[Constant] Make Constant::getSplatValue return poison on poison #141870
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This is a follow up from llvm#141845. I'm not sure if this actually NFC but it doesn't seem to affect any of the in-tree tests. I went through the users of getSplatValue to see if anything could be cleaned up but nothing immediately stuck out.
|
@llvm/pr-subscribers-llvm-analysis Author: Luke Lau (lukel97) ChangesThis is a follow up from #141845. I'm not sure if this actually NFC but it doesn't seem to affect any of the in-tree tests. I went through the users of getSplatValue to see if anything could be cleaned up but nothing immediately stuck out. Full diff: https://github.com/llvm/llvm-project/pull/141870.diff 2 Files Affected:
diff --git a/llvm/lib/Analysis/ConstantFolding.cpp b/llvm/lib/Analysis/ConstantFolding.cpp
index 40302fbc8ee52..2476dc58375e5 100644
--- a/llvm/lib/Analysis/ConstantFolding.cpp
+++ b/llvm/lib/Analysis/ConstantFolding.cpp
@@ -3794,11 +3794,6 @@ static Constant *ConstantFoldScalableVectorCall(
SplatOps.push_back(Op);
continue;
}
- // TODO: Should getSplatValue return a poison scalar for a poison vector?
- if (isa<PoisonValue>(Op)) {
- SplatOps.push_back(PoisonValue::get(Op->getType()->getScalarType()));
- continue;
- }
Constant *Splat = Op->getSplatValue();
if (!Splat)
return nullptr;
diff --git a/llvm/lib/IR/Constants.cpp b/llvm/lib/IR/Constants.cpp
index fa453309b34ee..a3c725b2af62a 100644
--- a/llvm/lib/IR/Constants.cpp
+++ b/llvm/lib/IR/Constants.cpp
@@ -1711,6 +1711,8 @@ void ConstantVector::destroyConstantImpl() {
Constant *Constant::getSplatValue(bool AllowPoison) const {
assert(this->getType()->isVectorTy() && "Only valid for vectors!");
+ if (isa<PoisonValue>(this))
+ return PoisonValue::get(cast<VectorType>(getType())->getElementType());
if (isa<ConstantAggregateZero>(this))
return getNullValue(cast<VectorType>(getType())->getElementType());
if (auto *CI = dyn_cast<ConstantInt>(this))
|
|
@llvm/pr-subscribers-llvm-ir Author: Luke Lau (lukel97) ChangesThis is a follow up from #141845. I'm not sure if this actually NFC but it doesn't seem to affect any of the in-tree tests. I went through the users of getSplatValue to see if anything could be cleaned up but nothing immediately stuck out. Full diff: https://github.com/llvm/llvm-project/pull/141870.diff 2 Files Affected:
diff --git a/llvm/lib/Analysis/ConstantFolding.cpp b/llvm/lib/Analysis/ConstantFolding.cpp
index 40302fbc8ee52..2476dc58375e5 100644
--- a/llvm/lib/Analysis/ConstantFolding.cpp
+++ b/llvm/lib/Analysis/ConstantFolding.cpp
@@ -3794,11 +3794,6 @@ static Constant *ConstantFoldScalableVectorCall(
SplatOps.push_back(Op);
continue;
}
- // TODO: Should getSplatValue return a poison scalar for a poison vector?
- if (isa<PoisonValue>(Op)) {
- SplatOps.push_back(PoisonValue::get(Op->getType()->getScalarType()));
- continue;
- }
Constant *Splat = Op->getSplatValue();
if (!Splat)
return nullptr;
diff --git a/llvm/lib/IR/Constants.cpp b/llvm/lib/IR/Constants.cpp
index fa453309b34ee..a3c725b2af62a 100644
--- a/llvm/lib/IR/Constants.cpp
+++ b/llvm/lib/IR/Constants.cpp
@@ -1711,6 +1711,8 @@ void ConstantVector::destroyConstantImpl() {
Constant *Constant::getSplatValue(bool AllowPoison) const {
assert(this->getType()->isVectorTy() && "Only valid for vectors!");
+ if (isa<PoisonValue>(this))
+ return PoisonValue::get(cast<VectorType>(getType())->getElementType());
if (isa<ConstantAggregateZero>(this))
return getNullValue(cast<VectorType>(getType())->getElementType());
if (auto *CI = dyn_cast<ConstantInt>(this))
|
nikic
approved these changes
May 29, 2025
…ing costmodel behaviour This fixes a regression in Analysis/CostModel/RISCV/rvv-load-store.ll where we were now trying to add a cost for materializing a poison in a store: -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <4 x i32> poison, ptr %p, align 16 +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: store <4 x i32> poison, ptr %p, align 16 ; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <4 x i32> undef, ptr %p, align 16 This came about because it had a check for getSplatValue, which now returns something for poison. So to preserve the existing behaviour, check if the value is undef and return non-constant so no materialization costs are added
svkeerthy
pushed a commit
that referenced
this pull request
May 29, 2025
) This is a follow up from #141845. TargetTransformInfo::getOperandInfo needs to be updated to check for undef values as otherwise a splat is considered a constant, and some RISC-V cost model tests will start adding a cost to materialize the constant.
google-yfyang
pushed a commit
to google-yfyang/llvm-project
that referenced
this pull request
May 29, 2025
…#141870) This is a follow up from llvm#141845. TargetTransformInfo::getOperandInfo needs to be updated to check for undef values as otherwise a splat is considered a constant, and some RISC-V cost model tests will start adding a cost to materialize the constant.
sivan-shani
pushed a commit
to sivan-shani/llvm-project
that referenced
this pull request
Jun 3, 2025
…#141870) This is a follow up from llvm#141845. TargetTransformInfo::getOperandInfo needs to be updated to check for undef values as otherwise a splat is considered a constant, and some RISC-V cost model tests will start adding a cost to materialize the constant.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This is a follow up from #141845.
TargetTransformInfo::getOperandInfo needs to be updated to check for undef values as otherwise a splat is considered a constant, and some RISC-V cost model tests will start adding a cost to materialize the constant.