aboutsummaryrefslogtreecommitdiff
path: root/lib/Driver/Driver.cpp
diff options
context:
space:
mode:
Diffstat (limited to 'lib/Driver/Driver.cpp')
-rw-r--r--lib/Driver/Driver.cpp367
1 files changed, 230 insertions, 137 deletions
diff --git a/lib/Driver/Driver.cpp b/lib/Driver/Driver.cpp
index 180c412bd791..4f8481c0beec 100644
--- a/lib/Driver/Driver.cpp
+++ b/lib/Driver/Driver.cpp
@@ -11,6 +11,7 @@
#include "InputInfo.h"
#include "ToolChains.h"
#include "clang/Basic/Version.h"
+#include "clang/Basic/VirtualFileSystem.h"
#include "clang/Config/config.h"
#include "clang/Driver/Action.h"
#include "clang/Driver/Compilation.h"
@@ -46,9 +47,11 @@ using namespace clang;
using namespace llvm::opt;
Driver::Driver(StringRef ClangExecutable, StringRef DefaultTargetTriple,
- DiagnosticsEngine &Diags)
- : Opts(createDriverOptTable()), Diags(Diags), Mode(GCCMode),
- SaveTemps(SaveTempsNone), ClangExecutable(ClangExecutable),
+ DiagnosticsEngine &Diags,
+ IntrusiveRefCntPtr<vfs::FileSystem> VFS)
+ : Opts(createDriverOptTable()), Diags(Diags), VFS(VFS), Mode(GCCMode),
+ SaveTemps(SaveTempsNone), LTOMode(LTOK_None),
+ ClangExecutable(ClangExecutable),
SysRoot(DEFAULT_SYSROOT), UseStdLib(true),
DefaultTargetTriple(DefaultTargetTriple),
DriverTitle("clang LLVM compiler"), CCPrintOptionsFilename(nullptr),
@@ -57,8 +60,13 @@ Driver::Driver(StringRef ClangExecutable, StringRef DefaultTargetTriple,
CCGenDiagnostics(false), CCCGenericGCCName(""), CheckInputsExist(true),
CCCUsePCH(true), SuppressMissingInputWarning(false) {
+ // Provide a sane fallback if no VFS is specified.
+ if (!this->VFS)
+ this->VFS = vfs::getRealFileSystem();
+
Name = llvm::sys::path::filename(ClangExecutable);
Dir = llvm::sys::path::parent_path(ClangExecutable);
+ InstalledDir = Dir; // Provide a sensible default installed dir.
// Compute the path to the resource directory.
StringRef ClangResourceDir(CLANG_RESOURCE_DIR);
@@ -174,10 +182,8 @@ phases::ID Driver::getFinalPhase(const DerivedArgList &DAL,
} else if ((PhaseArg = DAL.getLastArg(options::OPT_S))) {
FinalPhase = phases::Backend;
- // -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))) {
+ // -c compilation only runs up to the assembler.
+ } else if ((PhaseArg = DAL.getLastArg(options::OPT_c))) {
FinalPhase = phases::Assemble;
// Otherwise do everything.
@@ -203,6 +209,7 @@ DerivedArgList *Driver::TranslateInputArgs(const InputArgList &Args) const {
DerivedArgList *DAL = new DerivedArgList(Args);
bool HasNostdlib = Args.hasArg(options::OPT_nostdlib);
+ bool HasNodefaultlib = Args.hasArg(options::OPT_nodefaultlibs);
for (Arg *A : Args) {
// Unfortunately, we have to parse some forwarding options (-Xassembler,
// -Xlinker, -Xpreprocessor) because we either integrate their functionality
@@ -217,7 +224,7 @@ DerivedArgList *Driver::TranslateInputArgs(const InputArgList &Args) const {
DAL->AddFlagArg(A, Opts->getOption(options::OPT_Z_Xlinker__no_demangle));
// Add the remaining values as Xlinker arguments.
