[llvm] a6a6fca - [ubsan][pgo] Pass to remove ubsan checks based on profile data (#83471)
via llvm-commits
llvm-commits at lists.llvm.org
Thu Mar 7 10:52:48 PST 2024
Author: Vitaly Buka
Date: 2024-03-07T10:52:43-08:00
New Revision: a6a6fca7911feab8325129ea57247303b3c8d558
URL: https://github.com/llvm/llvm-project/commit/a6a6fca7911feab8325129ea57247303b3c8d558
DIFF: https://github.com/llvm/llvm-project/commit/a6a6fca7911feab8325129ea57247303b3c8d558.diff
LOG: [ubsan][pgo] Pass to remove ubsan checks based on profile data (#83471)
UBSAN checks can be too expensive to be used
in release binaries. However not all code affect
performace in the same way. Removing small
number of checks in hot code we can performance
loss, preserving most of the checks.
Added:
llvm/include/llvm/Transforms/Instrumentation/RemoveTrapsPass.h
llvm/lib/Transforms/Instrumentation/RemoveTrapsPass.cpp
llvm/test/Transforms/RemoveTraps/remove-traps.ll
Modified:
llvm/lib/Passes/PassBuilder.cpp
llvm/lib/Passes/PassRegistry.def
llvm/lib/Transforms/Instrumentation/CMakeLists.txt
Removed:
################################################################################
diff --git a/llvm/include/llvm/Transforms/Instrumentation/RemoveTrapsPass.h b/llvm/include/llvm/Transforms/Instrumentation/RemoveTrapsPass.h
new file mode 100644
index 00000000000000..58f6bbcec5dc9d
--- /dev/null
+++ b/llvm/include/llvm/Transforms/Instrumentation/RemoveTrapsPass.h
@@ -0,0 +1,32 @@
+//===- RemoveTrapsPass.h ----------------------------------------*- C++ -*-===//
+//
+// 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 provides the interface for the pass responsible for removing
+/// expensive ubsan checks.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_TRANSFORMS_INSTRUMENTATION_UBSANOPTIMIZATIONPASS_H
+#define LLVM_TRANSFORMS_INSTRUMENTATION_UBSANOPTIMIZATIONPASS_H
+
+#include "llvm/IR/Function.h"
+#include "llvm/IR/PassManager.h"
+#include "llvm/Pass.h"
+
+namespace llvm {
+
+// This pass is responsible for removing optional traps, like llvm.ubsantrap
+// from the hot code.
+class RemoveTrapsPass : public PassInfoMixin<RemoveTrapsPass> {
+public:
+ PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM);
+};
+
+} // namespace llvm
+
+#endif
diff --git a/llvm/lib/Passes/PassBuilder.cpp b/llvm/lib/Passes/PassBuilder.cpp
index e0bc57f8bf72f0..4d1eb10d2d41c6 100644
--- a/llvm/lib/Passes/PassBuilder.cpp
+++ b/llvm/lib/Passes/PassBuilder.cpp
@@ -177,6 +177,7 @@
#include "llvm/Transforms/Instrumentation/PGOForceFunctionAttrs.h"
#include "llvm/Transforms/Instrumentation/PGOInstrumentation.h"
#include "llvm/Transforms/Instrumentation/PoisonChecking.h"
+#include "llvm/Transforms/Instrumentation/RemoveTrapsPass.h"
#include "llvm/Transforms/Instrumentation/SanitizerBinaryMetadata.h"
#include "llvm/Transforms/Instrumentation/SanitizerCoverage.h"
