Commit graph

23891 commits

Author SHA1 Message Date
JMC47
48fd27cdab
Merge pull request #8451 from rlnilsen/motion-input-nunchuk
Add motion input support to nunchuk
2020-01-13 17:08:03 -05:00
JosJuice
966e1b31ba
Merge pull request #8394 from Pokechu22/misc-di-gpio
Various DI improvements
2020-01-13 17:17:24 +01:00
Connor McLaughlin
ae6d3be449
Merge pull request #8530 from s-daveb/master
MacOS: Fixes configuration hang; bump MacOS SDK.
2020-01-13 20:21:08 +10:00
Stenzek
d8b2be9d06 DolphinQt: Support compiling on ARM64 2020-01-12 15:23:42 +10:00
Scott Mansell
a8c33f4ef6
Fix trailing whitespace 2020-01-07 12:52:05 +13:00
Scott Mansell
21528c3e72
Document the evdev "interesting" heuristic
Was checking over this old code, and saw a comment calling me out for a lack of documentation.

It might be half a decade late, but better late then never.
2020-01-07 12:46:24 +13:00
Tilka
98f645daac
Merge pull request #8158 from Sintendo/jitopts
x64 micro-optimizations
2020-01-06 14:09:43 +01:00
Tilka
f17f03ea3c
Merge pull request #8551 from Sintendo/jit64addx
Jit64: addx optimizations
2020-01-06 13:15:17 +01:00
Tilka
6e18dfb600
Merge pull request #8133 from Sintendo/mov64imm32
x64Emitter: Emit shorter MOVs for 32-bit immediates
2020-01-06 13:12:56 +01:00
Sintendo
12fcbac2a3 Jit64: addx - Emit LEA for register + immediate
Prefer LEA over MOV + ADD when dealing with immediates.

Before:
44 8B EE             mov         r13d,esi
41 83 C5 20          add         r13d,20h

After:
44 8D 6E 20          lea         r13d,[rsi+20h]
2020-01-05 23:39:13 +01:00
Sintendo
8e7b6f4178 Jit64: addx - Prefer ADD over LEA when possible
The old logic would always emit LEA when both sources are in a register
and OE is disabled. However, ADD is still preferable when one of the
sources matches the destination.

Before:
45 8D 6C 35 00       lea         r13d,[r13+rsi]

After:
44 03 EE             add         r13d,esi
2020-01-05 23:23:56 +01:00
Léo Lam
c1f9bfce30 IOS/ES: Add SetUid exception for the Wii U Transfer Tool
The ES sysmodule in IOS62 (v6430) has an exception for the
Wii U Transfer Tool in the SetUid function.

If the active title is the Wii U Transfer Tool, then calling SetUid
is always allowed. (The UID is still checked first, though.)

