netdev - [PATCH net-next v3 9/9] xsk: support dynamic xmit.more control for batch xmit

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-Id: <20251021131209.41491-10-kerneljasonxing@gmail.com>
Date: Tue, 21 Oct 2025 21:12:09 +0800
From: Jason Xing <kerneljasonxing@...il.com>
To: davem@...emloft.net,
	edumazet@...gle.com,
	kuba@...nel.org,
	pabeni@...hat.com,
	bjorn@...nel.org,
	magnus.karlsson@...el.com,
	maciej.fijalkowski@...el.com,
	jonathan.lemon@...il.com,
	sdf@...ichev.me,
	ast@...nel.org,
	daniel@...earbox.net,
	hawk@...nel.org,
	john.fastabend@...il.com,
	joe@...a.to,
	willemdebruijn.kernel@...il.com
Cc: bpf@...r.kernel.org,
	netdev@...r.kernel.org,
	Jason Xing <kernelxing@...cent.com>
Subject: [PATCH net-next v3 9/9] xsk: support dynamic xmit.more control for batch xmit

From: Jason Xing <kernelxing@...cent.com>

Only set xmit.more false for the last skb.

In theory, only making xmit.more false for the last packets to be
sent in each round can bring much benefit like avoid triggering too
many irqs.

Compared to the numbers for batch mode, a huge improvement (26%) can
be seen on i40e/ixgbe driver since the cost of triggering irqs is
expensive.

Suggested-by: Jesper Dangaard Brouer <hawk@...nel.org>
Signed-off-by: Jason Xing <kernelxing@...cent.com>
---
 net/core/dev.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/net/core/dev.c b/net/core/dev.c
index 32de76c79d29..549a95b9d96f 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -4797,14 +4797,18 @@ int xsk_direct_xmit_batch(struct xdp_sock *xs, struct net_device *dev)
 {
 	u16 queue_id = xs->queue_id;
 	struct netdev_queue *txq = netdev_get_tx_queue(dev, queue_id);
+	struct sk_buff_head *send_queue = &xs->batch.send_queue;
 	int ret = NETDEV_TX_BUSY;
 	struct sk_buff *skb;
+	bool more = true;
 
 	local_bh_disable();
 	HARD_TX_LOCK(dev, txq, smp_processor_id());
-	while ((skb = __skb_dequeue(&xs->batch.send_queue)) != NULL) {
+	while ((skb = __skb_dequeue(send_queue)) != NULL) {
+		if (!skb_peek(send_queue))
+			more = false;
 		skb_set_queue_mapping(skb, queue_id);
-		ret = netdev_start_xmit(skb, dev, txq, false);
+		ret = netdev_start_xmit(skb, dev, txq, more);
 		if (ret != NETDEV_TX_OK)
 			break;
 	}
-- 
2.41.3