| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Send all possible nexthops to the parent process and correctly sync
the RIB, FIB and kernel routing table. Based on initial work by pyr@.
OK pyr@ norby@
PS: don't forget that you need to enable multipath support via a sysctl
|
|
ok claudio@
|
|
manpage bits from jmc@
ok norby@ claudio@
|
|
nois ratio on the routing socket. OK norby@ reyk@
|
|
the ospfd on the other side is dropping the packets because the sanity check
fails.
|
|
code which is very similar and especially include RTM_NEWADDR in
dispatch_rtmsg(). With this newly created interfaces will no longer cause
config reload problems.
OK norby@
|
|
structure. This gives greater flexibility and solves an upcoming problem
with multipath routes. OK norby@
|
|
the metric to the highest and least preferred possible value else no routes
are redistributed. OK norby@
|
|
Change kroute.c code so that it correctly tracks multipath routes.
Route redistribution is way more complex in a mulitpath environment.
It is not possible to redistribute multiple pathes and only a part of the
multipath routes may be redistributable. So a route is redistributed as soon
as one of the multipath routes is redistributable.
Tested and OK norby@ and pyr@ -- at least we see no regression
|
|
so don't unref before calling kroute_remove() or we end up with wrong
reference counts.
|
|
for stuff like RTM_IFINFO or RTM_IFANNOUNCE. So move check to a save location.
Always check rtm_type before accessing some random value inside the route
message.
|
|
that we can access it safely; ok claudio
in the long term we must deal with the 32/64 bit issues of if_buadrate and
friends
|
|
|
|
takes care that redistribution is correct after reload. OK norby@ pyr@
|
|
Found by Armin Wolfermann
|
|
Mostly based on a diff by Pierre-Yves Ritschard (I just simplified the parser
code). OK norby@
|
|
state, if known by the driver. this is required to check the full
duplex state without depending on the ifmedia ioctl which can't be
called in the kernel without process context.
ok henning@, brad@
|
|
not 6 any more (rtableid added). need to tell sysctl so.
|
|
address of a interface could be used. Now it is possible to specify a
interface more than once if multiple networks are configured. An alternative
network can be specified via e.g. interface em0:10.0.5.1. The old interface
syntax without the IP still works and uses the main/first configured IP
address.
ospfd now needs to include the IP header on outgoing messages as it is not
possible to specifiy the source address in sendto(2). Additionally all
multicast joins and leaves have to be tracked.
OK norby@
|
|
|
|
define which routes should be redistributed via OSPF. So now you can
define the export list on a fine graded basis. More to come... OK norby@
|
|
Having routes with incorrect ifindex makes it impossible to correctly validate
routes. -current ospfd needs a -current kernel.
|
|
routes. same as in bgpd now. norby ok
|
|
|
|
|
|
|
|
rt_msghdr, same as in bgpd, ok claudio norby
|
|
is known via kernel and ospf. If the kernel route is removed the ospf one
was not added to the FIB. This is an uncommon event so it is OK to request
an update from the RIB in this case.
Additionally ospfd treated bgpd inserted routes like static routes. This
is wrong because IGP routes habe a higher preference over EGP routes. In
this case a bgpd route needs to be overridden by a ospfd one.
With these to fixes bgpd and ospfd start to play nicely together.
OK norby@
|
|
|
|
resulted in a major bottleneck if bgpd was used on the same box -- not clever
to do linear searches over 175k entries :(. This now moves the redistribute
code back into kroute duty and kills the linear list. Also default routes are
now redistributed without the need for a kernel default route.
OK norby@
|
|
OK norby@
|
|
not be used to connect to a OSPF cloud and forced to be passive. With this
ospfd will only announce the carp interface route if the interface is master.
So you can connect a LAN in a redundant way to your ospf backbone. OK norby@
|
|
not a C string and strlcpy() only works on C strings -- returns length of
the source. Found by tedu@ OK deraadt@ tedu@
|
|
|
|
are e.g. created by the PMTU code and are removed after some time.
OK norby@
|
|
up and running. This makes redistribution of carp(4)-ed networks magically
work. OK norby@
|
|
state of static prefixes.
OK norby@
|
|
the buffer. While RTM_IFINFO starts with a struct if_msghdr RTM_NEWADDR
does not. In other words (struct sockaddr *)(next + sizeof(ifm)) is only
correct for RTM_IFINFO and not for RTM_NEWADDR. So move the ifm_type check
up else get_rtaddrs() would access memory outside of buf.
OK henning@
|
|
|
|
With "redistribute (static|connected|default|none)" it is possible to
tell ospfd which external routes should be announced.
Connected routes will be announced only if there is no corresponding
interface configured, in that case the prefix is not external.
Adding and removing of announced prefixes are done automaticaly.
OK norby@
|
|
ok and input claudio@
|
|
"ospfctl show fib interface".
OK henning@ norby@
|
|
OK henning@
|
|
so the strlcpy() will copy to much. Use sa_nlen instead with some additional
logic. This fixes a problem where interfaces where not found.
OK norby@
|
|
by using an in_addr_t as return type. It simplyfies also the usage as in most
cases the mask is used with a struct in_addr address which is also network
byte order. Add prototypes of prefixlen2mask() and mask2prefixlen() to
ospfd.h as it is used or will be used at different places.
OK norby@
|
|
|
|
|
|
OK norby@ deraadt@
|
