This is one possible path to fix the specific issue about fragmenting
packets, but it is preferred if we can get it fixed without touching the
IP kernel.

Me and Lennert is independently playing with a couple of different
approaches, trying to find the best approach in how to mix MAC level
routing (i.e. bridging) with IP processing.

Regards
Henrik Nordstr�m
MARA Systems AB, Sweden



lepton wrote:
> 
> Hi!
>         I know the following patch is not a clean method to resolve this 
>problem.Perhaps it
> will cause some problem or just waste some cpu time when you run your box as a 
>gateway
> mode firewall.I am not sure about this.But it seems it can work in bridge mode. so 
>anyone
> just need a bridge to work can try this.
>         I am looking foward to see someone else (Lennert or Brat ? :) )to give a 
>clean way to
> do this.
> 
> diff -ur linux-2.4.9/net/ipv4/ip_output.c uml/net/ipv4/ip_output.c
> --- linux-2.4.9/net/ipv4/ip_output.c    Fri Sep 28 14:01:59 2001
> +++ uml/net/ipv4/ip_output.c    Fri Sep 28 13:43:35 2001
> @@ -886,7 +886,9 @@
>                 iph->tot_len = htons(len + hlen);
> 
>                 ip_send_check(iph);
> -
> +               skb_push(skb2,ETH_HLEN);
> +               memcpy(skb2->data,skb->mac.raw,ETH_HLEN);
> +               skb_pull(skb2,ETH_HLEN);
>                 err = output(skb2);
>                 if (err)
>                         goto fail;
> 
>             lepton
>             [EMAIL PROTECTED]
> 
> _______________________________________________
> Bridge mailing list
> [EMAIL PROTECTED]
> http://www.math.leidenuniv.nl/mailman/listinfo/bridge
_______________________________________________
Bridge mailing list
[EMAIL PROTECTED]
http://www.math.leidenuniv.nl/mailman/listinfo/bridge

Reply via email to