Date: Wed, 1 Mar 2000 13:31:27 -0800 (PST) From: Archie Cobbs <archie@whistle.com> To: wollman@khavrinen.lcs.mit.edu (Garrett Wollman) Cc: julian@elischer.org, ales@megared.net.mx, net@FreeBSD.ORG Subject: Re: GRE Support in 4.X ??? Message-ID: <200003012131.NAA71857@bubba.whistle.com> In-Reply-To: <200003012104.QAA64880@khavrinen.lcs.mit.edu> from Garrett Wollman at "Mar 1, 2000 04:04:38 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Garrett Wollman writes:
> > Sorry, I don't understand your point here. What does TCP
> > have to do with anything?
>
> Look at the code for a moment:
>
> if (ntohl(gre->type) != GRE_PROTOCOL_TYPE) {
> rip_input(m, iphlen);
> return;
> }
>
> So, if it's not the protocol we're expecting, we just punt it to the
> SOCK_RAW interface.
>
> len = m->m_len - len;
> if (ip->ip_len != len) {
> if (len > ip->ip_len || len < sizeof(struct ip)) {
> m_freem(m);
> return;
> }
> m_adj(m, iphlen + sizeof(struct gre));
> }
> ip_input(m);
>
> Otherwise, decapsulate the packet and hand it directly to the IP layer
> to be acted upon as if it had been received directly. The user then
> configures IPFIREWALL_FORWARD to send the packet to lo0 where it gets
> interpreted by the local protocol stack.
What is your opinion about adding GRE support to the kernel?
I.e., should we do it and if so how should it work?
Should a process "bind" to a specific protocol? Should certain
protocols that are already defined (like the above encapsulation)
be handled automatically? Etc..
-Archie
___________________________________________________________________________
Archie Cobbs * Whistle Communications, Inc. * http://www.whistle.com
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?200003012131.NAA71857>