Wow, great sleuthing. Which program are you referring to that operates at non-standard rates?
In trying to decipher the old listings, I've repeatedly wondered if people had a different ROM disassembly available to them in the past. Of course, it's totally possible that folks were independently figuring things like this out and they simply didn't share the knowledge. By the way, I'm starting a side-project to gather all the known ROM/RAM/IO addresses for the Tandy 200 into a single machine and human readable file. The idea is to have a format that can be used for both disassembling existing code (for example, as "labels" in Ghidra) and as an equiv "library" for writing new assembly language programs. --b9 On May 4, 2026 1:20:19 PM PDT, Peter Noeth <[email protected]> wrote: >Group, > > I have been doing some work on a BASIC program I downloaded from the old >club100 library that uses the RS232 port at a non-standard baud rate. > > The BASIC program also POKES the value 31 into memory location 65421 >(FF8Dh). Since there is no documentation in the code or .DO file, I have >been researching this location. According to Ken Pettit's M100 ROM >disassembly (dated 2013), this memory location is only used twice, once in >the "UART config from text string" routine (17E6h) at 1832h and again at >the "RS232 Receive Interrupt" routine (6DACh) at location 6DB9h. > > The memory location FF8Dh is noted as "RS232 Control Byte" in both >locations in the disassembly, but in reality, is a "read data mask". >Microsoft included this ability, but who knows why. Without the original >commented ROM listing we will never know. > > In the "UART Config From Text String" routine, the value 255 (FFh) is >always written to memory at FF8Dh, which couldn't be a shadow of the RS232 >Control Byte (baud rate, parity, etc.). And in the "RS232 Receive >Interrupt" routine, after data is read from the UART Receive Data Register >(C8h), it is ANDED with the byte at FF8Dh before being stored in the C >Register. Definitely a "mask" operation. > > The BASIC program is putting a value of 31 (00011111b) in FF8Dh, masking >whatever data is received to a value between 0 and 31. This would make >sense to the BASIC program which is configuring the RS232 port for 5-bit >operation. But I would think it is not necessary. Unfortunately, I can't >contact the originator of the BASIC program, to find out why he did this >and how he found this memory location in RAM as I have found no Model 100 >documentation that references it (I own most all of the books written about >the Model 100). > > So, just a "word to the wise" about the function of memory location 65421 >(FF8Dh). Maybe Ken and others who have worked on commenting the disassembly >of the Model 100 ROM, will update their listings. > >Regards, >PeterN
