linux-kernel - Re: [PATCH] gpiolib: remove redundant callback check

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: <CAMRc=Mf_EXVBJzDY_R5Bge3kDdcANkpEqXz3RwgnfWEm4s5c0g@mail.gmail.com>
Date: Fri, 9 Jan 2026 06:36:21 -0800
From: Bartosz Golaszewski <brgl@...nel.org>
To: Michael Walle <mwalle@...nel.org>
Cc: linux-kernel@...r.kernel.org, linux-gpio@...r.kernel.org, 
	Bartosz Golaszewski <bartosz.golaszewski@....qualcomm.com>, Linus Walleij <linusw@...nel.org>, 
	Bartosz Golaszewski <brgl@...nel.org>
Subject: Re: [PATCH] gpiolib: remove redundant callback check

On Fri, 9 Jan 2026 15:05:36 +0100, Michael Walle <mwalle@...nel.org> said:
> Hi,
>
> On Fri Jan 9, 2026 at 11:55 AM CET, Bartosz Golaszewski wrote:
>> The presence of the .get_direction() callback is already checked in
>> gpiochip_get_direction(). Remove the duplicated check which also returns
>> the wrong error code to user-space.
>>
>> Fixes: e623c4303ed1 ("gpiolib: sanitize the return value of gpio_chip::get_direction()")
>> Reported-by: Michael Walle <mwalle@...nel.org>
>> Closes: https://lore.kernel.org/all/DFJAFK3DTBOZ.3G2P3A5IH34GF@kernel.org/
>> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@....qualcomm.com>
>> ---
>>  drivers/gpio/gpiolib.c | 3 ---
>>  1 file changed, 3 deletions(-)
>>
>> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
>> index 0a14085f3871..5eb918da7ea2 100644
>> --- a/drivers/gpio/gpiolib.c
>> +++ b/drivers/gpio/gpiolib.c
>> @@ -468,9 +468,6 @@ int gpiod_get_direction(struct gpio_desc *desc)
>>  	    test_bit(GPIOD_FLAG_IS_OUT, &flags))
>>  		return 0;
>>
>> -	if (!guard.gc->get_direction)
>> -		return -ENOTSUPP;
>> -
>
> Not sure, if that will make it better or worse though.
>
>>  	ret = gpiochip_get_direction(guard.gc, offset);
>
> Because that will then do a WARN_ON(!.get_direction) and will spam
> the kernel log in case of the gpio-shared-proxy. Also the return
> code will change from ENOTSUPP to EOPNOTSUPP.
>

>From checkpatch.pl:

--
# ENOTSUPP is not a standard error code and should be avoided in new patches.
# Folks usually mean EOPNOTSUPP (also called ENOTSUP), when they type ENOTSUPP.
--

This can end up propagated to user-space which doesn't know what ENOTSUPP is.
This is why we've already changed the return code to -EOPNOTSUPP before, this
is just a leftover.

I prefer to keep the WARN() because it's very unusual for a GPIO driver to not
implement this callback. Can you confirm you're still seeing the issue with
this patch[1]? If so, let's use [2] as the fix?

Bart

[1] https://lore.kernel.org/all/20260108-gpio-shared-false-positive-v1-1-5dbf8d1b2f7d@oss.qualcomm.com/
[2] https://lore.kernel.org/all/20260109130832.27326-1-bartosz.golaszewski@oss.qualcomm.com/