1: ===============================================================================
2: $Id: CHANGELOG,v 1.1.1.2 2013/07/22 11:54:41 misho Exp $
3: LIBNET 1.1 (c) 1998 - 2004 Mike D. Schiffman <mike@infonexus.com>
4: http://www.packetfactory.net/libnet
5: ===============================================================================
6:
7: - 1.1.6
8:
9: - Note about why including libnet.h breaks dnet.h/dumbnet.h (Sam Roberts)
10: - Prepare for 1.1.6-rc3. (Sam Roberts)
11: - Only use getifaddrs() if it exists. (reported by Dagobert Michelsen) (Sam Roberts)
12: - Clean up use of single and bracket quotes. (Sam Roberts)
13: - FreeBSD and Solaris volunteers to contibute. (Sam Roberts)
14: - ICMPv6 struct is too long, so use length macro (sickmind@lavabit.com) (Sam Roberts)
15: - ICMPv6 pblock sizes are needed to calculate IPv6's ip_len (reported by sickmind@lavabit.com) (Sam Roberts)
16: - Prep for 1.1.6 release, such as bumping version (Sam Roberts)
17: - Add people who have volunteered to check release candidates. (Sam Roberts)
18: - Rework libnet_autobuild_ipv6() to eliminate code duplication. (Sam Roberts)
19: - Implemented libnet_autobuild_ipv6() (repolho)
20: - Implemented unix version of libnet_get_ipaddr6() (repolho)
21: - Reintroduce libnet_pblock_record_ip_offset() which is empty. (Sam Roberts)
22: - Make clear that all contributions are under libnet copyright. (Sam Roberts)
23: - Fix doc comment format errors reported by doxygen. (Sam Roberts)
24: - Update for doxygen 1.7.4. (Sam Roberts)
25: - Note that PORTED is no longer maintained. (Sam Roberts)
26: - Note that CONTRIB is no longer maintained. (Sam Roberts)
27: - Prep for upcoming 1.1.6 release. (Sam Roberts)
28: - Remove BUGS, it referred to non-existent code. (Sam Roberts)
29: - Reworked icmpv6 patch to parallel the form of icmpv4 support. (Sam Roberts)
30: - Don't depend on netinet/ip.h. (Sam Roberts)
31: - Remove unused variable. (Sam Roberts)
32: - Support building ICMPv6 packets. (someone)
33: - Fixed typo in error message. (Thomas Habets)
34: - Use SO_BINDTODEVICE to force packets out opened device. (someone)
35: - Always use an IPPROTO of TCP when calculating TCP checksums. (Sam Roberts)
36: - Use correct addr type for addrlen calculation. (someone)
37: - libnet_build_igmp reserved field was mistakenly called 'code' (Sam Roberts)
38: - Documentation doesn't include any gif files. (Sam Roberts)
39: - Enable IPV6 support on Solaris 11. (Rich Burridge)
40: - Trying to fix write errors (Víctor Martínez)
41: - Presence of linux's PF_PACKET sockets is now detected. The acinclude.m4 merged in from packetfactory's 1.1.3-rc branch mysteriously assumed that that there was no PF_PACKET if the target OS was linux, which is the opposite of what we want. (Sam Roberts)
42: - Fix libnet_build_igmp() to not reverse the order of the ip address. libnet APIs that take IP addresses as a uint32_t expect them to already be in network byte order. (Sam Roberts)
43: - pblock_append deals with raw memory and structs, so declare it correctly (Sam Roberts)
44: - Clarified types and sizes of DHCP/BOOTP chaddr, sname, and file. chaddr is a hardware address, with size specified seperately, whereas sname and file are null terminated strings. (Sam Roberts)
45: - Explicitly ignore return value of write (some systems now warn about this). (Sam Roberts)
46: - Synchronize comment about h_len with parameter name in function. (Sam Roberts)
47: - Fixes a buffer overflow issue when copying chaddr, file, and sname fields to the DHCP header. (allfro)
48: - Fixes improper calculation of header size when libnet_pblock_probe is called. payload_s must be added to the header length in order to accommodate for the existence of a non-NULL payload. Otherwise the user is prompted with a 'memcpy would cause overflow' error and the program exits. (allfro)
49: - Fixes incorrect memory block size set in the timeexceed and redirect builders. The n variable does not add the size of the payload (payload_s) for proper allocation of the buffer when payload is not NULL and payload_s is greater than 0. This results in a memcpy buffer overflow error when libnet_pblock_append is called exiting the program. (allfro)
50: - Fixes a bug that incorrectly converts the addr, mask, and next_hop fields to network byte order. Users will usually call libnet_name2addr4 to fill these fields and this function already provides a network byte-ordered value. (allfro)
51: - snap parameter was getting copied into the dhost field. (Sam Roberts)
52: - h_len is no longer used, so pass zero. Coverity noticed that stack garbage was being passed instead of a valid value, its just that the value isn't used, and incluing l->total_size is wrong when the pblock is being updated (though it will work on pblock creation). (Sam Roberts)
