zebra: Nexthop tracking, route resolution recursive lookup

Description:
===========
Change is intended for fixing the NHT resolution logic.
While recursively resolving nexthop, keep looking for a valid/useable route in the rib,
by not stopping at the first/most-specific route in the rib.

Consider the following set of events taking place on R1:
R1(config)# ip route 2.2.2.0/24 ens192
R1# sharp watch nexthop 2.2.2.32 connected
R1# show ip nht
2.2.2.32(Connected)
 resolved via static
 is directly connected, ens192
 Client list: sharp(fd 33)

-2.2.2.32 NHT is resolved over the above valid static route.

R1# sharp install routes 2.2.2.32 nexthop 2.2.2.32 1
R1# 2.2.2.32(Connected)
 resolved via static
 is directly connected, ens192
 Client list: sharp(fd 33)

-.32/32 comes which is going to resolve through itself, but since this is an invalid route,
it will be marked as inactive and will not affect the NHT.

R1# sharp install routes 2.2.2.31 nexthop 2.2.2.32 1
R1# 2.2.2.32(Connected)
 unresolved(Connected)
 Client list: sharp(fd 50)

-Now a .31/32 comes which will resolve over .32 route, but as per the current logic,
this will trigger the NHT check, in turn making the NHT unresolved.

-With fix, NHT should stay in resolved state as long as the valid static or connected route stays installed

Fix:
====
-While resolving nexthops, walk up the tree from the most-specific match,
walk up the tree without any ZEBRA_NHT_CONNECTED check.

Co-authored-by: Vishal Dhingra <vdhingra@vmware.com>
Co-authored-by: Kantesh Mundaragi <kmundaragi@vmware.com>
Signed-off-by: Iqra Siddiqui <imujeebsiddi@vmware.com>

Merge pull request #10593 from donaldsharp/rip_cleanup

ripd: Cleanup ripd.c indentation levels and comments

Merge pull request #10625 from Jafaral/ospf-ns-vrf

topotests: give  a couple of ospf topotests better descriptive names

Merge pull request #10618 from louis-oui/topo-win-title

topotests: fix the usage of screen

topotests: set router name as title in screen and xterm

When running a topotest with the --shell or --vtysh argument, the
window titles of the routers are generic.

Set the router name as title to identify correctly the window.

Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>

topotests: fix usage of screen

Opening new tab in screen is not possible when using option --vtysh or
--shell. Error 'No such file or directory'.

Fix the issue.

Fixes: 6a5433ef0b ("tests: NEW micronet replacement for mininet")
Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>

tests: rename topotest ospf_topo1_vrf to ospf_netns_vrf

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

tests: rename topotest ospf_topo2 to ospf_unnumbered

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

Merge pull request #10539 from chiragshah6/evpn_dev1

bgpd: evpn route-map match ead type-1 route-type

Merge pull request #10604 from mjstapp/fix_staticd_nh_count

staticd: reject route config with too many nexthops

Merge pull request #10591 from donaldsharp/rip_packet_fix

ripd: Fix packet send for non primary addresses

doc: add description for match evpn route-type

Signed-off-by: Chirag Shah <chirag@nvidia.com>

bgpd: evpn route-map match esr type-4 route-type

Testing:

After fix:

bharat(config-route-map)# match evpn route-type
  1          EAD (Type-1) route
  2          MAC-IP (Type-2) route
  3          Multicast (Type-3) route
  4          Ethernet Segment (Type-4) route <----
  5          Prefix (Type-5) route
  ead        EAD (Type-1) route
  es         Ethernet Segment (Type-4) route
  macip      MAC-IP (Type-2) route
  multicast  Multicast (Type-3) route
  prefix     Prefix (Type-5) route

bharat(config-route-map)# match evpn route-type 4
bharat(config-route-map)# do show running-config
Building configuration...
...
!
route-map ALLOW_EVPN_R permit 10
 match evpn route-type es    <----
exit

BGP:
route-map: ALLOW_EVPN_R Invoked: 0 Optimization: enabled Processed Change: false
 permit, sequence 10 Invoked 0
  Match clauses:
    evpn route-type es  <-----
  Set clauses:
  Call clause:
  Action:
    Exit routemap

Signed-off-by: Chirag Shah <chirag@nvidia.com>

bgpd: evpn route-map match ead type-1 route-type

Add evpn mh route type-1 (EAD) to match clause
of route-map.

