All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v7 0/3] ath10k: improve throughout of TX of sdio
@ 2019-10-18  7:27 ` Wen Gong
  0 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-10-18  7:27 UTC (permalink / raw)
  To: ath10k; +Cc: linux-wireless

v7: change a few code style to meet ath10k-check script

v6: remove module parameters disable_tx_comp and alt_data
this is 3 patches of the 7 patches from ath10k: improve throughout of tcp/udp TX/RX of sdio

v5: change ath10k_htc_setup_tx_req to add check bundle_tx
to forbidden init 2 times

v4: add macro ATH10K_HTC_MSG_READY_EXT_ALT_DATA_MASK

v3: change some code style

v2: change some code style

The bottleneck of throughout on sdio chip is the bus bandwidth, to the
patches are all to increase the use ratio of sdio bus.

These patches only affect sdio bus chip, explanation is mentioned in each
patch's commit log.

Wen Gong (3):
  ath10k: disable TX complete indication of htt for sdio
  ath10k: add htt TX bundle for sdio
  ath10k: enable alt data of TX path for sdio

 drivers/net/wireless/ath/ath10k/core.c   |  27 +-
 drivers/net/wireless/ath/ath10k/core.h   |   4 +-
 drivers/net/wireless/ath/ath10k/hif.h    |   9 +
 drivers/net/wireless/ath/ath10k/htc.c    | 373 +++++++++++++++++++++--
 drivers/net/wireless/ath/ath10k/htc.h    |  38 ++-
 drivers/net/wireless/ath/ath10k/htt.c    |  13 +
 drivers/net/wireless/ath/ath10k/htt.h    |  17 +-
 drivers/net/wireless/ath/ath10k/htt_rx.c |  35 ++-
 drivers/net/wireless/ath/ath10k/htt_tx.c |  38 ++-
 drivers/net/wireless/ath/ath10k/hw.h     |   2 +-
 drivers/net/wireless/ath/ath10k/sdio.c   |  23 ++
 drivers/net/wireless/ath/ath10k/sdio.h   |   2 +-
 12 files changed, 528 insertions(+), 53 deletions(-)

-- 
2.23.0


^ permalink raw reply	[flat|nested] 10+ messages in thread
* [PATCH v7 0/3] ath10k: improve throughout of TX of sdio
@ 2019-10-18  7:27 ` Wen Gong
  0 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-10-18  7:27 UTC (permalink / raw)
  To: ath10k; +Cc: linux-wireless

v7: change a few code style to meet ath10k-check script

v6: remove module parameters disable_tx_comp and alt_data
this is 3 patches of the 7 patches from ath10k: improve throughout of tcp/udp TX/RX of sdio

v5: change ath10k_htc_setup_tx_req to add check bundle_tx
to forbidden init 2 times

v4: add macro ATH10K_HTC_MSG_READY_EXT_ALT_DATA_MASK

v3: change some code style

v2: change some code style

The bottleneck of throughout on sdio chip is the bus bandwidth, to the
patches are all to increase the use ratio of sdio bus.

These patches only affect sdio bus chip, explanation is mentioned in each
patch's commit log.

Wen Gong (3):
  ath10k: disable TX complete indication of htt for sdio
  ath10k: add htt TX bundle for sdio
  ath10k: enable alt data of TX path for sdio

 drivers/net/wireless/ath/ath10k/core.c   |  27 +-
 drivers/net/wireless/ath/ath10k/core.h   |   4 +-
 drivers/net/wireless/ath/ath10k/hif.h    |   9 +
 drivers/net/wireless/ath/ath10k/htc.c    | 373 +++++++++++++++++++++--
 drivers/net/wireless/ath/ath10k/htc.h    |  38 ++-
 drivers/net/wireless/ath/ath10k/htt.c    |  13 +
 drivers/net/wireless/ath/ath10k/htt.h    |  17 +-
 drivers/net/wireless/ath/ath10k/htt_rx.c |  35 ++-
 drivers/net/wireless/ath/ath10k/htt_tx.c |  38 ++-
 drivers/net/wireless/ath/ath10k/hw.h     |   2 +-
 drivers/net/wireless/ath/ath10k/sdio.c   |  23 ++
 drivers/net/wireless/ath/ath10k/sdio.h   |   2 +-
 12 files changed, 528 insertions(+), 53 deletions(-)

-- 
2.23.0


_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

^ permalink raw reply	[flat|nested] 10+ messages in thread
* [PATCH v7 1/3] ath10k: disable TX complete indication of htt for sdio
  2019-10-18  7:27 ` Wen Gong
@ 2019-10-18  7:27   ` Wen Gong
  -1 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-10-18  7:27 UTC (permalink / raw)
  To: ath10k; +Cc: linux-wireless

For sdio chip, it is high latency bus, all the TX packet's content will
be tranferred from HOST memory to firmware memory via sdio bus, then it
need much more memory in firmware than low latency bus chip, for low
latency chip, such as PCI-E, it only need to transfer the TX descriptor
via PCI-E bus to firmware memory. For sdio chip, reduce the complexity of
TX logic will help TX efficiency since its memory is limited, and it will
reduce the TX circle's time of each packet and then firmware will have more
memory for TX since TX complete also need memeory.

This patch disable TX complete indication from firmware for htt data
packet, it will not have TX complete indication from firmware to ath10k.
It will cut the cost of bus bandwidth of TX complete and make the TX
logic of firmware simpler, it results in significant performance
improvement on TX path.

Udp TX throughout is 130Mbps without this patch, and it arrives
400Mbps with this patch.

The downside of this patch is the command "iw wlan0 station dump" will
show 0 for "tx retries" and "tx failed" since all tx packet's status
is success.

This patch only effect sdio chip, it will not effect PCI, SNOC etc.

Tested with QCA6174 SDIO with firmware
WLAN.RMH.4.4.1-00017-QCARMSWPZ-1

Signed-off-by: Wen Gong <wgong@codeaurora.org>
---
 drivers/net/wireless/ath/ath10k/core.c   |  5 +---
 drivers/net/wireless/ath/ath10k/hif.h    |  9 +++++++
 drivers/net/wireless/ath/ath10k/htc.c    | 10 +++++++
 drivers/net/wireless/ath/ath10k/htc.h    |  3 +++
 drivers/net/wireless/ath/ath10k/htt.c    |  5 ++++
 drivers/net/wireless/ath/ath10k/htt.h    | 13 ++++++++-
 drivers/net/wireless/ath/ath10k/htt_rx.c | 34 +++++++++++++++++++++++-
 drivers/net/wireless/ath/ath10k/htt_tx.c | 31 +++++++++++++++++++++
 drivers/net/wireless/ath/ath10k/hw.h     |  2 +-
 drivers/net/wireless/ath/ath10k/sdio.c   | 23 ++++++++++++++++
 drivers/net/wireless/ath/ath10k/sdio.h   |  2 +-
 11 files changed, 129 insertions(+), 8 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/core.c b/drivers/net/wireless/ath/ath10k/core.c
index 36c62d66c19e..ebaaf18f6dc7 100644
--- a/drivers/net/wireless/ath/ath10k/core.c
+++ b/drivers/net/wireless/ath/ath10k/core.c
@@ -694,10 +694,7 @@ static int ath10k_init_sdio(struct ath10k *ar, enum ath10k_firmware_mode mode)
 	if (ret)
 		return ret;
 
-	/* Data transfer is not initiated, when reduced Tx completion
-	 * is used for SDIO. disable it until fixed
-	 */
-	param &= ~HI_ACS_FLAGS_SDIO_REDUCE_TX_COMPL_SET;
+	param |= HI_ACS_FLAGS_SDIO_REDUCE_TX_COMPL_SET;
 
 	/* Alternate credit size of 1544 as used by SDIO firmware is
 	 * not big enough for mac80211 / native wifi frames. disable it
diff --git a/drivers/net/wireless/ath/ath10k/hif.h b/drivers/net/wireless/ath/ath10k/hif.h
index 496ee34a4d78..0dd8973d0acf 100644
--- a/drivers/net/wireless/ath/ath10k/hif.h
+++ b/drivers/net/wireless/ath/ath10k/hif.h
@@ -56,6 +56,8 @@ struct ath10k_hif_ops {
 
 	int (*swap_mailbox)(struct ath10k *ar);
 
+	int (*get_htt_tx_complete)(struct ath10k *ar);
+
 	int (*map_service_to_pipe)(struct ath10k *ar, u16 service_id,
 				   u8 *ul_pipe, u8 *dl_pipe);
 
@@ -144,6 +146,13 @@ static inline int ath10k_hif_swap_mailbox(struct ath10k *ar)
 	return 0;
 }
 
+static inline int ath10k_hif_get_htt_tx_complete(struct ath10k *ar)
+{
+	if (ar->hif.ops->get_htt_tx_complete)
+		return ar->hif.ops->get_htt_tx_complete(ar);
+	return 0;
+}
+
 static inline int ath10k_hif_map_service_to_pipe(struct ath10k *ar,
 						 u16 service_id,
 						 u8 *ul_pipe, u8 *dl_pipe)
diff --git a/drivers/net/wireless/ath/ath10k/htc.c b/drivers/net/wireless/ath/ath10k/htc.c
index 1d4d1a1992fe..4c6cdc24d282 100644
--- a/drivers/net/wireless/ath/ath10k/htc.c
+++ b/drivers/net/wireless/ath/ath10k/htc.c
@@ -660,6 +660,16 @@ int ath10k_htc_wait_target(struct ath10k_htc *htc)
 	return 0;
 }
 
+void ath10k_htc_change_tx_credit_flow(struct ath10k_htc *htc,
+				      enum ath10k_htc_ep_id eid,
+				      bool enable)
+{
+	struct ath10k *ar = htc->ar;
+	struct ath10k_htc_ep *ep = &ar->htc.endpoint[eid];
+
+	ep->tx_credit_flow_enabled = enable;
+}
+
 int ath10k_htc_connect_service(struct ath10k_htc *htc,
 			       struct ath10k_htc_svc_conn_req *conn_req,
 			       struct ath10k_htc_svc_conn_resp *conn_resp)
diff --git a/drivers/net/wireless/ath/ath10k/htc.h b/drivers/net/wireless/ath/ath10k/htc.h
index f55d3caec61f..d0f37cbdab8b 100644
--- a/drivers/net/wireless/ath/ath10k/htc.h
+++ b/drivers/net/wireless/ath/ath10k/htc.h
@@ -367,6 +367,9 @@ int ath10k_htc_start(struct ath10k_htc *htc);
 int ath10k_htc_connect_service(struct ath10k_htc *htc,
 			       struct ath10k_htc_svc_conn_req  *conn_req,
 			       struct ath10k_htc_svc_conn_resp *conn_resp);
+void ath10k_htc_change_tx_credit_flow(struct ath10k_htc *htc,
+				      enum ath10k_htc_ep_id eid,
+				      bool enable);
 int ath10k_htc_send(struct ath10k_htc *htc, enum ath10k_htc_ep_id eid,
 		    struct sk_buff *packet);
 struct sk_buff *ath10k_htc_alloc_skb(struct ath10k *ar, int size);
diff --git a/drivers/net/wireless/ath/ath10k/htt.c b/drivers/net/wireless/ath/ath10k/htt.c
index 7b75200ceae5..4354bf285ff1 100644
--- a/drivers/net/wireless/ath/ath10k/htt.c
+++ b/drivers/net/wireless/ath/ath10k/htt.c
@@ -10,6 +10,7 @@
 #include "htt.h"
 #include "core.h"
 #include "debug.h"
+#include "hif.h"
 
 static const enum htt_t2h_msg_type htt_main_t2h_msg_types[] = {
 	[HTT_MAIN_T2H_MSG_TYPE_VERSION_CONF] = HTT_T2H_MSG_TYPE_VERSION_CONF,
@@ -153,6 +154,10 @@ int ath10k_htt_connect(struct ath10k_htt *htt)
 
 	htt->eid = conn_resp.eid;
 
+	htt->disable_tx_comp = ath10k_hif_get_htt_tx_complete(htt->ar);
+	if (htt->disable_tx_comp)
+		ath10k_htc_change_tx_credit_flow(&htt->ar->htc, htt->eid, true);
+
 	return 0;
 }
 
diff --git a/drivers/net/wireless/ath/ath10k/htt.h b/drivers/net/wireless/ath/ath10k/htt.h
index 30c080094af1..889bf9fe051a 100644
--- a/drivers/net/wireless/ath/ath10k/htt.h
+++ b/drivers/net/wireless/ath/ath10k/htt.h
@@ -150,9 +150,19 @@ enum htt_data_tx_desc_flags1 {
 	HTT_DATA_TX_DESC_FLAGS1_MORE_IN_BATCH    = 1 << 12,
 	HTT_DATA_TX_DESC_FLAGS1_CKSUM_L3_OFFLOAD = 1 << 13,
 	HTT_DATA_TX_DESC_FLAGS1_CKSUM_L4_OFFLOAD = 1 << 14,
-	HTT_DATA_TX_DESC_FLAGS1_RSVD1            = 1 << 15
+	HTT_DATA_TX_DESC_FLAGS1_TX_COMPLETE      = 1 << 15
 };
 
+#define HTT_TX_CREDIT_DELTA_ABS_M      0xffff0000
+#define HTT_TX_CREDIT_DELTA_ABS_S      16
+#define HTT_TX_CREDIT_DELTA_ABS_GET(word) \
+	    (((word) & HTT_TX_CREDIT_DELTA_ABS_M) >> HTT_TX_CREDIT_DELTA_ABS_S)
+
+#define HTT_TX_CREDIT_SIGN_BIT_M       0x00000100
+#define HTT_TX_CREDIT_SIGN_BIT_S       8
+#define HTT_TX_CREDIT_SIGN_BIT_GET(word) \
+	    (((word) & HTT_TX_CREDIT_SIGN_BIT_M) >> HTT_TX_CREDIT_SIGN_BIT_S)
+
 enum htt_data_tx_ext_tid {
 	HTT_DATA_TX_EXT_TID_NON_QOS_MCAST_BCAST = 16,
 	HTT_DATA_TX_EXT_TID_MGMT                = 17,
@@ -2019,6 +2029,7 @@ struct ath10k_htt {
 	bool tx_mem_allocated;
 	const struct ath10k_htt_tx_ops *tx_ops;
 	const struct ath10k_htt_rx_ops *rx_ops;
+	bool disable_tx_comp;
 };
 
 struct ath10k_htt_tx_ops {
diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c
index 9f0e7b4943ec..17ad6c51e1dd 100644
--- a/drivers/net/wireless/ath/ath10k/htt_rx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
@@ -3778,6 +3778,9 @@ bool ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb)
 	}
 	case HTT_T2H_MSG_TYPE_MGMT_TX_COMPLETION: {
 		struct htt_tx_done tx_done = {};
+		struct ath10k_htt *htt = &ar->htt;
+		struct ath10k_htc *htc = &ar->htc;
+		struct ath10k_htc_ep *ep = &ar->htc.endpoint[htt->eid];
 		int status = __le32_to_cpu(resp->mgmt_tx_completion.status);
 		int info = __le32_to_cpu(resp->mgmt_tx_completion.info);
 
