[llvm-branch-commits] [llvm] release/18.x: [InstCombine] Drop UB-implying attrs/metadata after speculating an instruction (#85542) (PR #85562)
via llvm-branch-commits
llvm-branch-commits at lists.llvm.org
Sat Mar 16 23:21:07 PDT 2024
https://github.com/llvmbot created https://github.com/llvm/llvm-project/pull/85562
Backport 252d01952c087cf0d141f7f281cf60efeb98be41
Requested by: @dtcxzyw
>From 7fd9979eb9cf9d93d76e12e85aa1ad847794ebbc Mon Sep 17 00:00:00 2001
From: Yingwei Zheng <dtcxzyw2333 at gmail.com>
Date: Sun, 17 Mar 2024 14:15:27 +0800
Subject: [PATCH] [InstCombine] Drop UB-implying attrs/metadata after
speculating an instruction (#85542)
When speculating an instruction in `InstCombinerImpl::FoldOpIntoSelect`,
the call may result in undefined behavior. This patch drops all
UB-implying attrs/metadata to fix this.
Fixes #85536.
(cherry picked from commit 252d01952c087cf0d141f7f281cf60efeb98be41)
---
.../InstCombine/InstructionCombining.cpp | 1 +
.../InstCombine/intrinsic-select.ll | 40 +++++++++++++++++++
2 files changed, 41 insertions(+)
diff --git a/llvm/lib/Transforms/InstCombine/InstructionCombining.cpp b/llvm/lib/Transforms/InstCombine/InstructionCombining.cpp
index 5d207dcfd18dd4..6f0cf9d9c8f187 100644
--- a/llvm/lib/Transforms/InstCombine/InstructionCombining.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstructionCombining.cpp
@@ -1455,6 +1455,7 @@ static Value *foldOperationIntoSelectOperand(Instruction &I, SelectInst *SI,
Value *NewOp, InstCombiner &IC) {
Instruction *Clone = I.clone();
Clone->replaceUsesOfWith(SI, NewOp);
+ Clone->dropUBImplyingAttrsAndMetadata();
IC.InsertNewInstBefore(Clone, SI->getIterator());
return Clone;
}
diff --git a/llvm/test/Transforms/InstCombine/intrinsic-select.ll b/llvm/test/Transforms/InstCombine/intrinsic-select.ll
index a203b28bcb82a8..f37226bbd5b09c 100644
--- a/llvm/test/Transforms/InstCombine/intrinsic-select.ll
+++ b/llvm/test/Transforms/InstCombine/intrinsic-select.ll
@@ -240,3 +240,43 @@ define i32 @vec_to_scalar_select_vector(<2 x i1> %b) {
%c = call i32 @llvm.vector.reduce.add.v2i32(<2 x i32> %s)
ret i32 %c
}
+
+define i8 @test_drop_noundef(i1 %cond, i8 %val) {
+; CHECK-LABEL: @test_drop_noundef(
+; CHECK-NEXT: entry:
+; CHECK-NEXT: [[TMP0:%.*]] = call i8 @llvm.smin.i8(i8 [[VAL:%.*]], i8 0)
+; CHECK-NEXT: [[RET:%.*]] = select i1 [[COND:%.*]], i8 -1, i8 [[TMP0]]
+; CHECK-NEXT: ret i8 [[RET]]
+;
+entry:
+ %sel = select i1 %cond, i8 -1, i8 %val
+ %ret = call noundef i8 @llvm.smin.i8(i8 %sel, i8 0)
+ ret i8 %ret
+}
+
+define i1 @pr85536(i32 %a) {
+; CHECK-LABEL: @pr85536(
+; CHECK-NEXT: entry:
+; CHECK-NEXT: [[CMP1:%.*]] = icmp ult i32 [[A:%.*]], 31
+; CHECK-NEXT: [[SHL1:%.*]] = shl nsw i32 -1, [[A]]
+; CHECK-NEXT: [[ZEXT:%.*]] = zext i32 [[SHL1]] to i64
+; CHECK-NEXT: [[SHL2:%.*]] = shl i64 [[ZEXT]], 48
+; CHECK-NEXT: [[SHR:%.*]] = ashr exact i64 [[SHL2]], 48
+; CHECK-NEXT: [[TMP0:%.*]] = call i64 @llvm.smin.i64(i64 [[SHR]], i64 0)
+; CHECK-NEXT: [[TMP1:%.*]] = and i64 [[TMP0]], 65535
+; CHECK-NEXT: [[RET1:%.*]] = icmp eq i64 [[TMP1]], 0
+; CHECK-NEXT: [[RET:%.*]] = select i1 [[CMP1]], i1 [[RET1]], i1 false
+; CHECK-NEXT: ret i1 [[RET]]
+;
+entry:
+ %cmp1 = icmp ugt i32 %a, 30
+ %shl1 = shl nsw i32 -1, %a
+ %zext = zext i32 %shl1 to i64
+ %shl2 = shl i64 %zext, 48
+ %shr = ashr exact i64 %shl2, 48
+ %sel = select i1 %cmp1, i64 -1, i64 %shr
+ %smin = call noundef i64 @llvm.smin.i64(i64 %sel, i64 0)
+ %masked = and i64 %smin, 65535
+ %ret = icmp eq i64 %masked, 0
+ ret i1 %ret
+}
More information about the llvm-branch-commits
mailing list