blob: 0e4c753362f46f1eb4e08d8a3aa7da2c41873a05 (
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
|
/** @file
*
* conv_id 2011 Robert Bullen
*
* Wireshark - Network traffic analyzer
* By Gerald Combs <gerald@wireshark.org>
* Copyright 1998 Gerald Combs
*
* SPDX-License-Identifier: GPL-2.0-or-later
*/
#ifndef __CONV_ID_H__
#define __CONV_ID_H__
/** conv_id_t is a type that can aid in conversation identification. When
* included in a "conversation key", whatever that may be, in addition to the
* typical { address, port, address, port } quadruple, it helps differentiate
* in case the quadruple is not sufficiently unique. For example, it is not
* uncommon to see a TCP quadruple reused these days, and employing a
* conv_id_t field ensures that each instance of a reused TCP conversation is
* tracked independently. Currently this type is used in both Wireshark's and
* tshark's conversation tables implementations (they are different, hence
* the need for a whole header file for this one silly type alias).
*
* The "protocol" or "statistic" code responsible for instantiating the
* "conversation key" is also responsible for assigning its conv_id_t, and
* therefore its interpretation is specific to its assignor. For example, the
* TCP conversations tables in Wireshark and tshark assign the value of
* tcp.stream. If a conv_id_t field is not used, it should be assigned the
* value CONV_ID_UNSET.
*/
typedef guint32 conv_id_t;
#define CONV_ID_UNSET G_MAXUINT32
#endif /* __CONV_ID_H__ */
|
