wireshark/epan/dissectors/packet-ncp.c

1118 lines
37 KiB
C
Raw Normal View History

/* packet-ncp.c
* Routines for NetWare Core Protocol
* Gilbert Ramirez <gram@alumni.rice.edu>
* Modified to allow NCP over TCP/IP decodes by James Coe <jammer@cin.net>
* Modified to decode server op-lock, packet signature,
* & NDS packets by Greg Morris <gmorris@novell.com>
*
* Portions Copyright (c) by Gilbert Ramirez 2000-2002
* Portions Copyright (c) by James Coe 2000-2002
* Portions Copyright (c) Novell, Inc. 2000-2003
*
* $Id$
*
* Wireshark - Network traffic analyzer
* By Gerald Combs <gerald@wireshark.org>
* Copyright 2000 Gerald Combs
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif
#ifdef HAVE_SYS_TYPES_H
# include <sys/types.h>
#endif
#ifdef HAVE_NETINET_IN_H
# include <netinet/in.h>
#endif
#include <string.h>
#include <glib.h>
#include <epan/packet.h>
#include <epan/emem.h>
#include <epan/prefs.h>
#include "packet-ipx.h"
#include "packet-tcp.h"
#include "packet-ncp-int.h"
#include <epan/reassemble.h>
#include <epan/conversation.h>
#include <epan/tap.h>
int proto_ncp = -1;
static int hf_ncp_ip_ver = -1;
static int hf_ncp_ip_length = -1;
static int hf_ncp_ip_rplybufsize = -1;
static int hf_ncp_ip_sig = -1;
static int hf_ncp_ip_packetsig = -1;
static int hf_ncp_type = -1;
static int hf_ncp_seq = -1;
static int hf_ncp_connection = -1;
static int hf_ncp_task = -1;
static int hf_ncp_stream_type = -1;
static int hf_ncp_system_flags = -1;
static int hf_ncp_system_flags_abt = -1;
static int hf_ncp_system_flags_eob = -1;
static int hf_ncp_system_flags_sys = -1;
static int hf_ncp_system_flags_bsy = -1;
static int hf_ncp_system_flags_lst = -1;
static int hf_ncp_src_connection = -1;
static int hf_ncp_dst_connection = -1;
static int hf_ncp_packet_seqno = -1;
static int hf_ncp_delay_time = -1;
static int hf_ncp_burst_seqno = -1;
static int hf_ncp_ack_seqno = -1;
static int hf_ncp_burst_len = -1;
static int hf_ncp_burst_offset = -1;
static int hf_ncp_data_offset = -1;
static int hf_ncp_data_bytes = -1;
static int hf_ncp_missing_fraglist_count = -1;
static int hf_ncp_missing_data_offset = -1;
static int hf_ncp_missing_data_count = -1;
static int hf_ncp_oplock_flag = -1;
static int hf_ncp_oplock_handle = -1;
static int hf_ncp_completion_code = -1;
static int hf_ncp_connection_status = -1;
static int hf_ncp_slot = -1;
static int hf_ncp_control_code = -1;
static int hf_ncp_fragment_handle = -1;
static int hf_lip_echo = -1;
static int hf_ncp_burst_command = -1;
static int hf_ncp_burst_file_handle = -1;
static int hf_ncp_burst_reserved = -1;
gint ett_ncp = -1;
gint ett_nds = -1;
gint ett_nds_segments = -1;
gint ett_nds_segment = -1;
static gint ett_ncp_system_flags = -1;
static struct novell_tap ncp_tap;
struct ncp_common_header header;
struct ncp_common_header *ncp_hdr;
/* Tables for reassembly of fragments. */
GHashTable *nds_fragment_table = NULL;
GHashTable *nds_reassembled_table = NULL;
dissector_handle_t nds_data_handle;
/* desegmentation of NCP over TCP */
static gboolean ncp_desegment = TRUE;
static dissector_handle_t data_handle;
static proto_item *expert_item = NULL;
#define TCP_PORT_NCP 524
#define UDP_PORT_NCP 524
#define NCP_RQST_HDR_LENGTH 7
#define NCP_RPLY_HDR_LENGTH 8
/* These are the header structures to handle NCP over IP */
#define NCPIP_RQST 0x446d6454 /* "DmdT" */
#define NCPIP_RPLY 0x744e6350 /* "tNcP" */
struct ncp_ip_header {
guint32 signature;
guint32 length;
};
/* This header only appears on NCP over IP request packets */
struct ncp_ip_rqhdr {
guint32 version;
guint32 rplybufsize;
};
static const value_string ncp_ip_signature[] = {
{ NCPIP_RQST, "Demand Transport (Request)" },
{ NCPIP_RPLY, "Transport is NCP (Reply)" },
{ 0, NULL },
};
static const value_string burst_command[] = {
{ 0x01000000, "Burst Read" },
{ 0x02000000, "Burst Write" },
{ 0, NULL },
};
/* The information in this module comes from:
NetWare LAN Analysis, Second Edition
Laura A. Chappell and Dan E. Hakes
(c) 1994 Novell, Inc.
Novell Press, San Jose.
ISBN: 0-7821-1362-1
And from the ncpfs source code by Volker Lendecke
And:
Programmer's Guide to the NetWare Core Protocol
Steve Conner & Diane Conner
(c) 1996 by Steve Conner & Diane Conner
Published by Annabooks, San Diego, California
ISBN: 0-929392-31-0
And:
http:developer.novell.com
NCP documentation
*/
static value_string ncp_type_vals[] = {
{ NCP_ALLOCATE_SLOT, "Create a service connection" },
{ NCP_SERVICE_REQUEST, "Service request" },
{ NCP_SERVICE_REPLY, "Service reply" },
{ NCP_WATCHDOG, "Watchdog" },
{ NCP_DEALLOCATE_SLOT, "Destroy service connection" },
{ NCP_BROADCAST_SLOT, "Server Broadcast" },
{ NCP_BURST_MODE_XFER, "Burst mode transfer" },
{ NCP_POSITIVE_ACK, "Request being processed" },
{ NCP_LIP_ECHO, "Large Internet Packet Echo" },
{ 0, NULL }
};
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
static value_string ncp_oplock_vals[] = {
{ 0x21, "Message Waiting" },
{ 0x24, "Clear Op-lock" },
{ 0, NULL }
};
/* Conversation Struct so we can detect NCP server sessions */
typedef struct {
conversation_t *conversation;
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
guint32 nwconnection;
guint8 nwtask;
} mncp_rhash_key;
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
/* Store the packet number for the start of the NCP session.
* Note sessions are defined as
* NCP Connection + NCP Task == Unique NCP server session
* It is normal for multiple sessions per connection to exist
* These are normally different applications running on multi-tasking
* Operating Systems.
*/
typedef struct {
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
guint32 session_start_packet_num;
} mncp_rhash_value;
static GHashTable *mncp_rhash = NULL;
/* Hash Functions */
static gint
mncp_equal(gconstpointer v, gconstpointer v2)
{
const mncp_rhash_key *val1 = (const mncp_rhash_key*)v;
const mncp_rhash_key *val2 = (const mncp_rhash_key*)v2;
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
if (val1->conversation == val2->conversation && val1->nwconnection == val2->nwconnection && val1->nwtask == val2->nwtask) {
return 1;
}
return 0;
}
static guint
mncp_hash(gconstpointer v)
{
const mncp_rhash_key *mncp_key = (const mncp_rhash_key*)v;
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
return GPOINTER_TO_UINT(mncp_key->conversation)+mncp_key->nwconnection+mncp_key->nwtask;
}
/* Initializes the hash table and the mem_chunk area each time a new
* file is loaded or re-loaded in wireshark */
static void
mncp_init_protocol(void)
{
if (mncp_rhash)
g_hash_table_destroy(mncp_rhash);
mncp_rhash = g_hash_table_new(mncp_hash, mncp_equal);
}
/* After the sequential run, we don't need the ncp_request hash and keys
* anymore; the lookups have already been done and the vital info
* saved in the reply-packets' private_data in the frame_data struct. */
static void
mncp_postseq_cleanup(void)
{
}
static mncp_rhash_value*
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
mncp_hash_insert(conversation_t *conversation, guint32 nwconnection, guint8 nwtask, packet_info *pinfo)
{
mncp_rhash_key *key;
mncp_rhash_value *value;
/* Now remember the request, so we can find it if we later
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
a reply to it. Track by conversation, connection, and task number.
in NetWare these values determine each unique session */
key = se_alloc(sizeof(mncp_rhash_key));
key->conversation = conversation;
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
key->nwconnection = nwconnection;
key->nwtask = nwtask;
value = se_alloc(sizeof(mncp_rhash_value));
g_hash_table_insert(mncp_rhash, key, value);
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
if (ncp_echo_conn && nwconnection != 65535) {
expert_add_info_format(pinfo, NULL, PI_RESPONSE_CODE, PI_CHAT, "Detected New Server Session. Connection %d, Task %d", nwconnection, nwtask);
value->session_start_packet_num = pinfo->fd->num;
}
return value;
}
/* Returns the ncp_rec*, or NULL if not found. */
static mncp_rhash_value*
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
mncp_hash_lookup(conversation_t *conversation, guint32 nwconnection, guint8 nwtask)
{
mncp_rhash_key key;
key.conversation = conversation;
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
key.nwconnection = nwconnection;
key.nwtask = nwtask;
return g_hash_table_lookup(mncp_rhash, &key);
}
/*
* Burst packet system flags.
