linux-kernel - RE: [PATCH 1/1] kthread: disable preemption during complete()

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]

Date:	Thu, 26 Jul 2012 14:16:51 -0700
From:	Peter Boonstoppel <pboonstoppel@...dia.com>
To:	Thomas Gleixner <tglx@...utronix.de>,
	Peter Zijlstra <peterz@...radead.org>
CC:	Tejun Heo <tj@...nel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Paul Gortmaker <paul.gortmaker@...driver.com>,
	Henrique de Moraes Holschuh <ibm-acpi@....eng.br>,
	Andy Walls <awalls@...metrocast.net>,
	Diwakar Tundlam <dtundlam@...dia.com>,
	Oleg Nesterov <oleg@...hat.com>, Ingo Molnar <mingo@...e.hu>
Subject: RE: [PATCH 1/1] kthread: disable preemption during complete()

> > tglx has patches that make the kthread create/destroy stuff from hotplug
> > go away.. that seems like the better approach.

> Right. That cpu hotplug setup/teardown stuff is ugly.

If that stuff gets removed complete that's great. The only change I'm aware of right now is the workqueue one: http://thread.gmane.org/gmane.linux.kernel/1329164

> > The main thing is avoiding the wakeup preemption from the complete()
> > because we're going to sleep right after anyway.

You are very likely to be preempted by the complete(), since the newly created thread has a relatively high vruntime.

> > The comment doesn't really make that clear.

> Right, the comment is crap. It has nothing to do with kthread_bind()
> and stuff. The whole purpose is to avoid the pointless preemption
> after wakeup.

The only case I want to solve is the kthread_bind()->wait_task_inactive() scenario. On our platforms this patch reduces average cpu_up() time from about 9ms to 8ms, but max time goes down from 37ms to 8.5ms. cpu_up() latency becomes much more predictable.


PeterB--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/