From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964895AbVIHQPc (ORCPT ); Thu, 8 Sep 2005 12:15:32 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S964891AbVIHQPc (ORCPT ); Thu, 8 Sep 2005 12:15:32 -0400 Received: from clock-tower.bc.nu ([81.2.110.250]:64426 "EHLO lxorguk.ukuu.org.uk") by vger.kernel.org with ESMTP id S964895AbVIHQPa (ORCPT ); Thu, 8 Sep 2005 12:15:30 -0400 Subject: Re: Ethernet IP multicast maximum packet size From: Alan Cox To: jagdmann@gmail.com Cc: linux-kernel@vger.kernel.org In-Reply-To: <5d0f6099050908083510aa9ab4@mail.gmail.com> References: <5d0f6099050908083510aa9ab4@mail.gmail.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit Date: Thu, 08 Sep 2005 17:40:25 +0100 Message-Id: <1126197625.19834.52.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Evolution 2.2.2 (2.2.2-5) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Iau, 2005-09-08 at 17:35 +0200, Dirk Jagdmann wrote: > Hello developers, > > I googled around all day, but did not find any satisfying answer. Is > there a maximum packet size when I use IP multicasting in a local > ethernet LAN? Or more precisely, does the multicast code in Linux > handle an IP fragmentation/defragmentation of a 63K UDP multicast > datagram, which is transmitted over an 1500bytes MTU ethernet? Or > should I stay on the safe edge and don't construct datagrams > 1500 so > I'll avoid fragmentation? It is considered good practice to keep local multicast within local MTU. For remote multicast 576 byte total frames used to be recommended by some people but MTU discovery and multicast is a real can of worms. (A google for "multicast implosion" should find some useful material on the question of discovery and multicast)