[Info-vax] clock problems with OpenVMS x86 on VirtualBox

Dan Cross cross at spitfire.i.gajendra.net
Thu May 18 10:45:24 EDT 2023


In article <0968d262-1b9c-46b9-b138-54387b83dafan at googlegroups.com>,
gah4  <gah4 at u.washington.edu> wrote:
>On Wednesday, May 17, 2023 at 5:25:40 PM UTC-7, Arne Vajhøj wrote:
>
>(snip)
>
>> A hypervisor is not a host OS. 
> 
>> Not in terminology and not in functionality. 
> 
>If MS-DOS is an OS, even though it doesn't do most things that OSs should do,
>a hypervisor doesn't seem so far off.  
>
>And if you consider the way CMS runs under VM, then VM is even more OS-like.
>VM does login/logout, disk space allocation, spooling, accounting, and probably
>a few other OS things that I forgot.
>
>And with a little work, you can run programs directly, without CMS in between,
>not so much different than the way MS-DOS runs programs. 

Just so.

I like Mothy Roscoe's definition of an operating system:

"The operating system is that body of software that:
 - Multiplexes the machine's hardware resources,
 - Abstracts the hardware platform,
 - Protects software principals from each other,
   (using the hardware)"
(https://www.usenix.org/conference/osdi21/presentation/fri-keynote)

If one considers "software principles" to mean the guest
operating system, and "abstracts the hardware platform" to mean,
say, provision of device models to the guest, then hypervisors
(even those of the so-called "Type-1" variety) meet this
definition.  Indeed, a type-1 HV may match this definition
even more closely, since it must provide some abstraction
whereby some _other_ software component can provide services to
guests _and_ direct control of the VMM itself vis, say,
scheduling and allocation policies; usually this is in the form
of a sepecially blessed VM (Hyper-V calls this the "root VM",
while Xen calls it "Dom0"), but that needn't be the case.
Importantly, Type-1 hypervisors also drive the host hardare in a
way that no other software component _can_.

This is why people who actually build hypervisors see the
distinction between "Type-1" and "Type-2" as more of an
convenient abstraction than a concrete difference.  For example
_technically_ Hyper-V is a type-1 hypervisor, even though in
practice it's completely inseparable from Windows; similarly,
KVM is often billed as a Type-1 HV but is intimately tied to the
OS it runs alongside with.  In contrast, z/VM provides many
services typically associated with a host OS and IBM calls it a
type-2 hypervisor, even though it runs directly on the bare
hardware.

	- Dan C.




More information about the Info-vax mailing list