On 29 May 2018 at 10:08, Auger Eric <eric.au...@redhat.com> wrote:
> Hi Peter,
> On 05/22/2018 02:27 PM, Peter Maydell wrote:
>> On 13 May 2018 at 15:35, Eric Auger <eric.au...@redhat.com> wrote:
>>> To prepare for multiple redistributor regions, we introduce
>>> an array of uint32_t properties that stores the redistributor
>>> count of each redistributor region.
>>>
>>> Non accelerated VGICv3 only supports a single redistributor region.
>>> The capacity of all redist regions is checked against the number of
>>> vcpus.
>>>
>>> Machvirt is updated to set the count to 123 vcpus for the unique
>>> redistributor region we currently have.
>>> +
>>> + if (type == 3) {
>>> + qdev_prop_set_uint32(gicdev, "len-redist-region-count", 1);
>>> + qdev_prop_set_uint32(gicdev , "redist-region-count[0]",
>>> + vms->memmap[VIRT_GIC_REDIST].size / 0x20000);
>>
>> We used to create a region which had num_cpu redistributors in it;
>> won't this cause us to create one which has as many redistributors
>> as will fit in the space ?
> Is that an issue? From a machine perspective the whole region is
> reserved for rdist. dt and ACPI will expose this whole region and the
> device will use a subset of it? I agree I need to document this change
> in the commit message though.
It's a difference, and a guest-visible difference too. This
patchset is supposed to be introducing split-redistributor-regions,
not changing the behaviour of other configs. Also, I don't think
it makes sense to model a GIC with more redistributors than CPUs:
real hardware doesn't look like that AFAIK.
thanks
-- PMM
