llvm-6502/tools/llc/llc.cpp
Chandler Carruth aeef83c6af Switch TargetTransformInfo from an immutable analysis pass that requires
a TargetMachine to construct (and thus isn't always available), to an
analysis group that supports layered implementations much like
AliasAnalysis does. This is a pretty massive change, with a few parts
that I was unable to easily separate (sorry), so I'll walk through it.

The first step of this conversion was to make TargetTransformInfo an
analysis group, and to sink the nonce implementations in
ScalarTargetTransformInfo and VectorTargetTranformInfo into
a NoTargetTransformInfo pass. This allows other passes to add a hard
requirement on TTI, and assume they will always get at least on
implementation.

The TargetTransformInfo analysis group leverages the delegation chaining
trick that AliasAnalysis uses, where the base class for the analysis
group delegates to the previous analysis *pass*, allowing all but tho
NoFoo analysis passes to only implement the parts of the interfaces they
support. It also introduces a new trick where each pass in the group
retains a pointer to the top-most pass that has been initialized. This
allows passes to implement one API in terms of another API and benefit
when some other pass above them in the stack has more precise results
for the second API.

The second step of this conversion is to create a pass that implements
the TargetTransformInfo analysis using the target-independent
abstractions in the code generator. This replaces the
ScalarTargetTransformImpl and VectorTargetTransformImpl classes in
lib/Target with a single pass in lib/CodeGen called
BasicTargetTransformInfo. This class actually provides most of the TTI
functionality, basing it upon the TargetLowering abstraction and other
information in the target independent code generator.

The third step of the conversion adds support to all TargetMachines to
register custom analysis passes. This allows building those passes with
access to TargetLowering or other target-specific classes, and it also
allows each target to customize the set of analysis passes desired in
the pass manager. The baseline LLVMTargetMachine implements this
interface to add the BasicTTI pass to the pass manager, and all of the
tools that want to support target-aware TTI passes call this routine on
whatever target machine they end up with to add the appropriate passes.

The fourth step of the conversion created target-specific TTI analysis
passes for the X86 and ARM backends. These passes contain the custom
logic that was previously in their extensions of the
ScalarTargetTransformInfo and VectorTargetTransformInfo interfaces.
I separated them into their own file, as now all of the interface bits
are private and they just expose a function to create the pass itself.
Then I extended these target machines to set up a custom set of analysis
passes, first adding BasicTTI as a fallback, and then adding their
customized TTI implementations.

The fourth step required logic that was shared between the target
independent layer and the specific targets to move to a different
interface, as they no longer derive from each other. As a consequence,
a helper functions were added to TargetLowering representing the common
logic needed both in the target implementation and the codegen
implementation of the TTI pass. While technically this is the only
change that could have been committed separately, it would have been
a nightmare to extract.

The final step of the conversion was just to delete all the old
boilerplate. This got rid of the ScalarTargetTransformInfo and
VectorTargetTransformInfo classes, all of the support in all of the
targets for producing instances of them, and all of the support in the
tools for manually constructing a pass based around them.

Now that TTI is a relatively normal analysis group, two things become
straightforward. First, we can sink it into lib/Analysis which is a more
natural layer for it to live. Second, clients of this interface can
depend on it *always* being available which will simplify their code and
behavior. These (and other) simplifications will follow in subsequent
commits, this one is clearly big enough.

Finally, I'm very aware that much of the comments and documentation
needs to be updated. As soon as I had this working, and plausibly well
commented, I wanted to get it committed and in front of the build bots.
I'll be doing a few passes over documentation later if it sticks.

Commits to update DragonEgg and Clang will be made presently.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171681 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-07 01:37:14 +00:00

