OpenVMS driver in C

[Alex Chupahin]

Thank you very much for all,
I'll try to study this information.

[anonys]

Hi Alex,
whether you find the most appropriate example? Can share the experience about it?thank you!

[Hoff]

Anonys, you are in an OpenVMS operating system forum, and OpenVMS drivers are wildly different from the HP-UX drivers (and Unix drivers in general) that are the subject of your other recent question:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1252977

If you want to see what an OpenVMS device driver looks like, gain access to an OpenVMS system and look for this file specification:

SYS$EXAMPLES:*DRIVER.*

The Deathrow cluster has open access to OpenVMS; you can sign up for a free account over there. HP offers access to various test drive systems, each of which also has free login accounts available.

There are also driver examples in the OpenVMS device driver documentation -- the older documentation stuff is available in the archived documents section in the left navigation of the OpenVMS documentation web site:

http://www.hp.com/go/openvms/doc/

The newer driver manual (writing drivers in C) is not available on-line AFAIK. It's a for-purchase book.

There are various OpenVMS device drivers available in source code form on the OpenVMS Freeware and the other similar resources located at:

http://mvb.saic.com/

You can use Google site:mvb.saic.com to target a Google search against that site.

And again and having written drivers for both OpenVMS and for various Unix platforms, the OpenVMS device drivers are completely different than Unix drivers.

Would you be seeking assistance on some sort of formal training or class assignment or homework here?

[Bill Pedersen]

Am interested in getting my hands on the most recent version of "Writing OpenVMS Alpha Device Drivers in C" I understand there was a second 2nd Edition but when I have tried to find it from a technical bookstore out in San Diego they say it was never published. Does anyone have any better information? My current copy (version unknown) is in storage in California so am looking to get one since I have relocated to the East Coast.

Thanks,

Bill.

[Robert Brooks_1]

Am interested in getting my hands on the most recent version of "Writing OpenVMS Alpha Device Drivers in C" I understand there was a second 2nd Edition . . .

--

I can say with absolute certainty that a 2nd edition was neither written nor published.

I can say with near certainty that a 2nd edition will not be written.


-- Rob

[Hoff]

I'm aware of two or three folks that had thought about writing an update to the driver book, and none of these ever came to fruition.

Regardless of the available edition, you'll very likely end up using the OpenVMS source listings here, both for areas such as the V8.4-vintage I/O intercept mechanism that will probably require some digging as would other kernel changes, and there are parts of the existing driver book that have omissions and errors.

Beyond what Robert has already mentioned, I'm not aware of a public errata for the book. Those of us that have written device drivers tend to have our own errata, and I expect the authors of the driver book have similar collections of errata. (I carefully collected errata when I had thoughts there might be a third edition of the DCL book, for instance.)

[Bill Pedersen]

I fully agree with you Hoff. There is no substitute for the source listings of a particular build you are working with on developing a driver.

Thanks to all for the updates.

Bill.

[Robert Brooks_1]

V8.4-vintage I/O intercept mechanism that will probably require some digging as would other kernel changes,

--

The I/O intercept mechanism first appeared in V8.2 (and backported immediately to V7.3-2).

It has not been well-documented, although I hope to get it into the I/O User's Guide for V8.4.

For anyone who wants to "play nice" with multipath devices, the use of this supported mechanism is an absolute must.

It was developed in response to a problem with a 3rd party caching product that thought it knew more about I/O internals than it did, because multipath turned certain assumptions on its head.

-- Rob