Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.
RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.

Bug 1713449

Summary: Quagga computes wrong broadcast for /31 point-to-point RFC 3021 implementation
Product: Red Hat Enterprise Linux 7 Reporter: Curtis Taylor <cutaylor>
Component: quaggaAssignee: Michal Ruprich <mruprich>
Status: CLOSED WONTFIX QA Contact: Daniel Rusek <drusek>
Severity: medium Docs Contact:
Priority: medium    
Version: 7.6CC: drusek, thozza
Target Milestone: rcKeywords: Patch, TestCaseProvided
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2019-08-06 15:30:28 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1709724, 1716960    
Attachments:
Description Flags
Patch for the calculation of broadcast address for /31 prefixes none

Description Curtis Taylor 2019-05-23 17:32:25 UTC 
Description of problem:
When using /31 bit netmask, quagga computes broadcast to be NETWORK instead of 255.255.255.255 as prescribed in RFC3021

Version-Release number of selected component (if applicable):
quagga-0.99.22.4-5.el7

How reproducible:
In environment with /31 bit point-to-point address, (i.e. network with 2 addresses, only .0 and .1 for host and router).

Steps to Reproduce:
1.  Setup system with point-to-point /31 network
   e.g. 192.168.100.240/31, host at 192.168.100.241, router at 192.168.100.240
2.  Setup quagga/zebra/BGP (or RIP or etc)
3.  BGP sees default route like:
      Network          Next Hop            Metric LocPrf Weight Path
*> 0.0.0.0          192.168.122.240               500      0 65009 4771 4648 i

Actual results:
Default route does not get added to zebra, nor kernel routing table.  
Zebra.log reports:
 2019/05/21 16:46:45 ZEBRA: warning: interface eth0 broadcast addr 255.255.255.255/31 != calculated 192.168.122.240, routing protocols may malfunction

Expected results:
Route is added to Zebra and kernel routing table.

Additional info:
Pertinent RFC 3021 excerpt:
RFC3021 - Using 31 bit prefixes on ipv4 point-to-point links

2.3. Impact on Current Routing Protocols

   Networks with 31-bit prefixes have no impact on current routing
   protocols.  Most of the currently deployed routing protocols have
   been designed to provide classless routing.  Furthermore, the
   communication between peers is done using multicast, limited
   broadcast or unicast addresses (all on the local network), none of
   which are affected with the use of 31-bit subnet masks.
...
3.3. "Requirements for IP Version 4 Routers" [RFC1812]

   Section 4.2.2.11 (d) is replaced with:
   ...
      The following text is added to section 4.3.3.9:

      The 255.255.255.255 IP broadcast address MUST be used for
      broadcast Address Mask Replies in point-to-point links with 31-bit
      subnet masks

Quagga explicitly makes Broadcast for 31 bit address (hostaddr ^ ~mask.s_addr)
     192.168.122.241 ^ ~(255.255.255.255/31) = 192.168.122.240
200 /* Add connected IPv4 route to the interface. */
201 void
202 connected_add_ipv4 (struct interface *ifp, int flags, struct in_addr *addr,
203                     u_char prefixlen, struct in_addr *broad,
204                     const char *label)
205 {
...
240           if (broad->s_addr != ipv4_broadcast_addr(addr->s_addr,prefixlen))
241             {
242               char buf[2][INET_ADDRSTRLEN];
243               struct in_addr bcalc;
244               bcalc.s_addr = ipv4_broadcast_addr(addr->s_addr,prefixlen);
245               zlog_warn("warning: interface %s broadcast addr %s/%d != "
246                         "calculated %s, routing protocols may malfunction",
247                         ifp->name,
248                         inet_ntop (AF_INET, broad, buf[0], sizeof(buf[0])),
249                         prefixlen,
250                         inet_ntop (AF_INET, &bcalc, buf[1], sizeof(buf[1])));
251             }

