[PATCH net 0/2] mlx5 misc fixes 2025-03-18

[PATCH net 0/2] mlx5 misc fixes 2025-03-18

[Tariq Toukan]

Hi,

This small patchset provides misc bug fixes to the mlx5 core driver.

Thanks,
Tariq.


Mark Bloch (1):
  net/mlx5: LAG, reload representors on LAG creation failure

Moshe Shemesh (1):
  net/mlx5: Start health poll after enable hca

 drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c |  4 ++++
 drivers/net/ethernet/mellanox/mlx5/core/main.c    | 15 +++++++--------
 2 files changed, 11 insertions(+), 8 deletions(-)


base-commit: daa624d3c2ddffdcbad140a9625a4064371db44f
-- 
2.31.1

[PATCH net 1/2] net/mlx5: LAG, reload representors on LAG creation failure

[Tariq Toukan]

From: Mark Bloch <mbloch@nvidia.com>

When LAG creation fails, the driver reloads the RDMA devices. If RDMA
representors are present, they should also be reloaded. This step was
missed in the cited commit.

Fixes: 598fe77df855 ("net/mlx5: Lag, Create shared FDB when in switchdev mode")
Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Shay Drori <shayd@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c
index ed2ba272946b..6c9737c53734 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c
@@ -1052,6 +1052,10 @@ static void mlx5_do_bond(struct mlx5_lag *ldev)
 		if (err) {
 			if (shared_fdb || roce_lag)
 				mlx5_lag_add_devices(ldev);
+			if (shared_fdb) {
+				mlx5_ldev_for_each(i, 0, ldev)
+					mlx5_eswitch_reload_ib_reps(ldev->pf[i].dev->priv.eswitch);
+			}
 
 			return;
 		} else if (roce_lag) {
-- 
2.31.1

[PATCH net 2/2] net/mlx5: Start health poll after enable hca

[Tariq Toukan]

From: Moshe Shemesh <moshe@nvidia.com>

The health poll mechanism performs periodic checks to detect firmware
errors. One of the checks verifies the function is still enabled on
firmware side, but the function is enabled only after enable_hca command
completed. Start health poll after enable_hca command to avoid a race
between function enabled and first health polling.

Fixes: 9b98d395b85d ("net/mlx5: Start health poll at earlier stage of driver load")
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Shay Drori <shayd@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/main.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index ec956c4bcebd..7c3312d6aed9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -1205,24 +1205,24 @@ static int mlx5_function_enable(struct mlx5_core_dev *dev, bool boot, u64 timeou
 	dev->caps.embedded_cpu = mlx5_read_embedded_cpu(dev);
 	mlx5_cmd_set_state(dev, MLX5_CMDIF_STATE_UP);
 
-	mlx5_start_health_poll(dev);
-
 	err = mlx5_core_enable_hca(dev, 0);
 	if (err) {
 		mlx5_core_err(dev, "enable hca failed\n");
-		goto stop_health_poll;
+		goto err_cmd_cleanup;
 	}
 
+	mlx5_start_health_poll(dev);
+
 	err = mlx5_core_set_issi(dev);
 	if (err) {
 		mlx5_core_err(dev, "failed to set issi\n");
-		goto err_disable_hca;
+		goto stop_health_poll;
 	}
 
 	err = mlx5_satisfy_startup_pages(dev, 1);
 	if (err) {
 		mlx5_core_err(dev, "failed to allocate boot pages\n");
-		goto err_disable_hca;
+		goto stop_health_poll;
 	}
 
 	err = mlx5_tout_query_dtor(dev);
@@ -1235,10 +1235,9 @@ static int mlx5_function_enable(struct mlx5_core_dev *dev, bool boot, u64 timeou
 
