[llvm] ef40ae4 - [SelectionDAG] Fix incorrect fold condition in foldSetCCWithFunnelShift. (#137637)

[via llvm-commits]

Author: Rux124
Date: 2025-05-12T13:25:07+08:00
New Revision: ef40ae4f4e273828e81de08a8b5ab71ac3f59c79

URL: https://github.com/llvm/llvm-project/commit/ef40ae4f4e273828e81de08a8b5ab71ac3f59c79
DIFF: https://github.com/llvm/llvm-project/commit/ef40ae4f4e273828e81de08a8b5ab71ac3f59c79.diff

LOG: [SelectionDAG] Fix incorrect fold condition in foldSetCCWithFunnelShift. (#137637)

Proposed by
[2ed1598](https://github.com/llvm/llvm-project/commit/2ed15984b49a1af87be37ec8bd6ee3ab7f724767):

`fshl X, (or X, Y), C ==/!= 0 --> or (srl Y, BW-C), X ==/!= 0`

This transformation is valid when (C%Bitwidth) != 0 , as verified by
[Alive2](https://alive2.llvm.org/ce/z/TQYM-m).

Fixes #136746

Added: 
    

Modified: 
    llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
    llvm/test/CodeGen/AArch64/setcc-fsh.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
index ba34c72156228..856e9d813bcdd 100644
--- a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
@@ -4462,11 +4462,14 @@ static SDValue foldSetCCWithFunnelShift(EVT VT, SDValue N0, SDValue N1,
 
   unsigned BitWidth = N0.getScalarValueSizeInBits();
   auto *ShAmtC = isConstOrConstSplat(N0.getOperand(2));
-  if (!ShAmtC || ShAmtC->getAPIntValue().uge(BitWidth))
+  if (!ShAmtC)
+    return SDValue();
+
+  uint64_t ShAmt = ShAmtC->getAPIntValue().urem(BitWidth);
+  if (ShAmt == 0)
     return SDValue();
 
   // Canonicalize fshr as fshl to reduce pattern-matching.
-  unsigned ShAmt = ShAmtC->getZExtValue();
   if (N0.getOpcode() == ISD::FSHR)
     ShAmt = BitWidth - ShAmt;
 

diff  --git a/llvm/test/CodeGen/AArch64/setcc-fsh.ll b/llvm/test/CodeGen/AArch64/setcc-fsh.ll
index 08bfe282703ff..472723b89a8d6 100644
--- a/llvm/test/CodeGen/AArch64/setcc-fsh.ll
+++ b/llvm/test/CodeGen/AArch64/setcc-fsh.ll
@@ -248,3 +248,27 @@ define i1 @fshl_or_ne_2(i32 %x, i32 %y) {
   %r = icmp ne i32 %f, 2
   ret i1 %r
 }
+
+define i1 @fshr_0_or_eq_0(i16 %x, i16 %y) {
+; CHECK-LABEL: fshr_0_or_eq_0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w0, #0xffff
+; CHECK-NEXT:    cset w0, eq
+; CHECK-NEXT:    ret
+  %or = or i16 %x, %y
+  %f = call i16 @llvm.fshr.i16(i16 %or, i16 %x, i16 0)
+  %r = icmp eq i16 %f, 0
+  ret i1 %r
+}
+
+define i1 @fshr_32_or_eq_0(i16 %x, i16 %y) {
+; CHECK-LABEL: fshr_32_or_eq_0:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w0, #0xffff
+; CHECK-NEXT:    cset w0, eq
+; CHECK-NEXT:    ret
+  %or = or i16 %x, %y
+  %f = call i16 @llvm.fshr.i16(i16 %or, i16 %x, i16 32)
+  %r = icmp eq i16 %f, 0
+  ret i1 %r
+}