-
Notifications
You must be signed in to change notification settings - Fork 1.1k
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
[NFC-ish] Finish MSAN handling (Refs. #6513) #6516
Conversation
Somehow, initially i missed that there was MSan support, so it might be good to actually mention that we don't need to run any MSan passes here, and that we didn't forget to run them. Secondly, it seems inconsistent not annotate the functions with `Attribute::SanitizeMemory`, like we do for others. I suppose it isn't strictly required, since they are used to actually drive the instrumentation passes, and we don't run MSan pass, but they are also used to disable some LLVM optimizations, and that //might// be important. Or not, but then i suppose there should be a comment about it?
Failures are again unrelated... same arm64/llvm14 failure. |
Requesting a review from @steven-johnson (whenever he's back) to address this. IIRC he implemented MSAN support and I'm not sure why this wouldn't be annotated. |
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.
LGTM pending green buildbots -- I think the arm failure are just stale from bad LLVM revisions, rerunning them now
Thanks for the fix! |
@alexreinking @steven-johnson Thank you! |
Somehow, initially i missed that there was MSan support, so it might be good to actually mention that we don't need to run any MSan passes here, and that we didn't forget to run them. Secondly, it seems inconsistent not annotate the functions with `Attribute::SanitizeMemory`, like we do for others. I suppose it isn't strictly required, since they are used to actually drive the instrumentation passes, and we don't run MSan pass, but they are also used to disable some LLVM optimizations, and that //might// be important. Or not, but then i suppose there should be a comment about it? Co-authored-by: Steven Johnson <[email protected]> (cherry picked from commit 7eb9949)
Somehow, initially i missed that there was MSan support, so it might be good to actually mention that we don't need to run any MSan passes here, and that we didn't forget to run them. Secondly, it seems inconsistent not annotate the functions with `Attribute::SanitizeMemory`, like we do for others. I suppose it isn't strictly required, since they are used to actually drive the instrumentation passes, and we don't run MSan pass, but they are also used to disable some LLVM optimizations, and that //might// be important. Or not, but then i suppose there should be a comment about it? Co-authored-by: Steven Johnson <[email protected]> (cherry picked from commit 7eb9949)
Somehow, initially i missed that there was MSan support,
so it might be good to actually mention that we don't need to run
any MSan passes here, and that we didn't forget to run them.
Secondly, it seems inconsistent not annotate the functions
with
Attribute::SanitizeMemory
, like we do for others.I suppose it isn't strictly required, since they are used
to actually drive the instrumentation passes, and we don't
run MSan pass, but they are also used to disable some LLVM optimizations,
and that //might// be important. Or not, but then i suppose
there should be a comment about it?
Let me know if i should replace the second change with a comment too.