From owner-freebsd-net@FreeBSD.ORG Tue Apr 22 15:13:11 2003 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E698637B401 for ; Tue, 22 Apr 2003 15:13:11 -0700 (PDT) Received: from gw.catspoiler.org (217-ip-163.nccn.net [209.79.217.163]) by mx1.FreeBSD.org (Postfix) with ESMTP id DD26543F85 for ; Tue, 22 Apr 2003 15:13:10 -0700 (PDT) (envelope-from truckman@FreeBSD.org) Received: from FreeBSD.org (scratch.catspoiler.org [192.168.101.3]) by gw.catspoiler.org (8.12.6/8.12.6) with ESMTP id h3MMD2XB027887; Tue, 22 Apr 2003 15:13:07 -0700 (PDT) (envelope-from truckman@FreeBSD.org) Message-Id: <200304222213.h3MMD2XB027887@gw.catspoiler.org> Date: Tue, 22 Apr 2003 15:13:02 -0700 (PDT) From: Don Lewis To: silby@silby.com In-Reply-To: <20030422004104.Q523@odysseus.silby.com> MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii cc: freebsd-net@FreeBSD.org Subject: Re: IP fragmentation disagreement between current and stable X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 22 Apr 2003 22:13:12 -0000 On 22 Apr, Mike Silbersack wrote: > > On Tue, 22 Apr 2003, Don Lewis wrote: > >> will contain 222 bytes of data. With a 14 byte Ethernet header and a 20 >> byte IP header, that just fits into one 256 byte mbuf. I wonder if >> there is a divide by 4 to calculate the number of words, and the >> >> Could this be a driver problem instead of a problem in the stack itself? >> Both ends are fxp cards. > > A similar problem was found with Via Rhine chips, where having multiple > packets exactly fill up the internal FIFO seemed to cause those packets to > be dropped. Naturally, this only cropped up with fragmented ping packets > of certain lengths. We never bothered patching it because, well, no clear > solution presented itself, and it wouldn't occur in normal usage. > (Actually, changing from store and forward back to smaller DMA sizes might > have fixed it, I'm not sure.) > > So, I would not rule out the possibility of a driver / chipset bug. That looks like the most likely problem. I swapped in an ancient de card on the -current box, and now it works. The fxp card that is is causing problems is a fairly recent Pro/100S desktop card that has an IBM OEM part number. I bought two of these cards from the same source over a period of several months, and both have another oddity. Neither one is probed by either the BIOS or the OS after a power-on cold boot. Not even the Intel diagnostic .exe program sees them. Hitting reset during the boot sequence or doing a reboot from the OS seems to reset something that allows the cards to work. I had to swap out the other card from the system that it was in and replace it with an older fxp that I had on hand because that system needed to be able to boot unattended after a power failure. I try the other card to see if it exhibits the same packet size problem. I'll also try a different flavor fxp, but that'll take some doing since the only remaining spare is on an older Asus motherboard that is on box running -stable.