H3cJP/citra
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
6,704 Commits 7 Branches 3 Tags
9936d1b9e28feeaa1a4c46bd3aa487fbeacbcdfd
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Subv cb47abecc6 GLCache: Corrected a mismatch between storing compressed sizes and verifying the uncompressed alignment in GetSurface. 2018-06-04 13:01:53 -05:00
Subv d57333406d GPU: Partial implementation of long GPU queries. 
Long queries write a 128-bit result value to memory, which consists of a 64 bit query value and a 64 bit timestamp.

In this implementation, only select=Zero of the Crop unit is implemented, this writes the query sequence as a 64 bit value, and a 0u64 value for the timestamp, since we emulate an infinitely fast GPU.

This specific type was hwtested, but more rigorous tests should be performed in the future for the other types.
 2018-06-03 19:17:31 -05:00
Subv 9cd87a6352 Kernel/Threads: A thread waking up by timeout from a WaitProcessWideKey may already have an assigned lock owner. 
This situation may happen like so:
Thread 1 with low priority calls WaitProcessWideKey with timeout.
Thread 2 with high priority calls WaitProcessWideKey without timeout.
Thread 3 calls SignalProcessWideKey
- Thread 2 acquires the lock and awakens.
- Thread 1 can't acquire the lock and is put to sleep with the lock owner being Thread 2.
Thread 1's timeout expires, with the lock owner still being set to Thread 2.
 2018-06-02 14:06:35 -05:00
Subv 99f9d47d16 GPU: Implemented the DXN1 (BC4) texture format. 2018-06-02 13:17:09 -05:00
Subv c02d7c8ce7 Kernel/Thread: Corrected a typo that caused the affinity mask to never be changed. 2018-05-30 21:36:29 -05:00
Subv 3957b0c34e Kernel/SVC: Support special core values -2 and -3 in svcSetThreadCoreMask. 
Also added some proper error handling.
 2018-05-30 21:36:29 -05:00
Subv d1f9c750a6 Kernel/Thread: Corrected a typo in an assert about the processor id. 2018-05-30 11:32:46 -05:00
Subv 99f12b05fa Shaders: Implemented reading the gl_InstanceID and gl_VertexID variables in the vertex shader. 2018-05-30 10:58:03 -05:00
Sebastian ValleandGitHub 8df011a57f Merge pull request #483 from bunnei/sonic 
Several GPU fixes to boot Sonic Mania
 2018-05-30 07:31:46 -05:00
Subv 734106dcb9 GPU: Implemented the R8 texture format (0x1D) 2018-05-29 21:49:37 -05:00
Sebastian ValleandGitHub 9e30f5574f Merge pull request #471 from bunnei/fmnmx 
shader_bytecode: Implement other variants of FMNMX.
 2018-05-27 09:04:47 -05:00
Subv 7ddc872b52 GPU: Implemented the A1B5G5R5 texture format (0x14) 2018-05-27 09:02:05 -05:00
Subv e2cdf54177 Shader: Implemented compound predicates in fset. 
You can specify a predicate in the fset instruction:

Result = ((Value1 Comp Value2) OP P0) ? 1.0 : 0.0;
 2018-05-24 17:39:59 -05:00
Subv e2db7a83f6 GPU: Allow command lists to rebind a channel to another engine in the middle of the command list. 2018-05-24 17:32:46 -05:00
Subv 126270d963 Shader: Implemented compound predicates in fsetp. 
You can specify three predicates in an fsetp instruction:

P1 = (Value1 Comp Value2) OP P0;
P2 = !(Value1 Comp Value2) OP P0;
 2018-05-24 17:22:36 -05:00
Sebastian ValleandGitHub 882111c4f2 Merge pull request #453 from Subv/thread_callstack 
Qt/WaitTree: Display the callstack for each thread in the wait tree widget
 2018-05-20 20:01:08 -05:00
Sebastian ValleandGitHub 6486544e09 Merge pull request #452 from Subv/psetp 
ShadersDecompiler: Added decoding for the PSETP instruction.
 2018-05-20 20:00:55 -05:00
Sebastian ValleandGitHub 2dbfcd32d7 Merge pull request #451 from Subv/gl_array_size 
GLRenderer: Remove unused vertex buffer and increase the size of the stream buffer to 128 MB.
 2018-05-20 20:00:40 -05:00
Subv 8440cef223 Shaders: Implemented the FMNMX shader instruction. 2018-05-20 17:53:06 -05:00
Subv fd500d3da6 Mutex: Do not assert when the mutex waiting threads list isn't empty on mutex release. 
A thread may own multiple mutexes at the same time, and only release one of them while other threads are waiting for the other mutexes.
 2018-05-20 14:41:58 -05:00
Subv 525492428d GPU: Implemented the nvmap Free ioctl. 
It releases a reference to an nvmap object
 2018-05-20 14:25:57 -05:00
Subv 72b5c448cf GPU: Implemented nvhost-as-gpu's UnmapBuffer ioctl. 
It removes a mapping previously created with the MapBufferEx ioctl.
 2018-05-20 14:25:56 -05:00
Sebastian ValleandGitHub 03388c3071 Merge pull request #450 from Subv/shader_link_error 
GLRenderer: Log the shader source code when program linking fails.
 2018-05-20 12:57:32 -05:00
Sebastian ValleandGitHub 353e1dd7e4 Merge pull request #443 from ogniK5377/ipc-500 
Added IPC RequestWithContext & ControlWithContext
 2018-05-19 17:03:30 -05:00
Sebastian ValleandGitHub dc26601860 Merge pull request #442 from Hexagon12/nfp-service-names 
Updated nfp to have more service names
 2018-05-19 17:01:23 -05:00
Subv 2a35a36251 Kernel/SVC: Signal the highest priority threads first in svcSignalProcessWideKey. 2018-05-19 16:58:30 -05:00
Subv c74f2555b6 Kernel/Threads: Reschedule the proper core when operating on that core's threads. 2018-05-19 16:57:44 -05:00
Subv fab3dd98fe SVC: Removed unused WaitSynchronization1 function 2018-05-19 16:56:33 -05:00
Subv c50393e066 Qt/WaitTree: Display the callstack for each thread in the wait tree widget. 2018-05-19 16:52:49 -05:00
Subv a056d5ad8c ShadersDecompiler: Added decoding for the PSETP instruction. 2018-05-19 11:41:14 -05:00
Subv 98b143c2d6 GLRenderer: Remove unused hw_vao_enabled_attributes variable. 2018-05-19 11:36:38 -05:00
Subv 370ab5df9b GLRenderer: Remove unused vertex buffer and increase the size of the stream buffer to 128 MB. 
The stream buffer is where all the vertex data is copied, some games require this to be much bigger than the 4 MB we used to have.
 2018-05-19 11:36:09 -05:00
Subv 21959ddfef GLRenderer: Log the shader source code when program linking fails. 2018-05-19 11:19:34 -05:00
Subv 0c8b7c00e8 GPU: Don't write to invalid memory locations when handling ioctls that don't have an output. 2018-05-01 14:54:15 -05:00
Subv da32c648bf Shaders: Implemented predicate condition 3 (LessEqual) in the fset and fsetp instructions. 2018-04-29 12:49:41 -05:00
Sebastian ValleandGitHub faa431b274 Merge pull request #418 from bunnei/copy-block-height 
fermi_2d: Fix surface copy block height.
 2018-04-29 09:49:33 -05:00
Subv 20d86d8a36 GPU: Partially implemented the Fermi2D surface copy operation. 
The hardware allows for some rather complicated operations to be performed on the data during the copy, this is not implemented.
Only same-format same-size raw copies are implemented for now.
 2018-04-25 12:54:26 -05:00
Subv e9ad8e9185 Shaders: Added bit decodings for the I2I instruction. 2018-04-25 12:52:55 -05:00
Subv 1740aa5444 Shaders: Implemented the FSET instruction. 
This instruction is similar to the FSETP instruction, but it doesn't set a predicate, it sets the destination register to 1.0 if the condition holds, and 0 otherwise.
 2018-04-25 12:52:32 -05:00
Subv 5ab597041f Memory: Added a missing shortcut for Memory::CopyBlock for the current process. 2018-04-25 11:55:30 -05:00
Subv 1dd4861d38 GPU: Make the Textures::CopySwizzledData function accessible from the outside of the file. 2018-04-25 11:55:30 -05:00
Subv a6da2b93c1 GPU: Added a function to retrieve the bytes per pixel of the render target formats. 2018-04-25 11:55:29 -05:00
Subv 378c881427 GPU: Added surface copy registers to Fermi2D 2018-04-25 11:55:29 -05:00
Subv b1109931b9 GPU: Added boilerplate code for the Fermi2D engine 2018-04-25 11:55:29 -05:00
Subv c16cfbbc6c GPU: Reduce the number of registers of Maxwell3D to 0xE00. 
The rest are just macro shim registers.
 2018-04-25 11:55:28 -05:00
Subv a994446b6e GPU: Move the Maxwell3D macro uploading code to the inside of the Maxwell3D processor. 
It doesn't belong in the PFIFO handler.
 2018-04-25 11:55:27 -05:00
Subv e2f2a49d2d GPU: Corrected the upper bound of the PFIFO method ids in the command processor. 2018-04-25 11:53:54 -05:00
Subv 0369ee7248 Shaders: Added decodings for the FSET instructions. 2018-04-24 22:42:54 -05:00
Subv f208953585 GPU: Added asserts to our code for handling the QUERY_GET GPU command. 
This is based on research from nouveau. Many things are currently unknown and will require hwtests in the future.
This commit also stubs QueryMode::Write2 to do the same as Write. Nouveau code treats them interchangeably, it is currently unknown what the difference is.
 2018-04-23 17:06:57 -05:00
Subv 9531a29283 GPU: Support multiple enabled vertex arrays. 
The vertex arrays will be copied to the stream buffer one after the other, and the attributes will be set using the ARB_vertex_attrib_binding extension.

