JIT: Avoid recording uncanonicalizable loops #115223
Merged
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
The example has a loop whose header is the beginning of a 'finally' clause and whose backedge is a CALLFINALLY edge. It is not possible to canonicalize such a loop: creating a new preheader will result in a new beginning of the finally clause, and by IR invariants the CALLFINALLY edge must jump to it. But that means the CALLFINALLY edge is no longer a backedge. The loop structure here is spurious and arises because of two-pass EH, so it is unlikely we will lose much by just discarding these cases.
|
Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch |
|
cc @dotnet/jit-contrib PTAL @AndyAyersMS @BruceForstall |
There was a problem hiding this comment.
Pull Request Overview
This PR addresses a JIT issue where loops that have their header in a 'finally' clause and backedges from CALLFINALLY edges cannot be canonicalized, and discards these cases.
- Updated the loop recording in FlowGraphNaturalLoops::Find() to include a canonicalizability check.
- Added the IsLoopCanonicalizable() helper function in flowgraph.cpp and declared it in compiler.h, with an updated comment explaining its purpose.
Reviewed Changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.
| File | Description |
|---|---|
| src/coreclr/jit/flowgraph.cpp | Added IsLoopCanonicalizable() and updated loop condition in Find() |
| src/coreclr/jit/compiler.h | Declared IsLoopCanonicalizable() and updated loop header comment |
Comments suppressed due to low confidence (1)
src/coreclr/jit/flowgraph.cpp:4882
- Consider adding unit tests to verify that loops with CALLFINALLY backedges are correctly marked as uncanonicalizable. This will help ensure that future changes do not reintroduce canonicalization issues for such loops.
if (!FindNaturalLoopBlocks(loop, worklist) || !IsLoopCanonicalizable(loop))
AndyAyersMS
approved these changes
May 1, 2025
BruceForstall
approved these changes
May 1, 2025
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 example has a loop whose header is the beginning of a 'finally' clause and whose backedge is a CALLFINALLY edge. It is not possible to canonicalize such a loop: creating a new preheader will result in a new beginning of the finally clause, and by IR invariants the CALLFINALLY edge must jump to it. But that means the CALLFINALLY edge is no longer a backedge.
The loop structure here is spurious and arises because of two-pass EH, so it is unlikely we will lose much by just discarding these cases.
Fix #114869