lkml.org 
[lkml]   [2006]   [Jul]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: 2.6.18-rc1-mm1
On Sun, 9 Jul 2006 12:32:27 +0200
"Michal Piotrowski" <michal.k.k.piotrowski@gmail.com> wrote:

> Hi,
>
> On 09/07/06, Andrew Morton <akpm@osdl.org> wrote:
> >
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18-rc1/2.6.18-rc1-mm1/
> >
>
> This looks like a problem with cpufreq.
>
> =======================================================
> [ INFO: possible circular locking dependency detected ]
> -------------------------------------------------------
> cpuspeed/1426 is trying to acquire lock:
> (&inode->i_data.tree_lock){.+..}, at: [<c0151dc7>] find_get_page+0x12/0x70
>
> but task is already holding lock:
> (&mm->mmap_sem){----}, at: [<c0116cab>] do_page_fault+0x10d/0x4ea
>
> which lock already depends on the new lock.
>

rofl. You broke lockdep.

Well. I guess it's barely conceivable that you earlier took an oops while
holding tree_lock, so lockdep decided that mmap_sem nests inside tree_lock.


> the existing dependency chain (in reverse order) is:
>
> -> #1 (cpucontrol){--..}:
> [<c0139a55>] lock_acquire+0x71/0x91
> [<c02ee288>] __mutex_lock_slowpath+0xd2/0x2f5
> [<c02ee4c7>] mutex_lock+0x1c/0x1f
> [<c013dd3b>] __lock_cpu_hotplug+0x34/0x4c
> [<c013dd6c>] lock_cpu_hotplug+0xa/0xc
> [<c029b587>] __cpufreq_driver_target+0x15/0x50
> [<c029c3ca>] cpufreq_governor_performance+0x1a/0x20
> [<c029a89b>] __cpufreq_governor+0x95/0x18c
> [<c029aa72>] __cpufreq_set_policy+0xe0/0x118
> [<c029af49>] cpufreq_set_policy+0x2d/0x6f
> [<c029bc45>] cpufreq_add_dev+0x3ee/0x4f3
> [<c024dccb>] sysdev_driver_register+0x5e/0x9e
> [<c029be70>] cpufreq_register_driver+0x80/0xf4
> [<fdba202a>] 0xfdba202a
> [<c0140f22>] sys_init_module+0xa6/0x21d
> [<c0103179>] sysenter_past_esp+0x56/0x8d
>
> -> #0 (&inode->i_data.tree_lock){.+..}:
> [<c0139a55>] lock_acquire+0x71/0x91
> [<c02ee288>] __mutex_lock_slowpath+0xd2/0x2f5
> [<c02ee4c7>] mutex_lock+0x1c/0x1f
> [<c029b7f2>] store_scaling_governor+0x14a/0x1a2
> [<c029b223>] store+0x37/0x48
> [<c01a9f4b>] sysfs_write_file+0xa6/0xcc
> [<c0172dab>] vfs_write+0xc9/0x172
> [<c017341d>] sys_write+0x3b/0x71
> [<c0103179>] sysenter_past_esp+0x56/0x8d

Straightforward ab/ba deadlock between cpufreq_policy.lock and
lock_cpu_hotplug(). But lockdep got confused about the identity of the
lock.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2006-07-09 12:55    [from the cache]
©2003-2011 Jasper Spaans