On Sat, Jun 6, 2015 at 2:20 PM, Ben Pfaff <b...@nicira.com> wrote:
> On Sun, May 31, 2015 at 05:35:14PM -0700, Alex Wang wrote:
> > > I'm not fully comfortable with assert-failing (killing ovs-vswitchd) if
> > > the flow table has unsupported features in it. Also, even when HSA
> > > doesn't kill the process due to assert-failing, it could still delay
> the
> > > process by an arbitrary amount of time. Do you have an idea for a way
> > > to keep curious admins from accidentally taking down their switch by
> > > running an HSA command ("hey, what does this ovs-appctl command do?")?
> >
> > Sure, the asserts are more for alerting things at development time.
> > Will change to VLOG.
>
> Great.
>
> > Also, I can only think of separating the logic info a thread to prevent
> > it from blocking the main thread?
>
> I don't know whether that is practical. Is it? If not, then let's not
> worry about it for now. If it becomes a problem, we'll come up with
> something.
>
Yeah, it is really not thoroughly thought about. Will keep exploring
Also, working on adding more tests, and addressing request for a new
function (detect OF flows that will never be used)
Will repost series later,
Thanks,
Alex Wang,
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
