Compare commits

...

161 Commits

Author SHA1 Message Date
5cdc08f6ea Vertex spirv 2023-01-05 17:00:43 +02:00
8779cb7785 renderer_opengl: Fix shader compilation
* Also use glCopyImageSubData to do texture copies'
2023-01-02 16:22:51 +02:00
a6ca7dca61 vk_instance: Don't specify uint8 extension twice
* I have no idea why this happened
2023-01-02 15:40:06 +02:00
cce6a79a91 vk_shader_gen_spv: Fix OpCompositeConstruct bug
* Fixes graphics on the 2d_shapes homebrew and maybe other games
2023-01-02 15:40:00 +02:00
230a463a39 msvc ci: Setup Vulkan SDK for glslangValidator 2023-01-02 00:25:28 +02:00
83e734bd6a vk_shader_gen: new is a reserved keyword on Metal 2022-12-31 20:24:31 +02:00
72ee29669a renderer_vulkan: Add support for VK_KHR_image_format_list
* May help drivers when using mutable images
2022-12-31 18:20:08 +02:00
b1a02e1710 renderer_vulkan: Remove upload_cmdbuf
* No longer needed with the new stream buffer
2022-12-31 17:44:15 +02:00
2c34f41747 vk_rasterizer: Don't bind redundant bindings 2022-12-31 10:07:35 +02:00
60d59730a9 Revert "ci: Fix macOS upload script calling wrong macpack."
This reverts commit 3b9ed5234d.
2022-12-30 18:08:20 +02:00
850ec1f8b8 vk_common: Enable beta extensions
* Required to access the portability subset'
2022-12-30 16:22:53 +02:00
3da6c25fd8 renderer_vulkan: Fallback to software shaders on unsupported topology
* MoltenVK does not support triangle fans
2022-12-30 15:25:27 +02:00
0e987959a6 renderer_vulkan: Rewrite data streaming
* Most GPUs nowadays provide a device local/host visible memory heap which is useful for avoiding copies between staging and local memory and especially beneficial for mobile and APUs that are mostly the target of this backend.

* This commit ports the old yuzu stream buffer with some changes to suit our needs and gets rid of the buffer flush methods
2022-12-30 11:10:49 +02:00
410b8b8809 vk_texture_runtime: Tune barriers
* Using eAllCommandBit is really bad for Mali GPUs. Also most access flags were redundant mostly for edge cases.

* To address this track surface usage and decide the best barrier flags at runtime. This gives a significant performance boost to mobile GPUs
2022-12-29 21:56:57 +02:00
d3392ae0b1 renderer_vulkan: Properly format structs 2022-12-29 20:11:57 +02:00
98e0ecf6a7 renderer_vulkan: Add fallback path for VK_EXT_index_type_uint8
* Also remove some flush barriers
2022-12-29 19:07:09 +02:00
ad45b9880d android: Add vulkan support to frontend 2022-12-28 14:01:50 +02:00
0d1646e4df Revert "vk_scheduler: Enable usage of jthread on macos"
This reverts commit 09dcd48257.
2022-12-26 17:15:21 +02:00
96f0746ab9 HACK: Skip normquat lerp to drop geometry shader requirement 2022-12-26 15:50:11 +02:00
09dcd48257 vk_scheduler: Enable usage of jthread on macos
- Import the yuzu polyfill libraries
2022-12-26 11:17:23 +02:00
62fc1f835e ci: Provide glslang 2022-12-26 09:01:05 +02:00
3b351c33d1 android: Fix build 2022-12-26 08:53:30 +02:00
793485d201 renderer_vulkan: Revert some stream buffer changes
* The previous design was much less prone to errors so switch back to that. Also make 16 byte alignment standard
2022-12-25 23:48:11 +02:00
3ef5ab7323 video_core: Move pixel format functions to cpp file 2022-12-25 23:37:28 +02:00
618c80c803 vk_instance: Address small issues 2022-12-25 23:21:44 +02:00
f7cb308243 video_core/host_shaders: Add CMake integration for string shaders 2022-12-25 23:17:39 +02:00
b8583f9af3 renderer_vulkan: Port per-game to vulkan renderer 2022-12-25 23:06:30 +02:00
33bf2b7c2d renderer_vulkan: Forward validation errors to logfile 2022-12-25 22:31:27 +02:00
d48e6c04ce video_core: Move most pica register handling to RasterizerAccelerated 2022-12-25 22:31:27 +02:00
88f34a7d69 vk_shader_gen_spv: Implement proctex sampler
* Fixes MHS menu and probably other games
2022-12-25 22:31:27 +02:00
c8e9b465e2 renderer_vulkan: Proper telemetry reporting 2022-12-25 22:31:27 +02:00
278198f5f5 vk_shader_gen_spv: Implement shadow plane sampling 2022-12-25 22:31:27 +02:00
58718e6bd6 renderer_vulkan: Fix LMDM crashes
* Cache vertex format lookups to avoid crashing the driver when the game does too many lookups

* Increase watcher count. For some reason the game uses way too many watchers

* Fix other misc SPIRV bugs and don't crash when shadow is requested
2022-12-25 22:31:27 +02:00
a814b21693 vk_shader_gen_spv: Implement alpha testing
* Should fix all current graphical bugs
2022-12-25 22:31:27 +02:00
5478a4d634 vk_instance: Make DynamicLoader static
* That way we don't construct/destroy it all the time
2022-12-25 22:31:27 +02:00
8f87586495 citra_qt: Add SPIR-V shader option 2022-12-25 22:31:27 +02:00
7e3a0f524c externals: Update sirit 2022-12-25 22:29:05 +02:00
922019cc22 renderer_vulkan: Move fragment shader microprofile to a better place 2022-12-25 22:29:05 +02:00
41e9cdb645 shader_cache: Fix type deduction 2022-12-25 22:29:05 +02:00
e90add52d2 cmake: Fork sirit
* Upstream is missing some required instructions
2022-12-25 22:29:04 +02:00
3c6ca2cc82 renderer_vulkan: Begin new fragment shader SPIR-V emitter 2022-12-25 22:29:04 +02:00
d1039d9a81 code: Address build issues 2022-12-25 22:29:04 +02:00
8b8cee1a5a vk_instance: Remove depth clip control feature
* To not crash drivers that don't support it since we don't require it anymore
2022-12-25 22:28:49 +02:00
5dc92cd72b ci: Bundle MoltenVK with macOS builds. 2022-12-25 22:28:49 +02:00
d1503605a7 vulkan: Fix supported extension check 2022-12-25 22:28:49 +02:00
6426f7a319 vulkan: Use required portability instance extension on macOS. 2022-12-25 22:28:49 +02:00
3b9ed5234d ci: Fix macOS upload script calling wrong macpack. 2022-12-25 22:28:49 +02:00
763127605e qt: Extract CAMetalLayer from NSView to pass to MoltenVK. 2022-12-25 22:28:49 +02:00
b86b19d366 renderer_vulkan: Drop requirement for VK_EXT_depth_clip_control 2022-12-25 22:28:49 +02:00
3dd74c69c5 Revert "CI: dont upload macos artifacts (#6121)"
This reverts commit 30831e6367.
2022-12-25 22:28:49 +02:00
01af8e3f2c renderer_vulkan: Integrate MacOS wsi 2022-12-25 22:28:49 +02:00
474cccda33 video_core: Fix build issues on macos 2022-12-25 22:28:49 +02:00
6057b18172 renderer_vulkan: Emulate 3-component vertex formats when unsupported
* This fixes the crashes on AMD
2022-12-25 22:28:48 +02:00
6a4ff8fa24 renderer_vulkan: Emulate logic op when unsupported
* Similar to GLES this is done to prepare for the android port
2022-12-25 22:28:48 +02:00
3c79360fd3 gl_rasterizer: Cleanup and fix bugs 2022-12-25 22:28:48 +02:00
939aafed40 vk_texture_runtime: Implement RGBA4 converter
* Fixes graphics in NES Remix. Need to also do a reinterpreter some time, but this will suffice for now
2022-12-25 22:28:48 +02:00
23417787f8 texture_downloader_es: Remove invalid operations 2022-12-25 22:28:48 +02:00
8946c1a7de gl_texture_runtime: Use OGLStreamBuffer for uploads/downloads
* Much better than the current implementation
2022-12-25 22:28:48 +02:00
5fe910b18f vk_stream_buffer: Cleanup flush barrier 2022-12-25 22:28:48 +02:00
3944cbdc19 video_core: Reorder microprofile defines 2022-12-25 22:28:48 +02:00
8ac3dd1840 renderer_vulkan: Rewrite stream buffer, again...
* The previous implemention was fine, but it wasted space. Buckets now are just ticks attached to a particular buffer region, which means we can flush/map arbitrary regions

* A bug in the texture runtime is also fixed which commited to the same buffer twice
2022-12-25 22:28:15 +02:00
ff34287e4b renderer_vulkan: Pack PicaFSConfig
* Using bitfields the struct size was reduced from 420 to 190 bytes. which should speed up hashing and copying to the worker thread
2022-12-25 22:28:15 +02:00
81f2a0eaa1 renderer_vulkan: Cleanup vertex array setup
* Also the function would commit more data then it requested leading to out of bound crashes
2022-12-25 22:28:15 +02:00
f11715a4f4 renderer_vulkan: Remove AttribType
* Use VertexAttributeFormat to avoid unnecessary enum casts
2022-12-25 22:28:15 +02:00
89c51371f7 video_core: Move HardwareVertex to RasterizerAccelerated 2022-12-25 22:28:15 +02:00
8076d893db video_core: Move api agnostic uniform updates to RasterizerAccelerated 2022-12-25 22:28:15 +02:00
72f8d520c9 renderer_vulkan: Fix swapchain resizing 2022-12-25 22:28:15 +02:00
f9274f8b9a renderer_vulkan: Add single-thread record ability to the scheduler
* Async is pretty nice but games that do a lot of flushes might have worse performance due to thread synchronization overhead

* I haven't noticed any cases of this yet but it doesn't hurt making this a UI option
2022-12-25 22:28:15 +02:00
3c09c03180 citra_qt: Refuse to enable debug option if the layers are not available 2022-12-25 22:24:28 +02:00
52251e3908 renderer_vulkan: Scheduler and presentation rewrite
* This commit ports yuzu's async scheduler replacing our older and crummier version
  Commands are recorded by the scheduler and processed by a separate worker thread.

* Queue submission is also moved to the worker thread which should alliviate slowdowns related to vkQueueSubmit stalls

* Fragment shader compilation and queue submission are also moved to that thread to reduce stutters
2022-12-25 22:23:46 +02:00
921444c2c9 externals: Update vulkan-headers 2022-12-25 22:23:46 +02:00
89d234f642 common: Remove concepts usage 2022-12-25 22:23:46 +02:00
13771b805b renderer_vulkan: Fix shader hash type 2022-12-25 22:23:46 +02:00
2a71059490 code: Remove usages of std::ranges
* MacOS is still runining my C++ 20 fun
2022-12-25 22:23:46 +02:00
053221f155 renderer_vulkan: Prefer immediate over mailbox present mode 2022-12-25 22:23:46 +02:00
4868c361e7 renderer_vulkan: Bump vertex buffer size
* So software shaders don't crash
2022-12-25 22:23:46 +02:00
0c30dbf33e renderer_vulkan: Add more microprofile targets 2022-12-25 22:23:46 +02:00
53370e81e2 renderer_vulkan: Improve StreamBuffer API and use it in TextureRuntime
* Also use separate upload and download buffers optimized for write and readback respectively. This gives a huge 20+ FPS boost in most games which were bottlenecked by slow reads
2022-12-25 22:23:46 +02:00
b3fb260c84 renderer_vulkan: Fix allocation caching bug 2022-12-25 22:23:46 +02:00
599ca7caf7 renderer_opengl: Port scaled upload/download code from vulkan 2022-12-25 22:23:46 +02:00
abc0fd5e7b renderer_vulkan: Include algorithm in vk_common
* Appears to be a bug in vulkan-hpp
2022-12-25 22:23:46 +02:00
309b25d201 renderer_vulkan: Use linear filtering when possible
* Fixes blocky artifacts in Samus Returns
2022-12-25 22:23:46 +02:00
9ac7ef20b0 renderer_vulkan: Abstract descriptor management
* The pipeline cache was starting to get cluttered
2022-12-25 22:23:46 +02:00
ebade3594d renderer_vulkan: Bump descriptor set allocation limit 2022-12-25 22:23:46 +02:00
dca159d79f renderer_vulkan: Fix storage descriptor binding and respect color mask
* RGBA8 surfaces now expose an additional R32Uint view used for storage descriptors. The format is guaranteed by the spec to support atomic loads/stores. This requires the mutable flag which incurs a performance cost, but might be better than breaking the current renderpass each draw when rendering shadows, especially on mobile

* Color mask is also implemented which fixes Street Fighter and Monster Hunter Stories
2022-12-25 22:23:46 +02:00
7007d5822a renderer_vulkan: Implement depth uploads with blit 2022-12-25 22:23:46 +02:00
6f0fdf037f renderer_vulkan: Use intermediate copy when framebuffer is used both as attachment and shader input 2022-12-25 22:23:46 +02:00
58621b0eb6 renderer_vulkan: Respect disk shader option 2022-12-25 22:23:46 +02:00
e8eef5c586 renderer_vulkan: Fix staging buffer size 2022-12-25 22:23:46 +02:00
2b37997a95 renderer_vulkan: Catch and log more runtime errors
* Also add the ability to enable command buffer dumping which is very useful
2022-12-25 22:23:15 +02:00
558062efd7 renderer_vulkan: Batch allocate descriptor sets
* Less driver calls should lead to better performance
2022-12-25 22:20:22 +02:00
5e880a4f26 renderer_vulkan: Emulate border color if possible 2022-12-25 22:20:22 +02:00
238956f773 renderer_vulkan: Implement scaled uploads and downloads
* This commit includes large changes to have textures are handling. Instead of using ImageAlloc, Surface is used instead which provides multiple benefits: automatic recycling on destruction and ability to use the TextureRuntime interface to simplify operations

* Layout tracking is also implemented which allows transitioning of individual mip levels without errors

* This fixes graphical errors in multiple games which relied on framebuffer uploads
2022-12-25 22:20:22 +02:00
a5351bc596 renderer_vulkan: Fix renderpass issues
* The cache didn't take into account the framebuffer and render area used, so if these changed the renderpass wouldn't restart. This caused graphical bugs in Pokemon X/Y
2022-12-25 22:20:22 +02:00
5b6e99b194 renderer_vulkan: Update stencil compare mask 2022-12-25 22:20:22 +02:00
3f3b4a2802 citra_qt: Fix graphics api indicator alignment 2022-12-25 22:20:22 +02:00
c19e8d36c9 renderer_opengl: Fix OpenGLES issues
* Always request a 4.4 context until I figure out how to get Qt to cooperate

* Use RGBA for BGR since the converted table will do that conversion
2022-12-25 22:20:22 +02:00
2ec31f404b renderer_vulkan: Report perf stats 2022-12-25 22:20:21 +02:00
28a2805450 renderer_vulkan: Better error handling 2022-12-25 22:20:21 +02:00
2601a1df6c renderer_vulkan: Allow direct allocation of images 2022-12-25 22:19:26 +02:00
96d5bb553b renderer_vulkan: Fix incorrect depth format detection
* Intel iGPUs don't support blit on all depth/stencil formats which caused issues since the runtime checks for this while the renderpass cache does not
2022-12-25 22:19:26 +02:00
b4d0f442c8 renderer_vulkan: Actually minize state changes
* Keep track of the current state and only update it when needed. Previously games would set the same state over and over cluttering renderdoc logs
2022-12-25 22:19:25 +02:00
2a68cab7d6 renderer_vulkan: Fix broken sync without timeline semaphores 2022-12-25 22:19:25 +02:00
009d73fdf6 renderer_vulkan: Allocate descriptor sets during reinterpretation 2022-12-25 22:19:25 +02:00
42af22f8fd renderer_vulkan: Enable logic ops and fix swapchain resizing 2022-12-25 22:19:25 +02:00
b85d15b035 renderer_vulkan: Clear stencil with renderpass
* Fixes outline retension in pokemon games
2022-12-25 22:19:25 +02:00
130e376c0c renderer_vulkan: Fix pipeline cache crashes 2022-12-25 22:19:25 +02:00
f884986257 renderer_vulkan: Optimize tiled format convertion + fix vertex buffer alignment
* Integrate format convertion to the morton copy function, removing the need for an intermediate copy and convertion pass. This should be beneficial for performance especially since most games use tiled textures

* Also bump vertex buffer size to avoid crashes with hardware shaders and provide correct offset on normal draws which fixes glitches in pokemon Y

* Reduce the local group size to 8 in the D24S8 compute shader which fixes graphical issues in the afformentioned pokemon games at native resolution

* Set LOD to 0 instead of 0.25 to fix another glitch in pokemon y
2022-12-25 22:19:25 +02:00
e23dc3efb1 renderer_opengl: Fix broken texture copy
* Resolves graphical bugs in Professor Layton vs Ace Attorney when using OpenGL
2022-12-25 22:19:25 +02:00
dd5e95d7c6 renderer_vulkan: Pipeline cache fixes
* Delete cache file if found invalid

* Name it after the vendor/device ids so each physical devices gets a separate cache
2022-12-25 22:19:25 +02:00
b72289eadd video_core: Fix renderpass cache bug and introduce RGBA -> BGR converter 2022-12-25 22:19:25 +02:00
7a5d4f03da renderer_opengl: Specify precision in compute shader and add RGB5A1 converter
* Fixes OpenGLES crash
2022-12-25 22:19:25 +02:00
18fa277c71 renderer_vulkan: Complete hardware shader support
* With these changes all commercial games I tested work fine and get a massive performance boost
2022-12-25 22:19:25 +02:00
73cc764091 renderer_vulkan: Begin hardware shader support
* Still experimental and works only with homebrew
2022-12-25 22:19:25 +02:00
9dea514d45 citra: Fix build issues with MinGW and MSVC 2022-12-25 22:19:25 +02:00
0bfaa035b9 renderer_vulkan: Fix warnings and cleanup 2022-12-25 22:19:06 +02:00
85df778785 code: Run clang-format 2022-12-25 22:19:06 +02:00
e54a92c252 code: Address build issues 2022-12-25 22:18:04 +02:00
9145d4cec8 video_core: Re-implement format reinterpretation
* Same as before but D24S8 to RGBA8 is switched to a compute shader which should provide better throughput and is much simpler to implement in Vulkan
2022-12-25 22:18:04 +02:00
c611592db6 citra_qt: Add physical device selection dialog 2022-12-25 22:18:04 +02:00
6aba809da8 renderer_opengl: Unbind unused framebuffer targets
* Fixes graphical glitches in many games for some reason
2022-12-25 22:12:42 +02:00
f0449d79fd renderer_opengl: Emulate texture copy with blit for now 2022-12-25 22:12:42 +02:00
33481ada7f renderer_opengl: Address buffer overflow 2022-12-25 22:12:42 +02:00
67195974e7 video_core: Small code improvements 2022-12-25 22:12:42 +02:00
70c2376fd0 renderer_vulkan: Don't sample from mipmaps when using texture cubes
* Mipmaps for texture cubes are unimplemented in the rasterizer cache, so sampling from mipmaps will return nothing
2022-12-25 22:12:42 +02:00
177c7de4f9 input_common: Small fix 2022-12-25 22:12:42 +02:00
eea914ba84 citra_qt: Improve graphics API intergration
* Add renderer debug option which toggles debug output in OpenGL/validation layers in Vulkan

* Fix many warnings and replace deprecated Qt functionailty with newer alternatives
2022-12-25 22:12:41 +02:00
62e88fbeb3 rasterizer_cache: Code cleanup
* Merge utils and types to a single header
2022-12-25 22:07:46 +02:00
5ce27d8341 texture_decode: Prefer std::memcpy where possible 2022-12-25 22:07:46 +02:00
eaf62eb635 renderer_vulkan: Rework format handling
* This is a pretty large commit that aims to solve some issues with the current format system
* The instance now builds at application initialization an array of format traits for each pixel format
  that includes information such as blit/attachment/storage support and fallback formats
* The runtime doesn't ask the instance for formats but receives these traits and can dedice on its own what to build
  For now we do the same as before, we require both blit and attachment support

* Morton swizzling also sees many bug fixes. The previous code was very hacky and didn't work for partial
  texture updates. It was also inconsistent, as it would take a tiled_buffer and write to the middle of linear
* Now the functions have been greatly simplified and adjusted to work better with std::span. This fixes out of bounds
  errors and texture glitches (like the display in Mario Kart 7)
2022-12-25 22:07:46 +02:00
9675811bbe renderer_vulkan: Add experimental Vulkan renderer 2022-12-25 22:07:46 +02:00
945faf8e92 externals: Add vulkan headers and vma 2022-12-25 22:02:59 +02:00
9403049671 rasterizer_cache: Refactor texture cube interface
* Reuse our Surface class instead of having a separate one, to avoid reimplementing stuff in the backend
2022-12-25 22:02:59 +02:00
40159d9779 gl_texture_runtime: Clean up texture upload/download code
* Improve readability and code clarity
2022-12-25 22:02:59 +02:00
f63653a5b9 rasterizer_cache: Use Common::Rectangle everywhere
* Make a nice alias for it and use it instead of having Rect2D/Region2D. Makes the new design less intrusive to the current cache
2022-12-25 22:02:59 +02:00
c71dbb5d19 rasterizer_cache: Make into template
* This is the final step, now RasterizerCache is compltely decoupled from OpenGL (technically not yet, but that's talking details). For now texture filtering and some GLES paths have been disabled and will be reimplemented in the following commits
2022-12-25 22:02:59 +02:00
0f4df2c012 rasterizer_cache: Use PBO staging buffer cache for texture uploads/downloads 2022-12-25 22:02:11 +02:00
c6fc4f5a87 rasterizer_cache: Reorder methods 2022-12-25 22:02:11 +02:00
916afa194d rasterizer_cache: Remove remnants of cached_pages 2022-12-25 22:02:11 +02:00
6f2cd11a85 rasterizer_cache: Fix texture cube blitting
* The target was GL_TEXTURE_2D instead of GL_TEXTURE_CUBE_MAP_*
2022-12-25 22:02:11 +02:00
14652d52bc morton_swizzle: Implement texture formats in UNSWIZZLE_TABLE
* I can now remove that loop that has been messing with my OCD
2022-12-25 22:02:11 +02:00
a57ee7cdf2 morton_swizzle: Use tiled_buffer instead of reading data from g_memory
* It's much safer and removes hardcoded global state usage
2022-12-25 22:02:11 +02:00
dbd3e6c29b rasterizer_accelerated: Zero intialize cached_pages
* Resolves random crashes because count takes random values
2022-12-25 22:02:11 +02:00
665cbca17c texture_runtime: Add staging buffer lock mechanism 2022-12-25 22:02:11 +02:00
efb9e9f40f cached_surface: Remove custom texture logic
* Makes things more complicated and is in the way. It's probably already
broken by recent changes, so I'll need to reimplement it anyway
2022-12-25 22:02:11 +02:00
8d35118f63 renderer_opengl: Add driver class to report info/bugs 2022-12-25 22:02:11 +02:00
553c85456e rasterizer_cache: Add staging buffer cache for uploads/downloads
*  In addition bump context version to 4.4 to enforce ARB_buffer_storage and use EXT_buffer_storage for GLES which is support on many mobile devices
2022-12-25 22:02:11 +02:00
68ca206d53 rasterizer_cache: Improve TextureRuntime API
* This makes every operation more explicit and mimics more the Vulkan API
2022-12-25 22:02:11 +02:00
e30e977140 renderer_opengl: Encapsulate sync objects in OGLSync 2022-12-25 22:02:10 +02:00
f13738d252 morton_swizzle: Optimize and use std::span 2022-12-25 22:02:10 +02:00
04b927ab7f morton_swizzle: Avoid buffer underflow
* Check the y coordinate before decrementing linear_buffer
2022-12-25 22:02:10 +02:00
993d172de9 morton_swizzle: Move out of bounds texture check out of the decode loop
* Running relative expensive checks like this on a hot path causes small but measurable performance loss. Tested SMD wit this and it doesn't crash
2022-12-25 22:02:10 +02:00
695447611e rasterizer_cache: Use SurfaceType instead of Aspect
* It was doing pointless enum conversions when both enums described the same thing
2022-12-25 22:02:10 +02:00
06bacfbd72 rasterizer_cache: Separate texture swizzling to utils 2022-12-25 22:02:10 +02:00
cf8bc35d46 rasterizer_cache: Remove OpenGL references from morton_swizzle 2022-12-25 22:02:10 +02:00
ef859bab84 citra_qt: Forbid renderer change during runtime
* It's an endless source of problems and isn't usefull
2022-12-25 22:02:10 +02:00
a2d0669562 rasterizer_cache: Touch up MatchFlags comments 2022-12-25 22:02:10 +02:00
95365ad6ba rasterizer_cache: Drop OpenGL postfix 2022-12-25 22:02:10 +02:00
1963b649e8 rasterizer_cache: Shorten filenames and general cleanup
* AllocateSurfaceTexture now takes the PixelFormat directly as FormatTuple is an OpenGL struct and will be moved there
2022-12-25 22:02:10 +02:00
db7cdb741c video_core: Move UpdatePagesCachedCount to RasterizerAccelerated 2022-12-25 22:02:10 +02:00
0fe61ba040 citra_qt: Prepare GUI for Vulkan support 2022-12-25 22:02:06 +02:00
195 changed files with 41314 additions and 7572 deletions

View File

@ -3,7 +3,7 @@
brew update
brew unlink python@2 || true
rm '/usr/local/bin/2to3' || true
brew install qt5 p7zip ccache ninja || true
brew install qt5 molten-vk glslang vulkan-loader p7zip ccache ninja || true
pip3 install macpack
export SDL_VER=2.0.16

View File

@ -12,20 +12,37 @@ cp build/bin/Release/citra "$REV_NAME"
cp -r build/bin/Release/citra-qt.app "$REV_NAME"
cp build/bin/Release/citra-room "$REV_NAME"
# move libs into folder for deployment
macpack "${REV_NAME}/citra-qt.app/Contents/MacOS/citra-qt" -d "../Frameworks"
# move qt frameworks into app bundle for deployment
$(brew --prefix)/opt/qt5/bin/macdeployqt "${REV_NAME}/citra-qt.app" -executable="${REV_NAME}/citra-qt.app/Contents/MacOS/citra-qt"
BUNDLE_PATH="$REV_NAME/citra-qt.app"
BUNDLE_CONTENTS_PATH="$BUNDLE_PATH/Contents"
BUNDLE_EXECUTABLE_PATH="$BUNDLE_CONTENTS_PATH/MacOS/citra-qt"
BUNDLE_LIB_PATH="$BUNDLE_CONTENTS_PATH/lib"
BUNDLE_RESOURCES_PATH="$BUNDLE_CONTENTS_PATH/Resources"
CITRA_STANDALONE_PATH="$REV_NAME/citra"
# move libs into folder for deployment
macpack "${REV_NAME}/citra" -d "libs"
macpack $BUNDLE_EXECUTABLE_PATH -d "../Frameworks"
# move qt frameworks into app bundle for deployment
$(brew --prefix)/opt/qt5/bin/macdeployqt $BUNDLE_PATH -executable=$BUNDLE_EXECUTABLE_PATH
# move libs into folder for deployment
macpack $CITRA_STANDALONE_PATH -d "libs"
# bundle MoltenVK
VULKAN_LOADER_PATH=$(brew --prefix vulkan-loader)
MOLTENVK_PATH=$(brew --prefix molten-vk)
mkdir $BUNDLE_LIB_PATH
cp $VULKAN_LOADER_PATH/lib/libvulkan.dylib $BUNDLE_LIB_PATH
cp $MOLTENVK_PATH/lib/libMoltenVK.dylib $BUNDLE_LIB_PATH
cp -r $MOLTENVK_PATH/share/vulkan $BUNDLE_RESOURCES_PATH
install_name_tool -add_rpath "@loader_path/../lib/" $BUNDLE_EXECUTABLE_PATH
# workaround for libc++
install_name_tool -change @loader_path/../Frameworks/libc++.1.0.dylib /usr/lib/libc++.1.dylib "${REV_NAME}/citra-qt.app/Contents/MacOS/citra-qt"
install_name_tool -change @loader_path/libs/libc++.1.0.dylib /usr/lib/libc++.1.dylib "${REV_NAME}/citra"
install_name_tool -change @loader_path/../Frameworks/libc++.1.0.dylib /usr/lib/libc++.1.dylib $BUNDLE_EXECUTABLE_PATH
install_name_tool -change @loader_path/libs/libc++.1.0.dylib /usr/lib/libc++.1.dylib $CITRA_STANDALONE_PATH
# Make the launching script executable
chmod +x ${REV_NAME}/citra-qt.app/Contents/MacOS/citra-qt
chmod +x $BUNDLE_EXECUTABLE_PATH
# Verify loader instructions
find "$REV_NAME" -type f -exec otool -L {} \;

View File

@ -95,6 +95,13 @@ jobs:
env:
MACOSX_DEPLOYMENT_TARGET: "10.13"
ENABLE_COMPATIBILITY_REPORTING: "ON"
- name: Pack
run: ./.ci/macos/upload.sh
- name: Upload
uses: actions/upload-artifact@v3
with:
name: macos
path: artifacts/
windows:
runs-on: windows-latest
steps:
@ -116,6 +123,14 @@ jobs:
shell: bash
- name: Set up MSVC
uses: ilammy/msvc-dev-cmd@v1
- name: Setup Vulkan SDK
uses: humbletim/setup-vulkan-sdk@v1.2.0
with:
vulkan-query-version: latest
vulkan-components: Glslang
vulkan-use-cache: true
- name: Test glslangValidator
run: glslangValidator --version
- name: Build
run: ./.ci/windows-msvc/build.sh
shell: bash
@ -150,7 +165,7 @@ jobs:
- name: Deps
run: |
sudo apt-get update
sudo apt-get install ccache apksigner -y
sudo apt-get install glslang-tools ccache apksigner -y
- name: Build
run: ./.ci/android/build.sh
- name: Copy and sign artifacts

12
.gitmodules vendored
View File

@ -58,3 +58,15 @@
[submodule "sdl2"]
path = externals/sdl2/SDL
url = https://github.com/libsdl-org/SDL
[submodule "vulkan-headers"]
path = externals/vulkan-headers
url = https://github.com/KhronosGroup/Vulkan-Headers
[submodule "glslang"]
path = externals/glslang
url = https://github.com/KhronosGroup/glslang
[submodule "glm"]
path = externals/glm
url = https://github.com/g-truc/glm
[submodule "sirit"]
path = externals/sirit
url = https://github.com/GPUCode/sirit

