Messages in this thread | | | Date | Mon, 21 Apr 2008 15:50:53 +0200 | From | Ingo Molnar <> | Subject | Re: [bug] SCSI/SLUB - latest -git: WARNING: at mm/slub.c:2443 kmem_cache_destroy, scsi_put_host_cmd_pool() |
| |
* Christoph Lameter <clameter@sgi.com> wrote:
> On Sat, 19 Apr 2008, Pekka J Enberg wrote: > > > > [a few .config options were turned off: just accept all the defaults > > > after 'make oldconfig'] > > > > I couldn't spot anything in particular in SLUB which makes me think > > SCSI code simply didn't free all objects before > > scsi_put_host_cmd_pool() called kmem_cache_destroy() to kill the > > cache. > > > > James, does this make sense or should I just look at SLUB harder? > > The WARN is intended to warn that a kmem_cache_destroy was run with > objects not freed.
i suspect if that warn-on triggers more frequently then it might make sense to turn it into a pretty SLUB warning about that cache, with a stackdump at the end. (that way people are not tricked into mistakenly believing that it's a SLUB bug)
Ingo
| |