1
0
mirror of https://github.com/TomHarte/CLK.git synced 2024-12-11 00:52:13 +00:00
Commit Graph

381 Commits

Author SHA1 Message Date
Thomas Harte
c5c722ae56 Generalise axis steps; begin HMMV. 2023-01-31 13:35:39 -05:00
Thomas Harte
872b9e5021 Predict Yamaha horizontal retrace interrupts. 2023-01-30 21:33:47 -05:00
Thomas Harte
492a170b20 Implement much of Yamaha line interrupts. 2023-01-30 21:24:53 -05:00
Thomas Harte
30a2b1611f Merge branch 'MSX2' of github.com:TomHarte/CLK into MSX2 2023-01-30 21:06:23 -05:00
Thomas Harte
29af5542f8 Make an effort at doing _something_ for G4. 2023-01-30 21:06:04 -05:00
Thomas Harte
bc4c54800e Type out just a little of status register 1. 2023-01-30 20:20:33 -05:00
Thomas Harte
8f20cb93e9 Note missed status accesses. 2023-01-29 21:20:50 -05:00
Thomas Harte
73e79b14ea Use Yamaha palette pervasively. 2023-01-29 21:17:00 -05:00
Thomas Harte
3142f5c21d Be overt about what replaces LineMode. 2023-01-29 21:04:15 -05:00
Thomas Harte
6d7f189ce7 Attempt the full panoply of logical pixel modes, across all graphics modes. 2023-01-29 18:28:49 -05:00
Thomas Harte
a91a5b8d07 Refer to actual field. 2023-01-29 18:02:09 -05:00
Thomas Harte
4cdcd3ac7d Retain logical operation, take colour combination outside the loop. 2023-01-29 13:29:19 -05:00
Thomas Harte
0576451102 Be overt about colour direction. 2023-01-29 13:22:56 -05:00
Thomas Harte
3f12a28f4f Locate first pixel correctly. 2023-01-28 22:50:04 -05:00
Thomas Harte
41ba883fb6 Honour direction, start transfer immediately. 2023-01-28 22:47:27 -05:00
Thomas Harte
1e646eb57b Improve transfer flag for LMMC. 2023-01-28 21:45:05 -05:00
Thomas Harte
2d6afe1013 Reduce repetition, tidy slightly. 2023-01-28 21:43:14 -05:00
Thomas Harte
d3c446d91b Take a shot at LMMC. 2023-01-28 21:30:45 -05:00
Thomas Harte
975ead5d01 Edge towards not assuming graphics mode. Much more to do here. 2023-01-28 13:54:26 -05:00
Thomas Harte
c6cda7c401 Switch to absolute placement of deferred events; properly serialise commands. 2023-01-28 11:55:12 -05:00
Thomas Harte
7e69e33ec2 Use paletted Yamaha border colour. 2023-01-27 22:42:02 -05:00
Thomas Harte
95e00dd958 Take slightly wrong-headed steps towards proper command timing. 2023-01-27 22:20:36 -05:00
Thomas Harte
8a673a697b Further adapt internal terminology. 2023-01-27 12:30:09 -05:00
Thomas Harte
75ad4cdb67 Fix line semantics. 2023-01-27 11:57:40 -05:00
Thomas Harte
d42b6df570 Retain extra pattern name address bits. 2023-01-27 06:44:11 -05:00
Thomas Harte
81b00c6d97 Add notes. 2023-01-26 22:25:10 -05:00
Thomas Harte
1e5f751bc0 Require STOP in order to stop. 2023-01-26 22:08:36 -05:00
Thomas Harte
9a65fffe16 That's PSET, not POINT. 2023-01-26 22:02:40 -05:00
Thomas Harte
515fa22bfe Implement point. 2023-01-26 21:52:41 -05:00
Thomas Harte
66ac089cc2 Deallocate. 2023-01-26 21:49:32 -05:00
Thomas Harte
5d5098acb2 Hack in vertical scrolling. 2023-01-26 21:38:51 -05:00
Thomas Harte
1bf8406e7e Correct deserialisation order. 2023-01-26 21:34:56 -05:00
Thomas Harte
