On Mon, Nov 14, 2016 at 12:18:07PM +0100, Rafał Miłecki wrote: > Thanks for explaining. It makes sense I guess, we should also simplify > initialization path. There is too much guessing on what method should > be used. Dropping monitor-based AP support would help with that a lot. Agreed.. There has been number of small changes to enable yet another combinations of driver capabilities to avoid accidentally hitting the monitor-based old design. Getting rid of this code would also remove quite a bit of code: src/drivers/driver_nl80211.c | 127 ++------- src/drivers/driver_nl80211.h | 11 - src/drivers/driver_nl80211_capa.c | 12 +- src/drivers/driver_nl80211_monitor.c | 491 ----------------------------------- src/drivers/drivers.mak | 1 - 5 files changed, 18 insertions(+), 624 deletions(-) That said, it is still possible that someone is using this interface, so I'm not completely sure when it would be fine to remove this from hostapd/wpa_supplicant.. wpa_supplicant side might be safer thing to drop since P2P/AP mode support was likely added at around the same time as monitor interface design was replaced. For hostapd, it might be safer to make the support build configurable and disable that configuration parameter for at least one hostapd release by default to see if anyone notices.. -- Jouni Malinen PGP id EFC895FA _______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
