Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   David Wu <david.wu@rock-chips.com>
To:     joabreu@synopsys.com, mcoquelin.stm32@gmail.com,
        alexandre.torgue@foss.st.com, davem@davemloft.net
Cc:     netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        David Wu <david.wu@rock-chips.com>
Subject: [PATCH RFC net-next] net: stmmac: Reduce the memory copy for rx data
Date:   Thu, 30 Jun 2022 14:24:53 +0800
Message-Id: <20220630062453.2530477-1-david.wu@rock-chips.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

Use skb_add_rx_frag() to reduce the memory copy for rx data, only copy
ethernet header.

Signed-off-by: David Wu <david.wu@rock-chips.com>
---
 drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index d1a7cf4567bc..ac8d92ec35c0 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -5226,7 +5226,7 @@ static int stmmac_rx(struct stmmac_priv *priv, int limit, u32 queue)
 			/* XDP program may expand or reduce tail */
 			buf1_len = xdp.data_end - xdp.data;
 
-			skb = napi_alloc_skb(&ch->rx_napi, buf1_len);
+			skb = napi_alloc_skb(&ch->rx_napi, ETH_HLEN);
 			if (!skb) {
 				priv->dev->stats.rx_dropped++;
 				count++;
@@ -5234,11 +5234,15 @@ static int stmmac_rx(struct stmmac_priv *priv, int limit, u32 queue)
 			}
 
 			/* XDP program may adjust header */
-			skb_copy_to_linear_data(skb, xdp.data, buf1_len);
-			skb_put(skb, buf1_len);
+			memcpy(skb->data, xdp.data, ETH_HLEN);
+			skb_put(skb, ETH_HLEN);
+			skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags,
+					buf->page, buf->page_offset + ETH_HLEN,
+					buf1_len - ETH_HLEN,
+					priv->dma_buf_sz);
 
-			/* Data payload copied into SKB, page ready for recycle */
-			page_pool_recycle_direct(rx_q->page_pool, buf->page);
+			/* Data payload appended into SKB */
+			page_pool_release_page(rx_q->page_pool, buf->page);
 			buf->page = NULL;
 		} else if (buf1_len) {
 			dma_sync_single_for_cpu(priv->device, buf->addr,
-- 
2.25.1