void *, memalign, SIZE_T alignment, SIZE_T size) {
+  GET_MALLOC_STACK_TRACE;
+  return hwasan_memalign(alignment, size, &stack);
+}
+#define HWASAN_MAYBE_INTERCEPT_MEMALIGN INTERCEPT_FUNCTION(memalign)
+#else
+#define HWASAN_MAYBE_INTERCEPT_MEMALIGN
+#endif
+
+INTERCEPTOR(void *, aligned_alloc, SIZE_T alignment, SIZE_T size) {
+  GET_MALLOC_STACK_TRACE;
+  return hwasan_aligned_alloc(alignment, size, &stack);
+}
+
+INTERCEPTOR(void *, __libc_memalign, SIZE_T alignment, SIZE_T size) {
+  GET_MALLOC_STACK_TRACE;
+  void *ptr = hwasan_memalign(alignment, size, &stack);
+  if (ptr)
+    DTLS_on_libc_memalign(ptr, size);
+  return ptr;
+}
+
+INTERCEPTOR(void *, valloc, SIZE_T size) {
+  GET_MALLOC_STACK_TRACE;
+  return hwasan_valloc(size, &stack);
+}
+
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
+INTERCEPTOR(void *, pvalloc, SIZE_T size) {
+  GET_MALLOC_STACK_TRACE;
+  return hwasan_pvalloc(size, &stack);
+}
+#define HWASAN_MAYBE_INTERCEPT_PVALLOC INTERCEPT_FUNCTION(pvalloc)
+#else
+#define HWASAN_MAYBE_INTERCEPT_PVALLOC
+#endif
+
+INTERCEPTOR(void, free, void *ptr) {
+  GET_MALLOC_STACK_TRACE;
+  if (!ptr || UNLIKELY(IsInDlsymAllocPool(ptr))) return;
+  HwasanDeallocate(&stack, ptr);
+}
+
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
+INTERCEPTOR(void, cfree, void *ptr) {
+  GET_MALLOC_STACK_TRACE;
+  if (!ptr || UNLIKELY(IsInDlsymAllocPool(ptr))) return;
+  HwasanDeallocate(&stack, ptr);
+}
+#define HWASAN_MAYBE_INTERCEPT_CFREE INTERCEPT_FUNCTION(cfree)
+#else
+#define HWASAN_MAYBE_INTERCEPT_CFREE
+#endif
+
+INTERCEPTOR(uptr, malloc_usable_size, void *ptr) {
+  return __sanitizer_get_allocated_size(ptr);
+}
+
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
+// This function actually returns a struct by value, but we can't unpoison a
+// temporary! The following is equivalent on all supported platforms but
+// aarch64 (which uses a different register for sret value).  We have a test
+// to confirm that.
+INTERCEPTOR(void, mallinfo, __sanitizer_mallinfo *sret) {
+#ifdef __aarch64__
+  uptr r8;
+  asm volatile("mov %0,x8" : "=r" (r8));
+  sret = reinterpret_cast<__sanitizer_mallinfo*>(r8);
+#endif
+  REAL(memset)(sret, 0, sizeof(*sret));
+}
+#define HWASAN_MAYBE_INTERCEPT_MALLINFO INTERCEPT_FUNCTION(mallinfo)
+#else
+#define HWASAN_MAYBE_INTERCEPT_MALLINFO
+#endif
+
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
+INTERCEPTOR(int, mallopt, int cmd, int value) {
+  return -1;
+}
+#define HWASAN_MAYBE_INTERCEPT_MALLOPT INTERCEPT_FUNCTION(mallopt)
+#else
+#define HWASAN_MAYBE_INTERCEPT_MALLOPT
+#endif
+
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
+INTERCEPTOR(void, malloc_stats, void) {
+  // FIXME: implement, but don't call REAL(malloc_stats)!
+}
+#define HWASAN_MAYBE_INTERCEPT_MALLOC_STATS INTERCEPT_FUNCTION(malloc_stats)
+#else
+#define HWASAN_MAYBE_INTERCEPT_MALLOC_STATS
+#endif
+
+
+INTERCEPTOR(void *, calloc, SIZE_T nmemb, SIZE_T 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);
+}
+
+INTERCEPTOR(void *, realloc, void *ptr, SIZE_T 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);
+}
+
+INTERCEPTOR(void *, malloc, SIZE_T size) {
+  GET_MALLOC_STACK_TRACE;
+  if (UNLIKELY(!hwasan_inited))
+    // Hack: dlsym calls malloc before REAL(malloc) is retrieved from dlsym.
+    return AllocateFromLocalPool(size);
+  return hwasan_malloc(size, &stack);
+}
+
+
+INTERCEPTOR(void *, mmap, void *addr, SIZE_T length, int prot, int flags,
+            int fd, OFF_T offset) {
+  if (hwasan_init_is_running)
+    return REAL(mmap)(addr, length, prot, flags, fd, offset);
+  ENSURE_HWASAN_INITED();
+  if (addr && !MEM_IS_APP(addr)) {
+    if (flags & map_fixed) {
+      errno = errno_EINVAL;
+      return (void *)-1;
+    } else {
+      addr = nullptr;
+    }
+  }
+  void *res = REAL(mmap)(addr, length, prot, flags, fd, offset);
+  return res;
+}
+
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
+INTERCEPTOR(void *, mmap64, void *addr, SIZE_T length, int prot, int flags,
+            int fd, OFF64_T offset) {
+  ENSURE_HWASAN_INITED();
+  if (addr && !MEM_IS_APP(addr)) {
+    if (flags & map_fixed) {
+      errno = errno_EINVAL;
+      return (void *)-1;
+    } else {
+      addr = nullptr;
+    }
+  }
+  void *res = REAL(mmap64)(addr, length, prot, flags, fd, offset);
+  return res;
+}
+#define HWASAN_MAYBE_INTERCEPT_MMAP64 INTERCEPT_FUNCTION(mmap64)
+#else
+#define HWASAN_MAYBE_INTERCEPT_MMAP64
+#endif
+
+extern "C" int pthread_attr_init(void *attr);
+extern "C" int pthread_attr_destroy(void *attr);
+
+static void *HwasanThreadStartFunc(void *arg) {
+  HwasanThread *t = (HwasanThread *)arg;
+  SetCurrentThread(t);
+  return t->ThreadStart();
+}
+
+INTERCEPTOR(int, pthread_create, void *th, void *attr, void *(*callback)(void*),
+            void * param) {
+  ENSURE_HWASAN_INITED(); // for GetTlsSize()
+  __sanitizer_pthread_attr_t myattr;
+  if (!attr) {
+    pthread_attr_init(&myattr);
+    attr = &myattr;
+  }
+
+  AdjustStackSize(attr);
+
+  HwasanThread *t = HwasanThread::Create(callback, param);
+
+  int res = REAL(pthread_create)(th, attr, HwasanThreadStartFunc, t);
+
+  if (attr == &myattr)
+    pthread_attr_destroy(&myattr);
+  return res;
+}
+
+static void BeforeFork() {
+  StackDepotLockAll();
+}
+
+static void AfterFork() {
+  StackDepotUnlockAll();
+}
+
+INTERCEPTOR(int, fork, void) {
+  ENSURE_HWASAN_INITED();
+  BeforeFork();
+  int pid = REAL(fork)();
+  AfterFork();
+  return pid;
+}
+
+
+struct HwasanInterceptorContext {
+  bool in_interceptor_scope;
+};
+
+namespace __hwasan {
+
+int OnExit() {
+  // FIXME: ask frontend whether we need to return failure.
+  return 0;
+}
+
+} // namespace __hwasan
+
+// A version of CHECK_UNPOISONED using a saved scope value. Used in common
+// interceptors.
+#define CHECK_UNPOISONED_CTX(ctx, x, n)                         \
+  do {                                                          \
+    if (!((HwasanInterceptorContext *)ctx)->in_interceptor_scope) \
+      CHECK_UNPOISONED_0(x, n);                                 \
+  } while (0)
+
+#define HWASAN_INTERCEPT_FUNC(name)                                       \
+  do {                                                                  \
+    if ((!INTERCEPT_FUNCTION(name) || !REAL(name)))                     \
+      VReport(1, "HWAddressSanitizer: failed to intercept '" #name "'\n"); \
+  } while (0)
+
+#define HWASAN_INTERCEPT_FUNC_VER(name, ver)                                    \
+  do {                                                                        \
+    if ((!INTERCEPT_FUNCTION_VER(name, ver) || !REAL(name)))                  \
+      VReport(                                                                \
+          1, "HWAddressSanitizer: failed to intercept '" #name "@@" #ver "'\n"); \
+  } while (0)
+
+#define COMMON_INTERCEPT_FUNCTION(name) HWASAN_INTERCEPT_FUNC(name)
+#define COMMON_INTERCEPT_FUNCTION_VER(name, ver)                          \
+  HWASAN_INTERCEPT_FUNC_VER(name, ver)
+#define COMMON_INTERCEPTOR_WRITE_RANGE(ctx, ptr, size) \
+  CHECK_UNPOISONED_CTX(ctx, ptr, size)
+#define COMMON_INTERCEPTOR_READ_RANGE(ctx, ptr, size) \
+  CHECK_UNPOISONED_CTX(ctx, ptr, size)
+#define COMMON_INTERCEPTOR_INITIALIZE_RANGE(ptr, size) \
+  HWASAN_WRITE_RANGE(ctx, ptr, size)
+#define COMMON_INTERCEPTOR_ENTER(ctx, func, ...)                  \
+  if (hwasan_init_is_running) return REAL(func)(__VA_ARGS__);       \
+  ENSURE_HWASAN_INITED();                                           \
+  HwasanInterceptorContext hwasan_ctx = {IsInInterceptorScope()};     \
+  ctx = (void *)&hwasan_ctx;                                        \
+  (void)ctx;                                                      \
+  InterceptorScope interceptor_scope;
+#define COMMON_INTERCEPTOR_DIR_ACQUIRE(ctx, path) \
+  do {                                            \
+  } while (false)
+#define COMMON_INTERCEPTOR_FD_ACQUIRE(ctx, fd) \
+  do {                                         \
+  } while (false)
+#define COMMON_INTERCEPTOR_FD_RELEASE(ctx, fd) \
+  do {                                         \
+  } while (false)
+#define COMMON_INTERCEPTOR_FD_SOCKET_ACCEPT(ctx, fd, newfd) \
+  do {                                                      \
+  } while (false)
+#define COMMON_INTERCEPTOR_SET_THREAD_NAME(ctx, name) \
+  do {                                                \
+  } while (false)  // FIXME
+#define COMMON_INTERCEPTOR_SET_PTHREAD_NAME(ctx, thread, name) \
+  do {                                                         \
+  } while (false)  // FIXME
+#define COMMON_INTERCEPTOR_BLOCK_REAL(name) REAL(name)
+#define COMMON_INTERCEPTOR_ON_EXIT(ctx) OnExit()
+
+#define COMMON_INTERCEPTOR_GET_TLS_RANGE(begin, end)                           \
+  if (HwasanThread *t = GetCurrentThread()) {                                    \
+    *begin = t->tls_begin();                                                   \
+    *end = t->tls_end();                                                       \
+  } else {                                                                     \
+    *begin = *end = 0;                                                         \
+  }
+
+#include "sanitizer_common/sanitizer_platform_interceptors.h"
+#include "sanitizer_common/sanitizer_common_interceptors.inc"
+#include "sanitizer_common/sanitizer_signal_interceptors.inc"
+
+#define COMMON_SYSCALL_PRE_READ_RANGE(p, s) CHECK_UNPOISONED(p, s)
+#define COMMON_SYSCALL_PRE_WRITE_RANGE(p, s) \
+  do {                                       \
+    (void)(p);                               \
+    (void)(s);                               \
+  } while (false)
+#define COMMON_SYSCALL_POST_READ_RANGE(p, s) \
+  do {                                       \
+    (void)(p);                               \
+    (void)(s);                               \
+  } while (false)
+#define COMMON_SYSCALL_POST_WRITE_RANGE(p, s) \
+  do {                                        \
+    (void)(p);                                \
+    (void)(s);                                \
+  } while (false)
+#include "sanitizer_common/sanitizer_common_syscalls.inc"
+
+
+
+namespace __hwasan {
+
+void InitializeInterceptors() {
+  static int inited = 0;
+  CHECK_EQ(inited, 0);
+  InitializeCommonInterceptors();
+  InitializeSignalInterceptors();
+
+  INTERCEPT_FUNCTION(mmap);
+  HWASAN_MAYBE_INTERCEPT_MMAP64;
+  INTERCEPT_FUNCTION(posix_memalign);
+  HWASAN_MAYBE_INTERCEPT_MEMALIGN;
+  INTERCEPT_FUNCTION(__libc_memalign);
+  INTERCEPT_FUNCTION(valloc);
+  HWASAN_MAYBE_INTERCEPT_PVALLOC;
+  INTERCEPT_FUNCTION(malloc);
+  INTERCEPT_FUNCTION(calloc);
+  INTERCEPT_FUNCTION(realloc);
+  INTERCEPT_FUNCTION(free);
+  HWASAN_MAYBE_INTERCEPT_CFREE;
+  INTERCEPT_FUNCTION(malloc_usable_size);
+  HWASAN_MAYBE_INTERCEPT_MALLINFO;
+  HWASAN_MAYBE_INTERCEPT_MALLOPT;
+  HWASAN_MAYBE_INTERCEPT_MALLOC_STATS;
+  INTERCEPT_FUNCTION(pthread_create);
+  INTERCEPT_FUNCTION(fork);
+
+  inited = 1;
+}
+} // namespace __hwasan
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_interface_internal.h b/contrib/compiler-rt/lib/hwasan/hwasan_interface_internal.h
new file mode 100644
index 000000000000..08b77534e0ae
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_interface_internal.h
@@ -0,0 +1,97 @@
+//===-- hwasan_interface_internal.h -------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+// Private Hwasan interface header.
+//===----------------------------------------------------------------------===//
+
+#ifndef HWASAN_INTERFACE_INTERNAL_H
+#define HWASAN_INTERFACE_INTERNAL_H
+
+#include "sanitizer_common/sanitizer_internal_defs.h"
+
+extern "C" {
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_init();
+
+using __sanitizer::uptr;
+using __sanitizer::sptr;
+using __sanitizer::uu64;
+using __sanitizer::uu32;
+using __sanitizer::uu16;
+using __sanitizer::u64;
+using __sanitizer::u32;
+using __sanitizer::u16;
+using __sanitizer::u8;
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_load(uptr, uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_load1(uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_load2(uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_load4(uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_load8(uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_load16(uptr);
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_store(uptr, uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_store1(uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_store2(uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_store4(uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_store8(uptr);
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_store16(uptr);
+
+// Returns the offset of the first tag mismatch or -1 if the whole range is
+// good.
+SANITIZER_INTERFACE_ATTRIBUTE
+sptr __hwasan_test_shadow(const void *x, uptr size);
+
+SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
+/* OPTIONAL */ const char* __hwasan_default_options();
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_print_shadow(const void *x, uptr size);
+
+SANITIZER_INTERFACE_ATTRIBUTE
+u16 __sanitizer_unaligned_load16(const uu16 *p);
+
+SANITIZER_INTERFACE_ATTRIBUTE
+u32 __sanitizer_unaligned_load32(const uu32 *p);
+
+SANITIZER_INTERFACE_ATTRIBUTE
+u64 __sanitizer_unaligned_load64(const uu64 *p);
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __sanitizer_unaligned_store16(uu16 *p, u16 x);
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __sanitizer_unaligned_store32(uu32 *p, u32 x);
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __sanitizer_unaligned_store64(uu64 *p, u64 x);
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_enable_allocator_tagging();
+
+SANITIZER_INTERFACE_ATTRIBUTE
+void __hwasan_disable_allocator_tagging();
+
+}  // extern "C"
+
+#endif  // HWASAN_INTERFACE_INTERNAL_H
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_linux.cc b/contrib/compiler-rt/lib/hwasan/hwasan_linux.cc
new file mode 100644
index 000000000000..9b8613171cbd
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_linux.cc
@@ -0,0 +1,251 @@
+//===-- hwasan_linux.cc -----------------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+// Linux-, NetBSD- and FreeBSD-specific code.
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_common/sanitizer_platform.h"
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD
+
+#include "hwasan.h"
+#include "hwasan_thread.h"
+
+#include <elf.h>
+#include <link.h>
+#include <pthread.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <unistd.h>
+#include <unwind.h>
+#include <sys/time.h>
+#include <sys/resource.h>
+
+#include "sanitizer_common/sanitizer_common.h"
+#include "sanitizer_common/sanitizer_procmaps.h"
+
+namespace __hwasan {
+
+void ReserveShadowMemoryRange(uptr beg, uptr end, const char *name) {
+  CHECK_EQ((beg % GetMmapGranularity()), 0);
+  CHECK_EQ(((end + 1) % GetMmapGranularity()), 0);
+  uptr size = end - beg + 1;
+  DecreaseTotalMmap(size);  // Don't count the shadow against mmap_limit_mb.
+  void *res = MmapFixedNoReserve(beg, size, name);
+  if (res != (void *)beg) {
+    Report(
+        "ReserveShadowMemoryRange failed while trying to map 0x%zx bytes. "
+        "Perhaps you're using ulimit -v\n",
+        size);
+    Abort();
+  }
+  if (common_flags()->no_huge_pages_for_shadow) NoHugePagesInRegion(beg, size);
+  if (common_flags()->use_madv_dontdump) DontDumpShadowMemory(beg, size);
+}
+
+static void ProtectGap(uptr addr, uptr size) {
+  void *res = MmapFixedNoAccess(addr, size, "shadow gap");
+  if (addr == (uptr)res) return;
+  // A few pages at the start of the address space can not be protected.
+  // But we really want to protect as much as possible, to prevent this memory
+  // being returned as a result of a non-FIXED mmap().
+  if (addr == 0) {
+    uptr step = GetMmapGranularity();
+    while (size > step) {
+      addr += step;
+      size -= step;
+      void *res = MmapFixedNoAccess(addr, size, "shadow gap");
+      if (addr == (uptr)res) return;
+    }
+  }
+
+  Report(
+      "ERROR: Failed to protect the shadow gap. "
+      "ASan cannot proceed correctly. ABORTING.\n");
+  DumpProcessMap();
+  Die();
+}
+
+bool InitShadow() {
+  const uptr maxVirtualAddress = GetMaxUserVirtualAddress();
+
+  // LowMem covers as much of the first 4GB as possible.
+  const uptr kLowMemEnd = 1UL<<32;
+  const uptr kLowShadowEnd = kLowMemEnd >> kShadowScale;
+  const uptr kLowShadowStart = kLowShadowEnd >> kShadowScale;
+
+  // HighMem covers the upper part of the address space.
+  const uptr kHighShadowEnd = (maxVirtualAddress >> kShadowScale) + 1;
+  const uptr kHighShadowStart = Max(kLowMemEnd, kHighShadowEnd >> kShadowScale);
+  CHECK(kHighShadowStart < kHighShadowEnd);
+
+  const uptr kHighMemStart = kHighShadowStart << kShadowScale;
+  CHECK(kHighShadowEnd <= kHighMemStart);
+
+  if (Verbosity()) {
+    Printf("|| `[%p, %p]` || HighMem    ||\n", (void *)kHighMemStart,
+           (void *)maxVirtualAddress);
+    if (kHighMemStart > kHighShadowEnd)
+      Printf("|| `[%p, %p]` || ShadowGap2 ||\n", (void *)kHighShadowEnd,
+             (void *)kHighMemStart);
+    Printf("|| `[%p, %p]` || HighShadow ||\n", (void *)kHighShadowStart,
+           (void *)kHighShadowEnd);
+    if (kHighShadowStart > kLowMemEnd)
+      Printf("|| `[%p, %p]` || ShadowGap2 ||\n", (void *)kHighShadowEnd,
+             (void *)kHighMemStart);
+    Printf("|| `[%p, %p]` || LowMem     ||\n", (void *)kLowShadowEnd,
+           (void *)kLowMemEnd);
+    Printf("|| `[%p, %p]` || LowShadow  ||\n", (void *)kLowShadowStart,
+           (void *)kLowShadowEnd);
+    Printf("|| `[%p, %p]` || ShadowGap1 ||\n", (void *)0,
+           (void *)kLowShadowStart);
+  }
+
+  ReserveShadowMemoryRange(kLowShadowStart, kLowShadowEnd - 1, "low shadow");
+  ReserveShadowMemoryRange(kHighShadowStart, kHighShadowEnd - 1, "high shadow");
+  ProtectGap(0, kLowShadowStart);
+  if (kHighShadowStart > kLowMemEnd)
+    ProtectGap(kLowMemEnd, kHighShadowStart - kLowMemEnd);
+  if (kHighMemStart > kHighShadowEnd)
+    ProtectGap(kHighShadowEnd, kHighMemStart - kHighShadowEnd);
+
+  return true;
+}
+
+static void HwasanAtExit(void) {
+  if (flags()->print_stats && (flags()->atexit || hwasan_report_count > 0))
+    ReportStats();
+  if (hwasan_report_count > 0) {
+    // ReportAtExitStatistics();
+    if (common_flags()->exitcode)
+      internal__exit(common_flags()->exitcode);
+  }
+}
+
+void InstallAtExitHandler() {
+  atexit(HwasanAtExit);
+}
+
+// ---------------------- TSD ---------------- {{{1
+
+static pthread_key_t tsd_key;
+static bool tsd_key_inited = false;
+
+void HwasanTSDInit(void (*destructor)(void *tsd)) {
+  CHECK(!tsd_key_inited);
+  tsd_key_inited = true;
+  CHECK_EQ(0, pthread_key_create(&tsd_key, destructor));
+}
+
+HwasanThread *GetCurrentThread() {
+  return (HwasanThread*)pthread_getspecific(tsd_key);
+}
+
+void SetCurrentThread(HwasanThread *t) {
+  // Make sure that HwasanTSDDtor gets called at the end.
+  CHECK(tsd_key_inited);
+  // Make sure we do not reset the current HwasanThread.
+  CHECK_EQ(0, pthread_getspecific(tsd_key));
+  pthread_setspecific(tsd_key, (void *)t);
+}
+
+void HwasanTSDDtor(void *tsd) {
+  HwasanThread *t = (HwasanThread*)tsd;
+  if (t->destructor_iterations_ > 1) {
+    t->destructor_iterations_--;
+    CHECK_EQ(0, pthread_setspecific(tsd_key, tsd));
+    return;
+  }
+  // Make sure that signal handler can not see a stale current thread pointer.
+  atomic_signal_fence(memory_order_seq_cst);
+  HwasanThread::TSDDtor(tsd);
+}
+
+struct AccessInfo {
+  uptr addr;
+  uptr size;
+  bool is_store;
+  bool is_load;
+};
+
+#if defined(__aarch64__)
+static AccessInfo GetAccessInfo(siginfo_t *info, ucontext_t *uc) {
+  // Access type is encoded in HLT immediate as 0x1XY,
+  // where X is 1 for store, 0 for load.
+  // Valid values of Y are 0 to 4, which are interpreted as log2(access_size),
+  // and 0xF, which means that access size is stored in X1 register.
+  // Access address is always in X0 register.
+  AccessInfo ai;
+  uptr pc = (uptr)info->si_addr;
+  unsigned code = ((*(u32 *)pc) >> 5) & 0xffff;
+  if ((code & 0xff00) != 0x100)
+    return AccessInfo{0, 0, false, false}; // Not ours.
+  bool is_store = code & 0x10;
+  unsigned size_log = code & 0xff;
+  if (size_log > 4 && size_log != 0xf)
+    return AccessInfo{0, 0, false, false}; // Not ours.
+
+  ai.is_store = is_store;
+  ai.is_load = !is_store;
+  ai.addr = uc->uc_mcontext.regs[0];
+  if (size_log == 0xf)
+    ai.size = uc->uc_mcontext.regs[1];
+  else
+    ai.size = 1U << size_log;
+  return ai;
+}
+#else
+static AccessInfo GetAccessInfo(siginfo_t *info, ucontext_t *uc) {
+  return AccessInfo{0, 0, false, false};
+}
+#endif
+
+static bool HwasanOnSIGILL(int signo, siginfo_t *info, ucontext_t *uc) {
+  SignalContext sig{info, uc};
+  AccessInfo ai = GetAccessInfo(info, uc);
+  if (!ai.is_store && !ai.is_load)
+    return false;
+
+  InternalScopedBuffer<BufferedStackTrace> stack_buffer(1);
+  BufferedStackTrace *stack = stack_buffer.data();
+  stack->Reset();
+  GetStackTrace(stack, kStackTraceMax, sig.pc, sig.bp, uc,
+                common_flags()->fast_unwind_on_fatal);
+
+  ReportTagMismatch(stack, ai.addr, ai.size, ai.is_store);
+
+  ++hwasan_report_count;
+  if (flags()->halt_on_error)
+    Die();
+
+  uc->uc_mcontext.pc += 4;
+  return true;
+}
+
+static void OnStackUnwind(const SignalContext &sig, const void *,
+                          BufferedStackTrace *stack) {
+  GetStackTrace(stack, kStackTraceMax, sig.pc, sig.bp, sig.context,
+                common_flags()->fast_unwind_on_fatal);
+}
+
+void HwasanOnDeadlySignal(int signo, void *info, void *context) {
+  // Probably a tag mismatch.
+  if (signo == SIGILL)
+    if (HwasanOnSIGILL(signo, (siginfo_t *)info, (ucontext_t*)context))
+      return;
+
+  HandleDeadlySignal(info, context, GetTid(), &OnStackUnwind, nullptr);
+}
+
+
+} // namespace __hwasan
+
+#endif // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_new_delete.cc b/contrib/compiler-rt/lib/hwasan/hwasan_new_delete.cc
new file mode 100644
index 000000000000..3ccc26734bf6
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_new_delete.cc
@@ -0,0 +1,66 @@
+//===-- hwasan_new_delete.cc ------------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+// Interceptors for operators new and delete.
+//===----------------------------------------------------------------------===//
+
+#include "hwasan.h"
+#include "interception/interception.h"
+#include "sanitizer_common/sanitizer_allocator.h"
+
+#if HWASAN_REPLACE_OPERATORS_NEW_AND_DELETE
+
+#include <stddef.h>
+
+using namespace __hwasan;  // NOLINT
+
+// Fake std::nothrow_t to avoid including <new>.
+namespace std {
+  struct nothrow_t {};
+}  // namespace std
+
+
+// TODO(alekseys): throw std::bad_alloc instead of dying on OOM.
+#define OPERATOR_NEW_BODY(nothrow) \
+  GET_MALLOC_STACK_TRACE; \
+  void *res = hwasan_malloc(size, &stack);\
+  if (!nothrow && UNLIKELY(!res)) DieOnFailure::OnOOM();\
+  return res
+
+INTERCEPTOR_ATTRIBUTE
+void *operator new(size_t size) { OPERATOR_NEW_BODY(false /*nothrow*/); }
+INTERCEPTOR_ATTRIBUTE
+void *operator new[](size_t size) { OPERATOR_NEW_BODY(false /*nothrow*/); }
+INTERCEPTOR_ATTRIBUTE
+void *operator new(size_t size, std::nothrow_t const&) {
+  OPERATOR_NEW_BODY(true /*nothrow*/);
+}
+INTERCEPTOR_ATTRIBUTE
+void *operator new[](size_t size, std::nothrow_t const&) {
+  OPERATOR_NEW_BODY(true /*nothrow*/);
+}
+
+#define OPERATOR_DELETE_BODY \
+  GET_MALLOC_STACK_TRACE; \
+  if (ptr) HwasanDeallocate(&stack, ptr)
+
+INTERCEPTOR_ATTRIBUTE
+void operator delete(void *ptr) NOEXCEPT { OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete[](void *ptr) NOEXCEPT { OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete(void *ptr, std::nothrow_t const&) { OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete[](void *ptr, std::nothrow_t const&) {
+  OPERATOR_DELETE_BODY;
+}
+
+#endif // HWASAN_REPLACE_OPERATORS_NEW_AND_DELETE
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_poisoning.cc b/contrib/compiler-rt/lib/hwasan/hwasan_poisoning.cc
new file mode 100644
index 000000000000..411fd05b10d5
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_poisoning.cc
@@ -0,0 +1,36 @@
+//===-- hwasan_poisoning.cc ---------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+//===----------------------------------------------------------------------===//
+
+#include "hwasan_poisoning.h"
+
+#include "interception/interception.h"
+#include "sanitizer_common/sanitizer_common.h"
+
+namespace __hwasan {
+
+uptr TagMemoryAligned(uptr p, uptr size, tag_t tag) {
+  CHECK(IsAligned(p, kShadowAlignment));
+  CHECK(IsAligned(size, kShadowAlignment));
+  uptr shadow_start = MEM_TO_SHADOW(p);
+  uptr shadow_size = MEM_TO_SHADOW_OFFSET(size);
+  internal_memset((void *)shadow_start, tag, shadow_size);
+  return AddTagToPointer(p, tag);
+}
+
+uptr TagMemory(uptr p, uptr size, tag_t tag) {
+  uptr start = RoundDownTo(p, kShadowAlignment);
+  uptr end = RoundUpTo(p + size, kShadowAlignment);
+  return TagMemoryAligned(start, end - start, tag);
+}
+
+}  // namespace __hwasan
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_poisoning.h b/contrib/compiler-rt/lib/hwasan/hwasan_poisoning.h
new file mode 100644
index 000000000000..b44a91f975f5
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_poisoning.h
@@ -0,0 +1,25 @@
+//===-- hwasan_poisoning.h ----------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef HWASAN_POISONING_H
+#define HWASAN_POISONING_H
+
+#include "hwasan.h"
+
+namespace __hwasan {
+uptr TagMemory(uptr p, uptr size, tag_t tag);
+uptr TagMemoryAligned(uptr p, uptr size, tag_t tag);
+
+}  // namespace __hwasan
+
+#endif  // HWASAN_POISONING_H
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_report.cc b/contrib/compiler-rt/lib/hwasan/hwasan_report.cc
new file mode 100644
index 000000000000..a3c6709491d3
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_report.cc
@@ -0,0 +1,133 @@
+//===-- hwasan_report.cc ----------------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+// Error reporting.
+//===----------------------------------------------------------------------===//
+
+#include "hwasan.h"
+#include "hwasan_allocator.h"
+#include "sanitizer_common/sanitizer_allocator_internal.h"
+#include "sanitizer_common/sanitizer_common.h"
+#include "sanitizer_common/sanitizer_flags.h"
+#include "sanitizer_common/sanitizer_mutex.h"
+#include "sanitizer_common/sanitizer_report_decorator.h"
+#include "sanitizer_common/sanitizer_stackdepot.h"
+#include "sanitizer_common/sanitizer_symbolizer.h"
+
+using namespace __sanitizer;
+
+namespace __hwasan {
+
+static StackTrace GetStackTraceFromId(u32 id) {
+  CHECK(id);
+  StackTrace res = StackDepotGet(id);
+  CHECK(res.trace);
+  return res;
+}
+
+class Decorator: public __sanitizer::SanitizerCommonDecorator {
+ public:
+  Decorator() : SanitizerCommonDecorator() { }
+  const char *Allocation() { return Magenta(); }
+  const char *Origin() { return Magenta(); }
+  const char *Name() { return Green(); }
+};
+
+struct HeapAddressDescription {
+  uptr addr;
+  u32 alloc_stack_id;
+  u32 free_stack_id;
+
+  void Print() const {
+    Decorator d;
+    if (free_stack_id) {
+      Printf("%sfreed here:%s\n", d.Allocation(), d.Default());
+      GetStackTraceFromId(free_stack_id).Print();
+      Printf("%spreviously allocated here:%s\n", d.Allocation(), d.Default());
+    } else {
+      Printf("%sallocated here:%s\n", d.Allocation(), d.Default());
+    }
+    GetStackTraceFromId(alloc_stack_id).Print();
+  }
+};
+
+bool GetHeapAddressInformation(uptr addr, uptr access_size,
+                               HeapAddressDescription *description) {
+  HwasanChunkView chunk = FindHeapChunkByAddress(addr);
+  if (!chunk.IsValid())
+    return false;
+  description->addr = addr;
+  description->alloc_stack_id = chunk.GetAllocStackId();
+  description->free_stack_id = chunk.GetFreeStackId();
+  return true;
+}
+
+void PrintAddressDescription(uptr addr, uptr access_size) {
+  HeapAddressDescription heap_description;
+  if (GetHeapAddressInformation(addr, access_size, &heap_description)) {
+    heap_description.Print();
+    return;
+  }
+  // We exhausted our possibilities. Bail out.
+  Printf("HWAddressSanitizer can not describe address in more detail.\n");
+}
+
+void ReportInvalidAccess(StackTrace *stack, u32 origin) {
+  ScopedErrorReportLock l;
+
+  Decorator d;
+  Printf("%s", d.Warning());
+  Report("WARNING: HWAddressSanitizer: invalid access\n");
+  Printf("%s", d.Default());
+  stack->Print();
+  ReportErrorSummary("invalid-access", stack);
+}
+
+void ReportStats() {}
+
+void ReportInvalidAccessInsideAddressRange(const char *what, const void *start,
+                                           uptr size, uptr offset) {
+  ScopedErrorReportLock l;
+
+  Decorator d;
+  Printf("%s", d.Warning());
+  Printf("%sTag mismatch in %s%s%s at offset %zu inside [%p, %zu)%s\n",
+         d.Warning(), d.Name(), what, d.Warning(), offset, start, size,
+         d.Default());
+  PrintAddressDescription((uptr)start + offset, 1);
+  // if (__sanitizer::Verbosity())
+  //   DescribeMemoryRange(start, size);
+}
+
+void ReportTagMismatch(StackTrace *stack, uptr addr, uptr access_size,
+                       bool is_store) {
+  ScopedErrorReportLock l;
+
+  Decorator d;
+  Printf("%s", d.Warning());
+  uptr address = GetAddressFromPointer(addr);
+  Printf("%s of size %zu at %p\n", is_store ? "WRITE" : "READ", access_size,
+         address);
+
+  tag_t ptr_tag = GetTagFromPointer(addr);
+  tag_t mem_tag = *(tag_t *)MEM_TO_SHADOW(address);
+  Printf("pointer tag 0x%x\nmemory tag  0x%x\n", ptr_tag, mem_tag);
+  Printf("%s", d.Default());
+
+  stack->Print();
+
+  PrintAddressDescription(address, access_size);
+
+  ReportErrorSummary("tag-mismatch", stack);
+}
+
+
+}  // namespace __hwasan
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_thread.cc b/contrib/compiler-rt/lib/hwasan/hwasan_thread.cc
new file mode 100644
index 000000000000..d6551ffc6fdc
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_thread.cc
@@ -0,0 +1,75 @@
+
+#include "hwasan.h"
+#include "hwasan_thread.h"
+#include "hwasan_poisoning.h"
+#include "hwasan_interface_internal.h"
+
+#include "sanitizer_common/sanitizer_tls_get_addr.h"
+
+namespace __hwasan {
+
+HwasanThread *HwasanThread::Create(thread_callback_t start_routine,
+                               void *arg) {
+  uptr PageSize = GetPageSizeCached();
+  uptr size = RoundUpTo(sizeof(HwasanThread), PageSize);
+  HwasanThread *thread = (HwasanThread*)MmapOrDie(size, __func__);
+  thread->start_routine_ = start_routine;
+  thread->arg_ = arg;
+  thread->destructor_iterations_ = GetPthreadDestructorIterations();
+
+  return thread;
+}
+
+void HwasanThread::SetThreadStackAndTls() {
+  uptr tls_size = 0;
+  uptr stack_size = 0;
+  GetThreadStackAndTls(IsMainThread(), &stack_bottom_, &stack_size,
+                       &tls_begin_, &tls_size);
+  stack_top_ = stack_bottom_ + stack_size;
+  tls_end_ = tls_begin_ + tls_size;
+
+  int local;
+  CHECK(AddrIsInStack((uptr)&local));
+}
+
+void HwasanThread::Init() {
+  SetThreadStackAndTls();
+  CHECK(MEM_IS_APP(stack_bottom_));
+  CHECK(MEM_IS_APP(stack_top_ - 1));
+}
+
+void HwasanThread::TSDDtor(void *tsd) {
+  HwasanThread *t = (HwasanThread*)tsd;
+  t->Destroy();
+}
+
+void HwasanThread::ClearShadowForThreadStackAndTLS() {
+  TagMemory(stack_bottom_, stack_top_ - stack_bottom_, 0);
+  if (tls_begin_ != tls_end_)
+    TagMemory(tls_begin_, tls_end_ - tls_begin_, 0);
+}
+
+void HwasanThread::Destroy() {
+  malloc_storage().CommitBack();
+  ClearShadowForThreadStackAndTLS();
+  uptr size = RoundUpTo(sizeof(HwasanThread), GetPageSizeCached());
+  UnmapOrDie(this, size);
+  DTLS_Destroy();
+}
+
+thread_return_t HwasanThread::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
+    // ThreadStart() for the worker threads.
+    return 0;
+  }
+
+  thread_return_t res = start_routine_(arg_);
+
+  return res;
+}
+
+} // namespace __hwasan
diff --git a/contrib/compiler-rt/lib/hwasan/hwasan_thread.h b/contrib/compiler-rt/lib/hwasan/hwasan_thread.h
new file mode 100644
index 000000000000..96f1bb813adf
--- /dev/null
+++ b/contrib/compiler-rt/lib/hwasan/hwasan_thread.h
@@ -0,0 +1,81 @@
+//===-- hwasan_thread.h -------------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of HWAddressSanitizer.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef HWASAN_THREAD_H
+#define HWASAN_THREAD_H
+
+#include "hwasan_allocator.h"
+#include "sanitizer_common/sanitizer_common.h"
+
+namespace __hwasan {
+
+class HwasanThread {
+ public:
+  static HwasanThread *Create(thread_callback_t start_routine, void *arg);
+  static void TSDDtor(void *tsd);
+  void Destroy();
+
+  void Init();  // Should be called from the thread itself.
+  thread_return_t ThreadStart();
+
+  uptr stack_top() { return stack_top_; }
+  uptr stack_bottom() { return stack_bottom_; }
+  uptr tls_begin() { return tls_begin_; }
+  uptr tls_end() { return tls_end_; }
+  bool IsMainThread() { return start_routine_ == nullptr; }
+
+  bool AddrIsInStack(uptr addr) {
+    return addr >= stack_bottom_ && addr < stack_top_;
+  }
+
+  bool InSignalHandler() { return in_signal_handler_; }
+  void EnterSignalHandler() { in_signal_handler_++; }
+  void LeaveSignalHandler() { in_signal_handler_--; }
+
+  bool InSymbolizer() { return in_symbolizer_; }
+  void EnterSymbolizer() { in_symbolizer_++; }
+  void LeaveSymbolizer() { in_symbolizer_--; }
+
+  bool InInterceptorScope() { return in_interceptor_scope_; }
+  void EnterInterceptorScope() { in_interceptor_scope_++; }
+  void LeaveInterceptorScope() { in_interceptor_scope_--; }
+
+  HwasanThreadLocalMallocStorage &malloc_storage() { return malloc_storage_; }
+
+  int destructor_iterations_;
+
+ private:
+  // NOTE: There is no HwasanThread constructor. It is allocated
+  // via mmap() and *must* be valid in zero-initialized state.
+  void SetThreadStackAndTls();
+  void ClearShadowForThreadStackAndTLS();
+  thread_callback_t start_routine_;
+  void *arg_;
+  uptr stack_top_;
+  uptr stack_bottom_;
+  uptr tls_begin_;
+  uptr tls_end_;
+
+  unsigned in_signal_handler_;
+  unsigned in_symbolizer_;
+  unsigned in_interceptor_scope_;
+
+  HwasanThreadLocalMallocStorage malloc_storage_;
+};
+
+HwasanThread *GetCurrentThread();
+void SetCurrentThread(HwasanThread *t);
+
+} // namespace __hwasan
+
+#endif // HWASAN_THREAD_H
diff --git a/contrib/compiler-rt/lib/interception/interception.h b/contrib/compiler-rt/lib/interception/interception.h
index d79fa67babfa..cba484936eac 100644
--- a/contrib/compiler-rt/lib/interception/interception.h
+++ b/contrib/compiler-rt/lib/interception/interception.h
@@ -15,13 +15,14 @@
 #ifndef INTERCEPTION_H
 #define INTERCEPTION_H
 
-#if !defined(__linux__) && !defined(__FreeBSD__) && \
-  !defined(__APPLE__) && !defined(_WIN32)
+#include "sanitizer_common/sanitizer_internal_defs.h"
+
+#if !SANITIZER_LINUX && !SANITIZER_FREEBSD && !SANITIZER_MAC && \
+    !SANITIZER_NETBSD && !SANITIZER_WINDOWS && !SANITIZER_FUCHSIA && \
+    !SANITIZER_SOLARIS
 # error "Interception doesn't work on this operating system."
 #endif
 
-#include "sanitizer_common/sanitizer_internal_defs.h"
-
 // These typedefs should be used only in the interceptor definitions to replace
 // the standard system types (e.g. SSIZE_T instead of ssize_t)
 typedef __sanitizer::uptr    SIZE_T;
@@ -87,7 +88,7 @@ typedef __sanitizer::OFF64_T OFF64_T;
 // As it's decided at compile time which functions are to be intercepted on Mac,
 // INTERCEPT_FUNCTION() is effectively a no-op on this system.
 
-#if defined(__APPLE__)
+#if SANITIZER_MAC
 #include <sys/cdefs.h>  // For __DARWIN_ALIAS_C().
 
 // Just a pair of pointers.
@@ -121,7 +122,7 @@ const interpose_substitution substitution_##func_name[] \
 # define INTERCEPTOR_ATTRIBUTE
 # define DECLARE_WRAPPER(ret_type, func, ...)
 
-#elif defined(_WIN32)
+#elif SANITIZER_WINDOWS
 # define WRAP(x) __asan_wrap_##x
 # define WRAPPER_NAME(x) "__asan_wrap_"#x
 # define INTERCEPTOR_ATTRIBUTE __declspec(dllexport)
@@ -129,7 +130,7 @@ const interpose_substitution substitution_##func_name[] \
     extern "C" ret_type func(__VA_ARGS__);
 # define DECLARE_WRAPPER_WINAPI(ret_type, func, ...) \
     extern "C" __declspec(dllimport) ret_type __stdcall func(__VA_ARGS__);
-#elif defined(__FreeBSD__)
+#elif SANITIZER_FREEBSD || SANITIZER_NETBSD
 # define WRAP(x) __interceptor_ ## x
 # define WRAPPER_NAME(x) "__interceptor_" #x
 # define INTERCEPTOR_ATTRIBUTE __attribute__((visibility("default")))
@@ -139,7 +140,7 @@ const interpose_substitution substitution_##func_name[] \
 # define DECLARE_WRAPPER(ret_type, func, ...) \
      extern "C" ret_type func(__VA_ARGS__) \
      __attribute__((alias("__interceptor_" #func), visibility("default")));
-#else
+#elif !SANITIZER_FUCHSIA
 # define WRAP(x) __interceptor_ ## x
 # define WRAPPER_NAME(x) "__interceptor_" #x
 # define INTERCEPTOR_ATTRIBUTE __attribute__((visibility("default")))
@@ -148,7 +149,15 @@ const interpose_substitution substitution_##func_name[] \
     __attribute__((weak, alias("__interceptor_" #func), visibility("default")));
 #endif
 
-#if !defined(__APPLE__)
+#if SANITIZER_FUCHSIA
+// There is no general interception at all on Fuchsia.
+// Sanitizer runtimes just define functions directly to preempt them,
+// and have bespoke ways to access the underlying libc functions.
+# include <zircon/sanitizer.h>
+# define INTERCEPTOR_ATTRIBUTE __attribute__((visibility("default")))
+# define REAL(x) __unsanitized_##x
+# define DECLARE_REAL(ret_type, func, ...)
+#elif !SANITIZER_MAC
 # define PTR_TO_REAL(x) real_##x
 # define REAL(x) __interception::PTR_TO_REAL(x)
 # define FUNC_TYPE(x) x##_f
@@ -159,22 +168,26 @@ const interpose_substitution substitution_##func_name[] \
       extern FUNC_TYPE(func) PTR_TO_REAL(func); \
     }
 # define ASSIGN_REAL(dst, src) REAL(dst) = REAL(src)
-#else  // __APPLE__
+#else  // SANITIZER_MAC
 # define REAL(x) x
 # define DECLARE_REAL(ret_type, func, ...) \
     extern "C" ret_type func(__VA_ARGS__);
 # define ASSIGN_REAL(x, y)
-#endif  // __APPLE__
+#endif  // SANITIZER_MAC
 
+#if !SANITIZER_FUCHSIA
 #define DECLARE_REAL_AND_INTERCEPTOR(ret_type, func, ...) \
   DECLARE_REAL(ret_type, func, __VA_ARGS__) \
   extern "C" ret_type WRAP(func)(__VA_ARGS__);
+#else
+#define DECLARE_REAL_AND_INTERCEPTOR(ret_type, func, ...)
+#endif
 
 // Generally, you don't need to use DEFINE_REAL by itself, as INTERCEPTOR
 // macros does its job. In exceptional cases you may need to call REAL(foo)
 // without defining INTERCEPTOR(..., foo, ...). For example, if you override
 // foo with an interceptor for other function.
-#if !defined(__APPLE__)
+#if !SANITIZER_MAC && !SANITIZER_FUCHSIA
 # define DEFINE_REAL(ret_type, func, ...) \
     typedef ret_type (*FUNC_TYPE(func))(__VA_ARGS__); \
     namespace __interception { \
@@ -184,7 +197,18 @@ const interpose_substitution substitution_##func_name[] \
 # define DEFINE_REAL(ret_type, func, ...)
 #endif
 
-#if !defined(__APPLE__)
+#if SANITIZER_FUCHSIA
+
+// We need to define the __interceptor_func name just to get
+// sanitizer_common/scripts/gen_dynamic_list.py to export func.
+// But we don't need to export __interceptor_func to get that.
+#define INTERCEPTOR(ret_type, func, ...)                                \
+  extern "C"[[ gnu::alias(#func), gnu::visibility("hidden") ]] ret_type \
+      __interceptor_##func(__VA_ARGS__);                                \
+  extern "C" INTERCEPTOR_ATTRIBUTE ret_type func(__VA_ARGS__)
+
+#elif !SANITIZER_MAC
+
 #define INTERCEPTOR(ret_type, func, ...) \
   DEFINE_REAL(ret_type, func, __VA_ARGS__) \
   DECLARE_WRAPPER(ret_type, func, __VA_ARGS__) \
@@ -196,7 +220,7 @@ const interpose_substitution substitution_##func_name[] \
 #define INTERCEPTOR_WITH_SUFFIX(ret_type, func, ...) \
   INTERCEPTOR(ret_type, func, __VA_ARGS__)
 
-#else  // __APPLE__
+#else  // SANITIZER_MAC
 
 #define INTERCEPTOR_ZZZ(suffix, ret_type, func, ...) \
   extern "C" ret_type func(__VA_ARGS__) suffix; \
@@ -215,7 +239,7 @@ const interpose_substitution substitution_##func_name[] \
   INTERPOSER_2(overridee, WRAP(overrider))
 #endif
 
-#if defined(_WIN32)
+#if SANITIZER_WINDOWS
 # define INTERCEPTOR_WINAPI(ret_type, func, ...) \
     typedef ret_type (__stdcall *FUNC_TYPE(func))(__VA_ARGS__); \
     namespace __interception { \
@@ -241,17 +265,19 @@ typedef unsigned long uptr;  // NOLINT
 
 #define INCLUDED_FROM_INTERCEPTION_LIB
 
-#if defined(__linux__) || defined(__FreeBSD__)
+#if SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
+
 # include "interception_linux.h"
 # define INTERCEPT_FUNCTION(func) INTERCEPT_FUNCTION_LINUX_OR_FREEBSD(func)
 # define INTERCEPT_FUNCTION_VER(func, symver) \
     INTERCEPT_FUNCTION_VER_LINUX_OR_FREEBSD(func, symver)
-#elif defined(__APPLE__)
+#elif SANITIZER_MAC
 # include "interception_mac.h"
 # define INTERCEPT_FUNCTION(func) INTERCEPT_FUNCTION_MAC(func)
 # define INTERCEPT_FUNCTION_VER(func, symver) \
     INTERCEPT_FUNCTION_VER_MAC(func, symver)
-#else  // defined(_WIN32)
+#elif SANITIZER_WINDOWS
 # include "interception_win.h"
 # define INTERCEPT_FUNCTION(func) INTERCEPT_FUNCTION_WIN(func)
 # define INTERCEPT_FUNCTION_VER(func, symver) \
diff --git a/contrib/compiler-rt/lib/interception/interception_linux.cc b/contrib/compiler-rt/lib/interception/interception_linux.cc
index 6e908ac017b2..c991550a4f72 100644
--- a/contrib/compiler-rt/lib/interception/interception_linux.cc
+++ b/contrib/compiler-rt/lib/interception/interception_linux.cc
@@ -12,25 +12,44 @@
 // Linux-specific interception methods.
 //===----------------------------------------------------------------------===//
 
-#if defined(__linux__) || defined(__FreeBSD__)
 #include "interception.h"
 
+#if SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
+
 #include <dlfcn.h>   // for dlsym() and dlvsym()
 
+#if SANITIZER_NETBSD
+#include "sanitizer_common/sanitizer_libc.h"
+#endif
+
 namespace __interception {
 bool GetRealFunctionAddress(const char *func_name, uptr *func_addr,
     uptr real, uptr wrapper) {
+#if SANITIZER_NETBSD
+  // XXX: Find a better way to handle renames
+  if (internal_strcmp(func_name, "sigaction") == 0) func_name = "__sigaction14";
+#endif
   *func_addr = (uptr)dlsym(RTLD_NEXT, func_name);
+  if (!*func_addr) {
+    // If the lookup using RTLD_NEXT failed, the sanitizer runtime library is
+    // later in the library search order than the DSO that we are trying to
+    // intercept, which means that we cannot intercept this function. We still
+    // want the address of the real definition, though, so look it up using
+    // RTLD_DEFAULT.
+    *func_addr = (uptr)dlsym(RTLD_DEFAULT, func_name);
+  }
   return real == wrapper;
 }
 
-#if !defined(__ANDROID__)  // android does not have dlvsym
+// Android and Solaris do not have dlvsym
+#if !SANITIZER_ANDROID && !SANITIZER_SOLARIS
 void *GetFuncAddrVer(const char *func_name, const char *ver) {
   return dlvsym(RTLD_NEXT, func_name, ver);
 }
-#endif  // !defined(__ANDROID__)
+#endif  // !SANITIZER_ANDROID
 
 }  // namespace __interception
 
-
-#endif  // __linux__ || __FreeBSD__
+#endif  // SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD ||
+        // SANITIZER_SOLARIS
diff --git a/contrib/compiler-rt/lib/interception/interception_linux.h b/contrib/compiler-rt/lib/interception/interception_linux.h
index 27a66c882041..98fe51b858fc 100644
--- a/contrib/compiler-rt/lib/interception/interception_linux.h
+++ b/contrib/compiler-rt/lib/interception/interception_linux.h
@@ -12,7 +12,8 @@
 // Linux-specific interception methods.
 //===----------------------------------------------------------------------===//
 
-#if defined(__linux__) || defined(__FreeBSD__)
+#if SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
 
 #if !defined(INCLUDED_FROM_INTERCEPTION_LIB)
 # error "interception_linux.h should be included from interception library only"
@@ -34,14 +35,16 @@ void *GetFuncAddrVer(const char *func_name, const char *ver);
       (::__interception::uptr) & (func),                                   \
       (::__interception::uptr) & WRAP(func))
 
-#if !defined(__ANDROID__)  // android does not have dlvsym
+// Android and Solaris do not have dlvsym
+#if !SANITIZER_ANDROID && !SANITIZER_SOLARIS
 #define INTERCEPT_FUNCTION_VER_LINUX_OR_FREEBSD(func, symver) \
   (::__interception::real_##func = (func##_f)(                \
        unsigned long)::__interception::GetFuncAddrVer(#func, symver))
 #else
 #define INTERCEPT_FUNCTION_VER_LINUX_OR_FREEBSD(func, symver) \
   INTERCEPT_FUNCTION_LINUX_OR_FREEBSD(func)
-#endif  // !defined(__ANDROID__)
+#endif  // !SANITIZER_ANDROID && !SANITIZER_SOLARIS
 
 #endif  // INTERCEPTION_LINUX_H
-#endif  // __linux__ || __FreeBSD__
+#endif  // SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD ||
+        // SANITIZER_SOLARIS
diff --git a/contrib/compiler-rt/lib/interception/interception_mac.cc b/contrib/compiler-rt/lib/interception/interception_mac.cc
index b035cf998140..ea8072f8dd73 100644
--- a/contrib/compiler-rt/lib/interception/interception_mac.cc
+++ b/contrib/compiler-rt/lib/interception/interception_mac.cc
@@ -12,9 +12,8 @@
 // Mac-specific interception methods.
 //===----------------------------------------------------------------------===//
 
-#ifdef __APPLE__
-
 #include "interception.h"
 
+#if SANITIZER_MAC
 
-#endif  // __APPLE__
+#endif  // SANITIZER_MAC
diff --git a/contrib/compiler-rt/lib/interception/interception_mac.h b/contrib/compiler-rt/lib/interception/interception_mac.h
index e5a35c6971cd..6f31fed47b84 100644
--- a/contrib/compiler-rt/lib/interception/interception_mac.h
+++ b/contrib/compiler-rt/lib/interception/interception_mac.h
@@ -12,7 +12,7 @@
 // Mac-specific interception methods.
 //===----------------------------------------------------------------------===//
 
-#ifdef __APPLE__
+#if SANITIZER_MAC
 
 #if !defined(INCLUDED_FROM_INTERCEPTION_LIB)
 # error "interception_mac.h should be included from interception.h only"
@@ -25,4 +25,4 @@
 #define INTERCEPT_FUNCTION_VER_MAC(func, symver)
 
 #endif  // INTERCEPTION_MAC_H
-#endif  // __APPLE__
+#endif  // SANITIZER_MAC
diff --git a/contrib/compiler-rt/lib/interception/interception_type_test.cc b/contrib/compiler-rt/lib/interception/interception_type_test.cc
index 98ce2e6afc1b..2b3a6d509bfb 100644
--- a/contrib/compiler-rt/lib/interception/interception_type_test.cc
+++ b/contrib/compiler-rt/lib/interception/interception_type_test.cc
@@ -12,9 +12,10 @@
 // Compile-time tests of the internal type definitions.
 //===----------------------------------------------------------------------===//
 
-#if defined(__linux__) || defined(__APPLE__)
-
 #include "interception.h"
+
+#if SANITIZER_LINUX || SANITIZER_MAC
+
 #include <sys/types.h>
 #include <stddef.h>
 #include <stdint.h>
@@ -24,14 +25,14 @@ COMPILER_CHECK(sizeof(::SSIZE_T) == sizeof(ssize_t));
 COMPILER_CHECK(sizeof(::PTRDIFF_T) == sizeof(ptrdiff_t));
 COMPILER_CHECK(sizeof(::INTMAX_T) == sizeof(intmax_t));
 
-#ifndef __APPLE__
+#if !SANITIZER_MAC
 COMPILER_CHECK(sizeof(::OFF64_T) == sizeof(off64_t));
 #endif
 
 // The following are the cases when pread (and friends) is used instead of
 // pread64. In those cases we need OFF_T to match off_t. We don't care about the
 // rest (they depend on _FILE_OFFSET_BITS setting when building an application).
-# if defined(__ANDROID__) || !defined _FILE_OFFSET_BITS || \
+# if SANITIZER_ANDROID || !defined _FILE_OFFSET_BITS || \
   _FILE_OFFSET_BITS != 64
 COMPILER_CHECK(sizeof(::OFF_T) == sizeof(off_t));
 # endif
diff --git a/contrib/compiler-rt/lib/interception/interception_win.cc b/contrib/compiler-rt/lib/interception/interception_win.cc
index b2902d57f542..dc3fe35242ef 100644
--- a/contrib/compiler-rt/lib/interception/interception_win.cc
+++ b/contrib/compiler-rt/lib/interception/interception_win.cc
@@ -125,9 +125,9 @@
 //                                      addr2:  .bytes <body>
 //===----------------------------------------------------------------------===//
 
-#ifdef _WIN32
-
 #include "interception.h"
+
+#if SANITIZER_WINDOWS
 #include "sanitizer_common/sanitizer_platform.h"
 #define WIN32_LEAN_AND_MEAN
 #include <windows.h>
@@ -223,9 +223,8 @@ static bool IsMemoryPadding(uptr address, uptr size) {
   return true;
 }
 
-static const u8 kHintNop10Bytes[] = {
-  0x66, 0x66, 0x0F, 0x1F, 0x84,
-  0x00, 0x00, 0x00, 0x00, 0x00
+static const u8 kHintNop9Bytes[] = {
+  0x66, 0x0F, 0x1F, 0x84, 0x00, 0x00, 0x00, 0x00, 0x00
 };
 
 template<class T>
@@ -240,8 +239,8 @@ static bool FunctionHasPrefix(uptr address, const T &pattern) {
 static bool FunctionHasPadding(uptr address, uptr size) {
   if (IsMemoryPadding(address - size, size))
     return true;
-  if (size <= sizeof(kHintNop10Bytes) &&
-      FunctionHasPrefix(address, kHintNop10Bytes))
+  if (size <= sizeof(kHintNop9Bytes) &&
+      FunctionHasPrefix(address, kHintNop9Bytes))
     return true;
   return false;
 }
@@ -554,7 +553,10 @@ static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
     case 0x246c8948:  // 48 89 6C 24 XX : mov QWORD ptr [rsp + XX], rbp
     case 0x245c8948:  // 48 89 5c 24 XX : mov QWORD PTR [rsp + XX], rbx
     case 0x24748948:  // 48 89 74 24 XX : mov QWORD PTR [rsp + XX], rsi
+    case 0x244C8948:  // 48 89 4C 24 XX : mov QWORD PTR [rsp + XX], rcx
       return 5;
+    case 0x24648348:  // 48 83 64 24 XX : and QWORD PTR [rsp + XX], YY
+      return 6;
   }
 
 #else
@@ -833,6 +835,7 @@ bool OverrideFunction(
 static void **InterestingDLLsAvailable() {
   static const char *InterestingDLLs[] = {
       "kernel32.dll",
+      "msvcr100.dll",      // VS2010
       "msvcr110.dll",      // VS2012
       "msvcr120.dll",      // VS2013
       "vcruntime140.dll",  // VS2015
@@ -1010,4 +1013,4 @@ bool OverrideImportedFunction(const char *module_to_patch,
 
 }  // namespace __interception
 
-#endif  // _WIN32
+#endif  // SANITIZER_MAC
diff --git a/contrib/compiler-rt/lib/interception/interception_win.h b/contrib/compiler-rt/lib/interception/interception_win.h
index 9061f9ed4c21..71a44428f5c7 100644
--- a/contrib/compiler-rt/lib/interception/interception_win.h
+++ b/contrib/compiler-rt/lib/interception/interception_win.h
@@ -12,7 +12,7 @@
 // Windows-specific interception methods.
 //===----------------------------------------------------------------------===//
 
-#ifdef _WIN32
+#if SANITIZER_WINDOWS
 
 #if !defined(INCLUDED_FROM_INTERCEPTION_LIB)
 # error "interception_win.h should be included from interception library only"
@@ -81,4 +81,4 @@ void TestOnlyReleaseTrampolineRegions();
       (::__interception::uptr *)&REAL(func))
 
 #endif  // INTERCEPTION_WIN_H
-#endif  // _WIN32
+#endif  // SANITIZER_WINDOWS
diff --git a/contrib/compiler-rt/lib/lsan/lsan.cc b/contrib/compiler-rt/lib/lsan/lsan.cc
index 6c4767d61252..a9f7e399e14c 100644
--- a/contrib/compiler-rt/lib/lsan/lsan.cc
+++ b/contrib/compiler-rt/lib/lsan/lsan.cc
@@ -56,6 +56,9 @@ static void InitializeFlags() {
   RegisterLsanFlags(&parser, f);
   RegisterCommonFlags(&parser);
 
+  // Override from user-specified string.
+  const char *lsan_default_options = MaybeCallLsanDefaultOptions();
+  parser.ParseString(lsan_default_options);
   parser.ParseString(GetEnv("LSAN_OPTIONS"));
 
   SetVerbosity(common_flags()->verbosity);
@@ -65,6 +68,17 @@ static void InitializeFlags() {
   if (common_flags()->help) parser.PrintFlagDescriptions();
 }
 
+static void OnStackUnwind(const SignalContext &sig, const void *,
+                          BufferedStackTrace *stack) {
+  GetStackTrace(stack, kStackTraceMax, sig.pc, sig.bp, sig.context,
+                common_flags()->fast_unwind_on_fatal);
+}
+
+static void LsanOnDeadlySignal(int signo, void *siginfo, void *context) {
+  HandleDeadlySignal(siginfo, context, GetCurrentThread(), &OnStackUnwind,
+                     nullptr);
+}
+
 extern "C" void __lsan_init() {
   CHECK(!lsan_init_is_running);
   if (lsan_inited)
@@ -80,6 +94,7 @@ extern "C" void __lsan_init() {
   InitTlsSize();
   InitializeInterceptors();
   InitializeThreadRegistry();
+  InstallDeadlySignalHandlers(LsanOnDeadlySignal);
   u32 tid = ThreadCreate(0, 0, true);
   CHECK_EQ(tid, 0);
   ThreadStart(tid, GetTid());
diff --git a/contrib/compiler-rt/lib/lsan/lsan.h b/contrib/compiler-rt/lib/lsan/lsan.h
index 7446e9507e0d..6baee81d00cd 100644
--- a/contrib/compiler-rt/lib/lsan/lsan.h
+++ b/contrib/compiler-rt/lib/lsan/lsan.h
@@ -12,24 +12,14 @@
 //
 //===----------------------------------------------------------------------===//
 
+#include "lsan_thread.h"
 #include "sanitizer_common/sanitizer_flags.h"
 #include "sanitizer_common/sanitizer_stacktrace.h"
 
-#define GET_STACK_TRACE(max_size, fast)                                        \
-  BufferedStackTrace stack;                                                    \
-  {                                                                            \
-    uptr stack_top = 0, stack_bottom = 0;                                      \
-    ThreadContext *t;                                                          \
-    if (fast && (t = CurrentThreadContext())) {                                \
-      stack_top = t->stack_end();                                              \
-      stack_bottom = t->stack_begin();                                         \
-    }                                                                          \
-    if (!SANITIZER_MIPS ||                                                     \
-        IsValidFrame(GET_CURRENT_FRAME(), stack_top, stack_bottom)) {          \
-      stack.Unwind(max_size, StackTrace::GetCurrentPc(), GET_CURRENT_FRAME(),  \
-                   /* context */ 0, stack_top, stack_bottom, fast);            \
-    }                                                                          \
-  }
+#define GET_STACK_TRACE(max_size, fast)                       \
+  __sanitizer::BufferedStackTrace stack;                      \
+  GetStackTrace(&stack, max_size, StackTrace::GetCurrentPc(), \
+                GET_CURRENT_FRAME(), nullptr, fast);
 
 #define GET_STACK_TRACE_FATAL \
   GET_STACK_TRACE(kStackTraceMax, common_flags()->fast_unwind_on_fatal)
@@ -51,9 +41,27 @@ void ReplaceSystemMalloc();
     __lsan_init();                \
 } while (0)
 
+// Get the stack trace with the given pc and bp.
+// The pc will be in the position 0 of the resulting stack trace.
+// The bp may refer to the current frame or to the caller's frame.
+ALWAYS_INLINE
+void GetStackTrace(__sanitizer::BufferedStackTrace *stack,
+                   __sanitizer::uptr max_depth, __sanitizer::uptr pc,
+                   __sanitizer::uptr bp, void *context, bool fast) {
+  uptr stack_top = 0, stack_bottom = 0;
+  ThreadContext *t;
+  if (fast && (t = CurrentThreadContext())) {
+    stack_top = t->stack_end();
+    stack_bottom = t->stack_begin();
+  }
+  if (!SANITIZER_MIPS || IsValidFrame(bp, stack_top, stack_bottom)) {
+    stack->Unwind(max_depth, pc, bp, context, stack_top, stack_bottom, fast);
+  }
+}
+
 }  // namespace __lsan
 
 extern bool lsan_inited;
 extern bool lsan_init_is_running;
 
-extern "C" void __lsan_init();
+extern "C" SANITIZER_INTERFACE_ATTRIBUTE void __lsan_init();
diff --git a/contrib/compiler-rt/lib/lsan/lsan_allocator.h b/contrib/compiler-rt/lib/lsan/lsan_allocator.h
index 5a0d94c71415..4006f7929269 100644
--- a/contrib/compiler-rt/lib/lsan/lsan_allocator.h
+++ b/contrib/compiler-rt/lib/lsan/lsan_allocator.h
@@ -68,9 +68,16 @@ struct AP32 {
 };
 typedef SizeClassAllocator32<AP32> PrimaryAllocator;
 #elif defined(__x86_64__) || defined(__powerpc64__)
+# if defined(__powerpc64__)
+const uptr kAllocatorSpace = 0xa0000000000ULL;
+const uptr kAllocatorSize  = 0x20000000000ULL;  // 2T.
+# else
+const uptr kAllocatorSpace = 0x600000000000ULL;
+const uptr kAllocatorSize  = 0x40000000000ULL;  // 4T.
+# endif
 struct AP64 {  // Allocator64 parameters. Deliberately using a short name.
-  static const uptr kSpaceBeg = 0x600000000000ULL;
-  static const uptr kSpaceSize =  0x40000000000ULL; // 4T.
+  static const uptr kSpaceBeg = kAllocatorSpace;
+  static const uptr kSpaceSize = kAllocatorSize;
   static const uptr kMetadataSize = sizeof(ChunkMetadata);
   typedef DefaultSizeClassMap SizeClassMap;
   typedef NoOpMapUnmapCallback MapUnmapCallback;
diff --git a/contrib/compiler-rt/lib/lsan/lsan_common.cc b/contrib/compiler-rt/lib/lsan/lsan_common.cc
index c121e6a8fb24..69ffda539a26 100644
--- a/contrib/compiler-rt/lib/lsan/lsan_common.cc
+++ b/contrib/compiler-rt/lib/lsan/lsan_common.cc
@@ -57,12 +57,12 @@ void RegisterLsanFlags(FlagParser *parser, Flags *f) {
 #define LOG_POINTERS(...)                           \
   do {                                              \
     if (flags()->log_pointers) Report(__VA_ARGS__); \
-  } while (0);
+  } while (0)
 
 #define LOG_THREADS(...)                           \
   do {                                             \
     if (flags()->log_threads) Report(__VA_ARGS__); \
-  } while (0);
+  } while (0)
 
 ALIGNED(64) static char suppression_placeholder[sizeof(SuppressionContext)];
 static SuppressionContext *suppression_ctx = nullptr;
@@ -107,6 +107,10 @@ void InitializeRootRegions() {
   root_regions = new(placeholder) InternalMmapVector<RootRegion>(1);
 }
 
+const char *MaybeCallLsanDefaultOptions() {
+  return (&__lsan_default_options) ? __lsan_default_options() : "";
+}
+
 void InitCommonLsan() {
   InitializeRootRegions();
   if (common_flags()->detect_leaks) {
@@ -122,7 +126,6 @@ class Decorator: public __sanitizer::SanitizerCommonDecorator {
   Decorator() : SanitizerCommonDecorator() { }
   const char *Error() { return Red(); }
   const char *Leak() { return Blue(); }
-  const char *End() { return Default(); }
 };
 
 static inline bool CanBeAHeapPointer(uptr p) {
@@ -408,8 +411,9 @@ static void MarkInvalidPCCb(uptr chunk, void *arg) {
   }
 }
 
-// On Linux, handles dynamically allocated TLS blocks by treating all chunks
-// allocated from ld-linux.so as reachable.
+// On Linux, treats all chunks allocated from ld-linux.so as reachable, which
+// covers dynamically allocated TLS blocks, internal dynamic loader's loaded
+// modules accounting etc.
 // Dynamic TLS blocks contain the TLS variables of dynamically loaded modules.
 // They are allocated with a __libc_memalign() call in allocate_and_init()
 // (elf/dl-tls.c). Glibc won't tell us the address ranges occupied by those
@@ -564,7 +568,7 @@ static bool CheckForLeaks() {
            "\n");
     Printf("%s", d.Error());
     Report("ERROR: LeakSanitizer: detected memory leaks\n");
-    Printf("%s", d.End());
+    Printf("%s", d.Default());
     param.leak_report.ReportTopLeaks(flags()->max_leaks);
   }
   if (common_flags()->print_suppressions)
@@ -594,6 +598,8 @@ static int DoRecoverableLeakCheck() {
   return have_leaks ? 1 : 0;
 }
 
+void DoRecoverableLeakCheckVoid() { DoRecoverableLeakCheck(); }
+
 static Suppression *GetSuppressionForAddr(uptr addr) {
   Suppression *s = nullptr;
 
@@ -698,7 +704,7 @@ void LeakReport::PrintReportForLeak(uptr index) {
   Printf("%s leak of %zu byte(s) in %zu object(s) allocated from:\n",
          leaks_[index].is_directly_leaked ? "Direct" : "Indirect",
          leaks_[index].total_size, leaks_[index].hit_count);
-  Printf("%s", d.End());
+  Printf("%s", d.Default());
 
   PrintStackTraceById(leaks_[index].stack_trace_id);
 
@@ -756,6 +762,7 @@ uptr LeakReport::UnsuppressedLeakCount() {
 namespace __lsan {
 void InitCommonLsan() { }
 void DoLeakCheck() { }
+void DoRecoverableLeakCheckVoid() { }
 void DisableInThisThread() { }
 void EnableInThisThread() { }
 }
@@ -854,6 +861,11 @@ int __lsan_do_recoverable_leak_check() {
 
 #if !SANITIZER_SUPPORTS_WEAK_HOOKS
 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
+const char * __lsan_default_options() {
+  return "";
+}
+
+SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 int __lsan_is_turned_off() {
   return 0;
 }
diff --git a/contrib/compiler-rt/lib/lsan/lsan_common.h b/contrib/compiler-rt/lib/lsan/lsan_common.h
index 31bf3eb1df42..f3863309d893 100644
--- a/contrib/compiler-rt/lib/lsan/lsan_common.h
+++ b/contrib/compiler-rt/lib/lsan/lsan_common.h
@@ -27,9 +27,9 @@
 // because of "small" (4 bytes) pointer size that leads to high false negative
 // ratio on large leaks. But we still want to have it for some 32 bit arches
 // (e.g. x86), see https://github.com/google/sanitizers/issues/403.
-// To enable LeakSanitizer on new architecture, one need to implement
-// internal_clone function as well as (probably) adjust TLS machinery for
-// new architecture inside sanitizer library.
+// To enable LeakSanitizer on a new architecture, one needs to implement the
+// internal_clone function as well as (probably) adjust the TLS machinery for
+// the new architecture inside the sanitizer library.
 #if (SANITIZER_LINUX && !SANITIZER_ANDROID || SANITIZER_MAC) && \
     (SANITIZER_WORDSIZE == 64) &&                               \
     (defined(__x86_64__) || defined(__mips64) || defined(__aarch64__) || \
@@ -143,8 +143,10 @@ enum IgnoreObjectResult {
 };
 
 // Functions called from the parent tool.
+const char *MaybeCallLsanDefaultOptions();
 void InitCommonLsan();
 void DoLeakCheck();
+void DoRecoverableLeakCheckVoid();
 void DisableCounterUnderflow();
 bool DisabledInThisThread();
 
@@ -250,6 +252,9 @@ class LsanMetadata {
 
 extern "C" {
 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
+const char *__lsan_default_options();
+
+SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 int __lsan_is_turned_off();
 
 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
diff --git a/contrib/compiler-rt/lib/lsan/lsan_common_linux.cc b/contrib/compiler-rt/lib/lsan/lsan_common_linux.cc
index 5042c7b3ada5..cdd7f032a192 100644
--- a/contrib/compiler-rt/lib/lsan/lsan_common_linux.cc
+++ b/contrib/compiler-rt/lib/lsan/lsan_common_linux.cc
@@ -20,6 +20,7 @@
 
 #include "sanitizer_common/sanitizer_common.h"
 #include "sanitizer_common/sanitizer_flags.h"
+#include "sanitizer_common/sanitizer_getauxval.h"
 #include "sanitizer_common/sanitizer_linux.h"
 #include "sanitizer_common/sanitizer_stackdepot.h"
 
@@ -30,8 +31,12 @@ static const char kLinkerName[] = "ld";
 static char linker_placeholder[sizeof(LoadedModule)] ALIGNED(64);
 static LoadedModule *linker = nullptr;
 
-static bool IsLinker(const char* full_name) {
-  return LibraryNameIs(full_name, kLinkerName);
+static bool IsLinker(const LoadedModule& module) {
+#if SANITIZER_USE_GETAUXVAL
+  return module.base_address() == getauxval(AT_BASE);
+#else
+  return LibraryNameIs(module.full_name(), kLinkerName);
+#endif  // SANITIZER_USE_GETAUXVAL
 }
 
 __attribute__((tls_model("initial-exec")))
@@ -49,22 +54,25 @@ void InitializePlatformSpecificModules() {
   ListOfModules modules;
   modules.init();
   for (LoadedModule &module : modules) {
-    if (!IsLinker(module.full_name())) continue;
+    if (!IsLinker(module))
+      continue;
     if (linker == nullptr) {
       linker = reinterpret_cast<LoadedModule *>(linker_placeholder);
       *linker = module;
       module = LoadedModule();
     } else {
       VReport(1, "LeakSanitizer: Multiple modules match \"%s\". "
-              "TLS will not be handled correctly.\n", kLinkerName);
+              "TLS and other allocations originating from linker might be "
+              "falsely reported as leaks.\n", kLinkerName);
       linker->clear();
       linker = nullptr;
       return;
     }
   }
   if (linker == nullptr) {
-    VReport(1, "LeakSanitizer: Dynamic linker not found. "
-               "TLS will not be handled correctly.\n");
+    VReport(1, "LeakSanitizer: Dynamic linker not found. TLS and other "
+               "allocations originating from linker might be falsely reported "
+                "as leaks.\n");
   }
 }
 
diff --git a/contrib/compiler-rt/lib/lsan/lsan_common_mac.cc b/contrib/compiler-rt/lib/lsan/lsan_common_mac.cc
index ade94340ae81..ac27c7af6e35 100644
--- a/contrib/compiler-rt/lib/lsan/lsan_common_mac.cc
+++ b/contrib/compiler-rt/lib/lsan/lsan_common_mac.cc
@@ -92,8 +92,25 @@ LoadedModule *GetLinker() { return nullptr; }
 // required on Darwin.
 void InitializePlatformSpecificModules() {}
 
+// Sections which can't contain contain global pointers. This list errs on the
+// side of caution to avoid false positives, at the expense of performance.
+//
+// Other potentially safe sections include:
+// __all_image_info, __crash_info, __const, __got, __interpose, __objc_msg_break
+//
+// Sections which definitely cannot be included here are:
+// __objc_data, __objc_const, __data, __bss, __common, __thread_data,
+// __thread_bss, __thread_vars, __objc_opt_rw, __objc_opt_ptrs
+static const char *kSkippedSecNames[] = {
+    "__cfstring",       "__la_symbol_ptr",  "__mod_init_func",
+    "__mod_term_func",  "__nl_symbol_ptr",  "__objc_classlist",
+    "__objc_classrefs", "__objc_imageinfo", "__objc_nlclslist",
+    "__objc_protolist", "__objc_selrefs",   "__objc_superrefs"};
+
 // Scans global variables for heap pointers.
 void ProcessGlobalRegions(Frontier *frontier) {
+  for (auto name : kSkippedSecNames) CHECK(ARRAY_SIZE(name) < kMaxSegName);
+
   MemoryMappingLayout memory_mapping(false);
   InternalMmapVector<LoadedModule> modules(/*initial_capacity*/ 128);
   memory_mapping.DumpListOfModules(&modules);
@@ -107,6 +124,10 @@ void ProcessGlobalRegions(Frontier *frontier) {
       // Sections storing global variables are writable and non-executable
       if (range.executable || !range.writable) continue;
 
+      for (auto name : kSkippedSecNames) {
+        if (!internal_strcmp(range.name, name)) continue;
+      }
+
       ScanGlobalRange(range.beg, range.end, frontier);
     }
   }
diff --git a/contrib/compiler-rt/lib/lsan/lsan_interceptors.cc b/contrib/compiler-rt/lib/lsan/lsan_interceptors.cc
index 168868b012bc..b3e73e3896d5 100644
--- a/contrib/compiler-rt/lib/lsan/lsan_interceptors.cc
+++ b/contrib/compiler-rt/lib/lsan/lsan_interceptors.cc
@@ -20,6 +20,7 @@
 #include "sanitizer_common/sanitizer_internal_defs.h"
 #include "sanitizer_common/sanitizer_linux.h"
 #include "sanitizer_common/sanitizer_platform_interceptors.h"
+#include "sanitizer_common/sanitizer_platform_limits_netbsd.h"
 #include "sanitizer_common/sanitizer_platform_limits_posix.h"
 #include "sanitizer_common/sanitizer_posix.h"
 #include "sanitizer_common/sanitizer_tls_get_addr.h"
@@ -44,6 +45,7 @@ int pthread_setspecific(unsigned key, const void *v);
 
 namespace std {
   struct nothrow_t;
+  enum class align_val_t: size_t;
 }
 
 #if !SANITIZER_MAC
@@ -203,13 +205,19 @@ INTERCEPTOR(int, mprobe, void *ptr) {
 #define OPERATOR_NEW_BODY(nothrow)                         \
   ENSURE_LSAN_INITED;                                      \
   GET_STACK_TRACE_MALLOC;                                  \
-  void *res = Allocate(stack, size, 1, kAlwaysClearMemory);\
-  if (!nothrow && UNLIKELY(!res)) DieOnFailure::OnOOM();\
+  void *res = lsan_malloc(size, stack);                    \
+  if (!nothrow && UNLIKELY(!res)) DieOnFailure::OnOOM();   \
+  return res;
+#define OPERATOR_NEW_BODY_ALIGN(nothrow)                   \
+  ENSURE_LSAN_INITED;                                      \
+  GET_STACK_TRACE_MALLOC;                                  \
+  void *res = lsan_memalign((uptr)align, size, stack);     \
+  if (!nothrow && UNLIKELY(!res)) DieOnFailure::OnOOM();   \
   return res;
 
 #define OPERATOR_DELETE_BODY \
   ENSURE_LSAN_INITED;        \
-  Deallocate(ptr);
+  lsan_free(ptr);
 
 // On OS X it's not enough to just provide our own 'operator new' and
 // 'operator delete' implementations, because they're going to be in the runtime
@@ -229,6 +237,18 @@ void *operator new(size_t size, std::nothrow_t const&)
 INTERCEPTOR_ATTRIBUTE
 void *operator new[](size_t size, std::nothrow_t const&)
 { OPERATOR_NEW_BODY(true /*nothrow*/); }
+INTERCEPTOR_ATTRIBUTE
+void *operator new(size_t size, std::align_val_t align)
+{ OPERATOR_NEW_BODY_ALIGN(false /*nothrow*/); }
+INTERCEPTOR_ATTRIBUTE
+void *operator new[](size_t size, std::align_val_t align)
+{ OPERATOR_NEW_BODY_ALIGN(false /*nothrow*/); }
+INTERCEPTOR_ATTRIBUTE
+void *operator new(size_t size, std::align_val_t align, std::nothrow_t const&)
+{ OPERATOR_NEW_BODY_ALIGN(true /*nothrow*/); }
+INTERCEPTOR_ATTRIBUTE
+void *operator new[](size_t size, std::align_val_t align, std::nothrow_t const&)
+{ OPERATOR_NEW_BODY_ALIGN(true /*nothrow*/); }
 
 INTERCEPTOR_ATTRIBUTE
 void operator delete(void *ptr) NOEXCEPT { OPERATOR_DELETE_BODY; }
@@ -239,6 +259,30 @@ void operator delete(void *ptr, std::nothrow_t const&) { OPERATOR_DELETE_BODY; }
 INTERCEPTOR_ATTRIBUTE
 void operator delete[](void *ptr, std::nothrow_t const &)
 { OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete(void *ptr, size_t size) NOEXCEPT
+{ OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete[](void *ptr, size_t size) NOEXCEPT
+{ OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete(void *ptr, std::align_val_t) NOEXCEPT
+{ OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete[](void *ptr, std::align_val_t) NOEXCEPT
+{ OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete(void *ptr, std::align_val_t, std::nothrow_t const&)
+{ OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete[](void *ptr, std::align_val_t, std::nothrow_t const&)
+{ OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete(void *ptr, size_t size, std::align_val_t) NOEXCEPT
+{ OPERATOR_DELETE_BODY; }
+INTERCEPTOR_ATTRIBUTE
+void operator delete[](void *ptr, size_t size, std::align_val_t) NOEXCEPT
+{ OPERATOR_DELETE_BODY; }
 
 #else  // SANITIZER_MAC
 
@@ -265,6 +309,7 @@ INTERCEPTOR(void, _ZdaPvRKSt9nothrow_t, void *ptr, std::nothrow_t const&)
 
 ///// Thread initialization and finalization. /////
 
+#if !SANITIZER_NETBSD
 static unsigned g_thread_finalize_key;
 
 static void thread_finalize(void *v) {
@@ -278,6 +323,18 @@ static void thread_finalize(void *v) {
   }
   ThreadFinish();
 }
+#endif
+
+#if SANITIZER_NETBSD
+INTERCEPTOR(void, _lwp_exit) {
+  ENSURE_LSAN_INITED;
+  ThreadFinish();
+  REAL(_lwp_exit)();
+}
+#define LSAN_MAYBE_INTERCEPT__LWP_EXIT INTERCEPT_FUNCTION(_lwp_exit)
+#else
+#define LSAN_MAYBE_INTERCEPT__LWP_EXIT
+#endif
 
 struct ThreadParam {
   void *(*callback)(void *arg);
@@ -291,11 +348,13 @@ extern "C" void *__lsan_thread_start_func(void *arg) {
   void *param = p->param;
   // Wait until the last iteration to maximize the chance that we are the last
   // destructor to run.
+#if !SANITIZER_NETBSD
   if (pthread_setspecific(g_thread_finalize_key,
                           (void*)GetPthreadDestructorIterations())) {
     Report("LeakSanitizer: failed to set thread key.\n");
     Die();
   }
+#endif
   int tid = 0;
   while ((tid = atomic_load(&p->tid, memory_order_acquire)) == 0)
     internal_sched_yield();
@@ -357,9 +416,14 @@ INTERCEPTOR(void, _exit, int status) {
   REAL(_exit)(status);
 }
 
+#define COMMON_INTERCEPT_FUNCTION(name) INTERCEPT_FUNCTION(name)
+#include "sanitizer_common/sanitizer_signal_interceptors.inc"
+
 namespace __lsan {
 
 void InitializeInterceptors() {
+  InitializeSignalInterceptors();
+
   INTERCEPT_FUNCTION(malloc);
   INTERCEPT_FUNCTION(free);
   LSAN_MAYBE_INTERCEPT_CFREE;
@@ -378,10 +442,14 @@ void InitializeInterceptors() {
   INTERCEPT_FUNCTION(pthread_join);
   INTERCEPT_FUNCTION(_exit);
 
+  LSAN_MAYBE_INTERCEPT__LWP_EXIT;
+
+#if !SANITIZER_NETBSD
   if (pthread_key_create(&g_thread_finalize_key, &thread_finalize)) {
     Report("LeakSanitizer: failed to create thread key.\n");
     Die();
   }
+#endif
 }
 
 } // namespace __lsan
diff --git a/contrib/compiler-rt/lib/lsan/weak_symbols.txt b/contrib/compiler-rt/lib/lsan/weak_symbols.txt
index da4f994da865..692255679bee 100644
--- a/contrib/compiler-rt/lib/lsan/weak_symbols.txt
+++ b/contrib/compiler-rt/lib/lsan/weak_symbols.txt
@@ -1,2 +1,3 @@
+___lsan_default_options
 ___lsan_default_suppressions
 ___lsan_is_turned_off
diff --git a/contrib/compiler-rt/lib/msan/msan.cc b/contrib/compiler-rt/lib/msan/msan.cc
index d2981f0b0edb..e6226ba7670f 100644
--- a/contrib/compiler-rt/lib/msan/msan.cc
+++ b/contrib/compiler-rt/lib/msan/msan.cc
@@ -218,14 +218,14 @@ static void InitializeFlags() {
 }
 
 void GetStackTrace(BufferedStackTrace *stack, uptr max_s, uptr pc, uptr bp,
-                   bool request_fast_unwind) {
+                   void *context, bool request_fast_unwind) {
   MsanThread *t = GetCurrentThread();
   if (!t || !StackTrace::WillUseFastUnwind(request_fast_unwind)) {
     // Block reports from our interceptors during _Unwind_Backtrace.
     SymbolizerScope sym_scope;
-    return stack->Unwind(max_s, pc, bp, nullptr, 0, 0, request_fast_unwind);
+    return stack->Unwind(max_s, pc, bp, context, 0, 0, request_fast_unwind);
   }
-  stack->Unwind(max_s, pc, bp, nullptr, t->stack_top(), t->stack_bottom(),
+  stack->Unwind(max_s, pc, bp, context, t->stack_top(), t->stack_bottom(),
                 request_fast_unwind);
 }
 
@@ -369,6 +369,16 @@ void __msan_warning_noreturn() {
   Die();
 }
 
+static void OnStackUnwind(const SignalContext &sig, const void *,
+                          BufferedStackTrace *stack) {
+  GetStackTrace(stack, kStackTraceMax, sig.pc, sig.bp, sig.context,
+                common_flags()->fast_unwind_on_fatal);
+}
+
+static void MsanOnDeadlySignal(int signo, void *siginfo, void *context) {
+  HandleDeadlySignal(siginfo, context, GetTid(), &OnStackUnwind, nullptr);
+}
+
 void __msan_init() {
   CHECK(!msan_init_is_running);
   if (msan_inited) return;
@@ -384,6 +394,7 @@ void __msan_init() {
   __sanitizer_set_report_path(common_flags()->log_path);
 
   InitializeInterceptors();
+  InstallDeadlySignalHandlers(MsanOnDeadlySignal);
   InstallAtExitHandler(); // Needs __cxa_atexit interceptor.
 
   DisableCoreDumperIfNecessary();
diff --git a/contrib/compiler-rt/lib/msan/msan.h b/contrib/compiler-rt/lib/msan/msan.h
index fa9c15b88bef..cbae444127ee 100644
--- a/contrib/compiler-rt/lib/msan/msan.h
+++ b/contrib/compiler-rt/lib/msan/msan.h
@@ -160,21 +160,25 @@ const MappingDesc kMemoryLayout[] = {
 # define SHADOW_TO_ORIGIN(shadow) (((uptr)(shadow)) + 0x1000000000ULL)
 
 #elif SANITIZER_LINUX && SANITIZER_PPC64
-
 const MappingDesc kMemoryLayout[] = {
-    {0x000000000000ULL, 0x000100000000ULL, MappingDesc::APP, "low memory"},
-    {0x000100000000ULL, 0x080000000000ULL, MappingDesc::INVALID, "invalid"},
-    {0x080000000000ULL, 0x180100000000ULL, MappingDesc::SHADOW, "shadow"},
-    {0x180100000000ULL, 0x1C0000000000ULL, MappingDesc::INVALID, "invalid"},
-    {0x1C0000000000ULL, 0x2C0100000000ULL, MappingDesc::ORIGIN, "origin"},
-    {0x2C0100000000ULL, 0x300000000000ULL, MappingDesc::INVALID, "invalid"},
-    {0x300000000000ULL, 0x400000000000ULL, MappingDesc::APP, "high memory"}};
+    {0x000000000000ULL, 0x000200000000ULL, MappingDesc::APP, "low memory"},
+    {0x000200000000ULL, 0x080000000000ULL, MappingDesc::INVALID, "invalid"},
+    {0x080000000000ULL, 0x180200000000ULL, MappingDesc::SHADOW, "shadow"},
+    {0x180200000000ULL, 0x1C0000000000ULL, MappingDesc::INVALID, "invalid"},
+    {0x1C0000000000ULL, 0x2C0200000000ULL, MappingDesc::ORIGIN, "origin"},
+    {0x2C0200000000ULL, 0x300000000000ULL, MappingDesc::INVALID, "invalid"},
+    {0x300000000000ULL, 0x800000000000ULL, MappingDesc::APP, "high memory"}};
 
+// Various kernels use different low end ranges but we can combine them into one
+// big range. They also use different high end ranges but we can map them all to
+// one range.
 // Maps low and high app ranges to contiguous space with zero base:
-//   Low:  0000 0000 0000 - 0000 ffff ffff  ->  1000 0000 0000 - 1000 ffff ffff
+//   Low:  0000 0000 0000 - 0001 ffff ffff  ->  1000 0000 0000 - 1001 ffff ffff
 //   High: 3000 0000 0000 - 3fff ffff ffff  ->  0000 0000 0000 - 0fff ffff ffff
+//   High: 4000 0000 0000 - 4fff ffff ffff  ->  0000 0000 0000 - 0fff ffff ffff
+//   High: 7000 0000 0000 - 7fff ffff ffff  ->  0000 0000 0000 - 0fff ffff ffff
 #define LINEARIZE_MEM(mem) \
-  (((uptr)(mem) & ~0x200000000000ULL) ^ 0x100000000000ULL)
+  (((uptr)(mem) & ~0xE00000000000ULL) ^ 0x100000000000ULL)
 #define MEM_TO_SHADOW(mem) (LINEARIZE_MEM((mem)) + 0x080000000000ULL)
 #define SHADOW_TO_ORIGIN(shadow) (((uptr)(shadow)) + 0x140000000000ULL)
 
@@ -199,7 +203,7 @@ const MappingDesc kMemoryLayout[] = {
 #define MEM_TO_SHADOW(mem) (LINEARIZE_MEM((mem)) + 0x100000000000ULL)
 #define SHADOW_TO_ORIGIN(shadow) (((uptr)(shadow)) + 0x280000000000)
 
-#elif SANITIZER_LINUX && SANITIZER_WORDSIZE == 64
+#elif SANITIZER_NETBSD || (SANITIZER_LINUX && SANITIZER_WORDSIZE == 64)
 
 #ifdef MSAN_LINUX_X86_64_OLD_MAPPING
 // Requries PIE binary and ASLR enabled.
@@ -310,7 +314,7 @@ void PrintWarning(uptr pc, uptr bp);
 void PrintWarningWithOrigin(uptr pc, uptr bp, u32 origin);
 
 void GetStackTrace(BufferedStackTrace *stack, uptr max_s, uptr pc, uptr bp,
-                   bool request_fast_unwind);
+                   void *context, bool request_fast_unwind);
 
 void ReportUMR(StackTrace *stack, u32 origin);
 void ReportExpectedUMRNotFound(StackTrace *stack);
@@ -330,32 +334,32 @@ u32 ChainOrigin(u32 id, StackTrace *stack);
 
 const int STACK_TRACE_TAG_POISON = StackTrace::TAG_CUSTOM + 1;
 
-#define GET_MALLOC_STACK_TRACE                                                 \
-  BufferedStackTrace stack;                                                    \
-  if (__msan_get_track_origins() && msan_inited)                               \
-  GetStackTrace(&stack, common_flags()->malloc_context_size,                   \
-                StackTrace::GetCurrentPc(), GET_CURRENT_FRAME(),               \
+#define GET_MALLOC_STACK_TRACE                                            \
+  BufferedStackTrace stack;                                               \
+  if (__msan_get_track_origins() && msan_inited)                          \
+  GetStackTrace(&stack, common_flags()->malloc_context_size,              \
+                StackTrace::GetCurrentPc(), GET_CURRENT_FRAME(), nullptr, \
                 common_flags()->fast_unwind_on_malloc)
 
 // For platforms which support slow unwinder only, we restrict the store context
 // size to 1, basically only storing the current pc. We do this because the slow
 // unwinder which is based on libunwind is not async signal safe and causes
 // random freezes in forking applications as well as in signal handlers.
-#define GET_STORE_STACK_TRACE_PC_BP(pc, bp)                                    \
-  BufferedStackTrace stack;                                                    \
-  if (__msan_get_track_origins() > 1 && msan_inited) {                         \
-    if (!SANITIZER_CAN_FAST_UNWIND)                                            \
-      GetStackTrace(&stack, Min(1, flags()->store_context_size), pc, bp,       \
-                    false);                                                    \
-    else                                                                       \
-      GetStackTrace(&stack, flags()->store_context_size, pc, bp,               \
-                    common_flags()->fast_unwind_on_malloc);                    \
+#define GET_STORE_STACK_TRACE_PC_BP(pc, bp)                               \
+  BufferedStackTrace stack;                                               \
+  if (__msan_get_track_origins() > 1 && msan_inited) {                    \
+    if (!SANITIZER_CAN_FAST_UNWIND)                                       \
+      GetStackTrace(&stack, Min(1, flags()->store_context_size), pc, bp,  \
+                    nullptr, false);                                      \
+    else                                                                  \
+      GetStackTrace(&stack, flags()->store_context_size, pc, bp, nullptr, \
+                    common_flags()->fast_unwind_on_malloc);               \
   }
 
-#define GET_FATAL_STACK_TRACE_PC_BP(pc, bp)                                    \
-  BufferedStackTrace stack;                                                    \
-  if (msan_inited)                                                             \
-  GetStackTrace(&stack, kStackTraceMax, pc, bp,                                \
+#define GET_FATAL_STACK_TRACE_PC_BP(pc, bp)              \
+  BufferedStackTrace stack;                              \
+  if (msan_inited)                                       \
+  GetStackTrace(&stack, kStackTraceMax, pc, bp, nullptr, \
                 common_flags()->fast_unwind_on_fatal)
 
 #define GET_STORE_STACK_TRACE \
diff --git a/contrib/compiler-rt/lib/msan/msan_allocator.cc b/contrib/compiler-rt/lib/msan/msan_allocator.cc
index 1034dbdf9b55..0f9942324931 100644
--- a/contrib/compiler-rt/lib/msan/msan_allocator.cc
+++ b/contrib/compiler-rt/lib/msan/msan_allocator.cc
@@ -62,7 +62,8 @@ struct MsanMapUnmapCallback {
   };
   typedef SizeClassAllocator32<AP32> PrimaryAllocator;
 #elif defined(__x86_64__)
-#if SANITIZER_LINUX && !defined(MSAN_LINUX_X86_64_OLD_MAPPING)
+#if SANITIZER_NETBSD || \
+    (SANITIZER_LINUX && !defined(MSAN_LINUX_X86_64_OLD_MAPPING))
   static const uptr kAllocatorSpace = 0x700000000000ULL;
 #else
   static const uptr kAllocatorSpace = 0x600000000000ULL;
@@ -255,8 +256,12 @@ void *msan_valloc(uptr size, StackTrace *stack) {
 
 void *msan_pvalloc(uptr size, StackTrace *stack) {
   uptr PageSize = GetPageSizeCached();
+  if (UNLIKELY(CheckForPvallocOverflow(size, PageSize))) {
+    errno = errno_ENOMEM;
+    return Allocator::FailureHandler::OnBadRequest();
+  }
   // pvalloc(0) should allocate one page.
-  size = size == 0 ? PageSize : RoundUpTo(size, PageSize);
+  size = size ? RoundUpTo(size, PageSize) : PageSize;
   return SetErrnoOnNull(MsanAllocate(stack, size, PageSize, false));
 }
 
diff --git a/contrib/compiler-rt/lib/msan/msan_interceptors.cc b/contrib/compiler-rt/lib/msan/msan_interceptors.cc
index b5d22baca08d..a7fe09b25ffb 100644
--- a/contrib/compiler-rt/lib/msan/msan_interceptors.cc
+++ b/contrib/compiler-rt/lib/msan/msan_interceptors.cc
@@ -22,6 +22,7 @@
 #include "msan_thread.h"
 #include "msan_poisoning.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_interface.h"
 #include "sanitizer_common/sanitizer_allocator_internal.h"
@@ -33,6 +34,11 @@
 #include "sanitizer_common/sanitizer_linux.h"
 #include "sanitizer_common/sanitizer_tls_get_addr.h"
 
+#if SANITIZER_NETBSD
+#define gettimeofday __gettimeofday50
+#define getrusage __getrusage50
+#endif
+
 #include <stdarg.h>
 // ACHTUNG! No other system header includes in this file.
 // Ideally, we should get rid of stdarg.h as well.
@@ -86,22 +92,21 @@ static void *AllocateFromLocalPool(uptr size_in_bytes) {
 } while (0)
 
 // Check that [x, x+n) range is unpoisoned.
-#define CHECK_UNPOISONED_0(x, n)                                               \
-  do {                                                                         \
-    sptr offset = __msan_test_shadow(x, n);                                    \
-    if (__msan::IsInSymbolizer())                                              \
-      break;                                                                   \
-    if (offset >= 0 && __msan::flags()->report_umrs) {                         \
-      GET_CALLER_PC_BP_SP;                                                     \
-      (void) sp;                                                               \
-      ReportUMRInsideAddressRange(__func__, x, n, offset);                     \
-      __msan::PrintWarningWithOrigin(                                          \
-          pc, bp, __msan_get_origin((const char *)x + offset));                \
-      if (__msan::flags()->halt_on_error) {                                    \
-        Printf("Exiting\n");                                                   \
-        Die();                                                                 \
-      }                                                                        \
-    }                                                                          \
+#define CHECK_UNPOISONED_0(x, n)                                  \
+  do {                                                            \
+    sptr __offset = __msan_test_shadow(x, n);                     \
+    if (__msan::IsInSymbolizer()) break;                          \
+    if (__offset >= 0 && __msan::flags()->report_umrs) {          \
+      GET_CALLER_PC_BP_SP;                                        \
+      (void)sp;                                                   \
+      ReportUMRInsideAddressRange(__func__, x, n, __offset);      \
+      __msan::PrintWarningWithOrigin(                             \
+          pc, bp, __msan_get_origin((const char *)x + __offset)); \
+      if (__msan::flags()->halt_on_error) {                       \
+        Printf("Exiting\n");                                      \
+        Die();                                                    \
+      }                                                           \
+    }                                                             \
   } while (0)
 
 // Check that [x, x+n) range is unpoisoned unless we are in a nested
@@ -109,7 +114,7 @@ static void *AllocateFromLocalPool(uptr size_in_bytes) {
 #define CHECK_UNPOISONED(x, n)                             \
   do {                                                     \
     if (!IsInInterceptorScope()) CHECK_UNPOISONED_0(x, n); \
-  } while (0);
+  } while (0)
 
 #define CHECK_UNPOISONED_STRING_OF_LEN(x, len, n)               \
   CHECK_UNPOISONED((x),                                         \
@@ -118,7 +123,7 @@ static void *AllocateFromLocalPool(uptr size_in_bytes) {
 #define CHECK_UNPOISONED_STRING(x, n)                           \
     CHECK_UNPOISONED_STRING_OF_LEN((x), internal_strlen(x), (n))
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(SIZE_T, fread_unlocked, void *ptr, SIZE_T size, SIZE_T nmemb,
             void *file) {
   ENSURE_MSAN_INITED();
@@ -134,16 +139,21 @@ INTERCEPTOR(SIZE_T, fread_unlocked, void *ptr, SIZE_T size, SIZE_T nmemb,
 
 INTERCEPTOR(SSIZE_T, readlink, const char *path, char *buf, SIZE_T bufsiz) {
   ENSURE_MSAN_INITED();
-  CHECK_UNPOISONED_STRING(path, 0)
+  CHECK_UNPOISONED_STRING(path, 0);
   SSIZE_T res = REAL(readlink)(path, buf, bufsiz);
   if (res > 0)
     __msan_unpoison(buf, res);
   return res;
 }
 
+#if !SANITIZER_NETBSD
 INTERCEPTOR(void *, mempcpy, void *dest, const void *src, SIZE_T n) {
   return (char *)__msan_memcpy(dest, src, n) + n;
 }
+#define MSAN_MAYBE_INTERCEPT_MEMPCPY INTERCEPT_FUNCTION(mempcpy)
+#else
+#define MSAN_MAYBE_INTERCEPT_MEMPCPY
+#endif
 
 INTERCEPTOR(void *, memccpy, void *dest, const void *src, int c, SIZE_T n) {
   ENSURE_MSAN_INITED();
@@ -168,7 +178,7 @@ INTERCEPTOR(int, posix_memalign, void **memptr, SIZE_T alignment, SIZE_T size) {
   return res;
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(void *, memalign, SIZE_T alignment, SIZE_T size) {
   GET_MALLOC_STACK_TRACE;
   return msan_memalign(alignment, size, &stack);
@@ -183,6 +193,7 @@ INTERCEPTOR(void *, aligned_alloc, SIZE_T alignment, SIZE_T size) {
   return msan_aligned_alloc(alignment, size, &stack);
 }
 
+#if !SANITIZER_NETBSD
 INTERCEPTOR(void *, __libc_memalign, SIZE_T alignment, SIZE_T size) {
   GET_MALLOC_STACK_TRACE;
   void *ptr = msan_memalign(alignment, size, &stack);
@@ -190,13 +201,17 @@ INTERCEPTOR(void *, __libc_memalign, SIZE_T alignment, SIZE_T size) {
     DTLS_on_libc_memalign(ptr, size);
   return ptr;
 }
+#define MSAN_MAYBE_INTERCEPT___LIBC_MEMALIGN INTERCEPT_FUNCTION(__libc_memalign)
+#else
+#define MSAN_MAYBE_INTERCEPT___LIBC_MEMALIGN
+#endif
 
 INTERCEPTOR(void *, valloc, SIZE_T size) {
   GET_MALLOC_STACK_TRACE;
   return msan_valloc(size, &stack);
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(void *, pvalloc, SIZE_T size) {
   GET_MALLOC_STACK_TRACE;
   return msan_pvalloc(size, &stack);
@@ -212,7 +227,7 @@ INTERCEPTOR(void, free, void *ptr) {
   MsanDeallocate(&stack, ptr);
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(void, cfree, void *ptr) {
   GET_MALLOC_STACK_TRACE;
   if (!ptr || UNLIKELY(IsInDlsymAllocPool(ptr))) return;
@@ -223,11 +238,17 @@ INTERCEPTOR(void, cfree, void *ptr) {
 #define MSAN_MAYBE_INTERCEPT_CFREE
 #endif
 
+#if !SANITIZER_NETBSD
 INTERCEPTOR(uptr, malloc_usable_size, void *ptr) {
   return __sanitizer_get_allocated_size(ptr);
 }
+#define MSAN_MAYBE_INTERCEPT_MALLOC_USABLE_SIZE \
+  INTERCEPT_FUNCTION(malloc_usable_size)
+#else
+#define MSAN_MAYBE_INTERCEPT_MALLOC_USABLE_SIZE
+#endif
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 // This function actually returns a struct by value, but we can't unpoison a
 // temporary! The following is equivalent on all supported platforms but
 // aarch64 (which uses a different register for sret value).  We have a test
@@ -246,7 +267,7 @@ INTERCEPTOR(void, mallinfo, __sanitizer_mallinfo *sret) {
 #define MSAN_MAYBE_INTERCEPT_MALLINFO
 #endif
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(int, mallopt, int cmd, int value) {
   return -1;
 }
@@ -255,7 +276,7 @@ INTERCEPTOR(int, mallopt, int cmd, int value) {
 #define MSAN_MAYBE_INTERCEPT_MALLOPT
 #endif
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(void, malloc_stats, void) {
   // FIXME: implement, but don't call REAL(malloc_stats)!
 }
@@ -286,6 +307,7 @@ INTERCEPTOR(char *, strncpy, char *dest, const char *src, SIZE_T n) {  // NOLINT
   return res;
 }
 
+#if !SANITIZER_NETBSD
 INTERCEPTOR(char *, stpcpy, char *dest, const char *src) {  // NOLINT
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
@@ -295,6 +317,10 @@ INTERCEPTOR(char *, stpcpy, char *dest, const char *src) {  // NOLINT
   CopyShadowAndOrigin(dest, src, n + 1, &stack);
   return res;
 }
+#define MSAN_MAYBE_INTERCEPT_STPCPY INTERCEPT_FUNCTION(stpcpy)
+#else
+#define MSAN_MAYBE_INTERCEPT_STPCPY
+#endif
 
 INTERCEPTOR(char *, strdup, char *src) {
   ENSURE_MSAN_INITED();
@@ -308,7 +334,7 @@ INTERCEPTOR(char *, strdup, char *src) {
   return res;
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(char *, __strdup, char *src) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
@@ -323,6 +349,7 @@ INTERCEPTOR(char *, __strdup, char *src) {
 #define MSAN_MAYBE_INTERCEPT___STRDUP
 #endif
 
+#if !SANITIZER_NETBSD
 INTERCEPTOR(char *, gcvt, double number, SIZE_T ndigit, char *buf) {
   ENSURE_MSAN_INITED();
   char *res = REAL(gcvt)(number, ndigit, buf);
@@ -330,6 +357,10 @@ INTERCEPTOR(char *, gcvt, double number, SIZE_T ndigit, char *buf) {
   __msan_unpoison(buf, n + 1);
   return res;
 }
+#define MSAN_MAYBE_INTERCEPT_GCVT INTERCEPT_FUNCTION(gcvt)
+#else
+#define MSAN_MAYBE_INTERCEPT_GCVT
+#endif
 
 INTERCEPTOR(char *, strcat, char *dest, const char *src) {  // NOLINT
   ENSURE_MSAN_INITED();
@@ -386,6 +417,16 @@ INTERCEPTOR(char *, strncat, char *dest, const char *src, SIZE_T n) {  // NOLINT
     INTERCEPTOR_STRTO_BODY(ret_type, func, nptr, endptr, base, loc);     \
   }
 
+#if SANITIZER_NETBSD
+#define INTERCEPTORS_STRTO(ret_type, func, char_type)      \
+  INTERCEPTOR_STRTO(ret_type, func, char_type)             \
+  INTERCEPTOR_STRTO_LOC(ret_type, func##_l, char_type)
+
+#define INTERCEPTORS_STRTO_BASE(ret_type, func, char_type)      \
+  INTERCEPTOR_STRTO_BASE(ret_type, func, char_type)             \
+  INTERCEPTOR_STRTO_BASE_LOC(ret_type, func##_l, char_type)
+
+#else
 #define INTERCEPTORS_STRTO(ret_type, func, char_type)      \
   INTERCEPTOR_STRTO(ret_type, func, char_type)             \
   INTERCEPTOR_STRTO_LOC(ret_type, func##_l, char_type)     \
@@ -397,6 +438,7 @@ INTERCEPTOR(char *, strncat, char *dest, const char *src, SIZE_T n) {  // NOLINT
   INTERCEPTOR_STRTO_BASE_LOC(ret_type, func##_l, char_type)     \
   INTERCEPTOR_STRTO_BASE_LOC(ret_type, __##func##_l, char_type) \
   INTERCEPTOR_STRTO_BASE_LOC(ret_type, __##func##_internal, char_type)
+#endif
 
 INTERCEPTORS_STRTO(double, strtod, char)                     // NOLINT
 INTERCEPTORS_STRTO(float, strtof, char)                      // NOLINT
@@ -405,6 +447,7 @@ INTERCEPTORS_STRTO_BASE(long, strtol, char)                  // NOLINT
 INTERCEPTORS_STRTO_BASE(long long, strtoll, char)            // NOLINT
 INTERCEPTORS_STRTO_BASE(unsigned long, strtoul, char)        // NOLINT
 INTERCEPTORS_STRTO_BASE(unsigned long long, strtoull, char)  // NOLINT
+INTERCEPTORS_STRTO_BASE(u64, strtouq, char)                  // NOLINT
 
 INTERCEPTORS_STRTO(double, wcstod, wchar_t)                     // NOLINT
 INTERCEPTORS_STRTO(float, wcstof, wchar_t)                      // NOLINT
@@ -414,11 +457,17 @@ INTERCEPTORS_STRTO_BASE(long long, wcstoll, wchar_t)            // NOLINT
 INTERCEPTORS_STRTO_BASE(unsigned long, wcstoul, wchar_t)        // NOLINT
 INTERCEPTORS_STRTO_BASE(unsigned long long, wcstoull, wchar_t)  // NOLINT
 
+#if SANITIZER_NETBSD
+#define INTERCEPT_STRTO(func) \
+  INTERCEPT_FUNCTION(func); \
+  INTERCEPT_FUNCTION(func##_l);
+#else
 #define INTERCEPT_STRTO(func) \
   INTERCEPT_FUNCTION(func); \
   INTERCEPT_FUNCTION(func##_l); \
   INTERCEPT_FUNCTION(__##func##_l); \
   INTERCEPT_FUNCTION(__##func##_internal);
+#endif
 
 
 // FIXME: support *wprintf in common format interceptors.
@@ -457,6 +506,20 @@ INTERCEPTOR(SIZE_T, strxfrm_l, char *dest, const char *src, SIZE_T n,
   return res;
 }
 
+#if SANITIZER_LINUX
+INTERCEPTOR(SIZE_T, __strxfrm_l, char *dest, const char *src, SIZE_T n,
+            void *loc) {
+  ENSURE_MSAN_INITED();
+  CHECK_UNPOISONED(src, REAL(strlen)(src) + 1);
+  SIZE_T res = REAL(__strxfrm_l)(dest, src, n, loc);
+  if (res < n) __msan_unpoison(dest, res + 1);
+  return res;
+}
+#define MSAN_MAYBE_INTERCEPT___STRXFRM_L INTERCEPT_FUNCTION(__strxfrm_l)
+#else
+#define MSAN_MAYBE_INTERCEPT___STRXFRM_L
+#endif
+
 #define INTERCEPTOR_STRFTIME_BODY(char_type, ret_type, func, s, ...) \
   ENSURE_MSAN_INITED();                                              \
   ret_type res = REAL(func)(s, __VA_ARGS__);                         \
@@ -473,7 +536,7 @@ INTERCEPTOR(SIZE_T, strftime_l, char *s, SIZE_T max, const char *format,
   INTERCEPTOR_STRFTIME_BODY(char, SIZE_T, strftime_l, s, max, format, tm, loc);
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(SIZE_T, __strftime_l, char *s, SIZE_T max, const char *format,
             __sanitizer_tm *tm, void *loc) {
   INTERCEPTOR_STRFTIME_BODY(char, SIZE_T, __strftime_l, s, max, format, tm,
@@ -495,7 +558,7 @@ INTERCEPTOR(SIZE_T, wcsftime_l, wchar_t *s, SIZE_T max, const wchar_t *format,
                             loc);
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(SIZE_T, __wcsftime_l, wchar_t *s, SIZE_T max, const wchar_t *format,
             __sanitizer_tm *tm, void *loc) {
   INTERCEPTOR_STRFTIME_BODY(wchar_t, SIZE_T, __wcsftime_l, s, max, format, tm,
@@ -513,7 +576,8 @@ INTERCEPTOR(int, mbtowc, wchar_t *dest, const char *src, SIZE_T n) {
   return res;
 }
 
-INTERCEPTOR(int, mbrtowc, wchar_t *dest, const char *src, SIZE_T n, void *ps) {
+INTERCEPTOR(SIZE_T, mbrtowc, wchar_t *dest, const char *src, SIZE_T n,
+            void *ps) {
   ENSURE_MSAN_INITED();
   SIZE_T res = REAL(mbrtowc)(dest, src, n, ps);
   if (res != (SIZE_T)-1 && dest) __msan_unpoison(dest, sizeof(wchar_t));
@@ -529,6 +593,7 @@ INTERCEPTOR(wchar_t *, wmemcpy, wchar_t *dest, const wchar_t *src, SIZE_T n) {
   return res;
 }
 
+#if !SANITIZER_NETBSD
 INTERCEPTOR(wchar_t *, wmempcpy, wchar_t *dest, const wchar_t *src, SIZE_T n) {
   ENSURE_MSAN_INITED();
   GET_STORE_STACK_TRACE;
@@ -536,6 +601,10 @@ INTERCEPTOR(wchar_t *, wmempcpy, wchar_t *dest, const wchar_t *src, SIZE_T n) {
   CopyShadowAndOrigin(dest, src, n * sizeof(wchar_t), &stack);
   return res;
 }
+#define MSAN_MAYBE_INTERCEPT_WMEMPCPY INTERCEPT_FUNCTION(wmempcpy)
+#else
+#define MSAN_MAYBE_INTERCEPT_WMEMPCPY
+#endif
 
 INTERCEPTOR(wchar_t *, wmemset, wchar_t *s, wchar_t c, SIZE_T n) {
   CHECK(MEM_IS_APP(s));
@@ -569,6 +638,7 @@ INTERCEPTOR(int, gettimeofday, void *tv, void *tz) {
   return res;
 }
 
+#if !SANITIZER_NETBSD
 INTERCEPTOR(char *, fcvt, double x, int a, int *b, int *c) {
   ENSURE_MSAN_INITED();
   char *res = REAL(fcvt)(x, a, b, c);
@@ -577,6 +647,10 @@ INTERCEPTOR(char *, fcvt, double x, int a, int *b, int *c) {
   if (res) __msan_unpoison(res, REAL(strlen)(res) + 1);
   return res;
 }
+#define MSAN_MAYBE_INTERCEPT_FCVT INTERCEPT_FUNCTION(fcvt)
+#else
+#define MSAN_MAYBE_INTERCEPT_FCVT
+#endif
 
 INTERCEPTOR(char *, getenv, char *name) {
   if (msan_init_is_running)
@@ -601,7 +675,7 @@ static void UnpoisonEnviron() {
 
 INTERCEPTOR(int, setenv, const char *name, const char *value, int overwrite) {
   ENSURE_MSAN_INITED();
-  CHECK_UNPOISONED_STRING(name, 0)
+  CHECK_UNPOISONED_STRING(name, 0);
   int res = REAL(setenv)(name, value, overwrite);
   if (!res) UnpoisonEnviron();
   return res;
@@ -614,7 +688,7 @@ INTERCEPTOR(int, putenv, char *string) {
   return res;
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(int, __fxstat, int magic, int fd, void *buf) {
   ENSURE_MSAN_INITED();
   int res = REAL(__fxstat)(magic, fd, buf);
@@ -627,7 +701,7 @@ INTERCEPTOR(int, __fxstat, int magic, int fd, void *buf) {
 #define MSAN_MAYBE_INTERCEPT___FXSTAT
 #endif
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(int, __fxstat64, int magic, int fd, void *buf) {
   ENSURE_MSAN_INITED();
   int res = REAL(__fxstat64)(magic, fd, buf);
@@ -640,7 +714,7 @@ INTERCEPTOR(int, __fxstat64, int magic, int fd, void *buf) {
 #define MSAN_MAYBE_INTERCEPT___FXSTAT64
 #endif
 
-#if SANITIZER_FREEBSD
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
 INTERCEPTOR(int, fstatat, int fd, char *pathname, void *buf, int flags) {
   ENSURE_MSAN_INITED();
   int res = REAL(fstatat)(fd, pathname, buf, flags);
@@ -659,7 +733,7 @@ INTERCEPTOR(int, __fxstatat, int magic, int fd, char *pathname, void *buf,
 # define MSAN_INTERCEPT_FSTATAT INTERCEPT_FUNCTION(__fxstatat)
 #endif
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(int, __fxstatat64, int magic, int fd, char *pathname, void *buf,
             int flags) {
   ENSURE_MSAN_INITED();
@@ -706,7 +780,7 @@ INTERCEPTOR(char *, fgets, char *s, int size, void *stream) {
   return res;
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(char *, fgets_unlocked, char *s, int size, void *stream) {
   ENSURE_MSAN_INITED();
   char *res = REAL(fgets_unlocked)(s, size, stream);
@@ -729,7 +803,7 @@ INTERCEPTOR(int, getrlimit, int resource, void *rlim) {
   return res;
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(int, getrlimit64, int resource, void *rlim) {
   if (msan_init_is_running) return REAL(getrlimit64)(resource, rlim);
   ENSURE_MSAN_INITED();
@@ -805,7 +879,7 @@ INTERCEPTOR(int, gethostname, char *name, SIZE_T len) {
   return res;
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(int, epoll_wait, int epfd, void *events, int maxevents,
     int timeout) {
   ENSURE_MSAN_INITED();
@@ -820,7 +894,7 @@ INTERCEPTOR(int, epoll_wait, int epfd, void *events, int maxevents,
 #define MSAN_MAYBE_INTERCEPT_EPOLL_WAIT
 #endif
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(int, epoll_pwait, int epfd, void *events, int maxevents,
     int timeout, void *sigmask) {
   ENSURE_MSAN_INITED();
@@ -909,7 +983,7 @@ INTERCEPTOR(void *, mmap, void *addr, SIZE_T length, int prot, int flags,
   return res;
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 INTERCEPTOR(void *, mmap64, void *addr, SIZE_T length, int prot, int flags,
             int fd, OFF64_T offset) {
   ENSURE_MSAN_INITED();
@@ -983,59 +1057,13 @@ static void SignalAction(int signo, void *si, void *uc) {
   cb(signo, si, uc);
 }
 
-INTERCEPTOR(int, sigaction, int signo, const __sanitizer_sigaction *act,
-            __sanitizer_sigaction *oldact) {
-  ENSURE_MSAN_INITED();
-  // FIXME: check that *act is unpoisoned.
-  // That requires intercepting all of sigemptyset, sigfillset, etc.
-  int res;
-  if (flags()->wrap_signals) {
-    SpinMutexLock lock(&sigactions_mu);
-    CHECK_LT(signo, kMaxSignals);
-    uptr old_cb = atomic_load(&sigactions[signo], memory_order_relaxed);
-    __sanitizer_sigaction new_act;
-    __sanitizer_sigaction *pnew_act = act ? &new_act : nullptr;
-    if (act) {
-      REAL(memcpy)(pnew_act, act, sizeof(__sanitizer_sigaction));
-      uptr cb = (uptr)pnew_act->sigaction;
-      uptr new_cb = (pnew_act->sa_flags & __sanitizer::sa_siginfo)
-                        ? (uptr)SignalAction
-                        : (uptr)SignalHandler;
-      if (cb != __sanitizer::sig_ign && cb != __sanitizer::sig_dfl) {
-        atomic_store(&sigactions[signo], cb, memory_order_relaxed);
-        pnew_act->sigaction = (void (*)(int, void *, void *))new_cb;
-      }
-    }
-    res = REAL(sigaction)(signo, pnew_act, oldact);
-    if (res == 0 && oldact) {
-      uptr cb = (uptr)oldact->sigaction;
-      if (cb != __sanitizer::sig_ign && cb != __sanitizer::sig_dfl) {
-        oldact->sigaction = (void (*)(int, void *, void *))old_cb;
-      }
-    }
-  } else {
-    res = REAL(sigaction)(signo, act, oldact);
-  }
-
-  if (res == 0 && oldact) {
-    __msan_unpoison(oldact, sizeof(__sanitizer_sigaction));
-  }
-  return res;
-}
-
-INTERCEPTOR(int, signal, int signo, uptr cb) {
-  ENSURE_MSAN_INITED();
-  if (flags()->wrap_signals) {
-    CHECK_LT(signo, kMaxSignals);
-    SpinMutexLock lock(&sigactions_mu);
-    if (cb != __sanitizer::sig_ign && cb != __sanitizer::sig_dfl) {
-      atomic_store(&sigactions[signo], cb, memory_order_relaxed);
-      cb = (uptr) SignalHandler;
-    }
-    return REAL(signal)(signo, cb);
-  } else {
-    return REAL(signal)(signo, cb);
-  }
+static void read_sigaction(const __sanitizer_sigaction *act) {
+  CHECK_UNPOISONED(&act->sa_flags, sizeof(act->sa_flags));
+  if (act->sa_flags & __sanitizer::sa_siginfo)
+    CHECK_UNPOISONED(&act->sigaction, sizeof(act->sigaction));
+  else
+    CHECK_UNPOISONED(&act->handler, sizeof(act->handler));
+  CHECK_UNPOISONED(&act->sa_mask, sizeof(act->sa_mask));
 }
 
 extern "C" int pthread_attr_init(void *attr);
@@ -1080,6 +1108,11 @@ INTERCEPTOR(int, pthread_key_create, __sanitizer_pthread_key_t *key,
   return res;
 }
 
+#if SANITIZER_NETBSD
+INTERCEPTOR(void, __libc_thr_keycreate, void *m, void (*dtor)(void *value)) \
+  ALIAS(WRAPPER_NAME(pthread_key_create));
+#endif
+
 INTERCEPTOR(int, pthread_join, void *th, void **retval) {
   ENSURE_MSAN_INITED();
   int res = REAL(pthread_join)(th, retval);
@@ -1124,21 +1157,6 @@ INTERCEPTOR(int, __cxa_atexit, void (*func)(void *), void *arg,
   return REAL(__cxa_atexit)(MSanAtExitWrapper, r, dso_handle);
 }
 
-DECLARE_REAL(int, shmctl, int shmid, int cmd, void *buf)
-
-INTERCEPTOR(void *, shmat, int shmid, const void *shmaddr, int shmflg) {
-  ENSURE_MSAN_INITED();
-  void *p = REAL(shmat)(shmid, shmaddr, shmflg);
-  if (p != (void *)-1) {
-    __sanitizer_shmid_ds ds;
-    int res = REAL(shmctl)(shmid, shmctl_ipc_stat, &ds);
-    if (!res) {
-      __msan_unpoison(p, ds.shm_segsz);
-    }
-  }
-  return p;
-}
-
 static void BeforeFork() {
   StackDepotLockAll();
   ChainedOriginDepotLockAll();
@@ -1293,6 +1311,73 @@ int OnExit() {
 #include "sanitizer_common/sanitizer_platform_interceptors.h"
 #include "sanitizer_common/sanitizer_common_interceptors.inc"
 
+static uptr signal_impl(int signo, uptr cb);
+static int sigaction_impl(int signo, const __sanitizer_sigaction *act,
+                          __sanitizer_sigaction *oldact);
+
+#define SIGNAL_INTERCEPTOR_SIGACTION_IMPL(signo, act, oldact) \
+  { return sigaction_impl(signo, act, oldact); }
+
+#define SIGNAL_INTERCEPTOR_SIGNAL_IMPL(func, signo, handler) \
+  {                                                          \
+    handler = signal_impl(signo, handler);                   \
+    return REAL(func)(signo, handler);                       \
+  }
+
+#include "sanitizer_common/sanitizer_signal_interceptors.inc"
+
+static int sigaction_impl(int signo, const __sanitizer_sigaction *act,
+                          __sanitizer_sigaction *oldact) {
+  ENSURE_MSAN_INITED();
+  if (act) read_sigaction(act);
+  int res;
+  if (flags()->wrap_signals) {
+    SpinMutexLock lock(&sigactions_mu);
+    CHECK_LT(signo, kMaxSignals);
+    uptr old_cb = atomic_load(&sigactions[signo], memory_order_relaxed);
+    __sanitizer_sigaction new_act;
+    __sanitizer_sigaction *pnew_act = act ? &new_act : nullptr;
+    if (act) {
+      REAL(memcpy)(pnew_act, act, sizeof(__sanitizer_sigaction));
+      uptr cb = (uptr)pnew_act->sigaction;
+      uptr new_cb = (pnew_act->sa_flags & __sanitizer::sa_siginfo)
+                        ? (uptr)SignalAction
+                        : (uptr)SignalHandler;
+      if (cb != __sanitizer::sig_ign && cb != __sanitizer::sig_dfl) {
+        atomic_store(&sigactions[signo], cb, memory_order_relaxed);
+        pnew_act->sigaction = (decltype(pnew_act->sigaction))new_cb;
+      }
+    }
+    res = REAL(SIGACTION_SYMNAME)(signo, pnew_act, oldact);
+    if (res == 0 && oldact) {
+      uptr cb = (uptr)oldact->sigaction;
+      if (cb == (uptr)SignalAction || cb == (uptr)SignalHandler) {
+        oldact->sigaction = (decltype(oldact->sigaction))old_cb;
+      }
+    }
+  } else {
+    res = REAL(SIGACTION_SYMNAME)(signo, act, oldact);
+  }
+
+  if (res == 0 && oldact) {
+    __msan_unpoison(oldact, sizeof(__sanitizer_sigaction));
+  }
+  return res;
+}
+
+static uptr signal_impl(int signo, uptr cb) {
+  ENSURE_MSAN_INITED();
+  if (flags()->wrap_signals) {
+    CHECK_LT(signo, kMaxSignals);
+    SpinMutexLock lock(&sigactions_mu);
+    if (cb != __sanitizer::sig_ign && cb != __sanitizer::sig_dfl) {
+      atomic_store(&sigactions[signo], cb, memory_order_relaxed);
+      cb = (uptr)&SignalHandler;
+    }
+  }
+  return cb;
+}
+
 #define COMMON_SYSCALL_PRE_READ_RANGE(p, s) CHECK_UNPOISONED(p, s)
 #define COMMON_SYSCALL_PRE_WRITE_RANGE(p, s) \
   do {                                       \
@@ -1353,6 +1438,19 @@ static int msan_dl_iterate_phdr_cb(__sanitizer_dl_phdr_info *info, SIZE_T size,
   return cbdata->callback(info, size, cbdata->data);
 }
 
+INTERCEPTOR(void *, shmat, int shmid, const void *shmaddr, int shmflg) {
+  ENSURE_MSAN_INITED();
+  void *p = REAL(shmat)(shmid, shmaddr, shmflg);
+  if (p != (void *)-1) {
+    __sanitizer_shmid_ds ds;
+    int res = REAL(shmctl)(shmid, shmctl_ipc_stat, &ds);
+    if (!res) {
+      __msan_unpoison(p, ds.shm_segsz);
+    }
+  }
+  return p;
+}
+
 INTERCEPTOR(int, dl_iterate_phdr, dl_iterate_phdr_cb callback, void *data) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, dl_iterate_phdr, callback, data);
@@ -1455,12 +1553,13 @@ void InitializeInterceptors() {
   static int inited = 0;
   CHECK_EQ(inited, 0);
   InitializeCommonInterceptors();
+  InitializeSignalInterceptors();
 