53: - Length n should include the value_hdr. (Sam Roberts)
54: - Coverity: UNINIT (Jiri Popelka)
55: - Coverity: REVERSE_INULL (Jiri Popelka)
56: - Coverity: RESOURCE_LEAK (Jiri Popelka)
57: - Coverity: OVERRUN_STATIC (Jiri Popelka)
58: - Coverity: OVERRUN_STATIC (Jiri Popelka)
59: - Coverity: OVERRUN_STATIC (Jiri Popelka)
60: - Coverity: FORWARD_NULL (Jiri Popelka)
61: - Coverity: FORWARD_NULL (Jiri Popelka)
62: - Coverity: CHECKED_RETURN (Jiri Popelka)
63: - mkinstalldirs is replaced by autogen.sh. Maybe it shouldn't be checked in? (Sam Roberts)
64: - autogen retries if /sw/... doesn't exist (Sam Roberts)
65: - build_ipv6: set higher traffic class bits (Ulrich Weber)
66: - Fix missing uint instead of u_int (Dagobert Michelsen)
67:
68: - 1.1.5
69:
70: - IRIX: Get MAC address from ioctl(), not by spawning shell. (Thomas Habets)
71: - Compile fix for IRIX (added includes) (Thomas Habets)
72: - Don't explicitly check for UID 0, we may have capabilities even if not root. (Thomas Habets) (Sam Roberts)
73: - Cleaned up implementations of libnet_get_hwaddr(), some leaked memory, one returned a pointer to data on the stack, and the others return a pointer to static data. I'm settling on the non-reentrant static data form. (Sam Roberts)
74: - Visual C++ compiler(v9.0) uses bitfield type as a hint to pad the bitfield, so struct was too long. (Sam Roberts)
75: - Further simplify autogen.sh (Sam Roberts)
76: - Removed dependency on net/bpf.h, and on pcap.h. (Sam Roberts)
77: - LBL_ALIGN check is unused, removing. (Sam Roberts)
78: - Don't include pcap.h if we've already got net/bpf.h, pcap has it's own bpf. (Sam Roberts)
79: - Make libnet_get_hwaddr work in the last few releases of OpenBSD (stu@spacehopper.org) (Sam Roberts)
80: - Get DLT types from the source, <pcap.h>. (Sam Roberts)
81: - Use autoconf -ivf in autogen.sh (suggested by alon.barlev@gmail.com) (Sam Roberts)
82: - Don't distribute libnet.h, it is generated by configure (alon.barlev@gmail.com) (Sam Roberts)
83: - AIX build failures fixed, cause was inclusion of system headers libnet no longer uses (alon.barlev@gmail.com) (Sam Roberts)
84: - Add a link to the old docs. (Sam Roberts)
85: - Added links to github and sourceforge. (Sam Roberts)
86: - Replace C99/C++ comments with traditional C comments. (alon.barlev@gmail.com) (Sam Roberts)
87: - checksum would segfault if a IP checksum was requested for a non-ipv4 header (Sam Roberts)
88: - Closer backwards compat, assume its ipv4 if it's not ipv6. This seems totally wrong, but so it goes. (Sam Roberts)
89: - Try using the ip_len header field to guess the input buffer's size. (Sam Roberts)
90: - libnet_do_checksum(), despite being "internal", is used by external code. libnet needs to maintain backwards API compatibility, tcpsic from the isic package is an example of a binary failing when calling the new API with the old arguments. (Sam Roberts)
91: - merged icmpv6 patch in, but I believe either it or the sample/icmp6_unreach generates the cksum incorrectly (victor@inliniac.net) (Sam Roberts)
92: - Use libtool-1 or libtool-2 whatever available (alon.barlev@gmail.com) (Sam Roberts)
93: - Avoid looking at /dev and /usr/include when cross-compiling (alon.barlev@gmail.com) Alon: The following code support cross compiling: 1. You CANNOT check for /usr/include stuff as cross compiler is installed else-where. Autoconf know how to do this, use its header detection logic and ask the result. 2. Testing for /dev/ can be done only when not cross compiling... (Sam Roberts)
94: - Corrected target_os check, it was broke for linuxgnu, and m4 syntax was invalid (alon.barlev@gmail.com) Alon:The following change is needed in order to solve two issues: 1. linuxgnu and such target os. 2. You cannot set variable with space before '=' as it tries to execute the variable... 3. Print result of test in case of linux (Sam Roberts)
95: - --with-link-layer broken, was using wrong macro name, and didn't include all link types (Sam Roberts)
96: - ac_cv_c_bigendian is yes, not "big" (alon.barlev@gmail.com) (Sam Roberts)
97: - Remove dead code. (Sam Roberts)
98: - h_len is calculated for ip, udp, tcp, icmp, and igmp, so is allowed to be zero. (Sam Roberts)
99: - libnet_t's fd should be initialized to an invalid value, or libnet_destroy() will close stdin. (Sam Roberts)
100: - Alon's use of AC_CHECK_HEADERS fails to detect headers. Reverted part of 57acd56f09158decb69f301e7547ce8cde6ac63f (Sam Roberts)
101: - With link_none, the link apis were failing with not error message. (Sam Roberts)
102: - man doc makefile wasn't correctly referring to the srcdir (alon.barlev@gmail.com) (Sam Roberts)
