Re: [PATCH] x86/xsave: initialization improvements

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 29/08/13 11:49, Jan Beulich wrote:
>>>> On 29.08.13 at 12:17, Andrew Cooper <andrew.cooper3@citrix.com> wrote:
>> On 29/08/13 10:39, Jan Beulich wrote:
>>> - properly validate available feature set on APs
>>> - also validate xsaveopt availability on APs
>>> - properly indicate whether the initialization is on the BSP (we
>>>   shouldn't be using "cpu == 0" checks for this)
>>>
>>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>> Why is "cpu == 0" insufficient?
>>
>> The pcpu with id 0 is necessarily the BSP because of set_processor_id(0)
>> early in __start_xen().
> But eventually we will want/need to be able to hot-remove
> CPU 0, and then some other CPU may later come back up
> getting set its ID to zero. That shouldn't trigger one-time init
> code paths.
>
> I know this is not the only place, but as I come across them and
> touch respective code anyway, I'm trying to fix those instances.
>
> Jan
>

Are we expecting that to actually work on real hardware?

One example is the BSP NMI routing which will only go to the CPU which
the BIOS booted as the BSP.

~Andrew