When bgp is coming up and is reading a non-integrated config.
The bgp connection to zebra has not fully had a chance to start.
As such when a redistribute line is parsed the attempt is
made to install it but it was erroring out with a warning.
This caused the `redistribute XXX` line to create a error
message to the end user.
Since bgp calls zclient_send_reg_requests which re-registers
the redistribute call once the actual zebra connection is up
and once bgp comes alive this is ok.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
In some situations we already know the ifp and by extension
the ifindex there is no need to look it up for every
route we send to zebra.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This improves code readability and also future-proofs our codebase
against new changes in the data structure used to store interfaces.
The FOR_ALL_INTERFACES_ADDRESSES macro was also moved to lib/ but
for now only babeld is using it.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Performance tests showed that, when running on a system with a large
number of interfaces, some daemons would spend a considerable amount
of time in the if_lookup_by_index() function. Introduce a new rb-tree
to solve this problem.
With this change, we need to use the if_set_index() function whenever
we want to change the ifindex of an interface. This is necessary to
ensure that the 'ifaces_by_index' rb-tree is updated accordingly. The
return value of all insert/remove operations in the interface rb-trees
is checked to ensure that an error is logged if a corruption is
detected.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
IFINDEX_DELETED is not necessary anymore as we moved from a global
list of interfaces to a list of interfaces per VRF.
This reverts commit 84361d615.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
This is an important optimization for users running FRR on systems with
a large number of interfaces (e.g. thousands of tunnels). Red-black
trees scale much better than sorted linked-lists and also store the
elements in an ordered way (contrary to hash tables).
This is a big patch but the interesting bits are all in lib/if.[ch].
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Problem reported with the log displaying error messages if bgpd was
enabled in /etc/frr/daemons but bgp wasn't actually configured. The
problem was due to operating on interfaces before if_create had been
called (which happens at "router bgp" not at frr starting. Moved the
checks for the bgp instance before operating on interfaces. Manual
testing successful and bgp-smoke completed with no new issues.
Ticket: CM-13504
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Reviewed-by: CCR-6738
list_free is occassionally being used to delete the
list and accidently not deleting all the nodes.
We keep running across this usage pattern. Let's
remove the temptation and only allow list_delete
to handle list deletion.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This is a fallout from PR #1022 (zapi consolidation). In the early days,
the client daemons would allocate enough memory to send all nexthops
to zebra. Then zebra would add all nexthops to the RIB and respect
MULTIPATH_NUM only when installing the routes in the kernel. Now things
are different and the client daemons can send at most MULTIPATH_NUM
nexthops to zebra, and failure to respect that will result in a buffer
overflow. The MULTIPATH_NUM limit in the new zebra API is a small price
we pay to avoid allocating memory for each route sent to zebra.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
The previous code assumed that all nexthops of an ECMP route were of
the same address-family. This is not always the case.
Reported-by: Don Slice <dslice@cumulusnetworks.com>
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
bgp_attr_deep_dup is based on a misunderstanding of how route-maps work.
They never change actual data, just pointers & fields in "struct attr".
The correct thing to do is copy struct attr and call bgp_attr_flush()
afterwards.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
FLAG_BLACKHOLE is used for different things in different places. remove
it from the zclient API, instead indicate blackholes as proper nexthops
inside the message.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Some differences compared to the old API:
* Now the redistributed routes are sent using address-family
independent messages (ZEBRA_REDISTRIBUTE_ROUTE_ADD and
ZEBRA_REDISTRIBUTE_ROUTE_DEL). This allows us to unify the ipv4/ipv6
zclient callbacks in the client daemons and thus remove a lot of
duplicate code;
* Now zebra sends all nexthops of the redistributed routes to the client
daemons, not only the first one. This shouldn't have any noticeable
performance implications and will allow us to remove an ugly exception
we had for ldpd (which needs to know all nexthops of the redistributed
routes). The other client daemons can simply ignore the nexthops if
they want or consult just the first one (e.g. ospfd/ospf6d/ripd/ripngd).
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
The code path for a deleted interface was calling
zebra with a IFINDEX_DELETED, which caused zebra
to bitch and moan about the issue. Since the
only thing this function does is call zebra
to deregister the RA stuff, don't set the
ifindex to DELETED till afterwords.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This reverts commit c14777c6bf.
clang 5 is not widely available enough for people to indent with. This
is particularly problematic when rebasing/adjusting branches.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
This allows frr-reload.py (or anything else that scripts via vtysh)
to know if the vtysh command worked or hit an error.
