Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly
Appearance settings

[SelectionDAG] Deal with POISON for INSERT_VECTOR_ELT/INSERT_SUBVECTOR (part 2) #143103

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 2 commits into
base: users/bjope/insertundef_2
Choose a base branch
Loading
from
Open

[SelectionDAG] Deal with POISON for INSERT_VECTOR_ELT/INSERT_SUBVECTOR (part 2) #143103

wants to merge 2 commits into from

Conversation

bjope
Copy link
Collaborator

@bjope bjope commented Jun 6, 2025

Add support in isGuaranteedNotToBeUndefOrPoison and
SimplifyDemandedVectorElts to avoid regressions seen
after a previous commit fixing #141034.

@llvmbot
Copy link
Member

llvmbot commented Jun 6, 2025 •
edited
Loading

@llvm/pr-subscribers-backend-amdgpu
@llvm/pr-subscribers-llvm-selectiondag

@llvm/pr-subscribers-backend-aarch64

Author: Björn Pettersson (bjope)

Changes

Add support in isGuaranteedNotToBeUndefOrPoison and
SimplifyDemandedVectorElts to avoid regressions seen
after a previous commit fixing #141034.

Patch is 158.07 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/143103.diff

14 Files Affected:

  • (modified) llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp (+51-13)
  • (modified) llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp (+109-10)
  • (modified) llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp (+4)
  • (modified) llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll (+67)
  • (modified) llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll (+75-6)
  • (modified) llvm/test/CodeGen/AArch64/sve-fixed-length-int-vselect.ll (+100-8)
  • (modified) llvm/test/CodeGen/AArch64/sve-fixed-length-masked-gather.ll (+5-1)
  • (modified) llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-vselect.ll (+21)
  • (modified) llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-vselect.ll (+28)
  • (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfw-web-simplification.ll (+60-30)
  • (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vw-web-simplification.ll (+36-19)
  • (modified) llvm/test/CodeGen/RISCV/rvv/vector-deinterleave.ll (+36)
  • (modified) llvm/test/CodeGen/X86/pr62286.ll (+8-6)
  • (modified) llvm/test/CodeGen/X86/vector-interleaved-load-i16-stride-4.ll (+693-597)
diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index aba3c0f80a024..a52edca64dbce 100644
--- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -22905,6 +22905,7 @@ SDValue DAGCombiner::visitINSERT_VECTOR_ELT(SDNode *N) {
   auto *IndexC = dyn_cast<ConstantSDNode>(EltNo);
 
   // Insert into out-of-bounds element is undefined.
+  // Code below relies on that we handle this special case early.
   if (IndexC && VT.isFixedLengthVector() &&
       IndexC->getZExtValue() >= VT.getVectorNumElements())
     return DAG.getUNDEF(VT);
@@ -22915,14 +22916,29 @@ SDValue DAGCombiner::visitINSERT_VECTOR_ELT(SDNode *N) {
       InVec == InVal.getOperand(0) && EltNo == InVal.getOperand(1))
     return InVec;
 
-  if (!IndexC) {
-    // If this is variable insert to undef vector, it might be better to splat:
-    // inselt undef, InVal, EltNo --> build_vector < InVal, InVal, ... >
-    if (InVec.isUndef() && TLI.shouldSplatInsEltVarIndex(VT))
-      return DAG.getSplat(VT, DL, InVal);
-    return SDValue();
+  // If this is variable insert to undef vector, it might be better to splat:
+  // inselt undef, InVal, EltNo --> build_vector < InVal, InVal, ... >
+  if (!IndexC && InVec.isUndef() && TLI.shouldSplatInsEltVarIndex(VT))
+    return DAG.getSplat(VT, DL, InVal);
+
+  // Try to drop insert of UNDEF/POISON elements. This is also done in getNode,
+  // but we also do it as a DAG combine since for example simplifications into
+  // SPLAT_VECTOR/BUILD_VECTOR may turn poison elements into undef/zero etc, and
+  // then suddenly the InVec is guaranteed to not be poison.
+  if (InVal.isUndef()) {
+    if (IndexC && VT.isFixedLengthVector()) {
+      APInt EltMask = APInt::getOneBitSet(VT.getVectorNumElements(),
+                                          IndexC->getZExtValue());
+      if (DAG.isGuaranteedNotToBePoison(InVec, EltMask))
+        return InVec;
+    } else if (DAG.isGuaranteedNotToBePoison(InVec)) {
+      return InVec;
+    }
   }
 
+  if (!IndexC)
+    return SDValue();
+
   if (VT.isScalableVector())
     return SDValue();
 
@@ -27355,18 +27371,40 @@ SDValue DAGCombiner::visitINSERT_SUBVECTOR(SDNode *N) {
   SDValue N2 = N->getOperand(2);
   uint64_t InsIdx = N->getConstantOperandVal(2);
 
-  // If inserting an UNDEF, just return the original vector.
-  if (N1.isUndef())
-    return N0;
+  // If inserting an UNDEF, just return the original vector (unless it makes the
+  // result more poisonous).
+  if (N1.isUndef()){
+    if (VT.isFixedLengthVector()) {
+      unsigned SubVecNumElts = N1.getValueType().getVectorNumElements();
+      APInt EltMask = APInt::getBitsSet(VT.getVectorNumElements(),
+                                        InsIdx, InsIdx + SubVecNumElts);
+      if (DAG.isGuaranteedNotToBePoison(N0, EltMask))
+        return N0;
+    } else if (DAG.isGuaranteedNotToBePoison(N0))
+      return N0;
+  }
 
-  // If this is an insert of an extracted vector into an undef vector, we can
-  // just use the input to the extract if the types match, and can simplify
+  // If this is an insert of an extracted vector into an undef/poison vector, we
+  // can just use the input to the extract if the types match, and can simplify
   // in some cases even if they don't.
   if (N0.isUndef() && N1.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
       N1.getOperand(1) == N2) {
+    EVT N1VT = N1.getValueType();
     EVT SrcVT = N1.getOperand(0).getValueType();
-    if (SrcVT == VT)
-      return N1.getOperand(0);
+    if (SrcVT == VT) {
+      // Need to ensure that result isn't more poisonous if skipping both the
+      // extract+insert.
+      if (N0.getOpcode() == ISD::POISON)
+        return N1.getOperand(0);
+      if (VT.isFixedLengthVector() && N1VT.isFixedLengthVector()) {
+        unsigned SubVecNumElts = N1VT.getVectorNumElements();
+        APInt EltMask = APInt::getBitsSet(VT.getVectorNumElements(),
+                                          InsIdx, InsIdx + SubVecNumElts);
+        if (DAG.isGuaranteedNotToBePoison(N1.getOperand(0), ~EltMask))
+          return N1.getOperand(0);
+      } else if (DAG.isGuaranteedNotToBePoison(N1.getOperand(0)))
+        return N1.getOperand(0);
+    }
     // TODO: To remove the zero check, need to adjust the offset to
     // a multiple of the new src type.
     if (isNullConstant(N2)) {
diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index 279c7daf71c33..fadfd35489e54 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -5454,6 +5454,60 @@ bool SelectionDAG::isGuaranteedNotToBeUndefOrPoison(SDValue Op,
     }
     return true;
 