   INTERCEPT_FUNCTION(mmap);
   MSAN_MAYBE_INTERCEPT_MMAP64;
   INTERCEPT_FUNCTION(posix_memalign);
   MSAN_MAYBE_INTERCEPT_MEMALIGN;
-  INTERCEPT_FUNCTION(__libc_memalign);
+  MSAN_MAYBE_INTERCEPT___LIBC_MEMALIGN;
   INTERCEPT_FUNCTION(valloc);
   MSAN_MAYBE_INTERCEPT_PVALLOC;
   INTERCEPT_FUNCTION(malloc);
@@ -1468,7 +1567,7 @@ void InitializeInterceptors() {
   INTERCEPT_FUNCTION(realloc);
   INTERCEPT_FUNCTION(free);
   MSAN_MAYBE_INTERCEPT_CFREE;
-  INTERCEPT_FUNCTION(malloc_usable_size);
+  MSAN_MAYBE_INTERCEPT_MALLOC_USABLE_SIZE;
   MSAN_MAYBE_INTERCEPT_MALLINFO;
   MSAN_MAYBE_INTERCEPT_MALLOPT;
   MSAN_MAYBE_INTERCEPT_MALLOC_STATS;
@@ -1476,18 +1575,18 @@ void InitializeInterceptors() {
   MSAN_MAYBE_INTERCEPT_FREAD_UNLOCKED;
   INTERCEPT_FUNCTION(readlink);
   INTERCEPT_FUNCTION(memccpy);
-  INTERCEPT_FUNCTION(mempcpy);
+  MSAN_MAYBE_INTERCEPT_MEMPCPY;
   INTERCEPT_FUNCTION(bcopy);
   INTERCEPT_FUNCTION(wmemset);
   INTERCEPT_FUNCTION(wmemcpy);
-  INTERCEPT_FUNCTION(wmempcpy);
+  MSAN_MAYBE_INTERCEPT_WMEMPCPY;
   INTERCEPT_FUNCTION(wmemmove);
   INTERCEPT_FUNCTION(strcpy);  // NOLINT
-  INTERCEPT_FUNCTION(stpcpy);  // NOLINT
+  MSAN_MAYBE_INTERCEPT_STPCPY;  // NOLINT
   INTERCEPT_FUNCTION(strdup);
   MSAN_MAYBE_INTERCEPT___STRDUP;
   INTERCEPT_FUNCTION(strncpy);  // NOLINT
-  INTERCEPT_FUNCTION(gcvt);
+  MSAN_MAYBE_INTERCEPT_GCVT;
   INTERCEPT_FUNCTION(strcat);  // NOLINT
   INTERCEPT_FUNCTION(strncat);  // NOLINT
   INTERCEPT_STRTO(strtod);
@@ -1497,6 +1596,7 @@ void InitializeInterceptors() {
   INTERCEPT_STRTO(strtoul);
   INTERCEPT_STRTO(strtoll);
   INTERCEPT_STRTO(strtoull);
+  INTERCEPT_STRTO(strtouq);
   INTERCEPT_STRTO(wcstod);
   INTERCEPT_STRTO(wcstof);
   INTERCEPT_STRTO(wcstold);
@@ -1513,6 +1613,7 @@ void InitializeInterceptors() {
 #endif
   INTERCEPT_FUNCTION(strxfrm);
   INTERCEPT_FUNCTION(strxfrm_l);
+  MSAN_MAYBE_INTERCEPT___STRXFRM_L;
   INTERCEPT_FUNCTION(strftime);
   INTERCEPT_FUNCTION(strftime_l);
   MSAN_MAYBE_INTERCEPT___STRFTIME_L;
@@ -1531,7 +1632,7 @@ void InitializeInterceptors() {
   INTERCEPT_FUNCTION(setenv);
   INTERCEPT_FUNCTION(putenv);
   INTERCEPT_FUNCTION(gettimeofday);
-  INTERCEPT_FUNCTION(fcvt);
+  MSAN_MAYBE_INTERCEPT_FCVT;
   MSAN_MAYBE_INTERCEPT___FXSTAT;
   MSAN_INTERCEPT_FSTATAT;
   MSAN_MAYBE_INTERCEPT___FXSTAT64;
@@ -1553,14 +1654,17 @@ void InitializeInterceptors() {
   INTERCEPT_FUNCTION(dlerror);
   INTERCEPT_FUNCTION(dl_iterate_phdr);
   INTERCEPT_FUNCTION(getrusage);
-  INTERCEPT_FUNCTION(sigaction);
-  INTERCEPT_FUNCTION(signal);
 #if defined(__mips__)
   INTERCEPT_FUNCTION_VER(pthread_create, "GLIBC_2.2");
 #else
   INTERCEPT_FUNCTION(pthread_create);
 #endif
   INTERCEPT_FUNCTION(pthread_key_create);
+
+#if SANITIZER_NETBSD
+  INTERCEPT_FUNCTION(__libc_thr_keycreate);
+#endif
+
   INTERCEPT_FUNCTION(pthread_join);
   INTERCEPT_FUNCTION(tzset);
   INTERCEPT_FUNCTION(__cxa_atexit);
diff --git a/contrib/compiler-rt/lib/msan/msan_linux.cc b/contrib/compiler-rt/lib/msan/msan_linux.cc
index 0a687f620c94..4e6321fcb918 100644
--- a/contrib/compiler-rt/lib/msan/msan_linux.cc
+++ b/contrib/compiler-rt/lib/msan/msan_linux.cc
@@ -9,11 +9,11 @@
 //
 // This file is a part of MemorySanitizer.
 //
-// Linux- and FreeBSD-specific code.
+// Linux-, NetBSD- and FreeBSD-specific code.
 //===----------------------------------------------------------------------===//
 
 #include "sanitizer_common/sanitizer_platform.h"
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD
 
 #include "msan.h"
 #include "msan_thread.h"
@@ -120,7 +120,7 @@ bool InitShadow(bool init_origins) {
     return false;
   }
 
-  const uptr maxVirtualAddress = GetMaxVirtualAddress();
+  const uptr maxVirtualAddress = GetMaxUserVirtualAddress();
 
   for (unsigned i = 0; i < kMemoryLayoutSize; ++i) {
     uptr start = kMemoryLayout[i].start;
@@ -213,4 +213,4 @@ void MsanTSDDtor(void *tsd) {
 
 } // namespace __msan
 
-#endif // SANITIZER_FREEBSD || SANITIZER_LINUX
+#endif // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD
diff --git a/contrib/compiler-rt/lib/msan/msan_report.cc b/contrib/compiler-rt/lib/msan/msan_report.cc
index 9a35c9c13de5..28c9bbabb3e9 100644
--- a/contrib/compiler-rt/lib/msan/msan_report.cc
+++ b/contrib/compiler-rt/lib/msan/msan_report.cc
@@ -30,10 +30,8 @@ namespace __msan {
 class Decorator: public __sanitizer::SanitizerCommonDecorator {
  public:
   Decorator() : SanitizerCommonDecorator() { }
-  const char *Warning()    { return Red(); }
   const char *Origin()     { return Magenta(); }
   const char *Name()   { return Green(); }
-  const char *End()    { return Default(); }
 };
 
 static void DescribeStackOrigin(const char *so, uptr pc) {
@@ -47,7 +45,7 @@ static void DescribeStackOrigin(const char *so, uptr pc) {
       "  %sUninitialized value was created by an allocation of '%s%s%s'"
       " in the stack frame of function '%s%s%s'%s\n",
       d.Origin(), d.Name(), s, d.Origin(), d.Name(), sep + 1, d.Origin(),
-      d.End());
+      d.Default());
   InternalFree(s);
 
   if (pc) {
@@ -66,7 +64,7 @@ static void DescribeOrigin(u32 id) {
     StackTrace stack;
     o = o.getNextChainedOrigin(&stack);
     Printf("  %sUninitialized value was stored to memory at%s\n", d.Origin(),
-        d.End());
+           d.Default());
     stack.Print();
   }
   if (o.isStackOrigin()) {
@@ -78,18 +76,19 @@ static void DescribeOrigin(u32 id) {
     switch (stack.tag) {
       case StackTrace::TAG_ALLOC:
         Printf("  %sUninitialized value was created by a heap allocation%s\n",
-               d.Origin(), d.End());
+               d.Origin(), d.Default());
         break;
       case StackTrace::TAG_DEALLOC:
         Printf("  %sUninitialized value was created by a heap deallocation%s\n",
-               d.Origin(), d.End());
+               d.Origin(), d.Default());
         break;
       case STACK_TRACE_TAG_POISON:
         Printf("  %sMemory was marked as uninitialized%s\n", d.Origin(),
-               d.End());
+               d.Default());
         break;
       default:
-        Printf("  %sUninitialized value was created%s\n", d.Origin(), d.End());
+        Printf("  %sUninitialized value was created%s\n", d.Origin(),
+               d.Default());
         break;
     }
     stack.Print();
@@ -99,12 +98,12 @@ static void DescribeOrigin(u32 id) {
 void ReportUMR(StackTrace *stack, u32 origin) {
   if (!__msan::flags()->report_umrs) return;
 
-  SpinMutexLock l(&CommonSanitizerReportMutex);
+  ScopedErrorReportLock l;
 
   Decorator d;
   Printf("%s", d.Warning());
   Report("WARNING: MemorySanitizer: use-of-uninitialized-value\n");
-  Printf("%s", d.End());
+  Printf("%s", d.Default());
   stack->Print();
   if (origin) {
     DescribeOrigin(origin);
@@ -113,14 +112,14 @@ void ReportUMR(StackTrace *stack, u32 origin) {
 }
 
 void ReportExpectedUMRNotFound(StackTrace *stack) {
-  SpinMutexLock l(&CommonSanitizerReportMutex);
+  ScopedErrorReportLock l;
 
   Printf("WARNING: Expected use of uninitialized value not found\n");
   stack->Print();
 }
 
 void ReportStats() {
-  SpinMutexLock l(&CommonSanitizerReportMutex);
+  ScopedErrorReportLock l;
 
   if (__msan_get_track_origins() > 0) {
     StackDepotStats *stack_depot_stats = StackDepotGetStats();
@@ -138,13 +137,13 @@ void ReportStats() {
 }
 
 void ReportAtExitStatistics() {
-  SpinMutexLock l(&CommonSanitizerReportMutex);
+  ScopedErrorReportLock l;
 
   if (msan_report_count > 0) {
     Decorator d;
     Printf("%s", d.Warning());
     Printf("MemorySanitizer: %d warnings reported.\n", msan_report_count);
-    Printf("%s", d.End());
+    Printf("%s", d.Default());
   }
 }
 
@@ -203,7 +202,7 @@ 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);
-  Printf("%s", d.End());
+  Printf("%s", d.Default());
   while (s < e) {
     // Line start.
     if (pos % 16 == 0) {
@@ -265,7 +264,7 @@ void ReportUMRInsideAddressRange(const char *what, const void *start, uptr size,
   Printf("%s", d.Warning());
   Printf("%sUninitialized bytes in %s%s%s at offset %zu inside [%p, %zu)%s\n",
          d.Warning(), d.Name(), what, d.Warning(), offset, start, size,
-         d.End());
+         d.Default());
   if (__sanitizer::Verbosity())
     DescribeMemoryRange(start, size);
 }
diff --git a/contrib/compiler-rt/lib/profile/GCDAProfiling.c b/contrib/compiler-rt/lib/profile/GCDAProfiling.c
index 138af6ec4033..ef299f002e20 100644
--- a/contrib/compiler-rt/lib/profile/GCDAProfiling.c
+++ b/contrib/compiler-rt/lib/profile/GCDAProfiling.c
@@ -20,9 +20,6 @@
 |*
 \*===----------------------------------------------------------------------===*/
 
-#include "InstrProfilingPort.h"
-#include "InstrProfilingUtil.h"
-
 #include <errno.h>
 #include <fcntl.h>
 #include <stdio.h>
@@ -34,9 +31,6 @@
 #else
 #include <sys/mman.h>
 #include <sys/file.h>
-#ifndef MAP_FILE
-#define MAP_FILE 0
-#endif
 #endif
 
 #if defined(__FreeBSD__) && defined(__i386__)
@@ -62,6 +56,9 @@ typedef unsigned int uint32_t;
 typedef unsigned long long uint64_t;
 #endif
 
+#include "InstrProfiling.h"
+#include "InstrProfilingUtil.h"
+
 /* #define DEBUG_GCDAPROFILING */
 
 /*
@@ -238,17 +235,17 @@ void llvm_gcda_start_file(const char *orig_filename, const char version[4],
 
   /* Try just opening the file. */
   new_file = 0;
-  fd = open(filename, O_RDWR);
+  fd = open(filename, O_RDWR | O_BINARY);
 
   if (fd == -1) {
     /* Try opening the file, creating it if necessary. */
     new_file = 1;
     mode = "w+b";
-    fd = open(filename, O_RDWR | O_CREAT, 0644);
+    fd = open(filename, O_RDWR | O_CREAT | O_BINARY, 0644);
     if (fd == -1) {
       /* Try creating the directories first then opening the file. */
       __llvm_profile_recursive_mkdir(filename);
-      fd = open(filename, O_RDWR | O_CREAT, 0644);
+      fd = open(filename, O_RDWR | O_CREAT | O_BINARY, 0644);
       if (fd == -1) {
         /* Bah! It's hopeless. */
         int errnum = errno;
@@ -263,7 +260,7 @@ void llvm_gcda_start_file(const char *orig_filename, const char version[4],
    * same GCDA. This can fail if the filesystem doesn't support it, but in that
    * case we'll just carry on with the old racy behaviour and hope for the best.
    */
-  flock(fd, LOCK_EX);
+  lprofLockFd(fd);
   output_file = fdopen(fd, mode);
 
   /* Initialize the write buffer. */
@@ -462,7 +459,7 @@ void llvm_gcda_end_file() {
       unmap_file();
     }
 
-    flock(fd, LOCK_UN);
+    lprofUnlockFd(fd);
     fclose(output_file);
     output_file = NULL;
     write_buffer = NULL;
diff --git a/contrib/compiler-rt/lib/profile/InstrProfData.inc b/contrib/compiler-rt/lib/profile/InstrProfData.inc
index be0dd4ad04bf..6a98dc7b9b85 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfData.inc
+++ b/contrib/compiler-rt/lib/profile/InstrProfData.inc
@@ -628,9 +628,9 @@ serializeValueProfDataFrom(ValueProfRecordClosure *Closure,
 /* Raw profile format version (start from 1). */
 #define INSTR_PROF_RAW_VERSION 4
 /* Indexed profile format version (start from 1). */
-#define INSTR_PROF_INDEX_VERSION 4
+#define INSTR_PROF_INDEX_VERSION 5
 /* Coverage mapping format vresion (start from 0). */
-#define INSTR_PROF_COVMAP_VERSION 1
+#define INSTR_PROF_COVMAP_VERSION 2
 
 /* Profile version is always of type uint64_t. Reserve the upper 8 bits in the
  * version for other variants of profile. We set the lowest bit of the upper 8
diff --git a/contrib/compiler-rt/lib/profile/InstrProfiling.c b/contrib/compiler-rt/lib/profile/InstrProfiling.c
index fe66fec50658..00b31e1ee7d4 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfiling.c
+++ b/contrib/compiler-rt/lib/profile/InstrProfiling.c
@@ -7,12 +7,14 @@
 |*
 \*===----------------------------------------------------------------------===*/
 
-#include "InstrProfiling.h"
-#include "InstrProfilingInternal.h"
 #include <limits.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
+
+#include "InstrProfiling.h"
+#include "InstrProfilingInternal.h"
+
 #define INSTR_PROF_VALUE_PROF_DATA
 #include "InstrProfData.inc"
 
diff --git a/contrib/compiler-rt/lib/profile/InstrProfilingFile.c b/contrib/compiler-rt/lib/profile/InstrProfilingFile.c
index d038bb9cb5b0..d7c0abbc16e3 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfilingFile.c
+++ b/contrib/compiler-rt/lib/profile/InstrProfilingFile.c
@@ -7,9 +7,6 @@
 |*
 \*===----------------------------------------------------------------------===*/
 
-#include "InstrProfiling.h"
-#include "InstrProfilingInternal.h"
-#include "InstrProfilingUtil.h"
 #include <errno.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -31,6 +28,10 @@
 #endif
 #endif
 
+#include "InstrProfiling.h"
+#include "InstrProfilingInternal.h"
+#include "InstrProfilingUtil.h"
+
 /* From where is profile name specified.
  * The order the enumerators define their
  * precedence. Re-order them may lead to
@@ -85,7 +86,6 @@ typedef struct lprofFilename {
 COMPILER_RT_WEAK lprofFilename lprofCurFilename = {0, 0, 0, {0}, {0},
                                                    0, 0, 0, PNS_unknown};
 
-int getpid(void);
 static int getCurFilenameLength();
 static const char *getCurFilename(char *FilenameBuf);
 static unsigned doMerging() { return lprofCurFilename.MergePoolSize; }
@@ -325,7 +325,7 @@ static int parseFilenamePattern(const char *FilenamePat,
     if (FilenamePat[I] == '%') {
       if (FilenamePat[++I] == 'p') {
         if (!NumPids++) {
-          if (snprintf(PidChars, MAX_PID_SIZE, "%d", getpid()) <= 0) {
+          if (snprintf(PidChars, MAX_PID_SIZE, "%ld", (long)getpid()) <= 0) {
             PROF_WARN("Unable to get pid for filename pattern %s. Using the "
                       "default name.",
                       FilenamePat);
@@ -519,8 +519,10 @@ void __llvm_profile_initialize_file(void) {
 
   EnvFilenamePat = getFilenamePatFromEnv();
   if (EnvFilenamePat) {
-    SelectedPat = EnvFilenamePat;
-    PNS = PNS_environment;
+    /* Pass CopyFilenamePat = 1, to ensure that the filename would be valid 
+       at the  moment when __llvm_profile_write_file() gets executed. */
+    parseAndSetFilename(EnvFilenamePat, PNS_environment, 1);
+    return;
   } else if (hasCommandLineOverrider) {
     SelectedPat = INSTR_PROF_PROFILE_NAME_VAR;
     PNS = PNS_command_line;
diff --git a/contrib/compiler-rt/lib/profile/InstrProfilingInternal.h b/contrib/compiler-rt/lib/profile/InstrProfilingInternal.h
index 36490ef7d433..40540ab5f205 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfilingInternal.h
+++ b/contrib/compiler-rt/lib/profile/InstrProfilingInternal.h
@@ -10,8 +10,9 @@
 #ifndef PROFILE_INSTRPROFILING_INTERNALH_
 #define PROFILE_INSTRPROFILING_INTERNALH_
 
+#include <stddef.h>
+
 #include "InstrProfiling.h"
-#include "stddef.h"
 
 /*!
  * \brief Write instrumentation data to the given buffer, given explicit
diff --git a/contrib/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c b/contrib/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c
index b6c780ff514f..89f1ab4cfeda 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c
+++ b/contrib/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c
@@ -7,11 +7,13 @@
 |*
 \*===----------------------------------------------------------------------===*/
 
-#include "InstrProfiling.h"
+#if defined(__linux__) || defined(__FreeBSD__) || \
+    (defined(__sun__) && defined(__svr4__))
 
-#if defined(__linux__) || defined(__FreeBSD__)
 #include <stdlib.h>
 
+#include "InstrProfiling.h"
+
 #define PROF_DATA_START INSTR_PROF_SECT_START(INSTR_PROF_DATA_SECT_NAME)
 #define PROF_DATA_STOP INSTR_PROF_SECT_STOP(INSTR_PROF_DATA_SECT_NAME)
 #define PROF_NAME_START INSTR_PROF_SECT_START(INSTR_PROF_NAME_SECT_NAME)
diff --git a/contrib/compiler-rt/lib/profile/InstrProfilingPlatformOther.c b/contrib/compiler-rt/lib/profile/InstrProfilingPlatformOther.c
index b25966487e91..a339abc7f883 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfilingPlatformOther.c
+++ b/contrib/compiler-rt/lib/profile/InstrProfilingPlatformOther.c
@@ -7,12 +7,13 @@
 |*
 \*===----------------------------------------------------------------------===*/
 
-#include "InstrProfiling.h"
-
-#if !defined(__APPLE__) && !defined(__linux__) && !defined(__FreeBSD__)
+#if !defined(__APPLE__) && !defined(__linux__) && !defined(__FreeBSD__) && \
+    !(defined(__sun__) && defined(__svr4__))
 
 #include <stdlib.h>
 
+#include "InstrProfiling.h"
+
 static const __llvm_profile_data *DataFirst = NULL;
 static const __llvm_profile_data *DataLast = NULL;
 static const char *NamesFirst = NULL;
diff --git a/contrib/compiler-rt/lib/profile/InstrProfilingPort.h b/contrib/compiler-rt/lib/profile/InstrProfilingPort.h
index 5789351956b9..cd1264b0e575 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfilingPort.h
+++ b/contrib/compiler-rt/lib/profile/InstrProfilingPort.h
@@ -7,6 +7,9 @@
 |*
 \*===----------------------------------------------------------------------===*/
 
+/* This header must be included after all others so it can provide fallback
+   definitions for stuff missing in system headers. */
+
 #ifndef PROFILE_INSTRPROFILING_PORT_H_
 #define PROFILE_INSTRPROFILING_PORT_H_
 
@@ -44,9 +47,6 @@
 #define COMPILER_RT_GETHOSTNAME(Name, Len) ((void)(Name), (void)(Len), (-1))
 #else
 #define COMPILER_RT_GETHOSTNAME(Name, Len) lprofGetHostName(Name, Len)
-#ifndef _MSC_VER
-#define COMPILER_RT_HAS_UNAME 1
-#endif
 #endif
 
 #if COMPILER_RT_HAS_ATOMICS == 1
@@ -107,6 +107,14 @@
 #define PROF_NOTE(Format, ...)                                                 \
   fprintf(stderr, "LLVM Profile Note: " Format, __VA_ARGS__);
 
+#ifndef MAP_FILE
+#define MAP_FILE 0
+#endif
+
+#ifndef O_BINARY
+#define O_BINARY 0
+#endif
+
 #if defined(__FreeBSD__)
 
 #include <inttypes.h>
diff --git a/contrib/compiler-rt/lib/profile/InstrProfilingUtil.c b/contrib/compiler-rt/lib/profile/InstrProfilingUtil.c
index fb68f30a5e1f..110562383412 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfilingUtil.c
+++ b/contrib/compiler-rt/lib/profile/InstrProfilingUtil.c
@@ -7,13 +7,10 @@
 |*
 \*===----------------------------------------------------------------------===*/
 
-#include "InstrProfilingUtil.h"
-#include "InstrProfiling.h"
-
 #ifdef _WIN32
 #include <direct.h>
-#include <io.h>
 #include <windows.h>
+#include "WindowsMMap.h"
 #else
 #include <sys/stat.h>
 #include <sys/types.h>
@@ -34,6 +31,9 @@
 #include <sys/prctl.h>
 #endif
 
+#include "InstrProfiling.h"
+#include "InstrProfilingUtil.h"
+
 COMPILER_RT_VISIBILITY
 void __llvm_profile_recursive_mkdir(char *path) {
   int i;
@@ -86,16 +86,16 @@ COMPILER_RT_VISIBILITY int lprofGetHostName(char *Name, int Len) {
 #elif defined(COMPILER_RT_HAS_UNAME)
 COMPILER_RT_VISIBILITY int lprofGetHostName(char *Name, int Len) {
   struct utsname N;
-  int R;
-  if (!(R = uname(&N)))
+  int R = uname(&N);
+  if (R >= 0) {
     strncpy(Name, N.nodename, Len);
+    return 0;
+  }
   return R;
 }
 #endif
 
-COMPILER_RT_VISIBILITY FILE *lprofOpenFileEx(const char *ProfileName) {
-  FILE *f;
-  int fd;
+COMPILER_RT_VISIBILITY int lprofLockFd(int fd) {
 #ifdef COMPILER_RT_HAS_FCNTL_LCK
   struct flock s_flock;
 
@@ -103,21 +103,59 @@ COMPILER_RT_VISIBILITY FILE *lprofOpenFileEx(const char *ProfileName) {
   s_flock.l_start = 0;
   s_flock.l_len = 0; /* Until EOF.  */
   s_flock.l_pid = getpid();
-
   s_flock.l_type = F_WRLCK;
-  fd = open(ProfileName, O_RDWR | O_CREAT, 0666);
-  if (fd < 0)
-    return NULL;
 
   while (fcntl(fd, F_SETLKW, &s_flock) == -1) {
     if (errno != EINTR) {
       if (errno == ENOLCK) {
-        PROF_WARN("Data may be corrupted during profile merging : %s\n",
-                  "Fail to obtain file lock due to system limit.");
+        return -1;
       }
       break;
     }
   }
+  return 0;
+#else
+  flock(fd, LOCK_EX);
+  return 0;
+#endif
+}
+
+COMPILER_RT_VISIBILITY int lprofUnlockFd(int fd) {
+#ifdef COMPILER_RT_HAS_FCNTL_LCK
+  struct flock s_flock;
+
+  s_flock.l_whence = SEEK_SET;
+  s_flock.l_start = 0;
+  s_flock.l_len = 0; /* Until EOF.  */
+  s_flock.l_pid = getpid();
+  s_flock.l_type = F_UNLCK;
+
+  while (fcntl(fd, F_SETLKW, &s_flock) == -1) {
+    if (errno != EINTR) {
+      if (errno == ENOLCK) {
+        return -1;
+      }
+      break;
+    }
+  }
+  return 0;
+#else
+  flock(fd, LOCK_UN);
+  return 0;
+#endif
+}
+
+COMPILER_RT_VISIBILITY FILE *lprofOpenFileEx(const char *ProfileName) {
+  FILE *f;
+  int fd;
+#ifdef COMPILER_RT_HAS_FCNTL_LCK
+  fd = open(ProfileName, O_RDWR | O_CREAT, 0666);
+  if (fd < 0)
+    return NULL;
+
+  if (lprofLockFd(fd) != 0)
+    PROF_WARN("Data may be corrupted during profile merging : %s\n",
+              "Fail to obtain file lock due to system limit.");
 
   f = fdopen(fd, "r+b");
 #elif defined(_WIN32)
diff --git a/contrib/compiler-rt/lib/profile/InstrProfilingUtil.h b/contrib/compiler-rt/lib/profile/InstrProfilingUtil.h
index 9698599606e0..2f2ea1b876a8 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfilingUtil.h
+++ b/contrib/compiler-rt/lib/profile/InstrProfilingUtil.h
@@ -16,6 +16,9 @@
 /*! \brief Create a directory tree. */
 void __llvm_profile_recursive_mkdir(char *Pathname);
 
+int lprofLockFd(int fd);
+int lprofUnlockFd(int fd);
+
 /*! Open file \c Filename for read+write with write
  * lock for exclusive access. The caller will block
  * if the lock is already held by another process. */
diff --git a/contrib/compiler-rt/lib/profile/InstrProfilingValue.c b/contrib/compiler-rt/lib/profile/InstrProfilingValue.c
index 44263da80097..3db0de8a6723 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfilingValue.c
+++ b/contrib/compiler-rt/lib/profile/InstrProfilingValue.c
@@ -7,13 +7,15 @@
 |*
 \*===----------------------------------------------------------------------===*/
 
-#include "InstrProfiling.h"
-#include "InstrProfilingInternal.h"
-#include "InstrProfilingUtil.h" /* For PS4 getenv shim. */
 #include <limits.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
+
+#include "InstrProfiling.h"
+#include "InstrProfilingInternal.h"
+#include "InstrProfilingUtil.h"
+
 #define INSTR_PROF_VALUE_PROF_DATA
 #define INSTR_PROF_COMMON_API_IMPL
 #include "InstrProfData.inc"
@@ -22,7 +24,7 @@ static int hasStaticCounters = 1;
 static int OutOfNodesWarnings = 0;
 static int hasNonDefaultValsPerSite = 0;
 #define INSTR_PROF_MAX_VP_WARNS 10
-#define INSTR_PROF_DEFAULT_NUM_VAL_PER_SITE 8
+#define INSTR_PROF_DEFAULT_NUM_VAL_PER_SITE 16
 #define INSTR_PROF_VNODE_POOL_SIZE 1024
 
 #ifndef _MSC_VER
diff --git a/contrib/compiler-rt/lib/profile/InstrProfilingWriter.c b/contrib/compiler-rt/lib/profile/InstrProfilingWriter.c
index d4c9b9bd663c..7c6061d2c65a 100644
--- a/contrib/compiler-rt/lib/profile/InstrProfilingWriter.c
+++ b/contrib/compiler-rt/lib/profile/InstrProfilingWriter.c
@@ -7,14 +7,15 @@
 |*
 \*===----------------------------------------------------------------------===*/
 
-#include "InstrProfiling.h"
-#include "InstrProfilingInternal.h"
 #ifdef _MSC_VER
 /* For _alloca */
 #include <malloc.h>
 #endif
 #include <string.h>
 
+#include "InstrProfiling.h"
+#include "InstrProfilingInternal.h"
+
 #define INSTR_PROF_VALUE_PROF_DATA
 #include "InstrProfData.inc"
 
diff --git a/contrib/compiler-rt/lib/profile/WindowsMMap.c b/contrib/compiler-rt/lib/profile/WindowsMMap.c
index f81d7da5389e..dc87a888ae7b 100644
--- a/contrib/compiler-rt/lib/profile/WindowsMMap.c
+++ b/contrib/compiler-rt/lib/profile/WindowsMMap.c
@@ -18,11 +18,12 @@
 #if defined(_WIN32)
 
 #include "WindowsMMap.h"
-#include "InstrProfiling.h"
 
 #define WIN32_LEAN_AND_MEAN
 #include <windows.h>
 
+#include "InstrProfiling.h"
+
 #ifdef __USE_FILE_OFFSET64
 # define DWORD_HI(x) (x >> 32)
 # define DWORD_LO(x) ((x) & 0xffffffff)
@@ -120,9 +121,61 @@ int msync(void *addr, size_t length, int flags)
 }
 
 COMPILER_RT_VISIBILITY
-int flock(int fd, int operation)
-{
-  return -1; /* Not supported. */
+int lock(HANDLE handle, DWORD lockType, BOOL blocking) {
+  DWORD flags = lockType;
+  if (!blocking)
+    flags |= LOCKFILE_FAIL_IMMEDIATELY;
+
+  OVERLAPPED overlapped;
+  ZeroMemory(&overlapped, sizeof(OVERLAPPED));
+  overlapped.hEvent = CreateEvent(NULL, FALSE, FALSE, NULL);
+  BOOL result = LockFileEx(handle, flags, 0, MAXDWORD, MAXDWORD, &overlapped);
+  if (!result) {
+    DWORD dw = GetLastError();
+
+    // In non-blocking mode, return an error if the file is locked.
+    if (!blocking && dw == ERROR_LOCK_VIOLATION)
+      return -1; // EWOULDBLOCK
+
+    // If the error is ERROR_IO_PENDING, we need to wait until the operation
+    // finishes. Otherwise, we return an error.
+    if (dw != ERROR_IO_PENDING)
+      return -1;
+
+    DWORD dwNumBytes;
+    if (!GetOverlappedResult(handle, &overlapped, &dwNumBytes, TRUE))
+      return -1;
+  }
+
+  return 0;
+}
+
+COMPILER_RT_VISIBILITY
+int flock(int fd, int operation) {
+  HANDLE handle = (HANDLE)_get_osfhandle(fd);
+  if (handle == INVALID_HANDLE_VALUE)
+    return -1;
+
+  BOOL blocking = (operation & LOCK_NB) == 0;
+  int op = operation & ~LOCK_NB;
+
+  switch (op) {
+  case LOCK_EX:
+    return lock(handle, LOCKFILE_EXCLUSIVE_LOCK, blocking);
+
+  case LOCK_SH:
+    return lock(handle, 0, blocking);
+
+  case LOCK_UN:
+    if (!UnlockFile(handle, 0, 0, MAXDWORD, MAXDWORD))
+      return -1;
+    break;
+
+  default:
+    return -1;
+  }
+
+  return 0;
 }
 
 #undef DWORD_HI
diff --git a/contrib/compiler-rt/lib/safestack/safestack.cc b/contrib/compiler-rt/lib/safestack/safestack.cc
index b194b6cfa8f9..d783cd5a9b29 100644
--- a/contrib/compiler-rt/lib/safestack/safestack.cc
+++ b/contrib/compiler-rt/lib/safestack/safestack.cc
@@ -21,7 +21,9 @@
 #include <unistd.h>
 #include <sys/resource.h>
 #include <sys/types.h>
+#if !defined(__NetBSD__)
 #include <sys/user.h>
+#endif
 
 #include "interception/interception.h"
 #include "sanitizer_common/sanitizer_common.h"
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator.cc
index 84f523c5e431..fc4f7a75ae34 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator.cc
@@ -178,11 +178,13 @@ void InternalFree(void *addr, InternalAllocatorCache *cache) {
 }
 
 // LowLevelAllocator
+constexpr uptr kLowLevelAllocatorDefaultAlignment = 8;
+static uptr low_level_alloc_min_alignment = kLowLevelAllocatorDefaultAlignment;
 static LowLevelAllocateCallback low_level_alloc_callback;
 
 void *LowLevelAllocator::Allocate(uptr size) {
   // Align allocation size.
-  size = RoundUpTo(size, 8);
+  size = RoundUpTo(size, low_level_alloc_min_alignment);
   if (allocated_end_ - allocated_current_ < (sptr)size) {
     uptr size_to_allocate = Max(size, GetPageSizeCached());
     allocated_current_ =
@@ -199,6 +201,11 @@ void *LowLevelAllocator::Allocate(uptr size) {
   return res;
 }
 
+void SetLowLevelAllocateMinAlignment(uptr alignment) {
+  CHECK(IsPowerOfTwo(alignment));
+  low_level_alloc_min_alignment = Max(alignment, low_level_alloc_min_alignment);
+}
+
 void SetLowLevelAllocateCallback(LowLevelAllocateCallback callback) {
   low_level_alloc_callback = callback;
 }
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator.h
index 8c5696ea789c..38368361de6f 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator.h
@@ -56,6 +56,19 @@ struct NoOpMapUnmapCallback {
 // Callback type for iterating over chunks.
 typedef void (*ForEachChunkCallback)(uptr chunk, void *arg);
 
+INLINE u32 Rand(u32 *state) {  // ANSI C linear congruential PRNG.
+  return (*state = *state * 1103515245 + 12345) >> 16;
+}
+
+INLINE u32 RandN(u32 *state, u32 n) { return Rand(state) % n; }  // [0, n)
+
+template<typename T>
+INLINE void RandomShuffle(T *a, u32 n, u32 *rand_state) {
+  if (n <= 1) return;
+  for (u32 i = n - 1; i > 0; i--)
+    Swap(a[i], a[RandN(rand_state, i + 1)]);
+}
+
 #include "sanitizer_allocator_size_class_map.h"
 #include "sanitizer_allocator_stats.h"
 #include "sanitizer_allocator_primary64.h"
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_checks.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_checks.cc
new file mode 100644
index 000000000000..dc263dbef5fc
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_checks.cc
@@ -0,0 +1,23 @@
+//===-- sanitizer_allocator_checks.cc ---------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Various checks shared between ThreadSanitizer, MemorySanitizer, etc. memory
+// allocators.
+//
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_errno.h"
+
+namespace __sanitizer {
+
+void SetErrnoToENOMEM() {
+  errno = errno_ENOMEM;
+}
+
+} // namespace __sanitizer
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_checks.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_checks.h
index 202916eae348..b61a8b2eb3b6 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_checks.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_checks.h
@@ -15,17 +15,22 @@
 #ifndef SANITIZER_ALLOCATOR_CHECKS_H
 #define SANITIZER_ALLOCATOR_CHECKS_H
 
-#include "sanitizer_errno.h"
 #include "sanitizer_internal_defs.h"
 #include "sanitizer_common.h"
 #include "sanitizer_platform.h"
 
 namespace __sanitizer {
 
+// The following is defined in a separate compilation unit to avoid pulling in
+// sanitizer_errno.h in this header, which leads to conflicts when other system
+// headers include errno.h. This is usually the result of an unlikely event,
+// and as such we do not care as much about having it inlined.
+void SetErrnoToENOMEM();
+
 // A common errno setting logic shared by almost all sanitizer allocator APIs.
 INLINE void *SetErrnoOnNull(void *ptr) {
   if (UNLIKELY(!ptr))
-    errno = errno_ENOMEM;
+    SetErrnoToENOMEM();
   return ptr;
 }
 
@@ -59,6 +64,12 @@ INLINE bool CheckForCallocOverflow(uptr size, uptr n) {
   return (max / size) < n;
 }
 
+// Returns true if the size passed to pvalloc overflows when rounded to the next
+// multiple of page_size.
+INLINE bool CheckForPvallocOverflow(uptr size, uptr page_size) {
+  return RoundUpTo(size, page_size) < size;
+}
+
 } // namespace __sanitizer
 
 #endif  // SANITIZER_ALLOCATOR_CHECKS_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_combined.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_combined.h
index efd25cadfe74..0d8a2a174bb5 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_combined.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_combined.h
@@ -77,6 +77,10 @@ class CombinedAllocator {
     primary_.SetReleaseToOSIntervalMs(release_to_os_interval_ms);
   }
 
+  void ForceReleaseToOS() {
+    primary_.ForceReleaseToOS();
+  }
+
   void Deallocate(AllocatorCache *cache, void *p) {
     if (!p) return;
     if (primary_.PointerIsMine(p))
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_interface.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_interface.h
index 13910e719e78..2f5ce3151ee4 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_interface.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_interface.h
@@ -39,6 +39,9 @@ SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
     void __sanitizer_free_hook(void *ptr);
 
 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE void
+__sanitizer_purge_allocator();
+
+SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE void
 __sanitizer_print_memory_profile(uptr top_percent, uptr max_number_of_contexts);
 }  // extern "C"
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_local_cache.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_local_cache.h
index ec0742c20be2..1b3c2c0c19d1 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_local_cache.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_local_cache.h
@@ -26,9 +26,6 @@ struct SizeClassAllocatorLocalCache
 template <class SizeClassAllocator>
 struct SizeClassAllocator64LocalCache {
   typedef SizeClassAllocator Allocator;
-  static const uptr kNumClasses = SizeClassAllocator::kNumClasses;
-  typedef typename Allocator::SizeClassMapT SizeClassMap;
-  typedef typename Allocator::CompactPtrT CompactPtrT;
 
   void Init(AllocatorGlobalStats *s) {
     stats_.Init();
@@ -76,14 +73,18 @@ struct SizeClassAllocator64LocalCache {
   }
 
   void Drain(SizeClassAllocator *allocator) {
-    for (uptr class_id = 0; class_id < kNumClasses; class_id++) {
-      PerClass *c = &per_class_[class_id];
+    for (uptr i = 0; i < kNumClasses; i++) {
+      PerClass *c = &per_class_[i];
       while (c->count > 0)
-        Drain(c, allocator, class_id, c->count);
+        Drain(c, allocator, i, c->count);
     }
   }
 
-  // private:
+ private:
+  typedef typename Allocator::SizeClassMapT SizeClassMap;
+  static const uptr kNumClasses = SizeClassMap::kNumClasses;
+  typedef typename Allocator::CompactPtrT CompactPtrT;
+
   struct PerClass {
     u32 count;
     u32 max_count;
@@ -94,7 +95,7 @@ struct SizeClassAllocator64LocalCache {
   AllocatorStats stats_;
 
   void InitCache() {
-    if (per_class_[1].max_count)
+    if (LIKELY(per_class_[1].max_count))
       return;
     for (uptr i = 0; i < kNumClasses; i++) {
       PerClass *c = &per_class_[i];
@@ -130,7 +131,6 @@ template <class SizeClassAllocator>
 struct SizeClassAllocator32LocalCache {
   typedef SizeClassAllocator Allocator;
   typedef typename Allocator::TransferBatch TransferBatch;
-  static const uptr kNumClasses = SizeClassAllocator::kNumClasses;
 
   void Init(AllocatorGlobalStats *s) {
     stats_.Init();
@@ -138,6 +138,21 @@ struct SizeClassAllocator32LocalCache {
       s->Register(&stats_);
   }
 
+  // Returns a TransferBatch suitable for class_id.
+  TransferBatch *CreateBatch(uptr class_id, SizeClassAllocator *allocator,
+                             TransferBatch *b) {
+    if (uptr batch_class_id = per_class_[class_id].batch_class_id)
+      return (TransferBatch*)Allocate(allocator, batch_class_id);
+    return b;
+  }
+
+  // Destroys TransferBatch b.
+  void DestroyBatch(uptr class_id, SizeClassAllocator *allocator,
+                    TransferBatch *b) {
+    if (uptr batch_class_id = per_class_[class_id].batch_class_id)
+      Deallocate(allocator, batch_class_id, b);
+  }
+
   void Destroy(SizeClassAllocator *allocator, AllocatorGlobalStats *s) {
     Drain(allocator);
     if (s)
@@ -173,66 +188,57 @@ struct SizeClassAllocator32LocalCache {
   }
 
   void Drain(SizeClassAllocator *allocator) {
-    for (uptr class_id = 0; class_id < kNumClasses; class_id++) {
-      PerClass *c = &per_class_[class_id];
+    for (uptr i = 0; i < kNumClasses; i++) {
+      PerClass *c = &per_class_[i];
       while (c->count > 0)
-        Drain(allocator, class_id);
+        Drain(allocator, i);
     }
   }
 
-  // private:
-  typedef typename SizeClassAllocator::SizeClassMapT SizeClassMap;
+ private:
+  typedef typename Allocator::SizeClassMapT SizeClassMap;
+  static const uptr kBatchClassID = SizeClassMap::kBatchClassID;
+  static const uptr kNumClasses = SizeClassMap::kNumClasses;
+  // If kUseSeparateSizeClassForBatch is true, all TransferBatch objects are
+  // allocated from kBatchClassID size class (except for those that are needed
+  // for kBatchClassID itself). The goal is to have TransferBatches in a totally
+  // different region of RAM to improve security.
+  static const bool kUseSeparateSizeClassForBatch =
+      Allocator::kUseSeparateSizeClassForBatch;
+
   struct PerClass {
     uptr count;
     uptr max_count;
     uptr class_size;
-    uptr class_id_for_transfer_batch;
+    uptr batch_class_id;
     void *batch[2 * TransferBatch::kMaxNumCached];
   };
   PerClass per_class_[kNumClasses];
   AllocatorStats stats_;
 
   void InitCache() {
-    if (per_class_[1].max_count)
+    if (LIKELY(per_class_[1].max_count))
       return;
-    // TransferBatch class is declared in SizeClassAllocator.
-    uptr class_id_for_transfer_batch =
-        SizeClassMap::ClassID(sizeof(TransferBatch));
+    const uptr batch_class_id = SizeClassMap::ClassID(sizeof(TransferBatch));
     for (uptr i = 0; i < kNumClasses; i++) {
       PerClass *c = &per_class_[i];
       uptr max_cached = TransferBatch::MaxCached(i);
       c->max_count = 2 * max_cached;
       c->class_size = Allocator::ClassIdToSize(i);
-      // We transfer chunks between central and thread-local free lists in
-      // batches. For small size classes we allocate batches separately. For
-      // large size classes we may use one of the chunks to store the batch.
-      // sizeof(TransferBatch) must be a power of 2 for more efficient
-      // allocation.
-      c->class_id_for_transfer_batch = (c->class_size <
+      // Precompute the class id to use to store batches for the current class
+      // id. 0 means the class size is large enough to store a batch within one
+      // of the chunks. If using a separate size class, it will always be
+      // kBatchClassID, except for kBatchClassID itself.
+      if (kUseSeparateSizeClassForBatch) {
+        c->batch_class_id = (i == kBatchClassID) ? 0 : kBatchClassID;
+      } else {
+        c->batch_class_id = (c->class_size <
           TransferBatch::AllocationSizeRequiredForNElements(max_cached)) ?
-              class_id_for_transfer_batch : 0;
+              batch_class_id : 0;
+      }
     }
   }
 
-  // Returns a TransferBatch suitable for class_id.
-  // For small size classes allocates the batch from the allocator.
-  // For large size classes simply returns b.
-  TransferBatch *CreateBatch(uptr class_id, SizeClassAllocator *allocator,
-                             TransferBatch *b) {
-    if (uptr batch_class_id = per_class_[class_id].class_id_for_transfer_batch)
-      return (TransferBatch*)Allocate(allocator, batch_class_id);
-    return b;
-  }
-
-  // Destroys TransferBatch b.
-  // For small size classes deallocates b to the allocator.
-  // Does notthing for large size classes.
-  void DestroyBatch(uptr class_id, SizeClassAllocator *allocator,
-                    TransferBatch *b) {
-    if (uptr batch_class_id = per_class_[class_id].class_id_for_transfer_batch)
-      Deallocate(allocator, batch_class_id, b);
-  }
-
   NOINLINE bool Refill(SizeClassAllocator *allocator, uptr class_id) {
     InitCache();
     PerClass *c = &per_class_[class_id];
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary32.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary32.h
index e85821543d27..6c682aff64a2 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary32.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary32.h
@@ -41,6 +41,7 @@ template<class SizeClassAllocator> struct SizeClassAllocator32LocalCache;
 struct SizeClassAllocator32FlagMasks {  //  Bit masks.
   enum {
     kRandomShuffleChunks = 1,
+    kUseSeparateSizeClassForBatch = 2,
   };
 };
 
@@ -55,8 +56,10 @@ class SizeClassAllocator32 {
   typedef typename Params::ByteMap ByteMap;
   typedef typename Params::MapUnmapCallback MapUnmapCallback;
 
-  static const bool kRandomShuffleChunks =
-      Params::kFlags & SizeClassAllocator32FlagMasks::kRandomShuffleChunks;
+  static const bool kRandomShuffleChunks = Params::kFlags &
+      SizeClassAllocator32FlagMasks::kRandomShuffleChunks;
+  static const bool kUseSeparateSizeClassForBatch = Params::kFlags &
+      SizeClassAllocator32FlagMasks::kUseSeparateSizeClassForBatch;
 
   struct TransferBatch {
     static const uptr kMaxNumCached = SizeClassMap::kMaxNumCachedHint - 2;
@@ -94,11 +97,11 @@ class SizeClassAllocator32 {
 
   static const uptr kBatchSize = sizeof(TransferBatch);
   COMPILER_CHECK((kBatchSize & (kBatchSize - 1)) == 0);
-  COMPILER_CHECK(sizeof(TransferBatch) ==
-                 SizeClassMap::kMaxNumCachedHint * sizeof(uptr));
+  COMPILER_CHECK(kBatchSize == SizeClassMap::kMaxNumCachedHint * sizeof(uptr));
 
   static uptr ClassIdToSize(uptr class_id) {
-    return SizeClassMap::Size(class_id);
+    return (class_id == SizeClassMap::kBatchClassID) ?
+        kBatchSize : SizeClassMap::Size(class_id);
   }
 
   typedef SizeClassAllocator32<Params> ThisT;
@@ -117,6 +120,10 @@ class SizeClassAllocator32 {
     // This is empty here. Currently only implemented in 64-bit allocator.
   }
 
+  void ForceReleaseToOS() {
+    // Currently implemented in 64-bit allocator only.
+  }
+
   void *MapWithCallback(uptr size) {
     void *res = MmapOrDie(size, "SizeClassAllocator32");
     MapUnmapCallback().OnMap((uptr)res, size);
@@ -161,9 +168,9 @@ class SizeClassAllocator32 {
   NOINLINE void DeallocateBatch(AllocatorStats *stat, uptr class_id,
                                 TransferBatch *b) {
     CHECK_LT(class_id, kNumClasses);
+    CHECK_GT(b->Count(), 0);
     SizeClassInfo *sci = GetSizeClassInfo(class_id);
     SpinMutexLock l(&sci->mutex);
-    CHECK_GT(b->Count(), 0);
     sci->free_list.push_front(b);
   }
 
@@ -261,7 +268,8 @@ class SizeClassAllocator32 {
   struct SizeClassInfo {
     SpinMutex mutex;
     IntrusiveList<TransferBatch> free_list;
-    char padding[kCacheLineSize - sizeof(uptr) -
+    u32 rand_state;
+    char padding[kCacheLineSize - 2 * sizeof(uptr) -
                  sizeof(IntrusiveList<TransferBatch>)];
   };
   COMPILER_CHECK(sizeof(SizeClassInfo) == kCacheLineSize);
@@ -294,29 +302,62 @@ class SizeClassAllocator32 {
     return &size_class_info_array[class_id];
   }
 
+  bool PopulateBatches(AllocatorCache *c, SizeClassInfo *sci, uptr class_id,
+                       TransferBatch **current_batch, uptr max_count,
+                       uptr *pointers_array, uptr count) {
+    // If using a separate class for batches, we do not need to shuffle it.
+    if (kRandomShuffleChunks && (!kUseSeparateSizeClassForBatch ||
+        class_id != SizeClassMap::kBatchClassID))
+      RandomShuffle(pointers_array, count, &sci->rand_state);
+    TransferBatch *b = *current_batch;
+    for (uptr i = 0; i < count; i++) {
+      if (!b) {
+        b = c->CreateBatch(class_id, this, (TransferBatch*)pointers_array[i]);
+        if (UNLIKELY(!b))
+          return false;
+        b->Clear();
+      }
+      b->Add((void*)pointers_array[i]);
+      if (b->Count() == max_count) {
+        sci->free_list.push_back(b);
+        b = nullptr;
+      }
+    }
+    *current_batch = b;
+    return true;
+  }
+
   bool PopulateFreeList(AllocatorStats *stat, AllocatorCache *c,
                         SizeClassInfo *sci, uptr class_id) {
     uptr size = ClassIdToSize(class_id);
     uptr reg = AllocateRegion(stat, class_id);
     if (UNLIKELY(!reg))
       return false;
+    if (kRandomShuffleChunks)
+      if (UNLIKELY(sci->rand_state == 0))
+        // The random state is initialized from ASLR (PIE) and time.
+        sci->rand_state = reinterpret_cast<uptr>(sci) ^ NanoTime();
     uptr n_chunks = kRegionSize / (size + kMetadataSize);
     uptr max_count = TransferBatch::MaxCached(class_id);
     CHECK_GT(max_count, 0);
     TransferBatch *b = nullptr;
+    const uptr kShuffleArraySize = 48;
+    uptr shuffle_array[kShuffleArraySize];
+    uptr count = 0;
     for (uptr i = reg; i < reg + n_chunks * size; i += size) {
-      if (!b) {
-        b = c->CreateBatch(class_id, this, (TransferBatch*)i);
-        if (UNLIKELY(!b))
+      shuffle_array[count++] = i;
+      if (count == kShuffleArraySize) {
+        if (UNLIKELY(!PopulateBatches(c, sci, class_id, &b, max_count,
+                                      shuffle_array, count)))
           return false;
-        b->Clear();
-      }
-      b->Add((void*)i);
-      if (b->Count() == max_count) {
-        sci->free_list.push_back(b);
-        b = nullptr;
+        count = 0;
       }
     }
+    if (count) {
+      if (UNLIKELY(!PopulateBatches(c, sci, class_id, &b, max_count,
+                                    shuffle_array, count)))
+        return false;
+    }
     if (b) {
       CHECK_GT(b->Count(), 0);
       sci->free_list.push_back(b);
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary64.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary64.h
index 0c2e72ce7eb1..651a64b04f2e 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary64.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_primary64.h
@@ -62,21 +62,20 @@ class SizeClassAllocator64 {
   // as a 4-byte integer (offset from the region start shifted right by 4).
   typedef u32 CompactPtrT;
   static const uptr kCompactPtrScale = 4;
-  CompactPtrT PointerToCompactPtr(uptr base, uptr ptr) {
+  CompactPtrT PointerToCompactPtr(uptr base, uptr ptr) const {
     return static_cast<CompactPtrT>((ptr - base) >> kCompactPtrScale);
   }
-  uptr CompactPtrToPointer(uptr base, CompactPtrT ptr32) {
+  uptr CompactPtrToPointer(uptr base, CompactPtrT ptr32) const {
     return base + (static_cast<uptr>(ptr32) << kCompactPtrScale);
   }
 
   void Init(s32 release_to_os_interval_ms) {
     uptr TotalSpaceSize = kSpaceSize + AdditionalSize();
     if (kUsingConstantSpaceBeg) {
-      CHECK_EQ(kSpaceBeg, reinterpret_cast<uptr>(
-                              MmapFixedNoAccess(kSpaceBeg, TotalSpaceSize)));
+      CHECK_EQ(kSpaceBeg, address_range.Init(TotalSpaceSize, AllocatorName(),
+                                             kSpaceBeg));
     } else {
-      NonConstSpaceBeg =
-          reinterpret_cast<uptr>(MmapNoAccess(TotalSpaceSize));
+      NonConstSpaceBeg = address_range.Init(TotalSpaceSize, AllocatorName());
       CHECK_NE(NonConstSpaceBeg, ~(uptr)0);
     }
     SetReleaseToOSIntervalMs(release_to_os_interval_ms);
@@ -92,6 +91,13 @@ class SizeClassAllocator64 {
                  memory_order_relaxed);
   }
 
+  void ForceReleaseToOS() {
+    for (uptr class_id = 1; class_id < kNumClasses; class_id++) {
+      BlockingMutexLock l(&GetRegionInfo(class_id)->mutex);
+      MaybeReleaseToOS(class_id, true /*force*/);
+    }
+  }
+
   static bool CanAllocate(uptr size, uptr alignment) {
     return size <= SizeClassMap::kMaxSize &&
       alignment <= SizeClassMap::kMaxSize;
@@ -116,7 +122,7 @@ class SizeClassAllocator64 {
     region->num_freed_chunks = new_num_freed_chunks;
     region->stats.n_freed += n_chunks;
 
-    MaybeReleaseToOS(class_id);
+    MaybeReleaseToOS(class_id, false /*force*/);
   }
 
   NOINLINE bool GetFromAllocator(AllocatorStats *stat, uptr class_id,
@@ -155,7 +161,7 @@ class SizeClassAllocator64 {
         space_beg;
   }
 
-  uptr GetRegionBeginBySizeClass(uptr class_id) {
+  uptr GetRegionBeginBySizeClass(uptr class_id) const {
     return SpaceBeg() + kRegionSize * class_id;
   }
 
@@ -223,30 +229,39 @@ class SizeClassAllocator64 {
     uptr avail_chunks = region->allocated_user / ClassIdToSize(class_id);
     Printf(
         "%s %02zd (%6zd): mapped: %6zdK allocs: %7zd frees: %7zd inuse: %6zd "
-        "num_freed_chunks %7zd avail: %6zd rss: %6zdK releases: %6zd\n",
+        "num_freed_chunks %7zd avail: %6zd rss: %6zdK releases: %6zd "
+        "last released: %6zdK 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,
-        rss >> 10, region->rtoi.num_releases);
+        rss >> 10, region->rtoi.num_releases,
+        region->rtoi.last_released_bytes >> 10,
+        SpaceBeg() + kRegionSize * class_id);
   }
 
   void PrintStats() {
+    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);
+
     uptr total_mapped = 0;
+    uptr total_rss = 0;
     uptr n_allocated = 0;
     uptr n_freed = 0;
     for (uptr class_id = 1; class_id < kNumClasses; class_id++) {
       RegionInfo *region = GetRegionInfo(class_id);
-      total_mapped += region->mapped_user;
+      if (region->mapped_user != 0) {
+        total_mapped += region->mapped_user;
+        total_rss += rss_stats[class_id];
+      }
       n_allocated += region->stats.n_allocated;
       n_freed += region->stats.n_freed;
     }
-    Printf("Stats: SizeClassAllocator64: %zdM mapped in %zd allocations; "
-           "remains %zd\n",
-           total_mapped >> 20, n_allocated, n_allocated - n_freed);
-    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);
+
+    Printf("Stats: SizeClassAllocator64: %zdM mapped (%zdM rss) in "
+           "%zd allocations; remains %zd\n", total_mapped >> 20,
+           total_rss >> 20, n_allocated, n_allocated - n_freed);
     for (uptr class_id = 1; class_id < kNumClasses; class_id++)
       PrintStats(class_id, rss_stats[class_id]);
   }
@@ -294,7 +309,243 @@ class SizeClassAllocator64 {
   static const uptr kNumClasses = SizeClassMap::kNumClasses;
   static const uptr kNumClassesRounded = SizeClassMap::kNumClassesRounded;
 
+  // A packed array of counters. Each counter occupies 2^n bits, enough to store
+  // counter's max_value. Ctor will try to allocate the required buffer via
+  // mapper->MapPackedCounterArrayBuffer and the caller is expected to check
+  // whether the initialization was successful by checking IsAllocated() result.
+  // For the performance sake, none of the accessors check the validity of the
+  // arguments, it is assumed that index is always in [0, n) range and the value
+  // is not incremented past max_value.
+  template<class MemoryMapperT>
+  class PackedCounterArray {
+   public:
+    PackedCounterArray(u64 num_counters, u64 max_value, MemoryMapperT *mapper)
+        : n(num_counters), memory_mapper(mapper) {
+      CHECK_GT(num_counters, 0);
+      CHECK_GT(max_value, 0);
+      constexpr u64 kMaxCounterBits = sizeof(*buffer) * 8ULL;
+      // Rounding counter storage size up to the power of two allows for using
+      // bit shifts calculating particular counter's index and offset.
+      uptr counter_size_bits =
+          RoundUpToPowerOfTwo(MostSignificantSetBitIndex(max_value) + 1);
+      CHECK_LE(counter_size_bits, kMaxCounterBits);
+      counter_size_bits_log = Log2(counter_size_bits);
+      counter_mask = ~0ULL >> (kMaxCounterBits - counter_size_bits);
+
+      uptr packing_ratio = kMaxCounterBits >> counter_size_bits_log;
+      CHECK_GT(packing_ratio, 0);
+      packing_ratio_log = Log2(packing_ratio);
+      bit_offset_mask = packing_ratio - 1;
+
+      buffer_size =
+          (RoundUpTo(n, 1ULL << packing_ratio_log) >> packing_ratio_log) *
+          sizeof(*buffer);
+      buffer = reinterpret_cast<u64*>(
+          memory_mapper->MapPackedCounterArrayBuffer(buffer_size));
+    }
+    ~PackedCounterArray() {
+      if (buffer) {
+        memory_mapper->UnmapPackedCounterArrayBuffer(
+            reinterpret_cast<uptr>(buffer), buffer_size);
+      }
+    }
+
+    bool IsAllocated() const {
+      return !!buffer;
+    }
+
+    u64 GetCount() const {
+      return n;
+    }
+
+    uptr Get(uptr i) const {
+      DCHECK_LT(i, n);
+      uptr index = i >> packing_ratio_log;
+      uptr bit_offset = (i & bit_offset_mask) << counter_size_bits_log;
+      return (buffer[index] >> bit_offset) & counter_mask;
+    }
+
+    void Inc(uptr i) const {
+      DCHECK_LT(Get(i), counter_mask);
+      uptr index = i >> packing_ratio_log;
+      uptr bit_offset = (i & bit_offset_mask) << counter_size_bits_log;
+      buffer[index] += 1ULL << bit_offset;
+    }
+
+    void IncRange(uptr from, uptr to) const {
+      DCHECK_LE(from, to);
+      for (uptr i = from; i <= to; i++)
+        Inc(i);
+    }
+
+   private:
+    const u64 n;
+    u64 counter_size_bits_log;
+    u64 counter_mask;
+    u64 packing_ratio_log;
+    u64 bit_offset_mask;
+
+    MemoryMapperT* const memory_mapper;
+    u64 buffer_size;
+    u64* buffer;
+  };
+
+  template<class MemoryMapperT>
+  class FreePagesRangeTracker {
+   public:
+    explicit FreePagesRangeTracker(MemoryMapperT* mapper)
+        : memory_mapper(mapper),
+          page_size_scaled_log(Log2(GetPageSizeCached() >> kCompactPtrScale)),
+          in_the_range(false), current_page(0), current_range_start_page(0) {}
+
+    void NextPage(bool freed) {
+      if (freed) {
+        if (!in_the_range) {
+          current_range_start_page = current_page;
+          in_the_range = true;
+        }
+      } else {
+        CloseOpenedRange();
+      }
+      current_page++;
+    }
+
+    void Done() {
+      CloseOpenedRange();
+    }
+
+   private:
+    void CloseOpenedRange() {
+      if (in_the_range) {
+        memory_mapper->ReleasePageRangeToOS(
+            current_range_start_page << page_size_scaled_log,
+            current_page << page_size_scaled_log);
+        in_the_range = false;
+      }
+    }
+
+    MemoryMapperT* const memory_mapper;
+    const uptr page_size_scaled_log;
+    bool in_the_range;
+    uptr current_page;
+    uptr current_range_start_page;
+  };
+
+  // Iterates over the free_array to identify memory pages containing freed
+  // chunks only and returns these pages back to OS.
+  // allocated_pages_count is the total number of pages allocated for the
+  // current bucket.
+  template<class MemoryMapperT>
+  static void ReleaseFreeMemoryToOS(CompactPtrT *free_array,
+                                    uptr free_array_count, uptr chunk_size,
+                                    uptr allocated_pages_count,
+                                    MemoryMapperT *memory_mapper) {
+    const uptr page_size = GetPageSizeCached();
+
+    // Figure out the number of chunks per page and whether we can take a fast
+    // path (the number of chunks per page is the same for all pages).
+    uptr full_pages_chunk_count_max;
+    bool same_chunk_count_per_page;
+    if (chunk_size <= page_size && page_size % chunk_size == 0) {
+      // Same number of chunks per page, no cross overs.
+      full_pages_chunk_count_max = page_size / chunk_size;
+      same_chunk_count_per_page = true;
+    } else if (chunk_size <= page_size && page_size % chunk_size != 0 &&
+        chunk_size % (page_size % chunk_size) == 0) {
+      // Some chunks are crossing page boundaries, which means that the page
+      // contains one or two partial chunks, but all pages contain the same
+      // number of chunks.
+      full_pages_chunk_count_max = page_size / chunk_size + 1;
+      same_chunk_count_per_page = true;
+    } else if (chunk_size <= page_size) {
+      // Some chunks are crossing page boundaries, which means that the page
+      // contains one or two partial chunks.
+      full_pages_chunk_count_max = page_size / chunk_size + 2;
+      same_chunk_count_per_page = false;
+    } else if (chunk_size > page_size && chunk_size % page_size == 0) {
+      // One chunk covers multiple pages, no cross overs.
+      full_pages_chunk_count_max = 1;
+      same_chunk_count_per_page = true;
+    } else if (chunk_size > page_size) {
+      // One chunk covers multiple pages, Some chunks are crossing page
+      // boundaries. Some pages contain one chunk, some contain two.
+      full_pages_chunk_count_max = 2;
+      same_chunk_count_per_page = false;
+    } else {
+      UNREACHABLE("All chunk_size/page_size ratios must be handled.");
+    }
+
+    PackedCounterArray<MemoryMapperT> counters(allocated_pages_count,
+                                               full_pages_chunk_count_max,
+                                               memory_mapper);
+    if (!counters.IsAllocated())
+      return;
+
+    const uptr chunk_size_scaled = chunk_size >> kCompactPtrScale;
+    const uptr page_size_scaled = page_size >> kCompactPtrScale;
+    const uptr page_size_scaled_log = Log2(page_size_scaled);
+
+    // Iterate over free chunks and count how many free chunks affect each
+    // allocated page.
+    if (chunk_size <= page_size && page_size % chunk_size == 0) {
+      // Each chunk affects one page only.
+      for (uptr i = 0; i < free_array_count; i++)
+        counters.Inc(free_array[i] >> page_size_scaled_log);
+    } else {
+      // In all other cases chunks might affect more than one page.
+      for (uptr i = 0; i < free_array_count; i++) {
+        counters.IncRange(
+            free_array[i] >> page_size_scaled_log,
+            (free_array[i] + chunk_size_scaled - 1) >> page_size_scaled_log);
+      }
+    }
+
+    // Iterate over pages detecting ranges of pages with chunk counters equal
+    // to the expected number of chunks for the particular page.
+    FreePagesRangeTracker<MemoryMapperT> range_tracker(memory_mapper);
+    if (same_chunk_count_per_page) {
+      // Fast path, every page has the same number of chunks affecting it.
+      for (uptr i = 0; i < counters.GetCount(); i++)
+        range_tracker.NextPage(counters.Get(i) == full_pages_chunk_count_max);
+    } else {
+      // Show path, go through the pages keeping count how many chunks affect
+      // each page.
+      const uptr pn =
+          chunk_size < page_size ? page_size_scaled / chunk_size_scaled : 1;
+      const uptr pnc = pn * chunk_size_scaled;
+      // The idea is to increment the current page pointer by the first chunk
+      // size, middle portion size (the portion of the page covered by chunks
+      // except the first and the last one) and then the last chunk size, adding
+      // up the number of chunks on the current page and checking on every step
+      // whether the page boundary was crossed.
+      uptr prev_page_boundary = 0;
+      uptr current_boundary = 0;
+      for (uptr i = 0; i < counters.GetCount(); i++) {
+        uptr page_boundary = prev_page_boundary + page_size_scaled;
+        uptr chunks_per_page = pn;
+        if (current_boundary < page_boundary) {
+          if (current_boundary > prev_page_boundary)
+            chunks_per_page++;
+          current_boundary += pnc;
+          if (current_boundary < page_boundary) {
+            chunks_per_page++;
+            current_boundary += chunk_size_scaled;
+          }
+        }
+        prev_page_boundary = page_boundary;
+
+        range_tracker.NextPage(counters.Get(i) == chunks_per_page);
+      }
+    }
+    range_tracker.Done();
+  }
+
  private:
+  friend class MemoryMapper;
+
+  ReservedAddressRange address_range;
+  static const char *AllocatorName() { return "sanitizer_allocator"; }
+
   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
@@ -330,6 +581,7 @@ class SizeClassAllocator64 {
     uptr n_freed_at_last_release;
     uptr num_releases;
     u64 last_release_at_ns;
+    u64 last_released_bytes;
   };
 
   struct RegionInfo {
@@ -347,30 +599,18 @@ class SizeClassAllocator64 {
   };
   COMPILER_CHECK(sizeof(RegionInfo) >= kCacheLineSize);
 
-  u32 Rand(u32 *state) {  // ANSI C linear congruential PRNG.
-    return (*state = *state * 1103515245 + 12345) >> 16;
-  }
-
-  u32 RandN(u32 *state, u32 n) { return Rand(state) % n; }  // [0, n)
-
-  void RandomShuffle(u32 *a, u32 n, u32 *rand_state) {
-    if (n <= 1) return;
-    for (u32 i = n - 1; i > 0; i--)
-      Swap(a[i], a[RandN(rand_state, i + 1)]);
-  }
-
-  RegionInfo *GetRegionInfo(uptr class_id) {
+  RegionInfo *GetRegionInfo(uptr class_id) const {
     CHECK_LT(class_id, kNumClasses);
     RegionInfo *regions =
         reinterpret_cast<RegionInfo *>(SpaceBeg() + kSpaceSize);
     return &regions[class_id];
   }
 
-  uptr GetMetadataEnd(uptr region_beg) {
+  uptr GetMetadataEnd(uptr region_beg) const {
     return region_beg + kRegionSize - kFreeArraySize;
   }
 
-  uptr GetChunkIdx(uptr chunk, uptr size) {
+  uptr GetChunkIdx(uptr chunk, uptr size) const {
     if (!kUsingConstantSpaceBeg)
       chunk -= SpaceBeg();
 
@@ -382,13 +622,12 @@ class SizeClassAllocator64 {
     return (u32)offset / (u32)size;
   }
 
-  CompactPtrT *GetFreeArray(uptr region_beg) {
-    return reinterpret_cast<CompactPtrT *>(region_beg + kRegionSize -
-                                           kFreeArraySize);
+  CompactPtrT *GetFreeArray(uptr region_beg) const {
+    return reinterpret_cast<CompactPtrT *>(GetMetadataEnd(region_beg));
   }
 
   bool MapWithCallback(uptr beg, uptr size) {
-    uptr mapped = reinterpret_cast<uptr>(MmapFixedOrDieOnFatalError(beg, size));
+    uptr mapped = address_range.Map(beg, size);
     if (UNLIKELY(!mapped))
       return false;
     CHECK_EQ(beg, mapped);
@@ -397,21 +636,21 @@ class SizeClassAllocator64 {
   }
 
   void MapWithCallbackOrDie(uptr beg, uptr size) {
-    CHECK_EQ(beg, reinterpret_cast<uptr>(MmapFixedOrDie(beg, size)));
+    CHECK_EQ(beg, address_range.MapOrDie(beg, size));
     MapUnmapCallback().OnMap(beg, size);
   }
 
   void UnmapWithCallbackOrDie(uptr beg, uptr size) {
     MapUnmapCallback().OnUnmap(beg, size);
-    UnmapOrDie(reinterpret_cast<void *>(beg), size);
+    address_range.Unmap(beg, size);
   }
 
   bool EnsureFreeArraySpace(RegionInfo *region, uptr region_beg,
                             uptr num_freed_chunks) {
     uptr needed_space = num_freed_chunks * sizeof(CompactPtrT);
     if (region->mapped_free_array < needed_space) {
-      CHECK_LE(needed_space, kFreeArraySize);
       uptr new_mapped_free_array = RoundUpTo(needed_space, kFreeArrayMapSize);
+      CHECK_LE(new_mapped_free_array, kFreeArraySize);
       uptr current_map_end = reinterpret_cast<uptr>(GetFreeArray(region_beg)) +
                              region->mapped_free_array;
       uptr new_map_size = new_mapped_free_array - region->mapped_free_array;
@@ -422,55 +661,74 @@ class SizeClassAllocator64 {
     return true;
   }
 
+  // Check whether this size class is exhausted.
+  bool IsRegionExhausted(RegionInfo *region, uptr class_id,
+                         uptr additional_map_size) {
+    if (LIKELY(region->mapped_user + region->mapped_meta +
+               additional_map_size <= kRegionSize - kFreeArraySize))
+      return false;
+    if (!region->exhausted) {
+      region->exhausted = true;
+      Printf("%s: Out of memory. ", SanitizerToolName);
+      Printf("The process has exhausted %zuMB for size class %zu.\n",
+             kRegionSize >> 20, ClassIdToSize(class_id));
+    }
+    return true;
+  }
+
   NOINLINE bool PopulateFreeArray(AllocatorStats *stat, uptr class_id,
                                   RegionInfo *region, uptr requested_count) {
     // region->mutex is held.
-    const uptr size = ClassIdToSize(class_id);
-    const uptr new_space_beg = region->allocated_user;
-    const uptr new_space_end = new_space_beg + requested_count * size;
     const uptr region_beg = GetRegionBeginBySizeClass(class_id);
+    const uptr size = ClassIdToSize(class_id);
 
+    const uptr total_user_bytes =
+        region->allocated_user + requested_count * size;
     // Map more space for chunks, if necessary.
-    if (new_space_end > region->mapped_user) {
-      if (!kUsingConstantSpaceBeg && region->mapped_user == 0)
-        region->rand_state = static_cast<u32>(region_beg >> 12);  // From ASLR.
+    if (LIKELY(total_user_bytes > region->mapped_user)) {
+      if (UNLIKELY(region->mapped_user == 0)) {
+        if (!kUsingConstantSpaceBeg && kRandomShuffleChunks)
+          // The random state is initialized from ASLR.
+          region->rand_state = static_cast<u32>(region_beg >> 12);
+        // Postpone the first release to OS attempt for ReleaseToOSIntervalMs,
+        // preventing just allocated memory from being released sooner than
+        // necessary and also preventing extraneous ReleaseMemoryPagesToOS calls
+        // for short lived processes.
+        // Do it only when the feature is turned on, to avoid a potentially
+        // extraneous syscall.
+        if (ReleaseToOSIntervalMs() >= 0)
+          region->rtoi.last_release_at_ns = MonotonicNanoTime();
+      }
       // Do the mmap for the user memory.
-      uptr map_size = kUserMapSize;
-      while (new_space_end > region->mapped_user + map_size)
-        map_size += kUserMapSize;
-      CHECK_GE(region->mapped_user + map_size, new_space_end);
+      const uptr user_map_size =
+          RoundUpTo(total_user_bytes - region->mapped_user, kUserMapSize);
+      if (UNLIKELY(IsRegionExhausted(region, class_id, user_map_size)))
+        return false;
       if (UNLIKELY(!MapWithCallback(region_beg + region->mapped_user,
-                                    map_size)))
+                                    user_map_size)))
         return false;
-      stat->Add(AllocatorStatMapped, map_size);
-      region->mapped_user += map_size;
+      stat->Add(AllocatorStatMapped, user_map_size);
+      region->mapped_user += user_map_size;
     }
-    const uptr new_chunks_count = (region->mapped_user - new_space_beg) / size;
+    const uptr new_chunks_count =
+        (region->mapped_user - region->allocated_user) / size;
 
-    // Calculate the required space for metadata.
-    const uptr requested_allocated_meta =
-        region->allocated_meta + new_chunks_count * kMetadataSize;
-    uptr requested_mapped_meta = region->mapped_meta;
-    while (requested_allocated_meta > requested_mapped_meta)
-      requested_mapped_meta += kMetaMapSize;
-    // Check whether this size class is exhausted.
-    if (region->mapped_user + requested_mapped_meta >
-        kRegionSize - kFreeArraySize) {
-      if (!region->exhausted) {
-        region->exhausted = true;
-        Printf("%s: Out of memory. ", SanitizerToolName);
-        Printf("The process has exhausted %zuMB for size class %zu.\n",
-               kRegionSize >> 20, size);
+    if (kMetadataSize) {
+      // Calculate the required space for metadata.
+      const uptr total_meta_bytes =
+          region->allocated_meta + new_chunks_count * kMetadataSize;
+      const uptr meta_map_size = (total_meta_bytes > region->mapped_meta) ?
+          RoundUpTo(total_meta_bytes - region->mapped_meta, kMetaMapSize) : 0;
+      // Map more space for metadata, if necessary.
+      if (meta_map_size) {
+        if (UNLIKELY(IsRegionExhausted(region, class_id, meta_map_size)))
+          return false;
+        if (UNLIKELY(!MapWithCallback(
+            GetMetadataEnd(region_beg) - region->mapped_meta - meta_map_size,
+            meta_map_size)))
+          return false;
+        region->mapped_meta += meta_map_size;
       }
-      return false;
-    }
-    // Map more space for metadata, if necessary.
-    if (requested_mapped_meta > region->mapped_meta) {
-      if (UNLIKELY(!MapWithCallback(
-              GetMetadataEnd(region_beg) - requested_mapped_meta,
-              requested_mapped_meta - region->mapped_meta)))
-        return false;
-      region->mapped_meta = requested_mapped_meta;
     }
 
     // If necessary, allocate more space for the free array and populate it with
@@ -479,7 +737,7 @@ class SizeClassAllocator64 {
     if (UNLIKELY(!EnsureFreeArraySpace(region, region_beg, total_freed_chunks)))
       return false;
     CompactPtrT *free_array = GetFreeArray(region_beg);
-    for (uptr i = 0, chunk = new_space_beg; i < new_chunks_count;
+    for (uptr i = 0, chunk = region->allocated_user; i < new_chunks_count;
          i++, chunk += size)
       free_array[total_freed_chunks - 1 - i] = PointerToCompactPtr(0, chunk);
     if (kRandomShuffleChunks)
@@ -491,27 +749,66 @@ class SizeClassAllocator64 {
     region->num_freed_chunks += new_chunks_count;
     region->allocated_user += new_chunks_count * size;
     CHECK_LE(region->allocated_user, region->mapped_user);
-    region->allocated_meta = requested_allocated_meta;
+    region->allocated_meta += new_chunks_count * kMetadataSize;
     CHECK_LE(region->allocated_meta, region->mapped_meta);
     region->exhausted = false;
 
+    // TODO(alekseyshl): Consider bumping last_release_at_ns here to prevent
+    // MaybeReleaseToOS from releasing just allocated pages or protect these
+    // not yet used chunks some other way.
+
     return true;
   }
 
-  void MaybeReleaseChunkRange(uptr region_beg, uptr chunk_size,
-                              CompactPtrT first, CompactPtrT last) {
-    uptr beg_ptr = CompactPtrToPointer(region_beg, first);
-    uptr end_ptr = CompactPtrToPointer(region_beg, last) + chunk_size;
-    ReleaseMemoryPagesToOS(beg_ptr, end_ptr);
-  }
+  class MemoryMapper {
+   public:
+    MemoryMapper(const ThisT& base_allocator, uptr class_id)
+        : allocator(base_allocator),
+          region_base(base_allocator.GetRegionBeginBySizeClass(class_id)),
+          released_ranges_count(0),
+          released_bytes(0) {
+    }
+
+    uptr GetReleasedRangesCount() const {
+      return released_ranges_count;
+    }
+
+    uptr GetReleasedBytes() const {
+      return released_bytes;
+    }
+
+    uptr MapPackedCounterArrayBuffer(uptr buffer_size) {
+      // TODO(alekseyshl): The idea to explore is to check if we have enough
+      // space between num_freed_chunks*sizeof(CompactPtrT) and
+      // mapped_free_array to fit buffer_size bytes and use that space instead
+      // of mapping a temporary one.
+      return reinterpret_cast<uptr>(
+          MmapOrDieOnFatalError(buffer_size, "ReleaseToOSPageCounters"));
+    }
 
-  // Attempts to release some RAM back to OS. The region is expected to be
-  // locked.
-  // Algorithm:
-  // * Sort the chunks.
-  // * Find ranges fully covered by free-d chunks
-  // * Release them to OS with madvise.
-  void MaybeReleaseToOS(uptr class_id) {
+    void UnmapPackedCounterArrayBuffer(uptr buffer, uptr buffer_size) {
+      UnmapOrDie(reinterpret_cast<void *>(buffer), buffer_size);
+    }
+
+    // Releases [from, to) range of pages back to OS.
+    void ReleasePageRangeToOS(CompactPtrT from, CompactPtrT to) {
+      const uptr from_page = allocator.CompactPtrToPointer(region_base, from);
+      const uptr to_page = allocator.CompactPtrToPointer(region_base, to);
+      ReleaseMemoryPagesToOS(from_page, to_page);
+      released_ranges_count++;
+      released_bytes += to_page - from_page;
+    }
+
+   private:
+    const ThisT& allocator;
+    const uptr region_base;
+    uptr released_ranges_count;
+    uptr released_bytes;
+  };
+
+  // Attempts to release RAM occupied by freed chunks back to OS. The region is
+  // expected to be locked.
+  void MaybeReleaseToOS(uptr class_id, bool force) {
     RegionInfo *region = GetRegionInfo(class_id);
     const uptr chunk_size = ClassIdToSize(class_id);
     const uptr page_size = GetPageSizeCached();
@@ -524,37 +821,29 @@ class SizeClassAllocator64 {
       return;  // Nothing new to release.
     }
 
-    s32 interval_ms = ReleaseToOSIntervalMs();
-    if (interval_ms < 0)
-      return;
+    if (!force) {
+      s32 interval_ms = ReleaseToOSIntervalMs();
+      if (interval_ms < 0)
+        return;
 
-    u64 now_ns = NanoTime();
-    if (region->rtoi.last_release_at_ns + interval_ms * 1000000ULL > now_ns)
-      return;  // Memory was returned recently.
-    region->rtoi.last_release_at_ns = now_ns;
+      if (region->rtoi.last_release_at_ns + interval_ms * 1000000ULL >
+          MonotonicNanoTime()) {
+        return;  // Memory was returned recently.
+      }
+    }
 
-    uptr region_beg = GetRegionBeginBySizeClass(class_id);
-    CompactPtrT *free_array = GetFreeArray(region_beg);
-    SortArray(free_array, n);
+    MemoryMapper memory_mapper(*this, class_id);
 
-    const uptr scaled_chunk_size = chunk_size >> kCompactPtrScale;
-    const uptr kScaledGranularity = page_size >> kCompactPtrScale;
+    ReleaseFreeMemoryToOS<MemoryMapper>(
+        GetFreeArray(GetRegionBeginBySizeClass(class_id)), n, chunk_size,
+        RoundUpTo(region->allocated_user, page_size) / page_size,
+        &memory_mapper);
 
-    uptr range_beg = free_array[0];
-    uptr prev = free_array[0];
-    for (uptr i = 1; i < n; i++) {
-      uptr chunk = free_array[i];
-      CHECK_GT(chunk, prev);
-      if (chunk - prev != scaled_chunk_size) {
-        CHECK_GT(chunk - prev, scaled_chunk_size);
-        if (prev + scaled_chunk_size - range_beg >= kScaledGranularity) {
-          MaybeReleaseChunkRange(region_beg, chunk_size, range_beg, prev);
-          region->rtoi.n_freed_at_last_release = region->stats.n_freed;
-          region->rtoi.num_releases++;
-        }
-        range_beg = chunk;
-      }
-      prev = chunk;
+    if (memory_mapper.GetReleasedRangesCount() > 0) {
+      region->rtoi.n_freed_at_last_release = region->stats.n_freed;
+      region->rtoi.num_releases += memory_mapper.GetReleasedRangesCount();
+      region->rtoi.last_released_bytes = memory_mapper.GetReleasedBytes();
     }
+    region->rtoi.last_release_at_ns = MonotonicNanoTime();
   }
 };
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_size_class_map.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_size_class_map.h
index 7151a4636056..2bd83b2eb5a4 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_size_class_map.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_allocator_size_class_map.h
@@ -134,8 +134,9 @@ class SizeClassMap {
 
   static const uptr kMaxSize = 1UL << kMaxSizeLog;
   static const uptr kNumClasses =
-      kMidClass + ((kMaxSizeLog - kMidSizeLog) << S) + 1;
+      kMidClass + ((kMaxSizeLog - kMidSizeLog) << S) + 1 + 1;
   static const uptr kLargestClassID = kNumClasses - 2;
+  static const uptr kBatchClassID = kNumClasses - 1;
   COMPILER_CHECK(kNumClasses >= 16 && kNumClasses <= 256);
   static const uptr kNumClassesRounded =
       kNumClasses <= 32  ? 32 :
@@ -143,6 +144,11 @@ class SizeClassMap {
       kNumClasses <= 128 ? 128 : 256;
 
   static uptr Size(uptr class_id) {
+    // Estimate the result for kBatchClassID because this class does not know
+    // the exact size of TransferBatch. It's OK since we are using the actual
+    // sizeof(TransferBatch) where it matters.
+    if (UNLIKELY(class_id == kBatchClassID))
+      return kMaxNumCachedHint * sizeof(uptr);
     if (class_id <= kMidClass)
       return kMinSize * class_id;
     class_id -= kMidClass;
@@ -151,9 +157,10 @@ class SizeClassMap {
   }
 
   static uptr ClassID(uptr size) {
+    if (UNLIKELY(size > kMaxSize))
+      return 0;
     if (size <= kMidSize)
       return (size + kMinSize - 1) >> kMinSizeLog;
-    if (size > kMaxSize) return 0;
     uptr l = MostSignificantSetBitIndex(size);
     uptr hbits = (size >> (l - S)) & M;
     uptr lbits = size & ((1 << (l - S)) - 1);
@@ -162,7 +169,13 @@ class SizeClassMap {
   }
 
   static uptr MaxCachedHint(uptr class_id) {
-    if (class_id == 0) return 0;
+    // Estimate the result for kBatchClassID because this class does not know
+    // the exact size of TransferBatch. We need to cache fewer batches than user
+    // chunks, so this number can be small.
+    if (UNLIKELY(class_id == kBatchClassID))
+      return 16;
+    if (UNLIKELY(class_id == 0))
+      return 0;
     uptr n = (1UL << kMaxBytesCachedLog) / Size(class_id);
     return Max<uptr>(1, Min(kMaxNumCachedHint, n));
   }
@@ -178,6 +191,8 @@ class SizeClassMap {
       uptr p = prev_s ? (d * 100 / prev_s) : 0;
       uptr l = s ? MostSignificantSetBitIndex(s) : 0;
       uptr cached = MaxCachedHint(i) * 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(i), cached, ClassID(s));
@@ -192,12 +207,13 @@ class SizeClassMap {
       // Printf("Validate: c%zd\n", c);
       uptr s = Size(c);
       CHECK_NE(s, 0U);
+      if (c == kBatchClassID)
+        continue;
       CHECK_EQ(ClassID(s), c);
-      if (c != kNumClasses - 1)
+      if (c < kLargestClassID)
         CHECK_EQ(ClassID(s + 1), c + 1);
       CHECK_EQ(ClassID(s - 1), c);
-      if (c)
-        CHECK_GT(Size(c), Size(c-1));
+      CHECK_GT(Size(c), Size(c - 1));
     }
     CHECK_EQ(ClassID(kMaxSize + 1), 0);
 
@@ -207,7 +223,7 @@ class SizeClassMap {
       CHECK_LT(c, kNumClasses);
       CHECK_GE(Size(c), s);
       if (c > 0)
-        CHECK_LT(Size(c-1), s);
+        CHECK_LT(Size(c - 1), s);
     }
   }
 };
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_asm.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_asm.h
index 47c2b12a2049..76a0bf71425d 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_asm.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_asm.h
@@ -47,12 +47,12 @@
 # define ASM_HIDDEN(symbol) .hidden symbol
 # define ASM_TYPE_FUNCTION(symbol) .type symbol, @function
 # define ASM_SIZE(symbol) .size symbol, .-symbol
-# define ASM_TSAN_SYMBOL(symbol) symbol
-# define ASM_TSAN_SYMBOL_INTERCEPTOR(symbol) symbol
+# define ASM_SYMBOL(symbol) symbol
+# define ASM_SYMBOL_INTERCEPTOR(symbol) symbol
 #else
 # define ASM_HIDDEN(symbol)
 # define ASM_TYPE_FUNCTION(symbol)
 # define ASM_SIZE(symbol)
-# define ASM_TSAN_SYMBOL(symbol) _##symbol
-# define ASM_TSAN_SYMBOL_INTERCEPTOR(symbol) _wrap_##symbol
+# define ASM_SYMBOL(symbol) _##symbol
+# define ASM_SYMBOL_INTERCEPTOR(symbol) _wrap_##symbol
 #endif
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_bitvector.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_bitvector.h
index d8472732ff21..0f65814c398f 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_bitvector.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_bitvector.h
@@ -22,7 +22,7 @@ namespace __sanitizer {
 template <class basic_int_t = uptr>
 class BasicBitVector {
  public:
-  enum SizeEnum { kSize = sizeof(basic_int_t) * 8 };
+  enum SizeEnum : uptr { kSize = sizeof(basic_int_t) * 8 };
 
   uptr size() const { return kSize; }
   // No CTOR.
@@ -115,7 +115,7 @@ class TwoLevelBitVector {
   // This structure allows O(kLevel1Size) time for clear() and empty(),
   // as well fast handling of sparse BVs.
  public:
-  enum SizeEnum { kSize = BV::kSize * BV::kSize * kLevel1Size };
+  enum SizeEnum : uptr { kSize = BV::kSize * BV::kSize * kLevel1Size };
   // No CTOR.
 
   uptr size() const { return kSize; }
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_bvgraph.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_bvgraph.h
index df72f1c2d471..9c2db9ae4954 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_bvgraph.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_bvgraph.h
@@ -25,7 +25,7 @@ namespace __sanitizer {
 template<class BV>
 class BVGraph {
  public:
-  enum SizeEnum { kSize = BV::kSize };
+  enum SizeEnum : uptr { kSize = BV::kSize };
   uptr size() const { return kSize; }
   // No CTOR.
   void clear() {
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common.cc
index 7e6f8fce76d2..ae26d5efc24f 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common.cc
@@ -26,72 +26,7 @@ const char *SanitizerToolName = "SanitizerTool";
 
 atomic_uint32_t current_verbosity;
 uptr PageSizeCached;
-
-StaticSpinMutex report_file_mu;
-ReportFile report_file = {&report_file_mu, kStderrFd, "", "", 0};
-
-void RawWrite(const char *buffer) {
-  report_file.Write(buffer, internal_strlen(buffer));
-}
-
-void ReportFile::ReopenIfNecessary() {
-  mu->CheckLocked();
-  if (fd == kStdoutFd || fd == kStderrFd) return;
-
-  uptr pid = internal_getpid();
-  // If in tracer, use the parent's file.
-  if (pid == stoptheworld_tracer_pid)
-    pid = stoptheworld_tracer_ppid;
-  if (fd != kInvalidFd) {
-    // If the report file is already opened by the current process,
-    // do nothing. Otherwise the report file was opened by the parent
-    // process, close it now.
-    if (fd_pid == pid)
-      return;
-    else
-      CloseFile(fd);
-  }
-
-  const char *exe_name = GetProcessName();
-  if (common_flags()->log_exe_name && exe_name) {
-    internal_snprintf(full_path, kMaxPathLength, "%s.%s.%zu", path_prefix,
-                      exe_name, pid);
-  } else {
-    internal_snprintf(full_path, kMaxPathLength, "%s.%zu", path_prefix, pid);
-  }
-  fd = OpenFile(full_path, WrOnly);
-  if (fd == kInvalidFd) {
-    const char *ErrorMsgPrefix = "ERROR: Can't open file: ";
-    WriteToFile(kStderrFd, ErrorMsgPrefix, internal_strlen(ErrorMsgPrefix));
-    WriteToFile(kStderrFd, full_path, internal_strlen(full_path));
-    Die();
-  }
-  fd_pid = pid;
-}
-
-void ReportFile::SetReportPath(const char *path) {
-  if (!path)
-    return;
-  uptr len = internal_strlen(path);
-  if (len > sizeof(path_prefix) - 100) {
-    Report("ERROR: Path is too long: %c%c%c%c%c%c%c%c...\n",
-           path[0], path[1], path[2], path[3],
-           path[4], path[5], path[6], path[7]);
-    Die();
-  }
-
-  SpinMutexLock l(mu);
-  if (fd != kStdoutFd && fd != kStderrFd && fd != kInvalidFd)
-    CloseFile(fd);
-  fd = kInvalidFd;
-  if (internal_strcmp(path, "stdout") == 0) {
-    fd = kStdoutFd;
-  } else if (internal_strcmp(path, "stderr") == 0) {
-    fd = kStderrFd;
-  } else {
-    internal_snprintf(path_prefix, kMaxPathLength, "%s", path);
-  }
-}
+u32 NumberOfCPUsCached;
 
 // PID of the tracer task in StopTheWorld. It shares the address space with the
 // main process, but has a different PID and thus requires special handling.
@@ -120,42 +55,6 @@ void NORETURN ReportMmapFailureAndDie(uptr size, const char *mem_type,
   UNREACHABLE("unable to mmap");
 }
 
-bool ReadFileToBuffer(const char *file_name, char **buff, uptr *buff_size,
-                      uptr *read_len, uptr max_len, error_t *errno_p) {
-  uptr PageSize = GetPageSizeCached();
-  uptr kMinFileLen = PageSize;
-  *buff = nullptr;
-  *buff_size = 0;
-  *read_len = 0;
-  // The files we usually open are not seekable, so try different buffer sizes.
-  for (uptr size = kMinFileLen; size <= max_len; size *= 2) {
-    fd_t fd = OpenFile(file_name, RdOnly, errno_p);
-    if (fd == kInvalidFd) return false;
-    UnmapOrDie(*buff, *buff_size);
-    *buff = (char*)MmapOrDie(size, __func__);
-    *buff_size = size;
-    *read_len = 0;
-    // Read up to one page at a time.
-    bool reached_eof = false;
-    while (*read_len + PageSize <= size) {
-      uptr just_read;
-      if (!ReadFromFile(fd, *buff + *read_len, PageSize, &just_read, errno_p)) {
-        UnmapOrDie(*buff, *buff_size);
-        return false;
-      }
-      if (just_read == 0) {
-        reached_eof = true;
-        break;
-      }
-      *read_len += just_read;
-    }
-    CloseFile(fd);
-    if (reached_eof)  // We've read the whole file.
-      break;
-  }
-  return true;
-}
-
 typedef bool UptrComparisonFunction(const uptr &a, const uptr &b);
 typedef bool U32ComparisonFunction(const u32 &a, const u32 &b);
 
@@ -285,9 +184,10 @@ void LoadedModule::clear() {
 }
 
 void LoadedModule::addAddressRange(uptr beg, uptr end, bool executable,
-                                   bool writable) {
+                                   bool writable, const char *name) {
   void *mem = InternalAlloc(sizeof(AddressRange));
-  AddressRange *r = new(mem) AddressRange(beg, end, executable, writable);
+  AddressRange *r =
+      new(mem) AddressRange(beg, end, executable, writable, name);
   ranges_.push_back(r);
   if (executable && end > max_executable_address_)
     max_executable_address_ = end;
@@ -359,36 +259,6 @@ bool TemplateMatch(const char *templ, const char *str) {
   return true;
 }
 
-static const char kPathSeparator = SANITIZER_WINDOWS ? ';' : ':';
-
-char *FindPathToBinary(const char *name) {
-  if (FileExists(name)) {
-    return internal_strdup(name);
-  }
-
-  const char *path = GetEnv("PATH");
-  if (!path)
-    return nullptr;
-  uptr name_len = internal_strlen(name);
-  InternalScopedBuffer<char> buffer(kMaxPathLength);
-  const char *beg = path;
-  while (true) {
-    const char *end = internal_strchrnul(beg, kPathSeparator);
-    uptr prefix_len = end - beg;
-    if (prefix_len + name_len + 2 <= kMaxPathLength) {
-      internal_memcpy(buffer.data(), beg, prefix_len);
-      buffer[prefix_len] = '/';
-      internal_memcpy(&buffer[prefix_len + 1], name, name_len);
-      buffer[prefix_len + 1 + name_len] = '\0';
-      if (FileExists(buffer.data()))
-        return internal_strdup(buffer.data());
-    }
-    if (*end == '\0') break;
-    beg = end + 1;
-  }
-  return nullptr;
-}
-
 static char binary_name_cache_str[kMaxPathLength];
 static char process_name_cache_str[kMaxPathLength];
 
@@ -482,15 +352,6 @@ static int InstallMallocFreeHooks(void (*malloc_hook)(const void *, uptr),
 using namespace __sanitizer;  // NOLINT
 
 extern "C" {
-void __sanitizer_set_report_path(const char *path) {
-  report_file.SetReportPath(path);
-}
-
-void __sanitizer_set_report_fd(void *fd) {
-  report_file.fd = (fd_t)reinterpret_cast<uptr>(fd);
-  report_file.fd_pid = internal_getpid();
-}
-
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_report_error_summary,
                              const char *error_summary) {
   Printf("%s\n", error_summary);
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common.h
index 89aae579856a..1fbaee7e39a1 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common.h
@@ -29,8 +29,11 @@ extern "C" void _ReadWriteBarrier();
 #endif
 
 namespace __sanitizer {
-struct StackTrace;
+
 struct AddressInfo;
+struct BufferedStackTrace;
+struct SignalContext;
+struct StackTrace;
 
 // Constants.
 const uptr kWordSize = SANITIZER_WORDSIZE / 8;
@@ -71,6 +74,7 @@ INLINE uptr GetPageSizeCached() {
 }
 uptr GetMmapGranularity();
 uptr GetMaxVirtualAddress();
+uptr GetMaxUserVirtualAddress();
 // Threads
 tid_t GetTid();
 uptr GetThreadSelf();
@@ -125,6 +129,30 @@ void CheckVMASize();
 void RunMallocHooks(const void *ptr, uptr size);
 void RunFreeHooks(const void *ptr);
 
+class ReservedAddressRange {
+ public:
+  uptr Init(uptr size, const char *name = nullptr, uptr fixed_addr = 0);
+  uptr Map(uptr fixed_addr, uptr size);
+  uptr MapOrDie(uptr fixed_addr, uptr size);
+  void Unmap(uptr addr, uptr size);
+  void *base() const { return base_; }
+  uptr size() const { return size_; }
+
+ private:
+  void* base_;
+  uptr size_;
+  const char* name_;
+  uptr os_handle_;
+};
+
+typedef void (*fill_profile_f)(uptr start, uptr rss, bool file,
+                               /*out*/uptr *stats, uptr stats_size);
+
+// 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);
+
 // InternalScopedBuffer can be used instead of large stack arrays to
 // keep frame size low.
 // FIXME: use InternalAlloc instead of MmapOrDie once
@@ -179,12 +207,15 @@ class LowLevelAllocator {
   char *allocated_end_;
   char *allocated_current_;
 };
+// Set the min alignment of LowLevelAllocator to at least alignment.
+void SetLowLevelAllocateMinAlignment(uptr alignment);
 typedef void (*LowLevelAllocateCallback)(uptr ptr, uptr size);
 // Allows to register tool-specific callbacks for LowLevelAllocator.
 // Passing NULL removes the callback.
 void SetLowLevelAllocateCallback(LowLevelAllocateCallback callback);
 
 // IO
+void CatastrophicErrorWrite(const char *buffer, uptr length);
 void RawWrite(const char *buffer);
 bool ColorizeReports();
 void RemoveANSIEscapeSequencesFromString(char *buffer);
@@ -200,67 +231,18 @@ void SetPrintfAndReportCallback(void (*callback)(const char *));
     if ((uptr)Verbosity() >= (level)) Printf(__VA_ARGS__); \
   } while (0)
 
-// Can be used to prevent mixing error reports from different sanitizers.
-extern StaticSpinMutex CommonSanitizerReportMutex;
-
-struct ReportFile {
-  void Write(const char *buffer, uptr length);
-  bool SupportsColors();
-  void SetReportPath(const char *path);
-
-  // Don't use fields directly. They are only declared public to allow
-  // aggregate initialization.
-
-  // Protects fields below.
-  StaticSpinMutex *mu;
-  // Opened file descriptor. Defaults to stderr. It may be equal to
-  // kInvalidFd, in which case new file will be opened when necessary.
-  fd_t fd;
-  // Path prefix of report file, set via __sanitizer_set_report_path.
-  char path_prefix[kMaxPathLength];
-  // Full path to report, obtained as <path_prefix>.PID
-  char full_path[kMaxPathLength];
-  // PID of the process that opened fd. If a fork() occurs,
-  // the PID of child will be different from fd_pid.
-  uptr fd_pid;
+// Lock sanitizer error reporting and protects against nested errors.
+class ScopedErrorReportLock {
+ public:
+  ScopedErrorReportLock();
+  ~ScopedErrorReportLock();
 
- private:
-  void ReopenIfNecessary();
+  static void CheckLocked();
 };
-extern ReportFile report_file;
 
 extern uptr stoptheworld_tracer_pid;
 extern uptr stoptheworld_tracer_ppid;
 
-enum FileAccessMode {
-  RdOnly,
-  WrOnly,
-  RdWr
-};
-
-// Returns kInvalidFd on error.
-fd_t OpenFile(const char *filename, FileAccessMode mode,
-              error_t *errno_p = nullptr);
-void CloseFile(fd_t);
-
-// Return true on success, false on error.
-bool ReadFromFile(fd_t fd, void *buff, uptr buff_size,
-                  uptr *bytes_read = nullptr, error_t *error_p = nullptr);
-bool WriteToFile(fd_t fd, const void *buff, uptr buff_size,
-                 uptr *bytes_written = nullptr, error_t *error_p = nullptr);
-
-bool RenameFile(const char *oldpath, const char *newpath,
-                error_t *error_p = nullptr);
-
-// Scoped file handle closer.
-struct FileCloser {
-  explicit FileCloser(fd_t fd) : fd(fd) {}
-  ~FileCloser() { CloseFile(fd); }
-  fd_t fd;
-};
-
-bool SupportsColoredOutput(fd_t fd);
-
 // Opens the file 'file_name" and reads up to 'max_len' bytes.
 // The resulting buffer is mmaped and stored in '*buff'.
 // The size of the mmaped region is stored in '*buff_size'.
@@ -269,11 +251,6 @@ bool SupportsColoredOutput(fd_t fd);
 bool ReadFileToBuffer(const char *file_name, char **buff, uptr *buff_size,
                       uptr *read_len, uptr max_len = 1 << 26,
                       error_t *errno_p = nullptr);
-// Maps given file to virtual memory, and returns pointer to it
-// (or NULL if mapping fails). Stores the size of mmaped region
-// in '*buff_size'.
-void *MapFileToMemory(const char *file_name, uptr *buff_size);
-void *MapWritableFileToMemory(void *addr, uptr size, fd_t fd, OFF_T offset);
 
 bool IsAccessibleMemoryRange(uptr beg, uptr size);
 
@@ -293,27 +270,8 @@ void CacheBinaryName();
 void DisableCoreDumperIfNecessary();
 void DumpProcessMap();
 void PrintModuleMap();
-bool FileExists(const char *filename);
 const char *GetEnv(const char *name);
 bool SetEnv(const char *name, const char *value);
-const char *GetPwd();
-char *FindPathToBinary(const char *name);
-bool IsPathSeparator(const char c);
-bool IsAbsolutePath(const char *path);
-// 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
-// in parent process even in case of an error.
-// The child process will close all fds after STDERR_FILENO
-// before passing control to a program.
-pid_t StartSubprocess(const char *filename, const char *const argv[],
-                      fd_t stdin_fd = kInvalidFd, fd_t stdout_fd = kInvalidFd,
-                      fd_t stderr_fd = kInvalidFd);
-// Checks if specified process is still running
-bool IsProcessRunning(pid_t pid);
-// Waits for the process to finish and returns its exit code.
-// Returns -1 in case of an error.
-int WaitForProcess(pid_t pid);
 
 u32 GetUid();
 void ReExec();
@@ -337,6 +295,7 @@ uptr GetTlsSize();
 void SleepForSeconds(int seconds);
 void SleepForMillis(int millis);
 u64 NanoTime();
+u64 MonotonicNanoTime();
 int Atexit(void (*function)(void));
 void SortArray(uptr *array, uptr size);
 void SortArray(u32 *array, uptr size);
@@ -385,7 +344,24 @@ void SetSoftRssLimitExceededCallback(void (*Callback)(bool exceeded));
 typedef void (*SignalHandlerType)(int, void *, void *);
 HandleSignalMode GetHandleSignalMode(int signum);
 void InstallDeadlySignalHandlers(SignalHandlerType handler);
-const char *DescribeSignalOrException(int signo);
+
+// Signal reporting.
+// Each sanitizer uses slightly different implementation of stack unwinding.
+typedef void (*UnwindSignalStackCallbackType)(const SignalContext &sig,
+                                              const void *callback_context,
+                                              BufferedStackTrace *stack);
+// Print deadly signal report and die.
+void HandleDeadlySignal(void *siginfo, void *context, u32 tid,
+                        UnwindSignalStackCallbackType unwind,
+                        const void *unwind_context);
+
+// Part of HandleDeadlySignal, exposed for asan.
+void StartReportDeadlySignal();
+// Part of HandleDeadlySignal, exposed for asan.
+void ReportDeadlySignal(const SignalContext &sig, u32 tid,
+                        UnwindSignalStackCallbackType unwind,
+                        const void *unwind_context);
+
 // Alternative signal stack (POSIX-only).
 void SetAlternateSignalStack();
 void UnsetAlternateSignalStack();
@@ -702,6 +678,7 @@ inline const char *ModuleArchToString(ModuleArch arch) {
 }
 
 const uptr kModuleUUIDSize = 16;
+const uptr kMaxSegName = 16;
 
 // Represents a binary loaded into virtual memory (e.g. this can be an
 // executable or a shared object).
@@ -720,7 +697,8 @@ class LoadedModule {
   void set(const char *module_name, uptr base_address, ModuleArch arch,
            u8 uuid[kModuleUUIDSize], bool instrumented);
   void clear();
-  void addAddressRange(uptr beg, uptr end, bool executable, bool writable);
+  void addAddressRange(uptr beg, uptr end, bool executable, bool writable,
+                       const char *name = nullptr);
   bool containsAddress(uptr address) const;
 
   const char *full_name() const { return full_name_; }
@@ -736,13 +714,17 @@ class LoadedModule {
     uptr end;
     bool executable;
     bool writable;
+    char name[kMaxSegName];
 
-    AddressRange(uptr beg, uptr end, bool executable, bool writable)
+    AddressRange(uptr beg, uptr end, bool executable, bool writable,
+                 const char *name)
         : next(nullptr),
           beg(beg),
           end(end),
           executable(executable),
-          writable(writable) {}
+          writable(writable) {
+      internal_strncpy(this->name, (name ? name : ""), ARRAY_SIZE(this->name));
+    }
   };
 
   const IntrusiveList<AddressRange> &ranges() const { return ranges_; }
@@ -761,9 +743,10 @@ class LoadedModule {
 // filling this information.
 class ListOfModules {
  public:
-  ListOfModules() : modules_(kInitialCapacity) {}
+  ListOfModules() : initialized(false) {}
   ~ListOfModules() { clear(); }
   void init();
+  void fallbackInit();  // Uses fallback init if available, otherwise clears
   const LoadedModule *begin() const { return modules_.begin(); }
   LoadedModule *begin() { return modules_.begin(); }
   const LoadedModule *end() const { return modules_.end(); }
@@ -779,10 +762,15 @@ class ListOfModules {
     for (auto &module : modules_) module.clear();
     modules_.clear();
   }
+  void clearOrInit() {
+    initialized ? clear() : modules_.Initialize(kInitialCapacity);
+    initialized = true;
+  }
 
-  InternalMmapVector<LoadedModule> modules_;
+  InternalMmapVectorNoCtor<LoadedModule> modules_;
   // We rarely have more than 16K loaded modules.
   static const uptr kInitialCapacity = 1 << 14;
+  bool initialized;
 };
 
 // Callback type for iterating over a set of memory ranges.
@@ -858,35 +846,49 @@ static inline void SanitizerBreakOptimization(void *arg) {
 }
 
 struct SignalContext {
+  void *siginfo;
   void *context;
   uptr addr;
   uptr pc;
   uptr sp;
   uptr bp;
   bool is_memory_access;
-
   enum WriteFlag { UNKNOWN, READ, WRITE } write_flag;
 
-  SignalContext(void *context, uptr addr, uptr pc, uptr sp, uptr bp,
-                bool is_memory_access, WriteFlag write_flag)
-      : context(context),
-        addr(addr),
-        pc(pc),
-        sp(sp),
-        bp(bp),
-        is_memory_access(is_memory_access),
-        write_flag(write_flag) {}
+  // VS2013 doesn't implement unrestricted unions, so we need a trivial default
+  // constructor
+  SignalContext() = default;
+
+  // Creates signal context in a platform-specific manner.
+  // SignalContext is going to keep pointers to siginfo and context without
+  // owning them.
+  SignalContext(void *siginfo, void *context)
+      : siginfo(siginfo),
+        context(context),
+        addr(GetAddress()),
+        is_memory_access(IsMemoryAccess()),
+        write_flag(GetWriteFlag()) {
+    InitPcSpBp();
+  }
 
   static void DumpAllRegisters(void *context);
 
-  // Creates signal context in a platform-specific manner.
-  static SignalContext Create(void *siginfo, void *context);
+  // Type of signal e.g. SIGSEGV or EXCEPTION_ACCESS_VIOLATION.
+  int GetType() const;
 
-  // Returns true if the "context" indicates a memory write.
-  static WriteFlag GetWriteFlag(void *context);
-};
+  // String description of the signal.
+  const char *Describe() const;
 
-void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp);
+  // Returns true if signal is stack overflow.
+  bool IsStackOverflow() const;
+
+ private:
+  // Platform specific initialization.
+  void InitPcSpBp();
+  uptr GetAddress() const;
+  WriteFlag GetWriteFlag() const;
+  bool IsMemoryAccess() const;
+};
 
 void MaybeReexec();
 
@@ -929,8 +931,17 @@ const s32 kReleaseToOSIntervalNever = -1;
 void CheckNoDeepBind(const char *filename, int flag);
 
 // Returns the requested amount of random data (up to 256 bytes) that can then
-// be used to seed a PRNG.
-bool GetRandom(void *buffer, uptr length);
+// be used to seed a PRNG. Defaults to blocking like the underlying syscall.
+bool GetRandom(void *buffer, uptr length, bool blocking = true);
+
+// Returns the number of logical processors on the system.
+u32 GetNumberOfCPUs();
+extern u32 NumberOfCPUsCached;
+INLINE u32 GetNumberOfCPUsCached() {
+  if (!NumberOfCPUsCached)
+    NumberOfCPUsCached = GetNumberOfCPUs();
+  return NumberOfCPUsCached;
+}
 
 }  // namespace __sanitizer
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc
index 8607bf44902d..592a8c7b032a 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc
@@ -43,6 +43,7 @@
 #include "sanitizer_errno.h"
 #include "sanitizer_placement_new.h"
 #include "sanitizer_platform_interceptors.h"
+#include "sanitizer_symbolizer.h"
 #include "sanitizer_tls_get_addr.h"
 
 #include <stdarg.h>
@@ -68,6 +69,46 @@
 #define iconv __bsd_iconv
 #endif
 
+#if SANITIZER_NETBSD
+#define clock_getres __clock_getres50
+#define clock_gettime __clock_gettime50
+#define clock_settime __clock_settime50
+#define ctime __ctime50
+#define ctime_r __ctime_r50
+#define getitimer __getitimer50
+#define getpwent __getpwent50
+#define getpwnam __getpwnam50
+#define getpwnam_r __getpwnam_r50
+#define getpwuid __getpwuid50
+#define getpwuid_r __getpwuid_r50
+#define getutent __getutent50
+#define getutxent __getutxent50
+#define getutxid __getutxid50
+#define getutxline __getutxline50
+#define glob __glob30
+#define gmtime __gmtime50
+#define gmtime_r __gmtime_r50
+#define localtime_r __localtime_r50
+#define mktime __mktime50
+#define opendir __opendir30
+#define readdir __readdir30
+#define readdir_r __readdir_r30
+#define scandir __scandir30
+#define setitimer __setitimer50
+#define setlocale __setlocale50
+#define shmctl __shmctl50
+#define sigemptyset __sigemptyset14
+#define sigfillset __sigfillset14
+#define sigpending __sigpending14
+#define sigprocmask __sigprocmask14
+#define sigtimedwait __sigtimedwait50
+#define stat __stat50
+#define time __time50
+#define times __times13
+#define wait3 __wait350
+#define wait4 __wait450
+#endif
+
 // Platform-specific options.
 #if SANITIZER_MAC
 namespace __sanitizer {
@@ -258,7 +299,7 @@ typedef AddrHashMap<CommonInterceptorMetadata, 31051> MetadataHashMap;
 
 static MetadataHashMap *interceptor_metadata_map;
 
-#if SI_NOT_WINDOWS
+#if SI_POSIX
 UNUSED static void SetInterceptorMetadata(__sanitizer_FILE *addr,
                                           const FileMetadata &file) {
   MetadataHashMap::Handle h(interceptor_metadata_map, (uptr)addr);
@@ -272,7 +313,7 @@ UNUSED static const FileMetadata *GetInterceptorMetadata(
   MetadataHashMap::Handle h(interceptor_metadata_map, (uptr)addr,
                             /* remove */ false,
                             /* create */ false);
-  if (h.exists()) {
+  if (addr && h.exists()) {
     CHECK(!h.created());
     CHECK(h->type == CommonInterceptorMetadata::CIMT_FILE);
     return &h->file;
@@ -285,7 +326,7 @@ UNUSED static void DeleteInterceptorMetadata(void *addr) {
   MetadataHashMap::Handle h(interceptor_metadata_map, (uptr)addr, true);
   CHECK(h.exists());
 }
-#endif  // SI_NOT_WINDOWS
+#endif  // SI_POSIX
 
 #if SANITIZER_INTERCEPT_STRLEN
 INTERCEPTOR(SIZE_T, strlen, const char *s) {
@@ -886,7 +927,7 @@ INTERCEPTOR(long double, frexpl, long double x, int *exp) {
 #define INIT_FREXPF_FREXPL
 #endif  // SANITIZER_INTERCEPT_FREXPF_FREXPL
 
-#if SI_NOT_WINDOWS
+#if SI_POSIX
 static void write_iovec(void *ctx, struct __sanitizer_iovec *iovec,
                         SIZE_T iovlen, SIZE_T maxlen) {
   for (SIZE_T i = 0; i < iovlen && maxlen; ++i) {
@@ -1176,12 +1217,14 @@ INTERCEPTOR(unsigned long, time, unsigned long *t) {
 #if SANITIZER_INTERCEPT_LOCALTIME_AND_FRIENDS
 static void unpoison_tm(void *ctx, __sanitizer_tm *tm) {
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, tm, sizeof(*tm));
+#if !SANITIZER_SOLARIS
   if (tm->tm_zone) {
     // 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);
   }
+#endif
 }
 INTERCEPTOR(__sanitizer_tm *, localtime, unsigned long *timep) {
   void *ctx;
@@ -1509,6 +1552,12 @@ INTERCEPTOR(int, vsnprintf, char *str, SIZE_T size, const char *format,
             va_list ap)
 VSNPRINTF_INTERCEPTOR_IMPL(vsnprintf, str, size, format, ap)
 
+#if SANITIZER_INTERCEPT___PRINTF_CHK
+INTERCEPTOR(int, __vsnprintf_chk, char *str, SIZE_T size, int flag,
+            SIZE_T size_to, const char *format, va_list ap)
+VSNPRINTF_INTERCEPTOR_IMPL(vsnprintf, str, size, format, ap)
+#endif
+
 #if SANITIZER_INTERCEPT_PRINTF_L
 INTERCEPTOR(int, vsnprintf_l, char *str, SIZE_T size, void *loc,
             const char *format, va_list ap)
@@ -1522,6 +1571,12 @@ FORMAT_INTERCEPTOR_IMPL(snprintf_l, vsnprintf_l, str, size, loc, format)
 INTERCEPTOR(int, vsprintf, char *str, const char *format, va_list ap)
 VSPRINTF_INTERCEPTOR_IMPL(vsprintf, str, format, ap)
 
+#if SANITIZER_INTERCEPT___PRINTF_CHK
+INTERCEPTOR(int, __vsprintf_chk, char *str, int flag, SIZE_T size_to,
+            const char *format, va_list ap)
+VSPRINTF_INTERCEPTOR_IMPL(vsprintf, str, format, ap)
+#endif
+
 INTERCEPTOR(int, vasprintf, char **strp, const char *format, va_list ap)
 VASPRINTF_INTERCEPTOR_IMPL(vasprintf, strp, format, ap)
 
@@ -1550,12 +1605,30 @@ FORMAT_INTERCEPTOR_IMPL(printf, vprintf, format)
 INTERCEPTOR(int, fprintf, __sanitizer_FILE *stream, const char *format, ...)
 FORMAT_INTERCEPTOR_IMPL(fprintf, vfprintf, stream, format)
 
+#if SANITIZER_INTERCEPT___PRINTF_CHK
+INTERCEPTOR(int, __fprintf_chk, __sanitizer_FILE *stream, SIZE_T size,
+            const char *format, ...)
+FORMAT_INTERCEPTOR_IMPL(__fprintf_chk, vfprintf, stream, format)
+#endif
+
 INTERCEPTOR(int, sprintf, char *str, const char *format, ...) // NOLINT
 FORMAT_INTERCEPTOR_IMPL(sprintf, vsprintf, str, format) // NOLINT
 
+#if SANITIZER_INTERCEPT___PRINTF_CHK
+INTERCEPTOR(int, __sprintf_chk, char *str, int flag, SIZE_T size_to,
+            const char *format, ...) // NOLINT
+FORMAT_INTERCEPTOR_IMPL(__sprintf_chk, vsprintf, str, format) // NOLINT
+#endif
+
 INTERCEPTOR(int, snprintf, char *str, SIZE_T size, const char *format, ...)
 FORMAT_INTERCEPTOR_IMPL(snprintf, vsnprintf, str, size, format)
 
+#if SANITIZER_INTERCEPT___PRINTF_CHK
+INTERCEPTOR(int, __snprintf_chk, char *str, SIZE_T size, int flag,
+            SIZE_T size_to, const char *format, ...) // NOLINT
+FORMAT_INTERCEPTOR_IMPL(__snprintf_chk, vsnprintf, str, size, format) // NOLINT
+#endif
+
 INTERCEPTOR(int, asprintf, char **strp, const char *format, ...)
 FORMAT_INTERCEPTOR_IMPL(asprintf, vasprintf, strp, format)
 
@@ -1595,6 +1668,17 @@ FORMAT_INTERCEPTOR_IMPL(__isoc99_snprintf, __isoc99_vsnprintf, str, size,
 #define INIT_PRINTF
 #endif
 
+#if SANITIZER_INTERCEPT___PRINTF_CHK
+#define INIT___PRINTF_CHK                     \
+  COMMON_INTERCEPT_FUNCTION(__sprintf_chk);   \
+  COMMON_INTERCEPT_FUNCTION(__snprintf_chk);  \
+  COMMON_INTERCEPT_FUNCTION(__vsprintf_chk);  \
+  COMMON_INTERCEPT_FUNCTION(__vsnprintf_chk); \
+  COMMON_INTERCEPT_FUNCTION(__fprintf_chk);
+#else
+#define INIT___PRINTF_CHK
+#endif
+
 #if SANITIZER_INTERCEPT_PRINTF_L
 #define INIT_PRINTF_L                     \
   COMMON_INTERCEPT_FUNCTION(snprintf_l);  \
@@ -1723,7 +1807,8 @@ static void unpoison_group(void *ctx, __sanitizer_group *grp) {
 INTERCEPTOR(__sanitizer_passwd *, getpwnam, const char *name) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, getpwnam, name);
-  COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
+  if (name)
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, name, REAL(strlen)(name) + 1);
   __sanitizer_passwd *res = REAL(getpwnam)(name);
   if (res) unpoison_passwd(ctx, res);
   return res;
@@ -2003,6 +2088,13 @@ INTERCEPTOR(int, clock_gettime, u32 clk_id, void *tp) {
   }
   return res;
 }
+namespace __sanitizer {
+extern "C" {
+int real_clock_gettime(u32 clk_id, void *tp) {
+  return REAL(clock_gettime)(clk_id, tp);
+}
+}  // extern "C"
+}  // namespace __sanitizer
 INTERCEPTOR(int, clock_settime, u32 clk_id, const void *tp) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, clock_settime, clk_id, tp);
@@ -2064,6 +2156,18 @@ static void unpoison_glob_t(void *ctx, __sanitizer_glob_t *pglob) {
   }
 }
 
+#if SANITIZER_SOLARIS
+INTERCEPTOR(int, glob, const char *pattern, int flags,
+            int (*errfunc)(const char *epath, int eerrno),
+            __sanitizer_glob_t *pglob) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, glob, pattern, flags, errfunc, pglob);
+  COMMON_INTERCEPTOR_READ_STRING(ctx, pattern, 0);
+  int res = REAL(glob)(pattern, flags, errfunc, pglob);
+  if ((!res || res == glob_nomatch) && pglob) unpoison_glob_t(ctx, pglob);
+  return res;
+}
+#else
 static THREADLOCAL __sanitizer_glob_t *pglob_copy;
 
 static void wrapped_gl_closedir(void *dir) {
@@ -2127,7 +2231,14 @@ INTERCEPTOR(int, glob, const char *pattern, int flags,
   if ((!res || res == glob_nomatch) && pglob) unpoison_glob_t(ctx, pglob);
   return res;
 }
+#endif  // SANITIZER_SOLARIS
+#define INIT_GLOB                  \
+  COMMON_INTERCEPT_FUNCTION(glob);
+#else  // SANITIZER_INTERCEPT_GLOB
+#define INIT_GLOB
+#endif  // SANITIZER_INTERCEPT_GLOB
 
+#if SANITIZER_INTERCEPT_GLOB64
 INTERCEPTOR(int, glob64, const char *pattern, int flags,
             int (*errfunc)(const char *epath, int eerrno),
             __sanitizer_glob_t *pglob) {
@@ -2156,12 +2267,11 @@ INTERCEPTOR(int, glob64, const char *pattern, int flags,
   if ((!res || res == glob_nomatch) && pglob) unpoison_glob_t(ctx, pglob);
   return res;
 }
-#define INIT_GLOB                  \
-  COMMON_INTERCEPT_FUNCTION(glob); \
+#define INIT_GLOB64                \
   COMMON_INTERCEPT_FUNCTION(glob64);
-#else  // SANITIZER_INTERCEPT_GLOB
-#define INIT_GLOB
-#endif  // SANITIZER_INTERCEPT_GLOB
+#else  // SANITIZER_INTERCEPT_GLOB64
+#define INIT_GLOB64
+#endif  // SANITIZER_INTERCEPT_GLOB64
 
 #if SANITIZER_INTERCEPT_WAIT
 // According to sys/wait.h, wait(), waitid(), waitpid() may have symbol version
@@ -2466,7 +2576,15 @@ INTERCEPTOR(struct __sanitizer_hostent *, gethostent, int fake) {
   if (res) write_hostent(ctx, res);
   return res;
 }
+#define INIT_GETHOSTBYNAME                  \
+  COMMON_INTERCEPT_FUNCTION(gethostent);    \
+  COMMON_INTERCEPT_FUNCTION(gethostbyaddr); \
+  COMMON_INTERCEPT_FUNCTION(gethostbyname);
+#else
+#define INIT_GETHOSTBYNAME
+#endif  // SANITIZER_INTERCEPT_GETHOSTBYNAME
 
+#if SANITIZER_INTERCEPT_GETHOSTBYNAME2
 INTERCEPTOR(struct __sanitizer_hostent *, gethostbyname2, char *name, int af) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, gethostbyname2, name, af);
@@ -2474,14 +2592,10 @@ INTERCEPTOR(struct __sanitizer_hostent *, gethostbyname2, char *name, int af) {
   if (res) write_hostent(ctx, res);
   return res;
 }
-#define INIT_GETHOSTBYNAME                  \
-  COMMON_INTERCEPT_FUNCTION(gethostent);    \
-  COMMON_INTERCEPT_FUNCTION(gethostbyaddr); \
-  COMMON_INTERCEPT_FUNCTION(gethostbyname); \
-  COMMON_INTERCEPT_FUNCTION(gethostbyname2);
+#define INIT_GETHOSTBYNAME2 COMMON_INTERCEPT_FUNCTION(gethostbyname2);
 #else
-#define INIT_GETHOSTBYNAME
-#endif
+#define INIT_GETHOSTBYNAME2
+#endif  // SANITIZER_INTERCEPT_GETHOSTBYNAME2
 
 #if SANITIZER_INTERCEPT_GETHOSTBYNAME_R
 INTERCEPTOR(int, gethostbyname_r, char *name, struct __sanitizer_hostent *ret,
@@ -3371,7 +3485,7 @@ INTERCEPTOR(char *, strerror, int errnum) {
 //  * GNU version returns message pointer, which points to either buf or some
 //    static storage.
 #if ((_POSIX_C_SOURCE >= 200112L || _XOPEN_SOURCE >= 600) && !_GNU_SOURCE) || \
-    SANITIZER_MAC || SANITIZER_ANDROID
+    SANITIZER_MAC || SANITIZER_ANDROID || SANITIZER_NETBSD
 // POSIX version. Spec is not clear on whether buf is NULL-terminated.
 // At least on OSX, buf contents are valid even when the call fails.
 INTERCEPTOR(int, strerror_r, int errnum, char *buf, SIZE_T buflen) {
@@ -3589,7 +3703,7 @@ INTERCEPTOR(int, ppoll, __sanitizer_pollfd *fds, __sanitizer_nfds_t nfds,
   if (fds && nfds) read_pollfd(ctx, fds, nfds);
   if (timeout_ts)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, timeout_ts, struct_timespec_sz);
-  // FIXME: read sigmask when all of sigemptyset, etc are intercepted.
+  if (sigmask) COMMON_INTERCEPTOR_READ_RANGE(ctx, sigmask, sizeof(*sigmask));
   int res =
       COMMON_INTERCEPTOR_BLOCK_REAL(ppoll)(fds, nfds, timeout_ts, sigmask);
   if (fds && nfds) write_pollfd(ctx, fds, nfds);
@@ -3630,7 +3744,7 @@ INTERCEPTOR(int, wordexp, char *s, __sanitizer_wordexp_t *p, int flags) {
 INTERCEPTOR(int, sigwait, __sanitizer_sigset_t *set, int *sig) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, sigwait, set, sig);
-  // FIXME: read sigset_t when all of sigemptyset, etc are intercepted
+  if (set) COMMON_INTERCEPTOR_READ_RANGE(ctx, set, sizeof(*set));
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -3647,7 +3761,7 @@ INTERCEPTOR(int, sigwait, __sanitizer_sigset_t *set, int *sig) {
 INTERCEPTOR(int, sigwaitinfo, __sanitizer_sigset_t *set, void *info) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, sigwaitinfo, set, info);
-  // FIXME: read sigset_t when all of sigemptyset, etc are intercepted
+  if (set) COMMON_INTERCEPTOR_READ_RANGE(ctx, set, sizeof(*set));
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -3666,7 +3780,7 @@ INTERCEPTOR(int, sigtimedwait, __sanitizer_sigset_t *set, void *info,
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, sigtimedwait, set, info, timeout);
   if (timeout) COMMON_INTERCEPTOR_READ_RANGE(ctx, timeout, struct_timespec_sz);
-  // FIXME: read sigset_t when all of sigemptyset, etc are intercepted
+  if (set) COMMON_INTERCEPTOR_READ_RANGE(ctx, set, sizeof(*set));
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -3729,7 +3843,7 @@ INTERCEPTOR(int, sigprocmask, int how, __sanitizer_sigset_t *set,
             __sanitizer_sigset_t *oldset) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, sigprocmask, how, set, oldset);
-  // FIXME: read sigset_t when all of sigemptyset, etc are intercepted
+  if (set) COMMON_INTERCEPTOR_READ_RANGE(ctx, set, sizeof(*set));
   // FIXME: under ASan the call below may write to freed memory and corrupt
   // its metadata. See
   // https://github.com/google/sanitizers/issues/321.
@@ -3827,6 +3941,15 @@ INTERCEPTOR(int, pthread_mutex_unlock, void *m) {
 #define INIT_PTHREAD_MUTEX_UNLOCK
 #endif
 