*/
#define ABT 0x04 /* Abort request */
#define BSY 0x08 /* Server Busy */
#define EOB 0x10 /* End of burst */
#define LST 0x40 /* Include Fragment List */
#define SYS 0x80 /* System packet */
Allow either old-style (pre-tvbuff) or new-style (tvbuffified) dissectors to be registered as dissectors for particular ports, registered as heuristic dissectors, and registered as dissectors for conversations, and have routines to be used both by old-style and new-style dissectors to call registered dissectors. Have the code that calls those dissectors translate the arguments as necessary. (For conversation dissectors, replace "find_conversation_dissector()", which just returns a pointer to the dissector, with "old_try_conversation_dissector()" and "try_conversation_dissector()", which actually call the dissector, so that there's a single place at which we can do that translation. Also make "dissector_lookup()" static and, instead of calling it and, if it returns a non-null pointer, calling that dissector, just use "old_dissector_try_port()" or "dissector_try_port()", for the same reason.) This allows some dissectors that took old-style arguments and immediately translated them to new-style arguments to just take new-style arguments; make them do so. It also allows some new-style dissectors not to have to translate arguments before calling routines to look up and call dissectors; make them not do so. Get rid of checks for too-short frames in new-style dissectors - the tvbuff code does those checks for you. Give the routines to register old-style dissectors, and to call dissectors from old-style dissectors, names beginning with "old_", with the routines for new-style dissectors not having the "old_". Update the dissectors that use those routines appropriately. Rename "dissect_data()" to "old_dissect_data()", and "dissect_data_tvb()" to "dissect_data()". svn path=/trunk/; revision=2218
2000-08-07 03:21:25 +00:00
static void
dissect_ncp_common(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree,
gboolean is_tcp)
{
proto_tree *ncp_tree = NULL;
proto_item *ti;
struct ncp_ip_header ncpiph;
struct ncp_ip_rqhdr ncpiphrq;
guint16 ncp_burst_seqno, ncp_ack_seqno;
guint16 flags = 0;
proto_tree *flags_tree = NULL;
int hdr_offset = 0;
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
int commhdr = 0;
int offset = 0;
gint length_remaining;
tvbuff_t *next_tvb;
guint32 testvar = 0, ncp_burst_command, burst_len, burst_off, burst_file;
guint8 subfunction;
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
guint32 nw_connection = 0, data_offset;
guint16 data_len = 0;
guint16 missing_fraglist_count = 0;
mncp_rhash_value *request_value = NULL;
conversation_t *conversation;
if (check_col(pinfo->cinfo, COL_PROTOCOL))
col_set_str(pinfo->cinfo, COL_PROTOCOL, "NCP");
if (check_col(pinfo->cinfo, COL_INFO))
col_clear(pinfo->cinfo, COL_INFO);
hdr_offset = 0;
ncp_hdr = &header;
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
commhdr = hdr_offset;
ti = proto_tree_add_item(tree, proto_ncp, tvb, 0, -1, FALSE);
ncp_tree = proto_item_add_subtree(ti, ett_ncp);
if (is_tcp) {
if (tvb_get_ntohl(tvb, hdr_offset) != NCPIP_RQST && tvb_get_ntohl(tvb, hdr_offset) != NCPIP_RPLY)
commhdr += 1;
/* Get NCPIP Header data */
ncpiph.signature = tvb_get_ntohl(tvb, commhdr);
proto_tree_add_uint(ncp_tree, hf_ncp_ip_sig, tvb, commhdr, 4, ncpiph.signature);
ncpiph.length = (0x7fffffff & tvb_get_ntohl(tvb, commhdr+4));
proto_tree_add_uint(ncp_tree, hf_ncp_ip_length, tvb, commhdr+4, 4, ncpiph.length);
commhdr += 8;
if (ncpiph.signature == NCPIP_RQST) {
ncpiphrq.version = tvb_get_ntohl(tvb, commhdr);
proto_tree_add_uint(ncp_tree, hf_ncp_ip_ver, tvb, commhdr, 4, ncpiphrq.version);
commhdr += 4;
ncpiphrq.rplybufsize = tvb_get_ntohl(tvb, commhdr);
proto_tree_add_uint(ncp_tree, hf_ncp_ip_rplybufsize, tvb, commhdr, 4, ncpiphrq.rplybufsize);
commhdr += 4;
}
/* Check to see if this is a valid offset, otherwise increment for packet signature */
if (match_strval(tvb_get_ntohs(tvb, commhdr), ncp_type_vals)==NULL) {
Fixes and updated files: File NCP2222.py: 1. Added NCP service type 0xffff = All types 2. Added evaluation of task states. (following bits are defined. 0=normal, 1 = TTS explicit transaction in progress, 2= TTS implicit transaction in progress, 4 = Shared file ste lock in progress) 3. Reversed the Volume Request Flags, this was backwards (s/b 0=do not return name with volume number, 1=Return name with volume number) 4. Fixed endianess of Creator ID in NetWare Information Struct. 5. File information structure incorrectly defined Current Block Being Decompressed. 6. Logical Lock Status structure incorrectly identified Task Number as a byte. It should be two bytes "word". 7. Fixed endianess of Modifier ID in NetWare Information Struct. 8. Fixed Name Space Information structure (was defined as Name Space, changed to Creator Name Space Number). 9. Fixed Semaphore Structure (Task Number was incorrectly defined as a byte, should be two bytes "word"). 10. Added Task structure (Task Number, Task State) 11. Fixed Volume structure (incorrectly defined Volume Name) 12. Added VolumeWithName structure (This struct contains both volume number and volume name) 13. Added error 0x8901 - "No purgable files available" 14. Added error 0x8977 - "Buffer too small" 15. Added error 0x899c - "No more trustees found" 16. Added error 0x89d9 - "Queue station is not a server" 17. Added NCP connection status values (0=Ok, 1=Bad service connection, 10=File server is down, 40=Broadcast Message Pending) 18. Fixed error definitions for NCP 22/50. 19. Fixed reply packet for NCP 22/52 to properly display volume information depending on reply struct, Volume or volume with name. 20. Fixed error definitions for NCP 22/52 21. Fixed NCP 23/26 reply packet to properly return internet address and display correctly. 22. Fixed NCP 23/27 Reply packet structure to properly repeat connection numbers array. 23. Fixed error definitions for NCP 23/33 24. Fixed NCP 23/114 Request (improperly defined Charge Information as a long value, s/b word) 25. Fixed NCP 23/120 Reply (Improperly defined Job Number as a long value, s/b word) 26. Fixed error definitions for NCP 23/123 27. Fixed error definitions for NCP 23/124 28. Fixed error definitions for NCP 23/131 29. Fixed error definitions for NCP 23/132 30. Fixed error definitions for NCP 23/135 31. Fixed error definitions for NCP 23/137 32. Fixed error definitions for NCP 23/138 33. Fixed NCP 23/205 Reply (User Login Allowed incorrectly defined as long value, s/b byte value) 34. Fixed NCP 23/234 Reply (Reply structure incorrectly defined based on old NCP documentation. Corrected per new docs) 35. Fixed endianess of NCP 23/237 Reply on value Number of Locks. 36. Fixed endianess of NCP 23/238 Reply on value Number of Locks. 37. Fixed error definitions for NCP 30 38. Fixed error definitions for NCP 36/6 39. Fixed error definitions for NCP 86/2 40. Fixed error definitions for NCP 86/3 41. Fixed error definitions for NCP 86/4 42. Fixed error definitions for NCP 86/5 43. Fixed error definitions for NCP 87/1 44. Fixed error definitions for NCP 87/4 45. Fixed error definitions for NCP 87/5 46. Fixed error definitions for NCP 87/10 47. Fixed error definitions for NCP 87/11 48. Fixed error definitions for NCP 87/12 49. Fixed error definitions for NCP 87/17 50. Fixed error definitions for NCP 87/18 51. Fixed NCP 87/20 Reply to properly decode multiple entries returned. This used to only dissect the first entry. 