+  case ISD::INSERT_SUBVECTOR: {
+    if (Op.getValueType().isScalableVector())
+      break;
+    SDValue Src = Op.getOperand(0);
+    SDValue Sub = Op.getOperand(1);
+    uint64_t Idx = Op.getConstantOperandVal(2);
+    unsigned NumSubElts = Sub.getValueType().getVectorNumElements();
+    APInt DemandedSubElts = DemandedElts.extractBits(NumSubElts, Idx);
+    APInt DemandedSrcElts = DemandedElts;
+    DemandedSrcElts.clearBits(Idx, Idx + NumSubElts);
+
+    if (!!DemandedSubElts &&
+        !isGuaranteedNotToBeUndefOrPoison(Sub, DemandedSubElts,
+                                          PoisonOnly, Depth + 1))
+        return false;
+    if (!!DemandedSrcElts &&
+        !isGuaranteedNotToBeUndefOrPoison(Src, DemandedSrcElts,
+                                          PoisonOnly, Depth + 1))
+      return false;
+    return true;
+  }
+
+  case ISD::INSERT_VECTOR_ELT: {
+    SDValue InVec = Op.getOperand(0);
+    SDValue InVal = Op.getOperand(1);
+    SDValue EltNo = Op.getOperand(2);
+    EVT VT = InVec.getValueType();
+    auto *IndexC = dyn_cast<ConstantSDNode>(EltNo);
+    if (IndexC && VT.isFixedLengthVector() &&
+        IndexC->getZExtValue() < VT.getVectorNumElements()) {
+      if (DemandedElts[IndexC->getZExtValue()] &&
+          !isGuaranteedNotToBeUndefOrPoison(InVal, PoisonOnly, Depth + 1))
+        return false;
+      APInt InVecDemandedElts = DemandedElts;
+      InVecDemandedElts.clearBit(IndexC->getZExtValue());
+      if (!!InVecDemandedElts &&
+          !isGuaranteedNotToBeUndefOrPoison(InVec, InVecDemandedElts,
+                                            PoisonOnly, Depth + 1))
+        return false;
+      return true;
+    }
+    break;
+  }
+
+  case ISD::SCALAR_TO_VECTOR:
+    // If only demanding upper (undef) elements.
+    if (DemandedElts.ugt(1))
+      return PoisonOnly;
+    // If only demanding element 0, or only considering poison.
+    if (PoisonOnly || DemandedElts == 0)
+      return isGuaranteedNotToBeUndefOrPoison(Op.getOperand(0), PoisonOnly,
+                                              Depth + 1);
+    return false;
+
   case ISD::SPLAT_VECTOR:
     return isGuaranteedNotToBeUndefOrPoison(Op.getOperand(0), PoisonOnly,
                                             Depth + 1);
@@ -7900,23 +7954,42 @@ SDValue SelectionDAG::getNode(unsigned Opcode, const SDLoc &DL, EVT VT,
     // INSERT_VECTOR_ELT into out-of-bounds element is an UNDEF, except
     // for scalable vectors where we will generate appropriate code to
     // deal with out-of-bounds cases correctly.
-    if (N3C && N1.getValueType().isFixedLengthVector() &&
-        N3C->getZExtValue() >= N1.getValueType().getVectorNumElements())
+    if (N3C && VT.isFixedLengthVector() &&
+        N3C->getZExtValue() >= VT.getVectorNumElements())
       return getUNDEF(VT);
 
     // Undefined index can be assumed out-of-bounds, so that's UNDEF too.
     if (N3.isUndef())
       return getUNDEF(VT);
 
-    // If the inserted element is an UNDEF, just use the input vector.
-    if (N2.isUndef())
+    // If inserting poison, just use the input vector.
+    if (N2.getOpcode() == ISD::POISON)
       return N1;
 
+    // Inserting undef into undef/poison is still undef.
+    if (N2.getOpcode() == ISD::UNDEF && N1.isUndef())
+      return getUNDEF(VT);
+
+    // If the inserted element is an UNDEF, just use the input vector.
+    // But not if skipping the insert could make the result more poisonous.
+    if (N2.isUndef()) {
+      if (N3C && VT.isFixedLengthVector()) {
+        APInt EltMask = APInt::getOneBitSet(VT.getVectorNumElements(),
+                                            N3C->getZExtValue());
+        if (isGuaranteedNotToBePoison(N1, EltMask))
+          return N1;
+      } else if (isGuaranteedNotToBePoison(N1))
+        return N1;
+    }
     break;
   }
   case ISD::INSERT_SUBVECTOR: {
-    // Inserting undef into undef is still undef.
-    if (N1.isUndef() && N2.isUndef())
+    // If inserting poison, just use the input vector,
+    if (N2.getOpcode() == ISD::POISON)
+      return N1;
+
+    // Inserting undef into undef/poison is still undef.
+    if (N2.getOpcode() == ISD::UNDEF && N1.isUndef())
       return getUNDEF(VT);
 
     EVT N2VT = N2.getValueType();
@@ -7945,11 +8018,37 @@ SDValue SelectionDAG::getNode(unsigned Opcode, const SDLoc &DL, EVT VT,
     if (VT == N2VT)
       return N2;
 
-    // If this is an insert of an extracted vector into an undef vector, we
-    // can just use the input to the extract.
+    // If this is an insert of an extracted vector into an undef/poison vector,
+    // we can just use the input to the extract. But not if skipping the
+    // extract+insert could make the result more poisonous.
     if (N1.isUndef() && N2.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
-        N2.getOperand(1) == N3 && N2.getOperand(0).getValueType() == VT)
-      return N2.getOperand(0);
+        N2.getOperand(1) == N3 && N2.getOperand(0).getValueType() == VT) {
+      if (N1.getOpcode() == ISD::POISON)
+        return N2.getOperand(0);
+      if (VT.isFixedLengthVector() && N2VT.isFixedLengthVector()) {
+        unsigned LoBit = N3->getAsZExtVal();
+        unsigned HiBit = LoBit + N2VT.getVectorNumElements();
+        APInt EltMask = APInt::getBitsSet(VT.getVectorNumElements(),
+                                          LoBit, HiBit);
+        if (isGuaranteedNotToBePoison(N2.getOperand(0), ~EltMask))
+          return N2.getOperand(0);
+      } else if (isGuaranteedNotToBePoison(N2.getOperand(0)))
+        return N2.getOperand(0);
+    }
+
+    // If the inserted subvector is UNDEF, just use the input vector.
+    // But not if skipping the insert could make the result more poisonous.
+    if (N2.isUndef()) {
+      if (VT.isFixedLengthVector()) {
+        unsigned LoBit = N3->getAsZExtVal();
+        unsigned HiBit = LoBit + N2VT.getVectorNumElements();
+        APInt EltMask = APInt::getBitsSet(VT.getVectorNumElements(),
+                                          LoBit, HiBit);
+        if (isGuaranteedNotToBePoison(N1, EltMask))
+          return N1;
+      } else if (isGuaranteedNotToBePoison(N1))
+        return N1;
+    }
     break;
   }
   case ISD::BITCAST:
diff --git a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
index e8e820ac1f695..643ec26bba3f7 100644
--- a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
@@ -3359,6 +3359,10 @@ bool TargetLowering::SimplifyDemandedVectorElts(
     APInt DemandedSrcElts = DemandedElts;
     DemandedSrcElts.clearBits(Idx, Idx + NumSubElts);
 