   107 #define IPV4_MAX_PREFIXLEN 32

   837 in_addr_t
   838 ipv4_broadcast_addr (in_addr_t hostaddr, int masklen)
   839 {
   840   struct in_addr mask;
   841 
   842   masklen2ip (masklen, &mask);
   843   return (masklen != IPV4_MAX_PREFIXLEN-1) ?
   844          /* normal case */
   845          (hostaddr | ~mask.s_addr) :
   846          /* special case for /31 */
   847          (hostaddr ^ ~mask.s_addr);
   848 }

ipcalc determines the Broadcast to be 255.255.255.255 following RFC 3021 correctly:
$ ipcalc -b -n 192.168.122.241/31
BROADCAST=255.255.255.255
NETWORK=192.168.122.240

initscripts : ipcacl does cacl_broadcast() and explicitly sets 255.255.255.255, as expected per RFC 3021
159 struct in_addr calc_broadcast(struct in_addr addr, int prefix)
160 {
161     struct in_addr mask = prefix2mask(prefix);
162     struct in_addr broadcast;
163 
164     memset(&broadcast, 0, sizeof(broadcast));
165 
166 /* if prefix is set to 31 return 255.255.255.255 (RFC3021) */
167     if (mask.s_addr ==  htonl(0xFFFFFFFE))
168         broadcast.s_addr = htonl(0xFFFFFFFF);
169     else
...
Comment 3 Tomáš Hozza 2019-07-30 11:28:12 UTC 
Hi.

I tried to reproduce the issue with BGP route not being propagated to the kernel routing table. I can reproduce the warning regarding broadcast address, but even with it, BGP routes are still exchanged just fine and propagated to the kernel routing table. I also tried RIP, which worked as well.

I tried the following scenario:
- two instances of zebra and bgpd running on the same system, but in different network namespaces (netns)
- veth device with each end put into different netns
- point-to-point IPv4 addresses with prefix /31 statically configured on each end of the veth device.
- using RPM quagga-0.99.22.4-5.el7_4.x86_64


"Server side" configuration (netns: "vns")
------------------------------------------
~~~
[root@localhost ~]# ip -n vns a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: dummy1: <BROADCAST,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000
    link/ether 06:cb:ae:64:13:32 brd ff:ff:ff:ff:ff:ff
    inet 192.168.155.1/24 scope global dummy1
       valid_lft forever preferred_lft forever
    inet6 fe80::4cb:aeff:fe64:1332/64 scope link 
       valid_lft forever preferred_lft forever
4: VNSIF@if3: <BROADCAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 82:f0:be:7f:3c:d8 brd ff:ff:ff:ff:ff:ff link-netnsid 1
    inet 192.168.100.240/31 brd 255.255.255.255 scope global VNSIF
       valid_lft forever preferred_lft forever
    inet6 fe80::80f0:beff:fe7f:3cd8/64 scope link 
       valid_lft forever preferred_lft forever
~~~

~~~
[root@localhost ~]# ip -n vns route
192.168.100.240/31 dev VNSIF proto kernel scope link src 192.168.100.240 
192.168.155.0/24 dev dummy1 proto kernel scope link src 192.168.155.1
~~~

~~~
[root@localhost ~]# cat /etc/quagga/zebra-vns.conf 
! -*- zebra -*-
!
! zebra configuration file
!
hostname RouterVNS
password zebra
enable password zebra
!
! Interface's description. 
!

interface dummy1
  link-detect
interface VNSIF
  link-detect

debug zebra events
debug zebra kernel

log file /var/log/quagga/zebra-vns.log debugging
log syslog debugging
~~~

~~~
[root@localhost ~]# cat /etc/quagga/bgpd-vns.conf 
! -*- bgp -*-
!
hostname bgpdVNS
password zebra
!
router bgp 7675
 bgp router-id 192.168.100.240
 network 0.0.0.0/0
 network 192.168.155.0/24
 neighbor 192.168.100.241 remote-as 7676
!
line vty
!
debug bgp events
debug bgp updates
debug bgp zebra
!
log syslog debugging
log file /var/log/quagga/bgpd-vns.log debugging
~~~


