On 2013.07.23 at 10:07 -0500, Ben Myers wrote:
> Hi Markus,
> On Tue, Jul 23, 2013 at 12:42:59PM +0200, Markus Trippelsdorf wrote:
> > On 2013.07.22 at 14:48 -0500, Mark Tinguely wrote:
> > > On 07/22/13 10:15, Markus Trippelsdorf wrote:
> > > > On 2013.07.22 at 09:40 -0500, Mark Tinguely wrote:
> > > >> On 07/22/13 06:07, Markus Trippelsdorf wrote:
> > > >>> On 2013.07.22 at 20:18 +1000, Dave Chinner wrote:
> > > >> It seems to me that since we cannot fix this for inode 1/2, then
> > > >> besides
> > > >> this patch we have to revert patch cca9f93a52d and make it inode 3+ /
> > > >> superblock 5+ (crc) dependent.
> > > >
> > > > Which is exactly what the hunk I've posted does.
> > > >
> > > > Here's the combined patch:
> > >
> > > Following Dave's instruction to recreate this problem, your patch works
> > > with an inode 2 and inode 3 (once I remembered to load the module before
> > > recovery). Dave's patch was successful on inode 3 - again after I
> > > remembered to load the module before recovery.
> > >
> > > Whomever makes the formal patch, consider it reviewed-by me.
> > To get this patch to Linus ASAP, here's the combined patch again.
> > Please apply.
> > Thanks.
> I'd prefer to have Dave take another look at the combined patch before pulling
> this in. Even so, we shouldn't have any trouble getting this into -rc3.
Ok, understood. I'll let Dave handle the rest.
BTW the comment above xfs_iread() probably needs an update, too.