52. Fixed error definitions for NCP 87/30 53. Fixed NCP 87/33 Reply to properly utilize the request flags to dissect the reply packet NetWare Info Struct 54. Fixed error definitions for NCP 87/33 55. Fixed error definitions for NCP 88/22 56. Fixed error definitions for NCP 89/1 57. Fixed error definitions for NCP 89/10 58. Fixed NCP 89/11 Request packet structure 59. Fixed error definitions for NCP 89/11 60. Fixed NCP 89/20 Reply to properly decode multiple entries returned. This used to only dissect the first entry. 61. Fixed error definitions for NCP 104/5 62. Added undefined NCP 112 63. Fixed NCP 123/11 Reply to properly dissect FileName, Name, and copyright. 64. Fixed NCP 123/17 Reply to properly display NCP Network Address values. 65. Fixed NCP 123/24 Reply to properly display Driver Board Name, Driver Short Name, and Dirver Logical Name. 66. Fixed error definitions for NCP 123/33 67. Fixed NCP 123/60 to properly display Set Command Name and Set Command Value. 68. Fixed error definitions for NCP 123/70 69. Fixed NCP 123/71 Reply to utilize new File Information Struct 70. Fixed error definitions for NCP 123/71 71. Fixed error definitions for NCP 123/72 72. Added NCP 123/249 73. Added NCP 123/251 74. Added NCP 123/252 75. Added NCP 123/253 76. Added NCP 123/254 77. Added NCP 123/255 78. Fixed error definitions for NCP 131/1 79. Fixed error definitions for NCP 131/2 File packet-ncp2222.inc 1. Added new NCP preference setting to tell Wireshark to decode the NetWare information structure as new or old style. 2. Set default NCP preference settings of echo connection and echo file to FALSE. 3. Added NDS verb 2 Request Flags (0=retain old object, 1=delete old object) 4. Fixed problem where NDS fragmentation could not be reassembled on reload. (fragment array needed to be re-initialized on reload) 5. Fixed NCP service types 1111, 5555, bbbb, and 1111/LIP to reflect real type number. (Was defined as 0xf1, 0xf2, 0xf3, etc... Now defined as 0x1, 0x5, 0xb, etc) 6. Fixed function build_expert_data to parse subtree memory structures to acquire NCP request value records 7. Fix NDS attribute type Boolean to byte value and properly aligned. 8. Fixed attribute zendmSearchOrder to display properly regardless of number or order of value entries. 9. Added check of length of packet prior to attempting to defragment 10. Added for Request packets logic to abort if NCP type isn't found. 11. Added logic to store packet length and then manually decode NCP function 123 based on length. 12. Added logic to manually dissect NCP 87/20 and 89/20 reply packets. 13. NDS resolve name replies with remote entry as the specifier should not store the EID returned (0x00000000) 14. Added logic to trap and echo to expert tap when connection status flags indicate an error. 15. Added manual dissect of NCP 23/26 replies 16. Added logic to Capture the EID returned form NCP 22/51 "Get vol info". 17. Fixed NDS verb 0x2a 18. Fixed NDS verb 0x2b File packet-ncp.c 1. Fix offset for packet signature. (This used to automatically set the offset to account for packet signature if it could not determine the correct NCP type. But for some failed fragment packets, retransmissions, etc, this would be wrongly identified. So first we check to see if we can read a valid type at the offset before we just automatically assume that packet signature is being used) 2. Register the new NCP preference for old/new NetWareInfoStruct. File packet-ncp-int.h 1. Add extern declaration for ncp_newstyle setting. 2. Add length, req_mask, and req_mast_ext to ncp_req_hash_value structure File packet-ncp-sss.c 1. Fix SecretStore request verb Write App Secrets, Client Put Data, to evaluate packet length. svn path=/trunk/; revision=19073
2006-08-29 08:45:53 +00:00
/* Check to see if we have a valid type after packet signature length */
if (match_strval(tvb_get_ntohs(tvb, commhdr+8), ncp_type_vals)!=NULL) {
proto_tree_add_item(ncp_tree, hf_ncp_ip_packetsig, tvb, commhdr, 8, FALSE);
commhdr += 8;
}
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
}
}
header.type = tvb_get_ntohs(tvb, commhdr);
header.sequence = tvb_get_guint8(tvb, commhdr+2);
header.conn_low = tvb_get_guint8(tvb, commhdr+3);
header.task = tvb_get_guint8(tvb, commhdr+4);
header.conn_high = tvb_get_guint8(tvb, commhdr+5);
proto_tree_add_uint(ncp_tree, hf_ncp_type, tvb, commhdr, 2, header.type);
nw_connection = (header.conn_high*256)+header.conn_low;
/* Ok, we need to track the conversation so that we can
* determine if a new server session is occuring for this
* connection.
*/
conversation = find_conversation(pinfo->fd->num, &pinfo->src, &pinfo->dst,
PT_NCP, (guint32) pinfo->srcport, (guint32) pinfo->destport,
0);
if ((ncpiph.length & 0x80000000) || ncpiph.signature == NCPIP_RPLY) {
/* First time through we will record the initial connection and task
* values
*/
if (!pinfo->fd->flags.visited) {
if (conversation != NULL) {
/* find the record telling us the
* request made that caused this
* reply
*/
request_value = mncp_hash_lookup(conversation, nw_connection, header.task);
/* if for some reason we have no
* conversation in our hash, create
* one */
if (request_value == NULL) {
request_value = mncp_hash_insert(conversation, nw_connection, header.task, pinfo);
}
} else {
/* It's not part of any conversation
* - create a new one.
*/
conversation = conversation_new(pinfo->fd->num, &pinfo->src,
&pinfo->dst, PT_NCP, (guint32) pinfo->srcport, (guint32) pinfo->destport, 0);
request_value = mncp_hash_insert(conversation, nw_connection, header.task, pinfo);
}
/* If this is a request packet then we
* might have a new task
*/
if (ncpiph.signature == NCPIP_RPLY) {
/* Now on reply packets we have to
* use the state of the original
* request packet, so look up the
* request value and check the task number
*/
request_value = mncp_hash_lookup(conversation, nw_connection, header.task);
}
} else {
/* Get request value data */
request_value = mncp_hash_lookup(conversation, nw_connection, header.task);
if (request_value) {
if ((request_value->session_start_packet_num == pinfo->fd->num) && ncp_echo_conn)
{
expert_add_info_format(pinfo, NULL, PI_RESPONSE_CODE, PI_CHAT, "Detected New Server Session. Connection %d, Task %d", nw_connection, header.task);
}
}
}
} else {
if (!pinfo->fd->flags.visited) {
if (conversation != NULL) {
/* find the record telling us the
* request made that caused this
* reply
*/
request_value = mncp_hash_lookup(conversation, nw_connection, header.task);
/* if for some reason we have no
* conversation in our hash, create
* one */
if (request_value == NULL) {
request_value = mncp_hash_insert(conversation, nw_connection, header.task, pinfo);
}
} else {
/* It's not part of any conversation
* - create a new one.
*/
conversation = conversation_new(pinfo->fd->num, &pinfo->src,
&pinfo->dst, PT_NCP, (guint32) pinfo->srcport, (guint32) pinfo->destport, 0);
request_value = mncp_hash_insert(conversation, nw_connection, header.task, pinfo);
}
/* find the record telling us the request
* made that caused this reply
*/
} else {
request_value = mncp_hash_lookup(conversation, nw_connection, header.task);
if (request_value) {
if ((request_value->session_start_packet_num == pinfo->fd->num) && ncp_echo_conn)
{
expert_add_info_format(pinfo, NULL, PI_RESPONSE_CODE, PI_CHAT, "Detected New Server Session. Connection %d, Task %d", nw_connection, header.task);
}
}
}
}
tap_queue_packet(ncp_tap.hdr, pinfo, ncp_hdr);
if (check_col(pinfo->cinfo, COL_INFO)) {
col_add_fstr(pinfo->cinfo, COL_INFO,
"%s",
val_to_str(header.type, ncp_type_vals, "Unknown type (0x%04x)"));
}
/*
* Process the packet-type-specific header.
