On Tue, 2019-10-08 at 13:50 -0500, Denis Kenzior wrote:
> Hi Johannes,
>
> > > But they shouldn't change due a mac address change? I wonder if we can
> > > further relax the requirements to allow mac change if
> > > NL80211_SCAN_FLAG_RANDOM_ADDR was used?
> >
> > No, at least with HW scan that would completely confuse the driver -
> > since from the driver's POV we'd remove the interface it's currently
> > managing the scan for.
>
> So help me understand this better.
include/net/mac80211.h:
int (*hw_scan)(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
struct ieee80211_scan_request *req);
How is it difficult to understand that with an API like that, it might
not be a good idea to remove the vif from the driver while it's
scanning?
> Just by virtue of copying the new
> mac into sdata->vif.addr
That's not what happens?
johannes
