diff options
Diffstat (limited to 'tools')
229 files changed, 15365 insertions, 5834 deletions
diff --git a/tools/CMakeLists.txt b/tools/CMakeLists.txt index f50382b08977e..b654b8c5cb8e0 100644 --- a/tools/CMakeLists.txt +++ b/tools/CMakeLists.txt @@ -34,8 +34,10 @@ endif() # requires targets specified in DEPENDS to exist before the call to # ExternalProject_Add. add_llvm_tool_subdirectory(lto) +add_llvm_tool_subdirectory(gold) add_llvm_tool_subdirectory(llvm-ar) add_llvm_tool_subdirectory(llvm-config) +add_llvm_tool_subdirectory(llvm-lto) add_llvm_tool_subdirectory(llvm-profdata) # Projects supported via LLVM_EXTERNAL_*_SOURCE_DIR need to be explicitly @@ -49,4 +51,9 @@ add_llvm_external_project(lldb) # file as external projects. add_llvm_implicit_projects() +# Add subprojects specified using LLVM_EXTERNAL_PROJECTS +foreach(p ${LLVM_EXTERNAL_PROJECTS}) + add_llvm_external_project(${p}) +endforeach(p) + set(LLVM_COMMON_DEPENDS ${LLVM_COMMON_DEPENDS} PARENT_SCOPE) diff --git a/tools/Makefile b/tools/Makefile deleted file mode 100644 index 92d495451879b..0000000000000 --- a/tools/Makefile +++ /dev/null @@ -1,81 +0,0 @@ -##===- tools/Makefile --------------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := .. - -include $(LEVEL)/Makefile.config - -# Build clang if present. - -ifneq ($(CLANG_SRC_ROOT),) - OPTIONAL_PARALLEL_DIRS := $(CLANG_SRC_ROOT) -else - OPTIONAL_PARALLEL_DIRS := clang -endif - -# Build LLDB if present. Note LLDB must be built last as it depends on -# the wider LLVM infrastructure (including Clang). -OPTIONAL_DIRS := lldb - -# NOTE: The tools are organized into five groups of four consisting of one -# large and three small executables. This is done to minimize memory load -# in parallel builds. Please retain this ordering. -DIRS := llvm-config -PARALLEL_DIRS := opt llvm-as llvm-dis llc llvm-ar llvm-nm llvm-link \ - lli llvm-extract llvm-mc bugpoint llvm-bcanalyzer llvm-diff \ - llvm-objdump llvm-readobj llvm-rtdyld \ - llvm-dwarfdump llvm-cov llvm-size llvm-stress llvm-mcmarkup \ - llvm-profdata llvm-symbolizer obj2yaml yaml2obj llvm-c-test \ - llvm-cxxdump verify-uselistorder dsymutil llvm-pdbdump \ - llvm-split sancov llvm-dwp - -# If Intel JIT Events support is configured, build an extra tool to test it. -ifeq ($(USE_INTEL_JITEVENTS), 1) - PARALLEL_DIRS += llvm-jitlistener -endif - -# Let users override the set of tools to build from the command line. -ifdef ONLY_TOOLS - OPTIONAL_PARALLEL_DIRS := - OPTIONAL_DIRS := $(findstring lldb,$(ONLY_TOOLS)) - PARALLEL_DIRS := $(filter-out lldb,$(ONLY_TOOLS)) -endif - -# These libraries build as dynamic libraries (.dylib /.so), they can only be -# built if ENABLE_PIC is set. -ifndef ONLY_TOOLS -ifeq ($(ENABLE_PIC),1) - # gold only builds if binutils is around. It requires "lto" to build before - # it so it is added to DIRS. llvm-lto also requires lto - DIRS += lto llvm-lto - ifdef BINUTILS_INCDIR - DIRS += gold - endif - - PARALLEL_DIRS += bugpoint-passes -endif - -ifdef LLVM_HAS_POLLY - PARALLEL_DIRS += polly -endif -endif - -# On Win32, loadable modules can be built with ENABLE_SHARED. -ifneq ($(ENABLE_SHARED),1) - ifneq (,$(filter $(HOST_OS), Cygwin MingW)) - PARALLEL_DIRS := $(filter-out bugpoint-passes, \ - $(PARALLEL_DIRS)) - endif -endif - -ifneq (,$(filter go,$(BINDINGS_TO_BUILD))) - PARALLEL_DIRS += llvm-go -endif - -include $(LEVEL)/Makefile.common diff --git a/tools/bugpoint-passes/Makefile b/tools/bugpoint-passes/Makefile deleted file mode 100644 index 61f96bc338590..0000000000000 --- a/tools/bugpoint-passes/Makefile +++ /dev/null @@ -1,23 +0,0 @@ -##===- tools/bugpoint-passes/Makefile -- -------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -LIBRARYNAME := BugpointPasses -LOADABLE_MODULE := 1 -USEDLIBS := - -# If we don't need RTTI or EH, there's no reason to export anything -# from this plugin. -ifneq ($(REQUIRES_RTTI), 1) -ifneq ($(REQUIRES_EH), 1) -EXPORTED_SYMBOL_FILE = $(PROJ_SRC_DIR)/bugpoint.exports -endif -endif - -include $(LEVEL)/Makefile.common diff --git a/tools/bugpoint/BugDriver.h b/tools/bugpoint/BugDriver.h index 20efff3fda5f7..52ec2c0056496 100644 --- a/tools/bugpoint/BugDriver.h +++ b/tools/bugpoint/BugDriver.h @@ -76,7 +76,7 @@ public: // command line arguments into instance variables of BugDriver. // bool addSources(const std::vector<std::string> &FileNames); - void addPass(std::string p) { PassesToRun.push_back(p); } + void addPass(std::string p) { PassesToRun.push_back(std::move(p)); } void setPassesToRun(const std::vector<std::string> &PTR) { PassesToRun = PTR; } @@ -130,12 +130,6 @@ public: /// bool isExecutingJIT(); - /// runPasses - Run all of the passes in the "PassesToRun" list, discard the - /// output, and return true if any of the passes crashed. - bool runPasses(Module *M) const { - return runPasses(M, PassesToRun); - } - Module *getProgram() const { return Program; } /// swapProgramIn - Set the current module to the specified module, returning @@ -183,7 +177,7 @@ public: /// Error. /// std::string executeProgramSafely(const Module *Program, - std::string OutputFile, + const std::string &OutputFile, std::string *Error) const; /// createReferenceFile - calls compileProgram and then records the output @@ -243,12 +237,8 @@ public: /// Carefully run the specified set of pass on the specified/ module, /// returning the transformed module on success, or a null pointer on failure. - /// If AutoDebugCrashes is set to true, then bugpoint will automatically - /// attempt to track down a crashing pass if one exists, and this method will - /// never return null. std::unique_ptr<Module> runPassesOn(Module *M, const std::vector<std::string> &Passes, - bool AutoDebugCrashes = false, unsigned NumExtraArgs = 0, const char *const *ExtraArgs = nullptr); @@ -266,6 +256,16 @@ public: std::string &OutputFilename, bool DeleteOutput = false, bool Quiet = false, unsigned NumExtraArgs = 0, const char * const *ExtraArgs = nullptr) const; + + /// runPasses - Just like the method above, but this just returns true or + /// false indicating whether or not the optimizer crashed on the specified + /// input (true = crashed). Does not produce any output. + /// + bool runPasses(Module *M, + const std::vector<std::string> &PassesToRun) const { + std::string Filename; + return runPasses(M, PassesToRun, Filename, true); + } /// runManyPasses - Take the specified pass list and create different /// combinations of passes to compile the program with. Compile the program with @@ -285,17 +285,6 @@ public: const Module *M) const; private: - /// runPasses - Just like the method above, but this just returns true or - /// false indicating whether or not the optimizer crashed on the specified - /// input (true = crashed). - /// - bool runPasses(Module *M, - const std::vector<std::string> &PassesToRun, - bool DeleteOutput = true) const { - std::string Filename; - return runPasses(M, PassesToRun, Filename, DeleteOutput); - } - /// initializeExecutionEnvironment - This method is used to set up the /// environment for executing LLVM programs. /// diff --git a/tools/bugpoint/CrashDebugger.cpp b/tools/bugpoint/CrashDebugger.cpp index 6cdc43ab8699c..b25e6ecec19b6 100644 --- a/tools/bugpoint/CrashDebugger.cpp +++ b/tools/bugpoint/CrashDebugger.cpp @@ -54,6 +54,9 @@ namespace { cl::opt<bool> NoNamedMDRM("disable-namedmd-remove", cl::desc("Do not remove global named metadata"), cl::init(false)); + cl::opt<bool> VerboseErrors("verbose-errors", + cl::desc("Print the output of crashing program"), + cl::init(false)); } namespace llvm { @@ -164,6 +167,7 @@ ReduceCrashingGlobalVariables::TestGlobalVariables( if (I.hasInitializer() && !GVSet.count(&I)) { DeleteGlobalInitializer(&I); I.setLinkage(GlobalValue::ExternalLinkage); + I.setComdat(nullptr); } // Try running the hacked up program... @@ -264,8 +268,8 @@ bool ReduceCrashingFunctions::TestFuncs(std::vector<Function*> &Funcs) { std::vector<GlobalValue*> ToRemove; // First, remove aliases to functions we're about to purge. for (GlobalAlias &Alias : M->aliases()) { - Constant *Root = Alias.getAliasee()->stripPointerCasts(); - Function *F = dyn_cast<Function>(Root); + GlobalObject *Root = Alias.getBaseObject(); + Function *F = dyn_cast_or_null<Function>(Root); if (F) { if (Functions.count(F)) // We're keeping this function. @@ -373,9 +377,9 @@ bool ReduceCrashingBlocks::TestBlocks(std::vector<const BasicBlock*> &BBs) { (*SI)->removePredecessor(&*BB); TerminatorInst *BBTerm = BB->getTerminator(); - if (BBTerm->isEHPad()) + if (BBTerm->isEHPad() || BBTerm->getType()->isTokenTy()) continue; - if (!BBTerm->getType()->isVoidTy() && !BBTerm->getType()->isTokenTy()) + if (!BBTerm->getType()->isVoidTy()) BBTerm->replaceAllUsesWith(Constant::getNullValue(BBTerm->getType())); // Replace the old terminator instruction. @@ -459,7 +463,7 @@ bool ReduceCrashingInstructions::TestInsts(std::vector<const Instruction*> Module *M = CloneModule(BD.getProgram(), VMap).release(); // Convert list to set for fast lookup... - SmallPtrSet<Instruction*, 64> Instructions; + SmallPtrSet<Instruction*, 32> Instructions; for (unsigned i = 0, e = Insts.size(); i != e; ++i) { assert(!isa<TerminatorInst>(Insts[i])); Instructions.insert(cast<Instruction>(VMap[Insts[i]])); @@ -476,8 +480,8 @@ bool ReduceCrashingInstructions::TestInsts(std::vector<const Instruction*> for (BasicBlock::iterator I = FI->begin(), E = FI->end(); I != E;) { Instruction *Inst = &*I++; if (!Instructions.count(Inst) && !isa<TerminatorInst>(Inst) && - !Inst->isEHPad()) { - if (!Inst->getType()->isVoidTy() && !Inst->getType()->isTokenTy()) + !Inst->isEHPad() && !Inst->getType()->isTokenTy()) { + if (!Inst->getType()->isVoidTy()) Inst->replaceAllUsesWith(UndefValue::get(Inst->getType())); Inst->eraseFromParent(); } @@ -552,7 +556,9 @@ bool ReduceCrashingNamedMD::TestNamedMDs(std::vector<std::string> &NamedMDs) { std::vector<NamedMDNode *> ToDelete; ToDelete.reserve(M->named_metadata_size() - Names.size()); for (auto &NamedMD : M->named_metadata()) - if (!Names.count(NamedMD.getName())) + // Always keep a nonempty llvm.dbg.cu because the Verifier would complain. + if (!Names.count(NamedMD.getName()) && + (!(NamedMD.getName() == "llvm.dbg.cu" && NamedMD.getNumOperands() > 0))) ToDelete.push_back(&NamedMD); for (auto *NamedMD : ToDelete) @@ -600,7 +606,7 @@ public: bool ReduceCrashingNamedMDOps::TestNamedMDOps( std::vector<const MDNode *> &NamedMDOps) { // Convert list to set for fast lookup... - SmallPtrSet<const MDNode *, 64> OldMDNodeOps; + SmallPtrSet<const MDNode *, 32> OldMDNodeOps; for (unsigned i = 0, e = NamedMDOps.size(); i != e; ++i) { OldMDNodeOps.insert(NamedMDOps[i]); } @@ -637,7 +643,7 @@ bool ReduceCrashingNamedMDOps::TestNamedMDOps( // module, and that they don't include any deleted blocks. NamedMDOps.clear(); for (const MDNode *Node : OldMDNodeOps) - NamedMDOps.push_back(cast<MDNode>(VMap.MD()[Node].get())); + NamedMDOps.push_back(cast<MDNode>(*VMap.getMappedMD(Node))); BD.setNewProgram(M); // It crashed, keep the trimmed version... return true; @@ -646,16 +652,10 @@ bool ReduceCrashingNamedMDOps::TestNamedMDOps( return false; } -/// DebugACrash - Given a predicate that determines whether a component crashes -/// on a program, try to destructively reduce the program while still keeping -/// the predicate true. -static bool DebugACrash(BugDriver &BD, - bool (*TestFn)(const BugDriver &, Module *), - std::string &Error) { - // See if we can get away with nuking some of the global variable initializers - // in the program... - if (!NoGlobalRM && - BD.getProgram()->global_begin() != BD.getProgram()->global_end()) { +static void ReduceGlobalInitializers(BugDriver &BD, + bool (*TestFn)(const BugDriver &, Module *), + std::string &Error) { + if (BD.getProgram()->global_begin() != BD.getProgram()->global_end()) { // Now try to reduce the number of global variable initializers in the // module to something small. Module *M = CloneModule(BD.getProgram()).release(); @@ -666,6 +666,7 @@ static bool DebugACrash(BugDriver &BD, if (I->hasInitializer()) { DeleteGlobalInitializer(&*I); I->setLinkage(GlobalValue::ExternalLinkage); + I->setComdat(nullptr); DeletedInit = true; } @@ -695,8 +696,7 @@ static bool DebugACrash(BugDriver &BD, unsigned OldSize = GVs.size(); ReduceCrashingGlobalVariables(BD, TestFn).reduceList(GVs, Error); - if (!Error.empty()) - return true; + assert(!Error.empty()); if (GVs.size() < OldSize) BD.EmitProgressBitcode(BD.getProgram(), "reduced-global-variables"); @@ -704,40 +704,11 @@ static bool DebugACrash(BugDriver &BD, } } } +} - // Now try to reduce the number of functions in the module to something small. - std::vector<Function*> Functions; - for (Function &F : *BD.getProgram()) - if (!F.isDeclaration()) - Functions.push_back(&F); - - if (Functions.size() > 1 && !BugpointIsInterrupted) { - outs() << "\n*** Attempting to reduce the number of functions " - "in the testcase\n"; - - unsigned OldSize = Functions.size(); - ReduceCrashingFunctions(BD, TestFn).reduceList(Functions, Error); - - if (Functions.size() < OldSize) - BD.EmitProgressBitcode(BD.getProgram(), "reduced-function"); - } - - // Attempt to delete entire basic blocks at a time to speed up - // convergence... this actually works by setting the terminator of the blocks - // to a return instruction then running simplifycfg, which can potentially - // shrinks the code dramatically quickly - // - if (!DisableSimplifyCFG && !BugpointIsInterrupted) { - std::vector<const BasicBlock*> Blocks; - for (Function &F : *BD.getProgram()) - for (BasicBlock &BB : F) - Blocks.push_back(&BB); - unsigned OldSize = Blocks.size(); - ReduceCrashingBlocks(BD, TestFn).reduceList(Blocks, Error); - if (Blocks.size() < OldSize) - BD.EmitProgressBitcode(BD.getProgram(), "reduced-blocks"); - } - +static void ReduceInsts(BugDriver &BD, + bool (*TestFn)(const BugDriver &, Module *), + std::string &Error) { // Attempt to delete instructions using bisection. This should help out nasty // cases with large basic blocks where the problem is at one end. if (!BugpointIsInterrupted) { @@ -751,11 +722,10 @@ static bool DebugACrash(BugDriver &BD, ReduceCrashingInstructions(BD, TestFn).reduceList(Insts, Error); } - // FIXME: This should use the list reducer to converge faster by deleting - // larger chunks of instructions at a time! unsigned Simplification = 2; do { - if (BugpointIsInterrupted) break; + if (BugpointIsInterrupted) + return; --Simplification; outs() << "\n*** Attempting to reduce testcase by deleting instruc" << "tions: Simplification Level #" << Simplification << '\n'; @@ -784,7 +754,8 @@ static bool DebugACrash(BugDriver &BD, if (InstructionsToSkipBeforeDeleting) { --InstructionsToSkipBeforeDeleting; } else { - if (BugpointIsInterrupted) goto ExitLoops; + if (BugpointIsInterrupted) + return; if (I->isEHPad() || I->getType()->isTokenTy()) continue; @@ -810,10 +781,60 @@ static bool DebugACrash(BugDriver &BD, } } while (Simplification); + BD.EmitProgressBitcode(BD.getProgram(), "reduced-instructions"); +} - if (!NoNamedMDRM) { - BD.EmitProgressBitcode(BD.getProgram(), "reduced-instructions"); +/// DebugACrash - Given a predicate that determines whether a component crashes +/// on a program, try to destructively reduce the program while still keeping +/// the predicate true. +static bool DebugACrash(BugDriver &BD, + bool (*TestFn)(const BugDriver &, Module *), + std::string &Error) { + // See if we can get away with nuking some of the global variable initializers + // in the program... + if (!NoGlobalRM) + ReduceGlobalInitializers(BD, TestFn, Error); + + // Now try to reduce the number of functions in the module to something small. + std::vector<Function*> Functions; + for (Function &F : *BD.getProgram()) + if (!F.isDeclaration()) + Functions.push_back(&F); + + if (Functions.size() > 1 && !BugpointIsInterrupted) { + outs() << "\n*** Attempting to reduce the number of functions " + "in the testcase\n"; + + unsigned OldSize = Functions.size(); + ReduceCrashingFunctions(BD, TestFn).reduceList(Functions, Error); + + if (Functions.size() < OldSize) + BD.EmitProgressBitcode(BD.getProgram(), "reduced-function"); + } + + // Attempt to delete entire basic blocks at a time to speed up + // convergence... this actually works by setting the terminator of the blocks + // to a return instruction then running simplifycfg, which can potentially + // shrinks the code dramatically quickly + // + if (!DisableSimplifyCFG && !BugpointIsInterrupted) { + std::vector<const BasicBlock*> Blocks; + for (Function &F : *BD.getProgram()) + for (BasicBlock &BB : F) + Blocks.push_back(&BB); + unsigned OldSize = Blocks.size(); + ReduceCrashingBlocks(BD, TestFn).reduceList(Blocks, Error); + if (Blocks.size() < OldSize) + BD.EmitProgressBitcode(BD.getProgram(), "reduced-blocks"); + } + + // Attempt to delete instructions using bisection. This should help out nasty + // cases with large basic blocks where the problem is at one end. + if (!BugpointIsInterrupted) + ReduceInsts(BD, TestFn, Error); + + if (!NoNamedMDRM) { if (!BugpointIsInterrupted) { // Try to reduce the amount of global metadata (particularly debug info), // by dropping global named metadata that anchors them @@ -833,10 +854,9 @@ static bool DebugACrash(BugDriver &BD, NamedMDOps.push_back(op); ReduceCrashingNamedMDOps(BD, TestFn).reduceList(NamedMDOps, Error); } + BD.EmitProgressBitcode(BD.getProgram(), "reduced-named-md"); } -ExitLoops: - // Try to clean up the testcase by running funcresolve and globaldce... if (!BugpointIsInterrupted) { outs() << "\n*** Attempting to perform final cleanups: "; @@ -857,7 +877,7 @@ ExitLoops: } static bool TestForOptimizerCrash(const BugDriver &BD, Module *M) { - return BD.runPasses(M); + return BD.runPasses(M, BD.getPassesToRun()); } /// debugOptimizerCrash - This method is called when some pass crashes on input. @@ -888,7 +908,10 @@ static bool TestForCodeGenCrash(const BugDriver &BD, Module *M) { std::string Error; BD.compileProgram(M, &Error); if (!Error.empty()) { - errs() << "<crash>\n"; + if (VerboseErrors) + errs() << Error << "\n"; + else + errs() << "<crash>\n"; return true; // Tool is still crashing. } errs() << '\n'; diff --git a/tools/bugpoint/ExecutionDriver.cpp b/tools/bugpoint/ExecutionDriver.cpp index 41b8ccc18e871..ab9c05fa924af 100644 --- a/tools/bugpoint/ExecutionDriver.cpp +++ b/tools/bugpoint/ExecutionDriver.cpp @@ -384,7 +384,7 @@ std::string BugDriver::executeProgram(const Module *Program, /// backend, if reference output is not provided. /// std::string BugDriver::executeProgramSafely(const Module *Program, - std::string OutputFile, + const std::string &OutputFile, std::string *Error) const { return executeProgram(Program, OutputFile, "", "", SafeInterpreter, Error); } diff --git a/tools/bugpoint/ExtractFunction.cpp b/tools/bugpoint/ExtractFunction.cpp index fe0ab69dc162d..de596a57d83cd 100644 --- a/tools/bugpoint/ExtractFunction.cpp +++ b/tools/bugpoint/ExtractFunction.cpp @@ -215,6 +215,8 @@ void llvm::DeleteGlobalInitializer(GlobalVariable *GV) { // void llvm::DeleteFunctionBody(Function *F) { eliminateAliases(F); + // Function declarations can't have comdats. + F->setComdat(nullptr); // delete the body of the function... F->deleteBody(); @@ -409,7 +411,7 @@ BugDriver::extractMappedBlocksFromModule(const std::vector<BasicBlock *> &BBs, std::vector<std::string> PI; PI.push_back("extract-blocks"); - std::unique_ptr<Module> Ret = runPassesOn(M, PI, false, 1, &ExtraArg); + std::unique_ptr<Module> Ret = runPassesOn(M, PI, 1, &ExtraArg); sys::fs::remove(Filename.c_str()); diff --git a/tools/bugpoint/Makefile b/tools/bugpoint/Makefile deleted file mode 100644 index 174f8d25161ee..0000000000000 --- a/tools/bugpoint/Makefile +++ /dev/null @@ -1,18 +0,0 @@ -##===- tools/bugpoint/Makefile -----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := bugpoint -LINK_COMPONENTS := asmparser instrumentation scalaropts ipo linker bitreader \ - bitwriter irreader vectorize objcarcopts codegen - -# Support plugins. -NO_DEAD_STRIP := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/bugpoint/Miscompilation.cpp b/tools/bugpoint/Miscompilation.cpp index 16919f5a44921..40955ed3caa93 100644 --- a/tools/bugpoint/Miscompilation.cpp +++ b/tools/bugpoint/Miscompilation.cpp @@ -26,12 +26,13 @@ #include "llvm/Support/CommandLine.h" #include "llvm/Support/FileUtilities.h" #include "llvm/Transforms/Utils/Cloning.h" + using namespace llvm; namespace llvm { extern cl::opt<std::string> OutputPrefix; extern cl::list<std::string> InputArgv; -} +} // end namespace llvm namespace { static llvm::cl::opt<bool> @@ -52,7 +53,7 @@ namespace { std::vector<std::string> &Suffix, std::string &Error) override; }; -} +} // end anonymous namespace /// TestResult - After passes have been split into a test group and a control /// group, see if they still break the program. @@ -208,7 +209,7 @@ namespace { bool TestFuncs(const std::vector<Function*> &Prefix, std::string &Error); }; -} +} // end anonymous namespace /// Given two modules, link them together and run the program, checking to see /// if the program matches the diff. If there is an error, return NULL. If not, @@ -469,7 +470,7 @@ namespace { bool TestFuncs(const std::vector<BasicBlock*> &BBs, std::string &Error); }; -} +} // end anonymous namespace /// TestFuncs - Extract all blocks for the miscompiled functions except for the /// specified blocks. If the problem still exists, return true. @@ -696,8 +697,14 @@ static bool TestOptimizer(BugDriver &BD, std::unique_ptr<Module> Test, // of the functions being tested. outs() << " Optimizing functions being tested: "; std::unique_ptr<Module> Optimized = - BD.runPassesOn(Test.get(), BD.getPassesToRun(), - /*AutoDebugCrashes*/ true); + BD.runPassesOn(Test.get(), BD.getPassesToRun()); + if (!Optimized) { + errs() << " Error running this sequence of passes" + << " on the input program!\n"; + delete BD.swapProgramIn(Test.get()); + BD.EmitProgressBitcode(Test.get(), "pass-error", false); + return BD.debugOptimizerCrash(); + } outs() << "done.\n"; outs() << " Checking to see if the merged program executes correctly: "; @@ -712,7 +719,6 @@ static bool TestOptimizer(BugDriver &BD, std::unique_ptr<Module> Test, return Broken; } - /// debugMiscompilation - This method is used when the passes selected are not /// crashing, but the generated output is semantically different from the /// input. @@ -752,8 +758,6 @@ void BugDriver::debugMiscompilation(std::string *Error) { outs() << " Portion that is input to optimizer: "; EmitProgressBitcode(ToOptimize, "tooptimize"); delete ToOptimize; // Delete hacked module. - - return; } /// Get the specified modules ready for code generator testing. @@ -984,7 +988,6 @@ static bool TestCodeGenerator(BugDriver &BD, std::unique_ptr<Module> Test, return Result; } - /// debugCodeGenerator - debug errors in LLC, LLI, or CBE. /// bool BugDriver::debugCodeGenerator(std::string *Error) { diff --git a/tools/bugpoint/OptimizerDriver.cpp b/tools/bugpoint/OptimizerDriver.cpp index 344e7b588fb86..2cc2f4471a581 100644 --- a/tools/bugpoint/OptimizerDriver.cpp +++ b/tools/bugpoint/OptimizerDriver.cpp @@ -267,18 +267,11 @@ bool BugDriver::runPasses(Module *Program, std::unique_ptr<Module> BugDriver::runPassesOn(Module *M, const std::vector<std::string> &Passes, - bool AutoDebugCrashes, unsigned NumExtraArgs, + unsigned NumExtraArgs, const char *const *ExtraArgs) { std::string BitcodeResult; if (runPasses(M, Passes, BitcodeResult, false/*delete*/, true/*quiet*/, NumExtraArgs, ExtraArgs)) { - if (AutoDebugCrashes) { - errs() << " Error running this sequence of passes" - << " on the input program!\n"; - delete swapProgramIn(M); - EmitProgressBitcode(M, "pass-error", false); - exit(debugOptimizerCrash()); - } return nullptr; } diff --git a/tools/bugpoint/ToolRunner.cpp b/tools/bugpoint/ToolRunner.cpp index 2ccd649051289..4af4b10c4d78a 100644 --- a/tools/bugpoint/ToolRunner.cpp +++ b/tools/bugpoint/ToolRunner.cpp @@ -21,6 +21,7 @@ #include "llvm/Support/raw_ostream.h" #include <fstream> #include <sstream> +#include <utility> using namespace llvm; #define DEBUG_TYPE "toolrunner" @@ -272,9 +273,9 @@ namespace { std::string CompilerCommand; std::vector<std::string> CompilerArgs; public: - CustomCompiler( - const std::string &CompilerCmd, std::vector<std::string> CompArgs) : - CompilerCommand(CompilerCmd), CompilerArgs(CompArgs) {} + CustomCompiler(const std::string &CompilerCmd, + std::vector<std::string> CompArgs) + : CompilerCommand(CompilerCmd), CompilerArgs(std::move(CompArgs)) {} void compileProgram(const std::string &Bitcode, std::string *Error, @@ -333,9 +334,9 @@ namespace { std::string ExecutionCommand; std::vector<std::string> ExecutorArgs; public: - CustomExecutor( - const std::string &ExecutionCmd, std::vector<std::string> ExecArgs) : - ExecutionCommand(ExecutionCmd), ExecutorArgs(ExecArgs) {} + CustomExecutor(const std::string &ExecutionCmd, + std::vector<std::string> ExecArgs) + : ExecutionCommand(ExecutionCmd), ExecutorArgs(std::move(ExecArgs)) {} int ExecuteProgram(const std::string &Bitcode, const std::vector<std::string> &Args, diff --git a/tools/bugpoint/bugpoint.cpp b/tools/bugpoint/bugpoint.cpp index 48f30e6709f80..28565f1daac33 100644 --- a/tools/bugpoint/bugpoint.cpp +++ b/tools/bugpoint/bugpoint.cpp @@ -113,7 +113,7 @@ void initializePollyPasses(llvm::PassRegistry &Registry); int main(int argc, char **argv) { #ifndef DEBUG_BUGPOINT - llvm::sys::PrintStackTraceOnErrorSignal(); + llvm::sys::PrintStackTraceOnErrorSignal(argv[0]); llvm::PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. #endif @@ -143,7 +143,7 @@ int main(int argc, char **argv) { sys::SetInterruptFunction(BugpointInterruptFunction); #endif - LLVMContext& Context = getGlobalContext(); + LLVMContext Context; // If we have an override, set it and then track the triple we want Modules // to use. if (!OverrideTriple.empty()) { diff --git a/tools/dsymutil/BinaryHolder.cpp b/tools/dsymutil/BinaryHolder.cpp index 4c2c1d195c3ca..e45f7fefe0844 100644 --- a/tools/dsymutil/BinaryHolder.cpp +++ b/tools/dsymutil/BinaryHolder.cpp @@ -19,15 +19,6 @@ namespace llvm { namespace dsymutil { -Triple BinaryHolder::getTriple(const object::MachOObjectFile &Obj) { - // If a ThumbTriple is returned, use it instead of the standard - // one. This is because the thumb triple always allows to create a - // target, whereas the non-thumb one might not. - Triple ThumbTriple; - Triple T = Obj.getArch(nullptr, &ThumbTriple); - return ThumbTriple.getArch() ? ThumbTriple : T; -} - static std::vector<MemoryBufferRef> getMachOFatMemoryBuffers(StringRef Filename, MemoryBuffer &Mem, object::MachOUniversalBinary &Fat) { @@ -82,13 +73,15 @@ BinaryHolder::GetMemoryBuffersForFile(StringRef Filename, auto ErrOrFat = object::MachOUniversalBinary::create( CurrentMemoryBuffer->getMemBufferRef()); - if (ErrOrFat.getError()) { + if (!ErrOrFat) { + consumeError(ErrOrFat.takeError()); // Not a fat binary must be a standard one. Return a one element vector. return std::vector<MemoryBufferRef>{CurrentMemoryBuffer->getMemBufferRef()}; } CurrentFatBinary = std::move(*ErrOrFat); - return getMachOFatMemoryBuffers(Filename, *CurrentMemoryBuffer, + CurrentFatBinaryName = Filename; + return getMachOFatMemoryBuffers(CurrentFatBinaryName, *CurrentMemoryBuffer, *CurrentFatBinary); } @@ -109,10 +102,8 @@ BinaryHolder::GetArchiveMemberBuffers(StringRef Filename, Buffers.reserve(CurrentArchives.size()); for (const auto &CurrentArchive : CurrentArchives) { - for (auto ChildOrErr : CurrentArchive->children()) { - if (std::error_code Err = ChildOrErr.getError()) - return Err; - const auto &Child = *ChildOrErr; + Error Err; + for (auto Child : CurrentArchive->children(Err)) { if (auto NameOrErr = Child.getName()) { if (*NameOrErr == Filename) { if (Timestamp != sys::TimeValue::PosixZeroTime() && @@ -130,6 +121,8 @@ BinaryHolder::GetArchiveMemberBuffers(StringRef Filename, } } } + if (Err) + return errorToErrorCode(std::move(Err)); } if (Buffers.empty()) @@ -153,19 +146,21 @@ BinaryHolder::MapArchiveAndGetMemberBuffers(StringRef Filename, std::vector<MemoryBufferRef> ArchiveBuffers; auto ErrOrFat = object::MachOUniversalBinary::create( CurrentMemoryBuffer->getMemBufferRef()); - if (ErrOrFat.getError()) { + if (!ErrOrFat) { + consumeError(ErrOrFat.takeError()); // Not a fat binary must be a standard one. ArchiveBuffers.push_back(CurrentMemoryBuffer->getMemBufferRef()); } else { CurrentFatBinary = std::move(*ErrOrFat); + CurrentFatBinaryName = ArchiveFilename; ArchiveBuffers = getMachOFatMemoryBuffers( - ArchiveFilename, *CurrentMemoryBuffer, *CurrentFatBinary); + CurrentFatBinaryName, *CurrentMemoryBuffer, *CurrentFatBinary); } for (auto MemRef : ArchiveBuffers) { auto ErrOrArchive = object::Archive::create(MemRef); - if (auto Err = ErrOrArchive.getError()) - return Err; + if (!ErrOrArchive) + return errorToErrorCode(ErrOrArchive.takeError()); CurrentArchives.push_back(std::move(*ErrOrArchive)); } return GetArchiveMemberBuffers(Filename, Timestamp); @@ -175,7 +170,7 @@ ErrorOr<const object::ObjectFile &> BinaryHolder::getObjfileForArch(const Triple &T) { for (const auto &Obj : CurrentObjectFiles) { if (const auto *MachO = dyn_cast<object::MachOObjectFile>(Obj.get())) { - if (getTriple(*MachO).str() == T.str()) + if (MachO->getArchTriple().str() == T.str()) return *MachO; } else if (Obj->getArch() == T.getArch()) return *Obj; @@ -196,8 +191,8 @@ BinaryHolder::GetObjectFiles(StringRef Filename, sys::TimeValue Timestamp) { CurrentObjectFiles.clear(); for (auto MemBuf : *ErrOrMemBufferRefs) { auto ErrOrObjectFile = object::ObjectFile::createObjectFile(MemBuf); - if (auto Err = ErrOrObjectFile.getError()) - return Err; + if (!ErrOrObjectFile) + return errorToErrorCode(ErrOrObjectFile.takeError()); Objects.push_back(ErrOrObjectFile->get()); CurrentObjectFiles.push_back(std::move(*ErrOrObjectFile)); diff --git a/tools/dsymutil/BinaryHolder.h b/tools/dsymutil/BinaryHolder.h index 9d7b4bd8787f3..97508b9fb09da 100644 --- a/tools/dsymutil/BinaryHolder.h +++ b/tools/dsymutil/BinaryHolder.h @@ -42,6 +42,7 @@ class BinaryHolder { std::unique_ptr<MemoryBuffer> CurrentMemoryBuffer; std::vector<std::unique_ptr<object::ObjectFile>> CurrentObjectFiles; std::unique_ptr<object::MachOUniversalBinary> CurrentFatBinary; + std::string CurrentFatBinaryName; bool Verbose; /// Get the MemoryBufferRefs for the file specification in \p @@ -128,8 +129,6 @@ public: return Err; return cast<ObjectFileType>(*ErrOrObj); } - - static Triple getTriple(const object::MachOObjectFile &Obj); }; } } diff --git a/tools/dsymutil/DebugMap.cpp b/tools/dsymutil/DebugMap.cpp index 4717085f43227..5130cd62ed2fc 100644 --- a/tools/dsymutil/DebugMap.cpp +++ b/tools/dsymutil/DebugMap.cpp @@ -24,13 +24,13 @@ DebugMapObject::DebugMapObject(StringRef ObjectFilename, sys::TimeValue Timestamp) : Filename(ObjectFilename), Timestamp(Timestamp) {} -bool DebugMapObject::addSymbol(StringRef Name, uint64_t ObjectAddress, +bool DebugMapObject::addSymbol(StringRef Name, Optional<uint64_t> ObjectAddress, uint64_t LinkedAddress, uint32_t Size) { auto InsertResult = Symbols.insert( std::make_pair(Name, SymbolMapping(ObjectAddress, LinkedAddress, Size))); - if (InsertResult.second) - AddressToMapping[ObjectAddress] = &*InsertResult.first; + if (ObjectAddress && InsertResult.second) + AddressToMapping[*ObjectAddress] = &*InsertResult.first; return InsertResult.second; } @@ -47,8 +47,11 @@ void DebugMapObject::print(raw_ostream &OS) const { Entries.begin(), Entries.end(), [](const Entry &LHS, const Entry &RHS) { return LHS.first < RHS.first; }); for (const auto &Sym : Entries) { - OS << format("\t%016" PRIx64 " => %016" PRIx64 "+0x%x\t%s\n", - uint64_t(Sym.second.ObjectAddress), + if (Sym.second.ObjectAddress) + OS << format("\t%016" PRIx64, uint64_t(*Sym.second.ObjectAddress)); + else + OS << "\t????????????????"; + OS << format(" => %016" PRIx64 "+0x%x\t%s\n", uint64_t(Sym.second.BinaryAddress), uint32_t(Sym.second.Size), Sym.first.data()); } @@ -136,7 +139,7 @@ struct MappingTraits<dsymutil::DebugMapObject>::YamlDMO { void MappingTraits<std::pair<std::string, DebugMapObject::SymbolMapping>>:: mapping(IO &io, std::pair<std::string, DebugMapObject::SymbolMapping> &s) { io.mapRequired("sym", s.first); - io.mapRequired("objAddr", s.second.ObjectAddress); + io.mapOptional("objAddr", s.second.ObjectAddress); io.mapRequired("binAddr", s.second.BinaryAddress); io.mapOptional("size", s.second.Size); } @@ -225,9 +228,14 @@ MappingTraits<dsymutil::DebugMapObject>::YamlDMO::denormalize(IO &IO) { // look them up here and rewrite them. for (const auto &Sym : ErrOrObjectFile->symbols()) { uint64_t Address = Sym.getValue(); - ErrorOr<StringRef> Name = Sym.getName(); - if (!Name) + Expected<StringRef> Name = Sym.getName(); + if (!Name || + (Sym.getFlags() & (SymbolRef::SF_Absolute | SymbolRef::SF_Common))) { + // TODO: Actually report errors helpfully. + if (!Name) + consumeError(Name.takeError()); continue; + } SymbolAddresses[*Name] = Address; } } @@ -237,7 +245,9 @@ MappingTraits<dsymutil::DebugMapObject>::YamlDMO::denormalize(IO &IO) { dsymutil::DebugMapObject Res(Path, TV); for (auto &Entry : Entries) { auto &Mapping = Entry.second; - uint64_t ObjAddress = Mapping.ObjectAddress; + Optional<uint64_t> ObjAddress; + if (Mapping.ObjectAddress) + ObjAddress = *Mapping.ObjectAddress; auto AddressIt = SymbolAddresses.find(Entry.first); if (AddressIt != SymbolAddresses.end()) ObjAddress = AddressIt->getValue(); diff --git a/tools/dsymutil/DebugMap.h b/tools/dsymutil/DebugMap.h index 06ac5a503dcd1..1a3d62b67b7b4 100644 --- a/tools/dsymutil/DebugMap.h +++ b/tools/dsymutil/DebugMap.h @@ -1,4 +1,4 @@ -//===- tools/dsymutil/DebugMap.h - Generic debug map representation -------===// +//=== tools/dsymutil/DebugMap.h - Generic debug map representation -*- C++ -*-// // // The LLVM Linker // @@ -117,12 +117,15 @@ public: class DebugMapObject { public: struct SymbolMapping { - yaml::Hex64 ObjectAddress; + Optional<yaml::Hex64> ObjectAddress; yaml::Hex64 BinaryAddress; yaml::Hex32 Size; - SymbolMapping(uint64_t ObjectAddress, uint64_t BinaryAddress, uint32_t Size) - : ObjectAddress(ObjectAddress), BinaryAddress(BinaryAddress), - Size(Size) {} + SymbolMapping(Optional<uint64_t> ObjectAddr, uint64_t BinaryAddress, + uint32_t Size) + : BinaryAddress(BinaryAddress), Size(Size) { + if (ObjectAddr) + ObjectAddress = *ObjectAddr; + } /// For YAML IO support SymbolMapping() = default; }; @@ -132,7 +135,7 @@ public: /// \brief Adds a symbol mapping to this DebugMapObject. /// \returns false if the symbol was already registered. The request /// is discarded in this case. - bool addSymbol(llvm::StringRef SymName, uint64_t ObjectAddress, + bool addSymbol(llvm::StringRef SymName, Optional<uint64_t> ObjectAddress, uint64_t LinkedAddress, uint32_t Size); /// \brief Lookup a symbol mapping. diff --git a/tools/dsymutil/DwarfLinker.cpp b/tools/dsymutil/DwarfLinker.cpp index 8c5f7946a4257..bea30ded591fc 100644 --- a/tools/dsymutil/DwarfLinker.cpp +++ b/tools/dsymutil/DwarfLinker.cpp @@ -315,16 +315,15 @@ public: const std::vector<AccelInfo> &getPubtypes() const { return Pubtypes; } /// Get the full path for file \a FileNum in the line table - const char *getResolvedPath(unsigned FileNum) { + StringRef getResolvedPath(unsigned FileNum) { if (FileNum >= ResolvedPaths.size()) - return nullptr; - return ResolvedPaths[FileNum].size() ? ResolvedPaths[FileNum].c_str() - : nullptr; + return StringRef(); + return ResolvedPaths[FileNum]; } /// Set the fully resolved path for the line-table's file \a FileNum /// to \a Path. - void setResolvedPath(unsigned FileNum, const std::string &Path) { + void setResolvedPath(unsigned FileNum, StringRef Path) { if (ResolvedPaths.size() <= FileNum) ResolvedPaths.resize(FileNum + 1); ResolvedPaths[FileNum] = Path; @@ -378,7 +377,10 @@ private: /// @} /// Cached resolved paths from the line table. - std::vector<std::string> ResolvedPaths; + /// Note, the StringRefs here point in to the intern (uniquing) string pool. + /// This means that a StringRef returned here doesn't need to then be uniqued + /// for the purposes of getting a unique address for each string. + std::vector<StringRef> ResolvedPaths; /// Is this unit subject to the ODR rule? bool HasODR; @@ -529,7 +531,7 @@ public: /// original \p Entries. void emitRangesEntries( int64_t UnitPcOffset, uint64_t OrigLowPc, - FunctionIntervals::const_iterator FuncRange, + const FunctionIntervals::const_iterator &FuncRange, const std::vector<DWARFDebugRangeList::RangeListEntry> &Entries, unsigned AddressSize); @@ -593,8 +595,7 @@ bool DwarfStreamer::init(Triple TheTriple, StringRef OutputFilename) { MOFI.reset(new MCObjectFileInfo); MC.reset(new MCContext(MAI.get(), MRI.get(), MOFI.get())); - MOFI->InitMCObjectFileInfo(TheTriple, Reloc::Default, CodeModel::Default, - *MC); + MOFI->InitMCObjectFileInfo(TheTriple, /*PIC*/ false, CodeModel::Default, *MC); MAB = TheTarget->createMCAsmBackend(*MRI, TripleName, ""); if (!MAB) @@ -628,7 +629,8 @@ bool DwarfStreamer::init(Triple TheTriple, StringRef OutputFilename) { return error("no object streamer for target " + TripleName, Context); // Finally create the AsmPrinter we'll use to emit the DIEs. - TM.reset(TheTarget->createTargetMachine(TripleName, "", "", TargetOptions())); + TM.reset(TheTarget->createTargetMachine(TripleName, "", "", TargetOptions(), + None)); if (!TM) return error("no target machine for target " + TripleName, Context); @@ -713,7 +715,7 @@ void DwarfStreamer::emitStrings(const NonRelocatableStringpool &Pool) { /// sized addresses describing the ranges. void DwarfStreamer::emitRangesEntries( int64_t UnitPcOffset, uint64_t OrigLowPc, - FunctionIntervals::const_iterator FuncRange, + const FunctionIntervals::const_iterator &FuncRange, const std::vector<DWARFDebugRangeList::RangeListEntry> &Entries, unsigned AddressSize) { MS->SwitchSection(MC->getObjectFileInfo()->getDwarfRangesSection()); @@ -790,7 +792,7 @@ void DwarfStreamer::emitUnitRangesEntries(CompileUnit &Unit, Asm->EmitInt8(AddressSize); // Address size Asm->EmitInt8(0); // Segment size - Asm->OutStreamer->EmitFill(Padding, 0x0); + Asm->OutStreamer->emitFill(Padding, 0x0); for (auto Range = Ranges.begin(), End = Ranges.end(); Range != End; ++Range) { @@ -1414,7 +1416,7 @@ private: /// \defgroup Helpers Various helper methods. /// /// @{ - bool createStreamer(Triple TheTriple, StringRef OutputFilename); + bool createStreamer(const Triple &TheTriple, StringRef OutputFilename); /// \brief Attempt to load a debug object from disk. ErrorOr<const object::ObjectFile &> loadObject(BinaryHolder &BinaryHolder, @@ -1458,6 +1460,9 @@ private: /// Mapping the PCM filename to the DwoId. StringMap<uint64_t> ClangModules; + + bool ModuleCacheHintDisplayed = false; + bool ArchiveHintDisplayed = false; }; /// Similar to DWARFUnitSection::getUnitForOffset(), but returning our @@ -1601,7 +1606,6 @@ PointerIntPair<DeclContext *, 1> DeclContextTree::getChildDeclContext( Tag != dwarf::DW_TAG_enumeration_type && NameRef.empty()) return PointerIntPair<DeclContext *, 1>(nullptr); - std::string File; unsigned Line = 0; unsigned ByteSize = UINT32_MAX; @@ -1629,25 +1633,27 @@ PointerIntPair<DeclContext *, 1> DeclContextTree::getChildDeclContext( // FIXME: Passing U.getOrigUnit().getCompilationDir() // instead of "" would allow more uniquing, but for now, do // it this way to match dsymutil-classic. + std::string File; if (LT->getFileNameByIndex( FileNum, "", DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath, File)) { Line = DIE->getAttributeValueAsUnsignedConstant( &U.getOrigUnit(), dwarf::DW_AT_decl_line, 0); -#ifdef HAVE_REALPATH // Cache the resolved paths, because calling realpath is expansive. - if (const char *ResolvedPath = U.getResolvedPath(FileNum)) { - File = ResolvedPath; + StringRef ResolvedPath = U.getResolvedPath(FileNum); + if (!ResolvedPath.empty()) { + FileRef = ResolvedPath; } else { +#ifdef HAVE_REALPATH char RealPath[PATH_MAX + 1]; RealPath[PATH_MAX] = 0; if (::realpath(File.c_str(), RealPath)) File = RealPath; - U.setResolvedPath(FileNum, File); - } #endif - FileRef = StringPool.internString(File); + FileRef = StringPool.internString(File); + U.setResolvedPath(FileNum, FileRef); + } } } } @@ -1738,7 +1744,8 @@ void DwarfLinker::reportWarning(const Twine &Warning, const DWARFUnit *Unit, 6 /* Indent */); } -bool DwarfLinker::createStreamer(Triple TheTriple, StringRef OutputFilename) { +bool DwarfLinker::createStreamer(const Triple &TheTriple, + StringRef OutputFilename) { if (Options.NoOutput) return true; @@ -1851,10 +1858,10 @@ void DwarfLinker::startDebugObject(DWARFContext &Dwarf, DebugMapObject &Obj) { // -gline-tables-only on Darwin. for (const auto &Entry : Obj.symbols()) { const auto &Mapping = Entry.getValue(); - if (Mapping.Size) - Ranges[Mapping.ObjectAddress] = std::make_pair( - Mapping.ObjectAddress + Mapping.Size, - int64_t(Mapping.BinaryAddress) - Mapping.ObjectAddress); + if (Mapping.Size && Mapping.ObjectAddress) + Ranges[*Mapping.ObjectAddress] = std::make_pair( + *Mapping.ObjectAddress + Mapping.Size, + int64_t(Mapping.BinaryAddress) - *Mapping.ObjectAddress); } } @@ -1872,6 +1879,26 @@ void DwarfLinker::endDebugObject() { DIEAlloc.Reset(); } +static bool isMachOPairedReloc(uint64_t RelocType, uint64_t Arch) { + switch (Arch) { + case Triple::x86: + return RelocType == MachO::GENERIC_RELOC_SECTDIFF || + RelocType == MachO::GENERIC_RELOC_LOCAL_SECTDIFF; + case Triple::x86_64: + return RelocType == MachO::X86_64_RELOC_SUBTRACTOR; + case Triple::arm: + case Triple::thumb: + return RelocType == MachO::ARM_RELOC_SECTDIFF || + RelocType == MachO::ARM_RELOC_LOCAL_SECTDIFF || + RelocType == MachO::ARM_RELOC_HALF || + RelocType == MachO::ARM_RELOC_HALF_SECTDIFF; + case Triple::aarch64: + return RelocType == MachO::ARM64_RELOC_SUBTRACTOR; + default: + return false; + } +} + /// \brief Iterate over the relocations of the given \p Section and /// store the ones that correspond to debug map entries into the /// ValidRelocs array. @@ -1882,10 +1909,24 @@ findValidRelocsMachO(const object::SectionRef &Section, StringRef Contents; Section.getContents(Contents); DataExtractor Data(Contents, Obj.isLittleEndian(), 0); + bool SkipNext = false; for (const object::RelocationRef &Reloc : Section.relocations()) { + if (SkipNext) { + SkipNext = false; + continue; + } + object::DataRefImpl RelocDataRef = Reloc.getRawDataRefImpl(); MachO::any_relocation_info MachOReloc = Obj.getRelocation(RelocDataRef); + + if (isMachOPairedReloc(Obj.getAnyRelocationType(MachOReloc), + Obj.getArch())) { + SkipNext = true; + Linker.reportWarning(" unsupported relocation in debug_info section."); + continue; + } + unsigned RelocSize = 1 << Obj.getAnyRelocationLength(MachOReloc); uint64_t Offset64 = Reloc.getOffset(); if ((RelocSize != 4 && RelocSize != 8)) { @@ -1895,21 +1936,35 @@ findValidRelocsMachO(const object::SectionRef &Section, uint32_t Offset = Offset64; // Mach-o uses REL relocations, the addend is at the relocation offset. uint64_t Addend = Data.getUnsigned(&Offset, RelocSize); + uint64_t SymAddress; + int64_t SymOffset; + + if (Obj.isRelocationScattered(MachOReloc)) { + // The address of the base symbol for scattered relocations is + // stored in the reloc itself. The actual addend will store the + // base address plus the offset. + SymAddress = Obj.getScatteredRelocationValue(MachOReloc); + SymOffset = int64_t(Addend) - SymAddress; + } else { + SymAddress = Addend; + SymOffset = 0; + } auto Sym = Reloc.getSymbol(); if (Sym != Obj.symbol_end()) { - ErrorOr<StringRef> SymbolName = Sym->getName(); + Expected<StringRef> SymbolName = Sym->getName(); if (!SymbolName) { + consumeError(SymbolName.takeError()); Linker.reportWarning("error getting relocation symbol name."); continue; } if (const auto *Mapping = DMO.lookupSymbol(*SymbolName)) ValidRelocs.emplace_back(Offset64, RelocSize, Addend, Mapping); - } else if (const auto *Mapping = DMO.lookupObjectAddress(Addend)) { + } else if (const auto *Mapping = DMO.lookupObjectAddress(SymAddress)) { // Do not store the addend. The addend was the address of the // symbol in the object file, the address in the binary that is // stored in the debug map doesn't need to be offseted. - ValidRelocs.emplace_back(Offset64, RelocSize, 0, Mapping); + ValidRelocs.emplace_back(Offset64, RelocSize, SymOffset, Mapping); } } } @@ -1985,14 +2040,16 @@ hasValidRelocation(uint32_t StartOffset, uint32_t EndOffset, const auto &ValidReloc = ValidRelocs[NextValidReloc++]; const auto &Mapping = ValidReloc.Mapping->getValue(); + uint64_t ObjectAddress = + Mapping.ObjectAddress ? uint64_t(*Mapping.ObjectAddress) : UINT64_MAX; if (Linker.Options.Verbose) outs() << "Found valid debug map entry: " << ValidReloc.Mapping->getKey() - << " " << format("\t%016" PRIx64 " => %016" PRIx64, - uint64_t(Mapping.ObjectAddress), + << " " << format("\t%016" PRIx64 " => %016" PRIx64, ObjectAddress, uint64_t(Mapping.BinaryAddress)); - Info.AddrAdjust = int64_t(Mapping.BinaryAddress) + ValidReloc.Addend - - Mapping.ObjectAddress; + Info.AddrAdjust = int64_t(Mapping.BinaryAddress) + ValidReloc.Addend; + if (Mapping.ObjectAddress) + Info.AddrAdjust -= ObjectAddress; Info.InDebugMap = true; return true; } @@ -3193,7 +3250,10 @@ bool DwarfLinker::registerModuleReference( auto Cached = ClangModules.find(PCMfile); if (Cached != ClangModules.end()) { - if (Cached->second != DwoId) + // FIXME: Until PR27449 (https://llvm.org/bugs/show_bug.cgi?id=27449) is + // fixed in clang, only warn about DWO_id mismatches in verbose mode. + // ASTFileSignatures will change randomly when a module is rebuilt. + if (Options.Verbose && (Cached->second != DwoId)) reportWarning(Twine("hash mismatch: this object file was built against a " "different version of the module ") + PCMfile); if (Options.Verbose) @@ -3238,7 +3298,37 @@ void DwarfLinker::loadClangModule(StringRef Filename, StringRef ModulePath, ModuleMap.addDebugMapObject(Path, sys::TimeValue::PosixZeroTime()); auto ErrOrObj = loadObject(ObjHolder, Obj, ModuleMap); if (!ErrOrObj) { - ClangModules.erase(ClangModules.find(Filename)); + // Try and emit more helpful warnings by applying some heuristics. + StringRef ObjFile = CurrentDebugObject->getObjectFilename(); + bool isClangModule = sys::path::extension(Filename).equals(".pcm"); + bool isArchive = ObjFile.endswith(")"); + if (isClangModule) { + StringRef ModuleCacheDir = sys::path::parent_path(Path); + if (sys::fs::exists(ModuleCacheDir)) { + // If the module's parent directory exists, we assume that the module + // cache has expired and was pruned by clang. A more adventurous + // dsymutil would invoke clang to rebuild the module now. + if (!ModuleCacheHintDisplayed) { + errs() << "note: The clang module cache may have expired since this " + "object file was built. Rebuilding the object file will " + "rebuild the module cache.\n"; + ModuleCacheHintDisplayed = true; + } + } else if (isArchive) { + // If the module cache directory doesn't exist at all and the object + // file is inside a static library, we assume that the static library + // was built on a different machine. We don't want to discourage module + // debugging for convenience libraries within a project though. + if (!ArchiveHintDisplayed) { + errs() << "note: Linking a static library that was built with " + "-gmodules, but the module cache was not found. " + "Redistributable static libraries should never be built " + "with module debugging enabled. The debug experience will " + "be degraded due to incomplete debug information.\n"; + ArchiveHintDisplayed = true; + } + } + } return; } @@ -3256,10 +3346,18 @@ void DwarfLinker::loadClangModule(StringRef Filename, StringRef ModulePath, << " 1 compile unit.\n"; exitDsymutil(1); } - if (getDwoId(*CUDie, *CU) != DwoId) - reportWarning( - Twine("hash mismatch: this object file was built against a " - "different version of the module ") + Filename); + // FIXME: Until PR27449 (https://llvm.org/bugs/show_bug.cgi?id=27449) is + // fixed in clang, only warn about DWO_id mismatches in verbose mode. + // ASTFileSignatures will change randomly when a module is rebuilt. + uint64_t PCMDwoId = getDwoId(*CUDie, *CU); + if (PCMDwoId != DwoId) { + if (Options.Verbose) + reportWarning( + Twine("hash mismatch: this object file was built against a " + "different version of the module ") + Filename); + // Update the cache entry with the DwoId of the module loaded from disk. + ClangModules[Filename] = PCMDwoId; + } // Add this module. Unit = llvm::make_unique<CompileUnit>(*CU, UnitID++, !Options.NoODR, diff --git a/tools/dsymutil/MachODebugMapParser.cpp b/tools/dsymutil/MachODebugMapParser.cpp index 4412db25426ca..22215200ed5f6 100644 --- a/tools/dsymutil/MachODebugMapParser.cpp +++ b/tools/dsymutil/MachODebugMapParser.cpp @@ -137,8 +137,7 @@ void MachODebugMapParser::switchToNewDebugMapObject(StringRef Filename, } static std::string getArchName(const object::MachOObjectFile &Obj) { - Triple ThumbTriple; - Triple T = Obj.getArch(nullptr, &ThumbTriple); + Triple T = Obj.getArchTriple(); return T.getArchName(); } @@ -146,8 +145,7 @@ std::unique_ptr<DebugMap> MachODebugMapParser::parseOneBinary(const MachOObjectFile &MainBinary, StringRef BinaryPath) { loadMainBinarySymbols(MainBinary); - Result = - make_unique<DebugMap>(BinaryHolder::getTriple(MainBinary), BinaryPath); + Result = make_unique<DebugMap>(MainBinary.getArchTriple(), BinaryPath); MainBinaryStrings = MainBinary.getStringTableData(); for (const SymbolRef &Symbol : MainBinary.symbols()) { const DataRefImpl &DRI = Symbol.getRawDataRefImpl(); @@ -296,7 +294,11 @@ static bool shouldLinkArch(SmallVectorImpl<StringRef> &Archs, StringRef Arch) { std::find(Archs.begin(), Archs.end(), "arm") != Archs.end()) return true; - return std::find(Archs.begin(), Archs.end(), Arch) != Archs.end(); + SmallString<16> ArchName = Arch; + if (Arch.startswith("thumb")) + ArchName = ("arm" + Arch.substr(5)).str(); + + return std::find(Archs.begin(), Archs.end(), ArchName) != Archs.end(); } bool MachODebugMapParser::dumpStab() { @@ -308,9 +310,8 @@ bool MachODebugMapParser::dumpStab() { return false; } - Triple T; for (const auto *Binary : *MainBinOrError) - if (shouldLinkArch(Archs, Binary->getArch(nullptr, &T).getArchName())) + if (shouldLinkArch(Archs, Binary->getArchTriple().getArchName())) dumpOneBinaryStab(*Binary, BinaryPath); return true; @@ -326,9 +327,8 @@ ErrorOr<std::vector<std::unique_ptr<DebugMap>>> MachODebugMapParser::parse() { return Error; std::vector<std::unique_ptr<DebugMap>> Results; - Triple T; for (const auto *Binary : *MainBinOrError) - if (shouldLinkArch(Archs, Binary->getArch(nullptr, &T).getArchName())) + if (shouldLinkArch(Archs, Binary->getArchTriple().getArchName())) Results.push_back(parseOneBinary(*Binary, BinaryPath)); return std::move(Results); @@ -389,9 +389,7 @@ void MachODebugMapParser::handleStabSymbolTableEntry(uint32_t StringIndex, if (ObjectSymIt == CurrentObjectAddresses.end()) return Warning("could not find object file symbol for symbol " + Twine(Name)); - if (!ObjectSymIt->getValue()) - return; - if (!CurrentDebugMapObject->addSymbol(Name, *ObjectSymIt->getValue(), Value, + if (!CurrentDebugMapObject->addSymbol(Name, ObjectSymIt->getValue(), Value, Size)) return Warning(Twine("failed to insert symbol '") + Name + "' in the debug map."); @@ -404,15 +402,21 @@ void MachODebugMapParser::loadCurrentObjectFileSymbols( for (auto Sym : Obj.symbols()) { uint64_t Addr = Sym.getValue(); - ErrorOr<StringRef> Name = Sym.getName(); - if (!Name) + Expected<StringRef> Name = Sym.getName(); + if (!Name) { + // TODO: Actually report errors helpfully. + consumeError(Name.takeError()); continue; - // Objective-C on i386 uses artificial absolute symbols to - // perform some link time checks. Those symbols have a fixed 0 - // address that might conflict with real symbols in the object - // file. As I cannot see a way for absolute symbols to find - // their way into the debug information, let's just ignore those. - if (Sym.getFlags() & SymbolRef::SF_Absolute) + } + // The value of some categories of symbols isn't meaningful. For + // example common symbols store their size in the value field, not + // their address. Absolute symbols have a fixed address that can + // conflict with standard symbols. These symbols (especially the + // common ones), might still be referenced by relocations. These + // relocations will use the symbol itself, and won't need an + // object file address. The object file address field is optional + // in the DebugMap, leave it unassigned for these symbols. + if (Sym.getFlags() & (SymbolRef::SF_Absolute | SymbolRef::SF_Common)) CurrentObjectAddresses[*Name] = None; else CurrentObjectAddresses[*Name] = Addr; @@ -436,7 +440,13 @@ void MachODebugMapParser::loadMainBinarySymbols( section_iterator Section = MainBinary.section_end(); MainBinarySymbolAddresses.clear(); for (const auto &Sym : MainBinary.symbols()) { - SymbolRef::Type Type = Sym.getType(); + Expected<SymbolRef::Type> TypeOrErr = Sym.getType(); + if (!TypeOrErr) { + // TODO: Actually report errors helpfully. + consumeError(TypeOrErr.takeError()); + continue; + } + SymbolRef::Type Type = *TypeOrErr; // Skip undefined and STAB entries. if ((Type & SymbolRef::ST_Debug) || (Type & SymbolRef::ST_Unknown)) continue; @@ -446,16 +456,22 @@ void MachODebugMapParser::loadMainBinarySymbols( // addresses should be fetched for the debug map. if (!(Sym.getFlags() & SymbolRef::SF_Global)) continue; - ErrorOr<section_iterator> SectionOrErr = Sym.getSection(); - if (!SectionOrErr) + Expected<section_iterator> SectionOrErr = Sym.getSection(); + if (!SectionOrErr) { + // TODO: Actually report errors helpfully. + consumeError(SectionOrErr.takeError()); continue; + } Section = *SectionOrErr; if (Section == MainBinary.section_end() || Section->isText()) continue; uint64_t Addr = Sym.getValue(); - ErrorOr<StringRef> NameOrErr = Sym.getName(); - if (!NameOrErr) + Expected<StringRef> NameOrErr = Sym.getName(); + if (!NameOrErr) { + // TODO: Actually report errors helpfully. + consumeError(NameOrErr.takeError()); continue; + } StringRef Name = *NameOrErr; if (Name.size() == 0 || Name[0] == '\0') continue; diff --git a/tools/dsymutil/MachOUtils.cpp b/tools/dsymutil/MachOUtils.cpp index 44cc528ed8f2b..8a730a1d0c8a2 100644 --- a/tools/dsymutil/MachOUtils.cpp +++ b/tools/dsymutil/MachOUtils.cpp @@ -16,6 +16,7 @@ #include "llvm/MC/MCAsmLayout.h" #include "llvm/MC/MCSectionMachO.h" #include "llvm/MC/MCObjectStreamer.h" +#include "llvm/MC/MCMachObjectWriter.h" #include "llvm/MC/MCStreamer.h" #include "llvm/Object/MachO.h" #include "llvm/Support/FileUtilities.h" @@ -241,7 +242,7 @@ static void transferSegmentAndSections( // start address leave a sufficient gap to store the __DWARF // segment. uint64_t PrevEndAddress = EndAddress; - EndAddress = RoundUpToAlignment(EndAddress, 0x1000); + EndAddress = alignTo(EndAddress, 0x1000); if (GapForDwarf == UINT64_MAX && Segment.vmaddr > EndAddress && Segment.vmaddr - EndAddress >= DwarfSegmentSize) GapForDwarf = EndAddress; @@ -268,8 +269,8 @@ static void createDwarfSegment(uint64_t VMAddr, uint64_t FileOffset, uint64_t FileSize, unsigned NumSections, MCAsmLayout &Layout, MachObjectWriter &Writer) { Writer.writeSegmentLoadCommand("__DWARF", NumSections, VMAddr, - RoundUpToAlignment(FileSize, 0x1000), - FileOffset, FileSize, /* MaxProt */ 7, + alignTo(FileSize, 0x1000), FileOffset, + FileSize, /* MaxProt */ 7, /* InitProt =*/3); for (unsigned int i = 0, n = Layout.getSectionOrder().size(); i != n; ++i) { @@ -279,8 +280,8 @@ static void createDwarfSegment(uint64_t VMAddr, uint64_t FileOffset, unsigned Align = Sec->getAlignment(); if (Align > 1) { - VMAddr = RoundUpToAlignment(VMAddr, Align); - FileOffset = RoundUpToAlignment(FileOffset, Align); + VMAddr = alignTo(VMAddr, Align); + FileOffset = alignTo(FileOffset, Align); } Writer.writeSection(Layout, *Sec, VMAddr, FileOffset, 0, 0, 0); @@ -394,8 +395,7 @@ bool generateDsymCompanion(const DebugMap &DM, MCStreamer &MS, continue; if (uint64_t Size = Layout.getSectionFileSize(Sec)) { - DwarfSegmentSize = - RoundUpToAlignment(DwarfSegmentSize, Sec->getAlignment()); + DwarfSegmentSize = alignTo(DwarfSegmentSize, Sec->getAlignment()); DwarfSegmentSize += Size; ++NumDwarfSections; } @@ -419,7 +419,7 @@ bool generateDsymCompanion(const DebugMap &DM, MCStreamer &MS, uint64_t SymtabStart = LoadCommandSize; SymtabStart += HeaderSize; - SymtabStart = RoundUpToAlignment(SymtabStart, 0x1000); + SymtabStart = alignTo(SymtabStart, 0x1000); // We gathered all the information we need, start emitting the output file. Writer.writeHeader(MachO::MH_DSYM, NumLoadCommands, LoadCommandSize, false); @@ -441,7 +441,7 @@ bool generateDsymCompanion(const DebugMap &DM, MCStreamer &MS, NewStringsSize); uint64_t DwarfSegmentStart = StringStart + NewStringsSize; - DwarfSegmentStart = RoundUpToAlignment(DwarfSegmentStart, 0x1000); + DwarfSegmentStart = alignTo(DwarfSegmentStart, 0x1000); // Write the load commands for the segments and sections we 'import' from // the original binary. @@ -460,7 +460,7 @@ bool generateDsymCompanion(const DebugMap &DM, MCStreamer &MS, DwarfSegmentSize, GapForDwarf, EndAddress); } - uint64_t DwarfVMAddr = RoundUpToAlignment(EndAddress, 0x1000); + uint64_t DwarfVMAddr = alignTo(EndAddress, 0x1000); uint64_t DwarfVMMax = Is64Bit ? UINT64_MAX : UINT32_MAX; if (DwarfVMAddr + DwarfSegmentSize > DwarfVMMax || DwarfVMAddr + DwarfSegmentSize < DwarfVMAddr /* Overflow */) { @@ -510,7 +510,7 @@ bool generateDsymCompanion(const DebugMap &DM, MCStreamer &MS, continue; uint64_t Pos = OutFile.tell(); - Writer.WriteZeros(RoundUpToAlignment(Pos, Sec.getAlignment()) - Pos); + Writer.WriteZeros(alignTo(Pos, Sec.getAlignment()) - Pos); MCAsm.writeSectionData(&Sec, Layout); } diff --git a/tools/dsymutil/Makefile b/tools/dsymutil/Makefile deleted file mode 100644 index c4365e0196b3e..0000000000000 --- a/tools/dsymutil/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/dsymutil/Makefile -----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-dsymutil -LINK_COMPONENTS := all-targets AsmPrinter DebugInfoDWARF MC Object Support - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/dsymutil/NonRelocatableStringpool.h b/tools/dsymutil/NonRelocatableStringpool.h index 9db1f3342bd3f..fffd02084e344 100644 --- a/tools/dsymutil/NonRelocatableStringpool.h +++ b/tools/dsymutil/NonRelocatableStringpool.h @@ -9,6 +9,8 @@ #ifndef LLVM_TOOLS_DSYMUTIL_NONRELOCATABLESTRINGPOOL_H #define LLVM_TOOLS_DSYMUTIL_NONRELOCATABLESTRINGPOOL_H +#include "llvm/ADT/StringMap.h" + namespace llvm { namespace dsymutil { diff --git a/tools/dsymutil/dsymutil.cpp b/tools/dsymutil/dsymutil.cpp index e9ee57f3dee67..1ce0aefeec2af 100644 --- a/tools/dsymutil/dsymutil.cpp +++ b/tools/dsymutil/dsymutil.cpp @@ -176,14 +176,17 @@ static std::error_code getUniqueFile(const llvm::Twine &Model, int &ResultFD, static std::string getOutputFileName(llvm::StringRef InputFile, bool TempFile = false) { if (TempFile) { + llvm::SmallString<128> TmpFile; + llvm::sys::path::system_temp_directory(true, TmpFile); llvm::StringRef Basename = OutputFileOpt.empty() ? InputFile : llvm::StringRef(OutputFileOpt); - llvm::Twine OutputFile = Basename + ".tmp%%%%%%.dwarf"; + llvm::sys::path::append(TmpFile, llvm::sys::path::filename(Basename)); + int FD; llvm::SmallString<128> UniqueFile; - if (auto EC = getUniqueFile(OutputFile, FD, UniqueFile)) { + if (auto EC = getUniqueFile(TmpFile + ".tmp%%%%%.dwarf", FD, UniqueFile)) { llvm::errs() << "error: failed to create temporary outfile '" - << OutputFile << "': " << EC.message() << '\n'; + << TmpFile << "': " << EC.message() << '\n'; return ""; } llvm::sys::RemoveFileOnSignal(UniqueFile); @@ -234,7 +237,7 @@ void llvm::dsymutil::exitDsymutil(int ExitStatus) { } int main(int argc, char **argv) { - llvm::sys::PrintStackTraceOnErrorSignal(); + llvm::sys::PrintStackTraceOnErrorSignal(argv[0]); llvm::PrettyStackTraceProgram StackPrinter(argc, argv); llvm::llvm_shutdown_obj Shutdown; LinkOptions Options; diff --git a/tools/gold/CMakeLists.txt b/tools/gold/CMakeLists.txt index 1a6169d65c2ca..e9029e1e7c107 100644 --- a/tools/gold/CMakeLists.txt +++ b/tools/gold/CMakeLists.txt @@ -10,6 +10,7 @@ if( LLVM_ENABLE_PIC AND LLVM_BINUTILS_INCDIR ) set(LLVM_LINK_COMPONENTS ${LLVM_TARGETS_TO_BUILD} Linker + LTO BitWriter IPO ) diff --git a/tools/gold/Makefile b/tools/gold/Makefile deleted file mode 100644 index aa006b0048fc7..0000000000000 --- a/tools/gold/Makefile +++ /dev/null @@ -1,31 +0,0 @@ -#===- tools/gold/Makefile ----------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -LIBRARYNAME := LLVMgold -LINK_LIBS_IN_SHARED := 1 -SHARED_LIBRARY := 1 -LOADABLE_MODULE := 1 - -EXPORTED_SYMBOL_FILE = $(PROJ_SRC_DIR)/gold.exports - -# Include this here so we can get the configuration of the targets -# that have been configured for construction. We have to do this -# early so we can set up LINK_COMPONENTS before including Makefile.rules -include $(LEVEL)/Makefile.config - -LINK_COMPONENTS := $(TARGETS_TO_BUILD) Linker BitWriter IPO - -# Because off_t is used in the public API, the largefile parts are required for -# ABI compatibility. -CXXFLAGS += -I$(BINUTILS_INCDIR) -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -LDFLAGS += -L$(SharedLibDir)/$(SharedPrefix) - -include $(LEVEL)/Makefile.common - diff --git a/tools/gold/gold-plugin.cpp b/tools/gold/gold-plugin.cpp index 31d7f156a7004..e3351c29fcdb7 100644 --- a/tools/gold/gold-plugin.cpp +++ b/tools/gold/gold-plugin.cpp @@ -12,8 +12,6 @@ // //===----------------------------------------------------------------------===// -#include "llvm/Config/config.h" // plugin-api.h requires HAVE_STDINT_H -#include "llvm/ADT/DenseSet.h" #include "llvm/ADT/StringSet.h" #include "llvm/Analysis/TargetLibraryInfo.h" #include "llvm/Analysis/TargetTransformInfo.h" @@ -21,6 +19,7 @@ #include "llvm/CodeGen/Analysis.h" #include "llvm/CodeGen/CommandFlags.h" #include "llvm/CodeGen/ParallelCG.h" +#include "llvm/Config/config.h" // plugin-api.h requires HAVE_STDINT_H #include "llvm/IR/AutoUpgrade.h" #include "llvm/IR/Constants.h" #include "llvm/IR/DiagnosticInfo.h" @@ -29,31 +28,37 @@ #include "llvm/IR/LegacyPassManager.h" #include "llvm/IR/Module.h" #include "llvm/IR/Verifier.h" +#include "llvm/LTO/LTO.h" #include "llvm/Linker/IRMover.h" #include "llvm/MC/SubtargetFeature.h" -#include "llvm/Object/FunctionIndexObjectFile.h" #include "llvm/Object/IRObjectFile.h" +#include "llvm/Object/ModuleSummaryIndexObjectFile.h" #include "llvm/Support/Host.h" #include "llvm/Support/ManagedStatic.h" #include "llvm/Support/MemoryBuffer.h" +#include "llvm/Support/Path.h" #include "llvm/Support/TargetRegistry.h" #include "llvm/Support/TargetSelect.h" +#include "llvm/Support/ThreadPool.h" #include "llvm/Support/raw_ostream.h" +#include "llvm/Support/thread.h" #include "llvm/Transforms/IPO.h" +#include "llvm/Transforms/IPO/FunctionImport.h" #include "llvm/Transforms/IPO/PassManagerBuilder.h" +#include "llvm/Transforms/Utils/FunctionImportUtils.h" #include "llvm/Transforms/Utils/GlobalStatus.h" -#include "llvm/Transforms/Utils/ModuleUtils.h" #include "llvm/Transforms/Utils/ValueMapper.h" #include <list> #include <plugin-api.h> #include <system_error> +#include <utility> #include <vector> -#ifndef LDPO_PIE // FIXME: remove this declaration when we stop maintaining Ubuntu Quantal and // Precise and Debian Wheezy (binutils 2.23 is required) -# define LDPO_PIE 3 -#endif +#define LDPO_PIE 3 + +#define LDPT_GET_SYMBOLS_V3 28 using namespace llvm; @@ -70,34 +75,68 @@ static ld_plugin_message message = discard_message; namespace { struct claimed_file { void *handle; + void *leader_handle; std::vector<ld_plugin_symbol> syms; + off_t filesize; + std::string name; }; /// RAII wrapper to manage opening and releasing of a ld_plugin_input_file. struct PluginInputFile { void *Handle; - ld_plugin_input_file File; + std::unique_ptr<ld_plugin_input_file> File; PluginInputFile(void *Handle) : Handle(Handle) { - if (get_input_file(Handle, &File) != LDPS_OK) + File = llvm::make_unique<ld_plugin_input_file>(); + if (get_input_file(Handle, File.get()) != LDPS_OK) message(LDPL_FATAL, "Failed to get file information"); } ~PluginInputFile() { - if (release_input_file(Handle) != LDPS_OK) - message(LDPL_FATAL, "Failed to release file information"); + // File would have been reset to nullptr if we moved this object + // to a new owner. + if (File) + if (release_input_file(Handle) != LDPS_OK) + message(LDPL_FATAL, "Failed to release file information"); } - ld_plugin_input_file &file() { return File; } + + ld_plugin_input_file &file() { return *File; } + + PluginInputFile(PluginInputFile &&RHS) = default; + PluginInputFile &operator=(PluginInputFile &&RHS) = default; }; struct ResolutionInfo { + uint64_t CommonSize = 0; + unsigned CommonAlign = 0; bool IsLinkonceOdr = true; - bool UnnamedAddr = true; + GlobalValue::UnnamedAddr UnnamedAddr = GlobalValue::UnnamedAddr::Global; GlobalValue::VisibilityTypes Visibility = GlobalValue::DefaultVisibility; bool CommonInternal = false; bool UseCommon = false; - unsigned CommonSize = 0; - unsigned CommonAlign = 0; - claimed_file *CommonFile = nullptr; +}; + +/// Class to own information used by a task or during its cleanup for a +/// ThinLTO backend instantiation. +class ThinLTOTaskInfo { + /// The output stream the task will codegen into. + std::unique_ptr<raw_fd_ostream> OS; + + /// The file name corresponding to the output stream, used during cleanup. + std::string Filename; + + /// Flag indicating whether the output file is a temp file that must be + /// added to the cleanup list during cleanup. + bool TempOutFile; + +public: + ThinLTOTaskInfo(std::unique_ptr<raw_fd_ostream> OS, std::string Filename, + bool TempOutFile) + : OS(std::move(OS)), Filename(std::move(Filename)), + TempOutFile(TempOutFile) {} + + /// Performs task related cleanup activities that must be done + /// single-threaded (i.e. call backs to gold). + void cleanup(); }; } @@ -106,12 +145,14 @@ static ld_plugin_get_symbols get_symbols = nullptr; static ld_plugin_add_input_file add_input_file = nullptr; static ld_plugin_set_extra_library_path set_extra_library_path = nullptr; static ld_plugin_get_view get_view = nullptr; -static Reloc::Model RelocationModel = Reloc::Default; +static Optional<Reloc::Model> RelocationModel; static std::string output_name = ""; static std::list<claimed_file> Modules; +static DenseMap<int, void *> FDToLeaderHandle; static StringMap<ResolutionInfo> ResInfo; static std::vector<std::string> Cleanup; static llvm::TargetOptions TargetOpts; +static std::string DefaultTriple = sys::getDefaultTargetTriple(); namespace options { enum OutputType { @@ -123,7 +164,11 @@ namespace options { static bool generate_api_file = false; static OutputType TheOutputType = OT_NORMAL; static unsigned OptLevel = 2; - static unsigned Parallelism = 1; + // Default parallelism of 0 used to indicate that user did not specify. + // Actual parallelism default value depends on implementation. + // Currently, code generation defaults to no parallelism, whereas + // ThinLTO uses the hardware_concurrency as the default. + static unsigned Parallelism = 0; #ifdef NDEBUG static bool DisableVerify = true; #else @@ -137,6 +182,24 @@ namespace options { // the information from intermediate files and write a combined // global index for the ThinLTO backends. static bool thinlto = false; + // If false, all ThinLTO backend compilations through code gen are performed + // using multiple threads in the gold-plugin, before handing control back to + // gold. If true, write individual backend index files which reflect + // the import decisions, and exit afterwards. The assumption is + // that the build system will launch the backend processes. + static bool thinlto_index_only = false; + // If true, when generating individual index files for distributed backends, + // also generate a "${bitcodefile}.imports" file at the same location for each + // bitcode file, listing the files it imports from in plain text. This is to + // support distributed build file staging. + static bool thinlto_emit_imports_files = false; + // Option to control where files for a distributed backend (the individual + // index files and optional imports files) are created. + // If specified, expects a string of the form "oldprefix:newprefix", and + // instead of generating these files in the same directory path as the + // corresponding bitcode file, will use a path formed by replacing the + // bitcode file's path prefix matching oldprefix with newprefix. + static std::string thinlto_prefix_replace; // Additional options to pass into the code generator. // Note: This array will contain all plugin options which are not claimed // as plugin exclusive to pass to the code generator. @@ -168,6 +231,14 @@ namespace options { TheOutputType = OT_DISABLE; } else if (opt == "thinlto") { thinlto = true; + } else if (opt == "thinlto-index-only") { + thinlto_index_only = true; + } else if (opt == "thinlto-emit-imports-files") { + thinlto_emit_imports_files = true; + } else if (opt.startswith("thinlto-prefix-replace=")) { + thinlto_prefix_replace = opt.substr(strlen("thinlto-prefix-replace=")); + if (thinlto_prefix_replace.find(";") == std::string::npos) + message(LDPL_FATAL, "thinlto-prefix-replace expects 'old;new' format"); } else if (opt.size() == 2 && opt[0] == 'O') { if (opt[1] < '0' || opt[1] > '3') message(LDPL_FATAL, "Optimization level must be between 0 and 3"); @@ -212,79 +283,87 @@ ld_plugin_status onload(ld_plugin_tv *tv) { bool RegisteredAllSymbolsRead = false; for (; tv->tv_tag != LDPT_NULL; ++tv) { - switch (tv->tv_tag) { - case LDPT_OUTPUT_NAME: - output_name = tv->tv_u.tv_string; - break; - case LDPT_LINKER_OUTPUT: - switch (tv->tv_u.tv_val) { - case LDPO_REL: // .o - case LDPO_DYN: // .so - case LDPO_PIE: // position independent executable - RelocationModel = Reloc::PIC_; - break; - case LDPO_EXEC: // .exe - RelocationModel = Reloc::Static; - break; - default: - message(LDPL_ERROR, "Unknown output file type %d", tv->tv_u.tv_val); - return LDPS_ERR; - } - break; - case LDPT_OPTION: - options::process_plugin_option(tv->tv_u.tv_string); - break; - case LDPT_REGISTER_CLAIM_FILE_HOOK: { - ld_plugin_register_claim_file callback; - callback = tv->tv_u.tv_register_claim_file; - - if (callback(claim_file_hook) != LDPS_OK) - return LDPS_ERR; - - registeredClaimFile = true; - } break; - case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: { - ld_plugin_register_all_symbols_read callback; - callback = tv->tv_u.tv_register_all_symbols_read; - - if (callback(all_symbols_read_hook) != LDPS_OK) - return LDPS_ERR; - - RegisteredAllSymbolsRead = true; - } break; - case LDPT_REGISTER_CLEANUP_HOOK: { - ld_plugin_register_cleanup callback; - callback = tv->tv_u.tv_register_cleanup; - - if (callback(cleanup_hook) != LDPS_OK) - return LDPS_ERR; - } break; - case LDPT_GET_INPUT_FILE: - get_input_file = tv->tv_u.tv_get_input_file; - break; - case LDPT_RELEASE_INPUT_FILE: - release_input_file = tv->tv_u.tv_release_input_file; - break; - case LDPT_ADD_SYMBOLS: - add_symbols = tv->tv_u.tv_add_symbols; - break; - case LDPT_GET_SYMBOLS_V2: - get_symbols = tv->tv_u.tv_get_symbols; - break; - case LDPT_ADD_INPUT_FILE: - add_input_file = tv->tv_u.tv_add_input_file; - break; - case LDPT_SET_EXTRA_LIBRARY_PATH: - set_extra_library_path = tv->tv_u.tv_set_extra_library_path; - break; - case LDPT_GET_VIEW: - get_view = tv->tv_u.tv_get_view; + // Cast tv_tag to int to allow values not in "enum ld_plugin_tag", like, for + // example, LDPT_GET_SYMBOLS_V3 when building against an older plugin-api.h + // header. + switch (static_cast<int>(tv->tv_tag)) { + case LDPT_OUTPUT_NAME: + output_name = tv->tv_u.tv_string; + break; + case LDPT_LINKER_OUTPUT: + switch (tv->tv_u.tv_val) { + case LDPO_REL: // .o + case LDPO_DYN: // .so + case LDPO_PIE: // position independent executable + RelocationModel = Reloc::PIC_; break; - case LDPT_MESSAGE: - message = tv->tv_u.tv_message; + case LDPO_EXEC: // .exe + RelocationModel = Reloc::Static; break; default: - break; + message(LDPL_ERROR, "Unknown output file type %d", tv->tv_u.tv_val); + return LDPS_ERR; + } + break; + case LDPT_OPTION: + options::process_plugin_option(tv->tv_u.tv_string); + break; + case LDPT_REGISTER_CLAIM_FILE_HOOK: { + ld_plugin_register_claim_file callback; + callback = tv->tv_u.tv_register_claim_file; + + if (callback(claim_file_hook) != LDPS_OK) + return LDPS_ERR; + + registeredClaimFile = true; + } break; + case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: { + ld_plugin_register_all_symbols_read callback; + callback = tv->tv_u.tv_register_all_symbols_read; + + if (callback(all_symbols_read_hook) != LDPS_OK) + return LDPS_ERR; + + RegisteredAllSymbolsRead = true; + } break; + case LDPT_REGISTER_CLEANUP_HOOK: { + ld_plugin_register_cleanup callback; + callback = tv->tv_u.tv_register_cleanup; + + if (callback(cleanup_hook) != LDPS_OK) + return LDPS_ERR; + } break; + case LDPT_GET_INPUT_FILE: + get_input_file = tv->tv_u.tv_get_input_file; + break; + case LDPT_RELEASE_INPUT_FILE: + release_input_file = tv->tv_u.tv_release_input_file; + break; + case LDPT_ADD_SYMBOLS: + add_symbols = tv->tv_u.tv_add_symbols; + break; + case LDPT_GET_SYMBOLS_V2: + // Do not override get_symbols_v3 with get_symbols_v2. + if (!get_symbols) + get_symbols = tv->tv_u.tv_get_symbols; + break; + case LDPT_GET_SYMBOLS_V3: + get_symbols = tv->tv_u.tv_get_symbols; + break; + case LDPT_ADD_INPUT_FILE: + add_input_file = tv->tv_u.tv_add_input_file; + break; + case LDPT_SET_EXTRA_LIBRARY_PATH: + set_extra_library_path = tv->tv_u.tv_set_extra_library_path; + break; + case LDPT_GET_VIEW: + get_view = tv->tv_u.tv_get_view; + break; + case LDPT_MESSAGE: + message = tv->tv_u.tv_message; + break; + default: + break; } } @@ -305,7 +384,7 @@ ld_plugin_status onload(ld_plugin_tv *tv) { return LDPS_ERR; } if (!release_input_file) { - message(LDPL_ERROR, "relesase_input_file not passed to LLVMgold."); + message(LDPL_ERROR, "release_input_file not passed to LLVMgold."); return LDPS_ERR; } @@ -344,7 +423,6 @@ static void diagnosticHandler(const DiagnosticInfo &DI) { case DS_Error: message(LDPL_FATAL, "LLVM gold plugin has failed to create LTO module: %s", ErrStorage.c_str()); - llvm_unreachable("Fatal doesn't return."); case DS_Warning: Level = LDPL_WARNING; break; @@ -428,11 +506,23 @@ static ld_plugin_status claim_file_hook(const ld_plugin_input_file *file, claimed_file &cf = Modules.back(); cf.handle = file->handle; - - // If we are doing ThinLTO compilation, don't need to process the symbols. - // Later we simply build a combined index file after all files are claimed. - if (options::thinlto) - return LDPS_OK; + // Keep track of the first handle for each file descriptor, since there are + // multiple in the case of an archive. This is used later in the case of + // ThinLTO parallel backends to ensure that each file is only opened and + // released once. + auto LeaderHandle = + FDToLeaderHandle.insert(std::make_pair(file->fd, file->handle)).first; + cf.leader_handle = LeaderHandle->second; + // Save the filesize since for parallel ThinLTO backends we can only + // invoke get_input_file once per archive (only for the leader handle). + cf.filesize = file->filesize; + // In the case of an archive library, all but the first member must have a + // non-zero offset, which we can append to the file name to obtain a + // unique name. + cf.name = file->name; + if (file->offset) + cf.name += ".llvm." + std::to_string(file->offset) + "." + + sys::path::filename(Obj->getModule().getSourceFileName()).str(); for (auto &Sym : Obj->symbols()) { uint32_t Symflags = Sym.getFlags(); @@ -456,21 +546,12 @@ static ld_plugin_status claim_file_hook(const ld_plugin_input_file *file, sym.visibility = LDPV_DEFAULT; if (GV) { - Res.UnnamedAddr &= GV->hasUnnamedAddr(); + Res.UnnamedAddr = + GlobalValue::getMinUnnamedAddr(Res.UnnamedAddr, GV->getUnnamedAddr()); Res.IsLinkonceOdr &= GV->hasLinkOnceLinkage(); - if (GV->hasCommonLinkage()) { - Res.CommonAlign = std::max(Res.CommonAlign, GV->getAlignment()); - const DataLayout &DL = GV->getParent()->getDataLayout(); - uint64_t Size = DL.getTypeAllocSize(GV->getType()->getElementType()); - if (Size >= Res.CommonSize) { - Res.CommonSize = Size; - Res.CommonFile = &cf; - } - } Res.Visibility = getMinVisibility(Res.Visibility, GV->getVisibility()); switch (GV->getVisibility()) { case GlobalValue::DefaultVisibility: - sym.visibility = LDPV_DEFAULT; break; case GlobalValue::HiddenVisibility: sym.visibility = LDPV_HIDDEN; @@ -561,52 +642,56 @@ static void freeSymName(ld_plugin_symbol &Sym) { Sym.comdat_key = nullptr; } -static std::unique_ptr<FunctionInfoIndex> -getFunctionIndexForFile(claimed_file &F, ld_plugin_input_file &Info) { +/// Helper to get a file's symbols and a view into it via gold callbacks. +static const void *getSymbolsAndView(claimed_file &F) { + ld_plugin_status status = get_symbols(F.handle, F.syms.size(), F.syms.data()); + if (status == LDPS_NO_SYMS) + return nullptr; - if (get_symbols(F.handle, F.syms.size(), &F.syms[0]) != LDPS_OK) + if (status != LDPS_OK) message(LDPL_FATAL, "Failed to get symbol information"); const void *View; if (get_view(F.handle, &View) != LDPS_OK) message(LDPL_FATAL, "Failed to get a view of file"); - MemoryBufferRef BufferRef(StringRef((const char *)View, Info.filesize), - Info.name); + return View; +} + +static std::unique_ptr<ModuleSummaryIndex> +getModuleSummaryIndexForFile(claimed_file &F) { + const void *View = getSymbolsAndView(F); + if (!View) + return nullptr; + + MemoryBufferRef BufferRef(StringRef((const char *)View, F.filesize), F.name); // Don't bother trying to build an index if there is no summary information // in this bitcode file. - if (!object::FunctionIndexObjectFile::hasFunctionSummaryInMemBuffer( + if (!object::ModuleSummaryIndexObjectFile::hasGlobalValueSummaryInMemBuffer( BufferRef, diagnosticHandler)) - return std::unique_ptr<FunctionInfoIndex>(nullptr); + return std::unique_ptr<ModuleSummaryIndex>(nullptr); - ErrorOr<std::unique_ptr<object::FunctionIndexObjectFile>> ObjOrErr = - object::FunctionIndexObjectFile::create(BufferRef, diagnosticHandler); + ErrorOr<std::unique_ptr<object::ModuleSummaryIndexObjectFile>> ObjOrErr = + object::ModuleSummaryIndexObjectFile::create(BufferRef, + diagnosticHandler); if (std::error_code EC = ObjOrErr.getError()) - message(LDPL_FATAL, "Could not read function index bitcode from file : %s", + message(LDPL_FATAL, + "Could not read module summary index bitcode from file : %s", EC.message().c_str()); - object::FunctionIndexObjectFile &Obj = **ObjOrErr; + object::ModuleSummaryIndexObjectFile &Obj = **ObjOrErr; return Obj.takeIndex(); } static std::unique_ptr<Module> -getModuleForFile(LLVMContext &Context, claimed_file &F, - ld_plugin_input_file &Info, raw_fd_ostream *ApiFile, - StringSet<> &Internalize, StringSet<> &Maybe, - std::vector<GlobalValue *> &Keep) { - - if (get_symbols(F.handle, F.syms.size(), F.syms.data()) != LDPS_OK) - message(LDPL_FATAL, "Failed to get symbol information"); - - const void *View; - if (get_view(F.handle, &View) != LDPS_OK) - message(LDPL_FATAL, "Failed to get a view of file"); - - MemoryBufferRef BufferRef(StringRef((const char *)View, Info.filesize), - Info.name); +getModuleForFile(LLVMContext &Context, claimed_file &F, const void *View, + StringRef Name, raw_fd_ostream *ApiFile, + StringSet<> &Internalize, std::vector<GlobalValue *> &Keep, + StringMap<unsigned> &Realign) { + MemoryBufferRef BufferRef(StringRef((const char *)View, F.filesize), Name); ErrorOr<std::unique_ptr<object::IRObjectFile>> ObjOrErr = object::IRObjectFile::create(BufferRef, Context); @@ -656,7 +741,6 @@ getModuleForFile(LLVMContext &Context, claimed_file &F, // Override gold's resolution for common symbols. We want the largest // one to win. if (GV->hasCommonLinkage()) { - cast<GlobalVariable>(GV)->setAlignment(Res.CommonAlign); if (Resolution == LDPR_PREVAILING_DEF_IRONLY) Res.CommonInternal = true; @@ -664,14 +748,29 @@ getModuleForFile(LLVMContext &Context, claimed_file &F, Resolution == LDPR_PREVAILING_DEF) Res.UseCommon = true; - if (Res.CommonFile == &F && Res.UseCommon) { + const DataLayout &DL = GV->getParent()->getDataLayout(); + uint64_t Size = DL.getTypeAllocSize(GV->getType()->getElementType()); + unsigned Align = GV->getAlignment(); + + if (Res.UseCommon && Size >= Res.CommonSize) { + // Take GV. if (Res.CommonInternal) Resolution = LDPR_PREVAILING_DEF_IRONLY; else Resolution = LDPR_PREVAILING_DEF; + cast<GlobalVariable>(GV)->setAlignment( + std::max(Res.CommonAlign, Align)); } else { + // Do not take GV, it's smaller than what we already have in the + // combined module. Resolution = LDPR_PREEMPTED_IR; + if (Align > Res.CommonAlign) + // Need to raise the alignment though. + Realign[Sym.name] = Align; } + + Res.CommonSize = std::max(Res.CommonSize, Size); + Res.CommonAlign = std::max(Res.CommonAlign, Align); } switch (Resolution) { @@ -715,12 +814,9 @@ getModuleForFile(LLVMContext &Context, claimed_file &F, break; case LDPR_PREVAILING_DEF_IRONLY_EXP: { - // We can only check for address uses after we merge the modules. The - // reason is that this GV might have a copy in another module - // and in that module the address might be significant, but that - // copy will be LDPR_PREEMPTED_IR. - Maybe.insert(GV->getName()); Keep.push_back(GV); + if (canBeOmittedFromSymbolTable(GV)) + Internalize.insert(GV->getName()); break; } } @@ -731,26 +827,6 @@ getModuleForFile(LLVMContext &Context, claimed_file &F, return Obj.takeModule(); } -static void runLTOPasses(Module &M, TargetMachine &TM) { - M.setDataLayout(TM.createDataLayout()); - - legacy::PassManager passes; - passes.add(createTargetTransformInfoWrapperPass(TM.getTargetIRAnalysis())); - - PassManagerBuilder PMB; - PMB.LibraryInfo = new TargetLibraryInfoImpl(Triple(TM.getTargetTriple())); - PMB.Inliner = createFunctionInliningPass(); - // Unconditionally verify input since it is not verified before this - // point and has unknown origin. - PMB.VerifyInput = true; - PMB.VerifyOutput = !options::DisableVerify; - PMB.LoopVectorize = true; - PMB.SLPVectorize = true; - PMB.OptLevel = options::OptLevel; - PMB.populateLTOPassManager(passes); - passes.run(M); -} - static void saveBCFile(StringRef Path, Module &M) { std::error_code EC; raw_fd_ostream OS(Path, EC, sys::fs::OpenFlags::F_None); @@ -759,156 +835,660 @@ static void saveBCFile(StringRef Path, Module &M) { WriteBitcodeToFile(&M, OS, /* ShouldPreserveUseListOrder */ false); } -static void codegen(std::unique_ptr<Module> M) { - const std::string &TripleStr = M->getTargetTriple(); - Triple TheTriple(TripleStr); +static void recordFile(std::string Filename, bool TempOutFile) { + if (add_input_file(Filename.c_str()) != LDPS_OK) + message(LDPL_FATAL, + "Unable to add .o file to the link. File left behind in: %s", + Filename.c_str()); + if (TempOutFile) + Cleanup.push_back(Filename.c_str()); +} - std::string ErrMsg; - const Target *TheTarget = TargetRegistry::lookupTarget(TripleStr, ErrMsg); - if (!TheTarget) - message(LDPL_FATAL, "Target not found: %s", ErrMsg.c_str()); +void ThinLTOTaskInfo::cleanup() { + // Close the output file descriptor before we pass it to gold. + OS->close(); - if (unsigned NumOpts = options::extra.size()) - cl::ParseCommandLineOptions(NumOpts, &options::extra[0]); + recordFile(Filename, TempOutFile); +} + +namespace { +/// Class to manage optimization and code generation for a module, possibly +/// in a thread (ThinLTO). +class CodeGen { + /// The module for which this will generate code. + std::unique_ptr<llvm::Module> M; + + /// The output stream to generate code into. + raw_fd_ostream *OS; + + /// The task ID when this was invoked in a thread (ThinLTO). + int TaskID; + + /// The module summary index for ThinLTO tasks. + const ModuleSummaryIndex *CombinedIndex; + + /// The target machine for generating code for this module. + std::unique_ptr<TargetMachine> TM; + + /// Filename to use as base when save-temps is enabled, used to get + /// a unique and identifiable save-temps output file for each ThinLTO backend. + std::string SaveTempsFilename; + + /// Map from a module name to the corresponding buffer holding a view of the + /// bitcode provided via the get_view gold callback. + StringMap<MemoryBufferRef> *ModuleMap; + // Functions to import into this module. + FunctionImporter::ImportMapTy *ImportList; + + // Map of globals defined in this module to their summary. + std::map<GlobalValue::GUID, GlobalValueSummary *> *DefinedGlobals; + +public: + /// Constructor used by full LTO. + CodeGen(std::unique_ptr<llvm::Module> M) + : M(std::move(M)), OS(nullptr), TaskID(-1), CombinedIndex(nullptr), + ModuleMap(nullptr) { + initTargetMachine(); + } + /// Constructor used by ThinLTO. + CodeGen(std::unique_ptr<llvm::Module> M, raw_fd_ostream *OS, int TaskID, + const ModuleSummaryIndex *CombinedIndex, std::string Filename, + StringMap<MemoryBufferRef> *ModuleMap, + FunctionImporter::ImportMapTy *ImportList, + std::map<GlobalValue::GUID, GlobalValueSummary *> *DefinedGlobals) + : M(std::move(M)), OS(OS), TaskID(TaskID), CombinedIndex(CombinedIndex), + SaveTempsFilename(std::move(Filename)), ModuleMap(ModuleMap), + ImportList(ImportList), DefinedGlobals(DefinedGlobals) { + assert(options::thinlto == !!CombinedIndex && + "Expected module summary index iff performing ThinLTO"); + initTargetMachine(); + } + + /// Invoke LTO passes and the code generator for the module. + void runAll(); + + /// Invoke the actual code generation to emit Module's object to file. + void runCodegenPasses(); + +private: + const Target *TheTarget; + std::string TripleStr; + std::string FeaturesString; + TargetOptions Options; + + /// Create a target machine for the module. Must be unique for each + /// module/task. + void initTargetMachine(); + + std::unique_ptr<TargetMachine> createTargetMachine(); + + /// Run all LTO passes on the module. + void runLTOPasses(); + + /// Sets up output files necessary to perform optional multi-threaded + /// split code generation, and invokes the code generation implementation. + /// If BCFileName is not empty, saves bitcode for module partitions into + /// {BCFileName}0 .. {BCFileName}N. + void runSplitCodeGen(const SmallString<128> &BCFilename); +}; +} + +static SubtargetFeatures getFeatures(Triple &TheTriple) { SubtargetFeatures Features; Features.getDefaultSubtargetFeatures(TheTriple); for (const std::string &A : MAttrs) Features.AddFeature(A); + return Features; +} - TargetOptions Options = InitTargetOptionsFromCodeGenFlags(); - CodeGenOpt::Level CGOptLevel; +static CodeGenOpt::Level getCGOptLevel() { switch (options::OptLevel) { case 0: - CGOptLevel = CodeGenOpt::None; - break; + return CodeGenOpt::None; case 1: - CGOptLevel = CodeGenOpt::Less; - break; + return CodeGenOpt::Less; case 2: - CGOptLevel = CodeGenOpt::Default; - break; + return CodeGenOpt::Default; case 3: - CGOptLevel = CodeGenOpt::Aggressive; - break; + return CodeGenOpt::Aggressive; } - std::unique_ptr<TargetMachine> TM(TheTarget->createTargetMachine( - TripleStr, options::mcpu, Features.getString(), Options, RelocationModel, + llvm_unreachable("Invalid optimization level"); +} + +void CodeGen::initTargetMachine() { + TripleStr = M->getTargetTriple(); + Triple TheTriple(TripleStr); + + std::string ErrMsg; + TheTarget = TargetRegistry::lookupTarget(TripleStr, ErrMsg); + if (!TheTarget) + message(LDPL_FATAL, "Target not found: %s", ErrMsg.c_str()); + + SubtargetFeatures Features = getFeatures(TheTriple); + FeaturesString = Features.getString(); + Options = InitTargetOptionsFromCodeGenFlags(); + + // Disable the new X86 relax relocations since gold might not support them. + // FIXME: Check the gold version or add a new option to enable them. + Options.RelaxELFRelocations = false; + + TM = createTargetMachine(); +} + +std::unique_ptr<TargetMachine> CodeGen::createTargetMachine() { + CodeGenOpt::Level CGOptLevel = getCGOptLevel(); + + return std::unique_ptr<TargetMachine>(TheTarget->createTargetMachine( + TripleStr, options::mcpu, FeaturesString, Options, RelocationModel, CodeModel::Default, CGOptLevel)); +} + +void CodeGen::runLTOPasses() { + M->setDataLayout(TM->createDataLayout()); + + if (CombinedIndex) { + // Apply summary-based LinkOnce/Weak resolution decisions. + thinLTOResolveWeakForLinkerModule(*M, *DefinedGlobals); + + // Apply summary-based internalization decisions. Skip if there are no + // defined globals from the summary since not only is it unnecessary, but + // if this module did not have a summary section the internalizer will + // assert if it finds any definitions in this module that aren't in the + // DefinedGlobals set. + if (!DefinedGlobals->empty()) + thinLTOInternalizeModule(*M, *DefinedGlobals); + + // Create a loader that will parse the bitcode from the buffers + // in the ModuleMap. + ModuleLoader Loader(M->getContext(), *ModuleMap); - runLTOPasses(*M, *TM); + // Perform function importing. + FunctionImporter Importer(*CombinedIndex, Loader); + Importer.importFunctions(*M, *ImportList); + } + + legacy::PassManager passes; + passes.add(createTargetTransformInfoWrapperPass(TM->getTargetIRAnalysis())); + + PassManagerBuilder PMB; + PMB.LibraryInfo = new TargetLibraryInfoImpl(Triple(TM->getTargetTriple())); + PMB.Inliner = createFunctionInliningPass(); + // Unconditionally verify input since it is not verified before this + // point and has unknown origin. + PMB.VerifyInput = true; + PMB.VerifyOutput = !options::DisableVerify; + PMB.LoopVectorize = true; + PMB.SLPVectorize = true; + PMB.OptLevel = options::OptLevel; + if (options::thinlto) + PMB.populateThinLTOPassManager(passes); + else + PMB.populateLTOPassManager(passes); + passes.run(*M); +} + +/// Open a file and return the new file descriptor given a base input +/// file name, a flag indicating whether a temp file should be generated, +/// and an optional task id. The new filename generated is +/// returned in \p NewFilename. +static int openOutputFile(SmallString<128> InFilename, bool TempOutFile, + SmallString<128> &NewFilename, int TaskID = -1) { + int FD; + if (TempOutFile) { + std::error_code EC = + sys::fs::createTemporaryFile("lto-llvm", "o", FD, NewFilename); + if (EC) + message(LDPL_FATAL, "Could not create temporary file: %s", + EC.message().c_str()); + } else { + NewFilename = InFilename; + if (TaskID >= 0) + NewFilename += utostr(TaskID); + std::error_code EC = + sys::fs::openFileForWrite(NewFilename, FD, sys::fs::F_None); + if (EC) + message(LDPL_FATAL, "Could not open file: %s", EC.message().c_str()); + } + return FD; +} - if (options::TheOutputType == options::OT_SAVE_TEMPS) - saveBCFile(output_name + ".opt.bc", *M); +void CodeGen::runCodegenPasses() { + assert(OS && "Output stream must be set before emitting to file"); + legacy::PassManager CodeGenPasses; + if (TM->addPassesToEmitFile(CodeGenPasses, *OS, + TargetMachine::CGFT_ObjectFile)) + report_fatal_error("Failed to setup codegen"); + CodeGenPasses.run(*M); +} +void CodeGen::runSplitCodeGen(const SmallString<128> &BCFilename) { SmallString<128> Filename; + // Note that openOutputFile will append a unique ID for each task if (!options::obj_path.empty()) Filename = options::obj_path; else if (options::TheOutputType == options::OT_SAVE_TEMPS) Filename = output_name + ".o"; - std::vector<SmallString<128>> Filenames(options::Parallelism); + // Note that the default parallelism is 1 instead of the + // hardware_concurrency, as there are behavioral differences between + // parallelism levels (e.g. symbol ordering will be different, and some uses + // of inline asm currently have issues with parallelism >1). + unsigned int MaxThreads = options::Parallelism ? options::Parallelism : 1; + + std::vector<SmallString<128>> Filenames(MaxThreads); + std::vector<SmallString<128>> BCFilenames(MaxThreads); bool TempOutFile = Filename.empty(); { - // Open a file descriptor for each backend thread. This is done in a block + // Open a file descriptor for each backend task. This is done in a block // so that the output file descriptors are closed before gold opens them. std::list<llvm::raw_fd_ostream> OSs; - std::vector<llvm::raw_pwrite_stream *> OSPtrs(options::Parallelism); - for (unsigned I = 0; I != options::Parallelism; ++I) { - int FD; - if (TempOutFile) { - std::error_code EC = - sys::fs::createTemporaryFile("lto-llvm", "o", FD, Filenames[I]); - if (EC) - message(LDPL_FATAL, "Could not create temporary file: %s", - EC.message().c_str()); - } else { - Filenames[I] = Filename; - if (options::Parallelism != 1) - Filenames[I] += utostr(I); - std::error_code EC = - sys::fs::openFileForWrite(Filenames[I], FD, sys::fs::F_None); - if (EC) - message(LDPL_FATAL, "Could not open file: %s", EC.message().c_str()); - } + std::vector<llvm::raw_pwrite_stream *> OSPtrs(MaxThreads); + for (unsigned I = 0; I != MaxThreads; ++I) { + int FD = openOutputFile(Filename, TempOutFile, Filenames[I], + // Only append ID if there are multiple tasks. + MaxThreads > 1 ? I : -1); OSs.emplace_back(FD, true); OSPtrs[I] = &OSs.back(); } - // Run backend threads. - splitCodeGen(std::move(M), OSPtrs, options::mcpu, Features.getString(), - Options, RelocationModel, CodeModel::Default, CGOptLevel); + std::list<llvm::raw_fd_ostream> BCOSs; + std::vector<llvm::raw_pwrite_stream *> BCOSPtrs; + if (!BCFilename.empty() && MaxThreads > 1) { + for (unsigned I = 0; I != MaxThreads; ++I) { + int FD = openOutputFile(BCFilename, false, BCFilenames[I], I); + BCOSs.emplace_back(FD, true); + BCOSPtrs.push_back(&BCOSs.back()); + } + } + + // Run backend tasks. + splitCodeGen(std::move(M), OSPtrs, BCOSPtrs, + [&]() { return createTargetMachine(); }); } - for (auto &Filename : Filenames) { - if (add_input_file(Filename.c_str()) != LDPS_OK) - message(LDPL_FATAL, - "Unable to add .o file to the link. File left behind in: %s", - Filename.c_str()); - if (TempOutFile) - Cleanup.push_back(Filename.c_str()); + for (auto &Filename : Filenames) + recordFile(Filename.c_str(), TempOutFile); +} + +void CodeGen::runAll() { + runLTOPasses(); + + SmallString<128> OptFilename; + if (options::TheOutputType == options::OT_SAVE_TEMPS) { + OptFilename = output_name; + // If the CodeGen client provided a filename, use it. Always expect + // a provided filename if we are in a task (i.e. ThinLTO backend). + assert(!SaveTempsFilename.empty() || TaskID == -1); + if (!SaveTempsFilename.empty()) + OptFilename = SaveTempsFilename; + OptFilename += ".opt.bc"; + saveBCFile(OptFilename, *M); } + + // If we are already in a thread (i.e. ThinLTO), just perform + // codegen passes directly. + if (TaskID >= 0) + runCodegenPasses(); + // Otherwise attempt split code gen. + else + runSplitCodeGen(OptFilename); } -/// gold informs us that all symbols have been read. At this point, we use -/// get_symbols to see if any of our definitions have been overridden by a -/// native object file. Then, perform optimization and codegen. -static ld_plugin_status allSymbolsReadHook(raw_fd_ostream *ApiFile) { - if (Modules.empty()) - return LDPS_OK; +/// Links the module in \p View from file \p F into the combined module +/// saved in the IRMover \p L. +static void linkInModule(LLVMContext &Context, IRMover &L, claimed_file &F, + const void *View, StringRef Name, + raw_fd_ostream *ApiFile, StringSet<> &Internalize, + bool SetName = false) { + std::vector<GlobalValue *> Keep; + StringMap<unsigned> Realign; + std::unique_ptr<Module> M = getModuleForFile(Context, F, View, Name, ApiFile, + Internalize, Keep, Realign); + if (!M.get()) + return; + if (!options::triple.empty()) + M->setTargetTriple(options::triple.c_str()); + else if (M->getTargetTriple().empty()) { + M->setTargetTriple(DefaultTriple); + } + + // For ThinLTO we want to propagate the source file name to ensure + // we can create the correct global identifiers matching those in the + // original module. + if (SetName) + L.getModule().setSourceFileName(M->getSourceFileName()); + + if (Error E = L.move(std::move(M), Keep, + [](GlobalValue &, IRMover::ValueAdder) {})) { + handleAllErrors(std::move(E), [&](const llvm::ErrorInfoBase &EIB) { + message(LDPL_FATAL, "Failed to link module %s: %s", Name.str().c_str(), + EIB.message().c_str()); + }); + } + + for (const auto &I : Realign) { + GlobalValue *Dst = L.getModule().getNamedValue(I.first()); + if (!Dst) + continue; + cast<GlobalVariable>(Dst)->setAlignment(I.second); + } +} + +/// Perform the ThinLTO backend on a single module, invoking the LTO and codegen +/// pipelines. +static void thinLTOBackendTask(claimed_file &F, const void *View, + StringRef Name, raw_fd_ostream *ApiFile, + const ModuleSummaryIndex &CombinedIndex, + raw_fd_ostream *OS, unsigned TaskID, + StringMap<MemoryBufferRef> &ModuleMap, + FunctionImporter::ImportMapTy &ImportList, + std::map<GlobalValue::GUID, GlobalValueSummary *> &DefinedGlobals) { + // Need to use a separate context for each task + LLVMContext Context; + Context.setDiscardValueNames(options::TheOutputType != + options::OT_SAVE_TEMPS); + Context.enableDebugTypeODRUniquing(); // Merge debug info types. + Context.setDiagnosticHandler(diagnosticHandlerForContext, nullptr, true); + + std::unique_ptr<llvm::Module> NewModule(new llvm::Module(Name, Context)); + IRMover L(*NewModule.get()); + + StringSet<> Dummy; + linkInModule(Context, L, F, View, Name, ApiFile, Dummy, true); + if (renameModuleForThinLTO(*NewModule, CombinedIndex)) + message(LDPL_FATAL, "Failed to rename module for ThinLTO"); - // If we are doing ThinLTO compilation, simply build the combined - // function index/summary and emit it. We don't need to parse the modules - // and link them in this case. - if (options::thinlto) { - FunctionInfoIndex CombinedIndex; - uint64_t NextModuleId = 0; + CodeGen codeGen(std::move(NewModule), OS, TaskID, &CombinedIndex, Name, + &ModuleMap, &ImportList, &DefinedGlobals); + codeGen.runAll(); +} + +/// Launch each module's backend pipeline in a separate task in a thread pool. +static void +thinLTOBackends(raw_fd_ostream *ApiFile, + const ModuleSummaryIndex &CombinedIndex, + StringMap<MemoryBufferRef> &ModuleMap, + StringMap<FunctionImporter::ImportMapTy> &ImportLists, + StringMap<std::map<GlobalValue::GUID, GlobalValueSummary *>> + &ModuleToDefinedGVSummaries) { + unsigned TaskCount = 0; + std::vector<ThinLTOTaskInfo> Tasks; + Tasks.reserve(Modules.size()); + unsigned int MaxThreads = options::Parallelism + ? options::Parallelism + : thread::hardware_concurrency(); + + // Create ThreadPool in nested scope so that threads will be joined + // on destruction. + { + ThreadPool ThinLTOThreadPool(MaxThreads); for (claimed_file &F : Modules) { - PluginInputFile InputFile(F.handle); + // Do all the gold callbacks in the main thread, since gold is not thread + // safe by default. + const void *View = getSymbolsAndView(F); + if (!View) + continue; + + SmallString<128> Filename; + if (!options::obj_path.empty()) + // Note that openOutputFile will append a unique ID for each task + Filename = options::obj_path; + else if (options::TheOutputType == options::OT_SAVE_TEMPS) { + // Use the input file name so that we get a unique and identifiable + // output file for each ThinLTO backend task. + Filename = F.name; + Filename += ".thinlto.o"; + } + bool TempOutFile = Filename.empty(); + + SmallString<128> NewFilename; + int FD = openOutputFile(Filename, TempOutFile, NewFilename, + // Only append the TaskID if we will use the + // non-unique obj_path. + !options::obj_path.empty() ? TaskCount : -1); + TaskCount++; + std::unique_ptr<raw_fd_ostream> OS = + llvm::make_unique<raw_fd_ostream>(FD, true); + + // Enqueue the task + ThinLTOThreadPool.async(thinLTOBackendTask, std::ref(F), View, F.name, + ApiFile, std::ref(CombinedIndex), OS.get(), + TaskCount, std::ref(ModuleMap), + std::ref(ImportLists[F.name]), + std::ref(ModuleToDefinedGVSummaries[F.name])); + + // Record the information needed by the task or during its cleanup + // to a ThinLTOTaskInfo instance. For information needed by the task + // the unique_ptr ownership is transferred to the ThinLTOTaskInfo. + Tasks.emplace_back(std::move(OS), NewFilename.c_str(), TempOutFile); + } + } + + for (auto &Task : Tasks) + Task.cleanup(); +} - std::unique_ptr<FunctionInfoIndex> Index = - getFunctionIndexForFile(F, InputFile.file()); +/// Parse the thinlto_prefix_replace option into the \p OldPrefix and +/// \p NewPrefix strings, if it was specified. +static void getThinLTOOldAndNewPrefix(std::string &OldPrefix, + std::string &NewPrefix) { + StringRef PrefixReplace = options::thinlto_prefix_replace; + assert(PrefixReplace.empty() || PrefixReplace.find(";") != StringRef::npos); + std::pair<StringRef, StringRef> Split = PrefixReplace.split(";"); + OldPrefix = Split.first.str(); + NewPrefix = Split.second.str(); +} + +/// Given the original \p Path to an output file, replace any path +/// prefix matching \p OldPrefix with \p NewPrefix. Also, create the +/// resulting directory if it does not yet exist. +static std::string getThinLTOOutputFile(const std::string &Path, + const std::string &OldPrefix, + const std::string &NewPrefix) { + if (OldPrefix.empty() && NewPrefix.empty()) + return Path; + SmallString<128> NewPath(Path); + llvm::sys::path::replace_path_prefix(NewPath, OldPrefix, NewPrefix); + StringRef ParentPath = llvm::sys::path::parent_path(NewPath.str()); + if (!ParentPath.empty()) { + // Make sure the new directory exists, creating it if necessary. + if (std::error_code EC = llvm::sys::fs::create_directories(ParentPath)) + llvm::errs() << "warning: could not create directory '" << ParentPath + << "': " << EC.message() << '\n'; + } + return NewPath.str(); +} + +/// Perform ThinLTO link, which creates the combined index file. +/// Also, either launch backend threads or (under thinlto-index-only) +/// emit individual index files for distributed backends and exit. +static ld_plugin_status thinLTOLink(raw_fd_ostream *ApiFile) { + // Map from a module name to the corresponding buffer holding a view of the + // bitcode provided via the get_view gold callback. + StringMap<MemoryBufferRef> ModuleMap; + // Map to own RAII objects that manage the file opening and releasing + // interfaces with gold. + DenseMap<void *, std::unique_ptr<PluginInputFile>> HandleToInputFile; + + // Keep track of symbols that must not be internalized because they + // are referenced outside of a single IR module. + DenseSet<GlobalValue::GUID> Preserve; + + // Keep track of the prevailing copy for each GUID, for use in resolving + // weak linkages. + DenseMap<GlobalValue::GUID, const GlobalValueSummary *> PrevailingCopy; + + ModuleSummaryIndex CombinedIndex; + uint64_t NextModuleId = 0; + for (claimed_file &F : Modules) { + if (!HandleToInputFile.count(F.leader_handle)) + HandleToInputFile.insert(std::make_pair( + F.leader_handle, llvm::make_unique<PluginInputFile>(F.handle))); + // Pass this into getModuleSummaryIndexForFile + const void *View = getSymbolsAndView(F); + if (!View) + continue; - // Skip files without a function summary. - if (Index) - CombinedIndex.mergeFrom(std::move(Index), ++NextModuleId); + MemoryBufferRef ModuleBuffer(StringRef((const char *)View, F.filesize), + F.name); + assert(ModuleMap.find(ModuleBuffer.getBufferIdentifier()) == + ModuleMap.end() && + "Expect unique Buffer Identifier"); + ModuleMap[ModuleBuffer.getBufferIdentifier()] = ModuleBuffer; + + std::unique_ptr<ModuleSummaryIndex> Index = getModuleSummaryIndexForFile(F); + + // Use gold's symbol resolution information to identify symbols referenced + // by more than a single IR module (i.e. referenced by multiple IR modules + // or by a non-IR module). Cross references introduced by importing are + // checked separately via the export lists. Also track the prevailing copy + // for later symbol resolution. + for (auto &Sym : F.syms) { + ld_plugin_symbol_resolution Resolution = + (ld_plugin_symbol_resolution)Sym.resolution; + GlobalValue::GUID SymGUID = GlobalValue::getGUID(Sym.name); + if (Resolution != LDPR_PREVAILING_DEF_IRONLY) + Preserve.insert(SymGUID); + + if (Index && (Resolution == LDPR_PREVAILING_DEF || + Resolution == LDPR_PREVAILING_DEF_IRONLY || + Resolution == LDPR_PREVAILING_DEF_IRONLY_EXP)) + PrevailingCopy[SymGUID] = Index->getGlobalValueSummary(SymGUID); } + // Skip files without a module summary. + if (Index) + CombinedIndex.mergeFrom(std::move(Index), ++NextModuleId); + } + + // Collect for each module the list of function it defines (GUID -> + // Summary). + StringMap<std::map<GlobalValue::GUID, GlobalValueSummary *>> + ModuleToDefinedGVSummaries(NextModuleId); + CombinedIndex.collectDefinedGVSummariesPerModule(ModuleToDefinedGVSummaries); + + StringMap<FunctionImporter::ImportMapTy> ImportLists(NextModuleId); + StringMap<FunctionImporter::ExportSetTy> ExportLists(NextModuleId); + ComputeCrossModuleImport(CombinedIndex, ModuleToDefinedGVSummaries, + ImportLists, ExportLists); + + auto isPrevailing = [&](GlobalValue::GUID GUID, const GlobalValueSummary *S) { + const auto &Prevailing = PrevailingCopy.find(GUID); + assert(Prevailing != PrevailingCopy.end()); + return Prevailing->second == S; + }; + + // Callback for internalization, to prevent internalization of symbols + // that were not candidates initially, and those that are being imported + // (which introduces new cross references). + auto isExported = [&](StringRef ModuleIdentifier, GlobalValue::GUID GUID) { + const auto &ExportList = ExportLists.find(ModuleIdentifier); + return (ExportList != ExportLists.end() && + ExportList->second.count(GUID)) || + Preserve.count(GUID); + }; + + thinLTOResolveWeakForLinkerInIndex( + CombinedIndex, isPrevailing, + [](StringRef ModuleIdentifier, GlobalValue::GUID GUID, + GlobalValue::LinkageTypes NewLinkage) {}); + + // Use global summary-based analysis to identify symbols that can be + // internalized (because they aren't exported or preserved as per callback). + // Changes are made in the index, consumed in the ThinLTO backends. + thinLTOInternalizeAndPromoteInIndex(CombinedIndex, isExported); + + if (options::thinlto_emit_imports_files && !options::thinlto_index_only) + message(LDPL_WARNING, + "thinlto-emit-imports-files ignored unless thinlto-index-only"); + + if (options::thinlto_index_only) { + // If the thinlto-prefix-replace option was specified, parse it and + // extract the old and new prefixes. + std::string OldPrefix, NewPrefix; + getThinLTOOldAndNewPrefix(OldPrefix, NewPrefix); + + // For each input bitcode file, generate an individual index that + // contains summaries only for its own global values, and for any that + // should be imported. + for (claimed_file &F : Modules) { + std::error_code EC; + + std::string NewModulePath = + getThinLTOOutputFile(F.name, OldPrefix, NewPrefix); + raw_fd_ostream OS((Twine(NewModulePath) + ".thinlto.bc").str(), EC, + sys::fs::OpenFlags::F_None); + if (EC) + message(LDPL_FATAL, "Unable to open %s.thinlto.bc for writing: %s", + NewModulePath.c_str(), EC.message().c_str()); + // Build a map of module to the GUIDs and summary objects that should + // be written to its index. + std::map<std::string, GVSummaryMapTy> ModuleToSummariesForIndex; + gatherImportedSummariesForModule(F.name, ModuleToDefinedGVSummaries, + ImportLists, ModuleToSummariesForIndex); + WriteIndexToFile(CombinedIndex, OS, &ModuleToSummariesForIndex); + + if (options::thinlto_emit_imports_files) { + if ((EC = EmitImportsFiles(F.name, + (Twine(NewModulePath) + ".imports").str(), + ImportLists))) + message(LDPL_FATAL, "Unable to open %s.imports", + NewModulePath.c_str(), EC.message().c_str()); + } + } + + cleanup_hook(); + exit(0); + } + + // Create OS in nested scope so that it will be closed on destruction. + { std::error_code EC; raw_fd_ostream OS(output_name + ".thinlto.bc", EC, sys::fs::OpenFlags::F_None); if (EC) message(LDPL_FATAL, "Unable to open %s.thinlto.bc for writing: %s", output_name.data(), EC.message().c_str()); - WriteFunctionSummaryToFile(CombinedIndex, OS); - OS.close(); - - cleanup_hook(); - exit(0); + WriteIndexToFile(CombinedIndex, OS); } + thinLTOBackends(ApiFile, CombinedIndex, ModuleMap, ImportLists, + ModuleToDefinedGVSummaries); + return LDPS_OK; +} + +/// gold informs us that all symbols have been read. At this point, we use +/// get_symbols to see if any of our definitions have been overridden by a +/// native object file. Then, perform optimization and codegen. +static ld_plugin_status allSymbolsReadHook(raw_fd_ostream *ApiFile) { + if (Modules.empty()) + return LDPS_OK; + + if (unsigned NumOpts = options::extra.size()) + cl::ParseCommandLineOptions(NumOpts, &options::extra[0]); + + if (options::thinlto) + return thinLTOLink(ApiFile); + LLVMContext Context; + Context.setDiscardValueNames(options::TheOutputType != + options::OT_SAVE_TEMPS); + Context.enableDebugTypeODRUniquing(); // Merge debug info types. Context.setDiagnosticHandler(diagnosticHandlerForContext, nullptr, true); std::unique_ptr<Module> Combined(new Module("ld-temp.o", Context)); IRMover L(*Combined); - std::string DefaultTriple = sys::getDefaultTargetTriple(); - StringSet<> Internalize; - StringSet<> Maybe; for (claimed_file &F : Modules) { + // RAII object to manage the file opening and releasing interfaces with + // gold. PluginInputFile InputFile(F.handle); - std::vector<GlobalValue *> Keep; - std::unique_ptr<Module> M = getModuleForFile( - Context, F, InputFile.file(), ApiFile, Internalize, Maybe, Keep); - if (!options::triple.empty()) - M->setTargetTriple(options::triple.c_str()); - else if (M->getTargetTriple().empty()) - M->setTargetTriple(DefaultTriple); - - if (L.move(*M, Keep, [](GlobalValue &, IRMover::ValueAdder) {})) - message(LDPL_FATAL, "Failed to link module"); + const void *View = getSymbolsAndView(F); + if (!View) + continue; + linkInModule(Context, L, F, View, F.name, ApiFile, Internalize); } for (const auto &Name : Internalize) { @@ -917,15 +1497,6 @@ static ld_plugin_status allSymbolsReadHook(raw_fd_ostream *ApiFile) { internalize(*GV); } - for (const auto &Name : Maybe) { - GlobalValue *GV = Combined->getNamedValue(Name.first()); - if (!GV) - continue; - GV->setLinkage(GlobalValue::LinkOnceODRLinkage); - if (canBeOmittedFromSymbolTable(GV)) - internalize(*GV); - } - if (options::TheOutputType == options::OT_DISABLE) return LDPS_OK; @@ -940,7 +1511,8 @@ static ld_plugin_status allSymbolsReadHook(raw_fd_ostream *ApiFile) { return LDPS_OK; } - codegen(std::move(Combined)); + CodeGen codeGen(std::move(Combined)); + codeGen.runAll(); if (!options::extra_library_path.empty() && set_extra_library_path(options::extra_library_path.c_str()) != LDPS_OK) @@ -966,10 +1538,14 @@ static ld_plugin_status all_symbols_read_hook(void) { if (options::TheOutputType == options::OT_BC_ONLY || options::TheOutputType == options::OT_DISABLE) { - if (options::TheOutputType == options::OT_DISABLE) + if (options::TheOutputType == options::OT_DISABLE) { // Remove the output file here since ld.bfd creates the output file // early. - sys::fs::remove(output_name); + std::error_code EC = sys::fs::remove(output_name); + if (EC) + message(LDPL_ERROR, "Failed to delete '%s': %s", output_name.c_str(), + EC.message().c_str()); + } exit(0); } diff --git a/tools/llc/Makefile b/tools/llc/Makefile deleted file mode 100644 index cd34c80d840a8..0000000000000 --- a/tools/llc/Makefile +++ /dev/null @@ -1,18 +0,0 @@ -#===- tools/llc/Makefile -----------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llc -LINK_COMPONENTS := all-targets bitreader asmparser irreader mirparser transformutils - -# Support plugins. -NO_DEAD_STRIP := 1 - -include $(LEVEL)/Makefile.common - diff --git a/tools/llc/llc.cpp b/tools/llc/llc.cpp index bffa39fd9e5c5..9b6086ad4ea83 100644 --- a/tools/llc/llc.cpp +++ b/tools/llc/llc.cpp @@ -21,7 +21,12 @@ #include "llvm/CodeGen/LinkAllAsmWriterComponents.h" #include "llvm/CodeGen/LinkAllCodegenComponents.h" #include "llvm/CodeGen/MIRParser/MIRParser.h" +#include "llvm/CodeGen/MachineFunctionPass.h" +#include "llvm/CodeGen/MachineModuleInfo.h" +#include "llvm/CodeGen/TargetPassConfig.h" #include "llvm/IR/DataLayout.h" +#include "llvm/IR/DiagnosticInfo.h" +#include "llvm/IR/DiagnosticPrinter.h" #include "llvm/IR/IRPrintingPasses.h" #include "llvm/IR/LLVMContext.h" #include "llvm/IR/LegacyPassManager.h" @@ -68,6 +73,11 @@ static cl::opt<bool> NoIntegratedAssembler("no-integrated-as", cl::Hidden, cl::desc("Disable integrated assembler")); +static cl::opt<bool> + PreserveComments("preserve-as-comments", cl::Hidden, + cl::desc("Preserve Comments in outputted assembly"), + cl::init(true)); + // Determine optimization level. static cl::opt<char> OptLevel("O", @@ -103,6 +113,33 @@ static cl::opt<bool> "manager and verify the result is the same."), cl::init(false)); +static cl::opt<bool> DiscardValueNames( + "discard-value-names", + cl::desc("Discard names from Value (other than GlobalValue)."), + cl::init(false), cl::Hidden); + +namespace { +static ManagedStatic<std::vector<std::string>> RunPassNames; + +struct RunPassOption { + void operator=(const std::string &Val) const { + if (Val.empty()) + return; + SmallVector<StringRef, 8> PassNames; + StringRef(Val).split(PassNames, ',', -1, false); + for (auto PassName : PassNames) + RunPassNames->push_back(PassName); + } +}; +} + +static RunPassOption RunPassOpt; + +static cl::opt<RunPassOption, true, cl::parser<std::string>> RunPass( + "run-pass", + cl::desc("Run compiler only for specified passes (comma separated list)"), + cl::value_desc("pass-name"), cl::ZeroOrMore, cl::location(RunPassOpt)); + static int compileModule(char **, LLVMContext &); static std::unique_ptr<tool_output_file> @@ -173,16 +210,27 @@ GetOutputStream(const char *TargetName, Triple::OSType OS, return FDOut; } +static void DiagnosticHandler(const DiagnosticInfo &DI, void *Context) { + bool *HasError = static_cast<bool *>(Context); + if (DI.getSeverity() == DS_Error) + *HasError = true; + + DiagnosticPrinterRawOStream DP(errs()); + errs() << LLVMContext::getDiagnosticMessagePrefix(DI.getSeverity()) << ": "; + DI.print(DP); + errs() << "\n"; +} + // main - Entry point for the llc compiler. // int main(int argc, char **argv) { - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); // Enable debug stream buffering. EnableDebugBuffering = true; - LLVMContext &Context = getGlobalContext(); + LLVMContext Context; llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. // Initialize targets first, so that --version shows registered targets. @@ -198,13 +246,19 @@ int main(int argc, char **argv) { initializeCodeGen(*Registry); initializeLoopStrengthReducePass(*Registry); initializeLowerIntrinsicsPass(*Registry); - initializeUnreachableBlockElimPass(*Registry); + initializeUnreachableBlockElimLegacyPassPass(*Registry); // Register the target printer for --version. cl::AddExtraVersionPrinter(TargetRegistry::printRegisteredTargetsForVersion); cl::ParseCommandLineOptions(argc, argv, "llvm system compiler\n"); + Context.setDiscardValueNames(DiscardValueNames); + + // Set a diagnostic handler that doesn't exit on the first error + bool HasError = false; + Context.setDiagnosticHandler(DiagnosticHandler, &HasError); + // Compile the module TimeCompilations times to give better compile time // metrics. for (unsigned I = TimeCompilations; I; --I) @@ -213,6 +267,34 @@ int main(int argc, char **argv) { return 0; } +static bool addPass(PassManagerBase &PM, const char *argv0, + StringRef PassName, TargetPassConfig &TPC) { + if (PassName == "none") + return false; + + const PassRegistry *PR = PassRegistry::getPassRegistry(); + const PassInfo *PI = PR->getPassInfo(PassName); + if (!PI) { + errs() << argv0 << ": run-pass " << PassName << " is not registered.\n"; + return true; + } + + Pass *P; + if (PI->getTargetMachineCtor()) + P = PI->getTargetMachineCtor()(&TPC.getTM<TargetMachine>()); + else if (PI->getNormalCtor()) + P = PI->getNormalCtor()(); + else { + errs() << argv0 << ": cannot create pass: " << PI->getPassName() << "\n"; + return true; + } + std::string Banner = std::string("After ") + std::string(P->getPassName()); + PM.add(P); + TPC.printAndVerify(Banner); + + return false; +} + static int compileModule(char **argv, LLVMContext &Context) { // Load the module to be compiled... SMDiagnostic Err; @@ -227,10 +309,8 @@ static int compileModule(char **argv, LLVMContext &Context) { if (!SkipModule) { if (StringRef(InputFilename).endswith_lower(".mir")) { MIR = createMIRParserFromFile(InputFilename, Err, Context); - if (MIR) { + if (MIR) M = MIR->parseLLVMModule(); - assert(M && "parseLLVMModule should exit on failure"); - } } else M = parseIRFile(InputFilename, Err, Context); if (!M) { @@ -285,10 +365,11 @@ static int compileModule(char **argv, LLVMContext &Context) { Options.MCOptions.ShowMCEncoding = ShowMCEncoding; Options.MCOptions.MCUseDwarfDirectory = EnableDwarfDirectory; Options.MCOptions.AsmVerbose = AsmVerbose; + Options.MCOptions.PreserveAsmComments = PreserveComments; std::unique_ptr<TargetMachine> Target( TheTarget->createTargetMachine(TheTriple.getTriple(), CPUStr, FeaturesStr, - Options, RelocModel, CMModel, OLvl)); + Options, getRelocModel(), CMModel, OLvl)); assert(Target && "Could not allocate target machine!"); @@ -348,18 +429,28 @@ static int compileModule(char **argv, LLVMContext &Context) { AnalysisID StartAfterID = nullptr; AnalysisID StopAfterID = nullptr; const PassRegistry *PR = PassRegistry::getPassRegistry(); - if (!RunPass.empty()) { + if (!RunPassNames->empty()) { if (!StartAfter.empty() || !StopAfter.empty()) { errs() << argv[0] << ": start-after and/or stop-after passes are " "redundant when run-pass is specified.\n"; return 1; } - const PassInfo *PI = PR->getPassInfo(RunPass); - if (!PI) { - errs() << argv[0] << ": run-pass pass is not registered.\n"; + if (!MIR) { + errs() << argv[0] << ": run-pass needs a .mir input.\n"; return 1; } - StopAfterID = StartBeforeID = PI->getTypeInfo(); + LLVMTargetMachine &LLVMTM = static_cast<LLVMTargetMachine&>(*Target); + TargetPassConfig &TPC = *LLVMTM.createPassConfig(PM); + PM.add(&TPC); + LLVMTM.addMachineModuleInfo(PM); + LLVMTM.addMachineFunctionAnalysis(PM, MIR.get()); + TPC.printAndVerify(""); + + for (const std::string &RunPassName : *RunPassNames) { + if (addPass(PM, argv[0], RunPassName, TPC)) + return 1; + } + PM.add(createPrintMIRPass(*OS)); } else { if (!StartAfter.empty()) { const PassInfo *PI = PR->getPassInfo(StartAfter); @@ -377,14 +468,15 @@ static int compileModule(char **argv, LLVMContext &Context) { } StopAfterID = PI->getTypeInfo(); } - } - // Ask the target to add backend passes as necessary. - if (Target->addPassesToEmitFile(PM, *OS, FileType, NoVerify, StartBeforeID, - StartAfterID, StopAfterID, MIR.get())) { - errs() << argv[0] << ": target does not support generation of this" - << " file type!\n"; - return 1; + // Ask the target to add backend passes as necessary. + if (Target->addPassesToEmitFile(PM, *OS, FileType, NoVerify, + StartBeforeID, StartAfterID, StopAfterID, + MIR.get())) { + errs() << argv[0] << ": target does not support generation of this" + << " file type!