Re: [Qemu-devel] [RFC PATCH 10/11] hw/arm/virt-acpi-build: Generation of DSDT table for virt devices

[Shannon Zhao <zhaoshenglong@huawei.com>]

On 2015/1/26 18:40, Igor Mammedov wrote:
> On Sat, 24 Jan 2015 17:21:19 +0800
> Shannon Zhao <zhaoshenglong@huawei.com> wrote:
> 
>> > DSDT consists of the usual common table header plus a definition
>> > block in AML encoding which describes all devices in the platform.
>> > 
>> > After initializing DSDT with header information the namespace is
>> > created which is followed by the device encodings. The devices are
>> > described using the Resource Template for the 32-Bit Fixed Memory
>> > Range and the Extended Interrupt Descriptors.
>> > 
>> > The following devices are included in the DSDT:
>> > - CPUs
>> > - UART
>> > - RTC
>> > - NAND Flash
>> > - virtio-mmio
>> > 
>> > Signed-off-by: Shannon Zhao <zhaoshenglong@huawei.com>
>> > ---
>> >  hw/arm/virt-acpi-build.c |  120 ++++++++++++++++++++++++++++++++++++++++++++++
>> >  1 files changed, 120 insertions(+), 0 deletions(-)
>> > 
>> > diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c
>> > index de1f307..5c76ca2 100644
>> > --- a/hw/arm/virt-acpi-build.c
>> > +++ b/hw/arm/virt-acpi-build.c
>> > @@ -98,6 +98,111 @@ static inline void acpi_add_table(GArray *table_offsets, GArray *table_data)
>> >      g_array_append_val(table_offsets, offset);
>> >  }
>> >  
>> > +static void acpi_dsdt_add_cpus(AcpiAml *scope, int smp_cpus)
>> > +{
>> > +    AcpiAml dev, crs;
>> > +    int i;
>> > +    char name[5];
>> > +    for (i = 0; i < smp_cpus; i++) {
> I'm not sure about ARM butm shouldn't not present but possble CPUs
> also described here?
> 

In struct VirtGuestInfo there are nb_cpus and max_cpus. nb_cpus stands present CPUs while
max_cpus stands possible CPUs. We can use them to create bitmap.

> PS:
> One more thing about CPU hotplug, I'd like current (x86) bitmap based
> QEMU<->APCI interface (which scales only upto 256 CPU) have redone
> to a one similar to memory hotplug first.
> So that ARM wouldn't have to support compatibility mode for it in the future.
> 

Good idea :-)

Thanks,
Shannon