From mboxrd@z Thu Jan  1 00:00:00 1970
From: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Subject: Re: Can I stop sound from driver as system error ?
Date: Tue, 15 Dec 2015 12:55:20 +0900
Message-ID: <566F8F28.2040208@sakamocchi.jp>
References: <87zixcpktb.wl%kuninori.morimoto.gx@renesas.com>
 <566F717D.3050701@sakamocchi.jp>
 <87y4cwph08.wl%kuninori.morimoto.gx@renesas.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; Format="flowed"
Content-Transfer-Encoding: 7bit
Return-path: <alsa-devel-bounces@alsa-project.org>
Received: from smtp301.phy.lolipop.jp (smtp301.phy.lolipop.jp [210.157.22.84])
 by alsa0.perex.cz (Postfix) with ESMTP id 941D9261295
 for <alsa-devel@alsa-project.org>; Tue, 15 Dec 2015 04:55:23 +0100 (CET)
In-Reply-To: <87y4cwph08.wl%kuninori.morimoto.gx@renesas.com>
List-Unsubscribe: <http://mailman.alsa-project.org/mailman/options/alsa-devel>,
 <mailto:alsa-devel-request@alsa-project.org?subject=unsubscribe>
List-Archive: <http://mailman.alsa-project.org/pipermail/alsa-devel/>
List-Post: <mailto:alsa-devel@alsa-project.org>
List-Help: <mailto:alsa-devel-request@alsa-project.org?subject=help>
List-Subscribe: <http://mailman.alsa-project.org/mailman/listinfo/alsa-devel>,
 <mailto:alsa-devel-request@alsa-project.org?subject=subscribe>
Errors-To: alsa-devel-bounces@alsa-project.org
Sender: alsa-devel-bounces@alsa-project.org
To: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Cc: Linux-ALSA <alsa-devel@alsa-project.org>, Mark Brown <broonie@kernel.org>
List-Id: alsa-devel@alsa-project.org

Hi,

On Dec 15 2015 11:47, Kuninori Morimoto wrote:
>>> I wonder can I stop sound from driver side (not from user-land)
>>> as system error during playback or capture ?
>>> How to do it if we can ?
>>
>> Stop PCM substreams with SNDRV_PCM_STATE_XRUN. In detail, see
>> implementation of snd_pcm_stop_xrun() available in 3.19 or
>> later. Then, userspace applications cannot continue to read/write PCM
>> frames via ALSA uapi.
>>
>> When using alsa-lib, then, userspace applications receive -EPIPE as
>> return value of calling PCM APIs. The applications can stop their
>> operation for PCM character devices, or can also recover the XRUN
>> state of PCM substream by calling snd_pcm_prepare(). In this case,
>> struct snd_pcm_ops.prepare() is called again in driver side.
>>
>> I think tinyalsa works almost the same.
>
> Thank you for detail explanation !
> I will try/investigate it

The state of PCM substreams can be changed by process context, software 
IRQ (softIRQ) context and hardware IRQ (hardIRQ) context. Therefore, 
ALSA PCM core uses spin_lock_irqsave()/spin_lock_irqrestore() to manage 
the race condition with softIRQ disabled, hardIRQ disabled and 
kernel-preemption disabled in local processor core.

As a glance, snd-soc-rcar in upstream pushes many operations to PCM 
trigger processing, while this trigger processing is protected with the 
critical section.

When you program the xrun handling in hardware/software IRQ context, 
it's better to pay enough attention to the lock, because any process 
contexts can spin in some processor cores.


Regards

Takashi Sakamoto