\n"; + return 1; + } } // Before executing passes, print the final values of the LLVM options. @@ -404,6 +496,10 @@ static int compileModule(char **argv, LLVMContext &Context) { PM.run(*M); + auto HasError = *static_cast<bool *>(Context.getDiagnosticContext()); + if (HasError) + return 1; + // Compare the two outputs and make sure they're the same if (CompileTwice) { if (Buffer.size() != CompileTwiceBuffer.size() || diff --git a/tools/lli/CMakeLists.txt b/tools/lli/CMakeLists.txt index e317db6b713c9..2bdd066eb7a62 100644 --- a/tools/lli/CMakeLists.txt +++ b/tools/lli/CMakeLists.txt @@ -1,4 +1,6 @@ -add_subdirectory(ChildTarget) +if ( LLVM_INCLUDE_UTILS ) + add_subdirectory(ChildTarget) +endif() set(LLVM_LINK_COMPONENTS CodeGen diff --git a/tools/lli/ChildTarget/CMakeLists.txt b/tools/lli/ChildTarget/CMakeLists.txt index 0f851d593444d..e4fe0c7197f24 100644 --- a/tools/lli/ChildTarget/CMakeLists.txt +++ b/tools/lli/ChildTarget/CMakeLists.txt @@ -4,8 +4,7 @@ set(LLVM_LINK_COMPONENTS Support ) -add_llvm_executable(lli-child-target +add_llvm_utility(lli-child-target ChildTarget.cpp ) -set_target_properties(lli-child-target PROPERTIES FOLDER "Misc") diff --git a/tools/lli/ChildTarget/ChildTarget.cpp b/tools/lli/ChildTarget/ChildTarget.cpp index 0b75e20f83e9a..f6d2413655eac 100644 --- a/tools/lli/ChildTarget/ChildTarget.cpp +++ b/tools/lli/ChildTarget/ChildTarget.cpp @@ -1,4 +1,4 @@ -#include "llvm/ExecutionEngine/Orc/OrcArchitectureSupport.h" +#include "llvm/ExecutionEngine/Orc/OrcABISupport.h" #include "llvm/ExecutionEngine/Orc/OrcRemoteTargetServer.h" #include "llvm/Support/Debug.h" #include "llvm/Support/DynamicLibrary.h" @@ -12,11 +12,13 @@ using namespace llvm::orc; using namespace llvm::sys; #ifdef __x86_64__ -typedef OrcX86_64 HostOrcArch; +typedef OrcX86_64_SysV HostOrcArch; #else -typedef OrcGenericArchitecture HostOrcArch; +typedef OrcGenericABI HostOrcArch; #endif +ExitOnError ExitOnErr; + int main(int argc, char *argv[]) { if (argc != 3) { @@ -24,6 +26,8 @@ int main(int argc, char *argv[]) { return 1; } + ExitOnErr.setBanner(std::string(argv[0]) + ":"); + int InFD; int OutFD; { @@ -41,24 +45,29 @@ int main(int argc, char *argv[]) { return RTDyldMemoryManager::getSymbolAddressInProcess(Name); }; + auto RegisterEHFrames = [](uint8_t *Addr, uint32_t Size) { + RTDyldMemoryManager::registerEHFramesInProcess(Addr, Size); + }; + + auto DeregisterEHFrames = [](uint8_t *Addr, uint32_t Size) { + RTDyldMemoryManager::deregisterEHFramesInProcess(Addr, Size); + }; + FDRPCChannel Channel(InFD, OutFD); typedef remote::OrcRemoteTargetServer<FDRPCChannel, HostOrcArch> JITServer; - JITServer Server(Channel, SymbolLookup); + JITServer Server(Channel, SymbolLookup, RegisterEHFrames, DeregisterEHFrames); while (1) { - JITServer::JITProcId Id = JITServer::InvalidId; - if (auto EC = Server.getNextProcId(Id)) { - errs() << "Error: " << EC.message() << "\n"; - return 1; - } + uint32_t RawId; + ExitOnErr(Server.startReceivingFunction(Channel, RawId)); + auto Id = static_cast<JITServer::JITFuncId>(RawId); switch (Id) { case JITServer::TerminateSessionId: + ExitOnErr(Server.handleTerminateSession()); return 0; default: - if (auto EC = Server.handleKnownProcedure(Id)) { - errs() << "Error: " << EC.message() << "\n"; - return 1; - } + ExitOnErr(Server.handleKnownFunction(Id)); + break; } } diff --git a/tools/lli/ChildTarget/Makefile b/tools/lli/ChildTarget/Makefile deleted file mode 100644 index 0767ba1d724cf..0000000000000 --- a/tools/lli/ChildTarget/Makefile +++ /dev/null @@ -1,19 +0,0 @@ -##===- tools/lli/Makefile ------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../../.. -TOOLNAME := lli-child-target - -include $(LEVEL)/Makefile.config - -LINK_COMPONENTS := support OrcJIT - -SOURCES := ChildTarget.cpp - -include $(LLVM_SRC_ROOT)/Makefile.rules diff --git a/tools/lli/Makefile b/tools/lli/Makefile deleted file mode 100644 index 70d8c80e5c0af..0000000000000 --- a/tools/lli/Makefile +++ /dev/null @@ -1,31 +0,0 @@ -##===- tools/lli/Makefile ------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := lli - -PARALLEL_DIRS := ChildTarget - -include $(LEVEL)/Makefile.config - -LINK_COMPONENTS := mcjit orcjit instrumentation interpreter nativecodegen bitreader asmparser irreader selectiondag native - -# If Intel JIT Events support is confiured, link against the LLVM Intel JIT -# Events interface library -ifeq ($(USE_INTEL_JITEVENTS), 1) - LINK_COMPONENTS += debuginfodwarf inteljitevents object -endif - -# If oprofile support is confiured, link against the LLVM oprofile interface -# library -ifeq ($(USE_OPROFILE), 1) - LINK_COMPONENTS += oprofilejit -endif - -include $(LLVM_SRC_ROOT)/Makefile.rules diff --git a/tools/lli/OrcLazyJIT.cpp b/tools/lli/OrcLazyJIT.cpp index 7f483f742b80d..b13e7696627f0 100644 --- a/tools/lli/OrcLazyJIT.cpp +++ b/tools/lli/OrcLazyJIT.cpp @@ -8,7 +8,7 @@ //===----------------------------------------------------------------------===// #include "OrcLazyJIT.h" -#include "llvm/ExecutionEngine/Orc/OrcArchitectureSupport.h" +#include "llvm/ExecutionEngine/Orc/OrcABISupport.h" #include "llvm/Support/Debug.h" #include "llvm/Support/DynamicLibrary.h" #include <cstdio> @@ -46,31 +46,6 @@ namespace { cl::init(true), cl::Hidden); } -std::unique_ptr<OrcLazyJIT::CompileCallbackMgr> -OrcLazyJIT::createCompileCallbackMgr(Triple T) { - switch (T.getArch()) { - default: return nullptr; - - case Triple::x86_64: { - typedef orc::LocalJITCompileCallbackManager<orc::OrcX86_64> CCMgrT; - return llvm::make_unique<CCMgrT>(0); - } - } -} - -OrcLazyJIT::IndirectStubsManagerBuilder -OrcLazyJIT::createIndirectStubsMgrBuilder(Triple T) { - switch (T.getArch()) { - default: return nullptr; - - case Triple::x86_64: - return [](){ - return llvm::make_unique< - orc::LocalIndirectStubsManager<orc::OrcX86_64>>(); - }; - } -} - OrcLazyJIT::TransformFtor OrcLazyJIT::createDebugDumper() { switch (OrcDumpKind) { @@ -142,8 +117,8 @@ int llvm::runOrcLazyJIT(std::unique_ptr<Module> M, int ArgC, char* ArgV[]) { EngineBuilder EB; EB.setOptLevel(getOptLevel()); auto TM = std::unique_ptr<TargetMachine>(EB.selectTarget()); - auto CompileCallbackMgr = - OrcLazyJIT::createCompileCallbackMgr(Triple(TM->getTargetTriple())); + Triple T(TM->getTargetTriple()); + auto CompileCallbackMgr = orc::createLocalCompileCallbackManager(T, 0); // If we couldn't build the factory function then there must not be a callback // manager for this target. Bail out. @@ -153,8 +128,7 @@ int llvm::runOrcLazyJIT(std::unique_ptr<Module> M, int ArgC, char* ArgV[]) { return 1; } - auto IndirectStubsMgrBuilder = - OrcLazyJIT::createIndirectStubsMgrBuilder(Triple(TM->getTargetTriple())); + auto IndirectStubsMgrBuilder = orc::createLocalIndirectStubsManagerBuilder(T); // If we couldn't build a stubs-manager-builder for this target then bail out. if (!IndirectStubsMgrBuilder) { @@ -181,3 +155,4 @@ int llvm::runOrcLazyJIT(std::unique_ptr<Module> M, int ArgC, char* ArgV[]) { auto Main = fromTargetAddress<MainFnPtr>(MainSym.getAddress()); return Main(ArgC, ArgV); } + diff --git a/tools/lli/OrcLazyJIT.h b/tools/lli/OrcLazyJIT.h index 2a5b31d1bfb8e..733bdd8c04d5f 100644 --- a/tools/lli/OrcLazyJIT.h +++ b/tools/lli/OrcLazyJIT.h @@ -62,9 +62,6 @@ public: DtorRunner.runViaLayer(CODLayer); } - static std::unique_ptr<CompileCallbackMgr> createCompileCallbackMgr(Triple T); - static IndirectStubsManagerBuilder createIndirectStubsMgrBuilder(Triple T); - ModuleHandleT addModule(std::unique_ptr<Module> M) { // Attach a data-layout if one isn't already present. if (M->getDataLayout().isDefault()) @@ -82,12 +79,11 @@ public: // 1) Search the JIT symbols. // 2) Check for C++ runtime overrides. // 3) Search the host process (LLI)'s symbol table. - std::shared_ptr<RuntimeDyld::SymbolResolver> Resolver = + auto Resolver = orc::createLambdaResolver( [this](const std::string &Name) { if (auto Sym = CODLayer.findSymbol(Name, true)) - return RuntimeDyld::SymbolInfo(Sym.getAddress(), - Sym.getFlags()); + return Sym.toRuntimeDyldSymbol(); if (auto Sym = CXXRuntimeOverrides.searchOverrides(Name)) return Sym; diff --git a/tools/lli/RemoteJITUtils.h b/tools/lli/RemoteJITUtils.h index a3f3fa0fd3b41..15068d2faf6fd 100644 --- a/tools/lli/RemoteJITUtils.h +++ b/tools/lli/RemoteJITUtils.h @@ -16,6 +16,7 @@ #include "llvm/ExecutionEngine/Orc/RPCChannel.h" #include "llvm/ExecutionEngine/RTDyldMemoryManager.h" +#include <mutex> #if !defined(_MSC_VER) && !defined(__MINGW32__) #include <unistd.h> @@ -24,27 +25,47 @@ #endif /// RPC channel that reads from and writes from file descriptors. -class FDRPCChannel : public llvm::orc::remote::RPCChannel { +class FDRPCChannel final : public llvm::orc::remote::RPCChannel { public: FDRPCChannel(int InFD, int OutFD) : InFD(InFD), OutFD(OutFD) {} - std::error_code readBytes(char *Dst, unsigned Size) override { + llvm::Error readBytes(char *Dst, unsigned Size) override { assert(Dst && "Attempt to read into null."); - ssize_t ReadResult = ::read(InFD, Dst, Size); - if (ReadResult != (ssize_t)Size) - return std::error_code(errno, std::generic_category()); - return std::error_code(); + ssize_t Completed = 0; + while (Completed < static_cast<ssize_t>(Size)) { + ssize_t Read = ::read(InFD, Dst + Completed, Size - Completed); + if (Read <= 0) { + auto ErrNo = errno; + if (ErrNo == EAGAIN || ErrNo == EINTR) + continue; + else + return llvm::errorCodeToError( + std::error_code(errno, std::generic_category())); + } + Completed += Read; + } + return llvm::Error::success(); } - std::error_code appendBytes(const char *Src, unsigned Size) override { + llvm::Error appendBytes(const char *Src, unsigned Size) override { assert(Src && "Attempt to append from null."); - ssize_t WriteResult = ::write(OutFD, Src, Size); - if (WriteResult != (ssize_t)Size) - std::error_code(errno, std::generic_category()); - return std::error_code(); + ssize_t Completed = 0; + while (Completed < static_cast<ssize_t>(Size)) { + ssize_t Written = ::write(OutFD, Src + Completed, Size - Completed); + if (Written < 0) { + auto ErrNo = errno; + if (ErrNo == EAGAIN || ErrNo == EINTR) + continue; + else + return llvm::errorCodeToError( + std::error_code(errno, std::generic_category())); + } + Completed += Written; + } + return llvm::Error::success(); } - std::error_code send() override { return std::error_code(); } + llvm::Error send() override { return llvm::Error::success(); } private: int InFD, OutFD; diff --git a/tools/lli/lli.cpp b/tools/lli/lli.cpp index a76ec11fb1daf..92de5da472166 100644 --- a/tools/lli/lli.cpp +++ b/tools/lli/lli.cpp @@ -187,20 +187,15 @@ namespace { cl::desc("Disable JIT lazy compilation"), cl::init(false)); - cl::opt<Reloc::Model> - RelocModel("relocation-model", - cl::desc("Choose relocation model"), - cl::init(Reloc::Default), - cl::values( - clEnumValN(Reloc::Default, "default", - "Target default relocation model"), - clEnumValN(Reloc::Static, "static", - "Non-relocatable code"), - clEnumValN(Reloc::PIC_, "pic", - "Fully relocatable, position independent code"), - clEnumValN(Reloc::DynamicNoPIC, "dynamic-no-pic", - "Relocatable external references, non-relocatable code"), - clEnumValEnd)); + cl::opt<Reloc::Model> RelocModel( + "relocation-model", cl::desc("Choose relocation model"), + cl::values( + clEnumValN(Reloc::Static, "static", "Non-relocatable code"), + clEnumValN(Reloc::PIC_, "pic", + "Fully relocatable, position independent code"), + clEnumValN(Reloc::DynamicNoPIC, "dynamic-no-pic", + "Relocatable external references, non-relocatable code"), + clEnumValEnd)); cl::opt<llvm::CodeModel::Model> CMModel("code-model", @@ -235,6 +230,8 @@ namespace { clEnumValN(FloatABI::Hard, "hard", "Hard float ABI (uses FP registers)"), clEnumValEnd)); + + ExitOnError ExitOnErr; } //===----------------------------------------------------------------------===// @@ -257,7 +254,7 @@ public: ~LLIObjectCache() override {} void notifyObjectCompiled(const Module *M, MemoryBufferRef Obj) override { - const std::string ModuleID = M->getModuleIdentifier(); + const std::string &ModuleID = M->getModuleIdentifier(); std::string CacheName; if (!getCacheFilename(ModuleID, CacheName)) return; @@ -272,7 +269,7 @@ public: } std::unique_ptr<MemoryBuffer> getObject(const Module* M) override { - const std::string ModuleID = M->getModuleIdentifier(); + const std::string &ModuleID = M->getModuleIdentifier(); std::string CacheName; if (!getCacheFilename(ModuleID, CacheName)) return nullptr; @@ -312,26 +309,12 @@ private: } }; -static ExecutionEngine *EE = nullptr; -static LLIObjectCache *CacheManager = nullptr; - -static void do_shutdown() { - // Cygwin-1.5 invokes DLL's dtors before atexit handler. -#ifndef DO_NOTHING_ATEXIT - delete EE; - if (CacheManager) - delete CacheManager; - llvm_shutdown(); -#endif -} - // On Mingw and Cygwin, an external symbol named '__main' is called from the // generated 'main' function to allow static intialization. To avoid linking // problems with remote targets (because lli's remote target support does not // currently handle external linking) we add a secondary module which defines // an empty '__main' function. -static void addCygMingExtraModule(ExecutionEngine *EE, - LLVMContext &Context, +static void addCygMingExtraModule(ExecutionEngine &EE, LLVMContext &Context, StringRef TargetTripleStr) { IRBuilder<> Builder(Context); Triple TargetTriple(TargetTripleStr); @@ -361,7 +344,7 @@ static void addCygMingExtraModule(ExecutionEngine *EE, Builder.CreateRet(ReturnVal); // Add this new module to the ExecutionEngine. - EE->addModule(std::move(M)); + EE.addModule(std::move(M)); } CodeGenOpt::Level getOptLevel() { @@ -382,11 +365,13 @@ CodeGenOpt::Level getOptLevel() { // main Driver function // int main(int argc, char **argv, char * const *envp) { - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); - LLVMContext &Context = getGlobalContext(); - atexit(do_shutdown); // Call llvm_shutdown() on exit. + atexit(llvm_shutdown); // Call llvm_shutdown() on exit. + + if (argc > 1) + ExitOnErr.setBanner(std::string(argv[0]) + ": "); // If we have a native target, initialize it to ensure it is linked in and // usable by the JIT. @@ -401,6 +386,8 @@ int main(int argc, char **argv, char * const *envp) { if (DisableCoreFiles) sys::Process::PreventCoreFiles(); + LLVMContext Context; + // Load the bitcode... SMDiagnostic Err; std::unique_ptr<Module> Owner = parseIRFile(InputFile, Err, Context); @@ -433,7 +420,8 @@ int main(int argc, char **argv, char * const *envp) { builder.setMArch(MArch); builder.setMCPU(MCPU); builder.setMAttrs(MAttrs); - builder.setRelocationModel(RelocModel); + if (RelocModel.getNumOccurrences()) + builder.setRelocationModel(RelocModel); builder.setCodeModel(CMModel); builder.setErrorStr(&ErrorMsg); builder.setEngineKind(ForceInterpreter @@ -471,7 +459,7 @@ int main(int argc, char **argv, char * const *envp) { builder.setTargetOptions(Options); - EE = builder.create(); + std::unique_ptr<ExecutionEngine> EE(builder.create()); if (!EE) { if (!ErrorMsg.empty()) errs() << argv[0] << ": error creating EE: " << ErrorMsg << "\n"; @@ -480,9 +468,10 @@ int main(int argc, char **argv, char * const *envp) { exit(1); } + std::unique_ptr<LLIObjectCache> CacheManager; if (EnableCacheManager) { - CacheManager = new LLIObjectCache(ObjectCacheDir); - EE->setObjectCache(CacheManager); + CacheManager.reset(new LLIObjectCache(ObjectCacheDir)); + EE->setObjectCache(CacheManager.get()); } // Load any additional modules specified on the command line. @@ -501,9 +490,11 @@ int main(int argc, char **argv, char * const *envp) { } for (unsigned i = 0, e = ExtraObjects.size(); i != e; ++i) { - ErrorOr<object::OwningBinary<object::ObjectFile>> Obj = + Expected<object::OwningBinary<object::ObjectFile>> Obj = object::ObjectFile::createObjectFile(ExtraObjects[i]); if (!Obj) { + // TODO: Actually report errors helpfully. + consumeError(Obj.takeError()); Err.print(argv[0], errs()); return 1; } @@ -520,10 +511,14 @@ int main(int argc, char **argv, char * const *envp) { } std::unique_ptr<MemoryBuffer> &ArBuf = ArBufOrErr.get(); - ErrorOr<std::unique_ptr<object::Archive>> ArOrErr = + Expected<std::unique_ptr<object::Archive>> ArOrErr = object::Archive::create(ArBuf->getMemBufferRef()); - if (std::error_code EC = ArOrErr.getError()) { - errs() << EC.message(); + if (!ArOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(ArOrErr.takeError(), OS, ""); + OS.flush(); + errs() << Buf; return 1; } std::unique_ptr<object::Archive> &Ar = ArOrErr.get(); @@ -536,7 +531,7 @@ int main(int argc, char **argv, char * const *envp) { // If the target is Cygwin/MingW and we are generating remote code, we // need an extra module to help out with linking. if (RemoteMCJIT && Triple(Mod->getTargetTriple()).isOSCygMing()) { - addCygMingExtraModule(EE, Context, Mod->getTargetTriple()); + addCygMingExtraModule(*EE, Context, Mod->getTargetTriple()); } // The following functions have no effect if their respective profiling @@ -580,14 +575,14 @@ int main(int argc, char **argv, char * const *envp) { // Reset errno to zero on entry to main. errno = 0; - int Result; + int Result = -1; // Sanity check use of remote-jit: LLI currently only supports use of the // remote JIT on Unix platforms. if (RemoteMCJIT) { #ifndef LLVM_ON_UNIX errs() << "Warning: host does not support external remote targets.\n" - << " Defaulting to local execution execution\n"; + << " Defaulting to local execution\n"; return -1; #else if (ChildExecPath.empty()) { @@ -658,18 +653,11 @@ int main(int argc, char **argv, char * const *envp) { // Create a remote target client running over the channel. typedef orc::remote::OrcRemoteTargetClient<orc::remote::RPCChannel> MyRemote; - ErrorOr<MyRemote> R = MyRemote::Create(*C); - if (!R) { - errs() << "Could not create remote: " << R.getError().message() << "\n"; - exit(1); - } + MyRemote R = ExitOnErr(MyRemote::Create(*C)); // Create a remote memory manager. std::unique_ptr<MyRemote::RCMemoryManager> RemoteMM; - if (auto EC = R->createRemoteMemoryManager(RemoteMM)) { - errs() << "Could not create remote memory manager: " << EC.message() << "\n"; - exit(1); - } + ExitOnErr(R.createRemoteMemoryManager(RemoteMM)); // Forward MCJIT's memory manager calls to the remote memory manager. static_cast<ForwardingMemoryManager*>(RTDyldMM)->setMemMgr( @@ -678,15 +666,12 @@ int main(int argc, char **argv, char * const *envp) { // Forward MCJIT's symbol resolution calls to the remote. static_cast<ForwardingMemoryManager*>(RTDyldMM)->setResolver( orc::createLambdaResolver( + [](const std::string &Name) { return nullptr; }, [&](const std::string &Name) { - orc::TargetAddress Addr = 0; - if (auto EC = R->getSymbolAddress(Addr, Name)) { - errs() << "Failure during symbol lookup: " << EC.message() << "\n"; - exit(1); - } - return RuntimeDyld::SymbolInfo(Addr, JITSymbolFlags::Exported); - }, - [](const std::string &Name) { return nullptr; } + if (auto Addr = ExitOnErr(R.getSymbolAddress(Name))) + return RuntimeDyld::SymbolInfo(Addr, JITSymbolFlags::Exported); + return RuntimeDyld::SymbolInfo(nullptr); + } )); // Grab the target address of the JIT'd main function on the remote and call @@ -696,8 +681,7 @@ int main(int argc, char **argv, char * const *envp) { EE->finalizeObject(); DEBUG(dbgs() << "Executing '" << EntryFn->getName() << "' at 0x" << format("%llx", Entry) << "\n"); - if (auto EC = R->callIntVoid(Result, Entry)) - errs() << "ERROR: " << EC.message() << "\n"; + Result = ExitOnErr(R.callIntVoid(Entry)); // Like static constructors, the remote target MCJIT support doesn't handle // this yet. It could. FIXME. @@ -705,11 +689,10 @@ int main(int argc, char **argv, char * const *envp) { // Delete the EE - we need to tear it down *before* we terminate the session // with the remote, otherwise it'll crash when it tries to release resources // on a remote that has already been disconnected. - delete EE; - EE = nullptr; + EE.reset(); // Signal the remote target that we're done JITing. - R->terminateSession(); + ExitOnErr(R.terminateSession()); } return Result; diff --git a/tools/llvm-ar/Makefile b/tools/llvm-ar/Makefile deleted file mode 100644 index 824bb9473000d..0000000000000 --- a/tools/llvm-ar/Makefile +++ /dev/null @@ -1,21 +0,0 @@ -##===- tools/llvm-ar/Makefile ------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-ar -TOOLALIAS = llvm-ranlib -LINK_COMPONENTS := all-targets bitreader libdriver support object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common - -all-local:: - $(Verb) $(AliasTool) $(notdir $(ToolBuildPath)) $(ToolDir)/llvm-lib$(EXEEXT) diff --git a/tools/llvm-ar/llvm-ar.cpp b/tools/llvm-ar/llvm-ar.cpp index ef5fab68b9479..865152b6af831 100644 --- a/tools/llvm-ar/llvm-ar.cpp +++ b/tools/llvm-ar/llvm-ar.cpp @@ -19,6 +19,7 @@ #include "llvm/LibDriver/LibDriver.h" #include "llvm/Object/Archive.h" #include "llvm/Object/ArchiveWriter.h" +#include "llvm/Object/MachO.h" #include "llvm/Object/ObjectFile.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/Errc.h" @@ -64,12 +65,25 @@ static void failIfError(std::error_code EC, Twine Context = "") { fail(Context + ": " + EC.message()); } +static void failIfError(Error E, Twine Context = "") { + if (!E) + return; + + handleAllErrors(std::move(E), [&](const llvm::ErrorInfoBase &EIB) { + std::string ContextStr = Context.str(); + if (ContextStr == "") + fail(EIB.message()); + fail(Context + ": " + EIB.message()); + }); +} + // llvm-ar/llvm-ranlib remaining positional arguments. static cl::list<std::string> RestOfArgs(cl::Positional, cl::ZeroOrMore, cl::desc("[relpos] [count] <archive-file> [members]...")); static cl::opt<bool> MRI("M", cl::desc("")); +static cl::opt<std::string> Plugin("plugin", cl::desc("plugin (ignored for compatibility")); namespace { enum Format { Default, GNU, BSD }; @@ -77,7 +91,7 @@ enum Format { Default, GNU, BSD }; static cl::opt<Format> FormatOpt("format", cl::desc("Archive format to create"), - cl::values(clEnumValN(Default, "defalut", "default"), + cl::values(clEnumValN(Default, "default", "default"), clEnumValN(GNU, "gnu", "gnu"), clEnumValN(BSD, "bsd", "bsd"), clEnumValEnd)); @@ -343,12 +357,9 @@ static void doDisplayTable(StringRef Name, const object::Archive::Child &C) { static void doExtract(StringRef Name, const object::Archive::Child &C) { // Retain the original mode. sys::fs::perms Mode = C.getAccessMode(); - SmallString<128> Storage = Name; int FD; - failIfError( - sys::fs::openFileForWrite(Storage.c_str(), FD, sys::fs::F_None, Mode), - Storage.c_str()); + failIfError(sys::fs::openFileForWrite(Name, FD, sys::fs::F_None, Mode), Name); { raw_fd_ostream file(FD, false); @@ -394,35 +405,37 @@ static void performReadOperation(ArchiveOperation Operation, fail("extracting from a thin archive is not supported"); bool Filter = !Members.empty(); - for (auto &ChildOrErr : OldArchive->children()) { - failIfError(ChildOrErr.getError()); - const object::Archive::Child &C = *ChildOrErr; - - ErrorOr<StringRef> NameOrErr = C.getName(); - failIfError(NameOrErr.getError()); - StringRef Name = NameOrErr.get(); - - if (Filter) { - auto I = std::find(Members.begin(), Members.end(), Name); - if (I == Members.end()) - continue; - Members.erase(I); - } + { + Error Err; + for (auto &C : OldArchive->children(Err)) { + ErrorOr<StringRef> NameOrErr = C.getName(); + failIfError(NameOrErr.getError()); + StringRef Name = NameOrErr.get(); - switch (Operation) { - default: - llvm_unreachable("Not a read operation"); - case Print: - doPrint(Name, C); - break; - case DisplayTable: - doDisplayTable(Name, C); - break; - case Extract: - doExtract(Name, C); - break; + if (Filter) { + auto I = std::find(Members.begin(), Members.end(), Name); + if (I == Members.end()) + continue; + Members.erase(I); + } + + switch (Operation) { + default: + llvm_unreachable("Not a read operation"); + case Print: + doPrint(Name, C); + break; + case DisplayTable: + doDisplayTable(Name, C); + break; + case Extract: + doExtract(Name, C); + break; + } } + failIfError(std::move(Err)); } + if (Members.empty()) return; for (StringRef Name : Members) @@ -430,25 +443,28 @@ static void performReadOperation(ArchiveOperation Operation, std::exit(1); } -static void addMember(std::vector<NewArchiveIterator> &Members, +static void addMember(std::vector<NewArchiveMember> &Members, StringRef FileName, int Pos = -1) { - NewArchiveIterator NI(FileName); + Expected<NewArchiveMember> NMOrErr = + NewArchiveMember::getFile(FileName, Deterministic); + failIfError(NMOrErr.takeError(), FileName); if (Pos == -1) - Members.push_back(NI); + Members.push_back(std::move(*NMOrErr)); else - Members[Pos] = NI; + Members[Pos] = std::move(*NMOrErr); } -static void addMember(std::vector<NewArchiveIterator> &Members, - const object::Archive::Child &M, StringRef Name, - int Pos = -1) { +static void addMember(std::vector<NewArchiveMember> &Members, + const object::Archive::Child &M, int Pos = -1) { if (Thin && !M.getParent()->isThin()) fail("Cannot convert a regular archive to a thin one"); - NewArchiveIterator NI(M, Name); + Expected<NewArchiveMember> NMOrErr = + NewArchiveMember::getOldMember(M, Deterministic); + failIfError(NMOrErr.takeError()); if (Pos == -1) - Members.push_back(NI); + Members.push_back(std::move(*NMOrErr)); else - Members[Pos] = NI; + Members[Pos] = std::move(*NMOrErr); } enum InsertAction { @@ -509,17 +525,16 @@ static InsertAction computeInsertAction(ArchiveOperation Operation, // We have to walk this twice and computing it is not trivial, so creating an // explicit std::vector is actually fairly efficient. -static std::vector<NewArchiveIterator> +static std::vector<NewArchiveMember> computeNewArchiveMembers(ArchiveOperation Operation, object::Archive *OldArchive) { - std::vector<NewArchiveIterator> Ret; - std::vector<NewArchiveIterator> Moved; + std::vector<NewArchiveMember> Ret; + std::vector<NewArchiveMember> Moved; int InsertPos = -1; StringRef PosName = sys::path::filename(RelPos); if (OldArchive) { - for (auto &ChildOrErr : OldArchive->children()) { - failIfError(ChildOrErr.getError()); - auto &Child = ChildOrErr.get(); + Error Err; + for (auto &Child : OldArchive->children(Err)) { int Pos = Ret.size(); ErrorOr<StringRef> NameOrErr = Child.getName(); failIfError(NameOrErr.getError()); @@ -537,7 +552,7 @@ computeNewArchiveMembers(ArchiveOperation Operation, computeInsertAction(Operation, Child, Name, MemberI); switch (Action) { case IA_AddOldMember: - addMember(Ret, Child, Name); + addMember(Ret, Child); break; case IA_AddNewMeber: addMember(Ret, *MemberI); @@ -545,7 +560,7 @@ computeNewArchiveMembers(ArchiveOperation Operation, case IA_Delete: break; case IA_MoveOldMember: - addMember(Moved, Child, Name); + addMember(Moved, Child); break; case IA_MoveNewMember: addMember(Moved, *MemberI); @@ -554,6 +569,7 @@ computeNewArchiveMembers(ArchiveOperation Operation, if (MemberI != Members.end()) Members.erase(MemberI); } + failIfError(std::move(Err)); } if (Operation == Delete) @@ -566,10 +582,15 @@ computeNewArchiveMembers(ArchiveOperation Operation, InsertPos = Ret.size(); assert(unsigned(InsertPos) <= Ret.size()); - Ret.insert(Ret.begin() + InsertPos, Moved.begin(), Moved.end()); - - Ret.insert(Ret.begin() + InsertPos, Members.size(), NewArchiveIterator("")); int Pos = InsertPos; + for (auto &M : Moved) { + Ret.insert(Ret.begin() + Pos, std::move(M)); + ++Pos; + } + + for (unsigned I = 0; I != Members.size(); ++I) + Ret.insert(Ret.begin() + InsertPos, NewArchiveMember()); + Pos = InsertPos; for (auto &Member : Members) { addMember(Ret, Member, Pos); ++Pos; @@ -578,36 +599,61 @@ computeNewArchiveMembers(ArchiveOperation Operation, return Ret; } +static object::Archive::Kind getDefaultForHost() { + return Triple(sys::getProcessTriple()).isOSDarwin() ? object::Archive::K_BSD + : object::Archive::K_GNU; +} + +static object::Archive::Kind getKindFromMember(const NewArchiveMember &Member) { + Expected<std::unique_ptr<object::ObjectFile>> OptionalObject = + object::ObjectFile::createObjectFile(Member.Buf->getMemBufferRef()); + + if (OptionalObject) + return isa<object::MachOObjectFile>(**OptionalObject) + ? object::Archive::K_BSD + : object::Archive::K_GNU; + + // squelch the error in case we had a non-object file + consumeError(OptionalObject.takeError()); + return getDefaultForHost(); +} + static void -performWriteOperation(ArchiveOperation Operation, object::Archive *OldArchive, - std::vector<NewArchiveIterator> *NewMembersP) { +performWriteOperation(ArchiveOperation Operation, + object::Archive *OldArchive, + std::unique_ptr<MemoryBuffer> OldArchiveBuf, + std::vector<NewArchiveMember> *NewMembersP) { + std::vector<NewArchiveMember> NewMembers; + if (!NewMembersP) + NewMembers = computeNewArchiveMembers(Operation, OldArchive); + object::Archive::Kind Kind; switch (FormatOpt) { - case Default: { - Triple T(sys::getProcessTriple()); - if (T.isOSDarwin()) - Kind = object::Archive::K_BSD; - else + case Default: + if (Thin) Kind = object::Archive::K_GNU; + else if (OldArchive) + Kind = OldArchive->kind(); + else if (NewMembersP) + Kind = NewMembersP->size() ? getKindFromMember(NewMembersP->front()) + : getDefaultForHost(); + else + Kind = NewMembers.size() ? getKindFromMember(NewMembers.front()) + : getDefaultForHost(); break; - } case GNU: Kind = object::Archive::K_GNU; break; case BSD: + if (Thin) + fail("Only the gnu format has a thin mode"); Kind = object::Archive::K_BSD; break; } - if (NewMembersP) { - std::pair<StringRef, std::error_code> Result = writeArchive( - ArchiveName, *NewMembersP, Symtab, Kind, Deterministic, Thin); - failIfError(Result.second, Result.first); - return; - } - std::vector<NewArchiveIterator> NewMembers = - computeNewArchiveMembers(Operation, OldArchive); - auto Result = - writeArchive(ArchiveName, NewMembers, Symtab, Kind, Deterministic, Thin); + + std::pair<StringRef, std::error_code> Result = + writeArchive(ArchiveName, NewMembersP ? *NewMembersP : NewMembers, Symtab, + Kind, Deterministic, Thin, std::move(OldArchiveBuf)); failIfError(Result.second, Result.first); } @@ -621,12 +667,13 @@ static void createSymbolTable(object::Archive *OldArchive) { if (OldArchive->hasSymbolTable()) return; - performWriteOperation(CreateSymTab, OldArchive, nullptr); + performWriteOperation(CreateSymTab, OldArchive, nullptr, nullptr); } static void performOperation(ArchiveOperation Operation, object::Archive *OldArchive, - std::vector<NewArchiveIterator> *NewMembers) { + std::unique_ptr<MemoryBuffer> OldArchiveBuf, + std::vector<NewArchiveMember> *NewMembers) { switch (Operation) { case Print: case DisplayTable: @@ -638,7 +685,8 @@ static void performOperation(ArchiveOperation Operation, case Move: case QuickAppend: case ReplaceOrInsert: - performWriteOperation(Operation, OldArchive, NewMembers); + performWriteOperation(Operation, OldArchive, std::move(OldArchiveBuf), + NewMembers); return; case CreateSymTab: createSymbolTable(OldArchive); @@ -648,7 +696,7 @@ static void performOperation(ArchiveOperation Operation, } static int performOperation(ArchiveOperation Operation, - std::vector<NewArchiveIterator> *NewMembers) { + std::vector<NewArchiveMember> *NewMembers) { // Create or open the archive object. ErrorOr<std::unique_ptr<MemoryBuffer>> Buf = MemoryBuffer::getFile(ArchiveName, -1, false); @@ -657,10 +705,12 @@ static int performOperation(ArchiveOperation Operation, fail("error opening '" + ArchiveName + "': " + EC.message() + "!"); if (!EC) { - object::Archive Archive(Buf.get()->getMemBufferRef(), EC); + Error Err; + object::Archive Archive(Buf.get()->getMemBufferRef(), Err); + EC = errorToErrorCode(std::move(Err)); failIfError(EC, "error loading '" + ArchiveName + "': " + EC.message() + "!"); - performOperation(Operation, &Archive, NewMembers); + performOperation(Operation, &Archive, std::move(Buf.get()), NewMembers); return 0; } @@ -675,7 +725,7 @@ static int performOperation(ArchiveOperation Operation, } } - performOperation(Operation, nullptr, NewMembers); + performOperation(Operation, nullptr, nullptr, NewMembers); return 0; } @@ -686,7 +736,7 @@ static void runMRIScript() { failIfError(Buf.getError()); const MemoryBuffer &Ref = *Buf.get(); bool Saved = false; - std::vector<NewArchiveIterator> NewMembers; + std::vector<NewArchiveMember> NewMembers; std::vector<std::unique_ptr<MemoryBuffer>> ArchiveBuffers; std::vector<std::unique_ptr<object::Archive>> Archives; @@ -712,15 +762,15 @@ static void runMRIScript() { ArchiveBuffers.push_back(std::move(*BufOrErr)); auto LibOrErr = object::Archive::create(ArchiveBuffers.back()->getMemBufferRef()); - failIfError(LibOrErr.getError(), "Could not parse library"); + failIfError(errorToErrorCode(LibOrErr.takeError()), + "Could not parse library"); Archives.push_back(std::move(*LibOrErr)); object::Archive &Lib = *Archives.back(); - for (auto &MemberOrErr : Lib.children()) { - failIfError(MemberOrErr.getError()); - auto &Member = MemberOrErr.get(); - ErrorOr<StringRef> NameOrErr = Member.getName(); - failIfError(NameOrErr.getError()); - addMember(NewMembers, Member, *NameOrErr); + { + Error Err; + for (auto &Member : Lib.children(Err)) + addMember(NewMembers, Member); + failIfError(std::move(Err)); } break; } @@ -760,7 +810,7 @@ static int ar_main() { static int ranlib_main() { if (RestOfArgs.size() != 1) - fail(ToolName + "takes just one archive as argument"); + fail(ToolName + " takes just one archive as an argument"); ArchiveName = RestOfArgs[0]; return performOperation(CreateSymTab, nullptr); } @@ -768,7 +818,7 @@ static int ranlib_main() { int main(int argc, char **argv) { ToolName = argv[0]; // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. diff --git a/tools/llvm-as-fuzzer/llvm-as-fuzzer.cpp b/tools/llvm-as-fuzzer/llvm-as-fuzzer.cpp index b4024bcaa99ae..52deca6e9cd77 100644 --- a/tools/llvm-as-fuzzer/llvm-as-fuzzer.cpp +++ b/tools/llvm-as-fuzzer/llvm-as-fuzzer.cpp @@ -53,7 +53,7 @@ extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { // parsed is always null terminated. std::unique_ptr<MemoryBuffer> MemBuf = MemoryBuffer::getMemBufferCopy(Input); SMDiagnostic Err; - LLVMContext &Context = getGlobalContext(); + LLVMContext Context; std::unique_ptr<Module> M; if (setjmp(JmpBuf)) diff --git a/tools/llvm-as/Makefile b/tools/llvm-as/Makefile deleted file mode 100644 index dfd71b295a9a8..0000000000000 --- a/tools/llvm-as/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-as/Makefile ------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-as -LINK_COMPONENTS := asmparser bitwriter - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-as/llvm-as.cpp b/tools/llvm-as/llvm-as.cpp index d4e4d8d71070a..89a6ee5edfd89 100644 --- a/tools/llvm-as/llvm-as.cpp +++ b/tools/llvm-as/llvm-as.cpp @@ -15,9 +15,9 @@ // //===----------------------------------------------------------------------===// -#include "llvm/IR/LLVMContext.h" #include "llvm/AsmParser/Parser.h" #include "llvm/Bitcode/ReaderWriter.h" +#include "llvm/IR/LLVMContext.h" #include "llvm/IR/Module.h" #include "llvm/IR/Verifier.h" #include "llvm/Support/CommandLine.h" @@ -31,29 +31,28 @@ #include <memory> using namespace llvm; -static cl::opt<std::string> -InputFilename(cl::Positional, cl::desc("<input .llvm file>"), cl::init("-")); +static cl::opt<std::string> InputFilename(cl::Positional, + cl::desc("<input .llvm file>"), + cl::init("-")); -static cl::opt<std::string> -OutputFilename("o", cl::desc("Override output filename"), - cl::value_desc("filename")); +static cl::opt<std::string> OutputFilename("o", + cl::desc("Override output filename"), + cl::value_desc("filename")); -static cl::opt<bool> -Force("f", cl::desc("Enable binary output on terminals")); +static cl::opt<bool> Force("f", cl::desc("Enable binary output on terminals")); -static cl::opt<bool> -DisableOutput("disable-output", cl::desc("Disable output"), cl::init(false)); +static cl::opt<bool> DisableOutput("disable-output", cl::desc("Disable output"), + cl::init(false)); -static cl::opt<bool> -EmitFunctionSummary("function-summary", cl::desc("Emit function summary index"), - cl::init(false)); +static cl::opt<bool> EmitModuleHash("module-hash", cl::desc("Emit module hash"), + cl::init(false)); -static cl::opt<bool> -DumpAsm("d", cl::desc("Print assembly as parsed"), cl::Hidden); +static cl::opt<bool> DumpAsm("d", cl::desc("Print assembly as parsed"), + cl::Hidden); static cl::opt<bool> -DisableVerify("disable-verify", cl::Hidden, - cl::desc("Do not run verifier on input LLVM (dangerous!)")); + DisableVerify("disable-verify", cl::Hidden, + cl::desc("Do not run verifier on input LLVM (dangerous!)")); static cl::opt<bool> PreserveBitcodeUseListOrder( "preserve-bc-uselistorder", @@ -81,8 +80,8 @@ static void WriteOutputFile(const Module *M) { } if (Force || !CheckBitcodeOutputToConsole(Out->os(), true)) - WriteBitcodeToFile(M, Out->os(), PreserveBitcodeUseListOrder, - EmitFunctionSummary); + WriteBitcodeToFile(M, Out->os(), PreserveBitcodeUseListOrder, nullptr, + EmitModuleHash); // Declare success. Out->keep(); @@ -90,10 +89,10 @@ static void WriteOutputFile(const Module *M) { int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); - LLVMContext &Context = getGlobalContext(); - llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. + LLVMContext Context; + llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. cl::ParseCommandLineOptions(argc, argv, "llvm .ll -> .bc assembler\n"); // Parse the file now... @@ -115,7 +114,8 @@ int main(int argc, char **argv) { } } - if (DumpAsm) errs() << "Here's the assembly:\n" << *M.get(); + if (DumpAsm) + errs() << "Here's the assembly:\n" << *M.get(); if (!DisableOutput) WriteOutputFile(M.get()); diff --git a/tools/llvm-bcanalyzer/Makefile b/tools/llvm-bcanalyzer/Makefile deleted file mode 100644 index 2fc61dbd62a7d..0000000000000 --- a/tools/llvm-bcanalyzer/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-bcanalyzer/Makefile ----------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-bcanalyzer -LINK_COMPONENTS := bitreader - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-bcanalyzer/llvm-bcanalyzer.cpp b/tools/llvm-bcanalyzer/llvm-bcanalyzer.cpp index fe68689def60a..319e34d882dcc 100644 --- a/tools/llvm-bcanalyzer/llvm-bcanalyzer.cpp +++ b/tools/llvm-bcanalyzer/llvm-bcanalyzer.cpp @@ -27,8 +27,8 @@ // //===----------------------------------------------------------------------===// +#include "llvm/ADT/StringExtras.h" #include "llvm/Bitcode/BitstreamReader.h" -#include "llvm/ADT/Optional.h" #include "llvm/Bitcode/LLVMBitCodes.h" #include "llvm/Bitcode/ReaderWriter.h" #include "llvm/IR/Verifier.h" @@ -37,6 +37,7 @@ #include "llvm/Support/ManagedStatic.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/PrettyStackTrace.h" +#include "llvm/Support/SHA1.h" #include "llvm/Support/Signals.h" #include "llvm/Support/raw_ostream.h" #include <algorithm> @@ -103,23 +104,24 @@ static const char *GetBlockName(unsigned BlockID, if (CurStreamType != LLVMIRBitstream) return nullptr; switch (BlockID) { - default: return nullptr; - case bitc::MODULE_BLOCK_ID: return "MODULE_BLOCK"; - case bitc::PARAMATTR_BLOCK_ID: return "PARAMATTR_BLOCK"; - case bitc::PARAMATTR_GROUP_BLOCK_ID: return "PARAMATTR_GROUP_BLOCK_ID"; - case bitc::TYPE_BLOCK_ID_NEW: return "TYPE_BLOCK_ID"; - case bitc::CONSTANTS_BLOCK_ID: return "CONSTANTS_BLOCK"; - case bitc::FUNCTION_BLOCK_ID: return "FUNCTION_BLOCK"; + default: return nullptr; + case bitc::OPERAND_BUNDLE_TAGS_BLOCK_ID: return "OPERAND_BUNDLE_TAGS_BLOCK"; + case bitc::MODULE_BLOCK_ID: return "MODULE_BLOCK"; + case bitc::PARAMATTR_BLOCK_ID: return "PARAMATTR_BLOCK"; + case bitc::PARAMATTR_GROUP_BLOCK_ID: return "PARAMATTR_GROUP_BLOCK_ID"; + case bitc::TYPE_BLOCK_ID_NEW: return "TYPE_BLOCK_ID"; + case bitc::CONSTANTS_BLOCK_ID: return "CONSTANTS_BLOCK"; + case bitc::FUNCTION_BLOCK_ID: return "FUNCTION_BLOCK"; case bitc::IDENTIFICATION_BLOCK_ID: - return "IDENTIFICATION_BLOCK_ID"; - case bitc::VALUE_SYMTAB_BLOCK_ID: return "VALUE_SYMTAB"; - case bitc::METADATA_BLOCK_ID: return "METADATA_BLOCK"; - case bitc::METADATA_KIND_BLOCK_ID: return "METADATA_KIND_BLOCK"; - case bitc::METADATA_ATTACHMENT_ID: return "METADATA_ATTACHMENT_BLOCK"; - case bitc::USELIST_BLOCK_ID: return "USELIST_BLOCK_ID"; - case bitc::FUNCTION_SUMMARY_BLOCK_ID: - return "FUNCTION_SUMMARY_BLOCK"; - case bitc::MODULE_STRTAB_BLOCK_ID: return "MODULE_STRTAB_BLOCK"; + return "IDENTIFICATION_BLOCK_ID"; + case bitc::VALUE_SYMTAB_BLOCK_ID: return "VALUE_SYMTAB"; + case bitc::METADATA_BLOCK_ID: return "METADATA_BLOCK"; + case bitc::METADATA_KIND_BLOCK_ID: return "METADATA_KIND_BLOCK"; + case bitc::METADATA_ATTACHMENT_ID: return "METADATA_ATTACHMENT_BLOCK"; + case bitc::USELIST_BLOCK_ID: return "USELIST_BLOCK_ID"; + case bitc::GLOBALVAL_SUMMARY_BLOCK_ID: + return "GLOBALVAL_SUMMARY_BLOCK"; + case bitc::MODULE_STRTAB_BLOCK_ID: return "MODULE_STRTAB_BLOCK"; } } @@ -172,7 +174,9 @@ static const char *GetCodeName(unsigned CodeID, unsigned BlockID, STRINGIFY_CODE(MODULE_CODE, PURGEVALS) STRINGIFY_CODE(MODULE_CODE, GCNAME) STRINGIFY_CODE(MODULE_CODE, VSTOFFSET) - STRINGIFY_CODE(MODULE_CODE, METADATA_VALUES) + STRINGIFY_CODE(MODULE_CODE, METADATA_VALUES_UNUSED) + STRINGIFY_CODE(MODULE_CODE, SOURCE_FILENAME) + STRINGIFY_CODE(MODULE_CODE, HASH) } case bitc::IDENTIFICATION_BLOCK_ID: switch (CodeID) { @@ -187,6 +191,10 @@ static const char *GetCodeName(unsigned CodeID, unsigned BlockID, // FIXME: Should these be different? case bitc::PARAMATTR_CODE_ENTRY_OLD: return "ENTRY"; case bitc::PARAMATTR_CODE_ENTRY: return "ENTRY"; + } + case bitc::PARAMATTR_GROUP_BLOCK_ID: + switch (CodeID) { + default: return nullptr; case bitc::PARAMATTR_GRP_CODE_ENTRY: return "ENTRY"; } case bitc::TYPE_BLOCK_ID_NEW: @@ -272,6 +280,7 @@ static const char *GetCodeName(unsigned CodeID, unsigned BlockID, STRINGIFY_CODE(FUNC_CODE, INST_CALL) STRINGIFY_CODE(FUNC_CODE, DEBUG_LOC) STRINGIFY_CODE(FUNC_CODE, INST_GEP) + STRINGIFY_CODE(FUNC_CODE, OPERAND_BUNDLE) } case bitc::VALUE_SYMTAB_BLOCK_ID: switch (CodeID) { @@ -279,20 +288,29 @@ static const char *GetCodeName(unsigned CodeID, unsigned BlockID, STRINGIFY_CODE(VST_CODE, ENTRY) STRINGIFY_CODE(VST_CODE, BBENTRY) STRINGIFY_CODE(VST_CODE, FNENTRY) - STRINGIFY_CODE(VST_CODE, COMBINED_FNENTRY) + STRINGIFY_CODE(VST_CODE, COMBINED_ENTRY) } case bitc::MODULE_STRTAB_BLOCK_ID: switch (CodeID) { default: return nullptr; STRINGIFY_CODE(MST_CODE, ENTRY) + STRINGIFY_CODE(MST_CODE, HASH) } - case bitc::FUNCTION_SUMMARY_BLOCK_ID: + case bitc::GLOBALVAL_SUMMARY_BLOCK_ID: switch (CodeID) { default: return nullptr; - STRINGIFY_CODE(FS_CODE, PERMODULE_ENTRY) - STRINGIFY_CODE(FS_CODE, COMBINED_ENTRY) + STRINGIFY_CODE(FS, PERMODULE) + STRINGIFY_CODE(FS, PERMODULE_PROFILE) + STRINGIFY_CODE(FS, PERMODULE_GLOBALVAR_INIT_REFS) + STRINGIFY_CODE(FS, COMBINED) + STRINGIFY_CODE(FS, COMBINED_PROFILE) + STRINGIFY_CODE(FS, COMBINED_GLOBALVAR_INIT_REFS) + STRINGIFY_CODE(FS, ALIAS) + STRINGIFY_CODE(FS, COMBINED_ALIAS) + STRINGIFY_CODE(FS, COMBINED_ORIGINAL_NAME) + STRINGIFY_CODE(FS, VERSION) } case bitc::METADATA_ATTACHMENT_ID: switch(CodeID) { @@ -302,7 +320,8 @@ static const char *GetCodeName(unsigned CodeID, unsigned BlockID, case bitc::METADATA_BLOCK_ID: switch(CodeID) { default:return nullptr; - STRINGIFY_CODE(METADATA, STRING) + STRINGIFY_CODE(METADATA, STRING_OLD) + STRINGIFY_CODE(METADATA, STRINGS) STRINGIFY_CODE(METADATA, NAME) STRINGIFY_CODE(METADATA, KIND) // Older bitcode has it in a MODULE_BLOCK STRINGIFY_CODE(METADATA, NODE) @@ -346,6 +365,12 @@ static const char *GetCodeName(unsigned CodeID, unsigned BlockID, case bitc::USELIST_CODE_DEFAULT: return "USELIST_CODE_DEFAULT"; case bitc::USELIST_CODE_BB: return "USELIST_CODE_BB"; } + + case bitc::OPERAND_BUNDLE_TAGS_BLOCK_ID: + switch(CodeID) { + default: return nullptr; + case bitc::OPERAND_BUNDLE_TAG: return "OPERAND_BUNDLE_TAG"; + } } #undef STRINGIFY_CODE } @@ -394,6 +419,57 @@ static bool Error(const Twine &Err) { return true; } +static bool decodeMetadataStringsBlob(BitstreamReader &Reader, StringRef Indent, + ArrayRef<uint64_t> Record, + StringRef Blob) { + if (Blob.empty()) + return true; + + if (Record.size() != 2) + return true; + + unsigned NumStrings = Record[0]; + unsigned StringsOffset = Record[1]; + outs() << " num-strings = " << NumStrings << " {\n"; + + StringRef Lengths = Blob.slice(0, StringsOffset); + SimpleBitstreamCursor R(Reader); + R.jumpToPointer(Lengths.begin()); + + // Ensure that Blob doesn't get invalidated, even if this is reading from a + // StreamingMemoryObject with corrupt data. + R.setArtificialByteLimit(R.getCurrentByteNo() + StringsOffset); + + StringRef Strings = Blob.drop_front(StringsOffset); + do { + if (R.AtEndOfStream()) + return Error("bad length"); + + unsigned Size = R.ReadVBR(6); + if (Strings.size() < Size) + return Error("truncated chars"); + + outs() << Indent << " '"; + outs().write_escaped(Strings.slice(0, Size), /*hex=*/true); + outs() << "'\n"; + Strings = Strings.drop_front(Size); + } while (--NumStrings); + + outs() << Indent << " }"; + return false; +} + +static bool decodeBlob(unsigned Code, unsigned BlockID, BitstreamReader &Reader, + StringRef Indent, ArrayRef<uint64_t> Record, + StringRef Blob) { + if (BlockID != bitc::METADATA_BLOCK_ID) + return true; + if (Code != bitc::METADATA_STRINGS) + return true; + + return decodeMetadataStringsBlob(Reader, Indent, Record, Blob); +} + /// ParseBlock - Read a block, updating statistics, etc. static bool ParseBlock(BitstreamCursor &Stream, unsigned BlockID, unsigned IndentLevel, CurStreamTypeType CurStreamType) { @@ -406,21 +482,24 @@ static bool ParseBlock(BitstreamCursor &Stream, unsigned BlockID, BlockStats.NumInstances++; // BLOCKINFO is a special part of the stream. + bool DumpRecords = Dump; if (BlockID == bitc::BLOCKINFO_BLOCK_ID) { if (Dump) outs() << Indent << "<BLOCKINFO_BLOCK/>\n"; - if (Stream.ReadBlockInfoBlock()) + if (BitstreamCursor(Stream).ReadBlockInfoBlock()) return Error("Malformed BlockInfoBlock"); - uint64_t BlockBitEnd = Stream.GetCurrentBitNo(); - BlockStats.NumBits += BlockBitEnd-BlockBitStart; - return false; + // It's not really interesting to dump the contents of the blockinfo block. + DumpRecords = false; } unsigned NumWords = 0; if (Stream.EnterSubBlock(BlockID, &NumWords)) return Error("Malformed block record"); + // Keep it for later, when we see a MODULE_HASH record + uint64_t BlockEntryPos = Stream.getCurrentByteNo(); + const char *BlockName = nullptr; - if (Dump) { + if (DumpRecords) { outs() << Indent << "<"; if ((BlockName = GetBlockName(BlockID, *Stream.getBitStreamReader(), CurStreamType))) @@ -453,7 +532,7 @@ static bool ParseBlock(BitstreamCursor &Stream, unsigned BlockID, case BitstreamEntry::EndBlock: { uint64_t BlockBitEnd = Stream.GetCurrentBitNo(); BlockStats.NumBits += BlockBitEnd-BlockBitStart; - if (Dump) { + if (DumpRecords) { outs() << Indent << "</"; if (BlockName) outs() << BlockName << ">\n"; @@ -490,6 +569,7 @@ static bool ParseBlock(BitstreamCursor &Stream, unsigned BlockID, ++BlockStats.NumRecords; StringRef Blob; + unsigned CurrentRecordPos = Stream.getCurrentByteNo(); unsigned Code = Stream.readRecord(Entry.ID, Record, &Blob); // Increment the # occurrences of this code. @@ -503,7 +583,7 @@ static bool ParseBlock(BitstreamCursor &Stream, unsigned BlockID, ++BlockStats.NumAbbreviatedRecords; } - if (Dump) { + if (DumpRecords) { outs() << Indent << " <"; if (const char *CodeName = GetCodeName(Code, BlockID, *Stream.getBitStreamReader(), @@ -524,6 +604,37 @@ static bool ParseBlock(BitstreamCursor &Stream, unsigned BlockID, for (unsigned i = 0, e = Record.size(); i != e; ++i) outs() << " op" << i << "=" << (int64_t)Record[i]; + // If we found a module hash, let's verify that it matches! + if (BlockID == bitc::MODULE_BLOCK_ID && Code == bitc::MODULE_CODE_HASH) { + if (Record.size() != 5) + outs() << " (invalid)"; + else { + // Recompute the hash and compare it to the one in the bitcode + SHA1 Hasher; + StringRef Hash; + { + int BlockSize = CurrentRecordPos - BlockEntryPos; + auto Ptr = Stream.getPointerToByte(BlockEntryPos, BlockSize); + Hasher.update(ArrayRef<uint8_t>(Ptr, BlockSize)); + Hash = Hasher.result(); + } + SmallString<20> RecordedHash; + RecordedHash.resize(20); + int Pos = 0; + for (auto &Val : Record) { + assert(!(Val >> 32) && "Unexpected high bits set"); + RecordedHash[Pos++] = (Val >> 24) & 0xFF; + RecordedHash[Pos++] = (Val >> 16) & 0xFF; + RecordedHash[Pos++] = (Val >> 8) & 0xFF; + RecordedHash[Pos++] = (Val >> 0) & 0xFF; + } + if (Hash == RecordedHash) + outs() << " (match)"; + else + outs() << " (!mismatch!)"; + } + } + outs() << "/>"; if (Abbv) { @@ -547,7 +658,8 @@ static bool ParseBlock(BitstreamCursor &Stream, unsigned BlockID, } } - if (Blob.data()) { + if (Blob.data() && decodeBlob(Code, BlockID, *Stream.getBitStreamReader(), + Indent, Record, Blob)) { outs() << " blob data = "; if (ShowBinaryBlobs) { outs() << "'"; @@ -600,9 +712,28 @@ static bool openBitcodeFile(StringRef Path, // If we have a wrapper header, parse it and ignore the non-bc file contents. // The magic number is 0x0B17C0DE stored in little endian. - if (isBitcodeWrapper(BufPtr, EndBufPtr)) + if (isBitcodeWrapper(BufPtr, EndBufPtr)) { + if (MemBuf->getBufferSize() < BWH_HeaderSize) + return Error("Invalid bitcode wrapper header"); + + if (Dump) { + unsigned Magic = support::endian::read32le(&BufPtr[BWH_MagicField]); + unsigned Version = support::endian::read32le(&BufPtr[BWH_VersionField]); + unsigned Offset = support::endian::read32le(&BufPtr[BWH_OffsetField]); + unsigned Size = support::endian::read32le(&BufPtr[BWH_SizeField]); + unsigned CPUType = support::endian::read32le(&BufPtr[BWH_CPUTypeField]); + + outs() << "<BITCODE_WRAPPER_HEADER" + << " Magic=" << format_hex(Magic, 10) + << " Version=" << format_hex(Version, 10) + << " Offset=" << format_hex(Offset, 10) + << " Size=" << format_hex(Size, 10) + << " CPUType=" << format_hex(CPUType, 10) << "/>\n"; + } + if (SkipBitcodeWrapperHeader(BufPtr, EndBufPtr, true)) return Error("Invalid bitcode wrapper header"); + } StreamFile = BitstreamReader(BufPtr, EndBufPtr); Stream = BitstreamCursor(StreamFile); @@ -745,7 +876,7 @@ static int AnalyzeBitcode() { std::reverse(FreqPairs.begin(), FreqPairs.end()); outs() << "\tRecord Histogram:\n"; - outs() << "\t\t Count # Bits %% Abv Record Kind\n"; + outs() << "\t\t Count # Bits b/Rec % Abv Record Kind\n"; for (unsigned i = 0, e = FreqPairs.size(); i != e; ++i) { const PerRecordStats &RecStats = Stats.CodeFreq[FreqPairs[i].second]; @@ -753,13 +884,20 @@ static int AnalyzeBitcode() { RecStats.NumInstances, (unsigned long)RecStats.TotalBits); + if (RecStats.NumInstances > 1) + outs() << format(" %9.1f", + (double)RecStats.TotalBits/RecStats.NumInstances); + else + outs() << " "; + if (RecStats.NumAbbrev) outs() << - format("%7.2f ", + format(" %7.2f", (double)RecStats.NumAbbrev/RecStats.NumInstances*100); else - outs() << " "; + outs() << " "; + outs() << " "; if (const char *CodeName = GetCodeName(FreqPairs[i].second, I->first, StreamFile, CurStreamType)) @@ -777,7 +915,7 @@ static int AnalyzeBitcode() { int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. cl::ParseCommandLineOptions(argc, argv, "llvm-bcanalyzer file analyzer\n"); diff --git a/tools/llvm-c-test/CMakeLists.txt b/tools/llvm-c-test/CMakeLists.txt index f22dffb30e8a4..858f2b07214a1 100644 --- a/tools/llvm-c-test/CMakeLists.txt +++ b/tools/llvm-c-test/CMakeLists.txt @@ -4,6 +4,7 @@ set(LLVM_LINK_COMPONENTS Core MCDisassembler Object + Support Target ) @@ -36,7 +37,9 @@ endif () add_llvm_tool(llvm-c-test calc.c + diagnostic.c disassemble.c + echo.cpp helpers.c include-all.c main.c diff --git a/tools/llvm-c-test/Makefile b/tools/llvm-c-test/Makefile deleted file mode 100644 index 08be7c3619fb2..0000000000000 --- a/tools/llvm-c-test/Makefile +++ /dev/null @@ -1,29 +0,0 @@ -##===- tools/llvm-c-test -----------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL = ../.. -TOOLNAME = llvm-c-test - -TOOL_NO_EXPORTS = 1 -NO_INSTALL = 1 - - -# If there is no shared lib, link all components... -ifneq ($(ENABLE_SHARED),1) -LINK_COMPONENTS = all -endif - -include $(LEVEL)/Makefile.common - -CFLAGS += -std=c99 -Wall -Wstrict-prototypes - -# ...but if it is built - use it -ifeq ($(ENABLE_SHARED),1) -LIBS = -lLLVM-$(LLVMVersion) -endif diff --git a/tools/llvm-c-test/calc.c b/tools/llvm-c-test/calc.c index 3119ccf8e9aa7..4c273cbf70a26 100644 --- a/tools/llvm-c-test/calc.c +++ b/tools/llvm-c-test/calc.c @@ -14,7 +14,6 @@ \*===----------------------------------------------------------------------===*/ #include "llvm-c-test.h" -#include "llvm-c/Core.h" #include <stdio.h> #include <stdlib.h> #include <string.h> @@ -140,9 +139,9 @@ static void handle_line(char **tokens, int ntokens) { LLVMDisposeModule(M); } -int calc(void) { +int llvm_calc(void) { - tokenize_stdin(handle_line); + llvm_tokenize_stdin(handle_line); return 0; } diff --git a/tools/llvm-c-test/diagnostic.c b/tools/llvm-c-test/diagnostic.c new file mode 100644 index 0000000000000..16d5174732360 --- /dev/null +++ b/tools/llvm-c-test/diagnostic.c @@ -0,0 +1,89 @@ +//===-- diagnostic.cpp - tool for testing libLLVM and llvm-c API ----------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements the --test-diagnostic-handler command in llvm-c-test. +// +// This command uses the C API to read a module with a custom diagnostic +// handler set to test the diagnostic handler functionality. +// +//===----------------------------------------------------------------------===// + +#include "llvm-c-test.h" +#include "llvm-c/BitReader.h" +#include "llvm-c/Core.h" + +#include <stdio.h> + +static void diagnosticHandler(LLVMDiagnosticInfoRef DI, void *C) { + fprintf(stderr, "Executing diagnostic handler\n"); + + fprintf(stderr, "Diagnostic severity is of type "); + switch (LLVMGetDiagInfoSeverity(DI)) { + case LLVMDSError: + fprintf(stderr, "error"); + break; + case LLVMDSWarning: + fprintf(stderr, "warning"); + break; + case LLVMDSRemark: + fprintf(stderr, "remark"); + break; + case LLVMDSNote: + fprintf(stderr, "note"); + break; + } + fprintf(stderr, "\n"); + + (*(int *)C) = 1; +} + +static int handlerCalled = 0; + +int llvm_test_diagnostic_handler(void) { + LLVMContextRef C = LLVMGetGlobalContext(); + LLVMContextSetDiagnosticHandler(C, diagnosticHandler, &handlerCalled); + + if (LLVMContextGetDiagnosticHandler(C) != diagnosticHandler) { + fprintf(stderr, "LLVMContext{Set,Get}DiagnosticHandler failed\n"); + return 1; + } + + int *DC = (int *)LLVMContextGetDiagnosticContext(C); + if (DC != &handlerCalled || *DC) { + fprintf(stderr, "LLVMContextGetDiagnosticContext failed\n"); + return 1; + } + + LLVMMemoryBufferRef MB; + char *msg = NULL; + if (LLVMCreateMemoryBufferWithSTDIN(&MB, &msg)) { + fprintf(stderr, "Error reading file: %s\n", msg); + LLVMDisposeMessage(msg); + return 1; + } + + + LLVMModuleRef M; + int Ret = LLVMGetBitcodeModule2(MB, &M); + if (Ret) { + // We do not return if the bitcode was invalid, as we want to test whether + // the diagnostic handler was executed. + fprintf(stderr, "Error parsing bitcode: %s\n", msg); + } + + LLVMDisposeMemoryBuffer(MB); + + if (handlerCalled) { + fprintf(stderr, "Diagnostic handler was called while loading module\n"); + } else { + fprintf(stderr, "Diagnostic handler was not called while loading module\n"); + } + + return 0; +} diff --git a/tools/llvm-c-test/disassemble.c b/tools/llvm-c-test/disassemble.c index 05a9218d014a6..7d6f1783aed9e 100644 --- a/tools/llvm-c-test/disassemble.c +++ b/tools/llvm-c-test/disassemble.c @@ -83,12 +83,12 @@ static void handle_line(char **tokens, int ntokens) { do_disassemble(triple, features, disbuf, disbuflen); } -int disassemble(void) { +int llvm_disassemble(void) { LLVMInitializeAllTargetInfos(); LLVMInitializeAllTargetMCs(); LLVMInitializeAllDisassemblers(); - tokenize_stdin(handle_line); + llvm_tokenize_stdin(handle_line); return 0; } diff --git a/tools/llvm-c-test/echo.cpp b/tools/llvm-c-test/echo.cpp new file mode 100644 index 0000000000000..72ff138c74e32 --- /dev/null +++ b/tools/llvm-c-test/echo.cpp @@ -0,0 +1,958 @@ +//===-- echo.cpp - tool for testing libLLVM and llvm-c API ----------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements the --echo command in llvm-c-test. +// +// This command uses the C API to read a module and output an exact copy of it +// as output. It is used to check that the resulting module matches the input +// to validate that the C API can read and write modules properly. +// +//===----------------------------------------------------------------------===// + +#include "llvm-c-test.h" +#include "llvm-c/Target.h" +#include "llvm/ADT/DenseMap.h" +#include "llvm/Support/ErrorHandling.h" + +#include <stdio.h> +#include <stdlib.h> + +using namespace llvm; + +// Provide DenseMapInfo for C API opaque types. +template<typename T> +struct CAPIDenseMap {}; + +// The default DenseMapInfo require to know about pointer alignement. +// Because the C API uses opaques pointer types, their alignement is unknown. +// As a result, we need to roll out our own implementation. +template<typename T> +struct CAPIDenseMap<T*> { + struct CAPIDenseMapInfo { + static inline T* getEmptyKey() { + uintptr_t Val = static_cast<uintptr_t>(-1); + return reinterpret_cast<T*>(Val); + } + static inline T* getTombstoneKey() { + uintptr_t Val = static_cast<uintptr_t>(-2); + return reinterpret_cast<T*>(Val); + } + static unsigned getHashValue(const T *PtrVal) { + return hash_value(PtrVal); + } + static bool isEqual(const T *LHS, const T *RHS) { return LHS == RHS; } + }; + + typedef DenseMap<T*, T*, CAPIDenseMapInfo> Map; +}; + +typedef CAPIDenseMap<LLVMValueRef>::Map ValueMap; +typedef CAPIDenseMap<LLVMBasicBlockRef>::Map BasicBlockMap; + +struct TypeCloner { + LLVMModuleRef M; + LLVMContextRef Ctx; + + TypeCloner(LLVMModuleRef M): M(M), Ctx(LLVMGetModuleContext(M)) {} + + LLVMTypeRef Clone(LLVMValueRef Src) { + return Clone(LLVMTypeOf(Src)); + } + + LLVMTypeRef Clone(LLVMTypeRef Src) { + LLVMTypeKind Kind = LLVMGetTypeKind(Src); + switch (Kind) { + case LLVMVoidTypeKind: + return LLVMVoidTypeInContext(Ctx); + case LLVMHalfTypeKind: + return LLVMHalfTypeInContext(Ctx); + case LLVMFloatTypeKind: + return LLVMFloatTypeInContext(Ctx); + case LLVMDoubleTypeKind: + return LLVMDoubleTypeInContext(Ctx); + case LLVMX86_FP80TypeKind: + return LLVMX86FP80TypeInContext(Ctx); + case LLVMFP128TypeKind: + return LLVMFP128TypeInContext(Ctx); + case LLVMPPC_FP128TypeKind: + return LLVMPPCFP128TypeInContext(Ctx); + case LLVMLabelTypeKind: + return LLVMLabelTypeInContext(Ctx); + case LLVMIntegerTypeKind: + return LLVMIntTypeInContext(Ctx, LLVMGetIntTypeWidth(Src)); + case LLVMFunctionTypeKind: { + unsigned ParamCount = LLVMCountParamTypes(Src); + LLVMTypeRef* Params = nullptr; + if (ParamCount > 0) { + Params = (LLVMTypeRef*) malloc(ParamCount * sizeof(LLVMTypeRef)); + LLVMGetParamTypes(Src, Params); + for (unsigned i = 0; i < ParamCount; i++) + Params[i] = Clone(Params[i]); + } + + LLVMTypeRef FunTy = LLVMFunctionType(Clone(LLVMGetReturnType(Src)), + Params, ParamCount, + LLVMIsFunctionVarArg(Src)); + if (ParamCount > 0) + free(Params); + return FunTy; + } + case LLVMStructTypeKind: { + LLVMTypeRef S = nullptr; + const char *Name = LLVMGetStructName(Src); + if (Name) { + S = LLVMGetTypeByName(M, Name); + if (S) + return S; + S = LLVMStructCreateNamed(Ctx, Name); + if (LLVMIsOpaqueStruct(Src)) + return S; + } + + unsigned EltCount = LLVMCountStructElementTypes(Src); + SmallVector<LLVMTypeRef, 8> Elts; + for (unsigned i = 0; i < EltCount; i++) + Elts.push_back(Clone(LLVMStructGetTypeAtIndex(Src, i))); + if (Name) + LLVMStructSetBody(S, Elts.data(), EltCount, LLVMIsPackedStruct(Src)); + else + S = LLVMStructTypeInContext(Ctx, Elts.data(), EltCount, + LLVMIsPackedStruct(Src)); + return S; + } + case LLVMArrayTypeKind: + return LLVMArrayType( + Clone(LLVMGetElementType(Src)), + LLVMGetArrayLength(Src) + ); + case LLVMPointerTypeKind: + return LLVMPointerType( + Clone(LLVMGetElementType(Src)), + LLVMGetPointerAddressSpace(Src) + ); + case LLVMVectorTypeKind: + return LLVMVectorType( + Clone(LLVMGetElementType(Src)), + LLVMGetVectorSize(Src) + ); + case LLVMMetadataTypeKind: + break; + case LLVMX86_MMXTypeKind: + return LLVMX86MMXTypeInContext(Ctx); + default: + break; + } + + fprintf(stderr, "%d is not a supported typekind\n", Kind); + exit(-1); + } +}; + +static ValueMap clone_params(LLVMValueRef Src, LLVMValueRef Dst) { + unsigned Count = LLVMCountParams(Src); + if (Count != LLVMCountParams(Dst)) + report_fatal_error("Parameter count mismatch"); + + ValueMap VMap; + if (Count == 0) + return VMap; + + LLVMValueRef SrcFirst = LLVMGetFirstParam(Src); + LLVMValueRef DstFirst = LLVMGetFirstParam(Dst); + LLVMValueRef SrcLast = LLVMGetLastParam(Src); + LLVMValueRef DstLast = LLVMGetLastParam(Dst); + + LLVMValueRef SrcCur = SrcFirst; + LLVMValueRef DstCur = DstFirst; + LLVMValueRef SrcNext = nullptr; + LLVMValueRef DstNext = nullptr; + while (true) { + const char *Name = LLVMGetValueName(SrcCur); + LLVMSetValueName(DstCur, Name); + + VMap[SrcCur] = DstCur; + + Count--; + SrcNext = LLVMGetNextParam(SrcCur); + DstNext = LLVMGetNextParam(DstCur); + if (SrcNext == nullptr && DstNext == nullptr) { + if (SrcCur != SrcLast) + report_fatal_error("SrcLast param does not match End"); + if (DstCur != DstLast) + report_fatal_error("DstLast param does not match End"); + break; + } + + if (SrcNext == nullptr) + report_fatal_error("SrcNext was unexpectedly null"); + if (DstNext == nullptr) + report_fatal_error("DstNext was unexpectedly null"); + + LLVMValueRef SrcPrev = LLVMGetPreviousParam(SrcNext); + if (SrcPrev != SrcCur) + report_fatal_error("SrcNext.Previous param is not Current"); + + LLVMValueRef DstPrev = LLVMGetPreviousParam(DstNext); + if (DstPrev != DstCur) + report_fatal_error("DstNext.Previous param is not Current"); + + SrcCur = SrcNext; + DstCur = DstNext; + } + + if (Count != 0) + report_fatal_error("Parameter count does not match iteration"); + + return VMap; +} + +static void check_value_kind(LLVMValueRef V, LLVMValueKind K) { + if (LLVMGetValueKind(V) != K) + report_fatal_error("LLVMGetValueKind returned incorrect type"); +} + +static LLVMValueRef clone_constant_impl(LLVMValueRef Cst, LLVMModuleRef M); + +static LLVMValueRef clone_constant(LLVMValueRef Cst, LLVMModuleRef M) { + LLVMValueRef Ret = clone_constant_impl(Cst, M); + check_value_kind(Ret, LLVMGetValueKind(Cst)); + return Ret; +} + +static LLVMValueRef clone_constant_impl(LLVMValueRef Cst, LLVMModuleRef M) { + if (!LLVMIsAConstant(Cst)) + report_fatal_error("Expected a constant"); + + // Maybe it is a symbol + if (LLVMIsAGlobalValue(Cst)) { + const char *Name = LLVMGetValueName(Cst); + + // Try function + if (LLVMIsAFunction(Cst)) { + check_value_kind(Cst, LLVMFunctionValueKind); + LLVMValueRef Dst = LLVMGetNamedFunction(M, Name); + if (Dst) + return Dst; + report_fatal_error("Could not find function"); + } + + // Try global variable + if (LLVMIsAGlobalVariable(Cst)) { + check_value_kind(Cst, LLVMGlobalVariableValueKind); + LLVMValueRef Dst = LLVMGetNamedGlobal(M, Name); + if (Dst) + return Dst; + report_fatal_error("Could not find function"); + } + + fprintf(stderr, "Could not find @%s\n", Name); + exit(-1); + } + + // Try integer literal + if (LLVMIsAConstantInt(Cst)) { + check_value_kind(Cst, LLVMConstantIntValueKind); + return LLVMConstInt(TypeCloner(M).Clone(Cst), + LLVMConstIntGetZExtValue(Cst), false); + } + + // Try zeroinitializer + if (LLVMIsAConstantAggregateZero(Cst)) { + check_value_kind(Cst, LLVMConstantAggregateZeroValueKind); + return LLVMConstNull(TypeCloner(M).Clone(Cst)); + } + + // Try constant array + if (LLVMIsAConstantArray(Cst)) { + check_value_kind(Cst, LLVMConstantArrayValueKind); + LLVMTypeRef Ty = TypeCloner(M).Clone(Cst); + unsigned EltCount = LLVMGetArrayLength(Ty); + SmallVector<LLVMValueRef, 8> Elts; + for (unsigned i = 0; i < EltCount; i++) + Elts.push_back(clone_constant(LLVMGetOperand(Cst, i), M)); + return LLVMConstArray(LLVMGetElementType(Ty), Elts.data(), EltCount); + } + + // Try contant data array + if (LLVMIsAConstantDataArray(Cst)) { + check_value_kind(Cst, LLVMConstantDataArrayValueKind); + LLVMTypeRef Ty = TypeCloner(M).Clone(Cst); + unsigned EltCount = LLVMGetArrayLength(Ty); + SmallVector<LLVMValueRef, 8> Elts; + for (unsigned i = 0; i < EltCount; i++) + Elts.push_back(clone_constant(LLVMGetElementAsConstant(Cst, i), M)); + return LLVMConstArray(LLVMGetElementType(Ty), Elts.data(), EltCount); + } + + // Try constant struct + if (LLVMIsAConstantStruct(Cst)) { + check_value_kind(Cst, LLVMConstantStructValueKind); + LLVMTypeRef Ty = TypeCloner(M).Clone(Cst); + unsigned EltCount = LLVMCountStructElementTypes(Ty); + SmallVector<LLVMValueRef, 8> Elts; + for (unsigned i = 0; i < EltCount; i++) + Elts.push_back(clone_constant(LLVMGetOperand(Cst, i), M)); + if (LLVMGetStructName(Ty)) + return LLVMConstNamedStruct(Ty, Elts.data(), EltCount); + return LLVMConstStructInContext(LLVMGetModuleContext(M), Elts.data(), + EltCount, LLVMIsPackedStruct(Ty)); + } + + // Try undef + if (LLVMIsUndef(Cst)) { + check_value_kind(Cst, LLVMUndefValueValueKind); + return LLVMGetUndef(TypeCloner(M).Clone(Cst)); + } + + // Try float literal + if (LLVMIsAConstantFP(Cst)) { + check_value_kind(Cst, LLVMConstantFPValueKind); + report_fatal_error("ConstantFP is not supported"); + } + + // This kind of constant is not supported + if (!LLVMIsAConstantExpr(Cst)) + report_fatal_error("Expected a constant expression"); + + // At this point, it must be a constant expression + check_value_kind(Cst, LLVMConstantExprValueKind); + + LLVMOpcode Op = LLVMGetConstOpcode(Cst); + switch(Op) { + case LLVMBitCast: + return LLVMConstBitCast(clone_constant(LLVMGetOperand(Cst, 0), M), + TypeCloner(M).Clone(Cst)); + default: + fprintf(stderr, "%d is not a supported opcode\n", Op); + exit(-1); + } +} + +struct FunCloner { + LLVMValueRef Fun; + LLVMModuleRef M; + + ValueMap VMap; + BasicBlockMap BBMap; + + FunCloner(LLVMValueRef Src, LLVMValueRef Dst): Fun(Dst), + M(LLVMGetGlobalParent(Fun)), VMap(clone_params(Src, Dst)) {} + + LLVMTypeRef CloneType(LLVMTypeRef Src) { + return TypeCloner(M).Clone(Src); + } + + LLVMTypeRef CloneType(LLVMValueRef Src) { + return TypeCloner(M).Clone(Src); + } + + // Try to clone everything in the llvm::Value hierarchy. + LLVMValueRef CloneValue(LLVMValueRef Src) { + // First, the value may be constant. + if (LLVMIsAConstant(Src)) + return clone_constant(Src, M); + + // Function argument should always be in the map already. + auto i = VMap.find(Src); + if (i != VMap.end()) + return i->second; + + if (!LLVMIsAInstruction(Src)) + report_fatal_error("Expected an instruction"); + + auto Ctx = LLVMGetModuleContext(M); + auto Builder = LLVMCreateBuilderInContext(Ctx); + auto BB = DeclareBB(LLVMGetInstructionParent(Src)); + LLVMPositionBuilderAtEnd(Builder, BB); + auto Dst = CloneInstruction(Src, Builder); + LLVMDisposeBuilder(Builder); + return Dst; + } + + void CloneAttrs(LLVMValueRef Src, LLVMValueRef Dst) { + auto Ctx = LLVMGetModuleContext(M); + int ArgCount = LLVMGetNumArgOperands(Src); + for (int i = LLVMAttributeReturnIndex; i <= ArgCount; i++) { + for (unsigned k = 0, e = LLVMGetLastEnumAttributeKind(); k < e; ++k) { + if (auto SrcA = LLVMGetCallSiteEnumAttribute(Src, i, k)) { + auto Val = LLVMGetEnumAttributeValue(SrcA); + auto A = LLVMCreateEnumAttribute(Ctx, k, Val); + LLVMAddCallSiteAttribute(Dst, i, A); + } + } + } + } + + LLVMValueRef CloneInstruction(LLVMValueRef Src, LLVMBuilderRef Builder) { + check_value_kind(Src, LLVMInstructionValueKind); + if (!LLVMIsAInstruction(Src)) + report_fatal_error("Expected an instruction"); + + const char *Name = LLVMGetValueName(Src); + + // Check if this is something we already computed. + { + auto i = VMap.find(Src); + if (i != VMap.end()) { + // If we have a hit, it means we already generated the instruction + // as a dependancy to somethign else. We need to make sure + // it is ordered properly. + auto I = i->second; + LLVMInstructionRemoveFromParent(I); + LLVMInsertIntoBuilderWithName(Builder, I, Name); + return I; + } + } + + // We tried everything, it must be an instruction + // that hasn't been generated already. + LLVMValueRef Dst = nullptr; + + LLVMOpcode Op = LLVMGetInstructionOpcode(Src); + switch(Op) { + case LLVMRet: { + int OpCount = LLVMGetNumOperands(Src); + if (OpCount == 0) + Dst = LLVMBuildRetVoid(Builder); + else + Dst = LLVMBuildRet(Builder, CloneValue(LLVMGetOperand(Src, 0))); + break; + } + case LLVMBr: { + if (!LLVMIsConditional(Src)) { + LLVMValueRef SrcOp = LLVMGetOperand(Src, 0); + LLVMBasicBlockRef SrcBB = LLVMValueAsBasicBlock(SrcOp); + Dst = LLVMBuildBr(Builder, DeclareBB(SrcBB)); + break; + } + + LLVMValueRef Cond = LLVMGetCondition(Src); + LLVMValueRef Else = LLVMGetOperand(Src, 1); + LLVMBasicBlockRef ElseBB = DeclareBB(LLVMValueAsBasicBlock(Else)); + LLVMValueRef Then = LLVMGetOperand(Src, 2); + LLVMBasicBlockRef ThenBB = DeclareBB(LLVMValueAsBasicBlock(Then)); + Dst = LLVMBuildCondBr(Builder, Cond, ThenBB, ElseBB); + break; + } + case LLVMSwitch: + case LLVMIndirectBr: + break; + case LLVMInvoke: { + SmallVector<LLVMValueRef, 8> Args; + int ArgCount = LLVMGetNumArgOperands(Src); + for (int i = 0; i < ArgCount; i++) + Args.push_back(CloneValue(LLVMGetOperand(Src, i))); + LLVMValueRef Fn = CloneValue(LLVMGetCalledValue(Src)); + LLVMBasicBlockRef Then = DeclareBB(LLVMGetNormalDest(Src)); + LLVMBasicBlockRef Unwind = DeclareBB(LLVMGetUnwindDest(Src)); + Dst = LLVMBuildInvoke(Builder, Fn, Args.data(), ArgCount, + Then, Unwind, Name); + CloneAttrs(Src, Dst); + break; + } + case LLVMUnreachable: + Dst = LLVMBuildUnreachable(Builder); + break; + case LLVMAdd: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildAdd(Builder, LHS, RHS, Name); + break; + } + case LLVMSub: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildSub(Builder, LHS, RHS, Name); + break; + } + case LLVMMul: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildMul(Builder, LHS, RHS, Name); + break; + } + case LLVMUDiv: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildUDiv(Builder, LHS, RHS, Name); + break; + } + case LLVMSDiv: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildSDiv(Builder, LHS, RHS, Name); + break; + } + case LLVMURem: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildURem(Builder, LHS, RHS, Name); + break; + } + case LLVMSRem: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildSRem(Builder, LHS, RHS, Name); + break; + } + case LLVMShl: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildShl(Builder, LHS, RHS, Name); + break; + } + case LLVMLShr: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildLShr(Builder, LHS, RHS, Name); + break; + } + case LLVMAShr: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildAShr(Builder, LHS, RHS, Name); + break; + } + case LLVMAnd: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildAnd(Builder, LHS, RHS, Name); + break; + } + case LLVMOr: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildOr(Builder, LHS, RHS, Name); + break; + } + case LLVMXor: { + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildXor(Builder, LHS, RHS, Name); + break; + } + case LLVMAlloca: { + LLVMTypeRef Ty = CloneType(LLVMGetAllocatedType(Src)); + Dst = LLVMBuildAlloca(Builder, Ty, Name); + break; + } + case LLVMLoad: { + LLVMValueRef Ptr = CloneValue(LLVMGetOperand(Src, 0)); + Dst = LLVMBuildLoad(Builder, Ptr, Name); + LLVMSetAlignment(Dst, LLVMGetAlignment(Src)); + break; + } + case LLVMStore: { + LLVMValueRef Val = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef Ptr = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildStore(Builder, Val, Ptr); + LLVMSetAlignment(Dst, LLVMGetAlignment(Src)); + break; + } + case LLVMGetElementPtr: { + LLVMValueRef Ptr = CloneValue(LLVMGetOperand(Src, 0)); + SmallVector<LLVMValueRef, 8> Idx; + int NumIdx = LLVMGetNumIndices(Src); + for (int i = 1; i <= NumIdx; i++) + Idx.push_back(CloneValue(LLVMGetOperand(Src, i))); + if (LLVMIsInBounds(Src)) + Dst = LLVMBuildInBoundsGEP(Builder, Ptr, Idx.data(), NumIdx, Name); + else + Dst = LLVMBuildGEP(Builder, Ptr, Idx.data(), NumIdx, Name); + break; + } + case LLVMAtomicCmpXchg: { + LLVMValueRef Ptr = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef Cmp = CloneValue(LLVMGetOperand(Src, 1)); + LLVMValueRef New = CloneValue(LLVMGetOperand(Src, 2)); + LLVMAtomicOrdering Succ = LLVMGetCmpXchgSuccessOrdering(Src); + LLVMAtomicOrdering Fail = LLVMGetCmpXchgFailureOrdering(Src); + LLVMBool SingleThread = LLVMIsAtomicSingleThread(Src); + + Dst = LLVMBuildAtomicCmpXchg(Builder, Ptr, Cmp, New, Succ, Fail, + SingleThread); + } break; + case LLVMBitCast: { + LLVMValueRef V = CloneValue(LLVMGetOperand(Src, 0)); + Dst = LLVMBuildBitCast(Builder, V, CloneType(Src), Name); + break; + } + case LLVMICmp: { + LLVMIntPredicate Pred = LLVMGetICmpPredicate(Src); + LLVMValueRef LHS = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef RHS = CloneValue(LLVMGetOperand(Src, 1)); + Dst = LLVMBuildICmp(Builder, Pred, LHS, RHS, Name); + break; + } + case LLVMPHI: { + // We need to agressively set things here because of loops. + VMap[Src] = Dst = LLVMBuildPhi(Builder, CloneType(Src), Name); + + SmallVector<LLVMValueRef, 8> Values; + SmallVector<LLVMBasicBlockRef, 8> Blocks; + + unsigned IncomingCount = LLVMCountIncoming(Src); + for (unsigned i = 0; i < IncomingCount; ++i) { + Blocks.push_back(DeclareBB(LLVMGetIncomingBlock(Src, i))); + Values.push_back(CloneValue(LLVMGetIncomingValue(Src, i))); + } + + LLVMAddIncoming(Dst, Values.data(), Blocks.data(), IncomingCount); + return Dst; + } + case LLVMCall: { + SmallVector<LLVMValueRef, 8> Args; + int ArgCount = LLVMGetNumArgOperands(Src); + for (int i = 0; i < ArgCount; i++) + Args.push_back(CloneValue(LLVMGetOperand(Src, i))); + LLVMValueRef Fn = CloneValue(LLVMGetCalledValue(Src)); + Dst = LLVMBuildCall(Builder, Fn, Args.data(), ArgCount, Name); + LLVMSetTailCall(Dst, LLVMIsTailCall(Src)); + CloneAttrs(Src, Dst); + break; + } + case LLVMResume: { + Dst = LLVMBuildResume(Builder, CloneValue(LLVMGetOperand(Src, 0))); + break; + } + case LLVMLandingPad: { + // The landing pad API is a bit screwed up for historical reasons. + Dst = LLVMBuildLandingPad(Builder, CloneType(Src), nullptr, 0, Name); + unsigned NumClauses = LLVMGetNumClauses(Src); + for (unsigned i = 0; i < NumClauses; ++i) + LLVMAddClause(Dst, CloneValue(LLVMGetClause(Src, i))); + LLVMSetCleanup(Dst, LLVMIsCleanup(Src)); + break; + } + case LLVMExtractValue: { + LLVMValueRef Agg = CloneValue(LLVMGetOperand(Src, 0)); + if (LLVMGetNumIndices(Src) != 1) + report_fatal_error("Expected only one indice"); + auto I = LLVMGetIndices(Src)[0]; + Dst = LLVMBuildExtractValue(Builder, Agg, I, Name); + break; + } + case LLVMInsertValue: { + LLVMValueRef Agg = CloneValue(LLVMGetOperand(Src, 0)); + LLVMValueRef V = CloneValue(LLVMGetOperand(Src, 1)); + if (LLVMGetNumIndices(Src) != 1) + report_fatal_error("Expected only one indice"); + auto I = LLVMGetIndices(Src)[0]; + Dst = LLVMBuildInsertValue(Builder, Agg, V, I, Name); + break; + } + default: + break; + } + + if (Dst == nullptr) { + fprintf(stderr, "%d is not a supported opcode\n", Op); + exit(-1); + } + + check_value_kind(Dst, LLVMInstructionValueKind); + return VMap[Src] = Dst; + } + + LLVMBasicBlockRef DeclareBB(LLVMBasicBlockRef Src) { + // Check if this is something we already computed. + { + auto i = BBMap.find(Src); + if (i != BBMap.end()) { + return i->second; + } + } + + LLVMValueRef V = LLVMBasicBlockAsValue(Src); + if (!LLVMValueIsBasicBlock(V) || LLVMValueAsBasicBlock(V) != Src) + report_fatal_error("Basic block is not a basic block"); + + const char *Name = LLVMGetBasicBlockName(Src); + const char *VName = LLVMGetValueName(V); + if (Name != VName) + report_fatal_error("Basic block name mismatch"); + + LLVMBasicBlockRef BB = LLVMAppendBasicBlock(Fun, Name); + return BBMap[Src] = BB; + } + + LLVMBasicBlockRef CloneBB(LLVMBasicBlockRef Src) { + LLVMBasicBlockRef BB = DeclareBB(Src); + + // Make sure ordering is correct. + LLVMBasicBlockRef Prev = LLVMGetPreviousBasicBlock(Src); + if (Prev) + LLVMMoveBasicBlockAfter(BB, DeclareBB(Prev)); + + LLVMValueRef First = LLVMGetFirstInstruction(Src); + LLVMValueRef Last = LLVMGetLastInstruction(Src); + + if (First == nullptr) { + if (Last != nullptr) + report_fatal_error("Has no first instruction, but last one"); + return BB; + } + + auto Ctx = LLVMGetModuleContext(M); + LLVMBuilderRef Builder = LLVMCreateBuilderInContext(Ctx); + LLVMPositionBuilderAtEnd(Builder, BB); + + LLVMValueRef Cur = First; + LLVMValueRef Next = nullptr; + while(true) { + CloneInstruction(Cur, Builder); + Next = LLVMGetNextInstruction(Cur); + if (Next == nullptr) { + if (Cur != Last) + report_fatal_error("Final instruction does not match Last"); + break; + } + + LLVMValueRef Prev = LLVMGetPreviousInstruction(Next); + if (Prev != Cur) + report_fatal_error("Next.Previous instruction is not Current"); + + Cur = Next; + } + + LLVMDisposeBuilder(Builder); + return BB; + } + + void CloneBBs(LLVMValueRef Src) { + unsigned Count = LLVMCountBasicBlocks(Src); + if (Count == 0) + return; + + LLVMBasicBlockRef First = LLVMGetFirstBasicBlock(Src); + LLVMBasicBlockRef Last = LLVMGetLastBasicBlock(Src); + + LLVMBasicBlockRef Cur = First; + LLVMBasicBlockRef Next = nullptr; + while(true) { + CloneBB(Cur); + Count--; + Next = LLVMGetNextBasicBlock(Cur); + if (Next == nullptr) { + if (Cur != Last) + report_fatal_error("Final basic block does not match Last"); + break; + } + + LLVMBasicBlockRef Prev = LLVMGetPreviousBasicBlock(Next); + if (Prev != Cur) + report_fatal_error("Next.Previous basic bloc is not Current"); + + Cur = Next; + } + + if (Count != 0) + report_fatal_error("Basic block count does not match iterration"); + } +}; + +static void declare_symbols(LLVMModuleRef Src, LLVMModuleRef M) { + LLVMValueRef Begin = LLVMGetFirstGlobal(Src); + LLVMValueRef End = LLVMGetLastGlobal(Src); + + LLVMValueRef Cur = Begin; + LLVMValueRef Next = nullptr; + if (!Begin) { + if (End != nullptr) + report_fatal_error("Range has an end but no begining"); + goto FunDecl; + } + + while (true) { + const char *Name = LLVMGetValueName(Cur); + if (LLVMGetNamedGlobal(M, Name)) + report_fatal_error("GlobalVariable already cloned"); + LLVMAddGlobal(M, LLVMGetElementType(TypeCloner(M).Clone(Cur)), Name); + + Next = LLVMGetNextGlobal(Cur); + if (Next == nullptr) { + if (Cur != End) + report_fatal_error(""); + break; + } + + LLVMValueRef Prev = LLVMGetPreviousGlobal(Next); + if (Prev != Cur) + report_fatal_error("Next.Previous global is not Current"); + + Cur = Next; + } + +FunDecl: + Begin = LLVMGetFirstFunction(Src); + End = LLVMGetLastFunction(Src); + if (!Begin) { + if (End != nullptr) + report_fatal_error("Range has an end but no begining"); + return; + } + + auto Ctx = LLVMGetModuleContext(M); + + Cur = Begin; + Next = nullptr; + while (true) { + const char *Name = LLVMGetValueName(Cur); + if (LLVMGetNamedFunction(M, Name)) + report_fatal_error("Function already cloned"); + auto Ty = LLVMGetElementType(TypeCloner(M).Clone(Cur)); + auto F = LLVMAddFunction(M, Name, Ty); + + // Copy attributes + for (int i = LLVMAttributeFunctionIndex, c = LLVMCountParams(F); + i <= c; ++i) { + for (unsigned k = 0, e = LLVMGetLastEnumAttributeKind(); k < e; ++k) { + if (auto SrcA = LLVMGetEnumAttributeAtIndex(Cur, i, k)) { + auto Val = LLVMGetEnumAttributeValue(SrcA); + auto DstA = LLVMCreateEnumAttribute(Ctx, k, Val); + LLVMAddAttributeAtIndex(F, i, DstA); + } + } + } + + Next = LLVMGetNextFunction(Cur); + if (Next == nullptr) { + if (Cur != End) + report_fatal_error("Last function does not match End"); + break; + } + + LLVMValueRef Prev = LLVMGetPreviousFunction(Next); + if (Prev != Cur) + report_fatal_error("Next.Previous function is not Current"); + + Cur = Next; + } +} + +static void clone_symbols(LLVMModuleRef Src, LLVMModuleRef M) { + LLVMValueRef Begin = LLVMGetFirstGlobal(Src); + LLVMValueRef End = LLVMGetLastGlobal(Src); + + LLVMValueRef Cur = Begin; + LLVMValueRef Next = nullptr; + if (!Begin) { + if (End != nullptr) + report_fatal_error("Range has an end but no begining"); + goto FunClone; + } + + while (true) { + const char *Name = LLVMGetValueName(Cur); + LLVMValueRef G = LLVMGetNamedGlobal(M, Name); + if (!G) + report_fatal_error("GlobalVariable must have been declared already"); + + if (auto I = LLVMGetInitializer(Cur)) + LLVMSetInitializer(G, clone_constant(I, M)); + + LLVMSetGlobalConstant(G, LLVMIsGlobalConstant(Cur)); + LLVMSetThreadLocal(G, LLVMIsThreadLocal(Cur)); + LLVMSetExternallyInitialized(G, LLVMIsExternallyInitialized(Cur)); + LLVMSetLinkage(G, LLVMGetLinkage(Cur)); + LLVMSetSection(G, LLVMGetSection(Cur)); + LLVMSetVisibility(G, LLVMGetVisibility(Cur)); + LLVMSetUnnamedAddr(G, LLVMHasUnnamedAddr(Cur)); + LLVMSetAlignment(G, LLVMGetAlignment(Cur)); + + Next = LLVMGetNextGlobal(Cur); + if (Next == nullptr) { + if (Cur != End) + report_fatal_error(""); + break; + } + + LLVMValueRef Prev = LLVMGetPreviousGlobal(Next); + if (Prev != Cur) + report_fatal_error("Next.Previous global is not Current"); + + Cur = Next; + } + +FunClone: + Begin = LLVMGetFirstFunction(Src); + End = LLVMGetLastFunction(Src); + if (!Begin) { + if (End != nullptr) + report_fatal_error("Range has an end but no begining"); + return; + } + + Cur = Begin; + Next = nullptr; + while (true) { + const char *Name = LLVMGetValueName(Cur); + LLVMValueRef Fun = LLVMGetNamedFunction(M, Name); + if (!Fun) + report_fatal_error("Function must have been declared already"); + + if (LLVMHasPersonalityFn(Cur)) { + const char *FName = LLVMGetValueName(LLVMGetPersonalityFn(Cur)); + LLVMValueRef P = LLVMGetNamedFunction(M, FName); + if (!P) + report_fatal_error("Could not find personality function"); + LLVMSetPersonalityFn(Fun, P); + } + + FunCloner FC(Cur, Fun); + FC.CloneBBs(Cur); + + Next = LLVMGetNextFunction(Cur); + if (Next == nullptr) { + if (Cur != End) + report_fatal_error("Last function does not match End"); + break; + } + + LLVMValueRef Prev = LLVMGetPreviousFunction(Next); + if (Prev != Cur) + report_fatal_error("Next.Previous function is not Current"); + + Cur = Next; + } +} + +int llvm_echo(void) { + LLVMEnablePrettyStackTrace(); + + LLVMModuleRef Src = llvm_load_module(false, true); + size_t Len; + const char *ModuleName = LLVMGetModuleIdentifier(Src, &Len); + LLVMContextRef Ctx = LLVMContextCreate(); + LLVMModuleRef M = LLVMModuleCreateWithNameInContext(ModuleName, Ctx); + + // This whole switcharound is done because the C API has no way to + // set the source_filename + LLVMSetModuleIdentifier(M, "", 0); + LLVMGetModuleIdentifier(M, &Len); + if (Len != 0) + report_fatal_error("LLVM{Set,Get}ModuleIdentifier failed"); + LLVMSetModuleIdentifier(M, ModuleName, strlen(ModuleName)); + + LLVMSetTarget(M, LLVMGetTarget(Src)); + LLVMSetModuleDataLayout(M, LLVMGetModuleDataLayout(Src)); + if (strcmp(LLVMGetDataLayoutStr(M), LLVMGetDataLayoutStr(Src))) + report_fatal_error("Inconsistent DataLayout string representation"); + + declare_symbols(Src, M); + clone_symbols(Src, M); + char *Str = LLVMPrintModuleToString(M); + fputs(Str, stdout); + + LLVMDisposeMessage(Str); + LLVMDisposeModule(M); + LLVMContextDispose(Ctx); + + return 0; +} diff --git a/tools/llvm-c-test/helpers.c b/tools/llvm-c-test/helpers.c index 1ea8a4f669952..97fbaab6d6c33 100644 --- a/tools/llvm-c-test/helpers.c +++ b/tools/llvm-c-test/helpers.c @@ -18,7 +18,7 @@ #define MAX_TOKENS 512 #define MAX_LINE_LEN 1024 -void tokenize_stdin(void (*cb)(char **tokens, int ntokens)) { +void llvm_tokenize_stdin(void (*cb)(char **tokens, int ntokens)) { char line[MAX_LINE_LEN]; char *tokbuf[MAX_TOKENS]; diff --git a/tools/llvm-c-test/llvm-c-test.h b/tools/llvm-c-test/llvm-c-test.h index 7929fc4d19be7..0d1ade093bc7f 100644 --- a/tools/llvm-c-test/llvm-c-test.h +++ b/tools/llvm-c-test/llvm-c-test.h @@ -14,30 +14,46 @@ #define LLVM_C_TEST_H #include <stdbool.h> +#include "llvm-c/Core.h" + +#ifdef __cplusplus +extern "C" { +#endif // helpers.c -void tokenize_stdin(void (*cb)(char **tokens, int ntokens)); +void llvm_tokenize_stdin(void (*cb)(char **tokens, int ntokens)); // module.c -int module_dump(bool Lazy, bool New); -int module_list_functions(void); -int module_list_globals(void); +LLVMModuleRef llvm_load_module(bool Lazy, bool New); +int llvm_module_dump(bool Lazy, bool New); +int llvm_module_list_functions(void); +int llvm_module_list_globals(void); // calc.c -int calc(void); +int llvm_calc(void); // disassemble.c -int disassemble(void); +int llvm_disassemble(void); // metadata.c -int add_named_metadata_operand(void); -int set_metadata(void); +int llvm_add_named_metadata_operand(void); +int llvm_set_metadata(void); // object.c -int object_list_sections(void); -int object_list_symbols(void); +int llvm_object_list_sections(void); +int llvm_object_list_symbols(void); // targets.c -int targets_list(void); +int llvm_targets_list(void); + +// echo.c +int llvm_echo(void); + +// diagnostic.c +int llvm_test_diagnostic_handler(void); + +#ifdef __cplusplus +} +#endif /* !defined(__cplusplus) */ #endif diff --git a/tools/llvm-c-test/main.c b/tools/llvm-c-test/main.c index e6b6e17098b36..90d3501778516 100644 --- a/tools/llvm-c-test/main.c +++ b/tools/llvm-c-test/main.c @@ -13,7 +13,6 @@ #include "llvm-c-test.h" #include "llvm-c/BitReader.h" -#include "llvm-c/Core.h" #include <stdio.h> #include <stdlib.h> #include <string.h> @@ -22,20 +21,20 @@ static void print_usage(void) { fprintf(stderr, "llvm-c-test command\n\n"); fprintf(stderr, " Commands:\n"); fprintf(stderr, " * --module-dump\n"); - fprintf(stderr, " Read bytecode from stdin - print disassembly\n\n"); + fprintf(stderr, " Read bitcode from stdin - print disassembly\n\n"); fprintf(stderr, " * --lazy-module-dump\n"); fprintf(stderr, - " Lazily read bytecode from stdin - print disassembly\n\n"); + " Lazily read bitcode from stdin - print disassembly\n\n"); fprintf(stderr, " * --new-module-dump\n"); - fprintf(stderr, " Read bytecode from stdin - print disassembly\n\n"); + fprintf(stderr, " Read bitcode from stdin - print disassembly\n\n"); fprintf(stderr, " * --lazy-new-module-dump\n"); fprintf(stderr, - " Lazily read bytecode from stdin - print disassembly\n\n"); + " Lazily read bitcode from stdin - print disassembly\n\n"); fprintf(stderr, " * --module-list-functions\n"); fprintf(stderr, - " Read bytecode from stdin - list summary of functions\n\n"); + " Read bitcode from stdin - list summary of functions\n\n"); fprintf(stderr, " * --module-list-globals\n"); - fprintf(stderr, " Read bytecode from stdin - list summary of globals\n\n"); + fprintf(stderr, " Read bitcode from stdin - list summary of globals\n\n"); fprintf(stderr, " * --targets-list\n"); fprintf(stderr, " List available targets\n\n"); fprintf(stderr, " * --object-list-sections\n"); @@ -50,6 +49,12 @@ static void print_usage(void) { fprintf( stderr, " Read lines of name, rpn from stdin - print generated module\n\n"); + fprintf(stderr, " * --echo\n"); + fprintf(stderr, + " Read bitcode file form stdin - print it back out\n\n"); + fprintf(stderr, " * --test-diagnostic-handler\n"); + fprintf(stderr, + " Read bitcode file form stdin with a diagnostic handler set\n\n"); } int main(int argc, char **argv) { @@ -58,31 +63,35 @@ int main(int argc, char **argv) { LLVMInitializeCore(pr); if (argc == 2 && !strcmp(argv[1], "--lazy-new-module-dump")) { - return module_dump(true, true); + return llvm_module_dump(true, true); } else if (argc == 2 && !strcmp(argv[1], "--new-module-dump")) { - return module_dump(false, true); + return llvm_module_dump(false, true); } else if (argc == 2 && !strcmp(argv[1], "--lazy-module-dump")) { - return module_dump(true, false); + return llvm_module_dump(true, false); } else if (argc == 2 && !strcmp(argv[1], "--module-dump")) { - return module_dump(false, false); + return llvm_module_dump(false, false); } else if (argc == 2 && !strcmp(argv[1], "--module-list-functions")) { - return module_list_functions(); + return llvm_module_list_functions(); } else if (argc == 2 && !strcmp(argv[1], "--module-list-globals")) { - return module_list_globals(); + return llvm_module_list_globals(); } else if (argc == 2 && !strcmp(argv[1], "--targets-list")) { - return targets_list(); + return llvm_targets_list(); } else if (argc == 2 && !strcmp(argv[1], "--object-list-sections")) { - return object_list_sections(); + return llvm_object_list_sections(); } else if (argc == 2 && !strcmp(argv[1], "--object-list-symbols")) { - return object_list_symbols(); + return llvm_object_list_symbols(); } else if (argc == 2 && !strcmp(argv[1], "--disassemble")) { - return disassemble(); + return llvm_disassemble(); } else if (argc == 2 && !strcmp(argv[1], "--calc")) { - return calc(); + return llvm_calc(); } else if (argc == 2 && !strcmp(argv[1], "--add-named-metadata-operand")) { - return add_named_metadata_operand(); + return llvm_add_named_metadata_operand(); } else if (argc == 2 && !strcmp(argv[1], "--set-metadata")) { - return set_metadata(); + return llvm_set_metadata(); + } else if (argc == 2 && !strcmp(argv[1], "--echo")) { + return llvm_echo(); + } else if (argc == 2 && !strcmp(argv[1], "--test-diagnostic-handler")) { + return llvm_test_diagnostic_handler(); } else { print_usage(); } diff --git a/tools/llvm-c-test/metadata.c b/tools/llvm-c-test/metadata.c index b64a696c52810..89215b8ebcbd8 100644 --- a/tools/llvm-c-test/metadata.c +++ b/tools/llvm-c-test/metadata.c @@ -13,9 +13,8 @@ \*===----------------------------------------------------------------------===*/ #include "llvm-c-test.h" -#include "llvm-c/Core.h" -int add_named_metadata_operand(void) { +int llvm_add_named_metadata_operand(void) { LLVMModuleRef m = LLVMModuleCreateWithName("Mod"); LLVMValueRef values[] = { LLVMConstInt(LLVMInt32Type(), 0, 0) }; @@ -27,7 +26,7 @@ int add_named_metadata_operand(void) { return 0; } -int set_metadata(void) { +int llvm_set_metadata(void) { LLVMBuilderRef b = LLVMCreateBuilder(); LLVMValueRef values[] = { LLVMConstInt(LLVMInt32Type(), 0, 0) }; diff --git a/tools/llvm-c-test/module.c b/tools/llvm-c-test/module.c index a6c47bf5fa168..c47b55d502949 100644 --- a/tools/llvm-c-test/module.c +++ b/tools/llvm-c-test/module.c @@ -14,7 +14,6 @@ #include "llvm-c-test.h" #include "llvm-c/BitReader.h" -#include "llvm-c/Core.h" #include <stdio.h> #include <stdlib.h> #include <string.h> @@ -26,7 +25,7 @@ static void diagnosticHandler(LLVMDiagnosticInfoRef DI, void *C) { exit(1); } -static LLVMModuleRef load_module(bool Lazy, bool New) { +LLVMModuleRef llvm_load_module(bool Lazy, bool New) { LLVMMemoryBufferRef MB; LLVMModuleRef M; char *msg = NULL; @@ -63,8 +62,8 @@ static LLVMModuleRef load_module(bool Lazy, bool New) { return M; } -int module_dump(bool Lazy, bool New) { - LLVMModuleRef M = load_module(Lazy, New); +int llvm_module_dump(bool Lazy, bool New) { + LLVMModuleRef M = llvm_load_module(Lazy, New); char *irstr = LLVMPrintModuleToString(M); puts(irstr); @@ -75,8 +74,8 @@ int module_dump(bool Lazy, bool New) { return 0; } -int module_list_functions(void) { - LLVMModuleRef M = load_module(false, false); +int llvm_module_list_functions(void) { + LLVMModuleRef M = llvm_load_module(false, false); LLVMValueRef f; f = LLVMGetFirstFunction(M); @@ -116,8 +115,8 @@ int module_list_functions(void) { return 0; } -int module_list_globals(void) { - LLVMModuleRef M = load_module(false, false); +int llvm_module_list_globals(void) { + LLVMModuleRef M = llvm_load_module(false, false); LLVMValueRef g; g = LLVMGetFirstGlobal(M); diff --git a/tools/llvm-c-test/object.c b/tools/llvm-c-test/object.c index 43521787b6099..809ad54f8721c 100644 --- a/tools/llvm-c-test/object.c +++ b/tools/llvm-c-test/object.c @@ -13,12 +13,11 @@ \*===----------------------------------------------------------------------===*/ #include "llvm-c-test.h" -#include "llvm-c/Core.h" #include "llvm-c/Object.h" #include <stdio.h> #include <stdlib.h> -int object_list_sections(void) { +int llvm_object_list_sections(void) { LLVMMemoryBufferRef MB; LLVMObjectFileRef O; LLVMSectionIteratorRef sect; @@ -50,7 +49,7 @@ int object_list_sections(void) { return 0; } -int object_list_symbols(void) { +int llvm_object_list_symbols(void) { LLVMMemoryBufferRef MB; LLVMObjectFileRef O; LLVMSectionIteratorRef sect; diff --git a/tools/llvm-c-test/targets.c b/tools/llvm-c-test/targets.c index 252c2e02a67ab..f2a9e924a74a3 100644 --- a/tools/llvm-c-test/targets.c +++ b/tools/llvm-c-test/targets.c @@ -14,7 +14,7 @@ #include "llvm-c/TargetMachine.h" #include <stdio.h> -int targets_list(void) { +int llvm_targets_list(void) { LLVMTargetRef t; LLVMInitializeAllTargetInfos(); LLVMInitializeAllTargets(); diff --git a/tools/llvm-config/BuildVariables.inc.in b/tools/llvm-config/BuildVariables.inc.in index 345f47e91b4a3..709ea35044c6a 100644 --- a/tools/llvm-config/BuildVariables.inc.in +++ b/tools/llvm-config/BuildVariables.inc.in @@ -29,5 +29,8 @@ #define LLVM_BUILD_SYSTEM "@LLVM_BUILD_SYSTEM@" #define LLVM_HAS_RTTI "@LLVM_HAS_RTTI@" #define LLVM_ENABLE_DYLIB "@LLVM_BUILD_LLVM_DYLIB@" +#define LLVM_LINK_DYLIB "@LLVM_LINK_LLVM_DYLIB@" #define LLVM_ENABLE_SHARED "@LLVM_ENABLE_SHARED@" #define LLVM_DYLIB_COMPONENTS "@LLVM_DYLIB_COMPONENTS@" +#define LLVM_DYLIB_VERSION "@LLVM_DYLIB_VERSION@" +#define LLVM_HAS_GLOBAL_ISEL "@LLVM_HAS_GLOBAL_ISEL@" diff --git a/tools/llvm-config/CMakeLists.txt b/tools/llvm-config/CMakeLists.txt index 83794bb3fddd6..d45877135ba2a 100644 --- a/tools/llvm-config/CMakeLists.txt +++ b/tools/llvm-config/CMakeLists.txt @@ -11,7 +11,17 @@ add_llvm_tool(llvm-config # Compute the substitution values for various items. get_property(LLVM_SYSTEM_LIBS_LIST TARGET LLVMSupport PROPERTY LLVM_SYSTEM_LIBS) foreach(l ${LLVM_SYSTEM_LIBS_LIST}) - set(SYSTEM_LIBS ${SYSTEM_LIBS} "-l${l}") + if(MSVC) + set(SYSTEM_LIBS ${SYSTEM_LIBS} "${l}.lib") + else() + if (l MATCHES "^-") + # If it's an option, pass it without changes. + set(SYSTEM_LIBS ${SYSTEM_LIBS} "${l}") + else() + # Otherwise assume it's a library name we need to link with. + set(SYSTEM_LIBS ${SYSTEM_LIBS} "-l${l}") + endif() + endif() endforeach() string(REPLACE ";" " " SYSTEM_LIBS "${SYSTEM_LIBS}") @@ -26,6 +36,8 @@ set(LLVM_CFLAGS "${CMAKE_C_FLAGS} ${CMAKE_C_FLAGS_${uppercase_CMAKE_BUILD_TYPE}} set(LLVM_CXXFLAGS "${CMAKE_CXX_FLAGS} ${CMAKE_CXX_FLAGS_${uppercase_CMAKE_BUILD_TYPE}} ${COMPILE_FLAGS} ${LLVM_DEFINITIONS}") set(LLVM_BUILD_SYSTEM cmake) set(LLVM_HAS_RTTI ${LLVM_CONFIG_HAS_RTTI}) +set(LLVM_DYLIB_VERSION "${LLVM_VERSION_MAJOR}.${LLVM_VERSION_MINOR}${LLVM_VERSION_SUFFIX}") +set(LLVM_HAS_GLOBAL_ISEL "${LLVM_BUILD_GLOBAL_ISEL}") # Use the C++ link flags, since they should be a superset of C link flags. set(LLVM_LDFLAGS "${CMAKE_CXX_LINK_FLAGS}") diff --git a/tools/llvm-config/Makefile b/tools/llvm-config/Makefile deleted file mode 100644 index d2fe2cfa34aad..0000000000000 --- a/tools/llvm-config/Makefile +++ /dev/null @@ -1,95 +0,0 @@ -##===- tools/llvm-config/Makefile---------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-config -USEDLIBS := LLVMSupport.a - -# We generate sources in the build directory, make sure it is in the include -# paths. -INCLUDE_BUILD_DIR := 1 - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -# Note that we have to use lazy expansion here. -BUILDVARIABLES_SRCPATH = $(PROJ_SRC_ROOT)/tools/$(TOOLNAME)/BuildVariables.inc.in -BUILDVARIABLES_OBJPATH = $(ObjDir)/BuildVariables.inc -BUILT_SOURCES = $(BUILDVARIABLES_OBJPATH) - -include $(LEVEL)/Makefile.common - -# Combine preprocessor flags (except for -I) and CXX flags. -SUB_CPPFLAGS := ${CPP.BaseFlags} -SUB_CFLAGS := ${CPP.BaseFlags} ${C.Flags} -SUB_CXXFLAGS := ${CPP.BaseFlags} ${CXX.Flags} - -# Override LIBS with TARGET's LIBS for cross compilation. -# FIXME: Host's llvm-config is not generated. It's for target's. -ifneq ($(TARGET_LIBS), ) - LLVM_SYSTEM_LIBS := $(TARGET_LIBS) -else - LLVM_SYSTEM_LIBS := $(LIBS) -endif - -ifneq ($(REQUIRES_RTTI), 1) - LLVM_HAS_RTTI := NO -else - LLVM_HAS_RTTI := YES -endif - -# This is blank for now. We need to be careful about adding stuff here: -# LDFLAGS tend not to be portable, and we don't currently require the -# user to use libtool when linking against LLVM. -SUB_LDFLAGS := - -$(ObjDir)/BuildVariables.inc: $(BUILDVARIABLES_SRCPATH) Makefile $(ObjDir)/.dir - $(Echo) "Building llvm-config BuildVariables.inc file." - $(Verb) $(ECHO) 's/@LLVM_SRC_ROOT@/$(subst /,\/,$(LLVM_SRC_ROOT))/' \ - > temp.sed - $(Verb) $(ECHO) 's/@LLVM_OBJ_ROOT@/$(subst /,\/,$(LLVM_OBJ_ROOT))/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_CPPFLAGS@/$(subst /,\/,$(SUB_CPPFLAGS))/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_CFLAGS@/$(subst /,\/,$(SUB_CFLAGS))/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_CXXFLAGS@/$(subst /,\/,$(SUB_CXXFLAGS))/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_LDFLAGS@/$(subst /,\/,$(SUB_LDFLAGS))/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_BUILDMODE@/$(subst /,\/,$(BuildMode))/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_LIBDIR_SUFFIX@//' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_SYSTEM_LIBS@/$(subst /,\/,$(LLVM_SYSTEM_LIBS))/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_TARGETS_BUILT@/$(subst /,\/,$(TARGETS_TO_BUILD))/' \ - >> temp.sed - $(if $(filter-out $(ENABLE_SHARED),0),\ - $(Verb) $(ECHO) 's/@LLVM_BUILD_LLVM_DYLIB@/ON/',\ - $(Verb) $(ECHO) 's/@LLVM_BUILD_LLVM_DYLIB@/OFF/') \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_ENABLE_SHARED@/OFF/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_DYLIB_COMPONENTS@/all/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_BUILD_SYSTEM@/autoconf/' \ - >> temp.sed - $(Verb) $(ECHO) 's/@LLVM_HAS_RTTI@/$(LLVM_HAS_RTTI)/' \ - >> temp.sed - $(Verb) $(SED) -f temp.sed < $< > $@ - $(Verb) $(RM) temp.sed - -# When cross-compiling, install a version of llvm-config that runs on the host. -ifeq ($(LLVM_CROSS_COMPILING),1) -install:: $(DESTDIR)$(PROJ_bindir) - $(Echo) Installing llvm-config-host - $(Verb) $(ProgInstall) $(BuildLLVMToolDir)/llvm-config \ - $(DESTDIR)$(PROJ_bindir)/$(program_prefix)llvm-config-host -endif diff --git a/tools/llvm-config/llvm-config.cpp b/tools/llvm-config/llvm-config.cpp index 80f627936d0cf..94d426be3613e 100644 --- a/tools/llvm-config/llvm-config.cpp +++ b/tools/llvm-config/llvm-config.cpp @@ -29,8 +29,8 @@ #include "llvm/Support/raw_ostream.h" #include <cstdlib> #include <set> -#include <vector> #include <unordered_set> +#include <vector> using namespace llvm; @@ -46,6 +46,22 @@ using namespace llvm; // create entries for pseudo groups like x86 or all-targets. #include "LibraryDependencies.inc" +// LinkMode determines what libraries and flags are returned by llvm-config. +enum LinkMode { + // LinkModeAuto will link with the default link mode for the installation, + // which is dependent on the value of LLVM_LINK_LLVM_DYLIB, and fall back + // to the alternative if the required libraries are not available. + LinkModeAuto = 0, + + // LinkModeShared will link with the dynamic component libraries if they + // exist, and return an error otherwise. + LinkModeShared = 1, + + // LinkModeStatic will link with the static component libraries if they + // exist, and return an error otherwise. + LinkModeStatic = 2, +}; + /// \brief Traverse a single component adding to the topological ordering in /// \arg RequiredLibs. /// @@ -56,14 +72,25 @@ using namespace llvm; /// libraries. /// \param GetComponentNames - Get the component names instead of the /// library name. -static void VisitComponent(const std::string& Name, - const StringMap<AvailableComponent*> &ComponentMap, - std::set<AvailableComponent*> &VisitedComponents, +static void VisitComponent(const std::string &Name, + const StringMap<AvailableComponent *> &ComponentMap, + std::set<AvailableComponent *> &VisitedComponents, std::vector<std::string> &RequiredLibs, bool IncludeNonInstalled, bool GetComponentNames, - const std::string *ActiveLibDir, bool *HasMissing) { + const std::function<std::string(const StringRef &)> + *GetComponentLibraryPath, + std::vector<std::string> *Missing, + const std::string &DirSep) { // Lookup the component. AvailableComponent *AC = ComponentMap.lookup(Name); + if (!AC) { + errs() << "Can't find component: '" << Name << "' in the map. Available components are: "; + for (const auto &Component : ComponentMap) { + errs() << "'" << Component.first() << "' "; + } + errs() << "\n"; + report_fatal_error("abort"); + } assert(AC && "Invalid component name!"); // Add to the visited table. @@ -80,7 +107,7 @@ static void VisitComponent(const std::string& Name, for (unsigned i = 0; AC->RequiredLibraries[i]; ++i) { VisitComponent(AC->RequiredLibraries[i], ComponentMap, VisitedComponents, RequiredLibs, IncludeNonInstalled, GetComponentNames, - ActiveLibDir, HasMissing); + GetComponentLibraryPath, Missing, DirSep); } if (GetComponentNames) { @@ -90,8 +117,13 @@ static void VisitComponent(const std::string& Name, // Add to the required library list. if (AC->Library) { - if (!IncludeNonInstalled && HasMissing && !*HasMissing && ActiveLibDir) { - *HasMissing = !sys::fs::exists(*ActiveLibDir + "/" + AC->Library); + if (Missing && GetComponentLibraryPath) { + std::string path = (*GetComponentLibraryPath)(AC->Library); + if (DirSep == "\\") { + std::replace(path.begin(), path.end(), '/', '\\'); + } + if (!sys::fs::exists(path)) + Missing->push_back(path); } RequiredLibs.push_back(AC->Library); } @@ -105,15 +137,16 @@ static void VisitComponent(const std::string& Name, /// \param IncludeNonInstalled - Whether non-installed components should be /// reported. /// \param GetComponentNames - True if one would prefer the component names. -static std::vector<std::string> -ComputeLibsForComponents(const std::vector<StringRef> &Components, - bool IncludeNonInstalled, bool GetComponentNames, - const std::string *ActiveLibDir, bool *HasMissing) { +static std::vector<std::string> ComputeLibsForComponents( + const std::vector<StringRef> &Components, bool IncludeNonInstalled, + bool GetComponentNames, const std::function<std::string(const StringRef &)> + *GetComponentLibraryPath, + std::vector<std::string> *Missing, const std::string &DirSep) { std::vector<std::string> RequiredLibs; std::set<AvailableComponent *> VisitedComponents; // Build a map of component names to information. - StringMap<AvailableComponent*> ComponentMap; + StringMap<AvailableComponent *> ComponentMap; for (unsigned i = 0; i != array_lengthof(AvailableComponents); ++i) { AvailableComponent *AC = &AvailableComponents[i]; ComponentMap[AC->Name] = AC; @@ -133,7 +166,7 @@ ComputeLibsForComponents(const std::vector<StringRef> &Components, VisitComponent(ComponentLower, ComponentMap, VisitedComponents, RequiredLibs, IncludeNonInstalled, GetComponentNames, - ActiveLibDir, HasMissing); + GetComponentLibraryPath, Missing, DirSep); } // The list is now ordered with leafs first, we want the libraries to printed @@ -176,9 +209,12 @@ Options:\n\ --host-target Target triple used to configure LLVM.\n\ --build-mode Print build mode of LLVM tree (e.g. Debug or Release).\n\ --assertion-mode Print assertion mode of LLVM tree (ON or OFF).\n\ - --build-system Print the build system used to build LLVM (autoconf or cmake).\n\ + --build-system Print the build system used to build LLVM (always cmake).\n\ --has-rtti Print whether or not LLVM was built with rtti (YES or NO).\n\ + --has-global-isel Print whether or not LLVM was built with global-isel support (YES or NO).\n\ --shared-mode Print how the provided components can be collectively linked (`shared` or `static`).\n\ + --link-shared Link the components as shared libraries.\n\ + --link-static Link the component libraries statically.\n\ Typical components:\n\ all All LLVM libraries (default).\n\ engine Either a native JIT or a bitcode interpreter.\n"; @@ -189,14 +225,15 @@ Typical components:\n\ std::string GetExecutablePath(const char *Argv0) { // This just needs to be some symbol in the binary; C++ doesn't // allow taking the address of ::main however. - void *P = (void*) (intptr_t) GetExecutablePath; + void *P = (void *)(intptr_t)GetExecutablePath; return llvm::sys::fs::getMainExecutable(Argv0, P); } /// \brief Expand the semi-colon delimited LLVM_DYLIB_COMPONENTS into /// the full list of components. std::vector<std::string> GetAllDyLibComponents(const bool IsInDevelopmentTree, - const bool GetComponentNames) { + const bool GetComponentNames, + const std::string &DirSep) { std::vector<StringRef> DyLibComponents; StringRef DyLibComponentsStr(LLVM_DYLIB_COMPONENTS); @@ -214,7 +251,7 @@ std::vector<std::string> GetAllDyLibComponents(const bool IsInDevelopmentTree, return ComputeLibsForComponents(DyLibComponents, /*IncludeNonInstalled=*/IsInDevelopmentTree, - GetComponentNames, nullptr, nullptr); + GetComponentNames, nullptr, nullptr, DirSep); } int main(int argc, char **argv) { @@ -228,7 +265,7 @@ int main(int argc, char **argv) { // that we can report the correct information when run from a development // tree. bool IsInDevelopmentTree; - enum { MakefileStyle, CMakeStyle, CMakeBuildModeStyle } DevelopmentTreeLayout; + enum { CMakeStyle, CMakeBuildModeStyle } DevelopmentTreeLayout; llvm::SmallString<256> CurrentPath(GetExecutablePath(argv[0])); std::string CurrentExecPrefix; std::string ActiveObjRoot; @@ -243,25 +280,12 @@ int main(int argc, char **argv) { // Create an absolute path, and pop up one directory (we expect to be inside a // bin dir). sys::fs::make_absolute(CurrentPath); - CurrentExecPrefix = sys::path::parent_path( - sys::path::parent_path(CurrentPath)).str(); + CurrentExecPrefix = + sys::path::parent_path(sys::path::parent_path(CurrentPath)).str(); // Check to see if we are inside a development tree by comparing to possible // locations (prefix style or CMake style). - if (sys::fs::equivalent(CurrentExecPrefix, - Twine(LLVM_OBJ_ROOT) + "/" + build_mode)) { - IsInDevelopmentTree = true; - DevelopmentTreeLayout = MakefileStyle; - - // If we are in a development tree, then check if we are in a BuildTools - // directory. This indicates we are built for the build triple, but we - // always want to provide information for the host triple. - if (sys::path::filename(LLVM_OBJ_ROOT) == "BuildTools") { - ActiveObjRoot = sys::path::parent_path(LLVM_OBJ_ROOT); - } else { - ActiveObjRoot = LLVM_OBJ_ROOT; - } - } else if (sys::fs::equivalent(CurrentExecPrefix, LLVM_OBJ_ROOT)) { + if (sys::fs::equivalent(CurrentExecPrefix, LLVM_OBJ_ROOT)) { IsInDevelopmentTree = true; DevelopmentTreeLayout = CMakeStyle; ActiveObjRoot = LLVM_OBJ_ROOT; @@ -272,7 +296,7 @@ int main(int argc, char **argv) { ActiveObjRoot = LLVM_OBJ_ROOT; } else { IsInDevelopmentTree = false; - DevelopmentTreeLayout = MakefileStyle; // Initialized to avoid warnings. + DevelopmentTreeLayout = CMakeStyle; // Initialized to avoid warnings. } // Compute various directory locations based on the derived location @@ -286,12 +310,6 @@ int main(int argc, char **argv) { // CMake organizes the products differently than a normal prefix style // layout. switch (DevelopmentTreeLayout) { - case MakefileStyle: - ActivePrefix = ActiveObjRoot; - ActiveBinDir = ActiveObjRoot + "/" + build_mode + "/bin"; - ActiveLibDir = - ActiveObjRoot + "/" + build_mode + "/lib" + LLVM_LIBDIR_SUFFIX; - break; case CMakeStyle: ActiveBinDir = ActiveObjRoot + "/bin"; ActiveLibDir = ActiveObjRoot + "/lib" + LLVM_LIBDIR_SUFFIX; @@ -305,8 +323,8 @@ int main(int argc, char **argv) { } // We need to include files from both the source and object trees. - ActiveIncludeOption = ("-I" + ActiveIncludeDir + " " + - "-I" + ActiveObjRoot + "/include"); + ActiveIncludeOption = + ("-I" + ActiveIncludeDir + " " + "-I" + ActiveObjRoot + "/include"); } else { ActivePrefix = CurrentExecPrefix; ActiveIncludeDir = ActivePrefix + "/include"; @@ -323,25 +341,36 @@ int main(int argc, char **argv) { /// in the first place. This can't be done at configure/build time. StringRef SharedExt, SharedVersionedExt, SharedDir, SharedPrefix, StaticExt, - StaticPrefix, StaticDir = "lib"; - const Triple HostTriple(Triple::normalize(LLVM_DEFAULT_TARGET_TRIPLE)); + StaticPrefix, StaticDir = "lib", DirSep = "/"; + const Triple HostTriple(Triple::normalize(LLVM_HOST_TRIPLE)); if (HostTriple.isOSWindows()) { SharedExt = "dll"; - SharedVersionedExt = PACKAGE_VERSION ".dll"; - StaticExt = "a"; + SharedVersionedExt = LLVM_DYLIB_VERSION ".dll"; + if (HostTriple.isOSCygMing()) { + StaticExt = "a"; + StaticPrefix = "lib"; + } else { + StaticExt = "lib"; + DirSep = "\\"; + std::replace(ActiveObjRoot.begin(), ActiveObjRoot.end(), '/', '\\'); + std::replace(ActivePrefix.begin(), ActivePrefix.end(), '/', '\\'); + std::replace(ActiveBinDir.begin(), ActiveBinDir.end(), '/', '\\'); + std::replace(ActiveLibDir.begin(), ActiveLibDir.end(), '/', '\\'); + std::replace(ActiveIncludeOption.begin(), ActiveIncludeOption.end(), '/', + '\\'); + } SharedDir = ActiveBinDir; StaticDir = ActiveLibDir; - StaticPrefix = SharedPrefix = "lib"; } else if (HostTriple.isOSDarwin()) { SharedExt = "dylib"; - SharedVersionedExt = PACKAGE_VERSION ".dylib"; + SharedVersionedExt = LLVM_DYLIB_VERSION ".dylib"; StaticExt = "a"; StaticDir = SharedDir = ActiveLibDir; StaticPrefix = SharedPrefix = "lib"; } else { // default to the unix values: SharedExt = "so"; - SharedVersionedExt = PACKAGE_VERSION ".so"; + SharedVersionedExt = LLVM_DYLIB_VERSION ".so"; StaticExt = "a"; StaticDir = SharedDir = ActiveLibDir; StaticPrefix = SharedPrefix = "lib"; @@ -349,24 +378,32 @@ int main(int argc, char **argv) { const bool BuiltDyLib = (std::strcmp(LLVM_ENABLE_DYLIB, "ON") == 0); - enum { CMake, AutoConf } ConfigTool; - if (std::strcmp(LLVM_BUILD_SYSTEM, "cmake") == 0) { - ConfigTool = CMake; - } else { - ConfigTool = AutoConf; - } - /// CMake style shared libs, ie each component is in a shared library. - const bool BuiltSharedLibs = - (ConfigTool == CMake && std::strcmp(LLVM_ENABLE_SHARED, "ON") == 0); + const bool BuiltSharedLibs = std::strcmp(LLVM_ENABLE_SHARED, "ON") == 0; bool DyLibExists = false; const std::string DyLibName = - (SharedPrefix + "LLVM-" + SharedVersionedExt).str(); + (SharedPrefix + "LLVM-" + SharedVersionedExt).str(); + + // If LLVM_LINK_DYLIB is ON, the single shared library will be returned + // for "--libs", etc, if they exist. This behaviour can be overridden with + // --link-static or --link-shared. + bool LinkDyLib = (std::strcmp(LLVM_LINK_DYLIB, "ON") == 0); if (BuiltDyLib) { - DyLibExists = sys::fs::exists(SharedDir + "/" + DyLibName); + std::string path((SharedDir + DirSep + DyLibName).str()); + if (DirSep == "\\") { + std::replace(path.begin(), path.end(), '/', '\\'); + } + DyLibExists = sys::fs::exists(path); + if (!DyLibExists) { + // The shared library does not exist: don't error unless the user + // explicitly passes --link-shared. + LinkDyLib = false; + } } + LinkMode LinkMode = + (LinkDyLib || BuiltSharedLibs) ? LinkModeShared : LinkModeAuto; /// Get the component's library name without the lib prefix and the /// extension. Returns true if Lib is in a recognized format. @@ -392,28 +429,24 @@ int main(int argc, char **argv) { }; /// Maps Unixizms to the host platform. auto GetComponentLibraryFileName = [&](const StringRef &Lib, - const bool ForceShared) { - std::string LibFileName = Lib; - StringRef LibName; - if (GetComponentLibraryNameSlice(Lib, LibName)) { - if (BuiltSharedLibs || ForceShared) { - LibFileName = (SharedPrefix + LibName + "." + SharedExt).str(); - } else { - // default to static - LibFileName = (StaticPrefix + LibName + "." + StaticExt).str(); - } + const bool Shared) { + std::string LibFileName; + if (Shared) { + LibFileName = (SharedPrefix + Lib + "." + SharedExt).str(); + } else { + // default to static + LibFileName = (StaticPrefix + Lib + "." + StaticExt).str(); } return LibFileName; }; /// Get the full path for a possibly shared component library. - auto GetComponentLibraryPath = [&](const StringRef &Name, - const bool ForceShared) { - auto LibFileName = GetComponentLibraryFileName(Name, ForceShared); - if (BuiltSharedLibs || ForceShared) { - return (SharedDir + "/" + LibFileName).str(); + auto GetComponentLibraryPath = [&](const StringRef &Name, const bool Shared) { + auto LibFileName = GetComponentLibraryFileName(Name, Shared); + if (Shared) { + return (SharedDir + DirSep + LibFileName).str(); } else { - return (StaticDir + "/" + LibFileName).str(); + return (StaticDir + DirSep + LibFileName).str(); } }; @@ -440,7 +473,8 @@ int main(int argc, char **argv) { } else if (Arg == "--cxxflags") { OS << ActiveIncludeOption << ' ' << LLVM_CXXFLAGS << '\n'; } else if (Arg == "--ldflags") { - OS << "-L" << ActiveLibDir << ' ' << LLVM_LDFLAGS << '\n'; + OS << ((HostTriple.isWindowsMSVCEnvironment()) ? "-LIBPATH:" : "-L") + << ActiveLibDir << ' ' << LLVM_LDFLAGS << '\n'; } else if (Arg == "--system-libs") { PrintSystemLibs = true; } else if (Arg == "--libs") { @@ -461,10 +495,14 @@ int main(int argc, char **argv) { Components.push_back(AvailableComponents[j].Name); if (AvailableComponents[j].Library && !IsInDevelopmentTree) { - if (DyLibExists && - !sys::fs::exists(GetComponentLibraryPath( - AvailableComponents[j].Library, false))) { - Components = GetAllDyLibComponents(IsInDevelopmentTree, true); + std::string path( + GetComponentLibraryPath(AvailableComponents[j].Library, false)); + if (DirSep == "\\") { + std::replace(path.begin(), path.end(), '/', '\\'); + } + if (DyLibExists && !sys::fs::exists(path)) { + Components = + GetAllDyLibComponents(IsInDevelopmentTree, true, DirSep); std::sort(Components.begin(), Components.end()); break; } @@ -495,12 +533,18 @@ int main(int argc, char **argv) { OS << LLVM_BUILD_SYSTEM << '\n'; } else if (Arg == "--has-rtti") { OS << LLVM_HAS_RTTI << '\n'; + } else if (Arg == "--has-global-isel") { + OS << LLVM_HAS_GLOBAL_ISEL << '\n'; } else if (Arg == "--shared-mode") { PrintSharedMode = true; } else if (Arg == "--obj-root") { OS << ActivePrefix << '\n'; } else if (Arg == "--src-root") { OS << LLVM_SRC_ROOT << '\n'; + } else if (Arg == "--link-shared") { + LinkMode = LinkModeShared; + } else if (Arg == "--link-static") { + LinkMode = LinkModeStatic; } else { usage(); } @@ -512,6 +556,11 @@ int main(int argc, char **argv) { if (!HasAnyOption) usage(); + if (LinkMode == LinkModeShared && !DyLibExists && !BuiltSharedLibs) { + errs() << "llvm-config: error: " << DyLibName << " is missing\n"; + return 1; + } + if (PrintLibs || PrintLibNames || PrintLibFiles || PrintSystemLibs || PrintSharedMode) { @@ -525,16 +574,45 @@ int main(int argc, char **argv) { Components.push_back("all"); // Construct the list of all the required libraries. - bool HasMissing = false; - std::vector<std::string> RequiredLibs = - ComputeLibsForComponents(Components, - /*IncludeNonInstalled=*/IsInDevelopmentTree, - false, &ActiveLibDir, &HasMissing); + std::function<std::string(const StringRef &)> + GetComponentLibraryPathFunction = [&](const StringRef &Name) { + return GetComponentLibraryPath(Name, LinkMode == LinkModeShared); + }; + std::vector<std::string> MissingLibs; + std::vector<std::string> RequiredLibs = ComputeLibsForComponents( + Components, + /*IncludeNonInstalled=*/IsInDevelopmentTree, false, + &GetComponentLibraryPathFunction, &MissingLibs, DirSep); + if (!MissingLibs.empty()) { + switch (LinkMode) { + case LinkModeShared: + if (DyLibExists && !BuiltSharedLibs) + break; + // Using component shared libraries. + for (auto &Lib : MissingLibs) + errs() << "llvm-config: error: missing: " << Lib << "\n"; + return 1; + case LinkModeAuto: + if (DyLibExists) { + LinkMode = LinkModeShared; + break; + } + errs() + << "llvm-config: error: component libraries and shared library\n\n"; + // fall through + case LinkModeStatic: + for (auto &Lib : MissingLibs) + errs() << "llvm-config: error: missing: " << Lib << "\n"; + return 1; + } + } else if (LinkMode == LinkModeAuto) { + LinkMode = LinkModeStatic; + } if (PrintSharedMode) { std::unordered_set<std::string> FullDyLibComponents; std::vector<std::string> DyLibComponents = - GetAllDyLibComponents(IsInDevelopmentTree, false); + GetAllDyLibComponents(IsInDevelopmentTree, false, DirSep); for (auto &Component : DyLibComponents) { FullDyLibComponents.insert(Component); @@ -549,7 +627,7 @@ int main(int argc, char **argv) { } FullDyLibComponents.clear(); - if (HasMissing && DyLibExists) { + if (LinkMode == LinkModeShared) { OS << "shared\n"; return 0; } else { @@ -560,36 +638,39 @@ int main(int argc, char **argv) { if (PrintLibs || PrintLibNames || PrintLibFiles) { - auto PrintForLib = [&](const StringRef &Lib, const bool ForceShared) { + auto PrintForLib = [&](const StringRef &Lib) { + const bool Shared = LinkMode == LinkModeShared; if (PrintLibNames) { - OS << GetComponentLibraryFileName(Lib, ForceShared); + OS << GetComponentLibraryFileName(Lib, Shared); } else if (PrintLibFiles) { - OS << GetComponentLibraryPath(Lib, ForceShared); + OS << GetComponentLibraryPath(Lib, Shared); } else if (PrintLibs) { - // If this is a typical library name, include it using -l. - StringRef LibName; - if (Lib.startswith("lib")) { + // On Windows, output full path to library without parameters. + // Elsewhere, if this is a typical library name, include it using -l. + if (HostTriple.isWindowsMSVCEnvironment()) { + OS << GetComponentLibraryPath(Lib, Shared); + } else { + StringRef LibName; if (GetComponentLibraryNameSlice(Lib, LibName)) { + // Extract library name (remove prefix and suffix). OS << "-l" << LibName; } else { - OS << "-l:" << GetComponentLibraryFileName(Lib, ForceShared); + // Lib is already a library name without prefix and suffix. + OS << "-l" << Lib; } - } else { - // Otherwise, print the full path. - OS << GetComponentLibraryPath(Lib, ForceShared); } } }; - if (HasMissing && DyLibExists) { - PrintForLib(DyLibName, true); + if (LinkMode == LinkModeShared && !BuiltSharedLibs) { + PrintForLib(DyLibName); } else { for (unsigned i = 0, e = RequiredLibs.size(); i != e; ++i) { auto Lib = RequiredLibs[i]; if (i) OS << ' '; - PrintForLib(Lib, false); + PrintForLib(Lib); } } OS << '\n'; diff --git a/tools/llvm-cov/CMakeLists.txt b/tools/llvm-cov/CMakeLists.txt index 193218a6639f2..e22828e11effe 100644 --- a/tools/llvm-cov/CMakeLists.txt +++ b/tools/llvm-cov/CMakeLists.txt @@ -1,4 +1,4 @@ -set(LLVM_LINK_COMPONENTS core support object profiledata) +set(LLVM_LINK_COMPONENTS core support object coverage profiledata) add_llvm_tool(llvm-cov llvm-cov.cpp @@ -8,5 +8,7 @@ add_llvm_tool(llvm-cov CoverageReport.cpp CoverageSummaryInfo.cpp SourceCoverageView.cpp + SourceCoverageViewHTML.cpp + SourceCoverageViewText.cpp TestingSupport.cpp ) diff --git a/tools/llvm-cov/CodeCoverage.cpp b/tools/llvm-cov/CodeCoverage.cpp index 8dc4d665f23ca..0a4d1a67d6105 100644 --- a/tools/llvm-cov/CodeCoverage.cpp +++ b/tools/llvm-cov/CodeCoverage.cpp @@ -13,24 +13,25 @@ // //===----------------------------------------------------------------------===// -#include "RenderingSupport.h" #include "CoverageFilters.h" #include "CoverageReport.h" #include "CoverageViewOptions.h" +#include "RenderingSupport.h" #include "SourceCoverageView.h" #include "llvm/ADT/SmallString.h" #include "llvm/ADT/StringRef.h" #include "llvm/ADT/Triple.h" -#include "llvm/ProfileData/CoverageMapping.h" +#include "llvm/ProfileData/Coverage/CoverageMapping.h" #include "llvm/ProfileData/InstrProfReader.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/Format.h" -#include "llvm/Support/ManagedStatic.h" +#include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/Path.h" -#include "llvm/Support/PrettyStackTrace.h" #include "llvm/Support/Process.h" -#include "llvm/Support/Signals.h" +#include "llvm/Support/Program.h" +#include "llvm/Support/ThreadPool.h" +#include "llvm/Support/ToolOutputFile.h" #include <functional> #include <system_error> @@ -51,28 +52,47 @@ public: /// \brief Print the error message to the error output stream. void error(const Twine &Message, StringRef Whence = ""); + /// \brief Record (but do not print) an error message in a thread-safe way. + void deferError(const Twine &Message, StringRef Whence = ""); + + /// \brief Record (but do not print) a warning message in a thread-safe way. + void deferWarning(const Twine &Message, StringRef Whence = ""); + + /// \brief Print (and then clear) all deferred error and warning messages. + void consumeDeferredMessages(); + + /// \brief Append a reference to a private copy of \p Path into SourceFiles. + void addCollectedPath(const std::string &Path); + /// \brief Return a memory buffer for the given source file. ErrorOr<const MemoryBuffer &> getSourceFile(StringRef SourceFile); /// \brief Create source views for the expansions of the view. void attachExpansionSubViews(SourceCoverageView &View, ArrayRef<ExpansionRecord> Expansions, - CoverageMapping &Coverage); + const CoverageMapping &Coverage); /// \brief Create the source view of a particular function. std::unique_ptr<SourceCoverageView> - createFunctionView(const FunctionRecord &Function, CoverageMapping &Coverage); + createFunctionView(const FunctionRecord &Function, + const CoverageMapping &Coverage); /// \brief Create the main source view of a particular source file. std::unique_ptr<SourceCoverageView> - createSourceFileView(StringRef SourceFile, CoverageMapping &Coverage); + createSourceFileView(StringRef SourceFile, const CoverageMapping &Coverage); - /// \brief Load the coverage mapping data. Return true if an error occured. + /// \brief Load the coverage mapping data. Return nullptr if an error occured. std::unique_ptr<CoverageMapping> load(); + /// \brief If a demangler is available, demangle all symbol names. + void demangleSymbols(const CoverageMapping &Coverage); + + /// \brief Demangle \p Sym if possible. Otherwise, just return \p Sym. + StringRef getSymbolForHumans(StringRef Sym) const; + int run(Command Cmd, int argc, const char **argv); - typedef std::function<int(int, const char **)> CommandLineParserType; + typedef llvm::function_ref<int(int, const char **)> CommandLineParserType; int show(int argc, const char **argv, CommandLineParserType commandLineParser); @@ -84,25 +104,69 @@ public: CoverageViewOptions ViewOpts; std::string PGOFilename; CoverageFiltersMatchAll Filters; - std::vector<std::string> SourceFiles; - std::vector<std::pair<std::string, std::unique_ptr<MemoryBuffer>>> - LoadedSourceFiles; + std::vector<StringRef> SourceFiles; bool CompareFilenamesOnly; StringMap<std::string> RemappedFilenames; std::string CoverageArch; + +private: + /// A cache for demangled symbol names. + StringMap<std::string> DemangledNames; + + /// File paths (absolute, or otherwise) to input source files. + std::vector<std::string> CollectedPaths; + + /// Errors and warnings which have not been printed. + std::mutex DeferredMessagesLock; + std::vector<std::string> DeferredMessages; + + /// A container for input source file buffers. + std::mutex LoadedSourceFilesLock; + std::vector<std::pair<std::string, std::unique_ptr<MemoryBuffer>>> + LoadedSourceFiles; }; } -void CodeCoverageTool::error(const Twine &Message, StringRef Whence) { - errs() << "error: "; +static std::string getErrorString(const Twine &Message, StringRef Whence, + bool Warning) { + std::string Str = (Warning ? "warning" : "error"); + Str += ": "; if (!Whence.empty()) - errs() << Whence << ": "; - errs() << Message << "\n"; + Str += Whence.str() + ": "; + Str += Message.str() + "\n"; + return Str; +} + +void CodeCoverageTool::error(const Twine &Message, StringRef Whence) { + errs() << getErrorString(Message, Whence, false); +} + +void CodeCoverageTool::deferError(const Twine &Message, StringRef Whence) { + std::unique_lock<std::mutex> Guard{DeferredMessagesLock}; + DeferredMessages.emplace_back(getErrorString(Message, Whence, false)); +} + +void CodeCoverageTool::deferWarning(const Twine &Message, StringRef Whence) { + std::unique_lock<std::mutex> Guard{DeferredMessagesLock}; + DeferredMessages.emplace_back(getErrorString(Message, Whence, true)); +} + +void CodeCoverageTool::consumeDeferredMessages() { + std::unique_lock<std::mutex> Guard{DeferredMessagesLock}; + for (const std::string &Message : DeferredMessages) + ViewOpts.colored_ostream(errs(), raw_ostream::RED) << Message; + DeferredMessages.clear(); +} + +void CodeCoverageTool::addCollectedPath(const std::string &Path) { + CollectedPaths.push_back(Path); + SourceFiles.emplace_back(CollectedPaths.back()); } ErrorOr<const MemoryBuffer &> CodeCoverageTool::getSourceFile(StringRef SourceFile) { // If we've remapped filenames, look up the real location for this file. + std::unique_lock<std::mutex> Guard{LoadedSourceFilesLock}; if (!RemappedFilenames.empty()) { auto Loc = RemappedFilenames.find(SourceFile); if (Loc != RemappedFilenames.end()) @@ -113,17 +177,16 @@ CodeCoverageTool::getSourceFile(StringRef SourceFile) { return *Files.second; auto Buffer = MemoryBuffer::getFile(SourceFile); if (auto EC = Buffer.getError()) { - error(EC.message(), SourceFile); + deferError(EC.message(), SourceFile); return EC; } LoadedSourceFiles.emplace_back(SourceFile, std::move(Buffer.get())); return *LoadedSourceFiles.back().second; } -void -CodeCoverageTool::attachExpansionSubViews(SourceCoverageView &View, - ArrayRef<ExpansionRecord> Expansions, - CoverageMapping &Coverage) { +void CodeCoverageTool::attachExpansionSubViews( + SourceCoverageView &View, ArrayRef<ExpansionRecord> Expansions, + const CoverageMapping &Coverage) { if (!ViewOpts.ShowExpandedRegions) return; for (const auto &Expansion : Expansions) { @@ -135,8 +198,9 @@ CodeCoverageTool::attachExpansionSubViews(SourceCoverageView &View, continue; auto SubViewExpansions = ExpansionCoverage.getExpansions(); - auto SubView = llvm::make_unique<SourceCoverageView>( - SourceBuffer.get(), ViewOpts, std::move(ExpansionCoverage)); + auto SubView = + SourceCoverageView::create(Expansion.Function.Name, SourceBuffer.get(), + ViewOpts, std::move(ExpansionCoverage)); attachExpansionSubViews(*SubView, SubViewExpansions, Coverage); View.addExpansion(Expansion.Region, std::move(SubView)); } @@ -144,7 +208,7 @@ CodeCoverageTool::attachExpansionSubViews(SourceCoverageView &View, std::unique_ptr<SourceCoverageView> CodeCoverageTool::createFunctionView(const FunctionRecord &Function, - CoverageMapping &Coverage) { + const CoverageMapping &Coverage) { auto FunctionCoverage = Coverage.getCoverageForFunction(Function); if (FunctionCoverage.empty()) return nullptr; @@ -153,8 +217,9 @@ CodeCoverageTool::createFunctionView(const FunctionRecord &Function, return nullptr; auto Expansions = FunctionCoverage.getExpansions(); - auto View = llvm::make_unique<SourceCoverageView>( - SourceBuffer.get(), ViewOpts, std::move(FunctionCoverage)); + auto View = SourceCoverageView::create(getSymbolForHumans(Function.Name), + SourceBuffer.get(), ViewOpts, + std::move(FunctionCoverage)); attachExpansionSubViews(*View, Expansions, Coverage); return View; @@ -162,7 +227,7 @@ CodeCoverageTool::createFunctionView(const FunctionRecord &Function, std::unique_ptr<SourceCoverageView> CodeCoverageTool::createSourceFileView(StringRef SourceFile, - CoverageMapping &Coverage) { + const CoverageMapping &Coverage) { auto SourceBuffer = getSourceFile(SourceFile); if (!SourceBuffer) return nullptr; @@ -171,15 +236,16 @@ CodeCoverageTool::createSourceFileView(StringRef SourceFile, return nullptr; auto Expansions = FileCoverage.getExpansions(); - auto View = llvm::make_unique<SourceCoverageView>( - SourceBuffer.get(), ViewOpts, std::move(FileCoverage)); + auto View = SourceCoverageView::create(SourceFile, SourceBuffer.get(), + ViewOpts, std::move(FileCoverage)); attachExpansionSubViews(*View, Expansions, Coverage); - for (auto Function : Coverage.getInstantiations(SourceFile)) { + for (const auto *Function : Coverage.getInstantiations(SourceFile)) { auto SubViewCoverage = Coverage.getCoverageForFunction(*Function); auto SubViewExpansions = SubViewCoverage.getExpansions(); - auto SubView = llvm::make_unique<SourceCoverageView>( - SourceBuffer.get(), ViewOpts, std::move(SubViewCoverage)); + auto SubView = SourceCoverageView::create( + getSymbolForHumans(Function->Name), SourceBuffer.get(), ViewOpts, + std::move(SubViewCoverage)); attachExpansionSubViews(*SubView, SubViewExpansions, Coverage); if (SubView) { @@ -210,10 +276,9 @@ std::unique_ptr<CoverageMapping> CodeCoverageTool::load() { errs() << "warning: profile data may be out of date - object is newer\n"; auto CoverageOrErr = CoverageMapping::load(ObjectFilename, PGOFilename, CoverageArch); - if (std::error_code EC = CoverageOrErr.getError()) { + if (Error E = CoverageOrErr.takeError()) { colored_ostream(errs(), raw_ostream::RED) - << "error: Failed to load coverage: " << EC.message(); - errs() << "\n"; + << "error: Failed to load coverage: " << toString(std::move(E)) << "\n"; return nullptr; } auto Coverage = std::move(CoverageOrErr.get()); @@ -237,15 +302,95 @@ std::unique_ptr<CoverageMapping> CodeCoverageTool::load() { } } + demangleSymbols(*Coverage); + return Coverage; } -int CodeCoverageTool::run(Command Cmd, int argc, const char **argv) { - // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); - PrettyStackTraceProgram X(argc, argv); - llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. +void CodeCoverageTool::demangleSymbols(const CoverageMapping &Coverage) { + if (!ViewOpts.hasDemangler()) + return; + + // Pass function names to the demangler in a temporary file. + int InputFD; + SmallString<256> InputPath; + std::error_code EC = + sys::fs::createTemporaryFile("demangle-in", "list", InputFD, InputPath); + if (EC) { + error(InputPath, EC.message()); + return; + } + tool_output_file InputTOF{InputPath, InputFD}; + + unsigned NumSymbols = 0; + for (const auto &Function : Coverage.getCoveredFunctions()) { + InputTOF.os() << Function.Name << '\n'; + ++NumSymbols; + } + InputTOF.os().close(); + + // Use another temporary file to store the demangler's output. + int OutputFD; + SmallString<256> OutputPath; + EC = sys::fs::createTemporaryFile("demangle-out", "list", OutputFD, + OutputPath); + if (EC) { + error(OutputPath, EC.message()); + return; + } + tool_output_file OutputTOF{OutputPath, OutputFD}; + OutputTOF.os().close(); + + // Invoke the demangler. + std::vector<const char *> ArgsV; + for (const std::string &Arg : ViewOpts.DemanglerOpts) + ArgsV.push_back(Arg.c_str()); + ArgsV.push_back(nullptr); + StringRef InputPathRef = InputPath.str(); + StringRef OutputPathRef = OutputPath.str(); + StringRef StderrRef; + const StringRef *Redirects[] = {&InputPathRef, &OutputPathRef, &StderrRef}; + std::string ErrMsg; + int RC = sys::ExecuteAndWait(ViewOpts.DemanglerOpts[0], ArgsV.data(), + /*env=*/nullptr, Redirects, /*secondsToWait=*/0, + /*memoryLimit=*/0, &ErrMsg); + if (RC) { + error(ErrMsg, ViewOpts.DemanglerOpts[0]); + return; + } + // Parse the demangler's output. + auto BufOrError = MemoryBuffer::getFile(OutputPath); + if (!BufOrError) { + error(OutputPath, BufOrError.getError().message()); + return; + } + + std::unique_ptr<MemoryBuffer> DemanglerBuf = std::move(*BufOrError); + + SmallVector<StringRef, 8> Symbols; + StringRef DemanglerData = DemanglerBuf->getBuffer(); + DemanglerData.split(Symbols, '\n', /*MaxSplit=*/NumSymbols, + /*KeepEmpty=*/false); + if (Symbols.size() != NumSymbols) { + error("Demangler did not provide expected number of symbols"); + return; + } + + // Cache the demangled names. + unsigned I = 0; + for (const auto &Function : Coverage.getCoveredFunctions()) + DemangledNames[Function.Name] = Symbols[I++]; +} + +StringRef CodeCoverageTool::getSymbolForHumans(StringRef Sym) const { + const auto DemangledName = DemangledNames.find(Sym); + if (DemangledName == DemangledNames.end()) + return Sym; + return DemangledName->getValue(); +} + +int CodeCoverageTool::run(Command Cmd, int argc, const char **argv) { cl::opt<std::string, true> ObjectFilename( cl::Positional, cl::Required, cl::location(this->ObjectFilename), cl::desc("Covered executable or object file.")); @@ -264,6 +409,15 @@ int CodeCoverageTool::run(Command Cmd, int argc, const char **argv) { cl::opt<bool> DebugDump("dump", cl::Optional, cl::desc("Show internal debug dump")); + cl::opt<CoverageViewOptions::OutputFormat> Format( + "format", cl::desc("Output format for line-based coverage reports"), + cl::values(clEnumValN(CoverageViewOptions::OutputFormat::Text, "text", + "Text output"), + clEnumValN(CoverageViewOptions::OutputFormat::HTML, "html", + "HTML output"), + clEnumValEnd), + cl::init(CoverageViewOptions::OutputFormat::Text)); + cl::opt<bool> FilenameEquivalence( "filename-equivalence", cl::Optional, cl::desc("Treat source files as equivalent to paths in the coverage data " @@ -310,14 +464,39 @@ int CodeCoverageTool::run(Command Cmd, int argc, const char **argv) { "use-color", cl::desc("Emit colored output (default=autodetect)"), cl::init(cl::BOU_UNSET)); + cl::list<std::string> DemanglerOpts( + "Xdemangler", cl::desc("<demangler-path>|<demangler-option>")); + auto commandLineParser = [&, this](int argc, const char **argv) -> int { cl::ParseCommandLineOptions(argc, argv, "LLVM code coverage tool\n"); ViewOpts.Debug = DebugDump; CompareFilenamesOnly = FilenameEquivalence; - ViewOpts.Colors = UseColor == cl::BOU_UNSET - ? sys::Process::StandardOutHasColors() - : UseColor == cl::BOU_TRUE; + ViewOpts.Format = Format; + switch (ViewOpts.Format) { + case CoverageViewOptions::OutputFormat::Text: + ViewOpts.Colors = UseColor == cl::BOU_UNSET + ? sys::Process::StandardOutHasColors() + : UseColor == cl::BOU_TRUE; + break; + case CoverageViewOptions::OutputFormat::HTML: + if (UseColor == cl::BOU_FALSE) + error("Color output cannot be disabled when generating html."); + ViewOpts.Colors = true; + break; + } + + // If a demangler is supplied, check if it exists and register it. + if (DemanglerOpts.size()) { + auto DemanglerPathOrErr = sys::findProgramByName(DemanglerOpts[0]); + if (!DemanglerPathOrErr) { + error("Could not find the demangler!", + DemanglerPathOrErr.getError().message()); + return 1; + } + DemanglerOpts[0] = *DemanglerPathOrErr; + ViewOpts.DemanglerOpts.swap(DemanglerOpts); + } // Create the function filters if (!NameFilters.empty() || !NameRegexFilters.empty()) { @@ -363,7 +542,7 @@ int CodeCoverageTool::run(Command Cmd, int argc, const char **argv) { errs() << "error: " << File << ": " << EC.message(); return 1; } - SourceFiles.push_back(Path.str()); + addCollectedPath(Path.str()); } return 0; }; @@ -406,6 +585,12 @@ int CodeCoverageTool::show(int argc, const char **argv, cl::desc("Show function instantiations"), cl::cat(ViewCategory)); + cl::opt<std::string> ShowOutputDirectory( + "output-dir", cl::init(""), + cl::desc("Directory in which coverage information is written out")); + cl::alias ShowOutputDirectoryA("o", cl::desc("Alias for --output-dir"), + cl::aliasopt(ShowOutputDirectory)); + auto Err = commandLineParser(argc, argv); if (Err) return Err; @@ -417,30 +602,46 @@ int CodeCoverageTool::show(int argc, const char **argv, ViewOpts.ShowLineStatsOrRegionMarkers = ShowBestLineRegionsCounts; ViewOpts.ShowExpandedRegions = ShowExpansions; ViewOpts.ShowFunctionInstantiations = ShowInstantiations; + ViewOpts.ShowOutputDirectory = ShowOutputDirectory; + + if (ViewOpts.hasOutputDirectory()) { + if (auto E = sys::fs::create_directories(ViewOpts.ShowOutputDirectory)) { + error("Could not create output directory!", E.message()); + return 1; + } + } auto Coverage = load(); if (!Coverage) return 1; + auto Printer = CoveragePrinter::create(ViewOpts); + if (!Filters.empty()) { - // Show functions + auto OSOrErr = Printer->createViewFile("functions", /*InToplevel=*/true); + if (Error E = OSOrErr.takeError()) { + error("Could not create view file!", toString(std::move(E))); + return 1; + } + auto OS = std::move(OSOrErr.get()); + + // Show functions. for (const auto &Function : Coverage->getCoveredFunctions()) { if (!Filters.matches(Function)) continue; auto mainView = createFunctionView(Function, *Coverage); if (!mainView) { - ViewOpts.colored_ostream(outs(), raw_ostream::RED) - << "warning: Could not read coverage for '" << Function.Name; - outs() << "\n"; + ViewOpts.colored_ostream(errs(), raw_ostream::RED) + << "warning: Could not read coverage for '" << Function.Name << "'." + << "\n"; continue; } - ViewOpts.colored_ostream(outs(), raw_ostream::CYAN) << Function.Name - << ":"; - outs() << "\n"; - mainView->render(outs(), /*WholeFile=*/false); - outs() << "\n"; + + mainView->print(*OS.get(), /*WholeFile=*/false, /*ShowSourceName=*/true); } + + Printer->closeViewFile(std::move(OS)); return 0; } @@ -448,28 +649,49 @@ int CodeCoverageTool::show(int argc, const char **argv, bool ShowFilenames = SourceFiles.size() != 1; if (SourceFiles.empty()) - // Get the source files from the function coverage mapping + // Get the source files from the function coverage mapping. for (StringRef Filename : Coverage->getUniqueSourceFiles()) SourceFiles.push_back(Filename); - for (const auto &SourceFile : SourceFiles) { - auto mainView = createSourceFileView(SourceFile, *Coverage); - if (!mainView) { - ViewOpts.colored_ostream(outs(), raw_ostream::RED) - << "warning: The file '" << SourceFile << "' isn't covered."; - outs() << "\n"; - continue; + // Create an index out of the source files. + if (ViewOpts.hasOutputDirectory()) { + if (Error E = Printer->createIndexFile(SourceFiles)) { + error("Could not create index file!", toString(std::move(E))); + return 1; } + } - if (ShowFilenames) { - ViewOpts.colored_ostream(outs(), raw_ostream::CYAN) << SourceFile << ":"; - outs() << "\n"; - } - mainView->render(outs(), /*Wholefile=*/true); - if (SourceFiles.size() > 1) - outs() << "\n"; + // In -output-dir mode, it's safe to use multiple threads to print files. + unsigned ThreadCount = 1; + if (ViewOpts.hasOutputDirectory()) + ThreadCount = std::thread::hardware_concurrency(); + ThreadPool Pool(ThreadCount); + + for (StringRef SourceFile : SourceFiles) { + Pool.async([this, SourceFile, &Coverage, &Printer, ShowFilenames] { + auto View = createSourceFileView(SourceFile, *Coverage); + if (!View) { + deferWarning("The file '" + SourceFile.str() + "' isn't covered."); + return; + } + + auto OSOrErr = Printer->createViewFile(SourceFile, /*InToplevel=*/false); + if (Error E = OSOrErr.takeError()) { + deferError("Could not create view file!", toString(std::move(E))); + return; + } + auto OS = std::move(OSOrErr.get()); + + View->print(*OS.get(), /*Wholefile=*/true, + /*ShowSourceName=*/ShowFilenames); + Printer->closeViewFile(std::move(OS)); + }); } + Pool.wait(); + + consumeDeferredMessages(); + return 0; } @@ -479,6 +701,9 @@ int CodeCoverageTool::report(int argc, const char **argv, if (Err) return Err; + if (ViewOpts.Format == CoverageViewOptions::OutputFormat::HTML) + error("HTML output for summary reports is not yet supported."); + auto Coverage = load(); if (!Coverage) return 1; diff --git a/tools/llvm-cov/CoverageFilters.h b/tools/llvm-cov/CoverageFilters.h index dc5dc98807b1a..756c4b47872c1 100644 --- a/tools/llvm-cov/CoverageFilters.h +++ b/tools/llvm-cov/CoverageFilters.h @@ -14,7 +14,7 @@ #ifndef LLVM_COV_COVERAGEFILTERS_H #define LLVM_COV_COVERAGEFILTERS_H -#include "llvm/ProfileData/CoverageMapping.h" +#include "llvm/ProfileData/Coverage/CoverageMapping.h" #include <memory> #include <vector> diff --git a/tools/llvm-cov/CoverageReport.cpp b/tools/llvm-cov/CoverageReport.cpp index ed01a2e154f1b..10e53b3f1f723 100644 --- a/tools/llvm-cov/CoverageReport.cpp +++ b/tools/llvm-cov/CoverageReport.cpp @@ -171,7 +171,7 @@ void CoverageReport::render(const FunctionCoverageSummary &Function, OS << "\n"; } -void CoverageReport::renderFunctionReports(ArrayRef<std::string> Files, +void CoverageReport::renderFunctionReports(ArrayRef<StringRef> Files, raw_ostream &OS) { adjustColumnWidths(Coverage.get()); bool isFirst = true; diff --git a/tools/llvm-cov/CoverageReport.h b/tools/llvm-cov/CoverageReport.h index 7ec3df90b8f92..bb3d734b52a54 100644 --- a/tools/llvm-cov/CoverageReport.h +++ b/tools/llvm-cov/CoverageReport.h @@ -32,7 +32,7 @@ public: std::unique_ptr<coverage::CoverageMapping> Coverage) : Options(Options), Coverage(std::move(Coverage)) {} - void renderFunctionReports(ArrayRef<std::string> Files, raw_ostream &OS); + void renderFunctionReports(ArrayRef<StringRef> Files, raw_ostream &OS); void renderFileReports(raw_ostream &OS); }; diff --git a/tools/llvm-cov/CoverageSummaryInfo.h b/tools/llvm-cov/CoverageSummaryInfo.h index c393b00d32a43..822742b635e96 100644 --- a/tools/llvm-cov/CoverageSummaryInfo.h +++ b/tools/llvm-cov/CoverageSummaryInfo.h @@ -15,7 +15,7 @@ #ifndef LLVM_COV_COVERAGESUMMARYINFO_H #define LLVM_COV_COVERAGESUMMARYINFO_H -#include "llvm/ProfileData/CoverageMapping.h" +#include "llvm/ProfileData/Coverage/CoverageMapping.h" #include "llvm/Support/raw_ostream.h" namespace llvm { @@ -47,6 +47,8 @@ struct RegionCoverageInfo { bool isFullyCovered() const { return Covered == NumRegions; } double getPercentCovered() const { + if (NumRegions == 0) + return 0.0; return double(Covered) / double(NumRegions) * 100.0; } }; @@ -83,6 +85,8 @@ struct LineCoverageInfo { bool isFullyCovered() const { return Covered == (NumLines - NonCodeLines); } double getPercentCovered() const { + if (NumLines - NonCodeLines == 0) + return 0.0; return double(Covered) / double(NumLines - NonCodeLines) * 100.0; } }; @@ -109,6 +113,8 @@ struct FunctionCoverageInfo { bool isFullyCovered() const { return Executed == NumFunctions; } double getPercentCovered() const { + if (NumFunctions == 0) + return 0.0; return double(Executed) / double(NumFunctions) * 100.0; } }; diff --git a/tools/llvm-cov/CoverageViewOptions.h b/tools/llvm-cov/CoverageViewOptions.h index 1208fad791766..350c264a28764 100644 --- a/tools/llvm-cov/CoverageViewOptions.h +++ b/tools/llvm-cov/CoverageViewOptions.h @@ -11,11 +11,17 @@ #define LLVM_COV_COVERAGEVIEWOPTIONS_H #include "RenderingSupport.h" +#include <vector> namespace llvm { /// \brief The options for displaying the code coverage information. struct CoverageViewOptions { + enum class OutputFormat { + Text, + HTML + }; + bool Debug; bool Colors; bool ShowLineNumbers; @@ -25,12 +31,21 @@ struct CoverageViewOptions { bool ShowExpandedRegions; bool ShowFunctionInstantiations; bool ShowFullFilenames; + OutputFormat Format; + std::string ShowOutputDirectory; + std::vector<std::string> DemanglerOpts; /// \brief Change the output's stream color if the colors are enabled. ColoredRawOstream colored_ostream(raw_ostream &OS, raw_ostream::Colors Color) const { return llvm::colored_ostream(OS, Color, Colors); } + + /// \brief Check if an output directory has been specified. + bool hasOutputDirectory() const { return !ShowOutputDirectory.empty(); } + + /// \brief Check if a demangler has been specified. + bool hasDemangler() const { return !DemanglerOpts.empty(); } }; } diff --git a/tools/llvm-cov/LLVMBuild.txt b/tools/llvm-cov/LLVMBuild.txt index d6eb74de0d4bf..33f51fb2ed5a9 100644 --- a/tools/llvm-cov/LLVMBuild.txt +++ b/tools/llvm-cov/LLVMBuild.txt @@ -19,4 +19,4 @@ type = Tool name = llvm-cov parent = Tools -required_libraries = ProfileData Support Instrumentation +required_libraries = Coverage Support Instrumentation diff --git a/tools/llvm-cov/Makefile b/tools/llvm-cov/Makefile deleted file mode 100644 index 6e32b4d233d7e..0000000000000 --- a/tools/llvm-cov/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-cov/Makefile -----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-cov -LINK_COMPONENTS := core support profiledata object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-cov/RenderingSupport.h b/tools/llvm-cov/RenderingSupport.h index 3ef155d3fd4bb..aa70fbc23e3c0 100644 --- a/tools/llvm-cov/RenderingSupport.h +++ b/tools/llvm-cov/RenderingSupport.h @@ -55,6 +55,7 @@ inline ColoredRawOstream colored_ostream(raw_ostream &OS, OS.changeColor(Color, Bold, BG); return ColoredRawOstream(OS, IsColorUsed); } -} + +} // namespace llvm #endif // LLVM_COV_RENDERINGSUPPORT_H diff --git a/tools/llvm-cov/SourceCoverageView.cpp b/tools/llvm-cov/SourceCoverageView.cpp index 58c8a67952946..baf7c148bb800 100644 --- a/tools/llvm-cov/SourceCoverageView.cpp +++ b/tools/llvm-cov/SourceCoverageView.cpp @@ -6,80 +6,82 @@ // License. See LICENSE.TXT for details. // //===----------------------------------------------------------------------===// -// -// This class implements rendering for code coverage of source code. -// +/// +/// \file This class implements rendering for code coverage of source code. +/// //===----------------------------------------------------------------------===// #include "SourceCoverageView.h" -#include "llvm/ADT/Optional.h" +#include "SourceCoverageViewHTML.h" +#include "SourceCoverageViewText.h" #include "llvm/ADT/SmallString.h" #include "llvm/ADT/StringExtras.h" +#include "llvm/Support/FileSystem.h" #include "llvm/Support/LineIterator.h" +#include "llvm/Support/Path.h" using namespace llvm; -void SourceCoverageView::renderLine( - raw_ostream &OS, StringRef Line, int64_t LineNumber, - const coverage::CoverageSegment *WrappedSegment, - ArrayRef<const coverage::CoverageSegment *> Segments, - unsigned ExpansionCol) { - Optional<raw_ostream::Colors> Highlight; - SmallVector<std::pair<unsigned, unsigned>, 2> HighlightedRanges; - - // The first segment overlaps from a previous line, so we treat it specially. - if (WrappedSegment && WrappedSegment->HasCount && WrappedSegment->Count == 0) - Highlight = raw_ostream::RED; - - // Output each segment of the line, possibly highlighted. - unsigned Col = 1; - for (const auto *S : Segments) { - unsigned End = std::min(S->Col, static_cast<unsigned>(Line.size()) + 1); - colored_ostream(OS, Highlight ? *Highlight : raw_ostream::SAVEDCOLOR, - Options.Colors && Highlight, /*Bold=*/false, /*BG=*/true) - << Line.substr(Col - 1, End - Col); - if (Options.Debug && Highlight) - HighlightedRanges.push_back(std::make_pair(Col, End)); - Col = End; - if (Col == ExpansionCol) - Highlight = raw_ostream::CYAN; - else if (S->HasCount && S->Count == 0) - Highlight = raw_ostream::RED; - else - Highlight = None; - } +void CoveragePrinter::StreamDestructor::operator()(raw_ostream *OS) const { + if (OS == &outs()) + return; + delete OS; +} - // Show the rest of the line - colored_ostream(OS, Highlight ? *Highlight : raw_ostream::SAVEDCOLOR, - Options.Colors && Highlight, /*Bold=*/false, /*BG=*/true) - << Line.substr(Col - 1, Line.size() - Col + 1); - OS << "\n"; - - if (Options.Debug) { - for (const auto &Range : HighlightedRanges) - errs() << "Highlighted line " << LineNumber << ", " << Range.first - << " -> " << Range.second << "\n"; - if (Highlight) - errs() << "Highlighted line " << LineNumber << ", " << Col << " -> ?