Re: linux-next: Tree for July 7

On Thu, 15 Jul 2010 10:46:39 +1000
Dave Chinner <david@fromorbit.com> wrote:

> On Wed, Jul 14, 2010 at 05:27:41PM -0700, Andrew Morton wrote:
> > On Thu, 15 Jul 2010 09:45:13 +1000
> > Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> > 
> > > Hi Andrew, Al,
> > > 
> > > On Wed, 14 Jul 2010 15:20:42 -0700 Andrew Morton <akpm@linux-foundation.org> wrote:
> > > >
> > > > Bug is still there.  Reverting ""Make ->drop_inode() just return
> > > > whether inode needs to be dropped" makes it go away.
> > > > 
> > > > Al, please don't just dump crap in linux-next and then ignore bug reports. 
> > > > People are trying to get work done here.
> > > 
> > > I will revert that commit in linux-next today if it is still there (and
> > > not fixed).
> > 
> > Reverting the patch didn't help - the patch moved the BUG_ON() from
> > generic_delete_inode() into __iput_final().  With the patch reverted
> > the kernel still goes BUG over the same test in generic_delete_inode().
> > 
> > It could be that the bug was caused by fs-writeback.c changes.  David,
> > do you have time to take a look at this?
> 
> I'll have a look at it. 

Thanks.

> What filesystem(s) are you seeing this on?

ext3 at least.

> Any particular workload that triggers it?

Booting FC6 :( I get an enormous storm of these things.




whee, progress!

With Rusty's patch for the modules bug, and reverting Greg
Vandal-Hartman's "driver core: remove CONFIG_SYSFS_DEPRECATED" and
deleting the BUG_ON from generic_delete_inode(), I have a login prompt!
Admittedly I don't have any networking any more, but that seems a
minor quibble.

What a clusterfsck.