notaz [Sat, 6 Jan 2018 23:20:00 +0000 (01:20 +0200)]
drc: support ms ABI
notaz [Sat, 6 Jan 2018 19:29:59 +0000 (21:29 +0200)]
improve 64bit portability
for win64 mostly
notaz [Wed, 3 Jan 2018 01:00:32 +0000 (03:00 +0200)]
libretro: pass required arg to VirtualProtect
notaz [Sat, 5 Apr 2014 19:08:06 +0000 (22:08 +0300)]
add a pitch variable
32x is still broken
notaz [Wed, 27 Dec 2017 17:53:32 +0000 (19:53 +0200)]
Merge pull request #72 from Chips-fr/master
Support latest raspberry GLES library name
Chips-fr [Mon, 25 Dec 2017 16:38:36 +0000 (17:38 +0100)]
Support latest raspberry GLES library name
notaz [Mon, 11 Dec 2017 00:53:17 +0000 (02:53 +0200)]
drc: revive x86 dynarec, support x86-64
notaz [Sun, 3 Dec 2017 16:23:27 +0000 (18:23 +0200)]
don't do idle log - it's not useful
notaz [Sun, 3 Dec 2017 15:41:30 +0000 (17:41 +0200)]
drc: arm: use movw/movt
it's about time...
notaz [Sun, 3 Dec 2017 15:39:59 +0000 (17:39 +0200)]
drc: do lit check before size_nolit is cleared
notaz [Sun, 3 Dec 2017 01:40:13 +0000 (03:40 +0200)]
drc: ignore cache-through on smc check
notaz [Sat, 2 Dec 2017 21:55:27 +0000 (23:55 +0200)]
drc: rm overlapped block entry points
otherwise we get duplicates in hash tables
notaz [Fri, 1 Dec 2017 22:10:08 +0000 (00:10 +0200)]
stop IdleDet from affecting save states
notaz [Wed, 29 Nov 2017 00:31:19 +0000 (02:31 +0200)]
32x: improve 68k bios handling
notaz [Tue, 28 Nov 2017 00:04:08 +0000 (02:04 +0200)]
32x: add other timing hacks
For sdram sync, like NJTE.
Still bad, but don't have a better solution for now (or ever?).
notaz [Sun, 26 Nov 2017 18:19:40 +0000 (20:19 +0200)]
32x: remove some comm hacks
they can (and do) easily break things
notaz [Sat, 21 Oct 2017 23:04:26 +0000 (02:04 +0300)]
move more globals to PicoInterface
similar reasons as before
notaz [Sat, 21 Oct 2017 21:39:43 +0000 (00:39 +0300)]
sms: do psg like md does
notaz [Fri, 20 Oct 2017 21:02:38 +0000 (00:02 +0300)]
handle 'bad' dma better
notaz [Thu, 19 Oct 2017 21:57:34 +0000 (00:57 +0300)]
pandora: mark prerelease versions
so they can live along released ones
notaz [Thu, 19 Oct 2017 21:41:12 +0000 (00:41 +0300)]
add 68k overclocking support
notaz [Wed, 18 Oct 2017 23:38:20 +0000 (02:38 +0300)]
rearrange globals
scripted find/replace
gives slightly better code on ARM, less unnecessary asm,
~400 bytes saved
notaz [Wed, 18 Oct 2017 23:44:56 +0000 (02:44 +0300)]
pandora: fix build
Fixes:
df9251536de "libretro: satisfy vita's dynarec needs in a cleaner way"
notaz [Mon, 16 Oct 2017 22:01:26 +0000 (01:01 +0300)]
sms: more md-consistent drawing
notaz [Mon, 16 Oct 2017 21:53:35 +0000 (00:53 +0300)]
sms: improve sr a bit
notaz [Sat, 14 Oct 2017 22:15:00 +0000 (01:15 +0300)]
famec: eliminate global context ptr
saves like 25-35K of .text
current compile resource usage on i5-6600K:
cpu mem
gcc 5.4.0: 17.0 1.1g
clang 3.8: 1686 2.3g
FAMEC_NO_GOTOS:
gcc 5.4.0: 8.4 0.4g
clang 3.8: 20.0 0.15g
vs2008/O2: ~1800 ?
vs2008/O1: ~720 ?
