Hal Rosenstock wrote:
On Fri, 2005-09-16 at 00:55, Eitan Zahavi wrote:
For the gen2 stack we could use the following "hack":
OpenSM scans all nodes for their description every time it does a full
sweep.
So we could cause an extra sweep after each node description change by
faking trap 144
(HCA port capability mask change) and sending it over.
Is a full sweep needed to handle a single HCA port capability mask
change ?
Current implementation does either a full sweep or light sweep (scan only
switch info for
change bit and non responsive ports).
Also, couldn't the NodeDescription change be handled with a light sweep
(if SM Get NodeDescription were added there) ? That might be better or
are there other issues with doing it that way ?
See above. No other issues but we might eventually define incremental algorithm.
-- Hal
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
