diff options
Diffstat (limited to 'lib/Driver')
-rw-r--r-- | lib/Driver/Action.cpp | 21 | ||||
-rw-r--r-- | lib/Driver/Driver.cpp | 374 | ||||
-rw-r--r-- | lib/Driver/Job.cpp | 12 | ||||
-rw-r--r-- | lib/Driver/MSVCToolChain.cpp | 3 | ||||
-rw-r--r-- | lib/Driver/MinGWToolChain.cpp | 60 | ||||
-rw-r--r-- | lib/Driver/SanitizerArgs.cpp | 15 | ||||
-rw-r--r-- | lib/Driver/ToolChain.cpp | 2 | ||||
-rw-r--r-- | lib/Driver/ToolChains.cpp | 78 | ||||
-rw-r--r-- | lib/Driver/ToolChains.h | 17 | ||||
-rw-r--r-- | lib/Driver/Tools.cpp | 322 | ||||
-rw-r--r-- | lib/Driver/Types.cpp | 21 |
11 files changed, 676 insertions, 249 deletions
diff --git a/lib/Driver/Action.cpp b/lib/Driver/Action.cpp index 360dbee..3219dc1 100644 --- a/lib/Driver/Action.cpp +++ b/lib/Driver/Action.cpp @@ -24,6 +24,8 @@ const char *Action::getClassName(ActionClass AC) { switch (AC) { case InputClass: return "input"; case BindArchClass: return "bind-arch"; + case CudaDeviceClass: return "cuda-device"; + case CudaHostClass: return "cuda-host"; case PreprocessJobClass: return "preprocessor"; case PrecompileJobClass: return "precompiler"; case AnalyzeJobClass: return "analyzer"; @@ -53,6 +55,25 @@ BindArchAction::BindArchAction(std::unique_ptr<Action> Input, const char *_ArchName) : Action(BindArchClass, std::move(Input)), ArchName(_ArchName) {} +void CudaDeviceAction::anchor() {} + +CudaDeviceAction::CudaDeviceAction(std::unique_ptr<Action> Input, + const char *ArchName, bool AtTopLevel) + : Action(CudaDeviceClass, std::move(Input)), GpuArchName(ArchName), + AtTopLevel(AtTopLevel) {} + +void CudaHostAction::anchor() {} + +CudaHostAction::CudaHostAction(std::unique_ptr<Action> Input, + const ActionList &_DeviceActions) + : Action(CudaHostClass, std::move(Input)), DeviceActions(_DeviceActions) {} + +CudaHostAction::~CudaHostAction() { + for (iterator it = DeviceActions.begin(), ie = DeviceActions.end(); it != ie; + ++it) + delete *it; +} + void JobAction::anchor() {} JobAction::JobAction(ActionClass Kind, std::unique_ptr<Action> Input, diff --git a/lib/Driver/Driver.cpp b/lib/Driver/Driver.cpp index b9dc35d..180c412 100644 --- a/lib/Driver/Driver.cpp +++ b/lib/Driver/Driver.cpp @@ -174,8 +174,10 @@ phases::ID Driver::getFinalPhase(const DerivedArgList &DAL, } else if ((PhaseArg = DAL.getLastArg(options::OPT_S))) { FinalPhase = phases::Backend; - // -c only runs up to the assembler. - } else if ((PhaseArg = DAL.getLastArg(options::OPT_c))) { + // -c and partial CUDA compilations only run up to the assembler. + } else if ((PhaseArg = DAL.getLastArg(options::OPT_c)) || + (PhaseArg = DAL.getLastArg(options::OPT_cuda_device_only)) || + (PhaseArg = DAL.getLastArg(options::OPT_cuda_host_only))) { FinalPhase = phases::Assemble; // Otherwise do everything. @@ -281,6 +283,83 @@ DerivedArgList *Driver::TranslateInputArgs(const InputArgList &Args) const { return DAL; } +/// \brief Compute target triple from args. +/// +/// This routine provides the logic to compute a target triple from various +/// args passed to the driver and the default triple string. +static llvm::Triple computeTargetTriple(StringRef DefaultTargetTriple, + const ArgList &Args, + StringRef DarwinArchName = "") { + // FIXME: Already done in Compilation *Driver::BuildCompilation + if (const Arg *A = Args.getLastArg(options::OPT_target)) + DefaultTargetTriple = A->getValue(); + + llvm::Triple Target(llvm::Triple::normalize(DefaultTargetTriple)); + + // Handle Apple-specific options available here. + if (Target.isOSBinFormatMachO()) { + // If an explict Darwin arch name is given, that trumps all. + if (!DarwinArchName.empty()) { + tools::darwin::setTripleTypeForMachOArchName(Target, DarwinArchName); + return Target; + } + + // Handle the Darwin '-arch' flag. + if (Arg *A = Args.getLastArg(options::OPT_arch)) { + StringRef ArchName = A->getValue(); + tools::darwin::setTripleTypeForMachOArchName(Target, ArchName); + } + } + + // Handle pseudo-target flags '-mlittle-endian'/'-EL' and + // '-mbig-endian'/'-EB'. + if (Arg *A = Args.getLastArg(options::OPT_mlittle_endian, + options::OPT_mbig_endian)) { + if (A->getOption().matches(options::OPT_mlittle_endian)) { + llvm::Triple LE = Target.getLittleEndianArchVariant(); + if (LE.getArch() != llvm::Triple::UnknownArch) + Target = std::move(LE); + } else { + llvm::Triple BE = Target.getBigEndianArchVariant(); + if (BE.getArch() != llvm::Triple::UnknownArch) + Target = std::move(BE); + } + } + + // Skip further flag support on OSes which don't support '-m32' or '-m64'. + if (Target.getArchName() == "tce" || Target.getOS() == llvm::Triple::Minix) + return Target; + + // Handle pseudo-target flags '-m64', '-mx32', '-m32' and '-m16'. + if (Arg *A = Args.getLastArg(options::OPT_m64, options::OPT_mx32, + options::OPT_m32, options::OPT_m16)) { + llvm::Triple::ArchType AT = llvm::Triple::UnknownArch; + + if (A->getOption().matches(options::OPT_m64)) { + AT = Target.get64BitArchVariant().getArch(); + if (Target.getEnvironment() == llvm::Triple::GNUX32) + Target.setEnvironment(llvm::Triple::GNU); + } else if (A->getOption().matches(options::OPT_mx32) && + Target.get64BitArchVariant().getArch() == llvm::Triple::x86_64) { + AT = llvm::Triple::x86_64; + Target.setEnvironment(llvm::Triple::GNUX32); + } else if (A->getOption().matches(options::OPT_m32)) { + AT = Target.get32BitArchVariant().getArch(); + if (Target.getEnvironment() == llvm::Triple::GNUX32) + Target.setEnvironment(llvm::Triple::GNU); + } else if (A->getOption().matches(options::OPT_m16) && + Target.get32BitArchVariant().getArch() == llvm::Triple::x86) { + AT = llvm::Triple::x86; + Target.setEnvironment(llvm::Triple::CODE16); + } + + if (AT != llvm::Triple::UnknownArch && AT != Target.getArch()) + Target.setArch(AT); + } + + return Target; +} + Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) { llvm::PrettyStackTraceString CrashInfo("Compilation construction"); @@ -367,7 +446,8 @@ Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) { DerivedArgList *TranslatedArgs = TranslateInputArgs(*UArgs); // Owned by the host. - const ToolChain &TC = getToolChain(*UArgs); + const ToolChain &TC = + getToolChain(*UArgs, computeTargetTriple(DefaultTargetTriple, *UArgs)); // The compilation takes ownership of Args. Compilation *C = new Compilation(*this, TC, UArgs.release(), TranslatedArgs); @@ -398,6 +478,19 @@ Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) { return C; } +static void printArgList(raw_ostream &OS, const llvm::opt::ArgList &Args) { + llvm::opt::ArgStringList ASL; + for (const auto *A : Args) + A->render(Args, ASL); + + for (auto I = ASL.begin(), E = ASL.end(); I != E; ++I) { + if (I != ASL.begin()) + OS << ' '; + Command::printArg(OS, *I, true); + } + OS << '\n'; +} + // When clang crashes, produce diagnostic information including the fully // preprocessed source file(s). Request that the developer attach the // diagnostic information to a bug report. @@ -546,7 +639,9 @@ void Driver::generateCompilationDiagnostics(Compilation &C, << "Error generating run script: " + Script + " " + EC.message(); } else { ScriptOS << "# Crash reproducer for " << getClangFullVersion() << "\n" - << "# Original command: "; + << "# Driver args: "; + printArgList(ScriptOS, C.getInputArgs()); + ScriptOS << "# Original command: "; Cmd.Print(ScriptOS, "\n", /*Quote=*/true); Cmd.Print(ScriptOS, "\n", /*Quote=*/true, &CrashInfo); Diag(clang::diag::note_drv_command_failed_diag_msg) << Script; @@ -787,14 +882,6 @@ bool Driver::HandleImmediateArgs(const Compilation &C) { } return false; } - - if (C.getArgs().hasArg(options::OPT_print_multi_os_directory)) { - // FIXME: This should print out "lib/../lib", "lib/../lib64", or - // "lib/../lib32" as appropriate for the toolchain. For now, print - // nothing because it's not supported yet. - return false; - } - return true; } @@ -815,9 +902,20 @@ static unsigned PrintActions1(const Compilation &C, Action *A, } else if (BindArchAction *BIA = dyn_cast<BindArchAction>(A)) { os << '"' << BIA->getArchName() << '"' << ", {" << PrintActions1(C, *BIA->begin(), Ids) << "}"; + } else if (CudaDeviceAction *CDA = dyn_cast<CudaDeviceAction>(A)) { + os << '"' << CDA->getGpuArchName() << '"' << ", {" + << PrintActions1(C, *CDA->begin(), Ids) << "}"; } else { + ActionList *AL; + if (CudaHostAction *CHA = dyn_cast<CudaHostAction>(A)) { + os << "{" << PrintActions1(C, *CHA->begin(), Ids) << "}" + << ", gpu binaries "; + AL = &CHA->getDeviceActions(); + } else + AL = &A->getInputs(); + const char *Prefix = "{"; - for (Action *PreRequisite : *A) { + for (Action *PreRequisite : *AL) { os << Prefix << PrintActions1(C, PreRequisite, Ids); Prefix = ", "; } @@ -1130,6 +1228,93 @@ void Driver::BuildInputs(const ToolChain &TC, DerivedArgList &Args, } } +// For each unique --cuda-gpu-arch= argument creates a TY_CUDA_DEVICE input +// action and then wraps each in CudaDeviceAction paired with appropriate GPU +// arch name. If we're only building device-side code, each action remains +// independent. Otherwise we pass device-side actions as inputs to a new +// CudaHostAction which combines both host and device side actions. +static std::unique_ptr<Action> +buildCudaActions(const Driver &D, const ToolChain &TC, DerivedArgList &Args, + const Arg *InputArg, const types::ID InputType, + std::unique_ptr<Action> Current, ActionList &Actions) { + + assert(InputType == types::TY_CUDA && + "CUDA Actions only apply to CUDA inputs."); + + // Collect all cuda_gpu_arch parameters, removing duplicates. + SmallVector<const char *, 4> GpuArchList; + llvm::StringSet<> GpuArchNames; + for (Arg *A : Args) { + if (A->getOption().matches(options::OPT_cuda_gpu_arch_EQ)) { + A->claim(); + if (GpuArchNames.insert(A->getValue()).second) + GpuArchList.push_back(A->getValue()); + } + } + + // Default to sm_20 which is the lowest common denominator for supported GPUs. + // sm_20 code should work correctly, if suboptimally, on all newer GPUs. + if (GpuArchList.empty()) + GpuArchList.push_back("sm_20"); + + // Replicate inputs for each GPU architecture. + Driver::InputList CudaDeviceInputs; + for (unsigned i = 0, e = GpuArchList.size(); i != e; ++i) + CudaDeviceInputs.push_back(std::make_pair(types::TY_CUDA_DEVICE, InputArg)); + + // Build actions for all device inputs. + ActionList CudaDeviceActions; + D.BuildActions(TC, Args, CudaDeviceInputs, CudaDeviceActions); + assert(GpuArchList.size() == CudaDeviceActions.size() && + "Failed to create actions for all devices"); + + // Check whether any of device actions stopped before they could generate PTX. + bool PartialCompilation = false; + bool DeviceOnlyCompilation = Args.hasArg(options::OPT_cuda_device_only); + for (unsigned i = 0, e = GpuArchList.size(); i != e; ++i) { + if (CudaDeviceActions[i]->getKind() != Action::BackendJobClass) { + PartialCompilation = true; + break; + } + } + + // Figure out what to do with device actions -- pass them as inputs to the + // host action or run each of them independently. + if (PartialCompilation || DeviceOnlyCompilation) { + // In case of partial or device-only compilation results of device actions + // are not consumed by the host action device actions have to be added to + // top-level actions list with AtTopLevel=true and run independently. + + // -o is ambiguous if we have more than one top-level action. + if (Args.hasArg(options::OPT_o) && + (!DeviceOnlyCompilation || GpuArchList.size() > 1)) { + D.Diag(clang::diag::err_drv_output_argument_with_multiple_files); + return nullptr; + } + + for (unsigned i = 0, e = GpuArchList.size(); i != e; ++i) + Actions.push_back( + new CudaDeviceAction(std::unique_ptr<Action>(CudaDeviceActions[i]), + GpuArchList[i], /* AtTopLevel */ true)); + // Kill host action in case of device-only compilation. + if (DeviceOnlyCompilation) + Current.reset(nullptr); + return Current; + } else { + // Outputs of device actions during complete CUDA compilation get created + // with AtTopLevel=false and become inputs for the host action. + ActionList DeviceActions; + for (unsigned i = 0, e = GpuArchList.size(); i != e; ++i) + DeviceActions.push_back( + new CudaDeviceAction(std::unique_ptr<Action>(CudaDeviceActions[i]), + GpuArchList[i], /* AtTopLevel */ false)); + // Return a new host action that incorporates original host action and all + // device actions. + return std::unique_ptr<Action>( + new CudaHostAction(std::move(Current), DeviceActions)); + } +} + void Driver::BuildActions(const ToolChain &TC, DerivedArgList &Args, const InputList &Inputs, ActionList &Actions) const { llvm::PrettyStackTraceString CrashInfo("Building compilation actions"); @@ -1227,6 +1412,25 @@ void Driver::BuildActions(const ToolChain &TC, DerivedArgList &Args, continue; } + phases::ID CudaInjectionPhase; + if (isSaveTempsEnabled()) { + // All phases are done independently, inject GPU blobs during compilation + // phase as that's where we generate glue code to init them. + CudaInjectionPhase = phases::Compile; + } else { + // Assumes that clang does everything up until linking phase, so we inject + // cuda device actions at the last step before linking. Otherwise CUDA + // host action forces preprocessor into a separate invocation. + if (FinalPhase == phases::Link) { + for (auto i = PL.begin(), e = PL.end(); i != e; ++i) { + auto next = i + 1; + if (next != e && *next == phases::Link) + CudaInjectionPhase = *i; + } + } else + CudaInjectionPhase = FinalPhase; + } + // Build the pipeline for this file. std::unique_ptr<Action> Current(new InputAction(*InputArg, InputType)); for (SmallVectorImpl<phases::ID>::iterator i = PL.begin(), e = PL.end(); @@ -1252,6 +1456,15 @@ void Driver::BuildActions(const ToolChain &TC, DerivedArgList &Args, // Otherwise construct the appropriate action. Current = ConstructPhaseAction(TC, Args, Phase, std::move(Current)); + + if (InputType == types::TY_CUDA && Phase == CudaInjectionPhase && + !Args.hasArg(options::OPT_cuda_host_only)) { + Current = buildCudaActions(*this, TC, Args, InputArg, InputType, + std::move(Current), Actions); + if (!Current) + break; + } + if (Current->getType() == types::TY_Nothing) break; } @@ -1491,7 +1704,13 @@ static const Tool *SelectToolForJob(Compilation &C, bool SaveTemps, if (isa<BackendJobAction>(JA)) { // Check if the compiler supports emitting LLVM IR. assert(Inputs->size() == 1); - JobAction *CompileJA = cast<CompileJobAction>(*Inputs->begin()); + JobAction *CompileJA; + // Extract real host action, if it's a CudaHostAction. + if (CudaHostAction *CudaHA = dyn_cast<CudaHostAction>(*Inputs->begin())) + CompileJA = cast<CompileJobAction>(*CudaHA->begin()); + else + CompileJA = cast<CompileJobAction>(*Inputs->begin()); + const Tool *Compiler = TC->SelectTool(*CompileJA); if (!Compiler) return nullptr; @@ -1525,6 +1744,20 @@ void Driver::BuildJobsForAction(Compilation &C, const Action *A, InputInfo &Result) const { llvm::PrettyStackTraceString CrashInfo("Building compilation jobs"); + InputInfoList CudaDeviceInputInfos; + if (const CudaHostAction *CHA = dyn_cast<CudaHostAction>(A)) { + InputInfo II; + // Append outputs of device jobs to the input list. + for (const Action *DA : CHA->getDeviceActions()) { + BuildJobsForAction(C, DA, TC, "", AtTopLevel, + /*MultipleArchs*/ false, LinkingOutput, II); + CudaDeviceInputInfos.push_back(II); + } + // Override current action with a real host compile action and continue + // processing it. + A = *CHA->begin(); + } + if (const InputAction *IA = dyn_cast<InputAction>(A)) { // FIXME: It would be nice to not claim this here; maybe the old scheme of // just using Args was better? @@ -1544,15 +1777,30 @@ void Driver::BuildJobsForAction(Compilation &C, const Action *A, const char *ArchName = BAA->getArchName(); if (ArchName) - TC = &getToolChain(C.getArgs(), ArchName); + TC = &getToolChain( + C.getArgs(), + computeTargetTriple(DefaultTargetTriple, C.getArgs(), ArchName)); else TC = &C.getDefaultToolChain(); - BuildJobsForAction(C, *BAA->begin(), TC, BAA->getArchName(), AtTopLevel, + BuildJobsForAction(C, *BAA->begin(), TC, ArchName, AtTopLevel, MultipleArchs, LinkingOutput, Result); return; } + if (const CudaDeviceAction *CDA = dyn_cast<CudaDeviceAction>(A)) { + // Figure out which NVPTX triple to use for device-side compilation based on + // whether host is 64-bit. + llvm::Triple DeviceTriple(C.getDefaultToolChain().getTriple().isArch64Bit() + ? "nvptx64-nvidia-cuda" + : "nvptx-nvidia-cuda"); + BuildJobsForAction(C, *CDA->begin(), + &getToolChain(C.getArgs(), DeviceTriple), + CDA->getGpuArchName(), CDA->isAtTopLevel(), + /*MultipleArchs*/ true, LinkingOutput, Result); + return; + } + const ActionList *Inputs = &A->getInputs(); const JobAction *JA = cast<JobAction>(A); @@ -1584,6 +1832,10 @@ void Driver::BuildJobsForAction(Compilation &C, const Action *A, if (JA->getType() == types::TY_dSYM) BaseInput = InputInfos[0].getFilename(); + // Append outputs of cuda device jobs to the input list + if (CudaDeviceInputInfos.size()) + InputInfos.append(CudaDeviceInputInfos.begin(), CudaDeviceInputInfos.end()); + // Determine the place to write output to, if any. if (JA->getType() == types::TY_Nothing) Result = InputInfo(A->getType(), BaseInput); @@ -1899,93 +2151,8 @@ std::string Driver::GetTemporaryPath(StringRef Prefix, return Path.str(); } -/// \brief Compute target triple from args. -/// -/// This routine provides the logic to compute a target triple from various -/// args passed to the driver and the default triple string. -static llvm::Triple computeTargetTriple(StringRef DefaultTargetTriple, - const ArgList &Args, - StringRef DarwinArchName) { - // FIXME: Already done in Compilation *Driver::BuildCompilation - if (const Arg *A = Args.getLastArg(options::OPT_target)) - DefaultTargetTriple = A->getValue(); - - llvm::Triple Target(llvm::Triple::normalize(DefaultTargetTriple)); - - // Handle Apple-specific options available here. - if (Target.isOSBinFormatMachO()) { - // If an explict Darwin arch name is given, that trumps all. - if (!DarwinArchName.empty()) { - tools::darwin::setTripleTypeForMachOArchName(Target, DarwinArchName); - return Target; - } - - // Handle the Darwin '-arch' flag. - if (Arg *A = Args.getLastArg(options::OPT_arch)) { - StringRef ArchName = A->getValue(); - tools::darwin::setTripleTypeForMachOArchName(Target, ArchName); - } - } - - // Handle pseudo-target flags '-mlittle-endian'/'-EL' and - // '-mbig-endian'/'-EB'. - if (Arg *A = Args.getLastArg(options::OPT_mlittle_endian, - options::OPT_mbig_endian)) { - if (A->getOption().matches(options::OPT_mlittle_endian)) { - if (Target.getArch() == llvm::Triple::mips) - Target.setArch(llvm::Triple::mipsel); - else if (Target.getArch() == llvm::Triple::mips64) - Target.setArch(llvm::Triple::mips64el); - else if (Target.getArch() == llvm::Triple::aarch64_be) - Target.setArch(llvm::Triple::aarch64); - } else { - if (Target.getArch() == llvm::Triple::mipsel) - Target.setArch(llvm::Triple::mips); - else if (Target.getArch() == llvm::Triple::mips64el) - Target.setArch(llvm::Triple::mips64); - else if (Target.getArch() == llvm::Triple::aarch64) - Target.setArch(llvm::Triple::aarch64_be); - } - } - - // Skip further flag support on OSes which don't support '-m32' or '-m64'. - if (Target.getArchName() == "tce" || Target.getOS() == llvm::Triple::Minix) - return Target; - - // Handle pseudo-target flags '-m64', '-mx32', '-m32' and '-m16'. - if (Arg *A = Args.getLastArg(options::OPT_m64, options::OPT_mx32, - options::OPT_m32, options::OPT_m16)) { - llvm::Triple::ArchType AT = llvm::Triple::UnknownArch; - - if (A->getOption().matches(options::OPT_m64)) { - AT = Target.get64BitArchVariant().getArch(); - if (Target.getEnvironment() == llvm::Triple::GNUX32) - Target.setEnvironment(llvm::Triple::GNU); - } else if (A->getOption().matches(options::OPT_mx32) && - Target.get64BitArchVariant().getArch() == llvm::Triple::x86_64) { - AT = llvm::Triple::x86_64; - Target.setEnvironment(llvm::Triple::GNUX32); - } else if (A->getOption().matches(options::OPT_m32)) { - AT = Target.get32BitArchVariant().getArch(); - if (Target.getEnvironment() == llvm::Triple::GNUX32) - Target.setEnvironment(llvm::Triple::GNU); - } else if (A->getOption().matches(options::OPT_m16) && - Target.get32BitArchVariant().getArch() == llvm::Triple::x86) { - AT = llvm::Triple::x86; - Target.setEnvironment(llvm::Triple::CODE16); - } - - if (AT != llvm::Triple::UnknownArch && AT != Target.getArch()) - Target.setArch(AT); - } - - return Target; -} - const ToolChain &Driver::getToolChain(const ArgList &Args, - StringRef DarwinArchName) const { - llvm::Triple Target = - computeTargetTriple(DefaultTargetTriple, Args, DarwinArchName); + const llvm::Triple &Target) const { ToolChain *&TC = ToolChains[Target.str()]; if (!TC) { @@ -2050,6 +2217,9 @@ const ToolChain &Driver::getToolChain(const ArgList &Args, break; } break; + case llvm::Triple::CUDA: + TC = new toolchains::CudaToolChain(*this, Target, Args); + break; default: // Of these targets, Hexagon is the only one that might have // an OS of Linux, in which case it got handled above already. diff --git a/lib/Driver/Job.cpp b/lib/Driver/Job.cpp index ac18e1e..42bba56 100644 --- a/lib/Driver/Job.cpp +++ b/lib/Driver/Job.cpp @@ -71,7 +71,7 @@ static int skipArgs(const char *Flag, bool HaveCrashVFS) { return 0; } -static void PrintArg(raw_ostream &OS, const char *Arg, bool Quote) { +void Command::printArg(raw_ostream &OS, const char *Arg, bool Quote) { const bool Escape = std::strpbrk(Arg, "\"\\$"); if (!Quote && !Escape) { @@ -146,7 +146,7 @@ void Command::Print(raw_ostream &OS, const char *Terminator, bool Quote, CrashReportInfo *CrashInfo) const { // Always quote the exe. OS << ' '; - PrintArg(OS, Executable, /*Quote=*/true); + printArg(OS, Executable, /*Quote=*/true); llvm::ArrayRef<const char *> Args = Arguments; llvm::SmallVector<const char *, 128> ArgsRespFile; @@ -175,20 +175,20 @@ void Command::Print(raw_ostream &OS, const char *Terminator, bool Quote, // Replace the input file name with the crashinfo's file name. OS << ' '; StringRef ShortName = llvm::sys::path::filename(CrashInfo->Filename); - PrintArg(OS, ShortName.str().c_str(), Quote); + printArg(OS, ShortName.str().c_str(), Quote); continue; } } OS << ' '; - PrintArg(OS, Arg, Quote); + printArg(OS, Arg, Quote); } if (CrashInfo && HaveCrashVFS) { OS << ' '; - PrintArg(OS, "-ivfsoverlay", Quote); + printArg(OS, "-ivfsoverlay", Quote); OS << ' '; - PrintArg(OS, CrashInfo->VFSPath.str().c_str(), Quote); + printArg(OS, CrashInfo->VFSPath.str().c_str(), Quote); } if (ResponseFile != nullptr) { diff --git a/lib/Driver/MSVCToolChain.cpp b/lib/Driver/MSVCToolChain.cpp index f20f58a..c816b29 100644 --- a/lib/Driver/MSVCToolChain.cpp +++ b/lib/Driver/MSVCToolChain.cpp @@ -526,8 +526,5 @@ MSVCToolChain::ComputeEffectiveClangTriple(const ArgList &Args, SanitizerMask MSVCToolChain::getSupportedSanitizers() const { SanitizerMask Res = ToolChain::getSupportedSanitizers(); Res |= SanitizerKind::Address; - // CFI checks are not implemented for MSVC ABI for now. - Res &= ~SanitizerKind::CFI; - Res &= ~SanitizerKind::CFICastStrict; return Res; } diff --git a/lib/Driver/MinGWToolChain.cpp b/lib/Driver/MinGWToolChain.cpp index 606508d..197c19e 100644 --- a/lib/Driver/MinGWToolChain.cpp +++ b/lib/Driver/MinGWToolChain.cpp @@ -1,5 +1,4 @@ -//===--- MinGWToolChain.cpp - MinGWToolChain Implementation -//-----------------------===// +//===--- MinGWToolChain.cpp - MinGWToolChain Implementation ---------------===// // // The LLVM Compiler Infrastructure // @@ -25,6 +24,9 @@ MinGW::MinGW(const Driver &D, const llvm::Triple &Triple, const ArgList &Args) : ToolChain(D, Triple, Args) { getProgramPaths().push_back(getDriver().getInstalledDir()); + llvm::SmallString<1024> LibDir; + +#ifdef LLVM_ON_WIN32 if (getDriver().SysRoot.size()) Base = getDriver().SysRoot; else if (llvm::ErrorOr<std::string> GPPName = @@ -34,8 +36,17 @@ MinGW::MinGW(const Driver &D, const llvm::Triple &Triple, const ArgList &Args) else Base = llvm::sys::path::parent_path(getDriver().getInstalledDir()); Base += llvm::sys::path::get_separator(); - llvm::SmallString<1024> LibDir(Base); + LibDir = Base; llvm::sys::path::append(LibDir, "lib", "gcc"); +#else + if (getDriver().SysRoot.size()) + Base = getDriver().SysRoot; + else + Base = "/usr/"; + LibDir = Base; + llvm::sys::path::append(LibDir, "lib64", "gcc"); +#endif + LibDir += llvm::sys::path::get_separator(); // First look for mingw-w64. @@ -45,6 +56,7 @@ MinGW::MinGW(const Driver &D, const llvm::Triple &Triple, const ArgList &Args) llvm::sys::fs::directory_iterator MingW64Entry(LibDir + Arch, EC); if (!EC) { GccLibDir = MingW64Entry->path(); + Ver = llvm::sys::path::filename(GccLibDir); } else { // If mingw-w64 not found, try looking for mingw.org. Arch = "mingw32"; @@ -58,6 +70,10 @@ MinGW::MinGW(const Driver &D, const llvm::Triple &Triple, const ArgList &Args) getFilePaths().push_back(GccLibDir); getFilePaths().push_back(Base + "lib"); getFilePaths().push_back(Base + Arch + "lib"); +#ifdef LLVM_ON_UNIX + // For openSUSE. + getFilePaths().push_back(Base + Arch + "sys-root/mingw/lib"); +#endif } bool MinGW::IsIntegratedAssemblerDefault() const { return true; } @@ -117,6 +133,11 @@ void MinGW::AddClangSystemIncludeArgs(const ArgList &DriverArgs, llvm::sys::path::append(IncludeDir, "include"); addSystemInclude(DriverArgs, CC1Args, IncludeDir.c_str()); IncludeDir += "-fixed"; +#ifdef LLVM_ON_UNIX + // For openSUSE. + addSystemInclude(DriverArgs, CC1Args, + "/usr/x86_64-w64-mingw32/sys-root/mingw/include"); +#endif addSystemInclude(DriverArgs, CC1Args, IncludeDir.c_str()); addSystemInclude(DriverArgs, CC1Args, Base + Arch + "include"); addSystemInclude(DriverArgs, CC1Args, Base + "include"); @@ -128,16 +149,27 @@ void MinGW::AddClangCXXStdlibIncludeArgs(const ArgList &DriverArgs, DriverArgs.hasArg(options::OPT_nostdincxx)) return; - llvm::SmallString<1024> IncludeDir; - for (bool MingW64 : {true, false}) { - if (MingW64) - IncludeDir = Base + Arch; - else - IncludeDir = GccLibDir; - llvm::sys::path::append(IncludeDir, "include", "c++"); - addSystemInclude(DriverArgs, CC1Args, IncludeDir.str()); - IncludeDir += llvm::sys::path::get_separator(); - addSystemInclude(DriverArgs, CC1Args, IncludeDir.str() + Arch); - addSystemInclude(DriverArgs, CC1Args, IncludeDir.str() + "backward"); + // C++ includes may be found in several locations depending on distribution. + // Windows + // ------- + // mingw-w64 mingw-builds: $sysroot/i686-w64-mingw32/include/c++. + // mingw-w64 msys2: $sysroot/include/c++/4.9.2 + // mingw.org: GccLibDir/include/c++ + // + // Linux + // ----- + // openSUSE: GccLibDir/include/c++ + llvm::SmallVector<llvm::SmallString<1024>, 3> CppIncludeBases; + CppIncludeBases.emplace_back(Base); + llvm::sys::path::append(CppIncludeBases[0], Arch, "include", "c++"); + CppIncludeBases.emplace_back(Base); + llvm::sys::path::append(CppIncludeBases[1], "include", "c++", Ver); + CppIncludeBases.emplace_back(GccLibDir); + llvm::sys::path::append(CppIncludeBases[2], "include", "c++"); + for (auto &CppIncludeBase : CppIncludeBases) { + CppIncludeBase += llvm::sys::path::get_separator(); + addSystemInclude(DriverArgs, CC1Args, CppIncludeBase); + addSystemInclude(DriverArgs, CC1Args, CppIncludeBase + Arch); + addSystemInclude(DriverArgs, CC1Args, CppIncludeBase + "backward"); } } diff --git a/lib/Driver/SanitizerArgs.cpp b/lib/Driver/SanitizerArgs.cpp index 3043481..c3ad8ef 100644 --- a/lib/Driver/SanitizerArgs.cpp +++ b/lib/Driver/SanitizerArgs.cpp @@ -176,6 +176,7 @@ void SanitizerArgs::clear() { BlacklistFiles.clear(); CoverageFeatures = 0; MsanTrackOrigins = 0; + MsanUseAfterDtor = false; AsanFieldPadding = 0; AsanZeroBaseShadow = false; AsanSharedRuntime = false; @@ -289,8 +290,12 @@ SanitizerArgs::SanitizerArgs(const ToolChain &TC, // toolchain. We don't have a good way to check the latter, so we just // check if the toolchan supports vptr. if (~Supported & Vptr) { - if (SanitizerMask KindsToDiagnose = - Kinds & ~TrappingKinds & NeedsUbsanCxxRt) { + SanitizerMask KindsToDiagnose = Kinds & ~TrappingKinds & NeedsUbsanCxxRt; + // The runtime library supports the Microsoft C++ ABI, but only well enough + // for CFI. FIXME: Remove this once we support vptr on Windows. + if (TC.getTriple().isOSWindows()) + KindsToDiagnose &= ~CFI; + if (KindsToDiagnose) { SanitizerSet S; S.Mask = KindsToDiagnose; D.Diag(diag::err_drv_unsupported_opt_for_target) @@ -413,6 +418,8 @@ SanitizerArgs::SanitizerArgs(const ToolChain &TC, } } } + MsanUseAfterDtor = + Args.hasArg(options::OPT_fsanitize_memory_use_after_dtor); } // Parse -f(no-)?sanitize-coverage flags if coverage is supported by the @@ -558,6 +565,10 @@ void SanitizerArgs::addArgs(const ToolChain &TC, const llvm::opt::ArgList &Args, if (MsanTrackOrigins) CmdArgs.push_back(Args.MakeArgString("-fsanitize-memory-track-origins=" + llvm::utostr(MsanTrackOrigins))); + + if (MsanUseAfterDtor) + CmdArgs.push_back(Args.MakeArgString("-fsanitize-memory-use-after-dtor")); + if (AsanFieldPadding) CmdArgs.push_back(Args.MakeArgString("-fsanitize-address-field-padding=" + llvm::utostr(AsanFieldPadding))); diff --git a/lib/Driver/ToolChain.cpp b/lib/Driver/ToolChain.cpp index da020a2..e6a1bc9 100644 --- a/lib/Driver/ToolChain.cpp +++ b/lib/Driver/ToolChain.cpp @@ -151,6 +151,8 @@ Tool *ToolChain::getTool(Action::ActionClass AC) const { case Action::InputClass: case Action::BindArchClass: + case Action::CudaDeviceClass: + case Action::CudaHostClass: case Action::LipoJobClass: case Action::DsymutilJobClass: case Action::VerifyDebugInfoJobClass: diff --git a/lib/Driver/ToolChains.cpp b/lib/Driver/ToolChains.cpp index df74b41..15e36a1 100644 --- a/lib/Driver/ToolChains.cpp +++ b/lib/Driver/ToolChains.cpp @@ -303,6 +303,7 @@ void Darwin::addProfileRTLibs(const ArgList &Args, if (!