The Latest Version of Toad Hall's TOADCPM.COM
(Version 2.2, 10 Dec 83)
Using various patches (credit given below), I've patched CPM
2.2 to add some things I've found very useful:
PRINTER: In Version 2.1, it's configured for a 9600 baud
printer, using the Data Ready protocol (donno if it's the best
for my Mannesman Tally, but it works!). Version 2.2 now has
XON/XOFF, which works just fine using exactly the same cabling.
Seems to cause a bit less delay when simultaneously printing and
making keyboard entries during UNSPOOL or WordStar's PRINT
utility. [Note: As of 10 Dec I went back to Xon/Xoff, just to
see if there was a difference. None, really.]
ASSEMBLY SETTINGS: (Look at the TOADBIOS.ASM listing for details
-- it's for the XON/XOFF, 19200 baud Version 2.2, but all changes
are fully documented.) You'll see it's for a 64K system with
DEBUG set to ON to give me various sizes and locations during MAC
assembly, and NONSTANDARD set to 1 (donno why; I thought Morrow's
BIOS was conventional, but I'm just following the advice of my
vendor/-confidant Jack Long at Cost Plus Computers -- and he's
batting a thousand so far!). The original CBIOS&.ASM had at one
point a value set to DECISION I I/O instead of Mult/IO; suspecting
this might be giving me some problems in using the clock, I
changed it to Mult/IO, which is exactly what's in my Decision I.
DUPLICATE *.COM FILES: I dislike having to duplicate .COM
programs (utilities and other files) in different User Areas and
disks when my trusty 5 Meg hard disk is just sitting there as the
default A disk most of the time.
I've used DUPUSR.COM (a program to put a kind of pointer in
each User area directory so you get the name of a program or file
in that directory, but not the actual file. This means no actual
disk storage space used.), but DUPUSR has one SEVERE AND
DANGEROUS problem: if you erase a DUPUSR-created file name and
don't hit CTRL C right after the erasure, you can have some
SEVERE directory damage! About the third time I trashed my hard
disk directory through carelessness like this, I decided there
had to be a better way!
I read an article out in Netland by Neil Maron that gave a
patch to make the system look on drive A: if your .COM file isn't
on the logged in drive. Additionally, if it isn't on A: in the
same User Area, it'll look to User Area 0 for that file before
giving up. This means I can keep all the common utilities and
.COM files in User Area 0 on my hard disk, configure my CP/M to
always use the hard disk as A:, and I'm in business.
There are only a couple of places where this fails me.
a. WordStar: You can have WS.COM on disk A, User Area 0,
all right, BUT you must have the *.OVR files it uses in the
default disk (A for me) and the User Area you're working in! WS
can be installed or patched to look to a default disk for its
.OVR files, but it does NOT know how to look to User Area 0! No
problem, really -- this is where I take the chance and use
DUPUSR. I DUPUSR the .OVR files to all appropriate User Areas
where I might want to run WS.COM. I naturally use STAT to make
the .OVR files Read/Only and $SYS so they don't clutter up my
directory and I won't erase them by accident. I don't bother
copying WS or the .OVR files to other disks, but instead depend
on the default disk (A).
b. MBASIC (Microsoft BASIC v. 5.1) doesn't seem to want to
accept a compound command line like MBASIC STARTREK if MBASIC
isn't actually on that disk or in that User Area. (I think it's
a function of Neil's patch only looking at the .COM file name and
not handling the STARTREK extension properly.) No big problem --
I have two options: (1) RUN MBASIC (which works just fine,
reaching out to A: and then if necessary User Area 0 on A: to
find MBASIC) and then when in MBASIC, LOAD or RUN whatever I
want; or (2) DUPUSR MBASIC into necessary User Areas so I'll know
it's there. (Again protect it with R/O and $SYS.)
c. SUBMIT may give some problems. So far I've had none, but
there are lots of fancy applications, modified SUBMITs, etc. out
there where this may hang up. If there's a problem with any
SUBMIT applications, just DUPUSR it into the User Area requiring
SUBMIT applications and forget about it.
d. My library is a bit limited, but if you have any other
fancy programs that do lots of overlaying, calling in other
programs, etc., you'd better not depend on this A: User Area 0
patch working correctly. I suspect Visicalc, DBaseII, and those
kinds of application programs won't take kindly to it!