384 lines
12 KiB
C++

//===-- llc.cpp - Implement the LLVM Native Code Generator ----------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This is the llc code generator driver. It provides a convenient
// command-line interface for generating native assembly-language code
// or C code, given LLVM bitcode.
//
//===----------------------------------------------------------------------===//
#include "llvm/IR/LLVMContext.h"
#include "llvm/ADT/Triple.h"
#include "llvm/Assembly/PrintModulePass.h"
#include "llvm/CodeGen/CommandFlags.h"
#include "llvm/CodeGen/LinkAllAsmWriterComponents.h"
#include "llvm/CodeGen/LinkAllCodegenComponents.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/Module.h"
#include "llvm/MC/SubtargetFeature.h"
#include "llvm/Pass.h"
#include "llvm/PassManager.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/FormattedStream.h"
#include "llvm/Support/Host.h"
#include "llvm/Support/IRReader.h"
#include "llvm/Support/ManagedStatic.h"
#include "llvm/Support/PluginLoader.h"
#include "llvm/Support/PrettyStackTrace.h"
#include "llvm/Support/Signals.h"
#include "llvm/Support/TargetRegistry.h"
#include "llvm/Support/TargetSelect.h"
#include "llvm/Support/ToolOutputFile.h"
#include "llvm/Target/TargetLibraryInfo.h"
#include "llvm/Target/TargetMachine.h"
#include <memory>
using namespace llvm;
// General options for llc. Other pass-specific options are specified
// within the corresponding llc passes, and target-specific options
// and back-end code generation options are specified with the target machine.
//
static cl::opt<std::string>
InputFilename(cl::Positional, cl::desc("<input bitcode>"), cl::init("-"));
static cl::opt<std::string>
OutputFilename("o", cl::desc("Output filename"), cl::value_desc("filename"));
static cl::opt<unsigned>
TimeCompilations("time-compilations", cl::Hidden, cl::init(1u),
cl::value_desc("N"),
cl::desc("Repeat compilation N times for timing"));
// Determine optimization level.
static cl::opt<char>
OptLevel("O",
cl::desc("Optimization level. [-O0, -O1, -O2, or -O3] "
"(default = '-O2')"),
cl::Prefix,
cl::ZeroOrMore,
cl::init(' '));
static cl::opt<std::string>
TargetTriple("mtriple", cl::desc("Override target triple for module"));
cl::opt<bool> NoVerify("disable-verify", cl::Hidden,
cl::desc("Do not verify input module"));
cl::opt<bool>
DisableSimplifyLibCalls("disable-simplify-libcalls",
cl::desc("Disable simplify-libcalls"),
cl::init(false));
static int compileModule(char**, LLVMContext&);
// GetFileNameRoot - Helper function to get the basename of a filename.
static inline std::string
GetFileNameRoot(const std::string &InputFilename) {
std::string IFN = InputFilename;
std::string outputFilename;
int Len = IFN.length();
if ((Len > 2) &&
IFN[Len-3] == '.' &&
((IFN[Len-2] == 'b' && IFN[Len-1] == 'c') ||
(IFN[Len-2] == 'l' && IFN[Len-1] == 'l'))) {
outputFilename = std::string(IFN.begin(), IFN.end()-3); // s/.bc/.s/
} else {
outputFilename = IFN;
}
return outputFilename;
}
static 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 (OutputFilename.empty()) {
if (InputFilename == "-")
OutputFilename = "-";
else {
OutputFilename = GetFileNameRoot(InputFilename);
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::string error;
unsigned OpenFlags = 0;
if (Binary) OpenFlags |= raw_fd_ostream::F_Binary;
tool_output_file *FDOut = new tool_output_file(OutputFilename.c_str(), error,
OpenFlags);
if (!error.empty()) {
errs() << error << '\n';
delete FDOut;
return 0;
}
return FDOut;
}
// main - Entry point for the llc compiler.
//
int main(int argc, char **argv) {
sys::PrintStackTraceOnErrorSignal();
PrettyStackTraceProgram X(argc, argv);
// Enable debug stream buffering.
EnableDebugBuffering = true;
LLVMContext &Context = getGlobalContext();
llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
// Initialize targets first, so that --version shows registered targets.
InitializeAllTargets();
InitializeAllTargetMCs();
InitializeAllAsmPrinters();
InitializeAllAsmParsers();
// Initialize codegen and IR passes used by llc so that the -print-after,
// -print-before, and -stop-after options work.
PassRegistry *Registry = PassRegistry::getPassRegistry();
initializeCore(*Registry);
initializeCodeGen(*Registry);
initializeLoopStrengthReducePass(*Registry);
initializeLowerIntrinsicsPass(*Registry);
initializeUnreachableBlockElimPass(*Registry);
// Register the target printer for --version.
