linux-kernel - Re: [PATCH 1/2] gpio: gpio-mmio: Allow volatile shadow regs

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CACRpkdbzhRMB1gr6CUVdref15v1S5DjHaUxi6=BOYoc3f_ZCrQ@mail.gmail.com>
Date:   Tue, 30 Oct 2018 13:06:27 +0100
From:   Linus Walleij <linus.walleij@...aro.org>
To:     kunyi@...gle.com
Cc:     "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
        Tomer Maimon <tmaimon77@...il.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        avifishman70@...il.com,
        OpenBMC Maillist <openbmc@...ts.ozlabs.org>,
        Mark Brown <broonie@...nel.org>
Subject: Re: [PATCH 1/2] gpio: gpio-mmio: Allow volatile shadow regs

On Wed, Oct 17, 2018 at 11:30 PM Kun Yi <kunyi@...gle.com> wrote:

> Currently the generic GPIO driver stores the direction and data shadow register
> when the driver probes. However, in embedded SOCs the GPIO pins are often
> interleaved with pins muxed to other functions, and pinctrl driver might
> toggle the direction/data register values for these pins. With GPIO
> driver being not the only owner, it should read the shadow registers
> before updating them, otherwise some pin states would be overwritten.
>
> This patch adds a flag BGPIOF_VOLATILE_REG to allow a pinctrl driver that uses
> the generic GPIO interface to indicate the need of read-before-update.
>
> Signed-off-by: Kun Yi <kunyi@...gle.com>

Hi Kun,

as you see the build robot has problems with the patch, please look
into it.

Architecturally I'm not so sure about this, we are introducing more
and more "shadow registers" and I feel what happens at the end
of the day is that we end up reimplementing regmap-mmio.

Please look into switching gpio-mmio.c to use regmap-mmio
and add more registers based on that.

It might be a bit complex but is a way better way forward for
everyone.

Yours,
Linus Walleij