+    // If none of the sub operand elements are demanded, bypass the insert.
+    if (!DemandedSubElts)
+      return TLO.CombineTo(Op, Src);
+
     APInt SubUndef, SubZero;
     if (SimplifyDemandedVectorElts(Sub, DemandedSubElts, SubUndef, SubZero, TLO,
                                    Depth + 1))
diff --git a/llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll b/llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll
index 94074d1689f6a..2c44f56316801 100644
--- a/llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll
+++ b/llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll
@@ -66,6 +66,36 @@ define <16 x i8> @test_insert_v16i8_insert_2_undef_base(i8 %a) {
   %v.6 = insertelement <16 x i8> %v.4, i8 %a, i32 6
   %v.7 = insertelement <16 x i8> %v.6, i8 %a, i32 7
   %v.8 = insertelement <16 x i8> %v.7, i8 %a, i32 8
+  %v.10 = insertelement <16 x i8> %v.8, i8 %a, i32 10
+  %v.11 = insertelement <16 x i8> %v.10, i8 %a, i32 11
+  %v.12 = insertelement <16 x i8> %v.11, i8 %a, i32 12
+  %v.13 = insertelement <16 x i8> %v.12, i8 %a, i32 13
+  %v.14 = insertelement <16 x i8> %v.13, i8 %a, i32 14
+  %v.15 = insertelement <16 x i8> %v.14, i8 %a, i32 15
+  ret <16 x i8> %v.15
+}
+
+; Similar to above, but we leave element 8 as undef. One interesting part with
+; this test case is that %a may be poison, so simply inserting %a also at
+; index 8 would make the result vector more poisonous.
+define <16 x i8> @test_insert_v16i8_insert_2_undef_base_skip8(i32 %a0) {
+; CHECK-LABEL: test_insert_v16i8_insert_2_undef_base_skip8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    lsr w8, w0, #5
+; CHECK-NEXT:    dup.16b v0, w8
+; CHECK-NEXT:    mov.b v0[5], wzr
+; CHECK-NEXT:    mov.b v0[8], w8
+; CHECK-NEXT:    mov.b v0[9], wzr
+; CHECK-NEXT:    ret
+  %a1 = lshr exact i32 %a0, 5
+  %a = trunc i32 %a1 to i8
+  %v.0 = insertelement <16 x i8> <i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef>  , i8 %a, i32 0
+  %v.1 = insertelement <16 x i8> %v.0, i8 %a, i32 1
+  %v.2 = insertelement <16 x i8> %v.1, i8 %a, i32 2
+  %v.3 = insertelement <16 x i8> %v.2, i8 %a, i32 3
+  %v.4 = insertelement <16 x i8> %v.3, i8 %a, i32 4
+  %v.6 = insertelement <16 x i8> %v.4, i8 %a, i32 6
+  %v.7 = insertelement <16 x i8> %v.6, i8 %a, i32 7
   %v.10 = insertelement <16 x i8> %v.7, i8 %a, i32 10
   %v.11 = insertelement <16 x i8> %v.10, i8 %a, i32 11
   %v.12 = insertelement <16 x i8> %v.11, i8 %a, i32 12
@@ -94,6 +124,43 @@ define <16 x i8> @test_insert_v16i8_insert_2_undef_base_different_valeus(i8 %a,
   %v.6 = insertelement <16 x i8> %v.4, i8 %a, i32 6
   %v.7 = insertelement <16 x i8> %v.6, i8 %b, i32 7
   %v.8 = insertelement <16 x i8> %v.7, i8 %a, i32 8
+  %v.10 = insertelement <16 x i8> %v.8, i8 %a, i32 10
+  %v.11 = insertelement <16 x i8> %v.10, i8 %a, i32 11
+  %v.12 = insertelement <16 x i8> %v.11, i8 %b, i32 12
+  %v.13 = insertelement <16 x i8> %v.12, i8 %a, i32 13
+  %v.14 = insertelement <16 x i8> %v.13, i8 %a, i32 14
+  %v.15 = insertelement <16 x i8> %v.14, i8 %b, i32 15
+  ret <16 x i8> %v.15
+}
+
+; Similar to above, but we leave element 8 as undef. One interesting part with
+; this test case is that %a and %b may be poison, so simply inserting %a or %b
+; at index 8 would make the result vector more poisonous.
+define <16 x i8> @test_insert_v16i8_insert_2_undef_base_different_valeus_skip8(i32 %a0, i32 %b0) {
+; CHECK-LABEL: test_insert_v16i8_insert_2_undef_base_different_valeus_skip8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    lsr w8, w0, #5
+; CHECK-NEXT:    dup.16b v0, w8
+; CHECK-NEXT:    lsr w8, w1, #5
+; CHECK-NEXT:    mov.b v0[2], w8
+; CHECK-NEXT:    mov.b v0[5], wzr
+; CHECK-NEXT:    mov.b v0[7], w8
+; CHECK-NEXT:    mov.b v0[8], w8
+; CHECK-NEXT:    mov.b v0[9], wzr
+; CHECK-NEXT:    mov.b v0[12], w8
+; CHECK-NEXT:    mov.b v0[15], w8
+; CHECK-NEXT:    ret
+  %a1 = lshr exact i32 %a0, 5
+  %a = trunc i32 %a1 to i8
+  %b1 = lshr exact i32 %b0, 5
+  %b = trunc i32 %b1 to i8
+  %v.0 = insertelement <16 x i8> <i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef>  , i8 %a, i32 0
+  %v.1 = insertelement <16 x i8> %v.0, i8 %a, i32 1
+  %v.2 = insertelement <16 x i8> %v.1, i8 %b, i32 2
+  %v.3 = insertelement <16 x i8> %v.2, i8 %a, i32 3
+  %v.4 = insertelement <16 x i8> %v.3, i8 %a, i32 4
+  %v.6 = insertelement <16 x i8> %v.4, i8 %a, i32 6
+  %v.7 = insertelement <16 x i8> %v.6, i8 %b, i32 7
   %v.10 = insertelement <16 x i8> %v.7, i8 %a, i32 10
   %v.11 = insertelement <16 x i8> %v.10, i8 %a, i32 11
   %v.12 = insertelement <16 x i8> %v.11, i8 %b, i32 12
diff --git a/llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll b/llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll
index 9efe0b33910c8..2905d707bdd09 100644
--- a/llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll
+++ b/llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll
@@ -37,6 +37,10 @@ define void @select_v16f16(ptr %a, ptr %b) vscale_range(2,0) #0 {
 ; CHECK-NEXT:    ld1h { z0.h }, p0/z, [x0]
 ; CHECK-NEXT:    ld1h { z1.h }, p0/z, [x1]
 ; CHECK-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
+; CHECK-NEXT:    mov z2.h, p1/z, #-1 // =0xffffffffffffffff
+; CHECK-NEXT:    ptrue p1.h
+; CHECK-NEXT:    and z2.h, z2.h, #0x1
+; CHECK-NEXT:    cmpne p1.h, p1/z, z2.h, #0
 ; CHECK-NEXT:    sel z0.h, p1, z0.h, z1.h
 ; CHECK-NEXT:    st1h { z0.h }, p0, [x0]
 ; CHECK-NEXT:    ret
@@ -59,8 +63,15 @@ define void @select_v32f16(ptr %a, ptr %b) #0 {
 ; VBITS_GE_256-NEXT:    ld1h { z3.h }, p0/z, [x1]
 ; VBITS_GE_256-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
 ; VBITS_GE_256-NEXT:    fcmeq p2.h, p0/z, z2.h, z3.h
-; VBITS_GE_256-NEXT:    sel z0.h, p1, z0.h, z1.h
-; VBITS_GE_256-NEXT:    sel z1.h, p2, z2.h, z3.h
+; VBITS_GE_256-NEXT:    mov z4.h, p1/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_256-NEXT:    ptrue p1.h
+; VBITS_GE_256-NEXT:    mov z5.h, p2/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_256-NEXT:    and z4.h, z4.h, #0x1
+; VBITS_GE_256-NEXT:    and z5.h, z5.h, #0x1
+; VBITS_GE_256-NEXT:    cmpne p2.h, p1/z, z4.h, #0
+; VBITS_GE_256-NEXT:    cmpne p1.h, p1/z, z5.h, #0
+; VBITS_GE_256-NEXT:    sel z0.h, p2, z0.h, z1.h
+; VBITS_GE_256-NEXT:    sel z1.h, p1, z2.h, z3.h
 ; VBITS_GE_256-NEXT:    st1h { z0.h }, p0, [x0, x8, lsl #1]
 ; VBITS_GE_256-NEXT:    st1h { z1.h }, p0, [x0]
 ; VBITS_GE_256-NEXT:    ret
@@ -71,6 +82,10 @@ define void @select_v32f16(ptr %a, ptr %b) #0 {
 ; VBITS_GE_512-NEXT:    ld1h { z0.h }, p0/z, [x0]
 ; VBITS_GE_512-NEXT:    ld1h { z1.h }, p0/z, [x1]
 ; VBITS_GE_512-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
+; VBITS_GE_512-NEXT:    mov z2.h, p1/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_512-NEXT:    ptrue p1.h
+; VBITS_GE_512-NEXT:    and z2.h, z2.h, #0x1
+; VBITS_GE_512-NEXT:    cmpne p1.h, p1/z, z2.h, #0
 ; VBITS_GE_512-NEXT:    sel z0.h, p1, z0.h, z1.h
 ; VBITS_GE_512-NEXT:    st1h { z0.h }, p0, [x0]
 ; VBITS_GE_512-NEXT:    ret
@@ -89,6 +104,10 @@ define void @select_v64f16(ptr %a, ptr %b) vscale_range(8,0) #0 {
 ; CHECK-NEXT:    ld1h { z0.h }, p0/z, [x0]
 ; CHECK-NEXT:    ld1h { z1.h }, p0/z, [x1]
 ; CHECK-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
+; CHECK-NEXT:    mov z2.h, p1/z, #-1 // =0xffffffffffffffff
+; CHECK-NEXT:    ptrue p1.h
+; CHECK-NEXT:    and z2.h, z2.h, #0x1
+; CHECK-NEXT:    cmpne p1.h, p1/z, z2.h, #0
 ; CHECK-NEXT:    sel z0.h, p1, z0.h, z1.h
 ; CHECK-NEXT:    st1h { z0.h }, p0, [x0]
 ; CHECK-NEXT:    ret
@@ -107,6 +126,10 @@ define void @select_v128f16(ptr %a, ptr %b) vscale_range(16,0) #0 {
 ; CHECK-NEXT:    ld1h { z0.h }, p0/z, [x0]
 ; CHECK-NEXT:    ld1h { z1.h }, p0/z, [x1]
 ; CHECK-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
+; CHECK-NEXT:    mov z2.h, p1/z, #-1 // =0xffffffffffffffff
+; CHECK-NEXT:    ptrue p1.h
+; CHECK-NEXT:    and z2.h, z2.h, #0x1
+; CHECK-NEXT:    cmpne p1.h, p1/z, z2.h, #0
 ; CHECK-NEXT:    sel z0.h, p1, z0.h, z1.h
 ; CHECK-NEXT:    st1h { z0.h }, p0, [x0]
 ; CHECK-NEXT:    ret
@@ -150,6 +173,10 @@ define void @select_v8f32(ptr %a, ptr %b) vscale_range(2,0) #0 {
 ; CHECK-NEXT:    ld1w { z0.s }, p0/z, [x0]
 ; CHECK-NEXT:    ld1w { z1.s }, p0/z, [x1]
 ; CHECK-NEXT:    fcmeq p1.s, p0/z, z0.s, z1.s
+; CHECK-NEXT:    mov z2.s, p1/z, #-1 // =0xffffffffffffffff
+; CHECK-NEXT:    ptrue p1.s
+; CHECK-NEXT:    and z2.s, z2.s, #0x1
+; CHECK-NEXT:    cmpne p1.s, p1/z, z2.s, #0
 ; CHECK-NEXT:    sel z0.s, p1, z0.s, z1.s
 ; CHECK-NEXT:    st1w { z0.s }, p0, [x0]
 ; CHECK-NEXT:    ret
@@ -172,8 +199,15 @@ define void @select_v16f32(ptr %a, ptr %b) #0 {
 ; VBITS_GE_256-NEXT:    ld1w { z3.s }, p0/z, [x1]
 ; VBITS_GE_256-NEXT:    fcmeq p1.s, p0/z, z0.s, z1.s
 ; VBITS_GE_256-NEXT:    fcmeq p2.s, p0/z, z2.s, z3.s
-; VBITS_GE_256-NEXT:    sel z0.s, p1, z0.s, z1.s
-; VBITS_GE_256-NEXT:    sel z1.s, p2, z2.s, z3.s
+; VBITS_GE_256-NEXT:    mov z4.s, p1/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_256-NEXT:    ptrue p1.s
+; VBITS_GE_256-NEXT:    mov z5.s, p2/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_256-NEXT:    and z4.s, z4.s, #0x1
+; VBITS_...
[truncated]

