3 This file is a HOWTO for Ethereal developers. It describes how to start coding
4 a Ethereal protocol dissector and the use some of the important functions and
7 1. Setting up your protocol dissector code.
9 This section provides skeleton code for a protocol dissector. It also explains
10 the basic functions needed to enter values in the traffic summary columns,
11 add to the protocol tree, and work with registered header fields.
17 Ethereal runs on many platforms, and can be compiled with a number of
18 different compilers; here are some rules for writing code that will work
19 on multiple platforms.
21 Don't use C++-style comments (comments beginning with "//" and running
22 to the end of the line); Ethereal's dissectors are written in C, and
23 thus run through C rather than C++ compilers, and not all C compilers
24 support C++-style comments (GCC does, but IBM's C compiler for AIX, for
25 example, doesn't do so by default).
27 Don't initialize variables in their declaration with non-constant
28 values. Not all compilers support this. E.g. don't use
29 guint32 i = somearray[2];
35 Don't use zero-length arrays; not all compilers support them. If an
36 array would have no members, just leave it out.
38 Don't declare variables in the middle of executable code; not all C
39 compilers support that. Variables should be declared outside a
40 function, or at the beginning of a function or compound statement.
42 Don't use "inline"; not all compilers support it. If you want to have a
43 function be an inline function if the compiler supports it, use
44 G_INLINE_FUNC, which is declared by <glib.h>. This may not work with
45 functions declared in header files; if it doesn't work, don't declare
46 the function in a header file, even if this requires that you not make
47 it inline on any platform.
49 Don't use "uchar", "u_char", "ushort", "u_short", "uint", "u_int",
50 "ulong", "u_long" or "boolean"; they aren't defined on all platforms.
51 If you want an 8-bit unsigned quantity, use "guint8"; if you want an
52 8-bit character value with the 8th bit not interpreted as a sign bit,
53 use "guchar"; if you want a 16-bit unsigned quantity, use "guint16";
54 if you want a 32-bit unsigned quantity, use "guint32"; and if you want
55 an "int-sized" unsigned quantity, use "guint"; if you want a boolean,
56 use "gboolean". Use "%d", "%u", "%x", and "%o" to print those types;
57 don't use "%ld", "%lu", "%lx", or "%lo", as longs are 64 bits long on
58 many platforms, but "guint32" is 32 bits long.
60 Don't use "long" to mean "signed 32-bit integer", and don't use
61 "unsigned long" to mean "unsigned 32-bit integer"; "long"s are 64 bits
62 long on many platforms. Use "gint32" for signed 32-bit integers and use
63 "guint32" for unsigned 32-bit integers.
65 Don't use "long" to mean "signed 64-bit integer" and don't use "unsigned
66 long" to mean "unsigned 64-bit integer"; "long"s are 32 bits long on
67 other many platforms. Don't use "long long" or "unsigned long long",
68 either, as not all platforms support them; use "gint64" or "guint64",
69 which will be defined as the appropriate types for 64-bit signed and
72 When printing or displaying the values of 64-bit integral data types,
73 don't assume use "%lld", "%llu", "%llx", or "%llo" - not all platforms
74 support "%ll" for printing 64-bit integral data types. Instead, use
75 PRId64, PRIu64, PRIx64, and PRIo64, for example
77 proto_tree_add_text(tree, tvb, offset, 8,
78 "Sequence Number: %" PRIu64, sequence_number);
80 When specifying an integral constant that doesn't fit in 32 bits, don't
81 use "LL" at the end of the constant - not all compilers use "LL" for
82 that. Instead, put the constant in a call to the "G_GINT64_CONSTANT()"
85 G_GINT64_CONSTANT(11644473600U)
91 Don't use a label without a statement following it. For example,
101 will not work with all compilers - you have to do
111 with some statement, even if it's a null statement, after the label.
113 Don't use "bzero()", "bcopy()", or "bcmp()"; instead, use the ANSI C
116 "memset()" (with zero as the second argument, so that it sets
117 all the bytes to zero);
119 "memcpy()" or "memmove()" (note that the first and second
120 arguments to "memcpy()" are in the reverse order to the
121 arguments to "bcopy()"; note also that "bcopy()" is typically
122 guaranteed to work on overlapping memory regions, while
123 "memcpy()" isn't, so if you may be copying from one region to a
124 region that overlaps it, use "memmove()", not "memcpy()" - but
125 "memcpy()" might be faster as a result of not guaranteeing
126 correct operation on overlapping memory regions);
128 and "memcmp()" (note that "memcmp()" returns 0, 1, or -1, doing
129 an ordered comparison, rather than just returning 0 for "equal"
130 and 1 for "not equal", as "bcmp()" does).
132 Not all platforms necessarily have "bzero()"/"bcopy()"/"bcmp()", and
133 those that do might not declare them in the header file on which they're
134 declared on your platform.
136 Don't use "index()" or "rindex()"; instead, use the ANSI C equivalents,
137 "strchr()" and "strrchr()". Not all platforms necessarily have
138 "index()" or "rindex()", and those that do might not declare them in the
139 header file on which they're declared on your platform.
141 Don't fetch data from packets by getting a pointer to data in the packet
142 with "tvb_get_ptr()", casting that pointer to a pointer to a structure,
143 and dereferencing that pointer. That point won't necessarily be aligned
144 on the proper boundary, which can cause crashes on some platforms (even
145 if it doesn't crash on an x86-based PC); furthermore, the data in a
146 packet is not necessarily in the byte order of the machine on which
147 Ethereal is running. Use the tvbuff routines to extract individual
148 items from the packet, or use "proto_tree_add_item()" and let it extract
151 Don't use "ntohs()", "ntohl()", "htons()", or "htonl()"; the header
152 files required to define or declare them differ between platforms, and
153 you might be able to get away with not including the appropriate header
154 file on your platform but that might not work on other platforms.
155 Instead, use "g_ntohs()", "g_ntohl()", "g_htons()", and "g_htonl()";
156 those are declared by <glib.h>, and you'll need to include that anyway,
157 as Ethereal header files that all dissectors must include use stuff from
160 Don't fetch a little-endian value using "tvb_get_ntohs() or
161 "tvb_get_ntohl()" and then using "g_ntohs()", "g_htons()", "g_ntohl()",
162 or "g_htonl()" on the resulting value - the g_ routines in question
163 convert between network byte order (big-endian) and *host* byte order,
164 not *little-endian* byte order; not all machines on which Ethereal runs
165 are little-endian, even though PC's are. Fetch those values using
166 "tvb_get_letohs()" and "tvb_get_letohl()".
168 Don't put a comma after the last element of an enum - some compilers may
169 either warn about it (producing extra noise) or refuse to accept it.
171 Don't include <unistd.h> without protecting it with
179 and, if you're including it to get routines such as "open()", "close()",
180 "read()", and "write()" declared, also include <io.h> if present:
186 in order to declare the Windows C library routines "_open()",
187 "_close()", "_read()", and "_write()". Your file must include <glib.h>
188 - which many of the Ethereal header files include, so you might not have
189 to include it explicitly - in order to get "open()", "close()",
190 "read()", "write()", etc. mapped to "_open()", "_close()", "_read()",
193 When opening a file with "fopen()", "freopen()", or "fdopen()", if the
194 file contains ASCII text, use "r", "w", "a", and so on as the open mode
195 - but if it contains binary data, use "rb", "wb", and so on. On
196 Windows, if a file is opened in a text mode, writing a byte with the
197 value of octal 12 (newline) to the file causes two bytes, one with the
198 value octal 15 (carriage return) and one with the value octal 12, to be
199 written to the file, and causes bytes with the value octal 15 to be
200 discarded when reading the file (to translate between C's UNIX-style
201 lines that end with newline and Windows' DEC-style lines that end with
202 carriage return/line feed).
204 In addition, that also means that when opening or creating a binary
205 file, you must use "open()" (with O_CREAT and possibly O_TRUNC if the
206 file is to be created if it doesn't exist), and OR in the O_BINARY flag.
207 That flag is not present on most, if not all, UNIX systems, so you must
214 to properly define it for UNIX (it's not necessary on UNIX).
216 Don't use forward declarations of static arrays without a specified size
217 in a fashion such as this:
219 static const value_string foo_vals[];
223 static const value_string foo_vals[] = {
230 as some compilers will reject the first of those statements. Instead,
231 initialize the array at the point at which it's first declared, so that
234 Don't put declarations in the middle of a block; put them before all
235 code. Not all compilers support declarations in the middle of code,
244 For #define names and enum member names, prefix the names with a tag so
245 as to avoid collisions with other names - this might be more of an issue
246 on Windows, as it appears to #define names such as DELETE and
249 Don't use the "numbered argument" feature that many UNIX printf's
252 sprintf(add_string, " - (%1$d) (0x%1$04x)", value);
254 as not all UNIX printf's implement it, and Windows printf doesn't appear
255 to implement it. Use something like
257 sprintf(add_string, " - (%d) (0x%04x)", value, value);
261 Don't use "variadic macros", such as
263 #define DBG(format, args...) fprintf(stderr, format, ## args)
265 as not all C compilers support them. Use macros that take a fixed
266 number of arguments, such as
268 #define DBG0(format) fprintf(stderr, format)
269 #define DBG1(format, arg1) fprintf(stderr, format, arg1)
270 #define DBG2(format, arg1, arg2) fprintf(stderr, format, arg1, arg2)
276 #define DBG(args) printf args
278 snprintf() -> g_snprintf()
279 snprintf() is not available on all platforms, so it's a good idea to use the
280 g_snprintf() function declared by <glib.h> instead.
282 tmpnam() -> mkstemp()
283 tmpnam is insecure and should not be used any more. Ethereal brings its
284 own mkstemp implementation for use on platforms that lack mkstemp.
285 Note: mkstemp does not accept NULL as a parameter.
287 The pointer retured by a call to "tvb_get_ptr()" is not guaranteed to be
288 aligned on any particular byte boundary; this means that you cannot
289 safely cast it to any data type other than a pointer to "char",
290 "unsigned char", "guint8", or other one-byte data types. You cannot,
291 for example, safely cast it to a pointer to a structure, and then access
292 the structure members directly; on some systems, unaligned accesses to
293 integral data types larger than 1 byte, and floating-point data types,
294 cause a trap, which will, at best, result in the OS slowly performing an
295 unaligned access for you, and will, on at least some platforms, cause
296 the program to be terminated.
298 Ethereal supports both platforms with GLib 1.2[.x]/GTK+ 1.2[.x] and GLib
299 2.x/GTK+ 1.3[.x] and 2.x. If at all possible, either use only
300 mechanisms that are present in GLib 1.2[.x] and GTK+ 1.2[.x], use #if's
301 to conditionally use older or newer mechanisms depending on the platform
302 on which Ethereal is being built, or, if the code in GLib or GTK+ that
303 implements that mechanism will build with GLib 1.2[.x]/GTK+ 1.2[.x],
304 conditionally include that code as part of the Ethereal source and use
305 the included version with GLib 1.2[.x] or GTK+ 1.2[.x]. In particular,
306 if the GLib 2.x or GTK+ 2.x mechanism indicates that a routine is
307 deprecated and shouldn't be used in new code, and that it was renamed in
308 GLib 2.x or GTK+ 2.x and the new name should be used, disregard that and
309 use the old name - it'll still work with GLib 2.x or GTK+ 2.x, but will
310 also work with GLib 1.2[.x] and GTK+ 1.2[.x].
312 When different code must be used on UN*X and Win32, use a #if or #ifdef
313 that tests _WIN32, not WIN32. Try to write code portably whenever
314 possible, however; note that there are some routines in Ethereal with
315 platform-dependent implementations and platform-independent APIs, such
316 as the routines in epan/filesystem.c, allowing the code that calls it to
317 be written portably without #ifdefs.
319 1.1.2 String handling
321 Do not use functions such as strcat() or strcpy().
322 A lot of work has been done to remove the existing calls to these functions and
323 we do not want any new callers of these functions.
325 Instead use g_snprintf() since that function will if used correctly prevent
326 buffer overflows for large strings.
328 When using a buffer to create a string, do not use a buffer stored on the stack.
329 I.e. do not use a buffer declared as
331 instead allocate a buffer dynamically using the emem routines (see README.malloc) such as
334 #define MAX_BUFFER 1024
335 buffer=ep_alloc(MAX_BUFFER);
338 g_snprintf(buffer, MAX_BUFFER, ...
340 This avoid the stack to be corrupted in case there is a bug in your code that
341 accidentally writes beyond the end of the buffer.
344 If you write a routine that will create and return a pointer to a filled in
345 string and if that buffer will not be further processed or appended to after
346 the routine returns (except being added to the proto tree),
347 do not preallocate the buffer to fill in and pass as a parameter instead
348 pass a pointer to a pointer to the function and return a pointer to an
349 emem allocated buffer that will be automatically freed. (see README.malloc)
351 I.e. do not write code such as
353 foo_to_str(char *string, ... ){
359 foo_to_str(buffer, ...
360 proto_tree_add_text(... buffer ...
362 instead write the code as
364 foo_to_str(char **buffer, ...
372 foo_to_str(&buffer, ...
373 proto_tree_add_text(... *buffer ...
375 Use ep_ allocated buffers. They are very fast and nice. These buffers are all
376 automatically free()d when the dissection of the current packet ends so you
377 dont have to worry about free()ing them explicitely in order to not leak memory.
378 Please read README.malloc .
383 Ethereal is not guaranteed to read only network traces that contain correctly-
384 formed packets. Ethereal is commonly used is to track down networking problems,
385 and the problems might be due to a buggy protocol implementation sending out
388 Therefore, protocol dissectors not only have to be able to handle
389 correctly-formed packets without, for example, crashing or looping
390 infinitely, they also have to be able to handle *incorrectly*-formed
391 packets without crashing or looping infinitely.
393 Here are some suggestions for making dissectors more robust in the face
394 of incorrectly-formed packets:
396 Do *NOT* use "g_assert()" or "g_assert_not_reached()" in dissectors.
397 *NO* value in a packet's data should be considered "wrong" in the sense
398 that it's a problem with the dissector if found; if it cannot do
399 anything else with a particular value from a packet's data, the
400 dissector should put into the protocol tree an indication that the
401 value is invalid, and should return.
