linux-kernel - Re: [PATCH 5/6] gpiolib: cdev: consolidate edge detector configuration flags

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: <CAHp75VeTyS_77LCp6Uab18M3hJ0dtf5jiSSrPN4tHaBHFEUTkg@mail.gmail.com>
Date:   Wed, 13 Jul 2022 12:07:29 +0200
From:   Andy Shevchenko <andy.shevchenko@...il.com>
To:     Kent Gibson <warthog618@...il.com>
Cc:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
        Bartosz Golaszewski <brgl@...ev.pl>,
        Linus Walleij <linus.walleij@...aro.org>,
        Dipen Patel <dipenp@...dia.com>
Subject: Re: [PATCH 5/6] gpiolib: cdev: consolidate edge detector
 configuration flags

On Wed, Jul 13, 2022 at 3:39 AM Kent Gibson <warthog618@...il.com> wrote:
>
> Combine the polarity_change flag, struct line eflags, and hte enable
> flag into a single flag variable.
>
> The combination of these flags describes the configuration state
> of the edge detector, so formalize and clarify that by combining
> them into a single variable, edflags, in struct line.
>
> The edflags is a subset of the GPIO_V2_LINE_FLAGsb relevant to
> the edge detector, and is also a superset of the eflags it replaces.
> The eflags name is still used to describe the subset of edflags
> corresponding to the rising/falling edge flags where edflags is
> masked down to that subset.
>
> This consolidation reduces the number of variables being passed,
> simplifies state comparisons, and provides a more extensible
> foundation should additional edge sources be integrated in the
> future.

I believe that you have checked this from a locking perspective, so we
won't have worse lock contamination, if any.

...

>         struct linereq *lr;
>         struct gpio_v2_line_event le;
>         int level;
> -       u64 eflags;
> +       u64 edflags;

I would at the same time move it up before `int level;`.

...

> +       int level = -1, diff_seqno;
> +       u64 eflags, edflags = READ_ONCE(line->edflags);

Ditto.

...

>         u32 debounce_period_us;
>         unsigned long irqflags = 0;
>         int irq, ret;
> +       u64 eflags;

Ditto for similarity.

-- 
With Best Regards,
Andy Shevchenko