@@ -3803,6 +3806,12 @@ bool ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb)
 			break;
 		}
 
+		if (htt->disable_tx_comp) {
+			spin_lock_bh(&htc->tx_lock);
+			ep->tx_credits++;
+			spin_unlock_bh(&htc->tx_lock);
+		}
+
 		status = ath10k_txrx_tx_unref(htt, &tx_done);
 		if (!status) {
 			spin_lock_bh(&htt->tx_lock);
@@ -3877,8 +3886,31 @@ bool ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb)
 		skb_queue_tail(&htt->rx_in_ord_compl_q, skb);
 		return false;
 	}
-	case HTT_T2H_MSG_TYPE_TX_CREDIT_UPDATE_IND:
+	case HTT_T2H_MSG_TYPE_TX_CREDIT_UPDATE_IND: {
+		struct ath10k_htt *htt = &ar->htt;
+		struct ath10k_htc *htc = &ar->htc;
+		struct ath10k_htc_ep *ep = &ar->htc.endpoint[htt->eid];
+		u32 msg_word = __le32_to_cpu(*(__le32 *)resp);
+		int htt_credit_delta;
+
+		htt_credit_delta = HTT_TX_CREDIT_DELTA_ABS_GET(msg_word);
+		if (HTT_TX_CREDIT_SIGN_BIT_GET(msg_word))
+			htt_credit_delta = -htt_credit_delta;
+
+		ath10k_dbg(ar, ATH10K_DBG_HTT,
+			   "credit update: delta:%d\n",
+			   htt_credit_delta);
+
+		if (htt->disable_tx_comp) {
+			spin_lock_bh(&htc->tx_lock);
+			ep->tx_credits += htt_credit_delta;
+			spin_unlock_bh(&htc->tx_lock);
+			ath10k_dbg(ar, ATH10K_DBG_HTT,
+				   "credit total:%d\n",
+				   ep->tx_credits);
+		}
 		break;
+	}
 	case HTT_T2H_MSG_TYPE_CHAN_CHANGE: {
 		u32 phymode = __le32_to_cpu(resp->chan_change.phymode);
 		u32 freq = __le32_to_cpu(resp->chan_change.freq);
diff --git a/drivers/net/wireless/ath/ath10k/htt_tx.c b/drivers/net/wireless/ath/ath10k/htt_tx.c
index a182c0944cc7..c6c4b2a4d20f 100644
--- a/drivers/net/wireless/ath/ath10k/htt_tx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_tx.c
@@ -543,7 +543,35 @@ void ath10k_htt_tx_free(struct ath10k_htt *htt)
 
 void ath10k_htt_htc_tx_complete(struct ath10k *ar, struct sk_buff *skb)
 {
+	struct ath10k_htt *htt = &ar->htt;
+	struct htt_tx_done tx_done = {0};
+	struct htt_cmd_hdr *htt_hdr;
+	struct htt_data_tx_desc *desc_hdr;
+	u16 flags1;
+
 	dev_kfree_skb_any(skb);
+
+	if (!htt->disable_tx_comp)
+		return;
+
+	htt_hdr = (struct htt_cmd_hdr *)skb->data;
+	if (htt_hdr->msg_type != HTT_H2T_MSG_TYPE_TX_FRM)
+		return;
+
+	desc_hdr = (struct htt_data_tx_desc *)
+		(skb->data + sizeof(*htt_hdr));
+	flags1 = __le16_to_cpu(desc_hdr->flags1);
+
+	ath10k_dbg(ar, ATH10K_DBG_HTT,
+		   "htt tx complete msdu id:%u ,flags1:%x\n",
+		   __le16_to_cpu(desc_hdr->id), flags1);
+
+	if (flags1 & HTT_DATA_TX_DESC_FLAGS1_TX_COMPLETE)
+		return;
+
+	tx_done.status = HTT_TX_COMPL_STATE_ACK;
+	tx_done.msdu_id = __le16_to_cpu(desc_hdr->id);
+	ath10k_txrx_tx_unref(&ar->htt, &tx_done);
 }
 
 void ath10k_htt_hif_tx_complete(struct ath10k *ar, struct sk_buff *skb)
@@ -1269,6 +1297,9 @@ static int ath10k_htt_tx_hl(struct ath10k_htt *htt, enum ath10k_hw_txrx_mode txm
 		flags0 |= SM(ATH10K_HW_TXRX_MGMT,
 			     HTT_DATA_TX_DESC_FLAGS0_PKT_TYPE);
 		flags0 |= HTT_DATA_TX_DESC_FLAGS0_MAC_HDR_PRESENT;
+
+		if (htt->disable_tx_comp)
+			flags1 |= HTT_DATA_TX_DESC_FLAGS1_TX_COMPLETE;
 		break;
 	}
 
diff --git a/drivers/net/wireless/ath/ath10k/hw.h b/drivers/net/wireless/ath/ath10k/hw.h
index 2ae57c1de7b5..6349665a9b34 100644
--- a/drivers/net/wireless/ath/ath10k/hw.h
+++ b/drivers/net/wireless/ath/ath10k/hw.h
@@ -759,7 +759,7 @@ ath10k_is_rssi_enable(struct ath10k_hw_params *hw,
 #define TARGET_TLV_NUM_TDLS_VDEVS		1
 #define TARGET_TLV_NUM_TIDS			((TARGET_TLV_NUM_PEERS) * 2)
 #define TARGET_TLV_NUM_MSDU_DESC		(1024 + 32)
-#define TARGET_TLV_NUM_MSDU_DESC_HL		64
+#define TARGET_TLV_NUM_MSDU_DESC_HL		1024
 #define TARGET_TLV_NUM_WOW_PATTERNS		22
 #define TARGET_TLV_MGMT_NUM_MSDU_DESC		(50)
 
diff --git a/drivers/net/wireless/ath/ath10k/sdio.c b/drivers/net/wireless/ath/ath10k/sdio.c
index 120200a93bcc..84910446df50 100644
--- a/drivers/net/wireless/ath/ath10k/sdio.c
+++ b/drivers/net/wireless/ath/ath10k/sdio.c
@@ -1674,6 +1674,28 @@ static int ath10k_sdio_hif_swap_mailbox(struct ath10k *ar)
 	return 0;
 }
 
+static int ath10k_sdio_get_htt_tx_complete(struct ath10k *ar)
+{
+	u32 addr, val;
+	int ret;
+
+	addr = host_interest_item_address(HI_ITEM(hi_acs_flags));
+
+	ret = ath10k_sdio_hif_diag_read32(ar, addr, &val);
+	if (ret) {
+		ath10k_warn(ar,
+			    "unable to read hi_acs_flags for htt tx comple : %d\n", ret);
+		return ret;
+	}
+
+	ret = (val & HI_ACS_FLAGS_SDIO_REDUCE_TX_COMPL_FW_ACK);
+
+	ath10k_dbg(ar, ATH10K_DBG_SDIO, "sdio reduce tx complete fw%sack\n",
+		   ret ? " " : " not ");
+
+	return ret;
+}
+
 /* HIF start/stop */
 
 static int ath10k_sdio_hif_start(struct ath10k *ar)
@@ -1943,6 +1965,7 @@ static const struct ath10k_hif_ops ath10k_sdio_hif_ops = {
 	.start			= ath10k_sdio_hif_start,
 	.stop			= ath10k_sdio_hif_stop,
 	.swap_mailbox		= ath10k_sdio_hif_swap_mailbox,
+	.get_htt_tx_complete	= ath10k_sdio_get_htt_tx_complete,
 	.map_service_to_pipe	= ath10k_sdio_hif_map_service_to_pipe,
 	.get_default_pipe	= ath10k_sdio_hif_get_default_pipe,
 	.send_complete_check	= ath10k_sdio_hif_send_complete_check,
diff --git a/drivers/net/wireless/ath/ath10k/sdio.h b/drivers/net/wireless/ath/ath10k/sdio.h
index b8c7ac0330bd..96a1c8e62257 100644
--- a/drivers/net/wireless/ath/ath10k/sdio.h
+++ b/drivers/net/wireless/ath/ath10k/sdio.h
@@ -37,7 +37,7 @@
 	(ATH10K_SDIO_MAX_BUFFER_SIZE - sizeof(struct ath10k_htc_hdr))
 
 #define ATH10K_HIF_MBOX_NUM_MAX                 4
-#define ATH10K_SDIO_BUS_REQUEST_MAX_NUM         64
+#define ATH10K_SDIO_BUS_REQUEST_MAX_NUM         1024
 
 #define ATH10K_SDIO_HIF_COMMUNICATION_TIMEOUT_HZ (100 * HZ)
 
-- 
2.23.0


^ permalink raw reply related	[flat|nested] 10+ messages in thread
* [PATCH v7 1/3] ath10k: disable TX complete indication of htt for sdio
@ 2019-10-18  7:27   ` Wen Gong
  0 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-10-18  7:27 UTC (permalink / raw)
  To: ath10k; +Cc: linux-wireless

For sdio chip, it is high latency bus, all the TX packet's content will
be tranferred from HOST memory to firmware memory via sdio bus, then it
need much more memory in firmware than low latency bus chip, for low
latency chip, such as PCI-E, it only need to transfer the TX descriptor
via PCI-E bus to firmware memory. For sdio chip, reduce the complexity of
TX logic will help TX efficiency since its memory is limited, and it will
reduce the TX circle's time of each packet and then firmware will have more
memory for TX since TX complete also need memeory.

This patch disable TX complete indication from firmware for htt data
packet, it will not have TX complete indication from firmware to ath10k.
It will cut the cost of bus bandwidth of TX complete and make the TX
logic of firmware simpler, it results in significant performance
improvement on TX path.

Udp TX throughout is 130Mbps without this patch, and it arrives
400Mbps with this patch.

The downside of this patch is the command "iw wlan0 station dump" will
show 0 for "tx retries" and "tx failed" since all tx packet's status
is success.

This patch only effect sdio chip, it will not effect PCI, SNOC etc.

Tested with QCA6174 SDIO with firmware
WLAN.RMH.4.4.1-00017-QCARMSWPZ-1

Signed-off-by: Wen Gong <wgong@codeaurora.org>
---
 drivers/net/wireless/ath/ath10k/core.c   |  5 +---
 drivers/net/wireless/ath/ath10k/hif.h    |  9 +++++++
 drivers/net/wireless/ath/ath10k/htc.c    | 10 +++++++
 drivers/net/wireless/ath/ath10k/htc.h    |  3 +++
 drivers/net/wireless/ath/ath10k/htt.c    |  5 ++++
 drivers/net/wireless/ath/ath10k/htt.h    | 13 ++++++++-
 drivers/net/wireless/ath/ath10k/htt_rx.c | 34 +++++++++++++++++++++++-
 drivers/net/wireless/ath/ath10k/htt_tx.c | 31 +++++++++++++++++++++
 drivers/net/wireless/ath/ath10k/hw.h     |  2 +-
 drivers/net/wireless/ath/ath10k/sdio.c   | 23 ++++++++++++++++
 drivers/net/wireless/ath/ath10k/sdio.h   |  2 +-
 11 files changed, 129 insertions(+), 8 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/core.c b/drivers/net/wireless/ath/ath10k/core.c
index 36c62d66c19e..ebaaf18f6dc7 100644
--- a/drivers/net/wireless/ath/ath10k/core.c
+++ b/drivers/net/wireless/ath/ath10k/core.c
@@ -694,10 +694,7 @@ static int ath10k_init_sdio(struct ath10k *ar, enum ath10k_firmware_mode mode)
 	if (ret)
 		return ret;
 
-	/* Data transfer is not initiated, when reduced Tx completion
-	 * is used for SDIO. disable it until fixed
-	 */
-	param &= ~HI_ACS_FLAGS_SDIO_REDUCE_TX_COMPL_SET;
+	param |= HI_ACS_FLAGS_SDIO_REDUCE_TX_COMPL_SET;
 
 	/* Alternate credit size of 1544 as used by SDIO firmware is
 	 * not big enough for mac80211 / native wifi frames. disable it
diff --git a/drivers/net/wireless/ath/ath10k/hif.h b/drivers/net/wireless/ath/ath10k/hif.h
index 496ee34a4d78..0dd8973d0acf 100644
--- a/drivers/net/wireless/ath/ath10k/hif.h
+++ b/drivers/net/wireless/ath/ath10k/hif.h
@@ -56,6 +56,8 @@ struct ath10k_hif_ops {
 
 	int (*swap_mailbox)(struct ath10k *ar);
 
+	int (*get_htt_tx_complete)(struct ath10k *ar);
+
 	int (*map_service_to_pipe)(struct ath10k *ar, u16 service_id,
 				   u8 *ul_pipe, u8 *dl_pipe);
 
@@ -144,6 +146,13 @@ static inline int ath10k_hif_swap_mailbox(struct ath10k *ar)
 	return 0;
 }
 
+static inline int ath10k_hif_get_htt_tx_complete(struct ath10k *ar)
+{
+	if (ar->hif.ops->get_htt_tx_complete)
+		return ar->hif.ops->get_htt_tx_complete(ar);
+	return 0;
+}
+
 static inline int ath10k_hif_map_service_to_pipe(struct ath10k *ar,
 						 u16 service_id,
 						 u8 *ul_pipe, u8 *dl_pipe)
diff --git a/drivers/net/wireless/ath/ath10k/htc.c b/drivers/net/wireless/ath/ath10k/htc.c
index 1d4d1a1992fe..4c6cdc24d282 100644
--- a/drivers/net/wireless/ath/ath10k/htc.c
+++ b/drivers/net/wireless/ath/ath10k/htc.c
@@ -660,6 +660,16 @@ int ath10k_htc_wait_target(struct ath10k_htc *htc)
 	return 0;
 }
 
+void ath10k_htc_change_tx_credit_flow(struct ath10k_htc *htc,
+				      enum ath10k_htc_ep_id eid,
+				      bool enable)
+{
+	struct ath10k *ar = htc->ar;
+	struct ath10k_htc_ep *ep = &ar->htc.endpoint[eid];
+
+	ep->tx_credit_flow_enabled = enable;
+}
+
 int ath10k_htc_connect_service(struct ath10k_htc *htc,
 			       struct ath10k_htc_svc_conn_req *conn_req,
 			       struct ath10k_htc_svc_conn_resp *conn_resp)
diff --git a/drivers/net/wireless/ath/ath10k/htc.h b/drivers/net/wireless/ath/ath10k/htc.h
index f55d3caec61f..d0f37cbdab8b 100644
--- a/drivers/net/wireless/ath/ath10k/htc.h
+++ b/drivers/net/wireless/ath/ath10k/htc.h
@@ -367,6 +367,9 @@ int ath10k_htc_start(struct ath10k_htc *htc);
 int ath10k_htc_connect_service(struct ath10k_htc *htc,
 			       struct ath10k_htc_svc_conn_req  *conn_req,
 			       struct ath10k_htc_svc_conn_resp *conn_resp);
+void ath10k_htc_change_tx_credit_flow(struct ath10k_htc *htc,
+				      enum ath10k_htc_ep_id eid,
+				      bool enable);
 int ath10k_htc_send(struct ath10k_htc *htc, enum ath10k_htc_ep_id eid,
 		    struct sk_buff *packet);
 struct sk_buff *ath10k_htc_alloc_skb(struct ath10k *ar, int size);
diff --git a/drivers/net/wireless/ath/ath10k/htt.c b/drivers/net/wireless/ath/ath10k/htt.c
index 7b75200ceae5..4354bf285ff1 100644
--- a/drivers/net/wireless/ath/ath10k/htt.c
+++ b/drivers/net/wireless/ath/ath10k/htt.c
@@ -10,6 +10,7 @@
 #include "htt.h"
 #include "core.h"
 #include "debug.h"
+#include "hif.h"
 
 static const enum htt_t2h_msg_type htt_main_t2h_msg_types[] = {
 	[HTT_MAIN_T2H_MSG_TYPE_VERSION_CONF] = HTT_T2H_MSG_TYPE_VERSION_CONF,
@@ -153,6 +154,10 @@ int ath10k_htt_connect(struct ath10k_htt *htt)
 
 	htt->eid = conn_resp.eid;
 