yuzu now thus requires OpenGL 4.3 or the ARB_vertex_attrib_binding extension.
 2018-04-23 11:34:50 -05:00
Subv 46572d027d Kernel: Implemented mutex priority inheritance. 
Verified with a hwtest and implemented based on reverse engineering.

Thread A's priority will get bumped to the highest priority among all the threads that are waiting for a mutex that A holds.
Once A releases the mutex and ownership is transferred to B, A's priority will return to normal and B's priority will be bumped.
 2018-04-23 11:23:44 -05:00
Subv 0d6eafe11a NvDrv/nvhost-as-gpu: Ensure that the object passed to MapBufferEx has already been allocated. 
Also added a consistency check and a comment for the case when the object id is different than its handle. The real nvservices doesn't make a distinction between ids and handles, each object gets an unique handle which doubles as its id.
 2018-04-23 11:21:46 -05:00
Subv e4bd0bddea Nvdrv/nvhost-as-gpu: Implemented the ioctl REMAP command. 
It takes a previously-reserved (AllocateSpace) GPU memory address and maps it to the address of the nvmap object passed to Remap.
 2018-04-23 11:21:46 -05:00
Subv e862c50a70 Nvdrv: Assert when receiving an unimplemented ioctl in the nv* handlers. 2018-04-23 11:13:53 -05:00
Subv f823c1d599 GPU: Make the GPU virtual memory manager use 16 page bits and 10 page table bits. 
Also removed some dead code and added memory map consistency asserts.
 2018-04-23 10:57:12 -05:00
Subv 010227e149 GPU: Implement the RGB10_A2 RenderTarget format, it will use the same format as the A2BGR10 texture format. 2018-04-23 10:50:28 -05:00
Subv c079cf4eec GPU: Implement the A2BGR10 texture format. 2018-04-21 17:32:25 -05:00
Subv 17a0ef1e1e ShaderGen: Implemented the KIL instruction, which is equivalent to 'discard'. 2018-04-20 21:09:34 -05:00
Subv c3a8ea76f1 ShaderGen: Implemented predicated instruction execution. 
Each predicated instruction will be wrapped in an `if (predicate) { instruction_body; }` in the GLSL, where `predicate` is one of the predicate boolean variables previously set by fsetp.
 2018-04-20 21:09:33 -05:00
Subv 0a5e01b710 ShaderGen: Implemented the fsetp instruction. 
Predicate variables are now added to the generated shader code in the form of 'pX' where X is the predicate id.
These predicate variables are initialized to false on shader startup and are set via the fsetp instructions.