@llvmbot
Copy link
Member

llvmbot commented Jun 6, 2025

@llvm/pr-subscribers-backend-x86

Author: Björn Pettersson (bjope)

Changes

Add support in isGuaranteedNotToBeUndefOrPoison and
SimplifyDemandedVectorElts to avoid regressions seen
after a previous commit fixing #141034.

Patch is 158.07 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/143103.diff

14 Files Affected:

  • (modified) llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp (+51-13)
  • (modified) llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp (+109-10)
  • (modified) llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp (+4)
  • (modified) llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll (+67)
  • (modified) llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll (+75-6)
  • (modified) llvm/test/CodeGen/AArch64/sve-fixed-length-int-vselect.ll (+100-8)
  • (modified) llvm/test/CodeGen/AArch64/sve-fixed-length-masked-gather.ll (+5-1)
  • (modified) llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-vselect.ll (+21)
  • (modified) llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-vselect.ll (+28)
  • (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfw-web-simplification.ll (+60-30)
  • (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vw-web-simplification.ll (+36-19)
  • (modified) llvm/test/CodeGen/RISCV/rvv/vector-deinterleave.ll (+36)
  • (modified) llvm/test/CodeGen/X86/pr62286.ll (+8-6)
  • (modified) llvm/test/CodeGen/X86/vector-interleaved-load-i16-stride-4.ll (+693-597)
diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index aba3c0f80a024..a52edca64dbce 100644
--- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -22905,6 +22905,7 @@ SDValue DAGCombiner::visitINSERT_VECTOR_ELT(SDNode *N) {
   auto *IndexC = dyn_cast<ConstantSDNode>(EltNo);
 
   // Insert into out-of-bounds element is undefined.
+  // Code below relies on that we handle this special case early.
   if (IndexC && VT.isFixedLengthVector() &&
       IndexC->getZExtValue() >= VT.getVectorNumElements())
     return DAG.getUNDEF(VT);
@@ -22915,14 +22916,29 @@ SDValue DAGCombiner::visitINSERT_VECTOR_ELT(SDNode *N) {
       InVec == InVal.getOperand(0) && EltNo == InVal.getOperand(1))
     return InVec;
 
-  if (!IndexC) {
-    // If this is variable insert to undef vector, it might be better to splat:
-    // inselt undef, InVal, EltNo --> build_vector < InVal, InVal, ... >
-    if (InVec.isUndef() && TLI.shouldSplatInsEltVarIndex(VT))
-      return DAG.getSplat(VT, DL, InVal);
-    return SDValue();
+  // If this is variable insert to undef vector, it might be better to splat:
+  // inselt undef, InVal, EltNo --> build_vector < InVal, InVal, ... >
+  if (!IndexC && InVec.isUndef() && TLI.shouldSplatInsEltVarIndex(VT))
+    return DAG.getSplat(VT, DL, InVal);
+
+  // Try to drop insert of UNDEF/POISON elements. This is also done in getNode,
+  // but we also do it as a DAG combine since for example simplifications into
+  // SPLAT_VECTOR/BUILD_VECTOR may turn poison elements into undef/zero etc, and
+  // then suddenly the InVec is guaranteed to not be poison.
+  if (InVal.isUndef()) {
+    if (IndexC && VT.isFixedLengthVector()) {
+      APInt EltMask = APInt::getOneBitSet(VT.getVectorNumElements(),
+                                          IndexC->getZExtValue());
+      if (DAG.isGuaranteedNotToBePoison(InVec, EltMask))
+        return InVec;
+    } else if (DAG.isGuaranteedNotToBePoison(InVec)) {
+      return InVec;
+    }
   }
 
+  if (!IndexC)
+    return SDValue();
+
   if (VT.isScalableVector())
     return SDValue();
 
@@ -27355,18 +27371,40 @@ SDValue DAGCombiner::visitINSERT_SUBVECTOR(SDNode *N) {
   SDValue N2 = N->getOperand(2);
   uint64_t InsIdx = N->getConstantOperandVal(2);
 
-  // If inserting an UNDEF, just return the original vector.
-  if (N1.isUndef())
-    return N0;
+  // If inserting an UNDEF, just return the original vector (unless it makes the
+  // result more poisonous).
+  if (N1.isUndef()){
+    if (VT.isFixedLengthVector()) {
+      unsigned SubVecNumElts = N1.getValueType().getVectorNumElements();
+      APInt EltMask = APInt::getBitsSet(VT.getVectorNumElements(),
+                                        InsIdx, InsIdx + SubVecNumElts);
+      if (DAG.isGuaranteedNotToBePoison(N0, EltMask))
+        return N0;
+    } else if (DAG.isGuaranteedNotToBePoison(N0))
+      return N0;
+  }
 
-  // If this is an insert of an extracted vector into an undef vector, we can
-  // just use the input to the extract if the types match, and can simplify
+  // If this is an insert of an extracted vector into an undef/poison vector, we
+  // can just use the input to the extract if the types match, and can simplify
   // in some cases even if they don't.
   if (N0.isUndef() && N1.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
       N1.getOperand(1) == N2) {
+    EVT N1VT = N1.getValueType();
     EVT SrcVT = N1.getOperand(0).getValueType();
-    if (SrcVT == VT)
-      return N1.getOperand(0);
+    if (SrcVT == VT) {
+      // Need to ensure that result isn't more poisonous if skipping both the
+      // extract+insert.
+      if (N0.getOpcode() == ISD::POISON)
+        return N1.getOperand(0);
+      if (VT.isFixedLengthVector() && N1VT.isFixedLengthVector()) {
+        unsigned SubVecNumElts = N1VT.getVectorNumElements();
+        APInt EltMask = APInt::getBitsSet(VT.getVectorNumElements(),
+                                          InsIdx, InsIdx + SubVecNumElts);
+        if (DAG.isGuaranteedNotToBePoison(N1.getOperand(0), ~EltMask))
+          return N1.getOperand(0);
+      } else if (DAG.isGuaranteedNotToBePoison(N1.getOperand(0)))
+        return N1.getOperand(0);
+    }
     // TODO: To remove the zero check, need to adjust the offset to
     // a multiple of the new src type.
     if (isNullConstant(N2)) {
diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index 279c7daf71c33..fadfd35489e54 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -5454,6 +5454,60 @@ bool SelectionDAG::isGuaranteedNotToBeUndefOrPoison(SDValue Op,
     }
     return true;
 
+  case ISD::INSERT_SUBVECTOR: {
+    if (Op.getValueType().isScalableVector())
+      break;
+    SDValue Src = Op.getOperand(0);
+    SDValue Sub = Op.getOperand(1);
+    uint64_t Idx = Op.getConstantOperandVal(2);
+    unsigned NumSubElts = Sub.getValueType().getVectorNumElements();
+    APInt DemandedSubElts = DemandedElts.extractBits(NumSubElts, Idx);
+    APInt DemandedSrcElts = DemandedElts;
+    DemandedSrcElts.clearBits(Idx, Idx + NumSubElts);
+
+    if (!!DemandedSubElts &&
+        !isGuaranteedNotToBeUndefOrPoison(Sub, DemandedSubElts,
+                                          PoisonOnly, Depth + 1))
+        return false;
+    if (!!DemandedSrcElts &&
+        !isGuaranteedNotToBeUndefOrPoison(Src, DemandedSrcElts,
+                                          PoisonOnly, Depth + 1))
+      return false;
+    return true;
+  }
+
+  case ISD::INSERT_VECTOR_ELT: {
+    SDValue InVec = Op.getOperand(0);
+    SDValue InVal = Op.getOperand(1);
+    SDValue EltNo = Op.getOperand(2);
+    EVT VT = InVec.getValueType();
+    auto *IndexC = dyn_cast<ConstantSDNode>(EltNo);
+    if (IndexC && VT.isFixedLengthVector() &&
+        IndexC->getZExtValue() < VT.getVectorNumElements()) {
+      if (DemandedElts[IndexC->getZExtValue()] &&
+          !isGuaranteedNotToBeUndefOrPoison(InVal, PoisonOnly, Depth + 1))
+        return false;
+      APInt InVecDemandedElts = DemandedElts;
+      InVecDemandedElts.clearBit(IndexC->getZExtValue());
+      if (!!InVecDemandedElts &&
+          !isGuaranteedNotToBeUndefOrPoison(InVec, InVecDemandedElts,
+                                            PoisonOnly, Depth + 1))
+        return false;
+      return true;
+    }
+    break;
+  }
+
+  case ISD::SCALAR_TO_VECTOR:
+    // If only demanding upper (undef) elements.
+    if (DemandedElts.ugt(1))
+      return PoisonOnly;
+    // If only demanding element 0, or only considering poison.
+    if (PoisonOnly || DemandedElts == 0)
+      return isGuaranteedNotToBeUndefOrPoison(Op.getOperand(0), PoisonOnly,
+                                              Depth + 1);
+    return false;
+
   case ISD::SPLAT_VECTOR:
     return isGuaranteedNotToBeUndefOrPoison(Op.getOperand(0), PoisonOnly,
                                             Depth + 1);
@@ -7900,23 +7954,42 @@ SDValue SelectionDAG::getNode(unsigned Opcode, const SDLoc &DL, EVT VT,
     // INSERT_VECTOR_ELT into out-of-bounds element is an UNDEF, except
     // for scalable vectors where we will generate appropriate code to
     // deal with out-of-bounds cases correctly.
-    if (N3C && N1.getValueType().isFixedLengthVector() &&
-        N3C->getZExtValue() >= N1.getValueType().getVectorNumElements())
+    if (N3C && VT.isFixedLengthVector() &&
+        N3C->getZExtValue() >= VT.getVectorNumElements())
       return getUNDEF(VT);
 
