llvm-project/llvm/lib/CodeGen/GlobalISel/Legalizer.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

337 lines
12 KiB
C++
Raw Normal View History

//===-- llvm/CodeGen/GlobalISel/Legalizer.cpp -----------------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
/// \file This file implements the LegalizerHelper class to legalize individual
/// instructions and the LegalizePass wrapper pass for the primary
/// legalization.
//
//===----------------------------------------------------------------------===//
#include "llvm/CodeGen/GlobalISel/Legalizer.h"
#include "llvm/ADT/PostOrderIterator.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/CodeGen/GlobalISel/CSEInfo.h"
#include "llvm/CodeGen/GlobalISel/CSEMIRBuilder.h"
#include "llvm/CodeGen/GlobalISel/GISelChangeObserver.h"
#include "llvm/CodeGen/GlobalISel/GISelWorkList.h"
#include "llvm/CodeGen/GlobalISel/LegalizationArtifactCombiner.h"
#include "llvm/CodeGen/GlobalISel/LegalizerHelper.h"
#include "llvm/CodeGen/GlobalISel/Utils.h"
#include "llvm/CodeGen/MachineOptimizationRemarkEmitter.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/TargetPassConfig.h"
#include "llvm/CodeGen/TargetSubtargetInfo.h"
#include "llvm/InitializePasses.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/Error.h"
[GlobalISel] Enable CSE in the IRTranslator & legalizer for -O0 with constants only. Other opcodes shouldn't be CSE'd until we can be sure debug info quality won't be degraded. This change also improves the IRTranslator so that in most places, but not all, it creates constants using the MIRBuilder directly instead of first creating a new destination vreg and then creating a constant. By doing this, the buildConstant() method can just return the vreg of an existing G_CONSTANT instead of having to create a COPY from it. I measured a 0.2% improvement in compile time and a 0.9% improvement in code size at -O0 ARM64. Compile time: Program base cse diff test-suite...ark/tramp3d-v4/tramp3d-v4.test 9.04 9.12 0.8% test-suite...Mark/mafft/pairlocalalign.test 2.68 2.66 -0.7% test-suite...-typeset/consumer-typeset.test 5.53 5.51 -0.4% test-suite :: CTMark/lencod/lencod.test 5.30 5.28 -0.3% test-suite :: CTMark/Bullet/bullet.test 25.82 25.76 -0.2% test-suite...:: CTMark/ClamAV/clamscan.test 6.92 6.90 -0.2% test-suite...TMark/7zip/7zip-benchmark.test 34.24 34.17 -0.2% test-suite :: CTMark/SPASS/SPASS.test 6.25 6.24 -0.1% test-suite...:: CTMark/sqlite3/sqlite3.test 1.66 1.66 -0.1% test-suite :: CTMark/kimwitu++/kc.test 13.61 13.60 -0.0% Geomean difference -0.2% Code size: Program base cse diff test-suite...-typeset/consumer-typeset.test 1315632 1266480 -3.7% test-suite...:: CTMark/ClamAV/clamscan.test 1313892 1297508 -1.2% test-suite :: CTMark/lencod/lencod.test 1439504 1423112 -1.1% test-suite...TMark/7zip/7zip-benchmark.test 2936980 2904172 -1.1% test-suite :: CTMark/Bullet/bullet.test 3478276 3445460 -0.9% test-suite...ark/tramp3d-v4/tramp3d-v4.test 8082868 8033492 -0.6% test-suite :: CTMark/kimwitu++/kc.test 3870380 3853972 -0.4% test-suite :: CTMark/SPASS/SPASS.test 1434904 1434896 -0.0% test-suite...Mark/mafft/pairlocalalign.test 764528 764528 0.0% test-suite...:: CTMark/sqlite3/sqlite3.test 782092 782092 0.0% Geomean difference -0.9% Differential Revision: https://reviews.llvm.org/D60580 llvm-svn: 358369
