Re: BUG: drop_pagecache_sb vs kjournald lockup

[Jan Kara]

On Wed 19-03-08 07:46:46, David Chinner wrote:
> On Tue, Mar 18, 2008 at 02:43:26PM +0100, Jan Kara wrote:
> > > 2.6.25-rc3, 4p ia64, ext3 root drive.
> > > 
> > > I was running an XFS stress test on one of the XFS partitions on
> > > the machine (zero load on the root ext3 drive), when the system
> > > locked up in kjournald with this on the console:
> > > 
> > > BUG: spinlock lockup on CPU#2, kjournald/2150, a000000100e022e0
> > > 
> >   <snip traces>
> 
> <snip other stuff>
> 
> > > Anyone know the reason why drop_pagecache_sb() uses such a brute-force
> > > mechanism to free up clean page cache pages?
> >   Yes, we know that drop_pagecache_sb() has locking issues but since it
> > is intended to be used for debugging purposes only, nobody cared enough
> > to fix it. Completely untested patch below if you dare to try ;)
> 
> It may be intended for debuging purposes, but it does get used in
> production HPC environments (a lot!). I guess I've never seen this
> lockup before because SGI customers don't use ext3, but they have
> complained about the system "stopping" while drop_caches is executed.
> This locking ..... strategy would explain it, though.
  ;) But in case your customers use it in production, shouldn't you push
for a better interface for such feature? Just a thought...

> I'll try the patch, but I can't guarantee anything - I only saw this
> lockup once in about 18 hours when dropping caches every 2 seconds.
  Well, if you enable lockdep, it should warn you about possible problems
much earlier (at least we already got several reports of lockdep warnings
when using drop_caches).

								Honza
-- 
Jan Kara <jack@xxxxxxx>
SUSE Labs, CR
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/