403 If you are allocating a chunk of memory to contain data from a packet,
404 or to contain information derived from data in a packet, and the size of
405 the chunk of memory is derived from a size field in the packet, make
406 sure all the data is present in the packet before allocating the buffer.
409 1) Ethereal won't leak that chunk of memory if an attempt to
410 fetch data not present in the packet throws an exception
414 2) it won't crash trying to allocate an absurdly-large chunk of
415 memory if the size field has a bogus large value.
417 If you're fetching into such a chunk of memory a string from the buffer,
418 and the string has a specified size, you can use "tvb_get_*_string()",
419 which will check whether the entire string is present before allocating
420 a buffer for the string, and will also put a trailing '\0' at the end of
423 If you're fetching into such a chunk of memory a 2-byte Unicode string
424 from the buffer, and the string has a specified size, you can use
425 "tvb_get_ephemeral_faked_unicode()", which will check whether the entire
426 string is present before allocating a buffer for the string, and will also
427 put a trailing '\0' at the end of the buffer. The resulting string will be
428 a sequence of single-byte characters; the only Unicode characters that
429 will be handled correctly are those in the ASCII range. (Ethereal's
430 ability to handle non-ASCII strings is limited; it needs to be
433 If you're fetching into such a chunk of memory a sequence of bytes from
434 the buffer, and the sequence has a specified size, you can use
435 "tvb_memdup()", which will check whether the entire sequence is present
436 before allocating a buffer for it.
438 Otherwise, you can check whether the data is present by using
439 "tvb_ensure_bytes_exist()" or by getting a pointer to the data by using
440 "tvb_get_ptr()", although note that there might be problems with using
441 the pointer from "tvb_get_ptr()" (see the item on this in the
442 Portability section above, and the next item below).
444 Note also that you should only fetch string data into a fixed-length
445 buffer if the code ensures that no more bytes than will fit into the
446 buffer are fetched ("the protocol ensures" isn't good enough, as
447 protocol specifications can't ensure only packets that conform to the
448 specification will be transmitted or that only packets for the protocol
449 in question will be interpreted as packets for that protocol by
450 Ethereal). If there's no maximum length of string data to be fetched,
451 routines such as "tvb_get_*_string()" are safer, as they allocate a buffer
452 large enough to hold the string. (Note that some variants of this call
453 require you to free the string once you're finished with it.)
455 If you have gotten a pointer using "tvb_get_ptr()", you must make sure
456 that you do not refer to any data past the length passed as the last
457 argument to "tvb_get_ptr()"; while the various "tvb_get" routines
458 perform bounds checking and throw an exception if you refer to data not
459 available in the tvbuff, direct references through a pointer gotten from
460 "tvb_get_ptr()" do not do any bounds checking.
462 If you have a loop that dissects a sequence of items, each of which has
463 a length field, with the offset in the tvbuff advanced by the length of
464 the item, then, if the length field is the total length of the item, and
465 thus can be zero, you *MUST* check for a zero-length item and abort the
466 loop if you see one. Otherwise, a zero-length item could cause the
467 dissector to loop infinitely. You should also check that the offset,
468 after having the length added to it, is greater than the offset before
469 the length was added to it, if the length field is greater than 24 bits
470 long, so that, if the length value is *very* large and adding it to the
471 offset causes an overflow, that overflow is detected.
473 If you are fetching a length field from the buffer, corresponding to the
474 length of a portion of the packet, and subtracting from that length a
475 value corresponding to the length of, for example, a header in the
476 packet portion in question, *ALWAYS* check that the value of the length
477 field is greater than or equal to the length you're subtracting from it,
478 and report an error in the packet and stop dissecting the packet if it's
479 less than the length you're subtracting from it. Otherwise, the
480 resulting length value will be negative, which will either cause errors
481 in the dissector or routines called by the dissector, or, if the value
482 is interpreted as an unsigned integer, will cause the value to be
483 interpreted as a very large positive value.
485 Any tvbuff offset that is added to as processing is done on a packet
486 should be stored in a 32-bit variable, such as an "int"; if you store it
487 in an 8-bit or 16-bit variable, you run the risk of the variable
490 sprintf() -> g_snprintf()
491 Prevent yourself from using the sprintf() function, as it does not test the
492 length of the given output buffer and might be writing into memory areas not
493 intended for. This function is one of the main causes of security problems
494 like buffer exploits and many other bugs that are very hard to find. It's
495 much better to use the g_snprintf() function declared by <glib.h> instead.
497 You should test your dissector against incorrectly-formed packets. This
498 can be done using the randpkt and editcap utilities that come with the
499 Ethereal distribution. Testing using randpkt can be done by generating
500 output at the same layer as your protocol, and forcing Ethereal/Tethereal
501 to decode it as your protocol, e.g. if your protocol sits on top of UDP:
503 randpkt -c 50000 -t dns randpkt.pcap
504 tethereal -nVr randpkt.pcap -d udp.port==53,<myproto>
506 Testing using editcap can be done using preexisting capture files and the
507 "-E" flag, which introduces errors in a capture file. E.g.:
509 editcap -E 0.03 infile.pcap outfile.pcap
510 tethereal -nVr outfile.pcap
512 1.1.4 Name convention.
514 Ethereal uses the underscore_convention rather than the InterCapConvention for
515 function names, so new code should probably use underscores rather than
516 intercaps for functions and variable names. This is especially important if you
517 are writing code that will be called from outside your code. We are just
518 trying to keep things consistent for other users.
520 1.1.5 White space convention.
522 Avoid using tab expansions different from 8 spaces, as not all text editors in
523 use by the developers support this.
525 When creating a new file, you are free to choose an indentation logic. Most of
526 the files in Ethereal tend to use 2-space or 4-space indentation. You are
527 encouraged to write a short comment on the indentation logic at the beginning
530 When editing an existing file, try following the existing indentation logic and
531 even if it very tempting, never ever use a restyler/reindenter utility on an
536 Ethereal requires certain things when setting up a protocol dissector.
537 Below is skeleton code for a dissector that you can copy to a file and
538 fill in. Your dissector should follow the naming convention of packet-
539 followed by the abbreviated name for the protocol. It is recommended
540 that where possible you keep to the IANA abbreviated name for the
541 protocol, if there is one, or a commonly-used abbreviation for the
544 Usually, you will put your newly created dissector file into the directory
545 epan/dissectors, just like all the other packet-....c files already in there.
547 Also, please add your dissector file to the corresponding makefile,
548 described in section "1.9 Editing Makefile.common to add your dissector" below.
550 Dissectors that use the dissector registration to register with a lower level
551 dissector don't need to define a prototype in the .h file. For other
552 dissectors the main dissector routine should have a prototype in a header
553 file whose name is "packet-", followed by the abbreviated name for the
554 protocol, followed by ".h"; any dissector file that calls your dissector
555 should be changed to include that file.
557 You may not need to include all the headers listed in the skeleton
558 below, and you may need to include additional headers. For example, the
567 is needed only if you are using a function from libpcre, e.g. the
568 "pcre_compile()" function.
571 in the comment will be updated by CVS when the file is
572 checked in; it will allow the RCS "ident" command to report which
573 version of the file is currently checked out.
575 When creating a new file, it is fine to just write "$Id$" as RCS will
576 automatically fill in the identifier at the time the file will be added to the
577 SVN repository (checked in).
579 ------------------------------------Cut here------------------------------------
580 /* packet-PROTOABBREV.c
581 * Routines for PROTONAME dissection
582 * Copyright 2000, YOUR_NAME <YOUR_EMAIL_ADDRESS>
586 * Ethereal - Network traffic analyzer
587 * By Gerald Combs <gerald@ethereal.com>
588 * Copyright 1998 Gerald Combs
590 * Copied from WHATEVER_FILE_YOU_USED (where "WHATEVER_FILE_YOU_USED"
591 * is a dissector file; if you just copied this from README.developer,
592 * don't bother with the "Copied from" - you don't even need to put
593 * in a "Copied from" if you copied an existing dissector, especially
594 * if the bulk of the code in the new dissector is your code)
596 * This program is free software; you can redistribute it and/or
597 * modify it under the terms of the GNU General Public License
598 * as published by the Free Software Foundation; either version 2
599 * of the License, or (at your option) any later version.
601 * This program is distributed in the hope that it will be useful,
602 * but WITHOUT ANY WARRANTY; without even the implied warranty of
603 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
604 * GNU General Public License for more details.
606 * You should have received a copy of the GNU General Public License
607 * along with this program; if not, write to the Free Software
608 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
621 #include <epan/packet.h>
622 #include <epan/prefs.h>
624 /* IF PROTO exposes code to other dissectors, then it must be exported
625 in a header file. If not, a header file is not needed at all. */
626 #include "packet-PROTOABBREV.h"
628 /* Forward declaration we need below */
629 void proto_reg_handoff_PROTOABBREV(void);
631 /* Initialize the protocol and registered fields */
632 static int proto_PROTOABBREV = -1;
633 static int hf_PROTOABBREV_FIELDABBREV = -1;
635 /* Global sample preference ("controls" display of numbers) */
636 static gboolean gPREF_HEX = FALSE;
638 /* Initialize the subtree pointers */
639 static gint ett_PROTOABBREV = -1;
641 /* Code to actually dissect the packets */
643 dissect_PROTOABBREV(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree)
646 /* Set up structures needed to add the protocol subtree and manage it */
648 proto_tree *PROTOABBREV_tree;
650 /* Make entries in Protocol column and Info column on summary display */
651 if (check_col(pinfo->cinfo, COL_PROTOCOL))
652 col_set_str(pinfo->cinfo, COL_PROTOCOL, "PROTOABBREV");
654 /* This field shows up as the "Info" column in the display; you should use
655 it, if possible, to summarize what's in the packet, so that a user looking
656 at the list of packets can tell what type of packet it is. See section 1.5
657 for more information.
659 Before changing the contents of a column you should make sure the column is
660 active by calling "check_col(pinfo->cinfo, COL_*)". If it is not active
661 don't bother setting it.
663 If you are setting the column to a constant string, use "col_set_str()",
664 as it's more efficient than the other "col_set_XXX()" calls.
666 If you're setting it to a string you've constructed, or will be
667 appending to the column later, use "col_add_str()".
669 "col_add_fstr()" can be used instead of "col_add_str()"; it takes
670 "printf()"-like arguments. Don't use "col_add_fstr()" with a format
671 string of "%s" - just use "col_add_str()" or "col_set_str()", as it's
672 more efficient than "col_add_fstr()".
674 If you will be fetching any data from the packet before filling in
675 the Info column, clear that column first, in case the calls to fetch
676 data from the packet throw an exception because they're fetching data
677 past the end of the packet, so that the Info column doesn't have data
678 left over from the previous dissector; do
680 if (check_col(pinfo->cinfo, COL_INFO))
681 col_clear(pinfo->cinfo, COL_INFO);
685 if (check_col(pinfo->cinfo, COL_INFO))
686 col_set_str(pinfo->cinfo, COL_INFO, "XXX Request");
688 /* A protocol dissector can be called in 2 different ways:
690 (a) Operational dissection
692 In this mode, Ethereal is only interested in the way protocols
693 interact, protocol conversations are created, packets are reassembled
694 and handed over to higher-level protocol dissectors.
695 In this mode Ethereal does not build a so-called "protocol tree".
697 (b) Detailed dissection
699 In this mode, Ethereal is also interested in all details of a given
700 protocol, so a "protocol tree" is created.
702 Ethereal distinguishes between the 2 modes with the proto_tree pointer:
706 In the interest of speed, if "tree" is NULL, avoid building a
707 protocol tree and adding stuff to it, or even looking at any packet
708 data needed only if you're building the protocol tree, if possible.
710 Note, however, that you must fill in column information, create
711 conversations, reassemble packets, build any other persistent state
712 needed for dissection, and call subdissectors regardless of whether
713 "tree" is NULL or not. This might be inconvenient to do without
714 doing most of the dissection work; the routines for adding items to
715 the protocol tree can be passed a null protocol tree pointer, in
716 which case they'll return a null item pointer, and
717 "proto_item_add_subtree()" returns a null tree pointer if passed a
718 null item pointer, so, if you're careful not to dereference any null
719 tree or item pointers, you can accomplish this by doing all the
720 dissection work. This might not be as efficient as skipping that
721 work if you're not building a protocol tree, but if the code would
722 have a lot of tests whether "tree" is null if you skipped that work,
723 you might still be better off just doing all that work regardless of
724 whether "tree" is null or not. */
727 /* NOTE: The offset and length values in the call to
728 "proto_tree_add_item()" define what data bytes to highlight in the hex
729 display window when the line in the protocol tree display
730 corresponding to that item is selected.
732 Supplying a length of -1 is the way to highlight all data from the
733 offset to the end of the packet. */
735 /* create display subtree for the protocol */
736 ti = proto_tree_add_item(tree, proto_PROTOABBREV, tvb, 0, -1, FALSE);
738 PROTOABBREV_tree = proto_item_add_subtree(ti, ett_PROTOABBREV);
740 /* add an item to the subtree, see section 1.6 for more information */
741 proto_tree_add_item(PROTOABBREV_tree,
742 hf_PROTOABBREV_FIELDABBREV, tvb, offset, len, FALSE)
745 /* Continue adding tree items to process the packet here */
750 /* If this protocol has a sub-dissector call it here, see section 1.8 */
754 /* Register the protocol with Ethereal */
756 /* this format is require because a script is used to build the C function
757 that calls all the protocol registration.
761 proto_register_PROTOABBREV(void)
763 module_t *PROTOABBREV_module;
765 /* Setup list of header fields See Section 1.6.1 for details*/
766 static hf_register_info hf[] = {
767 { &hf_PROTOABBREV_FIELDABBREV,
768 { "FIELDNAME", "PROTOABBREV.FIELDABBREV",
769 FIELDTYPE, FIELDBASE, FIELDCONVERT, BITMASK,
770 "FIELDDESCR", HFILL }
774 /* Setup protocol subtree array */
775 static gint *ett[] = {
779 /* Register the protocol name and description */
780 proto_PROTOABBREV = proto_register_protocol("PROTONAME",
781 "PROTOSHORTNAME", "PROTOABBREV");
783 /* Required function calls to register the header fields and subtrees used */
784 proto_register_field_array(proto_PROTOABBREV, hf, array_length(hf));
785 proto_register_subtree_array(ett, array_length(ett));
787 /* Register preferences module (See Section 2.6 for more on preferences) */
788 PROTOABBREV_module = prefs_register_protocol(proto_PROTOABBREV, proto_reg_handoff_PROTOABBREV);
790 /* Register a sample preference */
791 prefs_register_bool_preference(PROTOABBREV_module, "showHex",
792 "Display numbers in Hex",
793 "Enable to display numerical values in hexidecimal.",
798 /* If this dissector uses sub-dissector registration add a registration routine.
