[ofa-general] Re: having problems with the latest for-2.6.27 tree

[Steve Wise]

Steve Wise wrote:
> Roland Dreier wrote:
>> maybe a bug in "RDMA/cma: Simplify locking needed for serialization of
>> callbacks"?
>>
>> I'll read it over, or you could try running with lockdep enabled...
>>
>>  - R.
>>   
>
> I backed that patch out and things work again...
>
>
> _______________________________________________
> general mailing list
> general at lists.openfabrics.org
> http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
>
> To unsubscribe, please visit 
> http://openib.org/mailman/listinfo/openib-general
With lock dep stuff on:

> =============================================
> [ INFO: possible recursive locking detected ]
> 2.6.26-rc7 #3
> ---------------------------------------------
> iw_cm_wq/3030 is trying to acquire lock:
>  (&id_priv->handler_mutex){--..}, at: [<ffffffffa03a52a6>] 
> iw_conn_req_handler+0x76/0x1f6 [rdma_cm]
>
> but task is already holding lock:
>  (&id_priv->handler_mutex){--..}, at: [<ffffffffa03a31f5>] 
> cma_disable_callback+0x1b/0x38 [rdma_cm]
>
> other info that might help us debug this:
> 3 locks held by iw_cm_wq/3030:
>  #0:  (iw_cm_wq){--..}, at: [<ffffffff81040e78>] run_workqueue+0x95/0x1df
>  #1:  (&work->work){--..}, at: [<ffffffff81040e78>] 
> run_workqueue+0x95/0x1df
>  #2:  (&id_priv->handler_mutex){--..}, at: [<ffffffffa03a31f5>] 
> cma_disable_callback+0x1b/0x38 [rdma_cm]
>
> stack backtrace:
> Pid: 3030, comm: iw_cm_wq Not tainted 2.6.26-rc7 #3
>
> Call Trace:
>  [<ffffffff8104f4e0>] __lock_acquire+0x8fa/0xca0
>  [<ffffffffa03a52a6>] :rdma_cm:iw_conn_req_handler+0x76/0x1f6
>  [<ffffffff8104f913>] lock_acquire+0x8d/0xb1
>  [<ffffffffa03a52a6>] :rdma_cm:iw_conn_req_handler+0x76/0x1f6
>  [<ffffffff8117475e>] account+0xd8/0xe8
>  [<ffffffffa03a52a6>] :rdma_cm:iw_conn_req_handler+0x76/0x1f6
>  [<ffffffff81253352>] mutex_lock_nested+0xe8/0x260
>  [<ffffffffa03a30ee>] :rdma_cm:cma_listen_handler+0x0/0x1b
>  [<ffffffffa03a52a6>] :rdma_cm:iw_conn_req_handler+0x76/0x1f6
>  [<ffffffff81014495>] save_stack_trace+0x22/0x3e
>  [<ffffffff8104c3e0>] save_trace+0x37/0x91
>  [<ffffffff8104e5ae>] mark_held_locks+0x49/0x67
>  [<ffffffff8104e5ae>] mark_held_locks+0x49/0x67
>  [<ffffffff81094196>] kmem_cache_alloc+0x78/0xa2
>  [<ffffffff8104e75e>] trace_hardirqs_on+0xfb/0x126
>  [<ffffffffa039191a>] :iw_cm:cm_work_handler+0x0/0x3ef
>  [<ffffffffa039191a>] :iw_cm:cm_work_handler+0x0/0x3ef
>  [<ffffffffa0391a75>] :iw_cm:cm_work_handler+0x15b/0x3ef
>  [<ffffffffa039191a>] :iw_cm:cm_work_handler+0x0/0x3ef
>  [<ffffffff81040ec2>] run_workqueue+0xdf/0x1df
>  [<ffffffff81041946>] worker_thread+0xd8/0xe3
>  [<ffffffff810442db>] autoremove_wake_function+0x0/0x2e
>  [<ffffffff8104186e>] worker_thread+0x0/0xe3
>  [<ffffffff810441c0>] kthread+0x47/0x73
>  [<ffffffff81254497>] trace_hardirqs_on_thunk+0x35/0x3a
>  [<ffffffff8100ccb8>] child_rip+0xa/0x12
>  [<ffffffff8100c3cf>] restore_args+0x0/0x30
>  [<ffffffff81044179>] kthread+0x0/0x73
>  [<ffffffff8100ccae>] child_rip+0x0/0x12