Re: ext3_orphan_del may double-decrement bh->b_count

[Chris Mason <mason@suse.com>]

On Wed, 2004-06-02 at 21:00, Andrew Morton wrote:
> Chris Mason <mason@suse.com> wrote:
>  
> > > What was the "other bug"?
> > 
> > We've got many names for it, but none that could be posted here ;-)
> 
> intrigued.
> 
Nah, just a string of curses unfit for the general public, or even lkml.

> > Looks like HP came up with a simplified test case:
> > 
> > http://sourceforge.net/mailarchive/forum.php?thread_id=4536665&forum_id=6379
> 
> hm, I never received that.  If I'd known I wouldn't have removed the
> buffer_error() stuff.
> 
I've been trying to blame this one on various other bugs until
recently.  The test case posted above is the first simple test that is
supposed to be able to reproduce.  Since ext2 also sees data corruption
issues, it's either not ext3 specific or two different bugs.

> > I've got machines trying to reproduce now.
> 
> You need the buffer-tracing patch.  This is against 2.6.7-rc2.  It should
> spit a nice trace when you hit the problem.  It'll tell us how that buffer
> got itself not uptodate.

Thanks.  jeffm had worked out something similar that stored the EIP of
each bit operation, the uptodate bit seems to have turned off all on its
own.  Once we can reproduce reliably on local boxes, we'll start
layering on the debugging code.  

No triggers yet, I might have to grab a bigger machine in the morning.

-chris