Sometimes mac80211 doesn't push idle state downwards to the driver. Specifically, there are times in some functions where the "local->hw.conf.flags & IEEE80211_CONF_IDLE" may change but the equivalent of "drv_config(local, IEEE80211_CONF_CHANGE_IDLE)" does not end up being called. This is usually not all that problematic except, for example, suspending and resuming an idle ath9k device. If the device isn't marked idle when ieee80211_stop_device() is called, the device never gets put to sleep, and will end up in an unresponsive state on resume. As a precaution, explicitly call drv_config() before ieee80211_stop_device(), which should be a no-op under normal circumstances, but where this problem arises, it will shut down the ath9k where necessary. One example where this problem occurs is when I down an interface while a scan is in progress on ath9k. If the device was not shut down correctly and the system suspends and resumes repeatedly with ath9k, I end up with a fatal register read error (0x7000/deadbeef) when trying to bring the interface back up. Signed-off-by: Paul Stewart <pstew@xxxxxxxxxx> --- net/mac80211/iface.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index b6db237..5af5a89 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -536,6 +536,7 @@ static void ieee80211_do_stop(struct ieee80211_sub_if_data *sdata, if (local->ops->napi_poll) napi_disable(&local->napi); ieee80211_clear_tx_pending(local); + drv_config(local, IEEE80211_CONF_CHANGE_IDLE); ieee80211_stop_device(local); /* no reconfiguring after stop! */ -- 1.7.3.1 -- 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
