linux-kernel - Re: [PATCH v3 2/3] ASoC: codecs: wsa883x: Add devm action to safely disable regulator on device removal

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] [day] [month] [year] [list]

Message-ID: <4bc486cb-9d94-4bad-ae07-e9a7aeed481a@kernel.org>
Date: Mon, 28 Jul 2025 15:02:44 +0200
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Mohammad Rafi Shaik <mohammad.rafi.shaik@....qualcomm.com>,
 Srinivas Kandagatla <srini@...nel.org>, Liam Girdwood <lgirdwood@...il.com>,
 Mark Brown <broonie@...nel.org>, Rob Herring <robh@...nel.org>,
 Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley
 <conor+dt@...nel.org>, Jaroslav Kysela <perex@...ex.cz>,
 Takashi Iwai <tiwai@...e.com>, Philipp Zabel <p.zabel@...gutronix.de>,
 Linus Walleij <linus.walleij@...aro.org>, Bartosz Golaszewski <brgl@...ev.pl>
Cc: linux-arm-msm@...r.kernel.org, linux-sound@...r.kernel.org,
 devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
 linux-gpio@...r.kernel.org, quic_pkumpatl@...cinc.com,
 kernel@....qualcomm.com
Subject: Re: [PATCH v3 2/3] ASoC: codecs: wsa883x: Add devm action to safely
 disable regulator on device removal

On 28/07/2025 14:36, Mohammad Rafi Shaik wrote:
> 
> 
> On 7/27/2025 3:00 PM, Krzysztof Kozlowski wrote:
>> On 27/07/2025 10:31, Mohammad Rafi Shaik wrote:
>>> To prevent potential warnings from _regulator_put() during device
>>
>> Warning is either there or not. Either you fix real, specific issue or
>> not. The code looks correct at first glance, so please describe exactly
>> how these warnings happen or how what is the bug being fixed.
>>
> 
> The current wsa883x codec driver manually enables and disables 
> regulators during probe and remove.
> In patch v3-0003, reset functionality was added using 
> devm_reset_control_get_optional_shared_deasserted() for shared gpios.


There is no such code at this point. Each patch is a separate commit and
must stand on its own. With its own explanation. You cannot say that you
add bugs later, so you need to fix something now.

Describe actual problem here. If there is no problem here, describe why
you are doing this.

> 
> However, during cleanup, this led to a warning:
> "WARNING: CPU: 2 PID: 195 at drivers/regulator/core.c:2450 
> _regulator_put+0x50/0x58"
> 
> This occurs because the regulator is still enabled/released when the 
> devm-managed cleanup path attempts to release it.

So that patch was broken? You just did not properly clean up there?

> 
> To resolve this, remove the manual regulator disable logic and instead 
> register a devm-managed cleanup action using devm_add_action_or_reset(). 
> This ensures proper cleanup and avoids regulator misuse warnings.
> 
> For reference, the wsa884x codec driver already follows this approach by 
> using devm actions for regulator management.
> 
>>> removal, register a devm-managed cleanup action using
>>> devm_add_action_or_reset() to safely disable the regulator
>>> associated with the WSA883x codec, ensuring that the regulator
>>> is properly disabled when the device is removed, even if the
>>
>> Device cannot be removed/unloaded, AFAIK, because of suppressed bind.
>> Regulator is already disabled during error paths, so that part of above
>> sentences is just misleading.
>>
>> How can one trigger the warnings?
>>
> 
> The warning in _regulator_put() can be triggered by applying patch 
> v3-0003, which introduces reset functionality using 
> devm_reset_control_get_optional_shared_deasserted().


There is no such code now. You say "potential warnings" are here.

> 
> Since the existing driver handles regulator enable/disable manually, the 
> devm-managed reset cleanup path may attempt to release regulators that 
> are still enabled, leading to the warning.
> 
> This issue highlights the need to replace manual regulator handling with 
> devm_add_action_or_reset() to ensure proper cleanup and avoid such warnings.
> 
>>
>>> probe fails or the driver is unloaded unexpectedly.
>>
>> How driver can be unloaded unexpectedly?
>>
> 
> "Unloaded" might not be the most accurate term here. What I meant is 
> that the driver’s probe can fail due to an error—such as missing 
> resources or improper regulator handling.


Use standard Linux terms, e.g. probe failure, probe deferral etc.

Best regards,
Krzysztof