[SampleFDO] Extend profile-sample-accurate option to cover isFunctionColdInCallGraph

For SampleFDO, when a callsite doesn't appear in the profile, it will not be marked as cold callsite unless the option -profile-sample-accurate is specified.

But profile-sample-accurate doesn't cover function isFunctionColdInCallGraph which is used to decide whether a function should be put into text.unlikely section, so even if the user knows the profile is accurate and specifies profile-sample-accurate, those functions not appearing in the sample profile are still not be put into text.unlikely section right now.

The patch fixes that.

Differential Revision: https://reviews.llvm.org/D55567

llvm-svn: 348940

diff --git a/llvm/lib/Analysis/ProfileSummaryInfo.cpp b/llvm/lib/Analysis/ProfileSummaryInfo.cpp
index bd8459d..4326e53 100644 (file)
--- a/llvm/lib/Analysis/ProfileSummaryInfo.cpp
+++ b/llvm/lib/Analysis/ProfileSummaryInfo.cpp
@@ -261,7 +261,14 @@ bool ProfileSummaryInfo::isHotBlock(const BasicBlock *BB, BlockFrequencyInfo *BF
 bool ProfileSummaryInfo::isColdBlock(const BasicBlock *BB,
                                   BlockFrequencyInfo *BFI) {
   auto Count = BFI->getBlockProfileCount(BB);
-  return Count && isColdCount(*Count);
+  if (Count)
+    return isColdCount(*Count);
+  if (!hasSampleProfile())
+    return false;
+
+  const Function *F = BB->getParent();
+  return ProfileSampleAccurate ||
+         (F && F->hasFnAttribute("profile-sample-accurate"));
 }
 
 bool ProfileSummaryInfo::isHotCallSite(const CallSite &CS,

diff --git a/llvm/test/Transforms/CodeGenPrepare/section-samplepgo.ll b/llvm/test/Transforms/CodeGenPrepare/section-samplepgo.ll
index 93d2a5f..8b59e7d 100644 (file)
--- a/llvm/test/Transforms/CodeGenPrepare/section-samplepgo.ll
+++ b/llvm/test/Transforms/CodeGenPrepare/section-samplepgo.ll
@@ -1,16 +1,19 @@
 ; RUN: opt < %s -codegenprepare -S | FileCheck %s
+; RUN: opt < %s -codegenprepare -profile-sample-accurate -S | FileCheck %s --check-prefix ACCURATE
 
 target triple = "x86_64-pc-linux-gnu"
 
 ; This tests that hot/cold functions get correct section prefix assigned
 
 ; CHECK: hot_func{{.*}}!section_prefix ![[HOT_ID:[0-9]+]]
+; ACCURATE: hot_func{{.*}}!section_prefix ![[HOT_ID:[0-9]+]]
 ; The entry is hot
 define void @hot_func() !prof !15 {
   ret void
 }
 
 ; CHECK: hot_call_func{{.*}}!section_prefix ![[HOT_ID]]
+; ACCURATE: hot_call_func{{.*}}!section_prefix ![[HOT_ID]]
 ; The sum of 2 callsites are hot
 define void @hot_call_func() !prof !16 {
   call void @hot_func(), !prof !17
@@ -19,6 +22,7 @@ define void @hot_call_func() !prof !16 {
 }
 
 ; CHECK-NOT: normal_func{{.*}}!section_prefix
+; ACCURATE-NOT: normal_func{{.*}}!section_prefix
 ; The sum of all callsites are neither hot or cold
 define void @normal_func() !prof !16 {
   call void @hot_func(), !prof !17
@@ -28,12 +32,36 @@ define void @normal_func() !prof !16 {
 }
 
 ; CHECK: cold_func{{.*}}!section_prefix ![[COLD_ID:[0-9]+]]
+; ACCURATE: cold_func{{.*}}!section_prefix ![[COLD_ID:[0-9]+]]
 ; The entry and the callsite are both cold
 define void @cold_func() !prof !16 {
   call void @hot_func(), !prof !18
   ret void
 }
 
+
+; CHECK-NOT: foo_not_in_profile{{.*}}!section_prefix
+; The function not appearing in profile is neither hot nor cold
+;
+; ACCURATE: foo_not_in_profile{{.*}}!section_prefix ![[COLD_ID:[0-9]+]]
+; The function not appearing in profile is cold when -profile-sample-accurate
+; is on
+define void @foo_not_in_profile() !prof !19 {
+  call void @hot_func()
+  ret void
+}
+
+; CHECK: bar_not_in_profile{{.*}}!section_prefix ![[COLD_ID:[0-9]+]]
+; ACCURATE: bar_not_in_profile{{.*}}!section_prefix ![[COLD_ID:[0-9]+]]
+; The function not appearing in profile is cold when the func has
+; profile-sample-accurate attribute
+define void @bar_not_in_profile() #0 !prof !19 {
+  call void @hot_func()
+  ret void
+}
+
+attributes #0 = { "profile-sample-accurate" }
+
 ; CHECK: ![[HOT_ID]] = !{!"function_section_prefix", !".hot"}
 ; CHECK: ![[COLD_ID]] = !{!"function_section_prefix", !".unlikely"}
 !llvm.module.flags = !{!1}
@@ -55,3 +83,4 @@ define void @cold_func() !prof !16 {
 !16 = !{!"function_entry_count", i64 1}
 !17 = !{!"branch_weights", i32 80}
 !18 = !{!"branch_weights", i32 1}
+!19 = !{!"function_entry_count", i64 -1}