Skip to content

[Metadata] Handle memprof, callsite merging when one is missing. #132106

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Apr 2, 2025
Merged

[Metadata] Handle memprof, callsite merging when one is missing. #132106

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f3df524
[Metadata] Handle memprof, callsite merging when one is missing.
snehasish Mar 19, 2025
7772c93
Address comments.
snehasish Mar 21, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
34 changes: 25 additions & 9 deletions llvm/lib/Transforms/Utils/Local.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3355,8 +3355,12 @@ static void combineMetadata(Instruction *K, const Instruction *J,
case LLVMContext::MD_invariant_group:
// Preserve !invariant.group in K.
break;
// Keep empty cases for mmra, memprof, and callsite to prevent them from
// being removed as unknown metadata. The actual merging is handled
// separately below.
case LLVMContext::MD_mmra:
// Combine MMRAs
case LLVMContext::MD_memprof:
case LLVMContext::MD_callsite:
break;
case LLVMContext::MD_align:
if (!AAOnly && (DoesKMove || !K->hasMetadata(LLVMContext::MD_noundef)))
Expand All @@ -3369,14 +3373,6 @@ static void combineMetadata(Instruction *K, const Instruction *J,
K->setMetadata(Kind,
MDNode::getMostGenericAlignmentOrDereferenceable(JMD, KMD));
break;
case LLVMContext::MD_memprof:
if (!AAOnly)
K->setMetadata(Kind, MDNode::getMergedMemProfMetadata(KMD, JMD));
break;
case LLVMContext::MD_callsite:
if (!AAOnly)
K->setMetadata(Kind, MDNode::getMergedCallsiteMetadata(KMD, JMD));
break;
case LLVMContext::MD_preserve_access_index:
// Preserve !preserve.access.index in K.
break;
Expand Down Expand Up @@ -3420,6 +3416,26 @@ static void combineMetadata(Instruction *K, const Instruction *J,
K->setMetadata(LLVMContext::MD_mmra,
MMRAMetadata::combine(K->getContext(), JMMRA, KMMRA));
}

// Merge memprof metadata.
// Handle separately to support cases where only one instruction has the
// metadata.
auto *JMemProf = J->getMetadata(LLVMContext::MD_memprof);
auto *KMemProf = K->getMetadata(LLVMContext::MD_memprof);
if (!AAOnly && (JMemProf || KMemProf)) {
K->setMetadata(LLVMContext::MD_memprof,
MDNode::getMergedMemProfMetadata(KMemProf, JMemProf));
}

// Merge callsite metadata.
// Handle separately to support cases where only one instruction has the
// metadata.
auto *JCallSite = J->getMetadata(LLVMContext::MD_callsite);
auto *KCallSite = K->getMetadata(LLVMContext::MD_callsite);
if (!AAOnly && (JCallSite || KCallSite)) {
K->setMetadata(LLVMContext::MD_callsite,
MDNode::getMergedCallsiteMetadata(KCallSite, JCallSite));
}
}

void llvm::combineMetadataForCSE(Instruction *K, const Instruction *J,
Expand Down
46 changes: 46 additions & 0 deletions llvm/test/Transforms/SimplifyCFG/merge-calls-memprof-left.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,46 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
;; Test to ensure that memprof related metadata is not dropped when
;; instructions are combined. Currently the metadata from the first instruction
;; is kept, which prevents full loss of profile context information.