+#if SANITIZER_NETBSD
+INTERCEPTOR(int, __libc_mutex_lock, void *m) \
+  ALIAS(WRAPPER_NAME(pthread_mutex_lock));
+INTERCEPTOR(int, __libc_mutex_unlock, void *m) \
+  ALIAS(WRAPPER_NAME(pthread_mutex_unlock));
+INTERCEPTOR(int, __libc_thr_setcancelstate, int state, int *oldstate) \
+  ALIAS(WRAPPER_NAME(pthread_setcancelstate));
+#endif
+
 #if SANITIZER_INTERCEPT_GETMNTENT || SANITIZER_INTERCEPT_GETMNTENT_R
 static void write_mntent(void *ctx, __sanitizer_mntent *mnt) {
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, mnt, sizeof(*mnt));
@@ -4404,7 +4527,7 @@ INTERCEPTOR(char *, tempnam, char *dir, char *pfx) {
 #define INIT_TEMPNAM
 #endif
 
-#if SANITIZER_INTERCEPT_PTHREAD_SETNAME_NP
+#if SANITIZER_INTERCEPT_PTHREAD_SETNAME_NP && !SANITIZER_NETBSD
 INTERCEPTOR(int, pthread_setname_np, uptr thread, const char *name) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, pthread_setname_np, thread, name);
@@ -4413,6 +4536,17 @@ INTERCEPTOR(int, pthread_setname_np, uptr thread, const char *name) {
   return REAL(pthread_setname_np)(thread, name);
 }
 #define INIT_PTHREAD_SETNAME_NP COMMON_INTERCEPT_FUNCTION(pthread_setname_np);
+#elif SANITIZER_INTERCEPT_PTHREAD_SETNAME_NP && SANITIZER_NETBSD
+INTERCEPTOR(int, pthread_setname_np, uptr thread, const char *name, void *arg) {
+  void *ctx;
+  char newname[32]; // PTHREAD_MAX_NAMELEN_NP=32
+  COMMON_INTERCEPTOR_ENTER(ctx, pthread_setname_np, thread, name, arg);
+  COMMON_INTERCEPTOR_READ_STRING(ctx, name, 0);
+  internal_snprintf(newname, sizeof(newname), name, arg);
+  COMMON_INTERCEPTOR_SET_PTHREAD_NAME(ctx, thread, newname);
+  return REAL(pthread_setname_np)(thread, name, arg);
+}
+#define INIT_PTHREAD_SETNAME_NP COMMON_INTERCEPT_FUNCTION(pthread_setname_np);
 #else
 #define INIT_PTHREAD_SETNAME_NP
 #endif
@@ -5575,6 +5709,7 @@ INTERCEPTOR(void*, dlopen, const char *filename, int flag) {
   if (filename) COMMON_INTERCEPTOR_READ_STRING(ctx, filename, 0);
   COMMON_INTERCEPTOR_ON_DLOPEN(filename, flag);
   void *res = REAL(dlopen)(filename, flag);
+  Symbolizer::GetOrInit()->InvalidateModuleList();
   COMMON_INTERCEPTOR_LIBRARY_LOADED(filename, res);
   return res;
 }
@@ -5583,6 +5718,7 @@ INTERCEPTOR(int, dlclose, void *handle) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER_NOIGNORE(ctx, dlclose, handle);
   int res = REAL(dlclose)(handle);
+  Symbolizer::GetOrInit()->InvalidateModuleList();
   COMMON_INTERCEPTOR_LIBRARY_UNLOADED();
   return res;
 }
@@ -5821,7 +5957,7 @@ INTERCEPTOR(int, pthread_setcancelstate, int state, int *oldstate) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, pthread_setcancelstate, state, oldstate);
   int res = REAL(pthread_setcancelstate)(state, oldstate);
-  if (res == 0)
+  if (res == 0 && oldstate != nullptr)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, oldstate, sizeof(*oldstate));
   return res;
 }
@@ -5830,7 +5966,7 @@ INTERCEPTOR(int, pthread_setcanceltype, int type, int *oldtype) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, pthread_setcanceltype, type, oldtype);
   int res = REAL(pthread_setcanceltype)(type, oldtype);
-  if (res == 0)
+  if (res == 0 && oldtype != nullptr)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, oldtype, sizeof(*oldtype));
   return res;
 }
@@ -6313,6 +6449,7 @@ static void InitializeCommonInterceptors() {
   INIT_GETITIMER;
   INIT_TIME;
   INIT_GLOB;
+  INIT_GLOB64;
   INIT_WAIT;
   INIT_WAIT4;
   INIT_INET;
@@ -6321,6 +6458,7 @@ static void InitializeCommonInterceptors() {
   INIT_GETNAMEINFO;
   INIT_GETSOCKNAME;
   INIT_GETHOSTBYNAME;
+  INIT_GETHOSTBYNAME2;
   INIT_GETHOSTBYNAME_R;
   INIT_GETHOSTBYNAME2_R;
   INIT_GETHOSTBYADDR_R;
@@ -6457,4 +6595,12 @@ static void InitializeCommonInterceptors() {
   INIT_GETLOADAVG;
   INIT_WCSLEN;
   INIT_WCSCAT;
+
+#if SANITIZER_NETBSD
+  COMMON_INTERCEPT_FUNCTION(__libc_mutex_lock);
+  COMMON_INTERCEPT_FUNCTION(__libc_mutex_unlock);
+  COMMON_INTERCEPT_FUNCTION(__libc_thr_setcancelstate);
+#endif
+
+  INIT___PRINTF_CHK;
 }
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_ioctl.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_ioctl.inc
index 4ed9afedf84a..24e7548a597e 100755
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_ioctl.inc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_ioctl.inc
@@ -57,7 +57,9 @@ static void ioctl_table_fill() {
   _(SIOCGIFCONF, CUSTOM, 0);
   _(SIOCGPGRP, WRITE, sizeof(int));
   _(SIOCSPGRP, READ, sizeof(int));
+#if !SANITIZER_SOLARIS
   _(TIOCCONS, NONE, 0);
+#endif
   _(TIOCEXCL, NONE, 0);
   _(TIOCGETD, WRITE, sizeof(int));
   _(TIOCGPGRP, WRITE, pid_t_sz);
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interface.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interface.inc
index 550427c906a6..2568df3a5bff 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interface.inc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_interface.inc
@@ -34,6 +34,7 @@ INTERFACE_FUNCTION(__sanitizer_get_heap_size)
 INTERFACE_FUNCTION(__sanitizer_get_ownership)
 INTERFACE_FUNCTION(__sanitizer_get_unmapped_bytes)
 INTERFACE_FUNCTION(__sanitizer_install_malloc_and_free_hooks)
+INTERFACE_FUNCTION(__sanitizer_purge_allocator)
 INTERFACE_FUNCTION(__sanitizer_print_memory_profile)
 INTERFACE_WEAK_FUNCTION(__sanitizer_free_hook)
 INTERFACE_WEAK_FUNCTION(__sanitizer_malloc_hook)
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_libcdep.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_libcdep.cc
index cf200512de84..5cdfbbb275f7 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_libcdep.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_common_libcdep.cc
@@ -14,7 +14,10 @@
 #include "sanitizer_common.h"
 
 #include "sanitizer_allocator_interface.h"
+#include "sanitizer_file.h"
 #include "sanitizer_flags.h"
+#include "sanitizer_procmaps.h"
+#include "sanitizer_report_decorator.h"
 #include "sanitizer_stackdepot.h"
 #include "sanitizer_stacktrace.h"
 #include "sanitizer_symbolizer.h"
@@ -25,12 +28,25 @@
 
 namespace __sanitizer {
 
+#if !SANITIZER_FUCHSIA
+
 bool ReportFile::SupportsColors() {
   SpinMutexLock l(mu);
   ReopenIfNecessary();
   return SupportsColoredOutput(fd);
 }
 
+static INLINE bool ReportSupportsColors() {
+  return report_file.SupportsColors();
+}
+
+#else  // SANITIZER_FUCHSIA
+
+// Fuchsia's logs always go through post-processing that handles colorization.
+static INLINE bool ReportSupportsColors() { return true; }
+
+#endif  // !SANITIZER_FUCHSIA
+
 bool ColorizeReports() {
   // FIXME: Add proper Windows support to AnsiColorDecorator and re-enable color
   // printing on Windows.
@@ -39,7 +55,7 @@ bool ColorizeReports() {
 
   const char *flag = common_flags()->color;
   return internal_strcmp(flag, "always") == 0 ||
-         (internal_strcmp(flag, "auto") == 0 && report_file.SupportsColors());
+         (internal_strcmp(flag, "auto") == 0 && ReportSupportsColors());
 }
 
 static void (*sandboxing_callback)();
@@ -131,6 +147,127 @@ void BackgroundThread(void *arg) {
 }
 #endif
 
+#if !SANITIZER_FUCHSIA && !SANITIZER_GO
+void StartReportDeadlySignal() {
+  // Write the first message using fd=2, just in case.
+  // It may actually fail to write in case stderr is closed.
+  CatastrophicErrorWrite(SanitizerToolName, internal_strlen(SanitizerToolName));
+  static const char kDeadlySignal[] = ":DEADLYSIGNAL\n";
+  CatastrophicErrorWrite(kDeadlySignal, sizeof(kDeadlySignal) - 1);
+}
+
+static void MaybeReportNonExecRegion(uptr pc) {
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD
+  MemoryMappingLayout proc_maps(/*cache_enabled*/ true);
+  MemoryMappedSegment segment;
+  while (proc_maps.Next(&segment)) {
+    if (pc >= segment.start && pc < segment.end && !segment.IsExecutable())
+      Report("Hint: PC is at a non-executable region. Maybe a wild jump?\n");
+  }
+#endif
+}
+
+static void PrintMemoryByte(InternalScopedString *str, const char *before,
+                            u8 byte) {
+  SanitizerCommonDecorator d;
+  str->append("%s%s%x%x%s ", before, d.MemoryByte(), byte >> 4, byte & 15,
+              d.Default());
+}
+
+static void MaybeDumpInstructionBytes(uptr pc) {
+  if (!common_flags()->dump_instruction_bytes || (pc < GetPageSizeCached()))
+    return;
+  InternalScopedString str(1024);
+  str.append("First 16 instruction bytes at pc: ");
+  if (IsAccessibleMemoryRange(pc, 16)) {
+    for (int i = 0; i < 16; ++i) {
+      PrintMemoryByte(&str, "", ((u8 *)pc)[i]);
+    }
+    str.append("\n");
+  } else {
+    str.append("unaccessible\n");
+  }
+  Report("%s", str.data());
+}
+
+static void MaybeDumpRegisters(void *context) {
+  if (!common_flags()->dump_registers) return;
+  SignalContext::DumpAllRegisters(context);
+}
+
+static void ReportStackOverflowImpl(const SignalContext &sig, u32 tid,
+                                    UnwindSignalStackCallbackType unwind,
+                                    const void *unwind_context) {
+  SanitizerCommonDecorator d;
+  Printf("%s", d.Warning());
+  static const char kDescription[] = "stack-overflow";
+  Report("ERROR: %s: %s on address %p (pc %p bp %p sp %p T%d)\n",
+         SanitizerToolName, kDescription, (void *)sig.addr, (void *)sig.pc,
+         (void *)sig.bp, (void *)sig.sp, tid);
+  Printf("%s", d.Default());
+  InternalScopedBuffer<BufferedStackTrace> stack_buffer(1);
+  BufferedStackTrace *stack = stack_buffer.data();
+  stack->Reset();
+  unwind(sig, unwind_context, stack);
+  stack->Print();
+  ReportErrorSummary(kDescription, stack);
+}
+
+static void ReportDeadlySignalImpl(const SignalContext &sig, u32 tid,
+                                   UnwindSignalStackCallbackType unwind,
+                                   const void *unwind_context) {
+  SanitizerCommonDecorator d;
+  Printf("%s", d.Warning());
+  const char *description = sig.Describe();
+  Report("ERROR: %s: %s on unknown address %p (pc %p bp %p sp %p T%d)\n",
+         SanitizerToolName, description, (void *)sig.addr, (void *)sig.pc,
+         (void *)sig.bp, (void *)sig.sp, tid);
+  Printf("%s", d.Default());
+  if (sig.pc < GetPageSizeCached())
+    Report("Hint: pc points to the zero page.\n");
+  if (sig.is_memory_access) {
+    const char *access_type =
+        sig.write_flag == SignalContext::WRITE
+            ? "WRITE"
+            : (sig.write_flag == SignalContext::READ ? "READ" : "UNKNOWN");
+    Report("The signal is caused by a %s memory access.\n", access_type);
+    if (sig.addr < GetPageSizeCached())
+      Report("Hint: address points to the zero page.\n");
+  }
+  MaybeReportNonExecRegion(sig.pc);
+  InternalScopedBuffer<BufferedStackTrace> stack_buffer(1);
+  BufferedStackTrace *stack = stack_buffer.data();
+  stack->Reset();
+  unwind(sig, unwind_context, stack);
+  stack->Print();
+  MaybeDumpInstructionBytes(sig.pc);
+  MaybeDumpRegisters(sig.context);
+  Printf("%s can not provide additional info.\n", SanitizerToolName);
+  ReportErrorSummary(description, stack);
+}
+
+void ReportDeadlySignal(const SignalContext &sig, u32 tid,
+                        UnwindSignalStackCallbackType unwind,
+                        const void *unwind_context) {
+  if (sig.IsStackOverflow())
+    ReportStackOverflowImpl(sig, tid, unwind, unwind_context);
+  else
+    ReportDeadlySignalImpl(sig, tid, unwind, unwind_context);
+}
+
+void HandleDeadlySignal(void *siginfo, void *context, u32 tid,
+                        UnwindSignalStackCallbackType unwind,
+                        const void *unwind_context) {
+  StartReportDeadlySignal();
+  ScopedErrorReportLock rl;
+  SignalContext sig(siginfo, context);
+  ReportDeadlySignal(sig, tid, unwind, unwind_context);
+  Report("ABORTING\n");
+  Die();
+}
+
+#endif  // !SANITIZER_FUCHSIA && !SANITIZER_GO
+
 void WriteToSyslog(const char *msg) {
   InternalScopedString msg_copy(kErrorMessageBufferSize);
   msg_copy.append("%s", msg);
@@ -161,6 +298,47 @@ void MaybeStartBackgroudThread() {
 #endif
 }
 
+static atomic_uintptr_t reporting_thread = {0};
+static StaticSpinMutex CommonSanitizerReportMutex;
+
+ScopedErrorReportLock::ScopedErrorReportLock() {
+  uptr current = GetThreadSelf();
+  for (;;) {
+    uptr expected = 0;
+    if (atomic_compare_exchange_strong(&reporting_thread, &expected, current,
+                                       memory_order_relaxed)) {
+      // We've claimed reporting_thread so proceed.
+      CommonSanitizerReportMutex.Lock();
+      return;
+    }
+
+    if (expected == current) {
+      // This is either asynch signal or nested error during error reporting.
+      // Fail simple to avoid deadlocks in Report().
+
+      // Can't use Report() here because of potential deadlocks in nested
+      // signal handlers.
+      CatastrophicErrorWrite(SanitizerToolName,
+                             internal_strlen(SanitizerToolName));
+      static const char msg[] = ": nested bug in the same thread, aborting.\n";
+      CatastrophicErrorWrite(msg, sizeof(msg) - 1);
+
+      internal__exit(common_flags()->exitcode);
+    }
+
+    internal_sched_yield();
+  }
+}
+
+ScopedErrorReportLock::~ScopedErrorReportLock() {
+  CommonSanitizerReportMutex.Unlock();
+  atomic_store_relaxed(&reporting_thread, 0);
+}
+
+void ScopedErrorReportLock::CheckLocked() {
+  CommonSanitizerReportMutex.CheckLocked();
+}
+
 }  // namespace __sanitizer
 
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_sandbox_on_notify,
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_fuchsia.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_fuchsia.cc
new file mode 100644
index 000000000000..c5be48bceace
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_fuchsia.cc
@@ -0,0 +1,240 @@
+//===-- sanitizer_coverage_fuchsia.cc ------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// Sanitizer Coverage Controller for Trace PC Guard, Fuchsia-specific version.
+//
+// This Fuchsia-specific implementation uses the same basic scheme and the
+// same simple '.sancov' file format as the generic implementation.  The
+// difference is that we just produce a single blob of output for the whole
+// program, not a separate one per DSO.  We do not sort the PC table and do
+// not prune the zeros, so the resulting file is always as large as it
+// would be to report 100% coverage.  Implicit tracing information about
+// the address ranges of DSOs allows offline tools to split the one big
+// blob into separate files that the 'sancov' tool can understand.
+//
+// Unlike the traditional implementation that uses an atexit hook to write
+// out data files at the end, the results on Fuchsia do not go into a file
+// per se.  The 'coverage_dir' option is ignored.  Instead, they are stored
+// directly into a shared memory object (a Zircon VMO).  At exit, that VMO
+// is handed over to a system service that's responsible for getting the
+// data out to somewhere that it can be fed into the sancov tool (where and
+// how is not our problem).
+
+#include "sanitizer_platform.h"
+#if SANITIZER_FUCHSIA
+#include "sanitizer_atomic.h"
+#include "sanitizer_common.h"
+#include "sanitizer_internal_defs.h"
+
+#include <zircon/process.h>
+#include <zircon/sanitizer.h>
+#include <zircon/syscalls.h>
+
+using namespace __sanitizer;  // NOLINT
+
+namespace __sancov {
+namespace {
+
+// TODO(mcgrathr): Move the constant into a header shared with other impls.
+constexpr u64 Magic64 = 0xC0BFFFFFFFFFFF64ULL;
+static_assert(SANITIZER_WORDSIZE == 64, "Fuchsia is always LP64");
+
+constexpr const char kSancovSinkName[] = "sancov";
+
+// Collects trace-pc guard coverage.
+// This class relies on zero-initialization.
+class TracePcGuardController {
+ public:
+  // 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
+  // first call to TracePcGuard (below) will store the corresponding PC at
+  // that index in the array.  (Each later call with the same guard slot is
+  // presumed to be from the same PC.)  Then it clears the guard slot back
+  // to zero, which tells the compiler not to bother calling in again.  At
+  // the end of the run, we have a big array where each element is either
+  // zero or is a tracked PC location that was hit in the trace.
+
+  // This is called from global constructors.  Each translation unit has a
+  // contiguous array of guard slots, and a constructor that calls here
+  // with the bounds of its array.  Those constructors are allowed to call
+  // here more than once for the same array.  Usually all of these
+  // constructors run in the initial thread, but it's possible that a
+  // dlopen call on a secondary thread will run constructors that get here.
+  void InitTracePcGuard(u32 *start, u32 *end) {
+    if (end > start && *start == 0 && common_flags()->coverage) {
+      // Complete the setup before filling in any guards with indices.
+      // This avoids the possibility of code called from Setup reentering
+      // TracePcGuard.
+      u32 idx = Setup(end - start);
+      for (u32 *p = start; p < end; ++p) {
+        *p = idx++;
+      }
+    }
+  }
+
+  void TracePcGuard(u32 *guard, uptr pc) {
+    atomic_uint32_t *guard_ptr = reinterpret_cast<atomic_uint32_t *>(guard);
+    u32 idx = atomic_exchange(guard_ptr, 0, memory_order_relaxed);
+    if (idx > 0) array_[idx] = pc;
+  }
+
+  void Dump() {
+    BlockingMutexLock locked(&setup_lock_);
+    if (array_) {
+      CHECK_NE(vmo_, ZX_HANDLE_INVALID);
+
+      // Publish the VMO to the system, where it can be collected and
+      // analyzed after this process exits.  This always consumes the VMO
+      // handle.  Any failure is just logged and not indicated to us.
+      __sanitizer_publish_data(kSancovSinkName, vmo_);
+      vmo_ = ZX_HANDLE_INVALID;
+
+      // This will route to __sanitizer_log_write, which will ensure that
+      // information about shared libraries is written out.  This message
+      // uses the `dumpfile` symbolizer markup element to highlight the
+      // dump.  See the explanation for this in:
+      // https://fuchsia.googlesource.com/zircon/+/master/docs/symbolizer_markup.md
+      Printf("SanitizerCoverage: {{{dumpfile:%s:%s}}} with up to %u PCs\n",
+             kSancovSinkName, vmo_name_, next_index_ - 1);
+    }
+  }
+
+ private:
+  // We map in the largest possible view into the VMO: one word
+  // for every possible 32-bit index value.  This avoids the need
+  // to change the mapping when increasing the size of the VMO.
+  // We can always spare the 32G of address space.
+  static constexpr size_t MappingSize = sizeof(uptr) << 32;
+
+  BlockingMutex setup_lock_;
+  uptr *array_;
+  u32 next_index_;
+  zx_handle_t vmo_;
+  char vmo_name_[ZX_MAX_NAME_LEN];
+
+  size_t DataSize() const { return next_index_ * sizeof(uintptr_t); }
+
+  u32 Setup(u32 num_guards) {
+    BlockingMutexLock locked(&setup_lock_);
+    DCHECK(common_flags()->coverage);
+
+    if (next_index_ == 0) {
+      CHECK_EQ(vmo_, ZX_HANDLE_INVALID);
+      CHECK_EQ(array_, nullptr);
+
+      // The first sample goes at [1] to reserve [0] for the magic number.
+      next_index_ = 1 + num_guards;
+
+      zx_status_t status = _zx_vmo_create(DataSize(), 0, &vmo_);
+      CHECK_EQ(status, ZX_OK);
+
+      // Give the VMO a name including our process KOID so it's easy to spot.
+      internal_snprintf(vmo_name_, sizeof(vmo_name_), "%s.%zu", kSancovSinkName,
+                        internal_getpid());
+      _zx_object_set_property(vmo_, ZX_PROP_NAME, vmo_name_,
+                              internal_strlen(vmo_name_));
+
+      // Map the largest possible view we might need into the VMO.  Later
+      // we might need to increase the VMO's size before we can use larger
+      // indices, but we'll never move the mapping address so we don't have
+      // any multi-thread synchronization issues with that.
+      uintptr_t mapping;
+      status =
+          _zx_vmar_map(_zx_vmar_root_self(), 0, vmo_, 0, MappingSize,
+                       ZX_VM_FLAG_PERM_READ | ZX_VM_FLAG_PERM_WRITE, &mapping);
+      CHECK_EQ(status, ZX_OK);
+
+      // Hereafter other threads are free to start storing into
+      // elements [1, next_index_) of the big array.
+      array_ = reinterpret_cast<uptr *>(mapping);
+
+      // Store the magic number.
+      // Hereafter, the VMO serves as the contents of the '.sancov' file.
+      array_[0] = Magic64;
+
+      return 1;
+    } else {
+      // The VMO is already mapped in, but it's not big enough to use the
+      // new indices.  So increase the size to cover the new maximum index.
+
+      CHECK_NE(vmo_, ZX_HANDLE_INVALID);
+      CHECK_NE(array_, nullptr);
+
+      uint32_t first_index = next_index_;
+      next_index_ += num_guards;
+
+      zx_status_t status = _zx_vmo_set_size(vmo_, DataSize());
+      CHECK_EQ(status, ZX_OK);
+
+      return first_index;
+    }
+  }
+};
+
+static TracePcGuardController pc_guard_controller;
+
+}  // namespace
+}  // namespace __sancov
+
+namespace __sanitizer {
+void InitializeCoverage(bool enabled, const char *dir) {
+  CHECK_EQ(enabled, common_flags()->coverage);
+  CHECK_EQ(dir, common_flags()->coverage_dir);
+
+  static bool coverage_enabled = false;
+  if (!coverage_enabled) {
+    coverage_enabled = enabled;
+    Atexit(__sanitizer_cov_dump);
+    AddDieCallback(__sanitizer_cov_dump);
+  }
+}
+}  // namespace __sanitizer
+
+extern "C" {
+SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_dump_coverage(  // NOLINT
+    const uptr *pcs, uptr len) {
+  UNIMPLEMENTED();
+}
+
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_pc_guard, u32 *guard) {
+  if (!*guard) return;
+  __sancov::pc_guard_controller.TracePcGuard(guard, GET_CALLER_PC() - 1);
+}
+
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_pc_guard_init,
+                             u32 *start, u32 *end) {
+  if (start == end || *start) return;
+  __sancov::pc_guard_controller.InitTracePcGuard(start, end);
+}
+
+SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_dump_trace_pc_guard_coverage() {
+  __sancov::pc_guard_controller.Dump();
+}
+SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_cov_dump() {
+  __sanitizer_dump_trace_pc_guard_coverage();
+}
+// Default empty implementations (weak). 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) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_cmp4, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_cmp8, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_const_cmp1, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_const_cmp2, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_const_cmp4, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_const_cmp8, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_switch, void) {}
+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) {}
+}  // extern "C"
+
+#endif  // !SANITIZER_FUCHSIA
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_interface.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_interface.inc
index d4749000d80a..f909b74c9752 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_interface.inc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_interface.inc
@@ -9,6 +9,7 @@
 // Sanitizer Coverage interface list.
 //===----------------------------------------------------------------------===//
 INTERFACE_FUNCTION(__sanitizer_cov_dump)
+INTERFACE_FUNCTION(__sanitizer_cov_reset)
 INTERFACE_FUNCTION(__sanitizer_dump_coverage)
 INTERFACE_FUNCTION(__sanitizer_dump_trace_pc_guard_coverage)
 INTERFACE_WEAK_FUNCTION(__sancov_default_options)
@@ -17,6 +18,10 @@ INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_cmp1)
 INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_cmp2)
 INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_cmp4)
 INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_cmp8)
+INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_const_cmp1)
+INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_const_cmp2)
+INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_const_cmp4)
+INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_const_cmp8)
 INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_div4)
 INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_div8)
 INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_gep)
@@ -24,3 +29,5 @@ INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_pc_guard)
 INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_pc_guard_init)
 INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_pc_indir)
 INTERFACE_WEAK_FUNCTION(__sanitizer_cov_trace_switch)
+INTERFACE_WEAK_FUNCTION(__sanitizer_cov_8bit_counters_init)
+INTERFACE_WEAK_FUNCTION(__sanitizer_cov_pcs_init)
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cc
index 24433356c63d..3c5f29b2899b 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cc
@@ -8,10 +8,14 @@
 //===----------------------------------------------------------------------===//
 // Sanitizer Coverage Controller for Trace PC Guard.
 
+#include "sanitizer_platform.h"
+
+#if !SANITIZER_FUCHSIA
 #include "sancov_flags.h"
 #include "sanitizer_allocator_internal.h"
 #include "sanitizer_atomic.h"
 #include "sanitizer_common.h"
+#include "sanitizer_file.h"
 #include "sanitizer_symbolizer.h"
 
 using namespace __sanitizer;
@@ -124,11 +128,17 @@ class TracePcGuardController {
   }
 
   void TracePcGuard(u32* guard, uptr pc) {
-    atomic_uint32_t* guard_ptr = reinterpret_cast<atomic_uint32_t*>(guard);
-    u32 idx = atomic_exchange(guard_ptr, 0, memory_order_relaxed);
+    u32 idx = *guard;
     if (!idx) return;
     // we start indices from 1.
-    pc_vector[idx - 1] = pc;
+    atomic_uintptr_t* pc_ptr =
+        reinterpret_cast<atomic_uintptr_t*>(&pc_vector[idx - 1]);
+    if (atomic_load(pc_ptr, memory_order_relaxed) == 0)
+      atomic_store(pc_ptr, pc, memory_order_relaxed);
+  }
+
+  void Reset() {
+    internal_memset(&pc_vector[0], 0, sizeof(pc_vector[0]) * pc_vector.size());
   }
 
   void Dump() {
@@ -180,15 +190,31 @@ SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_dump_trace_pc_guard_coverage() {
 SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_cov_dump() {
   __sanitizer_dump_trace_pc_guard_coverage();
 }
+SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_cov_reset() {
+  __sancov::pc_guard_controller.Reset();
+}
 // Default empty implementations (weak). 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) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_cmp4, void) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_cmp8, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_const_cmp1, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_const_cmp2, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_const_cmp4, void) {}
+SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_const_cmp8, void) {}
 SANITIZER_INTERFACE_WEAK_DEF(void, __sanitizer_cov_trace_switch, void) {}
 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_pcs_init, void) {}
 }  // extern "C"
+// Weak definition for code instrumented with -fsanitize-coverage=stack-depth
+// and later linked with code containing a strong definition.
+// E.g., -fsanitize=fuzzer-no-link
+SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
+SANITIZER_TLS_INITIAL_EXEC_ATTRIBUTE uptr __sancov_lowest_stack;
+
+#endif  // !SANITIZER_FUCHSIA
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_errno.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_errno.h
index 7872b89c227c..42cc290502bf 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_errno.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_errno.h
@@ -24,8 +24,10 @@
 
 #if SANITIZER_FREEBSD || SANITIZER_MAC
 #  define __errno_location __error
-#elif SANITIZER_ANDROID
+#elif SANITIZER_ANDROID || SANITIZER_NETBSD
 #  define __errno_location __errno
+#elif SANITIZER_SOLARIS
+#  define __errno_location ___errno
 #elif SANITIZER_WINDOWS
 #  define __errno_location _errno
 #endif
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_file.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_file.cc
new file mode 100644
index 000000000000..cde54bfde804
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_file.cc
@@ -0,0 +1,177 @@
+//===-- sanitizer_file.cc ------------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// This file is shared between AddressSanitizer and ThreadSanitizer
+// run-time libraries.  It defines filesystem-related interfaces.  This
+// is separate from sanitizer_common.cc so that it's simpler to disable
+// all the filesystem support code for a port that doesn't use it.
+//
+//===---------------------------------------------------------------------===//
+
+#include "sanitizer_platform.h"
+
+#if !SANITIZER_FUCHSIA
+
+#include "sanitizer_common.h"
+#include "sanitizer_file.h"
+
+namespace __sanitizer {
+
+void CatastrophicErrorWrite(const char *buffer, uptr length) {
+  WriteToFile(kStderrFd, buffer, length);
+}
+
+StaticSpinMutex report_file_mu;
+ReportFile report_file = {&report_file_mu, kStderrFd, "", "", 0};
+
+void RawWrite(const char *buffer) {
+  report_file.Write(buffer, internal_strlen(buffer));
+}
+
+void ReportFile::ReopenIfNecessary() {
+  mu->CheckLocked();
+  if (fd == kStdoutFd || fd == kStderrFd) return;
+
+  uptr pid = internal_getpid();
+  // If in tracer, use the parent's file.
+  if (pid == stoptheworld_tracer_pid)
+    pid = stoptheworld_tracer_ppid;
+  if (fd != kInvalidFd) {
+    // If the report file is already opened by the current process,
+    // do nothing. Otherwise the report file was opened by the parent
+    // process, close it now.
+    if (fd_pid == pid)
+      return;
+    else
+      CloseFile(fd);
+  }
+
+  const char *exe_name = GetProcessName();
+  if (common_flags()->log_exe_name && exe_name) {
+    internal_snprintf(full_path, kMaxPathLength, "%s.%s.%zu", path_prefix,
+                      exe_name, pid);
+  } else {
+    internal_snprintf(full_path, kMaxPathLength, "%s.%zu", path_prefix, pid);
+  }
+  fd = OpenFile(full_path, WrOnly);
+  if (fd == kInvalidFd) {
+    const char *ErrorMsgPrefix = "ERROR: Can't open file: ";
+    WriteToFile(kStderrFd, ErrorMsgPrefix, internal_strlen(ErrorMsgPrefix));
+    WriteToFile(kStderrFd, full_path, internal_strlen(full_path));
+    Die();
+  }
+  fd_pid = pid;
+}
+
+void ReportFile::SetReportPath(const char *path) {
+  if (!path)
+    return;
+  uptr len = internal_strlen(path);
+  if (len > sizeof(path_prefix) - 100) {
+    Report("ERROR: Path is too long: %c%c%c%c%c%c%c%c...\n",
+           path[0], path[1], path[2], path[3],
+           path[4], path[5], path[6], path[7]);
+    Die();
+  }
+
+  SpinMutexLock l(mu);
+  if (fd != kStdoutFd && fd != kStderrFd && fd != kInvalidFd)
+    CloseFile(fd);
+  fd = kInvalidFd;
+  if (internal_strcmp(path, "stdout") == 0) {
+    fd = kStdoutFd;
+  } else if (internal_strcmp(path, "stderr") == 0) {
+    fd = kStderrFd;
+  } else {
+    internal_snprintf(path_prefix, kMaxPathLength, "%s", path);
+  }
+}
+
+bool ReadFileToBuffer(const char *file_name, char **buff, uptr *buff_size,
+                      uptr *read_len, uptr max_len, error_t *errno_p) {
+  uptr PageSize = GetPageSizeCached();
+  uptr kMinFileLen = PageSize;
+  *buff = nullptr;
+  *buff_size = 0;
+  *read_len = 0;
+  // The files we usually open are not seekable, so try different buffer sizes.
+  for (uptr size = kMinFileLen; size <= max_len; size *= 2) {
+    fd_t fd = OpenFile(file_name, RdOnly, errno_p);
+    if (fd == kInvalidFd) return false;
+    UnmapOrDie(*buff, *buff_size);
+    *buff = (char*)MmapOrDie(size, __func__);
+    *buff_size = size;
+    *read_len = 0;
+    // Read up to one page at a time.
+    bool reached_eof = false;
+    while (*read_len + PageSize <= size) {
+      uptr just_read;
+      if (!ReadFromFile(fd, *buff + *read_len, PageSize, &just_read, errno_p)) {
+        UnmapOrDie(*buff, *buff_size);
+        return false;
+      }
+      if (just_read == 0) {
+        reached_eof = true;
+        break;
+      }
+      *read_len += just_read;
+    }
+    CloseFile(fd);
+    if (reached_eof)  // We've read the whole file.
+      break;
+  }
+  return true;
+}
+
+static const char kPathSeparator = SANITIZER_WINDOWS ? ';' : ':';
+
+char *FindPathToBinary(const char *name) {
+  if (FileExists(name)) {
+    return internal_strdup(name);
+  }
+
+  const char *path = GetEnv("PATH");
+  if (!path)
+    return nullptr;
+  uptr name_len = internal_strlen(name);
+  InternalScopedBuffer<char> buffer(kMaxPathLength);
+  const char *beg = path;
+  while (true) {
+    const char *end = internal_strchrnul(beg, kPathSeparator);
+    uptr prefix_len = end - beg;
+    if (prefix_len + name_len + 2 <= kMaxPathLength) {
+      internal_memcpy(buffer.data(), beg, prefix_len);
+      buffer[prefix_len] = '/';
+      internal_memcpy(&buffer[prefix_len + 1], name, name_len);
+      buffer[prefix_len + 1 + name_len] = '\0';
+      if (FileExists(buffer.data()))
+        return internal_strdup(buffer.data());
+    }
+    if (*end == '\0') break;
+    beg = end + 1;
+  }
+  return nullptr;
+}
+
+} // namespace __sanitizer
+
+using namespace __sanitizer;  // NOLINT
+
+extern "C" {
+void __sanitizer_set_report_path(const char *path) {
+  report_file.SetReportPath(path);
+}
+
+void __sanitizer_set_report_fd(void *fd) {
+  report_file.fd = (fd_t)reinterpret_cast<uptr>(fd);
+  report_file.fd_pid = internal_getpid();
+}
+} // extern "C"
+
+#endif  // !SANITIZER_FUCHSIA
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_file.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_file.h
new file mode 100644
index 000000000000..9a12ab7342c7
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_file.h
@@ -0,0 +1,110 @@
+//===-- sanitizer_file.h ---------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// This file is shared between run-time libraries of sanitizers.
+// It declares filesystem-related interfaces.  This is separate from
+// sanitizer_common.h so that it's simpler to disable all the filesystem
+// support code for a port that doesn't use it.
+//
+//===---------------------------------------------------------------------===//
+#ifndef SANITIZER_FILE_H
+#define SANITIZER_FILE_H
+
+#include "sanitizer_interface_internal.h"
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_libc.h"
+#include "sanitizer_mutex.h"
+
+namespace __sanitizer {
+
+struct ReportFile {
+  void Write(const char *buffer, uptr length);
+  bool SupportsColors();
+  void SetReportPath(const char *path);
+
+  // Don't use fields directly. They are only declared public to allow
+  // aggregate initialization.
+
+  // Protects fields below.
+  StaticSpinMutex *mu;
+  // Opened file descriptor. Defaults to stderr. It may be equal to
+  // kInvalidFd, in which case new file will be opened when necessary.
+  fd_t fd;
+  // Path prefix of report file, set via __sanitizer_set_report_path.
+  char path_prefix[kMaxPathLength];
+  // Full path to report, obtained as <path_prefix>.PID
+  char full_path[kMaxPathLength];
+  // PID of the process that opened fd. If a fork() occurs,
+  // the PID of child will be different from fd_pid.
+  uptr fd_pid;
+
+ private:
+  void ReopenIfNecessary();
+};
+extern ReportFile report_file;
+
+enum FileAccessMode {
+  RdOnly,
+  WrOnly,
+  RdWr
+};
+
+// Returns kInvalidFd on error.
+fd_t OpenFile(const char *filename, FileAccessMode mode,
+              error_t *errno_p = nullptr);
+void CloseFile(fd_t);
+
+// Return true on success, false on error.
+bool ReadFromFile(fd_t fd, void *buff, uptr buff_size,
+                  uptr *bytes_read = nullptr, error_t *error_p = nullptr);
+bool WriteToFile(fd_t fd, const void *buff, uptr buff_size,
+                 uptr *bytes_written = nullptr, error_t *error_p = nullptr);
+
+bool RenameFile(const char *oldpath, const char *newpath,
+                error_t *error_p = nullptr);
+
+// Scoped file handle closer.
+struct FileCloser {
+  explicit FileCloser(fd_t fd) : fd(fd) {}
+  ~FileCloser() { CloseFile(fd); }
+  fd_t fd;
+};
+
+bool SupportsColoredOutput(fd_t fd);
+
+// OS
+const char *GetPwd();
+bool FileExists(const char *filename);
+char *FindPathToBinary(const char *name);
+bool IsPathSeparator(const char c);
+bool IsAbsolutePath(const char *path);
+// 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
+// in parent process even in case of an error.
+// The child process will close all fds after STDERR_FILENO
+// before passing control to a program.
+pid_t StartSubprocess(const char *filename, const char *const argv[],
+                      fd_t stdin_fd = kInvalidFd, fd_t stdout_fd = kInvalidFd,
+                      fd_t stderr_fd = kInvalidFd);
+// Checks if specified process is still running
+bool IsProcessRunning(pid_t pid);
+// Waits for the process to finish and returns its exit code.
+// Returns -1 in case of an error.
+int WaitForProcess(pid_t pid);
+
+// Maps given file to virtual memory, and returns pointer to it
+// (or NULL if mapping fails). Stores the size of mmaped region
+// in '*buff_size'.
+void *MapFileToMemory(const char *file_name, uptr *buff_size);
+void *MapWritableFileToMemory(void *addr, uptr size, fd_t fd, OFF_T offset);
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_FILE_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_flag_parser.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_flag_parser.h
index 4988fbb7a86d..f649f5bffea7 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_flag_parser.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_flag_parser.h
@@ -75,7 +75,7 @@ inline bool FlagHandler<HandleSignalMode>::Parse(const char *value) {
 
 template <>
 inline bool FlagHandler<const char *>::Parse(const char *value) {
-  *t_ = internal_strdup(value);
+  *t_ = value;
   return true;
 }
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_flags.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_flags.inc
index 9f71861efcdb..445e25f9df70 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_flags.inc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_flags.inc
@@ -62,6 +62,9 @@ COMMON_FLAG(
 COMMON_FLAG(
     int, verbosity, 0,
     "Verbosity level (0 - silent, 1 - a bit of output, 2+ - more output).")
+COMMON_FLAG(bool, strip_env, 1,
+            "Whether to remove the sanitizer from DYLD_INSERT_LIBRARIES to "
+            "avoid passing it to children. Default is true.")
 COMMON_FLAG(bool, detect_leaks, !SANITIZER_MAC, "Enable memory leak detection.")
 COMMON_FLAG(
     bool, leak_check_at_exit, true,
@@ -129,11 +132,11 @@ COMMON_FLAG(uptr, soft_rss_limit_mb, 0,
             " This limit does not affect memory allocations other than"
             " malloc/new.")
 COMMON_FLAG(bool, heap_profile, false, "Experimental heap profiler, asan-only")
-COMMON_FLAG(s32, allocator_release_to_os_interval_ms, kReleaseToOSIntervalNever,
-            "Experimental. Only affects a 64-bit allocator. If set, tries to "
-            "release unused memory to the OS, but not more often than this "
-            "interval (in milliseconds). Negative values mean do not attempt "
-            "to release memory to the OS.\n")
+COMMON_FLAG(s32, allocator_release_to_os_interval_ms, 5000,
+            "Only affects a 64-bit allocator. If set, tries to release unused "
+            "memory to the OS, but not more often than this interval (in "
+            "milliseconds). Negative values mean do not attempt to release "
+            "memory to the OS.\n")
 COMMON_FLAG(bool, can_use_proc_maps_statm, true,
             "If false, do not attempt to read /proc/maps/statm."
             " Mostly useful for testing sanitizers.")
@@ -229,3 +232,8 @@ COMMON_FLAG(bool, print_cmdline, false, "Print command line on crash "
             "(asan only).")
 COMMON_FLAG(bool, html_cov_report, false, "Generate html coverage report.")
 COMMON_FLAG(const char *, sancov_path, "sancov", "Sancov tool location.")
+COMMON_FLAG(bool, dump_instruction_bytes, false,
+          "If true, dump 16 bytes starting at the instruction that caused SEGV")
+COMMON_FLAG(bool, dump_registers, true,
+          "If true, dump values of CPU registers when SEGV happens. Only "
+          "available on OS X for now.")
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.cc
new file mode 100644
index 000000000000..936ec794b8e8
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.cc
@@ -0,0 +1,539 @@
+//===-- sanitizer_fuchsia.cc ---------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// This file is shared between AddressSanitizer and other sanitizer
+// run-time libraries and implements Fuchsia-specific functions from
+// sanitizer_common.h.
+//===---------------------------------------------------------------------===//
+
+#include "sanitizer_fuchsia.h"
+#if SANITIZER_FUCHSIA
+
+#include "sanitizer_common.h"
+#include "sanitizer_libc.h"
+#include "sanitizer_mutex.h"
+#include "sanitizer_stacktrace.h"
+
+#include <limits.h>
+#include <pthread.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <unwind.h>
+#include <zircon/errors.h>
+#include <zircon/process.h>
+#include <zircon/syscalls.h>
+
+namespace __sanitizer {
+
+void NORETURN internal__exit(int exitcode) { _zx_process_exit(exitcode); }
+
+uptr internal_sched_yield() {
+  zx_status_t status = _zx_nanosleep(0);
+  CHECK_EQ(status, ZX_OK);
+  return 0;  // Why doesn't this return void?
+}
+
+static void internal_nanosleep(zx_time_t ns) {
+  zx_status_t status = _zx_nanosleep(_zx_deadline_after(ns));
+  CHECK_EQ(status, ZX_OK);
+}
+
+unsigned int internal_sleep(unsigned int seconds) {
+  internal_nanosleep(ZX_SEC(seconds));
+  return 0;
+}
+
+u64 NanoTime() { return _zx_time_get(ZX_CLOCK_UTC); }
+
+u64 MonotonicNanoTime() { return _zx_time_get(ZX_CLOCK_MONOTONIC); }
+
+uptr internal_getpid() {
+  zx_info_handle_basic_t info;
+  zx_status_t status =
+      _zx_object_get_info(_zx_process_self(), ZX_INFO_HANDLE_BASIC, &info,
+                          sizeof(info), NULL, NULL);
+  CHECK_EQ(status, ZX_OK);
+  uptr pid = static_cast<uptr>(info.koid);
+  CHECK_EQ(pid, info.koid);
+  return pid;
+}
+
+uptr GetThreadSelf() { return reinterpret_cast<uptr>(thrd_current()); }
+
+uptr GetTid() { return GetThreadSelf(); }
+
+void Abort() { abort(); }
+
+int Atexit(void (*function)(void)) { return atexit(function); }
+
+void SleepForSeconds(int seconds) { internal_sleep(seconds); }
+
+void SleepForMillis(int millis) { internal_nanosleep(ZX_MSEC(millis)); }
+
+void GetThreadStackTopAndBottom(bool, uptr *stack_top, uptr *stack_bottom) {
+  pthread_attr_t attr;
+  CHECK_EQ(pthread_getattr_np(pthread_self(), &attr), 0);
+  void *base;
+  size_t size;
+  CHECK_EQ(pthread_attr_getstack(&attr, &base, &size), 0);
+  CHECK_EQ(pthread_attr_destroy(&attr), 0);
+
+  *stack_bottom = reinterpret_cast<uptr>(base);
+  *stack_top = *stack_bottom + size;
+}
+
+void MaybeReexec() {}
+void PrepareForSandboxing(__sanitizer_sandbox_arguments *args) {}
+void DisableCoreDumperIfNecessary() {}
+void InstallDeadlySignalHandlers(SignalHandlerType handler) {}
+void StartReportDeadlySignal() {}
+void ReportDeadlySignal(const SignalContext &sig, u32 tid,
+                        UnwindSignalStackCallbackType unwind,
+                        const void *unwind_context) {}
+void SetAlternateSignalStack() {}
+void UnsetAlternateSignalStack() {}
+void InitTlsSize() {}
+
+void PrintModuleMap() {}
+
+bool SignalContext::IsStackOverflow() const { return false; }
+void SignalContext::DumpAllRegisters(void *context) { UNIMPLEMENTED(); }
+const char *SignalContext::Describe() const { UNIMPLEMENTED(); }
+
+struct UnwindTraceArg {
+  BufferedStackTrace *stack;
+  u32 max_depth;
+};
+
+_Unwind_Reason_Code Unwind_Trace(struct _Unwind_Context *ctx, void *param) {
+  UnwindTraceArg *arg = static_cast<UnwindTraceArg *>(param);
+  CHECK_LT(arg->stack->size, arg->max_depth);
+  uptr pc = _Unwind_GetIP(ctx);
+  if (pc < PAGE_SIZE) return _URC_NORMAL_STOP;
+  arg->stack->trace_buffer[arg->stack->size++] = pc;
+  return (arg->stack->size == arg->max_depth ? _URC_NORMAL_STOP
+                                             : _URC_NO_REASON);
+}
+
+void BufferedStackTrace::SlowUnwindStack(uptr pc, u32 max_depth) {
+  CHECK_GE(max_depth, 2);
+  size = 0;
+  UnwindTraceArg arg = {this, Min(max_depth + 1, kStackTraceMax)};
+  _Unwind_Backtrace(Unwind_Trace, &arg);
+  CHECK_GT(size, 0);
+  // We need to pop a few frames so that pc is on top.
+  uptr to_pop = LocatePcInTrace(pc);
+  // trace_buffer[0] belongs to the current function so we always pop it,
+  // unless there is only 1 frame in the stack trace (1 frame is always better
+  // than 0!).
+  PopStackFrames(Min(to_pop, static_cast<uptr>(1)));
+  trace_buffer[0] = pc;
+}
+
+void BufferedStackTrace::SlowUnwindStackWithContext(uptr pc, void *context,
+                                                    u32 max_depth) {
+  CHECK_NE(context, nullptr);
+  UNREACHABLE("signal context doesn't exist");
+}
+
+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_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() {
+  atomic_uint32_t *m = reinterpret_cast<atomic_uint32_t *>(&opaque_storage_);
+  CHECK_NE(MtxUnlocked, atomic_load(m, memory_order_relaxed));
+}
+
+uptr GetPageSize() { return PAGE_SIZE; }
+
+uptr GetMmapGranularity() { return PAGE_SIZE; }
+
+sanitizer_shadow_bounds_t ShadowBounds;
+
+uptr GetMaxUserVirtualAddress() {
+  ShadowBounds = __sanitizer_shadow_bounds();
+  return ShadowBounds.memory_limit - 1;
+}
+
+uptr GetMaxVirtualAddress() { return GetMaxUserVirtualAddress(); }
+
+static void *DoAnonymousMmapOrDie(uptr size, const char *mem_type,
+                                  bool raw_report, bool die_for_nomem) {
+  size = RoundUpTo(size, PAGE_SIZE);
+
+  zx_handle_t vmo;
+  zx_status_t status = _zx_vmo_create(size, 0, &vmo);
+  if (status != ZX_OK) {
+    if (status != ZX_ERR_NO_MEMORY || die_for_nomem)
+      ReportMmapFailureAndDie(size, mem_type, "zx_vmo_create", status,
+                              raw_report);
+    return nullptr;
+  }
+  _zx_object_set_property(vmo, ZX_PROP_NAME, mem_type,
+                          internal_strlen(mem_type));
+
+  // TODO(mcgrathr): Maybe allocate a VMAR for all sanitizer heap and use that?
+  uintptr_t addr;
+  status = _zx_vmar_map(_zx_vmar_root_self(), 0, vmo, 0, size,
+                        ZX_VM_FLAG_PERM_READ | ZX_VM_FLAG_PERM_WRITE, &addr);
+  _zx_handle_close(vmo);
+
+  if (status != ZX_OK) {
+    if (status != ZX_ERR_NO_MEMORY || die_for_nomem)
+      ReportMmapFailureAndDie(size, mem_type, "zx_vmar_map", status,
+                              raw_report);
+    return nullptr;
+  }
+
+  IncreaseTotalMmap(size);
+
+  return reinterpret_cast<void *>(addr);
+}
+
+void *MmapOrDie(uptr size, const char *mem_type, bool raw_report) {
+  return DoAnonymousMmapOrDie(size, mem_type, raw_report, true);
+}
+
+void *MmapNoReserveOrDie(uptr size, const char *mem_type) {
+  return MmapOrDie(size, mem_type);
+}
+
+void *MmapOrDieOnFatalError(uptr size, const char *mem_type) {
+  return DoAnonymousMmapOrDie(size, mem_type, false, false);
+}
+
+uptr ReservedAddressRange::Init(uptr init_size, const char *name,
+                                uptr fixed_addr) {
+  init_size = RoundUpTo(init_size, PAGE_SIZE);
+  DCHECK_EQ(os_handle_, ZX_HANDLE_INVALID);
+  uintptr_t base;
+  zx_handle_t vmar;
+  zx_status_t status =
+      _zx_vmar_allocate(_zx_vmar_root_self(), 0, init_size,
+                        ZX_VM_FLAG_CAN_MAP_READ | ZX_VM_FLAG_CAN_MAP_WRITE |
+                            ZX_VM_FLAG_CAN_MAP_SPECIFIC,
+                        &vmar, &base);
+  if (status != ZX_OK)
+    ReportMmapFailureAndDie(init_size, name, "zx_vmar_allocate", status);
+  base_ = reinterpret_cast<void *>(base);
+  size_ = init_size;
+  name_ = name;
+  os_handle_ = vmar;
+
+  return reinterpret_cast<uptr>(base_);
+}
+
+static uptr DoMmapFixedOrDie(zx_handle_t vmar, uptr fixed_addr, uptr map_size,
+                             void *base, const char *name, bool die_for_nomem) {
+  uptr offset = fixed_addr - reinterpret_cast<uptr>(base);
+  map_size = RoundUpTo(map_size, PAGE_SIZE);
+  zx_handle_t vmo;
+  zx_status_t status = _zx_vmo_create(map_size, 0, &vmo);
+  if (status != ZX_OK) {
+    if (status != ZX_ERR_NO_MEMORY || die_for_nomem)
+      ReportMmapFailureAndDie(map_size, name, "zx_vmo_create", status);
+    return 0;
+  }
+  _zx_object_set_property(vmo, ZX_PROP_NAME, name, sizeof(name) - 1);
+  DCHECK_GE(base + size_, map_size + offset);
+  uintptr_t addr;
+
+  status = _zx_vmar_map(
+      vmar, offset, vmo, 0, map_size,
+      ZX_VM_FLAG_PERM_READ | ZX_VM_FLAG_PERM_WRITE | ZX_VM_FLAG_SPECIFIC,
+      &addr);
+  _zx_handle_close(vmo);
+  if (status != ZX_OK) {
+    if (status != ZX_ERR_NO_MEMORY || die_for_nomem) {
+      ReportMmapFailureAndDie(map_size, name, "zx_vmar_map", status);
+    }
+    return 0;
+  }
+  IncreaseTotalMmap(map_size);
+  return addr;
+}
+
+uptr ReservedAddressRange::Map(uptr fixed_addr, uptr map_size) {
+  return DoMmapFixedOrDie(os_handle_, fixed_addr, map_size, base_,
+                          name_, false);
+}
+
+uptr ReservedAddressRange::MapOrDie(uptr fixed_addr, uptr map_size) {
+  return DoMmapFixedOrDie(os_handle_, fixed_addr, map_size, base_,
+                          name_, true);
+}
+
+void UnmapOrDieVmar(void *addr, uptr size, zx_handle_t target_vmar) {
+  if (!addr || !size) return;
+  size = RoundUpTo(size, PAGE_SIZE);
+
+  zx_status_t status =
+      _zx_vmar_unmap(target_vmar, reinterpret_cast<uintptr_t>(addr), size);
+  if (status != ZX_OK) {
+    Report("ERROR: %s failed to deallocate 0x%zx (%zd) bytes at address %p\n",
+           SanitizerToolName, size, size, addr);
+    CHECK("unable to unmap" && 0);
+  }
+
+  DecreaseTotalMmap(size);
+}
+
+void ReservedAddressRange::Unmap(uptr fixed_addr, uptr size) {
+  uptr offset = fixed_addr - reinterpret_cast<uptr>(base_);
+  uptr addr = reinterpret_cast<uptr>(base_) + offset;
+  void *addr_as_void = reinterpret_cast<void *>(addr);
+  uptr base_as_uptr = reinterpret_cast<uptr>(base_);
+  // Only unmap at the beginning or end of the range.
+  CHECK((addr_as_void == base_) || (addr + size == base_as_uptr + size_));
+  CHECK_LE(size, size_);
+  UnmapOrDieVmar(reinterpret_cast<void *>(addr), size,
+                 static_cast<zx_handle_t>(os_handle_));
+  if (addr_as_void == base_) {
+    base_ = reinterpret_cast<void *>(addr + size);
+  }
+  size_ = size_ - size;
+}
+
+// This should never be called.
+void *MmapFixedNoAccess(uptr fixed_addr, uptr size, const char *name) {
+  UNIMPLEMENTED();
+}
+
+void *MmapAlignedOrDieOnFatalError(uptr size, uptr alignment,
+                                   const char *mem_type) {
+  CHECK_GE(size, PAGE_SIZE);
+  CHECK(IsPowerOfTwo(size));
+  CHECK(IsPowerOfTwo(alignment));
+
+  zx_handle_t vmo;
+  zx_status_t status = _zx_vmo_create(size, 0, &vmo);
+  if (status != ZX_OK) {
+    if (status != ZX_ERR_NO_MEMORY)
+      ReportMmapFailureAndDie(size, mem_type, "zx_vmo_create", status, false);
+    return nullptr;
+  }
+  _zx_object_set_property(vmo, ZX_PROP_NAME, mem_type,
+                          internal_strlen(mem_type));
+
+  // TODO(mcgrathr): Maybe allocate a VMAR for all sanitizer heap and use that?
+
+  // Map a larger size to get a chunk of address space big enough that
+  // it surely contains an aligned region of the requested size.  Then
+  // overwrite the aligned middle portion with a mapping from the
+  // beginning of the VMO, and unmap the excess before and after.
+  size_t map_size = size + alignment;
+  uintptr_t addr;
+  status = _zx_vmar_map(_zx_vmar_root_self(), 0, vmo, 0, map_size,
+                        ZX_VM_FLAG_PERM_READ | ZX_VM_FLAG_PERM_WRITE, &addr);
+  if (status == ZX_OK) {
+    uintptr_t map_addr = addr;
+    uintptr_t map_end = map_addr + map_size;
+    addr = RoundUpTo(map_addr, alignment);
+    uintptr_t end = addr + size;
+    if (addr != map_addr) {
+      zx_info_vmar_t info;
+      status = _zx_object_get_info(_zx_vmar_root_self(), ZX_INFO_VMAR, &info,
+                                   sizeof(info), NULL, NULL);
+      if (status == ZX_OK) {
+        uintptr_t new_addr;
+        status =
+            _zx_vmar_map(_zx_vmar_root_self(), addr - info.base, vmo, 0, size,
+                         ZX_VM_FLAG_PERM_READ | ZX_VM_FLAG_PERM_WRITE |
+                             ZX_VM_FLAG_SPECIFIC_OVERWRITE,
+                         &new_addr);
+        if (status == ZX_OK) CHECK_EQ(new_addr, addr);
+      }
+    }
+    if (status == ZX_OK && addr != map_addr)
+      status = _zx_vmar_unmap(_zx_vmar_root_self(), map_addr, addr - map_addr);
+    if (status == ZX_OK && end != map_end)
+      status = _zx_vmar_unmap(_zx_vmar_root_self(), end, map_end - end);
+  }
+  _zx_handle_close(vmo);
+
+  if (status != ZX_OK) {
+    if (status != ZX_ERR_NO_MEMORY)
+      ReportMmapFailureAndDie(size, mem_type, "zx_vmar_map", status, false);
+    return nullptr;
+  }
+
+  IncreaseTotalMmap(size);
+
+  return reinterpret_cast<void *>(addr);
+}
+
+void UnmapOrDie(void *addr, uptr size) {
+  UnmapOrDieVmar(addr, size, _zx_vmar_root_self());
+}
+
+// This is used on the shadow mapping, which cannot be changed.
+// Zircon doesn't have anything like MADV_DONTNEED.
+void ReleaseMemoryPagesToOS(uptr beg, uptr end) {}
+
+void DumpProcessMap() {
+  // TODO(mcgrathr): write it
+  return;
+}
+
+bool IsAccessibleMemoryRange(uptr beg, uptr size) {
+  // TODO(mcgrathr): Figure out a better way.
+  zx_handle_t vmo;
+  zx_status_t status = _zx_vmo_create(size, 0, &vmo);
+  if (status == ZX_OK) {
+    while (size > 0) {
+      size_t wrote;
+      status = _zx_vmo_write(vmo, reinterpret_cast<const void *>(beg), 0, size,
+                             &wrote);
+      if (status != ZX_OK) break;
+      CHECK_GT(wrote, 0);
+      CHECK_LE(wrote, size);
+      beg += wrote;
+      size -= wrote;
+    }
+    _zx_handle_close(vmo);
+  }
+  return status == ZX_OK;
+}
+
+// FIXME implement on this platform.
+void GetMemoryProfile(fill_profile_f cb, uptr *stats, uptr stats_size) {}
+
+bool ReadFileToBuffer(const char *file_name, char **buff, uptr *buff_size,
+                      uptr *read_len, uptr max_len, error_t *errno_p) {
+  zx_handle_t vmo;
+  zx_status_t status = __sanitizer_get_configuration(file_name, &vmo);
+  if (status == ZX_OK) {
+    uint64_t vmo_size;
+    status = _zx_vmo_get_size(vmo, &vmo_size);
+    if (status == ZX_OK) {
+      if (vmo_size < max_len) max_len = vmo_size;
+      size_t map_size = RoundUpTo(max_len, PAGE_SIZE);
+      uintptr_t addr;
+      status = _zx_vmar_map(_zx_vmar_root_self(), 0, vmo, 0, map_size,
+                            ZX_VM_FLAG_PERM_READ, &addr);
+      if (status == ZX_OK) {
+        *buff = reinterpret_cast<char *>(addr);
+        *buff_size = map_size;
+        *read_len = max_len;
+      }
+    }
+    _zx_handle_close(vmo);
+  }
+  if (status != ZX_OK && errno_p) *errno_p = status;
+  return status == ZX_OK;
+}
+
+void RawWrite(const char *buffer) {
+  __sanitizer_log_write(buffer, internal_strlen(buffer));
+}
+
+void CatastrophicErrorWrite(const char *buffer, uptr length) {
+  __sanitizer_log_write(buffer, length);
+}
+
+char **StoredArgv;
+char **StoredEnviron;
+
+char **GetArgv() { return StoredArgv; }
+
+const char *GetEnv(const char *name) {
+  if (StoredEnviron) {
+    uptr NameLen = internal_strlen(name);
+    for (char **Env = StoredEnviron; *Env != 0; Env++) {
+      if (internal_strncmp(*Env, name, NameLen) == 0 && (*Env)[NameLen] == '=')
+        return (*Env) + NameLen + 1;
+    }
+  }
+  return nullptr;
+}
+
+uptr ReadBinaryName(/*out*/ char *buf, uptr buf_len) {
+  const char *argv0 = StoredArgv[0];
+  if (!argv0) argv0 = "<UNKNOWN>";
+  internal_strncpy(buf, argv0, buf_len);
+  return internal_strlen(buf);
+}
+
+uptr ReadLongProcessName(/*out*/ char *buf, uptr buf_len) {
+  return ReadBinaryName(buf, buf_len);
+}
+
+uptr MainThreadStackBase, MainThreadStackSize;
+
+bool GetRandom(void *buffer, uptr length, bool blocking) {
+  CHECK_LE(length, ZX_CPRNG_DRAW_MAX_LEN);
+  size_t size;
+  CHECK_EQ(_zx_cprng_draw(buffer, length, &size), ZX_OK);
+  CHECK_EQ(size, length);
+  return true;
+}
+
+u32 GetNumberOfCPUs() {
+  return zx_system_get_num_cpus();
+}
+
+uptr GetRSS() { UNIMPLEMENTED(); }
+
+}  // namespace __sanitizer
+
+using namespace __sanitizer;  // NOLINT
+
+extern "C" {
+void __sanitizer_startup_hook(int argc, char **argv, char **envp,
+                              void *stack_base, size_t stack_size) {
+  __sanitizer::StoredArgv = argv;
+  __sanitizer::StoredEnviron = envp;
+  __sanitizer::MainThreadStackBase = reinterpret_cast<uintptr_t>(stack_base);
+  __sanitizer::MainThreadStackSize = stack_size;
+}
+
+void __sanitizer_set_report_path(const char *path) {
+  // Handle the initialization code in each sanitizer, but no other calls.
+  // This setting is never consulted on Fuchsia.
+  DCHECK_EQ(path, common_flags()->log_path);
+}
+
+void __sanitizer_set_report_fd(void *fd) {
+  UNREACHABLE("not available on Fuchsia");
+}
+}  // extern "C"
+
+#endif  // SANITIZER_FUCHSIA
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.h
new file mode 100644
index 000000000000..18821b4fd9f1
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_fuchsia.h
@@ -0,0 +1,31 @@
+//===-- sanitizer_fuchsia.h ------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===---------------------------------------------------------------------===//
+//
+// Fuchsia-specific sanitizer support.
+//
+//===---------------------------------------------------------------------===//
+#ifndef SANITIZER_FUCHSIA_H
+#define SANITIZER_FUCHSIA_H
+
+#include "sanitizer_platform.h"
+#if SANITIZER_FUCHSIA
+
+#include "sanitizer_common.h"
+
+#include <zircon/sanitizer.h>
+
+namespace __sanitizer {
+
+extern uptr MainThreadStackBase, MainThreadStackSize;
+extern sanitizer_shadow_bounds_t ShadowBounds;
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_FUCHSIA
+#endif  // SANITIZER_FUCHSIA_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_getauxval.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_getauxval.h
new file mode 100644
index 000000000000..934e311a7851
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_getauxval.h
@@ -0,0 +1,47 @@
+//===-- sanitizer_getauxval.h -----------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Common getauxval() guards and definitions.
+// getauxval() is not defined until glibc version 2.16, or until API level 21
+// for Android.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef SANITIZER_GETAUXVAL_H
+#define SANITIZER_GETAUXVAL_H
+
+#include "sanitizer_platform.h"
+
+#if SANITIZER_LINUX
+
+# include <features.h>
+
+# ifndef __GLIBC_PREREQ
+#  define __GLIBC_PREREQ(x, y) 0
+# endif
+
+# if __GLIBC_PREREQ(2, 16) || (SANITIZER_ANDROID && __ANDROID_API__ >= 21)
+#  define SANITIZER_USE_GETAUXVAL 1
+# else
+#  define SANITIZER_USE_GETAUXVAL 0
+# endif
+
+# if SANITIZER_USE_GETAUXVAL
+#  include <sys/auxv.h>
+# else
+// The weak getauxval definition allows to check for the function at runtime.
+// This is useful for Android, when compiled at a lower API level yet running
+// on a more recent platform that offers the function.
+extern "C" SANITIZER_WEAK_ATTRIBUTE
+unsigned long getauxval(unsigned long type);  // NOLINT
+# endif
+
+#endif // SANITIZER_LINUX
+
+#endif // SANITIZER_GETAUXVAL_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_interface_internal.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_interface_internal.h
index b28d8f08e7a3..942f0b1cb586 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_interface_internal.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_interface_internal.h
@@ -46,7 +46,6 @@ extern "C" {
   SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
   void __sanitizer_report_error_summary(const char *error_summary);
 
-  SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_cov_init();
   SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_cov_dump();
   SANITIZER_INTERFACE_ATTRIBUTE void __sanitizer_dump_coverage(
       const __sanitizer::uptr *pcs, const __sanitizer::uptr len);
@@ -81,6 +80,14 @@ extern "C" {
   SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
   void __sanitizer_cov_trace_cmp8();
   SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
+  void __sanitizer_cov_trace_const_cmp1();
+  SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
+  void __sanitizer_cov_trace_const_cmp2();
+  SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
+  void __sanitizer_cov_trace_const_cmp4();
+  SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
+  void __sanitizer_cov_trace_const_cmp8();
+  SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
   void __sanitizer_cov_trace_switch();
   SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
   void __sanitizer_cov_trace_div4();
@@ -95,6 +102,10 @@ 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_pcs_init();
 } // extern "C"
 
 #endif  // SANITIZER_INTERFACE_INTERNAL_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_internal_defs.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_internal_defs.h
index f35b095ee94e..dc480e75f98d 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_internal_defs.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_internal_defs.h
@@ -35,6 +35,14 @@
 # define SANITIZER_WEAK_ATTRIBUTE  __attribute__((weak))
 #endif
 
+// TLS is handled differently on different platforms
+#if SANITIZER_LINUX || SANITIZER_NETBSD
+# define SANITIZER_TLS_INITIAL_EXEC_ATTRIBUTE \
+    __attribute__((tls_model("initial-exec"))) thread_local
+#else
+# define SANITIZER_TLS_INITIAL_EXEC_ATTRIBUTE
+#endif
+
 //--------------------------- WEAK FUNCTIONS ---------------------------------//
 // When working with weak functions, to simplify the code and make it more
 // portable, when possible define a default implementation using this macro:
@@ -56,11 +64,13 @@
 
 // SANITIZER_SUPPORTS_WEAK_HOOKS means that we support real weak functions that
 // will evaluate to a null pointer when not defined.
-#if (SANITIZER_LINUX || SANITIZER_MAC) && !SANITIZER_GO
+#ifndef SANITIZER_SUPPORTS_WEAK_HOOKS
+#if (SANITIZER_LINUX || SANITIZER_MAC || SANITIZER_SOLARIS) && !SANITIZER_GO
 # define SANITIZER_SUPPORTS_WEAK_HOOKS 1
 #else
 # define SANITIZER_SUPPORTS_WEAK_HOOKS 0
 #endif
+#endif // SANITIZER_SUPPORTS_WEAK_HOOKS
 // For some weak hooks that will be called very often and we want to avoid the
 // overhead of executing the default implementation when it is not necessary,
 // we can use the flag SANITIZER_SUPPORTS_WEAK_HOOKS to only define the default
@@ -81,6 +91,10 @@
 // FIXME: do we have anything like this on Mac?
 #if SANITIZER_LINUX && !SANITIZER_ANDROID && !defined(PIC)
 # define SANITIZER_CAN_USE_PREINIT_ARRAY 1
+// Before Solaris 11.4, .preinit_array is fully supported only with GNU ld.
+// FIXME: Check for those conditions.
+#elif SANITIZER_SOLARIS && !defined(PIC)
+# define SANITIZER_CAN_USE_PREINIT_ARRAY 1
 #else
 # define SANITIZER_CAN_USE_PREINIT_ARRAY 0
 #endif
@@ -127,14 +141,14 @@ typedef unsigned error_t;
 typedef int fd_t;
 typedef int error_t;
 #endif
+#if SANITIZER_SOLARIS && !defined(_LP64)
+typedef long pid_t;
+#else
 typedef int pid_t;
+#endif
 
-// WARNING: OFF_T may be different from OS type off_t, depending on the value of
-// _FILE_OFFSET_BITS. This definition of OFF_T matches the ABI of system calls
-// like pread and mmap, as opposed to pread64 and mmap64.
-// FreeBSD, Mac and Linux/x86-64 are special.
-#if SANITIZER_FREEBSD || SANITIZER_MAC || \
-  (SANITIZER_LINUX && defined(__x86_64__))
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD || SANITIZER_MAC || \
+    (SANITIZER_LINUX && defined(__x86_64__))
 typedef u64 OFF_T;
 #else
 typedef uptr OFF_T;
@@ -257,8 +271,8 @@ void NORETURN CheckFailed(const char *file, int line, const char *cond,
 
 #define CHECK_IMPL(c1, op, c2) \
   do { \
-    __sanitizer::u64 v1 = (u64)(c1); \
-    __sanitizer::u64 v2 = (u64)(c2); \
+    __sanitizer::u64 v1 = (__sanitizer::u64)(c1); \
+    __sanitizer::u64 v2 = (__sanitizer::u64)(c2); \
     if (UNLIKELY(!(v1 op v2))) \
       __sanitizer::CheckFailed(__FILE__, __LINE__, \
         "(" #c1 ") " #op " (" #c2 ")", v1, v2); \
@@ -384,6 +398,7 @@ namespace __dfsan { using namespace __sanitizer; }  // NOLINT
 namespace __esan  { using namespace __sanitizer; }  // NOLINT
 namespace __lsan  { using namespace __sanitizer; }  // NOLINT
 namespace __msan  { using namespace __sanitizer; }  // NOLINT
+namespace __hwasan  { using namespace __sanitizer; }  // NOLINT
 namespace __tsan  { using namespace __sanitizer; }  // NOLINT
 namespace __scudo { using namespace __sanitizer; }  // NOLINT
 namespace __ubsan { using namespace __sanitizer; }  // NOLINT
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_libignore.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_libignore.cc
index aa4fa88efffa..0df055edae97 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_libignore.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_libignore.cc
@@ -9,7 +9,7 @@
 
 #include "sanitizer_platform.h"
 
-#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_MAC
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_MAC || SANITIZER_NETBSD
 
 #include "sanitizer_libignore.h"
 #include "sanitizer_flags.h"
@@ -125,4 +125,5 @@ void LibIgnore::OnLibraryUnloaded() {
 
 } // namespace __sanitizer
 
-#endif // #if SANITIZER_FREEBSD || SANITIZER_LINUX
+#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_MAC ||
+        // SANITIZER_NETBSD
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux.cc
index 8c3c1e5d6a5d..6c83e8db42a5 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux.cc
@@ -14,10 +14,12 @@
 
 #include "sanitizer_platform.h"
 
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
 
 #include "sanitizer_common.h"
 #include "sanitizer_flags.h"
+#include "sanitizer_getauxval.h"
 #include "sanitizer_internal_defs.h"
 #include "sanitizer_libc.h"
 #include "sanitizer_linux.h"
@@ -27,10 +29,14 @@
 #include "sanitizer_stacktrace.h"
 #include "sanitizer_symbolizer.h"
 
-#if !SANITIZER_FREEBSD
+#if SANITIZER_LINUX
 #include <asm/param.h>
 #endif
 
+#if SANITIZER_NETBSD
+#include <lwp.h>
+#endif
+
 // For mips64, syscall(__NR_stat) fills the buffer in the 'struct kernel_stat'
 // format. Struct kernel_stat is defined as 'struct stat' in asm/stat.h. To
 // access stat from asm/stat.h, without conflicting with definition in
@@ -50,7 +56,9 @@
 #include <pthread.h>
 #include <sched.h>
 #include <sys/mman.h>
+#if !SANITIZER_SOLARIS
 #include <sys/ptrace.h>
+#endif
 #include <sys/resource.h>
 #include <sys/stat.h>
 #include <sys/syscall.h>
@@ -79,22 +87,24 @@ extern "C" {
 extern char **environ;  // provided by crt1
 #endif  // SANITIZER_FREEBSD
 
-#if !SANITIZER_ANDROID
-#include <sys/signal.h>
-#endif
+#if SANITIZER_NETBSD
+#include <limits.h>  // For NAME_MAX
+#include <sys/sysctl.h>
+extern char **environ;  // provided by crt1
+#include <sys/exec.h>
+extern struct ps_strings *__ps_strings;
+#endif  // SANITIZER_NETBSD
 
-#ifndef __GLIBC_PREREQ
-#define __GLIBC_PREREQ(x, y) 0
-#endif
+#if SANITIZER_SOLARIS
+#include <stdlib.h>
+#include <thread.h>
 
-#if SANITIZER_LINUX && __GLIBC_PREREQ(2, 16)
-# define SANITIZER_USE_GETAUXVAL 1
-#else
-# define SANITIZER_USE_GETAUXVAL 0
+extern char **_environ;
+#define environ _environ
 #endif
 
-#if SANITIZER_USE_GETAUXVAL
-#include <sys/auxv.h>
+#if !SANITIZER_ANDROID
+#include <sys/signal.h>
 #endif
 
 #if SANITIZER_LINUX
@@ -125,21 +135,36 @@ extern void internal_sigreturn();
 }
 #endif
 
+#if SANITIZER_LINUX && defined(__NR_getrandom)
+# if !defined(GRND_NONBLOCK)
+#  define GRND_NONBLOCK 1
+# endif
+# define SANITIZER_USE_GETRANDOM 1
+#else
+# define SANITIZER_USE_GETRANDOM 0
+#endif  // SANITIZER_LINUX && defined(__NR_getrandom)
+
 namespace __sanitizer {
 
 #if SANITIZER_LINUX && defined(__x86_64__)
 #include "sanitizer_syscall_linux_x86_64.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
 
 // --------------- sanitizer_libc.h
+#if !SANITIZER_SOLARIS
 #if !SANITIZER_S390
 uptr internal_mmap(void *addr, uptr length, int prot, int flags, int fd,
                    OFF_T offset) {
-#if SANITIZER_FREEBSD || SANITIZER_LINUX_USES_64BIT_SYSCALLS
+#if SANITIZER_NETBSD
+  return internal_syscall_ptr(SYSCALL(mmap), addr, length, prot, flags, fd,
+                              (long)0, offset);
+#elif SANITIZER_FREEBSD || SANITIZER_LINUX_USES_64BIT_SYSCALLS
   return internal_syscall(SYSCALL(mmap), (uptr)addr, length, prot, flags, fd,
                           offset);
 #else
@@ -152,11 +177,11 @@ uptr internal_mmap(void *addr, uptr length, int prot, int flags, int fd,
 #endif // !SANITIZER_S390
 
 uptr internal_munmap(void *addr, uptr length) {
-  return internal_syscall(SYSCALL(munmap), (uptr)addr, length);
+  return internal_syscall_ptr(SYSCALL(munmap), (uptr)addr, length);
 }
 
 int internal_mprotect(void *addr, uptr length, int prot) {
-  return internal_syscall(SYSCALL(mprotect), (uptr)addr, length, prot);
+  return internal_syscall_ptr(SYSCALL(mprotect), (uptr)addr, length, prot);
 }
 
 uptr internal_close(fd_t fd) {
@@ -167,7 +192,7 @@ uptr internal_open(const char *filename, int flags) {
 #if SANITIZER_USES_CANONICAL_LINUX_SYSCALLS
   return internal_syscall(SYSCALL(openat), AT_FDCWD, (uptr)filename, flags);
 #else
-  return internal_syscall(SYSCALL(open), (uptr)filename, flags);
+  return internal_syscall_ptr(SYSCALL(open), (uptr)filename, flags);
 #endif
 }
 
@@ -176,32 +201,36 @@ uptr internal_open(const char *filename, int flags, u32 mode) {
   return internal_syscall(SYSCALL(openat), AT_FDCWD, (uptr)filename, flags,
                           mode);
 #else
-  return internal_syscall(SYSCALL(open), (uptr)filename, flags, mode);
+  return internal_syscall_ptr(SYSCALL(open), (uptr)filename, flags, mode);
 #endif
 }
 
 uptr internal_read(fd_t fd, void *buf, uptr count) {
   sptr res;
-  HANDLE_EINTR(res, (sptr)internal_syscall(SYSCALL(read), fd, (uptr)buf,
+  HANDLE_EINTR(res, (sptr)internal_syscall_ptr(SYSCALL(read), fd, (uptr)buf,
                count));
   return res;
 }
 
 uptr internal_write(fd_t fd, const void *buf, uptr count) {
   sptr res;
-  HANDLE_EINTR(res, (sptr)internal_syscall(SYSCALL(write), fd, (uptr)buf,
+  HANDLE_EINTR(res, (sptr)internal_syscall_ptr(SYSCALL(write), fd, (uptr)buf,
                count));
   return res;
 }
 
 uptr internal_ftruncate(fd_t fd, uptr size) {
   sptr res;
+#if SANITIZER_NETBSD
+  HANDLE_EINTR(res, internal_syscall64(SYSCALL(ftruncate), fd, 0, (s64)size));
+#else
   HANDLE_EINTR(res, (sptr)internal_syscall(SYSCALL(ftruncate), fd,
                (OFF_T)size));
+#endif
   return res;
 }
 
-#if !SANITIZER_LINUX_USES_64BIT_SYSCALLS && !SANITIZER_FREEBSD
+#if !SANITIZER_LINUX_USES_64BIT_SYSCALLS && SANITIZER_LINUX
 static void stat64_to_stat(struct stat64 *in, struct stat *out) {
   internal_memset(out, 0, sizeof(*out));
   out->st_dev = in->st_dev;
@@ -221,6 +250,21 @@ static void stat64_to_stat(struct stat64 *in, struct stat *out) {
 #endif
 
 #if defined(__mips64)
+// Undefine compatibility macros from <sys/stat.h>
+// so that they would not clash with the kernel_stat
+// st_[a|m|c]time fields
+#undef st_atime
+#undef st_mtime
+#undef st_ctime
+#if defined(SANITIZER_ANDROID)
+// Bionic sys/stat.h defines additional macros
+// for compatibility with the old NDKs and
+// they clash with the kernel_stat structure
+// st_[a|m|c]time_nsec fields.
+#undef st_atime_nsec
+#undef st_mtime_nsec
+#undef st_ctime_nsec
+#endif
 static void kernel_stat_to_stat(struct kernel_stat *in, struct stat *out) {
   internal_memset(out, 0, sizeof(*out));
   out->st_dev = in->st_dev;
@@ -233,15 +277,29 @@ static void kernel_stat_to_stat(struct kernel_stat *in, struct stat *out) {
   out->st_size = in->st_size;
   out->st_blksize = in->st_blksize;
   out->st_blocks = in->st_blocks;
-  out->st_atime = in->st_atime_nsec;
-  out->st_mtime = in->st_mtime_nsec;
-  out->st_ctime = in->st_ctime_nsec;
+#if defined(__USE_MISC)     || \
+    defined(__USE_XOPEN2K8) || \
+    defined(SANITIZER_ANDROID)
+  out->st_atim.tv_sec = in->st_atime;
+  out->st_atim.tv_nsec = in->st_atime_nsec;
+  out->st_mtim.tv_sec = in->st_mtime;
+  out->st_mtim.tv_nsec = in->st_mtime_nsec;
+  out->st_ctim.tv_sec = in->st_ctime;
+  out->st_ctim.tv_nsec = in->st_ctime_nsec;
+#else
+  out->st_atime = in->st_atime;
+  out->st_atimensec = in->st_atime_nsec;
+  out->st_mtime = in->st_mtime;
+  out->st_mtimensec = in->st_mtime_nsec;
+  out->st_ctime = in->st_ctime;
+  out->st_atimensec = in->st_ctime_nsec;
+#endif
 }
 #endif
 
 uptr internal_stat(const char *path, void *buf) {
-#if SANITIZER_FREEBSD
-  return internal_syscall(SYSCALL(fstatat), AT_FDCWD, (uptr)path,
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
+  return internal_syscall_ptr(SYSCALL(fstatat), AT_FDCWD, (uptr)path,
                           (uptr)buf, 0);
 #elif SANITIZER_USES_CANONICAL_LINUX_SYSCALLS
   return internal_syscall(SYSCALL(newfstatat), AT_FDCWD, (uptr)path,
@@ -265,7 +323,9 @@ uptr internal_stat(const char *path, void *buf) {
 }
 
 uptr internal_lstat(const char *path, void *buf) {
-#if SANITIZER_FREEBSD
+#if SANITIZER_NETBSD
+  return internal_syscall_ptr(SYSCALL(lstat), path, buf);
+#elif SANITIZER_FREEBSD
   return internal_syscall(SYSCALL(fstatat), AT_FDCWD, (uptr)path,
                           (uptr)buf, AT_SYMLINK_NOFOLLOW);
 #elif SANITIZER_USES_CANONICAL_LINUX_SYSCALLS
@@ -290,15 +350,15 @@ uptr internal_lstat(const char *path, void *buf) {
 }
 
 uptr internal_fstat(fd_t fd, void *buf) {
-#if SANITIZER_FREEBSD || SANITIZER_LINUX_USES_64BIT_SYSCALLS
-# if SANITIZER_MIPS64
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD || SANITIZER_LINUX_USES_64BIT_SYSCALLS
+# if SANITIZER_MIPS64 && !SANITIZER_NETBSD
   // For mips64, fstat syscall fills buffer in the format of kernel_stat
   struct kernel_stat kbuf;
   int res = internal_syscall(SYSCALL(fstat), fd, &kbuf);
   kernel_stat_to_stat(&kbuf, (struct stat *)buf);
   return res;
 # else
-  return internal_syscall(SYSCALL(fstat), fd, (uptr)buf);
+  return internal_syscall_ptr(SYSCALL(fstat), fd, (uptr)buf);
 # endif
 #else
   struct stat64 buf64;
@@ -328,7 +388,7 @@ uptr internal_readlink(const char *path, char *buf, uptr bufsize) {
   return internal_syscall(SYSCALL(readlinkat), AT_FDCWD,
                           (uptr)path, (uptr)buf, bufsize);
 #else
-  return internal_syscall(SYSCALL(readlink), (uptr)path, (uptr)buf, bufsize);
+  return internal_syscall_ptr(SYSCALL(readlink), path, buf, bufsize);
 #endif
 }
 
@@ -336,7 +396,7 @@ uptr internal_unlink(const char *path) {
 #if SANITIZER_USES_CANONICAL_LINUX_SYSCALLS
   return internal_syscall(SYSCALL(unlinkat), AT_FDCWD, (uptr)path, 0);
 #else
-  return internal_syscall(SYSCALL(unlink), (uptr)path);
+  return internal_syscall_ptr(SYSCALL(unlink), (uptr)path);
 #endif
 }
 
@@ -345,7 +405,7 @@ uptr internal_rename(const char *oldpath, const char *newpath) {
   return internal_syscall(SYSCALL(renameat), AT_FDCWD, (uptr)oldpath, AT_FDCWD,
                           (uptr)newpath);
 #else
-  return internal_syscall(SYSCALL(rename), (uptr)oldpath, (uptr)newpath);
+  return internal_syscall_ptr(SYSCALL(rename), (uptr)oldpath, (uptr)newpath);
 #endif
 }
 
@@ -354,7 +414,7 @@ uptr internal_sched_yield() {
 }
 
 void internal__exit(int exitcode) {
-#if SANITIZER_FREEBSD
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
   internal_syscall(SYSCALL(exit), exitcode);
 #else
   internal_syscall(SYSCALL(exit_group), exitcode);
@@ -366,16 +426,17 @@ unsigned int internal_sleep(unsigned int seconds) {
   struct timespec ts;
   ts.tv_sec = 1;
   ts.tv_nsec = 0;
-  int res = internal_syscall(SYSCALL(nanosleep), &ts, &ts);
+  int res = internal_syscall_ptr(SYSCALL(nanosleep), &ts, &ts);
   if (res) return ts.tv_sec;
   return 0;
 }
 
 uptr internal_execve(const char *filename, char *const argv[],
                      char *const envp[]) {
-  return internal_syscall(SYSCALL(execve), (uptr)filename, (uptr)argv,
+  return internal_syscall_ptr(SYSCALL(execve), (uptr)filename, (uptr)argv,
                           (uptr)envp);
 }
+#endif // !SANITIZER_SOLARIS
 
 // ----------------- sanitizer_common.h
 bool FileExists(const char *filename) {
@@ -393,27 +454,37 @@ bool FileExists(const char *filename) {
 tid_t GetTid() {
 #if SANITIZER_FREEBSD
   return (uptr)pthread_self();
+#elif SANITIZER_NETBSD
+  return _lwp_self();
+#elif SANITIZER_SOLARIS
+  return (uptr)thr_self();
 #else
   return internal_syscall(SYSCALL(gettid));
 #endif
 }
 
+#if !SANITIZER_SOLARIS
 u64 NanoTime() {
-#if SANITIZER_FREEBSD
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
   timeval tv;
 #else
   kernel_timeval tv;
 #endif
   internal_memset(&tv, 0, sizeof(tv));
-  internal_syscall(SYSCALL(gettimeofday), (uptr)&tv, 0);
+  internal_syscall_ptr(SYSCALL(gettimeofday), &tv, 0);
   return (u64)tv.tv_sec * 1000*1000*1000 + tv.tv_usec * 1000;
 }
 
+uptr internal_clock_gettime(__sanitizer_clockid_t clk_id, void *tp) {
+  return internal_syscall_ptr(SYSCALL(clock_gettime), clk_id, tp);
+}
+#endif // !SANITIZER_SOLARIS
+
 // Like getenv, but reads env directly from /proc (on Linux) or parses the
-// 'environ' array (on FreeBSD) and does not use libc. This function should be
-// called first inside __asan_init.
+// 'environ' array (on some others) and does not use libc. This function
+// should be called first inside __asan_init.
 const char *GetEnv(const char *name) {
-#if SANITIZER_FREEBSD
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD || SANITIZER_SOLARIS
   if (::environ != 0) {
     uptr NameLen = internal_strlen(name);
     for (char **Env = ::environ; *Env != 0; Env++) {
@@ -451,13 +522,13 @@ const char *GetEnv(const char *name) {
 #endif
 }
 
-#if !SANITIZER_FREEBSD
+#if !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 extern "C" {
   SANITIZER_WEAK_ATTRIBUTE extern void *__libc_stack_end;
 }
 #endif
 
-#if !SANITIZER_GO && !SANITIZER_FREEBSD
+#if !SANITIZER_GO && !SANITIZER_FREEBSD && !SANITIZER_NETBSD
 static void ReadNullSepFileToArray(const char *path, char ***arr,
                                    int arr_size) {
   char *buff;
@@ -483,7 +554,22 @@ static void ReadNullSepFileToArray(const char *path, char ***arr,
 #endif
 
 static void GetArgsAndEnv(char ***argv, char ***envp) {
-#if !SANITIZER_FREEBSD
+#if SANITIZER_FREEBSD
+  // On FreeBSD, retrieving the argument and environment arrays is done via the
+  // kern.ps_strings sysctl, which returns a pointer to a structure containing
+  // this information. See also <sys/exec.h>.
+  ps_strings *pss;
+  size_t sz = sizeof(pss);
+  if (sysctlbyname("kern.ps_strings", &pss, &sz, NULL, 0) == -1) {
+    Printf("sysctl kern.ps_strings failed\n");
+    Die();
+  }
+  *argv = pss->ps_argvstr;
+  *envp = pss->ps_envstr;
+#elif SANITIZER_NETBSD
+  *argv = __ps_strings->ps_argvstr;
+  *argv = __ps_strings->ps_envstr;
+#else
 #if !SANITIZER_GO
   if (&__libc_stack_end) {
 #endif
@@ -498,18 +584,6 @@ static void GetArgsAndEnv(char ***argv, char ***envp) {
     ReadNullSepFileToArray("/proc/self/environ", envp, kMaxEnvp);
   }
 #endif
-#else
-  // On FreeBSD, retrieving the argument and environment arrays is done via the
-  // kern.ps_strings sysctl, which returns a pointer to a structure containing
-  // this information. See also <sys/exec.h>.
-  ps_strings *pss;
-  size_t sz = sizeof(pss);
-  if (sysctlbyname("kern.ps_strings", &pss, &sz, NULL, 0) == -1) {
-    Printf("sysctl kern.ps_strings failed\n");
-    Die();
-  }
-  *argv = pss->ps_argvstr;
-  *envp = pss->ps_envstr;
 #endif
 }
 
@@ -521,14 +595,32 @@ char **GetArgv() {
 
 void ReExec() {
   char **argv, **envp;
+  const char *pathname = "/proc/self/exe";
+
+#if SANITIZER_NETBSD
+  static const int name[] = {
+    CTL_KERN, KERN_PROC_ARGS, -1, KERN_PROC_PATHNAME,
+  };
+  char path[400];
+  size_t len;
+
+  len = sizeof(path);
+  if (sysctl(name, ARRAY_SIZE(name), path, &len, NULL, 0) != -1)
+    pathname = path;
+#elif SANITIZER_SOLARIS
+  pathname = getexecname();
+  CHECK_NE(pathname, NULL);
+#endif
+
   GetArgsAndEnv(&argv, &envp);
-  uptr rv = internal_execve("/proc/self/exe", argv, envp);
+  uptr rv = internal_execve(pathname, argv, envp);
   int rverrno;
   CHECK_EQ(internal_iserror(rv, &rverrno), true);
   Printf("execve failed, errno %d\n", rverrno);
   Die();
 }
 
+#if !SANITIZER_SOLARIS
 enum MutexState {
   MtxUnlocked = 0,
   MtxLocked = 1,
@@ -547,6 +639,8 @@ void BlockingMutex::Lock() {
   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, MtxSleeping, 0, 0, 0);
 #endif
@@ -560,6 +654,8 @@ void BlockingMutex::Unlock() {
   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, 1, 0, 0, 0);
 #endif
@@ -570,11 +666,23 @@ void BlockingMutex::CheckLocked() {
   atomic_uint32_t *m = reinterpret_cast<atomic_uint32_t *>(&opaque_storage_);
   CHECK_NE(MtxUnlocked, atomic_load(m, memory_order_relaxed));
 }
+#endif // !SANITIZER_SOLARIS
 
 // ----------------- sanitizer_linux.h
 // The actual size of this structure is specified by d_reclen.
 // Note that getdents64 uses a different structure format. We only provide the
 // 32-bit syscall here.
+#if SANITIZER_NETBSD
+// struct dirent is different for Linux and us. At this moment, we use only
+// d_fileno (Linux call this d_ino), d_reclen, and d_name.
+struct linux_dirent {
+  u64 d_ino;  // d_fileno
+  u16 d_reclen;
+  u16 d_namlen;  // not used
+  u8 d_type;     // not used
+  char d_name[NAME_MAX + 1];
+};
+#else
 struct linux_dirent {
 #if SANITIZER_X32 || defined(__aarch64__)
   u64 d_ino;
@@ -589,15 +697,29 @@ struct linux_dirent {
 #endif
   char               d_name[256];
 };
+#endif
 
+#if !SANITIZER_SOLARIS
 // Syscall wrappers.
 uptr internal_ptrace(int request, int pid, void *addr, void *data) {
+#if SANITIZER_NETBSD
+  // XXX We need additional work for ptrace:
+  //   - for request, we use PT_FOO whereas Linux uses PTRACE_FOO
+  //   - data is int for us, but void * for Linux
+  //   - Linux sometimes uses data in the case where we use addr instead
+  // At this moment, this function is used only within
+  // "#if SANITIZER_LINUX && defined(__x86_64__)" block in
+  // sanitizer_stoptheworld_linux_libcdep.cc.
+  return internal_syscall_ptr(SYSCALL(ptrace), request, pid, (uptr)addr,
+                              (uptr)data);
+#else
   return internal_syscall(SYSCALL(ptrace), request, pid, (uptr)addr,
                           (uptr)data);
+#endif
 }
 
 uptr internal_waitpid(int pid, int *status, int options) {
-  return internal_syscall(SYSCALL(wait4), pid, (uptr)status, options,
+  return internal_syscall_ptr(SYSCALL(wait4), pid, (uptr)status, options,
                           0 /* rusage */);
 }
 
@@ -615,12 +737,16 @@ uptr internal_getdents(fd_t fd, struct linux_dirent *dirp, unsigned int count) {
 #elif SANITIZER_USES_CANONICAL_LINUX_SYSCALLS
   return internal_syscall(SYSCALL(getdents64), fd, (uptr)dirp, count);
 #else
-  return internal_syscall(SYSCALL(getdents), fd, (uptr)dirp, count);
+  return internal_syscall_ptr(SYSCALL(getdents), fd, (uptr)dirp, count);
 #endif
 }
 
 uptr internal_lseek(fd_t fd, OFF_T offset, int whence) {
+#if SANITIZER_NETBSD
+  return internal_syscall64(SYSCALL(lseek), fd, 0, offset, whence);
+#else
   return internal_syscall(SYSCALL(lseek), fd, offset, whence);
+#endif
 }
 
 #if SANITIZER_LINUX
@@ -630,7 +756,7 @@ uptr internal_prctl(int option, uptr arg2, uptr arg3, uptr arg4, uptr arg5) {
 #endif
 
 uptr internal_sigaltstack(const void *ss, void *oss) {
-  return internal_syscall(SYSCALL(sigaltstack), (uptr)ss, (uptr)oss);
+  return internal_syscall_ptr(SYSCALL(sigaltstack), (uptr)ss, (uptr)oss);
 }
 
 int internal_fork() {
@@ -711,8 +837,8 @@ int internal_sigaction_syscall(int signum, const void *act, void *oldact) {
 
 uptr internal_sigprocmask(int how, __sanitizer_sigset_t *set,
     __sanitizer_sigset_t *oldset) {
-#if SANITIZER_FREEBSD
-  return internal_syscall(SYSCALL(sigprocmask), how, set, oldset);
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
+  return internal_syscall_ptr(SYSCALL(sigprocmask), how, set, oldset);
 #else
   __sanitizer_kernel_sigset_t *k_set = (__sanitizer_kernel_sigset_t *)set;
   __sanitizer_kernel_sigset_t *k_oldset = (__sanitizer_kernel_sigset_t *)oldset;
@@ -751,6 +877,7 @@ bool internal_sigismember(__sanitizer_sigset_t *set, int signum) {
   return k_set->sig[idx] & (1 << bit);
 }
 #endif  // SANITIZER_LINUX
+#endif // !SANITIZER_SOLARIS
 
 // ThreadLister implementation.
 ThreadLister::ThreadLister(int pid)
@@ -856,7 +983,9 @@ static uptr GetKernelAreaSize() {
 #endif  // SANITIZER_WORDSIZE == 32
 
 uptr GetMaxVirtualAddress() {
-#if SANITIZER_WORDSIZE == 64
+#if SANITIZER_NETBSD && defined(__x86_64__)
+  return 0x7f7ffffff000ULL;  // (0x00007f8000000000 - PAGE_SIZE)
+#elif SANITIZER_WORDSIZE == 64
 # if defined(__powerpc64__) || defined(__aarch64__)
   // On PowerPC64 we have two different address space layouts: 44- and 46-bit.
   // We somehow need to figure out which one we are using now and choose
@@ -877,15 +1006,21 @@ uptr GetMaxVirtualAddress() {
 # if defined(__s390__)
   return (1ULL << 31) - 1;  // 0x7fffffff;
 # else
-  uptr res = (1ULL << 32) - 1;  // 0xffffffff;
-  if (!common_flags()->full_address_space)
-    res -= GetKernelAreaSize();
-  CHECK_LT(reinterpret_cast<uptr>(&res), res);
-  return res;
+  return (1ULL << 32) - 1;  // 0xffffffff;
 # endif
 #endif  // SANITIZER_WORDSIZE
 }
 
+uptr GetMaxUserVirtualAddress() {
+  uptr addr = GetMaxVirtualAddress();
+#if SANITIZER_WORDSIZE == 32 && !defined(__s390__)
+  if (!common_flags()->full_address_space)
+    addr -= GetKernelAreaSize();
+  CHECK_LT(reinterpret_cast<uptr>(&addr), addr);
+#endif
+  return addr;
+}
+
 uptr GetPageSize() {
 // Android post-M sysconf(_SC_PAGESIZE) crashes if called from .preinit_array.
 #if SANITIZER_ANDROID
@@ -900,8 +1035,17 @@ uptr GetPageSize() {
 }
 
 uptr ReadBinaryName(/*out*/char *buf, uptr buf_len) {
+#if SANITIZER_SOLARIS
+  const char *default_module_name = getexecname();
+  CHECK_NE(default_module_name, NULL);
+  return internal_snprintf(buf, buf_len, "%s", default_module_name);
+#else
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
 #if SANITIZER_FREEBSD
-  const int Mib[] = { CTL_KERN, KERN_PROC, KERN_PROC_PATHNAME, -1 };
+  const int Mib[4] = {CTL_KERN, KERN_PROC, KERN_PROC_PATHNAME, -1};
+#else
+  const int Mib[4] = {CTL_KERN, KERN_PROC_ARGS, -1, KERN_PROC_PATHNAME};
+#endif
   const char *default_module_name = "kern.proc.pathname";
   size_t Size = buf_len;
   bool IsErr = (sysctl(Mib, ARRAY_SIZE(Mib), buf, &Size, NULL, 0) != 0);
@@ -913,7 +1057,7 @@ uptr ReadBinaryName(/*out*/char *buf, uptr buf_len) {
       default_module_name, buf, buf_len);
   int readlink_error;
   bool IsErr = internal_iserror(module_name_len, &readlink_error);
-#endif
+#endif  // SANITIZER_SOLARIS
   if (IsErr) {
     // We can't read binary name for some reason, assume it's unknown.
     Report("WARNING: reading executable name failed with errno %d, "
@@ -923,6 +1067,7 @@ uptr ReadBinaryName(/*out*/char *buf, uptr buf_len) {
     CHECK_LT(module_name_len, buf_len);
   }
   return module_name_len;
+#endif
 }
 
 uptr ReadLongProcessName(/*out*/ char *buf, uptr buf_len) {
@@ -1539,12 +1684,17 @@ static bool Aarch64GetESR(ucontext_t *ucontext, u64 *esr) {
 }
 #endif
 
-SignalContext::WriteFlag SignalContext::GetWriteFlag(void *context) {
+SignalContext::WriteFlag SignalContext::GetWriteFlag() const {
   ucontext_t *ucontext = (ucontext_t *)context;
 #if defined(__x86_64__) || defined(__i386__)
   static const uptr PF_WRITE = 1U << 1;
 #if SANITIZER_FREEBSD
   uptr err = ucontext->uc_mcontext.mc_err;
+#elif SANITIZER_NETBSD
+  uptr err = ucontext->uc_mcontext.__gregs[_REG_ERR];
+#elif SANITIZER_SOLARIS && defined(__i386__)
+# define ERR 13
+  uptr err = ucontext->uc_mcontext.gregs[ERR];
 #else
   uptr err = ucontext->uc_mcontext.gregs[REG_ERR];
 #endif
@@ -1558,6 +1708,12 @@ SignalContext::WriteFlag SignalContext::GetWriteFlag(void *context) {
   u64 esr;
   if (!Aarch64GetESR(ucontext, &esr)) return UNKNOWN;
   return esr & ESR_ELx_WNR ? WRITE : READ;
+#elif SANITIZER_SOLARIS && defined(__sparc__)
+  // Decode the instruction to determine the access type.
+  // From OpenSolaris $SRC/uts/sun4/os/trap.c (get_accesstype).
+  uptr pc = ucontext->uc_mcontext.gregs[REG_PC];
+  u32 instr = *(u32 *)pc;
+  return (instr >> 21) & 1 ? WRITE: READ;
 #else
   (void)ucontext;
   return UNKNOWN;  // FIXME: Implement.
@@ -1568,7 +1724,7 @@ void SignalContext::DumpAllRegisters(void *context) {
   // FIXME: Implement this.
 }
 
-void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
+static void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
 #if defined(__arm__)
   ucontext_t *ucontext = (ucontext_t*)context;
   *pc = ucontext->uc_mcontext.arm_pc;
@@ -1591,6 +1747,11 @@ void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
   *pc = ucontext->uc_mcontext.mc_rip;
   *bp = ucontext->uc_mcontext.mc_rbp;
   *sp = ucontext->uc_mcontext.mc_rsp;
+#elif SANITIZER_NETBSD
+  ucontext_t *ucontext = (ucontext_t *)context;
+  *pc = ucontext->uc_mcontext.__gregs[_REG_RIP];
+  *bp = ucontext->uc_mcontext.__gregs[_REG_RBP];
+  *sp = ucontext->uc_mcontext.__gregs[_REG_RSP];
 # else
   ucontext_t *ucontext = (ucontext_t*)context;
   *pc = ucontext->uc_mcontext.gregs[REG_RIP];
@@ -1603,8 +1764,26 @@ void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
   *pc = ucontext->uc_mcontext.mc_eip;
   *bp = ucontext->uc_mcontext.mc_ebp;
   *sp = ucontext->uc_mcontext.mc_esp;
+#elif SANITIZER_NETBSD
+  ucontext_t *ucontext = (ucontext_t *)context;
+  *pc = ucontext->uc_mcontext.__gregs[_REG_EIP];
+  *bp = ucontext->uc_mcontext.__gregs[_REG_EBP];
+  *sp = ucontext->uc_mcontext.__gregs[_REG_ESP];
 # else
   ucontext_t *ucontext = (ucontext_t*)context;
+# if SANITIZER_SOLARIS
+  /* Use the numeric values: the symbolic ones are undefined by llvm
+     include/llvm/Support/Solaris.h.  */
+# ifndef REG_EIP
+#  define REG_EIP 14 // REG_PC
+# endif
+# ifndef REG_EBP
+#  define REG_EBP  6 // REG_FP
+# endif
+# ifndef REG_ESP
+#  define REG_ESP 17 // REG_SP
+# endif
+# endif
   *pc = ucontext->uc_mcontext.gregs[REG_EIP];
   *bp = ucontext->uc_mcontext.gregs[REG_EBP];
   *sp = ucontext->uc_mcontext.gregs[REG_ESP];
@@ -1619,7 +1798,16 @@ void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
 #elif defined(__sparc__)
   ucontext_t *ucontext = (ucontext_t*)context;
   uptr *stk_ptr;
-# if defined (__arch64__)
+# if defined (__sparcv9)
+# ifndef MC_PC
+#  define MC_PC REG_PC
+# endif
+# ifndef MC_O6
+#  define MC_O6 REG_O6
+# endif
+# ifdef SANITIZER_SOLARIS
+#  define mc_gregs gregs
+# endif
   *pc = ucontext->uc_mcontext.mc_gregs[MC_PC];
   *sp = ucontext->uc_mcontext.mc_gregs[MC_O6];
   stk_ptr = (uptr *) (*sp + 2047);
@@ -1649,6 +1837,8 @@ void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
 #endif
 }
 
+void SignalContext::InitPcSpBp() { GetPcSpBp(context, &pc, &sp, &bp); }
+
 void MaybeReexec() {
   // No need to re-exec on Linux.
 }
@@ -1676,25 +1866,27 @@ uptr FindAvailableMemoryRange(uptr size, uptr alignment, uptr left_padding,
   return 0;
 }
 
-bool GetRandom(void *buffer, uptr length) {
+bool GetRandom(void *buffer, uptr length, bool blocking) {
   if (!buffer || !length || length > 256)
     return false;
-#if defined(__NR_getrandom)
+#if SANITIZER_USE_GETRANDOM
   static atomic_uint8_t skip_getrandom_syscall;
   if (!atomic_load_relaxed(&skip_getrandom_syscall)) {
     // Up to 256 bytes, getrandom will not be interrupted.
-    uptr res = internal_syscall(SYSCALL(getrandom), buffer, length, 0);
+    uptr res = internal_syscall(SYSCALL(getrandom), buffer, length,
+                                blocking ? 0 : GRND_NONBLOCK);
     int rverrno = 0;
     if (internal_iserror(res, &rverrno) && rverrno == ENOSYS)
       atomic_store_relaxed(&skip_getrandom_syscall, 1);
     else if (res == length)
       return true;
   }
-#endif
+#endif  // SANITIZER_USE_GETRANDOM
+  // Up to 256 bytes, a read off /dev/urandom will not be interrupted.
+  // blocking is moot here, O_NONBLOCK has no effect when opening /dev/urandom.
   uptr fd = internal_open("/dev/urandom", O_RDONLY);
   if (internal_iserror(fd))
     return false;
-  // internal_read deals with EINTR.
   uptr res = internal_read(fd, buffer, length);
   if (internal_iserror(res))
     return false;
@@ -1704,4 +1896,6 @@ bool GetRandom(void *buffer, uptr length) {
 
 } // namespace __sanitizer
 
-#endif // SANITIZER_FREEBSD || SANITIZER_LINUX
+#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD ||
+        // SANITIZER_SOLARIS
+
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux.h
index 11cad6b80933..2d227f82a99e 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux.h
@@ -14,11 +14,14 @@
 #define SANITIZER_LINUX_H
 
 #include "sanitizer_platform.h"
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
 #include "sanitizer_common.h"
 #include "sanitizer_internal_defs.h"
-#include "sanitizer_posix.h"
+#include "sanitizer_platform_limits_netbsd.h"
 #include "sanitizer_platform_limits_posix.h"
+#include "sanitizer_platform_limits_solaris.h"
+#include "sanitizer_posix.h"
 
 struct link_map;  // Opaque type returned by dlopen().
 
@@ -27,11 +30,25 @@ namespace __sanitizer {
 // the one in <dirent.h>, which is used by readdir().
 struct linux_dirent;
 
+struct ProcSelfMapsBuff {
+  char *data;
+  uptr mmaped_size;
+  uptr len;
+};
+
+struct MemoryMappingLayoutData {
+  ProcSelfMapsBuff proc_self_maps;
+  const char *current;
+};
+
+void ReadProcMaps(ProcSelfMapsBuff *proc_maps);
+
 // Syscall wrappers.
 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);
+uptr internal_clock_gettime(__sanitizer_clockid_t clk_id, void *tp);
 
 // Linux-only syscalls.
 #if SANITIZER_LINUX
@@ -128,5 +145,6 @@ ALWAYS_INLINE uptr *get_android_tls_ptr() {
 
 }  // namespace __sanitizer
 
-#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX
+#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD ||
+        // SANITIZER_SOLARIS
 #endif  // SANITIZER_LINUX_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux_libcdep.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux_libcdep.cc
index 52196db12731..56fdfc8705f3 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux_libcdep.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_linux_libcdep.cc
@@ -14,11 +14,13 @@
 
 #include "sanitizer_platform.h"
 
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
 
 #include "sanitizer_allocator_internal.h"
 #include "sanitizer_atomic.h"
 #include "sanitizer_common.h"
+#include "sanitizer_file.h"
 #include "sanitizer_flags.h"
 #include "sanitizer_freebsd.h"
 #include "sanitizer_linux.h"
@@ -36,19 +38,35 @@
 #if SANITIZER_FREEBSD
 #include <pthread_np.h>
 #include <osreldate.h>
+#include <sys/sysctl.h>
 #define pthread_getattr_np pthread_attr_get_np
 #endif
 
+#if SANITIZER_NETBSD
+#include <sys/sysctl.h>
+#include <sys/tls.h>
+#endif
+
+#if SANITIZER_SOLARIS
+#include <thread.h>
+#endif
+
 #if SANITIZER_LINUX
 #include <sys/prctl.h>
 #endif
 
 #if SANITIZER_ANDROID
 #include <android/api-level.h>
+#if !defined(CPU_COUNT) && !defined(__aarch64__)
+#include <dirent.h>
+#include <fcntl.h>
+struct __sanitizer::linux_dirent {
+  long           d_ino;
+  off_t          d_off;
+  unsigned short d_reclen;
+  char           d_name[];
+};
 #endif
-
-#if SANITIZER_ANDROID && __ANDROID_API__ < 21
-#include <android/log.h>
 #endif
 
 #if !SANITIZER_ANDROID
@@ -102,13 +120,20 @@ void GetThreadStackTopAndBottom(bool at_initialization, uptr *stack_top,
     *stack_bottom = segment.end - stacksize;
     return;
   }
+  uptr stacksize = 0;
+  void *stackaddr = nullptr;
+#if SANITIZER_SOLARIS
+  stack_t ss;
+  CHECK_EQ(thr_stksegment(&ss), 0);
+  stacksize = ss.ss_size;
+  stackaddr = (char *)ss.ss_sp - stacksize;
+#else // !SANITIZER_SOLARIS
   pthread_attr_t attr;
   pthread_attr_init(&attr);
   CHECK_EQ(pthread_getattr_np(pthread_self(), &attr), 0);
-  uptr stacksize = 0;
-  void *stackaddr = nullptr;
   my_pthread_attr_getstack(&attr, &stackaddr, &stacksize);
   pthread_attr_destroy(&attr);
+#endif // SANITIZER_SOLARIS
 
   *stack_top = (uptr)stackaddr + stacksize;
   *stack_bottom = (uptr)stackaddr;
@@ -148,7 +173,8 @@ bool SanitizerGetThreadName(char *name, int max_len) {
 #endif
 }
 
-#if !SANITIZER_FREEBSD && !SANITIZER_ANDROID && !SANITIZER_GO
+#if !SANITIZER_FREEBSD && !SANITIZER_ANDROID && !SANITIZER_GO && \
+    !SANITIZER_NETBSD && !SANITIZER_SOLARIS
 static uptr g_tls_size;
 
 #ifdef __i386__
@@ -176,7 +202,8 @@ void InitTlsSize() {
 }
 #else
 void InitTlsSize() { }
-#endif  // !SANITIZER_FREEBSD && !SANITIZER_ANDROID && !SANITIZER_GO
+#endif  // !SANITIZER_FREEBSD && !SANITIZER_ANDROID && !SANITIZER_GO &&
+        // !SANITIZER_NETBSD && !SANITIZER_SOLARIS
 
 #if (defined(__x86_64__) || defined(__i386__) || defined(__mips__) \
     || defined(__aarch64__) || defined(__powerpc64__) || defined(__s390__) \
@@ -323,7 +350,7 @@ static void **ThreadSelfSegbase() {
   // sysarch(AMD64_GET_FSBASE, segbase);
   __asm __volatile("movq %%fs:0, %0" : "=r" (segbase));
 # else
-#  error "unsupported CPU arch for FreeBSD platform"
+#  error "unsupported CPU arch"
 # endif
   return segbase;
 }
@@ -333,6 +360,35 @@ uptr ThreadSelf() {
 }
 #endif  // SANITIZER_FREEBSD
 
+#if SANITIZER_NETBSD
+static struct tls_tcb * ThreadSelfTlsTcb() {
+  struct tls_tcb * tcb;
+# ifdef __HAVE___LWP_GETTCB_FAST
+  tcb = (struct tls_tcb *)__lwp_gettcb_fast();
+# elif defined(__HAVE___LWP_GETPRIVATE_FAST)
+  tcb = (struct tls_tcb *)__lwp_getprivate_fast();
+# endif
+  return tcb;
+}
+
+uptr ThreadSelf() {
+  return (uptr)ThreadSelfTlsTcb()->tcb_pthread;
+}
+
+int GetSizeFromHdr(struct dl_phdr_info *info, size_t size, void *data) {
+  const Elf_Phdr *hdr = info->dlpi_phdr;
+  const Elf_Phdr *last_hdr = hdr + info->dlpi_phnum;
+
+  for (; hdr != last_hdr; ++hdr) {
+    if (hdr->p_type == PT_TLS && info->dlpi_tls_modid == 1) {
+      *(uptr*)data = hdr->p_memsz;
+      break;
+    }
+  }
+  return 0;
+}
+#endif  // SANITIZER_NETBSD
+
 #if !SANITIZER_GO
 static void GetTls(uptr *addr, uptr *size) {
 #if SANITIZER_LINUX && !SANITIZER_ANDROID
@@ -362,9 +418,27 @@ static void GetTls(uptr *addr, uptr *size) {
     *addr = (uptr) dtv[2];
     *size = (*addr == 0) ? 0 : ((uptr) segbase[0] - (uptr) dtv[2]);
   }
+#elif SANITIZER_NETBSD
+  struct tls_tcb * const tcb = ThreadSelfTlsTcb();
+  *addr = 0;
+  *size = 0;
+  if (tcb != 0) {
+    // Find size (p_memsz) of dlpi_tls_modid 1 (TLS block of the main program).
+    // ld.elf_so hardcodes the index 1.
+    dl_iterate_phdr(GetSizeFromHdr, size);
+
+    if (*size != 0) {
+      // The block has been found and tcb_dtv[1] contains the base address
+      *addr = (uptr)tcb->tcb_dtv[1];
+    }
+  }
 #elif SANITIZER_ANDROID
   *addr = 0;
   *size = 0;
+#elif SANITIZER_SOLARIS
+  // FIXME
+  *addr = 0;
+  *size = 0;
 #else
 # error "Unknown OS"
 #endif
@@ -373,7 +447,8 @@ static void GetTls(uptr *addr, uptr *size) {
 
 #if !SANITIZER_GO
 uptr GetTlsSize() {
-#if SANITIZER_FREEBSD || SANITIZER_ANDROID
+#if SANITIZER_FREEBSD || SANITIZER_ANDROID || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
   uptr addr, size;
   GetTls(&addr, &size);
   return size;
@@ -419,7 +494,7 @@ typedef ElfW(Phdr) Elf_Phdr;
 # endif
 
 struct DlIteratePhdrData {
-  InternalMmapVector<LoadedModule> *modules;
+  InternalMmapVectorNoCtor<LoadedModule> *modules;
   bool first;
 };
 
@@ -457,21 +532,41 @@ extern "C" __attribute__((weak)) int dl_iterate_phdr(
     int (*)(struct dl_phdr_info *, size_t, void *), void *);
 #endif
 
-void ListOfModules::init() {
-  clear();
+static bool requiresProcmaps() {
 #if SANITIZER_ANDROID && __ANDROID_API__ <= 22
-  u32 api_level = AndroidGetApiLevel();
   // Fall back to /proc/maps if dl_iterate_phdr is unavailable or broken.
   // The runtime check allows the same library to work with
   // both K and L (and future) Android releases.
-  if (api_level <= ANDROID_LOLLIPOP_MR1) { // L or earlier
-    MemoryMappingLayout memory_mapping(false);
-    memory_mapping.DumpListOfModules(&modules_);
-    return;
-  }
+  return AndroidGetApiLevel() <= ANDROID_LOLLIPOP_MR1;
+#else
+  return false;
 #endif
-  DlIteratePhdrData data = {&modules_, true};
-  dl_iterate_phdr(dl_iterate_phdr_cb, &data);
+}
+
+static void procmapsInit(InternalMmapVectorNoCtor<LoadedModule> *modules) {
+  MemoryMappingLayout memory_mapping(/*cache_enabled*/true);
+  memory_mapping.DumpListOfModules(modules);
+}
+
+void ListOfModules::init() {
+  clearOrInit();
+  if (requiresProcmaps()) {
+    procmapsInit(&modules_);
+  } else {
+    DlIteratePhdrData data = {&modules_, true};
+    dl_iterate_phdr(dl_iterate_phdr_cb, &data);
+  }
+}
+
+// When a custom loader is used, dl_iterate_phdr may not contain the full
+// list of modules. Allow callers to fall back to using procmaps.
+void ListOfModules::fallbackInit() {
+  if (!requiresProcmaps()) {
+    clearOrInit();
+    procmapsInit(&modules_);
+  } else {
+    clear();
+  }
 }
 
 // getrusage does not give us the current RSS, only the max RSS.
@@ -513,12 +608,65 @@ uptr GetRSS() {
   return rss * GetPageSizeCached();
 }
 
-// 64-bit Android targets don't provide the deprecated __android_log_write.
-// Starting with the L release, syslog() works and is preferable to
-// __android_log_write.
+// sysconf(_SC_NPROCESSORS_{CONF,ONLN}) cannot be used on most platforms as
+// they allocate memory.
+u32 GetNumberOfCPUs() {
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
+  u32 ncpu;
+  int req[2];
+  size_t len = sizeof(ncpu);
+  req[0] = CTL_HW;
+  req[1] = HW_NCPU;
+  CHECK_EQ(sysctl(req, 2, &ncpu, &len, NULL, 0), 0);
+  return ncpu;
+#elif SANITIZER_ANDROID && !defined(CPU_COUNT) && !defined(__aarch64__)
+  // Fall back to /sys/devices/system/cpu on Android when cpu_set_t doesn't
+  // exist in sched.h. That is the case for toolchains generated with older
+  // NDKs.
+  // This code doesn't work on AArch64 because internal_getdents makes use of
+  // the 64bit getdents syscall, but cpu_set_t seems to always exist on AArch64.
+  uptr fd = internal_open("/sys/devices/system/cpu", O_RDONLY | O_DIRECTORY);
+  if (internal_iserror(fd))
+    return 0;
+  InternalScopedBuffer<u8> buffer(4096);
+  uptr bytes_read = buffer.size();
+  uptr n_cpus = 0;
+  u8 *d_type;
+  struct linux_dirent *entry = (struct linux_dirent *)&buffer[bytes_read];
+  while (true) {
+    if ((u8 *)entry >= &buffer[bytes_read]) {
+      bytes_read = internal_getdents(fd, (struct linux_dirent *)buffer.data(),
+                                     buffer.size());
+      if (internal_iserror(bytes_read) || !bytes_read)
+        break;
+      entry = (struct linux_dirent *)buffer.data();
+    }
+    d_type = (u8 *)entry + entry->d_reclen - 1;
+    if (d_type >= &buffer[bytes_read] ||
+        (u8 *)&entry->d_name[3] >= &buffer[bytes_read])
+      break;
+    if (entry->d_ino != 0 && *d_type == DT_DIR) {
+      if (entry->d_name[0] == 'c' && entry->d_name[1] == 'p' &&
+          entry->d_name[2] == 'u' &&
+          entry->d_name[3] >= '0' && entry->d_name[3] <= '9')
+        n_cpus++;
+    }
+    entry = (struct linux_dirent *)(((u8 *)entry) + entry->d_reclen);
+  }
+  internal_close(fd);
+  return n_cpus;
+#elif SANITIZER_SOLARIS
+  return sysconf(_SC_NPROCESSORS_ONLN);
+#else
+  cpu_set_t CPUs;
+  CHECK_EQ(sched_getaffinity(0, sizeof(cpu_set_t), &CPUs), 0);
+  return CPU_COUNT(&CPUs);
+#endif
+}
+
 #if SANITIZER_LINUX
 
-#if SANITIZER_ANDROID
+# if SANITIZER_ANDROID
 static atomic_uint8_t android_log_initialized;
 
 void AndroidLogInit() {
@@ -529,34 +677,97 @@ void AndroidLogInit() {
 static bool ShouldLogAfterPrintf() {
   return atomic_load(&android_log_initialized, memory_order_acquire);
 }
-#else
-void AndroidLogInit() {}
 
-static bool ShouldLogAfterPrintf() { return true; }
-#endif  // SANITIZER_ANDROID
+extern "C" SANITIZER_WEAK_ATTRIBUTE
+int async_safe_write_log(int pri, const char* tag, const char* msg);
+extern "C" SANITIZER_WEAK_ATTRIBUTE
+int __android_log_write(int prio, const char* tag, const char* msg);
+
+// ANDROID_LOG_INFO is 4, but can't be resolved at runtime.
+#define SANITIZER_ANDROID_LOG_INFO 4
 
+// async_safe_write_log is a new public version of __libc_write_log that is
+// used behind syslog. It is preferable to syslog as it will not do any dynamic
+// memory allocation or formatting.
+// If the function is not available, syslog is preferred for L+ (it was broken
+// pre-L) as __android_log_write triggers a racey behavior with the strncpy
+// interceptor. Fallback to __android_log_write pre-L.
 void WriteOneLineToSyslog(const char *s) {
-#if SANITIZER_ANDROID &&__ANDROID_API__ < 21
-  __android_log_write(ANDROID_LOG_INFO, NULL, s);
-#else
-  syslog(LOG_INFO, "%s", s);
-#endif
+  if (&async_safe_write_log) {
+    async_safe_write_log(SANITIZER_ANDROID_LOG_INFO, GetProcessName(), s);
+  } else if (AndroidGetApiLevel() > ANDROID_KITKAT) {
+    syslog(LOG_INFO, "%s", s);
+  } else {
+    CHECK(&__android_log_write);
+    __android_log_write(SANITIZER_ANDROID_LOG_INFO, nullptr, s);
+  }
 }
 
+extern "C" SANITIZER_WEAK_ATTRIBUTE
+void android_set_abort_message(const char *);
+
+void SetAbortMessage(const char *str) {
+  if (&android_set_abort_message)
+    android_set_abort_message(str);
+}
+# else
+void AndroidLogInit() {}
+
+static bool ShouldLogAfterPrintf() { return true; }
+
+void WriteOneLineToSyslog(const char *s) { syslog(LOG_INFO, "%s", s); }
+
+void SetAbortMessage(const char *str) {}
+# endif  // SANITIZER_ANDROID
+
 void LogMessageOnPrintf(const char *str) {
   if (common_flags()->log_to_syslog && ShouldLogAfterPrintf())
     WriteToSyslog(str);
 }
 
-#if SANITIZER_ANDROID && __ANDROID_API__ >= 21
-extern "C" void android_set_abort_message(const char *msg);
-void SetAbortMessage(const char *str) { android_set_abort_message(str); }
-#else
-void SetAbortMessage(const char *str) {}
-#endif
+#endif  // SANITIZER_LINUX
+
+#if SANITIZER_LINUX && !SANITIZER_GO
+// glibc crashes when using clock_gettime from a preinit_array function as the
+// vDSO function pointers haven't been initialized yet. __progname is
+// initialized after the vDSO function pointers, so if it exists, is not null
+// and is not empty, we can use clock_gettime.
+extern "C" SANITIZER_WEAK_ATTRIBUTE char *__progname;
+INLINE bool CanUseVDSO() {
+  // Bionic is safe, it checks for the vDSO function pointers to be initialized.
+  if (SANITIZER_ANDROID)
+    return true;
+  if (&__progname && __progname && *__progname)
+    return true;
+  return false;
+}
 
-#endif // SANITIZER_LINUX
+// MonotonicNanoTime is a timing function that can leverage the vDSO by calling
+// clock_gettime. real_clock_gettime only exists if clock_gettime is
+// intercepted, so define it weakly and use it if available.
+extern "C" SANITIZER_WEAK_ATTRIBUTE
+int real_clock_gettime(u32 clk_id, void *tp);
+u64 MonotonicNanoTime() {
+  timespec ts;
+  if (CanUseVDSO()) {
+    if (&real_clock_gettime)
+      real_clock_gettime(CLOCK_MONOTONIC, &ts);
+    else
+      clock_gettime(CLOCK_MONOTONIC, &ts);
+  } else {
+    internal_clock_gettime(CLOCK_MONOTONIC, &ts);
+  }
+  return (u64)ts.tv_sec * (1000ULL * 1000 * 1000) + ts.tv_nsec;
+}
+#else
+// Non-Linux & Go always use the syscall.
+u64 MonotonicNanoTime() {
+  timespec ts;
+  internal_clock_gettime(CLOCK_MONOTONIC, &ts);
+  return (u64)ts.tv_sec * (1000ULL * 1000 * 1000) + ts.tv_nsec;
+}
+#endif  // SANITIZER_LINUX && !SANITIZER_GO
 
 } // namespace __sanitizer
 
-#endif // SANITIZER_FREEBSD || SANITIZER_LINUX
+#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac.cc
index 1edd4157fd6b..e1c51f580874 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac.cc
@@ -23,6 +23,7 @@
 #include <stdio.h>
 
 #include "sanitizer_common.h"
+#include "sanitizer_file.h"
 #include "sanitizer_flags.h"
 #include "sanitizer_internal_defs.h"
 #include "sanitizer_libc.h"
@@ -184,7 +185,7 @@ uptr internal_getpid() {
 
 int internal_sigaction(int signum, const void *act, void *oldact) {
   return sigaction(signum,
-                   (struct sigaction *)act, (struct sigaction *)oldact);
+                   (const struct sigaction *)act, (struct sigaction *)oldact);
 }
 
 void internal_sigfillset(__sanitizer_sigset_t *set) { sigfillset(set); }
@@ -364,6 +365,10 @@ u64 NanoTime() {
   return 0;
 }
 
+u64 MonotonicNanoTime() {
+  return 0;
+}
+
 uptr GetTlsSize() {
   return 0;
 }
@@ -410,11 +415,13 @@ void GetThreadStackAndTls(bool main, uptr *stk_addr, uptr *stk_size,
 }
 
 void ListOfModules::init() {
-  clear();
+  clearOrInit();
   MemoryMappingLayout memory_mapping(false);
   memory_mapping.DumpListOfModules(&modules_);
 }
 
+void ListOfModules::fallbackInit() { clear(); }
+
 static HandleSignalMode GetHandleSignalModeImpl(int signum) {
   switch (signum) {
     case SIGABRT:
@@ -573,7 +580,7 @@ void LogFullErrorReport(const char *buffer) {
 #endif
 }
 
-SignalContext::WriteFlag SignalContext::GetWriteFlag(void *context) {
+SignalContext::WriteFlag SignalContext::GetWriteFlag() const {
 #if defined(__x86_64__) || defined(__i386__)
   ucontext_t *ucontext = static_cast<ucontext_t*>(context);
   return ucontext->uc_mcontext->__es.__err & 2 /*T_PF_WRITE*/ ? WRITE : READ;
@@ -582,7 +589,7 @@ SignalContext::WriteFlag SignalContext::GetWriteFlag(void *context) {
 #endif
 }
 
