Re: [PATCH 21/21] vdpasim: control virtqueue support

From: Jason Wang <jasowang@redhat.com>
To: Eugenio Perez Martin <eperezma@redhat.com>
Cc: Eli Cohen <elic@nvidia.com>, Michael Tsirkin <mst@redhat.com>,
	kvm list <kvm@vger.kernel.org>,
	virtualization@lists.linux-foundation.org,
	netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
	Cindy Lu <lulu@redhat.com>, Eli Cohen <eli@mellanox.com>,
	lingshan.zhu@intel.com, Rob Miller <rob.miller@broadcom.com>,
	Stefan Hajnoczi <stefanha@redhat.com>,
	Stefano Garzarella <sgarzare@redhat.com>
Subject: Re: [PATCH 21/21] vdpasim: control virtqueue support
Date: Mon, 25 Jan 2021 11:16:30 +0800	[thread overview]
Message-ID: <9be0b845-4d59-05fc-5575-347a03c5da52@redhat.com> (raw)
In-Reply-To: <CAJaqyWcRirQgz+n63rU2nYVH2RKqjQfwHGFLzOG=H46qRWuTog@mail.gmail.com>

On 2021/1/23 上午3:43, Eugenio Perez Martin wrote:
> On Tue, Jan 12, 2021 at 4:12 AM Jason Wang <jasowang@redhat.com> wrote:
>>
>> On 2021/1/11 下午8:26, Eli Cohen wrote:
>>> On Wed, Dec 16, 2020 at 02:48:18PM +0800, Jason Wang wrote:
>>>> This patch introduces the control virtqueue support for vDPA
>>>> simulator. This is a requirement for supporting advanced features like
>>>> multiqueue.
>>>>
>>>> A requirement for control virtqueue is to isolate its memory access
>>>> from the rx/tx virtqueues. This is because when using vDPA device
>>>> for VM, the control virqueue is not directly assigned to VM. Userspace
>>>> (Qemu) will present a shadow control virtqueue to control for
>>>> recording the device states.
>>>>
>>>> The isolation is done via the virtqueue groups and ASID support in
>>>> vDPA through vhost-vdpa. The simulator is extended to have:
>>>>
>>>> 1) three virtqueues: RXVQ, TXVQ and CVQ (control virtqueue)
>>>> 2) two virtqueue groups: group 0 contains RXVQ and TXVQ; group 1
>>>>      contains CVQ
>>>> 3) two address spaces and the simulator simply implements the address
>>>>      spaces by mapping it 1:1 to IOTLB.
>>>>
>>>> For the VM use cases, userspace(Qemu) may set AS 0 to group 0 and AS 1
>>>> to group 1. So we have:
>>>>
>>>> 1) The IOTLB for virtqueue group 0 contains the mappings of guest, so
>>>>      RX and TX can be assigned to guest directly.
>>>> 2) The IOTLB for virtqueue group 1 contains the mappings of CVQ which
>>>>      is the buffers that allocated and managed by VMM only. So CVQ of
>>>>      vhost-vdpa is visible to VMM only. And Guest can not access the CVQ
>>>>      of vhost-vdpa.
>>>>
>>>> For the other use cases, since AS 0 is associated to all virtqueue
>>>> groups by default. All virtqueues share the same mapping by default.
>>>>
>>>> To demonstrate the function, VIRITO_NET_F_CTRL_MACADDR is
>>>> implemented in the simulator for the driver to set mac address.
>>>>
>>> Hi Jason,
>>>
>>> is there any version of qemu/libvirt available that I can see the
>>> control virtqueue working in action?
>>
>> Not yet, the qemu part depends on the shadow virtqueue work of Eugenio.
>> But it will work as:
>>
>> 1) qemu will use a separated address space for the control virtqueue
>> (shadow) exposed through vhost-vDPA
>> 2) the commands sent through control virtqueue by guest driver will
>> intercept by qemu
>> 3) Qemu will send those commands to the shadow control virtqueue
>>
>> Eugenio, any ETA for the new version of shadow virtqueue support in Qemu?
>>
> Hi Jason. Sorry for the late response.
>
> For the notification part I have addressed all the issues of the RFC
> [1], except the potential race conditions Stefan pointed, and tested
> with vdpa devices. You can find at
> https://github.com/eugpermar/qemu/tree/vdpa_sw_live_migration.d/notifications.rfc
> . Since the shadow path is activated only through QMP and does not
> interfere with regular operation, I could post to the qemu list if you
> prefer. The series will be smaller if merged in steps.

Sure. Please post them.

>
> Adding the buffer forwarding on top should not take long.
>
> [1] https://lkml.org/lkml/2020/9/23/1243
>
> Thanks!

Thanks