799 This exact format is required because a script is used to find these routines
800 and create the code that calls these routines.
802 This function is also called by preferences whenever "Apply" is pressed
803 (see prefs_register_protocol above) so it should accommodate being called
807 proto_reg_handoff_PROTOABBREV(void)
809 static gboolean inited = FALSE;
813 dissector_handle_t PROTOABBREV_handle;
815 PROTOABBREV_handle = create_dissector_handle(dissect_PROTOABBREV,
817 dissector_add("PARENT_SUBFIELD", ID_VALUE, PROTOABBREV_handle);
823 If you perform registration functions which are dependant upon
824 prefs the you should de-register everything which was associated
825 with the previous settings and re-register using the new prefs settings
826 here. In general this means you need to keep track of what value the
827 preference had at the time you registered using a local static in this
830 static int currentPort = -1;
832 if( -1 != currentPort ) {
833 dissector_delete( "tcp.port", currentPort, PROTOABBREV_handle);
836 currentPort = gPortPref;
838 dissector_add("tcp.port", currentPort, PROTOABBREV_handle);
843 ------------------------------------Cut here------------------------------------
845 1.3 Explanation of needed substitutions in code skeleton.
847 In the above code block the following strings should be substituted with
850 YOUR_NAME Your name, of course. You do want credit, don't you?
851 It's the only payment you will receive....
852 YOUR_EMAIL_ADDRESS Keep those cards and letters coming.
853 WHATEVER_FILE_YOU_USED Add this line if you are using another file as a
855 PROTONAME The name of the protocol; this is displayed in the
856 top-level protocol tree item for that protocol.
857 PROTOSHORTNAME An abbreviated name for the protocol; this is displayed
858 in the "Preferences" dialog box if your dissector has
859 any preferences, and in the dialog box for filter fields
860 when constructing a filter expression.
861 PROTOABBREV A name for the protocol for use in filter expressions;
862 it should contain only lower-case letters, digits, and
864 FIELDNAME The displayed name for the header field.
865 FIELDABBREV The abbreviated name for the header field. (NO SPACES)
866 FIELDTYPE FT_NONE, FT_BOOLEAN, FT_UINT8, FT_UINT16, FT_UINT24,
867 FT_UINT32, FT_UINT64, FT_INT8, FT_INT16, FT_INT24, FT_INT32,
868 FT_INT64, FT_FLOAT, FT_DOUBLE, FT_ABSOLUTE_TIME,
869 FT_RELATIVE_TIME, FT_STRING, FT_STRINGZ, FT_UINT_STRING,
870 FT_ETHER, FT_BYTES, FT_IPv4, FT_IPv6, FT_IPXNET,
871 FT_FRAMENUM, FT_PROTOCOL
872 FIELDBASE BASE_NONE, BASE_DEC, BASE_HEX, BASE_OCT
873 FIELDCONVERT VALS(x), TFS(x), NULL
874 BITMASK Usually 0x0 unless using the TFS(x) field conversion.
875 FIELDDESCR A brief description of the field.
876 PARENT_SUBFIELD Lower level protocol field used for lookup, i.e. "tcp.port"
877 ID_VALUE Lower level protocol field value that identifies this protocol
878 For example the TCP or UDP port number
880 If, for example, PROTONAME is "Internet Bogosity Discovery Protocol",
881 PROTOSHORTNAME would be "IBDP", and PROTOABBREV would be "ibdp". Try to
882 conform with IANA names.
884 1.4 The dissector and the data it receives.
889 This is only needed if the dissector doesn't use self-registration to
890 register itself with the lower level dissector, or if the protocol dissector
891 wants/needs to expose code to other subdissectors.
893 The dissector must declared as exactly as follows in the file
894 packet-PROTOABBREV.h:
897 dissect_PROTOABBREV(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree);
900 1.4.2 Extracting data from packets.
902 NOTE: See the README.tvbuff file for more details
904 The "tvb" argument to a dissector points to a buffer containing the raw
905 data to be analyzed by the dissector; for example, for a protocol
906 running atop UDP, it contains the UDP payload (but not the UDP header,
907 or any protocol headers above it). A tvbuffer is a opaque data
908 structure, the internal data structures are hidden and the data must be
909 access via the tvbuffer accessors.
913 Single-byte accessor:
915 guint8 tvb_get_guint8(tvbuff_t*, gint offset);
917 Network-to-host-order accessors for 16-bit integers (guint16), 32-bit
918 integers (guint32), and 24-bit integers:
920 guint16 tvb_get_ntohs(tvbuff_t*, gint offset);
921 guint32 tvb_get_ntohl(tvbuff_t*, gint offset);
922 guint32 tvb_get_ntoh24(tvbuff_t*, gint offset);
924 Network-to-host-order accessors for single-precision and
925 double-precision IEEE floating-point numbers:
927 gfloat tvb_get_ntohieee_float(tvbuff_t*, gint offset);
928 gdouble tvb_get_ntohieee_double(tvbuff_t*, gint offset);
930 Little-Endian-to-host-order accessors for 16-bit integers (guint16),
931 32-bit integers (guint32), and 24-bit integers:
933 guint16 tvb_get_letohs(tvbuff_t*, gint offset);
934 guint32 tvb_get_letohl(tvbuff_t*, gint offset);
935 guint32 tvb_get_letoh24(tvbuff_t*, gint offset);
937 Little-Endian-to-host-order accessors for single-precision and
938 double-precision IEEE floating-point numbers:
940 gfloat tvb_get_letohieee_float(tvbuff_t*, gint offset);
941 gdouble tvb_get_letohieee_double(tvbuff_t*, gint offset);
943 Accessors for IPv4 and IPv6 addresses:
945 guint32 tvb_get_ipv4(tvbuff_t*, gint offset);
946 void tvb_get_ipv6(tvbuff_t*, gint offset, struct e_in6_addr *addr);
948 NOTE: IPv4 addresses are not to be converted to host byte order before
949 being passed to "proto_tree_add_ipv4()". You should use "tvb_get_ipv4()"
950 to fetch them, not "tvb_get_ntohl()" *OR* "tvb_get_letohl()" - don't,
951 for example, try to use "tvb_get_ntohl()", find that it gives you the
952 wrong answer on the PC on which you're doing development, and try
953 "tvb_get_letohl()" instead, as "tvb_get_letohl()" will give the wrong
954 answer on big-endian machines.
958 guint8 *tvb_get_string(tvbuff_t*, gint offset, gint length);
959 guint8 *tvb_get_ephemeral_string(tvbuff_t*, gint offset, gint length);
961 Returns a null-terminated buffer containing data from the specified
962 tvbuff, starting at the specified offset, and containing the specified
963 length worth of characters (the length of the buffer will be length+1,
964 as it includes a null character to terminate the string).
966 tvb_get_string() returns a buffer allocated by g_malloc() so you must
967 g_free() it when you are finished with the string. Failure to g_free() this
968 buffer will lead to memory leaks.
969 tvb_get_ephemeral_string() returns a buffer allocated from a special heap
970 with a lifetime until the next packet is dissected. You do not need to
971 free() this buffer, it will happen automatically once the next packet is
975 guint8 *tvb_get_stringz(tvbuff_t *tvb, gint offset, gint *lengthp);
976 guint8 *tvb_get_ephemeral_stringz(tvbuff_t *tvb, gint offset, gint *lengthp);
978 Returns a null-terminated buffer, allocated with "g_malloc()",
979 containing data from the specified tvbuff, starting with at the
980 specified offset, and containing all characters from the tvbuff up to
981 and including a terminating null character in the tvbuff. "*lengthp"
982 will be set to the length of the string, including the terminating null.
984 tvb_get_stringz() returns a buffer allocated by g_malloc() so you must
985 g_free() it when you are finished with the string. Failure to g_free() this
986 buffer will lead to memory leaks.
987 tvb_get_ephemeral_stringz() returns a buffer allocated from a special heap
988 with a lifetime until the next packet is dissected. You do not need to
989 free() this buffer, it will happen automatically once the next packet is
993 guint8 *tvb_fake_unicode(tvbuff_t*, gint offset, gint length);
994 guint8 *tvb_get_ephemeral_faked_unicode(tvbuff_t*, gint offset, gint length);
996 Converts a 2-byte unicode string to an ASCII string.
997 Returns a null-terminated buffer containing data from the specified
998 tvbuff, starting at the specified offset, and containing the specified
999 length worth of characters (the length of the buffer will be length+1,
1000 as it includes a null character to terminate the string).
1002 tvb_fake_unicode() returns a buffer allocated by g_malloc() so you must
1003 g_free() it when you are finished with the string. Failure to g_free() this
1004 buffer will lead to memory leaks.
1005 tvb_get_ephemeral_faked_unicode() returns a buffer allocated from a special
1006 heap with a lifetime until the next packet is dissected. You do not need to
1007 free() this buffer, it will happen automatically once the next packet is
1012 guint8* tvb_memcpy(tvbuff_t*, guint8* target, gint offset, gint length);
1014 Copies into the specified target the specified length's worth of data
1015 from the specified tvbuff, starting at the specified offset.
1017 guint8* tvb_memdup(tvbuff_t*, gint offset, gint length);
1019 Returns a buffer, allocated with "g_malloc()", containing the specified
1020 length's worth of data from the specified tvbuff, starting at the
1024 /* WARNING! This function is possibly expensive, temporarily allocating
1025 * another copy of the packet data. Furthermore, it's dangerous because once
1026 * this pointer is given to the user, there's no guarantee that the user will
1027 * honor the 'length' and not overstep the boundaries of the buffer.
1029 guint8* tvb_get_ptr(tvbuff_t*, gint offset, gint length);
1031 The reason that tvb_get_ptr() might have to allocate a copy of its data
1032 only occurs with TVBUFF_COMPOSITES, data that spans multiple tvbuffers.
1033 If the user request a pointer to a range of bytes that spans the member
1034 tvbuffs that make up the TVBUFF_COMPOSITE, the data will have to be
1035 copied to another memory region to assure that all the bytes are
1040 1.5 Functions to handle columns in the traffic summary window.
1042 The topmost pane of the main window is a list of the packets in the
1043 capture, possibly filtered by a display filter.
1045 Each line corresponds to a packet, and has one or more columns, as
1046 configured by the user.
1048 Many of the columns are handled by code outside individual dissectors;
1049 most dissectors need only specify the value to put in the "Protocol" and
1052 Columns are specified by COL_ values; the COL_ value for the "Protocol"
1053 field, typically giving an abbreviated name for the protocol (but not
1054 the all-lower-case abbreviation used elsewhere) is COL_PROTOCOL, and the
1055 COL_ value for the "Info" field, giving a summary of the contents of the
1056 packet for that protocol, is COL_INFO.
1058 A value for a column should only be added if the user specified that it
1059 be displayed; to check whether a given column is to be displayed, call
1060 'check_col' with the COL_ value for that field as an argument - it will
1061 return TRUE if the column is to be displayed and FALSE if it is not to
1064 The value for a column can be specified with one of several functions,
1065 all of which take the 'fd' argument to the dissector as their first
1066 argument, and the COL_ value for the column as their second argument.
1068 1.5.1 The col_set_str function.
1070 'col_set_str' takes a string as its third argument, and sets the value
1071 for the column to that value. It assumes that the pointer passed to it
1072 points to a string constant or a static "const" array, not to a
1073 variable, as it doesn't copy the string, it merely saves the pointer
1074 value; the argument can itself be a variable, as long as it always
1075 points to a string constant or a static "const" array.
1077 It is more efficient than 'col_add_str' or 'col_add_fstr'; however, if
1078 the dissector will be using 'col_append_str' or 'col_append_fstr" to
1079 append more information to the column, the string will have to be copied
1080 anyway, so it's best to use 'col_add_str' rather than 'col_set_str' in
1083 For example, to set the "Protocol" column
1086 if (check_col(pinfo->cinfo, COL_PROTOCOL))
1087 col_set_str(pinfo->cinfo, COL_PROTOCOL, "PROTOABBREV");
1090 1.5.2 The col_add_str function.
1092 'col_add_str' takes a string as its third argument, and sets the value
1093 for the column to that value. It takes the same arguments as
1094 'col_set_str', but copies the string, so that if the string is, for
1095 example, an automatic variable that won't remain in scope when the
1096 dissector returns, it's safe to use.
1099 1.5.3 The col_add_fstr function.
1101 'col_add_fstr' takes a 'printf'-style format string as its third
1102 argument, and 'printf'-style arguments corresponding to '%' format
1103 items in that string as its subsequent arguments. For example, to set
1104 the "Info" field to "<XXX> request, <N> bytes", where "reqtype" is a
1105 string containing the type of the request in the packet and "n" is an
1106 unsigned integer containing the number of bytes in the request:
1108 if (check_col(pinfo->cinfo, COL_INFO))
1109 col_add_fstr(pinfo->cinfo, COL_INFO, "%s request, %u bytes",
1112 Don't use 'col_add_fstr' with a format argument of just "%s" -
1113 'col_add_str', or possibly even 'col_set_str' if the string that matches
1114 the "%s" is a static constant string, will do the same job more
1118 1.5.4 The col_clear function.
1120 If the Info column will be filled with information from the packet, that
1121 means that some data will be fetched from the packet before the Info
1122 column is filled in. If the packet is so small that the data in
1123 question cannot be fetched, the routines to fetch the data will throw an
1124 exception (see the comment at the beginning about tvbuffers improving
1125 the handling of short packets - the tvbuffers keep track of how much
1126 data is in the packet, and throw an exception on an attempt to fetch
1127 data past the end of the packet, so that the dissector won't process
1128 bogus data), causing the Info column not to be filled in.
