[B.A.T.M.A.N.] echo counting in batmand

[B.A.T.M.A.N.] echo counting in batmand

[Dominic Follett-Smith]

HI there,

I am trying to trace where the protocol updates its count of echos via
a neighbour. So far this is what I am interpreting: If I know that the
OGM came from this station, then we can record that we got the packet
in bit_mark, and update our neighbour's bcast_own_sum in their
originator entry.

if ((has_directlink_flag) && (if_incoming->addr.sin_addr.s_addr ==
bat_packet->orig) && (bat_packet->seqno - if_incoming->out.seqno + 2
== 0)) {

					debug_output(4, "count own bcast (is_my_orig): old = %i, ",
orig_neigh_node->bcast_own_sum[if_incoming->if_num]);

					bit_mark((TYPE_OF_WORD
*)&(orig_neigh_node->bcast_own[if_incoming->if_num * num_words]), 0);
					orig_neigh_node->bcast_own_sum[if_incoming->if_num] =
bit_packet_count((TYPE_OF_WORD
*)&(orig_neigh_node->bcast_own[if_incoming->if_num * num_words]));

					debug_output(4, "new = %i \n",
orig_neigh_node->bcast_own_sum[if_incoming->if_num]);

				}

What I am trying to see is where the sliding window (of size
local_window_size) for echos is defined. This I cannot locate, and in
my report I am saying that bcast_own_sum is the sliding window for
echos, which I am not sure is correct or incorrect.

Thanks in advance for the help.

Regards
Dominic
UCT, SA

Re: [B.A.T.M.A.N.] echo counting in batmand

[Marek Lindner]

Hi,

> if ((has_directlink_flag) && (if_incoming->addr.sin_addr.s_addr ==
> bat_packet->orig) && (bat_packet->seqno - if_incoming->out.seqno + 2
> == 0)) {
> 
> 					debug_output(4, "count own bcast (is_my_orig): old = %i, 
",
> orig_neigh_node->bcast_own_sum[if_incoming->if_num]);
> 
> 					bit_mark((TYPE_OF_WORD
> *)&(orig_neigh_node->bcast_own[if_incoming->if_num * num_words]), 0);
> 					orig_neigh_node->bcast_own_sum[if_incoming->if_num] =
> bit_packet_count((TYPE_OF_WORD
> *)&(orig_neigh_node->bcast_own[if_incoming->if_num * num_words]));
> 
> 					debug_output(4, "new = %i \n",
> orig_neigh_node->bcast_own_sum[if_incoming->if_num]);
> 
> 				}

you may want to look at the corresponding kernel code because it is way 
cleaner and easier to understand.

Cheers,
Marek