View File

@ -9,6 +9,7 @@ cmake_policy(SET CMP0069 NEW)
list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/CMakeModules")
list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/externals/cmake-modules")
include(DownloadExternals)
include(GNUInstallDirs)
include(CMakeDependentOption)
project(citra LANGUAGES C CXX ASM)

View File

@ -13,6 +13,6 @@
<file alias="256x256/plus_folder.png">icons/256x256/plus_folder.png</file>
</qresource>
<qresource prefix="colorful">
<file>style.qss</file>
<file alias="style.qss">../default/style.qss</file>
</qresource>
</RCC>

View File

@ -1,4 +0,0 @@
/*
This file is intentionally left blank.
We do not want to apply any stylesheet for colorful, only icons.
*/

View File

@ -1,33 +1,22 @@
<RCC>
<qresource prefix="icons/default">
<file alias="index.theme">icons/index.theme</file>
<file alias="16x16/checked.png">icons/16x16/checked.png</file>
<file alias="16x16/failed.png">icons/16x16/failed.png</file>
<file alias="16x16/connected.png">icons/16x16/connected.png</file>
<file alias="16x16/disconnected.png">icons/16x16/disconnected.png</file>
<file alias="16x16/connected_notification.png">icons/16x16/connected_notification.png</file>
<file alias="16x16/lock.png">icons/16x16/lock.png</file>
<file alias="48x48/bad_folder.png">icons/48x48/bad_folder.png</file>
<file alias="48x48/chip.png">icons/48x48/chip.png</file>
<file alias="48x48/folder.png">icons/48x48/folder.png</file>
<file alias="48x48/no_avatar.png">icons/48x48/no_avatar.png</file>
<file alias="48x48/plus.png">icons/48x48/plus.png</file>
<file alias="48x48/sd_card.png">icons/48x48/sd_card.png</file>
<file alias="256x256/citra.png">icons/256x256/citra.png</file>
<file alias="256x256/plus_folder.png">icons/256x256/plus_folder.png</file>
</qresource>
<qresource prefix="default">
<file>style.qss</file>
</qresource>
</RCC>

13
dist/qt_themes/default/style.qss vendored Normal file
View File

@ -0,0 +1,13 @@
QPushButton#GraphicsAPIStatusBarButton {
color: #656565;
border: 1px solid transparent;
background-color: transparent;
padding: 0px 3px 0px 3px;
text-align: center;
min-width: 60px;
min-height: 20px;
}
QPushButton#GraphicsAPIStatusBarButton:hover {
border: 1px solid #76797C;
}

View File

@ -522,13 +522,12 @@ QToolButton#qt_toolbar_ext_button {
QPushButton {
color: #eff0f1;
border-width: 1px;
border-color: #54575B;
border-style: solid;
padding: 6px 4px;
border: 1px solid #54575B;
border-radius: 2px;
padding: 5px 0px 5px 0px;
outline: none;
min-width: 100px;
min-height: 13px;
background-color: #232629;
}
@ -1237,3 +1236,17 @@ QPlainTextEdit:disabled {
TouchScreenPreview {
qproperty-dotHighlightColor: #3daee9;
}
QPushButton#GraphicsAPIStatusBarButton {
color: #656565;
border: 1px solid transparent;
background-color: transparent;
padding: 0px 3px 0px 3px;
text-align: center;
min-width: 60px;
min-height: 20px;
}
QPushButton#GraphicsAPIStatusBarButton:hover {
border: 1px solid #76797C;
}

View File

@ -60,6 +60,19 @@ endif()
# Glad
add_subdirectory(glad)
# glslang
set(SKIP_GLSLANG_INSTALL ON)
set(ENABLE_GLSLANG_BINARIES OFF)
set(ENABLE_SPVREMAPPER OFF)
set(ENABLE_CTEST OFF)
add_subdirectory(glslang)
# Sirit
add_subdirectory(sirit)
# glm
add_subdirectory(glm)
# inih
add_subdirectory(inih)
@ -154,3 +167,12 @@ if(ANDROID)
add_subdirectory(libyuv)
target_include_directories(yuv INTERFACE ./libyuv/include)
endif()
# VMA
add_library(vma INTERFACE)
target_include_directories(vma INTERFACE ./vma)
# vulkan-headers
add_library(vulkan-headers INTERFACE)
target_include_directories(vulkan-headers INTERFACE ./vulkan-headers/include)

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff

1
externals/glm vendored Submodule

Submodule externals/glm added at cc98465e35

1
externals/glslang vendored Submodule

Submodule externals/glslang added at c0cf8ad876

View File

@ -39,9 +39,9 @@ set(SDL_JOYSTICK ON CACHE BOOL "")
set(SDL_HAPTIC OFF CACHE BOOL "")
set(SDL_HIDAPI ON CACHE BOOL "")
set(SDL_POWER OFF CACHE BOOL "")
set(SDL_THREADS ON CACHE BOOL "")
set(SDL_TIMERS ON CACHE BOOL "")
set(SDL_FILE ON CACHE BOOL "")
set(SDL_THREADS ON CACHE BOOL "")
set(SDL_LOADSO ON CACHE BOOL "")
set(SDL_CPUINFO OFF CACHE BOOL "")
set(SDL_FILESYSTEM OFF CACHE BOOL "")

1
externals/sirit vendored Submodule

Submodule externals/sirit added at f0b6bbe55b

19670
externals/vma/vk_mem_alloc.h vendored Normal file

File diff suppressed because it is too large Load Diff

1
externals/vulkan-headers vendored Submodule

View File

@ -122,6 +122,7 @@ else()
if (MINGW)
add_definitions(-DMINGW_HAS_SECURE_API)
add_compile_options("-Wa,-mbig-obj")
if (COMPILE_WITH_DWARF)
add_compile_options("-gdwarf")
endif()

View File

@ -30,7 +30,8 @@
android:supportsRtl="true"
android:isGame="true"
android:banner="@mipmap/ic_launcher"
android:requestLegacyExternalStorage="true">
android:requestLegacyExternalStorage="true"
android:debuggable="true">
<activity
android:name="org.citra.citra_emu.ui.main.MainActivity"

View File

@ -19,6 +19,8 @@ add_library(citra-android SHARED
default_ini.h
emu_window/emu_window.cpp
emu_window/emu_window.h
emu_window/emu_window_vk.cpp
emu_window/emu_window_vk.h
game_info.cpp
game_info.h
game_settings.cpp

View File

@ -114,7 +114,13 @@ void Config::ReadValues() {
sdl2_config->GetString("Premium", "texture_filter_name", "none");
// Renderer
Settings::values.use_gles = sdl2_config->GetBoolean("Renderer", "use_gles", true);
Settings::values.graphics_api =
static_cast<Settings::GraphicsAPI>(sdl2_config->GetInteger("Renderer", "graphics_api", 2));
Settings::values.async_command_recording =
sdl2_config->GetBoolean("Renderer", "async_command_recording", true);
Settings::values.spirv_shader_gen =
sdl2_config->GetBoolean("Renderer", "spirv_shader_gen", true);
Settings::values.renderer_debug = sdl2_config->GetBoolean("Renderer", "renderer_debug", false);
Settings::values.use_hw_renderer = sdl2_config->GetBoolean("Renderer", "use_hw_renderer", true);
Settings::values.use_hw_shader = sdl2_config->GetBoolean("Renderer", "use_hw_shader", true);
Settings::values.shaders_accurate_mul =

View File

@ -0,0 +1,176 @@
// Copyright 2019 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <algorithm>
#include <array>
#include <cstdlib>
#include <string>
#include <android/native_window_jni.h>
#include "common/logging/log.h"
#include "common/settings.h"
#include "input_common/main.h"
#include "jni/emu_window/emu_window_vk.h"
#include "jni/id_cache.h"
#include "jni/input_manager.h"
#include "network/network.h"
#include "video_core/renderer_base.h"
#include "video_core/video_core.h"
static bool IsPortraitMode() {
return JNI_FALSE != IDCache::GetEnvForThread()->CallStaticBooleanMethod(
IDCache::GetNativeLibraryClass(), IDCache::GetIsPortraitMode());
}
static void UpdateLandscapeScreenLayout() {
Settings::values.layout_option =
static_cast<Settings::LayoutOption>(IDCache::GetEnvForThread()->CallStaticIntMethod(
IDCache::GetNativeLibraryClass(), IDCache::GetLandscapeScreenLayout()));
}
void EmuWindow_Android_Vulkan::OnSurfaceChanged(ANativeWindow* surface) {
render_window = surface;
StopPresenting();
}
bool EmuWindow_Android_Vulkan::OnTouchEvent(int x, int y, bool pressed) {
if (pressed) {
return TouchPressed((unsigned)std::max(x, 0), (unsigned)std::max(y, 0));
}
TouchReleased();
return true;
}
void EmuWindow_Android_Vulkan::OnTouchMoved(int x, int y) {
TouchMoved((unsigned)std::max(x, 0), (unsigned)std::max(y, 0));
}
void EmuWindow_Android_Vulkan::OnFramebufferSizeChanged() {
UpdateLandscapeScreenLayout();
const bool is_portrait_mode{IsPortraitMode()};
const int bigger{window_width > window_height ? window_width : window_height};
const int smaller{window_width < window_height ? window_width : window_height};
if (is_portrait_mode) {
UpdateCurrentFramebufferLayout(smaller, bigger, is_portrait_mode);
} else {
UpdateCurrentFramebufferLayout(bigger, smaller, is_portrait_mode);
}
}
EmuWindow_Android_Vulkan::EmuWindow_Android_Vulkan(ANativeWindow* surface) {
LOG_DEBUG(Frontend, "Initializing EmuWindow_Android_Vulkan");
if (!surface) {
LOG_CRITICAL(Frontend, "surface is nullptr");
return;
}
Network::Init();
host_window = surface;
CreateWindowSurface();
if (core_context = CreateSharedContext(); !core_context) {
LOG_CRITICAL(Frontend, "CreateSharedContext() failed");
return;
}
OnFramebufferSizeChanged();
}
bool EmuWindow_Android_Vulkan::CreateWindowSurface() {
if (!host_window) {
return true;
}
window_info.type = Frontend::WindowSystemType::Android;
window_info.render_surface = host_window;
return true;
}
void EmuWindow_Android_Vulkan::DestroyWindowSurface() {
/*if (!egl_surface) {
return;
}
if (eglGetCurrentSurface(EGL_DRAW) == egl_surface) {
eglMakeCurrent(egl_display, EGL_NO_SURFACE, EGL_NO_SURFACE, EGL_NO_CONTEXT);
}
if (!eglDestroySurface(egl_display, egl_surface)) {
LOG_CRITICAL(Frontend, "eglDestroySurface() failed");
}
egl_surface = EGL_NO_SURFACE;*/
}
void EmuWindow_Android_Vulkan::DestroyContext() {
/*if (!egl_context) {
return;
}
if (eglGetCurrentContext() == egl_context) {
eglMakeCurrent(egl_display, EGL_NO_SURFACE, EGL_NO_SURFACE, EGL_NO_CONTEXT);
}
if (!eglDestroyContext(egl_display, egl_context)) {
LOG_CRITICAL(Frontend, "eglDestroySurface() failed");
}
if (!eglTerminate(egl_display)) {
LOG_CRITICAL(Frontend, "eglTerminate() failed");
}
egl_context = EGL_NO_CONTEXT;
egl_display = EGL_NO_DISPLAY;*/
}
EmuWindow_Android_Vulkan::~EmuWindow_Android_Vulkan() {
DestroyWindowSurface();
DestroyContext();
}
std::unique_ptr<Frontend::GraphicsContext> EmuWindow_Android_Vulkan::CreateSharedContext() const {
return std::make_unique<SharedContext_Android>();
}
void EmuWindow_Android_Vulkan::StopPresenting() {
/*if (presenting_state == PresentingState::Running) {
eglMakeCurrent(egl_display, EGL_NO_SURFACE, EGL_NO_SURFACE, EGL_NO_CONTEXT);
}*/
presenting_state = PresentingState::Stopped;
}
void EmuWindow_Android_Vulkan::TryPresenting() {
if (presenting_state != PresentingState::Running) {
if (presenting_state == PresentingState::Initial) {
/*eglMakeCurrent(egl_display, egl_surface, egl_surface, egl_context);
glBindFramebuffer(GL_DRAW_FRAMEBUFFER, 0);*/
presenting_state = PresentingState::Running;
} else {
return;
}
}
/*eglSwapInterval(egl_display, Settings::values.use_vsync_new ? 1 : 0);
if (VideoCore::g_renderer) {
VideoCore::g_renderer->TryPresent(0);
eglSwapBuffers(egl_display, egl_surface);
}*/
}
void EmuWindow_Android_Vulkan::PollEvents() {
if (!render_window) {
return;
}
host_window = render_window;
render_window = nullptr;
DestroyWindowSurface();
CreateWindowSurface();
OnFramebufferSizeChanged();
presenting_state = PresentingState::Initial;
}
void EmuWindow_Android_Vulkan::MakeCurrent() {
core_context->MakeCurrent();
}
void EmuWindow_Android_Vulkan::DoneCurrent() {
core_context->DoneCurrent();
}

View File

@ -0,0 +1,59 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <vector>
#include "core/frontend/emu_window.h"
struct ANativeWindow;
class SharedContext_Android : public Frontend::GraphicsContext {};
class EmuWindow_Android_Vulkan : public Frontend::EmuWindow {
public:
EmuWindow_Android_Vulkan(ANativeWindow* surface);
~EmuWindow_Android_Vulkan();
void Present();
/// Called by the onSurfaceChanges() method to change the surface
void OnSurfaceChanged(ANativeWindow* surface);
/// Handles touch event that occur.(Touched or released)
bool OnTouchEvent(int x, int y, bool pressed);
/// Handles movement of touch pointer
void OnTouchMoved(int x, int y);
void PollEvents() override;
void MakeCurrent() override;
void DoneCurrent() override;
void TryPresenting();
void StopPresenting();
std::unique_ptr<GraphicsContext> CreateSharedContext() const override;
private:
void OnFramebufferSizeChanged();
bool CreateWindowSurface();
void DestroyWindowSurface();
void DestroyContext();
ANativeWindow* render_window{};
ANativeWindow* host_window{};
int window_width{1080};
int window_height{2220};
std::unique_ptr<Frontend::GraphicsContext> core_context;
enum class PresentingState {
Initial,
Running,
Stopped,
};
PresentingState presenting_state{};
};

View File

@ -23,7 +23,6 @@
#include "core/frontend/applets/default_applets.h"
#include "core/frontend/camera/factory.h"
#include "core/frontend/mic.h"
#include "core/frontend/scope_acquire_context.h"
#include "core/hle/service/am/am.h"
#include "core/hle/service/nfc/nfc.h"
#include "core/savestate.h"
@ -33,7 +32,7 @@
#include "jni/camera/ndk_camera.h"
#include "jni/camera/still_image_camera.h"
#include "jni/config.h"
#include "jni/emu_window/emu_window.h"
#include "jni/emu_window/emu_window_vk.h"
#include "jni/game_info.h"
#include "jni/game_settings.h"
#include "jni/id_cache.h"
@ -49,7 +48,7 @@ namespace {
ANativeWindow* s_surf;
std::unique_ptr<EmuWindow_Android> window;
std::unique_ptr<EmuWindow_Android_Vulkan> window;
std::atomic<bool> stop_run{true};
std::atomic<bool> pause_emulation{false};
@ -147,7 +146,7 @@ static Core::System::ResultStatus RunCitra(const std::string& filepath) {
return Core::System::ResultStatus::ErrorLoader;
}
window = std::make_unique<EmuWindow_Android>(s_surf);
window = std::make_unique<EmuWindow_Android_Vulkan>(s_surf);
Core::System& system{Core::System::GetInstance()};

View File

@ -7,7 +7,7 @@ buildscript {
jcenter()
}
dependencies {
classpath 'com.android.tools.build:gradle:7.2.0'
classpath 'com.android.tools.build:gradle:7.3.1'
// NOTE: Do not place your application dependencies here; they belong
// in the individual module build.gradle files

View File

@ -7,10 +7,7 @@
#include <regex>
#include <string>
#include <thread>
// This needs to be included before getopt.h because the latter #defines symbols used by it
#include "common/microprofile.h"
#include "citra/config.h"
#include "citra/emu_window/emu_window_sdl2.h"
#include "citra/lodepng_image_interface.h"
@ -18,22 +15,18 @@
#include "common/detached_tasks.h"
#include "common/file_util.h"
#include "common/logging/backend.h"
#include "common/logging/filter.h"
#include "common/logging/log.h"
#include "common/microprofile.h"
#include "common/scm_rev.h"
#include "common/scope_exit.h"
#include "common/settings.h"
#include "common/string_util.h"
#include "core/core.h"
#include "core/dumping/backend.h"
#include "core/file_sys/cia_container.h"
#include "core/frontend/applets/default_applets.h"
#include "core/frontend/framebuffer_layout.h"
#include "core/frontend/scope_acquire_context.h"
#include "core/gdbstub/gdbstub.h"
#include "core/hle/service/am/am.h"
#include "core/hle/service/cfg/cfg.h"
#include "core/loader/loader.h"
#include "core/movie.h"
#include "input_common/main.h"
#include "network/network.h"
@ -368,7 +361,7 @@ int main(int argc, char** argv) {
const auto secondary_window =
use_secondary_window ? std::make_unique<EmuWindow_SDL2>(false, true) : nullptr;
Frontend::ScopeAcquireContext scope(*emu_window);
const auto scope = emu_window->Acquire();
LOG_INFO(Frontend, "Citra Version: {} | {}-{}", Common::g_build_fullname, Common::g_scm_branch,
Common::g_scm_desc);

View File

@ -109,7 +109,8 @@ void Config::ReadValues() {
sdl2_config->GetInteger("Core", "cpu_clock_percentage", 100);
// Renderer
Settings::values.use_gles = sdl2_config->GetBoolean("Renderer", "use_gles", false);
Settings::values.graphics_api =
static_cast<Settings::GraphicsAPI>(sdl2_config->GetInteger("Renderer", "graphics_api", 0));
Settings::values.use_hw_renderer = sdl2_config->GetBoolean("Renderer", "use_hw_renderer", true);
Settings::values.use_hw_shader = sdl2_config->GetBoolean("Renderer", "use_hw_shader", true);
#ifdef __APPLE__

View File

@ -137,7 +137,9 @@ void EmuWindow_SDL2::Fullscreen() {
EmuWindow_SDL2::EmuWindow_SDL2(bool fullscreen, bool is_secondary) : EmuWindow(is_secondary) {
// Initialize the window
if (Settings::values.use_gles) {
const bool is_opengles =
Settings::values.graphics_api.GetValue() == Settings::GraphicsAPI::OpenGLES;
if (is_opengles) {
SDL_GL_SetAttribute(SDL_GL_CONTEXT_MAJOR_VERSION, 3);
SDL_GL_SetAttribute(SDL_GL_CONTEXT_MINOR_VERSION, 2);
SDL_GL_SetAttribute(SDL_GL_CONTEXT_PROFILE_MASK, SDL_GL_CONTEXT_PROFILE_ES);
@ -192,7 +194,7 @@ EmuWindow_SDL2::EmuWindow_SDL2(bool fullscreen, bool is_secondary) : EmuWindow(i
}
render_window_id = SDL_GetWindowID(render_window);
auto gl_load_func = Settings::values.use_gles ? gladLoadGLES2Loader : gladLoadGLLoader;
auto gl_load_func = is_opengles ? gladLoadGLES2Loader : gladLoadGLLoader;
if (!gl_load_func(static_cast<GLADloadproc>(SDL_GL_GetProcAddress))) {
LOG_CRITICAL(Frontend, "Failed to initialize GL functions: {}", SDL_GetError());

View File

@ -250,6 +250,8 @@ if (APPLE)
set_target_properties(citra-qt PROPERTIES MACOSX_BUNDLE TRUE)
set_target_properties(citra-qt PROPERTIES MACOSX_BUNDLE_INFO_PLIST ${CMAKE_CURRENT_SOURCE_DIR}/Info.plist)
target_sources(citra-qt PRIVATE
applesurfacehelper.h
applesurfacehelper.mm
macos_authorization.h
macos_authorization.mm
)
@ -266,9 +268,13 @@ endif()
create_target_directory_groups(citra-qt)
target_link_libraries(citra-qt PRIVATE audio_core common core input_common network video_core)
target_link_libraries(citra-qt PRIVATE Boost::boost glad nihstro-headers Qt5::Widgets Qt5::Multimedia)
target_link_libraries(citra-qt PRIVATE Boost::boost glad vma vulkan-headers nihstro-headers Qt5::Widgets Qt5::Multimedia)
target_link_libraries(citra-qt PRIVATE ${PLATFORM_LIBRARIES} Threads::Threads)
if (NOT WIN32)
target_include_directories(citra-qt PRIVATE ${Qt5Gui_PRIVATE_INCLUDE_DIRS})
endif()
target_compile_definitions(citra-qt PRIVATE
# Use QStringBuilder for string concatenation to reduce
# the overall number of temporary strings created.

View File

@ -0,0 +1,11 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
namespace AppleSurfaceHelper {
void* GetSurfaceLayer(void* surface);
} // namespace AppleSurfaceHelper

View File

@ -0,0 +1,16 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#import <Cocoa/Cocoa.h>
#include "citra_qt/applesurfacehelper.h"
namespace AppleSurfaceHelper {
void* GetSurfaceLayer(void* surface) {
NSView* view = static_cast<NSView*>(surface);
return view.layer;
}
} // AppleSurfaceHelper

View File

@ -6,10 +6,10 @@
#include <QDragEnterEvent>
#include <QHBoxLayout>
#include <QKeyEvent>
#include <QMessageBox>
#include <QOffscreenSurface>
#include <QOpenGLContext>
#include <QOpenGLFunctions>
#include <QOpenGLFunctions_4_3_Core>
#include <QOpenGLExtraFunctions>
#include <fmt/format.h>
#include "citra_qt/bootmanager.h"
#include "citra_qt/main.h"
@ -18,15 +18,20 @@
#include "common/settings.h"
#include "core/3ds.h"
#include "core/core.h"
#include "core/frontend/scope_acquire_context.h"
#include "core/perf_stats.h"
#include "input_common/keyboard.h"
#include "input_common/main.h"
#include "input_common/motion_emu.h"
#include "network/network.h"
#include "video_core/renderer_base.h"
#include "video_core/video_core.h"
#if defined(__APPLE__)
#include "citra_qt/applesurfacehelper.h"
#endif
#if !defined(WIN32)
#include <qpa/qplatformnativeinterface.h>
#endif
EmuThread::EmuThread(Frontend::GraphicsContext& core_context) : core_context(core_context) {}
EmuThread::~EmuThread() = default;
@ -44,7 +49,7 @@ static GMainWindow* GetMainWindow() {
void EmuThread::run() {
MicroProfileOnThreadCreate("EmuThread");
Frontend::ScopeAcquireContext scope(core_context);
const auto scope = core_context.Acquire();
emit LoadProgress(VideoCore::LoadCallbackStage::Prepare, 0, 0);
@ -55,6 +60,7 @@ void EmuThread::run() {
});
emit LoadProgress(VideoCore::LoadCallbackStage::Complete, 0, 0);
emit HideLoadingScreen();
core_context.MakeCurrent();
@ -113,88 +119,242 @@ void EmuThread::run() {
#endif
}
OpenGLWindow::OpenGLWindow(QWindow* parent, QWidget* event_handler, QOpenGLContext* shared_context,
bool is_secondary)
: QWindow(parent), context(std::make_unique<QOpenGLContext>(shared_context->parent())),
event_handler(event_handler), is_secondary{is_secondary} {
class OpenGLSharedContext : public Frontend::GraphicsContext {
public:
/// Create the original context that should be shared from
explicit OpenGLSharedContext(QSurface* surface) : surface(surface) {
QSurfaceFormat format;
// disable vsync for any shared contexts
auto format = shared_context->format();
format.setSwapInterval(Settings::values.use_vsync_new ? 1 : 0);
this->setFormat(format);
format.setVersion(4, 4);
format.setProfile(QSurfaceFormat::CoreProfile);
context->setShareContext(shared_context);
context->setScreen(this->screen());
context->setFormat(format);
context->create();
if (Settings::values.renderer_debug) {
format.setOption(QSurfaceFormat::FormatOption::DebugContext);
}
setSurfaceType(QWindow::OpenGLSurface);
// TODO: expose a setting for buffer value (ie default/single/double/triple)
format.setSwapBehavior(QSurfaceFormat::DefaultSwapBehavior);
format.setSwapInterval(0);
// TODO: One of these flags might be interesting: WA_OpaquePaintEvent, WA_NoBackground,
// WA_DontShowOnScreen, WA_DeleteOnClose
}
OpenGLWindow::~OpenGLWindow() {
context->doneCurrent();
}
void OpenGLWindow::Present() {
if (!isExposed())
return;
context->makeCurrent(this);
if (VideoCore::g_renderer) {
VideoCore::g_renderer->TryPresent(100, is_secondary);
context = std::make_unique<QOpenGLContext>();
context->setFormat(format);
if (!context->create()) {
LOG_ERROR(Frontend, "Unable to create main openGL context");
}
}
context->swapBuffers(this);
auto f = context->versionFunctions<QOpenGLFunctions_4_3_Core>();
f->glFinish();
QWindow::requestUpdate();
}
bool OpenGLWindow::event(QEvent* event) {
switch (event->type()) {
case QEvent::UpdateRequest:
/// Create the shared contexts for rendering and presentation
explicit OpenGLSharedContext(QOpenGLContext* share_context, QSurface* main_surface = nullptr) {
// disable vsync for any shared contexts
auto format = share_context->format();
format.setSwapInterval(main_surface ? Settings::values.use_vsync_new.GetValue() : 0);
context = std::make_unique<QOpenGLContext>();
context->setShareContext(share_context);
context->setFormat(format);
if (!context->create()) {
LOG_ERROR(Frontend, "Unable to create shared openGL context");
}
if (!main_surface) {
offscreen_surface = std::make_unique<QOffscreenSurface>(nullptr);
offscreen_surface->setFormat(format);
offscreen_surface->create();
surface = offscreen_surface.get();
} else {
surface = main_surface;
}
}
~OpenGLSharedContext() {
context->doneCurrent();
}
void SwapBuffers() override {
context->swapBuffers(surface);
}
void MakeCurrent() override {
// We can't track the current state of the underlying context in this wrapper class because
// Qt may make the underlying context not current for one reason or another. In particular,
// the WebBrowser uses GL, so it seems to conflict if we aren't careful.
// Instead of always just making the context current (which does not have any caching to
// check if the underlying context is already current) we can check for the current context
// in the thread local data by calling `currentContext()` and checking if its ours.
if (QOpenGLContext::currentContext() != context.get()) {
context->makeCurrent(surface);
}
}
void DoneCurrent() override {
context->doneCurrent();
}
QOpenGLContext* GetShareContext() const {
return context.get();
}
private:
// Avoid using Qt parent system here since we might move the QObjects to new threads
// As a note, this means we should avoid using slots/signals with the objects too
std::unique_ptr<QOpenGLContext> context;
std::unique_ptr<QOffscreenSurface> offscreen_surface{};
QSurface* surface;
};
class DummyContext : public Frontend::GraphicsContext {};
class RenderWidget : public QWidget {
public:
RenderWidget(GRenderWindow* parent) : QWidget(parent), render_window(parent) {
setAttribute(Qt::WA_NativeWindow);
setAttribute(Qt::WA_PaintOnScreen);
}
virtual ~RenderWidget() = default;
virtual void Present() {}
void paintEvent(QPaintEvent* event) override {
Present();
return true;
case QEvent::MouseButtonPress:
case QEvent::MouseButtonRelease:
case QEvent::MouseButtonDblClick:
case QEvent::MouseMove:
case QEvent::KeyPress:
case QEvent::KeyRelease:
case QEvent::FocusIn:
case QEvent::FocusOut:
case QEvent::FocusAboutToChange:
case QEvent::Enter:
case QEvent::Leave:
case QEvent::Wheel:
case QEvent::TabletMove:
case QEvent::TabletPress:
case QEvent::TabletRelease:
case QEvent::TabletEnterProximity:
case QEvent::TabletLeaveProximity:
case QEvent::TouchBegin:
case QEvent::TouchUpdate:
case QEvent::TouchEnd:
case QEvent::InputMethodQuery:
case QEvent::TouchCancel:
return QCoreApplication::sendEvent(event_handler, event);
case QEvent::Drop:
GetMainWindow()->DropAction(static_cast<QDropEvent*>(event));
return true;
case QEvent::DragEnter:
case QEvent::DragMove:
GetMainWindow()->AcceptDropEvent(static_cast<QDropEvent*>(event));
return true;
default:
return QWindow::event(event);
update();
}
void resizeEvent(QResizeEvent* ev) override {
render_window->resize(ev->size());
render_window->OnFramebufferSizeChanged();
}
void keyPressEvent(QKeyEvent* event) override {
InputCommon::GetKeyboard()->PressKey(event->key());
}
void keyReleaseEvent(QKeyEvent* event) override {
InputCommon::GetKeyboard()->ReleaseKey(event->key());
}
void mousePressEvent(QMouseEvent* event) override {
if (event->source() == Qt::MouseEventSynthesizedBySystem)
return; // touch input is handled in TouchBeginEvent
const auto pos{event->pos()};
if (event->button() == Qt::LeftButton) {
const auto [x, y] = render_window->ScaleTouch(pos);
render_window->TouchPressed(x, y);
} else if (event->button() == Qt::RightButton) {
InputCommon::GetMotionEmu()->BeginTilt(pos.x(), pos.y());
}
}
void mouseMoveEvent(QMouseEvent* event) override {
if (event->source() == Qt::MouseEventSynthesizedBySystem)
return; // touch input is handled in TouchUpdateEvent
const auto pos{event->pos()};
const auto [x, y] = render_window->ScaleTouch(pos);
render_window->TouchMoved(x, y);
InputCommon::GetMotionEmu()->Tilt(pos.x(), pos.y());
}
void mouseReleaseEvent(QMouseEvent* event) override {
if (event->source() == Qt::MouseEventSynthesizedBySystem)
return; // touch input is handled in TouchEndEvent
if (event->button() == Qt::LeftButton)
render_window->TouchReleased();
else if (event->button() == Qt::RightButton)
InputCommon::GetMotionEmu()->EndTilt();
}
std::pair<unsigned, unsigned> GetSize() const {
return std::make_pair(width(), height());
}
QPaintEngine* paintEngine() const override {
return nullptr;
}
private:
GRenderWindow* render_window;
};
class OpenGLRenderWidget : public RenderWidget {
public:
explicit OpenGLRenderWidget(GRenderWindow* parent, bool is_secondary)
: RenderWidget(parent), is_secondary(is_secondary) {
windowHandle()->setSurfaceType(QWindow::OpenGLSurface);
}
void SetContext(std::unique_ptr<OpenGLSharedContext>&& context_) {
context = std::move(context_);
}
void Present() override {
if (!isVisible()) {
return;
}
if (!Core::System::GetInstance().IsPoweredOn()) {
return;
}
context->MakeCurrent();
const auto f = context->GetShareContext()->extraFunctions();
f->glBindFramebuffer(GL_DRAW_FRAMEBUFFER, 0);
VideoCore::g_renderer->TryPresent(100, is_secondary);
context->SwapBuffers();
f->glFinish();
}
private:
std::unique_ptr<OpenGLSharedContext> context{};
bool is_secondary;
};
class VulkanRenderWidget : public RenderWidget {
public:
explicit VulkanRenderWidget(GRenderWindow* parent) : RenderWidget(parent) {
windowHandle()->setSurfaceType(QWindow::VulkanSurface);
}
};
static Frontend::WindowSystemType GetWindowSystemType() {
// Determine WSI type based on Qt platform.
QString platform_name = QGuiApplication::platformName();
if (platform_name == QStringLiteral("windows"))
return Frontend::WindowSystemType::Windows;
else if (platform_name == QStringLiteral("xcb"))
return Frontend::WindowSystemType::X11;
else if (platform_name == QStringLiteral("wayland"))
return Frontend::WindowSystemType::Wayland;
else if (platform_name == QStringLiteral("cocoa"))
return Frontend::WindowSystemType::MacOS;
LOG_CRITICAL(Frontend, "Unknown Qt platform!");
return Frontend::WindowSystemType::Windows;
}
void OpenGLWindow::exposeEvent(QExposeEvent* event) {
QWindow::requestUpdate();
QWindow::exposeEvent(event);
static Frontend::EmuWindow::WindowSystemInfo GetWindowSystemInfo(QWindow* window) {
Frontend::EmuWindow::WindowSystemInfo wsi;
wsi.type = GetWindowSystemType();
// Our Win32 Qt external doesn't have the private API.
#if defined(WIN32)
wsi.render_surface = window ? reinterpret_cast<void*>(window->winId()) : nullptr;
#elif defined(__APPLE__)
wsi.render_surface =
window ? AppleSurfaceHelper::GetSurfaceLayer(reinterpret_cast<void*>(window->winId()))
: nullptr;
#else
QPlatformNativeInterface* pni = QGuiApplication::platformNativeInterface();
wsi.display_connection = pni->nativeResourceForWindow("display", window);
if (wsi.type == Frontend::WindowSystemType::Wayland)
wsi.render_surface = window ? pni->nativeResourceForWindow("surface", window) : nullptr;
else
wsi.render_surface = window ? reinterpret_cast<void*>(window->winId()) : nullptr;
#endif
wsi.render_surface_scale = window ? static_cast<float>(window->devicePixelRatio()) : 1.0f;
return wsi;
}
GRenderWindow::GRenderWindow(QWidget* parent_, EmuThread* emu_thread, bool is_secondary_)
@ -218,11 +378,11 @@ GRenderWindow::GRenderWindow(QWidget* parent_, EmuThread* emu_thread, bool is_se
GRenderWindow::~GRenderWindow() = default;
void GRenderWindow::MakeCurrent() {
core_context->MakeCurrent();
main_context->MakeCurrent();
}
void GRenderWindow::DoneCurrent() {
core_context->DoneCurrent();
main_context->DoneCurrent();
}
void GRenderWindow::PollEvents() {
@ -393,34 +553,70 @@ void GRenderWindow::resizeEvent(QResizeEvent* event) {
OnFramebufferSizeChanged();
}
void GRenderWindow::InitRenderTarget() {
std::unique_ptr<Frontend::GraphicsContext> GRenderWindow::CreateSharedContext() const {
const Settings::GraphicsAPI graphics_api = Settings::values.graphics_api.GetValue();
if (graphics_api == Settings::GraphicsAPI::OpenGL ||
graphics_api == Settings::GraphicsAPI::OpenGLES) {
auto c = static_cast<OpenGLSharedContext*>(main_context.get());
// Bind the shared contexts to the main surface in case the backend wants to take over
// presentation
return std::make_unique<OpenGLSharedContext>(c->GetShareContext(),
child_widget->windowHandle());
}
return std::make_unique<DummyContext>();
}
bool GRenderWindow::InitRenderTarget() {
ReleaseRenderTarget();
{
// Create a dummy render widget so that Qt
// places the render window at the correct position.
const RenderWidget dummy_widget{this};
}
first_frame = false;
GMainWindow* parent = GetMainWindow();
QWindow* parent_win_handle = parent ? parent->windowHandle() : nullptr;
child_window = new OpenGLWindow(parent_win_handle, this, QOpenGLContext::globalShareContext(),
is_secondary);
child_window->create();
child_widget = createWindowContainer(child_window, this);
const Settings::GraphicsAPI graphics_api = Settings::values.graphics_api.GetValue();
switch (graphics_api) {
case Settings::GraphicsAPI::OpenGL:
case Settings::GraphicsAPI::OpenGLES:
if (!InitializeOpenGL()) {
return false;
}
break;
case Settings::GraphicsAPI::Vulkan:
if (!InitializeVulkan()) {
return false;
}
break;
}
// Update the Window System information with the new render target
window_info = GetWindowSystemInfo(child_widget->windowHandle());
child_widget->resize(Core::kScreenTopWidth, Core::kScreenTopHeight + Core::kScreenBottomHeight);
layout()->addWidget(child_widget);
// Reset minimum required size to avoid resizing issues on the main window after restarting.
setMinimumSize(1, 1);
core_context = CreateSharedContext();
resize(Core::kScreenTopWidth, Core::kScreenTopHeight + Core::kScreenBottomHeight);
OnMinimalClientAreaChangeRequest(GetActiveConfig().min_client_area_size);
OnFramebufferSizeChanged();
BackupGeometry();
return true;
}
void GRenderWindow::ReleaseRenderTarget() {
if (child_widget) {
layout()->removeWidget(child_widget);
delete child_widget;
child_widget->deleteLater();
child_widget = nullptr;
}
main_context.reset();
}
void GRenderWindow::CaptureScreenshot(u32 res_scale, const QString& screenshot_path) {
@ -445,6 +641,29 @@ void GRenderWindow::OnMinimalClientAreaChangeRequest(std::pair<u32, u32> minimal
setMinimumSize(minimal_size.first, minimal_size.second);
}
bool GRenderWindow::InitializeOpenGL() {
// TODO: One of these flags might be interesting: WA_OpaquePaintEvent, WA_NoBackground,
// WA_DontShowOnScreen, WA_DeleteOnClose
auto child = new OpenGLRenderWidget(this, is_secondary);
child_widget = child;
child_widget->windowHandle()->create();
auto context = std::make_shared<OpenGLSharedContext>(child->windowHandle());
main_context = context;
child->SetContext(
std::make_unique<OpenGLSharedContext>(context->GetShareContext(), child->windowHandle()));
return true;
}
bool GRenderWindow::InitializeVulkan() {
auto child = new VulkanRenderWidget(this);
child_widget = child;
child_widget->windowHandle()->create();
main_context = std::make_unique<DummyContext>();
return true;
}
void GRenderWindow::OnEmulationStarting(EmuThread* emu_thread) {
this->emu_thread = emu_thread;
}
@ -456,31 +675,3 @@ void GRenderWindow::OnEmulationStopping() {
void GRenderWindow::showEvent(QShowEvent* event) {
QWidget::showEvent(event);
}
std::unique_ptr<Frontend::GraphicsContext> GRenderWindow::CreateSharedContext() const {
return std::make_unique<GLContext>(QOpenGLContext::globalShareContext());
}
GLContext::GLContext(QOpenGLContext* shared_context)
: context(std::make_unique<QOpenGLContext>(shared_context->parent())),
surface(std::make_unique<QOffscreenSurface>(nullptr)) {
// disable vsync for any shared contexts
auto format = shared_context->format();
format.setSwapInterval(0);
context->setShareContext(shared_context);
context->setFormat(format);
context->create();
surface->setParent(shared_context->parent());
surface->setFormat(format);
surface->create();
}
void GLContext::MakeCurrent() {
context->makeCurrent(surface.get());
}
void GLContext::DoneCurrent() {
context->doneCurrent();
}

