Change mad thread model to be 1 thread/port rather than 1 thread/port/CPU (Note that I have not applied this but am requesting comments).
Index: mad.c
===================================================================
--- mad.c (revision 1269)
+++ mad.c (working copy)
@@ -1900,7 +1900,7 @@
goto error7;
snprintf(name, sizeof name, "ib_mad%d", port_num);
- port_priv->wq = create_workqueue(name);
+ port_priv->wq = create_singlethread_workqueue(name);
if (!port_priv->wq) {
ret = -ENOMEM;
goto error8;
My guess is that this is probably preferable to having 1/port/CPU, especially on larger systems. It would depend on what the clients do when notified of a completion.
I guess one advantage of keeping it 1/port/CPU (for now) is that it would help test multi-threaded support.
- 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