- for (const StringRef Val : A->getValues())
+ for (StringRef Val : A->getValues())
if (Val != "--no-demangle")
DAL->AddSeparateArg(A, Opts->getOption(options::OPT_Xlinker), Val);
@@ -246,7 +253,7 @@ DerivedArgList *Driver::TranslateInputArgs(const InputArgList &Args) const {
StringRef Value = A->getValue();
// Rewrite unless -nostdlib is present.
- if (!HasNostdlib && Value == "stdc++") {
+ if (!HasNostdlib && !HasNodefaultlib && Value == "stdc++") {
DAL->AddFlagArg(A, Opts->getOption(options::OPT_Z_reserved_lib_stdcxx));
continue;
}
@@ -261,7 +268,7 @@ DerivedArgList *Driver::TranslateInputArgs(const InputArgList &Args) const {
// Pick up inputs via the -- option.
if (A->getOption().matches(options::OPT__DASH_DASH)) {
A->claim();
- for (const StringRef Val : A->getValues())
+ for (StringRef Val : A->getValues())
DAL->append(MakeInputArg(*DAL, Opts, Val));
continue;
}
@@ -327,7 +334,8 @@ static llvm::Triple computeTargetTriple(StringRef DefaultTargetTriple,
}
// Skip further flag support on OSes which don't support '-m32' or '-m64'.
- if (Target.getArchName() == "tce" || Target.getOS() == llvm::Triple::Minix)
+ if (Target.getArch() == llvm::Triple::tce ||
+ Target.getOS() == llvm::Triple::Minix)
return Target;
// Handle pseudo-target flags '-m64', '-mx32', '-m32' and '-m16'.
@@ -360,6 +368,32 @@ static llvm::Triple computeTargetTriple(StringRef DefaultTargetTriple,
return Target;
}
+// \brief Parse the LTO options and record the type of LTO compilation
+// based on which -f(no-)?lto(=.*)? option occurs last.
+void Driver::setLTOMode(const llvm::opt::ArgList &Args) {
+ LTOMode = LTOK_None;
+ if (!Args.hasFlag(options::OPT_flto, options::OPT_flto_EQ,
+ options::OPT_fno_lto, false))
+ return;
+
+ StringRef LTOName("full");
+
+ const Arg *A = Args.getLastArg(options::OPT_flto_EQ);
+ if (A)
+ LTOName = A->getValue();
+
+ LTOMode = llvm::StringSwitch<LTOKind>(LTOName)
+ .Case("full", LTOK_Full)
+ .Case("thin", LTOK_Thin)
+ .Default(LTOK_Unknown);
+
+ if (LTOMode == LTOK_Unknown) {
+ assert(A);
+ Diag(diag::err_drv_unsupported_option_argument) << A->getOption().getName()
+ << A->getValue();
+ }
+}
+
Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) {
llvm::PrettyStackTraceString CrashInfo("Compilation construction");
@@ -387,6 +421,9 @@ Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) {
InputArgList Args = ParseArgStrings(ArgList.slice(1));
+ // Silence driver warnings if requested
+ Diags.setIgnoreAllWarnings(Args.hasArg(options::OPT_w));
+
// -no-canonical-prefixes is used very early in main.
Args.ClaimAllArgs(options::OPT_no_canonical_prefixes);
@@ -411,6 +448,7 @@ Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) {
// clang-cl targets MSVC-style Win32.
llvm::Triple T(DefaultTargetTriple);
T.setOS(llvm::Triple::Win32);
+ T.setVendor(llvm::Triple::PC);
T.setEnvironment(llvm::Triple::MSVC);
DefaultTargetTriple = T.str();
}
@@ -439,6 +477,8 @@ Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) {
.Default(SaveTempsCwd);
}
+ setLTOMode(Args);
+
std::unique_ptr<llvm::opt::InputArgList> UArgs =
llvm::make_unique<InputArgList>(std::move(Args));
@@ -452,6 +492,10 @@ Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) {
// The compilation takes ownership of Args.
