[mlir][vector] Fix crash in vector.insert canonicalization
#97801
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
matthias-springer
requested review from
dcaballe and
nicolasvasilache
as code owners
|
@llvm/pr-subscribers-mlir Author: Matthias Springer (matthias-springer) ChangesThe Full diff: https://github.com/llvm/llvm-project/pull/97801.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Vector/IR/VectorOps.cpp b/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
index 53a6648de014c..7967aa1582fd8 100644
--- a/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
+++ b/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
@@ -2851,6 +2851,9 @@ class InsertOpConstantFolder final : public OpRewritePattern<InsertOp> {
Attribute vectorDestCst;
if (!matchPattern(destVector, m_Constant(&vectorDestCst)))
return failure();
+ auto denseDest = llvm::dyn_cast<DenseElementsAttr>(vectorDestCst);
+ if (!denseDest)
+ return failure();
VectorType destTy = destVector.getType();
if (destTy.isScalable())
@@ -2860,9 +2863,7 @@ class InsertOpConstantFolder final : public OpRewritePattern<InsertOp> {
if (destTy.getNumElements() > vectorSizeFoldThreshold &&
!destVector.hasOneUse())
return failure();
-
- auto denseDest = llvm::cast<DenseElementsAttr>(vectorDestCst);
-
+
Value sourceValue = op.getSource();
Attribute sourceCst;
if (!matchPattern(sourceValue, m_Constant(&sourceCst)))
diff --git a/mlir/test/Dialect/Vector/canonicalize.mlir b/mlir/test/Dialect/Vector/canonicalize.mlir
index 1a674d715ca61..9f16dfb5093d0 100644
--- a/mlir/test/Dialect/Vector/canonicalize.mlir
+++ b/mlir/test/Dialect/Vector/canonicalize.mlir
@@ -2729,3 +2729,14 @@ func.func @fold_vector_step_to_constant() -> vector<4xindex> {
%0 = vector.step : vector<4xindex>
return %0 : vector<4xindex>
}
+
+// -----
+
+// CHECK-LABEL: func @vector_insert_regression(
+// CHECK: llvm.mlir.undef
+// CHECK: vector.insert
+func.func @vector_insert_regression(%arg0: i8) -> vector<4xi8> {
+ %0 = llvm.mlir.undef : vector<4xi8>
+ %1 = vector.insert %arg0, %0 [0] : i8 into vector<4xi8>
+ return %1 : vector<4xi8>
+}
|
|
@llvm/pr-subscribers-mlir-vector Author: Matthias Springer (matthias-springer) ChangesThe Full diff: https://github.com/llvm/llvm-project/pull/97801.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Vector/IR/VectorOps.cpp b/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
index 53a6648de014c..7967aa1582fd8 100644
--- a/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
+++ b/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
@@ -2851,6 +2851,9 @@ class InsertOpConstantFolder final : public OpRewritePattern<InsertOp> {
Attribute vectorDestCst;
if (!matchPattern(destVector, m_Constant(&vectorDestCst)))
return failure();
+ auto denseDest = llvm::dyn_cast<DenseElementsAttr>(vectorDestCst);
+ if (!denseDest)
+ return failure();
VectorType destTy = destVector.getType();
if (destTy.isScalable())
@@ -2860,9 +2863,7 @@ class InsertOpConstantFolder final : public OpRewritePattern<InsertOp> {
if (destTy.getNumElements() > vectorSizeFoldThreshold &&
!destVector.hasOneUse())
return failure();
-
- auto denseDest = llvm::cast<DenseElementsAttr>(vectorDestCst);
-
+
Value sourceValue = op.getSource();
Attribute sourceCst;
if (!matchPattern(sourceValue, m_Constant(&sourceCst)))
diff --git a/mlir/test/Dialect/Vector/canonicalize.mlir b/mlir/test/Dialect/Vector/canonicalize.mlir
index 1a674d715ca61..9f16dfb5093d0 100644
--- a/mlir/test/Dialect/Vector/canonicalize.mlir
+++ b/mlir/test/Dialect/Vector/canonicalize.mlir
@@ -2729,3 +2729,14 @@ func.func @fold_vector_step_to_constant() -> vector<4xindex> {
%0 = vector.step : vector<4xindex>
return %0 : vector<4xindex>
}
+
+// -----
+
+// CHECK-LABEL: func @vector_insert_regression(
+// CHECK: llvm.mlir.undef
+// CHECK: vector.insert
+func.func @vector_insert_regression(%arg0: i8) -> vector<4xi8> {
+ %0 = llvm.mlir.undef : vector<4xi8>
+ %1 = vector.insert %arg0, %0 [0] : i8 into vector<4xi8>
+ return %1 : vector<4xi8>
+}
|
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
c-rhodes
approved these changes
Jul 5, 2024
The `InsertOpConstantFolder` assumed that whenever the destination can be folded to a constant attribute, that attribute must be a `DenseElementsAttr`. That is is not necessarily the case.
matthias-springer
force-pushed
the
users/matthias-springer/vector_insert_regression
branch
from
2a7b2e6 to
9db3421
Compare
matthias-springer
deleted the
users/matthias-springer/vector_insert_regression
branch
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.
The
InsertOpConstantFolderassumed that whenever the destination can be folded to a constant attribute, that attribute must be aDenseElementsAttr. That is is not necessarily the case.