2019-04-15 13:04:20 +08:00
#include "llvm/Target/TargetMachine.h"
#include <iterator>
#define DEBUG_TYPE "legalizer"
using namespace llvm;
static cl::opt<bool>
EnableCSEInLegalizer("enable-cse-in-legalizer",
cl::desc("Should enable CSE in Legalizer"),
cl::Optional, cl::init(false));
char Legalizer::ID = 0;
INITIALIZE_PASS_BEGIN(Legalizer, DEBUG_TYPE,
"Legalize the Machine IR a function's Machine IR", false,
false)
INITIALIZE_PASS_DEPENDENCY(TargetPassConfig)
INITIALIZE_PASS_DEPENDENCY(GISelCSEAnalysisWrapperPass)
INITIALIZE_PASS_END(Legalizer, DEBUG_TYPE,
"Legalize the Machine IR a function's Machine IR", false,
false)
Legalizer::Legalizer() : MachineFunctionPass(ID) { }
void Legalizer::getAnalysisUsage(AnalysisUsage &AU) const {
AU.addRequired<TargetPassConfig>();
AU.addRequired<GISelCSEAnalysisWrapperPass>();
AU.addPreserved<GISelCSEAnalysisWrapperPass>();
getSelectionDAGFallbackAnalysisUsage(AU);
MachineFunctionPass::getAnalysisUsage(AU);
}
void Legalizer::init(MachineFunction &MF) {
}
static bool isArtifact(const MachineInstr &MI) {
switch (MI.getOpcode()) {
default:
return false;
case TargetOpcode::G_TRUNC:
case TargetOpcode::G_ZEXT:
case TargetOpcode::G_ANYEXT:
case TargetOpcode::G_SEXT:
case TargetOpcode::G_MERGE_VALUES:
case TargetOpcode::G_UNMERGE_VALUES:
case TargetOpcode::G_CONCAT_VECTORS:
case TargetOpcode::G_BUILD_VECTOR:
case TargetOpcode::G_EXTRACT:
return true;
}
}
using InstListTy = GISelWorkList<256>;
using ArtifactListTy = GISelWorkList<128>;
namespace {
class LegalizerWorkListManager : public GISelChangeObserver {
InstListTy &InstList;
ArtifactListTy &ArtifactList;
#ifndef NDEBUG
SmallVector<MachineInstr *, 4> NewMIs;
#endif
public:
LegalizerWorkListManager(InstListTy &Insts, ArtifactListTy &Arts)
: InstList(Insts), ArtifactList(Arts) {}
void createdOrChangedInstr(MachineInstr &MI) {
// Only legalize pre-isel generic instructions.
// Legalization process could generate Target specific pseudo
// instructions with generic types. Don't record them
if (isPreISelGenericOpcode(MI.getOpcode())) {
if (isArtifact(MI))
ArtifactList.insert(&MI);
else
InstList.insert(&MI);
}
}
void createdInstr(MachineInstr &MI) override {
LLVM_DEBUG(dbgs() << ".. .. New MI: " << MI);
LLVM_DEBUG(NewMIs.push_back(&MI));
createdOrChangedInstr(MI);
}
void printNewInstrs() {
LLVM_DEBUG({
for (const auto *MI : NewMIs)
dbgs() << ".. .. New MI: " << *MI;
NewMIs.clear();
});
}
void erasingInstr(MachineInstr &MI) override {
LLVM_DEBUG(dbgs() << ".. .. Erasing: " << MI);
InstList.remove(&MI);
ArtifactList.remove(&MI);
}
void changingInstr(MachineInstr &MI) override {
LLVM_DEBUG(dbgs() << ".. .. Changing MI: " << MI);
}
void changedInstr(MachineInstr &MI) override {
// When insts change, we want to revisit them to legalize them again.
// We'll consider them the same as created.
LLVM_DEBUG(dbgs() << ".. .. Changed MI: " << MI);
createdOrChangedInstr(MI);
}
};
} // namespace
Legalizer::MFResult
Legalizer::legalizeMachineFunction(MachineFunction &MF, const LegalizerInfo &LI,
ArrayRef<GISelChangeObserver *> AuxObservers,
MachineIRBuilder &MIRBuilder) {
MachineRegisterInfo &MRI = MF.getRegInfo();
// Populate worklists.