1130 This means that the Info column will have data for the previous
1131 protocol, which would be confusing if, for example, the Protocol column
1132 had data for this protocol.
1134 Therefore, before a dissector fetches any data whatsoever from the
1135 packet (unless it's a heuristic dissector fetching data to determine
1136 whether the packet is one that it should dissect, in which case it
1137 should check, before fetching the data, whether there's any data to
1138 fetch; if there isn't, it should return FALSE), it should set the
1139 Protocol column and the Info column.
1141 If the Protocol column will ultimately be set to, for example, a value
1142 containing a protocol version number, with the version number being a
1143 field in the packet, the dissector should, before fetching the version
1144 number field or any other field from the packet, set it to a value
1145 without a version number, using 'col_set_str', and should later set it
1146 to a value with the version number after it's fetched the version
1149 If the Info column will ultimately be set to a value containing
1150 information from the packet, the dissector should, before fetching any
1151 fields from the packet, clear the column using 'col_clear' (which is
1152 more efficient than clearing it by calling 'col_set_str' or
1153 'col_add_str' with a null string), and should later set it to the real
1154 string after it's fetched the data to use when doing that.
1157 1.5.5 The col_append_str function.
1159 Sometimes the value of a column, especially the "Info" column, can't be
1160 conveniently constructed at a single point in the dissection process;
1161 for example, it might contain small bits of information from many of the
1162 fields in the packet. 'col_append_str' takes, as arguments, the same
1163 arguments as 'col_add_str', but the string is appended to the end of the
1164 current value for the column, rather than replacing the value for that
1165 column. (Note that no blank separates the appended string from the
1166 string to which it is appended; if you want a blank there, you must add
1167 it yourself as part of the string being appended.)
1170 1.5.6 The col_append_fstr function.
1172 'col_append_fstr' is to 'col_add_fstr' as 'col_append_str' is to
1173 'col_add_str' - it takes, as arguments, the same arguments as
1174 'col_add_fstr', but the formatted string is appended to the end of the
1175 current value for the column, rather than replacing the value for that
1178 1.5.7 The col_append_sep_str and col_append_sep_fstr functions.
1180 In specific situations the developer knows that a column's value will be
1181 created in a stepwise manner, where the appended values are listed. Both
1182 'col_append_sep_str' and 'col_append_sep_fstr' functions will add an item
1183 separator between two consecutive items, and will not add the separator at the
1184 beginning of the column. The remainder of the work both functions do is
1185 identical to what 'col_append_str' and 'col_append_fstr' do.
1187 1.6 Constructing the protocol tree.
1189 The middle pane of the main window, and the topmost pane of a packet
1190 popup window, are constructed from the "protocol tree" for a packet.
1192 The protocol tree, or proto_tree, is a GNode, the N-way tree structure
1193 available within GLIB. Of course the protocol dissectors don't care
1194 what a proto_tree really is; they just pass the proto_tree pointer as an
1195 argument to the routines which allow them to add items and new branches
1198 When a packet is selected in the packet-list pane, or a packet popup
1199 window is created, a new logical protocol tree (proto_tree) is created.
1200 The pointer to the proto_tree (in this case, 'protocol tree'), is passed
1201 to the top-level protocol dissector, and then to all subsequent protocol
1202 dissectors for that packet, and then the GUI tree is drawn via
1205 The logical proto_tree needs to know detailed information about the
1206 protocols and fields about which information will be collected from the
1207 dissection routines. By strictly defining (or "typing") the data that can
1208 be attached to a proto tree, searching and filtering becomes possible.
1209 This means that the for every protocol and field (which I also call
1210 "header fields", since they are fields in the protocol headers) which
1211 might be attached to a tree, some information is needed.
1213 Every dissector routine will need to register its protocols and fields
1214 with the central protocol routines (in proto.c). At first I thought I
1215 might keep all the protocol and field information about all the
1216 dissectors in one file, but decentralization seemed like a better idea.
1217 That one file would have gotten very large; one small change would have
1218 required a re-compilation of the entire file. Also, by allowing
1219 registration of protocols and fields at run-time, loadable modules of
1220 protocol dissectors (perhaps even user-supplied) is feasible.
1222 To do this, each protocol should have a register routine, which will be
1223 called when Ethereal starts. The code to call the register routines is
1224 generated automatically; to arrange that a protocol's register routine
1225 be called at startup:
1227 the file containing a dissector's "register" routine must be
1228 added to "DISSECTOR_SRC" in "epan/dissectors/Makefile.common";
1230 the "register" routine must have a name of the form
1231 "proto_register_XXX";
1233 the "register" routine must take no argument, and return no
1236 the "register" routine's name must appear in the source file
1237 either at the beginning of the line, or preceded only by "void "
1238 at the beginning of the line (that'd typically be the
1239 definition) - other white space shouldn't cause a problem, e.g.:
1241 void proto_register_XXX(void) {
1250 proto_register_XXX( void )
1257 and so on should work.
1259 For every protocol or field that a dissector wants to register, a variable of
1260 type int needs to be used to keep track of the protocol. The IDs are
1261 needed for establishing parent/child relationships between protocols and
1262 fields, as well as associating data with a particular field so that it
1263 can be stored in the logical tree and displayed in the GUI protocol
1266 Some dissectors will need to create branches within their tree to help
1267 organize header fields. These branches should be registered as header
1268 fields. Only true protocols should be registered as protocols. This is
1269 so that a display filter user interface knows how to distinguish
1270 protocols from fields.
1272 A protocol is registered with the name of the protocol and its
1275 Here is how the frame "protocol" is registered.
1279 proto_frame = proto_register_protocol (
1281 /* short name */ "Frame",
1282 /* abbrev */ "frame" );
1284 A header field is also registered with its name and abbreviation, but
1285 information about the its data type is needed. It helps to look at
1286 the header_field_info struct to see what information is expected:
1288 struct header_field_info {
1297 int id; /* calculated */
1299 int bitshift; /* calculated */
1304 A string representing the name of the field. This is the name
1305 that will appear in the graphical protocol tree.
1309 A string with an abbreviation of the field. We concatenate the
1310 abbreviation of the parent protocol with an abbreviation for the field,
1311 using a period as a separator. For example, the "src" field in an IP packet
1312 would have "ip.src" as an abbreviation. It is acceptable to have
1313 multiple levels of periods if, for example, you have fields in your
1314 protocol that are then subdivided into subfields. For example, TRMAC
1315 has multiple error fields, so the abbreviations follow this pattern:
1316 "trmac.errors.iso", "trmac.errors.noniso", etc.
1318 The abbreviation is the identifier used in a display filter.
1322 The type of value this field holds. The current field types are:
1324 FT_NONE No field type. Used for fields that
1325 aren't given a value, and that can only
1326 be tested for presence or absence; a
1327 field that represents a data structure,
1328 with a subtree below it containing
1329 fields for the members of the structure,
1330 or that represents an array with a
1331 subtree below it containing fields for
1332 the members of the array, might be an
1334 FT_PROTOCOL Used for protocols which will be placing
1335 themselves as top-level items in the
1336 "Packet Details" pane of the UI.
1337 FT_BOOLEAN 0 means "false", any other value means
1339 FT_FRAMENUM A frame number; if this is used, the "Go
1340 To Corresponding Frame" menu item can
1342 FT_UINT8 An 8-bit unsigned integer.
1343 FT_UINT16 A 16-bit unsigned integer.
1344 FT_UINT24 A 24-bit unsigned integer.
1345 FT_UINT32 A 32-bit unsigned integer.
1346 FT_UINT64 A 64-bit unsigned integer.
1347 FT_INT8 An 8-bit signed integer.
1348 FT_INT16 A 16-bit signed integer.
1349 FT_INT24 A 24-bit signed integer.
1350 FT_INT32 A 32-bit signed integer.
1351 FT_INT64 A 64-bit signed integer.
1352 FT_FLOAT A single-precision floating point number.
1353 FT_DOUBLE A double-precision floating point number.
1354 FT_ABSOLUTE_TIME Seconds (4 bytes) and nanoseconds (4 bytes)
1355 of time displayed as month name, month day,
1356 year, hours, minutes, and seconds with 9
1357 digits after the decimal point.
1358 FT_RELATIVE_TIME Seconds (4 bytes) and nanoseconds (4 bytes)
1359 of time displayed as seconds and 9 digits
1360 after the decimal point.
1361 FT_STRING A string of characters, not necessarily
1362 NUL-terminated, but possibly NUL-padded.
1363 This, and the other string-of-characters
1364 types, are to be used for text strings,
1365 not raw binary data.
1366 FT_STRINGZ A NUL-terminated string of characters.
1367 FT_UINT_STRING A counted string of characters, consisting
1368 of a count (represented as an integral
1369 value) followed immediately by the
1370 specified number of characters.
1371 FT_ETHER A six octet string displayed in
1372 Ethernet-address format.
1373 FT_BYTES A string of bytes with arbitrary values;
1374 used for raw binary data.
1375 FT_IPv4 A version 4 IP address (4 bytes) displayed
1376 in dotted-quad IP address format (4
1377 decimal numbers separated by dots).
1378 FT_IPv6 A version 6 IP address (16 bytes) displayed
1379 in standard IPv6 address format.
1380 FT_IPXNET An IPX address displayed in hex as a 6-byte
1381 network number followed by a 6-byte station
1384 Some of these field types are still not handled in the display filter
1385 routines, but the most common ones are. The FT_UINT* variables all
1386 represent unsigned integers, and the FT_INT* variables all represent
1387 signed integers; the number on the end represent how many bits are used
1388 to represent the number.
1392 The display field has a couple of overloaded uses. This is unfortunate,
1393 but since we're C as an application programming language, this sometimes
1394 makes for cleaner programs. Right now I still think that overloading
1395 this variable was okay.
1397 For integer fields (FT_UINT* and FT_INT*), this variable represents the
1398 base in which you would like the value displayed. The acceptable bases
1405 BASE_DEC, BASE_HEX, and BASE_OCT are decimal, hexadecimal, and octal,
1408 For FT_BOOLEAN fields that are also bitfields, 'display' is used to tell
1409 the proto_tree how wide the parent bitfield is. With integers this is
1410 not needed since the type of integer itself (FT_UINT8, FT_UINT16,
1411 FT_UINT24, FT_UINT32, etc.) tells the proto_tree how wide the parent
1414 Additionally, BASE_NONE is used for 'display' as a NULL-value. That is,
1415 for non-integers and non-bitfield FT_BOOLEANs, you'll want to use BASE_NONE
1416 in the 'display' field. You may not use BASE_NONE for integers.
1418 It is possible that in the future we will record the endianness of
1419 integers. If so, it is likely that we'll use a bitmask on the display field
1420 so that integers would be represented as BEND|BASE_DEC or LEND|BASE_HEX.
1421 But that has not happened yet.
1425 Some integer fields, of type FT_UINT*, need labels to represent the true
1426 value of a field. You could think of those fields as having an
1427 enumerated data type, rather than an integral data type.
1429 A 'value_string' structure is a way to map values to strings.
1431 typedef struct _value_string {
1436 For fields of that type, you would declare an array of "value_string"s:
1438 static const value_string valstringname[] = {
1439 { INTVAL1, "Descriptive String 1" },
1440 { INTVAL2, "Descriptive String 2" },
1444 (the last entry in the array must have a NULL 'strptr' value, to
1445 indicate the end of the array). The 'strings' field would be set to
1446 'VALS(valstringname)'.
1448 If the field has a numeric rather than an enumerated type, the 'strings'
1449 field would be set to NULL.
1451 FT_BOOLEANS have a default map of 0 = "False", 1 (or anything else) = "True".
1452 Sometimes it is useful to change the labels for boolean values (e.g.,
1453 to "Yes"/"No", "Fast"/"Slow", etc.). For these mappings, a struct called
1454 true_false_string is used. (This struct is new as of Ethereal 0.7.6).
1456 typedef struct true_false_string {
1459 } true_false_string;
1461 For Boolean fields for which "False" and "True" aren't the desired
1462 labels, you would declare a "true_false_string"s:
1464 static const true_false_string boolstringname = {
1469 Its two fields are pointers to the string representing truth, and the
1470 string representing falsehood. For FT_BOOLEAN fields that need a
1471 'true_false_string' struct, the 'strings' field would be set to
1472 'TFS(&boolstringname)'.
1474 If the Boolean field is to be displayed as "False" or "True", the
1475 'strings' field would be set to NULL.
1479 If the field is a bitfield, then the bitmask is the mask which will
1480 leave only the bits needed to make the field when ANDed with a value.
1481 The proto_tree routines will calculate 'bitshift' automatically
1482 from 'bitmask', by finding the rightmost set bit in the bitmask.
1483 If the field is not a bitfield, then bitmask should be set to 0.
1487 This is a string giving a proper description of the field.
1488 It should be at least one grammatically complete sentence.
1489 It is meant to provide a more detailed description of the field than the
1490 name alone provides. This information will be used in the man page, and
1491 in a future GUI display-filter creation tool. We might also add tooltips
1492 to the labels in the GUI protocol tree, in which case the blurb would
1493 be used as the tooltip text.
1496 1.6.1 Field Registration.
1498 Protocol registration is handled by creating an instance of the
1499 header_field_info struct (or an array of such structs), and
1500 calling the registration function along with the registration ID of
1501 the protocol that is the parent of the fields. Here is a complete example:
1503 static int proto_eg = -1;
1504 static int hf_field_a = -1;
1505 static int hf_field_b = -1;
1507 static hf_register_info hf[] = {
1510 { "Field A", "proto.field_a", FT_UINT8, BASE_HEX, NULL,
1511 0xf0, "Field A represents Apples", HFILL }},
1514 { "Field B", "proto.field_b", FT_UINT16, BASE_DEC, VALS(vs),
1515 0x0, "Field B represents Bananas", HFILL }}
1518 proto_eg = proto_register_protocol("Example Protocol",
1520 proto_register_field_array(proto_eg, hf, array_length(hf));
1522 Be sure that your array of hf_register_info structs is declared 'static',
1523 since the proto_register_field_array() function does not create a copy
1524 of the information in the array... it uses that static copy of the
1525 information that the compiler created inside your array. Here's the
1526 layout of the hf_register_info struct:
1528 typedef struct hf_register_info {
1529 int *p_id; /* pointer to parent variable */
1530 header_field_info hfinfo;
1533 Also be sure to use the handy array_length() macro found in packet.h
1534 to have the compiler compute the array length for you at compile time.
