Allow Ooga Booga and NFL 2K2 to connect to network lobbby at https://www.dreamcastlive.net.
Sadly all other tested games still fail the ppp handshake. Work in progress...
Some old 2D games change palettes very frequently, which causes the
texture to be updated for each render.
This change uses a hash of the palette to detect changes, and include
the palette type in the cache key to distinguish textures of different
depth.
Now it actually says something about what exactly it tries to open, also on "cancel" it exits gracefully.
Proposal: instead of checking one thousand return codes, which are easily ignored, I suggest using an exceptions (this part of emulator doesn't require execution speed, so exceptions here would be just fine).
The VMU which was written to disk was never initialized and therefore consisted only of zeroes.
After saving for the first time the VMU was corrupted.
Changes:
* Not only initialize the VMU in memory but also write the correct contents to disk
* If an existing VMU consists only of zeroes overwrite it with the initial empty VMU
Originally ported from nullDC to libretro in commits:
2fa562db1b46c52b663b3dd4bb33a64907357458
f8eb58ac16a9e5adf662b99be5d00729264808e0
Modified for use w/ reicast per-game configuration
A new function was introduced to abstract the setup of input devices for each os: os_SetupInput().
I hope I implemented this everywhere correctly and the behaviour is identical to before.
The new function mcfg_CreateNAOMIJamma() creates the NAOMI Jamma interface and is hidden behind a compile flag (same as before).
The previous function mcfg_CreateDevices() was renamed to mcfg_CreateDevicesFromConfig() because it creates the devices based on the configuration (``nb`` setting).
Don't expose mcfg_Create() directly, but use a proxy method mcfg_CreateController() to create the complete controller.
Using evdev the ``nb``-setting in the ``emu.cfg`` is no longer necessary (see #970) as the devices get created while initializing the controllers.
Moved the ``switch`` to it's own function.
The SH4 sets the signaling bit to 0 for qNaN: 7fbfffff instead of the
usual 7fffffff. Same games seem to rely on this.
Fixes Fur Fighters freeze and missing geometry in game.
The -ffast-math gcc option implies the -ffinite-math-only option, which
produces wrong results with Inf and NaN. Use integer math to detect the
sign of float numbers in FTRC to avoid these issues.
Also the upper cut off value for conversion was apparently wrong.
Also fixed the x86 dynarec but not tested.
Fixes wrong car color in Tokyo Xtreme Racer car selection screen.
div32 matching doesn't handle division by zero and edge cases, which
causes crashes with some games.
Setting enabled by default for Pro Pinball Trilogy.
Instead of ignoring TA data if ListInit hasn't been called, call
ListInit then accept the data. The data will still be probably ignored
but end of list interrupts are sent so game won't hang.
Fixes hang in King of Fighters New Wave (Naomi).
Some games do not align TA_ISP_BASE on 1 MB boundary so the rendering
context is not found at render time. Fixes black screen in NFL 2K2, NBA
2K2, Oooga Booga, Floigan Bros. and probably more Visual Concepts games.
Sadly Ooga Booga still crashes.
The previous algorithm to determine the sort mode for TR polys was
faulty. And the sort mode can vary between passes for multipass
rendering. This also removes the need for the hack to support Always
depth mode with autosorted TR polys. Should fix the Resident Evil 3 menu
problem.
Parse modifier volumes similarily to other polys (first, count, params).
Draw all triangles in one shot and use quad to sum up instead of
redrawing entire strip. Use OR operation for open volumes/quads
(Soulcalibur).
This allows for some very primitive auto-automated testing, by comparing known good frames.
As this happens on the TA level, it doesn't actually require rendering (and thus it's server friendly)
Two new config entries under the new [testing] namespace control behavior
- ta.HashLogFile, file where data should be logged, empty if not logging
- ta.HashCheckFile, past log to check against, empty if not checking
The emu will crash via verify if the logs don't match, and exit(1) if they do
- Rewrite mem ops to only modify regs after exception path
- Throw & catch logic for interpreter that raises the exception
- Re-enabled some commented mmu code
This adds support for separate config and data dirs.
