[pull request][net-next 00/14] mlx5 updates 2021-02-01

[pull request][net-next 00/14] mlx5 updates 2021-02-01

[Saeed Mahameed]

From: Saeed Mahameed <saeedm@nvidia.com>

Hi Jakub, Dave,

This series adds misc updates to mlx5 driver.
For more information please see tag log below.

Please pull and let me know if there is any problem.

Thanks,
Saeed.

---
The following changes since commit 9ae4bdc6e4c1281ddf8d6335bea35864d086cbf9:

  Merge branch 'rework-the-memory-barrier-for-scrq-entry' (2021-02-01 20:21:14 -0800)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-updates-2021-02-01

for you to fetch changes up to a283ea1b97163d21e0f1a3df387b71787042b990:

  net/mlx5: DR, Avoid unnecessary csum recalculation on supporting devices (2021-02-01 22:52:36 -0800)

----------------------------------------------------------------
mlx5-updates-2021-02-01

mlx5 netdev updates:

1) Trivial refactoring ahead of the upcoming uplink representor series.
2) Increased RSS table size to 256, for better results
3) Misc. Cleanup and very trivial improvements

----------------------------------------------------------------
Noam Stolero (1):
      net/mlx5e: Increase indirection RQ table size to 256

Roi Dayan (5):
      net/mlx5e: Refactor mlx5e_netdev_init/cleanup to mlx5e_priv_init/cleanup
      net/mlx5e: Move netif_carrier_off() out of mlx5e_priv_init()
      net/mlx5e: Move set vxlan nic info to profile init
      net/mlx5e: Avoid false lock depenency warning on tc_ht
      net/mlx5e: Move representor neigh init into profile enable

Saeed Mahameed (4):
      net/mlx5e: Separate between netdev objects and mlx5e profiles initialization
      net/mxl5e: Add change profile method
      net/mlx5e: accel, remove redundant space
      net/mlx5e: CT: remove useless conversion to PTR_ERR then ERR_PTR

Tariq Toukan (2):
      net/mlx5e: Enable napi in channel's activation stage
      net/mlx5e: kTLS, Improve TLS RX workqueue scope

Tom Rix (1):
      net/mlx5e: remove h from printk format specifier

Yevgeny Kliteynik (1):
      net/mlx5: DR, Avoid unnecessary csum recalculation on supporting devices

 drivers/net/ethernet/mellanox/mlx5/core/en.h       |  36 ++--
 .../net/ethernet/mellanox/mlx5/core/en/params.c    |   2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c   |  12 +-
 .../net/ethernet/mellanox/mlx5/core/en/rep/neigh.c |  18 +-
 drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c |   7 +-
 .../mellanox/mlx5/core/en_accel/en_accel.h         |   2 +-
 .../ethernet/mellanox/mlx5/core/en_accel/ktls.c    |  24 ++-
 .../net/ethernet/mellanox/mlx5/core/en_accel/tls.c |   7 -
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c  | 226 ++++++++++++++-------
 drivers/net/ethernet/mellanox/mlx5/core/en_rep.c   |  90 ++++----
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c    |  12 ++
 .../net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c  |  26 +--
 .../net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h  |   5 +-
 .../ethernet/mellanox/mlx5/core/ipoib/ipoib_vlan.c |   6 +-
 .../mellanox/mlx5/core/steering/dr_action.c        |   9 +-
 .../ethernet/mellanox/mlx5/core/steering/dr_ste.c  |   5 +
 .../mellanox/mlx5/core/steering/dr_types.h         |   2 +
 17 files changed, 304 insertions(+), 185 deletions(-)

[net-next 01/14] net/mlx5e: Separate between netdev objects and mlx5e profiles initialization

[Saeed Mahameed]

From: Saeed Mahameed <saeedm@nvidia.com>

1) Initialize netdevice features and structures on netdevice allocation
   and outside of the mlx5e profile.

2) As now mlx5e netdevice private params will be setup on profile init only
   after netdevice features are already set, we add  a call to
   netde_update_features() to resolve any conflict.
   This is nice since we reuse the fix_features ndo code if a profile
   wants different default features, instead of duplicating features
   conflict resolution code on profile initialization.

3) With this we achieve total separation between mlx5e profiles and
   netdevices, and will allow replacing mlx5e profiles on the fly to reuse
   the same netdevice for multiple profiles.
   e.g. for uplink representor profile as shown in the following patch

4) Profile callbacks are not allowed to touch netdev->features directly
   anymore, since in downstream patch we will detach/attach netdev
   dynamically to profile, hence we move the code dealing with
   netdev->features from profile->init() to fix_features ndo, and we
   will call netdev_update_features() on
   mlx5e_attach_netdev(profile, netdev);

Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en.h  |  23 ++--
 .../net/ethernet/mellanox/mlx5/core/en_main.c | 127 +++++++++---------
 .../net/ethernet/mellanox/mlx5/core/en_rep.c  |  51 ++++---
 .../ethernet/mellanox/mlx5/core/ipoib/ipoib.c |  23 ++--
 .../ethernet/mellanox/mlx5/core/ipoib/ipoib.h |   5 +-
 .../mellanox/mlx5/core/ipoib/ipoib_vlan.c     |   6 +-
 6 files changed, 121 insertions(+), 114 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 39f389cc40fc..bf5de1e79134 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -895,8 +895,7 @@ extern const struct mlx5e_rx_handlers mlx5e_rx_handlers_nic;
 
 struct mlx5e_profile {
 	int	(*init)(struct mlx5_core_dev *mdev,
-			struct net_device *netdev,
-			const struct mlx5e_profile *profile, void *ppriv);
+			struct net_device *netdev);
 	void	(*cleanup)(struct mlx5e_priv *priv);
 	int	(*init_rx)(struct mlx5e_priv *priv);
 	void	(*cleanup_rx)(struct mlx5e_priv *priv);
@@ -1155,24 +1154,22 @@ int mlx5e_ethtool_set_pauseparam(struct mlx5e_priv *priv,
 				 struct ethtool_pauseparam *pauseparam);
 
 /* mlx5e generic netdev management API */
+static inline unsigned int mlx5e_calc_max_nch(struct mlx5e_priv *priv)
+{
+	return priv->netdev->num_rx_queues / max_t(u8, priv->profile->rq_groups, 1);
+}
+
 int mlx5e_netdev_init(struct net_device *netdev,
 		      struct mlx5e_priv *priv,
-		      struct mlx5_core_dev *mdev,
-		      const struct mlx5e_profile *profile,
-		      void *ppriv);
+		      struct mlx5_core_dev *mdev);
 void mlx5e_netdev_cleanup(struct net_device *netdev, struct mlx5e_priv *priv);
-struct net_device*
-mlx5e_create_netdev(struct mlx5_core_dev *mdev, const struct mlx5e_profile *profile,
-		    int nch, void *ppriv);
+struct net_device *
+mlx5e_create_netdev(struct mlx5_core_dev *mdev, unsigned int txqs, unsigned int rxqs);
 int mlx5e_attach_netdev(struct mlx5e_priv *priv);
 void mlx5e_detach_netdev(struct mlx5e_priv *priv);
 void mlx5e_destroy_netdev(struct mlx5e_priv *priv);
 void mlx5e_set_netdev_mtu_boundaries(struct mlx5e_priv *priv);
-void mlx5e_build_nic_params(struct mlx5e_priv *priv,
-			    struct mlx5e_xsk *xsk,
-			    struct mlx5e_rss_params *rss_params,
-			    struct mlx5e_params *params,
-			    u16 mtu);
+void mlx5e_build_nic_params(struct mlx5e_priv *priv, struct mlx5e_xsk *xsk, u16 mtu);
 void mlx5e_build_rq_params(struct mlx5_core_dev *mdev,
 			   struct mlx5e_params *params);
 void mlx5e_build_rss_params(struct mlx5e_rss_params *rss_params,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index aad3887e3c1a..260ced27014d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -4093,6 +4093,7 @@ static netdev_features_t mlx5e_fix_features(struct net_device *netdev,
 		if (!params->vlan_strip_disable)
 			netdev_warn(netdev, "Dropping C-tag vlan stripping offload due to S-tag vlan\n");
 	}
+
 	if (!MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_STRIDING_RQ)) {
 		if (features & NETIF_F_LRO) {
 			netdev_warn(netdev, "Disabling LRO, not supported in legacy RQ\n");
@@ -4928,15 +4929,15 @@ void mlx5e_build_rss_params(struct mlx5e_rss_params *rss_params,
 			tirc_default_config[tt].rx_hash_fields;
 }
 
-void mlx5e_build_nic_params(struct mlx5e_priv *priv,
-			    struct mlx5e_xsk *xsk,
-			    struct mlx5e_rss_params *rss_params,
-			    struct mlx5e_params *params,
-			    u16 mtu)
+void mlx5e_build_nic_params(struct mlx5e_priv *priv, struct mlx5e_xsk *xsk, u16 mtu)
 {
+	struct mlx5e_rss_params *rss_params = &priv->rss_params;
+	struct mlx5e_params *params = &priv->channels.params;
 	struct mlx5_core_dev *mdev = priv->mdev;
 	u8 rx_cq_period_mode;
 
+	priv->max_nch = mlx5e_calc_max_nch(priv);
+
 	params->sw_mtu = mtu;
 	params->hard_mtu = MLX5E_ETH_HARD_MTU;
 	params->num_channels = min_t(unsigned int, MLX5E_MAX_NUM_CHANNELS / 2,
@@ -4994,6 +4995,11 @@ void mlx5e_build_nic_params(struct mlx5e_priv *priv,
 