InstListTy InstList;
ArtifactListTy ArtifactList;
ReversePostOrderTraversal<MachineFunction *> RPOT(&MF);
// Perform legalization bottom up so we can DCE as we legalize.
// Traverse BB in RPOT and within each basic block, add insts top down,
// so when we pop_back_val in the legalization process, we traverse bottom-up.
for (auto *MBB : RPOT) {
if (MBB->empty())
continue;
for (MachineInstr &MI : *MBB) {
// Only legalize pre-isel generic instructions: others don't have types
// and are assumed to be legal.
if (!isPreISelGenericOpcode(MI.getOpcode()))
continue;
if (isArtifact(MI))
ArtifactList.deferred_insert(&MI);
else
InstList.deferred_insert(&MI);
}
}
ArtifactList.finalize();
InstList.finalize();
// This observer keeps the worklists updated.
LegalizerWorkListManager WorkListObserver(InstList, ArtifactList);
// We want both WorkListObserver as well as all the auxiliary observers (e.g.
// CSEInfo) to observe all changes. Use the wrapper observer.
GISelObserverWrapper WrapperObserver(&WorkListObserver);
for (GISelChangeObserver *Observer : AuxObservers)
WrapperObserver.addObserver(Observer);
// Now install the observer as the delegate to MF.
// This will keep all the observers notified about new insertions/deletions.
RAIIMFObsDelInstaller Installer(MF, WrapperObserver);
LegalizerHelper Helper(MF, LI, WrapperObserver, MIRBuilder);
LegalizationArtifactCombiner ArtCombiner(MIRBuilder, MRI, LI);
auto RemoveDeadInstFromLists = [&WrapperObserver](MachineInstr *DeadMI) {
WrapperObserver.erasingInstr(*DeadMI);
};
bool Changed = false;
SmallVector<MachineInstr *, 128> RetryList;
do {
LLVM_DEBUG(dbgs() << "=== New Iteration ===\n");
assert(RetryList.empty() && "Expected no instructions in RetryList");
unsigned NumArtifacts = ArtifactList.size();
while (!InstList.empty()) {
MachineInstr &MI = *InstList.pop_back_val();
assert(isPreISelGenericOpcode(MI.getOpcode()) &&
"Expecting generic opcode");
if (isTriviallyDead(MI, MRI)) {
LLVM_DEBUG(dbgs() << MI << "Is dead; erasing.\n");
MI.eraseFromParentAndMarkDBGValuesForRemoval();
continue;
}
// Do the legalization for this instruction.
auto Res = Helper.legalizeInstrStep(MI);
// Error out if we couldn't legalize this instruction. We may want to
// fall back to DAG ISel instead in the future.
if (Res == LegalizerHelper::UnableToLegalize) {
// Move illegal artifacts to RetryList instead of aborting because
// legalizing InstList may generate artifacts that allow
// ArtifactCombiner to combine away them.
