1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-22 18:54:02 +01:00
llvm-mirror/lib/Analysis/InlineAdvisor.cpp
modimo 46d7a90a64 [InlineAdvisor] Allow replay of inline decisions for the CGSCC inliner from optimization remarks
This change leverages the work done in D83743 to replay in the SampleProfile inliner to also be used in the CGSCC inliner. NOTE: currently restricted to non-ML advisors only.

The added switch `-cgscc-inline-replay=<remarks file>` will replay the inlining decisions in that file where the remarks file is generated via `-Rpass=inline`. The aim here is to make it easier to analyze changes that would modify inlining heuristics to be separated from this behavior. Doing so allows easier examination of assembly and runtime behavior compared to the baseline rather than trying to dig through the large churn caused by inlining.

In LTO compilation, since inlining is done twice you can separately specify replay by passing the flag to the FE (`-cgscc-inline-replay=`) and to the linker (`-Wl,cgscc-inline-replay=`) with the remarks generated from their respective places.

Testing on mysqld by comparing the inline decisions between base (generates remarks.txt) and diff (replay using identical input/tools with remarks.txt) and examining the inlining sites with `diff` shows 14,000 mismatches out of 247,341 for a ~94% replay accuracy. I believe this gap can be narrowed further though for the general case we may never achieve full accuracy. For my personal use, this is close enough to be representative: I set the baseline as the one generated by the replay on identical input/toolset and compare that to my modified input/toolset using the same replay.

Testing:
ninja check-llvm
newly added test correctly replays CGSCC inlining decisions

Reviewed By: mtrofin, wenlei

Differential Revision: https://reviews.llvm.org/D94334
2021-01-25 15:38:57 -08:00

515 lines
19 KiB
C++

//===- InlineAdvisor.cpp - analysis pass implementation -------------------===//
//
// 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
//
//===----------------------------------------------------------------------===//
//
// This file implements InlineAdvisorAnalysis and DefaultInlineAdvisor, and
// related types.
//
//===----------------------------------------------------------------------===//
#include "llvm/Analysis/InlineAdvisor.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/InlineCost.h"
#include "llvm/Analysis/OptimizationRemarkEmitter.h"
#include "llvm/Analysis/ProfileSummaryInfo.h"
#include "llvm/Analysis/ReplayInlineAdvisor.h"
#include "llvm/Analysis/TargetLibraryInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/IR/DebugInfoMetadata.h"
#include "llvm/IR/Instructions.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/raw_ostream.h"
#include <sstream>
using namespace llvm;
#define DEBUG_TYPE "inline"
// This weirdly named statistic tracks the number of times that, when attempting
// to inline a function A into B, we analyze the callers of B in order to see
// if those would be more profitable and blocked inline steps.
STATISTIC(NumCallerCallersAnalyzed, "Number of caller-callers analyzed");
/// Flag to add inline messages as callsite attributes 'inline-remark'.
static cl::opt<bool>
InlineRemarkAttribute("inline-remark-attribute", cl::init(false),
cl::Hidden,
cl::desc("Enable adding inline-remark attribute to"
" callsites processed by inliner but decided"
" to be not inlined"));
// An integer used to limit the cost of inline deferral. The default negative
// number tells shouldBeDeferred to only take the secondary cost into account.