Compilation *C = new Compilation(*this, TC, UArgs.release(), TranslatedArgs);
+ C->setCudaDeviceToolChain(
+ &getToolChain(C->getArgs(), llvm::Triple(TC.getTriple().isArch64Bit()
+ ? "nvptx64-nvidia-cuda"
+ : "nvptx-nvidia-cuda")));
if (!HandleImmediateArgs(*C))
return C;
@@ -462,10 +506,9 @@ Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) {
// Construct the list of abstract actions to perform for this compilation. On
// MachO targets this uses the driver-driver and universal actions.
if (TC.getTriple().isOSBinFormatMachO())
- BuildUniversalActions(C->getDefaultToolChain(), C->getArgs(), Inputs,
- C->getActions());
+ BuildUniversalActions(*C, C->getDefaultToolChain(), Inputs);
else
- BuildActions(C->getDefaultToolChain(), C->getArgs(), Inputs,
+ BuildActions(*C, C->getDefaultToolChain(), C->getArgs(), Inputs,
C->getActions());
if (CCCPrintPhases) {
@@ -578,9 +621,9 @@ void Driver::generateCompilationDiagnostics(Compilation &C,
// Darwin OSes this uses the driver-driver and builds universal actions.
const ToolChain &TC = C.getDefaultToolChain();
if (TC.getTriple().isOSBinFormatMachO())
- BuildUniversalActions(TC, C.getArgs(), Inputs, C.getActions());
+ BuildUniversalActions(C, TC, Inputs);
else
- BuildActions(TC, C.getArgs(), Inputs, C.getActions());
+ BuildActions(C, TC, C.getArgs(), Inputs, C.getActions());
BuildJobs(C);
@@ -761,6 +804,9 @@ void Driver::PrintVersion(const Compilation &C, raw_ostream &OS) const {
} else
OS << "Thread model: " << TC.getThreadModel();
OS << '\n';
+
+ // Print out the install directory.
+ OS << "InstalledDir: " << InstalledDir << '\n';
}
/// PrintDiagnosticCategories - Implement the --print-diagnostic-categories
@@ -906,7 +952,7 @@ static unsigned PrintActions1(const Compilation &C, Action *A,
os << '"' << CDA->getGpuArchName() << '"' << ", {"
<< PrintActions1(C, *CDA->begin(), Ids) << "}";
} else {
- ActionList *AL;
+ const ActionList *AL;
if (CudaHostAction *CHA = dyn_cast<CudaHostAction>(A)) {
os << "{" << PrintActions1(C, *CHA->begin(), Ids) << "}"
<< ", gpu binaries ";
@@ -914,12 +960,15 @@ static unsigned PrintActions1(const Compilation &C, Action *A,
} else
AL = &A->getInputs();
- const char *Prefix = "{";
- for (Action *PreRequisite : *AL) {
- os << Prefix << PrintActions1(C, PreRequisite, Ids);
- Prefix = ", ";
- }
- os << "}";
+ if (AL->size()) {
+ const char *Prefix = "{";
+ for (Action *PreRequisite : *AL) {
+ os << Prefix << PrintActions1(C, PreRequisite, Ids);
+ Prefix = ", ";
+ }
+ os << "}";
+ } else
+ os << "{}";
}
unsigned Id = Ids.size();
@@ -945,16 +994,17 @@ static bool ContainsCompileOrAssembleAction(const Action *A) {
isa<AssembleJobAction>(A))
return true;
- for (Action::const_iterator it = A->begin(), ie = A->end(); it != ie; ++it)
- if (ContainsCompileOrAssembleAction(*it))
+ for (const Action *Input : *A)
+ if (ContainsCompileOrAssembleAction(Input))
return true;
return false;
}
-void Driver::BuildUniversalActions(const ToolChain &TC, DerivedArgList &Args,
- const InputList &BAInputs,
- ActionList &Actions) const {
+void Driver::BuildUniversalActions(Compilation &C, const ToolChain &TC,
+ const InputList &BAInputs) const {
+ DerivedArgList &Args = C.getArgs();
+ ActionList &Actions = C.getActions();
llvm::PrettyStackTraceString CrashInfo("Building universal build actions");
// Collect the list of architectures. Duplicates are allowed, but should only
// be handled once (in the order seen).