View File

@ -27,19 +27,6 @@ namespace VideoCore {
enum class LoadCallbackStage;
}
class GLContext : public Frontend::GraphicsContext {
public:
explicit GLContext(QOpenGLContext* shared_context);
void MakeCurrent() override;
void DoneCurrent() override;
private:
std::unique_ptr<QOpenGLContext> context;
std::unique_ptr<QOffscreenSurface> surface;
};
class EmuThread final : public QThread {
Q_OBJECT
@ -126,26 +113,6 @@ signals:
void HideLoadingScreen();
};
class OpenGLWindow : public QWindow {
Q_OBJECT
public:
explicit OpenGLWindow(QWindow* parent, QWidget* event_handler, QOpenGLContext* shared_context,
bool is_secondary = false);
~OpenGLWindow();
void Present();
protected:
bool event(QEvent* event) override;
void exposeEvent(QExposeEvent* event) override;
private:
std::unique_ptr<QOpenGLContext> context;
QWidget* event_handler;
bool is_secondary;
};
class GRenderWindow : public QWidget, public Frontend::EmuWindow {
Q_OBJECT
@ -185,13 +152,15 @@ public:
return has_focus;
}
void InitRenderTarget();
bool InitRenderTarget();
/// Destroy the previous run's child_widget which should also destroy the child_window
void ReleaseRenderTarget();
void CaptureScreenshot(u32 res_scale, const QString& screenshot_path);
std::pair<u32, u32> ScaleTouch(const QPointF pos) const;
public slots:
void OnEmulationStarting(EmuThread* emu_thread);
@ -211,29 +180,29 @@ signals:
void MouseActivity();
private:
std::pair<u32, u32> ScaleTouch(QPointF pos) const;
void TouchBeginEvent(const QTouchEvent* event);
void TouchUpdateEvent(const QTouchEvent* event);
void TouchEndEvent();
void OnMinimalClientAreaChangeRequest(std::pair<u32, u32> minimal_size) override;
std::unique_ptr<GraphicsContext> core_context;
QByteArray geometry;
/// Native window handle that backs this presentation widget
QWindow* child_window = nullptr;
/// In order to embed the window into GRenderWindow, you need to use createWindowContainer to
/// put the child_window into a widget then add it to the layout. This child_widget can be
/// parented to GRenderWindow and use Qt's lifetime system
QWidget* child_widget = nullptr;
bool InitializeOpenGL();
bool InitializeVulkan();
EmuThread* emu_thread;
// Main context that will be shared with all other contexts that are requested.
// If this is used in a shared context setting, then this should not be used directly, but
// should instead be shared from
std::shared_ptr<Frontend::GraphicsContext> main_context;
/// Temporary storage of the screenshot taken
QImage screenshot_image;
QByteArray geometry;
QWidget* child_widget = nullptr;
bool first_frame = false;
bool has_focus = false;

View File

@ -455,6 +455,8 @@ void Config::ReadDebuggingValues() {
qt_config->value(QStringLiteral("record_frame_times"), false).toBool();
ReadBasicSetting(Settings::values.use_gdbstub);
ReadBasicSetting(Settings::values.gdbstub_port);
ReadBasicSetting(Settings::values.renderer_debug);
ReadBasicSetting(Settings::values.dump_command_buffers);
qt_config->beginGroup(QStringLiteral("LLE"));
for (const auto& service_module : Service::service_module_map) {
@ -601,6 +603,10 @@ void Config::ReadPathValues() {
void Config::ReadRendererValues() {
qt_config->beginGroup(QStringLiteral("Renderer"));
ReadGlobalSetting(Settings::values.physical_device);
ReadGlobalSetting(Settings::values.async_command_recording);
ReadGlobalSetting(Settings::values.spirv_shader_gen);
ReadGlobalSetting(Settings::values.graphics_api);
ReadGlobalSetting(Settings::values.use_hw_renderer);
ReadGlobalSetting(Settings::values.use_hw_shader);
#ifdef __APPLE__
@ -965,6 +971,8 @@ void Config::SaveDebuggingValues() {
qt_config->setValue(QStringLiteral("record_frame_times"), Settings::values.record_frame_times);
WriteBasicSetting(Settings::values.use_gdbstub);
WriteBasicSetting(Settings::values.gdbstub_port);
WriteBasicSetting(Settings::values.renderer_debug);
WriteBasicSetting(Settings::values.dump_command_buffers);
qt_config->beginGroup(QStringLiteral("LLE"));
for (const auto& service_module : Settings::values.lle_modules) {
@ -1077,6 +1085,10 @@ void Config::SavePathValues() {
void Config::SaveRendererValues() {
qt_config->beginGroup(QStringLiteral("Renderer"));
WriteGlobalSetting(Settings::values.graphics_api);
WriteGlobalSetting(Settings::values.physical_device);
WriteGlobalSetting(Settings::values.async_command_recording);
WriteGlobalSetting(Settings::values.spirv_shader_gen);
WriteGlobalSetting(Settings::values.use_hw_renderer);
WriteGlobalSetting(Settings::values.use_hw_shader);
#ifdef __APPLE__

View File

@ -3,6 +3,7 @@
// Refer to the license.txt file included.
#include <QDesktopServices>
#include <QMessageBox>
#include <QUrl>
#include "citra_qt/configuration/configure_debug.h"
#include "citra_qt/debugger/console.h"
@ -11,7 +12,9 @@
#include "common/logging/log.h"
#include "common/settings.h"
#include "core/core.h"
#include "qcheckbox.h"
#include "ui_configure_debug.h"
#include "video_core/renderer_vulkan/vk_instance.h"
ConfigureDebug::ConfigureDebug(QWidget* parent)
: QWidget(parent), ui(std::make_unique<Ui::ConfigureDebug>()) {
@ -23,8 +26,40 @@ ConfigureDebug::ConfigureDebug(QWidget* parent)
QDesktopServices::openUrl(QUrl::fromLocalFile(path));
});
connect(ui->toggle_renderer_debug, &QCheckBox::clicked, this, [this](bool checked) {
if (checked && Settings::values.graphics_api.GetValue() == Settings::GraphicsAPI::Vulkan) {
try {
Vulkan::Instance debug_inst{true};
} catch (vk::LayerNotPresentError&) {
ui->toggle_renderer_debug->toggle();
QMessageBox::warning(this, tr("Validation layer not available"),
tr("Unable to enable debug renderer because the layer "
"<strong>VK_LAYER_KHRONOS_validation</strong> is missing. "
"Please install the Vulkan SDK or the appropriate package "
"of your distribution"));
}
}
});
connect(ui->toggle_dump_command_buffers, &QCheckBox::clicked, this, [this](bool checked) {
if (checked && Settings::values.graphics_api.GetValue() == Settings::GraphicsAPI::Vulkan) {
try {
Vulkan::Instance debug_inst{false, true};
} catch (vk::LayerNotPresentError&) {
ui->toggle_dump_command_buffers->toggle();
QMessageBox::warning(this, tr("Command buffer dumping not available"),
tr("Unable to enable command buffer dumping because the layer "
"<strong>VK_LAYER_LUNARG_api_dump</strong> is missing. "
"Please install the Vulkan SDK or the appropriate package "
"of your distribution"));
}
}
});
const bool is_powered_on = Core::System::GetInstance().IsPoweredOn();
ui->toggle_cpu_jit->setEnabled(!is_powered_on);
ui->toggle_renderer_debug->setEnabled(!is_powered_on);
ui->toggle_dump_command_buffers->setEnabled(!is_powered_on);
}
ConfigureDebug::~ConfigureDebug() = default;
@ -37,6 +72,8 @@ void ConfigureDebug::SetConfiguration() {
ui->toggle_console->setChecked(UISettings::values.show_console.GetValue());
ui->log_filter_edit->setText(QString::fromStdString(Settings::values.log_filter.GetValue()));
ui->toggle_cpu_jit->setChecked(Settings::values.use_cpu_jit.GetValue());
ui->toggle_renderer_debug->setChecked(Settings::values.renderer_debug.GetValue());
ui->toggle_dump_command_buffers->setChecked(Settings::values.dump_command_buffers.GetValue());
}
void ConfigureDebug::ApplyConfiguration() {
@ -49,6 +86,8 @@ void ConfigureDebug::ApplyConfiguration() {
filter.ParseFilterString(Settings::values.log_filter.GetValue());
Log::SetGlobalFilter(filter);
Settings::values.use_cpu_jit = ui->toggle_cpu_jit->isChecked();
Settings::values.renderer_debug = ui->toggle_renderer_debug->isChecked();
Settings::values.dump_command_buffers = ui->toggle_dump_command_buffers->isChecked();
}
void ConfigureDebug::RetranslateUI() {

View File

@ -22,5 +22,6 @@ public:
void RetranslateUI();
void SetConfiguration();
private:
std::unique_ptr<Ui::ConfigureDebug> ui;
};

View File

@ -6,8 +6,8 @@
<rect>
<x>0</x>
<y>0</y>
<width>443</width>
<height>300</height>
<width>454</width>
<height>356</height>
</rect>
</property>
<property name="windowTitle">
@ -122,6 +122,23 @@
</property>
</widget>
</item>
<item>
<widget class="QCheckBox" name="toggle_renderer_debug">
<property name="toolTip">
<string>&lt;html&gt;&lt;head/&gt;&lt;body&gt;&lt;p&gt;Enables debug reporting in the currently selected graphics API. Causes measurable performance loss, don't enable unless for debugging purposes&lt;/p&gt;&lt;/body&gt;&lt;/html&gt;</string>
</property>
<property name="text">
<string>Enable debug renderer</string>
</property>
</widget>
</item>
<item>
<widget class="QCheckBox" name="toggle_dump_command_buffers">
<property name="text">
<string>Dump command buffers</string>
</property>
</widget>
</item>
</layout>
</widget>
</item>

View File

@ -11,17 +11,29 @@
#include "common/settings.h"
#include "core/core.h"
#include "ui_configure_graphics.h"
#include "video_core/renderer_vulkan/vk_instance.h"
ConfigureGraphics::ConfigureGraphics(QWidget* parent)
: QWidget(parent), ui(std::make_unique<Ui::ConfigureGraphics>()) {
ui->setupUi(this);
DiscoverPhysicalDevices();
SetupPerGameUI();
SetConfiguration();
ui->hw_renderer_group->setEnabled(ui->hw_renderer_group->isEnabled() &&
ui->toggle_hw_renderer->isChecked());
ui->toggle_vsync_new->setEnabled(!Core::System::GetInstance().IsPoweredOn());
const bool not_running = !Core::System::GetInstance().IsPoweredOn();
const bool hw_renderer_enabled = ui->toggle_hw_renderer->isChecked();
ui->toggle_hw_renderer->setEnabled(not_running);
ui->hw_renderer_group->setEnabled(hw_renderer_enabled && not_running);
ui->toggle_vsync_new->setEnabled(not_running);
ui->graphics_api_combo->setEnabled(not_running);
ui->toggle_shader_jit->setEnabled(not_running);
ui->toggle_disk_shader_cache->setEnabled(hw_renderer_enabled && not_running);
ui->physical_device_combo->setEnabled(not_running);
SetPhysicalDeviceComboVisibility(ui->graphics_api_combo->currentIndex());
connect(ui->graphics_api_combo, qOverload<int>(&QComboBox::currentIndexChanged), this,
&ConfigureGraphics::SetPhysicalDeviceComboVisibility);
connect(ui->toggle_hw_renderer, &QCheckBox::toggled, this, [this] {
const bool checked = ui->toggle_hw_renderer->isChecked();
@ -71,6 +83,12 @@ void ConfigureGraphics::SetConfiguration() {
ui->toggle_accurate_mul->setChecked(Settings::values.shaders_accurate_mul.GetValue());
ui->toggle_disk_shader_cache->setChecked(Settings::values.use_disk_shader_cache.GetValue());
ui->toggle_vsync_new->setChecked(Settings::values.use_vsync_new.GetValue());
ui->graphics_api_combo->setCurrentIndex(
static_cast<int>(Settings::values.graphics_api.GetValue()));
ui->physical_device_combo->setCurrentIndex(
static_cast<int>(Settings::values.physical_device.GetValue()));
ui->toggle_async_recording->setChecked(Settings::values.async_command_recording.GetValue());
ui->spirv_shader_gen->setChecked(Settings::values.spirv_shader_gen.GetValue());
if (Settings::IsConfiguringGlobal()) {
ui->toggle_shader_jit->setChecked(Settings::values.use_shader_jit.GetValue());
@ -90,6 +108,14 @@ void ConfigureGraphics::ApplyConfiguration() {
ui->toggle_disk_shader_cache, use_disk_shader_cache);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_vsync_new, ui->toggle_vsync_new,
use_vsync_new);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.graphics_api,
ui->graphics_api_combo);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.physical_device,
ui->physical_device_combo);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.async_command_recording,
ui->toggle_async_recording, async_command_recording);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.spirv_shader_gen,
ui->spirv_shader_gen, spirv_shader_gen);
if (Settings::IsConfiguringGlobal()) {
Settings::values.use_shader_jit = ui->toggle_shader_jit->isChecked();
@ -129,3 +155,22 @@ void ConfigureGraphics::SetupPerGameUI() {
ConfigurationShared::SetColoredTristate(ui->toggle_vsync_new, Settings::values.use_vsync_new,
use_vsync_new);
}
void ConfigureGraphics::DiscoverPhysicalDevices() {
Vulkan::Instance instance{};
const auto physical_devices = instance.GetPhysicalDevices();
ui->physical_device_combo->clear();
for (const vk::PhysicalDevice& physical_device : physical_devices) {
const QString name = QString::fromLocal8Bit(physical_device.getProperties().deviceName);
ui->physical_device_combo->addItem(name);
}
}
void ConfigureGraphics::SetPhysicalDeviceComboVisibility(int index) {
const auto graphics_api = static_cast<Settings::GraphicsAPI>(index);
const bool is_visible = graphics_api == Settings::GraphicsAPI::Vulkan;
ui->physical_device_label->setVisible(is_visible);
ui->physical_device_combo->setVisible(is_visible);
ui->spirv_shader_gen->setVisible(is_visible);
}

View File

@ -30,12 +30,18 @@ public:
void SetupPerGameUI();
private:
void DiscoverPhysicalDevices();
void SetPhysicalDeviceComboVisibility(int index);
ConfigurationShared::CheckState use_hw_renderer;
ConfigurationShared::CheckState use_hw_shader;
ConfigurationShared::CheckState separable_shader;
ConfigurationShared::CheckState shaders_accurate_mul;
ConfigurationShared::CheckState use_disk_shader_cache;
ConfigurationShared::CheckState use_vsync_new;
ConfigurationShared::CheckState async_command_recording;
ConfigurationShared::CheckState spirv_shader_gen;
std::unique_ptr<Ui::ConfigureGraphics> ui;
QColor bg_color;
};

View File

@ -7,7 +7,7 @@
<x>0</x>
<y>0</y>
<width>400</width>
<height>430</height>
<height>513</height>
</rect>
</property>
<property name="minimumSize">
@ -20,6 +20,66 @@
<string>Form</string>
</property>
<layout class="QVBoxLayout" name="verticalLayout">
<item>
<widget class="QGroupBox" name="apiBox">
<property name="title">
<string>API Settings</string>
</property>
<layout class="QVBoxLayout" name="verticalLayout_3">
<item>
<layout class="QHBoxLayout" name="horizontalLayout">
<item>
<widget class="QLabel" name="graphics_api_label">
<property name="text">
<string>Graphics API</string>
</property>
</widget>
</item>
<item>
<widget class="QComboBox" name="graphics_api_combo">
<item>
<property name="text">
<string>OpenGL</string>
</property>
</item>
<item>
<property name="text">
<string>OpenGLES</string>
</property>
</item>
<item>
<property name="text">
<string>Vulkan</string>
</property>
</item>
</widget>
</item>
</layout>
</item>
<item>
<layout class="QHBoxLayout" name="horizontalLayout_3">
<item>
<widget class="QLabel" name="physical_device_label">
<property name="text">
<string>Physical device</string>
</property>
</widget>
</item>
<item>
<widget class="QComboBox" name="physical_device_combo"/>
</item>
</layout>
</item>
<item>
<widget class="QCheckBox" name="spirv_shader_gen">
<property name="text">
<string>SPIR-V Shader Generation</string>
</property>
</widget>
</item>
</layout>
</widget>
</item>
<item>
<widget class="QGroupBox" name="rendererBox">
<property name="title">
@ -118,6 +178,16 @@
<string>Advanced</string>
</property>
<layout class="QVBoxLayout" name="verticalLayout_2">
<item>
<widget class="QCheckBox" name="toggle_async_recording">
<property name="toolTip">
<string>&lt;html&gt;&lt;head/&gt;&lt;body&gt;&lt;p&gt;Offloads command buffer recording and fragment shader generation to a worker thread. Can improve performance especially on weaker systems. Disable if you notice better performance. If unsure leave it enabled,&lt;/p&gt;&lt;/body&gt;&lt;/html&gt;</string>
</property>
<property name="text">
<string>Async Command Recording</string>
</property>
</widget>
</item>
<item>
<widget class="QCheckBox" name="toggle_disk_shader_cache">
<property name="toolTip">

View File

@ -16,7 +16,6 @@
#include "citra_qt/util/spinbox.h"
#include "common/color.h"
#include "core/core.h"
#include "core/hw/gpu.h"
#include "core/memory.h"
#include "video_core/pica_state.h"
#include "video_core/regs_framebuffer.h"

View File

@ -18,12 +18,6 @@
#include <QtGui>
#include <QtWidgets>
#include <fmt/format.h>
#ifdef __APPLE__
#include <unistd.h> // for chdir
#endif
#ifdef _WIN32
#include <windows.h>
#endif
#include "citra_qt/aboutdialog.h"
#include "citra_qt/applets/mii_selector.h"
#include "citra_qt/applets/swkbd.h"
@ -65,9 +59,7 @@
#include "common/file_util.h"
#include "common/literals.h"
#include "common/logging/backend.h"
#include "common/logging/filter.h"
#include "common/logging/log.h"
#include "common/logging/text_formatter.h"
#include "common/memory_detect.h"
#include "common/microprofile.h"
#include "common/scm_rev.h"
@ -82,8 +74,6 @@
#include "core/file_sys/archive_extsavedata.h"
#include "core/file_sys/archive_source_sd_savedata.h"
#include "core/frontend/applets/default_applets.h"
#include "core/frontend/scope_acquire_context.h"
#include "core/gdbstub/gdbstub.h"
#include "core/hle/service/fs/archive.h"
#include "core/hle/service/nfc/nfc.h"
#include "core/loader/loader.h"
@ -93,9 +83,15 @@
#include "input_common/main.h"
#include "network/network_settings.h"
#include "ui_main.h"
#include "video_core/renderer_base.h"
#include "video_core/video_core.h"
#ifdef __APPLE__
#include <unistd.h> // for chdir
#endif
#ifdef _WIN32
#include <windows.h>
#endif
#ifdef USE_DISCORD_PRESENCE
#include "citra_qt/discord_impl.h"
#endif
@ -298,7 +294,6 @@ void GMainWindow::InitializeWidgets() {
// Create status bar
message_label = new QLabel();
// Configured separately for left alignment
message_label->setVisible(false);
message_label->setFrameStyle(QFrame::NoFrame);
message_label->setContentsMargins(4, 0, 4, 0);
message_label->setAlignment(Qt::AlignLeft);
@ -323,10 +318,28 @@ void GMainWindow::InitializeWidgets() {
label->setVisible(false);
label->setFrameStyle(QFrame::NoFrame);
label->setContentsMargins(4, 0, 4, 0);
statusBar()->addPermanentWidget(label, 0);
statusBar()->addPermanentWidget(label);
}
statusBar()->addPermanentWidget(multiplayer_state->GetStatusText(), 0);
statusBar()->addPermanentWidget(multiplayer_state->GetStatusIcon(), 0);
// Setup Graphics API button
graphics_api_button = new QPushButton();
graphics_api_button->setObjectName(QStringLiteral("GraphicsAPIStatusBarButton"));
graphics_api_button->setFocusPolicy(Qt::NoFocus);
UpdateAPIIndicator(false);
connect(graphics_api_button, &QPushButton::clicked, this, [this] {
if (emulation_running) {
return;
}
UpdateAPIIndicator(true);
});
statusBar()->insertPermanentWidget(0, graphics_api_button);
statusBar()->addPermanentWidget(multiplayer_state->GetStatusText());
statusBar()->addPermanentWidget(multiplayer_state->GetStatusIcon());
statusBar()->setVisible(true);
// Removes an ugly inner border from the status bar widgets under Linux
@ -954,16 +967,7 @@ bool GMainWindow::LoadROM(const QString& filename) {
render_window->InitRenderTarget();
secondary_window->InitRenderTarget();
Frontend::ScopeAcquireContext scope(*render_window);
const QString below_gl43_title = tr("OpenGL 4.3 Unsupported");
const QString below_gl43_message = tr("Your GPU may not support OpenGL 4.3, or you do not "
"have the latest graphics driver.");
if (!QOpenGLContext::globalShareContext()->versionFunctions<QOpenGLFunctions_4_3_Core>()) {
QMessageBox::critical(this, below_gl43_title, below_gl43_message);
return false;
}
const auto scope = render_window->Acquire();
Core::System& system{Core::System::GetInstance()};
@ -1017,7 +1021,7 @@ bool GMainWindow::LoadROM(const QString& filename) {
case Core::System::ResultStatus::ErrorVideoCore:
QMessageBox::critical(
this, tr("Video Core Error"),
tr("An error has occurred. Please <a "
tr("An error has occurred during intialization of the video backend. Please <a "
"href='https://community.citra-emu.org/t/how-to-upload-the-log-file/296'>see "
"the "
"log</a> for more details. "
@ -1032,10 +1036,6 @@ bool GMainWindow::LoadROM(const QString& filename) {
"proper drivers for your graphics card from the manufacturer's website."));
break;
case Core::System::ResultStatus::ErrorVideoCore_ErrorBelowGL43:
QMessageBox::critical(this, below_gl43_title, below_gl43_message);
break;
default:
QMessageBox::critical(
this, tr("Error while loading ROM!"),
@ -1268,7 +1268,6 @@ void GMainWindow::ShutdownGame() {
// Disable status bar updates
status_bar_update_timer.stop();
message_label->setVisible(false);
message_label_used_for_movie = false;
emu_speed_label->setVisible(false);
game_fps_label->setVisible(false);
@ -1937,6 +1936,7 @@ void GMainWindow::OnConfigure() {
setMouseTracking(false);
}
UpdateSecondaryWindowVisibility();
UpdateAPIIndicator(false);
} else {
Settings::values.input_profiles = old_input_profiles;
Settings::values.touch_from_button_maps = old_touch_from_button_maps;
@ -2248,6 +2248,26 @@ void GMainWindow::ShowMouseCursor() {
}
}
void GMainWindow::UpdateAPIIndicator(bool override) {
static std::array graphics_apis = {QStringLiteral("OPENGL"), QStringLiteral("OPENGLES"),
QStringLiteral("VULKAN")};
static std::array graphics_api_colors = {QStringLiteral("#00ccdd"), QStringLiteral("#ba2a8d"),
QStringLiteral("#91242a")};
u32 api_index = static_cast<u32>(Settings::values.graphics_api.GetValue());
if (override) {
api_index = (api_index + 1) % graphics_apis.size();
Settings::values.graphics_api = static_cast<Settings::GraphicsAPI>(api_index);
}
const QString style_sheet = QStringLiteral("QPushButton { font-weight: bold; color: %0; }")
.arg(graphics_api_colors[api_index]);
graphics_api_button->setText(graphics_apis[api_index]);
graphics_api_button->setStyleSheet(style_sheet);
}
void GMainWindow::OnMouseActivity() {
ShowMouseCursor();
}
@ -2435,8 +2455,16 @@ void GMainWindow::UpdateUITheme() {
QStringList theme_paths(default_theme_paths);
if (is_default_theme || current_theme.isEmpty()) {
qApp->setStyleSheet({});
setStyleSheet({});
const QString theme_uri(QStringLiteral(":default/style.qss"));
QFile f(theme_uri);
if (f.open(QFile::ReadOnly | QFile::Text)) {
QTextStream ts(&f);
qApp->setStyleSheet(ts.readAll());
setStyleSheet(ts.readAll());
} else {
qApp->setStyleSheet({});
setStyleSheet({});
}
theme_paths.append(default_icons);
QIcon::setThemeName(default_icons);
} else {
@ -2615,14 +2643,6 @@ int main(int argc, char* argv[]) {
QCoreApplication::setOrganizationName(QStringLiteral("Citra team"));
QCoreApplication::setApplicationName(QStringLiteral("Citra"));
QSurfaceFormat format;
format.setVersion(4, 3);
format.setProfile(QSurfaceFormat::CoreProfile);
format.setSwapInterval(0);
// TODO: expose a setting for buffer value (ie default/single/double/triple)
format.setSwapBehavior(QSurfaceFormat::DefaultSwapBehavior);
QSurfaceFormat::setDefaultFormat(format);
#ifdef __APPLE__
std::string bin_path = FileUtil::GetBundleDirectory() + DIR_SEP + "..";
chdir(bin_path.c_str());