75acbd2d6c A quick hack shows some part of the MSX logo. 2023-01-26 21:31:49 -05:00
Thomas Harte
baa6f9b3cd Implements the Command side of the line command. 2023-01-26 21:17:11 -05:00
Thomas Harte
1c6a0ad3f7 Clean up repetition. 2023-01-26 19:51:56 -05:00
Thomas Harte
fbfa26ad5e Minor steps towards implementing Line. 2023-01-26 12:55:08 -05:00
Thomas Harte
b12fd00145 Generate an appropriate instance for line drawing. 2023-01-26 12:09:06 -05:00
Thomas Harte
0c8815d6a0 Retain command-engine context. 2023-01-26 11:59:27 -05:00
Thomas Harte
700470915a Add pixel serialisation for Yamaha graphics mode 5. 2023-01-24 23:07:29 -05:00
Thomas Harte
f8b42d4107 While being lazy with types, implement 4/5/6/7 fetching. 2023-01-24 13:15:00 -05:00
Thomas Harte
63bd0f918d Be overt about buffer target and vertical position. 2023-01-24 12:46:09 -05:00
Thomas Harte
445b34933a Edge further towards actual fetching. 2023-01-23 23:19:04 -05:00
Thomas Harte
bbd8b011ba Remove temporarily-faulty check. 2023-01-23 22:52:03 -05:00
Thomas Harte
15fbf4cb7f Ensure Yamaha-style refresh is used in all modes. 2023-01-23 22:39:34 -05:00
Thomas Harte
7293f9dc10 Remove extraneous brackets, add comment to self. 2023-01-23 22:25:56 -05:00
Thomas Harte
8567c934b1 Ensure Yamaha refresh program is used. 2023-01-22 22:11:01 -05:00
Thomas Harte
2744a9b6b0 Tidy up. 2023-01-22 22:02:39 -05:00
Thomas Harte
91047e5b3a Start attempting to use table-based Yamaha fetch. 2023-01-22 22:00:28 -05:00
Thomas Harte
c6dd7d4726 Transcribe no-sprite event list. 2023-01-22 20:19:21 -05:00
Thomas Harte
b7d80f5ed1 Copy in some notes, expand line buffer. 2023-01-21 23:04:48 -05:00
Thomas Harte
a5765abbad Route into the Yamaha fetcher.
Albeit that it doesn't yet fetch.
2023-01-21 22:47:16 -05:00
Thomas Harte
696ec12516 Add address rotation for applicable modes. 2023-01-21 22:33:26 -05:00
Thomas Harte
c9734df65c Implement extended colour, sprite and RAM pointers. 2023-01-21 20:45:23 -05:00
Thomas Harte
13e490e7d7 Log selected screen mode. 2023-01-21 14:48:55 -05:00
Thomas Harte
cefcc1d443 Expand Yamaha graphics mode recognition. 2023-01-21 14:35:26 -05:00
Thomas Harte
d1f929e6f7 Just do a multiply and divide. Easy. 2023-01-21 14:19:52 -05:00
Thomas Harte
e289e6e757 Catch and map Yamaha palette entries.
It's one less thing in the uncaptured log.
2023-01-21 14:12:46 -05:00
Thomas Harte
a726c9d97a Enable indirect register writes. 2023-01-20 23:14:57 -05:00
Thomas Harte
c77e7c268f 1 = disable, 0 = enable. 2023-01-20 23:08:41 -05:00
Thomas Harte
9c57bfd58d Attempt to log dropped indirect writes. 2023-01-20 23:07:14 -05:00
Thomas Harte
4efda108c6 Transcribe the Yamaha 9938 register meanings. 2023-01-20 23:00:33 -05:00
Thomas Harte
191cf4829b Attempt real blank reporting. 2023-01-20 22:29:49 -05:00
Thomas Harte
9b7a925816 Give clearer names to the two pointers. 2023-01-20 20:29:15 -05:00
Thomas Harte
392b0acb58 Pull everything out of master_system_ struct.
Now that it's inherently collected in the relevant `Storage`.
