On Wed, Nov 13, 2013 at 08:13:44PM +0100, Johannes Berg wrote: > On Wed, 2013-11-13 at 19:57 +0100, Luis R. Rodriguez wrote: > > On Wed, Nov 13, 2013 at 7:48 PM, Johannes Berg > > <johannes@xxxxxxxxxxxxxxxx> wrote: > > > On Wed, 2013-11-13 at 19:12 +0100, Luis R. Rodriguez wrote: > > > > > >> Janusz, please test, and likely feel free to take on the last patch. > > >> We may need help from Johannes on the rtnl lock question now required > > >> for checks on enabling beacon operation. > > > > > > What's that rtnl thing about? > > > > Whether or not callers of cfg80211_reg_can_beacon() have it properly > > held already and if not best strategy to make this simple and concise. > > The rtnl is required as we need to now query the cfg80211 regdom dfs > > region for the case where a channel requires DFS. > > Can't you just use RCU for that part? If Janusz is right then we don't even need that patch at all :D The rtnl then would only be used for sending the wiphy, we'd just require hostapd to always check for the DFS region prior to starting IR for the first time, if the netdev was brought down (quisced), we'd have to have it check again the DFS region. This however does assume we can quiesce properly upon regulatory domain changes. Luis -- 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
