This repository has been archived by the owner on Jun 19, 2021. It is now read-only.
forked from gyurco/Minimig-AGA
-
Notifications
You must be signed in to change notification settings - Fork 2
/
TODO
49 lines (40 loc) · 3.01 KB
/
TODO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
FIX FIRST
- HELP key / HRTMON (NMI) button fix - DONE (OSD HRTmon disabled - F11 is HELP key)
- 512kB ROM mirror (just upload it twice) - DONE
- video dithering needs fixed - to be checked with C model - DONE
- fix reset problems, fastRAM problems - DONE
- try to fix clock/sdram issues - ONGOING
- add support for AROS - ONGOING (IDE problems)
- audio sigma-delta needs fixes - to be checked with C model
- implement audio volume control (full, half, quarter?)
- mapping CTRL, ALT, ... keys in mist.ini
- IDE OSD settings sometimes don't get applied
- A8 kickstart ROM mirror - WONTFIX (memory block is partly occupied by HRTmon, besides it is very rarely used)
IMPORTANT - AGA CHIPSET stuff
- increment hpos counter in denise by two LSB bits (35ns counter)
- hdiwstrt & hdiwstop (HDIWHIGH) also have two new high-resolution bits
- actually, all (horizontal) comparators in AGA chipset have 35ns resolution
- http://eab.abime.net/showpost.php?p=669147&postcount=77
- bitplane shifter problems
* more 'undocumented' stuff ... (looks like the output shifter is actually 256 bit to allow 64-pixel fetches in lowres to be shifted on shres resolution - implemented)
* upper is probably wrong, shifter is 64bits plus additional shres shifter (two bits?)
* even that bitplane shifter is probably wrong: http://eab.abime.net/showthread.php?t=71437
* check if shifter problems are the cause of minimig display shifted to the right too much in some cases (how is shres scrolling supposed to work - what to do if those bits are set to 0??)
- bitplane DMA register delays (TLC - PowerTrax demo, hint by Toni Wilen):
* Writing to BPLxPT when exactly next cycle has DMA to matching BPLxDAT: write goes nowhere. (Same happens with sprites and blitter registers too, I assume all DMA channels have similar internal 2 cycle pipeline)
* Another similar special feature is bitplane modulos, writing to BPLxMOD when exactly next cycle is matching BPLxDAT write and it also does modulo add, _old_ modulo value is used! (Write to BPLxMOD is still accepted, next time BPLxMOD value is needed it is used normally)
- sprites can be positioned on 35ns pixels - to add! - I think I did this - CHECK
- add CD32 gamepad emulation and / or real amiga cd32 gamepad support (http://gerdkautzmann.de/cd32gamepad/cd32gamepad.html) - ONGOING
- reimplement blitter - CONSIDERING
IMPORTANT - CPU stuff
- CPU chipRAM turbo - DONE, need to enable cache for chipRAM - CHECK
- CPU kickstart turbo - DONE, some problems remain - ONGOING
- CPU compatibility fixes are a must! - ONGOING (Till fixed most of the instructions, still need to check stack frames - see (http://lxr.free-electrons.com/source/include/asm-m68k/traps.h?v=2.2.26 ... http://ftp.dei.uc.pt/pub/linux/kernel/people/geert/uClinux-amiga-2.6.x-merging/008-frame.diff)
- CPU fix cart compatibility problems (overlay detection should be moved to TG68K.vhd); big problems with turbo kickstart - ONGOING
MEDIUM - mostly firmware stuff
- joystick autofire menu ?
LOW - waaay down the line
- ethernet - ONGOING
- turbo blitter
- RTG
- AHI audio