diff options
| author | 2015-03-08 21:31:15 -0500 | |
|---|---|---|
| committer | 2015-03-09 20:13:21 -0500 | |
| commit | 1248e291f0c9a29734b0f5175df8fa675cce930c (patch) | |
| tree | 18218cbe496e5a2f42022a5e6741fa0665218874 /src | |
| parent | Frontend/Qt: Allow the framebuffer widget to inspect the depth buffer (diff) | |
| download | yuzu-1248e291f0c9a29734b0f5175df8fa675cce930c.tar.gz yuzu-1248e291f0c9a29734b0f5175df8fa675cce930c.tar.xz yuzu-1248e291f0c9a29734b0f5175df8fa675cce930c.zip | |
GPU: Added the stencil test structure to the Pica Regs struct.
Diffstat (limited to 'src')
| -rw-r--r-- | src/citra_qt/debugger/graphics_framebuffer.cpp | 6 | ||||
| -rw-r--r-- | src/citra_qt/debugger/graphics_framebuffer.h | 6 | ||||
| -rw-r--r-- | src/core/hw/gpu.cpp | 4 | ||||
| -rw-r--r-- | src/core/hw/gpu.h | 6 | ||||
| -rw-r--r-- | src/video_core/color.h | 18 | ||||
| -rw-r--r-- | src/video_core/pica.h | 43 | ||||
| -rw-r--r-- | src/video_core/rasterizer.cpp | 54 |
7 files changed, 76 insertions, 61 deletions
diff --git a/src/citra_qt/debugger/graphics_framebuffer.cpp b/src/citra_qt/debugger/graphics_framebuffer.cpp index 2985cadb1..d621d7204 100644 --- a/src/citra_qt/debugger/graphics_framebuffer.cpp +++ b/src/citra_qt/debugger/graphics_framebuffer.cpp | |||
| @@ -176,7 +176,7 @@ void GraphicsFramebufferWidget::OnUpdate() | |||
| 176 | { | 176 | { |
| 177 | // TODO: Store a reference to the registers in the debug context instead of accessing them directly... | 177 | // TODO: Store a reference to the registers in the debug context instead of accessing them directly... |
| 178 | 178 | ||
| 179 | auto framebuffer = Pica::registers.framebuffer; | 179 | const auto& framebuffer = Pica::registers.framebuffer; |
| 180 | 180 | ||
| 181 | framebuffer_address = framebuffer.GetColorBufferPhysicalAddress(); | 181 | framebuffer_address = framebuffer.GetColorBufferPhysicalAddress(); |
| 182 | framebuffer_width = framebuffer.GetWidth(); | 182 | framebuffer_width = framebuffer.GetWidth(); |
| @@ -189,7 +189,7 @@ void GraphicsFramebufferWidget::OnUpdate() | |||
| 189 | 189 | ||
| 190 | case Source::DepthBuffer: | 190 | case Source::DepthBuffer: |
| 191 | { | 191 | { |
| 192 | auto framebuffer = Pica::registers.framebuffer; | 192 | const auto& framebuffer = Pica::registers.framebuffer; |
| 193 | 193 | ||
| 194 | framebuffer_address = framebuffer.GetDepthBufferPhysicalAddress(); | 194 | framebuffer_address = framebuffer.GetDepthBufferPhysicalAddress(); |
| 195 | framebuffer_width = framebuffer.GetWidth(); | 195 | framebuffer_width = framebuffer.GetWidth(); |
| @@ -294,4 +294,4 @@ u32 GraphicsFramebufferWidget::BytesPerPixel(GraphicsFramebufferWidget::Format f | |||
| 294 | case Format::D16: | 294 | case Format::D16: |
| 295 | return 2; | 295 | return 2; |
| 296 | } | 296 | } |
| 297 | } \ No newline at end of file | 297 | } |
diff --git a/src/citra_qt/debugger/graphics_framebuffer.h b/src/citra_qt/debugger/graphics_framebuffer.h index dff91d131..4cb396ffe 100644 --- a/src/citra_qt/debugger/graphics_framebuffer.h +++ b/src/citra_qt/debugger/graphics_framebuffer.h | |||
| @@ -20,9 +20,9 @@ class GraphicsFramebufferWidget : public BreakPointObserverDock { | |||
| 20 | using Event = Pica::DebugContext::Event; | 20 | using Event = Pica::DebugContext::Event; |
