Re: [PATCH net-next v3 5/5] net: mscc: ocelot: Add support for MRP

The 02/10/2021 10:18, Vladimir Oltean wrote:

Hi Vladimir,

Would you mind adding the switchdev MRP support for the DSA driver too,
and move the code to the common ocelot library? I would like to give it
a run. I think that's only fair, since I have to keep in sync the
vsc7514 driver too for features that get added through DSA :)

That is totally fair. I will do it in the next version :)

On Tue, Feb 09, 2021 at 09:21:12PM +0100, Horatiu Vultur wrote:

quoted

Add basic support for MRP. The HW will just trap all MRP frames on the
ring ports to CPU and allow the SW to process them. In this way it is
possible to for this node to behave both as MRM and MRC.

Current limitations are:
- it doesn't support Interconnect roles.
- it supports only a single ring.
- the HW should be able to do forwarding of MRP Test frames so the SW
  will not need to do this. So it would be able to have the role MRC
  without SW support.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
---
 drivers/net/ethernet/mscc/ocelot_net.c     | 154 +++++++++++++++++++++
 drivers/net/ethernet/mscc/ocelot_vsc7514.c |   6 +
 include/soc/mscc/ocelot.h                  |   6 +
 3 files changed, 166 insertions(+)

diff --git a/drivers/net/ethernet/mscc/ocelot_net.c b/drivers/net/ethernet/mscc/ocelot_net.c
index 8f12fa45b1b5..65971403e823 100644
--- a/drivers/net/ethernet/mscc/ocelot_net.c
+++ b/drivers/net/ethernet/mscc/ocelot_net.c

@@ -9,7 +9,10 @@
  */

 #include <linux/if_bridge.h>
+#include <linux/mrp_bridge.h>
 #include <net/pkt_cls.h>
+#include <soc/mscc/ocelot_vcap.h>
+#include <uapi/linux/mrp_bridge.h>
 #include "ocelot.h"
 #include "ocelot_vcap.h"

@@ -1069,6 +1072,139 @@ static int ocelot_port_obj_del_mdb(struct net_device *dev,
      return ocelot_port_mdb_del(ocelot, port, mdb);
 }