(Args.hasFlag(options::OPT_fprofile_arcs, options::OPT_fno_profile_arcs, false) || Args.hasArg(options::OPT_fprofile_generate) || + Args.hasArg(options::OPT_fprofile_generate_EQ) || Args.hasArg(options::OPT_fprofile_instr_generate) || Args.hasArg(options::OPT_fprofile_instr_generate_EQ) || Args.hasArg(options::OPT_fcreate_profile) || @@ -2337,6 +2338,13 @@ NaCl_TC::NaCl_TC(const Driver &D, const llvm::Triple &Triple, file_paths.push_back(ToolPath + "arm-nacl"); break; } + case llvm::Triple::mipsel: { + file_paths.push_back(FilePath + "mipsel-nacl/lib"); + file_paths.push_back(FilePath + "mipsel-nacl/usr/lib"); + prog_paths.push_back(ProgPath + "bin"); + file_paths.push_back(ToolPath + "mipsel-nacl"); + break; + } default: break; } @@ -2372,6 +2380,9 @@ void NaCl_TC::AddClangSystemIncludeArgs(const ArgList &DriverArgs, case llvm::Triple::x86_64: llvm::sys::path::append(P, "x86_64-nacl/usr/include"); break; + case llvm::Triple::mipsel: + llvm::sys::path::append(P, "mipsel-nacl/usr/include"); + break; default: return; } @@ -2416,6 +2427,10 @@ void NaCl_TC::AddClangCXXStdlibIncludeArgs(const ArgList &DriverArgs, llvm::sys::path::append(P, "x86_64-nacl/include/c++/v1"); addSystemInclude(DriverArgs, CC1Args, P.str()); break; + case llvm::Triple::mipsel: + llvm::sys::path::append(P, "mipsel-nacl/include/c++/v1"); + addSystemInclude(DriverArgs, CC1Args, P.str()); + break; default: break; } @@ -2868,6 +2883,7 @@ enum Distro { UbuntuTrusty, UbuntuUtopic, UbuntuVivid, + UbuntuWily, UnknownDistro }; @@ -2882,7 +2898,7 @@ static bool IsDebian(enum Distro Distro) { } static bool IsUbuntu(enum Distro Distro) { - return Distro >= UbuntuHardy && Distro <= UbuntuVivid; + return Distro >= UbuntuHardy && Distro <= UbuntuWily; } static Distro DetectDistro(llvm::Triple::ArchType Arch) { @@ -2911,6 +2927,7 @@ static Distro DetectDistro(llvm::Triple::ArchType Arch) { .Case("trusty", UbuntuTrusty) .Case("utopic", UbuntuUtopic) .Case("vivid", UbuntuVivid) + .Case("wily", UbuntuWily) .Default(UnknownDistro); return Version; } @@ -3635,6 +3652,65 @@ Tool *DragonFly::buildLinker() const { return new tools::dragonfly::Linker(*this); } +/// Stub for CUDA toolchain. At the moment we don't have assembler or +/// linker and need toolchain mainly to propagate device-side options +/// to CC1. + +CudaToolChain::CudaToolChain(const Driver &D, const llvm::Triple &Triple, + const ArgList &Args) + : Linux(D, Triple, Args) {} + +void +CudaToolChain::addClangTargetOptions(const llvm::opt::ArgList &DriverArgs, + llvm::opt::ArgStringList &CC1Args) const { + Linux::addClangTargetOptions(DriverArgs, CC1Args); + CC1Args.push_back("-fcuda-is-device"); +} + +llvm::opt::DerivedArgList * +CudaToolChain::TranslateArgs(const llvm::opt::DerivedArgList &Args, + const char *BoundArch) const { + DerivedArgList *DAL = new DerivedArgList(Args.getBaseArgs()); + const OptTable &Opts = getDriver().getOpts(); + + for (Arg *A : Args) { + if (A->getOption().matches(options::OPT_Xarch__)) { + // Skip this argument unless the architecture matches BoundArch + if (A->getValue(0) != StringRef(BoundArch)) + continue; + + unsigned Index = Args.getBaseArgs().MakeIndex(A->getValue(1)); + unsigned Prev = Index; + std::unique_ptr<Arg> XarchArg(Opts.ParseOneArg(Args, Index)); + + // If the argument parsing failed or more than one argument was + // consumed, the -Xarch_ argument's parameter tried to consume + // extra arguments. Emit an error and ignore. + // + // We also want to disallow any options which would alter the + // driver behavior; that isn't going to work in our model. We + // use isDriverOption() as an approximation, although things + // like -O4 are going to slip through. + if (!XarchArg || Index > Prev + 1) { + getDriver().Diag(diag::err_drv_invalid_Xarch_argument_with_args) + << A->getAsString(Args); + continue; + } else if (XarchArg->getOption().hasFlag(options::DriverOption)) { + getDriver().Diag(diag::err_drv_invalid_Xarch_argument_isdriver) + << A->getAsString(Args); + continue; + } + XarchArg->setBaseArg(A); + A = XarchArg.release(); + DAL->AddSynthesizedArg(A); + } + DAL->append(A); + } + + DAL->AddJoinedArg(nullptr, Opts.getOption(options::OPT_march_EQ), BoundArch); + return DAL; +} + /// XCore tool chain XCore::XCore(const Driver &D, const llvm::Triple &Triple, const ArgList &Args) : ToolChain(D, Triple, Args) { diff --git a/lib/Driver/ToolChains.h b/lib/Driver/ToolChains.h index 629f90a..327ff9b 100644 --- a/lib/Driver/ToolChains.h +++ b/lib/Driver/ToolChains.h @@ -552,6 +552,7 @@ protected: private: std::string Base; std::string GccLibDir; + std::string Ver; std::string Arch; mutable std::unique_ptr<tools::gcc::Preprocessor> Preprocessor; mutable std::unique_ptr<tools::gcc::Compiler> Compiler; @@ -698,6 +699,18 @@ private: std::string computeSysRoot() const; }; +class LLVM_LIBRARY_VISIBILITY CudaToolChain : public Linux { +public: + CudaToolChain(const Driver &D, const llvm::Triple &Triple, + const llvm::opt::ArgList &Args); + + llvm::opt::DerivedArgList * + TranslateArgs(const llvm::opt::DerivedArgList &Args, + const char *BoundArch) const override; + void addClangTargetOptions(const llvm::opt::ArgList &DriverArgs, + llvm::opt::ArgStringList &CC1Args) const override; +}; + class LLVM_LIBRARY_VISIBILITY Hexagon_TC : public Linux { protected: GCCVersion GCCLibAndIncVersion; @@ -746,7 +759,9 @@ public: void AddCXXStdlibLibArgs(const llvm::opt::ArgList &Args, llvm::opt::ArgStringList &CmdArgs) const override; - bool IsIntegratedAssemblerDefault() const override { return false; } + bool IsIntegratedAssemblerDefault() const override { + return getTriple().getArch() == llvm::Triple::mipsel; + } // Get the path to the file containing NaCl's ARM macros. It lives in NaCl_TC // because the AssembleARM tool needs a const char * that it can pass around diff --git a/lib/Driver/Tools.cpp b/lib/Driver/Tools.cpp index a2955db..8c11992 100644 --- a/lib/Driver/Tools.cpp +++ b/lib/Driver/Tools.cpp @@ -638,6 +638,8 @@ static void getARMTargetFeatures(const Driver &D, const llvm::Triple &Triple, const ArgList &Args, std::vector<const char *> &Features, bool ForAS) { + bool KernelOrKext = + Args.hasArg(options::OPT_mkernel, options::OPT_fapple_kext); StringRef FloatABI = tools::arm::getARMFloatABI(D, Args, Triple); if (!ForAS) { // FIXME: Note, this is a hack, the LLVM backend doesn't actually use these @@ -705,6 +707,17 @@ static void getARMTargetFeatures(const Driver &D, const llvm::Triple &Triple, if (Triple.getSubArch() == llvm::Triple::SubArchType::ARMSubArch_v8_1a) { Features.insert(Features.begin(), "+v8.1a"); } + + // Look for the last occurrence of -mlong-calls or -mno-long-calls. If + // neither options are specified, see if we are compiling for kernel/kext and + // decide whether to pass "+long-calls" based on the OS and its version. + if (Arg *A = Args.getLastArg(options::OPT_mlong_calls, + options::OPT_mno_long_calls)) { + if (A->getOption().matches(options::OPT_mlong_calls)) + Features.push_back("+long-calls"); + } else if (KernelOrKext && (!Triple.isiOS() || Triple.isOSVersionLT(6))) { + Features.push_back("+long-calls"); + } } void Clang::AddARMTargetArgs(const ArgList &Args, ArgStringList &CmdArgs, @@ -778,11 +791,6 @@ void Clang::AddARMTargetArgs(const ArgList &Args, ArgStringList &CmdArgs, // Kernel code has more strict alignment requirements. if (KernelOrKext) { - if (!Triple.isiOS() || Triple.isOSVersionLT(6)) { - CmdArgs.push_back("-backend-option"); - CmdArgs.push_back("-arm-long-calls"); - } - CmdArgs.push_back("-backend-option"); CmdArgs.push_back("-arm-strict-align"); @@ -1264,9 +1272,7 @@ void Clang::AddPPCTargetArgs(const ArgList &Args, ArgStringList &CmdArgs) const { // Select the ABI to use. const char *ABIName = nullptr; - if (Arg *A = Args.getLastArg(options::OPT_mabi_EQ)) { - ABIName = A->getValue(); - } else if (getToolChain().getTriple().isOSLinux()) + if (getToolChain().getTriple().isOSLinux()) switch (getToolChain().getArch()) { case llvm::Triple::ppc64: { // When targeting a processor that supports QPX, or if QPX is @@ -1291,6 +1297,13 @@ void Clang::AddPPCTargetArgs(const ArgList &Args, break; } + if (Arg *A = Args.getLastArg(options::OPT_mabi_EQ)) + // The ppc64 linux abis are all "altivec" abis by default. Accept and ignore + // the option if given as we don't have backend support for any targets + // that don't use the altivec abi. + if (StringRef(A->getValue()) != "altivec") + ABIName = A->getValue(); + if (ABIName) { CmdArgs.push_back("-target-abi"); CmdArgs.push_back(ABIName); @@ -1475,6 +1488,12 @@ static std::string getCPUName(const ArgList &Args, const llvm::Triple &T) { return CPUName; } + case llvm::Triple::nvptx: + case llvm::Triple::nvptx64: + if (const Arg *A = Args.getLastArg(options::OPT_march_EQ)) + return A->getValue(); + return ""; + case llvm::Triple::ppc: case llvm::Triple::ppc64: case llvm::Triple::ppc64le: { @@ -2028,17 +2047,6 @@ shouldUseExceptionTablesForObjCExceptions(const ObjCRuntime &runtime, Triple.getArch() == llvm::Triple::arm)); } -// exceptionSettings() exists to share the logic between -cc1 and linker -// invocations. -static bool exceptionSettings(const ArgList &Args, const llvm::Triple &Triple) { - if (Arg *A = Args.getLastArg(options::OPT_fexceptions, - options::OPT_fno_exceptions)) - if (A->getOption().matches(options::OPT_fexceptions)) - return true; - - return false; -} - /// Adds exception related arguments to the driver command arguments. There's a /// master flag, -fexceptions and also language specific flags to enable/disable /// C++ and Objective-C exceptions. This makes it possible to for example @@ -2062,8 +2070,9 @@ static void addExceptionArgs(const ArgList &Args, types::ID InputType, return; } - // Gather the exception settings from the command line arguments. - bool EH = exceptionSettings(Args, Triple); + // See if the user explicitly enabled exceptions. + bool EH = Args.hasFlag(options::OPT_fexceptions, options::OPT_fno_exceptions, + false); // Obj-C exceptions are enabled by default, regardless of -fexceptions. This // is not necessarily sensible, but follows GCC. @@ -2076,8 +2085,11 @@ static void addExceptionArgs(const ArgList &Args, types::ID InputType, } if (types::isCXX(InputType)) { - bool CXXExceptionsEnabled = - Triple.getArch() != llvm::Triple::xcore && !Triple.isPS4CPU(); + // Disable C++ EH by default on XCore, PS4, and MSVC. + // FIXME: Remove MSVC from this list once things work. + bool CXXExceptionsEnabled = Triple.getArch() != llvm::Triple::xcore && + !Triple.isPS4CPU() && + !Triple.isWindowsMSVCEnvironment(); Arg *ExceptionArg = Args.getLastArg( options::OPT_fcxx_exceptions, options::OPT_fno_cxx_exceptions, options::OPT_fexceptions, options::OPT_fno_exceptions); @@ -2290,6 +2302,7 @@ static void addProfileRT(const ToolChain &TC, const ArgList &Args, if (!(Args.hasFlag(options::OPT_fprofile_arcs, options::OPT_fno_profile_arcs, false) || Args.hasArg(options::OPT_fprofile_generate) || + Args.hasArg(options::OPT_fprofile_generate_EQ) || Args.hasArg(options::OPT_fprofile_instr_generate) || Args.hasArg(options::OPT_fprofile_instr_generate_EQ) || Args.hasArg(options::OPT_fcreate_profile) || @@ -2723,6 +2736,89 @@ VersionTuple visualstudio::getMSVCVersion(const Driver *D, return VersionTuple(); } +static void addPGOAndCoverageFlags(Compilation &C, const Driver &D, + const InputInfo &Output, const ArgList &Args, + ArgStringList &CmdArgs) { + auto *ProfileGenerateArg = Args.getLastArg( + options::OPT_fprofile_instr_generate, + options::OPT_fprofile_instr_generate_EQ, options::OPT_fprofile_generate, + options::OPT_fprofile_generate_EQ); + + auto *ProfileUseArg = Args.getLastArg( + options::OPT_fprofile_instr_use, options::OPT_fprofile_instr_use_EQ, + options::OPT_fprofile_use, options::OPT_fprofile_use_EQ); + + if (ProfileGenerateArg && ProfileUseArg) + D.Diag(diag::err_drv_argument_not_allowed_with) + << ProfileGenerateArg->getSpelling() + << ProfileUseArg->getSpelling(); + + if (ProfileGenerateArg && + ProfileGenerateArg->getOption().matches( + options::OPT_fprofile_instr_generate_EQ)) + ProfileGenerateArg->render(Args, CmdArgs); + else if (ProfileGenerateArg && + ProfileGenerateArg->getOption().matches( + options::OPT_fprofile_generate_EQ)) { + SmallString<128> Path(ProfileGenerateArg->getValue()); + llvm::sys::path::append(Path, "default.profraw"); + CmdArgs.push_back( + Args.MakeArgString(Twine("-fprofile-instr-generate=") + Path)); + } else + Args.AddAllArgs(CmdArgs, options::OPT_fprofile_instr_generate); + + if (ProfileUseArg && + ProfileUseArg->getOption().matches(options::OPT_fprofile_instr_use_EQ)) + ProfileUseArg->render(Args, CmdArgs); + else if (ProfileUseArg && + (ProfileUseArg->getOption().matches(options::OPT_fprofile_use_EQ) || + ProfileUseArg->getOption().matches( + options::OPT_fprofile_instr_use))) { + SmallString<128> Path( + ProfileUseArg->getNumValues() == 0 ? "" : ProfileUseArg->getValue()); + if (Path.empty() || llvm::sys::fs::is_directory(Path)) + llvm::sys::path::append(Path, "default.profdata"); + CmdArgs.push_back(Args.MakeArgString(Twine("-fprofile-instr-use=") + Path)); + } + + if (Args.hasArg(options::OPT_ftest_coverage) || + Args.hasArg(options::OPT_coverage)) + CmdArgs.push_back("-femit-coverage-notes"); + if (Args.hasFlag(options::OPT_fprofile_arcs, options::OPT_fno_profile_arcs, + false) || + Args.hasArg(options::OPT_coverage)) + CmdArgs.push_back("-femit-coverage-data"); + + if (Args.hasArg(options::OPT_fcoverage_mapping) && !ProfileGenerateArg) + D.Diag(diag::err_drv_argument_only_allowed_with) + << "-fcoverage-mapping" + << "-fprofile-instr-generate"; + + if (Args.hasArg(options::OPT_fcoverage_mapping)) + CmdArgs.push_back("-fcoverage-mapping"); + + if (C.getArgs().hasArg(options::OPT_c) || + C.getArgs().hasArg(options::OPT_S)) { + if (Output.isFilename()) { + CmdArgs.push_back("-coverage-file"); + SmallString<128> CoverageFilename; + if (Arg *FinalOutput = C.getArgs().getLastArg(options::OPT_o)) { + CoverageFilename = FinalOutput->getValue(); + } else { + CoverageFilename = llvm::sys::path::filename(Output.getBaseInput()); + } + if (llvm::sys::path::is_relative(CoverageFilename)) { + SmallString<128> Pwd; + if (!llvm::sys::fs::current_path(Pwd)) { + llvm::sys::path::append(Pwd, CoverageFilename); + CoverageFilename.swap(Pwd); + } + } + CmdArgs.push_back(Args.MakeArgString(CoverageFilename)); + } + } +} + void Clang::ConstructJob(Compilation &C, const JobAction &JA, const InputInfo &Output, const InputInfoList &Inputs, const ArgList &Args, const char *LinkingOutput) const { @@ -2736,8 +2832,14 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA, getToolChain().getTriple().isWindowsCygwinEnvironment(); bool IsWindowsMSVC = getToolChain().getTriple().isWindowsMSVCEnvironment(); - assert(Inputs.size() == 1 && "Unable to handle multiple inputs."); + // Check number of inputs for sanity. We need at least one input. + assert(Inputs.size() >= 1 && "Must have at least one input."); const InputInfo &Input = Inputs[0]; + // CUDA compilation may have multiple inputs (source file + results of + // device-side compilations). All other jobs are expected to have exactly one + // input. + bool IsCuda = types::isCuda(Input.getType()); + assert((IsCuda || Inputs.size() == 1) && "Unable to handle multiple inputs."); // Invoke ourselves in -cc1 mode. // @@ -2805,6 +2907,10 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA, assert((isa<CompileJobAction>(JA) || isa<BackendJobAction>(JA)) && "Invalid action for clang tool."); + if (JA.getType() == types::TY_LTO_IR || + JA.getType() == types::TY_LTO_BC) { + CmdArgs.push_back("-flto"); + } if (JA.getType() == types::TY_Nothing) { CmdArgs.push_back("-fsyntax-only"); } else if (JA.getType() == types::TY_LLVM_IR || @@ -3520,62 +3626,7 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA, Args.AddAllArgs(CmdArgs, options::OPT_finstrument_functions); - if ((Args.hasArg(options::OPT_fprofile_instr_generate) || - Args.hasArg(options::OPT_fprofile_instr_generate_EQ)) && - (Args.hasArg(options::OPT_fprofile_instr_use) || - Args.hasArg(options::OPT_fprofile_instr_use_EQ))) - D.Diag(diag::err_drv_argument_not_allowed_with) - << "-fprofile-instr-generate" - << "-fprofile-instr-use"; - - if (Arg *A = Args.getLastArg(options::OPT_fprofile_instr_generate_EQ)) - A->render(Args, CmdArgs); - else - Args.AddAllArgs(CmdArgs, options::OPT_fprofile_instr_generate); - - if (Arg *A = Args.getLastArg(options::OPT_fprofile_instr_use_EQ)) - A->render(Args, CmdArgs); - else if (Args.hasArg(options::OPT_fprofile_instr_use)) - CmdArgs.push_back("-fprofile-instr-use=pgo-data"); - - if (Args.hasArg(options::OPT_ftest_coverage) || - Args.hasArg(options::OPT_coverage)) - CmdArgs.push_back("-femit-coverage-notes"); - if (Args.hasFlag(options::OPT_fprofile_arcs, options::OPT_fno_profile_arcs, - false) || - Args.hasArg(options::OPT_coverage)) - CmdArgs.push_back("-femit-coverage-data"); - - if (Args.hasArg(options::OPT_fcoverage_mapping) && - !(Args.hasArg(options::OPT_fprofile_instr_generate) || - Args.hasArg(options::OPT_fprofile_instr_generate_EQ))) - D.Diag(diag::err_drv_argument_only_allowed_with) - << "-fcoverage-mapping" - << "-fprofile-instr-generate"; - - if (Args.hasArg(options::OPT_fcoverage_mapping)) - CmdArgs.push_back("-fcoverage-mapping"); - - if (C.getArgs().hasArg(options::OPT_c) || - C.getArgs().hasArg(options::OPT_S)) { - if (Output.isFilename()) { - CmdArgs.push_back("-coverage-file"); - SmallString<128> CoverageFilename; - if (Arg *FinalOutput = C.getArgs().getLastArg(options::OPT_o)) { - CoverageFilename = FinalOutput->getValue(); - } else { - CoverageFilename = llvm::sys::path::filename(Output.getBaseInput()); - } - if (llvm::sys::path::is_relative(CoverageFilename)) { - SmallString<128> Pwd; - if (!llvm::sys::fs::current_path(Pwd)) { - llvm::sys::path::append(Pwd, CoverageFilename); - CoverageFilename.swap(Pwd); - } - } - CmdArgs.push_back(Args.MakeArgString(CoverageFilename)); - } - } + addPGOAndCoverageFlags(C, D, Output, Args, CmdArgs); // Pass options for controlling the default header search paths. if (Args.hasArg(options::OPT_nostdinc)) { @@ -3915,6 +3966,16 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA, case OMPRT_IOMP5: // Clang can generate useful OpenMP code for these two runtime libraries. CmdArgs.push_back("-fopenmp"); + + // If no option regarding the use of TLS in OpenMP codegeneration is + // given, decide a default based on the target. Otherwise rely on the + // options and pass the right information to the frontend. + if (!Args.hasFlag(options::OPT_fopenmp_use_tls, + options::OPT_fnoopenmp_use_tls, + getToolChain().getArch() == llvm::Triple::ppc || + getToolChain().getArch() == llvm::Triple::ppc64 || + getToolChain().getArch() == llvm::Triple::ppc64le)) + CmdArgs.push_back("-fnoopenmp-use-tls"); break; default: // By default, if Clang doesn't know how to generate useful OpenMP code @@ -4070,17 +4131,6 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA, CmdArgs.push_back("-arm-restrict-it"); } - if (TT.getArch() == llvm::Triple::arm || - TT.getArch() == llvm::Triple::thumb) { - if (Arg *A = Args.getLastArg(options::OPT_mlong_calls, - options::OPT_mno_long_calls)) { - if (A->getOption().matches(options::OPT_mlong_calls)) { - CmdArgs.push_back("-backend-option"); - CmdArgs.push_back("-arm-long-calls"); - } - } - } - // Forward -f options with positive and negative forms; we translate // these by hand. if (Arg *A = Args.getLastArg(options::OPT_fprofile_sample_use_EQ)) { @@ -4172,27 +4222,27 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA, // -fmodule-cache-path specifies where our implicitly-built module files // should be written. - SmallString<128> ModuleCachePath; + SmallString<128> Path; if (Arg *A = Args.getLastArg(options::OPT_fmodules_cache_path)) - ModuleCachePath = A->getValue(); + Path = A->getValue(); if (HaveModules) { if (C.isForDiagnostics()) { // When generating crash reports, we want to emit the modules along with // the reproduction sources, so we ignore any provided module path. - ModuleCachePath = Output.getFilename(); - llvm::sys::path::replace_extension(ModuleCachePath, ".cache"); - llvm::sys::path::append(ModuleCachePath, "modules"); - } else if (ModuleCachePath.empty()) { + Path = Output.getFilename(); + llvm::sys::path::replace_extension(Path, ".cache"); + llvm::sys::path::append(Path, "modules"); + } else if (Path.empty()) { // No module path was provided: use the default. llvm::sys::path::system_temp_directory(/*erasedOnReboot=*/false, - ModuleCachePath); - llvm::sys::path::append(ModuleCachePath, "org.llvm.clang."); - appendUserToPath(ModuleCachePath); - llvm::sys::path::append(ModuleCachePath, "ModuleCache"); + Path); + llvm::sys::path::append(Path, "org.llvm.clang."); + appendUserToPath(Path); + llvm::sys::path::append(Path, "ModuleCache"); } const