One thing -- if you DON'T have your hard disk set up as A,
then you need to change those JMPs around. Make CP/M first look
to User Area 0 on the present disk, and then to Drive A. That
way, if you're working in your hard disk, it'll check out User
Area 0 there first before running off to look at some poor
innocent floppy.
USER AREA WITH > PROMPT: I HATE not knowing which User Area I'm
in, and I've had some real problems. Examples: You get the
infamous BDOS ERR ON d: error message; you get kicked back to
User Area 0 automatically; you don't know or remember it and
erase or do something to a file; and successfully trash the very
special, unique, and wonderful file in User Area 0 you had NO
intention of messing with!
In any case, it's a pain having to go back to the right User
Area again. The original code was written by Bruce Kendall
(address unknown) 7-12-80, and tightened up by Bruce Ratoff
11-17-80. I made all the Morrow Decision CP/M-specific stuff.
Basically, it patches a little CALL into the portion of your
CCP that prints the > prompt, reaches out to a wee little patch
that was poked into an area of NOPs in the CCP that gets the User
Area and then displays current disk, user area (in hex -- sorry
abot that -- had to keep it small), and > prompt, and goes right
on about its business! Real simple. [Note: I grabbed a little
routine from another program (donno where), and made this patch
now print the User Area in decimal after the disk. Couldn't do
this with a "hot patch" in the actual BIOS; had to do it in
assembler in the source code. If you HAVE source code, great!]
The TOADCPM.COM on this disk (assuming you have the disk)
includes that patch. The code is at the end of this article if
you want to implement it yourself in your own BIOS. It's
definitely nice to have, and not at all hard to implement. Note
that I did NOT make this a stand-alone little program that'll
reach out and patch your CP/M in memory. You'll have to use DDT
for this, and if you aren't sure of the procedures, a step-by-
step (hopefully eminently logical and easy to follow) procedure
is outlined right after the source code.
The memory locations given are for Morrow Design's CP/M for
their Decision I (where the CCP starts at 0B00H in the DDT image,
and at D400H in real memory). If your CCP starts somewhere else,
you'll have to figure out your own locations for COUT (console
out), GTCMD (Get Command), PATCH (my patch in the NOP area of the
CCP), etc. Not very hard -- you can use the H tool in DDT to
figure out some offsets. (If you're doing these hacks in your
own CCP and BIOS, it's time for you to figure out DDT, offsets,
DDT images, real images, etc. anyway!)
AUTOSTART: There have been several articles around on how to
make a "turn-key" system with CP/M and its "Autostart"
capability. Basically, they involve making a version of CP/M
that has a .COM file name already in the CCP where CP/M looks for
a program name to execute. When you boot up (cold or warm,
depending), that program will run. One article fully explaining
it is "'Turn-Key' CP/M Systems" by James J. Frantz, published in
Creative Computing, December 1979. (It's for CP/M Version 1.4,
but not to worry - that part still is the same in Ver 2.2.)
One little problem with this, though! Morrow's CBIOS& has a
kind of switch (variable name AUTOST) in the source code that you
have to turn on for Autostart to work! You have several options:
Never, only on cold boot, only on warm boot, and on both cold and
warm boot. This version is engaged for warm and cold boot. The
default command "AUTO" is plugged in at the front of the CCP, but
will have absolutely no effect so long as you don't have a
AUTO.COM file in User Area 0! If you do ... it'll run! (You can
get rid of this with DDT by replacing from 0B07H (DDT image) up
to the remaining 20's with 20, and then SAVE 48 TOADCPM.COM.