TODO:

* Not all the comparison types are implemented.
* Only the single-predicate version is implemented.
 2018-04-20 21:09:33 -05:00
Subv a70ed9c8ae Kernel: Use 0x2C as default main thread priority for homebrew and lone NRO/NSOs 2018-04-20 21:04:35 -05:00
Subv 013778aa21 Qt: Update the WaitTree widget to show info about the current mutex of each thread. 2018-04-20 21:04:34 -05:00
Subv be155f4d9d Kernel: Remove unused ConditionVariable class. 2018-04-20 21:04:33 -05:00
Subv 5fdfbfe25a Kernel: Remove old and unused Mutex code. 2018-04-20 21:04:32 -05:00
Subv b18ccf9399 Kernel: Properly implemented svcWaitProcessWideKey and svcSignalProcessWideKey 
They work in tandem with guest code to provide synchronization primitives along with svcArbitrateLock/Unlock
 2018-04-20 21:04:27 -05:00
Subv e81a2080eb Kernel: Corrected the implementation of svcArbitrateLock and svcArbitrateUnlock. 
Switch mutexes are no longer kernel objects, they are managed in userland and only use the kernel to handle the contention case.
Mutex addresses store a special flag value (0x40000000) to notify the guest code that there are still some threads waiting for the mutex to be released. This flag is updated when a thread calls ArbitrateUnlock.

