On 02/04/2019 11:27, Sebastian Gottschall wrote:
Am 01.04.2019 um 16:02 schrieb John Crispin:
On 01/04/2019 15:14, John Crispin wrote:
@@ -5185,6 +5214,23 @@ static int ath10k_add_interface(struct
ieee80211_hw *hw,
break;
}
+ switch (vif->type) {
+ case NL80211_IFTYPE_AP:
+ hw_encap = 1;
+ break;
+ case NL80211_IFTYPE_STATION:
+// if (!sdata->u.mgd.use_4addr)
+ hw_encap = 1;
+ break;
+ case NL80211_IFTYPE_AP_VLAN:
+// if (!sdata->wdev.use_4addr)
+ hw_encap = 1;
+ break;
+ default:
+ break;
+ }
+ ieee80211_set_hw_80211_encap(vif, ar->ethernetmode & hw_encap);
+
/* Using vdev_id as queue number will make it very easy to do
per-vif
* tx queue locking. This shouldn't wrap due to interface
combinations
* but do a modulo for correctness sake and prevent using
offchannel tx
Hi
I am not sure what the best way to handle the 4addr mode is. sdata is
part of the subsystem and is not exposed to the drivers.Calling
ieee80211_set_hw_80211_encap() from within the driver looked best to
me at first glance and would avoid adding an extra op call to figure
out if the if encap is supported. Any pointers on what the preferred
solution would look like are welcome.
John
you may do the 4addr check within ieee80211_set_hw_80211_encap, there
is no need for dealing with that in ath10k or do i miss something?
Sebastian
Correct, I'll add that in V4
John