notaz [Sat, 14 Oct 2017 21:45:55 +0000 (00:45 +0300)]
famec: split fm68k_emulate
in FAMEC_NO_GOTOS mode at least
notaz [Sat, 14 Oct 2017 18:10:25 +0000 (21:10 +0300)]
get rid of custom memcpy funcs
not used for anything important, just a maintenance burden
notaz [Fri, 13 Oct 2017 23:13:40 +0000 (02:13 +0300)]
fix clang build
yet another workaround for it...
notaz [Sat, 14 Oct 2017 18:28:24 +0000 (21:28 +0300)]
let it build on msvc
supposedly for the original XBox?
notaz [Fri, 13 Oct 2017 20:53:15 +0000 (23:53 +0300)]
drop draw_amips from libretro too
There's no proof it's any faster, it's only a maintenance burden.
See also
4aedc593008ca6e9230b700cec6483c3ecd73bef .
notaz [Fri, 13 Oct 2017 20:49:13 +0000 (23:49 +0300)]
drop some unnecessary inlines
apparently somebody compiles with msvc?
notaz [Fri, 13 Oct 2017 20:38:09 +0000 (23:38 +0300)]
android: make armeabi buildable
notaz [Fri, 13 Oct 2017 20:36:51 +0000 (23:36 +0300)]
libretro: satisfy vita's dynarec needs in a cleaner way
notaz [Fri, 13 Oct 2017 20:36:36 +0000 (23:36 +0300)]
libretro: update for core changes
also drops the broken "fps override" feature
notaz [Fri, 13 Oct 2017 20:31:22 +0000 (23:31 +0300)]
libretro: drop sram clear
It's done by the core. If the core is missing something, core itself
needs to be fixed.
notaz [Fri, 13 Oct 2017 20:30:44 +0000 (23:30 +0300)]
some portability cleanups
notaz [Thu, 12 Oct 2017 21:39:51 +0000 (00:39 +0300)]
re-import all libretro code from it's fork
Verbatim copy from https://github.com/libretro/picodrive/ commit
9ae88ef15ff00cacc3877c7ecc13b0092bab50b8 , so look there for the history
of libretro specific changes. Unfortunately there is too much noise and
divergence to merge this in a proper way.
iLag [Sun, 26 Mar 2017 03:28:08 +0000 (20:28 -0700)]
Fix remaining bugs and fix indentation
iLag [Sun, 26 Mar 2017 02:35:36 +0000 (19:35 -0700)]
Restore support for short GG cheats.
iLag [Sun, 26 Mar 2017 02:29:00 +0000 (19:29 -0700)]
Add support for Master System cheats.
notaz: drop wrong code (PicoWrite8_z80 are 68k functions, not for sms)
iLag [Wed, 22 Mar 2017 21:19:35 +0000 (14:19 -0700)]
Add support for single-line cheats and PAR codes.
notaz: include required header
David [Tue, 24 Nov 2015 12:49:31 +0000 (13:49 +0100)]
Fix the lprintf method (libretro side)
The lprintf method used the format string instead of the buffer build
from va_list
notaz [Thu, 12 Oct 2017 21:40:45 +0000 (00:40 +0300)]
build: some clean ups
notaz [Thu, 12 Oct 2017 21:40:30 +0000 (00:40 +0300)]
build: drop function-sections
this has provoked people to do some horrible ifdeffery in libretro fork,
while the benefit is minimal, if any.
