Acee [Thu, 18 May 2023 14:43:52 +0000 (10:43 -0400)]
ospfd: OSPF P2MP Delayed Reflooding configuration
Currently, delayed reflooding on P2MP interfaces for LSAs received
from neighbors on the interface is unconditionally (see commit c706f0e32ba8aa8780a0618b6fbba364c383ae05). In some cases, this
change wasn't desirable and this feature makes delayed reflooding
configurable for P2MP interfaces via the CLI command:
"ip ospf network point-to-multipoint delay-reflood" in interface
submode.
Chirag Shah [Fri, 19 May 2023 05:43:08 +0000 (22:43 -0700)]
bgpd: fix aggregate route display
Based on RFC-4760, if NEXT_HOP attribute is not
suppose to be set if MP_REACH_NLRI NLRI is used.
for IPv4 aggregate route only NEXT_HOP attribute
with ipv4 prefixlen needs to be set.
Testing Done:
Before fix:
----------
aggregate route:
*> 184.123.0.0/16 ::(TORC11) 0 32768 i
After fix:
---------
aggregate route:
*> 184.123.0.0/16 0.0.0.0(TORC11) 0 32768 i
* i peerlink-3 0 100 0 i
* uplink1 0 4435 5546 i
184.123.1.0/24 0.0.0.0(TORC11) 0 32768 i
s> 184.123.8.0/22 0.0.0.0(TORC11) 0 32768 i
Chirag Shah [Fri, 19 May 2023 04:49:03 +0000 (21:49 -0700)]
bgpd: fix aggregate route best path select
In ebgp+ ibgp deployment aggregate summary-only route
selected path should always be locally originated
summary route.
When aggregate route summary-only config is removed
The selected path is iBGP peer as its lower cost
Upon reconfiguring aggregate route summary-only,
the locally originated is not selected due to
always choosing first path attribute and bailing
out as no change in route update.
Donald Sharp [Fri, 19 May 2023 16:51:09 +0000 (12:51 -0400)]
tests: pytest does not like return True from a test
From running the test:
bgp_remove_private_as/test_bgp_remove_private_as.py::test_bgp_remove_private_as
/home/sharpd/.local/lib/python3.10/site-packages/_pytest/python.py:198: PytestReturnNotNoneWarning: Expected None, but bgp_remove_private_as/test_bgp_remove_private_as.py::test_bgp_remove_private_as returned True, which will be an error in a future version of pytest. Did you mean to use `assert` instead of `return`?
warnings.warn(
Donald Sharp [Thu, 18 May 2023 20:03:01 +0000 (16:03 -0400)]
zebra: On shutdown stop hook calls for fpm rmac updates
When shutting down zebra, the hook for the rmac update was
not being unregistered. As such it would be possible
to get into a condition where more rmacs are being
added to the queue for handling in the future after we
are told to shutdown.
Donald Sharp [Thu, 18 May 2023 19:59:43 +0000 (15:59 -0400)]
zebra: Properly handle zfpm_g->t_conn_down in zebra_fpm.c
The t_conn_down pointer was being set to NULL when it already
was. The t_conn_down pointer was being dropped( and leaving
a thread possibly running in the background ) which could
cause problems on shutdown. And finally when shutting down
the t_conn_down event was not being stopped at all.
Donald Sharp [Fri, 19 May 2023 13:54:05 +0000 (09:54 -0400)]
zebra: Do not allow old FPM to access freed memory after shutdown
On shutdown, the old FPM queues up dests to be sent to
the FPM listener. This is done through the rib_shutdown
hook. Which is called when the table that the routes are
stored in are being deleted. This dest has pointers
to the rnode. The rnode has pointers to the table it
is associated with as well as the table->info pointer for
the zebra data associated with this table.
The FPM after this attempts to tell this to it's listener
via events. Unfortunately the zvrf, table_id and nl_pid
was being grabbed from memory that had been freed! Since
all this can be grabbed from memory that has not been freed
on shutdown let's switch over to using that instead of freed
memory for gathering data.
Christian Hopps [Thu, 18 May 2023 02:26:49 +0000 (22:26 -0400)]
doc: configure: add configure option to generate .ccls file
`ccls` needs information from FRR build configuration to work,
so allow creation of a custom ccls config during autoconf.
