[Mlir-commits] [mlir] [mlir][IR] Fix overload resolution on MSVC build (PR #84589)
llvmlistbot at llvm.org
llvmlistbot at llvm.org
Fri Mar 8 16:46:44 PST 2024
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-mlir-linalg
Author: Matthias Springer (matthias-springer)
<details>
<summary>Changes</summary>
#<!-- -->82629 added additional overloads to `replaceAllUsesWith` and `replaceUsesWithIf`. This caused a build breakage with MSVC when called with ops that can implicitly convert to `Value`.
```
external/llvm-project/mlir/lib/Dialect/SCF/Transforms/TileUsingInterface.cpp(881): error C2666: 'mlir::RewriterBase::replaceAllUsesWith': 2 overloads have similar conversions
external/llvm-project/mlir/include\mlir/IR/PatternMatch.h(631): note: could be 'void mlir::RewriterBase::replaceAllUsesWith(mlir::Operation *,mlir::ValueRange)'
external/llvm-project/mlir/include\mlir/IR/PatternMatch.h(626): note: or 'void mlir::RewriterBase::replaceAllUsesWith(mlir::ValueRange,mlir::ValueRange)'
external/llvm-project/mlir/include\mlir/IR/PatternMatch.h(616): note: or 'void mlir::RewriterBase::replaceAllUsesWith(mlir::Value,mlir::Value)'
external/llvm-project/mlir/lib/Dialect/SCF/Transforms/TileUsingInterface.cpp(882): note: while trying to match the argument list '(mlir::tensor::ExtractSliceOp, T)'
with
[
T=mlir::Value
]
```
Note: The LLVM build bots (Linux and Windows) did not break, this seems to be an issue with `Tools\MSVC\14.29.30133\bin\HostX64\x64\cl.exe`.
This change renames the newly added overloads to `replaceAllOpUsesWith` and `replaceOpUsesWithIf`.
---
Full diff: https://github.com/llvm/llvm-project/pull/84589.diff
4 Files Affected:
- (modified) mlir/include/mlir/IR/PatternMatch.h (+13-7)
- (modified) mlir/lib/Dialect/Linalg/Transforms/DecomposeLinalgOps.cpp (+2-2)
- (modified) mlir/lib/IR/PatternMatch.cpp (+2-2)
- (modified) mlir/lib/Transforms/Utils/RegionUtils.cpp (+1-1)
``````````diff
diff --git a/mlir/include/mlir/IR/PatternMatch.h b/mlir/include/mlir/IR/PatternMatch.h
index e3500b3f9446d8..8d84ab6100007e 100644
--- a/mlir/include/mlir/IR/PatternMatch.h
+++ b/mlir/include/mlir/IR/PatternMatch.h
@@ -628,7 +628,10 @@ class RewriterBase : public OpBuilder {
for (auto it : llvm::zip(from, to))
replaceAllUsesWith(std::get<0>(it), std::get<1>(it));
}
- void replaceAllUsesWith(Operation *from, ValueRange to) {
+ // Note: This function cannot be called `replaceAllUsesWith` because the
+ // overload resolution, when called with an op that can be implicitly
+ // converted to a Value, would be ambiguous.
+ void replaceAllOpUsesWith(Operation *from, ValueRange to) {
replaceAllUsesWith(from->getResults(), to);
}
@@ -642,9 +645,12 @@ class RewriterBase : public OpBuilder {
void replaceUsesWithIf(ValueRange from, ValueRange to,
function_ref<bool(OpOperand &)> functor,
bool *allUsesReplaced = nullptr);
- void replaceUsesWithIf(Operation *from, ValueRange to,
- function_ref<bool(OpOperand &)> functor,
- bool *allUsesReplaced = nullptr) {
+ // Note: This function cannot be called `replaceOpUsesWithIf` because the
+ // overload resolution, when called with an op that can be implicitly
+ // converted to a Value, would be ambiguous.
