When we disassociate in managed mode, we flush the queues
after mac80211 has already removed the station.
During that time, the pointer to ieee80211_sta to the
fw_id_to_mac_id map is -EINVAL. In that case we should not
set the station as being drained when the last Tx of this
station has exited the shared Tx queue since we are
flushing all the queues anyway.
The draining logic is meant to be used in GO / AP mode only.
In GO / AP mode, we set -EBUSY in the fw_id_to_mac_id map.
This is why testing the ieee80211_sta pointer in the
fw_id_to_mac_id map with IS_ERR isn't enough to set the
station as draining, we need to check that it is -EBUSY.
The only impact of the bug was a print:
Drained sta 1, but it is internal?
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
*/
spin_lock_bh(&mvmsta->lock);
sta = rcu_dereference(mvm->fw_id_to_mac_id[sta_id]);
- if (IS_ERR_OR_NULL(sta)) {
+ if (!sta || PTR_ERR(sta) == -EBUSY) {
/*
* Station disappeared in the meantime:
* so we are draining.
schedule_work(&mvm->sta_drained_wk);
}
spin_unlock_bh(&mvmsta->lock);
- } else if (!mvmsta) {
+ } else if (!mvmsta && PTR_ERR(sta) == -EBUSY) {
/* Tx response without STA, so we are draining */
set_bit(sta_id, mvm->sta_drained);
schedule_work(&mvm->sta_drained_wk);