+#if IS_ENABLED(CONFIG_BRIDGE_MRP)
+static int ocelot_mrp_del_vcap(struct ocelot *ocelot, int port)
+{
+     struct ocelot_vcap_block *block_vcap_is2;
+     struct ocelot_vcap_filter *filter;
+
+     block_vcap_is2 = &ocelot->block[VCAP_IS2];
+     filter = ocelot_vcap_block_find_filter_by_id(block_vcap_is2, port,
+                                                  false);
+     if (!filter)
+             return 0;
+
+     return ocelot_vcap_filter_del(ocelot, filter);
+}
+
+static int ocelot_add_mrp(struct net_device *dev,
+                       const struct switchdev_obj_mrp *mrp)
+{
+     struct ocelot_port_private *priv = netdev_priv(dev);
+     struct ocelot_port *ocelot_port = &priv->port;
+     struct ocelot *ocelot = ocelot_port->ocelot;
+
+     if (mrp->p_port != dev && mrp->s_port != dev)
+             return 0;
+
+     if (ocelot->mrp_ring_id != 0 &&
+         ocelot->mrp_s_port &&
+         ocelot->mrp_p_port)
+             return -EINVAL;
+
+     if (mrp->p_port == dev)
+             ocelot->mrp_p_port = dev;
+
+     if (mrp->s_port == dev)
+             ocelot->mrp_s_port = dev;
+
+     ocelot->mrp_ring_id = mrp->ring_id;
+
+     return 0;
+}
+
+static int ocelot_del_mrp(struct net_device *dev,
+                       const struct switchdev_obj_mrp *mrp)
+{
+     struct ocelot_port_private *priv = netdev_priv(dev);
+     struct ocelot_port *ocelot_port = &priv->port;
+     struct ocelot *ocelot = ocelot_port->ocelot;
+
+     if (ocelot->mrp_p_port != dev && ocelot->mrp_s_port != dev)
+             return 0;
+
+     if (ocelot->mrp_ring_id == 0 &&
+         !ocelot->mrp_s_port &&
+         !ocelot->mrp_p_port)
+             return -EINVAL;
+
+     if (ocelot_mrp_del_vcap(ocelot, priv->chip_port))
+             return -EINVAL;
+
+     if (ocelot->mrp_p_port == dev)
+             ocelot->mrp_p_port = NULL;
+
+     if (ocelot->mrp_s_port == dev)
+             ocelot->mrp_s_port = NULL;
+
+     ocelot->mrp_ring_id = 0;
+
+     return 0;
+}
+
+static int ocelot_add_ring_role(struct net_device *dev,
+                             const struct switchdev_obj_ring_role_mrp *mrp)
+{
+     struct ocelot_port_private *priv = netdev_priv(dev);
+     struct ocelot_port *ocelot_port = &priv->port;
+     struct ocelot *ocelot = ocelot_port->ocelot;
+     struct ocelot_vcap_filter *filter;
+     int err;
+
+     if (ocelot->mrp_ring_id != mrp->ring_id)
+             return -EINVAL;
+
+     if (!mrp->sw_backup)
+             return -EOPNOTSUPP;
+
+     if (ocelot->mrp_p_port != dev && ocelot->mrp_s_port != dev)
+             return 0;
+
+     filter = kzalloc(sizeof(*filter), GFP_KERNEL);
+     if (!filter)
+             return -ENOMEM;
+
+     filter->key_type = OCELOT_VCAP_KEY_ETYPE;
+     filter->prio = 1;
+     filter->id.cookie = priv->chip_port;
+     filter->id.tc_offload = false;
+     filter->block_id = VCAP_IS2;
+     filter->type = OCELOT_VCAP_FILTER_OFFLOAD;
+     filter->ingress_port_mask = BIT(priv->chip_port);
+     *(__be16 *)filter->key.etype.etype.value = htons(ETH_P_MRP);
+     *(__be16 *)filter->key.etype.etype.mask = htons(0xffff);
+     filter->action.mask_mode = OCELOT_MASK_MODE_PERMIT_DENY;
+     filter->action.port_mask = 0x0;
+     filter->action.cpu_copy_ena = true;
+     filter->action.cpu_qu_num = 0;
+
+     err = ocelot_vcap_filter_add(ocelot, filter, NULL);
+     if (err)
+             kfree(filter);
+
+     return err;
+}
+
+static int ocelot_del_ring_role(struct net_device *dev,
+                             const struct switchdev_obj_ring_role_mrp *mrp)
+{
+     struct ocelot_port_private *priv = netdev_priv(dev);
+     struct ocelot_port *ocelot_port = &priv->port;
+     struct ocelot *ocelot = ocelot_port->ocelot;
+
+     if (ocelot->mrp_ring_id != mrp->ring_id)
+             return -EINVAL;
+
+     if (!mrp->sw_backup)
+             return -EOPNOTSUPP;
+
+     if (ocelot->mrp_p_port != dev && ocelot->mrp_s_port != dev)
+             return 0;
+
+     return ocelot_mrp_del_vcap(ocelot, priv->chip_port);
+}
+#endif
+

Would it make sense for this chunk of conditionally compiled code to
stay in a separate file like ocelot_mrp.c?

