aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2020-04-26cdriso: fix a disk switching deadlockcmitu
2020-04-24Merge pull request #397 from justinweiss/arm-lighting-blendingAutechre
unai: Add ARM-optimized lighting / blending functions
2020-04-20Clean up indentation / add comments for assembly functionsJustin Weiss
2020-03-17Merge pull request #394 from MrHuu/patch-1Autechre
(3DS) Disable dithering by default
2020-03-03(3DS) Disable dithering by defaultMrHuu
Have dithering disabled by default on 3DS.
2020-03-03Merge pull request #391 from justinweiss/async-cd-readsTwinaphex
Add an async CD access method
2020-03-01Preserve uSrc MSB across lighting and blendingJustin Weiss
This saves a few cycles because gcc stores / loads srcMSB when using ARM-optimized lighting.
2020-02-27WIP: Add ARM-assembly versions of lighting and blendingJustin Weiss
2020-02-27Merge pull request #392 from justinweiss/ctr-use-floatmathTwinaphex
[3DS] Tweak compile-time options to increase performance
2020-02-26[3DS] Tweak compile-time options to increase performanceJustin Weiss
2020-02-25Add async CD accessJustin Weiss
2020-02-25Merge pull request #390 from justinweiss/ctr-fix-dynarec-crashesTwinaphex
Fix dynarec crashes on 3DS
2020-02-25Revert "(3DS) Set pthread stack size to 12MB - possible solution for crashes"twinaphex
This reverts commit dfcd7153048c2f6d479dfcd328ee2c0e1e0784ed.
2020-02-24Fix dynarec crashes on 3DSJustin Weiss
After the dynarec writes new instructions, it has to flush the instruction and data caches. Some of these flush operations are privileged on the 3DS, so the clear cache functions have to run through svcBackdoor. The Nintendo implementation (and CFW reimplementation) of svcBackdoor has a problem where interrupts and context switches will cause crashes. Even though we can disable interrupts in the flush function, there's still a window of time between svcBackdoor being called and the function being run where an interrupt will corrupt the stack. Luma3DS implemements a svcCustomBackdoor call we can use that also runs a function in supervisor mode, but uses an implementation that avoids this problem.
2020-02-24Merge pull request #389 from notaz/masterTwinaphex
Merge upstream changes
2020-02-21Merge pull request #388 from ZachCook/masterTwinaphex
Fix second instance runahead without breaking reset
2020-02-21Fix second instance runahead without breaking resetZachary Cook
2020-02-21Revert "Use a more lightweight reset method for lightrec"twinaphex
This reverts commit 7070d5c660963aec56e1d667211b4a09a62d400d.
2020-02-21Merge pull request #386 from ZachCook/masterTwinaphex
Fix second instance runahead
2020-02-19Use a more lightweight reset method for lightrecZachary Cook
Reset is common when using runahead, this seems to prevent crashes when resets occur rapidly.
2020-02-18Add fallback addresses so second instance runahead can workZachary Cook
Also prefer MAP_FIXED_NOREPLACE over MAP_FIXED, as it won't remove old memory maps
2020-02-16(3DS) Set pthread stack size to 12MB - possible solution for crashesTwinaphex
2020-02-16Merge pull request #156 from pcercuei/fix-gte-dividenotaz
Replace existing gte_divider with MAME's version
2020-02-12Replace existing gte_divider with MAME's versionPaul Cercueil
The original code did not have a license header, and as a result was licensed under the GPLv2 which means that it was impossible to link it with LGPL code. The new code originates from MAME and is under a BSD-3-Clause license, which is GPL-compatible, and also much cleaner. Signed-off-by: Paul Cercueil <paul@crapouillou.net>
2020-02-12Silence some unused variable warningstwinaphex
2020-02-12Take out this override for OSXtwinaphex
2020-02-12Buildfixtwinaphex
2020-02-12gte_neon: fix compile under clang's assemblernotaz
2020-02-12add some missing license headersnotaz
2020-02-12gte_neon: fix compile under clang's assemblernotaz
2020-02-11License these files as MITtwinaphex
2020-02-09Update Android.mkTwinaphex
2020-02-09(Android) Aarch64/x86/x64 now has lightrec supporttwinaphex
2020-02-09Buildfix for LLVMtwinaphex
2020-02-09OSX should use lightrecTwinaphex
2020-02-09Fix buildTwinaphex
2020-02-09Update Makefile.libretroTwinaphex
2020-02-09Update psxmem.cTwinaphex
2020-02-08(NEON GPU) Don't hide this for non-ARMtwinaphex
2020-02-08Start getting NEON GPU to work on non-Armtwinaphex
2020-02-08License this MIT - should be compatible with everythingtwinaphex
2020-02-08Merge pull request #381 from negativeExponent/travis_updateTwinaphex
Update .travis.yml
2020-02-08Merge pull request #380 from negativeExponent/fix_makefileTwinaphex
Update Makefile.libretro
2020-02-08menu: fix a buffer overflownotaz
Reported by @wwhheerree, closes #152.
2020-02-09Update .travis.ymlnegativeExponent
2020-02-09Update Makefile.libretronegativeExponent
2020-02-08Merge pull request #378 from pcercuei/libretroTwinaphex
Fix Linux build
2020-02-08Fix Linux buildPaul Cercueil
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
2020-02-08Merge pull request #377 from pcercuei/libretroTwinaphex
Lightrec dynarec
2020-02-08cdrom: add missing BCD->int convert in cdrFreeze()Daniel Silsby
For some Eboot .PBP game images, when loading a savestate, a zlib error like 'uncompress failed with -5 for block 5383, sector 86132' would be output to stdout/stderr. The savestate would still be playable, though. Traced it down to cdrFreeze() failing to convert cdr.Prev from BCD->int when calling ReadTrack(). Note that the odd 'cdr.Prev[0]++;' line seems to be someone's hacky way of forcing ReadTrack() to avoid skipping what it thinks is a duplicate sector read.