linux-kernel - Re: [PATCH 1/2] iio: st_sensors: add always

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:   Fri, 11 Feb 2022 02:10:56 +0100
From:   Linus Walleij <linus.walleij@...aro.org>
To:     Massimo Toscanelli <massimo.toscanelli@...ca-geosystems.com>
Cc:     linux-kernel@...r.kernel.org, jic23@...nel.org, lars@...afoo.de,
        caihuoqing@...du.com, aardelean@...iqon.com,
        andy.shevchenko@...il.com, hdegoede@...hat.com,
        Qing-wu.Li@...ca-geosystems.com.cn, stephan@...hold.net,
        linux-iio@...r.kernel.org, bsp-development.geo@...ca-geosystems.com
Subject: Re: [PATCH 1/2] iio: st_sensors: add always_on flag

On Mon, Feb 7, 2022 at 10:05 AM Massimo Toscanelli
<massimo.toscanelli@...ca-geosystems.com> wrote:

> The st_sensors_read_info_raw() implementation allows to get raw data
> from st_sensors, enabling and disabling the device at every read.
> This leads to delays in data access, caused by the msleep that waits
> the hardware to be ready after every read.
>
> Introduced always_on flag in st_sensor_data, to allow the user to
> keep the device always enabled. In this way, every data access to the
> device can be performed with no delays.
>
> Add always_on sysfs attribute.
>
> Signed-off-by: Massimo Toscanelli <massimo.toscanelli@...ca-geosystems.com>

This creates special dependencies on sysfs poking etc.

Couldn't the runtime PM solve this problem in a better way?

If you look in for example:
drivers/iio/accel/kxsd9.c
how the different pm_runtime* primitives are used, you get an
idea.

Especially note

        /*
         * Set autosuspend to two orders of magnitude larger than the
         * start-up time. 20ms start-up time means 2000ms autosuspend,
         * i.e. 2 seconds.
         */
        pm_runtime_set_autosuspend_delay(dev, 2000);

This creates a "hysteresis window" around when the device is
on, so it is not repeatedly shut off and on, but only after 2 seconds
of inactivity.

This way no special userspace is needed to achieve what you want,
and it benefits everyone.

I wanted to fix this for all the ST sensors but never got around to.

Yours,
Linus Walleij