linux-kernel - Re: [PATCH v2] gpio: gpiolib: clear the array

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]

Date:   Thu, 4 May 2023 22:15:57 +0800
From:   Yan Wang <rk.code@...look.com>
To:     Andy Shevchenko <andy.shevchenko@...il.com>
Cc:     linus.walleij@...aro.org, brgl@...ev.pl,
        "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] gpio: gpiolib: clear the array_info's memory space



> On May 4, 2023, at 19:36, Andy Shevchenko <andy.shevchenko@...il.com> wrote:
> 
> On Thu, May 4, 2023 at 12:38 PM Yan Wang <rk.code@...look.com> wrote:
>> On 4/26/2023 10:42 PM, andy.shevchenko@...il.com wrote:
>>> Sun, Apr 23, 2023 at 09:59:43PM +0800, Yan Wang kirjoitti:
>>>> if hardware number different to array index,it needs to clear to points
>>>> memory space if the array_info have been assigned a value.
>>> Can you explain a bit more what's going on there?
> 
> ...
> 
>> I use gpiod_get_array() to get a gpio array form the node of DTS.
>> 
>> the node is as follows:
>> ...
>> gpios = <&gpio1 0 0>, <&gpio1 10 0>;
>> ...
>> 
>> First scan pin-0 of gpio1,its index and hardware number are 0,
>> 
>> if (descs->ndescs == 0 && gpio_chip_hwgpio(desc) == 0) {
>> ...
>> descs->info = array_info.
>> }
>> 
>> Then scan pin-10 , its index is 1 ,but hardware number is 10 .
>> 
>> if (gpio_chip_hwgpio(desc) != descs->ndescs) {
>> array_info = NULL;
>> }
>> just set array_info = NULL, Should the array_info point to memory be
>> cleared ?
> 
> This is a good question. The entire algorithm is a bit difficult to
> understand from the first glance. I need some time to check it myself.
Looking forward to your test results.
> 
>> if not cleared ,I use the gpiod_set_array_value_cansleep() to pull up or
>> down pin-10 is invalid.
> 
> I'm not sure I follow. The array operations are against the given
> array of the descriptors. If you ask to have that operation done, the
> all descriptors in the array should be considered.
> 
>> I found that the set_mask and get_mask vlaues of descs->info are seted
>> 0x03 in gpiod_get_array(),
> 
> Yes, this mask is for the argument. The 0x03 is the correct one.
> 
>> I think 0x401 is their correct value.
> 
> No. You have an array of two elements, and not 11.

Due to hardware number are 10 and 0 , so this mask is 0x401（bit10 and bit0 are 1）.

> 
>>>> if (bitmap_full(array_info->get_mask, descs->ndescs)) {
>>>> + /*clear descs->info*/
>>>> + memset(array_info, 0, sizeof(struct gpio_array));
>>>> array_info = NULL;
>>>> }
> 
> 
> -- 
> With Best Regards,
> Andy Shevchenko