Christian Hopps [Fri, 17 Jun 2022 06:04:51 +0000 (02:04 -0400)]
lib: cleanup red-herring memleaks in parent of daemonizing fork
- The parent of the daemonizing fork reports memleaks for the early
northbound allocations (libyang). If these were real memleaks these
would show up in the child as well; however, ignoring all memleaks in
the parent of the fork is too hard a sale. Instead, spend some CPU
cycles cleaning up the allocations in the parent after the fork and
immeidatley prior to exiting the parent after the daemonizing fork.
bgpd: Fix memory leak when setting [l]community at egress
```
==2209758== 7,791,480 (399,840 direct, 7,391,640 indirect) bytes in 9,996 blocks are definitely lost in loss record 102 of 103
==2209758== at 0x4C33B25: calloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==2209758== by 0x4EE264F: qcalloc (memory.c:116)
==2209758== by 0x22E62A: lcommunity_new (bgp_lcommunity.c:42)
==2209758== by 0x22E62A: lcommunity_dup (bgp_lcommunity.c:155)
==2209758== by 0x26F28F: route_set_lcommunity (bgp_routemap.c:2382)
==2209758== by 0x4EFF7EF: route_map_apply_ext (routemap.c:2663)
==2209758== by 0x250B3D: subgroup_announce_check (bgp_route.c:2202)
==2209758== by 0x27A195: subgroup_announce_table (bgp_updgrp_adv.c:690)
==2209758== by 0x27A509: subgroup_coalesce_timer (bgp_updgrp_adv.c:332)
==2209758== by 0x4F1C3FC: thread_call (thread.c:2002)
==2209758== by 0x4ED6D67: frr_run (libfrr.c:1196)
==2209758== by 0x1E921B: main (bgp_main.c:519)
```
To reproduce it's enough to have something like applied at egress:
On the other side doing:
```
spine1-debian-11# sharp install routes 100.100.100.1 nexthop 192.168.10.65 30
spine1-debian-11# sharp remove routes 100.100.100.1 30
spine1-debian-11# sharp install routes 100.100.100.1 nexthop 192.168.10.65 30
```
Before:
```
Large Community : 66 40 2816 66 2816
Large Community value : 66 variable 29728 66 29728
Large Community : 96 40 4096 96 4096
Large Community value : 96 variable 52048 96 52048
```
After:
```
Large Community : 28 40 1152 29 1192
Large Community value : 28 variable 1456 29 2200
Large Community : 28 40 1152 29 1192
Large Community value : 28 variable 1456 29 2200
```
bgpd: Withdraw implicitly old paths from VRFs when import/export list changes
If we overwrite import/export RT list via route-maps or even flush by using
`set extcommunity none`, then we must withdraw old paths from VRFs to avoid
stale paths.
lib: Update sysrepo code with the latest API changes
* sr_event_notif_send -> sr_notif_send
* sr_process_events -> sr_subscription_process_events
* sr_oper_get_items_subscribe -> sr_oper_get_subscribe
* Removed SR_SUBSCR_CTX_REUSE flag from the code at all
Donald Sharp [Wed, 8 Jun 2022 13:01:31 +0000 (09:01 -0400)]
babeld: Ignore sub-tlv's with mandatory bit set
RFC 8966 4.4
If the mandatory bit is set, then the
whole enclosing TLV MUST be silently ignored (except for updating the
parser state by a Router-Id, Next Hop, or Update TLV, as described in
the next section).
Donald Sharp [Wed, 8 Jun 2022 12:41:05 +0000 (08:41 -0400)]
babeld: Update to ignore unicast hellos
RFC 8966 requires a bit of work to make FRR forward compabitble with
it for the moment.
tl;dr -> Flags of hellos should be 0 except the main bit and
The Unicast Hello bit if set should cause the packet to be ignored.
Additionally FRR needs to ignore Hello packets with the interval field
set to 0.
Donald Sharp [Wed, 8 Jun 2022 20:51:48 +0000 (16:51 -0400)]
pimd: Fix pim_channel_oil_empty tests
The pim_channel_oil_empty() function was setting
the pimreg if it ever existed for NULL comparison
but of course the pimreg device is never pulled back
out again when it was needed to be when the pimreg
is not present.
Donald Sharp [Thu, 9 Jun 2022 14:29:04 +0000 (10:29 -0400)]
pimd: Show interface traffic even if interface is currently `down`
the `show ip pim interface [x] traffic` command was deciding
to skip display of interfaces if they happened to be down at
that moment. This of course does not make a bunch of sense
to limit the output for a interface that may have sent data
in the past.
