mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Andrey Ryabinin <arbn@yandex-team.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>,
	Stefan Hajnoczi <stefanha@redhat.com>,
	Stefano Garzarella <sgarzare@redhat.com>,
	Andrey Ryabinin <arbn@yandex-team.com>,
	Alexei Starovoitov <ast@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	"David S. Miller" <davem@davemloft.net>,
	Jakub Kicinski <kuba@kernel.org>,
	Jesper Dangaard Brouer <hawk@kernel.org>,
	John Fastabend <john.fastabend@gmail.com>,
	kvm@vger.kernel.org, virtualization@lists.linux-foundation.org,
	netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
	bpf@vger.kernel.org
Subject: [PATCH 6/6] vhost_net: use RCU callbacks instead of synchronize_rcu()
Date: Mon, 15 Nov 2021 18:30:03 +0300	[thread overview]
Message-ID: <20211115153003.9140-6-arbn@yandex-team.com> (raw)
In-Reply-To: <20211115153003.9140-1-arbn@yandex-team.com>

Currently vhost_net_release() uses synchronize_rcu() to synchronize
freeing with vhost_zerocopy_callback(). However synchronize_rcu()
is quite costly operation. It take more than 10 seconds
to shutdown qemu launched with couple net devices like this:
	-netdev tap,id=tap0,..,vhost=on,queues=80
because we end up calling synchronize_rcu() netdev_count*queues times.

Free vhost net structures in rcu callback instead of using
synchronize_rcu() to fix the problem.

Signed-off-by: Andrey Ryabinin <arbn@yandex-team.com>
---
 drivers/vhost/net.c | 22 ++++++++++++++--------
 1 file changed, 14 insertions(+), 8 deletions(-)

diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
index 97a209d6a527..0699d30e83d5 100644
--- a/drivers/vhost/net.c
+++ b/drivers/vhost/net.c
@@ -132,6 +132,7 @@ struct vhost_net {
 	struct vhost_dev dev;
 	struct vhost_net_virtqueue vqs[VHOST_NET_VQ_MAX];
 	struct vhost_poll poll[VHOST_NET_VQ_MAX];
+	struct rcu_head rcu;
 	/* Number of TX recently submitted.
 	 * Protected by tx vq lock. */
 	unsigned tx_packets;
@@ -1389,6 +1390,18 @@ static void vhost_net_flush(struct vhost_net *n)
 	}
 }
 
+static void vhost_net_free(struct rcu_head *rcu_head)
+{
+	struct vhost_net *n = container_of(rcu_head, struct vhost_net, rcu);
+
+	kfree(n->vqs[VHOST_NET_VQ_RX].rxq.queue);
+	kfree(n->vqs[VHOST_NET_VQ_TX].xdp);
+	kfree(n->dev.vqs);
+	if (n->page_frag.page)
+		__page_frag_cache_drain(n->page_frag.page, n->refcnt_bias);
+	kvfree(n);
+}
+
 static int vhost_net_release(struct inode *inode, struct file *f)
 {
 	struct vhost_net *n = f->private_data;
@@ -1404,15 +1417,8 @@ static int vhost_net_release(struct inode *inode, struct file *f)
 		sockfd_put(tx_sock);
 	if (rx_sock)
 		sockfd_put(rx_sock);
-	/* Make sure no callbacks are outstanding */
-	synchronize_rcu();
 
-	kfree(n->vqs[VHOST_NET_VQ_RX].rxq.queue);
-	kfree(n->vqs[VHOST_NET_VQ_TX].xdp);
-	kfree(n->dev.vqs);
-	if (n->page_frag.page)
-		__page_frag_cache_drain(n->page_frag.page, n->refcnt_bias);
-	kvfree(n);
+	call_rcu(&n->rcu, vhost_net_free);
 	return 0;
 }
 
-- 
2.32.0


  parent reply	other threads:[~2021-11-15 15:32 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-15 15:29 [PATCH 1/6] vhost: get rid of vhost_poll_flush() wrapper Andrey Ryabinin
2021-11-15 15:29 ` [PATCH 2/6] vhost_net: get rid of vhost_net_flush_vq() and extra flush calls Andrey Ryabinin
2021-11-16 14:33   ` Stefano Garzarella
2021-11-19 10:31     ` Andrey Ryabinin
2021-11-15 15:30 ` [PATCH 3/6] vhost_test: remove vhost_test_flush_vq() Andrey Ryabinin
2021-11-15 15:30 ` [PATCH 4/6] vhost_vsock: simplify vhost_vsock_flush() Andrey Ryabinin
2021-11-16 14:35   ` Stefano Garzarella
2021-11-15 15:30 ` [PATCH 5/6] vhost_net: remove NOP vhost_net_flush() in vhost_net_release() Andrey Ryabinin
2021-11-15 15:30 ` Andrey Ryabinin [this message]
2021-11-16  5:00   ` [PATCH 6/6] vhost_net: use RCU callbacks instead of synchronize_rcu() Jason Wang
2021-11-19 11:32     ` Andrey Ryabinin
2021-11-22  2:48       ` Jason Wang
2021-11-22  9:37       ` Michael S. Tsirkin
2021-11-22 15:12         ` Andrey Ryabinin
2021-11-16 14:33 ` [PATCH 1/6] vhost: get rid of vhost_poll_flush() wrapper Stefano Garzarella
2021-12-03 17:45   ` Mike Christie
2021-11-16 14:39 ` Denis Kirjanov
2021-11-16 14:41 ` Stefano Garzarella
2021-11-19 10:30   ` Andrey Ryabinin

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=20211115153003.9140-6-arbn@yandex-team.com \
    --to=arbn@yandex-team.com \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=davem@davemloft.net \
    --cc=hawk@kernel.org \
    --cc=jasowang@redhat.com \
    --cc=john.fastabend@gmail.com \
    --cc=kuba@kernel.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mst@redhat.com \
    --cc=netdev@vger.kernel.org \
    --cc=sgarzare@redhat.com \
    --cc=stefanha@redhat.com \
    --cc=virtualization@lists.linux-foundation.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

all inboxes | Powered by JetHome®