 	/* AF_XDP */
 	params->xsk = xsk;
+
+	/* Do not update netdev->features directly in here
+	 * on mlx5e_attach_netdev() we will call mlx5e_update_features()
+	 * To update netdev->features please modify mlx5e_fix_features()
+	 */
 }
 
 static void mlx5e_set_netdev_dev_addr(struct net_device *netdev)
@@ -5146,18 +5152,12 @@ static void mlx5e_build_nic_netdev(struct net_device *netdev)
 		netdev->hw_features |= NETIF_F_RXFCS;
 
 	netdev->features          = netdev->hw_features;
-	if (!priv->channels.params.lro_en)
-		netdev->features  &= ~NETIF_F_LRO;
 
+	/* Defaults */
 	if (fcs_enabled)
 		netdev->features  &= ~NETIF_F_RXALL;
-
-	if (!priv->channels.params.scatter_fcs_en)
-		netdev->features  &= ~NETIF_F_RXFCS;
-
-	/* prefere CQE compression over rxhash */
-	if (MLX5E_GET_PFLAG(&priv->channels.params, MLX5E_PFLAG_RX_CQE_COMPRESS))
-		netdev->features &= ~NETIF_F_RXHASH;
+	netdev->features  &= ~NETIF_F_LRO;
+	netdev->features  &= ~NETIF_F_RXFCS;
 
 #define FT_CAP(f) MLX5_CAP_FLOWTABLE(mdev, flow_table_properties_nic_receive.f)
 	if (FT_CAP(flow_modify_en) &&
@@ -5223,33 +5223,27 @@ void mlx5e_destroy_q_counters(struct mlx5e_priv *priv)
 }
 
 static int mlx5e_nic_init(struct mlx5_core_dev *mdev,
-			  struct net_device *netdev,
-			  const struct mlx5e_profile *profile,
-			  void *ppriv)
+			  struct net_device *netdev)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
-	struct mlx5e_rss_params *rss = &priv->rss_params;
 	int err;
 
-	err = mlx5e_netdev_init(netdev, priv, mdev, profile, ppriv);
-	if (err)
-		return err;
-
-	mlx5e_build_nic_params(priv, &priv->xsk, rss, &priv->channels.params,
-			       netdev->mtu);
+	mlx5e_build_nic_params(priv, &priv->xsk, netdev->mtu);
 
 	mlx5e_timestamp_init(priv);
 
 	err = mlx5e_ipsec_init(priv);
 	if (err)
 		mlx5_core_err(mdev, "IPSec initialization failed, %d\n", err);
+
 	err = mlx5e_tls_init(priv);
 	if (err)
 		mlx5_core_err(mdev, "TLS initialization failed, %d\n", err);
-	mlx5e_build_nic_netdev(netdev);
+
 	err = mlx5e_devlink_port_register(priv);
 	if (err)
 		mlx5_core_err(mdev, "mlx5e_devlink_port_register failed, %d\n", err);
+
 	mlx5e_health_create_reporters(priv);
 
 	return 0;
@@ -5261,7 +5255,6 @@ static void mlx5e_nic_cleanup(struct mlx5e_priv *priv)
 	mlx5e_devlink_port_unregister(priv);
 	mlx5e_tls_cleanup(priv);
 	mlx5e_ipsec_cleanup(priv);
-	mlx5e_netdev_cleanup(priv->netdev, priv);
 }
 
 static int mlx5e_init_nic_rx(struct mlx5e_priv *priv)
@@ -5464,21 +5457,14 @@ static const struct mlx5e_profile mlx5e_nic_profile = {
 };
 
 /* mlx5e generic netdev management API (move to en_common.c) */
-
-/* mlx5e_netdev_init/cleanup must be called from profile->init/cleanup callbacks */
 int mlx5e_netdev_init(struct net_device *netdev,
 		      struct mlx5e_priv *priv,
-		      struct mlx5_core_dev *mdev,
-		      const struct mlx5e_profile *profile,
-		      void *ppriv)
+		      struct mlx5_core_dev *mdev)
 {
 	/* priv init */
 	priv->mdev        = mdev;
 	priv->netdev      = netdev;
-	priv->profile     = profile;
-	priv->ppriv       = ppriv;
 	priv->msglevel    = MLX5E_MSG_LEVEL;
-	priv->max_nch     = netdev->num_rx_queues / max_t(u8, profile->rq_groups, 1);
 	priv->max_opened_tc = 1;
 
 	if (!alloc_cpumask_var(&priv->scratchpad.cpumask, GFP_KERNEL))
@@ -5518,35 +5504,24 @@ void mlx5e_netdev_cleanup(struct net_device *netdev, struct mlx5e_priv *priv)
 	kvfree(priv->htb.qos_sq_stats);
 }
 
-struct net_device *mlx5e_create_netdev(struct mlx5_core_dev *mdev,
-				       const struct mlx5e_profile *profile,
-				       int nch,
-				       void *ppriv)
+struct net_device *
+mlx5e_create_netdev(struct mlx5_core_dev *mdev, unsigned int txqs, unsigned int rxqs)
 {
 	struct net_device *netdev;
-	unsigned int ptp_txqs = 0;
-	int qos_sqs = 0;
 	int err;
 
-	if (MLX5_CAP_GEN(mdev, ts_cqe_to_dest_cqn))
-		ptp_txqs = profile->max_tc;
-
-	if (mlx5_qos_is_supported(mdev))
-		qos_sqs = mlx5e_qos_max_leaf_nodes(mdev);
-
-	netdev = alloc_etherdev_mqs(sizeof(struct mlx5e_priv),
-				    nch * profile->max_tc + ptp_txqs + qos_sqs,
-				    nch * profile->rq_groups);
+	netdev = alloc_etherdev_mqs(sizeof(struct mlx5e_priv), txqs, rxqs);
 	if (!netdev) {
 		mlx5_core_err(mdev, "alloc_etherdev_mqs() failed\n");
 		return NULL;
 	}
 
-	err = profile->init(mdev, netdev, profile, ppriv);
+	err = mlx5e_netdev_init(netdev, netdev_priv(netdev), mdev);
 	if (err) {
-		mlx5_core_err(mdev, "failed to init mlx5e profile %d\n", err);
+		mlx5_core_err(mdev, "mlx5e_netdev_init failed, err=%d\n", err);
 		goto err_free_netdev;
 	}
+	dev_net_set(netdev, mlx5_core_net(mdev));
 
 	return netdev;
 
@@ -5556,14 +5531,23 @@ struct net_device *mlx5e_create_netdev(struct mlx5_core_dev *mdev,
 	return NULL;
 }
 
+static void mlx5e_update_features(struct net_device *netdev)
+{
+	if (netdev->reg_state != NETREG_REGISTERED)
+		return; /* features will be updated on netdev registration */
+
+	rtnl_lock();
+	netdev_update_features(netdev);
+	rtnl_unlock();
+}
+
 int mlx5e_attach_netdev(struct mlx5e_priv *priv)
 {
 	const bool take_rtnl = priv->netdev->reg_state == NETREG_REGISTERED;
-	const struct mlx5e_profile *profile;
+	const struct mlx5e_profile *profile = priv->profile;
 	int max_nch;
 	int err;
 
-	profile = priv->profile;
 	clear_bit(MLX5E_STATE_DESTROYING, &priv->state);
 
 	/* max number of channels may have changed */
@@ -5603,6 +5587,8 @@ int mlx5e_attach_netdev(struct mlx5e_priv *priv)
 	if (profile->enable)
 		profile->enable(priv);
 
+	mlx5e_update_features(priv->netdev);
+
 	return 0;
 
 err_cleanup_tx:
@@ -5631,11 +5617,9 @@ void mlx5e_detach_netdev(struct mlx5e_priv *priv)
 
 void mlx5e_destroy_netdev(struct mlx5e_priv *priv)
 {
-	const struct mlx5e_profile *profile = priv->profile;
 	struct net_device *netdev = priv->netdev;
 
-	if (profile->cleanup)
-		profile->cleanup(priv);
+	mlx5e_netdev_cleanup(netdev, priv);
 	free_netdev(netdev);
 }
 
@@ -5681,28 +5665,48 @@ static int mlx5e_probe(struct auxiliary_device *adev,
 		       const struct auxiliary_device_id *id)
 {
 	struct mlx5_adev *edev = container_of(adev, struct mlx5_adev, adev);
+	const struct mlx5e_profile *profile = &mlx5e_nic_profile;
 	struct mlx5_core_dev *mdev = edev->mdev;
 	struct net_device *netdev;
 	pm_message_t state = {};
-	void *priv;
+	unsigned int txqs, rxqs, ptp_txqs = 0;
+	struct mlx5e_priv *priv;
+	int qos_sqs = 0;
 	int err;
 	int nch;
 
+	if (MLX5_CAP_GEN(mdev, ts_cqe_to_dest_cqn))
+		ptp_txqs = profile->max_tc;
+
+	if (mlx5_qos_is_supported(mdev))
+		qos_sqs = mlx5e_qos_max_leaf_nodes(mdev);
+
 	nch = mlx5e_get_max_num_channels(mdev);
-	netdev = mlx5e_create_netdev(mdev, &mlx5e_nic_profile, nch, NULL);
+	txqs = nch * profile->max_tc + ptp_txqs + qos_sqs;
+	rxqs = nch * profile->rq_groups;
+	netdev = mlx5e_create_netdev(mdev, txqs, rxqs);
 	if (!netdev) {
 		mlx5_core_err(mdev, "mlx5e_create_netdev failed\n");
 		return -ENOMEM;
 	}
 
-	dev_net_set(netdev, mlx5_core_net(mdev));
+	mlx5e_build_nic_netdev(netdev);
+
 	priv = netdev_priv(netdev);
 	dev_set_drvdata(&adev->dev, priv);
 
