// Out of line constructor provides default values for pass options and
// registers all common codegen passes.
TargetPassConfig::TargetPassConfig(LLVMTargetMachine &TM, PassManagerBase &pm)
: ImmutablePass(ID), PM(&pm), TM(&TM) {
Impl = new PassConfigImpl();
// Register all target independent codegen passes to activate their PassIDs,
// including this pass itself.
initializeCodeGen(*PassRegistry::getPassRegistry());
// Also register alias analysis passes required by codegen passes.
initializeBasicAAWrapperPassPass(*PassRegistry::getPassRegistry());
initializeAAResultsWrapperPassPass(*PassRegistry::getPassRegistry());
if (StringRef(PrintMachineInstrs.getValue()).equals(""))
TM.Options.PrintMachineCode = true;
if (EnableIPRA.getNumOccurrences())
TM.Options.EnableIPRA = EnableIPRA;
else {
// If not explicitly specified, use target default.
TM.Options.EnableIPRA |= TM.useIPRA();
}
if (TM.Options.EnableIPRA)
setRequiresCodeGenSCCOrder();
if (EnableGlobalISelAbort.getNumOccurrences())
TM.Options.GlobalISelAbort = EnableGlobalISelAbort;
setStartStopPasses();
}
int main(int argc, char **argv) {
sys::PrintStackTraceOnErrorSignal();
llvm::PrettyStackTraceProgram X(argc, argv);
llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
LLVMContext &Context = getGlobalContext();
cl::ParseCommandLineOptions(argc, argv, "Bitcode strip tool\n");
int input_fd = open(InputFilename.c_str(), O_RDONLY);
unsigned char *wrapper = NULL;
char *bitcode = NULL;
// Read bitcode wrapper
size_t bitcode_size = 0;
size_t wrapper_size = ReadBitcodeWrapper(input_fd, &wrapper, bitcode_size);
// Read bitcode
bitcode = (char*) calloc(1, bitcode_size);
size_t nread = read(input_fd, (void*) bitcode, bitcode_size);
if (nread != bitcode_size) {
errs() << "Could not read bitcode\n";
return 1;
}
// Strip bitcode
std::string BCString;
StripBitcode(bitcode, bitcode_size, BCString, Context);
// Update bitcode size
WriteInt32(wrapper, 12, BCString.length());
// Default to input filename
if (OutputFilename.empty())
OutputFilename.getValue() = InputFilename;
// Output stripped bitcode
std::error_code EC;
tool_output_file Out(OutputFilename.c_str(), EC, sys::fs::F_None);
if (EC) {
errs() << EC.message() << '\n';
return 1;
}
Out.os().write((const char *) wrapper, wrapper_size);
Out.os().write(BCString.c_str(), BCString.length());
Out.keep();
// Clean up
free((void *) wrapper);
free((void *) bitcode);
close(input_fd);
return 0;
}
int main(int argc, const char **argv) {
CommonOptionsParser OptionsParser(argc, argv, NoGlobalStyleCategory);
ClangTool Tool(OptionsParser.getCompilations(),
OptionsParser.getSourcePathList());
MatchFinder finder;
// Snarf abstract-only namespaces from the environment.
std::vector<std::string> abstract_namespaces_v, banned_namespaces_v;
std::copy(abstract_namespaces.begin(), abstract_namespaces.end(), std::back_inserter(abstract_namespaces_v));
std::copy(banned_namespaces.begin(), banned_namespaces.end(), std::back_inserter(banned_namespaces_v));
std::unique_ptr<RuleCheckerBase> rules[] = {
make_unique<DisallowNew>(),
make_unique<DisallowDelete>(),
make_unique<DisallowGlobals>(),
make_unique<DisallowNonAbstract>(abstract_namespaces_v),
make_unique<DisallowCoupling>(banned_namespaces_v)
};
size_t rules_size = sizeof(rules) / sizeof(rules[0]);
auto rules_begin = &rules[0];
auto rules_end = &rules[rules_size];
std::vector<std::string> analyze_paths_v;
std::copy(analyze_paths.begin(), analyze_paths.end(), std::back_inserter(analyze_paths_v));
for (size_t i = 0; i < sizeof(rules) / sizeof(rules[0]); ++i) {
auto &rule = rules[i];
rule->setAnalyzePaths(analyze_paths_v);
rule->SetupMatches(finder);
rule->getPrinter().setDebug(Debug.getValue());
}
#ifndef NDEBUG
llvm::DebugFlag = Debug.getValue();
#endif
return Tool.run(newFrontendActionFactory(&finder).get()) ||
(Werror.getValue() &&
std::any_of
(rules_begin, rules_end,
[] (const std::unique_ptr<RuleCheckerBase> &rule) {
return rule->getPrinter().getWarnings();
}));
}
/*!