Paraphrasing the doc entry: ccls is a very powerful tool that allows
dev environments to provide sophisticated IDE functionality, e.g.,
semantically aware jumps and code refactoring...
Christian Hopps [Wed, 17 May 2023 11:10:13 +0000 (07:10 -0400)]
tests: fix implicit config file and recently added logic error
- Restore default of looking for a daemon config underneath the router directory
if no config file was specified.
- Recent change for adding unified config support had a logic bug, fix
- Update the one test that conflicted with this default
- comment out asyncio option causing warnings if asyncio wasn't installed.
When FRR is built with the option `--disable-bfdd`, the build process
fails with the following error:
```
zebra/zebra_ptm.c: In function ‘zebra_ptm_init’:
zebra/zebra_ptm.c:119:35: error: ‘FRR_PTM_NAME’ undeclared (first use in this function)
119 | snprintf(buf, sizeof(buf), "%s", FRR_PTM_NAME);
| ^~~~~~~~~~~~
zebra/zebra_ptm.c:119:35: note: each undeclared identifier is reported only once for each function it appears in
make[1]: *** [Makefile:10520: zebra/zebra_ptm.o] Error 1
```
The reason is that `FRR_PTM_NAME` is defined in `version.h` which is not
imported.
Donald Sharp [Wed, 17 May 2023 11:56:27 +0000 (07:56 -0400)]
tests: Change order of config files
Our CI test system is configuring interfaces like this:
int A
<ospfX config>
router ospfX
router-id Z.Y.M.Q
On sufficiently loaded systems, the router-id might not be respected because
the interface A neighbor might have come up by the time the router-id
command is read in. This is a problem for our topotests in that
the tests expect neighbors to be formed with certain router-id,
but the FRR code has stored the new router-id but not accepted
it's use since a neighbor relationship has formed.
Modify the ci test system code to put the 'router ospfX' commands
before the interface commands, thus forcing the router-id to be
read in *before* any possibility that the neighbor can have come
up.
Also, I've filed issue #13452 to address the ordering of commands
with regards to router-id in our ospf protocols. Once that is
done this should be backed out, to return to a more natural ordering
of commands.
Donald Sharp [Tue, 16 May 2023 20:46:35 +0000 (16:46 -0400)]
tests: Clean up commands that do nothing
Recently clear commands were added to the tests that do nothing
because they are using the wrong way to input the command.
Since these do nothing remove them:
Christian Hopps [Sat, 13 May 2023 00:32:54 +0000 (20:32 -0400)]
tests: improve bgp test determinism
don't grep the tail of a log file after running a previous test, there could be
(and have been) other items added to the log in between.
add before and after count of shutdown messages to very the actual message shows
up as the test intended, and keep the search for the shutdown message.
Donatas Abraitis [Wed, 10 May 2023 20:37:47 +0000 (23:37 +0300)]
bgpd: Implement `match source-protocol` for route-maps
The main idea is to filter routes by matching source (originating) protocol
for outgoing direction. For instance, filter outgoing routes to an arbitrary
router that are static only. Or filter out only routes learned from RIP.
The vrf check should use the carefully adjusted `vrfid`, which is
based on globally/reliable interface. We can't believe the
`bvrf->vrf->vrf_id` because the `/proc/sys/net/ipv4/udp_l3mdev_accept`
maybe is set "1" in VRF-lite backend even with security drawback.
Donald Sharp [Tue, 9 May 2023 17:10:35 +0000 (13:10 -0400)]
ospfd: Respect loopback's cost that is set and set loopback costs to 0
When setting an loopback's cost, set the value to 0, unless the operator
has assigned a value for the loopback's cost.
RFC states:
If the state of the interface is Loopback, add a Type 3
link (stub network) as long as this is not an interface
to an unnumbered point-to-point network. The Link ID
should be set to the IP interface address, the Link Data
set to the mask 0xffffffff (indicating a host route),
and the cost set to 0.
FRR is going to allow this to be overridden if the operator specifically
sets a value too.
Fixes: #13472 Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Acee [Tue, 9 May 2023 20:51:03 +0000 (16:51 -0400)]
ospfd: OSPF opaque LSA stale processing fix and topotests.
