linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH net-next v2 0/3] hinic3: queue_api related fixes
@ 2025-05-22  6:54 Gur Stavi
  2025-05-22  6:54 ` [PATCH net-next v2 1/3] queue_api: add subqueue variant netif_subqueue_sent Gur Stavi
                   ` (3 more replies)
  0 siblings, 4 replies; 6+ messages in thread
From: Gur Stavi @ 2025-05-22  6:54 UTC (permalink / raw)
  To: Gur Stavi
  Cc: netdev, linux-kernel, David S. Miller, Eric Dumazet,
	Jakub Kicinski, Paolo Abeni, Simon Horman, Fan Gong

This patch series contains improvement to queue_api and 2 queue_api
related patches to the hinic3 driver.

Changes:

v1: http://lore.kernel.org/netdev/cover.1747824040.git.gur.stavi@huawei.com

v2:
* Update cover letter subject and text.
* Add 2 patches for user code related to queue api.

Gur Stavi (3):
  queue_api: add subqueue variant netif_subqueue_sent
  hinic3: use netif_subqueue_sent api
  hinic3: remove tx_q name collision hack

 .../net/ethernet/huawei/hinic3/hinic3_tx.c    | 23 ++++++++-----------
 include/net/netdev_queues.h                   |  8 +++++++
 2 files changed, 18 insertions(+), 13 deletions(-)


base-commit: 3da895b23901964fcf23450f10b529d45069f333
-- 
2.45.2


^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH net-next v2 1/3] queue_api: add subqueue variant netif_subqueue_sent
  2025-05-22  6:54 [PATCH net-next v2 0/3] hinic3: queue_api related fixes Gur Stavi
@ 2025-05-22  6:54 ` Gur Stavi
  2025-05-27 10:06   ` Paolo Abeni
  2025-05-22  6:54 ` [PATCH net-next v2 2/3] hinic3: use netif_subqueue_sent api Gur Stavi
                   ` (2 subsequent siblings)
  3 siblings, 1 reply; 6+ messages in thread
From: Gur Stavi @ 2025-05-22  6:54 UTC (permalink / raw)
  To: Gur Stavi
  Cc: netdev, linux-kernel, David S. Miller, Eric Dumazet,
	Jakub Kicinski, Paolo Abeni, Simon Horman, Fan Gong

Add a new macro, netif_subqueue_sent, which is a wrapper for
netdev_tx_sent_queue.

Drivers that use the subqueue variant macros, netif_subqueue_xxx,
identify queue by index and are not required to obtain
struct netdev_queue explicitly.

Such drivers still need to call netdev_tx_sent_queue which is a
counterpart of netif_subqueue_completed_wake. Allowing drivers to use a
subqueue variant for this purpose improves their code consistency by
always referring to queue by its index.

Signed-off-by: Gur Stavi <gur.stavi@huawei.com>
---
 include/net/netdev_queues.h | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/include/net/netdev_queues.h b/include/net/netdev_queues.h
index ba2eaf39089b..7b6656ee549f 100644
--- a/include/net/netdev_queues.h
+++ b/include/net/netdev_queues.h
@@ -294,6 +294,14 @@ netdev_txq_completed_mb(struct netdev_queue *dev_queue,
 		netif_txq_try_stop(_txq, get_desc, start_thrs);		\
 	})
 
+#define netif_subqueue_sent(dev, idx, bytes)				\
+	({								\
+		struct netdev_queue *_txq;				\
+									\
+		_txq = netdev_get_tx_queue(dev, idx);			\
+		netdev_tx_sent_queue(_txq, bytes);			\
+	})
+
 #define netif_subqueue_maybe_stop(dev, idx, get_desc, stop_thrs, start_thrs) \
 	({								\
 		struct netdev_queue *_txq;				\
-- 
2.45.2


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* [PATCH net-next v2 2/3] hinic3: use netif_subqueue_sent api
  2025-05-22  6:54 [PATCH net-next v2 0/3] hinic3: queue_api related fixes Gur Stavi
  2025-05-22  6:54 ` [PATCH net-next v2 1/3] queue_api: add subqueue variant netif_subqueue_sent Gur Stavi
@ 2025-05-22  6:54 ` Gur Stavi
  2025-05-22  6:54 ` [PATCH net-next v2 3/3] hinic3: remove tx_q name collision hack Gur Stavi
  2025-05-22 16:58 ` [PATCH net-next v2 0/3] hinic3: queue_api related fixes Jakub Kicinski
  3 siblings, 0 replies; 6+ messages in thread
From: Gur Stavi @ 2025-05-22  6:54 UTC (permalink / raw)
  To: Gur Stavi
  Cc: netdev, linux-kernel, David S. Miller, Eric Dumazet,
	Jakub Kicinski, Paolo Abeni, Simon Horman, Fan Gong

