开云体育

Yes, having A FP should be seamless, least it was on my old Altair
when it behaved.? There is no reason why CP/M should not run
with a FP.? Also the blinkin lights thing.

The Xmodem bug,? That may be a feature. ;)? Modem program in
the CP/M space were always troublesome as some wanted a well
configured IO byte, others did direct IO and had to be tweaked for
the IO device in use (SIO-A or B, 8251, 6850, 2651, 16450,
SIO/(drt or SIO2)? and many others often different enough.

DDT the likely clash is RST7 (address 38) as DDT uses that as a
debug return point.? ?Usually front panels had port FFh as an input
from the FP switches (sense, often used by non cpm for setting up
what IO exists).? Also FFH is the opcode for RST7 so missing memory
(altar/imsai) has pull-ups so by default no ram, is a RST7.

FYI the most nasty bug on the Altair was noise on all the front panel
wires to the bus causing the appearance of "flying bits" and crashes.?

DDT loads to 100H then senses the base of BDOS and relocated
to under that so the space from 100h to below DDT is available for?
user programs that would nominally be at loaded at 100H when
finished or debugged.

ajparent1/kb1gmx wrote:

Having a [Front Panel] should be seamless, least it was on my old
Altair when it behaved. There is no reason why CP/M should not run
with a FP. Also the blinkin lights thing.

Yes; the blinkin-lights are an important feature of a FP. "What are all those blinking lights doing, son?" "Impressing you, dad!"

A hardware front panel shouldn't have any effect on the computer. But a software-driven front panel (like the Z80MC or a Heath H8) has to interact. It needs a program in memory to run it, and an interrupt to periodically update the blinkin-lights. The interrupt routine requires time, and that there be a working stack, among other things.

The Z80MC uses RST7 for its timer/serial/front panel interrupt. We just discovered that CP/M's DDT also uses RST7 for breakpoints. Thus they conflict; so DDT will crash the system!

There are a couple ways to fix this:

1. Software: I've been looking at the DDT source, and there is an option to use RST6 instead of RST7. DRI apnote #7 (attached) describes the procedure for patching DDT to use RST6. It says to change address 01EB from FF to F7.

But, Bob Kircher reports that his copy of DDT had the FF at 01EC (not 01EB). So he patched to 01EC from FF to F7; and it worked. :-)

2. Hardware: It is also possible to move the hardware interrupt from RST7 to some other interrupt. But this requires some cuts-n-patches, plus some software changes to the ROM and BIOS to account for it. So I think we'll save that for a future revision.

DDT loads to 100H then senses the base of BDOS and relocated
to under that so the space from 100h to below DDT is available for
user programs that would nominally be at loaded at 100H when
finished or debugged.

Yes; DDT is a *very* clever program! The source has been published ro all to see. Besides relocating itself, it can also allow the program it is debugging to overlay parts of it, and still keep the remaining parts working.

Lee
--
There is a computer disease that anybody who works with computers knows
about. It's very serious, and interferes completely with your work. The
trouble with computers is that you 'play' with them! (Richard Feynman)
--
Lee Hart, 814 8th Ave N, Sartell MN 56377, www.sunrise-ev.com

With 8080 RST7 is free (bus to 1s using pullups) if you want any other you need to assert that on the bus
an octal bus driver with the inputs set to RST3 or 4 is suggested.? RST6 is also in a sorta sometimes
used space as well.?

DDT is a very compact program as well.

heath H8 the front panel was a source of issue for later software.? It was its weakness in the works
of Altair and IMSAI .? Both of them the lights for address were just buffers address the Data leds?
were also same and the processor wait stated between operations.

Most of the "smart" front panels the address leds were again off the address bus
but the data was a latched port? both input and output and the very small usually 256 bytes?
program interrogated switches or loaded the latch. Most all avoided interrupts as they were
difficult (who knows what CPU) and often got in the way.? When running the leds show data
and address off the bus and the FP know if it was running or suspended.

the 8080 FPs often used tricks like 1802 DMA load mode save for it was done via wait and?
jamming nop or jumps for a single cycle.

Allison

ajparent1/kb1gmx wrote:

With 8080 RST7 is free (bus to 1s using pullups)

Yep! That's how I did it on the Altaid 8800 (an 8080 version of Membership Card) <>. But that has to change to make stock versions of DDT (or SID or ZSID) work.

For the Z80MC, DRI application note #07 is the patch to change DDT to use RST6 instead of RST7. We made the patch, and put a copy of DDT with it (named DDT-V6.COM) on the DISK-1 image for the SD-CARD.

Josh has also updated his Windows CP/M Explorer program to version 1.3.

Both have been uploaded to the website <>

DDT is a very compact program as well.

It certain is! There are some very clever tricks to keep it small. Whoever the programmer was (Gary Kildall himself?), he certainly knew how to code for limited resources.

heath H8 the front panel was a source of issue for later software. It
was its weakness in the works of Altair and IMSAI.

It depends on what you want. Front panels are expensive and hard to use; after the initial novelty wears off, they don't get used much.

The Heath H8 keypad and 7-seg LED monitor was a lot easier to use; and cheaper, too. It remained useful on a regular basis. For one thing, it would continuously display the last register, memory address, or I/O port even while your main program ran. But as you say, it depended on an interrupt, and for the CPU to be running the monitor program. It took away a little execution time (but you could turn it off to make your programs run a little faster).

One of the cleverest early Front Panels was George Morrow's "Keyed-Up". It was an 8080 CPU board for the S-100 bus that included a keypad and 7-seg display. It ran software to work; but that software was a tiny 32-byte PROM. Instead of depending on an interrupt, it opened the 8080 data bus, and inserted its PROM. The card had a separate program counter for the PROM; it didn't depend on the CPU's PC. So his little program ran no matter where the CPU was, to let you examine and change memory, etc.

Lee

--
There is a computer disease that anybody who works with computers knows
about. It's very serious, and interferes completely with your work. The
trouble with computers is that you 'play' with them! (Richard Feynman)
--
Lee Hart, 814 8th Ave N, Sartell MN 56377, www.sunrise-ev.com