netdev/ieee802154_submac: add retransmission reporting

[benpicco]

Contribution description

If we do software retransmissions, we already keep a count.
Allow to query it with NETOPT_TX_RETRIES_NEEDED.

Testing procedure

Issues/PRs references

#14448

[jia200x]

Ah yes, indeed this works. But maybe we should handle the case where there are HW retrans but no tx_info in the SubMAC. That way, the SubMAC always gives this information

[jia200x]

alternatively, we can return an error when NETOPT_TX_RETRIES_NEEDED is called in such case

[benpicco]

Ah I wasn't sure if info could be NULL.

[jia200x]

Since you are there, could you add this patch?

diff --git a/sys/net/link_layer/ieee802154/submac.c b/sys/net/link_layer/ieee802154/submac.c
index 0584ba95a3..392714b995 100644
--- a/sys/net/link_layer/ieee802154/submac.c
+++ b/sys/net/link_layer/ieee802154/submac.c
@@ -194,6 +194,9 @@ static void _handle_tx_success(ieee802154_submac_t *submac,
 
     if (ieee802154_radio_has_frame_retrans(dev) ||
         ieee802154_radio_has_irq_ack_timeout(dev) || !submac->wait_for_ack) {
+        if (!ieee802154_radio_has_frame_retrans_info()) {
+            info->retrans = -1;
+        }
         _tx_end(submac, info->status, info);
     }
     else {

And change these lines to:

Suggested change

	if (info) {
	netdev_submac->retrans = info->retrans;
	} else {
	netdev_submac->retrans = -1;
	}
	if (info) {
	netdev_submac->retrans = info->retrans;
	}

?
That way the SubMAC will always give its best effort to report the retransmissions. And if there's no info (TX_MEDIUM_BUSY or TX_NO_ACK), then netdev_submac->retrans doesn't get updated.

[benpicco]

I think turning the _get_cap() function into a bitfield would be a good optimization.

[jia200x]

yes, we could!

[jia200x]

Looks good to me. Please squash.
Also, could you add its own commit for the one that touches the radio HAL API?

[jia200x]

ACK