@@ -983,13 +1033,11 @@ void Driver::BuildUniversalActions(const ToolChain &TC, DerivedArgList &Args,
Archs.push_back(Args.MakeArgString(TC.getDefaultUniversalArchName()));
ActionList SingleActions;
- BuildActions(TC, Args, BAInputs, SingleActions);
+ BuildActions(C, TC, Args, BAInputs, SingleActions);
// Add in arch bindings for every top level action, as well as lipo and
// dsymutil steps if needed.
- for (unsigned i = 0, e = SingleActions.size(); i != e; ++i) {
- Action *Act = SingleActions[i];
-
+ for (Action* Act : SingleActions) {
// Make sure we can lipo this kind of output. If not (and it is an actual
// output) then we disallow, since we can't create an output file with the
// right name without overwriting it. We could remove this oddity by just
@@ -1228,18 +1276,23 @@ 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.
+// 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. In case of partial (i.e preprocessing
+// only) or device-only compilation, each device action is added to /p
+// Actions and /p Current is released. Otherwise the function creates
+// and returns a new CudaHostAction which wraps /p Current 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.");
+buildCudaActions(Compilation &C, DerivedArgList &Args, const Arg *InputArg,
+ std::unique_ptr<Action> HostAction, ActionList &Actions) {
+ Arg *PartialCompilationArg = Args.getLastArg(options::OPT_cuda_host_only,
+ options::OPT_cuda_device_only);
+ // Host-only compilation case.
+ if (PartialCompilationArg &&
+ PartialCompilationArg->getOption().matches(options::OPT_cuda_host_only))
+ return std::unique_ptr<Action>(
+ new CudaHostAction(std::move(HostAction), {}));
// Collect all cuda_gpu_arch parameters, removing duplicates.
SmallVector<const char *, 4> GpuArchList;
@@ -1259,20 +1312,22 @@ buildCudaActions(const Driver &D, const ToolChain &TC, DerivedArgList &Args,
// Replicate inputs for each GPU architecture.
Driver::InputList CudaDeviceInputs;
- for (unsigned i = 0, e = GpuArchList.size(); i != e; ++i)
+ 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.
+ assert(C.getCudaDeviceToolChain() &&
+ "Missing toolchain for device-side compilation.");
ActionList CudaDeviceActions;
- D.BuildActions(TC, Args, CudaDeviceInputs, CudaDeviceActions);
+ C.getDriver().BuildActions(C, *C.getCudaDeviceToolChain(), 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) {
+ for (unsigned I = 0, E = GpuArchList.size(); I != E; ++I) {
+ if (CudaDeviceActions[I]->getKind() != Action::BackendJobClass) {
PartialCompilation = true;
break;
}
@@ -1280,6 +1335,7 @@ buildCudaActions(const Driver &D, const ToolChain &TC, DerivedArgList &Args,
// Figure out what to do with device actions -- pass them as inputs to the
// host action or run each of them independently.
+ bool DeviceOnlyCompilation = PartialCompilationArg != nullptr;
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
@@ -1288,35 +1344,37 @@ buildCudaActions(const Driver &D, const ToolChain &TC, DerivedArgList &Args,
// -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);
+ C.getDriver().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));
+ 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));
- }
+ HostAction.reset(nullptr);
+ return HostAction;
+ }
+
+ // 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(HostAction), DeviceActions));
}
-void Driver::BuildActions(const ToolChain &TC, DerivedArgList &Args,
- const InputList &Inputs, ActionList &Actions) const {
+void Driver::BuildActions(Compilation &C, const ToolChain &TC,
+ DerivedArgList &Args, const InputList &Inputs,
+ ActionList &Actions) const {
llvm::PrettyStackTraceString CrashInfo("Building compilation actions");
if (!SuppressMissingInputWarning && Inputs.empty()) {
@@ -1373,9 +1431,9 @@ void Driver::BuildActions(const ToolChain &TC, DerivedArgList &Args,
ActionList LinkerInputs;
llvm::SmallVector<phases::ID, phases::MaxNumberOfPhases> PL;
- for (unsigned i = 0, e = Inputs.size(); i != e; ++i) {
- types::ID InputType = Inputs[i].first;
- const Arg *InputArg = Inputs[i].second;
+ for (auto &I : Inputs) {
+ types::ID InputType = I.first;
+ const Arg *InputArg = I.second;
PL.clear();
types::getCompilationPhases(InputType, PL);
@@ -1412,24 +1470,12 @@ 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;
- }
+ phases::ID CudaInjectionPhase = FinalPhase;
+ for (const auto &Phase : PL)
+ if (Phase <= FinalPhase && Phase == phases::Compile) {
+ CudaInjectionPhase = Phase;
+ break;
+ }
// Build the pipeline for this file.