| 21 | 21 | ||
| 22 | enum class Source { | 22 | enum class Source { |
| 23 | PicaTarget = 0, | 23 | PicaTarget = 0, |
| 24 | DepthBuffer = 1, | 24 | DepthBuffer = 1, |
| 25 | Custom = 2, | 25 | Custom = 2, |
| 26 | 26 | ||
| 27 | // TODO: Add GPU framebuffer sources! | 27 | // TODO: Add GPU framebuffer sources! |
| 28 | }; | 28 | }; |
diff --git a/src/core/hw/gpu.cpp b/src/core/hw/gpu.cpp index b7102b874..424ce2ca7 100644 --- a/src/core/hw/gpu.cpp +++ b/src/core/hw/gpu.cpp | |||
| @@ -81,9 +81,9 @@ inline void Write(u32 addr, const T data) { | |||
| 81 | if (config.fill_24bit) { | 81 | if (config.fill_24bit) { |
| 82 | // fill with 24-bit values | 82 | // fill with 24-bit values |
| 83 | for (u8* ptr = start; ptr < end; ptr += 3) { | 83 | for (u8* ptr = start; ptr < end; ptr += 3) { |
| 84 | ptr[0] = config.value_24bit_r; | 84 | ptr[0] = config.value_24bit_b; |
| 85 | ptr[1] = config.value_24bit_g; | 85 | ptr[1] = config.value_24bit_g; |
| 86 | ptr[2] = config.value_24bit_b; | 86 | ptr[2] = config.value_24bit_r; |
| 87 | } | 87 | } |
| 88 | } else if (config.fill_32bit) { | 88 | } else if (config.fill_32bit) { |
| 89 | // fill with 32-bit values | 89 | // fill with 32-bit values |
diff --git a/src/core/hw/gpu.h b/src/core/hw/gpu.h index 5ca4a5450..737b1e968 100644 --- a/src/core/hw/gpu.h +++ b/src/core/hw/gpu.h | |||
| @@ -100,10 +100,10 @@ struct Regs { | |||
| 100 | // Set to 1 upon completion. | 100 | // Set to 1 upon completion. |
| 101 | BitField<0, 1, u32> finished; | 101 | BitField<0, 1, u32> finished; |
| 102 | 102 | ||
| 103 | // If both of these bits are unset, then it will fill the memory with a 16 bit value | 103 | // 0: fill with 16- or 32-bit wide values; 1: fill with 24-bit wide values |
| 104 | // 1: fill with 24-bit wide values | ||
| 105 | BitField<8, 1, u32> fill_24bit; | 104 | BitField<8, 1, u32> fill_24bit; |
| 106 | // 1: fill with 32-bit wide values | 105 | |
| 106 | // 0: fill with 16-bit wide values; 1: fill with 32-bit wide values | ||
| 107 | BitField<9, 1, u32> fill_32bit; | 107 | BitField<9, 1, u32> fill_32bit; |
| 108 | }; | 108 | }; |
| 109 | 109 | ||
diff --git a/src/video_core/color.h b/src/video_core/color.h index 35b56efc0..14ade74f2 100644 --- a/src/video_core/color.h +++ b/src/video_core/color.h | |||
| @@ -105,7 +105,7 @@ inline const Math::Vec4<u8> DecodeRGBA4(const u8* bytes) { | |||
| 105 | * @param bytes Pointer to encoded source value | 105 | * @param bytes Pointer to encoded source value |
| 106 | * @return Depth value as an u32 | 106 | * @return Depth value as an u32 |
| 107 | */ | 107 | */ |
| 108 | inline const u32 DecodeD16(const u8* bytes) { | 108 | inline u32 DecodeD16(const u8* bytes) { |
| 109 | return *reinterpret_cast<const u16_le*>(bytes); | 109 | return *reinterpret_cast<const u16_le*>(bytes); |
| 110 | } | 110 | } |
| 111 | 111 | ||
| @@ -114,7 +114,7 @@ inline const u32 DecodeD16(const u8* bytes) { | |||
| 114 | * @param bytes Pointer to encoded source value | 114 | * @param bytes Pointer to encoded source value |
| 115 | * @return Depth value as an u32 | 115 | * @return Depth value as an u32 |
| 116 | */ | 116 | */ |
| 117 | inline const u32 DecodeD24(const u8* bytes) { | 117 | inline u32 DecodeD24(const u8* bytes) { |
