[Bisected] Corruption of root fs during git bisect of drm system hang

[Dave Chinner]

On Sat, Jul 20, 2013 at 11:48:36AM +1000, Dave Chinner wrote:
> On Fri, Jul 19, 2013 at 02:22:35PM +0200, Markus Trippelsdorf wrote:
> > On 2013.07.15 at 08:47 +0200, Markus Trippelsdorf wrote:
> > I've bisected this issue to the following commit:
> > 
> >  commit cca9f93a52d2ead50b5da59ca83d5f469ee4be5f
> >  Author: Dave Chinner <dchinner at redhat.com>
> >  Date:   Thu Jun 27 16:04:49 2013 +1000
> > 
> >      xfs: don't do IO when creating an new inode
> >          
> > Reverting this commit on top of the Linus tree "solves" all problems for
> > me. IOW I no longer loose my KDE and LibreOffice config files during a
> > crash. Log recovery now works fine and xfs_repair shows no issues.
> 
> Thanks for bisecting this, Marcus.
> 
> I'll admit, right now it doesn't make a lot of sense to me - I don't
> immediately see a connection between not reading an inode during the
> create phase and unlinked list and directory corruption after a
> crash. But now you've identified a change that might be the cause,
> I have an avenue of investigation I can follow.
> 
> Indeed, in the time I've taken to write this mail I've thought of
> 2-3 possible causes that I need to investigate....

Hi Markus, can you test the patch I just posted to the list titled
"xfs: di_flushiter considered harmful" and see if it fixes your
problem? Archive link here:

http://oss.sgi.com/pipermail/xfs/2013-July/028331.html

Cheers,

Dave.
-- 
Dave Chinner
david at fromorbit.com