I've stumbled across something that looks like a bug in the IP
stack in linux relating to source routing.
I have 3 machines here, two linux machines A and B and a BSD
machine C. I have forwarding and source routing options turned
on in all three (in linux, find /proc/sys -name '*source*' |xargs cat
shows all files have a one in them). All three machines are on
the same wire and can freely communicate with each other. There
is no packet filtering enabled.
Using the netcat utility I run tests from machine A:
$ nc -g B C 3333
I see packets forwarded to B then on to C and the connection
completes successfully.
$ nc -g A -g B C 3333
I see packets forwarded over lo to A, and no further packets
are sent over lo or eth0. Machines B and C receive no packets.
$ nc -g B A 3333
I see packets forwarded out over eth0 to B and then back to A
but machine A never seems to accept the packets or reply with
a SYN or RST.
Similar tests from machine C work properly:
$ nc -g B A 3333
$ nc -g C -g B A 3333
$ nc -g B C 3333
and in fact I can bring up an alias interface C2 on C that is nonroutable
and reach the other machines with it:
$ nc -s C2 -g C -g B A 3333
It seems that source routing forwarding over eth0 is working on my
linux box (B), but is not working properly over lo on my other
linux box (A). Further it seems that machine A cannot communicate
with itself using a source route (although C is happy to do so).
What gives? Networking bug? Am I missing a configuration option
in /proc?
Tim Newsham
http://www.lava.net/~newsham/
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
