Christian Hopps [Wed, 19 Apr 2023 11:57:42 +0000 (07:57 -0400)]
tests: restore per-node logging functionality
After the munet switch we weren't passing the logger on to low-level
LinuxNamespace and thus Commander parent classes, so the lowest-level
`cmd_status` logs were missing from more specific log files in the run
directory.
Christian Hopps [Sat, 22 Apr 2023 01:59:33 +0000 (21:59 -0400)]
tests: fix check for daemon startup, remove sleep hack
- Remove the .pid and .vty files and then wait for them to show back up.
- Fix broken BGP GR test to not fail now that it's bug is exposed. It
only worked b/c when starting a daemon the pid file still existed and
blocked the bogus second BGP launch from happening.
David Lamparter [Fri, 21 Apr 2023 14:23:39 +0000 (16:23 +0200)]
lib: fix _hash_member() crash w/ empty hash
The typesafe hash _member() didn't check tabshift/count before
proceeding to look at the hash table, leading it to dereference a NULL
pointer when the hash table is in fact empty.
Test case added to tests/lib/test_typelist.
Note this function is not currently used anywhere. Only lib/cspf.c uses
_member(), but it does so on a RB-tree rather than a hash.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
David Lamparter [Fri, 21 Apr 2023 13:14:43 +0000 (15:14 +0200)]
pimd: harden MLD code loop boundaries
Coverity complains about these being tainted/untrusted loop boundaries.
The way the code works, it's counting up groups/sources, but keeps
checking against remaining data length in the packet - which is
perfectly fine IMHO. Except Coverity doesn't understand it :(
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Donald Sharp [Fri, 21 Apr 2023 12:54:51 +0000 (08:54 -0400)]
zebra: ctx has to be non NULL at this point
Remove the pointer check for ctx. At this point in the
function it has to be non null since we deref'ed it.
Additionally the alloc function that creates it cannot
fail.
Donald Sharp [Fri, 21 Apr 2023 11:57:19 +0000 (07:57 -0400)]
bgpd: Coverity mistakenly believes we can fall through
Coverity has incorrectly found that a switch statement
can be fallen through. It is missing the fact that the
previous case statement has a return through every path
it can get too. Let's just appease coverity
Donald Sharp [Thu, 20 Apr 2023 20:27:20 +0000 (16:27 -0400)]
bgpd: Fix lcom->str string length to correctly cover aliases
If you have a very large number of large communities whose
string length happened to be greater than BUFSIZ FRR's bgpd
would crash. This is because bgpd would write beyond
the end of the string.
Originally the code auto-calculated the string size appropriately
but commit ed0e57e3f079352714c3a3a8a5b0dddf4aadfe1d modified
the string length to be a hard coded BUFSIZ. When a route-map
like this is added:
Modify the code to correctly determine the string length of the communities
and to also double check if the string has an alias and ensure that the
string is still sufficiently large enough. If not auto size it again.
bgpd: Fix for ain->attr corruption during path update
1. Consider a established L2VPN EVPN BGP peer with soft-reconfiguartion
inbound configured
2. When the interface of this directly connected BGP peer is shutdown,
bgp_soft_reconfig_table_update() is called, which memsets the evpn buffer
and calls bgp_update() with received attributes stored in ain table(ain->attr).
In bgp_update(), evpn_overlay attribute in ain->attr (which is an interned
attr) was modified by doing a memcpy
3. Above action causes 2 attributes in the attrhash (which were previously different)
to match!
