Problem in MIRROR target?
Fri, 02 Nov 2001 12:22:59 +0530
No. It happens quite frequently with machines with multiple interfaces
to the internet. In fact the present code of ipt_MIRROR.c takes care of
this by checking if the switching of addresses causes a change in
outgoing device and changes the sk_buff->dst accordingly in route_mirror()
Imran S. Patel wrote:
> On Fri, 2 Nov 2001, Sandip Bhattacharya wrote:
>>I agree that in most of the cases it would be unnecessary. What if the
>>route for coming in is not the same as the route for going out? that is
>>in the input netdevice is different from the one going out? In this case
>>if the pmtu of these devices are different wouldn't that cause a problem?
> I think that it is highly improbable that the host will send the MIRRORed
> packet thru some other network interface, if you go by normal routing
> rules. i.e. the mirrored packet has to be bounced back thru the same