TODO:
* Fix svcWaitProcessWideKey
* Fix svcSignalProcessWideKey
* Remove the Mutex class.
 2018-04-20 21:04:25 -05:00
Subv d03fc77475 ShaderGen: Register id 255 is special and is hardcoded to return 0 (SR_ZERO). 2018-04-20 14:57:40 -05:00
Subv 2e0a9f66a0 ShaderGen: Ignore the 'sched' instruction when generating shaders. 
The 'sched' instruction has a very convoluted encoding, but fortunately it seems to only appear on a fixed interval (once every 4 instructions).
 2018-04-20 14:57:40 -05:00
Subv fe84842137 ShaderGen: Implemented the fmul32i shader instruction. 2018-04-19 13:46:32 -05:00
Subv 5367935d35 ShaderGen: Fixed a case where the TEXS instruction would use the same registers for the input and the output. 
It will now save the coords before writing the outputs in a subscope.
 2018-04-19 13:33:17 -05:00
Subv 057170928c GPU: Add support for the DXT23 and DXT45 compressed texture formats. 2018-04-18 20:48:53 -05:00
Subv 2985056340 GPU: Implemented the B5G6R5 format. 2018-04-18 18:16:45 -05:00
Subv 43d98ca8fe GLCache: Added boilerplate code to make supporting configurable texture component types. 
For now only the UNORM type is supported.
 2018-04-18 14:17:28 -05:00
Subv 5b3fab6766 GLCache: Unify texture and framebuffer formats when converting to OpenGL. 2018-04-18 14:17:28 -05:00
Subv b2c1672e10 GPU: Texture format 8 and framebuffer format 0xD5 are actually ABGR8. 2018-04-18 14:17:27 -05:00
Subv 48d4efbd69 GPU: Pitch textures are now supported, don't assert when encountering them. 2018-04-18 12:52:53 -05:00
Subv a3e82e8e1f GLCache: Take into account the texture's block height when caching and unswizzling. 2018-04-18 12:52:53 -05:00
Subv ac09b5a2e9 GLCache: Added a function to convert cached PixelFormats back to texture formats. 
TODO: The way we handle cached formats must change, framebuffer and texture formats are too different to keep them in the same place.
 2018-04-18 12:52:52 -05:00
Subv 6b63aaa5b4 GPU: Allow using a configurable block height when unswizzling textures. 2018-04-18 12:52:51 -05:00
Subv db5f2bfa7e GPU/TIC: Added the pitch and block height fields to the TIC structure. 2018-04-18 11:38:39 -05:00
Subv 636ad34707 MaxwellToGL: Implemented tex wrap mode 1 (Wrap, GL_REPEAT). 2018-04-17 10:17:18 -05:00
Subv 7fc516cc1a MaxwellToGL: Added a TODO and partial implementation of maxwell wrap mode 4 (Clamp, GL_CLAMP). 
This clamp mode was removed from OpenGL as of 3.1, we can emulate it by using GL_CLAMP_TO_BORDER to get the border color of the texture, and then manually sampling the edge to mix them in the fragment shader.
 2018-04-17 10:16:50 -05:00
Subv 477aab5960 GPU: Use the same buffer names in the generated GLSL and the buffer uploading code. 2018-04-15 15:02:50 -05:00
Subv 14ac40436e GPU: Don't use explicit binding points when uploading the constbuffers to opengl. 
The bindpoints will now be dynamically calculated based on the number of buffers used by the previous shader stage.
 2018-04-15 14:14:57 -05:00
Subv e128e90350 GPU: Don't use GetPointer when uploading the constbuffer data to the GPU. 2018-04-15 11:18:09 -05:00
Subv 7da47da66e GPU: Use the buffer hints from the shader decompiler to upload only the necessary const buffers for each shader stage. 2018-04-15 11:15:54 -05:00
Subv c9b511da08 GPU: Upload the entirety of each constbuffer for each shader stage as SSBOs. 
We're going to need the shader generator to give us a mapping of the actual used const buffers to properly bind them to the shader.
 2018-04-14 23:02:05 -05:00
Subv 1957640ea2 GPU: Allow configuring ssbos in the opengl state manager. 2018-04-14 22:54:23 -05:00
Subv ae58e46036 GPU: Added a function to determine whether a shader stage is enabled or not. 2018-04-14 22:54:23 -05:00
SubvandJames Rowe dcc27d6dc1 GPU: Assert when finding a texture with a format type other than UNORM. 2018-04-06 20:44:46 -06:00
SubvandJames Rowe b0ca330e14 GL: Set up the textures used for each draw call. 
Each Maxwell shader stage can have an arbitrary number of textures, but we're limited to a certain number in OpenGL. We try to only use the minimum amount of host textures by not keeping a 1:1 relation between guest texture ids and host texture ids, ie, guest texture id 8 can be host texture id 0 if it's the only texture used in the guest shader program.
This mapping will have to be passed to the shader decompiler so it can rewrite the texture accesses.
 2018-04-06 20:44:46 -06:00
SubvandJames Rowe cb3183212d GL: Bind the textures to the shaders used for drawing. 2018-04-06 20:44:46 -06:00
SubvandJames Rowe 65faeb9b2a GLCache: Specialize the MortonCopy function for the DXT1 texture format. 
It will now use the UnswizzleTexture function instead of the MortonCopyPixels128, which doesn't seem to work for textures.
 2018-04-06 20:44:46 -06:00
SubvandJames Rowe b258403f0d GLCache: Implemented GetTextureSurface. 2018-04-06 20:44:45 -06:00
SubvandJames Rowe 65ea52394b GLCache: Support uploading compressed textures to the GPU. 
Compressed texture formats like DXT1, DXT2, DXT3, etc will use this to ease the load on the CPU.
 2018-04-06 20:44:45 -06:00
SubvandJames Rowe 73eaef9c05 GL: Remove remaining references to 3DS-specific pixel formats 2018-04-06 20:44:42 -06:00
SubvandJames Rowe b305646c44 RasterizerCache: Remove 3DS-specific pixel formats. 
We're only left with RGB8 and DXT1 for now. More will be added as they are needed.
 2018-04-06 20:40:24 -06:00
SubvandJames Rowe c28ed85875 GL: Create the sampler objects when starting up the GL rasterizer. 2018-04-06 20:40:24 -06:00
SubvandJames Rowe ca96b04a0c GL: Ported the SamplerInfo struct from citra. 2018-04-06 20:40:24 -06:00
SubvandJames Rowe 0171ec606b GL: Rename PicaTexture to MaxwellTexture. 2018-04-06 20:40:24 -06:00