iwlwifi: mvm: close the SP if we send fewer frames than expected in SP
When we have holes in the BA window, there might be frames
that have been ACKed between the read and the right
pointers. This means that these frames won't be scheduled
again by the SCD and the firwmare won't see them.
This invalidates the number of frames we tell the firmware
to send. When we detect this case, tell mac80211 to close
the SP and to send an EOSP so that the firmware can be in
sync.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/tx.c b/drivers/net/wireless/intel/iwlwifi/mvm/tx.c
index aaebb5d..cc0cdce 100644
--- a/drivers/net/wireless/intel/iwlwifi/mvm/tx.c
+++ b/drivers/net/wireless/intel/iwlwifi/mvm/tx.c
@@ -788,13 +788,43 @@
if (tid != IWL_TID_NON_QOS) {
struct iwl_mvm_tid_data *tid_data =
&mvmsta->tid_data[tid];
+ bool send_eosp_ndp = false;
spin_lock_bh(&mvmsta->lock);
tid_data->next_reclaimed = next_reclaimed;
IWL_DEBUG_TX_REPLY(mvm, "Next reclaimed packet:%d\n",
next_reclaimed);
iwl_mvm_check_ratid_empty(mvm, sta, tid);
+
+ if (mvmsta->sleep_tx_count) {
+ mvmsta->sleep_tx_count--;
+ if (mvmsta->sleep_tx_count &&
+ !iwl_mvm_tid_queued(tid_data)) {
+ /*
+ * The number of frames in the queue
+ * dropped to 0 even if we sent less
+ * frames than we thought we had on the
+ * Tx queue.
+ * This means we had holes in the BA
+ * window that we just filled, ask
+ * mac80211 to send EOSP since the
+ * firmware won't know how to do that.
+ * Send NDP and the firmware will send
+ * EOSP notification that will trigger
+ * a call to ieee80211_sta_eosp().
+ */
+ send_eosp_ndp = true;
+ }
+ }
+
spin_unlock_bh(&mvmsta->lock);
+ if (send_eosp_ndp) {
+ iwl_mvm_sta_modify_sleep_tx_count(mvm, sta,
+ IEEE80211_FRAME_RELEASE_UAPSD,
+ 1, tid, false, false);
+ mvmsta->sleep_tx_count = 0;
+ ieee80211_send_eosp_nullfunc(sta, tid);
+ }
}
if (mvmsta->next_status_eosp) {