All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH 0/3] ath11k: improve TX path
@ 2019-11-20 20:29 John Crispin
  2019-11-20 20:29 ` [PATCH 1/3] ath11k: move some tx_status parsing to debugfs code John Crispin
                   ` (2 more replies)
  0 siblings, 3 replies; 5+ messages in thread
From: John Crispin @ 2019-11-20 20:29 UTC (permalink / raw)
  To: Kalle Valo; +Cc: ath11k, John Crispin

This series improves the TX path. The following numbers
are peak performance measured.

prior to series
[SUM]  30.00-31.00  sec   104 MBytes   874 Mbits/sec
series applied
[SUM]  35.00-36.00  sec   124 MBytes  1.04 Gbits/sec
series applied + tx encap
[SUM]  77.00-78.00  sec   167 MBytes  1.40 Gbits/sec

John Crispin (3):
  ath11k: move some tx_status parsing to debugfs code
  ath11k: optimise ath11k_dp_tx_completion_handler
  ath11k: optimize ath11k_hal_tx_status_parse

 drivers/net/wireless/ath/ath11k/debugfs_sta.c | 66 +++++++++-------
 drivers/net/wireless/ath/ath11k/dp.c          | 16 ++--
 drivers/net/wireless/ath/ath11k/dp.h          | 11 ++-
 drivers/net/wireless/ath/ath11k/dp_tx.c       | 79 ++++++++++++-------
 drivers/net/wireless/ath/ath11k/hal_tx.c      | 73 -----------------
 drivers/net/wireless/ath/ath11k/hal_tx.h      | 13 +--
 6 files changed, 101 insertions(+), 157 deletions(-)

-- 
2.20.1


_______________________________________________
ath11k mailing list
ath11k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath11k

^ permalink raw reply	[flat|nested] 5+ messages in thread
* [PATCH 1/3] ath11k: move some tx_status parsing to debugfs code
  2019-11-20 20:29 [PATCH 0/3] ath11k: improve TX path John Crispin
@ 2019-11-20 20:29 ` John Crispin
  2019-11-29  8:32   ` Kalle Valo
  2019-11-20 20:29 ` [PATCH 2/3] ath11k: optimise ath11k_dp_tx_completion_handler John Crispin
  2019-11-20 20:29 ` [PATCH 3/3] ath11k: optimize ath11k_hal_tx_status_parse John Crispin
  2 siblings, 1 reply; 5+ messages in thread
From: John Crispin @ 2019-11-20 20:29 UTC (permalink / raw)
  To: Kalle Valo; +Cc: ath11k, John Crispin

Some of the fields are only used by debugfs. Move the parsing of these
from the data hot path to the debugfs code.

Signed-off-by: John Crispin <john@phrozen.org>
---
 drivers/net/wireless/ath/ath11k/debugfs_sta.c | 66 +++++++++++--------
 drivers/net/wireless/ath/ath11k/hal_tx.c      | 19 +-----
 drivers/net/wireless/ath/ath11k/hal_tx.h      |  7 +-
 3 files changed, 39 insertions(+), 53 deletions(-)

diff --git a/drivers/net/wireless/ath/ath11k/debugfs_sta.c b/drivers/net/wireless/ath/ath11k/debugfs_sta.c
index 3c5f931e22a9..3cdc34218a7d 100644
--- a/drivers/net/wireless/ath/ath11k/debugfs_sta.c
+++ b/drivers/net/wireless/ath/ath11k/debugfs_sta.c
@@ -129,12 +129,17 @@ void ath11k_update_per_peer_stats_from_txcompl(struct ath11k *ar,
 {
 	struct ath11k_base *ab = ar->ab;
 	struct ath11k_per_peer_tx_stats *peer_stats = &ar->cached_stats;
+	enum hal_tx_rate_stats_pkt_type pkt_type;
+	enum hal_tx_rate_stats_sgi sgi;
+	enum hal_tx_rate_stats_bw bw;
 	struct ath11k_peer *peer;
 	struct ath11k_sta *arsta;
 	struct ieee80211_sta *sta;
+	u16 num_tones_in_ru;
 	u16 rate;
 	u8 rate_idx;
 	int ret;
+	u8 mcs;
 
 	rcu_read_lock();
 	spin_lock_bh(&ab->base_lock);
@@ -150,51 +155,54 @@ void ath11k_update_per_peer_stats_from_txcompl(struct ath11k *ar,
 	arsta = (struct ath11k_sta *)sta->drv_priv;
 
