Date: Thu, 15 Nov 2001 13:22:22 -0800 From: Brooks Davis <brooks@one-eyed-alien.net> To: CJTT <cjtt@employees.org> Cc: freebsd-net@FreeBSD.ORG Subject: Re: ti driver, vlan and tcpdump Message-ID: <20011115132222.B17252@Odin.AC.HMC.Edu> In-Reply-To: <Pine.GSO.3.96.1011115123414.13489A-100000@willers.employees.org>; from cjtt@employees.org on Thu, Nov 15, 2001 at 12:53:35PM -0800 References: <Pine.GSO.3.96.1011115123414.13489A-100000@willers.employees.org>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] On Thu, Nov 15, 2001 at 12:53:35PM -0800, CJTT wrote: > > While sending some test traffic to a > FreeBSD 4.4 (cvsupped yesterday), I noticed > that tcpdump doesn't seem to see VLAN traffic > received on the ti card as it does on the fxp card. > (however, tcpdump on the sending ti card > sees the traffic leave). > [traffic dumps]. > > Why do the two cards behave differently with regard to > tcpdump (libpcap)? I need to be able to see the VLAN traffic > without configuring VLANs (in a test lab). This is caused by the fact that ethernet adaptors provide bpf support via the ether_input() function and on cards like those supported by ti(4) which provide hardware vlan tag extraction, ether_input() is never called. It would be relativly simple to modify vlan_input_tag() to take an ifp argument and have it rebuild a packet to send to bpf. In current you wouldn't even have to change the drivers because the wrapper macro for locking already takes an if pointer. -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4 [-- Attachment #2 --] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE79DIOXY6L6fI4GtQRAoB7AKDlkS+OFYPiW6+2lLdghsQEYRzeNgCfSgbd m860WWqJDVbM/S3sVIwqlpk= =XQQC -----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011115132222.B17252>