[Info-vax] Software does wear out, was: Re: Raid Controller in I64 ans Alpha(MSA$UTIL)

[Stephen Hoffman]

On 2013-12-02 17:49:05 +0000, Simon Clubley said:

> Interesting. I knew some emulators provided replacement drivers to 
> allow the OS in the emulated environment to better interact with the 
> emulator, but I must admit I didn't realise the replacement drivers had 
> developed as far as providing (in this case) DPI on incoming network 
> packets.

This isn't deep packet inspection, it's that the emulator can include 
calls to host code or assists for various common operations.  It's a 
whole lot of work for the emulator to present the appearance of a 
hardware controller to a guest driver, and a whole lot less work for a 
guest driver to be modified to pass more directly through and into a 
host driver or host device.

IBM VM has been using virtual disks and other virtual devices for eons.

>From the perspective of VMS software, this might be an analog to the LD 
driver that goes directly to the underlying physical devices, rather 
than going directly to the VMS drivers as LD does.  VMS application 
software wouldn't notice that difference, except the I/O will be faster 
and with less overhead.

IIRC, the FX!32 translation environment worked by passing the calls 
through to the underlying NT, which is conceptually similar to 
paravirtualization with a VM,

Like ZFS does with file systems and drivers, paravirtualization 
"breaks" the traditional layering of a VM, the guest, and the host 
operating system.

> It would be interesting to read the documentation for a emulator which 
> provides this support and see what the drivers are now capable of 
> doing.  Do you have a example to hand ?

A search for paravirtualization should kick over at least few 
discussions.  <http://en.wikipedia.org/wiki/Paravirtualization>


-- 
Pure Personal Opinion | HoffmanLabs LLC