This fixes this test crash:
rnode = <lib.topogen.TopoRouter object at 0x7fc755be3880>, dut = 'c1', input_dict = {'c1': {'c1-l1-eth2': ['helloTx', 'helloRx']}}, output_dict = {'c1': {}}
Donald Sharp [Wed, 8 Jun 2022 15:33:32 +0000 (11:33 -0400)]
pimd: pim_addr_cmp returns 0 for same not -1 or 1
The test for setting pimDesignatedRouterLocal was comparing
the result of pim_addr_cmp to non zero to say that the
pim_dr_addr and primary_address are the same. This is
not correct in that pim_addr_cmp returns 0 for being the
same, so the correct test is !pim_addr_cmp
pimd: Handle receive of (*,G) register stop with src addr as 0
PIM conformance test case 11.12 sends a register stop msg with
source address as 0. This should trigger a register stop for
all the upstreams. It is not happening as such because we are not
considering 0.0.0.0 source address for starg_handling.
Christian Hopps [Sat, 8 Jan 2022 21:57:10 +0000 (16:57 -0500)]
ospfd: api: fix recovery of LSA after restart of api client
Prior to this fix, restarting the client just failed b/c the code tried to
"refresh" the existing LSA being added, except that code checked for meta-data
to exist, which was deleted when the client disconnected previously (or had
never connected and the LSA state was picked up from the network).
Christian Hopps [Wed, 1 Jun 2022 19:25:35 +0000 (15:25 -0400)]
ospfd: api: add reachable router notifications
Reachable router information is used by OSPF opaque clients in order
to determine if the router advertising the opaque LSA data is
reachable (i.e., 2-way conectivity check).
Christian Hopps [Sat, 15 Jan 2022 11:13:43 +0000 (06:13 -0500)]
ospfd: add all_rtrs route table when opaque enabled
The reachable router table is used by OSPF opaque clients in order to
determine if the router advertising the opaque LSA data is
reachable (i.e., 2-way conectivity check).
isisd: replace ipv4-unicast with standard in config and oper contexts
Only the multi-topology command can use 'ipv4-unicast' keyword to
configure standard topology.
The remaining code: dump from show commands, and yang definition,
uses 'standard' keyword instead.
The test have not been modified. The change would consists in
modifying test_fuzz_isis_tlv_tests.h.gz:
- replacing ipv4-unicast occurences with standard
\x69\x70\x76\x34\x2d\x75\x6e\x69\x63\x61\x73\x74
with
\x73\x74\x61\x6e\x64\x61\x72\x64
- align the buffer length by removing 4 bytes per occurence
Instead, a specific isis_mtid2str_fake() routing has been
put in place in isis_tlvs.c file.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
David Lamparter [Wed, 1 Jun 2022 15:10:11 +0000 (17:10 +0200)]
ospf6d: fix rx/tx buffer sizes
OSPFv3 packets can be fragmented and up to 64k long, regardless of
interface MTU. Trying to size these buffers to MTU is just plain wrong.
To not make this a super intrusive change during the 8.3 release freeze,
just code this into ospf6_iobuf_size().
Since the buffer is now always 64k, don't waste time zeroing the entire
thing in receive; instead just zero kind of a "sled" of 128 bytes after
the buffer as a security precaution.
Fixes: #11298 Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Philippe Guibert [Wed, 25 May 2022 09:47:36 +0000 (11:47 +0200)]
isisd: add ISIS_MT_STANDARD define
This define is added to reflect more what the standard
topology means. Actually, the standard topology may
contain not only IPv4 protocol but also IPv6 protocol.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Martin Buck [Wed, 1 Jun 2022 09:58:52 +0000 (11:58 +0200)]
tools: Really fix all daemon shutdown without complaining about PID files
Fix daemon shutdown broken by f0cccaa6bf7dda71d0e1ad1b0406a9b453a84bb3. Now
we still don't complain about missing PID files but actually stop the
running daemons.
The previous fix was broken because it passed a new "--all" option to
daemon_stop which wasn't handled properly (it assumed $1 contains the
daemon name when at that time it acutally contained the "--all" option).
Plus, "--all" wasn't actually necessary, because we already passed
"--reallyall" from all_stop to daemon_stop after the daemon name.
So remove "--all" again and simply check for "--reallyall" in $2. This
should *really* fix #11317.
Signed-off-by: Martin Buck <mb-tmp-tvguho.pbz@gromit.dyndns.org>
Donald Sharp [Tue, 31 May 2022 16:19:27 +0000 (12:19 -0400)]
bgpd: Add ability to match `match peer <PEERGROUPNAME>`
Allow the end operator to match `match peer <PEERGROUPNAME>`
in a route-map. If the end operator defines interface
based peers and peer-groups that have names that overlap
the interface based peer will match first.
Fixes: #11316 Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Martin Buck [Tue, 31 May 2022 12:16:34 +0000 (14:16 +0200)]
tools: Don't complain about missing PID files when stopping all damons
Since 1686b1d4864384ee508b58418f20970f2dff2587, we try to stop all daemons,
even those which are not (no longer) enabled in /etc/frr/daemons. But we
shouldn't complain about missing PID files for daemons which have never been
started and just silently ignore those.