103: - Avoid mallocing zero bytes, it perturbs electric fence. (Sam Roberts)
104: - autotools patches for cross compiling and seperate builddir (alon.barlev@gmail.com) (Sam Roberts)
105: - html doc makefile wasn't correctly referring to the srcdir (alon.barlev@gmail.com) (Sam Roberts)
106: - ip_offset is now calculated on the fly, and UDP and TCP no longer use h_len (Sam Roberts)
107: - IP offset calculation should allow nesting of IP protocols. (Sam Roberts)
108: - Remove gccisms in bitfield definitions. (Sam Roberts)
109: - injection type of LIBNET_NONE, for packet construction without injection (also, more const correctness) (Sam Roberts)
110: - Notes about checksumming. (Sam Roberts)
111: - libnet_build_tcp was not returning the ptag. (Sam Roberts)
112: - Packet boundaries are now passed to _do_checksum(), so it can validate its input. Hopefully, this will end the recurring segmentation faults due to buffer overruns. (Sam Roberts)
113: - TCP building is triggering memory overwrites; closer examination shows the link list manipulation to be wrong, and the checksumming approach to be incapable of working. I reworked code to simplify and clarify how it works currently, in preparation to fixing it. (Sam Roberts)
114: - Updated comments and notes. (Sam Roberts)
115: - Null the pointer in the about-to-be-freed structure, not the one on the stack. (Sam Roberts)
116: - Added missing pblock types, and made strings consistent with definitions. (Sam Roberts)
117: - Change version policy, we will be 1.1.4 until 1.1.5 is released. (Sam Roberts)
118: - Bring CHANGELOG up to date with today, and script used to generated it. (Sam Roberts)
119: - Begin implementation and tests for repairing pblocks after an update. (Sam Roberts)
120: - libnet_pblock_insert_before() didn't remove ptag2 from old location (Sam Roberts)
121: - Clarifications in document comments. (Sam Roberts)
122: - Don't doxygen process internal header libnet-headers.h (Sam Roberts)
123: - Declared many constant arguments as const, const-correct code spews warnings when built against libnet. (Sam Roberts)
124: - Include pcap DLT_ types from correct header, was using an internal one before. (Sam Roberts)
125: - Declared many constant arguments as const, const-correct code spews warnings when built against libnet. (Sam Roberts)
126: - Note about build_data, which doesn't update ip_offset, among other problems. (Sam Roberts)
127: - libnet_clear_packet() wasn't clearing all packet context. (Sam Roberts)
128: - Why don't TCP and UDP use the DATA pblock type? (Sam Roberts)
129: - whitespace cleanup (Sam Roberts)
130: - Add libnet_dll.c as extra, so its there for win32, and build libnet_link.c (Sam Roberts)
131: - This file wasn't being built, and needed to include bpf to build. (Sam Roberts)
132: - Forgot to make device a const string here, too. (Sam Roberts)
133: - make string argument constant (Sam Roberts)
134: - only ignore Makefile in libnet/ (Sam Roberts)
135: - Replace u_intX_t with C99 uintX_t. (Thomas Habets)
136: - pclose() following popen(), not fclose() (Thomas Habets)
137: - snoof & dlpi: don't free on libnet_link_close() (Thomas Habets)
138: - Summarize changes for log. (Sam Roberts)
139: - Reindented, removing hard tabs, and using consistent brace positioning. (Sam Roberts)
140: - The non-standard types are no longer used. (Sam Roberts)
141: - /sw/.. path doesn't always exist (Sam Roberts)
142: - src/libnet_link_snoop.c: Only fclose if f!=NULL (Thomas Habets)
143: - src/libnet_link_snoop.c: fixed snoop-based backend. Works on IRIX. (Thomas Habets)
144: - Use uint64_t, not u_int64_t (Thomas Habets)
145: - define a lying gethostbyname2() if it's not defined (Thomas Habets)
146: - define STDOUT_FILENO if it's not defined (Thomas Habets)
147: - configure.in: check for gethostbyname2 (Thomas Habets)
148: - configure.in: Check for uint{16,32,64}_t (Thomas Habets)
149: - Configure switch to install samples (Sam Roberts)
150: - Attempt at applying a patch to get installable samples, which doesn't work. (Sam Roberts)
151: - Pointers not cleared after free could lead to double deallocation. (Sam Roberts)
152: - Convert CRLF to LF. (Sam Roberts)
153: - Auto* changes to work on OS X from git checkout. (Sam Roberts)
154: - Update autobuild endianness and unaligned checks. (Mike Frysinger) (Sam Roberts)
155: - Add srcdir to include path. (Sam Roberts)
156: - Adjust srcdir and builddir so libnet can build out-of-tree (Robin Getz/Mike Frysinger) (Sam Roberts)
157: - Beginning 1.1.5 development. (Sam Roberts)
158:
159: - 1.1.4
160:
161: - Strip CRLF from files not in win32/ (Robert Scheck)
162: - libnet was using HAVE_CONFIG_H in a public header in order to deal with platform types.
