Sasha Khapyorsky wrote:
On 20:40 Sun 19 Jul , Nicolas Morey-Chaisemartin wrote:
+ * Port select count reading
+ */
+typedef struct {
+ uint64_t xmit_wait;
+ time_t time;
+} perfmgr_db_sel_reading_t;
+
Why do we need a separate structure for this counter?
Sasha
I don't think we really need one but in the event plugin interface xmit_wait
counters are defined using another struct so I used a new structure to keep the
whole thing homogeneous.
I guess that separate xmit wait counter in event plugin was done
assuming proprietary (vendor specific) counters implemented in some
switches and I think that it was done even before IBA 1.2.1 where
xmit_wait counter was included in a standard set.
Sasha
So what is best in your mind?
Bring everything path into the perf counter struct (meaning all plugin
will have to change to fit the new API) or keeping the approach I proposed.
I'd rather choose the second solution not only because it's already
done, but in the future I'd like to have as many information as possible
on contention related counters (xmit_wait per VL for examples) and
probably some day vendor specific info will appear in this struct. SO
we're probably better off keeping it splitted from the beginning.
Anyway it's your call so if you'd rather merge everything ion the
existing struct, I'll rewrite the patch.
Nicolas
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