Most of the attributes in 'struct attr_extra' allow for
the more interesting cases of using bgp. The extra
overhead of managing it will induce errors as we add
more attributes and the extra memory overhead is
negligible on anything but full bgp feeds.
Additionally this greatly simplifies the code for
the handling of data.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
bgpd: Fix missing label set
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Implement support for sticky (static) MACs. This includes the following:
- Recognize MAC is static (using NUD_NOARP flag) and inform BGP
- Construct MAC mobility extended community for sticky MACs as per
RFC 7432 section 15.2
- Inform to zebra that remote MAC is sticky, where appropriate
- Install sticky MACs into the kernel with the right flag
- Appropriate handling in route selection
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Core EVPN route handling functionality. This includes support for the
following:
- interface with zebra to learn about local VNIs and MACIPs as well as
to install remote VTEPs (per VNI) and remote MACIPs
- create/update/delete EVPN type-2 and type-3 routes
- attribute creation, route selection and install
- route handling per VNI and for the global routing table
- parsing of received EVPN routes and handling by route type
- encoding attributes for EVPN routes and EVPN prefix creation (for
Updates)
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Reviewed-by: Daniel Walton <dwalton@cumulusnetworks.com>
Implement the command 'advertise-all-vni' under the EVPN address-family
in order to allow the local system to learn about local VNIs (and MACs
and Neighbors corresponding to those VNIs) and exchange with other EVPN
speakers.
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Reviewed-by: Dinesh Dutt <ddutt@cumulusnetworks.com>
The if_update function was taking the interface name as
input and reapplying it, using strncpy to reapply the name.
This has several issues. strncpy should not be used
to copy memory in place. The second issue is that
the interface name is not actually changing when we
update interface to be in the new vrf.
Since every usage of if_update was just reapplying the same
name the interface actually had, just remove that part of
the function and rename it to if_update_to_new_vrf
to represent what it is actually doing.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The data for each nexthop is stored off of the info
pointer, instead of handling the first one and then looping
over the remaining, just loop over them all and handle the
first one as a special case.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
In the bgp code to send routes to zebra we were spending
a non trivial amount of time managing buffer sizes.
We know a priori the multipath supported by our system
so let's just use that value to appropriately size
the buffers.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
- All ipv4 labeled-unicast routes are now installed in the ipv4 unicast
table. This allows us to do things like take routes from an ipv4
unicast peer, allocate a label for them and TX them to a ipv4
labeled-unicast peer. We can do the opposite where we take routes from
a labeled-unicast peer, remove the label and advertise them to an ipv4
unicast peer.
- Multipath over a labeled route and non-labeled route is not allowed.
- You cannot activate a peer for both 'ipv4 unicast' and 'ipv4
labeled-unicast'
- The 'tag' variable was overloaded for zebra's route tag feature as
well as the mpls label. I added a 'mpls_label_t mpls' variable to
avoid this. This is much cleaner but resulted in touching a lot of
code.
This reverts commit fa14eb2c0b.
This was for stable/2.0 and wasn't intended to go on stable/3.0
-- my bad, missed this in the merge.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Problem reported that ecmp wasn't working correctly in a vrf with
ipv6. Issue was that originator of the routes were sending the updates
with a link-local nexthop and nhlen of 16. In this particular case,
bgp_zebra_announce was using the wrong call to get the ifindex and
was not supplying the vrf. This caused ecmp to work only in the case
of the default vrf.
Ticket: CM-15545
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Reviewed-by: CCR-6017
The FSF's address changed, and we had a mixture of comment styles for
the GPL file header. (The style with * at the beginning won out with
580 to 141 in existing files.)
Note: I've intentionally left intact other "variations" of the copyright
header, e.g. whether it says "Zebra", "Quagga", "FRR", or nothing.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Labeled-unicast updates were being sent with an ipv6 nexthop due to
not setting the mp_nexthop_len or nh_afi. On the receive side, the
prefix length was being incorrectly determined and has been fixed.
Also the stream for bgp_label_buf was not created. All resolved.
Ticket: CM-15260
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Reviewed-by:
Implement support for negotiating IPv4 or IPv6 labeled-unicast address
family, exchanging prefixes and installing them in the routing table, as
well as interactions with Zebra for FEC registration. This is the
implementation of RFC 3107.
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
