Date: Fri, 28 Jul 2006 00:06:04 +0200 From: Max Laier <max@love2party.net> To: freebsd-net@freebsd.org Cc: andre@freebsd.org, Julian Elischer <julian@elischer.org>, "Christian S.J. Peron" <csjp@freebsd.org> Subject: Re: [patch] RFC: allow divert from layer 2 ipfw (e.g. bridge) Message-ID: <200607280006.11151.max@love2party.net> In-Reply-To: <44C93425.60001@elischer.org> References: <44C7B5E2.5080001@elischer.org> <44C92278.5000901@FreeBSD.org> <44C93425.60001@elischer.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--nextPart3585932.ycM77z0gHq Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline On Thursday 27 July 2006 23:46, Julian Elischer wrote: > Christian S.J. Peron wrote: > > Max Laier wrote: > >> On Wednesday 26 July 2006 20:35, Julian Elischer wrote: > >>> This code is running on quite a few systems but in a very limited > >>> environment that may not test all possibilities.. > >>> > >>> Does anyone have comments or suggestions as to changes that I might > >>> make > >>> for checkin into generic FreeBSD? It was originally written for 4.x b= ut > >>> with 6.x in mind. > >>> It is now running on 6.1 and seems to be ok so far. > >>> > >>> Certainly I am interested in hearing from Robert and Luigi and I am > >>> particularly interested in > >>> what people think on how this will handle locking/SMP difficulies. > >> > >> Instead of putting more special processing to every L2-entry point in > >> the system, I'd prefer if we could finally get round to L2 pfil > >> hooks. That would make it much easier to add such functionality in a > >> common hook function and use it everywhere. > > > > I agree with Max here, I think it's time we look at getting together > > pfil hooks for layer 2. I would be interested in doing the leg work > > here if you guys are willing to review it. > > The divert code changes are sort-of irrelevant to this discussion. > it adds an ISR to handle divert input from L2. I was referring to the changes in bdg_forward() and ether_output_frame() wh= ich=20 could as well be placed inside a L2-hook inside ip_fw_pfil.c (where they=20 belong IMHO). We went through great length to clean the ip_input/output pa= th=20 off the IPFW/DIVERT entanglement. I don't think we should create the=20 same "mess" in L2 now, just to clean it up later. That said, I'm not opposed against L2 divert or anything. I just wanted to= =20 remind everybody about the plan of L2-pfil hooks that we talked about in th= e=20 past. I would love to see this happen prior the divert changes. =2D-=20 /"\ Best regards, | mlaier@freebsd.org \ / Max Laier | ICQ #67774661 X http://pf4freebsd.love2party.net/ | mlaier@EFnet / \ ASCII Ribbon Campaign | Against HTML Mail and News --nextPart3585932.ycM77z0gHq Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.4 (FreeBSD) iD8DBQBEyTjTXyyEoT62BG0RAmqoAJ9I5WX1bTRawKTCik5lcMyMElt48gCfYePX s/uHfbvSN6YXnQpWzoqRfPI= =YEk2 -----END PGP SIGNATURE----- --nextPart3585932.ycM77z0gHq--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200607280006.11151.max>