* Initialization of pointer analysis
*/
void PointerAnalysis::initialize(Module& module) {
/// whether we have already built PAG
if(pag == NULL) {
/// run class hierarchy analysis
chgraph = new CHGraph();
chgraph->buildCHG(module);
DBOUT(DGENERAL, outs() << pasMsg("Building Symbol table ...\n"));
SymbolTableInfo* symTable = SymbolTableInfo::Symbolnfo();
symTable->buildMemModel(module);
DBOUT(DGENERAL, outs() << pasMsg("Building PAG ...\n"));
if (!Graphtxt.getValue().empty()) {
PAGBuilderFromFile fileBuilder(Graphtxt.getValue());
pag = fileBuilder.build();
} else {
PAGBuilder builder;
pag = builder.build(module);
}
// dump the PAG graph
if (dumpGraph())
PAG::getPAG()->dump("pag_initial");
// print to command line of the PAG graph
if (PAGPrint)
pag->print();
}
typeSystem = new TypeSystem(pag);
mod = &module;
/// initialise pta call graph
if(EnableThreadCallGraph)
ptaCallGraph = new ThreadCallGraph(mod);
else
ptaCallGraph = new PTACallGraph(mod);
callGraphSCCDetection();
}
bool TFA::runOnModule(Module &M) {
if (Input.getValue() == llvm::cl::BOU_UNSET || Input.getValue()
== llvm::cl::BOU_TRUE) {
InputValues &IV = getAnalysis<InputValues> ();
inputDepValues = IV.getInputDepValues();
} else {
InputDep &IV = getAnalysis<InputDep> ();
inputDepValues = IV.getInputDepValues();
}
bSSA &bssa = getAnalysis<bSSA> ();
depGraph = bssa.newGraph;
DEBUG( // display dependence graph
string Error;
std::string tmp = M.getModuleIdentifier();
replace(tmp.begin(), tmp.end(), '\\', '_');
std::string Filename = "/tmp/" + tmp + ".dot";
//Print dependency graph (in dot format)
depGraph->toDot(M.getModuleIdentifier(), Filename);
DisplayGraph(Filename, true, GraphProgram::DOT);
);
void Executor::initTimers() {
static bool first = true;
if (first) {
first = false;
setupHandler();
}
if (MaxTime) {
addTimer(new HaltTimer(this), MaxTime.getValue());
}
}
/// \brief Emit an inline hint if \p F is globally hot or cold.
///
/// If \p F consumes a significant fraction of samples (indicated by
/// SampleProfileGlobalHotThreshold), apply the InlineHint attribute for the
/// inliner to consider the function hot.
///
/// If \p F consumes a small fraction of samples (indicated by
/// SampleProfileGlobalColdThreshold), apply the Cold attribute for the inliner
/// to consider the function cold.
///
/// FIXME - This setting of inline hints is sub-optimal. Instead of marking a
/// function globally hot or cold, we should be annotating individual callsites.
/// This is not currently possible, but work on the inliner will eventually
/// provide this ability. See http://reviews.llvm.org/D15003 for details and
/// discussion.
///
/// \returns True if either attribute was applied to \p F.
bool SampleProfileLoader::emitInlineHints(Function &F) {
if (TotalCollectedSamples == 0)
return false;
uint64_t FunctionSamples = Samples->getTotalSamples();
double SamplesPercent =
(double)FunctionSamples / (double)TotalCollectedSamples * 100.0;
// If the function collected more samples than the hot threshold, mark
// it globally hot.
if (SamplesPercent >= SampleProfileGlobalHotThreshold) {
F.addFnAttr(llvm::Attribute::InlineHint);
std::string Msg;
raw_string_ostream S(Msg);
S << "Applied inline hint to globally hot function '" << F.getName()
<< "' with " << format("%.2f", SamplesPercent)
<< "% of samples (threshold: "
<< format("%.2f", SampleProfileGlobalHotThreshold.getValue()) << "%)";
S.flush();
emitOptimizationRemark(F.getContext(), DEBUG_TYPE, F, DebugLoc(), Msg);
return true;
}
// If the function collected fewer samples than the cold threshold, mark
// it globally cold.
if (SamplesPercent <= SampleProfileGlobalColdThreshold) {
F.addFnAttr(llvm::Attribute::Cold);
std::string Msg;
raw_string_ostream S(Msg);
S << "Applied cold hint to globally cold function '" << F.getName()
<< "' with " << format("%.2f", SamplesPercent)
<< "% of samples (threshold: "
<< format("%.2f", SampleProfileGlobalColdThreshold.getValue()) << "%)";
S.flush();
emitOptimizationRemark(F.getContext(), DEBUG_TYPE, F, DebugLoc(), Msg);
return true;
}
return false;
}
// This checks whether the transformation is legal.
// Also returns false in cases where it's potentially legal, but
// we don't even want to try.
bool X86CallFrameOptimization::isLegal(MachineFunction &MF) {
if (NoX86CFOpt.getValue())
return false;
// We currently only support call sequences where *all* parameters.
// are passed on the stack.
// No point in running this in 64-bit mode, since some arguments are
// passed in-register in all common calling conventions, so the pattern
// we're looking for will never match.
if (STI->is64Bit())
return false;
// We can't encode multiple DW_CFA_GNU_args_size or DW_CFA_def_cfa_offset
// in the compact unwind encoding that Darwin uses. So, bail if there
// is a danger of that being generated.
if (STI->isTargetDarwin() &&
(!MF.getMMI().getLandingPads().empty() ||
(MF.getFunction()->needsUnwindTableEntry() && !TFL->hasFP(MF))))
return false;
// You would expect straight-line code between call-frame setup and
// call-frame destroy. You would be wrong. There are circumstances (e.g.
// CMOV_GR8 expansion of a select that feeds a function call!) where we can
// end up with the setup and the destroy in different basic blocks.
// This is bad, and breaks SP adjustment.
// So, check that all of the frames in the function are closed inside
// the same block, and, for good measure, that there are no nested frames.
unsigned FrameSetupOpcode = TII->getCallFrameSetupOpcode();
unsigned FrameDestroyOpcode = TII->getCallFrameDestroyOpcode();
for (MachineBasicBlock &BB : MF) {
bool InsideFrameSequence = false;
for (MachineInstr &MI : BB) {
if (MI.getOpcode() == FrameSetupOpcode) {
if (InsideFrameSequence)
return false;
InsideFrameSequence = true;
} else if (MI.getOpcode() == FrameDestroyOpcode) {
if (!InsideFrameSequence)
return false;
InsideFrameSequence = false;
}
}
if (InsideFrameSequence)
return false;
}
return true;
}
// This checks whether the transformation is legal.