 	memset(&arsta->txrate, 0, sizeof(arsta->txrate));
-
-	if (ts->pkt_type == HAL_TX_RATE_STATS_PKT_TYPE_11A ||
-	    ts->pkt_type == HAL_TX_RATE_STATS_PKT_TYPE_11B) {
-		ret = ath11k_mac_hw_ratecode_to_legacy_rate(ts->mcs,
-							    ts->pkt_type,
+	pkt_type = FIELD_GET(HAL_TX_RATE_STATS_INFO0_PKT_TYPE,
+			     ts->rate_stats);
+	mcs = FIELD_GET(HAL_TX_RATE_STATS_INFO0_MCS,
+			ts->rate_stats);
+	sgi = FIELD_GET(HAL_TX_RATE_STATS_INFO0_SGI,
+			ts->rate_stats);
+	num_tones_in_ru = FIELD_GET(HAL_TX_RATE_STATS_INFO0_TONES_IN_RU,
+				    ts->rate_stats);
+	bw = FIELD_GET(HAL_TX_RATE_STATS_INFO0_BW, ts->rate_stats);
+
+	if (pkt_type == HAL_TX_RATE_STATS_PKT_TYPE_11A ||
+	    pkt_type == HAL_TX_RATE_STATS_PKT_TYPE_11B) {
+		ret = ath11k_mac_hw_ratecode_to_legacy_rate(mcs,
+							    pkt_type,
 							    &rate_idx,
 							    &rate);
-		if (ret < 0) {
-			spin_unlock_bh(&ab->base_lock);
-			rcu_read_unlock();
-			return;
-		}
+		if (ret < 0)
+			goto err_out;
 		arsta->txrate.legacy = rate;
-	} else if (ts->pkt_type == HAL_TX_RATE_STATS_PKT_TYPE_11N) {
-		if (ts->mcs > 7) {
-			ath11k_warn(ab, "Invalid HT mcs index %d\n", ts->mcs);
-			spin_unlock_bh(&ab->base_lock);
-			rcu_read_unlock();
-			return;
+	} else if (pkt_type == HAL_TX_RATE_STATS_PKT_TYPE_11N) {
+		if (mcs > 7) {
+			ath11k_warn(ab, "Invalid HT mcs index %d\n", mcs);
+			goto err_out;
 		}
 
-		arsta->txrate.mcs = ts->mcs + 8 * (arsta->last_txrate.nss - 1);
+		arsta->txrate.mcs = mcs + 8 * (arsta->last_txrate.nss - 1);
 		arsta->txrate.flags = RATE_INFO_FLAGS_MCS;
-		if (ts->sgi)
+		if (sgi)
 			arsta->txrate.flags |= RATE_INFO_FLAGS_SHORT_GI;
-	} else if (ts->pkt_type == HAL_TX_RATE_STATS_PKT_TYPE_11AC) {
-		if (ts->mcs > 9) {
-			ath11k_warn(ab, "Invalid VHT mcs index %d\n", ts->mcs);
-			spin_unlock_bh(&ab->base_lock);
-			rcu_read_unlock();
-			return;
+	} else if (pkt_type == HAL_TX_RATE_STATS_PKT_TYPE_11AC) {
+		if (mcs > 9) {
+			ath11k_warn(ab, "Invalid VHT mcs index %d\n", mcs);
+			goto err_out;
 		}
 
-		arsta->txrate.mcs = ts->mcs;
+		arsta->txrate.mcs = mcs;
 		arsta->txrate.flags = RATE_INFO_FLAGS_VHT_MCS;
-		if (ts->sgi)
+		if (sgi)
 			arsta->txrate.flags |= RATE_INFO_FLAGS_SHORT_GI;
-	} else {
-		/*TODO: update HE rates */
+	} else if (pkt_type == HAL_TX_RATE_STATS_PKT_TYPE_11AX) {
+		/* TODO */
 	}
 
 	arsta->txrate.nss = arsta->last_txrate.nss;
-	arsta->txrate.bw = ath11k_mac_bw_to_mac80211_bw(ts->bw);
+	arsta->txrate.bw = ath11k_mac_bw_to_mac80211_bw(bw);
 
 	ath11k_accumulate_per_peer_tx_stats(arsta, peer_stats, rate_idx);
+err_out:
 	spin_unlock_bh(&ab->base_lock);
 	rcu_read_unlock();
 }
diff --git a/drivers/net/wireless/ath/ath11k/hal_tx.c b/drivers/net/wireless/ath/ath11k/hal_tx.c
index 72a51ed65a51..c2a3a9d2e534 100644
--- a/drivers/net/wireless/ath/ath11k/hal_tx.c
+++ b/drivers/net/wireless/ath/ath11k/hal_tx.c
@@ -127,24 +127,7 @@ void ath11k_hal_tx_status_parse(struct ath11k_base *ab,
 	if (!(desc->rate_stats.info0 & HAL_TX_RATE_STATS_INFO0_VALID))
 		return;
 
