[v4,net-next,4/9] net/sched: mqprio: add an extack message to mqprio_parse_opt()

Message ID 20230403103440.2895683-5-vladimir.oltean@nxp.com
State New
Headers
Series Add tc-mqprio and tc-taprio support for preemptible traffic classes |

Commit Message

Vladimir Oltean April 3, 2023, 10:34 a.m. UTC
  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.

The kernel was (and still is) rejecting this because iproute2 defaults
to "hw 1" if this command line option is not specified.

Link: https://lore.kernel.org/netdev/ede5e9a2f27bf83bfb86d3e8c4ca7b34093b99e2.camel@inf.elte.hu/
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ferenc Fejes <fejes@inf.elte.hu>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
---
v3->v4: none
v2->v3: change link from patchwork to lore
v1->v2: slightly reword last paragraph of commit message

 net/sched/sch_mqprio.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
  

Comments

Jamal Hadi Salim April 7, 2023, 4:08 p.m. UTC | #1
On Mon, Apr 3, 2023 at 6:35 AM Vladimir Oltean <vladimir.oltean@nxp.com> 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.
>
> The kernel was (and still is) rejecting this because iproute2 defaults
> to "hw 1" if this command line option is not specified.
>
> Link: https://lore.kernel.org/netdev/ede5e9a2f27bf83bfb86d3e8c4ca7b34093b99e2.camel@inf.elte.hu/
> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
> Reviewed-by: Ferenc Fejes <fejes@inf.elte.hu>
> Reviewed-by: Simon Horman <simon.horman@corigine.com>

Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>

cheers,
jamal
> ---
> v3->v4: none
> v2->v3: change link from patchwork to lore
> v1->v2: slightly reword last paragraph of commit message
>
>  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 5a9261c38b95..9ee5a9a9b9e9 100644
> --- a/net/sched/sch_mqprio.c
> +++ b/net/sched/sch_mqprio.c
> @@ -133,8 +133,11 @@ 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;
>  }
> --
> 2.34.1
>
  

Patch

diff --git a/net/sched/sch_mqprio.c b/net/sched/sch_mqprio.c
index 5a9261c38b95..9ee5a9a9b9e9 100644
--- a/net/sched/sch_mqprio.c
+++ b/net/sched/sch_mqprio.c
@@ -133,8 +133,11 @@  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;
 }