If you're hacking up a special CP/M for an Autostart
program, the warm boot and cold boot commands can be entered as
DB's right around page 12 of Morrow's CBIOS Revision E for
Version 2.2 (Mar 4 1982) as provided with my Decision I. In
memory, if you want to poke it into your CCP, it's right at the
very start of your CCP (0B00H in the DDT image on my system)
where you can see the Digital Research copyright.
I'm not going to give that full procedure here -- plenty of
other references out there already published. But you can put it
right in your CBIOS while editing it a lot easier than looking up
hex and poking into memory. Remember, though -- just poking in
the Autostart command (with its accompanying requirements, like
length of command in CCP+7, and a NOP or 00 after the command) is
NOT sufficient with Morrow's CBIOS! You HAVE to turn that Auto-
start switch on! (Drove me crazy for a while until I figured out
that was the problem!)
COLDBOOT TO YOUR HARD DISK: My system originally used an 8"
floppy to cold-boot, and then I had to run "BOOTMW" to make my
hard disk A:. This was a bit of bother, and I accidentally found
the answer: Load the Hard Disk version of CP/M on the floppy!
Morrow supplied me with MOVCPM-5 to create a CP/M for my 5
Meg hard disk, and that's what I used (SYSGENing it to the hard
disk, etc.). Fine and dandy -- saved a nice CPM-HD on file.
Then I accidentally used DDT to do some patches on THAT version
of CP/M, saved it, and SYSGENed it onto an 8" floppy! (Never
would have done it on purpose -- obviously a hard disk CP/M
shouldn't be on a floppy!) Damned if it didn't work just fine!
I put in the floppy, do a reset, she boots, and bingo -- the cold
boot message comes on, showing the HDCMA hard disk as A:, and the
8" as B:. And in fact, that's what's happened -- not just a
bogus message. Works fine, and I've had no problems whatsoever.
One little thing: You need the new improved CP/M with
improved TDBIOS on your hard disk too, else the warm boot just
boots up the old normal CP/M. So SYSGEN TOADCPM (or your version
of your hard disk CP/M with TOADBIOS.ASM) onto both your floppies
and your hard disk. Works - guaranteed. (Keep a couple of
floppies around with a normal (i.e., A: is a floppy) just in case
your hard disk goes down some day!)
TERMINAL/SERIAL BAUD UPGRADE: Version 2.2 relies on a different
approach. Jack Long at Cost-Plus said he tried to set the
switches on the CPU board to have the Decision I do a cold boot
from the hard disk. Worked, OK, but then refused to recognize
any other disks! This put me off for a while, but found my
technique of booting from a floppy and switching over to the hard
disk was giving me problems when I tried to upgrade my I/O board
and terminal to 19,200 baud.
The Mult/IO board has no baud rate switches at all, and will
initialize at 9600 baud. On a cold boot, the Morrow Designs CP/M
bots from the floppy, loads the monitor (the FFFF: business), and
then expects some sort of monitor command. ("B" is all that's
necessary to tell it to boot off the PROM.) Unfortunately, if
the Mult/IO board is running at a default 9600 baud, and your
terminal is set to 19200, the doggoned monitor can't read that
"B", and you never WILL get booted! Real hassle, that -- and if
you cold boot with everything set at 9600, then you have to use
Morrow Design's BAUD.COM to kick up the CPU board and Mult/IO to
19200, physically kick up the terminal to 19200 (can't send a
command to it, you know -- not if the serial port's putting out
twice as fast as the terminal can read!) ... a real pain!
Finally got it working by setting the console default baud
rate to 19200, setting the terminal switches to 19200, and
setting Switch 6 on the Morrow CPU board to ON (engages a cold
boot from the floppy, skips the monitor, and boots right up with
no terminal input required. The CBIOS initializes the serial
port to 19200, using that default baud rate, before the terminal
ever comes into play.