\n"; - } +std::string CoveragePrinter::getOutputPath(StringRef Path, StringRef Extension, + bool InToplevel, bool Relative) { + assert(Extension.size() && "The file extension may not be empty"); + + SmallString<256> FullPath; + + if (!Relative) + FullPath.append(Opts.ShowOutputDirectory); + + if (!InToplevel) + sys::path::append(FullPath, getCoverageDir()); + + SmallString<256> ParentPath = sys::path::parent_path(Path); + sys::path::remove_dots(ParentPath, /*remove_dot_dots=*/true); + sys::path::append(FullPath, sys::path::relative_path(ParentPath)); + + auto PathFilename = (sys::path::filename(Path) + "." + Extension).str(); + sys::path::append(FullPath, PathFilename); + + return FullPath.str(); } -void SourceCoverageView::renderIndent(raw_ostream &OS, unsigned Level) { - for (unsigned I = 0; I < Level; ++I) - OS << " |"; +Expected<CoveragePrinter::OwnedStream> +CoveragePrinter::createOutputStream(StringRef Path, StringRef Extension, + bool InToplevel) { + if (!Opts.hasOutputDirectory()) + return OwnedStream(&outs()); + + std::string FullPath = getOutputPath(Path, Extension, InToplevel, false); + + auto ParentDir = sys::path::parent_path(FullPath); + if (auto E = sys::fs::create_directories(ParentDir)) + return errorCodeToError(E); + + std::error_code E; + raw_ostream *RawStream = new raw_fd_ostream(FullPath, E, sys::fs::F_RW); + auto OS = CoveragePrinter::OwnedStream(RawStream); + if (E) + return errorCodeToError(E); + return std::move(OS); } -void SourceCoverageView::renderViewDivider(unsigned Level, unsigned Length, - raw_ostream &OS) { - assert(Level != 0 && "Cannot render divider at top level"); - renderIndent(OS, Level - 1); - OS.indent(2); - for (unsigned I = 0; I < Length; ++I) - OS << "-"; +std::unique_ptr<CoveragePrinter> +CoveragePrinter::create(const CoverageViewOptions &Opts) { + switch (Opts.Format) { + case CoverageViewOptions::OutputFormat::Text: + return llvm::make_unique<CoveragePrinterText>(Opts); + case CoverageViewOptions::OutputFormat::HTML: + return llvm::make_unique<CoveragePrinterHTML>(Opts); + } + llvm_unreachable("Unknown coverage output format!"); } -/// Format a count using engineering notation with 3 significant digits. -static std::string formatCount(uint64_t N) { +std::string SourceCoverageView::formatCount(uint64_t N) { std::string Number = utostr(N); int Len = Number.size(); if (Len <= 3) @@ -94,63 +96,49 @@ static std::string formatCount(uint64_t N) { return Result; } -void -SourceCoverageView::renderLineCoverageColumn(raw_ostream &OS, - const LineCoverageInfo &Line) { - if (!Line.isMapped()) { - OS.indent(LineCoverageColumnWidth) << '|'; - return; - } - std::string C = formatCount(Line.ExecutionCount); - OS.indent(LineCoverageColumnWidth - C.size()); - colored_ostream(OS, raw_ostream::MAGENTA, - Line.hasMultipleRegions() && Options.Colors) - << C; - OS << '|'; +bool SourceCoverageView::shouldRenderRegionMarkers( + bool LineHasMultipleRegions) const { + return getOptions().ShowRegionMarkers && + (!getOptions().ShowLineStatsOrRegionMarkers || LineHasMultipleRegions); } -void SourceCoverageView::renderLineNumberColumn(raw_ostream &OS, - unsigned LineNo) { - SmallString<32> Buffer; - raw_svector_ostream BufferOS(Buffer); - BufferOS << LineNo; - auto Str = BufferOS.str(); - // Trim and align to the right - Str = Str.substr(0, std::min(Str.size(), (size_t)LineNumberColumnWidth)); - OS.indent(LineNumberColumnWidth - Str.size()) << Str << '|'; +bool SourceCoverageView::hasSubViews() const { + return !ExpansionSubViews.empty() || !InstantiationSubViews.empty(); } -void SourceCoverageView::renderRegionMarkers( - raw_ostream &OS, ArrayRef<const coverage::CoverageSegment *> Segments) { - unsigned PrevColumn = 1; - for (const auto *S : Segments) { - if (!S->IsRegionEntry) - continue; - // Skip to the new region - if (S->Col > PrevColumn) - OS.indent(S->Col - PrevColumn); - PrevColumn = S->Col + 1; - std::string C = formatCount(S->Count); - PrevColumn += C.size(); - OS << '^' << C; +std::unique_ptr<SourceCoverageView> +SourceCoverageView::create(StringRef SourceName, const MemoryBuffer &File, + const CoverageViewOptions &Options, + coverage::CoverageData &&CoverageInfo) { + switch (Options.Format) { + case CoverageViewOptions::OutputFormat::Text: + return llvm::make_unique<SourceCoverageViewText>(SourceName, File, Options, + std::move(CoverageInfo)); + case CoverageViewOptions::OutputFormat::HTML: + return llvm::make_unique<SourceCoverageViewHTML>(SourceName, File, Options, + std::move(CoverageInfo)); } - OS << "\n"; + llvm_unreachable("Unknown coverage output format!"); +} + +void SourceCoverageView::addExpansion( + const coverage::CounterMappingRegion &Region, + std::unique_ptr<SourceCoverageView> View) { + ExpansionSubViews.emplace_back(Region, std::move(View)); +} - if (Options.Debug) - for (const auto *S : Segments) - errs() << "Marker at " << S->Line << ":" << S->Col << " = " - << formatCount(S->Count) << (S->IsRegionEntry ? "\n" : " (pop)\n"); +void SourceCoverageView::addInstantiation( + StringRef FunctionName, unsigned Line, + std::unique_ptr<SourceCoverageView> View) { + InstantiationSubViews.emplace_back(FunctionName, Line, std::move(View)); } -void SourceCoverageView::render(raw_ostream &OS, bool WholeFile, - unsigned IndentLevel) { - // The width of the leading columns - unsigned CombinedColumnWidth = - (Options.ShowLineStats ? LineCoverageColumnWidth + 1 : 0) + - (Options.ShowLineNumbers ? LineNumberColumnWidth + 1 : 0); - // The width of the line that is used to divide between the view and the - // subviews. - unsigned DividerWidth = CombinedColumnWidth + 4; +void SourceCoverageView::print(raw_ostream &OS, bool WholeFile, + bool ShowSourceName, unsigned ViewDepth) { + if (ShowSourceName) + renderSourceName(OS); + + renderViewHeader(OS); // We need the expansions and instantiations sorted so we can go through them // while we iterate lines. @@ -186,79 +174,60 @@ void SourceCoverageView::render(raw_ostream &OS, bool WholeFile, LineSegments.push_back(&*NextSegment++); // Calculate a count to be for the line as a whole. - LineCoverageInfo LineCount; + LineCoverageStats LineCount; if (WrappedSegment && WrappedSegment->HasCount) LineCount.addRegionCount(WrappedSegment->Count); for (const auto *S : LineSegments) if (S->HasCount && S->IsRegionEntry) - LineCount.addRegionStartCount(S->Count); + LineCount.addRegionStartCount(S->Count); - // Render the line prefix. - renderIndent(OS, IndentLevel); - if (Options.ShowLineStats) + renderLinePrefix(OS, ViewDepth); + if (getOptions().ShowLineStats) renderLineCoverageColumn(OS, LineCount); - if (Options.ShowLineNumbers) + if (getOptions().ShowLineNumbers) renderLineNumberColumn(OS, LI.line_number()); // If there are expansion subviews, we want to highlight the first one. unsigned ExpansionColumn = 0; if (NextESV != EndESV && NextESV->getLine() == LI.line_number() && - Options.Colors) + getOptions().Colors) ExpansionColumn = NextESV->getStartCol(); // Display the source code for the current line. - renderLine(OS, *LI, LI.line_number(), WrappedSegment, LineSegments, - ExpansionColumn); + renderLine(OS, {*LI, LI.line_number()}, WrappedSegment, LineSegments, + ExpansionColumn, ViewDepth); // Show the region markers. - if (Options.ShowRegionMarkers && (!Options.ShowLineStatsOrRegionMarkers || - LineCount.hasMultipleRegions()) && - !LineSegments.empty()) { - renderIndent(OS, IndentLevel); - OS.indent(CombinedColumnWidth); - renderRegionMarkers(OS, LineSegments); - } + if (shouldRenderRegionMarkers(LineCount.hasMultipleRegions())) + renderRegionMarkers(OS, LineSegments, ViewDepth); // Show the expansions and instantiations for this line. - unsigned NestedIndent = IndentLevel + 1; bool RenderedSubView = false; for (; NextESV != EndESV && NextESV->getLine() == LI.line_number(); ++NextESV) { - renderViewDivider(NestedIndent, DividerWidth, OS); - OS << "\n"; + renderViewDivider(OS, ViewDepth + 1); + + // Re-render the current line and highlight the expansion range for + // this subview. if (RenderedSubView) { - // Re-render the current line and highlight the expansion range for - // this subview. ExpansionColumn = NextESV->getStartCol(); - renderIndent(OS, IndentLevel); - OS.indent(CombinedColumnWidth + (IndentLevel == 0 ? 0 : 1)); - renderLine(OS, *LI, LI.line_number(), WrappedSegment, LineSegments, - ExpansionColumn); - renderViewDivider(NestedIndent, DividerWidth, OS); - OS << "\n"; + renderExpansionSite(OS, {*LI, LI.line_number()}, WrappedSegment, + LineSegments, ExpansionColumn, ViewDepth); + renderViewDivider(OS, ViewDepth + 1); } - // Render the child subview - if (Options.Debug) - errs() << "Expansion at line " << NextESV->getLine() << ", " - << NextESV->getStartCol() << " -> " << NextESV->getEndCol() - << "\n"; - NextESV->View->render(OS, false, NestedIndent); + + renderExpansionView(OS, *NextESV, ViewDepth + 1); RenderedSubView = true; } for (; NextISV != EndISV && NextISV->Line == LI.line_number(); ++NextISV) { - renderViewDivider(NestedIndent, DividerWidth, OS); - OS << "\n"; - renderIndent(OS, NestedIndent); - OS << ' '; - Options.colored_ostream(OS, raw_ostream::CYAN) << NextISV->FunctionName - << ":"; - OS << "\n"; - NextISV->View->render(OS, false, NestedIndent); + renderViewDivider(OS, ViewDepth + 1); + renderInstantiationView(OS, *NextISV, ViewDepth + 1); RenderedSubView = true; } - if (RenderedSubView) { - renderViewDivider(NestedIndent, DividerWidth, OS); - OS << "\n"; - } + if (RenderedSubView) + renderViewDivider(OS, ViewDepth + 1); + renderLineSuffix(OS, ViewDepth); } + + renderViewFooter(OS); } diff --git a/tools/llvm-cov/SourceCoverageView.h b/tools/llvm-cov/SourceCoverageView.h index 9e6fe5f350012..feef959f86bf1 100644 --- a/tools/llvm-cov/SourceCoverageView.h +++ b/tools/llvm-cov/SourceCoverageView.h @@ -6,16 +6,16 @@ // License. See LICENSE.TXT for details. // //===----------------------------------------------------------------------===// -// -// This class implements rendering for code coverage of source code. -// +/// +/// \file This class implements rendering for code coverage of source code. +/// //===----------------------------------------------------------------------===// #ifndef LLVM_COV_SOURCECOVERAGEVIEW_H #define LLVM_COV_SOURCECOVERAGEVIEW_H #include "CoverageViewOptions.h" -#include "llvm/ProfileData/CoverageMapping.h" +#include "llvm/ProfileData/Coverage/CoverageMapping.h" #include "llvm/Support/MemoryBuffer.h" #include <vector> @@ -23,7 +23,7 @@ namespace llvm { class SourceCoverageView; -/// \brief A view that represents a macro or include expansion +/// \brief A view that represents a macro or include expansion. struct ExpansionView { coverage::CounterMappingRegion Region; std::unique_ptr<SourceCoverageView> View; @@ -48,7 +48,7 @@ struct ExpansionView { } }; -/// \brief A view that represents a function instantiation +/// \brief A view that represents a function instantiation. struct InstantiationView { StringRef FunctionName; unsigned Line; @@ -73,87 +73,211 @@ struct InstantiationView { } }; -/// \brief A code coverage view of a specific source file. -/// It can have embedded coverage views. -class SourceCoverageView { -private: - /// \brief Coverage information for a single line. - struct LineCoverageInfo { - uint64_t ExecutionCount; - unsigned RegionCount; - bool Mapped; +/// \brief Coverage statistics for a single line. +struct LineCoverageStats { + uint64_t ExecutionCount; + unsigned RegionCount; + bool Mapped; + + LineCoverageStats() : ExecutionCount(0), RegionCount(0), Mapped(false) {} - LineCoverageInfo() : ExecutionCount(0), RegionCount(0), Mapped(false) {} + bool isMapped() const { return Mapped; } - bool isMapped() const { return Mapped; } + bool hasMultipleRegions() const { return RegionCount > 1; } - bool hasMultipleRegions() const { return RegionCount > 1; } + void addRegionStartCount(uint64_t Count) { + // The max of all region starts is the most interesting value. + addRegionCount(RegionCount ? std::max(ExecutionCount, Count) : Count); + ++RegionCount; + } + + void addRegionCount(uint64_t Count) { + Mapped = true; + ExecutionCount = Count; + } +}; - void addRegionStartCount(uint64_t Count) { - // The max of all region starts is the most interesting value. - addRegionCount(RegionCount ? std::max(ExecutionCount, Count) : Count); - ++RegionCount; - } +/// \brief A file manager that handles format-aware file creation. +class CoveragePrinter { + const CoverageViewOptions &Opts; - void addRegionCount(uint64_t Count) { - Mapped = true; - ExecutionCount = Count; - } +public: + struct StreamDestructor { + void operator()(raw_ostream *OS) const; }; + using OwnedStream = std::unique_ptr<raw_ostream, StreamDestructor>; + +protected: + CoveragePrinter(const CoverageViewOptions &Opts) : Opts(Opts) {} + + /// \brief Return `OutputDir/ToplevelDir/Path.Extension`. If \p InToplevel is + /// false, skip the ToplevelDir component. If \p Relative is false, skip the + /// OutputDir component. + std::string getOutputPath(StringRef Path, StringRef Extension, + bool InToplevel, bool Relative = true); + + /// \brief If directory output is enabled, create a file in that directory + /// at the path given by getOutputPath(). Otherwise, return stdout. + Expected<OwnedStream> createOutputStream(StringRef Path, StringRef Extension, + bool InToplevel); + + /// \brief Return the sub-directory name for file coverage reports. + static StringRef getCoverageDir() { return "coverage"; } + +public: + static std::unique_ptr<CoveragePrinter> + create(const CoverageViewOptions &Opts); + + virtual ~CoveragePrinter() {} + + /// @name File Creation Interface + /// @{ + + /// \brief Create a file to print a coverage view into. + virtual Expected<OwnedStream> createViewFile(StringRef Path, + bool InToplevel) = 0; + + /// \brief Close a file which has been used to print a coverage view. + virtual void closeViewFile(OwnedStream OS) = 0; + + /// \brief Create an index which lists reports for the given source files. + virtual Error createIndexFile(ArrayRef<StringRef> SourceFiles) = 0; + + /// @} +}; + +/// \brief A code coverage view of a source file or function. +/// +/// A source coverage view and its nested sub-views form a file-oriented +/// representation of code coverage data. This view can be printed out by a +/// renderer which implements the Rendering Interface. +class SourceCoverageView { + /// A function or file name. + StringRef SourceName; + + /// A memory buffer backing the source on display. const MemoryBuffer &File; + + /// Various options to guide the coverage renderer. const CoverageViewOptions &Options; + + /// Complete coverage information about the source on display. coverage::CoverageData CoverageInfo; + + /// A container for all expansions (e.g macros) in the source on display. std::vector<ExpansionView> ExpansionSubViews; + + /// A container for all instantiations (e.g template functions) in the source + /// on display. std::vector<InstantiationView> InstantiationSubViews; - /// \brief Render a source line with highlighting. - void renderLine(raw_ostream &OS, StringRef Line, int64_t LineNumber, - const coverage::CoverageSegment *WrappedSegment, - ArrayRef<const coverage::CoverageSegment *> Segments, - unsigned ExpansionCol); +protected: + struct LineRef { + StringRef Line; + int64_t LineNo; + + LineRef(StringRef Line, int64_t LineNo) : Line(Line), LineNo(LineNo) {} + }; + + using CoverageSegmentArray = ArrayRef<const coverage::CoverageSegment *>; + + /// @name Rendering Interface + /// @{ - void renderIndent(raw_ostream &OS, unsigned Level); + /// \brief Render a header for the view. + virtual void renderViewHeader(raw_ostream &OS) = 0; - void renderViewDivider(unsigned Offset, unsigned Length, raw_ostream &OS); + /// \brief Render a footer for the view. + virtual void renderViewFooter(raw_ostream &OS) = 0; + + /// \brief Render the source name for the view. + virtual void renderSourceName(raw_ostream &OS) = 0; + + /// \brief Render the line prefix at the given \p ViewDepth. + virtual void renderLinePrefix(raw_ostream &OS, unsigned ViewDepth) = 0; + + /// \brief Render the line suffix at the given \p ViewDepth. + virtual void renderLineSuffix(raw_ostream &OS, unsigned ViewDepth) = 0; + + /// \brief Render a view divider at the given \p ViewDepth. + virtual void renderViewDivider(raw_ostream &OS, unsigned ViewDepth) = 0; + + /// \brief Render a source line with highlighting. + virtual void renderLine(raw_ostream &OS, LineRef L, + const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, + unsigned ViewDepth) = 0; /// \brief Render the line's execution count column. - void renderLineCoverageColumn(raw_ostream &OS, const LineCoverageInfo &Line); + virtual void renderLineCoverageColumn(raw_ostream &OS, + const LineCoverageStats &Line) = 0; /// \brief Render the line number column. - void renderLineNumberColumn(raw_ostream &OS, unsigned LineNo); + virtual void renderLineNumberColumn(raw_ostream &OS, unsigned LineNo) = 0; /// \brief Render all the region's execution counts on a line. - void - renderRegionMarkers(raw_ostream &OS, - ArrayRef<const coverage::CoverageSegment *> Segments); + virtual void renderRegionMarkers(raw_ostream &OS, + CoverageSegmentArray Segments, + unsigned ViewDepth) = 0; - static const unsigned LineCoverageColumnWidth = 7; - static const unsigned LineNumberColumnWidth = 5; + /// \brief Render the site of an expansion. + virtual void + renderExpansionSite(raw_ostream &OS, LineRef L, + const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, + unsigned ViewDepth) = 0; -public: - SourceCoverageView(const MemoryBuffer &File, + /// \brief Render an expansion view and any nested views. + virtual void renderExpansionView(raw_ostream &OS, ExpansionView &ESV, + unsigned ViewDepth) = 0; + + /// \brief Render an instantiation view and any nested views. + virtual void renderInstantiationView(raw_ostream &OS, InstantiationView &ISV, + unsigned ViewDepth) = 0; + + /// @} + + /// \brief Format a count using engineering notation with 3 significant + /// digits. + static std::string formatCount(uint64_t N); + + /// \brief Check if region marker output is expected for a line. + bool shouldRenderRegionMarkers(bool LineHasMultipleRegions) const; + + /// \brief Check if there are any sub-views attached to this view. + bool hasSubViews() const; + + SourceCoverageView(StringRef SourceName, const MemoryBuffer &File, const CoverageViewOptions &Options, coverage::CoverageData &&CoverageInfo) - : File(File), Options(Options), CoverageInfo(std::move(CoverageInfo)) {} + : SourceName(SourceName), File(File), Options(Options), + CoverageInfo(std::move(CoverageInfo)) {} + +public: + static std::unique_ptr<SourceCoverageView> + create(StringRef SourceName, const MemoryBuffer &File, + const CoverageViewOptions &Options, + coverage::CoverageData &&CoverageInfo); + + virtual ~SourceCoverageView() {} + + StringRef getSourceName() const { return SourceName; } const CoverageViewOptions &getOptions() const { return Options; } /// \brief Add an expansion subview to this view. void addExpansion(const coverage::CounterMappingRegion &Region, - std::unique_ptr<SourceCoverageView> View) { - ExpansionSubViews.emplace_back(Region, std::move(View)); - } + std::unique_ptr<SourceCoverageView> View); /// \brief Add a function instantiation subview to this view. void addInstantiation(StringRef FunctionName, unsigned Line, - std::unique_ptr<SourceCoverageView> View) { - InstantiationSubViews.emplace_back(FunctionName, Line, std::move(View)); - } + std::unique_ptr<SourceCoverageView> View); - /// \brief Print the code coverage information for a specific - /// portion of a source file to the output stream. - void render(raw_ostream &OS, bool WholeFile, unsigned IndentLevel = 0); + /// \brief Print the code coverage information for a specific portion of a + /// source file to the output stream. + void print(raw_ostream &OS, bool WholeFile, bool ShowSourceName, + unsigned ViewDepth = 0); }; } // namespace llvm diff --git a/tools/llvm-cov/SourceCoverageViewHTML.cpp b/tools/llvm-cov/SourceCoverageViewHTML.cpp new file mode 100644 index 0000000000000..81963e5c54474 --- /dev/null +++ b/tools/llvm-cov/SourceCoverageViewHTML.cpp @@ -0,0 +1,436 @@ +//===- SourceCoverageViewHTML.cpp - A html code coverage view -------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +/// +/// \file This file implements the html coverage renderer. +/// +//===----------------------------------------------------------------------===// + +#include "SourceCoverageViewHTML.h" +#include "llvm/ADT/Optional.h" +#include "llvm/ADT/SmallString.h" +#include "llvm/ADT/StringExtras.h" +#include "llvm/Support/Path.h" + +using namespace llvm; + +namespace { + +const char *BeginHeader = + "<head>" + "<meta name='viewport' content='width=device-width,initial-scale=1'>" + "<meta charset='UTF-8'>"; + +const char *CSSForCoverage = + "<style>" +R"( + +.red { + background-color: #FFD0D0; +} +.cyan { + background-color: cyan; +} +.black { + background-color: black; + color: white; +} +.green { + background-color: #98FFA6; + color: white; +} +.magenta { + background-color: #F998FF; + color: white; +} +body { + font-family: -apple-system, sans-serif; +} +pre { + margin-top: 0px !important; + margin-bottom: 0px !important; +} +.source-name-title { + padding: 5px 10px; + border-bottom: 1px solid #dbdbdb; + background-color: #eee; +} +.centered { + display: table; + margin-left: auto; + margin-right: auto; + border: 1px solid #dbdbdb; + border-radius: 3px; +} +.expansion-view { + background-color: rgba(0, 0, 0, 0); + margin-left: 0px; + margin-top: 5px; + margin-right: 5px; + margin-bottom: 5px; + border: 1px solid #dbdbdb; + border-radius: 3px; +} +table { + border-collapse: collapse; +} +.line-number { + text-align: right; + color: #aaa; +} +.covered-line { + text-align: right; + color: #0080ff; +} +.uncovered-line { + text-align: right; + color: #ff3300; +} +.tooltip { + position: relative; + display: inline; + background-color: #b3e6ff; + text-decoration: none; +} +.tooltip span.tooltip-content { + position: absolute; + width: 100px; + margin-left: -50px; + color: #FFFFFF; + background: #000000; + height: 30px; + line-height: 30px; + text-align: center; + visibility: hidden; + border-radius: 6px; +} +.tooltip span.tooltip-content:after { + content: ''; + position: absolute; + top: 100%; + left: 50%; + margin-left: -8px; + width: 0; height: 0; + border-top: 8px solid #000000; + border-right: 8px solid transparent; + border-left: 8px solid transparent; +} +:hover.tooltip span.tooltip-content { + visibility: visible; + opacity: 0.8; + bottom: 30px; + left: 50%; + z-index: 999; +} +th, td { + vertical-align: top; + padding: 2px 5px; + border-collapse: collapse; + border-right: solid 1px #eee; + border-left: solid 1px #eee; +} +td:first-child { + border-left: none; +} +td:last-child { + border-right: none; +} + +)" + "</style>"; + +const char *EndHeader = "</head>"; + +const char *BeginCenteredDiv = "<div class='centered'>"; + +const char *EndCenteredDiv = "</div>"; + +const char *BeginSourceNameDiv = "<div class='source-name-title'>"; + +const char *EndSourceNameDiv = "</div>"; + +const char *BeginCodeTD = "<td class='code'>"; + +const char *EndCodeTD = "</td>"; + +const char *BeginPre = "<pre>"; + +const char *EndPre = "</pre>"; + +const char *BeginExpansionDiv = "<div class='expansion-view'>"; + +const char *EndExpansionDiv = "</div>"; + +const char *BeginTable = "<table>"; + +const char *EndTable = "</table>"; + +void emitPrelude(raw_ostream &OS) { + OS << "<!doctype html>" + "<html>" + << BeginHeader << CSSForCoverage << EndHeader << "<body>" + << BeginCenteredDiv; +} + +void emitEpilog(raw_ostream &OS) { + OS << EndCenteredDiv << "</body>" + "</html>"; +} + +// Return a string with the special characters in \p Str escaped. +std::string escape(StringRef Str) { + std::string Result; + for (char C : Str) { + if (C == '&') + Result += "&"; + else if (C == '<') + Result += "<"; + else if (C == '>') + Result += ">"; + else if (C == '\"') + Result += """; + else + Result += C; + } + return Result; +} + +// Create a \p Name tag around \p Str, and optionally set its \p ClassName. +std::string tag(const std::string &Name, const std::string &Str, + const std::string &ClassName = "") { + std::string Tag = "<" + Name; + if (ClassName != "") + Tag += " class='" + ClassName + "'"; + return Tag + ">" + Str + "</" + Name + ">"; +} + +// Create an anchor to \p Link with the label \p Str. +std::string a(const std::string &Link, const std::string &Str) { + return "<a href='" + Link + "'>" + Str + "</a>"; +} + +} // anonymous namespace + +Expected<CoveragePrinter::OwnedStream> +CoveragePrinterHTML::createViewFile(StringRef Path, bool InToplevel) { + auto OSOrErr = createOutputStream(Path, "html", InToplevel); + if (!OSOrErr) + return OSOrErr; + + OwnedStream OS = std::move(OSOrErr.get()); + emitPrelude(*OS.get()); + return std::move(OS); +} + +void CoveragePrinterHTML::closeViewFile(OwnedStream OS) { + emitEpilog(*OS.get()); +} + +Error CoveragePrinterHTML::createIndexFile(ArrayRef<StringRef> SourceFiles) { + auto OSOrErr = createOutputStream("index", "html", /*InToplevel=*/true); + if (Error E = OSOrErr.takeError()) + return E; + auto OS = std::move(OSOrErr.get()); + raw_ostream &OSRef = *OS.get(); + + // Emit a table containing links to reports for each file in the covmapping. + emitPrelude(OSRef); + OSRef << BeginSourceNameDiv << "Index" << EndSourceNameDiv; + OSRef << BeginTable; + for (StringRef SF : SourceFiles) { + std::string LinkText = escape(sys::path::relative_path(SF)); + std::string LinkTarget = + escape(getOutputPath(SF, "html", /*InToplevel=*/false)); + OSRef << tag("tr", tag("td", tag("pre", a(LinkTarget, LinkText), "code"))); + } + OSRef << EndTable; + emitEpilog(OSRef); + + return Error::success(); +} + +void SourceCoverageViewHTML::renderViewHeader(raw_ostream &OS) { + OS << BeginTable; +} + +void SourceCoverageViewHTML::renderViewFooter(raw_ostream &OS) { + OS << EndTable; +} + +void SourceCoverageViewHTML::renderSourceName(raw_ostream &OS) { + OS << BeginSourceNameDiv << tag("pre", escape(getSourceName())) + << EndSourceNameDiv; +} + +void SourceCoverageViewHTML::renderLinePrefix(raw_ostream &OS, unsigned) { + OS << "<tr>"; +} + +void SourceCoverageViewHTML::renderLineSuffix(raw_ostream &OS, unsigned) { + // If this view has sub-views, renderLine() cannot close the view's cell. + // Take care of it here, after all sub-views have been rendered. + if (hasSubViews()) + OS << EndCodeTD; + OS << "</tr>"; +} + +void SourceCoverageViewHTML::renderViewDivider(raw_ostream &, unsigned) { + // The table-based output makes view dividers unnecessary. +} + +void SourceCoverageViewHTML::renderLine( + raw_ostream &OS, LineRef L, const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, unsigned) { + StringRef Line = L.Line; + + // Steps for handling text-escaping, highlighting, and tooltip creation: + // + // 1. Split the line into N+1 snippets, where N = |Segments|. The first + // snippet starts from Col=1 and ends at the start of the first segment. + // The last snippet starts at the last mapped column in the line and ends + // at the end of the line. Both are required but may be empty. + + SmallVector<std::string, 8> Snippets; + + unsigned LCol = 1; + auto Snip = [&](unsigned Start, unsigned Len) { + assert(Start + Len <= Line.size() && "Snippet extends past the EOL"); + Snippets.push_back(Line.substr(Start, Len)); + LCol += Len; + }; + + Snip(LCol - 1, Segments.empty() ? 0 : (Segments.front()->Col - 1)); + + for (unsigned I = 1, E = Segments.size(); I < E; ++I) { + assert(LCol == Segments[I - 1]->Col && "Snippet start position is wrong"); + Snip(LCol - 1, Segments[I]->Col - LCol); + } + + // |Line| + 1 is needed to avoid underflow when, e.g |Line| = 0 and LCol = 1. + Snip(LCol - 1, Line.size() + 1 - LCol); + assert(LCol == Line.size() + 1 && "Final snippet doesn't reach the EOL"); + + // 2. Escape all of the snippets. + + for (unsigned I = 0, E = Snippets.size(); I < E; ++I) + Snippets[I] = escape(Snippets[I]); + + // 3. Use \p WrappedSegment to set the highlight for snippets 0 and 1. Use + // segment 1 to set the highlight for snippet 2, segment 2 to set the + // highlight for snippet 3, and so on. + + Optional<std::string> Color; + auto Highlight = [&](const std::string &Snippet) { + return tag("span", Snippet, Color.getValue()); + }; + + auto CheckIfUncovered = [](const coverage::CoverageSegment *S) { + return S && S->HasCount && S->Count == 0; + }; + + if (CheckIfUncovered(WrappedSegment) || + CheckIfUncovered(Segments.empty() ? nullptr : Segments.front())) { + Color = "red"; + Snippets[0] = Highlight(Snippets[0]); + Snippets[1] = Highlight(Snippets[1]); + } + + for (unsigned I = 1, E = Segments.size(); I < E; ++I) { + const auto *CurSeg = Segments[I]; + if (CurSeg->Col == ExpansionCol) + Color = "cyan"; + else if (CheckIfUncovered(CurSeg)) + Color = "red"; + else + Color = None; + + if (Color.hasValue()) + Snippets[I + 1] = Highlight(Snippets[I + 1]); + } + + // 4. Snippets[1:N+1] correspond to \p Segments[0:N]: use these to generate + // sub-line region count tooltips if needed. + + bool HasMultipleRegions = [&] { + unsigned RegionCount = 0; + for (const auto *S : Segments) + if (S->HasCount && S->IsRegionEntry) + if (++RegionCount > 1) + return true; + return false; + }(); + + if (shouldRenderRegionMarkers(HasMultipleRegions)) { + for (unsigned I = 0, E = Segments.size(); I < E; ++I) { + const auto *CurSeg = Segments[I]; + if (!CurSeg->IsRegionEntry || !CurSeg->HasCount) + continue; + + Snippets[I + 1] = + tag("div", Snippets[I + 1] + tag("span", formatCount(CurSeg->Count), + "tooltip-content"), + "tooltip"); + } + } + + OS << BeginCodeTD; + OS << BeginPre; + for (const auto &Snippet : Snippets) + OS << Snippet; + OS << EndPre; + + // If there are no sub-views left to attach to this cell, end the cell. + // Otherwise, end it after the sub-views are rendered (renderLineSuffix()). + if (!hasSubViews()) + OS << EndCodeTD; +} + +void SourceCoverageViewHTML::renderLineCoverageColumn( + raw_ostream &OS, const LineCoverageStats &Line) { + std::string Count = ""; + if (Line.isMapped()) + Count = tag("pre", formatCount(Line.ExecutionCount)); + std::string CoverageClass = + (Line.ExecutionCount > 0) ? "covered-line" : "uncovered-line"; + OS << tag("td", Count, CoverageClass); +} + +void SourceCoverageViewHTML::renderLineNumberColumn(raw_ostream &OS, + unsigned LineNo) { + OS << tag("td", tag("pre", utostr(uint64_t(LineNo))), "line-number"); +} + +void SourceCoverageViewHTML::renderRegionMarkers(raw_ostream &, + CoverageSegmentArray, + unsigned) { + // Region markers are rendered in-line using tooltips. +} + +void SourceCoverageViewHTML::renderExpansionSite( + raw_ostream &OS, LineRef L, const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, unsigned ViewDepth) { + // Render the line containing the expansion site. No extra formatting needed. + renderLine(OS, L, WrappedSegment, Segments, ExpansionCol, ViewDepth); +} + +void SourceCoverageViewHTML::renderExpansionView(raw_ostream &OS, + ExpansionView &ESV, + unsigned ViewDepth) { + OS << BeginExpansionDiv; + ESV.View->print(OS, /*WholeFile=*/false, /*ShowSourceName=*/false, + ViewDepth + 1); + OS << EndExpansionDiv; +} + +void SourceCoverageViewHTML::renderInstantiationView(raw_ostream &OS, + InstantiationView &ISV, + unsigned ViewDepth) { + OS << BeginExpansionDiv; + ISV.View->print(OS, /*WholeFile=*/false, /*ShowSourceName=*/true, ViewDepth); + OS << EndExpansionDiv; +} diff --git a/tools/llvm-cov/SourceCoverageViewHTML.h b/tools/llvm-cov/SourceCoverageViewHTML.h new file mode 100644 index 0000000000000..50ecf2bf89970 --- /dev/null +++ b/tools/llvm-cov/SourceCoverageViewHTML.h @@ -0,0 +1,83 @@ +//===- SourceCoverageViewHTML.h - A html code coverage view ---------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +/// +/// \file This file defines the interface to the html coverage renderer. +/// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_COV_SOURCECOVERAGEVIEWHTML_H +#define LLVM_COV_SOURCECOVERAGEVIEWHTML_H + +#include "SourceCoverageView.h" + +namespace llvm { + +/// \brief A coverage printer for html output. +class CoveragePrinterHTML : public CoveragePrinter { +public: + Expected<OwnedStream> createViewFile(StringRef Path, + bool InToplevel) override; + + void closeViewFile(OwnedStream OS) override; + + Error createIndexFile(ArrayRef<StringRef> SourceFiles) override; + + CoveragePrinterHTML(const CoverageViewOptions &Opts) + : CoveragePrinter(Opts) {} +}; + +/// \brief A code coverage view which supports html-based rendering. +class SourceCoverageViewHTML : public SourceCoverageView { + void renderViewHeader(raw_ostream &OS) override; + + void renderViewFooter(raw_ostream &OS) override; + + void renderSourceName(raw_ostream &OS) override; + + void renderLinePrefix(raw_ostream &OS, unsigned ViewDepth) override; + + void renderLineSuffix(raw_ostream &OS, unsigned ViewDepth) override; + + void renderViewDivider(raw_ostream &OS, unsigned ViewDepth) override; + + void renderLine(raw_ostream &OS, LineRef L, + const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, + unsigned ViewDepth) override; + + void renderExpansionSite(raw_ostream &OS, LineRef L, + const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, + unsigned ViewDepth) override; + + void renderExpansionView(raw_ostream &OS, ExpansionView &ESV, + unsigned ViewDepth) override; + + void renderInstantiationView(raw_ostream &OS, InstantiationView &ISV, + unsigned ViewDepth) override; + + void renderLineCoverageColumn(raw_ostream &OS, + const LineCoverageStats &Line) override; + + void renderLineNumberColumn(raw_ostream &OS, unsigned LineNo) override; + + void renderRegionMarkers(raw_ostream &OS, CoverageSegmentArray Segments, + unsigned ViewDepth) override; + +public: + SourceCoverageViewHTML(StringRef SourceName, const MemoryBuffer &File, + const CoverageViewOptions &Options, + coverage::CoverageData &&CoverageInfo) + : SourceCoverageView(SourceName, File, Options, std::move(CoverageInfo)) { + } +}; + +} // namespace llvm + +#endif // LLVM_COV_SOURCECOVERAGEVIEWHTML_H diff --git a/tools/llvm-cov/SourceCoverageViewText.cpp b/tools/llvm-cov/SourceCoverageViewText.cpp new file mode 100644 index 0000000000000..ae9d6daed08f7 --- /dev/null +++ b/tools/llvm-cov/SourceCoverageViewText.cpp @@ -0,0 +1,213 @@ +//===- SourceCoverageViewText.cpp - A text-based code coverage view -------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +/// +/// \file This file implements the text-based coverage renderer. +/// +//===----------------------------------------------------------------------===// + +#include "SourceCoverageViewText.h" +#include "llvm/ADT/Optional.h" +#include "llvm/ADT/SmallString.h" +#include "llvm/ADT/StringExtras.h" + +using namespace llvm; + +Expected<CoveragePrinter::OwnedStream> +CoveragePrinterText::createViewFile(StringRef Path, bool InToplevel) { + return createOutputStream(Path, "txt", InToplevel); +} + +void CoveragePrinterText::closeViewFile(OwnedStream OS) { + OS->operator<<('\n'); +} + +Error CoveragePrinterText::createIndexFile(ArrayRef<StringRef> SourceFiles) { + auto OSOrErr = createOutputStream("index", "txt", /*InToplevel=*/true); + if (Error E = OSOrErr.takeError()) + return E; + auto OS = std::move(OSOrErr.get()); + raw_ostream &OSRef = *OS.get(); + + for (StringRef SF : SourceFiles) + OSRef << getOutputPath(SF, "txt", /*InToplevel=*/false) << '\n'; + + return Error::success(); +} + +namespace { + +static const unsigned LineCoverageColumnWidth = 7; +static const unsigned LineNumberColumnWidth = 5; + +/// \brief Get the width of the leading columns. +unsigned getCombinedColumnWidth(const CoverageViewOptions &Opts) { + return (Opts.ShowLineStats ? LineCoverageColumnWidth + 1 : 0) + + (Opts.ShowLineNumbers ? LineNumberColumnWidth + 1 : 0); +} + +/// \brief The width of the line that is used to divide between the view and +/// the subviews. +unsigned getDividerWidth(const CoverageViewOptions &Opts) { + return getCombinedColumnWidth(Opts) + 4; +} + +} // anonymous namespace + +void SourceCoverageViewText::renderViewHeader(raw_ostream &) {} + +void SourceCoverageViewText::renderViewFooter(raw_ostream &) {} + +void SourceCoverageViewText::renderSourceName(raw_ostream &OS) { + getOptions().colored_ostream(OS, raw_ostream::CYAN) << getSourceName() + << ":\n"; +} + +void SourceCoverageViewText::renderLinePrefix(raw_ostream &OS, + unsigned ViewDepth) { + for (unsigned I = 0; I < ViewDepth; ++I) + OS << " |"; +} + +void SourceCoverageViewText::renderLineSuffix(raw_ostream &, unsigned) {} + +void SourceCoverageViewText::renderViewDivider(raw_ostream &OS, + unsigned ViewDepth) { + assert(ViewDepth != 0 && "Cannot render divider at top level"); + renderLinePrefix(OS, ViewDepth - 1); + OS.indent(2); + unsigned Length = getDividerWidth(getOptions()); + for (unsigned I = 0; I < Length; ++I) + OS << '-'; + OS << '\n'; +} + +void SourceCoverageViewText::renderLine( + raw_ostream &OS, LineRef L, + const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, unsigned ViewDepth) { + StringRef Line = L.Line; + unsigned LineNumber = L.LineNo; + + Optional<raw_ostream::Colors> Highlight; + SmallVector<std::pair<unsigned, unsigned>, 2> HighlightedRanges; + + // The first segment overlaps from a previous line, so we treat it specially. + if (WrappedSegment && WrappedSegment->HasCount && WrappedSegment->Count == 0) + Highlight = raw_ostream::RED; + + // Output each segment of the line, possibly highlighted. + unsigned Col = 1; + for (const auto *S : Segments) { + unsigned End = std::min(S->Col, static_cast<unsigned>(Line.size()) + 1); + colored_ostream(OS, Highlight ? *Highlight : raw_ostream::SAVEDCOLOR, + getOptions().Colors && Highlight, /*Bold=*/false, + /*BG=*/true) + << Line.substr(Col - 1, End - Col); + if (getOptions().Debug && Highlight) + HighlightedRanges.push_back(std::make_pair(Col, End)); + Col = End; + if (Col == ExpansionCol) + Highlight = raw_ostream::CYAN; + else if (S->HasCount && S->Count == 0) + Highlight = raw_ostream::RED; + else + Highlight = None; + } + + // Show the rest of the line. + colored_ostream(OS, Highlight ? *Highlight : raw_ostream::SAVEDCOLOR, + getOptions().Colors && Highlight, /*Bold=*/false, /*BG=*/true) + << Line.substr(Col - 1, Line.size() - Col + 1); + OS << '\n'; + + if (getOptions().Debug) { + for (const auto &Range : HighlightedRanges) + errs() << "Highlighted line " << LineNumber << ", " << Range.first + << " -> " << Range.second << '\n'; + if (Highlight) + errs() << "Highlighted line " << LineNumber << ", " << Col << " -> ?\n"; + } +} + +void SourceCoverageViewText::renderLineCoverageColumn( + raw_ostream &OS, const LineCoverageStats &Line) { + if (!Line.isMapped()) { + OS.indent(LineCoverageColumnWidth) << '|'; + return; + } + std::string C = formatCount(Line.ExecutionCount); + OS.indent(LineCoverageColumnWidth - C.size()); + colored_ostream(OS, raw_ostream::MAGENTA, + Line.hasMultipleRegions() && getOptions().Colors) + << C; + OS << '|'; +} + +void SourceCoverageViewText::renderLineNumberColumn(raw_ostream &OS, + unsigned LineNo) { + SmallString<32> Buffer; + raw_svector_ostream BufferOS(Buffer); + BufferOS << LineNo; + auto Str = BufferOS.str(); + // Trim and align to the right. + Str = Str.substr(0, std::min(Str.size(), (size_t)LineNumberColumnWidth)); + OS.indent(LineNumberColumnWidth - Str.size()) << Str << '|'; +} + +void SourceCoverageViewText::renderRegionMarkers( + raw_ostream &OS, CoverageSegmentArray Segments, unsigned ViewDepth) { + renderLinePrefix(OS, ViewDepth); + OS.indent(getCombinedColumnWidth(getOptions())); + + unsigned PrevColumn = 1; + for (const auto *S : Segments) { + if (!S->IsRegionEntry) + continue; + // Skip to the new region. + if (S->Col > PrevColumn) + OS.indent(S->Col - PrevColumn); + PrevColumn = S->Col + 1; + std::string C = formatCount(S->Count); + PrevColumn += C.size(); + OS << '^' << C; + } + OS << '\n'; + + if (getOptions().Debug) + for (const auto *S : Segments) + errs() << "Marker at " << S->Line << ":" << S->Col << " = " + << formatCount(S->Count) << (S->IsRegionEntry ? "\n" : " (pop)\n"); +} + +void SourceCoverageViewText::renderExpansionSite( + raw_ostream &OS, LineRef L, const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, unsigned ViewDepth) { + renderLinePrefix(OS, ViewDepth); + OS.indent(getCombinedColumnWidth(getOptions()) + (ViewDepth == 0 ? 0 : 1)); + renderLine(OS, L, WrappedSegment, Segments, ExpansionCol, ViewDepth); +} + +void SourceCoverageViewText::renderExpansionView(raw_ostream &OS, + ExpansionView &ESV, + unsigned ViewDepth) { + // Render the child subview. + if (getOptions().Debug) + errs() << "Expansion at line " << ESV.getLine() << ", " << ESV.getStartCol() + << " -> " << ESV.getEndCol() << '\n'; + ESV.View->print(OS, /*WholeFile=*/false, /*ShowSourceName=*/false, + ViewDepth + 1); +} + +void SourceCoverageViewText::renderInstantiationView(raw_ostream &OS, + InstantiationView &ISV, + unsigned ViewDepth) { + renderLinePrefix(OS, ViewDepth); + OS << ' '; + ISV.View->print(OS, /*WholeFile=*/false, /*ShowSourceName=*/true, ViewDepth); +} diff --git a/tools/llvm-cov/SourceCoverageViewText.h b/tools/llvm-cov/SourceCoverageViewText.h new file mode 100644 index 0000000000000..b2331247b37b7 --- /dev/null +++ b/tools/llvm-cov/SourceCoverageViewText.h @@ -0,0 +1,83 @@ +//===- SourceCoverageViewText.h - A text-based code coverage view ---------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +/// +/// \file This file defines the interface to the text-based coverage renderer. +/// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_COV_SOURCECOVERAGEVIEWTEXT_H +#define LLVM_COV_SOURCECOVERAGEVIEWTEXT_H + +#include "SourceCoverageView.h" + +namespace llvm { + +/// \brief A coverage printer for text output. +class CoveragePrinterText : public CoveragePrinter { +public: + Expected<OwnedStream> createViewFile(StringRef Path, + bool InToplevel) override; + + void closeViewFile(OwnedStream OS) override; + + Error createIndexFile(ArrayRef<StringRef> SourceFiles) override; + + CoveragePrinterText(const CoverageViewOptions &Opts) + : CoveragePrinter(Opts) {} +}; + +/// \brief A code coverage view which supports text-based rendering. +class SourceCoverageViewText : public SourceCoverageView { + void renderViewHeader(raw_ostream &OS) override; + + void renderViewFooter(raw_ostream &OS) override; + + void renderSourceName(raw_ostream &OS) override; + + void renderLinePrefix(raw_ostream &OS, unsigned ViewDepth) override; + + void renderLineSuffix(raw_ostream &OS, unsigned ViewDepth) override; + + void renderViewDivider(raw_ostream &OS, unsigned ViewDepth) override; + + void renderLine(raw_ostream &OS, LineRef L, + const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, + unsigned ViewDepth) override; + + void renderExpansionSite(raw_ostream &OS, LineRef L, + const coverage::CoverageSegment *WrappedSegment, + CoverageSegmentArray Segments, unsigned ExpansionCol, + unsigned ViewDepth) override; + + void renderExpansionView(raw_ostream &OS, ExpansionView &ESV, + unsigned ViewDepth) override; + + void renderInstantiationView(raw_ostream &OS, InstantiationView &ISV, + unsigned ViewDepth) override; + + void renderLineCoverageColumn(raw_ostream &OS, + const LineCoverageStats &Line) override; + + void renderLineNumberColumn(raw_ostream &OS, unsigned LineNo) override; + + void renderRegionMarkers(raw_ostream &OS, CoverageSegmentArray Segments, + unsigned ViewDepth) override; + +public: + SourceCoverageViewText(StringRef SourceName, const MemoryBuffer &File, + const CoverageViewOptions &Options, + coverage::CoverageData &&CoverageInfo) + : SourceCoverageView(SourceName, File, Options, std::move(CoverageInfo)) { + } +}; + +} // namespace llvm + +#endif // LLVM_COV_SOURCECOVERAGEVIEWTEXT_H diff --git a/tools/llvm-cov/TestingSupport.cpp b/tools/llvm-cov/TestingSupport.cpp index 6959897482ca2..72768f4fd583f 100644 --- a/tools/llvm-cov/TestingSupport.cpp +++ b/tools/llvm-cov/TestingSupport.cpp @@ -8,11 +8,9 @@ //===----------------------------------------------------------------------===// #include "llvm/Object/ObjectFile.h" +#include "llvm/ProfileData/InstrProf.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/LEB128.h" -#include "llvm/Support/ManagedStatic.h" -#include "llvm/Support/PrettyStackTrace.h" -#include "llvm/Support/Signals.h" #include "llvm/Support/raw_ostream.h" #include <functional> #include <system_error> @@ -21,10 +19,6 @@ using namespace llvm; using namespace object; int convertForTestingMain(int argc, const char *argv[]) { - sys::PrintStackTraceOnErrorSignal(); - PrettyStackTraceProgram X(argc, argv); - llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. - cl::opt<std::string> InputSourceFile(cl::Positional, cl::Required, cl::desc("<Source file>")); @@ -36,8 +30,12 @@ int convertForTestingMain(int argc, const char *argv[]) { cl::ParseCommandLineOptions(argc, argv, "LLVM code coverage tool\n"); auto ObjErr = llvm::object::ObjectFile::createObjectFile(InputSourceFile); - if (auto Err = ObjErr.getError()) { - errs() << "error: " << Err.message() << "\n"; + if (!ObjErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(ObjErr.takeError(), OS, ""); + OS.flush(); + errs() << "error: " << Buf; return 1; } ObjectFile *OF = ObjErr.get().getBinary(); @@ -54,9 +52,9 @@ int convertForTestingMain(int argc, const char *argv[]) { StringRef Name; if (Section.getName(Name)) return 1; - if (Name == "__llvm_prf_names") { + if (Name == llvm::getInstrProfNameSectionName(false)) { ProfileNames = Section; - } else if (Name == "__llvm_covmap") { + } else if (Name == llvm::getInstrProfCoverageSectionName(false)) { CoverageMapping = Section; } else continue; @@ -84,7 +82,11 @@ int convertForTestingMain(int argc, const char *argv[]) { OS << "llvmcovmtestdata"; encodeULEB128(ProfileNamesData.size(), OS); encodeULEB128(ProfileNamesAddress, OS); - OS << ProfileNamesData << CoverageMappingData; + OS << ProfileNamesData; + // Coverage mapping data is expected to have an alignment of 8. + for (unsigned Pad = OffsetToAlignment(OS.tell(), 8); Pad; --Pad) + OS.write(uint8_t(0)); + OS << CoverageMappingData; return 0; } diff --git a/tools/llvm-cov/gcov.cpp b/tools/llvm-cov/gcov.cpp index a5343fa29afc3..4652fed2a384e 100644 --- a/tools/llvm-cov/gcov.cpp +++ b/tools/llvm-cov/gcov.cpp @@ -16,10 +16,7 @@ #include "llvm/Support/Errc.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/GCOV.h" -#include "llvm/Support/ManagedStatic.h" #include "llvm/Support/Path.h" -#include "llvm/Support/PrettyStackTrace.h" -#include "llvm/Support/Signals.h" #include <system_error> using namespace llvm; @@ -85,11 +82,6 @@ static void reportCoverage(StringRef SourceFile, StringRef ObjectDir, } int gcovMain(int argc, const char *argv[]) { - // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); - PrettyStackTraceProgram X(argc, argv); - llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. - cl::list<std::string> SourceFiles(cl::Positional, cl::OneOrMore, cl::desc("SOURCEFILE")); diff --git a/tools/llvm-cov/llvm-cov.cpp b/tools/llvm-cov/llvm-cov.cpp index 8c5acaef63b25..ba60cd91da906 100644 --- a/tools/llvm-cov/llvm-cov.cpp +++ b/tools/llvm-cov/llvm-cov.cpp @@ -14,8 +14,11 @@ #include "llvm/ADT/StringRef.h" #include "llvm/ADT/StringSwitch.h" #include "llvm/Support/CommandLine.h" +#include "llvm/Support/ManagedStatic.h" #include "llvm/Support/Path.h" +#include "llvm/Support/PrettyStackTrace.h" #include "llvm/Support/Process.h" +#include "llvm/Support/Signals.h" #include "llvm/Support/raw_ostream.h" #include <string> @@ -51,6 +54,11 @@ static int versionMain(int argc, const char *argv[]) { } int main(int argc, const char **argv) { + // Print a stack trace if we signal out. + sys::PrintStackTraceOnErrorSignal(argv[0]); + PrettyStackTraceProgram X(argc, argv); + llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. + // If argv[0] is or ends with 'gcov', always be gcov compatible if (sys::path::stem(argv[0]).endswith_lower("gcov")) return gcovMain(argc, argv); diff --git a/tools/llvm-cxxdump/Error.cpp b/tools/llvm-cxxdump/Error.cpp index 16fed96e4834c..ff9f0f5790669 100644 --- a/tools/llvm-cxxdump/Error.cpp +++ b/tools/llvm-cxxdump/Error.cpp @@ -17,6 +17,9 @@ using namespace llvm; namespace { +// FIXME: This class is only here to support the transition to llvm::Error. It +// will be removed once this transition is complete. Clients should prefer to +// deal with the Error value directly, rather than converting to error_code. class cxxdump_error_category : public std::error_category { public: const char *name() const LLVM_NOEXCEPT override { return "llvm.cxxdump"; } diff --git a/tools/llvm-cxxdump/Makefile b/tools/llvm-cxxdump/Makefile deleted file mode 100644 index 02e8e5f901080..0000000000000 --- a/tools/llvm-cxxdump/Makefile +++ /dev/null @@ -1,18 +0,0 @@ -##===- tools/llvm-cxxdump/Makefile -------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-cxxdump -LINK_COMPONENTS := bitreader object all-targets - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common - diff --git a/tools/llvm-cxxdump/llvm-cxxdump.cpp b/tools/llvm-cxxdump/llvm-cxxdump.cpp index 3dda69266a2d8..c92d20d6ccf17 100644 --- a/tools/llvm-cxxdump/llvm-cxxdump.cpp +++ b/tools/llvm-cxxdump/llvm-cxxdump.cpp @@ -50,6 +50,14 @@ static void error(std::error_code EC) { exit(1); } +static void error(Error Err) { + if (Err) { + logAllUnhandledErrors(std::move(Err), outs(), "Error reading file: "); + outs().flush(); + exit(1); + } +} + } // namespace llvm static void reportError(StringRef Input, StringRef Message) { @@ -79,8 +87,8 @@ static void collectRelocatedSymbols(const ObjectFile *Obj, const object::symbol_iterator RelocSymI = Reloc.getSymbol(); if (RelocSymI == Obj->symbol_end()) continue; - ErrorOr<StringRef> RelocSymName = RelocSymI->getName(); - error(RelocSymName.getError()); + Expected<StringRef> RelocSymName = RelocSymI->getName(); + error(errorToErrorCode(RelocSymName.takeError())); uint64_t Offset = Reloc.getOffset(); if (Offset >= SymOffset && Offset < SymEnd) { *I = *RelocSymName; @@ -101,8 +109,8 @@ static void collectRelocationOffsets( const object::symbol_iterator RelocSymI = Reloc.getSymbol(); if (RelocSymI == Obj->symbol_end()) continue; - ErrorOr<StringRef> RelocSymName = RelocSymI->getName(); - error(RelocSymName.getError()); + Expected<StringRef> RelocSymName = RelocSymI->getName(); + error(errorToErrorCode(RelocSymName.takeError())); uint64_t Offset = Reloc.getOffset(); if (Offset >= SymOffset && Offset < SymEnd) Collection[std::make_pair(SymName, Offset - SymOffset)] = *RelocSymName; @@ -175,11 +183,11 @@ static void dumpCXXData(const ObjectFile *Obj) { for (auto &P : SymAddr) { object::SymbolRef Sym = P.first; uint64_t SymSize = P.second; - ErrorOr<StringRef> SymNameOrErr = Sym.getName(); - error(SymNameOrErr.getError()); + Expected<StringRef> SymNameOrErr = Sym.getName(); + error(errorToErrorCode(SymNameOrErr.takeError())); StringRef SymName = *SymNameOrErr; - ErrorOr<object::section_iterator> SecIOrErr = Sym.getSection(); - error(SecIOrErr.getError()); + Expected<object::section_iterator> SecIOrErr = Sym.getSection(); + error(errorToErrorCode(SecIOrErr.takeError())); object::section_iterator SecI = *SecIOrErr; // Skip external symbols. if (SecI == Obj->section_end()) @@ -190,8 +198,8 @@ static void dumpCXXData(const ObjectFile *Obj) { continue; StringRef SecContents; error(Sec.getContents(SecContents)); - ErrorOr<uint64_t> SymAddressOrErr = Sym.getAddress(); - error(SymAddressOrErr.getError()); + Expected<uint64_t> SymAddressOrErr = Sym.getAddress(); + error(errorToErrorCode(SymAddressOrErr.takeError())); uint64_t SymAddress = *SymAddressOrErr; uint64_t SecAddress = Sec.getAddress(); uint64_t SecSize = Sec.getSize(); @@ -482,14 +490,19 @@ static void dumpCXXData(const ObjectFile *Obj) { } static void dumpArchive(const Archive *Arc) { - for (auto &ErrorOrChild : Arc->children()) { - error(ErrorOrChild.getError()); - const Archive::Child &ArcC = *ErrorOrChild; - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = ArcC.getAsBinary(); - if (std::error_code EC = ChildOrErr.getError()) { + Error Err; + for (auto &ArcC : Arc->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = ArcC.getAsBinary(); + if (!ChildOrErr) { // Ignore non-object files. - if (EC != object_error::invalid_file_type) - reportError(Arc->getFileName(), EC.message()); + if (auto E = isNotObjectErrorInvalidFileType(ChildOrErr.takeError())) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(std::move(E), OS, ""); + OS.flush(); + reportError(Arc->getFileName(), Buf); + } + ChildOrErr.takeError(); continue; } @@ -498,12 +511,14 @@ static void dumpArchive(const Archive *Arc) { else reportError(Arc->getFileName(), cxxdump_error::unrecognized_file_format); } + error(std::move(Err)); } static void dumpInput(StringRef File) { // Attempt to open the binary. - ErrorOr<OwningBinary<Binary>> BinaryOrErr = createBinary(File); - if (std::error_code EC = BinaryOrErr.getError()) { + Expected<OwningBinary<Binary>> BinaryOrErr = createBinary(File); + if (!BinaryOrErr) { + auto EC = errorToErrorCode(BinaryOrErr.takeError()); reportError(File, EC); return; } @@ -518,7 +533,7 @@ static void dumpInput(StringRef File) { } int main(int argc, const char *argv[]) { - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; diff --git a/tools/llvm-diff/DiffConsumer.h b/tools/llvm-diff/DiffConsumer.h index 855f6884e65b4..82f5ce598b443 100644 --- a/tools/llvm-diff/DiffConsumer.h +++ b/tools/llvm-diff/DiffConsumer.h @@ -17,11 +17,12 @@ #include "DiffLog.h" #include "llvm/ADT/DenseMap.h" #include "llvm/ADT/SmallVector.h" -#include "llvm/ADT/StringRef.h" +#include "llvm/IR/Value.h" #include "llvm/Support/Casting.h" #include "llvm/Support/raw_ostream.h" namespace llvm { +class StringRef; class Module; class Value; class Function; diff --git a/tools/llvm-diff/DiffLog.cpp b/tools/llvm-diff/DiffLog.cpp index ed86058f1af10..898749e73bd20 100644 --- a/tools/llvm-diff/DiffLog.cpp +++ b/tools/llvm-diff/DiffLog.cpp @@ -13,7 +13,6 @@ #include "DiffLog.h" #include "DiffConsumer.h" -#include "llvm/ADT/SmallVector.h" #include "llvm/ADT/StringRef.h" #include "llvm/IR/Instructions.h" diff --git a/tools/llvm-diff/DifferenceEngine.cpp b/tools/llvm-diff/DifferenceEngine.cpp index 456560b093aba..df208a26ab7df 100644 --- a/tools/llvm-diff/DifferenceEngine.cpp +++ b/tools/llvm-diff/DifferenceEngine.cpp @@ -16,7 +16,6 @@ #include "llvm/ADT/DenseMap.h" #include "llvm/ADT/DenseSet.h" #include "llvm/ADT/SmallVector.h" -#include "llvm/ADT/StringRef.h" #include "llvm/ADT/StringSet.h" #include "llvm/IR/CFG.h" #include "llvm/IR/CallSite.h" @@ -210,7 +209,8 @@ class FunctionDifferenceEngine { if (!LeftI->use_empty()) TentativeValues.insert(std::make_pair(LeftI, RightI)); - ++LI, ++RI; + ++LI; + ++RI; } while (LI != LE); // This is sufficient: we can't get equality of // terminators if there are residual instructions. @@ -555,7 +555,9 @@ void FunctionDifferenceEngine::runBlockDiff(BasicBlock::iterator LStart, PI = Path.begin(), PE = Path.end(); while (PI != PE && *PI == DC_match) { unify(&*LI, &*RI); - ++PI, ++LI, ++RI; + ++PI; + ++LI; + ++RI; } for (; PI != PE; ++PI) { @@ -589,7 +591,8 @@ void FunctionDifferenceEngine::runBlockDiff(BasicBlock::iterator LStart, while (LI != LE) { assert(RI != RE); unify(&*LI, &*RI); - ++LI, ++RI; + ++LI; + ++RI; } // If the terminators have different kinds, but one is an invoke and the diff --git a/tools/llvm-diff/DifferenceEngine.h b/tools/llvm-diff/DifferenceEngine.h index f0d831144a4cb..7f084a377f0c7 100644 --- a/tools/llvm-diff/DifferenceEngine.h +++ b/tools/llvm-diff/DifferenceEngine.h @@ -17,7 +17,6 @@ #include "DiffConsumer.h" #include "DiffLog.h" -#include "llvm/ADT/SmallVector.h" #include "llvm/ADT/StringRef.h" #include <utility> diff --git a/tools/llvm-diff/Makefile b/tools/llvm-diff/Makefile deleted file mode 100644 index bd97a6a9f5e92..0000000000000 --- a/tools/llvm-diff/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-diff/Makefile ----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-diff -LINK_COMPONENTS := asmparser bitreader irreader - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-diff/llvm-diff.cpp b/tools/llvm-diff/llvm-diff.cpp index ae58f5caa9138..e449d6994784e 100644 --- a/tools/llvm-diff/llvm-diff.cpp +++ b/tools/llvm-diff/llvm-diff.cpp @@ -13,8 +13,6 @@ #include "DiffLog.h" #include "DifferenceEngine.h" -#include "llvm/ADT/DenseMap.h" -#include "llvm/ADT/SmallVector.h" #include "llvm/ADT/StringRef.h" #include "llvm/IR/LLVMContext.h" #include "llvm/IR/Module.h" diff --git a/tools/llvm-dis/Makefile b/tools/llvm-dis/Makefile deleted file mode 100644 index aeeeed0d68c9f..0000000000000 --- a/tools/llvm-dis/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-dis/Makefile ------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-dis -LINK_COMPONENTS := bitreader analysis - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-dis/llvm-dis.cpp b/tools/llvm-dis/llvm-dis.cpp index 9fdfcd4256c9a..88333aeb68892 100644 --- a/tools/llvm-dis/llvm-dis.cpp +++ b/tools/llvm-dis/llvm-dis.cpp @@ -27,6 +27,7 @@ #include "llvm/IR/Type.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/DataStream.h" +#include "llvm/Support/Error.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/FormattedStream.h" #include "llvm/Support/ManagedStatic.h" @@ -59,6 +60,11 @@ static cl::opt<bool> PreserveAssemblyUseListOrder( cl::desc("Preserve use-list order when writing LLVM assembly."), cl::init(false), cl::Hidden); +static cl::opt<bool> + MaterializeMetadata("materialize-metadata", + cl::desc("Load module without materializing metadata, " + "then materialize only the metadata")); + namespace { static void printDebugLoc(const DebugLoc &DL, formatted_raw_ostream &OS) { @@ -132,38 +138,59 @@ static void diagnosticHandler(const DiagnosticInfo &DI, void *Context) { exit(1); } +static Expected<std::unique_ptr<Module>> openInputFile(LLVMContext &Context) { + if (MaterializeMetadata) { + ErrorOr<std::unique_ptr<MemoryBuffer>> MBOrErr = + MemoryBuffer::getFileOrSTDIN(InputFilename); + if (!MBOrErr) + return errorCodeToError(MBOrErr.getError()); + ErrorOr<std::unique_ptr<Module>> MOrErr = + getLazyBitcodeModule(std::move(*MBOrErr), Context, + /*ShouldLazyLoadMetadata=*/true); + if (!MOrErr) + return errorCodeToError(MOrErr.getError()); + (*MOrErr)->materializeMetadata(); + return std::move(*MOrErr); + } else { + std::string ErrorMessage; + std::unique_ptr<DataStreamer> Streamer = + getDataFileStreamer(InputFilename, &ErrorMessage); + if (!Streamer) + return make_error<StringError>(ErrorMessage, inconvertibleErrorCode()); + std::string DisplayFilename; + if (InputFilename == "-") + DisplayFilename = "<stdin>"; + else + DisplayFilename = InputFilename; + ErrorOr<std::unique_ptr<Module>> MOrErr = + getStreamedBitcodeModule(DisplayFilename, std::move(Streamer), Context); + (*MOrErr)->materializeAll(); + return std::move(*MOrErr); + } +} + int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); - LLVMContext &Context = getGlobalContext(); + LLVMContext Context; llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. Context.setDiagnosticHandler(diagnosticHandler, argv[0]); cl::ParseCommandLineOptions(argc, argv, "llvm .bc -> .ll disassembler\n"); - std::string ErrorMessage; - std::unique_ptr<Module> M; - - // Use the bitcode streaming interface - std::unique_ptr<DataStreamer> Streamer = - getDataFileStreamer(InputFilename, &ErrorMessage); - if (Streamer) { - std::string DisplayFilename; - if (InputFilename == "-") - DisplayFilename = "<stdin>"; - else - DisplayFilename = InputFilename; - ErrorOr<std::unique_ptr<Module>> MOrErr = - getStreamedBitcodeModule(DisplayFilename, std::move(Streamer), Context); - M = std::move(*MOrErr); - M->materializeAll(); - } else { - errs() << argv[0] << ": " << ErrorMessage << '\n'; + Expected<std::unique_ptr<Module>> MOrErr = openInputFile(Context); + if (!MOrErr) { + handleAllErrors(MOrErr.takeError(), [&](ErrorInfoBase &EIB) { + errs() << argv[0] << ": "; + EIB.log(errs()); + errs() << '\n'; + }); return 1; } + std::unique_ptr<Module> M = std::move(*MOrErr); // Just use stdout. We won't actually print anything on it. if (DontPrint) diff --git a/tools/llvm-dwarfdump/Makefile b/tools/llvm-dwarfdump/Makefile deleted file mode 100644 index 00b18c60bb2ad..0000000000000 --- a/tools/llvm-dwarfdump/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-dwarfdump/Makefile -----------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-dwarfdump -LINK_COMPONENTS := DebugInfoDWARF Object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-dwarfdump/fuzzer/llvm-dwarfdump-fuzzer.cpp b/tools/llvm-dwarfdump/fuzzer/llvm-dwarfdump-fuzzer.cpp index af0ac36522875..32e173f9d1fce 100644 --- a/tools/llvm-dwarfdump/fuzzer/llvm-dwarfdump-fuzzer.cpp +++ b/tools/llvm-dwarfdump/fuzzer/llvm-dwarfdump-fuzzer.cpp @@ -24,10 +24,12 @@ extern "C" void LLVMFuzzerTestOneInput(uint8_t *data, size_t size) { std::unique_ptr<MemoryBuffer> Buff = MemoryBuffer::getMemBuffer( StringRef((const char *)data, size), "", false); - ErrorOr<std::unique_ptr<ObjectFile>> ObjOrErr = + Expected<std::unique_ptr<ObjectFile>> ObjOrErr = ObjectFile::createObjectFile(Buff->getMemBufferRef()); - if (!ObjOrErr) + if (auto E = ObjOrErr.takeError()) { + consumeError(std::move(E)); return; + } ObjectFile &Obj = *ObjOrErr.get(); std::unique_ptr<DIContext> DICtx(new DWARFContextInMemory(Obj)); DICtx->dump(nulls(), DIDT_All); diff --git a/tools/llvm-dwarfdump/llvm-dwarfdump.cpp b/tools/llvm-dwarfdump/llvm-dwarfdump.cpp index eaacc7c5f21b7..4b3a011f861d1 100644 --- a/tools/llvm-dwarfdump/llvm-dwarfdump.cpp +++ b/tools/llvm-dwarfdump/llvm-dwarfdump.cpp @@ -29,7 +29,6 @@ #include "llvm/Support/raw_ostream.h" #include <algorithm> #include <cstring> -#include <list> #include <string> #include <system_error> @@ -96,16 +95,17 @@ static void DumpInput(StringRef Filename) { error(Filename, BuffOrErr.getError()); std::unique_ptr<MemoryBuffer> Buff = std::move(BuffOrErr.get()); - ErrorOr<std::unique_ptr<Binary>> BinOrErr = + Expected<std::unique_ptr<Binary>> BinOrErr = object::createBinary(Buff->getMemBufferRef()); - error(Filename, BinOrErr.getError()); + if (!BinOrErr) + error(Filename, errorToErrorCode(BinOrErr.takeError())); if (auto *Obj = dyn_cast<ObjectFile>(BinOrErr->get())) DumpObjectFile(*Obj, Filename); else if (auto *Fat = dyn_cast<MachOUniversalBinary>(BinOrErr->get())) for (auto &ObjForArch : Fat->objects()) { auto MachOOrErr = ObjForArch.getAsObjectFile(); - error(Filename, MachOOrErr.getError()); + error(Filename, errorToErrorCode(MachOOrErr.takeError())); DumpObjectFile(**MachOOrErr, Filename + " (" + ObjForArch.getArchTypeName() + ")"); } @@ -114,7 +114,7 @@ static void DumpInput(StringRef Filename) { /// If the input path is a .dSYM bundle (as created by the dsymutil tool), /// replace it with individual entries for each of the object files inside the /// bundle otherwise return the input path. -static std::vector<std::string> expandBundle(std::string InputPath) { +static std::vector<std::string> expandBundle(const std::string &InputPath) { std::vector<std::string> BundlePaths; SmallString<256> BundlePath(InputPath); // Manually open up the bundle to avoid introducing additional dependencies. @@ -146,7 +146,7 @@ static std::vector<std::string> expandBundle(std::string InputPath) { int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. @@ -158,7 +158,7 @@ int main(int argc, char **argv) { // Expand any .dSYM bundles to the individual object files contained therein. std::vector<std::string> Objects; - for (auto F : InputFilenames) { + for (const auto &F : InputFilenames) { auto Objs = expandBundle(F); Objects.insert(Objects.end(), Objs.begin(), Objs.end()); } diff --git a/tools/llvm-dwp/CMakeLists.txt b/tools/llvm-dwp/CMakeLists.txt index b29c00d49c3d3..740ab5ca9040e 100644 --- a/tools/llvm-dwp/CMakeLists.txt +++ b/tools/llvm-dwp/CMakeLists.txt @@ -10,4 +10,5 @@ set(LLVM_LINK_COMPONENTS add_llvm_tool(llvm-dwp llvm-dwp.cpp + DWPError.cpp ) diff --git a/tools/llvm-dwp/DWPError.cpp b/tools/llvm-dwp/DWPError.cpp new file mode 100644 index 0000000000000..21d53ed6d198c --- /dev/null +++ b/tools/llvm-dwp/DWPError.cpp @@ -0,0 +1,3 @@ +#include "DWPError.h" +using namespace llvm; +char DWPError::ID; diff --git a/tools/llvm-dwp/DWPError.h b/tools/llvm-dwp/DWPError.h new file mode 100644 index 0000000000000..62025ed4caa55 --- /dev/null +++ b/tools/llvm-dwp/DWPError.h @@ -0,0 +1,23 @@ +#ifndef TOOLS_LLVM_DWP_DWPERROR +#define TOOLS_LLVM_DWP_DWPERROR + +#include "llvm/Support/Error.h" +#include "llvm/Support/ErrorHandling.h" +#include <string> + +namespace llvm { +class DWPError : public ErrorInfo<DWPError> { +public: + DWPError(std::string Info) : Info(std::move(Info)) {} + void log(raw_ostream &OS) const override { OS << Info; } + std::error_code convertToErrorCode() const override { + llvm_unreachable("Not implemented"); + } + static char ID; + +private: + std::string Info; +}; +} + +#endif diff --git a/tools/llvm-dwp/DWPStringPool.h b/tools/llvm-dwp/DWPStringPool.h new file mode 100644 index 0000000000000..7d41176b56197 --- /dev/null +++ b/tools/llvm-dwp/DWPStringPool.h @@ -0,0 +1,56 @@ +#ifndef TOOLS_LLVM_DWP_DWPSTRINGPOOL +#define TOOLS_LLVM_DWP_DWPSTRINGPOOL + +#include "llvm/ADT/DenseMap.h" +#include "llvm/MC/MCSection.h" +#include "llvm/MC/MCStreamer.h" +#include <cassert> + +namespace llvm { +class DWPStringPool { + + struct CStrDenseMapInfo { + static inline const char *getEmptyKey() { + return reinterpret_cast<const char *>(~static_cast<uintptr_t>(0)); + } + static inline const char *getTombstoneKey() { + return reinterpret_cast<const char *>(~static_cast<uintptr_t>(1)); + } + static unsigned getHashValue(const char *Val) { + assert(Val != getEmptyKey() && "Cannot hash the empty key!"); + assert(Val != getTombstoneKey() && "Cannot hash the tombstone key!"); + return (unsigned)hash_value(StringRef(Val)); + } + static bool isEqual(const char *LHS, const char *RHS) { + if (RHS == getEmptyKey()) + return LHS == getEmptyKey(); + if (RHS == getTombstoneKey()) + return LHS == getTombstoneKey(); + return strcmp(LHS, RHS) == 0; + } + }; + + MCStreamer &Out; + MCSection *Sec; + DenseMap<const char *, uint32_t, CStrDenseMapInfo> Pool; + uint32_t Offset = 0; + +public: + DWPStringPool(MCStreamer &Out, MCSection *Sec) : Out(Out), Sec(Sec) {} + + uint32_t getOffset(const char *Str, unsigned Length) { + assert(strlen(Str) + 1 == Length && "Ensure length hint is correct"); + + auto Pair = Pool.insert(std::make_pair(Str, Offset)); + if (Pair.second) { + Out.SwitchSection(Sec); + Out.EmitBytes(StringRef(Str, Length)); + Offset += Length; + } + + return Pair.first->second; + } +}; +} + +#endif diff --git a/tools/llvm-dwp/Makefile b/tools/llvm-dwp/Makefile deleted file mode 100644 index 826371ecf915f..0000000000000 --- a/tools/llvm-dwp/Makefile +++ /dev/null @@ -1,18 +0,0 @@ -##===- tools/llvm-dwp/Makefile -----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-dwp -LINK_COMPONENTS := all-targets AsmPrinter DebugInfoDWARF MC Object Support - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common - diff --git a/tools/llvm-dwp/llvm-dwp.cpp b/tools/llvm-dwp/llvm-dwp.cpp index 570854b849c17..7bd3d3f3f4d08 100644 --- a/tools/llvm-dwp/llvm-dwp.cpp +++ b/tools/llvm-dwp/llvm-dwp.cpp @@ -1,3 +1,19 @@ +//===-- llvm-dwp.cpp - Split DWARF merging tool for llvm ------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// A utility for merging DWARF 5 Split DWARF .dwo files into .dwp (DWARF +// package files). +// +//===----------------------------------------------------------------------===// +#include "DWPError.h" +#include "DWPStringPool.h" +#include "llvm/ADT/MapVector.h" #include "llvm/ADT/STLExtras.h" #include "llvm/ADT/StringSet.h" #include "llvm/CodeGen/AsmPrinter.h" @@ -12,7 +28,9 @@ #include "llvm/MC/MCStreamer.h" #include "llvm/MC/MCTargetOptionsCommandFlags.h" #include "llvm/Object/ObjectFile.h" +#include "llvm/Support/Compression.h" #include "llvm/Support/DataExtractor.h" +#include "llvm/Support/Error.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/MathExtras.h" #include "llvm/Support/MemoryBuffer.h" @@ -21,9 +39,9 @@ #include "llvm/Support/TargetSelect.h" #include "llvm/Support/raw_ostream.h" #include "llvm/Target/TargetMachine.h" -#include <list> +#include <deque> +#include <iostream> #include <memory> -#include <unordered_set> using namespace llvm; using namespace llvm::object; @@ -38,21 +56,14 @@ static opt<std::string> OutputFilename(Required, "o", value_desc("filename"), cat(DwpCategory)); -static int error(const Twine &Error, const Twine &Context) { - errs() << Twine("while processing ") + Context + ":\n"; - errs() << Twine("error: ") + Error + "\n"; - return 1; -} - -static std::error_code -writeStringsAndOffsets(MCStreamer &Out, StringMap<uint32_t> &Strings, - uint32_t &StringOffset, MCSection *StrSection, - MCSection *StrOffsetSection, StringRef CurStrSection, - StringRef CurStrOffsetSection) { +static void writeStringsAndOffsets(MCStreamer &Out, DWPStringPool &Strings, + MCSection *StrOffsetSection, + StringRef CurStrSection, + StringRef CurStrOffsetSection) { // Could possibly produce an error or warning if one of these was non-null but // the other was null. if (CurStrSection.empty() || CurStrOffsetSection.empty()) - return std::error_code(); + return; DenseMap<uint32_t, uint32_t> OffsetRemapping; @@ -60,15 +71,8 @@ writeStringsAndOffsets(MCStreamer &Out, StringMap<uint32_t> &Strings, uint32_t LocalOffset = 0; uint32_t PrevOffset = 0; while (const char *s = Data.getCStr(&LocalOffset)) { - StringRef Str(s, LocalOffset - PrevOffset - 1); - auto Pair = Strings.insert(std::make_pair(Str, StringOffset)); - if (Pair.second) { - Out.SwitchSection(StrSection); - Out.EmitBytes( - StringRef(Pair.first->getKeyData(), Pair.first->getKeyLength() + 1)); - StringOffset += Str.size() + 1; - } - OffsetRemapping[PrevOffset] = Pair.first->second; + OffsetRemapping[PrevOffset] = + Strings.getOffset(s, LocalOffset - PrevOffset); PrevOffset = LocalOffset; } @@ -83,8 +87,6 @@ writeStringsAndOffsets(MCStreamer &Out, StringMap<uint32_t> &Strings, auto NewOffset = OffsetRemapping[OldOffset]; Out.EmitIntValue(NewOffset, 4); } - - return std::error_code(); } static uint32_t getCUAbbrev(StringRef Abbrev, uint64_t AbbrCode) { @@ -103,7 +105,32 @@ static uint32_t getCUAbbrev(StringRef Abbrev, uint64_t AbbrCode) { return Offset; } -static uint64_t getCUSignature(StringRef Abbrev, StringRef Info) { +struct CompileUnitIdentifiers { + uint64_t Signature = 0; + const char *Name = ""; + const char *DWOName = ""; +}; + +static Expected<const char *> +getIndexedString(uint32_t Form, DataExtractor InfoData, uint32_t &InfoOffset, + StringRef StrOffsets, StringRef Str) { + if (Form == dwarf::DW_FORM_string) + return InfoData.getCStr(&InfoOffset); + if (Form != dwarf::DW_FORM_GNU_str_index) + return make_error<DWPError>( + "string field encoded without DW_FORM_string or DW_FORM_GNU_str_index"); + auto StrIndex = InfoData.getULEB128(&InfoOffset); + DataExtractor StrOffsetsData(StrOffsets, true, 0); + uint32_t StrOffsetsOffset = 4 * StrIndex; + uint32_t StrOffset = StrOffsetsData.getU32(&StrOffsetsOffset); + DataExtractor StrData(Str, true, 0); + return StrData.getCStr(&StrOffset); +} + +static Expected<CompileUnitIdentifiers> getCUIdentifiers(StringRef Abbrev, + StringRef Info, + StringRef StrOffsets, + StringRef Str) { uint32_t Offset = 0; DataExtractor InfoData(Info, true, 0); InfoData.getU32(&Offset); // Length @@ -116,79 +143,141 @@ static uint64_t getCUSignature(StringRef Abbrev, StringRef Info) { DataExtractor AbbrevData(Abbrev, true, 0); uint32_t AbbrevOffset = getCUAbbrev(Abbrev, AbbrCode); uint64_t Tag = AbbrevData.getULEB128(&AbbrevOffset); - (void)Tag; - // FIXME: Real error handling - assert(Tag == dwarf::DW_TAG_compile_unit); + if (Tag != dwarf::DW_TAG_compile_unit) + return make_error<DWPError>("top level DIE is not a compile unit"); // DW_CHILDREN AbbrevData.getU8(&AbbrevOffset); uint32_t Name; uint32_t Form; + CompileUnitIdentifiers ID; while ((Name = AbbrevData.getULEB128(&AbbrevOffset)) | (Form = AbbrevData.getULEB128(&AbbrevOffset)) && - Name != dwarf::DW_AT_GNU_dwo_id) { - DWARFFormValue::skipValue(Form, InfoData, &Offset, Version, AddrSize); + (Name != 0 || Form != 0)) { + switch (Name) { + case dwarf::DW_AT_name: { + Expected<const char *> EName = + getIndexedString(Form, InfoData, Offset, StrOffsets, Str); + if (!EName) + return EName.takeError(); + ID.Name = *EName; + break; + } + case dwarf::DW_AT_GNU_dwo_name: { + Expected<const char *> EName = + getIndexedString(Form, InfoData, Offset, StrOffsets, Str); + if (!EName) + return EName.takeError(); + ID.DWOName = *EName; + break; + } + case dwarf::DW_AT_GNU_dwo_id: + ID.Signature = InfoData.getU64(&Offset); + break; + default: + DWARFFormValue::skipValue(Form, InfoData, &Offset, Version, AddrSize); + } } - // FIXME: Real error handling - assert(Name == dwarf::DW_AT_GNU_dwo_id); - return InfoData.getU64(&Offset); + return ID; } struct UnitIndexEntry { - uint64_t Signature; DWARFUnitIndex::Entry::SectionContribution Contributions[8]; + std::string Name; + std::string DWOName; + StringRef DWPName; }; -static void addAllTypes(MCStreamer &Out, - std::vector<UnitIndexEntry> &TypeIndexEntries, - MCSection *OutputTypes, StringRef Types, - const UnitIndexEntry &CUEntry, uint32_t &TypesOffset) { - if (Types.empty()) - return; +static StringRef getSubsection(StringRef Section, + const DWARFUnitIndex::Entry &Entry, + DWARFSectionKind Kind) { + const auto *Off = Entry.getOffset(Kind); + if (!Off) + return StringRef(); + return Section.substr(Off->Offset, Off->Length); +} +static void addAllTypesFromDWP( + MCStreamer &Out, MapVector<uint64_t, UnitIndexEntry> &TypeIndexEntries, + const DWARFUnitIndex &TUIndex, MCSection *OutputTypes, StringRef Types, + const UnitIndexEntry &TUEntry, uint32_t &TypesOffset) { Out.SwitchSection(OutputTypes); - uint32_t Offset = 0; - DataExtractor Data(Types, true, 0); - while (Data.isValidOffset(Offset)) { - UnitIndexEntry Entry = CUEntry; + for (const DWARFUnitIndex::Entry &E : TUIndex.getRows()) { + auto *I = E.getOffsets(); + if (!I) + continue; + auto P = TypeIndexEntries.insert(std::make_pair(E.getSignature(), TUEntry)); + if (!P.second) + continue; + auto &Entry = P.first->second; // Zero out the debug_info contribution Entry.Contributions[0] = {}; + for (auto Kind : TUIndex.getColumnKinds()) { + auto &C = Entry.Contributions[Kind - DW_SECT_INFO]; + C.Offset += I->Offset; + C.Length = I->Length; + ++I; + } auto &C = Entry.Contributions[DW_SECT_TYPES - DW_SECT_INFO]; + Out.EmitBytes(Types.substr( + C.Offset - TUEntry.Contributions[DW_SECT_TYPES - DW_SECT_INFO].Offset, + C.Length)); C.Offset = TypesOffset; - auto PrevOffset = Offset; - // Length of the unit, including the 4 byte length field. - C.Length = Data.getU32(&Offset) + 4; - - Data.getU16(&Offset); // Version - Data.getU32(&Offset); // Abbrev offset - Data.getU8(&Offset); // Address size - Entry.Signature = Data.getU64(&Offset); - Offset = PrevOffset + C.Length; - - if (any_of(TypeIndexEntries, [&](const UnitIndexEntry &E) { - return E.Signature == Entry.Signature; - })) - continue; - - Out.EmitBytes(Types.substr(PrevOffset, C.Length)); TypesOffset += C.Length; + } +} - TypeIndexEntries.push_back(Entry); +static void addAllTypes(MCStreamer &Out, + MapVector<uint64_t, UnitIndexEntry> &TypeIndexEntries, + MCSection *OutputTypes, + const std::vector<StringRef> &TypesSections, + const UnitIndexEntry &CUEntry, uint32_t &TypesOffset) { + for (StringRef Types : TypesSections) { + Out.SwitchSection(OutputTypes); + uint32_t Offset = 0; + DataExtractor Data(Types, true, 0); + while (Data.isValidOffset(Offset)) { + UnitIndexEntry Entry = CUEntry; + // Zero out the debug_info contribution + Entry.Contributions[0] = {}; + auto &C = Entry.Contributions[DW_SECT_TYPES - DW_SECT_INFO]; + C.Offset = TypesOffset; + auto PrevOffset = Offset; + // Length of the unit, including the 4 byte length field. + C.Length = Data.getU32(&Offset) + 4; + + Data.getU16(&Offset); // Version + Data.getU32(&Offset); // Abbrev offset + Data.getU8(&Offset); // Address size + auto Signature = Data.getU64(&Offset); + Offset = PrevOffset + C.Length; + + auto P = TypeIndexEntries.insert(std::make_pair(Signature, Entry)); + if (!P.second) + continue; + + Out.EmitBytes(Types.substr(PrevOffset, C.Length)); + TypesOffset += C.Length; + } } } static void writeIndexTable(MCStreamer &Out, ArrayRef<unsigned> ContributionOffsets, - ArrayRef<UnitIndexEntry> IndexEntries, + const MapVector<uint64_t, UnitIndexEntry> &IndexEntries, uint32_t DWARFUnitIndex::Entry::SectionContribution::*Field) { for (const auto &E : IndexEntries) - for (size_t i = 0; i != array_lengthof(E.Contributions); ++i) + for (size_t i = 0; i != array_lengthof(E.second.Contributions); ++i) if (ContributionOffsets[i]) - Out.EmitIntValue(E.Contributions[i].*Field, 4); + Out.EmitIntValue(E.second.Contributions[i].*Field, 4); } -static void writeIndex(MCStreamer &Out, MCSection *Section, - ArrayRef<unsigned> ContributionOffsets, - ArrayRef<UnitIndexEntry> IndexEntries) { +static void +writeIndex(MCStreamer &Out, MCSection *Section, + ArrayRef<unsigned> ContributionOffsets, + const MapVector<uint64_t, UnitIndexEntry> &IndexEntries) { + if (IndexEntries.empty()) + return; + unsigned Columns = 0; for (auto &C : ContributionOffsets) if (C) @@ -196,15 +285,18 @@ static void writeIndex(MCStreamer &Out, MCSection *Section, std::vector<unsigned> Buckets(NextPowerOf2(3 * IndexEntries.size() / 2)); uint64_t Mask = Buckets.size() - 1; - for (size_t i = 0; i != IndexEntries.size(); ++i) { - auto S = IndexEntries[i].Signature; + size_t i = 0; + for (const auto &P : IndexEntries) { + auto S = P.first; auto H = S & Mask; + auto HP = ((S >> 32) & Mask) | 1; while (Buckets[H]) { - assert(S != IndexEntries[Buckets[H] - 1].Signature && - "Duplicate type unit"); - H += ((S >> 32) & Mask) | 1; + assert(S != IndexEntries.begin()[Buckets[H] - 1].first && + "Duplicate unit"); + H = (H + HP) & Mask; } Buckets[H] = i + 1; + ++i; } Out.SwitchSection(Section); @@ -215,7 +307,7 @@ static void writeIndex(MCStreamer &Out, MCSection *Section, // Write the signatures. for (const auto &I : Buckets) - Out.EmitIntValue(I ? IndexEntries[I - 1].Signature : 0, 8); + Out.EmitIntValue(I ? IndexEntries.begin()[I - 1].first : 0, 8); // Write the indexes. for (const auto &I : Buckets) @@ -234,11 +326,149 @@ static void writeIndex(MCStreamer &Out, MCSection *Section, writeIndexTable(Out, ContributionOffsets, IndexEntries, &DWARFUnitIndex::Entry::SectionContribution::Length); } -static std::error_code write(MCStreamer &Out, ArrayRef<std::string> Inputs) { +static bool consumeCompressedDebugSectionHeader(StringRef &data, + uint64_t &OriginalSize) { + // Consume "ZLIB" prefix. + if (!data.startswith("ZLIB")) + return false; + data = data.substr(4); + // Consume uncompressed section size (big-endian 8 bytes). + DataExtractor extractor(data, false, 8); + uint32_t Offset = 0; + OriginalSize = extractor.getU64(&Offset); + if (Offset == 0) + return false; + data = data.substr(Offset); + return true; +} + +std::string buildDWODescription(StringRef Name, StringRef DWPName, StringRef DWOName) { + std::string Text = "\'"; + Text += Name; + Text += '\''; + if (!DWPName.empty()) { + Text += " (from "; + if (!DWOName.empty()) { + Text += '\''; + Text += DWOName; + Text += "' in "; + } + Text += '\''; + Text += DWPName; + Text += "')"; + } + return Text; +} + +static Error handleCompressedSection( + std::deque<SmallString<32>> &UncompressedSections, StringRef &Name, + StringRef &Contents) { + if (!Name.startswith("zdebug_")) + return Error(); + UncompressedSections.emplace_back(); + uint64_t OriginalSize; + if (!zlib::isAvailable()) + return make_error<DWPError>("zlib not available"); + if (!consumeCompressedDebugSectionHeader(Contents, OriginalSize) || + zlib::uncompress(Contents, UncompressedSections.back(), OriginalSize) != + zlib::StatusOK) + return make_error<DWPError>( + ("failure while decompressing compressed section: '" + Name + "\'") + .str()); + Name = Name.substr(1); + Contents = UncompressedSections.back(); + return Error(); +} + +static Error handleSection( + const StringMap<std::pair<MCSection *, DWARFSectionKind>> &KnownSections, + const MCSection *StrSection, const MCSection *StrOffsetSection, + const MCSection *TypesSection, const MCSection *CUIndexSection, + const MCSection *TUIndexSection, const SectionRef &Section, MCStreamer &Out, + std::deque<SmallString<32>> &UncompressedSections, + uint32_t (&ContributionOffsets)[8], UnitIndexEntry &CurEntry, + StringRef &CurStrSection, StringRef &CurStrOffsetSection, + std::vector<StringRef> &CurTypesSection, StringRef &InfoSection, + StringRef &AbbrevSection, StringRef &CurCUIndexSection, + StringRef &CurTUIndexSection) { + if (Section.isBSS()) + return Error(); + + if (Section.isVirtual()) + return Error(); + + StringRef Name; + if (std::error_code Err = Section.getName(Name)) + return errorCodeToError(Err); + + Name = Name.substr(Name.find_first_not_of("._")); + + StringRef Contents; + if (auto Err = Section.getContents(Contents)) + return errorCodeToError(Err); + + if (auto Err = handleCompressedSection(UncompressedSections, Name, Contents)) + return Err; + + auto SectionPair = KnownSections.find(Name); + if (SectionPair == KnownSections.end()) + return Error(); + + if (DWARFSectionKind Kind = SectionPair->second.second) { + auto Index = Kind - DW_SECT_INFO; + if (Kind != DW_SECT_TYPES) { + CurEntry.Contributions[Index].Offset = ContributionOffsets[Index]; + ContributionOffsets[Index] += + (CurEntry.Contributions[Index].Length = Contents.size()); + } + + switch (Kind) { + case DW_SECT_INFO: + InfoSection = Contents; + break; + case DW_SECT_ABBREV: + AbbrevSection = Contents; + break; + default: + break; + } + } + + MCSection *OutSection = SectionPair->second.first; + if (OutSection == StrOffsetSection) + CurStrOffsetSection = Contents; + else if (OutSection == StrSection) + CurStrSection = Contents; + else if (OutSection == TypesSection) + CurTypesSection.push_back(Contents); + else if (OutSection == CUIndexSection) + CurCUIndexSection = Contents; + else if (OutSection == TUIndexSection) + CurTUIndexSection = Contents; + else { + Out.SwitchSection(OutSection); + Out.EmitBytes(Contents); + } + return Error(); +} + +static Error +buildDuplicateError(const std::pair<uint64_t, UnitIndexEntry> &PrevE, + const CompileUnitIdentifiers &ID, StringRef DWPName) { + return make_error<DWPError>( + std::string("Duplicate DWO ID (") + utohexstr(PrevE.first) + ") in " + + buildDWODescription(PrevE.second.Name, PrevE.second.DWPName, + PrevE.second.DWOName) + + " and " + buildDWODescription(ID.Name, DWPName, ID.DWOName)); +} + +static Error write(MCStreamer &Out, ArrayRef<std::string> Inputs) { const auto &MCOFI = *Out.getContext().getObjectFileInfo(); MCSection *const StrSection = MCOFI.getDwarfStrDWOSection(); MCSection *const StrOffsetSection = MCOFI.getDwarfStrOffDWOSection(); MCSection *const TypesSection = MCOFI.getDwarfTypesDWOSection(); + MCSection *const CUIndexSection = MCOFI.getDwarfCUIndexSection(); + MCSection *const TUIndexSection = MCOFI.getDwarfTUIndexSection(); const StringMap<std::pair<MCSection *, DWARFSectionKind>> KnownSections = { {"debug_info.dwo", {MCOFI.getDwarfInfoDWOSection(), DW_SECT_INFO}}, {"debug_types.dwo", {MCOFI.getDwarfTypesDWOSection(), DW_SECT_TYPES}}, @@ -246,96 +476,121 @@ static std::error_code write(MCStreamer &Out, ArrayRef<std::string> Inputs) { {"debug_str.dwo", {StrSection, static_cast<DWARFSectionKind>(0)}}, {"debug_loc.dwo", {MCOFI.getDwarfLocDWOSection(), DW_SECT_LOC}}, {"debug_line.dwo", {MCOFI.getDwarfLineDWOSection(), DW_SECT_LINE}}, - {"debug_abbrev.dwo", {MCOFI.getDwarfAbbrevDWOSection(), DW_SECT_ABBREV}}}; - - std::vector<UnitIndexEntry> IndexEntries; - std::vector<UnitIndexEntry> TypeIndexEntries; + {"debug_abbrev.dwo", {MCOFI.getDwarfAbbrevDWOSection(), DW_SECT_ABBREV}}, + {"debug_cu_index", {CUIndexSection, static_cast<DWARFSectionKind>(0)}}, + {"debug_tu_index", {TUIndexSection, static_cast<DWARFSectionKind>(0)}}}; - StringMap<uint32_t> Strings; - uint32_t StringOffset = 0; + MapVector<uint64_t, UnitIndexEntry> IndexEntries; + MapVector<uint64_t, UnitIndexEntry> TypeIndexEntries; uint32_t ContributionOffsets[8] = {}; + DWPStringPool Strings(Out, StrSection); + + SmallVector<OwningBinary<object::ObjectFile>, 128> Objects; + Objects.reserve(Inputs.size()); + + std::deque<SmallString<32>> UncompressedSections; + for (const auto &Input : Inputs) { auto ErrOrObj = object::ObjectFile::createObjectFile(Input); if (!ErrOrObj) - return ErrOrObj.getError(); + return ErrOrObj.takeError(); - IndexEntries.emplace_back(); - UnitIndexEntry &CurEntry = IndexEntries.back(); + auto &Obj = *ErrOrObj->getBinary(); + Objects.push_back(std::move(*ErrOrObj)); + + UnitIndexEntry CurEntry = {}; StringRef CurStrSection; StringRef CurStrOffsetSection; - StringRef CurTypesSection; + std::vector<StringRef> CurTypesSection; StringRef InfoSection; StringRef AbbrevSection; - - for (const auto &Section : ErrOrObj->getBinary()->sections()) { - StringRef Name; - if (std::error_code Err = Section.getName(Name)) + StringRef CurCUIndexSection; + StringRef CurTUIndexSection; + + for (const auto &Section : Obj.sections()) + if (auto Err = handleSection( + KnownSections, StrSection, StrOffsetSection, TypesSection, + CUIndexSection, TUIndexSection, Section, Out, + UncompressedSections, ContributionOffsets, CurEntry, + CurStrSection, CurStrOffsetSection, CurTypesSection, InfoSection, + AbbrevSection, CurCUIndexSection, CurTUIndexSection)) return Err; - auto SectionPair = - KnownSections.find(Name.substr(Name.find_first_not_of("._"))); - if (SectionPair == KnownSections.end()) - continue; + if (InfoSection.empty()) + continue; - StringRef Contents; - if (auto Err = Section.getContents(Contents)) - return Err; + writeStringsAndOffsets(Out, Strings, StrOffsetSection, CurStrSection, + CurStrOffsetSection); + + if (CurCUIndexSection.empty()) { + Expected<CompileUnitIdentifiers> EID = getCUIdentifiers( + AbbrevSection, InfoSection, CurStrOffsetSection, CurStrSection); + if (!EID) + return EID.takeError(); + const auto &ID = *EID; + auto P = IndexEntries.insert(std::make_pair(ID.Signature, CurEntry)); + if (!P.second) + return buildDuplicateError(*P.first, ID, ""); + P.first->second.Name = ID.Name; + P.first->second.DWOName = ID.DWOName; + addAllTypes(Out, TypeIndexEntries, TypesSection, CurTypesSection, + CurEntry, ContributionOffsets[DW_SECT_TYPES - DW_SECT_INFO]); + continue; + } - if (DWARFSectionKind Kind = SectionPair->second.second) { - auto Index = Kind - DW_SECT_INFO; - if (Kind != DW_SECT_TYPES) { - CurEntry.Contributions[Index].Offset = ContributionOffsets[Index]; - ContributionOffsets[Index] += - (CurEntry.Contributions[Index].Length = Contents.size()); - } - - switch (Kind) { - case DW_SECT_INFO: - InfoSection = Contents; - break; - case DW_SECT_ABBREV: - AbbrevSection = Contents; - break; - default: - break; - } - } + DWARFUnitIndex CUIndex(DW_SECT_INFO); + DataExtractor CUIndexData(CurCUIndexSection, Obj.isLittleEndian(), 0); + if (!CUIndex.parse(CUIndexData)) + return make_error<DWPError>("Failed to parse cu_index"); - MCSection *OutSection = SectionPair->second.first; - if (OutSection == StrOffsetSection) - CurStrOffsetSection = Contents; - else if (OutSection == StrSection) - CurStrSection = Contents; - else if (OutSection == TypesSection) - CurTypesSection = Contents; - else { - Out.SwitchSection(OutSection); - Out.EmitBytes(Contents); + for (const DWARFUnitIndex::Entry &E : CUIndex.getRows()) { + auto *I = E.getOffsets(); + if (!I) + continue; + auto P = IndexEntries.insert(std::make_pair(E.getSignature(), CurEntry)); + Expected<CompileUnitIdentifiers> EID = getCUIdentifiers( + getSubsection(AbbrevSection, E, DW_SECT_ABBREV), + getSubsection(InfoSection, E, DW_SECT_INFO), + getSubsection(CurStrOffsetSection, E, DW_SECT_STR_OFFSETS), + CurStrSection); + if (!EID) + return EID.takeError(); + const auto &ID = *EID; + if (!P.second) + return buildDuplicateError(*P.first, ID, Input); + auto &NewEntry = P.first->second; + NewEntry.Name = ID.Name; + NewEntry.DWOName = ID.DWOName; + NewEntry.DWPName = Input; + for (auto Kind : CUIndex.getColumnKinds()) { + auto &C = NewEntry.Contributions[Kind - DW_SECT_INFO]; + C.Offset += I->Offset; + C.Length = I->Length; + ++I; } } - assert(!AbbrevSection.empty()); - assert(!InfoSection.empty()); - CurEntry.Signature = getCUSignature(AbbrevSection, InfoSection); - addAllTypes(Out, TypeIndexEntries, TypesSection, CurTypesSection, CurEntry, - ContributionOffsets[DW_SECT_TYPES - DW_SECT_INFO]); - - if (auto Err = writeStringsAndOffsets(Out, Strings, StringOffset, - StrSection, StrOffsetSection, - CurStrSection, CurStrOffsetSection)) - return Err; + if (!CurTypesSection.empty()) { + if (CurTypesSection.size() != 1) + return make_error<DWPError>("multiple type unit sections in .dwp file"); + DWARFUnitIndex TUIndex(DW_SECT_TYPES); + DataExtractor TUIndexData(CurTUIndexSection, Obj.isLittleEndian(), 0); + if (!TUIndex.parse(TUIndexData)) + return make_error<DWPError>("Failed to parse tu_index"); + addAllTypesFromDWP(Out, TypeIndexEntries, TUIndex, TypesSection, + CurTypesSection.front(), CurEntry, + ContributionOffsets[DW_SECT_TYPES - DW_SECT_INFO]); + } } - if (!TypeIndexEntries.empty()) { - // Lie about there being no info contributions so the TU index only includes - // the type unit contribution - ContributionOffsets[0] = 0; - writeIndex(Out, MCOFI.getDwarfTUIndexSection(), ContributionOffsets, - TypeIndexEntries); - } + // Lie about there being no info contributions so the TU index only includes + // the type unit contribution + ContributionOffsets[0] = 0; + writeIndex(Out, MCOFI.getDwarfTUIndexSection(), ContributionOffsets, + TypeIndexEntries); // Lie about the type contribution ContributionOffsets[DW_SECT_TYPES - DW_SECT_INFO] = 0; @@ -345,7 +600,13 @@ static std::error_code write(MCStreamer &Out, ArrayRef<std::string> Inputs) { writeIndex(Out, MCOFI.getDwarfCUIndexSection(), ContributionOffsets, IndexEntries); - return std::error_code(); + return Error(); +} + +static int error(const Twine &Error, const Twine &Context) { + errs() << Twine("while processing ") + Context + ":\n"; + errs() << Twine("error: ") + Error + "\n"; + return 1; } int main(int argc, char **argv) { @@ -380,7 +641,7 @@ int main(int argc, char **argv) { MCObjectFileInfo MOFI; MCContext MC(MAI.get(), MRI.get(), &MOFI); - MOFI.InitMCObjectFileInfo(TheTriple, Reloc::Default, CodeModel::Default, MC); + MOFI.InitMCObjectFileInfo(TheTriple, /*PIC*/ false, CodeModel::Default, MC); auto MAB = TheTarget->createMCAsmBackend(*MRI, TripleName, ""); if (!MAB) @@ -413,8 +674,10 @@ int main(int argc, char **argv) { if (!MS) return error("no object streamer for target " + TripleName, Context); - if (auto Err = write(*MS, InputFiles)) - return error(Err.message(), "Writing DWP file"); + if (auto Err = write(*MS, InputFiles)) { + logAllUnhandledErrors(std::move(Err), errs(), "error: "); + return 1; + } MS->Finish(); } diff --git a/tools/llvm-extract/Makefile b/tools/llvm-extract/Makefile deleted file mode 100644 index d371c54759263..0000000000000 --- a/tools/llvm-extract/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-extract/Makefile -------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-extract -LINK_COMPONENTS := ipo bitreader bitwriter asmparser irreader - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-extract/llvm-extract.cpp b/tools/llvm-extract/llvm-extract.cpp index 1da456d33f52d..900c461ebd861 100644 --- a/tools/llvm-extract/llvm-extract.cpp +++ b/tools/llvm-extract/llvm-extract.cpp @@ -102,10 +102,10 @@ static cl::opt<bool> PreserveAssemblyUseListOrder( int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); - LLVMContext &Context = getGlobalContext(); + LLVMContext Context; llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. cl::ParseCommandLineOptions(argc, argv, "llvm extractor\n"); diff --git a/tools/llvm-go/Makefile b/tools/llvm-go/Makefile deleted file mode 100644 index 4465b2a4a9dbd..0000000000000 --- a/tools/llvm-go/Makefile +++ /dev/null @@ -1,16 +0,0 @@ -##===- tools/llvm-go/Makefile ------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -include $(LEVEL)/Makefile.common - -all:: $(ToolDir)/llvm-go$(EXEEXT) - -$(ToolDir)/llvm-go$(EXEEXT): $(PROJ_SRC_DIR)/llvm-go.go - $(GO) build -o $@ $< diff --git a/tools/llvm-go/llvm-go.go b/tools/llvm-go/llvm-go.go index ed79ca67b89d3..12f0d73b4cc48 100644 --- a/tools/llvm-go/llvm-go.go +++ b/tools/llvm-go/llvm-go.go @@ -88,17 +88,9 @@ func llvmConfig(args ...string) string { return outstr } -func llvmFlags(linkmode string) compilerFlags { - ldflags := llvmConfig("--ldflags") - switch linkmode { - case linkmodeComponentLibs: - ldflags += " " + llvmConfig(append([]string{"--libs"}, components...)...) - case linkmodeDylib: - ldflags += " -lLLVM" - default: - panic("invalid linkmode: " + linkmode) - } - ldflags += " " + llvmConfig("--system-libs") +func llvmFlags() compilerFlags { + args := append([]string{"--ldflags", "--libs", "--system-libs"}, components...) + ldflags := llvmConfig(args...) if runtime.GOOS != "darwin" { // OS X doesn't like -rpath with cgo. See: // https://code.google.com/p/go/issues/detail?id=7293 @@ -133,8 +125,8 @@ func printComponents() { fmt.Println(strings.Join(components, " ")) } -func printConfig(linkmode string) { - flags := llvmFlags(linkmode) +func printConfig() { + flags := llvmFlags() fmt.Printf(`// +build !byollvm @@ -153,7 +145,7 @@ type (run_build_sh int) `, flags.cpp, flags.cxx, flags.ld) } -func runGoWithLLVMEnv(args []string, cc, cxx, gocmd, llgo, cppflags, cxxflags, ldflags, linkmode string) { +func runGoWithLLVMEnv(args []string, cc, cxx, gocmd, llgo, cppflags, cxxflags, ldflags string) { args = addTag(args, "byollvm") srcdir := llvmConfig("--src-root") @@ -182,7 +174,7 @@ func runGoWithLLVMEnv(args []string, cc, cxx, gocmd, llgo, cppflags, cxxflags, l newgopathlist = append(newgopathlist, filepath.SplitList(os.Getenv("GOPATH"))...) newgopath := strings.Join(newgopathlist, string(filepath.ListSeparator)) - flags := llvmFlags(linkmode) + flags := llvmFlags() newenv := []string{ "CC=" + cc, @@ -250,7 +242,6 @@ func main() { ldflags := os.Getenv("CGO_LDFLAGS") gocmd := "go" llgo := "" - linkmode := linkmodeComponentLibs flags := []struct { name string @@ -262,7 +253,6 @@ func main() { {"llgo", &llgo}, {"cppflags", &cppflags}, {"ldflags", &ldflags}, - {"linkmode", &linkmode}, } args := os.Args[1:] @@ -283,11 +273,11 @@ LOOP: switch args[0] { case "build", "get", "install", "run", "test": - runGoWithLLVMEnv(args, cc, cxx, gocmd, llgo, cppflags, cxxflags, ldflags, linkmode) + runGoWithLLVMEnv(args, cc, cxx, gocmd, llgo, cppflags, cxxflags, ldflags) case "print-components": printComponents() case "print-config": - printConfig(linkmode) + printConfig() default: usage() } diff --git a/tools/llvm-jitlistener/Makefile b/tools/llvm-jitlistener/Makefile deleted file mode 100644 index 4acb6a5617157..0000000000000 --- a/tools/llvm-jitlistener/Makefile +++ /dev/null @@ -1,27 +0,0 @@ -##===- tools/llvm-jitlistener/Makefile ---------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-jitlistener - -include $(LEVEL)/Makefile.config - -LINK_COMPONENTS := mcjit interpreter nativecodegen bitreader asmparser irreader selectiondag Object - -# If Intel JIT Events support is configured, link against the LLVM Intel JIT -# Events interface library. If not, this tool will do nothing useful, but it -# will build correctly. -ifeq ($(USE_INTEL_JITEVENTS), 1) - LINK_COMPONENTS += debuginfodwarf inteljitevents -endif - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LLVM_SRC_ROOT)/Makefile.rules diff --git a/tools/llvm-jitlistener/llvm-jitlistener.cpp b/tools/llvm-jitlistener/llvm-jitlistener.cpp index af1a59bdbd377..6b72c17b1fede 100644 --- a/tools/llvm-jitlistener/llvm-jitlistener.cpp +++ b/tools/llvm-jitlistener/llvm-jitlistener.cpp @@ -105,8 +105,6 @@ unsigned int GetNewMethodID(void) { class JitEventListenerTest { protected: void InitEE(const std::string &IRFile) { - LLVMContext &Context = getGlobalContext(); - // If we have a native target, initialize it to ensure it is linked in and // usable by the JIT. InitializeNativeTarget(); @@ -181,7 +179,7 @@ InputFilename(cl::Positional, cl::desc("<input IR file>"), int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. diff --git a/tools/llvm-link/CMakeLists.txt b/tools/llvm-link/CMakeLists.txt index e4ced186336c6..f69df95947d95 100644 --- a/tools/llvm-link/CMakeLists.txt +++ b/tools/llvm-link/CMakeLists.txt @@ -5,6 +5,7 @@ set(LLVM_LINK_COMPONENTS Linker Object Support + TransformUtils ) add_llvm_tool(llvm-link diff --git a/tools/llvm-link/LLVMBuild.txt b/tools/llvm-link/LLVMBuild.txt index 807823820ed6b..1dba5c0adb3f0 100644 --- a/tools/llvm-link/LLVMBuild.txt +++ b/tools/llvm-link/LLVMBuild.txt @@ -19,4 +19,4 @@ type = Tool name = llvm-link parent = Tools -required_libraries = AsmParser BitReader BitWriter IRReader Linker Object +required_libraries = AsmParser BitReader BitWriter IRReader Linker Object TransformUtils diff --git a/tools/llvm-link/Makefile b/tools/llvm-link/Makefile deleted file mode 100644 index a08d2844ceca8..0000000000000 --- a/tools/llvm-link/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-link/Makefile ----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-link -LINK_COMPONENTS := linker bitreader bitwriter asmparser irreader object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-link/llvm-link.cpp b/tools/llvm-link/llvm-link.cpp index a32383028ae24..185ae2a82a17b 100644 --- a/tools/llvm-link/llvm-link.cpp +++ b/tools/llvm-link/llvm-link.cpp @@ -12,18 +12,18 @@ // //===----------------------------------------------------------------------===// -#include "llvm/Linker/Linker.h" #include "llvm/ADT/STLExtras.h" #include "llvm/Bitcode/ReaderWriter.h" #include "llvm/IR/AutoUpgrade.h" #include "llvm/IR/DiagnosticInfo.h" #include "llvm/IR/DiagnosticPrinter.h" -#include "llvm/IR/FunctionInfo.h" #include "llvm/IR/LLVMContext.h" #include "llvm/IR/Module.h" +#include "llvm/IR/ModuleSummaryIndex.h" #include "llvm/IR/Verifier.h" #include "llvm/IRReader/IRReader.h" -#include "llvm/Object/FunctionIndexObjectFile.h" +#include "llvm/Linker/Linker.h" +#include "llvm/Object/ModuleSummaryIndexObjectFile.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/ManagedStatic.h" @@ -33,7 +33,10 @@ #include "llvm/Support/SourceMgr.h" #include "llvm/Support/SystemUtils.h" #include "llvm/Support/ToolOutputFile.h" +#include "llvm/Transforms/Utils/FunctionImportUtils.h" + #include <memory> +#include <utility> using namespace llvm; static cl::list<std::string> @@ -52,15 +55,14 @@ static cl::list<std::string> Imports( cl::desc("Pair of function name and filename, where function should be " "imported from bitcode in filename")); -// Option to support testing of function importing. The function index +// Option to support testing of function importing. The module summary // must be specified in the case were we request imports via the -import // option, as well as when compiling any module with functions that may be // exported (imported by a different llvm-link -import invocation), to ensure // consistent promotion and renaming of locals. -static cl::opt<std::string> FunctionIndex("functionindex", - cl::desc("Function index filename"), - cl::init(""), - cl::value_desc("filename")); +static cl::opt<std::string> + SummaryIndex("summary-index", cl::desc("Module summary index filename"), + cl::init(""), cl::value_desc("filename")); static cl::opt<std::string> OutputFilename("o", cl::desc("Override output filename"), cl::init("-"), @@ -70,6 +72,10 @@ static cl::opt<bool> Internalize("internalize", cl::desc("Internalize linked symbols")); static cl::opt<bool> + DisableDITypeMap("disable-debug-info-type-map", + cl::desc("Don't use a uniquing type map for debug info")); + +static cl::opt<bool> OnlyNeeded("only-needed", cl::desc("Link only needed symbols")); static cl::opt<bool> @@ -89,10 +95,6 @@ static cl::opt<bool> SuppressWarnings("suppress-warnings", cl::desc("Suppress all linking warnings"), cl::init(false)); -static cl::opt<bool> - PreserveModules("preserve-modules", - cl::desc("Preserve linked modules for testing")); - static cl::opt<bool> PreserveBitcodeUseListOrder( "preserve-bc-uselistorder", cl::desc("Preserve use-list order when writing LLVM bitcode."), @@ -114,8 +116,10 @@ static std::unique_ptr<Module> loadFile(const char *argv0, if (Verbose) errs() << "Loading '" << FN << "'\n"; std::unique_ptr<Module> Result = getLazyIRFileModule(FN, Err, Context, !MaterializeMetadata); - if (!Result) + if (!Result) { Err.print(argv0, errs()); + return nullptr; + } if (MaterializeMetadata) { Result->materializeMetadata(); @@ -125,6 +129,48 @@ static std::unique_ptr<Module> loadFile(const char *argv0, return Result; } +namespace { + +/// Helper to load on demand a Module from file and cache it for subsequent +/// queries during function importing. +class ModuleLazyLoaderCache { + /// Cache of lazily loaded module for import. + StringMap<std::unique_ptr<Module>> ModuleMap; + + /// Retrieve a Module from the cache or lazily load it on demand. + std::function<std::unique_ptr<Module>(const char *argv0, + const std::string &FileName)> + createLazyModule; + +public: + /// Create the loader, Module will be initialized in \p Context. + ModuleLazyLoaderCache(std::function<std::unique_ptr<Module>( + const char *argv0, const std::string &FileName)> + createLazyModule) + : createLazyModule(std::move(createLazyModule)) {} + + /// Retrieve a Module from the cache or lazily load it on demand. + Module &operator()(const char *argv0, const std::string &FileName); + + std::unique_ptr<Module> takeModule(const std::string &FileName) { + auto I = ModuleMap.find(FileName); + assert(I != ModuleMap.end()); + std::unique_ptr<Module> Ret = std::move(I->second); + ModuleMap.erase(I); + return Ret; + } +}; + +// Get a Module for \p FileName from the cache, or load it lazily. +Module &ModuleLazyLoaderCache::operator()(const char *argv0, + const std::string &Identifier) { + auto &Module = ModuleMap[Identifier]; + if (!Module) + Module = createLazyModule(argv0, Identifier); + return *Module; +} +} // anonymous namespace + static void diagnosticHandler(const DiagnosticInfo &DI) { unsigned Severity = DI.getSeverity(); switch (Severity) { @@ -153,8 +199,24 @@ static void diagnosticHandlerWithContext(const DiagnosticInfo &DI, void *C) { /// Import any functions requested via the -import option. static bool importFunctions(const char *argv0, LLVMContext &Context, Linker &L) { - StringMap<std::unique_ptr<DenseMap<unsigned, MDNode *>>> - ModuleToTempMDValsMap; + if (SummaryIndex.empty()) + return true; + ErrorOr<std::unique_ptr<ModuleSummaryIndex>> IndexOrErr = + llvm::getModuleSummaryIndexForFile(SummaryIndex, diagnosticHandler); + std::error_code EC = IndexOrErr.getError(); + if (EC) { + errs() << EC.message() << '\n'; + return false; + } + auto Index = std::move(IndexOrErr.get()); + + // Map of Module -> List of globals to import from the Module + std::map<StringRef, DenseSet<const GlobalValue *>> ModuleToGlobalsToImportMap; + auto ModuleLoader = [&Context](const char *argv0, + const std::string &Identifier) { + return loadFile(argv0, Identifier, Context, false); + }; + ModuleLazyLoaderCache ModuleLoaderCache(ModuleLoader); for (const auto &Import : Imports) { // Identify the requested function and its bitcode source file. size_t Idx = Import.find(':'); @@ -166,19 +228,15 @@ static bool importFunctions(const char *argv0, LLVMContext &Context, std::string FileName = Import.substr(Idx + 1, std::string::npos); // Load the specified source module. - std::unique_ptr<Module> M = loadFile(argv0, FileName, Context, false); - if (!M.get()) { - errs() << argv0 << ": error loading file '" << FileName << "'\n"; - return false; - } + auto &SrcModule = ModuleLoaderCache(argv0, FileName); - if (verifyModule(*M, &errs())) { + if (verifyModule(SrcModule, &errs())) { errs() << argv0 << ": " << FileName << ": error: input module is broken!\n"; return false; } - Function *F = M->getFunction(FunctionName); + Function *F = SrcModule.getFunction(FunctionName); if (!F) { errs() << "Ignoring import request for non-existent function " << FunctionName << " from " << FileName << "\n"; @@ -196,53 +254,36 @@ static bool importFunctions(const char *argv0, LLVMContext &Context, if (Verbose) errs() << "Importing " << FunctionName << " from " << FileName << "\n"; - std::unique_ptr<FunctionInfoIndex> Index; - if (!FunctionIndex.empty()) { - ErrorOr<std::unique_ptr<FunctionInfoIndex>> IndexOrErr = - llvm::getFunctionIndexForFile(FunctionIndex, diagnosticHandler); - std::error_code EC = IndexOrErr.getError(); - if (EC) { - errs() << EC.message() << '\n'; - return false; - } - Index = std::move(IndexOrErr.get()); - } + auto &Entry = ModuleToGlobalsToImportMap[SrcModule.getModuleIdentifier()]; + Entry.insert(F); - // Save the mapping of value ids to temporary metadata created when - // importing this function. If we have already imported from this module, - // add new temporary metadata to the existing mapping. - auto &TempMDVals = ModuleToTempMDValsMap[FileName]; - if (!TempMDVals) - TempMDVals = llvm::make_unique<DenseMap<unsigned, MDNode *>>(); - - // Link in the specified function. - DenseSet<const GlobalValue *> FunctionsToImport; - FunctionsToImport.insert(F); - if (L.linkInModule(std::move(M), Linker::Flags::None, Index.get(), - &FunctionsToImport, TempMDVals.get())) - return false; + F->materialize(); } - // Now link in metadata for all modules from which we imported functions. - for (StringMapEntry<std::unique_ptr<DenseMap<unsigned, MDNode *>>> &SME : - ModuleToTempMDValsMap) { - // Load the specified source module. - std::unique_ptr<Module> M = loadFile(argv0, SME.getKey(), Context, true); - if (!M.get()) { - errs() << argv0 << ": error loading file '" << SME.getKey() << "'\n"; - return false; - } + // Do the actual import of globals now, one Module at a time + for (auto &GlobalsToImportPerModule : ModuleToGlobalsToImportMap) { + // Get the module for the import + auto &GlobalsToImport = GlobalsToImportPerModule.second; + std::unique_ptr<Module> SrcModule = + ModuleLoaderCache.takeModule(GlobalsToImportPerModule.first); + assert(&Context == &SrcModule->getContext() && "Context mismatch"); - if (verifyModule(*M, &errs())) { - errs() << argv0 << ": " << SME.getKey() - << ": error: input module is broken!\n"; - return false; - } + // If modules were created with lazy metadata loading, materialize it + // now, before linking it (otherwise this will be a noop). + SrcModule->materializeMetadata(); + UpgradeDebugInfo(*SrcModule); + + // Linkage Promotion and renaming + if (renameModuleForThinLTO(*SrcModule, *Index, &GlobalsToImport)) + return true; + + // Instruct the linker to not automatically import linkonce defintion. + unsigned Flags = Linker::Flags::DontForceLinkLinkonceODR; - // Link in all necessary metadata from this module. - if (L.linkInMetadata(*M, SME.getValue().get())) + if (L.linkInModule(std::move(SrcModule), Flags, &GlobalsToImport)) return false; } + return true; } @@ -258,41 +299,38 @@ static bool linkFiles(const char *argv0, LLVMContext &Context, Linker &L, return false; } - if (verifyModule(*M, &errs())) { + // Note that when ODR merging types cannot verify input files in here When + // doing that debug metadata in the src module might already be pointing to + // the destination. + if (DisableDITypeMap && verifyModule(*M, &errs())) { errs() << argv0 << ": " << File << ": error: input module is broken!\n"; return false; } - // If a function index is supplied, load it so linkInModule can treat + // If a module summary index is supplied, load it so linkInModule can treat // local functions/variables as exported and promote if necessary. - std::unique_ptr<FunctionInfoIndex> Index; - if (!FunctionIndex.empty()) { - ErrorOr<std::unique_ptr<FunctionInfoIndex>> IndexOrErr = - llvm::getFunctionIndexForFile(FunctionIndex, diagnosticHandler); + if (!SummaryIndex.empty()) { + ErrorOr<std::unique_ptr<ModuleSummaryIndex>> IndexOrErr = + llvm::getModuleSummaryIndexForFile(SummaryIndex, diagnosticHandler); std::error_code EC = IndexOrErr.getError(); if (EC) { errs() << EC.message() << '\n'; return false; } - Index = std::move(IndexOrErr.get()); + auto Index = std::move(IndexOrErr.get()); + + // Promotion + if (renameModuleForThinLTO(*M, *Index)) + return true; } if (Verbose) errs() << "Linking in '" << File << "'\n"; - if (L.linkInModule(std::move(M), ApplicableFlags, Index.get())) + if (L.linkInModule(std::move(M), ApplicableFlags)) return false; // All linker flags apply to linking of subsequent files. ApplicableFlags = Flags; - - // If requested for testing, preserve modules by releasing them from - // the unique_ptr before the are freed. This can help catch any - // cross-module references from e.g. unneeded metadata references - // that aren't properly set to null but instead mapped to the source - // module version. The bitcode writer will assert if it finds any such - // cross-module references. - if (PreserveModules) - M.release(); } return true; @@ -300,15 +338,18 @@ static bool linkFiles(const char *argv0, LLVMContext &Context, Linker &L, int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); - LLVMContext &Context = getGlobalContext(); + LLVMContext Context; Context.setDiagnosticHandler(diagnosticHandlerWithContext, nullptr, true); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. cl::ParseCommandLineOptions(argc, argv, "llvm linker\n"); + if (!DisableDITypeMap) + Context.enableDebugTypeODRUniquing(); + auto Composite = make_unique<Module>("llvm-link", Context); Linker L(*Composite); diff --git a/tools/llvm-lto/CMakeLists.txt b/tools/llvm-lto/CMakeLists.txt index 29b3339e6cea0..de888a74c530b 100644 --- a/tools/llvm-lto/CMakeLists.txt +++ b/tools/llvm-lto/CMakeLists.txt @@ -1,7 +1,9 @@ set(LLVM_LINK_COMPONENTS ${LLVM_TARGETS_TO_BUILD} + BitReader BitWriter Core + IRReader LTO MC Object diff --git a/tools/llvm-lto/LLVMBuild.txt b/tools/llvm-lto/LLVMBuild.txt index b36f4a981532b..7a86797899dea 100644 --- a/tools/llvm-lto/LLVMBuild.txt +++ b/tools/llvm-lto/LLVMBuild.txt @@ -19,4 +19,4 @@ type = Tool name = llvm-lto parent = Tools -required_libraries = BitWriter Core LTO Object Support all-targets +required_libraries = BitWriter Core IRReader LTO Object Support all-targets diff --git a/tools/llvm-lto/Makefile b/tools/llvm-lto/Makefile deleted file mode 100644 index f8ca7e1cac5d0..0000000000000 --- a/tools/llvm-lto/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-lto/Makefile -----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-lto -LINK_COMPONENTS := lto ipo scalaropts linker bitreader bitwriter mcdisassembler support target vectorize all-targets - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-lto/llvm-lto.cpp b/tools/llvm-lto/llvm-lto.cpp index 232051130cb20..b3b617b941d33 100644 --- a/tools/llvm-lto/llvm-lto.cpp +++ b/tools/llvm-lto/llvm-lto.cpp @@ -17,14 +17,19 @@ #include "llvm/CodeGen/CommandFlags.h" #include "llvm/IR/DiagnosticPrinter.h" #include "llvm/IR/LLVMContext.h" -#include "llvm/LTO/LTOCodeGenerator.h" -#include "llvm/LTO/LTOModule.h" -#include "llvm/Object/FunctionIndexObjectFile.h" +#include "llvm/IR/Verifier.h" +#include "llvm/IRReader/IRReader.h" +#include "llvm/LTO/legacy/LTOCodeGenerator.h" +#include "llvm/LTO/legacy/LTOModule.h" +#include "llvm/LTO/legacy/ThinLTOCodeGenerator.h" +#include "llvm/Object/ModuleSummaryIndexObjectFile.