-	ts->flags |= HAL_TX_STATUS_FLAGS_RATE_STATS_VALID;
-	ts->tsf = desc->rate_stats.tsf;
-	ts->bw = FIELD_GET(HAL_TX_RATE_STATS_INFO0_BW, desc->rate_stats.info0);
-	ts->pkt_type = FIELD_GET(HAL_TX_RATE_STATS_INFO0_PKT_TYPE,
-				 desc->rate_stats.info0);
-	if (desc->rate_stats.info0 & HAL_TX_RATE_STATS_INFO0_STBC)
-		ts->flags |= HAL_TX_STATUS_FLAGS_RATE_STBC;
-	if (desc->rate_stats.info0 & HAL_TX_RATE_STATS_INFO0_LDPC)
-		ts->flags |= HAL_TX_STATUS_FLAGS_RATE_LDPC;
-	if (desc->rate_stats.info0 & HAL_TX_RATE_STATS_INFO0_OFDMA_TX)
-		ts->flags |= HAL_TX_STATUS_FLAGS_OFDMA;
-
-	ts->sgi = FIELD_GET(HAL_TX_RATE_STATS_INFO0_SGI,
-			    desc->rate_stats.info0);
-	ts->mcs = FIELD_GET(HAL_TX_RATE_STATS_INFO0_MCS,
-			    desc->rate_stats.info0);
-	ts->num_tones_in_ru = FIELD_GET(HAL_TX_RATE_STATS_INFO0_TONES_IN_RU,
-					desc->rate_stats.info0);
+	ts->rate_stats = desc->rate_stats.info0;
 }
 
 void ath11k_hal_tx_set_dscp_tid_map(struct ath11k_base *ab, int id)
diff --git a/drivers/net/wireless/ath/ath11k/hal_tx.h b/drivers/net/wireless/ath/ath11k/hal_tx.h
index 5217eaf9da50..cf99e2b31f9f 100644
--- a/drivers/net/wireless/ath/ath11k/hal_tx.h
+++ b/drivers/net/wireless/ath/ath11k/hal_tx.h
@@ -51,17 +51,12 @@ struct hal_tx_status {
 	u32 desc_id;
 	enum hal_wbm_tqm_rel_reason status;
 	u8 ack_rssi;
-	enum hal_tx_rate_stats_bw bw;
-	enum hal_tx_rate_stats_pkt_type pkt_type;
-	enum hal_tx_rate_stats_sgi sgi;
-	u8 mcs;
-	u16 num_tones_in_ru;
 	u32 flags; /* %HAL_TX_STATUS_FLAGS_ */
-	u32 tsf;
 	u32 ppdu_id;
 	u8 try_cnt;
 	u8 tid;
 	u16 peer_id;
+	u32 rate_stats;
 };
 
 void ath11k_hal_tx_cmd_desc_setup(struct ath11k_base *ab, void *cmd,
-- 
2.20.1


_______________________________________________
ath11k mailing list
ath11k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath11k

^ permalink raw reply related	[flat|nested] 5+ messages in thread
* [PATCH 2/3] ath11k: optimise ath11k_dp_tx_completion_handler
  2019-11-20 20:29 [PATCH 0/3] ath11k: improve TX path John Crispin
  2019-11-20 20:29 ` [PATCH 1/3] ath11k: move some tx_status parsing to debugfs code John Crispin
@ 2019-11-20 20:29 ` John Crispin
  2019-11-20 20:29 ` [PATCH 3/3] ath11k: optimize ath11k_hal_tx_status_parse John Crispin
  2 siblings, 0 replies; 5+ messages in thread
From: John Crispin @ 2019-11-20 20:29 UTC (permalink / raw)
  To: Kalle Valo; +Cc: ath11k, John Crispin

the current code does 4 memcpys for each completion frame.
1) duplicate the desc
2 + 3) inside kfifo insertion
4) kfifo remove

The code simply drops the kfifo and uses a trivial ring buffer. This
requires a single memcpy for insertion. There is no removal needed as
we can simply use the inserted data for processing. As the code runs
inside the NAPI context it is atomic and there is no need for most of
the locking.

Signed-off-by: John Crispin <john@phrozen.org>
---
 drivers/net/wireless/ath/ath11k/dp.c     | 16 ++++-----
 drivers/net/wireless/ath/ath11k/dp.h     | 11 +++----
 drivers/net/wireless/ath/ath11k/dp_tx.c  | 41 +++++++++---------------
 drivers/net/wireless/ath/ath11k/hal_tx.c | 13 --------
 drivers/net/wireless/ath/ath11k/hal_tx.h |  2 --
 5 files changed, 27 insertions(+), 56 deletions(-)

diff --git a/drivers/net/wireless/ath/ath11k/dp.c b/drivers/net/wireless/ath/ath11k/dp.c
index 72c21cf6a352..a3e2b639cfcd 100644
--- a/drivers/net/wireless/ath/ath11k/dp.c
+++ b/drivers/net/wireless/ath/ath11k/dp.c
@@ -3,7 +3,6 @@
  * Copyright (c) 2018-2019 The Linux Foundation. All rights reserved.
  */
 
-#include <linux/kfifo.h>
 #include "core.h"
 #include "dp_tx.h"
 #include "hal_tx.h"
@@ -822,10 +821,7 @@ void ath11k_dp_free(struct ath11k_base *ab)
 			     ath11k_dp_tx_pending_cleanup, ab);
 		idr_destroy(&dp->tx_ring[i].txbuf_idr);
 		spin_unlock_bh(&dp->tx_ring[i].tx_idr_lock);
-
-		spin_lock_bh(&dp->tx_ring[i].tx_status_lock);
-		kfifo_free(&dp->tx_ring[i].tx_status_fifo);
-		spin_unlock_bh(&dp->tx_ring[i].tx_status_lock);
+		kfree(&dp->tx_ring[i].tx_status);
 	}
 
