Commit Graph

11260 Commits

Author SHA1 Message Date
Jules Blok c424aa1ac6 GeometryShaderManager: Boost convergence value.
Most game don't require much granularity regarding the convergence value.
The default convergence was also too low, this commit increases it tenfold.
2014-12-24 23:29:40 +01:00
Jules Blok 70b898e2c6 VideoConfig: Rename Separation to Depth and update descriptions.
More accurate since it's not a real IPD and Nvidia also calls it depth.
2014-12-24 23:28:23 +01:00
Jules Blok 9590ea0cf6 VideoConfig: Turn off stereoscopy when Real XFB is enabled. 2014-12-24 18:45:24 +01:00
Jules Blok 4a86234a79 D3D: Support stereoscopic XFB blit to screen. 2014-12-24 18:45:24 +01:00
Jules Blok a845aeeb3d FramebufferManager: Copy all EFB layers to the XFB. 2014-12-24 18:45:19 +01:00
Jules Blok 59bea317d7 GeometryShaderCache: Cosmetics. 2014-12-24 17:22:24 +01:00
Jules Blok b109b31f61 FramebufferManagerBase: Only allocate one layer for Real XFB. 2014-12-24 17:22:23 +01:00
Jules Blok a9364cd5db OGL: Support stereoscopic XFB blit to screen. 2014-12-24 17:22:23 +01:00
Jules Blok 12412ac5b7 FramebufferManager: Copy all EFB layers to the XFB framebuffer. 2014-12-24 17:22:22 +01:00
Markus Wick 5526b39320 Merge pull request #1748 from Armada651/stereo-format
FramebufferManager: Support stereoscopic EFB format changes.
2014-12-24 11:02:17 +01:00
JosJuice 329af1aea3 Always read 0x20 bytes of disc ID 2014-12-23 13:25:04 +01:00
Jules Blok b2efbdaf44 Cosmetics. 2014-12-23 13:16:09 +01:00
Jules Blok 01718eafa6 FramebufferManager: Use a separate layer variable. 2014-12-23 13:16:03 +01:00
JosJuice 61da252858 Mark cover as open when no disc is inserted 2014-12-23 09:43:51 +01:00
skidau e396286361 Merge pull request #1751 from FenrisulfrX/master
AR: Fix Memory Copy Without Pointer Support
2014-12-23 12:39:48 +11:00
degasus 01cd11a835 OGL: fix ssbo based bbox support 2014-12-22 19:10:35 +01:00
JosJuice 14ff0004b5 Check if there is a disc for all disc reads 2014-12-22 16:34:18 +01:00
JosJuice 867bd2eb34 Only use bounds check for DVDLowUnencryptedRead 2014-12-22 15:01:55 +01:00
Pierre Bourdon c5c10fe48b Merge pull request #1757 from JosJuice/gc-ipl-regression
Fix 4.0-4724 GC IPL regression
2014-12-22 13:03:08 +01:00
JosJuice c1ad4055a3 Fix 4.0-4724 GC IPL regression
Apparently the IPL actually writes to address 0.
2014-12-22 11:38:44 +01:00
Ryan Houdek 5af426df33 Merge pull request #1713 from degasus/vertex-loader
virtual vertex loader
2014-12-22 03:02:02 -06:00
FenrisulfrX e70baf43d4 AR: BugFix to Memory Copy Without Pointer
Using ReadUncheked to avoid changing the state of the emulation
2014-12-22 06:22:01 -02:00
skidau a00aa1a770 Merge pull request #1746 from skidau/dsp-thread-lock
Added lock guards around the DSP LLE cycle count updating.
2014-12-22 17:45:06 +11:00
skidau 4ede885702 Added lock guards around the DSP LLE cycle count updating.
Changed the isRunning flag to a Common::Flag.
2014-12-22 16:40:01 +11:00
Ryan Houdek 0d744cddf6 Merge pull request #1729 from Sonicadvance1/AArch64_compare_me
[AArch64] Add partial support for cmp* instructions.
2014-12-21 20:56:49 -06:00
Ryan Houdek d35f7231dc Merge pull request #1727 from Sonicadvance1/AArch64_conditional_continue
[AArch64] Adds support for conditional continue.
2014-12-21 20:56:36 -06:00
Ryan Houdek 47b67fe527 Merge pull request #1744 from Buddybenj/google-code
Update Outdated Google Code References
2014-12-21 20:55:56 -06:00
Ryan Houdek 02aea1b4c5 [AArch64] Adds support for conditional continue.
Requires PR #1723 prior to merging.
2014-12-21 20:48:40 -06:00
Ryan Houdek 1da6995d5b Merge pull request #1728 from Sonicadvance1/AArch64_BindToRegister
[AArch64] Adds support for BindToRegister to the integer instructions.
2014-12-21 20:48:24 -06:00
Ryan Houdek bd4fa951a9 Merge pull request #1726 from Sonicadvance1/AArch64_flush_if_not_needed
[AArch64] Flush registers from register cache if they won't be used again.
2014-12-21 20:46:38 -06:00
Ryan Houdek 4fad742da4 Merge pull request #1723 from Sonicadvance1/AArch64_loadstores
[AArch64] Implements loadstores with AArch64
2014-12-21 20:46:23 -06:00
Ryan Houdek 5cf79fd3b1 Merge pull request #1724 from Sonicadvance1/AArch64_dispatcher_optimization
[AArch64] Optimization in the dispatcher.
2014-12-21 20:08:40 -06:00
Ryan Houdek e423939c9f Merge pull request #1754 from FioraAeterna/mmulogspam
Eliminate some spammy log messages in MMU mode
2014-12-21 20:05:07 -06:00
Ryan Houdek 4247506c00 [AArch64] Implements loadstore instructions in the JIT recompiler.
These instructions are all implemented with fastmem support.
Currently loads with update are disabled due to an issue that I've yet to figure out.
I'm sure I'll figure that out later.
2014-12-21 19:43:59 -06:00
Ryan Houdek 17a4208fe5 [AArch64] Adds backpatching routines.
Currently supports only integer loadstores. Floating point loadstores will come later.

