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 for Android: free password hash cracker in your pocket
[<prev] [next>] [day] [month] [year] [list]
Message-ID: <Y8remhZnLCtr+y5s@linutronix.de>
Date:   Fri, 20 Jan 2023 19:34:02 +0100
From:   Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To:     Hillf Danton <hdanton@...a.com>
Cc:     Mel Gorman <mgorman@...hsingularity.net>,
        Peter Zijlstra <peterz@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...nel.org>,
        Davidlohr Bueso <dave@...olabs.net>, linux-mm@...ck.org,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] locking/rwbase: Prevent indefinite writer starvation

On 2023-01-20 17:37:11 [+0800], Hillf Danton wrote:
> 
> I am fine with either 4ms or 40ms, or a second.
> 
> Given the cure, does it still work when reader bias for RT tasks is allowed?
No.

> If not, why keep starving waiters after they pay the 40ms price?

That kind of starvation will also happen if you have only spinlock_t
locks and you say 3 RT tasks that acquire the lock back to back. And a
few SCHED_OTHER tasks. Those 3 will be always be in front of the queue
(as they skip the line) and the following SCHED_OTHER tasks will starve
and never get the lock.

So it is basically the same scenario.

Sebastian

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