Fixes https://bugs.dolphin-emu.org/issues/10985
2020-01-05 18:16:54 +01:00
Léo Lam
f35f4f2bf0
Merge pull request #8541 from jordan-woyak/float-parse-fix
StringUtil: Make TryParse of floats handle comma and dot decimal separators.
2020-01-05 12:12:09 +01:00
Pokechu22
3b5d20e12c Bump state version 2020-01-04 11:43:33 -08:00
Pokechu22
51f8a3606e Return error code to game when using unimplemented commands 2020-01-04 11:43:32 -08:00
Pokechu22
af5f0b20bb Report use of various unimplemented DI commands as game quirks 2020-01-04 11:43:32 -08:00
Pokechu22
f1dc908883 Clarify emulated behavior for ReadBCA 2020-01-04 11:43:32 -08:00
Pokechu22
6c0399103f Handle partitions in /dev/di, not DVDInterface
Partitions are Wii-exclusive, and don't happen at the DVDInterface level in
IOS.  This isn't quite the cleanest fix, but it gets rid of the assumption that
a partition is open on starting the game at least.
2020-01-04 11:43:32 -08:00
Pokechu22
71e8fb278f Return more errors from DTK 2020-01-04 11:43:31 -08:00
Pokechu22
55a88ba2ed Track drive state better, reporting errors if the state is wrong
Also, fix DVDLowStopMotor logging (which was based on the ioctl parameters)
2020-01-04 11:43:31 -08:00
Pokechu22
7d6b9bcb40 Check for error 001 out of bounds reads in DVDThread
All out of bounds reads should return the appropriate DI error, but it also
makes sense to have the error 001 read happen there.
2020-01-04 11:43:30 -08:00
Pokechu22
ef2fc5a49b Split /dev/di commands from DVDInterface
The various ioctls sometimes have different arguments than the DI command
registers, though they generally overlap.  There are also a bunch of ioctls
that don't even normally go into DVDInterface, just returning various data.
Some of the implemented ioctls are new to Dolphin.
2020-01-04 11:43:30 -08:00
Pokechu22
a8ae5fa21a Expose setting DVDInterface errors and split setting into two parts 2020-01-04 11:43:30 -08:00
Pokechu22
3110599559 Increase DIMAR by DILENGTH after a command 2020-01-04 11:43:30 -08:00
Pokechu22
d3aad1d6d5 DIMAR only ignores bits 0-4, not the upper bits
Based on a hardware test on a Wii.  The alignment code was originally added in 743641965a.
2020-01-04 11:43:29 -08:00
Pokechu22
c564d64104 Remove unused drive debug command constants
The corresponding code that used these was removed in 2009 with 93b83f8d65.  Now their only purpose is to generate warnings on osx.

Minimalistic documentation on these commands can be found at http://hitmen.c02.at/files/yagcd/yagcd/chap5.html#sec5.7.2 and https://web.archive.org/web/20070328200323/http://tmb.elitedvb.net/dvd-game/index.php/CMDFE.  Those constants only relate to the 0x11 subcommand, which is one of many.  Most can't be properly emulated unless we LLE the drive firmware (in which case, they don't need to be reimplemented).
2020-01-04 11:43:29 -08:00
Pokechu22
84f099cf62 Tidy and eliminate some of the DI register unions 2020-01-04 11:43:26 -08:00
Pokechu22
11bd132650 Implement RTC flag, which is counter-intuitively disc drive related 2020-01-04 11:43:26 -08:00
Pokechu22
77189e74cd Implement Broadway GPIOs
SLOT_LED and the AVE ones are not implemented yet, but the other Broadway ones are.
2020-01-04 11:43:26 -08:00
Pokechu22
a695b05b21 Add support for std::optional to PointerWrap 2020-01-04 11:43:26 -08:00
Jordan Woyak
0e8d4cb6ac StringUtil: Make TryParse of floats handle comma and dot decimal separators. 2020-01-04 07:19:15 -06:00
Léo Lam
ad75215bb0 Fix several warnings
A small, nonexhaustive set of warning fixes. The DiscIO Volume change
is a workaround for a GCC bug [1] that causes returning an unengaged
std::optional to emit annoying -Wmaybe-uninitialized warnings.
This last change alone fixes pages upon pages of warnings since
Volume.h is included from several files.

-Wstringop-truncation is another irrelevant warning for us, but
unfortunately there seems to be no way to disable it without
adding ugly pragmas wherever the warning appears.
2020-01-04 12:11:39 +01:00
Léo Lam
94c5460693
Merge pull request #8546 from jordan-woyak/default-num-mappings
HW: Add ticks to default mappings that are valid numeric literals.
2020-01-04 10:41:45 +01:00
Léo Lam
671defd8b9
Merge pull request #8543 from jordan-woyak/gcode-download-minor-fix
Core/GeckoCode: Fix DownloadCodes function assuming HTTP data is null terminated.
2020-01-04 10:18:44 +01:00
Jordan Woyak
2a9a04a1ec HW: Add ticks to default mappings that are valid numeric literals. 2020-01-03 22:08:49 -06:00
Jordan Woyak
b505004d61 Core/GeckoCode: Fix DownloadCodes function assuming HTTP data is null terminated. 2020-01-03 20:14:10 -06:00
George Talusan
b1a6cbc3b4 MacOS: Dispatch GL calls to main thread to prevent crashes on Catalina 2020-01-01 23:06:19 -05:00
David Korth
11339d77c6 IsSettingSaveable.cpp: Significant code size reduction.
- Refactor the Config::System::Main check so we check system once,
  then we check for the section.