*/
switch (header.type) {
case NCP_BROADCAST_SLOT: /* Server Broadcast */
proto_tree_add_uint(ncp_tree, hf_ncp_seq, tvb, commhdr + 2, 1, header.sequence);
proto_tree_add_uint(ncp_tree, hf_ncp_connection,tvb, commhdr + 3, 3, nw_connection);
proto_tree_add_item(ncp_tree, hf_ncp_task, tvb, commhdr + 4, 1, FALSE);
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
proto_tree_add_item(ncp_tree, hf_ncp_oplock_flag, tvb, commhdr + 9, 1, tvb_get_guint8(tvb, commhdr+9));
proto_tree_add_item(ncp_tree, hf_ncp_oplock_handle, tvb, commhdr + 10, 4, FALSE);
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
if ((tvb_get_guint8(tvb, commhdr+9)==0x24) && ncp_echo_file)
{
expert_add_info_format(pinfo, NULL, PI_RESPONSE_CODE, PI_CHAT, "Server requesting station to clear oplock on handle - %08x", tvb_get_ntohl(tvb, commhdr+10));
}
break;
case NCP_LIP_ECHO: /* Lip Echo Packet */
proto_tree_add_item(ncp_tree, hf_lip_echo, tvb, commhdr, 13, FALSE);
break;
case NCP_BURST_MODE_XFER: /* Packet Burst Packet */
/*
* XXX - we should keep track of whether there's a burst
* outstanding on a connection and, if not, treat the
* beginning of the data as a burst header.
*
* The burst header contains:
*
* 4 bytes of little-endian function number:
* 1 = read, 2 = write;
*
* 4 bytes of file handle;
*
* 8 reserved bytes;
*
* 4 bytes of big-endian file offset;
*
* 4 bytes of big-endian byte count.
*
* The data follows for a burst write operation.
*
* The first packet of a burst read reply contains:
*
* 4 bytes of little-endian result code:
* 0: No error
* 1: Initial error
* 2: I/O error
* 3: No data read;
*
* 4 bytes of returned byte count (big-endian?).
*
* The data follows.
*
* Each burst of a write request is responded to with a
* burst packet with a 2-byte little-endian result code:
*
* 0: Write successful
* 4: Write error
*/
flags = tvb_get_guint8(tvb, commhdr + 2);
ti = proto_tree_add_uint(ncp_tree, hf_ncp_system_flags,
tvb, commhdr + 2, 1, flags);
flags_tree = proto_item_add_subtree(ti, ett_ncp_system_flags);
proto_tree_add_item(flags_tree, hf_ncp_system_flags_abt,
tvb, commhdr + 2, 1, FALSE);
if (flags & ABT) {
proto_item_append_text(ti, " ABT");
}
flags&=(~( ABT ));
proto_tree_add_item(flags_tree, hf_ncp_system_flags_bsy,
tvb, commhdr + 2, 1, FALSE);
if (flags & BSY) {
proto_item_append_text(ti, " BSY");
}
flags&=(~( BSY ));
proto_tree_add_item(flags_tree, hf_ncp_system_flags_eob,
tvb, commhdr + 2, 1, FALSE);
if (flags & EOB) {
proto_item_append_text(ti, " EOB");
}
flags&=(~( EOB ));
proto_tree_add_item(flags_tree, hf_ncp_system_flags_lst,
tvb, commhdr + 2, 1, FALSE);
if (flags & LST) {
proto_item_append_text(ti, " LST");
}
flags&=(~( LST ));
proto_tree_add_item(flags_tree, hf_ncp_system_flags_sys,
tvb, commhdr + 2, 1, FALSE);
if (flags & SYS) {
proto_item_append_text(ti, " SYS");
}
flags&=(~( SYS ));
proto_tree_add_item(ncp_tree, hf_ncp_stream_type,
tvb, commhdr + 3, 1, FALSE);
proto_tree_add_item(ncp_tree, hf_ncp_src_connection,
tvb, commhdr + 4, 4, FALSE);
proto_tree_add_item(ncp_tree, hf_ncp_dst_connection,
tvb, commhdr + 8, 4, FALSE);
proto_tree_add_item(ncp_tree, hf_ncp_packet_seqno,
tvb, commhdr + 12, 4, FALSE);
proto_tree_add_item(ncp_tree, hf_ncp_delay_time,
tvb, commhdr + 16, 4, FALSE);
ncp_burst_seqno = tvb_get_ntohs(tvb, commhdr+20);
proto_tree_add_item(ncp_tree, hf_ncp_burst_seqno,
tvb, commhdr + 20, 2, FALSE);
ncp_ack_seqno = tvb_get_ntohs(tvb, commhdr+22);
proto_tree_add_item(ncp_tree, hf_ncp_ack_seqno,
tvb, commhdr + 22, 2, FALSE);
proto_tree_add_item(ncp_tree, hf_ncp_burst_len,
tvb, commhdr + 24, 4, FALSE);
data_offset = tvb_get_ntohl(tvb, commhdr + 28);
proto_tree_add_uint(ncp_tree, hf_ncp_data_offset,
tvb, commhdr + 28, 4, data_offset);
data_len = tvb_get_ntohs(tvb, commhdr + 32);
proto_tree_add_uint(ncp_tree, hf_ncp_data_bytes,
tvb, commhdr + 32, 2, data_len);
missing_fraglist_count = tvb_get_ntohs(tvb, commhdr + 34);
proto_tree_add_item(ncp_tree, hf_ncp_missing_fraglist_count,
tvb, commhdr + 34, 2, FALSE);
offset = commhdr + 36;
if (!(flags & SYS) && ncp_burst_seqno == ncp_ack_seqno &&
data_offset == 0) {
/*
* This is either a Burst Read or Burst Write
* command. The data length includes the burst
* mode header, plus any data in the command
* (there shouldn't be any in a read, but there
* might be some in a write).
*/
if (data_len < 4)
return;
ncp_burst_command = tvb_get_ntohl(tvb, offset);
proto_tree_add_item(ncp_tree, hf_ncp_burst_command,
tvb, offset, 4, FALSE);
offset += 4;
data_len -= 4;
if (data_len < 4)
return;
burst_file = tvb_get_ntohl(tvb, offset);
proto_tree_add_item(ncp_tree, hf_ncp_burst_file_handle,
tvb, offset, 4, FALSE);
offset += 4;
data_len -= 4;
if (data_len < 8)
return;
proto_tree_add_item(ncp_tree, hf_ncp_burst_reserved,
tvb, offset, 8, FALSE);
offset += 8;
data_len -= 8;
if (data_len < 4)
return;
burst_off = tvb_get_ntohl(tvb, offset);
proto_tree_add_uint(ncp_tree, hf_ncp_burst_offset,
tvb, offset, 4, burst_off);
offset += 4;
data_len -= 4;
if (data_len < 4)
return;
burst_len = tvb_get_ntohl(tvb, offset);
proto_tree_add_uint(ncp_tree, hf_ncp_burst_len,
tvb, offset, 4, burst_len);
offset += 4;
data_len -= 4;
if (check_col(pinfo->cinfo, COL_INFO)) {
col_add_fstr(pinfo->cinfo, COL_INFO,
"%s %d bytes starting at offset %d in file 0x%08x",
val_to_str(ncp_burst_command,
burst_command, "Unknown (0x%08x)"),
burst_len, burst_off, burst_file);
}
break;
} else {
if (tvb_get_guint8(tvb, commhdr + 2) & 0x10) {
if (check_col(pinfo->cinfo, COL_INFO)) {
col_set_str(pinfo->cinfo, COL_INFO,
"End of Burst");
}
}
}
break;
case NCP_ALLOCATE_SLOT: /* Allocate Slot Request */
length_remaining = tvb_length_remaining(tvb, commhdr + 4);
if (length_remaining > 4) {
testvar = tvb_get_ntohl(tvb, commhdr+4);
if (testvar == 0x4c495020) {
proto_tree_add_item(ncp_tree, hf_lip_echo, tvb, commhdr+4, 13, FALSE);
break;
}
}
/* otherwise fall through */
case NCP_POSITIVE_ACK: /* Positive Acknowledgement */
case NCP_SERVICE_REQUEST: /* Server NCP Request */
case NCP_SERVICE_REPLY: /* Server NCP Reply */
case NCP_WATCHDOG: /* Watchdog Packet */
case NCP_DEALLOCATE_SLOT: /* Deallocate Slot Request */
default:
proto_tree_add_uint(ncp_tree, hf_ncp_seq, tvb, commhdr + 2, 1, header.sequence);
proto_tree_add_uint(ncp_tree, hf_ncp_connection,tvb, commhdr + 3, 3, nw_connection);
proto_tree_add_item(ncp_tree, hf_ncp_task, tvb, commhdr + 4, 1, FALSE);
break;
}
/*
* Process the packet body.