+	priv->profile = profile;
+	priv->ppriv = NULL;
+	err = profile->init(mdev, netdev);
+	if (err) {
+		mlx5_core_err(mdev, "mlx5e_nic_profile init failed, %d\n", err);
+		goto err_destroy_netdev;
+	}
+
 	err = mlx5e_resume(adev);
 	if (err) {
 		mlx5_core_err(mdev, "mlx5e_resume failed, %d\n", err);
-		goto err_destroy_netdev;
+		goto err_profile_cleanup;
 	}
 
 	err = register_netdev(netdev);
@@ -5718,6 +5722,8 @@ static int mlx5e_probe(struct auxiliary_device *adev,
 
 err_resume:
 	mlx5e_suspend(adev, state);
+err_profile_cleanup:
+	profile->cleanup(priv);
 err_destroy_netdev:
 	mlx5e_destroy_netdev(priv);
 	return err;
@@ -5731,6 +5737,7 @@ static void mlx5e_remove(struct auxiliary_device *adev)
 	mlx5e_dcbnl_delete_app(priv);
 	unregister_netdev(priv->netdev);
 	mlx5e_suspend(adev, state);
+	priv->profile->cleanup(priv);
 	mlx5e_destroy_netdev(priv);
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index 629ae6ccb4cd..d94d2ff9d312 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -684,7 +684,10 @@ static void mlx5e_build_rep_params(struct net_device *netdev)
 					 MLX5_CQ_PERIOD_MODE_START_FROM_CQE :
 					 MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
 
+	priv->max_nch = mlx5e_calc_max_nch(priv);
 	params = &priv->channels.params;
+
+	params->num_channels = MLX5E_REP_PARAMS_DEF_NUM_CHANNELS;
 	params->hard_mtu    = MLX5E_ETH_HARD_MTU;
 	params->sw_mtu      = netdev->mtu;
 
@@ -710,12 +713,11 @@ static void mlx5e_build_rep_params(struct net_device *netdev)
 	mlx5e_build_rss_params(&priv->rss_params, params->num_channels);
 }
 
-static void mlx5e_build_rep_netdev(struct net_device *netdev)
+static void mlx5e_build_rep_netdev(struct net_device *netdev,
+				   struct mlx5_core_dev *mdev,
+				   struct mlx5_eswitch_rep *rep)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
-	struct mlx5e_rep_priv *rpriv = priv->ppriv;
-	struct mlx5_eswitch_rep *rep = rpriv->rep;
-	struct mlx5_core_dev *mdev = priv->mdev;
 
 	SET_NETDEV_DEV(netdev, mdev->device);
 	if (rep->vport == MLX5_VPORT_UPLINK) {
@@ -755,22 +757,11 @@ static void mlx5e_build_rep_netdev(struct net_device *netdev)
 }
 
 static int mlx5e_init_rep(struct mlx5_core_dev *mdev,
-			  struct net_device *netdev,
-			  const struct mlx5e_profile *profile,
-			  void *ppriv)
+			  struct net_device *netdev)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
-	int err;
-
-	err = mlx5e_netdev_init(netdev, priv, mdev, profile, ppriv);
-	if (err)
-		return err;
-
-	priv->channels.params.num_channels = MLX5E_REP_PARAMS_DEF_NUM_CHANNELS;
 
 	mlx5e_build_rep_params(netdev);
-	mlx5e_build_rep_netdev(netdev);
-
 	mlx5e_timestamp_init(priv);
 
 	return 0;
@@ -778,7 +769,6 @@ static int mlx5e_init_rep(struct mlx5_core_dev *mdev,
 
 static void mlx5e_cleanup_rep(struct mlx5e_priv *priv)
 {
-	mlx5e_netdev_cleanup(priv->netdev, priv);
 }
 
 static int mlx5e_create_rep_ttc_table(struct mlx5e_priv *priv)
@@ -1201,6 +1191,8 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 	struct mlx5e_rep_priv *rpriv;
 	struct devlink_port *dl_port;
 	struct net_device *netdev;
+	struct mlx5e_priv *priv;
+	unsigned int txqs, rxqs;
 	int nch, err;
 
 	rpriv = kzalloc(sizeof(*rpriv), GFP_KERNEL);
@@ -1210,10 +1202,13 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 	/* rpriv->rep to be looked up when profile->init() is called */
 	rpriv->rep = rep;
 
-	nch = mlx5e_get_max_num_channels(dev);
 	profile = (rep->vport == MLX5_VPORT_UPLINK) ?
 		  &mlx5e_uplink_rep_profile : &mlx5e_rep_profile;
-	netdev = mlx5e_create_netdev(dev, profile, nch, rpriv);
+
+	nch = mlx5e_get_max_num_channels(dev);
+	txqs = nch * profile->max_tc;
+	rxqs = nch * profile->rq_groups;
+	netdev = mlx5e_create_netdev(dev, txqs, rxqs);
 	if (!netdev) {
 		mlx5_core_warn(dev,
 			       "Failed to create representor netdev for vport %d\n",
@@ -1222,7 +1217,8 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 		return -EINVAL;
 	}
 
-	dev_net_set(netdev, mlx5_core_net(dev));
+	mlx5e_build_rep_netdev(netdev, dev, rep);
+
 	rpriv->netdev = netdev;
 	rep->rep_data[REP_ETH].priv = rpriv;
 	INIT_LIST_HEAD(&rpriv->vport_sqs_list);
@@ -1233,12 +1229,21 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 			goto err_destroy_netdev;
 	}
 
+	priv = netdev_priv(netdev);
+	priv->profile = profile;
+	priv->ppriv = rpriv;
+	err = profile->init(dev, netdev);
+	if (err) {
+		netdev_warn(netdev, "rep profile init failed, %d\n", err);
+		goto err_destroy_mdev_resources;
+	}
+
 	err = mlx5e_attach_netdev(netdev_priv(netdev));
 	if (err) {
 		netdev_warn(netdev,
 			    "Failed to attach representor netdev for vport %d\n",
 			    rep->vport);
-		goto err_destroy_mdev_resources;
+		goto err_cleanup_profile;
 	}
 
 	err = mlx5e_rep_neigh_init(rpriv);
@@ -1268,6 +1273,9 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 err_detach_netdev:
 	mlx5e_detach_netdev(netdev_priv(netdev));
 
+err_cleanup_profile:
+	priv->profile->cleanup(priv);
+
 err_destroy_mdev_resources:
 	if (rep->vport == MLX5_VPORT_UPLINK)
 		mlx5e_destroy_mdev_resources(dev);
@@ -1294,6 +1302,7 @@ mlx5e_vport_rep_unload(struct mlx5_eswitch_rep *rep)
 	unregister_netdev(netdev);
 	mlx5e_rep_neigh_cleanup(rpriv);
 	mlx5e_detach_netdev(priv);
+	priv->profile->cleanup(priv);
 	if (rep->vport == MLX5_VPORT_UPLINK)
 		mlx5e_destroy_mdev_resources(priv->mdev);
 	mlx5e_destroy_netdev(priv);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
index 97b5fcb1f406..5889029c2adf 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -72,23 +72,14 @@ static void mlx5i_build_nic_params(struct mlx5_core_dev *mdev,
 }
 
 /* Called directly after IPoIB netdevice was created to initialize SW structs */
-int mlx5i_init(struct mlx5_core_dev *mdev,
-	       struct net_device *netdev,
-	       const struct mlx5e_profile *profile,
-	       void *ppriv)
+int mlx5i_init(struct mlx5_core_dev *mdev, struct net_device *netdev)
 {
 	struct mlx5e_priv *priv  = mlx5i_epriv(netdev);
-	int err;
-
-	err = mlx5e_netdev_init(netdev, priv, mdev, profile, ppriv);
-	if (err)
-		return err;
 
 	mlx5e_set_netdev_mtu_boundaries(priv);
 	netdev->mtu = netdev->max_mtu;
 
-	mlx5e_build_nic_params(priv, NULL, &priv->rss_params, &priv->channels.params,
-			       netdev->mtu);
+	mlx5e_build_nic_params(priv, NULL, netdev->mtu);
 	mlx5i_build_nic_params(mdev, &priv->channels.params);
 
 	mlx5e_timestamp_init(priv);
@@ -753,7 +744,14 @@ static int mlx5_rdma_setup_rn(struct ib_device *ibdev, u8 port_num,
 			goto destroy_ht;
 	}
 
-	prof->init(mdev, netdev, prof, ipriv);
+	err = mlx5e_netdev_init(netdev, epriv, mdev);
+	if (err)
+		goto destroy_mdev_resources;
+
+	epriv->profile = prof;
+	epriv->ppriv = ipriv;
+
+	prof->init(mdev, netdev);
 
 	err = mlx5e_attach_netdev(epriv);
 	if (err)
@@ -777,6 +775,7 @@ static int mlx5_rdma_setup_rn(struct ib_device *ibdev, u8 port_num,
 	prof->cleanup(epriv);
 	if (ipriv->sub_interface)
 		return err;
+destroy_mdev_resources:
 	mlx5e_destroy_mdev_resources(mdev);
 destroy_ht:
 	mlx5i_pkey_qpn_ht_cleanup(netdev);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h
index b79dc1e28c41..99d46fda9f82 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h
@@ -87,10 +87,7 @@ void mlx5i_dev_cleanup(struct net_device *dev);
 int mlx5i_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd);
 