Improve consistency of code by using only netif_subqueue variant apis

Signed-off-by: Gur Stavi <gur.stavi@huawei.com>
---
 drivers/net/ethernet/huawei/hinic3/hinic3_tx.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/huawei/hinic3/hinic3_tx.c b/drivers/net/ethernet/huawei/hinic3/hinic3_tx.c
index ae08257dd1d2..7b6f101da313 100644
--- a/drivers/net/ethernet/huawei/hinic3/hinic3_tx.c
+++ b/drivers/net/ethernet/huawei/hinic3/hinic3_tx.c
@@ -542,8 +542,7 @@ static netdev_tx_t hinic3_send_one_skb(struct sk_buff *skb,
 		goto err_drop_pkt;
 	}
 
-	netdev_tx_sent_queue(netdev_get_tx_queue(netdev, txq->sq->q_id),
-			     skb->len);
+	netif_subqueue_sent(netdev, txq->sq->q_id, skb->len);
 	netif_subqueue_maybe_stop(netdev, tx_q->sq->q_id,
 				  hinic3_wq_free_wqebbs(&tx_q->sq->wq),
 				  tx_q->tx_stop_thrs,
-- 
2.45.2


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* [PATCH net-next v2 3/3] hinic3: remove tx_q name collision hack
  2025-05-22  6:54 [PATCH net-next v2 0/3] hinic3: queue_api related fixes Gur Stavi
  2025-05-22  6:54 ` [PATCH net-next v2 1/3] queue_api: add subqueue variant netif_subqueue_sent Gur Stavi
  2025-05-22  6:54 ` [PATCH net-next v2 2/3] hinic3: use netif_subqueue_sent api Gur Stavi
@ 2025-05-22  6:54 ` Gur Stavi
  2025-05-22 16:58 ` [PATCH net-next v2 0/3] hinic3: queue_api related fixes Jakub Kicinski
  3 siblings, 0 replies; 6+ messages in thread
From: Gur Stavi @ 2025-05-22  6:54 UTC (permalink / raw)
  To: Gur Stavi
  Cc: netdev, linux-kernel, David S. Miller, Eric Dumazet,
	Jakub Kicinski, Paolo Abeni, Simon Horman, Fan Gong

A local variable of tx_q worked around name collision with internal
txq variable in netif_subqueue macros.
This workaround is no longer needed.

Signed-off-by: Gur Stavi <gur.stavi@huawei.com>
---
 .../net/ethernet/huawei/hinic3/hinic3_tx.c    | 20 +++++++++----------
 1 file changed, 9 insertions(+), 11 deletions(-)

diff --git a/drivers/net/ethernet/huawei/hinic3/hinic3_tx.c b/drivers/net/ethernet/huawei/hinic3/hinic3_tx.c
index 7b6f101da313..3f7f73430be4 100644
--- a/drivers/net/ethernet/huawei/hinic3/hinic3_tx.c
+++ b/drivers/net/ethernet/huawei/hinic3/hinic3_tx.c
@@ -482,7 +482,6 @@ static netdev_tx_t hinic3_send_one_skb(struct sk_buff *skb,
 {
 	struct hinic3_sq_wqe_combo wqe_combo = {};
 	struct hinic3_tx_info *tx_info;
-	struct hinic3_txq *tx_q = txq;
 	u32 offload, queue_info = 0;
 	struct hinic3_sq_task task;
 	u16 wqebb_cnt, num_sge;
@@ -506,9 +505,9 @@ static netdev_tx_t hinic3_send_one_skb(struct sk_buff *skb,
 		if (likely(wqebb_cnt > txq->tx_stop_thrs))
 			txq->tx_stop_thrs = min(wqebb_cnt, txq->tx_start_thrs);
 
-		netif_subqueue_try_stop(netdev, tx_q->sq->q_id,
-					hinic3_wq_free_wqebbs(&tx_q->sq->wq),
-					tx_q->tx_start_thrs);
+		netif_subqueue_try_stop(netdev, txq->sq->q_id,
+					hinic3_wq_free_wqebbs(&txq->sq->wq),
+					txq->tx_start_thrs);
 
 		return NETDEV_TX_BUSY;
 	}
@@ -543,10 +542,10 @@ static netdev_tx_t hinic3_send_one_skb(struct sk_buff *skb,
 	}
 
 	netif_subqueue_sent(netdev, txq->sq->q_id, skb->len);
-	netif_subqueue_maybe_stop(netdev, tx_q->sq->q_id,
-				  hinic3_wq_free_wqebbs(&tx_q->sq->wq),
-				  tx_q->tx_stop_thrs,
-				  tx_q->tx_start_thrs);
+	netif_subqueue_maybe_stop(netdev, txq->sq->q_id,
+				  hinic3_wq_free_wqebbs(&txq->sq->wq),
+				  txq->tx_stop_thrs,
+				  txq->tx_start_thrs);
 
