-
Notifications
You must be signed in to change notification settings - Fork 10.8k
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.
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
release/18.x: [ArgPromotion] Remove incorrect TranspBlocks set for loads. (#84835) #84945
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
@nikic What do you think about merging this PR to the release branch? |
@llvm/pr-subscribers-llvm-transforms Author: None (llvmbot) ChangesRequested by: @nikic Full diff: https://github.com/llvm/llvm-project/pull/84945.diff 2 Files Affected:
diff --git a/llvm/lib/Transforms/IPO/ArgumentPromotion.cpp b/llvm/lib/Transforms/IPO/ArgumentPromotion.cpp
index 8058282c422503..062a3d341007ce 100644
--- a/llvm/lib/Transforms/IPO/ArgumentPromotion.cpp
+++ b/llvm/lib/Transforms/IPO/ArgumentPromotion.cpp
@@ -652,10 +652,6 @@ static bool findArgParts(Argument *Arg, const DataLayout &DL, AAResults &AAR,
// check to see if the pointer is guaranteed to not be modified from entry of
// the function to each of the load instructions.
- // Because there could be several/many load instructions, remember which
- // blocks we know to be transparent to the load.
- df_iterator_default_set<BasicBlock *, 16> TranspBlocks;
-
for (LoadInst *Load : Loads) {
// Check to see if the load is invalidated from the start of the block to
// the load itself.
@@ -669,7 +665,7 @@ static bool findArgParts(Argument *Arg, const DataLayout &DL, AAResults &AAR,
// To do this, we perform a depth first search on the inverse CFG from the
// loading block.
for (BasicBlock *P : predecessors(BB)) {
- for (BasicBlock *TranspBB : inverse_depth_first_ext(P, TranspBlocks))
+ for (BasicBlock *TranspBB : inverse_depth_first(P))
if (AAR.canBasicBlockModify(*TranspBB, Loc))
return false;
}
diff --git a/llvm/test/Transforms/ArgumentPromotion/aliasing-and-non-aliasing-loads-with-clobber.ll b/llvm/test/Transforms/ArgumentPromotion/aliasing-and-non-aliasing-loads-with-clobber.ll
new file mode 100644
index 00000000000000..1e1669b29b0db6
--- /dev/null
+++ b/llvm/test/Transforms/ArgumentPromotion/aliasing-and-non-aliasing-loads-with-clobber.ll
@@ -0,0 +1,100 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
+; RUN: opt -p argpromotion -S %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"
+
+@f = dso_local global { i16, i64 } { i16 1, i64 0 }, align 8
+
+; Test case for https://github.com/llvm/llvm-project/issues/84807.
+
+; Make sure the loads from @callee are not moved to @caller, as the store
+; in %then may aliases to load from %q.
+
+define i32 @caller1(i1 %c) {
+; CHECK-LABEL: define i32 @caller1(
+; CHECK-SAME: i1 [[C:%.*]]) {
+; CHECK-NEXT: entry:
+; CHECK-NEXT: call void @callee1(ptr noundef nonnull @f, i1 [[C]])
+; CHECK-NEXT: ret i32 0
+;
+entry:
+ call void @callee1(ptr noundef nonnull @f, i1 %c)
+ ret i32 0
+}
+
+define internal void @callee1(ptr nocapture noundef readonly %q, i1 %c) {
+; CHECK-LABEL: define internal void @callee1(
+; CHECK-SAME: ptr nocapture noundef readonly [[Q:%.*]], i1 [[C:%.*]]) {
+; CHECK-NEXT: entry:
+; CHECK-NEXT: br i1 [[C]], label [[THEN:%.*]], label [[EXIT:%.*]]
+; CHECK: then:
+; CHECK-NEXT: store i16 123, ptr @f, align 8
+; CHECK-NEXT: br label [[EXIT]]
+; CHECK: exit:
+; CHECK-NEXT: [[Q_0_VAL:%.*]] = load i16, ptr [[Q]], align 8
+; CHECK-NEXT: [[GEP_8:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 8
+; CHECK-NEXT: [[Q_8_VAL:%.*]] = load i64, ptr [[GEP_8]], align 8
+; CHECK-NEXT: call void @use(i16 [[Q_0_VAL]], i64 [[Q_8_VAL]])
+; CHECK-NEXT: ret void
+;
+entry:
+ br i1 %c, label %then, label %exit
+
+then:
+ store i16 123, ptr @f, align 8
+ br label %exit
+
+exit:
+ %l.0 = load i16, ptr %q, align 8
+ %gep.8 = getelementptr inbounds i8, ptr %q, i64 8
+ %l.1 = load i64, ptr %gep.8, align 8
+ call void @use(i16 %l.0, i64 %l.1)
+ ret void
+
+ uselistorder ptr %q, { 1, 0 }
+}
+
+; Same as @caller1/callee2, but with default uselist order.
+define i32 @caller2(i1 %c) {
+; CHECK-LABEL: define i32 @caller2(
+; CHECK-SAME: i1 [[C:%.*]]) {
+; CHECK-NEXT: entry:
+; CHECK-NEXT: call void @callee2(ptr noundef nonnull @f, i1 [[C]])
+; CHECK-NEXT: ret i32 0
+;
+entry:
+ call void @callee2(ptr noundef nonnull @f, i1 %c)
+ ret i32 0
+}
+
+define internal void @callee2(ptr nocapture noundef readonly %q, i1 %c) {
+; CHECK-LABEL: define internal void @callee2(
+; CHECK-SAME: ptr nocapture noundef readonly [[Q:%.*]], i1 [[C:%.*]]) {
+; CHECK-NEXT: entry:
+; CHECK-NEXT: br i1 [[C]], label [[THEN:%.*]], label [[EXIT:%.*]]
+; CHECK: then:
+; CHECK-NEXT: store i16 123, ptr @f, align 8
+; CHECK-NEXT: br label [[EXIT]]
+; CHECK: exit:
+; CHECK-NEXT: [[Q_0_VAL:%.*]] = load i16, ptr [[Q]], align 8
+; CHECK-NEXT: [[GEP_8:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 8
+; CHECK-NEXT: [[Q_8_VAL:%.*]] = load i64, ptr [[GEP_8]], align 8
+; CHECK-NEXT: call void @use(i16 [[Q_0_VAL]], i64 [[Q_8_VAL]])
+; CHECK-NEXT: ret void
+;
+entry:
+ br i1 %c, label %then, label %exit
+
+then:
+ store i16 123, ptr @f, align 8
+ br label %exit
+
+exit:
+ %l.0 = load i16, ptr %q, align 8
+ %gep.8 = getelementptr inbounds i8, ptr %q, i64 8
+ %l.1 = load i64, ptr %gep.8, align 8
+ call void @use(i16 %l.0, i64 %l.1)
+ ret void
+}
+
+declare void @use(i16, i64)
|
nikic
approved these changes
Mar 13, 2024
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Test case for llvm#84807, showing a mis-compile in ArgPromotion. (cherry picked from commit 31ffdb5)
The TranspBlocks set was used to cache aliasing decision for all processed loads in the parent loop. This is incorrect, because each load can access a different location, which means one load not being modified in a block doesn't translate to another load not being modified in the same block. All loads access the same underlying object, so we could perhaps use a location without size for all loads and retain the cache, but that would mean we loose precision. For now, just drop the cache. Fixes llvm#84807 PR: llvm#84835 (cherry picked from commit bba4a1d)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Backport 31ffdb5 bba4a1d
Requested by: @nikic