// Also returns false in cases where it's potentially legal, but
// we don't even want to try.
bool X86CallFrameOptimization::isLegal(MachineFunction &MF) {
if (NoX86CFOpt.getValue())
return false;
// We can't encode multiple DW_CFA_GNU_args_size or DW_CFA_def_cfa_offset
// in the compact unwind encoding that Darwin uses. So, bail if there
// is a danger of that being generated.
if (STI->isTargetDarwin() &&
(!MF.getLandingPads().empty() ||
(MF.getFunction().needsUnwindTableEntry() && !TFL->hasFP(MF))))
return false;
// It is not valid to change the stack pointer outside the prolog/epilog
// on 64-bit Windows.
if (STI->isTargetWin64())
return false;
// You would expect straight-line code between call-frame setup and
// call-frame destroy. You would be wrong. There are circumstances (e.g.
// CMOV_GR8 expansion of a select that feeds a function call!) where we can
// end up with the setup and the destroy in different basic blocks.
// This is bad, and breaks SP adjustment.
// So, check that all of the frames in the function are closed inside
// the same block, and, for good measure, that there are no nested frames.
unsigned FrameSetupOpcode = TII->getCallFrameSetupOpcode();
unsigned FrameDestroyOpcode = TII->getCallFrameDestroyOpcode();
for (MachineBasicBlock &BB : MF) {
bool InsideFrameSequence = false;
for (MachineInstr &MI : BB) {
if (MI.getOpcode() == FrameSetupOpcode) {
if (InsideFrameSequence)
return false;
InsideFrameSequence = true;
} else if (MI.getOpcode() == FrameDestroyOpcode) {
if (!InsideFrameSequence)
return false;
InsideFrameSequence = false;
}
}
if (InsideFrameSequence)
return false;
}
return true;
}
// This checks whether the transformation is legal.
// Also returns false in cases where it's potentially legal, but
// we don't even want to try.
bool X86CallFrameOptimization::isLegal(MachineFunction &MF) {
if (NoX86CFOpt.getValue())
return false;
// We currently only support call sequences where *all* parameters.
// are passed on the stack.
// No point in running this in 64-bit mode, since some arguments are
// passed in-register in all common calling conventions, so the pattern
// we're looking for will never match.
const X86Subtarget &STI = MF.getSubtarget<X86Subtarget>();
if (STI.is64Bit())
return false;
// You would expect straight-line code between call-frame setup and
// call-frame destroy. You would be wrong. There are circumstances (e.g.
// CMOV_GR8 expansion of a select that feeds a function call!) where we can
// end up with the setup and the destroy in different basic blocks.
// This is bad, and breaks SP adjustment.
// So, check that all of the frames in the function are closed inside
// the same block, and, for good measure, that there are no nested frames.
unsigned FrameSetupOpcode = TII->getCallFrameSetupOpcode();
unsigned FrameDestroyOpcode = TII->getCallFrameDestroyOpcode();
for (MachineBasicBlock &BB : MF) {
bool InsideFrameSequence = false;
for (MachineInstr &MI : BB) {
if (MI.getOpcode() == FrameSetupOpcode) {
if (InsideFrameSequence)
return false;
InsideFrameSequence = true;
} else if (MI.getOpcode() == FrameDestroyOpcode) {
if (!InsideFrameSequence)
return false;
InsideFrameSequence = false;
}
}
if (InsideFrameSequence)
return false;
}
return true;
}
int main(int argc, const char *argv[]) {
cl::ParseCommandLineOptions(argc, argv, "SPIR verifier");
if (InputFilename.empty()) {
errs() << HelpMessage;
return 1;
}
StringRef Path = InputFilename;
LLVMContext Ctx;
OwningPtr<MemoryBuffer> result;
// Parse the bitcode file into a module.
error_code ErrCode = MemoryBuffer::getFile(Path, result);
if (!result.get()) {
errs() << "Buffer Creation Error. " << ErrCode.message() << "\n";
return 1;
}
std::string ErrMsg;
Module *M = ParseBitcodeFile(result.get(), Ctx, &ErrMsg);
if (!M) {
outs() << "According to this SPIR Verifier, " << Path << " is an invalid SPIR module.\n";
errs() << "Bitcode parsing error. " << ErrMsg << "\n";
return 1;
}
// Run the verification pass, and report errors if necessary.
SpirValidation Validation;
Validation.runOnModule(*M);
const ErrorPrinter *EP = Validation.getErrorPrinter();
if (EP->hasErrors()) {
outs() << "According to this SPIR Verifier, " << Path << " is an invalid SPIR module.\n";
errs() << "The module contains the following errors:\n\n";
EP->print(errs(), LITMode.getValue());
return 1;
}
outs() << "According to this SPIR Verifier, " << Path << " is a valid SPIR module.\n";
return 0;
}
// Out of line constructor provides default values for pass options and
// registers all common codegen passes.
