On Saturday, November 23, 2019, Philippe Mathieu-Daudé <phi...@redhat.com>
wrote:

> On 11/22/19 9:37 PM, Aleksandar Markovic wrote:
>
>> On Fri, Nov 22, 2019 at 9:28 PM Philippe Mathieu-Daudé
>> <phi...@redhat.com> wrote:
>>
>>>
>>> On 11/13/19 2:47 PM, Aleksandar Markovic wrote:
>>>
>>>> From: Aleksandar Markovic <amarko...@wavecomp.com>
>>>>
>>>> Add two files that were recently introduced in a refactoring,
>>>> that Malta emulation relies on. They are added by this patch
>>>> to Malta section, but they are not added to the general MIPS
>>>> section, since they are really not MIPS-specific, and there
>>>> may be some non-MIPS hardware using them in future.
>>>>
>>>> Signed-off-by: Aleksandar Markovic <amarko...@wavecomp.com>
>>>> ---
>>>>    MAINTAINERS | 2 ++
>>>>    1 file changed, 2 insertions(+)
>>>>
>>>> diff --git a/MAINTAINERS b/MAINTAINERS
>>>> index ba9ca98..f8a1646 100644
>>>> --- a/MAINTAINERS
>>>> +++ b/MAINTAINERS
>>>> @@ -959,8 +959,10 @@ M: Philippe Mathieu-Daudé <phi...@redhat.com>
>>>>    R: Hervé Poussineau <hpous...@reactos.org>
>>>>    R: Aurelien Jarno <aurel...@aurel32.net>
>>>>    S: Maintained
>>>> +F: hw/isa/piix4.c
>>>>
>>>
>>> Maybe:
>>>
>>>      F: hw/*/piix4.c
>>>
>>> Or also add:
>>>
>>>      F: hw/acpi/piix4.c
>>>
>>>
>> Philippe, are you sure? hw/acpi/piix4.c is used in PC, not in Malta, no?
>>
>
> It is instantiated via:
>
>   mips_malta_init -> piix4_create -> piix4_pm_init
>
> The ACPI block is part of the PIIX4 chipset and should be mapped
> unconditionally.
>
>
Philippe, OK, sorry I failed to notice this detail. It just looked odd to
me that ACPI and MIPS could be connected, but stranger things happen. I am
going to add

F: hw/acpi/piix4.c

in the next version of this series. Should any change happen in the future,
I leave it to you to update the file list, of course.

Have a great sunny weekend!

Aleksandar



>
> The Malta indeed uses a PIIX4, which is a PIIX3 extended.
>
> The PC uses a Frankenstein PIIX3 with ACPI from PIIX4... I suppose because
> it started with a PIIX3, then we wanted power saving features but it was
> too late to switch to the PIIX4, so we abused of the fact than these blocks
> are exposed internally as PCI devices, so we can overload a PIIX3.
>
> See https://www.mail-archive.com/qemu-devel@nongnu.org/msg601512.html
>
>
> FYI I have a patch ready removing the 'if smbus' check in piix4_create().
>
>    F: hw/mips/mips_malta.c
>>>>    F: hw/mips/gt64xxx_pci.c
>>>> +F: include/hw/southbridge/piix.h
>>>>    F: tests/acceptance/linux_ssh_mips_malta.py
>>>>
>>>>    Mipssim
>>>>
>>>>
>>> Reviewed-by: Philippe Mathieu-Daudé <phi...@redhat.com>
>>>
>>>
>>>
>

Reply via email to