1. Fix OSPF opaque LSA processing to preserve the stale opaque
LSAs in the Link State Database for 60 seconds consistent with
what is done for other LSA types.
2. Add a topotest that tests for cases where ospfd is restarted
and a stale OSPF opaque LSA exists in the OSPF routing domain
both when the LSA is purged and when the LSA is reoriginagted
with a more recent instance.
Philippe Guibert [Tue, 31 Jan 2023 10:05:12 +0000 (11:05 +0100)]
doc: add bgp allocation-mode per nexthop information
Add bgp allocation-mode per nexthop information.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
PR=81857 Link: https://github.com/FRRouting/frr/pull/12646 Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com> Acked-by: Louis Scalbert <louis.scalbert@6wind.com>
This test demonstrates that a label is allocated for each
ipv6 next-hop. IPv6 test introduces link local ipv6 addresses
as next hops, and compared to IPv4, one can have two different
next-hops depending if the next-hop is defined by a global
address (static route redistributed) or a bgp peer.
This test checks that:
- The labels are correctly allocated per connected next-hop.
- The default label is used for non connected prefixes.
- The withdraw operation frees the MPLS entry.
- If a recursive route is redistributed by BGP, then the nexthop
tracking will find the appropriate nexthop entry, and the
associated label will be found out.
- When a prefix moves from one peer to one another behind the
vrf, then the MPLS switching operation for return
traffic is changing the outgoing interface to use.
- When the 'label vpn export <value>' MPLS label value is changed,
then the modification is propagated to prefixes which use that value.
- Also, when unconfiguring the per-nexthop allocation mode, check
that the MPLS entries and the VPNv4 entries of r1 are changed
accordingly.
- Reversely, when re-configuring the per-nexthop allocation mode,
check that the allocation mode reuses the other label values.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Philippe Guibert [Tue, 10 Jan 2023 13:53:54 +0000 (14:53 +0100)]
topotests: add bgp mpls allocation per next-hop test
A new test suite checks for the mpls label allocation
per nexthop mode. This test checks that:
- The labels are correctly allocated per connected
next-hop.
- The default label is used for non connected prefixes
- The withdraw operation frees the mpls entry.
- If a recursive route is redistributed by BGP, then the nexthop
tracking will find the appropriate nexthop entry, and the associated
label will be found out.
- When a prefix moves from one peer to one another behind the vrf,
then the MPLS switching operation for return traffic is changing
the outgoing interface to use.
- When the 'label vpn export <value>' MPLS label value is changed,
then the modification is propagated to prefixes which use that value.
- When unconfiguring the per-nexthop allocation mode, check
that the MPLS entries and the VPNv4 entries of r1 are changed
accordingly.
- Reversely, when re-configuring the per-nexthop allocation mode,
check that the allocation mode reuses the other label values.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Philippe Guibert [Thu, 16 Feb 2023 12:46:32 +0000 (13:46 +0100)]
bgpd: update time of last change when label nexthop entry changed
A timer attribute is added for each label nexthop entry, in order
to know when the last change occured.
The timer value will be used for troubleshooting by a show
command in the next commit.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
bgpd: export redistributed routes with label allocation per nexthop
The label allocation per nexthop mode requires to use a nexthop
tracking context. For redistributed routes, a nexthop tracking
context is created, and the resolution helps to know the real
nexthop ip address used. The below configuration example has
been used:
Without that patch, only the redistributed routes that rely on a
pre-existing nexthop tracking context could be exported.
Also, a command in the code about redistributed routes is modified
accordingly, to explain that redistribute routes may be submitted
to nexthop tracking in the case label allocation per next-hop is
used.
note:
VNC routes have been removed from the redistribution,
because of a test failure in the bgp_l3vpn_to_bgp_direct test.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Philippe Guibert [Fri, 13 Jan 2023 14:51:59 +0000 (15:51 +0100)]
bgpd: correctly initialize the IP nexthop of redistributed routes
This is a preliminary work to export redistributed routes from
a given VRF in an VPN network. The exportation works well, when
the label allocation is based on an per-vrf mode, but not on
a per nexthop mode.