static cl::opt<int>
InlineDeferralScale("inline-deferral-scale",
cl::desc("Scale to limit the cost of inline deferral"),
cl::init(2), cl::Hidden);
extern cl::opt<InlinerFunctionImportStatsOpts> InlinerFunctionImportStats;
void DefaultInlineAdvice::recordUnsuccessfulInliningImpl(
const InlineResult &Result) {
using namespace ore;
llvm::setInlineRemark(*OriginalCB, std::string(Result.getFailureReason()) +
"; " + inlineCostStr(*OIC));
ORE.emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE, "NotInlined", DLoc, Block)
<< NV("Callee", Callee) << " will not be inlined into "
<< NV("Caller", Caller) << ": "
<< NV("Reason", Result.getFailureReason());
});
}
void DefaultInlineAdvice::recordInliningWithCalleeDeletedImpl() {
if (EmitRemarks)
emitInlinedInto(ORE, DLoc, Block, *Callee, *Caller, *OIC);
}
void DefaultInlineAdvice::recordInliningImpl() {
if (EmitRemarks)
emitInlinedInto(ORE, DLoc, Block, *Callee, *Caller, *OIC);
}
llvm::Optional<llvm::InlineCost> static getDefaultInlineAdvice(
CallBase &CB, FunctionAnalysisManager &FAM, const InlineParams &Params) {
Function &Caller = *CB.getCaller();
ProfileSummaryInfo *PSI =
FAM.getResult<ModuleAnalysisManagerFunctionProxy>(Caller)
.getCachedResult<ProfileSummaryAnalysis>(
*CB.getParent()->getParent()->getParent());
auto &ORE = FAM.getResult<OptimizationRemarkEmitterAnalysis>(Caller);
auto GetAssumptionCache = [&](Function &F) -> AssumptionCache & {
return FAM.getResult<AssumptionAnalysis>(F);
};
auto GetBFI = [&](Function &F) -> BlockFrequencyInfo & {
return FAM.getResult<BlockFrequencyAnalysis>(F);
};
auto GetTLI = [&](Function &F) -> const TargetLibraryInfo & {
return FAM.getResult<TargetLibraryAnalysis>(F);
};
auto GetInlineCost = [&](CallBase &CB) {
Function &Callee = *CB.getCalledFunction();
auto &CalleeTTI = FAM.getResult<TargetIRAnalysis>(Callee);
bool RemarksEnabled =
Callee.getContext().getDiagHandlerPtr()->isMissedOptRemarkEnabled(
DEBUG_TYPE);
return getInlineCost(CB, Params, CalleeTTI, GetAssumptionCache, GetTLI,
GetBFI, PSI, RemarksEnabled ? &ORE : nullptr);
};
return llvm::shouldInline(CB, GetInlineCost, ORE,
Params.EnableDeferral.getValueOr(false));
}
std::unique_ptr<InlineAdvice>
DefaultInlineAdvisor::getAdviceImpl(CallBase &CB) {
auto OIC = getDefaultInlineAdvice(CB, FAM, Params);
return std::make_unique<DefaultInlineAdvice>(
this, CB, OIC,
FAM.getResult<OptimizationRemarkEmitterAnalysis>(*CB.getCaller()));
}
InlineAdvice::InlineAdvice(InlineAdvisor *Advisor, CallBase &CB,
OptimizationRemarkEmitter &ORE,
bool IsInliningRecommended)
: Advisor(Advisor), Caller(CB.getCaller()), Callee(CB.getCalledFunction()),
DLoc(CB.getDebugLoc()), Block(CB.getParent()), ORE(ORE),
IsInliningRecommended(IsInliningRecommended) {}
void InlineAdvisor::markFunctionAsDeleted(Function *F) {
assert((!DeletedFunctions.count(F)) &&
"Cannot put cause a function to become dead twice!");
DeletedFunctions.insert(F);
}
void InlineAdvisor::freeDeletedFunctions() {
for (auto *F : DeletedFunctions)
delete F;
DeletedFunctions.clear();
}
void InlineAdvice::recordInlineStatsIfNeeded() {
if (Advisor->ImportedFunctionsStats)
Advisor->ImportedFunctionsStats->recordInline(*Caller, *Callee);
}
void InlineAdvice::recordInlining() {
markRecorded();
recordInlineStatsIfNeeded();
recordInliningImpl();
}
void InlineAdvice::recordInliningWithCalleeDeleted() {
markRecorded();
recordInlineStatsIfNeeded();
Advisor->markFunctionAsDeleted(Callee);
recordInliningWithCalleeDeletedImpl();
}
AnalysisKey InlineAdvisorAnalysis::Key;
bool InlineAdvisorAnalysis::Result::tryCreate(InlineParams Params,
InliningAdvisorMode Mode,
StringRef ReplayFile) {
auto &FAM = MAM.getResult<FunctionAnalysisManagerModuleProxy>(M).getManager();
switch (Mode) {
case InliningAdvisorMode::Default:
Advisor.reset(new DefaultInlineAdvisor(M, FAM, Params));
// Restrict replay to default advisor, ML advisors are stateful so
// replay will need augmentations to interleave with them correctly.
if (!ReplayFile.empty()) {
Advisor = std::make_unique<ReplayInlineAdvisor>(
M, FAM, M.getContext(), std::move(Advisor), ReplayFile,
/* EmitRemarks =*/true);
}
break;
case InliningAdvisorMode::Development:
#ifdef LLVM_HAVE_TF_API
Advisor =
llvm::getDevelopmentModeAdvisor(M, MAM, [&FAM, Params](CallBase &CB) {
auto OIC = getDefaultInlineAdvice(CB, FAM, Params);
return OIC.hasValue();
});
#endif
break;
case InliningAdvisorMode::Release:
#ifdef LLVM_HAVE_TF_AOT
Advisor = llvm::getReleaseModeAdvisor(M, MAM);
#endif
break;
}
return !!Advisor;
}
/// Return true if inlining of CB can block the caller from being
/// inlined which is proved to be more beneficial. \p IC is the
/// estimated inline cost associated with callsite \p CB.
