aboutsummaryrefslogtreecommitdiffstats
path: root/doc/README.xml-output
blob: fc1d6e2db2c2bc6675b1c13711748472864766ea (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
Protocol Dissection in XML Format
=================================
$Id$
Copyright (c) 2003 by Gilbert Ramirez <gram@alumni.rice.edu>


Wireshark has the ability to export its protocol dissection in an
XML format, tshark has similar functionality by using the "-Tpdml" 
option. 

The XML that wireshark produces follows the Packet Details Markup
Language (PDML) specified by the group at the Politecnico Di Torino
working on Analyzer. The specification can be found at:

http://analyzer.polito.it/30alpha/docs/dissectors/PDMLSpec.htm

A related XML format, the Packet Summary Markup Language (PSML), is
also defined by the Analyzer group to provide packet summary information.
The PSML format is not documented in a publicly-available HTML document,
but its format is simple. Wireshark can export this format too. Some day it 
may be added to tshark so that "-Tpsml" would produce PSML.

One wonders if the "-T" option should read "-Txml" instead of "-Tpdml"
(and in the future, "-Tpsml"), but if tshark was required to produce
another XML-based format of its protocol dissection, then "-Txml" would
be ambiguous.

PDML
====
The PDML that wireshark produces is known not to be loadable into Analyzer.
It causes Analyzer to crash. As such, the PDML that wireshark produces
is be labeled with a version number of "0", which means that the PDML does
not fully follow the PDML spec. Furthermore, a creator attribute in the
"<pdml>" tag gives the version number of wireshark/tshark that produced the PDML.
In that way, as the PDML produced by wireshark matures, but still does not
meet the PDML spec, scripts can make intelligent decisions about how to
best parse the PDML, based on the "creator" attribute.

A PDML file is delimited by a "<pdml>" tag.
A PDML file contains multiple packets, denoted by the "<packet>" tag.
A packet will contain multiple protocols, denoted by the "<proto>" tag.
A protocol might contain one or more fields, denoted by the "<field>" tag.

A pseudo-protocol named "geninfo" is produced, as is required by the PDML
spec, and exported as the first protocol after the opening "<packet>" tag.
Its information comes from wireshark's "frame" protocol, which serves
the similar purpose of storing packet meta-data. Both "geninfo" and
"frame" protocols are provided in the PDML output.

The "<pdml>" tag
================
Example:
	<pdml version="0" creator="wireshark/0.9.17">

The creator is "wireshark" (i.e., the "wireshark" engine. It will always say
"wireshark", not "tshark") version 0.9.17.


The "<proto>" tag
=================
"<proto>" tags can have the following attributes:

	name - the display filter name for the protocol
	showname - the label used to describe this protocol in the protocol
		tree. This is usually the descriptive name of the protocol,
		but it can be modified by dissectors to include more data
		(tcp can do this)
	pos - the starting offset within the packet data where this
		protocol starts
	size - the number of octets in the packet data that this protocol
		covers.

The "<field>" tag
=================
"<field>" tags can have the following attributes:

	name - the display filter name for the field
	showname - the label used to describe this field in the protocol
		tree. This is usually the descriptive name of the protocol,
		followed by some representation of the value.
	pos - the starting offset within the packet data where this
		field starts
	size - the number of octets in the packet data that this field
		covers.
	value - the actual packet data, in hex, that this field covers
	show - the representation of the packet data ('value') as it would
		appear in a display filter.

Some dissectors sometimes place text into the protocol tree, without using
a field with a field-name. Those appear in PDML as "<field>" tags with no
'name' attribute, but with a 'show' attribute giving that text.

Many dissectors label the undissected payload of a protocol as belonging
to a "data" protocol, and the "data" protocol usually resided inside
that last protocol dissected. In the PDML, The "data" protocol becomes
a "data" field, placed exactly where the "data" protocol is in wireshark's
protocol tree. So, if wireshark would normally show:

+-- Frame
|
+-- Ethernet
|
+-- IP
|
+-- TCP
|
+-- HTTP
    |
    +-- Data

In PDML, the "Data" protocol would become another field under HTTP:

<packet>
	<proto name="frame">
	...
	</proto>

	<proto name="eth">
	...
	</proto>

	<proto name="ip">
	...
	</proto>

	<proto name="tcp">
	...
	</proto>

	<proto name="http">
	...
		<field name="data" value="........."/>
	</proto>
</packet>



tools/WiresharkXML.py
====================
This is a python module which provides some infrastructure for
Python developers who wish to parse PDML. It is designed to read
a PDML file and call a user's callback function every time a packet
is constructed from the protocols and fields for a single packet.

The python user should import the module, define a callback function
which accepts one argument, and call the parse_fh function:

------------------------------------------------------------
import WiresharkXML

def my_callback(packet):
	# do something

fh = open(xml_filename)
WiresharkXML.parse_fh(fh, my_callback)

# Now that the script has the packet data, do something.
------------------------------------------------------------

The object that is passed to the callback function is an
WiresharkXML.Packet object, which corresponds to a single packet.
WiresharkXML Provides 3 classes, each of which corresponds to a PDML tag:

	Packet	 - "<packet>" tag
	Protocol - "<proto>" tag
	Field    - "<field>" tag

Each of these classes has accessors which will return the defined attributes:

	get_name()
	get_showname()
	get_pos()
	get_size()
	get_value()
	get_show()

Protocols and fields can contain other fields. Thus, the Protocol and
Field class have a "children" member, which is a simple list of the
Field objects, if any, that are contained. The "children" list can be
directly accessed by calling users. It will be empty of this Protocol
or Field contains no Fields.

Furthermore, the Packet class is a sub-class of the PacketList class.
The PacketList class provides methods to look for protocols and fields.
The term "item" is used when the item being looked for can be
a protocol or a field:

	item_exists(name) - checks if an item exists in the PacketList
	get_items(name) - returns a PacketList of all matching items


General Notes
=============
Generally, parsing XML is slow. If you're writing a script to parse
the PDML output of tshark, pass a read filter with "-R" to tshark to
try to reduce as much as possible the number of packets coming out of tshark.
The less your script has to process, the faster it will be.

'tools/msnchat' is a sample Python program that uses WiresharkXML to parse
PDML. Given one or more capture files, it runs tshark on each of them,
providing a read filter to reduce tshark's output. It finds MSN Chat
conversations in the capture file and produces nice HTML showing the
conversations. It has only been tested with capture files containing
non-simultaneous chat sessions, but was written to more-or-less handle any
number of simultaneous chat sessions.