linux-kernel - Re: [PATCH v1 1/2] ASoC: simple-card: Support custom DAI system clock IDs

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: <MXQXY4d7ZMjCu1ChI1EL7daeg1zENP5H@localhost>
Date:   Mon, 24 Oct 2022 10:18:15 +0100
From:   Aidan MacDonald <aidanmacdonald.0x0@...il.com>
To:     Kuninori Morimoto <kuninori.morimoto.gx@...esas.com>
Cc:     broonie@...nel.org, lgirdwood@...il.com, robh+dt@...nel.org,
        krzysztof.kozlowski+dt@...aro.org, perex@...ex.cz, tiwai@...e.com,
        alsa-devel@...a-project.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1 1/2] ASoC: simple-card: Support custom DAI system
 clock IDs


Kuninori Morimoto <kuninori.morimoto.gx@...esas.com> writes:

> Hi Aidan
>
> Thank you for your patch
>
>> Some DAIs have multiple system clock sources, which can be chosen
>> using the "clk_id" argument to snd_soc_dai_set_sysclk(). Currently
>> this is hardcoded to 0 when using simple cards, but that choice is
>> not always suitable.
>>
>> Add the "system-clock-id" property to allow selecting a different
>> clock ID on a per-DAI basis.
>>
>> To simplify the logic on DPCM cards, add a dummy "asoc_simple_dai"
>> instance and use that for the dummy components on DPCM links. This
>> ensures that when we're iterating over DAIs in the PCM runtime there
>> is always a matching "asoc_simple_dai" we can dereference.
>>
>> Signed-off-by: Aidan MacDonald <aidanmacdonald.0x0@...il.com>
>> ---
>
> I think adding "system-clock-id" and adding "dummy asoc_simple_dai" are
> different topics. This patch should be separated into 2 patches.

Sounds good to me.

> And I couldn't understand the reason why we need to add dummy asoc_simple_dai.
> In my understanding, we don't parse DT for dummy connection.
> Which process are you talking about specifically here?
>
> 	This ensures that when we're iterating over DAIs in the PCM runtime there
> 	is always a matching "asoc_simple_dai" we can dereference.
> -
> Thank you for your help !!
>
> Best regards
> ---
> Kuninori Morimoto

DPCM DAI links have some real DAIs and one dummy DAI. Each real DAI has
an asoc_simple_dai associated with it to contain the information parsed
from the DT. The dummy DAI does not have an asoc_simple_dai. I'm adding
a dummy asoc_simple_dai for these dummy DAIs to make the mapping of
snd_soc_dai to asoc_simple_dai 1-to-1.

The non 1-to-1 mapping is problematic, because if I have a snd_soc_dai
and want to look up a simple-card property I would need to check if the
matching asoc_simple_dai exists first, and have a special case for DPCM
dummy DAIs. With a 1-to-1 mapping I can handle all DAIs the same way.

Regards,
Aidan