Return-Path: Subject: Re: [Fwd: Re: [Bluez-devel] PAN and bridge problem when bnep0 disconnects before bnep1] From: David Woodhouse To: Matthias Thomae Cc: Christoph Scholz , bluez-devel@lists.sourceforge.net, marcel@holtmann.org In-Reply-To: <40A480B8.2050400@thomae-privat.de> References: <1084467658.2322.37.camel@willie.informatik.uni-bonn.de> <40A480B8.2050400@thomae-privat.de> Content-Type: text/plain Message-Id: <1084539891.6181.68.camel@imladris.demon.co.uk> Mime-Version: 1.0 Date: Fri, 14 May 2004 14:04:51 +0100 List-ID: On Fri, 2004-05-14 at 10:18 +0200, Matthias Thomae wrote: > > I have not written to to the bridge people. My personal opinion is that > > BlueZ should do the forwarding (like Affix does). You can argue about > > that of course. > > I don't have a competent opinion about that (except: it's not only the > forwarding problem, but also DHCP, right?), and don't know what Affix does. I don't think it makes sense to do the bridging in our BNEP code. We already _have_ bridging code in the kernel, and there's no sense in duplicating it. Using the existing code allows us to do things like bridging directly onto real Ethernet, too. > Marcel, what do you think? Is there any sense to contact the bridge people? The mail at http://lists.osdl.org/pipermail/bridge/2004-May/000321.html to which Diego referred was from the 'bridge people', and says: "The bridge code was being to picky. The bridge will still work fine when there are two interfaces with the same local address, because it is only used for setting the source address on the outgoing bridge packets." It seems that we can read 'should still work fine' instead of 'will still work fine' in the above, given your report -- so yes, you should report it as a bug in the bridging code if you still can't get it to work on 2.6.6 + Stephen's patch. -- dwmw2