; RUN: opt < %s -passes=simplifycfg -S | FileCheck %s

target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please no include data layout, triple, attributes etc in tests if they are not necessary for the test.


define dso_local noundef nonnull ptr @_Z9test_leftb(i1 noundef zeroext %b) local_unnamed_addr #0 {
; CHECK-LABEL: define dso_local noundef nonnull ptr @_Z9test_leftb(
; CHECK-SAME: i1 noundef zeroext [[B:%.*]]) local_unnamed_addr {
; CHECK-NEXT: [[ENTRY:.*:]]
; CHECK-NEXT: [[CALL:%.*]] = call noalias noundef nonnull dereferenceable(4) ptr @_Znwm(i64 noundef 4), !memprof [[META0:![0-9]+]], !callsite [[META3:![0-9]+]]
; CHECK-NEXT: ret ptr [[CALL]]
;
entry:
br i1 %b, label %if.then, label %if.else

if.then: ; preds = %entry
%call = call noalias noundef nonnull dereferenceable(4) ptr @_Znwm(i64 noundef 4), !memprof !0, !callsite !3
br label %if.end

if.else: ; preds = %entry
%call1 = call noalias noundef nonnull dereferenceable(4) ptr @_Znwm(i64 noundef 4)
br label %if.end

if.end: ; preds = %if.else, %if.then
%x.0 = phi ptr [ %call, %if.then ], [ %call1, %if.else ]
ret ptr %x.0
}

declare ptr @_Znwm(i64) nounwind readonly

!0 = !{!1}
!1 = !{!2, !"notcold"}
!2 = !{i64 -852997907418798798, i64 -2101080423462424381, i64 5188446645037944434}
!3 = !{i64 -852997907418798798}

;.
; CHECK: [[META0]] = !{[[META1:![0-9]+]]}
; CHECK: [[META1]] = !{[[META2:![0-9]+]], !"notcold"}
; CHECK: [[META2]] = !{i64 -852997907418798798, i64 -2101080423462424381, i64 5188446645037944434}
; CHECK: [[META3]] = !{i64 -852997907418798798}
;.
46 changes: 46 additions & 0 deletions llvm/test/Transforms/SimplifyCFG/merge-calls-memprof-right.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,46 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
;; Test to ensure that memprof related metadata is not dropped when
;; instructions are combined. Currently the metadata from the first instruction
;; is kept, which prevents full loss of profile context information.

; RUN: opt < %s -passes=simplifycfg -S | FileCheck %s

target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"

define dso_local noundef nonnull ptr @_Z10test_rightb(i1 noundef zeroext %b) local_unnamed_addr #0 {
; CHECK-LABEL: define dso_local noundef nonnull ptr @_Z10test_rightb(
; CHECK-SAME: i1 noundef zeroext [[B:%.*]]) local_unnamed_addr {
; CHECK-NEXT: [[ENTRY:.*:]]
; CHECK-NEXT: [[CALL:%.*]] = call noalias noundef nonnull dereferenceable(4) ptr @_Znwm(i64 noundef 4), !memprof [[META0:![0-9]+]], !callsite [[META3:![0-9]+]]
; CHECK-NEXT: ret ptr [[CALL]]
;
entry:
br i1 %b, label %if.then, label %if.else

if.then: ; preds = %entry
%call = call noalias noundef nonnull dereferenceable(4) ptr @_Znwm(i64 noundef 4)
br label %if.end

if.else: ; preds = %entry
%call1 = call noalias noundef nonnull dereferenceable(4) ptr @_Znwm(i64 noundef 4), !memprof !4, !callsite !7
br label %if.end

if.end: ; preds = %if.else, %if.then
%x.0 = phi ptr [ %call, %if.then ], [ %call1, %if.else ]
ret ptr %x.0
}

declare ptr @_Znwm(i64) nounwind readonly

!4 = !{!5}
!5 = !{!6, !"cold"}
!6 = !{i64 123, i64 -2101080423462424381, i64 5188446645037944434}
!7 = !{i64 123}

;.
; CHECK: [[META0]] = !{[[META1:![0-9]+]]}
; CHECK: [[META1]] = !{[[META2:![0-9]+]], !"cold"}
; CHECK: [[META2]] = !{i64 123, i64 -2101080423462424381, i64 5188446645037944434}
; CHECK: [[META3]] = !{i64 123}
;.
Loading