163: https://bugzilla.redhat.com/show_bug.cgi?id=501633
164: - Patch to libnet.h.in for compilation on HURD (David Paleino)
165:
166: - 1.1.3
167:
168: Merged 1.1.3 release candidate from packet factory, 1.1.2, debian patches, and
169: my own fixes, including bugs causing memory corruption:
170:
171: - Fixed various errors, including memory corruption, when IPv4 options are
172: modified. (Sam Roberts)
173: - Convert from latin-1 to utf-8, from Robert Scheck. (Sam Roberts)
174: - Fixed doxygen errors and warnings, and added a deveoper script to prepare
175: libnet. (Sam Roberts)
176: - Removed CVS crud, again. (Sam Roberts)
177: - Applied autotools cleanup patch from Stefanos. (Sam Roberts)
178: - Applied patch from Stefanos to remove the autotools ephemera that leaked
179: back in. (Sam Roberts)
180: - Patches from Stefanos. (Sam Roberts)
181: - Updated .so revision to be one backwards compatible interface after
182: 1.1.2.1-fork's. (Sam Roberts)
183: - Fix for debian bug# 418975, IPv6 wasn't updating ip_offset. See
184: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=418975 and sample
185: test_ipv6_icmpv4.c for more information, and reproduction. (Sam Roberts)
186: - merged autogen.sh from 1.1.3, now ltmain.sh comes from autogen.sh (Sam
187: Roberts)
188: - 802.1Q and 802.1X header documentation was incorrectly using /** /** is
189: reserved for doxygen documentation comments, and they didn't have any.
190: That those packet headers, and no others, were marked that way was
191: causing man pages to be generated for them, incomplete man pages that
192: then were being hacked by debian patches 02- and 03-. (Sam Roberts)
193: - Update .so version to be one src change past the last debian release.
194: Debian patches to v1.1.2.1 used 4:0:3, in error, so we use 5:0:4, as per
195: the rules. See Makefile.am comments for reference. (Sam Roberts)
196: - Bug fixes and reproduction code for ip_offset accounting problem in
197: libnet_build_ipv4 (Sam Roberts)
198: - doxygen configuration updated, html seems fine - I don't know about the
199: man pages. (Sam Roberts)
200: - autotools merged from v1.1.3 to v1.1.2 (Sam Roberts)
201: - Remove autotools. And some garbage local files that should not have been
202: in upstream tarball. (Sam Roberts)
203: - Removed object files and cvs conflict residue contained in original
204: package. (Sam Roberts)
205: - debian patch 06 attempts to free the wrong pointer, and also leaks memory
206: from the inner loop. (Sam Roberts)
207: - libnet (inconsistently) uses various signed and/or unsigned typedefs
208: instead of char ANSI C uses char for string literals and the standard
209: library, so this generates many warnings. I've fixed a number of the
210: places where types representing null-terminated strings weren't typed
211: correctly. (Sam Roberts)
212: - 09-fix_hurd-i386_build.patch from debian source package See
213: http://packages.debian.org/source/sid/libnet (Sam Roberts)
214: - 08-fix_libnet_checksum.c.patch from debian source package See
215: http://packages.debian.org/source/sid/libnet (Sam Roberts)
216: - 07-add_libnet-cq-end-loop.patch from debian source package See
217: http://packages.debian.org/source/sid/libnet (Sam Roberts)
218: - 06-fix_libnet_pblock_coalesce_leak.patch from debian source package See
219: http://packages.debian.org/source/sid/libnet (Sam Roberts)
220: - 04-fix_libnet_build_ntp.patch from debian source package See
221: http://packages.debian.org/source/sid/libnet (Sam Roberts)
222: - 03-fix_libnet_802_1x_hdr.3.patch from debian source package See
223: http://packages.debian.org/source/sid/libnet (Sam Roberts)
224: - 02-fix_libnet_802_1q_hdr.3.patch See
225: http://packages.debian.org/source/sid/libnet (Sam Roberts)
226: - 01-fix_libnet-functions.h.3.patch from debian source package See
227: http://packages.debian.org/source/sid/libnet (Sam Roberts)
228: - strip CVS subdirectories from upstream package (Sam Roberts)
229:
230: From 1.1.3-rc:
231:
232: Added a libnet_version() function
233: Fixed a bug in libnet_build_ntp() where two arguments werent used due to a
234: typo
235: Fixed a bug ln libnet_name2addr4 in which it didnt call hstrerror
236: Fixed a memory leak in libnet_if_addr.c
237: Internals: added a payload builder macro
238: Added an HSRP builder
239: Fixed the cdp.c sample code
240: Added AC_PREREQ(2.50) to configure.in to come correct
241: Added a libnet udp header prototype. We need to add an entire exported
242: interface for the sole purpose of casting captured packets, this will
243: presumably be a part of the pcap integration.
244: Added libnet_adv_write_raw_ipv4()
245: Fixed the checksum function
246: Updated the autoconf/automake stuff to be up to date with the latest
247: versions. We now use libtool.
