From owner-freebsd-net Fri Jul 28 14:43:20 2000 Delivered-To: freebsd-net@freebsd.org Received: from verdi.nethelp.no (verdi.nethelp.no [158.36.41.162]) by hub.freebsd.org (Postfix) with SMTP id E9A5037B7C8 for ; Fri, 28 Jul 2000 14:43:15 -0700 (PDT) (envelope-from sthaug@nethelp.no) Received: (qmail 20396 invoked by uid 1001); 28 Jul 2000 21:43:10 +0000 (GMT) To: archie@whistle.com Cc: silby@silby.com, freebsd-net@FreeBSD.ORG, jlemon@FreeBSD.ORG Subject: Re: sub-optimal tcp_ouput() performance in the face of ENOBUFS From: sthaug@nethelp.no In-Reply-To: Your message of "Fri, 28 Jul 2000 14:34:48 -0700 (PDT)" References: <200007282134.OAA27448@bubba.whistle.com> X-Mailer: Mew version 1.05+ on Emacs 19.34.2 Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Date: Fri, 28 Jul 2000 23:43:09 +0200 Message-ID: <20394.964820589@verdi.nethelp.no> Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > I understand the scenario described by the commit message. What > I don't understand about this commit is this: suppose the exact > same scenario happens, except that instead of ip_output() returning > ENOBUFS, it returns zero, BUT the packet is dropped anyway because > of (say) an Ethernet collision. A normal Ethernet collision does *not* result in a dropped packet - simply a packet which is transmitted a few microseconds later by the Ethernet hardware. If the packet has not been successfully transmitted in 16 attempts, it is indeed dropped - but this is (or should be!) an extremely rare event. Steinar Haug, Nethelp consulting, sthaug@nethelp.no To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message