Re: [Qemu-devel] [PATCH 00/39] Delay destruction of memory regions to instance_finalize

[Paolo Bonzini]

Il 07/06/2013 03:45, Andreas Färber ha scritto:
> Am 07.06.2013 03:03, schrieb Paolo Bonzini:
>> Il 05/06/2013 05:50, Andreas Färber ha scritto:
>>> Also I notice some patches are accessing parent fields directly - please
>>> use BUS(), PCI_DEVICE() etc. to hide this.
>>
>> I'm always using them.  For example:
>>
>> +static void intel_hda_instance_finalize(Object *obj)
>> +{
>> +    PCIDevice *pci = PCI_DEVICE(obj);
>> +    IntelHDAState *d = DO_UPCAST(IntelHDAState, pci, pci);
> 
> I'll comment inline then. :)

Ah right, in those cases I was mostly doing what the
adjacent/pre-existing code did.  But I can definitely fix them up!

Paolo