+ void replaceOpUsesWithIf(Operation *from, ValueRange to,
+ function_ref<bool(OpOperand &)> functor,
+ bool *allUsesReplaced = nullptr) {
replaceUsesWithIf(from->getResults(), to, functor, allUsesReplaced);
}
@@ -652,9 +658,9 @@ class RewriterBase : public OpBuilder {
/// the listener about every in-place op modification (for every use that was
/// replaced). The optional `allUsesReplaced` flag is set to "true" if all
/// uses were replaced.
- void replaceUsesWithinBlock(Operation *op, ValueRange newValues, Block *block,
- bool *allUsesReplaced = nullptr) {
- replaceUsesWithIf(
+ void replaceOpUsesWithinBlock(Operation *op, ValueRange newValues,
+ Block *block, bool *allUsesReplaced = nullptr) {
+ replaceOpUsesWithIf(
op, newValues,
[block](OpOperand &use) {
return block->getParentOp()->isProperAncestor(use.getOwner());
diff --git a/mlir/lib/Dialect/Linalg/Transforms/DecomposeLinalgOps.cpp b/mlir/lib/Dialect/Linalg/Transforms/DecomposeLinalgOps.cpp
index 1658ea67a46077..999359c7fa8724 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/DecomposeLinalgOps.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/DecomposeLinalgOps.cpp
@@ -370,8 +370,8 @@ DecomposeLinalgOp::matchAndRewrite(GenericOp genericOp,
scalarReplacements.push_back(
residualGenericOpBody->getArgument(num + origNumInputs));
bool allUsesReplaced = false;
- rewriter.replaceUsesWithinBlock(peeledScalarOperation, scalarReplacements,
- residualGenericOpBody, &allUsesReplaced);
+ rewriter.replaceOpUsesWithinBlock(peeledScalarOperation, scalarReplacements,
+ residualGenericOpBody, &allUsesReplaced);
assert(!allUsesReplaced &&
"peeled scalar operation is erased when it wasnt expected to be");
}
diff --git a/mlir/lib/IR/PatternMatch.cpp b/mlir/lib/IR/PatternMatch.cpp
index 0a88e40f73ec6c..4079ccc7567256 100644
--- a/mlir/lib/IR/PatternMatch.cpp
+++ b/mlir/lib/IR/PatternMatch.cpp
@@ -122,7 +122,7 @@ void RewriterBase::replaceOp(Operation *op, ValueRange newValues) {
rewriteListener->notifyOperationReplaced(op, newValues);
// Replace all result uses. Also notifies the listener of modifications.
- replaceAllUsesWith(op, newValues);
+ replaceAllOpUsesWith(op, newValues);
// Erase op and notify listener.
eraseOp(op);
@@ -141,7 +141,7 @@ void RewriterBase::replaceOp(Operation *op, Operation *newOp) {
rewriteListener->notifyOperationReplaced(op, newOp);
// Replace all result uses. Also notifies the listener of modifications.
- replaceAllUsesWith(op, newOp->getResults());
+ replaceAllOpUsesWith(op, newOp->getResults());
// Erase op and notify listener.
eraseOp(op);
diff --git a/mlir/lib/Transforms/Utils/RegionUtils.cpp b/mlir/lib/Transforms/Utils/RegionUtils.cpp
index eff8acdfb33d20..e25867b527b716 100644
--- a/mlir/lib/Transforms/Utils/RegionUtils.cpp
+++ b/mlir/lib/Transforms/Utils/RegionUtils.cpp
@@ -161,7 +161,7 @@ SmallVector<Value> mlir::makeRegionIsolatedFromAbove(
rewriter.setInsertionPointToStart(newEntryBlock);
for (auto *clonedOp : clonedOperations) {
Operation *newOp = rewriter.clone(*clonedOp, map);
- rewriter.replaceUsesWithIf(clonedOp, newOp->getResults(), replaceIfFn);
+ rewriter.replaceOpUsesWithIf(clonedOp, newOp->getResults(), replaceIfFn);
}
rewriter.mergeBlocks(
entryBlock, newEntryBlock,
``````````
</details>
https://github.com/llvm/llvm-project/pull/84589
More information about the Mlir-commits
mailing list