+	htt->disable_tx_comp = ath10k_hif_get_htt_tx_complete(htt->ar);
+	if (htt->disable_tx_comp)
+		ath10k_htc_change_tx_credit_flow(&htt->ar->htc, htt->eid, true);
+
 	return 0;
 }
 
diff --git a/drivers/net/wireless/ath/ath10k/htt.h b/drivers/net/wireless/ath/ath10k/htt.h
index 30c080094af1..889bf9fe051a 100644
--- a/drivers/net/wireless/ath/ath10k/htt.h
+++ b/drivers/net/wireless/ath/ath10k/htt.h
@@ -150,9 +150,19 @@ enum htt_data_tx_desc_flags1 {
 	HTT_DATA_TX_DESC_FLAGS1_MORE_IN_BATCH    = 1 << 12,
 	HTT_DATA_TX_DESC_FLAGS1_CKSUM_L3_OFFLOAD = 1 << 13,
 	HTT_DATA_TX_DESC_FLAGS1_CKSUM_L4_OFFLOAD = 1 << 14,
-	HTT_DATA_TX_DESC_FLAGS1_RSVD1            = 1 << 15
+	HTT_DATA_TX_DESC_FLAGS1_TX_COMPLETE      = 1 << 15
 };
 
+#define HTT_TX_CREDIT_DELTA_ABS_M      0xffff0000
+#define HTT_TX_CREDIT_DELTA_ABS_S      16
+#define HTT_TX_CREDIT_DELTA_ABS_GET(word) \
+	    (((word) & HTT_TX_CREDIT_DELTA_ABS_M) >> HTT_TX_CREDIT_DELTA_ABS_S)
+
+#define HTT_TX_CREDIT_SIGN_BIT_M       0x00000100
+#define HTT_TX_CREDIT_SIGN_BIT_S       8
+#define HTT_TX_CREDIT_SIGN_BIT_GET(word) \
+	    (((word) & HTT_TX_CREDIT_SIGN_BIT_M) >> HTT_TX_CREDIT_SIGN_BIT_S)
+
 enum htt_data_tx_ext_tid {
 	HTT_DATA_TX_EXT_TID_NON_QOS_MCAST_BCAST = 16,
 	HTT_DATA_TX_EXT_TID_MGMT                = 17,
@@ -2019,6 +2029,7 @@ struct ath10k_htt {
 	bool tx_mem_allocated;
 	const struct ath10k_htt_tx_ops *tx_ops;
 	const struct ath10k_htt_rx_ops *rx_ops;
+	bool disable_tx_comp;
 };
 
 struct ath10k_htt_tx_ops {
diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c
index 9f0e7b4943ec..17ad6c51e1dd 100644
--- a/drivers/net/wireless/ath/ath10k/htt_rx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
@@ -3778,6 +3778,9 @@ bool ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb)
 	}
 	case HTT_T2H_MSG_TYPE_MGMT_TX_COMPLETION: {
 		struct htt_tx_done tx_done = {};
+		struct ath10k_htt *htt = &ar->htt;
+		struct ath10k_htc *htc = &ar->htc;
+		struct ath10k_htc_ep *ep = &ar->htc.endpoint[htt->eid];
 		int status = __le32_to_cpu(resp->mgmt_tx_completion.status);
 		int info = __le32_to_cpu(resp->mgmt_tx_completion.info);
 
@@ -3803,6 +3806,12 @@ bool ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb)
 			break;
 		}
 
+		if (htt->disable_tx_comp) {
+			spin_lock_bh(&htc->tx_lock);
+			ep->tx_credits++;
+			spin_unlock_bh(&htc->tx_lock);
+		}
+
 		status = ath10k_txrx_tx_unref(htt, &tx_done);
 		if (!status) {
 			spin_lock_bh(&htt->tx_lock);
@@ -3877,8 +3886,31 @@ bool ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb)
 		skb_queue_tail(&htt->rx_in_ord_compl_q, skb);
 		return false;
 	}
-	case HTT_T2H_MSG_TYPE_TX_CREDIT_UPDATE_IND:
+	case HTT_T2H_MSG_TYPE_TX_CREDIT_UPDATE_IND: {
+		struct ath10k_htt *htt = &ar->htt;
+		struct ath10k_htc *htc = &ar->htc;
+		struct ath10k_htc_ep *ep = &ar->htc.endpoint[htt->eid];
+		u32 msg_word = __le32_to_cpu(*(__le32 *)resp);
+		int htt_credit_delta;
+
+		htt_credit_delta = HTT_TX_CREDIT_DELTA_ABS_GET(msg_word);
+		if (HTT_TX_CREDIT_SIGN_BIT_GET(msg_word))
+			htt_credit_delta = -htt_credit_delta;
+
+		ath10k_dbg(ar, ATH10K_DBG_HTT,
+			   "credit update: delta:%d\n",
+			   htt_credit_delta);
+
+		if (htt->disable_tx_comp) {
+			spin_lock_bh(&htc->tx_lock);
+			ep->tx_credits += htt_credit_delta;
+			spin_unlock_bh(&htc->tx_lock);
+			ath10k_dbg(ar, ATH10K_DBG_HTT,
+				   "credit total:%d\n",
+				   ep->tx_credits);
+		}
 		break;
+	}
 	case HTT_T2H_MSG_TYPE_CHAN_CHANGE: {
 		u32 phymode = __le32_to_cpu(resp->chan_change.phymode);
 		u32 freq = __le32_to_cpu(resp->chan_change.freq);
diff --git a/drivers/net/wireless/ath/ath10k/htt_tx.c b/drivers/net/wireless/ath/ath10k/htt_tx.c
index a182c0944cc7..c6c4b2a4d20f 100644
--- a/drivers/net/wireless/ath/ath10k/htt_tx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_tx.c
@@ -543,7 +543,35 @@ void ath10k_htt_tx_free(struct ath10k_htt *htt)
 
 void ath10k_htt_htc_tx_complete(struct ath10k *ar, struct sk_buff *skb)
 {
+	struct ath10k_htt *htt = &ar->htt;
+	struct htt_tx_done tx_done = {0};
+	struct htt_cmd_hdr *htt_hdr;
+	struct htt_data_tx_desc *desc_hdr;
+	u16 flags1;
+
 	dev_kfree_skb_any(skb);
+
+	if (!htt->disable_tx_comp)
+		return;
+
+	htt_hdr = (struct htt_cmd_hdr *)skb->data;
+	if (htt_hdr->msg_type != HTT_H2T_MSG_TYPE_TX_FRM)
+		return;
+
+	desc_hdr = (struct htt_data_tx_desc *)
+		(skb->data + sizeof(*htt_hdr));
+	flags1 = __le16_to_cpu(desc_hdr->flags1);
+
+	ath10k_dbg(ar, ATH10K_DBG_HTT,
+		   "htt tx complete msdu id:%u ,flags1:%x\n",
+		   __le16_to_cpu(desc_hdr->id), flags1);
+
+	if (flags1 & HTT_DATA_TX_DESC_FLAGS1_TX_COMPLETE)
+		return;
+
+	tx_done.status = HTT_TX_COMPL_STATE_ACK;
+	tx_done.msdu_id = __le16_to_cpu(desc_hdr->id);
+	ath10k_txrx_tx_unref(&ar->htt, &tx_done);
 }
 
 void ath10k_htt_hif_tx_complete(struct ath10k *ar, struct sk_buff *skb)
@@ -1269,6 +1297,9 @@ static int ath10k_htt_tx_hl(struct ath10k_htt *htt, enum ath10k_hw_txrx_mode txm
 		flags0 |= SM(ATH10K_HW_TXRX_MGMT,
 			     HTT_DATA_TX_DESC_FLAGS0_PKT_TYPE);
 		flags0 |= HTT_DATA_TX_DESC_FLAGS0_MAC_HDR_PRESENT;
+
+		if (htt->disable_tx_comp)
+			flags1 |= HTT_DATA_TX_DESC_FLAGS1_TX_COMPLETE;
 		break;
 	}
 
diff --git a/drivers/net/wireless/ath/ath10k/hw.h b/drivers/net/wireless/ath/ath10k/hw.h
index 2ae57c1de7b5..6349665a9b34 100644
--- a/drivers/net/wireless/ath/ath10k/hw.h
+++ b/drivers/net/wireless/ath/ath10k/hw.h
@@ -759,7 +759,7 @@ ath10k_is_rssi_enable(struct ath10k_hw_params *hw,
 #define TARGET_TLV_NUM_TDLS_VDEVS		1
 #define TARGET_TLV_NUM_TIDS			((TARGET_TLV_NUM_PEERS) * 2)
 #define TARGET_TLV_NUM_MSDU_DESC		(1024 + 32)
-#define TARGET_TLV_NUM_MSDU_DESC_HL		64
+#define TARGET_TLV_NUM_MSDU_DESC_HL		1024
 #define TARGET_TLV_NUM_WOW_PATTERNS		22
 #define TARGET_TLV_MGMT_NUM_MSDU_DESC		(50)
 
diff --git a/drivers/net/wireless/ath/ath10k/sdio.c b/drivers/net/wireless/ath/ath10k/sdio.c
index 120200a93bcc..84910446df50 100644
--- a/drivers/net/wireless/ath/ath10k/sdio.c
+++ b/drivers/net/wireless/ath/ath10k/sdio.c
@@ -1674,6 +1674,28 @@ static int ath10k_sdio_hif_swap_mailbox(struct ath10k *ar)
 	return 0;
 }
 
+static int ath10k_sdio_get_htt_tx_complete(struct ath10k *ar)
+{
+	u32 addr, val;
+	int ret;
+
+	addr = host_interest_item_address(HI_ITEM(hi_acs_flags));
+
+	ret = ath10k_sdio_hif_diag_read32(ar, addr, &val);
+	if (ret) {
+		ath10k_warn(ar,
+			    "unable to read hi_acs_flags for htt tx comple : %d\n", ret);
+		return ret;
+	}
+
+	ret = (val & HI_ACS_FLAGS_SDIO_REDUCE_TX_COMPL_FW_ACK);
+
+	ath10k_dbg(ar, ATH10K_DBG_SDIO, "sdio reduce tx complete fw%sack\n",
+		   ret ? " " : " not ");
+
+	return ret;
+}
+
 /* HIF start/stop */
 
 static int ath10k_sdio_hif_start(struct ath10k *ar)
@@ -1943,6 +1965,7 @@ static const struct ath10k_hif_ops ath10k_sdio_hif_ops = {
 	.start			= ath10k_sdio_hif_start,
 	.stop			= ath10k_sdio_hif_stop,
 	.swap_mailbox		= ath10k_sdio_hif_swap_mailbox,
+	.get_htt_tx_complete	= ath10k_sdio_get_htt_tx_complete,
 	.map_service_to_pipe	= ath10k_sdio_hif_map_service_to_pipe,
 	.get_default_pipe	= ath10k_sdio_hif_get_default_pipe,
 	.send_complete_check	= ath10k_sdio_hif_send_complete_check,
diff --git a/drivers/net/wireless/ath/ath10k/sdio.h b/drivers/net/wireless/ath/ath10k/sdio.h
index b8c7ac0330bd..96a1c8e62257 100644
--- a/drivers/net/wireless/ath/ath10k/sdio.h
+++ b/drivers/net/wireless/ath/ath10k/sdio.h
@@ -37,7 +37,7 @@
 	(ATH10K_SDIO_MAX_BUFFER_SIZE - sizeof(struct ath10k_htc_hdr))
 
 #define ATH10K_HIF_MBOX_NUM_MAX                 4
-#define ATH10K_SDIO_BUS_REQUEST_MAX_NUM         64
+#define ATH10K_SDIO_BUS_REQUEST_MAX_NUM         1024
 
 #define ATH10K_SDIO_HIF_COMMUNICATION_TIMEOUT_HZ (100 * HZ)
 
-- 
2.23.0


_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

^ permalink raw reply related	[flat|nested] 10+ messages in thread
* [PATCH v7 2/3] ath10k: add htt TX bundle for sdio
  2019-10-18  7:27 ` Wen Gong
@ 2019-10-18  7:27   ` Wen Gong
  -1 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-10-18  7:27 UTC (permalink / raw)
  To: ath10k; +Cc: linux-wireless

The transmission utilization ratio for sdio bus for small packet is
slow, because the space and time cost for sdio bus is same for large
length packet and small length packet. So the speed of data for large
length packet is higher than small length.

Test result of different length of data:

data packet(byte)   cost time(us)   calculated rate(Mbps)
      256               28                73
      512               33               124
     1024               35               234
     1792               45               318
    14336              168               682
    28672              333               688
    57344              660               695

This patch change the TX packet from single packet to a large length
bundle packet, max size is 32, it results in significant performance
improvement on TX path.

This patch only effect sdio chip, it will not effect PCI, SNOC etc.
It only enable bundle for sdio chip.

Tested with QCA6174 SDIO with firmware
WLAN.RMH.4.4.1-00017-QCARMSWP-1.

