Re: [PATCH] ext4: fix decoding of raw_inode timestamps

[Jeff Layton]

On Wed, 2023-07-12 at 13:52 -0400, Theodore Ts'o wrote:
> On Wed, Jul 12, 2023 at 11:02:49AM -0400, Jeff Layton wrote:
> > When we covert a timestamp from raw disk format, we need to consider it
> > to be signed, as the value may represent a date earlier than 1970. This
> > fixes generic/258 on ext4.
> > 
> > Cc: Jan Kara <jack@xxxxxxx>
> > Fixes: f2ddb05870fb ("ext4: convert to ctime accessor functions")
> > Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
> 
> Acked-by: Theodore Ts'o <tytso@xxxxxxx>
> 
> Thanks for the fix!
> 
> It had been on my list to checking to see if the ext4 kunit tests
> would pass, since Jan had mentioned that he had done the work to make
> sure the ext4 kunit test would compile, but he hadn't gotten around to
> try run the kunit test.  Unfortunately, I hadn't gotten to it.
> 
> I *think* the ext4 kunit tests should have caught this as well; out of
> curiosity, have you tried running the ext4 kunit tests either before
> or after this patch?  If so, what were your findings?
> 
> Cheers,
> 
> 					- Ted

No, I haven't. I'm running fstests on it now. Is there a quickstart for
running those tests?
-- 
Jeff Layton <jlayton@xxxxxxxxxx>