*/
switch (header.type) {
case NCP_ALLOCATE_SLOT: /* Allocate Slot Request */
length_remaining = tvb_length_remaining(tvb, commhdr + 4);
if (length_remaining > 4) {
testvar = tvb_get_ntohl(tvb, commhdr+4);
if (testvar == 0x4c495020) {
proto_tree_add_text(ncp_tree, tvb, commhdr, -1,
"Lip Echo Packet");
/*break;*/
}
}
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
next_tvb = tvb_new_subset(tvb, commhdr, -1, -1);
dissect_ncp_request(next_tvb, pinfo, nw_connection,
header.sequence, header.type, ncp_tree);
break;
case NCP_DEALLOCATE_SLOT: /* Deallocate Slot Request */
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
next_tvb = tvb_new_subset(tvb, commhdr, -1, -1);
dissect_ncp_request(next_tvb, pinfo, nw_connection,
header.sequence, header.type, ncp_tree);
break;
case NCP_SERVICE_REQUEST: /* Server NCP Request */
case NCP_BROADCAST_SLOT: /* Server Broadcast Packet */
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
next_tvb = tvb_new_subset(tvb, commhdr, -1, -1);
if (tvb_get_guint8(tvb, commhdr+6) == 0x68) {
subfunction = tvb_get_guint8(tvb, commhdr+7);
switch (subfunction) {
case 0x02: /* NDS Frag Packet to decode */
dissect_nds_request(next_tvb, pinfo,
nw_connection, header.sequence,
header.type, ncp_tree);
break;
case 0x01: /* NDS Ping */
dissect_ping_req(next_tvb, pinfo,
nw_connection, header.sequence,
header.type, ncp_tree);
break;
default:
dissect_ncp_request(next_tvb, pinfo,
nw_connection, header.sequence,
header.type, ncp_tree);
break;
}
} else {
dissect_ncp_request(next_tvb, pinfo, nw_connection,
header.sequence, header.type, ncp_tree);
}
break;
case NCP_SERVICE_REPLY: /* Server NCP Reply */
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
next_tvb = tvb_new_subset(tvb, commhdr, -1, -1);
nds_defrag(next_tvb, pinfo, nw_connection, header.sequence,
header.type, ncp_tree, &ncp_tap);
break;
case NCP_POSITIVE_ACK: /* Positive Acknowledgement */
/*
* XXX - this used to call "nds_defrag()", which would
* clear out "frags". Was that the right thing to
* do?
*/
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
next_tvb = tvb_new_subset(tvb, commhdr, -1, -1);
dissect_ncp_reply(next_tvb, pinfo, nw_connection,
header.sequence, header.type, ncp_tree, &ncp_tap);
break;
case NCP_WATCHDOG: /* Watchdog Packet */
/*
* XXX - should the completion code be interpreted as
* it is in "packet-ncp2222.inc"? If so, this
* packet should be handled by "dissect_ncp_reply()".
*/
proto_tree_add_item(ncp_tree, hf_ncp_completion_code,
tvb, commhdr + 6, 1, TRUE);
proto_tree_add_item(ncp_tree, hf_ncp_connection_status,
tvb, commhdr + 7, 1, TRUE);
proto_tree_add_item(ncp_tree, hf_ncp_slot,
tvb, commhdr + 8, 1, TRUE);
proto_tree_add_item(ncp_tree, hf_ncp_control_code,
tvb, commhdr + 9, 1, TRUE);
/*
* Display the rest of the packet as data.
*/
if (tvb_offset_exists(tvb, commhdr + 10)) {
call_dissector(data_handle,
tvb_new_subset(tvb, commhdr + 10, -1, -1),
pinfo, ncp_tree);
}
break;
case NCP_BURST_MODE_XFER: /* Packet Burst Packet */
if (flags & SYS) {
/*
* System packet; show missing fragments if there
* are any.
*/
while (missing_fraglist_count != 0) {
proto_tree_add_item(ncp_tree, hf_ncp_missing_data_offset,
tvb, offset, 4, FALSE);
offset += 4;
proto_tree_add_item(ncp_tree, hf_ncp_missing_data_count,
tvb, offset, 2, FALSE);
offset += 2;
missing_fraglist_count--;
}
} else {
/*
* XXX - do this by using -1 and -1 as the length
* arguments to "tvb_new_subset()" and then calling
* "tvb_set_reported_length()"? That'll throw an
* exception if "data_len" goes past the reported
* length of the packet, but that's arguably a
* feature in this case.
*/
length_remaining = tvb_length_remaining(tvb, offset);
if (length_remaining > data_len)
length_remaining = data_len;
if (data_len != 0) {
call_dissector(data_handle,
tvb_new_subset(tvb, offset,
length_remaining, data_len),
pinfo, ncp_tree);
}
}
break;
case NCP_LIP_ECHO: /* LIP Echo Packet */
proto_tree_add_text(ncp_tree, tvb, commhdr, -1,
"Lip Echo Packet");
break;
default:
if (tree) {
expert_item = proto_tree_add_text(ncp_tree, tvb, commhdr + 6, -1,
"%s packets not supported yet",
val_to_str(header.type, ncp_type_vals,
"Unknown type (0x%04x)"));
if (ncp_echo_err) {
expert_add_info_format(pinfo, expert_item, PI_UNDECODED, PI_NOTE, "%s packets not supported yet", val_to_str(header.type, ncp_type_vals,
"Unknown type (0x%04x)"));
}
}
break;
}
}
static void
dissect_ncp(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree)
{
dissect_ncp_common(tvb, pinfo, tree, FALSE);
}
static guint
get_ncp_pdu_len(packet_info *pinfo _U_, tvbuff_t *tvb, int offset)
{
guint32 signature;
/*
* Check the NCP-over-TCP header signature, to make sure it's there.
* If it's not there, we cannot trust the next 4 bytes to be a
* packet length+"has signature" flag, so we just say the length is
* "what remains in the packet".
*/
signature = tvb_get_ntohl(tvb, offset);
if (signature != NCPIP_RQST && signature != NCPIP_RPLY)
return tvb_length_remaining(tvb, offset);
/*
* Get the length of the NCP-over-TCP packet. Strip off the "has
* signature" flag.
