Re: [PATCH v3 00/10] QOM: container_get() removal

[Philippe Mathieu-Daudé]

On 2/1/25 18:29, Peter Xu wrote:
> On Thu, Jan 02, 2025 at 04:12:34PM +0100, Philippe Mathieu-Daudé wrote:
> > Philippe Mathieu-Daudé (4):
> >    qdev: Expose qemu_create_machine()
> >    qdev: Implement qemu_create_machine() for user emulation
> >    qdev: Call qemu_create_machine() on user emulation
> >    qdev: Inline machine_containers[] in qemu_create_machine_containers()
>
> For these four, all look fine to me, feel free to take:
>
> Acked-by: Peter Xu <peterx@redhat.com>
>
> Said that, still one comment: maybe we don't need to make the function
> qemu_create_machine() to be an user+system API?  Instead we could have
> user_ss defines qemu_create_fake_machine() and invoke it iff USER_ONLY, the
> same as what patch 3 does.  Then we can drop patch 1, because exporting
> system's qemu_create_machine() isn't required, IIUC.

Good idea (not easy because -- again -- CONFIG_USER_ONLY isn't defined).

> We could add some comment above USER_ONLY's qemu_create_[fake_]machine() on
> why the fake machine is created (qdev_realize() of cpu devices relies on
> machine object, etc.).  It might still be helpful hint so that we know it
> can be dropped if qdev_realize() doesn't need to rely on machine one day.
>
> Thanks,