Re: Sider (Xebec/Datamac SASI) drive and UCSD Pascal 1.1 (Apple Pascal)

["John B. Matthews" <nospam@nospam.com>]

In article <40C3EFFA.B5AE5BAF@REMOVETHIScybertheque.org>,
 msg <msg@REMOVETHIScybertheque.org> wrote:

> Note: our work involves Xebec software V2.2 and firmware V3.1.
> 
> > Progress made; I can now access the Sider volumes from a floppy-booted
> > system.
> 
> In summary:  After setup of the Pascal partitions and volumes using
> "INSTALL PT#4", create a Pascal 1.1 boot floppy which contains the
> 'SYSTEM.ATTACH' and the proper 'ATTACH.DRIVERS' and 'ATTACH.DATA';
> boot from this floppy to access the Sider volumes.
> 
> One may now copy files from floppy to the hard disk's volumes.
> 
> Problems remain in creating the proper Pascal boot area on the
> Sider. The program 'XBOOT.COM.CODE' will write a copy of the
> running system (including the bound hard disk drivers) to the
> boot area; however when this system is cold booted, it expects
> volume #4 to be in floppy drive A regardless of what unit numbers
> are assigned to the hard disk volumes (using the Xebec utility
> for that purpose -- resident on the DOS3.3 VOL001).  This boot
> area is evidently separate and distinct from the root volume
> boot sectors which are unused.
> 
> The only file referenced on the floppy (unit #4) is 'SYSTEM.PASCAL';
> there is no need for the driver 'attach' files to reside on
> the floppy or on the 'root' hard disk volume; this explains
> their absence from our original hard disk.

Would this suggest that the ATTACHed driver was only needed to 
get files on to the hard drive in the first place? Could Xebec 
be using firmware for subsequent operation?

> If one could change the assignment of the root volume after
> booting from floppy, from #4 to (say) #9, and then write the
> boot area, the cold boot process should work.

This is how ATTACH itself works. It manipulates entried in a 
user driver jump vector table, pointed to by $E8/E9. In 1.1, 
this points to $FE80, and maps to the last block of 
SYSTEM.APPLE.  This table is followed by the unit table at 
$FEB0. Presumably Xebec is patching the table on disk to mount 
selected volumes at boot time.

The volume named ATTACH: has a utility named ATTACHUD.CODE to 
create new ATTACH.DATA files, and I see a utility named 
SHOWAD.CODE that looks like it might list the contents of your 
existing ATTACH.DATA file.

> Can anyone help with this?  Or perhaps someone versed in the
> internals could advise where to look in memory or in the
> written system image for the root volume parameters,
> and how to patch them.

The references already mentioned (Randy Hyde's book "p-Source", 
and Barry Hynes "Attach-Bios" paper) will only tell you how a 
system might be built. That might enable you to reverse engineer 
the Xebec setup.

> Regards,
> 
> Michael

John
----
jmatthews at wright dot edu
www dot wright dot edu/~john.matthews/