Signed-off-by: Martin Buck <mb-tmp-tvguho.pbz@gromit.dyndns.org>
Trey Aspelund [Sat, 28 May 2022 07:10:51 +0000 (03:10 -0400)]
zebra: Update advertise-svi-ip MACIPs w/ new MAC
When the kernel was sending an RTM_NEWLINK updating the MAC of a known
SVI, Type-2 routes created by advertise-svi-ip were not getting updated
with the new address.
This adds removal of any old Type-2 routes (with old MAC) and creation
of new Type-2 routes (with new MAC) into RTM_NEWLINK processing.
Donald Sharp [Tue, 24 May 2022 17:33:35 +0000 (13:33 -0400)]
pimd: Allow the LPM match work properly with prefix lists and normal rp's
All rp_info's are being saved in the pim->rp_list and the non prefix-list
rp's are supposed to be saved in table pim->rp_table. What was happening,
though, is that all the plists were being stored at the 224.0.0.0/4 node
of the tree(irrelevant to the fact that we should not be looking up the
non-prefix list rp's in the table) and since we can have multiple prefix lists
and only one rp_info stored at the 224.0.0.0/4 node of the tree, there can be situations
where the 224.0.0.0/4 node can be overwritten due to the order entered.
As such there exists situations where command enter ordering will cause
what we match to, change in pim_rp_find_match_group.
Fixes:
a) Do not store prefix list based rp_info's in the pim->rp_table
b) In pim_rp_find_match_group, ensure that the node lookup does
not correspond to a prefix list based node.
c) When in the situation there are both:
ip pim rp 4.5.6.7 224.0.0.0/16
ip pim rp 5.6.67.8 prefix-list FOO
ip prefix-list FOO permit 224.0.1.0/24
and we receive a group for 224.0.1.5, we were comparing the
224.0.0.0/16 to the 224.0.0.0/4 of the 5.6.67.8 group, when
FRR should have been comparing to entry that matched in the prefix-list
Donatas Abraitis [Thu, 26 May 2022 11:03:02 +0000 (14:03 +0300)]
bgpd: Fix memory leak for BGP community alias in CLI
Before:
```
root@spine1-debian-11:~/frr# vtysh -c 'show memory bgpd | include Large Community'
Large Community : 100 40 4000 100 4000
Large Community value : 100 12 2400 100 2400
root@spine1-debian-11:~/frr# for x in $(seq 1 100); do vtysh -c 'conf' -c 'bgp community alias 123:123:123 testas' > /dev/null; done
root@spine1-debian-11:~/frr# vtysh -c 'show memory bgpd | include Large Community'
Large Community : 200 40 8000 200 8000
Large Community value : 200 12 4800 200 4800
root@spine1-debian-11:~/frr# for x in $(seq 1 100); do vtysh -c 'conf' -c 'bgp community alias 123:123:123 testas' > /dev/null; done
root@spine1-debian-11:~/frr# vtysh -c 'show memory bgpd | include Large Community'
Large Community : 300 40 12000 300 12000
Large Community value : 300 12 7200 300 7200
root@spine1-debian-11:~/frr#
```
After:
```
root@spine1-debian-11:~/frr# vtysh -c 'show memory bgpd | include Large Community'
Large Community : 0 40 0 1 56
Large Community display string: 0 8192 0 1 8200
Large Community value : 0 12 0 1 24
root@spine1-debian-11:~/frr# for x in $(seq 1 100); do vtysh -c 'conf' -c 'bgp community alias 123:123:123 testas' > /dev/null; done
root@spine1-debian-11:~/frr# vtysh -c 'show memory bgpd | include Large Community'
Large Community : 0 40 0 1 56
Large Community display string: 0 8192 0 1 8200
Large Community value : 0 12 0 1 24
root@spine1-debian-11:~/frr#
```
After we call [l]community_str2com(), we should free the memory.
Donatas Abraitis [Wed, 25 May 2022 16:07:40 +0000 (19:07 +0300)]
github: Use pull_request_target as a target
And drop checkout action - not needed.
Due to the dangers inherent to automatic processing of PRs, GitHub’s standard
pull_request workflow trigger by default prevents write permissions and
secrets access to the target repository. However, in some scenarios such
access is needed to properly process the PR.
To this end the pull_request_target workflow trigger was introduced.
isisd: Stop fulfill MPLS table when SR is disabled
When Segment Routing is disabled, if isisd received LSP with Segment Routing
information, in particular prefix SIDs, it installs corresponding MPLS entries
while it should not as SR is disabled.
This patch adds extra control to verify if SR is enabled or not before
configuring MPLS LFIB & IP FIB with prefix SIDs and adjust SR & TI-LFA
tests accordingly.
projects / matthieu / frr.git / log