On Fri, Dec 06, 2024 at 04:32:04PM -0800, Darrick J. Wong wrote:
> On Fri, Dec 06, 2024 at 04:25:22PM -0800, Luis Chamberlain wrote:
> > On Thu, Oct 24, 2024 at 01:51:05PM +1100, Dave Chinner wrote:
> > > From: Dave Chinner <dchinner@xxxxxxxxxx>
> > >
> > > It's just read in from the superblock and used without doing any
> > > validity checks at all on the value.
> > >
> > > Fixes: fb4f2b4e5a82 ("xfs: add sparse inode chunk alignment superblock field")
> > > Signed-off-by: Dave Chinner <dchinner@xxxxxxxxxx>
> >
> > This is 59e43f5479cce106d71c0b91a297c7ad1913176c on v6.13-r1 now.
> >
> > This commit broke mounting 32k and 64k bs filesystems on 4k page size systems.
> > Oddly, it does not break 16k or 8k bs. I took a quick glance and I can't
> > easily identify a fix.
> >
> > I haven't had a chance yet to find a large page size system to see if
> > 32k page size and 64k page size systems are affected as well.
> >
> > CIs in place did not pick up on this given fstests check script just
> > bails out right away, we don't annotate this as a failure on fstests and
> > the tests don't even get listed as failures on xunit. You'd have to have
> > a trained curious eye to just monitor CIs and verify that all hosts
> > actually were chugging along. I suppose we can enhance this by just
> > assuming hosts which don't have results are assumed to be a failure.
> >
> > However if we want to enahnce this on fstests so that in the future we
> > pick up on these failures more easily it would be good time to evaluate
> > that now too.
>
> Known bug, already patched here:
> https://lore.kernel.org/linux-xfs/20241126202619.GO9438@frogsfrogsfrogs/
>
> and PR to the release manager here:
> https://lore.kernel.org/linux-xfs/173328206660.1159971.4540485910402305562.stg-ugh@frogsfrogsfrogs/
Woot, thanks!
Luis
