Fix several issues with interpreter and EH #119784
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
This change fixes few issues I have found while investigating coreclr tests failures when running interpreted * The GC info for reverse pinvoke stubs didn't have the revPInvokeOffset set * When SfiNext reached native marker state, it was doing one more unwind to get to the frame function state. That is benign for non-interpreter scenarios, as it doesn't change the regdisplay, but it is useless due to that. But for interpreter, it can actually incorrectly move to interpreted code. The fix is to not to do that unwind and just return. * The StackFrameIterator::SkipTo that is called when SfiNext collides with the previous EH was missing copying of the first argument register that can hold the InterpreterFrame pointer. * When SfiNext moved out of the last interpreted code frame and the caller of the interpreted code was not managed code (e.g. the CallDescrWorker), the StackFrameIterator's regdisplay was not updated to that context.
There was a problem hiding this comment.
Pull Request Overview
This PR fixes several issues with interpreter and exception handling (EH) in the CoreCLR runtime to address test failures when running interpreted code.
Key changes:
- Fixed GC info for reverse pinvoke stubs by setting the
revPInvokeOffset - Corrected stack frame iterator behavior when reaching native marker states
- Added proper register copying for interpreter frame pointers during stack unwinding
Reviewed Changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| src/coreclr/vm/stackwalk.cpp | Added first argument register copying for interpreter frame addresses and logging for debugging |
| src/coreclr/vm/exceptionhandling.cpp | Fixed stack frame iterator logic for native markers and interpreter frame unwinding |
| src/coreclr/interpreter/compiler.cpp | Added reverse pinvoke frame slot configuration for unmanaged callers only methods |
|
Tagging subscribers to this area: @BrzVlad, @janvorli, @kg |
69 tasks
davidwrighton
approved these changes
Sep 19, 2025
janvorli
added
the
NO-MERGE
|
Adding NO-MERGE until the #119863 gets in. We will need to backport the other to .NET 10 and this would make the backport more complicated. |
janvorli
removed
the
NO-MERGE
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 participants
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 change fixes few issues I have found while investigating coreclr tests failures when running interpreted
revPInvokeOffsetsetSfiNextreached native marker state, it was doing one more unwind to get to the frame function state. That is benign for non-interpreter scenarios, as it doesn't change the regdisplay, but it is useless due to that. But for interpreter, it can actually incorrectly move to interpreted code. The fix is to not to do that unwind and just return.StackFrameIterator::SkipTothat is called whenSfiNextcollides with the previous EH was missing copying of the first argument register that can hold theInterpreterFramepointer.SfiNextmoved out of the last interpreted code frame and the caller of the interpreted code was not managed code (e.g. theCallDescrWorker), theStackFrameIterator's regdisplay was not updated to that context.