One little side effect I STILL haven't figured out: every-
thing boots up fine (provided you gave the hard disk a few
seconds to come up to speed before hitting that reset button),
but after the cold boot message and A0> prompt, I get a ^Q!
Donno why - think I'm overflowing the buffer with my cold boot
message. No big thing; hope to fix it soon.
I have not yet tried to cold boot from the hard disk (by
setting switches 1 through 5 to ON ON ON ON ON. I don't really
like the idea since it demands a lot of a barely running hard
disk that MIGHT not be up to speed yet!
[Note: I've now added modem port (serial port 3) initialization
to 1200 baud on cold boot and port initializations. Sorry, no
code here since it's quite Morrow-specific; will add to this
document later.]
COMPANY CREDIT LINE ON COLD BOOT: You can see the little "Toad
Hall" logo come on, plus some patch comments, on the cold boot
message. Easy - just added it into the CBIOS when editing it.
That's on page 76 of Morrow's CBIOS& listing, way toward the end.
Just add in whatever you want (well, keep it reasonable -- don't
want to make this CBIOS TOO big, you know!) Don't forget ACRs
and ALFs (carriage returns and line feeds) where appropriate. I
mentioned above that ^Q problem after a cold boot. I suspect it
comes from an excessive cold boot message, so be warned! [It
didn't! Still don't know the answer. Suspect just some garbage
left around the Multi I/O board that wasn't purged correctly
after initialization that shows when running the terminal at
19200 baud!]
MISCELLANEOUS LOCATIONS: If you MAC TOADBIOS.ASM, you'll get most
of these as a DEBUG message. However, just for your information:
OFFSETC 3700 Offset in CP/M when R'ing
TOADBIOS.HEX on top of your
CPM64.COM in DDT.
CCP D400 Start of CCP in real memory
CCP 0B00 Start of CCP in DDT image
BDOS DC00 Start of BDOS in real memory
BIOS EA00 Start of TOADBIOS in real memory
BIOSLN 16 Length of TOADBIOS in pages.
BIOSHEXLN 0F00 Length of TOADBIOS in hex.
CCPLEN 800 Length of CCP in hex.
USRPTCH1 0E8D Actual location in DDT image
of first User Area Patch.
USRPTCH2 20F0 Beginning of second User Area
patch in NOP area of DDT image
of CCP.
CAD3 First patch in CCP to
jump to second patch in CCP
for A: User Area 0 default.
1DB3 Location in DDT image of
first patch.
PATCH 12F2 DDT image location of larger
patch in CCP for default A:.
PATCH1 EC50 Real memory location in
TOADBIOS of default code.
- - - - - - - -
USER AREA DISPLAY CODE:
DDT IMAGE INSTRS & REAL MEMORY LOCATIONS
0E82 LXI SP,DBAB ;original code
0E85 CALL D498 ; " ", flush
0E88 CALL D5D0 ; " ", get drive
0E8B ADI 41 ; " ", 'A'
- - - - - and here's the patch - - - - -
0E8D CALL E9F0 ;patch - was CALL D48C, console out
- - - - - end of patch - - - - -
MVI A,3E ;original code - '>'
CALL D48C ; " ", console out
CALL D539 ; " ", get cmd
ext patch is in an area of NOPs in the CCP - check to be sure!
(And don't get too low in that area: I think it's disk
parameters or something! Go too low with your patch and it'll
get overlaid and eaten up!) If something's there, just find
about 1 1/2 lines of NOPs somewhere in your CCP where this will
fit (maybe just before the BDOS where there's sometimes some
room). Use DDT's A (for Assembly) instruction to stick in this
stuff just like you see it. If your CCP does NOT start at D400,
then you'll have to figure out an offset. Best place to do this,
of course, is in your CBIOS in assembly -- makes it all much
easier to squeeze.