On Linux, these will be compliant XDG Basedir Specification, i.e.
XDG_CONFIG_HOME and XDG_CONFIG_DIRS (or XDG_DATA_HOME and XDG_DATA_DIRS
respectively). On all other platforms, there currently just set to the
homedir path (so no previous behaviour has been changed).
If reicast wants to read and write a data file, it just calls
get_data_path("/samplefile.txt"). If it does not need to write to
that file, it just uses get_data_path("/samplefile.txt", false). That
way, we can also use system-wide dirs (like /usr/share/reicast on
linux), that the user usually doesn't have write access to.
The same applies for config file, where you use get_config_path(args)
respectively.
Reicast doesn't support exceptions yet, so this isn't of much use now,
and is intended mostly as documentation. nullDC used some call stack
hooking magic to handle exceptions, which was never generic and clean
enough to be worth the effort to port to Reicast.
- Setups state, copies binary
- Binary locks up w/ a reboot loop
Naomi roms have a 512-byte header, executable length seems to be
at 368 or 3C0. The rom is copied from [0, len) to 0x0c020000.The
bios then hands over control at 0x0c021000
Here's the original compiler warning:
../../core/hw/sh4/dyna/shil.cpp:700:24: warning: '&&' within '||'
[-Wlogical-op-parentheses]
...if (op->rd.is_reg() && op->rd._reg==reg_sr_T || op->op==shop_ifb)
~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~ ~~
../../core/hw/sh4/dyna/shil.cpp:700:24: note: place parentheses around the '&&'
expression to silence this warning
...if (op->rd.is_reg() && op->rd._reg==reg_sr_T || op->op==shop_ifb)
^
( )
../../core/hw/sh4/dyna/shil.cpp:843:25: warning: '&&' within '||'
[-Wlogical-op-parentheses]
if (op->rs1.is_reg() && op->rs1._reg==reg_sr_T
~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
../../core/hw/sh4/dyna/shil.cpp:843:25: note: place parentheses around the '&&'
expression to silence this warning
if (op->rs1.is_reg() && op->rs1._reg==reg_sr_T
^
( )
../../core/hw/sh4/dyna/shil.cpp:844:25: warning: '&&' within '||'
[-Wlogical-op-parentheses]
|| op->rs2.is_reg() &&
op->rs2._reg==reg_sr_T
~~ ~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
../../core/hw/sh4/dyna/shil.cpp:844:25: note: place parentheses around the '&&'
expression to silence this warning
|| op->rs2.is_reg() && op->rs2._reg==reg_sr_T
^
( )
../../core/hw/sh4/dyna/shil.cpp:845:25: warning: '&&' within '||'
[-Wlogical-op-parentheses]
|| op->rs3.is_reg() && op->rs3._reg==reg_sr_T
~~ ~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
../../core/hw/sh4/dyna/shil.cpp:845:25: note: place parentheses around the '&&'
expression to silence this warning
|| op->rs3.is_reg() && op->rs3._reg==reg_sr_T
^
( )
This compiler warning has been fixed:
../../core/hw/sh4/dyna/decoder.cpp:1181:66: warning: '&&' within '||'
[-Wlogical-op-parentheses]
...|| blk->BlockType==BET_Cond_1 && blk->BranchBlock<=blk->addr)
~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../core/hw/sh4/dyna/decoder.cpp:1181:66: note: place parentheses around the
'&&' expression to silence this warning
...|| blk->BlockType==BET_Cond_1 && blk->BranchBlock<=blk->addr)
^
( )
This is the original warning message:
../../core/hw/holly/sb_mem.cpp:219:12: warning: comparison of unsigned
expression >= 0 is always true [-Wtautological-compare]
if ((base >=0x0000) && (base <=0x001F) /*&& (addr<=0x001FFFFF)*/...
~~~~ ^ ~~~~~~
This consolidates some of the work done for TARGET_NO_NVMEM and
feat/no-direct-memmap. If nvmem is disabled at compile time or alloc
fails _nvmem_enabled() will return false. Various other fixes
and cleanups all around.