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/ManagedStatic.h" +#include "llvm/Support/Path.h" #include "llvm/Support/PrettyStackTrace.h" #include "llvm/Support/Signals.h" +#include "llvm/Support/SourceMgr.h" #include "llvm/Support/TargetSelect.h" #include "llvm/Support/ToolOutputFile.h" #include "llvm/Support/raw_ostream.h" @@ -33,59 +38,108 @@ using namespace llvm; static cl::opt<char> -OptLevel("O", - cl::desc("Optimization level. [-O0, -O1, -O2, or -O3] " - "(default = '-O2')"), - cl::Prefix, - cl::ZeroOrMore, - cl::init('2')); + OptLevel("O", cl::desc("Optimization level. [-O0, -O1, -O2, or -O3] " + "(default = '-O2')"), + cl::Prefix, cl::ZeroOrMore, cl::init('2')); static cl::opt<bool> DisableVerify( "disable-verify", cl::init(false), cl::desc("Do not run the verifier during the optimization pipeline")); -static cl::opt<bool> -DisableInline("disable-inlining", cl::init(false), - cl::desc("Do not run the inliner pass")); +static cl::opt<bool> DisableInline("disable-inlining", cl::init(false), + cl::desc("Do not run the inliner pass")); static cl::opt<bool> -DisableGVNLoadPRE("disable-gvn-loadpre", cl::init(false), - cl::desc("Do not run the GVN load PRE pass")); + DisableGVNLoadPRE("disable-gvn-loadpre", cl::init(false), + cl::desc("Do not run the GVN load PRE pass")); -static cl::opt<bool> -DisableLTOVectorization("disable-lto-vectorization", cl::init(false), - cl::desc("Do not run loop or slp vectorization during LTO")); +static cl::opt<bool> DisableLTOVectorization( + "disable-lto-vectorization", cl::init(false), + cl::desc("Do not run loop or slp vectorization during LTO")); -static cl::opt<bool> -UseDiagnosticHandler("use-diagnostic-handler", cl::init(false), - cl::desc("Use a diagnostic handler to test the handler interface")); +static cl::opt<bool> UseDiagnosticHandler( + "use-diagnostic-handler", cl::init(false), + cl::desc("Use a diagnostic handler to test the handler interface")); static cl::opt<bool> ThinLTO("thinlto", cl::init(false), cl::desc("Only write combined global index for ThinLTO backends")); -static cl::opt<bool> -SaveModuleFile("save-merged-module", cl::init(false), - cl::desc("Write merged LTO module to file before CodeGen")); +enum ThinLTOModes { + THINLINK, + THINDISTRIBUTE, + THINEMITIMPORTS, + THINPROMOTE, + THINIMPORT, + THININTERNALIZE, + THINOPT, + THINCODEGEN, + THINALL +}; -static cl::list<std::string> -InputFilenames(cl::Positional, cl::OneOrMore, - cl::desc("<input bitcode files>")); +cl::opt<ThinLTOModes> ThinLTOMode( + "thinlto-action", cl::desc("Perform a single ThinLTO stage:"), + cl::values( + clEnumValN( + THINLINK, "thinlink", + "ThinLink: produces the index by linking only the summaries."), + clEnumValN(THINDISTRIBUTE, "distributedindexes", + "Produces individual indexes for distributed backends."), + clEnumValN(THINEMITIMPORTS, "emitimports", + "Emit imports files for distributed backends."), + clEnumValN(THINPROMOTE, "promote", + "Perform pre-import promotion (requires -thinlto-index)."), + clEnumValN(THINIMPORT, "import", "Perform both promotion and " + "cross-module importing (requires " + "-thinlto-index)."), + clEnumValN(THININTERNALIZE, "internalize", + "Perform internalization driven by -exported-symbol " + "(requires -thinlto-index)."), + clEnumValN(THINOPT, "optimize", "Perform ThinLTO optimizations."), + clEnumValN(THINCODEGEN, "codegen", "CodeGen (expected to match llc)"), + clEnumValN(THINALL, "run", "Perform ThinLTO end-to-end"), + clEnumValEnd)); static cl::opt<std::string> -OutputFilename("o", cl::init(""), - cl::desc("Override output filename"), - cl::value_desc("filename")); + ThinLTOIndex("thinlto-index", + cl::desc("Provide the index produced by a ThinLink, required " + "to perform the promotion and/or importing.")); + +static cl::opt<std::string> ThinLTOPrefixReplace( + "thinlto-prefix-replace", + cl::desc("Control where files for distributed backends are " + "created. Expects 'oldprefix;newprefix' and if path " + "prefix of output file is oldprefix it will be " + "replaced with newprefix.")); + +static cl::opt<std::string> ThinLTOModuleId( + "thinlto-module-id", + cl::desc("For the module ID for the file to process, useful to " + "match what is in the index.")); -static cl::list<std::string> -ExportedSymbols("exported-symbol", - cl::desc("Symbol to export from the resulting object file"), - cl::ZeroOrMore); +static cl::opt<std::string> + ThinLTOCacheDir("thinlto-cache-dir", cl::desc("Enable ThinLTO caching.")); + +static cl::opt<bool> + SaveModuleFile("save-merged-module", cl::init(false), + cl::desc("Write merged LTO module to file before CodeGen")); + +static cl::list<std::string> InputFilenames(cl::Positional, cl::OneOrMore, + cl::desc("<input bitcode files>")); + +static cl::opt<std::string> OutputFilename("o", cl::init(""), + cl::desc("Override output filename"), + cl::value_desc("filename")); + +static cl::list<std::string> ExportedSymbols( + "exported-symbol", + cl::desc("List of symbols to export from the resulting object file"), + cl::ZeroOrMore); static cl::list<std::string> -DSOSymbols("dso-symbol", - cl::desc("Symbol to put in the symtab in the resulting dso"), - cl::ZeroOrMore); + DSOSymbols("dso-symbol", + cl::desc("Symbol to put in the symtab in the resulting dso"), + cl::ZeroOrMore); static cl::opt<bool> ListSymbolsOnly( "list-symbols-only", cl::init(false), @@ -98,6 +152,14 @@ static cl::opt<bool> SetMergedModule( static cl::opt<unsigned> Parallelism("j", cl::Prefix, cl::init(1), cl::desc("Number of backend threads")); +static cl::opt<bool> RestoreGlobalsLinkage( + "restore-linkage", cl::init(false), + cl::desc("Restore original linkage of globals prior to CodeGen")); + +static cl::opt<bool> CheckHasObjC( + "check-for-objc", cl::init(false), + cl::desc("Only check if the module has objective-C defined in it")); + namespace { struct ModuleInfo { std::vector<bool> CanBeHidden; @@ -154,8 +216,8 @@ static void diagnosticHandler(const DiagnosticInfo &DI) { exit(1); } -static void diagnosticHandlerWithContenxt(const DiagnosticInfo &DI, - void *Context) { +static void diagnosticHandlerWithContext(const DiagnosticInfo &DI, + void *Context) { diagnosticHandler(DI); } @@ -174,6 +236,11 @@ static void error(const ErrorOr<T> &V, const Twine &Prefix) { error(V.getError(), Prefix); } +static void maybeVerifyModule(const Module &Mod) { + if (!DisableVerify && verifyModule(Mod)) + error("Broken Module"); +} + static std::unique_ptr<LTOModule> getLocalLTOModule(StringRef Path, std::unique_ptr<MemoryBuffer> &Buffer, const TargetOptions &Options) { @@ -182,9 +249,13 @@ getLocalLTOModule(StringRef Path, std::unique_ptr<MemoryBuffer> &Buffer, error(BufferOrErr, "error loading file '" + Path + "'"); Buffer = std::move(BufferOrErr.get()); CurrentActivity = ("loading file '" + Path + "'").str(); + std::unique_ptr<LLVMContext> Context = llvm::make_unique<LLVMContext>(); + Context->setDiagnosticHandler(diagnosticHandlerWithContext, nullptr, true); ErrorOr<std::unique_ptr<LTOModule>> Ret = LTOModule::createInLocalContext( - Buffer->getBufferStart(), Buffer->getBufferSize(), Options, Path); + std::move(Context), Buffer->getBufferStart(), Buffer->getBufferSize(), + Options, Path); CurrentActivity = ""; + maybeVerifyModule((*Ret)->getModule()); return std::move(*Ret); } @@ -211,16 +282,17 @@ static void listSymbols(const TargetOptions &Options) { /// /// This is meant to enable testing of ThinLTO combined index generation, /// currently available via the gold plugin via -thinlto. -static void createCombinedFunctionIndex() { - FunctionInfoIndex CombinedIndex; +static void createCombinedModuleSummaryIndex() { + ModuleSummaryIndex CombinedIndex; uint64_t NextModuleId = 0; for (auto &Filename : InputFilenames) { CurrentActivity = "loading file '" + Filename + "'"; - ErrorOr<std::unique_ptr<FunctionInfoIndex>> IndexOrErr = - llvm::getFunctionIndexForFile(Filename, diagnosticHandler); - std::unique_ptr<FunctionInfoIndex> Index = std::move(IndexOrErr.get()); + ErrorOr<std::unique_ptr<ModuleSummaryIndex>> IndexOrErr = + llvm::getModuleSummaryIndexForFile(Filename, diagnosticHandler); + error(IndexOrErr, "error " + CurrentActivity); + std::unique_ptr<ModuleSummaryIndex> Index = std::move(IndexOrErr.get()); CurrentActivity = ""; - // Skip files without a function summary. + // Skip files without a module summary. if (!Index) continue; CombinedIndex.mergeFrom(std::move(Index), ++NextModuleId); @@ -230,13 +302,400 @@ static void createCombinedFunctionIndex() { raw_fd_ostream OS(OutputFilename + ".thinlto.bc", EC, sys::fs::OpenFlags::F_None); error(EC, "error opening the file '" + OutputFilename + ".thinlto.bc'"); - WriteFunctionSummaryToFile(CombinedIndex, OS); + WriteIndexToFile(CombinedIndex, OS); OS.close(); } +/// Parse the thinlto_prefix_replace option into the \p OldPrefix and +/// \p NewPrefix strings, if it was specified. +static void getThinLTOOldAndNewPrefix(std::string &OldPrefix, + std::string &NewPrefix) { + assert(ThinLTOPrefixReplace.empty() || + ThinLTOPrefixReplace.find(";") != StringRef::npos); + StringRef PrefixReplace = ThinLTOPrefixReplace; + std::pair<StringRef, StringRef> Split = PrefixReplace.split(";"); + OldPrefix = Split.first.str(); + NewPrefix = Split.second.str(); +} + +/// Given the original \p Path to an output file, replace any path +/// prefix matching \p OldPrefix with \p NewPrefix. Also, create the +/// resulting directory if it does not yet exist. +static std::string getThinLTOOutputFile(const std::string &Path, + const std::string &OldPrefix, + const std::string &NewPrefix) { + if (OldPrefix.empty() && NewPrefix.empty()) + return Path; + SmallString<128> NewPath(Path); + llvm::sys::path::replace_path_prefix(NewPath, OldPrefix, NewPrefix); + StringRef ParentPath = llvm::sys::path::parent_path(NewPath.str()); + if (!ParentPath.empty()) { + // Make sure the new directory exists, creating it if necessary. + if (std::error_code EC = llvm::sys::fs::create_directories(ParentPath)) + error(EC, "error creating the directory '" + ParentPath + "'"); + } + return NewPath.str(); +} + +namespace thinlto { + +std::vector<std::unique_ptr<MemoryBuffer>> +loadAllFilesForIndex(const ModuleSummaryIndex &Index) { + std::vector<std::unique_ptr<MemoryBuffer>> InputBuffers; + + for (auto &ModPath : Index.modulePaths()) { + const auto &Filename = ModPath.first(); + auto CurrentActivity = "loading file '" + Filename + "'"; + auto InputOrErr = MemoryBuffer::getFile(Filename); + error(InputOrErr, "error " + CurrentActivity); + InputBuffers.push_back(std::move(*InputOrErr)); + } + return InputBuffers; +} + +std::unique_ptr<ModuleSummaryIndex> loadCombinedIndex() { + if (ThinLTOIndex.empty()) + report_fatal_error("Missing -thinlto-index for ThinLTO promotion stage"); + auto CurrentActivity = "loading file '" + ThinLTOIndex + "'"; + ErrorOr<std::unique_ptr<ModuleSummaryIndex>> IndexOrErr = + llvm::getModuleSummaryIndexForFile(ThinLTOIndex, diagnosticHandler); + error(IndexOrErr, "error " + CurrentActivity); + return std::move(IndexOrErr.get()); +} + +static std::unique_ptr<Module> loadModule(StringRef Filename, + LLVMContext &Ctx) { + SMDiagnostic Err; + std::unique_ptr<Module> M(parseIRFile(Filename, Err, Ctx)); + if (!M) { + Err.print("llvm-lto", errs()); + report_fatal_error("Can't load module for file " + Filename); + } + maybeVerifyModule(*M); + + if (ThinLTOModuleId.getNumOccurrences()) { + if (InputFilenames.size() != 1) + report_fatal_error("Can't override the module id for multiple files"); + M->setModuleIdentifier(ThinLTOModuleId); + } + return M; +} + +static void writeModuleToFile(Module &TheModule, StringRef Filename) { + std::error_code EC; + raw_fd_ostream OS(Filename, EC, sys::fs::OpenFlags::F_None); + error(EC, "error opening the file '" + Filename + "'"); + maybeVerifyModule(TheModule); + WriteBitcodeToFile(&TheModule, OS, /* ShouldPreserveUseListOrder */ true); +} + +class ThinLTOProcessing { +public: + ThinLTOCodeGenerator ThinGenerator; + + ThinLTOProcessing(const TargetOptions &Options) { + ThinGenerator.setCodePICModel(getRelocModel()); + ThinGenerator.setTargetOptions(Options); + ThinGenerator.setCacheDir(ThinLTOCacheDir); + + // Add all the exported symbols to the table of symbols to preserve. + for (unsigned i = 0; i < ExportedSymbols.size(); ++i) + ThinGenerator.preserveSymbol(ExportedSymbols[i]); + } + + void run() { + switch (ThinLTOMode) { + case THINLINK: + return thinLink(); + case THINDISTRIBUTE: + return distributedIndexes(); + case THINEMITIMPORTS: + return emitImports(); + case THINPROMOTE: + return promote(); + case THINIMPORT: + return import(); + case THININTERNALIZE: + return internalize(); + case THINOPT: + return optimize(); + case THINCODEGEN: + return codegen(); + case THINALL: + return runAll(); + } + } + +private: + /// Load the input files, create the combined index, and write it out. + void thinLink() { + // Perform "ThinLink": just produce the index + if (OutputFilename.empty()) + report_fatal_error( + "OutputFilename is necessary to store the combined index.\n"); + + LLVMContext Ctx; + std::vector<std::unique_ptr<MemoryBuffer>> InputBuffers; + for (unsigned i = 0; i < InputFilenames.size(); ++i) { + auto &Filename = InputFilenames[i]; + StringRef CurrentActivity = "loading file '" + Filename + "'"; + auto InputOrErr = MemoryBuffer::getFile(Filename); + error(InputOrErr, "error " + CurrentActivity); + InputBuffers.push_back(std::move(*InputOrErr)); + ThinGenerator.addModule(Filename, InputBuffers.back()->getBuffer()); + } + + auto CombinedIndex = ThinGenerator.linkCombinedIndex(); + std::error_code EC; + raw_fd_ostream OS(OutputFilename, EC, sys::fs::OpenFlags::F_None); + error(EC, "error opening the file '" + OutputFilename + "'"); + WriteIndexToFile(*CombinedIndex, OS); + return; + } + + /// Load the combined index from disk, then compute and generate + /// individual index files suitable for ThinLTO distributed backend builds + /// on the files mentioned on the command line (these must match the index + /// content). + void distributedIndexes() { + if (InputFilenames.size() != 1 && !OutputFilename.empty()) + report_fatal_error("Can't handle a single output filename and multiple " + "input files, do not provide an output filename and " + "the output files will be suffixed from the input " + "ones."); + + std::string OldPrefix, NewPrefix; + getThinLTOOldAndNewPrefix(OldPrefix, NewPrefix); + + auto Index = loadCombinedIndex(); + for (auto &Filename : InputFilenames) { + // Build a map of module to the GUIDs and summary objects that should + // be written to its index. + std::map<std::string, GVSummaryMapTy> ModuleToSummariesForIndex; + ThinLTOCodeGenerator::gatherImportedSummariesForModule( + Filename, *Index, ModuleToSummariesForIndex); + + std::string OutputName = OutputFilename; + if (OutputName.empty()) { + OutputName = Filename + ".thinlto.bc"; + } + OutputName = getThinLTOOutputFile(OutputName, OldPrefix, NewPrefix); + std::error_code EC; + raw_fd_ostream OS(OutputName, EC, sys::fs::OpenFlags::F_None); + error(EC, "error opening the file '" + OutputName + "'"); + WriteIndexToFile(*Index, OS, &ModuleToSummariesForIndex); + } + } + + /// Load the combined index from disk, compute the imports, and emit + /// the import file lists for each module to disk. + void emitImports() { + if (InputFilenames.size() != 1 && !OutputFilename.empty()) + report_fatal_error("Can't handle a single output filename and multiple " + "input files, do not provide an output filename and " + "the output files will be suffixed from the input " + "ones."); + + std::string OldPrefix, NewPrefix; + getThinLTOOldAndNewPrefix(OldPrefix, NewPrefix); + + auto Index = loadCombinedIndex(); + for (auto &Filename : InputFilenames) { + std::string OutputName = OutputFilename; + if (OutputName.empty()) { + OutputName = Filename + ".imports"; + } + OutputName = getThinLTOOutputFile(OutputName, OldPrefix, NewPrefix); + ThinLTOCodeGenerator::emitImports(Filename, OutputName, *Index); + } + } + + /// Load the combined index from disk, then load every file referenced by + /// the index and add them to the generator, finally perform the promotion + /// on the files mentioned on the command line (these must match the index + /// content). + void promote() { + if (InputFilenames.size() != 1 && !OutputFilename.empty()) + report_fatal_error("Can't handle a single output filename and multiple " + "input files, do not provide an output filename and " + "the output files will be suffixed from the input " + "ones."); + + auto Index = loadCombinedIndex(); + for (auto &Filename : InputFilenames) { + LLVMContext Ctx; + auto TheModule = loadModule(Filename, Ctx); + + ThinGenerator.promote(*TheModule, *Index); + + std::string OutputName = OutputFilename; + if (OutputName.empty()) { + OutputName = Filename + ".thinlto.promoted.bc"; + } + writeModuleToFile(*TheModule, OutputName); + } + } + + /// Load the combined index from disk, then load every file referenced by + /// the index and add them to the generator, then performs the promotion and + /// cross module importing on the files mentioned on the command line + /// (these must match the index content). + void import() { + if (InputFilenames.size() != 1 && !OutputFilename.empty()) + report_fatal_error("Can't handle a single output filename and multiple " + "input files, do not provide an output filename and " + "the output files will be suffixed from the input " + "ones."); + + auto Index = loadCombinedIndex(); + auto InputBuffers = loadAllFilesForIndex(*Index); + for (auto &MemBuffer : InputBuffers) + ThinGenerator.addModule(MemBuffer->getBufferIdentifier(), + MemBuffer->getBuffer()); + + for (auto &Filename : InputFilenames) { + LLVMContext Ctx; + auto TheModule = loadModule(Filename, Ctx); + + ThinGenerator.crossModuleImport(*TheModule, *Index); + + std::string OutputName = OutputFilename; + if (OutputName.empty()) { + OutputName = Filename + ".thinlto.imported.bc"; + } + writeModuleToFile(*TheModule, OutputName); + } + } + + void internalize() { + if (InputFilenames.size() != 1 && !OutputFilename.empty()) + report_fatal_error("Can't handle a single output filename and multiple " + "input files, do not provide an output filename and " + "the output files will be suffixed from the input " + "ones."); + + if (ExportedSymbols.empty()) + errs() << "Warning: -internalize will not perform without " + "-exported-symbol\n"; + + auto Index = loadCombinedIndex(); + auto InputBuffers = loadAllFilesForIndex(*Index); + for (auto &MemBuffer : InputBuffers) + ThinGenerator.addModule(MemBuffer->getBufferIdentifier(), + MemBuffer->getBuffer()); + + for (auto &Filename : InputFilenames) { + LLVMContext Ctx; + auto TheModule = loadModule(Filename, Ctx); + + ThinGenerator.internalize(*TheModule, *Index); + + std::string OutputName = OutputFilename; + if (OutputName.empty()) { + OutputName = Filename + ".thinlto.internalized.bc"; + } + writeModuleToFile(*TheModule, OutputName); + } + } + + void optimize() { + if (InputFilenames.size() != 1 && !OutputFilename.empty()) + report_fatal_error("Can't handle a single output filename and multiple " + "input files, do not provide an output filename and " + "the output files will be suffixed from the input " + "ones."); + if (!ThinLTOIndex.empty()) + errs() << "Warning: -thinlto-index ignored for optimize stage"; + + for (auto &Filename : InputFilenames) { + LLVMContext Ctx; + auto TheModule = loadModule(Filename, Ctx); + + ThinGenerator.optimize(*TheModule); + + std::string OutputName = OutputFilename; + if (OutputName.empty()) { + OutputName = Filename + ".thinlto.imported.bc"; + } + writeModuleToFile(*TheModule, OutputName); + } + } + + void codegen() { + if (InputFilenames.size() != 1 && !OutputFilename.empty()) + report_fatal_error("Can't handle a single output filename and multiple " + "input files, do not provide an output filename and " + "the output files will be suffixed from the input " + "ones."); + if (!ThinLTOIndex.empty()) + errs() << "Warning: -thinlto-index ignored for codegen stage"; + + for (auto &Filename : InputFilenames) { + LLVMContext Ctx; + auto TheModule = loadModule(Filename, Ctx); + + auto Buffer = ThinGenerator.codegen(*TheModule); + std::string OutputName = OutputFilename; + if (OutputName.empty()) { + OutputName = Filename + ".thinlto.o"; + } + if (OutputName == "-") { + outs() << Buffer->getBuffer(); + return; + } + + std::error_code EC; + raw_fd_ostream OS(OutputName, EC, sys::fs::OpenFlags::F_None); + error(EC, "error opening the file '" + OutputName + "'"); + OS << Buffer->getBuffer(); + } + } + + /// Full ThinLTO process + void runAll() { + if (!OutputFilename.empty()) + report_fatal_error("Do not provide an output filename for ThinLTO " + " processing, the output files will be suffixed from " + "the input ones."); + + if (!ThinLTOIndex.empty()) + errs() << "Warning: -thinlto-index ignored for full ThinLTO process"; + + LLVMContext Ctx; + std::vector<std::unique_ptr<MemoryBuffer>> InputBuffers; + for (unsigned i = 0; i < InputFilenames.size(); ++i) { + auto &Filename = InputFilenames[i]; + StringRef CurrentActivity = "loading file '" + Filename + "'"; + auto InputOrErr = MemoryBuffer::getFile(Filename); + error(InputOrErr, "error " + CurrentActivity); + InputBuffers.push_back(std::move(*InputOrErr)); + ThinGenerator.addModule(Filename, InputBuffers.back()->getBuffer()); + } + + ThinGenerator.run(); + + auto &Binaries = ThinGenerator.getProducedBinaries(); + if (Binaries.size() != InputFilenames.size()) + report_fatal_error("Number of output objects does not match the number " + "of inputs"); + + for (unsigned BufID = 0; BufID < Binaries.size(); ++BufID) { + auto OutputName = InputFilenames[BufID] + ".thinlto.o"; + std::error_code EC; + raw_fd_ostream OS(OutputName, EC, sys::fs::OpenFlags::F_None); + error(EC, "error opening the file '" + OutputName + "'"); + OS << Binaries[BufID]->getBuffer(); + } + } + + /// Load the combined index from disk, then load every file referenced by +}; + +} // namespace thinlto + int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. @@ -259,25 +718,49 @@ int main(int argc, char **argv) { return 0; } + if (CheckHasObjC) { + for (auto &Filename : InputFilenames) { + ErrorOr<std::unique_ptr<MemoryBuffer>> BufferOrErr = + MemoryBuffer::getFile(Filename); + error(BufferOrErr, "error loading file '" + Filename + "'"); + auto Buffer = std::move(BufferOrErr.get()); + LLVMContext Ctx; + if (llvm::isBitcodeContainingObjCCategory(*Buffer, Ctx)) + outs() << "Bitcode " << Filename << " contains ObjC\n"; + else + outs() << "Bitcode " << Filename << " does not contain ObjC\n"; + } + return 0; + } + + if (ThinLTOMode.getNumOccurrences()) { + if (ThinLTOMode.getNumOccurrences() > 1) + report_fatal_error("You can't specify more than one -thinlto-action"); + thinlto::ThinLTOProcessing ThinLTOProcessor(Options); + ThinLTOProcessor.run(); + return 0; + } + if (ThinLTO) { - createCombinedFunctionIndex(); + createCombinedModuleSummaryIndex(); return 0; } unsigned BaseArg = 0; LLVMContext Context; - Context.setDiagnosticHandler(diagnosticHandlerWithContenxt, nullptr, true); + Context.setDiagnosticHandler(diagnosticHandlerWithContext, nullptr, true); LTOCodeGenerator CodeGen(Context); if (UseDiagnosticHandler) CodeGen.setDiagnosticHandler(handleDiagnostics, nullptr); - CodeGen.setCodePICModel(RelocModel); + CodeGen.setCodePICModel(getRelocModel()); CodeGen.setDebugInfo(LTO_DEBUG_MODEL_DWARF); CodeGen.setTargetOptions(Options); + CodeGen.setShouldRestoreGlobalsLinkage(RestoreGlobalsLinkage); llvm::StringSet<llvm::MallocAllocator> DSOSymbolsSet; for (unsigned i = 0; i < DSOSymbols.size(); ++i) @@ -289,7 +772,6 @@ int main(int argc, char **argv) { CurrentActivity = "loading file '" + InputFilenames[i] + "'"; ErrorOr<std::unique_ptr<LTOModule>> ModuleOrErr = LTOModule::createFromFile(Context, InputFilenames[i].c_str(), Options); - error(ModuleOrErr, "error " + CurrentActivity); std::unique_ptr<LTOModule> &Module = *ModuleOrErr; CurrentActivity = ""; @@ -311,8 +793,7 @@ int main(int argc, char **argv) { CodeGen.setModule(std::move(Module)); } else if (!CodeGen.addModule(Module.get())) { // Print a message here so that we know addModule() did not abort. - errs() << argv[0] << ": error adding file '" << InputFilenames[i] << "'\n"; - return 1; + error("error adding file '" + InputFilenames[i] + "'"); } } @@ -346,8 +827,7 @@ int main(int argc, char **argv) { if (!CodeGen.optimize(DisableVerify, DisableInline, DisableGVNLoadPRE, DisableLTOVectorization)) { // Diagnostic messages should have been printed by the handler. - errs() << argv[0] << ": error optimizing the code\n"; - return 1; + error("error optimizing the code"); } if (SaveModuleFile) { @@ -355,10 +835,8 @@ int main(int argc, char **argv) { ModuleFilename += ".merged.bc"; std::string ErrMsg; - if (!CodeGen.writeMergedModules(ModuleFilename.c_str())) { - errs() << argv[0] << ": writing merged module failed.\n"; - return 1; - } + if (!CodeGen.writeMergedModules(ModuleFilename.c_str())) + error("writing merged module failed."); } std::list<tool_output_file> OSs; @@ -369,40 +847,29 @@ int main(int argc, char **argv) { PartFilename += "." + utostr(I); std::error_code EC; OSs.emplace_back(PartFilename, EC, sys::fs::F_None); - if (EC) { - errs() << argv[0] << ": error opening the file '" << PartFilename - << "': " << EC.message() << "\n"; - return 1; - } + if (EC) + error("error opening the file '" + PartFilename + "': " + EC.message()); OSPtrs.push_back(&OSs.back().os()); } - if (!CodeGen.compileOptimized(OSPtrs)) { + if (!CodeGen.compileOptimized(OSPtrs)) // Diagnostic messages should have been printed by the handler. - errs() << argv[0] << ": error compiling the code\n"; - return 1; - } + error("error compiling the code"); for (tool_output_file &OS : OSs) OS.keep(); } else { - if (Parallelism != 1) { - errs() << argv[0] << ": -j must be specified together with -o\n"; - return 1; - } + if (Parallelism != 1) + error("-j must be specified together with -o"); - if (SaveModuleFile) { - errs() << argv[0] << ": -save-merged-module must be specified with -o\n"; - return 1; - } + if (SaveModuleFile) + error(": -save-merged-module must be specified with -o"); const char *OutputName = nullptr; if (!CodeGen.compile_to_file(&OutputName, DisableVerify, DisableInline, - DisableGVNLoadPRE, DisableLTOVectorization)) { + DisableGVNLoadPRE, DisableLTOVectorization)) + error("error compiling the code"); // Diagnostic messages should have been printed by the handler. - errs() << argv[0] << ": error compiling the code\n"; - return 1; - } outs() << "Wrote native object file '" << OutputName << "'\n"; } diff --git a/tools/llvm-mc-fuzzer/CMakeLists.txt b/tools/llvm-mc-fuzzer/CMakeLists.txt index 0d6432b23dfa5..b42b3eee3c981 100644 --- a/tools/llvm-mc-fuzzer/CMakeLists.txt +++ b/tools/llvm-mc-fuzzer/CMakeLists.txt @@ -13,6 +13,6 @@ if( LLVM_USE_SANITIZE_COVERAGE ) add_llvm_tool(llvm-mc-fuzzer llvm-mc-fuzzer.cpp) target_link_libraries(llvm-mc-fuzzer - LLVMFuzzerNoMain + LLVMFuzzer ) endif() diff --git a/tools/llvm-mc-fuzzer/llvm-mc-fuzzer.cpp b/tools/llvm-mc-fuzzer/llvm-mc-fuzzer.cpp index 3f80e4582ee1b..df8e0112af5bf 100644 --- a/tools/llvm-mc-fuzzer/llvm-mc-fuzzer.cpp +++ b/tools/llvm-mc-fuzzer/llvm-mc-fuzzer.cpp @@ -9,13 +9,12 @@ // //===----------------------------------------------------------------------===// +#include "FuzzerInterface.h" #include "llvm-c/Disassembler.h" #include "llvm-c/Target.h" -#include "llvm/ADT/ArrayRef.h" #include "llvm/MC/SubtargetFeature.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/raw_ostream.h" -#include "FuzzerInterface.h" using namespace llvm; @@ -58,9 +57,10 @@ static cl::list<std::string> std::string FeaturesStr; static cl::list<std::string> - FuzzerArgv("fuzzer-args", cl::Positional, + FuzzerArgs("fuzzer-args", cl::Positional, cl::desc("Options to pass to the fuzzer"), cl::ZeroOrMore, cl::PositionalEatsArgs); +static std::vector<char *> ModifiedArgv; int DisassembleOneInput(const uint8_t *Data, size_t Size) { char AssemblyText[AssemblyTextBufSize]; @@ -88,7 +88,17 @@ int DisassembleOneInput(const uint8_t *Data, size_t Size) { return 0; } -int main(int argc, char **argv) { +int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Action == AC_Assemble) + errs() << "error: -assemble is not implemented\n"; + else if (Action == AC_Disassemble) + return DisassembleOneInput(Data, Size); + + llvm_unreachable("Unknown action"); + return 0; +} + +int LLVMFuzzerInitialize(int *argc, char ***argv) { // The command line is unusual compared to other fuzzers due to the need to // specify the target. Options like -triple, -mcpu, and -mattr work like // their counterparts in llvm-mc, while -fuzzer-args collects options for the @@ -112,11 +122,29 @@ int main(int argc, char **argv) { // individual instructions that test unique paths. Without this constraint, // there will be considerable redundancy in the corpus. + char **OriginalArgv = *argv; + LLVMInitializeAllTargetInfos(); LLVMInitializeAllTargetMCs(); LLVMInitializeAllDisassemblers(); - cl::ParseCommandLineOptions(argc, argv); + cl::ParseCommandLineOptions(*argc, OriginalArgv); + + // Rebuild the argv without the arguments llvm-mc-fuzzer consumed so that + // the driver can parse its arguments. + // + // FuzzerArgs cannot provide the non-const pointer that OriginalArgv needs. + // Re-use the strings from OriginalArgv instead of copying FuzzerArg to a + // non-const buffer to avoid the need to clean up when the fuzzer terminates. + ModifiedArgv.push_back(OriginalArgv[0]); + for (const auto &FuzzerArg : FuzzerArgs) { + for (int i = 1; i < *argc; ++i) { + if (FuzzerArg == OriginalArgv[i]) + ModifiedArgv.push_back(OriginalArgv[i]); + } + } + *argc = ModifiedArgv.size(); + *argv = ModifiedArgv.data(); // Package up features to be passed to target/subtarget // We have to pass it via a global since the callback doesn't @@ -128,11 +156,5 @@ int main(int argc, char **argv) { FeaturesStr = Features.getString(); } - if (Action == AC_Assemble) - errs() << "error: -assemble is not implemented\n"; - else if (Action == AC_Disassemble) - return fuzzer::FuzzerDriver(argc, argv, DisassembleOneInput); - - llvm_unreachable("Unknown action"); - return 1; + return 0; } diff --git a/tools/llvm-mc/Disassembler.cpp b/tools/llvm-mc/Disassembler.cpp index 5ffeffc076823..8185947fc5e57 100644 --- a/tools/llvm-mc/Disassembler.cpp +++ b/tools/llvm-mc/Disassembler.cpp @@ -16,7 +16,7 @@ #include "llvm/ADT/Triple.h" #include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCContext.h" -#include "llvm/MC/MCDisassembler.h" +#include "llvm/MC/MCDisassembler/MCDisassembler.h" #include "llvm/MC/MCInst.h" #include "llvm/MC/MCRegisterInfo.h" #include "llvm/MC/MCStreamer.h" diff --git a/tools/llvm-mc/Makefile b/tools/llvm-mc/Makefile deleted file mode 100644 index b147fadb5747b..0000000000000 --- a/tools/llvm-mc/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-mc/Makefile ------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-mc -LINK_COMPONENTS := all-targets MCDisassembler MCParser MC support - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-mc/llvm-mc.cpp b/tools/llvm-mc/llvm-mc.cpp index 96e3f7c21a517..00f19cb328c37 100644 --- a/tools/llvm-mc/llvm-mc.cpp +++ b/tools/llvm-mc/llvm-mc.cpp @@ -20,11 +20,11 @@ #include "llvm/MC/MCInstrInfo.h" #include "llvm/MC/MCObjectFileInfo.h" #include "llvm/MC/MCParser/AsmLexer.h" +#include "llvm/MC/MCParser/MCTargetAsmParser.h" #include "llvm/MC/MCRegisterInfo.h" #include "llvm/MC/MCSectionMachO.h" #include "llvm/MC/MCStreamer.h" #include "llvm/MC/MCSubtargetInfo.h" -#include "llvm/MC/MCTargetAsmParser.h" #include "llvm/MC/MCTargetOptionsCommandFlags.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/Compression.h" @@ -52,9 +52,22 @@ OutputFilename("o", cl::desc("Output filename"), static cl::opt<bool> ShowEncoding("show-encoding", cl::desc("Show instruction encodings")); -static cl::opt<bool> -CompressDebugSections("compress-debug-sections", - cl::desc("Compress DWARF debug sections")); +static cl::opt<bool> RelaxELFRel( + "relax-relocations", cl::init(true), + cl::desc("Emit R_X86_64_GOTPCRELX instead of R_X86_64_GOTPCREL")); + +static cl::opt<DebugCompressionType> +CompressDebugSections("compress-debug-sections", cl::ValueOptional, + cl::init(DebugCompressionType::DCT_None), + cl::desc("Choose DWARF debug sections compression:"), + cl::values( + clEnumValN(DebugCompressionType::DCT_None, "none", + "No compression"), + clEnumValN(DebugCompressionType::DCT_Zlib, "zlib", + "Use zlib compression"), + clEnumValN(DebugCompressionType::DCT_ZlibGnu, "zlib-gnu", + "Use zlib-gnu compression (depricated)"), + clEnumValEnd)); static cl::opt<bool> ShowInst("show-inst", cl::desc("Show internal instruction representation")); @@ -74,6 +87,10 @@ PrintImmHex("print-imm-hex", cl::init(false), static cl::list<std::string> DefineSymbol("defsym", cl::desc("Defines a symbol to be an integer constant")); +static cl::opt<bool> + PreserveComments("preserve-comments", + cl::desc("Preserve Comments in outputted assembly")); + enum OutputFileType { OFT_Null, OFT_AssemblyFile, @@ -115,20 +132,8 @@ MAttrs("mattr", cl::desc("Target specific attributes (-mattr=help for details)"), cl::value_desc("a1,+a2,-a3,...")); -static cl::opt<Reloc::Model> -RelocModel("relocation-model", - cl::desc("Choose relocation model"), - cl::init(Reloc::Default), - cl::values( - clEnumValN(Reloc::Default, "default", - "Target default relocation model"), - clEnumValN(Reloc::Static, "static", - "Non-relocatable code"), - clEnumValN(Reloc::PIC_, "pic", - "Fully relocatable, position independent code"), - clEnumValN(Reloc::DynamicNoPIC, "dynamic-no-pic", - "Relocatable external references, non-relocatable code"), - clEnumValEnd)); +static cl::opt<bool> PIC("position-independent", + cl::desc("Position independent"), cl::init(false)); static cl::opt<llvm::CodeModel::Model> CMModel("code-model", @@ -249,7 +254,7 @@ static int AsLexInput(SourceMgr &SrcMgr, MCAsmInfo &MAI, bool Error = false; while (Lexer.Lex().isNot(AsmToken::Eof)) { - AsmToken Tok = Lexer.getTok(); + const AsmToken &Tok = Lexer.getTok(); switch (Tok.getKind()) { default: @@ -368,7 +373,7 @@ static int AssembleInput(const char *ProgName, const Target *TheTarget, int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. @@ -419,20 +424,23 @@ int main(int argc, char **argv) { std::unique_ptr<MCAsmInfo> MAI(TheTarget->createMCAsmInfo(*MRI, TripleName)); assert(MAI && "Unable to create target asm info!"); - if (CompressDebugSections) { + MAI->setRelaxELFRelocations(RelaxELFRel); + + if (CompressDebugSections != DebugCompressionType::DCT_None) { if (!zlib::isAvailable()) { errs() << ProgName << ": build tools with zlib to enable -compress-debug-sections"; return 1; } - MAI->setCompressDebugSections(true); + MAI->setCompressDebugSections(CompressDebugSections); } + MAI->setPreserveAsmComments(PreserveComments); // FIXME: This is not pretty. MCContext has a ptr to MCObjectFileInfo and // MCObjectFileInfo needs a MCContext reference in order to initialize itself. MCObjectFileInfo MOFI; MCContext Ctx(MAI.get(), MRI.get(), &MOFI, &SrcMgr); - MOFI.InitMCObjectFileInfo(TheTriple, RelocModel, CMModel, Ctx); + MOFI.InitMCObjectFileInfo(TheTriple, PIC, CMModel, Ctx); if (SaveTempLabels) Ctx.setAllowTemporaryLabels(false); @@ -452,6 +460,12 @@ int main(int argc, char **argv) { Ctx.setDwarfDebugProducer(StringRef(DwarfDebugProducer)); if (!DebugCompilationDir.empty()) Ctx.setCompilationDir(DebugCompilationDir); + else { + // If no compilation dir is set, try to use the current directory. + SmallString<128> CWD; + if (!sys::fs::current_path(CWD)) + Ctx.setCompilationDir(CWD); + } if (!MainFileName.empty()) Ctx.setMainFileName(MainFileName); diff --git a/tools/llvm-mcmarkup/Makefile b/tools/llvm-mcmarkup/Makefile deleted file mode 100644 index 5633a9c301a3b..0000000000000 --- a/tools/llvm-mcmarkup/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-mcmarkup/Makefile ------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-mcmarkup -LINK_COMPONENTS := support - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS = 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-mcmarkup/llvm-mcmarkup.cpp b/tools/llvm-mcmarkup/llvm-mcmarkup.cpp index 56543139d6fac..0be3c715eee4e 100644 --- a/tools/llvm-mcmarkup/llvm-mcmarkup.cpp +++ b/tools/llvm-mcmarkup/llvm-mcmarkup.cpp @@ -208,7 +208,7 @@ static void parseMCMarkup(StringRef Filename) { int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. diff --git a/tools/llvm-nm/Makefile b/tools/llvm-nm/Makefile deleted file mode 100644 index ec20cef4258c1..0000000000000 --- a/tools/llvm-nm/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-nm/Makefile ------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-nm -LINK_COMPONENTS := all-targets bitreader object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-nm/llvm-nm.cpp b/tools/llvm-nm/llvm-nm.cpp index b812233034eee..9b2f5220badbc 100644 --- a/tools/llvm-nm/llvm-nm.cpp +++ b/tools/llvm-nm/llvm-nm.cpp @@ -16,6 +16,7 @@ // //===----------------------------------------------------------------------===// +#include "llvm/ADT/StringSwitch.h" #include "llvm/IR/Function.h" #include "llvm/IR/GlobalAlias.h" #include "llvm/IR/GlobalVariable.h" @@ -138,6 +139,15 @@ cl::opt<bool> ArchiveMap("print-armap", cl::desc("Print the archive map")); cl::alias ArchiveMaps("M", cl::desc("Alias for --print-armap"), cl::aliasopt(ArchiveMap), cl::Grouping); +enum Radix { d, o, x }; +cl::opt<Radix> + AddressRadix("radix", cl::desc("Radix (o/d/x) for printing symbol Values"), + cl::values(clEnumVal(d, "decimal"), clEnumVal(o, "octal"), + clEnumVal(x, "hexadecimal"), clEnumValEnd), + cl::init(x)); +cl::alias RadixAlias("t", cl::desc("Alias for --radix"), + cl::aliasopt(AddressRadix)); + cl::opt<bool> JustSymbolName("just-symbol-name", cl::desc("Print just the symbol's name")); cl::alias JustSymbolNames("j", cl::desc("Alias for --just-symbol-name"), @@ -180,6 +190,52 @@ static bool error(std::error_code EC, Twine Path = Twine()) { return false; } +// This version of error() prints the archive name and member name, for example: +// "libx.a(foo.o)" after the ToolName before the error message. It sets +// HadError but returns allowing the code to move on to other archive members. +static void error(llvm::Error E, StringRef FileName, const Archive::Child &C, + StringRef ArchitectureName = StringRef()) { + HadError = true; + errs() << ToolName << ": " << FileName; + + ErrorOr<StringRef> NameOrErr = C.getName(); + // TODO: if we have a error getting the name then it would be nice to print + // the index of which archive member this is and or its offset in the + // archive instead of "???" as the name. + if (NameOrErr.getError()) + errs() << "(" << "???" << ")"; + else + errs() << "(" << NameOrErr.get() << ")"; + + if (!ArchitectureName.empty()) + errs() << " (for architecture " << ArchitectureName << ") "; + + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(std::move(E), OS, ""); + OS.flush(); + errs() << " " << Buf << "\n"; +} + +// This version of error() prints the file name and which architecture slice it +// is from, for example: "foo.o (for architecture i386)" after the ToolName +// before the error message. It sets HadError but returns allowing the code to +// move on to other architecture slices. +static void error(llvm::Error E, StringRef FileName, + StringRef ArchitectureName = StringRef()) { + HadError = true; + errs() << ToolName << ": " << FileName; + + if (!ArchitectureName.empty()) + errs() << " (for architecture " << ArchitectureName << ") "; + + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(std::move(E), OS, ""); + OS.flush(); + errs() << " " << Buf << "\n"; +} + namespace { struct NMSymbol { uint64_t Address; @@ -260,15 +316,11 @@ static void darwinPrintSymbol(SymbolicFile &Obj, SymbolListT::iterator I, // use 1, 2 and 3 for section numbers. See below where they are used to // print out fake section names. NType |= MachO::N_SECT; - if(SymFlags & SymbolRef::SF_Const) + if (SymFlags & SymbolRef::SF_Const) NSect = 3; else { IRObjectFile *IRobj = dyn_cast<IRObjectFile>(&Obj); - char c = getSymbolNMTypeChar(*IRobj, I->Sym); - if (c == 't') - NSect = 1; - else - NSect = 2; + NSect = (getSymbolNMTypeChar(*IRobj, I->Sym) == 't') ? 1 : 2; } } if (SymFlags & SymbolRef::SF_Weak) @@ -338,7 +390,7 @@ static void darwinPrintSymbol(SymbolicFile &Obj, SymbolListT::iterator I, MachO::REFERENCE_FLAG_UNDEFINED_LAZY) outs() << "undefined [lazy bound]) "; else if ((NDesc & MachO::REFERENCE_TYPE) == - MachO::REFERENCE_FLAG_UNDEFINED_LAZY) + MachO::REFERENCE_FLAG_PRIVATE_UNDEFINED_LAZY) outs() << "undefined [private lazy bound]) "; else if ((NDesc & MachO::REFERENCE_TYPE) == MachO::REFERENCE_FLAG_PRIVATE_UNDEFINED_NON_LAZY) @@ -367,7 +419,14 @@ static void darwinPrintSymbol(SymbolicFile &Obj, SymbolListT::iterator I, outs() << "(?,?) "; break; } - section_iterator Sec = *MachO->getSymbolSection(I->Sym.getRawDataRefImpl()); + Expected<section_iterator> SecOrErr = + MachO->getSymbolSection(I->Sym.getRawDataRefImpl()); + if (!SecOrErr) { + consumeError(SecOrErr.takeError()); + outs() << "(?,?) "; + break; + } + section_iterator Sec = *SecOrErr; DataRefImpl Ref = Sec->getRawDataRefImpl(); StringRef SectionName; MachO->getSectionName(Ref, SectionName); @@ -538,8 +597,8 @@ static void darwinPrintStab(MachOObjectFile *MachO, SymbolListT::iterator I) { } static void sortAndPrintSymbolList(SymbolicFile &Obj, bool printName, - std::string ArchiveName, - std::string ArchitectureName) { + const std::string &ArchiveName, + const std::string &ArchitectureName) { if (!NoSort) { std::function<bool(const NMSymbol &, const NMSymbol &)> Cmp; if (NumericSort) @@ -570,32 +629,49 @@ static void sortAndPrintSymbolList(SymbolicFile &Obj, bool printName, if (isSymbolList64Bit(Obj)) { printBlanks = " "; printDashes = "----------------"; - printFormat = "%016" PRIx64; + switch (AddressRadix) { + case Radix::o: + printFormat = OutputFormat == posix ? "%" PRIo64 : "%016" PRIo64; + break; + case Radix::x: + printFormat = OutputFormat == posix ? "%" PRIx64 : "%016" PRIx64; + break; + default: + printFormat = OutputFormat == posix ? "%" PRId64 : "%016" PRId64; + } } else { printBlanks = " "; printDashes = "--------"; - printFormat = "%08" PRIx64; + switch (AddressRadix) { + case Radix::o: + printFormat = OutputFormat == posix ? "%" PRIo64 : "%08" PRIo64; + break; + case Radix::x: + printFormat = OutputFormat == posix ? "%" PRIx64 : "%08" PRIx64; + break; + default: + printFormat = OutputFormat == posix ? "%" PRId64 : "%08" PRId64; + } } for (SymbolListT::iterator I = SymbolList.begin(), E = SymbolList.end(); I != E; ++I) { uint32_t SymFlags = I->Sym.getFlags(); bool Undefined = SymFlags & SymbolRef::SF_Undefined; - if (!Undefined && UndefinedOnly) - continue; - if (Undefined && DefinedOnly) - continue; bool Global = SymFlags & SymbolRef::SF_Global; - if (!Global && ExternalOnly) - continue; - if (SizeSort && !PrintAddress) + if ((!Undefined && UndefinedOnly) || (Undefined && DefinedOnly) || + (!Global && ExternalOnly) || (SizeSort && !PrintAddress)) continue; if (PrintFileName) { if (!ArchitectureName.empty()) outs() << "(for architecture " << ArchitectureName << "):"; - if (!ArchiveName.empty()) - outs() << ArchiveName << ":"; - outs() << CurrentFilename << ": "; + if (OutputFormat == posix && !ArchiveName.empty()) + outs() << ArchiveName << "[" << CurrentFilename << "]: "; + else { + if (!ArchiveName.empty()) + outs() << ArchiveName << ":"; + outs() << CurrentFilename << ": "; + } } if ((JustSymbolName || (UndefinedOnly && isa<MachOObjectFile>(Obj) && OutputFormat != darwin)) && OutputFormat != posix) { @@ -606,8 +682,13 @@ static void sortAndPrintSymbolList(SymbolicFile &Obj, bool printName, char SymbolAddrStr[18] = ""; char SymbolSizeStr[18] = ""; - if (OutputFormat == sysv || I->TypeChar == 'U') - strcpy(SymbolAddrStr, printBlanks); + if (OutputFormat == sysv || I->TypeChar == 'U') { + if (OutputFormat == posix) + format(printFormat, I->Address) + .print(SymbolAddrStr, sizeof(SymbolAddrStr)); + else + strcpy(SymbolAddrStr, printBlanks); + } if (OutputFormat == sysv) strcpy(SymbolSizeStr, printBlanks); @@ -632,9 +713,9 @@ static void sortAndPrintSymbolList(SymbolicFile &Obj, bool printName, } else if (OutputFormat == posix) { outs() << I->Name << " " << I->TypeChar << " "; if (MachO) - outs() << I->Address << " " << "0" /* SymbolSizeStr */ << "\n"; + outs() << SymbolAddrStr << " " << "0" /* SymbolSizeStr */ << "\n"; else - outs() << SymbolAddrStr << SymbolSizeStr << "\n"; + outs() << SymbolAddrStr << " " << SymbolSizeStr << "\n"; } else if (OutputFormat == bsd || (OutputFormat == darwin && !MachO)) { if (PrintAddress) outs() << SymbolAddrStr << ' '; @@ -663,9 +744,11 @@ static char getSymbolNMTypeChar(ELFObjectFileBase &Obj, // OK, this is ELF elf_symbol_iterator SymI(I); - ErrorOr<elf_section_iterator> SecIOrErr = SymI->getSection(); - if (error(SecIOrErr.getError())) + Expected<elf_section_iterator> SecIOrErr = SymI->getSection(); + if (!SecIOrErr) { + consumeError(SecIOrErr.takeError()); return '?'; + } elf_section_iterator SecI = *SecIOrErr; if (SecI != Obj.section_end()) { @@ -690,9 +773,11 @@ static char getSymbolNMTypeChar(ELFObjectFileBase &Obj, } if (SymI->getELFType() == ELF::STT_SECTION) { - ErrorOr<StringRef> Name = SymI->getName(); - if (error(Name.getError())) + Expected<StringRef> Name = SymI->getName(); + if (!Name) { + consumeError(Name.takeError()); return '?'; + } return StringSwitch<char>(*Name) .StartsWith(".debug", 'N') .StartsWith(".note", 'n') @@ -707,9 +792,11 @@ static char getSymbolNMTypeChar(COFFObjectFile &Obj, symbol_iterator I) { // OK, this is COFF. symbol_iterator SymI(I); - ErrorOr<StringRef> Name = SymI->getName(); - if (error(Name.getError())) + Expected<StringRef> Name = SymI->getName(); + if (!Name) { + consumeError(Name.takeError()); return '?'; + } char Ret = StringSwitch<char>(*Name) .StartsWith(".debug", 'N') @@ -721,9 +808,11 @@ static char getSymbolNMTypeChar(COFFObjectFile &Obj, symbol_iterator I) { uint32_t Characteristics = 0; if (!COFF::isReservedSectionNumber(Symb.getSectionNumber())) { - ErrorOr<section_iterator> SecIOrErr = SymI->getSection(); - if (error(SecIOrErr.getError())) + Expected<section_iterator> SecIOrErr = SymI->getSection(); + if (!SecIOrErr) { + consumeError(SecIOrErr.takeError()); return '?'; + } section_iterator SecI = *SecIOrErr; const coff_section *Section = Obj.getCOFFSection(*SecI); Characteristics = Section->Characteristics; @@ -750,18 +839,10 @@ static char getSymbolNMTypeChar(COFFObjectFile &Obj, symbol_iterator I) { return '?'; } -static uint8_t getNType(MachOObjectFile &Obj, DataRefImpl Symb) { - if (Obj.is64Bit()) { - MachO::nlist_64 STE = Obj.getSymbol64TableEntry(Symb); - return STE.n_type; - } - MachO::nlist STE = Obj.getSymbolTableEntry(Symb); - return STE.n_type; -} - static char getSymbolNMTypeChar(MachOObjectFile &Obj, basic_symbol_iterator I) { DataRefImpl Symb = I->getRawDataRefImpl(); - uint8_t NType = getNType(Obj, Symb); + uint8_t NType = Obj.is64Bit() ? Obj.getSymbol64TableEntry(Symb).n_type + : Obj.getSymbolTableEntry(Symb).n_type; if (NType & MachO::N_STAB) return '-'; @@ -772,19 +853,23 @@ static char getSymbolNMTypeChar(MachOObjectFile &Obj, basic_symbol_iterator I) { case MachO::N_INDR: return 'i'; case MachO::N_SECT: { - section_iterator Sec = *Obj.getSymbolSection(Symb); + Expected<section_iterator> SecOrErr = Obj.getSymbolSection(Symb); + if (!SecOrErr) { + consumeError(SecOrErr.takeError()); + return 's'; + } + section_iterator Sec = *SecOrErr; DataRefImpl Ref = Sec->getRawDataRefImpl(); StringRef SectionName; Obj.getSectionName(Ref, SectionName); StringRef SegmentName = Obj.getSectionFinalSegmentName(Ref); if (SegmentName == "__TEXT" && SectionName == "__text") return 't'; - else if (SegmentName == "__DATA" && SectionName == "__data") + if (SegmentName == "__DATA" && SectionName == "__data") return 'd'; - else if (SegmentName == "__DATA" && SectionName == "__bss") + if (SegmentName == "__DATA" && SectionName == "__bss") return 'b'; - else - return 's'; + return 's'; } } @@ -792,35 +877,27 @@ static char getSymbolNMTypeChar(MachOObjectFile &Obj, basic_symbol_iterator I) { } static char getSymbolNMTypeChar(const GlobalValue &GV) { - if (GV.getType()->getElementType()->isFunctionTy()) - return 't'; // FIXME: should we print 'b'? At the IR level we cannot be sure if this // will be in bss or not, but we could approximate. - return 'd'; + return GV.getValueType()->isFunctionTy() ? 't' : 'd'; } static char getSymbolNMTypeChar(IRObjectFile &Obj, basic_symbol_iterator I) { const GlobalValue *GV = Obj.getSymbolGV(I->getRawDataRefImpl()); - if (!GV) - return 't'; - return getSymbolNMTypeChar(*GV); + return !GV ? 't' : getSymbolNMTypeChar(*GV); } static bool isObject(SymbolicFile &Obj, basic_symbol_iterator I) { - auto *ELF = dyn_cast<ELFObjectFileBase>(&Obj); - if (!ELF) - return false; - - return elf_symbol_iterator(I)->getELFType() == ELF::STT_OBJECT; + return !dyn_cast<ELFObjectFileBase>(&Obj) + ? false + : elf_symbol_iterator(I)->getELFType() == ELF::STT_OBJECT; } static char getNMTypeChar(SymbolicFile &Obj, basic_symbol_iterator I) { uint32_t Symflags = I->getFlags(); if ((Symflags & object::SymbolRef::SF_Weak) && !isa<MachOObjectFile>(Obj)) { char Ret = isObject(Obj, I) ? 'v' : 'w'; - if (!(Symflags & object::SymbolRef::SF_Undefined)) - Ret = toupper(Ret); - return Ret; + return (!(Symflags & object::SymbolRef::SF_Undefined)) ? toupper(Ret) : Ret; } if (Symflags & object::SymbolRef::SF_Undefined) @@ -835,10 +912,8 @@ static char getNMTypeChar(SymbolicFile &Obj, basic_symbol_iterator I) { else if (IRObjectFile *IR = dyn_cast<IRObjectFile>(&Obj)) { Ret = getSymbolNMTypeChar(*IR, I); Triple Host(sys::getDefaultTargetTriple()); - if (Ret == 'd' && Host.isOSDarwin()) { - if(Symflags & SymbolRef::SF_Const) - Ret = 's'; - } + if (Ret == 'd' && Host.isOSDarwin() && Symflags & SymbolRef::SF_Const) + Ret = 's'; } else if (COFFObjectFile *COFF = dyn_cast<COFFObjectFile>(&Obj)) Ret = getSymbolNMTypeChar(*COFF, I); @@ -861,9 +936,8 @@ static char getNMTypeChar(SymbolicFile &Obj, basic_symbol_iterator I) { // file or zero it is not present. static unsigned getNsectForSegSect(MachOObjectFile *Obj) { unsigned Nsect = 1; - for (section_iterator I = Obj->section_begin(), E = Obj->section_end(); - I != E; ++I) { - DataRefImpl Ref = I->getRawDataRefImpl(); + for (auto &S : Obj->sections()) { + DataRefImpl Ref = S.getRawDataRefImpl(); StringRef SectionName; Obj->getSectionName(Ref, SectionName); StringRef SegmentName = Obj->getSectionFinalSegmentName(Ref); @@ -883,20 +957,16 @@ static unsigned getNsectInMachO(MachOObjectFile &Obj, BasicSymbolRef Sym) { DataRefImpl Symb = Sym.getRawDataRefImpl(); if (Obj.is64Bit()) { MachO::nlist_64 STE = Obj.getSymbol64TableEntry(Symb); - if ((STE.n_type & MachO::N_TYPE) == MachO::N_SECT) - return STE.n_sect; - return 0; + return (STE.n_type & MachO::N_TYPE) == MachO::N_SECT ? STE.n_sect : 0; } MachO::nlist STE = Obj.getSymbolTableEntry(Symb); - if ((STE.n_type & MachO::N_TYPE) == MachO::N_SECT) - return STE.n_sect; - return 0; + return (STE.n_type & MachO::N_TYPE) == MachO::N_SECT ? STE.n_sect : 0; } -static void dumpSymbolNamesFromObject(SymbolicFile &Obj, bool printName, - std::string ArchiveName = std::string(), - std::string ArchitectureName = - std::string()) { +static void +dumpSymbolNamesFromObject(SymbolicFile &Obj, bool printName, + const std::string &ArchiveName = std::string(), + const std::string &ArchitectureName = std::string()) { auto Symbols = Obj.symbols(); if (DynamicSyms) { const auto *E = dyn_cast<ELFObjectFileBase>(&Obj); @@ -945,14 +1015,19 @@ static void dumpSymbolNamesFromObject(SymbolicFile &Obj, bool printName, } if (PrintAddress && isa<ObjectFile>(Obj)) { SymbolRef SymRef(Sym); - ErrorOr<uint64_t> AddressOrErr = SymRef.getAddress(); - if (error(AddressOrErr.getError())) + Expected<uint64_t> AddressOrErr = SymRef.getAddress(); + if (!AddressOrErr) { + consumeError(AddressOrErr.takeError()); break; + } S.Address = *AddressOrErr; } S.TypeChar = getNMTypeChar(Obj, Sym); - if (error(Sym.printName(OS))) - break; + std::error_code EC = Sym.printName(OS); + if (EC && MachO) + OS << "bad string index"; + else + error(EC); OS << '\0'; S.Sym = Sym; SymbolList.push_back(S); @@ -985,10 +1060,10 @@ static bool checkMachOAndArchFlags(SymbolicFile *O, std::string &Filename) { Triple T; if (MachO->is64Bit()) { H_64 = MachO->MachOObjectFile::getHeader64(); - T = MachOObjectFile::getArch(H_64.cputype, H_64.cpusubtype); + T = MachOObjectFile::getArchTriple(H_64.cputype, H_64.cpusubtype); } else { H = MachO->MachOObjectFile::getHeader(); - T = MachOObjectFile::getArch(H.cputype, H.cpusubtype); + T = MachOObjectFile::getArchTriple(H.cputype, H.cpusubtype); } if (std::none_of( ArchFlags.begin(), ArchFlags.end(), @@ -1005,11 +1080,13 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { if (error(BufferOrErr.getError(), Filename)) return; - LLVMContext &Context = getGlobalContext(); - ErrorOr<std::unique_ptr<Binary>> BinaryOrErr = createBinary( + LLVMContext Context; + Expected<std::unique_ptr<Binary>> BinaryOrErr = createBinary( BufferOrErr.get()->getMemBufferRef(), NoLLVMBitcode ? nullptr : &Context); - if (error(BinaryOrErr.getError(), Filename)) + if (!BinaryOrErr) { + error(errorToErrorCode(BinaryOrErr.takeError()), Filename); return; + } Binary &Bin = *BinaryOrErr.get(); if (Archive *A = dyn_cast<Archive>(&Bin)) { @@ -1032,27 +1109,31 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { } } - for (Archive::child_iterator I = A->child_begin(), E = A->child_end(); - I != E; ++I) { - if (error(I->getError())) - return; - auto &C = I->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(&Context); - if (ChildOrErr.getError()) - continue; - if (SymbolicFile *O = dyn_cast<SymbolicFile>(&*ChildOrErr.get())) { - if (!checkMachOAndArchFlags(O, Filename)) - return; - if (!PrintFileName) { - outs() << "\n"; - if (isa<MachOObjectFile>(O)) { - outs() << Filename << "(" << O->getFileName() << ")"; - } else - outs() << O->getFileName(); - outs() << ":\n"; + { + Error Err; + for (auto &C : A->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(&Context); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType(ChildOrErr.takeError())) + error(std::move(E), Filename, C); + continue; + } + if (SymbolicFile *O = dyn_cast<SymbolicFile>(&*ChildOrErr.get())) { + if (!checkMachOAndArchFlags(O, Filename)) + return; + if (!PrintFileName) { + outs() << "\n"; + if (isa<MachOObjectFile>(O)) { + outs() << Filename << "(" << O->getFileName() << ")"; + } else + outs() << O->getFileName(); + outs() << ":\n"; + } + dumpSymbolNamesFromObject(*O, false, Filename); } - dumpSymbolNamesFromObject(*O, false, Filename); } + if (Err) + error(std::move(Err), A->getFileName()); } return; } @@ -1068,7 +1149,7 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { I != E; ++I) { if (ArchFlags[i] == I->getArchTypeName()) { ArchFound = true; - ErrorOr<std::unique_ptr<ObjectFile>> ObjOrErr = + Expected<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); std::string ArchiveName; std::string ArchitectureName; @@ -1086,19 +1167,26 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { } dumpSymbolNamesFromObject(Obj, false, ArchiveName, ArchitectureName); - } else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = + } else if (auto E = isNotObjectErrorInvalidFileType( + ObjOrErr.takeError())) { + error(std::move(E), Filename, ArchFlags.size() > 1 ? + StringRef(I->getArchTypeName()) : StringRef()); + continue; + } else if (Expected<std::unique_ptr<Archive>> AOrErr = I->getAsArchive()) { std::unique_ptr<Archive> &A = *AOrErr; - for (Archive::child_iterator AI = A->child_begin(), - AE = A->child_end(); - AI != AE; ++AI) { - if (error(AI->getError())) - return; - auto &C = AI->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = + Error Err; + for (auto &C : A->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(&Context); - if (ChildOrErr.getError()) + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType( + ChildOrErr.takeError())) { + error(std::move(E), Filename, C, ArchFlags.size() > 1 ? + StringRef(I->getArchTypeName()) : StringRef()); + } continue; + } if (SymbolicFile *O = dyn_cast<SymbolicFile>(&*ChildOrErr.get())) { if (PrintFileName) { @@ -1118,6 +1206,14 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { ArchitectureName); } } + if (Err) + error(std::move(Err), A->getFileName()); + } else { + consumeError(AOrErr.takeError()); + error(Filename + " for architecture " + + StringRef(I->getArchTypeName()) + + " is not a Mach-O file or an archive file", + "Mach-O universal file"); } } } @@ -1137,25 +1233,29 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { E = UB->end_objects(); I != E; ++I) { if (HostArchName == I->getArchTypeName()) { - ErrorOr<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); + Expected<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); std::string ArchiveName; ArchiveName.clear(); if (ObjOrErr) { ObjectFile &Obj = *ObjOrErr.get(); dumpSymbolNamesFromObject(Obj, false); - } else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = + } else if (auto E = isNotObjectErrorInvalidFileType( + ObjOrErr.takeError())) { + error(std::move(E), Filename); + return; + } else if (Expected<std::unique_ptr<Archive>> AOrErr = I->getAsArchive()) { std::unique_ptr<Archive> &A = *AOrErr; - for (Archive::child_iterator AI = A->child_begin(), - AE = A->child_end(); - AI != AE; ++AI) { - if (error(AI->getError())) - return; - auto &C = AI->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = + Error Err; + for (auto &C : A->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(&Context); - if (ChildOrErr.getError()) + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType( + ChildOrErr.takeError())) + error(std::move(E), Filename, C); continue; + } if (SymbolicFile *O = dyn_cast<SymbolicFile>(&*ChildOrErr.get())) { if (PrintFileName) @@ -1167,6 +1267,14 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { dumpSymbolNamesFromObject(*O, false, ArchiveName); } } + if (Err) + error(std::move(Err), A->getFileName()); + } else { + consumeError(AOrErr.takeError()); + error(Filename + " for architecture " + + StringRef(I->getArchTypeName()) + + " is not a Mach-O file or an archive file", + "Mach-O universal file"); } return; } @@ -1178,7 +1286,7 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { for (MachOUniversalBinary::object_iterator I = UB->begin_objects(), E = UB->end_objects(); I != E; ++I) { - ErrorOr<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); + Expected<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); std::string ArchiveName; std::string ArchitectureName; ArchiveName.clear(); @@ -1197,16 +1305,25 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { outs() << ":\n"; } dumpSymbolNamesFromObject(Obj, false, ArchiveName, ArchitectureName); - } else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = I->getAsArchive()) { + } else if (auto E = isNotObjectErrorInvalidFileType( + ObjOrErr.takeError())) { + error(std::move(E), Filename, moreThanOneArch ? + StringRef(I->getArchTypeName()) : StringRef()); + continue; + } else if (Expected<std::unique_ptr<Archive>> AOrErr = + I->getAsArchive()) { std::unique_ptr<Archive> &A = *AOrErr; - for (Archive::child_iterator AI = A->child_begin(), AE = A->child_end(); - AI != AE; ++AI) { - if (error(AI->getError())) - return; - auto &C = AI->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(&Context); - if (ChildOrErr.getError()) + Error Err; + for (auto &C : A->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = + C.getAsBinary(&Context); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType( + ChildOrErr.takeError())) + error(std::move(E), Filename, C, moreThanOneArch ? + StringRef(ArchitectureName) : StringRef()); continue; + } if (SymbolicFile *O = dyn_cast<SymbolicFile>(&*ChildOrErr.get())) { if (PrintFileName) { ArchiveName = A->getFileName(); @@ -1226,6 +1343,14 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { dumpSymbolNamesFromObject(*O, false, ArchiveName, ArchitectureName); } } + if (Err) + error(std::move(Err), A->getFileName()); + } else { + consumeError(AOrErr.takeError()); + error(Filename + " for architecture " + + StringRef(I->getArchTypeName()) + + " is not a Mach-O file or an archive file", + "Mach-O universal file"); } } return; @@ -1234,14 +1359,12 @@ static void dumpSymbolNamesFromFile(std::string &Filename) { if (!checkMachOAndArchFlags(O, Filename)) return; dumpSymbolNamesFromObject(*O, true); - return; } - error("unrecognizable file type", Filename); } int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. @@ -1251,6 +1374,7 @@ int main(int argc, char **argv) { if (error(sys::ChangeStdinToBinary())) return 1; + // These calls are needed so that we can read bitcode correctly. llvm::InitializeAllTargetInfos(); llvm::InitializeAllTargetMCs(); llvm::InitializeAllAsmParsers(); @@ -1270,15 +1394,10 @@ int main(int argc, char **argv) { PrintAddress = false; if (OutputFormat == sysv || SizeSort) PrintSize = true; - - switch (InputFilenames.size()) { - case 0: + if (InputFilenames.empty()) InputFilenames.push_back("a.out"); - case 1: - break; - default: + if (InputFilenames.size() > 1) MultipleFiles = true; - } for (unsigned i = 0; i < ArchFlags.size(); ++i) { if (ArchFlags[i] == "all") { @@ -1299,6 +1418,4 @@ int main(int argc, char **argv) { if (HadError) return 1; - - return 0; } diff --git a/tools/llvm-objdump/CMakeLists.txt b/tools/llvm-objdump/CMakeLists.txt index d0dd4ac0182c3..551847f34eef0 100644 --- a/tools/llvm-objdump/CMakeLists.txt +++ b/tools/llvm-objdump/CMakeLists.txt @@ -17,3 +17,7 @@ add_llvm_tool(llvm-objdump ELFDump.cpp MachODump.cpp ) + +if(HAVE_LIBXAR) + target_link_libraries(llvm-objdump ${XAR_LIB}) +endif() diff --git a/tools/llvm-objdump/COFFDump.cpp b/tools/llvm-objdump/COFFDump.cpp index 5d21b3320e748..3ec6a1f73750e 100644 --- a/tools/llvm-objdump/COFFDump.cpp +++ b/tools/llvm-objdump/COFFDump.cpp @@ -161,13 +161,13 @@ static std::error_code resolveSectionAndAddress(const COFFObjectFile *Obj, const SymbolRef &Sym, const coff_section *&ResolvedSection, uint64_t &ResolvedAddr) { - ErrorOr<uint64_t> ResolvedAddrOrErr = Sym.getAddress(); - if (std::error_code EC = ResolvedAddrOrErr.getError()) - return EC; + Expected<uint64_t> ResolvedAddrOrErr = Sym.getAddress(); + if (!ResolvedAddrOrErr) + return errorToErrorCode(ResolvedAddrOrErr.takeError()); ResolvedAddr = *ResolvedAddrOrErr; - ErrorOr<section_iterator> Iter = Sym.getSection(); - if (std::error_code EC = Iter.getError()) - return EC; + Expected<section_iterator> Iter = Sym.getSection(); + if (!Iter) + return errorToErrorCode(Iter.takeError()); ResolvedSection = Obj->getCOFFSection(**Iter); return std::error_code(); } @@ -215,9 +215,9 @@ static std::error_code resolveSymbolName(const std::vector<RelocationRef> &Rels, SymbolRef Sym; if (std::error_code EC = resolveSymbol(Rels, Offset, Sym)) return EC; - ErrorOr<StringRef> NameOrErr = Sym.getName(); - if (std::error_code EC = NameOrErr.getError()) - return EC; + Expected<StringRef> NameOrErr = Sym.getName(); + if (!NameOrErr) + return errorToErrorCode(NameOrErr.takeError()); Name = *NameOrErr; return std::error_code(); } @@ -252,6 +252,56 @@ printSEHTable(const COFFObjectFile *Obj, uint32_t TableVA, int Count) { outs() << "\n\n"; } +template <typename T> +static void printTLSDirectoryT(const coff_tls_directory<T> *TLSDir) { + size_t FormatWidth = sizeof(T) * 2; + outs() << "TLS directory:" + << "\n StartAddressOfRawData: " + << format_hex(TLSDir->StartAddressOfRawData, FormatWidth) + << "\n EndAddressOfRawData: " + << format_hex(TLSDir->EndAddressOfRawData, FormatWidth) + << "\n AddressOfIndex: " + << format_hex(TLSDir->AddressOfIndex, FormatWidth) + << "\n AddressOfCallBacks: " + << format_hex(TLSDir->AddressOfCallBacks, FormatWidth) + << "\n SizeOfZeroFill: " + << TLSDir->SizeOfZeroFill + << "\n Characteristics: " + << TLSDir->Characteristics + << "\n Alignment: " + << TLSDir->getAlignment() + << "\n\n"; +} + +static void printTLSDirectory(const COFFObjectFile *Obj) { + const pe32_header *PE32Header; + error(Obj->getPE32Header(PE32Header)); + + const pe32plus_header *PE32PlusHeader; + error(Obj->getPE32PlusHeader(PE32PlusHeader)); + + // Skip if it's not executable. + if (!PE32Header && !PE32PlusHeader) + return; + + const data_directory *DataDir; + error(Obj->getDataDirectory(COFF::TLS_TABLE, DataDir)); + uintptr_t IntPtr = 0; + if (DataDir->RelativeVirtualAddress == 0) + return; + error(Obj->getRvaPtr(DataDir->RelativeVirtualAddress, IntPtr)); + + if (PE32Header) { + auto *TLSDir = reinterpret_cast<const coff_tls_directory32 *>(IntPtr); + printTLSDirectoryT(TLSDir); + } else { + auto *TLSDir = reinterpret_cast<const coff_tls_directory64 *>(IntPtr); + printTLSDirectoryT(TLSDir); + } + + outs() << "\n"; +} + static void printLoadConfiguration(const COFFObjectFile *Obj) { // Skip if it's not executable. const pe32_header *PE32Header; @@ -302,11 +352,11 @@ static void printImportTables(const COFFObjectFile *Obj) { if (I == E) return; outs() << "The Import Tables:\n"; - for (; I != E; I = ++I) { + for (const ImportDirectoryEntryRef &DirRef : Obj->import_directories()) { const import_directory_table_entry *Dir; StringRef Name; - if (I->getImportTableEntry(Dir)) return; - if (I->getName(Name)) return; + if (DirRef.getImportTableEntry(Dir)) return; + if (DirRef.getName(Name)) return; outs() << format(" lookup %08x time %08x fwd %08x name %08x addr %08x\n\n", static_cast<uint32_t>(Dir->ImportLookupTableRVA), @@ -316,17 +366,23 @@ static void printImportTables(const COFFObjectFile *Obj) { static_cast<uint32_t>(Dir->ImportAddressTableRVA)); outs() << " DLL Name: " << Name << "\n"; outs() << " Hint/Ord Name\n"; - const import_lookup_table_entry32 *entry; - if (I->getImportLookupEntry(entry)) - return; - for (; entry->Data; ++entry) { - if (entry->isOrdinal()) { - outs() << format(" % 6d\n", entry->getOrdinal()); + for (const ImportedSymbolRef &Entry : DirRef.imported_symbols()) { + bool IsOrdinal; + if (Entry.isOrdinal(IsOrdinal)) + return; + if (IsOrdinal) { + uint16_t Ordinal; + if (Entry.getOrdinal(Ordinal)) + return; + outs() << format(" % 6d\n", Ordinal); continue; } + uint32_t HintNameRVA; + if (Entry.getHintNameRVA(HintNameRVA)) + return; uint16_t Hint; StringRef Name; - if (Obj->getHintName(entry->getHintNameRVA(), Hint, Name)) + if (Obj->getHintName(HintNameRVA, Hint, Name)) return; outs() << format(" % 6d ", Hint) << Name << "\n"; } @@ -555,6 +611,7 @@ void llvm::printCOFFUnwindInfo(const COFFObjectFile *Obj) { void llvm::printCOFFFileHeader(const object::ObjectFile *Obj) { const COFFObjectFile *file = dyn_cast<const COFFObjectFile>(Obj); + printTLSDirectory(file); printLoadConfiguration(file); printImportTables(file); printExportTable(file); @@ -602,6 +659,13 @@ void llvm::printCOFFSymbolTable(const COFFObjectFile *coff) { SI = SI + Symbol->getNumberOfAuxSymbols(); break; + } else if (Symbol->isWeakExternal()) { + const coff_aux_weak_external *awe; + error(coff->getAuxSymbol<coff_aux_weak_external>(SI + 1, awe)); + + outs() << "AUX " << format("indx %d srch %d\n", + static_cast<uint32_t>(awe->TagIndex), + static_cast<uint32_t>(awe->Characteristics)); } else { outs() << "AUX Unknown\n"; } diff --git a/tools/llvm-objdump/MachODump.cpp b/tools/llvm-objdump/MachODump.cpp index 258c0b520a3d4..4d950f1d7bd7c 100644 --- a/tools/llvm-objdump/MachODump.cpp +++ b/tools/llvm-objdump/MachODump.cpp @@ -11,9 +11,9 @@ // //===----------------------------------------------------------------------===// +#include "llvm/Object/MachO.h" #include "llvm-objdump.h" #include "llvm-c/Disassembler.h" -#include "llvm/ADT/Optional.h" #include "llvm/ADT/STLExtras.h" #include "llvm/ADT/StringExtras.h" #include "llvm/ADT/Triple.h" @@ -22,14 +22,13 @@ #include "llvm/DebugInfo/DWARF/DWARFContext.h" #include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCContext.h" -#include "llvm/MC/MCDisassembler.h" +#include "llvm/MC/MCDisassembler/MCDisassembler.h" #include "llvm/MC/MCInst.h" #include "llvm/MC/MCInstPrinter.h" #include "llvm/MC/MCInstrDesc.h" #include "llvm/MC/MCInstrInfo.h" #include "llvm/MC/MCRegisterInfo.h" #include "llvm/MC/MCSubtargetInfo.h" -#include "llvm/Object/MachO.h" #include "llvm/Object/MachOUniversal.h" #include "llvm/Support/Casting.h" #include "llvm/Support/CommandLine.h" @@ -43,6 +42,7 @@ #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/TargetRegistry.h" #include "llvm/Support/TargetSelect.h" +#include "llvm/Support/ToolOutputFile.h" #include "llvm/Support/raw_ostream.h" #include <algorithm> #include <cstring> @@ -52,6 +52,12 @@ #include <cxxabi.h> #endif +#ifdef HAVE_LIBXAR +extern "C" { +#include <xar/xar.h> +} +#endif + using namespace llvm; using namespace object; @@ -143,11 +149,18 @@ static const Target *GetTarget(const MachOObjectFile *MachOObj, const char **McpuDefault, const Target **ThumbTarget) { // Figure out the target triple. + llvm::Triple TT(TripleName); if (TripleName.empty()) { - llvm::Triple TT("unknown-unknown-unknown"); - llvm::Triple ThumbTriple = Triple(); - TT = MachOObj->getArch(McpuDefault, &ThumbTriple); + TT = MachOObj->getArchTriple(McpuDefault); TripleName = TT.str(); + } + + if (TT.getArch() == Triple::arm) { + // We've inferred a 32-bit ARM target from the object file. All MachO CPUs + // that support ARM are also capable of Thumb mode. + llvm::Triple ThumbTriple = TT; + std::string ThumbName = (Twine("thumb") + TT.getArchName().substr(3)).str(); + ThumbTriple.setArchName(ThumbName); ThumbTripleName = ThumbTriple.str(); } @@ -172,8 +185,26 @@ static const Target *GetTarget(const MachOObjectFile *MachOObj, struct SymbolSorter { bool operator()(const SymbolRef &A, const SymbolRef &B) { - uint64_t AAddr = (A.getType() != SymbolRef::ST_Function) ? 0 : A.getValue(); - uint64_t BAddr = (B.getType() != SymbolRef::ST_Function) ? 0 : B.getValue(); + Expected<SymbolRef::Type> ATypeOrErr = A.getType(); + if (!ATypeOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(ATypeOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } + SymbolRef::Type AType = *ATypeOrErr; + Expected<SymbolRef::Type> BTypeOrErr = B.getType(); + if (!BTypeOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(BTypeOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } + SymbolRef::Type BType = *BTypeOrErr; + uint64_t AAddr = (AType != SymbolRef::ST_Function) ? 0 : A.getValue(); + uint64_t BAddr = (BType != SymbolRef::ST_Function) ? 0 : B.getValue(); return AAddr < BAddr; } }; @@ -266,9 +297,14 @@ static void getSectionsAndSymbols(MachOObjectFile *MachOObj, SmallVectorImpl<uint64_t> &FoundFns, uint64_t &BaseSegmentAddress) { for (const SymbolRef &Symbol : MachOObj->symbols()) { - ErrorOr<StringRef> SymName = Symbol.getName(); - if (std::error_code EC = SymName.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymName = Symbol.getName(); + if (!SymName) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymName.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } if (!SymName->startswith("ltmp")) Symbols.push_back(Symbol); } @@ -324,7 +360,7 @@ static void PrintIndirectSymbolTable(MachOObjectFile *O, bool verbose, if (cputype & MachO::CPU_ARCH_ABI64) outs() << format("0x%016" PRIx64, addr + j * stride) << " "; else - outs() << format("0x%08" PRIx32, addr + j * stride) << " "; + outs() << format("0x%08" PRIx32, (uint32_t)addr + j * stride) << " "; MachO::dysymtab_command Dysymtab = O->getDysymtabLoadCommand(); uint32_t indirect_symbol = O->getIndirectSymbolTableEntry(Dysymtab, n + j); if (indirect_symbol == MachO::INDIRECT_SYMBOL_LOCAL) { @@ -346,9 +382,14 @@ static void PrintIndirectSymbolTable(MachOObjectFile *O, bool verbose, if (indirect_symbol < Symtab.nsyms) { symbol_iterator Sym = O->getSymbolByIndex(indirect_symbol); SymbolRef Symbol = *Sym; - ErrorOr<StringRef> SymName = Symbol.getName(); - if (std::error_code EC = SymName.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymName = Symbol.getName(); + if (!SymName) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymName.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } outs() << *SymName; } else { outs() << "?"; @@ -573,13 +614,26 @@ static void CreateSymbolAddressMap(MachOObjectFile *O, SymbolAddressMap *AddrMap) { // Create a map of symbol addresses to symbol names. for (const SymbolRef &Symbol : O->symbols()) { - SymbolRef::Type ST = Symbol.getType(); + Expected<SymbolRef::Type> STOrErr = Symbol.getType(); + if (!STOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(STOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } + SymbolRef::Type ST = *STOrErr; if (ST == SymbolRef::ST_Function || ST == SymbolRef::ST_Data || ST == SymbolRef::ST_Other) { uint64_t Address = Symbol.getValue(); - ErrorOr<StringRef> SymNameOrErr = Symbol.getName(); - if (std::error_code EC = SymNameOrErr.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymNameOrErr = Symbol.getName(); + if (!SymNameOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymNameOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } StringRef SymName = *SymNameOrErr; if (!SymName.startswith(".objc")) (*AddrMap)[Address] = SymName; @@ -813,9 +867,14 @@ static void DumpLiteralPointerSection(MachOObjectFile *O, [&](const std::pair<uint64_t, SymbolRef> &P) { return P.first == i; }); if (Reloc != Relocs.end()) { symbol_iterator RelocSym = Reloc->second; - ErrorOr<StringRef> SymName = RelocSym->getName(); - if (std::error_code EC = SymName.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymName = RelocSym->getName(); + if (!SymName) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymName.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } outs() << "external relocation entry for symbol:" << *SymName << "\n"; continue; } @@ -964,10 +1023,10 @@ static void DumpRawSectionContents(MachOObjectFile *O, const char *sect, if (O->is64Bit()) outs() << format("%016" PRIx64, addr) << "\t"; else - outs() << format("%08" PRIx64, sect) << "\t"; + outs() << format("%08" PRIx64, addr) << "\t"; for (j = 0; j < 4 * sizeof(int32_t) && i + j < size; j += sizeof(int32_t)) { - if (i + j + sizeof(int32_t) < size) { + if (i + j + sizeof(int32_t) <= size) { uint32_t long_word; memcpy(&long_word, sect + i + j, sizeof(int32_t)); if (O->isLittleEndian() != sys::IsLittleEndianHost) @@ -975,7 +1034,7 @@ static void DumpRawSectionContents(MachOObjectFile *O, const char *sect, outs() << format("%08" PRIx32, long_word) << " "; } else { for (uint32_t k = 0; i + j + k < size; k++) { - uint8_t byte_word = *(sect + i + j); + uint8_t byte_word = *(sect + i + j + k); outs() << format("%02" PRIx32, (uint32_t)byte_word) << " "; } } @@ -989,6 +1048,12 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, StringRef DisSegName, StringRef DisSectName); static void DumpProtocolSection(MachOObjectFile *O, const char *sect, uint32_t size, uint32_t addr); +#ifdef HAVE_LIBXAR +static void DumpBitcodeSection(MachOObjectFile *O, const char *sect, + uint32_t size, bool verbose, + bool PrintXarHeader, bool PrintXarFileHeaders, + std::string XarMemberName); +#endif // defined(HAVE_LIBXAR) static void DumpSectionContents(StringRef Filename, MachOObjectFile *O, bool verbose) { @@ -1050,6 +1115,13 @@ static void DumpSectionContents(StringRef Filename, MachOObjectFile *O, DumpProtocolSection(O, sect, sect_size, sect_addr); continue; } +#ifdef HAVE_LIBXAR + if (SegName == "__LLVM" && SectName == "__bundle") { + DumpBitcodeSection(O, sect, sect_size, verbose, !NoSymbolicOperands, + ArchiveHeaders, ""); + continue; + } +#endif // defined(HAVE_LIBXAR) switch (section_type) { case MachO::S_REGULAR: DumpRawSectionContents(O, sect, sect_size, sect_addr); @@ -1127,10 +1199,10 @@ static bool checkMachOAndArchFlags(ObjectFile *O, StringRef Filename) { Triple T; if (MachO->is64Bit()) { H_64 = MachO->MachOObjectFile::getHeader64(); - T = MachOObjectFile::getArch(H_64.cputype, H_64.cpusubtype); + T = MachOObjectFile::getArchTriple(H_64.cputype, H_64.cpusubtype); } else { H = MachO->MachOObjectFile::getHeader(); - T = MachOObjectFile::getArch(H.cputype, H.cpusubtype); + T = MachOObjectFile::getArchTriple(H.cputype, H.cpusubtype); } unsigned i; for (i = 0; i < ArchFlags.size(); ++i) { @@ -1159,7 +1231,7 @@ static void ProcessMachO(StringRef Filename, MachOObjectFile *MachOOF, // If we are doing some processing here on the Mach-O file print the header // info. And don't print it otherwise like in the case of printing the // UniversalHeaders or ArchiveHeaders. - if (Disassemble || PrivateHeaders || ExportsTrie || Rebase || Bind || + if (Disassemble || PrivateHeaders || ExportsTrie || Rebase || Bind || SymbolTable || LazyBind || WeakBind || IndirectSymbols || DataInCode || LinkOptHints || DylibsUsed || DylibId || ObjcMetaData || (FilterSections.size() != 0)) { outs() << Filename; @@ -1192,8 +1264,10 @@ static void ProcessMachO(StringRef Filename, MachOObjectFile *MachOOF, PrintDylibs(MachOOF, false); if (DylibId) PrintDylibs(MachOOF, true); - if (SymbolTable) - PrintSymbolTable(MachOOF); + if (SymbolTable) { + StringRef ArchiveName = ArchiveMemberName == StringRef() ? "" : Filename; + PrintSymbolTable(MachOOF, ArchiveName, ArchitectureName); + } if (UnwindInfo) printMachOUnwindInfo(MachOOF); if (PrivateHeaders) { @@ -1214,6 +1288,12 @@ static void ProcessMachO(StringRef Filename, MachOObjectFile *MachOOF, printLazyBindTable(MachOOF); if (WeakBind) printWeakBindTable(MachOOF); + + if (DwarfDumpType != DIDT_Null) { + std::unique_ptr<DIContext> DICtx(new DWARFContextInMemory(*MachOOF)); + // Dump the complete DWARF structure. + DICtx->dump(outs(), DwarfDumpType, true /* DumpEH */); + } } // printUnknownCPUType() helps print_fat_headers for unknown CPU's. @@ -1323,9 +1403,12 @@ static void printCPUType(uint32_t cputype, uint32_t cpusubtype) { static void printMachOUniversalHeaders(const object::MachOUniversalBinary *UB, bool verbose) { outs() << "Fat headers\n"; - if (verbose) - outs() << "fat_magic FAT_MAGIC\n"; - else + if (verbose) { + if (UB->getMagic() == MachO::FAT_MAGIC) + outs() << "fat_magic FAT_MAGIC\n"; + else // UB->getMagic() == MachO::FAT_MAGIC_64 + outs() << "fat_magic FAT_MAGIC_64\n"; + } else outs() << "fat_magic " << format("0x%" PRIx32, MachO::FAT_MAGIC) << "\n"; uint32_t nfat_arch = UB->getNumberOfObjects(); @@ -1452,13 +1535,11 @@ static void printArchiveChild(const Archive::Child &C, bool verbose, } static void printArchiveHeaders(Archive *A, bool verbose, bool print_offset) { - for (Archive::child_iterator I = A->child_begin(false), E = A->child_end(); - I != E; ++I) { - if (std::error_code EC = I->getError()) - report_fatal_error(EC.message()); - const Archive::Child &C = **I; + Error Err; + for (const auto &C : A->children(Err, false)) printArchiveChild(C, verbose, print_offset); - } + if (Err) + report_fatal_error(std::move(Err)); } // ParseInputMachO() parses the named Mach-O file in Filename and handles the @@ -1480,29 +1561,31 @@ void llvm::ParseInputMachO(StringRef Filename) { } // Attempt to open the binary. - ErrorOr<OwningBinary<Binary>> BinaryOrErr = createBinary(Filename); - if (std::error_code EC = BinaryOrErr.getError()) - report_error(Filename, EC); + Expected<OwningBinary<Binary>> BinaryOrErr = createBinary(Filename); + if (!BinaryOrErr) + report_error(Filename, BinaryOrErr.takeError()); Binary &Bin = *BinaryOrErr.get().getBinary(); if (Archive *A = dyn_cast<Archive>(&Bin)) { outs() << "Archive : " << Filename << "\n"; if (ArchiveHeaders) printArchiveHeaders(A, !NonVerbose, ArchiveMemberOffsets); - for (Archive::child_iterator I = A->child_begin(), E = A->child_end(); - I != E; ++I) { - if (std::error_code EC = I->getError()) - report_error(Filename, EC); - auto &C = I->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); - if (ChildOrErr.getError()) + Error Err; + for (auto &C : A->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType(ChildOrErr.takeError())) + report_error(Filename, C, std::move(E)); continue; + } if (MachOObjectFile *O = dyn_cast<MachOObjectFile>(&*ChildOrErr.get())) { if (!checkMachOAndArchFlags(O, Filename)) return; ProcessMachO(Filename, O, O->getFileName()); } } + if (Err) + report_error(Filename, std::move(Err)); return; } if (UniversalHeaders) { @@ -1521,7 +1604,7 @@ void llvm::ParseInputMachO(StringRef Filename) { I != E; ++I) { if (ArchFlags[i] == I->getArchTypeName()) { ArchFound = true; - ErrorOr<std::unique_ptr<ObjectFile>> ObjOrErr = + Expected<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); std::string ArchitectureName = ""; if (ArchFlags.size() > 1) @@ -1530,7 +1613,12 @@ void llvm::ParseInputMachO(StringRef Filename) { ObjectFile &O = *ObjOrErr.get(); if (MachOObjectFile *MachOOF = dyn_cast<MachOObjectFile>(&O)) ProcessMachO(Filename, MachOOF, "", ArchitectureName); - } else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = + } else if (auto E = isNotObjectErrorInvalidFileType( + ObjOrErr.takeError())) { + report_error(Filename, StringRef(), std::move(E), + ArchitectureName); + continue; + } else if (Expected<std::unique_ptr<Archive>> AOrErr = I->getAsArchive()) { std::unique_ptr<Archive> &A = *AOrErr; outs() << "Archive : " << Filename; @@ -1539,19 +1627,25 @@ void llvm::ParseInputMachO(StringRef Filename) { outs() << "\n"; if (ArchiveHeaders) printArchiveHeaders(A.get(), !NonVerbose, ArchiveMemberOffsets); - for (Archive::child_iterator AI = A->child_begin(), - AE = A->child_end(); - AI != AE; ++AI) { - if (std::error_code EC = AI->getError()) - report_error(Filename, EC); - auto &C = AI->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); - if (ChildOrErr.getError()) + Error Err; + for (auto &C : A->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType(ChildOrErr.takeError())) + report_error(Filename, C, std::move(E), ArchitectureName); continue; + } if (MachOObjectFile *O = dyn_cast<MachOObjectFile>(&*ChildOrErr.get())) ProcessMachO(Filename, O, O->getFileName(), ArchitectureName); } + if (Err) + report_error(Filename, std::move(Err)); + } else { + consumeError(AOrErr.takeError()); + error("Mach-O universal file: " + Filename + " for " + + "architecture " + StringRef(I->getArchTypeName()) + + " is not a Mach-O file or an archive file"); } } } @@ -1571,32 +1665,42 @@ void llvm::ParseInputMachO(StringRef Filename) { I != E; ++I) { if (MachOObjectFile::getHostArch().getArchName() == I->getArchTypeName()) { - ErrorOr<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); + Expected<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); std::string ArchiveName; ArchiveName.clear(); if (ObjOrErr) { ObjectFile &O = *ObjOrErr.get(); if (MachOObjectFile *MachOOF = dyn_cast<MachOObjectFile>(&O)) ProcessMachO(Filename, MachOOF); - } else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = + } else if (auto E = isNotObjectErrorInvalidFileType( + ObjOrErr.takeError())) { + report_error(Filename, std::move(E)); + continue; + } else if (Expected<std::unique_ptr<Archive>> AOrErr = I->getAsArchive()) { std::unique_ptr<Archive> &A = *AOrErr; outs() << "Archive : " << Filename << "\n"; if (ArchiveHeaders) printArchiveHeaders(A.get(), !NonVerbose, ArchiveMemberOffsets); - for (Archive::child_iterator AI = A->child_begin(), - AE = A->child_end(); - AI != AE; ++AI) { - if (std::error_code EC = AI->getError()) - report_error(Filename, EC); - auto &C = AI->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); - if (ChildOrErr.getError()) + Error Err; + for (auto &C : A->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType(ChildOrErr.takeError())) + report_error(Filename, C, std::move(E)); continue; + } if (MachOObjectFile *O = dyn_cast<MachOObjectFile>(&*ChildOrErr.get())) ProcessMachO(Filename, O, O->getFileName()); } + if (Err) + report_error(Filename, std::move(Err)); + } else { + consumeError(AOrErr.takeError()); + error("Mach-O universal file: " + Filename + " for architecture " + + StringRef(I->getArchTypeName()) + + " is not a Mach-O file or an archive file"); } return; } @@ -1608,7 +1712,7 @@ void llvm::ParseInputMachO(StringRef Filename) { for (MachOUniversalBinary::object_iterator I = UB->begin_objects(), E = UB->end_objects(); I != E; ++I) { - ErrorOr<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); + Expected<std::unique_ptr<ObjectFile>> ObjOrErr = I->getAsObjectFile(); std::string ArchitectureName = ""; if (moreThanOneArch) ArchitectureName = I->getArchTypeName(); @@ -1616,7 +1720,12 @@ void llvm::ParseInputMachO(StringRef Filename) { ObjectFile &Obj = *ObjOrErr.get(); if (MachOObjectFile *MachOOF = dyn_cast<MachOObjectFile>(&Obj)) ProcessMachO(Filename, MachOOF, "", ArchitectureName); - } else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = I->getAsArchive()) { + } else if (auto E = isNotObjectErrorInvalidFileType( + ObjOrErr.takeError())) { + report_error(StringRef(), Filename, std::move(E), ArchitectureName); + continue; + } else if (Expected<std::unique_ptr<Archive>> AOrErr = + I->getAsArchive()) { std::unique_ptr<Archive> &A = *AOrErr; outs() << "Archive : " << Filename; if (!ArchitectureName.empty()) @@ -1624,14 +1733,14 @@ void llvm::ParseInputMachO(StringRef Filename) { outs() << "\n"; if (ArchiveHeaders) printArchiveHeaders(A.get(), !NonVerbose, ArchiveMemberOffsets); - for (Archive::child_iterator AI = A->child_begin(), AE = A->child_end(); - AI != AE; ++AI) { - if (std::error_code EC = AI->getError()) - report_error(Filename, EC); - auto &C = AI->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); - if (ChildOrErr.getError()) + Error Err; + for (auto &C : A->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType(ChildOrErr.takeError())) + report_error(Filename, C, std::move(E), ArchitectureName); continue; + } if (MachOObjectFile *O = dyn_cast<MachOObjectFile>(&*ChildOrErr.get())) { if (MachOObjectFile *MachOOF = dyn_cast<MachOObjectFile>(O)) @@ -1639,6 +1748,13 @@ void llvm::ParseInputMachO(StringRef Filename) { ArchitectureName); } } + if (Err) + report_error(Filename, std::move(Err)); + } else { + consumeError(AOrErr.takeError()); + error("Mach-O universal file: " + Filename + " for architecture " + + StringRef(I->getArchTypeName()) + + " is not a Mach-O file or an archive file"); } } return; @@ -1762,9 +1878,14 @@ static int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, } } if (reloc_found && isExtern) { - ErrorOr<StringRef> SymName = Symbol.getName(); - if (std::error_code EC = SymName.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymName = Symbol.getName(); + if (!SymName) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymName.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } const char *name = SymName->data(); op_info->AddSymbol.Present = 1; op_info->AddSymbol.Name = name; @@ -1832,9 +1953,14 @@ static int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, // is the offset from the external symbol. if (info->O->getAnyRelocationPCRel(RE)) op_info->Value -= Pc + Offset + Size; - ErrorOr<StringRef> SymName = Symbol.getName(); - if (std::error_code EC = SymName.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymName = Symbol.getName(); + if (!SymName) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymName.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } const char *name = SymName->data(); unsigned Type = info->O->getAnyRelocationType(RE); if (Type == MachO::X86_64_RELOC_SUBTRACTOR) { @@ -1849,9 +1975,14 @@ static int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, op_info->SubtractSymbol.Name = name; symbol_iterator RelocSymNext = info->O->getSymbolByIndex(SymbolNum); Symbol = *RelocSymNext; - ErrorOr<StringRef> SymNameNext = Symbol.getName(); - if (std::error_code EC = SymNameNext.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymNameNext = Symbol.getName(); + if (!SymNameNext) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymNameNext.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } name = SymNameNext->data(); } } @@ -1922,9 +2053,14 @@ static int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, } if (isExtern) { - ErrorOr<StringRef> SymName = Symbol.getName(); - if (std::error_code EC = SymName.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymName = Symbol.getName(); + if (!SymName) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymName.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } const char *name = SymName->data(); op_info->AddSymbol.Present = 1; op_info->AddSymbol.Name = name; @@ -2042,9 +2178,14 @@ static int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset, // NOTE: Scattered relocations don't exist on arm64. if (!info->O->getPlainRelocationExternal(RE)) return 0; - ErrorOr<StringRef> SymName = Reloc->getSymbol()->getName(); - if (std::error_code EC = SymName.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymName = Reloc->getSymbol()->getName(); + if (!SymName) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymName.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } const char *name = SymName->data(); op_info->AddSymbol.Present = 1; op_info->AddSymbol.Name = name; @@ -2172,9 +2313,14 @@ static const char *GuessIndirectSymbol(uint64_t ReferenceValue, if (indirect_symbol < Symtab.nsyms) { symbol_iterator Sym = info->O->getSymbolByIndex(indirect_symbol); SymbolRef Symbol = *Sym; - ErrorOr<StringRef> SymName = Symbol.getName(); - if (std::error_code EC = SymName.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymName = Symbol.getName(); + if (!SymName) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymName.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } const char *name = SymName->data(); return name; } @@ -2207,9 +2353,14 @@ static const char *GuessIndirectSymbol(uint64_t ReferenceValue, if (indirect_symbol < Symtab.nsyms) { symbol_iterator Sym = info->O->getSymbolByIndex(indirect_symbol); SymbolRef Symbol = *Sym; - ErrorOr<StringRef> SymName = Symbol.getName(); - if (std::error_code EC = SymName.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymName = Symbol.getName(); + if (!SymName) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymName.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } const char *name = SymName->data(); return name; } @@ -2436,9 +2587,14 @@ static const char *get_symbol_64(uint32_t sect_offset, SectionRef S, const char *SymbolName = nullptr; if (reloc_found && isExtern) { n_value = Symbol.getValue(); - ErrorOr<StringRef> NameOrError = Symbol.getName(); - if (std::error_code EC = NameOrError.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> NameOrError = Symbol.getName(); + if (!NameOrError) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(NameOrError.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } StringRef Name = *NameOrError; if (!Name.empty()) { SymbolName = Name.data(); @@ -5058,6 +5214,9 @@ static void print_image_info32(SectionRef S, struct DisassembleInfo *info) { struct objc_image_info32 o; const char *r; + if (S == SectionRef()) + return; + StringRef SectName; S.getName(SectName); DataRefImpl Ref = S.getRawDataRefImpl(); @@ -5530,6 +5689,373 @@ static void DumpProtocolSection(MachOObjectFile *O, const char *sect, } } +#ifdef HAVE_LIBXAR +inline void swapStruct(struct xar_header &xar) { + sys::swapByteOrder(xar.magic); + sys::swapByteOrder(xar.size); + sys::swapByteOrder(xar.version); + sys::swapByteOrder(xar.toc_length_compressed); + sys::swapByteOrder(xar.toc_length_uncompressed); + sys::swapByteOrder(xar.cksum_alg); +} + +static void PrintModeVerbose(uint32_t mode) { + switch(mode & S_IFMT){ + case S_IFDIR: + outs() << "d"; + break; + case S_IFCHR: + outs() << "c"; + break; + case S_IFBLK: + outs() << "b"; + break; + case S_IFREG: + outs() << "-"; + break; + case S_IFLNK: + outs() << "l"; + break; + case S_IFSOCK: + outs() << "s"; + break; + default: + outs() << "?"; + break; + } + + /* owner permissions */ + if(mode & S_IREAD) + outs() << "r"; + else + outs() << "-"; + if(mode & S_IWRITE) + outs() << "w"; + else + outs() << "-"; + if(mode & S_ISUID) + outs() << "s"; + else if(mode & S_IEXEC) + outs() << "x"; + else + outs() << "-"; + + /* group permissions */ + if(mode & (S_IREAD >> 3)) + outs() << "r"; + else + outs() << "-"; + if(mode & (S_IWRITE >> 3)) + outs() << "w"; + else + outs() << "-"; + if(mode & S_ISGID) + outs() << "s"; + else if(mode & (S_IEXEC >> 3)) + outs() << "x"; + else + outs() << "-"; + + /* other permissions */ + if(mode & (S_IREAD >> 6)) + outs() << "r"; + else + outs() << "-"; + if(mode & (S_IWRITE >> 6)) + outs() << "w"; + else + outs() << "-"; + if(mode & S_ISVTX) + outs() << "t"; + else if(mode & (S_IEXEC >> 6)) + outs() << "x"; + else + outs() << "-"; +} + +static void PrintXarFilesSummary(const char *XarFilename, xar_t xar) { + xar_iter_t xi; + xar_file_t xf; + xar_iter_t xp; + const char *key, *type, *mode, *user, *group, *size, *mtime, *name, *m; + char *endp; + uint32_t mode_value; + + xi = xar_iter_new(); + if (!xi) { + errs() << "Can't obtain an xar iterator for xar archive " + << XarFilename << "\n"; + return; + } + + // Go through the xar's files. + for (xf = xar_file_first(xar, xi); xf; xf = xar_file_next(xi)) { + xp = xar_iter_new(); + if(!xp){ + errs() << "Can't obtain an xar iterator for xar archive " + << XarFilename << "\n"; + return; + } + type = nullptr; + mode = nullptr; + user = nullptr; + group = nullptr; + size = nullptr; + mtime = nullptr; + name = nullptr; + for(key = xar_prop_first(xf, xp); key; key = xar_prop_next(xp)){ + const char *val = nullptr; + xar_prop_get(xf, key, &val); +#if 0 // Useful for debugging. + outs() << "key: " << key << " value: " << val << "\n"; +#endif + if(strcmp(key, "type") == 0) + type = val; + if(strcmp(key, "mode") == 0) + mode = val; + if(strcmp(key, "user") == 0) + user = val; + if(strcmp(key, "group") == 0) + group = val; + if(strcmp(key, "data/size") == 0) + size = val; + if(strcmp(key, "mtime") == 0) + mtime = val; + if(strcmp(key, "name") == 0) + name = val; + } + if(mode != nullptr){ + mode_value = strtoul(mode, &endp, 8); + if(*endp != '\0') + outs() << "(mode: \"" << mode << "\" contains non-octal chars) "; + if(strcmp(type, "file") == 0) + mode_value |= S_IFREG; + PrintModeVerbose(mode_value); + outs() << " "; + } + if(user != nullptr) + outs() << format("%10s/", user); + if(group != nullptr) + outs() << format("%-10s ", group); + if(size != nullptr) + outs() << format("%7s ", size); + if(mtime != nullptr){ + for(m = mtime; *m != 'T' && *m != '\0'; m++) + outs() << *m; + if(*m == 'T') + m++; + outs() << " "; + for( ; *m != 'Z' && *m != '\0'; m++) + outs() << *m; + outs() << " "; + } + if(name != nullptr) + outs() << name; + outs() << "\n"; + } +} + +static void DumpBitcodeSection(MachOObjectFile *O, const char *sect, + uint32_t size, bool verbose, + bool PrintXarHeader, bool PrintXarFileHeaders, + std::string XarMemberName) { + if(size < sizeof(struct xar_header)) { + outs() << "size of (__LLVM,__bundle) section too small (smaller than size " + "of struct xar_header)\n"; + return; + } + struct xar_header XarHeader; + memcpy(&XarHeader, sect, sizeof(struct xar_header)); + if (sys::IsLittleEndianHost) + swapStruct(XarHeader); + if (PrintXarHeader) { + if (!XarMemberName.empty()) + outs() << "In xar member " << XarMemberName << ": "; + else + outs() << "For (__LLVM,__bundle) section: "; + outs() << "xar header\n"; + if (XarHeader.magic == XAR_HEADER_MAGIC) + outs() << " magic XAR_HEADER_MAGIC\n"; + else + outs() << " magic " + << format_hex(XarHeader.magic, 10, true) + << " (not XAR_HEADER_MAGIC)\n"; + outs() << " size " << XarHeader.size << "\n"; + outs() << " version " << XarHeader.version << "\n"; + outs() << " toc_length_compressed " << XarHeader.toc_length_compressed + << "\n"; + outs() << "toc_length_uncompressed " << XarHeader.toc_length_uncompressed + << "\n"; + outs() << " cksum_alg "; + switch (XarHeader.cksum_alg) { + case XAR_CKSUM_NONE: + outs() << "XAR_CKSUM_NONE\n"; + break; + case XAR_CKSUM_SHA1: + outs() << "XAR_CKSUM_SHA1\n"; + break; + case XAR_CKSUM_MD5: + outs() << "XAR_CKSUM_MD5\n"; + break; +#ifdef XAR_CKSUM_SHA256 + case XAR_CKSUM_SHA256: + outs() << "XAR_CKSUM_SHA256\n"; + break; +#endif +#ifdef XAR_CKSUM_SHA512 + case XAR_CKSUM_SHA512: + outs() << "XAR_CKSUM_SHA512\n"; + break; +#endif + default: + outs() << XarHeader.cksum_alg << "\n"; + } + } + + SmallString<128> XarFilename; + int FD; + std::error_code XarEC = + sys::fs::createTemporaryFile("llvm-objdump", "xar", FD, XarFilename); + if (XarEC) { + errs() << XarEC.message() << "\n"; + return; + } + tool_output_file XarFile(XarFilename, FD); + raw_fd_ostream &XarOut = XarFile.os(); + StringRef XarContents(sect, size); + XarOut << XarContents; + XarOut.close(); + if (XarOut.has_error()) + return; + + xar_t xar = xar_open(XarFilename.c_str(), READ); + if (!xar) { + errs() << "Can't create temporary xar archive " << XarFilename << "\n"; + return; + } + + SmallString<128> TocFilename; + std::error_code TocEC = + sys::fs::createTemporaryFile("llvm-objdump", "toc", TocFilename); + if (TocEC) { + errs() << TocEC.message() << "\n"; + return; + } + xar_serialize(xar, TocFilename.c_str()); + + if (PrintXarFileHeaders) { + if (!XarMemberName.empty()) + outs() << "In xar member " << XarMemberName << ": "; + else + outs() << "For (__LLVM,__bundle) section: "; + outs() << "xar archive files:\n"; + PrintXarFilesSummary(XarFilename.c_str(), xar); + } + + ErrorOr<std::unique_ptr<MemoryBuffer>> FileOrErr = + MemoryBuffer::getFileOrSTDIN(TocFilename.c_str()); + if (std::error_code EC = FileOrErr.getError()) { + errs() << EC.message() << "\n"; + return; + } + std::unique_ptr<MemoryBuffer> &Buffer = FileOrErr.get(); + + if (!XarMemberName.empty()) + outs() << "In xar member " << XarMemberName << ": "; + else + outs() << "For (__LLVM,__bundle) section: "; + outs() << "xar table of contents:\n"; + outs() << Buffer->getBuffer() << "\n"; + + // TODO: Go through the xar's files. + xar_iter_t xi = xar_iter_new(); + if(!xi){ + errs() << "Can't obtain an xar iterator for xar archive " + << XarFilename.c_str() << "\n"; + xar_close(xar); + return; + } + for(xar_file_t xf = xar_file_first(xar, xi); xf; xf = xar_file_next(xi)){ + const char *key; + xar_iter_t xp; + const char *member_name, *member_type, *member_size_string; + size_t member_size; + + xp = xar_iter_new(); + if(!xp){ + errs() << "Can't obtain an xar iterator for xar archive " + << XarFilename.c_str() << "\n"; + xar_close(xar); + return; + } + member_name = NULL; + member_type = NULL; + member_size_string = NULL; + for(key = xar_prop_first(xf, xp); key; key = xar_prop_next(xp)){ + const char *val = nullptr; + xar_prop_get(xf, key, &val); +#if 0 // Useful for debugging. + outs() << "key: " << key << " value: " << val << "\n"; +#endif + if(strcmp(key, "name") == 0) + member_name = val; + if(strcmp(key, "type") == 0) + member_type = val; + if(strcmp(key, "data/size") == 0) + member_size_string = val; + } + /* + * If we find a file with a name, date/size and type properties + * and with the type being "file" see if that is a xar file. + */ + if (member_name != NULL && member_type != NULL && + strcmp(member_type, "file") == 0 && + member_size_string != NULL){ + // Extract the file into a buffer. + char *endptr; + member_size = strtoul(member_size_string, &endptr, 10); + if (*endptr == '\0' && member_size != 0) { + char *buffer = (char *) ::operator new (member_size); + if (xar_extract_tobuffersz(xar, xf, &buffer, &member_size) == 0) { +#if 0 // Useful for debugging. + outs() << "xar member: " << member_name << " extracted\n"; +#endif + // Set the XarMemberName we want to see printed in the header. + std::string OldXarMemberName; + // If XarMemberName is already set this is nested. So + // save the old name and create the nested name. + if (!XarMemberName.empty()) { + OldXarMemberName = XarMemberName; + XarMemberName = + (Twine("[") + XarMemberName + "]" + member_name).str(); + } else { + OldXarMemberName = ""; + XarMemberName = member_name; + } + // See if this is could be a xar file (nested). + if (member_size >= sizeof(struct xar_header)) { +#if 0 // Useful for debugging. + outs() << "could be a xar file: " << member_name << "\n"; +#endif + memcpy((char *)&XarHeader, buffer, sizeof(struct xar_header)); + if (sys::IsLittleEndianHost) + swapStruct(XarHeader); + if(XarHeader.magic == XAR_HEADER_MAGIC) + DumpBitcodeSection(O, buffer, member_size, verbose, + PrintXarHeader, PrintXarFileHeaders, + XarMemberName); + } + XarMemberName = OldXarMemberName; + } + delete buffer; + } + } + xar_iter_free(xp); + } + xar_close(xar); +} +#endif // defined(HAVE_LIBXAR) + static void printObjcMetaData(MachOObjectFile *O, bool verbose) { if (O->is64Bit()) printObjc2_64bit_MetaData(O, verbose); @@ -5948,7 +6474,7 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, return; } - // Set up thumb disassembler. + // Set up separate thumb disassembler if needed. std::unique_ptr<const MCRegisterInfo> ThumbMRI; std::unique_ptr<const MCAsmInfo> ThumbAsmInfo; std::unique_ptr<const MCSubtargetInfo> ThumbSTI; @@ -6077,13 +6603,26 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, SymbolAddressMap AddrMap; bool DisSymNameFound = false; for (const SymbolRef &Symbol : MachOOF->symbols()) { - SymbolRef::Type ST = Symbol.getType(); + Expected<SymbolRef::Type> STOrErr = Symbol.getType(); + if (!STOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(STOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } + SymbolRef::Type ST = *STOrErr; if (ST == SymbolRef::ST_Function || ST == SymbolRef::ST_Data || ST == SymbolRef::ST_Other) { uint64_t Address = Symbol.getValue(); - ErrorOr<StringRef> SymNameOrErr = Symbol.getName(); - if (std::error_code EC = SymNameOrErr.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymNameOrErr = Symbol.getName(); + if (!SymNameOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymNameOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } StringRef SymName = *SymNameOrErr; AddrMap[Address] = SymName; if (!DisSymName.empty() && DisSymName == SymName) @@ -6121,20 +6660,61 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, ThumbSymbolizerInfo.adrp_addr = 0; ThumbSymbolizerInfo.adrp_inst = 0; + unsigned int Arch = MachOOF->getArch(); + + // Skip all symbols if this is a stubs file. + if (Bytes.size() == 0) + return; + // Disassemble symbol by symbol. for (unsigned SymIdx = 0; SymIdx != Symbols.size(); SymIdx++) { - ErrorOr<StringRef> SymNameOrErr = Symbols[SymIdx].getName(); - if (std::error_code EC = SymNameOrErr.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymNameOrErr = Symbols[SymIdx].getName(); + if (!SymNameOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymNameOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } StringRef SymName = *SymNameOrErr; - SymbolRef::Type ST = Symbols[SymIdx].getType(); + Expected<SymbolRef::Type> STOrErr = Symbols[SymIdx].getType(); + if (!STOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(STOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } + SymbolRef::Type ST = *STOrErr; if (ST != SymbolRef::ST_Function && ST != SymbolRef::ST_Data) continue; // Make sure the symbol is defined in this section. bool containsSym = Sections[SectIdx].containsSymbol(Symbols[SymIdx]); - if (!containsSym) + if (!containsSym) { + if (!DisSymName.empty() && DisSymName == SymName) { + outs() << "-dis-symname: " << DisSymName << " not in the section\n"; + return; + } + continue; + } + // The __mh_execute_header is special and we need to deal with that fact + // this symbol is before the start of the (__TEXT,__text) section and at the + // address of the start of the __TEXT segment. This is because this symbol + // is an N_SECT symbol in the (__TEXT,__text) but its address is before the + // start of the section in a standard MH_EXECUTE filetype. + if (!DisSymName.empty() && DisSymName == "__mh_execute_header") { + outs() << "-dis-symname: __mh_execute_header not in any section\n"; + return; + } + // When this code is trying to disassemble a symbol at a time and in the + // case there is only the __mh_execute_header symbol left as in a stripped + // executable, we need to deal with this by ignoring this symbol so the + // whole section is disassembled and this symbol is then not displayed. + if (SymName == "__mh_execute_header" || SymName == "__mh_dylib_header" || + SymName == "__mh_bundle_header" || SymName == "__mh_object_header" || + SymName == "__mh_preload_header" || SymName == "__mh_dylinker_header") continue; // If we are only disassembling one symbol see if this is that symbol. @@ -6142,17 +6722,32 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, continue; // Start at the address of the symbol relative to the section's address. + uint64_t SectSize = Sections[SectIdx].getSize(); uint64_t Start = Symbols[SymIdx].getValue(); uint64_t SectionAddress = Sections[SectIdx].getAddress(); Start -= SectionAddress; + if (Start > SectSize) { + outs() << "section data ends, " << SymName + << " lies outside valid range\n"; + return; + } + // Stop disassembling either at the beginning of the next symbol or at // the end of the section. bool containsNextSym = false; uint64_t NextSym = 0; uint64_t NextSymIdx = SymIdx + 1; while (Symbols.size() > NextSymIdx) { - SymbolRef::Type NextSymType = Symbols[NextSymIdx].getType(); + Expected<SymbolRef::Type> STOrErr = Symbols[NextSymIdx].getType(); + if (!STOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(STOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } + SymbolRef::Type NextSymType = *STOrErr; if (NextSymType == SymbolRef::ST_Function) { containsNextSym = Sections[SectIdx].containsSymbol(Symbols[NextSymIdx]); @@ -6163,15 +6758,17 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, ++NextSymIdx; } - uint64_t SectSize = Sections[SectIdx].getSize(); - uint64_t End = containsNextSym ? NextSym : SectSize; + uint64_t End = containsNextSym ? std::min(NextSym, SectSize) : SectSize; uint64_t Size; symbolTableWorked = true; DataRefImpl Symb = Symbols[SymIdx].getRawDataRefImpl(); - bool isThumb = - (MachOOF->getSymbolFlags(Symb) & SymbolRef::SF_Thumb) && ThumbTarget; + bool IsThumb = MachOOF->getSymbolFlags(Symb) & SymbolRef::SF_Thumb; + + // We only need the dedicated Thumb target if there's a real choice + // (i.e. we're not targeting M-class) and the function is Thumb. + bool UseThumbTarget = IsThumb && ThumbTarget; outs() << SymName << ":\n"; DILineInfo lastLine; @@ -6189,7 +6786,7 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, outs() << format("%8" PRIx64 ":", PC); } } - if (!NoShowRawInsn) + if (!NoShowRawInsn || Arch == Triple::arm) outs() << "\t"; // Check the data in code table here to see if this is data not an @@ -6215,19 +6812,19 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, raw_svector_ostream Annotations(AnnotationsBytes); bool gotInst; - if (isThumb) + if (UseThumbTarget) gotInst = ThumbDisAsm->getInstruction(Inst, Size, Bytes.slice(Index), PC, DebugOut, Annotations); else gotInst = DisAsm->getInstruction(Inst, Size, Bytes.slice(Index), PC, DebugOut, Annotations); if (gotInst) { - if (!NoShowRawInsn) { + if (!NoShowRawInsn || Arch == Triple::arm) { dumpBytes(makeArrayRef(Bytes.data() + Index, Size), outs()); } formatted_raw_ostream FormattedOS(outs()); StringRef AnnotationsStr = Annotations.str(); - if (isThumb) + if (UseThumbTarget) ThumbIP->printInst(&Inst, FormattedOS, AnnotationsStr, *ThumbSTI); else IP->printInst(&Inst, FormattedOS, AnnotationsStr, *STI); @@ -6249,14 +6846,21 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, outs() << format("\t.byte 0x%02x #bad opcode\n", *(Bytes.data() + Index) & 0xff); Size = 1; // skip exactly one illegible byte and move on. - } else if (Arch == Triple::aarch64) { + } else if (Arch == Triple::aarch64 || + (Arch == Triple::arm && !IsThumb)) { uint32_t opcode = (*(Bytes.data() + Index) & 0xff) | (*(Bytes.data() + Index + 1) & 0xff) << 8 | (*(Bytes.data() + Index + 2) & 0xff) << 16 | (*(Bytes.data() + Index + 3) & 0xff) << 24; outs() << format("\t.long\t0x%08x\n", opcode); Size = 4; - } else { + } else if (Arch == Triple::arm) { + assert(IsThumb && "ARM mode should have been dealt with above"); + uint32_t opcode = (*(Bytes.data() + Index) & 0xff) | + (*(Bytes.data() + Index + 1) & 0xff) << 8; + outs() << format("\t.short\t0x%04x\n", opcode); + Size = 2; + } else{ errs() << "llvm-objdump: warning: invalid instruction encoding\n"; if (Size == 0) Size = 1; // skip illegible bytes @@ -6285,7 +6889,7 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF, outs() << format("%8" PRIx64 ":", PC); } } - if (!NoShowRawInsn) { + if (!NoShowRawInsn || Arch == Triple::arm) { outs() << "\t"; dumpBytes(makeArrayRef(Bytes.data() + Index, InstSize), outs()); } @@ -6387,9 +6991,14 @@ static void findUnwindRelocNameAddend(const MachOObjectFile *Obj, const RelocationRef &Reloc, uint64_t Addr, StringRef &Name, uint64_t &Addend) { if (Reloc.getSymbol() != Obj->symbol_end()) { - ErrorOr<StringRef> NameOrErr = Reloc.getSymbol()->getName(); - if (std::error_code EC = NameOrErr.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> NameOrErr = Reloc.getSymbol()->getName(); + if (!NameOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(NameOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } Name = *NameOrErr; Addend = Addr; return; @@ -6412,12 +7021,25 @@ static void findUnwindRelocNameAddend(const MachOObjectFile *Obj, // Go back one so that SymbolAddress <= Addr. --Sym; - section_iterator SymSection = *Sym->second.getSection(); + auto SectOrErr = Sym->second.getSection(); + if (!SectOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SectOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } + section_iterator SymSection = *SectOrErr; if (RelocSection == *SymSection) { // There's a valid symbol in the same section before this reference. - ErrorOr<StringRef> NameOrErr = Sym->second.getName(); - if (std::error_code EC = NameOrErr.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> NameOrErr = Sym->second.getName(); + if (!NameOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(NameOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } Name = *NameOrErr; Addend = Addr - Sym->first; return; @@ -6750,268 +7372,18 @@ static void printMachOUnwindInfoSection(const MachOObjectFile *Obj, } } -static unsigned getSizeForEncoding(bool is64Bit, - unsigned symbolEncoding) { - unsigned format = symbolEncoding & 0x0f; - switch (format) { - default: llvm_unreachable("Unknown Encoding"); - case dwarf::DW_EH_PE_absptr: - case dwarf::DW_EH_PE_signed: - return is64Bit ? 8 : 4; - case dwarf::DW_EH_PE_udata2: - case dwarf::DW_EH_PE_sdata2: - return 2; - case dwarf::DW_EH_PE_udata4: - case dwarf::DW_EH_PE_sdata4: - return 4; - case dwarf::DW_EH_PE_udata8: - case dwarf::DW_EH_PE_sdata8: - return 8; - } -} - -static uint64_t readPointer(const char *&Pos, bool is64Bit, unsigned Encoding) { - switch (getSizeForEncoding(is64Bit, Encoding)) { - case 2: - return readNext<uint16_t>(Pos); - break; - case 4: - return readNext<uint32_t>(Pos); - break; - case 8: - return readNext<uint64_t>(Pos); - break; - default: - llvm_unreachable("Illegal data size"); - } -} - -static void printMachOEHFrameSection(const MachOObjectFile *Obj, - std::map<uint64_t, SymbolRef> &Symbols, - const SectionRef &EHFrame) { - if (!Obj->isLittleEndian()) { - outs() << "warning: cannot handle big endian __eh_frame section\n"; - return; - } - - bool is64Bit = Obj->is64Bit(); - - outs() << "Contents of __eh_frame section:\n"; - - StringRef Contents; - EHFrame.getContents(Contents); - - /// A few fields of the CIE are used when decoding the FDE's. This struct - /// will cache those fields we need so that we don't have to decode it - /// repeatedly for each FDE that references it. - struct DecodedCIE { - Optional<uint32_t> FDEPointerEncoding; - Optional<uint32_t> LSDAPointerEncoding; - bool hasAugmentationLength; - }; - - // Map from the start offset of the CIE to the cached data for that CIE. - DenseMap<uint64_t, DecodedCIE> CachedCIEs; - - for (const char *Pos = Contents.data(), *End = Contents.end(); Pos != End; ) { - - const char *EntryStartPos = Pos; - - uint64_t Length = readNext<uint32_t>(Pos); - if (Length == 0xffffffff) - Length = readNext<uint64_t>(Pos); - - // Save the Pos so that we can check the length we encoded against what we - // end up decoding. - const char *PosAfterLength = Pos; - const char *EntryEndPos = PosAfterLength + Length; - - assert(EntryEndPos <= End && - "__eh_frame entry length exceeds section size"); - - uint32_t ID = readNext<uint32_t>(Pos); - if (ID == 0) { - // This is a CIE. - - uint32_t Version = readNext<uint8_t>(Pos); - - // Parse a null terminated augmentation string - SmallString<8> AugmentationString; - for (uint8_t Char = readNext<uint8_t>(Pos); Char; - Char = readNext<uint8_t>(Pos)) - AugmentationString.push_back(Char); - - // Optionally parse the EH data if the augmentation string says it's there. - Optional<uint64_t> EHData; - if (StringRef(AugmentationString).count("eh")) - EHData = is64Bit ? readNext<uint64_t>(Pos) : readNext<uint32_t>(Pos); - - unsigned ULEBByteCount; - uint64_t CodeAlignmentFactor = decodeULEB128((const uint8_t *)Pos, - &ULEBByteCount); - Pos += ULEBByteCount; - - int64_t DataAlignmentFactor = decodeSLEB128((const uint8_t *)Pos, - &ULEBByteCount); - Pos += ULEBByteCount; - - uint32_t ReturnAddressRegister = readNext<uint8_t>(Pos); - - Optional<uint64_t> AugmentationLength; - Optional<uint32_t> LSDAPointerEncoding; - Optional<uint32_t> PersonalityEncoding; - Optional<uint64_t> Personality; - Optional<uint32_t> FDEPointerEncoding; - if (!AugmentationString.empty() && AugmentationString.front() == 'z') { - AugmentationLength = decodeULEB128((const uint8_t *)Pos, - &ULEBByteCount); - Pos += ULEBByteCount; - - // Walk the augmentation string to get all the augmentation data. - for (unsigned i = 1, e = AugmentationString.size(); i != e; ++i) { - char Char = AugmentationString[i]; - switch (Char) { - case 'e': - assert((i + 1) != e && AugmentationString[i + 1] == 'h' && - "Expected 'eh' in augmentation string"); - break; - case 'L': - assert(!LSDAPointerEncoding && "Duplicate LSDA encoding"); - LSDAPointerEncoding = readNext<uint8_t>(Pos); - break; - case 'P': { - assert(!Personality && "Duplicate personality"); - PersonalityEncoding = readNext<uint8_t>(Pos); - Personality = readPointer(Pos, is64Bit, *PersonalityEncoding); - break; - } - case 'R': - assert(!FDEPointerEncoding && "Duplicate FDE encoding"); - FDEPointerEncoding = readNext<uint8_t>(Pos); - break; - case 'z': - llvm_unreachable("'z' must be first in the augmentation string"); - } - } - } - - outs() << "CIE:\n"; - outs() << " Length: " << Length << "\n"; - outs() << " CIE ID: " << ID << "\n"; - outs() << " Version: " << Version << "\n"; - outs() << " Augmentation String: " << AugmentationString << "\n"; - if (EHData) - outs() << " EHData: " << *EHData << "\n"; - outs() << " Code Alignment Factor: " << CodeAlignmentFactor << "\n"; - outs() << " Data Alignment Factor: " << DataAlignmentFactor << "\n"; - outs() << " Return Address Register: " << ReturnAddressRegister << "\n"; - if (AugmentationLength) { - outs() << " Augmentation Data Length: " << *AugmentationLength << "\n"; - if (LSDAPointerEncoding) { - outs() << " FDE LSDA Pointer Encoding: " - << *LSDAPointerEncoding << "\n"; - } - if (Personality) { - outs() << " Personality Encoding: " << *PersonalityEncoding << "\n"; - outs() << " Personality: " << *Personality << "\n"; - } - if (FDEPointerEncoding) { - outs() << " FDE Address Pointer Encoding: " - << *FDEPointerEncoding << "\n"; - } - } - // FIXME: Handle instructions. - // For now just emit some bytes - outs() << " Instructions:\n "; - dumpBytes(makeArrayRef((const uint8_t*)Pos, (const uint8_t*)EntryEndPos), - outs()); - outs() << "\n"; - Pos = EntryEndPos; - - // Cache this entry. - uint64_t Offset = EntryStartPos - Contents.data(); - CachedCIEs[Offset] = { FDEPointerEncoding, LSDAPointerEncoding, - AugmentationLength.hasValue() }; - continue; - } - - // This is an FDE. - // The CIE pointer for an FDE is the same location as the ID which we - // already read. - uint32_t CIEPointer = ID; - - const char *CIEStart = PosAfterLength - CIEPointer; - assert(CIEStart >= Contents.data() && - "FDE points to CIE before the __eh_frame start"); - - uint64_t CIEOffset = CIEStart - Contents.data(); - auto CIEIt = CachedCIEs.find(CIEOffset); - if (CIEIt == CachedCIEs.end()) - llvm_unreachable("Couldn't find CIE at offset in to __eh_frame section"); - - const DecodedCIE &CIE = CIEIt->getSecond(); - assert(CIE.FDEPointerEncoding && - "FDE references CIE which did not set pointer encoding"); - - uint64_t PCPointerSize = getSizeForEncoding(is64Bit, - *CIE.FDEPointerEncoding); - - uint64_t PCBegin = readPointer(Pos, is64Bit, *CIE.FDEPointerEncoding); - uint64_t PCRange = readPointer(Pos, is64Bit, *CIE.FDEPointerEncoding); - - Optional<uint64_t> AugmentationLength; - uint32_t LSDAPointerSize; - Optional<uint64_t> LSDAPointer; - if (CIE.hasAugmentationLength) { - unsigned ULEBByteCount; - AugmentationLength = decodeULEB128((const uint8_t *)Pos, - &ULEBByteCount); - Pos += ULEBByteCount; - - // Decode the LSDA if the CIE augmentation string said we should. - if (CIE.LSDAPointerEncoding) { - LSDAPointerSize = getSizeForEncoding(is64Bit, *CIE.LSDAPointerEncoding); - LSDAPointer = readPointer(Pos, is64Bit, *CIE.LSDAPointerEncoding); - } - } - - outs() << "FDE:\n"; - outs() << " Length: " << Length << "\n"; - outs() << " CIE Offset: " << CIEOffset << "\n"; - - if (PCPointerSize == 8) { - outs() << format(" PC Begin: %016" PRIx64, PCBegin) << "\n"; - outs() << format(" PC Range: %016" PRIx64, PCRange) << "\n"; - } else { - outs() << format(" PC Begin: %08" PRIx64, PCBegin) << "\n"; - outs() << format(" PC Range: %08" PRIx64, PCRange) << "\n"; - } - if (AugmentationLength) { - outs() << " Augmentation Data Length: " << *AugmentationLength << "\n"; - if (LSDAPointer) { - if (LSDAPointerSize == 8) - outs() << format(" LSDA Pointer: %016\n" PRIx64, *LSDAPointer); - else - outs() << format(" LSDA Pointer: %08\n" PRIx64, *LSDAPointer); - } - } - - // FIXME: Handle instructions. - // For now just emit some bytes - outs() << " Instructions:\n "; - dumpBytes(makeArrayRef((const uint8_t*)Pos, (const uint8_t*)EntryEndPos), - outs()); - outs() << "\n"; - Pos = EntryEndPos; - } -} - void llvm::printMachOUnwindInfo(const MachOObjectFile *Obj) { std::map<uint64_t, SymbolRef> Symbols; for (const SymbolRef &SymRef : Obj->symbols()) { // Discard any undefined or absolute symbols. They're not going to take part // in the convenience lookup for unwind info and just take up resources. - section_iterator Section = *SymRef.getSection(); + auto SectOrErr = SymRef.getSection(); + if (!SectOrErr) { + // TODO: Actually report errors helpfully. + consumeError(SectOrErr.takeError()); + continue; + } + section_iterator Section = *SectOrErr; if (Section == Obj->section_end()) continue; @@ -7026,8 +7398,6 @@ void llvm::printMachOUnwindInfo(const MachOObjectFile *Obj) { printMachOCompactUnwindSection(Obj, Symbols, Section); else if (SectName == "__unwind_info") printMachOUnwindInfoSection(Obj, Symbols, Section); - else if (SectName == "__eh_frame") - printMachOEHFrameSection(Obj, Symbols, Section); } } @@ -7145,6 +7515,10 @@ static void PrintMachHeader(uint32_t magic, uint32_t cputype, break; } break; + default: + outs() << format(" %7d", cputype); + outs() << format(" %10d", cpusubtype & ~MachO::CPU_SUBTYPE_MASK); + break; } if ((cpusubtype & MachO::CPU_SUBTYPE_MASK) == MachO::CPU_SUBTYPE_LIB64) { outs() << " LIB64"; @@ -8488,7 +8862,7 @@ static void PrintDylibCommand(MachO::dylib_command dl, const char *Ptr) { static void PrintLinkEditDataCommand(MachO::linkedit_data_command ld, uint32_t object_size) { if (ld.cmd == MachO::LC_CODE_SIGNATURE) - outs() << " cmd LC_FUNCTION_STARTS\n"; + outs() << " cmd LC_CODE_SIGNATURE\n"; else if (ld.cmd == MachO::LC_SEGMENT_SPLIT_INFO) outs() << " cmd LC_SEGMENT_SPLIT_INFO\n"; else if (ld.cmd == MachO::LC_FUNCTION_STARTS) diff --git a/tools/llvm-objdump/Makefile b/tools/llvm-objdump/Makefile deleted file mode 100644 index 7c165230dde49..0000000000000 --- a/tools/llvm-objdump/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-objdump/Makefile -------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-objdump -LINK_COMPONENTS := all-targets DebugInfoDWARF MC MCParser MCDisassembler Object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-objdump/llvm-objdump.cpp b/tools/llvm-objdump/llvm-objdump.cpp index d5ae5de4b5a33..ed55c918b58a5 100644 --- a/tools/llvm-objdump/llvm-objdump.cpp +++ b/tools/llvm-objdump/llvm-objdump.cpp @@ -22,20 +22,21 @@ #include "llvm/ADT/StringExtras.h" #include "llvm/ADT/Triple.h" #include "llvm/CodeGen/FaultMaps.h" +#include "llvm/DebugInfo/DWARF/DWARFContext.h" #include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCContext.h" -#include "llvm/MC/MCDisassembler.h" +#include "llvm/MC/MCDisassembler/MCDisassembler.h" +#include "llvm/MC/MCDisassembler/MCRelocationInfo.h" #include "llvm/MC/MCInst.h" #include "llvm/MC/MCInstPrinter.h" #include "llvm/MC/MCInstrAnalysis.h" #include "llvm/MC/MCInstrInfo.h" #include "llvm/MC/MCObjectFileInfo.h" #include "llvm/MC/MCRegisterInfo.h" -#include "llvm/MC/MCRelocationInfo.h" #include "llvm/MC/MCSubtargetInfo.h" #include "llvm/Object/Archive.h" -#include "llvm/Object/ELFObjectFile.h" #include "llvm/Object/COFF.h" +#include "llvm/Object/ELFObjectFile.h" #include "llvm/Object/MachO.h" #include "llvm/Object/ObjectFile.h" #include "llvm/Support/Casting.h" @@ -58,6 +59,7 @@ #include <cctype> #include <cstring> #include <system_error> +#include <utility> using namespace llvm; using namespace object; @@ -181,6 +183,11 @@ cl::opt<bool> cl::opt<bool> PrintFaultMaps("fault-map-section", cl::desc("Display contents of faultmap section")); +cl::opt<DIDumpType> llvm::DwarfDumpType( + "dwarf", cl::init(DIDT_Null), cl::desc("Dump of dwarf debug sections:"), + cl::values(clEnumValN(DIDT_Frames, "frames", ".debug_frame"), + clEnumValEnd)); + static StringRef ToolName; namespace { @@ -191,7 +198,7 @@ public: SectionFilterIterator(FilterPredicate P, llvm::object::section_iterator const &I, llvm::object::section_iterator const &E) - : Predicate(P), Iterator(I), End(E) { + : Predicate(std::move(P)), Iterator(I), End(E) { ScanPredicate(); } const llvm::object::SectionRef &operator*() const { return *Iterator; } @@ -218,7 +225,7 @@ private: class SectionFilter { public: SectionFilter(FilterPredicate P, llvm::object::ObjectFile const &O) - : Predicate(P), Object(O) {} + : Predicate(std::move(P)), Object(O) {} SectionFilterIterator begin() { return SectionFilterIterator(Predicate, Object.section_begin(), Object.section_end()); @@ -257,6 +264,12 @@ void llvm::error(std::error_code EC) { exit(1); } +LLVM_ATTRIBUTE_NORETURN void llvm::error(Twine Message) { + errs() << ToolName << ": " << Message << ".\n"; + errs().flush(); + exit(1); +} + LLVM_ATTRIBUTE_NORETURN void llvm::report_error(StringRef File, std::error_code EC) { assert(EC); @@ -264,6 +277,52 @@ LLVM_ATTRIBUTE_NORETURN void llvm::report_error(StringRef File, exit(1); } +LLVM_ATTRIBUTE_NORETURN void llvm::report_error(StringRef File, + llvm::Error E) { + assert(E); + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(std::move(E), OS, ""); + OS.flush(); + errs() << ToolName << ": '" << File << "': " << Buf; + exit(1); +} + +LLVM_ATTRIBUTE_NORETURN void llvm::report_error(StringRef ArchiveName, + StringRef FileName, + llvm::Error E, + StringRef ArchitectureName) { + assert(E); + errs() << ToolName << ": "; + if (ArchiveName != "") + errs() << ArchiveName << "(" << FileName << ")"; + else + errs() << FileName; + if (!ArchitectureName.empty()) + errs() << " (for architecture " << ArchitectureName << ")"; + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(std::move(E), OS, ""); + OS.flush(); + errs() << " " << Buf; + exit(1); +} + +LLVM_ATTRIBUTE_NORETURN void llvm::report_error(StringRef ArchiveName, + const object::Archive::Child &C, + llvm::Error E, + StringRef ArchitectureName) { + ErrorOr<StringRef> NameOrErr = C.getName(); + // TODO: if we have a error getting the name then it would be nice to print + // the index of which archive member this is and or its offset in the + // archive instead of "???" as the name. + if (NameOrErr.getError()) + llvm::report_error(ArchiveName, "???", std::move(E), ArchitectureName); + else + llvm::report_error(ArchiveName, NameOrErr.get(), std::move(E), + ArchitectureName); +} + static const Target *getTarget(const ObjectFile *Obj = nullptr) { // Figure out the target triple. llvm::Triple TheTriple("unknown-unknown-unknown"); @@ -308,12 +367,15 @@ public: ArrayRef<uint8_t> Bytes, uint64_t Address, raw_ostream &OS, StringRef Annot, MCSubtargetInfo const &STI) { - outs() << format("%8" PRIx64 ":", Address); + OS << format("%8" PRIx64 ":", Address); if (!NoShowRawInsn) { - outs() << "\t"; - dumpBytes(Bytes, outs()); + OS << "\t"; + dumpBytes(Bytes, OS); } - IP.printInst(MI, outs(), "", STI); + if (MI) + IP.printInst(MI, OS, "", STI); + else + OS << " <unknown>"; } }; PrettyPrinter PrettyPrinterInst; @@ -334,6 +396,11 @@ public: ArrayRef<uint8_t> Bytes, uint64_t Address, raw_ostream &OS, StringRef Annot, MCSubtargetInfo const &STI) override { + if (!MI) { + printLead(Bytes, Address, OS); + OS << " <unknown>"; + return; + } std::string Buffer; { raw_string_ostream TempStream(Buffer); @@ -370,12 +437,48 @@ public: } }; HexagonPrettyPrinter HexagonPrettyPrinterInst; + +class AMDGCNPrettyPrinter : public PrettyPrinter { +public: + void printInst(MCInstPrinter &IP, + const MCInst *MI, + ArrayRef<uint8_t> Bytes, + uint64_t Address, + raw_ostream &OS, + StringRef Annot, + MCSubtargetInfo const &STI) override { + if (!MI) { + OS << " <unknown>"; + return; + } + + SmallString<40> InstStr; + raw_svector_ostream IS(InstStr); + + IP.printInst(MI, IS, "", STI); + + OS << left_justify(IS.str(), 60) << format("// %012" PRIX64 ": ", Address); + typedef support::ulittle32_t U32; + for (auto D : makeArrayRef(reinterpret_cast<const U32*>(Bytes.data()), + Bytes.size() / sizeof(U32))) + // D should be explicitly casted to uint32_t here as it is passed + // by format to snprintf as vararg. + OS << format("%08" PRIX32 " ", static_cast<uint32_t>(D)); + + if (!Annot.empty()) + OS << "// " << Annot; + } +}; +AMDGCNPrettyPrinter AMDGCNPrettyPrinterInst; + PrettyPrinter &selectPrettyPrinter(Triple const &Triple) { switch(Triple.getArch()) { default: return PrettyPrinterInst; case Triple::hexagon: return HexagonPrettyPrinterInst; + case Triple::amdgcn: + return AMDGCNPrettyPrinterInst; } } } @@ -429,18 +532,18 @@ static std::error_code getRelocationValueString(const ELFObjectFile<ELFT> *Obj, const Elf_Sym *symb = Obj->getSymbol(SI->getRawDataRefImpl()); StringRef Target; if (symb->getType() == ELF::STT_SECTION) { - ErrorOr<section_iterator> SymSI = SI->getSection(); - if (std::error_code EC = SymSI.getError()) - return EC; + Expected<section_iterator> SymSI = SI->getSection(); + if (!SymSI) + return errorToErrorCode(SymSI.takeError()); const Elf_Shdr *SymSec = Obj->getSection((*SymSI)->getRawDataRefImpl()); ErrorOr<StringRef> SecName = EF.getSectionName(SymSec); if (std::error_code EC = SecName.getError()) return EC; Target = *SecName; } else { - ErrorOr<StringRef> SymName = symb->getName(StrTab); + Expected<StringRef> SymName = symb->getName(StrTab); if (!SymName) - return SymName.getError(); + return errorToErrorCode(SymName.takeError()); Target = *SymName; } switch (EF.getHeader()->e_machine) { @@ -470,6 +573,7 @@ static std::error_code getRelocationValueString(const ELFObjectFile<ELFT> *Obj, res = "Unknown"; } break; + case ELF::EM_LANAI: case ELF::EM_AARCH64: { std::string fmtbuf; raw_string_ostream fmt(fmtbuf); @@ -485,6 +589,7 @@ static std::error_code getRelocationValueString(const ELFObjectFile<ELFT> *Obj, case ELF::EM_ARM: case ELF::EM_HEXAGON: case ELF::EM_MIPS: + case ELF::EM_BPF: res = Target; break; case ELF::EM_WEBASSEMBLY: @@ -529,9 +634,9 @@ static std::error_code getRelocationValueString(const COFFObjectFile *Obj, const RelocationRef &Rel, SmallVectorImpl<char> &Result) { symbol_iterator SymI = Rel.getSymbol(); - ErrorOr<StringRef> SymNameOrErr = SymI->getName(); - if (std::error_code EC = SymNameOrErr.getError()) - return EC; + Expected<StringRef> SymNameOrErr = SymI->getName(); + if (!SymNameOrErr) + return errorToErrorCode(SymNameOrErr.takeError()); StringRef SymName = *SymNameOrErr; Result.append(SymName.begin(), SymName.end()); return std::error_code(); @@ -551,14 +656,24 @@ static void printRelocationTargetName(const MachOObjectFile *O, for (const SymbolRef &Symbol : O->symbols()) { std::error_code ec; - ErrorOr<uint64_t> Addr = Symbol.getAddress(); - if ((ec = Addr.getError())) - report_fatal_error(ec.message()); + Expected<uint64_t> Addr = Symbol.getAddress(); + if (!Addr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(Addr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } if (*Addr != Val) continue; - ErrorOr<StringRef> Name = Symbol.getName(); - if (std::error_code EC = Name.