[PATCH net-next 00/18] Mellanox mlx4 driver updates

[PATCH net-next 00/18] Mellanox mlx4 driver updates

[Or Gerlitz]

Hi Dave,

The main feature added by this series are Ido's changes to support 
Granular QoS for VFs, where for the time being only max rate is supported.

Muhammad added support for setting rx-fcs and rx-all through ethtool, 
and Ido did the interface identify work.

Last, add Ido as a maintainer for the mlx4 Ethernet driver!

Some of next week is the Passover holiday here and I will be 
mostly OOO. If needed (...) Ido or Amir will send V1 and such.

Rebased against net-next commit 033f46b "crypto: algif - 
explicitly mark end of data".

Or.

Ido Shamay (15):
  net/mlx4: Add RSS support for fragmented IP datagrams
  net/mlx4_en: Change loopback only upon feature change
  net/mlx4: Make mlx4_is_eth visible inline funcion
  net/mlx4: Aesthetic code changes in multi_func_init
  net/mlx4: New file for QoS related firmware commands
  net/mlx4: Add mlx4_ALLOCATE_VPP implementation
  net/mlx4: Add mlx4_SET_VPORT_QOS implementation
  net/mlx4: Query device for QoS per VF support
  net/mlx4: Allocate VPPs for each port on PF init
  net/mlx4: Added qos_vport QP configuration in VST mode
  net/mlx4_en: Enable TX rate limit per VF
  net/mlx4: Warn users of depracated QoS Firmware
  net/mlx4: Set enhanced QoS support by default when ETS supported
  net/mlx4: Add SET_PORT opcode modifiers enumeration
  net/mlx4_en: Add interface identify support

Muhammad Mahajna (2):
  net/mlx4_en: Add RX-FCS support
  net/mlx4_en: Add RX-ALL support

Or Gerlitz (1):
  MAINTAINERS: Update mlx4_en entry

 MAINTAINERS                                        |    1 +
 drivers/infiniband/hw/mlx4/main.c                  |   11 +-
 drivers/net/ethernet/mellanox/mlx4/Makefile        |    5 +-
 drivers/net/ethernet/mellanox/mlx4/cmd.c           |  292 +++++++++++++++++++-
 drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c     |    1 +
 drivers/net/ethernet/mellanox/mlx4/en_ethtool.c    |   27 ++
 drivers/net/ethernet/mellanox/mlx4/en_main.c       |    5 +
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c     |   70 ++++-
 drivers/net/ethernet/mellanox/mlx4/en_rx.c         |    5 +-
 drivers/net/ethernet/mellanox/mlx4/fw.c            |   53 ++++-
 drivers/net/ethernet/mellanox/mlx4/fw.h            |    1 +
 drivers/net/ethernet/mellanox/mlx4/fw_qos.c        |  289 +++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/fw_qos.h        |  145 ++++++++++
 drivers/net/ethernet/mellanox/mlx4/main.c          |   31 ++
 drivers/net/ethernet/mellanox/mlx4/mlx4.h          |   30 +--
 drivers/net/ethernet/mellanox/mlx4/mlx4_en.h       |    1 +
 drivers/net/ethernet/mellanox/mlx4/port.c          |  133 +++++-----
 drivers/net/ethernet/mellanox/mlx4/qp.c            |    5 +
 .../net/ethernet/mellanox/mlx4/resource_tracker.c  |   11 +-
 include/linux/mlx4/cmd.h                           |   13 +-
 include/linux/mlx4/device.h                        |   20 +-
 include/linux/mlx4/qp.h                            |    8 +-
 22 files changed, 1023 insertions(+), 134 deletions(-)
 create mode 100644 drivers/net/ethernet/mellanox/mlx4/fw_qos.c
 create mode 100644 drivers/net/ethernet/mellanox/mlx4/fw_qos.h

[PATCH net-next 01/18] net/mlx4: Add RSS support for fragmented IP datagrams

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Enable RSS support for fragmented IP packets, when device supports it.
Until now, fragmented IP packets were directed only to the default_qpn.
Since IP fragments (datagram) have no upper protocols (L3 IP packets),
hash is performed on 3-tuple - dst MAC, source IP and dest IP. The HW
makes sure that this holds for the 1st fragment too, so all fragments
go to the same QP.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/fw.c   |   12 ++++++++++++
 drivers/net/ethernet/mellanox/mlx4/fw.h   |    1 +
 drivers/net/ethernet/mellanox/mlx4/main.c |    2 ++
 include/linux/mlx4/device.h               |    1 +
 4 files changed, 16 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index 209a617..412e019 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -105,6 +105,7 @@ static void dump_dev_cap_flags(struct mlx4_dev *dev, u64 flags)
 		[41] = "Unicast VEP steering support",
 		[42] = "Multicast VEP steering support",
 		[48] = "Counters support",
+		[52] = "RSS IP fragments support",
 		[53] = "Port ETS Scheduler support",
 		[55] = "Port link type sensing support",
 		[59] = "Port management change event support",
@@ -1138,6 +1139,9 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave,
 	}
 	for (; slave_port < dev->caps.num_ports; ++slave_port)
 		flags &= ~(MLX4_DEV_CAP_FLAG_WOL_PORT1 << slave_port);
+
+	/* Not exposing RSS IP fragments to guests */
+	flags &= ~MLX4_DEV_CAP_FLAG_RSS_IP_FRAG;
 	MLX4_PUT(outbox->buf, flags, QUERY_DEV_CAP_EXT_FLAGS_OFFSET);
 
 	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_VL_PORT_OFFSET);
@@ -1701,6 +1705,10 @@ int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
 	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_COUNTERS)
 		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 4);
 
+	/* Enable RSS spread to fragmented IP packets when supported */
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_RSS_IP_FRAG)
+		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 13);
+
 	/* CX3 is capable of extending CQEs/EQEs from 32 to 64 bytes */
 	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_64B_EQE) {
 		*(inbox + INIT_HCA_EQE_CQE_OFFSETS / 4) |= cpu_to_be32(1 << 29);
@@ -1889,6 +1897,10 @@ int mlx4_QUERY_HCA(struct mlx4_dev *dev,
 		else
 			param->steering_mode = MLX4_STEERING_MODE_A0;
 	}
+
+	if (dword_field & (1 << 13))
+		param->rss_ip_frags = 1;
+
 	/* steering attributes */
 	if (param->steering_mode == MLX4_STEERING_MODE_DEVICE_MANAGED) {
 		MLX4_GET(param->mc_base, outbox, INIT_HCA_FS_BASE_OFFSET);
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.h b/drivers/net/ethernet/mellanox/mlx4/fw.h
index 863655b..07cb7c2 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.h
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.h
@@ -203,6 +203,7 @@ struct mlx4_init_hca_param {
 	u64 dev_cap_enabled;
 	u16 cqe_size; /* For use only when CQE stride feature enabled */
 	u16 eqe_size; /* For use only when EQE stride feature enabled */
+	u8 rss_ip_frags;
 };
 
 struct mlx4_init_ib_param {
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index 43aa767..ee0a670 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -885,6 +885,8 @@ static int mlx4_slave_cap(struct mlx4_dev *dev)
 	mlx4_warn(dev, "Timestamping is not supported in slave mode\n");
 
 	slave_adjust_steering_mode(dev, &dev_cap, &hca_param);
+	mlx4_dbg(dev, "RSS support for IP fragments is %s\n",
+		 hca_param.rss_ip_frags ? "on" : "off");
 
 	if (func_cap.extra_flags & MLX4_QUERY_FUNC_FLAGS_BF_RES_QP &&
 	    dev->caps.bf_reg_size)
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index ab7ebec..41eaafd 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -174,6 +174,7 @@ enum {
 	MLX4_DEV_CAP_FLAG_VEP_UC_STEER	= 1LL << 41,
 	MLX4_DEV_CAP_FLAG_VEP_MC_STEER	= 1LL << 42,
 	MLX4_DEV_CAP_FLAG_COUNTERS	= 1LL << 48,
+	MLX4_DEV_CAP_FLAG_RSS_IP_FRAG   = 1LL << 52,
 	MLX4_DEV_CAP_FLAG_SET_ETH_SCHED = 1LL << 53,
 	MLX4_DEV_CAP_FLAG_SENSE_SUPPORT	= 1LL << 55,
 	MLX4_DEV_CAP_FLAG_PORT_MNG_CHG_EV = 1LL << 59,
-- 
1.7.1

[PATCH net-next 02/18] net/mlx4_en: Change loopback only upon feature change

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Currently any change of netdev features results in a call to
mlx4_en_update_loopback_state(). Those calls are unnecessary,
and should be called only upon loopback feature change.

Also moved some of the logic into mlx4_en_update_loopback_state().

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_main.c   |    5 +++++
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c |   12 +++++-------
 2 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_main.c b/drivers/net/ethernet/mellanox/mlx4/en_main.c
index 58d5a07..913b716 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_main.c
@@ -103,6 +103,11 @@ void mlx4_en_update_loopback_state(struct net_device *dev,
 {
 	struct mlx4_en_priv *priv = netdev_priv(dev);
 
+	if (features & NETIF_F_LOOPBACK)
+		priv->ctrl_flags |= cpu_to_be32(MLX4_WQE_CTRL_FORCE_LOOPBACK);
+	else
+		priv->ctrl_flags &= cpu_to_be32(~MLX4_WQE_CTRL_FORCE_LOOPBACK);
+
 	priv->flags &= ~(MLX4_EN_FLAG_RX_FILTER_NEEDED|
 			MLX4_EN_FLAG_ENABLE_HW_LOOPBACK);
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 354e254..4ee564c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -2210,13 +2210,11 @@ static int mlx4_en_set_features(struct net_device *netdev,
 		en_info(priv, "Turn %s TX vlan strip offload\n",
 			(features & NETIF_F_HW_VLAN_CTAG_TX) ? "ON" : "OFF");
 
-	if (features & NETIF_F_LOOPBACK)
-		priv->ctrl_flags |= cpu_to_be32(MLX4_WQE_CTRL_FORCE_LOOPBACK);
-	else
-		priv->ctrl_flags &=
-			cpu_to_be32(~MLX4_WQE_CTRL_FORCE_LOOPBACK);
-
-	mlx4_en_update_loopback_state(netdev, features);
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_LOOPBACK)) {
+		en_info(priv, "Turn %s loopback\n",
+			(features & NETIF_F_LOOPBACK) ? "ON" : "OFF");
+		mlx4_en_update_loopback_state(netdev, features);
+	}
 
 	return 0;
 
-- 
1.7.1

[PATCH net-next 03/18] net/mlx4: Make mlx4_is_eth visible inline funcion

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Currently implemented as static function in resource_tracker.c --
this change will allow other files in mlx4_core to use it as well.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 .../net/ethernet/mellanox/mlx4/resource_tracker.c  |    5 -----
 include/linux/mlx4/device.h                        |    5 +++++
 2 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
index c258f86..2076e98 100644
--- a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
+++ b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
@@ -221,11 +221,6 @@ struct res_fs_rule {
 	int			qpn;
 };
 
-static int mlx4_is_eth(struct mlx4_dev *dev, int port)
-{
-	return dev->caps.port_mask[port] == MLX4_PORT_TYPE_IB ? 0 : 1;
-}
-
 static void *res_tracker_lookup(struct rb_root *root, u64 res_id)
 {
 	struct rb_node *node = root->rb_node;
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 41eaafd..0f7f13d 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -1002,6 +1002,11 @@ static inline int mlx4_is_slave(struct mlx4_dev *dev)
 	return dev->flags & MLX4_FLAG_SLAVE;
 }
 
+static inline int mlx4_is_eth(struct mlx4_dev *dev, int port)
+{
+	return dev->caps.port_type[port] == MLX4_PORT_TYPE_IB ? 0 : 1;
+}
+
 int mlx4_buf_alloc(struct mlx4_dev *dev, int size, int max_direct,
 		   struct mlx4_buf *buf, gfp_t gfp);
 void mlx4_buf_free(struct mlx4_dev *dev, int size, struct mlx4_buf *buf);
-- 
1.7.1