     // Undefined index can be assumed out-of-bounds, so that's UNDEF too.
     if (N3.isUndef())
       return getUNDEF(VT);
 
-    // If the inserted element is an UNDEF, just use the input vector.
-    if (N2.isUndef())
+    // If inserting poison, just use the input vector.
+    if (N2.getOpcode() == ISD::POISON)
       return N1;
 
+    // Inserting undef into undef/poison is still undef.
+    if (N2.getOpcode() == ISD::UNDEF && N1.isUndef())
+      return getUNDEF(VT);
+
+    // If the inserted element is an UNDEF, just use the input vector.
+    // But not if skipping the insert could make the result more poisonous.
+    if (N2.isUndef()) {
+      if (N3C && VT.isFixedLengthVector()) {
+        APInt EltMask = APInt::getOneBitSet(VT.getVectorNumElements(),
+                                            N3C->getZExtValue());
+        if (isGuaranteedNotToBePoison(N1, EltMask))
+          return N1;
+      } else if (isGuaranteedNotToBePoison(N1))
+        return N1;
+    }
     break;
   }
   case ISD::INSERT_SUBVECTOR: {
-    // Inserting undef into undef is still undef.
-    if (N1.isUndef() && N2.isUndef())
+    // If inserting poison, just use the input vector,
+    if (N2.getOpcode() == ISD::POISON)
+      return N1;
+
+    // Inserting undef into undef/poison is still undef.
+    if (N2.getOpcode() == ISD::UNDEF && N1.isUndef())
       return getUNDEF(VT);
 
     EVT N2VT = N2.getValueType();
@@ -7945,11 +8018,37 @@ SDValue SelectionDAG::getNode(unsigned Opcode, const SDLoc &DL, EVT VT,
     if (VT == N2VT)
       return N2;
 
-    // If this is an insert of an extracted vector into an undef vector, we
-    // can just use the input to the extract.
+    // If this is an insert of an extracted vector into an undef/poison vector,
+    // we can just use the input to the extract. But not if skipping the
+    // extract+insert could make the result more poisonous.
     if (N1.isUndef() && N2.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
-        N2.getOperand(1) == N3 && N2.getOperand(0).getValueType() == VT)
-      return N2.getOperand(0);
+        N2.getOperand(1) == N3 && N2.getOperand(0).getValueType() == VT) {
+      if (N1.getOpcode() == ISD::POISON)
+        return N2.getOperand(0);
+      if (VT.isFixedLengthVector() && N2VT.isFixedLengthVector()) {
+        unsigned LoBit = N3->getAsZExtVal();
+        unsigned HiBit = LoBit + N2VT.getVectorNumElements();
+        APInt EltMask = APInt::getBitsSet(VT.getVectorNumElements(),
+                                          LoBit, HiBit);
+        if (isGuaranteedNotToBePoison(N2.getOperand(0), ~EltMask))
+          return N2.getOperand(0);
+      } else if (isGuaranteedNotToBePoison(N2.getOperand(0)))
+        return N2.getOperand(0);
+    }
+
+    // If the inserted subvector is UNDEF, just use the input vector.
+    // But not if skipping the insert could make the result more poisonous.
+    if (N2.isUndef()) {
+      if (VT.isFixedLengthVector()) {
+        unsigned LoBit = N3->getAsZExtVal();
+        unsigned HiBit = LoBit + N2VT.getVectorNumElements();
+        APInt EltMask = APInt::getBitsSet(VT.getVectorNumElements(),
+                                          LoBit, HiBit);
+        if (isGuaranteedNotToBePoison(N1, EltMask))
+          return N1;
+      } else if (isGuaranteedNotToBePoison(N1))
+        return N1;
+    }
     break;
   }
   case ISD::BITCAST:
diff --git a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
index e8e820ac1f695..643ec26bba3f7 100644
--- a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
@@ -3359,6 +3359,10 @@ bool TargetLowering::SimplifyDemandedVectorElts(
     APInt DemandedSrcElts = DemandedElts;
     DemandedSrcElts.clearBits(Idx, Idx + NumSubElts);
 
