Re: [GIT PULL] SLAB changes for v3.10

[Glauber Costa <glommer@parallels.com>]

On 05/08/2013 10:20 AM, Pekka Enberg wrote:
>> > Hmm it's actually CONFIG_PROVE_LOCKING=y that causes the problem,
>> > not CONFIG_DEBUG_SPINLOCK=y. CONFIG_DEBUG_SPINLOCK=y was just selected
>> > by CONFIG_PROVE_LOCKING=y in my non-booting .config. I can now fix my
>> > non-booting .config by removing PROVE_LOCKING and DEBUG_SPINLOCK,
>> > but I cannot break my other working .config by adding PROVE_LOCKING
>> > and DEBUG_SPINLOCK. Hmm.
> That's almost certainly a slab bootstrap issue. Lockdep et al just
> change struct sizes which is why the problem shows up. Is this with
> SLAB or SLUB? Christoph, Glauber, care to take a look?
> 
>>>> >> > 

My first guess is that it hit a NULL cache. Being a NULL pointer
dereference, the thing among all that has the biggest chances of being
NULL and accessed unconditionally is the cache pointer itself.

Due to the size being too big. But if that were the case, he would have
hit the WARN_ON recently introduced:

                if (WARN_ON_ONCE(size > KMALLOC_MAX_SIZE))
                        return NULL;


Is this WARN hit ?