I don't see a bug here. I get correct TCP and other conversation lists with
your attached packet capture.
Regards,
Olivier
| -----Original Message-----
| From: Ulf Lamping
|
| Hi List!
|
| I was reporting an access violation in the
| Analyze->Statistics->Conversation List functions some while ago, but
| noone could reproduce this bug.
|
|
| The access violation is always at the same point while
| running it, but
| is dependant on the capture file I have opened,
| so it has to do something with the packets captured.
|
| A capture file which is ok, is always ok, a capture file which fails
| will always fail, so it's no transient bug :-)
|
|
| The access violation is always in endpoint_talkers_tcpip.c line 51
| (or the similar line in the other endpoint_talkers_xy sourcefiles,
| depending on the function called):
|
| add_ett_table_data(talkers, &tcphdr->ip_src, &tcphdr->ip_dst,
| tcphdr->th_sport, tcphdr->th_dport, 1, pinfo->fd->pkt_len, SAT_NONE,
| PT_TCP);
|
| As the pinfo->fd is 0x15, the pinfo->fd->pkt_len is doing the access
| violation.
|
| The callstack is looking like this:
| tcpip_talkers_packet(void * 0x021e1cb0, _packet_info * 0x021a3520,
| _epan_dissect_t * 0x021a3518, void * 0x00c21880) line 51 + 10 bytes
| tap_push_tapped_queue(_epan_dissect_t * 0x021a3518) line 263
| + 31 bytes
| retap_packet(_capture_file * 0x00c36b40, _frame_data * 0x02050f58,
| wtap_pseudo_header * 0x0011ed44, const unsigned char *
| 0x0011edd4, void
| * 0x00000000) line 1413 + 9 bytes
| process_specified_packets(_capture_file * 0x00c36b40,
| packet_range_tag *
| 0x0012ee24, const char * 0x00bc5310, const char * 0x00bc5304, int
| (_capture_file *, _frame_data *, wtap_pseudo_header *, const unsigned
| char *, void *)* 0x00727ae7 retap_packet(_capture_file *,
| _frame_data *,
| wtap_pseudo_header *, const unsigned char *, void *), void *
| 0x00000000)
| line 1385 + 29 bytes
| retap_packets(_capture_file * 0x00c36b40) line 1433 + 33 bytes
| init_ett_table(int 0, char * 0x00c01ea0, char * 0x00c01e9c, char *
| 0x00000000, void * 0x007cc365 tcpip_talkers_packet(void *,
| _packet_info
| *, _epan_dissect_t *, void *)) line 1071 + 10 bytes
| gtk_tcpip_talkers_init(char * 0x00c01e84) line 69 + 26 bytes
| gtk_tcpip_endpoints_cb(_GtkWidget * 0x01ae6eb0, void *
| 0x00000000) line
| 77 + 10 bytes
| LIBGTK-WIN32-2.0-0! 00e006a8()
| LIBGOBJECT-2.0-0! 00fb7ae9()
|
| Attached is a capture file which is causing this bug.
|
| Could someone take a look at this, as I have not enough
| knowledge of the
| conversation stuff to debug this myself.