Ticket: issue#10461
Reviewed By:
Testing Done:

With fix:

vtep1(config-route-map)# match evpn route-type
  1          EAD (Type-1) route
  2          MAC-IP (Type-2) route
  3          Multicast (Type-3) route
  5          Prefix (Type-5) route
  ead        EAD (Type-1) route
  macip      MAC-IP (Type-2) route
  multicast  Multicast (Type-3) route
  prefix     Prefix (Type-5) route

vtep1# show running-config bgpd
....
route-map HOST_ALLOW_1 permit 1
 match evpn route-type ead

vtep1# show route-map HOST_ALLOW_1

BGP:
route-map: HOST_ALLOW_1 Invoked: 6 Optimization: disabled Processed Change: false
 permit, sequence 1 Invoked 6
  Match clauses:
    ip address prefix-list LOCAL_HOST_VRF1
    evpn route-type ead

Signed-off-by: Chirag Shah <chirag@nvidia.com>

staticd: reject route config with too many nexthops

Restrict the number of nexthops for a route to the compiled-in
limit. Be careful with the zapi route struct's array of nexthops
too.

Signed-off-by: Mark Stapp <mstapp@nvidia.com>

staticd: capture zebra's advertised ECMP limit

Capture the ECMP limit advertised by zebra (via zapi).

Signed-off-by: Mark Stapp <mstapp@nvidia.com>

staticd: fix spelling in an error message

whitespace change to fix a typo.

Signed-off-by: Mark Stapp <mstapp@nvidia.com>

Merge pull request #10557 from alexk99/zebra-fpm-multihop-weight

Zebra FPM: don't lose next hop weights while exporting via FPM

Merge pull request #10582 from anlancs/pimd-remove-redundant

pimd: remove redundant code

Merge pull request #8968 from donaldsharp/route_map_test

tests: Cleanup test_route_map_topo1.py from pylint

Merge pull request #10547 from donaldsharp/10458

zebra: Keep the interface flags safe on multiple ioctl calls

tests: Cleanup test_route_map_topo1.py from pylint

noticed that pylint was complaining about some easily
fixable stuff in test_route_map_topo1.py so let's clean
it up some.

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

Merge pull request #10558 from Jafaral/ospf-net-or-iface

doc: mixing `ip ospf` and `network` commands is not supported

Merge pull request #10598 from mobash-rasool/pim-compile-fix

pimd: Fix Compilation issue in PIM

Merge pull request #10561 from mjstapp/nlsock_hash_lock

zebra: make netlink object hash threadsafe

pimd: Fix Compilation issue in PIM

A recent merge caused this, fixing it.

Signed-off-by: Mobashshera Rasool <mrasool@gmail.com>

Merge pull request #10590 from donaldsharp/bgp_error_codes

Bgp error codes

Merge pull request #9224 from SaiGomathiN/saig

pimd: IGMP Query Generation

Merge pull request #10586 from punithkumar-spk/master

bgpd: uninitialized compiler warning

Merge pull request #10356 from opensourcerouting/pim6-adjust-20220117

pim6d: conversion work batch

ripd: Cleanup ripd.c indentation levels and comments

The entirety of this file is heavily indented.  Work some coding
structure to make it easier to read and understand and not be
so heavily indented.

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

Merge pull request #10575 from donaldsharp/bgp_requires_policy

bgp: Add a 6 hour warning to missing policy

ripd: Fix packet send for non primary addresses

When rip is configured to work on secondary addresses
on an interface, rip was not properly sending out
the packets on secondary addresses because the source of the
packet was never properly being setup and rip would
send the packet out multiple times for the primary address
not once for each address on the interface that is setup to work.

tcpdump + rip debugs output with fix:

2022/02/15 19:59:50 RIP: [ZG173-BHW0K] turn on virbr0
2022/02/15 19:59:51 RIP: [PYB7S-80D89] multicast join at virbr0
2022/02/15 19:59:51 RIP: [GZR24-FCQGG] multicast request on virbr0
2022/02/15 19:59:51 RIP: [JTNCV-XD8S1] rip_send_packet 192.168.122.1 > 224.0.0.9 (virbr0)
2022/02/15 19:59:51 RIP: [VEJY5-67P5X] SEND to  224.0.0.9520
2022/02/15 19:59:51 RIP: [JTNCV-XD8S1] rip_send_packet 73.3.3.8 > 224.0.0.9 (virbr0)
2022/02/15 19:59:51 RIP: [VEJY5-67P5X] SEND to  224.0.0.9520
19:59:51.831128 IP 192.168.122.1.route > rip2-routers.mcast.net.route: RIPv2, Request, length: 24
19:59:51.831161 IP c-73-3-3-8.hsd1.mo.comcast.net.route > rip2-routers.mcast.net.route: RIPv2, Request, length: 24

