You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
[X ] Did you test it on the latest FRRouting/frr master branch?
Description
Found that if a route-map with "set ip address x.x.x.x" is applied outbound to an unnumbered peer, the update is sent with a single link-local ipv6 address (16 bytes) instead of the 32 byte link-local previously sent. This change was the result of a change made to allow setting ip next-hop for mpls-vpn routes (per the comment) but has the side-effect of modifying behavior not pertaining to that use-case.
Steps to Reproduce
Bring up bgp unnumbered peering between two routers
Define a route-map with "set ip next-hop x.x.x.x"
Apply the route-map to outbound updates to the unnumbered peer
Display the received bgp table entry for the change
Expected behavior:
Before commit commit d54600f, the received table entry looked like this:
r8# sh ip bgp 10.0.0.4
BGP routing table entry for 10.0.0.4/32
Paths: (1 available, best #1, table default)
Advertised to non peer-group peers:
r1(swp1)
10 40
fe80::202:ff:fe00:6 from r1(swp1) (90.0.0.1)
(fe80::202:ff:fe00:6) (used)
Origin IGP, valid, external, bestpath-from-AS 10, best
Last update: Wed Mar 6 13:51:05 2019
Actual behavior:
After the above commit, the following bgp table entry on the receiver:
r8# sh ip bgp 10.0.0.4
BGP routing table entry for 10.0.0.4/32
Paths: (1 available, best #1, table default)
Advertised to non peer-group peers:
r1(swp1)
10 40
fe80::202:ff:fe00:6 from r1(swp1) (90.0.0.1)
Origin IGP, valid, external, bestpath-from-AS 10, best
Last update: Wed Mar 6 20:05:11 2019
And pcap file snippet
Border Gateway Protocol - UPDATE Message
Marker: ffffffffffffffffffffffffffffffff
Length: 71
Type: UPDATE Message (2)
Withdrawn Routes Length: 0
Total Path Attribute Length: 48
Path attributes
Path Attribute - MP_REACH_NLRI
Flags: 0x90, Optional, Extended-Length, Non-transitive, Complete
Type Code: MP_REACH_NLRI (14)
Length: 26
Address family identifier (AFI): IPv4 (1)
Subsequent address family identifier (SAFI): Unicast (1)
Next hop network address (16 bytes)
Next Hop: fe80::202:ff:fe00:6
Number of Subnetwork points of attachment (SNPA): 0
Network layer reachability information (5 bytes)
10.0.0.4/32
MP Reach NLRI prefix length: 32
MP Reach NLRI IPv4 prefix: 10.0.0.4
Path Attribute - ORIGIN: IGP
Flags: 0x40, Transitive, Well-known, Complete
Type Code: ORIGIN (1)
Length: 1
Origin: IGP (0)
Path Attribute - AS_PATH: 10 40
Flags: 0x50, Transitive, Extended-Length, Well-known, Complete
Type Code: AS_PATH (2)
Length: 10
AS Path segment: 10 40
Description
Found that if a route-map with "set ip address x.x.x.x" is applied outbound to an unnumbered peer, the update is sent with a single link-local ipv6 address (16 bytes) instead of the 32 byte link-local previously sent. This change was the result of a change made to allow setting ip next-hop for mpls-vpn routes (per the comment) but has the side-effect of modifying behavior not pertaining to that use-case.
Steps to Reproduce
Configuration on sender
Expected behavior:
Before commit commit d54600f, the received table entry looked like this:
Pcap file snippet of the above update:
Actual behavior:
After the above commit, the following bgp table entry on the receiver:
And pcap file snippet
Portion of commit causing the changed behavior
Components
bgpd
Versions
The text was updated successfully, but these errors were encountered: