[mlir][Affine] Fix crash in affine-loop-fusion pass by guarding against an empty op list

[ayokunle321]

Related: #139231

This patch fixes a crash in the affine-loop-fusion pass when getInnermostCommonLoop is called with an empty list of operations.

The function expects at least one op to analyze, and passing an empty array of ops causes an assertion failure. This change ensures the pass checks for an empty op array before calling getInnermostCommonLoop.

@bondhugula @matthias-springer

[llvmbot]

@llvm/pr-subscribers-mlir-affine

Author: Ayokunle Amodu (ayokunle321)

Changes

Related: #139231

This patch fixes a crash in the affine-loop-fusion pass when getInnermostCommonLoop is called with an empty list of operations.

The function expects at least one op to analyze, and passing an empty array of ops causes an assertion failure. This change ensures the pass checks for an empty op array before calling getInnermostCommonLoop.

@bondhugula @matthias-springer

---

Full diff: https://github.com/llvm/llvm-project/pull/144841.diff

1 Files Affected:

• (modified) mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp (+2)

diff --git a/mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp b/mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp
index 4b4eb9ce37b4c..5c5d041ed6bdc 100644
--- a/mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp
+++ b/mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp
@@ -997,6 +997,8 @@ struct GreedyFusion {
           if (producerConsumerMemrefs.count(
                   cast<AffineWriteOpInterface>(op).getMemRef()))
             dstMemrefOps.push_back(op);
+        if (dstMemrefOps.size() == 0)
+          continue;
         unsigned dstLoopDepthTest =
             getInnermostCommonLoopDepth(dstMemrefOps) - numSurroundingLoops;
 

[llvmbot]

@llvm/pr-subscribers-mlir

Author: Ayokunle Amodu (ayokunle321)

Changes

Related: #139231

This patch fixes a crash in the affine-loop-fusion pass when getInnermostCommonLoop is called with an empty list of operations.

The function expects at least one op to analyze, and passing an empty array of ops causes an assertion failure. This change ensures the pass checks for an empty op array before calling getInnermostCommonLoop.

@bondhugula @matthias-springer

---

Full diff: https://github.com/llvm/llvm-project/pull/144841.diff

1 Files Affected:

• (modified) mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp (+2)

diff --git a/mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp b/mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp
index 4b4eb9ce37b4c..5c5d041ed6bdc 100644
--- a/mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp
+++ b/mlir/lib/Dialect/Affine/Transforms/LoopFusion.cpp
@@ -997,6 +997,8 @@ struct GreedyFusion {
           if (producerConsumerMemrefs.count(
                   cast<AffineWriteOpInterface>(op).getMemRef()))
             dstMemrefOps.push_back(op);
+        if (dstMemrefOps.size() == 0)
+          continue;
         unsigned dstLoopDepthTest =
             getInnermostCommonLoopDepth(dstMemrefOps) - numSurroundingLoops;
 

[github-actions]

⚠️ We detected that you are using a GitHub private e-mail address to contribute to the repo.
Please turn off Keep my email addresses private setting in your account.
See LLVM Developer Policy and LLVM Discourse for more information.

[ftynse]

Please add a test and reupload with a proper email address as instructed.

[ftynse]

clang-tidy: use .empty() instead of comparing size with 0.