Restrict call metadata based hotness detection to Sample PGO mode
Summary:
Don't use the metadata on call instructions for determining hotness
unless we are in sample PGO mode, where it is needed because profile
counts are not accurate. In instrumentation mode this is not necessary
and does more harm than good when calls have VP metadata that hasn't
been properly scaled after transformations or dropped after constant
prop based devirtualization (both should be fixed, but we don't need
to do this in the first place for instrumentation PGO).
This required adjusting a number of tests to distinguish between sample
and instrumentation PGO handling, and to add in profile summary metadata
so that getProfileCount can get the summary.
Reviewers: davidxl, danielcdh
Subscribers: aemerson, rengolin, mehdi_amini, Prazek, llvm-commits
Differential Revision: https://reviews.llvm.org/D32877
llvm-svn: 302844
diff --git a/llvm/test/Transforms/CodeGenPrepare/section-samplepgo.ll b/llvm/test/Transforms/CodeGenPrepare/section-samplepgo.ll
new file mode 100644
index 0000000..93d2a5f
--- /dev/null
+++ b/llvm/test/Transforms/CodeGenPrepare/section-samplepgo.ll
@@ -0,0 +1,57 @@
+; RUN: opt < %s -codegenprepare -S | FileCheck %s
+
+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]+]]
+; The entry is hot
+define void @hot_func() !prof !15 {
+ ret void
+}
+
+; CHECK: 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
+ call void @hot_func(), !prof !17
+ ret void
+}
+
+; CHECK-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
+ call void @hot_func(), !prof !18
+ call void @hot_func(), !prof !18
+ ret void
+}
+
+; CHECK: 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: ![[HOT_ID]] = !{!"function_section_prefix", !".hot"}
+; CHECK: ![[COLD_ID]] = !{!"function_section_prefix", !".unlikely"}
+!llvm.module.flags = !{!1}
+!1 = !{i32 1, !"ProfileSummary", !2}
+!2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
+!3 = !{!"ProfileFormat", !"SampleProfile"}
+!4 = !{!"TotalCount", i64 10000}
+!5 = !{!"MaxCount", i64 1000}
+!6 = !{!"MaxInternalCount", i64 1}
+!7 = !{!"MaxFunctionCount", i64 1000}
+!8 = !{!"NumCounts", i64 3}
+!9 = !{!"NumFunctions", i64 3}
+!10 = !{!"DetailedSummary", !11}
+!11 = !{!12, !13, !14}
+!12 = !{i32 10000, i64 100, i32 1}
+!13 = !{i32 999000, i64 100, i32 1}
+!14 = !{i32 999999, i64 1, i32 2}
+!15 = !{!"function_entry_count", i64 1000}
+!16 = !{!"function_entry_count", i64 1}
+!17 = !{!"branch_weights", i32 80}
+!18 = !{!"branch_weights", i32 1}
diff --git a/llvm/test/Transforms/CodeGenPrepare/section.ll b/llvm/test/Transforms/CodeGenPrepare/section.ll
index 775f473..4f3144e 100644
--- a/llvm/test/Transforms/CodeGenPrepare/section.ll
+++ b/llvm/test/Transforms/CodeGenPrepare/section.ll
@@ -10,26 +10,26 @@
ret void
}
-; CHECK: hot_call_func{{.*}}!section_prefix ![[HOT_ID]]
-; The sum of 2 callsites are hot
-define void @hot_call_func() !prof !16 {
+; For instrumentation based PGO, we should only look at entry counts,
+; not call site VP metadata (which can exist on value profiled memcpy,
+; or possibly left behind after static analysis based devirtualization).