[PATCH net-next 04/18] net/mlx4: Aesthetic code changes in multi_func_init

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Previous vf_oper and vf_admin code created very long lines, making it hard
to read the code. Added relevant in-struct pointers to reduce code
complexity and avoid code lines spread over 80 lines. Same logic is preserved.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/cmd.c |   19 +++++++++++++++----
 1 files changed, 15 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index 20b3c7b..52bf0b5 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -2213,6 +2213,9 @@ int mlx4_multi_func_init(struct mlx4_dev *dev)
 	}
 
 	if (mlx4_is_master(dev)) {
+		struct mlx4_vf_oper_state *vf_oper;
+		struct mlx4_vf_admin_state *vf_admin;
+
 		priv->mfunc.master.slave_state =
 			kzalloc(dev->num_slaves *
 				sizeof(struct mlx4_slave_state), GFP_KERNEL);
@@ -2232,6 +2235,8 @@ int mlx4_multi_func_init(struct mlx4_dev *dev)
 			goto err_comm_oper;
 
 		for (i = 0; i < dev->num_slaves; ++i) {
+			vf_admin = &priv->mfunc.master.vf_admin[i];
+			vf_oper = &priv->mfunc.master.vf_oper[i];
 			s_state = &priv->mfunc.master.slave_state[i];
 			s_state->last_cmd = MLX4_COMM_CMD_RESET;
 			for (j = 0; j < MLX4_EVENT_TYPES_NUM; ++j)
@@ -2242,6 +2247,9 @@ int mlx4_multi_func_init(struct mlx4_dev *dev)
 				     &priv->mfunc.comm[i].slave_read);
 			mmiowb();
 			for (port = 1; port <= MLX4_MAX_PORTS; port++) {
+				struct mlx4_vport_state *admin_vport;
+				struct mlx4_vport_state *oper_vport;
+
 				s_state->vlan_filter[port] =
 					kzalloc(sizeof(struct mlx4_vlan_fltr),
 						GFP_KERNEL);
@@ -2250,11 +2258,14 @@ int mlx4_multi_func_init(struct mlx4_dev *dev)
 						kfree(s_state->vlan_filter[port]);
 					goto err_slaves;
 				}
+
+				admin_vport = &vf_admin->vport[port];
+				oper_vport = &vf_oper->vport[port].state;
 				INIT_LIST_HEAD(&s_state->mcast_filters[port]);
-				priv->mfunc.master.vf_admin[i].vport[port].default_vlan = MLX4_VGT;
-				priv->mfunc.master.vf_oper[i].vport[port].state.default_vlan = MLX4_VGT;
-				priv->mfunc.master.vf_oper[i].vport[port].vlan_idx = NO_INDX;
-				priv->mfunc.master.vf_oper[i].vport[port].mac_idx = NO_INDX;
+				admin_vport->default_vlan = MLX4_VGT;
+				oper_vport->default_vlan = MLX4_VGT;
+				vf_oper->vport[port].vlan_idx = NO_INDX;
+				vf_oper->vport[port].mac_idx = NO_INDX;
 			}
 			spin_lock_init(&s_state->lock);
 		}
-- 
1.7.1

[PATCH net-next 05/18] net/mlx4: New file for QoS related firmware commands

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Create two new files fw_qos.h and fw_qos.c in mlx4_core module.

It gathers all relevant QoS firmware related commands etc, thus improving
encapsulation of the mlx4_core module. For now it contains the QoS existing
commands: mlx4_SET_PORT_SCHEDULER and mlx4_SET_PORT_PRIO2TC.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/Makefile    |    5 +-
 drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c |    1 +
 drivers/net/ethernet/mellanox/mlx4/fw_qos.c    |  125 ++++++++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/fw_qos.h    |   75 ++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/mlx4.h      |   15 ---
 drivers/net/ethernet/mellanox/mlx4/port.c      |   71 -------------
 include/linux/mlx4/device.h                    |    5 -
 7 files changed, 204 insertions(+), 93 deletions(-)
 create mode 100644 drivers/net/ethernet/mellanox/mlx4/fw_qos.c
 create mode 100644 drivers/net/ethernet/mellanox/mlx4/fw_qos.h

diff --git a/drivers/net/ethernet/mellanox/mlx4/Makefile b/drivers/net/ethernet/mellanox/mlx4/Makefile
index 3e9c70f..c82217e 100644
--- a/drivers/net/ethernet/mellanox/mlx4/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx4/Makefile
@@ -1,7 +1,8 @@
 obj-$(CONFIG_MLX4_CORE)		+= mlx4_core.o
 
-mlx4_core-y :=	alloc.o catas.o cmd.o cq.o eq.o fw.o icm.o intf.o main.o mcg.o \
-		mr.o pd.o port.o profile.o qp.o reset.o sense.o srq.o resource_tracker.o
+mlx4_core-y :=	alloc.o catas.o cmd.o cq.o eq.o fw.o fw_qos.o icm.o intf.o \
+		main.o mcg.o mr.o pd.o port.o profile.o qp.o reset.o sense.o \
+		srq.o resource_tracker.o
 
 obj-$(CONFIG_MLX4_EN)               += mlx4_en.o
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
index 8e3260c..f01918c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
@@ -35,6 +35,7 @@
 #include <linux/math64.h>
 
 #include "mlx4_en.h"
+#include "fw_qos.h"
 
 /* Definitions for QCN
  */
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.c b/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
new file mode 100644
index 0000000..0f5af7c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
@@ -0,0 +1,125 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies.
+ * All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/export.h>
+#include "fw_qos.h"
+
+struct mlx4_set_port_prio2tc_context {
+	u8 prio2tc[4];
+};
+
+struct mlx4_port_scheduler_tc_cfg_be {
+	__be16 pg;
+	__be16 bw_precentage;
+	__be16 max_bw_units; /* 3-100Mbps, 4-1Gbps, other values - reserved */
+	__be16 max_bw_value;
+};
+
+struct mlx4_set_port_scheduler_context {
+	struct mlx4_port_scheduler_tc_cfg_be tc[MLX4_NUM_TC];
+};
+
+int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_prio2tc_context *context;
+	int err;
+	u32 in_mod;
+	int i;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	context = mailbox->buf;
+
+	for (i = 0; i < MLX4_NUM_UP; i += 2)
+		context->prio2tc[i >> 1] = prio2tc[i] << 4 | prio2tc[i + 1];
+
+	in_mod = MLX4_SET_PORT_PRIO2TC << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_PRIO2TC);
+
+int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
+			    u8 *pg, u16 *ratelimit)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_scheduler_context *context;
+	int err;
+	u32 in_mod;
+	int i;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	context = mailbox->buf;
+
+	for (i = 0; i < MLX4_NUM_TC; i++) {
+		struct mlx4_port_scheduler_tc_cfg_be *tc = &context->tc[i];
+		u16 r;
+
+		if (ratelimit && ratelimit[i]) {
+			if (ratelimit[i] <= MLX4_MAX_100M_UNITS_VAL) {
+				r = ratelimit[i];
+				tc->max_bw_units =
+					htons(MLX4_RATELIMIT_100M_UNITS);
+			} else {
+				r = ratelimit[i] / 10;
+				tc->max_bw_units =
+					htons(MLX4_RATELIMIT_1G_UNITS);
+			}
+			tc->max_bw_value = htons(r);
+		} else {
+			tc->max_bw_value = htons(MLX4_RATELIMIT_DEFAULT);
+			tc->max_bw_units = htons(MLX4_RATELIMIT_1G_UNITS);
+		}
+
+		tc->pg = htons(pg[i]);
+		tc->bw_precentage = htons(tc_tx_bw[i]);
+	}
+
+	in_mod = MLX4_SET_PORT_SCHEDULER << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_SCHEDULER);
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
new file mode 100644
index 0000000..ab9be0f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
@@ -0,0 +1,75 @@
+/*
+ * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies.
+ * All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef MLX4_FW_QOS_H
+#define MLX4_FW_QOS_H
+
+#include <linux/mlx4/cmd.h>
+#include <linux/mlx4/device.h>
+
+#define MLX4_NUM_UP 8
+#define MLX4_NUM_TC 8
+
+/**
+ * mlx4_SET_PORT_PRIO2TC - This routine maps user priorities to traffic
+ * classes of a given port and device.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @prio2tc: Array of TC associated with each priorities.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc);
+
+/**
+ * mlx4_SET_PORT_SCHEDULER - This routine configures the arbitration between
+ * traffic classes (ETS) and configured rate limit for traffic classes.
+ * tc_tx_bw, pg and ratelimit are arrays where each index represents a TC.
+ * The description for those parameters below refers to a single TC.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @tc_tx_bw: The percentage of the bandwidth allocated for traffic class
+ *  within a TC group. The sum of the bw_percentage of all the traffic
+ *  classes within a TC group must equal 100% for correct operation.
+ * @pg: The TC group the traffic class is associated with.
+ * @ratelimit: The maximal bandwidth allowed for the use by this traffic class.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
+			    u8 *pg, u16 *ratelimit);
+
+#endif /* MLX4_FW_QOS_H */
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index 0b16db0..9e5236b 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -64,21 +64,6 @@
 
 #define INIT_HCA_TPT_MW_ENABLE          (1 << 7)
 
-struct mlx4_set_port_prio2tc_context {
-	u8 prio2tc[4];
-};
-
-struct mlx4_port_scheduler_tc_cfg_be {
-	__be16 pg;
-	__be16 bw_precentage;
-	__be16 max_bw_units; /* 3-100Mbps, 4-1Gbps, other values - reserved */
-	__be16 max_bw_value;
-};
-
-struct mlx4_set_port_scheduler_context {
-	struct mlx4_port_scheduler_tc_cfg_be tc[MLX4_NUM_TC];
-};
-
 enum {
 	MLX4_HCR_BASE		= 0x80680,
 	MLX4_HCR_SIZE		= 0x0001c,
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
index b97f173..6a53d42 100644
--- a/drivers/net/ethernet/mellanox/mlx4/port.c
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -1016,77 +1016,6 @@ int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
 }
 EXPORT_SYMBOL(mlx4_SET_PORT_qpn_calc);
 
-int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc)
-{
-	struct mlx4_cmd_mailbox *mailbox;
-	struct mlx4_set_port_prio2tc_context *context;
-	int err;
-	u32 in_mod;
-	int i;
-
-	mailbox = mlx4_alloc_cmd_mailbox(dev);
-	if (IS_ERR(mailbox))
-		return PTR_ERR(mailbox);
-	context = mailbox->buf;
-	for (i = 0; i < MLX4_NUM_UP; i += 2)
-		context->prio2tc[i >> 1] = prio2tc[i] << 4 | prio2tc[i + 1];
-
-	in_mod = MLX4_SET_PORT_PRIO2TC << 8 | port;
-	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
-		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
-
-	mlx4_free_cmd_mailbox(dev, mailbox);
-	return err;
-}
-EXPORT_SYMBOL(mlx4_SET_PORT_PRIO2TC);
-
-int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
-		u8 *pg, u16 *ratelimit)
-{
-	struct mlx4_cmd_mailbox *mailbox;
-	struct mlx4_set_port_scheduler_context *context;
-	int err;
-	u32 in_mod;
-	int i;
-
-	mailbox = mlx4_alloc_cmd_mailbox(dev);
-	if (IS_ERR(mailbox))
-		return PTR_ERR(mailbox);
-	context = mailbox->buf;
-
-	for (i = 0; i < MLX4_NUM_TC; i++) {
-		struct mlx4_port_scheduler_tc_cfg_be *tc = &context->tc[i];
-		u16 r;
-
-		if (ratelimit && ratelimit[i]) {
-			if (ratelimit[i] <= MLX4_MAX_100M_UNITS_VAL) {
-				r = ratelimit[i];
-				tc->max_bw_units =
-					htons(MLX4_RATELIMIT_100M_UNITS);
-			} else {
-				r = ratelimit[i]/10;
-				tc->max_bw_units =
-					htons(MLX4_RATELIMIT_1G_UNITS);
-			}
-			tc->max_bw_value = htons(r);
-		} else {
-			tc->max_bw_value = htons(MLX4_RATELIMIT_DEFAULT);
-			tc->max_bw_units = htons(MLX4_RATELIMIT_1G_UNITS);
-		}
-
-		tc->pg = htons(pg[i]);
-		tc->bw_precentage = htons(tc_tx_bw[i]);
-	}
-
-	in_mod = MLX4_SET_PORT_SCHEDULER << 8 | port;
-	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
-		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
-
-	mlx4_free_cmd_mailbox(dev, mailbox);
-	return err;
-}
-EXPORT_SYMBOL(mlx4_SET_PORT_SCHEDULER);
-
 enum {
 	VXLAN_ENABLE_MODIFY	= 1 << 7,
 	VXLAN_STEERING_MODIFY	= 1 << 6,
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 0f7f13d..b676d0c 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -49,8 +49,6 @@
 #define MSIX_LEGACY_SZ		4
 #define MIN_MSIX_P_PORT		5
 
-#define MLX4_NUM_UP			8
-#define MLX4_NUM_TC			8
 #define MLX4_MAX_100M_UNITS_VAL		255	/*
 						 * work around: can't set values
 						 * greater then this value when
@@ -1311,9 +1309,6 @@ int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu,
 			  u8 pptx, u8 pfctx, u8 pprx, u8 pfcrx);
 int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
 			   u8 promisc);
-int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc);
-int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
-		u8 *pg, u16 *ratelimit);
 int mlx4_SET_PORT_VXLAN(struct mlx4_dev *dev, u8 port, u8 steering, int enable);
 int mlx4_find_cached_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *idx);
 int mlx4_find_cached_vlan(struct mlx4_dev *dev, u8 port, u16 vid, int *idx);
-- 
1.7.1

[PATCH net-next 06/18] net/mlx4: Add mlx4_ALLOCATE_VPP implementation

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Implements device ALLOCATE_VPP command, to be used for granular QoS
configuration of VFs by the PF device. Defines and queries the amount
of VPPs assigned to each port, and the amount of VPPs assigned to each
priority of each port. Once the total VPPs are split between the priorities
of a port, they may be assigned with a share of the BW or a rate limit.