#include "llvm/Transforms/Instrumentation/ThreadSanitizer.h"
diff --git a/llvm/lib/Passes/PassRegistry.def b/llvm/lib/Passes/PassRegistry.def
index a345e8d72d9399..41f16d0915bf23 100644
--- a/llvm/lib/Passes/PassRegistry.def
+++ b/llvm/lib/Passes/PassRegistry.def
@@ -422,6 +422,7 @@ FUNCTION_PASS("print<uniformity>", UniformityInfoPrinterPass(dbgs()))
FUNCTION_PASS("reassociate", ReassociatePass())
FUNCTION_PASS("redundant-dbg-inst-elim", RedundantDbgInstEliminationPass())
FUNCTION_PASS("reg2mem", RegToMemPass())
+FUNCTION_PASS("remove-traps", RemoveTrapsPass())
FUNCTION_PASS("safe-stack", SafeStackPass(TM))
FUNCTION_PASS("scalarize-masked-mem-intrin", ScalarizeMaskedMemIntrinPass())
FUNCTION_PASS("scalarizer", ScalarizerPass())
diff --git a/llvm/lib/Transforms/Instrumentation/CMakeLists.txt b/llvm/lib/Transforms/Instrumentation/CMakeLists.txt
index ee9aa73ff03403..b23a6ed1f08415 100644
--- a/llvm/lib/Transforms/Instrumentation/CMakeLists.txt
+++ b/llvm/lib/Transforms/Instrumentation/CMakeLists.txt
@@ -17,6 +17,7 @@ add_llvm_component_library(LLVMInstrumentation
PGOInstrumentation.cpp
PGOMemOPSizeOpt.cpp
PoisonChecking.cpp
+ RemoveTrapsPass.cpp
SanitizerCoverage.cpp
SanitizerBinaryMetadata.cpp
ValueProfileCollector.cpp
diff --git a/llvm/lib/Transforms/Instrumentation/RemoveTrapsPass.cpp b/llvm/lib/Transforms/Instrumentation/RemoveTrapsPass.cpp
new file mode 100644
index 00000000000000..d87f7482a21d25
--- /dev/null
+++ b/llvm/lib/Transforms/Instrumentation/RemoveTrapsPass.cpp
@@ -0,0 +1,104 @@
+//===- RemoveTrapsPass.cpp --------------------------------------*- C++ -*-===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/Transforms/Instrumentation/RemoveTrapsPass.h"
+
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/Statistic.h"
+#include "llvm/Analysis/ProfileSummaryInfo.h"
+#include "llvm/IR/Instructions.h"
+#include "llvm/IR/IntrinsicInst.h"
+#include "llvm/IR/Intrinsics.h"
+#include "llvm/Support/RandomNumberGenerator.h"
+#include <memory>
+#include <random>
+
+using namespace llvm;
+
+#define DEBUG_TYPE "remove-traps"
+
+static cl::opt<int> HotPercentileCutoff(
+ "remove-traps-percentile-cutoff-hot", cl::init(0),
+ cl::desc("Alternative hot percentile cuttoff. By default "
+ "`-profile-summary-cutoff-hot` is used."));
+
+static cl::opt<float>
+ RandomRate("remove-traps-random-rate", cl::init(0.0),
+ cl::desc("Probability value in the range [0.0, 1.0] of "
+ "unconditional pseudo-random checks removal."));
+
+STATISTIC(NumChecksTotal, "Number of checks");
+STATISTIC(NumChecksRemoved, "Number of removed checks");
+
+static bool removeUbsanTraps(Function &F, const BlockFrequencyInfo &BFI,
+ const ProfileSummaryInfo *PSI) {
+ SmallVector<IntrinsicInst *, 16> Remove;
+ std::unique_ptr<RandomNumberGenerator> Rng;
+
+ auto ShouldRemove = [&](bool IsHot) {
+ if (!RandomRate.getNumOccurrences())
