On 24 April 2014 10:50, Johannes Berg <johannes@xxxxxxxxxxxxxxxx> wrote: > On Thu, 2014-04-24 at 10:44 +0200, Michal Kazior wrote: > >> I think Monitor operation should be performed on a best effort basis. >> This means monitor_start/stop should be attempted once number of >> non-monitor vdevs changes. >> >> We should probably introduce a ath10k_recalc_monitor() for that purpose. > > Doesn't mac80211 do that for you? > > See IEEE80211_HW_WANT_MONITOR_VIF. This is not sufficient in this case. E.g. If you add a disconnected 4addr sta interface to bridge the interface enters promisc mode. This attempts to start monitor vdev in ath10k before sta vdev is started internally. We could probably make it start earlier (in add_interface) but there still remains a problem when you stop last non-monitor interface (monitor vif will be created _after_ last non-monitor is removed which is too late). Michał -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
