Re: [BUG] at mm/slab.c:3320

[Nishanth Aravamudan <nacc@us.ibm.com>]

On 09.01.2008 [13:51:42 -0800], Christoph Lameter wrote:
> On Wed, 9 Jan 2008, Nishanth Aravamudan wrote:
> 
> > And given that the original mail has bug at mm/slab.c:3320, I assume we're
> > still hitting the
> > 
> > BUG_ON(ac->avail > 0 || !l3);
> 
> No we are in a different function here.

Ah you're right -- sorry for the noise.

> > Hrm, shouldn't we remove the !l3 bit from the BUG_ON? But even so,
> > unless for some reason the BUG_ON is being checked before the if
> > (!l3), are we hitting (ac->avail > 0)?
> 
> Yes we should remove the !l3 bit. There cannot be any objects in SLABs
> per cpu queue if there is no node structure. per cpu queues can only
> be refilled from the local node, not from foreign nodes. And in this
> particular case there is no memory available from the local node. So
> ac->avail == 0.

Makes sense, thanks for the clarification.

-Nish

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>