- Import naomi code from nullDC, modify and cleanup
- Only unprotected dimm-board support, custom lst files
- Still a compile option
- Boots naomi bios and some games, no input yet
This required moving the SH4_TCB outside the exe, as new tables that belong on the executable are ignored.
This isn't perfect, but there's a large area to scan for available address space so it shouldn't be a problem
- Basic pixel pipeline, a bit better triangle tests, specialized render handlers
- Textures w/ point filtering. Not very smart is it goes 32 -> 16 -> 32 bpp, but works.
- The texture cache is shared rather inelegantly w/ OpenGL one
- Culling
- PParam sorting (shared w/ GL)
The texturing and color blending paths are ugly and slow
- Adapted for indexed, async rendering, shared ta decoder
- This blits via a quick-and-hacky GDI blitter for now
- SSE/MMX intrins based, so not very portable
- A rather not good "reference" implementation
- At least, it's not terribly slow though
- GetTexture moved to Renderer interface
Using external function pointers is broken after the rec-cpp structural changes,
and relatively low priority so disabled for now.
Fixes builds on arm, non-android (see #720)
Since read_params, packet_cmd, read_buff, pio_buff, ata_cmd, cdda and
ByteCount have not been declared static, they are visible in other
translation units. Since their structs are anonymous, their types have
internal linkage only. Thus, accessing them in other translations units
is impossible anyway.
By declaring them static, we're giving it internal linkage so that
neither the type nor the variable are visible in other translation
units.
Here are the errors:
../../core/hw/gdrom/gdromv3.cpp:50:3: warning: anonymous type with no
linkage used to declare variable '<anonymous struct> read_params' with
linkage
} read_params;
^
../../core/hw/gdrom/gdromv3.cpp:83:3: warning: anonymous type with no
linkage used to declare variable '<anonymous struct> packet_cmd' with
linkage
} packet_cmd;
^
../../core/hw/gdrom/gdromv3.cpp:91:3: warning: anonymous type with no
linkage used to declare variable '<anonymous struct> read_buff' with
linkage
} read_buff;
^
../../core/hw/gdrom/gdromv3.cpp💯3: warning: anonymous type with no
linkage used to declare variable '<anonymous struct> pio_buff' with
linkage
} pio_buff;
^
../../core/hw/gdrom/gdromv3.cpp:106:3: warning: anonymous type with no
linkage used to declare variable '<anonymous struct> ata_cmd' with
linkage
} ata_cmd;
^
../../core/hw/gdrom/gdromv3.cpp:122:3: warning: anonymous type with no
linkage used to declare variable '<anonymous struct> cdda' with linkage
} cdda;
^
../../core/hw/gdrom/gdromv3.cpp:156:4: warning: anonymous type with no
linkage used to declare variable '<anonymous union> ByteCount' with
linkage
} ByteCount;
^
- Nest shil implementations to shilop_##name::form::impl for easier template matching
- Add more direct handlers
- Add prints if a direct handler isn't found
- Remove multiple indirect handlers, leave them for fallback
This adds the ability to simply close the emulator when launched
externally and prevents changes designed for linux from generating
conflicting controller configurations on Android (and may be needed for
iOS as well).
- 32 bit code was "fudgy" on this, so simplified via 64bit counters for edge detection
- Maths to count cycles were wrong
- Let's not do 32 bit wrapped maths after beers without any kind of pen-and-paper design
Should fix#594, also see #645
This works, but is extremelly hacky. Must be started without attached debugger, lldb doesn't want to let go of EXC_BAD_ADDRESS, but reicast really depends on it getting delivered as SIGSEGV/SIGBUS. Also xcode has a really bad day upon seeing the jit code. Oh well.
There's some dynarec bug that causes color corruption on bios logo/boot triagles, TA crash on ikaruga and infinitive loop on crazy taxi. I'd guess some fp-memory-write thingy, abi, or smth. Too bad.
