mlxsw: spectrum: Add IEEE 802.1Qbb PFC support

Implement the appropriate DCB ops and allow a user to configure certain
traffic classes as lossless.

The operation configures PFC for both the egress (respecting PFC frames)
and ingress (sending PFC frames) parts of the port.

At egress, when a PFC frame is received for a PFC enabled priority, then
all the priorities mapped to the same TC are stopped.

At ingress, the priority group (PG) buffers to which the enabled PFC
priorities are mapped are configured to be lossless. PFC frames will be
transmitted when the Xoff threshold is crossed.

The user-supplied delay parameter is used to determine the PG's size
according to the following formula:

PG_SIZE = PG_SIZE_LOSSY + delay * CELL_FACTOR + MTU

In the worst case scenario the delay will be made up of packets that
are all of size CELL_SIZE + 1, which means each packet will require
almost twice its true size when buffered in the switch. We therefore
multiply this value by the "cell factor", which is close to 2.

Another MTU is added in case the transmitting host already started
transmitting a maximum length frame when the PFC packet was received.

As with PAUSE enabled ports, when the port's MTU is changed both the
PGs' size and threshold are adjusted accordingly.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 116 insertions, 1 deletions

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_dcb.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_dcb.c
index 8786424f6191..0b323661c0b6 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_dcb.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_dcb.c
@@ -34,6 +34,7 @@
 
 #include <linux/netdevice.h>
 #include <linux/string.h>
+#include <linux/bitops.h>
 #include <net/dcbnl.h>
 
 #include "spectrum.h"
@@ -151,7 +152,8 @@ static int mlxsw_sp_port_headroom_set(struct mlxsw_sp_port *mlxsw_sp_port,
 	 * traffic is still directed to them.
 	 */
 	err = __mlxsw_sp_port_headroom_set(mlxsw_sp_port, dev->mtu,
-					   ets->prio_tc, pause_en);
+					   ets->prio_tc, pause_en,
+					   mlxsw_sp_port->dcb.pfc);
 	if (err) {
 		netdev_err(dev, "Failed to configure port's headroom\n");
 		return err;
@@ -291,11 +293,101 @@ err_port_ets_maxrate_set:
 	return err;
 }
 
+static int mlxsw_sp_port_pfc_cnt_get(struct mlxsw_sp_port *mlxsw_sp_port,
+				     u8 prio)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	struct ieee_pfc *my_pfc = mlxsw_sp_port->dcb.pfc;
+	char ppcnt_pl[MLXSW_REG_PPCNT_LEN];
+	int err;
+
+	mlxsw_reg_ppcnt_pack(ppcnt_pl, mlxsw_sp_port->local_port,
+			     MLXSW_REG_PPCNT_PRIO_CNT, prio);
+	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ppcnt), ppcnt_pl);
+	if (err)
+		return err;
+
+	my_pfc->requests[prio] = mlxsw_reg_ppcnt_tx_pause_get(ppcnt_pl);
+	my_pfc->indications[prio] = mlxsw_reg_ppcnt_rx_pause_get(ppcnt_pl);
+
+	return 0;
+}
+
+static int mlxsw_sp_dcbnl_ieee_getpfc(struct net_device *dev,
+				      struct ieee_pfc *pfc)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	int err, i;
+
+	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+		err = mlxsw_sp_port_pfc_cnt_get(mlxsw_sp_port, i);
+		if (err) {
+			netdev_err(dev, "Failed to get PFC count for priority %d\n",
+				   i);
+			return err;
+		}
+	}
+
+	memcpy(pfc, mlxsw_sp_port->dcb.pfc, sizeof(*pfc));
+
+	return 0;
+}
+
+static int mlxsw_sp_port_pfc_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				 struct ieee_pfc *pfc)
+{
+	char pfcc_pl[MLXSW_REG_PFCC_LEN];
+
+	mlxsw_reg_pfcc_pack(pfcc_pl, mlxsw_sp_port->local_port);
+	mlxsw_reg_pfcc_prio_pack(pfcc_pl, pfc->pfc_en);
+
+	return mlxsw_reg_write(mlxsw_sp_port->mlxsw_sp->core, MLXSW_REG(pfcc),
+			       pfcc_pl);
+}
+
+static int mlxsw_sp_dcbnl_ieee_setpfc(struct net_device *dev,
+				      struct ieee_pfc *pfc)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	int err;
+
+	if (mlxsw_sp_port->link.tx_pause || mlxsw_sp_port->link.rx_pause) {
+		netdev_err(dev, "PAUSE frames already enabled on port\n");
+		return -EINVAL;
+	}
+
+	err = __mlxsw_sp_port_headroom_set(mlxsw_sp_port, dev->mtu,
+					   mlxsw_sp_port->dcb.ets->prio_tc,
+					   false, pfc);
+	if (err) {
+		netdev_err(dev, "Failed to configure port's headroom for PFC\n");
+		return err;
+	}
+
+	err = mlxsw_sp_port_pfc_set(mlxsw_sp_port, pfc);
+	if (err) {
+		netdev_err(dev, "Failed to configure PFC\n");
+		goto err_port_pfc_set;
+	}
+
+	memcpy(mlxsw_sp_port->dcb.pfc, pfc, sizeof(*pfc));
+
+	return 0;
+
+err_port_pfc_set:
+	__mlxsw_sp_port_headroom_set(mlxsw_sp_port, dev->mtu,
+				     mlxsw_sp_port->dcb.ets->prio_tc, false,
+				     mlxsw_sp_port->dcb.pfc);
+	return err;
+}
+
 static const struct dcbnl_rtnl_ops mlxsw_sp_dcbnl_ops = {
 	.ieee_getets		= mlxsw_sp_dcbnl_ieee_getets,
 	.ieee_setets		= mlxsw_sp_dcbnl_ieee_setets,
 	.ieee_getmaxrate	= mlxsw_sp_dcbnl_ieee_getmaxrate,
 	.ieee_setmaxrate	= mlxsw_sp_dcbnl_ieee_setmaxrate,
+	.ieee_getpfc		= mlxsw_sp_dcbnl_ieee_getpfc,
+	.ieee_setpfc		= mlxsw_sp_dcbnl_ieee_setpfc,
 
