The attached patch should fix this. It's part of the upcoming 'stable'
release.
diff -u -r1.1 -r1.6
--- br_passthrough.c.20010907-2 2001/10/26 16:49:03 1.1
+++ br_passthrough.c.20010907-2 2001/10/27 18:54:27 1.6
@@ -159,12 +159,12 @@
dprintk(KERN_NOTICE "br_nf_local_in.. ");
if (skb->pkt_type != PACKET_OTHERHOST) {
- dprintk("ip_rcv_finish\n");
- IP_INC_STATS_BH(IpInReceives);
- if (skb->dst == &__fake_dst_entry)
+ dprintk("pass up\n");
+ if (skb->dst == &__fake_dst_entry) {
+ dst_release(&__fake_dst_entry);
skb->dst = NULL;
- ip_rcv_finish(skb);
- return NF_STOLEN;
+ }
+ return NF_ACCEPT;
}
dprintk("drop\n");
On Fri, Oct 26, 2001 at 01:10:43PM -0700, Dan Watson wrote:
> I have a bridge using kernel 2.4.12 with the 20010709-2 bridge patches;
> the bridge code is built as a module. The bridging and netfilter both work
> fine, but packet capture doesn't work right when the br_passthrough module
> is loaded. Running tcpdump on the bridge interface (br0) only reports
> broadcast traffic and traffic from (but not to) the bridge (IP address on
> br0). Without the br_passthrough module loaded, tcpdump shows all of the
> traffic crossing the bridge. The difference can be observed by using
> insmod/rmmod br_passthrough.
>
> I assume that this is a bug rather than a feature because it changes the
> expected behavior bridge interface. This is especially a problem because
> there is no way to capture packets on the individual interfaces in the
> bridge.
>
> _______________________________________________
> Bridge mailing list
> [EMAIL PROTECTED]
> http://www.math.leidenuniv.nl/mailman/listinfo/bridge
--
I are sigfile disease!!
All your quote are belong to us.
Copy us every "sig"!
_______________________________________________
Bridge mailing list
[EMAIL PROTECTED]
http://www.math.leidenuniv.nl/mailman/listinfo/bridge
