| Message ID | 20170917194013.8658-12-erik.stromdahl@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Kalle Valo |
| Headers | show |
diff --git a/drivers/net/wireless/ath/ath10k/htt_tx.c b/drivers/net/wireless/ath/ath10k/htt_tx.c index 82d01139ff92..c74fc137ac67 100644 --- a/drivers/net/wireless/ath/ath10k/htt_tx.c +++ b/drivers/net/wireless/ath/ath10k/htt_tx.c @@ -153,6 +153,9 @@ void ath10k_htt_tx_txq_update(struct ieee80211_hw *hw, void ath10k_htt_tx_dec_pending(struct ath10k_htt *htt) { + if (htt->ar->is_high_latency) + return; + lockdep_assert_held(&htt->tx_lock); htt->num_pending_tx--; @@ -162,6 +165,9 @@ void ath10k_htt_tx_dec_pending(struct ath10k_htt *htt) int ath10k_htt_tx_inc_pending(struct ath10k_htt *htt) { + if (htt->ar->is_high_latency) + return 0; + lockdep_assert_held(&htt->tx_lock); if (htt->num_pending_tx >= htt->max_num_pending_tx)
High latency chipsest does not seem to send any HTT_T2H_MSG_TYPE_TX_COMPL_IND for outgoing frames. This means that htt->num_pending_tx will never be decremented and we will eventually hit the maximum limit. All outgoing packets will then be discarded. Signed-off-by: Erik Stromdahl <erik.stromdahl@gmail.com> --- drivers/net/wireless/ath/ath10k/htt_tx.c | 6 ++++++ 1 file changed, 6 insertions(+)