- Use an std::array<> instead of std::vector<>.

- Use an array of pointers instead of an array of ConfigLocation.
  The latter contains two std::string objects, whereas pointers
  are only 8 bytes (on 64-bit).

Code size comparison: (64-bit Linux, gcc-9.2.0, release build)

   text    data     bss     dec     hex filename
  16136       0      40   16176    3f30 IsSettingSaveable.cpp.o [before]
   3933     720       0    4653    122d IsSettingSaveable.cpp.o [after]
 -12203    +720     -40  -11523   -2d03 Difference
2019-12-29 23:45:02 -05:00
David Korth
f5fe692842 Use pre-increment for iterators instead of post-increment.
Pre-increment is more efficient, since it doesn't have to return the
old iterator.
2019-12-29 23:45:02 -05:00
David Korth
c2dd2e8a2e Use std::istringstream or std::ostringstream instead of std::stringstream where possible.
This removes std::iostream from the inheritance chain, which reduces
overhead slightly.
2019-12-29 23:45:02 -05:00
David Korth
6e549bb668 InputConfig::LoadConfig(): Convert num[] to an array of std::string_view.
NOTE: The explicit std::string() conversions later are needed. Otherwise,
gcc-9.2.0 throws all sorts of errors because it can't find a matching
operator+() function.
2019-12-29 23:45:02 -05:00
David Korth
a23b3d26f4 GLExtensions.cpp: Use arrays of const char *const instead of std::string.
The strings end up being copied, so we might as well initialize the
std::string in the unordered_map directly.
2019-12-29 23:45:02 -05:00
David Korth
05101b251c OGL/Render.cpp, InitDriverInfo(): Use std::string_view to eliminate string copies.
Consolidate the NVIDIA Tegra and non-Tegra checks into one branch.
2019-12-29 23:45:02 -05:00
David Korth
1d7f128693 GetSysMenuVersionString(): Use a char instead of std::string for the region letter.
Append the region letter after determining the version number.
2019-12-29 23:45:02 -05:00
David Korth
9f3b9acad9 PowerPC.cpp: No need to explicitly initialize ppcState.
"ppcState{}" is stored in the .data segment, which means the full ~4 MB
is stored in the executable.

"ppcState" is stored in the .bss segment, which means it only stores a
note that tells it to allocate and zero ~4 MB at runtime.
2019-12-29 23:45:02 -05:00
David Korth
afe2e7de0f VolumeVerifier.cpp: Verify that the string arrays are sorted before using std::binary_sort().
Debug builds only, since this is wrapped in assert().
2019-12-29 23:42:55 -05:00
David Korth
d660aba20f VolumeVerifier.cpp: Use arrays of string_view objects instead of strings.
string_view is a thin wrapper around C strings, so it's more efficient
for constant strings than C++ strings.

The unordered_set<> also adds extra runtime overhead. For small arrays,
a simple linear search works. For larger arrays, std::binary_search()
works better than linear but without the unordered_set<> overhead.

ShouldBeDualLayer(): Removed a duplicate "SK8X52" entry.
2019-12-29 23:42:55 -05:00
Stenzek
d744c5a148 Compile fixes for Windows-on-ARM64 2019-12-28 19:20:41 +10:00
Stenzek
6fcb1c6c46 Add an ARM64 target to Visual Studio projects 2019-12-28 19:20:41 +10:00
Stenzek
5a65031611 Add a Win32 NoGUI platform and project 2019-12-28 08:56:27 +10:00