It was reported that "show ipv6 route vrf <vrfname>", "show ipv6 route
vrf <vrfname> ::/0 " or "show ipv6 route vrf <vrfname> json" all
displayed that the nexthop was in the default vrf. This was because
the kernel netlink messages would supply the RTA_OIF of the loopback
interface for the kernel-created default route for the vrf, where ipv4
did not supply any RTA_OIF. This fix suppresses the display if the
nexthop and route entry are in different vrfs and the nexthop is
NEXTHOP_TYPE_BLACKHOLE.
Ticket: CM-21722 Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Donald Sharp [Fri, 29 Jun 2018 13:49:08 +0000 (09:49 -0400)]
zebra: Remove need for ACK for netlink messages
Kernel requests via netlink are synchronous.
Therefore we do not need to specify a need for a ACK and
we can make the netlink_cmd NONBLOCKING
1) If the netlink message is going to cause an error
we will still get one. Since results from the kernel
are synchronous we will get the error message on the
netlink_cmd socket and handle it
2) If the netlink message is going to send more than
one packet we will still get them all. Since the results
from the kernel are synchronous we will receive all data.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Mark Stapp [Thu, 12 Jul 2018 13:51:52 +0000 (09:51 -0400)]
tests: align with srcdest lib changes, remove bgp binary
Align the srcdest table test with recent api changes that use
'const' more strictly. Remove test_bgp_table binary - looks
as if it was committed by mistake.
Quentin Young [Tue, 10 Jul 2018 19:38:11 +0000 (19:38 +0000)]
doc: cleanup maintainer release document
* Sphinx numbered lists use 1. not 1)
* Indents are three spaces
* Blocks under lists should be indented along with the rest of the list
content
* Sphinx converts :: to : when used as a code block marker
* Apply some preformatted markup to things that should be preformatted
* Remove trailing whitespace
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
zebra: mpls entry interface lookup adaptation for vrf backend
The search algorithm for interface based on ifindex only is adapted to
vrf netns based too. Only the default netns will be used to search the
interface index.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
zebra: pbr rule lookup index adaptation for vrf netns
the interface lookup based on ifindex in the case the target vrf is
unknown is using the generic vrf api. Like that, in the case of vrf
based netns, the search across different netns other than the default
one are not searched.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
zebra: upon RTM_NEWROUTE/DELROUTE, find appropriate interface
The interface lookup algorithm is different according to if we are on
netns vrf or not. If we are on the former case, then we only have to
parse the interfaces of the netns, while if we are on the other case, we
have to parse all the interfaces of all the vrfs ( since index is not
overlapping in the latter case).
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
This commit removes various parts of the bgpd implementation code which
are unused/useless, e.g. unused functions, unused variable
initializations, unused structs, ...
The current behavior of the `bgp default shutdown` command is to set the
state of all newly configured peers to shutdown. This leads to a problem
when restarting bgpd, because all peers will then be seen as newly
configured, which leads to all peers being set to shutdown after each
restart.
This behavior is undesired and not common when comparing the
implementation against other vendors. This commit moves the `bgp default
shutdown` configuration underneath the peer-group and peer
configuration, to ensure that existing peers will not be set to shutdown
after a daemon restart.
lib: add the pathname to get the realpath in case of failure
Sometimes, the file under /var/run/netns may not be authorised to be
read ( because it is not read permission for frr user, for instance).
so it is good to know what happened.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Arthur Jones [Thu, 14 Jun 2018 13:44:38 +0000 (06:44 -0700)]
unit tests: support code coverage instrumentation and reports
Currently, make check runs the unit tests and reports pass/fail,
but we have no way to guage how much of the code is covered by
these tests. gcov provides those statistics on a per source
file basis, but requires special CFLAGS and LDFLAGS. Here, we
add the --enable-gcov configure option to setup those options
correctly. We also add a make target called check-coverage,
which runs the unit tests, runs gcov and uploads the data to
the codecov.io cloud service for display.
