On 09/06/15 14:16, Peter Maydell wrote:
> On 9 June 2015 at 13:27, Marc Zyngier <marc.zyng...@arm.com> wrote:
>> On 09/06/15 12:24, Peter Maydell wrote:
>>> So either:
>>> * QEMU needs to tell the kernel the MPIDR for each vCPU
>>> * QEMU needs to ask the kernel the MPIDR for each vCPU
>>>
>>> Which is better? The latter is simpler and will work with
>>> existing kernels. The former would let us (for instance)
>>> use KVM when we're modelling (real world) boards which have
>>> particular cluster configurations (which might not match the
>>> kernel's current simplistic "always 16 CPUs at Aff0" setup).
>>
>> Given that it is userspace that provides the description of the
>> platform, it feels natural to let userspace set MPIDR accordingly.
>>
>> This will require some interesting rework in the kernel, but this
>> doesn't feel too bad.
>
> On the other hand QEMU needs to keep working with current kernels,
> so we need the "ask the kernel" code at least for the moment.
> Probably needs a capability for "kernel allows userspace to
> set MPIDR" so we can adapt.
Yeah, what I had in mind was something along the lines of:
- kernel computes its "default MPDIR"
- kernel exposes a new capability "KVM_ARM_ALLOW_MPIDR_OVERRIDE" (or
something along those lines)
- userspace does the right thing.
M.
--
Jazz is not dead. It just smells funny...