 /* Parent profile functions */
-int mlx5i_init(struct mlx5_core_dev *mdev,
-	       struct net_device *netdev,
-	       const struct mlx5e_profile *profile,
-	       void *ppriv);
+int mlx5i_init(struct mlx5_core_dev *mdev, struct net_device *netdev);
 void mlx5i_cleanup(struct mlx5e_priv *priv);
 
 int mlx5i_update_nic_rx(struct mlx5e_priv *priv);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib_vlan.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib_vlan.c
index 7163d9f6c4a6..3d0a18a0bed4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib_vlan.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib_vlan.c
@@ -276,14 +276,12 @@ static int mlx5i_pkey_change_mtu(struct net_device *netdev, int new_mtu)
 
 /* Called directly after IPoIB netdevice was created to initialize SW structs */
 static int mlx5i_pkey_init(struct mlx5_core_dev *mdev,
-			   struct net_device *netdev,
-			   const struct mlx5e_profile *profile,
-			   void *ppriv)
+			   struct net_device *netdev)
 {
 	struct mlx5e_priv *priv  = mlx5i_epriv(netdev);
 	int err;
 
-	err = mlx5i_init(mdev, netdev, profile, ppriv);
+	err = mlx5i_init(mdev, netdev);
 	if (err)
 		return err;
 
-- 
2.29.2

Re: [net-next 01/14] net/mlx5e: Separate between netdev objects and mlx5e profiles initialization

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (refs/heads/master):

On Mon,  1 Feb 2021 22:54:44 -0800 you wrote:
> From: Saeed Mahameed <saeedm@nvidia.com>
> 
> 1) Initialize netdevice features and structures on netdevice allocation
>    and outside of the mlx5e profile.
> 
> 2) As now mlx5e netdevice private params will be setup on profile init only
>    after netdevice features are already set, we add  a call to
>    netde_update_features() to resolve any conflict.
>    This is nice since we reuse the fix_features ndo code if a profile
>    wants different default features, instead of duplicating features
>    conflict resolution code on profile initialization.
> 
> [...]

Here is the summary with links:
  - [net-next,01/14] net/mlx5e: Separate between netdev objects and mlx5e profiles initialization
    https://git.kernel.org/netdev/net-next/c/3ef14e463f6e
  - [net-next,02/14] net/mxl5e: Add change profile method
    https://git.kernel.org/netdev/net-next/c/c4d7eb57687f
  - [net-next,03/14] net/mlx5e: Refactor mlx5e_netdev_init/cleanup to mlx5e_priv_init/cleanup
    https://git.kernel.org/netdev/net-next/c/c9fd1e33e989
  - [net-next,04/14] net/mlx5e: Move netif_carrier_off() out of mlx5e_priv_init()
    https://git.kernel.org/netdev/net-next/c/1227bbc5d09e
  - [net-next,05/14] net/mlx5e: Move set vxlan nic info to profile init
    https://git.kernel.org/netdev/net-next/c/84db66124714
  - [net-next,06/14] net/mlx5e: Avoid false lock depenency warning on tc_ht
    https://git.kernel.org/netdev/net-next/c/9ba33339c043
  - [net-next,07/14] net/mlx5e: Move representor neigh init into profile enable
    https://git.kernel.org/netdev/net-next/c/6b424e13b010
  - [net-next,08/14] net/mlx5e: Enable napi in channel's activation stage
    https://git.kernel.org/netdev/net-next/c/7637e499e219
  - [net-next,09/14] net/mlx5e: Increase indirection RQ table size to 256
    https://git.kernel.org/netdev/net-next/c/1dd55ba2fb70
  - [net-next,10/14] net/mlx5e: remove h from printk format specifier
    https://git.kernel.org/netdev/net-next/c/1d3a3f3bfe3c
  - [net-next,11/14] net/mlx5e: kTLS, Improve TLS RX workqueue scope
    https://git.kernel.org/netdev/net-next/c/26432001b5c4
  - [net-next,12/14] net/mlx5e: accel, remove redundant space
    https://git.kernel.org/netdev/net-next/c/8271e341ed63
  - [net-next,13/14] net/mlx5e: CT: remove useless conversion to PTR_ERR then ERR_PTR
    https://git.kernel.org/netdev/net-next/c/902c02458925
  - [net-next,14/14] net/mlx5: DR, Avoid unnecessary csum recalculation on supporting devices
    https://git.kernel.org/netdev/net-next/c/a283ea1b9716

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

[net-next 02/14] net/mxl5e: Add change profile method

[Saeed Mahameed]

From: Saeed Mahameed <saeedm@nvidia.com>

Port nic netdevice will be used as uplink representor in downstream
patches. Add change profile method to allow changing a mlx5e netdevice
profile dynamically.

Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en.h  |  7 +-
 .../net/ethernet/mellanox/mlx5/core/en_main.c | 68 ++++++++++++++++++-
 .../net/ethernet/mellanox/mlx5/core/en_rep.c  |  2 +-
 3 files changed, 73 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index bf5de1e79134..fa461cfd6410 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -1154,9 +1154,10 @@ int mlx5e_ethtool_set_pauseparam(struct mlx5e_priv *priv,
 				 struct ethtool_pauseparam *pauseparam);
 
 /* mlx5e generic netdev management API */
-static inline unsigned int mlx5e_calc_max_nch(struct mlx5e_priv *priv)
+static inline unsigned int
+mlx5e_calc_max_nch(struct mlx5e_priv *priv, const struct mlx5e_profile *profile)
 {
-	return priv->netdev->num_rx_queues / max_t(u8, priv->profile->rq_groups, 1);
+	return priv->netdev->num_rx_queues / max_t(u8, profile->rq_groups, 1);
 }
 
 int mlx5e_netdev_init(struct net_device *netdev,
@@ -1168,6 +1169,8 @@ mlx5e_create_netdev(struct mlx5_core_dev *mdev, unsigned int txqs, unsigned int
 int mlx5e_attach_netdev(struct mlx5e_priv *priv);
 void mlx5e_detach_netdev(struct mlx5e_priv *priv);
 void mlx5e_destroy_netdev(struct mlx5e_priv *priv);
+int mlx5e_netdev_change_profile(struct mlx5e_priv *priv,
+				const struct mlx5e_profile *new_profile, void *new_ppriv);
 void mlx5e_set_netdev_mtu_boundaries(struct mlx5e_priv *priv);
 void mlx5e_build_nic_params(struct mlx5e_priv *priv, struct mlx5e_xsk *xsk, u16 mtu);
 void mlx5e_build_rq_params(struct mlx5_core_dev *mdev,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 260ced27014d..91f23871ded5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -4936,7 +4936,7 @@ void mlx5e_build_nic_params(struct mlx5e_priv *priv, struct mlx5e_xsk *xsk, u16
 	struct mlx5_core_dev *mdev = priv->mdev;
 	u8 rx_cq_period_mode;
 
-	priv->max_nch = mlx5e_calc_max_nch(priv);
+	priv->max_nch = mlx5e_calc_max_nch(priv, priv->profile);
 
 	params->sw_mtu = mtu;
 	params->hard_mtu = MLX5E_ETH_HARD_MTU;
@@ -5461,6 +5461,8 @@ int mlx5e_netdev_init(struct net_device *netdev,
 		      struct mlx5e_priv *priv,
 		      struct mlx5_core_dev *mdev)
 {
+	memset(priv, 0, sizeof(*priv));
+
 	/* priv init */
 	priv->mdev        = mdev;
 	priv->netdev      = netdev;
@@ -5615,6 +5617,70 @@ void mlx5e_detach_netdev(struct mlx5e_priv *priv)
 	cancel_work_sync(&priv->update_stats_work);
 }
 
+static int
+mlx5e_netdev_attach_profile(struct mlx5e_priv *priv,
+			    const struct mlx5e_profile *new_profile, void *new_ppriv)
+{
+	struct net_device *netdev = priv->netdev;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	err = mlx5e_netdev_init(netdev, priv, mdev);
+	if (err) {
+		mlx5_core_err(mdev, "mlx5e_netdev_init failed, err=%d\n", err);
+		return err;
+	}
+	priv->profile = new_profile;
+	priv->ppriv = new_ppriv;
+	err = new_profile->init(priv->mdev, priv->netdev);
+	if (err)
+		return err;
+	err = mlx5e_attach_netdev(priv);
+	if (err)
+		new_profile->cleanup(priv);
+	return err;
+}
+
+int mlx5e_netdev_change_profile(struct mlx5e_priv *priv,
+				const struct mlx5e_profile *new_profile, void *new_ppriv)
+{
+	unsigned int new_max_nch = mlx5e_calc_max_nch(priv, new_profile);
+	const struct mlx5e_profile *orig_profile = priv->profile;
+	void *orig_ppriv = priv->ppriv;
+	int err, rollback_err;
+
+	/* sanity */
+	if (new_max_nch != priv->max_nch) {
+		netdev_warn(priv->netdev,
+			    "%s: Replacing profile with different max channles\n",
+			    __func__);
+		return -EINVAL;
+	}
+
+	/* cleanup old profile */
+	mlx5e_detach_netdev(priv);
+	priv->profile->cleanup(priv);
+	mlx5e_netdev_cleanup(priv->netdev, priv);
+
+	err = mlx5e_netdev_attach_profile(priv, new_profile, new_ppriv);
+	if (err) { /* roll back to original profile */
+		netdev_warn(priv->netdev, "%s: new profile init failed, %d\n",
+			    __func__, err);
+		goto rollback;
+	}
+
+	return 0;
+
+rollback:
+	rollback_err = mlx5e_netdev_attach_profile(priv, orig_profile, orig_ppriv);
+	if (rollback_err) {
+		netdev_err(priv->netdev,
+			   "%s: failed to rollback to orig profile, %d\n",
+			   __func__, rollback_err);
+	}
+	return err;
+}
+
 void mlx5e_destroy_netdev(struct mlx5e_priv *priv)
 {
 	struct net_device *netdev = priv->netdev;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index d94d2ff9d312..c8a0f4c88d4b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -684,7 +684,7 @@ static void mlx5e_build_rep_params(struct net_device *netdev)
 					 MLX5_CQ_PERIOD_MODE_START_FROM_CQE :
 					 MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
 
-	priv->max_nch = mlx5e_calc_max_nch(priv);
+	priv->max_nch = mlx5e_calc_max_nch(priv, priv->profile);
 	params = &priv->channels.params;
 
 	params->num_channels = MLX5E_REP_PARAMS_DEF_NUM_CHANNELS;
-- 
2.29.2

[net-next 03/14] net/mlx5e: Refactor mlx5e_netdev_init/cleanup to mlx5e_priv_init/cleanup

[Saeed Mahameed]

From: Roi Dayan <roid@nvidia.com>

We actually initialize priv and not netdev. The only call to
set netdev carrier will be moved in the following commit.

