80-Bus News

relating to RAM FA00H to FC00H was on the disk, the only possible culprit was the COLD BOOT. This little program is only about 50H bytes long and was soon taken to pieces and once again a sector counter was found to be responsible. The sector count byte at address 0009 had been 11H (17 decimal) and SIMON had already loaded the first sector for a total of 18. To make the BOOT load the full 20 I set this byte to 13H using a PEEK program on the actual disk. This time, when booted in drive A, the system worked, and that is nearly the end of the story. The byte at 909H in the CPMSYS64.COM file was changed to 13H so that all copies of the BOOT LOADER on the System track would now load the full BIOS.

(A COLD BOOT modified in this way should NOT be used to load the origional Gemini CP/M to a RAM address above D800H. Since the origional CBIOS is just over 2.5K jong, the Gemini CP/M would normally start higher in RAM, unless spcifically configured for say 63K. If loading were tried to the higher address using the modified BOOT, then the HL register pair, that hold the destination address of the bytes read in from the disk, would be incremented past 0FFFFH, and would overflow back to 0000H. The incoming data would load to low RAM, overwriting the BOOT, and soon cause a CRASH.)

To complete the job only one or two tasks remained. Richard Beal had included a little routine in the Load/​Relocate module of SYS which he calls “cinit”, which is there to remove any ‘stray’ input characters particularly when using Virtual Disk. I have successfully run CPMSYS64 without this routine, as SYS 11 does not support my MAP RAM card, and I do not use my G802 or any RAM B’s as V.Disk. It might be necessary to include it with a later version of SYS that does support my MAP RAM though. The other task concerns ‘Sign On’ messages. The origional CBIOS messages have been wiped out by the new BIOS, and since the Load/​Relocate Module of SYS is not now used, the comprehensive list of features that SYS displays on execution has also gone. It is nice to have an impressive list of system Software and Hardware features displayed on Cold Boot, to ‘awe’ the Plastic Box Brigade, and some form of Copyright mark should also be shown.

There are several places where it is possible to restore “signon” message routines, but “cinit” should be run by the COLD BOOT, or by SIMON after CP/M is in RAM. Since SIMON is overwritten by CP/M in larger size systems though, the latter course is not possible directly. There is currently enough room to put “signon” routines and messages into the BIOS area, but this would take up space needed for future expansions of SYS, so the BIOS is not the best choice.

There are app. 400 bytes free in the first disk sector that might be used for initialization routines, since the BOOT is only about 50H bytes long. There is another ‘snag’ though. SIMON loads the first disk sector into high memory, and then does a ‘copy’ to RAM at 0000H before making a jump to location 0002H, to start the COLD BOOT. The problem is that only 80H bytes are copied from 0FC2EH to 0000H so any text messages would be largely left behind in high memory to be overwritten by CP/M as it is loaded. This means that to use the free space in the first sector, it is necessary to modify SIMON so that the whole of the sector is copied to low memory. The boot could then be altered so that it executes “signon” and any other initialization routines before jumping to start CP/M.

–––––––––––––*****–––––––––––––

If however, SIMON is going to be altered, then the EPROM programming effort might as well be made worth while. For convenience my present SIMON and the short program that copies SIMON to RAM and pages out the EPROM card are all in a 2716 ROM on the card. There is nearly 1K of space free in the 2716. I decided to use up this free space by adding to and by modifying SIMON. The ‘Mods’ so far are :–