blob: 2d1d6b5b07f6f355fbf99f5429b85fa60d931a01 (
plain) (
blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
|
support vhost user without specifying vhostforce
[Impact]
* vhost-user falls back to virtio-net which causes performance lose without specifying the vhostforce option. But it should be the default behavior for vhost-user, since guests using PMD doesn't support msi-x.
[Test Case]
create a vhost-user virtio backend without specifying the vhostforce option, i.e. -netdev type=vhost-user,id=mynet1,chardev=<char_dev_for_the_controll_channel>
start the VM
vhost-user is not enabled
[Regression Potential]
* none
vhost user nic doesn't support non msi guests(like pxe stage) by default.
Vhost user nic can't fall back to qemu like normal vhost net nic does. So we should
enable it for non msi guests.
The problem has been fix in qemu upstream - http://git.qemu.org/?p=qemu.git;a=commitdiff;h=24f938a682d934b133863eb421aac33592f7a09e. And the patch needs to be backported to 1:2.2+dfsg-5expubuntu9.8 .
|