1536 If you don't have any fields to register, do *NOT* create a zero-length
1537 "hf" array; not all compilers used to compile Ethereal support them.
1538 Just omit the "hf" array, and the "proto_register_field_array()" call,
1541 It is OK to have header fields with a different format be registered with
1542 the same abbreviation. For instance, the following is valid:
1544 static hf_register_info hf[] = {
1546 { &hf_field_8bit, /* 8-bit version of proto.field */
1547 { "Field (8 bit)", "proto.field", FT_UINT8, BASE_DEC, NULL,
1548 0x00, "Field represents FOO", HFILL }},
1550 { &hf_field_32bit, /* 32-bit version of proto.field */
1551 { "Field (32 bit)", "proto.field", FT_UINT32, BASE_DEC, NULL,
1552 0x00, "Field represents FOO", HFILL }}
1555 This way a filter expression can match a header field, irrespective of the
1556 representation of it in the specific protocol context. This is interesting
1557 for protocols with variable-width header fields.
1559 The HFILL macro at the end of the struct will set resonable default values
1560 for internally used fields.
1562 1.6.2 Adding Items and Values to the Protocol Tree.
1564 A protocol item is added to an existing protocol tree with one of a
1565 handful of proto_XXX_DO_YYY() functions.
1567 Remember that it only makes sense to add items to a protocol tree if its
1568 proto_tree pointer is not null. Should you add an item to a NULL tree, then
1569 the proto_XXX_DO_YYY() function will immediately return. The cost of this
1570 function call can be avoided by checking for the tree pointer.
1572 Subtrees can be made with the proto_item_add_subtree() function:
1574 item = proto_tree_add_item(....);
1575 new_tree = proto_item_add_subtree(item, tree_type);
1577 This will add a subtree under the item in question; a subtree can be
1578 created under an item made by any of the "proto_tree_add_XXX" functions,
1579 so that the tree can be given an arbitrary depth.
1581 Subtree types are integers, assigned by
1582 "proto_register_subtree_array()". To register subtree types, pass an
1583 array of pointers to "gint" variables to hold the subtree type values to
1584 "proto_register_subtree_array()":
1586 static gint ett_eg = -1;
1587 static gint ett_field_a = -1;
1589 static gint *ett[] = {
1594 proto_register_subtree_array(ett, array_length(ett));
1596 in your "register" routine, just as you register the protocol and the
1597 fields for that protocol.
1599 There are several functions that the programmer can use to add either
1600 protocol or field labels to the proto_tree:
1603 proto_tree_add_item(tree, id, tvb, start, length, little_endian);
1606 proto_tree_add_item_hidden(tree, id, tvb, start, length, little_endian);
1609 proto_tree_add_none_format(tree, id, tvb, start, length, format, ...);
1612 proto_tree_add_protocol_format(tree, id, tvb, start, length,
1616 proto_tree_add_bytes(tree, id, tvb, start, length, start_ptr);
1619 proto_tree_add_bytes_hidden(tree, id, tvb, start, length, start_ptr);
1622 proto_tree_add_bytes_format(tree, id, tvb, start, length, start_ptr,
1626 proto_tree_add_time(tree, id, tvb, start, length, value_ptr);
1629 proto_tree_add_time_hidden(tree, id, tvb, start, length, value_ptr);
1632 proto_tree_add_time_format(tree, id, tvb, start, length, value_ptr,
1636 proto_tree_add_ipxnet(tree, id, tvb, start, length, value);
1639 proto_tree_add_ipxnet_hidden(tree, id, tvb, start, length, value);
1642 proto_tree_add_ipxnet_format(tree, id, tvb, start, length, value,
1646 proto_tree_add_ipv4(tree, id, tvb, start, length, value);
1649 proto_tree_add_ipv4_hidden(tree, id, tvb, start, length, value);
1652 proto_tree_add_ipv4_format(tree, id, tvb, start, length, value,
1656 proto_tree_add_ipv6(tree, id, tvb, start, length, value_ptr);
1659 proto_tree_add_ipv6_hidden(tree, id, tvb, start, length, value_ptr);
1662 proto_tree_add_ipv6_format(tree, id, tvb, start, length, value_ptr,
1666 proto_tree_add_ether(tree, id, tvb, start, length, value_ptr);
1669 proto_tree_add_ether_hidden(tree, id, tvb, start, length, value_ptr);
1672 proto_tree_add_ether_format(tree, id, tvb, start, length, value_ptr,
1676 proto_tree_add_string(tree, id, tvb, start, length, value_ptr);
1679 proto_tree_add_string_hidden(tree, id, tvb, start, length, value_ptr);
1682 proto_tree_add_string_format(tree, id, tvb, start, length, value_ptr,
1686 proto_tree_add_boolean(tree, id, tvb, start, length, value);
1689 proto_tree_add_boolean_hidden(tree, id, tvb, start, length, value);
1692 proto_tree_add_boolean_format(tree, id, tvb, start, length, value,
1696 proto_tree_add_float(tree, id, tvb, start, length, value);
1699 proto_tree_add_float_hidden(tree, id, tvb, start, length, value);
1702 proto_tree_add_float_format(tree, id, tvb, start, length, value,
1706 proto_tree_add_double(tree, id, tvb, start, length, value);
1709 proto_tree_add_double_hidden(tree, id, tvb, start, length, value);
1712 proto_tree_add_double_format(tree, id, tvb, start, length, value,
1716 proto_tree_add_uint(tree, id, tvb, start, length, value);
1719 proto_tree_add_uint_hidden(tree, id, tvb, start, length, value);
1722 proto_tree_add_uint_format(tree, id, tvb, start, length, value,
1726 proto_tree_add_uint64(tree, id, tvb, start, length, value);
1729 proto_tree_add_uint64_format(tree, id, tvb, start, length, value,
1733 proto_tree_add_int(tree, id, tvb, start, length, value);
1736 proto_tree_add_int_hidden(tree, id, tvb, start, length, value);
1739 proto_tree_add_int_format(tree, id, tvb, start, length, value,
1743 proto_tree_add_int64(tree, id, tvb, start, length, value);
1746 proto_tree_add_int64_format(tree, id, tvb, start, length, value,
1750 proto_tree_add_text(tree, tvb, start, length, format, ...);
1753 proto_tree_add_text_valist(tree, tvb, start, length, format, ap);
1755 The 'tree' argument is the tree to which the item is to be added. The
1756 'tvb' argument is the tvbuff from which the item's value is being
1757 extracted; the 'start' argument is the offset from the beginning of that
1758 tvbuff of the item being added, and the 'length' argument is the length,
1759 in bytes, of the item.
1761 The length of some items cannot be determined until the item has been
1762 dissected; to add such an item, add it with a length of -1, and, when the
1763 dissection is complete, set the length with 'proto_item_set_len()':
1766 proto_item_set_len(ti, length);
1768 The "ti" argument is the value returned by the call that added the item
1769 to the tree, and the "length" argument is the length of the item.
1771 proto_tree_add_item()
1772 ---------------------
1773 proto_tree_add_item is used when you wish to do no special formatting.
1774 The item added to the GUI tree will contain the name (as passed in the
1775 proto_register_*() function) and a value. The value will be fetched
1776 from the tvbuff by proto_tree_add_item(), based on the type of the field
1777 and, for integral and Boolean fields, the byte order of the value; the
1778 byte order is specified by the 'little_endian' argument, which is TRUE
1779 if the value is little-endian and FALSE if it is big-endian.
1781 Now that definitions of fields have detailed information about bitfield
1782 fields, you can use proto_tree_add_item() with no extra processing to
1783 add bitfield values to your tree. Here's an example. Take the Format
1784 Identifer (FID) field in the Transmission Header (TH) portion of the SNA
1785 protocol. The FID is the high nibble of the first byte of the TH. The
1786 FID would be registered like this:
1788 name = "Format Identifer"
1789 abbrev = "sna.th.fid"
1792 strings = sna_th_fid_vals
1795 The bitmask contains the value which would leave only the FID if bitwise-ANDed
1796 against the parent field, the first byte of the TH.
1798 The code to add the FID to the tree would be;
1800 proto_tree_add_item(bf_tree, hf_sna_th_fid, tvb, offset, 1, TRUE);
1802 The definition of the field already has the information about bitmasking
1803 and bitshifting, so it does the work of masking and shifting for us!
1804 This also means that you no longer have to create value_string structs
1805 with the values bitshifted. The value_string for FID looks like this,
1806 even though the FID value is actually contained in the high nibble.
1807 (You'd expect the values to be 0x0, 0x10, 0x20, etc.)
1809 /* Format Identifier */
1810 static const value_string sna_th_fid_vals[] = {
1811 { 0x0, "SNA device <--> Non-SNA Device" },
1812 { 0x1, "Subarea Node <--> Subarea Node" },
1813 { 0x2, "Subarea Node <--> PU2" },
1814 { 0x3, "Subarea Node or SNA host <--> Subarea Node" },
1817 { 0xf, "Adjaced Subarea Nodes" },
1821 The final implication of this is that display filters work the way you'd
1822 naturally expect them to. You'd type "sna.th.fid == 0xf" to find Adjacent
1823 Subarea Nodes. The user does not have to shift the value of the FID to
1824 the high nibble of the byte ("sna.th.fid == 0xf0") as was necessary
1825 before Ethereal 0.7.6.
1827 proto_tree_add_item_hidden()
1828 ----------------------------
1829 proto_tree_add_item_hidden is used to add fields and values to a tree,
1830 but not show them on a GUI tree. The caller may want a value to be
1831 included in a tree so that the packet can be filtered on this field, but
1832 the representation of that field in the tree is not appropriate. An
1833 example is the token-ring routing information field (RIF). The best way
1834 to show the RIF in a GUI is by a sequence of ring and bridge numbers.
1835 Rings are 3-digit hex numbers, and bridges are single hex digits:
1837 RIF: 001-A-013-9-C0F-B-555
1839 In the case of RIF, the programmer should use a field with no value and
1840 use proto_tree_add_none_format() to build the above representation. The
1841 programmer can then add the ring and bridge values, one-by-one, with
1842 proto_tree_add_item_hidden() so that the user can then filter on or
1843 search for a particular ring or bridge. Here's a skeleton of how the
1844 programmer might code this.
1847 rif = create_rif_string(...);
1849 proto_tree_add_none_format(tree, hf_tr_rif_label, ..., "RIF: %s", rif);
1851 for(i = 0; i < num_rings; i++) {
1852 proto_tree_add_item_hidden(tree, hf_tr_rif_ring, ..., FALSE);
1854 for(i = 0; i < num_rings - 1; i++) {
1855 proto_tree_add_item_hidden(tree, hf_tr_rif_bridge, ..., FALSE);
1858 The logical tree has these items:
1860 hf_tr_rif_label, text="RIF: 001-A-013-9-C0F-B-555", value = NONE
1861 hf_tr_rif_ring, hidden, value=0x001
1862 hf_tr_rif_bridge, hidden, value=0xA
1863 hf_tr_rif_ring, hidden, value=0x013
1864 hf_tr_rif_bridge, hidden, value=0x9
1865 hf_tr_rif_ring, hidden, value=0xC0F
1866 hf_tr_rif_bridge, hidden, value=0xB
1867 hf_tr_rif_ring, hidden, value=0x555
1869 GUI or print code will not display the hidden fields, but a display
1870 filter or "packet grep" routine will still see the values. The possible
1871 filter is then possible:
1873 tr.rif_ring eq 0x013
1875 proto_tree_add_protocol_format()
1876 ----------------------------
1877 proto_tree_add_protocol_format is used to add the top-level item for the
1878 protocol when the dissector routines wants complete control over how the
1879 field and value will be represented on the GUI tree. The ID value for
1880 the protocol is passed in as the "id" argument; the rest of the
1881 arguments are a "printf"-style format and any arguments for that format.
1882 The caller must include the name of the protocol in the format; it is
1883 not added automatically as in proto_tree_add_item().
1885 proto_tree_add_none_format()
1886 ----------------------------
1887 proto_tree_add_none_format is used to add an item of type FT_NONE.
1888 The caller must include the name of the field in the format; it is
1889 not added automatically as in proto_tree_add_item().
1891 proto_tree_add_bytes()
1892 proto_tree_add_time()
1893 proto_tree_add_ipxnet()
1894 proto_tree_add_ipv4()
1895 proto_tree_add_ipv6()
1896 proto_tree_add_ether()
1897 proto_tree_add_string()
1898 proto_tree_add_boolean()
1899 proto_tree_add_float()
1900 proto_tree_add_double()
1901 proto_tree_add_uint()
1902 proto_tree_add_uint64()
1903 proto_tree_add_int()
1904 proto_tree_add_int64()
1905 ----------------------------
1906 These routines are used to add items to the protocol tree if either:
1908 the value of the item to be added isn't just extracted from the
1909 packet data, but is computed from data in the packet;
1911 the value was fetched into a variable.
1913 The 'value' argument has the value to be added to the tree.
1915 NOTE: in all cases where the 'value' argument is a pointer, a copy is
1916 made of the object pointed to; if you have dynamically allocated a
1917 buffer for the object, that buffer will not be freed when the protocol
1918 tree is freed - you must free the buffer yourself when you don't need it
1921 For proto_tree_add_bytes(), the 'value_ptr' argument is a pointer to a
1924 For proto_tree_add_time(), the 'value_ptr' argument is a pointer to an
1925 "nstime_t", which is a structure containing the time to be added; it has
1926 'secs' and 'nsecs' members, giving the integral part and the fractional
1927 part of a time in units of seconds, with 'nsecs' being the number of
1928 nanoseconds. For absolute times, "secs" is a UNIX-style seconds since
1929 January 1, 1970, 00:00:00 GMT value.
1931 For proto_tree_add_ipxnet(), the 'value' argument is a 32-bit IPX
1934 For proto_tree_add_ipv4(), the 'value' argument is a 32-bit IPv4
1935 address, in network byte order.
1937 For proto_tree_add_ipv6(), the 'value_ptr' argument is a pointer to a
1938 128-bit IPv6 address.
