[Xorp-cvs] XORP cvs commit: xorp/libxorp xorp/libxipc xorp/contrib/olsr

[Bruce M Simpson]

Andrea,

There are definitely timing issues with OLSR and XORP, your changes have 
probably just exposed them further, that's totally OK -- if we need a 
kludge to keep things happy for the time being, so be it.

Andrea Bittau wrote:
> Log message:
> 	Force commit to correct CVS log message.  Message /tmp/log.r492 should read:
> 	
> 	All tests pass with new XRL code.  The only hack was adding a 1 second sleep to
> 	olsr's test_routing1.py test_c4_partial test.  I need to understand this better
> 	but I think that it has to do with the simulator trying to keep events
> 	synchronized with time and assuming that each eventloop run is a "single
> 	iteration" - this no longer holds with the aggressive eventloop.  We need to
> 	decide whether we want this new eventloop (+26% performance with XRL benchmark)
> 	or whether we should keep the old one for simplicity and safety.
>   

The regression tests are pretty aggressive and rigorous. It is a mobile 
ad-hoc protocol so timing issues really come into play, more so than 
with our existing protocols, as you can see a change in L1/L2 as a form 
of wireless handover; my effort from last year was the best I could do 
with the resources available to me as an individual.

I wrote a proposal for a time-contigent debugging facility which I gave 
to Atanu as part of the SNMP work, as there's overlap between it and the 
requirement to be able to raise SNMP traps in a way which doesn't 
require knowledge of SNMP. However, as was "feature creep" for that 
specific task, we haven't acted on it.

I'd be happy to supply this proposal for further digestion if people are 
interested in seeing it, as it would most likely help with things like 
measuring the latency of BGP updates.

cheers
BMS