[PATCH v5 net-next 4/9] net/sched: mqprio: add an extack message to mqprio_parse_opt()

[Vladimir Oltean]

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@xxxxxxxxxxx/
Signed-off-by: Vladimir Oltean <vladimir.oltean@xxxxxxx>
Reviewed-by: Ferenc Fejes <fejes@xxxxxxxxxxx>
Reviewed-by: Simon Horman <simon.horman@xxxxxxxxxxxx>
Acked-by: Jamal Hadi Salim <jhs@xxxxxxxxxxxx>
---
v3->v5: 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 8e8151ca8307..aae4d64dbf3f 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