netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Xi Wang <xii@google.com>
To: "David S. Miller" <davem@davemloft.net>, netdev@vger.kernel.org
Cc: Jason Wang <jasowang@redhat.com>,
	"Michael S. Tsirkin" <mst@redhat.com>,
	Maxim Krasnyansky <maxk@qti.qualcomm.com>,
	Neal Cardwell <ncardwell@google.com>,
	Eric Dumazet <edumazet@google.com>, Xi Wang <xii@google.com>
Subject: [PATCH v2] net-tun: restructure tun_do_read for better sleep/wakeup efficiency
Date: Fri, 16 May 2014 15:11:48 -0700	[thread overview]
Message-ID: <1400278308-25372-1-git-send-email-xii@google.com> (raw)

tun_do_read always adds current thread to wait queue, even if a packet
is ready to read. This is inefficient because both sleeper and waker
want to acquire the wait queue spin lock when packet rate is high.

We restructure the read function and use common kernel networking
routines to handle receive, sleep and wakeup. With the change
available packets are checked first before the reading thread is added
to the wait queue.

Ran performance tests with the following configuration:

 - my packet generator -> tap1 -> br0 -> tap0 -> my packet consumer
 - sender pinned to one core and receiver pinned to another core
 - sender send small UDP packets (64 bytes total) as fast as it can
 - sandy bridge cores
 - throughput are receiver side goodput numbers

The results are

baseline: 731k pkts/sec, cpu utilization at 1.50 cpus
 changed: 783k pkts/sec, cpu utilization at 1.53 cpus

The performance difference is largely determined by packet rate and
inter-cpu communication cost. For example, if the sender and
receiver are pinned to different cpu sockets, the results are

baseline: 558k pkts/sec, cpu utilization at 1.71 cpus
 changed: 690k pkts/sec, cpu utilization at 1.67 cpus

Co-authored-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Xi Wang <xii@google.com>
---

Changelog since v1:
 - Added back error code. NETREG_REGISTERED behavior is different but
   should be compatible with the previous implementation
 - Removed non essential changes


 drivers/net/tun.c | 54 ++++++++++++++++--------------------------------------
 1 file changed, 16 insertions(+), 38 deletions(-)

diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index ee328ba..98bad1f 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -498,12 +498,12 @@ static void tun_detach_all(struct net_device *dev)
 	for (i = 0; i < n; i++) {
 		tfile = rtnl_dereference(tun->tfiles[i]);
 		BUG_ON(!tfile);
-		wake_up_all(&tfile->wq.wait);
+		tfile->socket.sk->sk_data_ready(tfile->socket.sk);
 		RCU_INIT_POINTER(tfile->tun, NULL);
 		--tun->numqueues;
 	}
 	list_for_each_entry(tfile, &tun->disabled, next) {
-		wake_up_all(&tfile->wq.wait);
+		tfile->socket.sk->sk_data_ready(tfile->socket.sk);
 		RCU_INIT_POINTER(tfile->tun, NULL);
 	}
 	BUG_ON(tun->numqueues != 0);
@@ -807,8 +807,7 @@ static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
 	/* Notify and wake up reader process */
 	if (tfile->flags & TUN_FASYNC)
 		kill_fasync(&tfile->fasync, SIGIO, POLL_IN);
-	wake_up_interruptible_poll(&tfile->wq.wait, POLLIN |
-				   POLLRDNORM | POLLRDBAND);
+	tfile->socket.sk->sk_data_ready(tfile->socket.sk);
 
 	rcu_read_unlock();
 	return NETDEV_TX_OK;
@@ -965,7 +964,7 @@ static unsigned int tun_chr_poll(struct file *file, poll_table *wait)
 
 	tun_debug(KERN_INFO, tun, "tun_chr_poll\n");
 
