interrupts enabled during hypervisor setup (arch_entry)?

10 views
Skip to first unread message

Claudio Fontana

unread,
Jul 21, 2015, 11:03:29 AM7/21/15
to Jan Kiszka, jailho...@googlegroups.com
Hi,

sorry for maybe an obvious question, but I am a bit puzzled about the state of irqs (enabled/disabled) in the driver code during hypervisor setup:

I don't see them being disabled during the setup of the hype mode. Is it expected? Is there some precaution we have to take during the setup of the hype mode regarding incoming irqs from the kernel mode?

Thanks,

Claudio



Henning Schild

unread,
Jul 21, 2015, 11:21:45 AM7/21/15
to Claudio Fontana, Jan Kiszka, jailho...@googlegroups.com
The hypervisor disables interrupts right when it gets started by Linux.

hypervisor/arch/x86/entry.S:

>> arch_entry:
>> cli
...

Was that your question?

Henning

Claudio Fontana

unread,
Jul 21, 2015, 11:35:29 AM7/21/15
to Henning Schild, Jan Kiszka, jailho...@googlegroups.com
Doh.. sorry, I was looking only at the arm code. To me it looks like it is missing from the arm code..?

Thanks,

Claudio

--
Claudio Fontana
Server Virtualization Architect
Huawei Technologies Duesseldorf GmbH
Riesstraße 25 - 80992 München

Jan Kiszka

unread,
Jul 21, 2015, 12:12:45 PM7/21/15
to Claudio Fontana, Henning Schild, jailho...@googlegroups.com
On 2015-07-21 17:34, Claudio Fontana wrote:
> On 21.07.2015 17:22, Henning Schild wrote:
>> On Tue, 21 Jul 2015 17:03:23 +0200
>> Claudio Fontana <claudio...@huawei.com> wrote:
>>
>>> Hi,
>>>
>>> sorry for maybe an obvious question, but I am a bit puzzled about the
>>> state of irqs (enabled/disabled) in the driver code during hypervisor
>>> setup:
>>>
>>> I don't see them being disabled during the setup of the hype mode. Is
>>> it expected? Is there some precaution we have to take during the
>>> setup of the hype mode regarding incoming irqs from the kernel mode?
>>
>> The hypervisor disables interrupts right when it gets started by Linux.
>>
>> hypervisor/arch/x86/entry.S:
>>
>>>> arch_entry:
>>>> cli
>> ...
>>
>> Was that your question?
>>
>> Henning
>>
>
> Doh.. sorry, I was looking only at the arm code. To me it looks like it is missing from the arm code..?

Practically, you already get interrupt disabling via on_each_cpu. But
doing it explicitly also in the hypervisor entry is better, a safety
measure.

Jan

--
Siemens AG, Corporate Technology, CT RTC ITP SES-DE
Corporate Competence Center Embedded Linux
Reply all
Reply to author
Forward
0 new messages