On Tue, 2011-08-09 at 13:45 +0200, Johannes Berg wrote: > On Tue, 2011-08-09 at 13:43 +0200, Johannes Berg wrote: > > > > Yes, but __ieee80211_resume() calls directly ieee80211_reconfig(). > > > > Sure. Just trying to make sure I understand where it's coming from. So > > wiphy_unregister() is still running, before device_del(), but then we > > get the resume from the core... > > > > Shouldn't we implement this logic you just did in cfg80211 instead? > > Although I guess until wiphy_unregister() returns, drivers must still > expect callbacks, which clearly mac80211 didn't handle correctly here > with the timer... Actually... Doesn't that just mean we need to reorder the code in _free_hw()? If we do the timer stuff after wiphy_unregister(), we should be OK, and prevent other, similar problems in the future too. johannes -- 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