std::unique_ptr<Action> Current(new InputAction(*InputArg, InputType));
@@ -1457,10 +1503,9 @@ 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 (InputType == types::TY_CUDA && Phase == CudaInjectionPhase) {
+ Current =
+ buildCudaActions(C, Args, InputArg, std::move(Current), Actions);
if (!Current)
break;
}
@@ -1487,6 +1532,10 @@ void Driver::BuildActions(const ToolChain &TC, DerivedArgList &Args,
// Claim ignored clang-cl options.
Args.ClaimAllArgs(options::OPT_cl_ignored_Group);
+
+ // Claim --cuda-host-only arg which may be passed to non-CUDA
+ // compilations and should not trigger warnings there.
+ Args.ClaimAllArgs(options::OPT_cuda_host_only);
}
std::unique_ptr<Action>
@@ -1551,7 +1600,7 @@ Driver::ConstructPhaseAction(const ToolChain &TC, const ArgList &Args,
types::TY_LLVM_BC);
}
case phases::Backend: {
- if (IsUsingLTO(Args)) {
+ if (isUsingLTO()) {
types::ID Output =
Args.hasArg(options::OPT_S) ? types::TY_LTO_IR : types::TY_LTO_BC;
return llvm::make_unique<BackendJobAction>(std::move(Input), Output);
@@ -1572,10 +1621,6 @@ Driver::ConstructPhaseAction(const ToolChain &TC, const ArgList &Args,
llvm_unreachable("invalid phase in ConstructPhaseAction");
}
-bool Driver::IsUsingLTO(const ArgList &Args) const {
- return Args.hasFlag(options::OPT_flto, options::OPT_fno_lto, false);
-}
-
void Driver::BuildJobs(Compilation &C) const {
llvm::PrettyStackTraceString CrashInfo("Building compilation jobs");
@@ -1668,10 +1713,17 @@ void Driver::BuildJobs(Compilation &C) const {
}
}
-static const Tool *SelectToolForJob(Compilation &C, bool SaveTemps,
+// Returns a Tool for a given JobAction. In case the action and its
+// predecessors can be combined, updates Inputs with the inputs of the
+// first combined action. If one of the collapsed actions is a
+// CudaHostAction, updates CollapsedCHA with the pointer to it so the
+// caller can deal with extra handling such action requires.
+static const Tool *selectToolForJob(Compilation &C, bool SaveTemps,
const ToolChain *TC, const JobAction *JA,
- const ActionList *&Inputs) {
+ const ActionList *&Inputs,
+ const CudaHostAction *&CollapsedCHA) {
const Tool *ToolForJob = nullptr;
+ CollapsedCHA = nullptr;
// See if we should look for a compiler with an integrated assembler. We match
// bottom up, so what we are actually looking for is an assembler job with a
@@ -1688,13 +1740,19 @@ static const Tool *SelectToolForJob(Compilation &C, bool SaveTemps,
// checking the backend tool, check if the tool for the CompileJob
// has an integrated assembler.
const ActionList *BackendInputs = &(*Inputs)[0]->getInputs();
- JobAction *CompileJA = cast<CompileJobAction>(*BackendInputs->begin());
+ // Compile job may be wrapped in CudaHostAction, extract it if
+ // that's the case and update CollapsedCHA if we combine phases.
+ CudaHostAction *CHA = dyn_cast<CudaHostAction>(*BackendInputs->begin());
+ JobAction *CompileJA =
+ cast<CompileJobAction>(CHA ? *CHA->begin() : *BackendInputs->begin());
+ assert(CompileJA && "Backend job is not preceeded by compile job.");
const Tool *Compiler = TC->SelectTool(*CompileJA);
if (!Compiler)
return nullptr;
if (Compiler->hasIntegratedAssembler()) {
- Inputs = &(*BackendInputs)[0]->getInputs();
+ Inputs = &CompileJA->getInputs();
ToolForJob = Compiler;
+ CollapsedCHA = CHA;
}
}
@@ -1704,19 +1762,19 @@ 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;
- // 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());
-
+ // Compile job may be wrapped in CudaHostAction, extract it if
+ // that's the case and update CollapsedCHA if we combine phases.
+ CudaHostAction *CHA = dyn_cast<CudaHostAction>(*Inputs->begin());
+ JobAction *CompileJA =
+ cast<CompileJobAction>(CHA ? *CHA->begin() : *Inputs->begin());
+ assert(CompileJA && "Backend job is not preceeded by compile job.");
const Tool *Compiler = TC->SelectTool(*CompileJA);
if (!Compiler)
return nullptr;
if (!Compiler->canEmitIR() || !SaveTemps) {
- Inputs = &(*Inputs)[0]->getInputs();
+ Inputs = &CompileJA->getInputs();
ToolForJob = Compiler;
+ CollapsedCHA = CHA;
}
}
@@ -1749,7 +1807,7 @@ void Driver::BuildJobsForAction(Compilation &C, const Action *A,
InputInfo II;
// Append outputs of device jobs to the input list.
for (const Action *DA : CHA->getDeviceActions()) {
- BuildJobsForAction(C, DA, TC, "", AtTopLevel,
+ BuildJobsForAction(C, DA, TC, nullptr, AtTopLevel,
/*MultipleArchs*/ false, LinkingOutput, II);
CudaDeviceInputInfos.push_back(II);
}
@@ -1789,13 +1847,10 @@ void Driver::BuildJobsForAction(Compilation &C, const Action *A,
}
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),
+ // Initial processing of CudaDeviceAction carries host params.
+ // Call BuildJobsForAction() again, now with correct device parameters.
+ assert(CDA->getGpuArchName() && "No GPU name in device action.");
+ BuildJobsForAction(C, *CDA->begin(), C.getCudaDeviceToolChain(),
CDA->getGpuArchName(), CDA->isAtTopLevel(),
/*MultipleArchs*/ true, LinkingOutput, Result);
return;
@@ -1804,10 +1859,23 @@ void Driver::BuildJobsForAction(Compilation &C, const Action *A,
const ActionList *Inputs = &A->getInputs();
const JobAction *JA = cast<JobAction>(A);
- const Tool *T = SelectToolForJob(C, isSaveTempsEnabled(), TC, JA, Inputs);
+ const CudaHostAction *CollapsedCHA = nullptr;
+ const Tool *T =
+ selectToolForJob(C, isSaveTempsEnabled(), TC, JA, Inputs, CollapsedCHA);
if (!T)
return;
+ // If we've collapsed action list that contained CudaHostAction we
+ // need to build jobs for device-side inputs it may have held.
+ if (CollapsedCHA) {
+ InputInfo II;
+ for (const Action *DA : CollapsedCHA->getDeviceActions()) {
+ BuildJobsForAction(C, DA, TC, "", AtTopLevel,
+ /*MultipleArchs*/ false, LinkingOutput, II);
+ CudaDeviceInputInfos.push_back(II);
+ }
+ }
+
// Only use pipes when there is exactly one input.
InputInfoList InputInfos;
for (const Action *Input : *Inputs) {
@@ -2091,6 +2159,11 @@ void Driver::generatePrefixedToolNames(
// FIXME: Needs a better variable than DefaultTargetTriple
Names.emplace_back(DefaultTargetTriple + "-" + Tool);
Names.emplace_back(Tool);
+
+ // Allow the discovery of tools prefixed with LLVM's default target triple.
