From: Senthil Balasubramanian Date: Wed, 10 Jun 2009 21:55:32 +0530 Subject: [ath9k-devel] [PATCH 7/7] ath9k: schedule aggregation regardless of queue depth In-Reply-To: <1244651133-6537-7-git-send-email-senthilkumar@atheros.com> References: <1244651133-6537-1-git-send-email-senthilkumar@atheros.com> <1244651133-6537-2-git-send-email-senthilkumar@atheros.com> <1244651133-6537-3-git-send-email-senthilkumar@atheros.com> <1244651133-6537-4-git-send-email-senthilkumar@atheros.com> <1244651133-6537-5-git-send-email-senthilkumar@atheros.com> <1244651133-6537-6-git-send-email-senthilkumar@atheros.com> <1244651133-6537-7-git-send-email-senthilkumar@atheros.com> Message-ID: <1244651133-6537-8-git-send-email-senthilkumar@atheros.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ath9k-devel@lists.ath9k.org schedule aggregation regardless of queue depth. Also we shouldn't remove holding descriptor from AC queue if it happens to be the only descriptor. Signed-off-by: Senthil Balasubramanian --- drivers/net/wireless/ath/ath9k/xmit.c | 15 +-------------- 1 files changed, 1 insertions(+), 14 deletions(-) diff --git a/drivers/net/wireless/ath/ath9k/xmit.c b/drivers/net/wireless/ath/ath9k/xmit.c index 328391e..eda2458 100644 --- a/drivers/net/wireless/ath/ath9k/xmit.c +++ b/drivers/net/wireless/ath/ath9k/xmit.c @@ -1113,8 +1113,7 @@ void ath_txq_schedule(struct ath_softc *sc, struct ath_txq *txq) if (tid->paused) continue; - if ((txq->axq_depth % 2) == 0) - ath_tx_sched_aggr(sc, txq, tid); + ath_tx_sched_aggr(sc, txq, tid); /* * add tid to round-robin queue if more frames @@ -1957,19 +1956,7 @@ static void ath_tx_processq(struct ath_softc *sc, struct ath_txq *txq) if (bf->bf_stale) { bf_held = bf; if (list_is_last(&bf_held->list, &txq->axq_q)) { - txq->axq_link = NULL; - txq->axq_linkbuf = NULL; spin_unlock_bh(&txq->axq_lock); - - /* - * The holding descriptor is the last - * descriptor in queue. It's safe to remove - * the last holding descriptor in BH context. - */ - spin_lock_bh(&sc->tx.txbuflock); - list_move_tail(&bf_held->list, &sc->tx.txbuf); - spin_unlock_bh(&sc->tx.txbuflock); - break; } else { bf = list_entry(bf_held->list.next, -- 1.6.0.4