linux-kernel - Re: [BUG] Bw accounting warning on fair-servers' parameters change

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: <20250721093442.244074-1-yurand2000@gmail.com>
Date: Mon, 21 Jul 2025 11:34:42 +0200
From: Yuri Andriaccio <yurand2000@...il.com>
To: juri.lelli@...hat.com
Cc: bsegall@...gle.com,
	dietmar.eggemann@....com,
	linux-kernel@...r.kernel.org,
	luca.abeni@...tannapisa.it,
	mgorman@...e.de,
	mingo@...hat.com,
	peterz@...radead.org,
	rostedt@...dmis.org,
	vincent.guittot@...aro.org,
	vschneid@...hat.com,
	yurand2000@...il.com,
	yuri.andriaccio@...tannapisa.it
Subject: Re: [BUG] Bw accounting warning on fair-servers' parameters change

Hi,

On 18/07/25 16:53, Juri Lelli wrote:
> On 18/07/25 16:22, Juri Lelli wrote:
> > Hi,
> >
> > Thanks for reporting.
> >
> > On 18/07/25 13:38, Yuri Andriaccio wrote:
> > > Hi,
> > >
> > > I've been lately working on fair-servers and dl_servers for some patches and
> > > I've come across a bandwidth accounting warning on the latest tip/master (as of
> > > 2025-07-18, git sha ed0272f0675f). The warning is triggered by simply starting
> > > the machine, mounting debugfs and then just zeroing any fair-server's runtime.
> > >
> > >
> > > The warning:
> > >
> > > WARNING: kernel/sched/deadline.c:266 at dl_rq_change_utilization+0x208/0x230
> > > static inline void __sub_rq_bw(u64 dl_bw, struct dl_rq *dl_rq) {
> > >     ...
> > > 	WARN_ON_ONCE(dl_rq->running_bw > dl_rq->this_bw);
> > > }
> > >
> > > Steps to reproduce:
> > >
> > > mount -t debugfs none /sys/kernel/debug
> > > echo 0 > /sys/kernel/debug/sched/fair_server/cpu0/runtime
> > >
> > >
> > > It does not happen at every machine boot, but happens on most. Could it possibly
> > > be related to some of the deadline timers?
> >
> > I took a quick first look and currently suspect cccb45d7c4295
> > ("sched/deadline: Less agressive dl_server handling") could be playing a
> > role in this as it delays actual server stop.
> >
> > Could you please try to repro after having reverted such commit?
>
> After that (w/o the revert), could you please try to see if the
> following helps?

I've been performing some tests as you asked and indeed the culprit seems to be
cccb45d7c4295 ("sched/deadline: Less agressive dl_server handling"), as
reverting it on the current tip removes the issue.

I've also tested the fix you posted (w/o the reverted commit), and I can confirm
that the warning does not seem to be triggered anymore.

Thanks,
Yuri