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> Name = Symbol.getName(); + if (!Name) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(Name.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } fmt << *Name; return; } @@ -589,8 +704,8 @@ static void printRelocationTargetName(const MachOObjectFile *O, if (isExtern) { symbol_iterator SI = O->symbol_begin(); advance(SI, Val); - ErrorOr<StringRef> SOrErr = SI->getName(); - error(SOrErr.getError()); + Expected<StringRef> SOrErr = SI->getName(); + error(errorToErrorCode(SOrErr.takeError())); S = *SOrErr; } else { section_iterator SI = O->section_begin(); @@ -828,12 +943,10 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { const Target *TheTarget = getTarget(Obj); // Package up features to be passed to target/subtarget - std::string FeaturesStr; + SubtargetFeatures Features = Obj->getFeatures(); if (MAttrs.size()) { - SubtargetFeatures Features; for (unsigned i = 0; i != MAttrs.size(); ++i) Features.AddFeature(MAttrs[i]); - FeaturesStr = Features.getString(); } std::unique_ptr<const MCRegisterInfo> MRI( @@ -847,7 +960,7 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { if (!AsmInfo) report_fatal_error("error: no assembly info for target " + TripleName); std::unique_ptr<const MCSubtargetInfo> STI( - TheTarget->createMCSubtargetInfo(TripleName, MCPU, FeaturesStr)); + TheTarget->createMCSubtargetInfo(TripleName, MCPU, Features.getString())); if (!STI) report_fatal_error("error: no subtarget info for target " + TripleName); std::unique_ptr<const MCInstrInfo> MII(TheTarget->createMCInstrInfo()); @@ -891,17 +1004,17 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { typedef std::vector<std::pair<uint64_t, StringRef>> SectionSymbolsTy; std::map<SectionRef, SectionSymbolsTy> AllSymbols; for (const SymbolRef &Symbol : Obj->symbols()) { - ErrorOr<uint64_t> AddressOrErr = Symbol.getAddress(); - error(AddressOrErr.getError()); + Expected<uint64_t> AddressOrErr = Symbol.getAddress(); + error(errorToErrorCode(AddressOrErr.takeError())); uint64_t Address = *AddressOrErr; - ErrorOr<StringRef> Name = Symbol.getName(); - error(Name.getError()); + Expected<StringRef> Name = Symbol.getName(); + error(errorToErrorCode(Name.takeError())); if (Name->empty()) continue; - ErrorOr<section_iterator> SectionOrErr = Symbol.getSection(); - error(SectionOrErr.getError()); + Expected<section_iterator> SectionOrErr = Symbol.getSection(); + error(errorToErrorCode(SectionOrErr.takeError())); section_iterator SecI = *SectionOrErr; if (SecI == Obj->section_end()) continue; @@ -1031,6 +1144,18 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { if (Start >= End) continue; + if (Obj->isELF() && Obj->getArch() == Triple::amdgcn) { + // make size 4 bytes folded + End = Start + ((End - Start) & ~0x3ull); + Start += 256; // add sizeof(amd_kernel_code_t) + // cut trailing zeroes - up to 256 bytes (align) + const uint64_t EndAlign = 256; + const auto Limit = End - (std::min)(EndAlign, End - Start); + while (End > Limit && + *reinterpret_cast<const support::ulittle32_t*>(&Bytes[End - 4]) == 0) + End -= 4; + } + outs() << '\n' << Symbols[si].second << ":\n"; #ifndef NDEBUG @@ -1081,72 +1206,69 @@ static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) { if (Index >= End) break; - if (DisAsm->getInstruction(Inst, Size, Bytes.slice(Index), - SectionAddr + Index, DebugOut, - CommentStream)) { - PIP.printInst(*IP, &Inst, - Bytes.slice(Index, Size), - SectionAddr + Index, outs(), "", *STI); - outs() << CommentStream.str(); - Comments.clear(); - - // Try to resolve the target of a call, tail call, etc. to a specific - // symbol. - if (MIA && (MIA->isCall(Inst) || MIA->isUnconditionalBranch(Inst) || - MIA->isConditionalBranch(Inst))) { - uint64_t Target; - if (MIA->evaluateBranch(Inst, SectionAddr + Index, Size, Target)) { - // In a relocatable object, the target's section must reside in - // the same section as the call instruction or it is accessed - // through a relocation. - // - // In a non-relocatable object, the target may be in any section. - // - // N.B. We don't walk the relocations in the relocatable case yet. - auto *TargetSectionSymbols = &Symbols; - if (!Obj->isRelocatableObject()) { - auto SectionAddress = std::upper_bound( - SectionAddresses.begin(), SectionAddresses.end(), Target, - [](uint64_t LHS, - const std::pair<uint64_t, SectionRef> &RHS) { - return LHS < RHS.first; - }); - if (SectionAddress != SectionAddresses.begin()) { - --SectionAddress; - TargetSectionSymbols = &AllSymbols[SectionAddress->second]; - } else { - TargetSectionSymbols = nullptr; - } + bool Disassembled = DisAsm->getInstruction(Inst, Size, Bytes.slice(Index), + SectionAddr + Index, DebugOut, + CommentStream); + if (Size == 0) + Size = 1; + PIP.printInst(*IP, Disassembled ? &Inst : nullptr, + Bytes.slice(Index, Size), + SectionAddr + Index, outs(), "", *STI); + outs() << CommentStream.str(); + Comments.clear(); + + // Try to resolve the target of a call, tail call, etc. to a specific + // symbol. + if (MIA && (MIA->isCall(Inst) || MIA->isUnconditionalBranch(Inst) || + MIA->isConditionalBranch(Inst))) { + uint64_t Target; + if (MIA->evaluateBranch(Inst, SectionAddr + Index, Size, Target)) { + // In a relocatable object, the target's section must reside in + // the same section as the call instruction or it is accessed + // through a relocation. + // + // In a non-relocatable object, the target may be in any section. + // + // N.B. We don't walk the relocations in the relocatable case yet. + auto *TargetSectionSymbols = &Symbols; + if (!Obj->isRelocatableObject()) { + auto SectionAddress = std::upper_bound( + SectionAddresses.begin(), SectionAddresses.end(), Target, + [](uint64_t LHS, + const std::pair<uint64_t, SectionRef> &RHS) { + return LHS < RHS.first; + }); + if (SectionAddress != SectionAddresses.begin()) { + --SectionAddress; + TargetSectionSymbols = &AllSymbols[SectionAddress->second]; + } else { + TargetSectionSymbols = nullptr; } + } - // Find the first symbol in the section whose offset is less than - // or equal to the target. - if (TargetSectionSymbols) { - auto TargetSym = std::upper_bound( - TargetSectionSymbols->begin(), TargetSectionSymbols->end(), - Target, [](uint64_t LHS, - const std::pair<uint64_t, StringRef> &RHS) { - return LHS < RHS.first; - }); - if (TargetSym != TargetSectionSymbols->begin()) { - --TargetSym; - uint64_t TargetAddress = std::get<0>(*TargetSym); - StringRef TargetName = std::get<1>(*TargetSym); - outs() << " <" << TargetName; - uint64_t Disp = Target - TargetAddress; - if (Disp) - outs() << '+' << utohexstr(Disp); - outs() << '>'; - } + // Find the first symbol in the section whose offset is less than + // or equal to the target. + if (TargetSectionSymbols) { + auto TargetSym = std::upper_bound( + TargetSectionSymbols->begin(), TargetSectionSymbols->end(), + Target, [](uint64_t LHS, + const std::pair<uint64_t, StringRef> &RHS) { + return LHS < RHS.first; + }); + if (TargetSym != TargetSectionSymbols->begin()) { + --TargetSym; + uint64_t TargetAddress = std::get<0>(*TargetSym); + StringRef TargetName = std::get<1>(*TargetSym); + outs() << " <" << TargetName; + uint64_t Disp = Target - TargetAddress; + if (Disp) + outs() << "+0x" << utohexstr(Disp); + outs() << '>'; } } } - outs() << "\n"; - } else { - errs() << ToolName << ": warning: invalid instruction encoding\n"; - if (Size == 0) - Size = 1; // skip illegible bytes } + outs() << "\n"; // Print relocation for instruction. while (rel_cur != rel_end) { @@ -1270,7 +1392,8 @@ void llvm::PrintSectionContents(const ObjectFile *Obj) { } } -void llvm::PrintSymbolTable(const ObjectFile *o) { +void llvm::PrintSymbolTable(const ObjectFile *o, StringRef ArchiveName, + StringRef ArchitectureName) { outs() << "SYMBOL TABLE:\n"; if (const COFFObjectFile *coff = dyn_cast<const COFFObjectFile>(o)) { @@ -1278,20 +1401,26 @@ void llvm::PrintSymbolTable(const ObjectFile *o) { return; } for (const SymbolRef &Symbol : o->symbols()) { - ErrorOr<uint64_t> AddressOrError = Symbol.getAddress(); - error(AddressOrError.getError()); + Expected<uint64_t> AddressOrError = Symbol.getAddress(); + if (!AddressOrError) + report_error(ArchiveName, o->getFileName(), AddressOrError.takeError()); uint64_t Address = *AddressOrError; - SymbolRef::Type Type = Symbol.getType(); + Expected<SymbolRef::Type> TypeOrError = Symbol.getType(); + if (!TypeOrError) + report_error(ArchiveName, o->getFileName(), TypeOrError.takeError()); + SymbolRef::Type Type = *TypeOrError; uint32_t Flags = Symbol.getFlags(); - ErrorOr<section_iterator> SectionOrErr = Symbol.getSection(); - error(SectionOrErr.getError()); + Expected<section_iterator> SectionOrErr = Symbol.getSection(); + error(errorToErrorCode(SectionOrErr.takeError())); section_iterator Section = *SectionOrErr; StringRef Name; if (Type == SymbolRef::ST_Debug && Section != o->section_end()) { Section->getName(Name); } else { - ErrorOr<StringRef> NameOrErr = Symbol.getName(); - error(NameOrErr.getError()); + Expected<StringRef> NameOrErr = Symbol.getName(); + if (!NameOrErr) + report_error(ArchiveName, o->getFileName(), NameOrErr.takeError(), + ArchitectureName); Name = *NameOrErr; } @@ -1523,12 +1652,16 @@ static void printFirstPrivateFileHeader(const ObjectFile *o) { report_fatal_error("Invalid/Unsupported object file format"); } -static void DumpObject(const ObjectFile *o) { +static void DumpObject(const ObjectFile *o, const Archive *a = nullptr) { + StringRef ArchiveName = a != nullptr ? a->getFileName() : ""; // Avoid other output when using a raw option. if (!RawClangAST) { outs() << '\n'; - outs() << o->getFileName() - << ":\tfile format " << o->getFileFormatName() << "\n\n"; + if (a) + outs() << a->getFileName() << "(" << o->getFileName() << ")"; + else + outs() << o->getFileName(); + outs() << ":\tfile format " << o->getFileFormatName() << "\n\n"; } if (Disassemble) @@ -1540,7 +1673,7 @@ static void DumpObject(const ObjectFile *o) { if (SectionContents) PrintSectionContents(o); if (SymbolTable) - PrintSymbolTable(o); + PrintSymbolTable(o, ArchiveName); if (UnwindInfo) PrintUnwindInfo(o); if (PrivateHeaders) @@ -1561,23 +1694,30 @@ static void DumpObject(const ObjectFile *o) { printRawClangAST(o); if (PrintFaultMaps) printFaultMaps(o); + if (DwarfDumpType != DIDT_Null) { + std::unique_ptr<DIContext> DICtx(new DWARFContextInMemory(*o)); + // Dump the complete DWARF structure. + DICtx->dump(outs(), DwarfDumpType, true /* DumpEH */); + } } /// @brief Dump each object file in \a a; static void DumpArchive(const Archive *a) { - for (auto &ErrorOrChild : a->children()) { - if (std::error_code EC = ErrorOrChild.getError()) - report_error(a->getFileName(), EC); - const Archive::Child &C = *ErrorOrChild; - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); - if (std::error_code EC = ChildOrErr.getError()) - if (EC != object_error::invalid_file_type) - report_error(a->getFileName(), EC); + Error Err; + for (auto &C : a->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType(ChildOrErr.takeError())) + report_error(a->getFileName(), C, std::move(E)); + continue; + } if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) - DumpObject(o); + DumpObject(o, a); else report_error(a->getFileName(), object_error::invalid_file_type); } + if (Err) + report_error(a->getFileName(), std::move(Err)); } /// @brief Open file and figure out how to dump it. @@ -1592,9 +1732,9 @@ static void DumpInput(StringRef file) { } // Attempt to open the binary. - ErrorOr<OwningBinary<Binary>> BinaryOrErr = createBinary(file); - if (std::error_code EC = BinaryOrErr.getError()) - report_error(file, EC); + Expected<OwningBinary<Binary>> BinaryOrErr = createBinary(file); + if (!BinaryOrErr) + report_error(file, BinaryOrErr.takeError()); Binary &Binary = *BinaryOrErr.get().getBinary(); if (Archive *a = dyn_cast<Archive>(&Binary)) @@ -1607,7 +1747,7 @@ static void DumpInput(StringRef file) { int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. @@ -1654,7 +1794,8 @@ int main(int argc, char **argv) { && !(DylibId && MachOOpt) && !(ObjcMetaData && MachOOpt) && !(FilterSections.size() != 0 && MachOOpt) - && !PrintFaultMaps) { + && !PrintFaultMaps + && DwarfDumpType == DIDT_Null) { cl::PrintHelpMessage(); return 2; } diff --git a/tools/llvm-objdump/llvm-objdump.h b/tools/llvm-objdump/llvm-objdump.h index 60cabbc98e2f7..5b10ee87ca863 100644 --- a/tools/llvm-objdump/llvm-objdump.h +++ b/tools/llvm-objdump/llvm-objdump.h @@ -9,16 +9,20 @@ #ifndef LLVM_TOOLS_LLVM_OBJDUMP_LLVM_OBJDUMP_H #define LLVM_TOOLS_LLVM_OBJDUMP_LLVM_OBJDUMP_H -#include "llvm/ADT/StringRef.h" +#include "llvm/DebugInfo/DIContext.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/Compiler.h" #include "llvm/Support/DataTypes.h" +#include "llvm/Object/Archive.h" namespace llvm { +class StringRef; + namespace object { class COFFObjectFile; class MachOObjectFile; class ObjectFile; + class Archive; class RelocationRef; } @@ -55,6 +59,7 @@ extern cl::opt<bool> SectionContents; extern cl::opt<bool> SymbolTable; extern cl::opt<bool> UnwindInfo; extern cl::opt<bool> PrintImmHex; +extern cl::opt<DIDumpType> DwarfDumpType; // Various helper functions. void error(std::error_code ec); @@ -81,8 +86,21 @@ void printRawClangAST(const object::ObjectFile *o); void PrintRelocations(const object::ObjectFile *o); void PrintSectionHeaders(const object::ObjectFile *o); void PrintSectionContents(const object::ObjectFile *o); -void PrintSymbolTable(const object::ObjectFile *o); +void PrintSymbolTable(const object::ObjectFile *o, StringRef ArchiveName, + StringRef ArchitectureName = StringRef()); +LLVM_ATTRIBUTE_NORETURN void error(Twine Message); LLVM_ATTRIBUTE_NORETURN void report_error(StringRef File, std::error_code EC); +LLVM_ATTRIBUTE_NORETURN void report_error(StringRef File, llvm::Error E); +LLVM_ATTRIBUTE_NORETURN void report_error(StringRef FileName, + StringRef ArchiveName, + llvm::Error E, + StringRef ArchitectureName + = StringRef()); +LLVM_ATTRIBUTE_NORETURN void report_error(StringRef ArchiveName, + const object::Archive::Child &C, + llvm::Error E, + StringRef ArchitectureName + = StringRef()); } // end namespace llvm diff --git a/tools/llvm-pdbdump/BuiltinDumper.cpp b/tools/llvm-pdbdump/BuiltinDumper.cpp index 43270540f4534..2ce1a78391102 100644 --- a/tools/llvm-pdbdump/BuiltinDumper.cpp +++ b/tools/llvm-pdbdump/BuiltinDumper.cpp @@ -14,6 +14,7 @@ #include "llvm/DebugInfo/PDB/PDBSymbolTypeBuiltin.h" using namespace llvm; +using namespace llvm::pdb; BuiltinDumper::BuiltinDumper(LinePrinter &P) : PDBSymDumper(false), Printer(P) {} @@ -30,13 +31,31 @@ StringRef BuiltinDumper::getTypeName(const PDBSymbolTypeBuiltin &Symbol) { return "float"; return "double"; case PDB_BuiltinType::UInt: - if (Symbol.getLength() == 8) + switch (Symbol.getLength()) { + case 8: return "unsigned __int64"; - return "unsigned"; + case 4: + return "unsigned int"; + case 2: + return "unsigned short"; + case 1: + return "unsigned char"; + default: + return "unsigned"; + } case PDB_BuiltinType::Int: - if (Symbol.getLength() == 4) + switch (Symbol.getLength()) { + case 8: + return "__int64"; + case 4: + return "int"; + case 2: + return "short"; + case 1: + return "char"; + default: return "int"; - return "__int64"; + } case PDB_BuiltinType::Char: return "char"; case PDB_BuiltinType::WCharT: diff --git a/tools/llvm-pdbdump/BuiltinDumper.h b/tools/llvm-pdbdump/BuiltinDumper.h index ac666dbd059b7..7a2f1438669c3 100644 --- a/tools/llvm-pdbdump/BuiltinDumper.h +++ b/tools/llvm-pdbdump/BuiltinDumper.h @@ -10,9 +10,11 @@ #ifndef LLVM_TOOLS_LLVMPDBDUMP_BUILTINDUMPER_H #define LLVM_TOOLS_LLVMPDBDUMP_BUILTINDUMPER_H +#include "llvm/ADT/StringRef.h" #include "llvm/DebugInfo/PDB/PDBSymDumper.h" namespace llvm { +namespace pdb { class LinePrinter; @@ -28,5 +30,6 @@ private: LinePrinter &Printer; }; } +} #endif diff --git a/tools/llvm-pdbdump/CMakeLists.txt b/tools/llvm-pdbdump/CMakeLists.txt index 1907f917079ae..f725edfaf3dae 100644 --- a/tools/llvm-pdbdump/CMakeLists.txt +++ b/tools/llvm-pdbdump/CMakeLists.txt @@ -1,6 +1,8 @@ set(LLVM_LINK_COMPONENTS - Support + DebugInfoCodeView DebugInfoPDB + Object + Support ) add_llvm_tool(llvm-pdbdump @@ -12,7 +14,14 @@ add_llvm_tool(llvm-pdbdump ExternalSymbolDumper.cpp FunctionDumper.cpp LinePrinter.cpp + LLVMOutputStyle.cpp + PdbYaml.cpp TypeDumper.cpp TypedefDumper.cpp VariableDumper.cpp + YAMLOutputStyle.cpp ) + +if(LLVM_USE_SANITIZE_COVERAGE) + add_subdirectory(fuzzer) +endif() diff --git a/tools/llvm-pdbdump/ClassDefinitionDumper.cpp b/tools/llvm-pdbdump/ClassDefinitionDumper.cpp index 8abf3fa3912af..553bc0b267c2e 100644 --- a/tools/llvm-pdbdump/ClassDefinitionDumper.cpp +++ b/tools/llvm-pdbdump/ClassDefinitionDumper.cpp @@ -28,6 +28,7 @@ #include "llvm/Support/Format.h" using namespace llvm; +using namespace llvm::pdb; ClassDefinitionDumper::ClassDefinitionDumper(LinePrinter &P) : PDBSymDumper(true), Printer(P) {} @@ -84,7 +85,7 @@ void ClassDefinitionDumper::start(const PDBSymbolTypeUDT &Class) { auto &AccessGroup = Groups.find((int)Access)->second; if (auto Func = dyn_cast<PDBSymbolFunc>(Child.get())) { - if (Func->isCompilerGenerated() && opts::ExcludeCompilerGenerated) + if (Func->isCompilerGenerated() && opts::pretty::ExcludeCompilerGenerated) continue; if (Func->getLength() == 0 && !Func->isPureVirtual() && !Func->isIntroVirtualFunction()) diff --git a/tools/llvm-pdbdump/ClassDefinitionDumper.h b/tools/llvm-pdbdump/ClassDefinitionDumper.h index 5b48ba879cf7b..304e11dcb6c99 100644 --- a/tools/llvm-pdbdump/ClassDefinitionDumper.h +++ b/tools/llvm-pdbdump/ClassDefinitionDumper.h @@ -19,6 +19,7 @@ #include <unordered_map> namespace llvm { +namespace pdb { class LinePrinter; @@ -58,5 +59,5 @@ private: int dumpAccessGroup(PDB_MemberAccess Access, const SymbolGroup &Group); }; } - +} #endif diff --git a/tools/llvm-pdbdump/CompilandDumper.cpp b/tools/llvm-pdbdump/CompilandDumper.cpp index 68ceb620627b9..5ec37cbbb5a06 100644 --- a/tools/llvm-pdbdump/CompilandDumper.cpp +++ b/tools/llvm-pdbdump/CompilandDumper.cpp @@ -12,7 +12,9 @@ #include "llvm-pdbdump.h" #include "llvm/DebugInfo/PDB/IPDBEnumChildren.h" +#include "llvm/DebugInfo/PDB/IPDBLineNumber.h" #include "llvm/DebugInfo/PDB/IPDBSession.h" +#include "llvm/DebugInfo/PDB/IPDBSourceFile.h" #include "llvm/DebugInfo/PDB/PDBExtras.h" #include "llvm/DebugInfo/PDB/PDBSymbol.h" #include "llvm/DebugInfo/PDB/PDBSymbolCompiland.h" @@ -31,9 +33,9 @@ #include "FunctionDumper.h" #include <utility> -#include <vector> using namespace llvm; +using namespace llvm::pdb; CompilandDumper::CompilandDumper(LinePrinter &P) : PDBSymDumper(true), Printer(P) {} @@ -42,21 +44,65 @@ void CompilandDumper::dump(const PDBSymbolCompilandDetails &Symbol) {} void CompilandDumper::dump(const PDBSymbolCompilandEnv &Symbol) {} -void CompilandDumper::start(const PDBSymbolCompiland &Symbol, bool Children) { +void CompilandDumper::start(const PDBSymbolCompiland &Symbol, + CompilandDumpFlags opts) { std::string FullName = Symbol.getName(); if (Printer.IsCompilandExcluded(FullName)) return; Printer.NewLine(); WithColor(Printer, PDB_ColorItem::Path).get() << FullName; - if (!Children) - return; - auto ChildrenEnum = Symbol.findAllChildren(); - Printer.Indent(); - while (auto Child = ChildrenEnum->getNext()) - Child->dump(*this); - Printer.Unindent(); + if (opts & Flags::Lines) { + const IPDBSession &Session = Symbol.getSession(); + auto Files = Session.getSourceFilesForCompiland(Symbol); + Printer.Indent(); + while (auto File = Files->getNext()) { + Printer.NewLine(); + WithColor(Printer, PDB_ColorItem::Path).get() << File->getFileName(); + + auto Lines = Session.findLineNumbers(Symbol, *File); + Printer.Indent(); + while (auto Line = Lines->getNext()) { + Printer.NewLine(); + uint32_t LineStart = Line->getLineNumber(); + uint32_t LineEnd = Line->getLineNumberEnd(); + + Printer << "Line "; + PDB_ColorItem StatementColor = Line->isStatement() + ? PDB_ColorItem::Keyword + : PDB_ColorItem::LiteralValue; + WithColor(Printer, StatementColor).get() << LineStart; + if (LineStart != LineEnd) + WithColor(Printer, StatementColor).get() << " - " << LineEnd; + + Printer << ", Address: "; + if (Line->getLength() > 0) { + uint64_t AddrStart = Line->getVirtualAddress(); + uint64_t AddrEnd = AddrStart + Line->getLength() - 1; + WithColor(Printer, PDB_ColorItem::Address).get() + << "[" << format_hex(AddrStart, 10) << " - " + << format_hex(AddrEnd, 10) << "]"; + Printer << " (" << Line->getLength() << " bytes)"; + } else { + uint64_t AddrStart = Line->getVirtualAddress(); + WithColor(Printer, PDB_ColorItem::Address).get() + << "[" << format_hex(AddrStart, 10) << "] "; + Printer << "(0 bytes)"; + } + } + Printer.Unindent(); + } + Printer.Unindent(); + } + + if (opts & Flags::Children) { + auto ChildrenEnum = Symbol.findAllChildren(); + Printer.Indent(); + while (auto Child = ChildrenEnum->getNext()) + Child->dump(*this); + Printer.Unindent(); + } } void CompilandDumper::dump(const PDBSymbolData &Symbol) { @@ -112,9 +158,9 @@ void CompilandDumper::dump(const PDBSymbolThunk &Symbol) { Printer.NewLine(); Printer << "thunk "; - PDB_ThunkOrdinal Ordinal = Symbol.getThunkOrdinal(); + codeview::ThunkOrdinal Ordinal = Symbol.getThunkOrdinal(); uint64_t VA = Symbol.getVirtualAddress(); - if (Ordinal == PDB_ThunkOrdinal::TrampIncremental) { + if (Ordinal == codeview::ThunkOrdinal::TrampIncremental) { uint64_t Target = Symbol.getTargetVirtualAddress(); WithColor(Printer, PDB_ColorItem::Address).get() << format_hex(VA, 10); Printer << " -> "; diff --git a/tools/llvm-pdbdump/CompilandDumper.h b/tools/llvm-pdbdump/CompilandDumper.h index 0d1d27cd7a436..462aaeb2611fe 100644 --- a/tools/llvm-pdbdump/CompilandDumper.h +++ b/tools/llvm-pdbdump/CompilandDumper.h @@ -13,14 +13,18 @@ #include "llvm/DebugInfo/PDB/PDBSymDumper.h" namespace llvm { +namespace pdb { class LinePrinter; +typedef int CompilandDumpFlags; class CompilandDumper : public PDBSymDumper { public: + enum Flags { None = 0x0, Children = 0x1, Symbols = 0x2, Lines = 0x4 }; + CompilandDumper(LinePrinter &P); - void start(const PDBSymbolCompiland &Symbol, bool Children); + void start(const PDBSymbolCompiland &Symbol, CompilandDumpFlags flags); void dump(const PDBSymbolCompilandDetails &Symbol) override; void dump(const PDBSymbolCompilandEnv &Symbol) override; @@ -35,5 +39,6 @@ private: LinePrinter &Printer; }; } +} #endif diff --git a/tools/llvm-pdbdump/EnumDumper.cpp b/tools/llvm-pdbdump/EnumDumper.cpp index 3514c39f6e033..43b6018ffedf3 100644 --- a/tools/llvm-pdbdump/EnumDumper.cpp +++ b/tools/llvm-pdbdump/EnumDumper.cpp @@ -18,13 +18,14 @@ #include "llvm/DebugInfo/PDB/PDBSymbolTypeEnum.h" using namespace llvm; +using namespace llvm::pdb; EnumDumper::EnumDumper(LinePrinter &P) : PDBSymDumper(true), Printer(P) {} void EnumDumper::start(const PDBSymbolTypeEnum &Symbol) { WithColor(Printer, PDB_ColorItem::Keyword).get() << "enum "; WithColor(Printer, PDB_ColorItem::Type).get() << Symbol.getName(); - if (!opts::NoEnumDefs) { + if (!opts::pretty::NoEnumDefs) { auto BuiltinType = Symbol.getUnderlyingType(); if (BuiltinType->getBuiltinType() != PDB_BuiltinType::Int || BuiltinType->getLength() != 4) { diff --git a/tools/llvm-pdbdump/EnumDumper.h b/tools/llvm-pdbdump/EnumDumper.h index 23de0614247fe..0a34e1f89ada0 100644 --- a/tools/llvm-pdbdump/EnumDumper.h +++ b/tools/llvm-pdbdump/EnumDumper.h @@ -13,6 +13,7 @@ #include "llvm/DebugInfo/PDB/PDBSymDumper.h" namespace llvm { +namespace pdb { class LinePrinter; @@ -26,5 +27,5 @@ private: LinePrinter &Printer; }; } - +} #endif diff --git a/tools/llvm-pdbdump/ExternalSymbolDumper.cpp b/tools/llvm-pdbdump/ExternalSymbolDumper.cpp index c4e9f474880e3..508a2405772e4 100644 --- a/tools/llvm-pdbdump/ExternalSymbolDumper.cpp +++ b/tools/llvm-pdbdump/ExternalSymbolDumper.cpp @@ -15,6 +15,7 @@ #include "llvm/Support/Format.h" using namespace llvm; +using namespace llvm::pdb; ExternalSymbolDumper::ExternalSymbolDumper(LinePrinter &P) : PDBSymDumper(true), Printer(P) {} diff --git a/tools/llvm-pdbdump/ExternalSymbolDumper.h b/tools/llvm-pdbdump/ExternalSymbolDumper.h index d77b09cdebf17..b44b8a6fe98ae 100644 --- a/tools/llvm-pdbdump/ExternalSymbolDumper.h +++ b/tools/llvm-pdbdump/ExternalSymbolDumper.h @@ -13,6 +13,7 @@ #include "llvm/DebugInfo/PDB/PDBSymDumper.h" namespace llvm { +namespace pdb { class LinePrinter; @@ -28,5 +29,6 @@ private: LinePrinter &Printer; }; } +} #endif diff --git a/tools/llvm-pdbdump/FunctionDumper.cpp b/tools/llvm-pdbdump/FunctionDumper.cpp index 9584812e81a94..29ba15d521f0e 100644 --- a/tools/llvm-pdbdump/FunctionDumper.cpp +++ b/tools/llvm-pdbdump/FunctionDumper.cpp @@ -13,6 +13,7 @@ #include "llvm-pdbdump.h" #include "llvm/DebugInfo/PDB/IPDBSession.h" +#include "llvm/DebugInfo/PDB/PDBExtras.h" #include "llvm/DebugInfo/PDB/PDBSymbolData.h" #include "llvm/DebugInfo/PDB/PDBSymbolFunc.h" #include "llvm/DebugInfo/PDB/PDBSymbolFuncDebugEnd.h" @@ -27,11 +28,13 @@ #include "llvm/Support/Format.h" using namespace llvm; +using namespace llvm::codeview; +using namespace llvm::pdb; namespace { template <class T> void dumpClassParentWithScopeOperator(const T &Symbol, LinePrinter &Printer, - llvm::FunctionDumper &Dumper) { + FunctionDumper &Dumper) { uint32_t ClassParentId = Symbol.getClassParentId(); auto ClassParent = Symbol.getSession().template getConcreteSymbolById<PDBSymbolTypeUDT>( @@ -59,8 +62,8 @@ void FunctionDumper::start(const PDBSymbolTypeFunctionSig &Symbol, PDB_CallingConv CC = Symbol.getCallingConvention(); bool ShouldDumpCallingConvention = true; - if ((ClassParent && CC == PDB_CallingConv::Thiscall) || - (!ClassParent && CC == PDB_CallingConv::NearStdcall)) { + if ((ClassParent && CC == CallingConvention::ThisCall) || + (!ClassParent && CC == CallingConvention::NearStdCall)) { ShouldDumpCallingConvention = false; } @@ -152,12 +155,12 @@ void FunctionDumper::start(const PDBSymbolFunc &Symbol, PointerType Pointer) { Printer << " "; auto ClassParent = Symbol.getClassParent(); - PDB_CallingConv CC = Signature->getCallingConvention(); + CallingConvention CC = Signature->getCallingConvention(); if (Pointer != FunctionDumper::PointerType::None) Printer << "("; - if ((ClassParent && CC != PDB_CallingConv::Thiscall) || - (!ClassParent && CC != PDB_CallingConv::NearStdcall)) { + if ((ClassParent && CC != CallingConvention::ThisCall) || + (!ClassParent && CC != CallingConvention::NearStdCall)) { WithColor(Printer, PDB_ColorItem::Keyword).get() << Signature->getCallingConvention() << " "; } diff --git a/tools/llvm-pdbdump/FunctionDumper.h b/tools/llvm-pdbdump/FunctionDumper.h index 19a00145a1fd7..c71fafa18ed3b 100644 --- a/tools/llvm-pdbdump/FunctionDumper.h +++ b/tools/llvm-pdbdump/FunctionDumper.h @@ -13,7 +13,7 @@ #include "llvm/DebugInfo/PDB/PDBSymDumper.h" namespace llvm { - +namespace pdb { class LinePrinter; class FunctionDumper : public PDBSymDumper { @@ -38,5 +38,6 @@ private: LinePrinter &Printer; }; } +} #endif diff --git a/tools/llvm-pdbdump/LLVMOutputStyle.cpp b/tools/llvm-pdbdump/LLVMOutputStyle.cpp new file mode 100644 index 0000000000000..6d94295d1269b --- /dev/null +++ b/tools/llvm-pdbdump/LLVMOutputStyle.cpp @@ -0,0 +1,775 @@ +//===- LLVMOutputStyle.cpp ------------------------------------ *- C++ --*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "LLVMOutputStyle.h" + +#include "llvm-pdbdump.h" +#include "llvm/DebugInfo/CodeView/EnumTables.h" +#include "llvm/DebugInfo/CodeView/ModuleSubstreamVisitor.h" +#include "llvm/DebugInfo/CodeView/SymbolDumper.h" +#include "llvm/DebugInfo/PDB/PDBExtras.h" +#include "llvm/DebugInfo/PDB/Raw/DbiStream.h" +#include "llvm/DebugInfo/PDB/Raw/EnumTables.h" +#include "llvm/DebugInfo/PDB/Raw/ISectionContribVisitor.h" +#include "llvm/DebugInfo/PDB/Raw/IndexedStreamData.h" +#include "llvm/DebugInfo/PDB/Raw/InfoStream.h" +#include "llvm/DebugInfo/PDB/Raw/ModInfo.h" +#include "llvm/DebugInfo/PDB/Raw/ModStream.h" +#include "llvm/DebugInfo/PDB/Raw/PDBFile.h" +#include "llvm/DebugInfo/PDB/Raw/PublicsStream.h" +#include "llvm/DebugInfo/PDB/Raw/RawError.h" +#include "llvm/DebugInfo/PDB/Raw/TpiStream.h" +#include "llvm/Object/COFF.h" + +#include <unordered_map> + +using namespace llvm; +using namespace llvm::codeview; +using namespace llvm::pdb; + +static void printSectionOffset(llvm::raw_ostream &OS, + const SectionOffset &Off) { + OS << Off.Off << ", " << Off.Isect; +} + +LLVMOutputStyle::LLVMOutputStyle(PDBFile &File) + : File(File), P(outs()), TD(&P, false) {} + +Error LLVMOutputStyle::dump() { + if (auto EC = dumpFileHeaders()) + return EC; + + if (auto EC = dumpStreamSummary()) + return EC; + + if (auto EC = dumpStreamBlocks()) + return EC; + + if (auto EC = dumpStreamData()) + return EC; + + if (auto EC = dumpInfoStream()) + return EC; + + if (auto EC = dumpNamedStream()) + return EC; + + if (auto EC = dumpTpiStream(StreamTPI)) + return EC; + + if (auto EC = dumpTpiStream(StreamIPI)) + return EC; + + if (auto EC = dumpDbiStream()) + return EC; + + if (auto EC = dumpSectionContribs()) + return EC; + + if (auto EC = dumpSectionMap()) + return EC; + + if (auto EC = dumpPublicsStream()) + return EC; + + if (auto EC = dumpSectionHeaders()) + return EC; + + if (auto EC = dumpFpoStream()) + return EC; + + flush(); + + return Error::success(); +} + +Error LLVMOutputStyle::dumpFileHeaders() { + if (!opts::raw::DumpHeaders) + return Error::success(); + + DictScope D(P, "FileHeaders"); + P.printNumber("BlockSize", File.getBlockSize()); + P.printNumber("FreeBlockMap", File.getFreeBlockMapBlock()); + P.printNumber("NumBlocks", File.getBlockCount()); + P.printNumber("NumDirectoryBytes", File.getNumDirectoryBytes()); + P.printNumber("Unknown1", File.getUnknown1()); + P.printNumber("BlockMapAddr", File.getBlockMapIndex()); + P.printNumber("NumDirectoryBlocks", File.getNumDirectoryBlocks()); + + // The directory is not contiguous. Instead, the block map contains a + // contiguous list of block numbers whose contents, when concatenated in + // order, make up the directory. + P.printList("DirectoryBlocks", File.getDirectoryBlockArray()); + P.printNumber("NumStreams", File.getNumStreams()); + return Error::success(); +} + +Error LLVMOutputStyle::dumpStreamSummary() { + if (!opts::raw::DumpStreamSummary) + return Error::success(); + + // It's OK if we fail to load some of these streams, we still attempt to print + // what we can. + auto Dbi = File.getPDBDbiStream(); + auto Tpi = File.getPDBTpiStream(); + auto Ipi = File.getPDBIpiStream(); + auto Info = File.getPDBInfoStream(); + + ListScope L(P, "Streams"); + uint32_t StreamCount = File.getNumStreams(); + std::unordered_map<uint16_t, const ModuleInfoEx *> ModStreams; + std::unordered_map<uint16_t, std::string> NamedStreams; + + if (Dbi) { + for (auto &ModI : Dbi->modules()) { + uint16_t SN = ModI.Info.getModuleStreamIndex(); + ModStreams[SN] = &ModI; + } + } + if (Info) { + for (auto &NSE : Info->named_streams()) { + NamedStreams[NSE.second] = NSE.first(); + } + } + + for (uint16_t StreamIdx = 0; StreamIdx < StreamCount; ++StreamIdx) { + std::string Label("Stream "); + Label += to_string(StreamIdx); + std::string Value; + if (StreamIdx == OldMSFDirectory) + Value = "Old MSF Directory"; + else if (StreamIdx == StreamPDB) + Value = "PDB Stream"; + else if (StreamIdx == StreamDBI) + Value = "DBI Stream"; + else if (StreamIdx == StreamTPI) + Value = "TPI Stream"; + else if (StreamIdx == StreamIPI) + Value = "IPI Stream"; + else if (Dbi && StreamIdx == Dbi->getGlobalSymbolStreamIndex()) + Value = "Global Symbol Hash"; + else if (Dbi && StreamIdx == Dbi->getPublicSymbolStreamIndex()) + Value = "Public Symbol Hash"; + else if (Dbi && StreamIdx == Dbi->getSymRecordStreamIndex()) + Value = "Public Symbol Records"; + else if (Tpi && StreamIdx == Tpi->getTypeHashStreamIndex()) + Value = "TPI Hash"; + else if (Tpi && StreamIdx == Tpi->getTypeHashStreamAuxIndex()) + Value = "TPI Aux Hash"; + else if (Ipi && StreamIdx == Ipi->getTypeHashStreamIndex()) + Value = "IPI Hash"; + else if (Ipi && StreamIdx == Ipi->getTypeHashStreamAuxIndex()) + Value = "IPI Aux Hash"; + else if (Dbi && + StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::Exception)) + Value = "Exception Data"; + else if (Dbi && StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::Fixup)) + Value = "Fixup Data"; + else if (Dbi && StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::FPO)) + Value = "FPO Data"; + else if (Dbi && + StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::NewFPO)) + Value = "New FPO Data"; + else if (Dbi && + StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::OmapFromSrc)) + Value = "Omap From Source Data"; + else if (Dbi && + StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::OmapToSrc)) + Value = "Omap To Source Data"; + else if (Dbi && StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::Pdata)) + Value = "Pdata"; + else if (Dbi && + StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::SectionHdr)) + Value = "Section Header Data"; + else if (Dbi && + StreamIdx == + Dbi->getDebugStreamIndex(DbgHeaderType::SectionHdrOrig)) + Value = "Section Header Original Data"; + else if (Dbi && + StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::TokenRidMap)) + Value = "Token Rid Data"; + else if (Dbi && StreamIdx == Dbi->getDebugStreamIndex(DbgHeaderType::Xdata)) + Value = "Xdata"; + else { + auto ModIter = ModStreams.find(StreamIdx); + auto NSIter = NamedStreams.find(StreamIdx); + if (ModIter != ModStreams.end()) { + Value = "Module \""; + Value += ModIter->second->Info.getModuleName().str(); + Value += "\""; + } else if (NSIter != NamedStreams.end()) { + Value = "Named Stream \""; + Value += NSIter->second; + Value += "\""; + } else { + Value = "???"; + } + } + Value = "[" + Value + "]"; + Value = + Value + " (" + to_string(File.getStreamByteSize(StreamIdx)) + " bytes)"; + + P.printString(Label, Value); + } + + // Consume errors from missing streams. + if (!Dbi) + consumeError(Dbi.takeError()); + if (!Tpi) + consumeError(Tpi.takeError()); + if (!Ipi) + consumeError(Ipi.takeError()); + if (!Info) + consumeError(Info.takeError()); + + P.flush(); + return Error::success(); +} + +Error LLVMOutputStyle::dumpStreamBlocks() { + if (!opts::raw::DumpStreamBlocks) + return Error::success(); + + ListScope L(P, "StreamBlocks"); + uint32_t StreamCount = File.getNumStreams(); + for (uint32_t StreamIdx = 0; StreamIdx < StreamCount; ++StreamIdx) { + std::string Name("Stream "); + Name += to_string(StreamIdx); + auto StreamBlocks = File.getStreamBlockList(StreamIdx); + P.printList(Name, StreamBlocks); + } + return Error::success(); +} + +Error LLVMOutputStyle::dumpStreamData() { + uint32_t StreamCount = File.getNumStreams(); + StringRef DumpStreamStr = opts::raw::DumpStreamDataIdx; + uint32_t DumpStreamNum; + if (DumpStreamStr.getAsInteger(/*Radix=*/0U, DumpStreamNum)) + return Error::success(); + + if (DumpStreamNum >= StreamCount) + return make_error<RawError>(raw_error_code::no_stream); + + auto S = MappedBlockStream::createIndexedStream(DumpStreamNum, File); + if (!S) + return S.takeError(); + codeview::StreamReader R(**S); + while (R.bytesRemaining() > 0) { + ArrayRef<uint8_t> Data; + uint32_t BytesToReadInBlock = std::min( + R.bytesRemaining(), static_cast<uint32_t>(File.getBlockSize())); + if (auto EC = R.readBytes(Data, BytesToReadInBlock)) + return EC; + P.printBinaryBlock( + "Data", + StringRef(reinterpret_cast<const char *>(Data.begin()), Data.size())); + } + return Error::success(); +} + +Error LLVMOutputStyle::dumpInfoStream() { + if (!opts::raw::DumpHeaders) + return Error::success(); + auto IS = File.getPDBInfoStream(); + if (!IS) + return IS.takeError(); + + DictScope D(P, "PDB Stream"); + P.printNumber("Version", IS->getVersion()); + P.printHex("Signature", IS->getSignature()); + P.printNumber("Age", IS->getAge()); + P.printObject("Guid", IS->getGuid()); + return Error::success(); +} + +Error LLVMOutputStyle::dumpNamedStream() { + if (opts::raw::DumpStreamDataName.empty()) + return Error::success(); + + auto IS = File.getPDBInfoStream(); + if (!IS) + return IS.takeError(); + + uint32_t NameStreamIndex = + IS->getNamedStreamIndex(opts::raw::DumpStreamDataName); + if (NameStreamIndex == 0 || NameStreamIndex >= File.getNumStreams()) + return make_error<RawError>(raw_error_code::no_stream); + + if (NameStreamIndex != 0) { + std::string Name("Stream '"); + Name += opts::raw::DumpStreamDataName; + Name += "'"; + DictScope D(P, Name); + P.printNumber("Index", NameStreamIndex); + + auto NameStream = + MappedBlockStream::createIndexedStream(NameStreamIndex, File); + if (!NameStream) + return NameStream.takeError(); + codeview::StreamReader Reader(**NameStream); + + NameHashTable NameTable; + if (auto EC = NameTable.load(Reader)) + return EC; + + P.printHex("Signature", NameTable.getSignature()); + P.printNumber("Version", NameTable.getHashVersion()); + P.printNumber("Name Count", NameTable.getNameCount()); + ListScope L(P, "Names"); + for (uint32_t ID : NameTable.name_ids()) { + StringRef Str = NameTable.getStringForID(ID); + if (!Str.empty()) + P.printString(to_string(ID), Str); + } + } + return Error::success(); +} + +static void printTypeIndexOffset(raw_ostream &OS, + const TypeIndexOffset &TIOff) { + OS << "{" << TIOff.Type.getIndex() << ", " << TIOff.Offset << "}"; +} + +static void dumpTpiHash(ScopedPrinter &P, TpiStream &Tpi) { + if (!opts::raw::DumpTpiHash) + return; + DictScope DD(P, "Hash"); + P.printNumber("Number of Hash Buckets", Tpi.NumHashBuckets()); + P.printNumber("Hash Key Size", Tpi.getHashKeySize()); + P.printList("Values", Tpi.getHashValues()); + P.printList("Type Index Offsets", Tpi.getTypeIndexOffsets(), + printTypeIndexOffset); + P.printList("Hash Adjustments", Tpi.getHashAdjustments(), + printTypeIndexOffset); +} + +Error LLVMOutputStyle::dumpTpiStream(uint32_t StreamIdx) { + assert(StreamIdx == StreamTPI || StreamIdx == StreamIPI); + + bool DumpRecordBytes = false; + bool DumpRecords = false; + StringRef Label; + StringRef VerLabel; + if (StreamIdx == StreamTPI) { + DumpRecordBytes = opts::raw::DumpTpiRecordBytes; + DumpRecords = opts::raw::DumpTpiRecords; + Label = "Type Info Stream (TPI)"; + VerLabel = "TPI Version"; + } else if (StreamIdx == StreamIPI) { + DumpRecordBytes = opts::raw::DumpIpiRecordBytes; + DumpRecords = opts::raw::DumpIpiRecords; + Label = "Type Info Stream (IPI)"; + VerLabel = "IPI Version"; + } + if (!DumpRecordBytes && !DumpRecords && !opts::raw::DumpModuleSyms) + return Error::success(); + + auto Tpi = (StreamIdx == StreamTPI) ? File.getPDBTpiStream() + : File.getPDBIpiStream(); + if (!Tpi) + return Tpi.takeError(); + + if (DumpRecords || DumpRecordBytes) { + DictScope D(P, Label); + + P.printNumber(VerLabel, Tpi->getTpiVersion()); + P.printNumber("Record count", Tpi->NumTypeRecords()); + + ListScope L(P, "Records"); + + bool HadError = false; + for (auto &Type : Tpi->types(&HadError)) { + DictScope DD(P, ""); + + if (DumpRecords) { + if (auto EC = TD.dump(Type)) + return EC; + } + + if (DumpRecordBytes) + P.printBinaryBlock("Bytes", Type.Data); + } + dumpTpiHash(P, *Tpi); + if (HadError) + return make_error<RawError>(raw_error_code::corrupt_file, + "TPI stream contained corrupt record"); + } else if (opts::raw::DumpModuleSyms) { + // Even if the user doesn't want to dump type records, we still need to + // iterate them in order to build the list of types so that we can print + // them when dumping module symbols. So when they want to dump symbols + // but not types, use a null output stream. + ScopedPrinter *OldP = TD.getPrinter(); + TD.setPrinter(nullptr); + + bool HadError = false; + for (auto &Type : Tpi->types(&HadError)) { + if (auto EC = TD.dump(Type)) + return EC; + } + + TD.setPrinter(OldP); + dumpTpiHash(P, *Tpi); + if (HadError) + return make_error<RawError>(raw_error_code::corrupt_file, + "TPI stream contained corrupt record"); + } + P.flush(); + return Error::success(); +} + +Error LLVMOutputStyle::dumpDbiStream() { + bool DumpModules = opts::raw::DumpModules || opts::raw::DumpModuleSyms || + opts::raw::DumpModuleFiles || opts::raw::DumpLineInfo; + if (!opts::raw::DumpHeaders && !DumpModules) + return Error::success(); + + auto DS = File.getPDBDbiStream(); + if (!DS) + return DS.takeError(); + + DictScope D(P, "DBI Stream"); + P.printNumber("Dbi Version", DS->getDbiVersion()); + P.printNumber("Age", DS->getAge()); + P.printBoolean("Incremental Linking", DS->isIncrementallyLinked()); + P.printBoolean("Has CTypes", DS->hasCTypes()); + P.printBoolean("Is Stripped", DS->isStripped()); + P.printObject("Machine Type", DS->getMachineType()); + P.printNumber("Symbol Record Stream Index", DS->getSymRecordStreamIndex()); + P.printNumber("Public Symbol Stream Index", DS->getPublicSymbolStreamIndex()); + P.printNumber("Global Symbol Stream Index", DS->getGlobalSymbolStreamIndex()); + + uint16_t Major = DS->getBuildMajorVersion(); + uint16_t Minor = DS->getBuildMinorVersion(); + P.printVersion("Toolchain Version", Major, Minor); + + std::string DllName; + raw_string_ostream DllStream(DllName); + DllStream << "mspdb" << Major << Minor << ".dll version"; + DllStream.flush(); + P.printVersion(DllName, Major, Minor, DS->getPdbDllVersion()); + + if (DumpModules) { + ListScope L(P, "Modules"); + for (auto &Modi : DS->modules()) { + DictScope DD(P); + P.printString("Name", Modi.Info.getModuleName().str()); + P.printNumber("Debug Stream Index", Modi.Info.getModuleStreamIndex()); + P.printString("Object File Name", Modi.Info.getObjFileName().str()); + P.printNumber("Num Files", Modi.Info.getNumberOfFiles()); + P.printNumber("Source File Name Idx", Modi.Info.getSourceFileNameIndex()); + P.printNumber("Pdb File Name Idx", Modi.Info.getPdbFilePathNameIndex()); + P.printNumber("Line Info Byte Size", Modi.Info.getLineInfoByteSize()); + P.printNumber("C13 Line Info Byte Size", + Modi.Info.getC13LineInfoByteSize()); + P.printNumber("Symbol Byte Size", Modi.Info.getSymbolDebugInfoByteSize()); + P.printNumber("Type Server Index", Modi.Info.getTypeServerIndex()); + P.printBoolean("Has EC Info", Modi.Info.hasECInfo()); + if (opts::raw::DumpModuleFiles) { + std::string FileListName = + to_string(Modi.SourceFiles.size()) + " Contributing Source Files"; + ListScope LL(P, FileListName); + for (auto File : Modi.SourceFiles) + P.printString(File.str()); + } + bool HasModuleDI = + (Modi.Info.getModuleStreamIndex() < File.getNumStreams()); + bool ShouldDumpSymbols = + (opts::raw::DumpModuleSyms || opts::raw::DumpSymRecordBytes); + if (HasModuleDI && (ShouldDumpSymbols || opts::raw::DumpLineInfo)) { + auto ModStreamData = MappedBlockStream::createIndexedStream( + Modi.Info.getModuleStreamIndex(), File); + if (!ModStreamData) + return ModStreamData.takeError(); + ModStream ModS(Modi.Info, std::move(*ModStreamData)); + if (auto EC = ModS.reload()) + return EC; + + if (ShouldDumpSymbols) { + ListScope SS(P, "Symbols"); + codeview::CVSymbolDumper SD(P, TD, nullptr, false); + bool HadError = false; + for (const auto &S : ModS.symbols(&HadError)) { + DictScope DD(P, ""); + + if (opts::raw::DumpModuleSyms) + SD.dump(S); + if (opts::raw::DumpSymRecordBytes) + P.printBinaryBlock("Bytes", S.Data); + } + if (HadError) + return make_error<RawError>( + raw_error_code::corrupt_file, + "DBI stream contained corrupt symbol record"); + } + if (opts::raw::DumpLineInfo) { + ListScope SS(P, "LineInfo"); + bool HadError = false; + // Define a locally scoped visitor to print the different + // substream types types. + class RecordVisitor : public codeview::IModuleSubstreamVisitor { + public: + RecordVisitor(ScopedPrinter &P, PDBFile &F) : P(P), F(F) {} + Error visitUnknown(ModuleSubstreamKind Kind, + StreamRef Stream) override { + DictScope DD(P, "Unknown"); + ArrayRef<uint8_t> Data; + StreamReader R(Stream); + if (auto EC = R.readBytes(Data, R.bytesRemaining())) { + return make_error<RawError>( + raw_error_code::corrupt_file, + "DBI stream contained corrupt line info record"); + } + P.printBinaryBlock("Data", Data); + return Error::success(); + } + Error + visitFileChecksums(StreamRef Data, + const FileChecksumArray &Checksums) override { + DictScope DD(P, "FileChecksums"); + for (const auto &C : Checksums) { + DictScope DDD(P, "Checksum"); + if (auto Result = getFileNameForOffset(C.FileNameOffset)) + P.printString("FileName", Result.get()); + else + return Result.takeError(); + P.flush(); + P.printEnum("Kind", uint8_t(C.Kind), getFileChecksumNames()); + P.printBinaryBlock("Checksum", C.Checksum); + } + return Error::success(); + } + + Error visitLines(StreamRef Data, const LineSubstreamHeader *Header, + const LineInfoArray &Lines) override { + DictScope DD(P, "Lines"); + for (const auto &L : Lines) { + if (auto Result = getFileNameForOffset2(L.NameIndex)) + P.printString("FileName", Result.get()); + else + return Result.takeError(); + P.flush(); + for (const auto &N : L.LineNumbers) { + DictScope DDD(P, "Line"); + LineInfo LI(N.Flags); + P.printNumber("Offset", N.Offset); + if (LI.isAlwaysStepInto()) + P.printString("StepInto", StringRef("Always")); + else if (LI.isNeverStepInto()) + P.printString("StepInto", StringRef("Never")); + else + P.printNumber("LineNumberStart", LI.getStartLine()); + P.printNumber("EndDelta", LI.getLineDelta()); + P.printBoolean("IsStatement", LI.isStatement()); + } + for (const auto &C : L.Columns) { + DictScope DDD(P, "Column"); + P.printNumber("Start", C.StartColumn); + P.printNumber("End", C.EndColumn); + } + } + return Error::success(); + } + + private: + Expected<StringRef> getFileNameForOffset(uint32_t Offset) { + auto ST = F.getStringTable(); + if (!ST) + return ST.takeError(); + + return ST->getStringForID(Offset); + } + Expected<StringRef> getFileNameForOffset2(uint32_t Offset) { + auto DS = F.getPDBDbiStream(); + if (!DS) + return DS.takeError(); + return DS->getFileNameForIndex(Offset); + } + ScopedPrinter &P; + PDBFile &F; + }; + + RecordVisitor V(P, File); + for (const auto &L : ModS.lines(&HadError)) { + if (auto EC = codeview::visitModuleSubstream(L, V)) + return EC; + } + } + } + } + } + return Error::success(); +} + +Error LLVMOutputStyle::dumpSectionContribs() { + if (!opts::raw::DumpSectionContribs) + return Error::success(); + + auto Dbi = File.getPDBDbiStream(); + if (!Dbi) + return Dbi.takeError(); + + ListScope L(P, "Section Contributions"); + class Visitor : public ISectionContribVisitor { + public: + Visitor(ScopedPrinter &P, DbiStream &DS) : P(P), DS(DS) {} + void visit(const SectionContrib &SC) override { + DictScope D(P, "Contribution"); + P.printNumber("ISect", SC.ISect); + P.printNumber("Off", SC.Off); + P.printNumber("Size", SC.Size); + P.printFlags("Characteristics", SC.Characteristics, + codeview::getImageSectionCharacteristicNames(), + COFF::SectionCharacteristics(0x00F00000)); + { + DictScope DD(P, "Module"); + P.printNumber("Index", SC.Imod); + auto M = DS.modules(); + if (M.size() > SC.Imod) { + P.printString("Name", M[SC.Imod].Info.getModuleName()); + } + } + P.printNumber("Data CRC", SC.DataCrc); + P.printNumber("Reloc CRC", SC.RelocCrc); + P.flush(); + } + void visit(const SectionContrib2 &SC) override { + visit(SC.Base); + P.printNumber("ISect Coff", SC.ISectCoff); + P.flush(); + } + + private: + ScopedPrinter &P; + DbiStream &DS; + }; + Visitor V(P, *Dbi); + Dbi->visitSectionContributions(V); + return Error::success(); +} + +Error LLVMOutputStyle::dumpSectionMap() { + if (!opts::raw::DumpSectionMap) + return Error::success(); + + auto Dbi = File.getPDBDbiStream(); + if (!Dbi) + return Dbi.takeError(); + + ListScope L(P, "Section Map"); + for (auto &M : Dbi->getSectionMap()) { + DictScope D(P, "Entry"); + P.printFlags("Flags", M.Flags, getOMFSegMapDescFlagNames()); + P.printNumber("Flags", M.Flags); + P.printNumber("Ovl", M.Ovl); + P.printNumber("Group", M.Group); + P.printNumber("Frame", M.Frame); + P.printNumber("SecName", M.SecName); + P.printNumber("ClassName", M.ClassName); + P.printNumber("Offset", M.Offset); + P.printNumber("SecByteLength", M.SecByteLength); + P.flush(); + } + return Error::success(); +} + +Error LLVMOutputStyle::dumpPublicsStream() { + if (!opts::raw::DumpPublics) + return Error::success(); + + DictScope D(P, "Publics Stream"); + auto Publics = File.getPDBPublicsStream(); + if (!Publics) + return Publics.takeError(); + + auto Dbi = File.getPDBDbiStream(); + if (!Dbi) + return Dbi.takeError(); + + P.printNumber("Stream number", Dbi->getPublicSymbolStreamIndex()); + P.printNumber("SymHash", Publics->getSymHash()); + P.printNumber("AddrMap", Publics->getAddrMap()); + P.printNumber("Number of buckets", Publics->getNumBuckets()); + P.printList("Hash Buckets", Publics->getHashBuckets()); + P.printList("Address Map", Publics->getAddressMap()); + P.printList("Thunk Map", Publics->getThunkMap()); + P.printList("Section Offsets", Publics->getSectionOffsets(), + printSectionOffset); + ListScope L(P, "Symbols"); + codeview::CVSymbolDumper SD(P, TD, nullptr, false); + bool HadError = false; + for (auto S : Publics->getSymbols(&HadError)) { + DictScope DD(P, ""); + + SD.dump(S); + if (opts::raw::DumpSymRecordBytes) + P.printBinaryBlock("Bytes", S.Data); + } + if (HadError) + return make_error<RawError>( + raw_error_code::corrupt_file, + "Public symbol stream contained corrupt record"); + + return Error::success(); +} + +Error LLVMOutputStyle::dumpSectionHeaders() { + if (!opts::raw::DumpSectionHeaders) + return Error::success(); + + auto Dbi = File.getPDBDbiStream(); + if (!Dbi) + return Dbi.takeError(); + + ListScope D(P, "Section Headers"); + for (const object::coff_section &Section : Dbi->getSectionHeaders()) { + DictScope DD(P, ""); + + // If a name is 8 characters long, there is no NUL character at end. + StringRef Name(Section.Name, strnlen(Section.Name, sizeof(Section.Name))); + P.printString("Name", Name); + P.printNumber("Virtual Size", Section.VirtualSize); + P.printNumber("Virtual Address", Section.VirtualAddress); + P.printNumber("Size of Raw Data", Section.SizeOfRawData); + P.printNumber("File Pointer to Raw Data", Section.PointerToRawData); + P.printNumber("File Pointer to Relocations", Section.PointerToRelocations); + P.printNumber("File Pointer to Linenumbers", Section.PointerToLinenumbers); + P.printNumber("Number of Relocations", Section.NumberOfRelocations); + P.printNumber("Number of Linenumbers", Section.NumberOfLinenumbers); + P.printFlags("Characteristics", Section.Characteristics, + getImageSectionCharacteristicNames()); + } + return Error::success(); +} + +Error LLVMOutputStyle::dumpFpoStream() { + if (!opts::raw::DumpFpo) + return Error::success(); + + auto Dbi = File.getPDBDbiStream(); + if (!Dbi) + return Dbi.takeError(); + + ListScope D(P, "New FPO"); + for (const object::FpoData &Fpo : Dbi->getFpoRecords()) { + DictScope DD(P, ""); + P.printNumber("Offset", Fpo.Offset); + P.printNumber("Size", Fpo.Size); + P.printNumber("Number of locals", Fpo.NumLocals); + P.printNumber("Number of params", Fpo.NumParams); + P.printNumber("Size of Prolog", Fpo.getPrologSize()); + P.printNumber("Number of Saved Registers", Fpo.getNumSavedRegs()); + P.printBoolean("Has SEH", Fpo.hasSEH()); + P.printBoolean("Use BP", Fpo.useBP()); + P.printNumber("Frame Pointer", Fpo.getFP()); + } + return Error::success(); +} + +void LLVMOutputStyle::flush() { P.flush(); } diff --git a/tools/llvm-pdbdump/LLVMOutputStyle.h b/tools/llvm-pdbdump/LLVMOutputStyle.h new file mode 100644 index 0000000000000..77935d102205e --- /dev/null +++ b/tools/llvm-pdbdump/LLVMOutputStyle.h @@ -0,0 +1,50 @@ +//===- LLVMOutputStyle.h -------------------------------------- *- C++ --*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_TOOLS_LLVMPDBDUMP_LLVMOUTPUTSTYLE_H +#define LLVM_TOOLS_LLVMPDBDUMP_LLVMOUTPUTSTYLE_H + +#include "OutputStyle.h" + +#include "llvm/DebugInfo/CodeView/TypeDumper.h" +#include "llvm/Support/ScopedPrinter.h" + +namespace llvm { +namespace pdb { +class LLVMOutputStyle : public OutputStyle { +public: + LLVMOutputStyle(PDBFile &File); + + Error dump() override; + +private: + Error dumpFileHeaders(); + Error dumpStreamSummary(); + Error dumpStreamBlocks(); + Error dumpStreamData(); + Error dumpInfoStream(); + Error dumpNamedStream(); + Error dumpTpiStream(uint32_t StreamIdx); + Error dumpDbiStream(); + Error dumpSectionContribs(); + Error dumpSectionMap(); + Error dumpPublicsStream(); + Error dumpSectionHeaders(); + Error dumpFpoStream(); + + void flush(); + + PDBFile &File; + ScopedPrinter P; + codeview::CVTypeDumper TD; +}; +} +} + +#endif diff --git a/tools/llvm-pdbdump/LinePrinter.cpp b/tools/llvm-pdbdump/LinePrinter.cpp index a43727f02b5e5..47c7d3e3c0e74 100644 --- a/tools/llvm-pdbdump/LinePrinter.cpp +++ b/tools/llvm-pdbdump/LinePrinter.cpp @@ -16,6 +16,9 @@ #include <algorithm> +using namespace llvm; +using namespace llvm::pdb; + namespace { bool IsItemExcluded(llvm::StringRef Item, std::list<llvm::Regex> &IncludeFilters, @@ -41,19 +44,19 @@ using namespace llvm; LinePrinter::LinePrinter(int Indent, llvm::raw_ostream &Stream) : OS(Stream), IndentSpaces(Indent), CurrentIndent(0) { - SetFilters(ExcludeTypeFilters, opts::ExcludeTypes.begin(), - opts::ExcludeTypes.end()); - SetFilters(ExcludeSymbolFilters, opts::ExcludeSymbols.begin(), - opts::ExcludeSymbols.end()); - SetFilters(ExcludeCompilandFilters, opts::ExcludeCompilands.begin(), - opts::ExcludeCompilands.end()); - - SetFilters(IncludeTypeFilters, opts::IncludeTypes.begin(), - opts::IncludeTypes.end()); - SetFilters(IncludeSymbolFilters, opts::IncludeSymbols.begin(), - opts::IncludeSymbols.end()); - SetFilters(IncludeCompilandFilters, opts::IncludeCompilands.begin(), - opts::IncludeCompilands.end()); + SetFilters(ExcludeTypeFilters, opts::pretty::ExcludeTypes.begin(), + opts::pretty::ExcludeTypes.end()); + SetFilters(ExcludeSymbolFilters, opts::pretty::ExcludeSymbols.begin(), + opts::pretty::ExcludeSymbols.end()); + SetFilters(ExcludeCompilandFilters, opts::pretty::ExcludeCompilands.begin(), + opts::pretty::ExcludeCompilands.end()); + + SetFilters(IncludeTypeFilters, opts::pretty::IncludeTypes.begin(), + opts::pretty::IncludeTypes.end()); + SetFilters(IncludeSymbolFilters, opts::pretty::IncludeSymbols.begin(), + opts::pretty::IncludeSymbols.end()); + SetFilters(IncludeCompilandFilters, opts::pretty::IncludeCompilands.begin(), + opts::pretty::IncludeCompilands.end()); } void LinePrinter::Indent() { CurrentIndent += IndentSpaces; } diff --git a/tools/llvm-pdbdump/LinePrinter.h b/tools/llvm-pdbdump/LinePrinter.h index b0a9d2cfc4e82..a4401f8af9552 100644 --- a/tools/llvm-pdbdump/LinePrinter.h +++ b/tools/llvm-pdbdump/LinePrinter.h @@ -18,6 +18,7 @@ #include <list> namespace llvm { +namespace pdb { class LinePrinter { friend class WithColor; @@ -88,5 +89,6 @@ private: raw_ostream &OS; }; } +} #endif diff --git a/tools/llvm-pdbdump/Makefile b/tools/llvm-pdbdump/Makefile deleted file mode 100644 index 18aafc4076feb..0000000000000 --- a/tools/llvm-pdbdump/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-pdbdump/Makefile -------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-pdbdump -LINK_COMPONENTS := DebugInfoPDB Object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-pdbdump/OutputStyle.h b/tools/llvm-pdbdump/OutputStyle.h new file mode 100644 index 0000000000000..dfefc25a215e8 --- /dev/null +++ b/tools/llvm-pdbdump/OutputStyle.h @@ -0,0 +1,28 @@ +//===- OutputStyle.h ------------------------------------------ *- C++ --*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_TOOLS_LLVMPDBDUMP_OUTPUTSTYLE_H +#define LLVM_TOOLS_LLVMPDBDUMP_OUTPUTSTYLE_H + +#include "llvm/Support/Error.h" + +namespace llvm { +namespace pdb { +class PDBFile; + +class OutputStyle { +public: + virtual ~OutputStyle() {} + + virtual Error dump() = 0; +}; +} +} + +#endif diff --git a/tools/llvm-pdbdump/PdbYaml.cpp b/tools/llvm-pdbdump/PdbYaml.cpp new file mode 100644 index 0000000000000..e83f24e877422 --- /dev/null +++ b/tools/llvm-pdbdump/PdbYaml.cpp @@ -0,0 +1,162 @@ +//===- PdbYAML.cpp -------------------------------------------- *- C++ --*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "PdbYaml.h" + +#include "llvm/DebugInfo/PDB/PDBExtras.h" +#include "llvm/DebugInfo/PDB/Raw/PDBFile.h" + +using namespace llvm; +using namespace llvm::yaml; +using namespace llvm::pdb; +using namespace llvm::pdb::yaml; + +namespace llvm { +namespace yaml { +template <> struct ScalarTraits<llvm::pdb::PDB_UniqueId> { + static void output(const llvm::pdb::PDB_UniqueId &S, void *, + llvm::raw_ostream &OS) { + OS << S; + } + + static StringRef input(StringRef Scalar, void *Ctx, + llvm::pdb::PDB_UniqueId &S) { + if (Scalar.size() != 38) + return "GUID strings are 38 characters long"; + if (Scalar[0] != '{' || Scalar[37] != '}') + return "GUID is not enclosed in {}"; + if (Scalar[9] != '-' || Scalar[14] != '-' || Scalar[19] != '-' || + Scalar[24] != '-') + return "GUID sections are not properly delineated with dashes"; + + char *OutBuffer = S.Guid; + for (auto Iter = Scalar.begin(); Iter != Scalar.end();) { + if (*Iter == '-' || *Iter == '{' || *Iter == '}') { + ++Iter; + continue; + } + uint8_t Value = (llvm::hexDigitValue(*Iter) << 4); + ++Iter; + Value |= llvm::hexDigitValue(*Iter); + ++Iter; + *OutBuffer++ = Value; + } + + return ""; + } + + static bool mustQuote(StringRef Scalar) { return needsQuotes(Scalar); } +}; + +template <> struct ScalarEnumerationTraits<llvm::pdb::PDB_Machine> { + static void enumeration(IO &io, llvm::pdb::PDB_Machine &Value) { + io.enumCase(Value, "Invalid", PDB_Machine::Invalid); + io.enumCase(Value, "Am33", PDB_Machine::Am33); + io.enumCase(Value, "Amd64", PDB_Machine::Amd64); + io.enumCase(Value, "Arm", PDB_Machine::Arm); + io.enumCase(Value, "ArmNT", PDB_Machine::ArmNT); + io.enumCase(Value, "Ebc", PDB_Machine::Ebc); + io.enumCase(Value, "x86", PDB_Machine::x86); + io.enumCase(Value, "Ia64", PDB_Machine::Ia64); + io.enumCase(Value, "M32R", PDB_Machine::M32R); + io.enumCase(Value, "Mips16", PDB_Machine::Mips16); + io.enumCase(Value, "MipsFpu", PDB_Machine::MipsFpu); + io.enumCase(Value, "MipsFpu16", PDB_Machine::MipsFpu16); + io.enumCase(Value, "PowerPCFP", PDB_Machine::PowerPCFP); + io.enumCase(Value, "R4000", PDB_Machine::R4000); + io.enumCase(Value, "SH3", PDB_Machine::SH3); + io.enumCase(Value, "SH3DSP", PDB_Machine::SH3DSP); + io.enumCase(Value, "Thumb", PDB_Machine::Thumb); + io.enumCase(Value, "WceMipsV2", PDB_Machine::WceMipsV2); + } +}; + +template <> struct ScalarEnumerationTraits<llvm::pdb::PdbRaw_DbiVer> { + static void enumeration(IO &io, llvm::pdb::PdbRaw_DbiVer &Value) { + io.enumCase(Value, "V41", llvm::pdb::PdbRaw_DbiVer::PdbDbiVC41); + io.enumCase(Value, "V50", llvm::pdb::PdbRaw_DbiVer::PdbDbiV50); + io.enumCase(Value, "V60", llvm::pdb::PdbRaw_DbiVer::PdbDbiV60); + io.enumCase(Value, "V70", llvm::pdb::PdbRaw_DbiVer::PdbDbiV70); + io.enumCase(Value, "V110", llvm::pdb::PdbRaw_DbiVer::PdbDbiV110); + } +}; + +template <> struct ScalarEnumerationTraits<llvm::pdb::PdbRaw_ImplVer> { + static void enumeration(IO &io, llvm::pdb::PdbRaw_ImplVer &Value) { + io.enumCase(Value, "VC2", llvm::pdb::PdbRaw_ImplVer::PdbImplVC2); + io.enumCase(Value, "VC4", llvm::pdb::PdbRaw_ImplVer::PdbImplVC4); + io.enumCase(Value, "VC41", llvm::pdb::PdbRaw_ImplVer::PdbImplVC41); + io.enumCase(Value, "VC50", llvm::pdb::PdbRaw_ImplVer::PdbImplVC50); + io.enumCase(Value, "VC98", llvm::pdb::PdbRaw_ImplVer::PdbImplVC98); + io.enumCase(Value, "VC70Dep", llvm::pdb::PdbRaw_ImplVer::PdbImplVC70Dep); + io.enumCase(Value, "VC70", llvm::pdb::PdbRaw_ImplVer::PdbImplVC70); + io.enumCase(Value, "VC80", llvm::pdb::PdbRaw_ImplVer::PdbImplVC80); + io.enumCase(Value, "VC110", llvm::pdb::PdbRaw_ImplVer::PdbImplVC110); + io.enumCase(Value, "VC140", llvm::pdb::PdbRaw_ImplVer::PdbImplVC140); + } +}; +} +} + +void MappingTraits<PdbObject>::mapping(IO &IO, PdbObject &Obj) { + IO.mapOptional("MSF", Obj.Headers); + IO.mapOptional("StreamSizes", Obj.StreamSizes); + IO.mapOptional("StreamMap", Obj.StreamMap); + IO.mapOptional("PdbStream", Obj.PdbStream); + IO.mapOptional("DbiStream", Obj.DbiStream); +} + +void MappingTraits<MsfHeaders>::mapping(IO &IO, MsfHeaders &Obj) { + IO.mapRequired("SuperBlock", Obj.SuperBlock); + IO.mapRequired("NumDirectoryBlocks", Obj.NumDirectoryBlocks); + IO.mapRequired("DirectoryBlocks", Obj.DirectoryBlocks); + IO.mapRequired("NumStreams", Obj.NumStreams); + IO.mapRequired("FileSize", Obj.FileSize); +} + +void MappingTraits<msf::SuperBlock>::mapping(IO &IO, msf::SuperBlock &SB) { + if (!IO.outputting()) { + ::memcpy(SB.MagicBytes, msf::Magic, sizeof(msf::Magic)); + } + + IO.mapRequired("BlockSize", SB.BlockSize); + IO.mapRequired("FreeBlockMap", SB.FreeBlockMapBlock); + IO.mapRequired("NumBlocks", SB.NumBlocks); + IO.mapRequired("NumDirectoryBytes", SB.NumDirectoryBytes); + IO.mapRequired("Unknown1", SB.Unknown1); + IO.mapRequired("BlockMapAddr", SB.BlockMapAddr); +} + +void MappingTraits<StreamBlockList>::mapping(IO &IO, StreamBlockList &SB) { + IO.mapRequired("Stream", SB.Blocks); +} + +void MappingTraits<PdbInfoStream>::mapping(IO &IO, PdbInfoStream &Obj) { + IO.mapRequired("Age", Obj.Age); + IO.mapRequired("Guid", Obj.Guid); + IO.mapRequired("Signature", Obj.Signature); + IO.mapRequired("Version", Obj.Version); + IO.mapRequired("NamedStreams", Obj.NamedStreams); +} + +void MappingTraits<PdbDbiStream>::mapping(IO &IO, PdbDbiStream &Obj) { + IO.mapRequired("VerHeader", Obj.VerHeader); + IO.mapRequired("Age", Obj.Age); + IO.mapRequired("BuildNumber", Obj.BuildNumber); + IO.mapRequired("PdbDllVersion", Obj.PdbDllVersion); + IO.mapRequired("PdbDllRbld", Obj.PdbDllRbld); + IO.mapRequired("Flags", Obj.Flags); + IO.mapRequired("MachineType", Obj.MachineType); +} + +void MappingTraits<NamedStreamMapping>::mapping(IO &IO, + NamedStreamMapping &Obj) { + IO.mapRequired("Name", Obj.StreamName); + IO.mapRequired("StreamNum", Obj.StreamNumber); +} diff --git a/tools/llvm-pdbdump/PdbYaml.h b/tools/llvm-pdbdump/PdbYaml.h new file mode 100644 index 0000000000000..a7389af220572 --- /dev/null +++ b/tools/llvm-pdbdump/PdbYaml.h @@ -0,0 +1,112 @@ +//===- PdbYAML.h ---------------------------------------------- *- C++ --*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_TOOLS_LLVMPDBDUMP_PDBYAML_H +#define LLVM_TOOLS_LLVMPDBDUMP_PDBYAML_H + +#include "OutputStyle.h" + +#include "llvm/ADT/Optional.h" +#include "llvm/DebugInfo/PDB/PDBTypes.h" +#include "llvm/DebugInfo/PDB/Raw/MsfCommon.h" +#include "llvm/DebugInfo/PDB/Raw/PDBFile.h" +#include "llvm/DebugInfo/PDB/Raw/RawConstants.h" +#include "llvm/Support/Endian.h" +#include "llvm/Support/YAMLTraits.h" + +#include <vector> + +namespace llvm { +namespace pdb { + +namespace yaml { +struct MsfHeaders { + msf::SuperBlock SuperBlock; + uint32_t NumDirectoryBlocks; + std::vector<uint32_t> DirectoryBlocks; + uint32_t NumStreams; + uint32_t FileSize; +}; + +struct StreamBlockList { + std::vector<uint32_t> Blocks; +}; + +struct NamedStreamMapping { + StringRef StreamName; + uint32_t StreamNumber; +}; + +struct PdbInfoStream { + PdbRaw_ImplVer Version; + uint32_t Signature; + uint32_t Age; + PDB_UniqueId Guid; + std::vector<NamedStreamMapping> NamedStreams; +}; + +struct PdbDbiStream { + PdbRaw_DbiVer VerHeader; + uint32_t Age; + uint16_t BuildNumber; + uint32_t PdbDllVersion; + uint16_t PdbDllRbld; + uint16_t Flags; + PDB_Machine MachineType; +}; + +struct PdbObject { + Optional<MsfHeaders> Headers; + Optional<std::vector<uint32_t>> StreamSizes; + Optional<std::vector<StreamBlockList>> StreamMap; + Optional<PdbInfoStream> PdbStream; + Optional<PdbDbiStream> DbiStream; +}; +} +} +} + +namespace llvm { +namespace yaml { + +template <> struct MappingTraits<pdb::yaml::PdbObject> { + static void mapping(IO &IO, pdb::yaml::PdbObject &Obj); +}; + +template <> struct MappingTraits<pdb::yaml::MsfHeaders> { + static void mapping(IO &IO, pdb::yaml::MsfHeaders &Obj); +}; + +template <> struct MappingTraits<pdb::msf::SuperBlock> { + static void mapping(IO &IO, pdb::msf::SuperBlock &SB); +}; + +template <> struct MappingTraits<pdb::yaml::StreamBlockList> { + static void mapping(IO &IO, pdb::yaml::StreamBlockList &SB); +}; + +template <> struct MappingTraits<pdb::yaml::PdbInfoStream> { + static void mapping(IO &IO, pdb::yaml::PdbInfoStream &Obj); +}; + +template <> struct MappingTraits<pdb::yaml::PdbDbiStream> { + static void mapping(IO &IO, pdb::yaml::PdbDbiStream &Obj); +}; + +template <> struct MappingTraits<pdb::yaml::NamedStreamMapping> { + static void mapping(IO &IO, pdb::yaml::NamedStreamMapping &Obj); +}; +} +} + +LLVM_YAML_IS_SEQUENCE_VECTOR(uint32_t) +LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::pdb::yaml::NamedStreamMapping) +LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::pdb::yaml::StreamBlockList) + +#endif // LLVM_TOOLS_LLVMPDBDUMP_PDBYAML_H diff --git a/tools/llvm-pdbdump/TypeDumper.cpp b/tools/llvm-pdbdump/TypeDumper.cpp index 88c0bd65697ed..a49d640455539 100644 --- a/tools/llvm-pdbdump/TypeDumper.cpp +++ b/tools/llvm-pdbdump/TypeDumper.cpp @@ -24,6 +24,7 @@ #include "llvm/DebugInfo/PDB/PDBSymbolTypeUDT.h" using namespace llvm; +using namespace llvm::pdb; TypeDumper::TypeDumper(LinePrinter &P) : PDBSymDumper(true), Printer(P) {} @@ -62,7 +63,7 @@ void TypeDumper::dump(const PDBSymbolTypeEnum &Symbol) { if (Printer.IsTypeExcluded(Symbol.getName())) return; // Dump member enums when dumping their class definition. - if (Symbol.isNested()) + if (nullptr != Symbol.getClassParent()) return; Printer.NewLine(); @@ -87,7 +88,7 @@ void TypeDumper::dump(const PDBSymbolTypeUDT &Symbol) { Printer.NewLine(); - if (opts::NoClassDefs) { + if (opts::pretty::NoClassDefs) { WithColor(Printer, PDB_ColorItem::Keyword).get() << "class "; WithColor(Printer, PDB_ColorItem::Identifier).get() << Symbol.getName(); } else { diff --git a/tools/llvm-pdbdump/TypeDumper.h b/tools/llvm-pdbdump/TypeDumper.h index 5c0832eccaf98..76a477964f1f7 100644 --- a/tools/llvm-pdbdump/TypeDumper.h +++ b/tools/llvm-pdbdump/TypeDumper.h @@ -13,7 +13,7 @@ #include "llvm/DebugInfo/PDB/PDBSymDumper.h" namespace llvm { - +namespace pdb { class LinePrinter; class TypeDumper : public PDBSymDumper { @@ -30,5 +30,5 @@ private: LinePrinter &Printer; }; } - +} #endif diff --git a/tools/llvm-pdbdump/TypedefDumper.cpp b/tools/llvm-pdbdump/TypedefDumper.cpp index a6b5c53a6b3a9..b1e017613ce12 100644 --- a/tools/llvm-pdbdump/TypedefDumper.cpp +++ b/tools/llvm-pdbdump/TypedefDumper.cpp @@ -23,6 +23,7 @@ #include "llvm/DebugInfo/PDB/PDBSymbolTypeUDT.h" using namespace llvm; +using namespace llvm::pdb; TypedefDumper::TypedefDumper(LinePrinter &P) : PDBSymDumper(true), Printer(P) {} diff --git a/tools/llvm-pdbdump/TypedefDumper.h b/tools/llvm-pdbdump/TypedefDumper.h index 8cd578cc6d296..c22b58a7e41e4 100644 --- a/tools/llvm-pdbdump/TypedefDumper.h +++ b/tools/llvm-pdbdump/TypedefDumper.h @@ -13,6 +13,7 @@ #include "llvm/DebugInfo/PDB/PDBSymDumper.h" namespace llvm { +namespace pdb { class LinePrinter; @@ -33,5 +34,6 @@ private: LinePrinter &Printer; }; } +} #endif diff --git a/tools/llvm-pdbdump/VariableDumper.cpp b/tools/llvm-pdbdump/VariableDumper.cpp index e5665b5fcafa4..284d7e9b731f9 100644 --- a/tools/llvm-pdbdump/VariableDumper.cpp +++ b/tools/llvm-pdbdump/VariableDumper.cpp @@ -27,12 +27,13 @@ #include "llvm/Support/Format.h" using namespace llvm; +using namespace llvm::pdb; VariableDumper::VariableDumper(LinePrinter &P) : PDBSymDumper(true), Printer(P) {} void VariableDumper::start(const PDBSymbolData &Var) { - if (Var.isCompilerGenerated() && opts::ExcludeCompilerGenerated) + if (Var.isCompilerGenerated() && opts::pretty::ExcludeCompilerGenerated) return; if (Printer.IsSymbolExcluded(Var.getName())) return; diff --git a/tools/llvm-pdbdump/VariableDumper.h b/tools/llvm-pdbdump/VariableDumper.h index db8d8ea0e4305..4f00358878c94 100644 --- a/tools/llvm-pdbdump/VariableDumper.h +++ b/tools/llvm-pdbdump/VariableDumper.h @@ -11,10 +11,13 @@ #define LLVM_TOOLS_LLVMPDBDUMP_VARIABLEDUMPER_H #include "llvm/DebugInfo/PDB/PDBSymDumper.h" -#include "llvm/ADT/StringRef.h" namespace llvm { +class StringRef; + +namespace pdb { + class LinePrinter; class VariableDumper : public PDBSymDumper { @@ -37,5 +40,5 @@ private: LinePrinter &Printer; }; } - +} #endif diff --git a/tools/llvm-pdbdump/YAMLOutputStyle.cpp b/tools/llvm-pdbdump/YAMLOutputStyle.cpp new file mode 100644 index 0000000000000..a8b6202d5d161 --- /dev/null +++ b/tools/llvm-pdbdump/YAMLOutputStyle.cpp @@ -0,0 +1,142 @@ +//===- YAMLOutputStyle.cpp ------------------------------------ *- C++ --*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "YAMLOutputStyle.h" + +#include "PdbYaml.h" +#include "llvm-pdbdump.h" + +#include "llvm/DebugInfo/PDB/Raw/DbiStream.h" +#include "llvm/DebugInfo/PDB/Raw/InfoStream.h" +#include "llvm/DebugInfo/PDB/Raw/PDBFile.h" +#include "llvm/DebugInfo/PDB/Raw/RawConstants.h" + +using namespace llvm; +using namespace llvm::pdb; + +YAMLOutputStyle::YAMLOutputStyle(PDBFile &File) : File(File), Out(outs()) {} + +Error YAMLOutputStyle::dump() { + if (opts::pdb2yaml::StreamDirectory) + opts::pdb2yaml::StreamMetadata = true; + + if (auto EC = dumpFileHeaders()) + return EC; + + if (auto EC = dumpStreamMetadata()) + return EC; + + if (auto EC = dumpStreamDirectory()) + return EC; + + if (auto EC = dumpPDBStream()) + return EC; + + if (auto EC = dumpDbiStream()) + return EC; + + flush(); + return Error::success(); +} + +Error YAMLOutputStyle::dumpFileHeaders() { + if (opts::pdb2yaml::NoFileHeaders) + return Error::success(); + + yaml::MsfHeaders Headers; + Obj.Headers.emplace(); + Obj.Headers->SuperBlock.NumBlocks = File.getBlockCount(); + Obj.Headers->SuperBlock.BlockMapAddr = File.getBlockMapIndex(); + Obj.Headers->SuperBlock.BlockSize = File.getBlockSize(); + auto Blocks = File.getDirectoryBlockArray(); + Obj.Headers->DirectoryBlocks.assign(Blocks.begin(), Blocks.end()); + Obj.Headers->NumDirectoryBlocks = File.getNumDirectoryBlocks(); + Obj.Headers->SuperBlock.NumDirectoryBytes = File.getNumDirectoryBytes(); + Obj.Headers->NumStreams = + opts::pdb2yaml::StreamMetadata ? File.getNumStreams() : 0; + Obj.Headers->SuperBlock.FreeBlockMapBlock = File.getFreeBlockMapBlock(); + Obj.Headers->SuperBlock.Unknown1 = File.getUnknown1(); + Obj.Headers->FileSize = File.getFileSize(); + + return Error::success(); +} + +Error YAMLOutputStyle::dumpStreamMetadata() { + if (!opts::pdb2yaml::StreamMetadata) + return Error::success(); + + Obj.StreamSizes.emplace(); + Obj.StreamSizes->assign(File.getStreamSizes().begin(), + File.getStreamSizes().end()); + return Error::success(); +} + +Error YAMLOutputStyle::dumpStreamDirectory() { + if (!opts::pdb2yaml::StreamDirectory) + return Error::success(); + + auto StreamMap = File.getStreamMap(); + Obj.StreamMap.emplace(); + for (auto &Stream : StreamMap) { + pdb::yaml::StreamBlockList BlockList; + BlockList.Blocks.assign(Stream.begin(), Stream.end()); + Obj.StreamMap->push_back(BlockList); + } + + return Error::success(); +} + +Error YAMLOutputStyle::dumpPDBStream() { + if (!opts::pdb2yaml::PdbStream) + return Error::success(); + + auto IS = File.getPDBInfoStream(); + if (!IS) + return IS.takeError(); + + auto &InfoS = IS.get(); + Obj.PdbStream.emplace(); + Obj.PdbStream->Age = InfoS.getAge(); + Obj.PdbStream->Guid = InfoS.getGuid(); + Obj.PdbStream->Signature = InfoS.getSignature(); + Obj.PdbStream->Version = InfoS.getVersion(); + for (auto &NS : InfoS.named_streams()) { + yaml::NamedStreamMapping Mapping; + Mapping.StreamName = NS.getKey(); + Mapping.StreamNumber = NS.getValue(); + Obj.PdbStream->NamedStreams.push_back(Mapping); + } + + return Error::success(); +} + +Error YAMLOutputStyle::dumpDbiStream() { + if (!opts::pdb2yaml::DbiStream) + return Error::success(); + + auto DbiS = File.getPDBDbiStream(); + if (!DbiS) + return DbiS.takeError(); + + auto &DS = DbiS.get(); + Obj.DbiStream.emplace(); + Obj.DbiStream->Age = DS.getAge(); + Obj.DbiStream->BuildNumber = DS.getBuildNumber(); + Obj.DbiStream->Flags = DS.getFlags(); + Obj.DbiStream->MachineType = DS.getMachineType(); + Obj.DbiStream->PdbDllRbld = DS.getPdbDllRbld(); + Obj.DbiStream->PdbDllVersion = DS.getPdbDllVersion(); + Obj.DbiStream->VerHeader = DS.getDbiVersion(); + return Error::success(); +} + +void YAMLOutputStyle::flush() { + Out << Obj; + outs().flush(); +} diff --git a/tools/llvm-pdbdump/YAMLOutputStyle.h b/tools/llvm-pdbdump/YAMLOutputStyle.h new file mode 100644 index 0000000000000..d36dfec5f25ad --- /dev/null +++ b/tools/llvm-pdbdump/YAMLOutputStyle.h @@ -0,0 +1,45 @@ +//===- YAMLOutputStyle.h -------------------------------------- *- C++ --*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_TOOLS_LLVMPDBDUMP_YAMLOUTPUTSTYLE_H +#define LLVM_TOOLS_LLVMPDBDUMP_YAMLOUTPUTSTYLE_H + +#include "OutputStyle.h" +#include "PdbYaml.h" + +#include "llvm/DebugInfo/CodeView/TypeDumper.h" +#include "llvm/Support/ScopedPrinter.h" +#include "llvm/Support/YAMLTraits.h" + +namespace llvm { +namespace pdb { +class YAMLOutputStyle : public OutputStyle { +public: + YAMLOutputStyle(PDBFile &File); + + Error dump() override; + +private: + Error dumpFileHeaders(); + Error dumpStreamMetadata(); + Error dumpStreamDirectory(); + Error dumpPDBStream(); + Error dumpDbiStream(); + + void flush(); + + PDBFile &File; + llvm::yaml::Output Out; + + yaml::PdbObject Obj; +}; +} // namespace pdb +} // namespace llvm + +#endif // LLVM_TOOLS_LLVMPDBDUMP_YAMLOUTPUTSTYLE_H diff --git a/tools/llvm-pdbdump/fuzzer/CMakeLists.txt b/tools/llvm-pdbdump/fuzzer/CMakeLists.txt new file mode 100644 index 0000000000000..cf5a0f70aab38 --- /dev/null +++ b/tools/llvm-pdbdump/fuzzer/CMakeLists.txt @@ -0,0 +1,15 @@ +set(LLVM_LINK_COMPONENTS + DebugInfoCodeView + DebugInfoPDB + Object + Support + ) + +add_llvm_executable(llvm-pdbdump-fuzzer + EXCLUDE_FROM_ALL + llvm-pdbdump-fuzzer.cpp + ) + +target_link_libraries(llvm-pdbdump-fuzzer + LLVMFuzzer + ) diff --git a/tools/llvm-pdbdump/fuzzer/llvm-pdbdump-fuzzer.cpp b/tools/llvm-pdbdump/fuzzer/llvm-pdbdump-fuzzer.cpp new file mode 100644 index 0000000000000..e818dda32fc02 --- /dev/null +++ b/tools/llvm-pdbdump/fuzzer/llvm-pdbdump-fuzzer.cpp @@ -0,0 +1,104 @@ +//===-- llvm-pdbdump-fuzzer.cpp - Fuzz the llvm-pdbdump tool --------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +/// +/// \file +/// \brief This file implements a function that runs llvm-pdbdump +/// on a single input. This function is then linked into the Fuzzer library. +/// +//===----------------------------------------------------------------------===// +#include "llvm/ADT/STLExtras.h" +#include "llvm/DebugInfo/CodeView/ByteStream.h" +#include "llvm/DebugInfo/CodeView/SymbolDumper.h" +#include "llvm/DebugInfo/CodeView/TypeDumper.h" +#include "llvm/DebugInfo/PDB/Raw/DbiStream.h" +#include "llvm/DebugInfo/PDB/Raw/IPDBStreamData.h" +#include "llvm/DebugInfo/PDB/Raw/MappedBlockStream.h" +#include "llvm/DebugInfo/PDB/Raw/ModStream.h" +#include "llvm/DebugInfo/PDB/Raw/PDBFile.h" +#include "llvm/DebugInfo/PDB/Raw/RawSession.h" +#include "llvm/Support/MemoryBuffer.h" +#include "llvm/Support/ScopedPrinter.h" + +using namespace llvm; + +namespace { +// We need a class which behaves like an immutable ByteStream, but whose data +// is backed by an llvm::MemoryBuffer. It also needs to own the underlying +// MemoryBuffer, so this simple adapter is a good way to achieve that. +class InputByteStream : public codeview::ByteStream<false> { +public: + explicit InputByteStream(std::unique_ptr<MemoryBuffer> Buffer) + : ByteStream(ArrayRef<uint8_t>(Buffer->getBuffer().bytes_begin(), + Buffer->getBuffer().bytes_end())), + MemBuffer(std::move(Buffer)) {} + + std::unique_ptr<MemoryBuffer> MemBuffer; +}; +} + +extern "C" int LLVMFuzzerTestOneInput(uint8_t *data, size_t size) { + std::unique_ptr<MemoryBuffer> Buff = MemoryBuffer::getMemBuffer( + StringRef((const char *)data, size), "", false); + + ScopedPrinter P(nulls()); + codeview::CVTypeDumper TD(&P, false); + + auto InputStream = llvm::make_unique<InputByteStream>(std::move(Buff)); + std::unique_ptr<pdb::PDBFile> File(new pdb::PDBFile(std::move(InputStream))); + if (auto E = File->parseFileHeaders()) { + consumeError(std::move(E)); + return 0; + } + if (auto E = File->parseStreamData()) { + consumeError(std::move(E)); + return 0; + } + + auto DbiS = File->getPDBDbiStream(); + if (auto E = DbiS.takeError()) { + consumeError(std::move(E)); + return 0; + } + auto TpiS = File->getPDBTpiStream(); + if (auto E = TpiS.takeError()) { + consumeError(std::move(E)); + return 0; + } + auto IpiS = File->getPDBIpiStream(); + if (auto E = IpiS.takeError()) { + consumeError(std::move(E)); + return 0; + } + auto InfoS = File->getPDBInfoStream(); + if (auto E = InfoS.takeError()) { + consumeError(std::move(E)); + return 0; + } + pdb::DbiStream &DS = DbiS.get(); + + for (auto &Modi : DS.modules()) { + auto ModStreamData = pdb::MappedBlockStream::createIndexedStream( + Modi.Info.getModuleStreamIndex(), *File); + if (!ModStreamData) { + consumeError(ModStreamData.takeError()); + return 0; + } + pdb::ModStream ModS(Modi.Info, std::move(*ModStreamData)); + if (auto E = ModS.reload()) { + consumeError(std::move(E)); + return 0; + } + codeview::CVSymbolDumper SD(P, TD, nullptr, false); + bool HadError = false; + for (auto &S : ModS.symbols(&HadError)) { + SD.dump(S); + } + } + return 0; +} diff --git a/tools/llvm-pdbdump/llvm-pdbdump.cpp b/tools/llvm-pdbdump/llvm-pdbdump.cpp index 0e3f0b281fe02..62f215ddc8aef 100644 --- a/tools/llvm-pdbdump/llvm-pdbdump.cpp +++ b/tools/llvm-pdbdump/llvm-pdbdump.cpp @@ -17,15 +17,20 @@ #include "CompilandDumper.h" #include "ExternalSymbolDumper.h" #include "FunctionDumper.h" +#include "LLVMOutputStyle.h" #include "LinePrinter.h" +#include "OutputStyle.h" #include "TypeDumper.h" #include "VariableDumper.h" +#include "YAMLOutputStyle.h" #include "llvm/ADT/ArrayRef.h" #include "llvm/ADT/BitVector.h" #include "llvm/ADT/DenseMap.h" #include "llvm/ADT/StringExtras.h" #include "llvm/Config/config.h" +#include "llvm/DebugInfo/CodeView/ByteStream.h" +#include "llvm/DebugInfo/PDB/GenericError.h" #include "llvm/DebugInfo/PDB/IPDBEnumChildren.h" #include "llvm/DebugInfo/PDB/IPDBRawSymbol.h" #include "llvm/DebugInfo/PDB/IPDBSession.h" @@ -35,388 +40,380 @@ #include "llvm/DebugInfo/PDB/PDBSymbolExe.h" #include "llvm/DebugInfo/PDB/PDBSymbolFunc.h" #include "llvm/DebugInfo/PDB/PDBSymbolThunk.h" +#include "llvm/DebugInfo/PDB/Raw/DbiStream.h" +#include "llvm/DebugInfo/PDB/Raw/DbiStreamBuilder.h" +#include "llvm/DebugInfo/PDB/Raw/InfoStream.h" +#include "llvm/DebugInfo/PDB/Raw/InfoStreamBuilder.h" +#include "llvm/DebugInfo/PDB/Raw/PDBFile.h" +#include "llvm/DebugInfo/PDB/Raw/PDBFileBuilder.h" +#include "llvm/DebugInfo/PDB/Raw/RawConstants.h" +#include "llvm/DebugInfo/PDB/Raw/RawError.h" +#include "llvm/DebugInfo/PDB/Raw/RawSession.h" +#include "llvm/Support/COM.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/ConvertUTF.h" +#include "llvm/Support/FileOutputBuffer.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/Format.h" #include "llvm/Support/ManagedStatic.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/PrettyStackTrace.h" #include "llvm/Support/Process.h" -#include "llvm/Support/raw_ostream.h" +#include "llvm/Support/ScopedPrinter.h" #include "llvm/Support/Signals.h" - -#if defined(HAVE_DIA_SDK) -#ifndef NOMINMAX -#define NOMINMAX -#endif -#include <Windows.h> -#endif +#include "llvm/Support/raw_ostream.h" using namespace llvm; +using namespace llvm::codeview; +using namespace llvm::pdb; + +namespace { +// A simple adapter that acts like a ByteStream but holds ownership over +// and underlying FileOutputBuffer. +class FileBufferByteStream : public ByteStream<true> { +public: + FileBufferByteStream(std::unique_ptr<FileOutputBuffer> Buffer) + : ByteStream(MutableArrayRef<uint8_t>(Buffer->getBufferStart(), + Buffer->getBufferEnd())), + FileBuffer(std::move(Buffer)) {} + + Error commit() const override { + if (FileBuffer->commit()) + return llvm::make_error<RawError>(raw_error_code::not_writable); + return Error::success(); + } -namespace opts { +private: + std::unique_ptr<FileOutputBuffer> FileBuffer; +}; +} -enum class PDB_DumpType { ByType, ByObjFile, Both }; +namespace opts { -cl::list<std::string> InputFilenames(cl::Positional, - cl::desc("<input PDB files>"), - cl::OneOrMore); +cl::SubCommand RawSubcommand("raw", "Dump raw structure of the PDB file"); +cl::SubCommand + PrettySubcommand("pretty", + "Dump semantic information about types and symbols"); +cl::SubCommand + YamlToPdbSubcommand("yaml2pdb", + "Generate a PDB file from a YAML description"); +cl::SubCommand + PdbToYamlSubcommand("pdb2yaml", + "Generate a detailed YAML description of a PDB File"); cl::OptionCategory TypeCategory("Symbol Type Options"); cl::OptionCategory FilterCategory("Filtering Options"); cl::OptionCategory OtherOptions("Other Options"); +namespace pretty { +cl::list<std::string> InputFilenames(cl::Positional, + cl::desc("<input PDB files>"), + cl::OneOrMore, cl::sub(PrettySubcommand)); + cl::opt<bool> Compilands("compilands", cl::desc("Display compilands"), - cl::cat(TypeCategory)); + cl::cat(TypeCategory), cl::sub(PrettySubcommand)); cl::opt<bool> Symbols("symbols", cl::desc("Display symbols for each compiland"), - cl::cat(TypeCategory)); + cl::cat(TypeCategory), cl::sub(PrettySubcommand)); cl::opt<bool> Globals("globals", cl::desc("Dump global symbols"), - cl::cat(TypeCategory)); + cl::cat(TypeCategory), cl::sub(PrettySubcommand)); cl::opt<bool> Externals("externals", cl::desc("Dump external symbols"), - cl::cat(TypeCategory)); -cl::opt<bool> Types("types", cl::desc("Display types"), cl::cat(TypeCategory)); + cl::cat(TypeCategory), cl::sub(PrettySubcommand)); +cl::opt<bool> Types("types", cl::desc("Display types"), cl::cat(TypeCategory), + cl::sub(PrettySubcommand)); +cl::opt<bool> Lines("lines", cl::desc("Line tables"), cl::cat(TypeCategory), + cl::sub(PrettySubcommand)); cl::opt<bool> All("all", cl::desc("Implies all other options in 'Symbol Types' category"), - cl::cat(TypeCategory)); + cl::cat(TypeCategory), cl::sub(PrettySubcommand)); cl::opt<uint64_t> LoadAddress( "load-address", cl::desc("Assume the module is loaded at the specified address"), - cl::cat(OtherOptions)); - -cl::opt<bool> DumpHeaders("dump-headers", cl::desc("dump PDB headers"), - cl::cat(OtherOptions)); -cl::opt<bool> DumpStreamSizes("dump-stream-sizes", - cl::desc("dump PDB stream sizes"), - cl::cat(OtherOptions)); -cl::opt<bool> DumpStreamBlocks("dump-stream-blocks", - cl::desc("dump PDB stream blocks"), - cl::cat(OtherOptions)); -cl::opt<std::string> DumpStreamData("dump-stream", cl::desc("dump stream data"), - cl::cat(OtherOptions)); - -cl::list<std::string> - ExcludeTypes("exclude-types", - cl::desc("Exclude types by regular expression"), - cl::ZeroOrMore, cl::cat(FilterCategory)); -cl::list<std::string> - ExcludeSymbols("exclude-symbols", - cl::desc("Exclude symbols by regular expression"), - cl::ZeroOrMore, cl::cat(FilterCategory)); -cl::list<std::string> - ExcludeCompilands("exclude-compilands", - cl::desc("Exclude compilands by regular expression"), - cl::ZeroOrMore, cl::cat(FilterCategory)); + cl::cat(OtherOptions), cl::sub(PrettySubcommand)); +cl::list<std::string> ExcludeTypes( + "exclude-types", cl::desc("Exclude types by regular expression"), + cl::ZeroOrMore, cl::cat(FilterCategory), cl::sub(PrettySubcommand)); +cl::list<std::string> ExcludeSymbols( + "exclude-symbols", cl::desc("Exclude symbols by regular expression"), + cl::ZeroOrMore, cl::cat(FilterCategory), cl::sub(PrettySubcommand)); +cl::list<std::string> ExcludeCompilands( + "exclude-compilands", cl::desc("Exclude compilands by regular expression"), + cl::ZeroOrMore, cl::cat(FilterCategory), cl::sub(PrettySubcommand)); cl::list<std::string> IncludeTypes( "include-types", cl::desc("Include only types which match a regular expression"), - cl::ZeroOrMore, cl::cat(FilterCategory)); + cl::ZeroOrMore, cl::cat(FilterCategory), cl::sub(PrettySubcommand)); cl::list<std::string> IncludeSymbols( "include-symbols", cl::desc("Include only symbols which match a regular expression"), - cl::ZeroOrMore, cl::cat(FilterCategory)); + cl::ZeroOrMore, cl::cat(FilterCategory), cl::sub(PrettySubcommand)); cl::list<std::string> IncludeCompilands( "include-compilands", cl::desc("Include only compilands those which match a regular expression"), - cl::ZeroOrMore, cl::cat(FilterCategory)); + cl::ZeroOrMore, cl::cat(FilterCategory), cl::sub(PrettySubcommand)); cl::opt<bool> ExcludeCompilerGenerated( "no-compiler-generated", cl::desc("Don't show compiler generated types and symbols"), - cl::cat(FilterCategory)); + cl::cat(FilterCategory), cl::sub(PrettySubcommand)); cl::opt<bool> ExcludeSystemLibraries("no-system-libs", cl::desc("Don't show symbols from system libraries"), - cl::cat(FilterCategory)); + cl::cat(FilterCategory), cl::sub(PrettySubcommand)); cl::opt<bool> NoClassDefs("no-class-definitions", cl::desc("Don't display full class definitions"), - cl::cat(FilterCategory)); + cl::cat(FilterCategory), cl::sub(PrettySubcommand)); cl::opt<bool> NoEnumDefs("no-enum-definitions", cl::desc("Don't display full enum definitions"), - cl::cat(FilterCategory)); + cl::cat(FilterCategory), cl::sub(PrettySubcommand)); } +namespace raw { -static void reportError(StringRef Input, StringRef Message) { - if (Input == "-") - Input = "<stdin>"; - errs() << Input << ": " << Message << "\n"; - errs().flush(); - exit(1); -} +cl::OptionCategory MsfOptions("MSF Container Options"); +cl::OptionCategory TypeOptions("Type Record Options"); +cl::OptionCategory FileOptions("Module & File Options"); +cl::OptionCategory SymbolOptions("Symbol Options"); +cl::OptionCategory MiscOptions("Miscellaneous Options"); -static void reportError(StringRef Input, std::error_code EC) { - reportError(Input, EC.message()); -} +// MSF OPTIONS +cl::opt<bool> DumpHeaders("headers", cl::desc("dump PDB headers"), + cl::cat(MsfOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpStreamBlocks("stream-blocks", + cl::desc("dump PDB stream blocks"), + cl::cat(MsfOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpStreamSummary("stream-summary", + cl::desc("dump summary of the PDB streams"), + cl::cat(MsfOptions), cl::sub(RawSubcommand)); -static std::error_code checkOffset(MemoryBufferRef M, uintptr_t Addr, - const uint64_t Size) { - if (Addr + Size < Addr || Addr + Size < Size || - Addr + Size > uintptr_t(M.getBufferEnd()) || - Addr < uintptr_t(M.getBufferStart())) { - return std::make_error_code(std::errc::bad_address); - } - return std::error_code(); -} +// TYPE OPTIONS +cl::opt<bool> + DumpTpiRecords("tpi-records", + cl::desc("dump CodeView type records from TPI stream"), + cl::cat(TypeOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpTpiRecordBytes( + "tpi-record-bytes", + cl::desc("dump CodeView type record raw bytes from TPI stream"), + cl::cat(TypeOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpTpiHash("tpi-hash", cl::desc("dump CodeView TPI hash stream"), + cl::cat(TypeOptions), cl::sub(RawSubcommand)); +cl::opt<bool> + DumpIpiRecords("ipi-records", + cl::desc("dump CodeView type records from IPI stream"), + cl::cat(TypeOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpIpiRecordBytes( + "ipi-record-bytes", + cl::desc("dump CodeView type record raw bytes from IPI stream"), + cl::cat(TypeOptions), cl::sub(RawSubcommand)); + +// MODULE & FILE OPTIONS +cl::opt<bool> DumpModules("modules", cl::desc("dump compiland information"), + cl::cat(FileOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpModuleFiles("module-files", cl::desc("dump file information"), + cl::cat(FileOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpLineInfo("line-info", + cl::desc("dump file and line information"), + cl::cat(FileOptions), cl::sub(RawSubcommand)); + +// SYMBOL OPTIONS +cl::opt<bool> DumpModuleSyms("module-syms", cl::desc("dump module symbols"), + cl::cat(SymbolOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpPublics("publics", cl::desc("dump Publics stream data"), + cl::cat(SymbolOptions), cl::sub(RawSubcommand)); +cl::opt<bool> + DumpSymRecordBytes("sym-record-bytes", + cl::desc("dump CodeView symbol record raw bytes"), + cl::cat(SymbolOptions), cl::sub(RawSubcommand)); + +// MISCELLANEOUS OPTIONS +cl::opt<bool> DumpSectionContribs("section-contribs", + cl::desc("dump section contributions"), + cl::cat(MiscOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpSectionMap("section-map", cl::desc("dump section map"), + cl::cat(MiscOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpSectionHeaders("section-headers", + cl::desc("dump section headers"), + cl::cat(MiscOptions), cl::sub(RawSubcommand)); +cl::opt<bool> DumpFpo("fpo", cl::desc("dump FPO records"), cl::cat(MiscOptions), + cl::sub(RawSubcommand)); + +cl::opt<std::string> DumpStreamDataIdx("stream", cl::desc("dump stream data"), + cl::cat(MiscOptions), + cl::sub(RawSubcommand)); +cl::opt<std::string> DumpStreamDataName("stream-name", + cl::desc("dump stream data"), + cl::cat(MiscOptions), + cl::sub(RawSubcommand)); + +cl::opt<bool> RawAll("all", cl::desc("Implies most other options."), + cl::cat(MiscOptions), cl::sub(RawSubcommand)); -template <typename T> -static std::error_code checkOffset(MemoryBufferRef M, ArrayRef<T> AR) { - return checkOffset(M, uintptr_t(AR.data()), (uint64_t)AR.size() * sizeof(T)); +cl::list<std::string> InputFilenames(cl::Positional, + cl::desc("<input PDB files>"), + cl::OneOrMore, cl::sub(RawSubcommand)); } -static std::error_code checkOffset(MemoryBufferRef M, StringRef SR) { - return checkOffset(M, uintptr_t(SR.data()), SR.size()); -} +namespace yaml2pdb { +cl::opt<std::string> + YamlPdbOutputFile("pdb", cl::desc("the name of the PDB file to write"), + cl::sub(YamlToPdbSubcommand)); -// Sets Obj unless any bytes in [addr, addr + size) fall outsize of m. -// Returns unexpected_eof if error. -template <typename T> -static std::error_code getObject(const T *&Obj, MemoryBufferRef M, - const void *Ptr, - const uint64_t Size = sizeof(T)) { - uintptr_t Addr = uintptr_t(Ptr); - if (std::error_code EC = checkOffset(M, Addr, Size)) - return EC; - Obj = reinterpret_cast<const T *>(Addr); - return std::error_code(); +cl::list<std::string> InputFilename(cl::Positional, + cl::desc("<input YAML file>"), cl::Required, + cl::sub(YamlToPdbSubcommand)); } -static uint64_t bytesToBlocks(uint64_t NumBytes, uint64_t BlockSize) { - return RoundUpToAlignment(NumBytes, BlockSize) / BlockSize; +namespace pdb2yaml { +cl::opt<bool> + NoFileHeaders("no-file-headers", + cl::desc("Do not dump MSF file headers (you will not be able " + "to generate a fresh PDB from the resulting YAML)"), + cl::sub(PdbToYamlSubcommand), cl::init(false)); + +cl::opt<bool> StreamMetadata( + "stream-metadata", + cl::desc("Dump the number of streams and each stream's size"), + cl::sub(PdbToYamlSubcommand), cl::init(false)); +cl::opt<bool> StreamDirectory( + "stream-directory", + cl::desc("Dump each stream's block map (implies -stream-metadata)"), + cl::sub(PdbToYamlSubcommand), cl::init(false)); +cl::opt<bool> PdbStream("pdb-stream", + cl::desc("Dump the PDB Stream (Stream 1)"), + cl::sub(PdbToYamlSubcommand), cl::init(false)); +cl::opt<bool> DbiStream("dbi-stream", + cl::desc("Dump the DBI Stream (Stream 2)"), + cl::sub(PdbToYamlSubcommand), cl::init(false)); + +cl::list<std::string> InputFilename(cl::Positional, + cl::desc("<input PDB file>"), cl::Required, + cl::sub(PdbToYamlSubcommand)); } - -static uint64_t blockToOffset(uint64_t BlockNumber, uint64_t BlockSize) { - return BlockNumber * BlockSize; } -static void dumpStructure(MemoryBufferRef M) { - const PDB::SuperBlock *SB; - if (auto EC = getObject(SB, M, M.getBufferStart())) - reportError(M.getBufferIdentifier(), EC); - - if (opts::DumpHeaders) { - outs() << "BlockSize: " << SB->BlockSize << '\n'; - outs() << "Unknown0: " << SB->Unknown0 << '\n'; - outs() << "NumBlocks: " << SB->NumBlocks << '\n'; - outs() << "NumDirectoryBytes: " << SB->NumDirectoryBytes << '\n'; - outs() << "Unknown1: " << SB->Unknown1 << '\n'; - outs() << "BlockMapAddr: " << SB->BlockMapAddr << '\n'; - } +static ExitOnError ExitOnErr; - // We don't support blocksizes which aren't a multiple of four bytes. - if (SB->BlockSize % sizeof(support::ulittle32_t) != 0) - reportError(M.getBufferIdentifier(), - std::make_error_code(std::errc::illegal_byte_sequence)); - - // We don't support directories whose sizes aren't a multiple of four bytes. - if (SB->NumDirectoryBytes % sizeof(support::ulittle32_t) != 0) - reportError(M.getBufferIdentifier(), - std::make_error_code(std::errc::illegal_byte_sequence)); - - // The number of blocks which comprise the directory is a simple function of - // the number of bytes it contains. - uint64_t NumDirectoryBlocks = - bytesToBlocks(SB->NumDirectoryBytes, SB->BlockSize); - if (opts::DumpHeaders) - outs() << "NumDirectoryBlocks: " << NumDirectoryBlocks << '\n'; - - // The block map, as we understand it, is a block which consists of a list of - // block numbers. - // It is unclear what would happen if the number of blocks couldn't fit on a - // single block. - if (NumDirectoryBlocks > SB->BlockSize / sizeof(support::ulittle32_t)) - reportError(M.getBufferIdentifier(), - std::make_error_code(std::errc::illegal_byte_sequence)); - - - uint64_t BlockMapOffset = (uint64_t)SB->BlockMapAddr * SB->BlockSize; - if (opts::DumpHeaders) - outs() << "BlockMapOffset: " << BlockMapOffset << '\n'; - - // The directory is not contiguous. Instead, the block map contains a - // contiguous list of block numbers whose contents, when concatenated in - // order, make up the directory. - auto DirectoryBlocks = - makeArrayRef(reinterpret_cast<const support::ulittle32_t *>( - M.getBufferStart() + BlockMapOffset), - NumDirectoryBlocks); - if (auto EC = checkOffset(M, DirectoryBlocks)) - reportError(M.getBufferIdentifier(), EC); - - if (opts::DumpHeaders) { - outs() << "DirectoryBlocks: ["; - for (const support::ulittle32_t &DirectoryBlockAddr : DirectoryBlocks) { - if (&DirectoryBlockAddr != &DirectoryBlocks.front()) - outs() << ", "; - outs() << DirectoryBlockAddr; - } - outs() << "]\n"; +static void yamlToPdb(StringRef Path) { + ErrorOr<std::unique_ptr<MemoryBuffer>> ErrorOrBuffer = + MemoryBuffer::getFileOrSTDIN(Path, /*FileSize=*/-1, + /*RequiresNullTerminator=*/false); + + if (ErrorOrBuffer.getError()) { + ExitOnErr(make_error<GenericError>(generic_error_code::invalid_path, Path)); } - bool SeenNumStreams = false; - uint32_t NumStreams = 0; - std::vector<uint32_t> StreamSizes; - DenseMap<uint32_t, std::vector<uint32_t>> StreamMap; - uint32_t StreamIdx = 0; - uint64_t DirectoryBytesRead = 0; - // The structure of the directory is as follows: - // struct PDBDirectory { - // uint32_t NumStreams; - // uint32_t StreamSizes[NumStreams]; - // uint32_t StreamMap[NumStreams][]; - // }; - // - // Empty streams don't consume entries in the StreamMap. - for (uint32_t DirectoryBlockAddr : DirectoryBlocks) { - uint64_t DirectoryBlockOffset = - blockToOffset(DirectoryBlockAddr, SB->BlockSize); - auto DirectoryBlock = - makeArrayRef(reinterpret_cast<const support::ulittle32_t *>( - M.getBufferStart() + DirectoryBlockOffset), - SB->BlockSize / sizeof(support::ulittle32_t)); - if (auto EC = checkOffset(M, DirectoryBlock)) - reportError(M.getBufferIdentifier(), EC); - - // We read data out of the directory four bytes at a time. Depending on - // where we are in the directory, the contents may be: the number of streams - // in the directory, a stream's size, or a block in the stream map. - for (uint32_t Data : DirectoryBlock) { - // Don't read beyond the end of the directory. - if (DirectoryBytesRead == SB->NumDirectoryBytes) - break; - - DirectoryBytesRead += sizeof(Data); - - // This data must be the number of streams if we haven't seen it yet. - if (!SeenNumStreams) { - NumStreams = Data; - SeenNumStreams = true; - continue; - } - // This data must be a stream size if we have not seen them all yet. - if (StreamSizes.size() < NumStreams) { - // It seems like some streams have their set to -1 when their contents - // are not present. Treat them like empty streams for now. - if (Data == UINT32_MAX) - StreamSizes.push_back(0); - else - StreamSizes.push_back(Data); - continue; - } + std::unique_ptr<MemoryBuffer> &Buffer = ErrorOrBuffer.get(); + + llvm::yaml::Input In(Buffer->getBuffer()); + pdb::yaml::PdbObject YamlObj; + In >> YamlObj; + if (!YamlObj.Headers.hasValue()) + ExitOnErr(make_error<GenericError>(generic_error_code::unspecified, + "Yaml does not contain MSF headers")); + + auto OutFileOrError = FileOutputBuffer::create( + opts::yaml2pdb::YamlPdbOutputFile, YamlObj.Headers->FileSize); + if (OutFileOrError.getError()) + ExitOnErr(make_error<GenericError>(generic_error_code::invalid_path, + opts::yaml2pdb::YamlPdbOutputFile)); + + auto FileByteStream = + llvm::make_unique<FileBufferByteStream>(std::move(*OutFileOrError)); + PDBFileBuilder Builder(std::move(FileByteStream)); + + ExitOnErr(Builder.initialize(YamlObj.Headers->SuperBlock)); + ExitOnErr(Builder.getMsfBuilder().setDirectoryBlocksHint( + YamlObj.Headers->DirectoryBlocks)); + if (!YamlObj.StreamSizes.hasValue()) { + ExitOnErr(make_error<GenericError>( + generic_error_code::unspecified, + "Cannot generate a PDB when stream sizes are not known")); + } - // This data must be a stream block number if we have seen all of the - // stream sizes. - std::vector<uint32_t> *StreamBlocks = nullptr; - // Figure out which stream this block number belongs to. - while (StreamIdx < NumStreams) { - uint64_t NumExpectedStreamBlocks = - bytesToBlocks(StreamSizes[StreamIdx], SB->BlockSize); - StreamBlocks = &StreamMap[StreamIdx]; - if (NumExpectedStreamBlocks > StreamBlocks->size()) - break; - ++StreamIdx; - } - // It seems this block doesn't belong to any stream? The stream is either - // corrupt or something more mysterious is going on. - if (StreamIdx == NumStreams) - reportError(M.getBufferIdentifier(), - std::make_error_code(std::errc::illegal_byte_sequence)); + if (YamlObj.StreamMap.hasValue()) { + if (YamlObj.StreamMap->size() != YamlObj.StreamSizes->size()) { + ExitOnErr(make_error<GenericError>(generic_error_code::unspecified, + "YAML specifies different number of " + "streams in stream sizes and stream " + "map")); + } - StreamBlocks->push_back(Data); + auto &Sizes = *YamlObj.StreamSizes; + auto &Map = *YamlObj.StreamMap; + for (uint32_t I = 0; I < Sizes.size(); ++I) { + uint32_t Size = Sizes[I]; + std::vector<uint32_t> Blocks; + for (auto E : Map[I].Blocks) + Blocks.push_back(E); + ExitOnErr(Builder.getMsfBuilder().addStream(Size, Blocks)); + } + } else { + auto &Sizes = *YamlObj.StreamSizes; + for (auto S : Sizes) { + ExitOnErr(Builder.getMsfBuilder().addStream(S)); } } - // We should have read exactly SB->NumDirectoryBytes bytes. - assert(DirectoryBytesRead == SB->NumDirectoryBytes); - - if (opts::DumpHeaders) - outs() << "NumStreams: " << NumStreams << '\n'; - if (opts::DumpStreamSizes) - for (uint32_t StreamIdx = 0; StreamIdx < NumStreams; ++StreamIdx) - outs() << "StreamSizes[" << StreamIdx << "]: " << StreamSizes[StreamIdx] - << '\n'; - - if (opts::DumpStreamBlocks) { - for (uint32_t StreamIdx = 0; StreamIdx < NumStreams; ++StreamIdx) { - outs() << "StreamBlocks[" << StreamIdx << "]: ["; - std::vector<uint32_t> &StreamBlocks = StreamMap[StreamIdx]; - for (uint32_t &StreamBlock : StreamBlocks) { - if (&StreamBlock != &StreamBlocks.front()) - outs() << ", "; - outs() << StreamBlock; - } - outs() << "]\n"; - } + if (YamlObj.PdbStream.hasValue()) { + auto &InfoBuilder = Builder.getInfoBuilder(); + InfoBuilder.setAge(YamlObj.PdbStream->Age); + InfoBuilder.setGuid(YamlObj.PdbStream->Guid); + InfoBuilder.setSignature(YamlObj.PdbStream->Signature); + InfoBuilder.setVersion(YamlObj.PdbStream->Version); + for (auto &NM : YamlObj.PdbStream->NamedStreams) + InfoBuilder.getNamedStreamsBuilder().addMapping(NM.StreamName, + NM.StreamNumber); } - StringRef DumpStreamStr = opts::DumpStreamData; - uint32_t DumpStreamNum; - if (!DumpStreamStr.getAsInteger(/*Radix=*/0U, DumpStreamNum) && - DumpStreamNum < NumStreams) { - uint32_t StreamBytesRead = 0; - uint32_t StreamSize = StreamSizes[DumpStreamNum]; - std::vector<uint32_t> &StreamBlocks = StreamMap[DumpStreamNum]; - for (uint32_t &StreamBlockAddr : StreamBlocks) { - uint64_t StreamBlockOffset = blockToOffset(StreamBlockAddr, SB->BlockSize); - uint32_t BytesLeftToReadInStream = StreamSize - StreamBytesRead; - if (BytesLeftToReadInStream == 0) - break; - - uint32_t BytesToReadInBlock = std::min( - BytesLeftToReadInStream, static_cast<uint32_t>(SB->BlockSize)); - auto StreamBlockData = - StringRef(M.getBufferStart() + StreamBlockOffset, BytesToReadInBlock); - if (auto EC = checkOffset(M, StreamBlockData)) - reportError(M.getBufferIdentifier(), EC); - - outs() << StreamBlockData; - StreamBytesRead += StreamBlockData.size(); - } + if (YamlObj.DbiStream.hasValue()) { + auto &DbiBuilder = Builder.getDbiBuilder(); + DbiBuilder.setAge(YamlObj.DbiStream->Age); + DbiBuilder.setBuildNumber(YamlObj.DbiStream->BuildNumber); + DbiBuilder.setFlags(YamlObj.DbiStream->Flags); + DbiBuilder.setMachineType(YamlObj.DbiStream->MachineType); + DbiBuilder.setPdbDllRbld(YamlObj.DbiStream->PdbDllRbld); + DbiBuilder.setPdbDllVersion(YamlObj.DbiStream->PdbDllVersion); + DbiBuilder.setVersionHeader(YamlObj.DbiStream->VerHeader); } + + auto Pdb = Builder.build(); + ExitOnErr(Pdb.takeError()); + + auto &PdbFile = *Pdb; + ExitOnErr(PdbFile->commit()); } -static void dumpInput(StringRef Path) { - if (opts::DumpHeaders || !opts::DumpStreamData.empty()) { - ErrorOr<std::unique_ptr<MemoryBuffer>> ErrorOrBuffer = - MemoryBuffer::getFileOrSTDIN(Path, /*FileSize=*/-1, - /*RequiresNullTerminator=*/false); +static void pdb2Yaml(StringRef Path) { + std::unique_ptr<IPDBSession> Session; + ExitOnErr(loadDataForPDB(PDB_ReaderType::Raw, Path, Session)); + + RawSession *RS = static_cast<RawSession *>(Session.get()); + PDBFile &File = RS->getPDBFile(); + auto O = llvm::make_unique<YAMLOutputStyle>(File); + O = llvm::make_unique<YAMLOutputStyle>(File); - if (std::error_code EC = ErrorOrBuffer.getError()) - reportError(Path, EC); + ExitOnErr(O->dump()); +} - std::unique_ptr<MemoryBuffer> &Buffer = ErrorOrBuffer.get(); +static void dumpRaw(StringRef Path) { + std::unique_ptr<IPDBSession> Session; + ExitOnErr(loadDataForPDB(PDB_ReaderType::Raw, Path, Session)); - dumpStructure(Buffer->getMemBufferRef()); + RawSession *RS = static_cast<RawSession *>(Session.get()); + PDBFile &File = RS->getPDBFile(); + auto O = llvm::make_unique<LLVMOutputStyle>(File); - outs().flush(); - return; - } + ExitOnErr(O->dump()); +} +static void dumpPretty(StringRef Path) { std::unique_ptr<IPDBSession> Session; - PDB_ErrorCode Error = loadDataForPDB(PDB_ReaderType::DIA, Path, Session); - switch (Error) { - case PDB_ErrorCode::Success: - break; - case PDB_ErrorCode::NoPdbImpl: - outs() << "Reading PDBs is not supported on this platform.\n"; - return; - case PDB_ErrorCode::InvalidPath: - outs() << "Unable to load PDB at '" << Path - << "'. Check that the file exists and is readable.\n"; - return; - case PDB_ErrorCode::InvalidFileFormat: - outs() << "Unable to load PDB at '" << Path - << "'. The file has an unrecognized format.\n"; - return; - default: - outs() << "Unable to load PDB at '" << Path - << "'. An unknown error occured.\n"; - return; - } - if (opts::LoadAddress) - Session->setLoadAddress(opts::LoadAddress); + + ExitOnErr(loadDataForPDB(PDB_ReaderType::DIA, Path, Session)); + + if (opts::pretty::LoadAddress) + Session->setLoadAddress(opts::pretty::LoadAddress); LinePrinter Printer(2, outs()); @@ -453,19 +450,22 @@ static void dumpInput(StringRef Path) { outs() << "HasPrivateSymbols "; Printer.Unindent(); - if (opts::Compilands) { + if (opts::pretty::Compilands) { Printer.NewLine(); WithColor(Printer, PDB_ColorItem::SectionHeader).get() << "---COMPILANDS---"; Printer.Indent(); auto Compilands = GlobalScope->findAllChildren<PDBSymbolCompiland>(); CompilandDumper Dumper(Printer); + CompilandDumpFlags options = CompilandDumper::Flags::None; + if (opts::pretty::Lines) + options = options | CompilandDumper::Flags::Lines; while (auto Compiland = Compilands->getNext()) - Dumper.start(*Compiland, false); + Dumper.start(*Compiland, options); Printer.Unindent(); } - if (opts::Types) { + if (opts::pretty::Types) { Printer.NewLine(); WithColor(Printer, PDB_ColorItem::SectionHeader).get() << "---TYPES---"; Printer.Indent(); @@ -474,7 +474,7 @@ static void dumpInput(StringRef Path) { Printer.Unindent(); } - if (opts::Symbols) { + if (opts::pretty::Symbols) { Printer.NewLine(); WithColor(Printer, PDB_ColorItem::SectionHeader).get() << "---SYMBOLS---"; Printer.Indent(); @@ -485,7 +485,7 @@ static void dumpInput(StringRef Path) { Printer.Unindent(); } - if (opts::Globals) { + if (opts::pretty::Globals) { Printer.NewLine(); WithColor(Printer, PDB_ColorItem::SectionHeader).get() << "---GLOBALS---"; Printer.Indent(); @@ -511,59 +511,98 @@ static void dumpInput(StringRef Path) { } Printer.Unindent(); } - if (opts::Externals) { + if (opts::pretty::Externals) { Printer.NewLine(); WithColor(Printer, PDB_ColorItem::SectionHeader).get() << "---EXTERNALS---"; Printer.Indent(); ExternalSymbolDumper Dumper(Printer); Dumper.start(*GlobalScope); } + if (opts::pretty::Lines) { + Printer.NewLine(); + } outs().flush(); } int main(int argc_, const char *argv_[]) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv_[0]); PrettyStackTraceProgram X(argc_, argv_); + ExitOnErr.setBanner("llvm-pdbdump: "); + SmallVector<const char *, 256> argv; SpecificBumpPtrAllocator<char> ArgAllocator; - std::error_code EC = sys::Process::GetArgumentVector( - argv, makeArrayRef(argv_, argc_), ArgAllocator); - if (EC) { - errs() << "error: couldn't get arguments: " << EC.message() << '\n'; - return 1; - } + ExitOnErr(errorCodeToError(sys::Process::GetArgumentVector( + argv, makeArrayRef(argv_, argc_), ArgAllocator))); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. cl::ParseCommandLineOptions(argv.size(), argv.data(), "LLVM PDB Dumper\n"); - if (opts::All) { - opts::Compilands = true; - opts::Symbols = true; - opts::Globals = true; - opts::Types = true; - opts::Externals = true; - } - if (opts::ExcludeCompilerGenerated) { - opts::ExcludeTypes.push_back("__vc_attributes"); - opts::ExcludeCompilands.push_back("* Linker *"); - } - if (opts::ExcludeSystemLibraries) { - opts::ExcludeCompilands.push_back( - "f:\\binaries\\Intermediate\\vctools\\crt_bld"); - } -#if defined(HAVE_DIA_SDK) - CoInitializeEx(nullptr, COINIT_MULTITHREADED); -#endif + // These options are shared by two subcommands. + if ((opts::PdbToYamlSubcommand || opts::RawSubcommand) && opts::raw::RawAll) { + opts::raw::DumpHeaders = true; + opts::raw::DumpModules = true; + opts::raw::DumpModuleFiles = true; + opts::raw::DumpModuleSyms = true; + opts::raw::DumpPublics = true; + opts::raw::DumpSectionHeaders = true; + opts::raw::DumpStreamSummary = true; + opts::raw::DumpStreamBlocks = true; + opts::raw::DumpTpiRecords = true; + opts::raw::DumpTpiHash = true; + opts::raw::DumpIpiRecords = true; + opts::raw::DumpSectionMap = true; + opts::raw::DumpSectionContribs = true; + opts::raw::DumpLineInfo = true; + opts::raw::DumpFpo = true; + } - std::for_each(opts::InputFilenames.begin(), opts::InputFilenames.end(), - dumpInput); + llvm::sys::InitializeCOMRAII COM(llvm::sys::COMThreadingMode::MultiThreaded); + + if (opts::PdbToYamlSubcommand) { + pdb2Yaml(opts::pdb2yaml::InputFilename.front()); + } else if (opts::YamlToPdbSubcommand) { + yamlToPdb(opts::yaml2pdb::InputFilename.front()); + } else if (opts::PrettySubcommand) { + if (opts::pretty::Lines) + opts::pretty::Compilands = true; + + if (opts::pretty::All) { + opts::pretty::Compilands = true; + opts::pretty::Symbols = true; + opts::pretty::Globals = true; + opts::pretty::Types = true; + opts::pretty::Externals = true; + opts::pretty::Lines = true; + } -#if defined(HAVE_DIA_SDK) - CoUninitialize(); -#endif + // When adding filters for excluded compilands and types, we need to + // remember + // that these are regexes. So special characters such as * and \ need to be + // escaped in the regex. In the case of a literal \, this means it needs to + // be escaped again in the C++. So matching a single \ in the input + // requires + // 4 \es in the C++. + if (opts::pretty::ExcludeCompilerGenerated) { + opts::pretty::ExcludeTypes.push_back("__vc_attributes"); + opts::pretty::ExcludeCompilands.push_back("\\* Linker \\*"); + } + if (opts::pretty::ExcludeSystemLibraries) { + opts::pretty::ExcludeCompilands.push_back( + "f:\\\\binaries\\\\Intermediate\\\\vctools\\\\crt_bld"); + opts::pretty::ExcludeCompilands.push_back("f:\\\\dd\\\\vctools\\\\crt"); + opts::pretty::ExcludeCompilands.push_back( + "d:\\\\th.obj.x86fre\\\\minkernel"); + } + std::for_each(opts::pretty::InputFilenames.begin(), + opts::pretty::InputFilenames.end(), dumpPretty); + } else if (opts::RawSubcommand) { + std::for_each(opts::raw::InputFilenames.begin(), + opts::raw::InputFilenames.end(), dumpRaw); + } + outs().flush(); return 0; } diff --git a/tools/llvm-pdbdump/llvm-pdbdump.h b/tools/llvm-pdbdump/llvm-pdbdump.h index cb5bec64dec81..0a66515f4a015 100644 --- a/tools/llvm-pdbdump/llvm-pdbdump.h +++ b/tools/llvm-pdbdump/llvm-pdbdump.h @@ -14,12 +14,13 @@ #include "llvm/Support/raw_ostream.h" namespace opts { + +namespace pretty { extern llvm::cl::opt<bool> Compilands; extern llvm::cl::opt<bool> Symbols; extern llvm::cl::opt<bool> Globals; extern llvm::cl::opt<bool> Types; extern llvm::cl::opt<bool> All; - extern llvm::cl::opt<bool> ExcludeCompilerGenerated; extern llvm::cl::opt<bool> NoClassDefs; @@ -32,4 +33,37 @@ extern llvm::cl::list<std::string> IncludeSymbols; extern llvm::cl::list<std::string> IncludeCompilands; } -#endif
