Igor Ryzhov [Mon, 12 Jul 2021 19:51:49 +0000 (22:51 +0300)]
ospf6d: fix freebsd mcast group issues
There's a delay in FreeBSD between issuing a command to leave a
multicast group and an actual leave. If we execute "no router ospf6" and
"router ospf6" fast enough, we can end up in a situation when OS
performs the leave later than it performs the join and the interface
remains without a multicast group.
Instead of counting on a one second delay, we must wait until the
interface actually leaves the group.
Philippe Guibert [Mon, 12 Jul 2021 07:22:41 +0000 (09:22 +0200)]
bgpd: associate correct nexthop when using peer link-local
When setting bgp configuration using peers referencing link local
ipv6 addresses, the bgp should be able to handle incoming bgp
connections, and find out the appropriate interface where the
connection comes from.
ipv6 link local sessions work by using bgp unnumbered interfaces
config, but it does not work if we have a shared media with
multiple potential link local ipv6 addresses on the network.
The fix consists in finding out the appropriate interface, when
the local configuration references a link local ipv6 addresses,
and the source address used references an interface. below
configuration illustrates what can be done then:
note: this change does not solve the ability for such config to
create an outgoing connection to remote peer (as the link local
ipv6 address config does not indicate which interface to use).
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
staticd: fix late initialization of blackhole type
If a static route is added to a not-yet-existing VRF, the blackhole type
is not initialized. Initialization must be done before the VRF existence
check.
Rafael Zalamena [Thu, 8 Jul 2021 17:09:20 +0000 (14:09 -0300)]
lib,ospfd,ospf6d: remove duplicated function
Move `is_default_prefix` variations to `lib/prefix.h` and make the code
use the library version instead of implementing it again.
NOTE
----
The function was split into per family versions to cover all types.
Using `union prefixconstptr` is not possible due to static analyzer
warnings which cause CI to fail.
The specific cases that would cause this failure were:
- Caller used `struct prefix_ipv4` and called the generic function.
- `is_default_prefix` with signature using `const struct prefix *` or
`union prefixconstptr`.
The compiler would complain about reading bytes outside of the memory
bounds even though it did not take into account the `prefix->family`
part.
Signed-off-by: Rafael Zalamena <rzalamena@opensourcerouting.org>
Christian Hopps [Fri, 9 Jul 2021 07:58:02 +0000 (03:58 -0400)]
ospf6d: fix backlink check
This code has been wrong ~ever (according to git history). There are 3
conditional blocks with the added assertion that both the LSA and the
vertex being checked can't both be network LSAs.
The third block is clearly assuming both LSA and vertex are router
LSAs b/c it is accessing the backlink and lsdesc as router lsdesc's also
making sure both are p2p links (which they would have to be to point at
each other).
The programming error here is that (A && B) == False does NOT imply !A,
but the code is written that way.
So we end up in the third block one of LSA or vertex being network LSAs
rather easily (whenever that is the case and the desc isn't the backlink
being sought).
This was caught by ASAN b/c the lsdesc and backlinks are being accessed
(> 4 byte field offsets) as if they were router lsdesc's in the third
block, when in fact one of them is a network lsdesc which is only 4
bytes long -- so ASAN flags the access beyond bounds.
Problem: Sometimes the configured Local GR state is not reflected in
show command and peer node. This is causing failures in few of the
BGP-GR topotests.
RCA: This problem is seen when the configuration of local GR state
happens when the BGP session is in OpenSent state and moves to
Established after the configuration is complete.
When the session gets established, we move the GR state value from stub peer
to the config peer. This will result in overriding the GR state to
previous value.
Fix: The local GR state is modified only through CLI configuration and
does not change during BGP FSM transition. In this case it is not necessary
to transfer the GR state value from stub peer to config peer. This way we
can ensure that always the most recent config value is present in peer
datastructure.