Fixes: #10588
Signed-off-by: Donald Sharp <sharpd@nvidia.com>

bgpd: Fix uninitialized compiler warning

Signed-off-by: Punith Kumar <punithkumar.spk@gmail.com>

bgpd: Renumber bgp_create_error_code enum values

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

bgpd: Convert bgp error codes for cli input to an enum

Conversion of bgp error codes returned for cli input into
an enum and then properly handling all the error cases
in bgp_vty_return.

Because not all error codes returned were properly handled
in this function there existed configuration examples that
were accepted on the cli without an error message but not
saved.

Fixes: #10589
Signed-off-by: Donald Sharp <sharpd@nvidia.com>

bgpd: Remove unused BGP_ERR_MAX #define

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

bgpd: Move some error codes to bgp_vty_return handling

BGP_ERR_PEER_GROUP_MEMBER and BGP_ERR_PEER_GROUP_PEER_TYPE_DIFFERENT
both are not handled by bgp_vty_return, but both can be handled by
this function as that there is nothing special going on here.

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

bgpd: Remove impossible invalid state

confederations are checking to see that the bgp pointer
is non-null.  But it's impossible to have a null pointer
in the cli and in all paths we have already deref'ed the bgp
pointer.  Let's remove that error code as that it is impossible
to happen.

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

Merge pull request #10571 from rameshabhinay/ospf6_auth_trailer

ospf6d: fix coverity issues.

Merge pull request #10546 from ton31337/fix/check_for_null_inside_unintern

bgpd: Check for NULL inside aspath_unintern()

Merge pull request #10555 from anlancs/doc-comment

doc: add "cost" to ospfd "range" command

Merge pull request #10562 from donaldsharp/starv_warn

lib: Save number of times a thread is starved

bgp: Add a 15 minute warning to missing policy

Add a 15 minute warning to the logging system when
bgp policy is not setup properly.  Operators keep asking
about the missing policy( on upgrade typically ).  Let's
try to give them a bit more of a hint when something is
going wrong as that they are clearly missing the other
various places FRR tells them about it.

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

Merge pull request #10573 from ton31337/fix/doc_rc_tag

doc: An example how we do RC tagging for stabilization branch

pimd: remove redundant code

`ifchannel_rb` is already cleaned by `pim_ifchannel_delete_all`, so
it should be removed.

Signed-off-by: anlan_cs <vic.lan@pica8.com>

Merge pull request #10577 from donaldsharp/speeling

Speeling

zebra: Fix spelling mistake

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

yang: Fix spelling mistake

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

vtysh: Fix spelling mistakes

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

tests: Fix spelling and grammar mistakes

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

pceplib: Fix spelling mistakes

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

pathd: Fix spelling mistakes

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

ospfd: Fix spelling mistakes

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

ospf6d: Fix spelling mistakes

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

doc: Fix spelling mistakes

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

bgpd: Fix spelling mistakes

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

ospf6d: fix coverity issues.

Fixed below coverity issues
________________________________________________________________________________________________________
*** CID 1511366:    (TAINTED_SCALAR)
/ospf6d/ospf6_message.c: 2631 in ospf6_make_lsupdate_list()
2625                          + OSPF6_HEADER_SIZE)
2626                         > ospf6_packet_max(on->ospf6_if)) {
2627                             ospf6_fill_header(on->ospf6_if, (*op)->s,
2628                                               length + OSPF6_HEADER_SIZE);
2629                             (*op)->length = length + OSPF6_HEADER_SIZE;
2630                             ospf6_fill_lsupdate_header((*op)->s, *lsa_cnt);
>>>     CID 1511366:    (TAINTED_SCALAR)
>>>     Passing tainted variable "(*op)->length" to a tainted sink.
2631                             ospf6_send_lsupdate(on, NULL, *op);
2632
2633                             /* refresh packet */
2634                             *op = ospf6_packet_new(on->ospf6_if->ifmtu);
2635                             length = OSPF6_LS_UPD_MIN_SIZE;
2636                             *lsa_cnt = 0;
/ospf6d/ospf6_message.c: 2631 in ospf6_make_lsupdate_list()
2625                          + OSPF6_HEADER_SIZE)
2626                         > ospf6_packet_max(on->ospf6_if)) {
2627                             ospf6_fill_header(on->ospf6_if, (*op)->s,
2628                                               length + OSPF6_HEADER_SIZE);
2629                             (*op)->length = length + OSPF6_HEADER_SIZE;
2630                             ospf6_fill_lsupdate_header((*op)->s, *lsa_cnt);
>>>     CID 1511366:    (TAINTED_SCALAR)
>>>     Passing tainted variable "(*op)->length" to a tainted sink.
2631                             ospf6_send_lsupdate(on, NULL, *op);