TargetPassConfig::TargetPassConfig(TargetMachine *tm, PassManagerBase &pm)
: ImmutablePass(ID), PM(&pm), Started(true), Stopped(false),
AddingMachinePasses(false), TM(tm), Impl(nullptr), Initialized(false),
DisableVerify(false), EnableTailMerge(true) {
Impl = new PassConfigImpl();
// Register all target independent codegen passes to activate their PassIDs,
// including this pass itself.
initializeCodeGen(*PassRegistry::getPassRegistry());
// Also register alias analysis passes required by codegen passes.
initializeBasicAAWrapperPassPass(*PassRegistry::getPassRegistry());
initializeAAResultsWrapperPassPass(*PassRegistry::getPassRegistry());
// Substitute Pseudo Pass IDs for real ones.
substitutePass(&EarlyTailDuplicateID, &TailDuplicateID);
substitutePass(&PostRAMachineLICMID, &MachineLICMID);
if (StringRef(PrintMachineInstrs.getValue()).equals(""))
TM->Options.PrintMachineCode = true;
}
void RegionInfo::print(raw_ostream &OS, const Module *) const {
OS << "Region tree:\n";
TopLevelRegion->print(OS, true, 0, printStyle.getValue());
OS << "End region tree\n";
}
void Region::dump() const {
print(dbgs(), true, getDepth(), printStyle.getValue());
}
/// Add the complete set of target-independent postISel code generator passes.
///
/// This can be read as the standard order of major LLVM CodeGen stages. Stages
/// with nontrivial configuration or multiple passes are broken out below in
/// add%Stage routines.
///
/// Any TargetPassConfig::addXX routine may be overriden by the Target. The
/// addPre/Post methods with empty header implementations allow injecting
/// target-specific fixups just before or after major stages. Additionally,
/// targets have the flexibility to change pass order within a stage by
/// overriding default implementation of add%Stage routines below. Each
/// technique has maintainability tradeoffs because alternate pass orders are
/// not well supported. addPre/Post works better if the target pass is easily
/// tied to a common pass. But if it has subtle dependencies on multiple passes,
/// the target should override the stage instead.
///
/// TODO: We could use a single addPre/Post(ID) hook to allow pass injection
/// before/after any target-independent pass. But it's currently overkill.
void TargetPassConfig::addMachinePasses() {
// Insert a machine instr printer pass after the specified pass.
// If -print-machineinstrs specified, print machineinstrs after all passes.
if (StringRef(PrintMachineInstrs.getValue()).equals(""))
TM->Options.PrintMachineCode = true;
else if (!StringRef(PrintMachineInstrs.getValue())
.equals("option-unspecified")) {
const PassRegistry *PR = PassRegistry::getPassRegistry();
const PassInfo *TPI = PR->getPassInfo(PrintMachineInstrs.getValue());
const PassInfo *IPI = PR->getPassInfo(StringRef("print-machineinstrs"));
assert (TPI && IPI && "Pass ID not registered!");
const char *TID = (const char *)(TPI->getTypeInfo());
const char *IID = (const char *)(IPI->getTypeInfo());
insertPass(TID, IID);
}
// Print the instruction selected machine code...
printAndVerify("After Instruction Selection");
// Expand pseudo-instructions emitted by ISel.
if (addPass(&ExpandISelPseudosID))
printAndVerify("After ExpandISelPseudos");
// Add passes that optimize machine instructions in SSA form.
if (getOptLevel() != CodeGenOpt::None) {
addMachineSSAOptimization();
} else {
// If the target requests it, assign local variables to stack slots relative
// to one another and simplify frame index references where possible.
addPass(&LocalStackSlotAllocationID);
}
// Run pre-ra passes.
if (addPreRegAlloc())
printAndVerify("After PreRegAlloc passes");
// Run register allocation and passes that are tightly coupled with it,
// including phi elimination and scheduling.
if (getOptimizeRegAlloc())
addOptimizedRegAlloc(createRegAllocPass(true));
else
addFastRegAlloc(createRegAllocPass(false));
// Run post-ra passes.
if (addPostRegAlloc())
printAndVerify("After PostRegAlloc passes");
// Insert prolog/epilog code. Eliminate abstract frame index references...
addPass(&PrologEpilogCodeInserterID);
printAndVerify("After PrologEpilogCodeInserter");
/// Add passes that optimize machine instructions after register allocation.
if (getOptLevel() != CodeGenOpt::None)
addMachineLateOptimization();
// Expand pseudo instructions before second scheduling pass.
addPass(&ExpandPostRAPseudosID);
printAndVerify("After ExpandPostRAPseudos");
// Run pre-sched2 passes.
if (addPreSched2())
printAndVerify("After PreSched2 passes");
// Second pass scheduler.
if (getOptLevel() != CodeGenOpt::None) {
if (MISchedPostRA)
addPass(&PostMachineSchedulerID);
else
addPass(&PostRASchedulerID);
printAndVerify("After PostRAScheduler");
}
// GC
if (addGCPasses()) {
if (PrintGCInfo)
addPass(createGCInfoPrinter(dbgs()));
}
// Basic block placement.
if (getOptLevel() != CodeGenOpt::None)
addBlockPlacement();
if (addPreEmitPass())
printAndVerify("After PreEmit passes");
addPass(&StackMapLivenessID);
}
RegisterArguments() :
SpecializeArguments(ArgumentsFileName == "" ? "arguments" : ArgumentsFileName.getValue().c_str(),
SpecializeProgName.getValue() == "" ? NULL : SpecializeProgName.getValue().c_str())
{
}
void LatencyAccountant::exportStats(const XRayFileHeader &Header, F Fn) const {
std::vector<TupleType> Results;
Results.reserve(FunctionLatencies.size());
for (auto FT : FunctionLatencies) {
const auto &FuncId = FT.first;
auto &Timings = FT.second;
Results.emplace_back(FuncId, Timings.size(), getStats(Timings));
auto &Row = std::get<2>(Results.back());
if (Header.CycleFrequency) {
double CycleFrequency = Header.CycleFrequency;
Row.Min /= CycleFrequency;
Row.Median /= CycleFrequency;
Row.Pct90 /= CycleFrequency;
Row.Pct99 /= CycleFrequency;
Row.Max /= CycleFrequency;
Row.Sum /= CycleFrequency;
}
Row.Function = FuncIdHelper.SymbolOrNumber(FuncId);
Row.DebugInfo = FuncIdHelper.FileLineAndColumn(FuncId);
}
// Sort the data according to user-provided flags.