To associate a label with a connected nexthop, the nexthop
tracking contexts are used. Until today, there was no tracking
context for redistributed routes. But when using this vpn
allocation mode, one needs to know whether the route is directly
connected or not. When using the nexthop tracking context, the
nexthop attribute of the bgp update needs to have the nexthop
properly set. This was not the case for the mp_nexthop_global_in
attribute which was empty.
This commit is mandatory in order to later use nexthop tracking
context.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Philippe Guibert [Thu, 16 Feb 2023 09:39:40 +0000 (10:39 +0100)]
bgpd: use nexthop interface when adding LSP in BGP MPLSVPN
BGP MPLSVPN next hop label allocation was using only the next-hop
IP address. As MPLSVPN contexts rely on bnc contexts, the real
nexthop interface is known, and the LSP entry to enter can apply
to the specific interface. To illustrate, the BGP service is able
to handle the following two iproute2 commands:
> ip -f mpls route add 105 via inet 192.0.2.45 dev r1-eth1
> ip -f mpls route add 105 via inet 192.0.2.46 dev r1-eth2
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Philippe Guibert [Tue, 28 Feb 2023 13:25:02 +0000 (14:25 +0100)]
bgpd: add support for l3vpn per-nexthop label
This commit introduces a new method to associate a label to
prefixes to export to a VPNv4 backbone. All the methods to
associate a label to a BGP update is documented in rfc4364,
chapter 4.3.2. Initially, the "single label for an entire
VRF" method was available. This commit adds "single label
for each attachment circuit" method.
The change impacts the control-plane, because each BGP update
is checked to know if the nexthop has reachability in the VRF
or not. If this is the case, then a unique label for a given
destination IP in the VRF will be picked up. This label will
be reused for an other BGP update that will have the same
nexthop IP address.
The change impacts the data-plane, because the MPLs pop
mechanism applied to incoming labelled packets changes: the
MPLS label is popped, and the packet is directly sent to the
connected nexthop described in the previous outgoing BGP VPN
update.
By default per-vrf mode is done, but the user may choose
the per-nexthop mode, by using the vty command from the
previous commit. In the latter case, a per-vrf label
will however be allocated to handle networks that are not directly
connected. This is the case for local traffic for instance.
The change also include the following:
- ECMP case
In case a route is learnt in a given VRF, and is resolved via an
ECMP nexthop. This implies that when exporting the route as a BGP
update, if label allocation per nexthop is used, then two possible
MPLS values could be picked up, which is not possible with the
current implementation. Actually, the NLRI for VPNv4 stores one
prefix, and one single label value, not two. Today, RFC8277 with
multiple label capability is not yet available.
To avoid this corner case, when a route is resolved via more than one
nexthop, the label allocation per nexthop will not apply, and the
default per-vrf label will be chosen.
Let us imagine BGP redistributes a static route using the `172.31.0.20`
nexthop. The nexthop resolution will find two different nexthops fo a
unique BGP update.
To avoid this situation, BGP updates that resolve over multiple
nexthops are using the unique per-vrf label.
- recursive route case
Prefixes that need a recursive route to be resolved can
also be eligible for mpls allocation per nexthop. In that
case, the nexthop will be the recursive nexthop calculated.
To achieve this, all nexthop types in bnc contexts are valid,
except for the blackhole nexthops.
- network declared prefixes
Nexthop tracking is used to look for the reachability of the
prefixes. When the the 'no bgp network import-check' command
is used, network declared prefixes are maintained active,
even if there is no active nexthop.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Philippe Guibert [Tue, 28 Feb 2023 13:17:17 +0000 (14:17 +0100)]
bgpd: add the bgp_label_per_nexthop_cache struct and apis
This commit introduces the necessary structs and apis to
create the cache entries that store the label information
associated to a given nexthop.
A hash table is created in each BGP instance for all the
AFIs: IPv4 and IPv6. That hash table is initialised.
An API to look and/or create an entry based on a given
nexthop.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Philippe Guibert [Tue, 28 Feb 2023 13:11:30 +0000 (14:11 +0100)]
bgpd: introduce LP_TYPE_NEXTHOP label type
A new label type is introduced: LP_TYPE_NEXTHOP. This new
label type will be used in next commits to allocate labels
for a specific nexthop IP address.
The commit changes add vty and json outputs to display
the new label type and the label values associated.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
projects / mirror_frr.git / log