[llvm] 5126c38 - [CGP] Freeze condition when despeculating ctlz/cttz

[Nikita Popov via llvm-commits]

Author: Nikita Popov
Date: 2022-05-23T11:01:18+02:00
New Revision: 5126c38012c1d00729549ff4e4cb9d2be66ab9ab

URL: https://github.com/llvm/llvm-project/commit/5126c38012c1d00729549ff4e4cb9d2be66ab9ab
DIFF: https://github.com/llvm/llvm-project/commit/5126c38012c1d00729549ff4e4cb9d2be66ab9ab.diff

LOG: [CGP] Freeze condition when despeculating ctlz/cttz

Freeze the condition of the newly introduced conditional branch,
to avoid immediate undefined behavior if the input to ctlz/cttz
was originally poison.

Differential Revision: https://reviews.llvm.org/D125887

Added: 
    

Modified: 
    llvm/lib/CodeGen/CodeGenPrepare.cpp
    llvm/test/Transforms/CodeGenPrepare/X86/cttz-ctlz.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/CodeGen/CodeGenPrepare.cpp b/llvm/lib/CodeGen/CodeGenPrepare.cpp
index 9ed4f5bf43251..c71ca729391c8 100644
--- a/llvm/lib/CodeGen/CodeGenPrepare.cpp
+++ b/llvm/lib/CodeGen/CodeGenPrepare.cpp
@@ -2038,7 +2038,8 @@ static bool despeculateCountZeros(IntrinsicInst *CountZeros,
     return false;
 
   // Bail if the value is never zero.
-  if (llvm::isKnownNonZero(CountZeros->getOperand(0), *DL))
+  Value *Op = CountZeros->getOperand(0);
+  if (isKnownNonZero(Op, *DL))
     return false;
 
   // The intrinsic will be sunk behind a compare against zero and branch.
@@ -2059,7 +2060,10 @@ static bool despeculateCountZeros(IntrinsicInst *CountZeros,
   // Replace the unconditional branch that was created by the first split with
   // a compare against zero and a conditional branch.
   Value *Zero = Constant::getNullValue(Ty);
-  Value *Cmp = Builder.CreateICmpEQ(CountZeros->getOperand(0), Zero, "cmpz");
+  // Avoid introducing branch on poison.
+  if (!isGuaranteedNotToBeUndefOrPoison(Op))
+    Op = Builder.CreateFreeze(Op, Op->getName() + ".fr");
+  Value *Cmp = Builder.CreateICmpEQ(Op, Zero, "cmpz");
   Builder.CreateCondBr(Cmp, EndBlock, CallBlock);
   StartBlock->getTerminator()->eraseFromParent();
 

diff  --git a/llvm/test/Transforms/CodeGenPrepare/X86/cttz-ctlz.ll b/llvm/test/Transforms/CodeGenPrepare/X86/cttz-ctlz.ll
index 7b1246e83ea57..53156ab7324ec 100644
--- a/llvm/test/Transforms/CodeGenPrepare/X86/cttz-ctlz.ll
+++ b/llvm/test/Transforms/CodeGenPrepare/X86/cttz-ctlz.ll
@@ -13,7 +13,8 @@ target datalayout = "e-n32:64"
 define i64 @cttz(i64 %A) {
 ; SLOW-LABEL: @cttz(
 ; SLOW-NEXT:  entry:
-; SLOW-NEXT:    [[CMPZ:%.*]] = icmp eq i64 [[A:%.*]], 0
+; SLOW-NEXT:    [[A_FR:%.*]] = freeze i64 [[A:%.*]]
+; SLOW-NEXT:    [[CMPZ:%.*]] = icmp eq i64 [[A_FR]], 0
 ; SLOW-NEXT:    br i1 [[CMPZ]], label [[COND_END:%.*]], label [[COND_FALSE:%.*]]
 ; SLOW:       cond.false:
 ; SLOW-NEXT:    [[Z:%.*]] = call i64 @llvm.cttz.i64(i64 [[A]], i1 true)
@@ -29,7 +30,8 @@ define i64 @cttz(i64 %A) {
 ;
 ; FAST_LZ-LABEL: @cttz(
 ; FAST_LZ-NEXT:  entry:
-; FAST_LZ-NEXT:    [[CMPZ:%.*]] = icmp eq i64 [[A:%.*]], 0
+; FAST_LZ-NEXT:    [[A_FR:%.*]] = freeze i64 [[A:%.*]]
+; FAST_LZ-NEXT:    [[CMPZ:%.*]] = icmp eq i64 [[A_FR]], 0
 ; FAST_LZ-NEXT:    br i1 [[CMPZ]], label [[COND_END:%.*]], label [[COND_FALSE:%.*]]
 ; FAST_LZ:       cond.false:
 ; FAST_LZ-NEXT:    [[Z:%.*]] = call i64 @llvm.cttz.i64(i64 [[A]], i1 true)
@@ -46,7 +48,8 @@ entry:
 define i64 @ctlz(i64 %A) {
 ; SLOW-LABEL: @ctlz(
 ; SLOW-NEXT:  entry:
-; SLOW-NEXT:    [[CMPZ:%.*]] = icmp eq i64 [[A:%.*]], 0
+; SLOW-NEXT:    [[A_FR:%.*]] = freeze i64 [[A:%.*]]
+; SLOW-NEXT:    [[CMPZ:%.*]] = icmp eq i64 [[A_FR]], 0
 ; SLOW-NEXT:    br i1 [[CMPZ]], label [[COND_END:%.*]], label [[COND_FALSE:%.*]]
 ; SLOW:       cond.false:
 ; SLOW-NEXT:    [[Z:%.*]] = call i64 @llvm.ctlz.i64(i64 [[A]], i1 true)
@@ -57,7 +60,8 @@ define i64 @ctlz(i64 %A) {
 ;
 ; FAST_TZ-LABEL: @ctlz(
 ; FAST_TZ-NEXT:  entry:
-; FAST_TZ-NEXT:    [[CMPZ:%.*]] = icmp eq i64 [[A:%.*]], 0
+; FAST_TZ-NEXT:    [[A_FR:%.*]] = freeze i64 [[A:%.*]]
+; FAST_TZ-NEXT:    [[CMPZ:%.*]] = icmp eq i64 [[A_FR]], 0
 ; FAST_TZ-NEXT:    br i1 [[CMPZ]], label [[COND_END:%.*]], label [[COND_FALSE:%.*]]
 ; FAST_TZ:       cond.false:
 ; FAST_TZ-NEXT:    [[Z:%.*]] = call i64 @llvm.ctlz.i64(i64 [[A]], i1 true)