+ return IsHot;
+ if (!Rng)
+ Rng = F.getParent()->createRNG(F.getName());
+ std::bernoulli_distribution D(RandomRate);
+ return D(*Rng);
+ };
+
+ for (BasicBlock &BB : F) {
+ for (Instruction &I : BB) {
+ IntrinsicInst *II = dyn_cast<IntrinsicInst>(&I);
+ if (!II)
+ continue;
+ auto ID = II->getIntrinsicID();
+ switch (ID) {
+ case Intrinsic::ubsantrap: {
+ ++NumChecksTotal;
+
+ bool IsHot = false;
+ if (PSI) {
+ uint64_t Count = 0;
+ for (const auto *PR : predecessors(&BB))
+ Count += BFI.getBlockProfileCount(PR).value_or(0);
+
+ IsHot =
+ HotPercentileCutoff.getNumOccurrences()
+ ? (HotPercentileCutoff > 0 &&
+ PSI->isHotCountNthPercentile(HotPercentileCutoff, Count))
+ : PSI->isHotCount(Count);
+ }
+
+ if (ShouldRemove(IsHot)) {
+ Remove.push_back(II);
+ ++NumChecksRemoved;
+ }
+ break;
+ }
+ default:
+ break;
+ }
+ }
+ }
+
+ for (IntrinsicInst *I : Remove)
+ I->eraseFromParent();
+
+ return !Remove.empty();
+}
+
+PreservedAnalyses RemoveTrapsPass::run(Function &F,
+ FunctionAnalysisManager &AM) {
+ if (F.isDeclaration())
+ return PreservedAnalyses::all();
+ auto &MAMProxy = AM.getResult<ModuleAnalysisManagerFunctionProxy>(F);
+ ProfileSummaryInfo *PSI =
+ MAMProxy.getCachedResult<ProfileSummaryAnalysis>(*F.getParent());
+ BlockFrequencyInfo &BFI = AM.getResult<BlockFrequencyAnalysis>(F);
+
+ return removeUbsanTraps(F, BFI, PSI) ? PreservedAnalyses::none()
+ : PreservedAnalyses::all();
+}
diff --git a/llvm/test/Transforms/RemoveTraps/remove-traps.ll b/llvm/test/Transforms/RemoveTraps/remove-traps.ll
new file mode 100644
index 00000000000000..71549e7d9b4122
--- /dev/null
+++ b/llvm/test/Transforms/RemoveTraps/remove-traps.ll
@@ -0,0 +1,397 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
+; RUN: opt < %s -passes='function(remove-traps)' -S | FileCheck %s --check-prefixes=NOPROFILE
+; RUN: opt < %s -passes='function(remove-traps)' -remove-traps-random-rate=1 -S | FileCheck %s --check-prefixes=ALL
+; RUN: opt < %s -passes='require<profile-summary>,function(remove-traps)' -S | FileCheck %s --check-prefixes=HOT
+; RUN: opt < %s -passes='require<profile-summary>,function(remove-traps)' -remove-traps-percentile-cutoff-hot=700000 -S | FileCheck %s --check-prefixes=HOT70
+
+target triple = "x86_64-pc-linux-gnu"
+
+declare void @llvm.ubsantrap(i8 immarg)
+
+define dso_local noundef i32 @simple(ptr noundef readonly %0) {
+; NOPROFILE-LABEL: define dso_local noundef i32 @simple(
+; NOPROFILE-SAME: ptr noundef readonly [[TMP0:%.*]]) {
+; NOPROFILE-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; NOPROFILE-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; NOPROFILE: 3:
+; NOPROFILE-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; NOPROFILE-NEXT: unreachable
+; NOPROFILE: 4:
+; NOPROFILE-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; NOPROFILE-NEXT: ret i32 [[TMP5]]