notaz [Wed, 4 Oct 2017 20:01:56 +0000 (23:01 +0300)]
release 1.92
notaz [Sun, 8 Oct 2017 19:33:04 +0000 (22:33 +0300)]
fix cell scroll drawing
notaz [Mon, 9 Oct 2017 23:00:28 +0000 (02:00 +0300)]
fame: fix roxr
also take interrupts on exit, like other cores do
notaz [Mon, 9 Oct 2017 22:13:48 +0000 (01:13 +0300)]
rework sr
note to self:
h32 0x10A .. 0x127 0x1D2 .. 0x1FF 0x000 .. 0x109
pclk 30 | 46 | 266 = 342
hbset 0x126 ... 0x009
pclk 29 | 1 + 46 + 10 | 256
mclk 290 | 570 | 2560 = 3420
68kclk 41.4 81.4 365.7 ~= 488.5
h40 0x14A .. 0x16C 0x1C9 .. 0x1FF 0x000 .. 0x149
pclk 35 | 55 | 330 = 420
hbset 0x166 ... 0x00A
pclk 28 | 7 + 55 + 11 | 319
mclk 28*8 | 7*8 4*8+314+10+(18+11)*8 | 319*8 = 3420
68kclk 32 92 364.5 ~= 488.5
notaz [Fri, 6 Oct 2017 21:08:05 +0000 (00:08 +0300)]
some irq hacks
notes to self (derived from Nemesis):
H32: 0x10A .. 0x127 0x1D2 .. 0x1FF 0x000 0x001
pclk 30 46 1 |0x109
mclk 770 | 2650 = 3420
68kclk 110 378.5 ~= 488.5
H40: 0x14A .. 0x16C 0x1C9 .. 0x1FF 0x000 0x001
pclk 35 55 1
mclk 35*8 4*8+314+10+19*8 |0x149*8
788 | 2632 = 3420
68kclk 112.5 376 ~= 488.5
notaz [Thu, 5 Oct 2017 22:26:12 +0000 (01:26 +0300)]
update libpicofe
notaz [Thu, 5 Oct 2017 22:31:11 +0000 (01:31 +0300)]
fix t574 reset
notaz [Wed, 4 Oct 2017 23:06:21 +0000 (02:06 +0300)]
split memories away from Pico
saves ~3k of code on ARM because Pico no longer crosses ldr limit
notaz [Thu, 5 Oct 2017 22:34:24 +0000 (01:34 +0300)]
drop legacy save support
the format has been changed like 10 years ago
notaz [Wed, 4 Oct 2017 20:48:57 +0000 (23:48 +0300)]
drop unmaintained mips code
can always be brought back if anyone has proof it's faster
notaz [Tue, 3 Oct 2017 23:58:48 +0000 (02:58 +0300)]
vdp timing hacks
notaz [Tue, 3 Oct 2017 23:23:27 +0000 (02:23 +0300)]
adjust z80 timing a bit
notaz [Tue, 3 Oct 2017 20:02:47 +0000 (23:02 +0300)]
fix various logging issues
notaz [Tue, 3 Oct 2017 20:00:39 +0000 (23:00 +0300)]
menu: add reset hotkey
notaz [Mon, 2 Oct 2017 20:38:12 +0000 (23:38 +0300)]
some more risky timing changes
But should be closer to the real thing.
Let's see what this will break...
notaz [Sat, 30 Sep 2017 23:42:24 +0000 (02:42 +0300)]
fix underalloc
notaz [Sat, 30 Sep 2017 22:29:08 +0000 (01:29 +0300)]
some poor timing improvement attempts
notaz [Sat, 30 Sep 2017 21:59:44 +0000 (00:59 +0300)]
slightly better z80 vdp reads
notaz [Sat, 30 Sep 2017 21:03:54 +0000 (00:03 +0300)]
improve fast forward for cd
still broken, but a little bit better
notaz [Sat, 30 Sep 2017 19:55:57 +0000 (22:55 +0300)]
handle frequent psg updates
notaz [Sat, 30 Sep 2017 15:14:20 +0000 (18:14 +0300)]
sn76496: simplify writes
notaz [Fri, 29 Sep 2017 23:23:13 +0000 (02:23 +0300)]
clean up dac code a bit
notaz [Fri, 29 Sep 2017 21:17:04 +0000 (00:17 +0300)]
state: setup banks before CPUs on load
notaz [Fri, 6 Oct 2017 22:36:22 +0000 (01:36 +0300)]
simple ym2612 fixes
notaz [Fri, 29 Sep 2017 21:19:55 +0000 (00:19 +0300)]
fix ym2612 asm, rework EG
this should be split, but I'm lazy
EG saves ~900 bytes
notaz [Thu, 28 Sep 2017 00:16:21 +0000 (03:16 +0300)]
some support for vdp debug reg
notaz [Thu, 28 Sep 2017 00:19:36 +0000 (03:19 +0300)]
simplify tile drawing
