Better routing table calculation. We are ready to work withmultiple OSPF areas.
Better adjacency building, some minor bugfixes.
Better checking of configuration.
Bugfix in RT calculation.
It was too verbose.
Cleanup in packet.c. Deleted unused parameters.
Deleted unused parameters.
Be sure, that ospf_area is aged before routing table calculation.
Keep al lSAs invalidated.
Small cleanup, indentation and preparation for multiple areas routing table calculation.
This warning is stupid.
Struct area_net changed.
Cleanup in iface.c
Be more verbose in troubles.
Set size of the buffers. (Thanx MJ.)
Staticized lots of local functions.
One less unused variable.
`defaut' should be `default'.
... and a whole bunch of unused parameters and variables in ospf.
Moved the tests for socket existence here.
BTW, where do you exactly set the new buffer sizes?
Be more verbose.
cleanup in lsupd.c, indenting, "struct proto" removed...
'struct proto' removedFinally, I found the bug reported by Andreas Steinmetz. FIXED.
Useless logs removed.
Bugfix in last patch.
OSPF is ready for changing MTU.
dbdes indent an minor changes.
better log() usage.
lsreq.c indented and small 'struct proto' changes.
hello.* reindented, code cleanup.
Deleted useles "struct proto" sending.
Better log() usage.
Indented. No other change.
lsack.c cleaned. Better names for functions andDIRECT acks can be sent in one packet now.
ackd_timer_hook moded to neighbor.c
Fatal bug found. Sometimes BIRD did not originate router LSA. FIXED. :-)
Initialize iterator on the right place.
This can happen now.
Small reversing of previous patch.
Don't repeat "Sheduling rt calc....."
Code and comments cleanup.
hello.c and hello.h cleaned up. No design changes.
Added source addr for multicast socket.
Hotfix for router's parent without nexthop.It will probably work perfect, but I need to eliminate such situation.
Bugfix in external routes calculation.
Easier cleanup of an interface.
Don't free socket's resources.
Patch from Andreas Steinmetz <firstname.lastname@example.org>
Documented the pxlen parameter.
static declarations don't belong to includes. (And most of them wereredundant anyway.)
The code was broken for external /29 to /32 routes. Assuming that youhave one machine publishing a route to 10.1.1.3/32 and another onepublishing a route to 10.1.1.4/32. If the first machine went down theroute to 10.1.1.4/32 was wrongly killed by the old code, leading either...
The OSPF authentication type was sent in host byte order instead of ofnetwork byte order thus breaking interoperability with other routingdaemons on litte endian machines. The patch fixes this but note thatthis breaks compatability with older bird installations using OSPF and...
Useless include deleted.
Minor bug fix in neighbor state machine.
Minor bug in configuration.
Bug in DBDES receiving fixed.
OSPF was not able to be built stand-alone.
Endianity problem fixed. Thanx to S?r?s J?zsef
This prevents infinite loop when bird has more that 60 neighbors. Thanks to Rani Assaf
Endian-related bug fixes sent by Krzysztof Szuster <email@example.com>
Minor OSPF changes for faster startup.
Many spelling mistakes fixed.
Thanks you Harry Moyes.
ABS should be a macro.
Small typo fixed.
Age LSA DB after LSA origination and before routing table calculation.
New trace added.
Bugfix in router importation.
Removed some useless (hope!) code in next hop calculation.
Finally, next hop problem fixed.
Hope, bug in next hop calculation for stub routes fixed.
Fix %<something>I format strings.
Area networks added into configuration.
Changed definition of stub area.
Bugfix in B-bit setting in router LSA.
I will not originate the same lsa before MINLSINTERVAL.
Install route into main routing table just if it's necessary.
Bug in next-hop calculation fixed. (For dual connected neighbors.)
Patch from David Rohleder <firstname.lastname@example.org> thanx.
Hmm, nasty bug, LSUPD was not well flooded via PTP interfaces. :-(
Small bugfix in ext routes calculation.
Fixed many bugs in rt calculation and interface adding. Now it seems to beOK, I'm going to advertise 1.0.4.
I allocate struct ifa before unlocking. So route exported to OSPF are correct.I work also with interfaces that have some problems with socket opening.I declare them as stub.
Serious bug in ext lsa origination found. Going for 1.0.4.
Log some unusual situation.
Added stub interface. (Interface which is propagated to OSPF butit does not sends nor listens packets.)
I added some new options, please look at it and look into doc if it's OK.
Reconfigure poll interval.
Reconfiguration for new options.
Added hellos on NBMA networks. (I don't violate RFC now.)
Bugfix in ext lsa importing.
Fixed bug in nex hop calculation of external routes.
Stupid bug in next hop calculation of external routes.
Typo in rte_better.
My calculation of external routes violated rfc. :-(