qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] selecting VIRTIO_INPUT and VIRTIO_VGA


From: Laszlo Ersek
Subject: Re: [Qemu-devel] selecting VIRTIO_INPUT and VIRTIO_VGA
Date: Mon, 20 Jul 2015 20:52:15 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.1.0

On 07/13/15 13:49, Paolo Bonzini wrote:
> 
> 
> On 13/07/2015 13:45, Gerd Hoffmann wrote:
>>>>>>>> and virtio-vga is only compiled on 64-bit Intel?
>>>>>>
>>>>>> There is virtio-gpu-pci ...
>>>>>>
>>>>>> Any specific reason why we need vga compatibility on !x86?
>>>>
>>>> I was actually thinking about 32-bit x86. :)  I agree that !x86 is not
>>>> necessary.
>> Yea, setting it for i386 makes sense indeed.  Just went out of my focus,
>> last time I used qemu-system-i386 was a few years back ...
>>
>> Any reason why one would use qemu-system-i386 instead of
>> qemu-system-x86_64 btw?  I suspect a 32bit host machine is the only one?
> 
> Yes, I think so.
> 
> Well, for TCG there is a difference of course.  Laszlo was using
> qemu-system-i386 because OVMF doesn't support our x86_64 layout for the
> SMM state save area.

Right, for 64-bit processors, there's one definition from AMD and
another from Intel. They are incompatible.

qemu-system-i386, with TCG, exposes the 32-bit SMM save state area that
the 32-bit SMM drivers -- open sourced by Intel, and being ported to
OVMF by yours truly -- are compatible with.

qemu-system-x86_64, with TCG, exposes the AMD flavor of the 64-bit area.
The code open sourced by Intel thus far is absolutely incompatible with
it. (In the code there are traces of support for the Intel flavor.)

With KVM support now existing (thanks Paolo), this might not be that
important any longer. (Paolo taught me that -lm can disable long mode
under KVM with qemu-system-x86_64, determining the save state aera
layout too.)

Thanks
Laszlo



reply via email to

[Prev in Thread] Current Thread [Next in Thread]