qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Longpeng(Mike)" via <qemu-devel@nongnu.org>
To: <stefanha@redhat.com>, <mst@redhat.com>, <jasowang@redhat.com>,
	<sgarzare@redhat.com>
Cc: <cohuck@redhat.com>, <pbonzini@redhat.com>,
	<arei.gonglei@huawei.com>, <yechuan@huawei.com>,
	<huangzhichao@huawei.com>, <qemu-devel@nongnu.org>,
	Longpeng <longpeng2@huawei.com>
Subject: [PATCH] vhost: configure all host notifiers in a single MR transaction
Date: Fri, 18 Nov 2022 22:49:15 +0800	[thread overview]
Message-ID: <20221118144915.2009-1-longpeng2@huawei.com> (raw)

From: Longpeng <longpeng2@huawei.com>

This allows the vhost device to batch the setup of all its host notifiers.
This significantly reduces the device starting time, e.g. the vhost-vDPA
generic device [1] start time reduce from 376ms to 9.1ms for a VM with
64 vCPUs and 3 vDPA device(64vq per device).

[1] https://www.mail-archive.com/qemu-devel@nongnu.org/msg921541.html

Signed-off-by: Longpeng <longpeng2@huawei.com>
---
 hw/virtio/vhost.c | 39 ++++++++++++++++++++++++++++++++++++++-
 1 file changed, 38 insertions(+), 1 deletion(-)

diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index d1c4c20b8c..bf82d9b176 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -1507,6 +1507,7 @@ void vhost_dev_cleanup(struct vhost_dev *hdev)
 int vhost_dev_enable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
 {
     BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(vdev)));
+    int vq_init_count = 0;
     int i, r, e;
 
     /* We will pass the notifiers to the kernel, make sure that QEMU
@@ -1518,6 +1519,12 @@ int vhost_dev_enable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
         goto fail;
     }
 
+    /*
+     * Batch all the host notifiers in a single transaction to avoid
+     * quadratic time complexity in address_space_update_ioeventfds().
+     */
+    memory_region_transaction_begin();
+
     for (i = 0; i < hdev->nvqs; ++i) {
         r = virtio_bus_set_host_notifier(VIRTIO_BUS(qbus), hdev->vq_index + i,
                                          true);
@@ -1525,19 +1532,33 @@ int vhost_dev_enable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
             error_report("vhost VQ %d notifier binding failed: %d", i, -r);
             goto fail_vq;
         }
+
+        vq_init_count++;
     }
 
+    memory_region_transaction_commit();
+
     return 0;
 fail_vq:
-    while (--i >= 0) {
+    for (i = 0; i < vq_init_count; i++) {
         e = virtio_bus_set_host_notifier(VIRTIO_BUS(qbus), hdev->vq_index + i,
                                          false);
         if (e < 0) {
             error_report("vhost VQ %d notifier cleanup error: %d", i, -r);
         }
         assert (e >= 0);
+    }
+
+    /*
+     * The transaction expects the ioeventfds to be open when it
+     * commits. Do it now, before the cleanup loop.
+     */
+    memory_region_transaction_commit();
+
+    for (i = 0; i < vq_init_count; i++) {
         virtio_bus_cleanup_host_notifier(VIRTIO_BUS(qbus), hdev->vq_index + i);
     }
+
     virtio_device_release_ioeventfd(vdev);
 fail:
     return r;
@@ -1553,6 +1574,12 @@ void vhost_dev_disable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
     BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(vdev)));
     int i, r;
 
+    /*
+     * Batch all the host notifiers in a single transaction to avoid
+     * quadratic time complexity in address_space_update_ioeventfds().
+     */
+    memory_region_transaction_begin();
+
     for (i = 0; i < hdev->nvqs; ++i) {
         r = virtio_bus_set_host_notifier(VIRTIO_BUS(qbus), hdev->vq_index + i,
                                          false);
@@ -1560,8 +1587,18 @@ void vhost_dev_disable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
             error_report("vhost VQ %d notifier cleanup failed: %d", i, -r);
         }
         assert (r >= 0);
+    }
+
+    /*
+     * The transaction expects the ioeventfds to be open when it
+     * commits. Do it now, before the cleanup loop.
+     */
+    memory_region_transaction_commit();
+
+    for (i = 0; i < hdev->nvqs; ++i) {
         virtio_bus_cleanup_host_notifier(VIRTIO_BUS(qbus), hdev->vq_index + i);
     }
+
     virtio_device_release_ioeventfd(vdev);
 }
 
-- 
2.23.0



             reply	other threads:[~2022-11-18 14:50 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-18 14:49 Longpeng(Mike) via [this message]
2022-11-18 18:28 ` [PATCH] vhost: configure all host notifiers in a single MR transaction Stefan Hajnoczi
2022-11-21  4:01 ` Jason Wang
2022-11-28  9:08   ` longpeng2--- via
2022-11-29  2:55     ` Jason Wang

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=20221118144915.2009-1-longpeng2@huawei.com \
    --to=qemu-devel@nongnu.org \
    --cc=arei.gonglei@huawei.com \
    --cc=cohuck@redhat.com \
    --cc=huangzhichao@huawei.com \
    --cc=jasowang@redhat.com \
    --cc=longpeng2@huawei.com \
    --cc=mst@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=sgarzare@redhat.com \
    --cc=stefanha@redhat.com \
    --cc=yechuan@huawei.com \
    /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).