From owner-freebsd-stable@FreeBSD.ORG Sat Sep 27 11:32:57 2008 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 65F15106568B; Sat, 27 Sep 2008 11:32:57 +0000 (UTC) (envelope-from danny@cs.huji.ac.il) Received: from cs1.cs.huji.ac.il (cs1.cs.huji.ac.il [132.65.16.10]) by mx1.freebsd.org (Postfix) with ESMTP id 18DEE8FC1B; Sat, 27 Sep 2008 11:32:56 +0000 (UTC) (envelope-from danny@cs.huji.ac.il) Received: from pampa.cs.huji.ac.il ([132.65.80.32]) by cs1.cs.huji.ac.il with esmtp id 1KjY2h-0008GC-PP; Sat, 27 Sep 2008 14:32:55 +0300 X-Mailer: exmh version 2.7.2 01/07/2005 with nmh-1.2 To: Robert Watson In-reply-to: References: <20080926081806.GA19055@icarus.home.lan> <20080926095230.GA20789@icarus.home.lan> Comments: In-reply-to Robert Watson message dated "Sat, 27 Sep 2008 11:15:59 +0100." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sat, 27 Sep 2008 14:32:55 +0300 From: Danny Braniss Message-ID: Cc: freebsd-hackers@freebsd.org, Jeremy Chadwick , freebsd-stable@freebsd.org Subject: Re: bad NFS/UDP performance 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, 27 Sep 2008 11:32:57 -0000 > On Fri, 26 Sep 2008, Danny Braniss wrote: > > > after more testing, it seems it's related to changes made between Aug 4 and > > Aug 29 ie, a kernel built on Aug 4 works fine, Aug 29 is slow. I'l now try > > and close the gap. > > I think this is the best way forward -- skimming August changes, there are a > number of candidate commits, including retuning of UDP hashes by mav, my > rwlock changes, changes to mbuf chain handling, etc. it more difficult than I expected. for one, the kernel date was missleading, the actual source update is the key, so the window of changes is now 28/July to 19/August. I have the diffs, but nothing yet seems relevant. on the other hand, I tried NFS/TCP, and there things seem ok, ie the 'good' and the 'bad' give the same throughput, which seem to point to UDP changes ... danny