Re: [Qemu-devel] [PATCH v5 2/6] qdev: introduce FWPathProvider interface

[Alexey Kardashevskiy <aik@ozlabs.ru>]

On 03/13/2014 10:38 AM, Andreas Färber wrote:
> Am 13.03.2014 00:02, schrieb Alexey Kardashevskiy:
>> On 03/13/2014 05:07 AM, Andreas Färber wrote:
>>> Am 20.02.2014 14:50, schrieb Alexey Kardashevskiy:
>>>> QEMU supports firmware names for all devices in the QEMU tree but
>>>> some architectures expect some parts of firmware path names in different
>>>> format.
>>>>
>>>> This introduces a firmware-pathname-change interface definition.
>>>> If some machines needs to redefine the firmware path format, it has
>>>> to add the TYPE_FW_PATH_PROVIDER interface to an object that is above
>>>> the device on the QOM tree (typically /machine).
>>>>
>>>> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
>>>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>>>
>>> I do not see a reply to my question on v4 and didn't catch bonzini on
>>> IRC - guessing the two lines should be swapped since both v4 and v5 were
>>> submitted by Alexey?
>>
>>
>> I am actually waiting till machine QOM'fication to finish, then I'll repost
>> this patchset. Something changed?
> 
> Yes, machine QOM'ification is queued on qom-next and the pull request
> containing it is facing some issues still. But you can already rebase
> onto git://github.com/afaerber/qemu-cpu.git qom-next branch.


gosh :( so many trees. What tree/branch should I base my
"client-architecture-support" which needs new CPU suboptions support?


> In short, the pseries machine needs to use type_init(), call
> type_register_static(), and the TypeInfo needs to specify as .class_data
> the QEMUMachine and a .class_init function that assigns that QEMUMachine
> pointer to mc->qemu_machine as done in machine.c.




-- 
Alexey