80-Bus News |
November–December 1984 · Volume 3 · Issue 6 |
Page 44 of 55 |
---|
additional hardware is used, and in any case both CP/M and NAS-SYS use page one of memory, which is where the ‘RST’ locations are operative. Assuming a vector can be generated, there is still the problem of where to allow interrupts to be active within the program. It would be necessary to use EI and DI instructions carefully, avoiding the whole of the CCP and BDOS areas, and most of the CBIOS. The area diminishes until one is left with a few I/O areas such as the Keyboard. Fortunately this conclusion greatly simplifies the task, because computers spend nearly the whole time waiting for keyboard input.
There are occasions when the machine is not awaiting Keyboard input, as during Disk I/O, Formatting text, Compiling or Assembling a Source File, or perhaps copying files. Some of these activities can last for a few minutes. If one is content to allow the clock display to ‘freeze’ for these few times, but update regularly at all other times, then the solution to the problem becomes much simpler. The difficulty of implementing interrupts has been mentioned. Since clock updating is restricted to such a ‘small’ area of the program, in terms of code, then there is no real advantage in using Interrupts, and Polling will give the required performance.
The method that I have used is therefore based on integrating the clock ‘polling’ code into the keyboard scan code.
A number of short listings of source code are given. These show in full the extra code that I have added to SYS vers. 16. to permit conditional assembly for the Clock with options for a couple of extra features. Any code shown in lower case is ‘SYS’ code, included to show the position of my additions, while Upper case code is my added code. (Except for the ‘headings’ enclosed in “*” characters.) Modification for other SYS’s or the Gemini BIOSs should also be relatively easy, if the Source Code is to hand. In the example listings in this article, it is assumed that the system is using a keyboard attached to the Gemini GM812 IVC card, and that the RTC chip is addressed as 16 Ports from address 20H, as on the Gemini GM816 card. Modification of the clock software for PIO format should be along the lines of previous articles in the 80-BUS News, and other magazines. I have not looked at the problems that might arise in outputting the Clock to a Nascom format screen. If a Nascom keyboard is used the clock code can probably be fitted into that area, but I have not tried this method.
Code extract 1 lists additional ‘EQU’ates that need to be inserted into module one of SYS. If CLOCK is not true, then all clock related code will be omitted in subsequent modules. If Clock is TRUE, then all related clock routines will be included in the assembly. Setting LOCK to TRUE will cause code to be included into the WARM BOOT routines that will ‘lock’ the top line of the IVC screen display. This will stop the clock display from scrolling ‘through the top’ of the screen. I usually have both Nascom and Gemini keyboards on my machine, but the RTC code in this article is relevant to the keyboard on the IVC only.
Some of the clock code is a straight copy of the examples given in the GM816 manual and I make no claim as to its originality. The variable NMRREG is set to the number of RTC registers that are to be read. The variable OFFSET can be chosen such that the Clock display on the Screen is updated every 1 Sec., 10 Sec,. or 1 Minute as desired. The RTC chip is of course updating its own registers every .1 Second. There is only about 10 Bytes left in SYS at
Page 44 of 55 |
---|