DDT LOC REAL MEMORY LOCATIONS
20F0 CALL D48C ;console out routine
CALL D513 ;get user routine
ADI 90H ;routine to convert to number
DAA
ACI 40H
DAA
JMP D48C ;console out again, which returns
;us automatically to right after
;the first patch at 0E8D where
;things continue like normal.
[Note: above was per the original patch. My patch NOW reads
like this to produce the decimal user area:
USRPTCH:
CALL 0D48CH ; Console Out routine.
CALL 0D513H ; Get User routine.
CPI 10 ; If User # >9, print leading 1.
JC DUX
PUSH PSW ; Save A w/User #.
MVI A,'1' ; Load the 10 digit for Conout
CALL 0D48CH ; Conout.
POP PSW ; Get A w/User # back.
SUI 10
DUX: ADI '0'
ORA A ; Clear flags.
JMP 0D48CH ; Print second digit.
; End of patch
I include the original code below, just for reference. It has
the original locations, plus my patched Morrow locations.
; --- PATCH TO CP/M 2.X TO LIST USER # IN DRIVE PROMPT ---
; ( VALID FOR CP/M 2.0, 2.1, AND 2.2)
; BY BRUCE KENDALL (TKI)
; 7/12/80
; TIGHTENED UP BY BRUCE RATOFF
; 11/17/80
;
; IF YOU HAVE TRIED PLAYING WITH THE 'USER' COMMAND
; IN CP/M 2.X, YOU MAY HAVE BECOME ANNOYED THAT THERE
; WAS NO WAY OF TELLING WHAT USER AREA YOU WERE IN. THIS
; PATCH SOLVES THIS PROBLEM BY DISPLAYING THE USER NUMBER
; IN HEX ( A SINGLE CHARACTER SINCE USER # : 0-15 ARE VALID)
; BETWEEN THE DRIVE NAME LETTER AND THE '>'. THAT IS, A USER
; LOGGED INTO USER AREA #4 WOULD SEE THE STANDARD CP/M
; PROMPT (MODIFIED BY THIS PATCH) AS:
; A4> ( INSTEAD OF JUST A>)
;
MSIZE EQU 64 ; CP/M SYSTEM SIZE IN KB
;
DELTA EQU 1000H ; OFFSET FROM STD CP/M SIZE
; THIS WOULD BE SET TO 400H IF
; THE 20K CP/M WAS ACTUALLY A 19K
; CP/M (WHEN COMPARED TO THE STD
; 20K CP/M DESCRIBED IN THE CP/M
; MANUALS FROM DIGITAL RESEARCH).
;
BIAS EQU (MSIZE-20)*1024-DELTA ; OFFSET FROM 20K CP/M
CCP EQU 3400H+BIAS
;
OFFSET EQU 980H-CCP ; OFFSET USED WITH DDT IN
; SYSTEM CONFIGURATION (ASSUMES
; THAT 'CCP' OCCURES AT 980H IN THE
; SYSGEN MEMORY IMAGE).
;
COUT EQU CCP+8CH ; CCP CONSOLE OUTPUT ROUTINE
GTUSR EQU CCP+113H ; CCP GET USER # ROUTINE
;
ORG CCP+38DH
;
CALL PATCH ; THIS WAS A CALL COUT
;
; -----------------------------------------------
; NOTE THE CODE IN THE NEIGHBORHOOD OF THIS PATCH WAS
; USED TO PRINT OUT THE 'A>' PROMPT:
;
; CCP+382H:
; LXI SP,----
; CALL FLUSH ; RESET BUFFERS
; CALL GTDRV ; GET DRIVE #
; ADI 'A' ; ADD IN ASCII BIAS
; CALL COUT ; <--- MAKE PATCH HERE
; MVI A,'>' ; GET '>'
; CALL COUT ; PRINT IT OUT
; CALL GTCOMD ; GET CONSOLE COMMAND
; .
