Hal Rosenstock wrote:
Hi Sean,
With the latest CM changes (not sure when I last picked these up), I am
now seeing the following sequences. The times between the CM MADs are on
the order of 10s of usec. These sequences all refer to the same
connection. It seems that the retransmissions are occuring despite the
expected MAD being received.
This was a bug in the MAD code, where a request was resent when it was
canceled. The bug entered when ib_modify_mad() was added, and a fix went in
yesterday. It shouldn't affect the operation any, just results in a lot of
unnecessary MADs...
- Sean
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