Signed-off-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en.h  |  8 ++++----
 .../net/ethernet/mellanox/mlx5/core/en_main.c | 20 +++++++++----------
 .../ethernet/mellanox/mlx5/core/ipoib/ipoib.c |  4 ++--
 3 files changed, 16 insertions(+), 16 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index fa461cfd6410..8cc80c31341f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -1160,10 +1160,10 @@ mlx5e_calc_max_nch(struct mlx5e_priv *priv, const struct mlx5e_profile *profile)
 	return priv->netdev->num_rx_queues / max_t(u8, profile->rq_groups, 1);
 }
 
-int mlx5e_netdev_init(struct net_device *netdev,
-		      struct mlx5e_priv *priv,
-		      struct mlx5_core_dev *mdev);
-void mlx5e_netdev_cleanup(struct net_device *netdev, struct mlx5e_priv *priv);
+int mlx5e_priv_init(struct mlx5e_priv *priv,
+		    struct net_device *netdev,
+		    struct mlx5_core_dev *mdev);
+void mlx5e_priv_cleanup(struct mlx5e_priv *priv);
 struct net_device *
 mlx5e_create_netdev(struct mlx5_core_dev *mdev, unsigned int txqs, unsigned int rxqs);
 int mlx5e_attach_netdev(struct mlx5e_priv *priv);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 91f23871ded5..177e076f6cce 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -5457,9 +5457,9 @@ static const struct mlx5e_profile mlx5e_nic_profile = {
 };
 
 /* mlx5e generic netdev management API (move to en_common.c) */
-int mlx5e_netdev_init(struct net_device *netdev,
-		      struct mlx5e_priv *priv,
-		      struct mlx5_core_dev *mdev)
+int mlx5e_priv_init(struct mlx5e_priv *priv,
+		    struct net_device *netdev,
+		    struct mlx5_core_dev *mdev)
 {
 	memset(priv, 0, sizeof(*priv));
 
@@ -5494,7 +5494,7 @@ int mlx5e_netdev_init(struct net_device *netdev,
 	return -ENOMEM;
 }
 
-void mlx5e_netdev_cleanup(struct net_device *netdev, struct mlx5e_priv *priv)
+void mlx5e_priv_cleanup(struct mlx5e_priv *priv)
 {
 	int i;
 
@@ -5518,9 +5518,9 @@ mlx5e_create_netdev(struct mlx5_core_dev *mdev, unsigned int txqs, unsigned int
 		return NULL;
 	}
 
-	err = mlx5e_netdev_init(netdev, netdev_priv(netdev), mdev);
+	err = mlx5e_priv_init(netdev_priv(netdev), netdev, mdev);
 	if (err) {
-		mlx5_core_err(mdev, "mlx5e_netdev_init failed, err=%d\n", err);
+		mlx5_core_err(mdev, "mlx5e_priv_init failed, err=%d\n", err);
 		goto err_free_netdev;
 	}
 	dev_net_set(netdev, mlx5_core_net(mdev));
@@ -5625,9 +5625,9 @@ mlx5e_netdev_attach_profile(struct mlx5e_priv *priv,
 	struct mlx5_core_dev *mdev = priv->mdev;
 	int err;
 
-	err = mlx5e_netdev_init(netdev, priv, mdev);
+	err = mlx5e_priv_init(priv, netdev, mdev);
 	if (err) {
-		mlx5_core_err(mdev, "mlx5e_netdev_init failed, err=%d\n", err);
+		mlx5_core_err(mdev, "mlx5e_priv_init failed, err=%d\n", err);
 		return err;
 	}
 	priv->profile = new_profile;
@@ -5660,7 +5660,7 @@ int mlx5e_netdev_change_profile(struct mlx5e_priv *priv,
 	/* cleanup old profile */
 	mlx5e_detach_netdev(priv);
 	priv->profile->cleanup(priv);
-	mlx5e_netdev_cleanup(priv->netdev, priv);
+	mlx5e_priv_cleanup(priv);
 
 	err = mlx5e_netdev_attach_profile(priv, new_profile, new_ppriv);
 	if (err) { /* roll back to original profile */
@@ -5685,7 +5685,7 @@ void mlx5e_destroy_netdev(struct mlx5e_priv *priv)
 {
 	struct net_device *netdev = priv->netdev;
 
-	mlx5e_netdev_cleanup(netdev, priv);
+	mlx5e_priv_cleanup(priv);
 	free_netdev(netdev);
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
index 5889029c2adf..8641bd9bbb53 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -103,7 +103,7 @@ int mlx5i_init(struct mlx5_core_dev *mdev, struct net_device *netdev)
 /* Called directly before IPoIB netdevice is destroyed to cleanup SW structs */
 void mlx5i_cleanup(struct mlx5e_priv *priv)
 {
-	mlx5e_netdev_cleanup(priv->netdev, priv);
+	mlx5e_priv_cleanup(priv);
 }
 
 static void mlx5i_grp_sw_update_stats(struct mlx5e_priv *priv)
@@ -744,7 +744,7 @@ static int mlx5_rdma_setup_rn(struct ib_device *ibdev, u8 port_num,
 			goto destroy_ht;
 	}
 
-	err = mlx5e_netdev_init(netdev, epriv, mdev);
+	err = mlx5e_priv_init(epriv, netdev, mdev);
 	if (err)
 		goto destroy_mdev_resources;
 
-- 
2.29.2

[net-next 04/14] net/mlx5e: Move netif_carrier_off() out of mlx5e_priv_init()

[Saeed Mahameed]

From: Roi Dayan <roid@nvidia.com>

It's not part of priv initialization.

Signed-off-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c     | 6 +++---
 drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 1 +
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 177e076f6cce..e468d8329c2a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -5483,9 +5483,6 @@ int mlx5e_priv_init(struct mlx5e_priv *priv,
 	if (!priv->wq)
 		goto err_free_cpumask;
 
-	/* netdev init */
-	netif_carrier_off(netdev);
-
 	return 0;
 
 err_free_cpumask:
@@ -5523,6 +5520,8 @@ mlx5e_create_netdev(struct mlx5_core_dev *mdev, unsigned int txqs, unsigned int
 		mlx5_core_err(mdev, "mlx5e_priv_init failed, err=%d\n", err);
 		goto err_free_netdev;
 	}
+
+	netif_carrier_off(netdev);
 	dev_net_set(netdev, mlx5_core_net(mdev));
 
 	return netdev;
@@ -5630,6 +5629,7 @@ mlx5e_netdev_attach_profile(struct mlx5e_priv *priv,
 		mlx5_core_err(mdev, "mlx5e_priv_init failed, err=%d\n", err);
 		return err;
 	}
+	netif_carrier_off(netdev);
 	priv->profile = new_profile;
 	priv->ppriv = new_ppriv;
 	err = new_profile->init(priv->mdev, priv->netdev);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
index 8641bd9bbb53..1eeca45cfcdf 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -76,6 +76,7 @@ int mlx5i_init(struct mlx5_core_dev *mdev, struct net_device *netdev)
 {
 	struct mlx5e_priv *priv  = mlx5i_epriv(netdev);
 
+	netif_carrier_off(netdev);
 	mlx5e_set_netdev_mtu_boundaries(priv);
 	netdev->mtu = netdev->max_mtu;
 
-- 
2.29.2

[net-next 05/14] net/mlx5e: Move set vxlan nic info to profile init

[Saeed Mahameed]

From: Roi Dayan <roid@nvidia.com>

Since its profile dependent let's init the vxlan info
as part of profile initialization.

