Re: [PATCH v3 5/6] hw/pci-host: Add emulation of Marvell MV64361 PPC system controller

[Philippe Mathieu-Daudé]

On 2/22/21 6:01 PM, BALATON Zoltan wrote:
> On Mon, 22 Feb 2021, Philippe Mathieu-Daudé wrote:
>> On 2/22/21 4:22 PM, BALATON Zoltan wrote:
>>> The Marvell Discovery II aka. MV64361 is a PowerPC system controller
>>> chip that is used on the pegasos2 PPC board. This adds emulation of it
>>> that models the device enough to boot guests on this board. The
>>> mv643xx.h header with register definitions is taken from Linux 4.15.10
>>> only fixing end of line white space errors and removing not needed
>>> parts, it's otherwise keeps Linux formatting.
>>>
>>> Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
>>> ---
>>>  hw/pci-host/Kconfig           |   3 +
>>>  hw/pci-host/meson.build       |   2 +
>>>  hw/pci-host/mv64361.c         | 966 ++++++++++++++++++++++++++++++++++
>>>  hw/pci-host/mv643xx.h         | 919 ++++++++++++++++++++++++++++++++
>>>  hw/pci-host/trace-events      |   6 +
>>>  include/hw/pci-host/mv64361.h |   8 +
>>>  include/hw/pci/pci_ids.h      |   1 +
>>>  7 files changed, 1905 insertions(+)
>>>  create mode 100644 hw/pci-host/mv64361.c
>>>  create mode 100644 hw/pci-host/mv643xx.h
>>>  create mode 100644 include/hw/pci-host/mv64361.h
>>
>> Is this the datasheet for this controller?
>> ftp://ftp.freecalypso.org/PowerPC/support_ics/mv64360/datasheets/DS_64360_1_2.pdf.zip
>>
>>
>> It seems to have common parts with the GT64120.
> 
> Yes, as noted on https://osdn.net/projects/qmiga/wiki/SubprojectPegasos2
> the GT64120 seems to be an older (and MIPS specific) version of this
> chip that Marvell bought and likely made this PPC version based on that.
> (The whole pegasos2 seems to be like a MIPS board with a PPC CPU even
> down to the firmware which looks quite like PMON just has SmartFirmware
> instead of the commands of PMON. So whoever designed it, might have
> taken inspiration from some MIPS hardware.) However the emulation of
> GT64xxx in QEMU seemed to be not complete and different enough so I
> haven't looked at that when implementing this so I don't know how much
> commonality is there and if that would worth reusing. (If you think
> there could be some reuse then that should be done separate of this
> series, afterwards it's merged.)

Yes, I see your patch is more complete. Fair enough.