summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* IPFIX: handle EPERM for send()astaro_ulogdHolger Eitzenberger2010-10-211-1/+1
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* Remove old IPFIX CollectorHolger Eitzenberger2010-10-211-2/+0
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* IPFIX: use source key when accessing the markHolger Eitzenberger2010-10-211-1/+1
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* IPFIX: more improvementsHolger Eitzenberger2010-10-212-4/+13
| | | | | | | | o reset sequence number on a restart o free allocated messsages o log lost flows Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* IPFIX: implement restart logicHolger Eitzenberger2010-10-212-12/+28
| | | | | | Implement restart logic in case the IFIX Collector is restarted e. g. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* IPFIX: implement timer handlingHolger Eitzenberger2010-10-211-63/+68
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* IPFIX: misc fixesHolger Eitzenberger2010-10-213-7/+24
| | | | | | | o include the size of the Set header length in the Set length. o send IPv4 addresses as a 32bit value. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* IPFIX: the check for llist_empty() is unnecessaryHolger Eitzenberger2010-10-211-3/+0
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* IPFIX: implement proper commit logicHolger Eitzenberger2010-10-213-51/+208
| | | | | | | | | | | | | Implement a proper commit logic for any-size MTUs and any-size data records, therefore getting more more flexibility for implementing other templates. Solve it by introducing ipfix_msg, which abstracts away all the housekeeping work for a packet, such as sets and message length. Also prepare for better handling of cases where the peer is not available (restart etc.). Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* IPFIX: implement basic support for Set 256 (Vineyard)Holger Eitzenberger2010-10-213-46/+87
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* IPFIX: use UDP (required for Vineyard)Holger Eitzenberger2010-10-213-19/+156
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* IPFIX: send IPFIX packagesHolger Eitzenberger2010-10-211-1/+55
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Improve dlsym error messageHolger Eitzenberger2010-10-211-1/+1
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* IPFIX: TCP connectionHolger Eitzenberger2010-10-211-16/+97
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* IPFIX: initial protocol definitionsHolger Eitzenberger2010-10-214-1/+58
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* IPFIX: add initial IPFIX pluginHolger Eitzenberger2010-10-214-2/+114
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Remove the old/nonworking IPFIX codeHolger Eitzenberger2010-10-212-517/+2
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* OPRINT: use standard key printerHolger Eitzenberger2010-10-211-61/+2
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Add 'configure' check for Astaro logmarkHolger Eitzenberger2010-10-213-0/+22
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* NFLOG: enable logmark againHolger Eitzenberger2010-10-211-6/+2
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* BASE: use in_addr key typeHolger Eitzenberger2010-10-211-2/+2
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Add native key type in_addrHolger Eitzenberger2010-10-215-16/+46
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* ASTARO: fix obvious typoHolger Eitzenberger2010-10-211-1/+0
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* NFLOG: only write non zero lladdrHolger Eitzenberger2010-10-211-2/+5
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* ASTARO: correctly handle in/out lladdrHolger Eitzenberger2010-10-211-11/+22
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* TODO updatedHolger Eitzenberger2010-10-211-2/+1
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* ASTARO: use ICMPv6 type/code valuesHolger Eitzenberger2010-10-211-0/+20
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* BASE: log ICMPv6 type/code valuesHolger Eitzenberger2010-10-211-1/+25
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* db: the 'table' config switch is no longer mandatoryHolger Eitzenberger2010-10-211-1/+1
| | | | | | | | With the recent changes to implement a generic database mapping the 'table' attribute became optional. Codify it in the config table. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* db: micro-optimize row writingHolger Eitzenberger2010-10-212-4/+8
| | | | | | | It is specifically not needed to write the whole ulogd_value, just set its type to ULOGD_RET_NONE. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* db: improve temporal keymap handlingHolger Eitzenberger2010-10-213-37/+44
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* NFCT: cleanupHolger Eitzenberger2010-10-211-25/+16
| | | | | | | | | o turn conntrack refcount signed, thus being actually able to catch underflows. o fix obvious typo when accessing ct->last_seq. o use BUG_ON() instead of assert(). Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* db: remove column debugging aidHolger Eitzenberger2010-10-212-10/+1
| | | | | | Easy computable, therefore no need. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* ASTARO: fix one byte extra printHolger Eitzenberger2010-10-211-1/+1
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* Disable BUG_ON() if NDEBUG setHolger Eitzenberger2010-10-212-1/+11
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Replace assert() by BUG_ON()Holger Eitzenberger2010-10-211-9/+4
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Improve key_reset()Holger Eitzenberger2010-10-211-6/+2
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* PGSQL: update documentation for ->prepare()Holger Eitzenberger2010-10-211-4/+1
| | | | | | | No longer true, as the statement no comes from elsewhere with the new and with the old approach. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* db: use ulogd_free_keyset() instead of open-coded versionHolger Eitzenberger2010-10-214-16/+5
| | | | Signed-off-by: Holger Eitzenberger <heitzenberger@astaro.com>
* db: minor improvementHolger Eitzenberger2010-10-211-1/+1
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Fix type mismatch for raw.pktlenHolger Eitzenberger2010-10-211-1/+1
| | | | | | | | | | Seen this: ulogd: type mismatch raw.pktlen(sys1) <-> raw.pktlen(log1) Which fixes also the empty length="" in the Astaro log. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Reset key source pointer on errorHolger Eitzenberger2010-10-211-6/+13
| | | | | | | | | | | | | | | | | | | Also it is specifically not an error if a key pointer already has a source link. This may happen e. g. if a database plugin errors out in it's ->start() routine, having already the source links set. This is what happens: PGSQL: configure NFLOG: configure stack_resolve_keys(stack); PGSQL: start() /* error, back to PsInit */ PGSQL: configure() stack_resolve_keys(stack); /* key->source already set */ Fix it by removing the check in stack_resolve_keys(), also reset key->source correctly on other errors. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* db: add consistency check to db_columnHolger Eitzenberger2010-10-212-4/+12
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Correctly write empty columnsHolger Eitzenberger2010-10-212-1/+11
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Make key name a pointerHolger Eitzenberger2010-10-214-36/+38
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Fix key check by removing key.source overlayHolger Eitzenberger2010-10-2112-65/+59
| | | | | | | | | | | | | | The key check failed after introducing ulogd_key and more importantly by moving key.type to ulogd_value. Because of the overlay of key.u.source and key.u.val this broke the key check in stack_resolve_keys(), because the value.type was already set then. Fix it by removing this overlay (remove that union). This increases the ulogd_key size by 4 bytes, which is not a problem at all. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Enable again key checkHolger Eitzenberger2010-10-211-2/+0
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* ASTARO: do not print IPv6 payload lengthHolger Eitzenberger2010-10-211-3/+0
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* NFLOG: only set a valid interface indexHolger Eitzenberger2010-10-211-2/+4
| | | | Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>
* Fix utoa() return valueHolger Eitzenberger2010-10-211-14/+12
| | | | | | | The return value for utoa() is wrong for numbers with an even number of digits. This patch fixes it. Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org>