4. Later during fsm change event of the peer, bgp_adj_in_remove() is called
to clean up the ain->attr. But, because 2 attrs in attrhash match, it causes
BGP to assert in bgp_attr_unintern()
Before:
```
Routing Protocol is "rip"
Sending updates every 5 seconds with +/-50%, next due in 0 seconds
Timeout after 15 seconds, garbage collect after 10 seconds
Outgoing update filter list for all interface is not set
Incoming update filter list for all interface is not set
Default redistribution metric is 1
Redistributing:
Default version control: send version 2, receive any version
Interface Send Recv Key-chain
r1-eth0 2 1 2
Routing for Networks:
192.168.1.0/24
Routing Information Sources:
Gateway BadPackets BadRoutes Distance Last Update
192.168.1.2 0 0 120 00:00:05
192.168.1.3 0 0 120 00:00:04
Distance: (default is 120)
```
After:
```
Routing Protocol is "rip"
Sending updates every 5 seconds with +/-50%, next due in 4 seconds
Timeout after 15 seconds, garbage collect after 10 seconds
Outgoing update filter list for all interface is not set
Incoming update filter list for all interface is not set
Default redistribution metric is 1
Redistributing:
Default version control: send version 2, receive any version
Interface Send Recv Key-chain
r1-eth0 2 1 2
Routing for Networks:
192.168.1.0/24
Routing Information Sources:
Gateway BadPackets BadRoutes Distance Last Update
192.168.1.2 0 0 120 00:00:00
192.168.1.3 0 0 120 00:00:04
Distance: (default is 120)
```
Eric Kinzie [Fri, 2 Sep 2022 14:33:21 +0000 (16:33 +0200)]
lib: convert termtable to json
Add a function that returns a JSON-C structure containing a representation
of a termtable. This is intended to be a quick way to implement JSON
output to CLI commands.
Louis Scalbert [Thu, 10 Nov 2022 15:36:08 +0000 (16:36 +0100)]
isisd: add affinity-map configuration hooks
Add affinity-map hooks to check the utilization of affinity-map in
flex-algo contexts before its deletion and to update local TLVs when the
affinity-map bit-position is updated.
Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
Eric Kinzie [Fri, 1 Apr 2022 00:47:30 +0000 (00:47 +0000)]
isisd: fix merging flex-algo route nodes
As isis_route_verify_merge() builds a new route table, each (src,dst)
new node in the table is created by merging data from the same (src,dst)
pair present in all route tables. However, as it iterates over those
route nodes and copies SR data into the newly created node, it repeatedly
overwrites the isis_route_info pointer. This, in turn, pollutes the
isis_route_info structures in all of the other tables with partially
updated SR data.
After a route is deleted, this leaves pointers to freed memory in one
of the route trees.
Fixes: 26b0598f6b6f ("isisd: fix isis_route_merge_verify logic") Signed-off-by: Eric Kinzie <ekinzie@labn.net>
Louis Scalbert [Mon, 12 Dec 2022 15:00:58 +0000 (16:00 +0100)]
isisd: add ASLA support
Add the support of ASLA with the following TLV:
- Extended IS Reachability (already defined TLV 22)
- Application-Specific Link Attributes (Sub-TLV 16)
(to enable the Flex-Algo flag on a link)
- Admin-group (Sub-Sub-TLV 3)
- Extended Admin-group (Sub-Sub-TLV 14)
Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
Ospf segfault when Router Information is enabled in a non default VRF,
see issue #13144.
This patch forces vrf_id to default VRF for Opaque LSA and extension based
on Opaque LSA: Router Information, Traffic Engineering, Extended Prefix,
Extended Link and Segment Routing. Indeed, non default VRF is not yet
supported for Opaque LSA & co.
Hiroki Shirokura [Sat, 18 Dec 2021 08:11:58 +0000 (08:11 +0000)]
lib,vtysh,isisd,yang: algo cli/yang/callbacks
Define the IS-IS flex-algo structure in yang, the CLI configuration
commands and the skeletons of frontend and backend functions that are
called by the CLI code.
Signed-off-by: Hiroki Shirokura <hiroki.shirokura@linecorp.com> Signed-off-by: Eric Kinzie <ekinzie@labn.net> Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
isisd: add TLV_SIZE_MISMATCH for tlvs length mismatch
Basically in frrouting source code principle,
the log string should not be a complicated abstraction
or streamlined for grep.
But for log format for the "TLV size does not match ..."
can be unified, which makes development easier.
> $ grep "TLV size does not match expected size for" isisd/isis_tlvs.c
> "TLV size does not match expected size for Administrative Group!\n");
> "TLV size does not match expected size for Local IPv6 address!\n");
> ...(snip)...
