Il 29/09/2014 20:31, Andrey Korolyov ha scritto:
>> >
>> > Can you get a trace of the guest (echo kvm >
>> > /sys/kernel/debug/tracing/set_event; sleep 1s; cp
>> > /sys/kernel/debug/tracing/trace /tmp/trace.txt) while it is in the 100%
>> > cpu consumption state?
>> >
>> > Is it otherwise operation
On Mon, Sep 29, 2014 at 10:25 PM, Marcelo Tosatti wrote:
> Andrey,
>
> Can you get a trace of the guest (echo kvm >
> /sys/kernel/debug/tracing/set_event; sleep 1s; cp
> /sys/kernel/debug/tracing/trace /tmp/trace.txt) while it is in the 100%
> cpu consumption state?
>
> Is it otherwise operational
Andrey,
Can you get a trace of the guest (echo kvm >
/sys/kernel/debug/tracing/set_event; sleep 1s; cp
/sys/kernel/debug/tracing/trace /tmp/trace.txt) while it is in the 100%
cpu consumption state?
Is it otherwise operational?
On Thu, Sep 25, 2014 at 10:40:27PM +0400, Andrey Korolyov wrote:
> He
Hello,
there is a somewhat weird issue with 3.10, backported IOAPIC behavior
(which allowed windows guest to not hang in my case), 2.1.1 and webkit
suite which is well-known for all timer weirdness for Windows guests.
Here is a way to reproduce:
- start the guest,
- open webkit-based browser, cpu
