Re: [Qemu-devel] [PATCH v5] powerpc: add PVR mask support

[Paul Mackerras]

On Thu, Aug 29, 2013 at 01:33:07PM +1000, Paul Mackerras wrote:
> On Wed, Aug 28, 2013 at 12:49:27PM +0200, Andreas Färber wrote:
> 
> > According to the only reply I received, it's "POWER7", not "POWER7+" -
> > see my patch description. If that information was wrong, we'll need to
> > move POWER7P introduction before my fw_name patch and update it accordingly.
> 
> It should be "POWER7+" for consistency with other hypervisors and
> firmware.

Actually, on a POWER7+ machine under PowerVM here, I am seeing
PowerPC,address@hidden --- but PowerVM is running the partition in v2.06
architected mode rather than POWER7+ native mode.  I'm not sure at the
moment how to get it to run in POWER7+ native mode (or even if there
is such a thing).  In any case, I don't think it really matters
whether it's PowerPC,POWER7 or PowerPC,POWER7+.

Which brings me to another question.  PAPR specifies that when a
partition is being run in a compatibility mode for an older processor,
the cpu-version property in the cpu nodes has a "logical PVR" value
representing the architecture level of the compatibility mode; for
example, 0x0f000001 for architecture 2.04, 0x0f000002 for 2.05
(POWER6), 0x0f000003 for 2.06 (POWER7), 0x0f000004 for 2.07 (POWER8).

We are going to need a way to do that for pseries guests.  So for
example if we are running on a POWER8 machine and using KVM, we would
want to be able to tell qemu that we want a POWER7 machine and have
qemu put 0x0f000003 in the device tree, and also tell KVM to enable
POWER7 compatibility mode when running the guest, presumably by
setting the PVR for the guest to 0x0f000003.  That PVR setting doesn't
(and can't) modify the PVR seen by the guest, since the hardware
doesn't provide any way to virtualize the PVR, but we will make it set
the compatibility mode for the guest.

How do you think we should handle this in qemu?

Regards,
Paul.