+; CHECK: cold_func1{{.*}}!section_prefix ![[COLD_ID:[0-9]+]]
+define void @cold_func1() !prof !16 {
call void @hot_func(), !prof !17
call void @hot_func(), !prof !17
ret void
}
-; CHECK-NOT: normal_func{{.*}}!section_prefix
-; The sum of all callsites are neither hot or cold
-define void @normal_func() !prof !16 {
+; CHECK: cold_func2{{.*}}!section_prefix
+define void @cold_func2() !prof !16 {
call void @hot_func(), !prof !17
call void @hot_func(), !prof !18
call void @hot_func(), !prof !18
ret void
}
-; CHECK: cold_func{{.*}}!section_prefix ![[COLD_ID:[0-9]+]]
-; The entry and the callsite are both cold
-define void @cold_func() !prof !16 {
+; CHECK: cold_func3{{.*}}!section_prefix ![[COLD_ID]]
+define void @cold_func3() !prof !16 {
call void @hot_func(), !prof !18
ret void
}
diff --git a/llvm/test/Transforms/Inline/prof-update.ll b/llvm/test/Transforms/Inline/prof-update.ll
index 3fefa1c..4a4471e8 100644
--- a/llvm/test/Transforms/Inline/prof-update.ll
+++ b/llvm/test/Transforms/Inline/prof-update.ll
@@ -6,21 +6,21 @@
@func = global void ()* null
; CHECK: define void @callee(i32 %n) !prof ![[ENTRY_COUNT:[0-9]*]]
-define void @callee(i32 %n) !prof !1 {
+define void @callee(i32 %n) !prof !15 {
%cond = icmp sle i32 %n, 10
br i1 %cond, label %cond_true, label %cond_false
cond_true:
; ext1 is optimized away, thus not updated.
; CHECK: call void @ext1(), !prof ![[COUNT_CALLEE1:[0-9]*]]
- call void @ext1(), !prof !2
+ call void @ext1(), !prof !16
ret void
cond_false:
; ext is cloned and updated.
; CHECK: call void @ext(), !prof ![[COUNT_CALLEE:[0-9]*]]
- call void @ext(), !prof !2
+ call void @ext(), !prof !16
%f = load void ()*, void ()** @func
; CHECK: call void %f(), !prof ![[COUNT_IND_CALLEE:[0-9]*]]
- call void %f(), !prof !4
+ call void %f(), !prof !18
ret void
}
@@ -28,16 +28,29 @@
define void @caller() {
; CHECK: call void @ext(), !prof ![[COUNT_CALLER:[0-9]*]]
; CHECK: call void %f.i(), !prof ![[COUNT_IND_CALLER:[0-9]*]]
- call void @callee(i32 15), !prof !3
+ call void @callee(i32 15), !prof !17
ret void
}
-!llvm.module.flags = !{!0}
-!0 = !{i32 1, !"MaxFunctionCount", i32 2000}
-!1 = !{!"function_entry_count", i64 1000}
-!2 = !{!"branch_weights", i64 2000}
-!3 = !{!"branch_weights", i64 400}
-!4 = !{!"VP", i32 0, i64 140, i64 111, i64 80, i64 222, i64 40, i64 333, i64 20}
+!llvm.module.flags = !{!1}
+!1 = !{i32 1, !"ProfileSummary", !2}
+!2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
+!3 = !{!"ProfileFormat", !"SampleProfile"}
+!4 = !{!"TotalCount", i64 10000}
+!5 = !{!"MaxCount", i64 10}
+!6 = !{!"MaxInternalCount", i64 1}
+!7 = !{!"MaxFunctionCount", i64 2000}
+!8 = !{!"NumCounts", i64 2}
+!9 = !{!"NumFunctions", i64 2}
+!10 = !{!"DetailedSummary", !11}
+!11 = !{!12, !13, !14}
+!12 = !{i32 10000, i64 100, i32 1}
+!13 = !{i32 999000, i64 100, i32 1}
+!14 = !{i32 999999, i64 1, i32 2}
+!15 = !{!"function_entry_count", i64 1000}
+!16 = !{!"branch_weights", i64 2000}
+!17 = !{!"branch_weights", i64 400}
+!18 = !{!"VP", i32 0, i64 140, i64 111, i64 80, i64 222, i64 40, i64 333, i64 20}
attributes #0 = { alwaysinline }
; CHECK: ![[ENTRY_COUNT]] = !{!"function_entry_count", i64 600}
; CHECK: ![[COUNT_CALLEE1]] = !{!"branch_weights", i64 2000}