 	hinic3_prepare_sq_ctrl(&wqe_combo, queue_info, num_sge, owner);
 	hinic3_write_db(txq->sq, 0, DB_CFLAG_DP_SQ,
@@ -630,7 +629,6 @@ bool hinic3_tx_poll(struct hinic3_txq *txq, int budget)
 	struct net_device *netdev = txq->netdev;
 	u16 hw_ci, sw_ci, q_id = txq->sq->q_id;
 	struct hinic3_tx_info *tx_info;
-	struct hinic3_txq *tx_q = txq;
 	unsigned int bytes_compl = 0;
 	unsigned int pkts = 0;
 	u16 wqebb_cnt = 0;
@@ -662,8 +660,8 @@ bool hinic3_tx_poll(struct hinic3_txq *txq, int budget)
 	hinic3_wq_put_wqebbs(&txq->sq->wq, wqebb_cnt);
 
 	netif_subqueue_completed_wake(netdev, q_id, pkts, bytes_compl,
-				      hinic3_wq_free_wqebbs(&tx_q->sq->wq),
-				      tx_q->tx_start_thrs);
+				      hinic3_wq_free_wqebbs(&txq->sq->wq),
+				      txq->tx_start_thrs);
 
 	return pkts == HINIC3_TX_POLL_WEIGHT;
 }
-- 
2.45.2


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH net-next v2 0/3] hinic3: queue_api related fixes
  2025-05-22  6:54 [PATCH net-next v2 0/3] hinic3: queue_api related fixes Gur Stavi
                   ` (2 preceding siblings ...)
  2025-05-22  6:54 ` [PATCH net-next v2 3/3] hinic3: remove tx_q name collision hack Gur Stavi
@ 2025-05-22 16:58 ` Jakub Kicinski
  3 siblings, 0 replies; 6+ messages in thread
From: Jakub Kicinski @ 2025-05-22 16:58 UTC (permalink / raw)
  To: Gur Stavi
  Cc: netdev, linux-kernel, David S. Miller, Eric Dumazet, Paolo Abeni,
	Simon Horman, Fan Gong

On Thu, 22 May 2025 09:54:40 +0300 Gur Stavi wrote:
> This patch series contains improvement to queue_api and 2 queue_api
> related patches to the hinic3 driver.
> 
> Changes:
> 
> v1: http://lore.kernel.org/netdev/cover.1747824040.git.gur.stavi@huawei.com

Reviewed-by: Jakub Kicinski <kuba@kernel.org>

Thanks!

reminder: please respect the 24h reposting delay per 
https://www.kernel.org/doc/html/next/process/maintainer-netdev.html#resending-after-review
-- 
pv-bot: 24h

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH net-next v2 1/3] queue_api: add subqueue variant netif_subqueue_sent
  2025-05-22  6:54 ` [PATCH net-next v2 1/3] queue_api: add subqueue variant netif_subqueue_sent Gur Stavi
@ 2025-05-27 10:06   ` Paolo Abeni
  0 siblings, 0 replies; 6+ messages in thread
From: Paolo Abeni @ 2025-05-27 10:06 UTC (permalink / raw)
  To: Gur Stavi
  Cc: netdev, linux-kernel, David S. Miller, Eric Dumazet,
	Jakub Kicinski, Simon Horman, Fan Gong

On 5/22/25 8:54 AM, Gur Stavi wrote:
> diff --git a/include/net/netdev_queues.h b/include/net/netdev_queues.h
> index ba2eaf39089b..7b6656ee549f 100644
> --- a/include/net/netdev_queues.h
> +++ b/include/net/netdev_queues.h
> @@ -294,6 +294,14 @@ netdev_txq_completed_mb(struct netdev_queue *dev_queue,
>  		netif_txq_try_stop(_txq, get_desc, start_thrs);		\
>  	})
>  
> +#define netif_subqueue_sent(dev, idx, bytes)				\
> +	({								\
> +		struct netdev_queue *_txq;				\
> +									\
> +		_txq = netdev_get_tx_queue(dev, idx);			\
> +		netdev_tx_sent_queue(_txq, bytes);			\
> +	})

The above could be (and should be) a static inline function. Note that
(most of) the existing queue API helper are macros to allow quering the
tx ring without using indirect calls/function pointers.

Thanks,

Paolo


^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2025-05-27 10:07 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-05-22  6:54 [PATCH net-next v2 0/3] hinic3: queue_api related fixes Gur Stavi
2025-05-22  6:54 ` [PATCH net-next v2 1/3] queue_api: add subqueue variant netif_subqueue_sent Gur Stavi
2025-05-27 10:06   ` Paolo Abeni
2025-05-22  6:54 ` [PATCH net-next v2 2/3] hinic3: use netif_subqueue_sent api Gur Stavi
2025-05-22  6:54 ` [PATCH net-next v2 3/3] hinic3: remove tx_q name collision hack Gur Stavi
2025-05-22 16:58 ` [PATCH net-next v2 0/3] hinic3: queue_api related fixes Jakub Kicinski

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).