linux-kernel - sched/fair: documentation on fair server sysfs interface?

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>] [day] [month] [year] [list]

Message-ID: <82fed98e-a4ee-45e1-a4ae-e9fc303a4f67@windriver.com>
Date: Thu, 4 Sep 2025 08:34:23 -0600
From: Chris Friesen <chris.friesen@...driver.com>
To: LKML <linux-kernel@...r.kernel.org>
Subject: sched/fair: documentation on fair server sysfs interface?

Hi,

I'm not subscribed, please CC me on replies.

Is there any documentation available on how 
/sys/kernel/debug/sched/fair_server/cpuX/period and 
/sys/kernel/debug/sched/fair_server/cpuX/runtime are expected to be used 
and what behaviour should result?

I'm especially curious about the interactions with SCHED_FIFO and 
SCHED_DEADLINE tasks.

I had been using RT throttling to protect "regular" tasks from 
SCHED_FIFO tasks trying to hog the entire CPU, and it seems that doesn't 
work the same in 6.12.

What is the recommended way in 6.12 to have a realtime task that is 
effectively a CPU hog, and non-realtime tasks that sometimes need to 
run.   I'd like the realtime task to have as much CPU as possible, but 
not be able to totally starve out the non-realtime task.

Thanks,
Chris