Donald Sharp [Wed, 7 Jul 2021 20:00:12 +0000 (16:00 -0400)]
lib: Allow ZAPI_MESSAGE_OPAQUE_LENGTH length of data
We are sending up to ZAPI_MESSAGE_OPAQUE_LENGTH but checking
for one less. We know the data will fit in it to that size.
Also we have asserts on the write to ensure we don't go over
it
Fixes: #8995 Signed-off-by: Donald Sharp <sharpd@nvidia.com>
David Lamparter [Wed, 7 Jul 2021 12:57:36 +0000 (14:57 +0200)]
lib: fix coverity unused result warning
There's nothing that can be done here with an error. Try to make
Coverity understand that this is intentional.
(I don't know if the `(void)` will actually fix the coverity warning,
but I don't really have a better way to figure it out beyond just
getting this merged and waiting for a result...)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
ospf6d: Fix crash in ospf6_asbr_lsa_remove at ospf6d/ospf6_asbr.c:696
Issue: Crash observed when LSAs are removed from LSDB after max age
when there is no area configured.
(gdb) bt
0 raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:51
1 0x00007fdb190548bc in core_handler (signo=6, siginfo=0x7ffdd2f5a470, context=<optimized out>) at lib/sigevent.c:262
2 <signal handler called>
3 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:51
4 0x00007fdb185ad921 in __GI_abort () at abort.c:79
5 0x00007fdb1907f199 in _zlog_assert_failed (xref=xref@entry=0x55f30902aa20 <_xref.21999>, extra=extra@entry=0x0) at lib/zlog.c:581
6 0x000055f308dc4f78 in ospf6_asbr_lsa_remove (lsa=0x55f30a7546d0, asbr_entry=0x0) at ospf6d/ospf6_asbr.c:696
7 0x000055f308dd8f0d in ospf6_lsdb_remove (lsa=0x55f30a7546d0, lsdb=lsdb@entry=0x55f30a73d300) at ospf6d/ospf6_lsdb.c:166
8 0x000055f308dd9701 in ospf6_lsdb_maxage_remover (lsdb=0x55f30a73d300) at ospf6d/ospf6_lsdb.c:376
9 0x000055f308dee724 in ospf6_maxage_remover (thread=<optimized out>) at ospf6d/ospf6_top.c:603
10 0x00007fdb1906520d in thread_call (thread=thread@entry=0x7ffdd2f5ae90) at lib/thread.c:1919
11 0x00007fdb19023e48 in frr_run (master=0x55f30a569b70) at lib/libfrr.c:1155
12 0x000055f308dc09b6 in main (argc=6, argv=0x7ffdd2f5b198, envp=<optimized out>) at ospf6d/ospf6_main.c:235
(gdb)
Steps to reproduce the issue:
1. router ospf6
2. redistribute static
3. ipv6 route 1::1/128 Null0
4. no redistribute static
5. wait for Max aged LSA to flush
6. Check DB, crash occurs.
RCA:
Crash occurred while accessing listgetdata(listhead(ospf6->area_list))
When there is no area attached to any of the interface listhead(ospf6->area_list)
is NULL. Therefore it crashed due to NULL access.
zyxwvu Shi [Wed, 26 May 2021 02:33:55 +0000 (10:33 +0800)]
bgpd: Do not delete peer_af when deactivating peer-group.
There is no peer_af allocated in `peer_activate`. Trying to delete
the structure just results in an no-op and a error return value.
The error message "couldn't delete af structure for peer" is
unexpected.
Signed-off-by: zyxwvu Shi <shiyuchen.syc@bytedance.com>
Renato Westphal [Mon, 31 May 2021 13:27:51 +0000 (10:27 -0300)]
tests: add OSPF graceful restart topotest
Add a new topotest that features a topology with seven routers spread
across four OSPF areas:
* 1 backbone area;
* 1 regular non-backbone area (0.0.0.1);
* 1 stub area (0.0.0.2);
* 1 NSSA area (0.0.0.3).
All routers have both GR and GR helper functionality enabled in
the configuration. The test consists of restarting each router,
one at time, and checking that all forwarding planes (and LSDBs)
are kept intact during those restarts.