2023-01-19 15:09:16 -05:00
Thomas Harte
4b7606894e Move Master System state, and start simplifying. 2023-01-19 14:09:31 -05:00
Thomas Harte
1fb94d15ab No need for this-> ugliness in Base methods. 2023-01-19 12:32:42 -05:00
Thomas Harte
348c42bdea Start trying to bluff my way through extended status. 2023-01-18 22:23:19 -05:00
Thomas Harte
e450e53c4e Temporarily copy and paste my way to further logging. 2023-01-18 14:59:30 -05:00
Thomas Harte
355ee7fbc7 Adjust factoring of read and write per expanded V9938 scope. 2023-01-18 12:36:57 -05:00
Thomas Harte
4d96122884 Eliminate hard-coded assumption of 16kb.
Clearly I'll have to do something else to support 128k+, probably move the ram pointer?
2023-01-10 12:38:19 -05:00
Thomas Harte
f1f16d1f9a Clarify and simplify half_cycles_before_internal_cycles. 2023-01-09 22:55:46 -05:00
Thomas Harte
fd14829992 Avoid hand-writing all the various conversions. 2023-01-09 22:34:56 -05:00
Thomas Harte
c0fe88a5bb Apply clock conversion to existing usages of do_external_slot. 2023-01-09 13:54:49 -05:00
Thomas Harte
4d9d684618 Add TODO on dangling hard-coded conversion. 2023-01-08 21:44:25 -05:00
Thomas Harte
a0a835cf10 Export memory size into traits. 2023-01-08 21:37:20 -05:00
Thomas Harte
ef67205ce8 Set pixel count per mode. 2023-01-08 21:31:00 -05:00
Thomas Harte
794adf470b Break assumption that cycles = pixels; fix pixel clocking. 2023-01-08 21:25:22 -05:00
Thomas Harte
8cc20844a9 Clock convert for draw_ calls. 2023-01-08 17:31:08 -05:00
Thomas Harte
b522d65c50 Fix border lengths. 2023-01-08 17:04:19 -05:00
Thomas Harte
cb19c2ffb0 Honour internal-clocked timing constants. 2023-01-08 14:10:06 -05:00
Thomas Harte
5f6ddf8557 Avoid expressing the same thing at different clock rates. 2023-01-08 13:58:12 -05:00
Thomas Harte
72e0bfecc1 Edge towards clock-independent line composition. 2023-01-07 14:57:32 -05:00
Thomas Harte
cdf547ac82 Decline to provide synthetic text mode timing on the Mega Drive. 2023-01-07 14:37:06 -05:00
Thomas Harte
dd5b4b484a Avoid double responsibility for state. 2023-01-07 14:34:33 -05:00
Thomas Harte
56831e02fc Expand fixed timing constants. 2023-01-07 13:10:51 -05:00
Thomas Harte
5d2d3944ef Make VRAM access delay a timing property. 2023-01-07 12:48:43 -05:00
Thomas Harte
f9e21df701 Avoid further hard-coded 342s. 2023-01-07 09:13:34 -05:00
Thomas Harte
bb436204f6 Merge branch 'VDPs' of github.com:TomHarte/CLK into VDPs 2023-01-07 09:10:50 -05:00
Thomas Harte
de45536b5c Elucidate a magic constant, add an extra constexpr. 2023-01-07 09:10:41 -05:00
Thomas Harte
ebc1264c2c Create a common home for timing information. 2023-01-06 22:39:46 -05:00
Thomas Harte
7a82b76911 Ensure visibility of memset. 2023-01-05 13:21:03 -05:00
Thomas Harte
27d37f71ec Generalise and better factor bit reversal and TMS drawing. 2023-01-05 13:18:10 -05:00
Thomas Harte
c4a5a9763e Minor indentation improvement. 2023-01-02 15:04:50 -05:00
Thomas Harte
a9f97ac871 Fix nothing-to-do test. 2023-01-02 15:04:08 -05:00
Thomas Harte
475440dc70 Update ClockConverter for potential alternative clocks. 2023-01-02 14:59:36 -05:00
Thomas Harte
dc3f8f5e42 These are the three fetchers to implement.
They'll look fairly different from the TMS and SMS fetchers, I think, owing to the greater irregularity that comes with the smarter RAM accesses. I might need to play around for a while.