This system is semi based on the ARMv7 backpatching routine, where we need to initialize our backpatch routine sizes prior to actually using them so
we know we won't be overwriting any memory.
2014-12-21 19:40:29 -06:00
Jules Blok bcfd431e9a Frame: Fix a crash-on-exit on Mac OS X.
A null pointer exception is triggered when Dolphin is quit from the dock.
2014-12-22 02:06:23 +01:00
Fiora e7a49ae5f3 Eliminate some spammy log messages in MMU mode
dcbz: just don't use GetPointer, that can't be right anyways
ppcanalyst: don't print "instruction hex 0" messages in MMU mode, where ISIs
are expected.
2014-12-21 12:41:44 -08:00
Fiora 3b7d7da1ac JIT: Fix redundant MOVs in x86 trampolines
Fixes spammy log messages about redundant MOVs.
2014-12-21 12:06:07 -08:00
FenrisulfrX 46e74782a7 AR: Fix Memory Copy Without Pointer Support
Fix a bug when using AR codes with "Memory Copy Without Pointer Support"
code type which would write more data than it should.
2014-12-21 16:45:50 -02:00
Jules Blok 737bc0e7ad PixelShaderCache: Support stereoscopic EFB format changes. 2014-12-21 15:46:12 +01:00
Jules Blok ba242d27c8 FramebufferManager: Support stereoscopic EFB format changes. 2014-12-21 15:25:58 +01:00
degasus 1efd00227d VertexLoader: Skip vertices with position index = -1 2014-12-21 14:29:44 +01:00
degasus 325e8e370e VertexLoader: Merge dummy functions 2014-12-21 14:13:04 +01:00
degasus 7c486a8c24 VertexLoader: Add a VertexLoader pointer to each function call 2014-12-21 14:13:04 +01:00
degasus 7edf6ec4e4 VertexLoader: Add a test loader which compares two vertex loaders 2014-12-21 14:13:04 +01:00
degasus 809117102e VideoCommon: split VertexLoaderBase from VertexLoader 2014-12-21 14:12:43 +01:00
degasus a71c8158d9 VertexLoader: remove inlined getters
They just blow up the code.
2014-12-21 13:47:43 +01:00
degasus 6e3b2712d2 VideoCommon: Move NativeVertexFormat cache to VertexLoaderManager 2014-12-21 13:47:43 +01:00
degasus 1281798992 VertexLoaderUid: remove operator<
Not needed for unordered map.
2014-12-21 13:47:43 +01:00
degasus ec28a80e00 VideoLoader: remove VAT_*_FRACBITS
They are used to remove the flush amounts, but as we don't
flush anymore on vertex loader changes (only on native
vertex format right now), this optimization is now unneeded.