switch (AccountSortOutput) {
case SortField::FUNCID:
sortByKey(Results, [](const TupleType &X) { return std::get<0>(X); });
break;
case SortField::COUNT:
sortByKey(Results, [](const TupleType &X) { return std::get<1>(X); });
break;
case SortField::MIN:
sortByKey(Results, [](const TupleType &X) { return std::get<2>(X).Min; });
break;
case SortField::MED:
sortByKey(Results, [](const TupleType &X) { return std::get<2>(X).Median; });
break;
case SortField::PCT90:
sortByKey(Results, [](const TupleType &X) { return std::get<2>(X).Pct90; });
break;
case SortField::PCT99:
sortByKey(Results, [](const TupleType &X) { return std::get<2>(X).Pct99; });
break;
case SortField::MAX:
sortByKey(Results, [](const TupleType &X) { return std::get<2>(X).Max; });
break;
case SortField::SUM:
sortByKey(Results, [](const TupleType &X) { return std::get<2>(X).Sum; });
break;
case SortField::FUNC:
llvm_unreachable("Not implemented");
}
if (AccountTop > 0) {
auto MaxTop =
std::min(AccountTop.getValue(), static_cast<int>(Results.size()));
Results.erase(Results.begin() + MaxTop, Results.end());
}
for (const auto &R : Results)
Fn(std::get<0>(R), std::get<1>(R), std::get<2>(R));
}
/// Add the complete set of target-independent postISel code generator passes.
///
/// This can be read as the standard order of major LLVM CodeGen stages. Stages
/// with nontrivial configuration or multiple passes are broken out below in
/// add%Stage routines.
///
/// Any TargetPassConfig::addXX routine may be overriden by the Target. The
/// addPre/Post methods with empty header implementations allow injecting
/// target-specific fixups just before or after major stages. Additionally,
/// targets have the flexibility to change pass order within a stage by
/// overriding default implementation of add%Stage routines below. Each
/// technique has maintainability tradeoffs because alternate pass orders are
/// not well supported. addPre/Post works better if the target pass is easily
/// tied to a common pass. But if it has subtle dependencies on multiple passes,
/// the target should override the stage instead.
///
/// TODO: We could use a single addPre/Post(ID) hook to allow pass injection
/// before/after any target-independent pass. But it's currently overkill.
void TargetPassConfig::addMachinePasses() {
AddingMachinePasses = true;
// Insert a machine instr printer pass after the specified pass.
StringRef PrintMachineInstrsPassName = PrintMachineInstrs.getValue();
if (!PrintMachineInstrsPassName.equals("") &&
!PrintMachineInstrsPassName.equals("option-unspecified")) {
if (const PassInfo *TPI = getPassInfo(PrintMachineInstrsPassName)) {
const PassRegistry *PR = PassRegistry::getPassRegistry();
const PassInfo *IPI = PR->getPassInfo(StringRef("machineinstr-printer"));
assert(IPI && "failed to get \"machineinstr-printer\" PassInfo!");
const char *TID = (const char *)(TPI->getTypeInfo());
const char *IID = (const char *)(IPI->getTypeInfo());
insertPass(TID, IID);
}
}
// Print the instruction selected machine code...
printAndVerify("After Instruction Selection");
// Expand pseudo-instructions emitted by ISel.
addPass(&ExpandISelPseudosID);
// Add passes that optimize machine instructions in SSA form.
if (getOptLevel() != CodeGenOpt::None) {
addMachineSSAOptimization();
} else {
// If the target requests it, assign local variables to stack slots relative
// to one another and simplify frame index references where possible.
addPass(&LocalStackSlotAllocationID, false);
}
if (TM->Options.EnableIPRA)
addPass(createRegUsageInfoPropPass());
// Run pre-ra passes.
addPreRegAlloc();
// Run register allocation and passes that are tightly coupled with it,
// including phi elimination and scheduling.
if (getOptimizeRegAlloc())
addOptimizedRegAlloc(createRegAllocPass(true));
else {
if (RegAlloc != &useDefaultRegisterAllocator &&
RegAlloc != &createFastRegisterAllocator)
report_fatal_error("Must use fast (default) register allocator for unoptimized regalloc.");
addFastRegAlloc(createRegAllocPass(false));
}
// Run post-ra passes.
addPostRegAlloc();
// Insert prolog/epilog code. Eliminate abstract frame index references...
if (getOptLevel() != CodeGenOpt::None) {
addPass(&PostRAMachineSinkingID);
addPass(&ShrinkWrapID);
}
// Prolog/Epilog inserter needs a TargetMachine to instantiate. But only
// do so if it hasn't been disabled, substituted, or overridden.
if (!isPassSubstitutedOrOverridden(&PrologEpilogCodeInserterID))
addPass(createPrologEpilogInserterPass());
/// Add passes that optimize machine instructions after register allocation.
if (getOptLevel() != CodeGenOpt::None)
addMachineLateOptimization();
// Expand pseudo instructions before second scheduling pass.
addPass(&ExpandPostRAPseudosID);
// Run pre-sched2 passes.
addPreSched2();
if (EnableImplicitNullChecks)
addPass(&ImplicitNullChecksID);
// Second pass scheduler.