2023-01-01 22:44:06 -05:00
Thomas Harte
459ef39b08 constexpr the TMS palette. 2023-01-01 22:34:07 -05:00
Thomas Harte
27812fd0e2 Separate fetchers into their own header. 2023-01-01 22:26:50 -05:00
Thomas Harte
38eb4d36de Better explain cumulative nature of @c to_internal. 2023-01-01 22:18:39 -05:00
Thomas Harte
2bd20a0cf8 Add further exposition. 2023-01-01 22:17:21 -05:00
Thomas Harte
da61909ec5 Explain the purpose here. 2023-01-01 21:20:30 -05:00
Thomas Harte
5729ece7bb Incompletely transitions towards more flexible clock ratios. 2023-01-01 14:20:45 -05:00
Thomas Harte
151f60958e Relocate the 9918 implementation file. 2023-01-01 14:01:19 -05:00
Thomas Harte
180045ada6 Convert vram_access_delay into a free-standing function. 2023-01-01 13:51:52 -05:00
Thomas Harte
11542e7a7f Improve const correctness, simplify inheritance. 2023-01-01 13:49:11 -05:00
Thomas Harte
ffb0b2ce0b Eliminate runtime duplication of personality. 2022-12-31 21:50:57 -05:00
Thomas Harte
b7c315058f Also template Base. 2022-12-31 21:47:05 -05:00
Thomas Harte
d79aac3081 Shuffle the personality enum into the 'public' header. 2022-12-31 15:01:11 -05:00
Thomas Harte
711f7b2d75 C++17 makes this a single step. 2022-12-27 22:50:12 -05:00
Thomas Harte
dca8c51384 Prefer to avoid a macro. 2022-12-27 22:36:27 -05:00
Thomas Harte
2ab4b351ca Extend enum. 2022-12-27 22:20:47 -05:00
Thomas Harte
650b9a139b Tweak Master System blue scale. 2021-03-19 08:38:21 -04:00
Thomas Harte
3cb1072c29 Adds an explicit [[fallthrough]] tag. 2020-06-19 23:10:25 -04:00
Thomas Harte
267006782f Starts to add Qt target; resolves many build warnings. 2020-05-30 00:37:06 -04:00
Thomas Harte
66c2eb0414 Further tightens const and constexpr usage. 2020-05-12 22:22:21 -04:00
Thomas Harte
25996ce180 Further doubles down on construction syntax for type conversions. 2020-05-09 23:00:39 -04:00
Thomas Harte
929475d31e Minor correction: round down, not up. 2019-09-28 23:49:32 -04:00
Thomas Harte
ee89be6730 Removes many stray spaces. 2018-11-23 22:32:32 -05:00
Thomas Harte
9dff13cbbf Re-establishes output from the machines with 9918s and derivatives. 2018-11-14 22:25:19 -05:00
Thomas Harte
f65d80b7d1 Ensures offset and flags are initialised to 0.
This prevents a potential crash at startup.
2018-10-29 22:09:32 -04:00
Thomas Harte
e02aa885d8 Testing against the ColecoVision suggests this is probably always 7. 2018-10-26 20:59:12 -04:00
Thomas Harte
bb09762029 Introduces extra delays to VRAM access. 2018-10-26 20:19:08 -04:00
Thomas Harte
05a5c7120e Shunts CRAM dots into their proper place. 2018-10-26 20:06:51 -04:00
Thomas Harte
521d603902 Adds a first attempt at CRAM dot output. With a TODO. 2018-10-26 19:26:46 -04:00
Thomas Harte
916710353a Makes it explicit that I want the reference. 2018-10-25 23:18:34 -04:00
Thomas Harte
53b00dea3f Adds missing include. 2018-10-25 23:12:41 -04:00
Thomas Harte
0587b9f257 Edges to within millimetres of CRAM dots.
... but all the way up to bedtime.
2018-10-25 23:12:03 -04:00
Thomas Harte
2cdeaa2575 Moves misplaced bracket. 2018-10-23 22:37:19 -04:00
Thomas Harte
286783e880 Accepts GCC's suggestion of extra clarity brackets. 2018-10-23 22:36:23 -04:00
Thomas Harte
00e7958a97 Separates request for an SMS2 VDP from current graphics mode.