/// \p TotalSecondaryCost will be set to the estimated cost of inlining the
/// caller if \p CB is suppressed for inlining.
static bool
shouldBeDeferred(Function *Caller, InlineCost IC, int &TotalSecondaryCost,
function_ref<InlineCost(CallBase &CB)> GetInlineCost) {
// For now we only handle local or inline functions.
if (!Caller->hasLocalLinkage() && !Caller->hasLinkOnceODRLinkage())
return false;
// If the cost of inlining CB is non-positive, it is not going to prevent the
// caller from being inlined into its callers and hence we don't need to
// defer.
if (IC.getCost() <= 0)
return false;
// Try to detect the case where the current inlining candidate caller (call
// it B) is a static or linkonce-ODR function and is an inlining candidate
// elsewhere, and the current candidate callee (call it C) is large enough
// that inlining it into B would make B too big to inline later. In these
// circumstances it may be best not to inline C into B, but to inline B into
// its callers.
//
// This only applies to static and linkonce-ODR functions because those are
// expected to be available for inlining in the translation units where they
// are used. Thus we will always have the opportunity to make local inlining
// decisions. Importantly the linkonce-ODR linkage covers inline functions
// and templates in C++.
//
// FIXME: All of this logic should be sunk into getInlineCost. It relies on
// the internal implementation of the inline cost metrics rather than
// treating them as truly abstract units etc.
TotalSecondaryCost = 0;
// The candidate cost to be imposed upon the current function.
int CandidateCost = IC.getCost() - 1;
// If the caller has local linkage and can be inlined to all its callers, we
// can apply a huge negative bonus to TotalSecondaryCost.
bool ApplyLastCallBonus = Caller->hasLocalLinkage() && !Caller->hasOneUse();
// This bool tracks what happens if we DO inline C into B.
bool InliningPreventsSomeOuterInline = false;
unsigned NumCallerUsers = 0;
for (User *U : Caller->users()) {
CallBase *CS2 = dyn_cast<CallBase>(U);
// If this isn't a call to Caller (it could be some other sort
// of reference) skip it. Such references will prevent the caller
// from being removed.
if (!CS2 || CS2->getCalledFunction() != Caller) {
ApplyLastCallBonus = false;
continue;
}
InlineCost IC2 = GetInlineCost(*CS2);
++NumCallerCallersAnalyzed;
if (!IC2) {
ApplyLastCallBonus = false;
continue;
}
if (IC2.isAlways())
continue;
// See if inlining of the original callsite would erase the cost delta of
// this callsite. We subtract off the penalty for the call instruction,
// which we would be deleting.
if (IC2.getCostDelta() <= CandidateCost) {
InliningPreventsSomeOuterInline = true;
TotalSecondaryCost += IC2.getCost();
NumCallerUsers++;
}
}
if (!InliningPreventsSomeOuterInline)
return false;
// If all outer calls to Caller would get inlined, the cost for the last
// one is set very low by getInlineCost, in anticipation that Caller will
// be removed entirely. We did not account for this above unless there
// is only one caller of Caller.
if (ApplyLastCallBonus)
TotalSecondaryCost -= InlineConstants::LastCallToStaticBonus;
// If InlineDeferralScale is negative, then ignore the cost of primary
// inlining -- IC.getCost() multiplied by the number of callers to Caller.
