Public bug reported:
My host machine is a Lenovo X1 Carbon (3rd gen) laptop running 64-bit Ubuntu
Linux 14.04. My guest machine is running 64-bit Ubuntu Linux 12.04. My QEMU
was compiled moments ago from the git repository, and it says its version is:
qemu-x86_64 version 2.2.93, Copyright (c) 2003-2008 Fabrice Bellard
My issue is that when I create a passthrough for a USB host device to
the guest, it causes qemu's CPU utilization on the host to increase
significantly and stay there. After the passthrough is created, qemu's
CPU usage in top hovers between 20% and 30% (evenly spread across CPUs).
In virt-manager's CPU usage chart, it shows the guest's CPU usage
growing from 0% to about 5%, although running "top" in the guest does
not show an increase in CPU usage. These usage levels drop back down to
normal (0 to 1%) only after I physically unplug the device or remove the
passthrough mapping.
It doesn't seem to matter what the device is. I have tested it using:
A Realtek rtl8192cu-based 802.11n adapter
An HP optical mouse
A Samsung Galaxy S5 phone
The behavior is the same regardless of device, and they otherwise seem
to work properly in the guest machine.
** Affects: qemu
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1445633
Title:
Creating a passthrough USB device causes excessive CPU usage in the
guest
Status in QEMU:
New
Bug description:
My host machine is a Lenovo X1 Carbon (3rd gen) laptop running 64-bit Ubuntu
Linux 14.04. My guest machine is running 64-bit Ubuntu Linux 12.04. My QEMU
was compiled moments ago from the git repository, and it says its version is:
qemu-x86_64 version 2.2.93, Copyright (c) 2003-2008 Fabrice Bellard
My issue is that when I create a passthrough for a USB host device to
the guest, it causes qemu's CPU utilization on the host to increase
significantly and stay there. After the passthrough is created,
qemu's CPU usage in top hovers between 20% and 30% (evenly spread
across CPUs). In virt-manager's CPU usage chart, it shows the guest's
CPU usage growing from 0% to about 5%, although running "top" in the
guest does not show an increase in CPU usage. These usage levels drop
back down to normal (0 to 1%) only after I physically unplug the
device or remove the passthrough mapping.
It doesn't seem to matter what the device is. I have tested it using:
A Realtek rtl8192cu-based 802.11n adapter
An HP optical mouse
A Samsung Galaxy S5 phone
The behavior is the same regardless of device, and they otherwise seem
to work properly in the guest machine.
To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1445633/+subscriptions
