On Tue, 2016-02-23 at 11:20 -0800, pravin shelar wrote:
> On Tue, Feb 23, 2016 at 4:53 AM, Paolo Abeni <pab...@redhat.com> wrote:
> > This patch implements bookkeeping support to compute the maximum
> > headroom for all the devices in each datapath. When said value
> > changes, the underlying devs are notified via the
> > ndo_set_rx_headroom method.
> >
> > This also increases the internal vports xmit performance.
> >
> > Signed-off-by: Paolo Abeni <pab...@redhat.com>
> > ---
> > net/openvswitch/datapath.c | 48
> > ++++++++++++++++++++++++++++++++++++
> > net/openvswitch/datapath.h | 4 +++
> > net/openvswitch/vport-internal_dev.c | 10 +++++++-
> > 3 files changed, 61 insertions(+), 1 deletion(-)
> >
> > diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c
> > index c4e8455..7b37288 100644
> > --- a/net/openvswitch/datapath.c
> > +++ b/net/openvswitch/datapath.c
> > @@ -1908,6 +1908,34 @@ static struct vport *lookup_vport(struct net *net,
> > return ERR_PTR(-EINVAL);
> > }
> >
> > +/* Called with ovs_mutex */
> > +static void update_headroom(struct datapath *dp)
> > +{
> > + int i;
> > + struct vport *vport;
> > + struct net_device *dev;
> > + unsigned dev_headroom, max_headroom = 0;
> > +
> > + for (i = 0; i < DP_VPORT_HASH_BUCKETS; i++) {
> > + hlist_for_each_entry_rcu(vport, &dp->ports[i],
> > dp_hash_node) {
> > + dev = vport->dev;
> > + dev_headroom = netdev_get_fwd_headroom(dev);
> > + if (dev_headroom > max_headroom)
> > + max_headroom = dev_headroom;
> > + }
> > + }
> > +
> > + dp->max_headroom = max_headroom;
> > + for (i = 0; i < DP_VPORT_HASH_BUCKETS; i++) {
> > + hlist_for_each_entry_rcu(vport, &dp->ports[i],
> > dp_hash_node) {
> > + dev = vport->dev;
> > + if (dev->netdev_ops->ndo_set_rx_headroom)
> > + dev->netdev_ops->ndo_set_rx_headroom(dev,
> > +
> > max_headroom);
> > + }
> > + }
> > +}
> > +
> Code looks fine. But It could be kept in sync with bridge code from patch 2.
You are right.
I can had an helper to wrap the conditional ndo_set_rx_headroom()
invocation, so that all the code here and in the bridge will be devices
traversal only.
Paolo