+    // If none of the sub operand elements are demanded, bypass the insert.
+    if (!DemandedSubElts)
+      return TLO.CombineTo(Op, Src);
+
     APInt SubUndef, SubZero;
     if (SimplifyDemandedVectorElts(Sub, DemandedSubElts, SubUndef, SubZero, TLO,
                                    Depth + 1))
diff --git a/llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll b/llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll
index 94074d1689f6a..2c44f56316801 100644
--- a/llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll
+++ b/llvm/test/CodeGen/AArch64/arm64-vector-insertion.ll
@@ -66,6 +66,36 @@ define <16 x i8> @test_insert_v16i8_insert_2_undef_base(i8 %a) {
   %v.6 = insertelement <16 x i8> %v.4, i8 %a, i32 6
   %v.7 = insertelement <16 x i8> %v.6, i8 %a, i32 7
   %v.8 = insertelement <16 x i8> %v.7, i8 %a, i32 8
+  %v.10 = insertelement <16 x i8> %v.8, i8 %a, i32 10
+  %v.11 = insertelement <16 x i8> %v.10, i8 %a, i32 11
+  %v.12 = insertelement <16 x i8> %v.11, i8 %a, i32 12
+  %v.13 = insertelement <16 x i8> %v.12, i8 %a, i32 13
+  %v.14 = insertelement <16 x i8> %v.13, i8 %a, i32 14
+  %v.15 = insertelement <16 x i8> %v.14, i8 %a, i32 15
+  ret <16 x i8> %v.15
+}
+
+; Similar to above, but we leave element 8 as undef. One interesting part with
+; this test case is that %a may be poison, so simply inserting %a also at
+; index 8 would make the result vector more poisonous.
+define <16 x i8> @test_insert_v16i8_insert_2_undef_base_skip8(i32 %a0) {
+; CHECK-LABEL: test_insert_v16i8_insert_2_undef_base_skip8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    lsr w8, w0, #5
+; CHECK-NEXT:    dup.16b v0, w8
+; CHECK-NEXT:    mov.b v0[5], wzr
+; CHECK-NEXT:    mov.b v0[8], w8
+; CHECK-NEXT:    mov.b v0[9], wzr
+; CHECK-NEXT:    ret
+  %a1 = lshr exact i32 %a0, 5
+  %a = trunc i32 %a1 to i8
+  %v.0 = insertelement <16 x i8> <i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef>  , i8 %a, i32 0
+  %v.1 = insertelement <16 x i8> %v.0, i8 %a, i32 1
+  %v.2 = insertelement <16 x i8> %v.1, i8 %a, i32 2
+  %v.3 = insertelement <16 x i8> %v.2, i8 %a, i32 3
+  %v.4 = insertelement <16 x i8> %v.3, i8 %a, i32 4
+  %v.6 = insertelement <16 x i8> %v.4, i8 %a, i32 6
+  %v.7 = insertelement <16 x i8> %v.6, i8 %a, i32 7
   %v.10 = insertelement <16 x i8> %v.7, i8 %a, i32 10
   %v.11 = insertelement <16 x i8> %v.10, i8 %a, i32 11
   %v.12 = insertelement <16 x i8> %v.11, i8 %a, i32 12
@@ -94,6 +124,43 @@ define <16 x i8> @test_insert_v16i8_insert_2_undef_base_different_valeus(i8 %a,
   %v.6 = insertelement <16 x i8> %v.4, i8 %a, i32 6
   %v.7 = insertelement <16 x i8> %v.6, i8 %b, i32 7
   %v.8 = insertelement <16 x i8> %v.7, i8 %a, i32 8
+  %v.10 = insertelement <16 x i8> %v.8, i8 %a, i32 10
+  %v.11 = insertelement <16 x i8> %v.10, i8 %a, i32 11
+  %v.12 = insertelement <16 x i8> %v.11, i8 %b, i32 12
+  %v.13 = insertelement <16 x i8> %v.12, i8 %a, i32 13
+  %v.14 = insertelement <16 x i8> %v.13, i8 %a, i32 14
+  %v.15 = insertelement <16 x i8> %v.14, i8 %b, i32 15
+  ret <16 x i8> %v.15
+}
+
+; Similar to above, but we leave element 8 as undef. One interesting part with
+; this test case is that %a and %b may be poison, so simply inserting %a or %b
+; at index 8 would make the result vector more poisonous.
+define <16 x i8> @test_insert_v16i8_insert_2_undef_base_different_valeus_skip8(i32 %a0, i32 %b0) {
+; CHECK-LABEL: test_insert_v16i8_insert_2_undef_base_different_valeus_skip8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    lsr w8, w0, #5
+; CHECK-NEXT:    dup.16b v0, w8
+; CHECK-NEXT:    lsr w8, w1, #5
+; CHECK-NEXT:    mov.b v0[2], w8
+; CHECK-NEXT:    mov.b v0[5], wzr
+; CHECK-NEXT:    mov.b v0[7], w8
+; CHECK-NEXT:    mov.b v0[8], w8
+; CHECK-NEXT:    mov.b v0[9], wzr
+; CHECK-NEXT:    mov.b v0[12], w8
+; CHECK-NEXT:    mov.b v0[15], w8
+; CHECK-NEXT:    ret
+  %a1 = lshr exact i32 %a0, 5
+  %a = trunc i32 %a1 to i8
+  %b1 = lshr exact i32 %b0, 5
+  %b = trunc i32 %b1 to i8
+  %v.0 = insertelement <16 x i8> <i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 0, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef, i8 undef>  , i8 %a, i32 0
+  %v.1 = insertelement <16 x i8> %v.0, i8 %a, i32 1
+  %v.2 = insertelement <16 x i8> %v.1, i8 %b, i32 2
+  %v.3 = insertelement <16 x i8> %v.2, i8 %a, i32 3
+  %v.4 = insertelement <16 x i8> %v.3, i8 %a, i32 4
+  %v.6 = insertelement <16 x i8> %v.4, i8 %a, i32 6
+  %v.7 = insertelement <16 x i8> %v.6, i8 %b, i32 7
   %v.10 = insertelement <16 x i8> %v.7, i8 %a, i32 10
   %v.11 = insertelement <16 x i8> %v.10, i8 %a, i32 11
