Re: [Qemu-devel] [PATCH v2] qdev: Mark devices as non-hotpluggable by default

[Thomas Huth]

On 26.09.2017 04:59, Eduardo Habkost wrote:
> On Mon, Sep 25, 2017 at 07:05:26PM +0100, Peter Maydell wrote:
>> On 25 September 2017 at 18:59, Eduardo Habkost <address@hidden> wrote:
>>> Finding the full list of devices that can be instantiated
>>> internally at hotplug-time sounds tricky.
>>
>> If we just diff "list of devices marked hotplug before this patch"
>> against "list of devices marked hotplug after this patch" how
>> big is the list? Can we just eyeball it to see what needs
>> to be specialcased?
> 
> So, the full list quite big, ~1800 device types are affected by
> this patch:
> 
> https://gist.github.com/ehabkost/bd8e25c6811ac81d947ad8ad5b557f5c#file-dev-types-diff-json
> 
> If we ignore the "-cpu" classes, there ~640 affected device
> types.
> 
> However, if we look only at the direct children of TYPE_DEVICE,
> we have:

OK, thanks a lot for that list! But do you think that we can assume that
the devices which are not direct children of TYPE_DEVICE can not be
hotplugged internally during runtime? ... I currently don't think so
(but at least they are good candidates which need to be examined more
carefully).

But anyway, how can we continue here? Set hotpluggable = true on all 640
(or even all 1800) affected devices? That sounds very, very ugly to me.
Maybe we should just do it for the virtio-*-device devices and the
others that break during "make check" (btw. sorry for not noticing this
... I normally run "make check" regularly, but this time I apparently
missed to run it for aarch64), and if we later notice some more
problems, we know that we lack a "make check" test for that case, too,
and we should add such a test? That would at least eventually improve
our test coverage a little bit, I guess...

 Thomas