Skip to content
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

[InstCombine] Fix unexpected overwriting in foldSelectWithSRem #89539

Merged
merged 2 commits into from
Apr 21, 2024

Conversation

dtcxzyw
Copy link
Member

@dtcxzyw dtcxzyw commented Apr 21, 2024

Fixes #89516

@llvmbot
Copy link
Member

llvmbot commented Apr 21, 2024

@llvm/pr-subscribers-llvm-transforms

Author: Yingwei Zheng (dtcxzyw)

Changes

Fixes #89516


Full diff: https://github.com/llvm/llvm-project/pull/89539.diff

2 Files Affected:

  • (modified) llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp (+1-1)
  • (modified) llvm/test/Transforms/InstCombine/select-divrem.ll (+17)
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp b/llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp
index 0262af28068be7..73600206a55c14 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp
@@ -2720,7 +2720,7 @@ static Instruction *foldSelectWithSRem(SelectInst &SI, InstCombinerImpl &IC,
   // %cnd = icmp slt i32 %rem, 0
   // %add = add i32 %rem, %n
   // %sel = select i1 %cnd, i32 %add, i32 %rem
-  if (match(TrueVal, m_Add(m_Value(RemRes), m_Value(Remainder))) &&
+  if (match(TrueVal, m_Add(m_Specific(RemRes), m_Value(Remainder))) &&
       match(RemRes, m_SRem(m_Value(Op), m_Specific(Remainder))) &&
       IC.isKnownToBeAPowerOfTwo(Remainder, /*OrZero*/ true) &&
       FalseVal == RemRes)
diff --git a/llvm/test/Transforms/InstCombine/select-divrem.ll b/llvm/test/Transforms/InstCombine/select-divrem.ll
index f007c53359ca5a..e0c460c37451db 100644
--- a/llvm/test/Transforms/InstCombine/select-divrem.ll
+++ b/llvm/test/Transforms/InstCombine/select-divrem.ll
@@ -343,3 +343,20 @@ define i32 @rem_euclid_pow2_false_arm_folded(i32 %n) {
   %res = select i1 %nonneg, i32 %rem, i32 1
   ret i32 %res
 }
+
+define i8 @pr89516(i8 %n, i8 %x) {
+; CHECK-LABEL: @pr89516(
+; CHECK-NEXT:    [[COND:%.*]] = icmp slt i8 [[X:%.*]], 0
+; CHECK-NEXT:    [[POW2:%.*]] = shl nuw i8 1, [[N:%.*]]
+; CHECK-NEXT:    [[SREM:%.*]] = srem i8 1, [[POW2]]
+; CHECK-NEXT:    [[ADD:%.*]] = select i1 [[COND]], i8 [[POW2]], i8 0
+; CHECK-NEXT:    [[RES:%.*]] = add nuw i8 [[SREM]], [[ADD]]
+; CHECK-NEXT:    ret i8 [[RES]]
+;
+  %cond = icmp slt i8 %x, 0
+  %pow2 = shl nuw i8 1, %n
+  %srem = srem i8 1, %pow2
+  %add = add nuw i8 %srem, %pow2
+  %res = select i1 %cond, i8 %add, i8 %srem
+  ret i8 %res
+}

Copy link
Contributor

@antoniofrighetto antoniofrighetto left a comment

Choose a reason for hiding this comment

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

LGTM, thanks for fixing it!

@dtcxzyw dtcxzyw merged commit 6309440 into llvm:main Apr 21, 2024
6 checks passed
@dtcxzyw dtcxzyw deleted the fix-pr89516 branch April 21, 2024 14:41
@dtcxzyw dtcxzyw added this to the LLVM 18.X Release milestone Apr 21, 2024
@dtcxzyw
Copy link
Member Author

dtcxzyw commented Apr 21, 2024

/cherry-pick 6309440

llvmbot pushed a commit to llvmbot/llvm-project that referenced this pull request Apr 21, 2024
@llvmbot
Copy link
Member

llvmbot commented Apr 21, 2024

/pull-request #89546

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
Development

Successfully merging this pull request may close these issues.

miscompile of vanilla integer code by InstCombine
4 participants