This will allow us to hard code the frac factors within the
vertex loaders.
2014-12-21 13:47:42 +01:00
degasus 7e12fec7c3 VideoCommon: Make GPU syncing hack optional
This hack is there for quite a long time, and lots of games crashes if it's disabled.
But it's still a hack, so it shouldn't be enabled hard coded. This commit create a new
ini option for this hack which is enabled by default.
Maybe some games does still run very fine without this hack.
2014-12-21 11:54:53 +01:00
Ryan Houdek 9b2909357b Merge pull request #1731 from Sonicadvance1/ARMv7-minor-optimization
[ARMv7] Minor integer fastmem optimization.
2014-12-20 23:38:41 -06:00
Ryan Houdek 8ea6c62943 Merge pull request #1705 from Sonicadvance1/AArch64_emitter_fixes
[AArch64] Improvements to the AArch64 emitter.
2014-12-20 23:27:59 -06:00
Ryan Houdek 59e1a8a1a0 Merge pull request #1736 from degasus/osd
OSD
2014-12-20 23:21:24 -06:00
Benjamin Przybocki fba3c48ec4 Update Outdated Google Code References 2014-12-20 21:17:51 -06:00
Ryan Houdek ab2b499fa6 [ARMv7] Minor integer fastmem optimization.
This is a one instruction optimization for integer loadstores.
Makes sure to enable nop padding in some cases where a fault can still happen and cause us to overwrite other instructions that aren't meant to be.
2014-12-20 19:41:33 -06:00
Ryan Houdek c22e750f74 [AArch64] Add partial support for cmp* instructions.
This support is as far as ARMv7's support for these instructions.
Requires PR #1728 prior to merging.
2014-12-20 19:41:04 -06:00
Ryan Houdek 6f0fbac68f [AArch64] Adds support for BindToRegister to the integer instructions.
This requires PR #1723 prior to merging.
2014-12-20 19:40:14 -06:00
Ryan Houdek 73542ce93e [AArch64] Flush registers from register cache if they won't be used again.
This requires PR #1705 and #1723 prior to merging.
2014-12-20 19:38:34 -06:00
Ryan Houdek 360cf761d2 [AArch64] Optimization in the dispatcher.
Align our dispatcher to a page so we can jump to it with a ADRP+BR pair instead of ADRP+ADD+BR.
Also make sure to save /all/ of our callee saved registers that we are supposed to save.
Requires PR #1705 prior to merging.
2014-12-20 19:37:45 -06:00
Ryan Houdek 13b70c2fbd [AArch64] Add an exception exit function with the exit location already in PC.
This will be used with idle skipping later.
2014-12-20 19:36:48 -06:00
Ryan Houdek 32eb0a9d68 [AArch64] Removes CODE_SIZE variable.
This is in a global location now which conflicts with this one.
2014-12-20 19:36:48 -06:00
Ryan Houdek edfbb6ab9a [ARM] Adds a ArmCommon folder.
Moves the backpatch flags and struct to a common location for ARM.
I'm sure there will be more things that use this common ARM location in the future.
2014-12-20 19:36:48 -06:00
Ryan Houdek eaf17b7d7b [AArch64] Register cache improvements.
Adds the ability to flush the cache and maintain state.
Adds the BindToRegister ability.
Sorts register usage as callee saved used first, reduces dumping pressure when jumping to external routines/interpreter.
Adds a function to store a register, for use when flushing a register that won't be used during the rest of a block.
2014-12-20 19:36:48 -06:00
Ryan Houdek d3c2e8fb0a [AArch64] Improvements to the AArch64 emitter.
Fixes issues with negative offsets in loadstore instructions.
Adds ADRP/ADR instructions.
Optimizes MOVI2R function to take advantage of ADRP on pointers, can change a 3 instruction operation down to one.
Adds GPR push/pop operations for ABI related things.
2014-12-20 19:35:52 -06:00
Pierre Bourdon f2a07e43d1 Merge pull request #1638 from Stevoisiak/DSP-to-Audio
Rename DSP settings to Audio settings
2014-12-21 01:57:39 +01:00
Stevoisiak 91aeb7510f Changed config DSP references to config audio 2014-12-20 19:46:59 -05:00
Ryan Houdek 829132d465 Merge pull request #1732 from Sonicadvance1/stupid_android_garbage
Work around broken Android garbage.
2014-12-20 15:29:38 -06:00
Ryan Houdek f2d998c938 Work around broken Android garbage.
This fixes running Dolphin on the Nexus 9.

