[qpimd-users] Re: [quagga-users 11073] querry related with is there any MRT in pim

[Everton Marques]

Hi Bhavin,

   This time i have done some reading before sending any mail.
I have gone through the code of pim_mroute in pim daemon and it
shows that S,G entry is added to the MFC but what if it exceeds
it limits then is there any MRT or as such provision?

Which limit are you worried about?

One limit I am aware of is the number of interfaces in the
output interface list (OIL). You probably noticed that limit
is MAXVIFS as defined in <linux/mroute.h>. pimd
tracks the number of vif_index in use with the global
counter "qpim_mroute_oif_highest_vif_index".

I don't know what MRT stands for, please clarify...

Cheers,
Everton

[b***@iitb.ac.in]

Hello Everton,
Wha i mean is that pimd directly enteres the S,G entries in to
the MFC. I think there must be a limit on the no. of S,G entries which
can be accomodate in to MFC. MRT tends for multicastr routing table which
is maintained by router to forward multicast traffic. Is it mainly because
pimd deals with SSM so it does not have MRT? (MRT: for e.g. pimd can
have S,G Multicast routing table where it will have table containing
all S,G entries).

Regards,
Bhavin

Post by Everton Marques
Hi Bhavin,

   This time i have done some reading before sending any mail.
I have gone through the code of pim_mroute in pim daemon and it
shows that S,G entry is added to the MFC but what if it exceeds
it limits then is there any MRT or as such provision?

Which limit are you worried about?
One limit I am aware of is the number of interfaces in the
output interface list (OIL). You probably noticed that limit
is MAXVIFS as defined in <linux/mroute.h>. pimd
tracks the number of vif_index in use with the global
counter "qpim_mroute_oif_highest_vif_index".
I don't know what MRT stands for, please clarify...
Cheers,
Everton

[Everton Marques]

Hi Bhavin,

I think the Linux MFC is a data-plane, kernel-side, representation of the MRT.

I don't think there is a size limit for the MFC, but I am only guessing, since
I haven't studied its code. You can check the MFC code here:
http://lxr.linux.no/linux+v2.6.30.6/net/ipv4/ipmr.c

pimd does not have an explicit MRT table, but every global
S,G entry is hold as an "struct pim_upstream", and the list
of all global S,G entries is kept under "qpim_upstream_list".
By "global" (S,G) I mean it is not an interface-specific state.

Everton

Post by b***@iitb.ac.in
Hello Everton,
   Wha i mean is that pimd directly enteres the S,G entries in to
the MFC. I think there must be a limit on the no. of S,G entries which
can be accomodate in to MFC. MRT tends for multicastr routing table which
is maintained by router to forward multicast traffic. Is it mainly because
pimd deals with SSM so it does not have MRT? (MRT: for e.g. pimd can
have S,G Multicast routing table where it will have table containing
all S,G entries).
Regards,
Bhavin

Post by Everton Marques
Hi Bhavin,

   This time i have done some reading before sending any mail.
I have gone through the code of pim_mroute in pim daemon and it
shows that S,G entry is added to the MFC but what if it exceeds
it limits then is there any MRT or as such provision?

Which limit are you worried about?
One limit I am aware of is the number of interfaces in the
output interface list (OIL). You probably noticed that limit
is MAXVIFS as defined in <linux/mroute.h>. pimd
tracks the number of vif_index in use with the global
counter "qpim_mroute_oif_highest_vif_index".
I don't know what MRT stands for, please clarify...
Cheers,
Everton