-void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
+static void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
   ucontext_t *ucontext = (ucontext_t*)context;
 # if defined(__aarch64__)
   *pc = ucontext->uc_mcontext->__ss.__pc;
@@ -609,6 +616,8 @@ void GetPcSpBp(void *context, uptr *pc, uptr *sp, uptr *bp) {
 # endif
 }
 
+void SignalContext::InitPcSpBp() { GetPcSpBp(context, &pc, &sp, &bp); }
+
 #if !SANITIZER_GO
 static const char kDyldInsertLibraries[] = "DYLD_INSERT_LIBRARIES";
 LowLevelAllocator allocator_for_env;
@@ -736,6 +745,9 @@ void MaybeReexec() {
   if (!lib_is_in_env)
     return;
 
+  if (!common_flags()->strip_env)
+    return;
+
   // DYLD_INSERT_LIBRARIES is set and contains the runtime library. Let's remove
   // the dylib from the environment variable, because interceptors are installed
   // and we don't want our children to inherit the variable.
@@ -844,7 +856,7 @@ uptr GetTaskInfoMaxAddress() {
 }
 #endif
 
-uptr GetMaxVirtualAddress() {
+uptr GetMaxUserVirtualAddress() {
 #if SANITIZER_WORDSIZE == 64
 # if defined(__aarch64__) && SANITIZER_IOS && !SANITIZER_IOSSIM
   // Get the maximum VM address
@@ -859,6 +871,10 @@ uptr GetMaxVirtualAddress() {
 #endif  // SANITIZER_WORDSIZE
 }
 
+uptr GetMaxVirtualAddress() {
+  return GetMaxUserVirtualAddress();
+}
+
 uptr FindAvailableMemoryRange(uptr shadow_size,
                               uptr alignment,
                               uptr left_padding,
@@ -881,6 +897,11 @@ uptr FindAvailableMemoryRange(uptr shadow_size,
     mach_msg_type_number_t count = kRegionInfoSize;
     kr = mach_vm_region_recurse(mach_task_self(), &address, &vmsize, &depth,
                                 (vm_region_info_t)&vminfo, &count);
+    if (kr == KERN_INVALID_ADDRESS) {
+      // No more regions beyond "address", consider the gap at the end of VM.
+      address = GetMaxVirtualAddress() + 1;
+      vmsize = 0;
+    }
     if (free_begin != address) {
       // We found a free region [free_begin..address-1].
       uptr gap_start = RoundUpTo((uptr)free_begin + left_padding, alignment);
@@ -991,7 +1012,12 @@ void CheckNoDeepBind(const char *filename, int flag) {
 }
 
 // FIXME: implement on this platform.
-bool GetRandom(void *buffer, uptr length) {
+bool GetRandom(void *buffer, uptr length, bool blocking) {
+  UNIMPLEMENTED();
+}
+
+// FIXME: implement on this platform.
+u32 GetNumberOfCPUs() {
   UNIMPLEMENTED();
 }
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac.h
index 3f1c68c8610a..e022a2c0363c 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac.h
@@ -20,6 +20,17 @@
 
 namespace __sanitizer {
 
+struct MemoryMappingLayoutData {
+  int current_image;
+  u32 current_magic;
+  u32 current_filetype;
+  ModuleArch current_arch;
+  u8 current_uuid[kModuleUUIDSize];
+  int current_load_cmd_count;
+  const char *current_load_cmd_addr;
+  bool current_instrumented;
+};
+
 enum MacosVersion {
   MACOS_VERSION_UNINITIALIZED = 0,
   MACOS_VERSION_UNKNOWN,
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac_libcdep.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac_libcdep.cc
index c95daa9372ad..090f17c888b4 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac_libcdep.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mac_libcdep.cc
@@ -20,7 +20,7 @@
 namespace __sanitizer {
 
 void RestrictMemoryToMaxAddress(uptr max_address) {
-  uptr size_to_mmap = GetMaxVirtualAddress() + 1 - max_address;
+  uptr size_to_mmap = GetMaxUserVirtualAddress() + 1 - max_address;
   void *res = MmapFixedNoAccess(max_address, size_to_mmap, "high gap");
   CHECK(res != MAP_FAILED);
 }
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mutex.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mutex.h
index 1759bf13f689..4bb878ee19dc 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mutex.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_mutex.h
@@ -92,8 +92,10 @@ class BlockingMutex {
   // checks that the mutex is owned, and assumes callers to be generally
   // well-behaved.
   void CheckLocked();
+
  private:
-  uptr opaque_storage_[10];
+  // Solaris mutex_t has a member that requires 64-bit alignment.
+  ALIGNED(8) uptr opaque_storage_[10];
   uptr owner_;  // for debugging
 };
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform.h
index 396f7c9346d6..334903c26f9f 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform.h
@@ -14,7 +14,8 @@
 #define SANITIZER_PLATFORM_H
 
 #if !defined(__linux__) && !defined(__FreeBSD__) && !defined(__NetBSD__) && \
-  !defined(__APPLE__) && !defined(_WIN32)
+  !defined(__APPLE__) && !defined(_WIN32) && !defined(__Fuchsia__) && \
+  !(defined(__sun__) && defined(__svr4__))
 # error "This operating system is not supported"
 #endif
 
@@ -36,6 +37,12 @@
 # define SANITIZER_NETBSD 0
 #endif
 
+#if defined(__sun__) && defined(__svr4__)
+# define SANITIZER_SOLARIS 1
+#else
+# define SANITIZER_SOLARIS 0
+#endif
+
 #if defined(__APPLE__)
 # define SANITIZER_MAC     1
 # include <TargetConditionals.h>
@@ -44,7 +51,7 @@
 # else
 #  define SANITIZER_IOS    0
 # endif
-# if TARGET_IPHONE_SIMULATOR
+# if TARGET_OS_SIMULATOR
 #  define SANITIZER_IOSSIM 1
 # else
 #  define SANITIZER_IOSSIM 0
@@ -85,8 +92,15 @@
 # define SANITIZER_ANDROID 0
 #endif
 
+#if defined(__Fuchsia__)
+# define SANITIZER_FUCHSIA 1
+#else
+# define SANITIZER_FUCHSIA 0
+#endif
+
 #define SANITIZER_POSIX \
-  (SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_MAC || SANITIZER_NETBSD)
+  (SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_MAC || \
+    SANITIZER_NETBSD || SANITIZER_SOLARIS)
 
 #if __LP64__ || defined(_WIN64)
 #  define SANITIZER_WORDSIZE 64
@@ -175,13 +189,19 @@
 # define SANITIZER_ARM 0
 #endif
 
+#if SANITIZER_SOLARIS && SANITIZER_WORDSIZE == 32
+# define SANITIZER_SOLARIS32 1
+#else
+# define SANITIZER_SOLARIS32 0
+#endif
+
 // By default we allow to use SizeClassAllocator64 on 64-bit platform.
 // But in some cases (e.g. AArch64's 39-bit address space) SizeClassAllocator64
 // does not work well and we need to fallback to SizeClassAllocator32.
 // For such platforms build this code with -DSANITIZER_CAN_USE_ALLOCATOR64=0 or
 // change the definition of SANITIZER_CAN_USE_ALLOCATOR64 here.
 #ifndef SANITIZER_CAN_USE_ALLOCATOR64
-# if SANITIZER_ANDROID && defined(__aarch64__)
+# if (SANITIZER_ANDROID && defined(__aarch64__)) || SANITIZER_FUCHSIA
 #  define SANITIZER_CAN_USE_ALLOCATOR64 1
 # elif defined(__mips64) || defined(__aarch64__)
 #  define SANITIZER_CAN_USE_ALLOCATOR64 0
@@ -276,5 +296,10 @@
 # define SANITIZER_SUPPRESS_LEAK_ON_PTHREAD_EXIT 0
 #endif
 
+#if SANITIZER_FREEBSD || SANITIZER_MAC || SANITIZER_NETBSD || SANITIZER_SOLARIS
+# define SANITIZER_MADVISE_DONTNEED MADV_FREE
+#else
+# define SANITIZER_MADVISE_DONTNEED MADV_DONTNEED
+#endif
 
 #endif // SANITIZER_PLATFORM_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h
index 0380cee92a00..7e33fa59b312 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h
@@ -16,19 +16,26 @@
 
 #include "sanitizer_internal_defs.h"
 
+#if SANITIZER_POSIX
+# define SI_POSIX 1
+#else
+# define SI_POSIX 0
+#endif
+
 #if !SANITIZER_WINDOWS
 # define SI_WINDOWS 0
-# define SI_NOT_WINDOWS 1
-# include "sanitizer_platform_limits_posix.h"
 #else
 # define SI_WINDOWS 1
-# define SI_NOT_WINDOWS 0
 #endif
 
-#if SANITIZER_POSIX
-# define SI_POSIX 1
-#else
-# define SI_POSIX 0
+#if (SI_POSIX != 0) == (SI_WINDOWS != 0) && !SANITIZER_FUCHSIA
+# error "Windows is not POSIX!"
+#endif
+
+#if SI_POSIX
+# include "sanitizer_platform_limits_netbsd.h"
+# include "sanitizer_platform_limits_posix.h"
+# include "sanitizer_platform_limits_solaris.h"
 #endif
 
 #if SANITIZER_LINUX && !SANITIZER_ANDROID
@@ -75,10 +82,28 @@
 # define SI_IOS 0
 #endif
 
-#if !SANITIZER_WINDOWS && !SANITIZER_MAC
-# define SI_UNIX_NOT_MAC 1
+#if SANITIZER_FUCHSIA
+# define SI_NOT_FUCHSIA 0
+#else
+# define SI_NOT_FUCHSIA 1
+#endif
+
+#if SANITIZER_SOLARIS
+# define SI_SOLARIS 1
 #else
-# define SI_UNIX_NOT_MAC 0
+# define SI_SOLARIS 0
+#endif
+
+#if SANITIZER_SOLARIS32
+# define SI_SOLARIS32 1
+#else
+# define SI_SOLARIS32 0
+#endif
+
+#if SANITIZER_POSIX && !SANITIZER_MAC
+# define SI_POSIX_NOT_MAC 1
+#else
+# define SI_POSIX_NOT_MAC 0
 #endif
 
 #if SANITIZER_LINUX && !SANITIZER_FREEBSD
@@ -87,23 +112,23 @@
 # define SI_LINUX_NOT_FREEBSD 0
 #endif
 
-#define SANITIZER_INTERCEPT_STRLEN 1
-#define SANITIZER_INTERCEPT_STRNLEN SI_NOT_MAC
-#define SANITIZER_INTERCEPT_STRCMP 1
-#define SANITIZER_INTERCEPT_STRSTR 1
-#define SANITIZER_INTERCEPT_STRCASESTR SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_STRTOK 1
-#define SANITIZER_INTERCEPT_STRCHR 1
-#define SANITIZER_INTERCEPT_STRCHRNUL SI_UNIX_NOT_MAC
-#define SANITIZER_INTERCEPT_STRRCHR 1
-#define SANITIZER_INTERCEPT_STRSPN 1
-#define SANITIZER_INTERCEPT_STRPBRK 1
-#define SANITIZER_INTERCEPT_TEXTDOMAIN SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_STRCASECMP SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_STRLEN SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_STRNLEN (SI_NOT_MAC && SI_NOT_FUCHSIA)
+#define SANITIZER_INTERCEPT_STRCMP SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_STRSTR SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_STRCASESTR SI_POSIX
+#define SANITIZER_INTERCEPT_STRTOK SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_STRCHR SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_STRCHRNUL SI_POSIX_NOT_MAC
+#define SANITIZER_INTERCEPT_STRRCHR SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_STRSPN SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_STRPBRK SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_TEXTDOMAIN SI_LINUX_NOT_ANDROID || SI_SOLARIS
+#define SANITIZER_INTERCEPT_STRCASECMP SI_POSIX
 #define SANITIZER_INTERCEPT_MEMSET 1
 #define SANITIZER_INTERCEPT_MEMMOVE 1
 #define SANITIZER_INTERCEPT_MEMCPY 1
-#define SANITIZER_INTERCEPT_MEMCMP 1
+#define SANITIZER_INTERCEPT_MEMCMP SI_NOT_FUCHSIA
 #define SANITIZER_INTERCEPT_STRNDUP SI_POSIX
 #define SANITIZER_INTERCEPT___STRNDUP SI_LINUX_NOT_FREEBSD
 #if defined(__ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__) && \
@@ -114,24 +139,23 @@
 #endif
 // memmem on Darwin doesn't exist on 10.6
 // FIXME: enable memmem on Windows.
-#define SANITIZER_INTERCEPT_MEMMEM \
-  (SI_NOT_WINDOWS && !SI_MAC_DEPLOYMENT_BELOW_10_7)
-#define SANITIZER_INTERCEPT_MEMCHR 1
+#define SANITIZER_INTERCEPT_MEMMEM (SI_POSIX && !SI_MAC_DEPLOYMENT_BELOW_10_7)
+#define SANITIZER_INTERCEPT_MEMCHR SI_NOT_FUCHSIA
 #define SANITIZER_INTERCEPT_MEMRCHR (SI_FREEBSD || SI_LINUX || SI_NETBSD)
 
-#define SANITIZER_INTERCEPT_READ   SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_PREAD  SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_WRITE  SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_PWRITE SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_READ SI_POSIX
+#define SANITIZER_INTERCEPT_PREAD SI_POSIX
+#define SANITIZER_INTERCEPT_WRITE SI_POSIX
+#define SANITIZER_INTERCEPT_PWRITE SI_POSIX
 
-#define SANITIZER_INTERCEPT_FREAD SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_FWRITE SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_FREAD SI_POSIX
+#define SANITIZER_INTERCEPT_FWRITE SI_POSIX
 
-#define SANITIZER_INTERCEPT_PREAD64 SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_PWRITE64 SI_LINUX_NOT_ANDROID
+#define SANITIZER_INTERCEPT_PREAD64 SI_LINUX_NOT_ANDROID || SI_SOLARIS32
+#define SANITIZER_INTERCEPT_PWRITE64 SI_LINUX_NOT_ANDROID || SI_SOLARIS32
 
-#define SANITIZER_INTERCEPT_READV SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_WRITEV SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_READV SI_POSIX
+#define SANITIZER_INTERCEPT_WRITEV SI_POSIX
 
 #define SANITIZER_INTERCEPT_PREADV \
   (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
@@ -141,58 +165,68 @@
 
 #define SANITIZER_INTERCEPT_PRCTL   SI_LINUX
 
-#define SANITIZER_INTERCEPT_LOCALTIME_AND_FRIENDS SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_STRPTIME SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_LOCALTIME_AND_FRIENDS SI_POSIX
+#define SANITIZER_INTERCEPT_STRPTIME SI_POSIX
 
-#define SANITIZER_INTERCEPT_SCANF SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_SCANF SI_POSIX
 #define SANITIZER_INTERCEPT_ISOC99_SCANF SI_LINUX_NOT_ANDROID
 
 #ifndef SANITIZER_INTERCEPT_PRINTF
-# define SANITIZER_INTERCEPT_PRINTF SI_NOT_WINDOWS
+# define SANITIZER_INTERCEPT_PRINTF SI_POSIX
 # define SANITIZER_INTERCEPT_PRINTF_L (SI_FREEBSD || SI_NETBSD)
 # define SANITIZER_INTERCEPT_ISOC99_PRINTF SI_LINUX_NOT_ANDROID
 #endif
 
-#define SANITIZER_INTERCEPT_FREXP 1
-#define SANITIZER_INTERCEPT_FREXPF_FREXPL SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT___PRINTF_CHK \
+  (SANITIZER_INTERCEPT_PRINTF && SI_LINUX_NOT_ANDROID)
+
+#define SANITIZER_INTERCEPT_FREXP SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_FREXPF_FREXPL SI_POSIX
 
-#define SANITIZER_INTERCEPT_GETPWNAM_AND_FRIENDS SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_GETPWNAM_AND_FRIENDS SI_POSIX
 #define SANITIZER_INTERCEPT_GETPWNAM_R_AND_FRIENDS \
-  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_GETPWENT \
-  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_FGETPWENT SI_LINUX_NOT_ANDROID
+  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_FGETPWENT SI_LINUX_NOT_ANDROID || SI_SOLARIS
 #define SANITIZER_INTERCEPT_GETPWENT_R \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_SETPWENT (SI_MAC || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_CLOCK_GETTIME (SI_FREEBSD || SI_NETBSD || SI_LINUX)
-#define SANITIZER_INTERCEPT_GETITIMER SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_TIME SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_GLOB SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_WAIT SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_INET SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_PTHREAD_GETSCHEDPARAM SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_GETADDRINFO SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_GETNAMEINFO SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_GETSOCKNAME SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_GETHOSTBYNAME SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_GETHOSTBYNAME_R (SI_FREEBSD || SI_LINUX)
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_SETPWENT \
+  (SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_CLOCK_GETTIME \
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_GETITIMER SI_POSIX
+#define SANITIZER_INTERCEPT_TIME SI_POSIX
+#define SANITIZER_INTERCEPT_GLOB SI_LINUX_NOT_ANDROID || SI_SOLARIS
+#define SANITIZER_INTERCEPT_GLOB64 SI_LINUX_NOT_ANDROID
+#define SANITIZER_INTERCEPT_WAIT SI_POSIX
+#define SANITIZER_INTERCEPT_INET SI_POSIX
+#define SANITIZER_INTERCEPT_PTHREAD_GETSCHEDPARAM SI_POSIX
+#define SANITIZER_INTERCEPT_GETADDRINFO SI_POSIX
+#define SANITIZER_INTERCEPT_GETNAMEINFO SI_POSIX
+#define SANITIZER_INTERCEPT_GETSOCKNAME SI_POSIX
+#define SANITIZER_INTERCEPT_GETHOSTBYNAME SI_POSIX
+#define SANITIZER_INTERCEPT_GETHOSTBYNAME2 SI_POSIX && !SI_SOLARIS
+#define SANITIZER_INTERCEPT_GETHOSTBYNAME_R \
+  (SI_FREEBSD || SI_LINUX || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_GETHOSTBYNAME2_R \
   (SI_FREEBSD || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_GETHOSTBYADDR_R (SI_FREEBSD || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_GETHOSTENT_R (SI_FREEBSD || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_GETSOCKOPT SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_ACCEPT SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_GETHOSTBYADDR_R \
+  (SI_FREEBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_GETHOSTENT_R \
+  (SI_FREEBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_GETSOCKOPT SI_POSIX
+#define SANITIZER_INTERCEPT_ACCEPT SI_POSIX
 #define SANITIZER_INTERCEPT_ACCEPT4 SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_MODF SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_RECVMSG SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_SENDMSG SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_GETPEERNAME SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_IOCTL SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_INET_ATON SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_MODF SI_POSIX
+#define SANITIZER_INTERCEPT_RECVMSG SI_POSIX
+#define SANITIZER_INTERCEPT_SENDMSG SI_POSIX
+#define SANITIZER_INTERCEPT_GETPEERNAME SI_POSIX
+#define SANITIZER_INTERCEPT_IOCTL SI_POSIX
+#define SANITIZER_INTERCEPT_INET_ATON SI_POSIX
 #define SANITIZER_INTERCEPT_SYSINFO SI_LINUX
-#define SANITIZER_INTERCEPT_READDIR SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_READDIR64 SI_LINUX_NOT_ANDROID
+#define SANITIZER_INTERCEPT_READDIR SI_POSIX
+#define SANITIZER_INTERCEPT_READDIR64 SI_LINUX_NOT_ANDROID || SI_SOLARIS32
 #if SI_LINUX_NOT_ANDROID && \
   (defined(__i386) || defined(__x86_64) || defined(__mips64) || \
     defined(__powerpc64__) || defined(__aarch64__) || defined(__arm__) || \
@@ -201,116 +235,125 @@
 #else
 #define SANITIZER_INTERCEPT_PTRACE 0
 #endif
-#define SANITIZER_INTERCEPT_SETLOCALE SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_GETCWD SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_SETLOCALE SI_POSIX
+#define SANITIZER_INTERCEPT_GETCWD SI_POSIX
 #define SANITIZER_INTERCEPT_GET_CURRENT_DIR_NAME SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_STRTOIMAX SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_MBSTOWCS SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_MBSNRTOWCS (SI_MAC || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_WCSTOMBS SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_STRTOIMAX SI_POSIX
+#define SANITIZER_INTERCEPT_MBSTOWCS SI_POSIX
+#define SANITIZER_INTERCEPT_MBSNRTOWCS \
+  (SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_WCSTOMBS SI_POSIX
 #define SANITIZER_INTERCEPT_WCSNRTOMBS \
-  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_WCRTOMB \
-  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_TCGETATTR SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_REALPATH SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_CANONICALIZE_FILE_NAME SI_LINUX_NOT_ANDROID
+  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_TCGETATTR SI_LINUX_NOT_ANDROID || SI_SOLARIS
+#define SANITIZER_INTERCEPT_REALPATH SI_POSIX
+#define SANITIZER_INTERCEPT_CANONICALIZE_FILE_NAME \
+  (SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_CONFSTR \
-  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_SCHED_GETAFFINITY SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_SCHED_GETPARAM SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_STRERROR SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_STRERROR_R SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_SCHED_GETPARAM SI_LINUX_NOT_ANDROID || SI_SOLARIS
+#define SANITIZER_INTERCEPT_STRERROR SI_POSIX
+#define SANITIZER_INTERCEPT_STRERROR_R SI_POSIX
 #define SANITIZER_INTERCEPT_XPG_STRERROR_R SI_LINUX_NOT_ANDROID
 #define SANITIZER_INTERCEPT_SCANDIR \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_SCANDIR64 SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_GETGROUPS SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_POLL SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_PPOLL SI_LINUX_NOT_ANDROID
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_SCANDIR64 SI_LINUX_NOT_ANDROID || SI_SOLARIS32
+#define SANITIZER_INTERCEPT_GETGROUPS SI_POSIX
+#define SANITIZER_INTERCEPT_POLL SI_POSIX
+#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)
-#define SANITIZER_INTERCEPT_SIGWAIT SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_SIGWAITINFO SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_SIGTIMEDWAIT SI_LINUX_NOT_ANDROID
+  (SI_FREEBSD || SI_NETBSD || (SI_MAC && !SI_IOS) || SI_LINUX_NOT_ANDROID || \
+    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
 #define SANITIZER_INTERCEPT_SIGSETOPS \
-  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_SIGPENDING SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_SIGPROCMASK SI_NOT_WINDOWS
+  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_SIGPENDING SI_POSIX
+#define SANITIZER_INTERCEPT_SIGPROCMASK SI_POSIX
 #define SANITIZER_INTERCEPT_BACKTRACE \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_GETMNTENT SI_LINUX
 #define SANITIZER_INTERCEPT_GETMNTENT_R SI_LINUX_NOT_ANDROID
 #define SANITIZER_INTERCEPT_STATFS \
-  (SI_FREEBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_STATFS64 \
   ((SI_MAC && !SI_IOS) || SI_LINUX_NOT_ANDROID)
 #define SANITIZER_INTERCEPT_STATVFS \
   (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
 #define SANITIZER_INTERCEPT_STATVFS64 SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_INITGROUPS SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_ETHER_NTOA_ATON SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_INITGROUPS SI_POSIX
+#define SANITIZER_INTERCEPT_ETHER_NTOA_ATON SI_POSIX
 #define SANITIZER_INTERCEPT_ETHER_HOST \
   (SI_FREEBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
 #define SANITIZER_INTERCEPT_ETHER_R (SI_FREEBSD || SI_LINUX_NOT_ANDROID)
 #define SANITIZER_INTERCEPT_SHMCTL                       \
-  (SI_NETBSD || ((SI_FREEBSD || SI_LINUX_NOT_ANDROID) && \
+  (SI_NETBSD || SI_SOLARIS || ((SI_FREEBSD || SI_LINUX_NOT_ANDROID) && \
                  SANITIZER_WORDSIZE == 64))  // NOLINT
 #define SANITIZER_INTERCEPT_RANDOM_R SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_PTHREAD_ATTR_GET SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_PTHREAD_ATTR_GET SI_POSIX
 #define SANITIZER_INTERCEPT_PTHREAD_ATTR_GETINHERITSCHED \
-  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_PTHREAD_ATTR_GETAFFINITY_NP SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_PTHREAD_MUTEXATTR_GETPSHARED SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_PTHREAD_MUTEXATTR_GETTYPE SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_PTHREAD_MUTEXATTR_GETPSHARED \
+  (SI_POSIX && !SI_NETBSD)
+#define SANITIZER_INTERCEPT_PTHREAD_MUTEXATTR_GETTYPE SI_POSIX
 #define SANITIZER_INTERCEPT_PTHREAD_MUTEXATTR_GETPROTOCOL \
-  (SI_MAC || SI_NETBSD || SI_LINUX_NOT_ANDROID)
+  (SI_MAC || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_PTHREAD_MUTEXATTR_GETPRIOCEILING \
-  (SI_MAC || SI_NETBSD || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_PTHREAD_MUTEXATTR_GETROBUST SI_LINUX_NOT_ANDROID
+  (SI_MAC || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_PTHREAD_MUTEXATTR_GETROBUST \
+  (SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_PTHREAD_MUTEXATTR_GETROBUST_NP SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_PTHREAD_RWLOCKATTR_GETPSHARED SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_PTHREAD_RWLOCKATTR_GETPSHARED \
+  (SI_POSIX && !SI_NETBSD)
 #define SANITIZER_INTERCEPT_PTHREAD_RWLOCKATTR_GETKIND_NP SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_PTHREAD_CONDATTR_GETPSHARED SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_PTHREAD_CONDATTR_GETCLOCK SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_PTHREAD_BARRIERATTR_GETPSHARED SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_TMPNAM SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_TMPNAM_R SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_TTYNAME_R SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_TEMPNAM SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_SINCOS SI_LINUX
-#define SANITIZER_INTERCEPT_REMQUO SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_LGAMMA SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_LGAMMA_R (SI_FREEBSD || SI_LINUX)
-#define SANITIZER_INTERCEPT_LGAMMAL_R SI_LINUX_NOT_ANDROID
+#define SANITIZER_INTERCEPT_PTHREAD_CONDATTR_GETPSHARED \
+  (SI_POSIX && !SI_NETBSD)
+#define SANITIZER_INTERCEPT_PTHREAD_CONDATTR_GETCLOCK \
+  (SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_PTHREAD_BARRIERATTR_GETPSHARED \
+  (SI_LINUX_NOT_ANDROID && !SI_NETBSD)
+#define SANITIZER_INTERCEPT_TMPNAM SI_POSIX
+#define SANITIZER_INTERCEPT_TMPNAM_R SI_LINUX_NOT_ANDROID || SI_SOLARIS
+#define SANITIZER_INTERCEPT_TTYNAME_R SI_POSIX
+#define SANITIZER_INTERCEPT_TEMPNAM SI_POSIX
+#define SANITIZER_INTERCEPT_SINCOS SI_LINUX || SI_SOLARIS
+#define SANITIZER_INTERCEPT_REMQUO SI_POSIX
+#define SANITIZER_INTERCEPT_LGAMMA SI_POSIX
+#define SANITIZER_INTERCEPT_LGAMMA_R (SI_FREEBSD || SI_LINUX || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_LGAMMAL_R SI_LINUX_NOT_ANDROID || SI_SOLARIS
 #define SANITIZER_INTERCEPT_DRAND48_R SI_LINUX_NOT_ANDROID
 #define SANITIZER_INTERCEPT_RAND_R \
-  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_ICONV \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
-#define SANITIZER_INTERCEPT_TIMES SI_NOT_WINDOWS
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_TIMES SI_POSIX
 
 // FIXME: getline seems to be available on OSX 10.7
 #define SANITIZER_INTERCEPT_GETLINE \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 
 #define SANITIZER_INTERCEPT__EXIT \
-  (SI_LINUX || SI_FREEBSD || SI_NETBSD || SI_MAC)
+  (SI_LINUX || SI_FREEBSD || SI_NETBSD || SI_MAC || SI_SOLARIS)
 
-#define SANITIZER_INTERCEPT_PHTREAD_MUTEX SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_PHTREAD_MUTEX SI_POSIX
 #define SANITIZER_INTERCEPT_PTHREAD_SETNAME_NP \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 
 #define SANITIZER_INTERCEPT_TLS_GET_ADDR \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 
 #define SANITIZER_INTERCEPT_LISTXATTR SI_LINUX
 #define SANITIZER_INTERCEPT_GETXATTR SI_LINUX
 #define SANITIZER_INTERCEPT_GETRESID SI_LINUX
 #define SANITIZER_INTERCEPT_GETIFADDRS \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_MAC)
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_MAC || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_IF_INDEXTONAME \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_MAC)
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_MAC || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_CAPGET SI_LINUX_NOT_ANDROID
 #if SI_LINUX && defined(__arm__)
 #define SANITIZER_INTERCEPT_AEABI_MEM 1
@@ -318,64 +361,71 @@
 #define SANITIZER_INTERCEPT_AEABI_MEM 0
 #endif
 #define SANITIZER_INTERCEPT___BZERO SI_MAC
-#define SANITIZER_INTERCEPT_FTIME (!SI_FREEBSD && !SI_NETBSD && SI_NOT_WINDOWS)
-#define SANITIZER_INTERCEPT_XDR SI_LINUX_NOT_ANDROID
+#define SANITIZER_INTERCEPT_FTIME (!SI_FREEBSD && !SI_NETBSD && SI_POSIX)
+#define SANITIZER_INTERCEPT_XDR SI_LINUX_NOT_ANDROID || SI_SOLARIS
 #define SANITIZER_INTERCEPT_TSEARCH \
-  (SI_LINUX_NOT_ANDROID || SI_MAC || SI_NETBSD)
+  (SI_LINUX_NOT_ANDROID || SI_MAC || SI_NETBSD || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_LIBIO_INTERNALS SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_FOPEN SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_FOPEN64 SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_OPEN_MEMSTREAM (SI_LINUX_NOT_ANDROID || SI_NETBSD)
+#define SANITIZER_INTERCEPT_FOPEN SI_POSIX
+#define SANITIZER_INTERCEPT_FOPEN64 SI_LINUX_NOT_ANDROID || SI_SOLARIS32
+#define SANITIZER_INTERCEPT_OPEN_MEMSTREAM \
+  (SI_LINUX_NOT_ANDROID || SI_NETBSD || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_OBSTACK SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_FFLUSH SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_FCLOSE SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_FFLUSH SI_POSIX
+#define SANITIZER_INTERCEPT_FCLOSE SI_POSIX
 
 #ifndef SANITIZER_INTERCEPT_DLOPEN_DLCLOSE
 #define SANITIZER_INTERCEPT_DLOPEN_DLCLOSE \
-  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_MAC)
+  (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID || SI_MAC || SI_SOLARIS)
 #endif
 
 #define SANITIZER_INTERCEPT_GETPASS \
   (SI_LINUX_NOT_ANDROID || SI_MAC || SI_NETBSD)
 #define SANITIZER_INTERCEPT_TIMERFD SI_LINUX_NOT_ANDROID
 
-#define SANITIZER_INTERCEPT_MLOCKX SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_MLOCKX SI_POSIX
 #define SANITIZER_INTERCEPT_FOPENCOOKIE SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_SEM (SI_LINUX || SI_FREEBSD || SI_NETBSD)
-#define SANITIZER_INTERCEPT_PTHREAD_SETCANCEL SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_MINCORE (SI_LINUX || SI_NETBSD)
+#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_PROCESS_VM_READV SI_LINUX
 #define SANITIZER_INTERCEPT_CTERMID \
-  (SI_LINUX || SI_MAC || SI_FREEBSD || SI_NETBSD)
-#define SANITIZER_INTERCEPT_CTERMID_R (SI_MAC || SI_FREEBSD)
+  (SI_LINUX || SI_MAC || SI_FREEBSD || SI_NETBSD || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_CTERMID_R (SI_MAC || SI_FREEBSD || SI_SOLARIS)
 
 #define SANITIZER_INTERCEPTOR_HOOKS (SI_LINUX || SI_MAC || SI_WINDOWS)
-#define SANITIZER_INTERCEPT_RECV_RECVFROM SI_NOT_WINDOWS
-#define SANITIZER_INTERCEPT_SEND_SENDTO SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_RECV_RECVFROM SI_POSIX
+#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)
-#define SANITIZER_INTERCEPT___XSTAT \
-  (!SANITIZER_INTERCEPT_STAT && SI_NOT_WINDOWS)
+  (SI_FREEBSD || SI_MAC || SI_ANDROID || SI_NETBSD || SI_SOLARIS)
+#define SANITIZER_INTERCEPT___XSTAT (!SANITIZER_INTERCEPT_STAT && SI_POSIX)
 #define SANITIZER_INTERCEPT___XSTAT64 SI_LINUX_NOT_ANDROID
 #define SANITIZER_INTERCEPT___LXSTAT SANITIZER_INTERCEPT___XSTAT
 #define SANITIZER_INTERCEPT___LXSTAT64 SI_LINUX_NOT_ANDROID
 
-#define SANITIZER_INTERCEPT_UTMP (SI_NOT_WINDOWS && !SI_MAC && !SI_FREEBSD)
-#define SANITIZER_INTERCEPT_UTMPX (SI_LINUX_NOT_ANDROID || SI_MAC || SI_FREEBSD)
+#define SANITIZER_INTERCEPT_UTMP \
+  (SI_POSIX && !SI_MAC && !SI_FREEBSD && !SI_NETBSD)
+#define SANITIZER_INTERCEPT_UTMPX \
+  (SI_LINUX_NOT_ANDROID || SI_MAC || SI_FREEBSD || SI_NETBSD)
 
 #define SANITIZER_INTERCEPT_GETLOADAVG \
   (SI_LINUX_NOT_ANDROID || SI_MAC || SI_FREEBSD || SI_NETBSD)
 
 #define SANITIZER_INTERCEPT_MALLOPT_AND_MALLINFO \
-  (!SI_FREEBSD && !SI_MAC && !SI_NETBSD)
+  (!SI_FREEBSD && !SI_MAC && !SI_NETBSD && SI_NOT_FUCHSIA)
 #define SANITIZER_INTERCEPT_MEMALIGN (!SI_FREEBSD && !SI_MAC && !SI_NETBSD)
-#define SANITIZER_INTERCEPT_PVALLOC (!SI_FREEBSD && !SI_MAC && !SI_NETBSD)
-#define SANITIZER_INTERCEPT_CFREE (!SI_FREEBSD && !SI_MAC && !SI_NETBSD)
+#define SANITIZER_INTERCEPT_PVALLOC \
+  (!SI_FREEBSD && !SI_MAC && !SI_NETBSD && SI_NOT_FUCHSIA)
+#define SANITIZER_INTERCEPT_CFREE \
+  (!SI_FREEBSD && !SI_MAC && !SI_NETBSD && SI_NOT_FUCHSIA)
 #define SANITIZER_INTERCEPT_ALIGNED_ALLOC (!SI_MAC)
 #define SANITIZER_INTERCEPT_MALLOC_USABLE_SIZE (!SI_MAC)
 #define SANITIZER_INTERCEPT_MCHECK_MPROBE SI_LINUX_NOT_ANDROID
-#define SANITIZER_INTERCEPT_WCSCAT SI_NOT_WINDOWS
+#define SANITIZER_INTERCEPT_WCSCAT SI_POSIX
+#define SANITIZER_INTERCEPT_SIGNAL_AND_SIGACTION (!SI_WINDOWS && SI_NOT_FUCHSIA)
+#define SANITIZER_INTERCEPT_BSD_SIGNAL SI_ANDROID
 
 #endif  // #ifndef SANITIZER_PLATFORM_INTERCEPTORS_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.cc
new file mode 100644
index 000000000000..108e196e7a05
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.cc
@@ -0,0 +1,363 @@
+//===-- sanitizer_platform_limits_netbsd.cc -------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of Sanitizer common code.
+//
+// Sizes and layouts of platform-specific NetBSD data structures.
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_platform.h"
+
+#if SANITIZER_NETBSD
+#include <arpa/inet.h>
+#include <dirent.h>
+#include <glob.h>
+#include <grp.h>
+#include <ifaddrs.h>
+#include <limits.h>
+#include <link_elf.h>
+#include <net/if.h>
+#include <net/if_ether.h>
+#include <net/ppp_defs.h>
+#include <net/route.h>
+#include <netdb.h>
+#include <netinet/in.h>
+#include <netinet/ip_mroute.h>
+#include <poll.h>
+#include <pthread.h>
+#include <pwd.h>
+#include <semaphore.h>
+#include <signal.h>
+#include <stddef.h>
+#include <sys/filio.h>
+#include <sys/ipc.h>
+#include <sys/mman.h>
+#include <sys/mount.h>
+#include <sys/mqueue.h>
+#include <sys/msg.h>
+#include <sys/mtio.h>
+#include <sys/ptrace.h>
+#include <sys/resource.h>
+#include <sys/shm.h>
+#include <sys/signal.h>
+#include <sys/socket.h>
+#include <sys/sockio.h>
+#include <sys/soundcard.h>
+#include <sys/stat.h>
+#include <sys/statvfs.h>
+#include <sys/time.h>
+#include <sys/timeb.h>
+#include <sys/times.h>
+#include <sys/timespec.h>
+#include <sys/timex.h>
+#include <sys/types.h>
+#include <sys/ucontext.h>
+#include <sys/utsname.h>
+#include <term.h>
+#include <termios.h>
+#include <time.h>
+#include <utime.h>
+#include <utmp.h>
+#include <utmpx.h>
+#include <wchar.h>
+#include <wordexp.h>
+
+// Include these after system headers to avoid name clashes and ambiguities.
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_platform_limits_netbsd.h"
+
+namespace __sanitizer {
+unsigned struct_utsname_sz = sizeof(struct utsname);
+unsigned struct_stat_sz = sizeof(struct stat);
+unsigned struct_rusage_sz = sizeof(struct rusage);
+unsigned struct_tm_sz = sizeof(struct tm);
+unsigned struct_passwd_sz = sizeof(struct passwd);
+unsigned struct_group_sz = sizeof(struct group);
+unsigned siginfo_t_sz = sizeof(siginfo_t);
+unsigned struct_sigaction_sz = sizeof(struct sigaction);
+unsigned struct_itimerval_sz = sizeof(struct itimerval);
+unsigned pthread_t_sz = sizeof(pthread_t);
+unsigned pthread_mutex_t_sz = sizeof(pthread_mutex_t);
+unsigned pthread_cond_t_sz = sizeof(pthread_cond_t);
+unsigned pid_t_sz = sizeof(pid_t);
+unsigned timeval_sz = sizeof(timeval);
+unsigned uid_t_sz = sizeof(uid_t);
+unsigned gid_t_sz = sizeof(gid_t);
+unsigned mbstate_t_sz = sizeof(mbstate_t);
+unsigned sigset_t_sz = sizeof(sigset_t);
+unsigned struct_timezone_sz = sizeof(struct timezone);
+unsigned struct_tms_sz = sizeof(struct tms);
+unsigned struct_sigevent_sz = sizeof(struct sigevent);
+unsigned struct_sched_param_sz = sizeof(struct sched_param);
+unsigned struct_sockaddr_sz = sizeof(struct sockaddr);
+unsigned ucontext_t_sz = sizeof(ucontext_t);
+unsigned struct_rlimit_sz = sizeof(struct rlimit);
+unsigned struct_timespec_sz = sizeof(struct timespec);
+unsigned struct_utimbuf_sz = sizeof(struct utimbuf);
+unsigned struct_itimerspec_sz = sizeof(struct itimerspec);
+unsigned struct_timex_sz = sizeof(struct timex);
+unsigned struct_msqid_ds_sz = sizeof(struct msqid_ds);
+unsigned struct_mq_attr_sz = sizeof(struct mq_attr);
+unsigned struct_statvfs_sz = sizeof(struct statvfs);
+
+const uptr sig_ign = (uptr)SIG_IGN;
+const uptr sig_dfl = (uptr)SIG_DFL;
+const uptr sig_err = (uptr)SIG_ERR;
+const uptr sa_siginfo = (uptr)SA_SIGINFO;
+
+int shmctl_ipc_stat = (int)IPC_STAT;
+
+unsigned struct_utmp_sz = sizeof(struct utmp);
+unsigned struct_utmpx_sz = sizeof(struct utmpx);
+
+int map_fixed = MAP_FIXED;
+
+int af_inet = (int)AF_INET;
+int af_inet6 = (int)AF_INET6;
+
+uptr __sanitizer_in_addr_sz(int af) {
+  if (af == AF_INET)
+    return sizeof(struct in_addr);
+  else if (af == AF_INET6)
+    return sizeof(struct in6_addr);
+  else
+    return 0;
+}
+
+unsigned struct_ElfW_Phdr_sz = sizeof(Elf_Phdr);
+
+int glob_nomatch = GLOB_NOMATCH;
+int glob_altdirfunc = GLOB_ALTDIRFUNC;
+
+unsigned path_max = PATH_MAX;
+
+// ioctl arguments
+unsigned struct_ifreq_sz = sizeof(struct ifreq);
+unsigned struct_termios_sz = sizeof(struct termios);
+unsigned struct_winsize_sz = sizeof(struct winsize);
+unsigned struct_mtget_sz = sizeof(struct mtget);
+unsigned struct_mtop_sz = sizeof(struct mtop);
+unsigned struct_audio_buf_info_sz = sizeof(struct audio_buf_info);
+unsigned struct_ppp_stats_sz = sizeof(struct ppp_stats);
+unsigned struct_sioc_sg_req_sz = sizeof(struct sioc_sg_req);
+unsigned struct_sioc_vif_req_sz = sizeof(struct sioc_vif_req);
+
+const unsigned IOCTL_NOT_PRESENT = 0;
+
+unsigned IOCTL_FIOASYNC = FIOASYNC;
+unsigned IOCTL_FIOCLEX = FIOCLEX;
+unsigned IOCTL_FIOGETOWN = FIOGETOWN;
+unsigned IOCTL_FIONBIO = FIONBIO;
+unsigned IOCTL_FIONCLEX = FIONCLEX;
+unsigned IOCTL_FIOSETOWN = FIOSETOWN;
+unsigned IOCTL_SIOCADDMULTI = SIOCADDMULTI;
+unsigned IOCTL_SIOCATMARK = SIOCATMARK;
+unsigned IOCTL_SIOCDELMULTI = SIOCDELMULTI;
+unsigned IOCTL_SIOCGIFADDR = SIOCGIFADDR;
+unsigned IOCTL_SIOCGIFBRDADDR = SIOCGIFBRDADDR;
+unsigned IOCTL_SIOCGIFCONF = SIOCGIFCONF;
+unsigned IOCTL_SIOCGIFDSTADDR = SIOCGIFDSTADDR;
+unsigned IOCTL_SIOCGIFFLAGS = SIOCGIFFLAGS;
+unsigned IOCTL_SIOCGIFMETRIC = SIOCGIFMETRIC;
+unsigned IOCTL_SIOCGIFMTU = SIOCGIFMTU;
+unsigned IOCTL_SIOCGIFNETMASK = SIOCGIFNETMASK;
+unsigned IOCTL_SIOCGPGRP = SIOCGPGRP;
+unsigned IOCTL_SIOCSIFADDR = SIOCSIFADDR;
+unsigned IOCTL_SIOCSIFBRDADDR = SIOCSIFBRDADDR;
+unsigned IOCTL_SIOCSIFDSTADDR = SIOCSIFDSTADDR;
+unsigned IOCTL_SIOCSIFFLAGS = SIOCSIFFLAGS;
+unsigned IOCTL_SIOCSIFMETRIC = SIOCSIFMETRIC;
+unsigned IOCTL_SIOCSIFMTU = SIOCSIFMTU;
+unsigned IOCTL_SIOCSIFNETMASK = SIOCSIFNETMASK;
+unsigned IOCTL_SIOCSPGRP = SIOCSPGRP;
+unsigned IOCTL_TIOCCONS = TIOCCONS;
+unsigned IOCTL_TIOCEXCL = TIOCEXCL;
+unsigned IOCTL_TIOCGETD = TIOCGETD;
+unsigned IOCTL_TIOCGPGRP = TIOCGPGRP;
+unsigned IOCTL_TIOCGWINSZ = TIOCGWINSZ;
+unsigned IOCTL_TIOCMBIC = TIOCMBIC;
+unsigned IOCTL_TIOCMBIS = TIOCMBIS;
+unsigned IOCTL_TIOCMGET = TIOCMGET;
+unsigned IOCTL_TIOCMSET = TIOCMSET;
+unsigned IOCTL_TIOCNOTTY = TIOCNOTTY;
+unsigned IOCTL_TIOCNXCL = TIOCNXCL;
+unsigned IOCTL_TIOCOUTQ = TIOCOUTQ;
+unsigned IOCTL_TIOCPKT = TIOCPKT;
+unsigned IOCTL_TIOCSCTTY = TIOCSCTTY;
+unsigned IOCTL_TIOCSETD = TIOCSETD;
+unsigned IOCTL_TIOCSPGRP = TIOCSPGRP;
+unsigned IOCTL_TIOCSTI = TIOCSTI;
+unsigned IOCTL_TIOCSWINSZ = TIOCSWINSZ;
+unsigned IOCTL_SIOCGETSGCNT = SIOCGETSGCNT;
+unsigned IOCTL_SIOCGETVIFCNT = SIOCGETVIFCNT;
+
+const int si_SEGV_MAPERR = SEGV_MAPERR;
+const int si_SEGV_ACCERR = SEGV_ACCERR;
+}  // namespace __sanitizer
+
+using namespace __sanitizer;
+
+COMPILER_CHECK(sizeof(__sanitizer_pthread_attr_t) >= sizeof(pthread_attr_t));
+
+COMPILER_CHECK(sizeof(socklen_t) == sizeof(unsigned));
+CHECK_TYPE_SIZE(pthread_key_t);
+
+// There are more undocumented fields in dl_phdr_info that we are not interested
+// in.
+COMPILER_CHECK(sizeof(__sanitizer_dl_phdr_info) <= sizeof(dl_phdr_info));
+CHECK_SIZE_AND_OFFSET(dl_phdr_info, dlpi_addr);
+CHECK_SIZE_AND_OFFSET(dl_phdr_info, dlpi_name);
+CHECK_SIZE_AND_OFFSET(dl_phdr_info, dlpi_phdr);
+CHECK_SIZE_AND_OFFSET(dl_phdr_info, dlpi_phnum);
+
+CHECK_TYPE_SIZE(glob_t);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_pathc);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_pathv);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_offs);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_flags);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_closedir);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_readdir);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_opendir);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_lstat);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_stat);
+
+CHECK_TYPE_SIZE(addrinfo);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_flags);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_family);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_socktype);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_protocol);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_protocol);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_addrlen);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_canonname);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_addr);
+
+CHECK_TYPE_SIZE(hostent);
+CHECK_SIZE_AND_OFFSET(hostent, h_name);
+CHECK_SIZE_AND_OFFSET(hostent, h_aliases);
+CHECK_SIZE_AND_OFFSET(hostent, h_addrtype);
+CHECK_SIZE_AND_OFFSET(hostent, h_length);
+CHECK_SIZE_AND_OFFSET(hostent, h_addr_list);
+
+CHECK_TYPE_SIZE(iovec);
+CHECK_SIZE_AND_OFFSET(iovec, iov_base);
+CHECK_SIZE_AND_OFFSET(iovec, iov_len);
+
+CHECK_TYPE_SIZE(msghdr);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_name);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_namelen);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_iov);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_iovlen);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_control);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_controllen);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_flags);
+
+CHECK_TYPE_SIZE(cmsghdr);
+CHECK_SIZE_AND_OFFSET(cmsghdr, cmsg_len);
+CHECK_SIZE_AND_OFFSET(cmsghdr, cmsg_level);
+CHECK_SIZE_AND_OFFSET(cmsghdr, cmsg_type);
+
+COMPILER_CHECK(sizeof(__sanitizer_dirent) <= sizeof(dirent));
+CHECK_SIZE_AND_OFFSET(dirent, d_fileno);
+CHECK_SIZE_AND_OFFSET(dirent, d_reclen);
+
+CHECK_TYPE_SIZE(ifconf);
+CHECK_SIZE_AND_OFFSET(ifconf, ifc_len);
+CHECK_SIZE_AND_OFFSET(ifconf, ifc_ifcu);
+
+CHECK_TYPE_SIZE(pollfd);
+CHECK_SIZE_AND_OFFSET(pollfd, fd);
+CHECK_SIZE_AND_OFFSET(pollfd, events);
+CHECK_SIZE_AND_OFFSET(pollfd, revents);
+
+CHECK_TYPE_SIZE(nfds_t);
+
+CHECK_TYPE_SIZE(sigset_t);
+
+COMPILER_CHECK(sizeof(__sanitizer_sigaction) == sizeof(struct sigaction));
+// Can't write checks for sa_handler and sa_sigaction due to them being
+// preprocessor macros.
+CHECK_STRUCT_SIZE_AND_OFFSET(sigaction, sa_mask);
+
+CHECK_TYPE_SIZE(wordexp_t);
+CHECK_SIZE_AND_OFFSET(wordexp_t, we_wordc);
+CHECK_SIZE_AND_OFFSET(wordexp_t, we_wordv);
+CHECK_SIZE_AND_OFFSET(wordexp_t, we_offs);
+
+CHECK_TYPE_SIZE(tm);
+CHECK_SIZE_AND_OFFSET(tm, tm_sec);
+CHECK_SIZE_AND_OFFSET(tm, tm_min);
+CHECK_SIZE_AND_OFFSET(tm, tm_hour);
+CHECK_SIZE_AND_OFFSET(tm, tm_mday);
+CHECK_SIZE_AND_OFFSET(tm, tm_mon);
+CHECK_SIZE_AND_OFFSET(tm, tm_year);
+CHECK_SIZE_AND_OFFSET(tm, tm_wday);
+CHECK_SIZE_AND_OFFSET(tm, tm_yday);
+CHECK_SIZE_AND_OFFSET(tm, tm_isdst);
+CHECK_SIZE_AND_OFFSET(tm, tm_gmtoff);
+CHECK_SIZE_AND_OFFSET(tm, tm_zone);
+
+CHECK_TYPE_SIZE(ether_addr);
+
+CHECK_TYPE_SIZE(ipc_perm);
+CHECK_SIZE_AND_OFFSET(ipc_perm, _key);
+CHECK_SIZE_AND_OFFSET(ipc_perm, _seq);
+CHECK_SIZE_AND_OFFSET(ipc_perm, uid);
+CHECK_SIZE_AND_OFFSET(ipc_perm, gid);
+CHECK_SIZE_AND_OFFSET(ipc_perm, cuid);
+CHECK_SIZE_AND_OFFSET(ipc_perm, cgid);
+CHECK_SIZE_AND_OFFSET(ipc_perm, mode);
+
+CHECK_TYPE_SIZE(shmid_ds);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_perm);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_segsz);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_atime);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_dtime);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_ctime);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_cpid);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_lpid);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_nattch);
+
+CHECK_TYPE_SIZE(clock_t);
+
+CHECK_TYPE_SIZE(ifaddrs);
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_next);
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_name);
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_addr);
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_netmask);
+// Compare against the union, because we can't reach into the union in a
+// compliant way.
+#ifdef ifa_dstaddr
+#undef ifa_dstaddr
+#endif
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_dstaddr);
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_data);
+
+CHECK_TYPE_SIZE(timeb);
+CHECK_SIZE_AND_OFFSET(timeb, time);
+CHECK_SIZE_AND_OFFSET(timeb, millitm);
+CHECK_SIZE_AND_OFFSET(timeb, timezone);
+CHECK_SIZE_AND_OFFSET(timeb, dstflag);
+
+CHECK_TYPE_SIZE(passwd);
+CHECK_SIZE_AND_OFFSET(passwd, pw_name);
+CHECK_SIZE_AND_OFFSET(passwd, pw_passwd);
+CHECK_SIZE_AND_OFFSET(passwd, pw_uid);
+CHECK_SIZE_AND_OFFSET(passwd, pw_gid);
+CHECK_SIZE_AND_OFFSET(passwd, pw_dir);
+CHECK_SIZE_AND_OFFSET(passwd, pw_shell);
+
+CHECK_SIZE_AND_OFFSET(passwd, pw_gecos);
+
+CHECK_TYPE_SIZE(group);
+CHECK_SIZE_AND_OFFSET(group, gr_name);
+CHECK_SIZE_AND_OFFSET(group, gr_passwd);
+CHECK_SIZE_AND_OFFSET(group, gr_gid);
+CHECK_SIZE_AND_OFFSET(group, gr_mem);
+
+#endif  // SANITIZER_NETBSD
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.h
new file mode 100644
index 000000000000..6823004d7b41
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_netbsd.h
@@ -0,0 +1,582 @@
+//===-- sanitizer_platform_limits_netbsd.h --------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of Sanitizer common code.
+//
+// Sizes and layouts of platform-specific NetBSD data structures.
+//===----------------------------------------------------------------------===//
+
+#ifndef SANITIZER_PLATFORM_LIMITS_NETBSD_H
+#define SANITIZER_PLATFORM_LIMITS_NETBSD_H
+
+#if SANITIZER_NETBSD
+
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_platform.h"
+
+#define _GET_LINK_MAP_BY_DLOPEN_HANDLE(handle, shift) \
+  ((link_map *)((handle) == nullptr ? nullptr : ((char *)(handle) + (shift))))
+
+#if defined(__x86_64__)
+#define GET_LINK_MAP_BY_DLOPEN_HANDLE(handle) \
+  _GET_LINK_MAP_BY_DLOPEN_HANDLE(handle, 312)
+#elif defined(__i386__)
+#define GET_LINK_MAP_BY_DLOPEN_HANDLE(handle) \
+  _GET_LINK_MAP_BY_DLOPEN_HANDLE(handle, 164)
+#endif
+
+namespace __sanitizer {
+extern unsigned struct_utsname_sz;
+extern unsigned struct_stat_sz;
+extern unsigned struct_rusage_sz;
+extern unsigned siginfo_t_sz;
+extern unsigned struct_itimerval_sz;
+extern unsigned pthread_t_sz;
+extern unsigned pthread_mutex_t_sz;
+extern unsigned pthread_cond_t_sz;
+extern unsigned pid_t_sz;
+extern unsigned timeval_sz;
+extern unsigned uid_t_sz;
+extern unsigned gid_t_sz;
+extern unsigned mbstate_t_sz;
+extern unsigned struct_timezone_sz;
+extern unsigned struct_tms_sz;
+extern unsigned struct_itimerspec_sz;
+extern unsigned struct_sigevent_sz;
+extern unsigned struct_sched_param_sz;
+extern unsigned struct_statfs_sz;
+extern unsigned struct_sockaddr_sz;
+extern unsigned ucontext_t_sz;
+
+extern unsigned struct_rlimit_sz;
+extern unsigned struct_utimbuf_sz;
+extern unsigned struct_timespec_sz;
+
+struct __sanitizer_iocb {
+  u64 aio_offset;
+  uptr aio_buf;
+  long aio_nbytes;
+  u32 aio_fildes;
+  u32 aio_lio_opcode;
+  long aio_reqprio;
+#if SANITIZER_WORDSIZE == 64
+  u8 aio_sigevent[32];
+#else
+  u8 aio_sigevent[20];
+#endif
+  u32 _state;
+  u32 _errno;
+  long _retval;
+};
+
+struct __sanitizer___sysctl_args {
+  int *name;
+  int nlen;
+  void *oldval;
+  uptr *oldlenp;
+  void *newval;
+  uptr newlen;
+};
+
+struct __sanitizer_sem_t {
+  uptr data[5];
+};
+
+struct __sanitizer_ipc_perm {
+  u32 uid;
+  u32 gid;
+  u32 cuid;
+  u32 cgid;
+  u32 mode;
+  unsigned short _seq;
+  long _key;
+};
+
+struct __sanitizer_shmid_ds {
+  __sanitizer_ipc_perm shm_perm;
+  unsigned long shm_segsz;
+  u32 shm_lpid;
+  u32 shm_cpid;
+  unsigned int shm_nattch;
+  u64 shm_atime;
+  u64 shm_dtime;
+  u64 shm_ctime;
+  void *_shm_internal;
+};
+
+extern unsigned struct_msqid_ds_sz;
+extern unsigned struct_mq_attr_sz;
+extern unsigned struct_timex_sz;
+extern unsigned struct_statvfs_sz;
+
+struct __sanitizer_iovec {
+  void *iov_base;
+  uptr iov_len;
+};
+
+struct __sanitizer_ifaddrs {
+  struct __sanitizer_ifaddrs *ifa_next;
+  char *ifa_name;
+  unsigned int ifa_flags;
+  void *ifa_addr;     // (struct sockaddr *)
+  void *ifa_netmask;  // (struct sockaddr *)
+  void *ifa_dstaddr;  // (struct sockaddr *)
+  void *ifa_data;
+  unsigned int ifa_addrflags;
+};
+
+typedef unsigned __sanitizer_pthread_key_t;
+
+typedef long long __sanitizer_time_t;
+
+struct __sanitizer_passwd {
+  char *pw_name;
+  char *pw_passwd;
+  int pw_uid;
+  int pw_gid;
+  __sanitizer_time_t pw_change;
+  char *pw_class;
+  char *pw_gecos;
+  char *pw_dir;
+  char *pw_shell;
+  __sanitizer_time_t pw_expire;
+};
+
+struct __sanitizer_group {
+  char *gr_name;
+  char *gr_passwd;
+  int gr_gid;
+  char **gr_mem;
+};
+
+struct __sanitizer_timeb {
+  __sanitizer_time_t time;
+  unsigned short millitm;
+  short timezone;
+  short dstflag;
+};
+
+struct __sanitizer_ether_addr {
+  u8 octet[6];
+};
+
+struct __sanitizer_tm {
+  int tm_sec;
+  int tm_min;
+  int tm_hour;
+  int tm_mday;
+  int tm_mon;
+  int tm_year;
+  int tm_wday;
+  int tm_yday;
+  int tm_isdst;
+  long int tm_gmtoff;
+  const char *tm_zone;
+};
+
+struct __sanitizer_msghdr {
+  void *msg_name;
+  unsigned msg_namelen;
+  struct __sanitizer_iovec *msg_iov;
+  unsigned msg_iovlen;
+  void *msg_control;
+  unsigned msg_controllen;
+  int msg_flags;
+};
+struct __sanitizer_cmsghdr {
+  unsigned cmsg_len;
+  int cmsg_level;
+  int cmsg_type;
+};
+
+struct __sanitizer_dirent {
+  u64 d_fileno;
+  u16 d_reclen;
+  // more fields that we don't care about
+};
+
+typedef int __sanitizer_clock_t;
+typedef int __sanitizer_clockid_t;
+
+typedef u32 __sanitizer___kernel_uid_t;
+typedef u32 __sanitizer___kernel_gid_t;
+typedef u64 __sanitizer___kernel_off_t;
+typedef struct {
+  u32 fds_bits[8];
+} __sanitizer___kernel_fd_set;
+
+typedef struct {
+  unsigned int pta_magic;
+  int pta_flags;
+  void *pta_private;
+} __sanitizer_pthread_attr_t;
+
+struct __sanitizer_sigset_t {
+  // uint32_t * 4
+  unsigned int __bits[4];
+};
+
+struct __sanitizer_siginfo {
+  // The size is determined by looking at sizeof of real siginfo_t on linux.
+  u64 opaque[128 / sizeof(u64)];
+};
+
+using __sanitizer_sighandler_ptr = void (*)(int sig);
+using __sanitizer_sigactionhandler_ptr = void (*)(int sig,
+                                                  __sanitizer_siginfo *siginfo,
+                                                  void *uctx);
+
+struct __sanitizer_sigaction {
+  union {
+    __sanitizer_sighandler_ptr handler;
+    __sanitizer_sigactionhandler_ptr sigaction;
+  };
+  __sanitizer_sigset_t sa_mask;
+  int sa_flags;
+};
+
+typedef __sanitizer_sigset_t __sanitizer_kernel_sigset_t;
+
+struct __sanitizer_kernel_sigaction_t {
+  union {
+    void (*handler)(int signo);
+    void (*sigaction)(int signo, void *info, void *ctx);
+  };
+  unsigned long sa_flags;
+  void (*sa_restorer)(void);
+  __sanitizer_kernel_sigset_t sa_mask;
+};
+
+extern const uptr sig_ign;
+extern const uptr sig_dfl;
+extern const uptr sig_err;
+extern const uptr sa_siginfo;
+
+extern int af_inet;
+extern int af_inet6;
+uptr __sanitizer_in_addr_sz(int af);
+
+struct __sanitizer_dl_phdr_info {
+  uptr dlpi_addr;
+  const char *dlpi_name;
+  const void *dlpi_phdr;
+  short dlpi_phnum;
+};
+
+extern unsigned struct_ElfW_Phdr_sz;
+
+struct __sanitizer_addrinfo {
+  int ai_flags;
+  int ai_family;
+  int ai_socktype;
+  int ai_protocol;
+  unsigned ai_addrlen;
+  char *ai_canonname;
+  void *ai_addr;
+  struct __sanitizer_addrinfo *ai_next;
+};
+
+struct __sanitizer_hostent {
+  char *h_name;
+  char **h_aliases;
+  int h_addrtype;
+  int h_length;
+  char **h_addr_list;
+};
+
+struct __sanitizer_pollfd {
+  int fd;
+  short events;
+  short revents;
+};
+
+typedef unsigned __sanitizer_nfds_t;
+
+struct __sanitizer_glob_t {
+  uptr gl_pathc;
+  uptr gl_matchc;
+  uptr gl_offs;
+  int gl_flags;
+  char **gl_pathv;
+  int (*gl_errfunc)(const char *, int);
+  void (*gl_closedir)(void *dirp);
+  struct dirent *(*gl_readdir)(void *dirp);
+  void *(*gl_opendir)(const char *);
+  int (*gl_lstat)(const char *, void * /* struct stat* */);
+  int (*gl_stat)(const char *, void * /* struct stat* */);
+};
+
+extern int glob_nomatch;
+extern int glob_altdirfunc;
+
+extern unsigned path_max;
+
+struct __sanitizer_wordexp_t {
+  uptr we_wordc;
+  char **we_wordv;
+  uptr we_offs;
+  char *we_strings;
+  uptr we_nbytes;
+};
+
+typedef char __sanitizer_FILE;
+#define SANITIZER_HAS_STRUCT_FILE 0
+
+extern int shmctl_ipc_stat;
+
+// This simplifies generic code
+#define struct_shminfo_sz -1
+#define struct_shm_info_sz -1
+#define shmctl_shm_stat -1
+#define shmctl_ipc_info -1
+#define shmctl_shm_info -1
+
+extern unsigned struct_utmp_sz;
+extern unsigned struct_utmpx_sz;
+
+extern int map_fixed;
+
+// ioctl arguments
+struct __sanitizer_ifconf {
+  int ifc_len;
+  union {
+    void *ifcu_req;
+  } ifc_ifcu;
+};
+
+#define IOC_NRBITS 8
+#define IOC_TYPEBITS 8
+#define IOC_SIZEBITS 14
+#define IOC_DIRBITS 2
+#define IOC_NONE 0U
+#define IOC_WRITE 1U
+#define IOC_READ 2U
+#define IOC_NRMASK ((1 << IOC_NRBITS) - 1)
+#define IOC_TYPEMASK ((1 << IOC_TYPEBITS) - 1)
+#define IOC_SIZEMASK ((1 << IOC_SIZEBITS) - 1)
+#undef IOC_DIRMASK
+#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)
+
+extern unsigned struct_ifreq_sz;
+extern unsigned struct_termios_sz;
+extern unsigned struct_winsize_sz;
+
+extern unsigned struct_arpreq_sz;
+
+extern unsigned struct_mtget_sz;
+extern unsigned struct_mtop_sz;
+extern unsigned struct_rtentry_sz;
+extern unsigned struct_sbi_instrument_sz;
+extern unsigned struct_seq_event_rec_sz;
+extern unsigned struct_synth_info_sz;
+extern unsigned struct_vt_mode_sz;
+extern unsigned struct_audio_buf_info_sz;
+extern unsigned struct_ppp_stats_sz;
+extern unsigned struct_sioc_sg_req_sz;
+extern unsigned struct_sioc_vif_req_sz;
+
+// ioctl request identifiers
+
+// A special value to mark ioctls that are not present on the target platform,
+// when it can not be determined without including any system headers.
+extern const unsigned IOCTL_NOT_PRESENT;
+
+extern unsigned IOCTL_FIOASYNC;
+extern unsigned IOCTL_FIOCLEX;
+extern unsigned IOCTL_FIOGETOWN;
+extern unsigned IOCTL_FIONBIO;
+extern unsigned IOCTL_FIONCLEX;
+extern unsigned IOCTL_FIOSETOWN;
+extern unsigned IOCTL_SIOCADDMULTI;
+extern unsigned IOCTL_SIOCATMARK;
+extern unsigned IOCTL_SIOCDELMULTI;
+extern unsigned IOCTL_SIOCGIFADDR;
+extern unsigned IOCTL_SIOCGIFBRDADDR;
+extern unsigned IOCTL_SIOCGIFCONF;
+extern unsigned IOCTL_SIOCGIFDSTADDR;
+extern unsigned IOCTL_SIOCGIFFLAGS;
+extern unsigned IOCTL_SIOCGIFMETRIC;
+extern unsigned IOCTL_SIOCGIFMTU;
+extern unsigned IOCTL_SIOCGIFNETMASK;
+extern unsigned IOCTL_SIOCGPGRP;
+extern unsigned IOCTL_SIOCSIFADDR;
+extern unsigned IOCTL_SIOCSIFBRDADDR;
+extern unsigned IOCTL_SIOCSIFDSTADDR;
+extern unsigned IOCTL_SIOCSIFFLAGS;
+extern unsigned IOCTL_SIOCSIFMETRIC;
+extern unsigned IOCTL_SIOCSIFMTU;
+extern unsigned IOCTL_SIOCSIFNETMASK;
+extern unsigned IOCTL_SIOCSPGRP;
+extern unsigned IOCTL_TIOCCONS;
+extern unsigned IOCTL_TIOCEXCL;
+extern unsigned IOCTL_TIOCGETD;
+extern unsigned IOCTL_TIOCGPGRP;
+extern unsigned IOCTL_TIOCGWINSZ;
+extern unsigned IOCTL_TIOCMBIC;
+extern unsigned IOCTL_TIOCMBIS;
+extern unsigned IOCTL_TIOCMGET;
+extern unsigned IOCTL_TIOCMSET;
+extern unsigned IOCTL_TIOCNOTTY;
+extern unsigned IOCTL_TIOCNXCL;
+extern unsigned IOCTL_TIOCOUTQ;
+extern unsigned IOCTL_TIOCPKT;
+extern unsigned IOCTL_TIOCSCTTY;
+extern unsigned IOCTL_TIOCSETD;
+extern unsigned IOCTL_TIOCSPGRP;
+extern unsigned IOCTL_TIOCSTI;
+extern unsigned IOCTL_TIOCSWINSZ;
+extern unsigned IOCTL_SIOCGETSGCNT;
+extern unsigned IOCTL_SIOCGETVIFCNT;
+extern unsigned IOCTL_MTIOCGET;
+extern unsigned IOCTL_MTIOCTOP;
+extern unsigned IOCTL_SIOCADDRT;
+extern unsigned IOCTL_SIOCDELRT;
+extern unsigned IOCTL_SNDCTL_DSP_GETBLKSIZE;
+extern unsigned IOCTL_SNDCTL_DSP_GETFMTS;
+extern unsigned IOCTL_SNDCTL_DSP_NONBLOCK;
+extern unsigned IOCTL_SNDCTL_DSP_POST;
+extern unsigned IOCTL_SNDCTL_DSP_RESET;
+extern unsigned IOCTL_SNDCTL_DSP_SETFMT;
+extern unsigned IOCTL_SNDCTL_DSP_SETFRAGMENT;
+extern unsigned IOCTL_SNDCTL_DSP_SPEED;
+extern unsigned IOCTL_SNDCTL_DSP_STEREO;
+extern unsigned IOCTL_SNDCTL_DSP_SUBDIVIDE;
+extern unsigned IOCTL_SNDCTL_DSP_SYNC;
+extern unsigned IOCTL_SNDCTL_FM_4OP_ENABLE;
+extern unsigned IOCTL_SNDCTL_FM_LOAD_INSTR;
+extern unsigned IOCTL_SNDCTL_MIDI_INFO;
+extern unsigned IOCTL_SNDCTL_MIDI_PRETIME;
+extern unsigned IOCTL_SNDCTL_SEQ_CTRLRATE;
+extern unsigned IOCTL_SNDCTL_SEQ_GETINCOUNT;
+extern unsigned IOCTL_SNDCTL_SEQ_GETOUTCOUNT;
+extern unsigned IOCTL_SNDCTL_SEQ_NRMIDIS;
+extern unsigned IOCTL_SNDCTL_SEQ_NRSYNTHS;
+extern unsigned IOCTL_SNDCTL_SEQ_OUTOFBAND;
+extern unsigned IOCTL_SNDCTL_SEQ_PANIC;
+extern unsigned IOCTL_SNDCTL_SEQ_PERCMODE;
+extern unsigned IOCTL_SNDCTL_SEQ_RESET;
+extern unsigned IOCTL_SNDCTL_SEQ_RESETSAMPLES;
+extern unsigned IOCTL_SNDCTL_SEQ_SYNC;
+extern unsigned IOCTL_SNDCTL_SEQ_TESTMIDI;
+extern unsigned IOCTL_SNDCTL_SEQ_THRESHOLD;
+extern unsigned IOCTL_SNDCTL_SYNTH_INFO;
+extern unsigned IOCTL_SNDCTL_SYNTH_MEMAVL;
+extern unsigned IOCTL_SNDCTL_TMR_CONTINUE;
+extern unsigned IOCTL_SNDCTL_TMR_METRONOME;
+extern unsigned IOCTL_SNDCTL_TMR_SELECT;
+extern unsigned IOCTL_SNDCTL_TMR_SOURCE;
+extern unsigned IOCTL_SNDCTL_TMR_START;
+extern unsigned IOCTL_SNDCTL_TMR_STOP;
+extern unsigned IOCTL_SNDCTL_TMR_TEMPO;
+extern unsigned IOCTL_SNDCTL_TMR_TIMEBASE;
+extern unsigned IOCTL_SOUND_MIXER_READ_ALTPCM;
+extern unsigned IOCTL_SOUND_MIXER_READ_BASS;
+extern unsigned IOCTL_SOUND_MIXER_READ_CAPS;
+extern unsigned IOCTL_SOUND_MIXER_READ_CD;
+extern unsigned IOCTL_SOUND_MIXER_READ_DEVMASK;
+extern unsigned IOCTL_SOUND_MIXER_READ_ENHANCE;
+extern unsigned IOCTL_SOUND_MIXER_READ_IGAIN;
+extern unsigned IOCTL_SOUND_MIXER_READ_IMIX;
+extern unsigned IOCTL_SOUND_MIXER_READ_LINE1;
+extern unsigned IOCTL_SOUND_MIXER_READ_LINE2;
+extern unsigned IOCTL_SOUND_MIXER_READ_LINE3;
+extern unsigned IOCTL_SOUND_MIXER_READ_LINE;
+extern unsigned IOCTL_SOUND_MIXER_READ_LOUD;
+extern unsigned IOCTL_SOUND_MIXER_READ_MIC;
+extern unsigned IOCTL_SOUND_MIXER_READ_MUTE;
+extern unsigned IOCTL_SOUND_MIXER_READ_OGAIN;
+extern unsigned IOCTL_SOUND_MIXER_READ_PCM;
+extern unsigned IOCTL_SOUND_MIXER_READ_RECLEV;
+extern unsigned IOCTL_SOUND_MIXER_READ_RECMASK;
+extern unsigned IOCTL_SOUND_MIXER_READ_RECSRC;
+extern unsigned IOCTL_SOUND_MIXER_READ_SPEAKER;
+extern unsigned IOCTL_SOUND_MIXER_READ_STEREODEVS;
+extern unsigned IOCTL_SOUND_MIXER_READ_SYNTH;
+extern unsigned IOCTL_SOUND_MIXER_READ_TREBLE;
+extern unsigned IOCTL_SOUND_MIXER_READ_VOLUME;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_ALTPCM;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_BASS;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_CD;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_ENHANCE;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_IGAIN;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_IMIX;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_LINE1;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_LINE2;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_LINE3;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_LINE;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_LOUD;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_MIC;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_MUTE;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_OGAIN;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_PCM;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_RECLEV;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_RECSRC;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_SPEAKER;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_SYNTH;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_TREBLE;
+extern unsigned IOCTL_SOUND_MIXER_WRITE_VOLUME;
+extern unsigned IOCTL_SOUND_PCM_READ_BITS;
+extern unsigned IOCTL_SOUND_PCM_READ_CHANNELS;
+extern unsigned IOCTL_SOUND_PCM_READ_FILTER;
+extern unsigned IOCTL_SOUND_PCM_READ_RATE;
+extern unsigned IOCTL_SOUND_PCM_WRITE_CHANNELS;
+extern unsigned IOCTL_SOUND_PCM_WRITE_FILTER;
+extern unsigned IOCTL_VT_ACTIVATE;
+extern unsigned IOCTL_VT_GETMODE;
+extern unsigned IOCTL_VT_OPENQRY;
+extern unsigned IOCTL_VT_RELDISP;
+extern unsigned IOCTL_VT_SETMODE;
+extern unsigned IOCTL_VT_WAITACTIVE;
+extern unsigned IOCTL_KDDISABIO;
+extern unsigned IOCTL_KDENABIO;
+extern unsigned IOCTL_KDGETLED;
+extern unsigned IOCTL_KDGKBMODE;
+extern unsigned IOCTL_KDGKBTYPE;
+extern unsigned IOCTL_KDMKTONE;
+extern unsigned IOCTL_KDSETLED;
+extern unsigned IOCTL_KDSETMODE;
+extern unsigned IOCTL_KDSKBMODE;
+
+extern const int si_SEGV_MAPERR;
+extern const int si_SEGV_ACCERR;
+}  // namespace __sanitizer
+
+#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))
+
+// 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 SIGACTION_SYMNAME __sigaction14
+
+#endif  // SANITIZER_NETBSD
+
+#endif
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.cc
index 83f4fd22f623..f12e8206abe6 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.cc
@@ -210,6 +210,7 @@ namespace __sanitizer {
   unsigned struct_sigaction_sz = sizeof(struct sigaction);
   unsigned struct_itimerval_sz = sizeof(struct itimerval);
   unsigned pthread_t_sz = sizeof(pthread_t);
+  unsigned pthread_mutex_t_sz = sizeof(pthread_mutex_t);
   unsigned pthread_cond_t_sz = sizeof(pthread_cond_t);
   unsigned pid_t_sz = sizeof(pid_t);
   unsigned timeval_sz = sizeof(timeval);
@@ -264,9 +265,10 @@ namespace __sanitizer {
   unsigned struct_statvfs_sz = sizeof(struct statvfs);
 #endif // (SANITIZER_LINUX || SANITIZER_FREEBSD) && !SANITIZER_ANDROID
 
-  uptr sig_ign = (uptr)SIG_IGN;
-  uptr sig_dfl = (uptr)SIG_DFL;
-  uptr sa_siginfo = (uptr)SA_SIGINFO;
+  const uptr sig_ign = (uptr)SIG_IGN;
+  const uptr sig_dfl = (uptr)SIG_DFL;
+  const uptr sig_err = (uptr)SIG_ERR;
+  const uptr sa_siginfo = (uptr)SA_SIGINFO;
 
 #if SANITIZER_LINUX
   int e_tabsz = (int)E_TABSZ;
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
index 808418012d75..2853a84d43ae 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
@@ -15,6 +15,8 @@
 #ifndef SANITIZER_PLATFORM_LIMITS_POSIX_H
 #define SANITIZER_PLATFORM_LIMITS_POSIX_H
 
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_MAC
+
 #include "sanitizer_internal_defs.h"
 #include "sanitizer_platform.h"
 
@@ -44,6 +46,7 @@ namespace __sanitizer {
   extern unsigned siginfo_t_sz;
   extern unsigned struct_itimerval_sz;
   extern unsigned pthread_t_sz;
+  extern unsigned pthread_mutex_t_sz;
   extern unsigned pthread_cond_t_sz;
   extern unsigned pid_t_sz;
   extern unsigned timeval_sz;
@@ -535,7 +538,7 @@ namespace __sanitizer {
   typedef long __sanitizer_clock_t;
 #endif
 
-#if SANITIZER_LINUX
+#if SANITIZER_LINUX || SANITIZER_FREEBSD
   typedef int __sanitizer_clockid_t;
 #endif
 
@@ -596,13 +599,22 @@ namespace __sanitizer {
   };
 #endif
 
+  struct __sanitizer_siginfo {
+    // The size is determined by looking at sizeof of real siginfo_t on linux.
+    u64 opaque[128 / sizeof(u64)];
+  };
+
+  using __sanitizer_sighandler_ptr = void (*)(int sig);
+  using __sanitizer_sigactionhandler_ptr =
+      void (*)(int sig, __sanitizer_siginfo *siginfo, void *uctx);
+
   // Linux system headers define the 'sa_handler' and 'sa_sigaction' macros.
 #if SANITIZER_ANDROID && (SANITIZER_WORDSIZE == 64)
   struct __sanitizer_sigaction {
     unsigned sa_flags;
     union {
-      void (*sigaction)(int sig, void *siginfo, void *uctx);
-      void (*handler)(int sig);
+      __sanitizer_sigactionhandler_ptr sigaction;
+      __sanitizer_sighandler_ptr handler;
     };
     __sanitizer_sigset_t sa_mask;
     void (*sa_restorer)();
@@ -611,16 +623,16 @@ namespace __sanitizer {
   struct __sanitizer_sigaction {
     unsigned sa_flags;
     union {
-      void (*sigaction)(int sig, void *siginfo, void *uctx);
-      void (*handler)(int sig);
+      __sanitizer_sigactionhandler_ptr sigaction;
+      __sanitizer_sighandler_ptr handler;
     };
     __sanitizer_sigset_t sa_mask;
   };
 #elif SANITIZER_ANDROID && (SANITIZER_WORDSIZE == 32)
   struct __sanitizer_sigaction {
     union {
-      void (*sigaction)(int sig, void *siginfo, void *uctx);
-      void (*handler)(int sig);
+      __sanitizer_sigactionhandler_ptr sigaction;
+      __sanitizer_sighandler_ptr handler;
     };
     __sanitizer_sigset_t sa_mask;
     uptr sa_flags;
@@ -632,8 +644,8 @@ namespace __sanitizer {
     unsigned int sa_flags;
 #endif
     union {
-      void (*sigaction)(int sig, void *siginfo, void *uctx);
-      void (*handler)(int sig);
+      __sanitizer_sigactionhandler_ptr sigaction;
+      __sanitizer_sighandler_ptr handler;
     };
 #if SANITIZER_FREEBSD
     int sa_flags;
@@ -692,7 +704,7 @@ namespace __sanitizer {
     unsigned int sa_flags;
     union {
       void (*handler)(int signo);
-      void (*sigaction)(int signo, void *info, void *ctx);
+      void (*sigaction)(int signo, __sanitizer_siginfo *info, void *ctx);
     };
     __sanitizer_kernel_sigset_t sa_mask;
     void (*sa_restorer)(void);
@@ -701,7 +713,7 @@ namespace __sanitizer {
   struct __sanitizer_kernel_sigaction_t {
     union {
       void (*handler)(int signo);
-      void (*sigaction)(int signo, void *info, void *ctx);
+      void (*sigaction)(int signo, __sanitizer_siginfo *info, void *ctx);
     };
     unsigned long sa_flags;
     void (*sa_restorer)(void);
@@ -709,9 +721,10 @@ namespace __sanitizer {
   };
 #endif
 
-  extern uptr sig_ign;
-  extern uptr sig_dfl;
-  extern uptr sa_siginfo;
+  extern const uptr sig_ign;
+  extern const uptr sig_dfl;
+  extern const uptr sig_err;
+  extern const uptr sa_siginfo;
 
 #if SANITIZER_LINUX
   extern int e_tabsz;
@@ -1489,4 +1502,8 @@ struct __sanitizer_cookie_io_functions_t {
   COMPILER_CHECK(offsetof(struct __sanitizer_##CLASS, MEMBER) ==          \
                  offsetof(struct CLASS, MEMBER))
 
+#define SIGACTION_SYMNAME sigaction
+
+#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_MAC
+
 #endif
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.cc
new file mode 100644
index 000000000000..15b80a81ae3b
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.cc
@@ -0,0 +1,366 @@
+//===-- sanitizer_platform_limits_solaris.cc ------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of Sanitizer common code.
+//
+// Sizes and layouts of platform-specific Solaris data structures.
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_platform.h"
+
+#if SANITIZER_SOLARIS
+#include <arpa/inet.h>
+#include <dirent.h>
+#include <glob.h>
+#include <grp.h>
+#include <ifaddrs.h>
+#include <limits.h>
+#include <link.h>
+#include <net/if.h>
+#include <net/route.h>
+#include <netdb.h>
+#include <netinet/ip_mroute.h>
+#include <poll.h>
+#include <pthread.h>
+#include <pwd.h>
+#include <rpc/xdr.h>
+#include <semaphore.h>
+#include <signal.h>
+#include <stddef.h>
+#include <sys/ethernet.h>
+#include <sys/filio.h>
+#include <sys/ipc.h>
+#include <sys/mman.h>
+#include <sys/mount.h>
+#include <sys/mtio.h>
+#include <sys/ptyvar.h>
+#include <sys/resource.h>
+#include <sys/shm.h>
+#include <sys/socket.h>
+#include <sys/sockio.h>
+#include <sys/stat.h>
+#include <sys/statfs.h>
+#include <sys/statvfs.h>
+#include <sys/time.h>
+#include <sys/timeb.h>
+#include <sys/times.h>
+#include <sys/types.h>
+#include <sys/utsname.h>
+#include <termios.h>
+#include <time.h>
+#include <utmp.h>
+#include <utmpx.h>
+#include <wchar.h>
+#include <wordexp.h>
+
+// Include these after system headers to avoid name clashes and ambiguities.
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_platform_limits_solaris.h"
+
+namespace __sanitizer {
+  unsigned struct_utsname_sz = sizeof(struct utsname);
+  unsigned struct_stat_sz = sizeof(struct stat);
+  unsigned struct_stat64_sz = sizeof(struct stat64);
+  unsigned struct_rusage_sz = sizeof(struct rusage);
+  unsigned struct_tm_sz = sizeof(struct tm);
+  unsigned struct_passwd_sz = sizeof(struct passwd);
+  unsigned struct_group_sz = sizeof(struct group);
+  unsigned siginfo_t_sz = sizeof(siginfo_t);
+  unsigned struct_sigaction_sz = sizeof(struct sigaction);
+  unsigned struct_itimerval_sz = sizeof(struct itimerval);
+  unsigned pthread_t_sz = sizeof(pthread_t);
+  unsigned pthread_mutex_t_sz = sizeof(pthread_mutex_t);
+  unsigned pthread_cond_t_sz = sizeof(pthread_cond_t);
+  unsigned pid_t_sz = sizeof(pid_t);
+  unsigned timeval_sz = sizeof(timeval);
+  unsigned uid_t_sz = sizeof(uid_t);
+  unsigned gid_t_sz = sizeof(gid_t);
+  unsigned mbstate_t_sz = sizeof(mbstate_t);
+  unsigned sigset_t_sz = sizeof(sigset_t);
+  unsigned struct_timezone_sz = sizeof(struct timezone);
+  unsigned struct_tms_sz = sizeof(struct tms);
+  unsigned struct_sigevent_sz = sizeof(struct sigevent);
+  unsigned struct_sched_param_sz = sizeof(struct sched_param);
+  unsigned struct_statfs_sz = sizeof(struct statfs);
+  unsigned struct_sockaddr_sz = sizeof(struct sockaddr);
+  unsigned ucontext_t_sz = sizeof(ucontext_t);
+  unsigned struct_timespec_sz = sizeof(struct timespec);
+#if SANITIZER_SOLARIS32
+  unsigned struct_statvfs64_sz = sizeof(struct statvfs64);
+#endif
+  unsigned struct_statvfs_sz = sizeof(struct statvfs);
+
+  const uptr sig_ign = (uptr)SIG_IGN;
+  const uptr sig_dfl = (uptr)SIG_DFL;
+  const uptr sig_err = (uptr)SIG_ERR;
+  const uptr sa_siginfo = (uptr)SA_SIGINFO;
+
+  int shmctl_ipc_stat = (int)IPC_STAT;
+
+  unsigned struct_utmp_sz = sizeof(struct utmp);
+  unsigned struct_utmpx_sz = sizeof(struct utmpx);
+
+  int map_fixed = MAP_FIXED;
+
+  int af_inet = (int)AF_INET;
+  int af_inet6 = (int)AF_INET6;
+
+  uptr __sanitizer_in_addr_sz(int af) {
+    if (af == AF_INET)
+      return sizeof(struct in_addr);
+    else if (af == AF_INET6)
+      return sizeof(struct in6_addr);
+    else
+      return 0;
+  }
+
+  unsigned struct_ElfW_Phdr_sz = sizeof(ElfW(Phdr));
+
+  int glob_nomatch = GLOB_NOMATCH;
+
+  unsigned path_max = PATH_MAX;
+
+  // ioctl arguments
+  unsigned struct_ifreq_sz = sizeof(struct ifreq);
+  unsigned struct_termios_sz = sizeof(struct termios);
+  unsigned struct_winsize_sz = sizeof(struct winsize);
+
+  unsigned struct_sioc_sg_req_sz = sizeof(struct sioc_sg_req);
+  unsigned struct_sioc_vif_req_sz = sizeof(struct sioc_vif_req);
+
+  const unsigned IOCTL_NOT_PRESENT = 0;
+
+  unsigned IOCTL_FIOASYNC = FIOASYNC;
+  unsigned IOCTL_FIOCLEX = FIOCLEX;
+  unsigned IOCTL_FIOGETOWN = FIOGETOWN;
+  unsigned IOCTL_FIONBIO = FIONBIO;
+  unsigned IOCTL_FIONCLEX = FIONCLEX;
+  unsigned IOCTL_FIOSETOWN = FIOSETOWN;
+  unsigned IOCTL_SIOCADDMULTI = SIOCADDMULTI;
+  unsigned IOCTL_SIOCATMARK = SIOCATMARK;
+  unsigned IOCTL_SIOCDELMULTI = SIOCDELMULTI;
+  unsigned IOCTL_SIOCGIFADDR = SIOCGIFADDR;
+  unsigned IOCTL_SIOCGIFBRDADDR = SIOCGIFBRDADDR;
+  unsigned IOCTL_SIOCGIFCONF = SIOCGIFCONF;
+  unsigned IOCTL_SIOCGIFDSTADDR = SIOCGIFDSTADDR;
+  unsigned IOCTL_SIOCGIFFLAGS = SIOCGIFFLAGS;
+  unsigned IOCTL_SIOCGIFMETRIC = SIOCGIFMETRIC;
+  unsigned IOCTL_SIOCGIFMTU = SIOCGIFMTU;
+  unsigned IOCTL_SIOCGIFNETMASK = SIOCGIFNETMASK;
+  unsigned IOCTL_SIOCGPGRP = SIOCGPGRP;
+  unsigned IOCTL_SIOCSIFADDR = SIOCSIFADDR;
+  unsigned IOCTL_SIOCSIFBRDADDR = SIOCSIFBRDADDR;
+  unsigned IOCTL_SIOCSIFDSTADDR = SIOCSIFDSTADDR;
+  unsigned IOCTL_SIOCSIFFLAGS = SIOCSIFFLAGS;
+  unsigned IOCTL_SIOCSIFMETRIC = SIOCSIFMETRIC;
+  unsigned IOCTL_SIOCSIFMTU = SIOCSIFMTU;
+  unsigned IOCTL_SIOCSIFNETMASK = SIOCSIFNETMASK;
+  unsigned IOCTL_SIOCSPGRP = SIOCSPGRP;
+  unsigned IOCTL_TIOCEXCL = TIOCEXCL;
+  unsigned IOCTL_TIOCGETD = TIOCGETD;
+  unsigned IOCTL_TIOCGPGRP = TIOCGPGRP;
+  unsigned IOCTL_TIOCGWINSZ = TIOCGWINSZ;
+  unsigned IOCTL_TIOCMBIC = TIOCMBIC;
+  unsigned IOCTL_TIOCMBIS = TIOCMBIS;
+  unsigned IOCTL_TIOCMGET = TIOCMGET;
+  unsigned IOCTL_TIOCMSET = TIOCMSET;
+  unsigned IOCTL_TIOCNOTTY = TIOCNOTTY;
+  unsigned IOCTL_TIOCNXCL = TIOCNXCL;
+  unsigned IOCTL_TIOCOUTQ = TIOCOUTQ;
+  unsigned IOCTL_TIOCPKT = TIOCPKT;
+  unsigned IOCTL_TIOCSCTTY = TIOCSCTTY;
+  unsigned IOCTL_TIOCSETD = TIOCSETD;
+  unsigned IOCTL_TIOCSPGRP = TIOCSPGRP;
+  unsigned IOCTL_TIOCSTI = TIOCSTI;
+  unsigned IOCTL_TIOCSWINSZ = TIOCSWINSZ;
+
+  unsigned IOCTL_MTIOCGET = MTIOCGET;
+  unsigned IOCTL_MTIOCTOP = MTIOCTOP;
+
+  const int si_SEGV_MAPERR = SEGV_MAPERR;
+  const int si_SEGV_ACCERR = SEGV_ACCERR;
+} // namespace __sanitizer
+
+using namespace __sanitizer;
+
+COMPILER_CHECK(sizeof(__sanitizer_pthread_attr_t) >= sizeof(pthread_attr_t));
+
+COMPILER_CHECK(sizeof(socklen_t) == sizeof(unsigned));
+CHECK_TYPE_SIZE(pthread_key_t);
+
+// There are more undocumented fields in dl_phdr_info that we are not interested
+// in.
+COMPILER_CHECK(sizeof(__sanitizer_dl_phdr_info) <= sizeof(dl_phdr_info));
+CHECK_SIZE_AND_OFFSET(dl_phdr_info, dlpi_addr);
+CHECK_SIZE_AND_OFFSET(dl_phdr_info, dlpi_name);
+CHECK_SIZE_AND_OFFSET(dl_phdr_info, dlpi_phdr);
+CHECK_SIZE_AND_OFFSET(dl_phdr_info, dlpi_phnum);
+
+CHECK_TYPE_SIZE(glob_t);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_pathc);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_pathv);
+CHECK_SIZE_AND_OFFSET(glob_t, gl_offs);
+
+CHECK_TYPE_SIZE(addrinfo);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_flags);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_family);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_socktype);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_protocol);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_protocol);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_addrlen);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_canonname);
+CHECK_SIZE_AND_OFFSET(addrinfo, ai_addr);
+
+CHECK_TYPE_SIZE(hostent);
+CHECK_SIZE_AND_OFFSET(hostent, h_name);
+CHECK_SIZE_AND_OFFSET(hostent, h_aliases);
+CHECK_SIZE_AND_OFFSET(hostent, h_addrtype);
+CHECK_SIZE_AND_OFFSET(hostent, h_length);
+CHECK_SIZE_AND_OFFSET(hostent, h_addr_list);
+
+CHECK_TYPE_SIZE(iovec);
+CHECK_SIZE_AND_OFFSET(iovec, iov_base);
+CHECK_SIZE_AND_OFFSET(iovec, iov_len);
+
+CHECK_TYPE_SIZE(msghdr);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_name);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_namelen);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_iov);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_iovlen);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_control);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_controllen);
+CHECK_SIZE_AND_OFFSET(msghdr, msg_flags);
+
+CHECK_TYPE_SIZE(cmsghdr);
+CHECK_SIZE_AND_OFFSET(cmsghdr, cmsg_len);
+CHECK_SIZE_AND_OFFSET(cmsghdr, cmsg_level);
+CHECK_SIZE_AND_OFFSET(cmsghdr, cmsg_type);
+
+COMPILER_CHECK(sizeof(__sanitizer_dirent) <= sizeof(dirent));
+CHECK_SIZE_AND_OFFSET(dirent, d_ino);
+CHECK_SIZE_AND_OFFSET(dirent, d_off);
+CHECK_SIZE_AND_OFFSET(dirent, d_reclen);
+
+#if SANITIZER_SOLARIS32
+COMPILER_CHECK(sizeof(__sanitizer_dirent64) <= sizeof(dirent64));
+CHECK_SIZE_AND_OFFSET(dirent64, d_ino);
+CHECK_SIZE_AND_OFFSET(dirent64, d_off);
+CHECK_SIZE_AND_OFFSET(dirent64, d_reclen);
+#endif
+
+CHECK_TYPE_SIZE(ifconf);
+CHECK_SIZE_AND_OFFSET(ifconf, ifc_len);
+CHECK_SIZE_AND_OFFSET(ifconf, ifc_ifcu);
+
+CHECK_TYPE_SIZE(pollfd);
+CHECK_SIZE_AND_OFFSET(pollfd, fd);
+CHECK_SIZE_AND_OFFSET(pollfd, events);
+CHECK_SIZE_AND_OFFSET(pollfd, revents);
+
+CHECK_TYPE_SIZE(nfds_t);
+
+CHECK_TYPE_SIZE(sigset_t);
+
+COMPILER_CHECK(sizeof(__sanitizer_sigaction) == sizeof(struct sigaction));
+// Can't write checks for sa_handler and sa_sigaction due to them being
+// preprocessor macros.
+CHECK_STRUCT_SIZE_AND_OFFSET(sigaction, sa_mask);
+CHECK_STRUCT_SIZE_AND_OFFSET(sigaction, sa_flags);
+
+CHECK_TYPE_SIZE(wordexp_t);
+CHECK_SIZE_AND_OFFSET(wordexp_t, we_wordc);
+CHECK_SIZE_AND_OFFSET(wordexp_t, we_wordv);
+CHECK_SIZE_AND_OFFSET(wordexp_t, we_offs);
+
+CHECK_TYPE_SIZE(tm);
+CHECK_SIZE_AND_OFFSET(tm, tm_sec);
+CHECK_SIZE_AND_OFFSET(tm, tm_min);
+CHECK_SIZE_AND_OFFSET(tm, tm_hour);
+CHECK_SIZE_AND_OFFSET(tm, tm_mday);
+CHECK_SIZE_AND_OFFSET(tm, tm_mon);
+CHECK_SIZE_AND_OFFSET(tm, tm_year);
+CHECK_SIZE_AND_OFFSET(tm, tm_wday);
+CHECK_SIZE_AND_OFFSET(tm, tm_yday);
+CHECK_SIZE_AND_OFFSET(tm, tm_isdst);
+
+CHECK_TYPE_SIZE(ether_addr);
+
+CHECK_TYPE_SIZE(ipc_perm);
+CHECK_SIZE_AND_OFFSET(ipc_perm, key);
+CHECK_SIZE_AND_OFFSET(ipc_perm, seq);
+CHECK_SIZE_AND_OFFSET(ipc_perm, uid);
+CHECK_SIZE_AND_OFFSET(ipc_perm, gid);
+CHECK_SIZE_AND_OFFSET(ipc_perm, cuid);
+CHECK_SIZE_AND_OFFSET(ipc_perm, cgid);
+CHECK_SIZE_AND_OFFSET(ipc_perm, mode);
+
+CHECK_TYPE_SIZE(shmid_ds);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_perm);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_segsz);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_atime);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_dtime);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_ctime);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_cpid);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_lpid);
+CHECK_SIZE_AND_OFFSET(shmid_ds, shm_nattch);
+
+CHECK_TYPE_SIZE(clock_t);
+
+CHECK_TYPE_SIZE(ifaddrs);
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_next);
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_name);
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_addr);
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_netmask);
+// Compare against the union, because we can't reach into the union in a
+// compliant way.
+#ifdef ifa_dstaddr
+#undef ifa_dstaddr
+#endif
+COMPILER_CHECK(sizeof(((__sanitizer_ifaddrs *)nullptr)->ifa_dstaddr) ==
+               sizeof(((ifaddrs *)nullptr)->ifa_ifu));
+COMPILER_CHECK(offsetof(__sanitizer_ifaddrs, ifa_dstaddr) ==
+               offsetof(ifaddrs, ifa_ifu));
+CHECK_SIZE_AND_OFFSET(ifaddrs, ifa_data);
+
+CHECK_TYPE_SIZE(timeb);
+CHECK_SIZE_AND_OFFSET(timeb, time);
+CHECK_SIZE_AND_OFFSET(timeb, millitm);
+CHECK_SIZE_AND_OFFSET(timeb, timezone);
+CHECK_SIZE_AND_OFFSET(timeb, dstflag);
+
+CHECK_TYPE_SIZE(passwd);
+CHECK_SIZE_AND_OFFSET(passwd, pw_name);
+CHECK_SIZE_AND_OFFSET(passwd, pw_passwd);
+CHECK_SIZE_AND_OFFSET(passwd, pw_uid);
+CHECK_SIZE_AND_OFFSET(passwd, pw_gid);
+CHECK_SIZE_AND_OFFSET(passwd, pw_dir);
+CHECK_SIZE_AND_OFFSET(passwd, pw_shell);
+
+CHECK_SIZE_AND_OFFSET(passwd, pw_gecos);
+
+CHECK_TYPE_SIZE(group);
+CHECK_SIZE_AND_OFFSET(group, gr_name);
+CHECK_SIZE_AND_OFFSET(group, gr_passwd);
+CHECK_SIZE_AND_OFFSET(group, gr_gid);
+CHECK_SIZE_AND_OFFSET(group, gr_mem);
+
+CHECK_TYPE_SIZE(XDR);
+CHECK_SIZE_AND_OFFSET(XDR, x_op);
+CHECK_SIZE_AND_OFFSET(XDR, x_ops);
+CHECK_SIZE_AND_OFFSET(XDR, x_public);
+CHECK_SIZE_AND_OFFSET(XDR, x_private);
+CHECK_SIZE_AND_OFFSET(XDR, x_base);
+CHECK_SIZE_AND_OFFSET(XDR, x_handy);
+COMPILER_CHECK(__sanitizer_XDR_ENCODE == XDR_ENCODE);
+COMPILER_CHECK(__sanitizer_XDR_DECODE == XDR_DECODE);
+COMPILER_CHECK(__sanitizer_XDR_FREE == XDR_FREE);
+
+CHECK_TYPE_SIZE(sem_t);
+
+#endif  // SANITIZER_SOLARIS
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.h
new file mode 100644
index 000000000000..a9db71b99c81
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_solaris.h
@@ -0,0 +1,485 @@
+//===-- sanitizer_platform_limits_solaris.h -------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of Sanitizer common code.
+//
+// Sizes and layouts of platform-specific Solaris data structures.
+//===----------------------------------------------------------------------===//
+
+#ifndef SANITIZER_PLATFORM_LIMITS_SOLARIS_H
+#define SANITIZER_PLATFORM_LIMITS_SOLARIS_H
+
+#if SANITIZER_SOLARIS
+
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_platform.h"
+
+namespace __sanitizer {
+extern unsigned struct_utsname_sz;
+extern unsigned struct_stat_sz;
+extern unsigned struct_stat64_sz;
+extern unsigned struct_rusage_sz;
+extern unsigned siginfo_t_sz;
+extern unsigned struct_itimerval_sz;
+extern unsigned pthread_t_sz;
+extern unsigned pthread_mutex_t_sz;
+extern unsigned pthread_cond_t_sz;
+extern unsigned pid_t_sz;
+extern unsigned timeval_sz;
+extern unsigned uid_t_sz;
+extern unsigned gid_t_sz;
+extern unsigned mbstate_t_sz;
+extern unsigned struct_timezone_sz;
+extern unsigned struct_tms_sz;
+extern unsigned struct_itimerspec_sz;
+extern unsigned struct_sigevent_sz;
+extern unsigned struct_sched_param_sz;
+extern unsigned struct_statfs64_sz;
+extern unsigned struct_statfs_sz;
+extern unsigned struct_sockaddr_sz;
+extern unsigned ucontext_t_sz;
+
+extern unsigned struct_timespec_sz;
+extern unsigned struct_rlimit_sz;
+extern unsigned struct_utimbuf_sz;
+
+struct __sanitizer_sem_t {
+  //u64 data[6];
+  u32 sem_count;
+  u16 sem_type;
+  u16 sem_magic;
+  u64 sem_pad1[3];
+  u64 sem_pad2[2];
+};
+
+struct __sanitizer_ipc_perm {
+  unsigned int uid;           // uid_t
+  unsigned int gid;           // gid_t
+  unsigned int cuid;          // uid_t
+  unsigned int cgid;          // gid_t
+  unsigned int mode;          // mode_t
+  unsigned int seq;           // uint_t
+  int key;                    // key_t
+#if !defined(_LP64)
+  int pad[4];
+#endif
+  };
+
+struct __sanitizer_shmid_ds {
+  __sanitizer_ipc_perm shm_perm;
+  unsigned long shm_segsz;    // size_t
+  unsigned long shm_flags;    // uintptr_t
+  unsigned short shm_lkcnt;   // ushort_t
+  int shm_lpid;               // pid_t
+  int shm_cpid;               // pid_t
+  unsigned long shm_nattch;   // shmatt_t
+  unsigned long shm_cnattch;  // ulong_t
+#if defined(_LP64)
+  long shm_atime;             // time_t
+  long shm_dtime;
+  long shm_ctime;
+  void *shm_amp;
+  u64 shm_gransize;           // uint64_t
+  u64 shm_allocated;          // uint64_t
+  u64 shm_pad4[1];            // int64_t
+#else
+  long shm_atime;             // time_t
+  int shm_pad1;               // int32_t
+  long shm_dtime;             // time_t
+  int shm_pad2;               // int32_t
+  long shm_ctime;             // time_t
+  void *shm_amp;
+  u64 shm_gransize;           // uint64_t
+  u64 shm_allocated;          // uint64_t
+#endif
+};
+
+extern unsigned struct_statvfs_sz;
+#if SANITIZER_SOLARIS32
+extern unsigned struct_statvfs64_sz;
+#endif
+
+struct __sanitizer_iovec {
+  void *iov_base;
+  uptr iov_len;
+};
+
+struct __sanitizer_ifaddrs {
+  struct __sanitizer_ifaddrs *ifa_next;
+  char *ifa_name;
+  u64 ifa_flags;     // uint64_t
+  void *ifa_addr;    // (struct sockaddr *)
+  void *ifa_netmask; // (struct sockaddr *)
+  // This is a union on Linux.
+# ifdef ifa_dstaddr
+# undef ifa_dstaddr
+# endif
+  void *ifa_dstaddr; // (struct sockaddr *)
+  void *ifa_data;
+};
+
+typedef unsigned __sanitizer_pthread_key_t;
+
+struct __sanitizer_XDR {
+  int x_op;
+  void *x_ops;
+  uptr x_public;
+  uptr x_private;
+  uptr x_base;
+  unsigned x_handy;
+};
+
+const int __sanitizer_XDR_ENCODE = 0;
+const int __sanitizer_XDR_DECODE = 1;
+const int __sanitizer_XDR_FREE = 2;
+
+struct __sanitizer_passwd {
+  char *pw_name;
+  char *pw_passwd;
+  unsigned int pw_uid;    // uid_t
+  unsigned int pw_gid;    // gid_t
+  char *pw_age;
+  char *pw_comment;
+  char *pw_gecos;
+  char *pw_dir;
+  char *pw_shell;
+};
+
+struct __sanitizer_group {
+  char *gr_name;
+  char *gr_passwd;
+  int gr_gid;
+  char **gr_mem;
+};
+
+typedef long __sanitizer_time_t;
+
+struct __sanitizer_timeb {
+  __sanitizer_time_t time;
+  unsigned short millitm;
+  short timezone;
+  short dstflag;
+};
+
+struct __sanitizer_ether_addr {
+  u8 octet[6];
+};
+
+struct __sanitizer_tm {
+  int tm_sec;
+  int tm_min;
+  int tm_hour;
+  int tm_mday;
+  int tm_mon;
+  int tm_year;
+  int tm_wday;
+  int tm_yday;
+  int tm_isdst;
+};
+
+struct __sanitizer_msghdr {
+  void *msg_name;
+  unsigned msg_namelen;
+  struct __sanitizer_iovec *msg_iov;
+  unsigned msg_iovlen;
+  void *msg_control;
+  unsigned msg_controllen;
+  int msg_flags;
+};
+struct __sanitizer_cmsghdr {
+  unsigned cmsg_len;
+  int cmsg_level;
+  int cmsg_type;
+};
+
+#if SANITIZER_SOLARIS32 && 0
+// FIXME: need to deal with large file and non-large file cases
+struct __sanitizer_dirent {
+  unsigned long long d_ino;
+  long long d_off;
+  unsigned short d_reclen;
+  // more fields that we don't care about
+};
+#else
+struct __sanitizer_dirent {
+  unsigned long d_ino;
+  long d_off;
+  unsigned short d_reclen;
+  // more fields that we don't care about
+};
+#endif
+
+struct __sanitizer_dirent64 {
+  unsigned long long d_ino;
+  unsigned long long d_off;
+  unsigned short d_reclen;
+  // more fields that we don't care about
+};
+
+typedef long __sanitizer_clock_t;
+typedef int __sanitizer_clockid_t;
+
+// This thing depends on the platform. We are only interested in the upper
+// limit. Verified with a compiler assert in .cc.
+const int pthread_attr_t_max_sz = 128;
+union __sanitizer_pthread_attr_t {
+  char size[pthread_attr_t_max_sz]; // NOLINT
+  void *align;
+};
+
+struct __sanitizer_sigset_t {
+  // uint32_t * 4
+  unsigned int __bits[4];
+};
+
+struct __sanitizer_siginfo {
+  // The size is determined by looking at sizeof of real siginfo_t on linux.
+  u64 opaque[128 / sizeof(u64)];
+};
+
+using __sanitizer_sighandler_ptr = void (*)(int sig);
+using __sanitizer_sigactionhandler_ptr =
+    void (*)(int sig, __sanitizer_siginfo *siginfo, void *uctx);
+
+struct __sanitizer_sigaction {
+  int sa_flags;
+  union {
+    __sanitizer_sigactionhandler_ptr sigaction;
+    __sanitizer_sighandler_ptr handler;
+  };
+  __sanitizer_sigset_t sa_mask;
+#if !defined(_LP64)
+  int sa_resv[2];
+#endif
+};
+
+struct __sanitizer_kernel_sigset_t {
+  u8 sig[8];
+};
+
+struct __sanitizer_kernel_sigaction_t {
+  union {
+    void (*handler)(int signo);
+    void (*sigaction)(int signo, __sanitizer_siginfo *info, void *ctx);
+  };
+  unsigned long sa_flags;
+  void (*sa_restorer)(void);
+  __sanitizer_kernel_sigset_t sa_mask;
+};
+
+extern const uptr sig_ign;
+extern const uptr sig_dfl;
+extern const uptr sig_err;
+extern const uptr sa_siginfo;
+
+extern int af_inet;
+extern int af_inet6;
+uptr __sanitizer_in_addr_sz(int af);
+
+struct __sanitizer_dl_phdr_info {
+  uptr dlpi_addr;
+  const char *dlpi_name;
+  const void *dlpi_phdr;
+  short dlpi_phnum;
+};
+
+extern unsigned struct_ElfW_Phdr_sz;
+
+struct __sanitizer_addrinfo {
+  int ai_flags;
+  int ai_family;
+  int ai_socktype;
+  int ai_protocol;
+#if defined(__sparcv9)
+  int _ai_pad;
+#endif
+  unsigned ai_addrlen;
+  char *ai_canonname;
+  void *ai_addr;
+  struct __sanitizer_addrinfo *ai_next;
+};
+
+struct __sanitizer_hostent {
+  char *h_name;
+  char **h_aliases;
+  int h_addrtype;
+  int h_length;
+  char **h_addr_list;
+};
+
+struct __sanitizer_pollfd {
+  int fd;
+  short events;
+  short revents;
+};
+
+typedef unsigned long __sanitizer_nfds_t;
+
+struct __sanitizer_glob_t {
+  uptr gl_pathc;
+  char **gl_pathv;
+  uptr gl_offs;
+  char **gl_pathp;
+  int gl_pathn;
+};
+
+extern int glob_nomatch;
+extern int glob_altdirfunc;
+
+extern unsigned path_max;
+
+struct __sanitizer_wordexp_t {
+  uptr we_wordc;
+  char **we_wordv;
+  uptr we_offs;
+  char **we_wordp;
+  int we_wordn;
+};
+
+typedef void __sanitizer_FILE;
+#define SANITIZER_HAS_STRUCT_FILE 0
+
+// This simplifies generic code
+#define struct_shminfo_sz -1
+#define struct_shm_info_sz -1
+#define shmctl_shm_stat -1
+#define shmctl_ipc_info -1
+#define shmctl_shm_info -1
+
+extern int shmctl_ipc_stat;
+
+extern unsigned struct_utmp_sz;
+extern unsigned struct_utmpx_sz;
+
+extern int map_fixed;
+
+// ioctl arguments
+struct __sanitizer_ifconf {
+  int ifc_len;
+  union {
+    void *ifcu_req;
+  } ifc_ifcu;
+};
+
+// <sys/ioccom.h>
+#define IOC_NRBITS 8
+#define IOC_TYPEBITS 8
+#define IOC_SIZEBITS 12
+#define IOC_DIRBITS 4
+#undef IOC_NONE
+#define IOC_NONE 2U     // IOC_VOID
+#define IOC_READ 4U     // IOC_OUT
+#define IOC_WRITE 8U    // IOC_IN
+
+#define IOC_NRMASK ((1 << IOC_NRBITS) - 1)
+#define IOC_TYPEMASK ((1 << IOC_TYPEBITS) - 1)
+#define IOC_SIZEMASK ((1 << IOC_SIZEBITS) - 1)
+#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 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)
+
+#if defined(__sparc__)
+// In sparc the 14 bits SIZE field overlaps with the
+// least significant bit of DIR, so either IOC_READ or
+// IOC_WRITE shall be 1 in order to get a non-zero SIZE.
+#define IOC_SIZE(nr) \
+  ((((((nr) >> 29) & 0x7) & (4U | 2U)) == 0) ? 0 : (((nr) >> 16) & 0x3fff))
+#else
+#define IOC_SIZE(nr) (((nr) >> IOC_SIZESHIFT) & IOC_SIZEMASK)
+#endif
+
+extern unsigned struct_ifreq_sz;
+extern unsigned struct_termios_sz;
+extern unsigned struct_winsize_sz;
+
+extern unsigned struct_sioc_sg_req_sz;
+extern unsigned struct_sioc_vif_req_sz;
+
+// ioctl request identifiers
+
+// A special value to mark ioctls that are not present on the target platform,
+// when it can not be determined without including any system headers.
+extern const unsigned IOCTL_NOT_PRESENT;
+
+extern unsigned IOCTL_FIOASYNC;
+extern unsigned IOCTL_FIOCLEX;
+extern unsigned IOCTL_FIOGETOWN;
+extern unsigned IOCTL_FIONBIO;
+extern unsigned IOCTL_FIONCLEX;
+extern unsigned IOCTL_FIOSETOWN;
+extern unsigned IOCTL_SIOCADDMULTI;
+extern unsigned IOCTL_SIOCATMARK;
+extern unsigned IOCTL_SIOCDELMULTI;
+extern unsigned IOCTL_SIOCGIFADDR;
+extern unsigned IOCTL_SIOCGIFBRDADDR;
+extern unsigned IOCTL_SIOCGIFCONF;
+extern unsigned IOCTL_SIOCGIFDSTADDR;
+extern unsigned IOCTL_SIOCGIFFLAGS;
+extern unsigned IOCTL_SIOCGIFMETRIC;
+extern unsigned IOCTL_SIOCGIFMTU;
+extern unsigned IOCTL_SIOCGIFNETMASK;
+extern unsigned IOCTL_SIOCGPGRP;
+extern unsigned IOCTL_SIOCSIFADDR;
+extern unsigned IOCTL_SIOCSIFBRDADDR;
+extern unsigned IOCTL_SIOCSIFDSTADDR;
+extern unsigned IOCTL_SIOCSIFFLAGS;
+extern unsigned IOCTL_SIOCSIFMETRIC;
+extern unsigned IOCTL_SIOCSIFMTU;
+extern unsigned IOCTL_SIOCSIFNETMASK;
+extern unsigned IOCTL_SIOCSPGRP;
+extern unsigned IOCTL_TIOCEXCL;
+extern unsigned IOCTL_TIOCGETD;
+extern unsigned IOCTL_TIOCGPGRP;
+extern unsigned IOCTL_TIOCGWINSZ;
+extern unsigned IOCTL_TIOCMBIC;
+extern unsigned IOCTL_TIOCMBIS;
+extern unsigned IOCTL_TIOCMGET;
+extern unsigned IOCTL_TIOCMSET;
+extern unsigned IOCTL_TIOCNOTTY;
+extern unsigned IOCTL_TIOCNXCL;
+extern unsigned IOCTL_TIOCOUTQ;
+extern unsigned IOCTL_TIOCPKT;
+extern unsigned IOCTL_TIOCSCTTY;
+extern unsigned IOCTL_TIOCSETD;
+extern unsigned IOCTL_TIOCSPGRP;
+extern unsigned IOCTL_TIOCSTI;
+extern unsigned IOCTL_TIOCSWINSZ;
+extern unsigned IOCTL_MTIOCGET;
+extern unsigned IOCTL_MTIOCTOP;
+
+extern const int si_SEGV_MAPERR;
+extern const int si_SEGV_ACCERR;
+}  // namespace __sanitizer
+
+#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))
+
+// 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))
+
+#endif  // SANITIZER_SOLARIS
+
+#endif
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix.cc
index 8d3128ae199d..1fad71f3582b 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix.cc
@@ -17,6 +17,7 @@
 #if SANITIZER_POSIX
 
 #include "sanitizer_common.h"
+#include "sanitizer_file.h"
 #include "sanitizer_libc.h"
 #include "sanitizer_posix.h"
 #include "sanitizer_procmaps.h"
@@ -294,18 +295,22 @@ bool GetCodeRangeForFile(const char *module, uptr *start, uptr *end) {
   return false;
 }
 
-SignalContext SignalContext::Create(void *siginfo, void *context) {
-  auto si = (siginfo_t *)siginfo;
-  uptr addr = (uptr)si->si_addr;
-  uptr pc, sp, bp;
-  GetPcSpBp(context, &pc, &sp, &bp);
-  WriteFlag write_flag = GetWriteFlag(context);
-  bool is_memory_access = si->si_signo == SIGSEGV;
-  return SignalContext(context, addr, pc, sp, bp, is_memory_access, write_flag);
+uptr SignalContext::GetAddress() const {
+  auto si = static_cast<const siginfo_t *>(siginfo);
+  return (uptr)si->si_addr;
 }
 
-const char *DescribeSignalOrException(int signo) {
-  switch (signo) {
+bool SignalContext::IsMemoryAccess() const {
+  auto si = static_cast<const siginfo_t *>(siginfo);
+  return si->si_signo == SIGSEGV;
+}
+
+int SignalContext::GetType() const {
+  return static_cast<const siginfo_t *>(siginfo)->si_signo;
+}
+
+const char *SignalContext::Describe() const {
+  switch (GetType()) {
     case SIGFPE:
       return "FPE";
     case SIGILL:
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix.h
index e7d37cbf0882..adef08248004 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix.h
@@ -16,7 +16,9 @@
 // ----------- ATTENTION -------------
 // This header should NOT include any other headers from sanitizer runtime.
 #include "sanitizer_internal_defs.h"
+#include "sanitizer_platform_limits_netbsd.h"
 #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:
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix_libcdep.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix_libcdep.cc
index e113fb1093d4..db41cad6c318 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix_libcdep.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_posix_libcdep.cc
@@ -18,7 +18,9 @@
 
 #include "sanitizer_common.h"
 #include "sanitizer_flags.h"
+#include "sanitizer_platform_limits_netbsd.h"
 #include "sanitizer_platform_limits_posix.h"
+#include "sanitizer_platform_limits_solaris.h"
 #include "sanitizer_posix.h"
 #include "sanitizer_procmaps.h"
 #include "sanitizer_stacktrace.h"
@@ -61,7 +63,11 @@ void ReleaseMemoryPagesToOS(uptr beg, uptr end) {
   uptr beg_aligned = RoundUpTo(beg, page_size);
   uptr end_aligned = RoundDownTo(end, page_size);
   if (beg_aligned < end_aligned)
-    madvise((void*)beg_aligned, end_aligned - beg_aligned, MADV_DONTNEED);
+    // In the default Solaris compilation environment, madvise() is declared
+    // to take a caddr_t arg; casting it to void * results in an invalid
+    // conversion error, so use char * instead.
+    madvise((char *)beg_aligned, end_aligned - beg_aligned,
+            SANITIZER_MADVISE_DONTNEED);
 }
 
 void NoHugePagesInRegion(uptr addr, uptr size) {
@@ -213,6 +219,53 @@ void InstallDeadlySignalHandlers(SignalHandlerType handler) {
   MaybeInstallSigaction(SIGFPE, handler);
   MaybeInstallSigaction(SIGILL, handler);
 }
+
+bool SignalContext::IsStackOverflow() const {
+  // Access at a reasonable offset above SP, or slightly below it (to account
+  // for x86_64 or PowerPC redzone, ARM push of multiple registers, etc) is
+  // probably a stack overflow.
+#ifdef __s390__
+  // On s390, the fault address in siginfo points to start of the page, not
+  // to the precise word that was accessed.  Mask off the low bits of sp to
+  // take it into account.
+  bool IsStackAccess = addr >= (sp & ~0xFFF) && addr < sp + 0xFFFF;
+#else
+  bool IsStackAccess = addr + 512 > sp && addr < sp + 0xFFFF;
+#endif
+
+#if __powerpc__
+  // Large stack frames can be allocated with e.g.
+  //   lis r0,-10000
+  //   stdux r1,r1,r0 # store sp to [sp-10000] and update sp by -10000
+  // If the store faults then sp will not have been updated, so test above
+  // will not work, because the fault address will be more than just "slightly"
+  // below sp.
+  if (!IsStackAccess && IsAccessibleMemoryRange(pc, 4)) {
+    u32 inst = *(unsigned *)pc;
+    u32 ra = (inst >> 16) & 0x1F;
+    u32 opcd = inst >> 26;
+    u32 xo = (inst >> 1) & 0x3FF;
+    // Check for store-with-update to sp. The instructions we accept are:
+    //   stbu rs,d(ra)          stbux rs,ra,rb
+    //   sthu rs,d(ra)          sthux rs,ra,rb
+    //   stwu rs,d(ra)          stwux rs,ra,rb
+    //   stdu rs,ds(ra)         stdux rs,ra,rb
+    // where ra is r1 (the stack pointer).
+    if (ra == 1 &&
+        (opcd == 39 || opcd == 45 || opcd == 37 || opcd == 62 ||
+         (opcd == 31 && (xo == 247 || xo == 439 || xo == 183 || xo == 181))))
+      IsStackAccess = true;
+  }
+#endif  // __powerpc__
+
+  // We also check si_code to filter out SEGV caused by something else other
+  // then hitting the guard page or unmapped memory, like, for example,
+  // unaligned memory access.
+  auto si = static_cast<const siginfo_t *>(siginfo);
+  return IsStackAccess &&
+         (si->si_code == si_SEGV_MAPERR || si->si_code == si_SEGV_ACCERR);
+}
+
 #endif  // SANITIZER_GO
 
 bool IsAccessibleMemoryRange(uptr beg, uptr size) {
@@ -289,6 +342,46 @@ void *MmapFixedNoReserve(uptr fixed_addr, uptr size, const char *name) {
   return (void *)p;
 }
 
+uptr ReservedAddressRange::Init(uptr size, const char *name, uptr fixed_addr) {
+  // We don't pass `name` along because, when you enable `decorate_proc_maps`
+  // AND actually use a named mapping AND are using a sanitizer intercepting
+  // `open` (e.g. TSAN, ESAN), then you'll get a failure during initialization.
+  // TODO(flowerhack): Fix the implementation of GetNamedMappingFd to solve
+  // this problem.
+  if (fixed_addr) {
+    base_ = MmapFixedNoAccess(fixed_addr, size);
+  } else {
+    base_ = MmapNoAccess(size);
+  }
+  size_ = size;
+  name_ = name;
+  (void)os_handle_;  // unsupported
+  return reinterpret_cast<uptr>(base_);
+}
+
+// Uses fixed_addr for now.
+// Will use offset instead once we've implemented this function for real.
+uptr ReservedAddressRange::Map(uptr fixed_addr, uptr size) {
+  return reinterpret_cast<uptr>(MmapFixedOrDieOnFatalError(fixed_addr, size));
+}
+
+uptr ReservedAddressRange::MapOrDie(uptr fixed_addr, uptr size) {
+  return reinterpret_cast<uptr>(MmapFixedOrDie(fixed_addr, size));
+}
+
+void ReservedAddressRange::Unmap(uptr addr, uptr size) {
+  void* addr_as_void = reinterpret_cast<void*>(addr);
+  uptr base_as_uptr = reinterpret_cast<uptr>(base_);
+  // Only unmap at the beginning or end of the range.
+  CHECK((addr_as_void == base_) || (addr + size == base_as_uptr + size_));
+  CHECK_LE(size, size_);
+  UnmapOrDie(reinterpret_cast<void*>(addr), size);
+  if (addr_as_void == base_) {
+    base_ = reinterpret_cast<void*>(addr + size);
+  }
+  size_ = size_ - size;
+}
+
 void *MmapFixedNoAccess(uptr fixed_addr, uptr size, const char *name) {
   int fd = name ? GetNamedMappingFd(name, size) : -1;
   unsigned flags = MAP_PRIVATE | MAP_FIXED | MAP_NORESERVE;
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_printf.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_printf.cc
index 99b7ff1b55cf..5c2360043b07 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_printf.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_printf.cc
@@ -28,8 +28,6 @@
 
 namespace __sanitizer {
 
-StaticSpinMutex CommonSanitizerReportMutex;
-
 static int AppendChar(char **buff, const char *buff_end, char c) {
   if (*buff < buff_end) {
     **buff = c;
@@ -229,19 +227,16 @@ static void CallPrintfAndReportCallback(const char *str) {
     PrintfAndReportCallback(str);
 }
 
-static void SharedPrintfCode(bool append_pid, const char *format,
-                             va_list args) {
+static void NOINLINE SharedPrintfCodeNoBuffer(bool append_pid,
+                                              char *local_buffer,
+                                              int buffer_size,
+                                              const char *format,
+                                              va_list args) {
   va_list args2;
   va_copy(args2, args);
   const int kLen = 16 * 1024;
-  // |local_buffer| is small enough not to overflow the stack and/or violate
-  // the stack limit enforced by TSan (-Wframe-larger-than=512). On the other
-  // hand, the bigger the buffer is, the more the chance the error report will
-  // fit into it.
-  char local_buffer[400];
   int needed_length;
   char *buffer = local_buffer;
-  int buffer_size = ARRAY_SIZE(local_buffer);
   // First try to print a message using a local buffer, and then fall back to
   // mmaped buffer.
   for (int use_mmap = 0; use_mmap < 2; use_mmap++) {
@@ -259,7 +254,9 @@ static void SharedPrintfCode(bool append_pid, const char *format,
         RAW_CHECK_MSG(needed_length < kLen, \
                       "Buffer in Report is too short!\n"); \
       }
-    if (append_pid) {
+    // Fuchsia's logging infrastructure always keeps track of the logging
+    // process, thread, and timestamp, so never prepend such information.
+    if (!SANITIZER_FUCHSIA && append_pid) {
       int pid = internal_getpid();
       const char *exe_name = GetProcessName();
       if (common_flags()->log_exe_name && exe_name) {
@@ -267,9 +264,8 @@ static void SharedPrintfCode(bool append_pid, const char *format,
                                            "==%s", exe_name);
         CHECK_NEEDED_LENGTH
       }
-      needed_length += internal_snprintf(buffer + needed_length,
-                                         buffer_size - needed_length,
-                                         "==%d==", pid);
+      needed_length += internal_snprintf(
+          buffer + needed_length, buffer_size - needed_length, "==%d==", pid);
       CHECK_NEEDED_LENGTH
     }
     needed_length += VSNPrintf(buffer + needed_length,
@@ -292,6 +288,17 @@ static void SharedPrintfCode(bool append_pid, const char *format,
   va_end(args2);
 }
 
+static void NOINLINE SharedPrintfCode(bool append_pid, const char *format,
+                                      va_list args) {
+  // |local_buffer| is small enough not to overflow the stack and/or violate
+  // the stack limit enforced by TSan (-Wframe-larger-than=512). On the other
+  // hand, the bigger the buffer is, the more the chance the error report will
+  // fit into it.
+  char local_buffer[400];
+  SharedPrintfCodeNoBuffer(append_pid, local_buffer, ARRAY_SIZE(local_buffer),
+                           format, args);
+}
+
 FORMAT(1, 2)
 void Printf(const char *format, ...) {
   va_list args;
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps.h
index 06d072b4dc77..ea2cb7a5c68b 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps.h
@@ -14,22 +14,19 @@
 #ifndef SANITIZER_PROCMAPS_H
 #define SANITIZER_PROCMAPS_H
 
+#include "sanitizer_platform.h"
+
+#if SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD || \
+    SANITIZER_MAC || SANITIZER_SOLARIS
+
 #include "sanitizer_common.h"
 #include "sanitizer_internal_defs.h"
+#include "sanitizer_linux.h"
+#include "sanitizer_mac.h"
 #include "sanitizer_mutex.h"
 
 namespace __sanitizer {
 
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
-struct ProcSelfMapsBuff {
-  char *data;
-  uptr mmaped_size;
-  uptr len;
-};
-
-// Reads process memory map in an OS-specific way.
-void ReadProcMaps(ProcSelfMapsBuff *proc_maps);
-#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX
 
 // Memory protection masks.
 static const uptr kProtectionRead = 1;
@@ -37,15 +34,20 @@ static const uptr kProtectionWrite = 2;
 static const uptr kProtectionExecute = 4;
 static const uptr kProtectionShared = 8;
 
