Move accurate-sample-profile into the function attribute.

Summary: We need to have accurate-sample-profile in function attribute so that it works with LTO.

Reviewers: davidxl, rsmith

Reviewed By: davidxl

Subscribers: sanjoy, mehdi_amini, javed.absar, llvm-commits, eraman

Differential Revision: https://reviews.llvm.org/D37113

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@311706 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/include/llvm/IR/Attributes.td b/include/llvm/IR/Attributes.td
index 76284babc1aedb55e75e8f84382f681e693d44b8..2cf58b0fa41363f3c43b18349d7e4160422410e8 100644 (file)
--- a/include/llvm/IR/Attributes.td
+++ b/include/llvm/IR/Attributes.td
@@ -185,6 +185,7 @@ def NoInfsFPMath : StrBoolAttr<"no-infs-fp-math">;
 def NoNansFPMath : StrBoolAttr<"no-nans-fp-math">;
 def UnsafeFPMath : StrBoolAttr<"unsafe-fp-math">;
 def NoJumpTables : StrBoolAttr<"no-jump-tables">;
+def ProfileSampleAccurate : StrBoolAttr<"profile-sample-accurate">;
 
 class CompatRule<string F> {
   // The name of the function called to check the attribute of the caller and
@@ -216,6 +217,7 @@ def : MergeRule<"setAND<NoNansFPMathAttr>">;
 def : MergeRule<"setAND<UnsafeFPMathAttr>">;
 def : MergeRule<"setOR<NoImplicitFloatAttr>">;
 def : MergeRule<"setOR<NoJumpTablesAttr>">;
+def : MergeRule<"setOR<ProfileSampleAccurateAttr>">;
 def : MergeRule<"adjustCallerSSPLevel">;
 def : MergeRule<"adjustCallerStackProbes">;
 def : MergeRule<"adjustCallerStackProbeSize">;

diff --git a/lib/Analysis/ProfileSummaryInfo.cpp b/lib/Analysis/ProfileSummaryInfo.cpp
index 4bad196a91a5de397a2bfa09052fd75f43a3948e..671744f93fb8f7f4fb43bb448f8a19e08354beb4 100644 (file)
--- a/lib/Analysis/ProfileSummaryInfo.cpp
+++ b/lib/Analysis/ProfileSummaryInfo.cpp
@@ -39,8 +39,8 @@ static cl::opt<int> ProfileSummaryCutoffCold(
     cl::desc("A count is cold if it is below the minimum count"
              " to reach this percentile of total counts."));
 
-static cl::opt<bool> AccurateSampleProfile(
-    "accurate-sample-profile", cl::Hidden, cl::init(false),
+static cl::opt<bool> ProfileSampleAccurate(
+    "profile-sample-accurate", cl::Hidden, cl::init(false),
     cl::desc("If the sample profile is accurate, we will mark all un-sampled "
              "callsite as cold. Otherwise, treat un-sampled callsites as if "
              "we have no profile."));
@@ -231,7 +231,8 @@ bool ProfileSummaryInfo::isColdCallSite(const CallSite &CS,
   // If there is no profile for the caller, and we know the profile is
   // accurate, we consider the callsite as cold.
   return (hasSampleProfile() &&
-          (CS.getCaller()->getEntryCount() || AccurateSampleProfile));
+          (CS.getCaller()->getEntryCount() || ProfileSampleAccurate ||
+           CS.getCaller()->hasFnAttribute("profile-sample-accurate")));
 }
 
 INITIALIZE_PASS(ProfileSummaryInfoWrapperPass, "profile-summary-info",

diff --git a/test/Transforms/Inline/inline-cold-callsite-samplepgo.ll b/test/Transforms/Inline/inline-cold-callsite-samplepgo.ll
index 891781eb471d42399606c3de2831c8df9435de05..cf2e75757e7eaa0fc6e6e777dfedc0e705c2981f 100644 (file)
--- a/test/Transforms/Inline/inline-cold-callsite-samplepgo.ll
+++ b/test/Transforms/Inline/inline-cold-callsite-samplepgo.ll
@@ -1,8 +1,7 @@
-; For SamplePGO, if -accurate-sample-profile is specified, cold callsite
+; For SamplePGO, if -profile-sample-accurate is specified, cold callsite
 ; heuristics should be honored if the caller has no profile.
 
 ; RUN: opt < %s -inline -S -inline-cold-callsite-threshold=0 | FileCheck %s
-; RUN: opt < %s -inline -S -inline-cold-callsite-threshold=0 -accurate-sample-profile | FileCheck %s --check-prefix=ACCURATE
 
 define i32 @callee(i32 %x) {
   %x1 = add i32 %x, 1
@@ -14,14 +13,23 @@ define i32 @callee(i32 %x) {
 }
 
 define i32 @caller(i32 %y1) {
+; CHECK-LABEL: @caller
 ; CHECK-NOT: call i32 @callee
-; ACCURATE: call i32 @callee
+  %y2 = call i32 @callee(i32 %y1)
+  ret i32 %y2
+}
+
+define i32 @caller_accurate(i32 %y1) #0 {
+; CHECK-LABEL: @caller_accurate
+; CHECK: call i32 @callee
   %y2 = call i32 @callee(i32 %y1)
   ret i32 %y2
 }
 
 declare void @extern()
 
+attributes #0 = { "profile-sample-accurate" }
+
 !llvm.module.flags = !{!1}
 !1 = !{i32 1, !"ProfileSummary", !2}
 !2 = !{!3, !4, !5, !6, !7, !8, !9, !10}