summaryrefslogtreecommitdiff
path: root/usr.sbin/bgpd/session.c
AgeCommit message (Collapse)Author
2004-10-19allow neighbor definitions to depend on interface state.Henning Brauer
with this, if a neighbor is configured as dependent on carp0 for example, the neighbor will remain in state IDLE as long as carp0 is not master. once carp0 becomes master the session(s) depending on it immediately go to CONNECT (or ACTIVE, if they're configured passive), reducing failover time. claudio ok, with some input from ryan as well
2004-10-07use a static, const and prefilled u_int8_t[16] to check wether the markerHenning Brauer
at the beginning of each BGP message is all 1s instead of looping and comparing one by one
2004-10-05reject an all zero bgpid, following a similar suggestion fromHenning Brauer
pascal.gloor@spale.com
2004-09-22don't bother shrinking the pfd and index2peer arraysHenning Brauer
ok claudio
2004-09-16when a cloned peer goes back to IDLE state, don't remove him immediatelyHenning Brauer
but keep him around for some more time (an hour here). fixes an issue Arvid was reporting: when a cloned neighbor rejects our OPEN message due to capability negotiation (or missing implementation of those), we set a marker and retry without. However, if we remove the peer immediately, it gets freshly cloned on the next trial, with default settings, thus including capability negotiation, and thus failes again. this will also help with route flap dampening. ok claudio
2004-09-16imsg API cleanup:Henning Brauer
-kill imsg_compose_pid, imsg_compose_fdpass and imsg_create_pid -extend the original imsg_compose/_create API to take pid & fd too -make imsg_compose do imsg_create + imsg_add + imsg_close instead of duplicating the code -adjust all callers to the new API ok claudio
2004-09-16malloc the imsg buffers instead of having them staticly, suggested byHenning Brauer
micskye some time ago
2004-09-09correctly track peer count. fixes a memory corruption.Henning Brauer
exactly the same bug as in ntpd, but caused completely different... strange issues seen by arvid goetting who helped a lot in debugging, bug found after an ntpd issue showed up and I saw the relation. hunting with & ok claudio, otto, millert, ok deraadt
2004-08-11mrt dumps for cloned sessions did not work because the wrong address was used.Claudio Jeker
OK henning@
2004-08-11Cleanup on exit. OK henning@Claudio Jeker
2004-08-10Gahhh, why is this still around? It was never necessary to inform the parentClaudio Jeker
after fd passing. I thought I removed that some time ago... anyway now it is gone.
2004-08-06Forward IMSG_CTL_SHOW_NEIGHBOR messages to the rde so that we can reportClaudio Jeker
the current and max prefix count back to bgpctl. OK henning@
2004-08-04session_main() and rde_main() return... right, a pid. and a pid is a pid_tHenning Brauer
and not an int.
2004-07-13fix some typosJared Yanovich
ok henning otto
2004-07-042 more file descriptors for each RDE and SE inherited from the parentHenning Brauer
we should close
2004-07-03Switch mrt dumping to fd passing. This gives some speed up when extensiveClaudio Jeker
dumping is done. Acctually mrt dumps were broken because of the fd passing. The nice side effect is a much cleaner code, especially in the parent process. OK henning@
2004-06-22Precedence fix; ok henning@Alexander Guy
2004-06-22missing bzero in a function for this strange other address familyHenning Brauer
2004-06-22Templates with unknown AS where always configured as ebgp peers even ifClaudio Jeker
they were ibgp ones. OK henning@
2004-06-22so we call realloc() on our pollfd array and the peer_l one when they shrunkHenning Brauer
to save memory... yet, that realloc call can fail with ENOMEM ;) don't shrink when (needed + reserve < allocated), but (needed + 2 * reserve < allocated) the longer term goal is of course to not fail at all when a shrink-realloc fails... but that's for later
2004-06-20at least somewhat consistently name the TAILQ_ENTRYs... this confused meHenning Brauer
more than once
2004-06-20implement file descriptor passing in the imsg/msgbuf framework, and useHenning Brauer
it to let the main process to prepare new listening sockets (socket() and bind()) on behalf of the session engine, which of course cannot bind() to ports < 1024 any more once it dropped privileges. with some help from theo, claudio ok
2004-06-09move to a dynamically allocated struct pollfd array.Henning Brauer