-struct MemoryMappedSegment {
+struct MemoryMappedSegmentData;
+
+class MemoryMappedSegment {
+ public:
   MemoryMappedSegment(char *buff = nullptr, uptr size = 0)
-      : filename(buff), filename_size(size) {}
+      : filename(buff), filename_size(size), data_(nullptr) {}
   ~MemoryMappedSegment() {}
 
-  bool IsReadable() { return protection & kProtectionRead; }
-  bool IsWritable() { return protection & kProtectionWrite; }
-  bool IsExecutable() { return protection & kProtectionExecute; }
-  bool IsShared() { return protection & kProtectionShared; }
+  bool IsReadable() const { return protection & kProtectionRead; }
+  bool IsWritable() const { return protection & kProtectionWrite; }
+  bool IsExecutable() const { return protection & kProtectionExecute; }
+  bool IsShared() const { return protection & kProtectionShared; }
+
+  void AddAddressRanges(LoadedModule *module);
 
   uptr start;
   uptr end;
@@ -55,6 +57,12 @@ struct MemoryMappedSegment {
   uptr protection;
   ModuleArch arch;
   u8 uuid[kModuleUUIDSize];
+
+ private:
+  friend class MemoryMappingLayout;
+
+  // This field is assigned and owned by MemoryMappingLayout if needed
+  MemoryMappedSegmentData *data_;
 };
 
 class MemoryMappingLayout {
@@ -69,42 +77,14 @@ class MemoryMappingLayout {
   static void CacheMemoryMappings();
 
   // Adds all mapped objects into a vector.
-  void DumpListOfModules(InternalMmapVector<LoadedModule> *modules);
+  void DumpListOfModules(InternalMmapVectorNoCtor<LoadedModule> *modules);
 
  private:
   void LoadFromCache();
 
-  // FIXME: Hide implementation details for different platforms in
-  // platform-specific files.
-# if SANITIZER_FREEBSD || SANITIZER_LINUX
-  ProcSelfMapsBuff proc_self_maps_;
-  const char *current_;
-
-  // Static mappings cache.
-  static ProcSelfMapsBuff cached_proc_self_maps_;
-  static StaticSpinMutex cache_lock_;  // protects cached_proc_self_maps_.
-# elif SANITIZER_MAC
-  template <u32 kLCSegment, typename SegmentCommand>
-  bool NextSegmentLoad(MemoryMappedSegment *segment);
-  int current_image_;
-  u32 current_magic_;
-  u32 current_filetype_;
-  ModuleArch current_arch_;
-  u8 current_uuid_[kModuleUUIDSize];
-  int current_load_cmd_count_;
-  char *current_load_cmd_addr_;
-  bool current_instrumented_;
-# endif
+  MemoryMappingLayoutData data_;
 };
 
-typedef void (*fill_profile_f)(uptr start, uptr rss, bool file,
-                               /*out*/uptr *stats, uptr stats_size);
-
-// 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);
-
 // Returns code range for the specified module.
 bool GetCodeRangeForFile(const char *module, uptr *start, uptr *end);
 
@@ -115,4 +95,6 @@ uptr ParseHex(const char **p);
 
 }  // namespace __sanitizer
 
+#endif  // SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD ||
+        // SANITIZER_MAC || SANITIZER_SOLARIS
 #endif  // SANITIZER_PROCMAPS_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_common.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_common.cc
index b95f301a437d..0cd3e2458d87 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_common.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_common.cc
@@ -12,7 +12,8 @@
 
 #include "sanitizer_platform.h"
 
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
 
 #include "sanitizer_common.h"
 #include "sanitizer_placement_new.h"
@@ -20,9 +21,8 @@
 
 namespace __sanitizer {
 
-// Linker initialized.
-ProcSelfMapsBuff MemoryMappingLayout::cached_proc_self_maps_;
-StaticSpinMutex MemoryMappingLayout::cache_lock_;  // Linker initialized.
+static ProcSelfMapsBuff cached_proc_self_maps;
+static StaticSpinMutex cache_lock;
 
 static int TranslateDigit(char c) {
   if (c >= '0' && c <= '9')
@@ -64,60 +64,60 @@ uptr ParseHex(const char **p) {
   return ParseNumber(p, 16);
 }
 
+void MemoryMappedSegment::AddAddressRanges(LoadedModule *module) {
+  // data_ should be unused on this platform
+  CHECK(!data_);
+  module->addAddressRange(start, end, IsExecutable(), IsWritable());
+}
+
 MemoryMappingLayout::MemoryMappingLayout(bool cache_enabled) {
-  ReadProcMaps(&proc_self_maps_);
-  if (cache_enabled) {
-    if (proc_self_maps_.mmaped_size == 0) {
-      LoadFromCache();
-      CHECK_GT(proc_self_maps_.len, 0);
-    }
-  } else {
-    CHECK_GT(proc_self_maps_.mmaped_size, 0);
-  }
-  Reset();
   // FIXME: in the future we may want to cache the mappings on demand only.
   if (cache_enabled)
     CacheMemoryMappings();
+
+  // Read maps after the cache update to capture the maps/unmaps happening in
+  // the process of updating.
+  ReadProcMaps(&data_.proc_self_maps);
+  if (cache_enabled && data_.proc_self_maps.mmaped_size == 0)
+    LoadFromCache();
+  CHECK_GT(data_.proc_self_maps.mmaped_size, 0);
+  CHECK_GT(data_.proc_self_maps.len, 0);
+
+  Reset();
 }
 
 MemoryMappingLayout::~MemoryMappingLayout() {
   // Only unmap the buffer if it is different from the cached one. Otherwise
   // it will be unmapped when the cache is refreshed.
-  if (proc_self_maps_.data != cached_proc_self_maps_.data) {
-    UnmapOrDie(proc_self_maps_.data, proc_self_maps_.mmaped_size);
-  }
+  if (data_.proc_self_maps.data != cached_proc_self_maps.data)
+    UnmapOrDie(data_.proc_self_maps.data, data_.proc_self_maps.mmaped_size);
 }
 
 void MemoryMappingLayout::Reset() {
-  current_ = proc_self_maps_.data;
+  data_.current = data_.proc_self_maps.data;
 }
 
 // static
 void MemoryMappingLayout::CacheMemoryMappings() {
-  SpinMutexLock l(&cache_lock_);
+  ProcSelfMapsBuff new_proc_self_maps;
+  ReadProcMaps(&new_proc_self_maps);
   // Don't invalidate the cache if the mappings are unavailable.
-  ProcSelfMapsBuff old_proc_self_maps;
-  old_proc_self_maps = cached_proc_self_maps_;
-  ReadProcMaps(&cached_proc_self_maps_);
-  if (cached_proc_self_maps_.mmaped_size == 0) {
-    cached_proc_self_maps_ = old_proc_self_maps;
-  } else {
-    if (old_proc_self_maps.mmaped_size) {
-      UnmapOrDie(old_proc_self_maps.data,
-                 old_proc_self_maps.mmaped_size);
-    }
-  }
+  if (new_proc_self_maps.mmaped_size == 0)
+    return;
+  SpinMutexLock l(&cache_lock);
+  if (cached_proc_self_maps.mmaped_size)
+    UnmapOrDie(cached_proc_self_maps.data, cached_proc_self_maps.mmaped_size);
+  cached_proc_self_maps = new_proc_self_maps;
 }
 
 void MemoryMappingLayout::LoadFromCache() {
-  SpinMutexLock l(&cache_lock_);
-  if (cached_proc_self_maps_.data) {
-    proc_self_maps_ = cached_proc_self_maps_;
-  }
+  SpinMutexLock l(&cache_lock);
+  if (cached_proc_self_maps.data)
+    data_.proc_self_maps = cached_proc_self_maps;
 }
 
 void MemoryMappingLayout::DumpListOfModules(
-    InternalMmapVector<LoadedModule> *modules) {
+    InternalMmapVectorNoCtor<LoadedModule> *modules) {
   Reset();
   InternalScopedString module_name(kMaxPathLength);
   MemoryMappedSegment segment(module_name.data(), module_name.size());
@@ -139,8 +139,7 @@ void MemoryMappingLayout::DumpListOfModules(
     uptr base_address = (i ? segment.start : 0) - segment.offset;
     LoadedModule cur_module;
     cur_module.set(cur_name, base_address);
-    cur_module.addAddressRange(segment.start, segment.end,
-                               segment.IsExecutable(), segment.IsWritable());
+    segment.AddAddressRanges(&cur_module);
     modules->push_back(cur_module);
   }
 }
@@ -171,4 +170,5 @@ void GetMemoryProfile(fill_profile_f cb, uptr *stats, uptr stats_size) {
 
 } // namespace __sanitizer
 
-#endif // SANITIZER_FREEBSD || SANITIZER_LINUX
+#endif // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD ||
+       // SANITIZER_SOLARIS
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_freebsd.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_freebsd.cc
index f0cdbeb4483a..c26a6d47ea75 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_freebsd.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_freebsd.cc
@@ -7,18 +7,22 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// Information about the process mappings (FreeBSD-specific parts).
+// Information about the process mappings (FreeBSD and NetBSD-specific parts).
 //===----------------------------------------------------------------------===//
 
 #include "sanitizer_platform.h"
-#if SANITIZER_FREEBSD
+#if SANITIZER_FREEBSD || SANITIZER_NETBSD
 #include "sanitizer_common.h"
+#if SANITIZER_FREEBSD
 #include "sanitizer_freebsd.h"
+#endif
 #include "sanitizer_procmaps.h"
 
 #include <unistd.h>
 #include <sys/sysctl.h>
+#if SANITIZER_FREEBSD
 #include <sys/user.h>
+#endif
 
 // Fix 'kinfo_vmentry' definition on FreeBSD prior v9.2 in 32-bit mode.
 #if SANITIZER_FREEBSD && (SANITIZER_WORDSIZE == 32)
@@ -31,16 +35,30 @@
 namespace __sanitizer {
 
 void ReadProcMaps(ProcSelfMapsBuff *proc_maps) {
-  const int Mib[4] = { CTL_KERN, KERN_PROC, KERN_PROC_VMMAP, getpid() };
+  const int Mib[] = {
+#if SANITIZER_FREEBSD
+    CTL_KERN,
+    KERN_PROC,
+    KERN_PROC_VMMAP,
+    getpid()
+#else
+    CTL_VM,
+    VM_PROC,
+    VM_PROC_MAP,
+    getpid(),
+    sizeof(struct kinfo_vmentry)
+#endif
+  };
+
   size_t Size = 0;
-  int Err = sysctl(Mib, 4, NULL, &Size, NULL, 0);
+  int Err = sysctl(Mib, ARRAY_SIZE(Mib), NULL, &Size, NULL, 0);
   CHECK_EQ(Err, 0);
   CHECK_GT(Size, 0);
 
   size_t MmapedSize = Size * 4 / 3;
   void *VmMap = MmapOrDie(MmapedSize, "ReadProcMaps()");
   Size = MmapedSize;
-  Err = sysctl(Mib, 4, VmMap, &Size, NULL, 0);
+  Err = sysctl(Mib, ARRAY_SIZE(Mib), VmMap, &Size, NULL, 0);
   CHECK_EQ(Err, 0);
 
   proc_maps->data = (char*)VmMap;
@@ -49,9 +67,9 @@ void ReadProcMaps(ProcSelfMapsBuff *proc_maps) {
 }
 
 bool MemoryMappingLayout::Next(MemoryMappedSegment *segment) {
-  char *last = proc_self_maps_.data + proc_self_maps_.len;
-  if (current_ >= last) return false;
-  struct kinfo_vmentry *VmEntry = (struct kinfo_vmentry*)current_;
+  char *last = data_.proc_self_maps.data + data_.proc_self_maps.len;
+  if (data_.current >= last) return false;
+  struct kinfo_vmentry *VmEntry = (struct kinfo_vmentry *)data_.current;
 
   segment->start = (uptr)VmEntry->kve_start;
   segment->end = (uptr)VmEntry->kve_end;
@@ -71,11 +89,15 @@ bool MemoryMappingLayout::Next(MemoryMappedSegment *segment) {
                       VmEntry->kve_path);
   }
 
-  current_ += VmEntry->kve_structsize;
+#if SANITIZER_FREEBSD
+  data_.current += VmEntry->kve_structsize;
+#else
+  data_.current += sizeof(*VmEntry);
+#endif
 
   return true;
 }
 
 }  // namespace __sanitizer
 
-#endif  // SANITIZER_FREEBSD
+#endif  // SANITIZER_FREEBSD || SANITIZER_NETBSD
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_linux.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_linux.cc
index 1bcad2bf70e6..633e9393cce0 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_linux.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_linux.cc
@@ -18,8 +18,12 @@
 namespace __sanitizer {
 
 void ReadProcMaps(ProcSelfMapsBuff *proc_maps) {
-  ReadFileToBuffer("/proc/self/maps", &proc_maps->data, &proc_maps->mmaped_size,
-                   &proc_maps->len);
+  if (!ReadFileToBuffer("/proc/self/maps", &proc_maps->data,
+                        &proc_maps->mmaped_size, &proc_maps->len)) {
+    proc_maps->data = nullptr;
+    proc_maps->mmaped_size = 0;
+    proc_maps->len = 0;
+  }
 }
 
 static bool IsOneOf(char c, char c1, char c2) {
@@ -27,48 +31,48 @@ static bool IsOneOf(char c, char c1, char c2) {
 }
 
 bool MemoryMappingLayout::Next(MemoryMappedSegment *segment) {
-  char *last = proc_self_maps_.data + proc_self_maps_.len;
-  if (current_ >= last) return false;
-  char *next_line = (char*)internal_memchr(current_, '\n', last - current_);
+  char *last = data_.proc_self_maps.data + data_.proc_self_maps.len;
+  if (data_.current >= last) return false;
+  char *next_line =
+      (char *)internal_memchr(data_.current, '\n', last - data_.current);
   if (next_line == 0)
     next_line = last;
   // Example: 08048000-08056000 r-xp 00000000 03:0c 64593   /foo/bar
-  segment->start = ParseHex(&current_);
-  CHECK_EQ(*current_++, '-');
-  segment->end = ParseHex(&current_);
-  CHECK_EQ(*current_++, ' ');
-  CHECK(IsOneOf(*current_, '-', 'r'));
+  segment->start = ParseHex(&data_.current);
+  CHECK_EQ(*data_.current++, '-');
+  segment->end = ParseHex(&data_.current);
+  CHECK_EQ(*data_.current++, ' ');
+  CHECK(IsOneOf(*data_.current, '-', 'r'));
   segment->protection = 0;
-  if (*current_++ == 'r') segment->protection |= kProtectionRead;
-  CHECK(IsOneOf(*current_, '-', 'w'));
-  if (*current_++ == 'w') segment->protection |= kProtectionWrite;
-  CHECK(IsOneOf(*current_, '-', 'x'));
-  if (*current_++ == 'x') segment->protection |= kProtectionExecute;
-  CHECK(IsOneOf(*current_, 's', 'p'));
-  if (*current_++ == 's') segment->protection |= kProtectionShared;
-  CHECK_EQ(*current_++, ' ');
-  segment->offset = ParseHex(&current_);
-  CHECK_EQ(*current_++, ' ');
-  ParseHex(&current_);
-  CHECK_EQ(*current_++, ':');
-  ParseHex(&current_);
-  CHECK_EQ(*current_++, ' ');
-  while (IsDecimal(*current_))
-    current_++;
+  if (*data_.current++ == 'r') segment->protection |= kProtectionRead;
+  CHECK(IsOneOf(*data_.current, '-', 'w'));
+  if (*data_.current++ == 'w') segment->protection |= kProtectionWrite;
+  CHECK(IsOneOf(*data_.current, '-', 'x'));
+  if (*data_.current++ == 'x') segment->protection |= kProtectionExecute;
+  CHECK(IsOneOf(*data_.current, 's', 'p'));
+  if (*data_.current++ == 's') segment->protection |= kProtectionShared;
+  CHECK_EQ(*data_.current++, ' ');
+  segment->offset = ParseHex(&data_.current);
+  CHECK_EQ(*data_.current++, ' ');
+  ParseHex(&data_.current);
+  CHECK_EQ(*data_.current++, ':');
+  ParseHex(&data_.current);
+  CHECK_EQ(*data_.current++, ' ');
+  while (IsDecimal(*data_.current)) data_.current++;
   // Qemu may lack the trailing space.
   // https://github.com/google/sanitizers/issues/160
-  // CHECK_EQ(*current_++, ' ');
+  // CHECK_EQ(*data_.current++, ' ');
   // Skip spaces.
-  while (current_ < next_line && *current_ == ' ')
-    current_++;
+  while (data_.current < next_line && *data_.current == ' ') data_.current++;
   // Fill in the filename.
   if (segment->filename) {
-    uptr len = Min((uptr)(next_line - current_), segment->filename_size - 1);
-    internal_strncpy(segment->filename, current_, len);
+    uptr len =
+        Min((uptr)(next_line - data_.current), segment->filename_size - 1);
+    internal_strncpy(segment->filename, data_.current, len);
     segment->filename[len] = 0;
   }
 
-  current_ = next_line + 1;
+  data_.current = next_line + 1;
   return true;
 }
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_mac.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_mac.cc
index 560451a16d90..0167ab18ba13 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_mac.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_mac.cc
@@ -36,6 +36,51 @@
 
 namespace __sanitizer {
 
+// Contains information used to iterate through sections.
+struct MemoryMappedSegmentData {
+  char name[kMaxSegName];
+  uptr nsects;
+  const char *current_load_cmd_addr;
+  u32 lc_type;
+  uptr base_virt_addr;
+  uptr addr_mask;
+};
+
+template <typename Section>
+static void NextSectionLoad(LoadedModule *module, MemoryMappedSegmentData *data,
+                            bool isWritable) {
+  const Section *sc = (const Section *)data->current_load_cmd_addr;
+  data->current_load_cmd_addr += sizeof(Section);
+
+  uptr sec_start = (sc->addr & data->addr_mask) + data->base_virt_addr;
+  uptr sec_end = sec_start + sc->size;
+  module->addAddressRange(sec_start, sec_end, /*executable=*/false, isWritable,
+                          sc->sectname);
+}
+
+void MemoryMappedSegment::AddAddressRanges(LoadedModule *module) {
+  // Don't iterate over sections when the caller hasn't set up the
+  // data pointer, when there are no sections, or when the segment
+  // is executable. Avoid iterating over executable sections because
+  // it will confuse libignore, and because the extra granularity
+  // of information is not needed by any sanitizers.
+  if (!data_ || !data_->nsects || IsExecutable()) {
+    module->addAddressRange(start, end, IsExecutable(), IsWritable(),
+                            data_ ? data_->name : nullptr);
+    return;
+  }
+
+  do {
+    if (data_->lc_type == LC_SEGMENT) {
+      NextSectionLoad<struct section>(module, data_, IsWritable());
+#ifdef MH_MAGIC_64
+    } else if (data_->lc_type == LC_SEGMENT_64) {
+      NextSectionLoad<struct section_64>(module, data_, IsWritable());
+#endif
+    }
+  } while (--data_->nsects);
+}
+
 MemoryMappingLayout::MemoryMappingLayout(bool cache_enabled) {
   Reset();
 }
@@ -63,13 +108,13 @@ void MemoryMappingLayout::Reset() {
   // _dyld_image_count is thread-unsafe. We need to register callbacks for
   // adding and removing images which will invalidate the MemoryMappingLayout
   // state.
-  current_image_ = _dyld_image_count();
-  current_load_cmd_count_ = -1;
-  current_load_cmd_addr_ = 0;
-  current_magic_ = 0;
-  current_filetype_ = 0;
-  current_arch_ = kModuleArchUnknown;
-  internal_memset(current_uuid_, 0, kModuleUUIDSize);
+  data_.current_image = _dyld_image_count();
+  data_.current_load_cmd_count = -1;
+  data_.current_load_cmd_addr = 0;
+  data_.current_magic = 0;
+  data_.current_filetype = 0;
+  data_.current_arch = kModuleArchUnknown;
+  internal_memset(data_.current_uuid, 0, kModuleUUIDSize);
 }
 
 // The dyld load address should be unchanged throughout process execution,
@@ -138,39 +183,57 @@ const mach_header *get_dyld_hdr() {
 // segment.
 // Note that the segment addresses are not necessarily sorted.
 template <u32 kLCSegment, typename SegmentCommand>
-bool MemoryMappingLayout::NextSegmentLoad(MemoryMappedSegment *segment) {
-  const char *lc = current_load_cmd_addr_;
-  current_load_cmd_addr_ += ((const load_command *)lc)->cmdsize;
+static bool NextSegmentLoad(MemoryMappedSegment *segment,
+MemoryMappedSegmentData *seg_data, MemoryMappingLayoutData &layout_data) {
+  const char *lc = layout_data.current_load_cmd_addr;
+  layout_data.current_load_cmd_addr += ((const load_command *)lc)->cmdsize;
   if (((const load_command *)lc)->cmd == kLCSegment) {
     const SegmentCommand* sc = (const SegmentCommand *)lc;
-
-    if (current_image_ == kDyldImageIdx) {
+    uptr base_virt_addr, addr_mask;
+    if (layout_data.current_image == kDyldImageIdx) {
+      base_virt_addr = (uptr)get_dyld_hdr();
       // vmaddr is masked with 0xfffff because on macOS versions < 10.12,
       // it contains an absolute address rather than an offset for dyld.
       // To make matters even more complicated, this absolute address
       // isn't actually the absolute segment address, but the offset portion
       // of the address is accurate when combined with the dyld base address,
       // and the mask will give just this offset.
-      segment->start = (sc->vmaddr & 0xfffff) + (uptr)get_dyld_hdr();
-      segment->end = (sc->vmaddr & 0xfffff) + sc->vmsize + (uptr)get_dyld_hdr();
+      addr_mask = 0xfffff;
     } else {
-      const sptr dlloff = _dyld_get_image_vmaddr_slide(current_image_);
-      segment->start = sc->vmaddr + dlloff;
-      segment->end = sc->vmaddr + sc->vmsize + dlloff;
+      base_virt_addr =
+          (uptr)_dyld_get_image_vmaddr_slide(layout_data.current_image);
+      addr_mask = ~0;
+    }
+
+    segment->start = (sc->vmaddr & addr_mask) + base_virt_addr;
+    segment->end = segment->start + sc->vmsize;
+    // Most callers don't need section information, so only fill this struct
+    // when required.
+    if (seg_data) {
+      seg_data->nsects = sc->nsects;
+      seg_data->current_load_cmd_addr =
+          (const char *)lc + sizeof(SegmentCommand);
+      seg_data->lc_type = kLCSegment;
+      seg_data->base_virt_addr = base_virt_addr;
+      seg_data->addr_mask = addr_mask;
+      internal_strncpy(seg_data->name, sc->segname,
+                       ARRAY_SIZE(seg_data->name));
     }
 
     // Return the initial protection.
     segment->protection = sc->initprot;
-    segment->offset =
-        (current_filetype_ == /*MH_EXECUTE*/ 0x2) ? sc->vmaddr : sc->fileoff;
+    segment->offset = (layout_data.current_filetype ==
+                       /*MH_EXECUTE*/ 0x2)
+                          ? sc->vmaddr
+                          : sc->fileoff;
     if (segment->filename) {
-      const char *src = (current_image_ == kDyldImageIdx)
+      const char *src = (layout_data.current_image == kDyldImageIdx)
                             ? kDyldPath
-                            : _dyld_get_image_name(current_image_);
+                            : _dyld_get_image_name(layout_data.current_image);
       internal_strncpy(segment->filename, src, segment->filename_size);
     }
-    segment->arch = current_arch_;
-    internal_memcpy(segment->uuid, current_uuid_, kModuleUUIDSize);
+    segment->arch = layout_data.current_arch;
+    internal_memcpy(segment->uuid, layout_data.current_uuid, kModuleUUIDSize);
     return true;
   }
   return false;
@@ -202,7 +265,7 @@ ModuleArch ModuleArchFromCpuType(cpu_type_t cputype, cpu_subtype_t cpusubtype) {
 }
 
 static const load_command *NextCommand(const load_command *lc) {
-  return (const load_command *)((char *)lc + lc->cmdsize);
+  return (const load_command *)((const char *)lc + lc->cmdsize);
 }
 
 static void FindUUID(const load_command *first_lc, u8 *uuid_output) {
@@ -232,50 +295,53 @@ static bool IsModuleInstrumented(const load_command *first_lc) {
 }
 
 bool MemoryMappingLayout::Next(MemoryMappedSegment *segment) {
-  for (; current_image_ >= kDyldImageIdx; current_image_--) {
-    const mach_header *hdr = (current_image_ == kDyldImageIdx)
+  for (; data_.current_image >= kDyldImageIdx; data_.current_image--) {
+    const mach_header *hdr = (data_.current_image == kDyldImageIdx)
                                  ? get_dyld_hdr()
-                                 : _dyld_get_image_header(current_image_);
+                                 : _dyld_get_image_header(data_.current_image);
     if (!hdr) continue;
-    if (current_load_cmd_count_ < 0) {
+    if (data_.current_load_cmd_count < 0) {
       // Set up for this image;
-      current_load_cmd_count_ = hdr->ncmds;
-      current_magic_ = hdr->magic;
-      current_filetype_ = hdr->filetype;
-      current_arch_ = ModuleArchFromCpuType(hdr->cputype, hdr->cpusubtype);
-      switch (current_magic_) {
+      data_.current_load_cmd_count = hdr->ncmds;
+      data_.current_magic = hdr->magic;
+      data_.current_filetype = hdr->filetype;
+      data_.current_arch = ModuleArchFromCpuType(hdr->cputype, hdr->cpusubtype);
+      switch (data_.current_magic) {
 #ifdef MH_MAGIC_64
         case MH_MAGIC_64: {
-          current_load_cmd_addr_ = (char*)hdr + sizeof(mach_header_64);
+          data_.current_load_cmd_addr =
+              (const char *)hdr + sizeof(mach_header_64);
           break;
         }
 #endif
         case MH_MAGIC: {
-          current_load_cmd_addr_ = (char*)hdr + sizeof(mach_header);
+          data_.current_load_cmd_addr = (const char *)hdr + sizeof(mach_header);
           break;
         }
         default: {
           continue;
         }
       }
-      FindUUID((const load_command *)current_load_cmd_addr_, &current_uuid_[0]);
-      current_instrumented_ =
-          IsModuleInstrumented((const load_command *)current_load_cmd_addr_);
+      FindUUID((const load_command *)data_.current_load_cmd_addr,
+               data_.current_uuid);
+      data_.current_instrumented = IsModuleInstrumented(
+          (const load_command *)data_.current_load_cmd_addr);
     }
 
-    for (; current_load_cmd_count_ >= 0; current_load_cmd_count_--) {
-      switch (current_magic_) {
-        // current_magic_ may be only one of MH_MAGIC, MH_MAGIC_64.
+    for (; data_.current_load_cmd_count >= 0; data_.current_load_cmd_count--) {
+      switch (data_.current_magic) {
+        // data_.current_magic may be only one of MH_MAGIC, MH_MAGIC_64.
 #ifdef MH_MAGIC_64
         case MH_MAGIC_64: {
           if (NextSegmentLoad<LC_SEGMENT_64, struct segment_command_64>(
-                  segment))
+          segment, segment->data_, data_))
             return true;
           break;
         }
 #endif
         case MH_MAGIC: {
-          if (NextSegmentLoad<LC_SEGMENT, struct segment_command>(segment))
+          if (NextSegmentLoad<LC_SEGMENT, struct segment_command>(
+          segment, segment->data_, data_))
             return true;
           break;
         }
@@ -288,11 +354,13 @@ bool MemoryMappingLayout::Next(MemoryMappedSegment *segment) {
 }
 
 void MemoryMappingLayout::DumpListOfModules(
-    InternalMmapVector<LoadedModule> *modules) {
+    InternalMmapVectorNoCtor<LoadedModule> *modules) {
   Reset();
   InternalScopedString module_name(kMaxPathLength);
   MemoryMappedSegment segment(module_name.data(), kMaxPathLength);
-  for (uptr i = 0; Next(&segment); i++) {
+  MemoryMappedSegmentData data;
+  segment.data_ = &data;
+  while (Next(&segment)) {
     if (segment.filename[0] == '\0') continue;
     LoadedModule *cur_module = nullptr;
     if (!modules->empty() &&
@@ -302,10 +370,9 @@ void MemoryMappingLayout::DumpListOfModules(
       modules->push_back(LoadedModule());
       cur_module = &modules->back();
       cur_module->set(segment.filename, segment.start, segment.arch,
-                      segment.uuid, current_instrumented_);
+                      segment.uuid, data_.current_instrumented);
     }
-    cur_module->addAddressRange(segment.start, segment.end,
-                                segment.IsExecutable(), segment.IsWritable());
+    segment.AddAddressRanges(cur_module);
   }
 }
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_solaris.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_solaris.cc
new file mode 100644
index 000000000000..bfe83170f4e2
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_procmaps_solaris.cc
@@ -0,0 +1,57 @@
+//===-- sanitizer_procmaps_solaris.cc -------------------------------------===//
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Information about the process mappings (Solaris-specific parts).
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_platform.h"
+#if SANITIZER_SOLARIS
+#include "sanitizer_common.h"
+#include "sanitizer_procmaps.h"
+
+#include <procfs.h>
+#include <limits.h>
+
+namespace __sanitizer {
+
+void ReadProcMaps(ProcSelfMapsBuff *proc_maps) {
+  ReadFileToBuffer("/proc/self/xmap", &proc_maps->data, &proc_maps->mmaped_size,
+                   &proc_maps->len);
+}
+
+bool MemoryMappingLayout::Next(MemoryMappedSegment *segment) {
+  char *last = data_.proc_self_maps.data + data_.proc_self_maps.len;
+  if (data_.current >= last) return false;
+
+  prxmap_t *xmapentry = (prxmap_t*)data_.current;
+
+  segment->start = (uptr)xmapentry->pr_vaddr;
+  segment->end = (uptr)(xmapentry->pr_vaddr + xmapentry->pr_size);
+  segment->offset = (uptr)xmapentry->pr_offset;
+
+  segment->protection = 0;
+  if ((xmapentry->pr_mflags & MA_READ) != 0)
+    segment->protection |= kProtectionRead;
+  if ((xmapentry->pr_mflags & MA_WRITE) != 0)
+    segment->protection |= kProtectionWrite;
+  if ((xmapentry->pr_mflags & MA_EXEC) != 0)
+    segment->protection |= kProtectionExecute;
+
+  if (segment->filename != NULL && segment->filename_size > 0) {
+    internal_snprintf(segment->filename,
+                      Min(segment->filename_size, (uptr)PATH_MAX), "%s",
+                      xmapentry->pr_mapname);
+  }
+
+  data_.current += sizeof(prxmap_t);
+
+  return true;
+}
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_SOLARIS
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_quarantine.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_quarantine.h
index db38867ced28..886445a27322 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_quarantine.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_quarantine.h
@@ -87,15 +87,14 @@ class Quarantine {
     // is zero (it allows us to perform just one atomic read per Put() call).
     CHECK((size == 0 && cache_size == 0) || cache_size != 0);
 
-    atomic_store(&max_size_, size, memory_order_relaxed);
-    atomic_store(&min_size_, size / 10 * 9,
-                 memory_order_relaxed);  // 90% of max size.
-    atomic_store(&max_cache_size_, cache_size, memory_order_relaxed);
+    atomic_store_relaxed(&max_size_, size);
+    atomic_store_relaxed(&min_size_, size / 10 * 9);  // 90% of max size.
+    atomic_store_relaxed(&max_cache_size_, cache_size);
   }
 
-  uptr GetSize() const { return atomic_load(&max_size_, memory_order_relaxed); }
+  uptr GetSize() const { return atomic_load_relaxed(&max_size_); }
   uptr GetCacheSize() const {
-    return atomic_load(&max_cache_size_, memory_order_relaxed);
+    return atomic_load_relaxed(&max_cache_size_);
   }
 
   void Put(Cache *c, Callback cb, Node *ptr, uptr size) {
@@ -117,7 +116,16 @@ class Quarantine {
       cache_.Transfer(c);
     }
     if (cache_.Size() > GetSize() && recycle_mutex_.TryLock())
-      Recycle(cb);
+      Recycle(atomic_load_relaxed(&min_size_), cb);
+  }
+
+  void NOINLINE DrainAndRecycle(Cache *c, Callback cb) {
+    {
+      SpinMutexLock l(&cache_mutex_);
+      cache_.Transfer(c);
+    }
+    recycle_mutex_.Lock();
+    Recycle(0, cb);
   }
 
   void PrintStats() const {
@@ -139,9 +147,8 @@ class Quarantine {
   Cache cache_;
   char pad2_[kCacheLineSize];
 
-  void NOINLINE Recycle(Callback cb) {
+  void NOINLINE Recycle(uptr min_size, Callback cb) {
     Cache tmp;
-    uptr min_size = atomic_load(&min_size_, memory_order_relaxed);
     {
       SpinMutexLock l(&cache_mutex_);
       // Go over the batches and merge partially filled ones to
@@ -201,7 +208,7 @@ class QuarantineCache {
 
   // Total memory used, including internal accounting.
   uptr Size() const {
-    return atomic_load(&size_, memory_order_relaxed);
+    return atomic_load_relaxed(&size_);
   }
 
   // Memory used for internal accounting.
@@ -225,7 +232,7 @@ class QuarantineCache {
     list_.append_back(&from_cache->list_);
     SizeAdd(from_cache->Size());
 
-    atomic_store(&from_cache->size_, 0, memory_order_relaxed);
+    atomic_store_relaxed(&from_cache->size_, 0);
   }
 
   void EnqueueBatch(QuarantineBatch *b) {
@@ -296,10 +303,10 @@ class QuarantineCache {
   atomic_uintptr_t size_;
 
   void SizeAdd(uptr add) {
-    atomic_store(&size_, Size() + add, memory_order_relaxed);
+    atomic_store_relaxed(&size_, Size() + add);
   }
   void SizeSub(uptr sub) {
-    atomic_store(&size_, Size() - sub, memory_order_relaxed);
+    atomic_store_relaxed(&size_, Size() - sub);
   }
 };
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_report_decorator.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_report_decorator.h
index 86536aa19658..060b58d3f2d3 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_report_decorator.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_report_decorator.h
@@ -27,8 +27,9 @@ class SanitizerCommonDecorator {
   SanitizerCommonDecorator() : ansi_(ColorizeReports()) {}
   const char *Bold()    const { return ansi_ ? "\033[1m" : ""; }
   const char *Default() const { return ansi_ ? "\033[1m\033[0m"  : ""; }
-  const char *Warning()    { return Red(); }
-  const char *EndWarning() { return Default(); }
+  const char *Warning() const { return Red(); }
+  const char *MemoryByte() { return Magenta(); }
+
  protected:
   const char *Black()   const { return ansi_ ? "\033[1m\033[30m" : ""; }
   const char *Red()     const { return ansi_ ? "\033[1m\033[31m" : ""; }
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_signal_interceptors.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_signal_interceptors.inc
new file mode 100644
index 000000000000..f51fc049e327
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_signal_interceptors.inc
@@ -0,0 +1,87 @@
+//===-- sanitizer_signal_interceptors.inc -----------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Signal interceptors for sanitizers.
+//
+//===----------------------------------------------------------------------===//
+
+#include "interception/interception.h"
+#include "sanitizer_common.h"
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_platform_interceptors.h"
+
+using namespace __sanitizer;
+
+#if SANITIZER_NETBSD
+#define sigaction_symname __sigaction14
+#else
+#define sigaction_symname sigaction
+#endif
+
+#ifndef SIGNAL_INTERCEPTOR_SIGNAL_IMPL
+#define SIGNAL_INTERCEPTOR_SIGNAL_IMPL(func, signum, handler) \
+  { return REAL(func)(signum, handler); }
+#endif
+
+#ifndef SIGNAL_INTERCEPTOR_SIGACTION_IMPL
+#define SIGNAL_INTERCEPTOR_SIGACTION_IMPL(signum, act, oldact) \
+  { return REAL(sigaction_symname)(signum, act, oldact); }
+#endif
+
+#if SANITIZER_INTERCEPT_BSD_SIGNAL
+INTERCEPTOR(uptr, bsd_signal, int signum, uptr handler) {
+  if (GetHandleSignalMode(signum) == kHandleSignalExclusive) return 0;
+  SIGNAL_INTERCEPTOR_SIGNAL_IMPL(bsd_signal, signum, handler);
+}
+#define INIT_BSD_SIGNAL COMMON_INTERCEPT_FUNCTION(bsd_signal)
+#else  // SANITIZER_INTERCEPT_BSD_SIGNAL
+#define INIT_BSD_SIGNAL
+#endif  // SANITIZER_INTERCEPT_BSD_SIGNAL
+
+#if SANITIZER_INTERCEPT_SIGNAL_AND_SIGACTION
+INTERCEPTOR(uptr, signal, int signum, uptr handler) {
+  if (GetHandleSignalMode(signum) == kHandleSignalExclusive)
+    return (uptr) nullptr;
+  SIGNAL_INTERCEPTOR_SIGNAL_IMPL(signal, signum, handler);
+}
+#define INIT_SIGNAL COMMON_INTERCEPT_FUNCTION(signal)
+
+INTERCEPTOR(int, sigaction_symname, int signum,
+            const __sanitizer_sigaction *act, __sanitizer_sigaction *oldact) {
+  if (GetHandleSignalMode(signum) == kHandleSignalExclusive) return 0;
+  SIGNAL_INTERCEPTOR_SIGACTION_IMPL(signum, act, oldact);
+}
+#define INIT_SIGACTION COMMON_INTERCEPT_FUNCTION(sigaction_symname)
+
+namespace __sanitizer {
+int real_sigaction(int signum, const void *act, void *oldact) {
+  return REAL(sigaction_symname)(signum, (const __sanitizer_sigaction *)act,
+                         (__sanitizer_sigaction *)oldact);
+}
+}  // namespace __sanitizer
+#else  // SANITIZER_INTERCEPT_SIGNAL_AND_SIGACTION
+#define INIT_SIGNAL
+#define INIT_SIGACTION
+// We need to have defined REAL(sigaction) on other systems.
+namespace __sanitizer {
+struct __sanitizer_sigaction;
+}
+DEFINE_REAL(int, sigaction, int signum, const __sanitizer_sigaction *act,
+            __sanitizer_sigaction *oldact)
+#endif  // SANITIZER_INTERCEPT_SIGNAL_AND_SIGACTION
+
+static void InitializeSignalInterceptors() {
+  static bool was_called_once;
+  CHECK(!was_called_once);
+  was_called_once = true;
+
+  INIT_BSD_SIGNAL;
+  INIT_SIGNAL;
+  INIT_SIGACTION;
+}
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_solaris.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_solaris.cc
new file mode 100644
index 000000000000..a5db22994e0a
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_solaris.cc
@@ -0,0 +1,219 @@
+//===-- sanitizer_solaris.cc ----------------------------------------------===//
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is shared between various sanitizers' runtime libraries and
+// implements Solaris-specific functions.
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_platform.h"
+#if SANITIZER_SOLARIS
+
+#include <stdio.h>
+
+#include "sanitizer_common.h"
+#include "sanitizer_flags.h"
+#include "sanitizer_internal_defs.h"
+#include "sanitizer_libc.h"
+#include "sanitizer_placement_new.h"
+#include "sanitizer_platform_limits_posix.h"
+#include "sanitizer_procmaps.h"
+
+#include <fcntl.h>
+#include <pthread.h>
+#include <sched.h>
+#include <thread.h>
+#include <synch.h>
+#include <signal.h>
+#include <sys/mman.h>
+#include <sys/resource.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <dirent.h>
+#include <unistd.h>
+#include <errno.h>
+#include <stdlib.h>
+
+namespace __sanitizer {
+
+//#include "sanitizer_syscall_generic.inc"
+
+#define _REAL(func) _ ## func
+#define DECLARE__REAL(ret_type, func, ...) \
+  extern "C" ret_type _REAL(func)(__VA_ARGS__)
+#define DECLARE__REAL_AND_INTERNAL(ret_type, func, ...) \
+  DECLARE__REAL(ret_type, func, __VA_ARGS__); \
+  ret_type internal_ ## func(__VA_ARGS__)
+
+// ---------------------- sanitizer_libc.h
+DECLARE__REAL_AND_INTERNAL(uptr, mmap, void *addr, uptr /*size_t*/ length,
+                           int prot, int flags, int fd, OFF_T offset) {
+  return (uptr)_REAL(mmap)(addr, length, prot, flags, fd, offset);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, munmap, void *addr, uptr length) {
+  return _REAL(munmap)(addr, length);
+}
+
+DECLARE__REAL_AND_INTERNAL(int, mprotect, void *addr, uptr length, int prot) {
+  return _REAL(mprotect)(addr, length, prot);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, close, fd_t fd) {
+  return _REAL(close)(fd);
+}
+
+extern "C" int _REAL(open)(const char *, int, ...);
+
+uptr internal_open(const char *filename, int flags) {
+  return _REAL(open)(filename, flags);
+}
+
+uptr internal_open(const char *filename, int flags, u32 mode) {
+  return _REAL(open)(filename, flags, mode);
+}
+
+uptr OpenFile(const char *filename, bool write) {
+  return internal_open(filename,
+      write ? O_WRONLY | O_CREAT : O_RDONLY, 0660);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, read, fd_t fd, void *buf, uptr count) {
+  return _REAL(read)(fd, buf, count);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, write, fd_t fd, const void *buf, uptr count) {
+  return _REAL(write)(fd, buf, count);
+}
+
+// FIXME: There's only _ftruncate64 beginning with Solaris 11.
+DECLARE__REAL_AND_INTERNAL(uptr, ftruncate, fd_t fd, uptr size) {
+  return ftruncate(fd, size);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, stat, const char *path, void *buf) {
+  return _REAL(stat)(path, (struct stat *)buf);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, lstat, const char *path, void *buf) {
+  return _REAL(lstat)(path, (struct stat *)buf);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, fstat, fd_t fd, void *buf) {
+  return _REAL(fstat)(fd, (struct stat *)buf);
+}
+
+uptr internal_filesize(fd_t fd) {
+  struct stat st;
+  if (internal_fstat(fd, &st))
+    return -1;
+  return (uptr)st.st_size;
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, dup2, int oldfd, int newfd) {
+  return _REAL(dup2)(oldfd, newfd);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, readlink, const char *path, char *buf,
+                           uptr bufsize) {
+  return _REAL(readlink)(path, buf, bufsize);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, unlink, const char *path) {
+  return _REAL(unlink)(path);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, rename, const char *oldpath,
+                           const char *newpath) {
+  return _REAL(rename)(oldpath, newpath);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, sched_yield, void) {
+  return sched_yield();
+}
+
+DECLARE__REAL_AND_INTERNAL(void, _exit, int exitcode) {
+  _exit(exitcode);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, execve, const char *filename,
+                           char *const argv[], char *const envp[]) {
+  return _REAL(execve)(filename, argv, envp);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, waitpid, int pid, int *status, int options) {
+  return _REAL(waitpid)(pid, status, options);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, getpid, void) {
+  return _REAL(getpid)();
+}
+
+// FIXME: This might be wrong: _getdents doesn't take a struct linux_dirent *.
+DECLARE__REAL_AND_INTERNAL(uptr, getdents, fd_t fd, struct linux_dirent *dirp,
+                           unsigned int count) {
+  return _REAL(getdents)(fd, dirp, count);
+}
+
+DECLARE__REAL_AND_INTERNAL(uptr, lseek, fd_t fd, OFF_T offset, int whence) {
+  return _REAL(lseek)(fd, offset, whence);
+}
+
+// FIXME: This might be wrong: _sigfillset doesn't take a
+// __sanitizer_sigset_t *.
+DECLARE__REAL_AND_INTERNAL(void, sigfillset, __sanitizer_sigset_t *set) {
+  _REAL(sigfillset)(set);
+}
+
+// FIXME: This might be wrong: _sigprocmask doesn't take __sanitizer_sigset_t *.
+DECLARE__REAL_AND_INTERNAL(uptr, sigprocmask, int how,
+                           __sanitizer_sigset_t *set,
+                           __sanitizer_sigset_t *oldset) {
+  return _REAL(sigprocmask)(how, set, oldset);
+}
+
+DECLARE__REAL_AND_INTERNAL(int, fork, void) {
+  // TODO(glider): this may call user's pthread_atfork() handlers which is bad.
+  return _REAL(fork)();
+}
+
+u64 NanoTime() {
+  return gethrtime();
+}
+
+uptr internal_clock_gettime(__sanitizer_clockid_t clk_id, void *tp) {
+  // FIXME: No internal variant.
+  return clock_gettime(clk_id, (timespec *)tp);
+}
+
+// ----------------- sanitizer_common.h
+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() {
+  CHECK_EQ((uptr)thr_self(), owner_);
+}
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_SOLARIS
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.h
index 90142dffd5e4..f15196e4b725 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.h
@@ -97,6 +97,11 @@ struct BufferedStackTrace : public StackTrace {
   void Unwind(u32 max_depth, uptr pc, uptr bp, void *context, uptr stack_top,
               uptr stack_bottom, bool request_fast_unwind);
 
+  void Reset() {
+    *static_cast<StackTrace *>(this) = StackTrace(trace_buffer, 0);
+    top_frame_bp = 0;
+  }
+
  private:
   void FastUnwindStack(uptr pc, uptr bp, uptr stack_top, uptr stack_bottom,
                        u32 max_depth);
@@ -106,8 +111,8 @@ struct BufferedStackTrace : public StackTrace {
   void PopStackFrames(uptr count);
   uptr LocatePcInTrace(uptr pc);
 
-  BufferedStackTrace(const BufferedStackTrace &);
-  void operator=(const BufferedStackTrace &);
+  BufferedStackTrace(const BufferedStackTrace &) = delete;
+  void operator=(const BufferedStackTrace &) = delete;
 };
 
 // Check if given pointer points into allocated stack area.
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_printer.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_printer.cc
index 377f1ce755be..a271302708be 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_printer.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_printer.cc
@@ -12,9 +12,14 @@
 //===----------------------------------------------------------------------===//
 
 #include "sanitizer_stacktrace_printer.h"
+#include "sanitizer_file.h"
+#include "sanitizer_fuchsia.h"
 
 namespace __sanitizer {
 
+// sanitizer_symbolizer_fuchsia.cc implements these differently for Fuchsia.
+#if !SANITIZER_FUCHSIA
+
 static const char *StripFunctionName(const char *function, const char *prefix) {
   if (!function) return nullptr;
   if (!prefix) return function;
@@ -24,6 +29,80 @@ static const char *StripFunctionName(const char *function, const char *prefix) {
   return function;
 }
 
+static const char *DemangleFunctionName(const char *function) {
+  if (!function) return nullptr;
+
+  // NetBSD uses indirection for old threading functions for historical reasons
+  // The mangled names are internal implementation detail and should not be
+  // exposed even in backtraces.
+#if SANITIZER_NETBSD
+  if (!internal_strcmp(function, "__libc_mutex_init"))
+    return "pthread_mutex_init";
+  if (!internal_strcmp(function, "__libc_mutex_lock"))
+    return "pthread_mutex_lock";
+  if (!internal_strcmp(function, "__libc_mutex_trylock"))
+    return "pthread_mutex_trylock";
+  if (!internal_strcmp(function, "__libc_mutex_unlock"))
+    return "pthread_mutex_unlock";
+  if (!internal_strcmp(function, "__libc_mutex_destroy"))
+    return "pthread_mutex_destroy";
+  if (!internal_strcmp(function, "__libc_mutexattr_init"))
+    return "pthread_mutexattr_init";
+  if (!internal_strcmp(function, "__libc_mutexattr_settype"))
+    return "pthread_mutexattr_settype";
+  if (!internal_strcmp(function, "__libc_mutexattr_destroy"))
+    return "pthread_mutexattr_destroy";
+  if (!internal_strcmp(function, "__libc_cond_init"))
+    return "pthread_cond_init";
+  if (!internal_strcmp(function, "__libc_cond_signal"))
+    return "pthread_cond_signal";
+  if (!internal_strcmp(function, "__libc_cond_broadcast"))
+    return "pthread_cond_broadcast";
+  if (!internal_strcmp(function, "__libc_cond_wait"))
+    return "pthread_cond_wait";
+  if (!internal_strcmp(function, "__libc_cond_timedwait"))
+    return "pthread_cond_timedwait";
+  if (!internal_strcmp(function, "__libc_cond_destroy"))
+    return "pthread_cond_destroy";
+  if (!internal_strcmp(function, "__libc_rwlock_init"))
+    return "pthread_rwlock_init";
+  if (!internal_strcmp(function, "__libc_rwlock_rdlock"))
+    return "pthread_rwlock_rdlock";
+  if (!internal_strcmp(function, "__libc_rwlock_wrlock"))
+    return "pthread_rwlock_wrlock";
+  if (!internal_strcmp(function, "__libc_rwlock_tryrdlock"))
+    return "pthread_rwlock_tryrdlock";
+  if (!internal_strcmp(function, "__libc_rwlock_trywrlock"))
+    return "pthread_rwlock_trywrlock";
+  if (!internal_strcmp(function, "__libc_rwlock_unlock"))
+    return "pthread_rwlock_unlock";
+  if (!internal_strcmp(function, "__libc_rwlock_destroy"))
+    return "pthread_rwlock_destroy";
+  if (!internal_strcmp(function, "__libc_thr_keycreate"))
+    return "pthread_key_create";
+  if (!internal_strcmp(function, "__libc_thr_setspecific"))
+    return "pthread_setspecific";
+  if (!internal_strcmp(function, "__libc_thr_getspecific"))
+    return "pthread_getspecific";
+  if (!internal_strcmp(function, "__libc_thr_keydelete"))
+    return "pthread_key_delete";
+  if (!internal_strcmp(function, "__libc_thr_once"))
+    return "pthread_once";
+  if (!internal_strcmp(function, "__libc_thr_self"))
+    return "pthread_self";
+  if (!internal_strcmp(function, "__libc_thr_exit"))
+    return "pthread_exit";
+  if (!internal_strcmp(function, "__libc_thr_setcancelstate"))
+    return "pthread_setcancelstate";
+  if (!internal_strcmp(function, "__libc_thr_equal"))
+    return "pthread_equal";
+  if (!internal_strcmp(function, "__libc_thr_curcpu"))
+    return "pthread_curcpu_np";
+#endif
+
+  return function;
+}
+
 static const char kDefaultFormat[] = "    #%n %p %F %L";
 
 void RenderFrame(InternalScopedString *buffer, const char *format, int frame_no,
@@ -55,7 +134,9 @@ void RenderFrame(InternalScopedString *buffer, const char *format, int frame_no,
       buffer->append("0x%zx", info.module_offset);
       break;
     case 'f':
-      buffer->append("%s", StripFunctionName(info.function, strip_func_prefix));
+      buffer->append("%s",
+        DemangleFunctionName(
+          StripFunctionName(info.function, strip_func_prefix)));
       break;
     case 'q':
       buffer->append("0x%zx", info.function_offset != AddressInfo::kUnknown
@@ -76,7 +157,8 @@ void RenderFrame(InternalScopedString *buffer, const char *format, int frame_no,
       // Function name and offset, if file is unknown.
       if (info.function) {
         buffer->append("in %s",
-                       StripFunctionName(info.function, strip_func_prefix));
+                       DemangleFunctionName(
+                         StripFunctionName(info.function, strip_func_prefix)));
         if (!info.file && info.function_offset != AddressInfo::kUnknown)
           buffer->append("+0x%zx", info.function_offset);
       }
@@ -146,6 +228,8 @@ void RenderData(InternalScopedString *buffer, const char *format,
   }
 }
 
+#endif  // !SANITIZER_FUCHSIA
+
 void RenderSourceLocation(InternalScopedString *buffer, const char *file,
                           int line, int column, bool vs_style,
                           const char *strip_path_prefix) {
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cc
index d7fa5f6451d1..0f543d2d7c87 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cc
@@ -57,6 +57,14 @@
 #include "sanitizer_mutex.h"
 #include "sanitizer_placement_new.h"
 
+// Sufficiently old kernel headers don't provide this value, but we can still
+// call prctl with it. If the runtime kernel is new enough, the prctl call will
+// have the desired effect; if the kernel is too old, the call will error and we
+// can ignore said error.
+#ifndef PR_SET_PTRACER
+#define PR_SET_PTRACER 0x59616d61
+#endif
+
 // This module works by spawning a Linux task which then attaches to every
 // thread in the caller process with ptrace. This suspends the threads, and
 // PTRACE_GETREGS can then be used to obtain their register state. The callback
@@ -245,8 +253,9 @@ static void TracerThreadDieCallback() {
 }
 
 // Signal handler to wake up suspended threads when the tracer thread dies.
-static void TracerThreadSignalHandler(int signum, void *siginfo, void *uctx) {
-  SignalContext ctx = SignalContext::Create(siginfo, uctx);
+static void TracerThreadSignalHandler(int signum, __sanitizer_siginfo *siginfo,
+                                      void *uctx) {
+  SignalContext ctx(siginfo, uctx);
   Printf("Tracer caught signal %d: addr=0x%zx pc=0x%zx sp=0x%zx\n", signum,
          ctx.addr, ctx.pc, ctx.sp);
   ThreadSuspender *inst = thread_suspender_instance;
@@ -263,7 +272,7 @@ static void TracerThreadSignalHandler(int signum, void *siginfo, void *uctx) {
 }
 
 // Size of alternative stack for signal handlers in the tracer thread.
-static const int kHandlerStackSize = 4096;
+static const int kHandlerStackSize = 8192;
 
 // This function will be run as a cloned task.
 static int TracerThread(void* argument) {
@@ -433,9 +442,7 @@ void StopTheWorld(StopTheWorldCallback callback, void *argument) {
     ScopedSetTracerPID scoped_set_tracer_pid(tracer_pid);
     // On some systems we have to explicitly declare that we want to be traced
     // by the tracer thread.
-#ifdef PR_SET_PTRACER
     internal_prctl(PR_SET_PTRACER, tracer_pid, 0, 0, 0);
-#endif
     // Allow the tracer thread to start.
     tracer_thread_argument.mutex.Unlock();
     // NOTE: errno is shared between this thread and the tracer thread.
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_mac.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_mac.cc
index 0c27c472f02e..d84ebef3b7f9 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_mac.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_mac.cc
@@ -55,17 +55,9 @@ void RunThread(void *arg) {
   struct RunThreadArgs *run_args = (struct RunThreadArgs *)arg;
   SuspendedThreadsListMac suspended_threads_list;
 
-  mach_port_t task;
-  kern_return_t err = task_for_pid(mach_task_self(), internal_getpid(), &task);
-  if (err != KERN_SUCCESS) {
-    VReport(1, "Getting task from pid failed (errno %d).\n", err);
-    return;
-  }
-
   thread_array_t threads;
   mach_msg_type_number_t num_threads;
-
-  err = task_threads(task, &threads, &num_threads);
+  kern_return_t err = task_threads(mach_task_self(), &threads, &num_threads);
   if (err != KERN_SUCCESS) {
     VReport(1, "Failed to get threads for task (errno %d).\n", err);
     return;
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_suppressions.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_suppressions.cc
index f0f2c9c725a1..89ddfddd1108 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_suppressions.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_suppressions.cc
@@ -16,6 +16,7 @@
 #include "sanitizer_allocator_internal.h"
 #include "sanitizer_common.h"
 #include "sanitizer_flags.h"
+#include "sanitizer_file.h"
 #include "sanitizer_libc.h"
 #include "sanitizer_placement_new.h"
 
@@ -50,6 +51,7 @@ void SuppressionContext::ParseFromFile(const char *filename) {
   if (filename[0] == '\0')
     return;
 
+#if !SANITIZER_FUCHSIA
   // If we cannot find the file, check if its location is relative to
   // the location of the executable.
   InternalScopedString new_file_path(kMaxPathLength);
@@ -58,6 +60,7 @@ void SuppressionContext::ParseFromFile(const char *filename) {
                                           new_file_path.size())) {
     filename = new_file_path.data();
   }
+#endif  // !SANITIZER_FUCHSIA
 
   // Read the file.
   VPrintf(1, "%s: reading suppressions file at %s\n",
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.cc
index 1cd5b6ee24c0..672c936684be 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.cc
@@ -71,6 +71,10 @@ Symbolizer *Symbolizer::symbolizer_;
 StaticSpinMutex Symbolizer::init_mu_;
 LowLevelAllocator Symbolizer::symbolizer_allocator_;
 
+void Symbolizer::InvalidateModuleList() {
+  modules_fresh_ = false;
+}
+
 void Symbolizer::AddHooks(Symbolizer::StartSymbolizationHook start_hook,
                           Symbolizer::EndSymbolizationHook end_hook) {
   CHECK(start_hook_ == 0 && end_hook_ == 0);
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.h
index 4fc7742a21e3..208362d2f478 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer.h
@@ -119,8 +119,11 @@ class Symbolizer final {
   void AddHooks(StartSymbolizationHook start_hook,
                 EndSymbolizationHook end_hook);
 
+  void RefreshModules();
   const LoadedModule *FindModuleForAddress(uptr address);
 
+  void InvalidateModuleList();
+
  private:
   // GetModuleNameAndOffsetForPC has to return a string to the caller.
   // Since the corresponding module might get unloaded later, we should create
@@ -149,6 +152,7 @@ class Symbolizer final {
                                          uptr *module_offset,
                                          ModuleArch *module_arch);
   ListOfModules modules_;
+  ListOfModules fallback_modules_;
   // If stale, need to reload the modules before looking up addresses.
   bool modules_fresh_;
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_fuchsia.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_fuchsia.cc
new file mode 100644
index 000000000000..3d1117d9d3ca
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_fuchsia.cc
@@ -0,0 +1,107 @@
+//===-- sanitizer_symbolizer_fuchsia.cc -----------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is shared between various sanitizers' runtime libraries.
+//
+// Implementation of Fuchsia-specific symbolizer.
+//===----------------------------------------------------------------------===//
+
+#include "sanitizer_platform.h"
+#if SANITIZER_FUCHSIA
+
+#include "sanitizer_fuchsia.h"
+#include "sanitizer_symbolizer.h"
+
+namespace __sanitizer {
+
+// For Fuchsia we don't do any actual symbolization per se.
+// Instead, we emit text containing raw addresses and raw linkage
+// symbol names, embedded in Fuchsia's symbolization markup format.
+// Fuchsia's logging infrastructure emits enough information about
+// process memory layout that a post-processing filter can do the
+// symbolization and pretty-print the markup.  See the spec at:
+// https://fuchsia.googlesource.com/zircon/+/master/docs/symbolizer_markup.md
+
+// This is used by UBSan for type names, and by ASan for global variable names.
+constexpr const char *kFormatDemangle = "{{{symbol:%s}}}";
+constexpr uptr kFormatDemangleMax = 1024;  // Arbitrary.
+
+// Function name or equivalent from PC location.
+constexpr const char *kFormatFunction = "{{{pc:%p}}}";
+constexpr uptr kFormatFunctionMax = 64;  // More than big enough for 64-bit hex.
+
+// Global variable name or equivalent from data memory address.
+constexpr const char *kFormatData = "{{{data:%p}}}";
+
+// One frame in a backtrace (printed on a line by itself).
+constexpr const char *kFormatFrame = "{{{bt:%u:%p}}}";
+
+// This is used by UBSan for type names, and by ASan for global variable names.
+// It's expected to return a static buffer that will be reused on each call.
+const char *Symbolizer::Demangle(const char *name) {
+  static char buffer[kFormatDemangleMax];
+  internal_snprintf(buffer, sizeof(buffer), kFormatDemangle, name);
+  return buffer;
+}
+
+// This is used mostly for suppression matching.  Making it work
+// would enable "interceptor_via_lib" suppressions.  It's also used
+// once in UBSan to say "in module ..." in a message that also
+// includes an address in the module, so post-processing can already
+// pretty-print that so as to indicate the module.
+bool Symbolizer::GetModuleNameAndOffsetForPC(uptr pc, const char **module_name,
+                                             uptr *module_address) {
+  return false;
+}
+
+// This is used in some places for suppression checking, which we
+// don't really support for Fuchsia.  It's also used in UBSan to
+// identify a PC location to a function name, so we always fill in
+// the function member with a string containing markup around the PC
+// value.
+// TODO(mcgrathr): Under SANITIZER_GO, it's currently used by TSan
+// to render stack frames, but that should be changed to use
+// RenderStackFrame.
+SymbolizedStack *Symbolizer::SymbolizePC(uptr addr) {
+  SymbolizedStack *s = SymbolizedStack::New(addr);
+  char buffer[kFormatFunctionMax];
+  internal_snprintf(buffer, sizeof(buffer), kFormatFunction, addr);
+  s->info.function = internal_strdup(buffer);
+  return s;
+}
+
+// Always claim we succeeded, so that RenderDataInfo will be called.
+bool Symbolizer::SymbolizeData(uptr addr, DataInfo *info) {
+  info->Clear();
+  info->start = addr;
+  return true;
+}
+
+// We ignore the format argument to __sanitizer_symbolize_global.
+void RenderData(InternalScopedString *buffer, const char *format,
+                const DataInfo *DI, const char *strip_path_prefix) {
+  buffer->append(kFormatData, DI->start);
+}
+
+// We don't support the stack_trace_format flag at all.
+void RenderFrame(InternalScopedString *buffer, const char *format, int frame_no,
+                 const AddressInfo &info, bool vs_style,
+                 const char *strip_path_prefix, const char *strip_func_prefix) {
+  buffer->append(kFormatFrame, frame_no, info.address);
+}
+
+Symbolizer *Symbolizer::PlatformInit() {
+  return new (symbolizer_allocator_) Symbolizer({});
+}
+
+void Symbolizer::LateInitialize() { Symbolizer::GetOrInit(); }
+
+}  // namespace __sanitizer
+
+#endif  // SANITIZER_FUCHSIA
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_internal.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_internal.h
index 2ae42b33868b..a2ac1188241d 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_internal.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_internal.h
@@ -15,6 +15,7 @@
 #define SANITIZER_SYMBOLIZER_INTERNAL_H
 
 #include "sanitizer_symbolizer.h"
+#include "sanitizer_file.h"
 
 namespace __sanitizer {
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libbacktrace.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libbacktrace.cc
index 57354668bfbb..e98fab01d6bb 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libbacktrace.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libbacktrace.cc
@@ -95,7 +95,8 @@ struct SymbolizeCodeCallbackArg {
     if (frames_symbolized > 0) {
       SymbolizedStack *cur = SymbolizedStack::New(addr);
       AddressInfo *info = &cur->info;
-      info->FillModuleInfo(first->info.module, first->info.module_offset);
+      info->FillModuleInfo(first->info.module, first->info.module_offset,
+                           first->info.module_arch);
       last->next = cur;
       last = cur;
     }
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libcdep.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libcdep.cc
index 614470a633d0..a4bab668b27a 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libcdep.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_libcdep.cc
@@ -17,6 +17,18 @@
 
 namespace __sanitizer {
 
+Symbolizer *Symbolizer::GetOrInit() {
+  SpinMutexLock l(&init_mu_);
+  if (symbolizer_)
+    return symbolizer_;
+  symbolizer_ = PlatformInit();
+  CHECK(symbolizer_);
+  return symbolizer_;
+}
+
+// See sanitizer_symbolizer_fuchsia.cc.
+#if !SANITIZER_FUCHSIA
+
 const char *ExtractToken(const char *str, const char *delims, char **result) {
   uptr prefix_len = internal_strcspn(str, delims);
   *result = (char*)InternalAlloc(prefix_len + 1);
@@ -151,37 +163,47 @@ bool Symbolizer::FindModuleNameAndOffsetForAddress(uptr address,
   return true;
 }
 
+void Symbolizer::RefreshModules() {
+  modules_.init();
+  fallback_modules_.fallbackInit();
+  RAW_CHECK(modules_.size() > 0);
+  modules_fresh_ = true;
+}
+
+static const LoadedModule *SearchForModule(const ListOfModules &modules,
+                                           uptr address) {
+  for (uptr i = 0; i < modules.size(); i++) {
+    if (modules[i].containsAddress(address)) {
+      return &modules[i];
+    }
+  }
+  return nullptr;
+}
+
 const LoadedModule *Symbolizer::FindModuleForAddress(uptr address) {
   bool modules_were_reloaded = false;
   if (!modules_fresh_) {
-    modules_.init();
-    RAW_CHECK(modules_.size() > 0);
-    modules_fresh_ = true;
+    RefreshModules();
     modules_were_reloaded = true;
   }
-  for (uptr i = 0; i < modules_.size(); i++) {
-    if (modules_[i].containsAddress(address)) {
-      return &modules_[i];
-    }
-  }
-  // Reload the modules and look up again, if we haven't tried it yet.
+  const LoadedModule *module = SearchForModule(modules_, address);
+  if (module) return module;
+
+  // dlopen/dlclose interceptors invalidate the module list, but when
+  // interception is disabled, we need to retry if the lookup fails in
+  // case the module list changed.
+#if !SANITIZER_INTERCEPT_DLOPEN_DLCLOSE
   if (!modules_were_reloaded) {
-    // FIXME: set modules_fresh_ from dlopen()/dlclose() interceptors.
-    // It's too aggressive to reload the list of modules each time we fail
-    // to find a module for a given address.
-    modules_fresh_ = false;
-    return FindModuleForAddress(address);
+    RefreshModules();
+    module = SearchForModule(modules_, address);
+    if (module) return module;
   }
-  return 0;
-}
+#endif
 
-Symbolizer *Symbolizer::GetOrInit() {
-  SpinMutexLock l(&init_mu_);
-  if (symbolizer_)
-    return symbolizer_;
-  symbolizer_ = PlatformInit();
-  CHECK(symbolizer_);
-  return symbolizer_;
+  if (fallback_modules_.size()) {
+    module = SearchForModule(fallback_modules_, address);
+  }
+  return module;
 }
 
 // For now we assume the following protocol:
@@ -451,7 +473,7 @@ bool SymbolizerProcess::ReadFromSymbolizer(char *buffer, uptr max_length) {
     if (ReachedEndOfOutput(buffer, read_len))
       break;
     if (read_len + 1 == max_length) {
-      Report("WARNING: Symbolizer buffer too small");
+      Report("WARNING: Symbolizer buffer too small\n");
       read_len = 0;
       break;
     }
@@ -472,4 +494,6 @@ bool SymbolizerProcess::WriteToSymbolizer(const char *buffer, uptr length) {
   return true;
 }
 
+#endif  // !SANITIZER_FUCHSIA
+
 }  // namespace __sanitizer
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_posix_libcdep.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_posix_libcdep.cc
index 60ec7506c312..71d748d0515f 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_posix_libcdep.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_symbolizer_posix_libcdep.cc
@@ -16,6 +16,7 @@
 #if SANITIZER_POSIX
 #include "sanitizer_allocator_internal.h"
 #include "sanitizer_common.h"
+#include "sanitizer_file.h"
 #include "sanitizer_flags.h"
 #include "sanitizer_internal_defs.h"
 #include "sanitizer_linux.h"
@@ -76,6 +77,7 @@ static swift_demangle_ft swift_demangle_f;
 // symbolication.
 static void InitializeSwiftDemangler() {
   swift_demangle_f = (swift_demangle_ft)dlsym(RTLD_DEFAULT, "swift_demangle");
+  (void)dlerror(); // Cleanup error message in case of failure
 }
 
 // Attempts to demangle a Swift name. The demangler will return nullptr if a
@@ -271,6 +273,10 @@ class Addr2LineProcess : public SymbolizerProcess {
   bool ReadFromSymbolizer(char *buffer, uptr max_length) override {
     if (!SymbolizerProcess::ReadFromSymbolizer(buffer, max_length))
       return false;
+    // The returned buffer is empty when output is valid, but exceeds
+    // max_length.
+    if (*buffer == '\0')
+      return true;
     // We should cut out output_terminator_ at the end of given buffer,
     // appended by addr2line to mark the end of its meaningful output.
     // We cannot scan buffer from it's beginning, because it is legal for it
@@ -470,16 +476,16 @@ static SymbolizerTool *ChooseExternalSymbolizer(LowLevelAllocator *allocator) {
 
   // Otherwise symbolizer program is unknown, let's search $PATH
   CHECK(path == nullptr);
-  if (const char *found_path = FindPathToBinary("llvm-symbolizer")) {
-    VReport(2, "Using llvm-symbolizer found at: %s\n", found_path);
-    return new(*allocator) LLVMSymbolizer(found_path, allocator);
-  }
 #if SANITIZER_MAC
   if (const char *found_path = FindPathToBinary("atos")) {
     VReport(2, "Using atos found at: %s\n", found_path);
     return new(*allocator) AtosSymbolizer(found_path, allocator);
   }
 #endif  // SANITIZER_MAC
+  if (const char *found_path = FindPathToBinary("llvm-symbolizer")) {
+    VReport(2, "Using llvm-symbolizer found at: %s\n", found_path);
+    return new(*allocator) LLVMSymbolizer(found_path, allocator);
+  }
   if (common_flags()->allow_addr2line) {
     if (const char *found_path = FindPathToBinary("addr2line")) {
       VReport(2, "Using addr2line found at: %s\n", found_path);
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_generic.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_generic.inc
index 15cf05f06087..3c8e77867ce5 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_generic.inc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_generic.inc
@@ -7,20 +7,49 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// Generic implementations of internal_syscall and internal_iserror.
+// Generic implementations of internal_syscall* and internal_iserror.
 //
 //===----------------------------------------------------------------------===//
 
-#if SANITIZER_FREEBSD || SANITIZER_MAC
+#if SANITIZER_FREEBSD || SANITIZER_MAC || SANITIZER_NETBSD || SANITIZER_SOLARIS
 # define SYSCALL(name) SYS_ ## name
 #else
 # define SYSCALL(name) __NR_ ## name
 #endif
 
-#if (SANITIZER_FREEBSD || SANITIZER_MAC) && defined(__x86_64__)
+#if SANITIZER_NETBSD
+// We use 3 kinds of internal_syscall's for different types of retval in order
+// to address differences in calling conventions (e.g. registers to place the
+// return value in).
+//   - internal_syscall     for 32-bit length (int, pid_t)
+//   - internal_syscall64   for 64-bit length (off_t)
+//   - internal_syscall_ptr for pointer and (s)size_t
+# define  internal_syscall      syscall
+# define  internal_syscall64    __syscall
+// Handle syscall renames manually
+# define SYS_stat SYS___stat50
+# define SYS_lstat SYS___lstat50
+# define SYS_fstat SYS___fstat50
+# define SYS_gettimeofday SYS___gettimeofday50
+# define SYS_wait4 SYS___wait450
+# define SYS_getdents SYS___getdents30
+# define SYS_sigaltstack SYS___sigaltstack14
+# define SYS_sigprocmask SYS___sigprocmask14
+# define SYS_nanosleep SYS___nanosleep50
+# define SYS_clock_gettime SYS___clock_gettime50
+# if SANITIZER_WORDSIZE == 64
+#  define internal_syscall_ptr  __syscall
+# else
+#  define internal_syscall_ptr  syscall
+# endif
+#elif defined(__x86_64__) && (SANITIZER_FREEBSD || SANITIZER_MAC)
 # define internal_syscall __syscall
+# define internal_syscall64 __syscall
+# define internal_syscall_ptr __syscall
 # else
 # define internal_syscall syscall
+# define internal_syscall64 syscall
+# define internal_syscall_ptr syscall
 #endif
 
 bool internal_iserror(uptr retval, int *rverrno) {
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_aarch64.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_aarch64.inc
index 7ab1d7641449..1f05ed9b6c1c 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_aarch64.inc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_aarch64.inc
@@ -127,6 +127,9 @@ static uptr __internal_syscall(u64 nr, u64 arg1, long arg2, long arg3,
 
 #define internal_syscall(...) __SYSCALL_DISP(__internal_syscall, __VA_ARGS__)
 
+#define internal_syscall_ptr internal_syscall
+#define internal_syscall64 internal_syscall
+
 // Helper function used to avoid cobbler errno.
 bool internal_iserror(uptr retval, int *rverrno) {
   if (retval >= (uptr)-4095) {
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_arm.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_arm.inc
new file mode 100644
index 000000000000..a3fdb9e60d5d
--- /dev/null
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_arm.inc
@@ -0,0 +1,141 @@
+//===-- sanitizer_syscall_linux_arm.inc -------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Implementations of internal_syscall and internal_iserror for Linux/arm.
+//
+//===----------------------------------------------------------------------===//
+
+#define SYSCALL(name) __NR_ ## name
+
+static uptr __internal_syscall(u32 nr) {
+  register u32 r8 asm("r7") = nr;
+  register u32 r0 asm("r0");
+  asm volatile("swi #0"
+               : "=r"(r0)
+               : "r"(r8)
+               : "memory", "cc");
+  return r0;
+}
+#define __internal_syscall0(n) \
+  (__internal_syscall)(n)
+
+static uptr __internal_syscall(u32 nr, u32 arg1) {
+  register u32 r8 asm("r7") = nr;
+  register u32 r0 asm("r0") = arg1;
+  asm volatile("swi #0"
+               : "=r"(r0)
+               : "r"(r8), "0"(r0)
+               : "memory", "cc");
+  return r0;
+}
+#define __internal_syscall1(n, a1) \
+  (__internal_syscall)(n, (u32)(a1))
+
+static uptr __internal_syscall(u32 nr, u32 arg1, long arg2) {
+  register u32 r8 asm("r7") = nr;
+  register u32 r0 asm("r0") = arg1;
+  register u32 r1 asm("r1") = arg2;
+  asm volatile("swi #0"
+               : "=r"(r0)
+               : "r"(r8), "0"(r0), "r"(r1)
+               : "memory", "cc");
+  return r0;
+}
+#define __internal_syscall2(n, a1, a2) \
+  (__internal_syscall)(n, (u32)(a1), (long)(a2))
+
+static uptr __internal_syscall(u32 nr, u32 arg1, long arg2, long arg3) {
+  register u32 r8 asm("r7") = nr;
+  register u32 r0 asm("r0") = arg1;
+  register u32 r1 asm("r1") = arg2;
+  register u32 r2 asm("r2") = arg3;
+  asm volatile("swi #0"
+               : "=r"(r0)
+               : "r"(r8), "0"(r0), "r"(r1), "r"(r2)
+               : "memory", "cc");
+  return r0;
+}
+#define __internal_syscall3(n, a1, a2, a3) \
+  (__internal_syscall)(n, (u32)(a1), (long)(a2), (long)(a3))
+
+static uptr __internal_syscall(u32 nr, u32 arg1, long arg2, long arg3,
+                               u32 arg4) {
+  register u32 r8 asm("r7") = nr;
+  register u32 r0 asm("r0") = arg1;
+  register u32 r1 asm("r1") = arg2;
+  register u32 r2 asm("r2") = arg3;
+  register u32 r3 asm("r3") = arg4;
+  asm volatile("swi #0"
+               : "=r"(r0)
+               : "r"(r8), "0"(r0), "r"(r1), "r"(r2), "r"(r3)
+               : "memory", "cc");
+  return r0;
+}
+#define __internal_syscall4(n, a1, a2, a3, a4) \
+  (__internal_syscall)(n, (u32)(a1), (long)(a2), (long)(a3), (long)(a4))
+
+static uptr __internal_syscall(u32 nr, u32 arg1, long arg2, long arg3,
+                               u32 arg4, long arg5) {
+  register u32 r8 asm("r7") = nr;
+  register u32 r0 asm("r0") = arg1;
+  register u32 r1 asm("r1") = arg2;
+  register u32 r2 asm("r2") = arg3;
+  register u32 r3 asm("r3") = arg4;
+  register u32 r4 asm("r4") = arg5;
+  asm volatile("swi #0"
+               : "=r"(r0)
+               : "r"(r8), "0"(r0), "r"(r1), "r"(r2), "r"(r3), "r"(r4)
+               : "memory", "cc");
+  return r0;
+}
+#define __internal_syscall5(n, a1, a2, a3, a4, a5) \
+  (__internal_syscall)(n, (u32)(a1), (long)(a2), (long)(a3), (long)(a4), \
+                       (u32)(a5))
+
+static uptr __internal_syscall(u32 nr, u32 arg1, long arg2, long arg3,
+                               u32 arg4, long arg5, long arg6) {
+  register u32 r8 asm("r7") = nr;
+  register u32 r0 asm("r0") = arg1;
+  register u32 r1 asm("r1") = arg2;
+  register u32 r2 asm("r2") = arg3;
+  register u32 r3 asm("r3") = arg4;
+  register u32 r4 asm("r4") = arg5;
+  register u32 r5 asm("r5") = arg6;
+  asm volatile("swi #0"
+               : "=r"(r0)
+               : "r"(r8), "0"(r0), "r"(r1), "r"(r2), "r"(r3), "r"(r4), "r"(r5)
+               : "memory", "cc");
+  return r0;
+}
+#define __internal_syscall6(n, a1, a2, a3, a4, a5, a6) \
+  (__internal_syscall)(n, (u32)(a1), (long)(a2), (long)(a3), (long)(a4), \
+                       (u32)(a5), (long)(a6))
+
+#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__)
+
+#define internal_syscall_ptr internal_syscall
+#define internal_syscall64 internal_syscall
+
+// Helper function used to avoid cobbler errno.
+bool internal_iserror(uptr retval, int *rverrno) {
+  if (retval >= (uptr)-4095) {
+    if (rverrno)
+      *rverrno = -retval;
+    return true;
+  }
+  return false;
+}
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_x86_64.inc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_x86_64.inc
index 9853a6a675d3..327aaa80a674 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_x86_64.inc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_syscall_linux_x86_64.inc
@@ -20,6 +20,9 @@ static uptr internal_syscall(u64 nr) {
   return retval;
 }
 
+#define internal_syscall_ptr internal_syscall
+#define internal_syscall64 internal_syscall
+
 template <typename T1>
 static uptr internal_syscall(u64 nr, T1 arg1) {
   u64 retval;
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.cc
index 439e33a08d01..d9fd6549ba48 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.cc
@@ -21,6 +21,7 @@ ThreadContextBase::ThreadContextBase(u32 tid)
       status(ThreadStatusInvalid),
       detached(false), workerthread(false), parent_tid(0), next(0) {
   name[0] = '\0';
+  atomic_store(&thread_destroyed, 0, memory_order_release);
 }
 
 ThreadContextBase::~ThreadContextBase() {
@@ -44,6 +45,14 @@ void ThreadContextBase::SetDead() {
   OnDead();
 }
 
+void ThreadContextBase::SetDestroyed() {
+  atomic_store(&thread_destroyed, 1, memory_order_release);
+}
+
+bool ThreadContextBase::GetDestroyed() {
+  return !!atomic_load(&thread_destroyed, memory_order_acquire);
+}
+
 void ThreadContextBase::SetJoined(void *arg) {
   // FIXME(dvyukov): print message and continue (it's user error).
   CHECK_EQ(false, detached);
@@ -54,8 +63,11 @@ void ThreadContextBase::SetJoined(void *arg) {
 }
 
 void ThreadContextBase::SetFinished() {
-  if (!detached)
-    status = ThreadStatusFinished;
+  // ThreadRegistry::FinishThread calls here in ThreadStatusCreated state
+  // for a thread that never actually started.  In that case the thread
+  // should go to ThreadStatusFinished regardless of whether it was created
+  // as detached.
+  if (!detached || status == ThreadStatusCreated) status = ThreadStatusFinished;
   OnFinished();
 }
 
@@ -82,6 +94,7 @@ void ThreadContextBase::SetCreated(uptr _user_id, u64 _unique_id,
 void ThreadContextBase::Reset() {
   status = ThreadStatusInvalid;
   SetName(0);
+  atomic_store(&thread_destroyed, 0, memory_order_release);
   OnReset();
 }
 
@@ -204,7 +217,8 @@ void ThreadRegistry::SetThreadName(u32 tid, const char *name) {
   CHECK_LT(tid, n_contexts_);
   ThreadContextBase *tctx = threads_[tid];
   CHECK_NE(tctx, 0);
-  CHECK_EQ(ThreadStatusRunning, tctx->status);
+  CHECK_EQ(SANITIZER_FUCHSIA ? ThreadStatusCreated : ThreadStatusRunning,
+           tctx->status);
   tctx->SetName(name);
 }
 
@@ -239,33 +253,54 @@ void ThreadRegistry::DetachThread(u32 tid, void *arg) {
 }
 
 void ThreadRegistry::JoinThread(u32 tid, void *arg) {
-  BlockingMutexLock l(&mtx_);
-  CHECK_LT(tid, n_contexts_);
-  ThreadContextBase *tctx = threads_[tid];
-  CHECK_NE(tctx, 0);
-  if (tctx->status == ThreadStatusInvalid) {
-    Report("%s: Join of non-existent thread\n", SanitizerToolName);
-    return;
-  }
-  tctx->SetJoined(arg);
-  QuarantinePush(tctx);
+  bool destroyed = false;
+  do {
+    {
+      BlockingMutexLock l(&mtx_);
+      CHECK_LT(tid, n_contexts_);
+      ThreadContextBase *tctx = threads_[tid];
+      CHECK_NE(tctx, 0);
+      if (tctx->status == ThreadStatusInvalid) {
+        Report("%s: Join of non-existent thread\n", SanitizerToolName);
+        return;
+      }
+      if ((destroyed = tctx->GetDestroyed())) {
+        tctx->SetJoined(arg);
+        QuarantinePush(tctx);
+      }
+    }
+    if (!destroyed)
+      internal_sched_yield();
+  } while (!destroyed);
 }
 
+// Normally this is called when the thread is about to exit.  If
+// called in ThreadStatusCreated state, then this thread was never
+// really started.  We just did CreateThread for a prospective new
+// thread before trying to create it, and then failed to actually
+// create it, and so never called StartThread.
 void ThreadRegistry::FinishThread(u32 tid) {
   BlockingMutexLock l(&mtx_);
   CHECK_GT(alive_threads_, 0);
   alive_threads_--;
-  CHECK_GT(running_threads_, 0);
-  running_threads_--;
   CHECK_LT(tid, n_contexts_);
   ThreadContextBase *tctx = threads_[tid];
   CHECK_NE(tctx, 0);
-  CHECK_EQ(ThreadStatusRunning, tctx->status);
+  bool dead = tctx->detached;
+  if (tctx->status == ThreadStatusRunning) {
+    CHECK_GT(running_threads_, 0);
+    running_threads_--;
+  } else {
+    // The thread never really existed.
+    CHECK_EQ(tctx->status, ThreadStatusCreated);
+    dead = true;
+  }
   tctx->SetFinished();
-  if (tctx->detached) {
+  if (dead) {
     tctx->SetDead();
     QuarantinePush(tctx);
   }
+  tctx->SetDestroyed();
 }
 
 void ThreadRegistry::StartThread(u32 tid, tid_t os_id, bool workerthread,
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.h
index 9aae875c7360..b203be2f4dbc 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_thread_registry.h
@@ -50,6 +50,8 @@ class ThreadContextBase {
   u32 parent_tid;
   ThreadContextBase *next;  // For storing thread contexts in a list.
 
+  atomic_uint32_t thread_destroyed; // To address race of Joined vs Finished
+
   void SetName(const char *new_name);
 
   void SetDead();
@@ -60,6 +62,9 @@ class ThreadContextBase {
                   u32 _parent_tid, void *arg);
   void Reset();
 
+  void SetDestroyed();
+  bool GetDestroyed();
+
   // The following methods may be overriden by subclasses.
   // Some of them take opaque arg that may be optionally be used
   // by subclasses.
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_tls_get_addr.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_tls_get_addr.cc
index 29db37b8a464..6fa0e9298b31 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_tls_get_addr.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_tls_get_addr.cc
@@ -45,7 +45,7 @@ static const uptr kDestroyedThread = -1;
 
 static inline void DTLS_Deallocate(DTLS::DTV *dtv, uptr size) {
   if (!size) return;
-  VPrintf(2, "__tls_get_addr: DTLS_Deallocate %p %zd\n", dtv, size);
+  VReport(2, "__tls_get_addr: DTLS_Deallocate %p %zd\n", dtv, size);
   UnmapOrDie(dtv, size * sizeof(DTLS::DTV));
   atomic_fetch_sub(&number_of_live_dtls, 1, memory_order_relaxed);
 }
@@ -58,7 +58,7 @@ static inline void DTLS_Resize(uptr new_size) {
       (DTLS::DTV *)MmapOrDie(new_size * sizeof(DTLS::DTV), "DTLS_Resize");
   uptr num_live_dtls =
       atomic_fetch_add(&number_of_live_dtls, 1, memory_order_relaxed);
-  VPrintf(2, "__tls_get_addr: DTLS_Resize %p %zd\n", &dtls, num_live_dtls);
+  VReport(2, "__tls_get_addr: DTLS_Resize %p %zd\n", &dtls, num_live_dtls);
   CHECK_LT(num_live_dtls, 1 << 20);
   uptr old_dtv_size = dtls.dtv_size;
   DTLS::DTV *old_dtv = dtls.dtv;
@@ -72,7 +72,7 @@ static inline void DTLS_Resize(uptr new_size) {
 
 void DTLS_Destroy() {
   if (!common_flags()->intercept_tls_get_addr) return;
-  VPrintf(2, "__tls_get_addr: DTLS_Destroy %p %zd\n", &dtls, dtls.dtv_size);
+  VReport(2, "__tls_get_addr: DTLS_Destroy %p %zd\n", &dtls, dtls.dtv_size);
   uptr s = dtls.dtv_size;
   dtls.dtv_size = kDestroyedThread;  // Do this before unmap for AS-safety.
   DTLS_Deallocate(dtls.dtv, s);
@@ -97,28 +97,28 @@ DTLS::DTV *DTLS_on_tls_get_addr(void *arg_void, void *res,
   if (dtls.dtv[dso_id].beg) return 0;
   uptr tls_size = 0;
   uptr tls_beg = reinterpret_cast<uptr>(res) - arg->offset - kDtvOffset;
-  VPrintf(2, "__tls_get_addr: %p {%p,%p} => %p; tls_beg: %p; sp: %p "
+  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,
           atomic_load(&number_of_live_dtls, memory_order_relaxed));
   if (dtls.last_memalign_ptr == tls_beg) {
     tls_size = dtls.last_memalign_size;
-    VPrintf(2, "__tls_get_addr: glibc <=2.18 suspected; tls={%p,%p}\n",
+    VReport(2, "__tls_get_addr: glibc <=2.18 suspected; tls={%p,%p}\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.
-    VPrintf(2, "__tls_get_addr: static tls: %p\n", tls_beg);
+    VReport(2, "__tls_get_addr: static tls: %p\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;
-    VPrintf(2, "__tls_get_addr: glibc >=2.19 suspected; tls={%p %p}\n",
+    VReport(2, "__tls_get_addr: glibc >=2.19 suspected; tls={%p %p}\n",
         tls_beg, tls_size);
   } else {
-    VPrintf(2, "__tls_get_addr: Can't guess glibc version\n");
+    VReport(2, "__tls_get_addr: Can't guess glibc version\n");
     // This may happen inside the DTOR of main thread, so just ignore it.
     tls_size = 0;
   }
@@ -129,7 +129,7 @@ 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;
-  VPrintf(2, "DTLS_on_libc_memalign: %p %p\n", ptr, size);
+  VReport(2, "DTLS_on_libc_memalign: %p %p\n", ptr, size);
   dtls.last_memalign_ptr = reinterpret_cast<uptr>(ptr);
   dtls.last_memalign_size = size;
 }
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_unwind_linux_libcdep.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_unwind_linux_libcdep.cc
index 5943125c898c..24c94700173d 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_unwind_linux_libcdep.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_unwind_linux_libcdep.cc
@@ -8,11 +8,12 @@
 //===----------------------------------------------------------------------===//
 //
 // This file contains the unwind.h-based (aka "slow") stack unwinding routines
-// available to the tools on Linux, Android, and FreeBSD.
+// available to the tools on Linux, Android, NetBSD, FreeBSD, and Solaris.
 //===----------------------------------------------------------------------===//
 
 #include "sanitizer_platform.h"
-#if SANITIZER_FREEBSD || SANITIZER_LINUX
+#if SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD || \
+    SANITIZER_SOLARIS
 #include "sanitizer_common.h"
 #include "sanitizer_stacktrace.h"
 
@@ -78,7 +79,8 @@ void SanitizerInitializeUnwinder() {
 }
 #endif
 
-#ifdef __arm__
+#if defined(__arm__) && !SANITIZER_NETBSD
+// NetBSD uses dwarf EH
 #define UNWIND_STOP _URC_END_OF_STACK
 #define UNWIND_CONTINUE _URC_NO_REASON
 #else
@@ -165,4 +167,5 @@ void BufferedStackTrace::SlowUnwindStackWithContext(uptr pc, void *context,
 
 }  // namespace __sanitizer
 
-#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX
+#endif  // SANITIZER_FREEBSD || SANITIZER_LINUX || SANITIZER_NETBSD ||
+        // SANITIZER_SOLARIS
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_vector.h b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_vector.h
index a7fb3fa58d54..25cfeed35f22 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_vector.h
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_vector.h
@@ -1,4 +1,4 @@
-//===-- tsan_vector.h -------------------------------------------*- C++ -*-===//
+//===-- sanitizer_vector.h -------------------------------------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -7,38 +7,37 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// This file is a part of ThreadSanitizer (TSan), a race detector.
+// This file is shared between sanitizers run-time libraries.
 //
 //===----------------------------------------------------------------------===//
 
 // Low-fat STL-like vector container.
 
-#ifndef TSAN_VECTOR_H
-#define TSAN_VECTOR_H
+#ifndef SANITIZER_VECTOR_H
+#define SANITIZER_VECTOR_H
 
-#include "tsan_defs.h"
-#include "tsan_mman.h"
+#include "sanitizer_common/sanitizer_allocator_internal.h"
+#include "sanitizer_common/sanitizer_libc.h"
 
-namespace __tsan {
+namespace __sanitizer {
 
 template<typename T>
 class Vector {
  public:
-  explicit Vector(MBlockType typ)
-      : typ_(typ)
-      , begin_()
+  explicit Vector()
+      : begin_()
       , end_()
       , last_() {
   }
 
   ~Vector() {
     if (begin_)
-      internal_free(begin_);
+      InternalFree(begin_);
   }
 
   void Reset() {
     if (begin_)
-      internal_free(begin_);
+      InternalFree(begin_);
     begin_ = 0;
     end_ = 0;
     last_ = 0;
@@ -91,7 +90,6 @@ class Vector {
   }
 
  private:
-  const MBlockType typ_;
   T *begin_;
   T *end_;
   T *last_;
@@ -109,10 +107,10 @@ class Vector {
       cap = 16;
     if (cap < size)
       cap = size;
-    T *p = (T*)internal_alloc(typ_, cap * sizeof(T));
+    T *p = (T*)InternalAlloc(cap * sizeof(T));
     if (cap0) {
       internal_memcpy(p, begin_, cap0 * sizeof(T));
-      internal_free(begin_);
+      InternalFree(begin_);
     }
     begin_ = p;
     end_ = begin_ + size;
@@ -122,6 +120,6 @@ class Vector {
   Vector(const Vector&);
   void operator=(const Vector&);
 };
-}  // namespace __tsan
+}  // namespace __sanitizer
 
-#endif  // #ifndef TSAN_VECTOR_H
+#endif  // #ifndef SANITIZER_VECTOR_H
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_win.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_win.cc
index de01e8d119a1..34bf1812d605 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_win.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_win.cc
@@ -24,10 +24,10 @@
 
 #include "sanitizer_common.h"
 #include "sanitizer_dbghelp.h"
+#include "sanitizer_file.h"
 #include "sanitizer_libc.h"
 #include "sanitizer_mutex.h"
 #include "sanitizer_placement_new.h"
-#include "sanitizer_procmaps.h"
 #include "sanitizer_stacktrace.h"
 #include "sanitizer_symbolizer.h"
 #include "sanitizer_win_defs.h"
@@ -64,12 +64,16 @@ uptr GetMmapGranularity() {
   return si.dwAllocationGranularity;
 }
 
-uptr GetMaxVirtualAddress() {
+uptr GetMaxUserVirtualAddress() {
   SYSTEM_INFO si;
   GetSystemInfo(&si);
   return (uptr)si.lpMaximumApplicationAddress;
 }
 
+uptr GetMaxVirtualAddress() {
+  return GetMaxUserVirtualAddress();
+}
+
 bool FileExists(const char *filename) {
   return ::GetFileAttributesA(filename) != INVALID_FILE_ATTRIBUTES;
 }
@@ -235,6 +239,28 @@ void *MmapFixedOrDie(uptr fixed_addr, uptr size) {
   return p;
 }
 
+// Uses fixed_addr for now.
+// Will use offset instead once we've implemented this function for real.
+uptr ReservedAddressRange::Map(uptr fixed_addr, uptr size) {
+  return reinterpret_cast<uptr>(MmapFixedOrDieOnFatalError(fixed_addr, size));
+}
+
+uptr ReservedAddressRange::MapOrDie(uptr fixed_addr, uptr size) {
+  return reinterpret_cast<uptr>(MmapFixedOrDie(fixed_addr, size));
+}
+
+void ReservedAddressRange::Unmap(uptr addr, uptr size) {
+  void* addr_as_void = reinterpret_cast<void*>(addr);
+  uptr base_as_uptr = reinterpret_cast<uptr>(base_);
+  // Only unmap if it covers the entire range.
+  CHECK((addr == base_as_uptr) && (size == size_));
+  UnmapOrDie(addr_as_void, size);
+  if (addr_as_void == base_) {
+    base_ = reinterpret_cast<void*>(addr + size);
+  }
+  size_ = size_ - size;
+}
+
 void *MmapFixedOrDieOnFatalError(uptr fixed_addr, uptr size) {
   void *p = VirtualAlloc((LPVOID)fixed_addr, size,
       MEM_COMMIT, PAGE_READWRITE);
@@ -252,6 +278,19 @@ void *MmapNoReserveOrDie(uptr size, const char *mem_type) {
   return MmapOrDie(size, mem_type);
 }
 
+uptr ReservedAddressRange::Init(uptr size, const char *name, uptr fixed_addr) {
+  if (fixed_addr) {
+    base_ = MmapFixedNoAccess(fixed_addr, size, name);
+  } else {
+    base_ = MmapNoAccess(size);
+  }
+  size_ = size;
+  name_ = name;
+  (void)os_handle_;  // unsupported
+  return reinterpret_cast<uptr>(base_);
+}
+
+
 void *MmapFixedNoAccess(uptr fixed_addr, uptr size, const char *name) {
   (void)name; // unsupported
   void *res = VirtualAlloc((LPVOID)fixed_addr, size,
@@ -379,7 +418,7 @@ struct ModuleInfo {
 
 #if !SANITIZER_GO
 int CompareModulesBase(const void *pl, const void *pr) {
-  const ModuleInfo *l = (ModuleInfo *)pl, *r = (ModuleInfo *)pr;
+  const ModuleInfo *l = (const ModuleInfo *)pl, *r = (const ModuleInfo *)pr;
   if (l->base_address < r->base_address)
     return -1;
   return l->base_address > r->base_address;
@@ -466,6 +505,10 @@ u64 NanoTime() {
   return 0;
 }
 
+u64 MonotonicNanoTime() {
+  return 0;
+}
+
 void Abort() {
   internal__exit(3);
 }
@@ -524,7 +567,7 @@ static uptr GetPreferredBase(const char *modname) {
 }
 
 void ListOfModules::init() {
-  clear();
+  clearOrInit();
   HANDLE cur_process = GetCurrentProcess();
 
   // Query the list of modules.  Start by assuming there are no more than 256
@@ -583,7 +626,9 @@ void ListOfModules::init() {
     modules_.push_back(cur_module);
   }
   UnmapOrDie(hmodules, modules_buffer_size);
-};
+}
+
+void ListOfModules::fallbackInit() { clear(); }
 
 // We can't use atexit() directly at __asan_init time as the CRT is not fully
 // initialized at this point.  Place the functions into a vector and use
@@ -792,7 +837,7 @@ void BufferedStackTrace::SlowUnwindStack(uptr pc, u32 max_depth) {
   // FIXME: Compare with StackWalk64.
   // FIXME: Look at LLVMUnhandledExceptionFilter in Signals.inc
   size = CaptureStackBackTrace(1, Min(max_depth, kStackTraceMax),
-                               (void**)trace, 0);
+                               (void **)&trace_buffer[0], 0);
   if (size == 0)
     return;
 
@@ -889,32 +934,6 @@ bool IsHandledDeadlyException(DWORD exceptionCode) {
   return false;
 }
 
-const char *DescribeSignalOrException(int signo) {
-  unsigned code = signo;
-  // Get the string description of the exception if this is a known deadly
-  // exception.
-  switch (code) {
-    case EXCEPTION_ACCESS_VIOLATION: return "access-violation";
-    case EXCEPTION_ARRAY_BOUNDS_EXCEEDED: return "array-bounds-exceeded";
-    case EXCEPTION_STACK_OVERFLOW: return "stack-overflow";
-    case EXCEPTION_DATATYPE_MISALIGNMENT: return "datatype-misalignment";
-    case EXCEPTION_IN_PAGE_ERROR: return "in-page-error";
-    case EXCEPTION_ILLEGAL_INSTRUCTION: return "illegal-instruction";
-    case EXCEPTION_PRIV_INSTRUCTION: return "priv-instruction";
-    case EXCEPTION_BREAKPOINT: return "breakpoint";
-    case EXCEPTION_FLT_DENORMAL_OPERAND: return "flt-denormal-operand";
-    case EXCEPTION_FLT_DIVIDE_BY_ZERO: return "flt-divide-by-zero";
-    case EXCEPTION_FLT_INEXACT_RESULT: return "flt-inexact-result";
-    case EXCEPTION_FLT_INVALID_OPERATION: return "flt-invalid-operation";
-    case EXCEPTION_FLT_OVERFLOW: return "flt-overflow";
-    case EXCEPTION_FLT_STACK_CHECK: return "flt-stack-check";
-    case EXCEPTION_FLT_UNDERFLOW: return "flt-underflow";
-    case EXCEPTION_INT_DIVIDE_BY_ZERO: return "int-divide-by-zero";
-    case EXCEPTION_INT_OVERFLOW: return "int-overflow";
-  }
-  return "unknown exception";
-}
-
 bool IsAccessibleMemoryRange(uptr beg, uptr size) {
   SYSTEM_INFO si;
   GetNativeSystemInfo(&si);
@@ -940,39 +959,101 @@ bool IsAccessibleMemoryRange(uptr beg, uptr size) {
   return true;
 }
 
-SignalContext SignalContext::Create(void *siginfo, void *context) {
+bool SignalContext::IsStackOverflow() const {
+  return (DWORD)GetType() == EXCEPTION_STACK_OVERFLOW;
+}
+
+void SignalContext::InitPcSpBp() {
   EXCEPTION_RECORD *exception_record = (EXCEPTION_RECORD *)siginfo;
   CONTEXT *context_record = (CONTEXT *)context;
 
-  uptr pc = (uptr)exception_record->ExceptionAddress;
+  pc = (uptr)exception_record->ExceptionAddress;
 #ifdef _WIN64
-  uptr bp = (uptr)context_record->Rbp;
-  uptr sp = (uptr)context_record->Rsp;
+  bp = (uptr)context_record->Rbp;
+  sp = (uptr)context_record->Rsp;
 #else
-  uptr bp = (uptr)context_record->Ebp;
-  uptr sp = (uptr)context_record->Esp;
+  bp = (uptr)context_record->Ebp;
+  sp = (uptr)context_record->Esp;
 #endif
-  uptr access_addr = exception_record->ExceptionInformation[1];
+}
+
+uptr SignalContext::GetAddress() const {
+  EXCEPTION_RECORD *exception_record = (EXCEPTION_RECORD *)siginfo;
+  return exception_record->ExceptionInformation[1];
+}
+
+bool SignalContext::IsMemoryAccess() const {
+  return GetWriteFlag() != SignalContext::UNKNOWN;
+}
 
+SignalContext::WriteFlag SignalContext::GetWriteFlag() const {
+  EXCEPTION_RECORD *exception_record = (EXCEPTION_RECORD *)siginfo;
   // The contents of this array are documented at
   // https://msdn.microsoft.com/en-us/library/windows/desktop/aa363082(v=vs.85).aspx
   // The first element indicates read as 0, write as 1, or execute as 8.  The
   // second element is the faulting address.
-  WriteFlag write_flag = SignalContext::UNKNOWN;
   switch (exception_record->ExceptionInformation[0]) {
-  case 0: write_flag = SignalContext::READ; break;
-  case 1: write_flag = SignalContext::WRITE; break;
-  case 8: write_flag = SignalContext::UNKNOWN; break;
+    case 0:
+      return SignalContext::READ;
+    case 1:
+      return SignalContext::WRITE;
+    case 8:
+      return SignalContext::UNKNOWN;
   }
-  bool is_memory_access = write_flag != SignalContext::UNKNOWN;
-  return SignalContext(context, access_addr, pc, sp, bp, is_memory_access,
-                       write_flag);
+  return SignalContext::UNKNOWN;
 }
 
 void SignalContext::DumpAllRegisters(void *context) {
   // FIXME: Implement this.
 }
 
+int SignalContext::GetType() const {
+  return static_cast<const EXCEPTION_RECORD *>(siginfo)->ExceptionCode;
+}
+
+const char *SignalContext::Describe() const {
+  unsigned code = GetType();
+  // Get the string description of the exception if this is a known deadly
+  // exception.
+  switch (code) {
+    case EXCEPTION_ACCESS_VIOLATION:
+      return "access-violation";
+    case EXCEPTION_ARRAY_BOUNDS_EXCEEDED:
+      return "array-bounds-exceeded";
+    case EXCEPTION_STACK_OVERFLOW:
+      return "stack-overflow";
+    case EXCEPTION_DATATYPE_MISALIGNMENT:
+      return "datatype-misalignment";
+    case EXCEPTION_IN_PAGE_ERROR:
+      return "in-page-error";
+    case EXCEPTION_ILLEGAL_INSTRUCTION:
+      return "illegal-instruction";
+    case EXCEPTION_PRIV_INSTRUCTION:
+      return "priv-instruction";
+    case EXCEPTION_BREAKPOINT:
+      return "breakpoint";
+    case EXCEPTION_FLT_DENORMAL_OPERAND:
+      return "flt-denormal-operand";
+    case EXCEPTION_FLT_DIVIDE_BY_ZERO:
+      return "flt-divide-by-zero";
+    case EXCEPTION_FLT_INEXACT_RESULT:
+      return "flt-inexact-result";
+    case EXCEPTION_FLT_INVALID_OPERATION:
+      return "flt-invalid-operation";
+    case EXCEPTION_FLT_OVERFLOW:
+      return "flt-overflow";
+    case EXCEPTION_FLT_STACK_CHECK:
+      return "flt-stack-check";
+    case EXCEPTION_FLT_UNDERFLOW:
+      return "flt-underflow";
+    case EXCEPTION_INT_DIVIDE_BY_ZERO:
+      return "int-divide-by-zero";
+    case EXCEPTION_INT_OVERFLOW:
+      return "int-overflow";
+  }
+  return "unknown exception";
+}
+
 uptr ReadBinaryName(/*out*/char *buf, uptr buf_len) {
   // FIXME: Actually implement this function.
   CHECK_GT(buf_len, 0);
@@ -1021,7 +1102,12 @@ void CheckNoDeepBind(const char *filename, int flag) {
 }
 
 // FIXME: implement on this platform.
-bool GetRandom(void *buffer, uptr length) {
+bool GetRandom(void *buffer, uptr length, bool blocking) {
+  UNIMPLEMENTED();
+}
+
+// FIXME: implement on this platform.
+u32 GetNumberOfCPUs() {
   UNIMPLEMENTED();
 }
 
diff --git a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_win_weak_interception.cc b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_win_weak_interception.cc
index 364319398198..5711f5dc8b6a 100644
--- a/contrib/compiler-rt/lib/sanitizer_common/sanitizer_win_weak_interception.cc
+++ b/contrib/compiler-rt/lib/sanitizer_common/sanitizer_win_weak_interception.cc
@@ -12,7 +12,7 @@
 // definition is provided.
 //===----------------------------------------------------------------------===//
 
-#include "sanitizer_common/sanitizer_platform.h"
+#include "sanitizer_platform.h"
 #if SANITIZER_WINDOWS && SANITIZER_DYNAMIC
 #include "sanitizer_win_weak_interception.h"
 #include "sanitizer_allocator_interface.h"
diff --git a/contrib/compiler-rt/lib/sanitizer_common/symbolizer/scripts/build_symbolizer.sh b/contrib/compiler-rt/lib/sanitizer_common/symbolizer/scripts/build_symbolizer.sh
index c5865ecfee6c..0559a2e7eb05 100755
--- a/contrib/compiler-rt/lib/sanitizer_common/symbolizer/scripts/build_symbolizer.sh
+++ b/contrib/compiler-rt/lib/sanitizer_common/symbolizer/scripts/build_symbolizer.sh
@@ -110,12 +110,12 @@ 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"
 
 # Build LLVM.
 if [[ ! -d ${LLVM_BUILD} ]]; then
   mkdir -p ${LLVM_BUILD}
   cd ${LLVM_BUILD}
-  LLVM_FLAGS="${FLAGS} -I${ZLIB_BUILD} -I${LIBCXX_BUILD}/include/c++/v1"
   cmake -GNinja \
     -DCMAKE_BUILD_TYPE=Release \
     -DCMAKE_C_COMPILER=$CC \
@@ -137,7 +137,7 @@ mkdir ${SYMBOLIZER_BUILD}
 cd ${SYMBOLIZER_BUILD}
 
 echo "Compiling..."
-SYMBOLIZER_FLAGS="$FLAGS -std=c++11 -I${LLVM_SRC}/include -I${LLVM_BUILD}/include -I${LIBCXX_BUILD}/include/c++/v1"
+SYMBOLIZER_FLAGS="$LLVM_FLAGS -I${LLVM_SRC}/include -I${LLVM_BUILD}/include -std=c++11"
 $CXX $SYMBOLIZER_FLAGS ${SRC_DIR}/sanitizer_symbolize.cc ${SRC_DIR}/sanitizer_wrappers.cc -c
 $AR rc symbolizer.a sanitizer_symbolize.o sanitizer_wrappers.o
 
diff --git a/contrib/compiler-rt/lib/scudo/scudo_allocator.cpp b/contrib/compiler-rt/lib/scudo/scudo_allocator.cpp
index 6f30ee987513..e5a4d714c66e 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_allocator.cpp
+++ b/contrib/compiler-rt/lib/scudo/scudo_allocator.cpp
@@ -16,26 +16,27 @@
 
 #include "scudo_allocator.h"
 #include "scudo_crc32.h"
-#include "scudo_tls.h"
+#include "scudo_flags.h"
+#include "scudo_tsd.h"
 #include "scudo_utils.h"
 
 #include "sanitizer_common/sanitizer_allocator_checks.h"
 #include "sanitizer_common/sanitizer_allocator_interface.h"
-#include "sanitizer_common/sanitizer_errno.h"
 #include "sanitizer_common/sanitizer_quarantine.h"
 
+#include <errno.h>
 #include <string.h>
 
 namespace __scudo {
 
 // Global static cookie, initialized at start-up.
-static uptr Cookie;
+static u32 Cookie;
 
 // We default to software CRC32 if the alternatives are not supported, either
 // at compilation or at runtime.
 static atomic_uint8_t HashAlgorithm = { CRC32Software };
 
-INLINE u32 computeCRC32(uptr Crc, uptr Value, uptr *Array, uptr ArraySize) {
+INLINE u32 computeCRC32(u32 Crc, uptr Value, uptr *Array, uptr ArraySize) {
   // If the hardware CRC32 feature is defined here, it was enabled everywhere,
   // as opposed to only for scudo_crc32.cpp. This means that other hardware
   // specific instructions were likely emitted at other places, and as a
@@ -61,46 +62,60 @@ INLINE u32 computeCRC32(uptr Crc, uptr Value, uptr *Array, uptr ArraySize) {
 
 static ScudoBackendAllocator &getBackendAllocator();
 
-struct ScudoChunk : UnpackedHeader {
+namespace Chunk {
   // We can't use the offset member of the chunk itself, as we would double
   // fetch it without any warranty that it wouldn't have been tampered. To
   // prevent this, we work with a local copy of the header.
-  void *getAllocBeg(UnpackedHeader *Header) {
-    return reinterpret_cast<void *>(
-        reinterpret_cast<uptr>(this) - (Header->Offset << MinAlignmentLog));
+  static INLINE void *getBackendPtr(const void *Ptr, UnpackedHeader *Header) {
+    return reinterpret_cast<void *>(reinterpret_cast<uptr>(Ptr) -
+                                    AlignedChunkHeaderSize -
+                                    (Header->Offset << MinAlignmentLog));
+  }
+
+  static INLINE AtomicPackedHeader *getAtomicHeader(void *Ptr) {
+    return reinterpret_cast<AtomicPackedHeader *>(reinterpret_cast<uptr>(Ptr) -
+                                                  AlignedChunkHeaderSize);
+  }
+  static INLINE
+  const AtomicPackedHeader *getConstAtomicHeader(const void *Ptr) {
+    return reinterpret_cast<const AtomicPackedHeader *>(
+        reinterpret_cast<uptr>(Ptr) - AlignedChunkHeaderSize);
+  }
+
+  static INLINE bool isAligned(const void *Ptr) {
+    return IsAligned(reinterpret_cast<uptr>(Ptr), MinAlignment);
   }
 
   // Returns the usable size for a chunk, meaning the amount of bytes from the
   // beginning of the user data to the end of the backend allocated chunk.
-  uptr getUsableSize(UnpackedHeader *Header) {
-    uptr Size =
-        getBackendAllocator().getActuallyAllocatedSize(getAllocBeg(Header),
-                                                       Header->FromPrimary);
+  static INLINE uptr getUsableSize(const void *Ptr, UnpackedHeader *Header) {
+    const uptr Size = getBackendAllocator().getActuallyAllocatedSize(
+        getBackendPtr(Ptr, Header), Header->ClassId);
     if (Size == 0)
       return 0;
     return Size - AlignedChunkHeaderSize - (Header->Offset << MinAlignmentLog);
   }
 
-  // Compute the checksum of the Chunk pointer and its ChunkHeader.
-  u16 computeChecksum(UnpackedHeader *Header) const {
+  // Compute the checksum of the chunk pointer and its header.
+  static INLINE u16 computeChecksum(const void *Ptr, UnpackedHeader *Header) {
     UnpackedHeader ZeroChecksumHeader = *Header;
     ZeroChecksumHeader.Checksum = 0;
     uptr HeaderHolder[sizeof(UnpackedHeader) / sizeof(uptr)];
     memcpy(&HeaderHolder, &ZeroChecksumHeader, sizeof(HeaderHolder));
-    u32 Crc = computeCRC32(Cookie, reinterpret_cast<uptr>(this), HeaderHolder,
-                           ARRAY_SIZE(HeaderHolder));
+    const u32 Crc = computeCRC32(Cookie, reinterpret_cast<uptr>(Ptr),
+                                 HeaderHolder, ARRAY_SIZE(HeaderHolder));
     return static_cast<u16>(Crc);
   }
 
   // Checks the validity of a chunk by verifying its checksum. It doesn't
   // incur termination in the event of an invalid chunk.
-  bool isValid() {
-    UnpackedHeader NewUnpackedHeader;
-    const AtomicPackedHeader *AtomicHeader =
-        reinterpret_cast<const AtomicPackedHeader *>(this);
-    PackedHeader NewPackedHeader = atomic_load_relaxed(AtomicHeader);
-    NewUnpackedHeader = bit_cast<UnpackedHeader>(NewPackedHeader);
-    return (NewUnpackedHeader.Checksum == computeChecksum(&NewUnpackedHeader));
+  static INLINE bool isValid(const void *Ptr) {
+    PackedHeader NewPackedHeader =
+        atomic_load_relaxed(getConstAtomicHeader(Ptr));
+    UnpackedHeader NewUnpackedHeader =
+        bit_cast<UnpackedHeader>(NewPackedHeader);
+    return (NewUnpackedHeader.Checksum ==
+            computeChecksum(Ptr, &NewUnpackedHeader));
   }
 
   // Nulls out a chunk header. When returning the chunk to the backend, there
@@ -109,114 +124,46 @@ struct ScudoChunk : UnpackedHeader {
   // the header invalid. In the extremely rare event where 0 would be a valid
   // checksum for the chunk, the state of the chunk is ChunkAvailable anyway.
   COMPILER_CHECK(ChunkAvailable == 0);
-  void eraseHeader() {
-    PackedHeader NullPackedHeader = 0;
-    AtomicPackedHeader *AtomicHeader =
-        reinterpret_cast<AtomicPackedHeader *>(this);
-    atomic_store_relaxed(AtomicHeader, NullPackedHeader);
+  static INLINE void eraseHeader(void *Ptr) {
+    const PackedHeader NullPackedHeader = 0;
+    atomic_store_relaxed(getAtomicHeader(Ptr), NullPackedHeader);
   }
 
   // Loads and unpacks the header, verifying the checksum in the process.
-  void loadHeader(UnpackedHeader *NewUnpackedHeader) const {
-    const AtomicPackedHeader *AtomicHeader =
-        reinterpret_cast<const AtomicPackedHeader *>(this);
-    PackedHeader NewPackedHeader = atomic_load_relaxed(AtomicHeader);
+  static INLINE
+  void loadHeader(const void *Ptr, UnpackedHeader *NewUnpackedHeader) {
+    PackedHeader NewPackedHeader =
+        atomic_load_relaxed(getConstAtomicHeader(Ptr));
     *NewUnpackedHeader = bit_cast<UnpackedHeader>(NewPackedHeader);
     if (UNLIKELY(NewUnpackedHeader->Checksum !=
-        computeChecksum(NewUnpackedHeader))) {
-      dieWithMessage("ERROR: corrupted chunk header at address %p\n", this);
+        computeChecksum(Ptr, NewUnpackedHeader))) {
+      dieWithMessage("ERROR: corrupted chunk header at address %p\n", Ptr);
     }
   }
 
   // Packs and stores the header, computing the checksum in the process.
-  void storeHeader(UnpackedHeader *NewUnpackedHeader) {
-    NewUnpackedHeader->Checksum = computeChecksum(NewUnpackedHeader);
+  static INLINE void storeHeader(void *Ptr, UnpackedHeader *NewUnpackedHeader) {
+    NewUnpackedHeader->Checksum = computeChecksum(Ptr, NewUnpackedHeader);
     PackedHeader NewPackedHeader = bit_cast<PackedHeader>(*NewUnpackedHeader);
-    AtomicPackedHeader *AtomicHeader =
-        reinterpret_cast<AtomicPackedHeader *>(this);
-    atomic_store_relaxed(AtomicHeader, NewPackedHeader);
+    atomic_store_relaxed(getAtomicHeader(Ptr), NewPackedHeader);
   }
 
   // Packs and stores the header, computing the checksum in the process. We
   // compare the current header with the expected provided one to ensure that
   // we are not being raced by a corruption occurring in another thread.
-  void compareExchangeHeader(UnpackedHeader *NewUnpackedHeader,
-                             UnpackedHeader *OldUnpackedHeader) {
-    NewUnpackedHeader->Checksum = computeChecksum(NewUnpackedHeader);
+  static INLINE void compareExchangeHeader(void *Ptr,
+                                           UnpackedHeader *NewUnpackedHeader,
+                                           UnpackedHeader *OldUnpackedHeader) {
+    NewUnpackedHeader->Checksum = computeChecksum(Ptr, NewUnpackedHeader);
     PackedHeader NewPackedHeader = bit_cast<PackedHeader>(*NewUnpackedHeader);
     PackedHeader OldPackedHeader = bit_cast<PackedHeader>(*OldUnpackedHeader);
-    AtomicPackedHeader *AtomicHeader =
-        reinterpret_cast<AtomicPackedHeader *>(this);
-    if (UNLIKELY(!atomic_compare_exchange_strong(AtomicHeader,
-                                                 &OldPackedHeader,
-                                                 NewPackedHeader,
-                                                 memory_order_relaxed))) {
-      dieWithMessage("ERROR: race on chunk header at address %p\n", this);
+    if (UNLIKELY(!atomic_compare_exchange_strong(
+            getAtomicHeader(Ptr), &OldPackedHeader, NewPackedHeader,
+            memory_order_relaxed))) {
+      dieWithMessage("ERROR: race on chunk header at address %p\n", Ptr);
     }
   }
-};
-
-ScudoChunk *getScudoChunk(uptr UserBeg) {
-  return reinterpret_cast<ScudoChunk *>(UserBeg - AlignedChunkHeaderSize);
-}
-
-struct AllocatorOptions {
-  u32 QuarantineSizeMb;
-  u32 ThreadLocalQuarantineSizeKb;
-  bool MayReturnNull;
-  s32 ReleaseToOSIntervalMs;
-  bool DeallocationTypeMismatch;
-  bool DeleteSizeMismatch;
-  bool ZeroContents;
-
-  void setFrom(const Flags *f, const CommonFlags *cf);
-  void copyTo(Flags *f, CommonFlags *cf) const;
-};
-
-void AllocatorOptions::setFrom(const Flags *f, const CommonFlags *cf) {
-  MayReturnNull = cf->allocator_may_return_null;
-  ReleaseToOSIntervalMs = cf->allocator_release_to_os_interval_ms;
-  QuarantineSizeMb = f->QuarantineSizeMb;
-  ThreadLocalQuarantineSizeKb = f->ThreadLocalQuarantineSizeKb;
-  DeallocationTypeMismatch = f->DeallocationTypeMismatch;
-  DeleteSizeMismatch = f->DeleteSizeMismatch;
-  ZeroContents = f->ZeroContents;
-}
-
-void AllocatorOptions::copyTo(Flags *f, CommonFlags *cf) const {
-  cf->allocator_may_return_null = MayReturnNull;
-  cf->allocator_release_to_os_interval_ms = ReleaseToOSIntervalMs;
-  f->QuarantineSizeMb = QuarantineSizeMb;
-  f->ThreadLocalQuarantineSizeKb = ThreadLocalQuarantineSizeKb;
-  f->DeallocationTypeMismatch = DeallocationTypeMismatch;
-  f->DeleteSizeMismatch = DeleteSizeMismatch;
-  f->ZeroContents = ZeroContents;
-}
-
-static void initScudoInternal(const AllocatorOptions &Options);
-
-static bool ScudoInitIsRunning = false;
-
-void initScudo() {
-  SanitizerToolName = "Scudo";
-  CHECK(!ScudoInitIsRunning && "Scudo init calls itself!");
-  ScudoInitIsRunning = true;
-
-  // Check if hardware CRC32 is supported in the binary and by the platform, if
-  // so, opt for the CRC32 hardware version of the checksum.
-  if (computeHardwareCRC32 && testCPUFeature(CRC32CPUFeature))
-    atomic_store_relaxed(&HashAlgorithm, CRC32Hardware);
-
-  initFlags();
-
-  AllocatorOptions Options;
-  Options.setFrom(getFlags(), common_flags());
-  initScudoInternal(Options);
-
-  // TODO(kostyak): determine if MaybeStartBackgroudThread could be of some use.
-
-  ScudoInitIsRunning = false;
-}
+}  // namespace Chunk
 
 struct QuarantineCallback {
   explicit QuarantineCallback(AllocatorCache *Cache)
@@ -224,52 +171,46 @@ struct QuarantineCallback {
 
   // Chunk recycling function, returns a quarantined chunk to the backend,
   // first making sure it hasn't been tampered with.
-  void Recycle(ScudoChunk *Chunk) {
+  void Recycle(void *Ptr) {
     UnpackedHeader Header;
-    Chunk->loadHeader(&Header);
+    Chunk::loadHeader(Ptr, &Header);
     if (UNLIKELY(Header.State != ChunkQuarantine)) {
       dieWithMessage("ERROR: invalid chunk state when recycling address %p\n",
-                     Chunk);
+                     Ptr);
     }
-    Chunk->eraseHeader();
-    void *Ptr = Chunk->getAllocBeg(&Header);
-    if (Header.FromPrimary)
-      getBackendAllocator().deallocatePrimary(Cache_, Ptr);
+    Chunk::eraseHeader(Ptr);
+    void *BackendPtr = Chunk::getBackendPtr(Ptr, &Header);
+    if (Header.ClassId)
+      getBackendAllocator().deallocatePrimary(Cache_, BackendPtr,
+                                              Header.ClassId);
     else
-      getBackendAllocator().deallocateSecondary(Ptr);
+      getBackendAllocator().deallocateSecondary(BackendPtr);
   }
 
   // Internal quarantine allocation and deallocation functions. We first check
   // that the batches are indeed serviced by the Primary.
   // TODO(kostyak): figure out the best way to protect the batches.
-  COMPILER_CHECK(sizeof(QuarantineBatch) < SizeClassMap::kMaxSize);
   void *Allocate(uptr Size) {
-    return getBackendAllocator().allocatePrimary(Cache_, Size);
+    return getBackendAllocator().allocatePrimary(Cache_, BatchClassId);
   }
 
   void Deallocate(void *Ptr) {
-    getBackendAllocator().deallocatePrimary(Cache_, Ptr);
+    getBackendAllocator().deallocatePrimary(Cache_, Ptr, BatchClassId);
   }
 
   AllocatorCache *Cache_;
+  COMPILER_CHECK(sizeof(QuarantineBatch) < SizeClassMap::kMaxSize);
+  const uptr BatchClassId = SizeClassMap::ClassID(sizeof(QuarantineBatch));
 };
 
-typedef Quarantine<QuarantineCallback, ScudoChunk> ScudoQuarantine;
+typedef Quarantine<QuarantineCallback, void> ScudoQuarantine;
 typedef ScudoQuarantine::Cache ScudoQuarantineCache;
 COMPILER_CHECK(sizeof(ScudoQuarantineCache) <=
-               sizeof(ScudoThreadContext::QuarantineCachePlaceHolder));
+               sizeof(ScudoTSD::QuarantineCachePlaceHolder));
 
-AllocatorCache *getAllocatorCache(ScudoThreadContext *ThreadContext) {
-  return &ThreadContext->Cache;
-}
-
-ScudoQuarantineCache *getQuarantineCache(ScudoThreadContext *ThreadContext) {
-  return reinterpret_cast<
-      ScudoQuarantineCache *>(ThreadContext->QuarantineCachePlaceHolder);
-}
-
-ScudoPrng *getPrng(ScudoThreadContext *ThreadContext) {
-  return &ThreadContext->Prng;
+ScudoQuarantineCache *getQuarantineCache(ScudoTSD *TSD) {
+  return reinterpret_cast<ScudoQuarantineCache *>(
+      TSD->QuarantineCachePlaceHolder);
 }
 
 struct ScudoAllocator {
@@ -281,26 +222,22 @@ struct ScudoAllocator {
   ScudoBackendAllocator BackendAllocator;
   ScudoQuarantine AllocatorQuarantine;
 
-  StaticSpinMutex GlobalPrngMutex;
-  ScudoPrng GlobalPrng;
-
-  // The fallback caches are used when the thread local caches have been
-  // 'detroyed' on thread tear-down. They are protected by a Mutex as they can
-  // be accessed by different threads.
-  StaticSpinMutex FallbackMutex;
-  AllocatorCache FallbackAllocatorCache;
-  ScudoQuarantineCache FallbackQuarantineCache;
-  ScudoPrng FallbackPrng;
+  u32 QuarantineChunksUpToSize;
 
   bool DeallocationTypeMismatch;
   bool ZeroContents;
   bool DeleteSizeMismatch;
 
+  bool CheckRssLimit;
+  uptr HardRssLimitMb;
+  uptr SoftRssLimitMb;
+  atomic_uint8_t RssLimitExceeded;
+  atomic_uint64_t RssLastCheckedAtNS;
+
   explicit ScudoAllocator(LinkerInitialized)
-    : AllocatorQuarantine(LINKER_INITIALIZED),
-      FallbackQuarantineCache(LINKER_INITIALIZED) {}
+    : AllocatorQuarantine(LINKER_INITIALIZED) {}
 
-  void init(const AllocatorOptions &Options) {
+  void performSanityChecks() {
     // Verify that the header offset field can hold the maximum offset. In the
     // case of the Secondary allocator, it takes care of alignment and the
     // offset will always be 0. In the case of the Primary, the worst case
@@ -310,9 +247,9 @@ struct ScudoAllocator {
     // result, the maximum offset will be at most the maximum alignment for the
     // last size class minus the header size, in multiples of MinAlignment.
     UnpackedHeader Header = {};
-    uptr MaxPrimaryAlignment =
+    const uptr MaxPrimaryAlignment =
         1 << MostSignificantSetBitIndex(SizeClassMap::kMaxSize - MinAlignment);
-    uptr MaxOffset =
+    const uptr MaxOffset =
         (MaxPrimaryAlignment - AlignedChunkHeaderSize) >> MinAlignmentLog;
     Header.Offset = MaxOffset;
     if (Header.Offset != MaxOffset) {
@@ -324,36 +261,97 @@ struct ScudoAllocator {
     // case scenario happens in the Primary. It will depend on the second to
     // last and last class sizes, as well as the dynamic base for the Primary.
     // The following is an over-approximation that works for our needs.
-    uptr MaxSizeOrUnusedBytes = SizeClassMap::kMaxSize - 1;
+    const uptr MaxSizeOrUnusedBytes = SizeClassMap::kMaxSize - 1;
     Header.SizeOrUnusedBytes = MaxSizeOrUnusedBytes;
     if (Header.SizeOrUnusedBytes != MaxSizeOrUnusedBytes) {
       dieWithMessage("ERROR: the maximum possible unused bytes doesn't fit in "
                      "the header\n");
     }
 