if (isArtifact(MI)) {
LLVM_DEBUG(dbgs() << ".. Not legalized, moving to artifacts retry\n");
[Legalizer] Making artifact combining order-independent Legalization algorithm is complicated by two facts: 1) While regular instructions should be possible to legalize in an isolated, per-instruction, context-free manner, legalization artifacts can only be eliminated in pairs, which could be deeply, and ultimately arbitrary nested: { [ () ] }, where which paranthesis kind depicts an artifact kind, like extend, unmerge, etc. Such structure can only be fully eliminated by simple local combines if they are attempted in a particular order (inside out), or alternatively by repeated scans each eliminating only one innermost pair, resulting in O(n^2) complexity. 2) Some artifacts might in fact be regular instructions that could (and sometimes should) be legalized by the target-specific rules. Which means failure to eliminate all artifacts on the first iteration is not a failure, they need to be tried as instructions, which may produce more artifacts, including the ones that are in fact regular instructions, resulting in a non-constant number of iterations required to finish the process. I trust the recently introduced termination condition (no new artifacts were created during as-a-regular-instruction-retrial of artifacts not eliminated on the previous iteration) to be efficient in providing termination, but only performing the legalization in full if and only if at each step such chains of artifacts are successfully eliminated in full as well. Which is currently not guaranteed, as the artifact combines are applied only once and in an arbitrary order that has to do with the order of creation or insertion of artifacts into their worklist, which is a no particular order. In this patch I make a small change to the artifact combiner, making it to re-insert into the worklist immediate (modulo a look-through copies) artifact users of each vreg that changes its definition due to an artifact combine. Here the first scan through the artifacts worklist, while not being done in any guaranteed order, only needs to find the innermost pair(s) of artifacts that could be immediately combined out. After that the process follows def-use chains, making them shorter at each step, thus combining everything that can be combined in O(n) time. Reviewers: volkan, aditya_nandakumar, qcolombet, paquette, aemerson, dsanders Reviewed By: aditya_nandakumar, paquette Tags: #llvm Differential Revision: https://reviews.llvm.org/D71448
2019-12-13 05:20:41 +08:00
assert(NumArtifacts == 0 &&
"Artifacts are only expected in instruction list starting the "
"second iteration, but each iteration starting second must "
"start with an empty artifacts list");
(void)NumArtifacts;
RetryList.push_back(&MI);
continue;
}
Helper.MIRBuilder.stopObservingChanges();
return {Changed, &MI};
}
WorkListObserver.printNewInstrs();
Changed |= Res == LegalizerHelper::Legalized;
}
// Try to combine the instructions in RetryList again if there
// are new artifacts. If not, stop legalizing.
if (!RetryList.empty()) {
[Legalizer] Making artifact combining order-independent Legalization algorithm is complicated by two facts: 1) While regular instructions should be possible to legalize in an isolated, per-instruction, context-free manner, legalization artifacts can only be eliminated in pairs, which could be deeply, and ultimately arbitrary nested: { [ () ] }, where which paranthesis kind depicts an artifact kind, like extend, unmerge, etc. Such structure can only be fully eliminated by simple local combines if they are attempted in a particular order (inside out), or alternatively by repeated scans each eliminating only one innermost pair, resulting in O(n^2) complexity. 2) Some artifacts might in fact be regular instructions that could (and sometimes should) be legalized by the target-specific rules. Which means failure to eliminate all artifacts on the first iteration is not a failure, they need to be tried as instructions, which may produce more artifacts, including the ones that are in fact regular instructions, resulting in a non-constant number of iterations required to finish the process. I trust the recently introduced termination condition (no new artifacts were created during as-a-regular-instruction-retrial of artifacts not eliminated on the previous iteration) to be efficient in providing termination, but only performing the legalization in full if and only if at each step such chains of artifacts are successfully eliminated in full as well. Which is currently not guaranteed, as the artifact combines are applied only once and in an arbitrary order that has to do with the order of creation or insertion of artifacts into their worklist, which is a no particular order. In this patch I make a small change to the artifact combiner, making it to re-insert into the worklist immediate (modulo a look-through copies) artifact users of each vreg that changes its definition due to an artifact combine. Here the first scan through the artifacts worklist, while not being done in any guaranteed order, only needs to find the innermost pair(s) of artifacts that could be immediately combined out. After that the process follows def-use chains, making them shorter at each step, thus combining everything that can be combined in O(n) time. Reviewers: volkan, aditya_nandakumar, qcolombet, paquette, aemerson, dsanders Reviewed By: aditya_nandakumar, paquette Tags: #llvm Differential Revision: https://reviews.llvm.org/D71448
2019-12-13 05:20:41 +08:00
if (!ArtifactList.empty()) {
while (!RetryList.empty())
ArtifactList.insert(RetryList.pop_back_val());
} else {
LLVM_DEBUG(dbgs() << "No new artifacts created, not retrying!\n");
Helper.MIRBuilder.stopObservingChanges();
return {Changed, RetryList.front()};
}
}
while (!ArtifactList.empty()) {
MachineInstr &MI = *ArtifactList.pop_back_val();
assert(isPreISelGenericOpcode(MI.getOpcode()) &&
"Expecting generic opcode");
if (isTriviallyDead(MI, MRI)) {
LLVM_DEBUG(dbgs() << MI << "Is dead\n");
RemoveDeadInstFromLists(&MI);
MI.eraseFromParentAndMarkDBGValuesForRemoval();
continue;
}
SmallVector<MachineInstr *, 4> DeadInstructions;
LLVM_DEBUG(dbgs() << "Trying to combine: " << MI);
if (ArtCombiner.tryCombineInstruction(MI, DeadInstructions,
WrapperObserver)) {
WorkListObserver.printNewInstrs();
for (auto *DeadMI : DeadInstructions) {
LLVM_DEBUG(dbgs() << *DeadMI << "Is dead\n");
RemoveDeadInstFromLists(DeadMI);
DeadMI->eraseFromParentAndMarkDBGValuesForRemoval();
}
Changed = true;
continue;
}
// If this was not an artifact (that could be combined away), this might
// need special handling. Add it to InstList, so when it's processed
// there, it has to be legal or specially handled.