   %v.12 = insertelement <16 x i8> %v.11, i8 %b, i32 12
diff --git a/llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll b/llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll
index 9efe0b33910c8..2905d707bdd09 100644
--- a/llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll
+++ b/llvm/test/CodeGen/AArch64/sve-fixed-length-fp-vselect.ll
@@ -37,6 +37,10 @@ define void @select_v16f16(ptr %a, ptr %b) vscale_range(2,0) #0 {
 ; CHECK-NEXT:    ld1h { z0.h }, p0/z, [x0]
 ; CHECK-NEXT:    ld1h { z1.h }, p0/z, [x1]
 ; CHECK-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
+; CHECK-NEXT:    mov z2.h, p1/z, #-1 // =0xffffffffffffffff
+; CHECK-NEXT:    ptrue p1.h
+; CHECK-NEXT:    and z2.h, z2.h, #0x1
+; CHECK-NEXT:    cmpne p1.h, p1/z, z2.h, #0
 ; CHECK-NEXT:    sel z0.h, p1, z0.h, z1.h
 ; CHECK-NEXT:    st1h { z0.h }, p0, [x0]
 ; CHECK-NEXT:    ret
@@ -59,8 +63,15 @@ define void @select_v32f16(ptr %a, ptr %b) #0 {
 ; VBITS_GE_256-NEXT:    ld1h { z3.h }, p0/z, [x1]
 ; VBITS_GE_256-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
 ; VBITS_GE_256-NEXT:    fcmeq p2.h, p0/z, z2.h, z3.h
-; VBITS_GE_256-NEXT:    sel z0.h, p1, z0.h, z1.h
-; VBITS_GE_256-NEXT:    sel z1.h, p2, z2.h, z3.h
+; VBITS_GE_256-NEXT:    mov z4.h, p1/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_256-NEXT:    ptrue p1.h
+; VBITS_GE_256-NEXT:    mov z5.h, p2/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_256-NEXT:    and z4.h, z4.h, #0x1
+; VBITS_GE_256-NEXT:    and z5.h, z5.h, #0x1
+; VBITS_GE_256-NEXT:    cmpne p2.h, p1/z, z4.h, #0
+; VBITS_GE_256-NEXT:    cmpne p1.h, p1/z, z5.h, #0
+; VBITS_GE_256-NEXT:    sel z0.h, p2, z0.h, z1.h
+; VBITS_GE_256-NEXT:    sel z1.h, p1, z2.h, z3.h
 ; VBITS_GE_256-NEXT:    st1h { z0.h }, p0, [x0, x8, lsl #1]
 ; VBITS_GE_256-NEXT:    st1h { z1.h }, p0, [x0]
 ; VBITS_GE_256-NEXT:    ret
@@ -71,6 +82,10 @@ define void @select_v32f16(ptr %a, ptr %b) #0 {
 ; VBITS_GE_512-NEXT:    ld1h { z0.h }, p0/z, [x0]
 ; VBITS_GE_512-NEXT:    ld1h { z1.h }, p0/z, [x1]
 ; VBITS_GE_512-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
+; VBITS_GE_512-NEXT:    mov z2.h, p1/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_512-NEXT:    ptrue p1.h
+; VBITS_GE_512-NEXT:    and z2.h, z2.h, #0x1
+; VBITS_GE_512-NEXT:    cmpne p1.h, p1/z, z2.h, #0
 ; VBITS_GE_512-NEXT:    sel z0.h, p1, z0.h, z1.h
 ; VBITS_GE_512-NEXT:    st1h { z0.h }, p0, [x0]
 ; VBITS_GE_512-NEXT:    ret
@@ -89,6 +104,10 @@ define void @select_v64f16(ptr %a, ptr %b) vscale_range(8,0) #0 {
 ; CHECK-NEXT:    ld1h { z0.h }, p0/z, [x0]
 ; CHECK-NEXT:    ld1h { z1.h }, p0/z, [x1]
 ; CHECK-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
+; CHECK-NEXT:    mov z2.h, p1/z, #-1 // =0xffffffffffffffff
+; CHECK-NEXT:    ptrue p1.h
+; CHECK-NEXT:    and z2.h, z2.h, #0x1
+; CHECK-NEXT:    cmpne p1.h, p1/z, z2.h, #0
 ; CHECK-NEXT:    sel z0.h, p1, z0.h, z1.h
 ; CHECK-NEXT:    st1h { z0.h }, p0, [x0]
 ; CHECK-NEXT:    ret
@@ -107,6 +126,10 @@ define void @select_v128f16(ptr %a, ptr %b) vscale_range(16,0) #0 {
 ; CHECK-NEXT:    ld1h { z0.h }, p0/z, [x0]
 ; CHECK-NEXT:    ld1h { z1.h }, p0/z, [x1]
 ; CHECK-NEXT:    fcmeq p1.h, p0/z, z0.h, z1.h
+; CHECK-NEXT:    mov z2.h, p1/z, #-1 // =0xffffffffffffffff
+; CHECK-NEXT:    ptrue p1.h
+; CHECK-NEXT:    and z2.h, z2.h, #0x1
+; CHECK-NEXT:    cmpne p1.h, p1/z, z2.h, #0
 ; CHECK-NEXT:    sel z0.h, p1, z0.h, z1.h
 ; CHECK-NEXT:    st1h { z0.h }, p0, [x0]
 ; CHECK-NEXT:    ret
@@ -150,6 +173,10 @@ define void @select_v8f32(ptr %a, ptr %b) vscale_range(2,0) #0 {
 ; CHECK-NEXT:    ld1w { z0.s }, p0/z, [x0]
 ; CHECK-NEXT:    ld1w { z1.s }, p0/z, [x1]
 ; CHECK-NEXT:    fcmeq p1.s, p0/z, z0.s, z1.s
+; CHECK-NEXT:    mov z2.s, p1/z, #-1 // =0xffffffffffffffff
+; CHECK-NEXT:    ptrue p1.s
+; CHECK-NEXT:    and z2.s, z2.s, #0x1
+; CHECK-NEXT:    cmpne p1.s, p1/z, z2.s, #0
 ; CHECK-NEXT:    sel z0.s, p1, z0.s, z1.s
 ; CHECK-NEXT:    st1w { z0.s }, p0, [x0]
 ; CHECK-NEXT:    ret
@@ -172,8 +199,15 @@ define void @select_v16f32(ptr %a, ptr %b) #0 {
 ; VBITS_GE_256-NEXT:    ld1w { z3.s }, p0/z, [x1]
 ; VBITS_GE_256-NEXT:    fcmeq p1.s, p0/z, z0.s, z1.s
 ; VBITS_GE_256-NEXT:    fcmeq p2.s, p0/z, z2.s, z3.s
-; VBITS_GE_256-NEXT:    sel z0.s, p1, z0.s, z1.s
-; VBITS_GE_256-NEXT:    sel z1.s, p2, z2.s, z3.s
+; VBITS_GE_256-NEXT:    mov z4.s, p1/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_256-NEXT:    ptrue p1.s
+; VBITS_GE_256-NEXT:    mov z5.s, p2/z, #-1 // =0xffffffffffffffff
+; VBITS_GE_256-NEXT:    and z4.s, z4.s, #0x1
+; VBITS_...
[truncated]