"Client side" configuration (netns: "vnc")
------------------------------------------
~~~
[root@localhost ~]# ip -n vnc a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
3: VNCIF@if4: <BROADCAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 0a:c0:3d:c3:e1:7d brd ff:ff:ff:ff:ff:ff link-netnsid 0
    inet 192.168.100.241/31 brd 255.255.255.255 scope global VNCIF
       valid_lft forever preferred_lft forever
    inet6 fe80::8c0:3dff:fec3:e17d/64 scope link 
       valid_lft forever preferred_lft forever
~~~

~~~ (BEFORE DAEMONS ARE STARTED)
[root@localhost ~]# ip -n vnc route
192.168.100.240/31 dev VNCIF proto kernel scope link src 192.168.100.241
~~~

~~~
[root@localhost ~]# cat /etc/quagga/zebra-vnc.conf 
! -*- zebra -*-
!
! zebra configuration file
!
hostname RouterVNC
password zebra
enable password zebra
!
! Interface's description. 
!

interface VNCIF
  link-detect

debug zebra events
debug zebra kernel

log file /var/log/quagga/zebra-vnc.log debugging
log syslog debugging
~~~

~~~
[root@localhost ~]# cat /etc/quagga/bgpd-vnc.conf 
! -*- bgp -*-
!
hostname bgpdVNC
password zebra
!
router bgp 7676
 bgp router-id 192.168.100.241
 neighbor 192.168.100.240 remote-as 7675
 neighbor 192.168.100.240 next-hop-self
!
line vty
!
debug bgp events
debug bgp updates
debug bgp zebra
!
log syslog debugging
log file /var/log/quagga/bgpd-vnc.log debugging
~~~


Scenario #1:
============
- point-to-point address on VNCIF is configured before the bgpd and zebra daemons are started

~~~ ("Client" side log after starting bgpd)
júl 30 12:50:16 localhost.localdomain systemd[1]: Starting GNU Zebra routing manager...
júl 30 12:50:17 localhost.localdomain zebra[2862]: Zebra 0.99.22.4 starting: vty@2601
júl 30 12:50:17 localhost.localdomain systemd[1]: Started GNU Zebra routing manager.
júl 30 12:50:17 localhost.localdomain systemd[1]: Starting BGP routing daemon...
júl 30 12:50:17 localhost.localdomain bgpd[2864]: BGPd 0.99.22.4 starting: vty@2605, bgp@<all>:179
júl 30 12:50:17 localhost.localdomain systemd[1]: Started BGP routing daemon.
júl 30 12:50:17 localhost.localdomain bgpd[2864]: Import timer expired.
júl 30 12:50:17 localhost.localdomain zebra[2862]: zebra message comes from socket [16]
júl 30 12:50:17 localhost.localdomain zebra[2862]: client 16 says hello and bids fair to announce only bgp routes
júl 30 12:50:17 localhost.localdomain zebra[2862]: zebra message comes from socket [16]
júl 30 12:50:17 localhost.localdomain zebra[2862]: zebra message comes from socket [16]
júl 30 12:50:17 localhost.localdomain bgpd[2864]: Zebra rcvd: router id update 192.168.100.241/32
júl 30 12:50:17 localhost.localdomain bgpd[2864]: Zebra rcvd: interface add VNCIF
júl 30 12:50:17 localhost.localdomain bgpd[2864]: Zebra rcvd: interface VNCIF address add 192.168.100.241/31
júl 30 12:50:17 localhost.localdomain bgpd[2864]: Zebra rcvd: interface VNCIF address add fe80::8c0:3dff:fec3:e17d/64
júl 30 12:50:17 localhost.localdomain bgpd[2864]: Zebra rcvd: interface add lo
júl 30 12:50:17 localhost.localdomain bgpd[2864]: Zebra rcvd: interface lo address add 127.0.0.1/8
júl 30 12:50:17 localhost.localdomain bgpd[2864]: Zebra rcvd: interface lo address add ::1/128
júl 30 12:50:24 localhost.localdomain bgpd[2864]: 192.168.100.240 [Event] Connect start to 192.168.100.240 fd 16
júl 30 12:50:25 localhost.localdomain bgpd[2864]: 192.168.100.240 rcvd UPDATE w/ attr: nexthop 192.168.100.240, origin i, metric 0, path 7675
júl 30 12:50:25 localhost.localdomain bgpd[2864]: 192.168.100.240 rcvd 0.0.0.0/0
júl 30 12:50:25 localhost.localdomain bgpd[2864]: 192.168.100.240 rcvd 192.168.155.0/24
júl 30 12:50:25 localhost.localdomain bgpd[2864]: Zebra send: IPv4 route add 0.0.0.0/0 nexthop 192.168.100.240 metric 0 count 1
júl 30 12:50:25 localhost.localdomain bgpd[2864]: Zebra send: IPv4 route add 192.168.155.0/24 nexthop 192.168.100.240 metric 0 count 1
júl 30 12:50:25 localhost.localdomain zebra[2862]: zebra message comes from socket [16]
júl 30 12:50:25 localhost.localdomain zebra[2862]: zebra message comes from socket [16]
júl 30 12:50:25 localhost.localdomain zebra[2862]: netlink_route_multipath() (single hop): RTM_NEWROUTE 0.0.0.0/0, type IPv4 nexthop
júl 30 12:50:25 localhost.localdomain zebra[2862]: netlink_route_multipath() (single hop): nexthop via 192.168.100.240 if 3
júl 30 12:50:25 localhost.localdomain zebra[2862]: netlink_talk: netlink-cmd type RTM_NEWROUTE(24), seq=6
júl 30 12:50:25 localhost.localdomain zebra[2862]: netlink_parse_info: netlink-cmd ACK: type=RTM_NEWROUTE(24), seq=6, pid=0
júl 30 12:50:25 localhost.localdomain zebra[2862]: netlink_route_multipath() (single hop): RTM_NEWROUTE 192.168.155.0/24, type IPv4 nexthop
júl 30 12:50:25 localhost.localdomain zebra[2862]: netlink_route_multipath() (single hop): nexthop via 192.168.100.240 if 3
júl 30 12:50:25 localhost.localdomain zebra[2862]: netlink_talk: netlink-cmd type RTM_NEWROUTE(24), seq=7
júl 30 12:50:25 localhost.localdomain zebra[2862]: netlink_parse_info: netlink-cmd ACK: type=RTM_NEWROUTE(24), seq=7, pid=0
~~~

~~~ ("Client" side after starting bgpd)
[root@localhost ~]# vtysh -c "show ip bgp"
BGP table version is 0, local router ID is 192.168.100.241
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
              r RIB-failure, S Stale, R Removed
Origin codes: i - IGP, e - EGP, ? - incomplete

   Network          Next Hop            Metric LocPrf Weight Path
*> 0.0.0.0          192.168.100.240          0             0 7675 i
*> 192.168.155.0    192.168.100.240          0             0 7675 i

Total number of prefixes 2
~~~

~~~ ("Client" side after starting bgpd)
[root@localhost ~]# ip -n vnc route
default via 192.168.100.240 dev VNCIF proto zebra 
192.168.100.240/31 dev VNCIF proto kernel scope link src 192.168.100.241 
192.168.155.0/24 via 192.168.100.240 dev VNCIF proto zebra
~~~


Scenario #2:
============
- point-to-point address on VNCIF is NOT configured before the bgpd and zebra daemons are started
- once bgpd is running, the point-to-point address on VNCIF is configured

~~~ ("Client" side log after starting bgpd)
júl 30 13:01:54 localhost.localdomain systemd[1]: Starting GNU Zebra routing manager...
júl 30 13:01:54 localhost.localdomain zebra[2925]: Zebra 0.99.22.4 starting: vty@2601
júl 30 13:01:54 localhost.localdomain systemd[1]: Started GNU Zebra routing manager.
júl 30 13:01:54 localhost.localdomain systemd[1]: Starting BGP routing daemon...
júl 30 13:01:54 localhost.localdomain bgpd[2927]: BGPd 0.99.22.4 starting: vty@2605, bgp@<all>:179
júl 30 13:01:54 localhost.localdomain systemd[1]: Started BGP routing daemon.
júl 30 13:01:54 localhost.localdomain bgpd[2927]: Import timer expired.
júl 30 13:01:54 localhost.localdomain zebra[2925]: zebra message comes from socket [16]
júl 30 13:01:54 localhost.localdomain zebra[2925]: client 16 says hello and bids fair to announce only bgp routes
júl 30 13:01:54 localhost.localdomain zebra[2925]: zebra message comes from socket [16]
júl 30 13:01:54 localhost.localdomain zebra[2925]: zebra message comes from socket [16]
júl 30 13:01:54 localhost.localdomain bgpd[2927]: Zebra rcvd: router id update 0.0.0.0/32
júl 30 13:01:54 localhost.localdomain bgpd[2927]: Zebra rcvd: interface add VNCIF
júl 30 13:01:54 localhost.localdomain bgpd[2927]: Zebra rcvd: interface VNCIF address add fe80::8c0:3dff:fec3:e17d/64
júl 30 13:01:54 localhost.localdomain bgpd[2927]: Zebra rcvd: interface add lo
júl 30 13:01:54 localhost.localdomain bgpd[2927]: Zebra rcvd: interface lo address add 127.0.0.1/8
júl 30 13:01:54 localhost.localdomain bgpd[2927]: Zebra rcvd: interface lo address add ::1/128
júl 30 13:01:58 localhost.localdomain bgpd[2927]: 192.168.100.240 [Event] Connect start to 192.168.100.240 fd 16
júl 30 13:01:58 localhost.localdomain bgpd[2927]: can't connect to 192.168.100.240 fd 16 : Network is unreachable
júl 30 13:02:09 localhost.localdomain bgpd[2927]: Import timer expired.
júl 30 13:02:14 localhost.localdomain zebra[2925]: netlink_parse_info: netlink-listen type RTM_NEWADDR(20), seq=1564484535, pid=2928
júl 30 13:02:14 localhost.localdomain zebra[2925]: netlink_interface_addr RTM_NEWADDR VNCIF:
júl 30 13:02:14 localhost.localdomain zebra[2925]:   IFA_LOCAL     192.168.100.241/31
júl 30 13:02:14 localhost.localdomain zebra[2925]:   IFA_ADDRESS   192.168.100.241/31
júl 30 13:02:14 localhost.localdomain zebra[2925]:   IFA_BROADCAST 255.255.255.255/31
júl 30 13:02:14 localhost.localdomain zebra[2925]:   IFA_CACHEINFO pref -1, valid -1
júl 30 13:02:14 localhost.localdomain zebra[2925]: warning: interface VNCIF broadcast addr 255.255.255.255/31 != calculated 192.168.100.240, routing protocols may malfunction  <--- WARNING
júl 30 13:02:14 localhost.localdomain zebra[2925]: MESSAGE: ZEBRA_INTERFACE_ADDRESS_ADD 192.168.100.241/31 on VNCIF
júl 30 13:02:14 localhost.localdomain zebra[2925]: netlink_parse_info: netlink-listen type RTM_NEWROUTE(24), seq=0, pid=0
júl 30 13:02:14 localhost.localdomain zebra[2925]: RTM_NEWROUTE ipv4 multicast proto kernel
júl 30 13:02:14 localhost.localdomain zebra[2925]: netlink_parse_info: netlink-listen type RTM_NEWROUTE(24), seq=0, pid=0
júl 30 13:02:14 localhost.localdomain zebra[2925]: RTM_NEWROUTE ipv4 unicast proto kernel
júl 30 13:02:14 localhost.localdomain bgpd[2927]: Zebra rcvd: router id update 192.168.100.241/32
júl 30 13:02:14 localhost.localdomain bgpd[2927]: Zebra rcvd: interface VNCIF address add 192.168.100.241/31
júl 30 13:02:24 localhost.localdomain bgpd[2927]: Import timer expired.
júl 30 13:02:39 localhost.localdomain bgpd[2927]: Import timer expired.