Split into two functions (get/set) whoch are supplied with
mlx4_alloc_vpp_context and physical port number.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/cmd.c    |    9 +++
 drivers/net/ethernet/mellanox/mlx4/fw_qos.c |   75 +++++++++++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/fw_qos.h |   28 ++++++++++
 include/linux/mlx4/cmd.h                    |    1 +
 4 files changed, 113 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index 52bf0b5..e1a191d 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -1455,6 +1455,15 @@ static struct mlx4_cmd_info cmd_info[] = {
 		.wrapper = mlx4_CMD_EPERM_wrapper,
 	},
 	{
+		.opcode = MLX4_CMD_ALLOCATE_VPP,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper,
+	},
+	{
 		.opcode = MLX4_CMD_CONF_SPECIAL_QP,
 		.has_inbox = false,
 		.has_outbox = false,
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.c b/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
index 0f5af7c..5ce3440 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
@@ -34,6 +34,13 @@
 
 #include <linux/export.h>
 #include "fw_qos.h"
+#include "fw.h"
+
+enum {
+	/* allocate vpp opcode modifiers */
+	MLX4_ALLOCATE_VPP_ALLOCATE	= 0x0,
+	MLX4_ALLOCATE_VPP_QUERY		= 0x1
+};
 
 struct mlx4_set_port_prio2tc_context {
 	u8 prio2tc[4];
@@ -50,6 +57,12 @@ struct mlx4_set_port_scheduler_context {
 	struct mlx4_port_scheduler_tc_cfg_be tc[MLX4_NUM_TC];
 };
 
+/* Granular Qos (per VF) section */
+struct mlx4_alloc_vpp_param {
+	__be32 availible_vpp;
+	__be32 vpp_p_up[MLX4_NUM_UP];
+};
+
 int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc)
 {
 	struct mlx4_cmd_mailbox *mailbox;
@@ -123,3 +136,65 @@ int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
 	return err;
 }
 EXPORT_SYMBOL(mlx4_SET_PORT_SCHEDULER);
+
+int mlx4_ALLOCATE_VPP_get(struct mlx4_dev *dev, u8 port,
+			  u16 *availible_vpp, u8 *vpp_p_up)
+{
+	int i;
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_alloc_vpp_param *out_param;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	out_param = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, port,
+			   MLX4_ALLOCATE_VPP_QUERY,
+			   MLX4_CMD_ALLOCATE_VPP,
+			   MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	/* Total number of supported VPPs */
+	*availible_vpp = (u16)be32_to_cpu(out_param->availible_vpp);
+
+	for (i = 0; i < MLX4_NUM_UP; i++)
+		vpp_p_up[i] = (u8)be32_to_cpu(out_param->vpp_p_up[i]);
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx4_ALLOCATE_VPP_get);
+
+int mlx4_ALLOCATE_VPP_set(struct mlx4_dev *dev, u8 port, u8 *vpp_p_up)
+{
+	int i;
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_alloc_vpp_param *in_param;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	in_param = mailbox->buf;
+
+	for (i = 0; i < MLX4_NUM_UP; i++)
+		in_param->vpp_p_up[i] = cpu_to_be32(vpp_p_up[i]);
+
+	err = mlx4_cmd(dev, mailbox->dma, port,
+		       MLX4_ALLOCATE_VPP_ALLOCATE,
+		       MLX4_CMD_ALLOCATE_VPP,
+		       MLX4_CMD_TIME_CLASS_A,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_ALLOCATE_VPP_set);
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
index ab9be0f..be79951 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
@@ -71,5 +71,33 @@ int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc);
  **/
 int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
 			    u8 *pg, u16 *ratelimit);
+/**
+ * mlx4_ALLOCATE_VPP_get - Query port VPP availible resources and allocation.
+ * Before distribution of VPPs to priorities, only availible_vpp is returned.
+ * After initialization it returns the distribution of VPPs among priorities.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @availible_vpp: Pointer to variable where number of availible VPPs is stored
+ * @vpp_p_up: Distribution of VPPs to priorities is stored in this array
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_ALLOCATE_VPP_get(struct mlx4_dev *dev, u8 port,
+			  u16 *availible_vpp, u8 *vpp_p_up);
+/**
+ * mlx4_ALLOCATE_VPP_set - Distribution of VPPs among differnt priorities.
+ * The total number of VPPs assigned to all for a port must not exceed
+ * the value reported by availible_vpp in mlx4_ALLOCATE_VPP_get.
+ * VPP allocation is allowed only after the port type has been set,
+ * and while no QPs are open for this port.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @vpp_p_up: Allocation of VPPs to different priorities.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_ALLOCATE_VPP_set(struct mlx4_dev *dev, u8 port, u8 *vpp_p_up);
 
 #endif /* MLX4_FW_QOS_H */
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index 7299e95..e0f88a0 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -68,6 +68,7 @@ enum {
 	MLX4_CMD_UNMAP_ICM_AUX	 = 0xffb,
 	MLX4_CMD_SET_ICM_SIZE	 = 0xffd,
 	MLX4_CMD_ACCESS_REG	 = 0x3b,
+	MLX4_CMD_ALLOCATE_VPP	 = 0x80,
 
 	/*master notify fw on finish for slave's flr*/
 	MLX4_CMD_INFORM_FLR_DONE = 0x5b,
-- 
1.7.1

[PATCH net-next 07/18] net/mlx4: Add mlx4_SET_VPORT_QOS implementation

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Add the SET_VPORT_QOS device command, which is ntended for virtual
granular QoS configuration per VF in SRIOV mode. The SET_VPORT_QOS
command sets and queries QoS parameters of a VPort. Each priority
allowed for a VPort is assigned with a share of the BW, and a BW
limitation. QoS parameters can be modified at any time, but must be
initialized before any QP is associated with the VPort.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/cmd.c    |    9 +++
 drivers/net/ethernet/mellanox/mlx4/fw_qos.c |   89 +++++++++++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/fw_qos.h |   36 +++++++++++
 include/linux/mlx4/cmd.h                    |    1 +
 4 files changed, 135 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index e1a191d..23dc98e 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -1464,6 +1464,15 @@ static struct mlx4_cmd_info cmd_info[] = {
 		.wrapper = mlx4_CMD_EPERM_wrapper,
 	},
 	{
+		.opcode = MLX4_CMD_SET_VPORT_QOS,
+		.has_inbox = false,
+		.has_outbox = true,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = mlx4_CMD_EPERM_wrapper,
+	},
+	{
 		.opcode = MLX4_CMD_CONF_SPECIAL_QP,
 		.has_inbox = false,
 		.has_outbox = false,
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.c b/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
index 5ce3440..8f2fde0 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.c
@@ -42,6 +42,12 @@ enum {
 	MLX4_ALLOCATE_VPP_QUERY		= 0x1
 };
 
+enum {
+	/* set vport qos opcode modifiers */
+	MLX4_SET_VPORT_QOS_SET		= 0x0,
+	MLX4_SET_VPORT_QOS_QUERY	= 0x1
+};
+
 struct mlx4_set_port_prio2tc_context {
 	u8 prio2tc[4];
 };
@@ -63,6 +69,19 @@ struct mlx4_alloc_vpp_param {
 	__be32 vpp_p_up[MLX4_NUM_UP];
 };
 
+struct mlx4_prio_qos_param {
+	__be32 bw_share;
+	__be32 max_avg_bw;
+	__be32 reserved;
+	__be32 enable;
+	__be32 reserved1[4];
+};
+
+struct mlx4_set_vport_context {
+	__be32 reserved[8];
+	struct mlx4_prio_qos_param qos_p_up[MLX4_NUM_UP];
+};
+
 int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc)
 {
 	struct mlx4_cmd_mailbox *mailbox;
@@ -198,3 +217,73 @@ int mlx4_ALLOCATE_VPP_set(struct mlx4_dev *dev, u8 port, u8 *vpp_p_up)
 	return err;
 }
 EXPORT_SYMBOL(mlx4_ALLOCATE_VPP_set);
+
+int mlx4_SET_VPORT_QOS_get(struct mlx4_dev *dev, u8 port, u8 vport,
+			   struct mlx4_vport_qos_param *out_param)
+{
+	int i;
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_vport_context *ctx;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	ctx = mailbox->buf;
+
+	err = mlx4_cmd_box(dev, 0, mailbox->dma, (vport << 8) | port,
+			   MLX4_SET_VPORT_QOS_QUERY,
+			   MLX4_CMD_SET_VPORT_QOS,
+			   MLX4_CMD_TIME_CLASS_A,
+			   MLX4_CMD_NATIVE);
+	if (err)
+		goto out;
+
+	for (i = 0; i < MLX4_NUM_UP; i++) {
+		out_param[i].bw_share = be32_to_cpu(ctx->qos_p_up[i].bw_share);
+		out_param[i].max_avg_bw =
+			be32_to_cpu(ctx->qos_p_up[i].max_avg_bw);
+		out_param[i].enable =
+			!!(be32_to_cpu(ctx->qos_p_up[i].enable) & 31);
+	}
+
+out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_VPORT_QOS_get);
+
+int mlx4_SET_VPORT_QOS_set(struct mlx4_dev *dev, u8 port, u8 vport,
+			   struct mlx4_vport_qos_param *in_param)
+{
+	int i;
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_vport_context *ctx;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	ctx = mailbox->buf;
+
+	for (i = 0; i < MLX4_NUM_UP; i++) {
+		ctx->qos_p_up[i].bw_share = cpu_to_be32(in_param[i].bw_share);
+		ctx->qos_p_up[i].max_avg_bw =
+				cpu_to_be32(in_param[i].max_avg_bw);
+		ctx->qos_p_up[i].enable =
+				cpu_to_be32(in_param[i].enable << 31);
+	}
+
+	err = mlx4_cmd(dev, mailbox->dma, (vport << 8) | port,
+		       MLX4_SET_VPORT_QOS_SET,
+		       MLX4_CMD_SET_VPORT_QOS,
+		       MLX4_CMD_TIME_CLASS_A,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_VPORT_QOS_set);
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
index be79951..b3fffaf 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
@@ -41,6 +41,12 @@
 #define MLX4_NUM_UP 8
 #define MLX4_NUM_TC 8
 
+struct mlx4_vport_qos_param {
+	u32 bw_share;
+	u32 max_avg_bw;
+	u8 enable;
+};
+
 /**
  * mlx4_SET_PORT_PRIO2TC - This routine maps user priorities to traffic
  * classes of a given port and device.
@@ -100,4 +106,34 @@ int mlx4_ALLOCATE_VPP_get(struct mlx4_dev *dev, u8 port,
  **/
 int mlx4_ALLOCATE_VPP_set(struct mlx4_dev *dev, u8 port, u8 *vpp_p_up);
 
