linux-kernel - Re: [PATCH] sched/fair: Add null pointer check to pick_next

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

Message-ID: <3083fb0006f3f872ae8e7fee5b53e1bd6d3b2373.camel@gmx.de>
Date: Sat, 19 Apr 2025 04:53:08 +0200
From: Mike Galbraith <efault@....de>
To: Peter Zijlstra <peterz@...radead.org>, Rik van Riel <riel@...riel.com>
Cc: Chris Mason <clm@...a.com>, Pat Cody <pat@...cody.io>, mingo@...hat.com,
  juri.lelli@...hat.com, vincent.guittot@...aro.org,
 dietmar.eggemann@....com,  rostedt@...dmis.org, bsegall@...gle.com,
 mgorman@...e.de, vschneid@...hat.com,  linux-kernel@...r.kernel.org,
 patcody@...a.com, kernel-team@...a.com, Breno Leitao <leitao@...ian.org>
Subject: Re: [PATCH] sched/fair: Add null pointer check to pick_next_entity()

On Fri, 2025-04-18 at 17:44 +0200, Peter Zijlstra wrote:
> 
> However, due to PREEMPT_NONE, it is possible (Chris mentioned direct
> reclaim and OOM) to have very long (in-kernel) runtimes without
> scheduling.

Dunno what Chris considers 'very long', but FYI you can take some
pretty awful hits even sans PREEMPT_NONE.  In 6.12 time frame I caught
move_expired_inodes() eating 66ms, and isolate_lru_folios() holding
IRQs off for 78ms via dirt simple kbuild+bonnie on spinning rust.

(fugly bandaids for both linger in my 6.12+ trees compost heaps) 

> (I suppose this should be visible by tracing sched_switch)

(I used modified wakeup_rt that only traces max prio)

	-Mike