 	/* Deinit any SOC level resource */
@@ -865,17 +861,17 @@ int ath11k_dp_alloc(struct ath11k_base *ab)
 	if (ret)
 		goto fail_link_desc_cleanup;
 
-	size = roundup_pow_of_two(DP_TX_COMP_RING_SIZE);
+	size = sizeof(struct hal_wbm_release_ring) * DP_TX_COMP_RING_SIZE;
 
 	for (i = 0; i < DP_TCL_NUM_RING_MAX; i++) {
 		idr_init(&dp->tx_ring[i].txbuf_idr);
 		spin_lock_init(&dp->tx_ring[i].tx_idr_lock);
 		dp->tx_ring[i].tcl_data_ring_id = i;
 
-		spin_lock_init(&dp->tx_ring[i].tx_status_lock);
-		ret = kfifo_alloc(&dp->tx_ring[i].tx_status_fifo, size,
-				  GFP_KERNEL);
-		if (ret)
+		dp->tx_ring[i].tx_status_head = 0;
+		dp->tx_ring[i].tx_status_tail = DP_TX_COMP_RING_SIZE - 1;
+		dp->tx_ring[i].tx_status = kmalloc(size, GFP_KERNEL);
+		if (!dp->tx_ring[i].tx_status)
 			goto fail_cmn_srng_cleanup;
 	}
 
diff --git a/drivers/net/wireless/ath/ath11k/dp.h b/drivers/net/wireless/ath/ath11k/dp.h
index db216f055c42..02e55986e521 100644
--- a/drivers/net/wireless/ath/ath11k/dp.h
+++ b/drivers/net/wireless/ath/ath11k/dp.h
@@ -6,7 +6,6 @@
 #ifndef ATH11K_DP_H
 #define ATH11K_DP_H
 
-#include <linux/kfifo.h>
 #include "hal_rx.h"
 
 struct ath11k_base;
@@ -58,6 +57,8 @@ struct dp_rxdma_ring {
 	int bufs_max;
 };
 
+#define ATH11K_TX_COMPL_NEXT(x)	((x + 1) % DP_TX_COMP_RING_SIZE)
+
 struct dp_tx_ring {
 	u8 tcl_data_ring_id;
 	struct dp_srng tcl_data_ring;
@@ -65,11 +66,9 @@ struct dp_tx_ring {
 	struct idr txbuf_idr;
 	/* Protects txbuf_idr and num_pending */
 	spinlock_t tx_idr_lock;
-	DECLARE_KFIFO_PTR(tx_status_fifo, struct hal_wbm_release_ring);
-	/* lock to protect tx_status_fifo because tx_status_fifo can be
-	 * accessed concurrently.
-	 */
-	spinlock_t tx_status_lock;
+	struct hal_wbm_release_ring *tx_status;
+	int tx_status_head;
+	int tx_status_tail;
 };
 
 struct ath11k_pdev_mon_stats {
diff --git a/drivers/net/wireless/ath/ath11k/dp_tx.c b/drivers/net/wireless/ath/ath11k/dp_tx.c
index 28ebc414533e..5d6403cf99ab 100644
--- a/drivers/net/wireless/ath/ath11k/dp_tx.c
+++ b/drivers/net/wireless/ath/ath11k/dp_tx.c
@@ -79,7 +79,6 @@ int ath11k_dp_tx(struct ath11k *ar, struct ath11k_vif *arvif,
 	struct hal_srng *tcl_ring;
 	struct ieee80211_hdr *hdr = (void *)skb->data;
 	struct dp_tx_ring *tx_ring;
-	u8 cached_desc[HAL_TCL_DESC_LEN];
 	void *hal_tcl_desc;
 	u8 pool_id;
 	u8 hal_ring_id;
@@ -167,8 +166,6 @@ int ath11k_dp_tx(struct ath11k *ar, struct ath11k_vif *arvif,
 	skb_cb->vif = arvif->vif;
 	skb_cb->ar = ar;
 
-	ath11k_hal_tx_cmd_desc_setup(ab, cached_desc, &ti);
-
 	hal_ring_id = tx_ring->tcl_data_ring.ring_id;
 	tcl_ring = &ab->hal.srng_list[hal_ring_id];
 
@@ -188,7 +185,8 @@ int ath11k_dp_tx(struct ath11k *ar, struct ath11k_vif *arvif,
 		goto fail_unmap_dma;
 	}
 
-	ath11k_hal_tx_desc_sync(cached_desc, hal_tcl_desc);
+	ath11k_hal_tx_cmd_desc_setup(ab, hal_tcl_desc +
+					 sizeof(struct hal_tlv_hdr), &ti);
 
 	ath11k_hal_srng_access_end(ab, tcl_ring);
 
