Hi Eric,
rumours are that you could be responsible for the following lockdep message:
Happens on cpu hotplug on s390:
=============================================
[ INFO: possible recursive locking detected ]
2.6.33-rc4-00399-g24bc734 #1
---------------------------------------------
bash/2576 is trying to acquire lock:
(s_active){++++.+}, at: [<000000000016c506>] sysfs_addrm_finish+0x5a/0xa8
but task is already holding lock:
(s_active){++++.+}, at: [<000000000016c90c>] sysfs_get_active_two+0x64/0x88
other info that might help us debug this:
5 locks held by bash/2576:
#0: (&buffer->mutex){+.+.+.}, at: [<000000000016a912>] sysfs_write_file+0x4a/0x1a8
#1: (s_active){++++.+}, at: [<000000000016c90c>] sysfs_get_active_two+0x64/0x88
#2: (s_active){++++.+}, at: [<000000000016c8e6>] sysfs_get_active_two+0x3e/0x88
#3: (cpu_add_remove_lock){+.+.+.}, at: [<00000000003dd318>] cpu_down+0x44/0x88
#4: (cpu_hotplug.lock){+.+.+.}, at: [<0000000000056eee>] cpu_hotplug_begin+0x3e/0x74
stack backtrace:
CPU: 1 Not tainted 2.6.33-rc4-00399-g24bc734 #1
Process bash (pid: 2576, task: 000000003f0e25d0, ksp: 000000003d9339d8)
000000003d9338b0 000000003d933830 0000000000000002 0000000000000000
000000003d9338d0 000000003d933848 000000003d933848 00000000003e4996
0000000000000000 0000000000000000 000000003f0e2cb8 0000000000835208
000000000000000d 000000000000000c 000000003d933898 0000000000000000
0000000000000000 00000000000172aa 000000003d933830 000000003d933870
Call Trace:
([<00000000000171b2>] show_trace+0xee/0x144)
[<000000000008a1a0>] validate_chain+0xa2c/0x1100
[<000000000008ad84>] __lock_acquire+0x510/0xbfc
[<000000000008b51c>] lock_acquire+0xac/0xd4
[<000000000016bba8>] sysfs_deactivate+0x94/0x108
[<000000000016c506>] sysfs_addrm_finish+0x5a/0xa8
[<0000000000169f38>] sysfs_hash_and_remove+0x64/0x9c
[<000000000016dbae>] sysfs_remove_group+0x76/0x12c
[<00000000003dfaae>] smp_cpu_notify+0xae/0xc8
[<00000000003eb652>] notifier_call_chain+0x56/0x98
[<000000000007a1b2>] raw_notifier_call_chain+0x2a/0x3c
[<00000000003dd132>] _cpu_down+0x1aa/0x34c
[<00000000003dd356>] cpu_down+0x82/0x88
[<00000000003ddac0>] store_online+0x5c/0xc8
[<000000000016a9be>] sysfs_write_file+0xf6/0x1a8
[<00000000000ffd5c>] vfs_write+0xac/0x178
[<00000000000fff24>] SyS_write+0x58/0xa8
[<00000000000295b6>] sysc_noemu+0x10/0x16
On Thu, Jan 21, 2010 at 11:53 PM, Heiko Carstens
<[email protected]> wrote:
> Hi Eric,
>
> rumours are that you could be responsible for the following lockdep message:
>
> Happens on cpu hotplug on s390:
I can see the same warnings on x86_64
>
> =============================================
> [ INFO: possible recursive locking detected ]
> 2.6.33-rc4-00399-g24bc734 #1
> ---------------------------------------------
> bash/2576 is trying to acquire lock:
> (s_active){++++.+}, at: [<000000000016c506>] sysfs_addrm_finish+0x5a/0xa8
>
> but task is already holding lock:
> (s_active){++++.+}, at: [<000000000016c90c>] sysfs_get_active_two+0x64/0x88
>
> other info that might help us debug this:
> 5 locks held by bash/2576:
> #0: (&buffer->mutex){+.+.+.}, at: [<000000000016a912>] sysfs_write_file+0x4a/0x1a8
> #1: (s_active){++++.+}, at: [<000000000016c90c>] sysfs_get_active_two+0x64/0x88
> #2: (s_active){++++.+}, at: [<000000000016c8e6>] sysfs_get_active_two+0x3e/0x88
> #3: (cpu_add_remove_lock){+.+.+.}, at: [<00000000003dd318>] cpu_down+0x44/0x88
> #4: (cpu_hotplug.lock){+.+.+.}, at: [<0000000000056eee>] cpu_hotplug_begin+0x3e/0x74
>
> stack backtrace:
> CPU: 1 Not tainted 2.6.33-rc4-00399-g24bc734 #1
> Process bash (pid: 2576, task: 000000003f0e25d0, ksp: 000000003d9339d8)
> 000000003d9338b0 000000003d933830 0000000000000002 0000000000000000
> 000000003d9338d0 000000003d933848 000000003d933848 00000000003e4996
