[llvm] 0012320 - [LLVM][ComplexDeinterleaving] Update identifyDeinterleave to support all forms of zero. (#200400)

[via llvm-commits]

Author: Paul Walker
Date: 2026-06-02T13:38:30+01:00
New Revision: 0012320e60cddc1f98a9835f7d4ce8552fe12fe7

URL: https://github.com/llvm/llvm-project/commit/0012320e60cddc1f98a9835f7d4ce8552fe12fe7
DIFF: https://github.com/llvm/llvm-project/commit/0012320e60cddc1f98a9835f7d4ce8552fe12fe7.diff

LOG: [LLVM][ComplexDeinterleaving] Update identifyDeinterleave to support all forms of zero. (#200400)

The original code misses ConstantFP based zeros.

Added: 
    

Modified: 
    llvm/lib/CodeGen/ComplexDeinterleavingPass.cpp
    llvm/test/CodeGen/AArch64/complex-deinterleaving-f32-add.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/CodeGen/ComplexDeinterleavingPass.cpp b/llvm/lib/CodeGen/ComplexDeinterleavingPass.cpp
index a0d3bb7c23302..90b06cb9f902a 100644
--- a/llvm/lib/CodeGen/ComplexDeinterleavingPass.cpp
+++ b/llvm/lib/CodeGen/ComplexDeinterleavingPass.cpp
@@ -2066,12 +2066,12 @@ ComplexDeinterleavingGraph::identifyDeinterleave(ComplexValues &Vals) {
   }
 
   Value *RealOp1 = RealShuffle->getOperand(1);
-  if (!isa<UndefValue>(RealOp1) && !isa<ConstantAggregateZero>(RealOp1)) {
+  if (!isa<UndefValue>(RealOp1) && !match(RealOp1, m_Zero())) {
     LLVM_DEBUG(dbgs() << " - RealOp1 is not undef or zero.\n");
     return nullptr;
   }
   Value *ImagOp1 = ImagShuffle->getOperand(1);
-  if (!isa<UndefValue>(ImagOp1) && !isa<ConstantAggregateZero>(ImagOp1)) {
+  if (!isa<UndefValue>(ImagOp1) && !match(ImagOp1, m_Zero())) {
     LLVM_DEBUG(dbgs() << " - ImagOp1 is not undef or zero.\n");
     return nullptr;
   }

diff  --git a/llvm/test/CodeGen/AArch64/complex-deinterleaving-f32-add.ll b/llvm/test/CodeGen/AArch64/complex-deinterleaving-f32-add.ll
index c4d0c9364f1be..9b682e5ec12e1 100644
--- a/llvm/test/CodeGen/AArch64/complex-deinterleaving-f32-add.ll
+++ b/llvm/test/CodeGen/AArch64/complex-deinterleaving-f32-add.ll
@@ -75,3 +75,20 @@ entry:
   %interleaved.vec = shufflevector <8 x float> %0, <8 x float> %1, <16 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11, i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
   ret <16 x float> %interleaved.vec
 }
+
+; A variant of complex_add_v2f32 that uses ConstantFP(0.0) rather than ConstantAggregateZero.
+define <2 x float> @complex_add_v2f32_no_CAZ(<2 x float> %a, <2 x float> %b) {
+; CHECK-LABEL: complex_add_v2f32_no_CAZ:
+; CHECK:       // %bb.0: // %entry
+; CHECK-NEXT:    fcadd v0.2s, v1.2s, v0.2s, #90
+; CHECK-NEXT:    ret
+entry:
+  %a.real = shufflevector <2 x float> %a, <2 x float> splat (float +0.0), <1 x i32> <i32 0>
+  %a.imag = shufflevector <2 x float> %a, <2 x float> splat (float +0.0), <1 x i32> <i32 1>
+  %b.real = shufflevector <2 x float> %b, <2 x float> splat (float +0.0), <1 x i32> <i32 0>
+  %b.imag = shufflevector <2 x float> %b, <2 x float> splat (float +0.0), <1 x i32> <i32 1>
+  %0 = fsub fast <1 x float> %b.real, %a.imag
+  %1 = fadd fast <1 x float> %b.imag, %a.real
+  %interleaved.vec = shufflevector <1 x float> %0, <1 x float> %1, <2 x i32> <i32 0, i32 1>
+  ret <2 x float> %interleaved.vec
+}