Android's EGL stack has internal arrays that they use for tracking OpenGL function usage. Probably has something to do with their OpenGL profiling
garbage that used to be in ADT.

Android has three of these arrays, each statically allocated.
One array is for all GLES 1.x functions
One array is for all GLES 2.0/3.0/3.1 and a couple of extensions they deem worthy of being in this array.
The last array is for all function pointers grabbed via eglGetProcAddress that isn't in the other two arrays.

The last array is the issue that we are having problems with. This array is 256 members in length.
So if you are pulling more than 256 function pointers that Google doesn't track in their internal array, the function will return NULL and yell at you
in logcat.

The Nvidia Shield Tablet gets around this by replacing part of the EGL stack with their own implementation that doesn't have this garbage.
The Nexus 9 on the other hand doesn't get away with this. So we pull >100 more function pointers than the array can handle, and some of those we need
to use.

The workaround for this is to grab OpenGL 1.1 functions last because we won't actually be using those functions, so we get away with not grabbing the
function pointers.
2014-12-20 15:15:57 -06:00
Dolphin Bot f978d141c8 Merge pull request #1720 from Armada651/stereo-msaa
FramebufferManager: Support resolving a multi-layered EFB
2014-12-20 20:35:36 +01:00
degasus a858db1d27 OGL: move osd warning for not supported stereo mode into config validation 2014-12-20 19:54:00 +01:00
degasus ed9c14e0d5 OGL: Fix OGL3 with stereo enabled 2014-12-20 19:13:34 +01:00
Markus Wick 9122c868fe Merge pull request #1734 from Armada651/stereo-fixes
GeometryShaderManager: Set stereo parameters in a SetConstants() call.
2014-12-20 17:48:31 +01:00
Pierre Bourdon dad7911214 Merge pull request #1486 from rohit-n/goto
Remove some gotos.
2014-12-20 16:47:55 +01:00
Pierre Bourdon cee4a85a12 Merge pull request #1633 from JosJuice/dvd-speed-tweak
DVDInterface: Tweak speeds
2014-12-20 16:46:22 +01:00
Pierre Bourdon 24a2ca4d28 Merge pull request #1703 from FioraAeterna/saveregs
JIT: free up a register by eliminating RCODE_POINTERS
2014-12-20 16:45:08 +01:00
Pierre Bourdon e7eb4c06e2 Merge pull request #1634 from JosJuice/disc-drive-unification-lite
Make WII_IPC_HLE_Device_DI call DVDInterface
2014-12-20 16:41:52 +01:00
Jules Blok d37b65c117 FramebufferManager: Support resolving a multi-layered EFB in OGL. 2014-12-20 16:08:00 +01:00
degasus 3322c55484 OGL: don't clear the stencil buffer
We don't have one. This should remove some spamming of the nvidia driver.
2014-12-20 15:17:20 +01:00
degasus 3f9b52e555 OGL: draw shadows within rasterfont itself 2014-12-20 13:31:41 +01:00
Jules Blok d09af2dbba GeometryShaderManager: Set stereo parameters in a SetConstants() call.
Doing it in SetProjectionChanged() is too early because the projection type is not set yet.
2014-12-20 13:16:26 +01:00
degasus fb177ca04e VideoCommon: merge debug test generators 2014-12-20 13:06:29 +01:00
