Re: [Bug 42754] New: BUG: sleeping function called from invalid context at mm/memory.c:3924

Le jeudi 16 février 2012 à 16:10 +0100, Witold Baryluk a écrit :

On 02-15 14:02, Peter Zijlstra wrote:

quoted

On Sun, 2012-02-12 at 09:50 -0800, Paul E. McKenney wrote:

quoted

quoted

I do not know why lockdep is disabled, because I configured kernel to
keep telling my about lock dependency problems even after showing first
one. AFAIK. 

That setting is only relevant to RCU, normal lockdep reports will always
disable lockdep.

One more thing about a bug.

I was able to reproduce bug yestarday , and again today.

First one is again in NetworkManager process (just after loging in to
gnome, exactly like previously), and again NetworkManager process is
killed. And again restarting NM manually make it happy (and works
without further problems). So indeed it looks to be problem with
netlink, because netlink_recvmsg is common call here.

Second one is also in netlink, but process is this time ip, and message
is

[ 9136.204156] BUG: sleeping function called from invalid context at kernel/mutex.c:271

full stacks bellow.

Regards,
Witek

1)

[    0.000000] Linux version 3.3.0-rc3-t43-devel-smp-00188-g3ec1e88-dirty (baryluk@sredniczarny) (gcc version 4.6.2 (Debian 4.6.2-14) ) #22 SMP Sat Feb 11 23:16:44 CET 2012
..
[ 1078.282836] BUG: sleeping function called from invalid context at mm/memory.c:3924
[ 1078.282843] in_atomic(): 1, irqs_disabled(): 0, pid: 3953, name: NetworkManager
[ 1078.282847] INFO: lockdep is turned off.
[ 1078.282852] Pid: 3953, comm: NetworkManager Tainted: G        W  O 3.3.0-rc3-t43-devel-smp-00188-g3ec1e88-dirty #22
[ 1078.282856] Call Trace:
[ 1078.282868]  [<c107d554>] __might_sleep+0x124/0x1d0
[ 1078.282875]  [<c11400ee>] might_fault+0x2e/0xb0
[ 1078.282882]  [<c131d3ab>] copy_to_user+0x3b/0x60
[ 1078.282889]  [<c1568172>] put_cmsg+0x52/0xd0
[ 1078.282903]  [<c159a917>] netlink_recvmsg+0x287/0x340
[ 1078.282910]  [<c1557aa3>] sock_recvmsg+0xd3/0x110
[ 1078.282916]  [<c1140114>] ? might_fault+0x54/0xb0
[ 1078.282921]  [<c15663bc>] ? verify_iovec+0x4c/0xc0
[ 1078.282926]  [<c15579d0>] ? sock_sendmsg_nosec+0x100/0x100
[ 1078.282931]  [<c15588f4>] __sys_recvmsg+0x114/0x1e0
[ 1078.282938]  [<c1082752>] ? check_preempt_curr+0x72/0x90
[ 1078.282943]  [<c1082be2>] ? ttwu_queue+0x42/0xc0
[ 1078.282949]  [<c10b26f1>] ? __lock_acquired+0x2e1/0x330
[ 1078.282956]  [<c1177a65>] ? fget_light+0xc5/0x410
[ 1078.282961]  [<c1177a79>] ? fget_light+0xd9/0x410
[ 1078.282965]  [<c11779e5>] ? fget_light+0x45/0x410
[ 1078.282971]  [<c155a5b9>] sys_recvmsg+0x39/0x60
[ 1078.282976]  [<c155ab13>] sys_socketcall+0x2e3/0x330
[ 1078.282982]  [<c17adacd>] ? sysenter_exit+0xf/0x1a
[ 1078.282988]  [<c131cba8>] ? trace_hardirqs_on_thunk+0xc/0x10
[ 1078.282993]  [<c17ada98>] sysenter_do_call+0x12/0x38
[ 1078.282999]  [<c17a007b>] ? trace_packet+0x104/0x154
[ 1078.294339]  [<c17a92e0>] ? vmalloc_fault+0x190/0x190
[ 1078.294344]  [<c10b1298>] ? trace_hardirqs_on_caller+0x18/0x120
[ 1078.294349]  [<c1016ccd>] do_notify_resume+0x5d/0x80
[ 1078.294354]  [<c17a5d6a>] work_notifysig+0x25/0x2b
[ 1078.294359]  [<c17a92e0>] ? vmalloc_fault+0x190/0x190

Yes, I am aware of this and I'll take a look.

Thanks