________________________________________________________________________________________________________
*** CID 1511365:    (TAINTED_SCALAR)
/ospf6d/ospf6_message.c: 2674 in ospf6_make_ls_retrans_list()
2669                             if (on->ospf6_if->state == OSPF6_INTERFACE_POINTTOPOINT)
2670                                     (*op)->dst = allspfrouters6;
2671                             else
2672                                     (*op)->dst = on->linklocal_addr;
2673
>>>     CID 1511365:    (TAINTED_SCALAR)
>>>     Passing tainted variable "(*op)->length" to a tainted sink.
2674                             ospf6_fill_hdr_checksum(on->ospf6_if, *op);
2675                             ospf6_packet_add(on->ospf6_if, *op);
2676                             OSPF6_MESSAGE_WRITE_ON(on->ospf6_if);
/ospf6d/ospf6_message.c: 2674 in ospf6_make_ls_retrans_list()
2669                             if (on->ospf6_if->state == OSPF6_INTERFACE_POINTTOPOINT)
2670                                     (*op)->dst = allspfrouters6;
2671                             else
2672                                     (*op)->dst = on->linklocal_addr;
2673
>>>     CID 1511365:    (TAINTED_SCALAR)
>>>     Passing tainted variable "(*op)->length" to a tainted sink.
2674                             ospf6_fill_hdr_checksum(on->ospf6_if, *op);
2675                             ospf6_packet_add(on->ospf6_if, *op);
2676                             OSPF6_MESSAGE_WRITE_ON(on->ospf6_if);
/ospf6d/ospf6_message.c: 2674 in ospf6_make_ls_retrans_list()
2668                             ospf6_fill_lsupdate_header((*op)->s, *lsa_cnt);
2669                             if (on->ospf6_if->state == OSPF6_INTERFACE_POINTTOPOINT)
2670                                     (*op)->dst = allspfrouters6;
2671                             else
2672                                     (*op)->dst = on->linklocal_addr;
2673
>>>     CID 1511365:    (TAINTED_SCALAR)
>>>     Passing tainted variable "(*op)->length" to a tainted sink.
2674                             ospf6_fill_hdr_checksum(on->ospf6_if, *op);
2675                             ospf6_packet_add(on->ospf6_if, *op);
2676                             OSPF6_MESSAGE_WRITE_ON(on->ospf6_if);

________________________________________________________________________________________________________
*** CID 1511364:  Insecure data handling  (TAINTED_SCALAR)
/ospf6d/ospf6_message.c: 2125 in ospf6_write()
2120                     if (oi->at_data.flags != 0) {
2121                             at_len = ospf6_auth_len_get(oi);
2122                             if (at_len) {
2123                                     iovector[0].iov_len =
2124                                             ntohs(oh->length) + at_len;
>>>     CID 1511364:  Insecure data handling  (TAINTED_SCALAR)
>>>     Passing tainted variable "iovector[0].iov_len" to a tainted sink.
2125                                     ospf6_auth_digest_send(oi->linklocal_addr, oi,
2126                                                            oh, at_len,
2127                                                            iovector[0].iov_len);
2128                             } else {
2129                                     iovector[0].iov_len = ntohs(oh->length);
2130                             }

________________________________________________________________________________________________________
*** CID 1511363:    (DEADCODE)
/ospf6d/ospf6_auth_trailer.c: 275 in ospf6_hash_hmac_sha_digest()
269      case KEYCHAIN_ALGO_HMAC_SHA512:
270     #ifdef CRYPTO_OPENSSL
271              sha512_digest(mes, len, digest);
272     #endif
273              break;
274      case KEYCHAIN_ALGO_NULL:
>>>     CID 1511363:    (DEADCODE)
>>>     Execution cannot reach this statement: "case KEYCHAIN_ALGO_MAX:".
275      case KEYCHAIN_ALGO_MAX:
276      default:

