Re: [PATCH] mm/userfaultfd: Fix release hang over concurrent GUP

From: Peter Xu
Date: Wed Mar 12 2025 - 18:28:36 EST


On Wed, Mar 12, 2025 at 03:04:11PM -0700, Andrew Morton wrote:
> It would be good to have a Fixes: target but this bug seems to be so
> old that a bare cc:stable should be OK?

Yes I agree that should work for us, sorry I forgot to try attach that from
time to time..

AFAIU, such unwanted behavior (either GUP hangs, or if before commit
656710a60e36, the behavior could be a SIGBUS returned to GUP racing with
the userfaultfd_release()) should exist since the 1st day, but I didn't
verify the original commit.

Thanks,

--
Peter Xu