From mboxrd@z Thu Jan  1 00:00:00 1970
From: Bjorn Andersson <bjorn.andersson@linaro.org>
Subject: Re: [PATCH] qcom: apr: Make apr callbacks in non-atomic context
Date: Wed, 30 Jan 2019 17:16:49 -0800
Message-ID: <20190131011649.GA27190@builder>
References: <20181115184904.27223-1-srinivas.kandagatla@linaro.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Return-path: <linux-kernel-owner@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <20181115184904.27223-1-srinivas.kandagatla@linaro.org>
Sender: linux-kernel-owner@vger.kernel.org
To: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Cc: andy.gross@linaro.org, david.brown@linaro.org, linux-arm-msm@vger.kernel.org, linux-soc@vger.kernel.org, linux-kernel@vger.kernel.org, bgoswami@codeaurora.org, rohitkr@codeaurora.org
List-Id: linux-arm-msm@vger.kernel.org

On Thu 15 Nov 10:49 PST 2018, Srinivas Kandagatla wrote:

> APR communication with DSP is not atomic in nature.
> Its request-response type. Trying to pretend that these are atomic
> and invoking apr client callbacks directly under atomic/irq context has
> endless issues with soundcard. It makes more sense to convert these
> to nonatomic calls. This also coverts all the dais to be nonatomic.
> 

Hi Srinivas,

Sorry for not looking at this before.

Are you sure that you're meeting the latency requirements of low-latency
audio with this change?

[..]
> @@ -303,6 +363,10 @@ static int apr_remove_device(struct device *dev, void *null)
>  
>  static void apr_remove(struct rpmsg_device *rpdev)
>  {
> +	struct apr *apr = dev_get_drvdata(&rpdev->dev);
> +
> +	flush_workqueue(apr->rxwq);
> +	destroy_workqueue(apr->rxwq);

The devices may still be communicating until you remove them on the next
line, wouldn't it make more sense to destroy the work queue after
removing the APR devices?

>  	device_for_each_child(&rpdev->dev, NULL, apr_remove_device);
>  }

Regards,
Bjorn