@github-actions GitHub Actions
Copy link

github-actions bot commented Jun 6, 2025 •
edited
Loading

✅ With the latest revision this PR passed the undef deprecator.

@github-actions GitHub Actions
Copy link

github-actions bot commented Jun 6, 2025 •
edited
Loading

✅ With the latest revision this PR passed the C/C++ code formatter.

@bjope bjope mentioned this pull request Jun 8, 2025
Open
@bjope bjope force-pushed the users/bjope/insertundef_3 branch from 2693525 to 332a73f Compare June 10, 2025 12:29
@bjope bjope changed the base branch from main to users/bjope/insertundef_2 June 10, 2025 12:32
RKSimon
RKSimon reviewed Jun 10, 2025
View reviewed changes
llvm/include/llvm/CodeGen/SelectionDAGNodes.h Outdated
/// Recursively peek through INSERT_VECTOR_ELT nodes, returning the source
/// vector operand of \p V, as long as \p V is an does INSERT_VECTOR_ELT
/// operation that do not insert into any of the demanded vector elts.
LLVM_ABI SDValue peekThroughInsertVectorElt(SDValue V, APInt DemandedElts);
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

const APInt &DemandedElts (in case vector count > 64)

llvm/include/llvm/CodeGen/SelectionDAGNodes.h Outdated
@@ -1881,6 +1881,11 @@ LLVM_ABI SDValue peekThroughExtractSubvectors(SDValue V);
/// If \p V is not a truncation, it is returned as-is.
LLVM_ABI SDValue peekThroughTruncates(SDValue V);

/// Recursively peek through INSERT_VECTOR_ELT nodes, returning the source
/// vector operand of \p V, as long as \p V is an does INSERT_VECTOR_ELT
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

is an INSERT_VECTOR_ELT

@bjope
[SelectionDAG] Deal with POISON for INSERT_VECTOR_ELT/INSERT_SUBVECTO…
ff9e7aa 
…R (part 2)

Add support in isGuaranteedNotToBeUndefOrPoison and
SimplifyDemandedVectorElts to avoid regressions
seen after a previous commit fixing #141034.
@bjope bjope force-pushed the users/bjope/insertundef_2 branch from 33ebb39 to bf22ac6 Compare June 11, 2025 09:35
@bjope bjope force-pushed the users/bjope/insertundef_3 branch from 332a73f to ff9e7aa Compare June 11, 2025 09:35
@bjope bjope force-pushed the users/bjope/insertundef_3 branch from fc5b7aa to b635383 Compare June 11, 2025 10:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@RKSimon RKSimon RKSimon left review comments

Assignees
No one assigned
Labels
backend:AArch64 backend:AMDGPU backend:X86 llvm:SelectionDAG SelectionDAGISel as well
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@bjope @llvmbot @RKSimon
Morty Proxy This is a proxified and sanitized view of the page, visit original site.