+;
+; ALL-LABEL: define dso_local noundef i32 @simple(
+; ALL-SAME: ptr noundef readonly [[TMP0:%.*]]) {
+; ALL-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; ALL-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; ALL: 3:
+; ALL-NEXT: unreachable
+; ALL: 4:
+; ALL-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; ALL-NEXT: ret i32 [[TMP5]]
+;
+; HOT-LABEL: define dso_local noundef i32 @simple(
+; HOT-SAME: ptr noundef readonly [[TMP0:%.*]]) {
+; HOT-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; HOT-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; HOT: 3:
+; HOT-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; HOT-NEXT: unreachable
+; HOT: 4:
+; HOT-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; HOT-NEXT: ret i32 [[TMP5]]
+;
+; HOT70-LABEL: define dso_local noundef i32 @simple(
+; HOT70-SAME: ptr noundef readonly [[TMP0:%.*]]) {
+; HOT70-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; HOT70-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; HOT70: 3:
+; HOT70-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; HOT70-NEXT: unreachable
+; HOT70: 4:
+; HOT70-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; HOT70-NEXT: ret i32 [[TMP5]]
+;
+ %2 = icmp eq ptr %0, null
+ br i1 %2, label %3, label %4
+
+3:
+ tail call void @llvm.ubsantrap(i8 22)
+ unreachable
+
+4:
+ %5 = load i32, ptr %0, align 4
+ ret i32 %5
+}
+
+
+define dso_local noundef i32 @hot(ptr noundef readonly %0) !prof !36 {
+; NOPROFILE-LABEL: define dso_local noundef i32 @hot(
+; NOPROFILE-SAME: ptr noundef readonly [[TMP0:%.*]]) !prof [[PROF16:![0-9]+]] {
+; NOPROFILE-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; NOPROFILE-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; NOPROFILE: 3:
+; NOPROFILE-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; NOPROFILE-NEXT: unreachable
+; NOPROFILE: 4:
+; NOPROFILE-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; NOPROFILE-NEXT: ret i32 [[TMP5]]
+;
+; ALL-LABEL: define dso_local noundef i32 @hot(
+; ALL-SAME: ptr noundef readonly [[TMP0:%.*]]) !prof [[PROF16:![0-9]+]] {
+; ALL-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; ALL-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; ALL: 3:
+; ALL-NEXT: unreachable
+; ALL: 4:
+; ALL-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; ALL-NEXT: ret i32 [[TMP5]]
+;
+; HOT-LABEL: define dso_local noundef i32 @hot(
+; HOT-SAME: ptr noundef readonly [[TMP0:%.*]]) !prof [[PROF16:![0-9]+]] {
+; HOT-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; HOT-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; HOT: 3:
+; HOT-NEXT: unreachable
+; HOT: 4:
+; HOT-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; HOT-NEXT: ret i32 [[TMP5]]
+;
+; HOT70-LABEL: define dso_local noundef i32 @hot(
+; HOT70-SAME: ptr noundef readonly [[TMP0:%.*]]) !prof [[PROF16:![0-9]+]] {