- Force code to compile in arm mode (arm jit -> thumb mem functions is complicated)
- SIGILL, SIGBUS. Works w/o Mach exceptions and EXC_BAD_ADDRESS
- Code buffers move to __TEXT, munmapped && memmapped to actually work
- Primitive input. Button + start, or left (works to get out of bios date screen)
- Fixup emitter for thumb2/interworking (didn't work though, reverted to arm cc)
- Block Manager: Disable mem saving / page fault alloc-on-demand logic
- Move cycle counter to r11, r9 is not clean on iOS. Remove r11 from reg alloc list
- Cache flushes for iOS
- log to log.txt
- load game.chd
This is very simplistic, but can already boot and run many games
Known working so far
- Dead Or Alive 2 LE
- Head Hunter
- Marvel VS Capcom 2
Not working
- (Almost) all others
THIS IS A MESS. Sorry @angelXwind :p. We'll have to clean up and merge for r8, but most of the nasty work is there.
What works:
- full init, memalloc, mprotects, etc
- boot
- gles2
What doesn't:
- Exceptions for page faults
- Breaks texture cache
- Breaks dynarec because of fastpath
- segfaults don't work, needs special mach exc_server magic that still eludes me after 5 hours
- Dynarecs. RWX issues, and broken buffer mappings. Also some minnor linking stuff (most of it is taken care of)
- Audio. I just killed it off.
- All of the ui. Was getting in the way of me not knowing how to edit story boards...
- Input
Also, merged branch 'skmp/mac86-hackport' into skmp/ios-wip
Have fun~
- Texkill needs to be done before texture allocation, which is now done during parsing (texture-sync update)
- palette_update needs to be called from the emu thread, before kicking off the render
Fixes#141, "Main menus on POD and Speed Devil show up black"
Seems like the TA doesn't touch any of the memory structures before first write (and probably later on, but we're not gonna support corrupted ta streams for now)
According to p1pkin
«At least in SpeedDevils, at title screen game do a bit weird thing
- Init TA context
- Send lists to TA
- Init the same context again
- Start Render
so, if you are clear stored poly/vertex data at "TA_LIST_INIT" pvr2 reg write - you'll get black screen instead game title»
Fixes#141, "Main menus on POD and Speed Devil show up black"
Seems like the TA doesn't touch any of the memory structures before first write (and probably later on, but we're not gonna support corrupted ta streams for now)
According to p1pkin
«At least in SpeedDevils, at title screen game do a bit weird thing
- Init TA context
- Send lists to TA
- Init the same context again
- Start Render
so, if you are clear stored poly/vertex data at "TA_LIST_INIT" pvr2 reg write - you'll get black screen instead game title»
Fixes#141, "Main menus on POD and Speed Devil show up black"
Seems like the TA doesn't touch any of the memory structures before first write (and probably later on, but we're not gonna support corrupted ta streams for now)
According to p1pkin
«At least in SpeedDevils, at title screen game do a bit weird thing
- Init TA context
- Send lists to TA
- Init the same context again
- Start Render
so, if you are clear stored poly/vertex data at "TA_LIST_INIT" pvr2 reg write - you'll get black screen instead game title»
Notes:
Disc swapping should work on .gdi files (probably .chd files too).
Swapping with .cdi files is not currently supported because we currently create inaccurate TOC for them.
Only x86/Windows has a disc swap handler right now (ie: you can't swap on Android).
-Added config to input fragment
-Reduced logging
-Init the mic at the right time during startup so that port 2 either
gets vmu or mic
-Has not been tested on device without mic
-Needs more testing in games
- Clipping is done on the SH4 side, garbage might be outside the screen
- Mod vols aren't drawn anyway for now
- I wonder how much does this cost on the rend-thread?
Reverted changes in maple_cfg.cpp except a minor memory leak
Fixed an issue with the Xbox 360 controller support, that the first
press of the DPAD would not get handled
Lots of improvments on multi-controller support
- make ta_skip to ta frame skipping instead of skipping parsing entierly. it handles 1 out of every (N+1) frames.
- re purpose rend to unstable opts, as pvr.rend is of no real use right now
Flyinghead