+/**
+ * mlx4_SET_VPORT_QOS_get - Query QoS proporties of a Vport.
+ * Each priority allowed for the Vport is assigned with a share of the BW,
+ * and a BW limitation. This commands query the current QoS values.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @vport: Vport id.
+ * @out_param: Array of mlx4_vport_qos_param that will contain the values.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_SET_VPORT_QOS_get(struct mlx4_dev *dev, u8 port, u8 vport,
+			   struct mlx4_vport_qos_param *out_param);
+
+/**
+ * mlx4_SET_VPORT_QOS_set - Set QoS proporties of a Vport.
+ * QoS parameters can be modified at any time, but must be initialized
+ * before any QP is associated with the VPort.
+ *
+ * @dev: mlx4_dev.
+ * @port: Physical port number.
+ * @vport: Vport id.
+ * @out_param: Array of mlx4_vport_qos_param which holds the requested values.
+ *
+ * Returns 0 on success or a negative mlx4_core errno code.
+ **/
+int mlx4_SET_VPORT_QOS_set(struct mlx4_dev *dev, u8 port, u8 vport,
+			   struct mlx4_vport_qos_param *in_param);
+
 #endif /* MLX4_FW_QOS_H */
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index e0f88a0..88326ed 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -69,6 +69,7 @@ enum {
 	MLX4_CMD_SET_ICM_SIZE	 = 0xffd,
 	MLX4_CMD_ACCESS_REG	 = 0x3b,
 	MLX4_CMD_ALLOCATE_VPP	 = 0x80,
+	MLX4_CMD_SET_VPORT_QOS	 = 0x81,
 
 	/*master notify fw on finish for slave's flr*/
 	MLX4_CMD_INFORM_FLR_DONE = 0x5b,
-- 
1.7.1

[PATCH net-next 08/18] net/mlx4: Query device for QoS per VF support

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Checks in QUERY_DEV_CAP if the granular QoS per VF feature is
supported by the device. Disabled for guests.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/fw.c |   10 +++++++++-
 include/linux/mlx4/device.h             |    3 ++-
 2 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index 412e019..94e857e 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -147,7 +147,8 @@ static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags)
 		[21] = "Port Remap support",
 		[22] = "QCN support",
 		[23] = "QP rate limiting support",
-		[24] = "Ethernet Flow control statistics support"
+		[24] = "Ethernet Flow control statistics support",
+		[25] = "Granular QoS per VF support",
 	};
 	int i;
 
@@ -871,6 +872,8 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	MLX4_GET(size, outbox, QUERY_DEV_CAP_MAX_DESC_SZ_RQ_OFFSET);
 	dev_cap->max_rq_desc_sz = size;
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_CQ_EQ_CACHE_LINE_STRIDE);
+	if (field & (1 << 4))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_QOS_VPP;
 	if (field & (1 << 5))
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_ETH_PROT_CTRL;
 	if (field & (1 << 6))
@@ -1199,6 +1202,11 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave,
 	field16 = 0;
 	MLX4_PUT(outbox->buf, field16, QUERY_DEV_CAP_QP_RATE_LIMIT_NUM_OFFSET);
 
+	/* turn off QoS per VF support for guests */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_CQ_EQ_CACHE_LINE_STRIDE);
+	field &= 0xef;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_CQ_EQ_CACHE_LINE_STRIDE);
+
 	return 0;
 }
 
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index b676d0c..c37208f 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -205,7 +205,8 @@ enum {
 	MLX4_DEV_CAP_FLAG2_PORT_REMAP		= 1LL <<  21,
 	MLX4_DEV_CAP_FLAG2_QCN			= 1LL <<  22,
 	MLX4_DEV_CAP_FLAG2_QP_RATE_LIMIT	= 1LL <<  23,
-	MLX4_DEV_CAP_FLAG2_FLOWSTATS_EN         = 1LL <<  24
+	MLX4_DEV_CAP_FLAG2_FLOWSTATS_EN         = 1LL <<  24,
+	MLX4_DEV_CAP_FLAG2_QOS_VPP		= 1LL <<  25,
 };
 
 enum {
-- 
1.7.1

[PATCH net-next 09/18] net/mlx4: Allocate VPPs for each port on PF init

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Initialization of granular Qos per VF mechanism.

Query the port availible VPPs and allocates those on all supported
priorities in an equal share. Allocation is done only in SRIOV mode,
when the feature is supported by the device and port type is Ethernet.

Allocation currently is done only on the default priority 0.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/cmd.c    |   66 +++++++++++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/fw_qos.h |    3 +
 drivers/net/ethernet/mellanox/mlx4/mlx4.h   |    7 +++
 3 files changed, 76 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index 23dc98e..1e4a929 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -1883,6 +1883,63 @@ static int mlx4_master_immediate_activate_vlan_qos(struct mlx4_priv *priv,
 	return 0;
 }
 
+static void mlx4_set_default_port_qos(struct mlx4_dev *dev, int port)
+{
+	struct mlx4_qos_manager *port_qos_ctl;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	port_qos_ctl = &priv->mfunc.master.qos_ctl[port];
+	bitmap_zero(port_qos_ctl->priority_bm, MLX4_NUM_UP);
+
+	/* Enable only default prio at PF init routine */
+	set_bit(MLX4_DEFAULT_QOS_PRIO, port_qos_ctl->priority_bm);
+}
+
+static void mlx4_allocate_port_vpps(struct mlx4_dev *dev, int port)
+{
+	int i;
+	int err;
+	int num_vfs;
+	u16 availible_vpp;
+	u8 vpp_param[MLX4_NUM_UP];
+	struct mlx4_qos_manager *port_qos;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	err = mlx4_ALLOCATE_VPP_get(dev, port, &availible_vpp, vpp_param);
+	if (err) {
+		mlx4_info(dev, "Failed query availible VPPs\n");
+		return;
+	}
+
+	port_qos = &priv->mfunc.master.qos_ctl[port];
+	num_vfs = (availible_vpp /
+		   bitmap_weight(port_qos->priority_bm, MLX4_NUM_UP));
+
+	for (i = 0; i < MLX4_NUM_UP; i++) {
+		if (test_bit(i, port_qos->priority_bm))
+			vpp_param[i] = num_vfs;
+	}
+
+	err = mlx4_ALLOCATE_VPP_set(dev, port, vpp_param);
+	if (err) {
+		mlx4_info(dev, "Failed allocating VPPs\n");
+		return;
+	}
+
+	/* Query actual allocated VPP, just to make sure */
+	err = mlx4_ALLOCATE_VPP_get(dev, port, &availible_vpp, vpp_param);
+	if (err) {
+		mlx4_info(dev, "Failed query availible VPPs\n");
+		return;
+	}
+
+	port_qos->num_of_qos_vfs = num_vfs;
+	mlx4_dbg(dev, "Port %d Availible VPPs %d\n", port, availible_vpp);
+
+	for (i = 0; i < MLX4_NUM_UP; i++)
+		mlx4_dbg(dev, "Port %d UP %d Allocated %d VPPs\n", port, i,
+			 vpp_param[i]);
+}
 
 static int mlx4_master_activate_admin_state(struct mlx4_priv *priv, int slave)
 {
@@ -2288,6 +2345,15 @@ int mlx4_multi_func_init(struct mlx4_dev *dev)
 			spin_lock_init(&s_state->lock);
 		}
 
+		if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QOS_VPP) {
+			for (port = 1; port <= dev->caps.num_ports; port++) {
+				if (mlx4_is_eth(dev, port)) {
+					mlx4_set_default_port_qos(dev, port);
+					mlx4_allocate_port_vpps(dev, port);
+				}
+			}
+		}
+
 		memset(&priv->mfunc.master.cmd_eqe, 0, dev->caps.eqe_size);
 		priv->mfunc.master.cmd_eqe.type = MLX4_EVENT_TYPE_CMD;
 		INIT_WORK(&priv->mfunc.master.comm_work,
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
index b3fffaf..4a5320f 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
@@ -41,6 +41,9 @@
 #define MLX4_NUM_UP 8
 #define MLX4_NUM_TC 8
 
+/* Default supported priorities for VPP allocation */
+#define MLX4_DEFAULT_QOS_PRIO (0)
+
 struct mlx4_vport_qos_param {
 	u32 bw_share;
 	u32 max_avg_bw;
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index 9e5236b..1add698 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -50,6 +50,7 @@
 #include <linux/mlx4/driver.h>
 #include <linux/mlx4/doorbell.h>
 #include <linux/mlx4/cmd.h>
+#include "fw_qos.h"
 
 #define DRV_NAME	"mlx4_core"
 #define PFX		DRV_NAME ": "
@@ -553,6 +554,11 @@ struct mlx4_slave_event_eq {
 	struct mlx4_eqe event_eqe[SLAVE_EVENT_EQ_SIZE];
 };
 
+struct mlx4_qos_manager {
+	int num_of_qos_vfs;
+	DECLARE_BITMAP(priority_bm, MLX4_NUM_UP);
+};
+
 struct mlx4_master_qp0_state {
 	int proxy_qp0_active;
 	int qp0_active;
@@ -577,6 +583,7 @@ struct mlx4_mfunc_master_ctx {
 	struct mlx4_eqe		cmd_eqe;
 	struct mlx4_slave_event_eq slave_eq;
 	struct mutex		gen_eqe_mutex[MLX4_MFUNC_MAX];
+	struct mlx4_qos_manager qos_ctl[MLX4_MAX_PORTS + 1];
 };
 
 struct mlx4_mfunc {
-- 
1.7.1

[PATCH net-next 10/18] net/mlx4: Added qos_vport QP configuration in VST mode

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Granular QoS per VF feature introduce a new QP field, qos_vport.

PF administrator can connect VF QPs to a certain QoS Vport, to
inherit its proporties. Connecting QPs to the default QoS Vport
(defined as 0) is always allowed, even when there are no allocated VPPs.
At this point, only the default vport is connected to QPs.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/cmd.c           |    9 ++++++++-
 drivers/net/ethernet/mellanox/mlx4/fw_qos.h        |    3 +++
 drivers/net/ethernet/mellanox/mlx4/mlx4.h          |    2 ++
 drivers/net/ethernet/mellanox/mlx4/qp.c            |    5 +++++
 .../net/ethernet/mellanox/mlx4/resource_tracker.c  |    6 ++++++
 include/linux/mlx4/qp.h                            |    8 ++++++--
 6 files changed, 30 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index 1e4a929..ea8947a 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -48,6 +48,7 @@
 
 #include "mlx4.h"
 #include "fw.h"
+#include "fw_qos.h"
 
 #define CMD_POLL_TOKEN 0xffff
 #define INBOX_MASK	0xffffffffffffff00ULL
@@ -1808,7 +1809,8 @@ static int mlx4_master_immediate_activate_vlan_qos(struct mlx4_priv *priv,
 
 	if (vp_oper->state.default_vlan == vp_admin->default_vlan &&
 	    vp_oper->state.default_qos == vp_admin->default_qos &&
-	    vp_oper->state.link_state == vp_admin->link_state)
+	    vp_oper->state.link_state == vp_admin->link_state &&
+	    vp_oper->state.qos_vport == vp_admin->qos_vport)
 		return 0;
 
 	if (!(priv->mfunc.master.slave_state[slave].active &&
@@ -1866,6 +1868,7 @@ static int mlx4_master_immediate_activate_vlan_qos(struct mlx4_priv *priv,
 	vp_oper->state.default_vlan = vp_admin->default_vlan;
 	vp_oper->state.default_qos = vp_admin->default_qos;
 	vp_oper->state.link_state = vp_admin->link_state;
+	vp_oper->state.qos_vport = vp_admin->qos_vport;
 
 	if (vp_admin->link_state == IFLA_VF_LINK_STATE_DISABLE)
 		work->flags |= MLX4_VF_IMMED_VLAN_FLAG_LINK_DISABLE;
@@ -1874,6 +1877,7 @@ static int mlx4_master_immediate_activate_vlan_qos(struct mlx4_priv *priv,
 	work->port = port;
 	work->slave = slave;
 	work->qos = vp_oper->state.default_qos;
+	work->qos_vport = vp_oper->state.qos_vport;
 	work->vlan_id = vp_oper->state.default_vlan;
 	work->vlan_ix = vp_oper->vlan_idx;
 	work->priv = priv;
@@ -2339,6 +2343,9 @@ int mlx4_multi_func_init(struct mlx4_dev *dev)
 				INIT_LIST_HEAD(&s_state->mcast_filters[port]);
 				admin_vport->default_vlan = MLX4_VGT;
 				oper_vport->default_vlan = MLX4_VGT;
+				admin_vport->qos_vport =
+						MLX4_VPP_DEFAULT_VPORT;
+				oper_vport->qos_vport = MLX4_VPP_DEFAULT_VPORT;
 				vf_oper->vport[port].vlan_idx = NO_INDX;
 				vf_oper->vport[port].mac_idx = NO_INDX;
 			}
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
index 4a5320f..ac1f331 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
+++ b/drivers/net/ethernet/mellanox/mlx4/fw_qos.h
@@ -44,6 +44,9 @@
 /* Default supported priorities for VPP allocation */
 #define MLX4_DEFAULT_QOS_PRIO (0)
 
