On Wed, 2010-09-22 at 17:23 +0200, Stanislaw Gruszka wrote: > > > Yes, bg_restart() force scan abort, but not if no scan is pending. It looks > > > like that: > > > > > > cpu0 cpu1 > > > > > > iwl_bg_restart() > > > __iwl_down() > > > iwl_scan_cancel_timeout > > > (do nothing since no pending scan) > > > > > > __ieee80211_start_scan > > > __set_bit(SCAN_HW_SCANNING, &local->scanning); > > > > Ah, ok, makes sense. I think some mutex trickery could possibly help > > here though? But I'm not sure it's worth it... > > I showed possible patch earlier in this thread, I will rethink about it and > eventually post. Perhaps applying the patch is good idea, since we will stop > seeing the warning it that harmless case, and be sure something is really > wrong when see it again. Well I'm thinking that iwlwifi will reject or queue a scan while it's down, and if we do the check while holding the mutex that we also hold here, it should close the race? And then we'll only accept the scan again when we've been started by mac80211 again, or something? 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
