[net,1/1] igc: Fix double reset adapter triggered from a single taprio cmd

Following the implementation of "igc: Add TransmissionOverrun counter"
patch, when a taprio command is triggered by user, igc processes two
commands: TAPRIO_CMD_REPLACE followed by TAPRIO_CMD_STATS. However, both
commands unconditionally pass through igc_tsn_offload_apply() which
evaluates and triggers reset adapter. The double reset causes issues in
the calculation of adapter->qbv_count in igc.

TAPRIO_CMD_REPLACE command is expected to reset the adapter since it
activates qbv. It's unexpected for TAPRIO_CMD_STATS to do the same
because it doesn't configure any driver-specific TSN settings. So, the
evaluation in igc_tsn_offload_apply() isn't needed for TAPRIO_CMD_STATS.

To address this, commands parsing are relocated to
igc_tsn_enable_qbv_scheduling(). Commands that don't require an adapter
reset will exit after processing, thus avoiding igc_tsn_offload_apply().

Fixes: d3750076d464 ("igc: Add TransmissionOverrun counter")
Signed-off-by: Faizal Rahim <faizal.abdul.rahim@linux.intel.com>
---
 drivers/net/ethernet/intel/igc/igc_main.c | 33 ++++++++++++-----------
 1 file changed, 17 insertions(+), 16 deletions(-)

--
2.25.1

Added Vladimir and Husaini in CC.

On 25/6/2024 4:26 pm, Faizal Rahim wrote:
> Following the implementation of "igc: Add TransmissionOverrun counter"
> patch, when a taprio command is triggered by user, igc processes two
> commands: TAPRIO_CMD_REPLACE followed by TAPRIO_CMD_STATS. However, both
> commands unconditionally pass through igc_tsn_offload_apply() which
> evaluates and triggers reset adapter. The double reset causes issues in
> the calculation of adapter->qbv_count in igc.
> 
> TAPRIO_CMD_REPLACE command is expected to reset the adapter since it
> activates qbv. It's unexpected for TAPRIO_CMD_STATS to do the same
> because it doesn't configure any driver-specific TSN settings. So, the
> evaluation in igc_tsn_offload_apply() isn't needed for TAPRIO_CMD_STATS.
> 
> To address this, commands parsing are relocated to
> igc_tsn_enable_qbv_scheduling(). Commands that don't require an adapter
> reset will exit after processing, thus avoiding igc_tsn_offload_apply().
> 
> Fixes: d3750076d464 ("igc: Add TransmissionOverrun counter")
> Signed-off-by: Faizal Rahim <faizal.abdul.rahim@linux.intel.com>
> ---
>   drivers/net/ethernet/intel/igc/igc_main.c | 33 ++++++++++++-----------
>   1 file changed, 17 insertions(+), 16 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c
> index 87b655b839c1..33069880c86c 100644
> --- a/drivers/net/ethernet/intel/igc/igc_main.c
> +++ b/drivers/net/ethernet/intel/igc/igc_main.c
> @@ -6310,21 +6310,6 @@ static int igc_save_qbv_schedule(struct igc_adapter *adapter,
>   	size_t n;
>   	int i;
> 
> -	switch (qopt->cmd) {
> -	case TAPRIO_CMD_REPLACE:
> -		break;
> -	case TAPRIO_CMD_DESTROY:
> -		return igc_tsn_clear_schedule(adapter);
> -	case TAPRIO_CMD_STATS:
> -		igc_taprio_stats(adapter->netdev, &qopt->stats);
> -		return 0;
> -	case TAPRIO_CMD_QUEUE_STATS:
> -		igc_taprio_queue_stats(adapter->netdev, &qopt->queue_stats);
> -		return 0;
> -	default:
> -		return -EOPNOTSUPP;
> -	}
> -
>   	if (qopt->base_time < 0)
>   		return -ERANGE;
> 
> @@ -6433,7 +6418,23 @@ static int igc_tsn_enable_qbv_scheduling(struct igc_adapter *adapter,
>   	if (hw->mac.type != igc_i225)
>   		return -EOPNOTSUPP;
> 
> -	err = igc_save_qbv_schedule(adapter, qopt);
> +	switch (qopt->cmd) {
> +	case TAPRIO_CMD_REPLACE:
> +		err = igc_save_qbv_schedule(adapter, qopt);
> +		break;
> +	case TAPRIO_CMD_DESTROY:
> +		err = igc_tsn_clear_schedule(adapter);
> +		break;
> +	case TAPRIO_CMD_STATS:
> +		igc_taprio_stats(adapter->netdev, &qopt->stats);
> +		return 0;
> +	case TAPRIO_CMD_QUEUE_STATS:
> +		igc_taprio_queue_stats(adapter->netdev, &qopt->queue_stats);
> +		return 0;
> +	default:
> +		return -EOPNOTSUPP;
> +	}
> +
>   	if (err)
>   		return err;
> 
> --
> 2.25.1
>

Faizal Rahim <faizal.abdul.rahim@linux.intel.com> writes:

> Following the implementation of "igc: Add TransmissionOverrun counter"
> patch, when a taprio command is triggered by user, igc processes two
> commands: TAPRIO_CMD_REPLACE followed by TAPRIO_CMD_STATS. However, both
> commands unconditionally pass through igc_tsn_offload_apply() which
> evaluates and triggers reset adapter. The double reset causes issues in
> the calculation of adapter->qbv_count in igc.
>
> TAPRIO_CMD_REPLACE command is expected to reset the adapter since it
> activates qbv. It's unexpected for TAPRIO_CMD_STATS to do the same
> because it doesn't configure any driver-specific TSN settings. So, the
> evaluation in igc_tsn_offload_apply() isn't needed for TAPRIO_CMD_STATS.
>
> To address this, commands parsing are relocated to
> igc_tsn_enable_qbv_scheduling(). Commands that don't require an adapter
> reset will exit after processing, thus avoiding igc_tsn_offload_apply().
>
> Fixes: d3750076d464 ("igc: Add TransmissionOverrun counter")
> Signed-off-by: Faizal Rahim <faizal.abdul.rahim@linux.intel.com>
> ---
>  drivers/net/ethernet/intel/igc/igc_main.c | 33 ++++++++++++-----------
>  1 file changed, 17 insertions(+), 16 deletions(-)
>
> diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c
> index 87b655b839c1..33069880c86c 100644
> --- a/drivers/net/ethernet/intel/igc/igc_main.c
> +++ b/drivers/net/ethernet/intel/igc/igc_main.c
> @@ -6310,21 +6310,6 @@ static int igc_save_qbv_schedule(struct igc_adapter *adapter,
>  	size_t n;
>  	int i;
>
> -	switch (qopt->cmd) {
> -	case TAPRIO_CMD_REPLACE:
> -		break;
> -	case TAPRIO_CMD_DESTROY:
> -		return igc_tsn_clear_schedule(adapter);
> -	case TAPRIO_CMD_STATS:
> -		igc_taprio_stats(adapter->netdev, &qopt->stats);
> -		return 0;
> -	case TAPRIO_CMD_QUEUE_STATS:
> -		igc_taprio_queue_stats(adapter->netdev, &qopt->queue_stats);
> -		return 0;
> -	default:
> -		return -EOPNOTSUPP;
> -	}
> -
>  	if (qopt->base_time < 0)
>  		return -ERANGE;
>
> @@ -6433,7 +6418,23 @@ static int igc_tsn_enable_qbv_scheduling(struct igc_adapter *adapter,
>  	if (hw->mac.type != igc_i225)
>  		return -EOPNOTSUPP;
>
> -	err = igc_save_qbv_schedule(adapter, qopt);
> +	switch (qopt->cmd) {
> +	case TAPRIO_CMD_REPLACE:
> +		err = igc_save_qbv_schedule(adapter, qopt);
> +		break;
> +	case TAPRIO_CMD_DESTROY:
> +		err = igc_tsn_clear_schedule(adapter);
> +		break;
> +	case TAPRIO_CMD_STATS:
> +		igc_taprio_stats(adapter->netdev, &qopt->stats);
> +		return 0;
> +	case TAPRIO_CMD_QUEUE_STATS:
> +		igc_taprio_queue_stats(adapter->netdev, &qopt->queue_stats);
> +		return 0;
> +	default:
> +		return -EOPNOTSUPP;
> +	}
> +

Yeah, moving the command parsing to be done earlier sounds like the
right fix:

Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>


>  	if (err)
>  		return err;
>
> --
> 2.25.1
>

On Wed, Jun 26, 2024 at 09:19:12AM +0800, Abdul Rahim, Faizal wrote:
> Added Vladimir and Husaini in CC.
> 
> On 25/6/2024 4:26 pm, Faizal Rahim wrote:
> > Following the implementation of "igc: Add TransmissionOverrun counter"
> > patch, when a taprio command is triggered by user, igc processes two
> > commands: TAPRIO_CMD_REPLACE followed by TAPRIO_CMD_STATS. However, both
> > commands unconditionally pass through igc_tsn_offload_apply() which
> > evaluates and triggers reset adapter. The double reset causes issues in
> > the calculation of adapter->qbv_count in igc.
> > 
> > TAPRIO_CMD_REPLACE command is expected to reset the adapter since it
> > activates qbv. It's unexpected for TAPRIO_CMD_STATS to do the same
> > because it doesn't configure any driver-specific TSN settings. So, the
> > evaluation in igc_tsn_offload_apply() isn't needed for TAPRIO_CMD_STATS.
> > 
> > To address this, commands parsing are relocated to
> > igc_tsn_enable_qbv_scheduling(). Commands that don't require an adapter
> > reset will exit after processing, thus avoiding igc_tsn_offload_apply().
> > 
> > Fixes: d3750076d464 ("igc: Add TransmissionOverrun counter")
> > Signed-off-by: Faizal Rahim <faizal.abdul.rahim@linux.intel.com>
> > ---

Thank you for the patch. The code organization is much more logical this way.

Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>

On 25/06/2024 11:26, Faizal Rahim wrote:
> Following the implementation of "igc: Add TransmissionOverrun counter"
> patch, when a taprio command is triggered by user, igc processes two
> commands: TAPRIO_CMD_REPLACE followed by TAPRIO_CMD_STATS. However, both
> commands unconditionally pass through igc_tsn_offload_apply() which
> evaluates and triggers reset adapter. The double reset causes issues in
> the calculation of adapter->qbv_count in igc.
> 
> TAPRIO_CMD_REPLACE command is expected to reset the adapter since it
> activates qbv. It's unexpected for TAPRIO_CMD_STATS to do the same
> because it doesn't configure any driver-specific TSN settings. So, the
> evaluation in igc_tsn_offload_apply() isn't needed for TAPRIO_CMD_STATS.
> 
> To address this, commands parsing are relocated to
> igc_tsn_enable_qbv_scheduling(). Commands that don't require an adapter
> reset will exit after processing, thus avoiding igc_tsn_offload_apply().
> 
> Fixes: d3750076d464 ("igc: Add TransmissionOverrun counter")
> Signed-off-by: Faizal Rahim <faizal.abdul.rahim@linux.intel.com>
> Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
> Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
> ---
>   drivers/net/ethernet/intel/igc/igc_main.c | 33 ++++++++++++-----------
>   1 file changed, 17 insertions(+), 16 deletions(-)
> 
Tested-by: Mor Bar-Gabay <morx.bar.gabay@intel.com>