+/* Derived from FW feature definition, 0 is the default vport fo all QPs */
+#define MLX4_VPP_DEFAULT_VPORT (0)
+
 struct mlx4_vport_qos_param {
 	u32 bw_share;
 	u32 max_avg_bw;
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index 1add698..32469c6 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -498,6 +498,7 @@ struct mlx4_vport_state {
 	u32 tx_rate;
 	bool spoofchk;
 	u32 link_state;
+	u8 qos_vport;
 };
 
 struct mlx4_vf_admin_state {
@@ -636,6 +637,7 @@ struct mlx4_vf_immed_vlan_work {
 	int			orig_vlan_ix;
 	u8			port;
 	u8			qos;
+	u8                      qos_vport;
 	u16			vlan_id;
 	u16			orig_vlan_id;
 };
diff --git a/drivers/net/ethernet/mellanox/mlx4/qp.c b/drivers/net/ethernet/mellanox/mlx4/qp.c
index 69e4462..b75214a 100644
--- a/drivers/net/ethernet/mellanox/mlx4/qp.c
+++ b/drivers/net/ethernet/mellanox/mlx4/qp.c
@@ -447,6 +447,11 @@ int mlx4_update_qp(struct mlx4_dev *dev, u32 qpn,
 		cmd->qp_context.rate_limit_params = cpu_to_be16((params->rate_unit << 14) | params->rate_val);
 	}
 
+	if (attr & MLX4_UPDATE_QP_QOS_VPORT) {
+		qp_mask |= 1ULL << MLX4_UPD_QP_MASK_QOS_VPP;
+		cmd->qp_context.qos_vport = params->qos_vport;
+	}
+
 	cmd->primary_addr_path_mask = cpu_to_be64(pri_addr_path_mask);
 	cmd->qp_mask = cpu_to_be64(qp_mask);
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
index 2076e98..c82df82 100644
--- a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
+++ b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
@@ -765,6 +765,7 @@ static int update_vport_qp_param(struct mlx4_dev *dev,
 		qpc->pri_path.feup |= MLX4_FEUP_FORCE_ETH_UP | MLX4_FVL_FORCE_ETH_VLAN;
 		qpc->pri_path.sched_queue &= 0xC7;
 		qpc->pri_path.sched_queue |= (vp_oper->state.default_qos) << 3;
+		qpc->qos_vport = vp_oper->state.qos_vport;
 	}
 	if (vp_oper->state.spoofchk) {
 		qpc->pri_path.feup |= MLX4_FSM_FORCE_ETH_SRC_MAC;
@@ -4911,6 +4912,11 @@ void mlx4_vf_immed_vlan_work_handler(struct work_struct *_work)
 					qp->sched_queue & 0xC7;
 				upd_context->qp_context.pri_path.sched_queue |=
 					((work->qos & 0x7) << 3);
+				upd_context->qp_mask |=
+					cpu_to_be64(1ULL <<
+						    MLX4_UPD_QP_MASK_QOS_VPP);
+				upd_context->qp_context.qos_vport =
+					work->qos_vport;
 			}
 
 			err = mlx4_cmd(dev, mailbox->dma,
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index 1023ebe..6fed539 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -209,7 +209,8 @@ struct mlx4_qp_context {
 	__be16			sq_wqe_counter;
 	u32			reserved3;
 	__be16			rate_limit_params;
-	__be16			reserved4;
+	u8			reserved4;
+	u8			qos_vport;
 	__be32			param3;
 	__be32			nummmcpeers_basemkey;
 	u8			log_page_size;
@@ -231,6 +232,7 @@ struct mlx4_update_qp_context {
 enum {
 	MLX4_UPD_QP_MASK_PM_STATE	= 32,
 	MLX4_UPD_QP_MASK_VSD		= 33,
+	MLX4_UPD_QP_MASK_QOS_VPP	= 34,
 	MLX4_UPD_QP_MASK_RATE_LIMIT	= 35,
 };
 
@@ -432,7 +434,8 @@ enum mlx4_update_qp_attr {
 	MLX4_UPDATE_QP_SMAC		= 1 << 0,
 	MLX4_UPDATE_QP_VSD		= 1 << 1,
 	MLX4_UPDATE_QP_RATE_LIMIT	= 1 << 2,
-	MLX4_UPDATE_QP_SUPPORTED_ATTRS	= (1 << 3) - 1
+	MLX4_UPDATE_QP_QOS_VPORT	= 1 << 3,
+	MLX4_UPDATE_QP_SUPPORTED_ATTRS	= (1 << 4) - 1
 };
 
 enum mlx4_update_qp_params_flags {
@@ -441,6 +444,7 @@ enum mlx4_update_qp_params_flags {
 
 struct mlx4_update_qp_params {
 	u8	smac_index;
+	u8	qos_vport;
 	u32	flags;
 	u16	rate_unit;
 	u16	rate_val;
-- 
1.7.1

[PATCH net-next 11/18] net/mlx4_en: Enable TX rate limit per VF

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Support granular QoS per VF, by implementing the ndo_set_vf_rate.

Enforce a rate limit per VF when called, and enabled only for VFs in
VST mode with user priority supported by the device.

We don't enforce VFs to be in VST mode at the moment of configuration,
but rather save the given rate limit and enforce it when the VF is
moved to VST with user priority which is supported (currently 0).

VST<->VGT or VST qos value state changes are disallowed when a rate
limit is configured. Minimum BW share is not supported yet.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/cmd.c       |  177 +++++++++++++++++++++++-
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c |   11 ++
 include/linux/mlx4/cmd.h                       |    2 +
 3 files changed, 189 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index ea8947a..bfc3584 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -2781,6 +2781,103 @@ static int mlx4_slaves_closest_port(struct mlx4_dev *dev, int slave, int port)
 	return port;
 }
 
+static int mlx4_set_vport_qos(struct mlx4_priv *priv, int slave, int port,
+			      int max_tx_rate)
+{
+	int i;
+	int err;
+	struct mlx4_qos_manager *port_qos;
+	struct mlx4_dev *dev = &priv->dev;
+	struct mlx4_vport_qos_param vpp_qos[MLX4_NUM_UP];
+
+	port_qos = &priv->mfunc.master.qos_ctl[port];
+	memset(vpp_qos, 0, sizeof(struct mlx4_vport_qos_param) * MLX4_NUM_UP);
+
+	if (slave > port_qos->num_of_qos_vfs) {
+		mlx4_info(dev, "No availible VPP resources for this VF\n");
+		return -EINVAL;
+	}
+
+	/* Query for default QoS values from Vport 0 is needed */
+	err = mlx4_SET_VPORT_QOS_get(dev, port, 0, vpp_qos);
+	if (err) {
+		mlx4_info(dev, "Failed to query Vport 0 QoS values\n");
+		return err;
+	}
+
+	for (i = 0; i < MLX4_NUM_UP; i++) {
+		if (test_bit(i, port_qos->priority_bm) && max_tx_rate) {
+			vpp_qos[i].max_avg_bw = max_tx_rate;
+			vpp_qos[i].enable = 1;
+		} else {
+			/* if user supplied tx_rate == 0, meaning no rate limit
+			 * configuration is required. so we are leaving the
+			 * value of max_avg_bw as queried from Vport 0.
+			 */
+			vpp_qos[i].enable = 0;
+		}
+	}
+
+	err = mlx4_SET_VPORT_QOS_set(dev, port, slave, vpp_qos);
+	if (err) {
+		mlx4_info(dev, "Failed to set Vport %d QoS values\n", slave);
+		return err;
+	}
+
+	return 0;
+}
+
+static bool mlx4_is_vf_vst_and_prio_qos(struct mlx4_dev *dev, int port,
+					struct mlx4_vport_state *vf_admin)
+{
+	struct mlx4_qos_manager *info;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (!mlx4_is_master(dev) ||
+	    !(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QOS_VPP))
+		return false;
+
+	info = &priv->mfunc.master.qos_ctl[port];
+
+	if (vf_admin->default_vlan != MLX4_VGT &&
+	    test_bit(vf_admin->default_qos, info->priority_bm))
+		return true;
+
+	return false;
+}
+
+static bool mlx4_valid_vf_state_change(struct mlx4_dev *dev, int port,
+				       struct mlx4_vport_state *vf_admin,
+				       int vlan, int qos)
+{
+	struct mlx4_vport_state dummy_admin = {0};
+
+	if (!mlx4_is_vf_vst_and_prio_qos(dev, port, vf_admin) ||
+	    !vf_admin->tx_rate)
+		return true;
+
+	dummy_admin.default_qos = qos;
+	dummy_admin.default_vlan = vlan;
+
+	/* VF wants to move to other VST state which is valid with current
+	 * rate limit. Either differnt default vlan in VST or other
+	 * supported QoS priority. Otherwise we don't allow this change when
+	 * the TX rate is still configured.
+	 */
+	if (mlx4_is_vf_vst_and_prio_qos(dev, port, &dummy_admin))
+		return true;
+
+	mlx4_info(dev, "Cannot change VF state to %s while rate is set\n",
+		  (vlan == MLX4_VGT) ? "VGT" : "VST");
+
+	if (vlan != MLX4_VGT)
+		mlx4_info(dev, "VST priority %d not supported for QoS\n", qos);
+
+	mlx4_info(dev, "Please set rate to 0 prior to this VF state change\n");
+
+	return false;
+}
+
 int mlx4_set_vf_mac(struct mlx4_dev *dev, int port, int vf, u64 mac)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
@@ -2824,12 +2921,22 @@ int mlx4_set_vf_vlan(struct mlx4_dev *dev, int port, int vf, u16 vlan, u8 qos)
 	port = mlx4_slaves_closest_port(dev, slave, port);
 	vf_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
 
+	if (!mlx4_valid_vf_state_change(dev, port, vf_admin, vlan, qos))
+		return -EPERM;
+
 	if ((0 == vlan) && (0 == qos))
 		vf_admin->default_vlan = MLX4_VGT;
 	else
 		vf_admin->default_vlan = vlan;
 	vf_admin->default_qos = qos;
 
+	/* If rate was configured prior to VST, we saved the configured rate
+	 * in vf_admin->rate and now, if priority supported we enforce the QoS
+	 */
+	if (mlx4_is_vf_vst_and_prio_qos(dev, port, vf_admin) &&
+	    vf_admin->tx_rate)
+		vf_admin->qos_vport = slave;
+
 	if (mlx4_master_immediate_activate_vlan_qos(priv, slave, port))
 		mlx4_info(dev,
 			  "updating vf %d port %d config will take effect on next VF restart\n",
@@ -2838,6 +2945,69 @@ int mlx4_set_vf_vlan(struct mlx4_dev *dev, int port, int vf, u16 vlan, u8 qos)
 }
 EXPORT_SYMBOL_GPL(mlx4_set_vf_vlan);
 
