[net-next 12/14] net/mlx5: E-switch, mark devcom as not ready when all eswitches are unpaired

[Saeed Mahameed <saeed@kernel.org>]

From: Shay Drory <shayd@nvidia.com>

Whenever an eswitch is unpaired with another, the driver mark devcom
as not ready. While this is correct in case we are pairing only two
eswitches, in order to support pairing of more than two eswitches,
driver need to mark devcom as not ready only when all eswitches are
unpaired.

Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/eswitch.h         | 1 +
 .../net/ethernet/mellanox/mlx5/core/eswitch_offloads.c    | 8 +++++++-
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index 9833d1a587cc..d6e4ca436f39 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -343,6 +343,7 @@ struct mlx5_eswitch {
 	int                     mode;
 	u16                     manager_vport;
 	u16                     first_host_vport;
+	u8			num_peers;
 	struct mlx5_esw_functions esw_funcs;
 	struct {
 		u32             large_group_num;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index aeb15b10048e..09367a320741 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -2836,6 +2836,8 @@ static int mlx5_esw_offloads_devcom_event(int event,
 
 		esw->paired[mlx5_get_dev_index(peer_esw->dev)] = true;
 		peer_esw->paired[mlx5_get_dev_index(esw->dev)] = true;
+		esw->num_peers++;
+		peer_esw->num_peers++;
 		mlx5_devcom_comp_set_ready(devcom, MLX5_DEVCOM_ESW_OFFLOADS, true);
 		break;
 
@@ -2843,7 +2845,10 @@ static int mlx5_esw_offloads_devcom_event(int event,
 		if (!esw->paired[mlx5_get_dev_index(peer_esw->dev)])
 			break;
 
-		mlx5_devcom_comp_set_ready(devcom, MLX5_DEVCOM_ESW_OFFLOADS, false);
+		peer_esw->num_peers--;
+		esw->num_peers--;
+		if (!esw->num_peers && !peer_esw->num_peers)
+			mlx5_devcom_comp_set_ready(devcom, MLX5_DEVCOM_ESW_OFFLOADS, false);
 		esw->paired[mlx5_get_dev_index(peer_esw->dev)] = false;
 		peer_esw->paired[mlx5_get_dev_index(esw->dev)] = false;
 		mlx5_esw_offloads_unpair(peer_esw, esw);
@@ -2884,6 +2889,7 @@ void mlx5_esw_offloads_devcom_init(struct mlx5_eswitch *esw)
 				       mlx5_esw_offloads_devcom_event,
 				       esw);
 
+	esw->num_peers = 0;
 	mlx5_devcom_send_event(devcom,
 			       MLX5_DEVCOM_ESW_OFFLOADS,
 			       ESW_OFFLOADS_DEVCOM_PAIR, esw);
-- 
2.40.1