sockets - Wireshark conversation list sides -
when wireshark conversation list opened (statistics ->conversation list) wireshark showing column of "packets a->b" , column of "packets b->a". when i'm sniffing on 1 side of traffic (physically) can see half mac addresses have traffic on "a->b" , not on "b->a" , other half other way around, makes sense, because sniffed on 1 side of conversations.
the question is: how wireshark decide address call "side a" , address call "side b"? it's easy see described above doesn't depend on side sent more packets/sent first packet , find hard believe decided randomally.
any appreciated.
looking @ the code fills list (ui/gtk/conversations_table.c:1726
):
gtk_list_store_insert_with_values(store, &iter, g_maxint, conv_column_src_addr, src_addr, conv_column_src_port, src_port, conv_column_dst_addr, dst_addr, conv_column_dst_port, dst_port, conv_column_packets, conv_item->tx_frames+conv_item->rx_frames, conv_column_bytes, conv_item->tx_bytes+conv_item->rx_bytes, conv_column_pkt_ab, conv_item->tx_frames, conv_column_bytes_ab, conv_item->tx_bytes, conv_column_pkt_ba, conv_item->rx_frames, conv_column_bytes_ba, conv_item->rx_bytes, conv_column_start, start_time, conv_column_duration, duration, conv_column_bps_ab, tx_ptr, conv_column_bps_ba, rx_ptr, conv_index_column, idx, -1);
…we can see internally, don't use "address a" , "address b"—it's "source" , "destination" addresses of packet. can see "a→b" columns show tx (i.e. transmitted, outgoing) counts while "b→a" columns show rx (i.e. received, incoming) counts.
Comments
Post a Comment