1940 For proto_tree_add_ether(), the 'value_ptr' argument is a pointer to a
1943 For proto_tree_add_string(), the 'value_ptr' argument is a pointer to a
1946 For proto_tree_add_boolean(), the 'value' argument is a 32-bit integer;
1947 zero means "false", and non-zero means "true".
1949 For proto_tree_add_float(), the 'value' argument is a 'float' in the
1950 host's floating-point format.
1952 For proto_tree_add_double(), the 'value' argument is a 'double' in the
1953 host's floating-point format.
1955 For proto_tree_add_uint(), the 'value' argument is a 32-bit unsigned
1956 integer value, in host byte order. (This routine cannot be used to add
1959 For proto_tree_add_uint64(), the 'value' argument is a 64-bit unsigned
1960 integer value, in host byte order.
1962 For proto_tree_add_int(), the 'value' argument is a 32-bit signed
1963 integer value, in host byte order. (This routine cannot be used to add
1966 For proto_tree_add_int64(), the 'value' argument is a 64-bit signed
1967 integer value, in host byte order.
1969 proto_tree_add_bytes_hidden()
1970 proto_tree_add_time_hidden()
1971 proto_tree_add_ipxnet_hidden()
1972 proto_tree_add_ipv4_hidden()
1973 proto_tree_add_ipv6_hidden()
1974 proto_tree_add_ether_hidden()
1975 proto_tree_add_string_hidden()
1976 proto_tree_add_boolean_hidden()
1977 proto_tree_add_float_hidden()
1978 proto_tree_add_double_hidden()
1979 proto_tree_add_uint_hidden()
1980 proto_tree_add_int_hidden()
1981 ----------------------------
1982 These routines add fields and values to a tree, but don't show them in
1983 the GUI tree. They are used for the same reason that
1984 proto_tree_add_item() is used.
1986 proto_tree_add_bytes_format()
1987 proto_tree_add_time_format()
1988 proto_tree_add_ipxnet_format()
1989 proto_tree_add_ipv4_format()
1990 proto_tree_add_ipv6_format()
1991 proto_tree_add_ether_format()
1992 proto_tree_add_string_format()
1993 proto_tree_add_boolean_format()
1994 proto_tree_add_float_format()
1995 proto_tree_add_double_format()
1996 proto_tree_add_uint_format()
1997 proto_tree_add_uint64_format()
1998 proto_tree_add_int_format()
1999 proto_tree_add_int64_format()
2000 ----------------------------
2001 These routines are used to add items to the protocol tree when the
2002 dissector routines wants complete control over how the field and value
2003 will be represented on the GUI tree. The argument giving the value is
2004 the same as the corresponding proto_tree_add_XXX() function; the rest of
2005 the arguments are a "printf"-style format and any arguments for that
2006 format. The caller must include the name of the field in the format; it
2007 is not added automatically as in the proto_tree_add_XXX() functions.
2009 proto_tree_add_text()
2010 ---------------------
2011 proto_tree_add_text() is used to add a label to the GUI tree. It will
2012 contain no value, so it is not searchable in the display filter process.
2013 This function was needed in the transition from the old-style proto_tree
2014 to this new-style proto_tree so that Ethereal would still decode all
2015 protocols w/o being able to filter on all protocols and fields.
2016 Otherwise we would have had to cripple Ethereal's functionality while we
2017 converted all the old-style proto_tree calls to the new-style proto_tree
2020 This can also be used for items with subtrees, which may not have values
2021 themselves - the items in the subtree are the ones with values.
2023 For a subtree, the label on the subtree might reflect some of the items
2024 in the subtree. This means the label can't be set until at least some
2025 of the items in the subtree have been dissected. To do this, use
2026 'proto_item_set_text()' or 'proto_item_append_text()':
2029 proto_item_set_text(proto_item *ti, ...);
2032 proto_item_append_text(proto_item *ti, ...);
2034 'proto_item_set_text()' takes as an argument the value returned by
2035 'proto_tree_add_text()', a 'printf'-style format string, and a set of
2036 arguments corresponding to '%' format items in that string, and replaces
2037 the text for the item created by 'proto_tree_add_text()' with the result
2038 of applying the arguments to the format string.
2040 'proto_item_append_text()' is similar, but it appends to the text for
2041 the item the result of applying the arguments to the format string.
2043 For example, early in the dissection, one might do:
2045 ti = proto_tree_add_text(tree, tvb, offset, length, <label>);
2049 proto_item_set_text(ti, "%s: %s", type, value);
2051 after the "type" and "value" fields have been extracted and dissected.
2052 <label> would be a label giving what information about the subtree is
2053 available without dissecting any of the data in the subtree.
2055 Note that an exception might thrown when trying to extract the values of
2056 the items used to set the label, if not all the bytes of the item are
2057 available. Thus, one should create the item with text that is as
2058 meaningful as possible, and set it or append additional information to
2059 it as the values needed to supply that information is extracted.
2061 proto_tree_add_text_valist()
2062 ---------------------
2063 This is like proto_tree_add_text(), but takes, as the last argument, a
2064 'va_list'; it is used to allow routines that take a printf-like
2065 variable-length list of arguments to add a text item to the protocol
2068 1.7 Utility routines
2070 1.7.1 match_strval and val_to_str
2072 A dissector may need to convert a value to a string, using a
2073 'value_string' structure, by hand, rather than by declaring a field with
2074 an associated 'value_string' structure; this might be used, for example,
2075 to generate a COL_INFO line for a frame.
2077 'match_strval()' will do that:
2080 match_strval(guint32 val, const value_string *vs)
2082 It will look up the value 'val' in the 'value_string' table pointed to
2083 by 'vs', and return either the corresponding string, or NULL if the
2084 value could not be found in the table. Note that, unless 'val' is
2085 guaranteed to be a value in the 'value_string' table ("guaranteed" as in
2086 "the code has already checked that it's one of those values" or "the
2087 table handles all possible values of the size of 'val'", not "the
2088 protocol spec says it has to be" - protocol specs do not prevent invalid
2089 packets from being put onto a network or into a purported packet capture
2090 file), you must check whether 'match_strval()' returns NULL, and arrange
2091 that its return value not be dereferenced if it's NULL. In particular,
2092 don't use it in a call to generate a COL_INFO line for a frame such as
2094 col_add_fstr(COL_INFO, ", %s", match_strval(val, table));
2096 unless is it certain that 'val' is in 'table'.
2098 'val_to_str()' can be used to generate a string for values not found in
2102 val_to_str(guint32 val, const value_string *vs, const char *fmt)
2104 If the value 'val' is found in the 'value_string' table pointed to by
2105 'vs', 'val_to_str' will return the corresponding string; otherwise, it
2106 will use 'fmt' as an 'sprintf'-style format, with 'val' as an argument,
2107 to generate a string, and will return a pointer to that string.
2108 (Currently, it has three 64-byte static buffers, and cycles through
2109 them; this permits the results of up to three calls to 'val_to_str' to
2110 be passed as arguments to a routine using those strings.)
2113 1.8 Calling Other Dissectors
2115 NOTE: This is discussed in the README.tvbuff file. For more
2116 information on tvbuffers consult that file.
2118 As each dissector completes its portion of the protocol analysis, it
2119 is expected to create a new tvbuff of type TVBUFF_SUBSET which
2120 contains the payload portion of the protocol (that is, the bytes
2121 that are relevant to the next dissector).
2123 The syntax for creating a new TVBUFF_SUBSET is:
2125 next_tvb = tvb_new_subset(tvb, offset, length, reported_length)
2128 tvb is the tvbuff that the dissector has been working on. It
2129 can be a tvbuff of any type.
2131 next_tvb is the new TVBUFF_SUBSET.
2133 offset is the byte offset of 'tvb' at which the new tvbuff
2134 should start. The first byte is the 0th byte.
2136 length is the number of bytes in the new TVBUFF_SUBSET. A length
2137 argument of -1 says to use as many bytes as are available in
2140 reported_length is the number of bytes that the current protocol
2141 says should be in the payload. A reported_length of -1 says that
2142 the protocol doesn't say anything about the size of its payload.
2145 An example from packet-ipx.c -
2148 dissect_ipx(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree)
2151 int reported_length, available_length;
2154 /* Make the next tvbuff */
2156 /* IPX does have a length value in the header, so calculate report_length */
2157 Set this to -1 if there isn't any length information in the protocol
2159 reported_length = ipx_length - IPX_HEADER_LEN;
2161 /* Calculate the available data in the packet,
2162 set this to -1 to use all the data in the tv_buffer
2164 available_length = tvb_length(tvb) - IPX_HEADER_LEN;
2166 /* Create the tvbuffer for the next dissector */
2167 next_tvb = tvb_new_subset(tvb, IPX_HEADER_LEN,
2168 MIN(available_length, reported_length),
2171 /* call the next dissector */
2172 dissector_next( next_tvb, pinfo, tree);
2175 1.9 Editing Makefile.common to add your dissector.
2177 To arrange that your dissector will be built as part of Ethereal, you
2178 must add the name of the source file for your dissector to the
2179 'DISSECTOR_SRC' macro in the 'Makefile.common' file in the 'epan/dissectors'
2180 directory. (Note that this is for modern versions of UNIX, so there
2181 is no 14-character limitation on file names, and for modern versions of
2182 Windows, so there is no 8.3-character limitation on file names.)
2184 If your dissector also has its own header file or files, you must add
2185 them to the 'DISSECTOR_INCLUDES' macro in the 'Makefile.common' file in
2186 the 'epan/dissectors' directory, so that it's included when release source
2187 tarballs are built (otherwise, the source in the release tarballs won't
2190 1.10 Using the SVN source code tree.
2192 See <http://www.ethereal.com/development.html#source>
2194 1.11 Submitting code for your new dissector.
2196 - TEST YOUR DISSECTOR BEFORE SUBMITTING IT.
2197 Use fuzz-test.sh and/or randpkt against your dissector. These are
2198 described at <http://wiki.ethereal.com/FuzzTesting>.
2200 - Subscribe to <mailto:ethereal-dev@ethereal.com> by sending an email to
2201 <mailto:ethereal-dev-request@ethereal.com?body="help"> or visiting
2202 <http://www.ethereal.com/lists/>.
2204 - 'svn add' all the files of your new dissector.
2206 - 'svn diff' the workspace and save the result to a file.
2208 - Send the diff file along with a note requesting it's inclusion to
2209 <mailto:ethereal-dev@ethereal.com>. You can also use this procedure for
2210 providing patches to your dissector or any other part of ethereal.
2212 - If possible, add sample capture files to the sample captures page at
2213 <http://wiki.ethereal.com/SampleCaptures>. These files are used by
2214 the automated build system for fuzz testing.
2216 - If you find that you are contributing a lot to ethereal on an ongoing
2217 basis you can request to become a committer which will allow you to
2218 commit files to subversion directly.
2220 2. Advanced dissector topics.
2224 2.2 Following "conversations".
2226 In ethereal a conversation is defined as a series of data packet between two
2227 address:port combinations. A conversation is not sensitive to the direction of
2228 the packet. The same conversation will be returned for a packet bound from
2229 ServerA:1000 to ClientA:2000 and the packet from ClientA:2000 to ServerA:1000.
2231 There are five routines that you will use to work with a conversation:
2232 conversation_new, find_conversation, conversation_add_proto_data,
2233 conversation_get_proto_data, and conversation_delete_proto_data.
2236 2.2.1 The conversation_init function.
2238 This is an internal routine for the conversation code. As such the you
2239 will not have to call this routine. Just be aware that this routine is
2240 called at the start of each capture and before the packets are filtered
2241 with a display filter. The routine will destroy all stored
2242 conversations. This routine does NOT clean up any data pointers that are
2243 passed in the conversation_new 'data' variable. You are responsible for
2244 this clean up if you pass a malloc'ed pointer in this variable.
2246 See item 2.2.7 for more information about the 'data' pointer.
2249 2.2.2 The conversation_new function.
2251 This routine will create a new conversation based upon two address/port
2252 pairs. If you want to associate with the conversation a pointer to a
2253 private data structure you must use the conversation_add_proto_data
2254 function. The ptype variable is used to differentiate between
2255 conversations over different protocols, i.e. TCP and UDP. The options
2256 variable is used to define a conversation that will accept any destination
2257 address and/or port. Set options = 0 if the destination port and address
2258 are know when conversation_new is called. See section 2.4 for more
2259 information on usage of the options parameter.
2261 The conversation_new prototype:
2262 conversation_t *conversation_new(guint32 setup_frame, address *addr1,
2263 address *addr2, port_type ptype, guint32 port1, guint32 port2,
2267 guint32 setup_frame = The lowest numbered frame for this conversation
2268 address* addr1 = first data packet address
2269 address* addr2 = second data packet address
2270 port_type ptype = port type, this is defined in packet.h
2271 guint32 port1 = first data packet port
2272 guint32 port2 = second data packet port
2273 guint options = conversation options, NO_ADDR2 and/or NO_PORT2
2275 setup_frame indicates the first frame for this conversation, and is used to
2276 distinguish multiple conversations with the same addr1/port1 and addr2/port2
2277 pair that occur within the same capture session.
2279 "addr1" and "port1" are the first address/port pair; "addr2" and "port2"
2280 are the second address/port pair. A conversation doesn't have source
2281 and destination address/port pairs - packets in a conversation go in
2282 both directions - so "addr1"/"port1" may be the source or destination
2283 address/port pair; "addr2"/"port2" would be the other pair.
2285 If NO_ADDR2 is specified, the conversation is set up so that a
2286 conversation lookup will match only the "addr1" address; if NO_PORT2 is
2287 specified, the conversation is set up so that a conversation lookup will
2288 match only the "port1" port; if both are specified, i.e.
2289 NO_ADDR2|NO_PORT2, the conversation is set up so that the lookup will
2290 match only the "addr1"/"port1" address/port pair. This can be used if a
2291 packet indicates that, later in the capture, a conversation will be
2292 created using certain addresses and ports, in the case where the packet
2293 doesn't specify the addresses and ports of both sides.
2295 2.2.3 The find_conversation function.
2297 Call this routine to look up a conversation. If no conversation is found,
2298 the routine will return a NULL value.