+; HOT70-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; HOT70-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; HOT70: 3:
+; HOT70-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; HOT70-NEXT: unreachable
+; HOT70: 4:
+; HOT70-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; HOT70-NEXT: ret i32 [[TMP5]]
+;
+ %2 = icmp eq ptr %0, null
+ br i1 %2, label %3, label %4
+
+3:
+ tail call void @llvm.ubsantrap(i8 22)
+ unreachable
+
+4:
+ %5 = load i32, ptr %0, align 4
+ ret i32 %5
+}
+
+define dso_local noundef i32 @veryHot(ptr noundef readonly %0) !prof !39 {
+; NOPROFILE-LABEL: define dso_local noundef i32 @veryHot(
+; NOPROFILE-SAME: ptr noundef readonly [[TMP0:%.*]]) !prof [[PROF17:![0-9]+]] {
+; NOPROFILE-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; NOPROFILE-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; NOPROFILE: 3:
+; NOPROFILE-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; NOPROFILE-NEXT: unreachable
+; NOPROFILE: 4:
+; NOPROFILE-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; NOPROFILE-NEXT: ret i32 [[TMP5]]
+;
+; ALL-LABEL: define dso_local noundef i32 @veryHot(
+; ALL-SAME: ptr noundef readonly [[TMP0:%.*]]) !prof [[PROF17:![0-9]+]] {
+; ALL-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; ALL-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; ALL: 3:
+; ALL-NEXT: unreachable
+; ALL: 4:
+; ALL-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; ALL-NEXT: ret i32 [[TMP5]]
+;
+; HOT-LABEL: define dso_local noundef i32 @veryHot(
+; HOT-SAME: ptr noundef readonly [[TMP0:%.*]]) !prof [[PROF17:![0-9]+]] {
+; HOT-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; HOT-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; HOT: 3:
+; HOT-NEXT: unreachable
+; HOT: 4:
+; HOT-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; HOT-NEXT: ret i32 [[TMP5]]
+;
+; HOT70-LABEL: define dso_local noundef i32 @veryHot(
+; HOT70-SAME: ptr noundef readonly [[TMP0:%.*]]) !prof [[PROF17:![0-9]+]] {
+; HOT70-NEXT: [[TMP2:%.*]] = icmp eq ptr [[TMP0]], null
+; HOT70-NEXT: br i1 [[TMP2]], label [[TMP3:%.*]], label [[TMP4:%.*]]
+; HOT70: 3:
+; HOT70-NEXT: unreachable
+; HOT70: 4:
+; HOT70-NEXT: [[TMP5:%.*]] = load i32, ptr [[TMP0]], align 4
+; HOT70-NEXT: ret i32 [[TMP5]]
+;
+ %2 = icmp eq ptr %0, null
+ br i1 %2, label %3, label %4
+
+3:
+ tail call void @llvm.ubsantrap(i8 22)
+ unreachable
+
+4:
+ %5 = load i32, ptr %0, align 4
+ ret i32 %5
+}
+
+
+define dso_local noundef i32 @branchColdFnHot(i32 noundef %0, ptr noundef readonly %1) !prof !39 {
+; NOPROFILE-LABEL: define dso_local noundef i32 @branchColdFnHot(
+; NOPROFILE-SAME: i32 noundef [[TMP0:%.*]], ptr noundef readonly [[TMP1:%.*]]) !prof [[PROF17]] {
+; NOPROFILE-NEXT: [[TMP3:%.*]] = icmp eq i32 [[TMP0]], 0
+; NOPROFILE-NEXT: br i1 [[TMP3]], label [[TMP9:%.*]], label [[TMP4:%.*]], !prof [[PROF18:![0-9]+]]
