Re: [PATCH v2 0/5] qapi: Restrict machine (and migration) specific commands

[Markus Armbruster]

Philippe Mathieu-Daudé <philmd@redhat.com> writes:

> On 10/19/20 6:48 PM, Markus Armbruster wrote:
>> Eduardo Habkost <ehabkost@redhat.com> writes:
>> 
>>> On Mon, Oct 19, 2020 at 09:55:20AM +0200, Markus Armbruster wrote:
>>>> Eduardo Habkost <ehabkost@redhat.com> writes:
>>>>
>>>>> On Mon, Oct 12, 2020 at 02:15:31PM +0200, Philippe Mathieu-Daudé wrote:
>>>>>> Reduce the machine code pulled into qemu-storage-daemon.
>>>>>>
>>>>>> The series is fully Acked, but Markus wants it reviewed
>>>>>> by the Machine core maintainers.
>>>>>
>>>>> I've confirmed that all patches move QAPI schema code without
>>>>> introducing any additional changes.
>>>>>
>>>>> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
>>>>
>>>> I take this as "I agree the things moved to machine.json belong there".
>>>> Holler if I'm mistaken.
>>>
>>> I agree machine.json is better than misc.json for them, yes.
>>>
>>> I miss short descriptions of the purpose of each file, though.
>>> It would help us decide what's appropriate in the future.
>>
>> The QAPI modules are commonly aligned with sub-systems defined in
>> MAINTAINERS.
>>
>> Regardless, file comments would be nice.
>
> I don't understand what you mean/expect by "file comments".
> Example?

A comment explaining the file, at the beginning of the file.

> W.r.t. MAINTAINERS, I can move Xen code to qapi/migration-xen.json;

How much could be moved, and from where?

Sub-modules don't need to mirror MAINTAINERS slavishly.  We want
reasonably-sized modules, and we want useful get_maintainer.pl output.

> 'query-kvm' is used when no KVM built it, so I'll let it in
> machine.json; the others seem to belong in machine.json too,
> with no particular justification.