@@ -432,47 +430,44 @@ void ath11k_dp_tx_completion_handler(struct ath11k_base *ab, int ring_id)
 	int hal_ring_id = dp->tx_ring[ring_id].tcl_comp_ring.ring_id;
 	struct hal_srng *status_ring = &ab->hal.srng_list[hal_ring_id];
 	struct sk_buff *msdu;
-	struct hal_wbm_release_ring tx_status;
 	struct hal_tx_status ts;
 	struct dp_tx_ring *tx_ring = &dp->tx_ring[ring_id];
 	u32 *desc;
 	u32 msdu_id;
 	u8 mac_id;
 
-	spin_lock_bh(&status_ring->lock);
-
 	ath11k_hal_srng_access_begin(ab, status_ring);
 
-	spin_lock_bh(&tx_ring->tx_status_lock);
-	while (!kfifo_is_full(&tx_ring->tx_status_fifo) &&
+	while ((ATH11K_TX_COMPL_NEXT(tx_ring->tx_status_head) != tx_ring->tx_status_tail) &&
 	       (desc = ath11k_hal_srng_dst_get_next_entry(ab, status_ring))) {
-		ath11k_hal_tx_status_desc_sync((void *)desc,
-					       (void *)&tx_status);
-		kfifo_put(&tx_ring->tx_status_fifo, tx_status);
+		memcpy(&tx_ring->tx_status[tx_ring->tx_status_head],
+		       desc, sizeof(struct hal_wbm_release_ring));
+		tx_ring->tx_status_head =
+			ATH11K_TX_COMPL_NEXT(tx_ring->tx_status_head);
 	}
 
 	if ((ath11k_hal_srng_dst_peek(ab, status_ring) != NULL) &&
-	    kfifo_is_full(&tx_ring->tx_status_fifo)) {
+	    (ATH11K_TX_COMPL_NEXT(tx_ring->tx_status_head) == tx_ring->tx_status_tail)) {
 		/* TODO: Process pending tx_status messages when kfifo_is_full() */
 		ath11k_warn(ab, "Unable to process some of the tx_status ring desc because status_fifo is full\n");
 	}
 
-	spin_unlock_bh(&tx_ring->tx_status_lock);
-
 	ath11k_hal_srng_access_end(ab, status_ring);
-	spin_unlock_bh(&status_ring->lock);
 
-	spin_lock_bh(&tx_ring->tx_status_lock);
-	while (kfifo_get(&tx_ring->tx_status_fifo, &tx_status)) {
-		memset(&ts, 0, sizeof(ts));
-		ath11k_hal_tx_status_parse(ab, &tx_status, &ts);
+	while (ATH11K_TX_COMPL_NEXT(tx_ring->tx_status_tail) != tx_ring->tx_status_head) {
+		struct hal_wbm_release_ring *tx_status;
+
+		tx_ring->tx_status_tail =
+			ATH11K_TX_COMPL_NEXT(tx_ring->tx_status_tail);
+		tx_status = &tx_ring->tx_status[tx_ring->tx_status_tail];
+		ath11k_hal_tx_status_parse(ab, tx_status, &ts);
 
 		mac_id = FIELD_GET(DP_TX_DESC_ID_MAC_ID, ts.desc_id);
 		msdu_id = FIELD_GET(DP_TX_DESC_ID_MSDU_ID, ts.desc_id);
 
 		if (ts.buf_rel_source == HAL_WBM_REL_SRC_MODULE_FW) {
 			ath11k_dp_tx_process_htt_tx_complete(ab,
-							     (void *)&tx_status,
+							     (void *)tx_status,
 							     mac_id, msdu_id,
 							     tx_ring);
 			continue;
@@ -494,12 +489,8 @@ void ath11k_dp_tx_completion_handler(struct ath11k_base *ab, int ring_id)
 		if (atomic_dec_and_test(&ar->dp.num_tx_pending))
 			wake_up(&ar->dp.tx_empty_waitq);
 
-		/* TODO: Locking optimization so that tx_completion for an msdu
-		 * is not called with tx_status_lock acquired
-		 */
 		ath11k_dp_tx_complete_msdu(ar, msdu, &ts);
 	}
-	spin_unlock_bh(&tx_ring->tx_status_lock);
 }
 
 int ath11k_dp_tx_send_reo_cmd(struct ath11k_base *ab, struct dp_rx_tid *rx_tid,
diff --git a/drivers/net/wireless/ath/ath11k/hal_tx.c b/drivers/net/wireless/ath/ath11k/hal_tx.c
index c2a3a9d2e534..e8710bbbbc3a 100644
--- a/drivers/net/wireless/ath/ath11k/hal_tx.c
+++ b/drivers/net/wireless/ath/ath11k/hal_tx.c
@@ -74,19 +74,6 @@ void ath11k_hal_tx_cmd_desc_setup(struct ath11k_base *ab, void *cmd,
 	tcl_cmd->info4 = 0;
 }
 