+; NOPROFILE: 4:
+; NOPROFILE-NEXT: [[TMP5:%.*]] = icmp eq ptr [[TMP1]], null
+; NOPROFILE-NEXT: br i1 [[TMP5]], label [[TMP6:%.*]], label [[TMP7:%.*]]
+; NOPROFILE: 6:
+; NOPROFILE-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; NOPROFILE-NEXT: unreachable
+; NOPROFILE: 7:
+; NOPROFILE-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP1]], align 4
+; NOPROFILE-NEXT: br label [[TMP9]]
+; NOPROFILE: 9:
+; NOPROFILE-NEXT: [[TMP10:%.*]] = phi i32 [ [[TMP8]], [[TMP7]] ], [ 0, [[TMP2:%.*]] ]
+; NOPROFILE-NEXT: ret i32 [[TMP10]]
+;
+; ALL-LABEL: define dso_local noundef i32 @branchColdFnHot(
+; ALL-SAME: i32 noundef [[TMP0:%.*]], ptr noundef readonly [[TMP1:%.*]]) !prof [[PROF17]] {
+; ALL-NEXT: [[TMP3:%.*]] = icmp eq i32 [[TMP0]], 0
+; ALL-NEXT: br i1 [[TMP3]], label [[TMP9:%.*]], label [[TMP4:%.*]], !prof [[PROF18:![0-9]+]]
+; ALL: 4:
+; ALL-NEXT: [[TMP5:%.*]] = icmp eq ptr [[TMP1]], null
+; ALL-NEXT: br i1 [[TMP5]], label [[TMP6:%.*]], label [[TMP7:%.*]]
+; ALL: 6:
+; ALL-NEXT: unreachable
+; ALL: 7:
+; ALL-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP1]], align 4
+; ALL-NEXT: br label [[TMP9]]
+; ALL: 9:
+; ALL-NEXT: [[TMP10:%.*]] = phi i32 [ [[TMP8]], [[TMP7]] ], [ 0, [[TMP2:%.*]] ]
+; ALL-NEXT: ret i32 [[TMP10]]
+;
+; HOT-LABEL: define dso_local noundef i32 @branchColdFnHot(
+; HOT-SAME: i32 noundef [[TMP0:%.*]], ptr noundef readonly [[TMP1:%.*]]) !prof [[PROF17]] {
+; HOT-NEXT: [[TMP3:%.*]] = icmp eq i32 [[TMP0]], 0
+; HOT-NEXT: br i1 [[TMP3]], label [[TMP9:%.*]], label [[TMP4:%.*]], !prof [[PROF18:![0-9]+]]
+; HOT: 4:
+; HOT-NEXT: [[TMP5:%.*]] = icmp eq ptr [[TMP1]], null
+; HOT-NEXT: br i1 [[TMP5]], label [[TMP6:%.*]], label [[TMP7:%.*]]
+; HOT: 6:
+; HOT-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; HOT-NEXT: unreachable
+; HOT: 7:
+; HOT-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP1]], align 4
+; HOT-NEXT: br label [[TMP9]]
+; HOT: 9:
+; HOT-NEXT: [[TMP10:%.*]] = phi i32 [ [[TMP8]], [[TMP7]] ], [ 0, [[TMP2:%.*]] ]
+; HOT-NEXT: ret i32 [[TMP10]]
+;
+; HOT70-LABEL: define dso_local noundef i32 @branchColdFnHot(
+; HOT70-SAME: i32 noundef [[TMP0:%.*]], ptr noundef readonly [[TMP1:%.*]]) !prof [[PROF17]] {
+; HOT70-NEXT: [[TMP3:%.*]] = icmp eq i32 [[TMP0]], 0
+; HOT70-NEXT: br i1 [[TMP3]], label [[TMP9:%.*]], label [[TMP4:%.*]], !prof [[PROF18:![0-9]+]]
+; HOT70: 4:
+; HOT70-NEXT: [[TMP5:%.*]] = icmp eq ptr [[TMP1]], null
+; HOT70-NEXT: br i1 [[TMP5]], label [[TMP6:%.*]], label [[TMP7:%.*]]
+; HOT70: 6:
+; HOT70-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; HOT70-NEXT: unreachable
+; HOT70: 7:
+; HOT70-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP1]], align 4
+; HOT70-NEXT: br label [[TMP9]]
+; HOT70: 9:
+; HOT70-NEXT: [[TMP10:%.*]] = phi i32 [ [[TMP8]], [[TMP7]] ], [ 0, [[TMP2:%.*]] ]