2300 The find_conversation prototype:
2302 conversation_t *find_conversation(guint32 frame_num, address *addr_a,
2303 address *addr_b, port_type ptype, guint32 port_a, guint32 port_b,
2307 guint32 frame_num = a frame number to match
2308 address* addr_a = first address
2309 address* addr_b = second address
2310 port_type ptype = port type
2311 guint32 port_a = first data packet port
2312 guint32 port_b = second data packet port
2313 guint options = conversation options, NO_ADDR_B and/or NO_PORT_B
2315 frame_num is a frame number to match. The conversation returned is where
2316 (frame_num >= conversation->setup_frame
2317 && frame_num < conversation->next->setup_frame)
2318 Suppose there are a total of 3 conversations (A, B, and C) that match
2319 addr_a/port_a and addr_b/port_b, where the setup_frame used in
2320 conversation_new() for A, B and C are 10, 50, and 100 respectively. The
2321 frame_num passed in find_conversation is compared to the setup_frame of each
2322 conversation. So if (frame_num >= 10 && frame_num < 50), conversation A is
2323 returned. If (frame_num >= 50 && frame_num < 100), conversation B is returned.
2324 If (frame_num >= 100) conversation C is returned.
2326 "addr_a" and "port_a" are the first address/port pair; "addr_b" and
2327 "port_b" are the second address/port pair. Again, as a conversation
2328 doesn't have source and destination address/port pairs, so
2329 "addr_a"/"port_a" may be the source or destination address/port pair;
2330 "addr_b"/"port_b" would be the other pair. The search will match the
2331 "a" address/port pair against both the "1" and "2" address/port pairs,
2332 and match the "b" address/port pair against both the "2" and "1"
2333 address/port pairs; you don't have to worry about which side the "a" or
2334 "b" pairs correspond to.
2336 If the NO_ADDR_B flag was specified to "find_conversation()", the
2337 "addr_b" address will be treated as matching any "wildcarded" address;
2338 if the NO_PORT_B flag was specified, the "port_b" port will be treated
2339 as matching any "wildcarded" port. If both flags are specified, i.e.
2340 NO_ADDR_B|NO_PORT_B, the "addr_b" address will be treated as matching
2341 any "wildcarded" address and the "port_b" port will be treated as
2342 matching any "wildcarded" port.
2345 2.2.4 The conversation_add_proto_data function.
2347 Once you have created a conversation with conversation_new, you can
2348 associate data with it using this function.
2350 The conversation_add_proto_data prototype:
2352 void conversation_add_proto_data(conversation_t *conv, int proto,
2356 conversation_t *conv = the conversation in question
2357 int proto = registered protocol number
2358 void *data = dissector data structure
2360 "conversation" is the value returned by conversation_new. "proto" is a
2361 unique protocol number created with proto_register_protocol. Protocols
2362 are typically registered in the proto_register_XXXX section of your
2363 dissector. "data" is a pointer to the data you wish to associate with the
2364 conversation. Using the protocol number allows several dissectors to
2365 associate data with a given conversation.
2368 2.2.5 The conversation_get_proto_data function.
2370 After you have located a conversation with find_conversation, you can use
2371 this function to retrieve any data associated with it.
2373 The conversation_get_proto_data prototype:
2375 void *conversation_get_proto_data(conversation_t *conv, int proto);
2378 conversation_t *conv = the conversation in question
2379 int proto = registered protocol number
2381 "conversation" is the conversation created with conversation_new. "proto"
2382 is a unique protocol number acreated with proto_register_protocol,
2383 typically in the proto_register_XXXX portion of a dissector. The function
2384 returns a pointer to the data requested, or NULL if no data was found.
2387 2.2.6 The conversation_delete_proto_data function.
2389 After you are finished with a conversation, you can remove your assocation
2390 with this function. Please note that ONLY the conversation entry is
2391 removed. If you have allocated any memory for your data, you must free it
2394 The conversation_delete_proto_data prototype:
2396 void conversation_delete_proto_data(conversation_t *conv, int proto);
2399 conversation_t *conv = the conversation in question
2400 int proto = registered protocol number
2402 "conversation" is the conversation created with conversation_new. "proto"
2403 is a unique protocol number acreated with proto_register_protocol,
2404 typically in the proto_register_XXXX portion of a dissector.
2406 2.2.7 The example conversation code with GMemChunk's
2408 For a conversation between two IP addresses and ports you can use this as an
2409 example. This example uses the GMemChunk to allocate memory and stores the data
2410 pointer in the conversation 'data' variable.
2412 NOTE: Remember to register the init routine (my_dissector_init) in the
2413 protocol_register routine.
2416 /************************ Globals values ************************/
2418 /* the number of entries in the memory chunk array */
2419 #define my_init_count 10
2421 /* define your structure here */
2426 /* the GMemChunk base structure */
2427 static GMemChunk *my_vals = NULL;
2429 /* Registered protocol number
2430 static int my_proto = -1;
2433 /********************* in the dissector routine *********************/
2435 /* the local variables in the dissector */
2437 conversation_t *conversation;
2438 my_entry_t *data_ptr
2441 /* look up the conversation */
2443 conversation = find_conversation(pinfo->fd->num, &pinfo->src, &pinfo->dst,
2444 pinfo->ptype, pinfo->srcport, pinfo->destport, 0);
2446 /* if conversation found get the data pointer that you stored */
2448 data_ptr = (my_entry_t*)conversation_get_proto_data(conversation,
2452 /* new conversation create local data structure */
2454 data_ptr = g_mem_chunk_alloc(my_vals);
2456 /*** add your code here to setup the new data structure ***/
2458 /* create the conversation with your data pointer */
2460 conversation_new(pinfo->fd->num, &pinfo->src, &pinfo->dst, pinfo->ptype,
2461 pinfo->srcport, pinfo->destport, 0);
2462 conversation_add_proto_data(conversation, my_proto, (void *) data_ptr);
2465 /* at this point the conversation data is ready */
2468 /******************* in the dissector init routine *******************/
2470 #define my_init_count 20
2473 my_dissector_init( void){
2475 /* destroy memory chunks if needed */
2478 g_mem_chunk_destroy(my_vals);
2480 /* now create memory chunks */
2482 my_vals = g_mem_chunk_new( "my_proto_vals",
2484 my_init_count * sizeof( my_entry_t),
2488 /***************** in the protocol register routine *****************/
2490 /* register re-init routine */
2492 register_init_routine( &my_dissector_init);
2494 my_proto = proto_register_protocol("My Protocol", "My Protocol", "my_proto");
2497 2.2.8 An example conversation code that starts at a specific frame number
2499 Sometimes a disector has determined that a new conversation is needed that
2500 starts at a specific frame number, when a capture session encompasses multiple
2501 conversation that reuse the same src/dest ip/port pairs. You can use the
2502 compare the conversation->setup_frame returned by find_conversation with
2503 pinfo->fd->num to determine whether or not there already exists a conversation
2504 that starts at the specific frame number.
2506 /* in the dissector routine */
2508 conversation = find_conversation(pinfo->fd->num, &pinfo->src, &pinfo->dst,
2509 pinfo->ptype, pinfo->srcport, pinfo->destport, 0);
2510 if (conversation == NULL || (conversation->setup_frame != pinfo->fd->num)) {
2511 /* It's not part of any conversation or the returned
2512 * conversation->setup_frame doesn't match the current frame
2515 conversation = conversation_new(pinfo->fd->num, &pinfo->src,
2516 &pinfo->dst, pinfo->ptype, pinfo->srcport, pinfo->destport,
2521 2.2.9 The example conversation code using conversation index field
2523 Sometimes the conversation isn't enough to define a unique data storage
2524 value for the network traffic. For example if you are storing information
2525 about requests carried in a conversation, the request may have an
2526 identifier that is used to define the request. In this case the
2527 conversation and the identifier are required to find the data storage
2528 pointer. You can use the conversation data structure index value to
2529 uniquely define the conversation.
2531 See packet-afs.c for an example of how to use the conversation index. In
2532 this dissector multiple requests are sent in the same conversation. To store
2533 information for each request the dissector has an internal hash table based
2534 upon the conversation index and values inside the request packets.
2537 /* in the dissector routine */
2539 /* to find a request value, first lookup conversation to get index */
2540 /* then used the conversation index, and request data to find data */
2541 /* in the local hash table */
2543 conversation = find_conversation(pinfo->fd->num, &pinfo->src, &pinfo->dst,
2544 pinfo->ptype, pinfo->srcport, pinfo->destport, 0);
2545 if (conversation == NULL) {
2546 /* It's not part of any conversation - create a new one. */
2547 conversation = conversation_new(pinfo->fd->num, &pinfo->src,
2548 &pinfo->dst, pinfo->ptype, pinfo->srcport, pinfo->destport,
2552 request_key.conversation = conversation->index;
2553 request_key.service = pntohs(&rxh->serviceId);
2554 request_key.callnumber = pntohl(&rxh->callNumber);
2556 request_val = (struct afs_request_val *) g_hash_table_lookup(
2557 afs_request_hash, &request_key);
2559 /* only allocate a new hash element when it's a request */
2561 if ( !request_val && !reply)
2563 new_request_key = g_mem_chunk_alloc(afs_request_keys);
2564 *new_request_key = request_key;
2566 request_val = g_mem_chunk_alloc(afs_request_vals);
2567 request_val -> opcode = pntohl(&afsh->opcode);
2568 opcode = request_val->opcode;
2570 g_hash_table_insert(afs_request_hash, new_request_key,
2576 2.3 Dynamic conversation dissector registration
2579 NOTE: This sections assumes that all information is available to
2580 create a complete conversation, source port/address and
2581 destination port/address. If either the destination port or
2582 address is know, see section 2.4 Dynamic server port dissector
2585 For protocols that negotiate a secondary port connection, for example
2586 packet-msproxy.c, a conversation can install a dissector to handle
2587 the secondary protocol dissection. After the conversation is created
2588 for the negotiated ports use the conversation_set_dissector to define
2589 the dissection routine.
2590 Before we create these conversations or assign a dissector to them we should
2591 first check that the conversation does not already exist and if it exists
2592 whether it is registered to our protocol or not.
2593 We should do this because is uncommon but it does happen that multiple
2594 different protocols can use the same socketpair during different stages of
2595 an application cycle. By keeping track of the frame number a conversation
2596 was started in ethereal can still tell these different protocols apart.
2598 The second argument to conversation_set_dissector is a dissector handle,
2599 which is created with a call to create_dissector_handle or
2602 create_dissector_handle takes as arguments a pointer to the dissector
2603 function and a protocol ID as returned by proto_register_protocol;
2604 register_dissector takes as arguments a string giving a name for the
2605 dissector, a pointer to the dissector function, and a protocol ID.
2607 The protocol ID is the ID for the protocol dissected by the function.
2608 The function will not be called if the protocol has been disabled by the
2609 user; instead, the data for the protocol will be dissected as raw data.
2613 /* the handle for the dynamic dissector *
2614 static dissector_handle_t sub_dissector_handle;
2616 /* prototype for the dynamic dissector */
2617 static void sub_dissector( tvbuff_t *tvb, packet_info *pinfo,
2620 /* in the main protocol dissector, where the next dissector is setup */
2622 /* if conversation has a data field, create it and load structure */
2624 /* First check if a conversation already exists for this
2627 conversation = find_conversation(pinfo->fd->num,
2628 &pinfo->src, &pinfo->dst, protocol,
2629 src_port, dst_port, new_conv_info, 0);
2631 /* If there is no such conversation, or if there is one but for
2632 someone elses protocol then we just create a new conversation
2633 and assign our protocol to it.
2635 if( (conversation==NULL)
2636 || (conversation->dissector_handle!=sub_dissector_handle) ){
2637 new_conv_info = g_mem_chunk_alloc( new_conv_vals);
2638 new_conv_info->data1 = value1;
2640 /* create the conversation for the dynamic port */
2641 conversation = conversation_new(pinfo->fd->num,
2642 &pinfo->src, &pinfo->dst, protocol,
2643 src_port, dst_port, new_conv_info, 0);
2645 /* set the dissector for the new conversation */
2646 conversation_set_dissector(conversation, sub_dissector_handle);
2651 proto_register_PROTOABBREV(void)
2655 sub_dissector_handle = create_dissector_handle(sub_dissector,
2661 2.4 Dynamic server port dissector registration
2663 NOTE: While this example used both NO_ADDR2 and NO_PORT2 to create a
2664 conversation with only one port and address set, this isn't a
2665 requirement. Either the second port or the second address can be set
2666 when the conversation is created.
2668 For protocols that define a server address and port for a secondary
2669 protocol, a conversation can be used to link a protocol dissector to
2670 the server port and address. The key is to create the new
2671 conversation with the second address and port set to the "accept
2674 Some server applications can use the same port for different protocols during
2675 different stages of a transaction. For example it might initially use SNMP
2676 to perform some discovery and later switch to use TFTP using the same port.
2677 In order to handle this properly we must first check whether such a
2678 conversation already exists or not and if it exists we also check whether the
2679 registered dissector_handle for that conversation is "our" dissector or not.
2680 If not we create a new conversation ontop of the previous one and set this new
2681 conversation to use our protocol.
2682 Since ethereal keeps track of the frame number where a conversation started
2683 ethereal will still be able to keep the packets apart eventhough they do use
2684 the same socketpair.
2685 (See packet-tftp.c and packet-snmp.c for examples of this)
2687 There are two support routines that will allow the second port and/or
2688 address to be set latter.
2690 conversation_set_port2( conversation_t *conv, guint32 port);
2691 conversation_set_addr2( conversation_t *conv, address addr);
2693 These routines will change the second address or port for the
2694 conversation. So, the server port conversation will be converted into a
2695 more complete conversation definition. Don't use these routines if you
2696 want create a conversation between the server and client and retain the
2697 server port definition, you must create a new conversation.
2702 /* the handle for the dynamic dissector *
2703 static dissector_handle_t sub_dissector_handle;
2707 /* in the main protocol dissector, where the next dissector is setup */
2709 /* if conversation has a data field, create it and load structure */
2711 new_conv_info = g_mem_chunk_alloc( new_conv_vals);
2712 new_conv_info->data1 = value1;
2714 /* create the conversation for the dynamic server address and port */
2715 /* NOTE: The second address and port values don't matter because the */
2716 /* NO_ADDR2 and NO_PORT2 options are set. */
2718 /* First check if a conversation already exists for this
2721 conversation = find_conversation(pinfo->fd->num,
2722 &server_src_addr, 0, protocol,
2723 server_src_port, 0, NO_ADDR2 | NO_PORT_B);
2724 /* If there is no such conversation, or if there is one but for
2725 someone elses protocol then we just create a new conversation
2726 and assign our protocol to it.