View File

@ -7,6 +7,7 @@
#include <array>
#include <memory>
#include <QMainWindow>
#include <QPushButton>
#include <QTimer>
#include <QTranslator>
#include "citra_qt/compatibility_list.h"
@ -241,6 +242,7 @@ private:
void HideMouseCursor();
void ShowMouseCursor();
void OpenPerGameConfiguration(u64 title_id, const QString& file_name);
void UpdateAPIIndicator(bool override);
std::unique_ptr<Ui::MainWindow> ui;
@ -256,6 +258,7 @@ private:
QLabel* emu_speed_label = nullptr;
QLabel* game_fps_label = nullptr;
QLabel* emu_frametime_label = nullptr;
QPushButton* graphics_api_button = nullptr;
QTimer status_bar_update_timer;
bool message_label_used_for_movie = false;

View File

@ -59,6 +59,60 @@ __declspec(dllimport) void __stdcall DebugBreak(void);
#endif // _MSC_VER ndef
#define DECLARE_ENUM_FLAG_OPERATORS(type) \
[[nodiscard]] constexpr type operator|(type a, type b) noexcept { \
using T = std::underlying_type_t<type>; \
return static_cast<type>(static_cast<T>(a) | static_cast<T>(b)); \
} \
[[nodiscard]] constexpr type operator&(type a, type b) noexcept { \
using T = std::underlying_type_t<type>; \
return static_cast<type>(static_cast<T>(a) & static_cast<T>(b)); \
} \
[[nodiscard]] constexpr type operator^(type a, type b) noexcept { \
using T = std::underlying_type_t<type>; \
return static_cast<type>(static_cast<T>(a) ^ static_cast<T>(b)); \
} \
[[nodiscard]] constexpr type operator<<(type a, type b) noexcept { \
using T = std::underlying_type_t<type>; \
return static_cast<type>(static_cast<T>(a) << static_cast<T>(b)); \
} \
[[nodiscard]] constexpr type operator>>(type a, type b) noexcept { \
using T = std::underlying_type_t<type>; \
return static_cast<type>(static_cast<T>(a) >> static_cast<T>(b)); \
} \
constexpr type& operator|=(type& a, type b) noexcept { \
a = a | b; \
return a; \
} \
constexpr type& operator&=(type& a, type b) noexcept { \
a = a & b; \
return a; \
} \
constexpr type& operator^=(type& a, type b) noexcept { \
a = a ^ b; \
return a; \
} \
constexpr type& operator<<=(type& a, type b) noexcept { \
a = a << b; \
return a; \
} \
constexpr type& operator>>=(type& a, type b) noexcept { \
a = a >> b; \
return a; \
} \
[[nodiscard]] constexpr type operator~(type key) noexcept { \
using T = std::underlying_type_t<type>; \
return static_cast<type>(~static_cast<T>(key)); \
} \
[[nodiscard]] constexpr bool True(type key) noexcept { \
using T = std::underlying_type_t<type>; \
return static_cast<T>(key) != 0; \
} \
[[nodiscard]] constexpr bool False(type key) noexcept { \
using T = std::underlying_type_t<type>; \
return static_cast<T>(key) == 0; \
}
// Generic function to get last error message.
// Call directly after the command or use the error num.
// This function might change the error code.

View File

@ -4,6 +4,7 @@
#pragma once
#include <concepts>
#include <cstddef>
#include <cstring>
#include "common/cityhash.h"
@ -41,6 +42,13 @@ inline u64 HashCombine(std::size_t& seed, const u64 hash) {
return seed ^= hash + 0x9e3779b9 + (seed << 6) + (seed >> 2);
}
template <typename T>
struct IdentityHash {
T operator()(const T& value) const {
return value;
}
};
/// A helper template that ensures the padding in a struct is initialized by memsetting to 0.
template <typename T>
struct HashableStruct {

View File

@ -236,6 +236,7 @@ void DebuggerBackend::Write(const Entry& entry) {
CLS(Render) \
SUB(Render, Software) \
SUB(Render, OpenGL) \
SUB(Render, Vulkan) \
CLS(Audio) \
SUB(Audio, DSP) \
SUB(Audio, Sink) \

View File

@ -8,6 +8,7 @@
#include <array>
#include "common/common_types.h"
#include "common/logging/formatter.h"
namespace Log {
// trims up to and including the last of ../, ..\, src/, src\ in a string
@ -103,6 +104,7 @@ enum class Class : ClassType {
Render, ///< Emulator video output and hardware acceleration
Render_Software, ///< Software renderer backend
Render_OpenGL, ///< OpenGL backend
Render_Vulkan, ///< Vulkan backend
Audio, ///< Audio emulation
Audio_DSP, ///< The HLE and LLE implementations of the DSP
Audio_Sink, ///< Emulator audio output backend

View File

@ -4,6 +4,7 @@
#pragma once
#include <compare>
#include <cstdlib>
#include <type_traits>
@ -23,19 +24,34 @@ struct Rectangle {
constexpr Rectangle(T left, T top, T right, T bottom)
: left(left), top(top), right(right), bottom(bottom) {}
[[nodiscard]] T GetWidth() const {
constexpr auto operator<=>(const Rectangle&) const = default;
constexpr void operator*=(const T value) {
left *= value;
top *= value;
right *= value;
bottom *= value;
}
[[nodiscard]] constexpr Rectangle operator*(const T value) const {
return Rectangle{left * value, top * value, right * value, bottom * value};
}
[[nodiscard]] constexpr Rectangle operator/(const T value) const {
return Rectangle{left / value, top / value, right / value, bottom / value};
}
[[nodiscard]] constexpr T GetWidth() const {
return std::abs(static_cast<std::make_signed_t<T>>(right - left));
}
[[nodiscard]] T GetHeight() const {
[[nodiscard]] constexpr T GetHeight() const {
return std::abs(static_cast<std::make_signed_t<T>>(bottom - top));
}
[[nodiscard]] Rectangle<T> TranslateX(const T x) const {
[[nodiscard]] constexpr Rectangle<T> TranslateX(const T x) const {
return Rectangle{left + x, top, right + x, bottom};
}
[[nodiscard]] Rectangle<T> TranslateY(const T y) const {
[[nodiscard]] constexpr Rectangle<T> TranslateY(const T y) const {
return Rectangle{left, top + y, right, bottom + y};
}
[[nodiscard]] Rectangle<T> Scale(const float s) const {
[[nodiscard]] constexpr Rectangle<T> Scale(const float s) const {
return Rectangle{left, top, static_cast<T>(left + GetWidth() * s),
static_cast<T>(top + GetHeight() * s)};
}

View File

@ -3,8 +3,8 @@
// Refer to the license.txt file included.
#pragma once
#include <memory>
#include <span>
#include <vector>
#include <boost/serialization/export.hpp>
#include <boost/serialization/shared_ptr.hpp>
@ -79,6 +79,7 @@ public:
: backing_mem(std::move(backing_mem_)), offset(0) {
Init();
}
MemoryRef(std::shared_ptr<BackingMem> backing_mem_, u64 offset_)
: backing_mem(std::move(backing_mem_)), offset(offset_) {
ASSERT(offset <= backing_mem->GetSize());
@ -93,11 +94,11 @@ public:
return cptr;
}
u8* GetPtr() {
operator const u8*() const {
return cptr;
}
operator const u8*() const {
u8* GetPtr() {
return cptr;
}
@ -105,6 +106,14 @@ public:
return cptr;
}
auto GetWriteBytes(std::size_t size) {
return std::span{reinterpret_cast<std::byte*>(cptr), size > csize ? csize : size};
}
auto GetReadBytes(std::size_t size) const {
return std::span{reinterpret_cast<const std::byte*>(cptr), size > csize ? csize : size};
}
std::size_t GetSize() const {
return csize;
}

View File

@ -8,7 +8,6 @@
#include "common/settings.h"
#include "core/core.h"
#include "core/gdbstub/gdbstub.h"
#include "core/hle/kernel/shared_page.h"
#include "core/hle/service/cam/cam.h"
#include "core/hle/service/hid/hid.h"
#include "core/hle/service/ir/ir_rst.h"
@ -20,6 +19,17 @@
namespace Settings {
[[nodiscard]] std::string_view GetAPIName(GraphicsAPI api) {
switch (api) {
case GraphicsAPI::OpenGL:
return "OpenGL";
case GraphicsAPI::OpenGLES:
return "OpenGLES";
case GraphicsAPI::Vulkan:
return "Vulkan";
}
}
Values values = {};
static bool configuring_global = true;
@ -95,7 +105,8 @@ void LogSettings() {
LOG_INFO(Config, "Citra Configuration:");
log_setting("Core_UseCpuJit", values.use_cpu_jit.GetValue());
log_setting("Core_CPUClockPercentage", values.cpu_clock_percentage.GetValue());
log_setting("Renderer_UseGLES", values.use_gles.GetValue());
log_setting("Renderer_GraphicsAPI", GetAPIName(values.graphics_api.GetValue()));
log_setting("Renderer_AsyncRecording", values.async_command_recording.GetValue());
log_setting("Renderer_UseHwRenderer", values.use_hw_renderer.GetValue());
log_setting("Renderer_UseHwShader", values.use_hw_shader.GetValue());
log_setting("Renderer_SeparableShader", values.separable_shader.GetValue());

View File

@ -15,6 +15,12 @@
namespace Settings {
enum class GraphicsAPI {
OpenGL = 0,
OpenGLES = 1,
Vulkan = 2,
};
enum class InitClock : u32 {
SystemTime = 0,
FixedTime = 1,
@ -440,7 +446,12 @@ struct Values {
Setting<bool> allow_plugin_loader{true, "allow_plugin_loader"};
// Renderer
Setting<bool> use_gles{false, "use_gles"};
SwitchableSetting<GraphicsAPI> graphics_api{GraphicsAPI::OpenGL, "graphics_api"};
SwitchableSetting<u16> physical_device{0, "physical_device"};
SwitchableSetting<bool> spirv_shader_gen{true, "spirv_shader_gen"};
Setting<bool> renderer_debug{false, "renderer_debug"};
Setting<bool> dump_command_buffers{false, "dump_command_buffers"};
SwitchableSetting<bool> async_command_recording{true, "async_command_recording"};
SwitchableSetting<bool> use_hw_renderer{true, "use_hw_renderer"};
SwitchableSetting<bool> use_hw_shader{true, "use_hw_shader"};
SwitchableSetting<bool> separable_shader{false, "use_separable_shader"};

View File

@ -113,8 +113,6 @@ add_library(core STATIC
frontend/input.h
frontend/mic.cpp
frontend/mic.h
frontend/scope_acquire_context.cpp
frontend/scope_acquire_context.h
gdbstub/gdbstub.cpp
gdbstub/gdbstub.h
hle/applets/applet.cpp

View File

@ -431,8 +431,6 @@ System::ResultStatus System::Init(Frontend::EmuWindow& emu_window,
switch (result) {
case VideoCore::ResultStatus::ErrorGenericDrivers:
return ResultStatus::ErrorVideoCore_ErrorGenericDrivers;
case VideoCore::ResultStatus::ErrorBelowGL43:
return ResultStatus::ErrorVideoCore_ErrorBelowGL43;
default:
return ResultStatus::ErrorVideoCore;
}

View File

@ -88,8 +88,6 @@ public:
ErrorVideoCore, ///< Error in the video core
ErrorVideoCore_ErrorGenericDrivers, ///< Error in the video core due to the user having
/// generic drivers installed
ErrorVideoCore_ErrorBelowGL43, ///< Error in the video core due to the user not having
/// OpenGL 4.3 or higher
ErrorSavestate, ///< Error saving or loading
ShutdownRequested, ///< Emulated program requested a system shutdown
ErrorUnknown ///< Any other error

View File

@ -14,6 +14,17 @@
namespace Frontend {
/// Information for the Graphics Backends signifying what type of screen pointer is in
/// WindowInformation
enum class WindowSystemType : u8 {
Headless,
Android,
Windows,
MacOS,
X11,
Wayland,
};
struct Frame;
/**
* For smooth Vsync rendering, we want to always present the latest frame that the core generates,
@ -62,11 +73,33 @@ class GraphicsContext {
public:
virtual ~GraphicsContext();
/// Inform the driver to swap the front/back buffers and present the current image
virtual void SwapBuffers(){};
/// Makes the graphics context current for the caller thread
virtual void MakeCurrent() = 0;
virtual void MakeCurrent(){};
/// Releases (dunno if this is the "right" word) the context from the caller thread
virtual void DoneCurrent() = 0;
virtual void DoneCurrent(){};
class Scoped {
public:
explicit Scoped(GraphicsContext& context_) : context(context_) {
context.MakeCurrent();
}
~Scoped() {
context.DoneCurrent();
}
private:
GraphicsContext& context;
};
/// Calls MakeCurrent on the context and calls DoneCurrent when the scope for the returned value
/// ends
[[nodiscard]] Scoped Acquire() {
return Scoped{*this};
}
};
/**
@ -100,6 +133,23 @@ public:
Core::kScreenTopWidth, Core::kScreenTopHeight + Core::kScreenBottomHeight};
};
/// Data describing host window system information
struct WindowSystemInfo {
// Window system type. Determines which GL context or Vulkan WSI is used.
WindowSystemType type = WindowSystemType::Headless;
// Connection to a display server. This is used on X11 and Wayland platforms.
void* display_connection = nullptr;
// Render surface. This is a pointer to the native window handle, which depends
// on the platform. e.g. HWND for Windows, Window for X11. If the surface is
// set to nullptr, the video backend will run in headless mode.
void* render_surface = nullptr;
// Scale of the render surface. For hidpi systems, this will be >1.
float render_surface_scale = 1.0f;
};
/// Polls window events
virtual void PollEvents() = 0;
@ -163,6 +213,13 @@ public:
config = val;
}
/**
* Returns system information about the drawing area.
*/
const WindowSystemInfo& GetWindowInfo() const {
return window_info;
}
/**
* Gets the framebuffer layout (width, height, and screen regions)
* @note This method is thread-safe
@ -211,6 +268,7 @@ protected:
}
bool is_secondary{};
WindowSystemInfo window_info;
private:
/**

View File

@ -1,17 +0,0 @@
// Copyright 2019 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include "core/frontend/emu_window.h"
#include "core/frontend/scope_acquire_context.h"
namespace Frontend {
ScopeAcquireContext::ScopeAcquireContext(Frontend::GraphicsContext& context) : context{context} {
context.MakeCurrent();
}
ScopeAcquireContext::~ScopeAcquireContext() {
context.DoneCurrent();
}
} // namespace Frontend

View File

@ -1,23 +0,0 @@
// Copyright 2019 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include "common/common_types.h"
namespace Frontend {
class GraphicsContext;
/// Helper class to acquire/release window context within a given scope
class ScopeAcquireContext : NonCopyable {
public:
explicit ScopeAcquireContext(Frontend::GraphicsContext& context);
~ScopeAcquireContext();
private:
Frontend::GraphicsContext& context;
};
} // namespace Frontend

View File

@ -7,6 +7,7 @@
#include <array>
#include <memory>
#include <string>
#include <string_view>
#include <utility>
#include <vector>
#include "common/common_types.h"

View File

@ -3,8 +3,6 @@
// Refer to the license.txt file included.
#include <cryptopp/osrng.h>
#include "common/assert.h"
#include "common/common_types.h"
#include "common/file_util.h"
#include "common/logging/log.h"

View File

@ -1053,7 +1053,7 @@ public:
}
Common::ParamPackage GetNextInput() override {
SDL_Event event;
SDL_Event event{};
while (state.event_queue.Pop(event)) {
if (event.type != SDL_JOYAXISMOTION || std::abs(event.jaxis.value / 32767.0) < 0.5) {
continue;

View File

@ -1,3 +1,5 @@
add_subdirectory(host_shaders)
add_library(video_core STATIC
command_processor.cpp
command_processor.h
@ -13,6 +15,8 @@ add_library(video_core STATIC
precompiled_headers.h
primitive_assembly.cpp
primitive_assembly.h
rasterizer_accelerated.cpp
rasterizer_accelerated.h
rasterizer_interface.h
regs.cpp
regs.h
@ -24,21 +28,22 @@ add_library(video_core STATIC
regs_texturing.h
renderer_base.cpp
renderer_base.h
rasterizer_cache/cached_surface.cpp
rasterizer_cache/cached_surface.h
rasterizer_cache/morton_swizzle.h
rasterizer_cache/pixel_format.cpp
rasterizer_cache/pixel_format.h
rasterizer_cache/rasterizer_cache.cpp
rasterizer_cache/rasterizer_cache.h
rasterizer_cache/rasterizer_cache_types.h
rasterizer_cache/rasterizer_cache_utils.cpp
rasterizer_cache/rasterizer_cache_utils.h
rasterizer_cache/surface_base.h
rasterizer_cache/utils.cpp
rasterizer_cache/utils.h
rasterizer_cache/surface_params.cpp
rasterizer_cache/surface_params.h
rasterizer_cache/texture_runtime.cpp
rasterizer_cache/texture_runtime.h
renderer_opengl/frame_dumper_opengl.cpp
renderer_opengl/frame_dumper_opengl.h
renderer_opengl/gl_driver.cpp
renderer_opengl/gl_driver.h
renderer_opengl/gl_format_reinterpreter.cpp
renderer_opengl/gl_format_reinterpreter.h
renderer_opengl/gl_rasterizer.cpp
renderer_opengl/gl_rasterizer.h
renderer_opengl/gl_resource_manager.cpp
@ -57,6 +62,8 @@ add_library(video_core STATIC
renderer_opengl/gl_state.h
renderer_opengl/gl_stream_buffer.cpp
renderer_opengl/gl_stream_buffer.h
renderer_opengl/gl_texture_runtime.cpp
renderer_opengl/gl_texture_runtime.h
renderer_opengl/gl_vars.cpp
renderer_opengl/gl_vars.h
renderer_opengl/pica_to_gl.h
@ -64,8 +71,6 @@ add_library(video_core STATIC
renderer_opengl/post_processing_opengl.h
renderer_opengl/renderer_opengl.cpp
renderer_opengl/renderer_opengl.h
renderer_opengl/texture_downloader_es.cpp
renderer_opengl/texture_downloader_es.h
renderer_opengl/texture_filters/anime4k/anime4k_ultrafast.cpp
renderer_opengl/texture_filters/anime4k/anime4k_ultrafast.h
renderer_opengl/texture_filters/bicubic/bicubic.cpp
@ -79,14 +84,55 @@ add_library(video_core STATIC
renderer_opengl/texture_filters/texture_filterer.h
renderer_opengl/texture_filters/xbrz/xbrz_freescale.cpp
renderer_opengl/texture_filters/xbrz/xbrz_freescale.h
#temporary, move these back in alphabetical order before merging
renderer_opengl/gl_format_reinterpreter.cpp
renderer_opengl/gl_format_reinterpreter.h
renderer_vulkan/pica_to_vk.h
renderer_vulkan/renderer_vulkan.cpp
renderer_vulkan/renderer_vulkan.h
renderer_vulkan/vk_blit_helper.cpp
renderer_vulkan/vk_blit_helper.h
renderer_vulkan/vk_common.cpp
renderer_vulkan/vk_common.h
renderer_vulkan/vk_descriptor_manager.cpp
renderer_vulkan/vk_descriptor_manager.h
renderer_vulkan/vk_format_reinterpreter.cpp
renderer_vulkan/vk_format_reinterpreter.h
renderer_vulkan/vk_master_semaphore.cpp
renderer_vulkan/vk_master_semaphore.h
renderer_vulkan/vk_rasterizer.cpp
renderer_vulkan/vk_rasterizer.h
renderer_vulkan/vk_scheduler.cpp
renderer_vulkan/vk_scheduler.h
renderer_vulkan/vk_resource_pool.cpp
renderer_vulkan/vk_resource_pool.h
renderer_vulkan/vk_instance.cpp
renderer_vulkan/vk_instance.h
renderer_vulkan/vk_pipeline_cache.cpp
renderer_vulkan/vk_pipeline_cache.h
renderer_vulkan/vk_platform.cpp
renderer_vulkan/vk_platform.h
renderer_vulkan/vk_renderpass_cache.cpp
renderer_vulkan/vk_renderpass_cache.h
renderer_vulkan/vk_shader_decompiler.cpp
renderer_vulkan/vk_shader_decompiler.h
renderer_vulkan/vk_shader_gen.cpp
renderer_vulkan/vk_shader_gen.h
renderer_vulkan/vk_shader_gen_spv.cpp
renderer_vulkan/vk_shader_gen_spv.h
renderer_vulkan/vk_shader_util.cpp
renderer_vulkan/vk_shader_util.h
renderer_vulkan/vk_stream_buffer.cpp
renderer_vulkan/vk_stream_buffer.h
renderer_vulkan/vk_swapchain.cpp
renderer_vulkan/vk_swapchain.h
renderer_vulkan/vk_texture_runtime.cpp
renderer_vulkan/vk_texture_runtime.h
shader/debug_data.h
shader/shader.cpp
shader/shader.h
shader/shader_cache.h
shader/shader_interpreter.cpp
shader/shader_interpreter.h
shader/shader_uniforms.cpp
shader/shader_uniforms.h
swrasterizer/clipper.cpp
swrasterizer/clipper.h
swrasterizer/framebuffer.cpp
@ -112,36 +158,6 @@ add_library(video_core STATIC
video_core.h
)
set(SHADER_FILES
renderer_opengl/depth_to_color.frag
renderer_opengl/depth_to_color.vert
renderer_opengl/ds_to_color.frag
renderer_opengl/texture_filters/anime4k/refine.frag
renderer_opengl/texture_filters/anime4k/x_gradient.frag
renderer_opengl/texture_filters/anime4k/y_gradient.frag
renderer_opengl/texture_filters/bicubic/bicubic.frag
renderer_opengl/texture_filters/nearest_neighbor/nearest_neighbor.frag
renderer_opengl/texture_filters/scale_force/scale_force.frag
renderer_opengl/texture_filters/tex_coord.vert
renderer_opengl/texture_filters/xbrz/xbrz_freescale.frag
renderer_opengl/texture_filters/xbrz/xbrz_freescale.vert
)
include(${CMAKE_CURRENT_SOURCE_DIR}/generate_shaders.cmake)
foreach(shader_file ${SHADER_FILES})
get_filename_component(shader_file_name ${shader_file} NAME)
GetShaderHeaderFile(${shader_file_name})
list(APPEND SHADER_HEADERS ${shader_header_file})
endforeach()
add_custom_target(shaders
BYPRODUCTS ${SHADER_HEADERS}
COMMAND "${CMAKE_COMMAND}" -P ${CMAKE_CURRENT_SOURCE_DIR}/generate_shaders.cmake
SOURCES ${SHADER_FILES}
)
add_dependencies(video_core shaders)
target_include_directories(video_core PRIVATE ${CMAKE_CURRENT_BINARY_DIR})
if(ARCHITECTURE_x86_64)
@ -157,10 +173,19 @@ endif()
create_target_directory_groups(video_core)
# Ignore nullability warnings generated from VMA
if (NOT MSVC)
target_compile_options(vma INTERFACE -Wno-unused-variable -Wno-nullability-completeness)
endif()
target_link_libraries(video_core PUBLIC common core)
target_link_libraries(video_core PRIVATE glad nihstro-headers Boost::serialization)
set_target_properties(video_core PROPERTIES INTERPROCEDURAL_OPTIMIZATION ${ENABLE_LTO})
add_dependencies(video_core host_shaders)
target_include_directories(video_core PRIVATE ${HOST_SHADERS_INCLUDE})
target_link_libraries(video_core PRIVATE vulkan-headers vma sirit SPIRV glslang glad)
target_link_libraries(video_core PRIVATE nihstro-headers Boost::serialization glm::glm)
if (ARCHITECTURE_x86_64)
target_link_libraries(video_core PUBLIC xbyak)
endif()

View File

@ -0,0 +1,100 @@
# Copyright 2022 Citra Emulator Project
# Licensed under GPLv2 or any later version
# Refer to the license.txt file included.
set(SHADER_FILES
texture_filtering/bicubic.frag
texture_filtering/nearest_neighbor.frag
texture_filtering/refine.frag
texture_filtering/scale_force.frag
texture_filtering/tex_coord.vert
texture_filtering/xbrz_freescale.frag
texture_filtering/xbrz_freescale.vert
texture_filtering/x_gradient.frag
texture_filtering/y_gradient.frag
opengl_present.frag
opengl_present.vert
opengl_present_anaglyph.frag
opengl_present_interlaced.frag
vulkan_present.frag
vulkan_present.vert
vulkan_present_anaglyph.frag
vulkan_present_interlaced.frag
)
find_program(GLSLANGVALIDATOR "glslangValidator")
if ("${GLSLANGVALIDATOR}" STREQUAL "GLSLANGVALIDATOR-NOTFOUND")
message(FATAL_ERROR "Required program `glslangValidator` not found.")
endif()
set(MACROS "-Dgl_VertexID=gl_VertexIndex")
set(QUIET_FLAG "--quiet")
set(SHADER_INCLUDE ${CMAKE_CURRENT_BINARY_DIR}/include)
set(SHADER_DIR ${SHADER_INCLUDE}/video_core/host_shaders)
set(HOST_SHADERS_INCLUDE ${SHADER_INCLUDE} PARENT_SCOPE)
set(INPUT_FILE ${CMAKE_CURRENT_SOURCE_DIR}/source_shader.h.in)
set(HEADER_GENERATOR ${CMAKE_CURRENT_SOURCE_DIR}/StringShaderHeader.cmake)
# Check if `--quiet` is available on host's glslangValidator version
# glslangValidator prints to STDERR iff an unrecognized flag is passed to it
execute_process(
COMMAND
${GLSLANGVALIDATOR} ${QUIET_FLAG}
ERROR_VARIABLE
GLSLANG_ERROR
# STDOUT variable defined to silence unnecessary output during CMake configuration
OUTPUT_VARIABLE
GLSLANG_OUTPUT
)
if (NOT GLSLANG_ERROR STREQUAL "")
message(WARNING "Refusing to use unavailable flag `${QUIET_FLAG}` on `${GLSLANGVALIDATOR}`")
set(QUIET_FLAG "")
endif()
foreach(FILENAME IN ITEMS ${SHADER_FILES})
string(REPLACE "." "_" SHADER_NAME ${FILENAME})
set(SOURCE_FILE ${CMAKE_CURRENT_SOURCE_DIR}/${FILENAME})
# Skip generating source headers on Vulkan exclusive files
if (NOT ${FILENAME} MATCHES "vulkan.*")
set(SOURCE_HEADER_FILE ${SHADER_DIR}/${SHADER_NAME}.h)
add_custom_command(
OUTPUT
${SOURCE_HEADER_FILE}
COMMAND
${CMAKE_COMMAND} -P ${HEADER_GENERATOR} ${SOURCE_FILE} ${SOURCE_HEADER_FILE} ${INPUT_FILE}
MAIN_DEPENDENCY
${SOURCE_FILE}
DEPENDS
${INPUT_FILE}
# HEADER_GENERATOR should be included here but msbuild seems to assume it's always modified
)
set(SHADER_HEADERS ${SHADER_HEADERS} ${SOURCE_HEADER_FILE})
endif()
# Skip compiling to SPIR-V OpenGL exclusive files
if (NOT ${FILENAME} MATCHES "opengl.*")
string(TOUPPER ${SHADER_NAME}_SPV SPIRV_VARIABLE_NAME)
set(SPIRV_HEADER_FILE ${SHADER_DIR}/${SHADER_NAME}_spv.h)
add_custom_command(
OUTPUT
${SPIRV_HEADER_FILE}
COMMAND
${GLSLANGVALIDATOR} --target-env vulkan1.1 --glsl-version 450 ${QUIET_FLAG} ${MACROS} --variable-name ${SPIRV_VARIABLE_NAME} -o ${SPIRV_HEADER_FILE} ${SOURCE_FILE}
MAIN_DEPENDENCY
${SOURCE_FILE}
)
set(SHADER_HEADERS ${SHADER_HEADERS} ${SPIRV_HEADER_FILE})
endif()
endforeach()
set(SHADER_SOURCES ${SHADER_FILES})
list(APPEND SHADER_SOURCES ${GLSL_INCLUDES})
add_custom_target(host_shaders
DEPENDS
${SHADER_HEADERS}
SOURCES
${SHADER_SOURCES}
)