Signed-off-by: Wen Gong <wgong@codeaurora.org>
---
 drivers/net/wireless/ath/ath10k/core.c   |  14 +-
 drivers/net/wireless/ath/ath10k/core.h   |   4 +-
 drivers/net/wireless/ath/ath10k/htc.c    | 351 +++++++++++++++++++++--
 drivers/net/wireless/ath/ath10k/htc.h    |  22 +-
 drivers/net/wireless/ath/ath10k/htt.c    |   8 +
 drivers/net/wireless/ath/ath10k/htt.h    |   4 +
 drivers/net/wireless/ath/ath10k/htt_rx.c |   1 +
 drivers/net/wireless/ath/ath10k/htt_tx.c |   7 +-
 8 files changed, 374 insertions(+), 37 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/core.c b/drivers/net/wireless/ath/ath10k/core.c
index ebaaf18f6dc7..d9f1d5e75c86 100644
--- a/drivers/net/wireless/ath/ath10k/core.c
+++ b/drivers/net/wireless/ath/ath10k/core.c
@@ -3211,6 +3211,11 @@ struct ath10k *ath10k_core_create(size_t priv_size, struct device *dev,
 	if (!ar->workqueue_aux)
 		goto err_free_wq;
 
+	ar->workqueue_tx_complete =
+		create_singlethread_workqueue("ath10k_tx_complete_wq");
+	if (!ar->workqueue_tx_complete)
+		goto err_free_aux_wq;
+
 	mutex_init(&ar->conf_mutex);
 	mutex_init(&ar->dump_mutex);
 	spin_lock_init(&ar->data_lock);
@@ -3236,7 +3241,7 @@ struct ath10k *ath10k_core_create(size_t priv_size, struct device *dev,
 
 	ret = ath10k_coredump_create(ar);
 	if (ret)
-		goto err_free_aux_wq;
+		goto err_free_tx_complete;
 
 	ret = ath10k_debug_create(ar);
 	if (ret)
@@ -3246,12 +3251,12 @@ struct ath10k *ath10k_core_create(size_t priv_size, struct device *dev,
 
 err_free_coredump:
 	ath10k_coredump_destroy(ar);
-
+err_free_tx_complete:
+	destroy_workqueue(ar->workqueue_tx_complete);
 err_free_aux_wq:
 	destroy_workqueue(ar->workqueue_aux);
 err_free_wq:
 	destroy_workqueue(ar->workqueue);
-
 err_free_mac:
 	ath10k_mac_destroy(ar);
 
@@ -3267,6 +3272,9 @@ void ath10k_core_destroy(struct ath10k *ar)
 	flush_workqueue(ar->workqueue_aux);
 	destroy_workqueue(ar->workqueue_aux);
 
+	flush_workqueue(ar->workqueue_tx_complete);
+	destroy_workqueue(ar->workqueue_tx_complete);
+
 	ath10k_debug_destroy(ar);
 	ath10k_coredump_destroy(ar);
 	ath10k_htt_tx_destroy(&ar->htt);
diff --git a/drivers/net/wireless/ath/ath10k/core.h b/drivers/net/wireless/ath/ath10k/core.h
index c96d1a17fd00..0eebf3564aa4 100644
--- a/drivers/net/wireless/ath/ath10k/core.h
+++ b/drivers/net/wireless/ath/ath10k/core.h
@@ -1080,7 +1080,7 @@ struct ath10k {
 	struct workqueue_struct *workqueue;
 	/* Auxiliary workqueue */
 	struct workqueue_struct *workqueue_aux;
-
+	struct workqueue_struct *workqueue_tx_complete;
 	/* prevents concurrent FW reconfiguration */
 	struct mutex conf_mutex;
 
@@ -1121,6 +1121,8 @@ struct ath10k {
 
 	struct work_struct register_work;
 	struct work_struct restart_work;
+	struct work_struct bundle_tx_work;
+	struct work_struct tx_complete_work;
 
 	/* cycle count is reported twice for each visited channel during scan.
 	 * access protected by data_lock
diff --git a/drivers/net/wireless/ath/ath10k/htc.c b/drivers/net/wireless/ath/ath10k/htc.c
index 4c6cdc24d282..a2bf34e5a96e 100644
--- a/drivers/net/wireless/ath/ath10k/htc.c
+++ b/drivers/net/wireless/ath/ath10k/htc.c
@@ -51,10 +51,12 @@ void ath10k_htc_notify_tx_completion(struct ath10k_htc_ep *ep,
 				     struct sk_buff *skb)
 {
 	struct ath10k *ar = ep->htc->ar;
+	struct ath10k_htc_hdr *hdr;
 
 	ath10k_dbg(ar, ATH10K_DBG_HTC, "%s: ep %d skb %pK\n", __func__,
 		   ep->eid, skb);
 
+	hdr = (struct ath10k_htc_hdr *)skb->data;
 	ath10k_htc_restore_tx_skb(ep->htc, skb);
 
 	if (!ep->ep_ops.ep_tx_complete) {
@@ -63,6 +65,11 @@ void ath10k_htc_notify_tx_completion(struct ath10k_htc_ep *ep,
 		return;
 	}
 
+	if (hdr->flags & ATH10K_HTC_FLAG_SEND_BUNDLE) {
+		dev_kfree_skb_any(skb);
+		return;
+	}
+
 	ep->ep_ops.ep_tx_complete(ep->htc->ar, skb);
 }
 EXPORT_SYMBOL(ath10k_htc_notify_tx_completion);
@@ -78,7 +85,7 @@ static void ath10k_htc_prepare_tx_skb(struct ath10k_htc_ep *ep,
 	hdr->eid = ep->eid;
 	hdr->len = __cpu_to_le16(skb->len - sizeof(*hdr));
 	hdr->flags = 0;
-	if (ep->tx_credit_flow_enabled)
+	if (ep->tx_credit_flow_enabled && !ep->bundle_tx)
 		hdr->flags |= ATH10K_HTC_FLAG_NEED_CREDIT_UPDATE;
 
 	spin_lock_bh(&ep->htc->tx_lock);
@@ -86,6 +93,63 @@ static void ath10k_htc_prepare_tx_skb(struct ath10k_htc_ep *ep,
 	spin_unlock_bh(&ep->htc->tx_lock);
 }
 
+static int ath10k_htc_consume_credit(struct ath10k_htc_ep *ep,
+				     unsigned int len,
+				     bool consume)
+{
+	struct ath10k_htc *htc = ep->htc;
+	struct ath10k *ar = htc->ar;
+	enum ath10k_htc_ep_id eid = ep->eid;
+	int credits, ret = 0;
+
+	if (!ep->tx_credit_flow_enabled)
+		return 0;
+
+	credits = DIV_ROUND_UP(len, ep->tx_credit_size);
+	spin_lock_bh(&htc->tx_lock);
+
+	if (ep->tx_credits < credits) {
+		ath10k_dbg(ar, ATH10K_DBG_HTC,
+			   "htc insufficient credits ep %d required %d available %d consume %d\n",
+			   eid, credits, ep->tx_credits, consume);
+		ret = -EAGAIN;
+		goto unlock;
+	}
+
+	if (consume) {
+		ep->tx_credits -= credits;
+		ath10k_dbg(ar, ATH10K_DBG_HTC,
+			   "htc ep %d consumed %d credits (total %d)\n",
+			   eid, credits, ep->tx_credits);
+	}
+
+unlock:
+	spin_unlock_bh(&htc->tx_lock);
+	return ret;
+}
+
+static void ath10k_htc_release_credit(struct ath10k_htc_ep *ep, unsigned int len)
+{
+	struct ath10k_htc *htc = ep->htc;
+	struct ath10k *ar = htc->ar;
+	enum ath10k_htc_ep_id eid = ep->eid;
+	int credits;
+
+	if (!ep->tx_credit_flow_enabled)
+		return;
+
+	credits = DIV_ROUND_UP(len, ep->tx_credit_size);
+	spin_lock_bh(&htc->tx_lock);
+	ep->tx_credits += credits;
+	ath10k_dbg(ar, ATH10K_DBG_HTC,
+		   "htc ep %d reverted %d credits back (total %d)\n",
+		   eid, credits, ep->tx_credits);
+	spin_unlock_bh(&htc->tx_lock);
+
+	if (ep->ep_ops.ep_tx_credits)
+		ep->ep_ops.ep_tx_credits(htc->ar);
+}
+
 int ath10k_htc_send(struct ath10k_htc *htc,
 		    enum ath10k_htc_ep_id eid,
 		    struct sk_buff *skb)
@@ -95,8 +159,8 @@ int ath10k_htc_send(struct ath10k_htc *htc,
 	struct ath10k_skb_cb *skb_cb = ATH10K_SKB_CB(skb);
 	struct ath10k_hif_sg_item sg_item;
 	struct device *dev = htc->ar->dev;
-	int credits = 0;
 	int ret;
+	unsigned int skb_len;
 
 	if (htc->ar->state == ATH10K_STATE_WEDGED)
 		return -ECOMM;
@@ -108,23 +172,10 @@ int ath10k_htc_send(struct ath10k_htc *htc,
 
 	skb_push(skb, sizeof(struct ath10k_htc_hdr));
 
-	if (ep->tx_credit_flow_enabled) {
-		credits = DIV_ROUND_UP(skb->len, htc->target_credit_size);
-		spin_lock_bh(&htc->tx_lock);
-		if (ep->tx_credits < credits) {
-			ath10k_dbg(ar, ATH10K_DBG_HTC,
-				   "htc insufficient credits ep %d required %d available %d\n",
-				   eid, credits, ep->tx_credits);
-			spin_unlock_bh(&htc->tx_lock);
-			ret = -EAGAIN;
-			goto err_pull;
-		}
-		ep->tx_credits -= credits;
-		ath10k_dbg(ar, ATH10K_DBG_HTC,
-			   "htc ep %d consumed %d credits (total %d)\n",
-			   eid, credits, ep->tx_credits);
-		spin_unlock_bh(&htc->tx_lock);
-	}
+	skb_len = skb->len;
+	ret = ath10k_htc_consume_credit(ep, skb_len, true);
+	if (ret)
+		goto err_pull;
 
 	ath10k_htc_prepare_tx_skb(ep, skb);
 
@@ -155,17 +206,7 @@ int ath10k_htc_send(struct ath10k_htc *htc,
 	if (ar->bus_param.dev_type != ATH10K_DEV_TYPE_HL)
 		dma_unmap_single(dev, skb_cb->paddr, skb->len, DMA_TO_DEVICE);
 err_credits:
-	if (ep->tx_credit_flow_enabled) {
-		spin_lock_bh(&htc->tx_lock);
-		ep->tx_credits += credits;
-		ath10k_dbg(ar, ATH10K_DBG_HTC,
-			   "htc ep %d reverted %d credits back (total %d)\n",
-			   eid, credits, ep->tx_credits);
-		spin_unlock_bh(&htc->tx_lock);
-
-		if (ep->ep_ops.ep_tx_credits)
-			ep->ep_ops.ep_tx_credits(htc->ar);
-	}
+	ath10k_htc_release_credit(ep, skb_len);
 err_pull:
 	skb_pull(skb, sizeof(struct ath10k_htc_hdr));
 	return ret;
@@ -581,6 +622,252 @@ static u8 ath10k_htc_get_credit_allocation(struct ath10k_htc *htc,
 	return allocation;
 }
 
