linux-kernel - Re: [PATCH] gpio: Document the 'valid

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: <b4319228-44c1-4037-8474-2cbadcb839df@gmail.com>
Date: Fri, 28 Feb 2025 10:31:06 +0200
From: Matti Vaittinen <mazziesaccount@...il.com>
To: Linus Walleij <linus.walleij@...aro.org>
Cc: Matti Vaittinen <matti.vaittinen@...rohmeurope.com>,
 Bartosz Golaszewski <brgl@...ev.pl>, linux-gpio@...r.kernel.org,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH] gpio: Document the 'valid_mask' being internal

On 28/02/2025 10:23, Linus Walleij wrote:
> On Thu, Feb 27, 2025 at 9:24 AM Matti Vaittinen
> <mazziesaccount@...il.com> wrote:
> 
>> I did some quick testing. I used:
> (...)
>> which left GPIO0 ... GPIO6 masked (pins used for ADC) and only GPIO7
>> unmasked.
>>
>> Then I added:
>> gpiotst {
>>          compatible = "rohm,foo-bd72720-gpio";
>>          rohm,dvs-vsel-gpios = <&adc 5 0>, <&adc 6 0>;
>> };
>>
>> and a dummy driver which does:
>> gpio_array = devm_gpiod_get_array(&pdev->dev, "rohm,dvs-vsel",
>>                                    GPIOD_OUT_LOW);
>>
>> ...
>>
>> ret = gpiod_set_array_value_cansleep(gpio_array->ndescs,
>>                  gpio_array->desc, gpio_array->info, values);
>>
>> As a result the bd79124 gpio driver got it's set_multiple called with
>> masked pins. (Oh, and I had accidentally prepared to handle this as I
>> had added a sanity check for pinmux register in the set_multiple()).
> 
> But... how did you mask of the pins 0..5 in valid_mask in this
> example?

I will double-check this soon, but the BD79124 driver should use the 
init_valid_mask() to set all ADC channels 'invalid'. I believe I did 
print the gc->valid_mask in my test-run (0x80) and had the 
set_multiple() called with mask 0x60.

I need to rewind _my_ stack as I already switched to work with BD79104 
instead :) So, please give me couple of hours...

> If this is device tree, I would expect that at least you set up
> gpio-reserved-ranges = <0 5>; which will initialize the valid_mask.
> 
> You still need to tell the gpiolib that they are taken for other
> purposes somehow.
> 
> I think devm_gpiod_get_array() should have failed in that case.
> 
> The call graph should look like this:
> 
> devm_gpiod_get_array()
>      gpiod_get_array()
>          gpiod_get_index(0...n)
>              gpiod_find_and_request()
>                  gpiod_request()
>                      gpiod_request_commit()
>                          gpiochip_line_is_valid()

I remember trying to follow that call stack in the code. The beginning 
of it seems same, but for some reason I didn't end up in the 
gpiochip_line_is_valid(). This, however, requires confirmation :)

> 
> And gpiochip_line_is_valid() looks like this:
> 
> bool gpiochip_line_is_valid(const struct gpio_chip *gc,
>                  unsigned int offset)
> {
>      /* No mask means all valid */
>      if (likely(!gc->valid_mask))
>          return true;
>      return test_bit(offset, gc->valid_mask);
> }
> 
> So why is this not working?

couple of hours please, couple of hours ;)

Yours,
	-- Matti