Finally, we include a Dockerfile-coverage which creates a
container image in alpine linux to run the tests. To create
the image:
This commit fixes two issues:
- memcpy() using containers of different sizes when using addr2sa(), mixing
'struct sockaddr_storage' and 'union sockunion'.
- addr2sa() function not being thread safe (using a local static variable as
container.
Chirag Shah [Sat, 30 Jun 2018 06:58:54 +0000 (23:58 -0700)]
bgpd: l3vni del to free ip prefix routes from vrf
In Symmetric routing case, L3VNI stores evpn MAC_IP routes
as IP_PREFIX routes in associated bgp_vrf and fib.
When vxlan device for l3vni goes down, triggers l3vni delete
in bgp.
As part l3vni delete, evpn ip prefix routes associated
with the vni need to be withdrawn from zebra as well
bgpinfo needs to be freed.
bgp_delete does not free bgp_info associated
to evpn ip prefix routes (link to bgp_vrf).
Call to uninstall_evpn_route_entry_in_vrf() properly
cleanup bgp_info as well triggers appropriate updates.
Ticket:CM-21443
Testing Done:
On DUT, bringup symmetric routing configuration, learn
EVPN Type-2 and Type-3 Routes.
Type-2 MAC_IP routes will be stored as ip_prefix in vrf table
during l3vni bring up.
Remove L3vni, deletes all ip_prefix routes from the zebra, kernel
vrf route table and bgp_info is freed.
Check the show bgp memory stats for bgp_info post l3vni flap.
Signed-off-by: Chirag Shah <chirag@cumulusnetworks.com>
Quentin Young [Mon, 2 Jul 2018 18:07:28 +0000 (18:07 +0000)]
zebra: remove asserts on task nullity
While ZAPI I/O threads make a best effort to kill any scheduled tasks on
their threadmasters, after death another pthread can continue to
schedule onto the threadmaster. This isn't a problem per se since the
tasks will never run, but it also means that asserting that it hasn't
happened is pointless.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
F. Aragon [Mon, 2 Jul 2018 15:05:17 +0000 (17:05 +0200)]
bgpd zebra: element overflow (PVS-Studio)
The warning given by PVS-Studio is related to per-element overflow (there is
no real overflow, because of how elements are mapped in the union). This
same warning is typically reported by Coverity, too.
Don Slice [Wed, 27 Jun 2018 16:35:55 +0000 (16:35 +0000)]
zebra: duplicate nexthops for cross-vrf static routes on vrf bounce
Problem created by the fix for cm-21306 (inactive cross-vrf static routes
when vrfs were bounced.) Determined that in another case, that fix would
cause duplicate nexthops to appear in the table. Resolved the problem by
removing the vrf static route process from the zebra "add" process leaving
it in the zebra " if up" process as added in cm-21306 since that's the point
that the vrf device is now functional.
Ticket: CM-21429 Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
If a cache server was added after rpki was started it's tr_socket would
not be initialized. This would lead to a segfault if the rtr manager
ever decides to switch to that socket or if rpki support is stopped.
G. Paul Ziemba [Fri, 29 Jun 2018 21:55:59 +0000 (14:55 -0700)]
bgpd: fix NULL dereference in vrf-vpn leak config if before default instance
Issue 2473
VRF-VPN route-leaking configuration commands assumed existence of
default bgp instance. If a non-default vrf configuration occurred
before the default vrf configuration, it would result in a (NULL)
dereference to the non-existent default vrf.
This change 1) detects non-existence of default vrf and skips corresponding
RIB updating that normally occurs during configuration changes of the
route-leaking parameters; and 2) when the default bgp instance is defined,
runs the deferred RIB updating.
In vpn_leak_to_vrf_update_all(), replace early return if bgp_vpn is NULL
with assert, as the former would lead to subtly wrong RIB contents. The
underlying issue that led to commit 73aed5841a1fe31e8a1f251dfcf87e1f2b1f2463
should be fixed by the above changes.
projects / mirror / frr.git / log