View File

@ -0,0 +1,36 @@
# SPDX-FileCopyrightText: 2020 yuzu Emulator Project
# SPDX-License-Identifier: GPL-2.0-or-later
set(SOURCE_FILE ${CMAKE_ARGV3})
set(HEADER_FILE ${CMAKE_ARGV4})
set(INPUT_FILE ${CMAKE_ARGV5})
get_filename_component(CONTENTS_NAME ${SOURCE_FILE} NAME)
string(REPLACE "." "_" CONTENTS_NAME ${CONTENTS_NAME})
string(TOUPPER ${CONTENTS_NAME} CONTENTS_NAME)
FILE(READ ${SOURCE_FILE} line_contents)
# Replace double quotes with single quotes,
# as double quotes will be used to wrap the lines
STRING(REGEX REPLACE "\"" "'" line_contents "${line_contents}")
# CMake separates list elements with semicolons, but semicolons
# are used extensively in the shader code.
# Replace with a temporary marker, to be reverted later.
STRING(REGEX REPLACE ";" "{{SEMICOLON}}" line_contents "${line_contents}")
# Make every line an individual element in the CMake list.
STRING(REGEX REPLACE "\n" ";" line_contents "${line_contents}")
# Build the shader string, wrapping each line in double quotes.
foreach(line IN LISTS line_contents)
string(CONCAT CONTENTS "${CONTENTS}" \"${line}\\n\"\n)
endforeach()
# Revert the original semicolons in the source.
STRING(REGEX REPLACE "{{SEMICOLON}}" ";" CONTENTS "${CONTENTS}")
get_filename_component(OUTPUT_DIR ${HEADER_FILE} DIRECTORY)
make_directory(${OUTPUT_DIR})
configure_file(${INPUT_FILE} ${HEADER_FILE} @ONLY)

View File

@ -0,0 +1,18 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
layout(location = 0) in vec2 frag_tex_coord;
layout(location = 0) out vec4 color;
layout(binding = 0) uniform sampler2D color_texture;
uniform vec4 i_resolution;
uniform vec4 o_resolution;
uniform int layer;
void main() {
color = texture(color_texture, frag_tex_coord);
}

View File

@ -0,0 +1,23 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
layout(location = 0) in vec2 vert_position;
layout(location = 1) in vec2 vert_tex_coord;
layout(location = 0) out vec2 frag_tex_coord;
// This is a truncated 3x3 matrix for 2D transformations:
// The upper-left 2x2 submatrix performs scaling/rotation/mirroring.
// The third column performs translation.
// The third row could be used for projection, which we don't need in 2D. It hence is assumed to
// implicitly be [0, 0, 1]
uniform mat3x2 modelview_matrix;
void main() {
// Multiply input position by the rotscale part of the matrix and then manually translate by
// the last column. This is equivalent to using a full 3x3 matrix and expanding the vector
// to `vec3(vert_position.xy, 1.0)`
gl_Position = vec4(mat2(modelview_matrix) * vert_position + modelview_matrix[2], 0.0, 1.0);
frag_tex_coord = vert_tex_coord;
}

View File

@ -0,0 +1,32 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
// Anaglyph Red-Cyan shader based on Dubois algorithm
// Constants taken from the paper:
// "Conversion of a Stereo Pair to Anaglyph with
// the Least-Squares Projection Method"
// Eric Dubois, March 2009
const mat3 l = mat3( 0.437, 0.449, 0.164,
-0.062,-0.062,-0.024,
-0.048,-0.050,-0.017);
const mat3 r = mat3(-0.011,-0.032,-0.007,
0.377, 0.761, 0.009,
-0.026,-0.093, 1.234);
layout(location = 0) in vec2 frag_tex_coord;
layout(location = 0) out vec4 color;
layout(binding = 0) uniform sampler2D color_texture;
layout(binding = 1) uniform sampler2D color_texture_r;
uniform vec4 resolution;
uniform int layer;
void main() {
vec4 color_tex_l = texture(color_texture, frag_tex_coord);
vec4 color_tex_r = texture(color_texture_r, frag_tex_coord);
color = vec4(color_tex_l.rgb*l+color_tex_r.rgb*r, color_tex_l.a);
}

View File

@ -0,0 +1,22 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
layout(location = 0) in vec2 frag_tex_coord;
layout(location = 0) out vec4 color;
layout(binding = 0) uniform sampler2D color_texture;
layout(binding = 1) uniform sampler2D color_texture_r;
uniform vec4 o_resolution;
uniform int reverse_interlaced;
void main() {
float screen_row = o_resolution.x * frag_tex_coord.x;
if (int(screen_row) % 2 == reverse_interlaced)
color = texture(color_texture, frag_tex_coord);
else
color = texture(color_texture_r, frag_tex_coord);
}

View File

@ -0,0 +1,15 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <string_view>
namespace HostShaders {
constexpr std::string_view @CONTENTS_NAME@ = {
@CONTENTS@
};
} // namespace HostShaders

View File

@ -1,11 +1,14 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 330
precision mediump float;
in vec2 tex_coord;
layout(location = 0) in vec2 tex_coord;
layout(location = 0) out vec4 frag_color;
out vec4 frag_color;
uniform sampler2D input_texture;
layout(binding = 0) uniform sampler2D input_texture;
// from http://www.java-gaming.org/index.php?topic=35123.0
vec4 cubic(float v) {
@ -18,9 +21,9 @@ vec4 cubic(float v) {
return vec4(x, y, z, w) * (1.0 / 6.0);
}
vec4 textureBicubic(sampler2D sampler, vec2 texCoords) {
vec4 textureBicubic(sampler2D tex_sampler, vec2 texCoords) {
vec2 texSize = vec2(textureSize(sampler, 0));
vec2 texSize = vec2(textureSize(tex_sampler, 0));
vec2 invTexSize = 1.0 / texSize;
texCoords = texCoords * texSize - 0.5;
@ -38,10 +41,10 @@ vec4 textureBicubic(sampler2D sampler, vec2 texCoords) {
offset *= invTexSize.xxyy;
vec4 sample0 = texture(sampler, offset.xz);
vec4 sample1 = texture(sampler, offset.yz);
vec4 sample2 = texture(sampler, offset.xw);
vec4 sample3 = texture(sampler, offset.yw);
vec4 sample0 = texture(tex_sampler, offset.xz);
vec4 sample1 = texture(tex_sampler, offset.yz);
vec4 sample2 = texture(tex_sampler, offset.xw);
vec4 sample3 = texture(tex_sampler, offset.yw);
float sx = s.x / (s.x + s.y);
float sy = s.z / (s.z + s.w);

View File

@ -0,0 +1,15 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
precision mediump float;
layout(location = 0) in vec2 tex_coord;
layout(location = 0) out vec4 frag_color;
layout(binding = 0) uniform sampler2D input_texture;
void main() {
frag_color = texture(input_texture, tex_coord);
}

View File

@ -1,12 +1,15 @@
//? #version 330
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
precision mediump float;
in vec2 tex_coord;
layout(location = 0) in vec2 tex_coord;
layout(location = 0) out vec4 frag_color;
out vec4 frag_color;
uniform sampler2D HOOKED;
uniform sampler2D LUMAD;
layout(binding = 0) uniform sampler2D HOOKED;
layout(binding = 1) uniform sampler2D LUMAD;
const float LINE_DETECT_THRESHOLD = 0.4;
const float STRENGTH = 0.6;

View File

@ -1,5 +1,3 @@
//? #version 320 es
// from https://github.com/BreadFish64/ScaleFish/tree/master/scale_force
// MIT License
@ -24,13 +22,18 @@
// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
// SOFTWARE.
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 320 es
precision mediump float;
in vec2 tex_coord;
layout(location = 0) in vec2 tex_coord;
layout(location = 0) out vec4 frag_color;
out vec4 frag_color;
uniform sampler2D input_texture;
layout(binding = 0) uniform sampler2D input_texture;
vec2 tex_size;
vec2 inv_tex_size;

View File

@ -1,5 +1,9 @@
//? #version 330
out vec2 tex_coord;
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
layout(location = 0) out vec2 tex_coord;
const vec2 vertices[4] =
vec2[4](vec2(-1.0, -1.0), vec2(1.0, -1.0), vec2(-1.0, 1.0), vec2(1.0, 1.0));

View File

@ -1,11 +1,14 @@
//? #version 330
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
precision mediump float;
in vec2 tex_coord;
layout(location = 0) in vec2 tex_coord;
layout(location = 0) out vec2 frag_color;
out vec2 frag_color;
uniform sampler2D tex_input;
layout(binding = 0) uniform sampler2D tex_input;
const vec3 K = vec3(0.2627, 0.6780, 0.0593);
// TODO: improve handling of alpha channel

View File

@ -1,14 +1,25 @@
//? #version 330
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
precision mediump float;
in vec2 tex_coord;
in vec2 source_size;
in vec2 output_size;
layout(location = 0) in vec2 tex_coord;
layout(location = 1) in vec2 source_size;
layout(location = 2) in vec2 output_size;
out vec4 frag_color;
layout(location = 0) out vec4 frag_color;
uniform sampler2D tex;
layout(binding = 0) uniform sampler2D tex;
#ifdef VULKAN
layout(push_constant, std140) uniform XbrzInfo {
lowp float scale;
};
#else
uniform lowp float scale;
#endif
const int BLEND_NONE = 0;
const int BLEND_NORMAL = 1;

View File

@ -0,0 +1,28 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
layout(location = 0) out vec2 tex_coord;
layout(location = 1) out vec2 source_size;
layout(location = 2) out vec2 output_size;
layout(binding = 0) uniform sampler2D tex;
#ifdef VULKAN
layout(push_constant, std140) uniform XbrzInfo {
lowp float scale;
};
#else
uniform lowp float scale;
#endif
const vec2 vertices[4] =
vec2[4](vec2(-1.0, -1.0), vec2(1.0, -1.0), vec2(-1.0, 1.0), vec2(1.0, 1.0));
void main() {
gl_Position = vec4(vertices[gl_VertexID], 0.0, 1.0);
tex_coord = (vertices[gl_VertexID] + 1.0) / 2.0;
source_size = vec2(textureSize(tex, 0));
output_size = source_size * scale;
}

View File

@ -1,11 +1,14 @@
//? #version 330
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
//? #version 430 core
precision mediump float;
in vec2 tex_coord;
layout(location = 0) in vec2 tex_coord;
layout(location = 0) out float frag_color;
out float frag_color;
uniform sampler2D tex_input;
layout(binding = 0) uniform sampler2D tex_input;
void main() {
vec2 t = textureLodOffset(tex_input, tex_coord, 0.0, ivec2(0, 1)).xy;

View File

@ -0,0 +1,26 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#version 450 core
#extension GL_ARB_separate_shader_objects : enable
layout (location = 0) in vec2 frag_tex_coord;
layout (location = 0) out vec4 color;
layout (push_constant, std140) uniform DrawInfo {
mat4 modelview_matrix;
vec4 i_resolution;
vec4 o_resolution;
int screen_id_l;
int screen_id_r;
int layer;
int reverse_interlaced;
};
layout (set = 0, binding = 0) uniform texture2D screen_textures[3];
layout (set = 0, binding = 1) uniform sampler screen_sampler;
void main() {
color = texture(sampler2D(screen_textures[screen_id_l], screen_sampler), frag_tex_coord);
}

View File

@ -0,0 +1,25 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#version 450 core
#extension GL_ARB_separate_shader_objects : enable
layout (location = 0) in vec2 vert_position;
layout (location = 1) in vec2 vert_tex_coord;
layout (location = 0) out vec2 frag_tex_coord;
layout (push_constant, std140) uniform DrawInfo {
mat4 modelview_matrix;
vec4 i_resolution;
vec4 o_resolution;
int screen_id_l;
int screen_id_r;
int layer;
};
void main() {
vec4 position = vec4(vert_position, 0.0, 1.0) * modelview_matrix;
gl_Position = vec4(position.x, -position.y, 0.0, 1.0);
frag_tex_coord = vert_tex_coord;
}

View File

@ -0,0 +1,40 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#version 450 core
#extension GL_ARB_separate_shader_objects : enable
layout (location = 0) in vec2 frag_tex_coord;
layout (location = 0) out vec4 color;
// Anaglyph Red-Cyan shader based on Dubois algorithm
// Constants taken from the paper:
// "Conversion of a Stereo Pair to Anaglyph with
// the Least-Squares Projection Method"
// Eric Dubois, March 2009
const mat3 l = mat3( 0.437, 0.449, 0.164,
-0.062,-0.062,-0.024,
-0.048,-0.050,-0.017);
const mat3 r = mat3(-0.011,-0.032,-0.007,
0.377, 0.761, 0.009,
-0.026,-0.093, 1.234);
layout (push_constant, std140) uniform DrawInfo {
mat4 modelview_matrix;
vec4 i_resolution;
vec4 o_resolution;
int screen_id_l;
int screen_id_r;
int layer;
int reverse_interlaced;
};
layout (set = 0, binding = 0) uniform texture2D screen_textures[3];
layout (set = 0, binding = 1) uniform sampler screen_sampler;
void main() {
vec4 color_tex_l = texture(sampler2D(screen_textures[screen_id_l], screen_sampler), frag_tex_coord);
vec4 color_tex_r = texture(sampler2D(screen_textures[screen_id_r], screen_sampler), frag_tex_coord);
color = vec4(color_tex_l.rgb*l+color_tex_r.rgb*r, color_tex_l.a);
}

View File

@ -0,0 +1,30 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#version 450 core
#extension GL_ARB_separate_shader_objects : enable
layout (location = 0) in vec2 frag_tex_coord;
layout (location = 0) out vec4 color;
layout (push_constant, std140) uniform DrawInfo {
mat4 modelview_matrix;
vec4 i_resolution;
vec4 o_resolution;
int screen_id_l;
int screen_id_r;
int layer;
int reverse_interlaced;
};
layout (set = 0, binding = 0) uniform texture2D screen_textures[3];
layout (set = 0, binding = 1) uniform sampler screen_sampler;
void main() {
float screen_row = o_resolution.x * frag_tex_coord.x;
if (int(screen_row) % 2 == reverse_interlaced)
color = texture(sampler2D(screen_textures[screen_id_l], screen_sampler), frag_tex_coord);
else
color = texture(sampler2D(screen_textures[screen_id_r], screen_sampler), frag_tex_coord);
}

View File

@ -40,8 +40,8 @@ void Zero(T& o) {
State::State() : geometry_pipeline(*this) {
auto SubmitVertex = [this](const Shader::AttributeBuffer& vertex) {
using Pica::Shader::OutputVertex;
auto AddTriangle = [this](const OutputVertex& v0, const OutputVertex& v1,
const OutputVertex& v2) {
auto AddTriangle = [](const OutputVertex& v0, const OutputVertex& v1,
const OutputVertex& v2) {
VideoCore::g_renderer->Rasterizer()->AddTriangle(v0, v1, v2);
};
primitive_assembler.SubmitVertex(

