qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Gal Hammer <ghammer@redhat.com>
To: qemu-devel@nongnu.org
Cc: mst@redhat.com, Gal Hammer <ghammer@redhat.com>
Subject: [Qemu-devel] [PATCH v3 2/2] virtio: improve virtio devices initialization time
Date: Mon, 29 Jan 2018 16:20:57 +0200	[thread overview]
Message-ID: <1517235657-22547-3-git-send-email-ghammer@redhat.com> (raw)
In-Reply-To: <1517235657-22547-1-git-send-email-ghammer@redhat.com>

The loading time of a VM is quite significant when its virtio
devices use a large amount of virt-queues (e.g. a virtio-serial
device with max_ports=511). Most of the time is spend in the
creation of all the required event notifiers (ioeventfd and memory
regions).

This patch pack all the changes to the memory regions in a
single memory transaction.

Reported-by: Sitong Liu <address@hidden>
Reported-by: Xiaoling Gao <address@hidden>
Signed-off-by: Gal Hammer <ghammer@redhat.com>
---
 hw/virtio/virtio.c | 22 ++++++++++++++++++++--
 1 file changed, 20 insertions(+), 2 deletions(-)

diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index 3667cd6..006d3d1 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -2572,8 +2572,9 @@ static Property virtio_properties[] = {
 static int virtio_device_start_ioeventfd_impl(VirtIODevice *vdev)
 {
     VirtioBusState *qbus = VIRTIO_BUS(qdev_get_parent_bus(DEVICE(vdev)));
-    int n, r, err;
+    int i, n, r, err;
 
+    memory_region_transaction_begin();
     for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
         VirtQueue *vq = &vdev->vq[n];
         if (!virtio_queue_get_num(vdev, n)) {
@@ -2596,9 +2597,11 @@ static int virtio_device_start_ioeventfd_impl(VirtIODevice *vdev)
         }
         event_notifier_set(&vq->host_notifier);
     }
+    memory_region_transaction_commit();
     return 0;
 
 assign_error:
+    i = n; /* save n for a second iteration after transaction is committed. */
     while (--n >= 0) {
         VirtQueue *vq = &vdev->vq[n];
         if (!virtio_queue_get_num(vdev, n)) {
@@ -2608,7 +2611,14 @@ assign_error:
         event_notifier_set_handler(&vq->host_notifier, NULL);
         r = virtio_bus_set_host_notifier(qbus, n, false);
         assert(r >= 0);
-        virtio_bus_cleanup_host_notifier(qbus, n);
+    }
+    memory_region_transaction_commit();
+
+    while (--i >= 0) {
+        if (!virtio_queue_get_num(vdev, i)) {
+            continue;
+        }
+        virtio_bus_cleanup_host_notifier(qbus, i);
     }
     return err;
 }
@@ -2626,6 +2636,7 @@ static void virtio_device_stop_ioeventfd_impl(VirtIODevice *vdev)
     VirtioBusState *qbus = VIRTIO_BUS(qdev_get_parent_bus(DEVICE(vdev)));
     int n, r;
 
+    memory_region_transaction_begin();
     for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
         VirtQueue *vq = &vdev->vq[n];
 
@@ -2635,6 +2646,13 @@ static void virtio_device_stop_ioeventfd_impl(VirtIODevice *vdev)
         event_notifier_set_handler(&vq->host_notifier, NULL);
         r = virtio_bus_set_host_notifier(qbus, n, false);
         assert(r >= 0);
+    }
+    memory_region_transaction_commit();
+
+    for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
+        if (!virtio_queue_get_num(vdev, n)) {
+            continue;
+        }
         virtio_bus_cleanup_host_notifier(qbus, n);
     }
 }
-- 
2.7.5

  parent reply	other threads:[~2018-01-29 14:21 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-01-29 14:20 [Qemu-devel] [PATCH v3 0/2] virtio: improve virtio devices initialization time Gal Hammer
2018-01-29 14:20 ` [Qemu-devel] [PATCH v3 1/2] virtio: remove event notifier cleanup call on de-assign Gal Hammer
2018-01-29 14:20 ` Gal Hammer [this message]
2018-02-08 19:05   ` [Qemu-devel] [PATCH v3 2/2] virtio: improve virtio devices initialization time Michael S. Tsirkin
2018-01-29 17:47 ` [Qemu-devel] [PATCH v3 0/2] " Greg Kurz
2018-01-30  8:33   ` Gal Hammer

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=1517235657-22547-3-git-send-email-ghammer@redhat.com \
    --to=ghammer@redhat.com \
    --cc=mst@redhat.com \
    --cc=qemu-devel@nongnu.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).