-    DeallocationTypeMismatch = Options.DeallocationTypeMismatch;
-    DeleteSizeMismatch = Options.DeleteSizeMismatch;
-    ZeroContents = Options.ZeroContents;
-    SetAllocatorMayReturnNull(Options.MayReturnNull);
-    BackendAllocator.init(Options.ReleaseToOSIntervalMs);
+    const uptr LargestClassId = SizeClassMap::kLargestClassID;
+    Header.ClassId = LargestClassId;
+    if (Header.ClassId != LargestClassId) {
+      dieWithMessage("ERROR: the largest class ID doesn't fit in the header\n");
+    }
+  }
+
+  void init() {
+    SanitizerToolName = "Scudo";
+    initFlags();
+
+    performSanityChecks();
+
+    // Check if hardware CRC32 is supported in the binary and by the platform,
+    // if so, opt for the CRC32 hardware version of the checksum.
+    if (&computeHardwareCRC32 && hasHardwareCRC32())
+      atomic_store_relaxed(&HashAlgorithm, CRC32Hardware);
+
+    SetAllocatorMayReturnNull(common_flags()->allocator_may_return_null);
+    BackendAllocator.init(common_flags()->allocator_release_to_os_interval_ms);
+    HardRssLimitMb = common_flags()->hard_rss_limit_mb;
+    SoftRssLimitMb = common_flags()->soft_rss_limit_mb;
     AllocatorQuarantine.Init(
-        static_cast<uptr>(Options.QuarantineSizeMb) << 20,
-        static_cast<uptr>(Options.ThreadLocalQuarantineSizeKb) << 10);
-    GlobalPrng.init();
-    Cookie = GlobalPrng.getU64();
-    BackendAllocator.initCache(&FallbackAllocatorCache);
-    FallbackPrng.init();
+        static_cast<uptr>(getFlags()->QuarantineSizeKb) << 10,
+        static_cast<uptr>(getFlags()->ThreadLocalQuarantineSizeKb) << 10);
+    QuarantineChunksUpToSize = getFlags()->QuarantineChunksUpToSize;
+    DeallocationTypeMismatch = getFlags()->DeallocationTypeMismatch;
+    DeleteSizeMismatch = getFlags()->DeleteSizeMismatch;
+    ZeroContents = getFlags()->ZeroContents;
+
+    if (UNLIKELY(!GetRandom(reinterpret_cast<void *>(&Cookie), sizeof(Cookie),
+                            /*blocking=*/false))) {
+      Cookie = static_cast<u32>((NanoTime() >> 12) ^
+                                (reinterpret_cast<uptr>(this) >> 4));
+    }
+
+    CheckRssLimit = HardRssLimitMb || SoftRssLimitMb;
+    if (CheckRssLimit)
+      atomic_store_relaxed(&RssLastCheckedAtNS, MonotonicNanoTime());
   }
 
   // Helper function that checks for a valid Scudo chunk. nullptr isn't.
-  bool isValidPointer(const void *UserPtr) {
+  bool isValidPointer(const void *Ptr) {
     initThreadMaybe();
-    if (UNLIKELY(!UserPtr))
+    if (UNLIKELY(!Ptr))
       return false;
-    uptr UserBeg = reinterpret_cast<uptr>(UserPtr);
-    if (!IsAligned(UserBeg, MinAlignment))
+    if (!Chunk::isAligned(Ptr))
       return false;
-    return getScudoChunk(UserBeg)->isValid();
+    return Chunk::isValid(Ptr);
+  }
+
+  // Opportunistic RSS limit check. This will update the RSS limit status, if
+  // it can, every 100ms, otherwise it will just return the current one.
+  bool isRssLimitExceeded() {
+    u64 LastCheck = atomic_load_relaxed(&RssLastCheckedAtNS);
+    const u64 CurrentCheck = MonotonicNanoTime();
+    if (LIKELY(CurrentCheck < LastCheck + (100ULL * 1000000ULL)))
+      return atomic_load_relaxed(&RssLimitExceeded);
+    if (!atomic_compare_exchange_weak(&RssLastCheckedAtNS, &LastCheck,
+                                      CurrentCheck, memory_order_relaxed))
+      return atomic_load_relaxed(&RssLimitExceeded);
+    // TODO(kostyak): We currently use sanitizer_common's GetRSS which reads the
+    //                RSS from /proc/self/statm by default. We might want to
+    //                call getrusage directly, even if it's less accurate.
+    const uptr CurrentRssMb = GetRSS() >> 20;
+    if (HardRssLimitMb && HardRssLimitMb < CurrentRssMb) {
+      Report("%s: hard RSS limit exhausted (%zdMb vs %zdMb)\n",
+             SanitizerToolName, HardRssLimitMb, CurrentRssMb);
+      DumpProcessMap();
+      Die();
+    }
+    if (SoftRssLimitMb) {
+      if (atomic_load_relaxed(&RssLimitExceeded)) {
+        if (CurrentRssMb <= SoftRssLimitMb)
+          atomic_store_relaxed(&RssLimitExceeded, false);
+      } else {
+        if (CurrentRssMb > SoftRssLimitMb) {
+          atomic_store_relaxed(&RssLimitExceeded, true);
+          Report("%s: soft RSS limit exhausted (%zdMb vs %zdMb)\n",
+                 SanitizerToolName, SoftRssLimitMb, CurrentRssMb);
+        }
+      }
+    }
+    return atomic_load_relaxed(&RssLimitExceeded);
   }
 
   // Allocates a chunk.
@@ -375,207 +373,185 @@ struct ScudoAllocator {
     if (UNLIKELY(AlignedSize >= MaxAllowedMallocSize))
       return FailureHandler::OnBadRequest();
 
+    if (CheckRssLimit && UNLIKELY(isRssLimitExceeded()))
+      return FailureHandler::OnOOM();
+
     // Primary and Secondary backed allocations have a different treatment. We
     // deal with alignment requirements of Primary serviced allocations here,
     // but the Secondary will take care of its own alignment needs.
-    bool FromPrimary = PrimaryAllocator::CanAllocate(AlignedSize, MinAlignment);
-
-    void *Ptr;
-    u8 Salt;
-    uptr AllocSize;
-    if (FromPrimary) {
-      AllocSize = AlignedSize;
-      ScudoThreadContext *ThreadContext = getThreadContextAndLock();
-      if (LIKELY(ThreadContext)) {
-        Salt = getPrng(ThreadContext)->getU8();
-        Ptr = BackendAllocator.allocatePrimary(getAllocatorCache(ThreadContext),
-                                               AllocSize);
-        ThreadContext->unlock();
-      } else {
-        SpinMutexLock l(&FallbackMutex);
-        Salt = FallbackPrng.getU8();
-        Ptr = BackendAllocator.allocatePrimary(&FallbackAllocatorCache,
-                                               AllocSize);
-      }
+    void *BackendPtr;
+    uptr BackendSize;
+    u8 ClassId;
+    if (PrimaryAllocator::CanAllocate(AlignedSize, MinAlignment)) {
+      BackendSize = AlignedSize;
+      ClassId = SizeClassMap::ClassID(BackendSize);
+      ScudoTSD *TSD = getTSDAndLock();
+      BackendPtr = BackendAllocator.allocatePrimary(&TSD->Cache, ClassId);
+      TSD->unlock();
     } else {
-      {
-        SpinMutexLock l(&GlobalPrngMutex);
-        Salt = GlobalPrng.getU8();
-      }
-      AllocSize = NeededSize;
-      Ptr = BackendAllocator.allocateSecondary(AllocSize, Alignment);
+      BackendSize = NeededSize;
+      ClassId = 0;
+      BackendPtr = BackendAllocator.allocateSecondary(BackendSize, Alignment);
     }
-    if (UNLIKELY(!Ptr))
+    if (UNLIKELY(!BackendPtr))
       return FailureHandler::OnOOM();
 
     // If requested, we will zero out the entire contents of the returned chunk.
-    if ((ForceZeroContents || ZeroContents) && FromPrimary)
-      memset(Ptr, 0, BackendAllocator.getActuallyAllocatedSize(
-          Ptr, /*FromPrimary=*/true));
+    if ((ForceZeroContents || ZeroContents) && ClassId)
+      memset(BackendPtr, 0,
+             BackendAllocator.getActuallyAllocatedSize(BackendPtr, ClassId));
 
     UnpackedHeader Header = {};
-    uptr AllocBeg = reinterpret_cast<uptr>(Ptr);
-    uptr UserBeg = AllocBeg + AlignedChunkHeaderSize;
-    if (UNLIKELY(!IsAligned(UserBeg, Alignment))) {
+    uptr UserPtr = reinterpret_cast<uptr>(BackendPtr) + AlignedChunkHeaderSize;
+    if (UNLIKELY(!IsAligned(UserPtr, Alignment))) {
       // Since the Secondary takes care of alignment, a non-aligned pointer
       // means it is from the Primary. It is also the only case where the offset
       // field of the header would be non-zero.
-      CHECK(FromPrimary);
-      UserBeg = RoundUpTo(UserBeg, Alignment);
-      uptr Offset = UserBeg - AlignedChunkHeaderSize - AllocBeg;
-      Header.Offset = Offset >> MinAlignmentLog;
+      DCHECK(ClassId);
+      const uptr AlignedUserPtr = RoundUpTo(UserPtr, Alignment);
+      Header.Offset = (AlignedUserPtr - UserPtr) >> MinAlignmentLog;
+      UserPtr = AlignedUserPtr;
     }
-    CHECK_LE(UserBeg + Size, AllocBeg + AllocSize);
+    CHECK_LE(UserPtr + Size, reinterpret_cast<uptr>(BackendPtr) + BackendSize);
     Header.State = ChunkAllocated;
     Header.AllocType = Type;
-    if (FromPrimary) {
-      Header.FromPrimary = 1;
+    if (ClassId) {
+      Header.ClassId = ClassId;
       Header.SizeOrUnusedBytes = Size;
     } else {
       // The secondary fits the allocations to a page, so the amount of unused
       // bytes is the difference between the end of the user allocation and the
       // next page boundary.
-      uptr PageSize = GetPageSizeCached();
-      uptr TrailingBytes = (UserBeg + Size) & (PageSize - 1);
+      const uptr PageSize = GetPageSizeCached();
+      const uptr TrailingBytes = (UserPtr + Size) & (PageSize - 1);
       if (TrailingBytes)
         Header.SizeOrUnusedBytes = PageSize - TrailingBytes;
     }
-    Header.Salt = Salt;
-    getScudoChunk(UserBeg)->storeHeader(&Header);
-    void *UserPtr = reinterpret_cast<void *>(UserBeg);
-    // if (&__sanitizer_malloc_hook) __sanitizer_malloc_hook(UserPtr, Size);
-    return UserPtr;
+    void *Ptr = reinterpret_cast<void *>(UserPtr);
+    Chunk::storeHeader(Ptr, &Header);
+    // if (&__sanitizer_malloc_hook) __sanitizer_malloc_hook(Ptr, Size);
+    return Ptr;
   }
 
-  // Place a chunk in the quarantine. In the event of a zero-sized quarantine,
-  // we directly deallocate the chunk, otherwise the flow would lead to the
-  // chunk being loaded (and checked) twice, and stored (and checksummed) once,
-  // with no additional security value.
-  void quarantineOrDeallocateChunk(ScudoChunk *Chunk, UnpackedHeader *Header,
+  // Place a chunk in the quarantine or directly deallocate it in the event of
+  // a zero-sized quarantine, or if the size of the chunk is greater than the
+  // quarantine chunk size threshold.
+  void quarantineOrDeallocateChunk(void *Ptr, UnpackedHeader *Header,
                                    uptr Size) {
-    bool FromPrimary = Header->FromPrimary;
-    bool BypassQuarantine = (AllocatorQuarantine.GetCacheSize() == 0);
+    const bool BypassQuarantine = (AllocatorQuarantine.GetCacheSize() == 0) ||
+        (Size > QuarantineChunksUpToSize);
     if (BypassQuarantine) {
-      Chunk->eraseHeader();
-      void *Ptr = Chunk->getAllocBeg(Header);
-      if (FromPrimary) {
-        ScudoThreadContext *ThreadContext = getThreadContextAndLock();
-        if (LIKELY(ThreadContext)) {
-          getBackendAllocator().deallocatePrimary(
-              getAllocatorCache(ThreadContext), Ptr);
-          ThreadContext->unlock();
-        } else {
-          SpinMutexLock Lock(&FallbackMutex);
-          getBackendAllocator().deallocatePrimary(&FallbackAllocatorCache, Ptr);
-        }
+      Chunk::eraseHeader(Ptr);
+      void *BackendPtr = Chunk::getBackendPtr(Ptr, Header);
+      if (Header->ClassId) {
+        ScudoTSD *TSD = getTSDAndLock();
+        getBackendAllocator().deallocatePrimary(&TSD->Cache, BackendPtr,
+                                                Header->ClassId);
+        TSD->unlock();
       } else {
-        getBackendAllocator().deallocateSecondary(Ptr);
+        getBackendAllocator().deallocateSecondary(BackendPtr);
       }
     } else {
+      // If a small memory amount was allocated with a larger alignment, we want
+      // to take that into account. Otherwise the Quarantine would be filled
+      // with tiny chunks, taking a lot of VA memory. This is an approximation
+      // of the usable size, that allows us to not call
+      // GetActuallyAllocatedSize.
+      uptr EstimatedSize = Size + (Header->Offset << MinAlignmentLog);
       UnpackedHeader NewHeader = *Header;
       NewHeader.State = ChunkQuarantine;
-      Chunk->compareExchangeHeader(&NewHeader, Header);
-      ScudoThreadContext *ThreadContext = getThreadContextAndLock();
-      if (LIKELY(ThreadContext)) {
-        AllocatorQuarantine.Put(getQuarantineCache(ThreadContext),
-                                QuarantineCallback(
-                                    getAllocatorCache(ThreadContext)),
-                                Chunk, Size);
-        ThreadContext->unlock();
-      } else {
-        SpinMutexLock l(&FallbackMutex);
-        AllocatorQuarantine.Put(&FallbackQuarantineCache,
-                                QuarantineCallback(&FallbackAllocatorCache),
-                                Chunk, Size);
-      }
+      Chunk::compareExchangeHeader(Ptr, &NewHeader, Header);
+      ScudoTSD *TSD = getTSDAndLock();
+      AllocatorQuarantine.Put(getQuarantineCache(TSD),
+                              QuarantineCallback(&TSD->Cache), Ptr,
+                              EstimatedSize);
+      TSD->unlock();
     }
   }
 
-  // Deallocates a Chunk, which means adding it to the delayed free list (or
-  // Quarantine).
-  void deallocate(void *UserPtr, uptr DeleteSize, AllocType Type) {
-    initThreadMaybe();
-    // if (&__sanitizer_free_hook) __sanitizer_free_hook(UserPtr);
-    if (UNLIKELY(!UserPtr))
+  // Deallocates a Chunk, which means either adding it to the quarantine or
+  // directly returning it to the backend if criteria are met.
+  void deallocate(void *Ptr, uptr DeleteSize, AllocType Type) {
+    // For a deallocation, we only ensure minimal initialization, meaning thread
+    // local data will be left uninitialized for now (when using ELF TLS). The
+    // fallback cache will be used instead. This is a workaround for a situation
+    // where the only heap operation performed in a thread would be a free past
+    // the TLS destructors, ending up in initialized thread specific data never
+    // being destroyed properly. Any other heap operation will do a full init.
+    initThreadMaybe(/*MinimalInit=*/true);
+    // if (&__sanitizer_free_hook) __sanitizer_free_hook(Ptr);
+    if (UNLIKELY(!Ptr))
       return;
-    uptr UserBeg = reinterpret_cast<uptr>(UserPtr);
-    if (UNLIKELY(!IsAligned(UserBeg, MinAlignment))) {
+    if (UNLIKELY(!Chunk::isAligned(Ptr))) {
       dieWithMessage("ERROR: attempted to deallocate a chunk not properly "
-                     "aligned at address %p\n", UserPtr);
+                     "aligned at address %p\n", Ptr);
     }
-    ScudoChunk *Chunk = getScudoChunk(UserBeg);
-    UnpackedHeader OldHeader;
-    Chunk->loadHeader(&OldHeader);
-    if (UNLIKELY(OldHeader.State != ChunkAllocated)) {
+    UnpackedHeader Header;
+    Chunk::loadHeader(Ptr, &Header);
+    if (UNLIKELY(Header.State != ChunkAllocated)) {
       dieWithMessage("ERROR: invalid chunk state when deallocating address "
-                     "%p\n", UserPtr);
+                     "%p\n", Ptr);
     }
     if (DeallocationTypeMismatch) {
       // The deallocation type has to match the allocation one.
-      if (OldHeader.AllocType != Type) {
+      if (Header.AllocType != Type) {
         // With the exception of memalign'd Chunks, that can be still be free'd.
-        if (OldHeader.AllocType != FromMemalign || Type != FromMalloc) {
-          dieWithMessage("ERROR: allocation type mismatch on address %p\n",
-                         UserPtr);
+        if (Header.AllocType != FromMemalign || Type != FromMalloc) {
+          dieWithMessage("ERROR: allocation type mismatch when deallocating "
+                         "address %p\n", Ptr);
         }
       }
     }
-    uptr Size = OldHeader.FromPrimary ? OldHeader.SizeOrUnusedBytes :
-        Chunk->getUsableSize(&OldHeader) - OldHeader.SizeOrUnusedBytes;
+    uptr Size = Header.ClassId ? Header.SizeOrUnusedBytes :
+        Chunk::getUsableSize(Ptr, &Header) - Header.SizeOrUnusedBytes;
     if (DeleteSizeMismatch) {
       if (DeleteSize && DeleteSize != Size) {
         dieWithMessage("ERROR: invalid sized delete on chunk at address %p\n",
-                       UserPtr);
+                       Ptr);
       }
     }
-
-    // If a small memory amount was allocated with a larger alignment, we want
-    // to take that into account. Otherwise the Quarantine would be filled with
-    // tiny chunks, taking a lot of VA memory. This is an approximation of the
-    // usable size, that allows us to not call GetActuallyAllocatedSize.
-    uptr LiableSize = Size + (OldHeader.Offset << MinAlignment);
-    quarantineOrDeallocateChunk(Chunk, &OldHeader, LiableSize);
+    quarantineOrDeallocateChunk(Ptr, &Header, Size);
   }
 
   // Reallocates a chunk. We can save on a new allocation if the new requested
   // size still fits in the chunk.
   void *reallocate(void *OldPtr, uptr NewSize) {
     initThreadMaybe();
-    uptr UserBeg = reinterpret_cast<uptr>(OldPtr);
-    if (UNLIKELY(!IsAligned(UserBeg, MinAlignment))) {
+    if (UNLIKELY(!Chunk::isAligned(OldPtr))) {
       dieWithMessage("ERROR: attempted to reallocate a chunk not properly "
                      "aligned at address %p\n", OldPtr);
     }
-    ScudoChunk *Chunk = getScudoChunk(UserBeg);
     UnpackedHeader OldHeader;
-    Chunk->loadHeader(&OldHeader);
+    Chunk::loadHeader(OldPtr, &OldHeader);
     if (UNLIKELY(OldHeader.State != ChunkAllocated)) {
       dieWithMessage("ERROR: invalid chunk state when reallocating address "
                      "%p\n", OldPtr);
     }
-    if (UNLIKELY(OldHeader.AllocType != FromMalloc)) {
-      dieWithMessage("ERROR: invalid chunk type when reallocating address %p\n",
-                     OldPtr);
+    if (DeallocationTypeMismatch) {
+      if (UNLIKELY(OldHeader.AllocType != FromMalloc)) {
+        dieWithMessage("ERROR: allocation type mismatch when reallocating "
+                       "address %p\n", OldPtr);
+      }
     }
-    uptr UsableSize = Chunk->getUsableSize(&OldHeader);
+    const uptr UsableSize = Chunk::getUsableSize(OldPtr, &OldHeader);
     // The new size still fits in the current chunk, and the size difference
     // is reasonable.
     if (NewSize <= UsableSize &&
         (UsableSize - NewSize) < (SizeClassMap::kMaxSize / 2)) {
       UnpackedHeader NewHeader = OldHeader;
       NewHeader.SizeOrUnusedBytes =
-                OldHeader.FromPrimary ? NewSize : UsableSize - NewSize;
-      Chunk->compareExchangeHeader(&NewHeader, &OldHeader);
+          OldHeader.ClassId ? NewSize : UsableSize - NewSize;
+      Chunk::compareExchangeHeader(OldPtr, &NewHeader, &OldHeader);
       return OldPtr;
     }
     // Otherwise, we have to allocate a new chunk and copy the contents of the
     // old one.
     void *NewPtr = allocate(NewSize, MinAlignment, FromMalloc);
     if (NewPtr) {
-      uptr OldSize = OldHeader.FromPrimary ? OldHeader.SizeOrUnusedBytes :
+      uptr OldSize = OldHeader.ClassId ? OldHeader.SizeOrUnusedBytes :
           UsableSize - OldHeader.SizeOrUnusedBytes;
-      memcpy(NewPtr, OldPtr, Min(NewSize, OldSize));
-      quarantineOrDeallocateChunk(Chunk, &OldHeader, UsableSize);
+      memcpy(NewPtr, OldPtr, Min(NewSize, UsableSize));
+      quarantineOrDeallocateChunk(OldPtr, &OldHeader, OldSize);
     }
     return NewPtr;
   }
@@ -585,16 +561,14 @@ struct ScudoAllocator {
     initThreadMaybe();
     if (UNLIKELY(!Ptr))
       return 0;
-    uptr UserBeg = reinterpret_cast<uptr>(Ptr);
-    ScudoChunk *Chunk = getScudoChunk(UserBeg);
     UnpackedHeader Header;
-    Chunk->loadHeader(&Header);
+    Chunk::loadHeader(Ptr, &Header);
     // Getting the usable size of a chunk only makes sense if it's allocated.
     if (UNLIKELY(Header.State != ChunkAllocated)) {
       dieWithMessage("ERROR: invalid chunk state when sizing address %p\n",
                      Ptr);
     }
-    return Chunk->getUsableSize(&Header);
+    return Chunk::getUsableSize(Ptr, &Header);
   }
 
   void *calloc(uptr NMemB, uptr Size) {
@@ -604,11 +578,10 @@ struct ScudoAllocator {
     return allocate(NMemB * Size, MinAlignment, FromMalloc, true);
   }
 
-  void commitBack(ScudoThreadContext *ThreadContext) {
-    AllocatorCache *Cache = getAllocatorCache(ThreadContext);
-    AllocatorQuarantine.Drain(getQuarantineCache(ThreadContext),
-                              QuarantineCallback(Cache));
-    BackendAllocator.destroyCache(Cache);
+  void commitBack(ScudoTSD *TSD) {
+    AllocatorQuarantine.Drain(getQuarantineCache(TSD),
+                              QuarantineCallback(&TSD->Cache));
+    BackendAllocator.destroyCache(&TSD->Cache);
   }
 
   uptr getStats(AllocatorStat StatType) {
@@ -617,6 +590,19 @@ struct ScudoAllocator {
     BackendAllocator.getStats(stats);
     return stats[StatType];
   }
+
+  void *handleBadRequest() {
+    initThreadMaybe();
+    return FailureHandler::OnBadRequest();
+  }
+
+  void setRssLimit(uptr LimitMb, bool HardLimit) {
+    if (HardLimit)
+      HardRssLimitMb = LimitMb;
+    else
+      SoftRssLimitMb = LimitMb;
+    CheckRssLimit = HardRssLimitMb || SoftRssLimitMb;
+  }
 };
 
 static ScudoAllocator Instance(LINKER_INITIALIZED);
@@ -625,17 +611,17 @@ static ScudoBackendAllocator &getBackendAllocator() {
   return Instance.BackendAllocator;
 }
 
-static void initScudoInternal(const AllocatorOptions &Options) {
-  Instance.init(Options);
+void initScudo() {
+  Instance.init();
 }
 
-void ScudoThreadContext::init() {
+void ScudoTSD::init(bool Shared) {
+  UnlockRequired = Shared;
   getBackendAllocator().initCache(&Cache);
-  Prng.init();
   memset(QuarantineCachePlaceHolder, 0, sizeof(QuarantineCachePlaceHolder));
 }
 
-void ScudoThreadContext::commitBack() {
+void ScudoTSD::commitBack() {
   Instance.commitBack(this);
 }
 
@@ -672,6 +658,10 @@ void *scudoValloc(uptr Size) {
 
 void *scudoPvalloc(uptr Size) {
   uptr PageSize = GetPageSizeCached();
+  if (UNLIKELY(CheckForPvallocOverflow(Size, PageSize))) {
+    errno = ENOMEM;
+    return Instance.handleBadRequest();
+  }
   // pvalloc(0) should allocate one page.
   Size = Size ? RoundUpTo(Size, PageSize) : PageSize;
   return SetErrnoOnNull(Instance.allocate(Size, PageSize, FromMemalign));
@@ -679,28 +669,28 @@ void *scudoPvalloc(uptr Size) {
 
 void *scudoMemalign(uptr Alignment, uptr Size) {
   if (UNLIKELY(!IsPowerOfTwo(Alignment))) {
-    errno = errno_EINVAL;
-    return ScudoAllocator::FailureHandler::OnBadRequest();
+    errno = EINVAL;
+    return Instance.handleBadRequest();
   }
   return SetErrnoOnNull(Instance.allocate(Size, Alignment, FromMemalign));
 }
 
 int scudoPosixMemalign(void **MemPtr, uptr Alignment, uptr Size) {
   if (UNLIKELY(!CheckPosixMemalignAlignment(Alignment))) {
-    ScudoAllocator::FailureHandler::OnBadRequest();
-    return errno_EINVAL;
+    Instance.handleBadRequest();
+    return EINVAL;
   }
   void *Ptr = Instance.allocate(Size, Alignment, FromMemalign);
   if (UNLIKELY(!Ptr))
-    return errno_ENOMEM;
+    return ENOMEM;
   *MemPtr = Ptr;
   return 0;
 }
 
 void *scudoAlignedAlloc(uptr Alignment, uptr Size) {
   if (UNLIKELY(!CheckAlignedAllocAlignmentAndSize(Alignment, Size))) {
-    errno = errno_EINVAL;
-    return ScudoAllocator::FailureHandler::OnBadRequest();
+    errno = EINVAL;
+    return Instance.handleBadRequest();
   }
   return SetErrnoOnNull(Instance.allocate(Size, Alignment, FromMalloc));
 }
@@ -742,3 +732,13 @@ int __sanitizer_get_ownership(const void *Ptr) {
 uptr __sanitizer_get_allocated_size(const void *Ptr) {
   return Instance.getUsableSize(Ptr);
 }
+
+// Interface functions
+
+extern "C" {
+void __scudo_set_rss_limit(unsigned long LimitMb, int HardLimit) {  // NOLINT
+  if (!SCUDO_CAN_USE_PUBLIC_INTERFACE)
+    return;
+  Instance.setRssLimit(LimitMb, !!HardLimit);
+}
+}  // extern "C"
diff --git a/contrib/compiler-rt/lib/scudo/scudo_allocator.h b/contrib/compiler-rt/lib/scudo/scudo_allocator.h
index 29d85995a3ee..a561247def9c 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_allocator.h
+++ b/contrib/compiler-rt/lib/scudo/scudo_allocator.h
@@ -14,21 +14,15 @@
 #ifndef SCUDO_ALLOCATOR_H_
 #define SCUDO_ALLOCATOR_H_
 
-#include "scudo_flags.h"
-
-#include "sanitizer_common/sanitizer_allocator.h"
-
-#if !SANITIZER_LINUX
-# error "The Scudo hardened allocator is currently only supported on Linux."
-#endif
+#include "scudo_platform.h"
 
 namespace __scudo {
 
 enum AllocType : u8 {
-  FromMalloc    = 0, // Memory block came from malloc, realloc, calloc, etc.
-  FromNew       = 1, // Memory block came from operator new.
-  FromNewArray  = 2, // Memory block came from operator new [].
-  FromMemalign  = 3, // Memory block came from memalign, posix_memalign, etc.
+  FromMalloc    = 0,  // Memory block came from malloc, realloc, calloc, etc.
+  FromNew       = 1,  // Memory block came from operator new.
+  FromNewArray  = 2,  // Memory block came from operator new [].
+  FromMemalign  = 3,  // Memory block came from memalign, posix_memalign, etc.
 };
 
 enum ChunkState : u8 {
@@ -45,16 +39,15 @@ enum ChunkState : u8 {
 typedef u64 PackedHeader;
 struct UnpackedHeader {
   u64 Checksum          : 16;
-  u64 SizeOrUnusedBytes : 19; // Size for Primary backed allocations, amount of
-                              // unused bytes in the chunk for Secondary ones.
-  u64 FromPrimary       : 1;
-  u64 State             : 2;  // available, allocated, or quarantined
-  u64 AllocType         : 2;  // malloc, new, new[], or memalign
-  u64 Offset            : 16; // Offset from the beginning of the backend
-                              // allocation to the beginning of the chunk
-                              // itself, in multiples of MinAlignment. See
-                              // comment about its maximum value and in init().
-  u64 Salt              : 8;
+  u64 ClassId           : 8;
+  u64 SizeOrUnusedBytes : 20;  // Size for Primary backed allocations, amount of
+                               // unused bytes in the chunk for Secondary ones.
+  u64 State             : 2;   // available, allocated, or quarantined
+  u64 AllocType         : 2;   // malloc, new, new[], or memalign
+  u64 Offset            : 16;  // Offset from the beginning of the backend
+                               // allocation to the beginning of the chunk
+                               // itself, in multiples of MinAlignment. See
+                               // comment about its maximum value and in init().
 };
 
 typedef atomic_uint64_t AtomicPackedHeader;
@@ -72,14 +65,6 @@ const uptr AlignedChunkHeaderSize =
 
 #if SANITIZER_CAN_USE_ALLOCATOR64
 const uptr AllocatorSpace = ~0ULL;
-# if defined(__aarch64__) && SANITIZER_ANDROID
-const uptr AllocatorSize = 0x4000000000ULL;  // 256G.
-# elif defined(__aarch64__)
-const uptr AllocatorSize = 0x10000000000ULL;  // 1T.
-# else
-const uptr AllocatorSize = 0x40000000000ULL;  // 4T.
-# endif
-typedef DefaultSizeClassMap SizeClassMap;
 struct AP64 {
   static const uptr kSpaceBeg = AllocatorSpace;
   static const uptr kSpaceSize = AllocatorSize;
@@ -91,17 +76,12 @@ struct AP64 {
 };
 typedef SizeClassAllocator64<AP64> PrimaryAllocator;
 #else
-// Currently, the 32-bit Sanitizer allocator has not yet benefited from all the
-// security improvements brought to the 64-bit one. This makes the 32-bit
-// version of Scudo slightly less toughened.
-static const uptr RegionSizeLog = 20;
 static const uptr NumRegions = SANITIZER_MMAP_RANGE_SIZE >> RegionSizeLog;
 # if SANITIZER_WORDSIZE == 32
 typedef FlatByteMap<NumRegions> ByteMap;
 # elif SANITIZER_WORDSIZE == 64
 typedef TwoLevelByteMap<(NumRegions >> 12), 1 << 12> ByteMap;
 # endif  // SANITIZER_WORDSIZE
-typedef DefaultSizeClassMap SizeClassMap;
 struct AP32 {
   static const uptr kSpaceBeg = 0;
   static const u64 kSpaceSize = SANITIZER_MMAP_RANGE_SIZE;
@@ -111,7 +91,8 @@ struct AP32 {
   typedef __scudo::ByteMap ByteMap;
   typedef NoOpMapUnmapCallback MapUnmapCallback;
   static const uptr kFlags =
-      SizeClassAllocator32FlagMasks::kRandomShuffleChunks;
+      SizeClassAllocator32FlagMasks::kRandomShuffleChunks |
+      SizeClassAllocator32FlagMasks::kUseSeparateSizeClassForBatch;
 };
 typedef SizeClassAllocator32<AP32> PrimaryAllocator;
 #endif  // SANITIZER_CAN_USE_ALLOCATOR64
diff --git a/contrib/compiler-rt/lib/scudo/scudo_allocator_combined.h b/contrib/compiler-rt/lib/scudo/scudo_allocator_combined.h
index 7599c12abb6d..25e273114c23 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_allocator_combined.h
+++ b/contrib/compiler-rt/lib/scudo/scudo_allocator_combined.h
@@ -31,8 +31,8 @@ class ScudoCombinedAllocator {
 
   // Primary allocations are always MinAlignment aligned, and as such do not
   // require an Alignment parameter.
-  void *allocatePrimary(AllocatorCache *Cache, uptr Size) {
-    return Cache->Allocate(&Primary, Primary.ClassID(Size));
+  void *allocatePrimary(AllocatorCache *Cache, uptr ClassId) {
+    return Cache->Allocate(&Primary, ClassId);
   }
 
   // Secondary allocations do not require a Cache, but do require an Alignment
@@ -41,17 +41,17 @@ class ScudoCombinedAllocator {
     return Secondary.Allocate(&Stats, Size, Alignment);
   }
 
-  void deallocatePrimary(AllocatorCache *Cache, void *Ptr) {
-    Cache->Deallocate(&Primary, Primary.GetSizeClass(Ptr), Ptr);
+  void deallocatePrimary(AllocatorCache *Cache, void *Ptr, uptr ClassId) {
+    Cache->Deallocate(&Primary, ClassId, Ptr);
   }
 
   void deallocateSecondary(void *Ptr) {
     Secondary.Deallocate(&Stats, Ptr);
   }
 
-  uptr getActuallyAllocatedSize(void *Ptr, bool FromPrimary) {
-    if (FromPrimary)
-      return PrimaryAllocator::ClassIdToSize(Primary.GetSizeClass(Ptr));
+  uptr getActuallyAllocatedSize(void *Ptr, uptr ClassId) {
+    if (ClassId)
+      return PrimaryAllocator::ClassIdToSize(ClassId);
     return Secondary.GetActuallyAllocatedSize(Ptr);
   }
 
diff --git a/contrib/compiler-rt/lib/scudo/scudo_allocator_secondary.h b/contrib/compiler-rt/lib/scudo/scudo_allocator_secondary.h
index dbfb22565f9c..f2002ed986c3 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_allocator_secondary.h
+++ b/contrib/compiler-rt/lib/scudo/scudo_allocator_secondary.h
@@ -23,23 +23,24 @@
 
 class ScudoLargeMmapAllocator {
  public:
-
   void Init() {
-    PageSize = GetPageSizeCached();
+    PageSizeCached = GetPageSizeCached();
   }
 
   void *Allocate(AllocatorStats *Stats, uptr Size, uptr Alignment) {
-    uptr UserSize = Size - AlignedChunkHeaderSize;
+    const uptr UserSize = Size - AlignedChunkHeaderSize;
     // The Scudo frontend prevents us from allocating more than
     // MaxAllowedMallocSize, so integer overflow checks would be superfluous.
-    uptr MapSize = Size + SecondaryHeaderSize;
+    uptr MapSize = Size + AlignedReservedAddressRangeSize;
     if (Alignment > MinAlignment)
       MapSize += Alignment;
+    const uptr PageSize = PageSizeCached;
     MapSize = RoundUpTo(MapSize, PageSize);
     // Account for 2 guard pages, one before and one after the chunk.
     MapSize += 2 * PageSize;
 
-    uptr MapBeg = reinterpret_cast<uptr>(MmapNoAccess(MapSize));
+    ReservedAddressRange AddressRange;
+    uptr MapBeg = AddressRange.Init(MapSize);
     if (MapBeg == ~static_cast<uptr>(0))
       return ReturnNullOrDieOnFailure::OnOOM();
     // A page-aligned pointer is assumed after that, so check it now.
@@ -62,27 +63,27 @@ class ScudoLargeMmapAllocator {
             PageSize;
         CHECK_GE(NewMapBeg, MapBeg);
         if (NewMapBeg != MapBeg) {
-          UnmapOrDie(reinterpret_cast<void *>(MapBeg), NewMapBeg - MapBeg);
+          AddressRange.Unmap(MapBeg, NewMapBeg - MapBeg);
           MapBeg = NewMapBeg;
         }
         UserEnd = UserBeg + UserSize;
       }
       uptr NewMapEnd = RoundUpTo(UserEnd, PageSize) + PageSize;
       if (NewMapEnd != MapEnd) {
-        UnmapOrDie(reinterpret_cast<void *>(NewMapEnd), MapEnd - NewMapEnd);
+        AddressRange.Unmap(NewMapEnd, MapEnd - NewMapEnd);
         MapEnd = NewMapEnd;
       }
       MapSize = MapEnd - MapBeg;
     }
 
     CHECK_LE(UserEnd, MapEnd - PageSize);
-    // Actually mmap the memory, preserving the guard pages on either side.
-    CHECK_EQ(MapBeg + PageSize, reinterpret_cast<uptr>(
-        MmapFixedOrDie(MapBeg + PageSize, MapSize - 2 * PageSize)));
-    uptr Ptr = UserBeg - AlignedChunkHeaderSize;
-    SecondaryHeader *Header = getHeader(Ptr);
-    Header->MapBeg = MapBeg;
-    Header->MapSize = MapSize;
+    // Actually mmap the memory, preserving the guard pages on either side
+    CHECK_EQ(MapBeg + PageSize,
+             AddressRange.Map(MapBeg + PageSize, MapSize - 2 * PageSize));
+    const uptr Ptr = UserBeg - AlignedChunkHeaderSize;
+    ReservedAddressRange *StoredRange = getReservedAddressRange(Ptr);
+    *StoredRange = AddressRange;
+
     // The primary adds the whole class size to the stats when allocating a
     // chunk, so we will do something similar here. But we will not account for
     // the guard pages.
@@ -96,42 +97,43 @@ class ScudoLargeMmapAllocator {
   }
 
   void Deallocate(AllocatorStats *Stats, void *Ptr) {
-    SecondaryHeader *Header = getHeader(Ptr);
+    // Since we're unmapping the entirety of where the ReservedAddressRange
+    // actually is, copy onto the stack.
+    const uptr PageSize = PageSizeCached;
+    ReservedAddressRange AddressRange = *getReservedAddressRange(Ptr);
     {
       SpinMutexLock l(&StatsMutex);
-      Stats->Sub(AllocatorStatAllocated, Header->MapSize - 2 * PageSize);
-      Stats->Sub(AllocatorStatMapped, Header->MapSize - 2 * PageSize);
+      Stats->Sub(AllocatorStatAllocated, AddressRange.size() - 2 * PageSize);
+      Stats->Sub(AllocatorStatMapped, AddressRange.size() - 2 * PageSize);
     }
-    UnmapOrDie(reinterpret_cast<void *>(Header->MapBeg), Header->MapSize);
+    AddressRange.Unmap(reinterpret_cast<uptr>(AddressRange.base()),
+                       AddressRange.size());
   }
 
   uptr GetActuallyAllocatedSize(void *Ptr) {
-    SecondaryHeader *Header = getHeader(Ptr);
-    // Deduct PageSize as MapSize includes the trailing guard page.
-    uptr MapEnd = Header->MapBeg + Header->MapSize - PageSize;
+    ReservedAddressRange *StoredRange = getReservedAddressRange(Ptr);
+    // Deduct PageSize as ReservedAddressRange size includes the trailing guard
+    // page.
+    uptr MapEnd = reinterpret_cast<uptr>(StoredRange->base()) +
+        StoredRange->size() - PageSizeCached;
     return MapEnd - reinterpret_cast<uptr>(Ptr);
   }
 
  private:
-  // A Secondary allocated chunk header contains the base of the mapping and
-  // its size, which comprises the guard pages.
-  struct SecondaryHeader {
-    uptr MapBeg;
-    uptr MapSize;
-  };
-  // Check that sizeof(SecondaryHeader) is a multiple of MinAlignment.
-  COMPILER_CHECK((sizeof(SecondaryHeader) & (MinAlignment - 1)) == 0);
-
-  SecondaryHeader *getHeader(uptr Ptr) {
-    return reinterpret_cast<SecondaryHeader*>(Ptr - sizeof(SecondaryHeader));
+  ReservedAddressRange *getReservedAddressRange(uptr Ptr) {
+    return reinterpret_cast<ReservedAddressRange*>(
+        Ptr - sizeof(ReservedAddressRange));
   }
-  SecondaryHeader *getHeader(const void *Ptr) {
-    return getHeader(reinterpret_cast<uptr>(Ptr));
+  ReservedAddressRange *getReservedAddressRange(const void *Ptr) {
+    return getReservedAddressRange(reinterpret_cast<uptr>(Ptr));
   }
 
-  const uptr SecondaryHeaderSize = sizeof(SecondaryHeader);
-  const uptr HeadersSize = SecondaryHeaderSize + AlignedChunkHeaderSize;
-  uptr PageSize;
+  static constexpr uptr AlignedReservedAddressRangeSize =
+      (sizeof(ReservedAddressRange) + MinAlignment - 1) & ~(MinAlignment - 1);
+  static constexpr uptr HeadersSize =
+      AlignedReservedAddressRangeSize + AlignedChunkHeaderSize;
+
+  uptr PageSizeCached;
   SpinMutex StatsMutex;
 };
 
diff --git a/contrib/compiler-rt/lib/scudo/scudo_crc32.h b/contrib/compiler-rt/lib/scudo/scudo_crc32.h
index 5ffcc62658cc..e89e430f4085 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_crc32.h
+++ b/contrib/compiler-rt/lib/scudo/scudo_crc32.h
@@ -40,7 +40,7 @@ enum : u8 {
   CRC32Hardware = 1,
 };
 
-const static u32 CRC32Table[] = {
+static const u32 CRC32Table[] = {
   0x00000000, 0x77073096, 0xee0e612c, 0x990951ba, 0x076dc419, 0x706af48f,
   0xe963a535, 0x9e6495a3, 0x0edb8832, 0x79dcb8a4, 0xe0d5e91e, 0x97d2d988,
   0x09b64c2b, 0x7eb17cbd, 0xe7b82d07, 0x90bf1d91, 0x1db71064, 0x6ab020f2,
diff --git a/contrib/compiler-rt/lib/scudo/scudo_flags.cpp b/contrib/compiler-rt/lib/scudo/scudo_flags.cpp
index 90f0cbf4bb86..2aff3ef1e8fa 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_flags.cpp
+++ b/contrib/compiler-rt/lib/scudo/scudo_flags.cpp
@@ -17,12 +17,11 @@
 #include "sanitizer_common/sanitizer_flags.h"
 #include "sanitizer_common/sanitizer_flag_parser.h"
 
-extern "C" SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
-const char* __scudo_default_options();
+SANITIZER_INTERFACE_WEAK_DEF(const char*, __scudo_default_options, void);
 
 namespace __scudo {
 
-Flags ScudoFlags;  // Use via getFlags().
+static Flags ScudoFlags;  // Use via getFlags().
 
 void Flags::setDefaults() {
 #define SCUDO_FLAG(Type, Name, DefaultValue, Description) Name = DefaultValue;
@@ -37,7 +36,7 @@ static void RegisterScudoFlags(FlagParser *parser, Flags *f) {
 #undef SCUDO_FLAG
 }
 
-static const char *callGetScudoDefaultOptions() {
+static const char *getScudoDefaultOptions() {
   return (&__scudo_default_options) ? __scudo_default_options() : "";
 }
 
@@ -57,8 +56,7 @@ void initFlags() {
   RegisterCommonFlags(&ScudoParser);
 
   // Override from user-specified string.
-  const char *ScudoDefaultOptions = callGetScudoDefaultOptions();
-  ScudoParser.ParseString(ScudoDefaultOptions);
+  ScudoParser.ParseString(getScudoDefaultOptions());
 
   // Override from environment.
   ScudoParser.ParseString(GetEnv("SCUDO_OPTIONS"));
@@ -67,27 +65,52 @@ void initFlags() {
 
   // Sanity checks and default settings for the Quarantine parameters.
 
-  if (f->QuarantineSizeMb < 0) {
-    const int DefaultQuarantineSizeMb = FIRST_32_SECOND_64(4, 16);
-    f->QuarantineSizeMb = DefaultQuarantineSizeMb;
+  if (f->QuarantineSizeMb >= 0) {
+    // Backward compatible logic if QuarantineSizeMb is set.
+    if (f->QuarantineSizeKb >= 0) {
+      dieWithMessage("ERROR: please use either QuarantineSizeMb (deprecated) "
+          "or QuarantineSizeKb, but not both\n");
+    }
+    if (f->QuarantineChunksUpToSize >= 0) {
+      dieWithMessage("ERROR: QuarantineChunksUpToSize cannot be used in "
+          " conjunction with the deprecated QuarantineSizeMb option\n");
+    }
+    // If everything is in order, update QuarantineSizeKb accordingly.
+    f->QuarantineSizeKb = f->QuarantineSizeMb * 1024;
+  } else {
+    // Otherwise proceed with the new options.
+    if (f->QuarantineSizeKb < 0) {
+      const int DefaultQuarantineSizeKb = FIRST_32_SECOND_64(64, 256);
+      f->QuarantineSizeKb = DefaultQuarantineSizeKb;
+    }
+    if (f->QuarantineChunksUpToSize < 0) {
+      const int DefaultQuarantineChunksUpToSize = FIRST_32_SECOND_64(512, 2048);
+      f->QuarantineChunksUpToSize = DefaultQuarantineChunksUpToSize;
+    }
   }
-  // We enforce an upper limit for the quarantine size of 4Gb.
-  if (f->QuarantineSizeMb > (4 * 1024)) {
+
+  // We enforce an upper limit for the chunk quarantine threshold of 4Mb.
+  if (f->QuarantineChunksUpToSize > (4 * 1024 * 1024)) {
+    dieWithMessage("ERROR: the chunk quarantine threshold is too large\n");
+  }
+
+  // We enforce an upper limit for the quarantine size of 32Mb.
+  if (f->QuarantineSizeKb > (32 * 1024)) {
     dieWithMessage("ERROR: the quarantine size is too large\n");
   }
+
   if (f->ThreadLocalQuarantineSizeKb < 0) {
-    const int DefaultThreadLocalQuarantineSizeKb =
-        FIRST_32_SECOND_64(64, 256);
+    const int DefaultThreadLocalQuarantineSizeKb = FIRST_32_SECOND_64(16, 64);
     f->ThreadLocalQuarantineSizeKb = DefaultThreadLocalQuarantineSizeKb;
   }
-  // And an upper limit of 128Mb for the thread quarantine cache.
-  if (f->ThreadLocalQuarantineSizeKb > (128 * 1024)) {
+  // And an upper limit of 8Mb for the thread quarantine cache.
+  if (f->ThreadLocalQuarantineSizeKb > (8 * 1024)) {
     dieWithMessage("ERROR: the per thread quarantine cache size is too "
-                   "large\n");
+        "large\n");
   }
-  if (f->ThreadLocalQuarantineSizeKb == 0 && f->QuarantineSizeMb > 0) {
+  if (f->ThreadLocalQuarantineSizeKb == 0 && f->QuarantineSizeKb > 0) {
     dieWithMessage("ERROR: ThreadLocalQuarantineSizeKb can be set to 0 only "
-                   "when QuarantineSizeMb is set to 0\n");
+        "when QuarantineSizeKb is set to 0\n");
   }
 }
 
diff --git a/contrib/compiler-rt/lib/scudo/scudo_flags.inc b/contrib/compiler-rt/lib/scudo/scudo_flags.inc
index 45f9ea846e1a..f180478fdac3 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_flags.inc
+++ b/contrib/compiler-rt/lib/scudo/scudo_flags.inc
@@ -15,17 +15,27 @@
 # error "Define SCUDO_FLAG prior to including this file!"
 #endif
 
-// Default value is set in scudo_flags.cpp based on architecture.
 SCUDO_FLAG(int, QuarantineSizeMb, -1,
-           "Size (in Mb) of quarantine used to delay the actual deallocation "
-           "of chunks. Lower value may reduce memory usage but decrease the "
-           "effectiveness of the mitigation.")
+           "Deprecated. Please use QuarantineSizeKb.")
+
+// Default value is set in scudo_flags.cpp based on architecture.
+SCUDO_FLAG(int, QuarantineSizeKb, -1,
+           "Size in KB of quarantine used to delay the actual deallocation of "
+           "chunks. Lower value may reduce memory usage but decrease the "
+           "effectiveness of the mitigation. Defaults to 64KB (32-bit) or "
+           "256KB (64-bit)")
 
 // Default value is set in scudo_flags.cpp based on architecture.
 SCUDO_FLAG(int, ThreadLocalQuarantineSizeKb, -1,
-          "Size (in Kb) of per-thread cache used to offload the global "
+          "Size in KB of per-thread cache used to offload the global "
           "quarantine. Lower value may reduce memory usage but might increase "
-          "the contention on the global quarantine.")
+          "the contention on the global quarantine. Defaults to 16KB (32-bit) "
+          "or 64KB (64-bit)")
+
+// Default value is set in scudo_flags.cpp based on architecture.
+SCUDO_FLAG(int, QuarantineChunksUpToSize, -1,
+          "Size in bytes up to which chunks will be quarantined (if lower than"
+          "or equal to). Defaults to 256 (32-bit) or 2048 (64-bit)")
 
 SCUDO_FLAG(bool, DeallocationTypeMismatch, true,
           "Report errors on malloc/delete, new/free, new/delete[], etc.")
diff --git a/contrib/compiler-rt/lib/scudo/scudo_interface_internal.h b/contrib/compiler-rt/lib/scudo/scudo_interface_internal.h
new file mode 100644
index 000000000000..3f39e0c4ee0b
--- /dev/null
+++ b/contrib/compiler-rt/lib/scudo/scudo_interface_internal.h
@@ -0,0 +1,22 @@
+//===-- scudo_interface_internal.h ------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+///
+/// Private Scudo interface header.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef SCUDO_INTERFACE_INTERNAL_H_
+#define SCUDO_INTERFACE_INTERNAL_H_
+
+extern "C" {
+SANITIZER_INTERFACE_ATTRIBUTE
+void __scudo_set_rss_limit(unsigned long LimitMb, int HardLimit);  // NOLINT
+}  // extern "C"
+
+#endif  // SCUDO_INTERFACE_INTERNAL_H_
diff --git a/contrib/compiler-rt/lib/scudo/scudo_new_delete.cpp b/contrib/compiler-rt/lib/scudo/scudo_new_delete.cpp
index cdefb127b965..c5a1abbed82b 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_new_delete.cpp
+++ b/contrib/compiler-rt/lib/scudo/scudo_new_delete.cpp
@@ -15,7 +15,7 @@
 
 #include "interception/interception.h"
 
-#include <cstddef>
+#include <stddef.h>
 
 using namespace __scudo;
 
diff --git a/contrib/compiler-rt/lib/scudo/scudo_platform.h b/contrib/compiler-rt/lib/scudo/scudo_platform.h
new file mode 100644
index 000000000000..e1c9c32e9a62
--- /dev/null
+++ b/contrib/compiler-rt/lib/scudo/scudo_platform.h
@@ -0,0 +1,80 @@
+//===-- scudo_platform.h ----------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+///
+/// Scudo platform specific definitions.
+/// TODO(kostyak): add tests for the compile time defines.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef SCUDO_PLATFORM_H_
+#define SCUDO_PLATFORM_H_
+
+#include "sanitizer_common/sanitizer_allocator.h"
+
+#if !SANITIZER_LINUX && !SANITIZER_FUCHSIA
+# error "The Scudo hardened allocator is not supported on this platform."
+#endif
+
+#define SCUDO_TSD_EXCLUSIVE_SUPPORTED (!SANITIZER_ANDROID && !SANITIZER_FUCHSIA)
+
+#ifndef SCUDO_TSD_EXCLUSIVE
+// SCUDO_TSD_EXCLUSIVE wasn't defined, use a default TSD model for the platform.
+# if SANITIZER_ANDROID || SANITIZER_FUCHSIA
+// Android and Fuchsia use a pool of TSDs shared between threads.
+#  define SCUDO_TSD_EXCLUSIVE 0
+# elif SANITIZER_LINUX && !SANITIZER_ANDROID
+// Non-Android Linux use an exclusive TSD per thread.
+#  define SCUDO_TSD_EXCLUSIVE 1
+# else
+#  error "No default TSD model defined for this platform."
+# endif  // SANITIZER_ANDROID || SANITIZER_FUCHSIA
+#endif  // SCUDO_TSD_EXCLUSIVE
+
+// If the exclusive TSD model is chosen, make sure the platform supports it.
+#if SCUDO_TSD_EXCLUSIVE && !SCUDO_TSD_EXCLUSIVE_SUPPORTED
+# error "The exclusive TSD model is not supported on this platform."
+#endif
+
+// Maximum number of TSDs that can be created for the Shared model.
+#ifndef SCUDO_SHARED_TSD_POOL_SIZE
+# define SCUDO_SHARED_TSD_POOL_SIZE 32U
+#endif  // SCUDO_SHARED_TSD_POOL_SIZE
+
+// The following allows the public interface functions to be disabled.
+#ifndef SCUDO_CAN_USE_PUBLIC_INTERFACE
+# define SCUDO_CAN_USE_PUBLIC_INTERFACE 1
+#endif
+
+namespace __scudo {
+
+#if SANITIZER_CAN_USE_ALLOCATOR64
+# if defined(__aarch64__) && SANITIZER_ANDROID
+const uptr AllocatorSize = 0x4000000000ULL;  // 256G.
+# elif defined(__aarch64__)
+const uptr AllocatorSize = 0x10000000000ULL;  // 1T.
+# else
+const uptr AllocatorSize = 0x40000000000ULL;  // 4T.
+# endif
+#else
+const uptr RegionSizeLog = SANITIZER_ANDROID ? 19 : 20;
+#endif  // SANITIZER_CAN_USE_ALLOCATOR64
+
+#if !defined(SCUDO_SIZE_CLASS_MAP)
+# define SCUDO_SIZE_CLASS_MAP Default
+#endif
+
+#define SIZE_CLASS_MAP_TYPE SIZE_CLASS_MAP_TYPE_(SCUDO_SIZE_CLASS_MAP)
+#define SIZE_CLASS_MAP_TYPE_(T) SIZE_CLASS_MAP_TYPE__(T)
+#define SIZE_CLASS_MAP_TYPE__(T) T##SizeClassMap
+
+typedef SIZE_CLASS_MAP_TYPE SizeClassMap;
+
+}  // namespace __scudo
+
+#endif // SCUDO_PLATFORM_H_
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tls.h b/contrib/compiler-rt/lib/scudo/scudo_tls.h
deleted file mode 100644
index 20c49204cf13..000000000000
--- a/contrib/compiler-rt/lib/scudo/scudo_tls.h
+++ /dev/null
@@ -1,47 +0,0 @@
-//===-- scudo_tls.h ---------------------------------------------*- C++ -*-===//
-//
-//                     The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-///
-/// Scudo thread local structure definition.
-/// Implementation will differ based on the thread local storage primitives
-/// offered by the underlying platform.
-///
-//===----------------------------------------------------------------------===//
-
-#ifndef SCUDO_TLS_H_
-#define SCUDO_TLS_H_
-
-#include "scudo_allocator.h"
-#include "scudo_utils.h"
-
-#include "sanitizer_common/sanitizer_linux.h"
-#include "sanitizer_common/sanitizer_platform.h"
-
-namespace __scudo {
-
-// Platform specific base thread context definitions.
-#include "scudo_tls_context_android.inc"
-#include "scudo_tls_context_linux.inc"
-
-struct ALIGNED(64) ScudoThreadContext : public ScudoThreadContextPlatform {
-  AllocatorCache Cache;
-  ScudoPrng Prng;
-  uptr QuarantineCachePlaceHolder[4];
-  void init();
-  void commitBack();
-};
-
-void initThread();
-
-// Platform specific dastpath functions definitions.
-#include "scudo_tls_android.inc"
-#include "scudo_tls_linux.inc"
-
-}  // namespace __scudo
-
-#endif  // SCUDO_TLS_H_
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tls_android.cpp b/contrib/compiler-rt/lib/scudo/scudo_tls_android.cpp
deleted file mode 100644
index ec74e37c8dbc..000000000000
--- a/contrib/compiler-rt/lib/scudo/scudo_tls_android.cpp
+++ /dev/null
@@ -1,95 +0,0 @@
-//===-- scudo_tls_android.cpp -----------------------------------*- C++ -*-===//
-//
-//                     The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-///
-/// Scudo thread local structure implementation for Android.
-///
-//===----------------------------------------------------------------------===//
-
-#include "sanitizer_common/sanitizer_platform.h"
-
-#if SANITIZER_LINUX && SANITIZER_ANDROID
-
-#include "scudo_tls.h"
-
-#include <pthread.h>
-
-namespace __scudo {
-
-static pthread_once_t GlobalInitialized = PTHREAD_ONCE_INIT;
-static pthread_key_t PThreadKey;
-
-static atomic_uint32_t ThreadContextCurrentIndex;
-static ScudoThreadContext *ThreadContexts;
-static uptr NumberOfContexts;
-
-// sysconf(_SC_NPROCESSORS_{CONF,ONLN}) cannot be used as they allocate memory.
-static uptr getNumberOfCPUs() {
-  cpu_set_t CPUs;
-  CHECK_EQ(sched_getaffinity(0, sizeof(cpu_set_t), &CPUs), 0);
-  return CPU_COUNT(&CPUs);
-}
-
-static void initOnce() {
-  // Hack: TLS_SLOT_TSAN was introduced in N. To be able to use it on M for
-  // testing, we create an unused key. Since the key_data array follows the tls
-  // array, it basically gives us the extra entry we need.
-  // TODO(kostyak): remove and restrict to N and above.
-  CHECK_EQ(pthread_key_create(&PThreadKey, NULL), 0);
-  initScudo();
-  NumberOfContexts = getNumberOfCPUs();
-  ThreadContexts = reinterpret_cast<ScudoThreadContext *>(
-      MmapOrDie(sizeof(ScudoThreadContext) * NumberOfContexts, __func__));
-  for (uptr i = 0; i < NumberOfContexts; i++)
-    ThreadContexts[i].init();
-}
-
-void initThread() {
-  pthread_once(&GlobalInitialized, initOnce);
-  // Initial context assignment is done in a plain round-robin fashion.
-  u32 Index = atomic_fetch_add(&ThreadContextCurrentIndex, 1,
-                               memory_order_relaxed);
-  ScudoThreadContext *ThreadContext =
-      &ThreadContexts[Index % NumberOfContexts];
-  *get_android_tls_ptr() = reinterpret_cast<uptr>(ThreadContext);
-}
-
-ScudoThreadContext *getThreadContextAndLockSlow() {
-  ScudoThreadContext *ThreadContext;
-  // Go through all the contexts and find the first unlocked one. 
-  for (u32 i = 0; i < NumberOfContexts; i++) {
-    ThreadContext = &ThreadContexts[i];
-    if (ThreadContext->tryLock()) {
-      *get_android_tls_ptr() = reinterpret_cast<uptr>(ThreadContext);
-      return ThreadContext;
-    }
-  }
-  // No luck, find the one with the lowest precedence, and slow lock it.
-  u64 Precedence = UINT64_MAX;
-  for (u32 i = 0; i < NumberOfContexts; i++) {
-    u64 SlowLockPrecedence = ThreadContexts[i].getSlowLockPrecedence();
-    if (SlowLockPrecedence && SlowLockPrecedence < Precedence) {
-      ThreadContext = &ThreadContexts[i];
-      Precedence = SlowLockPrecedence;
-    }
-  }
-  if (LIKELY(Precedence != UINT64_MAX)) {
-    ThreadContext->lock();
-    *get_android_tls_ptr() = reinterpret_cast<uptr>(ThreadContext);
-    return ThreadContext;
-  }
-  // Last resort (can this happen?), stick with the current one.
-  ThreadContext =
-      reinterpret_cast<ScudoThreadContext *>(*get_android_tls_ptr());
-  ThreadContext->lock();
-  return ThreadContext;
-}
-
-}  // namespace __scudo
-
-#endif  // SANITIZER_LINUX && SANITIZER_ANDROID
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tls_android.inc b/contrib/compiler-rt/lib/scudo/scudo_tls_android.inc
deleted file mode 100644
index 8ecad7a30a6c..000000000000
--- a/contrib/compiler-rt/lib/scudo/scudo_tls_android.inc
+++ /dev/null
@@ -1,44 +0,0 @@
-//===-- scudo_tls_android.inc -----------------------------------*- C++ -*-===//
-//
-//                     The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-///
-/// Scudo thread local structure fastpath functions implementation for Android.
-///
-//===----------------------------------------------------------------------===//
-
-#ifndef SCUDO_TLS_ANDROID_H_
-#define SCUDO_TLS_ANDROID_H_
-
-#ifndef SCUDO_TLS_H_
-# error "This file must be included inside scudo_tls.h."
-#endif  // SCUDO_TLS_H_
-
-#if SANITIZER_LINUX && SANITIZER_ANDROID
-
-ALWAYS_INLINE void initThreadMaybe() {
-  if (LIKELY(*get_android_tls_ptr()))
-    return;
-  initThread();
-}
-
-ScudoThreadContext *getThreadContextAndLockSlow();
-
-ALWAYS_INLINE ScudoThreadContext *getThreadContextAndLock() {
-  ScudoThreadContext *ThreadContext =
-      reinterpret_cast<ScudoThreadContext *>(*get_android_tls_ptr());
-  CHECK(ThreadContext);
-  // Try to lock the currently associated context.
-  if (ThreadContext->tryLock())
-    return ThreadContext;
-  // If it failed, go the slow path.
-  return getThreadContextAndLockSlow();
-}
-
-#endif  // SANITIZER_LINUX && SANITIZER_ANDROID
-
-#endif  // SCUDO_TLS_ANDROID_H_
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tls_context_android.inc b/contrib/compiler-rt/lib/scudo/scudo_tls_context_android.inc
deleted file mode 100644
index f1951319d487..000000000000
--- a/contrib/compiler-rt/lib/scudo/scudo_tls_context_android.inc
+++ /dev/null
@@ -1,54 +0,0 @@
-//===-- scudo_tls_context_android.inc ---------------------------*- C++ -*-===//
-//
-//                     The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-///
-/// Android specific base thread context definition.
-///
-//===----------------------------------------------------------------------===//
-
-#ifndef SCUDO_TLS_CONTEXT_ANDROID_INC_
-#define SCUDO_TLS_CONTEXT_ANDROID_INC_
-
-#ifndef SCUDO_TLS_H_
-# error "This file must be included inside scudo_tls.h."
-#endif  // SCUDO_TLS_H_
-
-#if SANITIZER_LINUX && SANITIZER_ANDROID
-
-struct ScudoThreadContextPlatform {
-  INLINE bool tryLock() {
-    if (Mutex.TryLock()) {
-      atomic_store_relaxed(&SlowLockPrecedence, 0);
-      return true;
-    }
-    if (atomic_load_relaxed(&SlowLockPrecedence) == 0)
-      atomic_store_relaxed(&SlowLockPrecedence, NanoTime());
-    return false;
-  }
-
-  INLINE void lock() {
-    Mutex.Lock();
-    atomic_store_relaxed(&SlowLockPrecedence, 0);
-  }
-
-  INLINE void unlock() {
-    Mutex.Unlock();
-  }
-
-  INLINE u64 getSlowLockPrecedence() {
-    return atomic_load_relaxed(&SlowLockPrecedence);
-  }
-
- private:
-  StaticSpinMutex Mutex;
-  atomic_uint64_t SlowLockPrecedence;
-};
-
-#endif  // SANITIZER_LINUX && SANITIZER_ANDROID
-
-#endif  // SCUDO_TLS_CONTEXT_ANDROID_INC_
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tls_context_linux.inc b/contrib/compiler-rt/lib/scudo/scudo_tls_context_linux.inc
deleted file mode 100644
index 8d292bdbc932..000000000000
--- a/contrib/compiler-rt/lib/scudo/scudo_tls_context_linux.inc
+++ /dev/null
@@ -1,29 +0,0 @@
-//===-- scudo_tls_context_linux.inc -----------------------------*- C++ -*-===//
-//
-//                     The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-///
-/// Linux specific base thread context definition.
-///
-//===----------------------------------------------------------------------===//
-
-#ifndef SCUDO_TLS_CONTEXT_LINUX_INC_
-#define SCUDO_TLS_CONTEXT_LINUX_INC_
-
-#ifndef SCUDO_TLS_H_
-# error "This file must be included inside scudo_tls.h."
-#endif  // SCUDO_TLS_H_
-
-#if SANITIZER_LINUX && !SANITIZER_ANDROID
-
-struct ScudoThreadContextPlatform {
-  ALWAYS_INLINE void unlock() {}
-};
-
-#endif  // SANITIZER_LINUX && !SANITIZER_ANDROID
-
-#endif  // SCUDO_TLS_CONTEXT_LINUX_INC_
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tls_linux.inc b/contrib/compiler-rt/lib/scudo/scudo_tls_linux.inc
deleted file mode 100644
index 242ee3329ea8..000000000000
--- a/contrib/compiler-rt/lib/scudo/scudo_tls_linux.inc
+++ /dev/null
@@ -1,48 +0,0 @@
-//===-- scudo_tls_linux.inc -------------------------------------*- C++ -*-===//
-//
-//                     The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-///
-/// Scudo thread local structure fastpath functions implementation for platforms
-/// supporting thread_local.
-///
-//===----------------------------------------------------------------------===//
-
-#ifndef SCUDO_TLS_LINUX_H_
-#define SCUDO_TLS_LINUX_H_
-
-#ifndef SCUDO_TLS_H_
-# error "This file must be included inside scudo_tls.h."
-#endif  // SCUDO_TLS_H_
-
-#if SANITIZER_LINUX && !SANITIZER_ANDROID
-
-enum ThreadState : u8 {
-  ThreadNotInitialized = 0,
-  ThreadInitialized,
-  ThreadTornDown,
-};
-__attribute__((tls_model("initial-exec")))
-extern THREADLOCAL ThreadState ScudoThreadState;
-__attribute__((tls_model("initial-exec")))
-extern THREADLOCAL ScudoThreadContext ThreadLocalContext;
-
-ALWAYS_INLINE void initThreadMaybe() {
-  if (LIKELY(ScudoThreadState != ThreadNotInitialized))
-    return;
-  initThread();
-}
-
-ALWAYS_INLINE ScudoThreadContext *getThreadContextAndLock() {
-  if (UNLIKELY(ScudoThreadState == ThreadTornDown))
-    return nullptr;
-  return &ThreadLocalContext;
-}
-
-#endif  // SANITIZER_LINUX && !SANITIZER_ANDROID
-
-#endif  // SCUDO_TLS_LINUX_H_
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tsd.h b/contrib/compiler-rt/lib/scudo/scudo_tsd.h
new file mode 100644
index 000000000000..80464b5ea1e4
--- /dev/null
+++ b/contrib/compiler-rt/lib/scudo/scudo_tsd.h
@@ -0,0 +1,72 @@
+//===-- scudo_tsd.h ---------------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+///
+/// Scudo thread specific data definition.
+/// Implementation will differ based on the thread local storage primitives
+/// offered by the underlying platform.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef SCUDO_TSD_H_
+#define SCUDO_TSD_H_
+
+#include "scudo_allocator.h"
+#include "scudo_utils.h"
+
+#include <pthread.h>
+
+namespace __scudo {
+
+struct ALIGNED(64) ScudoTSD {
+  AllocatorCache Cache;
+  uptr QuarantineCachePlaceHolder[4];
+
+  void init(bool Shared);
+  void commitBack();
+
+  INLINE bool tryLock() {
+    if (Mutex.TryLock()) {
+      atomic_store_relaxed(&Precedence, 0);
+      return true;
+    }
+    if (atomic_load_relaxed(&Precedence) == 0)
+      atomic_store_relaxed(&Precedence, MonotonicNanoTime());
+    return false;
+  }
+
+  INLINE void lock() {
+    Mutex.Lock();
+    atomic_store_relaxed(&Precedence, 0);
+  }
+
+  INLINE void unlock() {
+    if (!UnlockRequired)
+      return;
+    Mutex.Unlock();
+  }
+
+  INLINE u64 getPrecedence() {
+    return atomic_load_relaxed(&Precedence);
+  }
+
+ private:
+  bool UnlockRequired;
+  StaticSpinMutex Mutex;
+  atomic_uint64_t Precedence;
+};
+
+void initThread(bool MinimalInit);
+
+// TSD model specific fastpath functions definitions.
+#include "scudo_tsd_exclusive.inc"
+#include "scudo_tsd_shared.inc"
+
+}  // namespace __scudo
+
+#endif  // SCUDO_TSD_H_
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tls_linux.cpp b/contrib/compiler-rt/lib/scudo/scudo_tsd_exclusive.cpp
index 1e38233f339c..1084dfac91e1 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_tls_linux.cpp
+++ b/contrib/compiler-rt/lib/scudo/scudo_tsd_exclusive.cpp
@@ -1,4 +1,4 @@
-//===-- scudo_tls_linux.cpp -------------------------------------*- C++ -*-===//
+//===-- scudo_tsd_exclusive.cpp ---------------------------------*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -7,18 +7,13 @@
 //
 //===----------------------------------------------------------------------===//
 ///
-/// Scudo thread local structure implementation for platforms supporting
-/// thread_local.
+/// Scudo exclusive TSD implementation.
 ///
 //===----------------------------------------------------------------------===//
 
-#include "sanitizer_common/sanitizer_platform.h"
+#include "scudo_tsd.h"
 
-#if SANITIZER_LINUX && !SANITIZER_ANDROID
-
-#include "scudo_tls.h"
-
-#include <pthread.h>
+#if SCUDO_TSD_EXCLUSIVE
 
 namespace __scudo {
 
@@ -28,7 +23,11 @@ static pthread_key_t PThreadKey;
 __attribute__((tls_model("initial-exec")))
 THREADLOCAL ThreadState ScudoThreadState = ThreadNotInitialized;
 __attribute__((tls_model("initial-exec")))
-THREADLOCAL ScudoThreadContext ThreadLocalContext;
+THREADLOCAL ScudoTSD TSD;
+
+// Fallback TSD for when the thread isn't initialized yet or is torn down. It
+// can be shared between multiple threads and as such must be locked.
+ScudoTSD FallbackTSD;
 
 static void teardownThread(void *Ptr) {
   uptr I = reinterpret_cast<uptr>(Ptr);
@@ -43,7 +42,7 @@ static void teardownThread(void *Ptr) {
                                    reinterpret_cast<void *>(I - 1)) == 0))
       return;
   }
-  ThreadLocalContext.commitBack();
+  TSD.commitBack();
   ScudoThreadState = ThreadTornDown;
 }
 
@@ -51,16 +50,19 @@ static void teardownThread(void *Ptr) {
 static void initOnce() {
   CHECK_EQ(pthread_key_create(&PThreadKey, teardownThread), 0);
   initScudo();
+  FallbackTSD.init(/*Shared=*/true);
 }
 
-void initThread() {
+void initThread(bool MinimalInit) {
   CHECK_EQ(pthread_once(&GlobalInitialized, initOnce), 0);
+  if (UNLIKELY(MinimalInit))
+    return;
   CHECK_EQ(pthread_setspecific(PThreadKey, reinterpret_cast<void *>(
       GetPthreadDestructorIterations())), 0);
-  ThreadLocalContext.init();
+  TSD.init(/*Shared=*/false);
   ScudoThreadState = ThreadInitialized;
 }
 
 }  // namespace __scudo
 
-#endif  // SANITIZER_LINUX && !SANITIZER_ANDROID
+#endif  // SCUDO_TSD_EXCLUSIVE
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tsd_exclusive.inc b/contrib/compiler-rt/lib/scudo/scudo_tsd_exclusive.inc
new file mode 100644
index 000000000000..567b6a1edd12
--- /dev/null
+++ b/contrib/compiler-rt/lib/scudo/scudo_tsd_exclusive.inc
@@ -0,0 +1,46 @@
+//===-- scudo_tsd_exclusive.inc ---------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+///
+/// Scudo exclusive TSD fastpath functions implementation.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef SCUDO_TSD_H_
+# error "This file must be included inside scudo_tsd.h."
+#endif  // SCUDO_TSD_H_
+
+#if SCUDO_TSD_EXCLUSIVE
+
+enum ThreadState : u8 {
+  ThreadNotInitialized = 0,
+  ThreadInitialized,
+  ThreadTornDown,
+};
+__attribute__((tls_model("initial-exec")))
+extern THREADLOCAL ThreadState ScudoThreadState;
+__attribute__((tls_model("initial-exec")))
+extern THREADLOCAL ScudoTSD TSD;
+
+extern ScudoTSD FallbackTSD;
+
+ALWAYS_INLINE void initThreadMaybe(bool MinimalInit = false) {
+  if (LIKELY(ScudoThreadState != ThreadNotInitialized))
+    return;
+  initThread(MinimalInit);
+}
+
+ALWAYS_INLINE ScudoTSD *getTSDAndLock() {
+  if (UNLIKELY(ScudoThreadState != ThreadInitialized)) {
+    FallbackTSD.lock();
+    return &FallbackTSD;
+  }
+  return &TSD;
+}
+
+#endif  // SCUDO_TSD_EXCLUSIVE
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tsd_shared.cpp b/contrib/compiler-rt/lib/scudo/scudo_tsd_shared.cpp
new file mode 100644
index 000000000000..3e13e5d3a109
--- /dev/null
+++ b/contrib/compiler-rt/lib/scudo/scudo_tsd_shared.cpp
@@ -0,0 +1,87 @@
+//===-- scudo_tsd_shared.cpp ------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+///
+/// Scudo shared TSD implementation.
+///
+//===----------------------------------------------------------------------===//
+
+#include "scudo_tsd.h"
+
+#if !SCUDO_TSD_EXCLUSIVE
+
+namespace __scudo {
+
+static pthread_once_t GlobalInitialized = PTHREAD_ONCE_INIT;
+pthread_key_t PThreadKey;
+
+static atomic_uint32_t CurrentIndex;
+static ScudoTSD *TSDs;
+static u32 NumberOfTSDs;
+
+static void initOnce() {
+  CHECK_EQ(pthread_key_create(&PThreadKey, NULL), 0);
+  initScudo();
+  NumberOfTSDs = Min(Max(1U, GetNumberOfCPUsCached()),
+                     static_cast<u32>(SCUDO_SHARED_TSD_POOL_SIZE));
+  TSDs = reinterpret_cast<ScudoTSD *>(
+      MmapOrDie(sizeof(ScudoTSD) * NumberOfTSDs, "ScudoTSDs"));
+  for (u32 i = 0; i < NumberOfTSDs; i++)
+    TSDs[i].init(/*Shared=*/true);
+}
+
+ALWAYS_INLINE void setCurrentTSD(ScudoTSD *TSD) {
+#if SANITIZER_ANDROID
+  *get_android_tls_ptr() = reinterpret_cast<uptr>(TSD);
+#else
+  CHECK_EQ(pthread_setspecific(PThreadKey, reinterpret_cast<void *>(TSD)), 0);
+#endif  // SANITIZER_ANDROID
+}
+
+void initThread(bool MinimalInit) {
+  pthread_once(&GlobalInitialized, initOnce);
+  // Initial context assignment is done in a plain round-robin fashion.
+  u32 Index = atomic_fetch_add(&CurrentIndex, 1, memory_order_relaxed);
+  setCurrentTSD(&TSDs[Index % NumberOfTSDs]);
+}
+
+ScudoTSD *getTSDAndLockSlow() {
+  ScudoTSD *TSD;
+  if (NumberOfTSDs > 1) {
+    // Go through all the contexts and find the first unlocked one.
+    for (u32 i = 0; i < NumberOfTSDs; i++) {
+      TSD = &TSDs[i];
+      if (TSD->tryLock()) {
+        setCurrentTSD(TSD);
+        return TSD;
+      }
+    }
+    // No luck, find the one with the lowest Precedence, and slow lock it.
+    u64 LowestPrecedence = UINT64_MAX;
+    for (u32 i = 0; i < NumberOfTSDs; i++) {
+      u64 Precedence = TSDs[i].getPrecedence();
+      if (Precedence && Precedence < LowestPrecedence) {
+        TSD = &TSDs[i];
+        LowestPrecedence = Precedence;
+      }
+    }
+    if (LIKELY(LowestPrecedence != UINT64_MAX)) {
+      TSD->lock();
+      setCurrentTSD(TSD);
+      return TSD;
+    }
+  }
+  // Last resort, stick with the current one.
+  TSD = getCurrentTSD();
+  TSD->lock();
+  return TSD;
+}
+
+}  // namespace __scudo
+
+#endif  // !SCUDO_TSD_EXCLUSIVE
diff --git a/contrib/compiler-rt/lib/scudo/scudo_tsd_shared.inc b/contrib/compiler-rt/lib/scudo/scudo_tsd_shared.inc
new file mode 100644
index 000000000000..79fcd651ed2d
--- /dev/null
+++ b/contrib/compiler-rt/lib/scudo/scudo_tsd_shared.inc
@@ -0,0 +1,48 @@
+//===-- scudo_tsd_shared.inc ------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+///
+/// Scudo shared TSD fastpath functions implementation.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef SCUDO_TSD_H_
+# error "This file must be included inside scudo_tsd.h."
+#endif  // SCUDO_TSD_H_
+
+#if !SCUDO_TSD_EXCLUSIVE
+
+extern pthread_key_t PThreadKey;
+
+ALWAYS_INLINE ScudoTSD* getCurrentTSD() {
+#if SANITIZER_ANDROID
+  return reinterpret_cast<ScudoTSD *>(*get_android_tls_ptr());
+#else
+  return reinterpret_cast<ScudoTSD *>(pthread_getspecific(PThreadKey));
+#endif  // SANITIZER_ANDROID
+}
+
+ALWAYS_INLINE void initThreadMaybe(bool MinimalInit = false) {
+  if (LIKELY(getCurrentTSD()))
+    return;
+  initThread(MinimalInit);
+}
+
+ScudoTSD *getTSDAndLockSlow();
+
+ALWAYS_INLINE ScudoTSD *getTSDAndLock() {
+  ScudoTSD *TSD = getCurrentTSD();
+  CHECK(TSD && "No TSD associated with the current thread!");
+  // Try to lock the currently associated context.
+  if (TSD->tryLock())
+    return TSD;
+  // If it failed, go the slow path.
+  return getTSDAndLockSlow();
+}
+
+#endif  // !SCUDO_TSD_EXCLUSIVE
diff --git a/contrib/compiler-rt/lib/scudo/scudo_utils.cpp b/contrib/compiler-rt/lib/scudo/scudo_utils.cpp
index f7903ff34c73..2f936bf9e780 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_utils.cpp
+++ b/contrib/compiler-rt/lib/scudo/scudo_utils.cpp
@@ -13,17 +13,18 @@
 
 #include "scudo_utils.h"
 
-#include <errno.h>
-#include <fcntl.h>
-#include <stdarg.h>
-#include <unistd.h>
 #if defined(__x86_64__) || defined(__i386__)
 # include <cpuid.h>
-#endif
-#if defined(__arm__) || defined(__aarch64__)
-# include <sys/auxv.h>
+#elif defined(__arm__) || defined(__aarch64__)
+# include "sanitizer_common/sanitizer_getauxval.h"
+# if SANITIZER_POSIX
+#  include "sanitizer_common/sanitizer_posix.h"
+#  include <fcntl.h>
+# endif
 #endif
 
+#include <stdarg.h>
+
 // TODO(kostyak): remove __sanitizer *Printf uses in favor for our own less
 //                complicated string formatting code. The following is a
 //                temporary workaround to be able to use __sanitizer::VSNPrintf.
@@ -36,13 +37,12 @@ extern int VSNPrintf(char *buff, int buff_length, const char *format,
 
 namespace __scudo {
 
-FORMAT(1, 2)
-void NORETURN dieWithMessage(const char *Format, ...) {
+FORMAT(1, 2) void NORETURN dieWithMessage(const char *Format, ...) {
   // Our messages are tiny, 256 characters is more than enough.
   char Message[256];
   va_list Args;
   va_start(Args, Format);
-  __sanitizer::VSNPrintf(Message, sizeof(Message), Format, Args);
+  VSNPrintf(Message, sizeof(Message), Format, Args);
   va_end(Args);
   RawWrite(Message);
   Die();
@@ -51,76 +51,68 @@ void NORETURN dieWithMessage(const char *Format, ...) {
 #if defined(__x86_64__) || defined(__i386__)
 // i386 and x86_64 specific code to detect CRC32 hardware support via CPUID.
 // CRC32 requires the SSE 4.2 instruction set.
-typedef struct {
-  u32 Eax;
-  u32 Ebx;
-  u32 Ecx;
-  u32 Edx;
-} CPUIDRegs;
-
-static void getCPUID(CPUIDRegs *Regs, u32 Level)
-{
-  __get_cpuid(Level, &Regs->Eax, &Regs->Ebx, &Regs->Ecx, &Regs->Edx);
+# ifndef bit_SSE4_2
+#  define bit_SSE4_2 bit_SSE42  // clang and gcc have different defines.
+# endif
+bool hasHardwareCRC32() {
+  u32 Eax, Ebx, Ecx, Edx;
+  __get_cpuid(0, &Eax, &Ebx, &Ecx, &Edx);
+  const bool IsIntel = (Ebx == signature_INTEL_ebx) &&
+                       (Edx == signature_INTEL_edx) &&
+                       (Ecx == signature_INTEL_ecx);
+  const bool IsAMD = (Ebx == signature_AMD_ebx) &&
+                     (Edx == signature_AMD_edx) &&
+                     (Ecx == signature_AMD_ecx);
+  if (!IsIntel && !IsAMD)
+    return false;
+  __get_cpuid(1, &Eax, &Ebx, &Ecx, &Edx);
+  return !!(Ecx & bit_SSE4_2);
 }
-
-CPUIDRegs getCPUFeatures() {
-  CPUIDRegs VendorRegs = {};
-  getCPUID(&VendorRegs, 0);
-  bool IsIntel =
-      (VendorRegs.Ebx == signature_INTEL_ebx) &&
-      (VendorRegs.Edx == signature_INTEL_edx) &&
-      (VendorRegs.Ecx == signature_INTEL_ecx);
-  bool IsAMD =
-      (VendorRegs.Ebx == signature_AMD_ebx) &&
-      (VendorRegs.Edx == signature_AMD_edx) &&
-      (VendorRegs.Ecx == signature_AMD_ecx);
-  // Default to an empty feature set if not on a supported CPU.
-  CPUIDRegs FeaturesRegs = {};
-  if (IsIntel || IsAMD) {
-    getCPUID(&FeaturesRegs, 1);
-  }
-  return FeaturesRegs;
-}
-
-#ifndef bit_SSE4_2
-# define bit_SSE4_2 bit_SSE42  // clang and gcc have different defines.
-#endif
-
-bool testCPUFeature(CPUFeature Feature)
-{
-  CPUIDRegs FeaturesRegs = getCPUFeatures();
-
-  switch (Feature) {
-    case CRC32CPUFeature:  // CRC32 is provided by SSE 4.2.
-      return !!(FeaturesRegs.Ecx & bit_SSE4_2);
-    default:
+#elif defined(__arm__) || defined(__aarch64__)
+// For ARM and AArch64, hardware CRC32 support is indicated in the AT_HWCAP
+// auxiliary vector.
+# ifndef AT_HWCAP
+#  define AT_HWCAP 16
+# endif
+# ifndef HWCAP_CRC32
+#  define HWCAP_CRC32 (1 << 7)  // HWCAP_CRC32 is missing on older platforms.
+# endif
+# if SANITIZER_POSIX
+bool hasHardwareCRC32ARMPosix() {
+  uptr F = internal_open("/proc/self/auxv", O_RDONLY);
+  if (internal_iserror(F))
+    return false;
+  struct { uptr Tag; uptr Value; } Entry = { 0, 0 };
+  for (;;) {
+    uptr N = internal_read(F, &Entry, sizeof(Entry));
+    if (internal_iserror(N) || N != sizeof(Entry) ||
+        (Entry.Tag == 0 && Entry.Value == 0) || Entry.Tag == AT_HWCAP)
       break;
   }
-  return false;
+  internal_close(F);
+  return (Entry.Tag == AT_HWCAP && (Entry.Value & HWCAP_CRC32) != 0);
 }
-#elif defined(__arm__) || defined(__aarch64__)
-// For ARM and AArch64, hardware CRC32 support is indicated in the
-// AT_HWVAL auxiliary vector.
-
-#ifndef HWCAP_CRC32
-# define HWCAP_CRC32 (1<<7)  // HWCAP_CRC32 is missing on older platforms.
-#endif
+# else
+bool hasHardwareCRC32ARMPosix() { return false; }
+# endif  // SANITIZER_POSIX
 
-bool testCPUFeature(CPUFeature Feature) {
-  uptr HWCap = getauxval(AT_HWCAP);
+// Bionic doesn't initialize its globals early enough. This causes issues when
+// trying to access them from a preinit_array (b/25751302) or from another
+// constructor called before the libc one (b/68046352). __progname is
+// initialized after the other globals, so we can check its value to know if
+// calling getauxval is safe.
+extern "C" SANITIZER_WEAK_ATTRIBUTE char *__progname;
+INLINE bool areBionicGlobalsInitialized() {
+  return !SANITIZER_ANDROID || (&__progname && __progname);
+}
 
-  switch (Feature) {
-    case CRC32CPUFeature:
-      return !!(HWCap & HWCAP_CRC32);
-    default:
-      break;
-  }
-  return false;
+bool hasHardwareCRC32() {
+  if (&getauxval && areBionicGlobalsInitialized())
+    return !!(getauxval(AT_HWCAP) & HWCAP_CRC32);
+  return hasHardwareCRC32ARMPosix();
 }
 #else
-bool testCPUFeature(CPUFeature Feature) {
-  return false;
-}
+bool hasHardwareCRC32() { return false; }
 #endif  // defined(__x86_64__) || defined(__i386__)
 
 }  // namespace __scudo
diff --git a/contrib/compiler-rt/lib/scudo/scudo_utils.h b/contrib/compiler-rt/lib/scudo/scudo_utils.h
index 6c6c9d893404..43448e0831e8 100644
--- a/contrib/compiler-rt/lib/scudo/scudo_utils.h
+++ b/contrib/compiler-rt/lib/scudo/scudo_utils.h
@@ -14,14 +14,14 @@
 #ifndef SCUDO_UTILS_H_
 #define SCUDO_UTILS_H_
 
-#include <string.h>
-
 #include "sanitizer_common/sanitizer_common.h"
 
+#include <string.h>
+
 namespace __scudo {
 
 template <class Dest, class Source>
-inline Dest bit_cast(const Source& source) {
+INLINE Dest bit_cast(const Source& source) {
   static_assert(sizeof(Dest) == sizeof(Source), "Sizes are not equal!");
   Dest dest;
   memcpy(&dest, &source, sizeof(dest));
@@ -30,63 +30,7 @@ inline Dest bit_cast(const Source& source) {
 
 void NORETURN dieWithMessage(const char *Format, ...);
 
-enum CPUFeature {
-  CRC32CPUFeature = 0,
-  MaxCPUFeature,
-};
-bool testCPUFeature(CPUFeature feature);
-
-INLINE u64 rotl(const u64 X, int K) {
-  return (X << K) | (X >> (64 - K));
-}
-
-// XoRoShiRo128+ PRNG (http://xoroshiro.di.unimi.it/).
-struct XoRoShiRo128Plus {
- public:
-  void init() {
-    if (UNLIKELY(!GetRandom(reinterpret_cast<void *>(State), sizeof(State)))) {
-      // Early processes (eg: init) do not have /dev/urandom yet, but we still
-      // have to provide them with some degree of entropy. Not having a secure
-      // seed is not as problematic for them, as they are less likely to be
-      // the target of heap based vulnerabilities exploitation attempts.
-      State[0] = NanoTime();
-      State[1] = 0;
-    }
-    fillCache();
-  }
-  u8 getU8() {
-    if (UNLIKELY(isCacheEmpty()))
-      fillCache();
-    const u8 Result = static_cast<u8>(CachedBytes & 0xff);
-    CachedBytes >>= 8;
-    CachedBytesAvailable--;
-    return Result;
-  }
-  u64 getU64() { return next(); }
-
- private:
-  u8 CachedBytesAvailable;
-  u64 CachedBytes;
-  u64 State[2];
-  u64 next() {
-    const u64 S0 = State[0];
-    u64 S1 = State[1];
-    const u64 Result = S0 + S1;
-    S1 ^= S0;
-    State[0] = rotl(S0, 55) ^ S1 ^ (S1 << 14);
-    State[1] = rotl(S1, 36);
-    return Result;
-  }
-  bool isCacheEmpty() {
-    return CachedBytesAvailable == 0;
-  }
-  void fillCache() {
-    CachedBytes = next();
-    CachedBytesAvailable = sizeof(CachedBytes);
-  }
-};
-
-typedef XoRoShiRo128Plus ScudoPrng;
+bool hasHardwareCRC32();
 
 }  // namespace __scudo
 
diff --git a/contrib/compiler-rt/lib/stats/stats.cc b/contrib/compiler-rt/lib/stats/stats.cc
index df9845a38de9..6a6eb3a3cdbc 100644
--- a/contrib/compiler-rt/lib/stats/stats.cc
+++ b/contrib/compiler-rt/lib/stats/stats.cc
@@ -13,6 +13,7 @@
 //===----------------------------------------------------------------------===//
 
 #include "sanitizer_common/sanitizer_common.h"
+#include "sanitizer_common/sanitizer_file.h"
 #include "sanitizer_common/sanitizer_internal_defs.h"
 #if SANITIZER_POSIX
 #include "sanitizer_common/sanitizer_posix.h"
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_fd.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_fd.cc
index d84df4a6413c..f13a7432ec95 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_fd.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_fd.cc
@@ -48,8 +48,8 @@ static bool bogusfd(int fd) {
 }
 
 static FdSync *allocsync(ThreadState *thr, uptr pc) {
-  FdSync *s = (FdSync*)user_alloc(thr, pc, sizeof(FdSync), kDefaultAlignment,
-      false);
+  FdSync *s = (FdSync*)user_alloc_internal(thr, pc, sizeof(FdSync),
+      kDefaultAlignment, false);
   atomic_store(&s->rc, 1, memory_order_relaxed);
   return s;
 }
@@ -79,7 +79,7 @@ static FdDesc *fddesc(ThreadState *thr, uptr pc, int fd) {
   if (l1 == 0) {
     uptr size = kTableSizeL2 * sizeof(FdDesc);
     // We need this to reside in user memory to properly catch races on it.
-    void *p = user_alloc(thr, pc, size, kDefaultAlignment, false);
+    void *p = user_alloc_internal(thr, pc, size, kDefaultAlignment, false);
     internal_memset(p, 0, size);
     MemoryResetRange(thr, (uptr)&fddesc, (uptr)p, size);
     if (atomic_compare_exchange_strong(pl1, &l1, (uptr)p, memory_order_acq_rel))
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors.cc
index 001123f4941e..d92976ad6e9e 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors.cc
@@ -17,6 +17,7 @@
 #include "sanitizer_common/sanitizer_errno.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_placement_new.h"
 #include "sanitizer_common/sanitizer_posix.h"
@@ -39,6 +40,21 @@ using namespace __tsan;  // NOLINT
 #define stderr __stderrp
 #endif
 
+#if SANITIZER_NETBSD
+#define dirfd(dirp) (*(int *)(dirp))
+#define fileno_unlocked fileno
+
+#if _LP64
+#define __sF_size 152
+#else
+#define __sF_size 88
+#endif
+
+#define stdout ((char*)&__sF + (__sF_size * 1))
+#define stderr ((char*)&__sF + (__sF_size * 2))
+
+#endif
+
 #if SANITIZER_ANDROID
 #define mallopt(a, b)
 #endif
@@ -49,11 +65,6 @@ const int kSigCount = 129;
 const int kSigCount = 65;
 #endif
 
-struct my_siginfo_t {
-  // The size is determined by looking at sizeof of real siginfo_t on linux.
-  u64 opaque[128 / sizeof(u64)];
-};
-
 #ifdef __mips__
 struct ucontext_t {
   u64 opaque[768 / sizeof(u64) + 1];
@@ -84,19 +95,25 @@ DECLARE_REAL_AND_INTERCEPTOR(void, free, void *ptr)
 extern "C" void *pthread_self();
 extern "C" void _exit(int status);
 extern "C" int fileno_unlocked(void *stream);
+#if !SANITIZER_NETBSD
 extern "C" int dirfd(void *dirp);
-#if !SANITIZER_FREEBSD && !SANITIZER_ANDROID
+#endif
+#if !SANITIZER_FREEBSD && !SANITIZER_ANDROID && !SANITIZER_NETBSD
 extern "C" int mallopt(int param, int value);
 #endif
+#if SANITIZER_NETBSD
+extern __sanitizer_FILE __sF[];
+#else
 extern __sanitizer_FILE *stdout, *stderr;
-#if !SANITIZER_FREEBSD && !SANITIZER_MAC
+#endif
+#if !SANITIZER_FREEBSD && !SANITIZER_MAC && !SANITIZER_NETBSD
 const int PTHREAD_MUTEX_RECURSIVE = 1;
 const int PTHREAD_MUTEX_RECURSIVE_NP = 1;
 #else
 const int PTHREAD_MUTEX_RECURSIVE = 2;
 const int PTHREAD_MUTEX_RECURSIVE_NP = 2;
 #endif
-#if !SANITIZER_FREEBSD && !SANITIZER_MAC
+#if !SANITIZER_FREEBSD && !SANITIZER_MAC && !SANITIZER_NETBSD
 const int EPOLL_CTL_ADD = 1;
 #endif
 const int SIGILL = 4;
@@ -105,7 +122,7 @@ const int SIGFPE = 8;
 const int SIGSEGV = 11;
 const int SIGPIPE = 13;
 const int SIGTERM = 15;
-#if defined(__mips__) || SANITIZER_FREEBSD || SANITIZER_MAC
+#if defined(__mips__) || SANITIZER_FREEBSD || SANITIZER_MAC || SANITIZER_NETBSD
 const int SIGBUS = 10;
 const int SIGSYS = 12;
 #else
@@ -113,7 +130,9 @@ const int SIGBUS = 7;
 const int SIGSYS = 31;
 #endif
 void *const MAP_FAILED = (void*)-1;
-#if !SANITIZER_MAC
+#if SANITIZER_NETBSD
+const int PTHREAD_BARRIER_SERIAL_THREAD = 1234567;
+#elif !SANITIZER_MAC
 const int PTHREAD_BARRIER_SERIAL_THREAD = -1;
 #endif
 const int MAP_FIXED = 0x10;
@@ -125,48 +144,7 @@ typedef long long_t;  // NOLINT
 # define F_TLOCK 2      /* Test and lock a region for exclusive use.  */
 # define F_TEST  3      /* Test a region for other processes locks.  */
 
-typedef void (*sighandler_t)(int sig);
-typedef void (*sigactionhandler_t)(int sig, my_siginfo_t *siginfo, void *uctx);
-
-#if SANITIZER_ANDROID
-struct sigaction_t {
-  u32 sa_flags;
-  union {
-    sighandler_t sa_handler;
-    sigactionhandler_t sa_sigaction;
-  };
-  __sanitizer_sigset_t sa_mask;
-  void (*sa_restorer)();
-};
-#else
-struct sigaction_t {
-#ifdef __mips__
-  u32 sa_flags;
-#endif
-  union {
-    sighandler_t sa_handler;
-    sigactionhandler_t sa_sigaction;
-  };
-#if SANITIZER_FREEBSD
-  int sa_flags;
-  __sanitizer_sigset_t sa_mask;
-#elif SANITIZER_MAC
-  __sanitizer_sigset_t sa_mask;
-  int sa_flags;
-#else
-  __sanitizer_sigset_t sa_mask;
-#ifndef __mips__
-  int sa_flags;
-#endif
-  void (*sa_restorer)();
-#endif
-};
-#endif
-
-const sighandler_t SIG_DFL = (sighandler_t)0;
-const sighandler_t SIG_IGN = (sighandler_t)1;
-const sighandler_t SIG_ERR = (sighandler_t)-1;
-#if SANITIZER_FREEBSD || SANITIZER_MAC
+#if SANITIZER_FREEBSD || SANITIZER_MAC || SANITIZER_NETBSD
 const int SA_SIGINFO = 0x40;
 const int SIG_SETMASK = 3;
 #elif defined(__mips__)
@@ -180,13 +158,11 @@ const int SIG_SETMASK = 2;
 #define COMMON_INTERCEPTOR_NOTHING_IS_INITIALIZED \
   (!cur_thread()->is_inited)
 
-static sigaction_t sigactions[kSigCount];
-
 namespace __tsan {
 struct SignalDesc {
   bool armed;
   bool sigaction;
-  my_siginfo_t siginfo;
+  __sanitizer_siginfo siginfo;
   ucontext_t ctx;
 };
 
@@ -200,11 +176,41 @@ struct ThreadSignalContext {
   __sanitizer_sigset_t oldset;
 };
 
-// The object is 64-byte aligned, because we want hot data to be located in
-// a single cache line if possible (it's accessed in every interceptor).
-static ALIGNED(64) char libignore_placeholder[sizeof(LibIgnore)];
+// The sole reason tsan wraps atexit callbacks is to establish synchronization
+// between callback setup and callback execution.
+struct AtExitCtx {
+  void (*f)();
+  void *arg;
+};
+
+// InterceptorContext holds all global data required for interceptors.
+// It's explicitly constructed in InitializeInterceptors with placement new
+// and is never destroyed. This allows usage of members with non-trivial
+// constructors and destructors.
+struct InterceptorContext {
+  // The object is 64-byte aligned, because we want hot data to be located
+  // in a single cache line if possible (it's accessed in every interceptor).
+  ALIGNED(64) LibIgnore libignore;
+  __sanitizer_sigaction sigactions[kSigCount];
+#if !SANITIZER_MAC && !SANITIZER_NETBSD
+  unsigned finalize_key;
+#endif
+
+  BlockingMutex atexit_mu;
+  Vector<struct AtExitCtx *> AtExitStack;
+
+  InterceptorContext()
+      : libignore(LINKER_INITIALIZED), AtExitStack() {
+  }
+};
+
+static ALIGNED(64) char interceptor_placeholder[sizeof(InterceptorContext)];
+InterceptorContext *interceptor_ctx() {
+  return reinterpret_cast<InterceptorContext*>(&interceptor_placeholder[0]);
+}
+
 LibIgnore *libignore() {
-  return reinterpret_cast<LibIgnore*>(&libignore_placeholder[0]);
+  return &interceptor_ctx()->libignore;
 }
 
 void InitializeLibIgnore() {
@@ -232,10 +238,6 @@ static ThreadSignalContext *SigCtx(ThreadState *thr) {
   return ctx;
 }
 
-#if !SANITIZER_MAC
-static unsigned g_thread_finalize_key;
-#endif
-
 ScopedInterceptor::ScopedInterceptor(ThreadState *thr, const char *fname,
                                      uptr pc)
     : thr_(thr), pc_(pc), in_ignored_lib_(false), ignoring_(false) {
@@ -284,8 +286,18 @@ void ScopedInterceptor::DisableIgnores() {
 #define TSAN_INTERCEPT(func) INTERCEPT_FUNCTION(func)
 #if SANITIZER_FREEBSD
 # define TSAN_INTERCEPT_VER(func, ver) INTERCEPT_FUNCTION(func)
+# define TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(func)
+# define TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS_THR(func)
+#elif SANITIZER_NETBSD
+# define TSAN_INTERCEPT_VER(func, ver) INTERCEPT_FUNCTION(func)
+# define TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(func) \
+         INTERCEPT_FUNCTION(__libc_##func)
+# define TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS_THR(func) \
+         INTERCEPT_FUNCTION(__libc_thr_##func)
 #else
 # define TSAN_INTERCEPT_VER(func, ver) INTERCEPT_FUNCTION_VER(func, ver)
+# define TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(func)
+# define TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS_THR(func)
 #endif
 
 #define READ_STRING_OF_LEN(thr, pc, s, len, n)                 \
@@ -346,17 +358,30 @@ TSAN_INTERCEPTOR(int, nanosleep, void *req, void *rem) {
   return res;
 }
 
-// The sole reason tsan wraps atexit callbacks is to establish synchronization
-// between callback setup and callback execution.
-struct AtExitCtx {
-  void (*f)();
-  void *arg;
-};
+TSAN_INTERCEPTOR(int, pause, int fake) {
+  SCOPED_TSAN_INTERCEPTOR(pause, fake);
+  return BLOCK_REAL(pause)(fake);
+}
 
-static void at_exit_wrapper(void *arg) {
-  ThreadState *thr = cur_thread();
-  uptr pc = 0;
-  Acquire(thr, pc, (uptr)arg);
+static void at_exit_wrapper() {
+  AtExitCtx *ctx;
+  {
+    // Ensure thread-safety.
+    BlockingMutexLock l(&interceptor_ctx()->atexit_mu);
+
+    // Pop AtExitCtx from the top of the stack of callback functions
+    uptr element = interceptor_ctx()->AtExitStack.Size() - 1;
+    ctx = interceptor_ctx()->AtExitStack[element];
+    interceptor_ctx()->AtExitStack.PopBack();
+  }
+
+  Acquire(cur_thread(), (uptr)0, (uptr)ctx);
+  ((void(*)())ctx->f)();
+  InternalFree(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);
@@ -367,7 +392,7 @@ static int setup_at_exit_wrapper(ThreadState *thr, uptr pc, void(*f)(),
 
 #if !SANITIZER_ANDROID
 TSAN_INTERCEPTOR(int, atexit, void (*f)()) {
-  if (cur_thread()->in_symbolizer)
+  if (UNLIKELY(cur_thread()->in_symbolizer))
     return 0;
   // We want to setup the atexit callback even if we are in ignored lib
   // or after fork.
@@ -377,7 +402,7 @@ TSAN_INTERCEPTOR(int, atexit, void (*f)()) {
 #endif
 
 TSAN_INTERCEPTOR(int, __cxa_atexit, void (*f)(void *a), void *arg, void *dso) {
-  if (cur_thread()->in_symbolizer)
+  if (UNLIKELY(cur_thread()->in_symbolizer))
     return 0;
   SCOPED_TSAN_INTERCEPTOR(__cxa_atexit, f, arg, dso);
   return setup_at_exit_wrapper(thr, pc, (void(*)())f, arg, dso);
@@ -392,12 +417,27 @@ static int setup_at_exit_wrapper(ThreadState *thr, uptr pc, void(*f)(),
   // Memory allocation in __cxa_atexit will race with free during exit,
   // because we do not see synchronization around atexit callback list.
   ThreadIgnoreBegin(thr, pc);
-  int res = REAL(__cxa_atexit)(at_exit_wrapper, ctx, dso);
+  int res;
+  if (!dso) {
+    // NetBSD does not preserve the 2nd argument if dso is equal to 0
+    // Store ctx in a local stack-like structure
+
+    // Ensure thread-safety.
+    BlockingMutexLock l(&interceptor_ctx()->atexit_mu);
+
+    res = REAL(__cxa_atexit)((void (*)(void *a))at_exit_wrapper, 0, 0);
+    // Push AtExitCtx on the top of the stack of callback functions
+    if (!res) {
+      interceptor_ctx()->AtExitStack.PushBack(ctx);
+    }
+  } else {
+    res = REAL(__cxa_atexit)(cxa_at_exit_wrapper, ctx, dso);
+  }
   ThreadIgnoreEnd(thr, pc);
   return res;
 }
 
-#if !SANITIZER_MAC
+#if !SANITIZER_MAC && !SANITIZER_NETBSD
 static void on_exit_wrapper(int status, void *arg) {
   ThreadState *thr = cur_thread();
   uptr pc = 0;
@@ -408,7 +448,7 @@ static void on_exit_wrapper(int status, void *arg) {
 }
 
 TSAN_INTERCEPTOR(int, on_exit, void(*f)(int, void*), void *arg) {
-  if (cur_thread()->in_symbolizer)
+  if (UNLIKELY(cur_thread()->in_symbolizer))
     return 0;
   SCOPED_TSAN_INTERCEPTOR(on_exit, f, arg);
   AtExitCtx *ctx = (AtExitCtx*)InternalAlloc(sizeof(AtExitCtx));
@@ -422,6 +462,9 @@ TSAN_INTERCEPTOR(int, on_exit, void(*f)(int, void*), void *arg) {
   ThreadIgnoreEnd(thr, pc);
   return res;
 }
+#define TSAN_MAYBE_INTERCEPT_ON_EXIT TSAN_INTERCEPT(on_exit)
+#else
+#define TSAN_MAYBE_INTERCEPT_ON_EXIT
 #endif
 
 // Cleanup old bufs.
@@ -461,13 +504,15 @@ static void LongJmp(ThreadState *thr, uptr *env) {
   uptr mangled_sp = env[0];
 #elif SANITIZER_FREEBSD
   uptr mangled_sp = env[2];
+#elif SANITIZER_NETBSD
+  uptr mangled_sp = env[6];
 #elif SANITIZER_MAC
 # ifdef __aarch64__
     uptr mangled_sp = env[13];
 # else
     uptr mangled_sp = env[2];
 # endif
-#elif defined(SANITIZER_LINUX)
+#elif SANITIZER_LINUX
 # ifdef __aarch64__
   uptr mangled_sp = env[13];
 # elif defined(__mips64)
@@ -510,10 +555,29 @@ TSAN_INTERCEPTOR(int, setjmp, void *env);
 TSAN_INTERCEPTOR(int, _setjmp, void *env);
 TSAN_INTERCEPTOR(int, sigsetjmp, void *env);
 #else  // SANITIZER_MAC
+
+#if SANITIZER_NETBSD
+#define setjmp_symname __setjmp14
+#define sigsetjmp_symname __sigsetjmp14
+#else
+#define setjmp_symname setjmp
+#define sigsetjmp_symname sigsetjmp
+#endif
+
+#define TSAN_INTERCEPTOR_SETJMP_(x) __interceptor_ ## x
+#define TSAN_INTERCEPTOR_SETJMP__(x) TSAN_INTERCEPTOR_SETJMP_(x)
+#define TSAN_INTERCEPTOR_SETJMP TSAN_INTERCEPTOR_SETJMP__(setjmp_symname)
+#define TSAN_INTERCEPTOR_SIGSETJMP TSAN_INTERCEPTOR_SETJMP__(sigsetjmp_symname)
+
+#define TSAN_STRING_SETJMP_(x) # x
+#define TSAN_STRING_SETJMP__(x) TSAN_STRING_SETJMP_(x)
+#define TSAN_STRING_SETJMP TSAN_STRING_SETJMP__(setjmp_symname)
+#define TSAN_STRING_SIGSETJMP TSAN_STRING_SETJMP__(sigsetjmp_symname)
+
 // Not called.  Merely to satisfy TSAN_INTERCEPT().
 extern "C" SANITIZER_INTERFACE_ATTRIBUTE
-int __interceptor_setjmp(void *env);
-extern "C" int __interceptor_setjmp(void *env) {
+int TSAN_INTERCEPTOR_SETJMP(void *env);
+extern "C" int TSAN_INTERCEPTOR_SETJMP(void *env) {
   CHECK(0);
   return 0;
 }
@@ -527,51 +591,75 @@ extern "C" int __interceptor__setjmp(void *env) {
 }
 
 extern "C" SANITIZER_INTERFACE_ATTRIBUTE
-int __interceptor_sigsetjmp(void *env);
-extern "C" int __interceptor_sigsetjmp(void *env) {
+int TSAN_INTERCEPTOR_SIGSETJMP(void *env);
+extern "C" int TSAN_INTERCEPTOR_SIGSETJMP(void *env) {
   CHECK(0);
   return 0;
 }
 
+#if !SANITIEZER_NETBSD
 extern "C" SANITIZER_INTERFACE_ATTRIBUTE
 int __interceptor___sigsetjmp(void *env);
 extern "C" int __interceptor___sigsetjmp(void *env) {
   CHECK(0);
   return 0;
 }
+#endif
 
-extern "C" int setjmp(void *env);
+extern "C" int setjmp_symname(void *env);
 extern "C" int _setjmp(void *env);
-extern "C" int sigsetjmp(void *env);
+extern "C" int sigsetjmp_symname(void *env);
+#if !SANITIEZER_NETBSD
 extern "C" int __sigsetjmp(void *env);
-DEFINE_REAL(int, setjmp, void *env)
+#endif
+DEFINE_REAL(int, setjmp_symname, void *env)
 DEFINE_REAL(int, _setjmp, void *env)
-DEFINE_REAL(int, sigsetjmp, void *env)
+DEFINE_REAL(int, sigsetjmp_symname, void *env)
+#if !SANITIEZER_NETBSD
 DEFINE_REAL(int, __sigsetjmp, void *env)
+#endif
 #endif  // SANITIZER_MAC
 
-TSAN_INTERCEPTOR(void, longjmp, uptr *env, int val) {
+#if SANITIZER_NETBSD
+#define longjmp_symname __longjmp14
+#define siglongjmp_symname __siglongjmp14
+#else
+#define longjmp_symname longjmp
+#define siglongjmp_symname siglongjmp
+#endif
+
+TSAN_INTERCEPTOR(void, longjmp_symname, uptr *env, int val) {
   // Note: if we call REAL(longjmp) in the context of ScopedInterceptor,
   // bad things will happen. We will jump over ScopedInterceptor dtor and can
   // leave thr->in_ignored_lib set.
   {
-    SCOPED_INTERCEPTOR_RAW(longjmp, env, val);
+    SCOPED_INTERCEPTOR_RAW(longjmp_symname, env, val);
   }
   LongJmp(cur_thread(), env);
-  REAL(longjmp)(env, val);
+  REAL(longjmp_symname)(env, val);
 }
 
-TSAN_INTERCEPTOR(void, siglongjmp, uptr *env, int val) {
+TSAN_INTERCEPTOR(void, siglongjmp_symname, uptr *env, int val) {
   {
-    SCOPED_INTERCEPTOR_RAW(siglongjmp, env, val);
+    SCOPED_INTERCEPTOR_RAW(siglongjmp_symname, env, val);
   }
   LongJmp(cur_thread(), env);
-  REAL(siglongjmp)(env, val);
+  REAL(siglongjmp_symname)(env, val);
 }
 
+#if SANITIZER_NETBSD
+TSAN_INTERCEPTOR(void, _longjmp, uptr *env, int val) {
+  {
+    SCOPED_INTERCEPTOR_RAW(_longjmp, env, val);
+  }
+  LongJmp(cur_thread(), env);
+  REAL(_longjmp)(env, val);
+}
+#endif
+
 #if !SANITIZER_MAC
 TSAN_INTERCEPTOR(void*, malloc, uptr size) {
-  if (cur_thread()->in_symbolizer)
+  if (UNLIKELY(cur_thread()->in_symbolizer))
     return InternalAlloc(size);
   void *p = 0;
   {
@@ -584,11 +672,11 @@ TSAN_INTERCEPTOR(void*, malloc, uptr size) {
 
 TSAN_INTERCEPTOR(void*, __libc_memalign, uptr align, uptr sz) {
   SCOPED_TSAN_INTERCEPTOR(__libc_memalign, align, sz);
-  return user_alloc(thr, pc, sz, align);
+  return user_memalign(thr, pc, align, sz);
 }
 
 TSAN_INTERCEPTOR(void*, calloc, uptr size, uptr n) {
-  if (cur_thread()->in_symbolizer)
+  if (UNLIKELY(cur_thread()->in_symbolizer))
     return InternalCalloc(size, n);
   void *p = 0;
   {
@@ -600,7 +688,7 @@ TSAN_INTERCEPTOR(void*, calloc, uptr size, uptr n) {
 }
 
 TSAN_INTERCEPTOR(void*, realloc, void *p, uptr size) {
-  if (cur_thread()->in_symbolizer)
+  if (UNLIKELY(cur_thread()->in_symbolizer))
     return InternalRealloc(p, size);
   if (p)
     invoke_free_hook(p);
@@ -615,7 +703,7 @@ TSAN_INTERCEPTOR(void*, realloc, void *p, uptr size) {
 TSAN_INTERCEPTOR(void, free, void *p) {
   if (p == 0)
     return;
-  if (cur_thread()->in_symbolizer)
+  if (UNLIKELY(cur_thread()->in_symbolizer))
     return InternalFree(p);
   invoke_free_hook(p);
   SCOPED_INTERCEPTOR_RAW(free, p);
@@ -625,7 +713,7 @@ TSAN_INTERCEPTOR(void, free, void *p) {
 TSAN_INTERCEPTOR(void, cfree, void *p) {
   if (p == 0)
     return;
-  if (cur_thread()->in_symbolizer)
+  if (UNLIKELY(cur_thread()->in_symbolizer))
     return InternalFree(p);
   invoke_free_hook(p);
   SCOPED_INTERCEPTOR_RAW(cfree, p);
@@ -730,7 +818,7 @@ TSAN_INTERCEPTOR(int, munmap, void *addr, long_t sz) {
 #if SANITIZER_LINUX
 TSAN_INTERCEPTOR(void*, memalign, uptr align, uptr sz) {
   SCOPED_INTERCEPTOR_RAW(memalign, align, sz);
-  return user_alloc(thr, pc, sz, align);
+  return user_memalign(thr, pc, align, sz);
 }
 #define TSAN_MAYBE_INTERCEPT_MEMALIGN TSAN_INTERCEPT(memalign)
 #else
@@ -739,21 +827,29 @@ TSAN_INTERCEPTOR(void*, memalign, uptr align, uptr sz) {
 
 #if !SANITIZER_MAC
 TSAN_INTERCEPTOR(void*, aligned_alloc, uptr align, uptr sz) {
-  SCOPED_INTERCEPTOR_RAW(memalign, align, sz);
-  return user_alloc(thr, pc, sz, align);
+  if (UNLIKELY(cur_thread()->in_symbolizer))
+    return InternalAlloc(sz, nullptr, align);
+  SCOPED_INTERCEPTOR_RAW(aligned_alloc, align, sz);
+  return user_aligned_alloc(thr, pc, align, sz);
 }
 
 TSAN_INTERCEPTOR(void*, valloc, uptr sz) {
+  if (UNLIKELY(cur_thread()->in_symbolizer))
+    return InternalAlloc(sz, nullptr, GetPageSizeCached());
   SCOPED_INTERCEPTOR_RAW(valloc, sz);
-  return user_alloc(thr, pc, sz, GetPageSizeCached());
+  return user_valloc(thr, pc, sz);
 }
 #endif
 
 #if SANITIZER_LINUX
 TSAN_INTERCEPTOR(void*, pvalloc, uptr sz) {
+  if (UNLIKELY(cur_thread()->in_symbolizer)) {
+    uptr PageSize = GetPageSizeCached();
+    sz = sz ? RoundUpTo(sz, PageSize) : PageSize;
+    return InternalAlloc(sz, nullptr, PageSize);
+  }
   SCOPED_INTERCEPTOR_RAW(pvalloc, sz);
-  sz = RoundUp(sz, GetPageSizeCached());
-  return user_alloc(thr, pc, sz, GetPageSizeCached());
+  return user_pvalloc(thr, pc, sz);
 }
 #define TSAN_MAYBE_INTERCEPT_PVALLOC TSAN_INTERCEPT(pvalloc)
 #else
@@ -762,9 +858,15 @@ TSAN_INTERCEPTOR(void*, pvalloc, uptr sz) {
 
 #if !SANITIZER_MAC
 TSAN_INTERCEPTOR(int, posix_memalign, void **memptr, uptr align, uptr sz) {
+  if (UNLIKELY(cur_thread()->in_symbolizer)) {
+    void *p = InternalAlloc(sz, nullptr, align);
+    if (!p)
+      return errno_ENOMEM;
+    *memptr = p;
+    return 0;
+  }
   SCOPED_INTERCEPTOR_RAW(posix_memalign, memptr, align, sz);
-  *memptr = user_alloc(thr, pc, sz, align);
-  return 0;
+  return user_posix_memalign(thr, pc, memptr, align, sz);
 }
 #endif
 
@@ -830,11 +932,12 @@ void DestroyThreadState() {
 }
 }  // namespace __tsan
 
-#if !SANITIZER_MAC
+#if !SANITIZER_MAC && !SANITIZER_NETBSD
 static void thread_finalize(void *v) {
   uptr iter = (uptr)v;
   if (iter > 1) {
-    if (pthread_setspecific(g_thread_finalize_key, (void*)(iter - 1))) {
+    if (pthread_setspecific(interceptor_ctx()->finalize_key,
+        (void*)(iter - 1))) {
       Printf("ThreadSanitizer: failed to set thread key\n");
       Die();
     }
@@ -860,9 +963,9 @@ extern "C" void *__tsan_thread_start_func(void *arg) {
     ThreadState *thr = cur_thread();
     // Thread-local state is not initialized yet.
     ScopedIgnoreInterceptors ignore;
-#if !SANITIZER_MAC
+#if !SANITIZER_MAC && !SANITIZER_NETBSD
     ThreadIgnoreBegin(thr, 0);
-    if (pthread_setspecific(g_thread_finalize_key,
+    if (pthread_setspecific(interceptor_ctx()->finalize_key,
                             (void *)GetPthreadDestructorIterations())) {
       Printf("ThreadSanitizer: failed to set thread key\n");
       Die();
@@ -887,6 +990,9 @@ extern "C" void *__tsan_thread_start_func(void *arg) {
 TSAN_INTERCEPTOR(int, pthread_create,
     void *th, void *attr, void *(*callback)(void*), void * param) {
   SCOPED_INTERCEPTOR_RAW(pthread_create, th, attr, callback, param);
+
+  MaybeSpawnBackgroundThread();
+
   if (ctx->after_multithreaded_fork) {
     if (flags()->die_after_fork) {
       Report("ThreadSanitizer: starting new threads after multi-threaded "
@@ -1312,10 +1418,15 @@ TSAN_INTERCEPTOR(int, pthread_once, void *o, void (*f)()) {
   if (o == 0 || f == 0)
     return errno_EINVAL;
   atomic_uint32_t *a;
-  if (!SANITIZER_MAC)
-    a = static_cast<atomic_uint32_t*>(o);
-  else  // On OS X, pthread_once_t has a header with a long-sized signature.
+
+  if (SANITIZER_MAC)
     a = static_cast<atomic_uint32_t*>((void *)((char *)o + sizeof(long_t)));
+  else if (SANITIZER_NETBSD)
+    a = static_cast<atomic_uint32_t*>
+          ((void *)((char *)o + __sanitizer::pthread_mutex_t_sz));
+  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)) {
@@ -1347,7 +1458,7 @@ TSAN_INTERCEPTOR(int, __fxstat, int version, int fd, void *buf) {
 #endif
 
 TSAN_INTERCEPTOR(int, fstat, int fd, void *buf) {
-#if SANITIZER_FREEBSD || SANITIZER_MAC || SANITIZER_ANDROID
+#if SANITIZER_FREEBSD || SANITIZER_MAC || SANITIZER_ANDROID || SANITIZER_NETBSD
   SCOPED_TSAN_INTERCEPTOR(fstat, fd, buf);
   if (fd > 0)
     FdAccess(thr, pc, fd);
@@ -1762,7 +1873,9 @@ TSAN_INTERCEPTOR(int, pthread_sigmask, int how, const __sanitizer_sigset_t *set,
 namespace __tsan {
 
 static void CallUserSignalHandler(ThreadState *thr, bool sync, bool acquire,
-    bool sigact, int sig, my_siginfo_t *info, void *uctx) {
+                                  bool sigact, 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
@@ -1784,14 +1897,13 @@ 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].sa_sigaction :
-     (uptr)sigactions[sig].sa_handler;
-  if (pc != (uptr)SIG_DFL && pc != (uptr)SIG_IGN) {
+  volatile uptr pc =
+      sigact ? (uptr)sigactions[sig].sigaction : (uptr)sigactions[sig].handler;
+  if (pc != sig_dfl && pc != sig_ign) {
     if (sigact)
-      ((sigactionhandler_t)pc)(sig, info, uctx);
+      ((__sanitizer_sigactionhandler_ptr)pc)(sig, info, uctx);
     else
-      ((sighandler_t)pc)(sig);
+      ((__sanitizer_sighandler_ptr)pc)(sig);
   }
   if (!ctx->after_multithreaded_fork) {
     thr->ignore_reads_and_writes = ignore_reads_and_writes;
@@ -1855,7 +1967,8 @@ static bool is_sync_signal(ThreadSignalContext *sctx, int sig) {
 }
 
 void ALWAYS_INLINE rtl_generic_sighandler(bool sigact, int sig,
-    my_siginfo_t *info, void *ctx) {
+                                          __sanitizer_siginfo *info,
+                                          void *ctx) {
   ThreadState *thr = cur_thread();
   ThreadSignalContext *sctx = SigCtx(thr);
   if (sig < 0 || sig >= kSigCount) {
@@ -1905,58 +2018,10 @@ static void rtl_sighandler(int sig) {
   rtl_generic_sighandler(false, sig, 0, 0);
 }
 
-static void rtl_sigaction(int sig, my_siginfo_t *info, void *ctx) {
+static void rtl_sigaction(int sig, __sanitizer_siginfo *info, void *ctx) {
   rtl_generic_sighandler(true, sig, info, ctx);
 }
 
-TSAN_INTERCEPTOR(int, sigaction, int sig, sigaction_t *act, sigaction_t *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 handler will run synchronously and corrupt tsan per-thread state.
-  SCOPED_INTERCEPTOR_RAW(sigaction, sig, act, old);
-  if (old)
-    internal_memcpy(old, &sigactions[sig], sizeof(*old));
-  if (act == 0)
-    return 0;
-  // Copy act into sigactions[sig].
-  // Can't use struct copy, because compiler can emit call to memcpy.
-  // Can't use internal_memcpy, because it copies byte-by-byte,
-  // and signal handler reads the sa_handler concurrently. It it can read
-  // some bytes from old value and some bytes from new value.
-  // Use volatile to prevent insertion of memcpy.
-  sigactions[sig].sa_handler = *(volatile sighandler_t*)&act->sa_handler;
-  sigactions[sig].sa_flags = *(volatile int*)&act->sa_flags;
-  internal_memcpy(&sigactions[sig].sa_mask, &act->sa_mask,
-      sizeof(sigactions[sig].sa_mask));
-#if !SANITIZER_FREEBSD && !SANITIZER_MAC
-  sigactions[sig].sa_restorer = act->sa_restorer;
-#endif
-  sigaction_t newact;
-  internal_memcpy(&newact, act, sizeof(newact));
-  internal_sigfillset(&newact.sa_mask);
-  if (act->sa_handler != SIG_IGN && act->sa_handler != SIG_DFL) {
-    if (newact.sa_flags & SA_SIGINFO)
-      newact.sa_sigaction = rtl_sigaction;
-    else
-      newact.sa_handler = rtl_sighandler;
-  }
-  ReleaseStore(thr, pc, (uptr)&sigactions[sig]);
-  int res = REAL(sigaction)(sig, &newact, 0);
-  return res;
-}
-
-TSAN_INTERCEPTOR(sighandler_t, signal, int sig, sighandler_t h) {
-  sigaction_t act;
-  act.sa_handler = h;
-  internal_memset(&act.sa_mask, -1, sizeof(act.sa_mask));
-  act.sa_flags = 0;
-  sigaction_t old;
-  int res = sigaction(sig, &act, &old);
-  if (res)
-    return SIG_ERR;
-  return old.sa_handler;
-}
-
 TSAN_INTERCEPTOR(int, raise, int sig) {
   SCOPED_TSAN_INTERCEPTOR(raise, sig);
   ThreadSignalContext *sctx = SigCtx(thr);
@@ -2020,7 +2085,7 @@ TSAN_INTERCEPTOR(int, getaddrinfo, void *node, void *service,
 }
 
 TSAN_INTERCEPTOR(int, fork, int fake) {
-  if (cur_thread()->in_symbolizer)
+  if (UNLIKELY(cur_thread()->in_symbolizer))
     return REAL(fork)(fake);
   SCOPED_INTERCEPTOR_RAW(fork, fake);
   ForkBefore(thr, pc);
@@ -2270,6 +2335,77 @@ static void HandleRecvmsg(ThreadState *thr, uptr pc,
 
 #include "sanitizer_common/sanitizer_common_interceptors.inc"
 
+static int sigaction_impl(int sig, const __sanitizer_sigaction *act,
+                          __sanitizer_sigaction *old);
+static __sanitizer_sighandler_ptr signal_impl(int sig,
+                                              __sanitizer_sighandler_ptr h);
+
+#define SIGNAL_INTERCEPTOR_SIGACTION_IMPL(signo, act, oldact) \
+  { return sigaction_impl(signo, act, oldact); }
+
+#define SIGNAL_INTERCEPTOR_SIGNAL_IMPL(func, signo, handler) \
+  { return (uptr)signal_impl(signo, (__sanitizer_sighandler_ptr)handler); }
+
+#include "sanitizer_common/sanitizer_signal_interceptors.inc"
+
+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 handler will run synchronously and corrupt tsan per-thread state.
+  SCOPED_INTERCEPTOR_RAW(sigaction, sig, act, old);
+  __sanitizer_sigaction *sigactions = interceptor_ctx()->sigactions;
+  __sanitizer_sigaction old_stored;
+  if (old) internal_memcpy(&old_stored, &sigactions[sig], sizeof(old_stored));
+  __sanitizer_sigaction newact;
+  if (act) {
+    // Copy act into sigactions[sig].
+    // Can't use struct copy, because compiler can emit call to memcpy.
+    // Can't use internal_memcpy, because it copies byte-by-byte,
+    // and signal handler reads the handler concurrently. It it can read
+    // some bytes from old value and some bytes from new value.
+    // Use volatile to prevent insertion of memcpy.
+    sigactions[sig].handler =
+        *(volatile __sanitizer_sighandler_ptr const *)&act->handler;
+    sigactions[sig].sa_flags = *(volatile int const *)&act->sa_flags;
+    internal_memcpy(&sigactions[sig].sa_mask, &act->sa_mask,
+                    sizeof(sigactions[sig].sa_mask));
+#if !SANITIZER_FREEBSD && !SANITIZER_MAC && !SANITIZER_NETBSD
+    sigactions[sig].sa_restorer = act->sa_restorer;
+#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;
+    }
+    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));
+    }
+  }
+  return res;
+}
+
+static __sanitizer_sighandler_ptr signal_impl(int sig,
+                                              __sanitizer_sighandler_ptr h) {
+  __sanitizer_sigaction act;
+  act.handler = h;
+  internal_memset(&act.sa_mask, -1, sizeof(act.sa_mask));
+  act.sa_flags = 0;
+  __sanitizer_sigaction old;
+  int res = sigaction_symname(sig, &act, &old);
+  if (res) return (__sanitizer_sighandler_ptr)sig_err;
+  return old.handler;
+}
+
 #define TSAN_SYSCALL() \
   ThreadState *thr = cur_thread(); \
   if (thr->ignore_interceptors) \
@@ -2290,7 +2426,7 @@ struct ScopedSyscall {
   }
 };
 