else {
LLVM_DEBUG(dbgs() << ".. Not combined, moving to instructions list\n");
InstList.insert(&MI);
}
}
} while (!InstList.empty());
return {Changed, /*FailedOn*/ nullptr};
}
bool Legalizer::runOnMachineFunction(MachineFunction &MF) {
// If the ISel pipeline failed, do not bother running that pass.
if (MF.getProperties().hasProperty(
MachineFunctionProperties::Property::FailedISel))
return false;
LLVM_DEBUG(dbgs() << "Legalize Machine IR for: " << MF.getName() << '\n');
init(MF);
const TargetPassConfig &TPC = getAnalysis<TargetPassConfig>();
GISelCSEAnalysisWrapper &Wrapper =
getAnalysis<GISelCSEAnalysisWrapperPass>().getCSEWrapper();
MachineOptimizationRemarkEmitter MORE(MF, /*MBFI=*/nullptr);
const size_t NumBlocks = MF.size();
std::unique_ptr<MachineIRBuilder> MIRBuilder;
GISelCSEInfo *CSEInfo = nullptr;
bool EnableCSE = EnableCSEInLegalizer.getNumOccurrences()
? EnableCSEInLegalizer
: TPC.isGISelCSEEnabled();
if (EnableCSE) {
MIRBuilder = std::make_unique<CSEMIRBuilder>();
CSEInfo = &Wrapper.get(TPC.getCSEConfig());
MIRBuilder->setCSEInfo(CSEInfo);
} else
MIRBuilder = std::make_unique<MachineIRBuilder>();
SmallVector<GISelChangeObserver *, 1> AuxObservers;
if (EnableCSE && CSEInfo) {
// We want CSEInfo in addition to WorkListObserver to observe all changes.
AuxObservers.push_back(CSEInfo);
}
assert(!CSEInfo || !errorToBool(CSEInfo->verify()));
const LegalizerInfo &LI = *MF.getSubtarget().getLegalizerInfo();
MFResult Result = legalizeMachineFunction(MF, LI, AuxObservers, *MIRBuilder);
if (Result.FailedOn) {
reportGISelFailure(MF, TPC, MORE, "gisel-legalize",
"unable to legalize instruction", *Result.FailedOn);
return false;
}
// For now don't support if new blocks are inserted - we would need to fix the
// outer loop for that.
if (MF.size() != NumBlocks) {
MachineOptimizationRemarkMissed R("gisel-legalize", "GISelFailure",
MF.getFunction().getSubprogram(),
/*MBB=*/nullptr);
R << "inserting blocks is not supported yet";
reportGISelFailure(MF, TPC, MORE, R);
return false;
}
// If for some reason CSE was not enabled, make sure that we invalidate the
// CSEInfo object (as we currently declare that the analysis is preserved).
// The next time get on the wrapper is called, it will force it to recompute
// the analysis.
if (!EnableCSE)
Wrapper.setComputed(false);
return Result.Changed;
}