From owner-freebsd-stable@FreeBSD.ORG Wed Feb 18 02:18:47 2015 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id C5381E2D for ; Wed, 18 Feb 2015 02:18:47 +0000 (UTC) Received: from esa-annu.net.uoguelph.ca (esa-annu.mail.uoguelph.ca [131.104.91.36]) by mx1.freebsd.org (Postfix) with ESMTP id 8B3BEA12 for ; Wed, 18 Feb 2015 02:18:46 +0000 (UTC) X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: A2ALBQDs9eNU/95baINbg1haBIJ/v0YKhSdKAoFhAQEBAQEBfIQMAQEBAwEBAQEgKyALBRYYAgINGQIpAQkmBggHBAEcBIgGCA25AZdwAQEBAQEBBAEBAQEBAQEBARmBIYluhB0BARs0B4JogUIFijqIVYNBgzc4hSCMHSKEDCAxAQaBBDl/AQEB X-IronPort-AV: E=Sophos;i="5.09,595,1418101200"; d="scan'208";a="193204163" Received: from muskoka.cs.uoguelph.ca (HELO zcs3.mail.uoguelph.ca) ([131.104.91.222]) by esa-annu.net.uoguelph.ca with ESMTP; 17 Feb 2015 21:18:45 -0500 Received: from zcs3.mail.uoguelph.ca (localhost.localdomain [127.0.0.1]) by zcs3.mail.uoguelph.ca (Postfix) with ESMTP id A75DCB3F94; Tue, 17 Feb 2015 21:18:45 -0500 (EST) Date: Tue, 17 Feb 2015 21:18:45 -0500 (EST) From: Rick Macklem To: Andris Message-ID: <518452225.5232520.1424225925671.JavaMail.root@uoguelph.ca> In-Reply-To: <005501d04aa7$5d78dc20$186a9460$@grese.net> Subject: Re: 10GbE with jumbo networking problems MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-Originating-IP: [172.17.95.11] X-Mailer: Zimbra 7.2.6_GA_2926 (ZimbraWebClient - FF3.0 (Win)/7.2.6_GA_2926) Cc: freebsd-stable@freebsd.org X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.18-1 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 18 Feb 2015 02:18:47 -0000 Andris wrote: > Hi, > > Is there any progress regarding this bug? > https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=183390 > > hardware and setup is similar to discussed in PR, except instead of > NFS iSCSI (ctl) is used. > OS - FreeBSD 10.1-STABLE r278064. > Well, the workaround fix mentioned by adrian as the last comment on that bug report should be in 10.1. As such, all that should be happening is that the ix driver does a lot of m_defrag() calls. Overhead, but it should work without "file too big" (EFBIG) errors. If you are seeing EFBIG errors, something else is broken. I believe lagg and vlan were both fixed to pass up the reduced value for if_hw_tsomax that avoids the EFBIG errors. Are you using anything else that might be sitting between TCP and the ix driver? (If tcp_output() doesn't see the reduced size of if_hw_tsomax, then the EFBIG errors will occur.) Also, hselasky@ has committed r271946 that added additional fields to "struct ifnet", so that drivers can specify how many transmit segments they can support for a TSO segment and how big each of these transmit segments (read data mbuf in the chain) can be. Unfortunately, the device driver authors haven't yet filled these in. I recently posted a request for this to freebsd-net@, so hopefully this will start happening soon. Finally, if disabling TSO doesn't make the problem go away, then it is some other issue that I don't believe has been isolated. (One responder in the above bug report does mention that disabling TSO didn't fix the problem, but I don't know why.) rick > WBR, > Andris > > > _______________________________________________ > freebsd-stable@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-stable > To unsubscribe, send any mail to > "freebsd-stable-unsubscribe@freebsd.org" >