[libvirt] [PATCH 1/2] Add VSCSI bus type and VSCSI controller type for pseries guest.

[Li Zhang]

On Fri 04 May 2012 05:55:46 PM CST, Daniel P. Berrange wrote:
> On Fri, May 04, 2012 at 05:49:54PM +0800, Li Zhang wrote:
>> On 05/04/2012 05:04 PM, Daniel P. Berrange wrote:
>>> On Fri, May 04, 2012 at 10:05:13AM +0800, Li Zhang wrote:
>>>> Now, there is only SCSI bus and controller type in libvirt.
>>>> And when configuring VSCSI controller, it needs to configure
>>>> the spapr-vio bus address type externally. It's a little
>>>> inconvenient to configure the controller type.
>>>>
>>>> This patch is to add VSCSI bus type and VSCSI controller type.
>>>> And handle with the problems when VSCSI and SCSI devices
>>>> working together, by assign the even number to index of
>>>> VSCSI controller and odd number to index of SCSI controller.
>>>>
>>>> And when the VSCSI controller is always assigned as SPAPRVIO
>>>> bus address type and SCSI controller will be always assigned
>>>> as PCI bus address, which is implemented according to the
>>>> controllers' type.
>>>>
>>>> So when one disk is based on VSCSI controller, then assign
>>>> the bus as 'vscsi', and one right VSCSI controller will be
>>>> selected.
>>>
>>> I'm not at all convinced by this description that we need
>>> a new controller type. At most we need a separate<address/>
>>> type for the VSCSI controller under an existing SCSI controller
>>> element.
>>>
>>
>> The problem is that when using two kinds of SCSI controller,
>> vscsi and scsi, which are based on different address type and
>> model at the same time, it needs to specify the model and address
>> type externally. And it's hard to find the right controller for
>> disk to use, only by the 'sdx'.
>
> Sure, but there's nothing special about s390 / vSCSI in that
> respect.  If you have a guest with a SATA controller and a
> SCSI controller you have the same problem. If you have multiple
> disk controllers assigned to the guest, and are adding a disk,
> then you are responsible for providing an<address>  element
> for the disk to indicate which controller it should be added
> to.
>
  OK. I see. I will try that way.

> The automagic guessing of controllers is only ever intended
> to work in the simple case where you have 1 kind of controller
>
 Got it.  Thanks. :)

> Daniel



--
Best Regards
Li

IBM LTC, China System&Technology Lab, Beijing