 reclaim_boot_pages:
 	mlx5_reclaim_startup_pages(dev);
-err_disable_hca:
-	mlx5_core_disable_hca(dev, 0);
 stop_health_poll:
 	mlx5_stop_health_poll(dev, boot);
+	mlx5_core_disable_hca(dev, 0);
 err_cmd_cleanup:
 	mlx5_cmd_set_state(dev, MLX5_CMDIF_STATE_DOWN);
 	mlx5_cmd_disable(dev);
@@ -1249,8 +1248,8 @@ static int mlx5_function_enable(struct mlx5_core_dev *dev, bool boot, u64 timeou
 static void mlx5_function_disable(struct mlx5_core_dev *dev, bool boot)
 {
 	mlx5_reclaim_startup_pages(dev);
-	mlx5_core_disable_hca(dev, 0);
 	mlx5_stop_health_poll(dev, boot);
+	mlx5_core_disable_hca(dev, 0);
 	mlx5_cmd_set_state(dev, MLX5_CMDIF_STATE_DOWN);
 	mlx5_cmd_disable(dev);
 }
-- 
2.31.1

Re: [PATCH net 1/2] net/mlx5: LAG, reload representors on LAG creation failure

[Michal Swiatkowski]

On Tue, Mar 18, 2025 at 10:51:16PM +0200, Tariq Toukan wrote:
> From: Mark Bloch <mbloch@nvidia.com>
> 
> When LAG creation fails, the driver reloads the RDMA devices. If RDMA
> representors are present, they should also be reloaded. This step was
> missed in the cited commit.
> 
> Fixes: 598fe77df855 ("net/mlx5: Lag, Create shared FDB when in switchdev mode")
> Signed-off-by: Mark Bloch <mbloch@nvidia.com>
> Reviewed-by: Shay Drori <shayd@nvidia.com>
> Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
> ---
>  drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c
> index ed2ba272946b..6c9737c53734 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c
> @@ -1052,6 +1052,10 @@ static void mlx5_do_bond(struct mlx5_lag *ldev)
>  		if (err) {
>  			if (shared_fdb || roce_lag)
>  				mlx5_lag_add_devices(ldev);
> +			if (shared_fdb) {
> +				mlx5_ldev_for_each(i, 0, ldev)
> +					mlx5_eswitch_reload_ib_reps(ldev->pf[i].dev->priv.eswitch);
> +			}
>  
>  			return;
>  		} else if (roce_lag) {
> -- 

Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>

> 2.31.1

Re: [PATCH net 2/2] net/mlx5: Start health poll after enable hca

[Michal Swiatkowski]

On Tue, Mar 18, 2025 at 10:51:17PM +0200, Tariq Toukan wrote:
> From: Moshe Shemesh <moshe@nvidia.com>
> 
> The health poll mechanism performs periodic checks to detect firmware
> errors. One of the checks verifies the function is still enabled on
> firmware side, but the function is enabled only after enable_hca command
> completed. Start health poll after enable_hca command to avoid a race
> between function enabled and first health polling.
> 
> Fixes: 9b98d395b85d ("net/mlx5: Start health poll at earlier stage of driver load")
> Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
> Reviewed-by: Shay Drori <shayd@nvidia.com>
> Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
> ---
>  drivers/net/ethernet/mellanox/mlx5/core/main.c | 15 +++++++--------
>  1 file changed, 7 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
> index ec956c4bcebd..7c3312d6aed9 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
> @@ -1205,24 +1205,24 @@ static int mlx5_function_enable(struct mlx5_core_dev *dev, bool boot, u64 timeou
>  	dev->caps.embedded_cpu = mlx5_read_embedded_cpu(dev);
>  	mlx5_cmd_set_state(dev, MLX5_CMDIF_STATE_UP);
>  
> -	mlx5_start_health_poll(dev);
> -
>  	err = mlx5_core_enable_hca(dev, 0);
>  	if (err) {
>  		mlx5_core_err(dev, "enable hca failed\n");
> -		goto stop_health_poll;
> +		goto err_cmd_cleanup;
>  	}
>  
> +	mlx5_start_health_poll(dev);
> +
>  	err = mlx5_core_set_issi(dev);
>  	if (err) {
>  		mlx5_core_err(dev, "failed to set issi\n");
> -		goto err_disable_hca;
> +		goto stop_health_poll;
>  	}
>  
>  	err = mlx5_satisfy_startup_pages(dev, 1);
>  	if (err) {
>  		mlx5_core_err(dev, "failed to allocate boot pages\n");
> -		goto err_disable_hca;
> +		goto stop_health_poll;
>  	}
>  
>  	err = mlx5_tout_query_dtor(dev);
> @@ -1235,10 +1235,9 @@ static int mlx5_function_enable(struct mlx5_core_dev *dev, bool boot, u64 timeou
>  
>  reclaim_boot_pages:
>  	mlx5_reclaim_startup_pages(dev);
> -err_disable_hca:
> -	mlx5_core_disable_hca(dev, 0);
>  stop_health_poll:
>  	mlx5_stop_health_poll(dev, boot);
> +	mlx5_core_disable_hca(dev, 0);
>  err_cmd_cleanup:
>  	mlx5_cmd_set_state(dev, MLX5_CMDIF_STATE_DOWN);
>  	mlx5_cmd_disable(dev);
> @@ -1249,8 +1248,8 @@ static int mlx5_function_enable(struct mlx5_core_dev *dev, bool boot, u64 timeou
>  static void mlx5_function_disable(struct mlx5_core_dev *dev, bool boot)
>  {
>  	mlx5_reclaim_startup_pages(dev);
> -	mlx5_core_disable_hca(dev, 0);
>  	mlx5_stop_health_poll(dev, boot);
> +	mlx5_core_disable_hca(dev, 0);
>  	mlx5_cmd_set_state(dev, MLX5_CMDIF_STATE_DOWN);
>  	mlx5_cmd_disable(dev);
>  }

Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>

> -- 
> 2.31.1

Re: [PATCH net 2/2] net/mlx5: Start health poll after enable hca

[Kalesh Anakkur Purayil]

[-- Attachment #1: Type: text/plain, Size: 798 bytes --]

On Wed, Mar 19, 2025 at 2:22 AM Tariq Toukan <tariqt@nvidia.com> wrote:
>
> From: Moshe Shemesh <moshe@nvidia.com>
>
> The health poll mechanism performs periodic checks to detect firmware
> errors. One of the checks verifies the function is still enabled on
> firmware side, but the function is enabled only after enable_hca command
> completed. Start health poll after enable_hca command to avoid a race
> between function enabled and first health polling.
>
> Fixes: 9b98d395b85d ("net/mlx5: Start health poll at earlier stage of driver load")
> Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
> Reviewed-by: Shay Drori <shayd@nvidia.com>
> Signed-off-by: Tariq Toukan <tariqt@nvidia.com>

Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
-- 
Regards,
Kalesh AP

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4226 bytes --]

Re: [PATCH net 1/2] net/mlx5: LAG, reload representors on LAG creation failure

[Kalesh Anakkur Purayil]

[-- Attachment #1: Type: text/plain, Size: 624 bytes --]

On Wed, Mar 19, 2025 at 2:22 AM Tariq Toukan <tariqt@nvidia.com> wrote:
>
> From: Mark Bloch <mbloch@nvidia.com>
>
> When LAG creation fails, the driver reloads the RDMA devices. If RDMA
> representors are present, they should also be reloaded. This step was
> missed in the cited commit.
>
> Fixes: 598fe77df855 ("net/mlx5: Lag, Create shared FDB when in switchdev mode")
> Signed-off-by: Mark Bloch <mbloch@nvidia.com>
> Reviewed-by: Shay Drori <shayd@nvidia.com>
> Signed-off-by: Tariq Toukan <tariqt@nvidia.com>

Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>


-- 
Regards,
Kalesh AP

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4226 bytes --]

Re: [PATCH net 0/2] mlx5 misc fixes 2025-03-18

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Tue, 18 Mar 2025 22:51:15 +0200 you wrote:
> Hi,
> 
> This small patchset provides misc bug fixes to the mlx5 core driver.
> 
> Thanks,
> Tariq.
> 
> [...]

Here is the summary with links:
  - [net,1/2] net/mlx5: LAG, reload representors on LAG creation failure
    https://git.kernel.org/netdev/net/c/bdf549a7a4d7
  - [net,2/2] net/mlx5: Start health poll after enable hca
    https://git.kernel.org/netdev/net/c/1726ad035cb0

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