[Coco] System gen problem

[Boisy G. Pitre]

On Feb 9, 2010, at 10:56 AM, Steven Hirsch wrote:

> On Tue, 9 Feb 2010, Aaron Wolfe wrote:
> 
>>> I just figured out that I needed to use the -d flag when taking a CVS update
>>> on nitros9.  Now, it has the sources for the DW modules so I'll try using
>>> those.
>>> 
>> 
>> You will not want to use the DW modules from CVS unless you also are
>> using DriveWire server from the DW CVS (the new Java version of the
>> server).  The new DW drivers will not work with any of the old DW
>> server versions (the individual linux/mac/win type).
> 
> Ok, thanks for the warning.

Steve,

FYI, I have just updated the SOURCE of the three platforms (win, linux & mac) to accommodate the OP_VPORT_READ command.  It just responds with two zeros and doesn't support any of the virtual port features of DriveWire 4.  I don't know what server you're using, but if it's the Linux one, you should be able to CVS update and recompile then use the new drivers. 

Again, this is untested.

> Thinking logically about the problem, what happens when you request, e.g. a directory of a drive?  I'm assuming that the kernel walks the RBF headers in memory until it finds one with the name of the requested volume.  Is there correct so far?
> 
> If I'm not off-base, then it appears that the kernel is either not finding the '/x0' module OR the module is missing an important bit of information required to flag it as operational.  If I run ident on the generated boot file, I can see 'x0' so I assume the name is correct.  What else needs to be there for the kernel to say "oh, yeah, you are a disk descriptor"?

The E$BNam error has something to do with linking to the dw3 subroutine module, I believe. 

One quick thing: once you boot to your hard drive, what happens if you just iniz /x0 (and not do a dir on it?)  Does an error come back?

> I'm sure no one else will ever have the problems I'm seeing.  Over the 25-years I've been involved with computers, I've become extremely good at breaking things.  I might go so far as to claim the title of "World's Most Dangerous Beta Tester"!  Unfortunately, the things that break for me generally never cause problems for anyone else and generally cannot even be reproduced by the author :-).
> 
> Ah well..  Back to earning my living.  Fortunately, the VLSI Design Automation software I get paid to write is proving more tractable. Today, that is.
> Steve
> 
> 
> -- 
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco

--
Boisy G. Pitre
http://www.tee-boy.com/