// Let Target optionally insert this pass by itself at some other
// point.
if (getOptLevel() != CodeGenOpt::None &&
!TM->targetSchedulesPostRAScheduling()) {
if (MISchedPostRA)
addPass(&PostMachineSchedulerID);
else
addPass(&PostRASchedulerID);
}
// GC
if (addGCPasses()) {
if (PrintGCInfo)
addPass(createGCInfoPrinter(dbgs()), false, false);
}
// Basic block placement.
if (getOptLevel() != CodeGenOpt::None)
addBlockPlacement();
addPreEmitPass();
if (TM->Options.EnableIPRA)
// Collect register usage information and produce a register mask of
// clobbered registers, to be used to optimize call sites.
addPass(createRegUsageInfoCollector());
addPass(&FuncletLayoutID, false);
addPass(&StackMapLivenessID, false);
addPass(&LiveDebugValuesID, false);
// Insert before XRay Instrumentation.
addPass(&FEntryInserterID, false);
addPass(&XRayInstrumentationID, false);
addPass(&PatchableFunctionID, false);
if (TM->Options.EnableMachineOutliner && getOptLevel() != CodeGenOpt::None &&
EnableMachineOutliner != NeverOutline) {
bool RunOnAllFunctions = (EnableMachineOutliner == AlwaysOutline);
bool AddOutliner = RunOnAllFunctions ||
TM->Options.SupportsDefaultOutlining;
if (AddOutliner)
addPass(createMachineOutlinerPass(RunOnAllFunctions));
}
// Add passes that directly emit MI after all other MI passes.
addPreEmitPass2();
AddingMachinePasses = false;
}
//Command line decoder control
void startCmdLine(){
LLVMContext &Context = getGlobalContext();
for (unsigned int i =0 ; i < PassList.size(); i++ ){
cout << "Pass added: "<< PassList[i]->getPassName() << endl;
cout << "Argument name :" << PassList[i]->getPassArgument() << endl;
}
clock_t timer = clock();
//Parsing XDF file
std::cout << "Parsing file " << XDFFile.getValue() << "." << endl;
XDFParser xdfParser(Verbose);
Network* network = xdfParser.parseFile(XDFFile, Context);
cout << "Network parsed in : "<< (clock() - timer) * 1000 / CLOCKS_PER_SEC << " ms, start engine" << endl;
//Parsing XCF file if needed
if(XCFFile != "") {
std::cout << "Parsing file " << XCFFile.getValue() << "." << endl;
XCFParser xcfParser(Verbose);
map<string, string>* mapping = xcfParser.parseFile(XCFFile);
network->setMapping(mapping);
}
if (enableTrace){
setTraces(network);
}
//Load network
engine->load(network);
// Optimizing decoder
if (optLevel > 0){
engine->optimize(network, optLevel);
}
// Verify the given decoder if needed
if (Verify){
engine->verify(network, "error.txt");
}
// Set input file
input_file = (char*)VidFile.c_str();
// Print the given decoder if needed
if (OutputDir != ""){
engine->print(network);
}
//Run network
engine->run(network);
cout << "End of Jade" << endl;
cout << "Total time: " << (clock() - timer) * 1000 / CLOCKS_PER_SEC << " ms" << endl;
if(XCFFile != "") {
cout << "Note: This execution time is calculated from CPU clock. When more than 1 thread were run, "
"the value displayed is higher than the real execution time." << endl;
}
}
static int compileModule(char **argv, LLVMContext &Context) {
// Load the module to be compiled...
SMDiagnostic Err;
std::unique_ptr<Module> M;
std::unique_ptr<MIRParser> MIR;
Triple TheTriple;
bool SkipModule = MCPU == "help" ||
(!MAttrs.empty() && MAttrs.front() == "help");
// If user just wants to list available options, skip module loading
if ((!SkipModule) && (!PrintInstructions.getValue())) {
if (StringRef(InputFilename).endswith_lower(".mir")) {
MIR = createMIRParserFromFile(InputFilename, Err, Context);
if (MIR) {
M = MIR->parseLLVMModule();
assert(M && "parseLLVMModule should exit on failure");
}
} else
M = parseIRFile(InputFilename, Err, Context);
if (!M) {
Err.print(argv[0], errs());
return 1;
}
// Verify module immediately to catch problems before doInitialization() is
// called on any passes.
if (!NoVerify && verifyModule(*M, &errs())) {
errs() << argv[0] << ": " << InputFilename
<< ": error: input module is broken!\n";
return 1;
}
// If we are supposed to override the target triple, do so now.
if (!TargetTriple.empty())
M->setTargetTriple(Triple::normalize(TargetTriple));
TheTriple = Triple(M->getTargetTriple());
} else {
TheTriple = Triple(Triple::normalize(TargetTriple));
}
if (TheTriple.getTriple().empty())
TheTriple.setTriple(sys::getDefaultTargetTriple());
// Get the target specific parser.