> "TLV size does not match expected size for Adjacency SID!\n");
> "TLV size does not match expected size for Adjacency SID!\n");
> "TLV size does not match expected size for Adjacency SID!\n");
> "TLV size does not match expected size for LAN-Adjacency SID!\n");
> "TLV size does not match expected size for LAN-Adjacency SID!\n");
> "TLV size does not match expected size for LAN-Adjacency SID!\n");
>
> $ grep "TLV size does not match expected size for" isisd/isis_tlvs.c | wc -l
> 25
Signed-off-by: Hiroki Shirokura <hiroki.shirokura@linecorp.com> Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
Many of the enum definitions defined in isis_tlvs.h
are often extended at the end. The c/c++ allows
commas at the end of a list. This commit simplifies
the patching of later extensions.
Signed-off-by: Hiroki Shirokura <hiroki.shirokura@linecorp.com> Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
Hiroki Shirokura [Sun, 12 Dec 2021 09:20:21 +0000 (09:20 +0000)]
isisd: update struct isis_route_info has multiple sr info by algorithm
Before this commit, there was only one sr psid info
included in route_info.
In fact, in RFC8667, Algorithm ID, which is a property of
Prefix-SID, has 8 bits of information. That is, each Prefix
can hold up to 256 Prefix-SIDs. This commit implements it.
The previously implemented single Prefix-SID will be
continued as Algorithm 0.
Signed-off-by: Hiroki Shirokura <hiroki.shirokura@linecorp.com> Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
isisd: update struct isis_sr_psid_info with nh context
Prefix-SID nexthops and backup nexthops are stored respectively in
isis_route_info->nexthops and isis_route_info->backup->nexthops.
With Flex-Algo, there are multiple Prefix-SIDs for a single prefix in
different algorithms. Each of these Prefix-SIDs performs SPF calculation
with a separate contract and sets a nexthops, so it is necessary to
store a different set nexthops for each Prefix-SID.
Add a nexthops and backup nethops list into the Prefix-SID
isis_sr_psid_info struct and use these lists instead of the when needed
After this commit, the nexthops for each Prefix-SID is not
taken from route_info, but the nexthop set inside the
Prefix-SID is taken. This works for backup nexthops as well.
Signed-off-by: Hiroki Shirokura <hiroki.shirokura@linecorp.com> Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
Hiroki Shirokura [Sat, 11 Dec 2021 06:22:42 +0000 (06:22 +0000)]
isisd: update isis_tlvs_add_*_reach() with multi algorithm
isis_tlvs_add_extended_ip_reach adds IS-IS Extended
IP reachability to the LSP. In this case, if the
pcfg argument is not NULL, you can add IGP
Prefix-SID as its sub tlv.
Before this commit, only one Prefix-SID can be added.
After this commit, the argument is not a single
pointer but an array of pointers, and multiple
Prefix-SIDs can be added.
This feature is necessary because Flex-Algo
requires multiple Prefix-SIDs for each Algorithm.
Signed-off-by: Hiroki Shirokura <hiroki.shirokura@linecorp.com> Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
Hiroki Shirokura [Sat, 18 Dec 2021 09:03:01 +0000 (09:03 +0000)]
isisd: update SR_ALGORITHM_COUNT to 256
Before this commit, SR_ALGORITHM_COUNT was set to 2,
and each was hardcoded with router capability tlv.
When Flex-Algo is supported, SR-Algorithm may be
variably supported up to 256.
Signed-off-by: Hiroki Shirokura <hiroki.shirokura@linecorp.com> Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
ospfd: use rib metric as the base for set metric +/-
When using route maps with external routes in OSPF as follows:
```
set metric +10
```
The current behavior is to use the default ospf metric as the base and then add
to 10 to it. The behavior isn't useful as-is. A value 30 (20 dfeault + 10) can
be set directly instead. the behavior is also not consistent with bgp. bgp does
use the rib metric in this case as the base. The current behavior also doesn't
allow the metric to accumulate when crossing different routing domains such as
vrfs causing the metric to reset every time the route enters a new vrf with a new
ospf network.
This PR changes the behavior such that the rib metric is used as a base for
ospf exteral routes when used with `set metric -/+`
projects / matthieu / frr.git / log