Thereby fixes various minor segments of Codemasters games.
2018-10-23 22:19:45 -04:00
Thomas Harte
2f995eb622 Adjusts vertical timing for display height. 2018-10-23 21:20:44 -04:00
Thomas Harte
90fbad0f1c Implements SMS2-style addressing if in a 224 or 240-line mode.
This isn't quite accurate, but it'll do for development.
2018-10-23 20:30:08 -04:00
Thomas Harte
2cbd28478d Allows the sprite terminator to be specified. 2018-10-23 20:01:47 -04:00
Thomas Harte
883680731a Uses explicit state to determine whether a pixel target has been requested. 2018-10-21 21:18:41 -04:00
Thomas Harte
0822c96ce0 Implements the proper row counter values for > 192 row modes. 2018-10-19 22:37:56 -04:00
Thomas Harte
4cd65eab5c Seeks to avoid bad macro expansion. 2018-10-18 22:36:25 -04:00
Thomas Harte
da00c832f5 Corrects colour fetching for multicolour text mode. 2018-10-18 20:38:00 -04:00
Thomas Harte
8ff265c3a1 Corrects multicolour text mode. 2018-10-18 20:25:42 -04:00
Thomas Harte
1fc88c4eff Corrects off-by-one error in line fetching coroutines. 2018-10-16 21:36:31 -04:00
Thomas Harte
5dfe7d8596 Corrects most of TMS sprite drawing. 2018-10-16 20:49:04 -04:00
Thomas Harte
231009b901 Makes faulty attempt to reintroduce TMS-mode sprites. 2018-10-16 20:00:06 -04:00
Thomas Harte
1c5f939aea Reintroduces tiles and some element of sprites in regular TMS mode. 2018-10-14 21:52:13 -04:00
Thomas Harte
c1e6406fc9 Corrects sprite accumulation. 2018-10-14 19:56:09 -04:00
Thomas Harte
d66979c68f Switched to a very large number of buffers, and resolved stupid attempt to reassign a reference. 2018-10-14 18:19:11 -04:00
Thomas Harte
6c09abc6cb Makes a flawed attempt to reformulate this exactly as two separate processes on a common clock with an interchange buffer.
Specifically because closer inspection of the TMS modes shows it isn't quite valid to model output of one line as having fully completed prior to fetching of the next. So some sort of extra buffer is required. At which point it is most natural to continue with the logic that each fetch routine is oriented around the fetching process for a single line, and each output routine has the same view, suggesting separate read/write addresses.

Something is wrong though, as video data is being output too rapidly (I think) and with occasional sync issues (again: subject to investigation).
2018-10-14 16:23:45 -04:00
Thomas Harte
9e52ead09a Ensures sprite scanning doesn't improperly set collision flag; that slot 151 is filled. 2018-10-12 19:50:48 -04:00
Thomas Harte
f6af6778ab Moves scrolling latch to proper position and implements 4-window fetching offset. 2018-10-11 22:36:27 -04:00
Thomas Harte
52e02db5c8 Introduces horizontal counter latching and reading.
Then makes a new guess at frame IRQ position. But gets it wrong. Hmmm.
2018-10-11 19:56:32 -04:00
Thomas Harte
2e379b0834 Adds latching of scroll values. 2018-10-10 21:28:18 -04:00
Thomas Harte
f00f6c8c23 Allows the frame interrupt to be placed anywhere in the frame. 2018-10-10 21:07:39 -04:00
Thomas Harte
50e23f4a2e Fixes 16px-high sprites. 2018-10-10 20:34:00 -04:00
Thomas Harte
e410302237 Switches to real SMS line output composition.
Including setting the sprite collision bit.
2018-10-08 22:43:10 -04:00
Thomas Harte
bca2161a05 Fixes TMS text mode for the new addressing order. 2018-10-07 21:09:01 -04:00
Thomas Harte
6975ed22c0 Doubles down on address-storage format, and implements the vertical scrolling lock. 2018-10-07 18:55:35 -04:00
Thomas Harte
3bead07043 Introduces proper indirection for sprite patterns.