Signed-off-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c |  3 +--
 drivers/net/ethernet/mellanox/mlx5/core/en_rep.c  | 14 ++++++++++----
 2 files changed, 11 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index e468d8329c2a..b9d2cb6f178d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -5101,8 +5101,6 @@ static void mlx5e_build_nic_netdev(struct net_device *netdev)
 	netdev->hw_features      |= NETIF_F_HW_VLAN_CTAG_FILTER;
 	netdev->hw_features      |= NETIF_F_HW_VLAN_STAG_TX;
 
-	mlx5e_vxlan_set_netdev_info(priv);
-
 	if (mlx5e_tunnel_any_tx_proto_supported(mdev)) {
 		netdev->hw_enc_features |= NETIF_F_HW_CSUM;
 		netdev->hw_enc_features |= NETIF_F_TSO;
@@ -5229,6 +5227,7 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev,
 	int err;
 
 	mlx5e_build_nic_params(priv, &priv->xsk, netdev->mtu);
+	mlx5e_vxlan_set_netdev_info(priv);
 
 	mlx5e_timestamp_init(priv);
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index c8a0f4c88d4b..45669a1db426 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -717,15 +717,12 @@ static void mlx5e_build_rep_netdev(struct net_device *netdev,
 				   struct mlx5_core_dev *mdev,
 				   struct mlx5_eswitch_rep *rep)
 {
-	struct mlx5e_priv *priv = netdev_priv(netdev);
-
 	SET_NETDEV_DEV(netdev, mdev->device);
 	if (rep->vport == MLX5_VPORT_UPLINK) {
 		netdev->netdev_ops = &mlx5e_netdev_ops_uplink_rep;
 		/* we want a persistent mac for the uplink rep */
 		mlx5_query_mac_address(mdev, netdev->dev_addr);
 		netdev->ethtool_ops = &mlx5e_uplink_rep_ethtool_ops;
-		mlx5e_vxlan_set_netdev_info(priv);
 		mlx5e_dcbnl_build_rep_netdev(netdev);
 	} else {
 		netdev->netdev_ops = &mlx5e_netdev_ops_rep;
@@ -767,6 +764,15 @@ static int mlx5e_init_rep(struct mlx5_core_dev *mdev,
 	return 0;
 }
 
+static int mlx5e_init_ul_rep(struct mlx5_core_dev *mdev,
+			     struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	mlx5e_vxlan_set_netdev_info(priv);
+	return mlx5e_init_rep(mdev, netdev);
+}
+
 static void mlx5e_cleanup_rep(struct mlx5e_priv *priv)
 {
 }
@@ -1165,7 +1171,7 @@ static const struct mlx5e_profile mlx5e_rep_profile = {
 };
 
 static const struct mlx5e_profile mlx5e_uplink_rep_profile = {
-	.init			= mlx5e_init_rep,
+	.init			= mlx5e_init_ul_rep,
 	.cleanup		= mlx5e_cleanup_rep,
 	.init_rx		= mlx5e_init_ul_rep_rx,
 	.cleanup_rx		= mlx5e_cleanup_ul_rep_rx,
-- 
2.29.2

[net-next 06/14] net/mlx5e: Avoid false lock depenency warning on tc_ht

[Saeed Mahameed]

From: Roi Dayan <roid@nvidia.com>

To avoid false lock dependency warning set the tc_ht lock
class different than the lock class of the ht being used when deleting
last flow from a group and then deleting a group, we get into del_sw_flow_group()
which call rhashtable_destroy on fg->ftes_hash which will take ht->mutex but
it's different than the ht->mutex here.

======================================================
WARNING: possible circular locking dependency detected
5.11.0-rc4_net_next_mlx5_949fdcc #1 Not tainted
------------------------------------------------------
modprobe/12950 is trying to acquire lock:
ffff88816510f910 (&node->lock){++++}-{3:3}, at: mlx5_del_flow_rules+0x2a/0x210 [mlx5_core]

but task is already holding lock:
ffff88815834e3e8 (&ht->mutex){+.+.}-{3:3}, at: rhashtable_free_and_destroy+0x37/0x340

which lock already depends on the new lock.

Signed-off-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 8fd38ad8113b..280ea1e1e039 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -190,6 +190,14 @@ struct mlx5e_tc_attr_to_reg_mapping mlx5e_tc_attr_to_reg_mappings[] = {
 	[NIC_ZONE_RESTORE_TO_REG] = nic_zone_restore_to_reg_ct,
 };
 
+/* To avoid false lock dependency warning set the tc_ht lock
+ * class different than the lock class of the ht being used when deleting
+ * last flow from a group and then deleting a group, we get into del_sw_flow_group()
+ * which call rhashtable_destroy on fg->ftes_hash which will take ht->mutex but
+ * it's different than the ht->mutex here.
+ */
+static struct lock_class_key tc_ht_lock_key;
+
 static void mlx5e_put_flow_tunnel_id(struct mlx5e_tc_flow *flow);
 
 void
@@ -5215,6 +5223,8 @@ int mlx5e_tc_nic_init(struct mlx5e_priv *priv)
 	if (err)
 		return err;
 
+	lockdep_set_class(&tc->ht.mutex, &tc_ht_lock_key);
+
 	if (MLX5_CAP_FLOWTABLE_NIC_RX(priv->mdev, ignore_flow_level)) {
 		attr.flags = MLX5_CHAINS_AND_PRIOS_SUPPORTED |
 			MLX5_CHAINS_IGNORE_FLOW_LEVEL_SUPPORTED;
@@ -5333,6 +5343,8 @@ int mlx5e_tc_esw_init(struct rhashtable *tc_ht)
 	if (err)
 		goto err_ht_init;
 
+	lockdep_set_class(&tc_ht->mutex, &tc_ht_lock_key);
+
 	return err;
 
 err_ht_init:
-- 
2.29.2

[net-next 07/14] net/mlx5e: Move representor neigh init into profile enable

[Saeed Mahameed]

From: Roi Dayan <roid@nvidia.com>

Also cleanup neigh in profile disable.
This is for logical separation.

Signed-off-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 .../mellanox/mlx5/core/en/rep/neigh.c         | 18 ++++++++----
 .../net/ethernet/mellanox/mlx5/core/en_rep.c  | 29 ++++++++++---------
 2 files changed, 29 insertions(+), 18 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c
index 58e27038c947..616ee585a985 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c
@@ -279,7 +279,7 @@ int mlx5e_rep_neigh_init(struct mlx5e_rep_priv *rpriv)
 
 	err = rhashtable_init(&neigh_update->neigh_ht, &mlx5e_neigh_ht_params);
 	if (err)
-		return err;
+		goto out_err;
 
 	INIT_LIST_HEAD(&neigh_update->neigh_list);
 	mutex_init(&neigh_update->encap_lock);
@@ -287,14 +287,19 @@ int mlx5e_rep_neigh_init(struct mlx5e_rep_priv *rpriv)
 			  mlx5e_rep_neigh_stats_work);
 	mlx5e_rep_neigh_update_init_interval(rpriv);
 
-	rpriv->neigh_update.netevent_nb.notifier_call = mlx5e_rep_netevent_event;
-	err = register_netevent_notifier(&rpriv->neigh_update.netevent_nb);
+	neigh_update->netevent_nb.notifier_call = mlx5e_rep_netevent_event;
+	err = register_netevent_notifier(&neigh_update->netevent_nb);
 	if (err)
-		goto out_err;
+		goto out_notifier;
 	return 0;
 
-out_err:
+out_notifier:
+	neigh_update->netevent_nb.notifier_call = NULL;
 	rhashtable_destroy(&neigh_update->neigh_ht);
+out_err:
+	netdev_warn(rpriv->netdev,
+		    "Failed to initialize neighbours handling for vport %d\n",
+		    rpriv->rep->vport);
 	return err;
 }
 
@@ -303,6 +308,9 @@ void mlx5e_rep_neigh_cleanup(struct mlx5e_rep_priv *rpriv)
 	struct mlx5e_neigh_update_table *neigh_update = &rpriv->neigh_update;
 	struct mlx5e_priv *priv = netdev_priv(rpriv->netdev);
 
+	if (!rpriv->neigh_update.netevent_nb.notifier_call)
+		return;
+
 	unregister_netevent_notifier(&neigh_update->netevent_nb);
 
 	flush_workqueue(priv->wq); /* flush neigh update works */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index 45669a1db426..84eeaa33033f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -1051,7 +1051,17 @@ static void mlx5e_cleanup_rep_tx(struct mlx5e_priv *priv)
 
 static void mlx5e_rep_enable(struct mlx5e_priv *priv)
 {
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+
 	mlx5e_set_netdev_mtu_boundaries(priv);
+	mlx5e_rep_neigh_init(rpriv);
+}
+
+static void mlx5e_rep_disable(struct mlx5e_priv *priv)
+{
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
+
+	mlx5e_rep_neigh_cleanup(rpriv);
 }
 
 static int mlx5e_update_rep_rx(struct mlx5e_priv *priv)
@@ -1086,6 +1096,7 @@ static int uplink_rep_async_event(struct notifier_block *nb, unsigned long event
 
 static void mlx5e_uplink_rep_enable(struct mlx5e_priv *priv)
 {
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
 	struct net_device *netdev = priv->netdev;
 	struct mlx5_core_dev *mdev = priv->mdev;
 	u16 max_mtu;
@@ -1104,12 +1115,15 @@ static void mlx5e_uplink_rep_enable(struct mlx5e_priv *priv)
 	mlx5_notifier_register(mdev, &priv->events_nb);
 	mlx5e_dcbnl_initialize(priv);
 	mlx5e_dcbnl_init_app(priv);
+	mlx5e_rep_neigh_init(rpriv);
 }
 
 static void mlx5e_uplink_rep_disable(struct mlx5e_priv *priv)
 {
+	struct mlx5e_rep_priv *rpriv = priv->ppriv;
 	struct mlx5_core_dev *mdev = priv->mdev;
 
+	mlx5e_rep_neigh_cleanup(rpriv);
 	mlx5e_dcbnl_delete_app(priv);
 	mlx5_notifier_unregister(mdev, &priv->events_nb);
 	mlx5e_rep_tc_disable(priv);
@@ -1161,6 +1175,7 @@ static const struct mlx5e_profile mlx5e_rep_profile = {
 	.init_tx		= mlx5e_init_rep_tx,
 	.cleanup_tx		= mlx5e_cleanup_rep_tx,
 	.enable		        = mlx5e_rep_enable,
+	.disable	        = mlx5e_rep_disable,
 	.update_rx		= mlx5e_update_rep_rx,
 	.update_stats           = mlx5e_stats_update_ndo_stats,
 	.rx_handlers            = &mlx5e_rx_handlers_rep,
@@ -1252,20 +1267,12 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 		goto err_cleanup_profile;
 	}
 
