JIT: fix x86 EH issue when removing an empty try #115731
Conversation
…tinuation Morph may have already cleared out the continuation block, so we won't find a GT_END_LFIN there. Fixes dotnet#115700
|
@dotnet/jit-contrib PTAL |
There was a problem hiding this comment.
Pull Request Overview
This PR addresses an x86 EH issue in the JIT by tolerating the case where the try/finally block has an unreachable continuation (e.g., when the finally throws and the continuation is cleared). The key changes include adding a new test to trigger the scenario and modifying the continuation cleanup logic in the JIT to account for empty throw blocks.
Reviewed Changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| src/tests/JIT/Regression/JitBlue/Runtime_115700/Runtime_115700.csproj | New project file for the regression test |
| src/tests/JIT/Regression/JitBlue/Runtime_115700/Runtime_115700.cs | Test case to reproduce the condition where the finally throws |
| src/coreclr/jit/fgehopt.cpp | Update to fgCleanupContinuation to tolerate an empty BBJ_THROW continuation |
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
|
Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch |
amanasifkhalid
left a comment
There was a problem hiding this comment.
LGTM. Looks like jit-format is broken again?
|
Formatting error is in codegenxarch, unrelated. I can fix it here I guess. |
|
Lots of time outs.. .this change should be very safe an unblocks formatting, so going to merge. |
|
/ba-g lots of timeouts |
If the try/finally has an unreachable continuation (eg the finally throws), morph may have already cleared out the continuation block, so we won't find a GT_END_LFIN there. Tolerate.
Fixes #115700