linux-kernel - Re: [PATCH v3 3/5] ASoC: qcom: add sdm845 sound card support

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 for Android: free password hash cracker in your pocket

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

Date:   Tue, 10 Jul 2018 11:59:26 +0100
From:   Srinivas Kandagatla <srinivas.kandagatla@...aro.org>
To:     Mark Brown <broonie@...nel.org>
Cc:     Rohit kumar <rohitkr@...eaurora.org>, lgirdwood@...il.com,
        robh+dt@...nel.org, mark.rutland@....com, plai@...eaurora.org,
        bgoswami@...eaurora.org, perex@...ex.cz, tiwai@...e.com,
        alsa-devel@...a-project.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 3/5] ASoC: qcom: add sdm845 sound card support



On 09/07/18 17:33, Mark Brown wrote:
> On Mon, Jul 09, 2018 at 03:02:11PM +0100, Srinivas Kandagatla wrote:
>> On 09/07/18 13:41, Mark Brown wrote:
>>>> AFAIU, The issue with that mechanism or EPROBEDEFER is that it works only
> 
>>> This is not the case, the card will be unbound at the ASoC level when
>>> any of the components are removed and then probed again when they
>>> reappear.
> 
>> I did try this and It works only for first time! May be am missing
>> something!
> 
>> snd_soc_component_del_unlocked() unregisters the sound card totally. so for
>> the second time (After DSP stop) there is no registered sound card in
>> place.. Am not sure how this is supposed to work?
> 
>> The reason I think it works for the first time is because of EPROBEDEFER
>> from the machine driver.
> 
> Ugh, right - we ripped out that code because there's no sensible use
> case for it so now we don't keep the cards on a list.  The expectation
> is that if someone is going around removing bits of the card they can
> probably figure out that they should be removing the card first.
> 
> In any case the place to implement this is in the core, there's nothing
> special about your cards here.  Either the core should be using the
> component framework or the card list should be resurrected and we open
> code it.  This isn't something that's unique to your device.
I totally agree with you, this functionality belongs to core!

I will explore both options and see how it goes.

thanks,
srini
>