From owner-freebsd-current@FreeBSD.ORG Mon Mar 10 20:04:55 2008 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id C38D0106566B for ; Mon, 10 Mar 2008 20:04:55 +0000 (UTC) (envelope-from freebsd-current@m.gmane.org) Received: from ciao.gmane.org (main.gmane.org [80.91.229.2]) by mx1.freebsd.org (Postfix) with ESMTP id 450DE8FC1A for ; Mon, 10 Mar 2008 20:04:55 +0000 (UTC) (envelope-from freebsd-current@m.gmane.org) Received: from list by ciao.gmane.org with local (Exim 4.43) id 1JYoEt-0000Wd-VY for freebsd-current@freebsd.org; Mon, 10 Mar 2008 20:04:51 +0000 Received: from mulderlab.f5.com ([205.229.151.151]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Mon, 10 Mar 2008 20:04:51 +0000 Received: from atkin901 by mulderlab.f5.com with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Mon, 10 Mar 2008 20:04:51 +0000 X-Injected-Via-Gmane: http://gmane.org/ To: freebsd-current@freebsd.org From: Mark Atkinson Date: Mon, 10 Mar 2008 13:04:43 -0700 Lines: 48 Message-ID: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7Bit X-Complaints-To: usenet@ger.gmane.org X-Gmane-NNTP-Posting-Host: mulderlab.f5.com User-Agent: KNode/0.10.5 Sender: news Subject: problem with fin+ack with TSO on em 82573 X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Mar 2008 20:04:55 -0000 I have some web servers, each with two of the following on them: em0@pci0:3:0:0: class=0x020000 card=0x61801462 chip=0x108b8086 rev=0x03 hdr=0x00 vendor = 'Intel Corporation' device = 'PC82573V Intel network controller (PCIE Gigabit Ethernet)' class = network subclass = ethernet When TSO is enabled on the interface I have some last ACK problems with some clients. Basically the ack no. doesn't appear to increment with TSO on. 172.16.17.1 is the freebsd-current box (as of 3/5/2008). This is a dump from the 172.16.18.144 box. When I dump from the view of (172.16.17.1) the freebsd box, the resulting appears to lie about the packet ordering, showing it sending the Fin first. Turn TSO off it appears to tell the truth. I don't trust tcpdump with tso/rxcsum/txcsum on anyway. 172.16.18.144.10408 > 172.16.17.1.http: F 312:312(0) ack 39247 win 65535 (DF) 172.16.17.1.http > 172.16.18.144.10408: F 39247:39247(0) ack 312 win 65535 (DF) 172.16.18.144.10408 > 172.16.17.1.http: F 312:312(0) ack 39247 win 65535 (DF) 172.16.17.1.http > 172.16.18.144.10408: F 39247:39247(0) ack 312 win 65535 (DF) 172.16.18.144.10408 > 172.16.17.1.http: F 312:312(0) ack 39247 win 65535 (DF) 172.16.17.1.http > 172.16.18.144.10408: F 39247:39247(0) ack 312 win 65535 (DF) 172.16.18.144.10408 > 172.16.17.1.http: F 312:312(0) ack 39247 win 65535 (DF) 172.16.17.1.http > 172.16.18.144.10408: F 39247:39247(0) ack 312 win 65535 (DF) 172.16.18.144.10408 > 172.16.17.1.http: F 312:312(0) ack 39247 win 65535 (DF) 172.16.17.1.http > 172.16.18.144.10408: F 39247:39247(0) ack 312 win 65535 (DF) 172.16.17.1.http > 172.16.18.144.10408: F 39247:39247(0) ack 312 win 65535 (DF) 172.16.17.1.http > 172.16.18.144.10408: F 39247:39247(0) ack 312 win 65535 172.16.17.1.http > 172.16.18.144.10408: F 39247:39247(0) ack 312 win 65535 -- Mark Atkinson atkin901@yahoo.com (!wired)?(coffee++):(wired);