[Quagga-bugs] [Bug 307] ospfd segfaulting

[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=307  
  




------- Additional Comments From skennedy at vcn.com  2006-10-17 20:15 -------
I explicity put in "passive-interface tunN" for each tunnel interface then
killed the ppp sessions to try and make ospfd crash, and it didn't crash!  So it
looks like it does indeed have something to do with ospf on the tunnel interfaces.

So if we could get a "passive-interface default" option that sets all the tun
interfaces to passive automatically, that would solve the problem for me.

------- Additional Comments From paul at dishone.st  2006-10-17 20:25 -------
Andy: I think we still have some index-by-local-address thingies somewhere, and
the add/delete of those is where we kill ospfd on Sean's setup. I had a quick
look to find where it's going wrong exactly, but can not - we should deal with
this more gracefully ;).

Enabling OSPF on only /one/ of the shared-address interfaces should work.

Interesting that setting passive on those interfaces fixes the crash. That
suggests the problem is in the NSM code somewhere (as passive interfaces still
get a struct ospf_interface setup for them).

Your CONNECTED_ID/PREFIX patch may help clean-up this stuff somewhat (if we
could agree to make the 'ID' always be the peer for PtP, when peer is specified).

For full unnumbered still need:

a) A table to detect 'unnumbered interfaces' (was it you (Andy) who was looking
at adding a flag for this to 'zebra' and Zserv?)

b) To describe such PtP interfaces by interface-ID rather than address, as the
RFC describes for such cases.

mise le meas,

Paul Jakma.  
  
  
  
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.