\ No newline at end of file +namespace raw { +extern llvm::cl::opt<bool> DumpHeaders; +extern llvm::cl::opt<bool> DumpStreamBlocks; +extern llvm::cl::opt<bool> DumpStreamSummary; +extern llvm::cl::opt<bool> DumpTpiHash; +extern llvm::cl::opt<bool> DumpTpiRecordBytes; +extern llvm::cl::opt<bool> DumpTpiRecords; +extern llvm::cl::opt<bool> DumpIpiRecords; +extern llvm::cl::opt<bool> DumpIpiRecordBytes; +extern llvm::cl::opt<std::string> DumpStreamDataIdx; +extern llvm::cl::opt<std::string> DumpStreamDataName; +extern llvm::cl::opt<bool> DumpModules; +extern llvm::cl::opt<bool> DumpModuleFiles; +extern llvm::cl::opt<bool> DumpModuleSyms; +extern llvm::cl::opt<bool> DumpPublics; +extern llvm::cl::opt<bool> DumpSectionContribs; +extern llvm::cl::opt<bool> DumpLineInfo; +extern llvm::cl::opt<bool> DumpSectionMap; +extern llvm::cl::opt<bool> DumpSymRecordBytes; +extern llvm::cl::opt<bool> DumpSectionHeaders; +extern llvm::cl::opt<bool> DumpFpo; +} + +namespace pdb2yaml { +extern llvm::cl::opt<bool> NoFileHeaders; +extern llvm::cl::opt<bool> StreamMetadata; +extern llvm::cl::opt<bool> StreamDirectory; +extern llvm::cl::opt<bool> PdbStream; +extern llvm::cl::opt<bool> DbiStream; +extern llvm::cl::list<std::string> InputFilename; +} +} + +#endif diff --git a/tools/llvm-profdata/Makefile b/tools/llvm-profdata/Makefile deleted file mode 100644 index 6f8c16265f37a..0000000000000 --- a/tools/llvm-profdata/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-profdata/Makefile ------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-profdata -LINK_COMPONENTS := profiledata support - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-profdata/llvm-profdata.cpp b/tools/llvm-profdata/llvm-profdata.cpp index dc6cd0a5784d6..8e4b4c3d4ed55 100644 --- a/tools/llvm-profdata/llvm-profdata.cpp +++ b/tools/llvm-profdata/llvm-profdata.cpp @@ -17,6 +17,7 @@ #include "llvm/IR/LLVMContext.h" #include "llvm/ProfileData/InstrProfReader.h" #include "llvm/ProfileData/InstrProfWriter.h" +#include "llvm/ProfileData/ProfileCommon.h" #include "llvm/ProfileData/SampleProfReader.h" #include "llvm/ProfileData/SampleProfWriter.h" #include "llvm/Support/CommandLine.h" @@ -30,7 +31,6 @@ #include "llvm/Support/Signals.h" #include "llvm/Support/raw_ostream.h" #include <algorithm> -#include <tuple> using namespace llvm; @@ -47,38 +47,50 @@ static void exitWithError(const Twine &Message, StringRef Whence = "", ::exit(1); } -static void exitWithErrorCode(const std::error_code &Error, - StringRef Whence = "") { - if (Error.category() == instrprof_category()) { - instrprof_error instrError = static_cast<instrprof_error>(Error.value()); - if (instrError == instrprof_error::unrecognized_format) { - // Hint for common error of forgetting -sample for sample profiles. - exitWithError(Error.message(), Whence, - "Perhaps you forgot to use the -sample option?"); - } +static void exitWithError(Error E, StringRef Whence = "") { + if (E.isA<InstrProfError>()) { + handleAllErrors(std::move(E), [&](const InstrProfError &IPE) { + instrprof_error instrError = IPE.get(); + StringRef Hint = ""; + if (instrError == instrprof_error::unrecognized_format) { + // Hint for common error of forgetting -sample for sample profiles. + Hint = "Perhaps you forgot to use the -sample option?"; + } + exitWithError(IPE.message(), Whence, Hint); + }); } - exitWithError(Error.message(), Whence); + + exitWithError(toString(std::move(E)), Whence); +} + +static void exitWithErrorCode(std::error_code EC, StringRef Whence = "") { + exitWithError(EC.message(), Whence); } namespace { enum ProfileKinds { instr, sample }; } -static void handleMergeWriterError(std::error_code &Error, - StringRef WhenceFile = "", +static void handleMergeWriterError(Error E, StringRef WhenceFile = "", StringRef WhenceFunction = "", bool ShowHint = true) { if (!WhenceFile.empty()) errs() << WhenceFile << ": "; if (!WhenceFunction.empty()) errs() << WhenceFunction << ": "; - errs() << Error.message() << "\n"; + + auto IPE = instrprof_error::success; + E = handleErrors(std::move(E), + [&IPE](std::unique_ptr<InstrProfError> E) -> Error { + IPE = E->get(); + return Error(std::move(E)); + }); + errs() << toString(std::move(E)) << "\n"; if (ShowHint) { StringRef Hint = ""; - if (Error.category() == instrprof_category()) { - instrprof_error instrError = static_cast<instrprof_error>(Error.value()); - switch (instrError) { + if (IPE != instrprof_error::success) { + switch (IPE) { case instrprof_error::hash_mismatch: case instrprof_error::count_mismatch: case instrprof_error::value_site_count_mismatch: @@ -107,7 +119,7 @@ typedef SmallVector<WeightedFile, 5> WeightedFileVector; static void mergeInstrProfile(const WeightedFileVector &Inputs, StringRef OutputFilename, - ProfileFormat OutputFormat) { + ProfileFormat OutputFormat, bool OutputSparse) { if (OutputFilename.compare("-") == 0) exitWithError("Cannot write indexed profdata format to stdout."); @@ -119,23 +131,29 @@ static void mergeInstrProfile(const WeightedFileVector &Inputs, if (EC) exitWithErrorCode(EC, OutputFilename); - InstrProfWriter Writer; - SmallSet<std::error_code, 4> WriterErrorCodes; + InstrProfWriter Writer(OutputSparse); + SmallSet<instrprof_error, 4> WriterErrorCodes; for (const auto &Input : Inputs) { auto ReaderOrErr = InstrProfReader::create(Input.Filename); - if (std::error_code ec = ReaderOrErr.getError()) - exitWithErrorCode(ec, Input.Filename); + if (Error E = ReaderOrErr.takeError()) + exitWithError(std::move(E), Input.Filename); auto Reader = std::move(ReaderOrErr.get()); + bool IsIRProfile = Reader->isIRLevelProfile(); + if (Writer.setIsIRLevelProfile(IsIRProfile)) + exitWithError("Merge IR generated profile with Clang generated profile."); + for (auto &I : *Reader) { - if (std::error_code EC = Writer.addRecord(std::move(I), Input.Weight)) { + if (Error E = Writer.addRecord(std::move(I), Input.Weight)) { // Only show hint the first time an error occurs. - bool firstTime = WriterErrorCodes.insert(EC).second; - handleMergeWriterError(EC, Input.Filename, I.Name, firstTime); + instrprof_error IPE = InstrProfError::take(std::move(E)); + bool firstTime = WriterErrorCodes.insert(IPE).second; + handleMergeWriterError(make_error<InstrProfError>(IPE), Input.Filename, + I.Name, firstTime); } } if (Reader->hasError()) - exitWithErrorCode(Reader->getError(), Input.Filename); + exitWithError(Reader->getError(), Input.Filename); } if (OutputFormat == PF_Text) Writer.writeText(Output); @@ -159,9 +177,9 @@ static void mergeSampleProfile(const WeightedFileVector &Inputs, auto Writer = std::move(WriterOrErr.get()); StringMap<FunctionSamples> ProfileMap; SmallVector<std::unique_ptr<sampleprof::SampleProfileReader>, 5> Readers; + LLVMContext Context; for (const auto &Input : Inputs) { - auto ReaderOrErr = - SampleProfileReader::create(Input.Filename, getGlobalContext()); + auto ReaderOrErr = SampleProfileReader::create(Input.Filename, Context); if (std::error_code EC = ReaderOrErr.getError()) exitWithErrorCode(EC, Input.Filename); @@ -183,7 +201,7 @@ static void mergeSampleProfile(const WeightedFileVector &Inputs, sampleprof_error Result = ProfileMap[FName].merge(Samples, Input.Weight); if (Result != sampleprof_error::success) { std::error_code EC = make_error_code(Result); - handleMergeWriterError(EC, Input.Filename, FName); + handleMergeWriterError(errorCodeToError(EC), Input.Filename, FName); } } } @@ -205,11 +223,53 @@ static WeightedFile parseWeightedFile(const StringRef &WeightedFilename) { return WeightedFile(FileName, Weight); } +static std::unique_ptr<MemoryBuffer> +getInputFilenamesFileBuf(const StringRef &InputFilenamesFile) { + if (InputFilenamesFile == "") + return {}; + + auto BufOrError = MemoryBuffer::getFileOrSTDIN(InputFilenamesFile); + if (!BufOrError) + exitWithErrorCode(BufOrError.getError(), InputFilenamesFile); + + return std::move(*BufOrError); +} + +static void parseInputFilenamesFile(MemoryBuffer *Buffer, + WeightedFileVector &WFV) { + if (!Buffer) + return; + + SmallVector<StringRef, 8> Entries; + StringRef Data = Buffer->getBuffer(); + Data.split(Entries, '\n', /*MaxSplit=*/-1, /*KeepEmpty=*/false); + for (const StringRef &FileWeightEntry : Entries) { + StringRef SanitizedEntry = FileWeightEntry.trim(" \t\v\f\r"); + // Skip comments. + if (SanitizedEntry.startswith("#")) + continue; + // If there's no comma, it's an unweighted profile. + else if (SanitizedEntry.find(',') == StringRef::npos) + WFV.emplace_back(SanitizedEntry, 1); + else + WFV.emplace_back(parseWeightedFile(SanitizedEntry)); + } +} + static int merge_main(int argc, const char *argv[]) { cl::list<std::string> InputFilenames(cl::Positional, cl::desc("<filename...>")); cl::list<std::string> WeightedInputFilenames("weighted-input", cl::desc("<weight>,<filename>")); + cl::opt<std::string> InputFilenamesFile( + "input-files", cl::init(""), + cl::desc("Path to file containing newline-separated " + "[<weight>,]<filename> entries")); + cl::alias InputFilenamesFileA("f", cl::desc("Alias for --input-files"), + cl::aliasopt(InputFilenamesFile)); + cl::opt<bool> DumpInputFileList( + "dump-input-file-list", cl::init(false), cl::Hidden, + cl::desc("Dump the list of input files and their weights, then exit")); cl::opt<std::string> OutputFilename("output", cl::value_desc("output"), cl::init("-"), cl::Required, cl::desc("Output file")); @@ -219,7 +279,6 @@ static int merge_main(int argc, const char *argv[]) { cl::desc("Profile kind:"), cl::init(instr), cl::values(clEnumVal(instr, "Instrumentation profile (default)"), clEnumVal(sample, "Sample profile"), clEnumValEnd)); - cl::opt<ProfileFormat> OutputFormat( cl::desc("Format of output profile"), cl::init(PF_Binary), cl::values(clEnumValN(PF_Binary, "binary", "Binary encoding (default)"), @@ -227,38 +286,63 @@ static int merge_main(int argc, const char *argv[]) { clEnumValN(PF_GCC, "gcc", "GCC encoding (only meaningful for -sample)"), clEnumValEnd)); + cl::opt<bool> OutputSparse("sparse", cl::init(false), + cl::desc("Generate a sparse profile (only meaningful for -instr)")); cl::ParseCommandLineOptions(argc, argv, "LLVM profile data merger\n"); - if (InputFilenames.empty() && WeightedInputFilenames.empty()) - exitWithError("No input files specified. See " + - sys::path::filename(argv[0]) + " -help"); - WeightedFileVector WeightedInputs; for (StringRef Filename : InputFilenames) - WeightedInputs.push_back(WeightedFile(Filename, 1)); + WeightedInputs.emplace_back(Filename, 1); for (StringRef WeightedFilename : WeightedInputFilenames) - WeightedInputs.push_back(parseWeightedFile(WeightedFilename)); + WeightedInputs.emplace_back(parseWeightedFile(WeightedFilename)); + + // Make sure that the file buffer stays alive for the duration of the + // weighted input vector's lifetime. + auto Buffer = getInputFilenamesFileBuf(InputFilenamesFile); + parseInputFilenamesFile(Buffer.get(), WeightedInputs); + + if (WeightedInputs.empty()) + exitWithError("No input files specified. See " + + sys::path::filename(argv[0]) + " -help"); + + if (DumpInputFileList) { + for (auto &WF : WeightedInputs) + outs() << WF.Weight << "," << WF.Filename << "\n"; + return 0; + } if (ProfileKind == instr) - mergeInstrProfile(WeightedInputs, OutputFilename, OutputFormat); + mergeInstrProfile(WeightedInputs, OutputFilename, OutputFormat, + OutputSparse); else mergeSampleProfile(WeightedInputs, OutputFilename, OutputFormat); return 0; } -static int showInstrProfile(std::string Filename, bool ShowCounts, - bool ShowIndirectCallTargets, bool ShowAllFunctions, - std::string ShowFunction, bool TextFormat, +static int showInstrProfile(const std::string &Filename, bool ShowCounts, + bool ShowIndirectCallTargets, + bool ShowDetailedSummary, + std::vector<uint32_t> DetailedSummaryCutoffs, + bool ShowAllFunctions, + const std::string &ShowFunction, bool TextFormat, raw_fd_ostream &OS) { auto ReaderOrErr = InstrProfReader::create(Filename); - if (std::error_code EC = ReaderOrErr.getError()) - exitWithErrorCode(EC, Filename); + std::vector<uint32_t> Cutoffs = std::move(DetailedSummaryCutoffs); + if (ShowDetailedSummary && Cutoffs.empty()) { + Cutoffs = {800000, 900000, 950000, 990000, 999000, 999900, 999990}; + } + InstrProfSummaryBuilder Builder(std::move(Cutoffs)); + if (Error E = ReaderOrErr.takeError()) + exitWithError(std::move(E), Filename); auto Reader = std::move(ReaderOrErr.get()); - uint64_t MaxFunctionCount = 0, MaxBlockCount = 0; - size_t ShownFunctions = 0, TotalFunctions = 0; + bool IsIRInstr = Reader->isIRLevelProfile(); + size_t ShownFunctions = 0; + uint64_t TotalNumValueSites = 0; + uint64_t TotalNumValueSitesWithValueProfile = 0; + uint64_t TotalNumValues = 0; for (const auto &Func : *Reader) { bool Show = ShowAllFunctions || (!ShowFunction.empty() && @@ -272,15 +356,8 @@ static int showInstrProfile(std::string Filename, bool ShowCounts, continue; } - ++TotalFunctions; assert(Func.Counts.size() > 0 && "function missing entry counter"); - if (Func.Counts[0] > MaxFunctionCount) - MaxFunctionCount = Func.Counts[0]; - - for (size_t I = 1, E = Func.Counts.size(); I < E; ++I) { - if (Func.Counts[I] > MaxBlockCount) - MaxBlockCount = Func.Counts[I]; - } + Builder.addRecord(Func); if (Show) { @@ -291,8 +368,9 @@ static int showInstrProfile(std::string Filename, bool ShowCounts, OS << " " << Func.Name << ":\n" << " Hash: " << format("0x%016" PRIx64, Func.Hash) << "\n" - << " Counters: " << Func.Counts.size() << "\n" - << " Function count: " << Func.Counts[0] << "\n"; + << " Counters: " << Func.Counts.size() << "\n"; + if (!IsIRInstr) + OS << " Function count: " << Func.Counts[0] << "\n"; if (ShowIndirectCallTargets) OS << " Indirect Call Site Count: " @@ -300,8 +378,9 @@ static int showInstrProfile(std::string Filename, bool ShowCounts, if (ShowCounts) { OS << " Block counts: ["; - for (size_t I = 1, E = Func.Counts.size(); I < E; ++I) { - OS << (I == 1 ? "" : ", ") << Func.Counts[I]; + size_t Start = (IsIRInstr ? 0 : 1); + for (size_t I = Start, E = Func.Counts.size(); I < E; ++I) { + OS << (I == Start ? "" : ", ") << Func.Counts[I]; } OS << "]\n"; } @@ -310,10 +389,14 @@ static int showInstrProfile(std::string Filename, bool ShowCounts, InstrProfSymtab &Symtab = Reader->getSymtab(); uint32_t NS = Func.getNumValueSites(IPVK_IndirectCallTarget); OS << " Indirect Target Results: \n"; + TotalNumValueSites += NS; for (size_t I = 0; I < NS; ++I) { uint32_t NV = Func.getNumValueDataForSite(IPVK_IndirectCallTarget, I); std::unique_ptr<InstrProfValueData[]> VD = Func.getValueForSite(IPVK_IndirectCallTarget, I); + TotalNumValues += NV; + if (NV) + TotalNumValueSitesWithValueProfile++; for (uint32_t V = 0; V < NV; V++) { OS << "\t[ " << I << ", "; OS << Symtab.getFuncName(VD[V].Value) << ", " << VD[V].Count @@ -323,26 +406,46 @@ static int showInstrProfile(std::string Filename, bool ShowCounts, } } } - if (Reader->hasError()) - exitWithErrorCode(Reader->getError(), Filename); + exitWithError(Reader->getError(), Filename); if (ShowCounts && TextFormat) return 0; - + std::unique_ptr<ProfileSummary> PS(Builder.getSummary()); if (ShowAllFunctions || !ShowFunction.empty()) OS << "Functions shown: " << ShownFunctions << "\n"; - OS << "Total functions: " << TotalFunctions << "\n"; - OS << "Maximum function count: " << MaxFunctionCount << "\n"; - OS << "Maximum internal block count: " << MaxBlockCount << "\n"; + OS << "Total functions: " << PS->getNumFunctions() << "\n"; + OS << "Maximum function count: " << PS->getMaxFunctionCount() << "\n"; + OS << "Maximum internal block count: " << PS->getMaxInternalCount() << "\n"; + if (ShownFunctions && ShowIndirectCallTargets) { + OS << "Total Number of Indirect Call Sites : " << TotalNumValueSites + << "\n"; + OS << "Total Number of Sites With Values : " + << TotalNumValueSitesWithValueProfile << "\n"; + OS << "Total Number of Profiled Values : " << TotalNumValues << "\n"; + } + + if (ShowDetailedSummary) { + OS << "Detailed summary:\n"; + OS << "Total number of blocks: " << PS->getNumCounts() << "\n"; + OS << "Total count: " << PS->getTotalCount() << "\n"; + for (auto Entry : PS->getDetailedSummary()) { + OS << Entry.NumCounts << " blocks with count >= " << Entry.MinCount + << " account for " + << format("%0.6g", (float)Entry.Cutoff / ProfileSummary::Scale * 100) + << " percentage of the total counts.\n"; + } + } return 0; } -static int showSampleProfile(std::string Filename, bool ShowCounts, - bool ShowAllFunctions, std::string ShowFunction, +static int showSampleProfile(const std::string &Filename, bool ShowCounts, + bool ShowAllFunctions, + const std::string &ShowFunction, raw_fd_ostream &OS) { using namespace sampleprof; - auto ReaderOrErr = SampleProfileReader::create(Filename, getGlobalContext()); + LLVMContext Context; + auto ReaderOrErr = SampleProfileReader::create(Filename, Context); if (std::error_code EC = ReaderOrErr.getError()) exitWithErrorCode(EC, Filename); @@ -370,6 +473,13 @@ static int show_main(int argc, const char *argv[]) { cl::opt<bool> ShowIndirectCallTargets( "ic-targets", cl::init(false), cl::desc("Show indirect call site target values for shown functions")); + cl::opt<bool> ShowDetailedSummary("detailed-summary", cl::init(false), + cl::desc("Show detailed profile summary")); + cl::list<uint32_t> DetailedSummaryCutoffs( + cl::CommaSeparated, "detailed-summary-cutoffs", + cl::desc( + "Cutoff percentages (times 10000) for generating detailed summary"), + cl::value_desc("800000,901000,999999")); cl::opt<bool> ShowAllFunctions("all-functions", cl::init(false), cl::desc("Details for every function")); cl::opt<std::string> ShowFunction("function", @@ -397,8 +507,11 @@ static int show_main(int argc, const char *argv[]) { if (ShowAllFunctions && !ShowFunction.empty()) errs() << "warning: -function argument ignored: showing all functions\n"; + std::vector<uint32_t> Cutoffs(DetailedSummaryCutoffs.begin(), + DetailedSummaryCutoffs.end()); if (ProfileKind == instr) return showInstrProfile(Filename, ShowCounts, ShowIndirectCallTargets, + ShowDetailedSummary, DetailedSummaryCutoffs, ShowAllFunctions, ShowFunction, TextFormat, OS); else return showSampleProfile(Filename, ShowCounts, ShowAllFunctions, @@ -407,7 +520,7 @@ static int show_main(int argc, const char *argv[]) { int main(int argc, const char *argv[]) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. diff --git a/tools/llvm-readobj/ARMAttributeParser.cpp b/tools/llvm-readobj/ARMAttributeParser.cpp index 688d349d7ec03..877dd71c90706 100644 --- a/tools/llvm-readobj/ARMAttributeParser.cpp +++ b/tools/llvm-readobj/ARMAttributeParser.cpp @@ -8,10 +8,10 @@ //===----------------------------------------------------------------------===// #include "ARMAttributeParser.h" -#include "StreamWriter.h" #include "llvm/ADT/STLExtras.h" #include "llvm/ADT/StringExtras.h" #include "llvm/Support/LEB128.h" +#include "llvm/Support/ScopedPrinter.h" using namespace llvm; using namespace llvm::ARMBuildAttrs; @@ -63,6 +63,7 @@ ARMAttributeParser::DisplayRoutines[] = { ATTRIBUTE_HANDLER(ABI_FP_16bit_format), ATTRIBUTE_HANDLER(MPextension_use), ATTRIBUTE_HANDLER(DIV_use), + ATTRIBUTE_HANDLER(DSP_extension), ATTRIBUTE_HANDLER(T2EE_use), ATTRIBUTE_HANDLER(Virtualization_use), ATTRIBUTE_HANDLER(nodefaults) @@ -340,7 +341,7 @@ void ARMAttributeParser::ABI_align_needed(AttrType Tag, const uint8_t *Data, if (Value < array_lengthof(Strings)) Description = std::string(Strings[Value]); else if (Value <= 12) - Description = std::string("8-byte alignment, ") + utostr(1 << Value) + Description = std::string("8-byte alignment, ") + utostr(1ULL << Value) + std::string("-byte extended alignment"); else Description = "Invalid"; @@ -361,8 +362,8 @@ void ARMAttributeParser::ABI_align_preserved(AttrType Tag, const uint8_t *Data, if (Value < array_lengthof(Strings)) Description = std::string(Strings[Value]); else if (Value <= 12) - Description = std::string("8-byte stack alignment, ") + utostr(1 << Value) - + std::string("-byte data alignment"); + Description = std::string("8-byte stack alignment, ") + + utostr(1ULL << Value) + std::string("-byte data alignment"); else Description = "Invalid"; @@ -517,6 +518,16 @@ void ARMAttributeParser::DIV_use(AttrType Tag, const uint8_t *Data, PrintAttribute(Tag, Value, ValueDesc); } +void ARMAttributeParser::DSP_extension(AttrType Tag, const uint8_t *Data, + uint32_t &Offset) { + static const char *const Strings[] = { "Not Permitted", "Permitted" }; + + uint64_t Value = ParseInteger(Data, Offset); + StringRef ValueDesc = + (Value < array_lengthof(Strings)) ? Strings[Value] : nullptr; + PrintAttribute(Tag, Value, ValueDesc); +} + void ARMAttributeParser::T2EE_use(AttrType Tag, const uint8_t *Data, uint32_t &Offset) { static const char *const Strings[] = { "Not Permitted", "Permitted" }; diff --git a/tools/llvm-readobj/ARMAttributeParser.h b/tools/llvm-readobj/ARMAttributeParser.h index f924c835d3eaf..6936b70ca1230 100644 --- a/tools/llvm-readobj/ARMAttributeParser.h +++ b/tools/llvm-readobj/ARMAttributeParser.h @@ -10,14 +10,14 @@ #ifndef LLVM_TOOLS_LLVM_READOBJ_ARMATTRIBUTEPARSER_H #define LLVM_TOOLS_LLVM_READOBJ_ARMATTRIBUTEPARSER_H -#include "StreamWriter.h" #include "llvm/Support/ARMBuildAttributes.h" +#include "llvm/Support/ScopedPrinter.h" namespace llvm { class StringRef; class ARMAttributeParser { - StreamWriter &SW; + ScopedPrinter &SW; struct DisplayHandler { ARMBuildAttrs::AttrType Attribute; @@ -100,6 +100,8 @@ class ARMAttributeParser { uint32_t &Offset); void DIV_use(ARMBuildAttrs::AttrType Tag, const uint8_t *Data, uint32_t &Offset); + void DSP_extension(ARMBuildAttrs::AttrType Tag, const uint8_t *Data, + uint32_t &Offset); void T2EE_use(ARMBuildAttrs::AttrType Tag, const uint8_t *Data, uint32_t &Offset); void Virtualization_use(ARMBuildAttrs::AttrType Tag, const uint8_t *Data, @@ -113,7 +115,7 @@ class ARMAttributeParser { SmallVectorImpl<uint8_t> &IndexList); void ParseSubsection(const uint8_t *Data, uint32_t Length); public: - ARMAttributeParser(StreamWriter &SW) : SW(SW) {} + ARMAttributeParser(ScopedPrinter &SW) : SW(SW) {} void Parse(ArrayRef<uint8_t> Section); }; diff --git a/tools/llvm-readobj/ARMEHABIPrinter.h b/tools/llvm-readobj/ARMEHABIPrinter.h index beb5fd4ea042b..59c9b713d85e1 100644 --- a/tools/llvm-readobj/ARMEHABIPrinter.h +++ b/tools/llvm-readobj/ARMEHABIPrinter.h @@ -11,7 +11,7 @@ #define LLVM_TOOLS_LLVM_READOBJ_ARMEHABIPRINTER_H #include "Error.h" -#include "StreamWriter.h" +#include "llvm-readobj.h" #include "llvm/ADT/STLExtras.h" #include "llvm/Object/ELF.h" #include "llvm/Object/ELFTypes.h" @@ -19,6 +19,7 @@ #include "llvm/Support/Debug.h" #include "llvm/Support/Endian.h" #include "llvm/Support/Format.h" +#include "llvm/Support/ScopedPrinter.h" #include "llvm/Support/type_traits.h" namespace llvm { @@ -26,7 +27,7 @@ namespace ARM { namespace EHABI { class OpcodeDecoder { - StreamWriter &SW; + ScopedPrinter &SW; raw_ostream &OS; struct RingEntry { @@ -63,7 +64,7 @@ class OpcodeDecoder { void PrintRegisters(uint32_t Mask, StringRef Prefix); public: - OpcodeDecoder(StreamWriter &SW) : SW(SW), OS(SW.getOStream()) {} + OpcodeDecoder(ScopedPrinter &SW) : SW(SW), OS(SW.getOStream()) {} void Decode(const uint8_t *Opcodes, off_t Offset, size_t Length); }; @@ -310,7 +311,7 @@ class PrinterContext { typedef typename object::ELFFile<ET>::Elf_Rel Elf_Rel; typedef typename object::ELFFile<ET>::Elf_Word Elf_Word; - StreamWriter &SW; + ScopedPrinter &SW; const object::ELFFile<ET> *ELF; const Elf_Shdr *Symtab; ArrayRef<Elf_Word> ShndxTable; @@ -334,7 +335,7 @@ class PrinterContext { void PrintOpcodes(const uint8_t *Entry, size_t Length, off_t Offset) const; public: - PrinterContext(StreamWriter &SW, const object::ELFFile<ET> *ELF, + PrinterContext(ScopedPrinter &SW, const object::ELFFile<ET> *ELF, const Elf_Shdr *Symtab) : SW(SW), ELF(ELF), Symtab(Symtab) {} @@ -354,8 +355,15 @@ PrinterContext<ET>::FunctionAtAddress(unsigned Section, for (const Elf_Sym &Sym : ELF->symbols(Symtab)) if (Sym.st_shndx == Section && Sym.st_value == Address && - Sym.getType() == ELF::STT_FUNC) - return Sym.getName(StrTable); + Sym.getType() == ELF::STT_FUNC) { + auto NameOrErr = Sym.getName(StrTable); + if (!NameOrErr) { + // TODO: Actually report errors helpfully. + consumeError(NameOrErr.takeError()); + return readobj_error::unknown_symbol; + } + return *NameOrErr; + } return readobj_error::unknown_symbol; } diff --git a/tools/llvm-readobj/ARMWinEHPrinter.cpp b/tools/llvm-readobj/ARMWinEHPrinter.cpp index 650955d1d75c0..1a033b1eb42e8 100644 --- a/tools/llvm-readobj/ARMWinEHPrinter.cpp +++ b/tools/llvm-readobj/ARMWinEHPrinter.cpp @@ -198,12 +198,15 @@ Decoder::getSectionContaining(const COFFObjectFile &COFF, uint64_t VA) { ErrorOr<object::SymbolRef> Decoder::getSymbol(const COFFObjectFile &COFF, uint64_t VA, bool FunctionOnly) { for (const auto &Symbol : COFF.symbols()) { - if (FunctionOnly && Symbol.getType() != SymbolRef::ST_Function) + Expected<SymbolRef::Type> Type = Symbol.getType(); + if (!Type) + return errorToErrorCode(Type.takeError()); + if (FunctionOnly && *Type != SymbolRef::ST_Function) continue; - ErrorOr<uint64_t> Address = Symbol.getAddress(); - if (std::error_code EC = Address.getError()) - return EC; + Expected<uint64_t> Address = Symbol.getAddress(); + if (!Address) + return errorToErrorCode(Address.takeError()); if (*Address == VA) return Symbol; } @@ -247,7 +250,7 @@ bool Decoder::opcode_10Lxxxxx(const uint8_t *OC, unsigned &Offset, printRegisters(std::make_pair(RegisterMask, 0)); OS << '\n'; - ++Offset, ++Offset; + Offset += 2; return false; } @@ -320,7 +323,7 @@ bool Decoder::opcode_111010xx(const uint8_t *OC, unsigned &Offset, static_cast<const char *>(Prologue ? "sub" : "add"), Imm); - ++Offset, ++Offset; + Offset += 2; return false; } @@ -334,7 +337,7 @@ bool Decoder::opcode_1110110L(const uint8_t *OC, unsigned &Offset, printRegisters(std::make_pair(GPRMask, 0)); OS << '\n'; - ++Offset, ++Offset; + Offset += 2; return false; } @@ -350,7 +353,7 @@ bool Decoder::opcode_11101110(const uint8_t *OC, unsigned &Offset, << format("0x%02x 0x%02x ; microsoft-specific (type: %u)\n", OC[Offset + 0], OC[Offset + 1], OC[Offset + 1] & 0x0f); - ++Offset, ++Offset; + Offset += 2; return false; } @@ -366,7 +369,7 @@ bool Decoder::opcode_11101111(const uint8_t *OC, unsigned &Offset, << format("0x%02x 0x%02x ; ldr.w lr, [sp], #%u\n", OC[Offset + 0], OC[Offset + 1], OC[Offset + 1] << 2); - ++Offset, ++Offset; + Offset += 2; return false; } @@ -381,7 +384,7 @@ bool Decoder::opcode_11110101(const uint8_t *OC, unsigned &Offset, printRegisters(std::make_pair(0, VFPMask)); OS << '\n'; - ++Offset, ++Offset; + Offset += 2; return false; } @@ -396,7 +399,7 @@ bool Decoder::opcode_11110110(const uint8_t *OC, unsigned &Offset, printRegisters(std::make_pair(0, VFPMask)); OS << '\n'; - ++Offset, ++Offset; + Offset += 2; return false; } @@ -409,7 +412,7 @@ bool Decoder::opcode_11110111(const uint8_t *OC, unsigned &Offset, static_cast<const char *>(Prologue ? "sub" : "add"), Imm); - ++Offset, ++Offset, ++Offset; + Offset += 3; return false; } @@ -424,7 +427,7 @@ bool Decoder::opcode_11111000(const uint8_t *OC, unsigned &Offset, OC[Offset + 0], OC[Offset + 1], OC[Offset + 2], OC[Offset + 3], static_cast<const char *>(Prologue ? "sub" : "add"), Imm); - ++Offset, ++Offset, ++Offset, ++Offset; + Offset += 4; return false; } @@ -437,7 +440,7 @@ bool Decoder::opcode_11111001(const uint8_t *OC, unsigned &Offset, OC[Offset + 0], OC[Offset + 1], OC[Offset + 2], static_cast<const char *>(Prologue ? "sub" : "add"), Imm); - ++Offset, ++Offset, ++Offset; + Offset += 3; return false; } @@ -452,7 +455,7 @@ bool Decoder::opcode_11111010(const uint8_t *OC, unsigned &Offset, OC[Offset + 0], OC[Offset + 1], OC[Offset + 2], OC[Offset + 3], static_cast<const char *>(Prologue ? "sub" : "add"), Imm); - ++Offset, ++Offset, ++Offset, ++Offset; + Offset += 4; return false; } @@ -567,9 +570,14 @@ bool Decoder::dumpXDataRecord(const COFFObjectFile &COFF, if (!Symbol) Symbol = getSymbol(COFF, Address, /*FunctionOnly=*/true); - ErrorOr<StringRef> Name = Symbol->getName(); - if (std::error_code EC = Name.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> Name = Symbol->getName(); + if (!Name) { + std::string Buf; + llvm::raw_string_ostream OS(Buf); + logAllUnhandledErrors(Name.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } ListScope EHS(SW, "ExceptionHandler"); SW.printString("Routine", formatSymbol(*Name, Address)); @@ -601,13 +609,23 @@ bool Decoder::dumpUnpackedEntry(const COFFObjectFile &COFF, StringRef FunctionName; uint64_t FunctionAddress; if (Function) { - ErrorOr<StringRef> FunctionNameOrErr = Function->getName(); - if (std::error_code EC = FunctionNameOrErr.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> FunctionNameOrErr = Function->getName(); + if (!FunctionNameOrErr) { + std::string Buf; + llvm::raw_string_ostream OS(Buf); + logAllUnhandledErrors(FunctionNameOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } FunctionName = *FunctionNameOrErr; - ErrorOr<uint64_t> FunctionAddressOrErr = Function->getAddress(); - if (std::error_code EC = FunctionAddressOrErr.getError()) - report_fatal_error(EC.message()); + Expected<uint64_t> FunctionAddressOrErr = Function->getAddress(); + if (!FunctionAddressOrErr) { + std::string Buf; + llvm::raw_string_ostream OS(Buf); + logAllUnhandledErrors(FunctionAddressOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } FunctionAddress = *FunctionAddressOrErr; } else { const pe32_header *PEHeader; @@ -619,20 +637,33 @@ bool Decoder::dumpUnpackedEntry(const COFFObjectFile &COFF, SW.printString("Function", formatSymbol(FunctionName, FunctionAddress)); if (XDataRecord) { - ErrorOr<StringRef> Name = XDataRecord->getName(); - if (std::error_code EC = Name.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> Name = XDataRecord->getName(); + if (!Name) { + std::string Buf; + llvm::raw_string_ostream OS(Buf); + logAllUnhandledErrors(Name.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } - ErrorOr<uint64_t> AddressOrErr = XDataRecord->getAddress(); - if (std::error_code EC = AddressOrErr.getError()) - report_fatal_error(EC.message()); + Expected<uint64_t> AddressOrErr = XDataRecord->getAddress(); + if (!AddressOrErr) { + std::string Buf; + llvm::raw_string_ostream OS(Buf); + logAllUnhandledErrors(AddressOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } uint64_t Address = *AddressOrErr; SW.printString("ExceptionRecord", formatSymbol(*Name, Address)); - ErrorOr<section_iterator> SIOrErr = XDataRecord->getSection(); - if (!SIOrErr) + Expected<section_iterator> SIOrErr = XDataRecord->getSection(); + if (!SIOrErr) { + // TODO: Actually report errors helpfully. + consumeError(SIOrErr.takeError()); return false; + } section_iterator SI = *SIOrErr; return dumpXDataRecord(COFF, *SI, FunctionAddress, Address); @@ -668,11 +699,23 @@ bool Decoder::dumpPackedEntry(const object::COFFObjectFile &COFF, StringRef FunctionName; uint64_t FunctionAddress; if (Function) { - ErrorOr<StringRef> FunctionNameOrErr = Function->getName(); - if (std::error_code EC = FunctionNameOrErr.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> FunctionNameOrErr = Function->getName(); + if (!FunctionNameOrErr) { + std::string Buf; + llvm::raw_string_ostream OS(Buf); + logAllUnhandledErrors(FunctionNameOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } FunctionName = *FunctionNameOrErr; - ErrorOr<uint64_t> FunctionAddressOrErr = Function->getAddress(); + Expected<uint64_t> FunctionAddressOrErr = Function->getAddress(); + if (!FunctionAddressOrErr) { + std::string Buf; + llvm::raw_string_ostream OS(Buf); + logAllUnhandledErrors(FunctionAddressOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } FunctionAddress = *FunctionAddressOrErr; } else { const pe32_header *PEHeader; diff --git a/tools/llvm-readobj/ARMWinEHPrinter.h b/tools/llvm-readobj/ARMWinEHPrinter.h index 274ef114841c1..95f521702268a 100644 --- a/tools/llvm-readobj/ARMWinEHPrinter.h +++ b/tools/llvm-readobj/ARMWinEHPrinter.h @@ -10,9 +10,9 @@ #ifndef LLVM_TOOLS_LLVM_READOBJ_ARMWINEHPRINTER_H #define LLVM_TOOLS_LLVM_READOBJ_ARMWINEHPRINTER_H -#include "StreamWriter.h" #include "llvm/Object/COFF.h" #include "llvm/Support/ErrorOr.h" +#include "llvm/Support/ScopedPrinter.h" namespace llvm { namespace ARM { @@ -22,7 +22,7 @@ class RuntimeFunction; class Decoder { static const size_t PDataEntrySize; - StreamWriter &SW; + ScopedPrinter &SW; raw_ostream &OS; struct RingEntry { @@ -107,7 +107,7 @@ class Decoder { const object::SectionRef Section); public: - Decoder(StreamWriter &SW) : SW(SW), OS(SW.getOStream()) {} + Decoder(ScopedPrinter &SW) : SW(SW), OS(SW.getOStream()) {} std::error_code dumpProcedureData(const object::COFFObjectFile &COFF); }; } diff --git a/tools/llvm-readobj/CMakeLists.txt b/tools/llvm-readobj/CMakeLists.txt index a06b678780cb4..7477f5b035d0c 100644 --- a/tools/llvm-readobj/CMakeLists.txt +++ b/tools/llvm-readobj/CMakeLists.txt @@ -1,6 +1,8 @@ set(LLVM_LINK_COMPONENTS + DebugInfoCodeView Object Support + DebugInfoCodeView ) add_llvm_tool(llvm-readobj @@ -13,6 +15,5 @@ add_llvm_tool(llvm-readobj llvm-readobj.cpp MachODumper.cpp ObjDumper.cpp - StreamWriter.cpp Win64EHDumper.cpp ) diff --git a/tools/llvm-readobj/COFFDumper.cpp b/tools/llvm-readobj/COFFDumper.cpp index d44da0d574661..348f5b4355643 100644 --- a/tools/llvm-readobj/COFFDumper.cpp +++ b/tools/llvm-readobj/COFFDumper.cpp @@ -12,16 +12,28 @@ /// //===----------------------------------------------------------------------===// -#include "llvm-readobj.h" #include "ARMWinEHPrinter.h" +#include "CodeView.h" #include "Error.h" #include "ObjDumper.h" #include "StackMapPrinter.h" -#include "StreamWriter.h" #include "Win64EHDumper.h" +#include "llvm-readobj.h" #include "llvm/ADT/DenseMap.h" #include "llvm/ADT/SmallString.h" #include "llvm/ADT/StringExtras.h" +#include "llvm/DebugInfo/CodeView/ByteStream.h" +#include "llvm/DebugInfo/CodeView/CodeView.h" +#include "llvm/DebugInfo/CodeView/Line.h" +#include "llvm/DebugInfo/CodeView/MemoryTypeTableBuilder.h" +#include "llvm/DebugInfo/CodeView/RecordSerialization.h" +#include "llvm/DebugInfo/CodeView/SymbolDumpDelegate.h" +#include "llvm/DebugInfo/CodeView/SymbolDumper.h" +#include "llvm/DebugInfo/CodeView/SymbolRecord.h" +#include "llvm/DebugInfo/CodeView/TypeDumper.h" +#include "llvm/DebugInfo/CodeView/TypeIndex.h" +#include "llvm/DebugInfo/CodeView/TypeRecord.h" +#include "llvm/DebugInfo/CodeView/TypeStreamMerger.h" #include "llvm/Object/COFF.h" #include "llvm/Object/ObjectFile.h" #include "llvm/Support/COFF.h" @@ -29,6 +41,7 @@ #include "llvm/Support/Compiler.h" #include "llvm/Support/DataExtractor.h" #include "llvm/Support/Format.h" +#include "llvm/Support/ScopedPrinter.h" #include "llvm/Support/SourceMgr.h" #include "llvm/Support/Win64EH.h" #include "llvm/Support/raw_ostream.h" @@ -39,16 +52,18 @@ using namespace llvm; using namespace llvm::object; +using namespace llvm::codeview; +using namespace llvm::support; using namespace llvm::Win64EH; namespace { class COFFDumper : public ObjDumper { public: - COFFDumper(const llvm::object::COFFObjectFile *Obj, StreamWriter& Writer) - : ObjDumper(Writer) - , Obj(Obj) { - } + friend class COFFObjectDumpDelegate; + COFFDumper(const llvm::object::COFFObjectFile *Obj, ScopedPrinter &Writer) + : ObjDumper(Writer), Obj(Obj), + CVTD(&Writer, opts::CodeViewSubsectionBytes) {} void printFileHeaders() override; void printSections() override; @@ -60,11 +75,15 @@ public: void printCOFFExports() override; void printCOFFDirectives() override; void printCOFFBaseReloc() override; + void printCOFFDebugDirectory() override; void printCodeViewDebugInfo() override; + void + mergeCodeViewTypes(llvm::codeview::MemoryTypeTableBuilder &CVTypes) override; void printStackMap() const override; private: void printSymbol(const SymbolRef &Sym); - void printRelocation(const SectionRef &Section, const RelocationRef &Reloc); + void printRelocation(const SectionRef &Section, const RelocationRef &Reloc, + uint64_t Bias = 0); void printDataDirectory(uint32_t Index, const std::string &FieldName); void printDOSHeader(const dos_header *DH); @@ -72,11 +91,33 @@ private: void printBaseOfDataField(const pe32_header *Hdr); void printBaseOfDataField(const pe32plus_header *Hdr); - void printCodeViewSection(const SectionRef &Section); + void printCodeViewSymbolSection(StringRef SectionName, const SectionRef &Section); + void printCodeViewTypeSection(StringRef SectionName, const SectionRef &Section); + StringRef getTypeName(TypeIndex Ty); + StringRef getFileNameForFileOffset(uint32_t FileOffset); + void printFileNameForOffset(StringRef Label, uint32_t FileOffset); + void printTypeIndex(StringRef FieldName, TypeIndex TI) { + // Forward to CVTypeDumper for simplicity. + CVTD.printTypeIndex(FieldName, TI); + } void printCodeViewSymbolsSubsection(StringRef Subsection, const SectionRef &Section, - uint32_t Offset); + StringRef SectionContents); + + void printCodeViewFileChecksums(StringRef Subsection); + + void printCodeViewInlineeLines(StringRef Subsection); + + void printRelocatedField(StringRef Label, const coff_section *Sec, + uint32_t RelocOffset, uint32_t Offset, + StringRef *RelocSym = nullptr); + + void printBinaryBlockWithRelocs(StringRef Label, const SectionRef &Sec, + StringRef SectionContents, StringRef Block); + + /// Given a .debug$S section, find the string table and file checksum table. + void initializeFileAndStringTables(StringRef Data); void cacheRelocations(); @@ -84,6 +125,9 @@ private: SymbolRef &Sym); std::error_code resolveSymbolName(const coff_section *Section, uint64_t Offset, StringRef &Name); + std::error_code resolveSymbolName(const coff_section *Section, + StringRef SectionContents, + const void *RelocPtr, StringRef &Name); void printImportedSymbols(iterator_range<imported_symbol_iterator> Range); void printDelayImportedSymbols( const DelayImportDirectoryEntryRef &I, @@ -94,17 +138,56 @@ private: const llvm::object::COFFObjectFile *Obj; bool RelocCached = false; RelocMapTy RelocMap; - StringRef CVFileIndexToStringOffsetTable; + StringRef CVFileChecksumTable; StringRef CVStringTable; + + CVTypeDumper CVTD; }; -} // namespace +class COFFObjectDumpDelegate : public SymbolDumpDelegate { +public: + COFFObjectDumpDelegate(COFFDumper &CD, const SectionRef &SR, + const COFFObjectFile *Obj, StringRef SectionContents) + : CD(CD), SR(SR), SectionContents(SectionContents) { + Sec = Obj->getCOFFSection(SR); + } + + uint32_t getRecordOffset(ArrayRef<uint8_t> Record) override { + return Record.data() - SectionContents.bytes_begin(); + } + + void printRelocatedField(StringRef Label, uint32_t RelocOffset, + uint32_t Offset, StringRef *RelocSym) override { + CD.printRelocatedField(Label, Sec, RelocOffset, Offset, RelocSym); + } + + void printBinaryBlockWithRelocs(StringRef Label, + ArrayRef<uint8_t> Block) override { + StringRef SBlock(reinterpret_cast<const char *>(Block.data()), + Block.size()); + if (opts::CodeViewSubsectionBytes) + CD.printBinaryBlockWithRelocs(Label, SR, SectionContents, SBlock); + } + + StringRef getFileNameForFileOffset(uint32_t FileOffset) override { + return CD.getFileNameForFileOffset(FileOffset); + } + + StringRef getStringTable() override { return CD.CVStringTable; } + +private: + COFFDumper &CD; + const SectionRef &SR; + const coff_section *Sec; + StringRef SectionContents; +}; +} // end namespace namespace llvm { std::error_code createCOFFDumper(const object::ObjectFile *Obj, - StreamWriter &Writer, + ScopedPrinter &Writer, std::unique_ptr<ObjDumper> &Result) { const COFFObjectFile *COFFObj = dyn_cast<COFFObjectFile>(Obj); if (!COFFObj) @@ -122,15 +205,19 @@ std::error_code COFFDumper::resolveSymbol(const coff_section *Section, uint64_t Offset, SymbolRef &Sym) { cacheRelocations(); const auto &Relocations = RelocMap[Section]; + auto SymI = Obj->symbol_end(); for (const auto &Relocation : Relocations) { uint64_t RelocationOffset = Relocation.getOffset(); if (RelocationOffset == Offset) { - Sym = *Relocation.getSymbol(); - return readobj_error::success; + SymI = Relocation.getSymbol(); + break; } } - return readobj_error::unknown_symbol; + if (SymI == Obj->symbol_end()) + return readobj_error::unknown_symbol; + Sym = *SymI; + return readobj_error::success; } // Given a section and an offset into this section the function returns the name @@ -141,13 +228,62 @@ std::error_code COFFDumper::resolveSymbolName(const coff_section *Section, SymbolRef Symbol; if (std::error_code EC = resolveSymbol(Section, Offset, Symbol)) return EC; - ErrorOr<StringRef> NameOrErr = Symbol.getName(); - if (std::error_code EC = NameOrErr.getError()) - return EC; + Expected<StringRef> NameOrErr = Symbol.getName(); + if (!NameOrErr) + return errorToErrorCode(NameOrErr.takeError()); Name = *NameOrErr; return std::error_code(); } +// Helper for when you have a pointer to real data and you want to know about +// relocations against it. +std::error_code COFFDumper::resolveSymbolName(const coff_section *Section, + StringRef SectionContents, + const void *RelocPtr, + StringRef &Name) { + assert(SectionContents.data() < RelocPtr && + RelocPtr < SectionContents.data() + SectionContents.size() && + "pointer to relocated object is not in section"); + uint64_t Offset = ptrdiff_t(reinterpret_cast<const char *>(RelocPtr) - + SectionContents.data()); + return resolveSymbolName(Section, Offset, Name); +} + +void COFFDumper::printRelocatedField(StringRef Label, const coff_section *Sec, + uint32_t RelocOffset, uint32_t Offset, + StringRef *RelocSym) { + StringRef SymStorage; + StringRef &Symbol = RelocSym ? *RelocSym : SymStorage; + if (!resolveSymbolName(Sec, RelocOffset, Symbol)) + W.printSymbolOffset(Label, Symbol, Offset); + else + W.printHex(Label, RelocOffset); +} + +void COFFDumper::printBinaryBlockWithRelocs(StringRef Label, + const SectionRef &Sec, + StringRef SectionContents, + StringRef Block) { + W.printBinaryBlock(Label, Block); + + assert(SectionContents.begin() < Block.begin() && + SectionContents.end() >= Block.end() && + "Block is not contained in SectionContents"); + uint64_t OffsetStart = Block.data() - SectionContents.data(); + uint64_t OffsetEnd = OffsetStart + Block.size(); + + W.flush(); + cacheRelocations(); + ListScope D(W, "BlockRelocations"); + const coff_section *Section = Obj->getCOFFSection(Sec); + const auto &Relocations = RelocMap[Section]; + for (const auto &Relocation : Relocations) { + uint64_t RelocationOffset = Relocation.getOffset(); + if (OffsetStart <= RelocationOffset && RelocationOffset < OffsetEnd) + printRelocation(Sec, Relocation, OffsetStart); + } +} + static const EnumEntry<COFF::MachineTypes> ImageFileMachineType[] = { LLVM_READOBJ_ENUM_ENT(COFF, IMAGE_FILE_MACHINE_UNKNOWN ), LLVM_READOBJ_ENUM_ENT(COFF, IMAGE_FILE_MACHINE_AM33 ), @@ -324,6 +460,26 @@ static const EnumEntry<COFF::COMDATType> ImageCOMDATSelect[] = { { "Newest" , COFF::IMAGE_COMDAT_SELECT_NEWEST } }; +static const EnumEntry<COFF::DebugType> ImageDebugType[] = { + { "Unknown" , COFF::IMAGE_DEBUG_TYPE_UNKNOWN }, + { "COFF" , COFF::IMAGE_DEBUG_TYPE_COFF }, + { "CodeView" , COFF::IMAGE_DEBUG_TYPE_CODEVIEW }, + { "FPO" , COFF::IMAGE_DEBUG_TYPE_FPO }, + { "Misc" , COFF::IMAGE_DEBUG_TYPE_MISC }, + { "Exception" , COFF::IMAGE_DEBUG_TYPE_EXCEPTION }, + { "Fixup" , COFF::IMAGE_DEBUG_TYPE_FIXUP }, + { "OmapToSrc" , COFF::IMAGE_DEBUG_TYPE_OMAP_TO_SRC }, + { "OmapFromSrc", COFF::IMAGE_DEBUG_TYPE_OMAP_FROM_SRC }, + { "Borland" , COFF::IMAGE_DEBUG_TYPE_BORLAND }, + { "Reserved10" , COFF::IMAGE_DEBUG_TYPE_RESERVED10 }, + { "CLSID" , COFF::IMAGE_DEBUG_TYPE_CLSID }, + { "VCFeature" , COFF::IMAGE_DEBUG_TYPE_VC_FEATURE }, + { "POGO" , COFF::IMAGE_DEBUG_TYPE_POGO }, + { "ILTCG" , COFF::IMAGE_DEBUG_TYPE_ILTCG }, + { "MPX" , COFF::IMAGE_DEBUG_TYPE_MPX }, + { "Repro" , COFF::IMAGE_DEBUG_TYPE_REPRO }, +}; + static const EnumEntry<COFF::WeakExternalCharacteristics> WeakExternalCharacteristics[] = { { "NoLibrary", COFF::IMAGE_WEAK_EXTERN_SEARCH_NOLIBRARY }, @@ -331,6 +487,35 @@ WeakExternalCharacteristics[] = { { "Alias" , COFF::IMAGE_WEAK_EXTERN_SEARCH_ALIAS } }; +static const EnumEntry<uint32_t> SubSectionTypes[] = { + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, Symbols), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, Lines), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, StringTable), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, FileChecksums), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, FrameData), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, InlineeLines), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, CrossScopeImports), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, CrossScopeExports), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, ILLines), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, FuncMDTokenMap), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, TypeMDTokenMap), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, MergedAssemblyInput), + LLVM_READOBJ_ENUM_CLASS_ENT(ModuleSubstreamKind, CoffSymbolRVA), +}; + +static const EnumEntry<uint32_t> FrameDataFlags[] = { + LLVM_READOBJ_ENUM_ENT(FrameData, HasSEH), + LLVM_READOBJ_ENUM_ENT(FrameData, HasEH), + LLVM_READOBJ_ENUM_ENT(FrameData, IsFunctionStart), +}; + +static const EnumEntry<uint8_t> FileChecksumKindNames[] = { + LLVM_READOBJ_ENUM_CLASS_ENT(FileChecksumKind, None), + LLVM_READOBJ_ENUM_CLASS_ENT(FileChecksumKind, MD5), + LLVM_READOBJ_ENUM_CLASS_ENT(FileChecksumKind, SHA1), + LLVM_READOBJ_ENUM_CLASS_ENT(FileChecksumKind, SHA256), +}; + template <typename T> static std::error_code getSymbolAuxData(const COFFObjectFile *Obj, COFFSymbolRef Symbol, @@ -463,8 +648,42 @@ void COFFDumper::printPEHeader(const PEHeader *Hdr) { "DelayImportDescriptor", "CLRRuntimeHeader", "Reserved" }; - for (uint32_t i = 0; i < Hdr->NumberOfRvaAndSize; ++i) { + for (uint32_t i = 0; i < Hdr->NumberOfRvaAndSize; ++i) printDataDirectory(i, directory[i]); + } +} + +void COFFDumper::printCOFFDebugDirectory() { + ListScope LS(W, "DebugDirectory"); + for (const debug_directory &D : Obj->debug_directories()) { + char FormattedTime[20] = {}; + time_t TDS = D.TimeDateStamp; + strftime(FormattedTime, 20, "%Y-%m-%d %H:%M:%S", gmtime(&TDS)); + DictScope S(W, "DebugEntry"); + W.printHex("Characteristics", D.Characteristics); + W.printHex("TimeDateStamp", FormattedTime, D.TimeDateStamp); + W.printHex("MajorVersion", D.MajorVersion); + W.printHex("MinorVersion", D.MinorVersion); + W.printEnum("Type", D.Type, makeArrayRef(ImageDebugType)); + W.printHex("SizeOfData", D.SizeOfData); + W.printHex("AddressOfRawData", D.AddressOfRawData); + W.printHex("PointerToRawData", D.PointerToRawData); + if (D.Type == COFF::IMAGE_DEBUG_TYPE_CODEVIEW) { + const debug_pdb_info *PDBInfo; + StringRef PDBFileName; + error(Obj->getDebugPDBInfo(&D, PDBInfo, PDBFileName)); + DictScope PDBScope(W, "PDBInfo"); + W.printHex("PDBSignature", PDBInfo->Signature); + W.printBinary("PDBGUID", makeArrayRef(PDBInfo->Guid)); + W.printNumber("PDBAge", PDBInfo->Age); + W.printString("PDBFileName", PDBFileName); + } else { + // FIXME: Type values of 12 and 13 are commonly observed but are not in + // the documented type enum. Figure out what they mean. + ArrayRef<uint8_t> RawData; + error( + Obj->getRvaAndSizeAsBytes(D.AddressOfRawData, D.SizeOfData, RawData)); + W.printBinaryBlock("RawData", RawData); } } } @@ -476,111 +695,180 @@ void COFFDumper::printBaseOfDataField(const pe32_header *Hdr) { void COFFDumper::printBaseOfDataField(const pe32plus_header *) {} void COFFDumper::printCodeViewDebugInfo() { + // Print types first to build CVUDTNames, then print symbols. + for (const SectionRef &S : Obj->sections()) { + StringRef SectionName; + error(S.getName(SectionName)); + if (SectionName == ".debug$T") + printCodeViewTypeSection(SectionName, S); + } for (const SectionRef &S : Obj->sections()) { - StringRef SecName; - error(S.getName(SecName)); - if (SecName == ".debug$S") - printCodeViewSection(S); + StringRef SectionName; + error(S.getName(SectionName)); + if (SectionName == ".debug$S") + printCodeViewSymbolSection(SectionName, S); } } -void COFFDumper::printCodeViewSection(const SectionRef &Section) { - StringRef Data; - error(Section.getContents(Data)); +void COFFDumper::initializeFileAndStringTables(StringRef Data) { + while (!Data.empty() && (CVFileChecksumTable.data() == nullptr || + CVStringTable.data() == nullptr)) { + // The section consists of a number of subsection in the following format: + // |SubSectionType|SubSectionSize|Contents...| + uint32_t SubType, SubSectionSize; + error(consume(Data, SubType)); + error(consume(Data, SubSectionSize)); + if (SubSectionSize > Data.size()) + return error(object_error::parse_failed); + switch (ModuleSubstreamKind(SubType)) { + case ModuleSubstreamKind::FileChecksums: + CVFileChecksumTable = Data.substr(0, SubSectionSize); + break; + case ModuleSubstreamKind::StringTable: + CVStringTable = Data.substr(0, SubSectionSize); + break; + default: + break; + } + uint32_t PaddedSize = alignTo(SubSectionSize, 4); + if (PaddedSize > Data.size()) + error(object_error::parse_failed); + Data = Data.drop_front(PaddedSize); + } +} + +void COFFDumper::printCodeViewSymbolSection(StringRef SectionName, + const SectionRef &Section) { + StringRef SectionContents; + error(Section.getContents(SectionContents)); + StringRef Data = SectionContents; SmallVector<StringRef, 10> FunctionNames; StringMap<StringRef> FunctionLineTables; ListScope D(W, "CodeViewDebugInfo"); - { - // FIXME: Add more offset correctness checks. - DataExtractor DE(Data, true, 4); - uint32_t Offset = 0, - Magic = DE.getU32(&Offset); - W.printHex("Magic", Magic); - if (Magic != COFF::DEBUG_SECTION_MAGIC) { - error(object_error::parse_failed); - return; - } + // Print the section to allow correlation with printSections. + W.printNumber("Section", SectionName, Obj->getSectionID(Section)); + + uint32_t Magic; + error(consume(Data, Magic)); + W.printHex("Magic", Magic); + if (Magic != COFF::DEBUG_SECTION_MAGIC) + return error(object_error::parse_failed); + + initializeFileAndStringTables(Data); + + // TODO: Convert this over to using ModuleSubstreamVisitor. + while (!Data.empty()) { + // The section consists of a number of subsection in the following format: + // |SubSectionType|SubSectionSize|Contents...| + uint32_t SubType, SubSectionSize; + error(consume(Data, SubType)); + error(consume(Data, SubSectionSize)); + + ListScope S(W, "Subsection"); + W.printEnum("SubSectionType", SubType, makeArrayRef(SubSectionTypes)); + W.printHex("SubSectionSize", SubSectionSize); + + // Get the contents of the subsection. + if (SubSectionSize > Data.size()) + return error(object_error::parse_failed); + StringRef Contents = Data.substr(0, SubSectionSize); + + // Add SubSectionSize to the current offset and align that offset to find + // the next subsection. + size_t SectionOffset = Data.data() - SectionContents.data(); + size_t NextOffset = SectionOffset + SubSectionSize; + NextOffset = alignTo(NextOffset, 4); + if (NextOffset > SectionContents.size()) + return error(object_error::parse_failed); + Data = SectionContents.drop_front(NextOffset); + + // Optionally print the subsection bytes in case our parsing gets confused + // later. + if (opts::CodeViewSubsectionBytes) + printBinaryBlockWithRelocs("SubSectionContents", Section, SectionContents, + Contents); + + switch (ModuleSubstreamKind(SubType)) { + case ModuleSubstreamKind::Symbols: + printCodeViewSymbolsSubsection(Contents, Section, SectionContents); + break; + + case ModuleSubstreamKind::InlineeLines: + printCodeViewInlineeLines(Contents); + break; + + case ModuleSubstreamKind::FileChecksums: + printCodeViewFileChecksums(Contents); + break; - bool Finished = false; - while (DE.isValidOffset(Offset) && !Finished) { - // The section consists of a number of subsection in the following format: - // |Type|PayloadSize|Payload...| - uint32_t SubSectionType = DE.getU32(&Offset), - PayloadSize = DE.getU32(&Offset); - ListScope S(W, "Subsection"); - W.printHex("Type", SubSectionType); - W.printHex("PayloadSize", PayloadSize); - if (PayloadSize > Data.size() - Offset) { + case ModuleSubstreamKind::Lines: { + // Holds a PC to file:line table. Some data to parse this subsection is + // stored in the other subsections, so just check sanity and store the + // pointers for deferred processing. + + if (SubSectionSize < 12) { + // There should be at least three words to store two function + // relocations and size of the code. error(object_error::parse_failed); return; } - StringRef Contents = Data.substr(Offset, PayloadSize); - if (opts::CodeViewSubsectionBytes) { - // Print the raw contents to simplify debugging if anything goes wrong - // afterwards. - W.printBinaryBlock("Contents", Contents); + StringRef LinkageName; + error(resolveSymbolName(Obj->getCOFFSection(Section), SectionOffset, + LinkageName)); + W.printString("LinkageName", LinkageName); + if (FunctionLineTables.count(LinkageName) != 0) { + // Saw debug info for this function already? + error(object_error::parse_failed); + return; } - switch (SubSectionType) { - case COFF::DEBUG_SYMBOL_SUBSECTION: - printCodeViewSymbolsSubsection(Contents, Section, Offset); - break; - case COFF::DEBUG_LINE_TABLE_SUBSECTION: { - // Holds a PC to file:line table. Some data to parse this subsection is - // stored in the other subsections, so just check sanity and store the - // pointers for deferred processing. - - if (PayloadSize < 12) { - // There should be at least three words to store two function - // relocations and size of the code. - error(object_error::parse_failed); - return; - } - - StringRef LinkageName; - error(resolveSymbolName(Obj->getCOFFSection(Section), Offset, - LinkageName)); - W.printString("LinkageName", LinkageName); - if (FunctionLineTables.count(LinkageName) != 0) { - // Saw debug info for this function already? + FunctionLineTables[LinkageName] = Contents; + FunctionNames.push_back(LinkageName); + break; + } + case ModuleSubstreamKind::FrameData: { + // First four bytes is a relocation against the function. + const uint32_t *CodePtr; + error(consumeObject(Contents, CodePtr)); + StringRef LinkageName; + error(resolveSymbolName(Obj->getCOFFSection(Section), SectionContents, + CodePtr, LinkageName)); + W.printString("LinkageName", LinkageName); + + // To find the active frame description, search this array for the + // smallest PC range that includes the current PC. + while (!Contents.empty()) { + const FrameData *FD; + error(consumeObject(Contents, FD)); + + if (FD->FrameFunc >= CVStringTable.size()) error(object_error::parse_failed); - return; - } - FunctionLineTables[LinkageName] = Contents; - FunctionNames.push_back(LinkageName); - break; + StringRef FrameFunc = + CVStringTable.drop_front(FD->FrameFunc).split('\0').first; + + DictScope S(W, "FrameData"); + W.printHex("RvaStart", FD->RvaStart); + W.printHex("CodeSize", FD->CodeSize); + W.printHex("LocalSize", FD->LocalSize); + W.printHex("ParamsSize", FD->ParamsSize); + W.printHex("MaxStackSize", FD->MaxStackSize); + W.printString("FrameFunc", FrameFunc); + W.printHex("PrologSize", FD->PrologSize); + W.printHex("SavedRegsSize", FD->SavedRegsSize); + W.printFlags("Flags", FD->Flags, makeArrayRef(FrameDataFlags)); } - case COFF::DEBUG_STRING_TABLE_SUBSECTION: - if (PayloadSize == 0 || CVStringTable.data() != nullptr || - Contents.back() != '\0') { - // Empty or duplicate or non-null-terminated subsection. - error(object_error::parse_failed); - return; - } - CVStringTable = Contents; - break; - case COFF::DEBUG_INDEX_SUBSECTION: - // Holds the translation table from file indices - // to offsets in the string table. - - if (PayloadSize == 0 || - CVFileIndexToStringOffsetTable.data() != nullptr) { - // Empty or duplicate subsection. - error(object_error::parse_failed); - return; - } - CVFileIndexToStringOffsetTable = Contents; - break; - } - Offset += PayloadSize; + break; + } - // Align the reading pointer by 4. - Offset += (-Offset) % 4; + // Do nothing for unrecognized subsections. + default: + break; } + W.flush(); } // Dump the line tables now that we've read all the subsections and know all @@ -594,8 +882,7 @@ void COFFDumper::printCodeViewSection(const SectionRef &Section) { uint32_t Offset = 6; // Skip relocations. uint16_t Flags = DE.getU16(&Offset); W.printHex("Flags", Flags); - bool HasColumnInformation = - Flags & COFF::DEBUG_LINE_TABLES_HAVE_COLUMN_RECORDS; + bool HasColumnInformation = Flags & codeview::LineFlags::HaveColumns; uint32_t FunctionSize = DE.getU32(&Offset); W.printHex("CodeSize", FunctionSize); while (DE.isValidOffset(Offset)) { @@ -615,28 +902,8 @@ void COFFDumper::printCodeViewSection(const SectionRef &Section) { return; } - uint32_t FilenameOffset; - { - DataExtractor SDE(CVFileIndexToStringOffsetTable, true, 4); - uint32_t OffsetInSDE = OffsetInIndex; - if (!SDE.isValidOffset(OffsetInSDE)) { - error(object_error::parse_failed); - return; - } - FilenameOffset = SDE.getU32(&OffsetInSDE); - } - - if (FilenameOffset == 0 || FilenameOffset + 1 >= CVStringTable.size() || - CVStringTable.data()[FilenameOffset - 1] != '\0') { - // Each string in an F3 subsection should be preceded by a null - // character. - error(object_error::parse_failed); - return; - } - - StringRef Filename(CVStringTable.data() + FilenameOffset); ListScope S(W, "FilenameSegment"); - W.printString("Filename", Filename); + printFileNameForOffset("Filename", OffsetInIndex); for (unsigned LineIdx = 0; LineIdx != NumLines && DE.isValidOffset(Offset); ++LineIdx) { // Then go the (PC, LineNumber) pairs. The line number is stored in the @@ -649,14 +916,15 @@ void COFFDumper::printCodeViewSection(const SectionRef &Section) { char Buffer[32]; format("+0x%X", PC).snprint(Buffer, 32); ListScope PCScope(W, Buffer); - uint32_t LineNumberStart = LineData & COFF::CVL_MaxLineNumber; - uint32_t LineNumberEndDelta = - (LineData >> COFF::CVL_LineNumberStartBits) & - COFF::CVL_LineNumberEndDeltaMask; - bool IsStatement = LineData & COFF::CVL_IsStatement; - W.printNumber("LineNumberStart", LineNumberStart); - W.printNumber("LineNumberEndDelta", LineNumberEndDelta); - W.printBoolean("IsStatement", IsStatement); + LineInfo LI(LineData); + if (LI.isAlwaysStepInto()) + W.printString("StepInto", StringRef("Always")); + else if (LI.isNeverStepInto()) + W.printString("StepInto", StringRef("Never")); + else + W.printNumber("LineNumberStart", LI.getStartLine()); + W.printNumber("LineNumberEndDelta", LI.getLineDelta()); + W.printBoolean("IsStatement", LI.isStatement()); if (HasColumnInformation && ColumnDE.isValidOffsetForDataOfSize(ColumnOffset, 4)) { uint16_t ColStart = ColumnDE.getU16(&ColumnOffset); @@ -678,85 +946,156 @@ void COFFDumper::printCodeViewSection(const SectionRef &Section) { void COFFDumper::printCodeViewSymbolsSubsection(StringRef Subsection, const SectionRef &Section, - uint32_t OffsetInSection) { - if (Subsection.size() == 0) { + StringRef SectionContents) { + ArrayRef<uint8_t> BinaryData(Subsection.bytes_begin(), + Subsection.bytes_end()); + auto CODD = llvm::make_unique<COFFObjectDumpDelegate>(*this, Section, Obj, + SectionContents); + + CVSymbolDumper CVSD(W, CVTD, std::move(CODD), opts::CodeViewSubsectionBytes); + ByteStream<> Stream(BinaryData); + CVSymbolArray Symbols; + StreamReader Reader(Stream); + if (auto EC = Reader.readArray(Symbols, Reader.getLength())) { + consumeError(std::move(EC)); + W.flush(); + error(object_error::parse_failed); + } + + if (!CVSD.dump(Symbols)) { + W.flush(); error(object_error::parse_failed); - return; } - DataExtractor DE(Subsection, true, 4); - uint32_t Offset = 0; - - // Function-level subsections have "procedure start" and "procedure end" - // commands that should come in pairs and surround relevant info. - bool InFunctionScope = false; - while (DE.isValidOffset(Offset)) { - // Read subsection segments one by one. - uint16_t Size = DE.getU16(&Offset); - // The section size includes the size of the type identifier. - if (Size < 2 || !DE.isValidOffsetForDataOfSize(Offset, Size)) { + W.flush(); +} + +void COFFDumper::printCodeViewFileChecksums(StringRef Subsection) { + StringRef Data = Subsection; + while (!Data.empty()) { + DictScope S(W, "FileChecksum"); + const FileChecksum *FC; + error(consumeObject(Data, FC)); + if (FC->FileNameOffset >= CVStringTable.size()) error(object_error::parse_failed); - return; - } - Size -= 2; - uint16_t Type = DE.getU16(&Offset); - switch (Type) { - case COFF::DEBUG_SYMBOL_TYPE_PROC_START: { - DictScope S(W, "ProcStart"); - if (InFunctionScope || Size < 36) { - error(object_error::parse_failed); - return; - } - InFunctionScope = true; - - // We're currently interested in a limited subset of fields in this - // segment, just ignore the rest of the fields for now. - uint8_t Unused[12]; - DE.getU8(&Offset, Unused, 12); - uint32_t CodeSize = DE.getU32(&Offset); - DE.getU8(&Offset, Unused, 12); - StringRef SectionName; - error(resolveSymbolName(Obj->getCOFFSection(Section), - OffsetInSection + Offset, SectionName)); - Offset += 4; - DE.getU8(&Offset, Unused, 3); - StringRef DisplayName = DE.getCStr(&Offset); - if (!DE.isValidOffset(Offset)) { - error(object_error::parse_failed); - return; - } - W.printString("DisplayName", DisplayName); - W.printString("Section", SectionName); - W.printHex("CodeSize", CodeSize); + StringRef Filename = + CVStringTable.drop_front(FC->FileNameOffset).split('\0').first; + W.printHex("Filename", Filename, FC->FileNameOffset); + W.printHex("ChecksumSize", FC->ChecksumSize); + W.printEnum("ChecksumKind", uint8_t(FC->ChecksumKind), + makeArrayRef(FileChecksumKindNames)); + if (FC->ChecksumSize >= Data.size()) + error(object_error::parse_failed); + StringRef ChecksumBytes = Data.substr(0, FC->ChecksumSize); + W.printBinary("ChecksumBytes", ChecksumBytes); + unsigned PaddedSize = alignTo(FC->ChecksumSize + sizeof(FileChecksum), 4) - + sizeof(FileChecksum); + if (PaddedSize > Data.size()) + error(object_error::parse_failed); + Data = Data.drop_front(PaddedSize); + } +} - break; - } - case COFF::DEBUG_SYMBOL_TYPE_PROC_END: { - W.startLine() << "ProcEnd\n"; - if (!InFunctionScope || Size > 0) { - error(object_error::parse_failed); - return; +void COFFDumper::printCodeViewInlineeLines(StringRef Subsection) { + StringRef Data = Subsection; + uint32_t Signature; + error(consume(Data, Signature)); + bool HasExtraFiles = Signature == unsigned(InlineeLinesSignature::ExtraFiles); + + while (!Data.empty()) { + const InlineeSourceLine *ISL; + error(consumeObject(Data, ISL)); + DictScope S(W, "InlineeSourceLine"); + printTypeIndex("Inlinee", ISL->Inlinee); + printFileNameForOffset("FileID", ISL->FileID); + W.printNumber("SourceLineNum", ISL->SourceLineNum); + + if (HasExtraFiles) { + uint32_t ExtraFileCount; + error(consume(Data, ExtraFileCount)); + W.printNumber("ExtraFileCount", ExtraFileCount); + ListScope ExtraFiles(W, "ExtraFiles"); + for (unsigned I = 0; I < ExtraFileCount; ++I) { + uint32_t FileID; + error(consume(Data, FileID)); + printFileNameForOffset("FileID", FileID); } - InFunctionScope = false; - break; } - default: { - if (opts::CodeViewSubsectionBytes) { - ListScope S(W, "Record"); - W.printHex("Size", Size); - W.printHex("Type", Type); - - StringRef Contents = DE.getData().substr(Offset, Size); - W.printBinaryBlock("Contents", Contents); + } +} + +StringRef COFFDumper::getFileNameForFileOffset(uint32_t FileOffset) { + // The file checksum subsection should precede all references to it. + if (!CVFileChecksumTable.data() || !CVStringTable.data()) + error(object_error::parse_failed); + // Check if the file checksum table offset is valid. + if (FileOffset >= CVFileChecksumTable.size()) + error(object_error::parse_failed); + + // The string table offset comes first before the file checksum. + StringRef Data = CVFileChecksumTable.drop_front(FileOffset); + uint32_t StringOffset; + error(consume(Data, StringOffset)); + + // Check if the string table offset is valid. + if (StringOffset >= CVStringTable.size()) + error(object_error::parse_failed); + + // Return the null-terminated string. + return CVStringTable.drop_front(StringOffset).split('\0').first; +} + +void COFFDumper::printFileNameForOffset(StringRef Label, uint32_t FileOffset) { + W.printHex(Label, getFileNameForFileOffset(FileOffset), FileOffset); +} + +void COFFDumper::mergeCodeViewTypes(MemoryTypeTableBuilder &CVTypes) { + for (const SectionRef &S : Obj->sections()) { + StringRef SectionName; + error(S.getName(SectionName)); + if (SectionName == ".debug$T") { + StringRef Data; + error(S.getContents(Data)); + uint32_t Magic; + error(consume(Data, Magic)); + if (Magic != 4) + error(object_error::parse_failed); + ArrayRef<uint8_t> Bytes(reinterpret_cast<const uint8_t *>(Data.data()), + Data.size()); + ByteStream<> Stream(Bytes); + CVTypeArray Types; + StreamReader Reader(Stream); + if (auto EC = Reader.readArray(Types, Reader.getLength())) { + consumeError(std::move(EC)); + W.flush(); + error(object_error::parse_failed); } - Offset += Size; - break; - } + if (!mergeTypeStreams(CVTypes, Types)) + return error(object_error::parse_failed); } } +} - if (InFunctionScope) - error(object_error::parse_failed); +void COFFDumper::printCodeViewTypeSection(StringRef SectionName, + const SectionRef &Section) { + ListScope D(W, "CodeViewTypes"); + W.printNumber("Section", SectionName, Obj->getSectionID(Section)); + + StringRef Data; + error(Section.getContents(Data)); + if (opts::CodeViewSubsectionBytes) + W.printBinaryBlock("Data", Data); + + uint32_t Magic; + error(consume(Data, Magic)); + W.printHex("Magic", Magic); + if (Magic != COFF::DEBUG_SECTION_MAGIC) + return error(object_error::parse_failed); + + if (auto EC = CVTD.dump({Data.bytes_begin(), Data.bytes_end()})) { + W.flush(); + error(llvm::errorToErrorCode(std::move(EC))); + } } void COFFDumper::printSections() { @@ -838,16 +1177,16 @@ void COFFDumper::printRelocations() { } void COFFDumper::printRelocation(const SectionRef &Section, - const RelocationRef &Reloc) { - uint64_t Offset = Reloc.getOffset(); + const RelocationRef &Reloc, uint64_t Bias) { + uint64_t Offset = Reloc.getOffset() - Bias; uint64_t RelocType = Reloc.getType(); SmallString<32> RelocName; StringRef SymbolName; Reloc.getTypeName(RelocName); symbol_iterator Symbol = Reloc.getSymbol(); if (Symbol != Obj->symbol_end()) { - ErrorOr<StringRef> SymbolNameOrErr = Symbol->getName(); - error(SymbolNameOrErr.getError()); + Expected<StringRef> SymbolNameOrErr = Symbol->getName(); + error(errorToErrorCode(SymbolNameOrErr.takeError())); SymbolName = *SymbolNameOrErr; } @@ -1190,3 +1529,18 @@ void COFFDumper::printStackMap() const { prettyPrintStackMap(llvm::outs(), StackMapV1Parser<support::big>(StackMapContentsArray)); } + +void llvm::dumpCodeViewMergedTypes( + ScopedPrinter &Writer, llvm::codeview::MemoryTypeTableBuilder &CVTypes) { + // Flatten it first, then run our dumper on it. + ListScope S(Writer, "MergedTypeStream"); + SmallString<0> Buf; + CVTypes.ForEachRecord([&](TypeIndex TI, StringRef Record) { + Buf.append(Record.begin(), Record.end()); + }); + CVTypeDumper CVTD(&Writer, opts::CodeViewSubsectionBytes); + if (auto EC = CVTD.dump({Buf.str().bytes_begin(), Buf.str().bytes_end()})) { + Writer.flush(); + error(llvm::errorToErrorCode(std::move(EC))); + } +} diff --git a/tools/llvm-readobj/CodeView.h b/tools/llvm-readobj/CodeView.h new file mode 100644 index 0000000000000..cf713962eb7fa --- /dev/null +++ b/tools/llvm-readobj/CodeView.h @@ -0,0 +1,54 @@ +//===-- CodeView.h - On-disk record types for CodeView ----------*- C++ -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +/// +/// \file +/// \brief This file provides data structures useful for consuming on-disk +/// CodeView. It is based on information published by Microsoft at +/// https://github.com/Microsoft/microsoft-pdb/. +/// +//===----------------------------------------------------------------------===// + +// FIXME: Find a home for this in include/llvm/DebugInfo/CodeView/. + +#ifndef LLVM_READOBJ_CODEVIEW_H +#define LLVM_READOBJ_CODEVIEW_H + +#include "llvm/DebugInfo/CodeView/CodeView.h" +#include "llvm/DebugInfo/CodeView/TypeIndex.h" +#include "llvm/Support/Endian.h" + +namespace llvm { +namespace codeview { + +using llvm::support::ulittle16_t; +using llvm::support::ulittle32_t; + +/// Data in the the SUBSEC_FRAMEDATA subection. +struct FrameData { + ulittle32_t RvaStart; + ulittle32_t CodeSize; + ulittle32_t LocalSize; + ulittle32_t ParamsSize; + ulittle32_t MaxStackSize; + ulittle32_t FrameFunc; + ulittle16_t PrologSize; + ulittle16_t SavedRegsSize; + ulittle32_t Flags; + enum : uint32_t { + HasSEH = 1 << 0, + HasEH = 1 << 1, + IsFunctionStart = 1 << 2, + }; +}; + + +} // namespace codeview +} // namespace llvm + +#endif // LLVM_READOBJ_CODEVIEW_H diff --git a/tools/llvm-readobj/ELFDumper.cpp b/tools/llvm-readobj/ELFDumper.cpp index be84f3c0f1bba..06fbe8d3fccba 100644 --- a/tools/llvm-readobj/ELFDumper.cpp +++ b/tools/llvm-readobj/ELFDumper.cpp @@ -12,13 +12,12 @@ /// //===----------------------------------------------------------------------===// -#include "llvm-readobj.h" #include "ARMAttributeParser.h" #include "ARMEHABIPrinter.h" #include "Error.h" #include "ObjDumper.h" #include "StackMapPrinter.h" -#include "StreamWriter.h" +#include "llvm-readobj.h" #include "llvm/ADT/Optional.h" #include "llvm/ADT/SmallString.h" #include "llvm/ADT/StringExtras.h" @@ -26,8 +25,10 @@ #include "llvm/Support/ARMBuildAttributes.h" #include "llvm/Support/Compiler.h" #include "llvm/Support/Format.h" +#include "llvm/Support/FormattedStream.h" #include "llvm/Support/MathExtras.h" #include "llvm/Support/MipsABIFlags.h" +#include "llvm/Support/ScopedPrinter.h" #include "llvm/Support/raw_ostream.h" using namespace llvm; @@ -37,12 +38,66 @@ using namespace ELF; #define LLVM_READOBJ_ENUM_CASE(ns, enum) \ case ns::enum: return #enum; +#define ENUM_ENT(enum, altName) \ + { #enum, altName, ELF::enum } + +#define ENUM_ENT_1(enum) \ + { #enum, #enum, ELF::enum } + +#define LLVM_READOBJ_PHDR_ENUM(ns, enum) \ + case ns::enum: \ + return std::string(#enum).substr(3); + +#define TYPEDEF_ELF_TYPES(ELFT) \ + typedef ELFFile<ELFT> ELFO; \ + typedef typename ELFO::Elf_Shdr Elf_Shdr; \ + typedef typename ELFO::Elf_Sym Elf_Sym; \ + typedef typename ELFO::Elf_Dyn Elf_Dyn; \ + typedef typename ELFO::Elf_Dyn_Range Elf_Dyn_Range; \ + typedef typename ELFO::Elf_Rel Elf_Rel; \ + typedef typename ELFO::Elf_Rela Elf_Rela; \ + typedef typename ELFO::Elf_Rela_Range Elf_Rela_Range; \ + typedef typename ELFO::Elf_Phdr Elf_Phdr; \ + typedef typename ELFO::Elf_Half Elf_Half; \ + typedef typename ELFO::Elf_Ehdr Elf_Ehdr; \ + typedef typename ELFO::Elf_Word Elf_Word; \ + typedef typename ELFO::Elf_Hash Elf_Hash; \ + typedef typename ELFO::Elf_GnuHash Elf_GnuHash; \ + typedef typename ELFO::uintX_t uintX_t; + namespace { +template <class ELFT> class DumpStyle; + +/// Represents a contiguous uniform range in the file. We cannot just create a +/// range directly because when creating one of these from the .dynamic table +/// the size, entity size and virtual address are different entries in arbitrary +/// order (DT_REL, DT_RELSZ, DT_RELENT for example). +struct DynRegionInfo { + DynRegionInfo() : Addr(nullptr), Size(0), EntSize(0) {} + DynRegionInfo(const void *A, uint64_t S, uint64_t ES) + : Addr(A), Size(S), EntSize(ES) {} + /// \brief Address in current address space. + const void *Addr; + /// \brief Size in bytes of the region. + uint64_t Size; + /// \brief Size of each entity in the region. + uint64_t EntSize; + + template <typename Type> ArrayRef<Type> getAsArrayRef() const { + const Type *Start = reinterpret_cast<const Type *>(Addr); + if (!Start) + return {Start, Start}; + if (EntSize != sizeof(Type) || Size % EntSize) + reportError("Invalid entity size"); + return {Start, Start + (Size / EntSize)}; + } +}; + template<typename ELFT> class ELFDumper : public ObjDumper { public: - ELFDumper(const ELFFile<ELFT> *Obj, StreamWriter &Writer); + ELFDumper(const ELFFile<ELFT> *Obj, ScopedPrinter &Writer); void printFileHeaders() override; void printSections() override; @@ -59,22 +114,29 @@ public: void printGnuHashTable() override; void printLoadName() override; void printVersionInfo() override; + void printGroupSections() override; void printAttributes() override; void printMipsPLTGOT() override; void printMipsABIFlags() override; void printMipsReginfo() override; + void printMipsOptions() override; void printStackMap() const override; + void printHashHistogram() override; + private: + std::unique_ptr<DumpStyle<ELFT>> ELFDumperStyle; typedef ELFFile<ELFT> ELFO; typedef typename ELFO::Elf_Shdr Elf_Shdr; typedef typename ELFO::Elf_Sym Elf_Sym; + typedef typename ELFO::Elf_Sym_Range Elf_Sym_Range; typedef typename ELFO::Elf_Dyn Elf_Dyn; typedef typename ELFO::Elf_Dyn_Range Elf_Dyn_Range; typedef typename ELFO::Elf_Rel Elf_Rel; typedef typename ELFO::Elf_Rela Elf_Rela; + typedef typename ELFO::Elf_Rel_Range Elf_Rel_Range; typedef typename ELFO::Elf_Rela_Range Elf_Rela_Range; typedef typename ELFO::Elf_Phdr Elf_Phdr; typedef typename ELFO::Elf_Half Elf_Half; @@ -89,55 +151,44 @@ private: typedef typename ELFO::Elf_Verdef Elf_Verdef; typedef typename ELFO::Elf_Verdaux Elf_Verdaux; - /// \brief Represents a region described by entries in the .dynamic table. - struct DynRegionInfo { - DynRegionInfo() : Addr(nullptr), Size(0), EntSize(0) {} - /// \brief Address in current address space. - const void *Addr; - /// \brief Size in bytes of the region. - uintX_t Size; - /// \brief Size of each entity in the region. - uintX_t EntSize; - }; + DynRegionInfo checkDRI(DynRegionInfo DRI) { + if (DRI.Addr < Obj->base() || + (const uint8_t *)DRI.Addr + DRI.Size > Obj->base() + Obj->getBufSize()) + error(llvm::object::object_error::parse_failed); + return DRI; + } + + DynRegionInfo createDRIFrom(const Elf_Phdr *P, uintX_t EntSize) { + return checkDRI({Obj->base() + P->p_offset, P->p_filesz, EntSize}); + } - void printSymbolsHelper(bool IsDynamic); - void printSymbol(const Elf_Sym *Symbol, const Elf_Shdr *SymTab, - StringRef StrTable, bool IsDynamic); + DynRegionInfo createDRIFrom(const Elf_Shdr *S) { + return checkDRI({Obj->base() + S->sh_offset, S->sh_size, S->sh_entsize}); + } + + void parseDynamicTable(ArrayRef<const Elf_Phdr *> LoadSegments); - void printRelocations(const Elf_Shdr *Sec); - void printRelocation(Elf_Rela Rel, const Elf_Shdr *SymTab); void printValue(uint64_t Type, uint64_t Value); - const Elf_Rela *dyn_rela_begin() const; - const Elf_Rela *dyn_rela_end() const; - Elf_Rela_Range dyn_relas() const; StringRef getDynamicString(uint64_t Offset) const; - const Elf_Dyn *dynamic_table_begin() const { - ErrorOr<const Elf_Dyn *> Ret = Obj->dynamic_table_begin(DynamicProgHeader); - error(Ret.getError()); - return *Ret; - } - const Elf_Dyn *dynamic_table_end() const { - ErrorOr<const Elf_Dyn *> Ret = Obj->dynamic_table_end(DynamicProgHeader); - error(Ret.getError()); - return *Ret; - } StringRef getSymbolVersion(StringRef StrTab, const Elf_Sym *symb, - bool &IsDefault); - void LoadVersionMap(); + bool &IsDefault) const; + void LoadVersionMap() const; void LoadVersionNeeds(const Elf_Shdr *ec) const; void LoadVersionDefs(const Elf_Shdr *sec) const; const ELFO *Obj; + DynRegionInfo DynRelRegion; DynRegionInfo DynRelaRegion; - const Elf_Phdr *DynamicProgHeader = nullptr; + DynRegionInfo DynPLTRelRegion; + DynRegionInfo DynSymRegion; + DynRegionInfo DynamicTable; StringRef DynamicStringTable; - const Elf_Sym *DynSymStart = nullptr; StringRef SOName; const Elf_Hash *HashTable = nullptr; const Elf_GnuHash *GnuHashTable = nullptr; - const Elf_Shdr *DotDynSymSec = nullptr; const Elf_Shdr *DotSymtabSec = nullptr; + StringRef DynSymtabName; ArrayRef<Elf_Word> ShndxTable; const Elf_Shdr *dot_gnu_version_sec = nullptr; // .gnu.version @@ -169,40 +220,176 @@ private: public: Elf_Dyn_Range dynamic_table() const { - ErrorOr<Elf_Dyn_Range> Ret = Obj->dynamic_table(DynamicProgHeader); - error(Ret.getError()); - return *Ret; + return DynamicTable.getAsArrayRef<Elf_Dyn>(); + } + + Elf_Sym_Range dynamic_symbols() const { + return DynSymRegion.getAsArrayRef<Elf_Sym>(); } + Elf_Rel_Range dyn_rels() const; + Elf_Rela_Range dyn_relas() const; std::string getFullSymbolName(const Elf_Sym *Symbol, StringRef StrTable, - bool IsDynamic); - const Elf_Shdr *getDotDynSymSec() const { return DotDynSymSec; } + bool IsDynamic) const; + + void printSymbolsHelper(bool IsDynamic) const; const Elf_Shdr *getDotSymtabSec() const { return DotSymtabSec; } - ArrayRef<Elf_Word> getShndxTable() { return ShndxTable; } + ArrayRef<Elf_Word> getShndxTable() const { return ShndxTable; } + StringRef getDynamicStringTable() const { return DynamicStringTable; } + const DynRegionInfo &getDynRelRegion() const { return DynRelRegion; } + const DynRegionInfo &getDynRelaRegion() const { return DynRelaRegion; } + const DynRegionInfo &getDynPLTRelRegion() const { return DynPLTRelRegion; } + const Elf_Hash *getHashTable() const { return HashTable; } + const Elf_GnuHash *getGnuHashTable() const { return GnuHashTable; } }; -template <class T> T errorOrDefault(ErrorOr<T> Val, T Default = T()) { - if (!Val) { - error(Val.getError()); - return Default; +template <class ELFT> +void ELFDumper<ELFT>::printSymbolsHelper(bool IsDynamic) const { + StringRef StrTable, SymtabName; + size_t Entries = 0; + Elf_Sym_Range Syms(nullptr, nullptr); + if (IsDynamic) { + StrTable = DynamicStringTable; + Syms = dynamic_symbols(); + SymtabName = DynSymtabName; + if (DynSymRegion.Addr) + Entries = DynSymRegion.Size / DynSymRegion.EntSize; + } else { + if (!DotSymtabSec) + return; + StrTable = unwrapOrError(Obj->getStringTableForSymtab(*DotSymtabSec)); + Syms = Obj->symbols(DotSymtabSec); + SymtabName = unwrapOrError(Obj->getSectionName(DotSymtabSec)); + Entries = DotSymtabSec->getEntityCount(); + } + if (Syms.begin() == Syms.end()) + return; + ELFDumperStyle->printSymtabMessage(Obj, SymtabName, Entries); + for (const auto &Sym : Syms) + ELFDumperStyle->printSymbol(Obj, &Sym, Syms.begin(), StrTable, IsDynamic); +} + +template <typename ELFT> class DumpStyle { +public: + using Elf_Shdr = typename ELFFile<ELFT>::Elf_Shdr; + using Elf_Sym = typename ELFFile<ELFT>::Elf_Sym; + + DumpStyle(ELFDumper<ELFT> *Dumper) : Dumper(Dumper) {} + virtual ~DumpStyle() {} + virtual void printFileHeaders(const ELFFile<ELFT> *Obj) = 0; + virtual void printGroupSections(const ELFFile<ELFT> *Obj) = 0; + virtual void printRelocations(const ELFFile<ELFT> *Obj) = 0; + virtual void printSections(const ELFFile<ELFT> *Obj) = 0; + virtual void printSymbols(const ELFFile<ELFT> *Obj) = 0; + virtual void printDynamicSymbols(const ELFFile<ELFT> *Obj) = 0; + virtual void printDynamicRelocations(const ELFFile<ELFT> *Obj) = 0; + virtual void printSymtabMessage(const ELFFile<ELFT> *obj, StringRef Name, + size_t Offset) { + return; } + virtual void printSymbol(const ELFFile<ELFT> *Obj, const Elf_Sym *Symbol, + const Elf_Sym *FirstSym, StringRef StrTable, + bool IsDynamic) = 0; + virtual void printProgramHeaders(const ELFFile<ELFT> *Obj) = 0; + virtual void printHashHistogram(const ELFFile<ELFT> *Obj) = 0; + const ELFDumper<ELFT> *dumper() const { return Dumper; } +private: + const ELFDumper<ELFT> *Dumper; +}; + +template <typename ELFT> class GNUStyle : public DumpStyle<ELFT> { + formatted_raw_ostream OS; +public: + TYPEDEF_ELF_TYPES(ELFT) + GNUStyle(ScopedPrinter &W, ELFDumper<ELFT> *Dumper) + : DumpStyle<ELFT>(Dumper), OS(W.getOStream()) {} + void printFileHeaders(const ELFO *Obj) override; + void printGroupSections(const ELFFile<ELFT> *Obj) override; + void printRelocations(const ELFO *Obj) override; + void printSections(const ELFO *Obj) override; + void printSymbols(const ELFO *Obj) override; + void printDynamicSymbols(const ELFO *Obj) override; + void printDynamicRelocations(const ELFO *Obj) override; + virtual void printSymtabMessage(const ELFO *Obj, StringRef Name, + size_t Offset) override; + void printProgramHeaders(const ELFO *Obj) override; + void printHashHistogram(const ELFFile<ELFT> *Obj) override; + +private: + struct Field { + StringRef Str; + unsigned Column; + Field(StringRef S, unsigned Col) : Str(S), Column(Col) {} + Field(unsigned Col) : Str(""), Column(Col) {} + }; + + template <typename T, typename TEnum> + std::string printEnum(T Value, ArrayRef<EnumEntry<TEnum>> EnumValues) { + for (const auto &EnumItem : EnumValues) + if (EnumItem.Value == Value) + return EnumItem.AltName; + return to_hexString(Value, false); + } + + formatted_raw_ostream &printField(struct Field F) { + if (F.Column != 0) + OS.PadToColumn(F.Column); + OS << F.Str; + OS.flush(); + return OS; + } + void printRelocation(const ELFO *Obj, const Elf_Shdr *SymTab, + const Elf_Rela &R, bool IsRela); + void printSymbol(const ELFO *Obj, const Elf_Sym *Symbol, const Elf_Sym *First, + StringRef StrTable, bool IsDynamic) override; + std::string getSymbolSectionNdx(const ELFO *Obj, const Elf_Sym *Symbol, + const Elf_Sym *FirstSym); + void printDynamicRelocation(const ELFO *Obj, Elf_Rela R, bool IsRela); + bool checkTLSSections(const Elf_Phdr &Phdr, const Elf_Shdr &Sec); + bool checkoffsets(const Elf_Phdr &Phdr, const Elf_Shdr &Sec); + bool checkVMA(const Elf_Phdr &Phdr, const Elf_Shdr &Sec); + bool checkPTDynamic(const Elf_Phdr &Phdr, const Elf_Shdr &Sec); +}; + +template <typename ELFT> class LLVMStyle : public DumpStyle<ELFT> { +public: + TYPEDEF_ELF_TYPES(ELFT) + LLVMStyle(ScopedPrinter &W, ELFDumper<ELFT> *Dumper) + : DumpStyle<ELFT>(Dumper), W(W) {} + + void printFileHeaders(const ELFO *Obj) override; + void printGroupSections(const ELFFile<ELFT> *Obj) override; + void printRelocations(const ELFO *Obj) override; + void printRelocations(const Elf_Shdr *Sec, const ELFO *Obj); + void printSections(const ELFO *Obj) override; + void printSymbols(const ELFO *Obj) override; + void printDynamicSymbols(const ELFO *Obj) override; + void printDynamicRelocations(const ELFO *Obj) override; + void printProgramHeaders(const ELFO *Obj) override; + void printHashHistogram(const ELFFile<ELFT> *Obj) override; + +private: + void printRelocation(const ELFO *Obj, Elf_Rela Rel, const Elf_Shdr *SymTab); + void printDynamicRelocation(const ELFO *Obj, Elf_Rela Rel); + void printSymbol(const ELFO *Obj, const Elf_Sym *Symbol, const Elf_Sym *First, + StringRef StrTable, bool IsDynamic) override; + ScopedPrinter &W; +}; - return *Val; -} } // namespace namespace llvm { template <class ELFT> static std::error_code createELFDumper(const ELFFile<ELFT> *Obj, - StreamWriter &Writer, + ScopedPrinter &Writer, std::unique_ptr<ObjDumper> &Result) { Result.reset(new ELFDumper<ELFT>(Obj, Writer)); return readobj_error::success; } std::error_code createELFDumper(const object::ObjectFile *Obj, - StreamWriter &Writer, + ScopedPrinter &Writer, std::unique_ptr<ObjDumper> &Result) { // Little-endian 32-bit if (const ELF32LEObjectFile *ELFObj = dyn_cast<ELF32LEObjectFile>(Obj)) @@ -284,9 +471,9 @@ void ELFDumper<ELFT>::LoadVersionDefs(const Elf_Shdr *sec) const { } } -template <class ELFT> void ELFDumper<ELFT>::LoadVersionMap() { +template <class ELFT> void ELFDumper<ELFT>::LoadVersionMap() const { // If there is no dynamic symtab or version table, there is nothing to do. - if (!DynSymStart || !dot_gnu_version_sec) + if (!DynSymRegion.Addr || !dot_gnu_version_sec) return; // Has the VersionMap already been loaded? @@ -305,99 +492,157 @@ template <class ELFT> void ELFDumper<ELFT>::LoadVersionMap() { LoadVersionNeeds(dot_gnu_version_r_sec); } - template <typename ELFO, class ELFT> -static void printVersionSymbolSection(ELFDumper<ELFT> *Dumper, - const ELFO *Obj, +static void printVersionSymbolSection(ELFDumper<ELFT> *Dumper, const ELFO *Obj, const typename ELFO::Elf_Shdr *Sec, - StreamWriter &W) { + ScopedPrinter &W) { DictScope SS(W, "Version symbols"); if (!Sec) return; - StringRef Name = errorOrDefault(Obj->getSectionName(Sec)); + StringRef Name = unwrapOrError(Obj->getSectionName(Sec)); W.printNumber("Section Name", Name, Sec->sh_name); W.printHex("Address", Sec->sh_addr); W.printHex("Offset", Sec->sh_offset); W.printNumber("Link", Sec->sh_link); - const typename ELFO::Elf_Shdr *DynSymSec = Dumper->getDotDynSymSec(); const uint8_t *P = (const uint8_t *)Obj->base() + Sec->sh_offset; - ErrorOr<StringRef> StrTableOrErr = - Obj->getStringTableForSymtab(*DynSymSec); - error(StrTableOrErr.getError()); + StringRef StrTable = Dumper->getDynamicStringTable(); // Same number of entries in the dynamic symbol table (DT_SYMTAB). ListScope Syms(W, "Symbols"); - for (const typename ELFO::Elf_Sym &Sym : Obj->symbols(DynSymSec)) { + for (const typename ELFO::Elf_Sym &Sym : Dumper->dynamic_symbols()) { DictScope S(W, "Symbol"); std::string FullSymbolName = - Dumper->getFullSymbolName(&Sym, *StrTableOrErr, true /* IsDynamic */); + Dumper->getFullSymbolName(&Sym, StrTable, true /* IsDynamic */); W.printNumber("Version", *P); W.printString("Name", FullSymbolName); P += sizeof(typename ELFO::Elf_Half); } } +static const EnumEntry<unsigned> SymVersionFlags[] = { + {"Base", "BASE", VER_FLG_BASE}, + {"Weak", "WEAK", VER_FLG_WEAK}, + {"Info", "INFO", VER_FLG_INFO}}; + template <typename ELFO, class ELFT> static void printVersionDefinitionSection(ELFDumper<ELFT> *Dumper, const ELFO *Obj, const typename ELFO::Elf_Shdr *Sec, - StreamWriter &W) { - DictScope SD(W, "Version definition"); + ScopedPrinter &W) { + typedef typename ELFO::Elf_Verdef VerDef; + typedef typename ELFO::Elf_Verdaux VerdAux; + + DictScope SD(W, "SHT_GNU_verdef"); if (!Sec) return; - StringRef Name = errorOrDefault(Obj->getSectionName(Sec)); - W.printNumber("Section Name", Name, Sec->sh_name); - W.printHex("Address", Sec->sh_addr); - W.printHex("Offset", Sec->sh_offset); - W.printNumber("Link", Sec->sh_link); - unsigned verdef_entries = 0; // The number of entries in the section SHT_GNU_verdef // is determined by DT_VERDEFNUM tag. + unsigned VerDefsNum = 0; for (const typename ELFO::Elf_Dyn &Dyn : Dumper->dynamic_table()) { if (Dyn.d_tag == DT_VERDEFNUM) - verdef_entries = Dyn.d_un.d_val; + VerDefsNum = Dyn.d_un.d_val; } const uint8_t *SecStartAddress = (const uint8_t *)Obj->base() + Sec->sh_offset; const uint8_t *SecEndAddress = SecStartAddress + Sec->sh_size; const uint8_t *P = SecStartAddress; - ErrorOr<const typename ELFO::Elf_Shdr *> StrTabOrErr = - Obj->getSection(Sec->sh_link); - error(StrTabOrErr.getError()); + const typename ELFO::Elf_Shdr *StrTab = + unwrapOrError(Obj->getSection(Sec->sh_link)); - ListScope Entries(W, "Entries"); - for (unsigned i = 0; i < verdef_entries; ++i) { - if (P + sizeof(typename ELFO::Elf_Verdef) > SecEndAddress) + while (VerDefsNum--) { + if (P + sizeof(VerDef) > SecEndAddress) report_fatal_error("invalid offset in the section"); - auto *VD = reinterpret_cast<const typename ELFO::Elf_Verdef *>(P); - DictScope Entry(W, "Entry"); - W.printHex("Offset", (uintptr_t)P - (uintptr_t)SecStartAddress); - W.printNumber("Rev", VD->vd_version); - // FIXME: print something more readable. - W.printNumber("Flags", VD->vd_flags); + + auto *VD = reinterpret_cast<const VerDef *>(P); + DictScope Def(W, "Definition"); + W.printNumber("Version", VD->vd_version); + W.printEnum("Flags", VD->vd_flags, makeArrayRef(SymVersionFlags)); W.printNumber("Index", VD->vd_ndx); - W.printNumber("Cnt", VD->vd_cnt); - W.printString("Name", StringRef((const char *)(Obj->base() + - (*StrTabOrErr)->sh_offset + - VD->getAux()->vda_name))); + W.printNumber("Hash", VD->vd_hash); + W.printString("Name", + StringRef((const char *)(Obj->base() + StrTab->sh_offset + + VD->getAux()->vda_name))); + if (!VD->vd_cnt) + report_fatal_error("at least one definition string must exist"); + if (VD->vd_cnt > 2) + report_fatal_error("more than one predecessor is not expected"); + + if (VD->vd_cnt == 2) { + const uint8_t *PAux = P + VD->vd_aux + VD->getAux()->vda_next; + const VerdAux *Aux = reinterpret_cast<const VerdAux *>(PAux); + W.printString("Predecessor", + StringRef((const char *)(Obj->base() + StrTab->sh_offset + + Aux->vda_name))); + } + P += VD->vd_next; } } +template <typename ELFO, class ELFT> +static void printVersionDependencySection(ELFDumper<ELFT> *Dumper, + const ELFO *Obj, + const typename ELFO::Elf_Shdr *Sec, + ScopedPrinter &W) { + typedef typename ELFO::Elf_Verneed VerNeed; + typedef typename ELFO::Elf_Vernaux VernAux; + + DictScope SD(W, "SHT_GNU_verneed"); + if (!Sec) + return; + + unsigned VerNeedNum = 0; + for (const typename ELFO::Elf_Dyn &Dyn : Dumper->dynamic_table()) + if (Dyn.d_tag == DT_VERNEEDNUM) + VerNeedNum = Dyn.d_un.d_val; + + const uint8_t *SecData = (const uint8_t *)Obj->base() + Sec->sh_offset; + const typename ELFO::Elf_Shdr *StrTab = + unwrapOrError(Obj->getSection(Sec->sh_link)); + + const uint8_t *P = SecData; + for (unsigned I = 0; I < VerNeedNum; ++I) { + const VerNeed *Need = reinterpret_cast<const VerNeed *>(P); + DictScope Entry(W, "Dependency"); + W.printNumber("Version", Need->vn_version); + W.printNumber("Count", Need->vn_cnt); + W.printString("FileName", + StringRef((const char *)(Obj->base() + StrTab->sh_offset + + Need->vn_file))); + + const uint8_t *PAux = P + Need->vn_aux; + for (unsigned J = 0; J < Need->vn_cnt; ++J) { + const VernAux *Aux = reinterpret_cast<const VernAux *>(PAux); + DictScope Entry(W, "Entry"); + W.printNumber("Hash", Aux->vna_hash); + W.printEnum("Flags", Aux->vna_flags, makeArrayRef(SymVersionFlags)); + W.printNumber("Index", Aux->vna_other); + W.printString("Name", + StringRef((const char *)(Obj->base() + StrTab->sh_offset + + Aux->vna_name))); + PAux += Aux->vna_next; + } + P += Need->vn_next; + } +} + template <typename ELFT> void ELFDumper<ELFT>::printVersionInfo() { // Dump version symbol section. printVersionSymbolSection(this, Obj, dot_gnu_version_sec, W); // Dump version definition section. printVersionDefinitionSection(this, Obj, dot_gnu_version_d_sec, W); + + // Dump version dependency section. + printVersionDependencySection(this, Obj, dot_gnu_version_r_sec, W); } template <typename ELFT> StringRef ELFDumper<ELFT>::getSymbolVersion(StringRef StrTab, const Elf_Sym *symb, - bool &IsDefault) { + bool &IsDefault) const { // This is a dynamic symbol. Look in the GNU symbol version table. if (!dot_gnu_version_sec) { // No version table. @@ -407,7 +652,7 @@ StringRef ELFDumper<ELFT>::getSymbolVersion(StringRef StrTab, // Determine the position in the symbol table of this entry. size_t entry_index = (reinterpret_cast<uintptr_t>(symb) - - reinterpret_cast<uintptr_t>(DynSymStart)) / + reinterpret_cast<uintptr_t>(DynSymRegion.Addr)) / sizeof(Elf_Sym); // Get the corresponding version index entry @@ -446,8 +691,8 @@ StringRef ELFDumper<ELFT>::getSymbolVersion(StringRef StrTab, template <typename ELFT> std::string ELFDumper<ELFT>::getFullSymbolName(const Elf_Sym *Symbol, StringRef StrTable, - bool IsDynamic) { - StringRef SymbolName = errorOrDefault(Symbol->getName(StrTable)); + bool IsDynamic) const { + StringRef SymbolName = unwrapOrError(Symbol->getName(StrTable)); if (!IsDynamic) return SymbolName; @@ -463,7 +708,7 @@ std::string ELFDumper<ELFT>::getFullSymbolName(const Elf_Sym *Symbol, template <typename ELFO> static void getSectionNameIndex(const ELFO &Obj, const typename ELFO::Elf_Sym *Symbol, - const typename ELFO::Elf_Shdr *SymTab, + const typename ELFO::Elf_Sym *FirstSym, ArrayRef<typename ELFO::Elf_Word> ShndxTable, StringRef &SectionName, unsigned &SectionIndex) { SectionIndex = Symbol->st_shndx; @@ -482,18 +727,18 @@ getSectionNameIndex(const ELFO &Obj, const typename ELFO::Elf_Sym *Symbol, else { if (SectionIndex == SHN_XINDEX) SectionIndex = - Obj.getExtendedSymbolTableIndex(Symbol, SymTab, ShndxTable); - ErrorOr<const typename ELFO::Elf_Shdr *> Sec = Obj.getSection(SectionIndex); - error(Sec.getError()); - SectionName = errorOrDefault(Obj.getSectionName(*Sec)); + Obj.getExtendedSymbolTableIndex(Symbol, FirstSym, ShndxTable); + const typename ELFO::Elf_Shdr *Sec = + unwrapOrError(Obj.getSection(SectionIndex)); + SectionName = unwrapOrError(Obj.getSectionName(Sec)); } } template <class ELFO> -static const typename ELFO::Elf_Shdr *findSectionByAddress(const ELFO *Obj, - uint64_t Addr) { +static const typename ELFO::Elf_Shdr * +findNotEmptySectionByAddress(const ELFO *Obj, uint64_t Addr) { for (const auto &Shdr : Obj->sections()) - if (Shdr.sh_addr == Addr) + if (Shdr.sh_addr == Addr && Shdr.sh_size > 0) return &Shdr; return nullptr; } @@ -502,233 +747,239 @@ template <class ELFO> static const typename ELFO::Elf_Shdr *findSectionByName(const ELFO &Obj, StringRef Name) { for (const auto &Shdr : Obj.sections()) { - if (Name == errorOrDefault(Obj.getSectionName(&Shdr))) + if (Name == unwrapOrError(Obj.getSectionName(&Shdr))) return &Shdr; } return nullptr; } static const EnumEntry<unsigned> ElfClass[] = { - { "None", ELF::ELFCLASSNONE }, - { "32-bit", ELF::ELFCLASS32 }, - { "64-bit", ELF::ELFCLASS64 }, + {"None", "none", ELF::ELFCLASSNONE}, + {"32-bit", "ELF32", ELF::ELFCLASS32}, + {"64-bit", "ELF64", ELF::ELFCLASS64}, }; static const EnumEntry<unsigned> ElfDataEncoding[] = { - { "None", ELF::ELFDATANONE }, - { "LittleEndian", ELF::ELFDATA2LSB }, - { "BigEndian", ELF::ELFDATA2MSB }, + {"None", "none", ELF::ELFDATANONE}, + {"LittleEndian", "2's complement, little endian", ELF::ELFDATA2LSB}, + {"BigEndian", "2's complement, big endian", ELF::ELFDATA2MSB}, }; static const EnumEntry<unsigned> ElfObjectFileType[] = { - { "None", ELF::ET_NONE }, - { "Relocatable", ELF::ET_REL }, - { "Executable", ELF::ET_EXEC }, - { "SharedObject", ELF::ET_DYN }, - { "Core", ELF::ET_CORE }, + {"None", "NONE (none)", ELF::ET_NONE}, + {"Relocatable", "REL (Relocatable file)", ELF::ET_REL}, + {"Executable", "EXEC (Executable file)", ELF::ET_EXEC}, + {"SharedObject", "DYN (Shared object file)", ELF::ET_DYN}, + {"Core", "CORE (Core file)", ELF::ET_CORE}, }; static const EnumEntry<unsigned> ElfOSABI[] = { - { "SystemV", ELF::ELFOSABI_NONE }, - { "HPUX", ELF::ELFOSABI_HPUX }, - { "NetBSD", ELF::ELFOSABI_NETBSD }, - { "GNU/Linux", ELF::ELFOSABI_LINUX }, - { "GNU/Hurd", ELF::ELFOSABI_HURD }, - { "Solaris", ELF::ELFOSABI_SOLARIS }, - { "AIX", ELF::ELFOSABI_AIX }, - { "IRIX", ELF::ELFOSABI_IRIX }, - { "FreeBSD", ELF::ELFOSABI_FREEBSD }, - { "TRU64", ELF::ELFOSABI_TRU64 }, - { "Modesto", ELF::ELFOSABI_MODESTO }, - { "OpenBSD", ELF::ELFOSABI_OPENBSD }, - { "OpenVMS", ELF::ELFOSABI_OPENVMS }, - { "NSK", ELF::ELFOSABI_NSK }, - { "AROS", ELF::ELFOSABI_AROS }, - { "FenixOS", ELF::ELFOSABI_FENIXOS }, - { "CloudABI", ELF::ELFOSABI_CLOUDABI }, - { "C6000_ELFABI", ELF::ELFOSABI_C6000_ELFABI }, - { "C6000_LINUX" , ELF::ELFOSABI_C6000_LINUX }, - { "ARM", ELF::ELFOSABI_ARM }, - { "Standalone" , ELF::ELFOSABI_STANDALONE } + {"SystemV", "UNIX - System V", ELF::ELFOSABI_NONE}, + {"HPUX", "UNIX - HP-UX", ELF::ELFOSABI_HPUX}, + {"NetBSD", "UNIX - NetBSD", ELF::ELFOSABI_NETBSD}, + {"GNU/Linux", "UNIX - GNU", ELF::ELFOSABI_LINUX}, + {"GNU/Hurd", "GNU/Hurd", ELF::ELFOSABI_HURD}, + {"Solaris", "UNIX - Solaris", ELF::ELFOSABI_SOLARIS}, + {"AIX", "UNIX - AIX", ELF::ELFOSABI_AIX}, + {"IRIX", "UNIX - IRIX", ELF::ELFOSABI_IRIX}, + {"FreeBSD", "UNIX - FreeBSD", ELF::ELFOSABI_FREEBSD}, + {"TRU64", "UNIX - TRU64", ELF::ELFOSABI_TRU64}, + {"Modesto", "Novell - Modesto", ELF::ELFOSABI_MODESTO}, + {"OpenBSD", "UNIX - OpenBSD", ELF::ELFOSABI_OPENBSD}, + {"OpenVMS", "VMS - OpenVMS", ELF::ELFOSABI_OPENVMS}, + {"NSK", "HP - Non-Stop Kernel", ELF::ELFOSABI_NSK}, + {"AROS", "AROS", ELF::ELFOSABI_AROS}, + {"FenixOS", "FenixOS", ELF::ELFOSABI_FENIXOS}, + {"CloudABI", "CloudABI", ELF::ELFOSABI_CLOUDABI}, + {"C6000_ELFABI", "Bare-metal C6000", ELF::ELFOSABI_C6000_ELFABI}, + {"C6000_LINUX", "Linux C6000", ELF::ELFOSABI_C6000_LINUX}, + {"ARM", "ARM", ELF::ELFOSABI_ARM}, + {"Standalone", "Standalone App", ELF::ELFOSABI_STANDALONE} }; static const EnumEntry<unsigned> ElfMachineType[] = { - LLVM_READOBJ_ENUM_ENT(ELF, EM_NONE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_M32 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SPARC ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_386 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_68K ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_88K ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_IAMCU ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_860 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MIPS ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_S370 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MIPS_RS3_LE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_PARISC ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_VPP500 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SPARC32PLUS ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_960 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_PPC ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_PPC64 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_S390 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SPU ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_V800 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_FR20 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_RH32 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_RCE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ARM ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ALPHA ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SH ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SPARCV9 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TRICORE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ARC ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_H8_300 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_H8_300H ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_H8S ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_H8_500 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_IA_64 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MIPS_X ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_COLDFIRE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_68HC12 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MMA ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_PCP ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_NCPU ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_NDR1 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_STARCORE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ME16 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ST100 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TINYJ ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_X86_64 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_PDSP ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_PDP10 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_PDP11 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_FX66 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ST9PLUS ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ST7 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_68HC16 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_68HC11 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_68HC08 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_68HC05 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SVX ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ST19 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_VAX ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_CRIS ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_JAVELIN ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_FIREPATH ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ZSP ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MMIX ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_HUANY ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_PRISM ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_AVR ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_FR30 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_D10V ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_D30V ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_V850 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_M32R ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MN10300 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MN10200 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_PJ ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_OPENRISC ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ARC_COMPACT ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_XTENSA ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_VIDEOCORE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TMM_GPP ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_NS32K ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TPC ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SNP1K ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ST200 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_IP2K ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MAX ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_CR ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_F2MC16 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MSP430 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_BLACKFIN ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SE_C33 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SEP ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ARCA ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_UNICORE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_EXCESS ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_DXP ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ALTERA_NIOS2 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_CRX ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_XGATE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_C166 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_M16C ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_DSPIC30F ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_CE ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_M32C ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TSK3000 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_RS08 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SHARC ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ECOG2 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SCORE7 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_DSP24 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_VIDEOCORE3 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_LATTICEMICO32), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SE_C17 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TI_C6000 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TI_C2000 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TI_C5500 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MMDSP_PLUS ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_CYPRESS_M8C ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_R32C ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TRIMEDIA ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_HEXAGON ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_8051 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_STXP7X ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_NDS32 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ECOG1 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ECOG1X ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MAXQ30 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_XIMO16 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MANIK ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_CRAYNV2 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_RX ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_METAG ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_MCST_ELBRUS ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ECOG16 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_CR16 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ETPU ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_SLE9X ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_L10M ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_K10M ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_AARCH64 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_AVR32 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_STM8 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TILE64 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TILEPRO ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_CUDA ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_TILEGX ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_CLOUDSHIELD ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_COREA_1ST ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_COREA_2ND ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_ARC_COMPACT2 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_OPEN8 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_RL78 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_VIDEOCORE5 ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_78KOR ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_56800EX ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_AMDGPU ), - LLVM_READOBJ_ENUM_ENT(ELF, EM_WEBASSEMBLY ), + ENUM_ENT(EM_NONE, "None"), + ENUM_ENT(EM_M32, "WE32100"), + ENUM_ENT(EM_SPARC, "Sparc"), + ENUM_ENT(EM_386, "Intel 80386"), + ENUM_ENT(EM_68K, "MC68000"), + ENUM_ENT(EM_88K, "MC88000"), + ENUM_ENT(EM_IAMCU, "EM_IAMCU"), + ENUM_ENT(EM_860, "Intel 80860"), + ENUM_ENT(EM_MIPS, "MIPS R3000"), + ENUM_ENT(EM_S370, "IBM System/370"), + ENUM_ENT(EM_MIPS_RS3_LE, "MIPS R3000 little-endian"), + ENUM_ENT(EM_PARISC, "HPPA"), + ENUM_ENT(EM_VPP500, "Fujitsu VPP500"), + ENUM_ENT(EM_SPARC32PLUS, "Sparc v8+"), + ENUM_ENT(EM_960, "Intel 80960"), + ENUM_ENT(EM_PPC, "PowerPC"), + ENUM_ENT(EM_PPC64, "PowerPC64"), + ENUM_ENT(EM_S390, "IBM S/390"), + ENUM_ENT(EM_SPU, "SPU"), + ENUM_ENT(EM_V800, "NEC V800 series"), + ENUM_ENT(EM_FR20, "Fujistsu FR20"), + ENUM_ENT(EM_RH32, "TRW RH-32"), + ENUM_ENT(EM_RCE, "Motorola RCE"), + ENUM_ENT(EM_ARM, "ARM"), + ENUM_ENT(EM_ALPHA, "EM_ALPHA"), + ENUM_ENT(EM_SH, "Hitachi SH"), + ENUM_ENT(EM_SPARCV9, "Sparc v9"), + ENUM_ENT(EM_TRICORE, "Siemens Tricore"), + ENUM_ENT(EM_ARC, "ARC"), + ENUM_ENT(EM_H8_300, "Hitachi H8/300"), + ENUM_ENT(EM_H8_300H, "Hitachi H8/300H"), + ENUM_ENT(EM_H8S, "Hitachi H8S"), + ENUM_ENT(EM_H8_500, "Hitachi H8/500"), + ENUM_ENT(EM_IA_64, "Intel IA-64"), + ENUM_ENT(EM_MIPS_X, "Stanford MIPS-X"), + ENUM_ENT(EM_COLDFIRE, "Motorola Coldfire"), + ENUM_ENT(EM_68HC12, "Motorola MC68HC12 Microcontroller"), + ENUM_ENT(EM_MMA, "Fujitsu Multimedia Accelerator"), + ENUM_ENT(EM_PCP, "Siemens PCP"), + ENUM_ENT(EM_NCPU, "Sony nCPU embedded RISC processor"), + ENUM_ENT(EM_NDR1, "Denso NDR1 microprocesspr"), + ENUM_ENT(EM_STARCORE, "Motorola Star*Core processor"), + ENUM_ENT(EM_ME16, "Toyota ME16 processor"), + ENUM_ENT(EM_ST100, "STMicroelectronics ST100 processor"), + ENUM_ENT(EM_TINYJ, "Advanced Logic Corp. TinyJ embedded processor"), + ENUM_ENT(EM_X86_64, "Advanced Micro Devices X86-64"), + ENUM_ENT(EM_PDSP, "Sony DSP processor"), + ENUM_ENT(EM_PDP10, "Digital Equipment Corp. PDP-10"), + ENUM_ENT(EM_PDP11, "Digital Equipment Corp. PDP-11"), + ENUM_ENT(EM_FX66, "Siemens FX66 microcontroller"), + ENUM_ENT(EM_ST9PLUS, "STMicroelectronics ST9+ 8/16 bit microcontroller"), + ENUM_ENT(EM_ST7, "STMicroelectronics ST7 8-bit microcontroller"), + ENUM_ENT(EM_68HC16, "Motorola MC68HC16 Microcontroller"), + ENUM_ENT(EM_68HC11, "Motorola MC68HC11 Microcontroller"), + ENUM_ENT(EM_68HC08, "Motorola MC68HC08 Microcontroller"), + ENUM_ENT(EM_68HC05, "Motorola MC68HC05 Microcontroller"), + ENUM_ENT(EM_SVX, "Silicon Graphics SVx"), + ENUM_ENT(EM_ST19, "STMicroelectronics ST19 8-bit microcontroller"), + ENUM_ENT(EM_VAX, "Digital VAX"), + ENUM_ENT(EM_CRIS, "Axis Communications 32-bit embedded processor"), + ENUM_ENT(EM_JAVELIN, "Infineon Technologies 32-bit embedded cpu"), + ENUM_ENT(EM_FIREPATH, "Element 14 64-bit DSP processor"), + ENUM_ENT(EM_ZSP, "LSI Logic's 16-bit DSP processor"), + ENUM_ENT(EM_MMIX, "Donald Knuth's educational 64-bit processor"), + ENUM_ENT(EM_HUANY, "Harvard Universitys's machine-independent object format"), + ENUM_ENT(EM_PRISM, "Vitesse Prism"), + ENUM_ENT(EM_AVR, "Atmel AVR 8-bit microcontroller"), + ENUM_ENT(EM_FR30, "Fujitsu FR30"), + ENUM_ENT(EM_D10V, "Mitsubishi D10V"), + ENUM_ENT(EM_D30V, "Mitsubishi D30V"), + ENUM_ENT(EM_V850, "NEC v850"), + ENUM_ENT(EM_M32R, "Renesas M32R (formerly Mitsubishi M32r)"), + ENUM_ENT(EM_MN10300, "Matsushita MN10300"), + ENUM_ENT(EM_MN10200, "Matsushita MN10200"), + ENUM_ENT(EM_PJ, "picoJava"), + ENUM_ENT(EM_OPENRISC, "OpenRISC 32-bit embedded processor"), + ENUM_ENT(EM_ARC_COMPACT, "EM_ARC_COMPACT"), + ENUM_ENT(EM_XTENSA, "Tensilica Xtensa Processor"), + ENUM_ENT(EM_VIDEOCORE, "Alphamosaic VideoCore processor"), + ENUM_ENT(EM_TMM_GPP, "Thompson Multimedia General Purpose Processor"), + ENUM_ENT(EM_NS32K, "National Semiconductor 32000 series"), + ENUM_ENT(EM_TPC, "Tenor Network TPC processor"), + ENUM_ENT(EM_SNP1K, "EM_SNP1K"), + ENUM_ENT(EM_ST200, "STMicroelectronics ST200 microcontroller"), + ENUM_ENT(EM_IP2K, "Ubicom IP2xxx 8-bit microcontrollers"), + ENUM_ENT(EM_MAX, "MAX Processor"), + ENUM_ENT(EM_CR, "National Semiconductor CompactRISC"), + ENUM_ENT(EM_F2MC16, "Fujitsu F2MC16"), + ENUM_ENT(EM_MSP430, "Texas Instruments msp430 microcontroller"), + ENUM_ENT(EM_BLACKFIN, "Analog Devices Blackfin"), + ENUM_ENT(EM_SE_C33, "S1C33 Family of Seiko Epson processors"), + ENUM_ENT(EM_SEP, "Sharp embedded microprocessor"), + ENUM_ENT(EM_ARCA, "Arca RISC microprocessor"), + ENUM_ENT(EM_UNICORE, "Unicore"), + ENUM_ENT(EM_EXCESS, "eXcess 16/32/64-bit configurable embedded CPU"), + ENUM_ENT(EM_DXP, "Icera Semiconductor Inc. Deep Execution Processor"), + ENUM_ENT(EM_ALTERA_NIOS2, "Altera Nios"), + ENUM_ENT(EM_CRX, "National Semiconductor CRX microprocessor"), + ENUM_ENT(EM_XGATE, "Motorola XGATE embedded processor"), + ENUM_ENT(EM_C166, "Infineon Technologies xc16x"), + ENUM_ENT(EM_M16C, "Renesas M16C"), + ENUM_ENT(EM_DSPIC30F, "Microchip Technology dsPIC30F Digital Signal Controller"), + ENUM_ENT(EM_CE, "Freescale Communication Engine RISC core"), + ENUM_ENT(EM_M32C, "Renesas M32C"), + ENUM_ENT(EM_TSK3000, "Altium TSK3000 core"), + ENUM_ENT(EM_RS08, "Freescale RS08 embedded processor"), + ENUM_ENT(EM_SHARC, "EM_SHARC"), + ENUM_ENT(EM_ECOG2, "Cyan Technology eCOG2 microprocessor"), + ENUM_ENT(EM_SCORE7, "SUNPLUS S+Core"), + ENUM_ENT(EM_DSP24, "New Japan Radio (NJR) 24-bit DSP Processor"), + ENUM_ENT(EM_VIDEOCORE3, "Broadcom VideoCore III processor"), + ENUM_ENT(EM_LATTICEMICO32, "Lattice Mico32"), + ENUM_ENT(EM_SE_C17, "Seiko Epson C17 family"), + ENUM_ENT(EM_TI_C6000, "Texas Instruments TMS320C6000 DSP family"), + ENUM_ENT(EM_TI_C2000, "Texas Instruments TMS320C2000 DSP family"), + ENUM_ENT(EM_TI_C5500, "Texas Instruments TMS320C55x DSP family"), + ENUM_ENT(EM_MMDSP_PLUS, "STMicroelectronics 64bit VLIW Data Signal Processor"), + ENUM_ENT(EM_CYPRESS_M8C, "Cypress M8C microprocessor"), + ENUM_ENT(EM_R32C, "Renesas R32C series microprocessors"), + ENUM_ENT(EM_TRIMEDIA, "NXP Semiconductors TriMedia architecture family"), + ENUM_ENT(EM_HEXAGON, "Qualcomm Hexagon"), + ENUM_ENT(EM_8051, "Intel 8051 and variants"), + ENUM_ENT(EM_STXP7X, "STMicroelectronics STxP7x family"), + ENUM_ENT(EM_NDS32, "Andes Technology compact code size embedded RISC processor family"), + ENUM_ENT(EM_ECOG1, "Cyan Technology eCOG1 microprocessor"), + ENUM_ENT(EM_ECOG1X, "Cyan Technology eCOG1X family"), + ENUM_ENT(EM_MAXQ30, "Dallas Semiconductor MAXQ30 Core microcontrollers"), + ENUM_ENT(EM_XIMO16, "New Japan Radio (NJR) 16-bit DSP Processor"), + ENUM_ENT(EM_MANIK, "M2000 Reconfigurable RISC Microprocessor"), + ENUM_ENT(EM_CRAYNV2, "Cray Inc. NV2 vector architecture"), + ENUM_ENT(EM_RX, "Renesas RX"), + ENUM_ENT(EM_METAG, "Imagination Technologies Meta processor architecture"), + ENUM_ENT(EM_MCST_ELBRUS, "MCST Elbrus general purpose hardware architecture"), + ENUM_ENT(EM_ECOG16, "Cyan Technology eCOG16 family"), + ENUM_ENT(EM_CR16, "Xilinx MicroBlaze"), + ENUM_ENT(EM_ETPU, "Freescale Extended Time Processing Unit"), + ENUM_ENT(EM_SLE9X, "Infineon Technologies SLE9X core"), + ENUM_ENT(EM_L10M, "EM_L10M"), + ENUM_ENT(EM_K10M, "EM_K10M"), + ENUM_ENT(EM_AARCH64, "AArch64"), + ENUM_ENT(EM_AVR32, "Atmel AVR 8-bit microcontroller"), + ENUM_ENT(EM_STM8, "STMicroeletronics STM8 8-bit microcontroller"), + ENUM_ENT(EM_TILE64, "Tilera TILE64 multicore architecture family"), + ENUM_ENT(EM_TILEPRO, "Tilera TILEPro multicore architecture family"), + ENUM_ENT(EM_CUDA, "NVIDIA CUDA architecture"), + ENUM_ENT(EM_TILEGX, "Tilera TILE-Gx multicore architecture family"), + ENUM_ENT(EM_CLOUDSHIELD, "EM_CLOUDSHIELD"), + ENUM_ENT(EM_COREA_1ST, "EM_COREA_1ST"), + ENUM_ENT(EM_COREA_2ND, "EM_COREA_2ND"), + ENUM_ENT(EM_ARC_COMPACT2, "EM_ARC_COMPACT2"), + ENUM_ENT(EM_OPEN8, "EM_OPEN8"), + ENUM_ENT(EM_RL78, "Renesas RL78"), + ENUM_ENT(EM_VIDEOCORE5, "Broadcom VideoCore V processor"), + ENUM_ENT(EM_78KOR, "EM_78KOR"), + ENUM_ENT(EM_56800EX, "EM_56800EX"), + ENUM_ENT(EM_AMDGPU, "EM_AMDGPU"), + ENUM_ENT(EM_WEBASSEMBLY, "EM_WEBASSEMBLY"), + ENUM_ENT(EM_LANAI, "EM_LANAI"), + ENUM_ENT(EM_BPF, "EM_BPF"), }; static const EnumEntry<unsigned> ElfSymbolBindings[] = { - { "Local", ELF::STB_LOCAL }, - { "Global", ELF::STB_GLOBAL }, - { "Weak", ELF::STB_WEAK }, - { "Unique", ELF::STB_GNU_UNIQUE } -}; + {"Local", "LOCAL", ELF::STB_LOCAL}, + {"Global", "GLOBAL", ELF::STB_GLOBAL}, + {"Weak", "WEAK", ELF::STB_WEAK}, + {"Unique", "UNIQUE", ELF::STB_GNU_UNIQUE}}; + +static const EnumEntry<unsigned> ElfSymbolVisibilities[] = { + {"DEFAULT", "DEFAULT", ELF::STV_DEFAULT}, + {"INTERNAL", "INTERNAL", ELF::STV_INTERNAL}, + {"HIDDEN", "HIDDEN", ELF::STV_HIDDEN}, + {"PROTECTED", "PROTECTED", ELF::STV_PROTECTED}}; static const EnumEntry<unsigned> ElfSymbolTypes[] = { - { "None", ELF::STT_NOTYPE }, - { "Object", ELF::STT_OBJECT }, - { "Function", ELF::STT_FUNC }, - { "Section", ELF::STT_SECTION }, - { "File", ELF::STT_FILE }, - { "Common", ELF::STT_COMMON }, - { "TLS", ELF::STT_TLS }, - { "GNU_IFunc", ELF::STT_GNU_IFUNC } -}; + {"None", "NOTYPE", ELF::STT_NOTYPE}, + {"Object", "OBJECT", ELF::STT_OBJECT}, + {"Function", "FUNC", ELF::STT_FUNC}, + {"Section", "SECTION", ELF::STT_SECTION}, + {"File", "FILE", ELF::STT_FILE}, + {"Common", "COMMON", ELF::STT_COMMON}, + {"TLS", "TLS", ELF::STT_TLS}, + {"GNU_IFunc", "IFUNC", ELF::STT_GNU_IFUNC}}; static const EnumEntry<unsigned> AMDGPUSymbolTypes[] = { { "AMDGPU_HSA_KERNEL", ELF::STT_AMDGPU_HSA_KERNEL }, @@ -786,27 +1037,92 @@ static const char *getElfSectionType(unsigned Arch, unsigned Type) { } } +static const char *getGroupType(uint32_t Flag) { + if (Flag & ELF::GRP_COMDAT) + return "COMDAT"; + else + return "(unknown)"; +} + static const EnumEntry<unsigned> ElfSectionFlags[] = { - LLVM_READOBJ_ENUM_ENT(ELF, SHF_WRITE ), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_ALLOC ), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_EXCLUDE ), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_EXECINSTR ), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_MERGE ), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_STRINGS ), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_INFO_LINK ), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_LINK_ORDER ), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_OS_NONCONFORMING), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_GROUP ), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_TLS ), + ENUM_ENT(SHF_WRITE, "W"), + ENUM_ENT(SHF_ALLOC, "A"), + ENUM_ENT(SHF_EXCLUDE, "E"), + ENUM_ENT(SHF_EXECINSTR, "X"), + ENUM_ENT(SHF_MERGE, "M"), + ENUM_ENT(SHF_STRINGS, "S"), + ENUM_ENT(SHF_INFO_LINK, "I"), + ENUM_ENT(SHF_LINK_ORDER, "L"), + ENUM_ENT(SHF_OS_NONCONFORMING, "o"), + ENUM_ENT(SHF_GROUP, "G"), + ENUM_ENT(SHF_TLS, "T"), + ENUM_ENT(SHF_MASKOS, "o"), + ENUM_ENT(SHF_MASKPROC, "p"), + ENUM_ENT_1(SHF_COMPRESSED), +}; + +static const EnumEntry<unsigned> ElfXCoreSectionFlags[] = { LLVM_READOBJ_ENUM_ENT(ELF, XCORE_SHF_CP_SECTION), - LLVM_READOBJ_ENUM_ENT(ELF, XCORE_SHF_DP_SECTION), - LLVM_READOBJ_ENUM_ENT(ELF, SHF_MIPS_NOSTRIP ), + LLVM_READOBJ_ENUM_ENT(ELF, XCORE_SHF_DP_SECTION) +}; + +static const EnumEntry<unsigned> ElfAMDGPUSectionFlags[] = { LLVM_READOBJ_ENUM_ENT(ELF, SHF_AMDGPU_HSA_GLOBAL), LLVM_READOBJ_ENUM_ENT(ELF, SHF_AMDGPU_HSA_READONLY), LLVM_READOBJ_ENUM_ENT(ELF, SHF_AMDGPU_HSA_CODE), LLVM_READOBJ_ENUM_ENT(ELF, SHF_AMDGPU_HSA_AGENT) }; +static const EnumEntry<unsigned> ElfHexagonSectionFlags[] = { + LLVM_READOBJ_ENUM_ENT(ELF, SHF_HEX_GPREL) +}; + +static const EnumEntry<unsigned> ElfMipsSectionFlags[] = { + LLVM_READOBJ_ENUM_ENT(ELF, SHF_MIPS_NODUPES), + LLVM_READOBJ_ENUM_ENT(ELF, SHF_MIPS_NAMES ), + LLVM_READOBJ_ENUM_ENT(ELF, SHF_MIPS_LOCAL ), + LLVM_READOBJ_ENUM_ENT(ELF, SHF_MIPS_NOSTRIP), + LLVM_READOBJ_ENUM_ENT(ELF, SHF_MIPS_GPREL ), + LLVM_READOBJ_ENUM_ENT(ELF, SHF_MIPS_MERGE ), + LLVM_READOBJ_ENUM_ENT(ELF, SHF_MIPS_ADDR ), + LLVM_READOBJ_ENUM_ENT(ELF, SHF_MIPS_STRING ) +}; + +static const EnumEntry<unsigned> ElfX86_64SectionFlags[] = { + LLVM_READOBJ_ENUM_ENT(ELF, SHF_X86_64_LARGE) +}; + +static std::string getGNUFlags(uint64_t Flags) { + std::string Str; + for (auto Entry : ElfSectionFlags) { + uint64_t Flag = Entry.Value & Flags; + Flags &= ~Entry.Value; + switch (Flag) { + case ELF::SHF_WRITE: + case ELF::SHF_ALLOC: + case ELF::SHF_EXECINSTR: + case ELF::SHF_MERGE: + case ELF::SHF_STRINGS: + case ELF::SHF_INFO_LINK: + case ELF::SHF_LINK_ORDER: + case ELF::SHF_OS_NONCONFORMING: + case ELF::SHF_GROUP: + case ELF::SHF_TLS: + case ELF::SHF_EXCLUDE: + Str += Entry.AltName; + break; + default: + if (Flag & ELF::SHF_MASKOS) + Str += "o"; + else if (Flag & ELF::SHF_MASKPROC) + Str += "p"; + else if (Flag) + Str += "x"; + } + } + return Str; +} + static const char *getElfSegmentType(unsigned Arch, unsigned Type) { // Check potentially overlapped processor-specific // program header type. @@ -851,6 +1167,53 @@ static const char *getElfSegmentType(unsigned Arch, unsigned Type) { } } +static std::string getElfPtType(unsigned Arch, unsigned Type) { + switch (Type) { + LLVM_READOBJ_PHDR_ENUM(ELF, PT_NULL) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_LOAD) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_DYNAMIC) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_INTERP) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_NOTE) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_SHLIB) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_PHDR) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_TLS) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_GNU_EH_FRAME) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_SUNW_UNWIND) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_GNU_STACK) + LLVM_READOBJ_PHDR_ENUM(ELF, PT_GNU_RELRO) + default: + // All machine specific PT_* types + switch (Arch) { + case ELF::EM_AMDGPU: + switch (Type) { + LLVM_READOBJ_ENUM_CASE(ELF, PT_AMDGPU_HSA_LOAD_GLOBAL_PROGRAM); + LLVM_READOBJ_ENUM_CASE(ELF, PT_AMDGPU_HSA_LOAD_GLOBAL_AGENT); + LLVM_READOBJ_ENUM_CASE(ELF, PT_AMDGPU_HSA_LOAD_READONLY_AGENT); + LLVM_READOBJ_ENUM_CASE(ELF, PT_AMDGPU_HSA_LOAD_CODE_AGENT); + } + return ""; + case ELF::EM_ARM: + if (Type == ELF::PT_ARM_EXIDX) + return "EXIDX"; + return ""; + case ELF::EM_MIPS: + case ELF::EM_MIPS_RS3_LE: + switch (Type) { + case PT_MIPS_REGINFO: + return "REGINFO"; + case PT_MIPS_RTPROC: + return "RTPROC"; + case PT_MIPS_OPTIONS: + return "OPTIONS"; + case PT_MIPS_ABIFLAGS: + return "ABIFLAGS"; + } + return ""; + } + } + return std::string("<unknown>: ") + to_string(format_hex(Type, 1)); +} + static const EnumEntry<unsigned> ElfSegmentFlags[] = { LLVM_READOBJ_ENUM_ENT(ELF, PF_X), LLVM_READOBJ_ENUM_ENT(ELF, PF_W), @@ -903,14 +1266,52 @@ static const EnumEntry<unsigned> ElfHeaderMipsFlags[] = { LLVM_READOBJ_ENUM_ENT(ELF, EF_MIPS_ARCH_64R6) }; +static const EnumEntry<unsigned> ElfSymOtherFlags[] = { + LLVM_READOBJ_ENUM_ENT(ELF, STV_INTERNAL), + LLVM_READOBJ_ENUM_ENT(ELF, STV_HIDDEN), + LLVM_READOBJ_ENUM_ENT(ELF, STV_PROTECTED) +}; + +static const EnumEntry<unsigned> ElfMipsSymOtherFlags[] = { + LLVM_READOBJ_ENUM_ENT(ELF, STO_MIPS_OPTIONAL), + LLVM_READOBJ_ENUM_ENT(ELF, STO_MIPS_PLT), + LLVM_READOBJ_ENUM_ENT(ELF, STO_MIPS_PIC), + LLVM_READOBJ_ENUM_ENT(ELF, STO_MIPS_MICROMIPS) +}; + +static const EnumEntry<unsigned> ElfMips16SymOtherFlags[] = { + LLVM_READOBJ_ENUM_ENT(ELF, STO_MIPS_OPTIONAL), + LLVM_READOBJ_ENUM_ENT(ELF, STO_MIPS_PLT), + LLVM_READOBJ_ENUM_ENT(ELF, STO_MIPS_MIPS16) +}; + +static const char *getElfMipsOptionsOdkType(unsigned Odk) { + switch (Odk) { + LLVM_READOBJ_ENUM_CASE(ELF, ODK_NULL); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_REGINFO); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_EXCEPTIONS); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_PAD); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_HWPATCH); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_FILL); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_TAGS); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_HWAND); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_HWOR); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_GP_GROUP); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_IDENT); + LLVM_READOBJ_ENUM_CASE(ELF, ODK_PAGESIZE); + default: + return "Unknown"; + } +} + template <typename ELFT> -ELFDumper<ELFT>::ELFDumper(const ELFFile<ELFT> *Obj, StreamWriter &Writer) +ELFDumper<ELFT>::ELFDumper(const ELFFile<ELFT> *Obj, ScopedPrinter &Writer) : ObjDumper(Writer), Obj(Obj) { SmallVector<const Elf_Phdr *, 4> LoadSegments; for (const Elf_Phdr &Phdr : Obj->program_headers()) { if (Phdr.p_type == ELF::PT_DYNAMIC) { - DynamicProgHeader = &Phdr; + DynamicTable = createDRIFrom(&Phdr, sizeof(Elf_Dyn)); continue; } if (Phdr.p_type != ELF::PT_LOAD || Phdr.p_filesz == 0) @@ -918,8 +1319,54 @@ ELFDumper<ELFT>::ELFDumper(const ELFFile<ELFT> *Obj, StreamWriter &Writer) LoadSegments.push_back(&Phdr); } + for (const Elf_Shdr &Sec : Obj->sections()) { + switch (Sec.sh_type) { + case ELF::SHT_SYMTAB: + if (DotSymtabSec != nullptr) + reportError("Multilpe SHT_SYMTAB"); + DotSymtabSec = &Sec; + break; + case ELF::SHT_DYNSYM: + if (DynSymRegion.Size) + reportError("Multilpe SHT_DYNSYM"); + DynSymRegion = createDRIFrom(&Sec); + // This is only used (if Elf_Shdr present)for naming section in GNU style + DynSymtabName = unwrapOrError(Obj->getSectionName(&Sec)); + break; + case ELF::SHT_SYMTAB_SHNDX: + ShndxTable = unwrapOrError(Obj->getSHNDXTable(Sec)); + break; + case ELF::SHT_GNU_versym: + if (dot_gnu_version_sec != nullptr) + reportError("Multiple SHT_GNU_versym"); + dot_gnu_version_sec = &Sec; + break; + case ELF::SHT_GNU_verdef: + if (dot_gnu_version_d_sec != nullptr) + reportError("Multiple SHT_GNU_verdef"); + dot_gnu_version_d_sec = &Sec; + break; + case ELF::SHT_GNU_verneed: + if (dot_gnu_version_r_sec != nullptr) + reportError("Multilpe SHT_GNU_verneed"); + dot_gnu_version_r_sec = &Sec; + break; + } + } + + parseDynamicTable(LoadSegments); + + if (opts::Output == opts::GNU) + ELFDumperStyle.reset(new GNUStyle<ELFT>(Writer, this)); + else + ELFDumperStyle.reset(new LLVMStyle<ELFT>(Writer, this)); +} + +template <typename ELFT> +void ELFDumper<ELFT>::parseDynamicTable( + ArrayRef<const Elf_Phdr *> LoadSegments) { auto toMappedAddr = [&](uint64_t VAddr) -> const uint8_t * { - const Elf_Phdr **I = std::upper_bound( + const Elf_Phdr *const *I = std::upper_bound( LoadSegments.begin(), LoadSegments.end(), VAddr, compareAddr<ELFT>); if (I == LoadSegments.begin()) report_fatal_error("Virtual address is not in any segment"); @@ -944,6 +1391,16 @@ ELFDumper<ELFT>::ELFDumper(const ELFFile<ELFT> *Obj, StreamWriter &Writer) GnuHashTable = reinterpret_cast<const Elf_GnuHash *>(toMappedAddr(Dyn.getPtr())); break; + case ELF::DT_STRTAB: + StringTableBegin = (const char *)toMappedAddr(Dyn.getPtr()); + break; + case ELF::DT_STRSZ: + StringTableSize = Dyn.getVal(); + break; + case ELF::DT_SYMTAB: + DynSymRegion.Addr = toMappedAddr(Dyn.getPtr()); + DynSymRegion.EntSize = sizeof(Elf_Sym); + break; case ELF::DT_RELA: DynRelaRegion.Addr = toMappedAddr(Dyn.getPtr()); break; @@ -956,15 +1413,29 @@ ELFDumper<ELFT>::ELFDumper(const ELFFile<ELFT> *Obj, StreamWriter &Writer) case ELF::DT_SONAME: SONameOffset = Dyn.getVal(); break; - case ELF::DT_STRTAB: - StringTableBegin = (const char *)toMappedAddr(Dyn.getPtr()); + case ELF::DT_REL: + DynRelRegion.Addr = toMappedAddr(Dyn.getPtr()); break; - case ELF::DT_STRSZ: - StringTableSize = Dyn.getVal(); + case ELF::DT_RELSZ: + DynRelRegion.Size = Dyn.getVal(); break; - case ELF::DT_SYMTAB: - DynSymStart = - reinterpret_cast<const Elf_Sym *>(toMappedAddr(Dyn.getPtr())); + case ELF::DT_RELENT: + DynRelRegion.EntSize = Dyn.getVal(); + break; + case ELF::DT_PLTREL: + if (Dyn.getVal() == DT_REL) + DynPLTRelRegion.EntSize = sizeof(Elf_Rel); + else if (Dyn.getVal() == DT_RELA) + DynPLTRelRegion.EntSize = sizeof(Elf_Rela); + else + reportError(Twine("unknown DT_PLTREL value of ") + + Twine((uint64_t)Dyn.getVal())); + break; + case ELF::DT_JMPREL: + DynPLTRelRegion.Addr = toMappedAddr(Dyn.getPtr()); + break; + case ELF::DT_PLTRELSZ: + DynPLTRelRegion.Size = Dyn.getVal(); break; } } @@ -972,326 +1443,54 @@ ELFDumper<ELFT>::ELFDumper(const ELFFile<ELFT> *Obj, StreamWriter &Writer) DynamicStringTable = StringRef(StringTableBegin, StringTableSize); if (SONameOffset) SOName = getDynamicString(SONameOffset); - - for (const Elf_Shdr &Sec : Obj->sections()) { - switch (Sec.sh_type) { - case ELF::SHT_GNU_versym: - if (dot_gnu_version_sec != nullptr) - reportError("Multiple SHT_GNU_versym"); - dot_gnu_version_sec = &Sec; - break; - case ELF::SHT_GNU_verdef: - if (dot_gnu_version_d_sec != nullptr) - reportError("Multiple SHT_GNU_verdef"); - dot_gnu_version_d_sec = &Sec; - break; - case ELF::SHT_GNU_verneed: - if (dot_gnu_version_r_sec != nullptr) - reportError("Multilpe SHT_GNU_verneed"); - dot_gnu_version_r_sec = &Sec; - break; - case ELF::SHT_DYNSYM: - if (DotDynSymSec != nullptr) - reportError("Multilpe SHT_DYNSYM"); - DotDynSymSec = &Sec; - break; - case ELF::SHT_SYMTAB: - if (DotSymtabSec != nullptr) - reportError("Multilpe SHT_SYMTAB"); - DotSymtabSec = &Sec; - break; - case ELF::SHT_SYMTAB_SHNDX: { - ErrorOr<ArrayRef<Elf_Word>> TableOrErr = Obj->getSHNDXTable(Sec); - error(TableOrErr.getError()); - ShndxTable = *TableOrErr; - break; - } - } - } -} - -template <typename ELFT> -const typename ELFDumper<ELFT>::Elf_Rela * -ELFDumper<ELFT>::dyn_rela_begin() const { - if (DynRelaRegion.Size && DynRelaRegion.EntSize != sizeof(Elf_Rela)) - report_fatal_error("Invalid relocation entry size"); - return reinterpret_cast<const Elf_Rela *>(DynRelaRegion.Addr); } template <typename ELFT> -const typename ELFDumper<ELFT>::Elf_Rela * -ELFDumper<ELFT>::dyn_rela_end() const { - uint64_t Size = DynRelaRegion.Size; - if (Size % sizeof(Elf_Rela)) - report_fatal_error("Invalid relocation table size"); - return dyn_rela_begin() + Size / sizeof(Elf_Rela); +typename ELFDumper<ELFT>::Elf_Rel_Range ELFDumper<ELFT>::dyn_rels() const { + return DynRelRegion.getAsArrayRef<Elf_Rel>(); } template <typename ELFT> typename ELFDumper<ELFT>::Elf_Rela_Range ELFDumper<ELFT>::dyn_relas() const { - return make_range(dyn_rela_begin(), dyn_rela_end()); + return DynRelaRegion.getAsArrayRef<Elf_Rela>(); } template<class ELFT> void ELFDumper<ELFT>::printFileHeaders() { - const Elf_Ehdr *Header = Obj->getHeader(); - - { - DictScope D(W, "ElfHeader"); - { - DictScope D(W, "Ident"); - W.printBinary("Magic", makeArrayRef(Header->e_ident).slice(ELF::EI_MAG0, - 4)); - W.printEnum ("Class", Header->e_ident[ELF::EI_CLASS], - makeArrayRef(ElfClass)); - W.printEnum ("DataEncoding", Header->e_ident[ELF::EI_DATA], - makeArrayRef(ElfDataEncoding)); - W.printNumber("FileVersion", Header->e_ident[ELF::EI_VERSION]); - - // Handle architecture specific OS/ABI values. - if (Header->e_machine == ELF::EM_AMDGPU && - Header->e_ident[ELF::EI_OSABI] == ELF::ELFOSABI_AMDGPU_HSA) - W.printHex("OS/ABI", "AMDGPU_HSA", ELF::ELFOSABI_AMDGPU_HSA); - else - W.printEnum ("OS/ABI", Header->e_ident[ELF::EI_OSABI], - makeArrayRef(ElfOSABI)); - W.printNumber("ABIVersion", Header->e_ident[ELF::EI_ABIVERSION]); - W.printBinary("Unused", makeArrayRef(Header->e_ident).slice(ELF::EI_PAD)); - } - - W.printEnum ("Type", Header->e_type, makeArrayRef(ElfObjectFileType)); - W.printEnum ("Machine", Header->e_machine, makeArrayRef(ElfMachineType)); - W.printNumber("Version", Header->e_version); - W.printHex ("Entry", Header->e_entry); - W.printHex ("ProgramHeaderOffset", Header->e_phoff); - W.printHex ("SectionHeaderOffset", Header->e_shoff); - if (Header->e_machine == EM_MIPS) - W.printFlags("Flags", Header->e_flags, makeArrayRef(ElfHeaderMipsFlags), - unsigned(ELF::EF_MIPS_ARCH), unsigned(ELF::EF_MIPS_ABI), - unsigned(ELF::EF_MIPS_MACH)); - else - W.printFlags("Flags", Header->e_flags); - W.printNumber("HeaderSize", Header->e_ehsize); - W.printNumber("ProgramHeaderEntrySize", Header->e_phentsize); - W.printNumber("ProgramHeaderCount", Header->e_phnum); - W.printNumber("SectionHeaderEntrySize", Header->e_shentsize); - W.printNumber("SectionHeaderCount", Header->e_shnum); - W.printNumber("StringTableSectionIndex", Header->e_shstrndx); - } + ELFDumperStyle->printFileHeaders(Obj); } template<class ELFT> void ELFDumper<ELFT>::printSections() { - ListScope SectionsD(W, "Sections"); - - int SectionIndex = -1; - for (const Elf_Shdr &Sec : Obj->sections()) { - ++SectionIndex; - - StringRef Name = errorOrDefault(Obj->getSectionName(&Sec)); - - DictScope SectionD(W, "Section"); - W.printNumber("Index", SectionIndex); - W.printNumber("Name", Name, Sec.sh_name); - W.printHex("Type", - getElfSectionType(Obj->getHeader()->e_machine, Sec.sh_type), - Sec.sh_type); - W.printFlags("Flags", Sec.sh_flags, makeArrayRef(ElfSectionFlags)); - W.printHex("Address", Sec.sh_addr); - W.printHex("Offset", Sec.sh_offset); - W.printNumber("Size", Sec.sh_size); - W.printNumber("Link", Sec.sh_link); - W.printNumber("Info", Sec.sh_info); - W.printNumber("AddressAlignment", Sec.sh_addralign); - W.printNumber("EntrySize", Sec.sh_entsize); - - if (opts::SectionRelocations) { - ListScope D(W, "Relocations"); - printRelocations(&Sec); - } - - if (opts::SectionSymbols) { - ListScope D(W, "Symbols"); - const Elf_Shdr *Symtab = DotSymtabSec; - ErrorOr<StringRef> StrTableOrErr = Obj->getStringTableForSymtab(*Symtab); - error(StrTableOrErr.getError()); - StringRef StrTable = *StrTableOrErr; - - for (const Elf_Sym &Sym : Obj->symbols(Symtab)) { - ErrorOr<const Elf_Shdr *> SymSec = - Obj->getSection(&Sym, Symtab, ShndxTable); - if (!SymSec) - continue; - if (*SymSec == &Sec) - printSymbol(&Sym, Symtab, StrTable, false); - } - } - - if (opts::SectionData && Sec.sh_type != ELF::SHT_NOBITS) { - ArrayRef<uint8_t> Data = errorOrDefault(Obj->getSectionContents(&Sec)); - W.printBinaryBlock("SectionData", - StringRef((const char *)Data.data(), Data.size())); - } - } + ELFDumperStyle->printSections(Obj); } template<class ELFT> void ELFDumper<ELFT>::printRelocations() { - ListScope D(W, "Relocations"); - - int SectionNumber = -1; - for (const Elf_Shdr &Sec : Obj->sections()) { - ++SectionNumber; - - if (Sec.sh_type != ELF::SHT_REL && Sec.sh_type != ELF::SHT_RELA) - continue; - - StringRef Name = errorOrDefault(Obj->getSectionName(&Sec)); - - W.startLine() << "Section (" << SectionNumber << ") " << Name << " {\n"; - W.indent(); - - printRelocations(&Sec); - - W.unindent(); - W.startLine() << "}\n"; - } -} - -template<class ELFT> -void ELFDumper<ELFT>::printDynamicRelocations() { - W.startLine() << "Dynamic Relocations {\n"; - W.indent(); - for (const Elf_Rela &Rel : dyn_relas()) { - SmallString<32> RelocName; - Obj->getRelocationTypeName(Rel.getType(Obj->isMips64EL()), RelocName); - StringRef SymbolName; - uint32_t SymIndex = Rel.getSymbol(Obj->isMips64EL()); - const Elf_Sym *Sym = DynSymStart + SymIndex; - SymbolName = errorOrDefault(Sym->getName(DynamicStringTable)); - if (opts::ExpandRelocs) { - DictScope Group(W, "Relocation"); - W.printHex("Offset", Rel.r_offset); - W.printNumber("Type", RelocName, (int)Rel.getType(Obj->isMips64EL())); - W.printString("Symbol", SymbolName.size() > 0 ? SymbolName : "-"); - W.printHex("Addend", Rel.r_addend); - } - else { - raw_ostream& OS = W.startLine(); - OS << W.hex(Rel.r_offset) << " " << RelocName << " " - << (SymbolName.size() > 0 ? SymbolName : "-") << " " - << W.hex(Rel.r_addend) << "\n"; - } - } - W.unindent(); - W.startLine() << "}\n"; -} - -template <class ELFT> -void ELFDumper<ELFT>::printRelocations(const Elf_Shdr *Sec) { - ErrorOr<const Elf_Shdr *> SymTabOrErr = Obj->getSection(Sec->sh_link); - error(SymTabOrErr.getError()); - const Elf_Shdr *SymTab = *SymTabOrErr; - - switch (Sec->sh_type) { - case ELF::SHT_REL: - for (const Elf_Rel &R : Obj->rels(Sec)) { - Elf_Rela Rela; - Rela.r_offset = R.r_offset; - Rela.r_info = R.r_info; - Rela.r_addend = 0; - printRelocation(Rela, SymTab); - } - break; - case ELF::SHT_RELA: - for (const Elf_Rela &R : Obj->relas(Sec)) - printRelocation(R, SymTab); - break; - } + ELFDumperStyle->printRelocations(Obj); } -template <class ELFT> -void ELFDumper<ELFT>::printRelocation(Elf_Rela Rel, const Elf_Shdr *SymTab) { - SmallString<32> RelocName; - Obj->getRelocationTypeName(Rel.getType(Obj->isMips64EL()), RelocName); - StringRef TargetName; - const Elf_Sym *Sym = Obj->getRelocationSymbol(&Rel, SymTab); - if (Sym && Sym->getType() == ELF::STT_SECTION) { - ErrorOr<const Elf_Shdr *> Sec = Obj->getSection(Sym, SymTab, ShndxTable); - error(Sec.getError()); - ErrorOr<StringRef> SecName = Obj->getSectionName(*Sec); - if (SecName) - TargetName = SecName.get(); - } else if (Sym) { - ErrorOr<StringRef> StrTableOrErr = Obj->getStringTableForSymtab(*SymTab); - error(StrTableOrErr.getError()); - TargetName = errorOrDefault(Sym->getName(*StrTableOrErr)); - } - - if (opts::ExpandRelocs) { - DictScope Group(W, "Relocation"); - W.printHex("Offset", Rel.r_offset); - W.printNumber("Type", RelocName, (int)Rel.getType(Obj->isMips64EL())); - W.printNumber("Symbol", TargetName.size() > 0 ? TargetName : "-", - Rel.getSymbol(Obj->isMips64EL())); - W.printHex("Addend", Rel.r_addend); - } else { - raw_ostream& OS = W.startLine(); - OS << W.hex(Rel.r_offset) << " " << RelocName << " " - << (TargetName.size() > 0 ? TargetName : "-") << " " - << W.hex(Rel.r_addend) << "\n"; - } +template <class ELFT> void ELFDumper<ELFT>::printProgramHeaders() { + ELFDumperStyle->printProgramHeaders(Obj); } -template<class ELFT> -void ELFDumper<ELFT>::printSymbolsHelper(bool IsDynamic) { - const Elf_Shdr *Symtab = (IsDynamic) ? DotDynSymSec : DotSymtabSec; - if (!Symtab) - return; - ErrorOr<StringRef> StrTableOrErr = Obj->getStringTableForSymtab(*Symtab); - error(StrTableOrErr.getError()); - StringRef StrTable = *StrTableOrErr; - for (const Elf_Sym &Sym : Obj->symbols(Symtab)) - printSymbol(&Sym, Symtab, StrTable, IsDynamic); +template <class ELFT> void ELFDumper<ELFT>::printDynamicRelocations() { + ELFDumperStyle->printDynamicRelocations(Obj); } template<class ELFT> void ELFDumper<ELFT>::printSymbols() { - ListScope Group(W, "Symbols"); - printSymbolsHelper(false); + ELFDumperStyle->printSymbols(Obj); } template<class ELFT> void ELFDumper<ELFT>::printDynamicSymbols() { - ListScope Group(W, "DynamicSymbols"); - printSymbolsHelper(true); + ELFDumperStyle->printDynamicSymbols(Obj); } -template <class ELFT> -void ELFDumper<ELFT>::printSymbol(const Elf_Sym *Symbol, const Elf_Shdr *SymTab, - StringRef StrTable, bool IsDynamic) { - unsigned SectionIndex = 0; - StringRef SectionName; - getSectionNameIndex(*Obj, Symbol, SymTab, ShndxTable, SectionName, - SectionIndex); - std::string FullSymbolName = getFullSymbolName(Symbol, StrTable, IsDynamic); - unsigned char SymbolType = Symbol->getType(); - - DictScope D(W, "Symbol"); - W.printNumber("Name", FullSymbolName, Symbol->st_name); - W.printHex ("Value", Symbol->st_value); - W.printNumber("Size", Symbol->st_size); - W.printEnum ("Binding", Symbol->getBinding(), - makeArrayRef(ElfSymbolBindings)); - if (Obj->getHeader()->e_machine == ELF::EM_AMDGPU && - SymbolType >= ELF::STT_LOOS && SymbolType < ELF::STT_HIOS) - W.printEnum ("Type", SymbolType, makeArrayRef(AMDGPUSymbolTypes)); - else - W.printEnum ("Type", SymbolType, makeArrayRef(ElfSymbolTypes)); - W.printNumber("Other", Symbol->st_other); - W.printHex("Section", SectionName, SectionIndex); +template <class ELFT> void ELFDumper<ELFT>::printHashHistogram() { + ELFDumperStyle->printHashHistogram(Obj); } - #define LLVM_READOBJ_TYPE_CASE(name) \ case DT_##name: return #name @@ -1336,8 +1535,11 @@ static const char *getTypeString(uint64_t Type) { LLVM_READOBJ_TYPE_CASE(VERNEED); LLVM_READOBJ_TYPE_CASE(VERNEEDNUM); LLVM_READOBJ_TYPE_CASE(VERSYM); + LLVM_READOBJ_TYPE_CASE(RELACOUNT); LLVM_READOBJ_TYPE_CASE(RELCOUNT); LLVM_READOBJ_TYPE_CASE(GNU_HASH); + LLVM_READOBJ_TYPE_CASE(TLSDESC_PLT); + LLVM_READOBJ_TYPE_CASE(TLSDESC_GOT); LLVM_READOBJ_TYPE_CASE(MIPS_RLD_VERSION); LLVM_READOBJ_TYPE_CASE(MIPS_RLD_MAP_REL); LLVM_READOBJ_TYPE_CASE(MIPS_FLAGS); @@ -1444,6 +1646,7 @@ StringRef ELFDumper<ELFT>::getDynamicString(uint64_t Value) const { template <class ELFT> void ELFDumper<ELFT>::printValue(uint64_t Type, uint64_t Value) { raw_ostream &OS = W.getOStream(); + const char* ConvChar = (opts::Output == opts::GNU) ? "0x%" PRIx64 : "0x%" PRIX64; switch (Type) { case DT_PLTREL: if (Value == DT_REL) { @@ -1478,8 +1681,9 @@ void ELFDumper<ELFT>::printValue(uint64_t Type, uint64_t Value) { case DT_MIPS_RLD_MAP_REL: case DT_MIPS_PLTGOT: case DT_MIPS_OPTIONS: - OS << format("0x%" PRIX64, Value); + OS << format(ConvChar, Value); break; + case DT_RELACOUNT: case DT_RELCOUNT: case DT_VERDEFNUM: case DT_VERNEEDNUM: @@ -1521,7 +1725,7 @@ void ELFDumper<ELFT>::printValue(uint64_t Type, uint64_t Value) { printFlags(Value, makeArrayRef(ElfDynamicDTFlags1), OS); break; default: - OS << format("0x%" PRIX64, Value); + OS << format(ConvChar, Value); break; } } @@ -1545,8 +1749,8 @@ template <> void ELFDumper<ELFType<support::little, false>>::printUnwindInfo() { template<class ELFT> void ELFDumper<ELFT>::printDynamicTable() { - auto I = dynamic_table_begin(); - auto E = dynamic_table_end(); + auto I = dynamic_table().begin(); + auto E = dynamic_table().end(); if (I == E) return; @@ -1574,7 +1778,7 @@ void ELFDumper<ELFT>::printDynamicTable() { const Elf_Dyn &Entry = *I; uintX_t Tag = Entry.getTag(); ++I; - W.startLine() << " " << format_hex(Tag, Is64 ? 18 : 10, true) << " " + W.startLine() << " " << format_hex(Tag, Is64 ? 18 : 10, opts::Output != opts::GNU) << " " << format("%-21s", getTypeString(Tag)); printValue(Tag, Entry.getVal()); OS << "\n"; @@ -1601,24 +1805,6 @@ void ELFDumper<ELFT>::printNeededLibraries() { } } -template<class ELFT> -void ELFDumper<ELFT>::printProgramHeaders() { - ListScope L(W, "ProgramHeaders"); - - for (const Elf_Phdr &Phdr : Obj->program_headers()) { - DictScope P(W, "ProgramHeader"); - W.printHex("Type", - getElfSegmentType(Obj->getHeader()->e_machine, Phdr.p_type), - Phdr.p_type); - W.printHex("Offset", Phdr.p_offset); - W.printHex("VirtualAddress", Phdr.p_vaddr); - W.printHex("PhysicalAddress", Phdr.p_paddr); - W.printNumber("FileSize", Phdr.p_filesz); - W.printNumber("MemSize", Phdr.p_memsz); - W.printFlags("Flags", Phdr.p_flags, makeArrayRef(ElfSegmentFlags)); - W.printNumber("Alignment", Phdr.p_align); - } -} template <typename ELFT> void ELFDumper<ELFT>::printHashTable() { @@ -1642,10 +1828,11 @@ void ELFDumper<ELFT>::printGnuHashTable() { W.printNumber("Shift Count", GnuHashTable->shift2); W.printHexList("Bloom Filter", GnuHashTable->filter()); W.printList("Buckets", GnuHashTable->buckets()); - if (!DotDynSymSec) + Elf_Sym_Range Syms = dynamic_symbols(); + unsigned NumSyms = std::distance(Syms.begin(), Syms.end()); + if (!NumSyms) reportError("No dynamic symbol section"); - W.printHexList("Values", - GnuHashTable->values(DotDynSymSec->getEntityCount())); + W.printHexList("Values", GnuHashTable->values(NumSyms)); } template <typename ELFT> void ELFDumper<ELFT>::printLoadName() { @@ -1669,21 +1856,18 @@ template <> void ELFDumper<ELFType<support::little, false>>::printAttributes() { if (Sec.sh_type != ELF::SHT_ARM_ATTRIBUTES) continue; - ErrorOr<ArrayRef<uint8_t>> Contents = Obj->getSectionContents(&Sec); - if (!Contents) - continue; - - if ((*Contents)[0] != ARMBuildAttrs::Format_Version) { - errs() << "unrecognised FormatVersion: 0x" << utohexstr((*Contents)[0]) + ArrayRef<uint8_t> Contents = unwrapOrError(Obj->getSectionContents(&Sec)); + if (Contents[0] != ARMBuildAttrs::Format_Version) { + errs() << "unrecognised FormatVersion: 0x" << utohexstr(Contents[0]) << '\n'; continue; } - W.printHex("FormatVersion", (*Contents)[0]); - if (Contents->size() == 1) + W.printHex("FormatVersion", Contents[0]); + if (Contents.size() == 1) continue; - ARMAttributeParser(W).Parse(*Contents); + ARMAttributeParser(W).Parse(Contents); } } } @@ -1700,7 +1884,7 @@ public: typedef typename ELFO::Elf_Rela Elf_Rela; MipsGOTParser(ELFDumper<ELFT> *Dumper, const ELFO *Obj, - Elf_Dyn_Range DynTable, StreamWriter &W); + Elf_Dyn_Range DynTable, ScopedPrinter &W); void parseGOT(); void parsePLT(); @@ -1708,7 +1892,7 @@ public: private: ELFDumper<ELFT> *Dumper; const ELFO *Obj; - StreamWriter &W; + ScopedPrinter &W; llvm::Optional<uint64_t> DtPltGot; llvm::Optional<uint64_t> DtLocalGotNum; llvm::Optional<uint64_t> DtGotSym; @@ -1733,7 +1917,7 @@ private: template <class ELFT> MipsGOTParser<ELFT>::MipsGOTParser(ELFDumper<ELFT> *Dumper, const ELFO *Obj, - Elf_Dyn_Range DynTable, StreamWriter &W) + Elf_Dyn_Range DynTable, ScopedPrinter &W) : Dumper(Dumper), Obj(Obj), W(W) { for (const auto &Entry : DynTable) { switch (Entry.getTag()) { @@ -1773,44 +1957,33 @@ template <class ELFT> void MipsGOTParser<ELFT>::parseGOT() { return; } - const Elf_Shdr *GOTShdr = findSectionByAddress(Obj, *DtPltGot); - if (!GOTShdr) { - W.startLine() << "There is no .got section in the file.\n"; - return; - } - - ErrorOr<ArrayRef<uint8_t>> GOT = Obj->getSectionContents(GOTShdr); - if (!GOT) { - W.startLine() << "The .got section is empty.\n"; - return; - } - - if (*DtLocalGotNum > getGOTTotal(*GOT)) { - W.startLine() << "MIPS_LOCAL_GOTNO exceeds a number of GOT entries.\n"; - return; - } - - const Elf_Shdr *DynSymSec = Dumper->getDotDynSymSec(); - ErrorOr<StringRef> StrTable = Obj->getStringTableForSymtab(*DynSymSec); - error(StrTable.getError()); - const Elf_Sym *DynSymBegin = Obj->symbol_begin(DynSymSec); - const Elf_Sym *DynSymEnd = Obj->symbol_end(DynSymSec); + StringRef StrTable = Dumper->getDynamicStringTable(); + const Elf_Sym *DynSymBegin = Dumper->dynamic_symbols().begin(); + const Elf_Sym *DynSymEnd = Dumper->dynamic_symbols().end(); std::size_t DynSymTotal = std::size_t(std::distance(DynSymBegin, DynSymEnd)); - if (*DtGotSym > DynSymTotal) { - W.startLine() << "MIPS_GOTSYM exceeds a number of dynamic symbols.\n"; - return; - } + if (*DtGotSym > DynSymTotal) + report_fatal_error("MIPS_GOTSYM exceeds a number of dynamic symbols"); std::size_t GlobalGotNum = DynSymTotal - *DtGotSym; - if (*DtLocalGotNum + GlobalGotNum > getGOTTotal(*GOT)) { - W.startLine() << "Number of global GOT entries exceeds the size of GOT.\n"; + if (*DtLocalGotNum + GlobalGotNum == 0) { + W.startLine() << "GOT is empty.\n"; return; } - const GOTEntry *GotBegin = makeGOTIter(*GOT, 0); - const GOTEntry *GotLocalEnd = makeGOTIter(*GOT, *DtLocalGotNum); + const Elf_Shdr *GOTShdr = findNotEmptySectionByAddress(Obj, *DtPltGot); + if (!GOTShdr) + report_fatal_error("There is no not empty GOT section at 0x" + + Twine::utohexstr(*DtPltGot)); + + ArrayRef<uint8_t> GOT = unwrapOrError(Obj->getSectionContents(GOTShdr)); + + if (*DtLocalGotNum + GlobalGotNum > getGOTTotal(GOT)) + report_fatal_error("Number of GOT entries exceeds the size of GOT section"); + + const GOTEntry *GotBegin = makeGOTIter(GOT, 0); + const GOTEntry *GotLocalEnd = makeGOTIter(GOT, *DtLocalGotNum); const GOTEntry *It = GotBegin; DictScope GS(W, "Primary GOT"); @@ -1842,16 +2015,16 @@ template <class ELFT> void MipsGOTParser<ELFT>::parseGOT() { ListScope GS(W, "Global entries"); const GOTEntry *GotGlobalEnd = - makeGOTIter(*GOT, *DtLocalGotNum + GlobalGotNum); + makeGOTIter(GOT, *DtLocalGotNum + GlobalGotNum); const Elf_Sym *GotDynSym = DynSymBegin + *DtGotSym; for (; It != GotGlobalEnd; ++It) { DictScope D(W, "Entry"); - printGlobalGotEntry(GOTShdr->sh_addr, GotBegin, It, GotDynSym++, - *StrTable, true); + printGlobalGotEntry(GOTShdr->sh_addr, GotBegin, It, GotDynSym++, StrTable, + true); } } - std::size_t SpecGotNum = getGOTTotal(*GOT) - *DtLocalGotNum - GlobalGotNum; + std::size_t SpecGotNum = getGOTTotal(GOT) - *DtLocalGotNum - GlobalGotNum; W.printNumber("Number of TLS and multi-GOT entries", uint64_t(SpecGotNum)); } @@ -1865,31 +2038,22 @@ template <class ELFT> void MipsGOTParser<ELFT>::parsePLT() { return; } - const Elf_Shdr *PLTShdr = findSectionByAddress(Obj, *DtMipsPltGot); - if (!PLTShdr) { - W.startLine() << "There is no .got.plt section in the file.\n"; - return; - } - ErrorOr<ArrayRef<uint8_t>> PLT = Obj->getSectionContents(PLTShdr); - if (!PLT) { - W.startLine() << "The .got.plt section is empty.\n"; - return; - } + const Elf_Shdr *PLTShdr = findNotEmptySectionByAddress(Obj, *DtMipsPltGot); + if (!PLTShdr) + report_fatal_error("There is no not empty PLTGOT section at 0x " + + Twine::utohexstr(*DtMipsPltGot)); + ArrayRef<uint8_t> PLT = unwrapOrError(Obj->getSectionContents(PLTShdr)); - const Elf_Shdr *PLTRelShdr = findSectionByAddress(Obj, *DtJmpRel); - if (!PLTShdr) { - W.startLine() << "There is no .rel.plt section in the file.\n"; - return; - } - ErrorOr<const Elf_Shdr *> SymTableOrErr = - Obj->getSection(PLTRelShdr->sh_link); - error(SymTableOrErr.getError()); - const Elf_Shdr *SymTable = *SymTableOrErr; - ErrorOr<StringRef> StrTable = Obj->getStringTableForSymtab(*SymTable); - error(StrTable.getError()); + const Elf_Shdr *PLTRelShdr = findNotEmptySectionByAddress(Obj, *DtJmpRel); + if (!PLTRelShdr) + report_fatal_error("There is no not empty RELPLT section at 0x" + + Twine::utohexstr(*DtJmpRel)); + const Elf_Shdr *SymTable = + unwrapOrError(Obj->getSection(PLTRelShdr->sh_link)); + StringRef StrTable = unwrapOrError(Obj->getStringTableForSymtab(*SymTable)); - const GOTEntry *PLTBegin = makeGOTIter(*PLT, 0); - const GOTEntry *PLTEnd = makeGOTIter(*PLT, getGOTTotal(*PLT)); + const GOTEntry *PLTBegin = makeGOTIter(PLT, 0); + const GOTEntry *PLTEnd = makeGOTIter(PLT, getGOTTotal(PLT)); const GOTEntry *It = PLTBegin; DictScope GS(W, "PLT GOT"); @@ -1908,7 +2072,7 @@ template <class ELFT> void MipsGOTParser<ELFT>::parsePLT() { *RE = Obj->rel_end(PLTRelShdr); RI != RE && It != PLTEnd; ++RI, ++It) { const Elf_Sym *Sym = Obj->getRelocationSymbol(&*RI, SymTable); - printPLTEntry(PLTShdr->sh_addr, PLTBegin, It, *StrTable, Sym); + printPLTEntry(PLTShdr->sh_addr, PLTBegin, It, StrTable, Sym); } break; case ELF::SHT_RELA: @@ -1916,7 +2080,7 @@ template <class ELFT> void MipsGOTParser<ELFT>::parsePLT() { *RE = Obj->rela_end(PLTRelShdr); RI != RE && It != PLTEnd; ++RI, ++It) { const Elf_Sym *Sym = Obj->getRelocationSymbol(&*RI, SymTable); - printPLTEntry(PLTShdr->sh_addr, PLTBegin, It, *StrTable, Sym); + printPLTEntry(PLTShdr->sh_addr, PLTBegin, It, StrTable, Sym); } break; } @@ -1956,7 +2120,7 @@ void MipsGOTParser<ELFT>::printGlobalGotEntry( unsigned SectionIndex = 0; StringRef SectionName; - getSectionNameIndex(*Obj, Sym, Dumper->getDotDynSymSec(), + getSectionNameIndex(*Obj, Sym, Dumper->dynamic_symbols().begin(), Dumper->getShndxTable(), SectionName, SectionIndex); W.printHex("Section", SectionName, SectionIndex); @@ -1990,7 +2154,7 @@ void MipsGOTParser<ELFT>::printPLTEntry(uint64_t PLTAddr, unsigned SectionIndex = 0; StringRef SectionName; - getSectionNameIndex(*Obj, Sym, Dumper->getDotDynSymSec(), + getSectionNameIndex(*Obj, Sym, Dumper->dynamic_symbols().begin(), Dumper->getShndxTable(), SectionName, SectionIndex); W.printHex("Section", SectionName, SectionIndex); @@ -2086,17 +2250,13 @@ template <class ELFT> void ELFDumper<ELFT>::printMipsABIFlags() { W.startLine() << "There is no .MIPS.abiflags section in the file.\n"; return; } - ErrorOr<ArrayRef<uint8_t>> Sec = Obj->getSectionContents(Shdr); - if (!Sec) { - W.startLine() << "The .MIPS.abiflags section is empty.\n"; - return; - } - if (Sec->size() != sizeof(Elf_Mips_ABIFlags<ELFT>)) { + ArrayRef<uint8_t> Sec = unwrapOrError(Obj->getSectionContents(Shdr)); + if (Sec.size() != sizeof(Elf_Mips_ABIFlags<ELFT>)) { W.startLine() << "The .MIPS.abiflags section has a wrong size.\n"; return; } - auto *Flags = reinterpret_cast<const Elf_Mips_ABIFlags<ELFT> *>(Sec->data()); + auto *Flags = reinterpret_cast<const Elf_Mips_ABIFlags<ELFT> *>(Sec.data()); raw_ostream &OS = W.getOStream(); DictScope GS(W, "MIPS ABI Flags"); @@ -2118,38 +2278,68 @@ template <class ELFT> void ELFDumper<ELFT>::printMipsABIFlags() { W.printHex("Flags 2", Flags->flags2); } +template <class ELFT> +static void printMipsReginfoData(ScopedPrinter &W, + const Elf_Mips_RegInfo<ELFT> &Reginfo) { + W.printHex("GP", Reginfo.ri_gp_value); + W.printHex("General Mask", Reginfo.ri_gprmask); + W.printHex("Co-Proc Mask0", Reginfo.ri_cprmask[0]); + W.printHex("Co-Proc Mask1", Reginfo.ri_cprmask[1]); + W.printHex("Co-Proc Mask2", Reginfo.ri_cprmask[2]); + W.printHex("Co-Proc Mask3", Reginfo.ri_cprmask[3]); +} + template <class ELFT> void ELFDumper<ELFT>::printMipsReginfo() { const Elf_Shdr *Shdr = findSectionByName(*Obj, ".reginfo"); if (!Shdr) { W.startLine() << "There is no .reginfo section in the file.\n"; return; } - ErrorOr<ArrayRef<uint8_t>> Sec = Obj->getSectionContents(Shdr); - if (!Sec) { - W.startLine() << "The .reginfo section is empty.\n"; + ArrayRef<uint8_t> Sec = unwrapOrError(Obj->getSectionContents(Shdr)); + if (Sec.size() != sizeof(Elf_Mips_RegInfo<ELFT>)) { + W.startLine() << "The .reginfo section has a wrong size.\n"; return; } - if (Sec->size() != sizeof(Elf_Mips_RegInfo<ELFT>)) { - W.startLine() << "The .reginfo section has a wrong size.\n"; + + DictScope GS(W, "MIPS RegInfo"); + auto *Reginfo = reinterpret_cast<const Elf_Mips_RegInfo<ELFT> *>(Sec.data()); + printMipsReginfoData(W, *Reginfo); +} + +template <class ELFT> void ELFDumper<ELFT>::printMipsOptions() { + const Elf_Shdr *Shdr = findSectionByName(*Obj, ".MIPS.options"); + if (!Shdr) { + W.startLine() << "There is no .MIPS.options section in the file.\n"; return; } - auto *Reginfo = reinterpret_cast<const Elf_Mips_RegInfo<ELFT> *>(Sec->data()); + DictScope GS(W, "MIPS Options"); - DictScope GS(W, "MIPS RegInfo"); - W.printHex("GP", Reginfo->ri_gp_value); - W.printHex("General Mask", Reginfo->ri_gprmask); - W.printHex("Co-Proc Mask0", Reginfo->ri_cprmask[0]); - W.printHex("Co-Proc Mask1", Reginfo->ri_cprmask[1]); - W.printHex("Co-Proc Mask2", Reginfo->ri_cprmask[2]); - W.printHex("Co-Proc Mask3", Reginfo->ri_cprmask[3]); + ArrayRef<uint8_t> Sec = unwrapOrError(Obj->getSectionContents(Shdr)); + while (!Sec.empty()) { + if (Sec.size() < sizeof(Elf_Mips_Options<ELFT>)) { + W.startLine() << "The .MIPS.options section has a wrong size.\n"; + return; + } + auto *O = reinterpret_cast<const Elf_Mips_Options<ELFT> *>(Sec.data()); + DictScope GS(W, getElfMipsOptionsOdkType(O->kind)); + switch (O->kind) { + case ODK_REGINFO: + printMipsReginfoData(W, O->getRegInfo()); + break; + default: + W.startLine() << "Unsupported MIPS options tag.\n"; + break; + } + Sec = Sec.slice(O->size); + } } template <class ELFT> void ELFDumper<ELFT>::printStackMap() const { const Elf_Shdr *StackMapSection = nullptr; for (const auto &Sec : Obj->sections()) { - ErrorOr<StringRef> Name = Obj->getSectionName(&Sec); - if (*Name == ".llvm_stackmaps") { + StringRef Name = unwrapOrError(Obj->getSectionName(&Sec)); + if (Name == ".llvm_stackmaps") { StackMapSection = &Sec; break; } @@ -2159,10 +2349,1184 @@ template <class ELFT> void ELFDumper<ELFT>::printStackMap() const { return; StringRef StackMapContents; - ErrorOr<ArrayRef<uint8_t>> StackMapContentsArray = - Obj->getSectionContents(StackMapSection); + ArrayRef<uint8_t> StackMapContentsArray = + unwrapOrError(Obj->getSectionContents(StackMapSection)); + + prettyPrintStackMap(llvm::outs(), StackMapV1Parser<ELFT::TargetEndianness>( + StackMapContentsArray)); +} + +template <class ELFT> void ELFDumper<ELFT>::printGroupSections() { + ELFDumperStyle->printGroupSections(Obj); +} + +static inline void printFields(formatted_raw_ostream &OS, StringRef Str1, + StringRef Str2) { + OS.PadToColumn(2u); + OS << Str1; + OS.PadToColumn(37u); + OS << Str2 << "\n"; + OS.flush(); +} + +template <class ELFT> void GNUStyle<ELFT>::printFileHeaders(const ELFO *Obj) { + const Elf_Ehdr *e = Obj->getHeader(); + OS << "ELF Header:\n"; + OS << " Magic: "; + std::string Str; + for (int i = 0; i < ELF::EI_NIDENT; i++) + OS << format(" %02x", static_cast<int>(e->e_ident[i])); + OS << "\n"; + Str = printEnum(e->e_ident[ELF::EI_CLASS], makeArrayRef(ElfClass)); + printFields(OS, "Class:", Str); + Str = printEnum(e->e_ident[ELF::EI_DATA], makeArrayRef(ElfDataEncoding)); + printFields(OS, "Data:", Str); + OS.PadToColumn(2u); + OS << "Version:"; + OS.PadToColumn(37u); + OS << to_hexString(e->e_ident[ELF::EI_VERSION]); + if (e->e_version == ELF::EV_CURRENT) + OS << " (current)"; + OS << "\n"; + Str = printEnum(e->e_ident[ELF::EI_OSABI], makeArrayRef(ElfOSABI)); + printFields(OS, "OS/ABI:", Str); + Str = "0x" + to_hexString(e->e_version); + Str = to_hexString(e->e_ident[ELF::EI_ABIVERSION]); + printFields(OS, "ABI Version:", Str); + Str = printEnum(e->e_type, makeArrayRef(ElfObjectFileType)); + printFields(OS, "Type:", Str); + Str = printEnum(e->e_machine, makeArrayRef(ElfMachineType)); + printFields(OS, "Machine:", Str); + Str = "0x" + to_hexString(e->e_version); + printFields(OS, "Version:", Str); + Str = "0x" + to_hexString(e->e_entry); + printFields(OS, "Entry point address:", Str); + Str = to_string(e->e_phoff) + " (bytes into file)"; + printFields(OS, "Start of program headers:", Str); + Str = to_string(e->e_shoff) + " (bytes into file)"; + printFields(OS, "Start of section headers:", Str); + Str = "0x" + to_hexString(e->e_flags); + printFields(OS, "Flags:", Str); + Str = to_string(e->e_ehsize) + " (bytes)"; + printFields(OS, "Size of this header:", Str); + Str = to_string(e->e_phentsize) + " (bytes)"; + printFields(OS, "Size of program headers:", Str); + Str = to_string(e->e_phnum); + printFields(OS, "Number of program headers:", Str); + Str = to_string(e->e_shentsize) + " (bytes)"; + printFields(OS, "Size of section headers:", Str); + Str = to_string(e->e_shnum); + printFields(OS, "Number of section headers:", Str); + Str = to_string(e->e_shstrndx); + printFields(OS, "Section header string table index:", Str); +} + +template <class ELFT> void GNUStyle<ELFT>::printGroupSections(const ELFO *Obj) { + uint32_t SectionIndex = 0; + bool HasGroups = false; + for (const Elf_Shdr &Sec : Obj->sections()) { + if (Sec.sh_type == ELF::SHT_GROUP) { + HasGroups = true; + const Elf_Shdr *Symtab = unwrapOrError(Obj->getSection(Sec.sh_link)); + StringRef StrTable = unwrapOrError(Obj->getStringTableForSymtab(*Symtab)); + const Elf_Sym *Signature = + Obj->template getEntry<Elf_Sym>(Symtab, Sec.sh_info); + ArrayRef<Elf_Word> Data = unwrapOrError( + Obj->template getSectionContentsAsArray<Elf_Word>(&Sec)); + StringRef Name = unwrapOrError(Obj->getSectionName(&Sec)); + OS << "\n" << getGroupType(Data[0]) << " group section [" + << format_decimal(SectionIndex, 5) << "] `" << Name << "' [" + << StrTable.data() + Signature->st_name << "] contains " + << (Data.size() - 1) << " sections:\n" + << " [Index] Name\n"; + for (auto &Ndx : Data.slice(1)) { + auto Sec = unwrapOrError(Obj->getSection(Ndx)); + const StringRef Name = unwrapOrError(Obj->getSectionName(Sec)); + OS << " [" << format_decimal(Ndx, 5) << "] " << Name + << "\n"; + } + } + ++SectionIndex; + } + if (!HasGroups) + OS << "There are no section groups in this file.\n"; +} + +template <class ELFT> +void GNUStyle<ELFT>::printRelocation(const ELFO *Obj, const Elf_Shdr *SymTab, + const Elf_Rela &R, bool IsRela) { + std::string Offset, Info, Addend = "", Value; + SmallString<32> RelocName; + StringRef StrTable = unwrapOrError(Obj->getStringTableForSymtab(*SymTab)); + StringRef TargetName; + const Elf_Sym *Sym = nullptr; + unsigned Width = ELFT::Is64Bits ? 16 : 8; + unsigned Bias = ELFT::Is64Bits ? 8 : 0; + + // First two fields are bit width dependent. The rest of them are after are + // fixed width. + Field Fields[5] = {0, 10 + Bias, 19 + 2 * Bias, 42 + 2 * Bias, 53 + 2 * Bias}; + Obj->getRelocationTypeName(R.getType(Obj->isMips64EL()), RelocName); + Sym = Obj->getRelocationSymbol(&R, SymTab); + if (Sym && Sym->getType() == ELF::STT_SECTION) { + const Elf_Shdr *Sec = unwrapOrError( + Obj->getSection(Sym, SymTab, this->dumper()->getShndxTable())); + TargetName = unwrapOrError(Obj->getSectionName(Sec)); + } else if (Sym) { + TargetName = unwrapOrError(Sym->getName(StrTable)); + } + + if (Sym && IsRela) { + if (R.r_addend < 0) + Addend = " - "; + else + Addend = " + "; + } + + Offset = to_string(format_hex_no_prefix(R.r_offset, Width)); + Info = to_string(format_hex_no_prefix(R.r_info, Width)); + + int64_t RelAddend = R.r_addend; + if (IsRela) + Addend += to_hexString(std::abs(RelAddend), false); + + if (Sym) + Value = to_string(format_hex_no_prefix(Sym->getValue(), Width)); + + Fields[0].Str = Offset; + Fields[1].Str = Info; + Fields[2].Str = RelocName; + Fields[3].Str = Value; + Fields[4].Str = TargetName; + for (auto &field : Fields) + printField(field); + OS << Addend; + OS << "\n"; +} + +static inline void printRelocHeader(raw_ostream &OS, bool Is64, bool IsRela) { + if (Is64) + OS << " Offset Info Type" + << " Symbol's Value Symbol's Name"; + else + OS << " Offset Info Type Sym. Value " + << "Symbol's Name"; + if (IsRela) + OS << (IsRela ? " + Addend" : ""); + OS << "\n"; +} + +template <class ELFT> void GNUStyle<ELFT>::printRelocations(const ELFO *Obj) { + bool HasRelocSections = false; + for (const Elf_Shdr &Sec : Obj->sections()) { + if (Sec.sh_type != ELF::SHT_REL && Sec.sh_type != ELF::SHT_RELA) + continue; + HasRelocSections = true; + StringRef Name = unwrapOrError(Obj->getSectionName(&Sec)); + unsigned Entries = Sec.getEntityCount(); + uintX_t Offset = Sec.sh_offset; + OS << "\nRelocation section '" << Name << "' at offset 0x" + << to_hexString(Offset, false) << " contains " << Entries + << " entries:\n"; + printRelocHeader(OS, ELFT::Is64Bits, (Sec.sh_type == ELF::SHT_RELA)); + const Elf_Shdr *SymTab = unwrapOrError(Obj->getSection(Sec.sh_link)); + if (Sec.sh_type == ELF::SHT_REL) { + for (const auto &R : Obj->rels(&Sec)) { + Elf_Rela Rela; + Rela.r_offset = R.r_offset; + Rela.r_info = R.r_info; + Rela.r_addend = 0; + printRelocation(Obj, SymTab, Rela, false); + } + } else { + for (const auto &R : Obj->relas(&Sec)) + printRelocation(Obj, SymTab, R, true); + } + } + if (!HasRelocSections) + OS << "\nThere are no relocations in this file.\n"; +} + +std::string getSectionTypeString(unsigned Arch, unsigned Type) { + using namespace ELF; + switch (Arch) { + case EM_ARM: + switch (Type) { + case SHT_ARM_EXIDX: + return "ARM_EXIDX"; + case SHT_ARM_PREEMPTMAP: + return "ARM_PREEMPTMAP"; + case SHT_ARM_ATTRIBUTES: + return "ARM_ATTRIBUTES"; + case SHT_ARM_DEBUGOVERLAY: + return "ARM_DEBUGOVERLAY"; + case SHT_ARM_OVERLAYSECTION: + return "ARM_OVERLAYSECTION"; + } + case EM_X86_64: + switch (Type) { + case SHT_X86_64_UNWIND: + return "X86_64_UNWIND"; + } + case EM_MIPS: + case EM_MIPS_RS3_LE: + switch (Type) { + case SHT_MIPS_REGINFO: + return "MIPS_REGINFO"; + case SHT_MIPS_OPTIONS: + return "MIPS_OPTIONS"; + case SHT_MIPS_ABIFLAGS: + return "MIPS_ABIFLAGS"; + } + } + switch (Type) { + case SHT_NULL: + return "NULL"; + case SHT_PROGBITS: + return "PROGBITS"; + case SHT_SYMTAB: + return "SYMTAB"; + case SHT_STRTAB: + return "STRTAB"; + case SHT_RELA: + return "RELA"; + case SHT_HASH: + return "HASH"; + case SHT_DYNAMIC: + return "DYNAMIC"; + case SHT_NOTE: + return "NOTE"; + case SHT_NOBITS: + return "NOBITS"; + case SHT_REL: + return "REL"; + case SHT_SHLIB: + return "SHLIB"; + case SHT_DYNSYM: + return "DYNSYM"; + case SHT_INIT_ARRAY: + return "INIT_ARRAY"; + case SHT_FINI_ARRAY: + return "FINI_ARRAY"; + case SHT_PREINIT_ARRAY: + return "PREINIT_ARRAY"; + case SHT_GROUP: + return "GROUP"; + case SHT_SYMTAB_SHNDX: + return "SYMTAB SECTION INDICES"; + // FIXME: Parse processor specific GNU attributes + case SHT_GNU_ATTRIBUTES: + return "ATTRIBUTES"; + case SHT_GNU_HASH: + return "GNU_HASH"; + case SHT_GNU_verdef: + return "VERDEF"; + case SHT_GNU_verneed: + return "VERNEED"; + case SHT_GNU_versym: + return "VERSYM"; + default: + return ""; + } + return ""; +} + +template <class ELFT> void GNUStyle<ELFT>::printSections(const ELFO *Obj) { + size_t SectionIndex = 0; + std::string Number, Type, Size, Address, Offset, Flags, Link, Info, EntrySize, + Alignment; + unsigned Bias; + unsigned Width; + + if (ELFT::Is64Bits) { + Bias = 0; + Width = 16; + } else { + Bias = 8; + Width = 8; + } + OS << "There are " << to_string(Obj->getHeader()->e_shnum) + << " section headers, starting at offset " + << "0x" << to_hexString(Obj->getHeader()->e_shoff, false) << ":\n\n"; + OS << "Section Headers:\n"; + Field Fields[11] = {{"[Nr]", 2}, + {"Name", 7}, + {"Type", 25}, + {"Address", 41}, + {"Off", 58 - Bias}, + {"Size", 65 - Bias}, + {"ES", 72 - Bias}, + {"Flg", 75 - Bias}, + {"Lk", 79 - Bias}, + {"Inf", 82 - Bias}, + {"Al", 86 - Bias}}; + for (auto &f : Fields) + printField(f); + OS << "\n"; + + for (const Elf_Shdr &Sec : Obj->sections()) { + Number = to_string(SectionIndex); + Fields[0].Str = Number; + Fields[1].Str = unwrapOrError(Obj->getSectionName(&Sec)); + Type = getSectionTypeString(Obj->getHeader()->e_machine, Sec.sh_type); + Fields[2].Str = Type; + Address = to_string(format_hex_no_prefix(Sec.sh_addr, Width)); + Fields[3].Str = Address; + Offset = to_string(format_hex_no_prefix(Sec.sh_offset, 6)); + Fields[4].Str = Offset; + Size = to_string(format_hex_no_prefix(Sec.sh_size, 6)); + Fields[5].Str = Size; + EntrySize = to_string(format_hex_no_prefix(Sec.sh_entsize, 2)); + Fields[6].Str = EntrySize; + Flags = getGNUFlags(Sec.sh_flags); + Fields[7].Str = Flags; + Link = to_string(Sec.sh_link); + Fields[8].Str = Link; + Info = to_string(Sec.sh_info); + Fields[9].Str = Info; + Alignment = to_string(Sec.sh_addralign); + Fields[10].Str = Alignment; + OS.PadToColumn(Fields[0].Column); + OS << "[" << right_justify(Fields[0].Str, 2) << "]"; + for (int i = 1; i < 7; i++) + printField(Fields[i]); + OS.PadToColumn(Fields[7].Column); + OS << right_justify(Fields[7].Str, 3); + OS.PadToColumn(Fields[8].Column); + OS << right_justify(Fields[8].Str, 2); + OS.PadToColumn(Fields[9].Column); + OS << right_justify(Fields[9].Str, 3); + OS.PadToColumn(Fields[10].Column); + OS << right_justify(Fields[10].Str, 2); + OS << "\n"; + ++SectionIndex; + } + OS << "Key to Flags:\n" + << " W (write), A (alloc), X (execute), M (merge), S (strings), l " + "(large)\n" + << " I (info), L (link order), G (group), T (TLS), E (exclude),\ + x (unknown)\n" + << " O (extra OS processing required) o (OS specific),\ + p (processor specific)\n"; +} + +template <class ELFT> +void GNUStyle<ELFT>::printSymtabMessage(const ELFO *Obj, StringRef Name, + size_t Entries) { + if (Name.size()) + OS << "\nSymbol table '" << Name << "' contains " << Entries + << " entries:\n"; + else + OS << "\n Symbol table for image:\n"; + + if (ELFT::Is64Bits) + OS << " Num: Value Size Type Bind Vis Ndx Name\n"; + else + OS << " Num: Value Size Type Bind Vis Ndx Name\n"; +} + +template <class ELFT> +std::string GNUStyle<ELFT>::getSymbolSectionNdx(const ELFO *Obj, + const Elf_Sym *Symbol, + const Elf_Sym *FirstSym) { + unsigned SectionIndex = Symbol->st_shndx; + switch (SectionIndex) { + case ELF::SHN_UNDEF: + return "UND"; + case ELF::SHN_ABS: + return "ABS"; + case ELF::SHN_COMMON: + return "COM"; + case ELF::SHN_XINDEX: + SectionIndex = Obj->getExtendedSymbolTableIndex( + Symbol, FirstSym, this->dumper()->getShndxTable()); + default: + // Find if: + // Processor specific + if (SectionIndex >= ELF::SHN_LOPROC && SectionIndex <= ELF::SHN_HIPROC) + return std::string("PRC[0x") + + to_string(format_hex_no_prefix(SectionIndex, 4)) + "]"; + // OS specific + if (SectionIndex >= ELF::SHN_LOOS && SectionIndex <= ELF::SHN_HIOS) + return std::string("OS[0x") + + to_string(format_hex_no_prefix(SectionIndex, 4)) + "]"; + // Architecture reserved: + if (SectionIndex >= ELF::SHN_LORESERVE && + SectionIndex <= ELF::SHN_HIRESERVE) + return std::string("RSV[0x") + + to_string(format_hex_no_prefix(SectionIndex, 4)) + "]"; + // A normal section with an index + return to_string(format_decimal(SectionIndex, 3)); + } +} + +template <class ELFT> +void GNUStyle<ELFT>::printSymbol(const ELFO *Obj, const Elf_Sym *Symbol, + const Elf_Sym *FirstSym, StringRef StrTable, + bool IsDynamic) { + static int Idx = 0; + static bool Dynamic = true; + size_t Width; + + // If this function was called with a different value from IsDynamic + // from last call, happens when we move from dynamic to static symbol + // table, "Num" field should be reset. + if (!Dynamic != !IsDynamic) { + Idx = 0; + Dynamic = false; + } + std::string Num, Name, Value, Size, Binding, Type, Visibility, Section; + unsigned Bias = 0; + if (ELFT::Is64Bits) { + Bias = 8; + Width = 16; + } else { + Bias = 0; + Width = 8; + } + Field Fields[8] = {0, 8, 17 + Bias, 23 + Bias, + 31 + Bias, 38 + Bias, 47 + Bias, 51 + Bias}; + Num = to_string(format_decimal(Idx++, 6)) + ":"; + Value = to_string(format_hex_no_prefix(Symbol->st_value, Width)); + Size = to_string(format_decimal(Symbol->st_size, 5)); + unsigned char SymbolType = Symbol->getType(); + if (Obj->getHeader()->e_machine == ELF::EM_AMDGPU && + SymbolType >= ELF::STT_LOOS && SymbolType < ELF::STT_HIOS) + Type = printEnum(SymbolType, makeArrayRef(AMDGPUSymbolTypes)); + else + Type = printEnum(SymbolType, makeArrayRef(ElfSymbolTypes)); + unsigned Vis = Symbol->getVisibility(); + Binding = printEnum(Symbol->getBinding(), makeArrayRef(ElfSymbolBindings)); + Visibility = printEnum(Vis, makeArrayRef(ElfSymbolVisibilities)); + Section = getSymbolSectionNdx(Obj, Symbol, FirstSym); + Name = this->dumper()->getFullSymbolName(Symbol, StrTable, IsDynamic); + Fields[0].Str = Num; + Fields[1].Str = Value; + Fields[2].Str = Size; + Fields[3].Str = Type; + Fields[4].Str = Binding; + Fields[5].Str = Visibility; + Fields[6].Str = Section; + Fields[7].Str = Name; + for (auto &Entry : Fields) + printField(Entry); + OS << "\n"; +} + +template <class ELFT> void GNUStyle<ELFT>::printSymbols(const ELFO *Obj) { + this->dumper()->printSymbolsHelper(true); + this->dumper()->printSymbolsHelper(false); +} + +template <class ELFT> +void GNUStyle<ELFT>::printDynamicSymbols(const ELFO *Obj) { + this->dumper()->printSymbolsHelper(true); +} + +static inline std::string printPhdrFlags(unsigned Flag) { + std::string Str; + Str = (Flag & PF_R) ? "R" : " "; + Str += (Flag & PF_W) ? "W" : " "; + Str += (Flag & PF_X) ? "E" : " "; + return Str; +} + +// SHF_TLS sections are only in PT_TLS, PT_LOAD or PT_GNU_RELRO +// PT_TLS must only have SHF_TLS sections +template <class ELFT> +bool GNUStyle<ELFT>::checkTLSSections(const Elf_Phdr &Phdr, + const Elf_Shdr &Sec) { + return (((Sec.sh_flags & ELF::SHF_TLS) && + ((Phdr.p_type == ELF::PT_TLS) || (Phdr.p_type == ELF::PT_LOAD) || + (Phdr.p_type == ELF::PT_GNU_RELRO))) || + (!(Sec.sh_flags & ELF::SHF_TLS) && Phdr.p_type != ELF::PT_TLS)); +} + +// Non-SHT_NOBITS must have its offset inside the segment +// Only non-zero section can be at end of segment +template <class ELFT> +bool GNUStyle<ELFT>::checkoffsets(const Elf_Phdr &Phdr, const Elf_Shdr &Sec) { + if (Sec.sh_type == ELF::SHT_NOBITS) + return true; + bool IsSpecial = + (Sec.sh_type == ELF::SHT_NOBITS) && ((Sec.sh_flags & ELF::SHF_TLS) != 0); + // .tbss is special, it only has memory in PT_TLS and has NOBITS properties + auto SectionSize = + (IsSpecial && Phdr.p_type != ELF::PT_TLS) ? 0 : Sec.sh_size; + if (Sec.sh_offset >= Phdr.p_offset) + return ((Sec.sh_offset + SectionSize <= Phdr.p_filesz + Phdr.p_offset) + /*only non-zero sized sections at end*/ && + (Sec.sh_offset + 1 <= Phdr.p_offset + Phdr.p_filesz)); + return false; +} + +// SHF_ALLOC must have VMA inside segment +// Only non-zero section can be at end of segment +template <class ELFT> +bool GNUStyle<ELFT>::checkVMA(const Elf_Phdr &Phdr, const Elf_Shdr &Sec) { + if (!(Sec.sh_flags & ELF::SHF_ALLOC)) + return true; + bool IsSpecial = + (Sec.sh_type == ELF::SHT_NOBITS) && ((Sec.sh_flags & ELF::SHF_TLS) != 0); + // .tbss is special, it only has memory in PT_TLS and has NOBITS properties + auto SectionSize = + (IsSpecial && Phdr.p_type != ELF::PT_TLS) ? 0 : Sec.sh_size; + if (Sec.sh_addr >= Phdr.p_vaddr) + return ((Sec.sh_addr + SectionSize <= Phdr.p_vaddr + Phdr.p_memsz) && + (Sec.sh_addr + 1 <= Phdr.p_vaddr + Phdr.p_memsz)); + return false; +} + +// No section with zero size must be at start or end of PT_DYNAMIC +template <class ELFT> +bool GNUStyle<ELFT>::checkPTDynamic(const Elf_Phdr &Phdr, const Elf_Shdr &Sec) { + if (Phdr.p_type != ELF::PT_DYNAMIC || Sec.sh_size != 0 || Phdr.p_memsz == 0) + return true; + // Is section within the phdr both based on offset and VMA ? + return ((Sec.sh_type == ELF::SHT_NOBITS) || + (Sec.sh_offset > Phdr.p_offset && + Sec.sh_offset < Phdr.p_offset + Phdr.p_filesz)) && + (!(Sec.sh_flags & ELF::SHF_ALLOC) || + (Sec.sh_addr > Phdr.p_vaddr && Sec.sh_addr < Phdr.p_memsz)); +} + +template <class ELFT> +void GNUStyle<ELFT>::printProgramHeaders(const ELFO *Obj) { + unsigned Bias = ELFT::Is64Bits ? 8 : 0; + unsigned Width = ELFT::Is64Bits ? 18 : 10; + unsigned SizeWidth = ELFT::Is64Bits ? 8 : 7; + std::string Type, Offset, VMA, LMA, FileSz, MemSz, Flag, Align; + + const Elf_Ehdr *Header = Obj->getHeader(); + Field Fields[8] = {2, 17, 26, 37 + Bias, + 48 + Bias, 56 + Bias, 64 + Bias, 68 + Bias}; + OS << "\nElf file type is " + << printEnum(Header->e_type, makeArrayRef(ElfObjectFileType)) << "\n" + << "Entry point " << format_hex(Header->e_entry, 3) << "\n" + << "There are " << Header->e_phnum << " program headers," + << " starting at offset " << Header->e_phoff << "\n\n" + << "Program Headers:\n"; + if (ELFT::Is64Bits) + OS << " Type Offset VirtAddr PhysAddr " + << " FileSiz MemSiz Flg Align\n"; + else + OS << " Type Offset VirtAddr PhysAddr FileSiz " + << "MemSiz Flg Align\n"; + for (const auto &Phdr : Obj->program_headers()) { + Type = getElfPtType(Header->e_machine, Phdr.p_type); + Offset = to_string(format_hex(Phdr.p_offset, 8)); + VMA = to_string(format_hex(Phdr.p_vaddr, Width)); + LMA = to_string(format_hex(Phdr.p_paddr, Width)); + FileSz = to_string(format_hex(Phdr.p_filesz, SizeWidth)); + MemSz = to_string(format_hex(Phdr.p_memsz, SizeWidth)); + Flag = printPhdrFlags(Phdr.p_flags); + Align = to_string(format_hex(Phdr.p_align, 1)); + Fields[0].Str = Type; + Fields[1].Str = Offset; + Fields[2].Str = VMA; + Fields[3].Str = LMA; + Fields[4].Str = FileSz; + Fields[5].Str = MemSz; + Fields[6].Str = Flag; + Fields[7].Str = Align; + for (auto Field : Fields) + printField(Field); + if (Phdr.p_type == ELF::PT_INTERP) { + OS << "\n [Requesting program interpreter: "; + OS << reinterpret_cast<const char *>(Obj->base()) + Phdr.p_offset << "]"; + } + OS << "\n"; + } + OS << "\n Section to Segment mapping:\n Segment Sections...\n"; + int Phnum = 0; + for (const Elf_Phdr &Phdr : Obj->program_headers()) { + std::string Sections; + OS << format(" %2.2d ", Phnum++); + for (const Elf_Shdr &Sec : Obj->sections()) { + // Check if each section is in a segment and then print mapping. + // readelf additionally makes sure it does not print zero sized sections + // at end of segments and for PT_DYNAMIC both start and end of section + // .tbss must only be shown in PT_TLS section. + bool TbssInNonTLS = (Sec.sh_type == ELF::SHT_NOBITS) && + ((Sec.sh_flags & ELF::SHF_TLS) != 0) && + Phdr.p_type != ELF::PT_TLS; + if (!TbssInNonTLS && checkTLSSections(Phdr, Sec) && + checkoffsets(Phdr, Sec) && checkVMA(Phdr, Sec) && + checkPTDynamic(Phdr, Sec) && (Sec.sh_type != ELF::SHT_NULL)) + Sections += unwrapOrError(Obj->getSectionName(&Sec)).str() + " "; + } + OS << Sections << "\n"; + OS.flush(); + } +} + +template <class ELFT> +void GNUStyle<ELFT>::printDynamicRelocation(const ELFO *Obj, Elf_Rela R, + bool IsRela) { + SmallString<32> RelocName; + StringRef SymbolName; + unsigned Width = ELFT::Is64Bits ? 16 : 8; + unsigned Bias = ELFT::Is64Bits ? 8 : 0; + // First two fields are bit width dependent. The rest of them are after are + // fixed width. + Field Fields[5] = {0, 10 + Bias, 19 + 2 * Bias, 42 + 2 * Bias, 53 + 2 * Bias}; + + uint32_t SymIndex = R.getSymbol(Obj->isMips64EL()); + const Elf_Sym *Sym = this->dumper()->dynamic_symbols().begin() + SymIndex; + Obj->getRelocationTypeName(R.getType(Obj->isMips64EL()), RelocName); + SymbolName = + unwrapOrError(Sym->getName(this->dumper()->getDynamicStringTable())); + std::string Addend = "", Info, Offset, Value; + Offset = to_string(format_hex_no_prefix(R.r_offset, Width)); + Info = to_string(format_hex_no_prefix(R.r_info, Width)); + Value = to_string(format_hex_no_prefix(Sym->getValue(), Width)); + int64_t RelAddend = R.r_addend; + if (SymbolName.size() && IsRela) { + if (R.r_addend < 0) + Addend = " - "; + else + Addend = " + "; + } + + if (!SymbolName.size() && Sym->getValue() == 0) + Value = ""; + + if (IsRela) + Addend += to_string(format_hex_no_prefix(std::abs(RelAddend), 1)); + + + Fields[0].Str = Offset; + Fields[1].Str = Info; + Fields[2].Str = RelocName.c_str(); + Fields[3].Str = Value; + Fields[4].Str = SymbolName; + for (auto &Field : Fields) + printField(Field); + OS << Addend; + OS << "\n"; +} + +template <class ELFT> +void GNUStyle<ELFT>::printDynamicRelocations(const ELFO *Obj) { + const DynRegionInfo &DynRelRegion = this->dumper()->getDynRelRegion(); + const DynRegionInfo &DynRelaRegion = this->dumper()->getDynRelaRegion(); + const DynRegionInfo &DynPLTRelRegion = this->dumper()->getDynPLTRelRegion(); + if (DynRelaRegion.Size > 0) { + OS << "\n'RELA' relocation section at offset " + << format_hex(reinterpret_cast<const uint8_t *>(DynRelaRegion.Addr) - + Obj->base(), + 1) << " contains " << DynRelaRegion.Size << " bytes:\n"; + printRelocHeader(OS, ELFT::Is64Bits, true); + for (const Elf_Rela &Rela : this->dumper()->dyn_relas()) + printDynamicRelocation(Obj, Rela, true); + } + if (DynRelRegion.Size > 0) { + OS << "\n'REL' relocation section at offset " + << format_hex(reinterpret_cast<const uint8_t *>(DynRelRegion.Addr) - + Obj->base(), + 1) << " contains " << DynRelRegion.Size << " bytes:\n"; + printRelocHeader(OS, ELFT::Is64Bits, false); + for (const Elf_Rel &Rel : this->dumper()->dyn_rels()) { + Elf_Rela Rela; + Rela.r_offset = Rel.r_offset; + Rela.r_info = Rel.r_info; + Rela.r_addend = 0; + printDynamicRelocation(Obj, Rela, false); + } + } + if (DynPLTRelRegion.Size) { + OS << "\n'PLT' relocation section at offset " + << format_hex(reinterpret_cast<const uint8_t *>(DynPLTRelRegion.Addr) - + Obj->base(), + 1) << " contains " << DynPLTRelRegion.Size << " bytes:\n"; + } + if (DynPLTRelRegion.EntSize == sizeof(Elf_Rela)) { + printRelocHeader(OS, ELFT::Is64Bits, true); + for (const Elf_Rela &Rela : DynPLTRelRegion.getAsArrayRef<Elf_Rela>()) + printDynamicRelocation(Obj, Rela, true); + } else { + printRelocHeader(OS, ELFT::Is64Bits, false); + for (const Elf_Rel &Rel : DynPLTRelRegion.getAsArrayRef<Elf_Rel>()) { + Elf_Rela Rela; + Rela.r_offset = Rel.r_offset; + Rela.r_info = Rel.r_info; + Rela.r_addend = 0; + printDynamicRelocation(Obj, Rela, false); + } + } +} + +// Hash histogram shows statistics of how efficient the hash was for the +// dynamic symbol table. The table shows number of hash buckets for different +// lengths of chains as absolute number and percentage of the total buckets. +// Additionally cumulative coverage of symbols for each set of buckets. +template <class ELFT> +void GNUStyle<ELFT>::printHashHistogram(const ELFFile<ELFT> *Obj) { + + const Elf_Hash *HashTable = this->dumper()->getHashTable(); + const Elf_GnuHash *GnuHashTable = this->dumper()->getGnuHashTable(); + + // Print histogram for .hash section + if (HashTable) { + size_t NBucket = HashTable->nbucket; + size_t NChain = HashTable->nchain; + ArrayRef<Elf_Word> Buckets = HashTable->buckets(); + ArrayRef<Elf_Word> Chains = HashTable->chains(); + size_t TotalSyms = 0; + // If hash table is correct, we have at least chains with 0 length + size_t MaxChain = 1; + size_t CumulativeNonZero = 0; + + if (NChain == 0 || NBucket == 0) + return; + + std::vector<size_t> ChainLen(NBucket, 0); + // Go over all buckets and and note chain lengths of each bucket (total + // unique chain lengths). + for (size_t B = 0; B < NBucket; B++) { + for (size_t C = Buckets[B]; C > 0 && C < NChain; C = Chains[C]) + if (MaxChain <= ++ChainLen[B]) + MaxChain++; + TotalSyms += ChainLen[B]; + } + + if (!TotalSyms) + return; + + std::vector<size_t> Count(MaxChain, 0) ; + // Count how long is the chain for each bucket + for (size_t B = 0; B < NBucket; B++) + ++Count[ChainLen[B]]; + // Print Number of buckets with each chain lengths and their cumulative + // coverage of the symbols + OS << "Histogram for bucket list length (total of " << NBucket + << " buckets)\n" + << " Length Number % of total Coverage\n"; + for (size_t I = 0; I < MaxChain; I++) { + CumulativeNonZero += Count[I] * I; + OS << format("%7lu %-10lu (%5.1f%%) %5.1f%%\n", I, Count[I], + (Count[I] * 100.0) / NBucket, + (CumulativeNonZero * 100.0) / TotalSyms); + } + } + + // Print histogram for .gnu.hash section + if (GnuHashTable) { + size_t NBucket = GnuHashTable->nbuckets; + ArrayRef<Elf_Word> Buckets = GnuHashTable->buckets(); + unsigned NumSyms = this->dumper()->dynamic_symbols().size(); + if (!NumSyms) + return; + ArrayRef<Elf_Word> Chains = GnuHashTable->values(NumSyms); + size_t Symndx = GnuHashTable->symndx; + size_t TotalSyms = 0; + size_t MaxChain = 1; + size_t CumulativeNonZero = 0; + + if (Chains.size() == 0 || NBucket == 0) + return; + + std::vector<size_t> ChainLen(NBucket, 0); + + for (size_t B = 0; B < NBucket; B++) { + if (!Buckets[B]) + continue; + size_t Len = 1; + for (size_t C = Buckets[B] - Symndx; + C < Chains.size() && (Chains[C] & 1) == 0; C++) + if (MaxChain < ++Len) + MaxChain++; + ChainLen[B] = Len; + TotalSyms += Len; + } + MaxChain++; + + if (!TotalSyms) + return; + + std::vector<size_t> Count(MaxChain, 0) ; + for (size_t B = 0; B < NBucket; B++) + ++Count[ChainLen[B]]; + // Print Number of buckets with each chain lengths and their cumulative + // coverage of the symbols + OS << "Histogram for `.gnu.hash' bucket list length (total of " << NBucket + << " buckets)\n" + << " Length Number % of total Coverage\n"; + for (size_t I = 0; I <MaxChain; I++) { + CumulativeNonZero += Count[I] * I; + OS << format("%7lu %-10lu (%5.1f%%) %5.1f%%\n", I, Count[I], + (Count[I] * 100.0) / NBucket, + (CumulativeNonZero * 100.0) / TotalSyms); + } + } +} + +template <class ELFT> void LLVMStyle<ELFT>::printFileHeaders(const ELFO *Obj) { + const Elf_Ehdr *e = Obj->getHeader(); + { + DictScope D(W, "ElfHeader"); + { + DictScope D(W, "Ident"); + W.printBinary("Magic", makeArrayRef(e->e_ident).slice(ELF::EI_MAG0, 4)); + W.printEnum("Class", e->e_ident[ELF::EI_CLASS], makeArrayRef(ElfClass)); + W.printEnum("DataEncoding", e->e_ident[ELF::EI_DATA], + makeArrayRef(ElfDataEncoding)); + W.printNumber("FileVersion", e->e_ident[ELF::EI_VERSION]); + + // Handle architecture specific OS/ABI values. + if (e->e_machine == ELF::EM_AMDGPU && + e->e_ident[ELF::EI_OSABI] == ELF::ELFOSABI_AMDGPU_HSA) + W.printHex("OS/ABI", "AMDGPU_HSA", ELF::ELFOSABI_AMDGPU_HSA); + else + W.printEnum("OS/ABI", e->e_ident[ELF::EI_OSABI], + makeArrayRef(ElfOSABI)); + W.printNumber("ABIVersion", e->e_ident[ELF::EI_ABIVERSION]); + W.printBinary("Unused", makeArrayRef(e->e_ident).slice(ELF::EI_PAD)); + } + + W.printEnum("Type", e->e_type, makeArrayRef(ElfObjectFileType)); + W.printEnum("Machine", e->e_machine, makeArrayRef(ElfMachineType)); + W.printNumber("Version", e->e_version); + W.printHex("Entry", e->e_entry); + W.printHex("ProgramHeaderOffset", e->e_phoff); + W.printHex("SectionHeaderOffset", e->e_shoff); + if (e->e_machine == EM_MIPS) + W.printFlags("Flags", e->e_flags, makeArrayRef(ElfHeaderMipsFlags), + unsigned(ELF::EF_MIPS_ARCH), unsigned(ELF::EF_MIPS_ABI), + unsigned(ELF::EF_MIPS_MACH)); + else + W.printFlags("Flags", e->e_flags); + W.printNumber("HeaderSize", e->e_ehsize); + W.printNumber("ProgramHeaderEntrySize", e->e_phentsize); + W.printNumber("ProgramHeaderCount", e->e_phnum); + W.printNumber("SectionHeaderEntrySize", e->e_shentsize); + W.printNumber("SectionHeaderCount", e->e_shnum); + W.printNumber("StringTableSectionIndex", e->e_shstrndx); + } +} + +template <class ELFT> +void LLVMStyle<ELFT>::printGroupSections(const ELFO *Obj) { + DictScope Lists(W, "Groups"); + uint32_t SectionIndex = 0; + bool HasGroups = false; + for (const Elf_Shdr &Sec : Obj->sections()) { + if (Sec.sh_type == ELF::SHT_GROUP) { + HasGroups = true; + const Elf_Shdr *Symtab = unwrapOrError(Obj->getSection(Sec.sh_link)); + StringRef StrTable = unwrapOrError(Obj->getStringTableForSymtab(*Symtab)); + const Elf_Sym *Sym = Obj->template getEntry<Elf_Sym>(Symtab, Sec.sh_info); + auto Data = unwrapOrError( + Obj->template getSectionContentsAsArray<Elf_Word>(&Sec)); + DictScope D(W, "Group"); + StringRef Name = unwrapOrError(Obj->getSectionName(&Sec)); + W.printNumber("Name", Name, Sec.sh_name); + W.printNumber("Index", SectionIndex); + W.printHex("Type", getGroupType(Data[0]), Data[0]); + W.startLine() << "Signature: " << StrTable.data() + Sym->st_name << "\n"; + { + ListScope L(W, "Section(s) in group"); + size_t Member = 1; + while (Member < Data.size()) { + auto Sec = unwrapOrError(Obj->getSection(Data[Member])); + const StringRef Name = unwrapOrError(Obj->getSectionName(Sec)); + W.startLine() << Name << " (" << Data[Member++] << ")\n"; + } + } + } + ++SectionIndex; + } + if (!HasGroups) + W.startLine() << "There are no group sections in the file.\n"; +} + +template <class ELFT> void LLVMStyle<ELFT>::printRelocations(const ELFO *Obj) { + ListScope D(W, "Relocations"); + + int SectionNumber = -1; + for (const Elf_Shdr &Sec : Obj->sections()) { + ++SectionNumber; + + if (Sec.sh_type != ELF::SHT_REL && Sec.sh_type != ELF::SHT_RELA) + continue; + + StringRef Name = unwrapOrError(Obj->getSectionName(&Sec)); + + W.startLine() << "Section (" << SectionNumber << ") " << Name << " {\n"; + W.indent(); + + printRelocations(&Sec, Obj); + + W.unindent(); + W.startLine() << "}\n"; + } +} + +template <class ELFT> +void LLVMStyle<ELFT>::printRelocations(const Elf_Shdr *Sec, const ELFO *Obj) { + const Elf_Shdr *SymTab = unwrapOrError(Obj->getSection(Sec->sh_link)); + + switch (Sec->sh_type) { + case ELF::SHT_REL: + for (const Elf_Rel &R : Obj->rels(Sec)) { + Elf_Rela Rela; + Rela.r_offset = R.r_offset; + Rela.r_info = R.r_info; + Rela.r_addend = 0; + printRelocation(Obj, Rela, SymTab); + } + break; + case ELF::SHT_RELA: + for (const Elf_Rela &R : Obj->relas(Sec)) + printRelocation(Obj, R, SymTab); + break; + } +} + +template <class ELFT> +void LLVMStyle<ELFT>::printRelocation(const ELFO *Obj, Elf_Rela Rel, + const Elf_Shdr *SymTab) { + SmallString<32> RelocName; + Obj->getRelocationTypeName(Rel.getType(Obj->isMips64EL()), RelocName); + StringRef TargetName; + const Elf_Sym *Sym = Obj->getRelocationSymbol(&Rel, SymTab); + if (Sym && Sym->getType() == ELF::STT_SECTION) { + const Elf_Shdr *Sec = unwrapOrError( + Obj->getSection(Sym, SymTab, this->dumper()->getShndxTable())); + TargetName = unwrapOrError(Obj->getSectionName(Sec)); + } else if (Sym) { + StringRef StrTable = unwrapOrError(Obj->getStringTableForSymtab(*SymTab)); + TargetName = unwrapOrError(Sym->getName(StrTable)); + } + + if (opts::ExpandRelocs) { + DictScope Group(W, "Relocation"); + W.printHex("Offset", Rel.r_offset); + W.printNumber("Type", RelocName, (int)Rel.getType(Obj->isMips64EL())); + W.printNumber("Symbol", TargetName.size() > 0 ? TargetName : "-", + Rel.getSymbol(Obj->isMips64EL())); + W.printHex("Addend", Rel.r_addend); + } else { + raw_ostream &OS = W.startLine(); + OS << W.hex(Rel.r_offset) << " " << RelocName << " " + << (TargetName.size() > 0 ? TargetName : "-") << " " + << W.hex(Rel.r_addend) << "\n"; + } +} + +template <class ELFT> void LLVMStyle<ELFT>::printSections(const ELFO *Obj) { + ListScope SectionsD(W, "Sections"); + + int SectionIndex = -1; + for (const Elf_Shdr &Sec : Obj->sections()) { + ++SectionIndex; + + StringRef Name = unwrapOrError(Obj->getSectionName(&Sec)); + + DictScope SectionD(W, "Section"); + W.printNumber("Index", SectionIndex); + W.printNumber("Name", Name, Sec.sh_name); + W.printHex("Type", + getElfSectionType(Obj->getHeader()->e_machine, Sec.sh_type), + Sec.sh_type); + std::vector<EnumEntry<unsigned>> SectionFlags(std::begin(ElfSectionFlags), + std::end(ElfSectionFlags)); + switch (Obj->getHeader()->e_machine) { + case EM_AMDGPU: + SectionFlags.insert(SectionFlags.end(), std::begin(ElfAMDGPUSectionFlags), + std::end(ElfAMDGPUSectionFlags)); + break; + case EM_HEXAGON: + SectionFlags.insert(SectionFlags.end(), + std::begin(ElfHexagonSectionFlags), + std::end(ElfHexagonSectionFlags)); + break; + case EM_MIPS: + SectionFlags.insert(SectionFlags.end(), std::begin(ElfMipsSectionFlags), + std::end(ElfMipsSectionFlags)); + break; + case EM_X86_64: + SectionFlags.insert(SectionFlags.end(), std::begin(ElfX86_64SectionFlags), + std::end(ElfX86_64SectionFlags)); + break; + case EM_XCORE: + SectionFlags.insert(SectionFlags.end(), std::begin(ElfXCoreSectionFlags), + std::end(ElfXCoreSectionFlags)); + break; + default: + // Nothing to do. + break; + } + W.printFlags("Flags", Sec.sh_flags, makeArrayRef(SectionFlags)); + W.printHex("Address", Sec.sh_addr); + W.printHex("Offset", Sec.sh_offset); + W.printNumber("Size", Sec.sh_size); + W.printNumber("Link", Sec.sh_link); + W.printNumber("Info", Sec.sh_info); + W.printNumber("AddressAlignment", Sec.sh_addralign); + W.printNumber("EntrySize", Sec.sh_entsize); + + if (opts::SectionRelocations) { + ListScope D(W, "Relocations"); + printRelocations(&Sec, Obj); + } + + if (opts::SectionSymbols) { + ListScope D(W, "Symbols"); + const Elf_Shdr *Symtab = this->dumper()->getDotSymtabSec(); + StringRef StrTable = unwrapOrError(Obj->getStringTableForSymtab(*Symtab)); + + for (const Elf_Sym &Sym : Obj->symbols(Symtab)) { + const Elf_Shdr *SymSec = unwrapOrError( + Obj->getSection(&Sym, Symtab, this->dumper()->getShndxTable())); + if (SymSec == &Sec) + printSymbol(Obj, &Sym, Obj->symbol_begin(Symtab), StrTable, false); + } + } + + if (opts::SectionData && Sec.sh_type != ELF::SHT_NOBITS) { + ArrayRef<uint8_t> Data = unwrapOrError(Obj->getSectionContents(&Sec)); + W.printBinaryBlock("SectionData", + StringRef((const char *)Data.data(), Data.size())); + } + } +} + +template <class ELFT> +void LLVMStyle<ELFT>::printSymbol(const ELFO *Obj, const Elf_Sym *Symbol, + const Elf_Sym *First, StringRef StrTable, + bool IsDynamic) { + unsigned SectionIndex = 0; + StringRef SectionName; + getSectionNameIndex(*Obj, Symbol, First, this->dumper()->getShndxTable(), + SectionName, SectionIndex); + std::string FullSymbolName = + this->dumper()->getFullSymbolName(Symbol, StrTable, IsDynamic); + unsigned char SymbolType = Symbol->getType(); + + DictScope D(W, "Symbol"); + W.printNumber("Name", FullSymbolName, Symbol->st_name); + W.printHex("Value", Symbol->st_value); + W.printNumber("Size", Symbol->st_size); + W.printEnum("Binding", Symbol->getBinding(), makeArrayRef(ElfSymbolBindings)); + if (Obj->getHeader()->e_machine == ELF::EM_AMDGPU && + SymbolType >= ELF::STT_LOOS && SymbolType < ELF::STT_HIOS) + W.printEnum("Type", SymbolType, makeArrayRef(AMDGPUSymbolTypes)); + else + W.printEnum("Type", SymbolType, makeArrayRef(ElfSymbolTypes)); + if (Symbol->st_other == 0) + // Usually st_other flag is zero. Do not pollute the output + // by flags enumeration in that case. + W.printNumber("Other", 0); + else { + std::vector<EnumEntry<unsigned>> SymOtherFlags(std::begin(ElfSymOtherFlags), + std::end(ElfSymOtherFlags)); + if (Obj->getHeader()->e_machine == EM_MIPS) { + // Someones in their infinite wisdom decided to make STO_MIPS_MIPS16 + // flag overlapped with other ST_MIPS_xxx flags. So consider both + // cases separately. + if ((Symbol->st_other & STO_MIPS_MIPS16) == STO_MIPS_MIPS16) + SymOtherFlags.insert(SymOtherFlags.end(), + std::begin(ElfMips16SymOtherFlags), + std::end(ElfMips16SymOtherFlags)); + else + SymOtherFlags.insert(SymOtherFlags.end(), + std::begin(ElfMipsSymOtherFlags), + std::end(ElfMipsSymOtherFlags)); + } + W.printFlags("Other", Symbol->st_other, makeArrayRef(SymOtherFlags), 0x3u); + } + W.printHex("Section", SectionName, SectionIndex); +} + +template <class ELFT> void LLVMStyle<ELFT>::printSymbols(const ELFO *Obj) { + ListScope Group(W, "Symbols"); + this->dumper()->printSymbolsHelper(false); +} + +template <class ELFT> +void LLVMStyle<ELFT>::printDynamicSymbols(const ELFO *Obj) { + ListScope Group(W, "DynamicSymbols"); + this->dumper()->printSymbolsHelper(true); +} + +template <class ELFT> +void LLVMStyle<ELFT>::printDynamicRelocations(const ELFO *Obj) { + const DynRegionInfo &DynRelRegion = this->dumper()->getDynRelRegion(); + const DynRegionInfo &DynRelaRegion = this->dumper()->getDynRelaRegion(); + const DynRegionInfo &DynPLTRelRegion = this->dumper()->getDynPLTRelRegion(); + if (DynRelRegion.Size && DynRelaRegion.Size) + report_fatal_error("There are both REL and RELA dynamic relocations"); + W.startLine() << "Dynamic Relocations {\n"; + W.indent(); + if (DynRelaRegion.Size > 0) + for (const Elf_Rela &Rela : this->dumper()->dyn_relas()) + printDynamicRelocation(Obj, Rela); + else + for (const Elf_Rel &Rel : this->dumper()->dyn_rels()) { + Elf_Rela Rela; + Rela.r_offset = Rel.r_offset; + Rela.r_info = Rel.r_info; + Rela.r_addend = 0; + printDynamicRelocation(Obj, Rela); + } + if (DynPLTRelRegion.EntSize == sizeof(Elf_Rela)) + for (const Elf_Rela &Rela : DynPLTRelRegion.getAsArrayRef<Elf_Rela>()) + printDynamicRelocation(Obj, Rela); + else + for (const Elf_Rel &Rel : DynPLTRelRegion.getAsArrayRef<Elf_Rel>()) { + Elf_Rela Rela; + Rela.r_offset = Rel.r_offset; + Rela.r_info = Rel.r_info; + Rela.r_addend = 0; + printDynamicRelocation(Obj, Rela); + } + W.unindent(); + W.startLine() << "}\n"; +} + +template <class ELFT> +void LLVMStyle<ELFT>::printDynamicRelocation(const ELFO *Obj, Elf_Rela Rel) { + SmallString<32> RelocName; + Obj->getRelocationTypeName(Rel.getType(Obj->isMips64EL()), RelocName); + StringRef SymbolName; + uint32_t SymIndex = Rel.getSymbol(Obj->isMips64EL()); + const Elf_Sym *Sym = this->dumper()->dynamic_symbols().begin() + SymIndex; + SymbolName = + unwrapOrError(Sym->getName(this->dumper()->getDynamicStringTable())); + if (opts::ExpandRelocs) { + DictScope Group(W, "Relocation"); + W.printHex("Offset", Rel.r_offset); + W.printNumber("Type", RelocName, (int)Rel.getType(Obj->isMips64EL())); + W.printString("Symbol", SymbolName.size() > 0 ? SymbolName : "-"); + W.printHex("Addend", Rel.r_addend); + } else { + raw_ostream &OS = W.startLine(); + OS << W.hex(Rel.r_offset) << " " << RelocName << " " + << (SymbolName.size() > 0 ? SymbolName : "-") << " " + << W.hex(Rel.r_addend) << "\n"; + } +} + +template <class ELFT> +void LLVMStyle<ELFT>::printProgramHeaders(const ELFO *Obj) { + ListScope L(W, "ProgramHeaders"); - prettyPrintStackMap( - llvm::outs(), - StackMapV1Parser<ELFT::TargetEndianness>(*StackMapContentsArray)); + for (const Elf_Phdr &Phdr : Obj->program_headers()) { + DictScope P(W, "ProgramHeader"); + W.printHex("Type", + getElfSegmentType(Obj->getHeader()->e_machine, Phdr.p_type), + Phdr.p_type); + W.printHex("Offset", Phdr.p_offset); + W.printHex("VirtualAddress", Phdr.p_vaddr); + W.printHex("PhysicalAddress", Phdr.p_paddr); + W.printNumber("FileSize", Phdr.p_filesz); + W.printNumber("MemSize", Phdr.p_memsz); + W.printFlags("Flags", Phdr.p_flags, makeArrayRef(ElfSegmentFlags)); + W.printNumber("Alignment", Phdr.p_align); + } +} +template <class ELFT> +void LLVMStyle<ELFT>::printHashHistogram(const ELFFile<ELFT> *Obj) { + W.startLine() << "Hash Histogram not implemented!\n"; } diff --git a/tools/llvm-readobj/Error.cpp b/tools/llvm-readobj/Error.cpp index 7e6f780c5d1a2..492eb33094976 100644 --- a/tools/llvm-readobj/Error.cpp +++ b/tools/llvm-readobj/Error.cpp @@ -17,6 +17,9 @@ using namespace llvm; namespace { +// FIXME: This class is only here to support the transition to llvm::Error. It +// will be removed once this transition is complete. Clients should prefer to +// deal with the Error value directly, rather than converting to error_code. class _readobj_error_category : public std::error_category { public: const char* name() const LLVM_NOEXCEPT override; diff --git a/tools/llvm-readobj/MachODumper.cpp b/tools/llvm-readobj/MachODumper.cpp index 58d2c9fca47dd..3773df250578e 100644 --- a/tools/llvm-readobj/MachODumper.cpp +++ b/tools/llvm-readobj/MachODumper.cpp @@ -11,15 +11,15 @@ // //===----------------------------------------------------------------------===// -#include "llvm-readobj.h" #include "Error.h" #include "ObjDumper.h" #include "StackMapPrinter.h" -#include "StreamWriter.h" +#include "llvm-readobj.h" #include "llvm/ADT/SmallString.h" #include "llvm/ADT/StringExtras.h" #include "llvm/Object/MachO.h" #include "llvm/Support/Casting.h" +#include "llvm/Support/ScopedPrinter.h" using namespace llvm; using namespace object; @@ -28,9 +28,8 @@ namespace { class MachODumper : public ObjDumper { public: - MachODumper(const MachOObjectFile *Obj, StreamWriter& Writer) - : ObjDumper(Writer) - , Obj(Obj) { } + MachODumper(const MachOObjectFile *Obj, ScopedPrinter &Writer) + : ObjDumper(Writer), Obj(Obj) {} void printFileHeaders() override; void printSections() override; @@ -69,7 +68,7 @@ private: namespace llvm { std::error_code createMachODumper(const object::ObjectFile *Obj, - StreamWriter &Writer, + ScopedPrinter &Writer, std::unique_ptr<ObjDumper> &Result) { const MachOObjectFile *MachOObj = dyn_cast<MachOObjectFile>(Obj); if (!MachOObj) @@ -239,7 +238,8 @@ static const EnumEntry<unsigned> MachOSymbolFlags[] = { { "ReferencedDynamically", 0x10 }, { "NoDeadStrip", 0x20 }, { "WeakRef", 0x40 }, - { "WeakDef", 0x80 } + { "WeakDef", 0x80 }, + { "AltEntry", 0x200 }, }; static const EnumEntry<unsigned> MachOSymbolTypes[] = { @@ -540,8 +540,9 @@ void MachODumper::printRelocation(const MachOObjectFile *Obj, if (IsExtern) { symbol_iterator Symbol = Reloc.getSymbol(); if (Symbol != Obj->symbol_end()) { - ErrorOr<StringRef> TargetNameOrErr = Symbol->getName(); - error(TargetNameOrErr.getError()); + Expected<StringRef> TargetNameOrErr = Symbol->getName(); + if (!TargetNameOrErr) + error(errorToErrorCode(TargetNameOrErr.takeError())); TargetName = *TargetNameOrErr; } } else if (!IsScattered) { @@ -604,15 +605,19 @@ void MachODumper::printDynamicSymbols() { void MachODumper::printSymbol(const SymbolRef &Symbol) { StringRef SymbolName; - if (ErrorOr<StringRef> SymbolNameOrErr = Symbol.getName()) + Expected<StringRef> SymbolNameOrErr = Symbol.getName(); + if (!SymbolNameOrErr) { + // TODO: Actually report errors helpfully. + consumeError(SymbolNameOrErr.takeError()); + } else SymbolName = *SymbolNameOrErr; MachOSymbol MOSymbol; getSymbol(Obj, Symbol.getRawDataRefImpl(), MOSymbol); StringRef SectionName = ""; - ErrorOr<section_iterator> SecIOrErr = Symbol.getSection(); - error(SecIOrErr.getError()); + Expected<section_iterator> SecIOrErr = Symbol.getSection(); + error(errorToErrorCode(SecIOrErr.takeError())); section_iterator SecI = *SecIOrErr; if (SecI != Obj->section_end()) error(SecI->getName(SectionName)); diff --git a/tools/llvm-readobj/Makefile b/tools/llvm-readobj/Makefile deleted file mode 100644 index 958bd0c5b3121..0000000000000 --- a/tools/llvm-readobj/Makefile +++ /dev/null @@ -1,18 +0,0 @@ -##===- tools/llvm-readobj/Makefile -----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-readobj -LINK_COMPONENTS := bitreader object all-targets - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common - diff --git a/tools/llvm-readobj/ObjDumper.cpp b/tools/llvm-readobj/ObjDumper.cpp index f500a05815532..2a0a90e5cfd5d 100644 --- a/tools/llvm-readobj/ObjDumper.cpp +++ b/tools/llvm-readobj/ObjDumper.cpp @@ -14,16 +14,13 @@ #include "ObjDumper.h" #include "Error.h" -#include "StreamWriter.h" -#include "llvm/ADT/StringRef.h" #include "llvm/Object/ObjectFile.h" +#include "llvm/Support/ScopedPrinter.h" #include "llvm/Support/raw_ostream.h" namespace llvm { -ObjDumper::ObjDumper(StreamWriter& Writer) - : W(Writer) { -} +ObjDumper::ObjDumper(ScopedPrinter &Writer) : W(Writer) {} ObjDumper::~ObjDumper() { } diff --git a/tools/llvm-readobj/ObjDumper.h b/tools/llvm-readobj/ObjDumper.h index db26d6983552d..a39fc260b5493 100644 --- a/tools/llvm-readobj/ObjDumper.h +++ b/tools/llvm-readobj/ObjDumper.h @@ -18,12 +18,15 @@ namespace object { class COFFImportFile; class ObjectFile; } +namespace codeview { +class MemoryTypeTableBuilder; +} -class StreamWriter; +class ScopedPrinter; class ObjDumper { public: - ObjDumper(StreamWriter& Writer); + ObjDumper(ScopedPrinter &Writer); virtual ~ObjDumper(); virtual void printFileHeaders() = 0; @@ -42,6 +45,8 @@ public: virtual void printGnuHashTable() { } virtual void printLoadName() {} virtual void printVersionInfo() {} + virtual void printGroupSections() {} + virtual void printHashHistogram() {} // Only implemented for ARM ELF at this time. virtual void printAttributes() { } @@ -50,13 +55,17 @@ public: virtual void printMipsPLTGOT() { } virtual void printMipsABIFlags() { } virtual void printMipsReginfo() { } + virtual void printMipsOptions() { } // Only implemented for PE/COFF. virtual void printCOFFImports() { } virtual void printCOFFExports() { } virtual void printCOFFDirectives() { } virtual void printCOFFBaseReloc() { } + virtual void printCOFFDebugDirectory() { } virtual void printCodeViewDebugInfo() { } + virtual void + mergeCodeViewTypes(llvm::codeview::MemoryTypeTableBuilder &CVTypes) {} // Only implemented for MachO. virtual void printMachODataInCode() { } @@ -69,23 +78,26 @@ public: virtual void printStackMap() const = 0; protected: - StreamWriter& W; + ScopedPrinter &W; }; std::error_code createCOFFDumper(const object::ObjectFile *Obj, - StreamWriter &Writer, + ScopedPrinter &Writer, std::unique_ptr<ObjDumper> &Result); std::error_code createELFDumper(const object::ObjectFile *Obj, - StreamWriter &Writer, + ScopedPrinter &Writer, std::unique_ptr<ObjDumper> &Result); std::error_code createMachODumper(const object::ObjectFile *Obj, - StreamWriter &Writer, + ScopedPrinter &Writer, std::unique_ptr<ObjDumper> &Result); void dumpCOFFImportFile(const object::COFFImportFile *File); +void dumpCodeViewMergedTypes(ScopedPrinter &Writer, + llvm::codeview::MemoryTypeTableBuilder &CVTypes); + } // namespace llvm #endif diff --git a/tools/llvm-readobj/StreamWriter.cpp b/tools/llvm-readobj/StreamWriter.cpp deleted file mode 100644 index 871811233a65f..0000000000000 --- a/tools/llvm-readobj/StreamWriter.cpp +++ /dev/null @@ -1,79 +0,0 @@ -#include "StreamWriter.h" -#include "llvm/ADT/StringExtras.h" -#include "llvm/Support/Format.h" -#include <cctype> - -using namespace llvm::support; - -namespace llvm { - -raw_ostream &operator<<(raw_ostream &OS, const HexNumber& Value) { - uint64_t N = Value.Value; - // Zero is a special case. - if (N == 0) - return OS << "0x0"; - - char NumberBuffer[20]; - char *EndPtr = NumberBuffer + sizeof(NumberBuffer); - char *CurPtr = EndPtr; - - while (N) { - uintptr_t X = N % 16; - *--CurPtr = (X < 10 ? '0' + X : 'A' + X - 10); - N /= 16; - } - - OS << "0x"; - return OS.write(CurPtr, EndPtr - CurPtr); -} - -void StreamWriter::printBinaryImpl(StringRef Label, StringRef Str, - ArrayRef<uint8_t> Data, bool Block) { - if (Data.size() > 16) - Block = true; - - if (Block) { - startLine() << Label; - if (Str.size() > 0) - OS << ": " << Str; - OS << " (\n"; - for (size_t addr = 0, end = Data.size(); addr < end; addr += 16) { - startLine() << format(" %04" PRIX64 ": ", uint64_t(addr)); - // Dump line of hex. - for (size_t i = 0; i < 16; ++i) { - if (i != 0 && i % 4 == 0) - OS << ' '; - if (addr + i < end) - OS << hexdigit((Data[addr + i] >> 4) & 0xF, false) - << hexdigit(Data[addr + i] & 0xF, false); - else - OS << " "; - } - // Print ascii. - OS << " |"; - for (std::size_t i = 0; i < 16 && addr + i < end; ++i) { - if (std::isprint(Data[addr + i] & 0xFF)) - OS << Data[addr + i]; - else - OS << "."; - } - OS << "|\n"; - } - - startLine() << ")\n"; - } else { - startLine() << Label << ":"; - if (Str.size() > 0) - OS << " " << Str; - OS << " ("; - for (size_t i = 0; i < Data.size(); ++i) { - if (i > 0) - OS << " "; - - OS << format("%02X", static_cast<int>(Data[i])); - } - OS << ")\n"; - } -} - -} // namespace llvm diff --git a/tools/llvm-readobj/StreamWriter.h b/tools/llvm-readobj/StreamWriter.h deleted file mode 100644 index d2dbb07af2566..0000000000000 --- a/tools/llvm-readobj/StreamWriter.h +++ /dev/null @@ -1,320 +0,0 @@ -//===-- StreamWriter.h ----------------------------------------------------===// -// -// The LLVM Compiler Infrastructure -// -// This file is distributed under the University of Illinois Open Source -// License. See LICENSE.TXT for details. -// -//===----------------------------------------------------------------------===// - -#ifndef LLVM_TOOLS_LLVM_READOBJ_STREAMWRITER_H -#define LLVM_TOOLS_LLVM_READOBJ_STREAMWRITER_H - -#include "llvm/ADT/ArrayRef.h" -#include "llvm/ADT/SmallVector.h" -#include "llvm/ADT/StringRef.h" -#include "llvm/Support/DataTypes.h" -#include "llvm/Support/Endian.h" -#include "llvm/Support/raw_ostream.h" -#include <algorithm> - -using namespace llvm; -using namespace llvm::support; - -namespace llvm { - -template<typename T> -struct EnumEntry { - StringRef Name; - T Value; -}; - -struct HexNumber { - // To avoid sign-extension we have to explicitly cast to the appropriate - // unsigned type. The overloads are here so that every type that is implicitly - // convertible to an integer (including enums and endian helpers) can be used - // without requiring type traits or call-site changes. - HexNumber(char Value) : Value(static_cast<unsigned char>(Value)) { } - HexNumber(signed char Value) : Value(static_cast<unsigned char>(Value)) { } - HexNumber(signed short Value) : Value(static_cast<unsigned short>(Value)) { } - HexNumber(signed int Value) : Value(static_cast<unsigned int>(Value)) { } - HexNumber(signed long Value) : Value(static_cast<unsigned long>(Value)) { } - HexNumber(signed long long Value) : Value(static_cast<unsigned long long>(Value)) { } - HexNumber(unsigned char Value) : Value(Value) { } - HexNumber(unsigned short Value) : Value(Value) { } - HexNumber(unsigned int Value) : Value(Value) { } - HexNumber(unsigned long Value) : Value(Value) { } - HexNumber(unsigned long long Value) : Value(Value) { } - uint64_t Value; -}; - -raw_ostream &operator<<(raw_ostream &OS, const HexNumber& Value); - -class StreamWriter { -public: - StreamWriter(raw_ostream &OS) - : OS(OS) - , IndentLevel(0) { - } - - void flush() { - OS.flush(); - } - - void indent(int Levels = 1) { - IndentLevel += Levels; - } - - void unindent(int Levels = 1) { - IndentLevel = std::max(0, IndentLevel - Levels); - } - - void printIndent() { - for (int i = 0; i < IndentLevel; ++i) - OS << " "; - } - - template<typename T> - HexNumber hex(T Value) { - return HexNumber(Value); - } - - template<typename T, typename TEnum> - void printEnum(StringRef Label, T Value, - ArrayRef<EnumEntry<TEnum> > EnumValues) { - StringRef Name; - bool Found = false; - for (const auto &EnumItem : EnumValues) { - if (EnumItem.Value == Value) { - Name = EnumItem.Name; - Found = true; - break; - } - } - - if (Found) { - startLine() << Label << ": " << Name << " (" << hex(Value) << ")\n"; - } else { - startLine() << Label << ": " << hex(Value) << "\n"; - } - } - - template <typename T, typename TFlag> - void printFlags(StringRef Label, T Value, ArrayRef<EnumEntry<TFlag>> Flags, - TFlag EnumMask1 = {}, TFlag EnumMask2 = {}, - TFlag EnumMask3 = {}) { - typedef EnumEntry<TFlag> FlagEntry; - typedef SmallVector<FlagEntry, 10> FlagVector; - FlagVector SetFlags; - - for (const auto &Flag : Flags) { - if (Flag.Value == 0) - continue; - - TFlag EnumMask{}; - if (Flag.Value & EnumMask1) - EnumMask = EnumMask1; - else if (Flag.Value & EnumMask2) - EnumMask = EnumMask2; - else if (Flag.Value & EnumMask3) - EnumMask = EnumMask3; - bool IsEnum = (Flag.Value & EnumMask) != 0; - if ((!IsEnum && (Value & Flag.Value) == Flag.Value) || - (IsEnum && (Value & EnumMask) == Flag.Value)) { - SetFlags.push_back(Flag); - } - } - - std::sort(SetFlags.begin(), SetFlags.end(), &flagName<TFlag>); - - startLine() << Label << " [ (" << hex(Value) << ")\n"; - for (const auto &Flag : SetFlags) { - startLine() << " " << Flag.Name << " (" << hex(Flag.Value) << ")\n"; - } - startLine() << "]\n"; - } - - template<typename T> - void printFlags(StringRef Label, T Value) { - startLine() << Label << " [ (" << hex(Value) << ")\n"; - uint64_t Flag = 1; - uint64_t Curr = Value; - while (Curr > 0) { - if (Curr & 1) - startLine() << " " << hex(Flag) << "\n"; - Curr >>= 1; - Flag <<= 1; - } - startLine() << "]\n"; - } - - void printNumber(StringRef Label, uint64_t Value) { - startLine() << Label << ": " << Value << "\n"; - } - - void printNumber(StringRef Label, uint32_t Value) { - startLine() << Label << ": " << Value << "\n"; - } - - void printNumber(StringRef Label, uint16_t Value) { - startLine() << Label << ": " << Value << "\n"; - } - - void printNumber(StringRef Label, uint8_t Value) { - startLine() << Label << ": " << unsigned(Value) << "\n"; - } - - void printNumber(StringRef Label, int64_t Value) { - startLine() << Label << ": " << Value << "\n"; - } - - void printNumber(StringRef Label, int32_t Value) { - startLine() << Label << ": " << Value << "\n"; - } - - void printNumber(StringRef Label, int16_t Value) { - startLine() << Label << ": " << Value << "\n"; - } - - void printNumber(StringRef Label, int8_t Value) { - startLine() << Label << ": " << int(Value) << "\n"; - } - - void printBoolean(StringRef Label, bool Value) { - startLine() << Label << ": " << (Value ? "Yes" : "No") << '\n'; - } - - template <typename T> - void printList(StringRef Label, const T &List) { - startLine() << Label << ": ["; - bool Comma = false; - for (const auto &Item : List) { - if (Comma) - OS << ", "; - OS << Item; - Comma = true; - } - OS << "]\n"; - } - - template <typename T> - void printHexList(StringRef Label, const T &List) { - startLine() << Label << ": ["; - bool Comma = false; - for (const auto &Item : List) { - if (Comma) - OS << ", "; - OS << hex(Item); - Comma = true; - } - OS << "]\n"; - } - - template<typename T> - void printHex(StringRef Label, T Value) { - startLine() << Label << ": " << hex(Value) << "\n"; - } - - template<typename T> - void printHex(StringRef Label, StringRef Str, T Value) { - startLine() << Label << ": " << Str << " (" << hex(Value) << ")\n"; - } - - void printString(StringRef Label, StringRef Value) { - startLine() << Label << ": " << Value << "\n"; - } - - void printString(StringRef Label, const std::string &Value) { - startLine() << Label << ": " << Value << "\n"; - } - - template<typename T> - void printNumber(StringRef Label, StringRef Str, T Value) { - startLine() << Label << ": " << Str << " (" << Value << ")\n"; - } - - void printBinary(StringRef Label, StringRef Str, ArrayRef<uint8_t> Value) { - printBinaryImpl(Label, Str, Value, false); - } - - void printBinary(StringRef Label, StringRef Str, ArrayRef<char> Value) { - auto V = makeArrayRef(reinterpret_cast<const uint8_t*>(Value.data()), - Value.size()); - printBinaryImpl(Label, Str, V, false); - } - - void printBinary(StringRef Label, ArrayRef<uint8_t> Value) { - printBinaryImpl(Label, StringRef(), Value, false); - } - - void printBinary(StringRef Label, ArrayRef<char> Value) { - auto V = makeArrayRef(reinterpret_cast<const uint8_t*>(Value.data()), - Value.size()); - printBinaryImpl(Label, StringRef(), V, false); - } - - void printBinary(StringRef Label, StringRef Value) { - auto V = makeArrayRef(reinterpret_cast<const uint8_t*>(Value.data()), - Value.size()); - printBinaryImpl(Label, StringRef(), V, false); - } - - void printBinaryBlock(StringRef Label, StringRef Value) { - auto V = makeArrayRef(reinterpret_cast<const uint8_t*>(Value.data()), - Value.size()); - printBinaryImpl(Label, StringRef(), V, true); - } - - raw_ostream& startLine() { - printIndent(); - return OS; - } - - raw_ostream& getOStream() { - return OS; - } - -private: - template<typename T> - static bool flagName(const EnumEntry<T>& lhs, const EnumEntry<T>& rhs) { - return lhs.Name < rhs.Name; - } - - void printBinaryImpl(StringRef Label, StringRef Str, ArrayRef<uint8_t> Value, - bool Block); - - raw_ostream &OS; - int IndentLevel; -}; - -struct DictScope { - DictScope(StreamWriter& W, StringRef N) : W(W) { - W.startLine() << N << " {\n"; - W.indent(); - } - - ~DictScope() { - W.unindent(); - W.startLine() << "}\n"; - } - - StreamWriter& W; -}; - -struct ListScope { - ListScope(StreamWriter& W, StringRef N) : W(W) { - W.startLine() << N << " [\n"; - W.indent(); - } - - ~ListScope() { - W.unindent(); - W.startLine() << "]\n"; - } - - StreamWriter& W; -}; - -} // namespace llvm - -#endif diff --git a/tools/llvm-readobj/Win64EHDumper.cpp b/tools/llvm-readobj/Win64EHDumper.cpp index 2da5ae3200fde..f7e56b3615425 100644 --- a/tools/llvm-readobj/Win64EHDumper.cpp +++ b/tools/llvm-readobj/Win64EHDumper.cpp @@ -120,13 +120,17 @@ static std::string formatSymbol(const Dumper::Context &Ctx, SymbolRef Symbol; if (!Ctx.ResolveSymbol(Section, Offset, Symbol, Ctx.UserData)) { - if (ErrorOr<StringRef> Name = Symbol.getName()) { + Expected<StringRef> Name = Symbol.getName(); + if (Name) { OS << *Name; if (Displacement > 0) OS << format(" +0x%X (0x%" PRIX64 ")", Displacement, Offset); else OS << format(" (0x%" PRIX64 ")", Offset); return OS.str(); + } else { + // TODO: Actually report errors helpfully. + consumeError(Name.takeError()); } } @@ -144,12 +148,14 @@ static std::error_code resolveRelocation(const Dumper::Context &Ctx, Ctx.ResolveSymbol(Section, Offset, Symbol, Ctx.UserData)) return EC; - ErrorOr<uint64_t> ResolvedAddressOrErr = Symbol.getAddress(); - if (std::error_code EC = ResolvedAddressOrErr.getError()) - return EC; + Expected<uint64_t> ResolvedAddressOrErr = Symbol.getAddress(); + if (!ResolvedAddressOrErr) + return errorToErrorCode(ResolvedAddressOrErr.takeError()); ResolvedAddress = *ResolvedAddressOrErr; - ErrorOr<section_iterator> SI = Symbol.getSection(); + Expected<section_iterator> SI = Symbol.getSection(); + if (!SI) + return errorToErrorCode(SI.takeError()); ResolvedSection = Ctx.COFF.getCOFFSection(**SI); return std::error_code(); } diff --git a/tools/llvm-readobj/Win64EHDumper.h b/tools/llvm-readobj/Win64EHDumper.h index a80df9c4f94dd..772f68bf283f7 100644 --- a/tools/llvm-readobj/Win64EHDumper.h +++ b/tools/llvm-readobj/Win64EHDumper.h @@ -10,7 +10,7 @@ #ifndef LLVM_TOOLS_LLVM_READOBJ_WIN64EHDUMPER_H #define LLVM_TOOLS_LLVM_READOBJ_WIN64EHDUMPER_H -#include "StreamWriter.h" +#include "llvm/Support/ScopedPrinter.h" #include "llvm/Support/Win64EH.h" namespace llvm { @@ -22,7 +22,7 @@ struct coff_section; namespace Win64EH { class Dumper { - StreamWriter &SW; + ScopedPrinter &SW; raw_ostream &OS; public: @@ -53,7 +53,7 @@ private: uint64_t SectionOffset, const RuntimeFunction &RF); public: - Dumper(StreamWriter &SW) : SW(SW), OS(SW.getOStream()) {} + Dumper(ScopedPrinter &SW) : SW(SW), OS(SW.getOStream()) {} void printData(const Context &Ctx); }; diff --git a/tools/llvm-readobj/llvm-readobj.cpp b/tools/llvm-readobj/llvm-readobj.cpp index fa8fee2b03ab7..c293919fd9527 100644 --- a/tools/llvm-readobj/llvm-readobj.cpp +++ b/tools/llvm-readobj/llvm-readobj.cpp @@ -22,7 +22,7 @@ #include "llvm-readobj.h" #include "Error.h" #include "ObjDumper.h" -#include "StreamWriter.h" +#include "llvm/DebugInfo/CodeView/MemoryTypeTableBuilder.h" #include "llvm/Object/Archive.h" #include "llvm/Object/COFFImportFile.h" #include "llvm/Object/ELFObjectFile.h" @@ -35,6 +35,7 @@ #include "llvm/Support/FileSystem.h" #include "llvm/Support/ManagedStatic.h" #include "llvm/Support/PrettyStackTrace.h" +#include "llvm/Support/ScopedPrinter.h" #include "llvm/Support/Signals.h" #include "llvm/Support/TargetRegistry.h" #include "llvm/Support/TargetSelect.h" @@ -144,6 +145,11 @@ namespace opts { cl::opt<bool> CodeView("codeview", cl::desc("Display CodeView debug information")); + // -codeview-merged-types + cl::opt<bool> + CodeViewMergedTypes("codeview-merged-types", + cl::desc("Display the merged CodeView type stream")); + // -codeview-subsection-bytes cl::opt<bool> CodeViewSubsectionBytes( "codeview-subsection-bytes", @@ -168,6 +174,10 @@ namespace opts { cl::opt<bool> MipsReginfo("mips-reginfo", cl::desc("Display the MIPS .reginfo section")); + // -mips-options + cl::opt<bool> MipsOptions("mips-options", + cl::desc("Display the MIPS .MIPS.options section")); + // -coff-imports cl::opt<bool> COFFImports("coff-imports", cl::desc("Display the PE/COFF import table")); @@ -186,6 +196,11 @@ namespace opts { COFFBaseRelocs("coff-basereloc", cl::desc("Display the PE/COFF .reloc section")); + // -coff-debug-directory + cl::opt<bool> + COFFDebugDirectory("coff-debug-directory", + cl::desc("Display the PE/COFF debug directory")); + // -macho-data-in-code cl::opt<bool> MachODataInCode("macho-data-in-code", @@ -227,6 +242,22 @@ namespace opts { cl::desc("Display ELF version sections (if present)")); cl::alias VersionInfoShort("V", cl::desc("Alias for -version-info"), cl::aliasopt(VersionInfo)); + + cl::opt<bool> SectionGroups("elf-section-groups", + cl::desc("Display ELF section group contents")); + cl::alias SectionGroupsShort("g", cl::desc("Alias for -elf-sections-groups"), + cl::aliasopt(SectionGroups)); + cl::opt<bool> HashHistogram( + "elf-hash-histogram", + cl::desc("Display bucket list histogram for hash sections")); + cl::alias HashHistogramShort("I", cl::desc("Alias for -elf-hash-histogram"), + cl::aliasopt(HashHistogram)); + + cl::opt<OutputStyleTy> + Output("elf-output-style", cl::desc("Specify ELF dump style"), + cl::values(clEnumVal(LLVM, "LLVM default style"), + clEnumVal(GNU, "GNU readelf style"), clEnumValEnd), + cl::init(LLVM)); } // namespace opts namespace llvm { @@ -264,6 +295,17 @@ static void reportError(StringRef Input, StringRef Message) { reportError(Twine(Input) + ": " + Message); } +static void reportError(StringRef Input, Error Err) { + if (Input == "-") + Input = "<stdin>"; + std::string ErrMsg; + { + raw_string_ostream ErrStream(ErrMsg); + logAllUnhandledErrors(std::move(Err), ErrStream, Input + ": "); + } + reportError(ErrMsg); +} + static bool isMipsArch(unsigned Arch) { switch (Arch) { case llvm::Triple::mips: @@ -276,8 +318,11 @@ static bool isMipsArch(unsigned Arch) { } } +static llvm::codeview::MemoryTypeTableBuilder CVTypes; + /// @brief Creates an format-specific object file dumper. -static std::error_code createDumper(const ObjectFile *Obj, StreamWriter &Writer, +static std::error_code createDumper(const ObjectFile *Obj, + ScopedPrinter &Writer, std::unique_ptr<ObjDumper> &Result) { if (!Obj) return readobj_error::unsupported_file_format; @@ -294,19 +339,20 @@ static std::error_code createDumper(const ObjectFile *Obj, StreamWriter &Writer, /// @brief Dumps the specified object file. static void dumpObject(const ObjectFile *Obj) { - StreamWriter Writer(outs()); + ScopedPrinter Writer(outs()); std::unique_ptr<ObjDumper> Dumper; if (std::error_code EC = createDumper(Obj, Writer, Dumper)) reportError(Obj->getFileName(), EC); - outs() << '\n'; - outs() << "File: " << Obj->getFileName() << "\n"; - outs() << "Format: " << Obj->getFileFormatName() << "\n"; - outs() << "Arch: " - << Triple::getArchTypeName((llvm::Triple::ArchType)Obj->getArch()) - << "\n"; - outs() << "AddressSize: " << (8*Obj->getBytesInAddress()) << "bit\n"; - Dumper->printLoadName(); + if (opts::Output == opts::LLVM) { + outs() << '\n'; + outs() << "File: " << Obj->getFileName() << "\n"; + outs() << "Format: " << Obj->getFileFormatName() << "\n"; + outs() << "Arch: " << Triple::getArchTypeName( + (llvm::Triple::ArchType)Obj->getArch()) << "\n"; + outs() << "AddressSize: " << (8 * Obj->getBytesInAddress()) << "bit\n"; + Dumper->printLoadName(); + } if (opts::FileHeaders) Dumper->printFileHeaders(); @@ -334,16 +380,24 @@ static void dumpObject(const ObjectFile *Obj) { Dumper->printGnuHashTable(); if (opts::VersionInfo) Dumper->printVersionInfo(); - if (Obj->getArch() == llvm::Triple::arm && Obj->isELF()) - if (opts::ARMAttributes) - Dumper->printAttributes(); - if (isMipsArch(Obj->getArch()) && Obj->isELF()) { - if (opts::MipsPLTGOT) - Dumper->printMipsPLTGOT(); - if (opts::MipsABIFlags) - Dumper->printMipsABIFlags(); - if (opts::MipsReginfo) - Dumper->printMipsReginfo(); + if (Obj->isELF()) { + if (Obj->getArch() == llvm::Triple::arm) + if (opts::ARMAttributes) + Dumper->printAttributes(); + if (isMipsArch(Obj->getArch())) { + if (opts::MipsPLTGOT) + Dumper->printMipsPLTGOT(); + if (opts::MipsABIFlags) + Dumper->printMipsABIFlags(); + if (opts::MipsReginfo) + Dumper->printMipsReginfo(); + if (opts::MipsOptions) + Dumper->printMipsOptions(); + } + if (opts::SectionGroups) + Dumper->printGroupSections(); + if (opts::HashHistogram) + Dumper->printHashHistogram(); } if (Obj->isCOFF()) { if (opts::COFFImports) @@ -354,8 +408,12 @@ static void dumpObject(const ObjectFile *Obj) { Dumper->printCOFFDirectives(); if (opts::COFFBaseRelocs) Dumper->printCOFFBaseReloc(); + if (opts::COFFDebugDirectory) + Dumper->printCOFFDebugDirectory(); if (opts::CodeView) Dumper->printCodeViewDebugInfo(); + if (opts::CodeViewMergedTypes) + Dumper->mergeCodeViewTypes(CVTypes); } if (Obj->isMachO()) { if (opts::MachODataInCode) @@ -377,35 +435,43 @@ static void dumpObject(const ObjectFile *Obj) { /// @brief Dumps each object file in \a Arc; static void dumpArchive(const Archive *Arc) { - for (auto &ErrorOrChild : Arc->children()) { - if (std::error_code EC = ErrorOrChild.getError()) - reportError(Arc->getFileName(), EC.message()); - const auto &Child = *ErrorOrChild; - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = Child.getAsBinary(); - if (std::error_code EC = ChildOrErr.getError()) { - // Ignore non-object files. - if (EC != object_error::invalid_file_type) - reportError(Arc->getFileName(), EC.message()); + Error Err; + for (auto &Child : Arc->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = Child.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType(ChildOrErr.takeError())) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(ChildOrErr.takeError(), OS, ""); + OS.flush(); + reportError(Arc->getFileName(), Buf); + } continue; } - if (ObjectFile *Obj = dyn_cast<ObjectFile>(&*ChildOrErr.get())) dumpObject(Obj); else reportError(Arc->getFileName(), readobj_error::unrecognized_file_format); } + if (Err) + reportError(Arc->getFileName(), std::move(Err)); } /// @brief Dumps each object file in \a MachO Universal Binary; static void dumpMachOUniversalBinary(const MachOUniversalBinary *UBinary) { for (const MachOUniversalBinary::ObjectForArch &Obj : UBinary->objects()) { - ErrorOr<std::unique_ptr<MachOObjectFile>> ObjOrErr = Obj.getAsObjectFile(); + Expected<std::unique_ptr<MachOObjectFile>> ObjOrErr = Obj.getAsObjectFile(); if (ObjOrErr) dumpObject(&*ObjOrErr.get()); - else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = Obj.getAsArchive()) + else if (auto E = isNotObjectErrorInvalidFileType(ObjOrErr.takeError())) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(ObjOrErr.takeError(), OS, ""); + OS.flush(); + reportError(UBinary->getFileName(), Buf); + } + else if (Expected<std::unique_ptr<Archive>> AOrErr = Obj.getAsArchive()) dumpArchive(&*AOrErr.get()); - else - reportError(UBinary->getFileName(), ObjOrErr.getError().message()); } } @@ -413,9 +479,9 @@ static void dumpMachOUniversalBinary(const MachOUniversalBinary *UBinary) { static void dumpInput(StringRef File) { // Attempt to open the binary. - ErrorOr<OwningBinary<Binary>> BinaryOrErr = createBinary(File); - if (std::error_code EC = BinaryOrErr.getError()) - reportError(File, EC); + Expected<OwningBinary<Binary>> BinaryOrErr = createBinary(File); + if (!BinaryOrErr) + reportError(File, errorToErrorCode(BinaryOrErr.takeError())); Binary &Binary = *BinaryOrErr.get().getBinary(); if (Archive *Arc = dyn_cast<Archive>(&Binary)) @@ -432,7 +498,7 @@ static void dumpInput(StringRef File) { } int main(int argc, const char *argv[]) { - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; @@ -448,5 +514,10 @@ int main(int argc, const char *argv[]) { std::for_each(opts::InputFilenames.begin(), opts::InputFilenames.end(), dumpInput); + if (opts::CodeViewMergedTypes) { + ScopedPrinter W(outs()); + dumpCodeViewMergedTypes(W, CVTypes); + } + return 0; } diff --git a/tools/llvm-readobj/llvm-readobj.h b/tools/llvm-readobj/llvm-readobj.h index 5a103920c165e..b169c00291838 100644 --- a/tools/llvm-readobj/llvm-readobj.h +++ b/tools/llvm-readobj/llvm-readobj.h @@ -12,6 +12,8 @@ #include "llvm/Support/CommandLine.h" #include "llvm/Support/Compiler.h" +#include "llvm/Support/ErrorOr.h" +#include "llvm/Support/Error.h" #include <string> namespace llvm { @@ -22,6 +24,20 @@ namespace llvm { // Various helper functions. LLVM_ATTRIBUTE_NORETURN void reportError(Twine Msg); void error(std::error_code ec); + template <class T> T unwrapOrError(ErrorOr<T> EO) { + if (EO) + return *EO; + reportError(EO.getError().message()); + } + template <class T> T unwrapOrError(Expected<T> EO) { + if (EO) + return *EO; + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(EO.takeError(), OS, ""); + OS.flush(); + reportError(Buf); + } bool relocAddressLess(object::RelocationRef A, object::RelocationRef B); } // namespace llvm @@ -42,9 +58,14 @@ namespace opts { extern llvm::cl::opt<bool> CodeViewSubsectionBytes; extern llvm::cl::opt<bool> ARMAttributes; extern llvm::cl::opt<bool> MipsPLTGOT; + enum OutputStyleTy { LLVM, GNU }; + extern llvm::cl::opt<OutputStyleTy> Output; } // namespace opts #define LLVM_READOBJ_ENUM_ENT(ns, enum) \ { #enum, ns::enum } +#define LLVM_READOBJ_ENUM_CLASS_ENT(enum_class, enum) \ + { #enum, std::underlying_type<enum_class>::type(enum_class::enum) } + #endif diff --git a/tools/llvm-rtdyld/Makefile b/tools/llvm-rtdyld/Makefile deleted file mode 100644 index 3e868b9858724..0000000000000 --- a/tools/llvm-rtdyld/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-rtdyld/Makefile --------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-rtdyld -LINK_COMPONENTS := all-targets support MC object RuntimeDyld MCJIT DebugInfoDWARF - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-rtdyld/llvm-rtdyld.cpp b/tools/llvm-rtdyld/llvm-rtdyld.cpp index 6ee3a44b63bf5..b1460e35de80f 100644 --- a/tools/llvm-rtdyld/llvm-rtdyld.cpp +++ b/tools/llvm-rtdyld/llvm-rtdyld.cpp @@ -19,7 +19,7 @@ #include "llvm/ExecutionEngine/RuntimeDyldChecker.h" #include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCContext.h" -#include "llvm/MC/MCDisassembler.h" +#include "llvm/MC/MCDisassembler/MCDisassembler.h" #include "llvm/MC/MCInstPrinter.h" #include "llvm/MC/MCInstrInfo.h" #include "llvm/MC/MCRegisterInfo.h" @@ -191,7 +191,7 @@ public: } uint8_t *allocateFromSlab(uintptr_t Size, unsigned Alignment, bool isCode) { - Size = RoundUpToAlignment(Size, Alignment); + Size = alignTo(Size, Alignment); if (CurrentSlabOffset + Size > SlabSize) report_fatal_error("Can't allocate enough memory. Tune --preallocate"); @@ -254,9 +254,9 @@ uint8_t *TrivialMemoryManager::allocateDataSection(uintptr_t Size, static const char *ProgramName; -static int Error(const Twine &Msg) { +static void ErrorAndExit(const Twine &Msg) { errs() << ProgramName << ": error: " << Msg << "\n"; - return 1; + exit(1); } static void loadDylibs() { @@ -290,13 +290,18 @@ static int printLineInfoForInput(bool LoadObjects, bool UseDebugObj) { ErrorOr<std::unique_ptr<MemoryBuffer>> InputBuffer = MemoryBuffer::getFileOrSTDIN(File); if (std::error_code EC = InputBuffer.getError()) - return Error("unable to read input: '" + EC.message() + "'"); + ErrorAndExit("unable to read input: '" + EC.message() + "'"); - ErrorOr<std::unique_ptr<ObjectFile>> MaybeObj( + Expected<std::unique_ptr<ObjectFile>> MaybeObj( ObjectFile::createObjectFile((*InputBuffer)->getMemBufferRef())); - if (std::error_code EC = MaybeObj.getError()) - return Error("unable to create object file: '" + EC.message() + "'"); + if (!MaybeObj) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(MaybeObj.takeError(), OS, ""); + OS.flush(); + ErrorAndExit("unable to create object file: '" + Buf + "'"); + } ObjectFile &Obj = **MaybeObj; @@ -309,7 +314,7 @@ static int printLineInfoForInput(bool LoadObjects, bool UseDebugObj) { Dyld.loadObject(Obj); if (Dyld.hasError()) - return Error(Dyld.getErrorString()); + ErrorAndExit(Dyld.getErrorString()); // Resolve all the relocations we can. Dyld.resolveRelocations(); @@ -330,13 +335,26 @@ static int printLineInfoForInput(bool LoadObjects, bool UseDebugObj) { // Use symbol info to iterate functions in the object. for (const auto &P : SymAddr) { object::SymbolRef Sym = P.first; - if (Sym.getType() == object::SymbolRef::ST_Function) { - ErrorOr<StringRef> Name = Sym.getName(); - if (!Name) + Expected<SymbolRef::Type> TypeOrErr = Sym.getType(); + if (!TypeOrErr) { + // TODO: Actually report errors helpfully. + consumeError(TypeOrErr.takeError()); + continue; + } + SymbolRef::Type Type = *TypeOrErr; + if (Type == object::SymbolRef::ST_Function) { + Expected<StringRef> Name = Sym.getName(); + if (!Name) { + // TODO: Actually report errors helpfully. + consumeError(Name.takeError()); continue; - ErrorOr<uint64_t> AddrOrErr = Sym.getAddress(); - if (!AddrOrErr) + } + Expected<uint64_t> AddrOrErr = Sym.getAddress(); + if (!AddrOrErr) { + // TODO: Actually report errors helpfully. + consumeError(AddrOrErr.takeError()); continue; + } uint64_t Addr = *AddrOrErr; uint64_t Size = P.second; @@ -344,7 +362,13 @@ static int printLineInfoForInput(bool LoadObjects, bool UseDebugObj) { // symbol in memory (rather than that in the unrelocated object file) // and use that to query the DWARFContext. if (!UseDebugObj && LoadObjects) { - object::section_iterator Sec = *Sym.getSection(); + auto SecOrErr = Sym.getSection(); + if (!SecOrErr) { + // TODO: Actually report errors helpfully. + consumeError(SecOrErr.takeError()); + continue; + } + object::section_iterator Sec = *SecOrErr; StringRef SecName; Sec->getName(SecName); uint64_t SectionLoadAddress = @@ -396,19 +420,24 @@ static int executeInput() { ErrorOr<std::unique_ptr<MemoryBuffer>> InputBuffer = MemoryBuffer::getFileOrSTDIN(File); if (std::error_code EC = InputBuffer.getError()) - return Error("unable to read input: '" + EC.message() + "'"); - ErrorOr<std::unique_ptr<ObjectFile>> MaybeObj( + ErrorAndExit("unable to read input: '" + EC.message() + "'"); + Expected<std::unique_ptr<ObjectFile>> MaybeObj( ObjectFile::createObjectFile((*InputBuffer)->getMemBufferRef())); - if (std::error_code EC = MaybeObj.getError()) - return Error("unable to create object file: '" + EC.message() + "'"); + if (!MaybeObj) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(MaybeObj.takeError(), OS, ""); + OS.flush(); + ErrorAndExit("unable to create object file: '" + Buf + "'"); + } ObjectFile &Obj = **MaybeObj; // Load the object file Dyld.loadObject(Obj); if (Dyld.hasError()) { - return Error(Dyld.getErrorString()); + ErrorAndExit(Dyld.getErrorString()); } } @@ -419,7 +448,7 @@ static int executeInput() { // Get the address of the entry point (_main by default). void *MainAddress = Dyld.getSymbolLocalAddress(EntryPoint); if (!MainAddress) - return Error("no definition for '" + EntryPoint + "'"); + ErrorAndExit("no definition for '" + EntryPoint + "'"); // Invalidate the instruction cache for each loaded function. for (auto &FM : MemMgr.FunctionMemory) { @@ -428,7 +457,7 @@ static int executeInput() { // setExecutable will call InvalidateInstructionCache. std::string ErrorStr; if (!sys::Memory::setExecutable(FM, &ErrorStr)) - return Error("unable to mark function executable: '" + ErrorStr + "'"); + ErrorAndExit("unable to mark function executable: '" + ErrorStr + "'"); } // Dispatch to _main(). @@ -448,12 +477,12 @@ static int checkAllExpressions(RuntimeDyldChecker &Checker) { ErrorOr<std::unique_ptr<MemoryBuffer>> CheckerFileBuf = MemoryBuffer::getFileOrSTDIN(CheckerFileName); if (std::error_code EC = CheckerFileBuf.getError()) - return Error("unable to read input '" + CheckerFileName + "': " + + ErrorAndExit("unable to read input '" + CheckerFileName + "': " + EC.message()); if (!Checker.checkAllRulesInBuffer("# rtdyld-check:", CheckerFileBuf.get().get())) - return Error("some checks in '" + CheckerFileName + "' failed"); + ErrorAndExit("some checks in '" + CheckerFileName + "' failed"); } return 0; } @@ -602,7 +631,7 @@ static int linkAndVerify() { // Check for missing triple. if (TripleName == "") - return Error("-triple required when running in -verify mode."); + ErrorAndExit("-triple required when running in -verify mode."); // Look up the target and build the disassembler. Triple TheTriple(Triple::normalize(TripleName)); @@ -610,29 +639,29 @@ static int linkAndVerify() { const Target *TheTarget = TargetRegistry::lookupTarget("", TheTriple, ErrorStr); if (!TheTarget) - return Error("Error accessing target '" + TripleName + "': " + ErrorStr); + ErrorAndExit("Error accessing target '" + TripleName + "': " + ErrorStr); TripleName = TheTriple.getTriple(); std::unique_ptr<MCSubtargetInfo> STI( TheTarget->createMCSubtargetInfo(TripleName, MCPU, "")); if (!STI) - return Error("Unable to create subtarget info!"); + ErrorAndExit("Unable to create subtarget info!"); std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName)); if (!MRI) - return Error("Unable to create target register info!"); + ErrorAndExit("Unable to create target register info!"); std::unique_ptr<MCAsmInfo> MAI(TheTarget->createMCAsmInfo(*MRI, TripleName)); if (!MAI) - return Error("Unable to create target asm info!"); + ErrorAndExit("Unable to create target asm info!"); MCContext Ctx(MAI.get(), MRI.get(), nullptr); std::unique_ptr<MCDisassembler> Disassembler( TheTarget->createMCDisassembler(*STI, Ctx)); if (!Disassembler) - return Error("Unable to create disassembler!"); + ErrorAndExit("Unable to create disassembler!"); std::unique_ptr<MCInstrInfo> MII(TheTarget->createMCInstrInfo()); @@ -659,20 +688,25 @@ static int linkAndVerify() { MemoryBuffer::getFileOrSTDIN(Filename); if (std::error_code EC = InputBuffer.getError()) - return Error("unable to read input: '" + EC.message() + "'"); + ErrorAndExit("unable to read input: '" + EC.message() + "'"); - ErrorOr<std::unique_ptr<ObjectFile>> MaybeObj( + Expected<std::unique_ptr<ObjectFile>> MaybeObj( ObjectFile::createObjectFile((*InputBuffer)->getMemBufferRef())); - if (std::error_code EC = MaybeObj.getError()) - return Error("unable to create object file: '" + EC.message() + "'"); + if (!MaybeObj) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(MaybeObj.takeError(), OS, ""); + OS.flush(); + ErrorAndExit("unable to create object file: '" + Buf + "'"); + } ObjectFile &Obj = **MaybeObj; // Load the object file Dyld.loadObject(Obj); if (Dyld.hasError()) { - return Error(Dyld.getErrorString()); + ErrorAndExit(Dyld.getErrorString()); } } @@ -688,14 +722,14 @@ static int linkAndVerify() { int ErrorCode = checkAllExpressions(Checker); if (Dyld.hasError()) - return Error("RTDyld reported an error applying relocations:\n " + + ErrorAndExit("RTDyld reported an error applying relocations:\n " + Dyld.getErrorString()); return ErrorCode; } int main(int argc, char **argv) { - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); ProgramName = argv[0]; diff --git a/tools/llvm-shlib/CMakeLists.txt b/tools/llvm-shlib/CMakeLists.txt index 2356103a9cd51..3fe672d679a3d 100644 --- a/tools/llvm-shlib/CMakeLists.txt +++ b/tools/llvm-shlib/CMakeLists.txt @@ -2,8 +2,6 @@ # library is enabled by setting LLVM_BUILD_LLVM_DYLIB=yes on the CMake # commandline. By default the shared library only exports the LLVM C API. -add_definitions( -DLLVM_VERSION_INFO=\"${PACKAGE_VERSION}\" ) - set(SOURCES libllvm.cpp ) diff --git a/tools/llvm-shlib/Makefile b/tools/llvm-shlib/Makefile deleted file mode 100644 index 19077a3858a6a..0000000000000 --- a/tools/llvm-shlib/Makefile +++ /dev/null @@ -1,116 +0,0 @@ -##===- tools/shlib/Makefile --------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. - -LIBRARYNAME = LLVM-$(LLVM_VERSION_MAJOR).$(LLVM_VERSION_MINOR)$(LLVM_VERSION_SUFFIX) -LIBRARYALIASNAME = LLVM-$(LLVMVersion) - -NO_BUILD_ARCHIVE := 1 -LINK_LIBS_IN_SHARED := 1 -SHARED_LIBRARY := 1 -SHARED_ALIAS := 1 - -include $(LEVEL)/Makefile.config - -ifeq ($(HOST_OS), $(filter $(HOST_OS), Cygwin MingW)) - EXPORTED_SYMBOL_FILE = $(ObjDir)/$(LIBRARYNAME).exports - - ifeq (1,$(ENABLE_EMBED_STDCXX)) - # It is needed to force static-stdc++.a linked. - SHLIB_FRAG_NAMES += stdc++.a.o - endif - -endif - -include $(LEVEL)/Makefile.common - -# Include all archives in libLLVM.(so|dylib) except the ones that have -# their own dynamic libraries and TableGen. -Archives := $(wildcard $(LibDir)/libLLVM*.a) -SharedLibraries := $(wildcard $(LibDir)/libLLVM*$(SHLIBEXT)) -ExcludeFromLibLlvm := $(basename $(SharedLibraries)).a %/libLLVMTableGen.a -IncludeInLibLlvm := $(filter-out $(ExcludeFromLibLlvm), $(Archives)) -LLVMLibsOptions := $(IncludeInLibLlvm:$(LibDir)/lib%.a=-l%) -LLVMLibsPaths := $(IncludeInLibLlvm) - -$(LibName.SO): $(LLVMLibsPaths) - -ifeq ($(HOST_OS),Darwin) - # set dylib internal version number to llvmCore submission number - ifdef LLVM_SUBMIT_VERSION - LLVMLibsOptions := $(LLVMLibsOptions) -Wl,-current_version \ - -Wl,$(LLVM_SUBMIT_VERSION).$(LLVM_SUBMIT_SUBVERSION) \ - -Wl,-compatibility_version -Wl,1 - endif - # Include everything from the .a's into the shared library. - LLVMLibsOptions := $(LLVMLibsOptions) -all_load -endif - -ifeq ($(HOST_OS), $(filter $(HOST_OS), DragonFly Linux FreeBSD GNU/kFreeBSD OpenBSD GNU Bitrig)) - # Include everything from the .a's into the shared library. - LLVMLibsOptions := -Wl,--whole-archive $(LLVMLibsOptions) \ - -Wl,--no-whole-archive -endif - -ifeq ($(HOST_OS), $(filter $(HOST_OS), DragonFly Linux FreeBSD GNU/kFreeBSD GNU)) - # Add soname to the library. - LLVMLibsOptions += -Wl,--soname,lib$(LIBRARYNAME)$(SHLIBEXT) -endif - -ifeq ($(HOST_OS), $(filter $(HOST_OS), Linux GNU GNU/kFreeBSD)) - # Don't allow unresolved symbols. - LLVMLibsOptions += -Wl,--no-undefined -endif - -ifeq ($(HOST_OS),SunOS) - # add -z allextract ahead of other libraries on Solaris - LLVMLibsOptions := -Wl,-z -Wl,allextract $(LLVMLibsOptions) -endif - -ifeq ($(HOST_OS), $(filter $(HOST_OS), Cygwin MingW)) - -SHLIB_STUBS := $(addprefix $(ObjDir)/, $(SHLIB_FRAG_NAMES)) -SHLIB_FRAGS := $(patsubst %.a.o, $(ObjDir)/%.syms.txt, $(LIBRARYNAME).a.o $(SHLIB_FRAG_NAMES)) -LLVMLibsOptions := $(SHLIB_STUBS) $(LLVMLibsOptions) - -$(LibName.SO): $(SHLIB_STUBS) - -%.syms.txt: %.a.o - $(Echo) Collecting global symbols of $(notdir $*) - $(Verb) $(NM_PATH) -g $< > $@ - -$(ObjDir)/$(LIBRARYNAME).exports: $(SHLIB_FRAGS) $(ObjDir)/.dir - $(Echo) Generating exports for $(LIBRARYNAME) - $(Verb) ($(SED) -n \ - -e "s/^.* T _\([^.][^.]*\)$$/\1/p" \ - -e "s/^.* [BDR] _\([^.][^.]*\)$$/\1 DATA/p" \ - $(SHLIB_FRAGS) \ - | sort -u) > $@ - -$(ObjDir)/$(LIBRARYNAME).a.o: $(LLVMLibsPaths) $(ObjDir)/.dir - $(Echo) Linking all LLVMLibs together for $(LIBRARYNAME) - $(Verb) $(Link) -nostartfiles -Wl,-r -nodefaultlibs -o $@ \ - -Wl,--whole-archive $(LLVMLibsPaths) \ - -Wl,--no-whole-archive - -$(ObjDir)/stdc++.a.o: $(ObjDir)/.dir - $(Echo) Linking all libs together for static libstdc++.a - $(Verb) $(Link) -nostartfiles -Wl,-r -nodefaultlibs -o $@ \ - -Wl,--whole-archive -lstdc++ \ - -Wl,--no-whole-archive -# FIXME: workaround to invalidate -lstdc++ - $(Echo) Making dummy -lstdc++ to lib - $(Verb) $(AR) rc $(ToolDir)/libstdc++.dll.a -# FIXME: Is install-local needed? - -clean-local:: - $(Verb) $(RM) -f $(ToolDir)/libstdc++.dll.a - -endif diff --git a/tools/llvm-shlib/libllvm.cpp b/tools/llvm-shlib/libllvm.cpp index 8424d660c9d0e..40b4f66b07331 100644 --- a/tools/llvm-shlib/libllvm.cpp +++ b/tools/llvm-shlib/libllvm.cpp @@ -11,10 +11,3 @@ // you can't define a target with no sources. // //===----------------------------------------------------------------------===// - -#include "llvm/Config/config.h" - -#if defined(DISABLE_LLVM_DYLIB_ATEXIT) -extern "C" int __cxa_atexit(); -extern "C" int __cxa_atexit() { return 0; } -#endif diff --git a/tools/llvm-size/Makefile b/tools/llvm-size/Makefile deleted file mode 100644 index 0622eb108978f..0000000000000 --- a/tools/llvm-size/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-size/Makefile ----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-size -LINK_COMPONENTS := object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS = 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-size/llvm-size.cpp b/tools/llvm-size/llvm-size.cpp index 069cc621f6157..c5966ead4b6c5 100644 --- a/tools/llvm-size/llvm-size.cpp +++ b/tools/llvm-size/llvm-size.cpp @@ -15,6 +15,7 @@ #include "llvm/ADT/APInt.h" #include "llvm/Object/Archive.h" +#include "llvm/Object/ELFObjectFile.h" #include "llvm/Object/MachO.h" #include "llvm/Object/MachOUniversal.h" #include "llvm/Object/ObjectFile.h" @@ -49,13 +50,19 @@ static cl::opt<OutputFormatTy> OutputFormatShort( clEnumValN(darwin, "m", "Darwin -m format"), clEnumValEnd), cl::init(berkeley)); -static bool berkeleyHeaderPrinted = false; -static bool moreThanOneFile = false; +static bool BerkeleyHeaderPrinted = false; +static bool MoreThanOneFile = false; cl::opt<bool> DarwinLongFormat("l", cl::desc("When format is darwin, use long format " "to include addresses and offsets.")); +cl::opt<bool> + ELFCommons("common", + cl::desc("Print common symbols in the ELF file. When using " + "Berkely format, this is added to bss."), + cl::init(false)); + static cl::list<std::string> ArchFlags("arch", cl::desc("architecture(s) from a Mach-O file to dump"), cl::ZeroOrMore); @@ -77,20 +84,75 @@ RadixShort(cl::desc("Print size in radix:"), static cl::list<std::string> InputFilenames(cl::Positional, cl::desc("<input files>"), cl::ZeroOrMore); +bool HadError = false; + static std::string ToolName; -/// @brief If ec is not success, print the error and return true. +/// If ec is not success, print the error and return true. static bool error(std::error_code ec) { if (!ec) return false; - outs() << ToolName << ": error reading file: " << ec.message() << ".\n"; - outs().flush(); + HadError = true; + errs() << ToolName << ": error reading file: " << ec.message() << ".\n"; + errs().flush(); + return true; +} + +static bool error(Twine Message) { + HadError = true; + errs() << ToolName << ": " << Message << ".\n"; + errs().flush(); return true; } -/// @brief Get the length of the string that represents @p num in Radix -/// including the leading 0x or 0 for hexadecimal and octal respectively. +// This version of error() prints the archive name and member name, for example: +// "libx.a(foo.o)" after the ToolName before the error message. It sets +// HadError but returns allowing the code to move on to other archive members. +static void error(llvm::Error E, StringRef FileName, const Archive::Child &C, + StringRef ArchitectureName = StringRef()) { + HadError = true; + errs() << ToolName << ": " << FileName; + + ErrorOr<StringRef> NameOrErr = C.getName(); + // TODO: if we have a error getting the name then it would be nice to print + // the index of which archive member this is and or its offset in the + // archive instead of "???" as the name. + if (NameOrErr.getError()) + errs() << "(" << "???" << ")"; + else + errs() << "(" << NameOrErr.get() << ")"; + + if (!ArchitectureName.empty()) + errs() << " (for architecture " << ArchitectureName << ") "; + + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(std::move(E), OS, ""); + OS.flush(); + errs() << " " << Buf << "\n"; +} + +// This version of error() prints the file name and which architecture slice it // is from, for example: "foo.o (for architecture i386)" after the ToolName +// before the error message. It sets HadError but returns allowing the code to +// move on to other architecture slices. +static void error(llvm::Error E, StringRef FileName, + StringRef ArchitectureName = StringRef()) { + HadError = true; + errs() << ToolName << ": " << FileName; + + if (!ArchitectureName.empty()) + errs() << " (for architecture " << ArchitectureName << ") "; + + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(std::move(E), OS, ""); + OS.flush(); + errs() << " " << Buf << "\n"; +} + +/// Get the length of the string that represents @p num in Radix including the +/// leading 0x or 0 for hexadecimal and octal respectively. static size_t getNumLengthAsString(uint64_t num) { APInt conv(64, num); SmallString<32> result; @@ -98,7 +160,7 @@ static size_t getNumLengthAsString(uint64_t num) { return result.size(); } -/// @brief Return the printing format for the Radix. +/// Return the printing format for the Radix. static const char *getRadixFmt() { switch (Radix) { case octal: @@ -111,11 +173,35 @@ static const char *getRadixFmt() { return nullptr; } -/// @brief Print the size of each Mach-O segment and section in @p MachO. +/// Remove unneeded ELF sections from calculation +static bool considerForSize(ObjectFile *Obj, SectionRef Section) { + if (!Obj->isELF()) + return true; + switch (static_cast<ELFSectionRef>(Section).getType()) { + case ELF::SHT_NULL: + case ELF::SHT_SYMTAB: + case ELF::SHT_STRTAB: + case ELF::SHT_REL: + case ELF::SHT_RELA: + return false; + } + return true; +} + +/// Total size of all ELF common symbols +static uint64_t getCommonSize(ObjectFile *Obj) { + uint64_t TotalCommons = 0; + for (auto &Sym : Obj->symbols()) + if (Obj->getSymbolFlags(Sym.getRawDataRefImpl()) & SymbolRef::SF_Common) + TotalCommons += Obj->getCommonSymbolSize(Sym.getRawDataRefImpl()); + return TotalCommons; +} + +/// Print the size of each Mach-O segment and section in @p MachO. /// /// This is when used when @c OutputFormat is darwin and produces the same /// output as darwin's size(1) -m output. -static void PrintDarwinSectionSizes(MachOObjectFile *MachO) { +static void printDarwinSectionSizes(MachOObjectFile *MachO) { std::string fmtbuf; raw_string_ostream fmt(fmtbuf); const char *radix_fmt = getRadixFmt(); @@ -155,10 +241,11 @@ static void PrintDarwinSectionSizes(MachOObjectFile *MachO) { outs() << "\ttotal " << format(fmt.str().c_str(), sec_total) << "\n"; } else if (Load.C.cmd == MachO::LC_SEGMENT) { MachO::segment_command Seg = MachO->getSegmentLoadCommand(Load); + uint64_t Seg_vmsize = Seg.vmsize; outs() << "Segment " << Seg.segname << ": " - << format(fmt.str().c_str(), Seg.vmsize); + << format(fmt.str().c_str(), Seg_vmsize); if (DarwinLongFormat) - outs() << " (vmaddr 0x" << format("%" PRIx64, Seg.vmaddr) << " fileoff " + outs() << " (vmaddr 0x" << format("%" PRIx32, Seg.vmaddr) << " fileoff " << Seg.fileoff << ")"; outs() << "\n"; total += Seg.vmsize; @@ -170,9 +257,10 @@ static void PrintDarwinSectionSizes(MachOObjectFile *MachO) { << format("%.16s", &Sec.sectname) << "): "; else outs() << "\tSection " << format("%.16s", &Sec.sectname) << ": "; - outs() << format(fmt.str().c_str(), Sec.size); + uint64_t Sec_size = Sec.size; + outs() << format(fmt.str().c_str(), Sec_size); if (DarwinLongFormat) - outs() << " (addr 0x" << format("%" PRIx64, Sec.addr) << " offset " + outs() << " (addr 0x" << format("%" PRIx32, Sec.addr) << " offset " << Sec.offset << ")"; outs() << "\n"; sec_total += Sec.size; @@ -184,11 +272,11 @@ static void PrintDarwinSectionSizes(MachOObjectFile *MachO) { outs() << "total " << format(fmt.str().c_str(), total) << "\n"; } -/// @brief Print the summary sizes of the standard Mach-O segments in @p MachO. +/// Print the summary sizes of the standard Mach-O segments in @p MachO. /// /// This is when used when @c OutputFormat is berkeley with a Mach-O file and /// produces the same output as darwin's size(1) default output. -static void PrintDarwinSegmentSizes(MachOObjectFile *MachO) { +static void printDarwinSegmentSizes(MachOObjectFile *MachO) { uint64_t total_text = 0; uint64_t total_data = 0; uint64_t total_objc = 0; @@ -250,19 +338,19 @@ static void PrintDarwinSegmentSizes(MachOObjectFile *MachO) { } uint64_t total = total_text + total_data + total_objc + total_others; - if (!berkeleyHeaderPrinted) { + if (!BerkeleyHeaderPrinted) { outs() << "__TEXT\t__DATA\t__OBJC\tothers\tdec\thex\n"; - berkeleyHeaderPrinted = true; + BerkeleyHeaderPrinted = true; } outs() << total_text << "\t" << total_data << "\t" << total_objc << "\t" << total_others << "\t" << total << "\t" << format("%" PRIx64, total) << "\t"; } -/// @brief Print the size of each section in @p Obj. +/// Print the size of each section in @p Obj. /// /// The format used is determined by @c OutputFormat and @c Radix. -static void PrintObjectSectionSizes(ObjectFile *Obj) { +static void printObjectSectionSizes(ObjectFile *Obj) { uint64_t total = 0; std::string fmtbuf; raw_string_ostream fmt(fmtbuf); @@ -273,11 +361,11 @@ static void PrintObjectSectionSizes(ObjectFile *Obj) { // let it fall through to OutputFormat berkeley. MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(Obj); if (OutputFormat == darwin && MachO) - PrintDarwinSectionSizes(MachO); + printDarwinSectionSizes(MachO); // If we have a MachOObjectFile and the OutputFormat is berkeley print as // darwin's default berkeley format for Mach-O files. else if (MachO && OutputFormat == berkeley) - PrintDarwinSegmentSizes(MachO); + printDarwinSegmentSizes(MachO); else if (OutputFormat == sysv) { // Run two passes over all sections. The first gets the lengths needed for // formatting the output. The second actually does the output. @@ -285,6 +373,8 @@ static void PrintObjectSectionSizes(ObjectFile *Obj) { std::size_t max_size_len = strlen("size"); std::size_t max_addr_len = strlen("addr"); for (const SectionRef &Section : Obj->sections()) { + if (!considerForSize(Obj, Section)) + continue; uint64_t size = Section.getSize(); total += size; @@ -320,6 +410,8 @@ static void PrintObjectSectionSizes(ObjectFile *Obj) { // Print each section. for (const SectionRef &Section : Obj->sections()) { + if (!considerForSize(Obj, Section)) + continue; StringRef name; if (error(Section.getName(name))) return; @@ -330,6 +422,13 @@ static void PrintObjectSectionSizes(ObjectFile *Obj) { outs() << format(fmt.str().c_str(), namestr.c_str(), size, addr); } + if (ELFCommons) { + uint64_t CommonSize = getCommonSize(Obj); + total += CommonSize; + outs() << format(fmt.str().c_str(), std::string("*COM*").c_str(), + CommonSize, static_cast<uint64_t>(0)); + } + // Print total. fmtbuf.clear(); fmt << "%-" << max_name_len << "s " @@ -357,12 +456,15 @@ static void PrintObjectSectionSizes(ObjectFile *Obj) { total_bss += size; } + if (ELFCommons) + total_bss += getCommonSize(Obj); + total = total_text + total_data + total_bss; - if (!berkeleyHeaderPrinted) { + if (!BerkeleyHeaderPrinted) { outs() << " text data bss " << (Radix == octal ? "oct" : "dec") << " hex filename\n"; - berkeleyHeaderPrinted = true; + BerkeleyHeaderPrinted = true; } // Print result. @@ -377,11 +479,11 @@ static void PrintObjectSectionSizes(ObjectFile *Obj) { } } -/// @brief Checks to see if the @p o ObjectFile is a Mach-O file and if it is -/// and there is a list of architecture flags specified then check to -/// make sure this Mach-O file is one of those architectures or all -/// architectures was specificed. If not then an error is generated and -/// this routine returns false. Else it returns true. +/// Checks to see if the @p o ObjectFile is a Mach-O file and if it is and there +/// is a list of architecture flags specified then check to make sure this +/// Mach-O file is one of those architectures or all architectures was +/// specificed. If not then an error is generated and this routine returns +/// false. Else it returns true. static bool checkMachOAndArchFlags(ObjectFile *o, StringRef file) { if (isa<MachOObjectFile>(o) && !ArchAll && ArchFlags.size() != 0) { MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o); @@ -391,10 +493,10 @@ static bool checkMachOAndArchFlags(ObjectFile *o, StringRef file) { Triple T; if (MachO->is64Bit()) { H_64 = MachO->MachOObjectFile::getHeader64(); - T = MachOObjectFile::getArch(H_64.cputype, H_64.cpusubtype); + T = MachOObjectFile::getArchTriple(H_64.cputype, H_64.cpusubtype); } else { H = MachO->MachOObjectFile::getHeader(); - T = MachOObjectFile::getArch(H.cputype, H.cpusubtype); + T = MachOObjectFile::getArchTriple(H.cputype, H.cpusubtype); } unsigned i; for (i = 0; i < ArchFlags.size(); ++i) { @@ -411,32 +513,26 @@ static bool checkMachOAndArchFlags(ObjectFile *o, StringRef file) { return true; } -/// @brief Print the section sizes for @p file. If @p file is an archive, print -/// the section sizes for each archive member. -static void PrintFileSectionSizes(StringRef file) { +/// Print the section sizes for @p file. If @p file is an archive, print the +/// section sizes for each archive member. +static void printFileSectionSizes(StringRef file) { // Attempt to open the binary. - ErrorOr<OwningBinary<Binary>> BinaryOrErr = createBinary(file); - if (std::error_code EC = BinaryOrErr.getError()) { - errs() << ToolName << ": " << file << ": " << EC.message() << ".\n"; + Expected<OwningBinary<Binary>> BinaryOrErr = createBinary(file); + if (!BinaryOrErr) { + error(errorToErrorCode(BinaryOrErr.takeError())); return; } Binary &Bin = *BinaryOrErr.get().getBinary(); if (Archive *a = dyn_cast<Archive>(&Bin)) { // This is an archive. Iterate over each member and display its sizes. - for (object::Archive::child_iterator i = a->child_begin(), - e = a->child_end(); - i != e; ++i) { - if (i->getError()) { - errs() << ToolName << ": " << file << ": " << i->getError().message() - << ".\n"; - exit(1); - } - auto &c = i->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = c.getAsBinary(); - if (std::error_code EC = ChildOrErr.getError()) { - errs() << ToolName << ": " << file << ": " << EC.message() << ".\n"; + Error Err; + for (auto &C : a->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType(ChildOrErr.takeError())) + error(std::move(E), a->getFileName(), C); continue; } if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) { @@ -447,7 +543,7 @@ static void PrintFileSectionSizes(StringRef file) { outs() << o->getFileName() << " (ex " << a->getFileName() << "):\n"; else if (MachO && OutputFormat == darwin) outs() << a->getFileName() << "(" << o->getFileName() << "):\n"; - PrintObjectSectionSizes(o); + printObjectSectionSizes(o); if (OutputFormat == berkeley) { if (MachO) outs() << a->getFileName() << "(" << o->getFileName() << ")\n"; @@ -456,6 +552,8 @@ static void PrintFileSectionSizes(StringRef file) { } } } + if (Err) + error(std::move(Err), a->getFileName()); } else if (MachOUniversalBinary *UB = dyn_cast<MachOUniversalBinary>(&Bin)) { // If we have a list of architecture flags specified dump only those. @@ -469,43 +567,44 @@ static void PrintFileSectionSizes(StringRef file) { I != E; ++I) { if (ArchFlags[i] == I->getArchTypeName()) { ArchFound = true; - ErrorOr<std::unique_ptr<ObjectFile>> UO = I->getAsObjectFile(); + Expected<std::unique_ptr<ObjectFile>> UO = I->getAsObjectFile(); if (UO) { if (ObjectFile *o = dyn_cast<ObjectFile>(&*UO.get())) { MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o); if (OutputFormat == sysv) outs() << o->getFileName() << " :\n"; else if (MachO && OutputFormat == darwin) { - if (moreThanOneFile || ArchFlags.size() > 1) + if (MoreThanOneFile || ArchFlags.size() > 1) outs() << o->getFileName() << " (for architecture " << I->getArchTypeName() << "): \n"; } - PrintObjectSectionSizes(o); + printObjectSectionSizes(o); if (OutputFormat == berkeley) { - if (!MachO || moreThanOneFile || ArchFlags.size() > 1) + if (!MachO || MoreThanOneFile || ArchFlags.size() > 1) outs() << o->getFileName() << " (for architecture " << I->getArchTypeName() << ")"; outs() << "\n"; } } - } else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = + } else if (auto E = isNotObjectErrorInvalidFileType( + UO.takeError())) { + error(std::move(E), file, ArchFlags.size() > 1 ? + StringRef(I->getArchTypeName()) : StringRef()); + return; + } else if (Expected<std::unique_ptr<Archive>> AOrErr = I->getAsArchive()) { std::unique_ptr<Archive> &UA = *AOrErr; // This is an archive. Iterate over each member and display its // sizes. - for (object::Archive::child_iterator i = UA->child_begin(), - e = UA->child_end(); - i != e; ++i) { - if (std::error_code EC = i->getError()) { - errs() << ToolName << ": " << file << ": " << EC.message() - << ".\n"; - exit(1); - } - auto &c = i->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = c.getAsBinary(); - if (std::error_code EC = ChildOrErr.getError()) { - errs() << ToolName << ": " << file << ": " << EC.message() - << ".\n"; + Error Err; + for (auto &C : UA->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType( + ChildOrErr.takeError())) + error(std::move(E), UA->getFileName(), C, + ArchFlags.size() > 1 ? + StringRef(I->getArchTypeName()) : StringRef()); continue; } if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) { @@ -518,7 +617,7 @@ static void PrintFileSectionSizes(StringRef file) { << ")" << " (for architecture " << I->getArchTypeName() << "):\n"; - PrintObjectSectionSizes(o); + printObjectSectionSizes(o); if (OutputFormat == berkeley) { if (MachO) { outs() << UA->getFileName() << "(" << o->getFileName() @@ -533,6 +632,13 @@ static void PrintFileSectionSizes(StringRef file) { } } } + if (Err) + error(std::move(Err), UA->getFileName()); + } else { + consumeError(AOrErr.takeError()); + error("Mach-O universal file: " + file + " for architecture " + + StringRef(I->getArchTypeName()) + + " is not a Mach-O file or an archive file"); } } } @@ -552,43 +658,40 @@ static void PrintFileSectionSizes(StringRef file) { E = UB->end_objects(); I != E; ++I) { if (HostArchName == I->getArchTypeName()) { - ErrorOr<std::unique_ptr<ObjectFile>> UO = I->getAsObjectFile(); + Expected<std::unique_ptr<ObjectFile>> UO = I->getAsObjectFile(); if (UO) { if (ObjectFile *o = dyn_cast<ObjectFile>(&*UO.get())) { MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o); if (OutputFormat == sysv) outs() << o->getFileName() << " :\n"; else if (MachO && OutputFormat == darwin) { - if (moreThanOneFile) + if (MoreThanOneFile) outs() << o->getFileName() << " (for architecture " << I->getArchTypeName() << "):\n"; } - PrintObjectSectionSizes(o); + printObjectSectionSizes(o); if (OutputFormat == berkeley) { - if (!MachO || moreThanOneFile) + if (!MachO || MoreThanOneFile) outs() << o->getFileName() << " (for architecture " << I->getArchTypeName() << ")"; outs() << "\n"; } } - } else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = + } else if (auto E = isNotObjectErrorInvalidFileType(UO.takeError())) { + error(std::move(E), file); + return; + } else if (Expected<std::unique_ptr<Archive>> AOrErr = I->getAsArchive()) { std::unique_ptr<Archive> &UA = *AOrErr; // This is an archive. Iterate over each member and display its // sizes. - for (object::Archive::child_iterator i = UA->child_begin(), - e = UA->child_end(); - i != e; ++i) { - if (std::error_code EC = i->getError()) { - errs() << ToolName << ": " << file << ": " << EC.message() - << ".\n"; - exit(1); - } - auto &c = i->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = c.getAsBinary(); - if (std::error_code EC = ChildOrErr.getError()) { - errs() << ToolName << ": " << file << ": " << EC.message() - << ".\n"; + Error Err; + for (auto &C : UA->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType( + ChildOrErr.takeError())) + error(std::move(E), UA->getFileName(), C); continue; } if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) { @@ -600,7 +703,7 @@ static void PrintFileSectionSizes(StringRef file) { outs() << UA->getFileName() << "(" << o->getFileName() << ")" << " (for architecture " << I->getArchTypeName() << "):\n"; - PrintObjectSectionSizes(o); + printObjectSectionSizes(o); if (OutputFormat == berkeley) { if (MachO) outs() << UA->getFileName() << "(" << o->getFileName() @@ -611,6 +714,13 @@ static void PrintFileSectionSizes(StringRef file) { } } } + if (Err) + error(std::move(Err), UA->getFileName()); + } else { + consumeError(AOrErr.takeError()); + error("Mach-O universal file: " + file + " for architecture " + + StringRef(I->getArchTypeName()) + + " is not a Mach-O file or an archive file"); } return; } @@ -618,45 +728,46 @@ static void PrintFileSectionSizes(StringRef file) { } // Either all architectures have been specified or none have been specified // and this does not contain the host architecture so dump all the slices. - bool moreThanOneArch = UB->getNumberOfObjects() > 1; + bool MoreThanOneArch = UB->getNumberOfObjects() > 1; for (MachOUniversalBinary::object_iterator I = UB->begin_objects(), E = UB->end_objects(); I != E; ++I) { - ErrorOr<std::unique_ptr<ObjectFile>> UO = I->getAsObjectFile(); + Expected<std::unique_ptr<ObjectFile>> UO = I->getAsObjectFile(); if (UO) { if (ObjectFile *o = dyn_cast<ObjectFile>(&*UO.get())) { MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o); if (OutputFormat == sysv) outs() << o->getFileName() << " :\n"; else if (MachO && OutputFormat == darwin) { - if (moreThanOneFile || moreThanOneArch) + if (MoreThanOneFile || MoreThanOneArch) outs() << o->getFileName() << " (for architecture " << I->getArchTypeName() << "):"; outs() << "\n"; } - PrintObjectSectionSizes(o); + printObjectSectionSizes(o); if (OutputFormat == berkeley) { - if (!MachO || moreThanOneFile || moreThanOneArch) + if (!MachO || MoreThanOneFile || MoreThanOneArch) outs() << o->getFileName() << " (for architecture " << I->getArchTypeName() << ")"; outs() << "\n"; } } - } else if (ErrorOr<std::unique_ptr<Archive>> AOrErr = + } else if (auto E = isNotObjectErrorInvalidFileType(UO.takeError())) { + error(std::move(E), file, MoreThanOneArch ? + StringRef(I->getArchTypeName()) : StringRef()); + return; + } else if (Expected<std::unique_ptr<Archive>> AOrErr = I->getAsArchive()) { std::unique_ptr<Archive> &UA = *AOrErr; // This is an archive. Iterate over each member and display its sizes. - for (object::Archive::child_iterator i = UA->child_begin(), - e = UA->child_end(); - i != e; ++i) { - if (std::error_code EC = i->getError()) { - errs() << ToolName << ": " << file << ": " << EC.message() << ".\n"; - exit(1); - } - auto &c = i->get(); - ErrorOr<std::unique_ptr<Binary>> ChildOrErr = c.getAsBinary(); - if (std::error_code EC = ChildOrErr.getError()) { - errs() << ToolName << ": " << file << ": " << EC.message() << ".\n"; + Error Err; + for (auto &C : UA->children(Err)) { + Expected<std::unique_ptr<Binary>> ChildOrErr = C.getAsBinary(); + if (!ChildOrErr) { + if (auto E = isNotObjectErrorInvalidFileType( + ChildOrErr.takeError())) + error(std::move(E), UA->getFileName(), C, MoreThanOneArch ? + StringRef(I->getArchTypeName()) : StringRef()); continue; } if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) { @@ -667,7 +778,7 @@ static void PrintFileSectionSizes(StringRef file) { else if (MachO && OutputFormat == darwin) outs() << UA->getFileName() << "(" << o->getFileName() << ")" << " (for architecture " << I->getArchTypeName() << "):\n"; - PrintObjectSectionSizes(o); + printObjectSectionSizes(o); if (OutputFormat == berkeley) { if (MachO) outs() << UA->getFileName() << "(" << o->getFileName() << ")" @@ -679,6 +790,13 @@ static void PrintFileSectionSizes(StringRef file) { } } } + if (Err) + error(std::move(Err), UA->getFileName()); + } else { + consumeError(AOrErr.takeError()); + error("Mach-O universal file: " + file + " for architecture " + + StringRef(I->getArchTypeName()) + + " is not a Mach-O file or an archive file"); } } } else if (ObjectFile *o = dyn_cast<ObjectFile>(&Bin)) { @@ -686,10 +804,10 @@ static void PrintFileSectionSizes(StringRef file) { return; if (OutputFormat == sysv) outs() << o->getFileName() << " :\n"; - PrintObjectSectionSizes(o); + printObjectSectionSizes(o); if (OutputFormat == berkeley) { MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o); - if (!MachO || moreThanOneFile) + if (!MachO || MoreThanOneFile) outs() << o->getFileName(); outs() << "\n"; } @@ -704,7 +822,7 @@ static void PrintFileSectionSizes(StringRef file) { int main(int argc, char **argv) { // Print a stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. @@ -731,9 +849,10 @@ int main(int argc, char **argv) { if (InputFilenames.size() == 0) InputFilenames.push_back("a.out"); - moreThanOneFile = InputFilenames.size() > 1; + MoreThanOneFile = InputFilenames.size() > 1; std::for_each(InputFilenames.begin(), InputFilenames.end(), - PrintFileSectionSizes); + printFileSectionSizes); - return 0; + if (HadError) + return 1; } diff --git a/tools/llvm-split/Makefile b/tools/llvm-split/Makefile deleted file mode 100644 index ef1243dbf1ffa..0000000000000 --- a/tools/llvm-split/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-split/Makefile ---------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-split -LINK_COMPONENTS := transformutils bitwriter core irreader support - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-split/llvm-split.cpp b/tools/llvm-split/llvm-split.cpp index 059770fbf4ac3..024363547f937 100644 --- a/tools/llvm-split/llvm-split.cpp +++ b/tools/llvm-split/llvm-split.cpp @@ -14,6 +14,7 @@ #include "llvm/ADT/StringExtras.h" #include "llvm/Bitcode/ReaderWriter.h" #include "llvm/IR/LLVMContext.h" +#include "llvm/IR/Verifier.h" #include "llvm/IRReader/IRReader.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/FileSystem.h" @@ -35,8 +36,12 @@ OutputFilename("o", cl::desc("Override output filename"), static cl::opt<unsigned> NumOutputs("j", cl::Prefix, cl::init(2), cl::desc("Number of output files")); +static cl::opt<bool> + PreserveLocals("preserve-locals", cl::Prefix, cl::init(false), + cl::desc("Split without externalizing locals")); + int main(int argc, char **argv) { - LLVMContext &Context = getGlobalContext(); + LLVMContext Context; SMDiagnostic Err; cl::ParseCommandLineOptions(argc, argv, "LLVM module splitter\n"); @@ -57,11 +62,12 @@ int main(int argc, char **argv) { exit(1); } + verifyModule(*MPart); WriteBitcodeToFile(MPart.get(), Out->os()); // Declare success. Out->keep(); - }); + }, PreserveLocals); return 0; } diff --git a/tools/llvm-stress/Makefile b/tools/llvm-stress/Makefile deleted file mode 100644 index 29245af7298cb..0000000000000 --- a/tools/llvm-stress/Makefile +++ /dev/null @@ -1,15 +0,0 @@ -##===- tools/llvm-stress/Makefile --------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-stress -LINK_COMPONENTS := object -LINK_COMPONENTS := bitreader bitwriter asmparser irreader instrumentation scalaropts ipo - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-stress/llvm-stress.cpp b/tools/llvm-stress/llvm-stress.cpp index 99d2afdcd3012..0b887ea9b4cd4 100644 --- a/tools/llvm-stress/llvm-stress.cpp +++ b/tools/llvm-stress/llvm-stress.cpp @@ -17,10 +17,10 @@ #include "llvm/IR/IRPrintingPasses.h" #include "llvm/IR/Instruction.h" #include "llvm/IR/LLVMContext.h" +#include "llvm/IR/LegacyPassManager.h" #include "llvm/IR/LegacyPassNameParser.h" #include "llvm/IR/Module.h" #include "llvm/IR/Verifier.h" -#include "llvm/IR/LegacyPassManager.h" #include "llvm/Support/Debug.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/ManagedStatic.h" @@ -28,8 +28,6 @@ #include "llvm/Support/PrettyStackTrace.h" #include "llvm/Support/ToolOutputFile.h" #include <algorithm> -#include <set> -#include <sstream> #include <vector> namespace llvm { @@ -43,6 +41,8 @@ static cl::opt<std::string> OutputFilename("o", cl::desc("Override output filename"), cl::value_desc("filename")); +static LLVMContext Context; + namespace cl { template <> class parser<Type*> final : public basic_parser<Type*> { public: @@ -50,7 +50,6 @@ public: // Parse options as IR types. Return true on error. bool parse(Option &O, StringRef, StringRef Arg, Type *&Value) { - auto &Context = getGlobalContext(); if (Arg == "half") Value = Type::getHalfTy(Context); else if (Arg == "fp128") Value = Type::getFP128Ty(Context); else if (Arg == "x86_fp80") Value = Type::getX86_FP80Ty(Context); @@ -687,7 +686,7 @@ int main(int argc, char **argv) { cl::ParseCommandLineOptions(argc, argv, "llvm codegen stress-tester\n"); llvm_shutdown_obj Y; - auto M = make_unique<Module>("/tmp/autogen.bc", getGlobalContext()); + auto M = make_unique<Module>("/tmp/autogen.bc", Context); Function *F = GenEmptyFunction(M.get()); // Pick an initial seed value diff --git a/tools/llvm-symbolizer/Makefile b/tools/llvm-symbolizer/Makefile deleted file mode 100644 index 8272d61280c75..0000000000000 --- a/tools/llvm-symbolizer/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/llvm-symbolizer/Makefile ----------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := llvm-symbolizer -LINK_COMPONENTS := DebugInfoDWARF DebugInfoPDB Object Support Symbolize - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/llvm-symbolizer/llvm-symbolizer.cpp b/tools/llvm-symbolizer/llvm-symbolizer.cpp index 950349377bf79..7ee4ba19603b7 100644 --- a/tools/llvm-symbolizer/llvm-symbolizer.cpp +++ b/tools/llvm-symbolizer/llvm-symbolizer.cpp @@ -86,10 +86,12 @@ static cl::opt<int> ClPrintSourceContextLines( "print-source-context-lines", cl::init(0), cl::desc("Print N number of source file context")); -static bool error(std::error_code ec) { - if (!ec) +template<typename T> +static bool error(Expected<T> &ResOrErr) { + if (ResOrErr) return false; - errs() << "LLVMSymbolizer: error reading file: " << ec.message() << ".\n"; + logAllUnhandledErrors(ResOrErr.takeError(), errs(), + "LLVMSymbolizer: error reading file: "); return true; } @@ -138,7 +140,7 @@ static bool parseCommand(StringRef InputString, bool &IsData, int main(int argc, char **argv) { // Print stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. @@ -185,14 +187,14 @@ int main(int argc, char **argv) { } if (IsData) { auto ResOrErr = Symbolizer.symbolizeData(ModuleName, ModuleOffset); - Printer << (error(ResOrErr.getError()) ? DIGlobal() : ResOrErr.get()); + Printer << (error(ResOrErr) ? DIGlobal() : ResOrErr.get()); } else if (ClPrintInlining) { auto ResOrErr = Symbolizer.symbolizeInlinedCode(ModuleName, ModuleOffset); - Printer << (error(ResOrErr.getError()) ? DIInliningInfo() + Printer << (error(ResOrErr) ? DIInliningInfo() : ResOrErr.get()); } else { auto ResOrErr = Symbolizer.symbolizeCode(ModuleName, ModuleOffset); - Printer << (error(ResOrErr.getError()) ? DILineInfo() : ResOrErr.get()); + Printer << (error(ResOrErr) ? DILineInfo() : ResOrErr.get()); } outs() << "\n"; outs().flush(); diff --git a/tools/lto/CMakeLists.txt b/tools/lto/CMakeLists.txt index 62447eec389bc..8a1571ba03f12 100644 --- a/tools/lto/CMakeLists.txt +++ b/tools/lto/CMakeLists.txt @@ -1,5 +1,6 @@ set(LLVM_LINK_COMPONENTS ${LLVM_TARGETS_TO_BUILD} + BitReader Core LTO MC @@ -18,7 +19,8 @@ set(LLVM_EXPORTED_SYMBOL_FILE ${CMAKE_CURRENT_SOURCE_DIR}/lto.exports) add_llvm_library(LTO SHARED ${SOURCES}) install(FILES ${LLVM_MAIN_INCLUDE_DIR}/llvm-c/lto.h - DESTINATION include/llvm-c) + DESTINATION include/llvm-c + COMPONENT LTO) if (APPLE) set(LTO_VERSION ${LLVM_VERSION_MAJOR}) diff --git a/tools/lto/Makefile b/tools/lto/Makefile deleted file mode 100644 index 530c05a47318b..0000000000000 --- a/tools/lto/Makefile +++ /dev/null @@ -1,42 +0,0 @@ -##===- tools/lto/Makefile ----------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -LIBRARYNAME := LTO -LINK_COMPONENTS := all-targets core lto mc mcdisassembler support -LINK_LIBS_IN_SHARED := 1 -SHARED_LIBRARY := 1 - -EXPORTED_SYMBOL_FILE = $(PROJ_SRC_DIR)/lto.exports - -include $(LEVEL)/Makefile.common - -ifeq ($(HOST_OS),Darwin) - # Special hack to allow libLTO to have an offset version number. - ifdef LLVM_LTO_VERSION_OFFSET - LTO_LIBRARY_VERSION := $(shell expr $(LLVM_SUBMIT_VERSION) + \ - $(LLVM_LTO_VERSION_OFFSET)) - else - LTO_LIBRARY_VERSION := $(LLVM_SUBMIT_VERSION) - endif - - # set dylib internal version number to llvmCore submission number - ifdef LLVM_SUBMIT_VERSION - LLVMLibsOptions := $(LLVMLibsOptions) -Wl,-current_version \ - -Wl,$(LTO_LIBRARY_VERSION).$(LLVM_SUBMIT_SUBVERSION) \ - -Wl,-compatibility_version -Wl,1 - endif - - # If we're doing an Apple-style build, add the LTO object path. - ifeq ($(RC_XBS),YES) - TempFile := $(shell mkdir -p ${OBJROOT}/dSYMs ; mktemp ${OBJROOT}/dSYMs/llvm-lto.XXXXXX) - LLVMLibsOptions := $(LLVMLibsOptions) \ - -Wl,-object_path_lto -Wl,$(TempFile) - endif -endif diff --git a/tools/lto/lto.cpp b/tools/lto/lto.cpp index d8f99c050a34c..7109cf4d90970 100644 --- a/tools/lto/lto.cpp +++ b/tools/lto/lto.cpp @@ -14,15 +14,18 @@ #include "llvm-c/lto.h" #include "llvm/ADT/STLExtras.h" +#include "llvm/Bitcode/ReaderWriter.h" #include "llvm/CodeGen/CommandFlags.h" #include "llvm/IR/DiagnosticInfo.h" #include "llvm/IR/DiagnosticPrinter.h" #include "llvm/IR/LLVMContext.h" -#include "llvm/LTO/LTOCodeGenerator.h" -#include "llvm/LTO/LTOModule.h" +#include "llvm/LTO/legacy/LTOCodeGenerator.h" +#include "llvm/LTO/legacy/LTOModule.h" +#include "llvm/LTO/legacy/ThinLTOCodeGenerator.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/Signals.h" #include "llvm/Support/TargetSelect.h" +#include "llvm/Support/raw_ostream.h" // extra command-line flags needed for LTOCodeGenerator static cl::opt<char> @@ -81,7 +84,6 @@ static void diagnosticHandler(const DiagnosticInfo &DI, void *Context) { DiagnosticPrinterRawOStream DP(Stream); DI.print(DP); } - sLastErrorString += '\n'; } // Initialize the configured targets if they have not been initialized. @@ -100,7 +102,8 @@ static void lto_initialize() { InitializeAllAsmPrinters(); InitializeAllDisassemblers(); - LTOContext = &getGlobalContext(); + static LLVMContext Context; + LTOContext = &Context; LTOContext->setDiagnosticHandler(diagnosticHandler, nullptr, true); initialized = true; } @@ -111,23 +114,24 @@ namespace { static void handleLibLTODiagnostic(lto_codegen_diagnostic_severity_t Severity, const char *Msg, void *) { sLastErrorString = Msg; - sLastErrorString += "\n"; } // This derived class owns the native object file. This helps implement the // libLTO API semantics, which require that the code generator owns the object // file. struct LibLTOCodeGenerator : LTOCodeGenerator { - LibLTOCodeGenerator() : LTOCodeGenerator(*LTOContext) { - setDiagnosticHandler(handleLibLTODiagnostic, nullptr); } + LibLTOCodeGenerator() : LTOCodeGenerator(*LTOContext) { init(); } LibLTOCodeGenerator(std::unique_ptr<LLVMContext> Context) : LTOCodeGenerator(*Context), OwnedContext(std::move(Context)) { - setDiagnosticHandler(handleLibLTODiagnostic, nullptr); } + init(); + } // Reset the module first in case MergedModule is created in OwnedContext. // Module must be destructed before its context gets destructed. ~LibLTOCodeGenerator() { resetMergedModule(); } + void init() { setDiagnosticHandler(handleLibLTODiagnostic, nullptr); } + std::unique_ptr<MemoryBuffer> NativeObjectFile; std::unique_ptr<LLVMContext> OwnedContext; }; @@ -135,6 +139,7 @@ struct LibLTOCodeGenerator : LTOCodeGenerator { } DEFINE_SIMPLE_CONVERSION_FUNCTIONS(LibLTOCodeGenerator, lto_code_gen_t) +DEFINE_SIMPLE_CONVERSION_FUNCTIONS(ThinLTOCodeGenerator, thinlto_code_gen_t) DEFINE_SIMPLE_CONVERSION_FUNCTIONS(LTOModule, lto_module_t) // Convert the subtarget features into a string to pass to LTOCodeGenerator. @@ -176,6 +181,14 @@ bool lto_module_is_object_file_for_target(const char* path, return LTOModule::isBitcodeForTarget(Buffer->get(), target_triplet_prefix); } +bool lto_module_has_objc_category(const void *mem, size_t length) { + std::unique_ptr<MemoryBuffer> Buffer(LTOModule::makeBuffer(mem, length)); + if (!Buffer) + return false; + LLVMContext Ctx; + return llvm::isBitcodeContainingObjCCategory(*Buffer, Ctx); +} + bool lto_module_is_object_file_in_memory(const void* mem, size_t length) { return LTOModule::isBitcodeFile(mem, length); } @@ -249,8 +262,14 @@ lto_module_t lto_module_create_in_local_context(const void *mem, size_t length, const char *path) { lto_initialize(); llvm::TargetOptions Options = InitTargetOptionsFromCodeGenFlags(); + + // Create a local context. Ownership will be transfered to LTOModule. + std::unique_ptr<LLVMContext> Context = llvm::make_unique<LLVMContext>(); + Context->setDiagnosticHandler(diagnosticHandler, nullptr, true); + ErrorOr<std::unique_ptr<LTOModule>> M = - LTOModule::createInLocalContext(mem, length, Options, path); + LTOModule::createInLocalContext(std::move(Context), mem, length, Options, + path); if (!M) return nullptr; return wrap(M->release()); @@ -262,8 +281,8 @@ lto_module_t lto_module_create_in_codegen_context(const void *mem, lto_code_gen_t cg) { lto_initialize(); llvm::TargetOptions Options = InitTargetOptionsFromCodeGenFlags(); - ErrorOr<std::unique_ptr<LTOModule>> M = LTOModule::createInContext( - mem, length, Options, path, &unwrap(cg)->getContext()); + ErrorOr<std::unique_ptr<LTOModule>> M = LTOModule::createFromBuffer( + unwrap(cg)->getContext(), mem, length, Options, path); return wrap(M->release()); } @@ -347,7 +366,7 @@ bool lto_codegen_set_pic_model(lto_code_gen_t cg, lto_codegen_model model) { unwrap(cg)->setCodePICModel(Reloc::DynamicNoPIC); return false; case LTO_CODEGEN_PIC_MODEL_DEFAULT: - unwrap(cg)->setCodePICModel(Reloc::Default); + unwrap(cg)->setCodePICModel(None); return false; } sLastErrorString = "Unknown PIC model"; @@ -435,3 +454,116 @@ void lto_codegen_set_should_embed_uselists(lto_code_gen_t cg, lto_bool_t ShouldEmbedUselists) { unwrap(cg)->setShouldEmbedUselists(ShouldEmbedUselists); } + +// ThinLTO API below + +thinlto_code_gen_t thinlto_create_codegen(void) { + lto_initialize(); + ThinLTOCodeGenerator *CodeGen = new ThinLTOCodeGenerator(); + CodeGen->setTargetOptions(InitTargetOptionsFromCodeGenFlags()); + + return wrap(CodeGen); +} + +void thinlto_codegen_dispose(thinlto_code_gen_t cg) { delete unwrap(cg); } + +void thinlto_codegen_add_module(thinlto_code_gen_t cg, const char *Identifier, + const char *Data, int Length) { + unwrap(cg)->addModule(Identifier, StringRef(Data, Length)); +} + +void thinlto_codegen_process(thinlto_code_gen_t cg) { unwrap(cg)->run(); } + +unsigned int thinlto_module_get_num_objects(thinlto_code_gen_t cg) { + return unwrap(cg)->getProducedBinaries().size(); +} +LTOObjectBuffer thinlto_module_get_object(thinlto_code_gen_t cg, + unsigned int index) { + assert(index < unwrap(cg)->getProducedBinaries().size() && "Index overflow"); + auto &MemBuffer = unwrap(cg)->getProducedBinaries()[index]; + return LTOObjectBuffer{MemBuffer->getBufferStart(), + MemBuffer->getBufferSize()}; +} + +void thinlto_codegen_disable_codegen(thinlto_code_gen_t cg, + lto_bool_t disable) { + unwrap(cg)->disableCodeGen(disable); +} + +void thinlto_codegen_set_codegen_only(thinlto_code_gen_t cg, + lto_bool_t CodeGenOnly) { + unwrap(cg)->setCodeGenOnly(CodeGenOnly); +} + +void thinlto_debug_options(const char *const *options, int number) { + // if options were requested, set them + if (number && options) { + std::vector<const char *> CodegenArgv(1, "libLTO"); + for (auto Arg : ArrayRef<const char *>(options, number)) + CodegenArgv.push_back(Arg); + cl::ParseCommandLineOptions(CodegenArgv.size(), CodegenArgv.data()); + } +} + +lto_bool_t lto_module_is_thinlto(lto_module_t mod) { + return unwrap(mod)->isThinLTO(); +} + +void thinlto_codegen_add_must_preserve_symbol(thinlto_code_gen_t cg, + const char *Name, int Length) { + unwrap(cg)->preserveSymbol(StringRef(Name, Length)); +} + +void thinlto_codegen_add_cross_referenced_symbol(thinlto_code_gen_t cg, + const char *Name, int Length) { + unwrap(cg)->crossReferenceSymbol(StringRef(Name, Length)); +} + +void thinlto_codegen_set_cpu(thinlto_code_gen_t cg, const char *cpu) { + return unwrap(cg)->setCpu(cpu); +} + +void thinlto_codegen_set_cache_dir(thinlto_code_gen_t cg, + const char *cache_dir) { + return unwrap(cg)->setCacheDir(cache_dir); +} + +void thinlto_codegen_set_cache_pruning_interval(thinlto_code_gen_t cg, + int interval) { + return unwrap(cg)->setCachePruningInterval(interval); +} + +void thinlto_codegen_set_cache_entry_expiration(thinlto_code_gen_t cg, + unsigned expiration) { + return unwrap(cg)->setCacheEntryExpiration(expiration); +} + +void thinlto_codegen_set_final_cache_size_relative_to_available_space( + thinlto_code_gen_t cg, unsigned Percentage) { + return unwrap(cg)->setMaxCacheSizeRelativeToAvailableSpace(Percentage); +} + +void thinlto_codegen_set_savetemps_dir(thinlto_code_gen_t cg, + const char *save_temps_dir) { + return unwrap(cg)->setSaveTempsDir(save_temps_dir); +} + +lto_bool_t thinlto_codegen_set_pic_model(thinlto_code_gen_t cg, + lto_codegen_model model) { + switch (model) { + case LTO_CODEGEN_PIC_MODEL_STATIC: + unwrap(cg)->setCodePICModel(Reloc::Static); + return false; + case LTO_CODEGEN_PIC_MODEL_DYNAMIC: + unwrap(cg)->setCodePICModel(Reloc::PIC_); + return false; + case LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC: + unwrap(cg)->setCodePICModel(Reloc::DynamicNoPIC); + return false; + case LTO_CODEGEN_PIC_MODEL_DEFAULT: + unwrap(cg)->setCodePICModel(None); + return false; + } + sLastErrorString = "Unknown PIC model"; + return true; +} diff --git a/tools/lto/lto.exports b/tools/lto/lto.exports index 8bc2b0f9d312f..74091c2641b77 100644 --- a/tools/lto/lto.exports +++ b/tools/lto/lto.exports @@ -18,6 +18,7 @@ lto_module_is_object_file lto_module_is_object_file_for_target lto_module_is_object_file_in_memory lto_module_is_object_file_in_memory_for_target +lto_module_has_objc_category lto_module_dispose lto_api_version lto_codegen_set_diagnostic_handler @@ -45,3 +46,22 @@ LLVMCreateDisasmCPU LLVMDisasmDispose LLVMDisasmInstruction LLVMSetDisasmOptions +thinlto_create_codegen +thinlto_codegen_dispose +thinlto_codegen_add_module +thinlto_codegen_process +thinlto_module_get_num_objects +thinlto_module_get_object +thinlto_codegen_set_pic_model +thinlto_codegen_set_cache_dir +thinlto_codegen_set_cache_pruning_interval +thinlto_codegen_set_cache_entry_expiration +thinlto_codegen_set_savetemps_dir +thinlto_codegen_set_cpu +thinlto_debug_options +lto_module_is_thinlto +thinlto_codegen_add_must_preserve_symbol +thinlto_codegen_add_cross_referenced_symbol +thinlto_codegen_set_final_cache_size_relative_to_available_space +thinlto_codegen_set_codegen_only +thinlto_codegen_disable_codegen
