Ameya Dharkar [Thu, 7 May 2020 20:33:02 +0000 (13:33 -0700)]
bgpd: EVPN RT-2 advertised with 2 labels for prefix-routes-only config
L3VNI is configured with "prefix-routes-only" flag. Even in this case,
intermittently, we observed that local EVPN MACIP routes are installed and
advertised with 2 labels and 2 export RTs.
This is a sequencing issue. Consider following case where L2VNI 200 and L3VNI
1000 are configured for tenant vrf vrf-blue.
Bug is observed for following sequence of events:
1. vrf-blue BGP instance is created.
2. L2VNI is created in bgp for vni 200. It is linked to the tenant vrf vrf-blue
in function bgpevpn_link_to_l3vni.
Following code sets "VNI_FLAG_USE_TWO_LABELS" flag for vni 200 as L3VNI is not
yet attached to vrf-blue BGP instance.
/* check if we are advertising two labels for this vpn */
if (!CHECK_FLAG(bgp_vrf->vrf_flags, BGP_VRF_L3VNI_PREFIX_ROUTES_ONLY))
SET_FLAG(vpn->flags, VNI_FLAG_USE_TWO_LABELS);
2. Now L3VNI is attached to vrf-blue BGP instance. In this case, we set
BGP_VRF_L3VNI_PREFIX_ROUTES_ONLY flag for vrf-blue but we do not clear
VNI_FLAG_USE_TWO_LABELS flag set on the corresponding L2VNIs.
This fix resolves following 2 issues observed above.
1. When L2VNI is created in BGP, flag VNI_FLAG_USE_TWO_LABELS should not be set
for this VNI if BGP vrf is not attached to any L3VNI.
2. When L3VNI is attached to the BGP vrf, set "VNI_FLAG_USE_TWO_LABELS" flag
if "prefix-routes-only" is not for the vrf.
UT cases:
1. Flap "prefix-routes-only" config for a vrf.
2. Test following triggers for vrfs with and without "prefix-routes-only"
- Flap L2VNI from kernel.
- Flap L3VNI from kernel.
Donald Sharp [Wed, 6 May 2020 14:13:48 +0000 (10:13 -0400)]
zebra: Loosen ONLINK restrictions a tiny bit
Loosen the ONLINK restrictions such that when an upper
level protocol sends us a nexthop with an ONLINK attribute
just ensure that interface is up and usable. ONLINK effectively
means we know what we are doing to the kernel.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Josh Cox [Tue, 5 May 2020 17:09:14 +0000 (13:09 -0400)]
bgpd: Check to ensure community attributes exist before freeing them
Community attributes might have been removed by an inbound route map, so we
should check to ensure they still exist before trying to free them.
This fixes a segfault described in issue #6345.
David Lamparter [Mon, 4 May 2020 15:30:24 +0000 (17:30 +0200)]
tools: frr-llvm-cg
This dumps call graph data from LLVM bitcode files into a JSON file.
Specifically for FRR, it understands thread_add_*(), hook_*() and
install_element() so it can provide extra information in these cases.
As a general feature, it tries to track down function pointers as far as
easily feasible.
Signed-off-by: David Lamparter <equinox@diac24.net>
Stephen Worley [Fri, 1 May 2020 16:34:43 +0000 (12:34 -0400)]
zebra: force off kernel NHG install with netns VRFs
Force off kernel NHG install with netns-based VRFs for
now. There is not really a good solution for allowing
kernel nexthop groups in namespaced based vrfs.
Signed-off-by: Stephen Worley <sworley@cumulusnetworks.com>
Donald Sharp [Thu, 30 Apr 2020 15:16:28 +0000 (11:16 -0400)]
bgpd: Ensure that we have a ifp pointer
It is possible that the if_lookup_by_index() call will return
a NULL value and calling zclient_send_interface_radv_req. Just
test that we have a valid interface pointer.
Found by Coverity
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The netlink_vrf_change() function is called both when a VRF device
is created in the Linux kernel and when it is activated. This
commit changes this function to perform the VRF misconfiguration
detection only when the VRF device is created, as doing the check
twice would cause a false positive followed by a hard failure (not
to mention the double check is unnecessary since the VRF table ID
can't change once the device is created).
lib: update the CLI xpath index when exiting from the VRF node
All custom "exit-*" commands that exit from a YANG-modeled
CLI node need to use cmd_exit() to ensure the CLI xpath index
(vty->xpath_index) will be updated accordingly.
Changes to ACL rules were not applied to LDP. This fix allows
LDP to be notified when a rule in an ACL filter is modified by
the user. The filter is properly applied to the LDP session.
The filter may cause a LDP session to go down/up or to remove/add
labels being advertised/received from a neighbor.
