fbc388faa1
more moving crud
2025-01-01 19:43:51 -05:00
aa2e827ce1
small improvement to card read timings
2025-01-01 12:18:36 -05:00
e7cba2ef55
attempt to make read/write timings more accurate?
...
idk what i just wrote tbh
2024-12-31 20:57:44 -05:00
1bd9e87311
move main ram contention junk
2024-12-31 19:18:06 -05:00
9e60f45481
implement savestate serialization
2024-12-31 12:30:49 -05:00
4071099a1f
Revert "redo a bunch of gx handling for marginal accuracy benefits"
...
This reverts commit cc9a6f8ec1
2024-12-29 07:08:14 -05:00
d2737f666c
minor error
2024-12-29 00:04:49 -05:00
cc9a6f8ec1
redo a bunch of gx handling for marginal accuracy benefits
2024-12-28 23:27:42 -05:00
b9472d430e
fix main ram dmas not halting for higher priority dmas
2024-12-28 12:42:44 -05:00
a86a3f8566
fix main ram dmas not stopping for major events
2024-12-27 18:30:52 -05:00
c39db83d5b
tentatively improve gxfifo irq timings?
2024-12-27 17:46:35 -05:00
4d4f374e70
minor optimization
2024-12-27 16:48:49 -05:00
459c088345
Merge remote-tracking branch 'upstream/master' into chemical-x2
2024-12-27 15:45:46 -05:00
a919e7b6af
improve timer irqs
2024-12-27 15:39:09 -05:00
c9e7eec2b5
put out some fires
2024-12-26 18:22:42 -05:00
cec4c2c880
fix handling of halts
2024-12-26 17:06:34 -05:00
21323bd1b7
clunky fix to get the order of interlocks and irqs right
2024-12-26 16:32:12 -05:00
3bd25ae58f
use timestamps for irq handling
...
Also add a (probably not fully correct) 4 cycle delay across the board.
Should improve the timing of when IRQs are noticed by the CPU a fair bit?
2024-12-26 12:52:26 -05:00
0c5dd28b1c
just case the string length to int to make std::min happy in all cases
2024-12-26 09:17:46 +01:00
c41951d49c
Fix almost every warning ( #2195 )
...
Fix almost every warning as of Clang 19
* <codecvt> is deprecated, we can use QString's UTF-16 conversion
instead
* remove sem_timedwait implementation as we don't need it anymore
* remove a useless shift that has its result discarded
* change usages of deprecated sprintf to snprintf
2024-12-25 16:54:10 +01:00
be26878b4c
FIX: this should be namespace std:: to preserve compatibilty with non-glibc when building without gdb stub
2024-12-25 16:48:46 +01:00
66d1091330
improve audio handling at non-60 fps targets ( #2246 )
2024-12-25 16:34:30 +01:00
286240894a
tweaks to card read timings
...
now with actual testing to back them up!
2024-12-24 12:37:38 -05:00
01ba503373
fix fixed src main ram dma timings being off by 1
2024-12-23 19:31:45 -05:00
8d94384907
minor fix?
2024-12-23 19:08:34 -05:00
72c86ade31
Fix gdbstub not activating until the console is reset ( #2245 )
...
The check for initialising the gdbstub depending on whether the JIT was
enabled or not was the wrong way around: previously, it would only
enable the gdbstub if the JIT was enabled.
The stub started working again if you reset the console, as
NDS::SetGdbArgs didn't have any such check and it was called by
EmuInstance::updateConsole.
2024-12-24 00:29:21 +01:00
85dc9f71d9
small improvements to arm9 dma timings
2024-12-23 13:39:59 -05:00
9981a6b98f
fix a timestamp
2024-12-23 12:47:15 -05:00
03dc0c4284
fix another crash
2024-12-21 19:09:03 -05:00
93dd0f47bd
fix an issue related to gx stalls
2024-12-21 18:06:03 -05:00
4940f940c7
improve accuracy and also performance
2024-12-21 17:28:32 -05:00
caa009ca03
better logic for gamecard reads
2024-12-21 13:19:34 -05:00
65e2f64695
"immediate mode" dma start should be delayed 1 cycle
2024-12-21 12:29:37 -05:00
22f1b4d90c
theoretically improve dma responsiveness?
...
in practice seems to make no difference...
2024-12-20 20:10:23 -05:00
98d0a6b371
fix gxfifo stalls being borked under certain circumstances
2024-12-20 10:42:34 -05:00
e0ac68c9f2
Reapply "improve dma accuracy slightly"
...
This reverts commit 14c765d5ed
2024-12-20 10:00:47 -05:00
40527f8897
remove leftover junk, fix an error with resets
2024-12-20 09:49:38 -05:00
dd857e89ef
fix an oversight
2024-12-20 00:33:08 -05:00
6897e4a8be
rework card read timings
2024-12-19 21:55:50 -05:00
14c765d5ed
Revert "improve dma accuracy slightly"
...
apparently it completely broke sm64ds for some reason
2024-12-19 21:54:37 -05:00
0686523543
card reads are "double buffered"
...
means they can fetch a word in the background while a word is still waiting to be read.
Thanks Gericom!
2024-12-19 16:54:12 -05:00
4493634a5b
improve dma accuracy slightly
2024-12-19 15:39:16 -05:00
4a598294a5
clean up a few errors
2024-12-17 23:29:47 -05:00
6ef7a337cb
small optimization to main ram dma
2024-12-17 11:07:37 -05:00
0a5499cb19
fix wrong "running" variable being updated during 16 bit main ram dmas
...
this probably fixes something
2024-12-17 11:02:41 -05:00
e3c8746229
fix a bug preventing main ram dmas from doing their first cycle properly
2024-12-16 20:50:39 -05:00
cacf891330
Revert "run arm 7 first?"
...
Caused performance issues in games???
2024-12-16 13:13:45 -05:00
d383381260
ONLY queue missed gamecard dmas
...
gxfifo dma did not like that prior fix
2024-12-16 13:02:56 -05:00
6d0ab97612
fix gamecard dma breaking w/ main ram contention
...
main ram dma on the arm9 is officially fully operational.
2024-12-16 12:33:13 -05:00
7590c48471
Revert "revert arm9 main ram dma again (again (again))"
...
This reverts commit c90f10d056
2024-12-16 11:45:14 -05:00
Jaklyy