Also initially I was thinking to add this code in a different file but I
was not sure. But I will do that in the next version.

quoted

 static int ocelot_port_obj_add(struct net_device *dev,
                             const struct switchdev_obj *obj,
                             struct netlink_ext_ack *extack)

@@ -1083,6 +1219,15 @@ static int ocelot_port_obj_add(struct net_device *dev,
      case SWITCHDEV_OBJ_ID_PORT_MDB:
              ret = ocelot_port_obj_add_mdb(dev, SWITCHDEV_OBJ_PORT_MDB(obj));
              break;
+#if IS_ENABLED(CONFIG_BRIDGE_MRP)
+     case SWITCHDEV_OBJ_ID_MRP:
+             ret = ocelot_add_mrp(dev, SWITCHDEV_OBJ_MRP(obj));
+             break;
+     case SWITCHDEV_OBJ_ID_RING_ROLE_MRP:
+             ret = ocelot_add_ring_role(dev,
+                                        SWITCHDEV_OBJ_RING_ROLE_MRP(obj));
+             break;
+#endif

I'm not really sure why SWITCHDEV_OBJ_ID_MRP is conditionally defined.
If you look at SWITCHDEV_ATTR_ID_BRIDGE_VLAN_FILTERING, that isn't
conditionally defined, even though it depends on CONFIG_BRIDGE_VLAN_FILTERING
at runtime.

Then should I just create another patch for removing this? To be honest
I also prefer that SWITCHDEV_OBJ_ID_MRP will not be conditionally
defined.

quoted

      default:
              return -EOPNOTSUPP;
      }

@@ -1103,6 +1248,15 @@ static int ocelot_port_obj_del(struct net_device *dev,
      case SWITCHDEV_OBJ_ID_PORT_MDB:
              ret = ocelot_port_obj_del_mdb(dev, SWITCHDEV_OBJ_PORT_MDB(obj));
              break;
+#if IS_ENABLED(CONFIG_BRIDGE_MRP)
+     case SWITCHDEV_OBJ_ID_MRP:
+             ret = ocelot_del_mrp(dev, SWITCHDEV_OBJ_MRP(obj));
+             break;
+     case SWITCHDEV_OBJ_ID_RING_ROLE_MRP:
+             ret = ocelot_del_ring_role(dev,
+                                        SWITCHDEV_OBJ_RING_ROLE_MRP(obj));
+             break;
+#endif
      default:
              return -EOPNOTSUPP;
      }

diff --git a/drivers/net/ethernet/mscc/ocelot_vsc7514.c b/drivers/net/ethernet/mscc/ocelot_vsc7514.c
index 6b6eb92149ba..96a9c9f98060 100644
--- a/drivers/net/ethernet/mscc/ocelot_vsc7514.c
+++ b/drivers/net/ethernet/mscc/ocelot_vsc7514.c

@@ -698,6 +698,12 @@ static irqreturn_t ocelot_xtr_irq_handler(int irq, void *arg)
                      skb->offload_fwd_mark = 1;

              skb->protocol = eth_type_trans(skb, dev);
+#if IS_ENABLED(CONFIG_BRIDGE_MRP)
+             if (skb->protocol == ntohs(ETH_P_MRP) &&
+                 (priv->dev == ocelot->mrp_p_port ||
+                  priv->dev == ocelot->mrp_s_port))
+                     skb->offload_fwd_mark = 0;
+#endif

I wonder if you could just reserve a certain CPUQ for trapped traffic,
and just generically check for that, instead of MRP port roles?

That is a good idea, I would try to have a look.

quoted

              if (!skb_defer_rx_timestamp(skb))
                      netif_rx(skb);
              dev->stats.rx_bytes += len;

diff --git a/include/soc/mscc/ocelot.h b/include/soc/mscc/ocelot.h
index d0d48e9620fb..d95c019ad84e 100644
--- a/include/soc/mscc/ocelot.h
+++ b/include/soc/mscc/ocelot.h

@@ -682,6 +682,12 @@ struct ocelot {
      /* Protects the PTP clock */
      spinlock_t                      ptp_clock_lock;
      struct ptp_pin_desc             ptp_pins[OCELOT_PTP_PINS_NUM];
+
+#if IS_ENABLED(CONFIG_BRIDGE_MRP)
+     u16                             mrp_ring_id;
+     struct net_device               *mrp_p_port;
+     struct net_device               *mrp_s_port;
+#endif
 };

 struct ocelot_policer {
--

2.27.0

-- 
/Horatiu