-/* Commit the descriptor to hardware */
-void ath11k_hal_tx_desc_sync(void *tx_desc_cached, void *hw_desc)
-{
-	memcpy(hw_desc + sizeof(struct hal_tlv_hdr), tx_desc_cached,
-	       sizeof(struct hal_tcl_data_cmd));
-}
-
-/* Get the descriptor status from hardware */
-void ath11k_hal_tx_status_desc_sync(void *hw_desc, void *local_desc)
-{
-	memcpy(local_desc, hw_desc, HAL_TX_STATUS_DESC_LEN);
-}
-
 void ath11k_hal_tx_status_parse(struct ath11k_base *ab,
 				struct hal_wbm_release_ring *desc,
 				struct hal_tx_status *ts)
diff --git a/drivers/net/wireless/ath/ath11k/hal_tx.h b/drivers/net/wireless/ath/ath11k/hal_tx.h
index cf99e2b31f9f..5fe89b729a6e 100644
--- a/drivers/net/wireless/ath/ath11k/hal_tx.h
+++ b/drivers/net/wireless/ath/ath11k/hal_tx.h
@@ -61,11 +61,9 @@ struct hal_tx_status {
 
 void ath11k_hal_tx_cmd_desc_setup(struct ath11k_base *ab, void *cmd,
 				  struct hal_tx_info *ti);
-void ath11k_hal_tx_desc_sync(void *tx_desc_cached, void *hw_desc);
 void ath11k_hal_tx_status_parse(struct ath11k_base *ab,
 				struct hal_wbm_release_ring *desc,
 				struct hal_tx_status *ts);
-void ath11k_hal_tx_status_desc_sync(void *hw_desc, void *local_desc);
 void ath11k_hal_tx_set_dscp_tid_map(struct ath11k_base *ab, int id);
 int ath11k_hal_reo_cmd_send(struct ath11k_base *ab, struct hal_srng *srng,
 			    enum hal_reo_cmd_type type,
-- 
2.20.1


_______________________________________________
ath11k mailing list
ath11k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath11k

^ permalink raw reply related	[flat|nested] 5+ messages in thread
* [PATCH 3/3] ath11k: optimize ath11k_hal_tx_status_parse
  2019-11-20 20:29 [PATCH 0/3] ath11k: improve TX path John Crispin
  2019-11-20 20:29 ` [PATCH 1/3] ath11k: move some tx_status parsing to debugfs code John Crispin
  2019-11-20 20:29 ` [PATCH 2/3] ath11k: optimise ath11k_dp_tx_completion_handler John Crispin
@ 2019-11-20 20:29 ` John Crispin
  2 siblings, 0 replies; 5+ messages in thread
From: John Crispin @ 2019-11-20 20:29 UTC (permalink / raw)
  To: Kalle Valo; +Cc: ath11k, John Crispin

Moving the function into dp_tx.c allows gcc to optimize the code better
and also avoid chace invalidates and context switches.

Signed-off-by: John Crispin <john@phrozen.org>
---
 drivers/net/wireless/ath/ath11k/dp.h     |  2 +-
 drivers/net/wireless/ath/ath11k/dp_tx.c  | 38 +++++++++++++++++++--
 drivers/net/wireless/ath/ath11k/hal_tx.c | 43 ------------------------
 drivers/net/wireless/ath/ath11k/hal_tx.h |  4 ---
 4 files changed, 37 insertions(+), 50 deletions(-)

diff --git a/drivers/net/wireless/ath/ath11k/dp.h b/drivers/net/wireless/ath/ath11k/dp.h
index 02e55986e521..4917f2bd46f1 100644
--- a/drivers/net/wireless/ath/ath11k/dp.h
+++ b/drivers/net/wireless/ath/ath11k/dp.h
@@ -57,7 +57,7 @@ struct dp_rxdma_ring {
 	int bufs_max;
 };
 
-#define ATH11K_TX_COMPL_NEXT(x)	((x + 1) % DP_TX_COMP_RING_SIZE)
+#define ATH11K_TX_COMPL_NEXT(x)	(((x) + 1) % DP_TX_COMP_RING_SIZE)
 
 struct dp_tx_ring {
 	u8 tcl_data_ring_id;
diff --git a/drivers/net/wireless/ath/ath11k/dp_tx.c b/drivers/net/wireless/ath/ath11k/dp_tx.c
index 5d6403cf99ab..828f8984235e 100644
--- a/drivers/net/wireless/ath/ath11k/dp_tx.c
+++ b/drivers/net/wireless/ath/ath11k/dp_tx.c
@@ -423,6 +423,37 @@ static void ath11k_dp_tx_complete_msdu(struct ath11k *ar,
 	rcu_read_unlock();
 }
 
