[EXTERNAL] RE: [PATCH v2 0/4] Add new headers for Hyper-V Dom0

[Michael Kelley]

From: MUKESH RATHOR <mukeshrathor at microsoft.com> Sent: Monday, November 11, 2024 3:11 PM
> 
> On 11/11/24 13:28, Michael Kelley wrote:
>  > From: MUKESH RATHOR <mukeshrathor at microsoft.com> Sent: Monday, November 11, 2024 10:53 AM
>  >>
>  >> On 11/10/24 20:12, Michael Kelley wrote:
>  >>   > From: Nuno Das Neves <nunodasneves at linux.microsoft.com> Sent:
>  >> Thursday, November 7, 2024 2:32 PM
>  >>   >>
>  >>   >> To support Hyper-V Dom0 (aka Linux as root partition), many new
>  >>   >> definitions are required.
>  >>   >
>  >>   > Using "dom0" terminology here and in the Subject: line is likely to
>  >>   > be confusing to folks who aren't intimately involved in Hyper-V work.
>  >>   > Previous Linux kernel commit messages and code for running in the
>  >>   > Hyper-V root partition use "root partition" terminology, and I couldn't
>  >>   > find "dom0" having been used before. "root partition" would be more
>  >>   > consistent, and it also matches the public documentation for Hyper-V.
>  >>   > "dom0" is Xen specific terminology, and having it show up in Hyper-V
>  >>   > patches would be confusing for the casual reader. I know "dom0" has
>  >>   > been used internally at Microsoft as shorthand for "Hyper-V root
>  >>   > partition", but it's probably best to completely avoid such shorthand
>  >>   > in public Linux kernel patches and code.
>  >>   >
>  >>   > Just my $.02 ....
>  >>
>  >> Hi Michael,
>  >>
>  >> FWIW, hyperv team and us are using the term "dom0" more and more to
>  >> avoid confusion between windows root and linux root, as dom0 is
>  >> always linux root. I did a quick search, and "dom0" is neither
>  >> copyrighted nor trademarked by xen, and I'm sure the fine folks
>  >> there won't be offended. Hopefully, [Hyper-V] tag would reduce
>  >> the confusion.
>  >>
>  >> Just my $0.1
>  >>
>  >
>  > Yeah, "dom0" certainly fits as shorthand for the rather ponderous
>  > "Linux running in a Hyper-V root partition". :-)
>  >
>  > But even using "Hyper-V dom0" to add clarity vs. Xen dom0 seems
>  > to me to be a misnomer because Hyper-V dom0 is only conceptually
>  > like Xen dom0. It's not actually an implementation of Xen dom0.
>  > Let me give two examples:
>  >
>  > 1) Hyper-V provides VMBus, which is conceptually similar to virtio.
>  > But VMBus is not an implementation of virtio, and we don't call it
>  > "Hyper-V virtio".  Of course, "VMBus" is a lot shorter than "Hyper-V
>  > root partition" so the motivation for a shorthand isn't there, but still.
>  > If Hyper-V should ever implement actual virtio interfaces, then it
>  > would be valid to call that "Hyper-V virtio".
>  >
>  > 2) KVM has "KVM Hyper-V", which I think is valid. It's an
>  > implementation of Hyper-V interfaces in KVM so that Windows
>  > guests can run as if they are running on Hyper-V.
>  >
>  > I won't speculate on what the Xen folks would think of "Hyper-V
>  > dom0", especially if it isn't an implementation that's compatible
>  > with Xen dom0 functionality.
>  >
>  > As for "more and more" usage of "dom0" by your team and the
>  > Hyper-V team:  Is that internal usage only? Or usage in public mailing
>  > lists or open source projects like Cloud Hypervisor? Again, from
>  > my standpoint, internal is internal and can be whatever is convenient
>  > and properly understood internally. But in public mailing lists and
>  > projects, I think "Hyper-V dom0" should be avoided unless it's
>  > truly an implementation of the dom0 interfaces.
>  >
>  > That's probably now $0.10 worth instead of $0.02. :-) And I'm
>  > not the decider here -- I'm just offering a perspective.
> 
> "dom0" is neither a technology nor a protocol. It simply means initial
> domain (which on xen happened to be domid of 0, could have been 1). This
> is created during boot, same as linux root on hyperv, and is privileged
> domain same as xen. Even in KVM world, I've heard many folks refer to
> the host as kvm dom0...
> 
> Given the mix of windows and linux with l1vh and nested, dom0 is helping
> in conversations internally, and I'm sure it will keep percolating
> externally.
> 

OK, fair enough. My perspective is probably more limited than
yours as my experience with "dom0" is exclusively with the Xen
code in a Linux kernel environment. I just haven't seen "dom0"
used elsewhere, but that certainly doesn't mean it's not being
done.

If the decision on the Microsoft side is that introducing "Hyper-V
dom0" terminology makes sense, I won't object further, though
I would think there should be some kind of reconciliation with
existing code/comments/documentation that uses "root partition".

Michael