Re: [virtio-dev] [PATCH v2 1/1] virtio-ism: introduce new device virtio-ism

[Cornelia Huck <cohuck@redhat.com>]

On Thu, Jan 12 2023, Xuan Zhuo <xuanzhuo@linux.alibaba.com> wrote:

> On Thu, 12 Jan 2023 09:42:05 +0100, Cornelia Huck <cohuck@redhat.com> wrote:
>> On Thu, Jan 12 2023, "Michael S. Tsirkin" <mst@redhat.com> wrote:
>>
>> > On Thu, Jan 12, 2023 at 10:01:25AM +0800, Jason Wang wrote:
>> >> On Wed, Jan 11, 2023 at 11:11 PM Halil Pasic <pasic@linux.ibm.com> wrote:
>> >> >
>> >> > On Wed, 11 Jan 2023 19:08:53 +0800
>> >> > Xuan Zhuo <xuanzhuo@linux.alibaba.com> wrote:
>> >> >
>> >> > > > > +ISM(Internal Shared Memory) device provides the ability to share memory between
>> >> > > > > +different VMs launched from the same entity.
>> >> > > >
>> >> > > > Launched by instead of from? Maybe introduce a catchy name for the
>> >> > > > "entity that launched the VMs" and prevent oversimplification by
>> >> > > > explaining any shortcomings of the name if any in one place. Host would
>> >> > > > be one candidate, VMM another.
>> >> > >
>> >> > > Cornelia Huck wrote:
>> >> > >
>> >> > >       Is there a way to avoid the term "host" (throughout this document)?
>> >> > >       IIUC, you need the uniqueness within the scope of the entity that
>> >> > >       launches the different instances that get shared access to the regions
>> >> > >       (which could conceivably a unit of hardware?)
>> >> > >
>> >> > > And I think she is right, so I am trying to remove the term HOST.
>> >> > >
>> >> > > Do you have better opinions? I think VMM is not particularly suitable.
>> >
>> > I think fundamentally from spec POV memory is shared between devices.
>> > How sharing is accomplished guest does not care so neither should the
>> > spec. Can some RDMA tricks be used for synchronisation behind the
>> > scenes? Maybe, the spec does not care. But we can give an example.
>> >
>> > So something like:
>> >
>> > 	An ISM(Internal Shared Memory) device provides the ability to
>> > 	access memory shared between multiple devices. This allows low-overhead
>> > 	communication in presence of such memory. For example, memory can be
>> > 	shared with guests of multiple virtual machines running on the same
>> > 	host, with each virtual machine including an ISM device and with
>> > 	the guests using the ISM devices to access the shared memory.
>> >
>> > what do others think?
>>
>> I like that: we don't want to talk about hosts/VMMs/etc. as we
>> fundamentally deal with devices and drivers, but sharing between guests
>> is of course the obvious use case.
>>
>> I'm just wondering how best to express the uniqueness scope, is it per
>> (ISM) device?
>
> No, each vm has at least one separate device. The devices in a host form
> an uniqueness scope.

Should we call it a 'group', then? A host would be an example of such a
group.