Messages in this thread | | | From | Sitsofe Wheeler <> | Subject | Re: BUG: unable to handle kernel NULL pointer dereference (drm_getunique) | Date | Thu, 05 Jun 2008 08:08:34 +0100 |
| |
Hello,
Johannes Weiner wrote:
> Sitsofe Wheeler <sitsofe@yahoo.com> writes: > >> [ 4305.767503] Pid: 8373, comm: unix2_chkpwd Tainted: G W >> [ (2.6.26-rc4-next-20080604skw #177) > > There was a warning before that oops already, could you post that too?
The warning is: [ 42.949862] ------------[ cut here ]------------ [ 42.949873] WARNING: at kernel/lockdep.c:2680 check_flags+0x8a/0x12d() [ 42.949880] Modules linked in: [ 42.949887] Pid: 5, comm: watchdog/0 Not tainted 2.6.26-rc4-next-20080604skw #178 [ 42.949898] [<c01226f0>] warn_on_slowpath+0x41/0x6a [ 42.949917] [<c013c584>] ? trace_hardirqs_off+0xb/0xd [ 42.949935] [<c013e3b4>] ? trace_hardirqs_on+0xb/0xd [ 42.951740] [<c013e36d>] ? trace_hardirqs_on_caller+0xe8/0x124 [ 42.951759] [<c013e3b4>] ? trace_hardirqs_on+0xb/0xd [ 42.951778] [<c011c8f7>] ? hrtick_set+0xce/0xd6 [ 42.951799] [<c013c1d2>] check_flags+0x8a/0x12d [ 42.951816] [<c013f2f8>] lock_acquire+0x3b/0x89 [ 42.951830] [<c0370c0a>] _read_lock+0x1c/0x49 [ 42.951848] [<c01531fe>] ? watchdog+0x97/0x1a9 [ 42.951867] [<c0153167>] ? watchdog+0x0/0x1a9 [ 42.951883] [<c01531fe>] watchdog+0x97/0x1a9 [ 42.951900] [<c0153167>] ? watchdog+0x0/0x1a9 [ 42.951916] [<c01329a0>] kthread+0x3b/0x63 [ 42.951933] [<c0132965>] ? kthread+0x0/0x63 [ 42.951952] [<c01038ab>] kernel_thread_helper+0x7/0x10 [ 42.951969] ======================= [ 42.951976] ---[ end trace 199a1fe68fc13dfd ]--- [ 42.951983] possible reason: unannotated irqs-on. [ 42.951991] irq event stamp: 18 [ 42.951998] hardirqs last enabled at (17): [<c013e3b4>] trace_hardirqs_on+0xb/0xd [ 42.952018] hardirqs last disabled at (18): [<c013c584>] trace_hardirqs_off+0xb/0xd [ 42.952043] softirqs last enabled at (0): [<c0120e1b>] copy_process+0x2dd/0xf9a [ 42.952062] softirqs last disabled at (0): [<00000000>] 0x0
I've mentioned it before ( http://lkml.org/lkml/2008/5/29/502 ) and even bisected it ( http://lkml.org/lkml/2008/6/4/313 ). I guess the problem is that everyone is busy and it's hard to tell if people didn't see it, if it's just not important. Perhaps there needs to be a rule that lockdep warnings always go into bugzilla (or that if you are willing to chase an issue for a minimum of six months it should be bugzillad...).
Thanks for your work looking at the real bug though!
-- Sitsofe | http://sucs.org/~sits/
| |