-
Notifications
You must be signed in to change notification settings - Fork 12.7k
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
Couldn't join subrange unreachable hit #32499
Comments
I've encountered this too for AMDGPU. See bug 33524 I've got a potential fix for this (although I'm not sure the approach I've taken is necessarily the right one) which I've uploaded for review (see bug 33524 for details). |
The latest patch for the fix for 33524 (https://reviews.llvm.org/D34391) also appears to fix the second example here (Another example). Attachment 18663 |
Another example of this type of bug |
My fix for the attachment I've just uploaded appears to also fix the first example uploaded to this bug. |
Issue appears to be during the subrange join phase. %other = COPY %ext A check already exists in JoinVals::analyzeValue but didn't work when in |
The test cases here don’t seem useful anymore. They compile for me once manually upgraded to trunk compatible IR, although it’s likely the original MIR might still fail |
Matt - do you want to close this one then? I'd expect the examples to pass given that there have been several fixes submitted around this area. I think the comments indicate that the examples here are expected to pass with latest trunk - given that the examples represented here were replicated in tests used to debug and implement fixes. It might be worth either closing this bug and starting a new one for new cases of subrange failures or updating this one with a fresh example if you've discovered a different case. Either way, I think the examples here might be a little bit stale. |
Extended Description
The text was updated successfully, but these errors were encountered: