[Quagga-bugs] [Bug 195] RIPv1: Auto-summary on classful boundaries doesn't work

[bugzilla-daemon at allevil.dishone.st]

Please do not reply directly to this email. All additional  
comments should be made in the comments box of this bug  
report.  
  
http://bugzilla.quagga.net/show_bug.cgi?id=195  
  

paul at dishone.st changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|0.99.1                      |unspecified
           Severity|normal                      |enhancement




------- Additional Comments From paul at dishone.st  2007-04-01 19:14 -------
My notes from looking into this:

The case where R1-R2 link is in 10/8, and hence clearly R1 and R2 are in
different networks, falls into:

RFC2453 3.7 or RFC1058, 3.2:

  "Normally hosts only
   know the subnet masks for directly-connected networks.  Therefore,
   unless special provisions have been made, routes to a subnet must not
   be sent outside the network of which the subnet is a part."

and RFC2453 4.3:

  "On an interface where a RIP-1 router may hear and operate on the
   information in a RIP-2 routing entry the following rules apply:

   1) information internal to one network must never be advertised into
      another network,

   2) information about a more specific subnet may not be advertised
      where RIP-1 routers would consider it a host route, and

   3) supernet routes (routes with a netmask less specific than the
      "natural" network mask) must not be advertised where they could be
      misinterpreted by RIP-1 routers."

The case where R1-R2 link is in 172.16/16, and hence R1 and R2 are part of the
same, subnetted, network, falls into:

RFC2453 3.7 or RFC1058, 3.2:

  "For neighbors connected to the subnetted
   network, it generates a list of all subnets to which it is directly
   connected, using the subnet number.  For neighbors connected to other
   networks, it makes a single entry for the network as a whole, showing
   the metric associated with that network.  (This metric would normally
   be the smallest metric for the subnets to which the gateway is
   attached.)"

In essence R1 must only advertise the subnetted 172.16/16 networks to R2 if it
knows R1 can interpret them as something other than host routes. R1 can only
know that if it knows it is connected to R2 via a subnet of 172.16. Any other
case, it must suppress the routes.

So the behaviour is correct, at least in terms of how it /suppresses/ routes.
What's missing is the summarisation, which requires a bit more state than the
approach below.



------- Additional Comments From paul at dishone.st  2007-04-01 22:26 -------
The problem here isn't any bad behaviour, ripd does exactly what it is supposed
to and filters out subnet routes when considering sending them to routers on
other networks. Rather, ripd simply does not do auto-summarisation - the
administrator must /manually/ install routes for the class-boundary network.

I.e. R1 above, being a border router for the 172.16 class B network, should have:

router rip
 route 172.16.0.0/16

in its configuration.

Hence, this bug should be 'enhancement', and it should not block 1.0.
  
  
  
  
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.