+static int ath10k_htc_send_bundle(struct ath10k_htc_ep *ep,
+				  struct sk_buff *bundle_skb,
+				  struct sk_buff_head *tx_save_head)
+{
+	struct ath10k_hif_sg_item sg_item;
+	struct ath10k_htc *htc = ep->htc;
+	struct ath10k *ar = htc->ar;
+	struct sk_buff *skb;
+	int ret, cn = 0;
+	unsigned int skb_len;
+
+	ath10k_dbg(ar, ATH10K_DBG_HTC, "bundle skb: len:%d\n", bundle_skb->len);
+	skb_len = bundle_skb->len;
+	ret = ath10k_htc_consume_credit(ep, skb_len, true);
+
+	if (!ret) {
+		sg_item.transfer_id = ep->eid;
+		sg_item.transfer_context = bundle_skb;
+		sg_item.vaddr = bundle_skb->data;
+		sg_item.len = bundle_skb->len;
+
+		ret = ath10k_hif_tx_sg(htc->ar, ep->ul_pipe_id, &sg_item, 1);
+		if (ret)
+			ath10k_htc_release_credit(ep, skb_len);
+	}
+
+	if (ret)
+		dev_kfree_skb_any(bundle_skb);
+
+	for (cn = 0; (skb = skb_dequeue_tail(tx_save_head)); cn++) {
+		if (ret) {
+			skb_pull(skb, sizeof(struct ath10k_htc_hdr));
+			skb_queue_head(&ep->tx_req_head, skb);
+		} else {
+			skb_queue_tail(&ep->tx_complete_head, skb);
+		}
+	}
+
+	if (!ret)
+		queue_work(ar->workqueue_tx_complete, &ar->tx_complete_work);
+
+	ath10k_dbg(ar, ATH10K_DBG_HTC,
+		   "bundle tx status:%d, eid:%d, req count:%d, count:%d, len:%d\n",
+		   ret, ep->eid, skb_queue_len(&ep->tx_req_head), cn, bundle_skb->len);
+	return ret;
+}
+
+static void ath10k_htc_send_one_skb(struct ath10k_htc_ep *ep, struct sk_buff *skb)
+{
+	struct ath10k_htc *htc = ep->htc;
+	struct ath10k *ar = htc->ar;
+	int ret;
+
+	ret = ath10k_htc_send(htc, ep->eid, skb);
+
+	if (ret)
+		skb_queue_head(&ep->tx_req_head, skb);
+
+	ath10k_dbg(ar, ATH10K_DBG_HTC, "tx one status:%d, eid:%d, len:%d, pending count:%d\n",
+		   ret, ep->eid, skb->len, skb_queue_len(&ep->tx_req_head));
+}
+
+static int ath10k_htc_send_bundle_skbs(struct ath10k_htc_ep *ep)
+{
+	struct ath10k_htc *htc = ep->htc;
+	struct sk_buff *bundle_skb, *skb;
+	struct sk_buff_head tx_save_head;
+	struct ath10k_htc_hdr *hdr;
+	u8 *bundle_buf;
+	int ret = 0, credit_pad, credit_remainder, trans_len, bundles_left = 0;
+
+	if (htc->ar->state == ATH10K_STATE_WEDGED)
+		return -ECOMM;
+
+	if (ep->tx_credit_flow_enabled &&
+	    ep->tx_credits < ATH10K_MIN_CREDIT_PER_HTC_TX_BUNDLE)
+		return 0;
+
+	bundles_left = ATH10K_MAX_MSG_PER_HTC_TX_BUNDLE * ep->tx_credit_size;
+	bundle_skb = dev_alloc_skb(bundles_left);
+
+	if (!bundle_skb)
+		return -ENOMEM;
+
+	bundle_buf = bundle_skb->data;
+	skb_queue_head_init(&tx_save_head);
+
+	while (true) {
+		skb = skb_dequeue(&ep->tx_req_head);
+		if (!skb)
+			break;
+
+		credit_pad = 0;
+		trans_len = skb->len + sizeof(*hdr);
+		credit_remainder = trans_len % ep->tx_credit_size;
+
+		if (credit_remainder != 0) {
+			credit_pad = ep->tx_credit_size - credit_remainder;
+			trans_len += credit_pad;
+		}
+
+		ret = ath10k_htc_consume_credit(ep,
+						bundle_buf + trans_len - bundle_skb->data,
+						false);
+		if (ret) {
+			skb_queue_head(&ep->tx_req_head, skb);
+			break;
+		}
+
+		if (bundles_left < trans_len) {
+			bundle_skb->len = bundle_buf - bundle_skb->data;
+			ret = ath10k_htc_send_bundle(ep, bundle_skb, &tx_save_head);
+
+			if (ret) {
+				skb_queue_head(&ep->tx_req_head, skb);
+				return ret;
+			}
+
+			if (skb_queue_len(&ep->tx_req_head) == 0) {
+				ath10k_htc_send_one_skb(ep, skb);
+				return ret;
+			}
+
+			if (ep->tx_credit_flow_enabled &&
+			    ep->tx_credits < ATH10K_MIN_CREDIT_PER_HTC_TX_BUNDLE) {
+				skb_queue_head(&ep->tx_req_head, skb);
+				return 0;
+			}
+
+			bundles_left =
+				ATH10K_MAX_MSG_PER_HTC_TX_BUNDLE * ep->tx_credit_size;
+			bundle_skb = dev_alloc_skb(bundles_left);
+
+			if (!bundle_skb) {
+				skb_queue_head(&ep->tx_req_head, skb);
+				return -ENOMEM;
+			}
+			bundle_buf = bundle_skb->data;
+			skb_queue_head_init(&tx_save_head);
+		}
+
+		skb_push(skb, sizeof(struct ath10k_htc_hdr));
+		ath10k_htc_prepare_tx_skb(ep, skb);
+
+		memcpy(bundle_buf, skb->data, skb->len);
+		hdr = (struct ath10k_htc_hdr *)bundle_buf;
+		hdr->flags |= ATH10K_HTC_FLAG_SEND_BUNDLE;
+		hdr->pad_len = __cpu_to_le16(credit_pad);
+		bundle_buf += trans_len;
+		bundles_left -= trans_len;
+		skb_queue_tail(&tx_save_head, skb);
+	}
+
+	if (bundle_buf != bundle_skb->data) {
+		bundle_skb->len = bundle_buf - bundle_skb->data;
+		ret = ath10k_htc_send_bundle(ep, bundle_skb, &tx_save_head);
+	} else {
+		dev_kfree_skb_any(bundle_skb);
+	}
+
+	return ret;
+}
+
+static void ath10k_htc_bundle_tx_work(struct work_struct *work)
+{
+	struct ath10k *ar = container_of(work, struct ath10k, bundle_tx_work);
+	struct ath10k_htc_ep *ep;
+	struct sk_buff *skb;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ar->htc.endpoint); i++) {
+		ep = &ar->htc.endpoint[i];
+
+		if (!ep->bundle_tx)
+			continue;
+
+		ath10k_dbg(ar, ATH10K_DBG_HTC, "bundle tx work, eid:%d, count:%d\n",
+			   ep->eid, skb_queue_len(&ep->tx_req_head));
+
+		if (skb_queue_len(&ep->tx_req_head) >=
+		    ATH10K_MIN_MSG_PER_HTC_TX_BUNDLE) {
+			ath10k_htc_send_bundle_skbs(ep);
+		} else {
+			skb = skb_dequeue(&ep->tx_req_head);
+
+			if (!skb)
+				continue;
+			ath10k_htc_send_one_skb(ep, skb);
+		}
+	}
+}
+
+static void ath10k_htc_tx_complete_work(struct work_struct *work)
+{
+	struct ath10k *ar = container_of(work, struct ath10k, tx_complete_work);
+	struct ath10k_htc_ep *ep;
+	enum ath10k_htc_ep_id eid;
+	struct sk_buff *skb;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ar->htc.endpoint); i++) {
+		ep = &ar->htc.endpoint[i];
+		eid = ep->eid;
+		if (ep->bundle_tx && eid == ar->htt.eid) {
+			ath10k_dbg(ar, ATH10K_DBG_HTC, "bundle tx complete, eid:%d, pending complete count:%d\n",
+				   ep->eid, skb_queue_len(&ep->tx_complete_head));
+
+			while (true) {
+				skb = skb_dequeue(&ep->tx_complete_head);
+				if (!skb)
+					break;
+				ath10k_htc_notify_tx_completion(ep, skb);
+			}
+		}
+	}
+}
+
+int ath10k_htc_send_hl(struct ath10k_htc *htc,
+		       enum ath10k_htc_ep_id eid,
+		       struct sk_buff *skb)
+{
+	struct ath10k_htc_ep *ep = &htc->endpoint[eid];
+	struct ath10k *ar = htc->ar;
+
+	ath10k_dbg(ar, ATH10K_DBG_HTC, "htc send hl: eid:%d, bundle:%d, tx count:%d, len:%d\n",
+		   eid, ep->bundle_tx, skb_queue_len(&ep->tx_req_head), skb->len);
+
+	if (ep->bundle_tx) {
+		skb_queue_tail(&ep->tx_req_head, skb);
+		queue_work(ar->workqueue, &ar->bundle_tx_work);
+		return 0;
+	} else {
+		return ath10k_htc_send(htc, eid, skb);
+	}
+}
+
+void ath10k_htc_setup_tx_req(struct ath10k_htc_ep *ep)
+{
+	if (ep->htc->max_msgs_per_htc_bundle >= ATH10K_MIN_MSG_PER_HTC_TX_BUNDLE &&
+	    !ep->bundle_tx) {
+		ep->bundle_tx = true;
+		skb_queue_head_init(&ep->tx_req_head);
+		skb_queue_head_init(&ep->tx_complete_head);
+	}
+}
+
 int ath10k_htc_wait_target(struct ath10k_htc *htc)
 {
 	struct ath10k *ar = htc->ar;
@@ -657,6 +944,9 @@ int ath10k_htc_wait_target(struct ath10k_htc *htc)
 			   htc->max_msgs_per_htc_bundle);
 	}
 
+	INIT_WORK(&ar->bundle_tx_work, ath10k_htc_bundle_tx_work);
+	INIT_WORK(&ar->tx_complete_work, ath10k_htc_tx_complete_work);
+
 	return 0;
 }
 
@@ -801,6 +1091,7 @@ int ath10k_htc_connect_service(struct ath10k_htc *htc,
 	ep->max_tx_queue_depth = conn_req->max_send_queue_depth;
 	ep->max_ep_message_len = __le16_to_cpu(resp_msg->max_msg_size);
 	ep->tx_credits = tx_alloc;
+	ep->tx_credit_size = htc->target_credit_size;
 
 	/* copy all the callbacks */
 	ep->ep_ops = conn_req->ep_ops;
diff --git a/drivers/net/wireless/ath/ath10k/htc.h b/drivers/net/wireless/ath/ath10k/htc.h
index d0f37cbdab8b..7ce47cb29def 100644
--- a/drivers/net/wireless/ath/ath10k/htc.h
+++ b/drivers/net/wireless/ath/ath10k/htc.h
@@ -64,8 +64,14 @@ struct ath10k_htc_hdr {
 		u8 seq_no; /* for tx */
 		u8 control_byte1;
 	} __packed;
-	u8 pad0;
-	u8 pad1;
+	union {
+		__le16 pad_len;
+		struct {
+			u8 pad0;
+			u8 pad1;
+		} __packed;
+	} __packed;
+
 } __packed __aligned(4);
 
 enum ath10k_ath10k_htc_msg_id {
@@ -102,6 +108,10 @@ enum ath10k_htc_conn_svc_status {
 	ATH10K_HTC_CONN_SVC_STATUS_NO_MORE_EP   = 4
 };
 
+#define ATH10K_MAX_MSG_PER_HTC_TX_BUNDLE        32
+#define ATH10K_MIN_MSG_PER_HTC_TX_BUNDLE        2
+#define ATH10K_MIN_CREDIT_PER_HTC_TX_BUNDLE     2
+
 enum ath10k_htc_setup_complete_flags {
 	ATH10K_HTC_SETUP_COMPLETE_FLAGS_RX_BNDL_EN = 1
 };
@@ -334,7 +344,12 @@ struct ath10k_htc_ep {
 
 	u8 seq_no; /* for debugging */
 	int tx_credits;
+	int tx_credit_size;
 	bool tx_credit_flow_enabled;
+	bool bundle_tx;
+	struct sk_buff_head tx_req_head;
+	struct sk_buff_head tx_complete_head;
+
 };
 
 struct ath10k_htc_svc_tx_credits {
@@ -363,6 +378,7 @@ struct ath10k_htc {
 
 int ath10k_htc_init(struct ath10k *ar);
 int ath10k_htc_wait_target(struct ath10k_htc *htc);
+void ath10k_htc_setup_tx_req(struct ath10k_htc_ep *ep);
 int ath10k_htc_start(struct ath10k_htc *htc);
 int ath10k_htc_connect_service(struct ath10k_htc *htc,
 			       struct ath10k_htc_svc_conn_req  *conn_req,
@@ -372,6 +388,8 @@ void ath10k_htc_change_tx_credit_flow(struct ath10k_htc *htc,
 				      bool enable);
 int ath10k_htc_send(struct ath10k_htc *htc, enum ath10k_htc_ep_id eid,
 		    struct sk_buff *packet);
+int ath10k_htc_send_hl(struct ath10k_htc *htc, enum ath10k_htc_ep_id eid,
+		       struct sk_buff *packet);
 struct sk_buff *ath10k_htc_alloc_skb(struct ath10k *ar, int size);
 void ath10k_htc_tx_completion_handler(struct ath10k *ar, struct sk_buff *skb);
 void ath10k_htc_rx_completion_handler(struct ath10k *ar, struct sk_buff *skb);
diff --git a/drivers/net/wireless/ath/ath10k/htt.c b/drivers/net/wireless/ath/ath10k/htt.c
index 4354bf285ff1..127b4e4980ef 100644
--- a/drivers/net/wireless/ath/ath10k/htt.c
+++ b/drivers/net/wireless/ath/ath10k/htt.c
@@ -135,6 +135,8 @@ int ath10k_htt_connect(struct ath10k_htt *htt)
 {
 	struct ath10k_htc_svc_conn_req conn_req;
 	struct ath10k_htc_svc_conn_resp conn_resp;
+	struct ath10k *ar = htt->ar;
+	struct ath10k_htc_ep *ep;
 	int status;
 
 	memset(&conn_req, 0, sizeof(conn_req));
@@ -142,6 +144,7 @@ int ath10k_htt_connect(struct ath10k_htt *htt)
 
 	conn_req.ep_ops.ep_tx_complete = ath10k_htt_htc_tx_complete;
 	conn_req.ep_ops.ep_rx_complete = ath10k_htt_htc_t2h_msg_handler;
+	conn_req.ep_ops.ep_tx_credits = ath10k_htt_op_ep_tx_credits;
 
 	/* connect to control service */
 	conn_req.service_id = ATH10K_HTC_SVC_ID_HTT_DATA_MSG;
@@ -154,6 +157,11 @@ int ath10k_htt_connect(struct ath10k_htt *htt)
 
 	htt->eid = conn_resp.eid;
 
+	if (ar->bus_param.dev_type == ATH10K_DEV_TYPE_HL) {
+		ep = &ar->htc.endpoint[htt->eid];
+		ath10k_htc_setup_tx_req(ep);
+	}
+
 	htt->disable_tx_comp = ath10k_hif_get_htt_tx_complete(htt->ar);
 	if (htt->disable_tx_comp)
 		ath10k_htc_change_tx_credit_flow(&htt->ar->htc, htt->eid, true);
diff --git a/drivers/net/wireless/ath/ath10k/htt.h b/drivers/net/wireless/ath/ath10k/htt.h
index 889bf9fe051a..4851a2e3345f 100644
--- a/drivers/net/wireless/ath/ath10k/htt.h
+++ b/drivers/net/wireless/ath/ath10k/htt.h
@@ -2030,6 +2030,9 @@ struct ath10k_htt {
 	const struct ath10k_htt_tx_ops *tx_ops;
 	const struct ath10k_htt_rx_ops *rx_ops;
 	bool disable_tx_comp;
+	bool bundle_tx;
+	struct sk_buff_head tx_req_head;
+	struct sk_buff_head tx_complete_head;
 };
 
 struct ath10k_htt_tx_ops {
@@ -2276,6 +2279,7 @@ int ath10k_htt_tx_fetch_resp(struct ath10k *ar,
 			     __le16 fetch_seq_num,
 			     struct htt_tx_fetch_record *records,
 			     size_t num_records);
+void ath10k_htt_op_ep_tx_credits(struct ath10k *ar);
 
 void ath10k_htt_tx_txq_update(struct ieee80211_hw *hw,
 			      struct ieee80211_txq *txq);
diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c
index 17ad6c51e1dd..d0b6950ca99e 100644
--- a/drivers/net/wireless/ath/ath10k/htt_rx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
@@ -3908,6 +3908,7 @@ bool ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb)
 			ath10k_dbg(ar, ATH10K_DBG_HTT,
 				   "credit total:%d\n",
 				   ep->tx_credits);
+			ep->ep_ops.ep_tx_credits(htc->ar);
 		}
 		break;
 	}
diff --git a/drivers/net/wireless/ath/ath10k/htt_tx.c b/drivers/net/wireless/ath/ath10k/htt_tx.c
index c6c4b2a4d20f..d77379b3d3dc 100644
--- a/drivers/net/wireless/ath/ath10k/htt_tx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_tx.c
@@ -541,6 +541,11 @@ void ath10k_htt_tx_free(struct ath10k_htt *htt)
 	ath10k_htt_tx_destroy(htt);
 }
 
+void ath10k_htt_op_ep_tx_credits(struct ath10k *ar)
+{
+	queue_work(ar->workqueue, &ar->bundle_tx_work);
+}
+
 void ath10k_htt_htc_tx_complete(struct ath10k *ar, struct sk_buff *skb)
 {
 	struct ath10k_htt *htt = &ar->htt;
@@ -1365,7 +1370,7 @@ static int ath10k_htt_tx_hl(struct ath10k_htt *htt, enum ath10k_hw_txrx_mode txm
 	 */
 	tx_desc->peerid = __cpu_to_le32(HTT_INVALID_PEERID);
 
-	res = ath10k_htc_send(&htt->ar->htc, htt->eid, msdu);
+	res = ath10k_htc_send_hl(&htt->ar->htc, htt->eid, msdu);
 
 out:
 	return res;
-- 
2.23.0


^ permalink raw reply related	[flat|nested] 10+ messages in thread
* [PATCH v7 2/3] ath10k: add htt TX bundle for sdio
@ 2019-10-18  7:27   ` Wen Gong
  0 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-10-18  7:27 UTC (permalink / raw)
  To: ath10k; +Cc: linux-wireless

The transmission utilization ratio for sdio bus for small packet is
slow, because the space and time cost for sdio bus is same for large
length packet and small length packet. So the speed of data for large
length packet is higher than small length.

Test result of different length of data:

data packet(byte)   cost time(us)   calculated rate(Mbps)
      256               28                73
      512               33               124
     1024               35               234
     1792               45               318
    14336              168               682
    28672              333               688
    57344              660               695

This patch change the TX packet from single packet to a large length
bundle packet, max size is 32, it results in significant performance
improvement on TX path.

This patch only effect sdio chip, it will not effect PCI, SNOC etc.
It only enable bundle for sdio chip.

Tested with QCA6174 SDIO with firmware
WLAN.RMH.4.4.1-00017-QCARMSWP-1.

