Re: [RFC PATCH-for-6.1 0/9] hw/clock: Strengthen machine (non-qdev) clock propagation

[Philippe Mathieu-Daudé]

On 5/3/21 5:20 PM, Igor Mammedov wrote:
> On Tue, 13 Apr 2021 15:43:19 -0400
> Eduardo Habkost <ehabkost@redhat.com> wrote:
> 
>> On Mon, Apr 12, 2021 at 11:44:29AM +0100, Peter Maydell wrote:
>>> On Mon, 12 Apr 2021 at 11:31, Philippe Mathieu-Daudé <f4bug@amsat.org> 
>>> wrote:  
>>>> TIL MachineClass::reset().
>>>>
>>>> - hw/hppa/machine.c
>>>> - hw/i386/pc.c
>>>>
>>>>   Used to reset CPUs manually because CPUs aren't sysbus-reset.  
>>>
>>> pc_machine_reset() is not resetting the CPUs -- it is
>>> re-resetting the APIC devices, which looks like it is a
>>> workaround for a reset-ordering or other problem. I'm not
>>> sure where the CPUs are being reset...  
>>
>> CPU reset code was moved from pc.c:pc_cpu_reset() to
>> cpu.c:x86_cpu_machine_reset_cb() in commit 65dee3805259
>> ("target-i386: move cpu_reset and reset callback to cpu.c").
>> It's not clear to me why.
> 
> it was for cpu hotplug support, so that is we would have
> CPU in well know initial state after realize is complete.

It makes sense, but I don't see why this is considered x86 specific.