linux-kernel - Re: [PATCH 1/2] ASoC: codecs: wsa883x: do not set can_multi

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: <bdceb6e4-9de8-07ab-502b-c526e97a592d@linaro.org>
Date:   Wed, 24 May 2023 11:09:32 +0100
From:   Srinivas Kandagatla <srinivas.kandagatla@...aro.org>
To:     Mark Brown <broonie@...nel.org>
Cc:     perex@...ex.cz, tiwai@...e.com, lgirdwood@...il.com,
        ckeepax@...nsource.cirrus.com, kuninori.morimoto.gx@...esas.com,
        linux-kernel@...r.kernel.org, pierre-louis.bossart@...ux.intel.com,
        alsa-devel@...a-project.org
Subject: Re: [PATCH 1/2] ASoC: codecs: wsa883x: do not set can_multi_write
 flag



On 24/05/2023 10:48, Mark Brown wrote:
> On Wed, May 24, 2023 at 10:42:21AM +0100, Srinivas Kandagatla wrote:
>> On 24/05/2023 09:57, Mark Brown wrote:
> 
>>> I'm unclear, is this a limitation of the hardware or of the current
>>> Soundwire code?
> 
>> Its both.
> 
>> Codec itself does not have any private write callback to support this and
>> AFAIU Qualcomm Soundwire controller does not have any such hw facility to
>> program multi-registers for device at one shot.
> 
> How about the *CODEC* hardware?
> 
No, Codec does not have any such interface to write to device registers 
directly without going via Soundwire.

>> Is write callback used for both Bulk writes and multi-writes?
> 
> Only multi-write at this point but we probably should consider redoing
> bulk writes to use it as well.

By the looks of the code, its other way around.

> 
>> Is multi-write feature of regmap bus or device?
> 
> Well, I don't think any buses that understand registers have implemented
> it yet but there's nothing fundamental that means that a bus couldn't
> usefully do something with multi-write.  The current users all use raw
> buses that don't know anything about registers or values.