248: Fixed a signed/unsigned comparison warning in the LIBNET_DO_PAYLOAD() macro
249: Changed all empty function prototypes to contain the void keyword
250: Removed all C++ style comments
251: Removed the configure.in check for strerror()
252:
253:
254: Mon Mar 29 09:23:49 PST 2004 1.1.2.1
255:
256: Fixed a typo in the ICMP patch mentioned below
257:
258: Thu Mar 25 10:49:04 PST 2004 1.1.2
259:
260: Fixed the ICMP error message builders (there was a pblock assembly bug
261: that would prevent you from building more than one ICMP {unreach, time
262: exceed, redirect} in succession; the order of operations has changed
263: slightly for building these packets, see the documenation and sample code
264: Added a Sebek builder
265: Fixed a bug in libnet_autobuild_arp() that had it pulling in the address to
266: a pointer instead of just the address
267: Added AM_MAINTAINER_MODE to configure.in
268: Changed the __libnet_dump* namespace to the more descriptive libnet_diag*
269: Added libnet_getpacket_size() to return the size of a packet in a given l
270: Removed "protocol" from the libnet context. It was a waste of four bytes
271: The raw socket interface always uses the "IPPROTO_RAW" protocol
272: Fixed a memory leak in the advanced interface; there is now a function
273: libnet_adv_free_packet() to free the memory allocated for the packet
274: when libnet_adv_cull_packet() is called
275: Fixed a bug on big endian boxes that had TCP and UDP checksums with odd
276: payloads come out incorrect
277: Changes all error messages to look and feel the same:
278: "%s(): foo\n", __func__
279: Added a bunch of htons/htonl fixes
280: Continued to add to the doxygen-based documentation
281: Added support for unconfigured interfaces
282: Changed the number of interfaces libnet can handle from 32 to 512
283: Removed uneeded control structure cruft from libnet_link_dlpi.c
284: Removed sample/ip.c and added sample/ip_link.c and sample/ip_raw.c
285: Added IPv6 fragmentation header builder
286: Added IPv6 routing information header builder
287: Added IPv6 destination options header builder
288: Fixed IPv6 flowlabel and traffic class bitwise math
289:
290: Tue Nov 25 15:33:27 PST 2003 1.1.1
291:
292: Fixed a bug in libnet_build_icmp_redirect: htonl(gateway) --> gateway.
293: Added icmp_redirect.c sample code.
294: Added libnet_autobuild_arp().
295: Added a slightly faster checksum.
296: Added a GRE builder.
297: Fixed a buffer overflow in libnet_build_dhcp().
298: Added more sanity checks to ensure we have proper link or network layer
299: headers when not in advanced mode.
300: Fixed a bug that would sometimes make __libnet_dump_context() crash under
301: linux.
302: Migration from sprintf and strcpy snprintf and strncpy.
303: Fixed bug in libnet_build_ipv4() when calculating size of memory block
304: Removed the support directory -- if you're an OLD version of OpenBSD or
305: FreeBSD you deserve what you get.
306: Added a BGP builder.
307: Changed the error handing functions to be more consistent and use
308: __FUNCTION__.
309: Fixed a bug in libnet_pblock_free() -- replaced it with
310: libnet_pblock_delete().
311: Fixed all of the inconsistencies inside all of the builders and pblock code
312: where some fringe conditions could result in u_longs being crunched into
313: u_shorts.
314: Fixed libnet_pblock_coalesce() to only require one pass through the list.
315: Added better diagnostics (__libnet_dump_context(), __libnet_dump_pblock()).
316: Added Token Ring and FDDI builders (Linux and Solaris only).
317: Added Token Ring and FDDI sample programs.
318: Fixed the handling of TCP and IP payloads when reusing a pblocks.
319: Fixed the handling of IP headers such that if a TCP packet changes size
320: via subsequent calls to libnet_build_tcp(), the IP header automatically
321: changes size as well.
322: Added libnet_pblock_delete() to remove a pblock from the list.
323: Added ip.c sample program (builds an arbitrary IP packet).
324: Added additional payload sanity checks to libnet_build_*.
325: Added a payload to sample/icmp_echo_cq.c.
326: Added an MPLS builder.
327: Added an 802.1x builder.
328: Added an RPC builder! Bout time eh?
329: Fixed do1x.c sample code to make the frame valid.
330: Fixed link-interface semantics under Mac/OSX
331: Changed libnet_stats to all be unsigned long longs to accomodate all of
332: hardcore packet writers.
333: Fixed IPv6 support (to some extent) removed the IP_HDRINCL stuff and
334: reworked the resolver stuff to use net_pton() and inet_ntop().
335: Fixed libnet_build_icmpv4_*() to properly handle the IP header in the
336: payload.
337: Fixed libnet_build_igmp() to handle checksums properly.
338: Fixed a bug in libnet_build_dnsv4() -- now it will work for TCP or UDP --
339: see the sample program for details...
340: Fixed a bug in sample/dhcp_discover.c
341: Added multiple packet interface (called the context queue interface).
342: Until I finish the manpage, see the sample code and README files
343: for instructions on how it works.
344: Fixed Cygwin support.
345: Fixed an OS/X compilation error due to lack of system header files.
346: Fixed OS/X link layer bug.
347: Fixed a bug in pblock_coalesce() that resulted in bad checksums when the
348: advanced mode was enabled.
349: Fixed a potential memory leak in pblock_coalesce().
350: Fixed a potential memory leak in libnet_select_device().
351: Fixed a potential memory leak in libnet_plist_chain_new().
352: Fixed Solaris support for IPv6 address support.
353: Fixed minor bugs in libnet_advanced.c.
354: Added loopback device support.
355:
356:
357: Mon Aug 5 15:18:52 PDT 2002 1.1.0
358:
359: First 1.1.0 non-beta release.
360:
361: Added libnet_adv_write_link() which allows an advanced user to access
362: libnet's low-level frame injection functionality directly.
363:
364:
365: Wed Jul 10 08:18:15 PDT 2002 1.1.0 Beta 07b
366:
367: Added some words to the manpage.
368:
369: Fixed a typo in libnet-functions.h -- forgot a comma.
370:
371:
372: Sun Jul 7 10:37:12 PDT 2002 1.1.0 Beta 07a
373:
374: My bad. Forgot to `make distclean` before last release resulting in some
375: compilation errors.
376:
377: My bad. Forgot to add advanced *_ADV writing support to libnet_write().
378: Simple fix.
