[ovs-dev] [RFC 03/14] revalidator: Use 'cmap' for storing ukeys.

Joe Stringer joestringer at nicira.com
Mon Aug 25 04:27:27 UTC 2014


On 23 August 2014 06:47, Ben Pfaff <blp at nicira.com> wrote:

> On Thu, Aug 21, 2014 at 05:41:58PM +1200, Joe Stringer wrote:
> > Signed-off-by: Joe Stringer <joestringer at nicira.com>
>
> As of this patch, there isn't an advantage to using a cmap yet, right?
>

Not yet, no. Cmaps become useful when access is spread across handler
threads as well as revalidators.


>
> Here, I'd be inclined to add the ovsrcu_quiesce() call outside the
> 'if' statement, unconditionally, because otherwise it's just a matter
> of luck whether the number of ops happened to be a multiple of
> REVALIDATE_MAX_BATCH:
> > @@ -1582,6 +1587,7 @@ revalidate(struct revalidator *revalidator)
> >
> >          if (n_ops) {
> >              push_dump_ops__(udpif, ops, n_ops);
> > +            ovsrcu_quiesce();
> >          }
> >      }
> >      dpif_flow_dump_thread_destroy(dump_thread);
>

I'll fold this in.



>  Acked-by: Ben Pfaff <blp at nicira.com>
>

Thanks, I'll hold onto this patch until I recieve further review.



More information about the dev mailing list