+int mlx4_set_vf_rate(struct mlx4_dev *dev, int port, int vf, int min_tx_rate,
+		     int max_tx_rate)
+{
+	int err;
+	int slave;
+	struct mlx4_vport_state *vf_admin;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (!mlx4_is_master(dev) ||
+	    !(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_QOS_VPP))
+		return -EPROTONOSUPPORT;
+
+	if (min_tx_rate) {
+		mlx4_info(dev, "Minimum BW share not supported\n");
+		return -EPROTONOSUPPORT;
+	}
+
+	slave = mlx4_get_slave_indx(dev, vf);
+	if (slave < 0)
+		return -EINVAL;
+
+	port = mlx4_slaves_closest_port(dev, slave, port);
+	vf_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
+
+	err = mlx4_set_vport_qos(priv, slave, port, max_tx_rate);
+	if (err) {
+		mlx4_info(dev, "vf %d failed to set rate %d\n", vf,
+			  max_tx_rate);
+		return err;
+	}
+
+	vf_admin->tx_rate = max_tx_rate;
+	/* if VF is not in supported mode (VST with supported prio),
+	 * we do not change vport configuration for its QPs, but save
+	 * the rate, so it will be enforced when it moves to supported
+	 * mode next time.
+	 */
+	if (!mlx4_is_vf_vst_and_prio_qos(dev, port, vf_admin)) {
+		mlx4_info(dev,
+			  "rate set for VF %d when not in valid state\n", vf);
+
+		if (vf_admin->default_vlan != MLX4_VGT)
+			mlx4_info(dev, "VST priority not supported by QoS\n");
+		else
+			mlx4_info(dev, "VF in VGT mode (needed VST)\n");
+
+		mlx4_info(dev,
+			  "rate %d take affect when VF moves to valid state\n",
+			  max_tx_rate);
+		return 0;
+	}
+
+	/* If user sets rate 0 assigning default vport for its QPs */
+	vf_admin->qos_vport = max_tx_rate ? slave : MLX4_VPP_DEFAULT_VPORT;
+
+	if (priv->mfunc.master.slave_state[slave].active &&
+	    dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_UPDATE_QP)
+		mlx4_master_immediate_activate_vlan_qos(priv, slave, port);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_set_vf_rate);
+
  /* mlx4_get_slave_default_vlan -
  * return true if VST ( default vlan)
  * if VST, will return vlan & qos (if not NULL)
@@ -2911,7 +3081,12 @@ int mlx4_get_vf_config(struct mlx4_dev *dev, int port, int vf, struct ifla_vf_in
 
 	ivf->vlan		= s_info->default_vlan;
 	ivf->qos		= s_info->default_qos;
-	ivf->max_tx_rate	= s_info->tx_rate;
+
+	if (mlx4_is_vf_vst_and_prio_qos(dev, port, s_info))
+		ivf->max_tx_rate = s_info->tx_rate;
+	else
+		ivf->max_tx_rate = 0;
+
 	ivf->min_tx_rate	= 0;
 	ivf->spoofchk		= s_info->spoofchk;
 	ivf->linkstate		= s_info->link_state;
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 4ee564c..0fd5ec9 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -2240,6 +2240,16 @@ static int mlx4_en_set_vf_vlan(struct net_device *dev, int vf, u16 vlan, u8 qos)
 	return mlx4_set_vf_vlan(mdev->dev, en_priv->port, vf, vlan, qos);
 }
 
+static int mlx4_en_set_vf_rate(struct net_device *dev, int vf, int min_tx_rate,
+			       int max_tx_rate)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_set_vf_rate(mdev->dev, en_priv->port, vf, min_tx_rate,
+				max_tx_rate);
+}
+
 static int mlx4_en_set_vf_spoofchk(struct net_device *dev, int vf, bool setting)
 {
 	struct mlx4_en_priv *en_priv = netdev_priv(dev);
@@ -2458,6 +2468,7 @@ static const struct net_device_ops mlx4_netdev_ops_master = {
 	.ndo_vlan_rx_kill_vid	= mlx4_en_vlan_rx_kill_vid,
 	.ndo_set_vf_mac		= mlx4_en_set_vf_mac,
 	.ndo_set_vf_vlan	= mlx4_en_set_vf_vlan,
+	.ndo_set_vf_rate	= mlx4_en_set_vf_rate,
 	.ndo_set_vf_spoofchk	= mlx4_en_set_vf_spoofchk,
 	.ndo_set_vf_link_state	= mlx4_en_set_vf_link_state,
 	.ndo_get_vf_config	= mlx4_en_get_vf_config,
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index 88326ed..a788839 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -296,6 +296,8 @@ void mlx4_free_cmd_mailbox(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbo
 u32 mlx4_comm_get_version(void);
 int mlx4_set_vf_mac(struct mlx4_dev *dev, int port, int vf, u64 mac);
 int mlx4_set_vf_vlan(struct mlx4_dev *dev, int port, int vf, u16 vlan, u8 qos);
+int mlx4_set_vf_rate(struct mlx4_dev *dev, int port, int vf, int min_tx_rate,
+		     int max_tx_rate);
 int mlx4_set_vf_spoofchk(struct mlx4_dev *dev, int port, int vf, bool setting);
 int mlx4_get_vf_config(struct mlx4_dev *dev, int port, int vf, struct ifla_vf_info *ivf);
 int mlx4_set_vf_link_state(struct mlx4_dev *dev, int port, int vf, int link_state);
-- 
1.7.1

[PATCH net-next 12/18] net/mlx4: Warn users of depracated QoS Firmware

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

A new capability bit was introduced in the past to to differ devices
using the QoS ETS feature. The old was deprecated since then.
If driver sees device which set only the old capabilty, it will print
warning to user suggesting to upgrade the FW.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c |    2 +-
 drivers/net/ethernet/mellanox/mlx4/fw.c        |    7 +++++--
 drivers/net/ethernet/mellanox/mlx4/main.c      |    8 ++++++++
 include/linux/mlx4/device.h                    |    1 +
 4 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 0fd5ec9..b02a2e8 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -2814,7 +2814,7 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	priv->msg_enable = MLX4_EN_MSG_LEVEL;
 #ifdef CONFIG_MLX4_EN_DCB
 	if (!mlx4_is_slave(priv->mdev->dev)) {
-		if (mdev->dev->caps.flags & MLX4_DEV_CAP_FLAG_SET_ETH_SCHED) {
+		if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ETS_CFG) {
 			dev->dcbnl_ops = &mlx4_en_dcbnl_ops;
 		} else {
 			en_info(priv, "enabling only PFC DCB ops\n");
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index 94e857e..16bbd6c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -149,6 +149,7 @@ static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags)
 		[23] = "QP rate limiting support",
 		[24] = "Ethernet Flow control statistics support",
 		[25] = "Granular QoS per VF support",
+		[26] = "Port ETS Scheduler support",
 	};
 	int i;
 
@@ -900,6 +901,8 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_VXLAN);
 	if (field & 1<<3)
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS;
+	if (field & (1 << 5))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_ETS_CFG;
 	MLX4_GET(dev_cap->max_icm_sz, outbox,
 		 QUERY_DEV_CAP_MAX_ICM_SZ_OFFSET);
 	if (dev_cap->flags & MLX4_DEV_CAP_FLAG_COUNTERS)
@@ -1157,9 +1160,9 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave,
 	field &= 0x7f;
 	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_CQ_TS_SUPPORT_OFFSET);
 
-	/* For guests, disable vxlan tunneling */
+	/* For guests, disable vxlan tunneling and QoS support */
 	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_VXLAN);
-	field &= 0xf7;
+	field &= 0xd7;
 	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_VXLAN);
 
 	/* For guests, report Blueflame disabled */
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index ee0a670..3becad3 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -528,6 +528,14 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 		dev->caps.alloc_res_qp_mask =
 			(dev->caps.bf_reg_size ? MLX4_RESERVE_ETH_BF_QP : 0) |
 			MLX4_RESERVE_A0_QP;
+
+		if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ETS_CFG) &&
+		    dev->caps.flags & MLX4_DEV_CAP_FLAG_SET_ETH_SCHED) {
+			mlx4_warn(dev, "Old device ETS support detected\n");
+			mlx4_warn(dev, "Consider upgrading device FW.\n");
+			dev->caps.flags2 |= MLX4_DEV_CAP_FLAG2_ETS_CFG;
+		}
+
 	} else {
 		dev->caps.alloc_res_qp_mask = 0;
 	}
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index c37208f..15f4676 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -207,6 +207,7 @@ enum {
 	MLX4_DEV_CAP_FLAG2_QP_RATE_LIMIT	= 1LL <<  23,
 	MLX4_DEV_CAP_FLAG2_FLOWSTATS_EN         = 1LL <<  24,
 	MLX4_DEV_CAP_FLAG2_QOS_VPP		= 1LL <<  25,
+	MLX4_DEV_CAP_FLAG2_ETS_CFG		= 1LL <<  26,
 };
 
 enum {
-- 
1.7.1

[PATCH net-next 13/18] net/mlx4: Set enhanced QoS support by default when ETS supported

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

If HCA supports ETS QoS feature, set enhanced QoS bit in init_hca as default.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/fw.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index 16bbd6c..b93e23b 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -49,9 +49,9 @@ enum {
 extern void __buggy_use_of_MLX4_GET(void);
 extern void __buggy_use_of_MLX4_PUT(void);
 
-static bool enable_qos;
+static bool enable_qos = true;
 module_param(enable_qos, bool, 0444);
-MODULE_PARM_DESC(enable_qos, "Enable Quality of Service support in the HCA (default: off)");
+MODULE_PARM_DESC(enable_qos, "Enable Enhanced QoS support (default: on)");
 
 #define MLX4_GET(dest, source, offset)				      \
 	do {							      \
@@ -1709,7 +1709,7 @@ int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
 		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 3);
 
 	/* Enable QoS support if module parameter set */
-	if (enable_qos)
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_ETS_CFG && enable_qos)
 		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 2);
 
 	/* enable counters */
-- 
1.7.1

[PATCH net-next 14/18] net/mlx4: Add SET_PORT opcode modifiers enumeration

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

The calls to SET_PORT used hard-code numbers, when supplying command's
opcode modifiers, fix that to use well defined constants.

Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/infiniband/hw/mlx4/main.c         |   11 +++++----
 drivers/net/ethernet/mellanox/mlx4/cmd.c  |    3 +-
 drivers/net/ethernet/mellanox/mlx4/port.c |   36 +++++++++++++++++------------
 include/linux/mlx4/cmd.h                  |    8 +++++-
 4 files changed, 36 insertions(+), 22 deletions(-)

diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index b972c0b..976bea7 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -587,8 +587,9 @@ static int mlx4_ib_SET_PORT(struct mlx4_ib_dev *dev, u8 port, int reset_qkey_vio
 		((__be32 *) mailbox->buf)[1] = cpu_to_be32(cap_mask);
 	}
 
-	err = mlx4_cmd(dev->dev, mailbox->dma, port, 0, MLX4_CMD_SET_PORT,
-		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
+	err = mlx4_cmd(dev->dev, mailbox->dma, port, MLX4_SET_PORT_IB_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_WRAPPED);
 
 	mlx4_free_cmd_mailbox(dev->dev, mailbox);
 	return err;
@@ -1525,8 +1526,8 @@ static void update_gids_task(struct work_struct *work)
 	memcpy(gids, gw->gids, sizeof gw->gids);
 
 	err = mlx4_cmd(dev, mailbox->dma, MLX4_SET_PORT_GID_TABLE << 8 | gw->port,
-		       1, MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
-		       MLX4_CMD_WRAPPED);
+		       MLX4_SET_PORT_ETH_OPCODE, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
 	if (err)
 		pr_warn("set port command failed\n");
 	else
@@ -1564,7 +1565,7 @@ static void reset_gids_task(struct work_struct *work)
 				    IB_LINK_LAYER_ETHERNET) {
 		err = mlx4_cmd(dev, mailbox->dma,
 			       MLX4_SET_PORT_GID_TABLE << 8 | gw->port,
-			       1, MLX4_CMD_SET_PORT,
+			       MLX4_SET_PORT_ETH_OPCODE, MLX4_CMD_SET_PORT,
 			       MLX4_CMD_TIME_CLASS_B,
 			       MLX4_CMD_WRAPPED);
 		if (err)
diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index bfc3584..0403fc0 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -726,7 +726,8 @@ static int mlx4_cmd_wait(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
 		 * specific command/input_mod/opcode_mod/fw-status to be debug.
 		 */
 		if (op == MLX4_CMD_SET_PORT && in_modifier == 1 &&
-		    op_modifier == 0 && context->fw_status == CMD_STAT_BAD_SIZE)
+		    op_modifier == MLX4_SET_PORT_IB_OPCODE &&
+		    context->fw_status == CMD_STAT_BAD_SIZE)
 			mlx4_dbg(dev, "command 0x%x failed: fw status = 0x%x\n",
 				 op, context->fw_status);
 		else
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
index 6a53d42..ca4488e 100644
--- a/drivers/net/ethernet/mellanox/mlx4/port.c
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -123,8 +123,9 @@ static int mlx4_set_port_mac_table(struct mlx4_dev *dev, u8 port,
 
 	in_mod = MLX4_SET_PORT_MAC_TABLE << 8 | port;
 
-	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
-		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
 
 	mlx4_free_cmd_mailbox(dev, mailbox);
 	return err;
@@ -337,8 +338,9 @@ static int mlx4_set_port_vlan_table(struct mlx4_dev *dev, u8 port,
 
 	memcpy(mailbox->buf, entries, MLX4_VLAN_TABLE_SIZE);
 	in_mod = MLX4_SET_PORT_VLAN_TABLE << 8 | port;
-	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
-		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
 
 	mlx4_free_cmd_mailbox(dev, mailbox);
 