379:
380:
381: Tue Jul 2 08:42:57 PDT 2002 1.1.0 Beta 07
382:
383: BETA support for IPv6.
384:
385: Fixed the IP and TCP options bugs that bound the TCP and IP payloads to
386: the IP and TCP headers respectively and saw options being appended
387: after the payload.
388:
389: Added libnet_hex_aton(). This functions reads in arbirtrarily long hex
390: strings from the command line and returns the equivalent byte string. It
391: does an implicit malloc() so make sure to free().
392:
393: Frédéric Raynal submitted a patch to break the coalesce loop down to two
394: passes using realloc resulting in a modest performance increase! Cool!
395:
396: Added "Advanced Mode" which will initialize the library with additional
397: functionality for advanced users who "know what they're doing". Basically
398: this feature will remove some of the sanity checks libnet does when
399: building and injecting packets, at the programmer's peril. It also exposes
400: the libnet_adv() functions.
401:
402: FINALLY changed that irritating struct ether_addr redefintion problem.
403: I internalized the name space of it (-> libnet_ether_addr) so there will
404: be no more issues there. Please update your code accordingly!
405:
406: Added IGMP checksum support which was omitted by accident.
407:
408: Removed netinet/ip_icmp.h from include list. This was causing problems
409: when including dnet.h which includes other system headers. We can
410: probably stand to remove several headers from libnet.h.in.
411:
412: Added sanity check to ensure that when *build_ethernet() is called the
413: injection method is LIBNET_LINK (except when advanced mode is on).
414:
415:
416: Thu Mar 28 22:18:46 PST 2002 1.1.0 Beta 06
417:
418: Fixed ICMP unreachable checksum error and payload issues. Now using the
419: payload interface with unreachables will append the payload to the IPv4
420: header of the "offending packet".
421:
422: Split STP builder into two; libnet_build_stp_conf() and
423: libnet_build_stp_tcn().
424:
425: New CHANGELOG format. :)
426:
427:
428: Mar 24 2002 1.1.0 Beta 05
429:
430: New building logic. Top down. Much smarter, we now build packets and
431: frames like an OS kernel.
432:
433: Added Cisco ISL builder.
434:
435:
436: Mar 18 2002 1.1.0 Beta 04
437:
438: Added an STP builder.
439:
440: Hooks for Cisco ISL builder.
441:
442: Changed libnet_init() to now accept an IP address for the device (so
443: either "fxp0" or "192.168.0.1" will work).
444:
445: Added libnet_clear_packet() to free packet memory when we're done with it.
446:
447:
448: Feb 28 2002 1.1.0 Beta 03
449:
450: Added 802.1q, 802.2, 802.3 builders.
451:
452:
453: Feb 25 2002 1.1.0 Beta 02
454:
455: Fixed Cygwin support.
456:
457:
458: Feb 01 2002 1.1.0 Beta 01
459:
460: Complete new API and overhaul of most everything.
461:
462: Improved linux packet socket support.
463:
464: Renamed libnet_host_lookup() and libnet_name_resolve() to the more
465: intuitive libnet_addr2name() and libnet_name2addr().
466:
467: All of the address resolution functions return host byte order (which
468: is what the build functions want).
469:
470: Removed a ton of code from every corner of libnet.
471:
472: Removed alot of useless crap:
473: misc directory
474: ports directory
475: util directory
476: cleaned out the test directory and moved it to sample
477:
478: The libnet-config script is no longer needed to specify machine
479: endianess -- that has been moved to libnet.h and done at compile time.
480: You can still use it to specify other CPP constants as well as
481: libraries.
482:
483: Added cygwin support.
484:
485: Hooks for a few ieee 802 builders.
486:
487: Added NTP builder.
488:
489: Added DHCP builder.
490:
491: Added BOOTP builder.
492:
493: Added Cisco CDP builder (needs work).
494:
495: Added IPSEC builder (needs work).
496:
497: - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
498:
499: 1.0.2a 02.06.2001 Oops! Messed up the install stuff. Fixed now.
500: Fixed the config.sub to correctly look for arm*
501: architecture.
502: Fixed the test.sh script
503: (Thankz again to syke).
504:
505: 1.0.2 02.03.2001 Added OpenBSD 2.7 etherspoof lkm and kernel patch.
506: (Thankz to obecian).
507: Added FreeBSD 4.0-STABLE (and 5.0-CURRENT?)
508: etherspoof kernel patch.
509: (Thankz to Matt Bing).
510: Added FreeBSD 4 support for automagic MAC address
511: spoofing (via ioctl). No more lkm!
512: (Thankz to Toni Andjelkovic).
513: Added VRRP support.
514: Fixed a NULL pointer check in libnet_checksum.c.
515: (Thankz to syke).
516: Fixed a function naming problem in libnet_if_addr.c.
517: (Thankz to gigisull).
518: Fixed a potential byte error in libnet_version.
519: (Thankz to wotan).
520: Fixed a potential overflow in
521: libnet_link_sockpacket.c and libnet_link_dlpi.c.
522: (Thankz to Jarno Huuskonen).
523: Fixed a manpage discrepancy (get_ip_addr returns
524: host-byte, not network-byte).
525: Fixed arena allocation code (misalignments and
526: whatnot) and arena manpage entry (2 arguments
527: were swapped).
528: (Thankz to Bryan T. Schmersal).