-#if !SANITIZER_FREEBSD && !SANITIZER_MAC
+#if !SANITIZER_FREEBSD && !SANITIZER_MAC && !SANITIZER_NETBSD
 static void syscall_access_range(uptr pc, uptr p, uptr s, bool write) {
   TSAN_SYSCALL();
   MemoryAccessRange(thr, pc, p, s, write);
@@ -2409,6 +2545,34 @@ TSAN_INTERCEPTOR(void *, __tls_get_addr, void *arg) {
 }
 #endif
 
+#if SANITIZER_NETBSD
+TSAN_INTERCEPTOR(void, _lwp_exit) {
+  SCOPED_TSAN_INTERCEPTOR(_lwp_exit);
+  DestroyThreadState();
+  REAL(_lwp_exit)();
+}
+#define TSAN_MAYBE_INTERCEPT__LWP_EXIT TSAN_INTERCEPT(_lwp_exit)
+#else
+#define TSAN_MAYBE_INTERCEPT__LWP_EXIT
+#endif
+
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, cond_init, void *c, void *a);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, cond_signal, void *c);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, cond_broadcast, void *c);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, cond_wait, void *c, void *m);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, cond_destroy, void *c);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, mutex_init, void *m, void *a);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, mutex_destroy, void *m);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, mutex_trylock, void *m);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, rwlock_init, void *m, void *a);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, rwlock_destroy, void *m);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, rwlock_rdlock, void *m);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, rwlock_tryrdlock, void *m);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, rwlock_wrlock, void *m);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, rwlock_trywrlock, void *m);
+TSAN_INTERCEPTOR_NETBSD_ALIAS(int, rwlock_unlock, void *m);
+TSAN_INTERCEPTOR_NETBSD_ALIAS_THR(int, once, void *o, void (*f)());
+
 namespace __tsan {
 
 static void finalize(void *arg) {
@@ -2440,20 +2604,30 @@ void InitializeInterceptors() {
   mallopt(-3, 32*1024);  // M_MMAP_THRESHOLD
 #endif
 
+  new(interceptor_ctx()) InterceptorContext();
+
   InitializeCommonInterceptors();
+  InitializeSignalInterceptors();
 
 #if !SANITIZER_MAC
   // We can not use TSAN_INTERCEPT to get setjmp addr,
   // because it does &setjmp and setjmp is not present in some versions of libc.
   using __interception::GetRealFunctionAddress;
-  GetRealFunctionAddress("setjmp", (uptr*)&REAL(setjmp), 0, 0);
+  GetRealFunctionAddress(TSAN_STRING_SETJMP,
+                         (uptr*)&REAL(setjmp_symname), 0, 0);
   GetRealFunctionAddress("_setjmp", (uptr*)&REAL(_setjmp), 0, 0);
-  GetRealFunctionAddress("sigsetjmp", (uptr*)&REAL(sigsetjmp), 0, 0);
+  GetRealFunctionAddress(TSAN_STRING_SIGSETJMP,
+                         (uptr*)&REAL(sigsetjmp_symname), 0, 0);
+#if !SANITIZER_NETBSD
   GetRealFunctionAddress("__sigsetjmp", (uptr*)&REAL(__sigsetjmp), 0, 0);
 #endif
+#endif
 
-  TSAN_INTERCEPT(longjmp);
-  TSAN_INTERCEPT(siglongjmp);
+  TSAN_INTERCEPT(longjmp_symname);
+  TSAN_INTERCEPT(siglongjmp_symname);
+#if SANITIZER_NETBSD
+  TSAN_INTERCEPT(_longjmp);
+#endif
 
   TSAN_INTERCEPT(malloc);
   TSAN_INTERCEPT(__libc_memalign);
@@ -2548,8 +2722,6 @@ void InitializeInterceptors() {
   TSAN_INTERCEPT(rmdir);
   TSAN_INTERCEPT(closedir);
 
-  TSAN_INTERCEPT(sigaction);
-  TSAN_INTERCEPT(signal);
   TSAN_INTERCEPT(sigsuspend);
   TSAN_INTERCEPT(sigblock);
   TSAN_INTERCEPT(sigsetmask);
@@ -2560,6 +2732,7 @@ void InitializeInterceptors() {
   TSAN_INTERCEPT(sleep);
   TSAN_INTERCEPT(usleep);
   TSAN_INTERCEPT(nanosleep);
+  TSAN_INTERCEPT(pause);
   TSAN_INTERCEPT(gettimeofday);
   TSAN_INTERCEPT(getaddrinfo);
 
@@ -2568,7 +2741,7 @@ void InitializeInterceptors() {
 #if !SANITIZER_ANDROID
   TSAN_INTERCEPT(dl_iterate_phdr);
 #endif
-  TSAN_INTERCEPT(on_exit);
+  TSAN_MAYBE_INTERCEPT_ON_EXIT;
   TSAN_INTERCEPT(__cxa_atexit);
   TSAN_INTERCEPT(_exit);
 
@@ -2576,6 +2749,8 @@ void InitializeInterceptors() {
   TSAN_INTERCEPT(__tls_get_addr);
 #endif
 
+  TSAN_MAYBE_INTERCEPT__LWP_EXIT;
+
 #if !SANITIZER_MAC && !SANITIZER_ANDROID
   // Need to setup it, because interceptors check that the function is resolved.
   // But atexit is emitted directly into the module, so can't be resolved.
@@ -2587,13 +2762,30 @@ void InitializeInterceptors() {
     Die();
   }
 
-#if !SANITIZER_MAC
-  if (pthread_key_create(&g_thread_finalize_key, &thread_finalize)) {
+#if !SANITIZER_MAC && !SANITIZER_NETBSD
+  if (pthread_key_create(&interceptor_ctx()->finalize_key, &thread_finalize)) {
     Printf("ThreadSanitizer: failed to create thread key\n");
     Die();
   }
 #endif
 
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(cond_init);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(cond_signal);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(cond_broadcast);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(cond_wait);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(cond_destroy);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(mutex_init);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(mutex_destroy);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(mutex_trylock);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(rwlock_init);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(rwlock_destroy);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(rwlock_rdlock);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(rwlock_tryrdlock);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(rwlock_wrlock);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(rwlock_trywrlock);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS(rwlock_unlock);
+  TSAN_MAYBE_INTERCEPT_NETBSD_ALIAS_THR(once);
+
   FdInit();
 }
 
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors.h b/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors.h
index de47466501da..959a39465e33 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors.h
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors.h
@@ -49,4 +49,16 @@ LibIgnore *libignore();
 
 #define TSAN_INTERCEPTOR(ret, func, ...) INTERCEPTOR(ret, func, __VA_ARGS__)
 
+#if SANITIZER_NETBSD
+# define TSAN_INTERCEPTOR_NETBSD_ALIAS(ret, func, ...) \
+  TSAN_INTERCEPTOR(ret, __libc_##func, __VA_ARGS__) \
+  ALIAS(WRAPPER_NAME(pthread_##func));
+# define TSAN_INTERCEPTOR_NETBSD_ALIAS_THR(ret, func, ...) \
+  TSAN_INTERCEPTOR(ret, __libc_thr_##func, __VA_ARGS__) \
+  ALIAS(WRAPPER_NAME(pthread_##func));
+#else
+# define TSAN_INTERCEPTOR_NETBSD_ALIAS(ret, func, ...)
+# define TSAN_INTERCEPTOR_NETBSD_ALIAS_THR(ret, func, ...)
+#endif
+
 #endif  // TSAN_INTERCEPTORS_H
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors_mac.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors_mac.cc
index 4f1079467331..3b3a4a2815c3 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors_mac.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_interceptors_mac.cc
@@ -100,7 +100,7 @@ OSATOMIC_INTERCEPTORS_BITWISE(OSAtomicXor, fetch_xor,
   TSAN_INTERCEPTOR(bool, f, t old_value, t new_value, t volatile *ptr) {    \
     SCOPED_TSAN_INTERCEPTOR(f, old_value, new_value, ptr);                  \
     return tsan_atomic_f##_compare_exchange_strong(                         \
-        (tsan_t *)ptr, (tsan_t *)&old_value, (tsan_t)new_value,             \
+        (volatile tsan_t *)ptr, (tsan_t *)&old_value, (tsan_t)new_value,    \
         kMacOrderNonBarrier, kMacOrderNonBarrier);                          \
   }                                                                         \
                                                                             \
@@ -108,7 +108,7 @@ OSATOMIC_INTERCEPTORS_BITWISE(OSAtomicXor, fetch_xor,
                    t volatile *ptr) {                                       \
     SCOPED_TSAN_INTERCEPTOR(f##Barrier, old_value, new_value, ptr);         \
     return tsan_atomic_f##_compare_exchange_strong(                         \
-        (tsan_t *)ptr, (tsan_t *)&old_value, (tsan_t)new_value,             \
+        (volatile tsan_t *)ptr, (tsan_t *)&old_value, (tsan_t)new_value,    \
         kMacOrderBarrier, kMacOrderNonBarrier);                             \
   }
 
@@ -122,14 +122,14 @@ OSATOMIC_INTERCEPTORS_CAS(OSAtomicCompareAndSwap32, __tsan_atomic32, a32,
 OSATOMIC_INTERCEPTORS_CAS(OSAtomicCompareAndSwap64, __tsan_atomic64, a64,
                           int64_t)
 
-#define OSATOMIC_INTERCEPTOR_BITOP(f, op, clear, mo)          \
-  TSAN_INTERCEPTOR(bool, f, uint32_t n, volatile void *ptr) { \
-    SCOPED_TSAN_INTERCEPTOR(f, n, ptr);                       \
-    char *byte_ptr = ((char *)ptr) + (n >> 3);                \
-    char bit = 0x80u >> (n & 7);                              \
-    char mask = clear ? ~bit : bit;                           \
-    char orig_byte = op((a8 *)byte_ptr, mask, mo);            \
-    return orig_byte & bit;                                   \
+#define OSATOMIC_INTERCEPTOR_BITOP(f, op, clear, mo)             \
+  TSAN_INTERCEPTOR(bool, f, uint32_t n, volatile void *ptr) {    \
+    SCOPED_TSAN_INTERCEPTOR(f, n, ptr);                          \
+    volatile char *byte_ptr = ((volatile char *)ptr) + (n >> 3); \
+    char bit = 0x80u >> (n & 7);                                 \
+    char mask = clear ? ~bit : bit;                              \
+    char orig_byte = op((volatile a8 *)byte_ptr, mask, mo);      \
+    return orig_byte & bit;                                      \
   }
 
 #define OSATOMIC_INTERCEPTORS_BITOP(f, op, clear)               \
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_interface_ann.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_interface_ann.cc
index f68a0468de53..a7922a42e42c 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_interface_ann.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_interface_ann.cc
@@ -14,6 +14,7 @@
 #include "sanitizer_common/sanitizer_internal_defs.h"
 #include "sanitizer_common/sanitizer_placement_new.h"
 #include "sanitizer_common/sanitizer_stacktrace.h"
+#include "sanitizer_common/sanitizer_vector.h"
 #include "tsan_interface_ann.h"
 #include "tsan_mutex.h"
 #include "tsan_report.h"
@@ -21,7 +22,6 @@
 #include "tsan_mman.h"
 #include "tsan_flags.h"
 #include "tsan_platform.h"
-#include "tsan_vector.h"
 
 #define CALLERPC ((uptr)__builtin_return_address(0))
 
@@ -185,10 +185,10 @@ void PrintMatchedBenignRaces() {
   int unique_count = 0;
   int hit_count = 0;
   int add_count = 0;
-  Vector<ExpectRace> hit_matched(MBlockScopedBuf);
+  Vector<ExpectRace> hit_matched;
   CollectMatchedBenignRaces(&hit_matched, &unique_count, &hit_count,
       &ExpectRace::hitcount);
-  Vector<ExpectRace> add_matched(MBlockScopedBuf);
+  Vector<ExpectRace> add_matched;
   CollectMatchedBenignRaces(&add_matched, &unique_count, &add_count,
       &ExpectRace::addcount);
   if (hit_matched.Size()) {
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_libdispatch_mac.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_libdispatch_mac.cc
index 8c759a3be4e1..eb22e4baa710 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_libdispatch_mac.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_libdispatch_mac.cc
@@ -86,7 +86,8 @@ static tsan_block_context_t *AllocContext(ThreadState *thr, uptr pc,
                                           void *orig_context,
                                           dispatch_function_t orig_work) {
   tsan_block_context_t *new_context =
-      (tsan_block_context_t *)user_alloc(thr, pc, sizeof(tsan_block_context_t));
+      (tsan_block_context_t *)user_alloc_internal(thr, pc,
+                                                  sizeof(tsan_block_context_t));
   new_context->queue = queue;
   new_context->orig_context = orig_context;
   new_context->orig_work = orig_work;
@@ -176,7 +177,8 @@ static void invoke_and_release_block(void *param) {
   }
 
 #define DISPATCH_INTERCEPT_SYNC_B(name, barrier)                             \
-  TSAN_INTERCEPTOR(void, name, dispatch_queue_t q, dispatch_block_t block) { \
+  TSAN_INTERCEPTOR(void, name, dispatch_queue_t q,                           \
+                   DISPATCH_NOESCAPE dispatch_block_t block) {               \
     SCOPED_TSAN_INTERCEPTOR(name, q, block);                                 \
     SCOPED_TSAN_INTERCEPTOR_USER_CALLBACK_START();                           \
     dispatch_block_t heap_block = Block_copy(block);                         \
@@ -266,7 +268,7 @@ TSAN_INTERCEPTOR(void, dispatch_after_f, dispatch_time_t when,
 // need to undefine the macro.
 #undef dispatch_once
 TSAN_INTERCEPTOR(void, dispatch_once, dispatch_once_t *predicate,
-                 dispatch_block_t block) {
+                 DISPATCH_NOESCAPE dispatch_block_t block) {
   SCOPED_INTERCEPTOR_RAW(dispatch_once, predicate, block);
   atomic_uint32_t *a = reinterpret_cast<atomic_uint32_t *>(predicate);
   u32 v = atomic_load(a, memory_order_acquire);
@@ -476,7 +478,8 @@ TSAN_INTERCEPTOR(void, dispatch_source_set_registration_handler_f,
 }
 
 TSAN_INTERCEPTOR(void, dispatch_apply, size_t iterations,
-                 dispatch_queue_t queue, void (^block)(size_t)) {
+                 dispatch_queue_t queue,
+                 DISPATCH_NOESCAPE void (^block)(size_t)) {
   SCOPED_TSAN_INTERCEPTOR(dispatch_apply, iterations, queue, block);
 
   void *parent_to_child_sync = nullptr;
@@ -519,9 +522,9 @@ TSAN_INTERCEPTOR(dispatch_data_t, dispatch_data_create, const void *buffer,
     return REAL(dispatch_data_create)(buffer, size, q, destructor);
 
   if (destructor == DISPATCH_DATA_DESTRUCTOR_FREE)
-    destructor = ^(void) { WRAP(free)((void *)buffer); };
+    destructor = ^(void) { WRAP(free)((void *)(uintptr_t)buffer); };
   else if (destructor == DISPATCH_DATA_DESTRUCTOR_MUNMAP)
-    destructor = ^(void) { WRAP(munmap)((void *)buffer, size); };
+    destructor = ^(void) { WRAP(munmap)((void *)(uintptr_t)buffer, size); };
 
   SCOPED_TSAN_INTERCEPTOR_USER_CALLBACK_START();
   dispatch_block_t heap_block = Block_copy(destructor);
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_malloc_mac.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_malloc_mac.cc
index 8d31ccbcaaca..455c95df6c50 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_malloc_mac.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_malloc_mac.cc
@@ -26,7 +26,7 @@ using namespace __tsan;
 #define COMMON_MALLOC_FORCE_UNLOCK()
 #define COMMON_MALLOC_MEMALIGN(alignment, size) \
   void *p =                                     \
-      user_alloc(cur_thread(), StackTrace::GetCurrentPc(), size, alignment)
+      user_memalign(cur_thread(), StackTrace::GetCurrentPc(), alignment, size)
 #define COMMON_MALLOC_MALLOC(size)                             \
   if (cur_thread()->in_symbolizer) return InternalAlloc(size); \
   SCOPED_INTERCEPTOR_RAW(malloc, size);                        \
@@ -43,7 +43,7 @@ using namespace __tsan;
   if (cur_thread()->in_symbolizer)                            \
     return InternalAlloc(size, nullptr, GetPageSizeCached()); \
   SCOPED_INTERCEPTOR_RAW(valloc, size);                       \
-  void *p = user_alloc(thr, pc, size, GetPageSizeCached())
+  void *p = user_valloc(thr, pc, size)
 #define COMMON_MALLOC_FREE(ptr)                              \
   if (cur_thread()->in_symbolizer) return InternalFree(ptr); \
   SCOPED_INTERCEPTOR_RAW(free, ptr);                         \
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_mman.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_mman.cc
index f79dccddba9f..19680238bf76 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_mman.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_mman.cc
@@ -13,6 +13,7 @@
 #include "sanitizer_common/sanitizer_allocator_checks.h"
 #include "sanitizer_common/sanitizer_allocator_interface.h"
 #include "sanitizer_common/sanitizer_common.h"
+#include "sanitizer_common/sanitizer_errno.h"
 #include "sanitizer_common/sanitizer_placement_new.h"
 #include "tsan_mman.h"
 #include "tsan_rtl.h"
@@ -149,11 +150,12 @@ static void SignalUnsafeCall(ThreadState *thr, uptr pc) {
   OutputReport(thr, rep);
 }
 
-void *user_alloc(ThreadState *thr, uptr pc, uptr sz, uptr align, bool signal) {
+void *user_alloc_internal(ThreadState *thr, uptr pc, uptr sz, uptr align,
+                          bool signal) {
   if ((sz >= (1ull << 40)) || (align >= (1ull << 40)))
     return Allocator::FailureHandler::OnBadRequest();
   void *p = allocator()->Allocate(&thr->proc()->alloc_cache, sz, align);
-  if (p == 0)
+  if (UNLIKELY(p == 0))
     return 0;
   if (ctx && ctx->initialized)
     OnUserAlloc(thr, pc, (uptr)p, sz, true);
@@ -162,15 +164,6 @@ void *user_alloc(ThreadState *thr, uptr pc, uptr sz, uptr align, bool signal) {
   return p;
 }
 
-void *user_calloc(ThreadState *thr, uptr pc, uptr size, uptr n) {
-  if (CheckForCallocOverflow(size, n))
-    return Allocator::FailureHandler::OnBadRequest();
-  void *p = user_alloc(thr, pc, n * size);
-  if (p)
-    internal_memset(p, 0, n * size);
-  return p;
-}
-
 void user_free(ThreadState *thr, uptr pc, void *p, bool signal) {
   ScopedGlobalProcessor sgp;
   if (ctx && ctx->initialized)
@@ -180,6 +173,19 @@ void user_free(ThreadState *thr, uptr pc, void *p, bool signal) {
     SignalUnsafeCall(thr, pc);
 }
 
+void *user_alloc(ThreadState *thr, uptr pc, uptr sz) {
+  return SetErrnoOnNull(user_alloc_internal(thr, pc, sz, kDefaultAlignment));
+}
+
+void *user_calloc(ThreadState *thr, uptr pc, uptr size, uptr n) {
+  if (UNLIKELY(CheckForCallocOverflow(size, n)))
+    return SetErrnoOnNull(Allocator::FailureHandler::OnBadRequest());
+  void *p = user_alloc_internal(thr, pc, n * size);
+  if (p)
+    internal_memset(p, 0, n * size);
+  return SetErrnoOnNull(p);
+}
+
 void OnUserAlloc(ThreadState *thr, uptr pc, uptr p, uptr sz, bool write) {
   DPrintf("#%d: alloc(%zu) = %p\n", thr->tid, sz, p);
   ctx->metamap.AllocBlock(thr, pc, p, sz);
@@ -200,15 +206,64 @@ void OnUserFree(ThreadState *thr, uptr pc, uptr p, bool write) {
 void *user_realloc(ThreadState *thr, uptr pc, void *p, uptr sz) {
   // FIXME: Handle "shrinking" more efficiently,
   // it seems that some software actually does this.
-  void *p2 = user_alloc(thr, pc, sz);
-  if (p2 == 0)
-    return 0;
-  if (p) {
-    uptr oldsz = user_alloc_usable_size(p);
-    internal_memcpy(p2, p, min(oldsz, sz));
+  if (!p)
+    return SetErrnoOnNull(user_alloc_internal(thr, pc, sz));
+  if (!sz) {
+    user_free(thr, pc, p);
+    return nullptr;
+  }
+  void *new_p = user_alloc_internal(thr, pc, sz);
+  if (new_p) {
+    uptr old_sz = user_alloc_usable_size(p);
+    internal_memcpy(new_p, p, min(old_sz, sz));
     user_free(thr, pc, p);
   }
-  return p2;
+  return SetErrnoOnNull(new_p);
+}
+
+void *user_memalign(ThreadState *thr, uptr pc, uptr align, uptr sz) {
+  if (UNLIKELY(!IsPowerOfTwo(align))) {
+    errno = errno_EINVAL;
+    return Allocator::FailureHandler::OnBadRequest();
+  }
+  return SetErrnoOnNull(user_alloc_internal(thr, pc, sz, align));
+}
+
+int user_posix_memalign(ThreadState *thr, uptr pc, void **memptr, uptr align,
+                        uptr sz) {
+  if (UNLIKELY(!CheckPosixMemalignAlignment(align))) {
+    Allocator::FailureHandler::OnBadRequest();
+    return errno_EINVAL;
+  }
+  void *ptr = user_alloc_internal(thr, pc, sz, align);
+  if (UNLIKELY(!ptr))
+    return errno_ENOMEM;
+  CHECK(IsAligned((uptr)ptr, align));
+  *memptr = ptr;
+  return 0;
+}
+
+void *user_aligned_alloc(ThreadState *thr, uptr pc, uptr align, uptr sz) {
+  if (UNLIKELY(!CheckAlignedAllocAlignmentAndSize(align, sz))) {
+    errno = errno_EINVAL;
+    return Allocator::FailureHandler::OnBadRequest();
+  }
+  return SetErrnoOnNull(user_alloc_internal(thr, pc, sz, align));
+}
+
+void *user_valloc(ThreadState *thr, uptr pc, uptr sz) {
+  return SetErrnoOnNull(user_alloc_internal(thr, pc, sz, GetPageSizeCached()));
+}
+
+void *user_pvalloc(ThreadState *thr, uptr pc, uptr sz) {
+  uptr PageSize = GetPageSizeCached();
+  if (UNLIKELY(CheckForPvallocOverflow(sz, PageSize))) {
+    errno = errno_ENOMEM;
+    return Allocator::FailureHandler::OnBadRequest();
+  }
+  // pvalloc(0) should allocate one page.
+  sz = sz ? RoundUpTo(sz, PageSize) : PageSize;
+  return SetErrnoOnNull(user_alloc_internal(thr, pc, sz, PageSize));
 }
 
 uptr user_alloc_usable_size(const void *p) {
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_mman.h b/contrib/compiler-rt/lib/tsan/rtl/tsan_mman.h
index 8cdeeb35aa6b..6042c5c5d96d 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_mman.h
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_mman.h
@@ -27,13 +27,20 @@ void AllocatorProcFinish(Processor *proc);
 void AllocatorPrintStats();
 
 // For user allocations.
-void *user_alloc(ThreadState *thr, uptr pc, uptr sz,
-                 uptr align = kDefaultAlignment, bool signal = true);
-void *user_calloc(ThreadState *thr, uptr pc, uptr sz, uptr n);
+void *user_alloc_internal(ThreadState *thr, uptr pc, uptr sz,
+                          uptr align = kDefaultAlignment, bool signal = true);
 // Does not accept NULL.
 void user_free(ThreadState *thr, uptr pc, void *p, bool signal = true);
+// Interceptor implementations.
+void *user_alloc(ThreadState *thr, uptr pc, uptr sz);
+void *user_calloc(ThreadState *thr, uptr pc, uptr sz, uptr n);
 void *user_realloc(ThreadState *thr, uptr pc, void *p, uptr sz);
-void *user_alloc_aligned(ThreadState *thr, uptr pc, uptr sz, uptr align);
+void *user_memalign(ThreadState *thr, uptr pc, uptr align, uptr sz);
+int user_posix_memalign(ThreadState *thr, uptr pc, void **memptr, uptr align,
+                        uptr sz);
+void *user_aligned_alloc(ThreadState *thr, uptr pc, uptr align, uptr sz);
+void *user_valloc(ThreadState *thr, uptr pc, uptr sz);
+void *user_pvalloc(ThreadState *thr, uptr pc, uptr sz);
 uptr user_alloc_usable_size(const void *p);
 
 // Invoking malloc/free hooks that may be installed by the user.
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_platform.h b/contrib/compiler-rt/lib/tsan/rtl/tsan_platform.h
index bea1daba3952..bfac70df12d9 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_platform.h
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_platform.h
@@ -42,6 +42,19 @@ C/C++ on linux/x86_64 and freebsd/x86_64
 7b00 0000 0000 - 7c00 0000 0000: heap
 7c00 0000 0000 - 7e80 0000 0000: -
 7e80 0000 0000 - 8000 0000 0000: modules and main thread stack
+
+C/C++ on netbsd/amd64 can reuse the same mapping:
+ * The address space starts from 0x1000 (option with 0x0) and ends with
+   0x7f7ffffff000.
+ * LoAppMem-kHeapMemEnd can be reused as it is.
+ * No VDSO support.
+ * No MidAppMem region.
+ * No additional HeapMem region.
+ * HiAppMem contains the stack, loader, shared libraries and heap.
+ * Stack on NetBSD/amd64 has prereserved 128MB.
+ * Heap grows downwards (top-down).
+ * ASLR must be disabled per-process or globally.
+
 */
 struct Mapping {
   static const uptr kMetaShadowBeg = 0x300000000000ull;
@@ -111,7 +124,8 @@ C/C++ on Darwin/iOS/ARM64 (36-bit VMA, 64 GB VM)
 0c00 0000 00 - 0d00 0000 00: -                                    (4 GB)
 0d00 0000 00 - 0e00 0000 00: metainfo                             (4 GB)
 0e00 0000 00 - 0f00 0000 00: -                                    (4 GB)
-0f00 0000 00 - 1000 0000 00: traces                               (4 GB)
+0f00 0000 00 - 0fc0 0000 00: traces                               (3 GB)
+0fc0 0000 00 - 1000 0000 00: -
 */
 struct Mapping {
   static const uptr kLoAppMemBeg   = 0x0100000000ull;
@@ -123,9 +137,9 @@ struct Mapping {
   static const uptr kMetaShadowBeg = 0x0d00000000ull;
   static const uptr kMetaShadowEnd = 0x0e00000000ull;
   static const uptr kTraceMemBeg   = 0x0f00000000ull;
-  static const uptr kTraceMemEnd   = 0x1000000000ull;
-  static const uptr kHiAppMemBeg   = 0x1000000000ull;
-  static const uptr kHiAppMemEnd   = 0x1000000000ull;
+  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 kVdsoBeg       = 0x7000000000000000ull;
@@ -303,6 +317,38 @@ struct Mapping46 {
   static const uptr kVdsoBeg       = 0x7800000000000000ull;
 };
 
+/*
+C/C++ on linux/powerpc64 (47-bit VMA)
+0000 0000 1000 - 0100 0000 0000: main binary
+0100 0000 0000 - 0200 0000 0000: -
+0100 0000 0000 - 1000 0000 0000: shadow
+1000 0000 0000 - 1000 0000 0000: -
+1000 0000 0000 - 2000 0000 0000: metainfo (memory blocks and sync objects)
+2000 0000 0000 - 2000 0000 0000: -
+2000 0000 0000 - 2200 0000 0000: traces
+2200 0000 0000 - 7d00 0000 0000: -
+7d00 0000 0000 - 7e00 0000 0000: heap
+7e00 0000 0000 - 7e80 0000 0000: -
+7e80 0000 0000 - 8000 0000 0000: modules and main thread stack
+*/
+struct Mapping47 {
+  static const uptr kMetaShadowBeg = 0x100000000000ull;
+  static const uptr kMetaShadowEnd = 0x200000000000ull;
+  static const uptr kTraceMemBeg   = 0x200000000000ull;
+  static const uptr kTraceMemEnd   = 0x220000000000ull;
+  static const uptr kShadowBeg     = 0x010000000000ull;
+  static const uptr kShadowEnd     = 0x100000000000ull;
+  static const uptr kHeapMemBeg    = 0x7d0000000000ull;
+  static const uptr kHeapMemEnd    = 0x7e0000000000ull;
+  static const uptr kLoAppMemBeg   = 0x000000001000ull;
+  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 kVdsoBeg       = 0x7800000000000000ull;
+};
+
 // Indicates the runtime will define the memory regions at runtime.
 #define TSAN_RUNTIME_VMA 1
 #endif
@@ -429,11 +475,13 @@ uptr MappingArchImpl(void) {
   DCHECK(0);
   return 0;
 #elif defined(__powerpc64__)
-  if (vmaSize == 44)
-    return MappingImpl<Mapping44, Type>();
-  else
-    return MappingImpl<Mapping46, Type>();
+  switch (vmaSize) {
+    case 44: return MappingImpl<Mapping44, Type>();
+    case 46: return MappingImpl<Mapping46, Type>();
+    case 47: return MappingImpl<Mapping47, Type>();
+  }
   DCHECK(0);
+  return 0;
 #else
   return MappingImpl<Mapping, Type>();
 #endif
@@ -582,11 +630,13 @@ bool IsAppMem(uptr mem) {
   DCHECK(0);
   return false;
 #elif defined(__powerpc64__)
-  if (vmaSize == 44)
-    return IsAppMemImpl<Mapping44>(mem);
-  else
-    return IsAppMemImpl<Mapping46>(mem);
+  switch (vmaSize) {
+    case 44: return IsAppMemImpl<Mapping44>(mem);
+    case 46: return IsAppMemImpl<Mapping46>(mem);
+    case 47: return IsAppMemImpl<Mapping47>(mem);
+  }
   DCHECK(0);
+  return false;
 #else
   return IsAppMemImpl<Mapping>(mem);
 #endif
@@ -609,11 +659,13 @@ bool IsShadowMem(uptr mem) {
   DCHECK(0);
   return false;
 #elif defined(__powerpc64__)
-  if (vmaSize == 44)
-    return IsShadowMemImpl<Mapping44>(mem);
-  else
-    return IsShadowMemImpl<Mapping46>(mem);
+  switch (vmaSize) {
+    case 44: return IsShadowMemImpl<Mapping44>(mem);
+    case 46: return IsShadowMemImpl<Mapping46>(mem);
+    case 47: return IsShadowMemImpl<Mapping47>(mem);
+  }
   DCHECK(0);
+  return false;
 #else
   return IsShadowMemImpl<Mapping>(mem);
 #endif
@@ -636,11 +688,13 @@ bool IsMetaMem(uptr mem) {
   DCHECK(0);
   return false;
 #elif defined(__powerpc64__)
-  if (vmaSize == 44)
-    return IsMetaMemImpl<Mapping44>(mem);
-  else
-    return IsMetaMemImpl<Mapping46>(mem);
+  switch (vmaSize) {
+    case 44: return IsMetaMemImpl<Mapping44>(mem);
+    case 46: return IsMetaMemImpl<Mapping46>(mem);
+    case 47: return IsMetaMemImpl<Mapping47>(mem);
+  }
   DCHECK(0);
+  return false;
 #else
   return IsMetaMemImpl<Mapping>(mem);
 #endif
@@ -673,11 +727,13 @@ uptr MemToShadow(uptr x) {
   DCHECK(0);
   return 0;
 #elif defined(__powerpc64__)
-  if (vmaSize == 44)
-    return MemToShadowImpl<Mapping44>(x);
-  else
-    return MemToShadowImpl<Mapping46>(x);
+  switch (vmaSize) {
+    case 44: return MemToShadowImpl<Mapping44>(x);
+    case 46: return MemToShadowImpl<Mapping46>(x);
+    case 47: return MemToShadowImpl<Mapping47>(x);
+  }
   DCHECK(0);
+  return 0;
 #else
   return MemToShadowImpl<Mapping>(x);
 #endif
@@ -712,11 +768,13 @@ u32 *MemToMeta(uptr x) {
   DCHECK(0);
   return 0;
 #elif defined(__powerpc64__)
-  if (vmaSize == 44)
-    return MemToMetaImpl<Mapping44>(x);
-  else
-    return MemToMetaImpl<Mapping46>(x);
+  switch (vmaSize) {
+    case 44: return MemToMetaImpl<Mapping44>(x);
+    case 46: return MemToMetaImpl<Mapping46>(x);
+    case 47: return MemToMetaImpl<Mapping47>(x);
+  }
   DCHECK(0);
+  return 0;
 #else
   return MemToMetaImpl<Mapping>(x);
 #endif
@@ -764,11 +822,13 @@ uptr ShadowToMem(uptr s) {
   DCHECK(0);
   return 0;
 #elif defined(__powerpc64__)
-  if (vmaSize == 44)
-    return ShadowToMemImpl<Mapping44>(s);
-  else
-    return ShadowToMemImpl<Mapping46>(s);
+  switch (vmaSize) {
+    case 44: return ShadowToMemImpl<Mapping44>(s);
+    case 46: return ShadowToMemImpl<Mapping46>(s);
+    case 47: return ShadowToMemImpl<Mapping47>(s);
+  }
   DCHECK(0);
+  return 0;
 #else
   return ShadowToMemImpl<Mapping>(s);
 #endif
@@ -799,11 +859,13 @@ uptr GetThreadTrace(int tid) {
   DCHECK(0);
   return 0;
 #elif defined(__powerpc64__)
-  if (vmaSize == 44)
-    return GetThreadTraceImpl<Mapping44>(tid);
-  else
-    return GetThreadTraceImpl<Mapping46>(tid);
+  switch (vmaSize) {
+    case 44: return GetThreadTraceImpl<Mapping44>(tid);
+    case 46: return GetThreadTraceImpl<Mapping46>(tid);
+    case 47: return GetThreadTraceImpl<Mapping47>(tid);
+  }
   DCHECK(0);
+  return 0;
 #else
   return GetThreadTraceImpl<Mapping>(tid);
 #endif
@@ -829,11 +891,13 @@ uptr GetThreadTraceHeader(int tid) {
   DCHECK(0);
   return 0;
 #elif defined(__powerpc64__)
-  if (vmaSize == 44)
-    return GetThreadTraceHeaderImpl<Mapping44>(tid);
-  else
-    return GetThreadTraceHeaderImpl<Mapping46>(tid);
+  switch (vmaSize) {
+    case 44: return GetThreadTraceHeaderImpl<Mapping44>(tid);
+    case 46: return GetThreadTraceHeaderImpl<Mapping46>(tid);
+    case 47: return GetThreadTraceHeaderImpl<Mapping47>(tid);
+  }
   DCHECK(0);
+  return 0;
 #else
   return GetThreadTraceHeaderImpl<Mapping>(tid);
 #endif
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cc
index ead1e5704989..e14d5f575a2e 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cc
@@ -14,11 +14,12 @@
 
 
 #include "sanitizer_common/sanitizer_platform.h"
-#if SANITIZER_LINUX || SANITIZER_FREEBSD
+#if SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD
 
 #include "sanitizer_common/sanitizer_common.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_posix.h"
 #include "sanitizer_common/sanitizer_procmaps.h"
@@ -216,9 +217,9 @@ void InitializePlatformEarly() {
     Die();
   }
 #elif defined(__powerpc64__)
-  if (vmaSize != 44 && vmaSize != 46) {
+  if (vmaSize != 44 && vmaSize != 46 && vmaSize != 47) {
     Printf("FATAL: ThreadSanitizer: unsupported VMA range\n");
-    Printf("FATAL: Found %d - Supported 44 and 46\n", vmaSize);
+    Printf("FATAL: Found %d - Supported 44, 46, and 47\n", vmaSize);
     Die();
   }
 #endif
@@ -401,4 +402,4 @@ void cur_thread_finalize() {
 
 }  // namespace __tsan
 
-#endif  // SANITIZER_LINUX || SANITIZER_FREEBSD
+#endif  // SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_platform_mac.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_platform_mac.cc
index 73a656ffca5e..f8d7324b763c 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_platform_mac.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_platform_mac.cc
@@ -167,8 +167,8 @@ void WriteMemoryProfile(char *buf, uptr buf_size, uptr nthread, uptr nlive) {
 #else  // !SANITIZER_GO
     "app      (0x%016zx-0x%016zx): resident %zd kB, dirty %zd kB\n"
 #endif
-    "stacks: %ld unique IDs, %ld kB allocated\n"
-    "threads: %ld total, %ld live\n"
+    "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,
@@ -231,7 +231,7 @@ static void my_pthread_introspection_hook(unsigned int event, pthread_t thread,
 
 void InitializePlatformEarly() {
 #if defined(__aarch64__)
-  uptr max_vm = GetMaxVirtualAddress() + 1;
+  uptr max_vm = GetMaxUserVirtualAddress() + 1;
   if (max_vm != Mapping::kHiAppMemEnd) {
     Printf("ThreadSanitizer: unsupported vm address limit %p, expected %p.\n",
            max_vm, Mapping::kHiAppMemEnd);
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_report.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_report.cc
index 32cc3325fe40..af47076968d3 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_report.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_report.cc
@@ -13,6 +13,7 @@
 #include "tsan_report.h"
 #include "tsan_platform.h"
 #include "tsan_rtl.h"
+#include "sanitizer_common/sanitizer_file.h"
 #include "sanitizer_common/sanitizer_placement_new.h"
 #include "sanitizer_common/sanitizer_report_decorator.h"
 #include "sanitizer_common/sanitizer_stacktrace_printer.h"
@@ -38,34 +39,27 @@ ReportLocation *ReportLocation::New(ReportLocationType type) {
 class Decorator: public __sanitizer::SanitizerCommonDecorator {
  public:
   Decorator() : SanitizerCommonDecorator() { }
-  const char *Warning()    { return Red(); }
-  const char *EndWarning() { return Default(); }
   const char *Access()     { return Blue(); }
-  const char *EndAccess()  { return Default(); }
   const char *ThreadDescription()    { return Cyan(); }
-  const char *EndThreadDescription() { return Default(); }
   const char *Location()   { return Green(); }
-  const char *EndLocation() { return Default(); }
   const char *Sleep()   { return Yellow(); }
-  const char *EndSleep() { return Default(); }
   const char *Mutex()   { return Magenta(); }
-  const char *EndMutex() { return Default(); }
 };
 
 ReportDesc::ReportDesc()
     : tag(kExternalTagNone)
-    , stacks(MBlockReportStack)
-    , mops(MBlockReportMop)
-    , locs(MBlockReportLoc)
-    , mutexes(MBlockReportMutex)
-    , threads(MBlockReportThread)
-    , unique_tids(MBlockReportThread)
+    , stacks()
+    , mops()
+    , locs()
+    , mutexes()
+    , threads()
+    , unique_tids()
     , sleep()
     , count() {
 }
 
 ReportMop::ReportMop()
-    : mset(MBlockReportMutex) {
+    : mset() {
 }
 
 ReportDesc::~ReportDesc() {
@@ -180,7 +174,7 @@ static void PrintMop(const ReportMop *mop, bool first) {
   }
   PrintMutexSet(mop->mset);
   Printf(":\n");
-  Printf("%s", d.EndAccess());
+  Printf("%s", d.Default());
   PrintStack(mop->stack);
 }
 
@@ -221,20 +215,20 @@ static void PrintLocation(const ReportLocation *loc) {
         loc->fd, thread_name(thrbuf, loc->tid));
     print_stack = true;
   }
-  Printf("%s", d.EndLocation());
+  Printf("%s", d.Default());
   if (print_stack)
     PrintStack(loc->stack);
 }
 
 static void PrintMutexShort(const ReportMutex *rm, const char *after) {
   Decorator d;
-  Printf("%sM%zd%s%s", d.Mutex(), rm->id, d.EndMutex(), after);
+  Printf("%sM%zd%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.EndMutex(), after);
+  Printf("%sM%zd (%p)%s%s", d.Mutex(), rm->id, rm->addr, d.Default(), after);
 }
 
 static void PrintMutex(const ReportMutex *rm) {
@@ -242,11 +236,11 @@ static void PrintMutex(const ReportMutex *rm) {
   if (rm->destroyed) {
     Printf("%s", d.Mutex());
     Printf("  Mutex M%llu is already destroyed.\n\n", rm->id);
-    Printf("%s", d.EndMutex());
+    Printf("%s", d.Default());
   } else {
     Printf("%s", d.Mutex());
     Printf("  Mutex M%llu (%p) created at:\n", rm->id, rm->addr);
-    Printf("%s", d.EndMutex());
+    Printf("%s", d.Default());
     PrintStack(rm->stack);
   }
 }
@@ -264,7 +258,7 @@ static void PrintThread(const ReportThread *rt) {
   if (rt->workerthread) {
     Printf(" (tid=%zu, %s) is a GCD worker thread\n", rt->os_id, thread_status);
     Printf("\n");
-    Printf("%s", d.EndThreadDescription());
+    Printf("%s", d.Default());
     return;
   }
   Printf(" (tid=%zu, %s) created by %s", rt->os_id, thread_status,
@@ -272,7 +266,7 @@ static void PrintThread(const ReportThread *rt) {
   if (rt->stack)
     Printf(" at:");
   Printf("\n");
-  Printf("%s", d.EndThreadDescription());
+  Printf("%s", d.Default());
   PrintStack(rt->stack);
 }
 
@@ -280,7 +274,7 @@ static void PrintSleep(const ReportStack *s) {
   Decorator d;
   Printf("%s", d.Sleep());
   Printf("  As if synchronized via sleep:\n");
-  Printf("%s", d.EndSleep());
+  Printf("%s", d.Default());
   PrintStack(s);
 }
 
@@ -324,7 +318,7 @@ void PrintReport(const ReportDesc *rep) {
   Printf("%s", d.Warning());
   Printf("WARNING: ThreadSanitizer: %s (pid=%d)\n", rep_typ_str,
          (int)internal_getpid());
-  Printf("%s", d.EndWarning());
+  Printf("%s", d.Default());
 
   if (rep->typ == ReportTypeDeadlock) {
     char thrbuf[kThreadBufSize];
@@ -342,7 +336,7 @@ void PrintReport(const ReportDesc *rep) {
       PrintMutexShort(rep->mutexes[i], " in ");
       Printf("%s", d.ThreadDescription());
       Printf("%s:\n", thread_name(thrbuf, rep->unique_tids[i]));
-      Printf("%s", d.EndThreadDescription());
+      Printf("%s", d.Default());
       if (flags()->second_deadlock_stack) {
         PrintStack(rep->stacks[2*i]);
         Printf("  Mutex ");
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_report.h b/contrib/compiler-rt/lib/tsan/rtl/tsan_report.h
index bc1582f90fe4..cdc999c6af1d 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_report.h
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_report.h
@@ -14,8 +14,8 @@
 #define TSAN_REPORT_H
 
 #include "sanitizer_common/sanitizer_symbolizer.h"
+#include "sanitizer_common/sanitizer_vector.h"
 #include "tsan_defs.h"
-#include "tsan_vector.h"
 
 namespace __tsan {
 
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl.cc
index a01525302b02..17b820977c26 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl.cc
@@ -14,6 +14,7 @@
 
 #include "sanitizer_common/sanitizer_atomic.h"
 #include "sanitizer_common/sanitizer_common.h"
+#include "sanitizer_common/sanitizer_file.h"
 #include "sanitizer_common/sanitizer_libc.h"
 #include "sanitizer_common/sanitizer_stackdepot.h"
 #include "sanitizer_common/sanitizer_placement_new.h"
@@ -101,8 +102,8 @@ Context::Context()
   , thread_registry(new(thread_registry_placeholder) ThreadRegistry(
       CreateThreadContext, kMaxTid, kThreadQuarantineSize, kMaxTidReuse))
   , racy_mtx(MutexTypeRacy, StatMtxRacy)
-  , racy_stacks(MBlockRacyStacks)
-  , racy_addresses(MBlockRacyAddresses)
+  , racy_stacks()
+  , racy_addresses()
   , fired_suppressions_mtx(MutexTypeFired, StatMtxFired)
   , fired_suppressions(8)
   , clock_alloc("clock allocator") {
@@ -120,7 +121,7 @@ ThreadState::ThreadState(Context *ctx, int tid, int unique_id, u64 epoch,
   // , ignore_interceptors()
   , clock(tid, reuse_count)
 #if !SANITIZER_GO
-  , jmp_bufs(MBlockJmpBuf)
+  , jmp_bufs()
 #endif
   , tid(tid)
   , unique_id(unique_id)
@@ -213,7 +214,7 @@ static void BackgroundThread(void *arg) {
                              memory_order_relaxed);
       if (last != 0 && last + flags()->flush_symbolizer_ms * kMs2Ns < now) {
         Lock l(&ctx->report_mtx);
-        SpinMutexLock l2(&CommonSanitizerReportMutex);
+        ScopedErrorReportLock l2;
         SymbolizeFlush();
         atomic_store(&ctx->last_symbolize_time_ns, 0, memory_order_relaxed);
       }
@@ -323,6 +324,21 @@ static void CheckShadowMapping() {
   }
 }
 
+#if !SANITIZER_GO
+static void OnStackUnwind(const SignalContext &sig, const void *,
+                          BufferedStackTrace *stack) {
+  uptr top = 0;
+  uptr bottom = 0;
+  bool fast = common_flags()->fast_unwind_on_fatal;
+  if (fast) GetThreadStackTopAndBottom(false, &top, &bottom);
+  stack->Unwind(kStackTraceMax, sig.pc, sig.bp, sig.context, top, bottom, fast);
+}
+
+static void TsanOnDeadlySignal(int signo, void *siginfo, void *context) {
+  HandleDeadlySignal(siginfo, context, GetTid(), &OnStackUnwind, nullptr);
+}
+#endif
+
 void Initialize(ThreadState *thr) {
   // Thread safe because done before all threads exist.
   static bool is_initialized = false;
@@ -360,6 +376,7 @@ void Initialize(ThreadState *thr) {
 #if !SANITIZER_GO
   InitializeShadowMemory();
   InitializeAllocatorLate();
+  InstallDeadlySignalHandlers(TsanOnDeadlySignal);
 #endif
   // Setup correct file descriptor for error reports.
   __sanitizer_set_report_path(common_flags()->log_path);
@@ -367,13 +384,6 @@ void Initialize(ThreadState *thr) {
 #if !SANITIZER_GO
   InitializeLibIgnore();
   Symbolizer::GetOrInit()->AddHooks(EnterSymbolizer, ExitSymbolizer);
-  // On MIPS, TSan initialization is run before
-  // __pthread_initialize_minimal_internal() is finished, so we can not spawn
-  // new threads.
-#ifndef __mips__
-  StartBackgroundThread();
-  SetSandboxingCallback(StopBackgroundThread);
-#endif
 #endif
 
   VPrintf(1, "***** Running under ThreadSanitizer v2 (pid %d) *****\n",
@@ -402,6 +412,21 @@ void Initialize(ThreadState *thr) {
   OnInitialize();
 }
 
+void MaybeSpawnBackgroundThread() {
+  // On MIPS, TSan initialization is run before
+  // __pthread_initialize_minimal_internal() is finished, so we can not spawn
+  // new threads.
+#if !SANITIZER_GO && !defined(__mips__)
+  static atomic_uint32_t bg_thread = {};
+  if (atomic_load(&bg_thread, memory_order_relaxed) == 0 &&
+      atomic_exchange(&bg_thread, 1, memory_order_relaxed) == 0) {
+    StartBackgroundThread();
+    SetSandboxingCallback(StopBackgroundThread);
+  }
+#endif
+}
+
+
 int Finalize(ThreadState *thr) {
   bool failed = false;
 
@@ -412,8 +437,7 @@ int Finalize(ThreadState *thr) {
 
   // Wait for pending reports.
   ctx->report_mtx.Lock();
-  CommonSanitizerReportMutex.Lock();
-  CommonSanitizerReportMutex.Unlock();
+  { ScopedErrorReportLock l; }
   ctx->report_mtx.Unlock();
 
 #if !SANITIZER_GO
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl.h b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl.h
index 2cf2e168454d..7d44ccae9893 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl.h
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl.h
@@ -34,12 +34,13 @@
 #include "sanitizer_common/sanitizer_libignore.h"
 #include "sanitizer_common/sanitizer_suppressions.h"
 #include "sanitizer_common/sanitizer_thread_registry.h"
+#include "sanitizer_common/sanitizer_vector.h"
 #include "tsan_clock.h"
 #include "tsan_defs.h"
 #include "tsan_flags.h"
+#include "tsan_mman.h"
 #include "tsan_sync.h"
 #include "tsan_trace.h"
-#include "tsan_vector.h"
 #include "tsan_report.h"
 #include "tsan_platform.h"
 #include "tsan_mutexset.h"
@@ -574,11 +575,8 @@ const char *GetObjectTypeFromTag(uptr tag);
 const char *GetReportHeaderFromTag(uptr tag);
 uptr TagFromShadowStackFrame(uptr pc);
 
-class ScopedReport {
+class ScopedReportBase {
  public:
-  explicit ScopedReport(ReportType typ, uptr tag = kExternalTagNone);
-  ~ScopedReport();
-
   void AddMemoryAccess(uptr addr, uptr external_tag, Shadow s, StackTrace stack,
                        const MutexSet *mset);
   void AddStack(StackTrace stack, bool suppressable = false);
@@ -593,6 +591,10 @@ class ScopedReport {
 
   const ReportDesc *GetReport() const;
 
+ protected:
+  ScopedReportBase(ReportType typ, uptr tag);
+  ~ScopedReportBase();
+
  private:
   ReportDesc *rep_;
   // Symbolizer makes lots of intercepted calls. If we try to process them,
@@ -601,8 +603,17 @@ class ScopedReport {
 
   void AddDeadMutex(u64 id);
 
-  ScopedReport(const ScopedReport&);
-  void operator = (const ScopedReport&);
+  ScopedReportBase(const ScopedReportBase &) = delete;
+  void operator=(const ScopedReportBase &) = delete;
+};
+
+class ScopedReport : public ScopedReportBase {
+ public:
+  explicit ScopedReport(ReportType typ, uptr tag = kExternalTagNone);
+  ~ScopedReport();
+
+ private:
+  ScopedErrorReportLock lock_;
 };
 
 ThreadContext *IsThreadStackOrTls(uptr addr, bool *is_stack);
@@ -690,6 +701,7 @@ void PrintCurrentStack(ThreadState *thr, uptr pc);
 void PrintCurrentStackSlow(uptr pc);  // uses libunwind
 
 void Initialize(ThreadState *thr);
+void MaybeSpawnBackgroundThread();
 int Finalize(ThreadState *thr);
 
 void OnUserAlloc(ThreadState *thr, uptr pc, uptr p, uptr sz, bool write);
@@ -825,7 +837,7 @@ void ALWAYS_INLINE TraceAddEvent(ThreadState *thr, FastState fs,
     return;
   DCHECK_GE((int)typ, 0);
   DCHECK_LE((int)typ, 7);
-  DCHECK_EQ(GetLsb(addr, 61), addr);
+  DCHECK_EQ(GetLsb(addr, kEventPCBits), addr);
   StatInc(thr, StatEvents);
   u64 pos = fs.GetTracePos();
   if (UNLIKELY((pos % kTracePartSize) == 0)) {
@@ -837,7 +849,7 @@ void ALWAYS_INLINE TraceAddEvent(ThreadState *thr, FastState fs,
   }
   Event *trace = (Event*)GetThreadTrace(fs.tid());
   Event *evp = &trace[pos];
-  Event ev = (u64)addr | ((u64)typ << 61);
+  Event ev = (u64)addr | ((u64)typ << kEventPCBits);
   *evp = ev;
 }
 
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_aarch64.S b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_aarch64.S
index f8c6453a5da4..f1e4dbb50fbb 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_aarch64.S
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_aarch64.S
@@ -6,7 +6,7 @@
 #if !defined(__APPLE__)
 .section .bss
 .type	__tsan_pointer_chk_guard, %object
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(__tsan_pointer_chk_guard))
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(__tsan_pointer_chk_guard))
 __tsan_pointer_chk_guard:
 .zero	8
 #endif
@@ -51,7 +51,7 @@ _sigsetjmp$non_lazy_ptr:
 // original ones.
 ASM_HIDDEN(_Z18InitializeGuardPtrv)
 .global _Z18InitializeGuardPtrv
-ASM_TYPE_FUNCTION(ASM_TSAN_SYMBOL_INTERCEPTOR(_Z18InitializeGuardPtrv))
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(_Z18InitializeGuardPtrv))
 _Z18InitializeGuardPtrv:
   CFI_STARTPROC
   // Allocates a jmp_buf for the setjmp call.
@@ -88,14 +88,14 @@ _Z18InitializeGuardPtrv:
   CFI_DEF_CFA (31, 0)
   ret
   CFI_ENDPROC
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(_Z18InitializeGuardPtrv))
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(_Z18InitializeGuardPtrv))
 #endif
 
 ASM_HIDDEN(__tsan_setjmp)
 .comm _ZN14__interception11real_setjmpE,8,8
-.globl ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp)
-ASM_TYPE_FUNCTION(ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp))
-ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp):
+.globl ASM_SYMBOL_INTERCEPTOR(setjmp)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(setjmp))
+ASM_SYMBOL_INTERCEPTOR(setjmp):
   CFI_STARTPROC
 
   // save env parameters for function call
@@ -125,7 +125,7 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp):
 #endif
 
   // call tsan interceptor
-  bl      ASM_TSAN_SYMBOL(__tsan_setjmp)
+  bl      ASM_SYMBOL(__tsan_setjmp)
 
   // restore env parameter
   mov     x0, x19
@@ -148,12 +148,12 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp):
   br      x1
 
   CFI_ENDPROC
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp))
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(setjmp))
 
 .comm _ZN14__interception12real__setjmpE,8,8
-.globl ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp)
-ASM_TYPE_FUNCTION(ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp))
-ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp):
+.globl ASM_SYMBOL_INTERCEPTOR(_setjmp)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(_setjmp))
+ASM_SYMBOL_INTERCEPTOR(_setjmp):
   CFI_STARTPROC
 
   // save env parameters for function call
@@ -183,7 +183,7 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp):
 #endif
 
   // call tsan interceptor
-  bl      ASM_TSAN_SYMBOL(__tsan_setjmp)
+  bl      ASM_SYMBOL(__tsan_setjmp)
 
   // Restore jmp_buf parameter
   mov     x0, x19
@@ -206,12 +206,12 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp):
   br      x1
 
   CFI_ENDPROC
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp))
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(_setjmp))
 
 .comm _ZN14__interception14real_sigsetjmpE,8,8
-.globl ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp)
-ASM_TYPE_FUNCTION(ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp))
-ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp):
+.globl ASM_SYMBOL_INTERCEPTOR(sigsetjmp)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(sigsetjmp))
+ASM_SYMBOL_INTERCEPTOR(sigsetjmp):
   CFI_STARTPROC
 
   // save env parameters for function call
@@ -243,7 +243,7 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp):
 #endif
 
   // call tsan interceptor
-  bl      ASM_TSAN_SYMBOL(__tsan_setjmp)
+  bl      ASM_SYMBOL(__tsan_setjmp)
 
   // restore env parameter
   mov     w1, w20
@@ -268,13 +268,13 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp):
 #endif
   br      x2
   CFI_ENDPROC
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp))
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(sigsetjmp))
 
 #if !defined(__APPLE__)
 .comm _ZN14__interception16real___sigsetjmpE,8,8
-.globl ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp)
-ASM_TYPE_FUNCTION(ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp))
-ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp):
+.globl ASM_SYMBOL_INTERCEPTOR(__sigsetjmp)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(__sigsetjmp))
+ASM_SYMBOL_INTERCEPTOR(__sigsetjmp):
   CFI_STARTPROC
 
   // save env parameters for function call
@@ -303,7 +303,7 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp):
 #endif
 
   // call tsan interceptor
-  bl      ASM_TSAN_SYMBOL(__tsan_setjmp)
+  bl      ASM_SYMBOL(__tsan_setjmp)
 
   mov     w1, w20
   mov     x0, x19
@@ -321,12 +321,12 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp):
   ldr     x2, [x2, #:got_lo12:_ZN14__interception16real___sigsetjmpE]
   ldr     x2, [x2]
 #else
-  adrp    x2, ASM_TSAN_SYMBOL(__sigsetjmp)@page
-  add     x2, x2, ASM_TSAN_SYMBOL(__sigsetjmp)@pageoff
+  adrp    x2, ASM_SYMBOL(__sigsetjmp)@page
+  add     x2, x2, ASM_SYMBOL(__sigsetjmp)@pageoff
 #endif
   br      x2
   CFI_ENDPROC
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp))
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(__sigsetjmp))
 #endif
 
 #if defined(__FreeBSD__) || defined(__linux__)
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_amd64.S b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_amd64.S
index 98947fd2a1ba..8af61bf0e892 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_amd64.S
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_amd64.S
@@ -10,8 +10,8 @@
 #endif
 
 ASM_HIDDEN(__tsan_trace_switch)
-.globl ASM_TSAN_SYMBOL(__tsan_trace_switch_thunk)
-ASM_TSAN_SYMBOL(__tsan_trace_switch_thunk):
+.globl ASM_SYMBOL(__tsan_trace_switch_thunk)
+ASM_SYMBOL(__tsan_trace_switch_thunk):
   CFI_STARTPROC
   # Save scratch registers.
   push %rax
@@ -50,7 +50,7 @@ ASM_TSAN_SYMBOL(__tsan_trace_switch_thunk):
   shr $4, %rsp  # clear 4 lsb, align to 16
   shl $4, %rsp
 
-  call ASM_TSAN_SYMBOL(__tsan_trace_switch)
+  call ASM_SYMBOL(__tsan_trace_switch)
 
   # Unalign stack frame back.
   mov %rbx, %rsp  # restore the original rsp
@@ -90,8 +90,8 @@ ASM_TSAN_SYMBOL(__tsan_trace_switch_thunk):
   CFI_ENDPROC
 
 ASM_HIDDEN(__tsan_report_race)
-.globl ASM_TSAN_SYMBOL(__tsan_report_race_thunk)
-ASM_TSAN_SYMBOL(__tsan_report_race_thunk):
+.globl ASM_SYMBOL(__tsan_report_race_thunk)
+ASM_SYMBOL(__tsan_report_race_thunk):
   CFI_STARTPROC
   # Save scratch registers.
   push %rax
@@ -130,7 +130,7 @@ ASM_TSAN_SYMBOL(__tsan_report_race_thunk):
   shr $4, %rsp  # clear 4 lsb, align to 16
   shl $4, %rsp
 
-  call ASM_TSAN_SYMBOL(__tsan_report_race)
+  call ASM_SYMBOL(__tsan_report_race)
 
   # Unalign stack frame back.
   mov %rbx, %rsp  # restore the original rsp
@@ -170,19 +170,27 @@ ASM_TSAN_SYMBOL(__tsan_report_race_thunk):
   CFI_ENDPROC
 
 ASM_HIDDEN(__tsan_setjmp)
-#if !defined(__APPLE__)
+#if defined(__NetBSD__)
+.comm _ZN14__interception15real___setjmp14E,8,8
+#elif !defined(__APPLE__)
 .comm _ZN14__interception11real_setjmpE,8,8
 #endif
-.globl ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp)
-ASM_TYPE_FUNCTION(ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp))
-ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp):
+#if defined(__NetBSD__)
+.globl ASM_SYMBOL_INTERCEPTOR(__setjmp14)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(__setjmp14))
+ASM_SYMBOL_INTERCEPTOR(__setjmp14):
+#else
+.globl ASM_SYMBOL_INTERCEPTOR(setjmp)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(setjmp))
+ASM_SYMBOL_INTERCEPTOR(setjmp):
+#endif
   CFI_STARTPROC
   // save env parameter
   push %rdi
   CFI_ADJUST_CFA_OFFSET(8)
   CFI_REL_OFFSET(%rdi, 0)
   // obtain %rsp
-#if defined(__FreeBSD__)
+#if defined(__FreeBSD__) || defined(__NetBSD__)
   lea 8(%rsp), %rdi
   mov %rdi, %rsi
 #elif defined(__APPLE__)
@@ -197,33 +205,40 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp):
 # error "Unknown platform"
 #endif
   // call tsan interceptor
-  call ASM_TSAN_SYMBOL(__tsan_setjmp)
+  call ASM_SYMBOL(__tsan_setjmp)
   // restore env parameter
   pop %rdi
   CFI_ADJUST_CFA_OFFSET(-8)
   CFI_RESTORE(%rdi)
   // tail jump to libc setjmp
   movl $0, %eax
-#if !defined(__APPLE__)
+#if defined(__NetBSD__)
+  movq _ZN14__interception15real___setjmp14E@GOTPCREL(%rip), %rdx
+  jmp *(%rdx)
+#elif !defined(__APPLE__)
   movq _ZN14__interception11real_setjmpE@GOTPCREL(%rip), %rdx
   jmp *(%rdx)
 #else
-  jmp ASM_TSAN_SYMBOL(setjmp)
+  jmp ASM_SYMBOL(setjmp)
 #endif
   CFI_ENDPROC
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(setjmp))
+#if defined(__NetBSD__)
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(__setjmp14))
+#else
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(setjmp))
+#endif
 
 .comm _ZN14__interception12real__setjmpE,8,8
-.globl ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp)
-ASM_TYPE_FUNCTION(ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp))
-ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp):
+.globl ASM_SYMBOL_INTERCEPTOR(_setjmp)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(_setjmp))
+ASM_SYMBOL_INTERCEPTOR(_setjmp):
   CFI_STARTPROC
   // save env parameter
   push %rdi
   CFI_ADJUST_CFA_OFFSET(8)
   CFI_REL_OFFSET(%rdi, 0)
   // obtain %rsp
-#if defined(__FreeBSD__)
+#if defined(__FreeBSD__) || defined(__NetBSD__)
   lea 8(%rsp), %rdi
   mov %rdi, %rsi
 #elif defined(__APPLE__)
@@ -238,7 +253,7 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp):
 # error "Unknown platform"
 #endif
   // call tsan interceptor
-  call ASM_TSAN_SYMBOL(__tsan_setjmp)
+  call ASM_SYMBOL(__tsan_setjmp)
   // restore env parameter
   pop %rdi
   CFI_ADJUST_CFA_OFFSET(-8)
@@ -249,15 +264,22 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp):
   movq _ZN14__interception12real__setjmpE@GOTPCREL(%rip), %rdx
   jmp *(%rdx)
 #else
-  jmp ASM_TSAN_SYMBOL(_setjmp)
+  jmp ASM_SYMBOL(_setjmp)
 #endif
   CFI_ENDPROC
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(_setjmp))
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(_setjmp))
 
+#if defined(__NetBSD__)
+.comm _ZN14__interception18real___sigsetjmp14E,8,8
+.globl ASM_SYMBOL_INTERCEPTOR(__sigsetjmp14)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(__sigsetjmp14))
+ASM_SYMBOL_INTERCEPTOR(__sigsetjmp14):
+#else
 .comm _ZN14__interception14real_sigsetjmpE,8,8
-.globl ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp)
-ASM_TYPE_FUNCTION(ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp))
-ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp):
+.globl ASM_SYMBOL_INTERCEPTOR(sigsetjmp)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(sigsetjmp))
+ASM_SYMBOL_INTERCEPTOR(sigsetjmp):
+#endif
   CFI_STARTPROC
   // save env parameter
   push %rdi
@@ -271,7 +293,7 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp):
   sub $8, %rsp
   CFI_ADJUST_CFA_OFFSET(8)
   // obtain %rsp
-#if defined(__FreeBSD__)
+#if defined(__FreeBSD__) || defined(__NetBSD__)
   lea 24(%rsp), %rdi
   mov %rdi, %rsi
 #elif defined(__APPLE__)
@@ -286,7 +308,7 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp):
 # error "Unknown platform"
 #endif
   // call tsan interceptor
-  call ASM_TSAN_SYMBOL(__tsan_setjmp)
+  call ASM_SYMBOL(__tsan_setjmp)
   // unalign stack frame
   add $8, %rsp
   CFI_ADJUST_CFA_OFFSET(-8)
@@ -300,20 +322,27 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp):
   CFI_RESTORE(%rdi)
   // tail jump to libc sigsetjmp
   movl $0, %eax
-#if !defined(__APPLE__)
+#if defined(__NetBSD__)
+  movq _ZN14__interception18real___sigsetjmp14E@GOTPCREL(%rip), %rdx
+  jmp *(%rdx)
+#elif !defined(__APPLE__)
   movq _ZN14__interception14real_sigsetjmpE@GOTPCREL(%rip), %rdx
   jmp *(%rdx)
 #else
-  jmp ASM_TSAN_SYMBOL(sigsetjmp)
+  jmp ASM_SYMBOL(sigsetjmp)
 #endif
   CFI_ENDPROC
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(sigsetjmp))
+#if defined(__NetBSD__)
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(__sigsetjmp14))
+#else
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(sigsetjmp))
+#endif
 
-#if !defined(__APPLE__)
+#if !defined(__APPLE__) && !defined(__NetBSD__)
 .comm _ZN14__interception16real___sigsetjmpE,8,8
-.globl ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp)
-ASM_TYPE_FUNCTION(ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp))
-ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp):
+.globl ASM_SYMBOL_INTERCEPTOR(__sigsetjmp)
+ASM_TYPE_FUNCTION(ASM_SYMBOL_INTERCEPTOR(__sigsetjmp))
+ASM_SYMBOL_INTERCEPTOR(__sigsetjmp):
   CFI_STARTPROC
   // save env parameter
   push %rdi
@@ -337,7 +366,7 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp):
   rol $0x11, %rsi
 #endif
   // call tsan interceptor
-  call ASM_TSAN_SYMBOL(__tsan_setjmp)
+  call ASM_SYMBOL(__tsan_setjmp)
   // unalign stack frame
   add $8, %rsp
   CFI_ADJUST_CFA_OFFSET(-8)
@@ -354,11 +383,12 @@ ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp):
   movq _ZN14__interception16real___sigsetjmpE@GOTPCREL(%rip), %rdx
   jmp *(%rdx)
   CFI_ENDPROC
-ASM_SIZE(ASM_TSAN_SYMBOL_INTERCEPTOR(__sigsetjmp))
-#endif  // !defined(__APPLE__)
+ASM_SIZE(ASM_SYMBOL_INTERCEPTOR(__sigsetjmp))
+#endif  // !defined(__APPLE__) && !defined(__NetBSD__)
 
 #if defined(__FreeBSD__) || defined(__linux__)
 /* We do not need executable stack.  */
+/* This note is not needed on NetBSD. */
 .section        .note.GNU-stack,"",@progbits
 #endif
 
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cc
index 2f85811620f1..152b965ad535 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cc
@@ -84,7 +84,9 @@ void MutexDestroy(ThreadState *thr, uptr pc, uptr addr, u32 flagz) {
   SyncVar *s = ctx->metamap.GetIfExistsAndLock(addr, true);
   if (s == 0)
     return;
-  if ((flagz & MutexFlagLinkerInit) || s->IsFlagSet(MutexFlagLinkerInit)) {
+  if ((flagz & MutexFlagLinkerInit)
+      || s->IsFlagSet(MutexFlagLinkerInit)
+      || ((flagz & MutexFlagNotStatic) && !s->IsFlagSet(MutexFlagNotStatic))) {
     // Destroy is no-op for linker-initialized mutexes.
     s->mtx.Unlock();
     return;
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cc
index 85a982941ed7..cc582ab50190 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cc
@@ -143,30 +143,28 @@ static ReportStack *SymbolizeStack(StackTrace trace) {
   return stack;
 }
 
-ScopedReport::ScopedReport(ReportType typ, uptr tag) {
+ScopedReportBase::ScopedReportBase(ReportType typ, uptr tag) {
   ctx->thread_registry->CheckLocked();
   void *mem = internal_alloc(MBlockReport, sizeof(ReportDesc));
   rep_ = new(mem) ReportDesc;
   rep_->typ = typ;
   rep_->tag = tag;
   ctx->report_mtx.Lock();
-  CommonSanitizerReportMutex.Lock();
 }
 
-ScopedReport::~ScopedReport() {
-  CommonSanitizerReportMutex.Unlock();
+ScopedReportBase::~ScopedReportBase() {
   ctx->report_mtx.Unlock();
   DestroyAndFree(rep_);
 }
 
-void ScopedReport::AddStack(StackTrace stack, bool suppressable) {
+void ScopedReportBase::AddStack(StackTrace stack, bool suppressable) {
   ReportStack **rs = rep_->stacks.PushBack();
   *rs = SymbolizeStack(stack);
   (*rs)->suppressable = suppressable;
 }
 
-void ScopedReport::AddMemoryAccess(uptr addr, uptr external_tag, Shadow s,
-                                   StackTrace stack, const MutexSet *mset) {
+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;
   rep_->mops.PushBack(mop);
@@ -187,11 +185,11 @@ void ScopedReport::AddMemoryAccess(uptr addr, uptr external_tag, Shadow s,
   }
 }
 
-void ScopedReport::AddUniqueTid(int unique_tid) {
+void ScopedReportBase::AddUniqueTid(int unique_tid) {
   rep_->unique_tids.PushBack(unique_tid);
 }
 
-void ScopedReport::AddThread(const ThreadContext *tctx, bool suppressable) {
+void ScopedReportBase::AddThread(const ThreadContext *tctx, bool suppressable) {
   for (uptr i = 0; i < rep_->threads.Size(); i++) {
     if ((u32)rep_->threads[i]->id == tctx->tid)
       return;
@@ -255,14 +253,14 @@ ThreadContext *IsThreadStackOrTls(uptr addr, bool *is_stack) {
 }
 #endif
 
-void ScopedReport::AddThread(int unique_tid, bool suppressable) {
+void ScopedReportBase::AddThread(int unique_tid, bool suppressable) {
 #if !SANITIZER_GO
   if (const ThreadContext *tctx = FindThreadByUidLocked(unique_tid))
     AddThread(tctx, suppressable);
 #endif
 }
 
-void ScopedReport::AddMutex(const SyncVar *s) {
+void ScopedReportBase::AddMutex(const SyncVar *s) {
   for (uptr i = 0; i < rep_->mutexes.Size(); i++) {
     if (rep_->mutexes[i]->id == s->uid)
       return;
@@ -276,7 +274,7 @@ void ScopedReport::AddMutex(const SyncVar *s) {
   rm->stack = SymbolizeStackId(s->creation_stack_id);
 }
 
-u64 ScopedReport::AddMutex(u64 id) {
+u64 ScopedReportBase::AddMutex(u64 id) {
   u64 uid = 0;
   u64 mid = id;
   uptr addr = SyncVar::SplitId(id, &uid);
@@ -295,7 +293,7 @@ u64 ScopedReport::AddMutex(u64 id) {
   return mid;
 }
 
-void ScopedReport::AddDeadMutex(u64 id) {
+void ScopedReportBase::AddDeadMutex(u64 id) {
   for (uptr i = 0; i < rep_->mutexes.Size(); i++) {
     if (rep_->mutexes[i]->id == id)
       return;
@@ -309,7 +307,7 @@ void ScopedReport::AddDeadMutex(u64 id) {
   rm->stack = 0;
 }
 
-void ScopedReport::AddLocation(uptr addr, uptr size) {
+void ScopedReportBase::AddLocation(uptr addr, uptr size) {
   if (addr == 0)
     return;
 #if !SANITIZER_GO
@@ -364,18 +362,19 @@ void ScopedReport::AddLocation(uptr addr, uptr size) {
 }
 
 #if !SANITIZER_GO
-void ScopedReport::AddSleep(u32 stack_id) {
+void ScopedReportBase::AddSleep(u32 stack_id) {
   rep_->sleep = SymbolizeStackId(stack_id);
 }
 #endif
 
-void ScopedReport::SetCount(int count) {
-  rep_->count = count;
-}
+void ScopedReportBase::SetCount(int count) { rep_->count = count; }
 
-const ReportDesc *ScopedReport::GetReport() const {
-  return rep_;
-}
+const ReportDesc *ScopedReportBase::GetReport() const { return rep_; }
+
+ScopedReport::ScopedReport(ReportType typ, uptr tag)
+    : ScopedReportBase(typ, tag) {}
+
+ScopedReport::~ScopedReport() {}
 
 void RestoreStack(int tid, const u64 epoch, VarSizeStackTrace *stk,
                   MutexSet *mset, uptr *tag) {
@@ -394,7 +393,7 @@ void RestoreStack(int tid, const u64 epoch, VarSizeStackTrace *stk,
   const u64 ebegin = RoundDown(eend, kTracePartSize);
   DPrintf("#%d: RestoreStack epoch=%zu ebegin=%zu eend=%zu partidx=%d\n",
           tid, (uptr)epoch, (uptr)ebegin, (uptr)eend, partidx);
-  Vector<uptr> stack(MBlockReportStack);
+  Vector<uptr> stack;
   stack.Resize(hdr->stack0.size + 64);
   for (uptr i = 0; i < hdr->stack0.size; i++) {
     stack[i] = hdr->stack0.trace[i];
@@ -406,8 +405,8 @@ void RestoreStack(int tid, const u64 epoch, VarSizeStackTrace *stk,
   Event *events = (Event*)GetThreadTrace(tid);
   for (uptr i = ebegin; i <= eend; i++) {
     Event ev = events[i];
-    EventType typ = (EventType)(ev >> 61);
-    uptr pc = (uptr)(ev & ((1ull << 61) - 1));
+    EventType typ = (EventType)(ev >> kEventPCBits);
+    uptr pc = (uptr)(ev & ((1ull << kEventPCBits) - 1));
     DPrintf2("  %zu typ=%d pc=%zx\n", i, typ, pc);
     if (typ == EventTypeMop) {
       stack[pos] = pc;
@@ -634,13 +633,33 @@ void ReportRace(ThreadState *thr) {
   const uptr kMop = 2;
   VarSizeStackTrace traces[kMop];
   uptr tags[kMop] = {kExternalTagNone};
-  const uptr toppc = TraceTopPC(thr);
+  uptr toppc = TraceTopPC(thr);
+  if (toppc >> kEventPCBits) {
+    // This is a work-around for a known issue.
+    // The scenario where this happens is rather elaborate and requires
+    // an instrumented __sanitizer_report_error_summary callback and
+    // a __tsan_symbolize_external callback and a race during a range memory
+    // access larger than 8 bytes. MemoryAccessRange adds the current PC to
+    // the trace and starts processing memory accesses. A first memory access
+    // triggers a race, we report it and call the instrumented
+    // __sanitizer_report_error_summary, which adds more stuff to the trace
+    // since it is intrumented. Then a second memory access in MemoryAccessRange
+    // also triggers a race and we get here and call TraceTopPC to get the
+    // current PC, however now it contains some unrelated events from the
+    // callback. Most likely, TraceTopPC will now return a EventTypeFuncExit
+    // event. Later we subtract -1 from it (in GetPreviousInstructionPc)
+    // and the resulting PC has kExternalPCBit set, so we pass it to
+    // __tsan_symbolize_external. __tsan_symbolize_external is within its rights
+    // to crash since the PC is completely bogus.
+    // test/tsan/double_race.cc contains a test case for this.
+    toppc = 0;
+  }
   ObtainCurrentStack(thr, toppc, &traces[0], &tags[0]);
   if (IsFiredSuppression(ctx, typ, traces[0]))
     return;
 
   // MutexSet is too large to live on stack.
-  Vector<u64> mset_buffer(MBlockScopedBuf);
+  Vector<u64> mset_buffer;
   mset_buffer.Resize(sizeof(MutexSet) / sizeof(u64) + 1);
   MutexSet *mset2 = new(&mset_buffer[0]) MutexSet();
 
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cc b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cc
index 83fab082afe3..e4d65b9a909b 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cc
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cc
@@ -211,7 +211,7 @@ void ThreadFinalize(ThreadState *thr) {
   if (!flags()->report_thread_leaks)
     return;
   ThreadRegistryLock l(ctx->thread_registry);
-  Vector<ThreadLeak> leaks(MBlockScopedBuf);
+  Vector<ThreadLeak> leaks;
   ctx->thread_registry->RunCallbackForEachThreadLocked(
       MaybeReportThreadLeak, &leaks);
   for (uptr i = 0; i < leaks.Size(); i++) {
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_sync.h b/contrib/compiler-rt/lib/tsan/rtl/tsan_sync.h
index b83c09ff784e..9039970bcf86 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_sync.h
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_sync.h
@@ -34,6 +34,7 @@ enum MutexFlags {
   MutexFlagTryLockFailed       = 1 << 5, // __tsan_mutex_try_lock_failed
   MutexFlagRecursiveLock       = 1 << 6, // __tsan_mutex_recursive_lock
   MutexFlagRecursiveUnlock     = 1 << 7, // __tsan_mutex_recursive_unlock
+  MutexFlagNotStatic           = 1 << 8, // __tsan_mutex_not_static
 
   // The following flags are runtime private.
   // Mutex API misuse was detected, so don't report any more.
@@ -43,7 +44,8 @@ enum MutexFlags {
   // Must list all mutex creation flags.
   MutexCreationFlagMask        = MutexFlagLinkerInit |
                                  MutexFlagWriteReentrant |
-                                 MutexFlagReadReentrant,
+                                 MutexFlagReadReentrant |
+                                 MutexFlagNotStatic,
 };
 
 struct SyncVar {
diff --git a/contrib/compiler-rt/lib/tsan/rtl/tsan_trace.h b/contrib/compiler-rt/lib/tsan/rtl/tsan_trace.h
index 96a18ac4101a..9aef375cbfe6 100644
--- a/contrib/compiler-rt/lib/tsan/rtl/tsan_trace.h
+++ b/contrib/compiler-rt/lib/tsan/rtl/tsan_trace.h
@@ -41,6 +41,8 @@ enum EventType {
 // u64 addr : 61;  // Associated pc.
 typedef u64 Event;
 
+const uptr kEventPCBits = 61;
+
 struct TraceHeader {
 #if !SANITIZER_GO
   BufferedStackTrace stack0;  // Start stack for the trace.
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_checks.inc b/contrib/compiler-rt/lib/ubsan/ubsan_checks.inc
index 0a87e6e8e3a4..73c69363215a 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_checks.inc
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_checks.inc
@@ -29,6 +29,7 @@ UBSAN_CHECK(UnsignedIntegerOverflow, "unsigned-integer-overflow",
 UBSAN_CHECK(IntegerDivideByZero, "integer-divide-by-zero",
             "integer-divide-by-zero")
 UBSAN_CHECK(FloatDivideByZero, "float-divide-by-zero", "float-divide-by-zero")
+UBSAN_CHECK(InvalidBuiltin, "invalid-builtin-use", "invalid-builtin-use")
 UBSAN_CHECK(InvalidShiftBase, "invalid-shift-base", "shift-base")
 UBSAN_CHECK(InvalidShiftExponent, "invalid-shift-exponent", "shift-exponent")
 UBSAN_CHECK(OutOfBoundsIndex, "out-of-bounds-index", "bounds")
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_diag.cc b/contrib/compiler-rt/lib/ubsan/ubsan_diag.cc
index 742802b8f341..d5b68407bb37 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_diag.cc
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_diag.cc
@@ -26,21 +26,24 @@
 
 using namespace __ubsan;
 
+void __ubsan::GetStackTrace(BufferedStackTrace *stack, uptr max_depth, uptr pc,
+                            uptr bp, void *context, bool fast) {
+  uptr top = 0;
+  uptr bottom = 0;
+  if (fast)
+    GetThreadStackTopAndBottom(false, &top, &bottom);
+  stack->Unwind(max_depth, pc, bp, context, top, bottom, fast);
+}
+
 static void MaybePrintStackTrace(uptr pc, uptr bp) {
   // We assume that flags are already parsed, as UBSan runtime
   // will definitely be called when we print the first diagnostics message.
   if (!flags()->print_stacktrace)
     return;
 
-  uptr top = 0;
-  uptr bottom = 0;
-  bool request_fast_unwind = common_flags()->fast_unwind_on_fatal;
-  if (request_fast_unwind)
-    __sanitizer::GetThreadStackTopAndBottom(false, &top, &bottom);
-
   BufferedStackTrace stack;
-  stack.Unwind(kStackTraceMax, pc, bp, nullptr, top, bottom,
-               request_fast_unwind);
+  GetStackTrace(&stack, kStackTraceMax, pc, bp, nullptr,
+                common_flags()->fast_unwind_on_fatal);
   stack.Print();
 }
 
@@ -97,9 +100,7 @@ class Decorator : public SanitizerCommonDecorator {
  public:
   Decorator() : SanitizerCommonDecorator() {}
   const char *Highlight() const { return Green(); }
-  const char *EndHighlight() const { return Default(); }
   const char *Note() const { return Black(); }
-  const char *EndNote() const { return Default(); }
 };
 }
 
@@ -295,7 +296,7 @@ static void PrintMemorySnippet(const Decorator &Decor, MemoryLocation Loc,
     Buffer.append("%c", P == Loc ? '^' : Byte);
     Buffer.append("%c", Byte);
   }
-  Buffer.append("%s\n", Decor.EndHighlight());
+  Buffer.append("%s\n", Decor.Default());
 
   // Go over the line again, and print names for the ranges.
   InRange = 0;
@@ -335,7 +336,7 @@ static void PrintMemorySnippet(const Decorator &Decor, MemoryLocation Loc,
 
 Diag::~Diag() {
   // All diagnostics should be printed under report mutex.
-  CommonSanitizerReportMutex.CheckLocked();
+  ScopedReport::CheckLocked();
   Decorator Decor;
   InternalScopedString Buffer(1024);
 
@@ -345,12 +346,12 @@ Diag::~Diag() {
 
   switch (Level) {
   case DL_Error:
-    Buffer.append("%s runtime error: %s%s", Decor.Warning(), Decor.EndWarning(),
+    Buffer.append("%s runtime error: %s%s", Decor.Warning(), Decor.Default(),
                   Decor.Bold());
     break;
 
   case DL_Note:
-    Buffer.append("%s note: %s", Decor.Note(), Decor.EndNote());
+    Buffer.append("%s note: %s", Decor.Note(), Decor.Default());
     break;
   }
 
@@ -363,17 +364,15 @@ Diag::~Diag() {
     PrintMemorySnippet(Decor, Loc.getMemoryLocation(), Ranges, NumRanges, Args);
 }
 
+ScopedReport::Initializer::Initializer() { InitAsStandaloneIfNecessary(); }
+
 ScopedReport::ScopedReport(ReportOptions Opts, Location SummaryLoc,
                            ErrorType Type)
-    : Opts(Opts), SummaryLoc(SummaryLoc), Type(Type) {
-  InitAsStandaloneIfNecessary();
-  CommonSanitizerReportMutex.Lock();
-}
+    : Opts(Opts), SummaryLoc(SummaryLoc), Type(Type) {}
 
 ScopedReport::~ScopedReport() {
   MaybePrintStackTrace(Opts.pc, Opts.bp);
   MaybeReportErrorSummary(SummaryLoc, Type);
-  CommonSanitizerReportMutex.Unlock();
   if (flags()->halt_on_error)
     Die();
 }
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_diag.h b/contrib/compiler-rt/lib/ubsan/ubsan_diag.h
index 3edb67a03c1f..7370b1b62421 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_diag.h
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_diag.h
@@ -231,10 +231,19 @@ bool ignoreReport(SourceLocation SLoc, ReportOptions Opts, ErrorType ET);
     GET_CALLER_PC_BP; \
     ReportOptions Opts = {unrecoverable_handler, pc, bp}
 
+void GetStackTrace(BufferedStackTrace *stack, uptr max_depth, uptr pc, uptr bp,
+                   void *context, bool fast);
+
 /// \brief Instantiate this class before printing diagnostics in the error
 /// report. This class ensures that reports from different threads and from
 /// different sanitizers won't be mixed.
 class ScopedReport {
+  struct Initializer {
+    Initializer();
+  };
+  Initializer initializer_;
+  ScopedErrorReportLock report_lock_;
+
   ReportOptions Opts;
   Location SummaryLoc;
   ErrorType Type;
@@ -242,6 +251,8 @@ class ScopedReport {
 public:
   ScopedReport(ReportOptions Opts, Location SummaryLoc, ErrorType Type);
   ~ScopedReport();
+
+  static void CheckLocked() { ScopedErrorReportLock::CheckLocked(); }
 };
 
 void InitializeSuppressions();
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_diag_standalone.cc b/contrib/compiler-rt/lib/ubsan/ubsan_diag_standalone.cc
index df8ed5fcdf6d..1f4a5bd4062b 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_diag_standalone.cc
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_diag_standalone.cc
@@ -26,9 +26,10 @@ void __sanitizer_print_stack_trace() {
   if (request_fast_unwind)
     __sanitizer::GetThreadStackTopAndBottom(false, &top, &bottom);
 
-  GET_REPORT_OPTIONS(false);
+  GET_CURRENT_PC_BP_SP;
+  (void)sp;
   BufferedStackTrace stack;
-  stack.Unwind(kStackTraceMax, Opts.pc, Opts.bp, nullptr, top, bottom,
+  stack.Unwind(kStackTraceMax, pc, bp, nullptr, top, bottom,
                request_fast_unwind);
   stack.Print();
 }
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_flags.cc b/contrib/compiler-rt/lib/ubsan/ubsan_flags.cc
index 8e1f40885a58..c4fb278c8367 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_flags.cc
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_flags.cc
@@ -18,6 +18,8 @@
 #include "sanitizer_common/sanitizer_flags.h"
 #include "sanitizer_common/sanitizer_flag_parser.h"
 
+#include <stdlib.h>
+
 namespace __ubsan {
 
 const char *MaybeCallUbsanDefaultOptions() {
@@ -45,7 +47,7 @@ void InitializeFlags() {
     CommonFlags cf;
     cf.CopyFrom(*common_flags());
     cf.print_summary = false;
-    cf.external_symbolizer_path = GetEnv("UBSAN_SYMBOLIZER_PATH");
+    cf.external_symbolizer_path = getenv("UBSAN_SYMBOLIZER_PATH");
     OverrideCommonFlags(cf);
   }
 
@@ -59,7 +61,7 @@ void InitializeFlags() {
   // Override from user-specified string.
   parser.ParseString(MaybeCallUbsanDefaultOptions());
   // Override from environment variable.
-  parser.ParseString(GetEnv("UBSAN_OPTIONS"));
+  parser.ParseString(getenv("UBSAN_OPTIONS"));
   InitializeCommonFlags();
   if (Verbosity()) ReportUnrecognizedFlags();
 
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_handlers.cc b/contrib/compiler-rt/lib/ubsan/ubsan_handlers.cc
index 75a4490a1843..1112ce1cc5ff 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_handlers.cc
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_handlers.cc
@@ -437,6 +437,30 @@ void __ubsan::__ubsan_handle_load_invalid_value_abort(InvalidValueData *Data,
   Die();
 }
 
+static void handleInvalidBuiltin(InvalidBuiltinData *Data, ReportOptions Opts) {
+  SourceLocation Loc = Data->Loc.acquire();
+  ErrorType ET = ErrorType::InvalidBuiltin;
+
+  if (ignoreReport(Loc, Opts, ET))
+    return;
+
+  ScopedReport R(Opts, Loc, ET);
+
+  Diag(Loc, DL_Error,
+       "passing zero to %0, which is not a valid argument")
+    << ((Data->Kind == BCK_CTZPassedZero) ? "ctz()" : "clz()");
+}
+
+void __ubsan::__ubsan_handle_invalid_builtin(InvalidBuiltinData *Data) {
+  GET_REPORT_OPTIONS(true);
+  handleInvalidBuiltin(Data, Opts);
+}
+void __ubsan::__ubsan_handle_invalid_builtin_abort(InvalidBuiltinData *Data) {
+  GET_REPORT_OPTIONS(true);
+  handleInvalidBuiltin(Data, Opts);
+  Die();
+}
+
 static void handleFunctionTypeMismatch(FunctionTypeMismatchData *Data,
                                        ValueHandle Function,
                                        ReportOptions Opts) {
@@ -628,16 +652,32 @@ static void handleCFIBadIcall(CFICheckFailData *Data, ValueHandle Function,
 }
 
 namespace __ubsan {
+
 #ifdef UBSAN_CAN_USE_CXXABI
+
+#ifdef _WIN32
+
+extern "C" void __ubsan_handle_cfi_bad_type_default(CFICheckFailData *Data,
+                                                    ValueHandle Vtable,
+                                                    bool ValidVtable,
+                                                    ReportOptions Opts) {
+  Die();
+}
+
+WIN_WEAK_ALIAS(__ubsan_handle_cfi_bad_type, __ubsan_handle_cfi_bad_type_default)
+#else
 SANITIZER_WEAK_ATTRIBUTE
-void HandleCFIBadType(CFICheckFailData *Data, ValueHandle Vtable,
-                      bool ValidVtable, ReportOptions Opts);
+#endif
+void __ubsan_handle_cfi_bad_type(CFICheckFailData *Data, ValueHandle Vtable,
+                                 bool ValidVtable, ReportOptions Opts);
+
 #else
-static void HandleCFIBadType(CFICheckFailData *Data, ValueHandle Vtable,
-                             bool ValidVtable, ReportOptions Opts) {
+void __ubsan_handle_cfi_bad_type(CFICheckFailData *Data, ValueHandle Vtable,
+                                 bool ValidVtable, ReportOptions Opts) {
   Die();
 }
 #endif
+
 }  // namespace __ubsan
 
 void __ubsan::__ubsan_handle_cfi_check_fail(CFICheckFailData *Data,
@@ -647,7 +687,7 @@ void __ubsan::__ubsan_handle_cfi_check_fail(CFICheckFailData *Data,
   if (Data->CheckKind == CFITCK_ICall)
     handleCFIBadIcall(Data, Value, Opts);
   else
-    HandleCFIBadType(Data, Value, ValidVtable, Opts);
+    __ubsan_handle_cfi_bad_type(Data, Value, ValidVtable, Opts);
 }
 
 void __ubsan::__ubsan_handle_cfi_check_fail_abort(CFICheckFailData *Data,
@@ -657,7 +697,7 @@ void __ubsan::__ubsan_handle_cfi_check_fail_abort(CFICheckFailData *Data,
   if (Data->CheckKind == CFITCK_ICall)
     handleCFIBadIcall(Data, Value, Opts);
   else
-    HandleCFIBadType(Data, Value, ValidVtable, Opts);
+    __ubsan_handle_cfi_bad_type(Data, Value, ValidVtable, Opts);
   Die();
 }
 
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_handlers.h b/contrib/compiler-rt/lib/ubsan/ubsan_handlers.h
index 796321b81889..311776b9f22c 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_handlers.h
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_handlers.h
@@ -122,6 +122,21 @@ struct InvalidValueData {
 /// \brief Handle a load of an invalid value for the type.
 RECOVERABLE(load_invalid_value, InvalidValueData *Data, ValueHandle Val)
 
+/// Known builtin check kinds.
+/// Keep in sync with the enum of the same name in CodeGenFunction.h
+enum BuiltinCheckKind : unsigned char {
+  BCK_CTZPassedZero,
+  BCK_CLZPassedZero,
+};
+
+struct InvalidBuiltinData {
+  SourceLocation Loc;
+  unsigned char Kind;
+};
+
+/// Handle a builtin called in an invalid way.
+RECOVERABLE(invalid_builtin, InvalidBuiltinData *Data)
+
 struct FunctionTypeMismatchData {
   SourceLocation Loc;
   const TypeDescriptor &Type;
@@ -177,6 +192,13 @@ struct CFICheckFailData {
 /// \brief Handle control flow integrity failures.
 RECOVERABLE(cfi_check_fail, CFICheckFailData *Data, ValueHandle Function,
             uptr VtableIsValid)
+
+struct ReportOptions;
+
+extern "C" SANITIZER_INTERFACE_ATTRIBUTE void __ubsan_handle_cfi_bad_type(
+    CFICheckFailData *Data, ValueHandle Vtable, bool ValidVtable,
+    ReportOptions Opts);
+
 }
 
 #endif // UBSAN_HANDLERS_H
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_handlers_cxx.cc b/contrib/compiler-rt/lib/ubsan/ubsan_handlers_cxx.cc
index d97ec4813ccd..e15abc64ecca 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_handlers_cxx.cc
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_handlers_cxx.cc
@@ -95,8 +95,8 @@ void __ubsan::__ubsan_handle_dynamic_type_cache_miss_abort(
 }
 
 namespace __ubsan {
-void HandleCFIBadType(CFICheckFailData *Data, ValueHandle Vtable,
-                      bool ValidVtable, ReportOptions Opts) {
+void __ubsan_handle_cfi_bad_type(CFICheckFailData *Data, ValueHandle Vtable,
+                                 bool ValidVtable, ReportOptions Opts) {
   SourceLocation Loc = Data->Loc.acquire();
   ErrorType ET = ErrorType::CFIBadType;
 
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_init.cc b/contrib/compiler-rt/lib/ubsan/ubsan_init.cc
index 307bca37680e..32fc434ad2bc 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_init.cc
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_init.cc
@@ -27,11 +27,7 @@ const char *__ubsan::GetSanititizerToolName() {
   return "UndefinedBehaviorSanitizer";
 }
 
-static enum {
-  UBSAN_MODE_UNKNOWN = 0,
-  UBSAN_MODE_STANDALONE,
-  UBSAN_MODE_PLUGIN
-} ubsan_mode;
+static bool ubsan_initialized;
 static StaticSpinMutex ubsan_init_mu;
 
 static void CommonInit() {
@@ -40,44 +36,30 @@ static void CommonInit() {
 
 static void CommonStandaloneInit() {
   SanitizerToolName = GetSanititizerToolName();
-  InitializeFlags();
   CacheBinaryName();
+  InitializeFlags();
   __sanitizer_set_report_path(common_flags()->log_path);
   AndroidLogInit();
   InitializeCoverage(common_flags()->coverage, common_flags()->coverage_dir);
   CommonInit();
-  ubsan_mode = UBSAN_MODE_STANDALONE;
 }
 
 void __ubsan::InitAsStandalone() {
-  if (SANITIZER_CAN_USE_PREINIT_ARRAY) {
-    CHECK_EQ(UBSAN_MODE_UNKNOWN, ubsan_mode);
-    CommonStandaloneInit();
-    return;
-  }
   SpinMutexLock l(&ubsan_init_mu);
-  CHECK_NE(UBSAN_MODE_PLUGIN, ubsan_mode);
-  if (ubsan_mode == UBSAN_MODE_UNKNOWN)
+  if (!ubsan_initialized) {
     CommonStandaloneInit();
-}
-
-void __ubsan::InitAsStandaloneIfNecessary() {
-  if (SANITIZER_CAN_USE_PREINIT_ARRAY) {
-    CHECK_NE(UBSAN_MODE_UNKNOWN, ubsan_mode);
-    return;
+    ubsan_initialized = true;
   }
-  SpinMutexLock l(&ubsan_init_mu);
-  if (ubsan_mode == UBSAN_MODE_UNKNOWN)
-    CommonStandaloneInit();
 }
 
+void __ubsan::InitAsStandaloneIfNecessary() { return InitAsStandalone(); }
+
 void __ubsan::InitAsPlugin() {
-#if !SANITIZER_CAN_USE_PREINIT_ARRAY
   SpinMutexLock l(&ubsan_init_mu);
-#endif
-  CHECK_EQ(UBSAN_MODE_UNKNOWN, ubsan_mode);
-  CommonInit();
-  ubsan_mode = UBSAN_MODE_PLUGIN;
+  if (!ubsan_initialized) {
+    CommonInit();
+    ubsan_initialized = true;
+  }
 }
 
 #endif  // CAN_SANITIZE_UB
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_init_standalone.cc b/contrib/compiler-rt/lib/ubsan/ubsan_init_standalone.cc
index ff1a20efea3d..8bd500025cbc 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_init_standalone.cc
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_init_standalone.cc
@@ -18,18 +18,17 @@
 
 #include "sanitizer_common/sanitizer_internal_defs.h"
 #include "ubsan_init.h"
+#include "ubsan_signals_standalone.h"
+
+namespace __ubsan {
 
-#if SANITIZER_CAN_USE_PREINIT_ARRAY
-__attribute__((section(".preinit_array"), used))
-void (*__local_ubsan_preinit)(void) = __ubsan::InitAsStandalone;
-#else
-// Use a dynamic initializer.
 class UbsanStandaloneInitializer {
  public:
   UbsanStandaloneInitializer() {
-    __ubsan::InitAsStandalone();
+    InitAsStandalone();
+    InitializeDeadlySignals();
   }
 };
 static UbsanStandaloneInitializer ubsan_standalone_initializer;
-#endif  // SANITIZER_CAN_USE_PREINIT_ARRAY
 
+} // namespace __ubsan
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_interface.inc b/contrib/compiler-rt/lib/ubsan/ubsan_interface.inc
index a69ca57cd7af..9beb3e2ff953 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_interface.inc
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_interface.inc
@@ -11,14 +11,19 @@
 INTERFACE_FUNCTION(__ubsan_handle_add_overflow)
 INTERFACE_FUNCTION(__ubsan_handle_add_overflow_abort)
 INTERFACE_FUNCTION(__ubsan_handle_builtin_unreachable)
+INTERFACE_FUNCTION(__ubsan_handle_cfi_bad_type)
 INTERFACE_FUNCTION(__ubsan_handle_cfi_check_fail)
 INTERFACE_FUNCTION(__ubsan_handle_cfi_check_fail_abort)
 INTERFACE_FUNCTION(__ubsan_handle_divrem_overflow)
 INTERFACE_FUNCTION(__ubsan_handle_divrem_overflow_abort)
+INTERFACE_FUNCTION(__ubsan_handle_dynamic_type_cache_miss)
+INTERFACE_FUNCTION(__ubsan_handle_dynamic_type_cache_miss_abort)
 INTERFACE_FUNCTION(__ubsan_handle_float_cast_overflow)
 INTERFACE_FUNCTION(__ubsan_handle_float_cast_overflow_abort)
 INTERFACE_FUNCTION(__ubsan_handle_function_type_mismatch)
 INTERFACE_FUNCTION(__ubsan_handle_function_type_mismatch_abort)
+INTERFACE_FUNCTION(__ubsan_handle_invalid_builtin)
+INTERFACE_FUNCTION(__ubsan_handle_invalid_builtin_abort)
 INTERFACE_FUNCTION(__ubsan_handle_load_invalid_value)
 INTERFACE_FUNCTION(__ubsan_handle_load_invalid_value_abort)
 INTERFACE_FUNCTION(__ubsan_handle_missing_return)
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_platform.h b/contrib/compiler-rt/lib/ubsan/ubsan_platform.h
index 1a3bfd6afb81..72eb4199960e 100644
--- a/contrib/compiler-rt/lib/ubsan/ubsan_platform.h
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_platform.h
@@ -14,12 +14,13 @@
 #define UBSAN_PLATFORM_H
 
 // Other platforms should be easy to add, and probably work as-is.
-#if (defined(__linux__) || defined(__FreeBSD__) || defined(__APPLE__)) && \
-    (defined(__x86_64__) || defined(__i386__) || defined(__arm__) || \
-     defined(__aarch64__) || defined(__mips__) || defined(__powerpc64__) || \
-     defined(__s390__))
+#if (defined(__linux__) || defined(__FreeBSD__) || defined(__APPLE__) ||       \
+     defined(__NetBSD__)) &&                                                   \
+    (defined(__x86_64__) || defined(__i386__) || defined(__arm__) ||           \
+     defined(__aarch64__) || defined(__mips__) || defined(__powerpc64__) ||    \
+     defined(__s390__)) || (defined(__sun__) && defined(__svr4__))
 # define CAN_SANITIZE_UB 1
-#elif defined(_WIN32)
+#elif defined(_WIN32) || defined(__Fuchsia__)
 # define CAN_SANITIZE_UB 1
 #else
 # define CAN_SANITIZE_UB 0
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_signals_standalone.cc b/contrib/compiler-rt/lib/ubsan/ubsan_signals_standalone.cc
new file mode 100644
index 000000000000..4c6646dd8b0b
--- /dev/null
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_signals_standalone.cc
@@ -0,0 +1,53 @@
+//=-- ubsan_signals_standalone.cc
+//------------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Installs signal handlers and related interceptors for UBSan standalone.
+//
+//===----------------------------------------------------------------------===//
+
+#include "ubsan_platform.h"
+#if CAN_SANITIZE_UB
+#include "interception/interception.h"
+#include "sanitizer_common/sanitizer_stacktrace.h"
+#include "ubsan_diag.h"
+#include "ubsan_init.h"
+
+#define COMMON_INTERCEPT_FUNCTION(name) INTERCEPT_FUNCTION(name)
+#include "sanitizer_common/sanitizer_signal_interceptors.inc"
+
+namespace __ubsan {
+
+#if SANITIZER_FUCHSIA
+void InitializeDeadlySignals() {}
+#else
+static void OnStackUnwind(const SignalContext &sig, const void *,
+                          BufferedStackTrace *stack) {
+  GetStackTrace(stack, kStackTraceMax, sig.pc, sig.bp, sig.context,
+                common_flags()->fast_unwind_on_fatal);
+}
+
+static void UBsanOnDeadlySignal(int signo, void *siginfo, void *context) {
+  HandleDeadlySignal(siginfo, context, GetTid(), &OnStackUnwind, nullptr);
+}
+
+static bool is_initialized = false;
+
+void InitializeDeadlySignals() {
+  if (is_initialized)
+    return;
+  is_initialized = true;
+  InitializeSignalInterceptors();
+  InstallDeadlySignalHandlers(&UBsanOnDeadlySignal);
+}
+#endif
+
+} // namespace __ubsan
+
+#endif // CAN_SANITIZE_UB
diff --git a/contrib/compiler-rt/lib/ubsan/ubsan_signals_standalone.h b/contrib/compiler-rt/lib/ubsan/ubsan_signals_standalone.h
new file mode 100644
index 000000000000..b29c29482ec8
--- /dev/null
+++ b/contrib/compiler-rt/lib/ubsan/ubsan_signals_standalone.h
@@ -0,0 +1,25 @@
+//=-- ubsan_signals_standalone.h
+//------------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Installs signal handlers and related interceptors for UBSan standalone.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef UBSAN_SIGNALS_STANDALONE_H
+#define UBSAN_SIGNALS_STANDALONE_H
+
+namespace __ubsan {
+
+// Initializes signal handlers and interceptors.
+void InitializeDeadlySignals();
+
+} // namespace __ubsan
+
+#endif // UBSAN_SIGNALS_STANDALONE_H
diff --git a/contrib/compiler-rt/lib/ubsan_minimal/ubsan.syms.extra b/contrib/compiler-rt/lib/ubsan_minimal/ubsan.syms.extra
new file mode 100644
index 000000000000..7f8be694401a
--- /dev/null
+++ b/contrib/compiler-rt/lib/ubsan_minimal/ubsan.syms.extra
@@ -0,0 +1 @@
+__ubsan_*
diff --git a/contrib/compiler-rt/lib/ubsan_minimal/ubsan_minimal_handlers.cc b/contrib/compiler-rt/lib/ubsan_minimal/ubsan_minimal_handlers.cc
new file mode 100644
index 000000000000..5a5675c983fe
--- /dev/null
+++ b/contrib/compiler-rt/lib/ubsan_minimal/ubsan_minimal_handlers.cc
@@ -0,0 +1,104 @@
+#include "sanitizer_common/sanitizer_atomic.h"
+
+#include <stdlib.h>
+#include <stdint.h>
+#include <string.h>
+#include <unistd.h>
+
+#ifdef KERNEL_USE
+extern "C" void ubsan_message(const char *msg);
+static void message(const char *msg) { ubsan_message(msg); }
+#else
+static void message(const char *msg) {
+  write(2, msg, strlen(msg));
+}
+#endif
+
+static const int kMaxCallerPcs = 20;
+static __sanitizer::atomic_uintptr_t caller_pcs[kMaxCallerPcs];
+// Number of elements in caller_pcs. A special value of kMaxCallerPcs + 1 means
+// that "too many errors" has already been reported.
+static __sanitizer::atomic_uint32_t caller_pcs_sz;
+
+__attribute__((noinline)) static bool report_this_error(void *caller_p) {
+  uintptr_t caller = reinterpret_cast<uintptr_t>(caller_p);
+  if (caller == 0) return false;
+  while (true) {
+    unsigned sz = __sanitizer::atomic_load_relaxed(&caller_pcs_sz);
+    if (sz > kMaxCallerPcs) return false;  // early exit
+    // when sz==kMaxCallerPcs print "too many errors", but only when cmpxchg
+    // succeeds in order to not print it multiple times.
+    if (sz > 0 && sz < kMaxCallerPcs) {
+      uintptr_t p;
+      for (unsigned i = 0; i < sz; ++i) {
+        p = __sanitizer::atomic_load_relaxed(&caller_pcs[i]);
+        if (p == 0) break;  // Concurrent update.
+        if (p == caller) return false;
+      }
+      if (p == 0) continue;  // FIXME: yield?
+    }
+
+    if (!__sanitizer::atomic_compare_exchange_strong(
+            &caller_pcs_sz, &sz, sz + 1, __sanitizer::memory_order_seq_cst))
+      continue;  // Concurrent update! Try again from the start.
+
+    if (sz == kMaxCallerPcs) {
+      message("ubsan: too many errors\n");
+      return false;
+    }
+    __sanitizer::atomic_store_relaxed(&caller_pcs[sz], caller);
+    return true;
+  }
+}
+
+#if defined(__ANDROID__)
+extern "C" __attribute__((weak)) void android_set_abort_message(const char *);
+static void abort_with_message(const char *msg) {
+  if (&android_set_abort_message) android_set_abort_message(msg);
+  abort();
+}
+#else
+static void abort_with_message(const char *) { abort(); }
+#endif
+
+#define INTERFACE extern "C" __attribute__((visibility("default")))
+
+// FIXME: add caller pc to the error message (possibly as "ubsan: error-type
+// @1234ABCD").
+#define HANDLER_RECOVER(name, msg)                               \
+  INTERFACE void __ubsan_handle_##name##_minimal() {             \
+    if (!report_this_error(__builtin_return_address(0))) return; \
+    message("ubsan: " msg "\n");                                 \
+  }
+
+#define HANDLER_NORECOVER(name, msg)                             \
+  INTERFACE void __ubsan_handle_##name##_minimal_abort() {       \
+    message("ubsan: " msg "\n");                                 \
+    abort_with_message("ubsan: " msg);                           \
+  }
+
+#define HANDLER(name, msg)                                       \
+  HANDLER_RECOVER(name, msg)                                     \
+  HANDLER_NORECOVER(name, msg)
+
+HANDLER(type_mismatch, "type-mismatch")
+HANDLER(add_overflow, "add-overflow")
+HANDLER(sub_overflow, "sub-overflow")
+HANDLER(mul_overflow, "mul-overflow")
+HANDLER(negate_overflow, "negate-overflow")
+HANDLER(divrem_overflow, "divrem-overflow")
+HANDLER(shift_out_of_bounds, "shift-out-of-bounds")
+HANDLER(out_of_bounds, "out-of-bounds")
+HANDLER_RECOVER(builtin_unreachable, "builtin-unreachable")
+HANDLER_RECOVER(missing_return, "missing-return")
+HANDLER(vla_bound_not_positive, "vla-bound-not-positive")
+HANDLER(float_cast_overflow, "float-cast-overflow")
+HANDLER(load_invalid_value, "load-invalid-value")
+HANDLER(invalid_builtin, "invalid-builtin")
+HANDLER(function_type_mismatch, "function-type-mismatch")
+HANDLER(nonnull_arg, "nonnull-arg")
+HANDLER(nonnull_return, "nonnull-return")
+HANDLER(nullability_arg, "nullability-arg")
+HANDLER(nullability_return, "nullability-return")
+HANDLER(pointer_overflow, "pointer-overflow")
+HANDLER(cfi_check_fail, "cfi-check-fail")
diff --git a/contrib/compiler-rt/lib/xray/weak_symbols.txt b/contrib/compiler-rt/lib/xray/weak_symbols.txt
new file mode 100644
index 000000000000..963fff2d697e
--- /dev/null
+++ b/contrib/compiler-rt/lib/xray/weak_symbols.txt
@@ -0,0 +1,4 @@
+___start_xray_fn_idx
+___start_xray_instr_map
+___stop_xray_fn_idx
+___stop_xray_instr_map
diff --git a/contrib/compiler-rt/lib/xray/xray_buffer_queue.cc b/contrib/compiler-rt/lib/xray/xray_buffer_queue.cc
index 7ba755ac3069..a0018f6b0cba 100644
--- a/contrib/compiler-rt/lib/xray/xray_buffer_queue.cc
+++ b/contrib/compiler-rt/lib/xray/xray_buffer_queue.cc
@@ -13,28 +13,34 @@
 //
 //===----------------------------------------------------------------------===//
 #include "xray_buffer_queue.h"
+#include "sanitizer_common/sanitizer_allocator_internal.h"
 #include "sanitizer_common/sanitizer_common.h"
 #include "sanitizer_common/sanitizer_libc.h"
 
-#include <cstdlib>
-#include <tuple>
-
 using namespace __xray;
 using namespace __sanitizer;
 
-BufferQueue::BufferQueue(std::size_t B, std::size_t N, bool &Success)
-    : BufferSize(B), Buffers(N), Mutex(), OwnedBuffers(), Finalizing{0} {
-  for (auto &T : Buffers) {
-    void *Tmp = malloc(BufferSize);
+BufferQueue::BufferQueue(size_t B, size_t N, bool &Success)
+    : BufferSize(B), Buffers(new BufferRep[N]()), BufferCount(N), Finalizing{0},
+      OwnedBuffers(new void *[N]()), Next(Buffers), First(Buffers),
+      LiveBuffers(0) {
+  for (size_t i = 0; i < N; ++i) {
+    auto &T = Buffers[i];
+    void *Tmp = InternalAlloc(BufferSize, nullptr, 64);
     if (Tmp == nullptr) {
       Success = false;
       return;
     }
-
-    auto &Buf = std::get<0>(T);
+    void *Extents = InternalAlloc(sizeof(BufferExtents), nullptr, 64);
+    if (Extents == nullptr) {
+      Success = false;
+      return;
+    }
+    auto &Buf = T.Buff;
     Buf.Buffer = Tmp;
     Buf.Size = B;
-    OwnedBuffers.emplace(Tmp);
+    Buf.Extents = reinterpret_cast<BufferExtents *>(Extents);
+    OwnedBuffers[i] = Tmp;
   }
   Success = true;
 }
@@ -42,27 +48,50 @@ BufferQueue::BufferQueue(std::size_t B, std::size_t N, bool &Success)
 BufferQueue::ErrorCode BufferQueue::getBuffer(Buffer &Buf) {
   if (__sanitizer::atomic_load(&Finalizing, __sanitizer::memory_order_acquire))
     return ErrorCode::QueueFinalizing;
-  __sanitizer::BlockingMutexLock Guard(&Mutex);
-  if (Buffers.empty())
+  __sanitizer::SpinMutexLock Guard(&Mutex);
+  if (LiveBuffers == BufferCount)
     return ErrorCode::NotEnoughMemory;
-  auto &T = Buffers.front();
-  auto &B = std::get<0>(T);
+
+  auto &T = *Next;
+  auto &B = T.Buff;
   Buf = B;
-  B.Buffer = nullptr;
-  B.Size = 0;
-  Buffers.pop_front();
+  T.Used = true;
+  ++LiveBuffers;
+
+  if (++Next == (Buffers + BufferCount))
+    Next = Buffers;
+
   return ErrorCode::Ok;
 }
 
 BufferQueue::ErrorCode BufferQueue::releaseBuffer(Buffer &Buf) {
-  if (OwnedBuffers.count(Buf.Buffer) == 0)
+  // Blitz through the buffers array to find the buffer.
+  bool Found = false;
+  for (auto I = OwnedBuffers, E = OwnedBuffers + BufferCount; I != E; ++I) {
+    if (*I == Buf.Buffer) {
+      Found = true;
+      break;
+    }
+  }
+  if (!Found)
     return ErrorCode::UnrecognizedBuffer;
-  __sanitizer::BlockingMutexLock Guard(&Mutex);
+
+  __sanitizer::SpinMutexLock Guard(&Mutex);
+
+  // This points to a semantic bug, we really ought to not be releasing more
+  // buffers than we actually get.
+  if (LiveBuffers == 0)
+    return ErrorCode::NotEnoughMemory;
 
   // Now that the buffer has been released, we mark it as "used".
-  Buffers.emplace(Buffers.end(), Buf, true /* used */);
+  First->Buff = Buf;
+  First->Used = true;
   Buf.Buffer = nullptr;
   Buf.Size = 0;
+  --LiveBuffers;
+  if (++First == (Buffers + BufferCount))
+    First = Buffers;
+
   return ErrorCode::Ok;
 }
 
@@ -74,8 +103,12 @@ BufferQueue::ErrorCode BufferQueue::finalize() {
 }
 
 BufferQueue::~BufferQueue() {
-  for (auto &T : Buffers) {
-    auto &Buf = std::get<0>(T);
-    free(Buf.Buffer);
+  for (auto I = Buffers, E = Buffers + BufferCount; I != E; ++I) {
+    auto &T = *I;
+    auto &Buf = T.Buff;
+    InternalFree(Buf.Buffer);
+    InternalFree(Buf.Extents);
   }
+  delete[] Buffers;
+  delete[] OwnedBuffers;
 }
diff --git a/contrib/compiler-rt/lib/xray/xray_buffer_queue.h b/contrib/compiler-rt/lib/xray/xray_buffer_queue.h
index e051695a297b..1ceb58274616 100644
--- a/contrib/compiler-rt/lib/xray/xray_buffer_queue.h
+++ b/contrib/compiler-rt/lib/xray/xray_buffer_queue.h
@@ -15,11 +15,9 @@
 #ifndef XRAY_BUFFER_QUEUE_H
 #define XRAY_BUFFER_QUEUE_H
 
+#include <cstddef>
 #include "sanitizer_common/sanitizer_atomic.h"
 #include "sanitizer_common/sanitizer_mutex.h"
-#include <deque>
-#include <unordered_set>
-#include <utility>
 
 namespace __xray {
 
@@ -29,23 +27,50 @@ namespace __xray {
 /// the "flight data recorder" (FDR) mode to support ongoing XRay function call
 /// trace collection.
 class BufferQueue {
-public:
+ public:
+  struct alignas(64) BufferExtents {
+    __sanitizer::atomic_uint64_t Size;
+  };
+
   struct Buffer {
     void *Buffer = nullptr;
     size_t Size = 0;
+    BufferExtents* Extents;
+  };
+
+ private:
+  struct BufferRep {
+    // The managed buffer.
+    Buffer Buff;
+
+    // This is true if the buffer has been returned to the available queue, and
+    // is considered "used" by another thread.
+    bool Used = false;
   };
 
-private:
+  // Size of each individual Buffer.
   size_t BufferSize;
 
-  // We use a bool to indicate whether the Buffer has been used in this
-  // freelist implementation.
-  std::deque<std::tuple<Buffer, bool>> Buffers;
-  __sanitizer::BlockingMutex Mutex;
-  std::unordered_set<void *> OwnedBuffers;
+  BufferRep *Buffers;
+  size_t BufferCount;
+
+  __sanitizer::SpinMutex Mutex;
   __sanitizer::atomic_uint8_t Finalizing;
 
-public:
+  // Pointers to buffers managed/owned by the BufferQueue.
+  void **OwnedBuffers;
+
+  // Pointer to the next buffer to be handed out.
+  BufferRep *Next;
+
+  // Pointer to the entry in the array where the next released buffer will be
+  // placed.
+  BufferRep *First;
+
+  // Count of buffers that have been handed out through 'getBuffer'.
+  size_t LiveBuffers;
+
+ public:
   enum class ErrorCode : unsigned {
     Ok,
     NotEnoughMemory,
@@ -56,16 +81,16 @@ public:
 
   static const char *getErrorString(ErrorCode E) {
     switch (E) {
-    case ErrorCode::Ok:
-      return "(none)";
-    case ErrorCode::NotEnoughMemory:
-      return "no available buffers in the queue";
-    case ErrorCode::QueueFinalizing:
-      return "queue already finalizing";
-    case ErrorCode::UnrecognizedBuffer:
-      return "buffer being returned not owned by buffer queue";
-    case ErrorCode::AlreadyFinalized:
-      return "queue already finalized";
+      case ErrorCode::Ok:
+        return "(none)";
+      case ErrorCode::NotEnoughMemory:
+        return "no available buffers in the queue";
+      case ErrorCode::QueueFinalizing:
+        return "queue already finalizing";
+      case ErrorCode::UnrecognizedBuffer:
+        return "buffer being returned not owned by buffer queue";
+      case ErrorCode::AlreadyFinalized:
+        return "queue already finalized";
     }
     return "unknown error";
   }
@@ -82,15 +107,18 @@ public:
   ///   - BufferQueue is not finalising.
   ///
   /// Returns:
-  ///   - std::errc::not_enough_memory on exceeding MaxSize.
-  ///   - no error when we find a Buffer.
-  ///   - std::errc::state_not_recoverable on finalising BufferQueue.
+  ///   - ErrorCode::NotEnoughMemory on exceeding MaxSize.
+  ///   - ErrorCode::Ok when we find a Buffer.
+  ///   - ErrorCode::QueueFinalizing or ErrorCode::AlreadyFinalized on
+  ///     a finalizing/finalized BufferQueue.
   ErrorCode getBuffer(Buffer &Buf);
 
   /// Updates |Buf| to point to nullptr, with size 0.
   ///
   /// Returns:
-  ///   - ...
+  ///   - ErrorCode::Ok when we successfully release the buffer.
+  ///   - ErrorCode::UnrecognizedBuffer for when this BufferQueue does not own
+  ///     the buffer being released.
   ErrorCode releaseBuffer(Buffer &Buf);
 
   bool finalizing() const {
@@ -107,17 +135,18 @@ public:
   ///   - All releaseBuffer operations will not fail.
   ///
   /// After a call to finalize succeeds, all subsequent calls to finalize will
-  /// fail with std::errc::state_not_recoverable.
+  /// fail with ErrorCode::QueueFinalizing.
   ErrorCode finalize();
 
   /// Applies the provided function F to each Buffer in the queue, only if the
   /// Buffer is marked 'used' (i.e. has been the result of getBuffer(...) and a
-  /// releaseBuffer(...) operation.
-  template <class F> void apply(F Fn) {
-    __sanitizer::BlockingMutexLock G(&Mutex);
-    for (const auto &T : Buffers) {
-      if (std::get<1>(T))
-        Fn(std::get<0>(T));
+  /// releaseBuffer(...) operation).
+  template <class F>
+  void apply(F Fn) {
+    __sanitizer::SpinMutexLock G(&Mutex);
+    for (auto I = Buffers, E = Buffers + BufferCount; I != E; ++I) {
+      const auto &T = *I;
+      if (T.Used) Fn(T.Buff);
     }
   }
 
@@ -125,6 +154,6 @@ public:
   ~BufferQueue();
 };
 
-} // namespace __xray
+}  // namespace __xray
 
-#endif // XRAY_BUFFER_QUEUE_H
+#endif  // XRAY_BUFFER_QUEUE_H
diff --git a/contrib/compiler-rt/lib/xray/xray_fdr_log_records.h b/contrib/compiler-rt/lib/xray/xray_fdr_log_records.h
index 3d6d38892c76..324208db82ca 100644
--- a/contrib/compiler-rt/lib/xray/xray_fdr_log_records.h
+++ b/contrib/compiler-rt/lib/xray/xray_fdr_log_records.h
@@ -30,7 +30,10 @@ struct alignas(16) MetadataRecord {
     TSCWrap,
     WalltimeMarker,
     CustomEventMarker,
+    CallArgument,
+    BufferExtents,
   };
+
   // Use 7 bits to identify this record type.
   /* RecordKinds */ uint8_t RecordKind : 7;
   char Data[15];
diff --git a/contrib/compiler-rt/lib/xray/xray_fdr_logging.cc b/contrib/compiler-rt/lib/xray/xray_fdr_logging.cc
index a7e1382c3865..1bfa10c21f5c 100644
--- a/contrib/compiler-rt/lib/xray/xray_fdr_logging.cc
+++ b/contrib/compiler-rt/lib/xray/xray_fdr_logging.cc
@@ -15,15 +15,11 @@
 //
 //===----------------------------------------------------------------------===//
 #include "xray_fdr_logging.h"
-#include <algorithm>
-#include <bitset>
-#include <cerrno>
-#include <cstring>
+#include <errno.h>
 #include <sys/syscall.h>
 #include <sys/time.h>
 #include <time.h>
 #include <unistd.h>
-#include <unordered_map>
 
 #include "sanitizer_common/sanitizer_atomic.h"
 #include "sanitizer_common/sanitizer_common.h"
@@ -39,7 +35,7 @@
 namespace __xray {
 
 // Global BufferQueue.
-std::shared_ptr<BufferQueue> BQ;
+BufferQueue *BQ = nullptr;
 
 __sanitizer::atomic_sint32_t LogFlushStatus = {
     XRayLogFlushStatus::XRAY_LOG_NOT_FLUSHING};
@@ -52,19 +48,31 @@ __sanitizer::SpinMutex FDROptionsMutex;
 XRayLogFlushStatus fdrLoggingFlush() XRAY_NEVER_INSTRUMENT {
   if (__sanitizer::atomic_load(&LoggingStatus,
                                __sanitizer::memory_order_acquire) !=
-      XRayLogInitStatus::XRAY_LOG_FINALIZED)
+      XRayLogInitStatus::XRAY_LOG_FINALIZED) {
+    if (__sanitizer::Verbosity())
+      Report("Not flushing log, implementation is not finalized.\n");
     return XRayLogFlushStatus::XRAY_LOG_NOT_FLUSHING;
+  }
 
   s32 Result = XRayLogFlushStatus::XRAY_LOG_NOT_FLUSHING;
   if (!__sanitizer::atomic_compare_exchange_strong(
           &LogFlushStatus, &Result, XRayLogFlushStatus::XRAY_LOG_FLUSHING,
-          __sanitizer::memory_order_release))
+          __sanitizer::memory_order_release)) {
+
+    if (__sanitizer::Verbosity())
+      Report("Not flushing log, implementation is still finalizing.\n");
     return static_cast<XRayLogFlushStatus>(Result);
+  }
 
-  // Make a copy of the BufferQueue pointer to prevent other threads that may be
-  // resetting it from blowing away the queue prematurely while we're dealing
-  // with it.
-  auto LocalBQ = BQ;
+  if (BQ == nullptr) {
+    if (__sanitizer::Verbosity())
+      Report("Cannot flush when global buffer queue is null.\n");
+    return XRayLogFlushStatus::XRAY_LOG_NOT_FLUSHING;
+  }
+
+  // We wait a number of milliseconds to allow threads to see that we've
+  // finalised before attempting to flush the log.
+  __sanitizer::SleepForMillis(flags()->xray_fdr_log_grace_period_ms);
 
   // We write out the file in the following format:
   //
@@ -95,24 +103,44 @@ XRayLogFlushStatus fdrLoggingFlush() XRAY_NEVER_INSTRUMENT {
       TSCSupported ? getTSCFrequency() : __xray::NanosecondsPerSecond;
 
   XRayFileHeader Header;
-  Header.Version = 1;
+
+  // Version 2 of the log writes the extents of the buffer, instead of relying
+  // on an end-of-buffer record.
+  Header.Version = 2;
   Header.Type = FileTypes::FDR_LOG;
   Header.CycleFrequency = CycleFrequency;
+
   // FIXME: Actually check whether we have 'constant_tsc' and 'nonstop_tsc'
   // before setting the values in the header.
   Header.ConstantTSC = 1;
   Header.NonstopTSC = 1;
-  Header.FdrData = FdrAdditionalHeaderData{LocalBQ->ConfiguredBufferSize()};
+  Header.FdrData = FdrAdditionalHeaderData{BQ->ConfiguredBufferSize()};
   retryingWriteAll(Fd, reinterpret_cast<char *>(&Header),
                    reinterpret_cast<char *>(&Header) + sizeof(Header));
 
-  LocalBQ->apply([&](const BufferQueue::Buffer &B) {
-    uint64_t BufferSize = B.Size;
-    if (BufferSize > 0) {
+  BQ->apply([&](const BufferQueue::Buffer &B) {
+    // Starting at version 2 of the FDR logging implementation, we only write
+    // the records identified by the extents of the buffer. We use the Extents
+    // from the Buffer and write that out as the first record in the buffer.
+    // We still use a Metadata record, but fill in the extents instead for the
+    // data.
+    MetadataRecord ExtentsRecord;
+    auto BufferExtents = __sanitizer::atomic_load(
+        &B.Extents->Size, __sanitizer::memory_order_acquire);
+    assert(BufferExtents <= B.Size);
+    ExtentsRecord.Type = uint8_t(RecordType::Metadata);
+    ExtentsRecord.RecordKind =
+        uint8_t(MetadataRecord::RecordKinds::BufferExtents);
+    std::memcpy(ExtentsRecord.Data, &BufferExtents, sizeof(BufferExtents));
+    if (BufferExtents > 0) {
+      retryingWriteAll(Fd, reinterpret_cast<char *>(&ExtentsRecord),
+                       reinterpret_cast<char *>(&ExtentsRecord) +
+                           sizeof(MetadataRecord));
       retryingWriteAll(Fd, reinterpret_cast<char *>(B.Buffer),
-                       reinterpret_cast<char *>(B.Buffer) + B.Size);
+                       reinterpret_cast<char *>(B.Buffer) + BufferExtents);
     }
   });
+
   __sanitizer::atomic_store(&LogFlushStatus,
                             XRayLogFlushStatus::XRAY_LOG_FLUSHED,
                             __sanitizer::memory_order_release);
@@ -124,8 +152,11 @@ XRayLogInitStatus fdrLoggingFinalize() XRAY_NEVER_INSTRUMENT {
   if (!__sanitizer::atomic_compare_exchange_strong(
           &LoggingStatus, &CurrentStatus,
           XRayLogInitStatus::XRAY_LOG_FINALIZING,
-          __sanitizer::memory_order_release))
+          __sanitizer::memory_order_release)) {
+    if (__sanitizer::Verbosity())
+      Report("Cannot finalize log, implementation not initialized.\n");
     return static_cast<XRayLogInitStatus>(CurrentStatus);
+  }
 