+; HOT70-NEXT: ret i32 [[TMP10]]
+;
+ %3 = icmp eq i32 %0, 0
+ br i1 %3, label %9, label %4, !prof !38
+
+4:
+ %5 = icmp eq ptr %1, null
+ br i1 %5, label %6, label %7
+
+6:
+ tail call void @llvm.ubsantrap(i8 22) #2
+ unreachable
+
+7:
+ %8 = load i32, ptr %1, align 4
+ br label %9
+
+9:
+ %10 = phi i32 [ %8, %7 ], [ 0, %2 ]
+ ret i32 %10
+}
+
+define dso_local noundef i32 @branchHotFnCold(i32 noundef %0, ptr noundef readonly %1) !prof !36 {
+; NOPROFILE-LABEL: define dso_local noundef i32 @branchHotFnCold(
+; NOPROFILE-SAME: i32 noundef [[TMP0:%.*]], ptr noundef readonly [[TMP1:%.*]]) !prof [[PROF16]] {
+; NOPROFILE-NEXT: [[TMP3:%.*]] = icmp eq i32 [[TMP0]], 0
+; NOPROFILE-NEXT: br i1 [[TMP3]], label [[TMP9:%.*]], label [[TMP4:%.*]], !prof [[PROF19:![0-9]+]]
+; NOPROFILE: 4:
+; NOPROFILE-NEXT: [[TMP5:%.*]] = icmp eq ptr [[TMP1]], null
+; NOPROFILE-NEXT: br i1 [[TMP5]], label [[TMP6:%.*]], label [[TMP7:%.*]]
+; NOPROFILE: 6:
+; NOPROFILE-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; NOPROFILE-NEXT: unreachable
+; NOPROFILE: 7:
+; NOPROFILE-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP1]], align 4
+; NOPROFILE-NEXT: br label [[TMP9]]
+; NOPROFILE: 9:
+; NOPROFILE-NEXT: [[TMP10:%.*]] = phi i32 [ [[TMP8]], [[TMP7]] ], [ 0, [[TMP2:%.*]] ]
+; NOPROFILE-NEXT: ret i32 [[TMP10]]
+;
+; ALL-LABEL: define dso_local noundef i32 @branchHotFnCold(
+; ALL-SAME: i32 noundef [[TMP0:%.*]], ptr noundef readonly [[TMP1:%.*]]) !prof [[PROF16]] {
+; ALL-NEXT: [[TMP3:%.*]] = icmp eq i32 [[TMP0]], 0
+; ALL-NEXT: br i1 [[TMP3]], label [[TMP9:%.*]], label [[TMP4:%.*]], !prof [[PROF19:![0-9]+]]
+; ALL: 4:
+; ALL-NEXT: [[TMP5:%.*]] = icmp eq ptr [[TMP1]], null
+; ALL-NEXT: br i1 [[TMP5]], label [[TMP6:%.*]], label [[TMP7:%.*]]
+; ALL: 6:
+; ALL-NEXT: unreachable
+; ALL: 7:
+; ALL-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP1]], align 4
+; ALL-NEXT: br label [[TMP9]]
+; ALL: 9:
+; ALL-NEXT: [[TMP10:%.*]] = phi i32 [ [[TMP8]], [[TMP7]] ], [ 0, [[TMP2:%.*]] ]
+; ALL-NEXT: ret i32 [[TMP10]]
+;
+; HOT-LABEL: define dso_local noundef i32 @branchHotFnCold(
+; HOT-SAME: i32 noundef [[TMP0:%.*]], ptr noundef readonly [[TMP1:%.*]]) !prof [[PROF16]] {
+; HOT-NEXT: [[TMP3:%.*]] = icmp eq i32 [[TMP0]], 0
+; HOT-NEXT: br i1 [[TMP3]], label [[TMP9:%.*]], label [[TMP4:%.*]], !prof [[PROF19:![0-9]+]]
+; HOT: 4:
+; HOT-NEXT: [[TMP5:%.*]] = icmp eq ptr [[TMP1]], null
+; HOT-NEXT: br i1 [[TMP5]], label [[TMP6:%.*]], label [[TMP7:%.*]]
+; HOT: 6:
+; HOT-NEXT: unreachable
+; HOT: 7:
+; HOT-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP1]], align 4
+; HOT-NEXT: br label [[TMP9]]
+; HOT: 9:
+; HOT-NEXT: [[TMP10:%.*]] = phi i32 [ [[TMP8]], [[TMP7]] ], [ 0, [[TMP2:%.*]] ]
+; HOT-NEXT: ret i32 [[TMP10]]
+;
+; HOT70-LABEL: define dso_local noundef i32 @branchHotFnCold(
