summaryrefslogtreecommitdiffstats
path: root/include
Commit message (Collapse)AuthorAgeFilesLines
* some old libc's don't define IPPROTO_SCTP/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-11-141-0/+6
|
* don't use kernel headers installed on system, but include our own ↵/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-11-114-3/+140
| | | | linux_nnfnetlink_conntrack.h
* o nfct_build_conntrack is too much, the only clients are new_conntrack and svn_t_libnetfilter_conntrack-0.0.26/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-11-091-17/+17
| | | | | | | update_conntrack, and it doesn't even fit well for both cases. So I decided to kill it and inline the code adapting it when was necessary. o Convert all unsigned int/long to POSIX types u_int32_t. Better now than later :(.
* o move nfct_handler to libnetfilter_conntrack.c, better for encapsulation/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-11-061-4/+0
| | | | | | | | o fixed ICMP ID handling o fix -> libtool: link: libtool library `nfct_proto_*.la' must begin with `lib' o remove wrong flag at extensions/Makefile.am o bumped version to 0.0.26 o fixed versioning :(
* destination ipv6 address is also 128bits, not 64./C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-11-041-1/+1
|
* add nfct_fd() function/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-11-041-0/+2
|
* add extra 'data' argument to callback functions/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-11-041-6/+6
|
* o Kill non-portable NIPQUAD, use inet_ntoa instead./C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-11-031-1/+1
| | | | | | | | | | | | o nfct_build_conntrack flit bits of status, timeout, id and mark. o finish incomplete support for marks: kernel part missing (patch on the way) o network byte order translation in build_conntrack instead of nfct_conntrack_alloc. Now this translation is transparent to the clients of the library. o Kill last blank space output in nfct_sprintf_[conntrack|expect] (Thanks to Krzysztof Oledzk for reporting this). o add missing initialization of buffer in nfct_default_*_display. o Bumped version to 0.2.5.
* o Bumped version to 0.2.4/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-11-011-3/+3
| | | | | | | o Redefine NFCT_ALL_CT_GROUPS o Fix typemsg2enum prototype, flags and type are 16 bits long, not 8 bits o Fix wrong expectation timeout and ID output o Fix getting and killing conntracks by ID
* Special thanks to Deti Fiegl from the Leibniz Supercomputing Centre in ↵/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-311-4/+7
| | | | | | | | | | | | | Munich, Germany for providing the "fast" hardware to reproduce spurious bugs ;) List of changes: o Replace misleading flag NFCT_ANY_GROUP by NFCT_ALL_GROUPS o Update test file to use NFCT_ALL_GROUPS o Add missing check of CTA_PROTOINFO_TCP that resulted in a segfault in conjuction with events. o Fix ICMP conntracks output o Add missing prototype definition of nfct_default_expect_display_id in libnetfilter_conntrack.h
* o Added the expectation printing API/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-301-4/+10
| | | | o Bumped version to 0.2.2
* o make static protocol handler functions/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-302-1/+3
| | | | | | | | o move build_tuple_proto and build_protoinfo to the extensions where it really belongs to. o Reworked the conntrack and expect netlink handlers o Fix expectation table output, now it's similar to the /proc output o Bumped version to 0.2.1
* Thanks to Harald for all the comments./C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-292-25/+81
| | | | | | | | | | o libnetfilter_conntrack.h splitted into two parts: what is visible to application programs and what is visible to extensions. o Killed includes asm/types.h and linux/if.h o Fixed nasty wrong ipv6 definition o Stolen the status bits from ip_conntrack.h, we don't include ip_conntrack.h anymore. o move nfct_handle to libnetfilter_conntrack.c: better for encapsulation
* o Renamed nfct_[set|unset]_callback to nfct_[register|unregister]_callback/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-281-6/+49