*/
return tvb_get_ntohl(tvb, offset + 4) & 0x7fffffff;
}
static void
dissect_ncp_tcp_pdu(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree)
{
dissect_ncp_common(tvb, pinfo, tree, TRUE);
}
static void
dissect_ncp_tcp(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree)
{
tcp_dissect_pdus(tvb, pinfo, tree, ncp_desegment, 8, get_ncp_pdu_len,
dissect_ncp_tcp_pdu);
}
void
proto_register_ncp(void)
{
static hf_register_info hf[] = {
{ &hf_ncp_ip_sig,
{ "NCP over IP signature", "ncp.ip.signature",
FT_UINT32, BASE_HEX, VALS(ncp_ip_signature), 0x0,
"", HFILL }},
{ &hf_ncp_ip_length,
{ "NCP over IP length", "ncp.ip.length",
FT_UINT32, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_ip_ver,
{ "NCP over IP Version", "ncp.ip.version",
FT_UINT32, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_ip_rplybufsize,
{ "NCP over IP Reply Buffer Size", "ncp.ip.replybufsize",
FT_UINT32, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_ip_packetsig,
{ "NCP over IP Packet Signature", "ncp.ip.packetsig",
FT_BYTES, BASE_NONE, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_type,
{ "Type", "ncp.type",
FT_UINT16, BASE_HEX, VALS(ncp_type_vals), 0x0,
"NCP message type", HFILL }},
{ &hf_ncp_seq,
{ "Sequence Number", "ncp.seq",
FT_UINT8, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_connection,
{ "Connection Number", "ncp.connection",
FT_UINT16, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_task,
{ "Task Number", "ncp.task",
FT_UINT8, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_oplock_flag,
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
{ "Broadcast Message Flag", "ncp.msg_flag",
FT_UINT8, BASE_HEX, VALS(ncp_oplock_vals), 0x0,
"", HFILL }},
{ &hf_ncp_oplock_handle,
{ "File Handle", "ncp.oplock_handle",
FT_UINT16, BASE_HEX, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_stream_type,
{ "Stream Type", "ncp.stream_type",
FT_UINT8, BASE_HEX, NULL, 0x0,
"Type of burst", HFILL }},
{ &hf_ncp_system_flags,
{ "System Flags", "ncp.system_flags",
FT_UINT8, BASE_HEX, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_system_flags_abt,
{ "ABT", "ncp.system_flags.abt",
FT_BOOLEAN, 8, NULL, ABT,
"Is this an abort request?", HFILL }},
{ &hf_ncp_system_flags_eob,
{ "EOB", "ncp.system_flags.eob",
FT_BOOLEAN, 8, NULL, EOB,
"Is this the last packet of the burst?", HFILL }},
{ &hf_ncp_system_flags_sys,
{ "SYS", "ncp.system_flags.sys",
FT_BOOLEAN, 8, NULL, SYS,
"Is this a system packet?", HFILL }},
{ &hf_ncp_system_flags_bsy,
{ "BSY", "ncp.system_flags.bsy",
FT_BOOLEAN, 8, NULL, BSY,
"Is the server busy?", HFILL }},
{ &hf_ncp_system_flags_lst,
{ "LST", "ncp.system_flags.lst",
FT_BOOLEAN, 8, NULL, LST,
"Return Fragment List?", HFILL }},
{ &hf_ncp_src_connection,
{ "Source Connection ID", "ncp.src_connection",
FT_UINT32, BASE_DEC, NULL, 0x0,
"The workstation's connection identification number", HFILL }},
{ &hf_ncp_dst_connection,
{ "Destination Connection ID", "ncp.dst_connection",
FT_UINT32, BASE_DEC, NULL, 0x0,
"The server's connection identification number", HFILL }},
{ &hf_ncp_packet_seqno,
{ "Packet Sequence Number", "ncp.packet_seqno",
FT_UINT32, BASE_DEC, NULL, 0x0,
"Sequence number of this packet in a burst", HFILL }},
{ &hf_ncp_delay_time,
{ "Delay Time", "ncp.delay_time", /* in 100 us increments */
FT_UINT32, BASE_DEC, NULL, 0x0,
"Delay time between consecutive packet sends (100 us increments)", HFILL }},
{ &hf_ncp_burst_seqno,
{ "Burst Sequence Number", "ncp.burst_seqno",
FT_UINT16, BASE_DEC, NULL, 0x0,
"Sequence number of this packet in the burst", HFILL }},
{ &hf_ncp_ack_seqno,
{ "ACK Sequence Number", "ncp.ack_seqno",
FT_UINT16, BASE_DEC, NULL, 0x0,
"Next expected burst sequence number", HFILL }},
{ &hf_ncp_burst_len,
{ "Burst Length", "ncp.burst_len",
FT_UINT32, BASE_DEC, NULL, 0x0,
"Total length of data in this burst", HFILL }},
{ &hf_ncp_burst_offset,
{ "Burst Offset", "ncp.burst_offset",
FT_UINT32, BASE_DEC, NULL, 0x0,
"Offset of data in the burst", HFILL }},
{ &hf_ncp_data_offset,
{ "Data Offset", "ncp.data_offset",
FT_UINT32, BASE_DEC, NULL, 0x0,
"Offset of this packet", HFILL }},
{ &hf_ncp_data_bytes,
{ "Data Bytes", "ncp.data_bytes",
FT_UINT16, BASE_DEC, NULL, 0x0,
"Number of data bytes in this packet", HFILL }},
{ &hf_ncp_missing_fraglist_count,
{ "Missing Fragment List Count", "ncp.missing_fraglist_count",
FT_UINT16, BASE_DEC, NULL, 0x0,
"Number of missing fragments reported", HFILL }},
{ &hf_ncp_missing_data_offset,
{ "Missing Data Offset", "ncp.missing_data_offset",
FT_UINT32, BASE_DEC, NULL, 0x0,
"Offset of beginning of missing data", HFILL }},
{ &hf_ncp_missing_data_count,
{ "Missing Data Count", "ncp.missing_data_count",
FT_UINT16, BASE_DEC, NULL, 0x0,
"Number of bytes of missing data", HFILL }},
{ &hf_ncp_completion_code,
{ "Completion Code", "ncp.completion_code",
FT_UINT8, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_connection_status,
{ "Connection Status", "ncp.connection_status",
FT_UINT8, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_slot,
{ "Slot", "ncp.slot",
FT_UINT8, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_control_code,
{ "Control Code", "ncp.control_code",
FT_UINT8, BASE_DEC, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_fragment_handle,
{ "Fragment Handle", "ncp.fragger_hndl",
FT_UINT16, BASE_HEX, NULL, 0x0,
"", HFILL }},
{ &hf_lip_echo,
{ "Large Internet Packet Echo", "ncp.lip_echo",
FT_STRING, BASE_NONE, NULL, 0x0,
"", HFILL }},
{ &hf_ncp_burst_command,
{ "Burst Command", "ncp.burst_command",
FT_UINT32, BASE_HEX, VALS(burst_command), 0x0,
"Packet Burst Command", HFILL }},
{ &hf_ncp_burst_file_handle,
{ "Burst File Handle", "ncp.file_handle",
FT_UINT32, BASE_HEX, NULL, 0x0,
"Packet Burst File Handle", HFILL }},
{ &hf_ncp_burst_reserved,
{ "Reserved", "ncp.burst_reserved",
FT_BYTES, BASE_HEX, NULL, 0x0, "", HFILL }},
};
static gint *ett[] = {
&ett_ncp,
&ett_ncp_system_flags,
&ett_nds,
&ett_nds_segments,
&ett_nds_segment,
};
module_t *ncp_module;
proto_ncp = proto_register_protocol("NetWare Core Protocol", "NCP", "ncp");
proto_register_field_array(proto_ncp, hf, array_length(hf));
proto_register_subtree_array(ett, array_length(ett));
ncp_module = prefs_register_protocol(proto_ncp, NULL);
prefs_register_obsolete_preference(ncp_module, "initial_hash_size");
prefs_register_bool_preference(ncp_module, "desegment",
"Reassemble NCP-over-TCP messages spanning multiple TCP segments",
"Whether the NCP dissector should reassemble messages spanning multiple TCP segments."
