[Info-vax] X86 first boot? - A Really Stupid Question

[Stephen Hoffman]

On 2019-03-31 23:58:44 +0000, Simon Clubley said:

> On 2019-03-29, Dave Froble <davef at tsoft-inc.com> wrote:
>> 
>> So, if many of the device drivers have been re-written by VSI, perhaps 
>> they could make them available.  Whether they would do so is another 
>> question.
> 
> Something more than the existing LRDRIVER would be nice. :-)

DQDRIVER and LRDRIVER work, but they're fairly limited in their designs 
and implementations.

And being hardware drivers, the designs are somewhat obfuscated by the 
details of the seemingly-inevitable hardware weirdnesses.

>> Whether they could or would make documentation available so third 
>> parties could write drivers might be a good question.
> 
> That's actually a rather good question.
> 
> When people write drivers for Itanium VMS at the moment, what 
> information sources do they use ?

The C driver book, the "step 2" driver book in the archived 
documentation, Jamie Hanrahan's Advanced Driver Techniques book, and 
other device drivers I've written.  And the errata that's been 
collected.

Access to the source listings can be useful for some aspects of the 
work, though whether VSI can or will offer that?

> Is the Writing OpenVMS Alpha Device Drivers in C book still the 
> starting point for Itanium ?
> My copy of that book says Copyright (c) 1996 Digital Equipment 
> Corporation.  Is there a later version of that book ?

That's the "latest" edition.

> Will the information in that book still be applicable for x86-64 VMS or 
> will there need to be (and if so, will there actually be) another 
> version of that book ?

There's also the question around whether third-party drivers and other 
kernel extensions can or will or should even be supported, absent VSI 
assistance and/or absent code-signing.

There do seem to be fewer third-party device drivers around.  There are 
various potential reasons for that apparent decline.  Pragmatically, 
USB hardware and that networking hardware can provide a fair amount of 
what used to require a dedicated adapter or controller.

Wouldn't surprise me to learn that we'll have to work with VSI 
directly, at least for the first wad of third-party device drivers 
written for OpenVMS x86-64.  That was the case with the OpenVMS Itanium 
port and the "step 1" device driver era.

All fodder for VSI to ruminate upon.


-- 
Pure Personal Opinion | HoffmanLabs LLC