-	poll_wait(file, &tfile->wq.wait, wait);
+	poll_wait(file, sk_sleep(sk), wait);
 
 	if (!skb_queue_empty(&sk->sk_receive_queue))
 		mask |= POLLIN | POLLRDNORM;
@@ -1330,47 +1329,26 @@ done:
 static ssize_t tun_do_read(struct tun_struct *tun, struct tun_file *tfile,
 			   const struct iovec *iv, ssize_t len, int noblock)
 {
-	DECLARE_WAITQUEUE(wait, current);
 	struct sk_buff *skb;
 	ssize_t ret = 0;
+	int peeked, err, off = 0;
 
 	tun_debug(KERN_INFO, tun, "tun_do_read\n");
 
-	if (unlikely(!noblock))
-		add_wait_queue(&tfile->wq.wait, &wait);
-	while (len) {
-		if (unlikely(!noblock))
-			current->state = TASK_INTERRUPTIBLE;
+	if (!len)
+		return ret;
 
-		/* Read frames from the queue */
-		if (!(skb = skb_dequeue(&tfile->socket.sk->sk_receive_queue))) {
-			if (noblock) {
-				ret = -EAGAIN;
-				break;
-			}
-			if (signal_pending(current)) {
-				ret = -ERESTARTSYS;
-				break;
-			}
-			if (tun->dev->reg_state != NETREG_REGISTERED) {
-				ret = -EIO;
-				break;
-			}
-
-			/* Nothing to read, let's sleep */
-			schedule();
-			continue;
-		}
+	if (tun->dev->reg_state != NETREG_REGISTERED)
+		return -EIO;
 
+	/* Read frames from queue */
+	skb = __skb_recv_datagram(tfile->socket.sk, noblock ? MSG_DONTWAIT : 0,
+				  &peeked, &off, &err);
+	if (skb) {
 		ret = tun_put_user(tun, tfile, skb, iv, len);
 		kfree_skb(skb);
-		break;
-	}
-
-	if (unlikely(!noblock)) {
-		current->state = TASK_RUNNING;
-		remove_wait_queue(&tfile->wq.wait, &wait);
-	}
+	} else
+		ret = err;
 
 	return ret;
 }
@@ -2199,8 +2177,8 @@ static int tun_chr_open(struct inode *inode, struct file * file)
 	tfile->flags = 0;
 	tfile->ifindex = 0;
 
-	rcu_assign_pointer(tfile->socket.wq, &tfile->wq);
 	init_waitqueue_head(&tfile->wq.wait);
+	RCU_INIT_POINTER(tfile->socket.wq, &tfile->wq);
 
 	tfile->socket.file = file;
 	tfile->socket.ops = &tun_socket_ops;
-- 
1.9.1.423.g4596e3a

             reply	other threads:[~2014-05-16 22:12 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-05-16 22:11 Xi Wang [this message]
2014-05-19  9:27 ` [PATCH v2] net-tun: restructure tun_do_read for better sleep/wakeup efficiency Jason Wang
2014-05-19 14:09   ` Eric Dumazet
2014-05-20  4:44     ` Jason Wang
2014-05-20  4:52       ` Eric Dumazet
2014-05-20  6:35         ` Michael S. Tsirkin
2014-05-20  5:11       ` Eric Dumazet
2014-05-20  6:03         ` Jason Wang
2014-05-20  6:34           ` Michael S. Tsirkin
2014-05-20  6:55             ` Jason Wang
2014-05-20 13:59           ` Eric Dumazet
2014-05-21  4:45             ` Jason Wang
2014-05-19 16:06   ` Michael S. Tsirkin
2014-05-20  4:51     ` Jason Wang
2014-05-20  6:22       ` Michael S. Tsirkin
2014-05-20  6:40         ` Jason Wang
2014-05-21  7:54 ` Michael S. Tsirkin
2014-05-21 19:51 ` David Miller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1400278308-25372-1-git-send-email-xii@google.com \
    --to=xii@google.com \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=jasowang@redhat.com \
    --cc=maxk@qti.qualcomm.com \
    --cc=mst@redhat.com \
    --cc=ncardwell@google.com \
    --cc=netdev@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).