linux-kernel - Re: [PATCH 1/6] arm64: dts: qcom: sc8280xp-x13s: disable soundcard

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]

Message-ID: <5de95075-ca62-3cae-ce07-d263ea3aa264@linaro.org>
Date:   Mon, 2 Jan 2023 16:46:40 +0100
From:   Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
To:     Johan Hovold <johan@...nel.org>
Cc:     Johan Hovold <johan+linaro@...nel.org>,
        Bjorn Andersson <andersson@...nel.org>,
        Andy Gross <agross@...nel.org>,
        Konrad Dybcio <konrad.dybcio@...aro.org>,
        Rob Herring <robh+dt@...nel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        Srinivas Kandagatla <srinivas.kandagatla@...aro.org>,
        linux-arm-msm@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/6] arm64: dts: qcom: sc8280xp-x13s: disable soundcard

On 02/01/2023 16:39, Johan Hovold wrote:
>>>>>>>  	wcd_tx: wcd9380-tx@0,3 {
>>>>>>>  		compatible = "sdw20217010d00";
>>>>>>> @@ -781,6 +787,8 @@ &vamacro {
>>>>>>>  	pinctrl-names = "default";
>>>>>>>  	vdd-micb-supply = <&vreg_s10b>;
>>>>>>>  	qcom,dmic-sample-rate = <600000>;
>>>>>>> +
>>>>>>> +	status = "disabled";
>>>>>>
>>>>>> That's a double disable.
>>>>>
>>>>> Yes, that's on purpose. We're temporarily disabling these nodes instead
>>>>> of reverting the series which should not have been merged.
>>>>
>>>> I don't get why disabling something twice is anyhow related to
>>>> "temporarily disable". One disable is enough for temporary or permanent
>>>> disables.
>>>
>>> It clearly shows that this was done on purpose and indicates which
>>> properties need to be changed to "okay" once we have actual support.
>>
>> No, it shows nothing clearly as from time to time we got duplicated
>> properties and it's a simply mistake. The double disable without any
>> comment looks like mistake, not intentional code.
> 
> It's not a mistake. It's intentional. And I don't want to spend hours on
> this because of someone else's cock-up.

To you it looks intentional, but for the reader of DTS which has
disabled node in DTSI and in DTS - so in two places - it looks like a
pure bug. Just because you know the reason behind the change does not
make the code readable.

> 
>>>>>
>>>>> Once we have driver support, these properties will be updated again.
>>>>
>>>> Linux kernel is not the only consumer of DTS, thus having or not having
>>>> the support in the kernel is not reason to disable pieces of it.
>>>> Assuming the DTS is correct, of course, because maybe that's the problem?
>>>
>>> Okay, let's revert these sound dts changes then until we have support.
>>> We have no idea if the dts changes are correct as sound still depends
>>> on out-of-tree hacks.
>>>
>>> People are using -next for development and I don't want to see them
>>> toast their speakers because we failed get the dependencies merged
>>> before merging the dts changes which is how we normally do this.
>>
>> If the error is in DTS, yeah, revert or disable is a way. But if the
>> issue is in the incomplete or broken Linux drivers, then these should be
>> changed, e.g. intentionally fail probing, skip new devices, drop new
>> compatible etc.
> 
> And how long does it take for that to propagate and isn't the response
> just going go to be "well then fix the driver".
> 
> I think you're just being unreasonable here.

I did not propose to fix the driver. I proposed to fail the driver's
probe or remove the compatible from it.

Such change propagate the same speed as DTS change.

> If Bjorn could rebase his tree, he could simply drop these for now as
> sound support was clearly not ready. Since that isn't the case we need
> to at least try to be constructive and figure out a reasonable
> alternative. While "Linux isn't the only consumer" is a true statement,
> it really is not relevant just because there are some dts changes in
> Bjorn's tree which should not be there.

The SC8280XP audio DTS looks in general correct, except some style
issues, redundant properties and never tested against DT bindings.
Therefore it looks as accurate and more-or-less correct representation
of the hardware, unless you have some more details on this.

Is the driver is to blame, focus there, not on DTS.

Best regards,
Krzysztof