+static inline void ath11k_hal_tx_status_parse(struct ath11k_base *ab,
+					      struct hal_wbm_release_ring *desc,
+					      struct hal_tx_status *ts)
+{
+	ts->buf_rel_source =
+		FIELD_GET(HAL_WBM_RELEASE_INFO0_REL_SRC_MODULE, desc->info0);
+	if (ts->buf_rel_source != HAL_WBM_REL_SRC_MODULE_FW &&
+	    ts->buf_rel_source != HAL_WBM_REL_SRC_MODULE_TQM)
+		return;
+
+	if (ts->buf_rel_source == HAL_WBM_REL_SRC_MODULE_FW)
+		return;
+
+	ts->status = FIELD_GET(HAL_WBM_RELEASE_INFO0_TQM_RELEASE_REASON,
+			       desc->info0);
+	ts->ppdu_id = FIELD_GET(HAL_WBM_RELEASE_INFO1_TQM_STATUS_NUMBER,
+				desc->info1);
+	ts->try_cnt = FIELD_GET(HAL_WBM_RELEASE_INFO1_TRANSMIT_COUNT,
+				desc->info1);
+	ts->ack_rssi = FIELD_GET(HAL_WBM_RELEASE_INFO2_ACK_FRAME_RSSI,
+				 desc->info2);
+	if (desc->info2 & HAL_WBM_RELEASE_INFO2_FIRST_MSDU)
+		ts->flags |= HAL_TX_STATUS_FLAGS_FIRST_MSDU;
+	ts->peer_id = FIELD_GET(HAL_WBM_RELEASE_INFO3_PEER_ID, desc->info3);
+	ts->tid = FIELD_GET(HAL_WBM_RELEASE_INFO3_TID, desc->info3);
+	if (desc->rate_stats.info0 & HAL_TX_RATE_STATS_INFO0_VALID)
+		ts->rate_stats = desc->rate_stats.info0;
+	else
+		ts->rate_stats = 0;
+}
+
 void ath11k_dp_tx_completion_handler(struct ath11k_base *ab, int ring_id)
 {
 	struct ath11k *ar;
@@ -456,14 +487,17 @@ void ath11k_dp_tx_completion_handler(struct ath11k_base *ab, int ring_id)
 
 	while (ATH11K_TX_COMPL_NEXT(tx_ring->tx_status_tail) != tx_ring->tx_status_head) {
 		struct hal_wbm_release_ring *tx_status;
+		u32 desc_id;
 
 		tx_ring->tx_status_tail =
 			ATH11K_TX_COMPL_NEXT(tx_ring->tx_status_tail);
 		tx_status = &tx_ring->tx_status[tx_ring->tx_status_tail];
 		ath11k_hal_tx_status_parse(ab, tx_status, &ts);
 
-		mac_id = FIELD_GET(DP_TX_DESC_ID_MAC_ID, ts.desc_id);
-		msdu_id = FIELD_GET(DP_TX_DESC_ID_MSDU_ID, ts.desc_id);
+		desc_id = FIELD_GET(BUFFER_ADDR_INFO1_SW_COOKIE,
+				    tx_status->buf_addr_info.info1);
+		mac_id = FIELD_GET(DP_TX_DESC_ID_MAC_ID, desc_id);
+		msdu_id = FIELD_GET(DP_TX_DESC_ID_MSDU_ID, desc_id);
 
 		if (ts.buf_rel_source == HAL_WBM_REL_SRC_MODULE_FW) {
 			ath11k_dp_tx_process_htt_tx_complete(ab,
diff --git a/drivers/net/wireless/ath/ath11k/hal_tx.c b/drivers/net/wireless/ath/ath11k/hal_tx.c
index e8710bbbbc3a..e4aa7e8a1284 100644
--- a/drivers/net/wireless/ath/ath11k/hal_tx.c
+++ b/drivers/net/wireless/ath/ath11k/hal_tx.c
@@ -74,49 +74,6 @@ void ath11k_hal_tx_cmd_desc_setup(struct ath11k_base *ab, void *cmd,
 	tcl_cmd->info4 = 0;
 }
 
