linux-kernel - Re: [PATCH 1/1] thermal: sysfs: avoid actual readings from sysfs

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: <f6d023e4-c22b-32fe-f7c6-51a988c43864@gmail.com>
Date:   Sat, 10 Jun 2023 12:24:39 -0500
From:   Russell Haley <yumpusamongus@...il.com>
To:     Eduardo Valentin <evalenti@...nel.org>,
        Daniel Lezcano <daniel.lezcano@...aro.org>
Cc:     eduval@...zon.com, rafael@...nel.org, linux-pm@...r.kernel.org,
        linux-kernel@...r.kernel.org, Amit Kucheria <amitk@...nel.org>,
        Zhang Rui <rui.zhang@...el.com>
Subject: Re: [PATCH 1/1] thermal: sysfs: avoid actual readings from sysfs

On 6/7/23 11:38, Eduardo Valentin wrote:
>> Can you elaborate 'the timing requirement for the governors' ? I'm
>> missing the point
> 
> 
> The point is to avoid contention on the device update path.
> Governor that use differential equations on temperature over time
> will be very time sensitive. Step wise, power allocator, or any
> PID will be very sensitive to time. So, If userspace is hitting
> this API too often we can see cases where the updates needed to
> service userspace may defer/delay the execution of the governor
> logic.
> 
> Despite that, there is really no point to have more updates than
> what was configured for the thermal zone to support. Say that
> we configure a thermal zone to update itself every 500ms, yet
> userspace keeps sending reads every 100ms, we do not need necessarily
> to do a trip to the device every single time to update the temperature,
> as per the design for the thermal zone.

A userspace governor might *also* use PID or filter multiple samples
taken at high rate. I specifically switched my python fan control script
from the Intel coretemp hwmon to the x86_pkg_tmp thermal zone because of
the coretemp driver's annoying 1-second caching behavior.