 	.getdcbx		= mlxsw_sp_dcbnl_getdcbx,
 	.setdcbx		= mlxsw_sp_dcbnl_setdcbx,
@@ -338,6 +430,23 @@ static void mlxsw_sp_port_maxrate_fini(struct mlxsw_sp_port *mlxsw_sp_port)
 	kfree(mlxsw_sp_port->dcb.maxrate);
 }
 
+static int mlxsw_sp_port_pfc_init(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	mlxsw_sp_port->dcb.pfc = kzalloc(sizeof(*mlxsw_sp_port->dcb.pfc),
+					 GFP_KERNEL);
+	if (!mlxsw_sp_port->dcb.pfc)
+		return -ENOMEM;
+
+	mlxsw_sp_port->dcb.pfc->pfc_cap = IEEE_8021QAZ_MAX_TCS;
+
+	return 0;
+}
+
+static void mlxsw_sp_port_pfc_fini(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	kfree(mlxsw_sp_port->dcb.pfc);
+}
+
 int mlxsw_sp_port_dcb_init(struct mlxsw_sp_port *mlxsw_sp_port)
 {
 	int err;
@@ -348,11 +457,16 @@ int mlxsw_sp_port_dcb_init(struct mlxsw_sp_port *mlxsw_sp_port)
 	err = mlxsw_sp_port_maxrate_init(mlxsw_sp_port);
 	if (err)
 		goto err_port_maxrate_init;
+	err = mlxsw_sp_port_pfc_init(mlxsw_sp_port);
+	if (err)
+		goto err_port_pfc_init;
 
 	mlxsw_sp_port->dev->dcbnl_ops = &mlxsw_sp_dcbnl_ops;
 
 	return 0;
 
+err_port_pfc_init:
+	mlxsw_sp_port_maxrate_fini(mlxsw_sp_port);
 err_port_maxrate_init:
 	mlxsw_sp_port_ets_fini(mlxsw_sp_port);
 	return err;
@@ -360,6 +474,7 @@ err_port_maxrate_init:
 
 void mlxsw_sp_port_dcb_fini(struct mlxsw_sp_port *mlxsw_sp_port)
 {
+	mlxsw_sp_port_pfc_fini(mlxsw_sp_port);
 	mlxsw_sp_port_maxrate_fini(mlxsw_sp_port);
 	mlxsw_sp_port_ets_fini(mlxsw_sp_port);
 }