View File

@ -0,0 +1,840 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <limits>
#include "common/alignment.h"
#include "core/memory.h"
#include "video_core/pica_state.h"
#include "video_core/rasterizer_accelerated.h"
#include "video_core/video_core.h"
namespace VideoCore {
static Common::Vec4f ColorRGBA8(const u32 color) {
const auto rgba =
Common::Vec4u{color >> 0 & 0xFF, color >> 8 & 0xFF, color >> 16 & 0xFF, color >> 24 & 0xFF};
return rgba / 255.0f;
}
static Common::Vec3f LightColor(const Pica::LightingRegs::LightColor& color) {
return Common::Vec3u{color.r, color.g, color.b} / 255.0f;
}
RasterizerAccelerated::HardwareVertex::HardwareVertex(const Pica::Shader::OutputVertex& v,
bool flip_quaternion) {
position[0] = v.pos.x.ToFloat32();
position[1] = v.pos.y.ToFloat32();
position[2] = v.pos.z.ToFloat32();
position[3] = v.pos.w.ToFloat32();
color[0] = v.color.x.ToFloat32();
color[1] = v.color.y.ToFloat32();
color[2] = v.color.z.ToFloat32();
color[3] = v.color.w.ToFloat32();
tex_coord0[0] = v.tc0.x.ToFloat32();
tex_coord0[1] = v.tc0.y.ToFloat32();
tex_coord1[0] = v.tc1.x.ToFloat32();
tex_coord1[1] = v.tc1.y.ToFloat32();
tex_coord2[0] = v.tc2.x.ToFloat32();
tex_coord2[1] = v.tc2.y.ToFloat32();
tex_coord0_w = v.tc0_w.ToFloat32();
normquat[0] = v.quat.x.ToFloat32();
normquat[1] = v.quat.y.ToFloat32();
normquat[2] = v.quat.z.ToFloat32();
normquat[3] = v.quat.w.ToFloat32();
view[0] = v.view.x.ToFloat32();
view[1] = v.view.y.ToFloat32();
view[2] = v.view.z.ToFloat32();
if (flip_quaternion) {
normquat = -normquat;
}
}
RasterizerAccelerated::RasterizerAccelerated() {
uniform_block_data.lighting_lut_dirty.fill(true);
}
/**
* This is a helper function to resolve an issue when interpolating opposite quaternions. See below
* for a detailed description of this issue (yuriks):
*
* For any rotation, there are two quaternions Q, and -Q, that represent the same rotation. If you
* interpolate two quaternions that are opposite, instead of going from one rotation to another
* using the shortest path, you'll go around the longest path. You can test if two quaternions are
* opposite by checking if Dot(Q1, Q2) < 0. In that case, you can flip either of them, therefore
* making Dot(Q1, -Q2) positive.
*
* This solution corrects this issue per-vertex before passing the quaternions to OpenGL. This is
* correct for most cases but can still rotate around the long way sometimes. An implementation
* which did `lerp(lerp(Q1, Q2), Q3)` (with proper weighting), applying the dot product check
* between each step would work for those cases at the cost of being more complex to implement.
*
* Fortunately however, the 3DS hardware happens to also use this exact same logic to work around
* these issues, making this basic implementation actually more accurate to the hardware.
*/
static bool AreQuaternionsOpposite(Common::Vec4<Pica::float24> qa, Common::Vec4<Pica::float24> qb) {
Common::Vec4f a{qa.x.ToFloat32(), qa.y.ToFloat32(), qa.z.ToFloat32(), qa.w.ToFloat32()};
Common::Vec4f b{qb.x.ToFloat32(), qb.y.ToFloat32(), qb.z.ToFloat32(), qb.w.ToFloat32()};
return (Common::Dot(a, b) < 0.f);
}
void RasterizerAccelerated::AddTriangle(const Pica::Shader::OutputVertex& v0,
const Pica::Shader::OutputVertex& v1,
const Pica::Shader::OutputVertex& v2) {
vertex_batch.emplace_back(v0, false);
vertex_batch.emplace_back(v1, AreQuaternionsOpposite(v0.quat, v1.quat));
vertex_batch.emplace_back(v2, AreQuaternionsOpposite(v0.quat, v2.quat));
}
void RasterizerAccelerated::UpdatePagesCachedCount(PAddr addr, u32 size, int delta) {
const u32 page_start = addr >> Memory::CITRA_PAGE_BITS;
const u32 page_end = ((addr + size - 1) >> Memory::CITRA_PAGE_BITS) + 1;
u32 uncache_start_addr = 0;
u32 cache_start_addr = 0;
u32 uncache_bytes = 0;
u32 cache_bytes = 0;
for (u32 page = page_start; page != page_end; page++) {
auto& count = cached_pages.at(page);
// Ensure no overflow happens
if (delta > 0) {
ASSERT_MSG(count < std::numeric_limits<u16>::max(), "Count will overflow!");
} else if (delta < 0) {
ASSERT_MSG(count > 0, "Count will underflow!");
} else {
ASSERT_MSG(false, "Delta must be non-zero!");
}
// Adds or subtracts 1, as count is a unsigned 8-bit value
count += delta;
// Assume delta is either -1 or 1
if (count == 0) {
if (uncache_bytes == 0) {
uncache_start_addr = page << Memory::CITRA_PAGE_BITS;
}
uncache_bytes += Memory::CITRA_PAGE_SIZE;
} else if (uncache_bytes > 0) {
VideoCore::g_memory->RasterizerMarkRegionCached(uncache_start_addr, uncache_bytes,
false);
uncache_bytes = 0;
}
if (count == 1 && delta > 0) {
if (cache_bytes == 0) {
cache_start_addr = page << Memory::CITRA_PAGE_BITS;
}
cache_bytes += Memory::CITRA_PAGE_SIZE;
} else if (cache_bytes > 0) {
VideoCore::g_memory->RasterizerMarkRegionCached(cache_start_addr, cache_bytes, true);
cache_bytes = 0;
}
}
if (uncache_bytes > 0) {
VideoCore::g_memory->RasterizerMarkRegionCached(uncache_start_addr, uncache_bytes, false);
}
if (cache_bytes > 0) {
VideoCore::g_memory->RasterizerMarkRegionCached(cache_start_addr, cache_bytes, true);
}
}
void RasterizerAccelerated::ClearAll(bool flush) {
// Force flush all surfaces from the cache
if (flush) {
FlushRegion(0x0, 0xFFFFFFFF);
}
u32 uncache_start_addr = 0;
u32 uncache_bytes = 0;
for (u32 page = 0; page != cached_pages.size(); page++) {
auto& count = cached_pages.at(page);
// Assume delta is either -1 or 1
if (count != 0) {
if (uncache_bytes == 0) {
uncache_start_addr = page << Memory::CITRA_PAGE_BITS;
}
uncache_bytes += Memory::CITRA_PAGE_SIZE;
} else if (uncache_bytes > 0) {
VideoCore::g_memory->RasterizerMarkRegionCached(uncache_start_addr, uncache_bytes,
false);
uncache_bytes = 0;
}
}
if (uncache_bytes > 0) {
VideoCore::g_memory->RasterizerMarkRegionCached(uncache_start_addr, uncache_bytes, false);
}
cached_pages = {};
}
RasterizerAccelerated::VertexArrayInfo RasterizerAccelerated::AnalyzeVertexArray(bool is_indexed) {
const auto& regs = Pica::g_state.regs;
const auto& vertex_attributes = regs.pipeline.vertex_attributes;
u32 vertex_min;
u32 vertex_max;
if (is_indexed) {
const auto& index_info = regs.pipeline.index_array;
const PAddr address = vertex_attributes.GetPhysicalBaseAddress() + index_info.offset;
const u8* index_address_8 = VideoCore::g_memory->GetPhysicalPointer(address);
const u16* index_address_16 = reinterpret_cast<const u16*>(index_address_8);
const bool index_u16 = index_info.format != 0;
vertex_min = 0xFFFF;
vertex_max = 0;
const u32 size = regs.pipeline.num_vertices * (index_u16 ? 2 : 1);
FlushRegion(address, size);
for (u32 index = 0; index < regs.pipeline.num_vertices; ++index) {
const u32 vertex = index_u16 ? index_address_16[index] : index_address_8[index];
vertex_min = std::min(vertex_min, vertex);
vertex_max = std::max(vertex_max, vertex);
}
} else {
vertex_min = regs.pipeline.vertex_offset;
vertex_max = regs.pipeline.vertex_offset + regs.pipeline.num_vertices - 1;
}
const u32 vertex_num = vertex_max - vertex_min + 1;
u32 vs_input_size = 0;
for (const auto& loader : vertex_attributes.attribute_loaders) {
if (loader.component_count != 0) {
vs_input_size += Common::AlignUp(loader.byte_count * vertex_num, 4);
}
}
return {vertex_min, vertex_max, vs_input_size};
}
void RasterizerAccelerated::NotifyPicaRegisterChanged(u32 id) {
const auto& regs = Pica::g_state.regs;
switch (id) {
// Depth modifiers
case PICA_REG_INDEX(rasterizer.viewport_depth_range):
SyncDepthScale();
break;
case PICA_REG_INDEX(rasterizer.viewport_depth_near_plane):
SyncDepthOffset();
break;
// Depth buffering
case PICA_REG_INDEX(rasterizer.depthmap_enable):
shader_dirty = true;
break;
// Shadow texture
case PICA_REG_INDEX(texturing.shadow):
SyncShadowTextureBias();
break;
// Fog state
case PICA_REG_INDEX(texturing.fog_color):
SyncFogColor();
break;
case PICA_REG_INDEX(texturing.fog_lut_data[0]):
case PICA_REG_INDEX(texturing.fog_lut_data[1]):
case PICA_REG_INDEX(texturing.fog_lut_data[2]):
case PICA_REG_INDEX(texturing.fog_lut_data[3]):
case PICA_REG_INDEX(texturing.fog_lut_data[4]):
case PICA_REG_INDEX(texturing.fog_lut_data[5]):
case PICA_REG_INDEX(texturing.fog_lut_data[6]):
case PICA_REG_INDEX(texturing.fog_lut_data[7]):
uniform_block_data.fog_lut_dirty = true;
break;
// ProcTex state
case PICA_REG_INDEX(texturing.proctex):
case PICA_REG_INDEX(texturing.proctex_lut):
case PICA_REG_INDEX(texturing.proctex_lut_offset):
SyncProcTexBias();
shader_dirty = true;
break;
case PICA_REG_INDEX(texturing.proctex_noise_u):
case PICA_REG_INDEX(texturing.proctex_noise_v):
case PICA_REG_INDEX(texturing.proctex_noise_frequency):
SyncProcTexNoise();
break;
case PICA_REG_INDEX(texturing.proctex_lut_data[0]):
case PICA_REG_INDEX(texturing.proctex_lut_data[1]):
case PICA_REG_INDEX(texturing.proctex_lut_data[2]):
case PICA_REG_INDEX(texturing.proctex_lut_data[3]):
case PICA_REG_INDEX(texturing.proctex_lut_data[4]):
case PICA_REG_INDEX(texturing.proctex_lut_data[5]):
case PICA_REG_INDEX(texturing.proctex_lut_data[6]):
case PICA_REG_INDEX(texturing.proctex_lut_data[7]):
using Pica::TexturingRegs;
switch (regs.texturing.proctex_lut_config.ref_table.Value()) {
case TexturingRegs::ProcTexLutTable::Noise:
uniform_block_data.proctex_noise_lut_dirty = true;
break;
case TexturingRegs::ProcTexLutTable::ColorMap:
uniform_block_data.proctex_color_map_dirty = true;
break;
case TexturingRegs::ProcTexLutTable::AlphaMap:
uniform_block_data.proctex_alpha_map_dirty = true;
break;
case TexturingRegs::ProcTexLutTable::Color:
uniform_block_data.proctex_lut_dirty = true;
break;
case TexturingRegs::ProcTexLutTable::ColorDiff:
uniform_block_data.proctex_diff_lut_dirty = true;
break;
}
break;
// Alpha test
case PICA_REG_INDEX(framebuffer.output_merger.alpha_test):
SyncAlphaTest();
shader_dirty = true;
break;
case PICA_REG_INDEX(framebuffer.shadow):
SyncShadowBias();
break;
// Scissor test
case PICA_REG_INDEX(rasterizer.scissor_test.mode):
shader_dirty = true;
break;
case PICA_REG_INDEX(texturing.main_config):
shader_dirty = true;
break;
// Texture 0 type
case PICA_REG_INDEX(texturing.texture0.type):
shader_dirty = true;
break;
// TEV stages
// (This also syncs fog_mode and fog_flip which are part of tev_combiner_buffer_input)
case PICA_REG_INDEX(texturing.tev_stage0.color_source1):
case PICA_REG_INDEX(texturing.tev_stage0.color_modifier1):
case PICA_REG_INDEX(texturing.tev_stage0.color_op):
case PICA_REG_INDEX(texturing.tev_stage0.color_scale):
case PICA_REG_INDEX(texturing.tev_stage1.color_source1):
case PICA_REG_INDEX(texturing.tev_stage1.color_modifier1):
case PICA_REG_INDEX(texturing.tev_stage1.color_op):
case PICA_REG_INDEX(texturing.tev_stage1.color_scale):
case PICA_REG_INDEX(texturing.tev_stage2.color_source1):
case PICA_REG_INDEX(texturing.tev_stage2.color_modifier1):
case PICA_REG_INDEX(texturing.tev_stage2.color_op):
case PICA_REG_INDEX(texturing.tev_stage2.color_scale):
case PICA_REG_INDEX(texturing.tev_stage3.color_source1):
case PICA_REG_INDEX(texturing.tev_stage3.color_modifier1):
case PICA_REG_INDEX(texturing.tev_stage3.color_op):
case PICA_REG_INDEX(texturing.tev_stage3.color_scale):
case PICA_REG_INDEX(texturing.tev_stage4.color_source1):
case PICA_REG_INDEX(texturing.tev_stage4.color_modifier1):
case PICA_REG_INDEX(texturing.tev_stage4.color_op):
case PICA_REG_INDEX(texturing.tev_stage4.color_scale):
case PICA_REG_INDEX(texturing.tev_stage5.color_source1):
case PICA_REG_INDEX(texturing.tev_stage5.color_modifier1):
case PICA_REG_INDEX(texturing.tev_stage5.color_op):
case PICA_REG_INDEX(texturing.tev_stage5.color_scale):
case PICA_REG_INDEX(texturing.tev_combiner_buffer_input):
shader_dirty = true;
break;
case PICA_REG_INDEX(texturing.tev_stage0.const_r):
SyncTevConstColor(0, regs.texturing.tev_stage0);
break;
case PICA_REG_INDEX(texturing.tev_stage1.const_r):
SyncTevConstColor(1, regs.texturing.tev_stage1);
break;
case PICA_REG_INDEX(texturing.tev_stage2.const_r):
SyncTevConstColor(2, regs.texturing.tev_stage2);
break;
case PICA_REG_INDEX(texturing.tev_stage3.const_r):
SyncTevConstColor(3, regs.texturing.tev_stage3);
break;
case PICA_REG_INDEX(texturing.tev_stage4.const_r):
SyncTevConstColor(4, regs.texturing.tev_stage4);
break;
case PICA_REG_INDEX(texturing.tev_stage5.const_r):
SyncTevConstColor(5, regs.texturing.tev_stage5);
break;
// TEV combiner buffer color
case PICA_REG_INDEX(texturing.tev_combiner_buffer_color):
SyncCombinerColor();
break;
// Fragment lighting switches
case PICA_REG_INDEX(lighting.disable):
case PICA_REG_INDEX(lighting.max_light_index):
case PICA_REG_INDEX(lighting.config0):
case PICA_REG_INDEX(lighting.config1):
case PICA_REG_INDEX(lighting.abs_lut_input):
case PICA_REG_INDEX(lighting.lut_input):
case PICA_REG_INDEX(lighting.lut_scale):
case PICA_REG_INDEX(lighting.light_enable):
break;
// Fragment lighting specular 0 color
case PICA_REG_INDEX(lighting.light[0].specular_0):
SyncLightSpecular0(0);
break;
case PICA_REG_INDEX(lighting.light[1].specular_0):
SyncLightSpecular0(1);
break;
case PICA_REG_INDEX(lighting.light[2].specular_0):
SyncLightSpecular0(2);
break;
case PICA_REG_INDEX(lighting.light[3].specular_0):
SyncLightSpecular0(3);
break;
case PICA_REG_INDEX(lighting.light[4].specular_0):
SyncLightSpecular0(4);
break;
case PICA_REG_INDEX(lighting.light[5].specular_0):
SyncLightSpecular0(5);
break;
case PICA_REG_INDEX(lighting.light[6].specular_0):
SyncLightSpecular0(6);
break;
case PICA_REG_INDEX(lighting.light[7].specular_0):
SyncLightSpecular0(7);
break;
// Fragment lighting specular 1 color
case PICA_REG_INDEX(lighting.light[0].specular_1):
SyncLightSpecular1(0);
break;
case PICA_REG_INDEX(lighting.light[1].specular_1):
SyncLightSpecular1(1);
break;
case PICA_REG_INDEX(lighting.light[2].specular_1):
SyncLightSpecular1(2);
break;
case PICA_REG_INDEX(lighting.light[3].specular_1):
SyncLightSpecular1(3);
break;
case PICA_REG_INDEX(lighting.light[4].specular_1):
SyncLightSpecular1(4);
break;
case PICA_REG_INDEX(lighting.light[5].specular_1):
SyncLightSpecular1(5);
break;
case PICA_REG_INDEX(lighting.light[6].specular_1):
SyncLightSpecular1(6);
break;
case PICA_REG_INDEX(lighting.light[7].specular_1):
SyncLightSpecular1(7);
break;
// Fragment lighting diffuse color
case PICA_REG_INDEX(lighting.light[0].diffuse):
SyncLightDiffuse(0);
break;
case PICA_REG_INDEX(lighting.light[1].diffuse):
SyncLightDiffuse(1);
break;
case PICA_REG_INDEX(lighting.light[2].diffuse):
SyncLightDiffuse(2);
break;
case PICA_REG_INDEX(lighting.light[3].diffuse):
SyncLightDiffuse(3);
break;
case PICA_REG_INDEX(lighting.light[4].diffuse):
SyncLightDiffuse(4);
break;
case PICA_REG_INDEX(lighting.light[5].diffuse):
SyncLightDiffuse(5);
break;
case PICA_REG_INDEX(lighting.light[6].diffuse):
SyncLightDiffuse(6);
break;
case PICA_REG_INDEX(lighting.light[7].diffuse):
SyncLightDiffuse(7);
break;
// Fragment lighting ambient color
case PICA_REG_INDEX(lighting.light[0].ambient):
SyncLightAmbient(0);
break;
case PICA_REG_INDEX(lighting.light[1].ambient):
SyncLightAmbient(1);
break;
case PICA_REG_INDEX(lighting.light[2].ambient):
SyncLightAmbient(2);
break;
case PICA_REG_INDEX(lighting.light[3].ambient):
SyncLightAmbient(3);
break;
case PICA_REG_INDEX(lighting.light[4].ambient):
SyncLightAmbient(4);
break;
case PICA_REG_INDEX(lighting.light[5].ambient):
SyncLightAmbient(5);
break;
case PICA_REG_INDEX(lighting.light[6].ambient):
SyncLightAmbient(6);
break;
case PICA_REG_INDEX(lighting.light[7].ambient):
SyncLightAmbient(7);
break;
// Fragment lighting position
case PICA_REG_INDEX(lighting.light[0].x):
case PICA_REG_INDEX(lighting.light[0].z):
SyncLightPosition(0);
break;
case PICA_REG_INDEX(lighting.light[1].x):
case PICA_REG_INDEX(lighting.light[1].z):
SyncLightPosition(1);
break;
case PICA_REG_INDEX(lighting.light[2].x):
case PICA_REG_INDEX(lighting.light[2].z):
SyncLightPosition(2);
break;
case PICA_REG_INDEX(lighting.light[3].x):
case PICA_REG_INDEX(lighting.light[3].z):
SyncLightPosition(3);
break;
case PICA_REG_INDEX(lighting.light[4].x):
case PICA_REG_INDEX(lighting.light[4].z):
SyncLightPosition(4);
break;
case PICA_REG_INDEX(lighting.light[5].x):
case PICA_REG_INDEX(lighting.light[5].z):
SyncLightPosition(5);
break;
case PICA_REG_INDEX(lighting.light[6].x):
case PICA_REG_INDEX(lighting.light[6].z):
SyncLightPosition(6);
break;
case PICA_REG_INDEX(lighting.light[7].x):
case PICA_REG_INDEX(lighting.light[7].z):
SyncLightPosition(7);
break;
// Fragment spot lighting direction
case PICA_REG_INDEX(lighting.light[0].spot_x):
case PICA_REG_INDEX(lighting.light[0].spot_z):
SyncLightSpotDirection(0);
break;
case PICA_REG_INDEX(lighting.light[1].spot_x):
case PICA_REG_INDEX(lighting.light[1].spot_z):
SyncLightSpotDirection(1);
break;
case PICA_REG_INDEX(lighting.light[2].spot_x):
case PICA_REG_INDEX(lighting.light[2].spot_z):
SyncLightSpotDirection(2);
break;
case PICA_REG_INDEX(lighting.light[3].spot_x):
case PICA_REG_INDEX(lighting.light[3].spot_z):
SyncLightSpotDirection(3);
break;
case PICA_REG_INDEX(lighting.light[4].spot_x):
case PICA_REG_INDEX(lighting.light[4].spot_z):
SyncLightSpotDirection(4);
break;
case PICA_REG_INDEX(lighting.light[5].spot_x):
case PICA_REG_INDEX(lighting.light[5].spot_z):
SyncLightSpotDirection(5);
break;
case PICA_REG_INDEX(lighting.light[6].spot_x):
case PICA_REG_INDEX(lighting.light[6].spot_z):
SyncLightSpotDirection(6);
break;
case PICA_REG_INDEX(lighting.light[7].spot_x):
case PICA_REG_INDEX(lighting.light[7].spot_z):
SyncLightSpotDirection(7);
break;
// Fragment lighting light source config
case PICA_REG_INDEX(lighting.light[0].config):
case PICA_REG_INDEX(lighting.light[1].config):
case PICA_REG_INDEX(lighting.light[2].config):
case PICA_REG_INDEX(lighting.light[3].config):
case PICA_REG_INDEX(lighting.light[4].config):
case PICA_REG_INDEX(lighting.light[5].config):
case PICA_REG_INDEX(lighting.light[6].config):
case PICA_REG_INDEX(lighting.light[7].config):
shader_dirty = true;
break;
// Fragment lighting distance attenuation bias
case PICA_REG_INDEX(lighting.light[0].dist_atten_bias):
SyncLightDistanceAttenuationBias(0);
break;
case PICA_REG_INDEX(lighting.light[1].dist_atten_bias):
SyncLightDistanceAttenuationBias(1);
break;
case PICA_REG_INDEX(lighting.light[2].dist_atten_bias):
SyncLightDistanceAttenuationBias(2);
break;
case PICA_REG_INDEX(lighting.light[3].dist_atten_bias):
SyncLightDistanceAttenuationBias(3);
break;
case PICA_REG_INDEX(lighting.light[4].dist_atten_bias):
SyncLightDistanceAttenuationBias(4);
break;
case PICA_REG_INDEX(lighting.light[5].dist_atten_bias):
SyncLightDistanceAttenuationBias(5);
break;
case PICA_REG_INDEX(lighting.light[6].dist_atten_bias):
SyncLightDistanceAttenuationBias(6);
break;
case PICA_REG_INDEX(lighting.light[7].dist_atten_bias):
SyncLightDistanceAttenuationBias(7);
break;
// Fragment lighting distance attenuation scale
case PICA_REG_INDEX(lighting.light[0].dist_atten_scale):
SyncLightDistanceAttenuationScale(0);
break;
case PICA_REG_INDEX(lighting.light[1].dist_atten_scale):
SyncLightDistanceAttenuationScale(1);
break;
case PICA_REG_INDEX(lighting.light[2].dist_atten_scale):
SyncLightDistanceAttenuationScale(2);
break;
case PICA_REG_INDEX(lighting.light[3].dist_atten_scale):
SyncLightDistanceAttenuationScale(3);
break;
case PICA_REG_INDEX(lighting.light[4].dist_atten_scale):
SyncLightDistanceAttenuationScale(4);
break;
case PICA_REG_INDEX(lighting.light[5].dist_atten_scale):
SyncLightDistanceAttenuationScale(5);
break;
case PICA_REG_INDEX(lighting.light[6].dist_atten_scale):
SyncLightDistanceAttenuationScale(6);
break;
case PICA_REG_INDEX(lighting.light[7].dist_atten_scale):
SyncLightDistanceAttenuationScale(7);
break;
// Fragment lighting global ambient color (emission + ambient * ambient)
case PICA_REG_INDEX(lighting.global_ambient):
SyncGlobalAmbient();
break;
// Fragment lighting lookup tables
case PICA_REG_INDEX(lighting.lut_data[0]):
case PICA_REG_INDEX(lighting.lut_data[1]):
case PICA_REG_INDEX(lighting.lut_data[2]):
case PICA_REG_INDEX(lighting.lut_data[3]):
case PICA_REG_INDEX(lighting.lut_data[4]):
case PICA_REG_INDEX(lighting.lut_data[5]):
case PICA_REG_INDEX(lighting.lut_data[6]):
case PICA_REG_INDEX(lighting.lut_data[7]): {
const auto& lut_config = regs.lighting.lut_config;
uniform_block_data.lighting_lut_dirty[lut_config.type] = true;
uniform_block_data.lighting_lut_dirty_any = true;
break;
}
default:
// Forward registers that map to fixed function API features to the video backend
NotifyFixedFunctionPicaRegisterChanged(id);
}
}
void RasterizerAccelerated::SyncDepthScale() {
float depth_scale =
Pica::float24::FromRaw(Pica::g_state.regs.rasterizer.viewport_depth_range).ToFloat32();
if (depth_scale != uniform_block_data.data.depth_scale) {
uniform_block_data.data.depth_scale = depth_scale;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncDepthOffset() {
float depth_offset =
Pica::float24::FromRaw(Pica::g_state.regs.rasterizer.viewport_depth_near_plane).ToFloat32();
if (depth_offset != uniform_block_data.data.depth_offset) {
uniform_block_data.data.depth_offset = depth_offset;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncFogColor() {
const auto& regs = Pica::g_state.regs;
uniform_block_data.data.fog_color = {
regs.texturing.fog_color.r.Value() / 255.0f,
regs.texturing.fog_color.g.Value() / 255.0f,
regs.texturing.fog_color.b.Value() / 255.0f,
};
uniform_block_data.dirty = true;
}
void RasterizerAccelerated::SyncProcTexNoise() {
const auto& regs = Pica::g_state.regs.texturing;
uniform_block_data.data.proctex_noise_f = {
Pica::float16::FromRaw(regs.proctex_noise_frequency.u).ToFloat32(),
Pica::float16::FromRaw(regs.proctex_noise_frequency.v).ToFloat32(),
};
uniform_block_data.data.proctex_noise_a = {
regs.proctex_noise_u.amplitude / 4095.0f,
regs.proctex_noise_v.amplitude / 4095.0f,
};
uniform_block_data.data.proctex_noise_p = {
Pica::float16::FromRaw(regs.proctex_noise_u.phase).ToFloat32(),
Pica::float16::FromRaw(regs.proctex_noise_v.phase).ToFloat32(),
};
uniform_block_data.dirty = true;
}
void RasterizerAccelerated::SyncProcTexBias() {
const auto& regs = Pica::g_state.regs.texturing;
uniform_block_data.data.proctex_bias =
Pica::float16::FromRaw(regs.proctex.bias_low | (regs.proctex_lut.bias_high << 8))
.ToFloat32();
uniform_block_data.dirty = true;
}
void RasterizerAccelerated::SyncAlphaTest() {
const auto& regs = Pica::g_state.regs;
if (regs.framebuffer.output_merger.alpha_test.ref != uniform_block_data.data.alphatest_ref) {
uniform_block_data.data.alphatest_ref = regs.framebuffer.output_merger.alpha_test.ref;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncCombinerColor() {
auto combiner_color = ColorRGBA8(Pica::g_state.regs.texturing.tev_combiner_buffer_color.raw);
if (combiner_color != uniform_block_data.data.tev_combiner_buffer_color) {
uniform_block_data.data.tev_combiner_buffer_color = combiner_color;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncTevConstColor(
std::size_t stage_index, const Pica::TexturingRegs::TevStageConfig& tev_stage) {
const auto const_color = ColorRGBA8(tev_stage.const_color);
if (const_color == uniform_block_data.data.const_color[stage_index]) {
return;
}
uniform_block_data.data.const_color[stage_index] = const_color;
uniform_block_data.dirty = true;
}
void RasterizerAccelerated::SyncGlobalAmbient() {
auto color = LightColor(Pica::g_state.regs.lighting.global_ambient);
if (color != uniform_block_data.data.lighting_global_ambient) {
uniform_block_data.data.lighting_global_ambient = color;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncLightSpecular0(int light_index) {
auto color = LightColor(Pica::g_state.regs.lighting.light[light_index].specular_0);
if (color != uniform_block_data.data.light_src[light_index].specular_0) {
uniform_block_data.data.light_src[light_index].specular_0 = color;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncLightSpecular1(int light_index) {
auto color = LightColor(Pica::g_state.regs.lighting.light[light_index].specular_1);
if (color != uniform_block_data.data.light_src[light_index].specular_1) {
uniform_block_data.data.light_src[light_index].specular_1 = color;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncLightDiffuse(int light_index) {
auto color = LightColor(Pica::g_state.regs.lighting.light[light_index].diffuse);
if (color != uniform_block_data.data.light_src[light_index].diffuse) {
uniform_block_data.data.light_src[light_index].diffuse = color;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncLightAmbient(int light_index) {
auto color = LightColor(Pica::g_state.regs.lighting.light[light_index].ambient);
if (color != uniform_block_data.data.light_src[light_index].ambient) {
uniform_block_data.data.light_src[light_index].ambient = color;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncLightPosition(int light_index) {
const Common::Vec3f position = {
Pica::float16::FromRaw(Pica::g_state.regs.lighting.light[light_index].x).ToFloat32(),
Pica::float16::FromRaw(Pica::g_state.regs.lighting.light[light_index].y).ToFloat32(),
Pica::float16::FromRaw(Pica::g_state.regs.lighting.light[light_index].z).ToFloat32()};
if (position != uniform_block_data.data.light_src[light_index].position) {
uniform_block_data.data.light_src[light_index].position = position;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncLightSpotDirection(int light_index) {
const auto& light = Pica::g_state.regs.lighting.light[light_index];
const auto spot_direction =
Common::Vec3f{light.spot_x / 2047.0f, light.spot_y / 2047.0f, light.spot_z / 2047.0f};
if (spot_direction != uniform_block_data.data.light_src[light_index].spot_direction) {
uniform_block_data.data.light_src[light_index].spot_direction = spot_direction;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncLightDistanceAttenuationBias(int light_index) {
float dist_atten_bias =
Pica::float20::FromRaw(Pica::g_state.regs.lighting.light[light_index].dist_atten_bias)
.ToFloat32();
if (dist_atten_bias != uniform_block_data.data.light_src[light_index].dist_atten_bias) {
uniform_block_data.data.light_src[light_index].dist_atten_bias = dist_atten_bias;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncLightDistanceAttenuationScale(int light_index) {
float dist_atten_scale =
Pica::float20::FromRaw(Pica::g_state.regs.lighting.light[light_index].dist_atten_scale)
.ToFloat32();
if (dist_atten_scale != uniform_block_data.data.light_src[light_index].dist_atten_scale) {
uniform_block_data.data.light_src[light_index].dist_atten_scale = dist_atten_scale;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncShadowBias() {
const auto& shadow = Pica::g_state.regs.framebuffer.shadow;
float constant = Pica::float16::FromRaw(shadow.constant).ToFloat32();
float linear = Pica::float16::FromRaw(shadow.linear).ToFloat32();
if (constant != uniform_block_data.data.shadow_bias_constant ||
linear != uniform_block_data.data.shadow_bias_linear) {
uniform_block_data.data.shadow_bias_constant = constant;
uniform_block_data.data.shadow_bias_linear = linear;
uniform_block_data.dirty = true;
}
}
void RasterizerAccelerated::SyncShadowTextureBias() {
int bias = Pica::g_state.regs.texturing.shadow.bias << 1;
if (bias != uniform_block_data.data.shadow_texture_bias) {
uniform_block_data.data.shadow_texture_bias = bias;
uniform_block_data.dirty = true;
}
}
} // namespace VideoCore

View File

@ -0,0 +1,142 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include "common/vector_math.h"
#include "video_core/rasterizer_interface.h"
#include "video_core/regs_texturing.h"
#include "video_core/shader/shader_uniforms.h"
namespace VideoCore {
class RasterizerAccelerated : public RasterizerInterface {
public:
RasterizerAccelerated();
virtual ~RasterizerAccelerated() = default;
void AddTriangle(const Pica::Shader::OutputVertex& v0, const Pica::Shader::OutputVertex& v1,
const Pica::Shader::OutputVertex& v2) override;
void UpdatePagesCachedCount(PAddr addr, u32 size, int delta) override;
void NotifyPicaRegisterChanged(u32 id) override;
void ClearAll(bool flush) override;
protected:
/// Notifies that a fixed function PICA register changed to the video backend
virtual void NotifyFixedFunctionPicaRegisterChanged(u32 id) = 0;
/// Syncs the depth scale to match the PICA register
void SyncDepthScale();
/// Syncs the depth offset to match the PICA register
void SyncDepthOffset();
/// Syncs the fog states to match the PICA register
void SyncFogColor();
/// Sync the procedural texture noise configuration to match the PICA register
void SyncProcTexNoise();
/// Sync the procedural texture bias configuration to match the PICA register
void SyncProcTexBias();
/// Syncs the alpha test states to match the PICA register
void SyncAlphaTest();
/// Syncs the TEV combiner color buffer to match the PICA register
void SyncCombinerColor();
/// Syncs the TEV constant color to match the PICA register
void SyncTevConstColor(std::size_t tev_index,
const Pica::TexturingRegs::TevStageConfig& tev_stage);
/// Syncs the lighting global ambient color to match the PICA register
void SyncGlobalAmbient();
/// Syncs the specified light's specular 0 color to match the PICA register
void SyncLightSpecular0(int light_index);
/// Syncs the specified light's specular 1 color to match the PICA register
void SyncLightSpecular1(int light_index);
/// Syncs the specified light's diffuse color to match the PICA register
void SyncLightDiffuse(int light_index);
/// Syncs the specified light's ambient color to match the PICA register
void SyncLightAmbient(int light_index);
/// Syncs the specified light's position to match the PICA register
void SyncLightPosition(int light_index);
/// Syncs the specified spot light direcition to match the PICA register
void SyncLightSpotDirection(int light_index);
/// Syncs the specified light's distance attenuation bias to match the PICA register
void SyncLightDistanceAttenuationBias(int light_index);
/// Syncs the specified light's distance attenuation scale to match the PICA register
void SyncLightDistanceAttenuationScale(int light_index);
/// Syncs the shadow rendering bias to match the PICA register
void SyncShadowBias();
/// Syncs the shadow texture bias to match the PICA register
void SyncShadowTextureBias();
protected:
/// Structure that keeps tracks of the uniform state
struct UniformBlockData {
Pica::Shader::UniformData data{};
std::array<bool, Pica::LightingRegs::NumLightingSampler> lighting_lut_dirty{};
bool lighting_lut_dirty_any = true;
bool fog_lut_dirty = true;
bool proctex_noise_lut_dirty = true;
bool proctex_color_map_dirty = true;
bool proctex_alpha_map_dirty = true;
bool proctex_lut_dirty = true;
bool proctex_diff_lut_dirty = true;
bool dirty = true;
};
/// Structure that the hardware rendered vertices are composed of
struct HardwareVertex {
HardwareVertex() = default;
HardwareVertex(const Pica::Shader::OutputVertex& v, bool flip_quaternion);
Common::Vec4f position;
Common::Vec4f color;
Common::Vec2f tex_coord0;
Common::Vec2f tex_coord1;
Common::Vec2f tex_coord2;
float tex_coord0_w;
Common::Vec4f normquat;
Common::Vec3f view;
};
struct VertexArrayInfo {
u32 vs_input_index_min;
u32 vs_input_index_max;
u32 vs_input_size;
};
/// Retrieve the range and the size of the input vertex
VertexArrayInfo AnalyzeVertexArray(bool is_indexed);
protected:
std::array<u16, 0x30000> cached_pages{};
std::vector<HardwareVertex> vertex_batch;
bool shader_dirty = true;
UniformBlockData uniform_block_data{};
std::array<std::array<Common::Vec2f, 256>, Pica::LightingRegs::NumLightingSampler>
lighting_lut_data{};
std::array<Common::Vec2f, 128> fog_lut_data{};
std::array<Common::Vec2f, 128> proctex_noise_lut_data{};
std::array<Common::Vec2f, 128> proctex_color_map_data{};
std::array<Common::Vec2f, 128> proctex_alpha_map_data{};
std::array<Common::Vec4f, 256> proctex_lut_data{};
std::array<Common::Vec4f, 256> proctex_diff_lut_data{};
};
} // namespace VideoCore

