Re: IGMP packets?

Re: IGMP packets?

[FoObArf00]

I have 3 machines.  A (eth0) is connected to B (eth1) with a cross over cable and B (eth0) and C (eth0) are on a switch.  I want to forward the igmp queries and reports that A (eth1) generates when joining/leaving a multicast group on B (eth1) from B(eth0) to C (eth0).  I put the code to do this on different places such as ip_rcv, ip_route_input, etc and no luck.  The weird thing is that it works when I do a tcpdump on B's eth1. 
 
A-(eth0)----(eth1)-B-(eth0)----(eth0)-C

Thanks.

Chris Wright <chrisw@osdl.org> wrote:

>* FoObArf00@netscape.net (FoObArf00@netscape.net) wrote:
>> I have been trying to analyze igmp packets (queries, reports) with
>> ttl of 1 ,of course, in the kernel and ran into a weird situation.
>> Only when an interface is in promiscuous mode (i.e. tcpdump), the igmp
>> packets get to ip_rcv on ip_input.c.  I was wondering if someone can
>> point in the right direction on how/where to get the these packets when
>> not doing a tcpdump.  Thanks
>
>Could you be more specific?  Which machines are involved (e.g. IIRC the
>igmp report is not looped back locally).  What do you mean by analyze in
>the kernel?
>
>thanks,
>-chris
>

__________________________________________________________________
Switch to Netscape Internet Service.
As low as $9.95 a month -- Sign up today at http://isp.netscape.com/register

Netscape. Just the Net You Need.

New! Netscape Toolbar for Internet Explorer
Search from anywhere on the Web and block those annoying pop-ups.
Download now at http://channels.netscape.com/ns/search/install.jsp

Re: IGMP packets?

[Chris Wright]

* FoObArf00@netscape.net (FoObArf00@netscape.net) wrote:
> I have 3 machines.  A (eth0) is connected to B (eth1) with a cross over
> cable and B (eth0) and C (eth0) are on a switch.  I want to forward the
> igmp queries and reports that A (eth1) generates when joining/leaving a
> multicast group on B (eth1) from B(eth0) to C (eth0).  I put the code to
> do this on different places such as ip_rcv, ip_route_input, etc and no
> luck.  The weird thing is that it works when I do a tcpdump on B's eth1.

Sounds like B(eth1) is filtering the packets.  Does it also work if you
enable IFF_ALLMULTI on eth1 (instead of IFF_PROMISC)?  Also, what kernel
are you using?  Does a group join on A generate a v1 or v3 report?

thanks,
-chris

IGMP packets?

[FoObArf00]

I have been trying to analyze igmp packets (queries, reports) with ttl of 1 ,of course, in the kernel and ran into a weird situation.  Only when an interface is in promiscuous mode (i.e. tcpdump), the igmp packets get to ip_rcv on ip_input.c.  I was wondering if someone can point in the right direction on how/where to get the these packets when not doing a tcpdump.  Thanks 

__________________________________________________________________
Switch to Netscape Internet Service.
As low as $9.95 a month -- Sign up today at http://isp.netscape.com/register

Netscape. Just the Net You Need.

New! Netscape Toolbar for Internet Explorer
Search from anywhere on the Web and block those annoying pop-ups.
Download now at http://channels.netscape.com/ns/search/install.jsp

Re: IGMP packets?

[Chris Wright]

* FoObArf00@netscape.net (FoObArf00@netscape.net) wrote:
> I have been trying to analyze igmp packets (queries, reports) with
> ttl of 1 ,of course, in the kernel and ran into a weird situation.
> Only when an interface is in promiscuous mode (i.e. tcpdump), the igmp
> packets get to ip_rcv on ip_input.c.  I was wondering if someone can
> point in the right direction on how/where to get the these packets when
> not doing a tcpdump.  Thanks

Could you be more specific?  Which machines are involved (e.g. IIRC the
igmp report is not looped back locally).  What do you mean by analyze in
the kernel?

thanks,
-chris