   // Do special things to make the log finalize itself, and not allow any more
   // operations to be performed until re-initialized.
@@ -146,7 +177,8 @@ XRayLogInitStatus fdrLoggingReset() XRAY_NEVER_INSTRUMENT {
     return static_cast<XRayLogInitStatus>(CurrentStatus);
 
   // Release the in-memory buffer queue.
-  BQ.reset();
+  delete BQ;
+  BQ = nullptr;
 
   // Spin until the flushing status is flushed.
   s32 CurrentFlushingStatus = XRayLogFlushStatus::XRAY_LOG_FLUSHED;
@@ -163,19 +195,22 @@ XRayLogInitStatus fdrLoggingReset() XRAY_NEVER_INSTRUMENT {
   return XRayLogInitStatus::XRAY_LOG_UNINITIALIZED;
 }
 
-static std::tuple<uint64_t, unsigned char>
-getTimestamp() XRAY_NEVER_INSTRUMENT {
+struct TSCAndCPU {
+  uint64_t TSC = 0;
+  unsigned char CPU = 0;
+};
+
+static TSCAndCPU getTimestamp() XRAY_NEVER_INSTRUMENT {
   // We want to get the TSC as early as possible, so that we can check whether
   // we've seen this CPU before. We also do it before we load anything else, to
   // allow for forward progress with the scheduling.
-  unsigned char CPU;
-  uint64_t TSC;
+  TSCAndCPU Result;
 
   // Test once for required CPU features
   static bool TSCSupported = probeRequiredCPUFeatures();
 
   if (TSCSupported) {
-    TSC = __xray::readTSC(CPU);
+    Result.TSC = __xray::readTSC(Result.CPU);
   } else {
     // FIXME: This code needs refactoring as it appears in multiple locations
     timespec TS;
@@ -184,32 +219,35 @@ getTimestamp() XRAY_NEVER_INSTRUMENT {
       Report("clock_gettime(2) return %d, errno=%d", result, int(errno));
       TS = {0, 0};
     }
-    CPU = 0;
-    TSC = TS.tv_sec * __xray::NanosecondsPerSecond + TS.tv_nsec;
+    Result.CPU = 0;
+    Result.TSC = TS.tv_sec * __xray::NanosecondsPerSecond + TS.tv_nsec;
   }
-  return std::make_tuple(TSC, CPU);
+  return Result;
 }
 
 void fdrLoggingHandleArg0(int32_t FuncId,
                           XRayEntryType Entry) XRAY_NEVER_INSTRUMENT {
-  auto TSC_CPU = getTimestamp();
-  __xray_fdr_internal::processFunctionHook(FuncId, Entry, std::get<0>(TSC_CPU),
-                                           std::get<1>(TSC_CPU), clock_gettime,
-                                           LoggingStatus, BQ);
+  auto TC = getTimestamp();
+  __xray_fdr_internal::processFunctionHook(FuncId, Entry, TC.TSC, TC.CPU, 0,
+                                           clock_gettime, BQ);
+}
+
+void fdrLoggingHandleArg1(int32_t FuncId, XRayEntryType Entry,
+                          uint64_t Arg) XRAY_NEVER_INSTRUMENT {
+  auto TC = getTimestamp();
+  __xray_fdr_internal::processFunctionHook(FuncId, Entry, TC.TSC, TC.CPU, Arg,
+                                           clock_gettime, BQ);
 }
 
 void fdrLoggingHandleCustomEvent(void *Event,
                                  std::size_t EventSize) XRAY_NEVER_INSTRUMENT {
   using namespace __xray_fdr_internal;
-  auto TSC_CPU = getTimestamp();
-  auto &TSC = std::get<0>(TSC_CPU);
-  auto &CPU = std::get<1>(TSC_CPU);
-  thread_local bool Running = false;
+  auto TC = getTimestamp();
+  auto &TSC = TC.TSC;
+  auto &CPU = TC.CPU;
   RecursionGuard Guard{Running};
-  if (!Guard) {
-    assert(Running && "RecursionGuard is buggy!");
+  if (!Guard)
     return;
-  }
   if (EventSize > std::numeric_limits<int32_t>::max()) {
     using Empty = struct {};
     static Empty Once = [&] {
@@ -220,15 +258,16 @@ void fdrLoggingHandleCustomEvent(void *Event,
     (void)Once;
   }
   int32_t ReducedEventSize = static_cast<int32_t>(EventSize);
-  if (!isLogInitializedAndReady(LocalBQ, TSC, CPU, clock_gettime))
+  auto &TLD = getThreadLocalData();
+  if (!isLogInitializedAndReady(TLD.BQ, TSC, CPU, clock_gettime))
     return;
 
   // Here we need to prepare the log to handle:
   //   - The metadata record we're going to write. (16 bytes)
   //   - The additional data we're going to write. Currently, that's the size of
   //   the event we're going to dump into the log as free-form bytes.
-  if (!prepareBuffer(clock_gettime, MetadataRecSize + EventSize)) {
-    LocalBQ = nullptr;
+  if (!prepareBuffer(TSC, CPU, clock_gettime, MetadataRecSize + EventSize)) {
+    TLD.BQ = nullptr;
     return;
   }
 
@@ -240,27 +279,35 @@ void fdrLoggingHandleCustomEvent(void *Event,
   CustomEvent.Type = uint8_t(RecordType::Metadata);
   CustomEvent.RecordKind =
       uint8_t(MetadataRecord::RecordKinds::CustomEventMarker);
-  constexpr auto TSCSize = sizeof(std::get<0>(TSC_CPU));
+  constexpr auto TSCSize = sizeof(TC.TSC);
   std::memcpy(&CustomEvent.Data, &ReducedEventSize, sizeof(int32_t));
   std::memcpy(&CustomEvent.Data[sizeof(int32_t)], &TSC, TSCSize);
-  std::memcpy(RecordPtr, &CustomEvent, sizeof(CustomEvent));
-  RecordPtr += sizeof(CustomEvent);
-  std::memcpy(RecordPtr, Event, ReducedEventSize);
+  std::memcpy(TLD.RecordPtr, &CustomEvent, sizeof(CustomEvent));
+  TLD.RecordPtr += sizeof(CustomEvent);
+  std::memcpy(TLD.RecordPtr, Event, ReducedEventSize);
+  incrementExtents(MetadataRecSize + EventSize);
   endBufferIfFull();
 }
 
 XRayLogInitStatus fdrLoggingInit(std::size_t BufferSize, std::size_t BufferMax,
                                  void *Options,
                                  size_t OptionsSize) XRAY_NEVER_INSTRUMENT {
-  if (OptionsSize != sizeof(FDRLoggingOptions))
+  if (OptionsSize != sizeof(FDRLoggingOptions)) {
+    if (__sanitizer::Verbosity())
+      Report("Cannot initialize FDR logging; wrong size for options: %d\n",
+             OptionsSize);
     return static_cast<XRayLogInitStatus>(__sanitizer::atomic_load(
         &LoggingStatus, __sanitizer::memory_order_acquire));
+  }
   s32 CurrentStatus = XRayLogInitStatus::XRAY_LOG_UNINITIALIZED;
   if (!__sanitizer::atomic_compare_exchange_strong(
           &LoggingStatus, &CurrentStatus,
           XRayLogInitStatus::XRAY_LOG_INITIALIZING,
-          __sanitizer::memory_order_release))
+          __sanitizer::memory_order_release)) {
+    if (__sanitizer::Verbosity())
+      Report("Cannot initialize already initialized implementation.\n");
     return static_cast<XRayLogInitStatus>(CurrentStatus);
+  }
 
   {
     __sanitizer::SpinMutexLock Guard(&FDROptionsMutex);
@@ -268,12 +315,40 @@ XRayLogInitStatus fdrLoggingInit(std::size_t BufferSize, std::size_t BufferMax,
   }
 
   bool Success = false;
-  BQ = std::make_shared<BufferQueue>(BufferSize, BufferMax, Success);
+
+  if (BQ != nullptr) {
+    delete BQ;
+    BQ = nullptr;
+  }
+
+  if (BQ == nullptr)
+    BQ = new BufferQueue(BufferSize, BufferMax, Success);
+
   if (!Success) {
     Report("BufferQueue init failed.\n");
+    if (BQ != nullptr) {
+      delete BQ;
+      BQ = nullptr;
+    }
     return XRayLogInitStatus::XRAY_LOG_UNINITIALIZED;
   }
 
+  static bool UNUSED Once = [] {
+    pthread_key_create(&__xray_fdr_internal::Key, +[](void *) {
+      auto &TLD = __xray_fdr_internal::getThreadLocalData();
+      if (TLD.BQ == nullptr)
+        return;
+      auto EC = TLD.BQ->releaseBuffer(TLD.Buffer);
+      if (EC != BufferQueue::ErrorCode::Ok)
+        Report("At thread exit, failed to release buffer at %p; error=%s\n",
+               TLD.Buffer.Buffer, BufferQueue::getErrorString(EC));
+    });
+    return false;
+  }();
+
+  // Arg1 handler should go in first to avoid concurrent code accidentally
+  // falling back to arg0 when it should have ran arg1.
+  __xray_set_handler_arg1(fdrLoggingHandleArg1);
   // Install the actual handleArg0 handler after initialising the buffers.
   __xray_set_handler(fdrLoggingHandleArg0);
   __xray_set_customevent_handler(fdrLoggingHandleCustomEvent);
@@ -281,20 +356,31 @@ XRayLogInitStatus fdrLoggingInit(std::size_t BufferSize, std::size_t BufferMax,
   __sanitizer::atomic_store(&LoggingStatus,
                             XRayLogInitStatus::XRAY_LOG_INITIALIZED,
                             __sanitizer::memory_order_release);
-  Report("XRay FDR init successful.\n");
+
+  if (__sanitizer::Verbosity())
+    Report("XRay FDR init successful.\n");
   return XRayLogInitStatus::XRAY_LOG_INITIALIZED;
 }
 
-} // namespace __xray
-
-static auto UNUSED Unused = [] {
+bool fdrLogDynamicInitializer() XRAY_NEVER_INSTRUMENT {
   using namespace __xray;
-  if (flags()->xray_fdr_log) {
-    XRayLogImpl Impl{
-        fdrLoggingInit, fdrLoggingFinalize, fdrLoggingHandleArg0,
-        fdrLoggingFlush,
-    };
+  XRayLogImpl Impl{
+      fdrLoggingInit,
+      fdrLoggingFinalize,
+      fdrLoggingHandleArg0,
+      fdrLoggingFlush,
+  };
+  auto RegistrationResult = __xray_log_register_mode("xray-fdr", Impl);
+  if (RegistrationResult != XRayLogRegisterStatus::XRAY_REGISTRATION_OK &&
+      __sanitizer::Verbosity())
+    Report("Cannot register XRay FDR mode to 'xray-fdr'; error = %d\n",
+           RegistrationResult);
+  if (flags()->xray_fdr_log ||
+      !__sanitizer::internal_strcmp(flags()->xray_mode, "xray-fdr"))
     __xray_set_log_impl(Impl);
-  }
   return true;
-}();
+}
+
+} // namespace __xray
+
+static auto UNUSED Unused = __xray::fdrLogDynamicInitializer();
diff --git a/contrib/compiler-rt/lib/xray/xray_fdr_logging.h b/contrib/compiler-rt/lib/xray/xray_fdr_logging.h
index 426b54dc7884..1639d550a44c 100644
--- a/contrib/compiler-rt/lib/xray/xray_fdr_logging.h
+++ b/contrib/compiler-rt/lib/xray/xray_fdr_logging.h
@@ -30,6 +30,7 @@ XRayLogInitStatus fdrLoggingInit(size_t BufferSize, size_t BufferMax,
                                  void *Options, size_t OptionsSize);
 XRayLogInitStatus fdrLoggingFinalize();
 void fdrLoggingHandleArg0(int32_t FuncId, XRayEntryType Entry);
+void fdrLoggingHandleArg1(int32_t FuncId, XRayEntryType Entry, uint64_t Arg1);
 XRayLogFlushStatus fdrLoggingFlush();
 XRayLogInitStatus fdrLoggingReset();
 
diff --git a/contrib/compiler-rt/lib/xray/xray_fdr_logging_impl.h b/contrib/compiler-rt/lib/xray/xray_fdr_logging_impl.h
index 4a1d80fd0eba..59eab55b2573 100644
--- a/contrib/compiler-rt/lib/xray/xray_fdr_logging_impl.h
+++ b/contrib/compiler-rt/lib/xray/xray_fdr_logging_impl.h
@@ -18,13 +18,13 @@
 #define XRAY_XRAY_FDR_LOGGING_IMPL_H
 
 #include <cassert>
-#include <cstdint>
+#include <cstddef>
 #include <cstring>
 #include <limits>
-#include <memory>
-#include <string>
+#include <pthread.h>
 #include <sys/syscall.h>
 #include <time.h>
+#include <type_traits>
 #include <unistd.h>
 
 #include "sanitizer_common/sanitizer_common.h"
@@ -52,132 +52,120 @@ __sanitizer::atomic_sint32_t LoggingStatus = {
 /// cooperation with xray_fdr_logging class, so be careful and think twice.
 namespace __xray_fdr_internal {
 
-/// Writes the new buffer record and wallclock time that begin a buffer for a
-/// thread to MemPtr and increments MemPtr. Bypasses the thread local state
-/// machine and writes directly to memory without checks.
-static void writeNewBufferPreamble(pid_t Tid, timespec TS, char *&MemPtr);
+/// Writes the new buffer record and wallclock time that begin a buffer for the
+/// current thread.
+static void writeNewBufferPreamble(pid_t Tid, timespec TS);
 
-/// Write a metadata record to switch to a new CPU to MemPtr and increments
-/// MemPtr. Bypasses the thread local state machine and writes directly to
-/// memory without checks.
-static void writeNewCPUIdMetadata(uint16_t CPU, uint64_t TSC, char *&MemPtr);
-
-/// Writes an EOB metadata record to MemPtr and increments MemPtr. Bypasses the
-/// thread local state machine and writes directly to memory without checks.
-static void writeEOBMetadata(char *&MemPtr);
-
-/// Writes a TSC Wrap metadata record to MemPtr and increments MemPtr. Bypasses
-/// the thread local state machine and directly writes to memory without checks.
-static void writeTSCWrapMetadata(uint64_t TSC, char *&MemPtr);
-
-/// Writes a Function Record to MemPtr and increments MemPtr. Bypasses the
-/// thread local state machine and writes the function record directly to
-/// memory.
+/// Writes a Function Record to the buffer associated with the current thread.
 static void writeFunctionRecord(int FuncId, uint32_t TSCDelta,
-                                XRayEntryType EntryType, char *&MemPtr);
+                                XRayEntryType EntryType);
 
 /// Sets up a new buffer in thread_local storage and writes a preamble. The
 /// wall_clock_reader function is used to populate the WallTimeRecord entry.
 static void setupNewBuffer(int (*wall_clock_reader)(clockid_t,
                                                     struct timespec *));
 
-/// Called to record CPU time for a new CPU within the current thread.
-static void writeNewCPUIdMetadata(uint16_t CPU, uint64_t TSC);
-
-/// Called to close the buffer when the thread exhausts the buffer or when the
-/// thread exits (via a thread local variable destructor).
-static void writeEOBMetadata();
-
 /// TSC Wrap records are written when a TSC delta encoding scheme overflows.
 static void writeTSCWrapMetadata(uint64_t TSC);
 
-/// Here's where the meat of the processing happens. The writer captures
-/// function entry, exit and tail exit points with a time and will create
-/// TSCWrap, NewCPUId and Function records as necessary. The writer might
-/// walk backward through its buffer and erase trivial functions to avoid
-/// polluting the log and may use the buffer queue to obtain or release a
-/// buffer.
-static void processFunctionHook(int32_t FuncId, XRayEntryType Entry,
-                                uint64_t TSC, unsigned char CPU,
-                                int (*wall_clock_reader)(clockid_t,
-                                                         struct timespec *),
-                                __sanitizer::atomic_sint32_t &LoggingStatus,
-                                const std::shared_ptr<BufferQueue> &BQ);
+// Group together thread-local-data in a struct, then hide it behind a function
+// call so that it can be initialized on first use instead of as a global. We
+// force the alignment to 64-bytes for x86 cache line alignment, as this
+// structure is used in the hot path of implementation.
+struct alignas(64) ThreadLocalData {
+  BufferQueue::Buffer Buffer;
+  char *RecordPtr = nullptr;
+  // The number of FunctionEntry records immediately preceding RecordPtr.
+  uint8_t NumConsecutiveFnEnters = 0;
 
-//-----------------------------------------------------------------------------|
-// The rest of the file is implementation.                                     |
-//-----------------------------------------------------------------------------|
-// Functions are implemented in the header for inlining since we don't want    |
-// to grow the stack when we've hijacked the binary for logging.               |
-//-----------------------------------------------------------------------------|
+  // The number of adjacent, consecutive pairs of FunctionEntry, Tail Exit
+  // records preceding RecordPtr.
+  uint8_t NumTailCalls = 0;
 
-namespace {
+  // We use a thread_local variable to keep track of which CPUs we've already
+  // run, and the TSC times for these CPUs. This allows us to stop repeating the
+  // CPU field in the function records.
+  //
+  // We assume that we'll support only 65536 CPUs for x86_64.
+  uint16_t CurrentCPU = std::numeric_limits<uint16_t>::max();
+  uint64_t LastTSC = 0;
+  uint64_t LastFunctionEntryTSC = 0;
 
-thread_local BufferQueue::Buffer Buffer;
-thread_local char *RecordPtr = nullptr;
+  // Make sure a thread that's ever called handleArg0 has a thread-local
+  // live reference to the buffer queue for this particular instance of
+  // FDRLogging, and that we're going to clean it up when the thread exits.
+  BufferQueue *BQ = nullptr;
+};
 
-// The number of FunctionEntry records immediately preceding RecordPtr.
-thread_local uint8_t NumConsecutiveFnEnters = 0;
+static_assert(std::is_trivially_destructible<ThreadLocalData>::value,
+              "ThreadLocalData must be trivially destructible");
 
-// The number of adjacent, consecutive pairs of FunctionEntry, Tail Exit
-// records preceding RecordPtr.
-thread_local uint8_t NumTailCalls = 0;
+static constexpr auto MetadataRecSize = sizeof(MetadataRecord);
+static constexpr auto FunctionRecSize = sizeof(FunctionRecord);
 
-constexpr auto MetadataRecSize = sizeof(MetadataRecord);
-constexpr auto FunctionRecSize = sizeof(FunctionRecord);
+// Use a global pthread key to identify thread-local data for logging.
+static pthread_key_t Key;
 
-// We use a thread_local variable to keep track of which CPUs we've already
-// run, and the TSC times for these CPUs. This allows us to stop repeating the
-// CPU field in the function records.
+// This function will initialize the thread-local data structure used by the FDR
+// logging implementation and return a reference to it. The implementation
+// details require a bit of care to maintain.
 //
-// We assume that we'll support only 65536 CPUs for x86_64.
-thread_local uint16_t CurrentCPU = std::numeric_limits<uint16_t>::max();
-thread_local uint64_t LastTSC = 0;
-thread_local uint64_t LastFunctionEntryTSC = 0;
-
-class ThreadExitBufferCleanup {
-  std::shared_ptr<BufferQueue> &Buffers;
-  BufferQueue::Buffer &Buffer;
-
-public:
-  explicit ThreadExitBufferCleanup(std::shared_ptr<BufferQueue> &BQ,
-                                   BufferQueue::Buffer &Buffer)
-      XRAY_NEVER_INSTRUMENT : Buffers(BQ),
-                              Buffer(Buffer) {}
-
-  ~ThreadExitBufferCleanup() noexcept XRAY_NEVER_INSTRUMENT {
-    if (RecordPtr == nullptr)
-      return;
+// First, some requirements on the implementation in general:
+//
+//   - XRay handlers should not call any memory allocation routines that may
+//     delegate to an instrumented implementation. This means functions like
+//     malloc() and free() should not be called while instrumenting.
+//
+//   - We would like to use some thread-local data initialized on first-use of
+//     the XRay instrumentation. These allow us to implement unsynchronized
+//     routines that access resources associated with the thread.
+//
+// The implementation here uses a few mechanisms that allow us to provide both
+// the requirements listed above. We do this by:
+//
+//   1. Using a thread-local aligned storage buffer for representing the
+//      ThreadLocalData struct. This data will be uninitialized memory by
+//      design.
+//
+//   2. Not requiring a thread exit handler/implementation, keeping the
+//      thread-local as purely a collection of references/data that do not
+//      require cleanup.
+//
+// We're doing this to avoid using a `thread_local` object that has a
+// non-trivial destructor, because the C++ runtime might call std::malloc(...)
+// to register calls to destructors. Deadlocks may arise when, for example, an
+// externally provided malloc implementation is XRay instrumented, and
+// initializing the thread-locals involves calling into malloc. A malloc
+// implementation that does global synchronization might be holding a lock for a
+// critical section, calling a function that might be XRay instrumented (and
+// thus in turn calling into malloc by virtue of registration of the
+// thread_local's destructor).
+static ThreadLocalData &getThreadLocalData() {
+  static_assert(alignof(ThreadLocalData) >= 64,
+                "ThreadLocalData must be cache line aligned.");
+  thread_local ThreadLocalData TLD;
+  thread_local bool UNUSED ThreadOnce = [] {
+    pthread_setspecific(Key, &TLD);
+    return false;
+  }();
+  return TLD;
+}
 
-    // We make sure that upon exit, a thread will write out the EOB
-    // MetadataRecord in the thread-local log, and also release the buffer to
-    // the queue.
-    assert((RecordPtr + MetadataRecSize) - static_cast<char *>(Buffer.Buffer) >=
-           static_cast<ptrdiff_t>(MetadataRecSize));
-    if (Buffers) {
-      writeEOBMetadata();
-      auto EC = Buffers->releaseBuffer(Buffer);
-      if (EC != BufferQueue::ErrorCode::Ok)
-        Report("Failed to release buffer at %p; error=%s\n", Buffer.Buffer,
-               BufferQueue::getErrorString(EC));
-      Buffers = nullptr;
-      return;
-    }
-  }
-};
+//-----------------------------------------------------------------------------|
+// The rest of the file is implementation.                                     |
+//-----------------------------------------------------------------------------|
+// Functions are implemented in the header for inlining since we don't want    |
+// to grow the stack when we've hijacked the binary for logging.               |
+//-----------------------------------------------------------------------------|
 
-// Make sure a thread that's ever called handleArg0 has a thread-local
-// live reference to the buffer queue for this particular instance of
-// FDRLogging, and that we're going to clean it up when the thread exits.
-thread_local std::shared_ptr<BufferQueue> LocalBQ = nullptr;
-thread_local ThreadExitBufferCleanup Cleanup(LocalBQ, Buffer);
+namespace {
 
 class RecursionGuard {
-  bool &Running;
+  volatile bool &Running;
   const bool Valid;
 
 public:
-  explicit RecursionGuard(bool &R) : Running(R), Valid(!R) {
+  explicit RecursionGuard(volatile bool &R) : Running(R), Valid(!R) {
     if (Valid)
       Running = true;
   }
@@ -195,34 +183,29 @@ public:
   }
 };
 
-inline bool loggingInitialized(
-    const __sanitizer::atomic_sint32_t &LoggingStatus) XRAY_NEVER_INSTRUMENT {
-  return __sanitizer::atomic_load(&LoggingStatus,
-                                  __sanitizer::memory_order_acquire) ==
-         XRayLogInitStatus::XRAY_LOG_INITIALIZED;
-}
-
 } // namespace
 
-inline void writeNewBufferPreamble(pid_t Tid, timespec TS,
-                                   char *&MemPtr) XRAY_NEVER_INSTRUMENT {
+static void writeNewBufferPreamble(pid_t Tid,
+                                   timespec TS) XRAY_NEVER_INSTRUMENT {
   static constexpr int InitRecordsCount = 2;
-  std::aligned_storage<sizeof(MetadataRecord)>::type Records[InitRecordsCount];
+  auto &TLD = getThreadLocalData();
+  MetadataRecord Metadata[InitRecordsCount];
   {
     // Write out a MetadataRecord to signify that this is the start of a new
     // buffer, associated with a particular thread, with a new CPU.  For the
     // data, we have 15 bytes to squeeze as much information as we can.  At this
     // point we only write down the following bytes:
     //   - Thread ID (pid_t, 4 bytes)
-    auto &NewBuffer = *reinterpret_cast<MetadataRecord *>(&Records[0]);
+    auto &NewBuffer = Metadata[0];
     NewBuffer.Type = uint8_t(RecordType::Metadata);
     NewBuffer.RecordKind = uint8_t(MetadataRecord::RecordKinds::NewBuffer);
     std::memcpy(&NewBuffer.Data, &Tid, sizeof(pid_t));
   }
+
   // Also write the WalltimeMarker record.
   {
     static_assert(sizeof(time_t) <= 8, "time_t needs to be at most 8 bytes");
-    auto &WalltimeMarker = *reinterpret_cast<MetadataRecord *>(&Records[1]);
+    auto &WalltimeMarker = Metadata[1];
     WalltimeMarker.Type = uint8_t(RecordType::Metadata);
     WalltimeMarker.RecordKind =
         uint8_t(MetadataRecord::RecordKinds::WalltimeMarker);
@@ -235,26 +218,48 @@ inline void writeNewBufferPreamble(pid_t Tid, timespec TS,
     std::memcpy(WalltimeMarker.Data, &Seconds, sizeof(Seconds));
     std::memcpy(WalltimeMarker.Data + sizeof(Seconds), &Micros, sizeof(Micros));
   }
-  std::memcpy(MemPtr, Records, sizeof(MetadataRecord) * InitRecordsCount);
-  MemPtr += sizeof(MetadataRecord) * InitRecordsCount;
-  NumConsecutiveFnEnters = 0;
-  NumTailCalls = 0;
+
+  TLD.NumConsecutiveFnEnters = 0;
+  TLD.NumTailCalls = 0;
+  if (TLD.BQ == nullptr || TLD.BQ->finalizing())
+    return;
+  std::memcpy(TLD.RecordPtr, Metadata, sizeof(Metadata));
+  TLD.RecordPtr += sizeof(Metadata);
+  // Since we write out the extents as the first metadata record of the
+  // buffer, we need to write out the extents including the extents record.
+  __sanitizer::atomic_store(&TLD.Buffer.Extents->Size, sizeof(Metadata),
+                            __sanitizer::memory_order_release);
 }
 
 inline void setupNewBuffer(int (*wall_clock_reader)(
     clockid_t, struct timespec *)) XRAY_NEVER_INSTRUMENT {
-  RecordPtr = static_cast<char *>(Buffer.Buffer);
+  auto &TLD = getThreadLocalData();
+  auto &B = TLD.Buffer;
+  TLD.RecordPtr = static_cast<char *>(B.Buffer);
   pid_t Tid = syscall(SYS_gettid);
   timespec TS{0, 0};
   // This is typically clock_gettime, but callers have injection ability.
   wall_clock_reader(CLOCK_MONOTONIC, &TS);
-  writeNewBufferPreamble(Tid, TS, RecordPtr);
-  NumConsecutiveFnEnters = 0;
-  NumTailCalls = 0;
+  writeNewBufferPreamble(Tid, TS);
+  TLD.NumConsecutiveFnEnters = 0;
+  TLD.NumTailCalls = 0;
+}
+
+static void incrementExtents(size_t Add) {
+  auto &TLD = getThreadLocalData();
+  __sanitizer::atomic_fetch_add(&TLD.Buffer.Extents->Size, Add,
+                                __sanitizer::memory_order_acq_rel);
 }
 
-inline void writeNewCPUIdMetadata(uint16_t CPU, uint64_t TSC,
-                                  char *&MemPtr) XRAY_NEVER_INSTRUMENT {
+static void decrementExtents(size_t Subtract) {
+  auto &TLD = getThreadLocalData();
+  __sanitizer::atomic_fetch_sub(&TLD.Buffer.Extents->Size, Subtract,
+                                __sanitizer::memory_order_acq_rel);
+}
+
+inline void writeNewCPUIdMetadata(uint16_t CPU,
+                                  uint64_t TSC) XRAY_NEVER_INSTRUMENT {
+  auto &TLD = getThreadLocalData();
   MetadataRecord NewCPUId;
   NewCPUId.Type = uint8_t(RecordType::Metadata);
   NewCPUId.RecordKind = uint8_t(MetadataRecord::RecordKinds::NewCPUId);
@@ -265,34 +270,15 @@ inline void writeNewCPUIdMetadata(uint16_t CPU, uint64_t TSC,
   // Total = 10 bytes.
   std::memcpy(&NewCPUId.Data, &CPU, sizeof(CPU));
   std::memcpy(&NewCPUId.Data[sizeof(CPU)], &TSC, sizeof(TSC));
-  std::memcpy(MemPtr, &NewCPUId, sizeof(MetadataRecord));
-  MemPtr += sizeof(MetadataRecord);
-  NumConsecutiveFnEnters = 0;
-  NumTailCalls = 0;
-}
-
-inline void writeNewCPUIdMetadata(uint16_t CPU,
-                                  uint64_t TSC) XRAY_NEVER_INSTRUMENT {
-  writeNewCPUIdMetadata(CPU, TSC, RecordPtr);
-}
-
-inline void writeEOBMetadata(char *&MemPtr) XRAY_NEVER_INSTRUMENT {
-  MetadataRecord EOBMeta;
-  EOBMeta.Type = uint8_t(RecordType::Metadata);
-  EOBMeta.RecordKind = uint8_t(MetadataRecord::RecordKinds::EndOfBuffer);
-  // For now we don't write any bytes into the Data field.
-  std::memcpy(MemPtr, &EOBMeta, sizeof(MetadataRecord));
-  MemPtr += sizeof(MetadataRecord);
-  NumConsecutiveFnEnters = 0;
-  NumTailCalls = 0;
+  std::memcpy(TLD.RecordPtr, &NewCPUId, sizeof(MetadataRecord));
+  TLD.RecordPtr += sizeof(MetadataRecord);
+  TLD.NumConsecutiveFnEnters = 0;
+  TLD.NumTailCalls = 0;
+  incrementExtents(sizeof(MetadataRecord));
 }
 
-inline void writeEOBMetadata() XRAY_NEVER_INSTRUMENT {
-  writeEOBMetadata(RecordPtr);
-}
-
-inline void writeTSCWrapMetadata(uint64_t TSC,
-                                 char *&MemPtr) XRAY_NEVER_INSTRUMENT {
+inline void writeTSCWrapMetadata(uint64_t TSC) XRAY_NEVER_INSTRUMENT {
+  auto &TLD = getThreadLocalData();
   MetadataRecord TSCWrap;
   TSCWrap.Type = uint8_t(RecordType::Metadata);
   TSCWrap.RecordKind = uint8_t(MetadataRecord::RecordKinds::TSCWrap);
@@ -301,58 +287,67 @@ inline void writeTSCWrapMetadata(uint64_t TSC,
   //   - Full TSC (uint64_t, 8 bytes)
   // Total = 8 bytes.
   std::memcpy(&TSCWrap.Data, &TSC, sizeof(TSC));
-  std::memcpy(MemPtr, &TSCWrap, sizeof(MetadataRecord));
-  MemPtr += sizeof(MetadataRecord);
-  NumConsecutiveFnEnters = 0;
-  NumTailCalls = 0;
+  std::memcpy(TLD.RecordPtr, &TSCWrap, sizeof(MetadataRecord));
+  TLD.RecordPtr += sizeof(MetadataRecord);
+  TLD.NumConsecutiveFnEnters = 0;
+  TLD.NumTailCalls = 0;
+  incrementExtents(sizeof(MetadataRecord));
 }
 
-inline void writeTSCWrapMetadata(uint64_t TSC) XRAY_NEVER_INSTRUMENT {
-  writeTSCWrapMetadata(TSC, RecordPtr);
+// Call Argument metadata records store the arguments to a function in the
+// order of their appearance; holes are not supported by the buffer format.
+static inline void writeCallArgumentMetadata(uint64_t A) XRAY_NEVER_INSTRUMENT {
+  auto &TLD = getThreadLocalData();
+  MetadataRecord CallArg;
+  CallArg.Type = uint8_t(RecordType::Metadata);
+  CallArg.RecordKind = uint8_t(MetadataRecord::RecordKinds::CallArgument);
+
+  std::memcpy(CallArg.Data, &A, sizeof(A));
+  std::memcpy(TLD.RecordPtr, &CallArg, sizeof(MetadataRecord));
+  TLD.RecordPtr += sizeof(MetadataRecord);
+  incrementExtents(sizeof(MetadataRecord));
 }
 
-inline void writeFunctionRecord(int FuncId, uint32_t TSCDelta,
-                                XRayEntryType EntryType,
-                                char *&MemPtr) XRAY_NEVER_INSTRUMENT {
-  std::aligned_storage<sizeof(FunctionRecord), alignof(FunctionRecord)>::type
-      AlignedFuncRecordBuffer;
-  auto &FuncRecord =
-      *reinterpret_cast<FunctionRecord *>(&AlignedFuncRecordBuffer);
+static inline void
+writeFunctionRecord(int FuncId, uint32_t TSCDelta,
+                    XRayEntryType EntryType) XRAY_NEVER_INSTRUMENT {
+  FunctionRecord FuncRecord;
   FuncRecord.Type = uint8_t(RecordType::Function);
   // Only take 28 bits of the function id.
   FuncRecord.FuncId = FuncId & ~(0x0F << 28);
   FuncRecord.TSCDelta = TSCDelta;
 
+  auto &TLD = getThreadLocalData();
   switch (EntryType) {
   case XRayEntryType::ENTRY:
-    ++NumConsecutiveFnEnters;
+    ++TLD.NumConsecutiveFnEnters;
     FuncRecord.RecordKind = uint8_t(FunctionRecord::RecordKinds::FunctionEnter);
     break;
   case XRayEntryType::LOG_ARGS_ENTRY:
     // We should not rewind functions with logged args.
-    NumConsecutiveFnEnters = 0;
-    NumTailCalls = 0;
+    TLD.NumConsecutiveFnEnters = 0;
+    TLD.NumTailCalls = 0;
     FuncRecord.RecordKind = uint8_t(FunctionRecord::RecordKinds::FunctionEnter);
     break;
   case XRayEntryType::EXIT:
     // If we've decided to log the function exit, we will never erase the log
     // before it.
-    NumConsecutiveFnEnters = 0;
-    NumTailCalls = 0;
+    TLD.NumConsecutiveFnEnters = 0;
+    TLD.NumTailCalls = 0;
     FuncRecord.RecordKind = uint8_t(FunctionRecord::RecordKinds::FunctionExit);
     break;
   case XRayEntryType::TAIL:
     // If we just entered the function we're tail exiting from or erased every
     // invocation since then, this function entry tail pair is a candidate to
     // be erased when the child function exits.
-    if (NumConsecutiveFnEnters > 0) {
-      ++NumTailCalls;
-      NumConsecutiveFnEnters = 0;
+    if (TLD.NumConsecutiveFnEnters > 0) {
+      ++TLD.NumTailCalls;
+      TLD.NumConsecutiveFnEnters = 0;
     } else {
       // We will never be able to erase this tail call since we have logged
       // something in between the function entry and tail exit.
-      NumTailCalls = 0;
-      NumConsecutiveFnEnters = 0;
+      TLD.NumTailCalls = 0;
+      TLD.NumConsecutiveFnEnters = 0;
     }
     FuncRecord.RecordKind =
         uint8_t(FunctionRecord::RecordKinds::FunctionTailExit);
@@ -370,8 +365,9 @@ inline void writeFunctionRecord(int FuncId, uint32_t TSCDelta,
   }
   }
 
-  std::memcpy(MemPtr, &AlignedFuncRecordBuffer, sizeof(FunctionRecord));
-  MemPtr += sizeof(FunctionRecord);
+  std::memcpy(TLD.RecordPtr, &FuncRecord, sizeof(FunctionRecord));
+  TLD.RecordPtr += sizeof(FunctionRecord);
+  incrementExtents(sizeof(FunctionRecord));
 }
 
 static uint64_t thresholdTicks() {
@@ -387,23 +383,21 @@ static uint64_t thresholdTicks() {
 // "Function Entry" record and any "Tail Call Exit" records after that.
 static void rewindRecentCall(uint64_t TSC, uint64_t &LastTSC,
                              uint64_t &LastFunctionEntryTSC, int32_t FuncId) {
-  using AlignedFuncStorage =
-      std::aligned_storage<sizeof(FunctionRecord),
-                           alignof(FunctionRecord)>::type;
-  RecordPtr -= FunctionRecSize;
-  AlignedFuncStorage AlignedFuncRecordBuffer;
-  const auto &FuncRecord = *reinterpret_cast<FunctionRecord *>(
-      std::memcpy(&AlignedFuncRecordBuffer, RecordPtr, FunctionRecSize));
+  auto &TLD = getThreadLocalData();
+  TLD.RecordPtr -= FunctionRecSize;
+  decrementExtents(FunctionRecSize);
+  FunctionRecord FuncRecord;
+  std::memcpy(&FuncRecord, TLD.RecordPtr, FunctionRecSize);
   assert(FuncRecord.RecordKind ==
              uint8_t(FunctionRecord::RecordKinds::FunctionEnter) &&
          "Expected to find function entry recording when rewinding.");
   assert(FuncRecord.FuncId == (FuncId & ~(0x0F << 28)) &&
          "Expected matching function id when rewinding Exit");
-  --NumConsecutiveFnEnters;
+  --TLD.NumConsecutiveFnEnters;
   LastTSC -= FuncRecord.TSCDelta;
 
   // We unwound one call. Update the state and return without writing a log.
-  if (NumConsecutiveFnEnters != 0) {
+  if (TLD.NumConsecutiveFnEnters != 0) {
     LastFunctionEntryTSC -= FuncRecord.TSCDelta;
     return;
   }
@@ -413,22 +407,19 @@ static void rewindRecentCall(uint64_t TSC, uint64_t &LastTSC,
   // exited from via this exit.
   LastFunctionEntryTSC = 0;
   auto RewindingTSC = LastTSC;
-  auto RewindingRecordPtr = RecordPtr - FunctionRecSize;
-  while (NumTailCalls > 0) {
-    AlignedFuncStorage TailExitRecordBuffer;
+  auto RewindingRecordPtr = TLD.RecordPtr - FunctionRecSize;
+  while (TLD.NumTailCalls > 0) {
     // Rewind the TSC back over the TAIL EXIT record.
-    const auto &ExpectedTailExit =
-        *reinterpret_cast<FunctionRecord *>(std::memcpy(
-            &TailExitRecordBuffer, RewindingRecordPtr, FunctionRecSize));
+    FunctionRecord ExpectedTailExit;
+    std::memcpy(&ExpectedTailExit, RewindingRecordPtr, FunctionRecSize);
 
     assert(ExpectedTailExit.RecordKind ==
                uint8_t(FunctionRecord::RecordKinds::FunctionTailExit) &&
            "Expected to find tail exit when rewinding.");
     RewindingRecordPtr -= FunctionRecSize;
     RewindingTSC -= ExpectedTailExit.TSCDelta;
-    AlignedFuncStorage FunctionEntryBuffer;
-    const auto &ExpectedFunctionEntry = *reinterpret_cast<FunctionRecord *>(
-        std::memcpy(&FunctionEntryBuffer, RewindingRecordPtr, FunctionRecSize));
+    FunctionRecord ExpectedFunctionEntry;
+    std::memcpy(&ExpectedFunctionEntry, RewindingRecordPtr, FunctionRecSize);
     assert(ExpectedFunctionEntry.RecordKind ==
                uint8_t(FunctionRecord::RecordKinds::FunctionEnter) &&
            "Expected to find function entry when rewinding tail call.");
@@ -437,80 +428,87 @@ static void rewindRecentCall(uint64_t TSC, uint64_t &LastTSC,
 
     // This tail call exceeded the threshold duration. It will not be erased.
     if ((TSC - RewindingTSC) >= thresholdTicks()) {
-      NumTailCalls = 0;
+      TLD.NumTailCalls = 0;
       return;
     }
 
     // We can erase a tail exit pair that we're exiting through since
     // its duration is under threshold.
-    --NumTailCalls;
+    --TLD.NumTailCalls;
     RewindingRecordPtr -= FunctionRecSize;
     RewindingTSC -= ExpectedFunctionEntry.TSCDelta;
-    RecordPtr -= 2 * FunctionRecSize;
+    TLD.RecordPtr -= 2 * FunctionRecSize;
     LastTSC = RewindingTSC;
+    decrementExtents(2 * FunctionRecSize);
   }
 }
 
-inline bool releaseThreadLocalBuffer(BufferQueue *BQ) {
-  auto EC = BQ->releaseBuffer(Buffer);
+inline bool releaseThreadLocalBuffer(BufferQueue &BQArg) {
+  auto &TLD = getThreadLocalData();
+  auto EC = BQArg.releaseBuffer(TLD.Buffer);
   if (EC != BufferQueue::ErrorCode::Ok) {
-    Report("Failed to release buffer at %p; error=%s\n", Buffer.Buffer,
+    Report("Failed to release buffer at %p; error=%s\n", TLD.Buffer.Buffer,
            BufferQueue::getErrorString(EC));
     return false;
   }
   return true;
 }
 
-inline bool prepareBuffer(int (*wall_clock_reader)(clockid_t,
+inline bool prepareBuffer(uint64_t TSC, unsigned char CPU,
+                          int (*wall_clock_reader)(clockid_t,
                                                    struct timespec *),
                           size_t MaxSize) XRAY_NEVER_INSTRUMENT {
-  char *BufferStart = static_cast<char *>(Buffer.Buffer);
-  if ((RecordPtr + MaxSize) > (BufferStart + Buffer.Size - MetadataRecSize)) {
-    writeEOBMetadata();
-    if (!releaseThreadLocalBuffer(LocalBQ.get()))
+  auto &TLD = getThreadLocalData();
+  char *BufferStart = static_cast<char *>(TLD.Buffer.Buffer);
+  if ((TLD.RecordPtr + MaxSize) > (BufferStart + TLD.Buffer.Size)) {
+    if (!releaseThreadLocalBuffer(*TLD.BQ))
       return false;
-    auto EC = LocalBQ->getBuffer(Buffer);
+    auto EC = TLD.BQ->getBuffer(TLD.Buffer);
     if (EC != BufferQueue::ErrorCode::Ok) {
       Report("Failed to acquire a buffer; error=%s\n",
              BufferQueue::getErrorString(EC));
       return false;
     }
     setupNewBuffer(wall_clock_reader);
+
+    // Always write the CPU metadata as the first record in the buffer.
+    writeNewCPUIdMetadata(CPU, TSC);
   }
   return true;
 }
 
-inline bool isLogInitializedAndReady(
-    std::shared_ptr<BufferQueue> &LocalBQ, uint64_t TSC, unsigned char CPU,
-    int (*wall_clock_reader)(clockid_t,
-                             struct timespec *)) XRAY_NEVER_INSTRUMENT {
+inline bool
+isLogInitializedAndReady(BufferQueue *LBQ, uint64_t TSC, unsigned char CPU,
+                         int (*wall_clock_reader)(clockid_t, struct timespec *))
+    XRAY_NEVER_INSTRUMENT {
   // Bail out right away if logging is not initialized yet.
   // We should take the opportunity to release the buffer though.
   auto Status = __sanitizer::atomic_load(&LoggingStatus,
                                          __sanitizer::memory_order_acquire);
+  auto &TLD = getThreadLocalData();
   if (Status != XRayLogInitStatus::XRAY_LOG_INITIALIZED) {
-    if (RecordPtr != nullptr &&
+    if (TLD.RecordPtr != nullptr &&
         (Status == XRayLogInitStatus::XRAY_LOG_FINALIZING ||
          Status == XRayLogInitStatus::XRAY_LOG_FINALIZED)) {
-      writeEOBMetadata();
-      if (!releaseThreadLocalBuffer(LocalBQ.get()))
+      if (!releaseThreadLocalBuffer(*LBQ))
         return false;
-      RecordPtr = nullptr;
-      LocalBQ = nullptr;
+      TLD.RecordPtr = nullptr;
       return false;
     }
     return false;
   }
 
-  if (!loggingInitialized(LoggingStatus) || LocalBQ->finalizing()) {
-    writeEOBMetadata();
-    if (!releaseThreadLocalBuffer(LocalBQ.get()))
+  if (__sanitizer::atomic_load(&LoggingStatus,
+                               __sanitizer::memory_order_acquire) !=
+          XRayLogInitStatus::XRAY_LOG_INITIALIZED ||
+      LBQ->finalizing()) {
+    if (!releaseThreadLocalBuffer(*LBQ))
       return false;
-    RecordPtr = nullptr;
+    TLD.RecordPtr = nullptr;
   }
 
-  if (Buffer.Buffer == nullptr) {
-    auto EC = LocalBQ->getBuffer(Buffer);
+  if (TLD.Buffer.Buffer == nullptr) {
+    auto EC = LBQ->getBuffer(TLD.Buffer);
     if (EC != BufferQueue::ErrorCode::Ok) {
       auto LS = __sanitizer::atomic_load(&LoggingStatus,
                                          __sanitizer::memory_order_acquire);
@@ -522,51 +520,100 @@ inline bool isLogInitializedAndReady(
     }
 
     setupNewBuffer(wall_clock_reader);
+
+    // Always write the CPU metadata as the first record in the buffer.
+    writeNewCPUIdMetadata(CPU, TSC);
   }
 
-  if (CurrentCPU == std::numeric_limits<uint16_t>::max()) {
+  if (TLD.CurrentCPU == std::numeric_limits<uint16_t>::max()) {
     // This means this is the first CPU this thread has ever run on. We set
     // the current CPU and record this as the first TSC we've seen.
-    CurrentCPU = CPU;
+    TLD.CurrentCPU = CPU;
     writeNewCPUIdMetadata(CPU, TSC);
   }
 
   return true;
 } // namespace __xray_fdr_internal
 
+// Compute the TSC difference between the time of measurement and the previous
+// event. There are a few interesting situations we need to account for:
+//
+//   - The thread has migrated to a different CPU. If this is the case, then
+//     we write down the following records:
+//
+//       1. A 'NewCPUId' Metadata record.
+//       2. A FunctionRecord with a 0 for the TSCDelta field.
+//
+//   - The TSC delta is greater than the 32 bits we can store in a
+//     FunctionRecord. In this case we write down the following records:
+//
+//       1. A 'TSCWrap' Metadata record.
+//       2. A FunctionRecord with a 0 for the TSCDelta field.
+//
+//   - The TSC delta is representable within the 32 bits we can store in a
+//     FunctionRecord. In this case we write down just a FunctionRecord with
+//     the correct TSC delta.
+inline uint32_t writeCurrentCPUTSC(ThreadLocalData &TLD, uint64_t TSC,
+                                   uint8_t CPU) {
+  if (CPU != TLD.CurrentCPU) {
+    // We've moved to a new CPU.
+    writeNewCPUIdMetadata(CPU, TSC);
+    return 0;
+  }
+  // If the delta is greater than the range for a uint32_t, then we write out
+  // the TSC wrap metadata entry with the full TSC, and the TSC for the
+  // function record be 0.
+  uint64_t Delta = TSC - TLD.LastTSC;
+  if (Delta <= std::numeric_limits<uint32_t>::max())
+    return Delta;
+
+  writeTSCWrapMetadata(TSC);
+  return 0;
+}
+
 inline void endBufferIfFull() XRAY_NEVER_INSTRUMENT {
-  auto BufferStart = static_cast<char *>(Buffer.Buffer);
-  if ((RecordPtr + MetadataRecSize) - BufferStart == MetadataRecSize) {
-    writeEOBMetadata();
-    if (!releaseThreadLocalBuffer(LocalBQ.get()))
+  auto &TLD = getThreadLocalData();
+  auto BufferStart = static_cast<char *>(TLD.Buffer.Buffer);
+  if ((TLD.RecordPtr + MetadataRecSize) - BufferStart <=
+      ptrdiff_t{MetadataRecSize}) {
+    if (!releaseThreadLocalBuffer(*TLD.BQ))
       return;
-    RecordPtr = nullptr;
+    TLD.RecordPtr = nullptr;
   }
 }
 
-inline void processFunctionHook(
-    int32_t FuncId, XRayEntryType Entry, uint64_t TSC, unsigned char CPU,
-    int (*wall_clock_reader)(clockid_t, struct timespec *),
-    __sanitizer::atomic_sint32_t &LoggingStatus,
-    const std::shared_ptr<BufferQueue> &BQ) XRAY_NEVER_INSTRUMENT {
+thread_local volatile bool Running = false;
+
+/// Here's where the meat of the processing happens. The writer captures
+/// function entry, exit and tail exit points with a time and will create
+/// TSCWrap, NewCPUId and Function records as necessary. The writer might
+/// walk backward through its buffer and erase trivial functions to avoid
+/// polluting the log and may use the buffer queue to obtain or release a
+/// buffer.
+inline void processFunctionHook(int32_t FuncId, XRayEntryType Entry,
+                                uint64_t TSC, unsigned char CPU, uint64_t Arg1,
+                                int (*wall_clock_reader)(clockid_t,
+                                                         struct timespec *),
+                                BufferQueue *BQ) XRAY_NEVER_INSTRUMENT {
   // Prevent signal handler recursion, so in case we're already in a log writing
   // mode and the signal handler comes in (and is also instrumented) then we
   // don't want to be clobbering potentially partial writes already happening in
   // the thread. We use a simple thread_local latch to only allow one on-going
   // handleArg0 to happen at any given time.
-  thread_local bool Running = false;
   RecursionGuard Guard{Running};
   if (!Guard) {
     assert(Running == true && "RecursionGuard is buggy!");
     return;
   }
 
+  auto &TLD = getThreadLocalData();
+
   // In case the reference has been cleaned up before, we make sure we
   // initialize it to the provided BufferQueue.
-  if (LocalBQ == nullptr)
-    LocalBQ = BQ;
+  if (TLD.BQ == nullptr)
+    TLD.BQ = BQ;
 
-  if (!isLogInitializedAndReady(LocalBQ, TSC, CPU, wall_clock_reader))
+  if (!isLogInitializedAndReady(TLD.BQ, TSC, CPU, wall_clock_reader))
     return;
 
   // Before we go setting up writing new function entries, we need to be really
@@ -579,10 +626,10 @@ inline void processFunctionHook(
   //   - The least number of bytes we will ever write is 8
   //     (sizeof(FunctionRecord)) only if the delta between the previous entry
   //     and this entry is within 32 bits.
-  //   - The most number of bytes we will ever write is 8 + 16 = 24. This is
-  //     computed by:
+  //   - The most number of bytes we will ever write is 8 + 16 + 16 = 40.
+  //     This is computed by:
   //
-  //       sizeof(FunctionRecord) + sizeof(MetadataRecord)
+  //       MaxSize = sizeof(FunctionRecord) + 2 * sizeof(MetadataRecord)
   //
   //     These arise in the following cases:
   //
@@ -596,77 +643,39 @@ inline void processFunctionHook(
   //          FunctionRecord.
   //       3. When we learn about a new CPU ID, we need to write down a "new cpu
   //          id" MetadataRecord before writing out the actual FunctionRecord.
+  //       4. The second MetadataRecord is the optional function call argument.
   //
-  //   - An End-of-Buffer (EOB) MetadataRecord is 16 bytes.
-  //
-  // So the math we need to do is to determine whether writing 24 bytes past the
-  // current pointer leaves us with enough bytes to write the EOB
-  // MetadataRecord. If we don't have enough space after writing as much as 24
-  // bytes in the end of the buffer, we need to write out the EOB, get a new
-  // Buffer, set it up properly before doing any further writing.
-  //
-  if (!prepareBuffer(wall_clock_reader, FunctionRecSize + MetadataRecSize)) {
-    LocalBQ = nullptr;
+  // So the math we need to do is to determine whether writing 40 bytes past the
+  // current pointer exceeds the buffer's maximum size. If we don't have enough
+  // space to write 40 bytes in the buffer, we need get a new Buffer, set it up
+  // properly before doing any further writing.
+  size_t MaxSize = FunctionRecSize + 2 * MetadataRecSize;
+  if (!prepareBuffer(TSC, CPU, wall_clock_reader, MaxSize)) {
+    TLD.BQ = nullptr;
     return;
   }
 
-  // By this point, we are now ready to write at most 24 bytes (one metadata
-  // record and one function record).
-  assert((RecordPtr + (MetadataRecSize + FunctionRecSize)) -
-                 static_cast<char *>(Buffer.Buffer) >=
+  // By this point, we are now ready to write up to 40 bytes (explained above).
+  assert((TLD.RecordPtr + MaxSize) - static_cast<char *>(TLD.Buffer.Buffer) >=
              static_cast<ptrdiff_t>(MetadataRecSize) &&
          "Misconfigured BufferQueue provided; Buffer size not large enough.");
 
-  // Here we compute the TSC Delta. There are a few interesting situations we
-  // need to account for:
-  //
-  //   - The thread has migrated to a different CPU. If this is the case, then
-  //     we write down the following records:
-  //
-  //       1. A 'NewCPUId' Metadata record.
-  //       2. A FunctionRecord with a 0 for the TSCDelta field.
-  //
-  //   - The TSC delta is greater than the 32 bits we can store in a
-  //     FunctionRecord. In this case we write down the following records:
-  //
-  //       1. A 'TSCWrap' Metadata record.
-  //       2. A FunctionRecord with a 0 for the TSCDelta field.
-  //
-  //   - The TSC delta is representable within the 32 bits we can store in a
-  //     FunctionRecord. In this case we write down just a FunctionRecord with
-  //     the correct TSC delta.
-  //
-  uint32_t RecordTSCDelta = 0;
-  if (CPU != CurrentCPU) {
-    // We've moved to a new CPU.
-    writeNewCPUIdMetadata(CPU, TSC);
-  } else {
-    // If the delta is greater than the range for a uint32_t, then we write out
-    // the TSC wrap metadata entry with the full TSC, and the TSC for the
-    // function record be 0.
-    auto Delta = TSC - LastTSC;
-    if (Delta > (1ULL << 32) - 1)
-      writeTSCWrapMetadata(TSC);
-    else
-      RecordTSCDelta = Delta;
-  }
-
-  LastTSC = TSC;
-  CurrentCPU = CPU;
+  auto RecordTSCDelta = writeCurrentCPUTSC(TLD, TSC, CPU);
+  TLD.LastTSC = TSC;
+  TLD.CurrentCPU = CPU;
   switch (Entry) {
   case XRayEntryType::ENTRY:
   case XRayEntryType::LOG_ARGS_ENTRY:
     // Update the thread local state for the next invocation.
-    LastFunctionEntryTSC = TSC;
+    TLD.LastFunctionEntryTSC = TSC;
     break;
   case XRayEntryType::TAIL:
-    break;
   case XRayEntryType::EXIT:
     // Break out and write the exit record if we can't erase any functions.
-    if (NumConsecutiveFnEnters == 0 ||
-        (TSC - LastFunctionEntryTSC) >= thresholdTicks())
+    if (TLD.NumConsecutiveFnEnters == 0 ||
+        (TSC - TLD.LastFunctionEntryTSC) >= thresholdTicks())
       break;
-    rewindRecentCall(TSC, LastTSC, LastFunctionEntryTSC, FuncId);
+    rewindRecentCall(TSC, TLD.LastTSC, TLD.LastFunctionEntryTSC, FuncId);
     return; // without writing log.
   case XRayEntryType::CUSTOM_EVENT: {
     // This is a bug in patching, so we'll report it once and move on.
@@ -681,7 +690,9 @@ inline void processFunctionHook(
   }
   }
 
-  writeFunctionRecord(FuncId, RecordTSCDelta, Entry, RecordPtr);
+  writeFunctionRecord(FuncId, RecordTSCDelta, Entry);
+  if (Entry == XRayEntryType::LOG_ARGS_ENTRY)
+    writeCallArgumentMetadata(Arg1);
 
   // If we've exhausted the buffer by this time, we then release the buffer to
   // make sure that other threads may start using this buffer.
diff --git a/contrib/compiler-rt/lib/xray/xray_flags.h b/contrib/compiler-rt/lib/xray/xray_flags.h
index f4e30283b8de..3ed5b8844cb4 100644
--- a/contrib/compiler-rt/lib/xray/xray_flags.h
+++ b/contrib/compiler-rt/lib/xray/xray_flags.h
@@ -16,6 +16,7 @@
 #define XRAY_FLAGS_H
 
 #include "sanitizer_common/sanitizer_flag_parser.h"
+#include "sanitizer_common/sanitizer_internal_defs.h"
 
 namespace __xray {
 
diff --git a/contrib/compiler-rt/lib/xray/xray_flags.inc b/contrib/compiler-rt/lib/xray/xray_flags.inc
index 7ddce78eb413..29f1fce7d7f4 100644
--- a/contrib/compiler-rt/lib/xray/xray_flags.inc
+++ b/contrib/compiler-rt/lib/xray/xray_flags.inc
@@ -16,12 +16,34 @@
 
 XRAY_FLAG(bool, patch_premain, false,
           "Whether to patch instrumentation points before main.")
-XRAY_FLAG(bool, xray_naive_log, true,
-          "Whether to install the naive log implementation.")
 XRAY_FLAG(const char *, xray_logfile_base, "xray-log.",
           "Filename base for the xray logfile.")
+XRAY_FLAG(const char *, xray_mode, "", "Mode to install by default.")
+XRAY_FLAG(uptr, xray_page_size_override, 0,
+          "Override the default page size for the system, in bytes. The size "
+          "should be a power-of-two.")
+
+// Basic (Naive) Mode logging options.
+XRAY_FLAG(bool, xray_naive_log, false,
+          "DEPRECATED: Use xray_mode=xray-basic instead.")
+XRAY_FLAG(int, xray_naive_log_func_duration_threshold_us, 5,
+          "Naive logging will try to skip functions that execute for fewer "
+          "microseconds than this threshold.")
+XRAY_FLAG(int, xray_naive_log_max_stack_depth, 64,
+          "Naive logging will keep track of at most this deep a call stack, "
+          "any more and the recordings will be droppped.")
+XRAY_FLAG(int, xray_naive_log_thread_buffer_size, 1024,
+          "The number of entries to keep on a per-thread buffer.")
+
+// FDR (Flight Data Recorder) Mode logging options.
 XRAY_FLAG(bool, xray_fdr_log, false,
-          "Whether to install the flight data recorder logging implementation.")
+          "DEPRECATED: Use xray_mode=xray-fdr instead.")
 XRAY_FLAG(int, xray_fdr_log_func_duration_threshold_us, 5,
           "FDR logging will try to skip functions that execute for fewer "
           "microseconds than this threshold.")
+XRAY_FLAG(int, xray_fdr_log_grace_period_us, 0,
+          "DEPRECATED: use xray_fdr_log_grace_period_ms instead.")
+XRAY_FLAG(int, xray_fdr_log_grace_period_ms, 100,
+          "FDR logging will wait this much time in microseconds before "
+          "actually flushing the log; this gives a chance for threads to "
+          "notice that the log has been finalized and clean up.")
diff --git a/contrib/compiler-rt/lib/xray/xray_init.cc b/contrib/compiler-rt/lib/xray/xray_init.cc
index aa660baa9920..11892cb8b7a3 100644
--- a/contrib/compiler-rt/lib/xray/xray_init.cc
+++ b/contrib/compiler-rt/lib/xray/xray_init.cc
@@ -44,12 +44,31 @@ __sanitizer::atomic_uint8_t XRayInitialized{0};
 __sanitizer::SpinMutex XRayInstrMapMutex;
 XRaySledMap XRayInstrMap;
 
+// Global flag to determine whether the flags have been initialized.
+__sanitizer::atomic_uint8_t XRayFlagsInitialized{0};
+
+// A mutex to allow only one thread to initialize the XRay data structures.
+__sanitizer::SpinMutex XRayInitMutex;
+
 // __xray_init() will do the actual loading of the current process' memory map
 // and then proceed to look for the .xray_instr_map section/segment.
 void __xray_init() XRAY_NEVER_INSTRUMENT {
-  initializeFlags();
+  __sanitizer::SpinMutexLock Guard(&XRayInitMutex);
+  // Short-circuit if we've already initialized XRay before.
+  if (__sanitizer::atomic_load(&XRayInitialized,
+                               __sanitizer::memory_order_acquire))
+    return;
+
+  if (!__sanitizer::atomic_load(&XRayFlagsInitialized,
+                                __sanitizer::memory_order_acquire)) {
+    initializeFlags();
+    __sanitizer::atomic_store(&XRayFlagsInitialized, true,
+                              __sanitizer::memory_order_release);
+  }
+
   if (__start_xray_instr_map == nullptr) {
-    Report("XRay instrumentation map missing. Not initializing XRay.\n");
+    if (Verbosity())
+      Report("XRay instrumentation map missing. Not initializing XRay.\n");
     return;
   }
 
@@ -63,9 +82,21 @@ void __xray_init() XRAY_NEVER_INSTRUMENT {
   __sanitizer::atomic_store(&XRayInitialized, true,
                             __sanitizer::memory_order_release);
 
+#ifndef XRAY_NO_PREINIT
   if (flags()->patch_premain)
     __xray_patch();
+#endif
 }
 
+#if !defined(XRAY_NO_PREINIT) && SANITIZER_CAN_USE_PREINIT_ARRAY
+// Only add the preinit array initialization if the sanitizers can.
 __attribute__((section(".preinit_array"),
                used)) void (*__local_xray_preinit)(void) = __xray_init;
+#else
+// If we cannot use the .preinit_array section, we should instead use dynamic
+// initialisation.
+static bool UNUSED __local_xray_dyninit = [] {
+  __xray_init();
+  return true;
+}();
+#endif
diff --git a/contrib/compiler-rt/lib/xray/xray_inmemory_log.cc b/contrib/compiler-rt/lib/xray/xray_inmemory_log.cc
index 83aecfaf7700..a27ffbcbd12e 100644
--- a/contrib/compiler-rt/lib/xray/xray_inmemory_log.cc
+++ b/contrib/compiler-rt/lib/xray/xray_inmemory_log.cc
@@ -16,80 +16,85 @@
 //===----------------------------------------------------------------------===//
 
 #include <cassert>
+#include <cstring>
+#include <errno.h>
 #include <fcntl.h>
-#include <mutex>
+#include <pthread.h>
 #include <sys/stat.h>
 #include <sys/syscall.h>
 #include <sys/types.h>
-#include <thread>
+#include <time.h>
 #include <unistd.h>
 
+#include "sanitizer_common/sanitizer_allocator_internal.h"
 #include "sanitizer_common/sanitizer_libc.h"
 #include "xray/xray_records.h"
 #include "xray_defs.h"
 #include "xray_flags.h"
+#include "xray_inmemory_log.h"
 #include "xray_interface_internal.h"
 #include "xray_tsc.h"
 #include "xray_utils.h"
 
-// __xray_InMemoryRawLog will use a thread-local aligned buffer capped to a
-// certain size (32kb by default) and use it as if it were a circular buffer for
-// events. We store simple fixed-sized entries in the log for external analysis.
-
-extern "C" {
-void __xray_InMemoryRawLog(int32_t FuncId,
-                           XRayEntryType Type) XRAY_NEVER_INSTRUMENT;
-}
-
 namespace __xray {
 
-std::mutex LogMutex;
+__sanitizer::SpinMutex LogMutex;
 
-class ThreadExitFlusher {
-  int Fd;
-  XRayRecord *Start;
-  size_t &Offset;
+// We use elements of this type to record the entry TSC of every function ID we
+// see as we're tracing a particular thread's execution.
+struct alignas(16) StackEntry {
+  int32_t FuncId;
+  uint16_t Type;
+  uint8_t CPU;
+  uint8_t Padding;
+  uint64_t TSC;
+};
 
-public:
-  explicit ThreadExitFlusher(int Fd, XRayRecord *Start,
-                             size_t &Offset) XRAY_NEVER_INSTRUMENT
-      : Fd(Fd),
-        Start(Start),
-        Offset(Offset) {}
+static_assert(sizeof(StackEntry) == 16, "Wrong size for StackEntry");
 
-  ~ThreadExitFlusher() XRAY_NEVER_INSTRUMENT {
-    std::lock_guard<std::mutex> L(LogMutex);
-    if (Fd > 0 && Start != nullptr) {
-      retryingWriteAll(Fd, reinterpret_cast<char *>(Start),
-                       reinterpret_cast<char *>(Start + Offset));
-      // Because this thread's exit could be the last one trying to write to the
-      // file and that we're not able to close out the file properly, we sync
-      // instead and hope that the pending writes are flushed as the thread
-      // exits.
-      fsync(Fd);
-    }
-  }
+struct alignas(64) ThreadLocalData {
+  void *InMemoryBuffer = nullptr;
+  size_t BufferSize = 0;
+  size_t BufferOffset = 0;
+  void *ShadowStack = nullptr;
+  size_t StackSize = 0;
+  size_t StackEntries = 0;
+  int Fd = -1;
+  pid_t TID = 0;
 };
 
-} // namespace __xray
+static pthread_key_t PThreadKey;
+
+static __sanitizer::atomic_uint8_t BasicInitialized{0};
+
+BasicLoggingOptions GlobalOptions;
 
-using namespace __xray;
+thread_local volatile bool RecursionGuard = false;
+
+static uint64_t thresholdTicks() XRAY_NEVER_INSTRUMENT {
+  static uint64_t TicksPerSec = probeRequiredCPUFeatures()
+                                    ? getTSCFrequency()
+                                    : __xray::NanosecondsPerSecond;
+  static const uint64_t ThresholdTicks =
+      TicksPerSec * GlobalOptions.DurationFilterMicros / 1000000;
+  return ThresholdTicks;
+}
 
-static int __xray_OpenLogFile() XRAY_NEVER_INSTRUMENT {
+static int openLogFile() XRAY_NEVER_INSTRUMENT {
   int F = getLogFD();
   if (F == -1)
     return -1;
 
   // Test for required CPU features and cache the cycle frequency
   static bool TSCSupported = probeRequiredCPUFeatures();
-  static uint64_t CycleFrequency = TSCSupported ? getTSCFrequency()
-                                   : __xray::NanosecondsPerSecond;
+  static uint64_t CycleFrequency =
+      TSCSupported ? getTSCFrequency() : __xray::NanosecondsPerSecond;
 
   // Since we're here, we get to write the header. We set it up so that the
   // header will only be written once, at the start, and let the threads
   // logging do writes which just append.
   XRayFileHeader Header;
-  Header.Version = 1;
+  Header.Version = 2; // Version 2 includes tail exit records.
   Header.Type = FileTypes::NAIVE_LOG;
   Header.CycleFrequency = CycleFrequency;
 
@@ -102,47 +107,210 @@ static int __xray_OpenLogFile() XRAY_NEVER_INSTRUMENT {
   return F;
 }
 
+int getGlobalFd() XRAY_NEVER_INSTRUMENT {
+  static int Fd = openLogFile();
+  return Fd;
+}
+
+ThreadLocalData &getThreadLocalData() XRAY_NEVER_INSTRUMENT {
+  thread_local ThreadLocalData TLD;
+  thread_local bool UNUSED TOnce = [] {
+    if (GlobalOptions.ThreadBufferSize == 0) {
+      if (__sanitizer::Verbosity())
+        Report("Not initializing TLD since ThreadBufferSize == 0.\n");
+      return false;
+    }
+    TLD.TID = __sanitizer::GetTid();
+    pthread_setspecific(PThreadKey, &TLD);
+    TLD.Fd = getGlobalFd();
+    TLD.InMemoryBuffer = reinterpret_cast<XRayRecord *>(
+        InternalAlloc(sizeof(XRayRecord) * GlobalOptions.ThreadBufferSize,
+                      nullptr, alignof(XRayRecord)));
+    TLD.BufferSize = GlobalOptions.ThreadBufferSize;
+    TLD.BufferOffset = 0;
+    if (GlobalOptions.MaxStackDepth == 0) {
+      if (__sanitizer::Verbosity())
+        Report("Not initializing the ShadowStack since MaxStackDepth == 0.\n");
+      TLD.StackSize = 0;
+      TLD.StackEntries = 0;
+      TLD.ShadowStack = nullptr;
+      return false;
+    }
+    TLD.ShadowStack = reinterpret_cast<StackEntry *>(
+        InternalAlloc(sizeof(StackEntry) * GlobalOptions.MaxStackDepth, nullptr,
+                      alignof(StackEntry)));
+    TLD.StackSize = GlobalOptions.MaxStackDepth;
+    TLD.StackEntries = 0;
+    if (__sanitizer::Verbosity() >= 2) {
+      static auto UNUSED Once = [] {
+        auto ticks = thresholdTicks();
+        Report("Ticks threshold: %d\n", ticks);
+        return false;
+      }();
+    }
+    return false;
+  }();
+  return TLD;
+}
+
 template <class RDTSC>
-void __xray_InMemoryRawLog(int32_t FuncId, XRayEntryType Type,
-                           RDTSC ReadTSC) XRAY_NEVER_INSTRUMENT {
-  using Buffer =
-      std::aligned_storage<sizeof(XRayRecord), alignof(XRayRecord)>::type;
-  static constexpr size_t BuffLen = 1024;
-  thread_local static Buffer InMemoryBuffer[BuffLen] = {};
-  thread_local static size_t Offset = 0;
-  static int Fd = __xray_OpenLogFile();
+void InMemoryRawLog(int32_t FuncId, XRayEntryType Type,
+                    RDTSC ReadTSC) XRAY_NEVER_INSTRUMENT {
+  auto &TLD = getThreadLocalData();
+  auto &InMemoryBuffer = TLD.InMemoryBuffer;
+  int Fd = getGlobalFd();
   if (Fd == -1)
     return;
-  thread_local __xray::ThreadExitFlusher Flusher(
-      Fd, reinterpret_cast<__xray::XRayRecord *>(InMemoryBuffer), Offset);
-  thread_local pid_t TId = syscall(SYS_gettid);
 
-  // First we get the useful data, and stuff it into the already aligned buffer
-  // through a pointer offset.
-  auto &R = reinterpret_cast<__xray::XRayRecord *>(InMemoryBuffer)[Offset];
+  // Use a simple recursion guard, to handle cases where we're already logging
+  // and for one reason or another, this function gets called again in the same
+  // thread.
+  if (RecursionGuard)
+    return;
+  RecursionGuard = true;
+  auto ExitGuard = __sanitizer::at_scope_exit([] { RecursionGuard = false; });
+
+  uint8_t CPU = 0;
+  uint64_t TSC = ReadTSC(CPU);
+
+  switch (Type) {
+  case XRayEntryType::ENTRY:
+  case XRayEntryType::LOG_ARGS_ENTRY: {
+    // Short circuit if we've reached the maximum depth of the stack.
+    if (TLD.StackEntries++ >= TLD.StackSize)
+      return;
+
+    // When we encounter an entry event, we keep track of the TSC and the CPU,
+    // and put it in the stack.
+    StackEntry E;
+    E.FuncId = FuncId;
+    E.CPU = CPU;
+    E.Type = Type;
+    E.TSC = TSC;
+    auto StackEntryPtr = static_cast<char *>(TLD.ShadowStack) +
+                         (sizeof(StackEntry) * (TLD.StackEntries - 1));
+    __sanitizer::internal_memcpy(StackEntryPtr, &E, sizeof(StackEntry));
+    break;
+  }
+  case XRayEntryType::EXIT:
+  case XRayEntryType::TAIL: {
+    if (TLD.StackEntries == 0)
+      break;
+
+    if (--TLD.StackEntries >= TLD.StackSize)
+      return;
+
+    // When we encounter an exit event, we check whether all the following are
+    // true:
+    //
+    // - The Function ID is the same as the most recent entry in the stack.
+    // - The CPU is the same as the most recent entry in the stack.
+    // - The Delta of the TSCs is less than the threshold amount of time we're
+    //   looking to record.
+    //
+    // If all of these conditions are true, we pop the stack and don't write a
+    // record and move the record offset back.
+    StackEntry StackTop;
+    auto StackEntryPtr = static_cast<char *>(TLD.ShadowStack) +
+                         (sizeof(StackEntry) * TLD.StackEntries);
+    __sanitizer::internal_memcpy(&StackTop, StackEntryPtr, sizeof(StackEntry));
+    if (StackTop.FuncId == FuncId && StackTop.CPU == CPU &&
+        StackTop.TSC < TSC) {
+      auto Delta = TSC - StackTop.TSC;
+      if (Delta < thresholdTicks()) {
+        assert(TLD.BufferOffset > 0);
+        TLD.BufferOffset -= StackTop.Type == XRayEntryType::ENTRY ? 1 : 2;
+        return;
+      }
+    }
+    break;
+  }
+  default:
+    // Should be unreachable.
+    assert(false && "Unsupported XRayEntryType encountered.");
+    break;
+  }
+
+  // First determine whether the delta between the function's enter record and
+  // the exit record is higher than the threshold.
+  __xray::XRayRecord R;
   R.RecordType = RecordTypes::NORMAL;
-  R.TSC = ReadTSC(R.CPU);
-  R.TId = TId;
+  R.CPU = CPU;
+  R.TSC = TSC;
+  R.TId = TLD.TID;
   R.Type = Type;
   R.FuncId = FuncId;
-  ++Offset;
-  if (Offset == BuffLen) {
-    std::lock_guard<std::mutex> L(LogMutex);
+  auto EntryPtr = static_cast<char *>(InMemoryBuffer) +
+                  (sizeof(__xray::XRayRecord) * TLD.BufferOffset);
+  __sanitizer::internal_memcpy(EntryPtr, &R, sizeof(R));
+  if (++TLD.BufferOffset == TLD.BufferSize) {
+    __sanitizer::SpinMutexLock L(&LogMutex);
+    auto RecordBuffer = reinterpret_cast<__xray::XRayRecord *>(InMemoryBuffer);
+    retryingWriteAll(Fd, reinterpret_cast<char *>(RecordBuffer),
+                     reinterpret_cast<char *>(RecordBuffer + TLD.BufferOffset));
+    TLD.BufferOffset = 0;
+    TLD.StackEntries = 0;
+  }
+}
+
+template <class RDTSC>
+void InMemoryRawLogWithArg(int32_t FuncId, XRayEntryType Type, uint64_t Arg1,
+                           RDTSC ReadTSC) XRAY_NEVER_INSTRUMENT {
+  auto &TLD = getThreadLocalData();
+  auto &InMemoryBuffer = TLD.InMemoryBuffer;
+  auto &Offset = TLD.BufferOffset;
+  const auto &BuffLen = TLD.BufferSize;
+  int Fd = getGlobalFd();
+  if (Fd == -1)
+    return;
+
+  // First we check whether there's enough space to write the data consecutively
+  // in the thread-local buffer. If not, we first flush the buffer before
+  // attempting to write the two records that must be consecutive.
+  if (Offset + 2 > BuffLen) {
+    __sanitizer::SpinMutexLock L(&LogMutex);
     auto RecordBuffer = reinterpret_cast<__xray::XRayRecord *>(InMemoryBuffer);
     retryingWriteAll(Fd, reinterpret_cast<char *>(RecordBuffer),
                      reinterpret_cast<char *>(RecordBuffer + Offset));
     Offset = 0;
+    TLD.StackEntries = 0;
+  }
+
+  // Then we write the "we have an argument" record.
+  InMemoryRawLog(FuncId, Type, ReadTSC);
+
+  if (RecursionGuard)
+    return;
+  RecursionGuard = true;
+  auto ExitGuard = __sanitizer::at_scope_exit([] { RecursionGuard = false; });
+
+  // And from here on write the arg payload.
+  __xray::XRayArgPayload R;
+  R.RecordType = RecordTypes::ARG_PAYLOAD;
+  R.FuncId = FuncId;
+  R.TId = TLD.TID;
+  R.Arg = Arg1;
+  auto EntryPtr =
+      &reinterpret_cast<__xray::XRayArgPayload *>(&InMemoryBuffer)[Offset];
+  std::memcpy(EntryPtr, &R, sizeof(R));
+  if (++Offset == BuffLen) {
+    __sanitizer::SpinMutexLock L(&LogMutex);
+    auto RecordBuffer = reinterpret_cast<__xray::XRayRecord *>(InMemoryBuffer);
+    retryingWriteAll(Fd, reinterpret_cast<char *>(RecordBuffer),
+                     reinterpret_cast<char *>(RecordBuffer + Offset));
+    Offset = 0;
+    TLD.StackEntries = 0;
   }
 }
 
-void __xray_InMemoryRawLogRealTSC(int32_t FuncId,
-                                  XRayEntryType Type) XRAY_NEVER_INSTRUMENT {
-  __xray_InMemoryRawLog(FuncId, Type, __xray::readTSC);
+void basicLoggingHandleArg0RealTSC(int32_t FuncId,
+                                   XRayEntryType Type) XRAY_NEVER_INSTRUMENT {
+  InMemoryRawLog(FuncId, Type, __xray::readTSC);
 }
 
-void __xray_InMemoryEmulateTSC(int32_t FuncId,
-                               XRayEntryType Type) XRAY_NEVER_INSTRUMENT {
-  __xray_InMemoryRawLog(FuncId, Type, [](uint8_t &CPU) XRAY_NEVER_INSTRUMENT {
+void basicLoggingHandleArg0EmulateTSC(int32_t FuncId, XRayEntryType Type)
+    XRAY_NEVER_INSTRUMENT {
+  InMemoryRawLog(FuncId, Type, [](uint8_t &CPU) XRAY_NEVER_INSTRUMENT {
     timespec TS;
     int result = clock_gettime(CLOCK_REALTIME, &TS);
     if (result != 0) {
@@ -154,13 +322,150 @@ void __xray_InMemoryEmulateTSC(int32_t FuncId,
   });
 }
 
-static auto UNUSED Unused = [] {
-  auto UseRealTSC = probeRequiredCPUFeatures();
-  if (!UseRealTSC)
+void basicLoggingHandleArg1RealTSC(int32_t FuncId, XRayEntryType Type,
+                                   uint64_t Arg1) XRAY_NEVER_INSTRUMENT {
+  InMemoryRawLogWithArg(FuncId, Type, Arg1, __xray::readTSC);
+}
+
+void basicLoggingHandleArg1EmulateTSC(int32_t FuncId, XRayEntryType Type,
+                                      uint64_t Arg1) XRAY_NEVER_INSTRUMENT {
+  InMemoryRawLogWithArg(
+      FuncId, Type, Arg1, [](uint8_t &CPU) XRAY_NEVER_INSTRUMENT {
+        timespec TS;
+        int result = clock_gettime(CLOCK_REALTIME, &TS);
+        if (result != 0) {
+          Report("clock_gettimg(2) return %d, errno=%d.", result, int(errno));
+          TS = {0, 0};
+        }
+        CPU = 0;
+        return TS.tv_sec * __xray::NanosecondsPerSecond + TS.tv_nsec;
+      });
+}
+
+static void TLDDestructor(void *P) XRAY_NEVER_INSTRUMENT {
+  ThreadLocalData &TLD = *reinterpret_cast<ThreadLocalData *>(P);
+  auto ExitGuard = __sanitizer::at_scope_exit([&TLD] {
+    // Clean up dynamic resources.
+    if (TLD.InMemoryBuffer)
+      InternalFree(TLD.InMemoryBuffer);
+    if (TLD.ShadowStack)
+      InternalFree(TLD.ShadowStack);
+    if (__sanitizer::Verbosity())
+      Report("Cleaned up log for TID: %d\n", TLD.TID);
+  });
+
+  if (TLD.Fd == -1 || TLD.BufferOffset == 0) {
+    if (__sanitizer::Verbosity())
+      Report("Skipping buffer for TID: %d; Fd = %d; Offset = %llu\n", TLD.TID,
+             TLD.Fd, TLD.BufferOffset);
+    return;
+  }
+
+  {
+    __sanitizer::SpinMutexLock L(&LogMutex);
+    retryingWriteAll(TLD.Fd, reinterpret_cast<char *>(TLD.InMemoryBuffer),
+                     reinterpret_cast<char *>(TLD.InMemoryBuffer) +
+                         (sizeof(__xray::XRayRecord) * TLD.BufferOffset));
+  }
+
+  // Because this thread's exit could be the last one trying to write to
+  // the file and that we're not able to close out the file properly, we
+  // sync instead and hope that the pending writes are flushed as the
+  // thread exits.
+  fsync(TLD.Fd);
+}
+
+XRayLogInitStatus basicLoggingInit(size_t BufferSize, size_t BufferMax,
+                                   void *Options,
+                                   size_t OptionsSize) XRAY_NEVER_INSTRUMENT {
+  static bool UNUSED Once = [] {
+    pthread_key_create(&PThreadKey, TLDDestructor);
+    return false;
+  }();
+
+  uint8_t Expected = 0;
+  if (!__sanitizer::atomic_compare_exchange_strong(
+          &BasicInitialized, &Expected, 1, __sanitizer::memory_order_acq_rel)) {
+    if (__sanitizer::Verbosity())
+      Report("Basic logging already initialized.\n");
+    return XRayLogInitStatus::XRAY_LOG_INITIALIZED;
+  }
+
+  if (OptionsSize != sizeof(BasicLoggingOptions)) {
+    Report("Invalid options size, potential ABI mismatch; expected %d got %d",
+           sizeof(BasicLoggingOptions), OptionsSize);
+    return XRayLogInitStatus::XRAY_LOG_UNINITIALIZED;
+  }
+
+  static auto UseRealTSC = probeRequiredCPUFeatures();
+  if (!UseRealTSC && __sanitizer::Verbosity())
     Report("WARNING: Required CPU features missing for XRay instrumentation, "
            "using emulation instead.\n");
-  if (flags()->xray_naive_log)
-    __xray_set_handler(UseRealTSC ? __xray_InMemoryRawLogRealTSC
-                                  : __xray_InMemoryEmulateTSC);
+
+  GlobalOptions = *reinterpret_cast<BasicLoggingOptions *>(Options);
+  __xray_set_handler_arg1(UseRealTSC ? basicLoggingHandleArg1RealTSC
+                                     : basicLoggingHandleArg1EmulateTSC);
+  __xray_set_handler(UseRealTSC ? basicLoggingHandleArg0RealTSC
+                                : basicLoggingHandleArg0EmulateTSC);
+  __xray_remove_customevent_handler();
+
+  return XRayLogInitStatus::XRAY_LOG_INITIALIZED;
+}
+
+XRayLogInitStatus basicLoggingFinalize() XRAY_NEVER_INSTRUMENT {
+  uint8_t Expected = 0;
+  if (!__sanitizer::atomic_compare_exchange_strong(
+          &BasicInitialized, &Expected, 0, __sanitizer::memory_order_acq_rel) &&
+      __sanitizer::Verbosity())
+    Report("Basic logging already finalized.\n");
+
+  // Nothing really to do aside from marking state of the global to be
+  // uninitialized.
+
+  return XRayLogInitStatus::XRAY_LOG_FINALIZED;
+}
+
+XRayLogFlushStatus basicLoggingFlush() XRAY_NEVER_INSTRUMENT {
+  // This really does nothing, since flushing the logs happen at the end of a
+  // thread's lifetime, or when the buffers are full.
+  return XRayLogFlushStatus::XRAY_LOG_FLUSHED;
+}
+
+// This is a handler that, effectively, does nothing.
+void basicLoggingHandleArg0Empty(int32_t, XRayEntryType) XRAY_NEVER_INSTRUMENT {
+}
+
+bool basicLogDynamicInitializer() XRAY_NEVER_INSTRUMENT {
+  XRayLogImpl Impl{
+      basicLoggingInit,
+      basicLoggingFinalize,
+      basicLoggingHandleArg0Empty,
+      basicLoggingFlush,
+  };
+  auto RegistrationResult = __xray_log_register_mode("xray-basic", Impl);
+  if (RegistrationResult != XRayLogRegisterStatus::XRAY_REGISTRATION_OK &&
+      __sanitizer::Verbosity())
+    Report("Cannot register XRay Basic Mode to 'xray-basic'; error = %d\n",
+           RegistrationResult);
+  if (flags()->xray_naive_log ||
+      !__sanitizer::internal_strcmp(flags()->xray_mode, "xray-basic")) {
+    __xray_set_log_impl(Impl);
+    BasicLoggingOptions Options;
+    Options.DurationFilterMicros =
+        flags()->xray_naive_log_func_duration_threshold_us;
+    Options.MaxStackDepth = flags()->xray_naive_log_max_stack_depth;
+    Options.ThreadBufferSize = flags()->xray_naive_log_thread_buffer_size;
+    __xray_log_init(flags()->xray_naive_log_thread_buffer_size, 0, &Options,
+                    sizeof(BasicLoggingOptions));
+    static auto UNUSED Once = [] {
+      static auto UNUSED &TLD = getThreadLocalData();
+      __sanitizer::Atexit(+[] { TLDDestructor(&TLD); });
+      return false;
+    }();
+  }
   return true;
-}();
+}
+
+} // namespace __xray
+
+static auto UNUSED Unused = __xray::basicLogDynamicInitializer();
diff --git a/contrib/compiler-rt/lib/xray/xray_inmemory_log.h b/contrib/compiler-rt/lib/xray/xray_inmemory_log.h
new file mode 100644
index 000000000000..e4fcb8ca5ffd
--- /dev/null
+++ b/contrib/compiler-rt/lib/xray/xray_inmemory_log.h
@@ -0,0 +1,44 @@
+//===-- xray_inmemory_log.h
+//------------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of XRay, a function call tracing system.
+//
+//===----------------------------------------------------------------------===//
+#ifndef XRAY_XRAY_INMEMORY_LOG_H
+#define XRAY_XRAY_INMEMORY_LOG_H
+
+#include "xray/xray_log_interface.h"
+
+/// Basic (Naive) Mode
+/// ==================
+///
+/// This implementation hooks in through the XRay logging implementation
+/// framework. The Basic Mode implementation will keep appending to a file as
+/// soon as the thread-local buffers are full. It keeps minimal in-memory state
+/// and does the minimum filtering required to keep log files smaller.
+
+namespace __xray {
+
+XRayLogInitStatus basicLoggingInit(size_t BufferSize, size_t BufferMax,
+                                   void *Options, size_t OptionsSize);
+XRayLogInitStatus basicLoggingFinalize();
+
+void basicLoggingHandleArg0RealTSC(int32_t FuncId, XRayEntryType Entry);
+void basicLoggingHandleArg0EmulateTSC(int32_t FuncId, XRayEntryType Entry);
+void basicLoggingHandleArg1RealTSC(int32_t FuncId, XRayEntryType Entry,
+                                   uint64_t Arg1);
+void basicLoggingHandleArg1EmulateTSC(int32_t FuncId, XRayEntryType Entry,
+                                      uint64_t Arg1);
+XRayLogFlushStatus basicLoggingFlush();
+XRayLogInitStatus basicLoggingReset();
+
+} // namespace __xray
+
+#endif // XRAY_XRAY_INMEMORY_LOG_H
diff --git a/contrib/compiler-rt/lib/xray/xray_interface.cc b/contrib/compiler-rt/lib/xray/xray_interface.cc
index 694d34c0102b..766313e85c58 100644
--- a/contrib/compiler-rt/lib/xray/xray_interface.cc
+++ b/contrib/compiler-rt/lib/xray/xray_interface.cc
@@ -23,12 +23,15 @@
 
 #include "sanitizer_common/sanitizer_common.h"
 #include "xray_defs.h"
+#include "xray_flags.h"
+
+extern __sanitizer::SpinMutex XRayInstrMapMutex;
+extern __sanitizer::atomic_uint8_t XRayInitialized;
+extern __xray::XRaySledMap XRayInstrMap;
 
 namespace __xray {
 
 #if defined(__x86_64__)
-// FIXME: The actual length is 11 bytes. Why was length 12 passed to mprotect()
-// ?
 static const int16_t cSledLength = 12;
 #elif defined(__aarch64__)
 static const int16_t cSledLength = 32;
@@ -53,6 +56,10 @@ __sanitizer::atomic_uintptr_t XRayArgLogger{0};
 // This is the function to call when we encounter a custom event log call.
 __sanitizer::atomic_uintptr_t XRayPatchedCustomEvent{0};
 
+// This is the global status to determine whether we are currently
+// patching/unpatching.
+__sanitizer::atomic_uint8_t XRayPatching{0};
+
 // MProtectHelper is an RAII wrapper for calls to mprotect(...) that will undo
 // any successful mprotect(...) changes. This is used to make a page writeable
 // and executable, and upon destruction if it was successful in doing so returns
@@ -88,85 +95,10 @@ public:
   }
 };
 
-} // namespace __xray
-
-extern __sanitizer::SpinMutex XRayInstrMapMutex;
-extern __sanitizer::atomic_uint8_t XRayInitialized;
-extern __xray::XRaySledMap XRayInstrMap;
-
-int __xray_set_handler(void (*entry)(int32_t,
-                                     XRayEntryType)) XRAY_NEVER_INSTRUMENT {
-  if (__sanitizer::atomic_load(&XRayInitialized,
-                               __sanitizer::memory_order_acquire)) {
-
-    __sanitizer::atomic_store(&__xray::XRayPatchedFunction,
-                              reinterpret_cast<uintptr_t>(entry),
-                              __sanitizer::memory_order_release);
-    return 1;
-  }
-  return 0;
-}
-
-int __xray_set_customevent_handler(void (*entry)(void *, size_t))
-    XRAY_NEVER_INSTRUMENT {
-  if (__sanitizer::atomic_load(&XRayInitialized,
-                               __sanitizer::memory_order_acquire)) {
-    __sanitizer::atomic_store(&__xray::XRayPatchedCustomEvent,
-                              reinterpret_cast<uintptr_t>(entry),
-                              __sanitizer::memory_order_release);
-    return 1;
-  }
-  return 0;
-}
-
-
-int __xray_remove_handler() XRAY_NEVER_INSTRUMENT {
-  return __xray_set_handler(nullptr);
-}
-
-int __xray_remove_customevent_handler() XRAY_NEVER_INSTRUMENT {
-  return __xray_set_customevent_handler(nullptr);
-}
-
-__sanitizer::atomic_uint8_t XRayPatching{0};
-
-using namespace __xray;
-
-// FIXME: Figure out whether we can move this class to sanitizer_common instead
-// as a generic "scope guard".
-template <class Function> class CleanupInvoker {
-  Function Fn;
-
-public:
-  explicit CleanupInvoker(Function Fn) XRAY_NEVER_INSTRUMENT : Fn(Fn) {}
-  CleanupInvoker(const CleanupInvoker &) XRAY_NEVER_INSTRUMENT = default;
-  CleanupInvoker(CleanupInvoker &&) XRAY_NEVER_INSTRUMENT = default;
-  CleanupInvoker &
-  operator=(const CleanupInvoker &) XRAY_NEVER_INSTRUMENT = delete;
-  CleanupInvoker &operator=(CleanupInvoker &&) XRAY_NEVER_INSTRUMENT = delete;
-  ~CleanupInvoker() XRAY_NEVER_INSTRUMENT { Fn(); }
-};
-
-template <class Function>
-CleanupInvoker<Function> scopeCleanup(Function Fn) XRAY_NEVER_INSTRUMENT {
-  return CleanupInvoker<Function>{Fn};
-}
-
-inline bool patchSled(const XRaySledEntry &Sled, bool Enable,
-                      int32_t FuncId) XRAY_NEVER_INSTRUMENT {
-  // While we're here, we should patch the nop sled. To do that we mprotect
-  // the page containing the function to be writeable.
-  const uint64_t PageSize = GetPageSizeCached();
-  void *PageAlignedAddr =
-      reinterpret_cast<void *>(Sled.Address & ~(PageSize - 1));
-  std::size_t MProtectLen = (Sled.Address + cSledLength) -
-                            reinterpret_cast<uint64_t>(PageAlignedAddr);
-  MProtectHelper Protector(PageAlignedAddr, MProtectLen);
-  if (Protector.MakeWriteable() == -1) {
-    printf("Failed mprotect: %d\n", errno);
-    return XRayPatchingStatus::FAILED;
-  }
+namespace {
 
+bool patchSled(const XRaySledEntry &Sled, bool Enable,
+               int32_t FuncId) XRAY_NEVER_INSTRUMENT {
   bool Success = false;
   switch (Sled.Kind) {
   case XRayEntryType::ENTRY:
@@ -191,6 +123,55 @@ inline bool patchSled(const XRaySledEntry &Sled, bool Enable,
   return Success;
 }
 
+XRayPatchingStatus patchFunction(int32_t FuncId,
+                                 bool Enable) XRAY_NEVER_INSTRUMENT {
+  if (!__sanitizer::atomic_load(&XRayInitialized,
+                                __sanitizer::memory_order_acquire))
+    return XRayPatchingStatus::NOT_INITIALIZED; // Not initialized.
+
+  uint8_t NotPatching = false;
+  if (!__sanitizer::atomic_compare_exchange_strong(
+          &XRayPatching, &NotPatching, true, __sanitizer::memory_order_acq_rel))
+    return XRayPatchingStatus::ONGOING; // Already patching.
+
+  // Next, we look for the function index.
+  XRaySledMap InstrMap;
+  {
+    __sanitizer::SpinMutexLock Guard(&XRayInstrMapMutex);
+    InstrMap = XRayInstrMap;
+  }
+
+  // If we don't have an index, we can't patch individual functions.
+  if (InstrMap.Functions == 0)
+    return XRayPatchingStatus::NOT_INITIALIZED;
+
+  // FuncId must be a positive number, less than the number of functions
+  // instrumented.
+  if (FuncId <= 0 || static_cast<size_t>(FuncId) > InstrMap.Functions) {
+    Report("Invalid function id provided: %d\n", FuncId);
+    return XRayPatchingStatus::FAILED;
+  }
+
+  // Now we patch ths sleds for this specific function.
+  auto SledRange = InstrMap.SledsIndex[FuncId - 1];
+  auto *f = SledRange.Begin;
+  auto *e = SledRange.End;
+
+  bool SucceedOnce = false;
+  while (f != e)
+    SucceedOnce |= patchSled(*f++, Enable, FuncId);
+
+  __sanitizer::atomic_store(&XRayPatching, false,
+                            __sanitizer::memory_order_release);
+
+  if (!SucceedOnce) {
+    Report("Failed patching any sled for function '%d'.", FuncId);
+    return XRayPatchingStatus::FAILED;
+  }
+
+  return XRayPatchingStatus::SUCCESS;
+}
+
 // controlPatching implements the common internals of the patching/unpatching
 // implementation. |Enable| defines whether we're enabling or disabling the
 // runtime XRay instrumentation.
@@ -205,14 +186,13 @@ XRayPatchingStatus controlPatching(bool Enable) XRAY_NEVER_INSTRUMENT {
     return XRayPatchingStatus::ONGOING; // Already patching.
 
   uint8_t PatchingSuccess = false;
-  auto XRayPatchingStatusResetter = scopeCleanup([&PatchingSuccess] {
-    if (!PatchingSuccess)
-      __sanitizer::atomic_store(&XRayPatching, false,
-                                __sanitizer::memory_order_release);
-  });
+  auto XRayPatchingStatusResetter =
+      __sanitizer::at_scope_exit([&PatchingSuccess] {
+        if (!PatchingSuccess)
+          __sanitizer::atomic_store(&XRayPatching, false,
+                                    __sanitizer::memory_order_release);
+      });
 
-  // Step 1: Compute the function id, as a unique identifier per function in the
-  // instrumentation map.
   XRaySledMap InstrMap;
   {
     __sanitizer::SpinMutexLock Guard(&XRayInstrMapMutex);
@@ -221,16 +201,47 @@ XRayPatchingStatus controlPatching(bool Enable) XRAY_NEVER_INSTRUMENT {
   if (InstrMap.Entries == 0)
     return XRayPatchingStatus::NOT_INITIALIZED;
 
-  const uint64_t PageSize = GetPageSizeCached();
+  uint32_t FuncId = 1;
+  uint64_t CurFun = 0;
+
+  // First we want to find the bounds for which we have instrumentation points,
+  // and try to get as few calls to mprotect(...) as possible. We're assuming
+  // that all the sleds for the instrumentation map are contiguous as a single
+  // set of pages. When we do support dynamic shared object instrumentation,
+  // we'll need to do this for each set of page load offsets per DSO loaded. For
+  // now we're assuming we can mprotect the whole section of text between the
+  // minimum sled address and the maximum sled address (+ the largest sled
+  // size).
+  auto MinSled = InstrMap.Sleds[0];
+  auto MaxSled = InstrMap.Sleds[InstrMap.Entries - 1];
+  for (std::size_t I = 0; I < InstrMap.Entries; I++) {
+    const auto &Sled = InstrMap.Sleds[I];
+    if (Sled.Address < MinSled.Address)
+      MinSled = Sled;
+    if (Sled.Address > MaxSled.Address)
+      MaxSled = Sled;
+  }
+
+  const size_t PageSize = flags()->xray_page_size_override > 0
+                              ? flags()->xray_page_size_override
+                              : GetPageSizeCached();
   if ((PageSize == 0) || ((PageSize & (PageSize - 1)) != 0)) {
     Report("System page size is not a power of two: %lld\n", PageSize);
     return XRayPatchingStatus::FAILED;
   }
 
-  uint32_t FuncId = 1;
-  uint64_t CurFun = 0;
-  for (std::size_t I = 0; I < InstrMap.Entries; I++) {
-    auto Sled = InstrMap.Sleds[I];
+  void *PageAlignedAddr =
+      reinterpret_cast<void *>(MinSled.Address & ~(PageSize - 1));
+  size_t MProtectLen =
+      (MaxSled.Address - reinterpret_cast<uptr>(PageAlignedAddr)) + cSledLength;
+  MProtectHelper Protector(PageAlignedAddr, MProtectLen);
+  if (Protector.MakeWriteable() == -1) {
+    Report("Failed mprotect: %d\n", errno);
+    return XRayPatchingStatus::FAILED;
+  }
+
+  for (std::size_t I = 0; I < InstrMap.Entries; ++I) {
+    auto &Sled = InstrMap.Sleds[I];
     auto F = Sled.Function;
     if (CurFun == 0)
       CurFun = F;
@@ -246,36 +257,14 @@ XRayPatchingStatus controlPatching(bool Enable) XRAY_NEVER_INSTRUMENT {
   return XRayPatchingStatus::SUCCESS;
 }
 
-XRayPatchingStatus __xray_patch() XRAY_NEVER_INSTRUMENT {
-  return controlPatching(true);
-}
-
-XRayPatchingStatus __xray_unpatch() XRAY_NEVER_INSTRUMENT {
-  return controlPatching(false);
-}
-
-XRayPatchingStatus patchFunction(int32_t FuncId,
-                                 bool Enable) XRAY_NEVER_INSTRUMENT {
-  if (!__sanitizer::atomic_load(&XRayInitialized,
-                                __sanitizer::memory_order_acquire))
-    return XRayPatchingStatus::NOT_INITIALIZED; // Not initialized.
-
-  uint8_t NotPatching = false;
-  if (!__sanitizer::atomic_compare_exchange_strong(
-          &XRayPatching, &NotPatching, true, __sanitizer::memory_order_acq_rel))
-    return XRayPatchingStatus::ONGOING; // Already patching.
-
-  // Next, we look for the function index.
+XRayPatchingStatus mprotectAndPatchFunction(int32_t FuncId,
+                                            bool Enable) XRAY_NEVER_INSTRUMENT {
   XRaySledMap InstrMap;
   {
     __sanitizer::SpinMutexLock Guard(&XRayInstrMapMutex);
     InstrMap = XRayInstrMap;
   }
 
-  // If we don't have an index, we can't patch individual functions.
-  if (InstrMap.Functions == 0)
-    return XRayPatchingStatus::NOT_INITIALIZED;
-
   // FuncId must be a positive number, less than the number of functions
   // instrumented.
   if (FuncId <= 0 || static_cast<size_t>(FuncId) > InstrMap.Functions) {
@@ -283,33 +272,98 @@ XRayPatchingStatus patchFunction(int32_t FuncId,
     return XRayPatchingStatus::FAILED;
   }
 
-  // Now we patch ths sleds for this specific function.
+  const size_t PageSize = flags()->xray_page_size_override > 0
+                              ? flags()->xray_page_size_override
+                              : GetPageSizeCached();
+  if ((PageSize == 0) || ((PageSize & (PageSize - 1)) != 0)) {
+    Report("Provided page size is not a power of two: %lld\n", PageSize);
+    return XRayPatchingStatus::FAILED;
+  }
+
+  // Here we compute the minumum sled and maximum sled associated with a
+  // particular function ID.
   auto SledRange = InstrMap.SledsIndex[FuncId - 1];
   auto *f = SledRange.Begin;
   auto *e = SledRange.End;
+  auto MinSled = *f;
+  auto MaxSled = *(SledRange.End - 1);
+  while (f != e) {
+    if (f->Address < MinSled.Address)
+      MinSled = *f;
+    if (f->Address > MaxSled.Address)
+      MaxSled = *f;
+    ++f;
+  }
 
-  bool SucceedOnce = false;
-  while (f != e)
-    SucceedOnce |= patchSled(*f++, Enable, FuncId);
+  void *PageAlignedAddr =
+      reinterpret_cast<void *>(MinSled.Address & ~(PageSize - 1));
+  size_t MProtectLen =
+      (MaxSled.Address - reinterpret_cast<uptr>(PageAlignedAddr)) + cSledLength;
+  MProtectHelper Protector(PageAlignedAddr, MProtectLen);
+  if (Protector.MakeWriteable() == -1) {
+    Report("Failed mprotect: %d\n", errno);
+    return XRayPatchingStatus::FAILED;
+  }
+  return patchFunction(FuncId, Enable);
+}
 
-  __sanitizer::atomic_store(&XRayPatching, false,
-                            __sanitizer::memory_order_release);
+} // namespace
 
-  if (!SucceedOnce) {
-    Report("Failed patching any sled for function '%d'.", FuncId);
-    return XRayPatchingStatus::FAILED;
+} // namespace __xray
+
+using namespace __xray;
+
+// The following functions are declared `extern "C" {...}` in the header, hence
+// they're defined in the global namespace.
+
+int __xray_set_handler(void (*entry)(int32_t,
+                                     XRayEntryType)) XRAY_NEVER_INSTRUMENT {
+  if (__sanitizer::atomic_load(&XRayInitialized,
+                               __sanitizer::memory_order_acquire)) {
+
+    __sanitizer::atomic_store(&__xray::XRayPatchedFunction,
+                              reinterpret_cast<uintptr_t>(entry),
+                              __sanitizer::memory_order_release);
+    return 1;
   }
+  return 0;
+}
 
-  return XRayPatchingStatus::SUCCESS;
+int __xray_set_customevent_handler(void (*entry)(void *, size_t))
+    XRAY_NEVER_INSTRUMENT {
+  if (__sanitizer::atomic_load(&XRayInitialized,
+                               __sanitizer::memory_order_acquire)) {
+    __sanitizer::atomic_store(&__xray::XRayPatchedCustomEvent,
+                              reinterpret_cast<uintptr_t>(entry),
+                              __sanitizer::memory_order_release);
+    return 1;
+  }
+  return 0;
+}
+
+int __xray_remove_handler() XRAY_NEVER_INSTRUMENT {
+  return __xray_set_handler(nullptr);
+}
+
+int __xray_remove_customevent_handler() XRAY_NEVER_INSTRUMENT {
+  return __xray_set_customevent_handler(nullptr);
+}
+
+XRayPatchingStatus __xray_patch() XRAY_NEVER_INSTRUMENT {
+  return controlPatching(true);
+}
+
+XRayPatchingStatus __xray_unpatch() XRAY_NEVER_INSTRUMENT {
+  return controlPatching(false);
 }
 
 XRayPatchingStatus __xray_patch_function(int32_t FuncId) XRAY_NEVER_INSTRUMENT {
-  return patchFunction(FuncId, true);
+  return mprotectAndPatchFunction(FuncId, true);
 }
 
 XRayPatchingStatus
 __xray_unpatch_function(int32_t FuncId) XRAY_NEVER_INSTRUMENT {
-  return patchFunction(FuncId, false);
+  return mprotectAndPatchFunction(FuncId, false);
 }
 
 int __xray_set_handler_arg1(void (*entry)(int32_t, XRayEntryType, uint64_t)) {
@@ -331,7 +385,7 @@ uintptr_t __xray_function_address(int32_t FuncId) XRAY_NEVER_INSTRUMENT {
   __sanitizer::SpinMutexLock Guard(&XRayInstrMapMutex);
   if (FuncId <= 0 || static_cast<size_t>(FuncId) > XRayInstrMap.Functions)
     return 0;
-  return XRayInstrMap.SledsIndex[FuncId - 1].Begin->Address
+  return XRayInstrMap.SledsIndex[FuncId - 1].Begin->Function
 // On PPC, function entries are always aligned to 16 bytes. The beginning of a
 // sled might be a local entry, which is always +8 based on the global entry.
 // Always return the global entry.
diff --git a/contrib/compiler-rt/lib/xray/xray_interface_internal.h b/contrib/compiler-rt/lib/xray/xray_interface_internal.h
index 4a2784612fcb..5811e2b7300a 100644
--- a/contrib/compiler-rt/lib/xray/xray_interface_internal.h
+++ b/contrib/compiler-rt/lib/xray/xray_interface_internal.h
@@ -28,13 +28,15 @@ struct XRaySledEntry {
   uint64_t Function;
   unsigned char Kind;
   unsigned char AlwaysInstrument;
-  unsigned char Padding[14]; // Need 32 bytes
+  unsigned char Version;
+  unsigned char Padding[13]; // Need 32 bytes
 #elif SANITIZER_WORDSIZE == 32
   uint32_t Address;
   uint32_t Function;
   unsigned char Kind;
   unsigned char AlwaysInstrument;
-  unsigned char Padding[6]; // Need 16 bytes
+  unsigned char Version;
+  unsigned char Padding[5]; // Need 16 bytes
 #else
 #error "Unsupported word size."
 #endif
diff --git a/contrib/compiler-rt/lib/xray/xray_log_interface.cc b/contrib/compiler-rt/lib/xray/xray_log_interface.cc
index ee14ae4b1b62..783f004d292a 100644
--- a/contrib/compiler-rt/lib/xray/xray_log_interface.cc
+++ b/contrib/compiler-rt/lib/xray/xray_log_interface.cc
@@ -12,35 +12,84 @@
 //===----------------------------------------------------------------------===//
 #include "xray/xray_log_interface.h"
 
+#include "sanitizer_common/sanitizer_allocator_internal.h"
 #include "sanitizer_common/sanitizer_atomic.h"
 #include "sanitizer_common/sanitizer_mutex.h"
 #include "xray/xray_interface.h"
 #include "xray_defs.h"
 
-#include <memory>
-
 __sanitizer::SpinMutex XRayImplMutex;
-std::unique_ptr<XRayLogImpl> GlobalXRayImpl;
+XRayLogImpl CurrentXRayImpl{nullptr, nullptr, nullptr, nullptr};
+XRayLogImpl *GlobalXRayImpl = nullptr;
+
+// We use a linked list of Mode to XRayLogImpl mappings. This is a linked list
+// when it should be a map because we're avoiding having to depend on C++
+// standard library data structures at this level of the implementation.
+struct ModeImpl {
+  ModeImpl *Next;
+  const char *Mode;
+  XRayLogImpl Impl;
+};
+
+ModeImpl SentinelModeImpl{
+    nullptr, nullptr, {nullptr, nullptr, nullptr, nullptr}};
+ModeImpl *ModeImpls = &SentinelModeImpl;
+
+XRayLogRegisterStatus
+__xray_log_register_mode(const char *Mode,
+                         XRayLogImpl Impl) XRAY_NEVER_INSTRUMENT {
+  if (Impl.flush_log == nullptr || Impl.handle_arg0 == nullptr ||
+      Impl.log_finalize == nullptr || Impl.log_init == nullptr)
+    return XRayLogRegisterStatus::XRAY_INCOMPLETE_IMPL;
+
+  __sanitizer::SpinMutexLock Guard(&XRayImplMutex);
+  // First, look for whether the mode already has a registered implementation.
+  for (ModeImpl *it = ModeImpls; it != &SentinelModeImpl; it = it->Next) {
+    if (!__sanitizer::internal_strcmp(Mode, it->Mode))
+      return XRayLogRegisterStatus::XRAY_DUPLICATE_MODE;
+  }
+  auto *NewModeImpl =
+      static_cast<ModeImpl *>(__sanitizer::InternalAlloc(sizeof(ModeImpl)));
+  NewModeImpl->Next = ModeImpls;
+  NewModeImpl->Mode = __sanitizer::internal_strdup(Mode);
+  NewModeImpl->Impl = Impl;
+  ModeImpls = NewModeImpl;
+  return XRayLogRegisterStatus::XRAY_REGISTRATION_OK;
+}
+
+XRayLogRegisterStatus
+__xray_log_select_mode(const char *Mode) XRAY_NEVER_INSTRUMENT {
+  __sanitizer::SpinMutexLock Guard(&XRayImplMutex);
+  for (ModeImpl *it = ModeImpls; it != &SentinelModeImpl; it = it->Next) {
+    if (!__sanitizer::internal_strcmp(Mode, it->Mode)) {
+      CurrentXRayImpl = it->Impl;
+      GlobalXRayImpl = &CurrentXRayImpl;
+      __xray_set_handler(it->Impl.handle_arg0);
+      return XRayLogRegisterStatus::XRAY_REGISTRATION_OK;
+    }
+  }
+  return XRayLogRegisterStatus::XRAY_MODE_NOT_FOUND;
+}
 
 void __xray_set_log_impl(XRayLogImpl Impl) XRAY_NEVER_INSTRUMENT {
   if (Impl.log_init == nullptr || Impl.log_finalize == nullptr ||
       Impl.handle_arg0 == nullptr || Impl.flush_log == nullptr) {
     __sanitizer::SpinMutexLock Guard(&XRayImplMutex);
-    GlobalXRayImpl.reset();
+    GlobalXRayImpl = nullptr;
     __xray_remove_handler();
     __xray_remove_handler_arg1();
     return;
   }
 
   __sanitizer::SpinMutexLock Guard(&XRayImplMutex);
-  GlobalXRayImpl.reset(new XRayLogImpl);
-  *GlobalXRayImpl = Impl;
+  CurrentXRayImpl = Impl;
+  GlobalXRayImpl = &CurrentXRayImpl;
   __xray_set_handler(Impl.handle_arg0);
 }
 
 void __xray_remove_log_impl() XRAY_NEVER_INSTRUMENT {
   __sanitizer::SpinMutexLock Guard(&XRayImplMutex);
-  GlobalXRayImpl.reset();
+  GlobalXRayImpl = nullptr;
   __xray_remove_handler();
   __xray_remove_handler_arg1();
 }
diff --git a/contrib/compiler-rt/lib/xray/xray_trampoline_x86_64.S b/contrib/compiler-rt/lib/xray/xray_trampoline_x86_64.S
index b59eedc4bb1b..350afd9265fd 100644
--- a/contrib/compiler-rt/lib/xray/xray_trampoline_x86_64.S
+++ b/contrib/compiler-rt/lib/xray/xray_trampoline_x86_64.S
@@ -14,10 +14,13 @@
 //===----------------------------------------------------------------------===//
 
 #include "../builtins/assembly.h"
+#include "../sanitizer_common/sanitizer_asm.h"
+
+
 
 .macro SAVE_REGISTERS
 	subq $192, %rsp
-	.cfi_def_cfa_offset 200
+	CFI_DEF_CFA_OFFSET(200)
 	// At this point, the stack pointer should be aligned to an 8-byte boundary,
 	// because any call instructions that come after this will add another 8
 	// bytes and therefore align it to 16-bytes.
@@ -57,63 +60,82 @@
 	movq	8(%rsp), %r8
 	movq	0(%rsp), %r9
 	addq	$192, %rsp
-	.cfi_def_cfa_offset 8
+	CFI_DEF_CFA_OFFSET(8)
+.endm
+
+.macro ALIGNED_CALL_RAX
+	// Call the logging handler, after aligning the stack to a 16-byte boundary.
+	// The approach we're taking here uses additional stack space to stash the
+	// stack pointer twice before aligning the pointer to 16-bytes. If the stack
+	// was 8-byte aligned, it will become 16-byte aligned -- when restoring the
+	// pointer, we can always look -8 bytes from the current position to get
+	// either of the values we've stashed in the first place.
+	pushq %rsp
+	pushq (%rsp)
+	andq $-0x10, %rsp
+  callq *%rax
+	movq 8(%rsp), %rsp
 .endm
 
 	.text
+#if !defined(__APPLE__)
+	.section .text
+#else
+	.section __TEXT,__text
+#endif
 	.file "xray_trampoline_x86.S"
 
 //===----------------------------------------------------------------------===//
 
-	.globl __xray_FunctionEntry
+	.globl ASM_SYMBOL(__xray_FunctionEntry)
 	.align 16, 0x90
-	.type __xray_FunctionEntry,@function
-
-__xray_FunctionEntry:
-	.cfi_startproc
+	ASM_TYPE_FUNCTION(__xray_FunctionEntry)
+ASM_SYMBOL(__xray_FunctionEntry):
+	CFI_STARTPROC
 	SAVE_REGISTERS
 
 	// This load has to be atomic, it's concurrent with __xray_patch().
 	// On x86/amd64, a simple (type-aligned) MOV instruction is enough.
-	movq	_ZN6__xray19XRayPatchedFunctionE(%rip), %rax
+	movq	ASM_SYMBOL(_ZN6__xray19XRayPatchedFunctionE)(%rip), %rax
 	testq	%rax, %rax
 	je	.Ltmp0
 
 	// The patched function prolog puts its xray_instr_map index into %r10d.
 	movl	%r10d, %edi
 	xor	%esi,%esi
-	callq	*%rax
+	ALIGNED_CALL_RAX
+
 .Ltmp0:
 	RESTORE_REGISTERS
 	retq
-.Ltmp1:
-	.size __xray_FunctionEntry, .Ltmp1-__xray_FunctionEntry
-	.cfi_endproc
+	ASM_SIZE(__xray_FunctionEntry)
+	CFI_ENDPROC
 
 //===----------------------------------------------------------------------===//
 
-	.globl __xray_FunctionExit
+	.globl ASM_SYMBOL(__xray_FunctionExit)
 	.align 16, 0x90
-	.type __xray_FunctionExit,@function
-__xray_FunctionExit:
-	.cfi_startproc
+	ASM_TYPE_FUNCTION(__xray_FunctionExit)
+ASM_SYMBOL(__xray_FunctionExit):
+	CFI_STARTPROC
 	// Save the important registers first. Since we're assuming that this
 	// function is only jumped into, we only preserve the registers for
 	// returning.
 	subq	$56, %rsp
-	.cfi_def_cfa_offset 64
+	CFI_DEF_CFA_OFFSET(64)
 	movq  %rbp, 48(%rsp)
 	movupd	%xmm0, 32(%rsp)
 	movupd	%xmm1, 16(%rsp)
 	movq	%rax, 8(%rsp)
 	movq	%rdx, 0(%rsp)
-	movq	_ZN6__xray19XRayPatchedFunctionE(%rip), %rax
+	movq	ASM_SYMBOL(_ZN6__xray19XRayPatchedFunctionE)(%rip), %rax
 	testq %rax,%rax
 	je	.Ltmp2
 
 	movl	%r10d, %edi
 	movl	$1, %esi
-	callq	*%rax
+  ALIGNED_CALL_RAX
+
 .Ltmp2:
 	// Restore the important registers.
 	movq  48(%rsp), %rbp
@@ -122,111 +144,94 @@ __xray_FunctionExit:
 	movq	8(%rsp), %rax
 	movq	0(%rsp), %rdx
 	addq	$56, %rsp
-	.cfi_def_cfa_offset 8
+	CFI_DEF_CFA_OFFSET(8)
 	retq
-.Ltmp3:
-	.size __xray_FunctionExit, .Ltmp3-__xray_FunctionExit
-	.cfi_endproc
+	ASM_SIZE(__xray_FunctionExit)
+	CFI_ENDPROC
 
 //===----------------------------------------------------------------------===//
 
-	.global __xray_FunctionTailExit
+	.globl ASM_SYMBOL(__xray_FunctionTailExit)
 	.align 16, 0x90
-	.type __xray_FunctionTailExit,@function
-__xray_FunctionTailExit:
-	.cfi_startproc
-	// Save the important registers as in the entry trampoline, but indicate that
-	// this is an exit. In the future, we will introduce a new entry type that
-	// differentiates between a normal exit and a tail exit, but we'd have to do
-	// this and increment the version number for the header.
+	ASM_TYPE_FUNCTION(__xray_FunctionTailExit)
+ASM_SYMBOL(__xray_FunctionTailExit):
+	CFI_STARTPROC
 	SAVE_REGISTERS
 
-	movq	_ZN6__xray19XRayPatchedFunctionE(%rip), %rax
+	movq	ASM_SYMBOL(_ZN6__xray19XRayPatchedFunctionE)(%rip), %rax
 	testq %rax,%rax
 	je	.Ltmp4
 
 	movl	%r10d, %edi
-	movl	$1, %esi
-	callq	*%rax
+	movl	$2, %esi
+
+  ALIGNED_CALL_RAX
 
 .Ltmp4:
 	RESTORE_REGISTERS
 	retq
-.Ltmp5:
-	.size __xray_FunctionTailExit, .Ltmp5-__xray_FunctionTailExit
-	.cfi_endproc
+	ASM_SIZE(__xray_FunctionTailExit)
+	CFI_ENDPROC
 
 //===----------------------------------------------------------------------===//
 
-	.globl __xray_ArgLoggerEntry
+	.globl ASM_SYMBOL(__xray_ArgLoggerEntry)
 	.align 16, 0x90
-	.type __xray_ArgLoggerEntry,@function
-__xray_ArgLoggerEntry:
-	.cfi_startproc
+	ASM_TYPE_FUNCTION(__xray_ArgLoggerEntry)
+ASM_SYMBOL(__xray_ArgLoggerEntry):
+	CFI_STARTPROC
 	SAVE_REGISTERS
 
 	// Again, these function pointer loads must be atomic; MOV is fine.
-	movq	_ZN6__xray13XRayArgLoggerE(%rip), %rax
+	movq	ASM_SYMBOL(_ZN6__xray13XRayArgLoggerE)(%rip), %rax
 	testq	%rax, %rax
 	jne	.Larg1entryLog
 
 	// If [arg1 logging handler] not set, defer to no-arg logging.
-	movq	_ZN6__xray19XRayPatchedFunctionE(%rip), %rax
+	movq	ASM_SYMBOL(_ZN6__xray19XRayPatchedFunctionE)(%rip), %rax
 	testq	%rax, %rax
 	je	.Larg1entryFail
 
 .Larg1entryLog:
 
-  // First argument will become the third
+	// First argument will become the third
 	movq	%rdi, %rdx
 
-  // XRayEntryType::ENTRY into the second
-	xorq	%rsi, %rsi
+	// XRayEntryType::LOG_ARGS_ENTRY into the second
+	mov	$0x3, %esi
 
 	// 32-bit function ID becomes the first
 	movl	%r10d, %edi
-	callq	*%rax
+	ALIGNED_CALL_RAX
 
 .Larg1entryFail:
 	RESTORE_REGISTERS
 	retq
-
-.Larg1entryEnd:
-	.size __xray_ArgLoggerEntry, .Larg1entryEnd-__xray_ArgLoggerEntry
-	.cfi_endproc
+	ASM_SIZE(__xray_ArgLoggerEntry)
+	CFI_ENDPROC
 
 //===----------------------------------------------------------------------===//
 
-	.global __xray_CustomEvent
+	.global ASM_SYMBOL(__xray_CustomEvent)
 	.align 16, 0x90
-	.type __xray_CustomEvent,@function
-__xray_CustomEvent:
-  .cfi_startproc
-	subq $16, %rsp
-	.cfi_def_cfa_offset 24
-	movq %rbp, 8(%rsp)
-	movq %rax, 0(%rsp)
+	ASM_TYPE_FUNCTION(__xray_CustomEvent)
+ASM_SYMBOL(__xray_CustomEvent):
+	CFI_STARTPROC
+	SAVE_REGISTERS
 
 	// We take two arguments to this trampoline, which should be in rdi	and rsi
 	// already. We also make sure that we stash %rax because we use that register
 	// to call the logging handler.
-	movq _ZN6__xray22XRayPatchedCustomEventE(%rip), %rax
+	movq ASM_SYMBOL(_ZN6__xray22XRayPatchedCustomEventE)(%rip), %rax
 	testq %rax,%rax
 	je .LcustomEventCleanup
 
-	// At this point we know that rcx and rdx already has the data, so we just
-	// call the logging handler.
-  callq *%rax
+	ALIGNED_CALL_RAX
 
 .LcustomEventCleanup:
-	movq 0(%rsp), %rax
-	movq 8(%rsp), %rbp
-	addq $16, %rsp
-	.cfi_def_cfa_offset 8
+	RESTORE_REGISTERS
 	retq
-
-.Ltmp8:
-	.size __xray_CustomEvent, .Ltmp8-__xray_CustomEvent
-	.cfi_endproc
+	ASM_SIZE(__xray_CustomEvent)
+	CFI_ENDPROC
 
 NO_EXEC_STACK_DIRECTIVE
diff --git a/contrib/compiler-rt/lib/xray/xray_utils.cc b/contrib/compiler-rt/lib/xray/xray_utils.cc
index b9a38d1b98eb..cf800d3aeaf8 100644
--- a/contrib/compiler-rt/lib/xray/xray_utils.cc
+++ b/contrib/compiler-rt/lib/xray/xray_utils.cc
@@ -117,7 +117,8 @@ int getLogFD() XRAY_NEVER_INSTRUMENT {
            TmpFilename);
     return -1;
   }
-  Report("XRay: Log file in '%s'\n", TmpFilename);
+  if (__sanitizer::Verbosity())
+    Report("XRay: Log file in '%s'\n", TmpFilename);
 
   return Fd;
 }
diff --git a/contrib/compiler-rt/lib/xray/xray_x86_64.cc b/contrib/compiler-rt/lib/xray/xray_x86_64.cc
index e34806fa1cea..e17f00ac3a62 100644
--- a/contrib/compiler-rt/lib/xray/xray_x86_64.cc
+++ b/contrib/compiler-rt/lib/xray/xray_x86_64.cc
@@ -76,6 +76,7 @@ static constexpr uint8_t CallOpCode = 0xe8;
 static constexpr uint16_t MovR10Seq = 0xba41;
 static constexpr uint16_t Jmp9Seq = 0x09eb;
 static constexpr uint16_t Jmp20Seq = 0x14eb;
+static constexpr uint16_t Jmp15Seq = 0x0feb;
 static constexpr uint8_t JmpOpCode = 0xe9;
 static constexpr uint8_t RetOpCode = 0xc3;
 static constexpr uint16_t NopwSeq = 0x9066;
@@ -207,8 +208,10 @@ bool patchCustomEvent(const bool Enable, const uint32_t FuncId,
                       const XRaySledEntry &Sled) XRAY_NEVER_INSTRUMENT {
   // Here we do the dance of replacing the following sled:
   //
+  // In Version 0:
+  //
   // xray_sled_n:
-  //   jmp +19          // 2 bytes
+  //   jmp +20          // 2 bytes
   //   ...
   //
   // With the following:
@@ -216,24 +219,35 @@ bool patchCustomEvent(const bool Enable, const uint32_t FuncId,
   //   nopw             // 2 bytes*
   //   ...
   //
-  // We need to do this in the following order:
   //
-  // 1. Overwrite the 5-byte nop with the call (relative), where (relative) is
-  //    the relative offset to the __xray_CustomEvent trampoline.
-  // 2. Do a two-byte atomic write over the 'jmp +24' to turn it into a 'nopw'.
-  //    This allows us to "enable" this code once the changes have committed.
+  // The "unpatch" should just turn the 'nopw' back to a 'jmp +20'.
+  //
+  // ---
   //
-  // The "unpatch" should just turn the 'nopw' back to a 'jmp +24'.
+  // In Version 1:
+  //
+  //   The jump offset is now 15 bytes (0x0f), so when restoring the nopw back
+  //   to a jmp, use 15 bytes instead.
   //
   if (Enable) {
     std::atomic_store_explicit(
         reinterpret_cast<std::atomic<uint16_t> *>(Sled.Address), NopwSeq,
         std::memory_order_release);
   } else {
-    std::atomic_store_explicit(
-        reinterpret_cast<std::atomic<uint16_t> *>(Sled.Address), Jmp20Seq,
-        std::memory_order_release);
-  }
+    switch (Sled.Version) {
+    case 1:
+      std::atomic_store_explicit(
+          reinterpret_cast<std::atomic<uint16_t> *>(Sled.Address), Jmp15Seq,
+          std::memory_order_release);
+      break;
+    case 0:
+    default:
+      std::atomic_store_explicit(
+          reinterpret_cast<std::atomic<uint16_t> *>(Sled.Address), Jmp20Seq,
+          std::memory_order_release);
+      break;
+    }
+    }
   return false;
 }
 
@@ -244,9 +258,9 @@ bool probeRequiredCPUFeatures() XRAY_NEVER_INSTRUMENT {
 
   // We check whether rdtscp support is enabled. According to the x86_64 manual,
   // level should be set at 0x80000001, and we should have a look at bit 27 in
-  // EDX. That's 0x8000000 (or 1u << 26).
+  // EDX. That's 0x8000000 (or 1u << 27).
   __get_cpuid(0x80000001, &EAX, &EBX, &ECX, &EDX);
-  if (!(EDX & (1u << 26))) {
+  if (!(EDX & (1u << 27))) {
     Report("Missing rdtscp support.\n");
     return false;
   }