From owner-freebsd-stable@FreeBSD.ORG Sat Dec 17 22:00:21 2011 Return-Path: Delivered-To: freebsd-stable@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 8E5E010656B3; Sat, 17 Dec 2011 22:00:21 +0000 (UTC) (envelope-from avg@FreeBSD.org) Received: from citadel.icyb.net.ua (citadel.icyb.net.ua [212.40.38.140]) by mx1.freebsd.org (Postfix) with ESMTP id C68A68FC21; Sat, 17 Dec 2011 22:00:18 +0000 (UTC) Received: from porto.starpoint.kiev.ua (porto-e.starpoint.kiev.ua [212.40.38.100]) by citadel.icyb.net.ua (8.8.8p3/ICyb-2.3exp) with ESMTP id AAA10029; Sun, 18 Dec 2011 00:00:16 +0200 (EET) (envelope-from avg@FreeBSD.org) Received: from localhost ([127.0.0.1]) by porto.starpoint.kiev.ua with esmtp (Exim 4.34 (FreeBSD)) id 1Rc2Ii-0009wA-G0; Sun, 18 Dec 2011 00:00:16 +0200 Message-ID: <4EED10EF.1030108@FreeBSD.org> Date: Sun, 18 Dec 2011 00:00:15 +0200 From: Andriy Gapon User-Agent: Mozilla/5.0 (X11; FreeBSD amd64; rv:8.0) Gecko/20111206 Thunderbird/8.0 MIME-Version: 1.0 To: Adrian Chadd References: <4EE1EAFE.3070408@m5p.com> <4EE2AE64.9060802@m5p.com> <4EE88343.2050302@m5p.com> <4EE933C6.4020209@zedat.fu-berlin.de> <4EECD261.2080208@m5p.com> <4EED05EC.8050103@FreeBSD.org> In-Reply-To: X-Enigmail-Version: undefined Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: George Mitchell , freebsd-stable@FreeBSD.org, Oliver Pinter Subject: Re: SCHED_ULE should not be the default X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 17 Dec 2011 22:00:21 -0000 on 17/12/2011 23:20 Adrian Chadd said the following: > Erm, just as a random question - since device drivers (and GEOM) run > as separate threads, has anyone looked into what kind of effects the > scheduler has on these? > > I definitely have measurable throughput/responsiveness differences > between ULE and 4BSD (and preempt/non-preempt on 4BSD) on my MIPS > boards when they're bridging traffic. I wonder if there's something > strange going on with the scheduling and preemption of driver netisrs, > taskqueues, the fast interrupt handlers, etc. > > This may -not- be a userland specific problem.. That's an interesting idea. From the recent discussion about USB I can conclude that USB threads run at higher priority than GEOM threads: PI_NET/PI_DISK vs PRIBIO. The former is from the ithread range, the latter is from the regular kernel range. Maybe it would make sense to give the GEOM threads a priority from the ithread range too - given their role and importance. -- Andriy Gapon