Kaypro 10 keyboard. [Vas Crabb, rfka01, TeamEurope]
(nw) This exposes other issues in MAME. Some incredibly poorly-written
code in the keyboard MCU depends on the fact that the T register is
updated during S4 then copied to A during S4 when mov a,t is executed.
If it doesn't see zero in this register when mov a,t is executed
immediately before the timer interrupt is taken, it hangs.
Really, the MCS-48 core needs to be re-written so it works at S-cycle
level (if not clock cycle leve), but for now I've worked around this
case in the least intrusive way possible.
This also exposes issues in the Z80DART code. The keyboard communicates
0.8% slower than the nominal 300 Baud. This works perfectly in real
life, but it causes occasional corrupted characters with MAME's
implementation, particularly if you hold a key down and let it repeat.
The Z80DART device needs to be fixed so it samples closer to the middle
of the bit intervals and re-aquires on each start bit.
Finally, I haven't hooked up the Kaypro's serial port Baud rate
genrators, or exposed the serial ports as slots. I'll leave that for
someone else to worry about.
* t4426: Replaced a bad dump caused by corrosion with a new verified dump. The firmware is no more bailing out to BASIC prompt
* t4426: Removed reference to old BAD_DUMP
Many changes in this PR, including the following:
* added support for InterPro 20x0, 24x0 and 27x0 systems
* added support for SR bus/slot devices
* added support for Ethernet, graphics, keyboard and mouse
* added support for serial DMA
* improved timer, interrupt and DMA handling
* improved comments/documentation
Note: ioga is still a major work in progress, and most of the other *ga's don't do anything useful yet, however this build can do some new and useful things, and it's closer to OS boot than before, as well as cleaned up in some of the better known areas.
New Not Working Clone Added
-----------------------------------------------
Gran Trak 10/Trak 10/Formula K (older) [TTL] [Ed Fries, Tim Giddens, Andy Welburn]