Signed-off-by: Wen Gong <wgong@codeaurora.org>
---
 drivers/net/wireless/ath/ath10k/core.c   |  14 +-
 drivers/net/wireless/ath/ath10k/core.h   |   4 +-
 drivers/net/wireless/ath/ath10k/htc.c    | 351 +++++++++++++++++++++--
 drivers/net/wireless/ath/ath10k/htc.h    |  22 +-
 drivers/net/wireless/ath/ath10k/htt.c    |   8 +
 drivers/net/wireless/ath/ath10k/htt.h    |   4 +
 drivers/net/wireless/ath/ath10k/htt_rx.c |   1 +
 drivers/net/wireless/ath/ath10k/htt_tx.c |   7 +-
 8 files changed, 374 insertions(+), 37 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/core.c b/drivers/net/wireless/ath/ath10k/core.c
index ebaaf18f6dc7..d9f1d5e75c86 100644
--- a/drivers/net/wireless/ath/ath10k/core.c
+++ b/drivers/net/wireless/ath/ath10k/core.c
@@ -3211,6 +3211,11 @@ struct ath10k *ath10k_core_create(size_t priv_size, struct device *dev,
 	if (!ar->workqueue_aux)
 		goto err_free_wq;
 
+	ar->workqueue_tx_complete =
+		create_singlethread_workqueue("ath10k_tx_complete_wq");
+	if (!ar->workqueue_tx_complete)
+		goto err_free_aux_wq;
+
 	mutex_init(&ar->conf_mutex);
 	mutex_init(&ar->dump_mutex);
 	spin_lock_init(&ar->data_lock);
@@ -3236,7 +3241,7 @@ struct ath10k *ath10k_core_create(size_t priv_size, struct device *dev,
 
 	ret = ath10k_coredump_create(ar);
 	if (ret)
-		goto err_free_aux_wq;
+		goto err_free_tx_complete;
 
 	ret = ath10k_debug_create(ar);
 	if (ret)
@@ -3246,12 +3251,12 @@ struct ath10k *ath10k_core_create(size_t priv_size, struct device *dev,
 
 err_free_coredump:
 	ath10k_coredump_destroy(ar);
-
+err_free_tx_complete:
+	destroy_workqueue(ar->workqueue_tx_complete);
 err_free_aux_wq:
 	destroy_workqueue(ar->workqueue_aux);
 err_free_wq:
 	destroy_workqueue(ar->workqueue);
-
 err_free_mac:
 	ath10k_mac_destroy(ar);
 
@@ -3267,6 +3272,9 @@ void ath10k_core_destroy(struct ath10k *ar)
 	flush_workqueue(ar->workqueue_aux);
 	destroy_workqueue(ar->workqueue_aux);
 
+	flush_workqueue(ar->workqueue_tx_complete);
+	destroy_workqueue(ar->workqueue_tx_complete);
+
 	ath10k_debug_destroy(ar);
 	ath10k_coredump_destroy(ar);
 	ath10k_htt_tx_destroy(&ar->htt);
diff --git a/drivers/net/wireless/ath/ath10k/core.h b/drivers/net/wireless/ath/ath10k/core.h
index c96d1a17fd00..0eebf3564aa4 100644
--- a/drivers/net/wireless/ath/ath10k/core.h
+++ b/drivers/net/wireless/ath/ath10k/core.h
@@ -1080,7 +1080,7 @@ struct ath10k {
 	struct workqueue_struct *workqueue;
 	/* Auxiliary workqueue */
 	struct workqueue_struct *workqueue_aux;
-
+	struct workqueue_struct *workqueue_tx_complete;
 	/* prevents concurrent FW reconfiguration */
 	struct mutex conf_mutex;
 
@@ -1121,6 +1121,8 @@ struct ath10k {
 
 	struct work_struct register_work;
 	struct work_struct restart_work;
+	struct work_struct bundle_tx_work;
+	struct work_struct tx_complete_work;
 
 	/* cycle count is reported twice for each visited channel during scan.
 	 * access protected by data_lock
diff --git a/drivers/net/wireless/ath/ath10k/htc.c b/drivers/net/wireless/ath/ath10k/htc.c
index 4c6cdc24d282..a2bf34e5a96e 100644
--- a/drivers/net/wireless/ath/ath10k/htc.c
+++ b/drivers/net/wireless/ath/ath10k/htc.c
@@ -51,10 +51,12 @@ void ath10k_htc_notify_tx_completion(struct ath10k_htc_ep *ep,
 				     struct sk_buff *skb)
 {
 	struct ath10k *ar = ep->htc->ar;
+	struct ath10k_htc_hdr *hdr;
 
 	ath10k_dbg(ar, ATH10K_DBG_HTC, "%s: ep %d skb %pK\n", __func__,
 		   ep->eid, skb);
 
+	hdr = (struct ath10k_htc_hdr *)skb->data;
 	ath10k_htc_restore_tx_skb(ep->htc, skb);
 
 	if (!ep->ep_ops.ep_tx_complete) {
@@ -63,6 +65,11 @@ void ath10k_htc_notify_tx_completion(struct ath10k_htc_ep *ep,
 		return;
 	}
 
+	if (hdr->flags & ATH10K_HTC_FLAG_SEND_BUNDLE) {
+		dev_kfree_skb_any(skb);
+		return;
+	}
+
 	ep->ep_ops.ep_tx_complete(ep->htc->ar, skb);
 }
 EXPORT_SYMBOL(ath10k_htc_notify_tx_completion);
@@ -78,7 +85,7 @@ static void ath10k_htc_prepare_tx_skb(struct ath10k_htc_ep *ep,
 	hdr->eid = ep->eid;
 	hdr->len = __cpu_to_le16(skb->len - sizeof(*hdr));
 	hdr->flags = 0;
-	if (ep->tx_credit_flow_enabled)
+	if (ep->tx_credit_flow_enabled && !ep->bundle_tx)
 		hdr->flags |= ATH10K_HTC_FLAG_NEED_CREDIT_UPDATE;
 
 	spin_lock_bh(&ep->htc->tx_lock);
@@ -86,6 +93,63 @@ static void ath10k_htc_prepare_tx_skb(struct ath10k_htc_ep *ep,
 	spin_unlock_bh(&ep->htc->tx_lock);
 }
 
+static int ath10k_htc_consume_credit(struct ath10k_htc_ep *ep,
+				     unsigned int len,
+				     bool consume)
+{
+	struct ath10k_htc *htc = ep->htc;
+	struct ath10k *ar = htc->ar;
+	enum ath10k_htc_ep_id eid = ep->eid;
+	int credits, ret = 0;
+
+	if (!ep->tx_credit_flow_enabled)
+		return 0;
+
+	credits = DIV_ROUND_UP(len, ep->tx_credit_size);
+	spin_lock_bh(&htc->tx_lock);
+
+	if (ep->tx_credits < credits) {
+		ath10k_dbg(ar, ATH10K_DBG_HTC,
+			   "htc insufficient credits ep %d required %d available %d consume %d\n",
+			   eid, credits, ep->tx_credits, consume);
+		ret = -EAGAIN;
+		goto unlock;
+	}
+
+	if (consume) {
+		ep->tx_credits -= credits;
+		ath10k_dbg(ar, ATH10K_DBG_HTC,
+			   "htc ep %d consumed %d credits (total %d)\n",
+			   eid, credits, ep->tx_credits);
+	}
+
+unlock:
+	spin_unlock_bh(&htc->tx_lock);
+	return ret;
+}
+
+static void ath10k_htc_release_credit(struct ath10k_htc_ep *ep, unsigned int len)
+{
+	struct ath10k_htc *htc = ep->htc;
+	struct ath10k *ar = htc->ar;
+	enum ath10k_htc_ep_id eid = ep->eid;
+	int credits;
+
+	if (!ep->tx_credit_flow_enabled)
+		return;
+
+	credits = DIV_ROUND_UP(len, ep->tx_credit_size);
+	spin_lock_bh(&htc->tx_lock);
+	ep->tx_credits += credits;
+	ath10k_dbg(ar, ATH10K_DBG_HTC,
+		   "htc ep %d reverted %d credits back (total %d)\n",
+		   eid, credits, ep->tx_credits);
+	spin_unlock_bh(&htc->tx_lock);
+
+	if (ep->ep_ops.ep_tx_credits)
+		ep->ep_ops.ep_tx_credits(htc->ar);
+}
+
 int ath10k_htc_send(struct ath10k_htc *htc,
 		    enum ath10k_htc_ep_id eid,
 		    struct sk_buff *skb)
@@ -95,8 +159,8 @@ int ath10k_htc_send(struct ath10k_htc *htc,
 	struct ath10k_skb_cb *skb_cb = ATH10K_SKB_CB(skb);
 	struct ath10k_hif_sg_item sg_item;
 	struct device *dev = htc->ar->dev;
-	int credits = 0;
 	int ret;
+	unsigned int skb_len;
 
 	if (htc->ar->state == ATH10K_STATE_WEDGED)
 		return -ECOMM;
@@ -108,23 +172,10 @@ int ath10k_htc_send(struct ath10k_htc *htc,
 
 	skb_push(skb, sizeof(struct ath10k_htc_hdr));
 
-	if (ep->tx_credit_flow_enabled) {
-		credits = DIV_ROUND_UP(skb->len, htc->target_credit_size);
-		spin_lock_bh(&htc->tx_lock);
-		if (ep->tx_credits < credits) {
-			ath10k_dbg(ar, ATH10K_DBG_HTC,
-				   "htc insufficient credits ep %d required %d available %d\n",
-				   eid, credits, ep->tx_credits);
-			spin_unlock_bh(&htc->tx_lock);
-			ret = -EAGAIN;
-			goto err_pull;
-		}
-		ep->tx_credits -= credits;
-		ath10k_dbg(ar, ATH10K_DBG_HTC,
-			   "htc ep %d consumed %d credits (total %d)\n",
-			   eid, credits, ep->tx_credits);
-		spin_unlock_bh(&htc->tx_lock);
-	}
+	skb_len = skb->len;
+	ret = ath10k_htc_consume_credit(ep, skb_len, true);
+	if (ret)
+		goto err_pull;
 
 	ath10k_htc_prepare_tx_skb(ep, skb);
 
@@ -155,17 +206,7 @@ int ath10k_htc_send(struct ath10k_htc *htc,
 	if (ar->bus_param.dev_type != ATH10K_DEV_TYPE_HL)
 		dma_unmap_single(dev, skb_cb->paddr, skb->len, DMA_TO_DEVICE);
 err_credits:
-	if (ep->tx_credit_flow_enabled) {
-		spin_lock_bh(&htc->tx_lock);
-		ep->tx_credits += credits;
-		ath10k_dbg(ar, ATH10K_DBG_HTC,
-			   "htc ep %d reverted %d credits back (total %d)\n",
-			   eid, credits, ep->tx_credits);
-		spin_unlock_bh(&htc->tx_lock);
-
-		if (ep->ep_ops.ep_tx_credits)
-			ep->ep_ops.ep_tx_credits(htc->ar);
-	}
+	ath10k_htc_release_credit(ep, skb_len);
 err_pull:
 	skb_pull(skb, sizeof(struct ath10k_htc_hdr));
 	return ret;
@@ -581,6 +622,252 @@ static u8 ath10k_htc_get_credit_allocation(struct ath10k_htc *htc,
 	return allocation;
 }
 