/ospf6d/ospf6_auth_trailer.c: 274 in ospf6_hash_hmac_sha_digest()
269      case KEYCHAIN_ALGO_HMAC_SHA512:
270     #ifdef CRYPTO_OPENSSL
271              sha512_digest(mes, len, digest);
272     #endif
273              break;
>>>     CID 1511363:    (DEADCODE)
>>>     Execution cannot reach this statement: "case KEYCHAIN_ALGO_NULL:".
274      case KEYCHAIN_ALGO_NULL:
275      case KEYCHAIN_ALGO_MAX:
276      default:

________________________________________________________________________________________________________
*** CID 1511362:  Insecure data handling  (TAINTED_SCALAR)
/ospf6d/ospf6_auth_trailer.c: 541 in ospf6_auth_check_digest()
535
536      auth_len = ntohs(ospf6_auth->length);
537
538      memcpy(temp_hash, ospf6_auth->data, hash_len);
539      memcpy(ospf6_auth->data, apad, hash_len);
540
>>>     CID 1511362:  Insecure data handling  (TAINTED_SCALAR)
>>>     Passing tainted variable "oh_len + auth_len + lls_block_len" to a tainted sink.
541      ospf6_auth_update_digest(oi, oh, ospf6_auth, auth_str,
542                               (oh_len + auth_len + lls_block_len),
543                               hash_algo);

________________________________________________________________________________________________________
*** CID 1511361:  Insecure data handling  (TAINTED_SCALAR)
/ospf6d/ospf6_auth_trailer.c: 124 in ospf6_auth_hdr_dump_recv()
118      at_len = length - (oh_len + lls_len);
119      if (at_len > 0) {
120              ospf6_at_hdr =
121                      (struct ospf6_auth_hdr *)((uint8_t *)ospfh + oh_len);
122              at_hdr_len = ntohs(ospf6_at_hdr->length);
123              hash_len = at_hdr_len - OSPF6_AUTH_HDR_MIN_SIZE;
>>>     CID 1511361:  Insecure data handling  (TAINTED_SCALAR)
>>>     Passing tainted variable "hash_len" to a tainted sink.
124              memcpy(temp, ospf6_at_hdr->data, hash_len);
125              temp[hash_len] = '\0';

________________________________________________________________________________________________________
*** CID 1482146:  Insecure data handling  (TAINTED_SCALAR)
/ospf6d/ospf6_message.c: 2787 in ospf6_lsupdate_send_neighbor_now()
2781
2782             if (IS_OSPF6_DEBUG_FLOODING
2783                 || IS_OSPF6_DEBUG_MESSAGE(OSPF6_MESSAGE_TYPE_LSUPDATE, SEND_HDR))
2784                     zlog_debug("%s: Send lsupdate with lsa %s (age %u)", __func__,
2785                                lsa->name, ntohs(lsa->header->age));
2786
>>>     CID 1482146:  Insecure data handling  (TAINTED_SCALAR)
>>>     Passing tainted variable "op->length" to a tainted sink.
2787             ospf6_send_lsupdate(on, NULL, op);

Signed-off-by: Abhinay Ramesh <rabhinay@vmware.com>

pim6d: IPv6-adjust pim_msg_send() and related

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: IPv6-adjust pim_ifchannel_*

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: IPv6-adjust assert-related addrs

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: IPv6-adjust jp_agg->group

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: IPv6-adjust iface primary/DR addrs

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: IPv6-adjust pim_upstream addr

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: IPv6-adjust Hello/TLV processing

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: IPv6-adjust neigh->source_addr

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: IPv6-adjust various pim_sgaddr uses

Since `pim_sgaddr` is `pim_addr` now, that causes a whole lot of fallout
anywhere S,G pairs are handled.

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: prepare SSM/filter functions

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: prepare/stub pim_rp_* functions for IPv6

These are sprinkled relatively widely through the PIM codebase, so for
the time being reduce the "compiler warning surface" by moving them
forward to proper types without actual implementations.

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: more TLV parse/encode IPv6 preparation

More proliferation of pim_addr to work towards IPV6.