\ No newline at end of file diff --git a/tools/obj2yaml/CMakeLists.txt b/tools/obj2yaml/CMakeLists.txt index 3cdac5c748752..9b89552506025 100644 --- a/tools/obj2yaml/CMakeLists.txt +++ b/tools/obj2yaml/CMakeLists.txt @@ -1,8 +1,13 @@ set(LLVM_LINK_COMPONENTS Object + ObjectYAML Support ) add_llvm_tool(obj2yaml - obj2yaml.cpp coff2yaml.cpp elf2yaml.cpp Error.cpp + obj2yaml.cpp + coff2yaml.cpp + elf2yaml.cpp + macho2yaml.cpp + Error.cpp ) diff --git a/tools/obj2yaml/Error.cpp b/tools/obj2yaml/Error.cpp index abef8af58cbfe..9d1af680a7399 100644 --- a/tools/obj2yaml/Error.cpp +++ b/tools/obj2yaml/Error.cpp @@ -13,6 +13,9 @@ using namespace llvm; namespace { +// FIXME: This class is only here to support the transition to llvm::Error. It +// will be removed once this transition is complete. Clients should prefer to +// deal with the Error value directly, rather than converting to error_code. class _obj2yaml_error_category : public std::error_category { public: const char *name() const LLVM_NOEXCEPT override; @@ -34,14 +37,26 @@ std::string _obj2yaml_error_category::message(int ev) const { return "Unrecognized file type."; case obj2yaml_error::unsupported_obj_file_format: return "Unsupported object file format."; + case obj2yaml_error::not_implemented: + return "Feature not yet implemented."; } llvm_unreachable("An enumerator of obj2yaml_error does not have a message " "defined."); } namespace llvm { - const std::error_category &obj2yaml_category() { + +const std::error_category &obj2yaml_category() { static _obj2yaml_error_category o; return o; } + +char Obj2YamlError::ID = 0; + +void Obj2YamlError::log(raw_ostream &OS) const { OS << ErrMsg << "\n"; } + +std::error_code Obj2YamlError::convertToErrorCode() const { + return std::error_code(static_cast<int>(Code), obj2yaml_category()); +} + } // namespace llvm diff --git a/tools/obj2yaml/Error.h b/tools/obj2yaml/Error.h index 982f59e236cc0..f5111f257ce36 100644 --- a/tools/obj2yaml/Error.h +++ b/tools/obj2yaml/Error.h @@ -10,6 +10,8 @@ #ifndef LLVM_TOOLS_OBJ2YAML_ERROR_H #define LLVM_TOOLS_OBJ2YAML_ERROR_H +#include "llvm/Support/Error.h" + #include <system_error> namespace llvm { @@ -19,13 +21,30 @@ enum class obj2yaml_error { success = 0, file_not_found, unrecognized_file_format, - unsupported_obj_file_format + unsupported_obj_file_format, + not_implemented }; inline std::error_code make_error_code(obj2yaml_error e) { return std::error_code(static_cast<int>(e), obj2yaml_category()); } +class Obj2YamlError : public ErrorInfo<Obj2YamlError> { +public: + static char ID; + Obj2YamlError(obj2yaml_error C) : Code(C) {} + Obj2YamlError(std::string ErrMsg) : ErrMsg(std::move(ErrMsg)) {} + Obj2YamlError(obj2yaml_error C, std::string ErrMsg) + : ErrMsg(std::move(ErrMsg)), Code(C) {} + void log(raw_ostream &OS) const override; + const std::string &getErrorMessage() const { return ErrMsg; } + std::error_code convertToErrorCode() const override; + +private: + std::string ErrMsg; + obj2yaml_error Code; +}; + } // namespace llvm namespace std { diff --git a/tools/obj2yaml/Makefile b/tools/obj2yaml/Makefile deleted file mode 100644 index 6cbef6998dc6c..0000000000000 --- a/tools/obj2yaml/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- utils/obj2yaml/Makefile ----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL = ../.. -TOOLNAME = obj2yaml -LINK_COMPONENTS := object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS = 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/obj2yaml/coff2yaml.cpp b/tools/obj2yaml/coff2yaml.cpp index f675bfe4e618e..c734601ede76b 100644 --- a/tools/obj2yaml/coff2yaml.cpp +++ b/tools/obj2yaml/coff2yaml.cpp @@ -9,7 +9,7 @@ #include "obj2yaml.h" #include "llvm/Object/COFF.h" -#include "llvm/Object/COFFYAML.h" +#include "llvm/ObjectYAML/COFFYAML.h" #include "llvm/Support/ErrorHandling.h" #include "llvm/Support/YAMLTraits.h" @@ -109,6 +109,7 @@ void COFFDumper::dumpSections(unsigned NumSections) { NewYAMLSection.Header.VirtualAddress = ObjSection.getAddress(); NewYAMLSection.Header.VirtualSize = COFFSection->VirtualSize; NewYAMLSection.Alignment = ObjSection.getAlignment(); + assert(NewYAMLSection.Alignment <= 8192); ArrayRef<uint8_t> sectionData; if (!ObjSection.isBSS()) @@ -120,9 +121,14 @@ void COFFDumper::dumpSections(unsigned NumSections) { const object::coff_relocation *reloc = Obj.getCOFFRelocation(Reloc); COFFYAML::Relocation Rel; object::symbol_iterator Sym = Reloc.getSymbol(); - ErrorOr<StringRef> SymbolNameOrErr = Sym->getName(); - if (std::error_code EC = SymbolNameOrErr.getError()) - report_fatal_error(EC.message()); + Expected<StringRef> SymbolNameOrErr = Sym->getName(); + if (!SymbolNameOrErr) { + std::string Buf; + raw_string_ostream OS(Buf); + logAllUnhandledErrors(SymbolNameOrErr.takeError(), OS, ""); + OS.flush(); + report_fatal_error(Buf); + } Rel.SymbolName = *SymbolNameOrErr; Rel.VirtualAddress = reloc->VirtualAddress; Rel.Type = reloc->Type; diff --git a/tools/obj2yaml/elf2yaml.cpp b/tools/obj2yaml/elf2yaml.cpp index f2b0138073451..782832d54571f 100644 --- a/tools/obj2yaml/elf2yaml.cpp +++ b/tools/obj2yaml/elf2yaml.cpp @@ -11,7 +11,7 @@ #include "obj2yaml.h" #include "llvm/ADT/STLExtras.h" #include "llvm/Object/ELFObjectFile.h" -#include "llvm/Object/ELFYAML.h" +#include "llvm/ObjectYAML/ELFYAML.h" #include "llvm/Support/ErrorHandling.h" #include "llvm/Support/YAMLTraits.h" @@ -179,10 +179,10 @@ ELFDumper<ELFT>::dumpSymbol(const Elf_Sym *Sym, const Elf_Shdr *SymTab, S.Size = Sym->st_size; S.Other = Sym->st_other; - ErrorOr<StringRef> NameOrErr = Sym->getName(StrTable); - if (std::error_code EC = NameOrErr.getError()) - return EC; - S.Name = NameOrErr.get(); + Expected<StringRef> SymbolNameOrErr = Sym->getName(StrTable); + if (!SymbolNameOrErr) + return errorToErrorCode(SymbolNameOrErr.takeError()); + S.Name = SymbolNameOrErr.get(); ErrorOr<const Elf_Shdr *> ShdrOrErr = Obj.getSection(Sym, SymTab, ShndxTable); if (std::error_code EC = ShdrOrErr.getError()) @@ -191,7 +191,7 @@ ELFDumper<ELFT>::dumpSymbol(const Elf_Sym *Sym, const Elf_Shdr *SymTab, if (!Shdr) return obj2yaml_error::success; - NameOrErr = Obj.getSectionName(Shdr); + ErrorOr<StringRef> NameOrErr = Obj.getSectionName(Shdr); if (std::error_code EC = NameOrErr.getError()) return EC; S.Section = NameOrErr.get(); @@ -217,9 +217,9 @@ std::error_code ELFDumper<ELFT>::dumpRelocation(const RelT *Rel, return EC; StringRef StrTab = *StrTabOrErr; - ErrorOr<StringRef> NameOrErr = Sym->getName(StrTab); - if (std::error_code EC = NameOrErr.getError()) - return EC; + Expected<StringRef> NameOrErr = Sym->getName(StrTab); + if (!NameOrErr) + return errorToErrorCode(NameOrErr.takeError()); R.Symbol = NameOrErr.get(); return obj2yaml_error::success; @@ -368,9 +368,9 @@ ErrorOr<ELFYAML::Group *> ELFDumper<ELFT>::dumpGroup(const Elf_Shdr *Shdr) { auto sectionContents = Obj.getSectionContents(Shdr); if (std::error_code ec = sectionContents.getError()) return ec; - ErrorOr<StringRef> symbolName = symbol->getName(StrTab); - if (std::error_code EC = symbolName.getError()) - return EC; + Expected<StringRef> symbolName = symbol->getName(StrTab); + if (!symbolName) + return errorToErrorCode(symbolName.takeError()); S->Info = *symbolName; const Elf_Word *groupMembers = reinterpret_cast<const Elf_Word *>(sectionContents->data()); diff --git a/tools/obj2yaml/macho2yaml.cpp b/tools/obj2yaml/macho2yaml.cpp new file mode 100644 index 0000000000000..c9a1385f17300 --- /dev/null +++ b/tools/obj2yaml/macho2yaml.cpp @@ -0,0 +1,527 @@ +//===------ macho2yaml.cpp - obj2yaml conversion tool -----------*- C++ -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "Error.h" +#include "obj2yaml.h" +#include "llvm/Object/MachOUniversal.h" +#include "llvm/ObjectYAML/ObjectYAML.h" +#include "llvm/Support/ErrorHandling.h" +#include "llvm/Support/LEB128.h" + +#include <string.h> // for memcpy + +using namespace llvm; + +class MachODumper { + + template <typename StructType> + const char *processLoadCommandData( + MachOYAML::LoadCommand &LC, + const llvm::object::MachOObjectFile::LoadCommandInfo &LoadCmd); + + const object::MachOObjectFile &Obj; + void dumpHeader(std::unique_ptr<MachOYAML::Object> &Y); + void dumpLoadCommands(std::unique_ptr<MachOYAML::Object> &Y); + void dumpLinkEdit(std::unique_ptr<MachOYAML::Object> &Y); + void dumpRebaseOpcodes(std::unique_ptr<MachOYAML::Object> &Y); + void dumpBindOpcodes(std::vector<MachOYAML::BindOpcode> &BindOpcodes, + ArrayRef<uint8_t> OpcodeBuffer, bool Lazy = false); + void dumpExportTrie(std::unique_ptr<MachOYAML::Object> &Y); + void dumpSymbols(std::unique_ptr<MachOYAML::Object> &Y); + +public: + MachODumper(const object::MachOObjectFile &O) : Obj(O) {} + Expected<std::unique_ptr<MachOYAML::Object>> dump(); +}; + +#define HANDLE_LOAD_COMMAND(LCName, LCValue, LCStruct) \ + case MachO::LCName: \ + memcpy((void *) & (LC.Data.LCStruct##_data), LoadCmd.Ptr, \ + sizeof(MachO::LCStruct)); \ + if (Obj.isLittleEndian() != sys::IsLittleEndianHost) \ + MachO::swapStruct(LC.Data.LCStruct##_data); \ + EndPtr = processLoadCommandData<MachO::LCStruct>(LC, LoadCmd); \ + break; + +template <typename SectionType> +MachOYAML::Section constructSectionCommon(SectionType Sec) { + MachOYAML::Section TempSec; + memcpy(reinterpret_cast<void *>(&TempSec.sectname[0]), &Sec.sectname[0], 16); + memcpy(reinterpret_cast<void *>(&TempSec.segname[0]), &Sec.segname[0], 16); + TempSec.addr = Sec.addr; + TempSec.size = Sec.size; + TempSec.offset = Sec.offset; + TempSec.align = Sec.align; + TempSec.reloff = Sec.reloff; + TempSec.nreloc = Sec.nreloc; + TempSec.flags = Sec.flags; + TempSec.reserved1 = Sec.reserved1; + TempSec.reserved2 = Sec.reserved2; + TempSec.reserved3 = 0; + return TempSec; +} + +template <typename SectionType> +MachOYAML::Section constructSection(SectionType Sec); + +template <> MachOYAML::Section constructSection(MachO::section Sec) { + MachOYAML::Section TempSec = constructSectionCommon(Sec); + TempSec.reserved3 = 0; + return TempSec; +} + +template <> MachOYAML::Section constructSection(MachO::section_64 Sec) { + MachOYAML::Section TempSec = constructSectionCommon(Sec); + TempSec.reserved3 = Sec.reserved3; + return TempSec; +} + +template <typename SectionType, typename SegmentType> +const char * +extractSections(const llvm::object::MachOObjectFile::LoadCommandInfo &LoadCmd, + std::vector<MachOYAML::Section> &Sections, + bool IsLittleEndian) { + auto End = LoadCmd.Ptr + LoadCmd.C.cmdsize; + const SectionType *Curr = + reinterpret_cast<const SectionType *>(LoadCmd.Ptr + sizeof(SegmentType)); + for (; reinterpret_cast<const void *>(Curr) < End; Curr++) { + if (IsLittleEndian != sys::IsLittleEndianHost) { + SectionType Sec; + memcpy((void *)&Sec, Curr, sizeof(SectionType)); + MachO::swapStruct(Sec); + Sections.push_back(constructSection(Sec)); + } else { + Sections.push_back(constructSection(*Curr)); + } + } + return reinterpret_cast<const char *>(Curr); +} + +template <typename StructType> +const char *MachODumper::processLoadCommandData( + MachOYAML::LoadCommand &LC, + const llvm::object::MachOObjectFile::LoadCommandInfo &LoadCmd) { + return LoadCmd.Ptr + sizeof(StructType); +} + +template <> +const char *MachODumper::processLoadCommandData<MachO::segment_command>( + MachOYAML::LoadCommand &LC, + const llvm::object::MachOObjectFile::LoadCommandInfo &LoadCmd) { + return extractSections<MachO::section, MachO::segment_command>( + LoadCmd, LC.Sections, Obj.isLittleEndian()); +} + +template <> +const char *MachODumper::processLoadCommandData<MachO::segment_command_64>( + MachOYAML::LoadCommand &LC, + const llvm::object::MachOObjectFile::LoadCommandInfo &LoadCmd) { + return extractSections<MachO::section_64, MachO::segment_command_64>( + LoadCmd, LC.Sections, Obj.isLittleEndian()); +} + +template <typename StructType> +const char * +readString(MachOYAML::LoadCommand &LC, + const llvm::object::MachOObjectFile::LoadCommandInfo &LoadCmd) { + auto Start = LoadCmd.Ptr + sizeof(StructType); + auto MaxSize = LoadCmd.C.cmdsize - sizeof(StructType); + auto Size = strnlen(Start, MaxSize); + LC.PayloadString = StringRef(Start, Size).str(); + return Start + Size; +} + +template <> +const char *MachODumper::processLoadCommandData<MachO::dylib_command>( + MachOYAML::LoadCommand &LC, + const llvm::object::MachOObjectFile::LoadCommandInfo &LoadCmd) { + return readString<MachO::dylib_command>(LC, LoadCmd); +} + +template <> +const char *MachODumper::processLoadCommandData<MachO::dylinker_command>( + MachOYAML::LoadCommand &LC, + const llvm::object::MachOObjectFile::LoadCommandInfo &LoadCmd) { + return readString<MachO::dylinker_command>(LC, LoadCmd); +} + +template <> +const char *MachODumper::processLoadCommandData<MachO::rpath_command>( + MachOYAML::LoadCommand &LC, + const llvm::object::MachOObjectFile::LoadCommandInfo &LoadCmd) { + return readString<MachO::rpath_command>(LC, LoadCmd); +} + +Expected<std::unique_ptr<MachOYAML::Object>> MachODumper::dump() { + auto Y = make_unique<MachOYAML::Object>(); + dumpHeader(Y); + dumpLoadCommands(Y); + dumpLinkEdit(Y); + return std::move(Y); +} + +void MachODumper::dumpHeader(std::unique_ptr<MachOYAML::Object> &Y) { + Y->Header.magic = Obj.getHeader().magic; + Y->Header.cputype = Obj.getHeader().cputype; + Y->Header.cpusubtype = Obj.getHeader().cpusubtype; + Y->Header.filetype = Obj.getHeader().filetype; + Y->Header.ncmds = Obj.getHeader().ncmds; + Y->Header.sizeofcmds = Obj.getHeader().sizeofcmds; + Y->Header.flags = Obj.getHeader().flags; + Y->Header.reserved = 0; +} + +void MachODumper::dumpLoadCommands(std::unique_ptr<MachOYAML::Object> &Y) { + for (auto LoadCmd : Obj.load_commands()) { + MachOYAML::LoadCommand LC; + const char *EndPtr = LoadCmd.Ptr; + switch (LoadCmd.C.cmd) { + default: + memcpy((void *)&(LC.Data.load_command_data), LoadCmd.Ptr, + sizeof(MachO::load_command)); + if (Obj.isLittleEndian() != sys::IsLittleEndianHost) + MachO::swapStruct(LC.Data.load_command_data); + EndPtr = processLoadCommandData<MachO::load_command>(LC, LoadCmd); + break; +#include "llvm/Support/MachO.def" + } + auto RemainingBytes = LoadCmd.C.cmdsize - (EndPtr - LoadCmd.Ptr); + if (!std::all_of(EndPtr, &EndPtr[RemainingBytes], + [](const char C) { return C == 0; })) { + LC.PayloadBytes.insert(LC.PayloadBytes.end(), EndPtr, + &EndPtr[RemainingBytes]); + RemainingBytes = 0; + } + LC.ZeroPadBytes = RemainingBytes; + Y->LoadCommands.push_back(std::move(LC)); + } +} + +void MachODumper::dumpLinkEdit(std::unique_ptr<MachOYAML::Object> &Y) { + dumpRebaseOpcodes(Y); + dumpBindOpcodes(Y->LinkEdit.BindOpcodes, Obj.getDyldInfoBindOpcodes()); + dumpBindOpcodes(Y->LinkEdit.WeakBindOpcodes, + Obj.getDyldInfoWeakBindOpcodes()); + dumpBindOpcodes(Y->LinkEdit.LazyBindOpcodes, Obj.getDyldInfoLazyBindOpcodes(), + true); + dumpExportTrie(Y); + dumpSymbols(Y); +} + +void MachODumper::dumpRebaseOpcodes(std::unique_ptr<MachOYAML::Object> &Y) { + MachOYAML::LinkEditData &LEData = Y->LinkEdit; + + auto RebaseOpcodes = Obj.getDyldInfoRebaseOpcodes(); + for (auto OpCode = RebaseOpcodes.begin(); OpCode != RebaseOpcodes.end(); + ++OpCode) { + MachOYAML::RebaseOpcode RebaseOp; + RebaseOp.Opcode = + static_cast<MachO::RebaseOpcode>(*OpCode & MachO::REBASE_OPCODE_MASK); + RebaseOp.Imm = *OpCode & MachO::REBASE_IMMEDIATE_MASK; + + unsigned Count; + uint64_t ULEB = 0; + + switch (RebaseOp.Opcode) { + case MachO::REBASE_OPCODE_DO_REBASE_ULEB_TIMES_SKIPPING_ULEB: + + ULEB = decodeULEB128(OpCode + 1, &Count); + RebaseOp.ExtraData.push_back(ULEB); + OpCode += Count; + // Intentionally no break here -- This opcode has two ULEB values + case MachO::REBASE_OPCODE_SET_SEGMENT_AND_OFFSET_ULEB: + case MachO::REBASE_OPCODE_ADD_ADDR_ULEB: + case MachO::REBASE_OPCODE_DO_REBASE_ULEB_TIMES: + case MachO::REBASE_OPCODE_DO_REBASE_ADD_ADDR_ULEB: + + ULEB = decodeULEB128(OpCode + 1, &Count); + RebaseOp.ExtraData.push_back(ULEB); + OpCode += Count; + break; + default: + break; + } + + LEData.RebaseOpcodes.push_back(RebaseOp); + + if (RebaseOp.Opcode == MachO::REBASE_OPCODE_DONE) + break; + } +} + +StringRef ReadStringRef(const uint8_t *Start) { + const uint8_t *Itr = Start; + for (; *Itr; ++Itr) + ; + return StringRef(reinterpret_cast<const char *>(Start), Itr - Start); +} + +void MachODumper::dumpBindOpcodes( + std::vector<MachOYAML::BindOpcode> &BindOpcodes, + ArrayRef<uint8_t> OpcodeBuffer, bool Lazy) { + for (auto OpCode = OpcodeBuffer.begin(); OpCode != OpcodeBuffer.end(); + ++OpCode) { + MachOYAML::BindOpcode BindOp; + BindOp.Opcode = + static_cast<MachO::BindOpcode>(*OpCode & MachO::BIND_OPCODE_MASK); + BindOp.Imm = *OpCode & MachO::BIND_IMMEDIATE_MASK; + + unsigned Count; + uint64_t ULEB = 0; + int64_t SLEB = 0; + + switch (BindOp.Opcode) { + case MachO::BIND_OPCODE_DO_BIND_ULEB_TIMES_SKIPPING_ULEB: + ULEB = decodeULEB128(OpCode + 1, &Count); + BindOp.ULEBExtraData.push_back(ULEB); + OpCode += Count; + // Intentionally no break here -- this opcode has two ULEB values + + case MachO::BIND_OPCODE_SET_DYLIB_ORDINAL_ULEB: + case MachO::BIND_OPCODE_SET_SEGMENT_AND_OFFSET_ULEB: + case MachO::BIND_OPCODE_ADD_ADDR_ULEB: + case MachO::BIND_OPCODE_DO_BIND_ADD_ADDR_ULEB: + ULEB = decodeULEB128(OpCode + 1, &Count); + BindOp.ULEBExtraData.push_back(ULEB); + OpCode += Count; + break; + + case MachO::BIND_OPCODE_SET_ADDEND_SLEB: + SLEB = decodeSLEB128(OpCode + 1, &Count); + BindOp.SLEBExtraData.push_back(SLEB); + OpCode += Count; + break; + + case MachO::BIND_OPCODE_SET_SYMBOL_TRAILING_FLAGS_IMM: + BindOp.Symbol = ReadStringRef(OpCode + 1); + OpCode += BindOp.Symbol.size() + 1; + break; + default: + break; + } + + BindOpcodes.push_back(BindOp); + + // Lazy bindings have DONE opcodes between operations, so we need to keep + // processing after a DONE. + if (!Lazy && BindOp.Opcode == MachO::BIND_OPCODE_DONE) + break; + } +} + +/*! + * /brief processes a node from the export trie, and its children. + * + * To my knowledge there is no documentation of the encoded format of this data + * other than in the heads of the Apple linker engineers. To that end hopefully + * this comment and the implementation below can serve to light the way for + * anyone crazy enough to come down this path in the future. + * + * This function reads and preserves the trie structure of the export trie. To + * my knowledge there is no code anywhere else that reads the data and preserves + * the Trie. LD64 (sources available at opensource.apple.com) has a similar + * implementation that parses the export trie into a vector. That code as well + * as LLVM's libObject MachO implementation were the basis for this. + * + * The export trie is an encoded trie. The node serialization is a bit awkward. + * The below pseudo-code is the best description I've come up with for it. + * + * struct SerializedNode { + * ULEB128 TerminalSize; + * struct TerminalData { <-- This is only present if TerminalSize > 0 + * ULEB128 Flags; + * ULEB128 Address; <-- Present if (! Flags & REEXPORT ) + * ULEB128 Other; <-- Present if ( Flags & REEXPORT || + * Flags & STUB_AND_RESOLVER ) + * char[] ImportName; <-- Present if ( Flags & REEXPORT ) + * } + * uint8_t ChildrenCount; + * Pair<char[], ULEB128> ChildNameOffsetPair[ChildrenCount]; + * SerializedNode Children[ChildrenCount] + * } + * + * Terminal nodes are nodes that represent actual exports. They can appear + * anywhere in the tree other than at the root; they do not need to be leaf + * nodes. When reading the data out of the trie this routine reads it in-order, + * but it puts the child names and offsets directly into the child nodes. This + * results in looping over the children twice during serialization and + * de-serialization, but it makes the YAML representation more human readable. + * + * Below is an example of the graph from a "Hello World" executable: + * + * ------- + * | '' | + * ------- + * | + * ------- + * | '_' | + * ------- + * | + * |----------------------------------------| + * | | + * ------------------------ --------------------- + * | '_mh_execute_header' | | 'main' | + * | Flags: 0x00000000 | | Flags: 0x00000000 | + * | Addr: 0x00000000 | | Addr: 0x00001160 | + * ------------------------ --------------------- + * + * This graph represents the trie for the exports "__mh_execute_header" and + * "_main". In the graph only the "_main" and "__mh_execute_header" nodes are + * terminal. +*/ + +const uint8_t *processExportNode(const uint8_t *CurrPtr, + const uint8_t *const End, + MachOYAML::ExportEntry &Entry) { + if (CurrPtr >= End) + return CurrPtr; + unsigned Count = 0; + Entry.TerminalSize = decodeULEB128(CurrPtr, &Count); + CurrPtr += Count; + if (Entry.TerminalSize != 0) { + Entry.Flags = decodeULEB128(CurrPtr, &Count); + CurrPtr += Count; + if (Entry.Flags & MachO::EXPORT_SYMBOL_FLAGS_REEXPORT) { + Entry.Address = 0; + Entry.Other = decodeULEB128(CurrPtr, &Count); + CurrPtr += Count; + Entry.ImportName = std::string(reinterpret_cast<const char *>(CurrPtr)); + } else { + Entry.Address = decodeULEB128(CurrPtr, &Count); + CurrPtr += Count; + if (Entry.Flags & MachO::EXPORT_SYMBOL_FLAGS_STUB_AND_RESOLVER) { + Entry.Other = decodeULEB128(CurrPtr, &Count); + CurrPtr += Count; + } else + Entry.Other = 0; + } + } + uint8_t childrenCount = *CurrPtr++; + if (childrenCount == 0) + return CurrPtr; + + Entry.Children.insert(Entry.Children.begin(), (size_t)childrenCount, + MachOYAML::ExportEntry()); + for (auto &Child : Entry.Children) { + Child.Name = std::string(reinterpret_cast<const char *>(CurrPtr)); + CurrPtr += Child.Name.length() + 1; + Child.NodeOffset = decodeULEB128(CurrPtr, &Count); + CurrPtr += Count; + } + for (auto &Child : Entry.Children) { + CurrPtr = processExportNode(CurrPtr, End, Child); + } + return CurrPtr; +} + +void MachODumper::dumpExportTrie(std::unique_ptr<MachOYAML::Object> &Y) { + MachOYAML::LinkEditData &LEData = Y->LinkEdit; + auto ExportsTrie = Obj.getDyldInfoExportsTrie(); + processExportNode(ExportsTrie.begin(), ExportsTrie.end(), LEData.ExportTrie); +} + +template <typename nlist_t> +MachOYAML::NListEntry constructNameList(const nlist_t &nlist) { + MachOYAML::NListEntry NL; + NL.n_strx = nlist.n_strx; + NL.n_type = nlist.n_type; + NL.n_sect = nlist.n_sect; + NL.n_desc = nlist.n_desc; + NL.n_value = nlist.n_value; + return NL; +} + +void MachODumper::dumpSymbols(std::unique_ptr<MachOYAML::Object> &Y) { + MachOYAML::LinkEditData &LEData = Y->LinkEdit; + + for (auto Symbol : Obj.symbols()) { + MachOYAML::NListEntry NLE = + Obj.is64Bit() ? constructNameList<MachO::nlist_64>( + *reinterpret_cast<const MachO::nlist_64 *>( + Symbol.getRawDataRefImpl().p)) + : constructNameList<MachO::nlist>( + *reinterpret_cast<const MachO::nlist *>( + Symbol.getRawDataRefImpl().p)); + LEData.NameList.push_back(NLE); + } + + StringRef RemainingTable = Obj.getStringTableData(); + while (RemainingTable.size() > 0) { + auto SymbolPair = RemainingTable.split('\0'); + RemainingTable = SymbolPair.second; + if (SymbolPair.first.empty()) + break; + LEData.StringTable.push_back(SymbolPair.first); + } +} + +Error macho2yaml(raw_ostream &Out, const object::MachOObjectFile &Obj) { + MachODumper Dumper(Obj); + Expected<std::unique_ptr<MachOYAML::Object>> YAML = Dumper.dump(); + if (!YAML) + return YAML.takeError(); + + yaml::YamlObjectFile YAMLFile; + YAMLFile.MachO = std::move(YAML.get()); + + yaml::Output Yout(Out); + Yout << YAMLFile; + return Error::success(); +} + +Error macho2yaml(raw_ostream &Out, const object::MachOUniversalBinary &Obj) { + yaml::YamlObjectFile YAMLFile; + YAMLFile.FatMachO.reset(new MachOYAML::UniversalBinary()); + MachOYAML::UniversalBinary &YAML = *YAMLFile.FatMachO; + YAML.Header.magic = Obj.getMagic(); + YAML.Header.nfat_arch = Obj.getNumberOfObjects(); + + for (auto Slice : Obj.objects()) { + MachOYAML::FatArch arch; + arch.cputype = Slice.getCPUType(); + arch.cpusubtype = Slice.getCPUSubType(); + arch.offset = Slice.getOffset(); + arch.size = Slice.getSize(); + arch.align = Slice.getAlign(); + arch.reserved = Slice.getReserved(); + YAML.FatArchs.push_back(arch); + + auto SliceObj = Slice.getAsObjectFile(); + if (!SliceObj) + return SliceObj.takeError(); + + MachODumper Dumper(*SliceObj.get()); + Expected<std::unique_ptr<MachOYAML::Object>> YAMLObj = Dumper.dump(); + if (!YAMLObj) + return YAMLObj.takeError(); + YAML.Slices.push_back(*YAMLObj.get()); + } + + yaml::Output Yout(Out); + Yout << YAML; + return Error::success(); +} + +std::error_code macho2yaml(raw_ostream &Out, const object::Binary &Binary) { + if (const auto *MachOObj = dyn_cast<object::MachOUniversalBinary>(&Binary)) { + if (auto Err = macho2yaml(Out, *MachOObj)) { + return errorToErrorCode(std::move(Err)); + } + return obj2yaml_error::success; + } + + if (const auto *MachOObj = dyn_cast<object::MachOObjectFile>(&Binary)) { + if (auto Err = macho2yaml(Out, *MachOObj)) { + return errorToErrorCode(std::move(Err)); + } + return obj2yaml_error::success; + } + + return obj2yaml_error::unsupported_obj_file_format; +} diff --git a/tools/obj2yaml/obj2yaml.cpp b/tools/obj2yaml/obj2yaml.cpp index ee6284da6e417..3f9373ee17e38 100644 --- a/tools/obj2yaml/obj2yaml.cpp +++ b/tools/obj2yaml/obj2yaml.cpp @@ -29,11 +29,15 @@ static std::error_code dumpObject(const ObjectFile &Obj) { } static std::error_code dumpInput(StringRef File) { - ErrorOr<OwningBinary<Binary>> BinaryOrErr = createBinary(File); - if (std::error_code EC = BinaryOrErr.getError()) - return EC; + Expected<OwningBinary<Binary>> BinaryOrErr = createBinary(File); + if (!BinaryOrErr) + return errorToErrorCode(BinaryOrErr.takeError()); Binary &Binary = *BinaryOrErr.get().getBinary(); + // Universal MachO is not a subclass of ObjectFile, so it needs to be handled + // here with the other binary types. + if (Binary.isMachO() || Binary.isMachOUniversalBinary()) + return macho2yaml(outs(), Binary); // TODO: If this is an archive, then burst it and dump each entry if (ObjectFile *Obj = dyn_cast<ObjectFile>(&Binary)) return dumpObject(*Obj); @@ -46,7 +50,7 @@ cl::opt<std::string> InputFilename(cl::Positional, cl::desc("<input file>"), int main(int argc, char *argv[]) { cl::ParseCommandLineOptions(argc, argv); - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. diff --git a/tools/obj2yaml/obj2yaml.h b/tools/obj2yaml/obj2yaml.h index 643ab7bc434d3..28c74751c0d95 100644 --- a/tools/obj2yaml/obj2yaml.h +++ b/tools/obj2yaml/obj2yaml.h @@ -21,5 +21,7 @@ std::error_code coff2yaml(llvm::raw_ostream &Out, const llvm::object::COFFObjectFile &Obj); std::error_code elf2yaml(llvm::raw_ostream &Out, const llvm::object::ObjectFile &Obj); +std::error_code macho2yaml(llvm::raw_ostream &Out, + const llvm::object::Binary &Obj); #endif diff --git a/tools/opt/AnalysisWrappers.cpp b/tools/opt/AnalysisWrappers.cpp index 4bdc268f8d759..cfdd2cf1582b0 100644 --- a/tools/opt/AnalysisWrappers.cpp +++ b/tools/opt/AnalysisWrappers.cpp @@ -71,23 +71,3 @@ char ExternalFunctionsPassedConstants::ID = 0; static RegisterPass<ExternalFunctionsPassedConstants> P1("print-externalfnconstants", "Print external fn callsites passed constants"); - -namespace { - struct CallGraphPrinter : public ModulePass { - static char ID; // Pass ID, replacement for typeid - CallGraphPrinter() : ModulePass(ID) {} - - void getAnalysisUsage(AnalysisUsage &AU) const override { - AU.setPreservesAll(); - AU.addRequiredTransitive<CallGraphWrapperPass>(); - } - bool runOnModule(Module &M) override { - getAnalysis<CallGraphWrapperPass>().print(errs(), &M); - return false; - } - }; -} - -char CallGraphPrinter::ID = 0; -static RegisterPass<CallGraphPrinter> - P2("print-callgraph", "Print a call graph"); diff --git a/tools/opt/BreakpointPrinter.cpp b/tools/opt/BreakpointPrinter.cpp index 363a7cd80074e..33b3edcd12378 100644 --- a/tools/opt/BreakpointPrinter.cpp +++ b/tools/opt/BreakpointPrinter.cpp @@ -25,7 +25,6 @@ namespace { struct BreakpointPrinter : public ModulePass { raw_ostream &Out; static char ID; - DITypeIdentifierMap TypeIdentifierMap; BreakpointPrinter(raw_ostream &out) : ModulePass(ID), Out(out) {} @@ -37,18 +36,13 @@ struct BreakpointPrinter : public ModulePass { } } else if (auto *TY = dyn_cast<DIType>(Context)) { if (!TY->getName().empty()) { - getContextName(TY->getScope().resolve(TypeIdentifierMap), N); + getContextName(TY->getScope().resolve(), N); N = N + TY->getName().str() + "::"; } } } bool runOnModule(Module &M) override { - TypeIdentifierMap.clear(); - NamedMDNode *CU_Nodes = M.getNamedMetadata("llvm.dbg.cu"); - if (CU_Nodes) - TypeIdentifierMap = generateDITypeIdentifierMap(CU_Nodes); - StringSet<> Processed; if (NamedMDNode *NMD = M.getNamedMetadata("llvm.dbg.sp")) for (unsigned i = 0, e = NMD->getNumOperands(); i != e; ++i) { @@ -56,7 +50,7 @@ struct BreakpointPrinter : public ModulePass { auto *SP = cast_or_null<DISubprogram>(NMD->getOperand(i)); if (!SP) continue; - getContextName(SP->getScope().resolve(TypeIdentifierMap), Name); + getContextName(SP->getScope().resolve(), Name); Name = Name + SP->getDisplayName().str(); if (!Name.empty() && Processed.insert(Name).second) { Out << Name << "\n"; diff --git a/tools/opt/Makefile b/tools/opt/Makefile deleted file mode 100644 index 2422eb4e40567..0000000000000 --- a/tools/opt/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/opt/Makefile ----------------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := opt -LINK_COMPONENTS := bitreader bitwriter asmparser irreader instrumentation scalaropts objcarcopts ipo vectorize all-targets codegen passes - -# Support plugins. -NO_DEAD_STRIP := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/opt/NewPMDriver.cpp b/tools/opt/NewPMDriver.cpp index 3030d65743af2..f7b2f18d09d90 100644 --- a/tools/opt/NewPMDriver.cpp +++ b/tools/opt/NewPMDriver.cpp @@ -15,7 +15,9 @@ #include "NewPMDriver.h" #include "llvm/ADT/StringRef.h" +#include "llvm/Analysis/AliasAnalysis.h" #include "llvm/Analysis/CGSCCPassManager.h" +#include "llvm/Analysis/LoopPassManager.h" #include "llvm/Bitcode/BitcodeWriterPass.h" #include "llvm/IR/Dominators.h" #include "llvm/IR/IRPrintingPasses.h" @@ -36,6 +38,15 @@ static cl::opt<bool> DebugPM("debug-pass-manager", cl::Hidden, cl::desc("Print pass management debugging information")); +// This flag specifies a textual description of the alias analysis pipeline to +// use when querying for aliasing information. It only works in concert with +// the "passes" flag above. +static cl::opt<std::string> + AAPipeline("aa-pipeline", + cl::desc("A textual description of the alias analysis " + "pipeline for handling managed aliasing queries"), + cl::Hidden); + bool llvm::runPassPipeline(StringRef Arg0, LLVMContext &Context, Module &M, TargetMachine *TM, tool_output_file *Out, StringRef PassPipeline, OutputKind OK, @@ -44,22 +55,28 @@ bool llvm::runPassPipeline(StringRef Arg0, LLVMContext &Context, Module &M, bool ShouldPreserveBitcodeUseListOrder) { PassBuilder PB(TM); + // Specially handle the alias analysis manager so that we can register + // a custom pipeline of AA passes with it. + AAManager AA; + if (!PB.parseAAPipeline(AA, AAPipeline)) { + errs() << Arg0 << ": unable to parse AA pipeline description.\n"; + return false; + } + + LoopAnalysisManager LAM(DebugPM); FunctionAnalysisManager FAM(DebugPM); CGSCCAnalysisManager CGAM(DebugPM); ModuleAnalysisManager MAM(DebugPM); + // Register the AA manager first so that our version is the one used. + FAM.registerPass([&] { return std::move(AA); }); + // Register all the basic analyses with the managers. PB.registerModuleAnalyses(MAM); PB.registerCGSCCAnalyses(CGAM); PB.registerFunctionAnalyses(FAM); - - // Cross register the analysis managers through their proxies. - MAM.registerPass(FunctionAnalysisManagerModuleProxy(FAM)); - MAM.registerPass(CGSCCAnalysisManagerModuleProxy(CGAM)); - CGAM.registerPass(FunctionAnalysisManagerCGSCCProxy(FAM)); - CGAM.registerPass(ModuleAnalysisManagerCGSCCProxy(MAM)); - FAM.registerPass(CGSCCAnalysisManagerFunctionProxy(CGAM)); - FAM.registerPass(ModuleAnalysisManagerFunctionProxy(MAM)); + PB.registerLoopAnalyses(LAM); + PB.crossRegisterProxies(LAM, FAM, CGAM, MAM); ModulePassManager MPM(DebugPM); if (VK > VK_NoVerifier) @@ -92,7 +109,7 @@ bool llvm::runPassPipeline(StringRef Arg0, LLVMContext &Context, Module &M, cl::PrintOptionValues(); // Now that we have all of the passes ready, run them. - MPM.run(M, &MAM); + MPM.run(M, MAM); // Declare success. if (OK != OK_NoOutput) diff --git a/tools/opt/NewPMDriver.h b/tools/opt/NewPMDriver.h index 349a7b1267f8e..66a470d49ba8b 100644 --- a/tools/opt/NewPMDriver.h +++ b/tools/opt/NewPMDriver.h @@ -21,9 +21,8 @@ #ifndef LLVM_TOOLS_OPT_NEWPMDRIVER_H #define LLVM_TOOLS_OPT_NEWPMDRIVER_H -#include "llvm/ADT/StringRef.h" - namespace llvm { +class StringRef; class LLVMContext; class Module; class TargetMachine; diff --git a/tools/opt/opt.cpp b/tools/opt/opt.cpp index fe1605aa8436f..0a4ce1dfe8efa 100644 --- a/tools/opt/opt.cpp +++ b/tools/opt/opt.cpp @@ -96,12 +96,16 @@ static cl::opt<bool> OutputAssembly("S", cl::desc("Write output as LLVM assembly")); static cl::opt<bool> -NoVerify("disable-verify", cl::desc("Do not verify result module"), cl::Hidden); +NoVerify("disable-verify", cl::desc("Do not run the verifier"), cl::Hidden); static cl::opt<bool> VerifyEach("verify-each", cl::desc("Verify after each transform")); static cl::opt<bool> + DisableDITypeMap("disable-debug-info-type-map", + cl::desc("Don't use a uniquing type map for debug info")); + +static cl::opt<bool> StripDebug("strip-debug", cl::desc("Strip debugger symbol info from translation unit")); @@ -136,6 +140,10 @@ static cl::opt<bool> OptLevelO3("O3", cl::desc("Optimization level 3. Similar to clang -O3")); +static cl::opt<unsigned> +CodeGenOptLevel("codegen-opt-level", + cl::desc("Override optimization level for codegen hooks")); + static cl::opt<std::string> TargetTriple("mtriple", cl::desc("Override target triple for module")); @@ -158,6 +166,12 @@ DisableSLPVectorization("disable-slp-vectorization", cl::desc("Disable the slp vectorization pass"), cl::init(false)); +static cl::opt<bool> EmitSummaryIndex("module-summary", + cl::desc("Emit module summary index"), + cl::init(false)); + +static cl::opt<bool> EmitModuleHash("module-hash", cl::desc("Emit module hash"), + cl::init(false)); static cl::opt<bool> DisableSimplifyLibCalls("disable-simplify-libcalls", @@ -196,6 +210,16 @@ static cl::opt<bool> cl::desc("Run all passes twice, re-using the same pass manager."), cl::init(false), cl::Hidden); +static cl::opt<bool> DiscardValueNames( + "discard-value-names", + cl::desc("Discard names from Value (other than GlobalValue)."), + cl::init(false), cl::Hidden); + +static cl::opt<bool> PassRemarksWithHotness( + "pass-remarks-with-hotness", + cl::desc("With PGO, include profile count in optimization remarks"), + cl::Hidden); + static inline void addPass(legacy::PassManagerBase &PM, Pass *P) { // Add the pass to the pass manager... PM.add(P); @@ -211,8 +235,10 @@ static inline void addPass(legacy::PassManagerBase &PM, Pass *P) { /// OptLevel - Optimization Level static void AddOptimizationPasses(legacy::PassManagerBase &MPM, legacy::FunctionPassManager &FPM, - unsigned OptLevel, unsigned SizeLevel) { - FPM.add(createVerifierPass()); // Verify that input is correct + TargetMachine *TM, unsigned OptLevel, + unsigned SizeLevel) { + if (!NoVerify || VerifyEach) + FPM.add(createVerifierPass()); // Verify that input is correct PassManagerBuilder Builder; Builder.OptLevel = OptLevel; @@ -240,6 +266,14 @@ static void AddOptimizationPasses(legacy::PassManagerBase &MPM, Builder.SLPVectorize = DisableSLPVectorization ? false : OptLevel > 1 && SizeLevel < 2; + // Add target-specific passes that need to run as early as possible. + if (TM) + Builder.addExtension( + PassManagerBuilder::EP_EarlyAsPossible, + [&](const PassManagerBuilder &, legacy::PassManagerBase &PM) { + TM->addEarlyAsPossiblePasses(PM); + }); + Builder.populateFunctionPassManager(FPM); Builder.populateModulePassManager(MPM); } @@ -260,6 +294,8 @@ static void AddStandardLinkPasses(legacy::PassManagerBase &PM) { // static CodeGenOpt::Level GetCodeGenOptLevel() { + if (CodeGenOptLevel.getNumOccurrences()) + return static_cast<CodeGenOpt::Level>(unsigned(CodeGenOptLevel)); if (OptLevelO1) return CodeGenOpt::Less; if (OptLevelO2) @@ -281,10 +317,9 @@ static TargetMachine* GetTargetMachine(Triple TheTriple, StringRef CPUStr, return nullptr; } - return TheTarget->createTargetMachine(TheTriple.getTriple(), - CPUStr, FeaturesStr, Options, - RelocModel, CMModel, - GetCodeGenOptLevel()); + return TheTarget->createTargetMachine(TheTriple.getTriple(), CPUStr, + FeaturesStr, Options, getRelocModel(), + CMModel, GetCodeGenOptLevel()); } #ifdef LINK_POLLY_INTO_TOOLS @@ -297,14 +332,14 @@ void initializePollyPasses(llvm::PassRegistry &Registry); // main for opt // int main(int argc, char **argv) { - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); llvm::PrettyStackTraceProgram X(argc, argv); // Enable debug stream buffering. EnableDebugBuffering = true; llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. - LLVMContext &Context = getGlobalContext(); + LLVMContext Context; InitializeAllTargets(); InitializeAllTargetMCs(); @@ -329,7 +364,12 @@ int main(int argc, char **argv) { initializeRewriteSymbolsPass(Registry); initializeWinEHPreparePass(Registry); initializeDwarfEHPreparePass(Registry); + initializeSafeStackPass(Registry); initializeSjLjEHPreparePass(Registry); + initializePreISelIntrinsicLoweringLegacyPassPass(Registry); + initializeGlobalMergePass(Registry); + initializeInterleavedAccessPass(Registry); + initializeUnreachableBlockElimLegacyPassPass(Registry); #ifdef LINK_POLLY_INTO_TOOLS polly::initializePollyPasses(Registry); @@ -345,6 +385,13 @@ int main(int argc, char **argv) { SMDiagnostic Err; + Context.setDiscardValueNames(DiscardValueNames); + if (!DisableDITypeMap) + Context.enableDebugTypeODRUniquing(); + + if (PassRemarksWithHotness) + Context.setDiagnosticHotnessRequested(true); + // Load the input module... std::unique_ptr<Module> M = parseIRFile(InputFilename, Err, Context); @@ -491,27 +538,27 @@ int main(int argc, char **argv) { } if (OptLevelO1 && OptLevelO1.getPosition() < PassList.getPosition(i)) { - AddOptimizationPasses(Passes, *FPasses, 1, 0); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 1, 0); OptLevelO1 = false; } if (OptLevelO2 && OptLevelO2.getPosition() < PassList.getPosition(i)) { - AddOptimizationPasses(Passes, *FPasses, 2, 0); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 2, 0); OptLevelO2 = false; } if (OptLevelOs && OptLevelOs.getPosition() < PassList.getPosition(i)) { - AddOptimizationPasses(Passes, *FPasses, 2, 1); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 2, 1); OptLevelOs = false; } if (OptLevelOz && OptLevelOz.getPosition() < PassList.getPosition(i)) { - AddOptimizationPasses(Passes, *FPasses, 2, 2); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 2, 2); OptLevelOz = false; } if (OptLevelO3 && OptLevelO3.getPosition() < PassList.getPosition(i)) { - AddOptimizationPasses(Passes, *FPasses, 3, 0); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 3, 0); OptLevelO3 = false; } @@ -563,21 +610,21 @@ int main(int argc, char **argv) { } if (OptLevelO1) - AddOptimizationPasses(Passes, *FPasses, 1, 0); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 1, 0); if (OptLevelO2) - AddOptimizationPasses(Passes, *FPasses, 2, 0); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 2, 0); if (OptLevelOs) - AddOptimizationPasses(Passes, *FPasses, 2, 1); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 2, 1); if (OptLevelOz) - AddOptimizationPasses(Passes, *FPasses, 2, 2); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 2, 2); if (OptLevelO3) - AddOptimizationPasses(Passes, *FPasses, 3, 0); + AddOptimizationPasses(Passes, *FPasses, TM.get(), 3, 0); - if (OptLevelO1 || OptLevelO2 || OptLevelOs || OptLevelOz || OptLevelO3) { + if (FPasses) { FPasses->doInitialization(); for (Function &F : *M) FPasses->run(F); @@ -605,10 +652,15 @@ int main(int argc, char **argv) { BOS = make_unique<raw_svector_ostream>(Buffer); OS = BOS.get(); } - if (OutputAssembly) + if (OutputAssembly) { + if (EmitSummaryIndex) + report_fatal_error("Text output is incompatible with -module-summary"); + if (EmitModuleHash) + report_fatal_error("Text output is incompatible with -module-hash"); Passes.add(createPrintModulePass(*OS, "", PreserveAssemblyUseListOrder)); - else - Passes.add(createBitcodeWriterPass(*OS, PreserveBitcodeUseListOrder)); + } else + Passes.add(createBitcodeWriterPass(*OS, PreserveBitcodeUseListOrder, + EmitSummaryIndex, EmitModuleHash)); } // Before executing passes, print the final values of the LLVM options. diff --git a/tools/sancov/Makefile b/tools/sancov/Makefile deleted file mode 100644 index 7dba1a7a594a8..0000000000000 --- a/tools/sancov/Makefile +++ /dev/null @@ -1,18 +0,0 @@ -##===- tools/sancov/Makefile ----------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := sancov -LINK_COMPONENTS := all-targets DebugInfoDWARF DebugInfoPDB MC MCParser \ - MCDisassembler Object Support Symbolize - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/sancov/sancov.cc b/tools/sancov/sancov.cc index a07cdbe097a3b..55b03709dde34 100644 --- a/tools/sancov/sancov.cc +++ b/tools/sancov/sancov.cc @@ -11,10 +11,11 @@ // coverage. //===----------------------------------------------------------------------===// #include "llvm/ADT/STLExtras.h" +#include "llvm/ADT/Twine.h" #include "llvm/DebugInfo/Symbolize/Symbolize.h" #include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCContext.h" -#include "llvm/MC/MCDisassembler.h" +#include "llvm/MC/MCDisassembler/MCDisassembler.h" #include "llvm/MC/MCInst.h" #include "llvm/MC/MCInstPrinter.h" #include "llvm/MC/MCInstrAnalysis.h" @@ -24,16 +25,20 @@ #include "llvm/MC/MCSubtargetInfo.h" #include "llvm/Object/Archive.h" #include "llvm/Object/Binary.h" +#include "llvm/Object/ELFObjectFile.h" #include "llvm/Object/ObjectFile.h" +#include "llvm/Support/Casting.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/Errc.h" #include "llvm/Support/ErrorOr.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/LineIterator.h" +#include "llvm/Support/MD5.h" #include "llvm/Support/ManagedStatic.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/Path.h" #include "llvm/Support/PrettyStackTrace.h" +#include "llvm/Support/Regex.h" #include "llvm/Support/Signals.h" #include "llvm/Support/SpecialCaseList.h" #include "llvm/Support/TargetRegistry.h" @@ -41,9 +46,11 @@ #include "llvm/Support/ToolOutputFile.h" #include "llvm/Support/raw_ostream.h" +#include <algorithm> #include <set> #include <stdio.h> #include <string> +#include <utility> #include <vector> using namespace llvm; @@ -54,29 +61,34 @@ namespace { enum ActionType { PrintAction, + PrintCovPointsAction, CoveredFunctionsAction, - NotCoveredFunctionsAction + NotCoveredFunctionsAction, + HtmlReportAction, + StatsAction }; cl::opt<ActionType> Action( cl::desc("Action (required)"), cl::Required, cl::values(clEnumValN(PrintAction, "print", "Print coverage addresses"), + clEnumValN(PrintCovPointsAction, "print-coverage-pcs", + "Print coverage instrumentation points addresses."), clEnumValN(CoveredFunctionsAction, "covered-functions", "Print all covered funcions."), clEnumValN(NotCoveredFunctionsAction, "not-covered-functions", "Print all not covered funcions."), + clEnumValN(HtmlReportAction, "html-report", + "Print HTML coverage report."), + clEnumValN(StatsAction, "print-coverage-stats", + "Print coverage statistics."), clEnumValEnd)); -static cl::list<std::string> ClInputFiles(cl::Positional, cl::OneOrMore, - cl::desc("<filenames...>")); +static cl::list<std::string> + ClInputFiles(cl::Positional, cl::OneOrMore, + cl::desc("(<binary file>|<.sancov file>)...")); -static cl::opt<std::string> - ClBinaryName("obj", cl::Required, - cl::desc("Path to object file to be symbolized")); - -static cl::opt<bool> - ClDemangle("demangle", cl::init(true), - cl::desc("Print demangled function name.")); +static cl::opt<bool> ClDemangle("demangle", cl::init(true), + cl::desc("Print demangled function name.")); static cl::opt<std::string> ClStripPathPrefix( "strip_path_prefix", cl::init(""), @@ -90,7 +102,9 @@ static cl::opt<bool> ClUseDefaultBlacklist( "use_default_blacklist", cl::init(true), cl::Hidden, cl::desc("Controls if default blacklist should be used.")); -static const char *const DefaultBlacklist = "fun:__sanitizer_*"; +static const char *const DefaultBlacklistStr = "fun:__sanitizer_.*\n" + "src:/usr/include/.*\n" + "src:.*/libc\\+\\+/.*\n"; // --------- FORMAT SPECIFICATION --------- @@ -103,7 +117,12 @@ static const uint32_t BinCoverageMagic = 0xC0BFFFFF; static const uint32_t Bitness32 = 0xFFFFFF32; static const uint32_t Bitness64 = 0xFFFFFF64; -// --------- +// --------- ERROR HANDLING --------- + +static void Fail(const llvm::Twine &E) { + errs() << "Error: " << E << "\n"; + exit(1); +} static void FailIfError(std::error_code Error) { if (!Error) @@ -116,11 +135,21 @@ template <typename T> static void FailIfError(const ErrorOr<T> &E) { FailIfError(E.getError()); } -static void FailIfNotEmpty(const std::string &E) { - if (E.empty()) +static void FailIfError(Error Err) { + if (Err) { + logAllUnhandledErrors(std::move(Err), errs(), "Error: "); + exit(1); + } +} + +template <typename T> static void FailIfError(Expected<T> &E) { + FailIfError(E.takeError()); +} + +static void FailIfNotEmpty(const llvm::Twine &E) { + if (E.str().empty()) return; - errs() << "Error: " << E << "\n"; - exit(1); + Fail(E); } template <typename T> @@ -128,8 +157,24 @@ static void FailIfEmpty(const std::unique_ptr<T> &Ptr, const std::string &Message) { if (Ptr.get()) return; - errs() << "Error: " << Message << "\n"; - exit(1); + Fail(Message); +} + +// --------- + +// Produces std::map<K, std::vector<E>> grouping input +// elements by FuncTy result. +template <class RangeTy, class FuncTy> +static inline auto group_by(const RangeTy &R, FuncTy F) + -> std::map<typename std::decay<decltype(F(*R.begin()))>::type, + std::vector<typename std::decay<decltype(*R.begin())>::type>> { + std::map<typename std::decay<decltype(F(*R.begin()))>::type, + std::vector<typename std::decay<decltype(*R.begin())>::type>> + Result; + for (const auto &E : R) { + Result[F(E)].push_back(E); + } + return Result; } template <typename T> @@ -149,8 +194,18 @@ struct FileLoc { uint32_t Line; }; -struct FunctionLoc { - bool operator<(const FunctionLoc &RHS) const { +struct FileFn { + bool operator<(const FileFn &RHS) const { + return std::tie(FileName, FunctionName) < + std::tie(RHS.FileName, RHS.FunctionName); + } + + std::string FileName; + std::string FunctionName; +}; + +struct FnLoc { + bool operator<(const FnLoc &RHS) const { return std::tie(Loc, FunctionName) < std::tie(RHS.Loc, RHS.FunctionName); } @@ -167,53 +222,98 @@ std::string stripPathPrefix(std::string Path) { return Path.substr(Pos + ClStripPathPrefix.size()); } -// Compute [FileLoc -> FunctionName] map for given addresses. -static std::map<FileLoc, std::string> -computeFunctionsMap(const std::set<uint64_t> &Addrs) { - std::map<FileLoc, std::string> Fns; - +static std::unique_ptr<symbolize::LLVMSymbolizer> createSymbolizer() { symbolize::LLVMSymbolizer::Options SymbolizerOptions; SymbolizerOptions.Demangle = ClDemangle; SymbolizerOptions.UseSymbolTable = true; - symbolize::LLVMSymbolizer Symbolizer(SymbolizerOptions); + return std::unique_ptr<symbolize::LLVMSymbolizer>( + new symbolize::LLVMSymbolizer(SymbolizerOptions)); +} - // Fill in Fns map. - for (auto Addr : Addrs) { - auto InliningInfo = Symbolizer.symbolizeInlinedCode(ClBinaryName, Addr); - FailIfError(InliningInfo); - for (uint32_t I = 0; I < InliningInfo->getNumberOfFrames(); ++I) { - auto FrameInfo = InliningInfo->getFrame(I); - SmallString<256> FileName(FrameInfo.FileName); - sys::path::remove_dots(FileName, /* remove_dot_dot */ true); - FileLoc Loc = {FileName.str(), FrameInfo.Line}; - Fns[Loc] = FrameInfo.FunctionName; - } +// A DILineInfo with address. +struct AddrInfo : public DILineInfo { + uint64_t Addr; + + AddrInfo(const DILineInfo &DI, uint64_t Addr) : DILineInfo(DI), Addr(Addr) { + FileName = normalizeFilename(FileName); } - return Fns; -} +private: + static std::string normalizeFilename(const std::string &FileName) { + SmallString<256> S(FileName); + sys::path::remove_dots(S, /* remove_dot_dot */ true); + return S.str().str(); + } +}; -// Compute functions for given addresses. It keeps only the first -// occurence of a function within a file. -std::set<FunctionLoc> computeFunctionLocs(const std::set<uint64_t> &Addrs) { - std::map<FileLoc, std::string> Fns = computeFunctionsMap(Addrs); +class Blacklists { +public: + Blacklists() + : DefaultBlacklist(createDefaultBlacklist()), + UserBlacklist(createUserBlacklist()) {} + + // AddrInfo contains normalized filename. It is important to check it rather + // than DILineInfo. + bool isBlacklisted(const AddrInfo &AI) { + if (DefaultBlacklist && DefaultBlacklist->inSection("fun", AI.FunctionName)) + return true; + if (DefaultBlacklist && DefaultBlacklist->inSection("src", AI.FileName)) + return true; + if (UserBlacklist && UserBlacklist->inSection("fun", AI.FunctionName)) + return true; + if (UserBlacklist && UserBlacklist->inSection("src", AI.FileName)) + return true; + return false; + } - std::set<FunctionLoc> Result; - std::string LastFileName; - std::set<std::string> ProcessedFunctions; +private: + static std::unique_ptr<SpecialCaseList> createDefaultBlacklist() { + if (!ClUseDefaultBlacklist) + return std::unique_ptr<SpecialCaseList>(); + std::unique_ptr<MemoryBuffer> MB = + MemoryBuffer::getMemBuffer(DefaultBlacklistStr); + std::string Error; + auto Blacklist = SpecialCaseList::create(MB.get(), Error); + FailIfNotEmpty(Error); + return Blacklist; + } - for (const auto &P : Fns) { - std::string FileName = P.first.FileName; - std::string FunctionName = P.second; + static std::unique_ptr<SpecialCaseList> createUserBlacklist() { + if (ClBlacklist.empty()) + return std::unique_ptr<SpecialCaseList>(); - if (LastFileName != FileName) - ProcessedFunctions.clear(); - LastFileName = FileName; + return SpecialCaseList::createOrDie({{ClBlacklist}}); + } + std::unique_ptr<SpecialCaseList> DefaultBlacklist; + std::unique_ptr<SpecialCaseList> UserBlacklist; +}; - if (!ProcessedFunctions.insert(FunctionName).second) - continue; +// Collect all debug info for given addresses. +static std::vector<AddrInfo> getAddrInfo(const std::string &ObjectFile, + const std::set<uint64_t> &Addrs, + bool InlinedCode) { + std::vector<AddrInfo> Result; + auto Symbolizer(createSymbolizer()); + Blacklists B; - Result.insert(FunctionLoc{P.first, P.second}); + for (auto Addr : Addrs) { + auto LineInfo = Symbolizer->symbolizeCode(ObjectFile, Addr); + FailIfError(LineInfo); + auto LineAddrInfo = AddrInfo(*LineInfo, Addr); + if (B.isBlacklisted(LineAddrInfo)) + continue; + Result.push_back(LineAddrInfo); + if (InlinedCode) { + auto InliningInfo = Symbolizer->symbolizeInlinedCode(ObjectFile, Addr); + FailIfError(InliningInfo); + for (uint32_t I = 0; I < InliningInfo->getNumberOfFrames(); ++I) { + auto FrameInfo = InliningInfo->getFrame(I); + auto FrameAddrInfo = AddrInfo(FrameInfo, Addr); + if (B.isBlacklisted(FrameAddrInfo)) + continue; + Result.push_back(FrameAddrInfo); + } + } } return Result; @@ -226,22 +326,20 @@ findSanitizerCovFunctions(const object::ObjectFile &O) { std::set<uint64_t> Result; for (const object::SymbolRef &Symbol : O.symbols()) { - ErrorOr<uint64_t> AddressOrErr = Symbol.getAddress(); - FailIfError(AddressOrErr); + Expected<uint64_t> AddressOrErr = Symbol.getAddress(); + FailIfError(errorToErrorCode(AddressOrErr.takeError())); - ErrorOr<StringRef> NameOrErr = Symbol.getName(); - FailIfError(NameOrErr); + Expected<StringRef> NameOrErr = Symbol.getName(); + FailIfError(errorToErrorCode(NameOrErr.takeError())); StringRef Name = NameOrErr.get(); if (Name == "__sanitizer_cov" || Name == "__sanitizer_cov_with_check" || Name == "__sanitizer_cov_trace_func_enter") { - Result.insert(AddressOrErr.get()); + if (!(Symbol.getFlags() & object::BasicSymbolRef::SF_Undefined)) + Result.insert(AddressOrErr.get()); } } - if (Result.empty()) - FailIfNotEmpty("__sanitizer_cov* functions not found"); - return Result; } @@ -284,8 +382,10 @@ static void getObjectCoveragePoints(const object::ObjectFile &O, FailIfEmpty(MIA, "no instruction analysis info for target " + TripleName); auto SanCovAddrs = findSanitizerCovFunctions(O); + if (SanCovAddrs.empty()) + Fail("__sanitizer_cov* functions not found"); - for (const auto Section : O.sections()) { + for (object::SectionRef Section : O.sections()) { if (Section.isVirtual() || !Section.isText()) // llvm-objdump does the same. continue; uint64_t SectionAddr = Section.getAddress(); @@ -293,9 +393,6 @@ static void getObjectCoveragePoints(const object::ObjectFile &O, if (!SectSize) continue; - StringRef SectionName; - FailIfError(Section.getName(SectionName)); - StringRef BytesStr; FailIfError(Section.getContents(BytesStr)); ArrayRef<uint8_t> Bytes(reinterpret_cast<const uint8_t *>(BytesStr.data()), @@ -310,98 +407,170 @@ static void getObjectCoveragePoints(const object::ObjectFile &O, Size = 1; continue; } + uint64_t Addr = Index + SectionAddr; + // Sanitizer coverage uses the address of the next instruction - 1. + uint64_t CovPoint = Addr + Size - 1; uint64_t Target; if (MIA->isCall(Inst) && - MIA->evaluateBranch(Inst, SectionAddr + Index, Size, Target)) { - if (SanCovAddrs.find(Target) != SanCovAddrs.end()) { - // Sanitizer coverage uses the address of the next instruction - 1. - Addrs->insert(Index + SectionAddr + Size - 1); - } - } + MIA->evaluateBranch(Inst, SectionAddr + Index, Size, Target) && + SanCovAddrs.find(Target) != SanCovAddrs.end()) + Addrs->insert(CovPoint); } } } -static void getArchiveCoveragePoints(const object::Archive &A, - std::set<uint64_t> *Addrs) { - for (auto &ErrorOrChild : A.children()) { - FailIfError(ErrorOrChild); - const object::Archive::Child &C = *ErrorOrChild; - ErrorOr<std::unique_ptr<object::Binary>> ChildOrErr = C.getAsBinary(); - FailIfError(ChildOrErr); - if (object::ObjectFile *O = - dyn_cast<object::ObjectFile>(&*ChildOrErr.get())) - getObjectCoveragePoints(*O, Addrs); +static void +visitObjectFiles(const object::Archive &A, + function_ref<void(const object::ObjectFile &)> Fn) { + Error Err; + for (auto &C : A.children(Err)) { + Expected<std::unique_ptr<object::Binary>> ChildOrErr = C.getAsBinary(); + FailIfError(errorToErrorCode(ChildOrErr.takeError())); + if (auto *O = dyn_cast<object::ObjectFile>(&*ChildOrErr.get())) + Fn(*O); else FailIfError(object::object_error::invalid_file_type); } + FailIfError(std::move(Err)); } -// Locate addresses of all coverage points in a file. Coverage point -// is defined as the 'address of instruction following __sanitizer_cov -// call - 1'. -std::set<uint64_t> getCoveragePoints(std::string FileName) { - std::set<uint64_t> Result; - - ErrorOr<object::OwningBinary<object::Binary>> BinaryOrErr = +static void +visitObjectFiles(const std::string &FileName, + function_ref<void(const object::ObjectFile &)> Fn) { + Expected<object::OwningBinary<object::Binary>> BinaryOrErr = object::createBinary(FileName); - FailIfError(BinaryOrErr); + if (!BinaryOrErr) + FailIfError(errorToErrorCode(BinaryOrErr.takeError())); object::Binary &Binary = *BinaryOrErr.get().getBinary(); if (object::Archive *A = dyn_cast<object::Archive>(&Binary)) - getArchiveCoveragePoints(*A, &Result); + visitObjectFiles(*A, Fn); else if (object::ObjectFile *O = dyn_cast<object::ObjectFile>(&Binary)) - getObjectCoveragePoints(*O, &Result); + Fn(*O); else FailIfError(object::object_error::invalid_file_type); +} +std::set<uint64_t> findSanitizerCovFunctions(const std::string &FileName) { + std::set<uint64_t> Result; + visitObjectFiles(FileName, [&](const object::ObjectFile &O) { + auto Addrs = findSanitizerCovFunctions(O); + Result.insert(Addrs.begin(), Addrs.end()); + }); return Result; } -static std::unique_ptr<SpecialCaseList> createDefaultBlacklist() { - if (!ClUseDefaultBlacklist) - return std::unique_ptr<SpecialCaseList>(); - std::unique_ptr<MemoryBuffer> MB = - MemoryBuffer::getMemBuffer(DefaultBlacklist); - std::string Error; - auto Blacklist = SpecialCaseList::create(MB.get(), Error); - FailIfNotEmpty(Error); - return Blacklist; +// Locate addresses of all coverage points in a file. Coverage point +// is defined as the 'address of instruction following __sanitizer_cov +// call - 1'. +std::set<uint64_t> getCoveragePoints(const std::string &FileName) { + std::set<uint64_t> Result; + visitObjectFiles(FileName, [&](const object::ObjectFile &O) { + getObjectCoveragePoints(O, &Result); + }); + return Result; +} + +static void printCovPoints(const std::string &ObjFile, raw_ostream &OS) { + for (uint64_t Addr : getCoveragePoints(ObjFile)) { + OS << "0x"; + OS.write_hex(Addr); + OS << "\n"; + } +} + +static std::string escapeHtml(const std::string &S) { + std::string Result; + Result.reserve(S.size()); + for (char Ch : S) { + switch (Ch) { + case '&': + Result.append("&"); + break; + case '\'': + Result.append("'"); + break; + case '"': + Result.append("""); + break; + case '<': + Result.append("<"); + break; + case '>': + Result.append(">"); + break; + default: + Result.push_back(Ch); + break; + } + } + return Result; } -static std::unique_ptr<SpecialCaseList> createUserBlacklist() { - if (ClBlacklist.empty()) - return std::unique_ptr<SpecialCaseList>(); +// Adds leading zeroes wrapped in 'lz' style. +// Leading zeroes help locate 000% coverage. +static std::string formatHtmlPct(size_t Pct) { + Pct = std::max(std::size_t{0}, std::min(std::size_t{100}, Pct)); - return SpecialCaseList::createOrDie({{ClBlacklist}}); + std::string Num = std::to_string(Pct); + std::string Zeroes(3 - Num.size(), '0'); + if (!Zeroes.empty()) + Zeroes = "<span class='lz'>" + Zeroes + "</span>"; + + return Zeroes + Num; } -static void printFunctionLocs(const std::set<FunctionLoc> &FnLocs, - raw_ostream &OS) { - std::unique_ptr<SpecialCaseList> DefaultBlacklist = createDefaultBlacklist(); - std::unique_ptr<SpecialCaseList> UserBlacklist = createUserBlacklist(); +static std::string anchorName(const std::string &Anchor) { + llvm::MD5 Hasher; + llvm::MD5::MD5Result Hash; + Hasher.update(Anchor); + Hasher.final(Hash); - for (const FunctionLoc &FnLoc : FnLocs) { - if (DefaultBlacklist && - DefaultBlacklist->inSection("fun", FnLoc.FunctionName)) - continue; - if (DefaultBlacklist && - DefaultBlacklist->inSection("src", FnLoc.Loc.FileName)) - continue; - if (UserBlacklist && UserBlacklist->inSection("fun", FnLoc.FunctionName)) - continue; - if (UserBlacklist && UserBlacklist->inSection("src", FnLoc.Loc.FileName)) - continue; + SmallString<32> HexString; + llvm::MD5::stringifyResult(Hash, HexString); + return HexString.str().str(); +} - OS << stripPathPrefix(FnLoc.Loc.FileName) << ":" << FnLoc.Loc.Line << " " - << FnLoc.FunctionName << "\n"; +static ErrorOr<bool> isCoverageFile(const std::string &FileName) { + ErrorOr<std::unique_ptr<MemoryBuffer>> BufOrErr = + MemoryBuffer::getFile(FileName); + if (!BufOrErr) { + errs() << "Warning: " << BufOrErr.getError().message() << "(" + << BufOrErr.getError().value() + << "), filename: " << llvm::sys::path::filename(FileName) << "\n"; + return BufOrErr.getError(); + } + std::unique_ptr<MemoryBuffer> Buf = std::move(BufOrErr.get()); + if (Buf->getBufferSize() < 8) { + return false; } + const FileHeader *Header = + reinterpret_cast<const FileHeader *>(Buf->getBufferStart()); + return Header->Magic == BinCoverageMagic; +} + +struct CoverageStats { + CoverageStats() : AllPoints(0), CovPoints(0), AllFns(0), CovFns(0) {} + + size_t AllPoints; + size_t CovPoints; + size_t AllFns; + size_t CovFns; +}; + +static raw_ostream &operator<<(raw_ostream &OS, const CoverageStats &Stats) { + OS << "all-edges: " << Stats.AllPoints << "\n"; + OS << "cov-edges: " << Stats.CovPoints << "\n"; + OS << "all-functions: " << Stats.AllFns << "\n"; + OS << "cov-functions: " << Stats.CovFns << "\n"; + return OS; } class CoverageData { - public: +public: // Read single file coverage data. - static ErrorOr<std::unique_ptr<CoverageData>> read(std::string FileName) { + static ErrorOr<std::unique_ptr<CoverageData>> + read(const std::string &FileName) { ErrorOr<std::unique_ptr<MemoryBuffer>> BufOrErr = MemoryBuffer::getFile(FileName); if (!BufOrErr) @@ -471,37 +640,570 @@ class CoverageData { } } +protected: + explicit CoverageData(std::unique_ptr<std::set<uint64_t>> Addrs) + : Addrs(std::move(Addrs)) {} + + friend class CoverageDataWithObjectFile; + + std::unique_ptr<std::set<uint64_t>> Addrs; +}; + +// Coverage data translated into source code line-level information. +// Fetches debug info in constructor and calculates various information per +// request. +class SourceCoverageData { +public: + enum LineStatus { + // coverage information for the line is not available. + // default value in maps. + UNKNOWN = 0, + // the line is fully covered. + COVERED = 1, + // the line is fully uncovered. + NOT_COVERED = 2, + // some points in the line a covered, some are not. + MIXED = 3 + }; + + SourceCoverageData(std::string ObjectFile, const std::set<uint64_t> &Addrs) + : AllCovPoints(getCoveragePoints(ObjectFile)) { + if (!std::includes(AllCovPoints.begin(), AllCovPoints.end(), Addrs.begin(), + Addrs.end())) { + Fail("Coverage points in binary and .sancov file do not match."); + } + + AllAddrInfo = getAddrInfo(ObjectFile, AllCovPoints, true); + CovAddrInfo = getAddrInfo(ObjectFile, Addrs, true); + } + + // Compute number of coverage points hit/total in a file. + // file_name -> <coverage, all_coverage> + std::map<std::string, std::pair<size_t, size_t>> computeFileCoverage() { + std::map<std::string, std::pair<size_t, size_t>> FileCoverage; + auto AllCovPointsByFile = + group_by(AllAddrInfo, [](const AddrInfo &AI) { return AI.FileName; }); + auto CovPointsByFile = + group_by(CovAddrInfo, [](const AddrInfo &AI) { return AI.FileName; }); + + for (const auto &P : AllCovPointsByFile) { + const std::string &FileName = P.first; + + FileCoverage[FileName] = + std::make_pair(CovPointsByFile[FileName].size(), + AllCovPointsByFile[FileName].size()); + } + return FileCoverage; + } + + // line_number -> line_status. + typedef std::map<int, LineStatus> LineStatusMap; + // file_name -> LineStatusMap + typedef std::map<std::string, LineStatusMap> FileLineStatusMap; + + // fills in the {file_name -> {line_no -> status}} map. + FileLineStatusMap computeLineStatusMap() { + FileLineStatusMap StatusMap; + + auto AllLocs = group_by(AllAddrInfo, [](const AddrInfo &AI) { + return FileLoc{AI.FileName, AI.Line}; + }); + auto CovLocs = group_by(CovAddrInfo, [](const AddrInfo &AI) { + return FileLoc{AI.FileName, AI.Line}; + }); + + for (const auto &P : AllLocs) { + const FileLoc &Loc = P.first; + auto I = CovLocs.find(Loc); + + if (I == CovLocs.end()) { + StatusMap[Loc.FileName][Loc.Line] = NOT_COVERED; + } else { + StatusMap[Loc.FileName][Loc.Line] = + (I->second.size() == P.second.size()) ? COVERED : MIXED; + } + } + return StatusMap; + } + + std::set<FileFn> computeAllFunctions() const { + std::set<FileFn> Fns; + for (const auto &AI : AllAddrInfo) { + Fns.insert(FileFn{AI.FileName, AI.FunctionName}); + } + return Fns; + } + + std::set<FileFn> computeCoveredFunctions() const { + std::set<FileFn> Fns; + auto CovFns = group_by(CovAddrInfo, [](const AddrInfo &AI) { + return FileFn{AI.FileName, AI.FunctionName}; + }); + + for (const auto &P : CovFns) { + Fns.insert(P.first); + } + return Fns; + } + + std::set<FileFn> computeNotCoveredFunctions() const { + std::set<FileFn> Fns; + + auto AllFns = group_by(AllAddrInfo, [](const AddrInfo &AI) { + return FileFn{AI.FileName, AI.FunctionName}; + }); + auto CovFns = group_by(CovAddrInfo, [](const AddrInfo &AI) { + return FileFn{AI.FileName, AI.FunctionName}; + }); + + for (const auto &P : AllFns) { + if (CovFns.find(P.first) == CovFns.end()) { + Fns.insert(P.first); + } + } + return Fns; + } + + // Compute % coverage for each function. + std::map<FileFn, int> computeFunctionsCoverage() const { + std::map<FileFn, int> FnCoverage; + auto AllFns = group_by(AllAddrInfo, [](const AddrInfo &AI) { + return FileFn{AI.FileName, AI.FunctionName}; + }); + + auto CovFns = group_by(CovAddrInfo, [](const AddrInfo &AI) { + return FileFn{AI.FileName, AI.FunctionName}; + }); + + for (const auto &P : AllFns) { + FileFn F = P.first; + FnCoverage[F] = CovFns[F].size() * 100 / P.second.size(); + } + + return FnCoverage; + } + + typedef std::map<FileLoc, std::set<std::string>> FunctionLocs; + // finds first line number in a file for each function. + FunctionLocs resolveFunctions(const std::set<FileFn> &Fns) const { + std::vector<AddrInfo> FnAddrs; + for (const auto &AI : AllAddrInfo) { + if (Fns.find(FileFn{AI.FileName, AI.FunctionName}) != Fns.end()) + FnAddrs.push_back(AI); + } + + auto GroupedAddrs = group_by(FnAddrs, [](const AddrInfo &AI) { + return FnLoc{FileLoc{AI.FileName, AI.Line}, AI.FunctionName}; + }); + + FunctionLocs Result; + std::string LastFileName; + std::set<std::string> ProcessedFunctions; + + for (const auto &P : GroupedAddrs) { + const FnLoc &Loc = P.first; + std::string FileName = Loc.Loc.FileName; + std::string FunctionName = Loc.FunctionName; + + if (LastFileName != FileName) + ProcessedFunctions.clear(); + LastFileName = FileName; + + if (!ProcessedFunctions.insert(FunctionName).second) + continue; + + auto FLoc = FileLoc{FileName, Loc.Loc.Line}; + Result[FLoc].insert(FunctionName); + } + return Result; + } + + std::set<std::string> files() const { + std::set<std::string> Files; + for (const auto &AI : AllAddrInfo) { + Files.insert(AI.FileName); + } + return Files; + } + + void collectStats(CoverageStats *Stats) const { + Stats->AllPoints += AllCovPoints.size(); + Stats->AllFns += computeAllFunctions().size(); + Stats->CovFns += computeCoveredFunctions().size(); + } + +private: + const std::set<uint64_t> AllCovPoints; + + std::vector<AddrInfo> AllAddrInfo; + std::vector<AddrInfo> CovAddrInfo; +}; + +static void printFunctionLocs(const SourceCoverageData::FunctionLocs &FnLocs, + raw_ostream &OS) { + for (const auto &Fns : FnLocs) { + for (const auto &Fn : Fns.second) { + OS << stripPathPrefix(Fns.first.FileName) << ":" << Fns.first.Line << " " + << Fn << "\n"; + } + } +} + +// Holder for coverage data + filename of corresponding object file. +class CoverageDataWithObjectFile : public CoverageData { +public: + static ErrorOr<std::unique_ptr<CoverageDataWithObjectFile>> + readAndMerge(const std::string &ObjectFile, + const std::vector<std::string> &FileNames) { + auto MergedDataOrError = CoverageData::readAndMerge(FileNames); + if (!MergedDataOrError) + return MergedDataOrError.getError(); + return std::unique_ptr<CoverageDataWithObjectFile>( + new CoverageDataWithObjectFile(ObjectFile, + std::move(MergedDataOrError.get()))); + } + + std::string object_file() const { return ObjectFile; } + // Print list of covered functions. // Line format: <file_name>:<line> <function_name> - void printCoveredFunctions(raw_ostream &OS) { - printFunctionLocs(computeFunctionLocs(*Addrs), OS); + void printCoveredFunctions(raw_ostream &OS) const { + SourceCoverageData SCovData(ObjectFile, *Addrs); + auto CoveredFns = SCovData.computeCoveredFunctions(); + printFunctionLocs(SCovData.resolveFunctions(CoveredFns), OS); } // Print list of not covered functions. // Line format: <file_name>:<line> <function_name> - void printNotCoveredFunctions(raw_ostream &OS) { - std::set<FunctionLoc> AllFns = - computeFunctionLocs(getCoveragePoints(ClBinaryName)); - std::set<FunctionLoc> CoveredFns = computeFunctionLocs(*Addrs); + void printNotCoveredFunctions(raw_ostream &OS) const { + SourceCoverageData SCovData(ObjectFile, *Addrs); + auto NotCoveredFns = SCovData.computeNotCoveredFunctions(); + printFunctionLocs(SCovData.resolveFunctions(NotCoveredFns), OS); + } + + void printReport(raw_ostream &OS) const { + SourceCoverageData SCovData(ObjectFile, *Addrs); + auto LineStatusMap = SCovData.computeLineStatusMap(); + + std::set<FileFn> AllFns = SCovData.computeAllFunctions(); + // file_loc -> set[function_name] + auto AllFnsByLoc = SCovData.resolveFunctions(AllFns); + auto FileCoverage = SCovData.computeFileCoverage(); + + auto FnCoverage = SCovData.computeFunctionsCoverage(); + auto FnCoverageByFile = + group_by(FnCoverage, [](const std::pair<FileFn, int> &FileFn) { + return FileFn.first.FileName; + }); + + // TOC + + size_t NotCoveredFilesCount = 0; + std::set<std::string> Files = SCovData.files(); + + // Covered Files. + OS << "<details open><summary>Touched Files</summary>\n"; + OS << "<table>\n"; + OS << "<tr><th>File</th><th>Coverage %</th>"; + OS << "<th>Hit (Total) Fns</th></tr>\n"; + for (const auto &FileName : Files) { + std::pair<size_t, size_t> FC = FileCoverage[FileName]; + if (FC.first == 0) { + NotCoveredFilesCount++; + continue; + } + size_t CovPct = FC.second == 0 ? 100 : 100 * FC.first / FC.second; + + OS << "<tr><td><a href=\"#" << anchorName(FileName) << "\">" + << stripPathPrefix(FileName) << "</a></td>" + << "<td>" << formatHtmlPct(CovPct) << "%</td>" + << "<td>" << FC.first << " (" << FC.second << ")" + << "</tr>\n"; + } + OS << "</table>\n"; + OS << "</details>\n"; + + // Not covered files. + if (NotCoveredFilesCount) { + OS << "<details><summary>Not Touched Files</summary>\n"; + OS << "<table>\n"; + for (const auto &FileName : Files) { + std::pair<size_t, size_t> FC = FileCoverage[FileName]; + if (FC.first == 0) + OS << "<tr><td>" << stripPathPrefix(FileName) << "</td>\n"; + } + OS << "</table>\n"; + OS << "</details>\n"; + } else { + OS << "<p>Congratulations! All source files are touched.</p>\n"; + } + + // Source + for (const auto &FileName : Files) { + std::pair<size_t, size_t> FC = FileCoverage[FileName]; + if (FC.first == 0) + continue; + OS << "<a name=\"" << anchorName(FileName) << "\"></a>\n"; + OS << "<h2>" << stripPathPrefix(FileName) << "</h2>\n"; + OS << "<details open><summary>Function Coverage</summary>"; + OS << "<div class='fnlist'>\n"; + + auto &FileFnCoverage = FnCoverageByFile[FileName]; + + for (const auto &P : FileFnCoverage) { + std::string FunctionName = P.first.FunctionName; + + OS << "<div class='fn' style='order: " << P.second << "'>"; + OS << "<span class='pct'>" << formatHtmlPct(P.second) + << "%</span> "; + OS << "<span class='name'><a href=\"#" + << anchorName(FileName + "::" + FunctionName) << "\">"; + OS << escapeHtml(FunctionName) << "</a></span>"; + OS << "</div>\n"; + } + OS << "</div></details>\n"; + + ErrorOr<std::unique_ptr<MemoryBuffer>> BufOrErr = + MemoryBuffer::getFile(FileName); + if (!BufOrErr) { + OS << "Error reading file: " << FileName << " : " + << BufOrErr.getError().message() << "(" + << BufOrErr.getError().value() << ")\n"; + continue; + } + + OS << "<pre>\n"; + const auto &LineStatuses = LineStatusMap[FileName]; + for (line_iterator I = line_iterator(*BufOrErr.get(), false); + !I.is_at_eof(); ++I) { + uint32_t Line = I.line_number(); + { // generate anchors (if any); + FileLoc Loc = FileLoc{FileName, Line}; + auto It = AllFnsByLoc.find(Loc); + if (It != AllFnsByLoc.end()) { + for (const std::string &Fn : It->second) { + OS << "<a name=\"" << anchorName(FileName + "::" + Fn) + << "\"></a>"; + }; + } + } - std::set<FunctionLoc> NotCoveredFns; - std::set_difference(AllFns.begin(), AllFns.end(), CoveredFns.begin(), - CoveredFns.end(), - std::inserter(NotCoveredFns, NotCoveredFns.end())); - printFunctionLocs(NotCoveredFns, OS); + OS << "<span "; + auto LIT = LineStatuses.find(I.line_number()); + auto Status = (LIT != LineStatuses.end()) ? LIT->second + : SourceCoverageData::UNKNOWN; + switch (Status) { + case SourceCoverageData::UNKNOWN: + OS << "class=unknown"; + break; + case SourceCoverageData::COVERED: + OS << "class=covered"; + break; + case SourceCoverageData::NOT_COVERED: + OS << "class=notcovered"; + break; + case SourceCoverageData::MIXED: + OS << "class=mixed"; + break; + } + OS << ">"; + OS << escapeHtml(*I) << "</span>\n"; + } + OS << "</pre>\n"; + } + } + + void collectStats(CoverageStats *Stats) const { + Stats->CovPoints += Addrs->size(); + + SourceCoverageData SCovData(ObjectFile, *Addrs); + SCovData.collectStats(Stats); } private: - explicit CoverageData(std::unique_ptr<std::set<uint64_t>> Addrs) - : Addrs(std::move(Addrs)) {} + CoverageDataWithObjectFile(std::string ObjectFile, + std::unique_ptr<CoverageData> Coverage) + : CoverageData(std::move(Coverage->Addrs)), + ObjectFile(std::move(ObjectFile)) {} + const std::string ObjectFile; +}; - std::unique_ptr<std::set<uint64_t>> Addrs; +// Multiple coverage files data organized by object file. +class CoverageDataSet { +public: + static ErrorOr<std::unique_ptr<CoverageDataSet>> + readCmdArguments(std::vector<std::string> FileNames) { + // Short name => file name. + std::map<std::string, std::string> ObjFiles; + std::string FirstObjFile; + std::set<std::string> CovFiles; + + // Partition input values into coverage/object files. + for (const auto &FileName : FileNames) { + auto ErrorOrIsCoverage = isCoverageFile(FileName); + if (!ErrorOrIsCoverage) + continue; + if (ErrorOrIsCoverage.get()) { + CovFiles.insert(FileName); + } else { + auto ShortFileName = llvm::sys::path::filename(FileName); + if (ObjFiles.find(ShortFileName) != ObjFiles.end()) { + Fail("Duplicate binary file with a short name: " + ShortFileName); + } + + ObjFiles[ShortFileName] = FileName; + if (FirstObjFile.empty()) + FirstObjFile = FileName; + } + } + + Regex SancovRegex("(.*)\\.[0-9]+\\.sancov"); + SmallVector<StringRef, 2> Components; + + // Object file => list of corresponding coverage file names. + auto CoverageByObjFile = group_by(CovFiles, [&](std::string FileName) { + auto ShortFileName = llvm::sys::path::filename(FileName); + auto Ok = SancovRegex.match(ShortFileName, &Components); + if (!Ok) { + Fail("Can't match coverage file name against " + "<module_name>.<pid>.sancov pattern: " + + FileName); + } + + auto Iter = ObjFiles.find(Components[1]); + if (Iter == ObjFiles.end()) { + Fail("Object file for coverage not found: " + FileName); + } + return Iter->second; + }); + + // Read coverage. + std::vector<std::unique_ptr<CoverageDataWithObjectFile>> MergedCoverage; + for (const auto &Pair : CoverageByObjFile) { + if (findSanitizerCovFunctions(Pair.first).empty()) { + for (const auto &FileName : Pair.second) { + CovFiles.erase(FileName); + } + + errs() + << "Ignoring " << Pair.first + << " and its coverage because __sanitizer_cov* functions were not " + "found.\n"; + continue; + } + + auto DataOrError = + CoverageDataWithObjectFile::readAndMerge(Pair.first, Pair.second); + FailIfError(DataOrError); + MergedCoverage.push_back(std::move(DataOrError.get())); + } + + return std::unique_ptr<CoverageDataSet>( + new CoverageDataSet(FirstObjFile, &MergedCoverage, CovFiles)); + } + + void printCoveredFunctions(raw_ostream &OS) const { + for (const auto &Cov : Coverage) { + Cov->printCoveredFunctions(OS); + } + } + + void printNotCoveredFunctions(raw_ostream &OS) const { + for (const auto &Cov : Coverage) { + Cov->printNotCoveredFunctions(OS); + } + } + + void printStats(raw_ostream &OS) const { + CoverageStats Stats; + for (const auto &Cov : Coverage) { + Cov->collectStats(&Stats); + } + OS << Stats; + } + + void printReport(raw_ostream &OS) const { + auto Title = + (llvm::sys::path::filename(MainObjFile) + " Coverage Report").str(); + + OS << "<html>\n"; + OS << "<head>\n"; + + // Stylesheet + OS << "<style>\n"; + OS << ".covered { background: #7F7; }\n"; + OS << ".notcovered { background: #F77; }\n"; + OS << ".mixed { background: #FF7; }\n"; + OS << "summary { font-weight: bold; }\n"; + OS << "details > summary + * { margin-left: 1em; }\n"; + OS << ".fnlist { display: flex; flex-flow: column nowrap; }\n"; + OS << ".fn { display: flex; flex-flow: row nowrap; }\n"; + OS << ".pct { width: 3em; text-align: right; margin-right: 1em; }\n"; + OS << ".name { flex: 2; }\n"; + OS << ".lz { color: lightgray; }\n"; + OS << "</style>\n"; + OS << "<title>" << Title << "</title>\n"; + OS << "</head>\n"; + OS << "<body>\n"; + + // Title + OS << "<h1>" << Title << "</h1>\n"; + + // Modules TOC. + if (Coverage.size() > 1) { + for (const auto &CovData : Coverage) { + OS << "<li><a href=\"#module_" << anchorName(CovData->object_file()) + << "\">" << llvm::sys::path::filename(CovData->object_file()) + << "</a></li>\n"; + } + } + + for (const auto &CovData : Coverage) { + if (Coverage.size() > 1) { + OS << "<h2>" << llvm::sys::path::filename(CovData->object_file()) + << "</h2>\n"; + } + OS << "<a name=\"module_" << anchorName(CovData->object_file()) + << "\"></a>\n"; + CovData->printReport(OS); + } + + // About + OS << "<details><summary>About</summary>\n"; + OS << "Coverage files:<ul>"; + for (const auto &InputFile : CoverageFiles) { + llvm::sys::fs::file_status Status; + llvm::sys::fs::status(InputFile, Status); + OS << "<li>" << stripPathPrefix(InputFile) << " (" + << Status.getLastModificationTime().str() << ")</li>\n"; + } + OS << "</ul></details>\n"; + + OS << "</body>\n"; + OS << "</html>\n"; + } + + bool empty() const { return Coverage.empty(); } + +private: + explicit CoverageDataSet( + const std::string &MainObjFile, + std::vector<std::unique_ptr<CoverageDataWithObjectFile>> *Data, + const std::set<std::string> &CoverageFiles) + : MainObjFile(MainObjFile), CoverageFiles(CoverageFiles) { + Data->swap(this->Coverage); + } + + const std::string MainObjFile; + std::vector<std::unique_ptr<CoverageDataWithObjectFile>> Coverage; + const std::set<std::string> CoverageFiles; }; + } // namespace int main(int argc, char **argv) { // Print stack trace if we signal out. - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. @@ -511,23 +1213,46 @@ int main(int argc, char **argv) { cl::ParseCommandLineOptions(argc, argv, "Sanitizer Coverage Processing Tool"); - auto CovData = CoverageData::readAndMerge(ClInputFiles); - FailIfError(CovData); - - switch (Action) { - case PrintAction: { + // -print doesn't need object files. + if (Action == PrintAction) { + auto CovData = CoverageData::readAndMerge(ClInputFiles); + FailIfError(CovData); CovData.get()->printAddrs(outs()); return 0; + } else if (Action == PrintCovPointsAction) { + // -print-coverage-points doesn't need coverage files. + for (const std::string &ObjFile : ClInputFiles) { + printCovPoints(ObjFile, outs()); + } + return 0; + } + + auto CovDataSet = CoverageDataSet::readCmdArguments(ClInputFiles); + FailIfError(CovDataSet); + + if (CovDataSet.get()->empty()) { + Fail("No coverage files specified."); } + + switch (Action) { case CoveredFunctionsAction: { - CovData.get()->printCoveredFunctions(outs()); + CovDataSet.get()->printCoveredFunctions(outs()); return 0; } case NotCoveredFunctionsAction: { - CovData.get()->printNotCoveredFunctions(outs()); + CovDataSet.get()->printNotCoveredFunctions(outs()); return 0; } + case HtmlReportAction: { + CovDataSet.get()->printReport(outs()); + return 0; + } + case StatsAction: { + CovDataSet.get()->printStats(outs()); + return 0; + } + case PrintAction: + case PrintCovPointsAction: + llvm_unreachable("unsupported action"); } - - llvm_unreachable("unsupported action"); } diff --git a/tools/sanstats/CMakeLists.txt b/tools/sanstats/CMakeLists.txt new file mode 100644 index 0000000000000..abfbdab3a0451 --- /dev/null +++ b/tools/sanstats/CMakeLists.txt @@ -0,0 +1,8 @@ +set(LLVM_LINK_COMPONENTS + Support + Symbolize + ) + +add_llvm_tool(sanstats + sanstats.cpp + ) diff --git a/tools/sanstats/sanstats.cpp b/tools/sanstats/sanstats.cpp new file mode 100644 index 0000000000000..b2216eab119e1 --- /dev/null +++ b/tools/sanstats/sanstats.cpp @@ -0,0 +1,138 @@ +//===- sanstats.cpp - Sanitizer statistics dumper -------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This tool dumps statistics information from files in the format produced +// by clang's -fsanitize-stats feature. +// +//===----------------------------------------------------------------------===// + +#include "llvm/DebugInfo/Symbolize/Symbolize.h" +#include "llvm/Support/CommandLine.h" +#include "llvm/Support/ErrorOr.h" +#include "llvm/Support/MemoryBuffer.h" +#include "llvm/Transforms/Utils/SanitizerStats.h" +#include <stdint.h> + +using namespace llvm; + +static cl::opt<std::string> ClInputFile(cl::Positional, cl::Required, + cl::desc("<filename>")); + +static cl::opt<bool> ClDemangle("demangle", cl::init(false), + cl::desc("Print demangled function name.")); + +inline uint64_t KindFromData(uint64_t Data, char SizeofPtr) { + return Data >> (SizeofPtr * 8 - kSanitizerStatKindBits); +} + +inline uint64_t CountFromData(uint64_t Data, char SizeofPtr) { + return Data & ((1ull << (SizeofPtr * 8 - kSanitizerStatKindBits)) - 1); +} + +uint64_t ReadLE(char Size, const char *Begin, const char *End) { + uint64_t Result = 0; + char Pos = 0; + while (Begin < End && Pos != Size) { + Result |= uint64_t(uint8_t(*Begin)) << (Pos * 8); + ++Begin; + ++Pos; + } + return Result; +} + +const char *ReadModule(char SizeofPtr, const char *Begin, const char *End) { + const char *FilenameBegin = Begin; + while (Begin != End && *Begin) + ++Begin; + if (Begin == End) + return nullptr; + StringRef Filename(FilenameBegin, Begin - FilenameBegin); + + ++Begin; + if (Begin == End) + return nullptr; + + symbolize::LLVMSymbolizer::Options SymbolizerOptions; + SymbolizerOptions.Demangle = ClDemangle; + SymbolizerOptions.UseSymbolTable = true; + symbolize::LLVMSymbolizer Symbolizer(SymbolizerOptions); + + while (1) { + uint64_t Addr = ReadLE(SizeofPtr, Begin, End); + Begin += SizeofPtr; + uint64_t Data = ReadLE(SizeofPtr, Begin, End); + Begin += SizeofPtr; + + if (Begin > End) + return nullptr; + if (Addr == 0 && Data == 0) + return Begin; + if (Begin == End) + return nullptr; + + if (Expected<DILineInfo> LineInfo = + Symbolizer.symbolizeCode(Filename, Addr)) { + llvm::outs() << LineInfo->FileName << ':' << LineInfo->Line << ' ' + << LineInfo->FunctionName << ' '; + } else { + logAllUnhandledErrors(LineInfo.takeError(), llvm::outs(), "<error> "); + } + + switch (KindFromData(Data, SizeofPtr)) { + case SanStat_CFI_VCall: + llvm::outs() << "cfi-vcall"; + break; + case SanStat_CFI_NVCall: + llvm::outs() << "cfi-nvcall"; + break; + case SanStat_CFI_DerivedCast: + llvm::outs() << "cfi-derived-cast"; + break; + case SanStat_CFI_UnrelatedCast: + llvm::outs() << "cfi-unrelated-cast"; + break; + case SanStat_CFI_ICall: + llvm::outs() << "cfi-icall"; + break; + default: + llvm::outs() << "<unknown>"; + break; + } + + llvm::outs() << " " << CountFromData(Data, SizeofPtr) << '\n'; + } +} + +int main(int argc, char **argv) { + cl::ParseCommandLineOptions(argc, argv, + "Sanitizer Statistics Processing Tool"); + + ErrorOr<std::unique_ptr<MemoryBuffer>> MBOrErr = + MemoryBuffer::getFile(ClInputFile, -1, false); + if (!MBOrErr) { + errs() << argv[0] << ": " << ClInputFile << ": " + << MBOrErr.getError().message() << '\n'; + return 1; + } + std::unique_ptr<MemoryBuffer> MB = std::move(MBOrErr.get()); + const char *Begin = MB->getBufferStart(), *End = MB->getBufferEnd(); + if (Begin == End) { + errs() << argv[0] << ": " << ClInputFile << ": short read\n"; + return 1; + } + char SizeofPtr = *Begin++; + while (Begin != End) { + Begin = ReadModule(SizeofPtr, Begin, End); + if (Begin == nullptr) { + errs() << argv[0] << ": " << ClInputFile << ": short read\n"; + return 1; + } + assert(Begin <= End); + } +} diff --git a/tools/verify-uselistorder/Makefile b/tools/verify-uselistorder/Makefile deleted file mode 100644 index 90d2aa89128b8..0000000000000 --- a/tools/verify-uselistorder/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- tools/verify-uselistorder/Makefile ------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL := ../.. -TOOLNAME := verify-uselistorder -LINK_COMPONENTS := AsmParser BitReader BitWriter Core IRReader Support - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS := 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/verify-uselistorder/verify-uselistorder.cpp b/tools/verify-uselistorder/verify-uselistorder.cpp index c3fec1326acb0..b19ecdaf6698f 100644 --- a/tools/verify-uselistorder/verify-uselistorder.cpp +++ b/tools/verify-uselistorder/verify-uselistorder.cpp @@ -191,6 +191,8 @@ ValueMapping::ValueMapping(const Module &M) { map(&G); for (const GlobalAlias &A : M.aliases()) map(&A); + for (const GlobalIFunc &IF : M.ifuncs()) + map(&IF); for (const Function &F : M) map(&F); @@ -200,6 +202,8 @@ ValueMapping::ValueMapping(const Module &M) { map(G.getInitializer()); for (const GlobalAlias &A : M.aliases()) map(A.getAliasee()); + for (const GlobalIFunc &IF : M.ifuncs()) + map(IF.getResolver()); for (const Function &F : M) { if (F.hasPrefixData()) map(F.getPrefixData()); @@ -463,6 +467,8 @@ static void changeUseLists(Module &M, Changer changeValueUseList) { changeValueUseList(&G); for (GlobalAlias &A : M.aliases()) changeValueUseList(&A); + for (GlobalIFunc &IF : M.ifuncs()) + changeValueUseList(&IF); for (Function &F : M) changeValueUseList(&F); @@ -472,6 +478,8 @@ static void changeUseLists(Module &M, Changer changeValueUseList) { changeValueUseList(G.getInitializer()); for (GlobalAlias &A : M.aliases()) changeValueUseList(A.getAliasee()); + for (GlobalIFunc &IF : M.ifuncs()) + changeValueUseList(IF.getResolver()); for (Function &F : M) { if (F.hasPrefixData()) changeValueUseList(F.getPrefixData()); @@ -518,14 +526,14 @@ static void reverseUseLists(Module &M) { } int main(int argc, char **argv) { - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); llvm::PrettyStackTraceProgram X(argc, argv); // Enable debug stream buffering. EnableDebugBuffering = true; llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. - LLVMContext &Context = getGlobalContext(); + LLVMContext Context; cl::ParseCommandLineOptions(argc, argv, "llvm tool to verify use-list order\n"); diff --git a/tools/xcode-toolchain/CMakeLists.txt b/tools/xcode-toolchain/CMakeLists.txt index 4eacbd320efcb..92b24581962b3 100644 --- a/tools/xcode-toolchain/CMakeLists.txt +++ b/tools/xcode-toolchain/CMakeLists.txt @@ -69,4 +69,24 @@ add_custom_target(install-xcode-toolchain COMMAND "${CMAKE_COMMAND}" -DCMAKE_INSTALL_PREFIX=${LLVMToolchainDir}/usr/ -P "${CMAKE_BINARY_DIR}/cmake_install.cmake" - ${cmake_3_2_USES_TERMINAL}) + USES_TERMINAL) + +if(LLVM_DISTRIBUTION_COMPONENTS) + if(CMAKE_CONFIGURATION_TYPES) + message(FATAL_ERROR "LLVM_DISTRIBUTION_COMPONENTS cannot be specified with multi-configuration generators (i.e. Xcode or Visual Studio)") + endif() + + add_custom_target(install-distribution-toolchain + DEPENDS ${LLVMToolchainDir}/ToolchainInfo.plist distribution) + + foreach(target ${LLVM_DISTRIBUTION_COMPONENTS}) + add_custom_target(install-distribution-${target} + DEPENDS ${target} + COMMAND "${CMAKE_COMMAND}" + -DCMAKE_INSTALL_COMPONENT=${target} + -DCMAKE_INSTALL_PREFIX=${LLVMToolchainDir}/usr/ + -P "${CMAKE_BINARY_DIR}/cmake_install.cmake" + USES_TERMINAL) + add_dependencies(install-distribution-toolchain install-distribution-${target}) + endforeach() +endif() diff --git a/tools/yaml2obj/CMakeLists.txt b/tools/yaml2obj/CMakeLists.txt index 52e9df4d766ae..885a69f5d3c36 100644 --- a/tools/yaml2obj/CMakeLists.txt +++ b/tools/yaml2obj/CMakeLists.txt @@ -1,6 +1,7 @@ set(LLVM_LINK_COMPONENTS MC Object + ObjectYAML Support ) @@ -8,4 +9,5 @@ add_llvm_tool(yaml2obj yaml2obj.cpp yaml2coff.cpp yaml2elf.cpp + yaml2macho.cpp ) diff --git a/tools/yaml2obj/Makefile b/tools/yaml2obj/Makefile deleted file mode 100644 index 912f0e31ae7c1..0000000000000 --- a/tools/yaml2obj/Makefile +++ /dev/null @@ -1,17 +0,0 @@ -##===- utils/yaml2obj/Makefile ----------------------------*- Makefile -*-===## -# -# The LLVM Compiler Infrastructure -# -# This file is distributed under the University of Illinois Open Source -# License. See LICENSE.TXT for details. -# -##===----------------------------------------------------------------------===## - -LEVEL = ../.. -TOOLNAME = yaml2obj -LINK_COMPONENTS := object - -# This tool has no plugins, optimize startup time. -TOOL_NO_EXPORTS = 1 - -include $(LEVEL)/Makefile.common diff --git a/tools/yaml2obj/yaml2coff.cpp b/tools/yaml2obj/yaml2coff.cpp index 61d985192552b..8f3f52179528d 100644 --- a/tools/yaml2obj/yaml2coff.cpp +++ b/tools/yaml2obj/yaml2coff.cpp @@ -14,12 +14,11 @@ #include "yaml2obj.h" #include "llvm/ADT/STLExtras.h" -#include "llvm/ADT/SmallString.h" #include "llvm/ADT/StringExtras.h" #include "llvm/ADT/StringMap.h" #include "llvm/ADT/StringSwitch.h" #include "llvm/Object/COFF.h" -#include "llvm/Object/COFFYAML.h" +#include "llvm/ObjectYAML/ObjectYAML.h" #include "llvm/Support/Endian.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/SourceMgr.h" @@ -76,14 +75,24 @@ struct COFFParser { unsigned Index = getStringIndex(Name); std::string str = utostr(Index); if (str.size() > 7) { - errs() << "String table got too large"; + errs() << "String table got too large\n"; return false; } Sec.Header.Name[0] = '/'; std::copy(str.begin(), str.end(), Sec.Header.Name + 1); } - Sec.Header.Characteristics |= (Log2_32(Sec.Alignment) + 1) << 20; + if (Sec.Alignment) { + if (Sec.Alignment > 8192) { + errs() << "Section alignment is too large\n"; + return false; + } + if (!isPowerOf2_32(Sec.Alignment)) { + errs() << "Section alignment is not a power of 2\n"; + return false; + } + Sec.Header.Characteristics |= (Log2_32(Sec.Alignment) + 1) << 20; + } } return true; } @@ -173,12 +182,12 @@ static bool layoutCOFF(COFFParser &CP) { // Assign each section data address consecutively. for (COFFYAML::Section &S : CP.Obj.Sections) { if (S.SectionData.binary_size() > 0) { - CurrentSectionDataOffset = RoundUpToAlignment( - CurrentSectionDataOffset, CP.isPE() ? CP.getFileAlignment() : 4); + CurrentSectionDataOffset = alignTo(CurrentSectionDataOffset, + CP.isPE() ? CP.getFileAlignment() : 4); S.Header.SizeOfRawData = S.SectionData.binary_size(); if (CP.isPE()) S.Header.SizeOfRawData = - RoundUpToAlignment(S.Header.SizeOfRawData, CP.getFileAlignment()); + alignTo(S.Header.SizeOfRawData, CP.getFileAlignment()); S.Header.PointerToRawData = CurrentSectionDataOffset; CurrentSectionDataOffset += S.Header.SizeOfRawData; if (!S.Relocations.empty()) { @@ -292,10 +301,9 @@ static uint32_t initializeOptionalHeader(COFFParser &CP, uint16_t Magic, T Heade Header->FileAlignment = CP.Obj.OptionalHeader->Header.FileAlignment; uint32_t SizeOfCode = 0, SizeOfInitializedData = 0, SizeOfUninitializedData = 0; - uint32_t SizeOfHeaders = RoundUpToAlignment( - CP.SectionTableStart + CP.SectionTableSize, Header->FileAlignment); - uint32_t SizeOfImage = - RoundUpToAlignment(SizeOfHeaders, Header->SectionAlignment); + uint32_t SizeOfHeaders = alignTo(CP.SectionTableStart + CP.SectionTableSize, + Header->FileAlignment); + uint32_t SizeOfImage = alignTo(SizeOfHeaders, Header->SectionAlignment); uint32_t BaseOfData = 0; for (const COFFYAML::Section &S : CP.Obj.Sections) { if (S.Header.Characteristics & COFF::IMAGE_SCN_CNT_CODE) @@ -309,8 +317,7 @@ static uint32_t initializeOptionalHeader(COFFParser &CP, uint16_t Magic, T Heade else if (S.Name.equals(".data")) BaseOfData = S.Header.VirtualAddress; // RVA if (S.Header.VirtualAddress) - SizeOfImage += - RoundUpToAlignment(S.Header.VirtualSize, Header->SectionAlignment); + SizeOfImage += alignTo(S.Header.VirtualSize, Header->SectionAlignment); } Header->SizeOfCode = SizeOfCode; Header->SizeOfInitializedData = SizeOfInitializedData; @@ -525,14 +532,7 @@ static bool writeCOFF(COFFParser &CP, raw_ostream &OS) { return true; } -int yaml2coff(yaml::Input &YIn, raw_ostream &Out) { - COFFYAML::Object Doc; - YIn >> Doc; - if (YIn.error()) { - errs() << "yaml2obj: Failed to parse YAML file!\n"; - return 1; - } - +int yaml2coff(llvm::COFFYAML::Object &Doc, raw_ostream &Out) { COFFParser CP(Doc); if (!CP.parse()) { errs() << "yaml2obj: Failed to parse YAML file!\n"; diff --git a/tools/yaml2obj/yaml2elf.cpp b/tools/yaml2obj/yaml2elf.cpp index d03986e9c88d5..c98093431a7ef 100644 --- a/tools/yaml2obj/yaml2elf.cpp +++ b/tools/yaml2obj/yaml2elf.cpp @@ -16,7 +16,7 @@ #include "llvm/ADT/ArrayRef.h" #include "llvm/MC/StringTableBuilder.h" #include "llvm/Object/ELFObjectFile.h" -#include "llvm/Object/ELFYAML.h" +#include "llvm/ObjectYAML/ELFYAML.h" #include "llvm/Support/ELF.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/YAMLTraits.h" @@ -38,7 +38,7 @@ class ContiguousBlobAccumulator { if (Align == 0) Align = 1; uint64_t CurrentOffset = InitialOffset + OS.tell(); - uint64_t AlignedOffset = RoundUpToAlignment(CurrentOffset, Align); + uint64_t AlignedOffset = alignTo(CurrentOffset, Align); for (; CurrentOffset != AlignedOffset; ++CurrentOffset) OS.write('\0'); return AlignedOffset; // == CurrentOffset; @@ -558,13 +558,7 @@ static bool isLittleEndian(const ELFYAML::Object &Doc) { return Doc.Header.Data == ELFYAML::ELF_ELFDATA(ELF::ELFDATA2LSB); } -int yaml2elf(yaml::Input &YIn, raw_ostream &Out) { - ELFYAML::Object Doc; - YIn >> Doc; - if (YIn.error()) { - errs() << "yaml2obj: Failed to parse YAML file!\n"; - return 1; - } +int yaml2elf(llvm::ELFYAML::Object &Doc, raw_ostream &Out) { using object::ELFType; typedef ELFType<support::little, true> LE64; typedef ELFType<support::big, true> BE64; diff --git a/tools/yaml2obj/yaml2macho.cpp b/tools/yaml2obj/yaml2macho.cpp new file mode 100644 index 0000000000000..fb29e206be35e --- /dev/null +++ b/tools/yaml2obj/yaml2macho.cpp @@ -0,0 +1,539 @@ +//===- yaml2macho - Convert YAML to a Mach object file --------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +/// +/// \file +/// \brief The Mach component of yaml2obj. +/// +//===----------------------------------------------------------------------===// + +#include "yaml2obj.h" +#include "llvm/ObjectYAML/ObjectYAML.h" +#include "llvm/Support/Error.h" +#include "llvm/Support/LEB128.h" +#include "llvm/Support/MachO.h" +#include "llvm/Support/YAMLTraits.h" +#include "llvm/Support/raw_ostream.h" + +#include "llvm/Support/Format.h" + +using namespace llvm; + +namespace { + +class MachOWriter { +public: + MachOWriter(MachOYAML::Object &Obj) : Obj(Obj), is64Bit(true), fileStart(0) { + is64Bit = Obj.Header.magic == MachO::MH_MAGIC_64 || + Obj.Header.magic == MachO::MH_CIGAM_64; + memset(reinterpret_cast<void *>(&Header), 0, sizeof(MachO::mach_header_64)); + } + + Error writeMachO(raw_ostream &OS); + +private: + Error writeHeader(raw_ostream &OS); + Error writeLoadCommands(raw_ostream &OS); + Error writeSectionData(raw_ostream &OS); + Error writeLinkEditData(raw_ostream &OS); + void writeBindOpcodes(raw_ostream &OS, + std::vector<MachOYAML::BindOpcode> &BindOpcodes); + // LinkEdit writers + Error writeRebaseOpcodes(raw_ostream &OS); + Error writeBasicBindOpcodes(raw_ostream &OS); + Error writeWeakBindOpcodes(raw_ostream &OS); + Error writeLazyBindOpcodes(raw_ostream &OS); + Error writeNameList(raw_ostream &OS); + Error writeStringTable(raw_ostream &OS); + Error writeExportTrie(raw_ostream &OS); + + void dumpExportEntry(raw_ostream &OS, MachOYAML::ExportEntry &Entry); + void ZeroToOffset(raw_ostream &OS, size_t offset); + + MachOYAML::Object &Obj; + bool is64Bit; + uint64_t fileStart; + + MachO::mach_header_64 Header; +}; + +Error MachOWriter::writeMachO(raw_ostream &OS) { + fileStart = OS.tell(); + if (auto Err = writeHeader(OS)) + return Err; + if (auto Err = writeLoadCommands(OS)) + return Err; + if (auto Err = writeSectionData(OS)) + return Err; + return Error::success(); +} + +Error MachOWriter::writeHeader(raw_ostream &OS) { + Header.magic = Obj.Header.magic; + Header.cputype = Obj.Header.cputype; + Header.cpusubtype = Obj.Header.cpusubtype; + Header.filetype = Obj.Header.filetype; + Header.ncmds = Obj.Header.ncmds; + Header.sizeofcmds = Obj.Header.sizeofcmds; + Header.flags = Obj.Header.flags; + Header.reserved = Obj.Header.reserved; + + auto header_size = + is64Bit ? sizeof(MachO::mach_header_64) : sizeof(MachO::mach_header); + OS.write((const char *)&Header, header_size); + + return Error::success(); +} + +template <typename SectionType> +SectionType constructSection(MachOYAML::Section Sec) { + SectionType TempSec; + memcpy(reinterpret_cast<void *>(&TempSec.sectname[0]), &Sec.sectname[0], 16); + memcpy(reinterpret_cast<void *>(&TempSec.segname[0]), &Sec.segname[0], 16); + TempSec.addr = Sec.addr; + TempSec.size = Sec.size; + TempSec.offset = Sec.offset; + TempSec.align = Sec.align; + TempSec.reloff = Sec.reloff; + TempSec.nreloc = Sec.nreloc; + TempSec.flags = Sec.flags; + TempSec.reserved1 = Sec.reserved1; + TempSec.reserved2 = Sec.reserved2; + return TempSec; +} + +template <typename StructType> +size_t writeLoadCommandData(MachOYAML::LoadCommand &LC, raw_ostream &OS) { + return 0; +} + +template <> +size_t writeLoadCommandData<MachO::segment_command>(MachOYAML::LoadCommand &LC, + raw_ostream &OS) { + size_t BytesWritten = 0; + for (const auto &Sec : LC.Sections) { + auto TempSec = constructSection<MachO::section>(Sec); + OS.write(reinterpret_cast<const char *>(&(TempSec)), + sizeof(MachO::section)); + BytesWritten += sizeof(MachO::section); + } + return BytesWritten; +} + +template <> +size_t +writeLoadCommandData<MachO::segment_command_64>(MachOYAML::LoadCommand &LC, + raw_ostream &OS) { + size_t BytesWritten = 0; + for (const auto &Sec : LC.Sections) { + auto TempSec = constructSection<MachO::section_64>(Sec); + TempSec.reserved3 = Sec.reserved3; + OS.write(reinterpret_cast<const char *>(&(TempSec)), + sizeof(MachO::section_64)); + BytesWritten += sizeof(MachO::section_64); + } + return BytesWritten; +} + +size_t writePayloadString(MachOYAML::LoadCommand &LC, raw_ostream &OS) { + size_t BytesWritten = 0; + if (!LC.PayloadString.empty()) { + OS.write(LC.PayloadString.c_str(), LC.PayloadString.length()); + BytesWritten = LC.PayloadString.length(); + } + return BytesWritten; +} + +template <> +size_t writeLoadCommandData<MachO::dylib_command>(MachOYAML::LoadCommand &LC, + raw_ostream &OS) { + return writePayloadString(LC, OS); +} + +template <> +size_t writeLoadCommandData<MachO::dylinker_command>(MachOYAML::LoadCommand &LC, + raw_ostream &OS) { + return writePayloadString(LC, OS); +} + +template <> +size_t writeLoadCommandData<MachO::rpath_command>(MachOYAML::LoadCommand &LC, + raw_ostream &OS) { + return writePayloadString(LC, OS); +} + +void ZeroFillBytes(raw_ostream &OS, size_t Size) { + std::vector<uint8_t> FillData; + FillData.insert(FillData.begin(), Size, 0); + OS.write(reinterpret_cast<char *>(FillData.data()), Size); +} + +void Fill(raw_ostream &OS, size_t Size, uint32_t Data) { + std::vector<uint32_t> FillData; + FillData.insert(FillData.begin(), (Size / 4) + 1, Data); + OS.write(reinterpret_cast<char *>(FillData.data()), Size); +} + +void MachOWriter::ZeroToOffset(raw_ostream &OS, size_t Offset) { + auto currOffset = OS.tell() - fileStart; + if (currOffset < Offset) + ZeroFillBytes(OS, Offset - currOffset); +} + +Error MachOWriter::writeLoadCommands(raw_ostream &OS) { + for (auto &LC : Obj.LoadCommands) { + size_t BytesWritten = 0; +#define HANDLE_LOAD_COMMAND(LCName, LCValue, LCStruct) \ + case MachO::LCName: \ + OS.write(reinterpret_cast<const char *>(&(LC.Data.LCStruct##_data)), \ + sizeof(MachO::LCStruct)); \ + BytesWritten = sizeof(MachO::LCStruct); \ + BytesWritten += writeLoadCommandData<MachO::LCStruct>(LC, OS); \ + break; + + switch (LC.Data.load_command_data.cmd) { + default: + OS.write(reinterpret_cast<const char *>(&(LC.Data.load_command_data)), + sizeof(MachO::load_command)); + BytesWritten = sizeof(MachO::load_command); + BytesWritten += writeLoadCommandData<MachO::load_command>(LC, OS); + break; +#include "llvm/Support/MachO.def" + } + + if (LC.PayloadBytes.size() > 0) { + OS.write(reinterpret_cast<const char *>(LC.PayloadBytes.data()), + LC.PayloadBytes.size()); + BytesWritten += LC.PayloadBytes.size(); + } + + if (LC.ZeroPadBytes > 0) { + ZeroFillBytes(OS, LC.ZeroPadBytes); + BytesWritten += LC.ZeroPadBytes; + } + + // Fill remaining bytes with 0. This will only get hit in partially + // specified test cases. + auto BytesRemaining = LC.Data.load_command_data.cmdsize - BytesWritten; + if (BytesRemaining > 0) { + ZeroFillBytes(OS, BytesRemaining); + } + } + return Error::success(); +} + +Error MachOWriter::writeSectionData(raw_ostream &OS) { + for (auto &LC : Obj.LoadCommands) { + switch (LC.Data.load_command_data.cmd) { + case MachO::LC_SEGMENT: + case MachO::LC_SEGMENT_64: + auto currOffset = OS.tell() - fileStart; + auto segname = LC.Data.segment_command_data.segname; + uint64_t segOff = is64Bit ? LC.Data.segment_command_64_data.fileoff + : LC.Data.segment_command_data.fileoff; + + if (0 == strncmp(&segname[0], "__LINKEDIT", 16)) { + if (auto Err = writeLinkEditData(OS)) + return Err; + } else { + // Zero Fill any data between the end of the last thing we wrote and the + // start of this section. + if (currOffset < segOff) { + ZeroFillBytes(OS, segOff - currOffset); + } + + for (auto &Sec : LC.Sections) { + // Zero Fill any data between the end of the last thing we wrote and + // the + // start of this section. + assert( + OS.tell() - fileStart <= Sec.offset && + "Wrote too much data somewhere, section offsets don't line up."); + currOffset = OS.tell() - fileStart; + if (currOffset < Sec.offset) { + ZeroFillBytes(OS, Sec.offset - currOffset); + } + + // Fills section data with 0xDEADBEEF + Fill(OS, Sec.size, 0xDEADBEEFu); + } + } + uint64_t segSize = is64Bit ? LC.Data.segment_command_64_data.filesize + : LC.Data.segment_command_data.filesize; + ZeroToOffset(OS, segOff + segSize); + break; + } + } + return Error::success(); +} + +void MachOWriter::writeBindOpcodes( + raw_ostream &OS, std::vector<MachOYAML::BindOpcode> &BindOpcodes) { + + for (auto Opcode : BindOpcodes) { + uint8_t OpByte = Opcode.Opcode | Opcode.Imm; + OS.write(reinterpret_cast<char *>(&OpByte), 1); + for (auto Data : Opcode.ULEBExtraData) { + encodeULEB128(Data, OS); + } + for (auto Data : Opcode.SLEBExtraData) { + encodeSLEB128(Data, OS); + } + if (!Opcode.Symbol.empty()) { + OS.write(Opcode.Symbol.data(), Opcode.Symbol.size()); + OS.write('\0'); + } + } +} + +void MachOWriter::dumpExportEntry(raw_ostream &OS, + MachOYAML::ExportEntry &Entry) { + encodeSLEB128(Entry.TerminalSize, OS); + if (Entry.TerminalSize > 0) { + encodeSLEB128(Entry.Flags, OS); + if (Entry.Flags & MachO::EXPORT_SYMBOL_FLAGS_REEXPORT) { + encodeSLEB128(Entry.Other, OS); + OS << Entry.ImportName; + OS.write('\0'); + } else { + encodeSLEB128(Entry.Address, OS); + if (Entry.Flags & MachO::EXPORT_SYMBOL_FLAGS_STUB_AND_RESOLVER) + encodeSLEB128(Entry.Other, OS); + } + } + OS.write(static_cast<uint8_t>(Entry.Children.size())); + for (auto EE : Entry.Children) { + OS << EE.Name; + OS.write('\0'); + encodeSLEB128(EE.NodeOffset, OS); + } + for (auto EE : Entry.Children) + dumpExportEntry(OS, EE); +} + +Error MachOWriter::writeExportTrie(raw_ostream &OS) { + dumpExportEntry(OS, Obj.LinkEdit.ExportTrie); + return Error::success(); +} + +template <typename NListType> +void writeNListEntry(MachOYAML::NListEntry &NLE, raw_ostream &OS) { + NListType ListEntry; + ListEntry.n_strx = NLE.n_strx; + ListEntry.n_type = NLE.n_type; + ListEntry.n_sect = NLE.n_sect; + ListEntry.n_desc = NLE.n_desc; + ListEntry.n_value = NLE.n_value; + OS.write(reinterpret_cast<const char *>(&ListEntry), sizeof(NListType)); +} + +Error MachOWriter::writeLinkEditData(raw_ostream &OS) { + typedef Error (MachOWriter::*writeHandler)(raw_ostream &); + typedef std::pair<uint64_t, writeHandler> writeOperation; + std::vector<writeOperation> WriteQueue; + + MachO::dyld_info_command *DyldInfoOnlyCmd = 0; + MachO::symtab_command *SymtabCmd = 0; + for (auto &LC : Obj.LoadCommands) { + switch (LC.Data.load_command_data.cmd) { + case MachO::LC_SYMTAB: + SymtabCmd = &LC.Data.symtab_command_data; + WriteQueue.push_back( + std::make_pair(SymtabCmd->symoff, &MachOWriter::writeNameList)); + WriteQueue.push_back( + std::make_pair(SymtabCmd->stroff, &MachOWriter::writeStringTable)); + break; + case MachO::LC_DYLD_INFO_ONLY: + DyldInfoOnlyCmd = &LC.Data.dyld_info_command_data; + WriteQueue.push_back(std::make_pair(DyldInfoOnlyCmd->rebase_off, + &MachOWriter::writeRebaseOpcodes)); + WriteQueue.push_back(std::make_pair(DyldInfoOnlyCmd->bind_off, + &MachOWriter::writeBasicBindOpcodes)); + WriteQueue.push_back(std::make_pair(DyldInfoOnlyCmd->weak_bind_off, + &MachOWriter::writeWeakBindOpcodes)); + WriteQueue.push_back(std::make_pair(DyldInfoOnlyCmd->lazy_bind_off, + &MachOWriter::writeLazyBindOpcodes)); + WriteQueue.push_back(std::make_pair(DyldInfoOnlyCmd->export_off, + &MachOWriter::writeExportTrie)); + break; + } + } + + std::sort(WriteQueue.begin(), WriteQueue.end(), + [](const writeOperation &a, const writeOperation &b) { + return a.first < b.first; + }); + + for (auto writeOp : WriteQueue) { + ZeroToOffset(OS, writeOp.first); + if (auto Err = (this->*writeOp.second)(OS)) + return Err; + } + + return Error::success(); +} + +Error MachOWriter::writeRebaseOpcodes(raw_ostream &OS) { + MachOYAML::LinkEditData &LinkEdit = Obj.LinkEdit; + + for (auto Opcode : LinkEdit.RebaseOpcodes) { + uint8_t OpByte = Opcode.Opcode | Opcode.Imm; + OS.write(reinterpret_cast<char *>(&OpByte), 1); + for (auto Data : Opcode.ExtraData) { + encodeULEB128(Data, OS); + } + } + return Error::success(); +} + +Error MachOWriter::writeBasicBindOpcodes(raw_ostream &OS) { + writeBindOpcodes(OS, Obj.LinkEdit.BindOpcodes); + return Error::success(); +} + +Error MachOWriter::writeWeakBindOpcodes(raw_ostream &OS) { + writeBindOpcodes(OS, Obj.LinkEdit.WeakBindOpcodes); + return Error::success(); +} + +Error MachOWriter::writeLazyBindOpcodes(raw_ostream &OS) { + writeBindOpcodes(OS, Obj.LinkEdit.LazyBindOpcodes); + return Error::success(); +} + +Error MachOWriter::writeNameList(raw_ostream &OS) { + for (auto NLE : Obj.LinkEdit.NameList) { + if (is64Bit) + writeNListEntry<MachO::nlist_64>(NLE, OS); + else + writeNListEntry<MachO::nlist>(NLE, OS); + } + return Error::success(); +} + +Error MachOWriter::writeStringTable(raw_ostream &OS) { + for (auto Str : Obj.LinkEdit.StringTable) { + OS.write(Str.data(), Str.size()); + OS.write('\0'); + } + return Error::success(); +} + +class UniversalWriter { +public: + UniversalWriter(yaml::YamlObjectFile &ObjectFile) + : ObjectFile(ObjectFile), fileStart(0) {} + + Error writeMachO(raw_ostream &OS); + +private: + Error writeFatHeader(raw_ostream &OS); + Error writeFatArchs(raw_ostream &OS); + + void ZeroToOffset(raw_ostream &OS, size_t offset); + + yaml::YamlObjectFile &ObjectFile; + uint64_t fileStart; +}; + +Error UniversalWriter::writeMachO(raw_ostream &OS) { + fileStart = OS.tell(); + if (ObjectFile.MachO) { + MachOWriter Writer(*ObjectFile.MachO); + return Writer.writeMachO(OS); + } + if (auto Err = writeFatHeader(OS)) + return Err; + if (auto Err = writeFatArchs(OS)) + return Err; + auto &FatFile = *ObjectFile.FatMachO; + assert(FatFile.FatArchs.size() == FatFile.Slices.size()); + for (size_t i = 0; i < FatFile.Slices.size(); i++) { + ZeroToOffset(OS, FatFile.FatArchs[i].offset); + MachOWriter Writer(FatFile.Slices[i]); + if (auto Err = Writer.writeMachO(OS)) + return Err; + auto SliceEnd = FatFile.FatArchs[i].offset + FatFile.FatArchs[i].size; + ZeroToOffset(OS, SliceEnd); + } + return Error::success(); +} + +Error UniversalWriter::writeFatHeader(raw_ostream &OS) { + auto &FatFile = *ObjectFile.FatMachO; + MachO::fat_header header; + header.magic = FatFile.Header.magic; + header.nfat_arch = FatFile.Header.nfat_arch; + if (sys::IsLittleEndianHost) + swapStruct(header); + OS.write(reinterpret_cast<const char *>(&header), sizeof(MachO::fat_header)); + return Error::success(); +} + +template <typename FatArchType> +FatArchType constructFatArch(MachOYAML::FatArch &Arch) { + FatArchType FatArch; + FatArch.cputype = Arch.cputype; + FatArch.cpusubtype = Arch.cpusubtype; + FatArch.offset = Arch.offset; + FatArch.size = Arch.size; + FatArch.align = Arch.align; + return FatArch; +} + +template <typename StructType> +void writeFatArch(MachOYAML::FatArch &LC, raw_ostream &OS) {} + +template <> +void writeFatArch<MachO::fat_arch>(MachOYAML::FatArch &Arch, raw_ostream &OS) { + auto FatArch = constructFatArch<MachO::fat_arch>(Arch); + if (sys::IsLittleEndianHost) + swapStruct(FatArch); + OS.write(reinterpret_cast<const char *>(&FatArch), sizeof(MachO::fat_arch)); +} + +template <> +void writeFatArch<MachO::fat_arch_64>(MachOYAML::FatArch &Arch, + raw_ostream &OS) { + auto FatArch = constructFatArch<MachO::fat_arch_64>(Arch); + FatArch.reserved = Arch.reserved; + if (sys::IsLittleEndianHost) + swapStruct(FatArch); + OS.write(reinterpret_cast<const char *>(&FatArch), + sizeof(MachO::fat_arch_64)); +} + +Error UniversalWriter::writeFatArchs(raw_ostream &OS) { + auto &FatFile = *ObjectFile.FatMachO; + bool is64Bit = FatFile.Header.magic == MachO::FAT_MAGIC_64; + for (auto Arch : FatFile.FatArchs) { + if (is64Bit) + writeFatArch<MachO::fat_arch_64>(Arch, OS); + else + writeFatArch<MachO::fat_arch>(Arch, OS); + } + + return Error::success(); +} + +void UniversalWriter::ZeroToOffset(raw_ostream &OS, size_t Offset) { + auto currOffset = OS.tell() - fileStart; + if (currOffset < Offset) + ZeroFillBytes(OS, Offset - currOffset); +} + +} // end anonymous namespace + +int yaml2macho(yaml::YamlObjectFile &Doc, raw_ostream &Out) { + UniversalWriter Writer(Doc); + if (auto Err = Writer.writeMachO(Out)) { + errs() << toString(std::move(Err)); + return 1; + } + return 0; +} diff --git a/tools/yaml2obj/yaml2obj.cpp b/tools/yaml2obj/yaml2obj.cpp index af4d868906724..f746d84a38985 100644 --- a/tools/yaml2obj/yaml2obj.cpp +++ b/tools/yaml2obj/yaml2obj.cpp @@ -16,6 +16,7 @@ #include "yaml2obj.h" #include "llvm/ADT/StringExtras.h" +#include "llvm/ObjectYAML/ObjectYAML.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/ManagedStatic.h" @@ -32,27 +33,6 @@ using namespace llvm; static cl::opt<std::string> Input(cl::Positional, cl::desc("<input>"), cl::init("-")); -// TODO: The "right" way to tell what kind of object file a given YAML file -// corresponds to is to look at YAML "tags" (e.g. `!Foo`). Then, different -// tags (`!ELF`, `!COFF`, etc.) would be used to discriminate between them. -// Interpreting the tags is needed eventually for when writing test cases, -// so that we can e.g. have `!Archive` contain a sequence of `!ELF`, and -// just Do The Right Thing. However, interpreting these tags and acting on -// them appropriately requires some work in the YAML parser and the YAMLIO -// library. -enum YAMLObjectFormat { - YOF_COFF, - YOF_ELF -}; - -cl::opt<YAMLObjectFormat> Format( - "format", - cl::desc("Interpret input as this type of object file"), - cl::values( - clEnumValN(YOF_COFF, "coff", "COFF object file format"), - clEnumValN(YOF_ELF, "elf", "ELF object file format"), - clEnumValEnd)); - cl::opt<unsigned> DocNum("docnum", cl::init(1), cl::desc("Read specified document from input (default = 1)")); @@ -60,14 +40,26 @@ DocNum("docnum", cl::init(1), static cl::opt<std::string> OutputFilename("o", cl::desc("Output filename"), cl::value_desc("filename")); -typedef int (*ConvertFuncPtr)(yaml::Input & YIn, raw_ostream &Out); - -static int convertYAML(yaml::Input &YIn, raw_ostream &Out, - ConvertFuncPtr Convert) { +static int convertYAML(yaml::Input &YIn, raw_ostream &Out) { unsigned CurDocNum = 0; do { - if (++CurDocNum == DocNum) - return Convert(YIn, Out); + if (++CurDocNum == DocNum) { + yaml::YamlObjectFile Doc; + YIn >> Doc; + if (YIn.error()) { + errs() << "yaml2obj: Failed to parse YAML file!\n"; + return 1; + } + + if (Doc.Elf) + return yaml2elf(*Doc.Elf, Out); + if (Doc.Coff) + return yaml2coff(*Doc.Coff, Out); + if (Doc.MachO || Doc.FatMachO) + return yaml2macho(Doc, Out); + errs() << "yaml2obj: Unknown document type!\n"; + return 1; + } } while (YIn.nextDocument()); errs() << "yaml2obj: Cannot find the " << DocNum @@ -77,7 +69,7 @@ static int convertYAML(yaml::Input &YIn, raw_ostream &Out, int main(int argc, char **argv) { cl::ParseCommandLineOptions(argc, argv); - sys::PrintStackTraceOnErrorSignal(); + sys::PrintStackTraceOnErrorSignal(argv[0]); PrettyStackTraceProgram X(argc, argv); llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. @@ -97,19 +89,9 @@ int main(int argc, char **argv) { if (!Buf) return 1; - ConvertFuncPtr Convert = nullptr; - if (Format == YOF_COFF) - Convert = yaml2coff; - else if (Format == YOF_ELF) - Convert = yaml2elf; - else { - errs() << "Not yet implemented\n"; - return 1; - } - yaml::Input YIn(Buf.get()->getBuffer()); - int Res = convertYAML(YIn, Out->os(), Convert); + int Res = convertYAML(YIn, Out->os()); if (Res == 0) Out->keep(); diff --git a/tools/yaml2obj/yaml2obj.h b/tools/yaml2obj/yaml2obj.h index 7290a9af2c649..b5025e860bd75 100644 --- a/tools/yaml2obj/yaml2obj.h +++ b/tools/yaml2obj/yaml2obj.h @@ -14,11 +14,23 @@ namespace llvm { class raw_ostream; + +namespace COFFYAML { +struct Object; +} + +namespace ELFYAML { +struct Object; +} + namespace yaml { class Input; +struct YamlObjectFile; } } -int yaml2coff(llvm::yaml::Input &YIn, llvm::raw_ostream &Out); -int yaml2elf(llvm::yaml::Input &YIn, llvm::raw_ostream &Out); + +int yaml2coff(llvm::COFFYAML::Object &Doc, llvm::raw_ostream &Out); +int yaml2elf(llvm::ELFYAML::Object &Doc, llvm::raw_ostream &Out); +int yaml2macho(llvm::yaml::YamlObjectFile &Doc, llvm::raw_ostream &Out); #endif |