cl::AddExtraVersionPrinter(TargetRegistry::printRegisteredTargetsForVersion);
cl::ParseCommandLineOptions(argc, argv, "llvm system compiler\n");
// Compile the module TimeCompilations times to give better compile time
// metrics.
for (unsigned I = TimeCompilations; I; --I)
if (int RetVal = compileModule(argv, Context))
return RetVal;
return 0;
}
static int compileModule(char **argv, LLVMContext &Context) {
// Load the module to be compiled...
SMDiagnostic Err;
std::auto_ptr<Module> M;
Module *mod = 0;
Triple TheTriple;
bool SkipModule = MCPU == "help" ||
(!MAttrs.empty() && MAttrs.front() == "help");
// If user just wants to list available options, skip module loading
if (!SkipModule) {
M.reset(ParseIRFile(InputFilename, Err, Context));
mod = M.get();
if (mod == 0) {
Err.print(argv[0], errs());
return 1;
}
// If we are supposed to override the target triple, do so now.
if (!TargetTriple.empty())
mod->setTargetTriple(Triple::normalize(TargetTriple));
TheTriple = Triple(mod->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;
}
// Package up features to be passed to target/subtarget
std::string FeaturesStr;
if (MAttrs.size()) {
SubtargetFeatures Features;
for (unsigned i = 0; i != MAttrs.size(); ++i)
Features.AddFeature(MAttrs[i]);
FeaturesStr = Features.getString();
}
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;
Options.LessPreciseFPMADOption = EnableFPMAD;
Options.NoFramePointerElim = DisableFPElim;
Options.NoFramePointerElimNonLeaf = DisableFPElimNonLeaf;
Options.AllowFPOpFusion = FuseFPOps;
Options.UnsafeFPMath = EnableUnsafeFPMath;
Options.NoInfsFPMath = EnableNoInfsFPMath;
Options.NoNaNsFPMath = EnableNoNaNsFPMath;
Options.HonorSignDependentRoundingFPMathOption =
EnableHonorSignDependentRoundingFPMath;
Options.UseSoftFloat = GenerateSoftFloatCalls;
if (FloatABIForCalls != FloatABI::Default)
Options.FloatABIType = FloatABIForCalls;
Options.NoZerosInBSS = DontPlaceZerosInBSS;
Options.GuaranteedTailCallOpt = EnableGuaranteedTailCallOpt;
Options.DisableTailCalls = DisableTailCalls;
Options.StackAlignmentOverride = OverrideStackAlignment;
Options.RealignStack = EnableRealignStack;
Options.TrapFuncName = TrapFuncName;
Options.PositionIndependentExecutable = EnablePIE;
Options.EnableSegmentedStacks = SegmentedStacks;
Options.UseInitArray = UseInitArray;
Options.SSPBufferSize = SSPBufferSize;
std::auto_ptr<TargetMachine>
target(TheTarget->createTargetMachine(TheTriple.getTriple(),
MCPU, FeaturesStr, Options,
RelocModel, CMModel, OLvl));
assert(target.get() && "Could not allocate target machine!");
assert(mod && "Should have exited after outputting help!");
TargetMachine &Target = *target.get();
if (DisableDotLoc)
Target.setMCUseLoc(false);
if (DisableCFI)
Target.setMCUseCFI(false);
if (EnableDwarfDirectory)
Target.setMCUseDwarfDirectory(true);
if (GenerateSoftFloatCalls)
FloatABIForCalls = FloatABI::Soft;
// Disable .loc support for older OS X versions.
if (TheTriple.isMacOSX() &&
TheTriple.isMacOSXVersionLT(10, 6))
Target.setMCUseLoc(false);
// Figure out where we are going to send the output.
OwningPtr<tool_output_file> Out
(GetOutputStream(TheTarget->getName(), TheTriple.getOS(), argv[0]));
if (!Out) return 1;
// Build up all of the passes that we want to do to the module.
PassManager PM;
// Add an appropriate TargetLibraryInfo pass for the module's triple.
TargetLibraryInfo *TLI = new TargetLibraryInfo(TheTriple);
if (DisableSimplifyLibCalls)
TLI->disableAllFunctions();
PM.add(TLI);
// Add intenal analysis passes from the target machine.
Target.addAnalysisPasses(PM);
// Add the target data from the target machine, if it exists, or the module.
if (const DataLayout *TD = Target.getDataLayout())
PM.add(new DataLayout(*TD));
else
PM.add(new DataLayout(mod));
// Override default to generate verbose assembly.
Target.setAsmVerbosityDefault(true);
if (RelaxAll) {
if (FileType != TargetMachine::CGFT_ObjectFile)
errs() << argv[0]
<< ": warning: ignoring -mc-relax-all because filetype != obj";
else
Target.setMCRelaxAll(true);
}
{
formatted_raw_ostream FOS(Out->os());
AnalysisID StartAfterID = 0;
AnalysisID StopAfterID = 0;
const PassRegistry *PR = PassRegistry::getPassRegistry();
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, FOS, FileType, NoVerify,
StartAfterID, StopAfterID)) {
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(*mod);
}
// Declare success.
Out->keep();
return 0;
}