Skip to content

Make longjmp over managed frames work #111259

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 8 commits into from
Jan 21, 2025
Merged

Make longjmp over managed frames work #111259

Changes from 1 commit
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
82b67ac
Make longjmp over managed frames work
janvorli Jan 10, 2025
3a019a2
Fully simulate legacy EH for longjmp on Windows
janvorli Jan 17, 2025
ec133dc
Fix Unix build
janvorli Jan 17, 2025
a7366aa
Fix bad merge
janvorli Jan 20, 2025
19b02b1
Fix regression issue id in the test path / name
janvorli Jan 20, 2025
7fe2b47
Add comment explaining this is best effort for back compat
janvorli Jan 21, 2025
2e11229
Update src/coreclr/vm/exceptionhandling.cpp
jkotas Jan 21, 2025
e0a9b38
Update src/coreclr/vm/exceptionhandling.cpp
jkotas Jan 21, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 7 additions & 0 deletions src/coreclr/vm/exceptionhandling.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -932,6 +932,13 @@ ProcessCLRExceptionNew(IN PEXCEPTION_RECORD pExceptionRecord,

Thread* pThread = GetThread();

if (pExceptionRecord->ExceptionCode == STATUS_LONGJUMP)
{
// This is an exception used to unwind during longjmp function. We just let it pass through managed
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we need to unwind Frames to avoid corrupting runtime data structures?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

And if we are unwinding Frames, is it that hard to call finallys too to match the old behavior?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I had it implemented by propagating the exception through the managed frames as usual, so the finallys were called, and then re-issued the longjmp after processing them. The only missing part to do was preventing calling catch handlers. The old EH doesn't call those "naturally", as the longjmp runs only the unwind pass.
And then you've mentioned it was not supported, so I have deleted that code :-).
But anyways, I can change it back to that, the change wasn't complicated. The only thing is that the Unix handling would then be different.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The only thing is that the Unix handling would then be different.

longjump over managed code does not and cannot work well on Unix, so I do not think it is a problem

// frames without any interaction.
return ExceptionContinueSearch;
}

if (pThread->HasThreadStateNC(Thread::TSNC_ProcessedUnhandledException))
{
if ((pExceptionRecord->ExceptionFlags & EXCEPTION_UNWINDING))
Expand Down
Loading