View File

@ -1,480 +0,0 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include "common/microprofile.h"
#include "common/scope_exit.h"
#include "common/texture.h"
#include "core/core.h"
#include "video_core/rasterizer_cache/cached_surface.h"
#include "video_core/rasterizer_cache/morton_swizzle.h"
#include "video_core/rasterizer_cache/rasterizer_cache.h"
#include "video_core/renderer_opengl/gl_state.h"
#include "video_core/renderer_opengl/texture_downloader_es.h"
#include "video_core/renderer_opengl/texture_filters/texture_filterer.h"
namespace OpenGL {
static Aspect ToAspect(SurfaceType type) {
switch (type) {
case SurfaceType::Color:
case SurfaceType::Texture:
case SurfaceType::Fill:
return Aspect::Color;
case SurfaceType::Depth:
return Aspect::Depth;
case SurfaceType::DepthStencil:
return Aspect::DepthStencil;
default:
LOG_CRITICAL(Render_OpenGL, "Unknown SurfaceType {}", type);
UNREACHABLE();
}
return Aspect::Color;
}
CachedSurface::~CachedSurface() {
if (texture.handle) {
auto tag = is_custom ? HostTextureTag{GetFormatTuple(PixelFormat::RGBA8),
custom_tex_info.width, custom_tex_info.height}
: HostTextureTag{GetFormatTuple(pixel_format), GetScaledWidth(),
GetScaledHeight()};
owner.host_texture_recycler.emplace(tag, std::move(texture));
}
}
MICROPROFILE_DEFINE(RasterizerCache_SurfaceLoad, "RasterizerCache", "Surface Load",
MP_RGB(128, 192, 64));
void CachedSurface::LoadGLBuffer(PAddr load_start, PAddr load_end) {
ASSERT(type != SurfaceType::Fill);
const bool need_swap =
GLES && (pixel_format == PixelFormat::RGBA8 || pixel_format == PixelFormat::RGB8);
const u8* const texture_src_data = VideoCore::g_memory->GetPhysicalPointer(addr);
if (texture_src_data == nullptr)
return;
if (gl_buffer.empty()) {
gl_buffer.resize(width * height * GetBytesPerPixel(pixel_format));
}
// TODO: Should probably be done in ::Memory:: and check for other regions too
if (load_start < Memory::VRAM_VADDR_END && load_end > Memory::VRAM_VADDR_END)
load_end = Memory::VRAM_VADDR_END;
if (load_start < Memory::VRAM_VADDR && load_end > Memory::VRAM_VADDR)
load_start = Memory::VRAM_VADDR;
MICROPROFILE_SCOPE(RasterizerCache_SurfaceLoad);
ASSERT(load_start >= addr && load_end <= end);
const u32 start_offset = load_start - addr;
if (!is_tiled) {
ASSERT(type == SurfaceType::Color);
if (need_swap) {
// TODO(liushuyu): check if the byteswap here is 100% correct
// cannot fully test this
if (pixel_format == PixelFormat::RGBA8) {
for (std::size_t i = start_offset; i < load_end - addr; i += 4) {
gl_buffer[i] = texture_src_data[i + 3];
gl_buffer[i + 1] = texture_src_data[i + 2];
gl_buffer[i + 2] = texture_src_data[i + 1];
gl_buffer[i + 3] = texture_src_data[i];
}
} else if (pixel_format == PixelFormat::RGB8) {
for (std::size_t i = start_offset; i < load_end - addr; i += 3) {
gl_buffer[i] = texture_src_data[i + 2];
gl_buffer[i + 1] = texture_src_data[i + 1];
gl_buffer[i + 2] = texture_src_data[i];
}
}
} else {
std::memcpy(&gl_buffer[start_offset], texture_src_data + start_offset,
load_end - load_start);
}
} else {
if (type == SurfaceType::Texture) {
Pica::Texture::TextureInfo tex_info{};
tex_info.width = width;
tex_info.height = height;
tex_info.format = static_cast<Pica::TexturingRegs::TextureFormat>(pixel_format);
tex_info.SetDefaultStride();
tex_info.physical_address = addr;
const SurfaceInterval load_interval(load_start, load_end);
const auto rect = GetSubRect(FromInterval(load_interval));
ASSERT(FromInterval(load_interval).GetInterval() == load_interval);
for (unsigned y = rect.bottom; y < rect.top; ++y) {
for (unsigned x = rect.left; x < rect.right; ++x) {
auto vec4 =
Pica::Texture::LookupTexture(texture_src_data, x, height - 1 - y, tex_info);
const std::size_t offset = (x + (width * y)) * 4;
std::memcpy(&gl_buffer[offset], vec4.AsArray(), 4);
}
}
} else {
morton_to_gl_fns[static_cast<std::size_t>(pixel_format)](stride, height, &gl_buffer[0],
addr, load_start, load_end);
}
}
}
MICROPROFILE_DEFINE(RasterizerCache_SurfaceFlush, "RasterizerCache", "Surface Flush",
MP_RGB(128, 192, 64));
void CachedSurface::FlushGLBuffer(PAddr flush_start, PAddr flush_end) {
u8* const dst_buffer = VideoCore::g_memory->GetPhysicalPointer(addr);
if (dst_buffer == nullptr)
return;
ASSERT(gl_buffer.size() == width * height * GetBytesPerPixel(pixel_format));
// TODO: Should probably be done in ::Memory:: and check for other regions too
// same as loadglbuffer()
if (flush_start < Memory::VRAM_VADDR_END && flush_end > Memory::VRAM_VADDR_END)
flush_end = Memory::VRAM_VADDR_END;
if (flush_start < Memory::VRAM_VADDR && flush_end > Memory::VRAM_VADDR)
flush_start = Memory::VRAM_VADDR;
MICROPROFILE_SCOPE(RasterizerCache_SurfaceFlush);
ASSERT(flush_start >= addr && flush_end <= end);
const u32 start_offset = flush_start - addr;
const u32 end_offset = flush_end - addr;
if (type == SurfaceType::Fill) {
const u32 coarse_start_offset = start_offset - (start_offset % fill_size);
const u32 backup_bytes = start_offset % fill_size;
std::array<u8, 4> backup_data;
if (backup_bytes)
std::memcpy(&backup_data[0], &dst_buffer[coarse_start_offset], backup_bytes);
for (u32 offset = coarse_start_offset; offset < end_offset; offset += fill_size) {
std::memcpy(&dst_buffer[offset], &fill_data[0],
std::min(fill_size, end_offset - offset));
}
if (backup_bytes)
std::memcpy(&dst_buffer[coarse_start_offset], &backup_data[0], backup_bytes);
} else if (!is_tiled) {
ASSERT(type == SurfaceType::Color);
if (pixel_format == PixelFormat::RGBA8 && GLES) {
for (std::size_t i = start_offset; i < flush_end - addr; i += 4) {
dst_buffer[i] = gl_buffer[i + 3];
dst_buffer[i + 1] = gl_buffer[i + 2];
dst_buffer[i + 2] = gl_buffer[i + 1];
dst_buffer[i + 3] = gl_buffer[i];
}
} else if (pixel_format == PixelFormat::RGB8 && GLES) {
for (std::size_t i = start_offset; i < flush_end - addr; i += 3) {
dst_buffer[i] = gl_buffer[i + 2];
dst_buffer[i + 1] = gl_buffer[i + 1];
dst_buffer[i + 2] = gl_buffer[i];
}
} else {
std::memcpy(dst_buffer + start_offset, &gl_buffer[start_offset],
flush_end - flush_start);
}
} else {
gl_to_morton_fns[static_cast<std::size_t>(pixel_format)](stride, height, &gl_buffer[0],
addr, flush_start, flush_end);
}
}
bool CachedSurface::LoadCustomTexture(u64 tex_hash) {
auto& custom_tex_cache = Core::System::GetInstance().CustomTexCache();
const auto& image_interface = Core::System::GetInstance().GetImageInterface();
if (custom_tex_cache.IsTextureCached(tex_hash)) {
custom_tex_info = custom_tex_cache.LookupTexture(tex_hash);
return true;
}
if (!custom_tex_cache.CustomTextureExists(tex_hash)) {
return false;
}
const auto& path_info = custom_tex_cache.LookupTexturePathInfo(tex_hash);
if (!image_interface->DecodePNG(custom_tex_info.tex, custom_tex_info.width,
custom_tex_info.height, path_info.path)) {
LOG_ERROR(Render_OpenGL, "Failed to load custom texture {}", path_info.path);
return false;
}
const std::bitset<32> width_bits(custom_tex_info.width);
const std::bitset<32> height_bits(custom_tex_info.height);
if (width_bits.count() != 1 || height_bits.count() != 1) {
LOG_ERROR(Render_OpenGL, "Texture {} size is not a power of 2", path_info.path);
return false;
}
LOG_DEBUG(Render_OpenGL, "Loaded custom texture from {}", path_info.path);
Common::FlipRGBA8Texture(custom_tex_info.tex, custom_tex_info.width, custom_tex_info.height);
custom_tex_cache.CacheTexture(tex_hash, custom_tex_info.tex, custom_tex_info.width,
custom_tex_info.height);
return true;
}
void CachedSurface::DumpTexture(GLuint target_tex, u64 tex_hash) {
// Make sure the texture size is a power of 2
// If not, the surface is actually a framebuffer
std::bitset<32> width_bits(width);
std::bitset<32> height_bits(height);
if (width_bits.count() != 1 || height_bits.count() != 1) {
LOG_WARNING(Render_OpenGL, "Not dumping {:016X} because size isn't a power of 2 ({}x{})",
tex_hash, width, height);
return;
}
// Dump texture to RGBA8 and encode as PNG
const auto& image_interface = Core::System::GetInstance().GetImageInterface();
auto& custom_tex_cache = Core::System::GetInstance().CustomTexCache();
std::string dump_path =
fmt::format("{}textures/{:016X}/", FileUtil::GetUserPath(FileUtil::UserPath::DumpDir),
Core::System::GetInstance().Kernel().GetCurrentProcess()->codeset->program_id);
if (!FileUtil::CreateFullPath(dump_path)) {
LOG_ERROR(Render, "Unable to create {}", dump_path);
return;
}
dump_path += fmt::format("tex1_{}x{}_{:016X}_{}.png", width, height, tex_hash, pixel_format);
if (!custom_tex_cache.IsTextureDumped(tex_hash) && !FileUtil::Exists(dump_path)) {
custom_tex_cache.SetTextureDumped(tex_hash);
LOG_INFO(Render_OpenGL, "Dumping texture to {}", dump_path);
std::vector<u8> decoded_texture;
decoded_texture.resize(width * height * 4);
OpenGLState state = OpenGLState::GetCurState();
GLuint old_texture = state.texture_units[0].texture_2d;
state.Apply();
/*
GetTexImageOES is used even if not using OpenGL ES to work around a small issue that
happens if using custom textures with texture dumping at the same.
Let's say there's 2 textures that are both 32x32 and one of them gets replaced with a
higher quality 256x256 texture. If the 256x256 texture is displayed first and the
32x32 texture gets uploaded to the same underlying OpenGL texture, the 32x32 texture
will appear in the corner of the 256x256 texture. If texture dumping is enabled and
the 32x32 is undumped, Citra will attempt to dump it. Since the underlying OpenGL
texture is still 256x256, Citra crashes because it thinks the texture is only 32x32.
GetTexImageOES conveniently only dumps the specified region, and works on both
desktop and ES.
*/
owner.texture_downloader_es->GetTexImage(GL_TEXTURE_2D, 0, GL_RGBA, GL_UNSIGNED_BYTE,
height, width, &decoded_texture[0]);
state.texture_units[0].texture_2d = old_texture;
state.Apply();
Common::FlipRGBA8Texture(decoded_texture, width, height);
if (!image_interface->EncodePNG(dump_path, decoded_texture, width, height))
LOG_ERROR(Render_OpenGL, "Failed to save decoded texture");
}
}
MICROPROFILE_DEFINE(RasterizerCache_TextureUL, "RasterizerCache", "Texture Upload",
MP_RGB(128, 192, 64));
void CachedSurface::UploadGLTexture(Common::Rectangle<u32> rect) {
if (type == SurfaceType::Fill) {
return;
}
MICROPROFILE_SCOPE(RasterizerCache_TextureUL);
ASSERT(gl_buffer.size() == width * height * GetBytesPerPixel(pixel_format));
u64 tex_hash = 0;
if (Settings::values.dump_textures || Settings::values.custom_textures) {
tex_hash = Common::ComputeHash64(gl_buffer.data(), gl_buffer.size());
}
if (Settings::values.custom_textures) {
is_custom = LoadCustomTexture(tex_hash);
}
// Load data from memory to the surface
GLint x0 = static_cast<GLint>(rect.left);
GLint y0 = static_cast<GLint>(rect.bottom);
std::size_t buffer_offset = (y0 * stride + x0) * GetBytesPerPixel(pixel_format);
const FormatTuple& tuple = GetFormatTuple(pixel_format);
GLuint target_tex = texture.handle;
// If not 1x scale, create 1x texture that we will blit from to replace texture subrect in
// surface
OGLTexture unscaled_tex;
if (res_scale != 1) {
x0 = 0;
y0 = 0;
if (is_custom) {
const auto& tuple = GetFormatTuple(PixelFormat::RGBA8);
unscaled_tex =
owner.AllocateSurfaceTexture(tuple, custom_tex_info.width, custom_tex_info.height);
} else {
unscaled_tex = owner.AllocateSurfaceTexture(tuple, rect.GetWidth(), rect.GetHeight());
}
target_tex = unscaled_tex.handle;
}
OpenGLState cur_state = OpenGLState::GetCurState();
GLuint old_tex = cur_state.texture_units[0].texture_2d;
cur_state.texture_units[0].texture_2d = target_tex;
cur_state.Apply();
// Ensure no bad interactions with GL_UNPACK_ALIGNMENT
ASSERT(stride * GetBytesPerPixel(pixel_format) % 4 == 0);
if (is_custom) {
if (res_scale == 1) {
texture = owner.AllocateSurfaceTexture(GetFormatTuple(PixelFormat::RGBA8),
custom_tex_info.width, custom_tex_info.height);
cur_state.texture_units[0].texture_2d = texture.handle;
cur_state.Apply();
}
// Always going to be using rgba8
glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(custom_tex_info.width));
glActiveTexture(GL_TEXTURE0);
glTexSubImage2D(GL_TEXTURE_2D, 0, x0, y0, custom_tex_info.width, custom_tex_info.height,
GL_RGBA, GL_UNSIGNED_BYTE, custom_tex_info.tex.data());
} else {
glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(stride));
glActiveTexture(GL_TEXTURE0);
glTexSubImage2D(GL_TEXTURE_2D, 0, x0, y0, static_cast<GLsizei>(rect.GetWidth()),
static_cast<GLsizei>(rect.GetHeight()), tuple.format, tuple.type,
&gl_buffer[buffer_offset]);
}
glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
if (Settings::values.dump_textures && !is_custom) {
DumpTexture(target_tex, tex_hash);
}
cur_state.texture_units[0].texture_2d = old_tex;
cur_state.Apply();
if (res_scale != 1) {
auto scaled_rect = rect;
scaled_rect.left *= res_scale;
scaled_rect.top *= res_scale;
scaled_rect.right *= res_scale;
scaled_rect.bottom *= res_scale;
const u32 width = is_custom ? custom_tex_info.width : rect.GetWidth();
const u32 height = is_custom ? custom_tex_info.height : rect.GetHeight();
const Common::Rectangle<u32> from_rect{0, height, width, 0};
if (is_custom ||
!owner.texture_filterer->Filter(unscaled_tex, from_rect, texture, scaled_rect, type)) {
const Aspect aspect = ToAspect(type);
runtime.BlitTextures(unscaled_tex, {aspect, from_rect}, texture, {aspect, scaled_rect});
}
}
InvalidateAllWatcher();
}
MICROPROFILE_DEFINE(RasterizerCache_TextureDL, "RasterizerCache", "Texture Download",
MP_RGB(128, 192, 64));
void CachedSurface::DownloadGLTexture(const Common::Rectangle<u32>& rect) {
if (type == SurfaceType::Fill) {
return;
}
MICROPROFILE_SCOPE(RasterizerCache_TextureDL);
if (gl_buffer.empty()) {
gl_buffer.resize(width * height * GetBytesPerPixel(pixel_format));
}
OpenGLState state = OpenGLState::GetCurState();
OpenGLState prev_state = state;
SCOPE_EXIT({ prev_state.Apply(); });
const FormatTuple& tuple = GetFormatTuple(pixel_format);
// Ensure no bad interactions with GL_PACK_ALIGNMENT
ASSERT(stride * GetBytesPerPixel(pixel_format) % 4 == 0);
glPixelStorei(GL_PACK_ROW_LENGTH, static_cast<GLint>(stride));
const std::size_t buffer_offset =
(rect.bottom * stride + rect.left) * GetBytesPerPixel(pixel_format);
// If not 1x scale, blit scaled texture to a new 1x texture and use that to flush
const Aspect aspect = ToAspect(type);
if (res_scale != 1) {
auto scaled_rect = rect;
scaled_rect.left *= res_scale;
scaled_rect.top *= res_scale;
scaled_rect.right *= res_scale;
scaled_rect.bottom *= res_scale;
const Common::Rectangle<u32> unscaled_tex_rect{0, rect.GetHeight(), rect.GetWidth(), 0};
auto unscaled_tex = owner.AllocateSurfaceTexture(tuple, rect.GetWidth(), rect.GetHeight());
// Blit scaled texture to the unscaled one
runtime.BlitTextures(texture, {aspect, scaled_rect}, unscaled_tex,
{aspect, unscaled_tex_rect});
state.texture_units[0].texture_2d = unscaled_tex.handle;
state.Apply();
glActiveTexture(GL_TEXTURE0);
if (GLES) {
owner.texture_downloader_es->GetTexImage(GL_TEXTURE_2D, 0, tuple.format, tuple.type,
rect.GetHeight(), rect.GetWidth(),
&gl_buffer[buffer_offset]);
} else {
glGetTexImage(GL_TEXTURE_2D, 0, tuple.format, tuple.type, &gl_buffer[buffer_offset]);
}
} else {
runtime.ReadTexture(texture, {aspect, rect}, tuple, gl_buffer.data());
}
glPixelStorei(GL_PACK_ROW_LENGTH, 0);
}
bool CachedSurface::CanFill(const SurfaceParams& dest_surface,
SurfaceInterval fill_interval) const {
if (type == SurfaceType::Fill && IsRegionValid(fill_interval) &&
boost::icl::first(fill_interval) >= addr &&
boost::icl::last_next(fill_interval) <= end && // dest_surface is within our fill range
dest_surface.FromInterval(fill_interval).GetInterval() ==
fill_interval) { // make sure interval is a rectangle in dest surface
if (fill_size * 8 != dest_surface.GetFormatBpp()) {
// Check if bits repeat for our fill_size
const u32 dest_bytes_per_pixel = std::max(dest_surface.GetFormatBpp() / 8, 1u);
std::vector<u8> fill_test(fill_size * dest_bytes_per_pixel);
for (u32 i = 0; i < dest_bytes_per_pixel; ++i)
std::memcpy(&fill_test[i * fill_size], &fill_data[0], fill_size);
for (u32 i = 0; i < fill_size; ++i)
if (std::memcmp(&fill_test[dest_bytes_per_pixel * i], &fill_test[0],
dest_bytes_per_pixel) != 0)
return false;
if (dest_surface.GetFormatBpp() == 4 && (fill_test[0] & 0xF) != (fill_test[0] >> 4))
return false;
}
return true;
}
return false;
}
bool CachedSurface::CanCopy(const SurfaceParams& dest_surface,
SurfaceInterval copy_interval) const {
SurfaceParams subrect_params = dest_surface.FromInterval(copy_interval);
ASSERT(subrect_params.GetInterval() == copy_interval);
if (CanSubRect(subrect_params))
return true;
if (CanFill(dest_surface, copy_interval))
return true;
return false;
}
} // namespace OpenGL

View File

@ -1,137 +0,0 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <list>
#include "common/assert.h"
#include "core/custom_tex_cache.h"
#include "video_core/rasterizer_cache/surface_params.h"
#include "video_core/rasterizer_cache/texture_runtime.h"
namespace OpenGL {
/**
* A watcher that notifies whether a cached surface has been changed. This is useful for caching
* surface collection objects, including texture cube and mipmap.
*/
class SurfaceWatcher {
friend class CachedSurface;
public:
explicit SurfaceWatcher(std::weak_ptr<CachedSurface>&& surface) : surface(std::move(surface)) {}
/// Checks whether the surface has been changed.
bool IsValid() const {
return !surface.expired() && valid;
}
/// Marks that the content of the referencing surface has been updated to the watcher user.
void Validate() {
ASSERT(!surface.expired());
valid = true;
}
/// Gets the referencing surface. Returns null if the surface has been destroyed
Surface Get() const {
return surface.lock();
}
private:
std::weak_ptr<CachedSurface> surface;
bool valid = false;
};
class RasterizerCacheOpenGL;
class CachedSurface : public SurfaceParams, public std::enable_shared_from_this<CachedSurface> {
public:
CachedSurface(SurfaceParams params, RasterizerCacheOpenGL& owner, TextureRuntime& runtime)
: SurfaceParams(params), owner(owner), runtime(runtime) {}
~CachedSurface();
/// Read/Write data in 3DS memory to/from gl_buffer
void LoadGLBuffer(PAddr load_start, PAddr load_end);
void FlushGLBuffer(PAddr flush_start, PAddr flush_end);
/// Custom texture loading and dumping
bool LoadCustomTexture(u64 tex_hash);
void DumpTexture(GLuint target_tex, u64 tex_hash);
/// Upload/Download data in gl_buffer in/to this surface's texture
void UploadGLTexture(Common::Rectangle<u32> rect);
void DownloadGLTexture(const Common::Rectangle<u32>& rect);
bool CanFill(const SurfaceParams& dest_surface, SurfaceInterval fill_interval) const;
bool CanCopy(const SurfaceParams& dest_surface, SurfaceInterval copy_interval) const;
bool IsRegionValid(SurfaceInterval interval) const {
return (invalid_regions.find(interval) == invalid_regions.end());
}
bool IsSurfaceFullyInvalid() const {
auto interval = GetInterval();
return *invalid_regions.equal_range(interval).first == interval;
}
std::shared_ptr<SurfaceWatcher> CreateWatcher() {
auto watcher = std::make_shared<SurfaceWatcher>(weak_from_this());
watchers.push_front(watcher);
return watcher;
}
void InvalidateAllWatcher() {
for (const auto& watcher : watchers) {
if (auto locked = watcher.lock()) {
locked->valid = false;
}
}
}
void UnlinkAllWatcher() {
for (const auto& watcher : watchers) {
if (auto locked = watcher.lock()) {
locked->valid = false;
locked->surface.reset();
}
}
watchers.clear();
}
public:
bool registered = false;
SurfaceRegions invalid_regions;
std::vector<u8> gl_buffer;
// Number of bytes to read from fill_data
u32 fill_size = 0;
std::array<u8, 4> fill_data;
OGLTexture texture;
// level_watchers[i] watches the (i+1)-th level mipmap source surface
std::array<std::shared_ptr<SurfaceWatcher>, 7> level_watchers;
u32 max_level = 0;
// Information about custom textures
bool is_custom = false;
Core::CustomTexInfo custom_tex_info;
private:
RasterizerCacheOpenGL& owner;
TextureRuntime& runtime;
std::list<std::weak_ptr<SurfaceWatcher>> watchers;
};
struct CachedTextureCube {
OGLTexture texture;
u16 res_scale = 1;
std::shared_ptr<SurfaceWatcher> px;
std::shared_ptr<SurfaceWatcher> nx;
std::shared_ptr<SurfaceWatcher> py;
std::shared_ptr<SurfaceWatcher> ny;
std::shared_ptr<SurfaceWatcher> pz;
std::shared_ptr<SurfaceWatcher> nz;
};
} // namespace OpenGL

View File

@ -3,151 +3,306 @@
// Refer to the license.txt file included.
#pragma once
#include <algorithm>
#include <bit>
#include <span>
#include "common/alignment.h"
#include "core/memory.h"
#include "common/color.h"
#include "video_core/rasterizer_cache/pixel_format.h"
#include "video_core/renderer_opengl/gl_vars.h"
#include "video_core/texture/etc1.h"
#include "video_core/utils.h"
#include "video_core/video_core.h"
namespace OpenGL {
namespace VideoCore {
template <bool morton_to_gl, PixelFormat format>
static void MortonCopyTile(u32 stride, u8* tile_buffer, u8* gl_buffer) {
template <typename T>
inline T MakeInt(const std::byte* bytes) {
T integer{};
std::memcpy(&integer, bytes, sizeof(T));
return integer;
}
template <PixelFormat format, bool converted>
constexpr void DecodePixel(const std::byte* source, std::byte* dest) {
constexpr u32 bytes_per_pixel = GetFormatBpp(format) / 8;
constexpr u32 aligned_bytes_per_pixel = GetBytesPerPixel(format);
for (u32 y = 0; y < 8; ++y) {
for (u32 x = 0; x < 8; ++x) {
u8* tile_ptr = tile_buffer + VideoCore::MortonInterleave(x, y) * bytes_per_pixel;
u8* gl_ptr = gl_buffer + ((7 - y) * stride + x) * aligned_bytes_per_pixel;
if constexpr (morton_to_gl) {
if constexpr (format == PixelFormat::D24S8) {
gl_ptr[0] = tile_ptr[3];
std::memcpy(gl_ptr + 1, tile_ptr, 3);
} else if (format == PixelFormat::RGBA8 && GLES) {
// because GLES does not have ABGR format
// so we will do byteswapping here
gl_ptr[0] = tile_ptr[3];
gl_ptr[1] = tile_ptr[2];
gl_ptr[2] = tile_ptr[1];
gl_ptr[3] = tile_ptr[0];
} else if (format == PixelFormat::RGB8 && GLES) {
gl_ptr[0] = tile_ptr[2];
gl_ptr[1] = tile_ptr[1];
gl_ptr[2] = tile_ptr[0];
if constexpr (format == PixelFormat::D24S8) {
const u32 d24s8 = std::rotl(MakeInt<u32>(source), 8);
std::memcpy(dest, &d24s8, sizeof(u32));
} else if constexpr (format == PixelFormat::RGBA8 && converted) {
const u32 rgba = MakeInt<u32>(source);
const u32 abgr = Common::swap32(rgba);
std::memcpy(dest, &abgr, 4);
} else if constexpr (format == PixelFormat::RGB8 && converted) {
u32 rgb{};
std::memcpy(&rgb, source, 3);
const u32 abgr = Common::swap32(rgb << 8) | 0xFF000000;
std::memcpy(dest, &abgr, 4);
} else if constexpr (format == PixelFormat::RGB565 && converted) {
const auto abgr = Common::Color::DecodeRGB565(reinterpret_cast<const u8*>(source));
std::memcpy(dest, abgr.AsArray(), 4);
} else if constexpr (format == PixelFormat::RGB5A1 && converted) {
const auto abgr = Common::Color::DecodeRGB5A1(reinterpret_cast<const u8*>(source));
std::memcpy(dest, abgr.AsArray(), 4);
} else if constexpr (format == PixelFormat::RGBA4 && converted) {
const auto abgr = Common::Color::DecodeRGBA4(reinterpret_cast<const u8*>(source));
std::memcpy(dest, abgr.AsArray(), 4);
} else if constexpr (format == PixelFormat::IA8) {
std::memset(dest, static_cast<int>(source[1]), 3);
dest[3] = source[0];
} else if constexpr (format == PixelFormat::RG8) {
const auto rgba = Common::Color::DecodeRG8(reinterpret_cast<const u8*>(source));
std::memcpy(dest, rgba.AsArray(), 4);
} else if constexpr (format == PixelFormat::I8) {
std::memset(dest, static_cast<int>(source[0]), 3);
dest[3] = std::byte{255};
} else if constexpr (format == PixelFormat::A8) {
std::memset(dest, 0, 3);
dest[3] = source[0];
} else if constexpr (format == PixelFormat::IA4) {
const u8 ia4 = static_cast<const u8>(source[0]);
std::memset(dest, Common::Color::Convert4To8(ia4 >> 4), 3);
dest[3] = std::byte{Common::Color::Convert4To8(ia4 & 0xF)};
} else {
std::memcpy(dest, source, bytes_per_pixel);
}
}
template <PixelFormat format>
constexpr void DecodePixel4(u32 x, u32 y, const std::byte* source_tile, std::byte* dest_pixel) {
const u32 morton_offset = VideoCore::MortonInterleave(x, y);
const u8 value = static_cast<const u8>(source_tile[morton_offset >> 1]);
const u8 pixel = Common::Color::Convert4To8((morton_offset % 2) ? (value >> 4) : (value & 0xF));
if constexpr (format == PixelFormat::I4) {
std::memset(dest_pixel, static_cast<int>(pixel), 3);
dest_pixel[3] = std::byte{255};
} else {
std::memset(dest_pixel, 0, 3);
dest_pixel[3] = std::byte{pixel};
}
}
template <PixelFormat format>
constexpr void DecodePixelETC1(u32 x, u32 y, const std::byte* source_tile, std::byte* dest_pixel) {
constexpr u32 subtile_width = 4;
constexpr u32 subtile_height = 4;
constexpr bool has_alpha = format == PixelFormat::ETC1A4;
constexpr std::size_t subtile_size = has_alpha ? 16 : 8;
const u32 subtile_index = (x / subtile_width) + 2 * (y / subtile_height);
x %= subtile_width;
y %= subtile_height;
const std::byte* subtile_ptr = source_tile + subtile_index * subtile_size;
u8 alpha = 255;
if constexpr (has_alpha) {
u64_le packed_alpha;
std::memcpy(&packed_alpha, subtile_ptr, sizeof(u64));
subtile_ptr += sizeof(u64);
alpha = Common::Color::Convert4To8((packed_alpha >> (4 * (x * subtile_width + y))) & 0xF);
}
const u64_le subtile_data = MakeInt<u64_le>(subtile_ptr);
const auto rgb = Pica::Texture::SampleETC1Subtile(subtile_data, x, y);
// Copy the uncompressed pixel to the destination
std::memcpy(dest_pixel, rgb.AsArray(), 3);
dest_pixel[3] = std::byte{alpha};
}
template <PixelFormat format, bool converted>
constexpr void EncodePixel(const std::byte* source, std::byte* dest) {
constexpr u32 bytes_per_pixel = GetFormatBpp(format) / 8;
if constexpr (format == PixelFormat::D24S8) {
const u32 s8d24 = std::rotr(MakeInt<u32>(source), 8);
std::memcpy(dest, &s8d24, sizeof(u32));
} else if constexpr (format == PixelFormat::RGBA8 && converted) {
const u32 abgr = MakeInt<u32>(source);
const u32 rgba = Common::swap32(abgr);
std::memcpy(dest, &rgba, 4);
} else if constexpr (format == PixelFormat::RGB8 && converted) {
const u32 abgr = MakeInt<u32>(source);
const u32 rgb = Common::swap32(abgr << 8);
std::memcpy(dest, &rgb, 3);
} else if constexpr (format == PixelFormat::RGB565 && converted) {
Common::Vec4<u8> rgba;
std::memcpy(rgba.AsArray(), source, 4);
Common::Color::EncodeRGB565(rgba, reinterpret_cast<u8*>(dest));
} else if constexpr (format == PixelFormat::RGB5A1 && converted) {
Common::Vec4<u8> rgba;
std::memcpy(rgba.AsArray(), source, 4);
Common::Color::EncodeRGB5A1(rgba, reinterpret_cast<u8*>(dest));
} else if constexpr (format == PixelFormat::RGBA4 && converted) {
Common::Vec4<u8> rgba;
std::memcpy(rgba.AsArray(), source, 4);
Common::Color::EncodeRGBA4(rgba, reinterpret_cast<u8*>(dest));
} else {
std::memcpy(dest, source, bytes_per_pixel);
}
}
template <bool morton_to_linear, PixelFormat format, bool converted>
constexpr void MortonCopyTile(u32 stride, std::span<std::byte> tile_buffer,
std::span<std::byte> linear_buffer) {
constexpr u32 bytes_per_pixel = GetFormatBpp(format) / 8;
constexpr u32 linear_bytes_per_pixel = converted ? 4 : GetBytesPerPixel(format);
constexpr bool is_compressed = format == PixelFormat::ETC1 || format == PixelFormat::ETC1A4;
constexpr bool is_4bit = format == PixelFormat::I4 || format == PixelFormat::A4;
for (u32 y = 0; y < 8; y++) {
for (u32 x = 0; x < 8; x++) {
const auto tiled_pixel = tile_buffer.subspan(
VideoCore::MortonInterleave(x, y) * bytes_per_pixel, bytes_per_pixel);
const auto linear_pixel = linear_buffer.subspan(
((7 - y) * stride + x) * linear_bytes_per_pixel, linear_bytes_per_pixel);
if constexpr (morton_to_linear) {
if constexpr (is_compressed) {
DecodePixelETC1<format>(x, y, tile_buffer.data(), linear_pixel.data());
} else if constexpr (is_4bit) {
DecodePixel4<format>(x, y, tile_buffer.data(), linear_pixel.data());
} else {
std::memcpy(gl_ptr, tile_ptr, bytes_per_pixel);
DecodePixel<format, converted>(tiled_pixel.data(), linear_pixel.data());
}
} else {
if constexpr (format == PixelFormat::D24S8) {
std::memcpy(tile_ptr, gl_ptr + 1, 3);
tile_ptr[3] = gl_ptr[0];
} else if (format == PixelFormat::RGBA8 && GLES) {
// because GLES does not have ABGR format
// so we will do byteswapping here
tile_ptr[0] = gl_ptr[3];
tile_ptr[1] = gl_ptr[2];
tile_ptr[2] = gl_ptr[1];
tile_ptr[3] = gl_ptr[0];
} else if (format == PixelFormat::RGB8 && GLES) {
tile_ptr[0] = gl_ptr[2];
tile_ptr[1] = gl_ptr[1];
tile_ptr[2] = gl_ptr[0];
} else {
std::memcpy(tile_ptr, gl_ptr, bytes_per_pixel);
}
EncodePixel<format, converted>(linear_pixel.data(), tiled_pixel.data());
}
}
}
}
template <bool morton_to_gl, PixelFormat format>
static void MortonCopy(u32 stride, u32 height, u8* gl_buffer, PAddr base, PAddr start, PAddr end) {
/**
* @brief Performs morton to/from linear convertions on the provided pixel data
* @param converted If true performs RGBA8 to/from convertion to all color formats
* @param width, height The dimentions of the rectangular region of pixels in linear_buffer
* @param start_offset The number of bytes from the start of the first tile to the start of
* tiled_buffer
* @param end_offset The number of bytes from the start of the first tile to the end of tiled_buffer
* @param linear_buffer The linear pixel data
* @param tiled_buffer The tiled pixel data
*
* The MortonCopy is at the heart of the PICA texture implementation, as it's responsible for
* converting between linear and morton tiled layouts. The function handles both convertions but
* there are slightly different paths and inputs for each:
*
* Morton to Linear:
* During uploads, tiled_buffer is always aligned to the tile or scanline boundary depending if the
* linear rectangle spans multiple vertical tiles. linear_buffer does not reference the entire
* texture area, but rather the specific rectangle affected by the upload.
*
* Linear to Morton:
* This is similar to the other convertion but with some differences. In this case tiled_buffer is
* not required to be aligned to any specific boundary which requires special care.
* start_offset/end_offset are useful here as they tell us exactly where the data should be placed
* in the linear_buffer.
*/
template <bool morton_to_linear, PixelFormat format, bool converted = false>
static constexpr void MortonCopy(u32 width, u32 height, u32 start_offset, u32 end_offset,
std::span<std::byte> linear_buffer,
std::span<std::byte> tiled_buffer) {
constexpr u32 bytes_per_pixel = GetFormatBpp(format) / 8;
constexpr u32 tile_size = bytes_per_pixel * 64;
constexpr u32 aligned_bytes_per_pixel = GetBytesPerPixel(format);
constexpr u32 aligned_bytes_per_pixel = converted ? 4 : GetBytesPerPixel(format);
constexpr u32 tile_size = GetFormatBpp(format) * 64 / 8;
static_assert(aligned_bytes_per_pixel >= bytes_per_pixel, "");
gl_buffer += aligned_bytes_per_pixel - bytes_per_pixel;
const PAddr aligned_down_start = base + Common::AlignDown(start - base, tile_size);
const PAddr aligned_start = base + Common::AlignUp(start - base, tile_size);
const PAddr aligned_end = base + Common::AlignDown(end - base, tile_size);
const u32 linear_tile_stride = (7 * width + 8) * aligned_bytes_per_pixel;
const u32 aligned_down_start_offset = Common::AlignDown(start_offset, tile_size);
const u32 aligned_start_offset = Common::AlignUp(start_offset, tile_size);
const u32 aligned_end_offset = Common::AlignDown(end_offset, tile_size);
ASSERT(!morton_to_gl || (aligned_start == start && aligned_end == end));
ASSERT(!morton_to_linear ||
(aligned_start_offset == start_offset && aligned_end_offset == end_offset));
const u32 begin_pixel_index = (aligned_down_start - base) / bytes_per_pixel;
u32 x = (begin_pixel_index % (stride * 8)) / 8;
u32 y = (begin_pixel_index / (stride * 8)) * 8;
// In OpenGL the texture origin is in the bottom left corner as opposed to other
// APIs that have it at the top left. To avoid flipping texture coordinates in
// the shader we read/write the linear buffer from the bottom up
u32 linear_offset = ((height - 8) * width) * aligned_bytes_per_pixel;
u32 tiled_offset = 0;
u32 x = 0;
u32 y = 0;
gl_buffer += ((height - 8 - y) * stride + x) * aligned_bytes_per_pixel;
auto glbuf_next_tile = [&] {
x = (x + 8) % stride;
gl_buffer += 8 * aligned_bytes_per_pixel;
const auto LinearNextTile = [&] {
x = (x + 8) % width;
linear_offset += 8 * aligned_bytes_per_pixel;
if (!x) {
y += 8;
gl_buffer -= stride * 9 * aligned_bytes_per_pixel;
y = (y + 8) % height;
if (!y) {
return;
}
linear_offset -= width * 9 * aligned_bytes_per_pixel;
}
};
u8* tile_buffer = VideoCore::g_memory->GetPhysicalPointer(start);
// If during a texture download the start coordinate is not tile aligned, swizzle
// the tile affected to a temporary buffer and copy the part we are interested in
if (start_offset < aligned_start_offset && !morton_to_linear) {
std::array<std::byte, tile_size> tmp_buf;
auto linear_data = linear_buffer.subspan(linear_offset, linear_tile_stride);
MortonCopyTile<morton_to_linear, format, converted>(width, tmp_buf, linear_data);
if (start < aligned_start && !morton_to_gl) {
std::array<u8, tile_size> tmp_buf;
MortonCopyTile<morton_to_gl, format>(stride, &tmp_buf[0], gl_buffer);
std::memcpy(tile_buffer, &tmp_buf[start - aligned_down_start],
std::min(aligned_start, end) - start);
std::memcpy(tiled_buffer.data(), tmp_buf.data() + start_offset - aligned_down_start_offset,
std::min(aligned_start_offset, end_offset) - start_offset);
tile_buffer += aligned_start - start;
glbuf_next_tile();
tiled_offset += aligned_start_offset - start_offset;
LinearNextTile();
}
const u8* const buffer_end = tile_buffer + aligned_end - aligned_start;
PAddr current_paddr = aligned_start;
while (tile_buffer < buffer_end) {
// Pokemon Super Mystery Dungeon will try to use textures that go beyond
// the end address of VRAM. Stop reading if reaches invalid address
if (!VideoCore::g_memory->IsValidPhysicalAddress(current_paddr) ||
!VideoCore::g_memory->IsValidPhysicalAddress(current_paddr + tile_size)) {
LOG_ERROR(Render_OpenGL, "Out of bound texture");
break;
}
MortonCopyTile<morton_to_gl, format>(stride, tile_buffer, gl_buffer);
tile_buffer += tile_size;
current_paddr += tile_size;
glbuf_next_tile();
const u32 buffer_end = tiled_offset + aligned_end_offset - aligned_start_offset;
while (tiled_offset < buffer_end) {
auto linear_data = linear_buffer.subspan(linear_offset, linear_tile_stride);
auto tiled_data = tiled_buffer.subspan(tiled_offset, tile_size);
MortonCopyTile<morton_to_linear, format, converted>(width, tiled_data, linear_data);
tiled_offset += tile_size;
LinearNextTile();
}
if (end > std::max(aligned_start, aligned_end) && !morton_to_gl) {
std::array<u8, tile_size> tmp_buf;
MortonCopyTile<morton_to_gl, format>(stride, &tmp_buf[0], gl_buffer);
std::memcpy(tile_buffer, &tmp_buf[0], end - aligned_end);
// If during a texture download the end coordinate is not tile aligned, swizzle
// the tile affected to a temporary buffer and copy the part we are interested in
if (end_offset > std::max(aligned_start_offset, aligned_end_offset) && !morton_to_linear) {
std::array<std::byte, tile_size> tmp_buf;
auto linear_data = linear_buffer.subspan(linear_offset, linear_tile_stride);
MortonCopyTile<morton_to_linear, format, converted>(width, tmp_buf, linear_data);
std::memcpy(tiled_buffer.data() + tiled_offset, tmp_buf.data(),
end_offset - aligned_end_offset);
}
}
static constexpr std::array<void (*)(u32, u32, u8*, PAddr, PAddr, PAddr), 18> morton_to_gl_fns = {
using MortonFunc = void (*)(u32, u32, u32, u32, std::span<std::byte>, std::span<std::byte>);
static constexpr std::array<MortonFunc, 18> UNSWIZZLE_TABLE = {
MortonCopy<true, PixelFormat::RGBA8>, // 0
MortonCopy<true, PixelFormat::RGB8>, // 1
MortonCopy<true, PixelFormat::RGB5A1>, // 2
MortonCopy<true, PixelFormat::RGB565>, // 3
MortonCopy<true, PixelFormat::RGBA4>, // 4
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
nullptr,
nullptr, // 5 - 13
MortonCopy<true, PixelFormat::D16>, // 14
nullptr, // 15
MortonCopy<true, PixelFormat::D24>, // 16
MortonCopy<true, PixelFormat::D24S8> // 17
MortonCopy<true, PixelFormat::IA8>, // 5
MortonCopy<true, PixelFormat::RG8>, // 6
MortonCopy<true, PixelFormat::I8>, // 7
MortonCopy<true, PixelFormat::A8>, // 8
MortonCopy<true, PixelFormat::IA4>, // 9
MortonCopy<true, PixelFormat::I4>, // 10
MortonCopy<true, PixelFormat::A4>, // 11
MortonCopy<true, PixelFormat::ETC1>, // 12
MortonCopy<true, PixelFormat::ETC1A4>, // 13
MortonCopy<true, PixelFormat::D16>, // 14
nullptr, // 15
MortonCopy<true, PixelFormat::D24>, // 16
MortonCopy<true, PixelFormat::D24S8> // 17
};
static constexpr std::array<void (*)(u32, u32, u8*, PAddr, PAddr, PAddr), 18> gl_to_morton_fns = {
static constexpr std::array<MortonFunc, 18> UNSWIZZLE_TABLE_CONVERTED = {
MortonCopy<true, PixelFormat::RGBA8, true>, // 0
MortonCopy<true, PixelFormat::RGB8, true>, // 1
MortonCopy<true, PixelFormat::RGB5A1, true>, // 2
MortonCopy<true, PixelFormat::RGB565, true>, // 3
MortonCopy<true, PixelFormat::RGBA4, true> // 4
};
static constexpr std::array<MortonFunc, 18> SWIZZLE_TABLE = {
MortonCopy<false, PixelFormat::RGBA8>, // 0
MortonCopy<false, PixelFormat::RGB8>, // 1
MortonCopy<false, PixelFormat::RGB5A1>, // 2
@ -168,4 +323,12 @@ static constexpr std::array<void (*)(u32, u32, u8*, PAddr, PAddr, PAddr), 18> gl
MortonCopy<false, PixelFormat::D24S8> // 17
};
} // namespace OpenGL
static constexpr std::array<MortonFunc, 18> SWIZZLE_TABLE_CONVERTED = {
MortonCopy<false, PixelFormat::RGBA8, true>, // 0
MortonCopy<false, PixelFormat::RGB8, true>, // 1
MortonCopy<false, PixelFormat::RGB5A1, true>, // 2
MortonCopy<false, PixelFormat::RGB565, true>, // 3
MortonCopy<false, PixelFormat::RGBA4, true> // 4
};
} // namespace VideoCore