| | | | | | | | | | | | | | | o Added some very brief comments to libnetfilter_conntrack.h o Implemented the conntrack printers API nfct_sprintf_* o Now nfct_default_conntrack_display display the classical /proc output, and nfct_default_conntrack_display the classical + conntrack ids o Use nfnl_talk if there's no data expected from kernel space to be processed, that is the case of nfct_[get|delete]_conntrack o Added some missing memset's zeroing o Code simplification: killed some char *buf where struct nfnlhdr is enough o Killed protocol handler destructors (fini) and nfct_unregister_proto: The library is unloaded if something goes wrong (different library versions), the modules never gets inserted in the proto_list. Fixes a segfault. o Bumped version to 0.2.0
* o new nfct_handler prototype/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-271-5/+10
| | | | | | | | | | | o NFCT_COUNTERS splitted in NFCT_COUNTERS_[ORIG|RPLY] o all global vars are now static o kill nfct_set_handler, it was too much o fixed very stupid bug in counters printing o fixed conntrack getting: invalid netlink flags NLM_F_[ROOT|MATCH] o nfnl_send returns the proper error to the client, instead of returning -1 o some cleanup's: killed the ret, it was useless o test for the conntrack API completed, still missing the expectation test
* o Add conntrack event notification test/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-231-3/+10
| | | | | | o Define NFCT_ANY_GROUP flag o Now callback can return a value to stop receiving events o implement nfct_unset_callback()
* o Bumped version to 0.1.3/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-211-23/+39
| | | | | | | | o Add support for ID's o Fixed stupid bug in NFCT_* flags, I'm stupid o Simplify handler logic o Define event message NFCT_MSG_* o Add support for conntrack marking (kernelspace part still missing)
* Some API changes, still some slighty modification are about to come before/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-211-27/+21
| | | | the first 1.0 release
* o Missing flags fro nfct_open()/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-201-0/+5
| | | | | o fix some indentation o fix a leak on error path in ncft_open()
* Commit libnetfilter_conntrack.h, I missed it :(/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-201-57/+126
|
* Rename libnfnetlink_conntrack to libnetfilter_conntrack, for some/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-201-0/+0
| | | | unknown reason this didn't happen in the lastest commit.
* Major changes, this library isn't libnfnetlink_conntrack anymore./C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=pablo/emailAddress=pablo@netfilter.org2005-10-165-7/+728
| | | | | We provide an high level interface that abstracts from the netlink sockets. Now users don't need to know anything about them.
* - Add missing files: include/libnfnetlink_conntrack/Makefile.am and ↵/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-10-072-1/+7
| | | | | | | | | | | | | | include/Makefile.am. - Rename list_conntrack_handler to callback_handler, IMHO a proper name for such function. - Use new nfnl_open prototype: Now it's got four parameters. - Kill recurrent definition of the structure nfnlhdr: Actually this should go somewhere in libnfnetlink, later. - Ignore utils subdirectory. It contains a testsuite that is currently broken. Yes I know you're aware of it ;) it's on the TODO list. I'll fix later. - ctnl_error now has a nicer definition. - kill some unneeded ctnl_error messages on failure. (Pablo Neira)
* include ip_conntrack.h for status bit enum/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-09-241-0/+3
|
* fix include paths, use correct automake version, add missing Makefile.am/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-09-231-0/+2
|
* This patch includes the following updates for the userspace/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-08-051-1/+6
| | | | | | | | | | | | | | libnfnetlink_conntrack library: a) change ctnl_open prototype: Now the subsystem is passed as parameter to select if we work working with given subsystem, say NFNL_CTNETLINK_CONNTRACK[_EXP]. b) added functions ctnl_[new|get|del]_expect c) minor change in ctnl_build_tuple that let us create tuples based on CTA_* and CTA_EXPECT_* attributes. (Pablo Neira)
* fix up include path names/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-07-301-1/+2
|
* restructuring libctnetlink -> libnfnetlink_conntrack/C=DE/ST=Berlin/L=Berlin/O=Netfilter Project/OU=Development/CN=laforge/emailAddress=laforge@netfilter.org2005-07-301-0/+122