Pat Ruddy [Wed, 29 Apr 2020 11:22:14 +0000 (12:22 +0100)]
zebra: fix json for show evpn arp-cache vni vtep
The function zebra_vxlan_print_neigh_vni_vtep does not create
a json object when json has been requested from the CLI and as a
result it prints out the information in normal CLI format.
Fix is to allocate the json object when required.
Don Slice [Tue, 28 Apr 2020 11:54:14 +0000 (11:54 +0000)]
zebra: add ability to set retransmit timer for IPv6 RAs
Reported by testing agency that rfc 4861 section 6.2.1 states
that all implementations must have a configuration knob to change
the setting of the advertised retransmit timer sent in RA packets.
This fix adds that capability.
Ticket: CM-29199 Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Intermittently, there is a 30 second delay for a LDP pseudowire to become
operational.
One way to reproduce the issue is: Once PW is up, shutdown link to trigger
a change to the pseudowire's next hop, and then restore link to cause
pseudowire to return to original NH.
Problem Descripton:
The Zebra PW manager installs pseudowires in the data plane when the
following two conditions are met:
1. Pseudowire is labeled via LDP mapping messages
2. A labeled NH route exists to reach the remote pseudowire endpoint
The Zebra PW manager registers a NHT callback when a pseudowire is enabled.
This allows the Zebra PW manager to install or reinstall the pseudowire.
The Zebra PW manager deregisters for the NHT callback when the pseudowire is
disabled. When LDP learns the remote-pseudowire status is 'not forwarding',
LDP notifies Zebra that the pseudowire is disabled.
This creates a race condition where a new labeled NH can be resolved after the
Zebra PW manager deregistered for the NHT callback.
For static pseudowires, it makes sense for Zebra PW manager to deregister for
NHT callbacks for disabled pseudowires. Static pseudowires become disabled
via CLI configuration commands.
For LDP pseudowires, the Zebra PW manager should not deregister for NHT
callbacks for disabled pseudowires.
Overview of changes:
1. Zebra PW manager should not deregister for NHT callbacks when an LDP
pseudowire is disabled.
Zebra PW manager will register for NHT callbacks when the LDP pseudowire
is first enabled.
Zebra PW manager will deregister for NHT callbacks when the LDP
pseudowire is deleted.
2. Remove the 30 second timer that was added in PR4122.
PR4122 tried to fix this race condition with a timer.
Once we eliminate the race condition (by keeping the Zebra PW manager
registered for NHT callbacks), this timer can be removed.
3. Zebra PW manager handling of static pseudowires will remain as-is.
Zebra PW manager will register for NHT callbacks when the static
pseudowire is enabled.
Zebra PW manager will deregister for NHT callbacks when the static
pseudowire is disabled.
Don Slice [Tue, 21 Apr 2020 19:01:35 +0000 (19:01 +0000)]
bgpd: turn off RAs when numbered peers are deleted
Problem reported that in many circumstances, RAs created in the
process of bringing up numbered IPv6 peers with extended-nexthop
capability enabled (for ipv4 over ipv6) were not stopped on the
interface when those peers were deleted. Found several circumstances
where this occurred and fix them in this patch.
Ticket: CM-26875 Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
when removing a whole address-family block from ldpd config
we were erroneously trying to also remove each of the interface
sub-sub-contexts that belonged to it; this would effectively
re-enable the AF we just removed. Work around this by ignoring
these sub-sub-contexts if we detect that we are already
removing the parent block.
Signed-off-by: Emanuele Di Pascale <emanuele@voltanet.io>
lib: Delete the entire access-list only if there are no more entries
When you enter the access-list with the same sequence number but with a
different prefix AND access-list has only a single entry, then the entry
is deleted and the whole access-list is deleted.
That means that "replace entry" never be re-inserted.
David Lamparter [Thu, 23 Apr 2020 18:34:16 +0000 (20:34 +0200)]
bfdd: fix frr-format false warning
Dealing with PRIu64 is unfortunately a bit hacky in the frr-format
plugin, as in, it works correctly with snprintfrr, but breaks on plain
snprintf. There's no good solution unfortunately :/.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
IPv4 Unicast Summary:
BGP router identifier 10.10.10.1, local AS number 65001 vrf-id 0
BGP table version 100003
RIB entries 200005, using 37 MiB of memory
Peers 1, using 21 KiB of memory
isisd: fix display of "isis bfd" in the running configuration
Since this command is modeled using YANG, it's already displayed
as part of the call to nb_cli_show_dnode_cmds(). Calling the
'isis_circuit_config_write' hook was only making that command
to be displayed twice.
The aforementioned hook is still necessary for fabricd, which wasn't
converted to the new northbound model yet.
projects / matthieu / frr.git / log