júl 30 13:02:54 localhost.localdomain bgpd[2927]: Performing BGP general scanning
júl 30 13:02:54 localhost.localdomain bgpd[2927]: scanning IPv4 Unicast routing tables
júl 30 13:02:54 localhost.localdomain bgpd[2927]: scanning IPv6 Unicast routing tables
júl 30 13:02:54 localhost.localdomain bgpd[2927]: Import timer expired.
júl 30 13:03:09 localhost.localdomain bgpd[2927]: Import timer expired.
júl 30 13:03:24 localhost.localdomain bgpd[2927]: Import timer expired.
júl 30 13:03:39 localhost.localdomain bgpd[2927]: Import timer expired.
júl 30 13:03:54 localhost.localdomain bgpd[2927]: Performing BGP general scanning
júl 30 13:03:54 localhost.localdomain bgpd[2927]: scanning IPv4 Unicast routing tables
júl 30 13:03:54 localhost.localdomain bgpd[2927]: scanning IPv6 Unicast routing tables
júl 30 13:03:54 localhost.localdomain bgpd[2927]: Import timer expired.
júl 30 13:03:58 localhost.localdomain bgpd[2927]: 192.168.100.240 [Event] Connect start to 192.168.100.240 fd 16
júl 30 13:03:59 localhost.localdomain bgpd[2927]: 192.168.100.240 rcvd UPDATE w/ attr: nexthop 192.168.100.240, origin i, metric 0, path 7675
júl 30 13:03:59 localhost.localdomain bgpd[2927]: 192.168.100.240 rcvd 0.0.0.0/0
júl 30 13:03:59 localhost.localdomain bgpd[2927]: 192.168.100.240 rcvd 192.168.155.0/24
júl 30 13:03:59 localhost.localdomain bgpd[2927]: Zebra send: IPv4 route add 0.0.0.0/0 nexthop 192.168.100.240 metric 0 count 1
júl 30 13:03:59 localhost.localdomain bgpd[2927]: Zebra send: IPv4 route add 192.168.155.0/24 nexthop 192.168.100.240 metric 0 count 1
júl 30 13:03:59 localhost.localdomain zebra[2925]: zebra message comes from socket [16]
júl 30 13:03:59 localhost.localdomain zebra[2925]: zebra message comes from socket [16]
júl 30 13:03:59 localhost.localdomain zebra[2925]: netlink_route_multipath() (single hop): RTM_NEWROUTE 0.0.0.0/0, type IPv4 nexthop
júl 30 13:03:59 localhost.localdomain zebra[2925]: netlink_route_multipath() (single hop): nexthop via 192.168.100.240 if 3
júl 30 13:03:59 localhost.localdomain zebra[2925]: netlink_talk: netlink-cmd type RTM_NEWROUTE(24), seq=6
júl 30 13:03:59 localhost.localdomain zebra[2925]: netlink_parse_info: netlink-cmd ACK: type=RTM_NEWROUTE(24), seq=6, pid=0
júl 30 13:03:59 localhost.localdomain zebra[2925]: netlink_route_multipath() (single hop): RTM_NEWROUTE 192.168.155.0/24, type IPv4 nexthop
júl 30 13:03:59 localhost.localdomain zebra[2925]: netlink_route_multipath() (single hop): nexthop via 192.168.100.240 if 3
júl 30 13:03:59 localhost.localdomain zebra[2925]: netlink_talk: netlink-cmd type RTM_NEWROUTE(24), seq=7
júl 30 13:03:59 localhost.localdomain zebra[2925]: netlink_parse_info: netlink-cmd ACK: type=RTM_NEWROUTE(24), seq=7, pid=0
júl 30 13:04:10 localhost.localdomain bgpd[2927]: Import timer expired.
~~~

~~~ ("Client" side after starting bgpd and after IP on VNCIF is configured)
[root@localhost ~]# vtysh -c "show ip bgp"
BGP table version is 0, local router ID is 192.168.100.241
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
              r RIB-failure, S Stale, R Removed
Origin codes: i - IGP, e - EGP, ? - incomplete

   Network          Next Hop            Metric LocPrf Weight Path
*> 0.0.0.0          192.168.100.240          0             0 7675 i
*> 192.168.155.0    192.168.100.240          0             0 7675 i

Total number of prefixes 2
~~~

~~~ ("Client" side after starting bgpd and after IP on VNCIF is configured)
[root@localhost ~]# ip -n vnc route
default via 192.168.100.240 dev VNCIF proto zebra 
192.168.100.240/31 dev VNCIF proto kernel scope link src 192.168.100.241 
192.168.155.0/24 via 192.168.100.240 dev VNCIF proto zebra
~~~