+ std::string LLVMDefaultTargetTriple = llvm::sys::getDefaultTargetTriple();
+ if (LLVMDefaultTargetTriple != DefaultTargetTriple)
+ Names.emplace_back(LLVMDefaultTargetTriple + "-" + Tool);
}
static bool ScanDirForExecutable(SmallString<128> &Dir,
@@ -2163,6 +2236,8 @@ const ToolChain &Driver::getToolChain(const ArgList &Args,
case llvm::Triple::Darwin:
case llvm::Triple::MacOSX:
case llvm::Triple::IOS:
+ case llvm::Triple::TvOS:
+ case llvm::Triple::WatchOS:
TC = new toolchains::DarwinClang(*this, Target, Args);
break;
case llvm::Triple::DragonFly:
@@ -2185,16 +2260,22 @@ const ToolChain &Driver::getToolChain(const ArgList &Args,
break;
case llvm::Triple::Linux:
if (Target.getArch() == llvm::Triple::hexagon)
- TC = new toolchains::Hexagon_TC(*this, Target, Args);
+ TC = new toolchains::HexagonToolChain(*this, Target, Args);
+ else if ((Target.getVendor() == llvm::Triple::MipsTechnologies) &&
+ !Target.hasEnvironment())
+ TC = new toolchains::MipsLLVMToolChain(*this, Target, Args);
else
TC = new toolchains::Linux(*this, Target, Args);
break;
case llvm::Triple::NaCl:
- TC = new toolchains::NaCl_TC(*this, Target, Args);
+ TC = new toolchains::NaClToolChain(*this, Target, Args);
break;
case llvm::Triple::Solaris:
TC = new toolchains::Solaris(*this, Target, Args);
break;
+ case llvm::Triple::AMDHSA:
+ TC = new toolchains::AMDGPUToolChain(*this, Target, Args);
+ break;
case llvm::Triple::Win32:
switch (Target.getEnvironment()) {
default:
@@ -2220,24 +2301,36 @@ const ToolChain &Driver::getToolChain(const ArgList &Args,
case llvm::Triple::CUDA:
TC = new toolchains::CudaToolChain(*this, Target, Args);
break;
+ case llvm::Triple::PS4:
+ TC = new toolchains::PS4CPU(*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.
- if (Target.getArchName() == "tce")
+ switch (Target.getArch()) {
+ case llvm::Triple::tce:
TC = new toolchains::TCEToolChain(*this, Target, Args);
- else if (Target.getArch() == llvm::Triple::hexagon)
- TC = new toolchains::Hexagon_TC(*this, Target, Args);
- else if (Target.getArch() == llvm::Triple::xcore)
- TC = new toolchains::XCore(*this, Target, Args);
- else if (Target.getArch() == llvm::Triple::shave)
- TC = new toolchains::SHAVEToolChain(*this, Target, Args);
- else if (Target.isOSBinFormatELF())
- TC = new toolchains::Generic_ELF(*this, Target, Args);
- else if (Target.isOSBinFormatMachO())
- TC = new toolchains::MachO(*this, Target, Args);
- else
- TC = new toolchains::Generic_GCC(*this, Target, Args);
- break;
+ break;
+ case llvm::Triple::hexagon:
+ TC = new toolchains::HexagonToolChain(*this, Target, Args);
+ break;
+ case llvm::Triple::xcore:
+ TC = new toolchains::XCoreToolChain(*this, Target, Args);
+ break;
+ case llvm::Triple::wasm32:
+ case llvm::Triple::wasm64:
+ TC = new toolchains::WebAssembly(*this, Target, Args);
+ break;
+ default:
+ if (Target.getVendor() == llvm::Triple::Myriad)
+ TC = new toolchains::MyriadToolChain(*this, Target, Args);
+ else if (Target.isOSBinFormatELF())
+ TC = new toolchains::Generic_ELF(*this, Target, Args);
+ else if (Target.isOSBinFormatMachO())
+ TC = new toolchains::MachO(*this, Target, Args);
+ else
+ TC = new toolchains::Generic_GCC(*this, Target, Args);
+ }
}
}
return *TC;
generated by cgit v1.2.3 (git 2.25.1) at 2024-10-23 07:26:39 +0000