Re: [Qemu-devel] [PATCH RESEND 0/2 V2] Virtual Machine Generation ID

[Gal Hammer <ghammer@redhat.com>]

On 14/09/2014 11:48, Paolo Bonzini wrote:
> Il 14/09/2014 08:25, Gal Hammer ha scritto:
>> Hi,
>>
>> A two parts patch to add a QEmu support for Microsoft's Virtual Machine
>> Generation ID device.
>>
>> The first one add a new ACPI directive which allow to use a 16-bytes
>> buffer in an ACPI table. This buffer is for storing the VM's UUID.
>>
>> The second is the ACPI tables changes and the actual device.
>>
>> Your comment are welcomed.
>>
>> Thanks,
>>
>>      Gal.
>>
>> V2: - Remove "-uuid" command line parameter.
>>      - Move device's description from SSDT to DDST table.
>>      - Add new "vmgenid" sysbus device.
>>
>> Gal Hammer (2):
>>    i386: Add an ACPI_EXTRACT_NAME_BUFFER16 directive.
>>    i386: Add a Virtual Machine Generation ID device.
>>
>>   default-configs/i386-softmmu.mak   |  1 +
>>   default-configs/x86_64-softmmu.mak |  1 +
>>   hw/i386/acpi-build.c               | 23 ++++++++++-
>>   hw/i386/acpi-dsdt.dsl              | 37 +++++++++++++++++
>>   hw/misc/Makefile.objs              |  1 +
>>   hw/misc/vmgenid.c                  | 85 ++++++++++++++++++++++++++++++++++++++
>>   include/hw/i386/pc.h               |  3 ++
>>   scripts/acpi_extract.py            | 23 +++++++----
>>   8 files changed, 163 insertions(+), 11 deletions(-)
>>   create mode 100644 hw/misc/vmgenid.c
>>
>
> The idea is okay, but the device has to be added to the Q35 DSDT as
> well.  If you just #included a file in the same style as
> acpi-dsdt-hpet.dsl, you would have a namespace clash between the PIIX
> and Q35 DSDTs.  acpi-dsdt-hpet.dsl can do it because it has no
> ACPI_EXTRACT directives.
>
> I think because of this, it is a bit simpler to add the device
> optionally to the SSDT, similar to how the TPM is handled.

Back to SSDT? Didn't you say something about migration issues that 
prevent me from using SSDT because the table's data is modified?

>
> Paolo
>

     Gal.