Conclusion
==========
- While I acknowledge that the broadcast address in zebra for /31 prefix IPv4 addresses is not calculated according to RFC3021, its impact on BGP and RIP seems to be none, or at least I can not reproduce any issue with BGP or RIP caused by this bug.
- BGP routes are propagated to the kernel routing table without issues and I'm not able to reproduce the issue which is assumed to be caused by the problem with wrong broadcast address calculation for /31 prefix IPv4 addresses.
- While fixing the broadcast address calculation in upstream is definitely desirable, I don't think this is necessary a priority for RHEL-7, unless we can prove that it actually has real impct.


Questions
=========
1) Were we actually able to reproduce the issue (BGP routes not propagated to kernel routing table) internally? If yes, can we please get the description of the setup and configuration files OR access to the environment, where the issue is reproducible?

2) Any idea what may I been doing wrong in my setup and as a result not being able to reproduce the issue?
Comment 5 Tomáš Hozza 2019-07-31 14:47:58 UTC 
Created attachment 1595112 [details]
Patch for the calculation of broadcast address for /31 prefixes
Comment 6 Tomáš Hozza 2019-08-02 13:45:25 UTC 
Submitted patch to FRR (https://github.com/FRRouting/frr/pull/4773) and Quagga (will provide link later).
Comment 8 Red Hat Bugzilla 2023-09-14 05:29:05 UTC 
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days