Il 26/11/2012 12:52, Andreas Färber ha scritto:
> Am 26.11.2012 08:20, schrieb Paolo Bonzini:
>> Il 26/11/2012 01:12, Andreas Färber ha scritto:
>>> qdev_free() was trivially wrapping object_delete(). Inline it and drop
>>> the wrapper as a further step to QOM'ify qdev.
>>
>> The reference counts o
Am 26.11.2012 08:20, schrieb Paolo Bonzini:
> Il 26/11/2012 01:12, Andreas Färber ha scritto:
>> qdev_free() was trivially wrapping object_delete(). Inline it and drop
>> the wrapper as a further step to QOM'ify qdev.
>
> The reference counts of devices and buses are completely disconnected
> from
Il 26/11/2012 01:12, Andreas Färber ha scritto:
> qdev_free() was trivially wrapping object_delete(). Inline it and drop
> the wrapper as a further step to QOM'ify qdev.
The reference counts of devices and buses are completely disconnected
from reality, and that's why object_delete() works.
I'd p
qdev_free() was trivially wrapping object_delete(). Inline it and drop
the wrapper as a further step to QOM'ify qdev.
Signed-off-by: Andreas Färber
Cc: Paolo Bonzini
---
hw/acpi_piix4.c |2 +-
hw/pci-hotplug.c |2 +-
hw/pci_bridge.c |2 +-
hw/pcie.c|2 +