" To use this option, you must also enable \"Allow subdissectors to reassemble TCP streams\" in the TCP protocol settings.",
&ncp_desegment);
prefs_register_bool_preference(ncp_module, "defragment_nds",
"Reassemble fragmented NDS messages spanning multiple reply packets",
"Whether the NCP dissector should defragment NDS messages spanning multiple reply packets.",
&nds_defragment);
Fixes and updated files: File NCP2222.py: 1. Added NCP service type 0xffff = All types 2. Added evaluation of task states. (following bits are defined. 0=normal, 1 = TTS explicit transaction in progress, 2= TTS implicit transaction in progress, 4 = Shared file ste lock in progress) 3. Reversed the Volume Request Flags, this was backwards (s/b 0=do not return name with volume number, 1=Return name with volume number) 4. Fixed endianess of Creator ID in NetWare Information Struct. 5. File information structure incorrectly defined Current Block Being Decompressed. 6. Logical Lock Status structure incorrectly identified Task Number as a byte. It should be two bytes "word". 7. Fixed endianess of Modifier ID in NetWare Information Struct. 8. Fixed Name Space Information structure (was defined as Name Space, changed to Creator Name Space Number). 9. Fixed Semaphore Structure (Task Number was incorrectly defined as a byte, should be two bytes "word"). 10. Added Task structure (Task Number, Task State) 11. Fixed Volume structure (incorrectly defined Volume Name) 12. Added VolumeWithName structure (This struct contains both volume number and volume name) 13. Added error 0x8901 - "No purgable files available" 14. Added error 0x8977 - "Buffer too small" 15. Added error 0x899c - "No more trustees found" 16. Added error 0x89d9 - "Queue station is not a server" 17. Added NCP connection status values (0=Ok, 1=Bad service connection, 10=File server is down, 40=Broadcast Message Pending) 18. Fixed error definitions for NCP 22/50. 19. Fixed reply packet for NCP 22/52 to properly display volume information depending on reply struct, Volume or volume with name. 20. Fixed error definitions for NCP 22/52 21. Fixed NCP 23/26 reply packet to properly return internet address and display correctly. 22. Fixed NCP 23/27 Reply packet structure to properly repeat connection numbers array. 23. Fixed error definitions for NCP 23/33 24. Fixed NCP 23/114 Request (improperly defined Charge Information as a long value, s/b word) 25. Fixed NCP 23/120 Reply (Improperly defined Job Number as a long value, s/b word) 26. Fixed error definitions for NCP 23/123 27. Fixed error definitions for NCP 23/124 28. Fixed error definitions for NCP 23/131 29. Fixed error definitions for NCP 23/132 30. Fixed error definitions for NCP 23/135 31. Fixed error definitions for NCP 23/137 32. Fixed error definitions for NCP 23/138 33. Fixed NCP 23/205 Reply (User Login Allowed incorrectly defined as long value, s/b byte value) 34. Fixed NCP 23/234 Reply (Reply structure incorrectly defined based on old NCP documentation. Corrected per new docs) 35. Fixed endianess of NCP 23/237 Reply on value Number of Locks. 36. Fixed endianess of NCP 23/238 Reply on value Number of Locks. 37. Fixed error definitions for NCP 30 38. Fixed error definitions for NCP 36/6 39. Fixed error definitions for NCP 86/2 40. Fixed error definitions for NCP 86/3 41. Fixed error definitions for NCP 86/4 42. Fixed error definitions for NCP 86/5 43. Fixed error definitions for NCP 87/1 44. Fixed error definitions for NCP 87/4 45. Fixed error definitions for NCP 87/5 46. Fixed error definitions for NCP 87/10 47. Fixed error definitions for NCP 87/11 48. Fixed error definitions for NCP 87/12 49. Fixed error definitions for NCP 87/17 50. Fixed error definitions for NCP 87/18 51. Fixed NCP 87/20 Reply to properly decode multiple entries returned. This used to only dissect the first entry. 52. Fixed error definitions for NCP 87/30 53. Fixed NCP 87/33 Reply to properly utilize the request flags to dissect the reply packet NetWare Info Struct 54. Fixed error definitions for NCP 87/33 55. Fixed error definitions for NCP 88/22 56. Fixed error definitions for NCP 89/1 57. Fixed error definitions for NCP 89/10 58. Fixed NCP 89/11 Request packet structure 59. Fixed error definitions for NCP 89/11 60. Fixed NCP 89/20 Reply to properly decode multiple entries returned. This used to only dissect the first entry. 61. Fixed error definitions for NCP 104/5 62. Added undefined NCP 112 63. Fixed NCP 123/11 Reply to properly dissect FileName, Name, and copyright. 64. Fixed NCP 123/17 Reply to properly display NCP Network Address values. 65. Fixed NCP 123/24 Reply to properly display Driver Board Name, Driver Short Name, and Dirver Logical Name. 66. Fixed error definitions for NCP 123/33 67. Fixed NCP 123/60 to properly display Set Command Name and Set Command Value. 68. Fixed error definitions for NCP 123/70 69. Fixed NCP 123/71 Reply to utilize new File Information Struct 70. Fixed error definitions for NCP 123/71 71. Fixed error definitions for NCP 123/72 72. Added NCP 123/249 73. Added NCP 123/251 74. Added NCP 123/252 75. Added NCP 123/253 76. Added NCP 123/254 77. Added NCP 123/255 78. Fixed error definitions for NCP 131/1 79. Fixed error definitions for NCP 131/2 File packet-ncp2222.inc 1. Added new NCP preference setting to tell Wireshark to decode the NetWare information structure as new or old style. 2. Set default NCP preference settings of echo connection and echo file to FALSE. 3. Added NDS verb 2 Request Flags (0=retain old object, 1=delete old object) 4. Fixed problem where NDS fragmentation could not be reassembled on reload. (fragment array needed to be re-initialized on reload) 5. Fixed NCP service types 1111, 5555, bbbb, and 1111/LIP to reflect real type number. (Was defined as 0xf1, 0xf2, 0xf3, etc... Now defined as 0x1, 0x5, 0xb, etc) 6. Fixed function build_expert_data to parse subtree memory structures to acquire NCP request value records 7. Fix NDS attribute type Boolean to byte value and properly aligned. 8. Fixed attribute zendmSearchOrder to display properly regardless of number or order of value entries. 9. Added check of length of packet prior to attempting to defragment 10. Added for Request packets logic to abort if NCP type isn't found. 11. Added logic to store packet length and then manually decode NCP function 123 based on length. 12. Added logic to manually dissect NCP 87/20 and 89/20 reply packets. 13. NDS resolve name replies with remote entry as the specifier should not store the EID returned (0x00000000) 14. Added logic to trap and echo to expert tap when connection status flags indicate an error. 15. Added manual dissect of NCP 23/26 replies 16. Added logic to Capture the EID returned form NCP 22/51 "Get vol info". 17. Fixed NDS verb 0x2a 18. Fixed NDS verb 0x2b File packet-ncp.c 1. Fix offset for packet signature. (This used to automatically set the offset to account for packet signature if it could not determine the correct NCP type. But for some failed fragment packets, retransmissions, etc, this would be wrongly identified. So first we check to see if we can read a valid type at the offset before we just automatically assume that packet signature is being used) 2. Register the new NCP preference for old/new NetWareInfoStruct. File packet-ncp-int.h 1. Add extern declaration for ncp_newstyle setting. 2. Add length, req_mask, and req_mast_ext to ncp_req_hash_value structure File packet-ncp-sss.c 1. Fix SecretStore request verb Write App Secrets, Client Put Data, to evaluate packet length. svn path=/trunk/; revision=19073
2006-08-29 08:45:53 +00:00
prefs_register_bool_preference(ncp_module, "newstyle",
"Dissect New Netware Information Structure",
"Dissect the NetWare Information Structure as NetWare 5.x or higher or as older NetWare 3.x.",
&ncp_newstyle);
prefs_register_bool_preference(ncp_module, "eid_2_expert",
"Expert: EID to Name lookups?",
"Whether the NCP dissector should echo the NDS Entry ID to name resolves to the expert table.",
&nds_echo_eid);
prefs_register_bool_preference(ncp_module, "connection_2_expert",
"Expert: NCP Connections?",
"Whether the NCP dissector should echo NCP connection information to the expert table.",
&ncp_echo_conn);
prefs_register_bool_preference(ncp_module, "error_2_expert",
"Expert: NCP Errors?",
"Whether the NCP dissector should echo protocol errors to the expert table.",
&ncp_echo_err);
prefs_register_bool_preference(ncp_module, "server_2_expert",
"Expert: Server Information?",
"Whether the NCP dissector should echo server information to the expert table.",
&ncp_echo_server);
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
prefs_register_bool_preference(ncp_module, "file_2_expert",
"Expert: File Information?",
