On Sun, Nov 04, 2007 at 12:19:19PM +0100, Torsten Kaiser wrote:
quoted text
> On 11/2/07, David Chinner <dgc@sgi.com> wrote:
> > That's stalled waiting on the inode cluster buffer lock. That implies
> > that the inode lcuser is already being written out and the inode has
> > been redirtied during writeout.
> >
> > Does the kernel you are testing have the "flush inodes in ascending
> > inode number order" patches applied? If so, can you remove that
> > patch and see if the problem goes away?
> 
> I can now confirm, that I see this also with the current mainline-git-version
> I used 2.6.24-rc1-git-b4f555081fdd27d13e6ff39d455d5aefae9d2c0c
> plus the fix for the sg changes in ieee1394.

Ok, so it's probably a side effect of the writeback changes.

Attached are two patches (two because one was in a separate patchset as
a standalone change) that should prevent async writeback from blocking
on locked inode cluster buffers. Apply the xfs-factor-inotobp patch first.
Can you see if this fixes the problem?

Cheers,

Dave.
-- 
Dave Chinner
Principal Engineer
SGI Australian Software Group