we used a ststic one with OPEN_MAX entries, which is a rather arbitary limit as OPEN_MAX is _not_ the max # of open fds we can have, but just a default for that setting. in the same move we have to allocate the peer_l array, basically there for pfd-index to peer pointers to prevent peer list scans all time, dynamiccaly to. we overallocate a little and use that reserve until we have to realloc again later to prevent reallocs for every single control connection or a single flapping peer. help & ok claudio
2004-06-06rework bgpd's handling of listening sockets. instead of one for eachHenning Brauer
supported address familiy, keep a tailq of an arbitary number of them. the new struct listen_addr contains the sockaddr and the fd. this fixes quite some nasty behaviour which was a consequence of the previous model. looks right deraadt@, and discussed with claudio
2004-05-28detect absence of PF_KEY interface and/or the TCP_MD5SIG setsockoptsHenning Brauer
at runtime and disable said subsystems if so. helps the guys porting bgpd to $otherBSD, and is actually the right thing to do. claudio ok
2004-05-28don't want SO_REUSEPORT...Henning Brauer
2004-05-21Add support for dynamic announcements. Usefule to annouce temporaryClaudio Jeker
blackhole routes or to make network announcements dependent on a external state (e.g. for carp setups) OK henning@
2004-05-08remove unused argument to control_dispatch_msg(), lintHenning Brauer
2004-05-06oups, spacesHenning Brauer
2004-05-06the Address Family Identifier field in the capability announcement isHenning Brauer
2 octets, thus we need to transform it from/to network byte order... fixes capability announcement and -parsing
2004-05-06do not punish the peer (by holding him in IDLE for IdleHoldTime seconds) whenHenning Brauer
receiving a "unsupported capabilities" notification. Speeds capability negotiation up quite a bit with peers that like to whine about caoabilities they don't understand
2004-05-06improve logging in teh capabilities parserHenning Brauer
2004-05-06scale socket buffer sizes (and thus window size) up to up to 64k,Henning Brauer
but only of tcp md5sig or ipsec is in use. excellent idea by ryan some time ago, claudio and theo agree
2004-05-06there's a little race condition: a session is taken down and its associatedHenning Brauer
write buffers are cleared, but there could be imsgs from the RDE for that peer (e. g. UPDATEs) in the read buffers for the pipe to the RDE or buffered in the RDE or somesuch. Thus, in session_update(), explicitely check for the session state and just drop the message if the session is not in state ESTABLISHED. claudio ok
2004-04-29sock -> fd; ok henningTheo de Raadt
2004-04-28rename sock to fdTheo de Raadt
2004-04-28if a peer follows the extremely misgiuded path that the RFCs just barely allowHenning Brauer
to send a NOTIFICATION and thus ternminating the session when it sees a capability it doesn't support (who would guess: zebra does so), parse the data section of the notifcication to find out what what capabilties it didn't like and do not advertise them the next time the session gets up. In case we get a notification about unsupported capabilities with an empty data part (don't ask for RFCs... and guess who does that), disable capabilty announcement alltogether. claudio ok
2004-04-28prefix the auth related defines by AUTH_, we had a name clash, markus okHenning Brauer
2004-04-27rename the ipsec struct to auth, move all tcpmd5 related fields in there, andHenning Brauer
add a generic "method" field that expresses what method (none/md5sig/ipsec manual/ipsec ike) is in use markus ok
2004-04-27unbreak^2Henning Brauer
2004-04-27repairHenning Brauer
2004-04-27crud stripping; henning okTheo de Raadt
2004-04-27announce route refresh capability, and parse announced route refresh capaHenning Brauer
2004-04-27initial route refresh support per RFC2918Henning Brauer
process incoming route refresh request and notify the RDE not advertised via capabilities yet, claudio ok
2004-04-26load ipsec SAs into the kernel and enable them.Markus Friedl
2004-04-26announce multiprotocol capabilities - IPv4 multicast for nowHenning Brauer
this implies ourgoing capabilities annoucnement is there and just needs the values to be filled in for other shitz we'll support soonish
2004-04-25better error logging in parse_capabilities()Henning Brauer
2004-04-25sizeof(op_len) when dealing with op_len, not sizeof(op_type) - same sizeHenning Brauer
so noop
2004-04-25if remote-as was not configured, inherit the AS from the neighbors openHenning Brauer
message for cloned neighbors, claudio ok
2004-04-25reserve upper half of the (internal) ID space for cloned neighbors, claudio okHenning Brauer