+static int ath10k_htc_send_bundle(struct ath10k_htc_ep *ep,
+				  struct sk_buff *bundle_skb,
+				  struct sk_buff_head *tx_save_head)
+{
+	struct ath10k_hif_sg_item sg_item;
+	struct ath10k_htc *htc = ep->htc;
+	struct ath10k *ar = htc->ar;
+	struct sk_buff *skb;
+	int ret, cn = 0;
+	unsigned int skb_len;
+
+	ath10k_dbg(ar, ATH10K_DBG_HTC, "bundle skb: len:%d\n", bundle_skb->len);
+	skb_len = bundle_skb->len;
+	ret = ath10k_htc_consume_credit(ep, skb_len, true);
+
+	if (!ret) {
+		sg_item.transfer_id = ep->eid;
+		sg_item.transfer_context = bundle_skb;
+		sg_item.vaddr = bundle_skb->data;
+		sg_item.len = bundle_skb->len;
+
+		ret = ath10k_hif_tx_sg(htc->ar, ep->ul_pipe_id, &sg_item, 1);
+		if (ret)
+			ath10k_htc_release_credit(ep, skb_len);
+	}
+
+	if (ret)
+		dev_kfree_skb_any(bundle_skb);
+
+	for (cn = 0; (skb = skb_dequeue_tail(tx_save_head)); cn++) {
+		if (ret) {
+			skb_pull(skb, sizeof(struct ath10k_htc_hdr));
+			skb_queue_head(&ep->tx_req_head, skb);
+		} else {
+			skb_queue_tail(&ep->tx_complete_head, skb);
+		}
+	}
+
+	if (!ret)
+		queue_work(ar->workqueue_tx_complete, &ar->tx_complete_work);
+
+	ath10k_dbg(ar, ATH10K_DBG_HTC,
+		   "bundle tx status:%d, eid:%d, req count:%d, count:%d, len:%d\n",
+		   ret, ep->eid, skb_queue_len(&ep->tx_req_head), cn, bundle_skb->len);
+	return ret;
+}
+
+static void ath10k_htc_send_one_skb(struct ath10k_htc_ep *ep, struct sk_buff *skb)
+{
+	struct ath10k_htc *htc = ep->htc;
+	struct ath10k *ar = htc->ar;
+	int ret;
+
+	ret = ath10k_htc_send(htc, ep->eid, skb);
+
+	if (ret)
+		skb_queue_head(&ep->tx_req_head, skb);
+
+	ath10k_dbg(ar, ATH10K_DBG_HTC, "tx one status:%d, eid:%d, len:%d, pending count:%d\n",
+		   ret, ep->eid, skb->len, skb_queue_len(&ep->tx_req_head));
+}
+
+static int ath10k_htc_send_bundle_skbs(struct ath10k_htc_ep *ep)
+{
+	struct ath10k_htc *htc = ep->htc;
+	struct sk_buff *bundle_skb, *skb;
+	struct sk_buff_head tx_save_head;
+	struct ath10k_htc_hdr *hdr;
+	u8 *bundle_buf;
+	int ret = 0, credit_pad, credit_remainder, trans_len, bundles_left = 0;
+
+	if (htc->ar->state == ATH10K_STATE_WEDGED)
+		return -ECOMM;
+
+	if (ep->tx_credit_flow_enabled &&
+	    ep->tx_credits < ATH10K_MIN_CREDIT_PER_HTC_TX_BUNDLE)
+		return 0;
+
+	bundles_left = ATH10K_MAX_MSG_PER_HTC_TX_BUNDLE * ep->tx_credit_size;
+	bundle_skb = dev_alloc_skb(bundles_left);
+
+	if (!bundle_skb)
+		return -ENOMEM;
+
+	bundle_buf = bundle_skb->data;
+	skb_queue_head_init(&tx_save_head);
+
+	while (true) {
+		skb = skb_dequeue(&ep->tx_req_head);
+		if (!skb)
+			break;
+
+		credit_pad = 0;
+		trans_len = skb->len + sizeof(*hdr);
+		credit_remainder = trans_len % ep->tx_credit_size;
+
+		if (credit_remainder != 0) {
+			credit_pad = ep->tx_credit_size - credit_remainder;
+			trans_len += credit_pad;
+		}
+
+		ret = ath10k_htc_consume_credit(ep,
+						bundle_buf + trans_len - bundle_skb->data,
+						false);
+		if (ret) {
+			skb_queue_head(&ep->tx_req_head, skb);
+			break;
+		}
+
+		if (bundles_left < trans_len) {
+			bundle_skb->len = bundle_buf - bundle_skb->data;
+			ret = ath10k_htc_send_bundle(ep, bundle_skb, &tx_save_head);
+
+			if (ret) {
+				skb_queue_head(&ep->tx_req_head, skb);
+				return ret;
+			}
+
+			if (skb_queue_len(&ep->tx_req_head) == 0) {
+				ath10k_htc_send_one_skb(ep, skb);
+				return ret;
+			}
+
+			if (ep->tx_credit_flow_enabled &&
+			    ep->tx_credits < ATH10K_MIN_CREDIT_PER_HTC_TX_BUNDLE) {
+				skb_queue_head(&ep->tx_req_head, skb);
+				return 0;
+			}
+
+			bundles_left =
+				ATH10K_MAX_MSG_PER_HTC_TX_BUNDLE * ep->tx_credit_size;
+			bundle_skb = dev_alloc_skb(bundles_left);
+
+			if (!bundle_skb) {
+				skb_queue_head(&ep->tx_req_head, skb);
+				return -ENOMEM;
+			}
+			bundle_buf = bundle_skb->data;
+			skb_queue_head_init(&tx_save_head);
+		}
+
+		skb_push(skb, sizeof(struct ath10k_htc_hdr));
+		ath10k_htc_prepare_tx_skb(ep, skb);
+
+		memcpy(bundle_buf, skb->data, skb->len);
+		hdr = (struct ath10k_htc_hdr *)bundle_buf;
+		hdr->flags |= ATH10K_HTC_FLAG_SEND_BUNDLE;
+		hdr->pad_len = __cpu_to_le16(credit_pad);
+		bundle_buf += trans_len;
+		bundles_left -= trans_len;
+		skb_queue_tail(&tx_save_head, skb);
+	}
+
+	if (bundle_buf != bundle_skb->data) {
+		bundle_skb->len = bundle_buf - bundle_skb->data;
+		ret = ath10k_htc_send_bundle(ep, bundle_skb, &tx_save_head);
+	} else {
+		dev_kfree_skb_any(bundle_skb);
+	}
+
+	return ret;
+}
+
+static void ath10k_htc_bundle_tx_work(struct work_struct *work)
+{
+	struct ath10k *ar = container_of(work, struct ath10k, bundle_tx_work);
+	struct ath10k_htc_ep *ep;
+	struct sk_buff *skb;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ar->htc.endpoint); i++) {
+		ep = &ar->htc.endpoint[i];
+
+		if (!ep->bundle_tx)
+			continue;
+
+		ath10k_dbg(ar, ATH10K_DBG_HTC, "bundle tx work, eid:%d, count:%d\n",
+			   ep->eid, skb_queue_len(&ep->tx_req_head));
+
+		if (skb_queue_len(&ep->tx_req_head) >=
+		    ATH10K_MIN_MSG_PER_HTC_TX_BUNDLE) {
+			ath10k_htc_send_bundle_skbs(ep);
+		} else {
+			skb = skb_dequeue(&ep->tx_req_head);
+
+			if (!skb)
+				continue;
+			ath10k_htc_send_one_skb(ep, skb);
+		}
+	}
+}
+
+static void ath10k_htc_tx_complete_work(struct work_struct *work)
+{
+	struct ath10k *ar = container_of(work, struct ath10k, tx_complete_work);
+	struct ath10k_htc_ep *ep;
+	enum ath10k_htc_ep_id eid;
+	struct sk_buff *skb;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ar->htc.endpoint); i++) {
+		ep = &ar->htc.endpoint[i];
+		eid = ep->eid;
+		if (ep->bundle_tx && eid == ar->htt.eid) {
+			ath10k_dbg(ar, ATH10K_DBG_HTC, "bundle tx complete, eid:%d, pending complete count:%d\n",
+				   ep->eid, skb_queue_len(&ep->tx_complete_head));
+
+			while (true) {
+				skb = skb_dequeue(&ep->tx_complete_head);
+				if (!skb)
+					break;
+				ath10k_htc_notify_tx_completion(ep, skb);
+			}
+		}
+	}
+}
+
+int ath10k_htc_send_hl(struct ath10k_htc *htc,
+		       enum ath10k_htc_ep_id eid,
+		       struct sk_buff *skb)
+{
+	struct ath10k_htc_ep *ep = &htc->endpoint[eid];
+	struct ath10k *ar = htc->ar;
+
+	ath10k_dbg(ar, ATH10K_DBG_HTC, "htc send hl: eid:%d, bundle:%d, tx count:%d, len:%d\n",
+		   eid, ep->bundle_tx, skb_queue_len(&ep->tx_req_head), skb->len);
+
+	if (ep->bundle_tx) {
+		skb_queue_tail(&ep->tx_req_head, skb);
+		queue_work(ar->workqueue, &ar->bundle_tx_work);
+		return 0;
+	} else {
+		return ath10k_htc_send(htc, eid, skb);
+	}
+}
+
+void ath10k_htc_setup_tx_req(struct ath10k_htc_ep *ep)
+{
+	if (ep->htc->max_msgs_per_htc_bundle >= ATH10K_MIN_MSG_PER_HTC_TX_BUNDLE &&
+	    !ep->bundle_tx) {
+		ep->bundle_tx = true;
+		skb_queue_head_init(&ep->tx_req_head);
+		skb_queue_head_init(&ep->tx_complete_head);
+	}
+}
+
 int ath10k_htc_wait_target(struct ath10k_htc *htc)
 {
 	struct ath10k *ar = htc->ar;
@@ -657,6 +944,9 @@ int ath10k_htc_wait_target(struct ath10k_htc *htc)
 			   htc->max_msgs_per_htc_bundle);
 	}
 
+	INIT_WORK(&ar->bundle_tx_work, ath10k_htc_bundle_tx_work);
+	INIT_WORK(&ar->tx_complete_work, ath10k_htc_tx_complete_work);
+
 	return 0;
 }
 
@@ -801,6 +1091,7 @@ int ath10k_htc_connect_service(struct ath10k_htc *htc,
 	ep->max_tx_queue_depth = conn_req->max_send_queue_depth;
 	ep->max_ep_message_len = __le16_to_cpu(resp_msg->max_msg_size);
 	ep->tx_credits = tx_alloc;
+	ep->tx_credit_size = htc->target_credit_size;
 
 	/* copy all the callbacks */
 	ep->ep_ops = conn_req->ep_ops;
diff --git a/drivers/net/wireless/ath/ath10k/htc.h b/drivers/net/wireless/ath/ath10k/htc.h
index d0f37cbdab8b..7ce47cb29def 100644
--- a/drivers/net/wireless/ath/ath10k/htc.h
+++ b/drivers/net/wireless/ath/ath10k/htc.h
@@ -64,8 +64,14 @@ struct ath10k_htc_hdr {
 		u8 seq_no; /* for tx */
 		u8 control_byte1;
 	} __packed;
-	u8 pad0;
-	u8 pad1;
+	union {
+		__le16 pad_len;
+		struct {
+			u8 pad0;
+			u8 pad1;
+		} __packed;
+	} __packed;
+
 } __packed __aligned(4);
 
 enum ath10k_ath10k_htc_msg_id {
@@ -102,6 +108,10 @@ enum ath10k_htc_conn_svc_status {
 	ATH10K_HTC_CONN_SVC_STATUS_NO_MORE_EP   = 4
 };
 
+#define ATH10K_MAX_MSG_PER_HTC_TX_BUNDLE        32
+#define ATH10K_MIN_MSG_PER_HTC_TX_BUNDLE        2
+#define ATH10K_MIN_CREDIT_PER_HTC_TX_BUNDLE     2
+
 enum ath10k_htc_setup_complete_flags {
 	ATH10K_HTC_SETUP_COMPLETE_FLAGS_RX_BNDL_EN = 1
 };
@@ -334,7 +344,12 @@ struct ath10k_htc_ep {
 
 	u8 seq_no; /* for debugging */
 	int tx_credits;
+	int tx_credit_size;
 	bool tx_credit_flow_enabled;
+	bool bundle_tx;
+	struct sk_buff_head tx_req_head;
+	struct sk_buff_head tx_complete_head;
+
 };
 
 struct ath10k_htc_svc_tx_credits {
@@ -363,6 +378,7 @@ struct ath10k_htc {
 
 int ath10k_htc_init(struct ath10k *ar);
 int ath10k_htc_wait_target(struct ath10k_htc *htc);
+void ath10k_htc_setup_tx_req(struct ath10k_htc_ep *ep);
 int ath10k_htc_start(struct ath10k_htc *htc);
 int ath10k_htc_connect_service(struct ath10k_htc *htc,
 			       struct ath10k_htc_svc_conn_req  *conn_req,
@@ -372,6 +388,8 @@ void ath10k_htc_change_tx_credit_flow(struct ath10k_htc *htc,
 				      bool enable);
 int ath10k_htc_send(struct ath10k_htc *htc, enum ath10k_htc_ep_id eid,
 		    struct sk_buff *packet);
+int ath10k_htc_send_hl(struct ath10k_htc *htc, enum ath10k_htc_ep_id eid,
+		       struct sk_buff *packet);
 struct sk_buff *ath10k_htc_alloc_skb(struct ath10k *ar, int size);
 void ath10k_htc_tx_completion_handler(struct ath10k *ar, struct sk_buff *skb);
 void ath10k_htc_rx_completion_handler(struct ath10k *ar, struct sk_buff *skb);
diff --git a/drivers/net/wireless/ath/ath10k/htt.c b/drivers/net/wireless/ath/ath10k/htt.c
index 4354bf285ff1..127b4e4980ef 100644
--- a/drivers/net/wireless/ath/ath10k/htt.c
+++ b/drivers/net/wireless/ath/ath10k/htt.c
@@ -135,6 +135,8 @@ int ath10k_htt_connect(struct ath10k_htt *htt)
 {
 	struct ath10k_htc_svc_conn_req conn_req;
 	struct ath10k_htc_svc_conn_resp conn_resp;
+	struct ath10k *ar = htt->ar;
+	struct ath10k_htc_ep *ep;
 	int status;
 
 	memset(&conn_req, 0, sizeof(conn_req));
@@ -142,6 +144,7 @@ int ath10k_htt_connect(struct ath10k_htt *htt)
 
 	conn_req.ep_ops.ep_tx_complete = ath10k_htt_htc_tx_complete;
 	conn_req.ep_ops.ep_rx_complete = ath10k_htt_htc_t2h_msg_handler;
+	conn_req.ep_ops.ep_tx_credits = ath10k_htt_op_ep_tx_credits;
 
 	/* connect to control service */
 	conn_req.service_id = ATH10K_HTC_SVC_ID_HTT_DATA_MSG;
@@ -154,6 +157,11 @@ int ath10k_htt_connect(struct ath10k_htt *htt)
 
 	htt->eid = conn_resp.eid;
 
+	if (ar->bus_param.dev_type == ATH10K_DEV_TYPE_HL) {
+		ep = &ar->htc.endpoint[htt->eid];
+		ath10k_htc_setup_tx_req(ep);
+	}
+
 	htt->disable_tx_comp = ath10k_hif_get_htt_tx_complete(htt->ar);
 	if (htt->disable_tx_comp)
 		ath10k_htc_change_tx_credit_flow(&htt->ar->htc, htt->eid, true);
diff --git a/drivers/net/wireless/ath/ath10k/htt.h b/drivers/net/wireless/ath/ath10k/htt.h
index 889bf9fe051a..4851a2e3345f 100644
--- a/drivers/net/wireless/ath/ath10k/htt.h
+++ b/drivers/net/wireless/ath/ath10k/htt.h
@@ -2030,6 +2030,9 @@ struct ath10k_htt {
 	const struct ath10k_htt_tx_ops *tx_ops;
 	const struct ath10k_htt_rx_ops *rx_ops;
 	bool disable_tx_comp;
+	bool bundle_tx;
+	struct sk_buff_head tx_req_head;
+	struct sk_buff_head tx_complete_head;
 };
 
 struct ath10k_htt_tx_ops {
@@ -2276,6 +2279,7 @@ int ath10k_htt_tx_fetch_resp(struct ath10k *ar,
 			     __le16 fetch_seq_num,
 			     struct htt_tx_fetch_record *records,
 			     size_t num_records);
+void ath10k_htt_op_ep_tx_credits(struct ath10k *ar);
 
 void ath10k_htt_tx_txq_update(struct ieee80211_hw *hw,
 			      struct ieee80211_txq *txq);
diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c
index 17ad6c51e1dd..d0b6950ca99e 100644
--- a/drivers/net/wireless/ath/ath10k/htt_rx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
@@ -3908,6 +3908,7 @@ bool ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb)
 			ath10k_dbg(ar, ATH10K_DBG_HTT,
 				   "credit total:%d\n",
 				   ep->tx_credits);
+			ep->ep_ops.ep_tx_credits(htc->ar);
 		}
 		break;
 	}
diff --git a/drivers/net/wireless/ath/ath10k/htt_tx.c b/drivers/net/wireless/ath/ath10k/htt_tx.c
index c6c4b2a4d20f..d77379b3d3dc 100644
--- a/drivers/net/wireless/ath/ath10k/htt_tx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_tx.c
@@ -541,6 +541,11 @@ void ath10k_htt_tx_free(struct ath10k_htt *htt)
 	ath10k_htt_tx_destroy(htt);
 }
 
+void ath10k_htt_op_ep_tx_credits(struct ath10k *ar)
+{
+	queue_work(ar->workqueue, &ar->bundle_tx_work);
+}
+
 void ath10k_htt_htc_tx_complete(struct ath10k *ar, struct sk_buff *skb)
 {
 	struct ath10k_htt *htt = &ar->htt;
@@ -1365,7 +1370,7 @@ static int ath10k_htt_tx_hl(struct ath10k_htt *htt, enum ath10k_hw_txrx_mode txm
 	 */
 	tx_desc->peerid = __cpu_to_le32(HTT_INVALID_PEERID);
 
-	res = ath10k_htc_send(&htt->ar->htc, htt->eid, msdu);
+	res = ath10k_htc_send_hl(&htt->ar->htc, htt->eid, msdu);
 
 out:
 	return res;
-- 
2.23.0


_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

^ permalink raw reply related	[flat|nested] 10+ messages in thread
* [PATCH v7 3/3] ath10k: enable alt data of TX path for sdio
  2019-10-18  7:27 ` Wen Gong
@ 2019-10-18  7:27   ` Wen Gong
  -1 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-10-18  7:27 UTC (permalink / raw)
  To: ath10k; +Cc: linux-wireless

The default credit size is 1792 bytes, but the IP mtu is 1500 bytes,
then it has about 290 bytes's waste for each data packet on sdio
transfer path for TX bundle, it will reduce the transmission utilization
ratio for data packet.

