[llvm] 98f1e4a - [DAG] SelectionDAG::canCreateUndefOrPoison - AVGFLOOR/AVGCEIL don't create undef/poison (#157056)

[via llvm-commits]

Author: Simon Pilgrim
Date: 2025-09-05T11:11:15Z
New Revision: 98f1e4a57f0573fc9651423dd068146934499d20

URL: https://github.com/llvm/llvm-project/commit/98f1e4a57f0573fc9651423dd068146934499d20
DIFF: https://github.com/llvm/llvm-project/commit/98f1e4a57f0573fc9651423dd068146934499d20.diff

LOG: [DAG] SelectionDAG::canCreateUndefOrPoison - AVGFLOOR/AVGCEIL don't create undef/poison (#157056)

AVGFLOORS: https://alive2.llvm.org/ce/z/6TdoQ_
AVGFLOORU: https://alive2.llvm.org/ce/z/4pfi4i
AVGCEILS: https://alive2.llvm.org/ce/z/nWu8WM
AVGCEILU: https://alive2.llvm.org/ce/z/CGvWiA

Fixes #147696

Added: 
    

Modified: 
    llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
    llvm/test/CodeGen/AArch64/freeze.ll
    llvm/test/CodeGen/X86/avg-mask.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index 9668d253d52ae..bcf25958d0982 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -5672,6 +5672,10 @@ bool SelectionDAG::canCreateUndefOrPoison(SDValue Op, const APInt &DemandedElts,
   case ISD::USUBSAT:
   case ISD::MULHU:
   case ISD::MULHS:
+  case ISD::AVGFLOORS:
+  case ISD::AVGFLOORU:
+  case ISD::AVGCEILS:
+  case ISD::AVGCEILU:
   case ISD::ABDU:
   case ISD::ABDS:
   case ISD::SMIN:

diff  --git a/llvm/test/CodeGen/AArch64/freeze.ll b/llvm/test/CodeGen/AArch64/freeze.ll
index 2a33a4c061dce..7a9d6b7e52457 100644
--- a/llvm/test/CodeGen/AArch64/freeze.ll
+++ b/llvm/test/CodeGen/AArch64/freeze.ll
@@ -430,16 +430,13 @@ define <8 x i16> @freeze_abds(<8 x i16> %a, <8 x i16> %b) {
   ret <8 x i16> %r
 }
 
-; TODO: Unnecessary final and
 define <8 x i16> @freeze_uhadd(<8 x i16> %a0, <8 x i16> %a1) {
 ; CHECK-LABEL: freeze_uhadd:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    movi v2.8h, #15
 ; CHECK-NEXT:    and v0.16b, v0.16b, v2.16b
 ; CHECK-NEXT:    and v1.16b, v1.16b, v2.16b
-; CHECK-NEXT:    movi v2.8h, #31
 ; CHECK-NEXT:    uhadd v0.8h, v0.8h, v1.8h
-; CHECK-NEXT:    and v0.16b, v0.16b, v2.16b
 ; CHECK-NEXT:    ret
   %m0 = and <8 x i16> %a0, splat (i16 15)
   %m1 = and <8 x i16> %a1, splat (i16 15)
@@ -449,16 +446,13 @@ define <8 x i16> @freeze_uhadd(<8 x i16> %a0, <8 x i16> %a1) {
   ret <8 x i16> %masked
 }
 
-; TODO: Unnecessary final and
 define <8 x i16> @freeze_urhadd(<8 x i16> %a0, <8 x i16> %a1) {
 ; CHECK-LABEL: freeze_urhadd:
 ; CHECK:       // %bb.0:
 ; CHECK-NEXT:    movi v2.8h, #15
 ; CHECK-NEXT:    and v0.16b, v0.16b, v2.16b
 ; CHECK-NEXT:    and v1.16b, v1.16b, v2.16b
-; CHECK-NEXT:    movi v2.8h, #31
 ; CHECK-NEXT:    urhadd v0.8h, v0.8h, v1.8h
-; CHECK-NEXT:    and v0.16b, v0.16b, v2.16b
 ; CHECK-NEXT:    ret
   %m0 = and <8 x i16> %a0, splat (i16 15)
   %m1 = and <8 x i16> %a1, splat (i16 15)

diff  --git a/llvm/test/CodeGen/X86/avg-mask.ll b/llvm/test/CodeGen/X86/avg-mask.ll
index e8866393e8b62..b148cd3d42df6 100644
--- a/llvm/test/CodeGen/X86/avg-mask.ll
+++ b/llvm/test/CodeGen/X86/avg-mask.ll
@@ -177,11 +177,11 @@ define <64 x i8> @avg_v64i8_maskz(<64 x i8> %a, <64 x i8> %b, i64 %mask) nounwin
 ; AVX512F-NEXT:    shrq $32, %rdi
 ; AVX512F-NEXT:    shrq $48, %rax
 ; AVX512F-NEXT:    shrl $16, %ecx
-; AVX512F-NEXT:    vpavgb %ymm1, %ymm0, %ymm2
-; AVX512F-NEXT:    vextracti64x4 $1, %zmm1, %ymm1
-; AVX512F-NEXT:    vextracti64x4 $1, %zmm0, %ymm0
+; AVX512F-NEXT:    vextracti64x4 $1, %zmm1, %ymm2
+; AVX512F-NEXT:    vextracti64x4 $1, %zmm0, %ymm3
+; AVX512F-NEXT:    vpavgb %ymm2, %ymm3, %ymm2
 ; AVX512F-NEXT:    vpavgb %ymm1, %ymm0, %ymm0
-; AVX512F-NEXT:    vinserti64x4 $1, %ymm0, %zmm2, %zmm0
+; AVX512F-NEXT:    vinserti64x4 $1, %ymm2, %zmm0, %zmm0
 ; AVX512F-NEXT:    kmovw %ecx, %k2
 ; AVX512F-NEXT:    kmovw %eax, %k3
 ; AVX512F-NEXT:    kmovw %edi, %k4
@@ -364,11 +364,11 @@ define <32 x i16> @avg_v32i16_maskz(<32 x i16> %a, <32 x i16> %b, i32 %mask) nou
 ; AVX512F:       # %bb.0:
 ; AVX512F-NEXT:    kmovw %edi, %k1
 ; AVX512F-NEXT:    shrl $16, %edi
-; AVX512F-NEXT:    vpavgw %ymm1, %ymm0, %ymm2
-; AVX512F-NEXT:    vextracti64x4 $1, %zmm1, %ymm1
-; AVX512F-NEXT:    vextracti64x4 $1, %zmm0, %ymm0
+; AVX512F-NEXT:    vextracti64x4 $1, %zmm1, %ymm2
+; AVX512F-NEXT:    vextracti64x4 $1, %zmm0, %ymm3
+; AVX512F-NEXT:    vpavgw %ymm2, %ymm3, %ymm2
 ; AVX512F-NEXT:    vpavgw %ymm1, %ymm0, %ymm0
-; AVX512F-NEXT:    vinserti64x4 $1, %ymm0, %zmm2, %zmm0
+; AVX512F-NEXT:    vinserti64x4 $1, %ymm2, %zmm0, %zmm0
 ; AVX512F-NEXT:    kmovw %edi, %k2
 ; AVX512F-NEXT:    vpternlogd {{.*#+}} zmm1 {%k1} {z} = -1
 ; AVX512F-NEXT:    vpmovdw %zmm1, %ymm1