Fixes and updated files: File NCP2222.py: 1. Added NCP service type 0xffff = All types 2. Added evaluation of task states. (following bits are defined. 0=normal, 1 = TTS explicit transaction in progress, 2= TTS implicit transaction in progress, 4 = Shared file ste lock in progress) 3. Reversed the Volume Request Flags, this was backwards (s/b 0=do not return name with volume number, 1=Return name with volume number) 4. Fixed endianess of Creator ID in NetWare Information Struct. 5. File information structure incorrectly defined Current Block Being Decompressed. 6. Logical Lock Status structure incorrectly identified Task Number as a byte. It should be two bytes "word". 7. Fixed endianess of Modifier ID in NetWare Information Struct. 8. Fixed Name Space Information structure (was defined as Name Space, changed to Creator Name Space Number). 9. Fixed Semaphore Structure (Task Number was incorrectly defined as a byte, should be two bytes "word"). 10. Added Task structure (Task Number, Task State) 11. Fixed Volume structure (incorrectly defined Volume Name) 12. Added VolumeWithName structure (This struct contains both volume number and volume name) 13. Added error 0x8901 - "No purgable files available" 14. Added error 0x8977 - "Buffer too small" 15. Added error 0x899c - "No more trustees found" 16. Added error 0x89d9 - "Queue station is not a server" 17. Added NCP connection status values (0=Ok, 1=Bad service connection, 10=File server is down, 40=Broadcast Message Pending) 18. Fixed error definitions for NCP 22/50. 19. Fixed reply packet for NCP 22/52 to properly display volume information depending on reply struct, Volume or volume with name. 20. Fixed error definitions for NCP 22/52 21. Fixed NCP 23/26 reply packet to properly return internet address and display correctly. 22. Fixed NCP 23/27 Reply packet structure to properly repeat connection numbers array. 23. Fixed error definitions for NCP 23/33 24. Fixed NCP 23/114 Request (improperly defined Charge Information as a long value, s/b word) 25. Fixed NCP 23/120 Reply (Improperly defined Job Number as a long value, s/b word) 26. Fixed error definitions for NCP 23/123 27. Fixed error definitions for NCP 23/124 28. Fixed error definitions for NCP 23/131 29. Fixed error definitions for NCP 23/132 30. Fixed error definitions for NCP 23/135 31. Fixed error definitions for NCP 23/137 32. Fixed error definitions for NCP 23/138 33. Fixed NCP 23/205 Reply (User Login Allowed incorrectly defined as long value, s/b byte value) 34. Fixed NCP 23/234 Reply (Reply structure incorrectly defined based on old NCP documentation. Corrected per new docs) 35. Fixed endianess of NCP 23/237 Reply on value Number of Locks. 36. Fixed endianess of NCP 23/238 Reply on value Number of Locks. 37. Fixed error definitions for NCP 30 38. Fixed error definitions for NCP 36/6 39. Fixed error definitions for NCP 86/2 40. Fixed error definitions for NCP 86/3 41. Fixed error definitions for NCP 86/4 42. Fixed error definitions for NCP 86/5 43. Fixed error definitions for NCP 87/1 44. Fixed error definitions for NCP 87/4 45. Fixed error definitions for NCP 87/5 46. Fixed error definitions for NCP 87/10 47. Fixed error definitions for NCP 87/11 48. Fixed error definitions for NCP 87/12 49. Fixed error definitions for NCP 87/17 50. Fixed error definitions for NCP 87/18 51. Fixed NCP 87/20 Reply to properly decode multiple entries returned. This used to only dissect the first entry. 52. Fixed error definitions for NCP 87/30 53. Fixed NCP 87/33 Reply to properly utilize the request flags to dissect the reply packet NetWare Info Struct 54. Fixed error definitions for NCP 87/33 55. Fixed error definitions for NCP 88/22 56. Fixed error definitions for NCP 89/1 57. Fixed error definitions for NCP 89/10 58. Fixed NCP 89/11 Request packet structure 59. Fixed error definitions for NCP 89/11 60. Fixed NCP 89/20 Reply to properly decode multiple entries returned. This used to only dissect the first entry. 61. Fixed error definitions for NCP 104/5 62. Added undefined NCP 112 63. Fixed NCP 123/11 Reply to properly dissect FileName, Name, and copyright. 64. Fixed NCP 123/17 Reply to properly display NCP Network Address values. 65. Fixed NCP 123/24 Reply to properly display Driver Board Name, Driver Short Name, and Dirver Logical Name. 66. Fixed error definitions for NCP 123/33 67. Fixed NCP 123/60 to properly display Set Command Name and Set Command Value. 68. Fixed error definitions for NCP 123/70 69. Fixed NCP 123/71 Reply to utilize new File Information Struct 70. Fixed error definitions for NCP 123/71 71. Fixed error definitions for NCP 123/72 72. Added NCP 123/249 73. Added NCP 123/251 74. Added NCP 123/252 75. Added NCP 123/253 76. Added NCP 123/254 77. Added NCP 123/255 78. Fixed error definitions for NCP 131/1 79. Fixed error definitions for NCP 131/2 File packet-ncp2222.inc 1. Added new NCP preference setting to tell Wireshark to decode the NetWare information structure as new or old style. 2. Set default NCP preference settings of echo connection and echo file to FALSE. 3. Added NDS verb 2 Request Flags (0=retain old object, 1=delete old object) 4. Fixed problem where NDS fragmentation could not be reassembled on reload. (fragment array needed to be re-initialized on reload) 5. Fixed NCP service types 1111, 5555, bbbb, and 1111/LIP to reflect real type number. (Was defined as 0xf1, 0xf2, 0xf3, etc... Now defined as 0x1, 0x5, 0xb, etc) 6. Fixed function build_expert_data to parse subtree memory structures to acquire NCP request value records 7. Fix NDS attribute type Boolean to byte value and properly aligned. 8. Fixed attribute zendmSearchOrder to display properly regardless of number or order of value entries. 9. Added check of length of packet prior to attempting to defragment 10. Added for Request packets logic to abort if NCP type isn't found. 11. Added logic to store packet length and then manually decode NCP function 123 based on length. 12. Added logic to manually dissect NCP 87/20 and 89/20 reply packets. 13. NDS resolve name replies with remote entry as the specifier should not store the EID returned (0x00000000) 14. Added logic to trap and echo to expert tap when connection status flags indicate an error. 15. Added manual dissect of NCP 23/26 replies 16. Added logic to Capture the EID returned form NCP 22/51 "Get vol info". 17. Fixed NDS verb 0x2a 18. Fixed NDS verb 0x2b File packet-ncp.c 1. Fix offset for packet signature. (This used to automatically set the offset to account for packet signature if it could not determine the correct NCP type. But for some failed fragment packets, retransmissions, etc, this would be wrongly identified. So first we check to see if we can read a valid type at the offset before we just automatically assume that packet signature is being used) 2. Register the new NCP preference for old/new NetWareInfoStruct. File packet-ncp-int.h 1. Add extern declaration for ncp_newstyle setting. 2. Add length, req_mask, and req_mast_ext to ncp_req_hash_value structure File packet-ncp-sss.c 1. Fix SecretStore request verb Write App Secrets, Client Put Data, to evaluate packet length. svn path=/trunk/; revision=19073
2006-08-29 08:45:53 +00:00
"Whether the NCP dissector should echo file open/close/oplock information to the expert table.",
From Greg Morris: Modification to (proto.h) is made to add an additional expert group type of PI_REQUEST_CODE to allow Request tag information to be passed to the expert tap. This is for such reasons where a dissector would like to echo specific information about certain types of requests. For example: NCP connection request is really a request not a REPLY_CODE. Same is true for the TCP SYN request. Changes to packet-ncp.c 1. Server broadcast message flag. Now indicates if the message is a pending message or an oplock clear notification. 2. Cleanup of packet signature detection process. Previous method had some flaws so I redesigned it. Appears to be solid now. 3. Echo NCP Server Session information to expert tap. Note on item #3: NCP Connection+Task = NCP Session, a Single connection can have many tasks. The server sees each connection/task as a unique session. For this reason the NCP session information is now echoed to the expert composite statistics so that you can easily identify the different NCP processes and sessions. It is important to NCP analysis to understand that each session is most likely a different program on the requesting host sharing the same NCP connection. Changes to packet-ncp2222.inc 1. Comment out the echo of NCP connection info to expert tap. Replaced by NCP sessions. 2. Add displayEID in request decode (resolves Coverity defect for dead code in NCP dissector) Changes to ncp2222.py 1. Fix for endian display of bindery object type in NCP 0x1720. 2. Fix for size of bindery object type to 2 bytes instead of 4 to match other bindery NCP's. svn path=/trunk/; revision=17636
2006-03-15 18:56:35 +00:00
&ncp_echo_file);
register_init_routine(&mncp_init_protocol);
ncp_tap.stat=register_tap("ncp_srt");
ncp_tap.hdr=register_tap("ncp_hdr");
register_postseq_cleanup_routine(&mncp_postseq_cleanup);
}
void
proto_reg_handoff_ncp(void)
{
dissector_handle_t ncp_handle;
dissector_handle_t ncp_tcp_handle;
ncp_handle = create_dissector_handle(dissect_ncp, proto_ncp);
ncp_tcp_handle = create_dissector_handle(dissect_ncp_tcp, proto_ncp);
dissector_add("tcp.port", TCP_PORT_NCP, ncp_tcp_handle);
dissector_add("udp.port", UDP_PORT_NCP, ncp_handle);
dissector_add("ipx.packet_type", IPX_PACKET_TYPE_NCP, ncp_handle);
dissector_add("ipx.socket", IPX_SOCKET_NCP, ncp_handle);
data_handle = find_dissector("data");
}