529: Fixed datatype discrepancies (u_char was used
530: liberally when char should have been used).
531: (Thankz to Kyle Hargraves).
532: Fixed the PF_PACKET interface to work correctly.
533: (Thankz to Smiler).
534:
535: 1.0.1b 04.07.2000 Fixed portlist chaining code to allow for more than
536: one active plist chain at a time (as per twitch's
537: patch).
538: Fixed discrepancy between the manpage and code for
539: libnet_close_link_interface. It now returns 1 on
540: success as per libnet standard (thankz to Toni
541: Andjelkovic for pointing this out).
542:
543: 1.0.1a 03.29.2000 Fixed a small bug in libnet_link_dlpi.c.
544:
545: 1.0.1 12.19.1999 Fixed a typo in libnet-headers.h ARH_H -> ARP_H.
546: Fixed a small typo in ether_mod-2.5.c.
547: Pre-happy BD to libnet! She'z almost 2 yearz old!
548:
549: 1.0.0 10.27.1999 Added verbose html documentation.
550: Added verbosely commented example code.
551: Fixed OSPF testcode compile issues.
552: Added ping of death ICMP test code module.
553: Fixed manpage installation wrongness.
554: Fixed a reported bug in OpenBSD etherspoof lkm.
555: Merged OSPF lsa checksum code into main checksum
556: module.
557: Fixed a reported bug in the Makefile.in under
558: Solaris when make install was invoked, the ln
559: failed.
560: Fixed linux-based IP broadcasting using the
561: raw sockets interface.
562:
563: 0.99g 09.13.1999 Added an OSPF builder (which is still in beta).
564: Fixed the Linux/configure.in bug. This was an odd
565: bug that affected Linux-based boxes, but not
566: BSD-based machines. The configure script refused
567: to expand most of the Makefile.in macros in every file
568: because of a conditional check.
569: Fixed some Makefile.in issues.
570:
571: 0.99f 09.09.1999 Changed test/poink.c to not rely on a previous
572: install of libnet to compile.
573: Added a redhat RPM.
574: Changed sourcefile naming scheme to libnet_*.
575: Removed all assertions from the tree.
576: We are moving closer to a 1.0 release and
577: assertions have no place in production code.
578: Furthermore, there should be no exit points
579: inside a library. Currently, all functions that
580: made assertions now return an integral 1 upon
581: success and a -1 when the assertion would have
582: failed (some had to be changed from returning
583: void to returning int). This will not break
584: backward compatbility.
585: Fixed bugs in the arena code.
586: next_packet_from_arena would never return the
587: first chunk of memory, only the "next". It
588: now handles this special case. Thanks to
589: Sascha Gresk for locating this bug.
590: Fixed another potential bug when attempting
591: to allocate large packet sizes inside an arena.
592: Added an OpenBSD 2.5 ether_spoof lkm.
593: Fixed TCP options bugs.
594:
595: 0.99e 07.21.1999 Modified the libnet-config script to work more
596: intutively now. It accepts multiple arguments.
597: See README.libnet-config.
598: Solaris m4/sh fixes (autoconf phase).
599: Internal error handling changed to use libnet_error.
600:
601: 0.99d 06.24.1999 Added: build_icmp_redirect().
602: Added: FreeBSD 3.x support for spoofing source.
603: Added: libnet_error().
604: Added: port list chaining code.
605: MAC addrresses (see README.bpf).
606: Bugfix: libnet_select_device correctly accepts
607: NULL device arguments.
608: Bugfix: build_icmp.c now copies the correct amount
609: of header information.
610: Bugfix: OpenBSD needs HAVE_SOCKADDR_SA_LEN.
611: Changed: write_ip internal semantics. Cleaner
612: and faster now.
613: Changed: init_packet argument parameters. More
614: correct now. Takes a u_short vs. a size_t.
615:
616: 0.99c 05.28.1999 link_int -> libnet_link_int.
617: Misc small testcode fixes.
618: Added libnet_tcp_header and libnet_ip_header.
619: Added libnet-config shell script, see
620: README.libnet-config and the manpage.
621: Updated ports.
622: Revamped checksum module -- it's much simpler
623: and more efficient (ripped out arch specific
624: code which seemed to be buggy with series' of
625: very large packets). Dug Song wrote it, with
626: small fixes/changes by MDS.
627:
628: 0.99b 05.06.1999 Fixed a nasty UDP/TCP + data checksum bug.
629: Header structure further divided into subfiles.
630: Moved get_hwaddr into low-level interface locales.
631: Fixed the BSD get_hwaddr (dugsong@anzen.com).
632: Ported to BSD/OS 3.x.
633: Added `LIBNET_VERSION` symbolic constant.
634: build_ip with payload semantics changed (now
635: requires a payload length which is more
636: intuitive).
637: Fixed the `disappearing MAC address problem`
638: within the linux version of get_hwaddr().
639:
640: 0.99a 04.14.1999 Linux 2.0.x kernels don't have <net/ethernet.h>
641: 0.99 included this header file without checking
642: to see if it present. This is now fixed.
643: Non-x86 systems have no tcp_check function but the
644: stub.c sourcefile did not check this. This is now
645: fixed.
646: Added the utilities directory and get_mac.c.
647:
648: 0.99 04.13.1999 Major manpage redux.
649: Added (broken?) PF_PACKET support for Linux (see
650: README.linux).
