Date: Wed, 29 Aug 2001 08:29:09 +0200 From: "Peter Blok" <pblok@inter.NL.net> To: "'Mike Tancsa'" <mike@sentex.net>, "'Garrett Wollman'" <wollman@khavrinen.lcs.mit.edu> Cc: <freebsd-net@FreeBSD.ORG> Subject: RE: Runt frames = broken VLAN ? Message-ID: <001001c13053$e9ea74c0$8a02a8c0@ntpc> In-Reply-To: <5.1.0.14.0.20010828160054.01accec0@marble.sentex.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
Mike, et al. I had the panic too. The problem is the order of initialization. It only happened when you compile NETGRAPH support in the kernel instead of using it as a module. When initialize netgraph a little bit later it works fine. In netgraph.h there is a macro called NETGRAPH_INIT. I have changed the SI_ORDER_ANY to SI_ORDER_LAST or SI_SUB_PSEUDO to something else (I can't remember). I am using the module. I'll check on this today and get a proper answer. Peter -----Original Message----- From: owner-freebsd-net@FreeBSD.ORG [mailto:owner-freebsd-net@FreeBSD.ORG]On Behalf Of Mike Tancsa Sent: Tuesday, August 28, 2001 22:11 To: Garrett Wollman Cc: freebsd-net@FreeBSD.ORG Subject: Re: Runt frames = broken VLAN ? At 12:54 PM 8/28/01 -0400, Garrett Wollman wrote: ><<On Tue, 28 Aug 2001 01:05:32 -0400, Mike Tancsa <mike@sentex.net> said: > > > Can anyone tell me why the VLAN code might be causing my switches (ciscos) > > to see a lot of runt frames when the interface is in 802.1q trunking > mode ? > >It's possible that the Cisco is (bogusly, IMHO) trying to enforce the >Ethernet minimum frame length on the *de*capsulated frames. If you >send a frame that's less than 60 octets long, it gets encapsulated >(adding another four octets) and then padded by the interface up to 64 >octets. After the encapsulation is removed by the receiver, the frame >appears to only be 60 octets long. > >I'd call it a Cisco bug. The minimum frame length in Ethernet arises >from the electrical parameters of the original CSMA/CD Ethernet >design; what matters is the number of clocks the transmitter is >active, not the length of the payload. If its a Cisco bug, would it not manifest it self consistently ? On other ports, I have a 3620 and another catalyst both in 802.1q trunking mode, but I dont see any runt frames there. Also, who is the VLAN maintainer these days ? I ran into a panic that I thought was netgraph related, but Archie Cobbs thinks it might be in the VLAN code. I filed a PR on the issue http://www.freebsd.org/cgi/query-pr.cgi?pr=30149 I saved the debug kernel as well as the core dump in case its needed, but the problem is easy enough to repeat. ---Mike To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?001001c13053$e9ea74c0$8a02a8c0>