On 30 September 2014 10:24, Ben Pfaff <b...@nicira.com> wrote: > On Fri, Sep 26, 2014 at 09:28:17PM +1200, Joe Stringer wrote: > > One of the limiting factors on the number of flows that can be supported > > in the datapath is the overhead of assembling flow dump messages in the > > datapath. This patch modifies the flow_dump interface to allow > > revalidators to skip dumping the key, mask and actions from the > > datapath, by making use of the unique identifiers introduced in earlier > > patches. > > > > For each flow dump, the dpif user specifies whether to skip these > > attributes, allowing the common case to only dump a pair of 128-bit ID > > and flow stats. This increases the number of flows that a revalidator > > can handle per second by 50% or more. > > > > Signed-off-by: Joe Stringer <joestrin...@nicira.com> > > Just now I noticed that this memcpy() in > dp_netdev_flow_to_dpif_flow(): > memcpy(&flow->uid, &netdev_flow->uid, sizeof flow->uid); > could just be written as an assignment: > flow->uid = netdev_flow->uid; > (It was introduced in some earlier patch.) >
I'll roll this into the appropriate patch. > In dpif_netlink_flow_dump_create(), I wouldn't mark 'dump_flags' as > static; its address is never taken. > OK, I'll fix this up. > I suspect that check_recirc() and check_uid() should pass > DPIF_FP_MODIFY as well as DPIF_FP_CREATE, just in case a previous run > of OVS was killed at just the right time to leave a straggler probe > flow in the datapath. > OK, I'll make the check_recirc() change in a separate patch and include the check_uid() change in this patch. > Acked-by: Ben Pfaff <b...@nicira.com> > Thanks, holding onto this patch like the others. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
