Messages in this thread | | | Date | Thu, 12 Jun 2008 10:13:19 -0700 (PDT) | From | Christoph Lameter <> | Subject | Re: repeatable slab corruption with LTP msgctl08 |
| |
On Thu, 12 Jun 2008, Pekka J Enberg wrote:
> I really don't understand how your bufctl chains has so many BUFCTL_END > elements in the first place. It's doesn't look like the memory has been > stomped on (slab->s_mem, for example, is 0xf2906088), so I'd look for a > double kfree() of size 128 somewhere...
Looks pretty strange. Could this be rerun with SLAB_DEBUG or with SLUB with full debugging?
| |