if (InlineDeferralScale < 0)
return TotalSecondaryCost < IC.getCost();
int TotalCost = TotalSecondaryCost + IC.getCost() * NumCallerUsers;
int Allowance = IC.getCost() * InlineDeferralScale;
return TotalCost < Allowance;
}
namespace llvm {
static std::basic_ostream<char> &operator<<(std::basic_ostream<char> &R,
const ore::NV &Arg) {
return R << Arg.Val;
}
template <class RemarkT>
RemarkT &operator<<(RemarkT &&R, const InlineCost &IC) {
using namespace ore;
if (IC.isAlways()) {
R << "(cost=always)";
} else if (IC.isNever()) {
R << "(cost=never)";
} else {
R << "(cost=" << ore::NV("Cost", IC.getCost())
<< ", threshold=" << ore::NV("Threshold", IC.getThreshold()) << ")";
}
if (const char *Reason = IC.getReason())
R << ": " << ore::NV("Reason", Reason);
return R;
}
} // namespace llvm
std::string llvm::inlineCostStr(const InlineCost &IC) {
std::stringstream Remark;
Remark << IC;
return Remark.str();
}
void llvm::setInlineRemark(CallBase &CB, StringRef Message) {
if (!InlineRemarkAttribute)
return;
Attribute Attr = Attribute::get(CB.getContext(), "inline-remark", Message);
CB.addAttribute(AttributeList::FunctionIndex, Attr);
}
/// Return the cost only if the inliner should attempt to inline at the given
/// CallSite. If we return the cost, we will emit an optimisation remark later
/// using that cost, so we won't do so from this function. Return None if
/// inlining should not be attempted.
Optional<InlineCost>
llvm::shouldInline(CallBase &CB,
function_ref<InlineCost(CallBase &CB)> GetInlineCost,
OptimizationRemarkEmitter &ORE, bool EnableDeferral) {
using namespace ore;
InlineCost IC = GetInlineCost(CB);
Instruction *Call = &CB;
Function *Callee = CB.getCalledFunction();
Function *Caller = CB.getCaller();
if (IC.isAlways()) {
LLVM_DEBUG(dbgs() << " Inlining " << inlineCostStr(IC)
<< ", Call: " << CB << "\n");
return IC;
}
if (!IC) {
LLVM_DEBUG(dbgs() << " NOT Inlining " << inlineCostStr(IC)
<< ", Call: " << CB << "\n");
if (IC.isNever()) {
ORE.emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE, "NeverInline", Call)
<< NV("Callee", Callee) << " not inlined into "
<< NV("Caller", Caller) << " because it should never be inlined "
<< IC;
});
} else {
ORE.emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE, "TooCostly", Call)
<< NV("Callee", Callee) << " not inlined into "
<< NV("Caller", Caller) << " because too costly to inline "
<< IC;
});
}
setInlineRemark(CB, inlineCostStr(IC));
return None;
}
int TotalSecondaryCost = 0;
if (EnableDeferral &&
shouldBeDeferred(Caller, IC, TotalSecondaryCost, GetInlineCost)) {
LLVM_DEBUG(dbgs() << " NOT Inlining: " << CB
<< " Cost = " << IC.getCost()
<< ", outer Cost = " << TotalSecondaryCost << '\n');
ORE.emit([&]() {
return OptimizationRemarkMissed(DEBUG_TYPE, "IncreaseCostInOtherContexts",
Call)
<< "Not inlining. Cost of inlining " << NV("Callee", Callee)
<< " increases the cost of inlining " << NV("Caller", Caller)
<< " in other contexts";
});
setInlineRemark(CB, "deferred");
// IC does not bool() to false, so get an InlineCost that will.
// This will not be inspected to make an error message.
return None;
}
LLVM_DEBUG(dbgs() << " Inlining " << inlineCostStr(IC) << ", Call: " << CB
<< '\n');
return IC;
}
std::string llvm::getCallSiteLocation(DebugLoc DLoc) {
std::ostringstream CallSiteLoc;
bool First = true;
for (DILocation *DIL = DLoc.get(); DIL; DIL = DIL->getInlinedAt()) {
if (!First)
CallSiteLoc << " @ ";
// Note that negative line offset is actually possible, but we use
// unsigned int to match line offset representation in remarks so
// it's directly consumable by relay advisor.