651: Moved alot of m4 from configure.in to aclocal.m4.
652: Added Linux m4 macro to detect PF_PACKET.
653: Added build_icmp_unreach
654: Added build_icmp_timestamp
655: Added standard nomenclature for all the ICMP
656: type/code symbolic constants (see the manpage).
657: Changed internal network structure nomenclature.
658: Decided to stop using the word nomenclature so much.
659: Fixed semantics of get_ipaddr (s/PF_INET/AF_INET).
660: Added a symlink in the install directory so libnet
661: is also named `libpwrite`.
662: Added ASN.1 conversion routines, mostly pilfered
663: from ucd snmplib.
664: Removed get_hwaddr from sockpacket.c and made the
665: existing one portable to Linux.
666: Added more testcode and changed testcode structure
667: to be more intuitive.
668: Added init_packet and destroy_packet.
669: Added an arena allocator.
670: Fixed alignment issues on SPARC and Alpha
671: (possibly others with strict alignment
672: requirements).
673: Added a packet dumping routine. Not fully tested.
674: Testcode updates including a master testcode shell
675: script.
676: Added stub functions to ease the eventual
677: transition to a more proper `libnet_*` function
678: naming convention. See README.stubs for more
679: info.
680:
681: 0.10a 02.04.1999 Added the libnet.s2h configuration file to the
682: distribution.
683: GLIBC fix.
684:
685: 0.10 01.31.1999 Many low-level changes, same interface though.
686: Split up the main libnet.h file into two files.
687: Autoconf changes:
688: checks to see if the underlying architecture
689: needs to be aligned.
690: flexible install location.
691: explicitly set $CC option in Makefile.in.
692: Added ensure-dir.sh.
693: Changes DEBUG semantics as I was told the previous
694: stuff broke on some compilers.
695: Created a FreeBSD/OpenBSD ports entry.
696: SGI snoop (drain) interface fixed.
697: Solaris/HPUX DLPI interface fixed.
698: Support for getting local IP addresses.
699: Support for getting local hardware addresses.
700: Added a DNS packet builder.
701: Added an RIP packet builder.
702: Added an ICMP MASKREQ/REPLY packet builder.
703: Added ICMP at the link layer test code.
704: Changed GLIBC version detection semantics.
705:
706: 0.9 12.15.1998 Major changes/additions here...
707: Added lowlevel packet building and writing
708: routines with a codebase from libpcap.
709: Broken DPLI support (fixme!).
710: Added ethernet and ARP building routines.
711: Added ICMP_ECHO building routine which led to the
712: Discovery of an odd kernel panic bug under
713: OpenBSD (see hook.c in test dir).
714: Added IGMP building routine.
715: Reworked autoconf script.
716: Fixed Linux ip_sum vs. ip_csum naming issue.
717: Fixed Solaris checksums (2.4, 2.5.x).
718: Added NetBSD autoconf entry.
719: Changed BSD_BYTE_SWAP semantics to correctly handle
720: IP datagrams through BPF (see write_ip.c).
721:
722: 0.8c 11.10.1998 Added more testcode.
723:
724: 0.8b 10.21.1998 OK. I THINK WE'VE FINALLY FIXED THAT GLIBC THING.
725: Ported to alpha Linux.
726:
727: 0.8a 10.15.1998 Added support for IP TOS bits (oops. Broke
728: backward compatibility again. Like I said, don't
729: rely on this until 1.x).
730:
731: 0.8 10.13.1998 Added support for IP options.
732: Added support for TCP options.
733: Added a dummy version function.
734: Fixed linux libc vs. glibc nomenclature discrepancy.
735: Solaris checksums fixed for non-payload laden
736: packets?
737:
738: 0.7b 09.22.1998 Linux glibc/libc nomenclature anomoly still there.
739: Fixed Linux/BSD icmp header size problem.
740:
741: 0.7a 08.26.1998 Fixed payload support (see test code).
742:
743: 0.7 08.25.1998 Solaris port (checksums broken -- Solaris has gay
744: fucking quirks when it comes to checksumming on
745: raw sockets).
746: Added autoconf scripts.
747: Added psuedorandom number generation code.
748: Added payload support (breaks backward
749: compatability and is untested).
750:
751: 0.6 06.21.1998 Fixed UDP checksum. Removed USE_NAME CPP option
752: made it a run time decision. Makes code more
753: extensible, but breaks backward compatibility.
754:
755: 0.5 06.02.1998 Added TCP/UDP/IP packet assembly routines.
756: Added a checksum function.
757: Added a manpage.
758: Removed daemonizing function (BSD has one).
759: UDP checksums broken...
760:
761: 0.4 01.12.1998 IP checksum (x86 assembly implementation).
762:
763: 0.3 01.12.1998 daemonizing function.
764:
765: 0.2 01.11.1998 raw socket function changed to allow user
766: designated protocol for raw socket
767:
768: 0.1 01.05.1998 Initial release, contains:
769: network byte order -> human readable IP address,
770: human readable IP address -> network byte order,
771: simple raw socket / IP_HDRINCL wrapper,
772: TCP checksum (x86 assembly implementation)
773: EOF
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/icons/cvsweb/back.gif){kind=icon}
Return to CHANGELOG CVS log ![[TXT]](/icons/cvsweb/text.gif){kind=icon}
![[DIR]](/icons/cvsweb/dir.gif){kind=icon}
Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / libnet / doc