Re: [PATCH net-next 08/12] net/sched: mqprio: add an extack message to mqprio_parse_opt()

[Ferenc Fejes]

Hi Vladimir!

On Fri, 2023-02-17 at 01:21 +0200, Vladimir Oltean wrote:
> Ferenc reports that a combination of poor iproute2 defaults and
> obscure
> cases where the kernel returns -EINVAL make it difficult to
> understand
> what is wrong with this command:
> 
> $ ip link add veth0 numtxqueues 8 numrxqueues 8 type veth peer name
> veth1
> $ tc qdisc add dev veth0 root mqprio num_tc 8 map 0 1 2 3 4 5 6 7 \
>         queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7
> RTNETLINK answers: Invalid argument
> 
> Hopefully with this patch, the cause is clearer:
> 
> Error: Device does not support hardware offload.

Much better, great improvement!

> 
> This was rejected because iproute2 defaults to "hw 1" if the option
> is
> not specified.
> 
> Link:
> https://patchwork.kernel.org/project/netdevbpf/patch/20230204135307.1036988-3-vladimir.oltean@xxxxxxx/#25215636
> Signed-off-by: Vladimir Oltean <vladimir.oltean@xxxxxxx>
> ---
>  net/sched/sch_mqprio.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/net/sched/sch_mqprio.c b/net/sched/sch_mqprio.c
> index 18eda5fade81..52cfc0ec2e23 100644
> --- a/net/sched/sch_mqprio.c
> +++ b/net/sched/sch_mqprio.c
> @@ -134,8 +134,11 @@ static int mqprio_parse_opt(struct net_device
> *dev, struct tc_mqprio_qopt *qopt,
>         /* If ndo_setup_tc is not present then hardware doesn't
> support offload
>          * and we should return an error.
>          */
> -       if (qopt->hw && !dev->netdev_ops->ndo_setup_tc)
> +       if (qopt->hw && !dev->netdev_ops->ndo_setup_tc) {
> +               NL_SET_ERR_MSG(extack,
> +                              "Device does not support hardware
> offload");
>                 return -EINVAL;
> +       }
>  
>         return 0;
>  }

Thanks for doing this!

Best,
Ferenc