cuts away 1126 bytes of code on x86
notaz [Tue, 26 Sep 2017 23:59:50 +0000 (02:59 +0300)]
scroll size improvement
notaz [Tue, 26 Sep 2017 21:20:52 +0000 (00:20 +0300)]
some support for 128k mode
notaz [Mon, 25 Sep 2017 23:59:31 +0000 (02:59 +0300)]
allow limited z80 vdp access
lame and totally wrong timing
notaz [Mon, 25 Sep 2017 20:41:34 +0000 (23:41 +0300)]
recognize the MED ssf2 header
notaz [Mon, 25 Sep 2017 20:37:54 +0000 (23:37 +0300)]
relax the bank check
there is 512K padding on load
notaz [Mon, 25 Sep 2017 20:34:34 +0000 (23:34 +0300)]
fame: always update the fetch map
should just get rid of that thing, it's annoying
notaz [Mon, 25 Sep 2017 00:00:20 +0000 (03:00 +0300)]
don't miss the sprite check
DrawAllSprites depends on HighLnSpr, so prepare in DrawAllSprites was a
bad idea as lots of things may have changed
notaz [Sat, 23 Sep 2017 22:27:31 +0000 (01:27 +0300)]
some dma improvements
no idea if anything needs this, shouldn't hurt either
notaz [Thu, 21 Sep 2017 23:16:15 +0000 (02:16 +0300)]
make dma honour banking
I was sure I had it done before, turns out not.
notaz [Thu, 21 Sep 2017 23:00:50 +0000 (02:00 +0300)]
integrate SPI EEPROM
notaz [Thu, 21 Sep 2017 21:46:42 +0000 (00:46 +0300)]
import SPI EEPROM from Genesis-Plus-GX
Author: EkeEke
notaz [Wed, 20 Sep 2017 20:46:28 +0000 (23:46 +0300)]
update libpicofe
notaz [Wed, 20 Sep 2017 20:43:38 +0000 (23:43 +0300)]
memory: allow SRAM word writes
Was allowing both low and high byte writes by byte, but not word,
which was stupid.
notaz [Wed, 20 Sep 2017 20:39:20 +0000 (23:39 +0300)]
workaround idledet breaking op test
notaz [Wed, 20 Sep 2017 20:37:58 +0000 (23:37 +0300)]
68k cores: fix bcd instructions
passing flamewing's test now
notaz [Tue, 12 Sep 2017 22:58:44 +0000 (01:58 +0300)]
famec: adjust divu timing
it's wrong, but I need it to be consistent with other cores now
notaz [Tue, 12 Sep 2017 22:57:37 +0000 (01:57 +0300)]
famec: fix CHK
notaz [Tue, 12 Sep 2017 22:55:25 +0000 (01:55 +0300)]
famec: improve trace a bit
or break it (nobody uses it anyway), just want m68k_opcode_sizes to pass
notaz [Tue, 12 Sep 2017 22:46:30 +0000 (01:46 +0300)]
musashi: don't generate unneeded handlers
at least move16 was incorrectly hooked up
notaz [Mon, 11 Sep 2017 22:26:49 +0000 (01:26 +0300)]
musashi: run same amount of cycles as other cores
for consistency
notaz [Mon, 11 Sep 2017 22:25:26 +0000 (01:25 +0300)]
musashi: fix build on newer gcc
notaz [Fri, 18 Aug 2017 21:38:03 +0000 (00:38 +0300)]
32x: skip unnecessary bios work in cd mode
notaz [Fri, 18 Aug 2017 00:44:25 +0000 (03:44 +0300)]
sh2: handle some branch exceptions
notaz [Thu, 10 Aug 2017 23:12:57 +0000 (02:12 +0300)]
drz80: drop fast_sp for compatibility
no measurable improvement in most cases anyway
notaz [Mon, 7 Aug 2017 23:25:37 +0000 (02:25 +0300)]
android: solve "text segment is not shareable"
Some asm is disabled and some performance is lost by this.
notaz [Sun, 6 Aug 2017 20:22:53 +0000 (23:22 +0300)]
get some gp2x stuff out of the way for others
notaz [Sun, 6 Aug 2017 20:02:28 +0000 (23:02 +0300)]
remove dead code
notaz [Sun, 6 Aug 2017 19:25:26 +0000 (22:25 +0300)]
eliminate texrels, part 3