Return-Path: <cyrus@holtmann.org>
Subject: Re: BUG: spinlock wrong CPU on CPU#3, krfcommd/1500 (resend)
Mime-Version: 1.0 (Mac OS X Mail 7.3 \(1878.2\))
Content-Type: text/plain; charset=us-ascii
From: Marcel Holtmann <marcel@holtmann.org>
In-Reply-To: <135873620.20140517000623@eikelenboom.it>
Date: Fri, 16 May 2014 15:40:11 -0700
Cc: gianluca@sottospazio.it, linux-bluetooth@vger.kernel.org
Message-Id: <20E11CF5-FA5F-42C1-8896-0798E055AE47@holtmann.org>
References: <1433096136.20140516162252@eikelenboom.it> <25887AA2-0BA2-4BD1-ACA6-C43E34F1284B@holtmann.org> <135873620.20140517000623@eikelenboom.it>
To: Sander Eikelenboom <linux@eikelenboom.it>
Sender: linux-bluetooth-owner@vger.kernel.org
List-ID: <linux-bluetooth.vger.kernel.org>

Hi Sander,

>>> I previously reported the splat below i encountered on boot,
>>> but haven't had a response yet. It is still present on 3.15-rc5.
>>> 
>>> 
>>> [   78.248869] BUG: spinlock wrong CPU on CPU#3, krfcommd/1500
>>> [   78.255126]  lock: 0xffff880056d8ded8, .magic: dead4ead, .owner: krfcommd/1500, .owner_cpu: 0
>>> [   78.261375] CPU: 3 PID: 1500 Comm: krfcommd Not tainted 3.15.0-rc3-20140501-nbdebug5+ #1
>>> [   78.267767] Hardware name: MSI MS-7640/890FXA-GD70 (MS-7640)  , BIOS V1.8B1 09/13/2010
>>> [   78.274173]  ffff880056d8ded8 ffff880057217d18 ffffffff81b8d666 ffff88005f6cfd88
>>> [   78.280707]  ffff880056c3a3c0 ffff880057217d38 ffffffff81b8974c ffff880056d8ded8
>>> [   78.287091]  ffffffff81f56fab ffff880057217d58 ffffffff81b89777 ffff880056d8ded8
>>> [   78.293371] Call Trace:
>>> [   78.299496]  [<ffffffff81b8d666>] dump_stack+0x46/0x58
>>> [   78.305601]  [<ffffffff81b8974c>] spin_dump+0x91/0x96
>>> [   78.311662]  [<ffffffff81b89777>] spin_bug+0x26/0x2b
>>> [   78.317599]  [<ffffffff8111a5ce>] do_raw_spin_unlock+0xae/0x100
>>> [   78.323557]  [<ffffffff81b974cb>] _raw_spin_unlock+0x2b/0x40
>>> [   78.329571]  [<ffffffff81b636d3>] rfcomm_run+0x1063/0x1310
>>> [   78.335608]  [<ffffffff81b9754d>] ? _raw_spin_unlock_irqrestore+0x6d/0x90
>>> [   78.341715]  [<ffffffff81b62670>] ? rfcomm_check_accept+0xc0/0xc0
>>> [   78.347830]  [<ffffffff810ee1a4>] kthread+0xe4/0x100
>>> [   78.353916]  [<ffffffff810ee0c0>] ? __init_kthread_worker+0x70/0x70
>>> [   78.360121]  [<ffffffff81b9893c>] ret_from_fork+0x7c/0xb0
>>> [   78.366191]  [<ffffffff810ee0c0>] ? __init_kthread_worker+0x70/0x70
> 
>> can you bisect which patch has introduced this issue. It is the first time I see it.
> 
> Well it's a nope .. since it doesn't occur on every boot and i have no way of 
> triggering it, it's unbisectable.
> Probably only appears when you have the spinlock (sleep) debugging on.

have you tried 86aae6c7b577654b7293374973985a153e0c147e from bluetooth-next since that fixes an issue with CONFIG_DEBUG_ATOMIC_SLEEP enabled.

Regards

Marcel