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: prepare IPv6 address encoding functions

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

pim6d: convert address comparison in I_am_DR macro

Signed-off-by: David Lamparter <equinox@opensourcerouting.org>

doc: An example how we do RC tagging for stabilization branch

Signed-off-by: Donatas Abraitis <donatas.abraitis@gmail.com>

zebra: make netlink object hash threadsafe

The recently-added hashtable of nlsock objects needs to be
thread-safe: it's accessed from the main and dplane pthreads.
Add a mutex for it, use wrapper apis when accessing it. Add
a per-OS init/terminate api so we can do init that's not
per-vrf or per-namespace.

Signed-off-by: Mark Stapp <mstapp@nvidia.com>

Merge pull request #10533 from taspelund/ip_rule_nl_debugs

zebra: add netlink debugs for ip rules

Merge pull request #9697 from SaiGomathiN/igmp-sources

pimd: json support added

zebra: cleanup multiline strings in debug_nl.c

NetDEF CI has been whining about multiline string style.
Make the strings single-line and call it a day.

Signed-off-by: Trey Aspelund <taspelund@nvidia.com>

zebra: add netlink debugs for ip rules

Adds functions to parse + decode netlink rules.
Adds RTM_NEWRULE + RTM_DELRULE to "debug zebra kernel".

Signed-off-by: Trey Aspelund <taspelund@nvidia.com>

Merge pull request #8890 from rameshabhinay/ospf6_auth_trailer

OSPF6: Support OSPF6 authentication trailer RFC 7166

lib: Save number of times a thread is starved

Add a counter to the number of times a thread is starved from
a timer event and add the output to `show thread cpu`

Signed-off-by: Donald Sharp <sharpd@nvidia.com>

doc: mixing `ip ospf` and `network` commands is not supported

Signed-off-by: Jafar Al-Gharaibeh <jafar@atcorp.com>

zebra: FPM next hop weights

Don't lose next hop weights while exporting via FPM

Signed-off-by: Alex Kiselev <alex@bisonrouter.com>

Merge pull request #10537 from mjstapp/fix_dplane_strdup

zebra: use frr mem apis in dplane

Merge pull request #10552 from idryzhov/frr-reload-nodes

tools: fix frr-reload context keywords

Merge pull request #10540 from idryzhov/attr-extra-revert

revert recent bgp attr->extra changes

Merge pull request #10545 from ton31337/feature/get_set_for_lcommunity

bgpd: Use get/set helpers for attr->lcommunity

bgpd: Use get/set helpers for attr->lcommunity

Signed-off-by: Donatas Abraitis <donatas.abraitis@gmail.com>

Revert "bgpd: Move out ipv6_ecommunity struct from attr to attr_extra"

This reverts commit 2703b7db19cebad82772d5210132dab412b855c3.

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

Revert "bgpd: Move attr->pmsi_tnl_type to attr->extra->pmsi_tnl_type"

This reverts commit fc6ba64f04696416a1216bb57a7fd72efc5a0904.

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

Revert "bgpd: Free only subattributes, not the whole attr_extra pointer"

This reverts commit 0911206097dc04315bcd6b3accd6236a047a1e06.

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

Merge pull request #10549 from idryzhov/bgp-coverity

bgpd: fix null pointer dereferences

Merge pull request #10550 from idryzhov/bgp-attr-undup

bgpd: remove bgp_attr_undup

Merge pull request #10551 from idryzhov/bgpd-memleaks

bgpd: fix a couple of memleaks

tools: fix frr-reload context keywords

There are singline-line commands inside `router bgp` that start with
`vnc ` or `bmp `. Those commands are currently treated as node-entering
commands. We need to specify such commands more precisely.

Fixes #10548.

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

bgpd: remove bgp_attr_undup

bgp_attr_undup does the same thing as bgp_attr_flush – frees the
temporary data that might be allocated when applying a route-map. There
is no need to have two separate functions for that.

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

bgpd: fix aspath memleak on error in vnc_direct_bgp_add_nve

bgp_attr_default_set creates a new empty aspath. If family error happens,
this aspath is not freed. Move attr initialization after we checked the
family.

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

bgpd: fix aspath memory leak in aggr_suppress_map_test

aspth_empty a couple of lines earlier creates an aspath and it must be
freed.

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

bgpd: fix missing bgp_attr_flush on errors in bgp_update

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

bgpd: fix null pointer dereferences

They were introduced during recent ecommunity rework in b53e67a3.

CID 1511347 and 1511348.

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>

Merge pull request #10482 from donaldsharp/zebra_buffering

Zebra buffering