On Aug 12, 2016 05:24, "Julien Grall" <julien.grall@arm.com> wrote:
>
> Hello Tamas,
>
>
> On 10/08/2016 17:00, Tamas K Lengyel wrote:
>>
>> diff --git a/tools/libxl/libxl_types.idl b/tools/libxl/libxl_types.idl
>> index ef614be..97948fd 100644
>> --- a/tools/libxl/libxl_types.idl
>> +++ b/tools/libxl/libxl_types.idl
>> @@ -439,6 +439,13 @@ libxl_rdm_reserve = Struct("rdm_reserve", [
>>      ("policy",      libxl_rdm_reserve_policy),
>>      ])
>>
>> +# Consistent with the values defined for HVM_PARAM_ALTP2M
>> +libxl_altp2m_mode = Enumeration("altp2m_mode", [
>> +    (0, "disabled"),
>> +    (1, "mixed"),
>> +    (2, "external_only"),
>> +    ], init_val = "LIBXL_ALTP2M_MODE_DISABLED")
>> +
>>  libxl_domain_build_info = Struct("domain_build_info",[
>>      ("max_vcpus",       integer),
>>      ("avail_vcpus",     libxl_bitmap),
>> @@ -512,7 +519,7 @@ libxl_domain_build_info = Struct("domain_build_info",[
>>                                         ("mmio_hole_memkb",  MemKB),
>>                                         ("timer_mode",       libxl_timer_mode),
>>                                         ("nested_hvm",       libxl_defbool),
>> -                                       ("altp2m",           libxl_defbool),
>> +                                       ("altp2m",           libxl_altp2m_mode),
>
>
> Should we move altp2m directly outside of the structure "hvm" to avoid yet another change when altp2m will be supported on ARM? (see [1])
>
> Regards,
>
> [1] https://lists.xen.org/archives/html/xen-devel/2016-08/msg00147.html
>
> --
> Julien Grall

Hi Julien,
We can do that here if preferred by the tools maintainers though I'm on the opinion that it should be adjusted in the ARM altp2m series once it actually becomes available there.

Tamas