char Arg[] = "-fmodules-cache-path="; - ModuleCachePath.insert(ModuleCachePath.begin(), Arg, Arg + strlen(Arg)); - CmdArgs.push_back(Args.MakeArgString(ModuleCachePath)); + Path.insert(Path.begin(), Arg, Arg + strlen(Arg)); + CmdArgs.push_back(Args.MakeArgString(Path)); } // When building modules and generating crashdumps, we need to dump a module @@ -4774,14 +4824,12 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA, assert(Output.isNothing() && "Invalid output."); } - for (const auto &II : Inputs) { - addDashXForInput(Args, II, CmdArgs); + addDashXForInput(Args, Input, CmdArgs); - if (II.isFilename()) - CmdArgs.push_back(II.getFilename()); - else - II.getInputArg().renderAsInput(Args, CmdArgs); - } + if (Input.isFilename()) + CmdArgs.push_back(Input.getFilename()); + else + Input.getInputArg().renderAsInput(Args, CmdArgs); Args.AddAllArgs(CmdArgs, options::OPT_undef); @@ -4819,6 +4867,16 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA, CmdArgs.push_back(SplitDwarfOut); } + // Host-side cuda compilation receives device-side outputs as Inputs[1...]. + // Include them with -fcuda-include-gpubinary. + if (IsCuda && Inputs.size() > 1) + for (InputInfoList::const_iterator it = std::next(Inputs.begin()), + ie = Inputs.end(); + it != ie; ++it) { + CmdArgs.push_back("-fcuda-include-gpubinary"); + CmdArgs.push_back(it->getFilename()); + } + // Finally add the compile command to the compilation. if (Args.hasArg(options::OPT__SLASH_fallback) && Output.getType() == types::TY_Object && @@ -5005,6 +5063,7 @@ struct EHFlags { /// The default is /EHs-c-, meaning cleanups are disabled. static EHFlags parseClangCLEHFlags(const Driver &D, const ArgList &Args) { EHFlags EH; + std::vector<std::string> EHArgs = Args.getAllArgValues(options::OPT__SLASH_EH); for (auto EHVal : EHArgs) { @@ -5026,6 +5085,11 @@ static EHFlags parseClangCLEHFlags(const Driver &D, const ArgList &Args) { break; } } + + // FIXME: Disable C++ EH completely, until it becomes more reliable. Users + // can use -Xclang to manually enable C++ EH until then. + EH = EHFlags(); + return EH; } @@ -8199,6 +8263,8 @@ void nacltools::Linker::ConstructJob(Compilation &C, const JobAction &JA, CmdArgs.push_back("armelf_nacl"); else if (Arch == llvm::Triple::x86_64) CmdArgs.push_back("elf_x86_64_nacl"); + else if (Arch == llvm::Triple::mipsel) + CmdArgs.push_back("mipselelf_nacl"); else D.Diag(diag::err_target_unsupported_arch) << ToolChain.getArchName() << "Native Client"; @@ -8260,6 +8326,13 @@ void nacltools::Linker::ConstructJob(Compilation &C, const JobAction &JA, // in the group for C++. if (Args.hasArg(options::OPT_pthread) || Args.hasArg(options::OPT_pthreads) || D.CCCIsCXX()) { + // Gold, used by Mips, handles nested groups differently than ld, and + // without '-lnacl' it prefers symbols from libpthread.a over libnacl.a, + // which is not a desired behaviour here. + // See https://sourceware.org/ml/binutils/2015-03/msg00034.html + if (getToolChain().getArch() == llvm::Triple::mipsel) + CmdArgs.push_back("-lnacl"); + CmdArgs.push_back("-lpthread"); } @@ -8270,6 +8343,13 @@ void nacltools::Linker::ConstructJob(Compilation &C, const JobAction &JA, else CmdArgs.push_back("-lgcc_s"); CmdArgs.push_back("--no-as-needed"); + + // Mips needs to create and use pnacl_legacy library that contains + // definitions from bitcode/pnaclmm.c and definitions for + // __nacl_tp_tls_offset() and __nacl_tp_tdb_offset(). + if (getToolChain().getArch() == llvm::Triple::mipsel) + CmdArgs.push_back("-lpnacl_legacy"); + CmdArgs.push_back("--end-group"); } @@ -8863,6 +8943,12 @@ void MinGW::Linker::ConstructJob(Compilation &C, const JobAction &JA, // handled somewhere else. Args.ClaimAllArgs(options::OPT_w); + StringRef LinkerName = Args.getLastArgValue(options::OPT_fuse_ld_EQ, "ld"); + if (LinkerName.equals_lower("lld")) { + CmdArgs.push_back("-flavor"); + CmdArgs.push_back("gnu"); + } + if (!D.SysRoot.empty()) CmdArgs.push_back(Args.MakeArgString("--sysroot=" + D.SysRoot)); @@ -8874,6 +8960,8 @@ void MinGW::Linker::ConstructJob(Compilation &C, const JobAction &JA, CmdArgs.push_back("i386pe"); if (TC.getArch() == llvm::Triple::x86_64) CmdArgs.push_back("i386pep"); + if (TC.getArch() == llvm::Triple::arm) + CmdArgs.push_back("thumb2pe"); if (Args.hasArg(options::OPT_mwindows)) { CmdArgs.push_back("--subsystem"); @@ -8968,10 +9056,8 @@ void MinGW::Linker::ConstructJob(Compilation &C, const JobAction &JA, if (Args.hasArg(options::OPT_pg)) CmdArgs.push_back("-lgmon"); - // FIXME: what to do about pthreads library? - // Currently required for OpenMP and posix-threading libgcc, - // does not exists in mingw.org. - //CmdArgs.push_back("-lpthread"); + if (Args.hasArg(options::OPT_pthread)) + CmdArgs.push_back("-lpthread"); // add system libraries if (Args.hasArg(options::OPT_mwindows)) { @@ -8985,7 +9071,7 @@ void MinGW::Linker::ConstructJob(Compilation &C, const JobAction &JA, if (Args.hasArg(options::OPT_static)) CmdArgs.push_back("--end-group"); - else + else if (!LinkerName.equals_lower("lld")) AddLibGCC(Args, CmdArgs); } @@ -8996,7 +9082,7 @@ void MinGW::Linker::ConstructJob(Compilation &C, const JobAction &JA, CmdArgs.push_back(Args.MakeArgString(TC.GetFilePath("crtend.o"))); } } - const char *Exec = Args.MakeArgString(TC.GetProgramPath("ld")); + const char *Exec = Args.MakeArgString(TC.GetProgramPath(LinkerName.data())); C.addCommand(llvm::make_unique<Command>(JA, *this, Exec, CmdArgs)); } @@ -9053,7 +9139,9 @@ void XCore::Linker::ConstructJob(Compilation &C, const JobAction &JA, if (Args.hasArg(options::OPT_v)) CmdArgs.push_back("-v"); - if (exceptionSettings(Args, getToolChain().getTriple())) + // Pass -fexceptions through to the linker if it was present. + if (Args.hasFlag(options::OPT_fexceptions, options::OPT_fno_exceptions, + false)) CmdArgs.push_back("-fexceptions"); AddLinkerInputs(getToolChain(), Inputs, Args, CmdArgs); diff --git a/lib/Driver/Types.cpp b/lib/Driver/Types.cpp index 7b28145..2085b01 100644 --- a/lib/Driver/Types.cpp +++ b/lib/Driver/Types.cpp @@ -86,6 +86,7 @@ bool types::isAcceptedByClang(ID Id) { case TY_C: case TY_PP_C: case TY_CL: case TY_CUDA: case TY_PP_CUDA: + case TY_CUDA_DEVICE: case TY_ObjC: case TY_PP_ObjC: case TY_PP_ObjC_Alias: case TY_CXX: case TY_PP_CXX: case TY_ObjCXX: case TY_PP_ObjCXX: case TY_PP_ObjCXX_Alias: @@ -122,7 +123,19 @@ bool types::isCXX(ID Id) { case TY_ObjCXX: case TY_PP_ObjCXX: case TY_PP_ObjCXX_Alias: case TY_CXXHeader: case TY_PP_CXXHeader: case TY_ObjCXXHeader: case TY_PP_ObjCXXHeader: - case TY_CUDA: case TY_PP_CUDA: + case TY_CUDA: case TY_PP_CUDA: case TY_CUDA_DEVICE: + return true; + } +} + +bool types::isCuda(ID Id) { + switch (Id) { + default: + return false; + + case TY_CUDA: + case TY_PP_CUDA: + case TY_CUDA_DEVICE: return true; } } @@ -206,10 +219,12 @@ void types::getCompilationPhases(ID Id, llvm::SmallVectorImpl<phases::ID> &P) { P.push_back(phases::Compile); P.push_back(phases::Backend); } - P.push_back(phases::Assemble); + if (Id != TY_CUDA_DEVICE) + P.push_back(phases::Assemble); } } - if (!onlyPrecompileType(Id)) { + + if (!onlyPrecompileType(Id) && Id != TY_CUDA_DEVICE) { P.push_back(phases::Link); } assert(0 < P.size() && "Not enough phases in list"); |