std::string Error;
const Target *TheTarget = TargetRegistry::lookupTarget(MArch, TheTriple,
Error);
if (!TheTarget) {
errs() << argv[0] << ": " << Error;
return 1;
}
std::string CPUStr = getCPUStr(), FeaturesStr = getFeaturesStr();
CodeGenOpt::Level OLvl = CodeGenOpt::Default;
switch (OptLevel) {
default:
errs() << argv[0] << ": invalid optimization level.\n";
return 1;
case ' ': break;
case '0': OLvl = CodeGenOpt::None; break;
case '1': OLvl = CodeGenOpt::Less; break;
case '2': OLvl = CodeGenOpt::Default; break;
case '3': OLvl = CodeGenOpt::Aggressive; break;
}
TargetOptions Options = InitTargetOptionsFromCodeGenFlags();
Options.DisableIntegratedAS = NoIntegratedAssembler;
Options.MCOptions.ShowMCEncoding = ShowMCEncoding;
Options.MCOptions.MCUseDwarfDirectory = EnableDwarfDirectory;
Options.MCOptions.AsmVerbose = AsmVerbose;
std::unique_ptr<TargetMachine> Target(
TheTarget->createTargetMachine(TheTriple.getTriple(), CPUStr, FeaturesStr,
Options, RelocModel, CMModel, OLvl));
assert(Target && "Could not allocate target machine!");
// If we don't have a module then just exit now. We do this down
// here since the CPU/Feature help is underneath the target machine
// creation.
if (SkipModule)
return 0;
// Figure out where we are going to send the output.
std::unique_ptr<tool_output_file> Out =
GetOutputStream(TheTarget->getName(), TheTriple.getOS(), argv[0]);
if (!Out) return 1;
{
raw_pwrite_stream *OS = &Out->os();
std::unique_ptr<buffer_ostream> BOS;
if (FileType != TargetMachine::CGFT_AssemblyFile &&
!Out->os().supportsSeeking()) {
BOS = make_unique<buffer_ostream>(*OS);
OS = BOS.get();
}
if (PrintInstructions.getValue()) {
printInstructions(TheTarget, OS);
Out->keep();
return 0;
}
assert(M && "Should have exited if we didn't have a module!");
if (FloatABIForCalls != FloatABI::Default)
Options.FloatABIType = FloatABIForCalls;
// Build up all of the passes that we want to do to the module.
legacy::PassManager PM;
// Add an appropriate TargetLibraryInfo pass for the module's triple.
TargetLibraryInfoImpl TLII(Triple(M->getTargetTriple()));
// The -disable-simplify-libcalls flag actually disables all builtin optzns.
if (DisableSimplifyLibCalls)
TLII.disableAllFunctions();
PM.add(new TargetLibraryInfoWrapperPass(TLII));
// Add the target data from the target machine, if it exists, or the module.
M->setDataLayout(Target->createDataLayout());
// Override function attributes based on CPUStr, FeaturesStr, and command line
// flags.
setFunctionAttributes(CPUStr, FeaturesStr, *M);
if (RelaxAll.getNumOccurrences() > 0 &&
FileType != TargetMachine::CGFT_ObjectFile)
errs() << argv[0]
<< ": warning: ignoring -mc-relax-all because filetype != obj";
AnalysisID StartBeforeID = nullptr;
AnalysisID StartAfterID = nullptr;
AnalysisID StopAfterID = nullptr;
const PassRegistry *PR = PassRegistry::getPassRegistry();
if (!RunPass.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";
return 1;
}
StopAfterID = StartBeforeID = PI->getTypeInfo();
} else {
if (!StartAfter.empty()) {
const PassInfo *PI = PR->getPassInfo(StartAfter);
if (!PI) {
errs() << argv[0] << ": start-after pass is not registered.\n";
return 1;
}
StartAfterID = PI->getTypeInfo();
}
if (!StopAfter.empty()) {
const PassInfo *PI = PR->getPassInfo(StopAfter);
if (!PI) {
errs() << argv[0] << ": stop-after pass is not registered.\n";
return 1;
}
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;
}
// Before executing passes, print the final values of the LLVM options.
cl::PrintOptionValues();
PM.run(*M);
}
// Declare success.
Out->keep();
return 0;
}
static void processOptionImpl(cl::opt<bool> &O, const cl::opt<bool> &Default) {
if (!O.getNumOccurrences() || O.getPosition() < Default.getPosition())
O = Default.getValue();
}
// This checks whether the transformation is legal and profitable
bool X86CallFrameOptimization::shouldPerformTransformation(MachineFunction &MF) {
if (NoX86CFOpt.getValue())
return false;
// We currently only support call sequences where *all* parameters.
// are passed on the stack.
// No point in running this in 64-bit mode, since some arguments are
// passed in-register in all common calling conventions, so the pattern
// we're looking for will never match.
const X86Subtarget &STI = MF.getSubtarget<X86Subtarget>();
if (STI.is64Bit())
return false;
// You would expect straight-line code between call-frame setup and
// call-frame destroy. You would be wrong. There are circumstances (e.g.
// CMOV_GR8 expansion of a select that feeds a function call!) where we can
// end up with the setup and the destroy in different basic blocks.
// This is bad, and breaks SP adjustment.
// So, check that all of the frames in the function are closed inside
// the same block, and, for good measure, that there are no nested frames.
int FrameSetupOpcode = TII->getCallFrameSetupOpcode();
int FrameDestroyOpcode = TII->getCallFrameDestroyOpcode();
for (MachineBasicBlock &BB : MF) {
bool InsideFrameSequence = false;
for (MachineInstr &MI : BB) {
if (MI.getOpcode() == FrameSetupOpcode) {
if (InsideFrameSequence)
return false;
InsideFrameSequence = true;
}
else if (MI.getOpcode() == FrameDestroyOpcode) {
if (!InsideFrameSequence)
return false;
InsideFrameSequence = false;
}
}
if (InsideFrameSequence)
return false;
}
// Now that we know the transformation is legal, check if it is
// profitable.
