Messages in this thread | | | From | David Howells <> | Subject | Re: [CRED bug?] 2.6.29-rc3 don't survive on stress workload | Date | Wed, 11 Feb 2009 12:54:06 +0000 |
| |
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:
> f6 : 1003e6b6b6b6b6b6b6b6b f7 : 0ffe6af8e85a500000000
I'd guess you now have CONFIG_DEBUG_SLAB turned on. That looks like slab poisoning in register f6.
> r17 : 00000000dead4ead r18 : a000000100cd2274 r19 : a000000100d6b6c0
And that looks like SPINLOCK_MAGIC in r17, so I'd guess something just successfully validated a spinlock, possibly atomic_dec_and_lock() in free_uid().
This may mean that the user_struct being released is valid, but the user_namespace it refers to is already released.
David
| |