Hi Pravin and Ben, Thanks for the feedback and yes, it's better to make truncate without associating to the output action. I will make the modification and resubmit v3 patch.
Regards, William On Mon, Apr 11, 2016 at 6:49 PM, Ben Pfaff <[email protected]> wrote: > On Mon, Apr 11, 2016 at 06:46:27PM -0700, pravin shelar wrote: > > On Mon, Apr 11, 2016 at 5:58 PM, Ben Pfaff <[email protected]> wrote: > > > On Mon, Apr 11, 2016 at 01:14:26PM -0700, pravin shelar wrote: > > >> The datapath truncate action should not be associated with output > > >> action. It should be general truncate action with truncate length > > >> parameter. This way we can apply it to actions other than output. > > > > > > This may require more care than perhaps you expect. For example, if > the > > > truncation removes a header, or part of a header, then the extracted > > > flow may need to be updated. It might also mean that datapath actions > > > to modify fields become invalid. > > > > There is no need to truncate packet immediately. We can keep the > > truncate len in skb_ovs_cb and use it to truncate packet right before > > seeing it over port or to user-space up-call. > > OK, that's a different situation, I agree. > _______________________________________________ dev mailing list [email protected] http://openvswitch.org/mailman/listinfo/dev