View File

@ -0,0 +1,98 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include "video_core/rasterizer_cache/pixel_format.h"
namespace VideoCore {
std::string_view PixelFormatAsString(PixelFormat format) {
switch (format) {
case PixelFormat::RGBA8:
return "RGBA8";
case PixelFormat::RGB8:
return "RGB8";
case PixelFormat::RGB5A1:
return "RGB5A1";
case PixelFormat::RGB565:
return "RGB565";
case PixelFormat::RGBA4:
return "RGBA4";
case PixelFormat::IA8:
return "IA8";
case PixelFormat::RG8:
return "RG8";
case PixelFormat::I8:
return "I8";
case PixelFormat::A8:
return "A8";
case PixelFormat::IA4:
return "IA4";
case PixelFormat::I4:
return "I4";
case PixelFormat::A4:
return "A4";
case PixelFormat::ETC1:
return "ETC1";
case PixelFormat::ETC1A4:
return "ETC1A4";
case PixelFormat::D16:
return "D16";
case PixelFormat::D24:
return "D24";
case PixelFormat::D24S8:
return "D24S8";
default:
return "NotReal";
}
}
bool CheckFormatsBlittable(PixelFormat source_format, PixelFormat dest_format) {
SurfaceType source_type = GetFormatType(source_format);
SurfaceType dest_type = GetFormatType(dest_format);
if ((source_type == SurfaceType::Color || source_type == SurfaceType::Texture) &&
(dest_type == SurfaceType::Color || dest_type == SurfaceType::Texture)) {
return true;
}
if (source_type == SurfaceType::Depth && dest_type == SurfaceType::Depth) {
return true;
}
if (source_type == SurfaceType::DepthStencil && dest_type == SurfaceType::DepthStencil) {
return true;
}
return false;
}
PixelFormat PixelFormatFromTextureFormat(Pica::TexturingRegs::TextureFormat format) {
const u32 format_index = static_cast<u32>(format);
return (format_index < 14) ? static_cast<PixelFormat>(format) : PixelFormat::Invalid;
}
PixelFormat PixelFormatFromColorFormat(Pica::FramebufferRegs::ColorFormat format) {
const u32 format_index = static_cast<u32>(format);
return (format_index < 5) ? static_cast<PixelFormat>(format) : PixelFormat::Invalid;
}
PixelFormat PixelFormatFromDepthFormat(Pica::FramebufferRegs::DepthFormat format) {
const u32 format_index = static_cast<u32>(format);
return (format_index < 4) ? static_cast<PixelFormat>(format_index + 14) : PixelFormat::Invalid;
}
PixelFormat PixelFormatFromGPUPixelFormat(GPU::Regs::PixelFormat format) {
const u32 format_index = static_cast<u32>(format);
switch (format) {
// RGB565 and RGB5A1 are switched in PixelFormat compared to ColorFormat
case GPU::Regs::PixelFormat::RGB565:
return PixelFormat::RGB565;
case GPU::Regs::PixelFormat::RGB5A1:
return PixelFormat::RGB5A1;
default:
return (format_index < 5) ? static_cast<PixelFormat>(format) : PixelFormat::Invalid;
}
}
} // namespace VideoCore

View File

@ -3,23 +3,22 @@
// Refer to the license.txt file included.
#pragma once
#include <string_view>
#include "core/hw/gpu.h"
#include "video_core/regs_framebuffer.h"
#include "video_core/regs_texturing.h"
namespace OpenGL {
namespace VideoCore {
constexpr u32 PIXEL_FORMAT_COUNT = 18;
constexpr std::size_t PIXEL_FORMAT_COUNT = 18;
enum class PixelFormat : u8 {
// First 5 formats are shared between textures and color buffers
enum class PixelFormat : u32 {
RGBA8 = 0,
RGB8 = 1,
RGB5A1 = 2,
RGB565 = 3,
RGBA4 = 4,
// Texture-only formats
IA8 = 5,
RG8 = 6,
I8 = 7,
@ -29,163 +28,81 @@ enum class PixelFormat : u8 {
A4 = 11,
ETC1 = 12,
ETC1A4 = 13,
// Depth buffer-only formats
D16 = 14,
D24 = 16,
D24S8 = 17,
Max = 18,
Invalid = 255,
};
enum class SurfaceType {
enum class SurfaceType : u32 {
Color = 0,
Texture = 1,
Depth = 2,
DepthStencil = 3,
Fill = 4,
Invalid = 5
Invalid = 5,
};
constexpr std::string_view PixelFormatAsString(PixelFormat format) {
switch (format) {
case PixelFormat::RGBA8:
return "RGBA8";
case PixelFormat::RGB8:
return "RGB8";
case PixelFormat::RGB5A1:
return "RGB5A1";
case PixelFormat::RGB565:
return "RGB565";
case PixelFormat::RGBA4:
return "RGBA4";
case PixelFormat::IA8:
return "IA8";
case PixelFormat::RG8:
return "RG8";
case PixelFormat::I8:
return "I8";
case PixelFormat::A8:
return "A8";
case PixelFormat::IA4:
return "IA4";
case PixelFormat::I4:
return "I4";
case PixelFormat::A4:
return "A4";
case PixelFormat::ETC1:
return "ETC1";
case PixelFormat::ETC1A4:
return "ETC1A4";
case PixelFormat::D16:
return "D16";
case PixelFormat::D24:
return "D24";
case PixelFormat::D24S8:
return "D24S8";
default:
return "NotReal";
}
}
enum class TextureType : u32 {
Texture2D = 0,
CubeMap = 1,
};
constexpr PixelFormat PixelFormatFromTextureFormat(Pica::TexturingRegs::TextureFormat format) {
const u32 format_index = static_cast<u32>(format);
return (format_index < 14) ? static_cast<PixelFormat>(format) : PixelFormat::Invalid;
}
constexpr PixelFormat PixelFormatFromColorFormat(Pica::FramebufferRegs::ColorFormat format) {
const u32 format_index = static_cast<u32>(format);
return (format_index < 5) ? static_cast<PixelFormat>(format) : PixelFormat::Invalid;
}
constexpr PixelFormat PixelFormatFromDepthFormat(Pica::FramebufferRegs::DepthFormat format) {
const u32 format_index = static_cast<u32>(format);
return (format_index < 4) ? static_cast<PixelFormat>(format_index + 14) : PixelFormat::Invalid;
}
constexpr PixelFormat PixelFormatFromGPUPixelFormat(GPU::Regs::PixelFormat format) {
const u32 format_index = static_cast<u32>(format);
switch (format) {
// RGB565 and RGB5A1 are switched in PixelFormat compared to ColorFormat
case GPU::Regs::PixelFormat::RGB565:
return PixelFormat::RGB565;
case GPU::Regs::PixelFormat::RGB5A1:
return PixelFormat::RGB5A1;
default:
return (format_index < 5) ? static_cast<PixelFormat>(format) : PixelFormat::Invalid;
}
}
constexpr SurfaceType GetFormatType(PixelFormat pixel_format) {
const u32 format_index = static_cast<u32>(pixel_format);
if (format_index < 5) {
return SurfaceType::Color;
}
if (format_index < 14) {
return SurfaceType::Texture;
}
if (pixel_format == PixelFormat::D16 || pixel_format == PixelFormat::D24) {
return SurfaceType::Depth;
}
if (pixel_format == PixelFormat::D24S8) {
return SurfaceType::DepthStencil;
}
return SurfaceType::Invalid;
}
constexpr bool CheckFormatsBlittable(PixelFormat source_format, PixelFormat dest_format) {
SurfaceType source_type = GetFormatType(source_format);
SurfaceType dest_type = GetFormatType(dest_format);
if ((source_type == SurfaceType::Color || source_type == SurfaceType::Texture) &&
(dest_type == SurfaceType::Color || dest_type == SurfaceType::Texture)) {
return true;
}
if (source_type == SurfaceType::Depth && dest_type == SurfaceType::Depth) {
return true;
}
if (source_type == SurfaceType::DepthStencil && dest_type == SurfaceType::DepthStencil) {
return true;
}
return false;
}
constexpr std::array<u8, PIXEL_FORMAT_COUNT> BITS_PER_BLOCK_TABLE = {{
32, // RGBA8
24, // RGB8
16, // RGB5A1
16, // RGB565
16, // RGBA4
16, // IA8
16, // RG8
8, // I8
8, // A8
8, // IA4
4, // I4
4, // A4
4, // ETC1
8, // ETC1A4
16, // D16
0,
24, // D24
32, // D24S8
}};
constexpr u32 GetFormatBpp(PixelFormat format) {
switch (format) {
case PixelFormat::RGBA8:
case PixelFormat::D24S8:
return 32;
case PixelFormat::RGB8:
case PixelFormat::D24:
return 24;
case PixelFormat::RGB5A1:
case PixelFormat::RGB565:
case PixelFormat::RGBA4:
case PixelFormat::IA8:
case PixelFormat::RG8:
case PixelFormat::D16:
return 16;
case PixelFormat::I8:
case PixelFormat::A8:
case PixelFormat::IA4:
case PixelFormat::ETC1A4:
return 8;
case PixelFormat::I4:
case PixelFormat::A4:
case PixelFormat::ETC1:
return 4;
default:
return 0;
}
ASSERT(static_cast<std::size_t>(format) < BITS_PER_BLOCK_TABLE.size());
return BITS_PER_BLOCK_TABLE[static_cast<std::size_t>(format)];
}
constexpr std::array<SurfaceType, PIXEL_FORMAT_COUNT> FORMAT_TYPE_TABLE = {{
SurfaceType::Color, // RGBA8
SurfaceType::Color, // RGB8
SurfaceType::Color, // RGB5A1
SurfaceType::Color, // RGB565
SurfaceType::Color, // RGBA4
SurfaceType::Texture, // IA8
SurfaceType::Texture, // RG8
SurfaceType::Texture, // I8
SurfaceType::Texture, // A8
SurfaceType::Texture, // IA4
SurfaceType::Texture, // I4
SurfaceType::Texture, // A4
SurfaceType::Texture, // ETC1
SurfaceType::Texture, // ETC1A4
SurfaceType::Depth, // D16
SurfaceType::Invalid,
SurfaceType::Depth, // D24
SurfaceType::DepthStencil, // D24S8
}};
constexpr SurfaceType GetFormatType(PixelFormat format) {
ASSERT(static_cast<std::size_t>(format) < FORMAT_TYPE_TABLE.size());
return FORMAT_TYPE_TABLE[static_cast<std::size_t>(format)];
}
constexpr u32 GetBytesPerPixel(PixelFormat format) {
// OpenGL needs 4 bpp alignment for D24 since using GL_UNSIGNED_INT as type
// Modern GPUs need 4 bpp alignment for D24
if (format == PixelFormat::D24 || GetFormatType(format) == SurfaceType::Texture) {
return 4;
}
@ -193,4 +110,16 @@ constexpr u32 GetBytesPerPixel(PixelFormat format) {
return GetFormatBpp(format) / 8;
}
} // namespace OpenGL
std::string_view PixelFormatAsString(PixelFormat format);
bool CheckFormatsBlittable(PixelFormat source_format, PixelFormat dest_format);
PixelFormat PixelFormatFromTextureFormat(Pica::TexturingRegs::TextureFormat format);
PixelFormat PixelFormatFromColorFormat(Pica::FramebufferRegs::ColorFormat format);
PixelFormat PixelFormatFromDepthFormat(Pica::FramebufferRegs::DepthFormat format);
PixelFormat PixelFormatFromGPUPixelFormat(GPU::Regs::PixelFormat format);
} // namespace VideoCore

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff

View File

@ -1,38 +0,0 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <memory>
#include <set>
#include <tuple>
#include <boost/icl/interval_map.hpp>
#include <boost/icl/interval_set.hpp>
#include "common/common_types.h"
#include "common/math_util.h"
namespace OpenGL {
class CachedSurface;
using Surface = std::shared_ptr<CachedSurface>;
// Declare rasterizer interval types
using SurfaceInterval = boost::icl::right_open_interval<PAddr>;
using SurfaceSet = std::set<Surface>;
using SurfaceRegions = boost::icl::interval_set<PAddr, std::less, SurfaceInterval>;
using SurfaceMap =
boost::icl::interval_map<PAddr, Surface, boost::icl::partial_absorber, std::less,
boost::icl::inplace_plus, boost::icl::inter_section, SurfaceInterval>;
using SurfaceCache =
boost::icl::interval_map<PAddr, SurfaceSet, boost::icl::partial_absorber, std::less,
boost::icl::inplace_plus, boost::icl::inter_section, SurfaceInterval>;
static_assert(std::is_same<SurfaceRegions::interval_type, SurfaceCache::interval_type>() &&
std::is_same<SurfaceMap::interval_type, SurfaceCache::interval_type>(),
"Incorrect interval types");
using SurfaceRect_Tuple = std::tuple<Surface, Common::Rectangle<u32>>;
using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, Common::Rectangle<u32>>;
using PageMap = boost::icl::interval_map<u32, int>;
} // namespace OpenGL

View File

@ -1,56 +0,0 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <glad/glad.h>
#include "video_core/rasterizer_cache/rasterizer_cache_utils.h"
#include "video_core/renderer_opengl/gl_vars.h"
namespace OpenGL {
constexpr FormatTuple tex_tuple = {GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE};
static constexpr std::array<FormatTuple, 4> depth_format_tuples = {{
{GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT}, // D16
{},
{GL_DEPTH_COMPONENT24, GL_DEPTH_COMPONENT, GL_UNSIGNED_INT}, // D24
{GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // D24S8
}};
static constexpr std::array<FormatTuple, 5> fb_format_tuples = {{
{GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8}, // RGBA8
{GL_RGB8, GL_BGR, GL_UNSIGNED_BYTE}, // RGB8
{GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_5_5_5_1}, // RGB5A1
{GL_RGB565, GL_RGB, GL_UNSIGNED_SHORT_5_6_5}, // RGB565
{GL_RGBA4, GL_RGBA, GL_UNSIGNED_SHORT_4_4_4_4}, // RGBA4
}};
// Same as above, with minor changes for OpenGL ES. Replaced
// GL_UNSIGNED_INT_8_8_8_8 with GL_UNSIGNED_BYTE and
// GL_BGR with GL_RGB
static constexpr std::array<FormatTuple, 5> fb_format_tuples_oes = {{
{GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE}, // RGBA8
{GL_RGB8, GL_RGB, GL_UNSIGNED_BYTE}, // RGB8
{GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_5_5_5_1}, // RGB5A1
{GL_RGB565, GL_RGB, GL_UNSIGNED_SHORT_5_6_5}, // RGB565
{GL_RGBA4, GL_RGBA, GL_UNSIGNED_SHORT_4_4_4_4}, // RGBA4
}};
const FormatTuple& GetFormatTuple(PixelFormat pixel_format) {
const SurfaceType type = GetFormatType(pixel_format);
const std::size_t format_index = static_cast<std::size_t>(pixel_format);
if (type == SurfaceType::Color) {
ASSERT(format_index < fb_format_tuples.size());
return (GLES ? fb_format_tuples_oes : fb_format_tuples)[format_index];
} else if (type == SurfaceType::Depth || type == SurfaceType::DepthStencil) {
const std::size_t tuple_idx = format_index - 14;
ASSERT(tuple_idx < depth_format_tuples.size());
return depth_format_tuples[tuple_idx];
}
return tex_tuple;
}
} // namespace OpenGL

View File

@ -1,73 +0,0 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <functional>
#include "common/hash.h"
#include "video_core/rasterizer_cache/pixel_format.h"
namespace OpenGL {
struct FormatTuple {
int internal_format;
u32 format;
u32 type;
};
const FormatTuple& GetFormatTuple(PixelFormat pixel_format);
struct HostTextureTag {
FormatTuple format_tuple{};
u32 width = 0;
u32 height = 0;
bool operator==(const HostTextureTag& rhs) const noexcept {
return std::memcmp(this, &rhs, sizeof(HostTextureTag)) == 0;
};
const u64 Hash() const {
return Common::ComputeHash64(this, sizeof(HostTextureTag));
}
};
struct TextureCubeConfig {
PAddr px;
PAddr nx;
PAddr py;
PAddr ny;
PAddr pz;
PAddr nz;
u32 width;
Pica::TexturingRegs::TextureFormat format;
bool operator==(const TextureCubeConfig& rhs) const {
return std::memcmp(this, &rhs, sizeof(TextureCubeConfig)) == 0;
}
bool operator!=(const TextureCubeConfig& rhs) const {
return std::memcmp(this, &rhs, sizeof(TextureCubeConfig)) != 0;
}
const u64 Hash() const {
return Common::ComputeHash64(this, sizeof(TextureCubeConfig));
}
};
} // namespace OpenGL
namespace std {
template <>
struct hash<OpenGL::HostTextureTag> {
std::size_t operator()(const OpenGL::HostTextureTag& tag) const noexcept {
return tag.Hash();
}
};
template <>
struct hash<OpenGL::TextureCubeConfig> {
std::size_t operator()(const OpenGL::TextureCubeConfig& config) const noexcept {
return config.Hash();
}
};
} // namespace std

View File

@ -0,0 +1,217 @@
// Copyright 2022 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <memory>
#include "common/alignment.h"
#include "common/assert.h"
#include "video_core/rasterizer_cache/surface_params.h"
namespace VideoCore {
using SurfaceRegions = boost::icl::interval_set<PAddr, std::less, SurfaceInterval>;
/**
* A watcher that notifies whether a cached surface has been changed. This is useful for caching
* surface collection objects, including texture cube and mipmap.
*/
template <class S>
class SurfaceWatcher {
public:
explicit SurfaceWatcher(std::weak_ptr<S>&& surface) : surface(std::move(surface)) {}
/// Checks whether the surface has been changed.
bool IsValid() const {
return !surface.expired() && valid;
}
/// Marks that the content of the referencing surface has been updated to the watcher user.
void Validate() {
ASSERT(!surface.expired());
valid = true;
}
/// Gets the referencing surface. Returns null if the surface has been destroyed
std::shared_ptr<S> Get() const {
return surface.lock();
}
public:
std::weak_ptr<S> surface;
bool valid = false;
};
template <class S>
class SurfaceBase : public SurfaceParams, public std::enable_shared_from_this<S> {
using Watcher = SurfaceWatcher<S>;
public:
SurfaceBase() = default;
SurfaceBase(const SurfaceParams& params) : SurfaceParams{params} {}
virtual ~SurfaceBase() = default;
/// Returns true when this surface can be used to fill the fill_interval of dest_surface
bool CanFill(const SurfaceParams& dest_surface, SurfaceInterval fill_interval) const;
/// Returns true when copy_interval of dest_surface can be validated by copying from this
/// surface
bool CanCopy(const SurfaceParams& dest_surface, SurfaceInterval copy_interval) const;
/// Returns the region of the biggest valid rectange within interval
SurfaceInterval GetCopyableInterval(const SurfaceParams& params) const;
/// Creates a surface watcher linked to this surface
std::shared_ptr<Watcher> CreateWatcher();
/// Invalidates all watchers linked to this surface
void InvalidateAllWatcher();
/// Removes any linked watchers from this surface
void UnlinkAllWatcher();
/// Returns true when the region denoted by interval is valid
bool IsRegionValid(SurfaceInterval interval) const {
return (invalid_regions.find(interval) == invalid_regions.end());
}
/// Returns true when the entire surface is invalid
bool IsSurfaceFullyInvalid() const {
auto interval = GetInterval();
return *invalid_regions.equal_range(interval).first == interval;
}
public:
bool registered = false;
SurfaceRegions invalid_regions;
std::array<std::shared_ptr<Watcher>, 7> level_watchers;
u32 max_level = 0;
std::array<u8, 4> fill_data;
u32 fill_size = 0;
public:
std::vector<std::weak_ptr<Watcher>> watchers;
};
template <class S>
bool SurfaceBase<S>::CanFill(const SurfaceParams& dest_surface,
SurfaceInterval fill_interval) const {
if (type == SurfaceType::Fill && IsRegionValid(fill_interval) &&
boost::icl::first(fill_interval) >= addr &&
boost::icl::last_next(fill_interval) <= end && // dest_surface is within our fill range
dest_surface.FromInterval(fill_interval).GetInterval() ==
fill_interval) { // make sure interval is a rectangle in dest surface
if (fill_size * 8 != dest_surface.GetFormatBpp()) {
// Check if bits repeat for our fill_size
const u32 dest_bytes_per_pixel = std::max(dest_surface.GetFormatBpp() / 8, 1u);
std::vector<u8> fill_test(fill_size * dest_bytes_per_pixel);
for (u32 i = 0; i < dest_bytes_per_pixel; ++i)
std::memcpy(&fill_test[i * fill_size], &fill_data[0], fill_size);
for (u32 i = 0; i < fill_size; ++i)
if (std::memcmp(&fill_test[dest_bytes_per_pixel * i], &fill_test[0],
dest_bytes_per_pixel) != 0)
return false;
if (dest_surface.GetFormatBpp() == 4 && (fill_test[0] & 0xF) != (fill_test[0] >> 4))
return false;
}
return true;
}
return false;
}
template <class S>
bool SurfaceBase<S>::CanCopy(const SurfaceParams& dest_surface,
SurfaceInterval copy_interval) const {
SurfaceParams subrect_params = dest_surface.FromInterval(copy_interval);
ASSERT(subrect_params.GetInterval() == copy_interval);
if (CanSubRect(subrect_params))
return true;
if (CanFill(dest_surface, copy_interval))
return true;
return false;
}
template <class S>
SurfaceInterval SurfaceBase<S>::GetCopyableInterval(const SurfaceParams& params) const {
SurfaceInterval result{};
const u32 tile_align = params.BytesInPixels(params.is_tiled ? 8 * 8 : 1);
const auto valid_regions =
SurfaceRegions{params.GetInterval() & GetInterval()} - invalid_regions;
for (auto& valid_interval : valid_regions) {
const SurfaceInterval aligned_interval{
params.addr +
Common::AlignUp(boost::icl::first(valid_interval) - params.addr, tile_align),
params.addr +
Common::AlignDown(boost::icl::last_next(valid_interval) - params.addr, tile_align)};
if (params.BytesInPixels(tile_align) > boost::icl::length(valid_interval) ||
boost::icl::length(aligned_interval) == 0) {
continue;
}
// Get the rectangle within aligned_interval
const u32 stride_bytes = params.BytesInPixels(params.stride) * (params.is_tiled ? 8 : 1);
SurfaceInterval rect_interval{
params.addr +
Common::AlignUp(boost::icl::first(aligned_interval) - params.addr, stride_bytes),
params.addr + Common::AlignDown(boost::icl::last_next(aligned_interval) - params.addr,
stride_bytes),
};
if (boost::icl::first(rect_interval) > boost::icl::last_next(rect_interval)) {
// 1 row
rect_interval = aligned_interval;
} else if (boost::icl::length(rect_interval) == 0) {
// 2 rows that do not make a rectangle, return the larger one
const SurfaceInterval row1{boost::icl::first(aligned_interval),
boost::icl::first(rect_interval)};
const SurfaceInterval row2{boost::icl::first(rect_interval),
boost::icl::last_next(aligned_interval)};
rect_interval = (boost::icl::length(row1) > boost::icl::length(row2)) ? row1 : row2;
}
if (boost::icl::length(rect_interval) > boost::icl::length(result)) {
result = rect_interval;
}
}
return result;
}
template <class S>
auto SurfaceBase<S>::CreateWatcher() -> std::shared_ptr<Watcher> {
auto weak_ptr = reinterpret_cast<S*>(this)->weak_from_this();
auto watcher = std::make_shared<Watcher>(std::move(weak_ptr));
watchers.push_back(watcher);
return watcher;
}
template <class S>
void SurfaceBase<S>::InvalidateAllWatcher() {
for (const auto& watcher : watchers) {
if (auto locked = watcher.lock()) {
locked->valid = false;
}
}
}
template <class S>
void SurfaceBase<S>::UnlinkAllWatcher() {
for (const auto& watcher : watchers) {
if (auto locked = watcher.lock()) {
locked->valid = false;
locked->surface.reset();
}
}
watchers.clear();
}
} // namespace VideoCore

Some files were not shown because too many files have changed in this diff Show More