uint32_t Offset =
DIL->getLine() - DIL->getScope()->getSubprogram()->getLine();
uint32_t Discriminator = DIL->getBaseDiscriminator();
StringRef Name = DIL->getScope()->getSubprogram()->getLinkageName();
if (Name.empty())
Name = DIL->getScope()->getSubprogram()->getName();
CallSiteLoc << Name.str() << ":" << llvm::utostr(Offset) << ":"
<< llvm::utostr(DIL->getColumn());
if (Discriminator)
CallSiteLoc << "." << llvm::utostr(Discriminator);
First = false;
}
return CallSiteLoc.str();
}
void llvm::addLocationToRemarks(OptimizationRemark &Remark, DebugLoc DLoc) {
if (!DLoc.get()) {
return;
}
bool First = true;
Remark << " at callsite ";
for (DILocation *DIL = DLoc.get(); DIL; DIL = DIL->getInlinedAt()) {
if (!First)
Remark << " @ ";
unsigned int Offset = DIL->getLine();
Offset -= DIL->getScope()->getSubprogram()->getLine();
unsigned int Discriminator = DIL->getBaseDiscriminator();
StringRef Name = DIL->getScope()->getSubprogram()->getLinkageName();
if (Name.empty())
Name = DIL->getScope()->getSubprogram()->getName();
Remark << Name << ":" << ore::NV("Line", Offset) << ":"
<< ore::NV("Column", DIL->getColumn());
if (Discriminator)
Remark << "." << ore::NV("Disc", Discriminator);
First = false;
}
Remark << ";";
}
void llvm::emitInlinedInto(OptimizationRemarkEmitter &ORE, DebugLoc DLoc,
const BasicBlock *Block, const Function &Callee,
const Function &Caller, const InlineCost &IC,
bool ForProfileContext, const char *PassName) {
ORE.emit([&]() {
bool AlwaysInline = IC.isAlways();
StringRef RemarkName = AlwaysInline ? "AlwaysInline" : "Inlined";
OptimizationRemark Remark(PassName ? PassName : DEBUG_TYPE, RemarkName,
DLoc, Block);
Remark << ore::NV("Callee", &Callee) << " inlined into ";
Remark << ore::NV("Caller", &Caller);
if (ForProfileContext)
Remark << " to match profiling context";
Remark << " with " << IC;
addLocationToRemarks(Remark, DLoc);
return Remark;
});
}
InlineAdvisor::InlineAdvisor(Module &M, FunctionAnalysisManager &FAM)
: M(M), FAM(FAM) {
if (InlinerFunctionImportStats != InlinerFunctionImportStatsOpts::No) {
ImportedFunctionsStats =
std::make_unique<ImportedFunctionsInliningStatistics>();
ImportedFunctionsStats->setModuleInfo(M);
}
}
InlineAdvisor::~InlineAdvisor() {
if (ImportedFunctionsStats) {
assert(InlinerFunctionImportStats != InlinerFunctionImportStatsOpts::No);
ImportedFunctionsStats->dump(InlinerFunctionImportStats ==
InlinerFunctionImportStatsOpts::Verbose);
}
freeDeletedFunctions();
}
std::unique_ptr<InlineAdvice> InlineAdvisor::getMandatoryAdvice(CallBase &CB,
bool Advice) {
return std::make_unique<InlineAdvice>(this, CB, getCallerORE(CB), Advice);
}
InlineAdvisor::MandatoryInliningKind
InlineAdvisor::getMandatoryKind(CallBase &CB, FunctionAnalysisManager &FAM,
OptimizationRemarkEmitter &ORE) {
auto &Callee = *CB.getCalledFunction();
auto GetTLI = [&](Function &F) -> const TargetLibraryInfo & {
return FAM.getResult<TargetLibraryAnalysis>(F);
};
auto &TIR = FAM.getResult<TargetIRAnalysis>(Callee);
auto TrivialDecision =
llvm::getAttributeBasedInliningDecision(CB, &Callee, TIR, GetTLI);
if (TrivialDecision.hasValue()) {
if (TrivialDecision->isSuccess())
return MandatoryInliningKind::Always;
else
return MandatoryInliningKind::Never;
}
return MandatoryInliningKind::NotMandatory;
}
std::unique_ptr<InlineAdvice> InlineAdvisor::getAdvice(CallBase &CB,
bool MandatoryOnly) {
if (!MandatoryOnly)
return getAdviceImpl(CB);
bool Advice = CB.getCaller() != CB.getCalledFunction() &&
MandatoryInliningKind::Always ==
getMandatoryKind(CB, FAM, getCallerORE(CB));
return getMandatoryAdvice(CB, Advice);
}
OptimizationRemarkEmitter &InlineAdvisor::getCallerORE(CallBase &CB) {
return FAM.getResult<OptimizationRemarkEmitterAnalysis>(*CB.getCaller());
}