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/[email protected]/
Signed-off-by: Vladimir Oltean <[email protected]>
Reviewed-by: Ferenc Fejes <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Acked-by: Jamal Hadi Salim <[email protected]>
---
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