-	err = mlx5e_rep_neigh_init(rpriv);
-	if (err) {
-		netdev_warn(netdev,
-			    "Failed to initialized neighbours handling for vport %d\n",
-			    rep->vport);
-		goto err_detach_netdev;
-	}
-
 	err = register_netdev(netdev);
 	if (err) {
 		netdev_warn(netdev,
 			    "Failed to register representor netdev for vport %d\n",
 			    rep->vport);
-		goto err_neigh_cleanup;
+		goto err_detach_netdev;
 	}
 
 	dl_port = mlx5_esw_offloads_devlink_port(dev->priv.eswitch, rpriv->rep->vport);
@@ -1273,9 +1280,6 @@ mlx5e_vport_rep_load(struct mlx5_core_dev *dev, struct mlx5_eswitch_rep *rep)
 		devlink_port_type_eth_set(dl_port, netdev);
 	return 0;
 
-err_neigh_cleanup:
-	mlx5e_rep_neigh_cleanup(rpriv);
-
 err_detach_netdev:
 	mlx5e_detach_netdev(netdev_priv(netdev));
 
@@ -1306,7 +1310,6 @@ mlx5e_vport_rep_unload(struct mlx5_eswitch_rep *rep)
 	if (dl_port)
 		devlink_port_type_clear(dl_port);
 	unregister_netdev(netdev);
-	mlx5e_rep_neigh_cleanup(rpriv);
 	mlx5e_detach_netdev(priv);
 	priv->profile->cleanup(priv);
 	if (rep->vport == MLX5_VPORT_UPLINK)
-- 
2.29.2

[net-next 08/14] net/mlx5e: Enable napi in channel's activation stage

[Saeed Mahameed]

From: Tariq Toukan <tariqt@mellanox.com>

The channel's napi is first needed upon activation, not creation.
Minimize its enabled scope by moving it from the channel's open/close
stage into the activate/deactivate stage.

Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c  | 12 ++++++------
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 14 ++++++--------
 2 files changed, 12 insertions(+), 14 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c
index eeddd1137dda..a76cfefec708 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c
@@ -428,16 +428,13 @@ static int mlx5e_ptp_open_queues(struct mlx5e_port_ptp *c,
 	if (err)
 		return err;
 
-	napi_enable(&c->napi);
-
 	err = mlx5e_ptp_open_txqsqs(c, cparams);
 	if (err)
-		goto disable_napi;
+		goto close_cqs;
 
 	return 0;
 
-disable_napi:
-	napi_disable(&c->napi);
+close_cqs:
 	mlx5e_ptp_close_cqs(c);
 
 	return err;
@@ -446,7 +443,6 @@ static int mlx5e_ptp_open_queues(struct mlx5e_port_ptp *c,
 static void mlx5e_ptp_close_queues(struct mlx5e_port_ptp *c)
 {
 	mlx5e_ptp_close_txqsqs(c);
-	napi_disable(&c->napi);
 	mlx5e_ptp_close_cqs(c);
 }
 
@@ -515,6 +511,8 @@ void mlx5e_ptp_activate_channel(struct mlx5e_port_ptp *c)
 {
 	int tc;
 
+	napi_enable(&c->napi);
+
 	for (tc = 0; tc < c->num_tc; tc++)
 		mlx5e_activate_txqsq(&c->ptpsq[tc].txqsq);
 }
@@ -525,4 +523,6 @@ void mlx5e_ptp_deactivate_channel(struct mlx5e_port_ptp *c)
 
 	for (tc = 0; tc < c->num_tc; tc++)
 		mlx5e_deactivate_txqsq(&c->ptpsq[tc].txqsq);
+
+	napi_disable(&c->napi);
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index b9d2cb6f178d..41c611197211 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -1892,13 +1892,11 @@ static int mlx5e_open_queues(struct mlx5e_channel *c,
 	if (err)
 		goto err_close_rx_cq;
 
-	napi_enable(&c->napi);
-
 	spin_lock_init(&c->async_icosq_lock);
 
 	err = mlx5e_open_icosq(c, params, &cparam->async_icosq, &c->async_icosq);
 	if (err)
-		goto err_disable_napi;
+		goto err_close_xdpsq_cq;
 
 	err = mlx5e_open_icosq(c, params, &cparam->icosq, &c->icosq);
 	if (err)
@@ -1941,9 +1939,7 @@ static int mlx5e_open_queues(struct mlx5e_channel *c,
 err_close_async_icosq:
 	mlx5e_close_icosq(&c->async_icosq);
 
-err_disable_napi:
-	napi_disable(&c->napi);
-
+err_close_xdpsq_cq:
 	if (c->xdp)
 		mlx5e_close_cq(&c->rq_xdpsq.cq);
 
@@ -1974,7 +1970,6 @@ static void mlx5e_close_queues(struct mlx5e_channel *c)
 	mlx5e_close_sqs(c);
 	mlx5e_close_icosq(&c->icosq);
 	mlx5e_close_icosq(&c->async_icosq);
-	napi_disable(&c->napi);
 	if (c->xdp)
 		mlx5e_close_cq(&c->rq_xdpsq.cq);
 	mlx5e_close_cq(&c->rq.cq);
@@ -2059,6 +2054,8 @@ static void mlx5e_activate_channel(struct mlx5e_channel *c)
 {
 	int tc;
 
+	napi_enable(&c->napi);
+
 	for (tc = 0; tc < c->num_tc; tc++)
 		mlx5e_activate_txqsq(&c->sq[tc]);
 	mlx5e_activate_icosq(&c->icosq);
@@ -2081,8 +2078,9 @@ static void mlx5e_deactivate_channel(struct mlx5e_channel *c)
 	mlx5e_deactivate_icosq(&c->icosq);
 	for (tc = 0; tc < c->num_tc; tc++)
 		mlx5e_deactivate_txqsq(&c->sq[tc]);
-
 	mlx5e_qos_deactivate_queues(c);
+
+	napi_disable(&c->napi);
 }
 
 static void mlx5e_close_channel(struct mlx5e_channel *c)
-- 
2.29.2

[net-next 09/14] net/mlx5e: Increase indirection RQ table size to 256

[Saeed Mahameed]

From: Noam Stolero <noams@nvidia.com>

Increasing the indirection RQ table size from 128 to 256 improves the
packet distribution over the NIC HW queues for various cases.

Let's take a look at the following scenario:
Assuming RSS result distributed uniformly and indirection table is filled
with queues in a cyclic manner.
Let N be the number of queues on a given setup.
If 256%N = 128%N = 0, then all queues have the same probability to be
chosen for a given RSS result.
This case doesn't improves nor degrade by this change.

If 256%N != 0 and 128%N != 0, there is a remainder which will favor some
queues. Increasing the indirection RQ table size to 256 reduce the ratio
between the favored queues probability to be selected to the rest of the
queues and improves the distribution.

For example, let's assume the number of queues is 56.
For a table size of 128, we have 128%56=16 queues which will have a 3/128
probability to be chosen and 2/128 for the rest 40.
16 queues have 1.5 times the probability to be chosen over the other 40.

For a table size of 256, we have 256%56=32 queues which will have a 5/256
probability to be chosen and 4/256 probability for the rest 24 queues.
Here 32 queues have 1.25 more probability to be chosen over the other 24.

This shows that the larger indirection table size would more likely cause
an even distribution.

This change also aligns our mlx5 driver's indirection table size with
other vendors.

Signed-off-by: Noam Stolero <noams@nvidia.com>
Reviewed-by: Tal Gilboa <talgi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 8cc80c31341f..a8e31cdd4a4e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -137,10 +137,10 @@ struct page_pool;
 #define MLX5E_PARAMS_DEFAULT_MIN_RX_WQES                0x80
 #define MLX5E_PARAMS_DEFAULT_MIN_RX_WQES_MPW            0x2
 
-#define MLX5E_LOG_INDIR_RQT_SIZE       0x7
+#define MLX5E_LOG_INDIR_RQT_SIZE       0x8
 #define MLX5E_INDIR_RQT_SIZE           BIT(MLX5E_LOG_INDIR_RQT_SIZE)
 #define MLX5E_MIN_NUM_CHANNELS         0x1
-#define MLX5E_MAX_NUM_CHANNELS         MLX5E_INDIR_RQT_SIZE
+#define MLX5E_MAX_NUM_CHANNELS         (MLX5E_INDIR_RQT_SIZE / 2)
 #define MLX5E_MAX_NUM_SQS              (MLX5E_MAX_NUM_CHANNELS * MLX5E_MAX_NUM_TC)
 #define MLX5E_TX_CQ_POLL_BUDGET        128
 #define MLX5E_TX_XSK_POLL_BUDGET       64
-- 
2.29.2

[net-next 10/14] net/mlx5e: remove h from printk format specifier

[Saeed Mahameed]

From: Tom Rix <trix@redhat.com>

This change fixes the checkpatch warning described in this commit
commit cbacb5ab0aa0 ("docs: printk-formats: Stop encouraging use of unnecessary %h[xudi] and %hh[xudi]")

Standard integer promotion is already done and %hx and %hhx is useless
so do not encourage the use of %hh[xudi] or %h[xudi].