-void ath11k_hal_tx_status_parse(struct ath11k_base *ab,
-				struct hal_wbm_release_ring *desc,
-				struct hal_tx_status *ts)
-{
-	ts->buf_rel_source =
-		FIELD_GET(HAL_WBM_RELEASE_INFO0_REL_SRC_MODULE, desc->info0);
-	if (ts->buf_rel_source != HAL_WBM_REL_SRC_MODULE_FW &&
-	    ts->buf_rel_source != HAL_WBM_REL_SRC_MODULE_TQM)
-		return;
-
-	ts->desc_id = FIELD_GET(BUFFER_ADDR_INFO1_SW_COOKIE,
-				desc->buf_addr_info.info1);
-
-	if (ts->buf_rel_source == HAL_WBM_REL_SRC_MODULE_FW)
-		return;
-
-	ts->status = FIELD_GET(HAL_WBM_RELEASE_INFO0_TQM_RELEASE_REASON,
-			       desc->info0);
-	ts->ppdu_id = FIELD_GET(HAL_WBM_RELEASE_INFO1_TQM_STATUS_NUMBER,
-				desc->info1);
-	ts->try_cnt = FIELD_GET(HAL_WBM_RELEASE_INFO1_TRANSMIT_COUNT,
-				desc->info1);
-
-	ts->ack_rssi = FIELD_GET(HAL_WBM_RELEASE_INFO2_ACK_FRAME_RSSI,
-				 desc->info2);
-	if (desc->info2 & HAL_WBM_RELEASE_INFO2_FIRST_MSDU)
-		ts->flags |= HAL_TX_STATUS_FLAGS_FIRST_MSDU;
-
-	if (desc->info2 & HAL_WBM_RELEASE_INFO2_LAST_MSDU)
-		ts->flags |= HAL_TX_STATUS_FLAGS_LAST_MSDU;
-
-	if (desc->info2 & HAL_WBM_RELEASE_INFO2_MSDU_IN_AMSDU)
-		ts->flags |= HAL_TX_STATUS_FLAGS_MSDU_IN_AMSDU;
-
-	ts->peer_id = FIELD_GET(HAL_WBM_RELEASE_INFO3_PEER_ID, desc->info3);
-	ts->tid = FIELD_GET(HAL_WBM_RELEASE_INFO3_TID, desc->info3);
-
-	if (!(desc->rate_stats.info0 & HAL_TX_RATE_STATS_INFO0_VALID))
-		return;
-
-	ts->rate_stats = desc->rate_stats.info0;
-}
-
 void ath11k_hal_tx_set_dscp_tid_map(struct ath11k_base *ab, int id)
 {
 	u32 ctrl_reg_val;
diff --git a/drivers/net/wireless/ath/ath11k/hal_tx.h b/drivers/net/wireless/ath/ath11k/hal_tx.h
index 5fe89b729a6e..ce48a61bfb66 100644
--- a/drivers/net/wireless/ath/ath11k/hal_tx.h
+++ b/drivers/net/wireless/ath/ath11k/hal_tx.h
@@ -48,7 +48,6 @@ struct hal_tx_info {
 /* Tx status parsed from srng desc */
 struct hal_tx_status {
 	enum hal_wbm_rel_src_module buf_rel_source;
-	u32 desc_id;
 	enum hal_wbm_tqm_rel_reason status;
 	u8 ack_rssi;
 	u32 flags; /* %HAL_TX_STATUS_FLAGS_ */
@@ -61,9 +60,6 @@ struct hal_tx_status {
 
 void ath11k_hal_tx_cmd_desc_setup(struct ath11k_base *ab, void *cmd,
 				  struct hal_tx_info *ti);
-void ath11k_hal_tx_status_parse(struct ath11k_base *ab,
-				struct hal_wbm_release_ring *desc,
-				struct hal_tx_status *ts);
 void ath11k_hal_tx_set_dscp_tid_map(struct ath11k_base *ab, int id);
 int ath11k_hal_reo_cmd_send(struct ath11k_base *ab, struct hal_srng *srng,
 			    enum hal_reo_cmd_type type,
-- 
2.20.1


_______________________________________________
ath11k mailing list
ath11k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath11k

^ permalink raw reply related	[flat|nested] 5+ messages in thread
* Re: [PATCH 1/3] ath11k: move some tx_status parsing to debugfs code
  2019-11-20 20:29 ` [PATCH 1/3] ath11k: move some tx_status parsing to debugfs code John Crispin
@ 2019-11-29  8:32   ` Kalle Valo
  0 siblings, 0 replies; 5+ messages in thread
From: Kalle Valo @ 2019-11-29  8:32 UTC (permalink / raw)
  To: John Crispin; +Cc: ath11k

John Crispin <john@phrozen.org> wrote:

> Some of the fields are only used by debugfs. Move the parsing of these
> from the data hot path to the debugfs code.
> 
> Signed-off-by: John Crispin <john@phrozen.org>

Does not apply, please rebase.

error: patch failed: drivers/net/wireless/ath/ath11k/dp_tx.c:167
error: drivers/net/wireless/ath/ath11k/dp_tx.c: patch does not apply
error: patch failed: drivers/net/wireless/ath/ath11k/hal_tx.c:74
error: drivers/net/wireless/ath/ath11k/hal_tx.c: patch does not apply
stg import: Diff does not apply cleanly

3 patches set to Changes Requested.

11254761 [1/3] ath11k: move some tx_status parsing to debugfs code
11254765 [2/3] ath11k: optimise ath11k_dp_tx_completion_handler
11254767 [3/3] ath11k: optimize ath11k_hal_tx_status_parse

-- 
https://patchwork.kernel.org/patch/11254761/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches


_______________________________________________
ath11k mailing list
ath11k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath11k

^ permalink raw reply	[flat|nested] 5+ messages in thread
end of thread, other threads:[~2019-11-29  8:32 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-11-20 20:29 [PATCH 0/3] ath11k: improve TX path John Crispin
2019-11-20 20:29 ` [PATCH 1/3] ath11k: move some tx_status parsing to debugfs code John Crispin
2019-11-29  8:32   ` Kalle Valo
2019-11-20 20:29 ` [PATCH 2/3] ath11k: optimise ath11k_dp_tx_completion_handler John Crispin
2019-11-20 20:29 ` [PATCH 3/3] ath11k: optimize ath11k_hal_tx_status_parse John Crispin

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.