linux-kernel - Re: [PATCH] gpio: shared: propagate configuration to pinctrl

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: <CAMRc=McCsEtddu-ANfDX2LDbhLLzztXTuWkZvnweqnobqBp4-Q@mail.gmail.com>
Date: Wed, 21 Jan 2026 14:07:40 +0100
From: Bartosz Golaszewski <brgl@...nel.org>
To: Linus Walleij <linusw@...nel.org>
Cc: Bartosz Golaszewski <bartosz.golaszewski@....qualcomm.com>, 
	Srinivas Kandagatla <srini@...nel.org>, linux-gpio@...r.kernel.org, linux-kernel@...r.kernel.org, 
	Mohammad Rafi Shaik <mohs@....qualcomm.com>, Ravi Hothi <raviravi@....qualcomm.com>
Subject: Re: [PATCH] gpio: shared: propagate configuration to pinctrl

On Wed, Jan 21, 2026 at 1:10 PM Linus Walleij <linusw@...nel.org> wrote:
>
> On Tue, Jan 20, 2026 at 4:49 PM Bartosz Golaszewski
> <bartosz.golaszewski@....qualcomm.com> wrote:
>
> > Just toggling the descriptor's "requested" flag is not enough. We need
> > to properly request it in order to potentially propagate any
> > configuration to pinctrl via the .request() callback.
> >
> > We must not take the reference to the device at this point (the device
> > is not ready but we're also requesting the device's own descriptor) so
> > make the _commit() variants of request and free functions available to
> > GPIO core in order to use them instead of their regular counterparts.
> >
> > This fixes an audio issue reported on one of the Qualcomm platforms.
> >
> > Fixes: a060b8c511ab ("gpiolib: implement low-level, shared GPIO support")
> > Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@....qualcomm.com>
>
> I don't quite understand all semantics but I guess it works
> like this:  gpiod_request_commit() is only called once, during
> proxy device creation.
>
> If I understand it correct:
> Reviewed-by: Linus Walleij <linusw@...nel.org>
>

Yes, we effectively hog the pin and allow access to it only though the proxy.

Bartosz