Signed-off-by: Tom Rix <trix@redhat.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en/params.c | 2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c   | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
index 43271a3856ca..36381a2ed5a5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
@@ -179,7 +179,7 @@ int mlx5e_validate_params(struct mlx5e_priv *priv, struct mlx5e_params *params)
 
 	stop_room = mlx5e_calc_sq_stop_room(priv->mdev, params);
 	if (stop_room >= sq_size) {
-		netdev_err(priv->netdev, "Stop room %hu is bigger than the SQ size %zu\n",
+		netdev_err(priv->netdev, "Stop room %u is bigger than the SQ size %zu\n",
 			   stop_room, sq_size);
 		return -EINVAL;
 	}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 41c611197211..2a6f9d042f51 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -3888,7 +3888,7 @@ static int set_feature_lro(struct net_device *netdev, bool enable)
 	mutex_lock(&priv->state_lock);
 
 	if (enable && priv->xsk.refcnt) {
-		netdev_warn(netdev, "LRO is incompatible with AF_XDP (%hu XSKs are active)\n",
+		netdev_warn(netdev, "LRO is incompatible with AF_XDP (%u XSKs are active)\n",
 			    priv->xsk.refcnt);
 		err = -EINVAL;
 		goto out;
@@ -4139,7 +4139,7 @@ static bool mlx5e_xsk_validate_mtu(struct net_device *netdev,
 			max_mtu_page = mlx5e_xdp_max_mtu(new_params, &xsk);
 			max_mtu = min(max_mtu_frame, max_mtu_page);
 
-			netdev_err(netdev, "MTU %d is too big for an XSK running on channel %hu. Try MTU <= %d\n",
+			netdev_err(netdev, "MTU %d is too big for an XSK running on channel %u. Try MTU <= %d\n",
 				   new_params->sw_mtu, ix, max_mtu);
 			return false;
 		}
-- 
2.29.2

[net-next 11/14] net/mlx5e: kTLS, Improve TLS RX workqueue scope

[Saeed Mahameed]

From: Tariq Toukan <tariqt@nvidia.com>

The TLS RX workqueue is needed only when kTLS RX device offload
is supported.

Move its creation from the general TLS init function to the
kTLS RX init.
Create it once at init time if supported, avoid creation/destroy
everytime the feature bit is toggled.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 .../mellanox/mlx5/core/en_accel/ktls.c        | 24 ++++++++++++++++---
 .../mellanox/mlx5/core/en_accel/tls.c         |  7 ------
 2 files changed, 21 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.c
index 1b392696280d..95293ee0d38d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.c
@@ -2,6 +2,7 @@
 // Copyright (c) 2019 Mellanox Technologies.
 
 #include "en.h"
+#include "en_accel/tls.h"
 #include "en_accel/ktls.h"
 #include "en_accel/ktls_utils.h"
 #include "en_accel/fs_tcp.h"
@@ -86,16 +87,33 @@ int mlx5e_ktls_set_feature_rx(struct net_device *netdev, bool enable)
 
 int mlx5e_ktls_init_rx(struct mlx5e_priv *priv)
 {
-	int err = 0;
+	int err;
 
-	if (priv->netdev->features & NETIF_F_HW_TLS_RX)
+	if (!mlx5_accel_is_ktls_rx(priv->mdev))
+		return 0;
+
+	priv->tls->rx_wq = create_singlethread_workqueue("mlx5e_tls_rx");
+	if (!priv->tls->rx_wq)
+		return -ENOMEM;
+
+	if (priv->netdev->features & NETIF_F_HW_TLS_RX) {
 		err = mlx5e_accel_fs_tcp_create(priv);
+		if (err) {
+			destroy_workqueue(priv->tls->rx_wq);
+			return err;
+		}
+	}
 
-	return err;
+	return 0;
 }
 
 void mlx5e_ktls_cleanup_rx(struct mlx5e_priv *priv)
 {
+	if (!mlx5_accel_is_ktls_rx(priv->mdev))
+		return;
+
 	if (priv->netdev->features & NETIF_F_HW_TLS_RX)
 		mlx5e_accel_fs_tcp_destroy(priv);
+
+	destroy_workqueue(priv->tls->rx_wq);
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.c
index fee991f5ee7c..d6b21b899dbc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.c
@@ -231,12 +231,6 @@ int mlx5e_tls_init(struct mlx5e_priv *priv)
 	if (!tls)
 		return -ENOMEM;
 
-	tls->rx_wq = create_singlethread_workqueue("mlx5e_tls_rx");
-	if (!tls->rx_wq) {
-		kfree(tls);
-		return -ENOMEM;
-	}
-
 	priv->tls = tls;
 	return 0;
 }
@@ -248,7 +242,6 @@ void mlx5e_tls_cleanup(struct mlx5e_priv *priv)
 	if (!tls)
 		return;
 
-	destroy_workqueue(tls->rx_wq);
 	kfree(tls);
 	priv->tls = NULL;
 }
-- 
2.29.2

[net-next 12/14] net/mlx5e: accel, remove redundant space

[Saeed Mahameed]

From: Saeed Mahameed <saeedm@nvidia.com>

Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
index 6488098d2700..959bb6cd7203 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
@@ -85,7 +85,7 @@ mlx5e_tx_tunnel_accel(struct sk_buff *skb, struct mlx5_wqe_eth_seg *eseg, u16 ih
 	}
 
 	mlx5e_set_eseg_swp(skb, eseg, &swp_spec);
-	if (skb_vlan_tag_present(skb) &&  ihs)
+	if (skb_vlan_tag_present(skb) && ihs)
 		mlx5e_eseg_swp_offsets_add_vlan(eseg);
 }
 
-- 
2.29.2

[net-next 13/14] net/mlx5e: CT: remove useless conversion to PTR_ERR then ERR_PTR

[Saeed Mahameed]

From: Saeed Mahameed <saeedm@nvidia.com>

Just return the ptr directly.

Reported-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
index e417904ae17f..40aaa105b2fc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
@@ -770,7 +770,6 @@ mlx5_tc_ct_shared_counter_get(struct mlx5_tc_ct_priv *ct_priv,
 	struct mlx5_ct_counter *shared_counter;
 	struct mlx5_ct_entry *rev_entry;
 	__be16 tmp_port;
-	int ret;
 
 	/* get the reversed tuple */
 	tmp_port = rev_tuple.port.src;
@@ -804,10 +803,8 @@ mlx5_tc_ct_shared_counter_get(struct mlx5_tc_ct_priv *ct_priv,
 	mutex_unlock(&ct_priv->shared_counter_lock);
 
 	shared_counter = mlx5_tc_ct_counter_create(ct_priv);
-	if (IS_ERR(shared_counter)) {
-		ret = PTR_ERR(shared_counter);
-		return ERR_PTR(ret);
-	}
+	if (IS_ERR(shared_counter))
+		return shared_counter;
 
 	shared_counter->is_shared = true;
 	refcount_set(&shared_counter->refcount, 1);
-- 
2.29.2

[net-next 14/14] net/mlx5: DR, Avoid unnecessary csum recalculation on supporting devices

[Saeed Mahameed]

From: Yevgeny Kliteynik <kliteyn@nvidia.com>

If as part of the actions the TTL of the packet is modified, the packet's
checksum needs to be recalculated. Connect-X6DX can handle this csum
recalculation natively. Older devices require this additional recalculation.

Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Reviewed-by: Alex Vesker <valex@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 .../net/ethernet/mellanox/mlx5/core/steering/dr_action.c | 9 +++++----
 .../net/ethernet/mellanox/mlx5/core/steering/dr_ste.c    | 5 +++++
 .../net/ethernet/mellanox/mlx5/core/steering/dr_types.h  | 2 ++
 3 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
index 27c2b8416d02..28a7971cac6a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
@@ -447,7 +447,8 @@ int mlx5dr_actions_build_ste_arr(struct mlx5dr_matcher *matcher,
 		case DR_ACTION_TYP_MODIFY_HDR:
 			attr.modify_index = action->rewrite.index;
 			attr.modify_actions = action->rewrite.num_of_actions;
-			recalc_cs_required = action->rewrite.modify_ttl;
+			recalc_cs_required = action->rewrite.modify_ttl &&
+					     !mlx5dr_ste_supp_ttl_cs_recalc(&dmn->info.caps);
 			break;
 		case DR_ACTION_TYP_L2_TO_TNL_L2:
 		case DR_ACTION_TYP_L2_TO_TNL_L3:
@@ -501,9 +502,9 @@ int mlx5dr_actions_build_ste_arr(struct mlx5dr_matcher *matcher,
 	*new_hw_ste_arr_sz = nic_matcher->num_of_builders;
 	last_ste = ste_arr + DR_STE_SIZE * (nic_matcher->num_of_builders - 1);
 
-	/* Due to a HW bug, modifying TTL on RX flows will cause an incorrect
-	 * checksum calculation. In this case we will use a FW table to
-	 * recalculate.
+	/* Due to a HW bug in some devices, modifying TTL on RX flows will
+	 * cause an incorrect checksum calculation. In this case we will
+	 * use a FW table to recalculate.
 	 */
 	if (dmn->type == MLX5DR_DOMAIN_TYPE_FDB &&
 	    rx_rule && recalc_cs_required && dest_action) {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c
index 9cd5c50c5d42..f49abc7a4b9b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c
@@ -18,6 +18,11 @@ static u32 dr_ste_crc32_calc(const void *input_data, size_t length)
 	return (__force u32)htonl(crc);
 }
 
+bool mlx5dr_ste_supp_ttl_cs_recalc(struct mlx5dr_cmd_caps *caps)
+{
+	return caps->sw_format_ver > MLX5_STEERING_FORMAT_CONNECTX_5;
+}
+
 u32 mlx5dr_ste_calc_hash_index(u8 *hw_ste_p, struct mlx5dr_ste_htbl *htbl)
 {
 	struct dr_hw_ste_format *hw_ste = (struct dr_hw_ste_format *)hw_ste_p;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h
index a8b497cbb844..4af0e4e6a13c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h
@@ -1211,6 +1211,8 @@ int mlx5dr_cmd_set_fte(struct mlx5_core_dev *dev,
 		       u32 group_id,
 		       struct mlx5dr_cmd_fte_info *fte);
 
+bool mlx5dr_ste_supp_ttl_cs_recalc(struct mlx5dr_cmd_caps *caps);
+
 struct mlx5dr_fw_recalc_cs_ft {
 	u64 rx_icm_addr;
 	u32 table_id;
-- 
2.29.2