+; HOT70-SAME: i32 noundef [[TMP0:%.*]], ptr noundef readonly [[TMP1:%.*]]) !prof [[PROF16]] {
+; HOT70-NEXT: [[TMP3:%.*]] = icmp eq i32 [[TMP0]], 0
+; HOT70-NEXT: br i1 [[TMP3]], label [[TMP9:%.*]], label [[TMP4:%.*]], !prof [[PROF19:![0-9]+]]
+; HOT70: 4:
+; HOT70-NEXT: [[TMP5:%.*]] = icmp eq ptr [[TMP1]], null
+; HOT70-NEXT: br i1 [[TMP5]], label [[TMP6:%.*]], label [[TMP7:%.*]]
+; HOT70: 6:
+; HOT70-NEXT: tail call void @llvm.ubsantrap(i8 22)
+; HOT70-NEXT: unreachable
+; HOT70: 7:
+; HOT70-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP1]], align 4
+; HOT70-NEXT: br label [[TMP9]]
+; HOT70: 9:
+; HOT70-NEXT: [[TMP10:%.*]] = phi i32 [ [[TMP8]], [[TMP7]] ], [ 0, [[TMP2:%.*]] ]
+; HOT70-NEXT: ret i32 [[TMP10]]
+;
+ %3 = icmp eq i32 %0, 0
+ br i1 %3, label %9, label %4, !prof !37
+
+4:
+ %5 = icmp eq ptr %1, null
+ br i1 %5, label %6, label %7
+
+6:
+ tail call void @llvm.ubsantrap(i8 22) #2
+ unreachable
+
+7:
+ %8 = load i32, ptr %1, align 4
+ br label %9
+
+9:
+ %10 = phi i32 [ %8, %7 ], [ 0, %2 ]
+ ret i32 %10
+}
+
+!llvm.module.flags = !{!6}
+!6 = !{i32 1, !"ProfileSummary", !7}
+!7 = !{!8, !9, !10, !11, !12, !13, !14, !17}
+!8 = !{!"ProfileFormat", !"InstrProf"}
+!9 = !{!"TotalCount", i64 30000}
+!10 = !{!"MaxCount", i64 10000}
+!11 = !{!"MaxInternalCount", i64 10000}
+!12 = !{!"MaxFunctionCount", i64 10000}
+!13 = !{!"NumCounts", i64 3}
+!14 = !{!"NumFunctions", i64 5}
+!17 = !{!"DetailedSummary", !18}
+!18 = !{!19, !29, !30, !32, !34}
+!19 = !{i32 10000, i64 10000, i32 3}
+!29 = !{i32 950000, i64 5000, i32 3}
+!30 = !{i32 990000, i64 500, i32 4}
+!32 = !{i32 999900, i64 250, i32 4}
+!34 = !{i32 999999, i64 1, i32 6}
+
+!36 = !{!"function_entry_count", i64 1000}
+!39 = !{!"function_entry_count", i64 7000}
+
+!37 = !{!"branch_weights", i32 1, i32 1000}
+!38 = !{!"branch_weights", i32 1000, i32 1}
+
+;.
+; NOPROFILE: [[PROF16]] = !{!"function_entry_count", i64 1000}
+; NOPROFILE: [[PROF17]] = !{!"function_entry_count", i64 7000}
+; NOPROFILE: [[PROF18]] = !{!"branch_weights", i32 1000, i32 1}
+; NOPROFILE: [[PROF19]] = !{!"branch_weights", i32 1, i32 1000}
+;.
+; ALL: [[PROF16]] = !{!"function_entry_count", i64 1000}
+; ALL: [[PROF17]] = !{!"function_entry_count", i64 7000}
+; ALL: [[PROF18]] = !{!"branch_weights", i32 1000, i32 1}
+; ALL: [[PROF19]] = !{!"branch_weights", i32 1, i32 1000}
+;.
+; HOT: [[PROF16]] = !{!"function_entry_count", i64 1000}
+; HOT: [[PROF17]] = !{!"function_entry_count", i64 7000}
+; HOT: [[PROF18]] = !{!"branch_weights", i32 1000, i32 1}
+; HOT: [[PROF19]] = !{!"branch_weights", i32 1, i32 1000}
+;.
+; HOT70: [[PROF16]] = !{!"function_entry_count", i64 1000}
+; HOT70: [[PROF17]] = !{!"function_entry_count", i64 7000}
+; HOT70: [[PROF18]] = !{!"branch_weights", i32 1000, i32 1}
+; HOT70: [[PROF19]] = !{!"branch_weights", i32 1, i32 1000}
+;.
More information about the llvm-commits
mailing list