// TODO: Add a heuristic that actually looks at the function,
// and enable this for more cases.
// This transformation is always a win when we expected to have
// a reserved call frame. Under other circumstances, it may be either
// a win or a loss, and requires a heuristic.
// For now, enable it only for the relatively clear win cases.
bool CannotReserveFrame = MF.getFrameInfo()->hasVarSizedObjects();
if (CannotReserveFrame)
return true;
// Don't do this when not optimizing for size.
bool OptForSize =
MF.getFunction()->hasFnAttribute(Attribute::OptimizeForSize) ||
MF.getFunction()->hasFnAttribute(Attribute::MinSize);
if (!OptForSize)
return false;
// Stack re-alignment can make this unprofitable even in terms of size.
// As mentioned above, a better heuristic is needed. For now, don't do this
// when the required alignment is above 8. (4 would be the safe choice, but
// some experimentation showed 8 is generally good).
if (TFL->getStackAlignment() > 8)
return false;
return true;
}
int main(int argc, char **argv) {
llvm::sys::PrintStackTraceOnErrorSignal();
llvm::PrettyStackTraceProgram X(argc, argv);
llvm_shutdown_obj Y;
cl::ParseCommandLineOptions(argc, argv, "linker script randomizer\n");
sys::Path ldPath = sys::Program::FindProgramByName("ld");
if (ldPath.isEmpty()) {
errs() << "Couldn't find system linker";
llvm_shutdown();
return 1;
}
sys::Path scriptPath = sys::Path::GetTemporaryDirectory();
if (scriptPath.isEmpty()) {
errs() << "Error accessing temporary directory";
llvm_shutdown();
return 1;
}
scriptPath.appendComponent("ld_script");
scriptPath.makeUnique(false, NULL);
sys::Path devNull;
std::string errMsg;
const char *ldArgs[] = { ldPath.c_str(), "--verbose", NULL };
const sys::Path *redirects[] = { &devNull, &scriptPath, &scriptPath, NULL };
if (sys::Program::ExecuteAndWait(ldPath, ldArgs, 0, redirects, 0, 0, &errMsg)) {
errs() << "Error executing linker";
llvm_shutdown();
return 1;
}
OwningPtr< MemoryBuffer > scriptFile;
error_code ec = MemoryBuffer::getFile(scriptPath.c_str(), scriptFile);
if (ec) {
errs() << "Error reading script file: " << ec.message();
llvm_shutdown();
return 1;
}
uint32_t minPage = (MinBaseAddress + PAGE_SIZE - 1) / PAGE_SIZE,
maxPage = MaxBaseAddress / PAGE_SIZE;
if (minPage > maxPage) {
errs() << "Base address interval is empty";
llvm_shutdown();
return 1;
}
uint32_t newAddr = PAGE_SIZE * (minPage +
RandomNumberGenerator::Generator().Random(maxPage - minPage + 1));
char oldAddrStr[12];
sprintf(oldAddrStr, "0x%08x", OldBaseAddress.getValue());
llvm::Regex oldAddrRegex(oldAddrStr);
StringRef scriptText = scriptFile->getBuffer();
StringRef oldScriptText = scriptText;
char newAddrStr[12];
sprintf(newAddrStr, "0x%08x", newAddr);
do {
// Replace one occurrence of old address with new one
oldScriptText = scriptText;
scriptText = oldAddrRegex.sub(newAddrStr, scriptText);
} while (!scriptText.equals(oldScriptText));
std::cout << scriptText.str();
return 0;
}
static std::unique_ptr<tool_output_file>
GetOutputStream(const char *TargetName, Triple::OSType OS,
const char *ProgName) {
// If we don't yet have an output filename, make one.
if (PrintInstructions.getValue()) {
if (OutputFilename.empty()) {
OutputFilename = "hotspot.out";
}
} else { if (OutputFilename.empty()) {
if (InputFilename == "-")
OutputFilename = "-";
else {
// If InputFilename ends in .bc or .ll, remove it.
StringRef IFN = InputFilename;
if (IFN.endswith(".bc") || IFN.endswith(".ll"))
OutputFilename = IFN.drop_back(3);
else if (IFN.endswith(".mir"))
OutputFilename = IFN.drop_back(4);
else
OutputFilename = IFN;
switch (FileType) {
case TargetMachine::CGFT_AssemblyFile:
if (TargetName[0] == 'c') {
if (TargetName[1] == 0)
OutputFilename += ".cbe.c";
else if (TargetName[1] == 'p' && TargetName[2] == 'p')
OutputFilename += ".cpp";
else
OutputFilename += ".s";
} else
OutputFilename += ".s";
break;
case TargetMachine::CGFT_ObjectFile:
if (OS == Triple::Win32)
OutputFilename += ".obj";
else
OutputFilename += ".o";
break;
case TargetMachine::CGFT_Null:
OutputFilename += ".null";
break;
}
}
}
}
// Decide if we need "binary" output.
bool Binary = false;
switch (FileType) {
case TargetMachine::CGFT_AssemblyFile:
break;
case TargetMachine::CGFT_ObjectFile:
case TargetMachine::CGFT_Null:
Binary = true;
break;
}
// Open the file.
std::error_code EC;
sys::fs::OpenFlags OpenFlags = sys::fs::F_None;
if (!Binary)
OpenFlags |= sys::fs::F_Text;
auto FDOut = llvm::make_unique<tool_output_file>(OutputFilename, EC,
OpenFlags);
if (EC) {
errs() << EC.message() << '\n';
return nullptr;
}
return FDOut;
}
