Re: [virtio-comment] Re: [PATCH V2 2/5] Introduce feature bit VIRTIO_F_TRANSPT_VQ_MDEV

["Zhu, Lingshan" <lingshan.zhu@intel.com>]

On 8/2/2022 8:55 PM, Michael S. Tsirkin wrote:
> On Tue, Aug 02, 2022 at 07:54:41PM +0800, Zhu, Lingshan wrote:
>>
>> On 8/2/2022 2:52 PM, Michael S. Tsirkin wrote:
>>> On Tue, Aug 02, 2022 at 11:17:51AM +0800, Zhu, Lingshan wrote:
>>>> On 8/2/2022 4:27 AM, Michael S. Tsirkin wrote:
>>>>> On Mon, Aug 01, 2022 at 05:32:13PM +0800, Zhu Lingshan wrote:
>>>>>> This commit introduces feature bit VIRTIO_F_TRANSPT_VQ_MDEV,
>>>>>> which depends on VIRTIO_F_TRANSPT_VQ.
>>>>>>
>>>>>> Feature bit VIRTIO_F_TRANSPT_VQ_MDEV indicates that the managed
>>>>>> devices are created, configured and destroyed through the
>>>>>> transport virtqueue, the transport virtqueue is the
>>>>>> transport layer for the managed devices.
>>>>> I'd avoid the term MDEV if possible, it's already used in virt contexts.
>>>> yes, let me work out a better name, it was vdev (virtual device) before, and
>>>> maybe
>>>> sdev(sub device) or a better one.
>>> Besides mdev can thinkably stand for both managed and management.
>>> Nvidia patches have the concept of a device group and
>>> group member but do not have a name for the device that controls
>>> the group. Admin device? Control device?
>> maybe management device?
> This term has been overused in this context with people saying things
> like "the management interface is used by management tools to
> manage managed devices" to the point where I'd
> rather we avoided it altogether.
They are like the parent device and the sub-device, but this is not good 
enough.
It has ever been suggest to use "upstream device" and "downstream 
device" to replace "master device" and "slave device",
but looks not good enough either.

The concept should be the management device and the managed device, we 
never use these terms in virtio spec,
any better suggestions?

>
>>> Also, creating/destroying is one thing that seems to only apply to
>>> SIOV. configuring seems to apply to SRIOV as well?
>> In this series, yes, because actually we can not provision a SRIOV VF with
>> dynamic config, and VF provisioning is done through SRIOV cap
>>
>> Thanks
>> Zhu Lingshan
> I am not sure what dynamic config means here exactly. Yes creating a VF
> itself is not possible. But allocating VQs to it and
> e.g. configuring the MAC address of a networking VF can be
> done dynamically.
I think this config can only be done through the CVQ after provisioning.
>
>>>
>>>
>>>>>> Signed-off-by: Jason Wang <jasowang@redhat.com>
>>>>>> Signed-off-by: Zhu Lingshan <lingshan.zhu@intel.com>
>>>>>> ---
>>>>>>     content.tex | 12 ++++++++++++
>>>>>>     1 file changed, 12 insertions(+)
>>>>>>
>>>>>> diff --git a/content.tex b/content.tex
>>>>>> index 969ca46..143949e 100644
>>>>>> --- a/content.tex
>>>>>> +++ b/content.tex
>>>>>> @@ -2968,6 +2968,18 @@ \subsection{Managed Devices Dscovery}\label{sec:Virtio Transport Options /Virtio
>>>>>>     Managed devices are created and discovered through a transport virtqueue.
>>>>>> +\subsection{Transport Virtqueue Features}\label{sec:Virtio Transport Options /Virtio Over Transport Virtqueue / Transport Virtqueue Features}
>>>>>> +
>>>>>> +Feature bit VIRTIO_F_TRANSPT_VQ_MDEV indicates that the managed devices are created, configured
>>>>>> +and destroyed through the transport virtqueue.
>>>>> Feature bit being present where? On the management device?
>>>> Yes, the management device, I will add this.
>>>>>> +This means the transport virtqueue is the transport layer for the managed devices.
>>>>> confused. how does this differ from VIRTIO_F_TRANSPT_VQ?
>>>> VIRTIO_F_TRANSPT_VQ means the management device offers a transport vq.
>>>> VIRTIO_F_TRANSPT_VQ_MDEV is a feature of the transport vq, means the
>>>> management device create / destroy and configure the managed device via the
>>>> transport vq.
>>>>
>>>> This reminds me that maybe I should use the term vdev(virtual device), so
>>>> that other kinds of devices can add their own feature bits under
>>>> VIRTIO_F_TRANSPT_VQ
>>>>
>>>> Thanks,
>>>> Zhu Lingshan
>>>>>> +
>>>>>> +Feature bit VIRTIO_F_TRANSPT_VQ_MDEV depends on VIRTIO_F_TRANSPT_VQ
>>>>>> +
>>>>>> +\devicenormative{\subsubsection}{Transport Virtqueue Features}{Virtio Transport Options / Virtio Over Transport Virtqueue / Transport Virtqueue Features}
>>>>>> +
>>>>>> +The management device MUST not accept VIRTIO_F_TRANSPT_VQ_MDEV if VIRTIO_F_TRANSPT_VQ is not negotiated.
>>>>>> +
>>>>>>     \chapter{Device Types}\label{sec:Device Types}
>>>>>>     On top of the queues, config space and feature negotiation facilities
>>>>>> -- 
>>>>>> 2.35.3
>>>>> This publicly archived list offers a means to provide input to the
>>>>> OASIS Virtual I/O Device (VIRTIO) TC.
>>>>>
>>>>> In order to verify user consent to the Feedback License terms and
>>>>> to minimize spam in the list archive, subscription is required
>>>>> before posting.
>>>>>
>>>>> Subscribe: virtio-comment-subscribe@lists.oasis-open.org
>>>>> Unsubscribe: virtio-comment-unsubscribe@lists.oasis-open.org
>>>>> List help: virtio-comment-help@lists.oasis-open.org
>>>>> List archive: https://lists.oasis-open.org/archives/virtio-comment/
>>>>> Feedback License: https://www.oasis-open.org/who/ipr/feedback_license.pdf
>>>>> List Guidelines: https://www.oasis-open.org/policies-guidelines/mailing-lists
>>>>> Committee: https://www.oasis-open.org/committees/virtio/
>>>>> Join OASIS: https://www.oasis-open.org/join/
>>>>>