[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1571210231-29154-8-git-send-email-tanhuazhong@huawei.com>
Date: Wed, 16 Oct 2019 15:17:06 +0800
From: Huazhong Tan <tanhuazhong@...wei.com>
To: <davem@...emloft.net>
CC: <netdev@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
<salil.mehta@...wei.com>, <yisen.zhuang@...wei.com>,
<linuxarm@...wei.com>, <jakub.kicinski@...ronome.com>,
Yunsheng Lin <linyunsheng@...wei.com>,
Huazhong Tan <tanhuazhong@...wei.com>
Subject: [PATCH net-next 07/12] net: hns3: minor optimization for barrier in IO path
From: Yunsheng Lin <linyunsheng@...wei.com>
Currently, the TX and RX ring in a queue is bounded to the
same IRQ, there may be unnecessary barrier op when only one of
the ring need to be processed.
This patch adjusts the location of rmb() in hns3_clean_tx_ring()
and adds a checking in hns3_clean_rx_ring() to avoid unnecessary
barrier op when there is nothing to do for the ring.
Signed-off-by: Yunsheng Lin <linyunsheng@...wei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@...wei.com>
---
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index 5a237f94..a0810b3 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -2488,11 +2488,12 @@ void hns3_clean_tx_ring(struct hns3_enet_ring *ring)
int head;
head = readl_relaxed(ring->tqp->io_base + HNS3_RING_TX_RING_HEAD_REG);
- rmb(); /* Make sure head is ready before touch any data */
if (is_ring_empty(ring) || head == ring->next_to_clean)
return; /* no data to poll */
+ rmb(); /* Make sure head is ready before touch any data */
+
if (unlikely(!is_valid_clean_head(ring, head))) {
netdev_err(netdev, "wrong head (%d, %d-%d)\n", head,
ring->next_to_use, ring->next_to_clean);
@@ -3108,11 +3109,14 @@ int hns3_clean_rx_ring(struct hns3_enet_ring *ring, int budget,
int err, num;
num = readl_relaxed(ring->tqp->io_base + HNS3_RING_RX_RING_FBDNUM_REG);
- rmb(); /* Make sure num taken effect before the other data is touched */
-
num -= unused_count;
unused_count -= ring->pending_buf;
+ if (num <= 0)
+ goto out;
+
+ rmb(); /* Make sure num taken effect before the other data is touched */
+
while (recv_pkts < budget && recv_bds < num) {
/* Reuse or realloc buffers */
if (unused_count >= RCB_NOF_ALLOC_RX_BUFF_ONCE) {
--
2.7.4
Powered by blists - more mailing lists