2728 if( (conversation==NULL)
2729 || (conversation->dissector_handle!=sub_dissector_handle) ){
2730 conversation = conversation_new(pinfo->fd->num,
2731 &server_src_addr, 0, protocol,
2732 server_src_port, 0, new_conv_info, NO_ADDR2 | NO_PORT2);
2734 /* set the dissector for the new conversation */
2735 conversation_set_dissector(conversation, sub_dissector_handle);
2738 2.5 Per packet information
2740 Information can be stored for each data packet that is processed by the dissector.
2741 The information is added with the p_add_proto_data function and retreived with the
2742 p_get_proto_data function. The data pointers passed into the p_add_proto_data are
2743 not managed by the proto_data routines. If you use malloc or any other dynamic
2744 memory allocation scheme, you must release the data when it isn't required.
2747 p_add_proto_data(frame_data *fd, int proto, void *proto_data)
2749 p_get_proto_data(frame_data *fd, int proto)
2752 fd - The fd pointer in the pinfo structure, pinfo->fd
2753 proto - Protocol id returned by the proto_register_protocol call during initialization
2754 proto_data - pointer to the dissector data.
2757 2.6 User Preferences
2759 If the dissector has user options, there is support for adding these preferences
2760 to a configuration dialog.
2762 You must register the module with the preferences routine with -
2764 module_t *prefs_register_protocol(proto_id, void (*apply_cb)(void))
2766 Where: proto_id - the value returned by "proto_register_protocol()" when
2767 the protocol was registered
2768 apply_cb - Callback routine that is call when preferences are applied
2771 Then you can register the fields that can be configured by the user with these routines -
2773 /* Register a preference with an unsigned integral value. */
2774 void prefs_register_uint_preference(module_t *module, const char *name,
2775 const char *title, const char *description, guint base, guint *var);
2777 /* Register a preference with an Boolean value. */
2778 void prefs_register_bool_preference(module_t *module, const char *name,
2779 const char *title, const char *description, gboolean *var);
2781 /* Register a preference with an enumerated value. */
2782 void prefs_register_enum_preference(module_t *module, const char *name,
2783 const char *title, const char *description, gint *var,
2784 const enum_val_t *enumvals, gboolean radio_buttons)
2786 /* Register a preference with a character-string value. */
2787 void prefs_register_string_preference(module_t *module, const char *name,
2788 const char *title, const char *description, char **var)
2790 /* Register a preference with a range of unsigned integers (e.g.,
2793 void prefs_register_range_preference(module_t *module, const char *name,
2794 const char *title, const char *description, range_t *var,
2797 Where: module - Returned by the prefs_register_protocol routine
2798 name - This is appended to the name of the protocol, with a
2799 "." between them, to construct a name that identifies
2800 the field in the preference file; the name itself
2801 should not include the protocol name, as the name in
2802 the preference file will already have it
2803 title - Field title in the preferences dialog
2804 description - Comments added to the preference file above the
2806 var - pointer to the storage location that is updated when the
2807 field is changed in the preference dialog box
2808 enumvals - an array of enum_val_t structures. This must be
2809 NULL-terminated; the members of that structure are:
2811 a short name, to be used with the "-o" flag - it
2812 should not contain spaces or upper-case letters,
2813 so that it's easier to put in a command line;
2815 a description, which is used in the GUI (and
2816 which, for compatibility reasons, is currently
2817 what's written to the preferences file) - it can
2818 contain spaces, capital letters, punctuation,
2821 the numerical value corresponding to that name
2823 radio_buttons - TRUE if the field is to be displayed in the
2824 preferences dialog as a set of radio buttons,
2825 FALSE if it is to be displayed as an option
2827 max_value - The maximum allowed value for a range (0 is the minimum).
2829 An example from packet-beep.c -
2831 proto_beep = proto_register_protocol("Blocks Extensible Exchange Protocol",
2836 /* Register our configuration options for BEEP, particularly our port */
2838 beep_module = prefs_register_protocol(proto_beep, proto_reg_handoff_beep);
2840 prefs_register_uint_preference(beep_module, "tcp.port", "BEEP TCP Port",
2841 "Set the port for BEEP messages (if other"
2842 " than the default of 10288)",
2843 10, &global_beep_tcp_port);
2845 prefs_register_bool_preference(beep_module, "strict_header_terminator",
2846 "BEEP Header Requires CRLF",
2847 "Specifies that BEEP requires CRLF as a "
2848 "terminator, and not just CR or LF",
2849 &global_beep_strict_term);
2851 This will create preferences "beep.tcp.port" and
2852 "beep.strict_header_terminator", the first of which is an unsigned
2853 integer and the second of which is a Boolean.
2855 2.7 Reassembly/desegmentation for protocols running atop TCP
2857 There are two main ways of reassembling a Protocol Data Unit (PDU) which
2858 spans across multiple TCP segments. The first approach is simpler, but
2859 assumes you are running atop of TCP when this occurs (but your dissector
2860 might run atop of UDP, too, for example), and that your PDUs consist of a
2861 fixed amount of data that includes enough information to determine the PDU
2862 length, possibly followed by additional data. The second method is more
2863 generic but requires more code and is less efficient.
2865 2.7.1 Using tcp_dissect_pdus()
2867 For the first method, you register two different dissection methods, one
2868 for the TCP case, and one for the other cases. It is a good idea to
2869 also have a dissect_PROTO_common function which will parse the generic
2870 content that you can find in all PDUs which is called from
2871 dissect_PROTO_tcp when the reassembly is complete and from
2872 dissect_PROTO_udp (or dissect_PROTO_other).
2874 To register the distinct dissector functions, consider the following
2875 example, stolen from packet-dns.c:
2877 dissector_handle_t dns_udp_handle;
2878 dissector_handle_t dns_tcp_handle;
2879 dissector_handle_t mdns_udp_handle;
2881 dns_udp_handle = create_dissector_handle(dissect_dns_udp,
2883 dns_tcp_handle = create_dissector_handle(dissect_dns_tcp,
2885 mdns_udp_handle = create_dissector_handle(dissect_mdns_udp,
2888 dissector_add("udp.port", UDP_PORT_DNS, dns_udp_handle);
2889 dissector_add("tcp.port", TCP_PORT_DNS, dns_tcp_handle);
2890 dissector_add("udp.port", UDP_PORT_MDNS, mdns_udp_handle);
2891 dissector_add("tcp.port", TCP_PORT_MDNS, dns_tcp_handle);
2893 The dissect_dns_udp function does very little work and calls
2894 dissect_dns_common, while dissect_dns_tcp calls tcp_dissect_pdus with a
2895 reference to a callback which will be called with reassembled data:
2898 dissect_dns_tcp(tvbuff_t *tvb, packet_info *pinfo, proto_tree *tree)
2900 tcp_dissect_pdus(tvb, pinfo, tree, dns_desegment, 2,
2901 get_dns_pdu_len, dissect_dns_tcp_pdu);
2904 (The dissect_dns_tcp_pdu function acts similarly to dissect_dns_udp.)
2905 The arguments to tcp_dissect_pdus are:
2907 the tvbuff pointer, packet_info pointer, and proto_tree pointer
2908 passed to the dissector;
2910 a gboolean flag indicating whether desegmentation is enabled for
2913 the number of bytes of PDU data required to determine the length
2916 a routine that takes as arguments a tvbuff pointer and an offset
2917 value representing the offset into the tvbuff at which a PDU
2918 begins and should return - *without* throwing an exception (it
2919 is guaranteed that the number of bytes specified by the previous
2920 argument to tcp_dissect_pdus is available, but more data might
2921 not be available, so don't refer to any data past that) - the
2922 total length of the PDU, in bytes;
2924 a routine that's passed a tvbuff pointer, packet_info pointer,
2925 and proto_tree pointer, with the tvbuff containing a
2926 possibly-reassembled PDU, and that should dissect that PDU.
2928 2.7.2 Modifying the pinfo struct
2930 The second reassembly mode is prefered when the dissector cannot determine
2931 how many bytes it will need to read in order to determine the size of a PDU.
2932 For this mode it is reccommended that your dissector be the newer dissector
2933 type which returns "int" rather than the older type which returned "void".
2935 This reassembly mode relies on Ethereal's mechanism for processing multiple PDUs
2936 per frame. When a dissector processes a PDU from a tvbuff the PDU may not be
2937 aligned to a frame of the underlying protocol. Ethereal allows dissectors to
2938 process PDUs in an idempotent way--dissectors only need to consider one PDU at a
2939 time. If your dissector discovers that it can not process a complete PDU from
2940 the current tvbuff the dissector should halt processing and request additional
2941 bytes from the lower level dissector.
2943 Your dissect_PROTO will be called by the lower level dissector whenever
2944 sufficient new bytes become available. Each time your dissector is called it is
2945 provided a different tvbuff, though the tvbuffs may contain data that your
2946 dissector declined to process during a previous call. When called a dissector
2947 should examine the tvbuff provided and determine if an entire PDU is available.
2948 If sufficient bytes are available the dissector processes the PDU and returns
2949 the length of the PDU from your dissect_PROTO.
2951 Completion of a PDU is signified by dissect_PROTO returning a positive value.
2952 The value is the number of bytes which were processed from the tvbuff. If there
2953 were insufficient bytes in the tvbuff to complete a PDU then the dissect_PROTO
2954 returns a negative value requesting additional bytes. The negative return value
2955 indicates how many additional bytes are required. Additionally dissect_PROTO
2956 must update the pinfo structure to indicate that more bytes are required. The
2957 desegment_offset field is the offset in the tvbuff at which the dissector will
2958 continue processing when next called. The desegment_len field should contain the
2959 estimated number of additional bytes required for completing the PDU. The
2960 dissect_PROTO will not be called again until the specified number of bytes are
2961 available. pinfo->desegment_len may be set to -1 if dissect_PROTO cannot
2962 determine how many additional bytes are required. Dissectors should set the
2963 desegment_len to a reasonable value when possible rather than always setting
2964 -1 as it will generally be more efficient.
2966 static hf_register_info hf[] = {
2968 {"C String", "c.string", FT_STRING, BASE_NONE, NULL, 0x0,
2974 * Dissect a buffer containing a C string.
2976 * @param tvb The buffer to dissect.
2977 * @param pinfo Packet Info.
2978 * @param tree The protocol tree.
2979 * @return Number of bytes from the tvbuff_t which were processed or a negative
2980 * value indicating more bytes are needed.
2982 static int dissect_cstr(tvbuff_t * tvb, packet_info * pinfo, proto_tree * tree)
2985 gint available = tvb_reported_length_remaining(tvb, offset);
2986 gint len = tvb_strnlen( tvb, offset, available );
2989 /* No '\0' found, ask for another byte. */
2990 pinfo->desegment_offset = offset;
2991 pinfo->desegment_len = 1;
2995 if (check_col(pinfo->cinfo, COL_INFO)) {
2996 col_set_str(pinfo->cinfo, COL_INFO, "C String");
2999 len += 1; /* Add one for the '\0' */
3002 proto_tree_add_item(tree, hf_cstring, tvb, offset, len, FALSE);
3008 This simple dissector will repeatedly return -1 requesting one more byte until
3009 the tvbuff contains a complete C string. The C string will then be added to the
3010 protocol tree. Unfortunately since there is no way to guess the size of C String
3011 without seeing the entire string this dissector can never request more than one
3016 See the README.plugins for more information on how to "pluginize"
3019 4.0 Extending Wiretap.
3021 5.0 How the Display Filter Engine works
3024 epan/dfilter/* - the display filter engine, including
3025 scanner, parser, syntax-tree semantics checker, DFVM bytecode
3026 generator, and DFVM engine.
3027 epan/ftypes/* - the definitions of the various FT_* field types.
3028 epan/proto.c - proto_tree-related routines
3032 The scanner/parser pair read the string representing the display filter
3033 and convert it into a very simple syntax tree. The syntax tree is very
3034 simple in that it is possible that many of the nodes contain unparsed
3035 chunks of text from the display filter.
3037 5.1 Enhancing the syntax tree.
3039 The semantics of the simple syntax tree are checked to make sure that
3040 the fields that are being compared are being compared to appropriate
3041 values. For example, if a field is an integer, it can't be compared to
3042 a string, unless a value_string has been defined for that field.
3044 During the process of checking the semantics, the simple syntax tree is
3045 fleshed out and no longer contains nodes with unparsed information. The
3046 syntax tree is no longer in its simple form, but in its complete form.
3048 5.2 Converting to DFVM bytecode
3050 The syntax tree is analyzed to create a sequence of bytecodes in the
3051 "DFVM" language. "DFVM" stands for Display Filter Virtual Machine. The
3052 DFVM is similar in spirit, but not in definition, to the BPF VM that
3053 libpcap uses to analyze packets.
3055 A virtual bytecode is created and used so that the actual process of
3056 filtering packets will be fast. That is, it should be faster to process
3057 a list of VM bytecodes than to attempt to filter packets directly from
3058 the syntax tree. (heh... no measurement has been made to support this
3063 Once the DFVM bytecode has been produced, it's a simple matter of
3064 running the DFVM engine against the proto_tree from the packet
3065 dissection, using the DFVM bytecodes as instructions. If the DFVM
3066 bytecode is known before packet dissection occurs, the
3067 proto_tree-related code can be "primed" to store away pointers to
3068 field_info structures that are interesting to the display filter. This
3069 makes lookup of those field_info structures during the filtering process
3073 6.0 Adding new capabilities.
3078 James Coe <jammer@cin.net>
3079 Gilbert Ramirez <gram@alumni.rice.edu>
3080 Jeff Foster <jfoste@woodward.com>
3081 Olivier Abad <oabad@cybercable.fr>
3082 Laurent Deniel <laurent.deniel@free.fr>
3083 Gerald Combs <gerald@ethereal.com>
3084 Guy Harris <guy@alum.mit.edu>