-
Notifications
You must be signed in to change notification settings - Fork 789
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.
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
Fix #17903 #18025
Fix #17903 #18025
Conversation
❗ Release notes required
|
/run fantomas |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
A test would be nice.
Co-authored-by: vzarytovskii <1260985+vzarytovskii@users.noreply.github.com>
Co-authored-by: Petr Pokorny <petr@innit.cz>
f3548b8
to
e7ac187
Compare
It's hitting some awkward ILVerify error, which I don't yet understand. Gonna dig more into it. There might be a flaw in the "can I realloc it" logic, which I introduced. |
Alright, so the failing baseline was due to the offset of the instruction (by 2 bytes), which can be a signal of the local realloc. I would like to run it here a bunch of times to see if it's some kind of "parallel ilxgen" issue with realloc caches. Update I hope it was just a local mistake of me not properly rebuilding FCS/fslib before checking baselines. |
Fixes #17903. Test is missing, checking the waters.
The idea of the fix is pretty dumb. The
copyOfStruct
address is leaking outside the scope which trips the locals reuse algorithm. After looking more into it, and chatting with Don, I decided to markcopyOfStruct
as no-reuse, but keeping it in cgbuf cache.