@@ -625,9 +627,9 @@ static int mlx4_reset_roce_port_gids(struct mlx4_dev *dev, int slave,
 		       MLX4_ROCE_GID_ENTRY_SIZE);
 
 	err = mlx4_cmd(dev, mailbox->dma,
-		       ((u32)port) | (MLX4_SET_PORT_GID_TABLE << 8), 1,
-		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
-		       MLX4_CMD_NATIVE);
+		       ((u32)port) | (MLX4_SET_PORT_GID_TABLE << 8),
+		       MLX4_SET_PORT_ETH_OPCODE, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
 	mutex_unlock(&(priv->port[port].gid_table.mutex));
 	return err;
 }
@@ -941,8 +943,9 @@ int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port, int pkey_tbl_sz)
 			(pkey_tbl_flag << MLX4_CHANGE_PORT_PKEY_TBL_SZ) |
 			(dev->caps.port_ib_mtu[port] << MLX4_SET_PORT_MTU_CAP) |
 			(vl_cap << MLX4_SET_PORT_VL_CAP));
-		err = mlx4_cmd(dev, mailbox->dma, port, 0, MLX4_CMD_SET_PORT,
-				MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
+		err = mlx4_cmd(dev, mailbox->dma, port,
+			       MLX4_SET_PORT_IB_OPCODE, MLX4_CMD_SET_PORT,
+			       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
 		if (err != -ENOMEM)
 			break;
 	}
@@ -971,8 +974,9 @@ int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu,
 	context->pfcrx = pfcrx;
 
 	in_mod = MLX4_SET_PORT_GENERAL << 8 | port;
-	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
-		       MLX4_CMD_TIME_CLASS_B,  MLX4_CMD_WRAPPED);
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_WRAPPED);
 
 	mlx4_free_cmd_mailbox(dev, mailbox);
 	return err;
@@ -1008,8 +1012,9 @@ int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
 	context->vlan_miss = MLX4_VLAN_MISS_IDX;
 
 	in_mod = MLX4_SET_PORT_RQP_CALC << 8 | port;
-	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
-		       MLX4_CMD_TIME_CLASS_B,  MLX4_CMD_WRAPPED);
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_WRAPPED);
 
 	mlx4_free_cmd_mailbox(dev, mailbox);
 	return err;
@@ -1050,8 +1055,9 @@ int mlx4_SET_PORT_VXLAN(struct mlx4_dev *dev, u8 port, u8 steering, int enable)
 	context->steering  = steering;
 
 	in_mod = MLX4_SET_PORT_VXLAN << 8 | port;
-	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
-		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
 
 	mlx4_free_cmd_mailbox(dev, mailbox);
 	return err;
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index a788839..62c4d4d 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -188,7 +188,13 @@ enum {
 };
 
 enum {
-	/* set port opcode modifiers */
+	/* Set port opcode modifiers */
+	MLX4_SET_PORT_IB_OPCODE		= 0x0,
+	MLX4_SET_PORT_ETH_OPCODE	= 0x1,
+};
+
+enum {
+	/* Set port Ethernet input modifiers */
 	MLX4_SET_PORT_GENERAL   = 0x0,
 	MLX4_SET_PORT_RQP_CALC  = 0x1,
 	MLX4_SET_PORT_MAC_TABLE = 0x2,
-- 
1.7.1

[PATCH net-next 15/18] net/mlx4_en: Add interface identify support

[Or Gerlitz]

From: Ido Shamay <idos@mellanox.com>

Add support for the interface ethtool identify feature.

Make the physical port LED to blink with green and yellow colors.

The device handles the LED blink by itself (synchrous use of
set_phys_id), by returning 0 to ETHTOOL_ID_ACTIVE command.

Signed-off-by: Eyal Grossman <eyalgr@mellanox.com>
Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_ethtool.c |   27 +++++++++++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/fw.c         |   10 ++++++++
 drivers/net/ethernet/mellanox/mlx4/mlx4_en.h    |    1 +
 drivers/net/ethernet/mellanox/mlx4/port.c       |   26 ++++++++++++++++++++++
 include/linux/mlx4/cmd.h                        |    1 +
 include/linux/mlx4/device.h                     |    2 +
 6 files changed, 67 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
index eba969b..3f44e2b 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
@@ -1939,6 +1939,32 @@ static int mlx4_en_get_module_eeprom(struct net_device *dev,
 	return 0;
 }
 
+static int mlx4_en_set_phys_id(struct net_device *dev,
+			       enum ethtool_phys_id_state state)
+{
+	int err;
+	u16 beacon_duration;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
+
+	if (!(mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PORT_BEACON))
+		return -EOPNOTSUPP;
+
+	switch (state) {
+	case ETHTOOL_ID_ACTIVE:
+		beacon_duration = PORT_BEACON_MAX_LIMIT;
+		break;
+	case ETHTOOL_ID_INACTIVE:
+		beacon_duration = 0;
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	err = mlx4_SET_PORT_BEACON(mdev->dev, priv->port, beacon_duration);
+	return err;
+}
+
 const struct ethtool_ops mlx4_en_ethtool_ops = {
 	.get_drvinfo = mlx4_en_get_drvinfo,
 	.get_settings = mlx4_en_get_settings,
@@ -1948,6 +1974,7 @@ const struct ethtool_ops mlx4_en_ethtool_ops = {
 	.get_sset_count = mlx4_en_get_sset_count,
 	.get_ethtool_stats = mlx4_en_get_ethtool_stats,
 	.self_test = mlx4_en_self_test,
+	.set_phys_id = mlx4_en_set_phys_id,
 	.get_wol = mlx4_en_get_wol,
 	.set_wol = mlx4_en_set_wol,
 	.get_msglevel = mlx4_en_get_msglevel,
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index b93e23b..6317844 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -150,6 +150,7 @@ static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags)
 		[24] = "Ethernet Flow control statistics support",
 		[25] = "Granular QoS per VF support",
 		[26] = "Port ETS Scheduler support",
+		[27] = "Port beacon support",
 	};
 	int i;
 
@@ -647,6 +648,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 #define QUERY_DEV_CAP_RSS_OFFSET		0x2e
 #define QUERY_DEV_CAP_MAX_RDMA_OFFSET		0x2f
 #define QUERY_DEV_CAP_RSZ_SRQ_OFFSET		0x33
+#define QUERY_DEV_CAP_PORT_BEACON_OFFSET	0x34
 #define QUERY_DEV_CAP_ACK_DELAY_OFFSET		0x35
 #define QUERY_DEV_CAP_MTU_WIDTH_OFFSET		0x36
 #define QUERY_DEV_CAP_VL_PORT_OFFSET		0x37
@@ -786,6 +788,9 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	if (field & 0x80)
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_FS_EN;
 	dev_cap->fs_log_max_ucast_qp_range_size = field & 0x1f;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_PORT_BEACON_OFFSET);
+	if (field & 0x80)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_PORT_BEACON;
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_FLOW_STEERING_IPOIB_OFFSET);
 	if (field & 0x80)
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_DMFS_IPOIB;
@@ -1165,6 +1170,11 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave,
 	field &= 0xd7;
 	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_VXLAN);
 
+	/* For guests, disable port BEACON */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_PORT_BEACON_OFFSET);
+	field &= 0x7f;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_PORT_BEACON_OFFSET);
+
 	/* For guests, report Blueflame disabled */
 	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_BF_OFFSET);
 	field &= 0x7f;
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index 67eeea2..9de3021 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -482,6 +482,7 @@ enum {
 	MLX4_EN_FLAG_RX_CSUM_NON_TCP_UDP	= (1 << 5),
 };
 
+#define PORT_BEACON_MAX_LIMIT (65535)
 #define MLX4_EN_MAC_HASH_SIZE (1 << BITS_PER_BYTE)
 #define MLX4_EN_MAC_HASH_IDX 5
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
index ca4488e..e8d716d 100644
--- a/drivers/net/ethernet/mellanox/mlx4/port.c
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -835,6 +835,12 @@ static int mlx4_common_set_port(struct mlx4_dev *dev, int slave, u32 in_mod,
 				MLX4_CMD_NATIVE);
 	}
 
+	/* Slaves are not allowed to SET_PORT beacon (LED) blink */
+	if (op_mod == MLX4_SET_PORT_BEACON_OPCODE) {
+		mlx4_warn(dev, "denying SET_PORT Beacon slave:%d\n", slave);
+		return -EPERM;
+	}
+
 	/* For IB, we only consider:
 	 * - The capability mask, which is set to the aggregate of all
 	 *   slave function capabilities
@@ -1064,6 +1070,26 @@ int mlx4_SET_PORT_VXLAN(struct mlx4_dev *dev, u8 port, u8 steering, int enable)
 }
 EXPORT_SYMBOL(mlx4_SET_PORT_VXLAN);
 
+int mlx4_SET_PORT_BEACON(struct mlx4_dev *dev, u8 port, u16 time)
+{
+	int err;
+	struct mlx4_cmd_mailbox *mailbox;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	*((__be32 *)mailbox->buf) = cpu_to_be32(time);
+
+	err = mlx4_cmd(dev, mailbox->dma, port, MLX4_SET_PORT_BEACON_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_BEACON);
+
 int mlx4_SET_MCAST_FLTR_wrapper(struct mlx4_dev *dev, int slave,
 				struct mlx4_vhcr *vhcr,
 				struct mlx4_cmd_mailbox *inbox,
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index 62c4d4d..f62e7cf 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -191,6 +191,7 @@ enum {
 	/* Set port opcode modifiers */
 	MLX4_SET_PORT_IB_OPCODE		= 0x0,
 	MLX4_SET_PORT_ETH_OPCODE	= 0x1,
+	MLX4_SET_PORT_BEACON_OPCODE	= 0x4,
 };
 
 enum {
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 15f4676..b761be7 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -208,6 +208,7 @@ enum {
 	MLX4_DEV_CAP_FLAG2_FLOWSTATS_EN         = 1LL <<  24,
 	MLX4_DEV_CAP_FLAG2_QOS_VPP		= 1LL <<  25,
 	MLX4_DEV_CAP_FLAG2_ETS_CFG		= 1LL <<  26,
+	MLX4_DEV_CAP_FLAG2_PORT_BEACON		= 1LL <<  27,
 };
 
 enum {
@@ -1311,6 +1312,7 @@ int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu,
 			  u8 pptx, u8 pfctx, u8 pprx, u8 pfcrx);
 int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
 			   u8 promisc);
+int mlx4_SET_PORT_BEACON(struct mlx4_dev *dev, u8 port, u16 time);
 int mlx4_SET_PORT_VXLAN(struct mlx4_dev *dev, u8 port, u8 steering, int enable);
 int mlx4_find_cached_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *idx);
 int mlx4_find_cached_vlan(struct mlx4_dev *dev, u8 port, u16 vid, int *idx);
-- 
1.7.1

[PATCH net-next 16/18] net/mlx4_en: Add RX-FCS support

[Or Gerlitz]

From: Muhammad Mahajna <muhammadm@mellanox.com>

Enabled when device supports KEEP FCS. When the flag is set, Ethernet FCS
is appended to the end of the frame, controlled by ethtool.