JosJuice d1c8a8bd9f Make WII_IPC_HLE_Device_DI call DVDInterface 2014-12-20 13:02:04 +01:00
degasus ffa014dd48 VideoCommon: merge debug info generators 2014-12-20 12:25:10 +01:00
Dolphin Bot a560d8f150 Merge pull request #1716 from Armada651/geom-wireframe
VideoCommon: Handle wireframe mode in the geometry shader.
2014-12-20 12:22:40 +01:00
degasus e6137407ba GLX: try to get an OpenGL 3.3 core context 2014-12-20 10:57:34 +01:00
Matthew Parlane 6a785af280 Merge pull request #1704 from phire/wii_ssl_segfault
De-memsetify Wii_SSL and fix segfault.
2014-12-20 21:16:55 +13:00
comex fba3abe4cf Manually revert d34418100b 2014-12-20 00:33:15 -05:00
skidau 31bc51072b Merge pull request #1712 from skidau/GC-Adapter-thread-fix
Corrected the buffer that is used by the GCAdapter thread.
2014-12-20 13:09:19 +11:00
Jules Blok bc3ed44050 OGL: Disable geometry shaders on Intel's Windows drivers due to broken interface blocks. 2014-12-19 23:56:02 +01:00
skidau f913cb6ab6 Removed redundant code from the GC Adapter interface. 2014-12-20 09:44:35 +11:00
skidau 9eecfca71d Corrected the buffer that is used by the GCAdapter thread. 2014-12-20 09:44:35 +11:00
Ryan Houdek b30802e2f0 Merge pull request #1721 from Sonicadvance1/fix_adreno_400
Fixes Adreno 400 performance.
2014-12-19 15:55:50 -06:00
Jules Blok 0d79e8f32b VideoCommon: Don't specify the redundant in/out qualifier if GL_ARB_shading_language_420pack is supported.
Some driver developers interpreted "can" as "must" in the OpenGL specs. (I'm looking at you AMD)
2014-12-19 22:45:39 +01:00
Ryan Houdek 8c0e26d969 Minor changes to DriverDetails
Fixes a typo where the official IMGTec drivers were said to be the OSS driver support.
Removes Mali GPU family detection just like I removed the Adreno family detection.
We don't support Mali Utgard anyway.
If we need family detection we can properly add it, right now it isn't needed.
2014-12-19 21:41:12 +00:00
Jules Blok 761749e07f FramebufferManager: Support resolving a multi-layered EFB in D3D. 2014-12-19 22:37:28 +01:00
Jules Blok 0ae082fb61 FramebufferManager: Return the depth texture instead of the color texture in GetResolvedEFBDepthTexture() when AA is enabled. 2014-12-19 22:37:28 +01:00
Ryan Houdek 0fec69001b Fixes Adreno 400 slow performance.
Adreno 300 and 400 have the same video driver performance issues because they are very similar architectures which use basically the same thing with
everything.
There isn't any need to detect the family of the driver with Qualcomm anyway. If we ever need family specific bugs then we can implement real support
for that.
Performance issue on Adreno 400 series was due to us only detecting Adreno 300 series, and with Adreno 400 it wouldn't use the bugs, which would cause
it to use glBufferSubData, causing the huge performance hit.
2014-12-19 21:31:37 +00:00
Lioncash 6adfa92a7e ControllerConfigDiag: Potentially fix scaling problems on high DPI
Fixes issue 7940 hopefully.
2014-12-19 14:50:36 -05:00
Jules Blok 531b3941ff GeometryShaderGen: Re-emit the first vertex when wireframe mode is enabled. 2014-12-19 14:24:08 +01:00