; .
; -------------------------------------------------
;
ORG CCP+15F0H ; PATCH AREA AT END OF BDOS
;
PATCH: CALL COUT ; OUTPUT CHAR. IN ACC TO CONSOLE
CALL GTUSR ; GET USER #
ADI 90H ; USE INTEL HEX/ASCII TRICK
DAA
ACI 40H
DAA
JMP COUT ; PRINT OUT AND RETURN
;
; ------------------------------------------------------
; NOTE: THE 'GTUSR' COMMAND IS JUST A SHORT ROUTINE:
;
;GTUSR: MVI E,0FFH
; MVI C,32
; JMP 05
; ------------------------------------------------------
;
END
- - - - - -
DEFAULT TO A: 0 CODE
I gave you CCP+n locations this time so you can offset from your
DDT image of the start of your CCP. You still have my own
values, given my real memory CCP at D400 and BIOS at EA00.
REAL MEMORY LOCATIONS
IN MY SYSTEM
CCP EQU 0D400 D400 ;start of CCP
GTUSR EQU CCP+113H D513 ;Get user number routine
STUSR EQU CCP+115H D515 ;Set user number
CCPMFCB EQU CCP+0D0H D4D0 ;Open file at CPMFCB$
CPMTYPE$ EQU CCP+7D6H DBD6 ;type field in CPMFCB$
CMDSK$ EQU CCP+7F0H DBF0 ;loc of disk given in cmd
CMDERR EQU CCP+76BH DB6B ;loc to type error in cmd
WIN EQU CCP+6DEH DADE ;go here if we get file open
DDT IMAGE INSTRS & REAL MEMORY LOCATIONS
11DB JZ CCP+6DBH DADB ;replaces JZ DB6B in my
;system. Loc of cmd err
12F2 PATCH: DDF2 ;replaces an unused area
;of NOPs - check!!
LXI H,CMDSK$ DBF0 ;get drive from curr cmd
ORA M ;A=0 on entry, so fetches
;drive
;in next line, if explicit drive given go try User 0
;(this will be escape even if we force A: in our cmd line
JNZ PATCH1 EC50 ;Wherever you put that
;new code in your CBIOS.
;I put it here -- you can
;find it by looking up
;PATCH1 in the SYM table
;generated by MAC
INR M ;force explicit reference
;to drive A
LXI D,CPMTYPE$ DBD6 ;need DE set up to this
;on entry to CCP
JMP CCP+6CDH DACD ;Now to reenter CCP
The following code must be patched into a blank area of your
CBIOS or entered into CBIOS&.ASM prior to assembly. Be sure to
call it PATCH1 or something distinctive (and NOT like anything in
CBIOS& or you'll have problems!) so you can find it in the SYM
table.
Arrive here because explicit drive set or can't find file on A:
PATCH1: EC50 ;in my TOADBIOS version
CALL GTUSR D513 ;get user code
ORA A ;set flags
JZ CMDERR DB6B ;already user 0 so we
;fail - .COM file isn't
;anywhere!
MOV E,A ;get old value into E
;for later
PUSH D ;save it
MVI E,0 ;set USER=0
CALL STUSR D515
CALL OCPMFCB D4D0 ;try Open again
POP D ;get old user code back
;before we save flags
PUSH PSW ;now save flags from call
CALL STUSR D515 ;now go set back to old
;user number
POP PSW ;get flags back from
;OPEN call
JNZ WIN DADE ;we found it!
JMP CMDERR DB6B ;nope, nowhere!
- - - - - - - -
That's the code. Hopefully you can make it work. I'm only a
novice Assembler programmer, and managed just fine! Kind of fun,
too! (Kind of nice at parties -- "What, you don't read hex? And
don't work in Assembler either?")
David P Kirschbaum
SGM, US Army
Toad Hall
7573 Jennings Lane
Fayetteville NC 28303