[ovs-dev] [bond megaflow v4 2/4] ofproto-dpif: Added Per backer recirculation ID management

[Ben Pfaff]

On Tue, Mar 25, 2014 at 12:39:49PM -0700, Andy Zhou wrote:
> On Tue, Mar 25, 2014 at 8:35 AM, Ben Pfaff <blp at nicira.com> wrote:
> > On Mon, Mar 24, 2014 at 06:58:42PM -0700, Andy Zhou wrote:
> >> Recirculation ID needs to be unique per datapath. Its usage will be
> >> tracked by the backer that corresponds to the datapath.
> >>
> >> In theory, Recirculation ID can be any uint32_t value, except 0. This
> >> implementation limits to a smaller range just for ease of debugging.
> >> Make the range size 0 effectively disables recirculation.
> >>
> >> Signed-off-by: Andy Zhou <azhou at nicira.com>
> >
> > There's some trailing whitespace:
> >     /home/blp/nicira/ovs/.git/rebase-apply/patch:79: trailing whitespace.
> >         struct ovs_mutex lock;
> >     /home/blp/nicira/ovs/.git/rebase-apply/patch:261: trailing whitespace.
> >      * ID pool keeps track recirculation ids.
> >     /home/blp/nicira/ovs/.git/rebase-apply/patch:273: trailing whitespace.
> >      *
> >     warning: 3 lines add whitespace errors.
> >
> > Missing {} here in rid_pool_alloc_id():
> > +    if (rids->n_ids == 0)
> > +        return 0;
> > also here:
> > +        if ((rid_pool_find(rids, id)))
> > +            goto found_free_id;
> > Extra () here in rid_pool_alloc_id():
> > +    if (!(rid_pool_find(rids, rids->next_free_id))) {
> > and here:
> > +        if ((rid_pool_find(rids, id)))
> >
> > It looks like open_dpif_backer() creates the recirculation pool after it
> > starts udpif threads.  It might be a good idea to use the other order (I
> > guess receiving packets could eventually trigger recirc id allocation).
> >
> > Acked-by: Ben Pfaff <blp at nicira.com>
> >
> Thanks. I will apply with following fixes:

Thank you.