1
0
mirror of https://github.com/hsoft/collapseos.git synced 2024-11-27 09:38:06 +11:00

Add user guide

This commit is contained in:
Virgil Dupras 2019-04-15 12:06:27 -04:00
parent 26b125b337
commit 9aae4ea040
3 changed files with 108 additions and 2 deletions

View File

@ -104,13 +104,17 @@ There's very little done so far, but here's how it's organized:
* `parts`: Pieces of code to be assembled by the user into an OS.
* `recipes`: collection of recipes that assemble parts together on a specific
machine.
* `doc`: User guide for when you've successfully installed Collapse OS.
Each folder has a README with more details.
## Roadmap
I'm still fiddling with things, honing my skills and knowledge, so the
project's roadmap is still hazy.
The project is progressing well and I already have a working shell (see `doc`
to see what it can do) on a classic RC2014.
However, such a vast project involves quite a lot of fiddling and I can't really
have a precise roadmap, only a general direction:
The primary target for Collapse OS is the z80 architecture. There's a good
amount of great z80-related hacks all around the internet, and the z80 CPU is

10
doc/README.md Normal file
View File

@ -0,0 +1,10 @@
# User Guide
This collection of document is intended to be a user guide, not assembly
instructions. It is therefore assumed that you have a machine with Collapse OS
properly running.
## Table of Contents
* [The shell](shell.md)

92
doc/shell.md Normal file
View File

@ -0,0 +1,92 @@
# shell
The shell is a text interface giving you access to commands to control your
machine. It is not built to be user friendly, but to minimize binary space and
maximize code simplicity.
We expect the user of this shell to work with a copy of the user guide within
reach.
It is its design goal, however, to give you the levers you need to control your
machine fully.
## Commands and arguments
You invoke a command by typing its name, followed by a list of arguments. All
numerical arguments have to be typed in hexadecimal form, without prefix or
suffix. Lowercase is fine. Single digit is fine for byte (not word) arguments
smaller than `0x10`. Example calls:
seek 01ff
peek 4
load 1f
call 00 0123
All numbers printed by the shell are in hexadecimals form.
Whenever a command is malformed, the shell will print `ERR` with a code. This
table describes those codes:
| Code | Description |
|------|---------------------------|
| `01` | Unknown command |
| `02` | Badly formatted arguments |
## seek
The shell has a global memory pointer (let's call it `memptr`) that is used by
other commands. This pointer is 2 bytes long and starts at `0x0000`. To move
it, you use the seek command with the new pointer position. The command
prints out the new `memptr` (just to confirm that it has run). Example:
> seek 42ff
42FF
## peek
Read memory targeted by `memptr` and prints its contents in hexadecimal form.
This command takes one byte argument (optional, default to 1), the number of
bytes we want to read. Example:
> seek 0040
0040
> peek 2
ED56
## load
Puts the serial console in input mode and waits for a specific number of
characters to be typed (that number being specified by a byte argument). These
characters will be literally placed in memory, one after the other, starting at
`memptr`.
This command is, for now, of limited use because it's tied to the active
console, but a method for selecting I/O sources is planned and this command will
become much more useful.
Example:
> load 5
Hello
> peek 5
48656C6C6F
## call
Calls the routine at `memptr`, setting the `A` and `HL` registers to the value
specified by its optional arguments (default to 0).
Be aware that this results in a call, not a jump, so your routine needs to
return if you don't want to break your system.
The following example works in the case where you've made yourself a jump table
in your glue code a `jp printstr` at `0x0004`:
> seek a000
A000
> load 6
Hello\0 (you can send a null char through a terminal with CTRL+@)
> seek 0004
0004
> call 00 a000
Hello>