This seems to work, so the onus is now back on the rendering loop.
2018-10-07 17:15:42 -04:00
Thomas Harte
ee20e42372 Makes initial attempt at collecting sprite contents.
With test plotting, indicating some sort of issue.
2018-10-07 16:53:25 -04:00
Thomas Harte
df411b4ede Corrects storage of visible sprites. 2018-10-07 16:40:32 -04:00
Thomas Harte
bfb9d8ccb6 At least attempts to use proper addressing for sprite info fetches. 2018-10-07 14:32:20 -04:00
Thomas Harte
338aec2930 Groups background fetches and experimentally seeks to daub sprites as white. 2018-10-06 22:07:04 -04:00
Thomas Harte
e6510dc87b Attempts to get at least as far as picking visible sprite indices. 2018-10-06 19:27:19 -04:00
Thomas Harte
7830cda912 Implements line querying and most of line interrupts. 2018-10-04 22:50:35 -04:00
Thomas Harte
858721a7a5 Added left border hiding. 2018-10-04 18:52:23 -04:00
Thomas Harte
de4e5c40aa Implements horizontal scrolling lock. 2018-10-03 23:28:33 -04:00
Thomas Harte
05248ab990 Starts to reimplement Master System output. 2018-10-03 23:13:21 -04:00
Thomas Harte
499fc62187 Sets things up for implementation of the inner mode-specific logic. 2018-10-02 21:58:09 -04:00
Thomas Harte
5361120353 Restores a stable frame. 2018-10-02 21:05:30 -04:00
Thomas Harte
60bab8fdf1 Starts to reformulate TMS collection as coroutines.
For the time being, thereby breaks all video. A static screen of the border colour is all you'll see.
2018-10-01 23:03:17 -04:00
Thomas Harte
91aa8f9295 Amps up colour content a little. 2018-09-30 20:47:26 -04:00
Thomas Harte
23191efc05 Starts writing and referring to colour RAM for colours. 2018-09-29 19:50:13 -04:00
Thomas Harte
0d8af010b6 Takes a stab at tile reversal and vertical scrolling. 2018-09-28 22:37:10 -04:00
Thomas Harte
7b9bb772ca Corrected to give a not-exactly-indexed-correctly approximation of what's on display. 2018-09-28 21:03:51 -04:00
Thomas Harte
f7e211c245 Makes first attempt to put something vaguely like the Master System tile map on screen. 2018-09-28 20:39:14 -04:00
Thomas Harte
35c2e74af8 Attempts to establish a coroutine-ish structure for access patterns.
The Master System mode, inevitably, is the test case.
2018-09-27 22:33:41 -04:00
Thomas Harte
19482a563f Attempts to explicitly make room for the SMS VDP mode. 2018-09-27 21:22:57 -04:00
Thomas Harte
9683c8f664 Advances towards the Master System actually receiving interrupts. 2018-09-23 15:58:23 -04:00
Thomas Harte
e7f4babf41 Starts taking steps towards SMS/GG and V9938/9958 support.
Specifically: routine namespace stuff, plus the intention to move to a table-based operation+cost version of timing. Reordering works fine for the TMS, and probably would also for the SMS/GG, but it'd be problematic with the command engine of the V9938/9958 and maintaining a consistent set of code is easier.
2018-09-17 22:59:16 -04:00
Thomas Harte
0b771ce61a Removes all instances of the copyright symbol. 2018-05-13 15:19:52 -04:00
Thomas Harte
bcc7ad0c30 Corrects bad TMS sprite selections on the top row of the screen. 2018-01-06 16:26:11 -05:00
Thomas Harte
ac80d10cd8 Separates the component parts of running an audio stream: task deferral, filtering and generation.
Walking towards improving opportunities for composition.
2017-12-17 21:26:06 -05:00
Thomas Harte
ec4c259695 Removes unused file. 2017-12-14 21:19:09 -05:00
Thomas Harte
aca7842ca4 Better documents and tidies the TMS9918. 2017-12-14 20:27:26 -05:00