[v2] net/bonding: fix error in bonding mode 4 with dedicated queues enabled

  when dedicated queues are enable with bonding mode 4 (mlx5), the
application sets the flow, which cannot be set if the device is not
started. This fixed the issue by starting the device just before
setting the flow. Because device should be started to set the flow.
Also it does not effect other driver codes (I have tried on ixgbe).

Bugzilla ID: 759

Signed-off-by: Usman Tanveer <usman.tanveer@emumba.com>
---
 drivers/net/bonding/rte_eth_bond_pmd.c | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)
  

We cannot configure the queues when the device is started. So, that
part is still before the device starts. I just moved
bond_ethdev_8023ad_flow_set() after the device start.


On Thu, Sep 29, 2022 at 9:05 AM Usman Tanveer <usman.tanveer@emumba.com> wrote:
>
> when dedicated queues are enable with bonding mode 4 (mlx5), the
> application sets the flow, which cannot be set if the device is not
> started. This fixed the issue by starting the device just before
> setting the flow. Because device should be started to set the flow.
> Also it does not effect other driver codes (I have tried on ixgbe).
>
> Bugzilla ID: 759
>
> Signed-off-by: Usman Tanveer <usman.tanveer@emumba.com>
> ---
>  drivers/net/bonding/rte_eth_bond_pmd.c | 19 +++++++++++--------
>  1 file changed, 11 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c
> index fd2d95a751..69cbbe19ff 100644
> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
> @@ -1828,7 +1828,18 @@ slave_start(struct rte_eth_dev *bonded_eth_dev,
>                         RTE_BOND_LOG(ERR, "bond_ethdev_8023ad_flow_destroy: port=%d, err (%d)",
>                                 slave_eth_dev->data->port_id, errval);
>                 }
> +       }
> +
> +       /* Start device */
> +       errval = rte_eth_dev_start(slave_eth_dev->data->port_id);
> +       if (errval != 0) {
> +               RTE_BOND_LOG(ERR, "rte_eth_dev_start: port=%u, err (%d)",
> +                               slave_eth_dev->data->port_id, errval);
> +               return -1;
> +       }
>
> +       if (internals->mode == BONDING_MODE_8023AD &&
> +                       internals->mode4.dedicated_queues.enabled == 1) {
>                 errval = bond_ethdev_8023ad_flow_set(bonded_eth_dev,
>                                 slave_eth_dev->data->port_id);
>                 if (errval != 0) {
> @@ -1839,14 +1850,6 @@ slave_start(struct rte_eth_dev *bonded_eth_dev,
>                 }
>         }
>
> -       /* Start device */
> -       errval = rte_eth_dev_start(slave_eth_dev->data->port_id);
> -       if (errval != 0) {
> -               RTE_BOND_LOG(ERR, "rte_eth_dev_start: port=%u, err (%d)",
> -                               slave_eth_dev->data->port_id, errval);
> -               return -1;
> -       }
> -
>         /* If RSS is enabled for bonding, synchronize RETA */
>         if (bonded_eth_dev->data->dev_conf.rxmode.mq_mode & RTE_ETH_MQ_RX_RSS) {
>                 int i;
> --
> 2.25.1
>
  

Thanks for making the change, IMHO this is much clearer.

Signed-off-by: Chas Williams <3chas3@gmail.com>

On 9/29/22 00:05, Usman Tanveer wrote:
> when dedicated queues are enable with bonding mode 4 (mlx5), the
> application sets the flow, which cannot be set if the device is not
> started. This fixed the issue by starting the device just before
> setting the flow. Because device should be started to set the flow.
> Also it does not effect other driver codes (I have tried on ixgbe).
> 
> Bugzilla ID: 759
> 
> Signed-off-by: Usman Tanveer <usman.tanveer@emumba.com>
> ---
>   drivers/net/bonding/rte_eth_bond_pmd.c | 19 +++++++++++--------
>   1 file changed, 11 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c
> index fd2d95a751..69cbbe19ff 100644
> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
> @@ -1828,7 +1828,18 @@ slave_start(struct rte_eth_dev *bonded_eth_dev,
>   			RTE_BOND_LOG(ERR, "bond_ethdev_8023ad_flow_destroy: port=%d, err (%d)",
>   				slave_eth_dev->data->port_id, errval);
>   		}
> +	}
> +
> +	/* Start device */
> +	errval = rte_eth_dev_start(slave_eth_dev->data->port_id);
> +	if (errval != 0) {
> +		RTE_BOND_LOG(ERR, "rte_eth_dev_start: port=%u, err (%d)",
> +				slave_eth_dev->data->port_id, errval);
> +		return -1;
> +	}
>   
> +	if (internals->mode == BONDING_MODE_8023AD &&
> +			internals->mode4.dedicated_queues.enabled == 1) {
>   		errval = bond_ethdev_8023ad_flow_set(bonded_eth_dev,
>   				slave_eth_dev->data->port_id);
>   		if (errval != 0) {
> @@ -1839,14 +1850,6 @@ slave_start(struct rte_eth_dev *bonded_eth_dev,
>   		}
>   	}
>   
> -	/* Start device */
> -	errval = rte_eth_dev_start(slave_eth_dev->data->port_id);
> -	if (errval != 0) {
> -		RTE_BOND_LOG(ERR, "rte_eth_dev_start: port=%u, err (%d)",
> -				slave_eth_dev->data->port_id, errval);
> -		return -1;
> -	}
> -
>   	/* If RSS is enabled for bonding, synchronize RETA */
>   	if (bonded_eth_dev->data->dev_conf.rxmode.mq_mode & RTE_ETH_MQ_RX_RSS) {
>   		int i;
  

On 10/15/22 18:28, Chas Williams wrote:
> Thanks for making the change, IMHO this is much clearer.
> 
> Signed-off-by: Chas Williams <3chas3@gmail.com>

Taking above sentence into account I treat it as:
Acked-by: Chas Williams <3chas3@gmail.com>

> 
> On 9/29/22 00:05, Usman Tanveer wrote:
>> when dedicated queues are enable with bonding mode 4 (mlx5), the
>> application sets the flow, which cannot be set if the device is not
>> started. This fixed the issue by starting the device just before
>> setting the flow. Because device should be started to set the flow.
>> Also it does not effect other driver codes (I have tried on ixgbe).
>>
>> Bugzilla ID: 759
>>
>> Signed-off-by: Usman Tanveer <usman.tanveer@emumba.com>

Applied to dpdk-next-net/main, thanks.