This patch enable the small credit size in firmware, firmware will use
the new credit size 1556 bytes, it will increase the transmission
utilization ratio for data packet on TX patch. It results in significant
performance improvement on TX path.

This patch only effect sdio chip, it will not effect PCI, SNOC etc.

Tested with QCA6174 SDIO with firmware
WLAN.RMH.4.4.1-00017-QCARMSWP-1.

Signed-off-by: Wen Gong <wgong@codeaurora.org>
---
 drivers/net/wireless/ath/ath10k/core.c |  8 ++++----
 drivers/net/wireless/ath/ath10k/htc.c  | 12 ++++++++++--
 drivers/net/wireless/ath/ath10k/htc.h  | 13 +++++++++++--
 3 files changed, 25 insertions(+), 8 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/core.c b/drivers/net/wireless/ath/ath10k/core.c
index d9f1d5e75c86..72648bf7a613 100644
--- a/drivers/net/wireless/ath/ath10k/core.c
+++ b/drivers/net/wireless/ath/ath10k/core.c
@@ -696,10 +696,10 @@ static int ath10k_init_sdio(struct ath10k *ar, enum ath10k_firmware_mode mode)
 
 	param |= HI_ACS_FLAGS_SDIO_REDUCE_TX_COMPL_SET;
 
-	/* Alternate credit size of 1544 as used by SDIO firmware is
-	 * not big enough for mac80211 / native wifi frames. disable it
-	 */
-	param &= ~HI_ACS_FLAGS_ALT_DATA_CREDIT_SIZE;
+	if (mode == ATH10K_FIRMWARE_MODE_NORMAL)
+		param |= HI_ACS_FLAGS_ALT_DATA_CREDIT_SIZE;
+	else
+		param &= ~HI_ACS_FLAGS_ALT_DATA_CREDIT_SIZE;
 
 	if (mode == ATH10K_FIRMWARE_MODE_UTF)
 		param &= ~HI_ACS_FLAGS_SDIO_SWAP_MAILBOX_SET;
diff --git a/drivers/net/wireless/ath/ath10k/htc.c b/drivers/net/wireless/ath/ath10k/htc.c
index a2bf34e5a96e..03c358cd16d4 100644
--- a/drivers/net/wireless/ath/ath10k/htc.c
+++ b/drivers/net/wireless/ath/ath10k/htc.c
@@ -936,12 +936,16 @@ int ath10k_htc_wait_target(struct ath10k_htc *htc)
 	 */
 	if (htc->control_resp_len >=
 	    sizeof(msg->hdr) + sizeof(msg->ready_ext)) {
+		htc->alt_data_credit_size =
+			__le16_to_cpu(msg->ready_ext.reserved) &
+			ATH10K_HTC_MSG_READY_EXT_ALT_DATA_MASK;
 		htc->max_msgs_per_htc_bundle =
 			min_t(u8, msg->ready_ext.max_msgs_per_htc_bundle,
 			      HTC_HOST_MAX_MSG_PER_RX_BUNDLE);
 		ath10k_dbg(ar, ATH10K_DBG_HTC,
-			   "Extended ready message. RX bundle size: %d\n",
-			   htc->max_msgs_per_htc_bundle);
+			   "Extended ready message. RX bundle size: %d, alt size:%d\n",
+			   htc->max_msgs_per_htc_bundle,
+			   htc->alt_data_credit_size);
 	}
 
 	INIT_WORK(&ar->bundle_tx_work, ath10k_htc_bundle_tx_work);
@@ -1093,6 +1097,10 @@ int ath10k_htc_connect_service(struct ath10k_htc *htc,
 	ep->tx_credits = tx_alloc;
 	ep->tx_credit_size = htc->target_credit_size;
 
+	if (conn_req->service_id == ATH10K_HTC_SVC_ID_HTT_DATA_MSG &&
+	    htc->alt_data_credit_size != 0)
+		ep->tx_credit_size = htc->alt_data_credit_size;
+
 	/* copy all the callbacks */
 	ep->ep_ops = conn_req->ep_ops;
 
diff --git a/drivers/net/wireless/ath/ath10k/htc.h b/drivers/net/wireless/ath/ath10k/htc.h
index 7ce47cb29def..31640c44777c 100644
--- a/drivers/net/wireless/ath/ath10k/htc.h
+++ b/drivers/net/wireless/ath/ath10k/htc.h
@@ -100,6 +100,8 @@ enum ath10k_htc_conn_flags {
 #define ATH10K_HTC_CONN_FLAGS_RECV_ALLOC_LSB  8
 };
 
+#define ATH10K_HTC_MSG_READY_EXT_ALT_DATA_MASK 0xFFF
+
 enum ath10k_htc_conn_svc_status {
 	ATH10K_HTC_CONN_SVC_STATUS_SUCCESS      = 0,
 	ATH10K_HTC_CONN_SVC_STATUS_NOT_FOUND    = 1,
@@ -136,8 +138,14 @@ struct ath10k_htc_ready_extended {
 	struct ath10k_htc_ready base;
 	u8 htc_version; /* @enum ath10k_htc_version */
 	u8 max_msgs_per_htc_bundle;
-	u8 pad0;
-	u8 pad1;
+	union {
+		__le16 reserved;
+		struct {
+			u8 pad0;
+			u8 pad1;
+		} __packed;
+	} __packed;
+
 } __packed;
 
 struct ath10k_htc_conn_svc {
@@ -374,6 +382,7 @@ struct ath10k_htc {
 	int total_transmit_credits;
 	int target_credit_size;
 	u8 max_msgs_per_htc_bundle;
+	int alt_data_credit_size;
 };
 
 int ath10k_htc_init(struct ath10k *ar);
-- 
2.23.0


^ permalink raw reply related	[flat|nested] 10+ messages in thread
* [PATCH v7 3/3] ath10k: enable alt data of TX path for sdio
@ 2019-10-18  7:27   ` Wen Gong
  0 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-10-18  7:27 UTC (permalink / raw)
  To: ath10k; +Cc: linux-wireless

The default credit size is 1792 bytes, but the IP mtu is 1500 bytes,
then it has about 290 bytes's waste for each data packet on sdio
transfer path for TX bundle, it will reduce the transmission utilization
ratio for data packet.

This patch enable the small credit size in firmware, firmware will use
the new credit size 1556 bytes, it will increase the transmission
utilization ratio for data packet on TX patch. It results in significant
performance improvement on TX path.

This patch only effect sdio chip, it will not effect PCI, SNOC etc.

Tested with QCA6174 SDIO with firmware
WLAN.RMH.4.4.1-00017-QCARMSWP-1.

Signed-off-by: Wen Gong <wgong@codeaurora.org>
---
 drivers/net/wireless/ath/ath10k/core.c |  8 ++++----
 drivers/net/wireless/ath/ath10k/htc.c  | 12 ++++++++++--
 drivers/net/wireless/ath/ath10k/htc.h  | 13 +++++++++++--
 3 files changed, 25 insertions(+), 8 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/core.c b/drivers/net/wireless/ath/ath10k/core.c
index d9f1d5e75c86..72648bf7a613 100644
--- a/drivers/net/wireless/ath/ath10k/core.c
+++ b/drivers/net/wireless/ath/ath10k/core.c
@@ -696,10 +696,10 @@ static int ath10k_init_sdio(struct ath10k *ar, enum ath10k_firmware_mode mode)
 
 	param |= HI_ACS_FLAGS_SDIO_REDUCE_TX_COMPL_SET;
 
-	/* Alternate credit size of 1544 as used by SDIO firmware is
-	 * not big enough for mac80211 / native wifi frames. disable it
-	 */
-	param &= ~HI_ACS_FLAGS_ALT_DATA_CREDIT_SIZE;
+	if (mode == ATH10K_FIRMWARE_MODE_NORMAL)
+		param |= HI_ACS_FLAGS_ALT_DATA_CREDIT_SIZE;
+	else
+		param &= ~HI_ACS_FLAGS_ALT_DATA_CREDIT_SIZE;
 
 	if (mode == ATH10K_FIRMWARE_MODE_UTF)
 		param &= ~HI_ACS_FLAGS_SDIO_SWAP_MAILBOX_SET;
diff --git a/drivers/net/wireless/ath/ath10k/htc.c b/drivers/net/wireless/ath/ath10k/htc.c
index a2bf34e5a96e..03c358cd16d4 100644
--- a/drivers/net/wireless/ath/ath10k/htc.c
+++ b/drivers/net/wireless/ath/ath10k/htc.c
@@ -936,12 +936,16 @@ int ath10k_htc_wait_target(struct ath10k_htc *htc)
 	 */
 	if (htc->control_resp_len >=
 	    sizeof(msg->hdr) + sizeof(msg->ready_ext)) {
+		htc->alt_data_credit_size =
+			__le16_to_cpu(msg->ready_ext.reserved) &
+			ATH10K_HTC_MSG_READY_EXT_ALT_DATA_MASK;
 		htc->max_msgs_per_htc_bundle =
 			min_t(u8, msg->ready_ext.max_msgs_per_htc_bundle,
 			      HTC_HOST_MAX_MSG_PER_RX_BUNDLE);
 		ath10k_dbg(ar, ATH10K_DBG_HTC,
-			   "Extended ready message. RX bundle size: %d\n",
-			   htc->max_msgs_per_htc_bundle);
+			   "Extended ready message. RX bundle size: %d, alt size:%d\n",
+			   htc->max_msgs_per_htc_bundle,
+			   htc->alt_data_credit_size);
 	}
 
 	INIT_WORK(&ar->bundle_tx_work, ath10k_htc_bundle_tx_work);
@@ -1093,6 +1097,10 @@ int ath10k_htc_connect_service(struct ath10k_htc *htc,
 	ep->tx_credits = tx_alloc;
 	ep->tx_credit_size = htc->target_credit_size;
 
+	if (conn_req->service_id == ATH10K_HTC_SVC_ID_HTT_DATA_MSG &&
+	    htc->alt_data_credit_size != 0)
+		ep->tx_credit_size = htc->alt_data_credit_size;
+
 	/* copy all the callbacks */
 	ep->ep_ops = conn_req->ep_ops;
 
diff --git a/drivers/net/wireless/ath/ath10k/htc.h b/drivers/net/wireless/ath/ath10k/htc.h
index 7ce47cb29def..31640c44777c 100644
--- a/drivers/net/wireless/ath/ath10k/htc.h
+++ b/drivers/net/wireless/ath/ath10k/htc.h
@@ -100,6 +100,8 @@ enum ath10k_htc_conn_flags {
 #define ATH10K_HTC_CONN_FLAGS_RECV_ALLOC_LSB  8
 };
 
+#define ATH10K_HTC_MSG_READY_EXT_ALT_DATA_MASK 0xFFF
+
 enum ath10k_htc_conn_svc_status {
 	ATH10K_HTC_CONN_SVC_STATUS_SUCCESS      = 0,
 	ATH10K_HTC_CONN_SVC_STATUS_NOT_FOUND    = 1,
@@ -136,8 +138,14 @@ struct ath10k_htc_ready_extended {
 	struct ath10k_htc_ready base;
 	u8 htc_version; /* @enum ath10k_htc_version */
 	u8 max_msgs_per_htc_bundle;
-	u8 pad0;
-	u8 pad1;
+	union {
+		__le16 reserved;
+		struct {
+			u8 pad0;
+			u8 pad1;
+		} __packed;
+	} __packed;
+
 } __packed;
 
 struct ath10k_htc_conn_svc {
@@ -374,6 +382,7 @@ struct ath10k_htc {
 	int total_transmit_credits;
 	int target_credit_size;
 	u8 max_msgs_per_htc_bundle;
+	int alt_data_credit_size;
 };
 
 int ath10k_htc_init(struct ath10k *ar);
-- 
2.23.0


_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

^ permalink raw reply related	[flat|nested] 10+ messages in thread
* RE: [PATCH v7 0/3] ath10k: improve throughout of TX of sdio
  2019-10-18  7:27 ` Wen Gong
@ 2019-12-02  8:49   ` Wen Gong
  -1 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-12-02  8:49 UTC (permalink / raw)
  To: Wen Gong, ath10k; +Cc: linux-wireless

> -----Original Message-----
> From: ath10k <ath10k-bounces@lists.infradead.org> On Behalf Of Wen Gong
> Sent: Friday, October 18, 2019 3:28 PM
> To: ath10k@lists.infradead.org
> Cc: linux-wireless@vger.kernel.org
> Subject: [EXT] [PATCH v7 0/3] ath10k: improve throughout of TX of sdio
> 
V8 sent:
https://patchwork.kernel.org/cover/11265701/
[v8,0/4] ath10k: improve throughout of TX of sdio 
[v8,1/4] ath10k: disable TX complete indication of htt for sdio  
[v8,2/4] ath10k: change ATH10K_SDIO_BUS_REQUEST_MAX_NUM from 64 to 1024  
[v8,3/4] ath10k: add htt TX bundle for sdio  
[v8,4/4] ath10k: enable alt data of TX path for sdio

^ permalink raw reply	[flat|nested] 10+ messages in thread
* RE: [PATCH v7 0/3] ath10k: improve throughout of TX of sdio
@ 2019-12-02  8:49   ` Wen Gong
  0 siblings, 0 replies; 10+ messages in thread
From: Wen Gong @ 2019-12-02  8:49 UTC (permalink / raw)
  To: Wen Gong, ath10k; +Cc: linux-wireless

> -----Original Message-----
> From: ath10k <ath10k-bounces@lists.infradead.org> On Behalf Of Wen Gong
> Sent: Friday, October 18, 2019 3:28 PM
> To: ath10k@lists.infradead.org
> Cc: linux-wireless@vger.kernel.org
> Subject: [EXT] [PATCH v7 0/3] ath10k: improve throughout of TX of sdio
> 
V8 sent:
https://patchwork.kernel.org/cover/11265701/
[v8,0/4] ath10k: improve throughout of TX of sdio 
[v8,1/4] ath10k: disable TX complete indication of htt for sdio  
[v8,2/4] ath10k: change ATH10K_SDIO_BUS_REQUEST_MAX_NUM from 64 to 1024  
[v8,3/4] ath10k: add htt TX bundle for sdio  
[v8,4/4] ath10k: enable alt data of TX path for sdio

_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

^ permalink raw reply	[flat|nested] 10+ messages in thread
end of thread, other threads:[~2019-12-02  8:55 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-10-18  7:27 [PATCH v7 0/3] ath10k: improve throughout of TX of sdio Wen Gong
2019-10-18  7:27 ` Wen Gong
2019-10-18  7:27 ` [PATCH v7 1/3] ath10k: disable TX complete indication of htt for sdio Wen Gong
2019-10-18  7:27   ` Wen Gong
2019-10-18  7:27 ` [PATCH v7 2/3] ath10k: add htt TX bundle " Wen Gong
2019-10-18  7:27   ` Wen Gong
2019-10-18  7:27 ` [PATCH v7 3/3] ath10k: enable alt data of TX path " Wen Gong
2019-10-18  7:27   ` Wen Gong
2019-12-02  8:49 ` [PATCH v7 0/3] ath10k: improve throughout of TX of sdio Wen Gong
2019-12-02  8:49   ` Wen Gong

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.