> 0000000000000000 0000000000000000 000000003f0e2cb8 0000000000835208
> 000000000000000d 000000000000000c 000000003d933898 0000000000000000
> 0000000000000000 00000000000172aa 000000003d933830 000000003d933870
> Call Trace:
> ([<00000000000171b2>] show_trace+0xee/0x144)
> [<000000000008a1a0>] validate_chain+0xa2c/0x1100
> [<000000000008ad84>] __lock_acquire+0x510/0xbfc
> [<000000000008b51c>] lock_acquire+0xac/0xd4
> [<000000000016bba8>] sysfs_deactivate+0x94/0x108
> [<000000000016c506>] sysfs_addrm_finish+0x5a/0xa8
> [<0000000000169f38>] sysfs_hash_and_remove+0x64/0x9c
> [<000000000016dbae>] sysfs_remove_group+0x76/0x12c
> [<00000000003dfaae>] smp_cpu_notify+0xae/0xc8
> [<00000000003eb652>] notifier_call_chain+0x56/0x98
> [<000000000007a1b2>] raw_notifier_call_chain+0x2a/0x3c
> [<00000000003dd132>] _cpu_down+0x1aa/0x34c
> [<00000000003dd356>] cpu_down+0x82/0x88
> [<00000000003ddac0>] store_online+0x5c/0xc8
> [<000000000016a9be>] sysfs_write_file+0xf6/0x1a8
> [<00000000000ffd5c>] vfs_write+0xac/0x178
> [<00000000000fff24>] SyS_write+0x58/0xa8
> [<00000000000295b6>] sysc_noemu+0x10/0x16
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
On Thu, Jan 21, 2010 at 11:53 PM, Heiko Carstens
<[email protected]> wrote:
> Hi Eric,
>
> rumours are that you could be responsible for the following lockdep message:
>
> Happens on cpu hotplug on s390:
>
> =============================================
> [ INFO: possible recursive locking detected ]
> 2.6.33-rc4-00399-g24bc734 #1
> ---------------------------------------------
> bash/2576 is trying to acquire lock:
> (s_active){++++.+}, at: [<000000000016c506>] sysfs_addrm_finish+0x5a/0xa8
>
> but task is already holding lock:
> (s_active){++++.+}, at: [<000000000016c90c>] sysfs_get_active_two+0x64/0x88
>
> other info that might help us debug this:
> 5 locks held by bash/2576:
> #0: (&buffer->mutex){+.+.+.}, at: [<000000000016a912>] sysfs_write_file+0x4a/0x1a8
> #1: (s_active){++++.+}, at: [<000000000016c90c>] sysfs_get_active_two+0x64/0x88
> #2: (s_active){++++.+}, at: [<000000000016c8e6>] sysfs_get_active_two+0x3e/0x88
> #3: (cpu_add_remove_lock){+.+.+.}, at: [<00000000003dd318>] cpu_down+0x44/0x88
> #4: (cpu_hotplug.lock){+.+.+.}, at: [<0000000000056eee>] cpu_hotplug_begin+0x3e/0x74
>
I think this is the same with the problem we met some days ago,
http://lkml.org/lkml/2010/1/10/282
does this patch from Erich help?
Thanks.
On Fri, Jan 22, 2010 at 12:42:22PM +0800, Am?rico Wang wrote:
> On Thu, Jan 21, 2010 at 11:53 PM, Heiko Carstens
> <[email protected]> wrote:
> > Hi Eric,
> >
> > rumours are that you could be responsible for the following lockdep message:
> >
> > Happens on cpu hotplug on s390:
> >
> > =============================================
> > [ INFO: possible recursive locking detected ]
> > 2.6.33-rc4-00399-g24bc734 #1
> > ---------------------------------------------
> > bash/2576 is trying to acquire lock:
> > ?(s_active){++++.+}, at: [<000000000016c506>] sysfs_addrm_finish+0x5a/0xa8
> >
> > but task is already holding lock:
> > ?(s_active){++++.+}, at: [<000000000016c90c>] sysfs_get_active_two+0x64/0x88
> >
> > other info that might help us debug this:
> > 5 locks held by bash/2576:
> > ?#0: ?(&buffer->mutex){+.+.+.}, at: [<000000000016a912>] sysfs_write_file+0x4a/0x1a8
> > ?#1: ?(s_active){++++.+}, at: [<000000000016c90c>] sysfs_get_active_two+0x64/0x88
> > ?#2: ?(s_active){++++.+}, at: [<000000000016c8e6>] sysfs_get_active_two+0x3e/0x88
> > ?#3: ?(cpu_add_remove_lock){+.+.+.}, at: [<00000000003dd318>] cpu_down+0x44/0x88
> > ?#4: ?(cpu_hotplug.lock){+.+.+.}, at: [<0000000000056eee>] cpu_hotplug_begin+0x3e/0x74
> >
>
> I think this is the same with the problem we met some days ago,
>
> http://lkml.org/lkml/2010/1/10/282
>
> does this patch from Erich help?
No, it doesn't help.