Signed-off-by: Muhammad Mahajna <muhammadm@mellanox.com>
Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c |   33 +++++++++++++++++++----
 drivers/net/ethernet/mellanox/mlx4/en_rx.c     |    5 +++-
 2 files changed, 31 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index b02a2e8..2cf8af7 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -2195,15 +2195,19 @@ static int mlx4_en_set_features(struct net_device *netdev,
 		netdev_features_t features)
 {
 	struct mlx4_en_priv *priv = netdev_priv(netdev);
+	bool reset = false;
 	int ret = 0;
 
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_RXFCS)) {
+		en_info(priv, "Turn %s RX-FCS\n",
+			(features & NETIF_F_RXFCS) ? "ON" : "OFF");
+		reset = true;
+	}
+
 	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_HW_VLAN_CTAG_RX)) {
 		en_info(priv, "Turn %s RX vlan strip offload\n",
 			(features & NETIF_F_HW_VLAN_CTAG_RX) ? "ON" : "OFF");
-		ret = mlx4_en_reset_config(netdev, priv->hwtstamp_config,
-					   features);
-		if (ret)
-			return ret;
+		reset = true;
 	}
 
 	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_HW_VLAN_CTAG_TX))
@@ -2216,8 +2220,14 @@ static int mlx4_en_set_features(struct net_device *netdev,
 		mlx4_en_update_loopback_state(netdev, features);
 	}
 
-	return 0;
+	if (reset) {
+		ret = mlx4_en_reset_config(netdev, priv->hwtstamp_config,
+					   features);
+		if (ret)
+			return ret;
+	}
 
+	return 0;
 }
 
 static int mlx4_en_set_vf_mac(struct net_device *dev, int queue, u8 *mac)
@@ -2901,6 +2911,9 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	dev->hw_features |= NETIF_F_LOOPBACK |
 			NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX;
 
+	if (mdev->dev->caps.flags & MLX4_DEV_CAP_FLAG_FCS_KEEP)
+		dev->hw_features |= NETIF_F_RXFCS;
+
 	if (mdev->dev->caps.steering_mode ==
 	    MLX4_STEERING_MODE_DEVICE_MANAGED &&
 	    mdev->dev->caps.dmfs_high_steer_mode != MLX4_STEERING_DMFS_A0_STATIC)
@@ -2996,7 +3009,8 @@ int mlx4_en_reset_config(struct net_device *dev,
 
 	if (priv->hwtstamp_config.tx_type == ts_config.tx_type &&
 	    priv->hwtstamp_config.rx_filter == ts_config.rx_filter &&
-	    !DEV_FEATURE_CHANGED(dev, features, NETIF_F_HW_VLAN_CTAG_RX))
+	    !DEV_FEATURE_CHANGED(dev, features, NETIF_F_HW_VLAN_CTAG_RX) &&
+	    !DEV_FEATURE_CHANGED(dev, features, NETIF_F_RXFCS))
 		return 0; /* Nothing to change */
 
 	if (DEV_FEATURE_CHANGED(dev, features, NETIF_F_HW_VLAN_CTAG_RX) &&
@@ -3035,6 +3049,13 @@ int mlx4_en_reset_config(struct net_device *dev,
 			dev->features &= ~NETIF_F_HW_VLAN_CTAG_RX;
 	}
 
+	if (DEV_FEATURE_CHANGED(dev, features, NETIF_F_RXFCS)) {
+		if (features & NETIF_F_RXFCS)
+			dev->features |= NETIF_F_RXFCS;
+		else
+			dev->features &= ~NETIF_F_RXFCS;
+	}
+
 	/* RX vlan offload and RX time-stamping can't co-exist !
 	 * Regardless of the caller's choice,
 	 * Turn Off RX vlan offload in case of time-stamping is ON
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
index 698d60d..79b1501 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -1116,7 +1116,10 @@ static int mlx4_en_config_rss_qp(struct mlx4_en_priv *priv, int qpn,
 	/* Cancel FCS removal if FW allows */
 	if (mdev->dev->caps.flags & MLX4_DEV_CAP_FLAG_FCS_KEEP) {
 		context->param3 |= cpu_to_be32(1 << 29);
-		ring->fcs_del = ETH_FCS_LEN;
+		if (priv->dev->features & NETIF_F_RXFCS)
+			ring->fcs_del = 0;
+		else
+			ring->fcs_del = ETH_FCS_LEN;
 	} else
 		ring->fcs_del = 0;
 
-- 
1.7.1

[PATCH net-next 17/18] net/mlx4_en: Add RX-ALL support

[Or Gerlitz]

From: Muhammad Mahajna <muhammadm@mellanox.com>

Enabled when the device supports KEEP FCS and IGNORE FCS.

When the flag is set, pass all received frames up the stack,
even ones with invalid FCS, controlled by ethtool.

Signed-off-by: Muhammad Mahajna <muhammadm@mellanox.com>
Signed-off-by: Ido Shamay <idos@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/en_netdev.c |   14 ++++++++++
 drivers/net/ethernet/mellanox/mlx4/fw.c        |    8 +++++
 drivers/net/ethernet/mellanox/mlx4/main.c      |   21 ++++++++++++++
 drivers/net/ethernet/mellanox/mlx4/mlx4.h      |    6 +++-
 drivers/net/ethernet/mellanox/mlx4/port.c      |   34 ++++++++++++++++++++++++
 include/linux/mlx4/device.h                    |    3 ++
 6 files changed, 84 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 2cf8af7..b13aa03 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -2204,6 +2204,17 @@ static int mlx4_en_set_features(struct net_device *netdev,
 		reset = true;
 	}
 
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_RXALL)) {
+		u8 ignore_fcs_value = (features & NETIF_F_RXALL) ? 1 : 0;
+
+		en_info(priv, "Turn %s RX-ALL\n",
+			ignore_fcs_value ? "ON" : "OFF");
+		ret = mlx4_SET_PORT_fcs_check(priv->mdev->dev,
+					      priv->port, ignore_fcs_value);
+		if (ret)
+			return ret;
+	}
+
 	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_HW_VLAN_CTAG_RX)) {
 		en_info(priv, "Turn %s RX vlan strip offload\n",
 			(features & NETIF_F_HW_VLAN_CTAG_RX) ? "ON" : "OFF");
@@ -2914,6 +2925,9 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	if (mdev->dev->caps.flags & MLX4_DEV_CAP_FLAG_FCS_KEEP)
 		dev->hw_features |= NETIF_F_RXFCS;
 
+	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_IGNORE_FCS)
+		dev->hw_features |= NETIF_F_RXALL;
+
 	if (mdev->dev->caps.steering_mode ==
 	    MLX4_STEERING_MODE_DEVICE_MANAGED &&
 	    mdev->dev->caps.dmfs_high_steer_mode != MLX4_STEERING_DMFS_A0_STATIC)
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index 6317844..b9881fc 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -151,6 +151,7 @@ static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags)
 		[25] = "Granular QoS per VF support",
 		[26] = "Port ETS Scheduler support",
 		[27] = "Port beacon support",
+		[28] = "RX-ALL support",
 	};
 	int i;
 
@@ -893,6 +894,8 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_CONFIG_DEV_OFFSET);
 	if (field & 0x20)
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_CONFIG_DEV;
+	if (field & (1 << 2))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_IGNORE_FCS;
 	MLX4_GET(dev_cap->reserved_lkey, outbox,
 		 QUERY_DEV_CAP_RSVD_LKEY_OFFSET);
 	MLX4_GET(field32, outbox, QUERY_DEV_CAP_ETH_BACKPL_OFFSET);
@@ -1220,6 +1223,11 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave,
 	field &= 0xef;
 	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_CQ_EQ_CACHE_LINE_STRIDE);
 
+	/* turn off ignore FCS feature for guests */
+	MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_CONFIG_DEV_OFFSET);
+	field &= 0xfb;
+	MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_CONFIG_DEV_OFFSET);
+
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index 3becad3..acceb75 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -297,6 +297,25 @@ static int mlx4_dev_port(struct mlx4_dev *dev, int port,
 	return err;
 }
 
+static inline void mlx4_enable_ignore_fcs(struct mlx4_dev *dev)
+{
+	if (!(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_IGNORE_FCS))
+		return;
+
+	if (mlx4_is_mfunc(dev)) {
+		mlx4_dbg(dev, "SRIOV mode - Disabling Ignore FCS");
+		dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_IGNORE_FCS;
+		return;
+	}
+
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_FCS_KEEP)) {
+		mlx4_dbg(dev,
+			 "Keep FCS is not supported - Disabling Ignore FCS");
+		dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_IGNORE_FCS;
+		return;
+	}
+}
+
 #define MLX4_A0_STEERING_TABLE_SIZE	256
 static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 {
@@ -540,6 +559,8 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 		dev->caps.alloc_res_qp_mask = 0;
 	}
 
+	mlx4_enable_ignore_fcs(dev);
+
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index 32469c6..f30eeb7 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -763,9 +763,11 @@ enum {
 
 
 struct mlx4_set_port_general_context {
-	u8 reserved[3];
+	u16 reserved1;
+	u8 v_ignore_fcs;
 	u8 flags;
-	u16 reserved2;
+	u8 ignore_fcs;
+	u8 reserved2;
 	__be16 mtu;
 	u8 pptx;
 	u8 pfctx;
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
index e8d716d..c2b2131 100644
--- a/drivers/net/ethernet/mellanox/mlx4/port.c
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -45,6 +45,14 @@
 #define MLX4_VLAN_VALID		(1u << 31)
 #define MLX4_VLAN_MASK		0xfff
 
+#define MLX4_STATS_TRAFFIC_COUNTERS_MASK	0xfULL
+#define MLX4_STATS_TRAFFIC_DROPS_MASK		0xc0ULL
+#define MLX4_STATS_ERROR_COUNTERS_MASK		0x1ffc30ULL
+#define MLX4_STATS_PORT_COUNTERS_MASK		0x1fe00000ULL
+
+#define MLX4_FLAG_V_IGNORE_FCS_MASK		0x2
+#define MLX4_IGNORE_FCS_MASK			0x1
+
 void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table)
 {
 	int i;
@@ -1027,6 +1035,32 @@ int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
 }
 EXPORT_SYMBOL(mlx4_SET_PORT_qpn_calc);
 
+int mlx4_SET_PORT_fcs_check(struct mlx4_dev *dev, u8 port, u8 ignore_fcs_value)
+{
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_general_context *context;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+	context->v_ignore_fcs |= MLX4_FLAG_V_IGNORE_FCS_MASK;
+	if (ignore_fcs_value)
+		context->ignore_fcs |= MLX4_IGNORE_FCS_MASK;
+	else
+		context->ignore_fcs &= ~MLX4_IGNORE_FCS_MASK;
+
+	in_mod = MLX4_SET_PORT_GENERAL << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+		       MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_fcs_check);
+
 enum {
 	VXLAN_ENABLE_MODIFY	= 1 << 7,
 	VXLAN_STEERING_MODIFY	= 1 << 6,
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index b761be7..f9ce34b 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -209,6 +209,7 @@ enum {
 	MLX4_DEV_CAP_FLAG2_QOS_VPP		= 1LL <<  25,
 	MLX4_DEV_CAP_FLAG2_ETS_CFG		= 1LL <<  26,
 	MLX4_DEV_CAP_FLAG2_PORT_BEACON		= 1LL <<  27,
+	MLX4_DEV_CAP_FLAG2_IGNORE_FCS		= 1LL <<  28,
 };
 
 enum {
@@ -1313,6 +1314,8 @@ int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu,
 int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
 			   u8 promisc);
 int mlx4_SET_PORT_BEACON(struct mlx4_dev *dev, u8 port, u16 time);
+int mlx4_SET_PORT_fcs_check(struct mlx4_dev *dev, u8 port,
+			    u8 ignore_fcs_value);
 int mlx4_SET_PORT_VXLAN(struct mlx4_dev *dev, u8 port, u8 steering, int enable);
 int mlx4_find_cached_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *idx);
 int mlx4_find_cached_vlan(struct mlx4_dev *dev, u8 port, u16 vid, int *idx);
-- 
1.7.1

[PATCH net-next 18/18] MAINTAINERS: Update mlx4_en entry

[Or Gerlitz]

Add Ido Shamay as co-maintainer for the mlx4 Ethernet driver.

Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
 MAINTAINERS |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/MAINTAINERS b/MAINTAINERS
index c201655..6ba55f2 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -6316,6 +6316,7 @@ F:	drivers/scsi/megaraid/
 
 MELLANOX ETHERNET DRIVER (mlx4_en)
 M:	Amir Vadai <amirv@mellanox.com>
+M:	Ido Shamay <idos@mellanox.com>
 L:	netdev@vger.kernel.org
 S:	Supported
 W:	http://www.mellanox.com
-- 
1.7.1

Re: [PATCH net-next 00/18] Mellanox mlx4 driver updates

[David Miller]

From: Or Gerlitz <ogerlitz@mellanox.com>
Date: Thu,  2 Apr 2015 16:31:05 +0300

> The main feature added by this series are Ido's changes to support 
> Granular QoS for VFs, where for the time being only max rate is supported.
> 
> Muhammad added support for setting rx-fcs and rx-all through ethtool, 
> and Ido did the interface identify work.
> 
> Last, add Ido as a maintainer for the mlx4 Ethernet driver!
> 
> Some of next week is the Passover holiday here and I will be 
> mostly OOO. If needed (...) Ido or Amir will send V1 and such.
> 
> Rebased against net-next commit 033f46b "crypto: algif - 
> explicitly mark end of data".

Series applied, thanks.