Skip to contentSkip to conversation

Patch "net/mlx5: E-Switch, Correctly deal with inline mode on ConnectX-5" has been added to the 4.10-stable tree

PATCH [email protected] 1 participant 1 message Apr 29, 2017

Subsystems (4)
lore
Share & tools
lore

[email protected]

author maintainer Applied
#0 lore raw

This is a note to let you know that I've just added the patch titled

net/mlx5: E-Switch, Correctly deal with inline mode on ConnectX-5

to the 4.10-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
net-mlx5-e-switch-correctly-deal-with-inline-mode-on-connectx-5.patch
and it can be found in the queue-4.10 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.

Quoted 1 line

From foo@baz Sat Apr 29 08:22:40 CEST 2017

From: Or Gerlitz <[email protected]>
Date: Thu, 30 Mar 2017 15:56:10 +0200
Subject: net/mlx5: E-Switch, Correctly deal with inline mode on ConnectX-5

From: Or Gerlitz <[email protected]>

[ Upstream commit c415f704c8276bc686abcb0497bf2606038ca73c ]

On ConnectX5 the wqe inline mode is "none" and hence the FW reports MLX5_CAP_INLINE_MODE_NOT_REQUIRED.

Fix our devlink callbacks to deal with that on get and set.

Also fix the tc flow parsing code not to fail anything when inline isn't required.

Fixes bffaa916588e ('net/mlx5: E-Switch, Add control for inline mode')
---
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c            |    3 -
 drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c |   36 ++++++++-----
 2 files changed, 26 insertions(+), 13 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +2 -1
@@ -611,7 +611,8 @@ static int parse_cls_flower(struct mlx5e
 
 	if (!err && esw->mode == SRIOV_OFFLOADS &&
 	    rep->vport != FDB_UPLINK_VPORT) {
-		if (min_inline > esw->offloads.inline_mode) {
+		if (esw->offloads.inline_mode != MLX5_INLINE_MODE_NONE &&
+		    esw->offloads.inline_mode < min_inline) {
 			netdev_warn(priv->netdev,
 				    "Flow is not offloaded due to min inline setting, required %d actual %d\n",
 				    min_inline, esw->offloads.inline_mode);
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +24 -12
@@ -908,8 +908,7 @@ int mlx5_devlink_eswitch_inline_mode_set
 	struct mlx5_core_dev *dev = devlink_priv(devlink);
 	struct mlx5_eswitch *esw = dev->priv.eswitch;
 	int num_vports = esw->enabled_vports;
-	int err;
-	int vport;
+	int err, vport;
 	u8 mlx5_mode;
 
 	if (!MLX5_CAP_GEN(dev, vport_group_manager))
@@ -918,9 +917,17 @@ int mlx5_devlink_eswitch_inline_mode_set
 	if (esw->mode == SRIOV_NONE)
 		return -EOPNOTSUPP;
 
-	if (MLX5_CAP_ETH(dev, wqe_inline_mode) !=
-	    MLX5_CAP_INLINE_MODE_VPORT_CONTEXT)
+	switch (MLX5_CAP_ETH(dev, wqe_inline_mode)) {
+	case MLX5_CAP_INLINE_MODE_NOT_REQUIRED:
+		if (mode == DEVLINK_ESWITCH_INLINE_MODE_NONE)
+			return 0;
+		/* fall through */
+	case MLX5_CAP_INLINE_MODE_L2:
+		esw_warn(dev, "Inline mode can't be set\n");
 		return -EOPNOTSUPP;
+	case MLX5_CAP_INLINE_MODE_VPORT_CONTEXT:
+		break;
+	}
 
 	if (esw->offloads.num_flows > 0) {
 		esw_warn(dev, "Can't set inline mode when flows are configured\n");
@@ -963,18 +970,14 @@ int mlx5_devlink_eswitch_inline_mode_get
 	if (esw->mode == SRIOV_NONE)
 		return -EOPNOTSUPP;
 
-	if (MLX5_CAP_ETH(dev, wqe_inline_mode) !=
-	    MLX5_CAP_INLINE_MODE_VPORT_CONTEXT)
-		return -EOPNOTSUPP;
-
 	return esw_inline_mode_to_devlink(esw->offloads.inline_mode, mode);
 }
 
 int mlx5_eswitch_inline_mode_get(struct mlx5_eswitch *esw, int nvfs, u8 *mode)
 {
+	u8 prev_mlx5_mode, mlx5_mode = MLX5_INLINE_MODE_L2;
 	struct mlx5_core_dev *dev = esw->dev;
 	int vport;
-	u8 prev_mlx5_mode, mlx5_mode = MLX5_INLINE_MODE_L2;
 
 	if (!MLX5_CAP_GEN(dev, vport_group_manager))
 		return -EOPNOTSUPP;
@@ -982,10 +985,18 @@ int mlx5_eswitch_inline_mode_get(struct
 	if (esw->mode == SRIOV_NONE)
 		return -EOPNOTSUPP;
 
-	if (MLX5_CAP_ETH(dev, wqe_inline_mode) !=
-	    MLX5_CAP_INLINE_MODE_VPORT_CONTEXT)
-		return -EOPNOTSUPP;
+	switch (MLX5_CAP_ETH(dev, wqe_inline_mode)) {
+	case MLX5_CAP_INLINE_MODE_NOT_REQUIRED:
+		mlx5_mode = MLX5_INLINE_MODE_NONE;
+		goto out;
+	case MLX5_CAP_INLINE_MODE_L2:
+		mlx5_mode = MLX5_INLINE_MODE_L2;
+		goto out;
+	case MLX5_CAP_INLINE_MODE_VPORT_CONTEXT:
+		goto query_vports;
+	}
 
+query_vports:
 	for (vport = 1; vport <= nvfs; vport++) {
 		mlx5_query_nic_vport_min_inline(dev, vport, &mlx5_mode);
 		if (vport > 1 && prev_mlx5_mode != mlx5_mode)
@@ -993,6 +1004,7 @@ int mlx5_eswitch_inline_mode_get(struct
 		prev_mlx5_mode = mlx5_mode;
 	}
 
+out:
 	*mode = mlx5_mode;
 	return 0;
 }

Patches currently in stable-queue which might be from [email protected] are

queue-4.10/net-mlx5-e-switch-correctly-deal-with-inline-mode-on-connectx-5.patch