Re: BUG at mm/slab.c:2990 with 2.6.33-rc5-tuxonice

[Pekka Enberg]

Pedro Ribeiro kirjoitti:
> 2010/1/26 Pedro Ribeiro <pedrib@xxxxxxxxx>:
> > 2010/1/26 Pekka Enberg <penberg@xxxxxxxxxxxxxx>:
> > > Matt Mackall wrote:
> > > > On Wed, 2010-01-27 at 07:39 +1100, Nigel Cunningham wrote:
> > > > > Hi.
> > > > >
> > > > > Pedro Ribeiro wrote:
> > > > > > Hi,
> > > > > >
> > > > > > I hit a bug at mm/slab.c:2990 with .33-rc5.
> > > > > > Unfortunately nothing more is available than a screen picture with a
> > > > > > crash dump, although it is a good one.
> > > > > > The bug was hit almost at the end of a hibernation cycle with
> > > > > > Tux-on-Ice, while saving memory contents to an encrypted swap
> > > > > > partition.
> > > > > >
> > > > > > The image is here http://img264.imageshack.us/img264/9634/mmslab.jpg
> > > > > > (150 kb)
> > > > > >
> > > > > > Hopefully it is of any use for you. Please let me know if you need any
> > > > > > more info.
> > > > > Looks to me to be completely unrelated to TuxOnIce - at least at a first
> > > > > glance.
> > > > >
> > > > > Ccing the slab allocator maintainers listed in MAINTAINERS.
> > > > Not sure if this will do us any good, it's the second oops.
> > > Looks like slab corruption to me which is usually not a slab bug but caused
> > > by buggy callers. Is CONFIG_DEBUG_SLAB enabled?
> > I have enabled it and compiled the kernel. As soon as I hit the bug, I
> > will send a photo here.
> >
> > Regards,
> > Pedro
>
> The pic is here.
> http://img43.imageshack.us/img43/3644/dsc01061ko.jpg
>
> There was a buttload of output before that, which I tried capturing in
> video, but its too crappy to post.

Can you try passing "pause_on_oops=15" as kernel parameter? It should 
delay the next oops for 15 seconds so there's enough time to take a 
picture of the first one?

Also, you could try CONFIG_SLUB and passing "slub_debug" as kernel 
parameter to get nicer diagnostics of the bug.

> Do you know if/where I can get patches for KDB in .33?

Nope, sorry.

			Pekka
--
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/