Roland, thank you.
Sorry for the muddle, I'm going to use another mailer.
As to the event handling, you are right.
I think we should review the event handling design in IPoIB and change
it a bit, to make it smarter.
Regards,
Leonid
Roland Dreier wrote:
I applied this, although I had to fix the patch up by hand. In
addition to your mailer doing some sort of quoted-printable mangling
(that turns "==" into "=3D=3D"), you seem to have generated this
against an old version of the driver: your patch has
> schedule_work(&priv->flush_task);
but the current source has:
queue_work(ipoib_workqueue, &priv->flush_task);
Also, for the future, is there some way we can be smarter about link
down events? There's no sense in starting to try and join multicast
groups, etc. if we know that the port is down.
Thanks,
Roland
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
