linux-kernel - Re: [PATCH v3 1/6] ASoC: dmaengine: do not use a NULL prepare_slave

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <968f9bfb-bddf-a8f0-6d8f-18b92d865aa2@microchip.com>
Date:   Wed, 20 Apr 2022 09:58:06 +0000
From:   <Codrin.Ciubotariu@...rochip.com>
To:     <sha@...gutronix.de>
CC:     <alsa-devel@...a-project.org>, <linux-kernel@...r.kernel.org>,
        <devicetree@...r.kernel.org>,
        <linux-arm-kernel@...ts.infradead.org>, <lars@...afoo.de>,
        <broonie@...nel.org>, <perex@...ex.cz>, <tiwai@...e.com>,
        <robh+dt@...nel.org>, <Nicolas.Ferre@...rochip.com>
Subject: Re: [PATCH v3 1/6] ASoC: dmaengine: do not use a NULL
 prepare_slave_config() callback

On 20.04.2022 12:15, Sascha Hauer wrote:
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
> 
> Hi,

Hi Sascha,

> 
> On Mon, Mar 07, 2022 at 02:21:57PM +0200, Codrin Ciubotariu wrote:
>> Even if struct snd_dmaengine_pcm_config is used, prepare_slave_config()
>> callback might not be set. Check if this callback is set before using it.
>>
>> Fixes: fa654e085300 ("ASoC: dmaengine-pcm: Provide default config")
>> Signed-off-by: Codrin Ciubotariu <codrin.ciubotariu@...rochip.com>
>> ---
>>
>> Changes in v2,v3:
>>   - none
>>
>>   sound/soc/soc-generic-dmaengine-pcm.c | 6 +++---
>>   1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/sound/soc/soc-generic-dmaengine-pcm.c b/sound/soc/soc-generic-dmaengine-pcm.c
>> index 285441d6aeed..2ab2ddc1294d 100644
>> --- a/sound/soc/soc-generic-dmaengine-pcm.c
>> +++ b/sound/soc/soc-generic-dmaengine-pcm.c
>> @@ -86,10 +86,10 @@ static int dmaengine_pcm_hw_params(struct snd_soc_component *component,
>>
>>        memset(&slave_config, 0, sizeof(slave_config));
>>
>> -     if (!pcm->config)
>> -             prepare_slave_config = snd_dmaengine_pcm_prepare_slave_config;
>> -     else
>> +     if (pcm->config && pcm->config->prepare_slave_config)
>>                prepare_slave_config = pcm->config->prepare_slave_config;
>> +     else
>> +             prepare_slave_config = snd_dmaengine_pcm_prepare_slave_config;
>>
>>        if (prepare_slave_config) {
>>                int ret = prepare_slave_config(substream, params, &slave_config);
> 
> I wonder if this patch is correct. There are drivers like
> sound/soc/mxs/mxs-pcm.c which call snd_dmaengine_pcm_register() with a
> config which has the prepare_slave_config callback unset. For these
> drivers dmaengine_pcm_hw_params() previously was a no-op. Now with this
> patch snd_dmaengine_pcm_prepare_slave_config() and
> dmaengine_slave_config() are called. At least for the mxs-pcm driver
> calling dmaengine_slave_config() will return -ENOSYS.
> 
> At least the "Check if this callback is set before using it" part is
> wrong, the callback is checked before using it with
> 
>          if (prepare_slave_config) {
>                  ...
>          }
> 
> I don't have any mxs hardware at hand to test this. I just stumbled upon
> the change of behaviour when rebasing
> https://patchwork.kernel.org/project/alsa-devel/patch/20220301122111.1073174-1-s.hauer@pengutronix.de/
> on current master.

You are right. I changed the behavior from:
if (pmc->config && !pcm->config->prepare_slave_config)
	<do nothing>
to:
if (pmc->config && !pcm->config->prepare_slave_config)
	snd_dmaengine_pcm_prepare_slave_config()

It was not intended and I agree that the commit message is not accurate. 
I guess some drivers might not need dmaengine_slave_config()...
However, in my case, for the mchp-pdmc driver, I do have pcm->config 
with pcm->config->prepare_slave_config NULL, but I still need 
snd_dmaengine_pcm_prepare_slave_config() to be called. Should we add a 
separate flag to call snd_dmaengine_pcm_prepare_slave_config() if 
pcm->config->prepare_slave_config is NULL?

Nice catch!

Best regards,
Codrin