A successful run takes about three minutes to finish.
Renato Westphal [Mon, 31 May 2021 13:27:51 +0000 (10:27 -0300)]
tests: add "save_config" parameter to kill_router_daemons()
Using "write memory" to save the daemons' configurations before
restarting them can cause log files to stop working correctly. Add
a new "save_config" to the kill_router_daemons() function to prevent
that from happening when saving the configurations isn't necessary.
Renato Westphal [Mon, 31 May 2021 13:27:51 +0000 (10:27 -0300)]
ospfd: introduce support for Graceful Restart (restarting mode)
RFC 3623 specifies the Graceful Restart enhancement to the OSPF
routing protocol. This PR implements support for the restarting mode,
whereas the helper mode was implemented by #6811.
This work is based on #6782, which implemented the pre-restart part
and settled the foundations for the post-restart part (behavioral
changes, GR exit conditions, and on-exit actions).
Here's a quick summary of how the GR restarting mode works:
* GR can be enabled on a per-instance basis using the `graceful-restart
[grace-period (1-1800)]` command;
* To perform a graceful shutdown, the `graceful-restart prepare ospf`
EXEC-level command needs to be issued before restarting the ospfd
daemon (there's no specific requirement on how the daemon should
be restarted);
* `graceful-restart prepare ospf` will initiate the graceful restart
for all GR-enabled instances by taking the following actions:
o Flooding Grace-LSAs over all interfaces
o Freezing the OSPF routes in the RIB
o Saving the end of the grace period in non-volatile memory (a JSON
file stored in `$frr_statedir`)
* Once ospfd is started again, it will follow the procedures
described in RFC 3623 until it detects it's time to exit the graceful
restart (either successfully or unsuccessfully).
Testing done:
* New topotest featuring a multi-area OSPF topology (including stub
and NSSA areas);
* Successful interop tests against IOS-XR routers acting as helpers.
Renato Westphal [Mon, 31 May 2021 13:27:51 +0000 (10:27 -0300)]
ospfd: rename the graceful restart header
Both the GR helper code and the upcoming GR restarting code are going
to share a lot of definitions. As such, rename ospf_gr_helper.h to
ospf_gr.h, which will be the central point of all GR definitions
and prototypes.
Renato Westphal [Mon, 31 May 2021 13:27:51 +0000 (10:27 -0300)]
ospfd: print extra LSA information in some log messages
Log the LSA advertising router in addition to the LSA type and
ID in the places where that information is necessary to uniquely
identify the LSA in the LSDB.
This is useful, for example, to know exactly which LSA has changed
when the router is exiting from the GR helper mode when a topology
change was detected.
Ondřej Surý [Mon, 15 Feb 2021 07:43:26 +0000 (08:43 +0100)]
doc: Use dpkg-buildpackage to build packages (add note about debuild)
The debuild command fails when we are doing source package only build
because it expects the arch-dependent .changes file to be present. Thus
in the instructions we switch to using dpkg-buildpackage directly and
add a note about using debuild in more complicated scenarios.
Ondřej Surý [Mon, 15 Feb 2021 07:40:08 +0000 (08:40 +0100)]
doc: Add instructions on how to build only source packages
In the CI, it's better to build the source package only once and then
instead of checking out the whole repository, only distribute the source
packages to the individual jobs.
Philippe Guibert [Wed, 30 Jun 2021 12:07:52 +0000 (14:07 +0200)]
bgpd: nht unresolved with global address next-hop
When bgp peers with ipv6 link local addresses, it may receive a
BGP update with next-hop containing both LL and GA information.
By default, nexthop tracking applies to GA, and ignores presence
of LL, when both addresses are present. This is a problem for
resolving GA as next-hop as the next-hop information can be solved
by using the LL address only.
The solution consists in defaulting the nexthop ipv6 choice to LL
when available, and moving back to GA if a route-map is locally
configured at inbound.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
projects / mirror_frr.git / log