| 118 | return (bytes[2] << 16) | (bytes[1] << 8) | bytes[0]; | 118 | return (bytes[2] << 16) | (bytes[1] << 8) | bytes[0]; |
| 119 | } | 119 | } |
| 120 | 120 | ||
| @@ -181,8 +181,8 @@ inline void EncodeRGBA4(const Math::Vec4<u8>& color, u8* bytes) { | |||
| 181 | } | 181 | } |
| 182 | 182 | ||
| 183 | /** | 183 | /** |
| 184 | * Encode a depth value as D16 format | 184 | * Encode a 16 bit depth value as D16 format |
| 185 | * @param value Source depth value to encode | 185 | * @param value 16 bit source depth value to encode |
| 186 | * @param bytes Pointer where to store the encoded value | 186 | * @param bytes Pointer where to store the encoded value |
| 187 | */ | 187 | */ |
| 188 | inline void EncodeD16(u32 value, u8* bytes) { | 188 | inline void EncodeD16(u32 value, u8* bytes) { |
| @@ -190,8 +190,8 @@ inline void EncodeD16(u32 value, u8* bytes) { | |||
| 190 | } | 190 | } |
| 191 | 191 | ||
| 192 | /** | 192 | /** |
| 193 | * Encode a depth value as D24 format | 193 | * Encode a 24 bit depth value as D24 format |
| 194 | * @param value Source depth value to encode | 194 | * @param value 24 bit source depth value to encode |
| 195 | * @param bytes Pointer where to store the encoded value | 195 | * @param bytes Pointer where to store the encoded value |
| 196 | */ | 196 | */ |
| 197 | inline void EncodeD24(u32 value, u8* bytes) { | 197 | inline void EncodeD24(u32 value, u8* bytes) { |
| @@ -201,9 +201,9 @@ inline void EncodeD24(u32 value, u8* bytes) { | |||
| 201 | } | 201 | } |
| 202 | 202 | ||
| 203 | /** | 203 | /** |
| 204 | * Encode depth and stencil values as D24S8 format | 204 | * Encode a 24 bit depth and 8 bit stencil values as D24S8 format |
| 205 | * @param depth Source depth values to encode | 205 | * @param depth 24 bit source depth value to encode |
| 206 | * @param stencil Source stencil value to encode | 206 | * @param stencil 8 bit source stencil value to encode |
| 207 | * @param bytes Pointer where to store the encoded value | 207 | * @param bytes Pointer where to store the encoded value |
| 208 | */ | 208 | */ |
| 209 | inline void EncodeD24S8(u32 depth, u8 stencil, u8* bytes) { | 209 | inline void EncodeD24S8(u32 depth, u8 stencil, u8* bytes) { |
diff --git a/src/video_core/pica.h b/src/video_core/pica.h index 6549693f5..fe20cd77d 100644 --- a/src/video_core/pica.h +++ b/src/video_core/pica.h | |||
| @@ -393,7 +393,15 @@ struct Regs { | |||
| 393 | BitField< 8, 8, u32> ref; | 393 | BitField< 8, 8, u32> ref; |
| 394 | } alpha_test; | 394 | } alpha_test; |
| 395 | 395 | ||
| 396 | INSERT_PADDING_WORDS(0x2); | 396 | union { |
| 397 | BitField< 0, 1, u32> stencil_test_enable; | ||
| 398 | BitField< 4, 3, CompareFunc> stencil_test_func; | ||
| 399 | BitField< 8, 8, u32> stencil_replacement_value; | ||
| 400 | BitField<16, 8, u32> stencil_reference_value; | ||
| 401 | BitField<24, 8, u32> stencil_mask; | ||
| 402 | } stencil_test; | ||
| 403 | |||
| 404 | INSERT_PADDING_WORDS(0x1); | ||
| 397 | 405 | ||
| 398 | union { | 406 | union { |
| 399 | BitField< 0, 1, u32> depth_test_enable; | 407 | BitField< 0, 1, u32> depth_test_enable; |
| @@ -408,6 +416,30 @@ struct Regs { | |||
| 408 | INSERT_PADDING_WORDS(0x8); | 416 | INSERT_PADDING_WORDS(0x8); |
| 409 | } output_merger; | 417 | } output_merger; |
| 410 | 418 | ||
| 419 | enum DepthFormat : u32 { | ||
| 420 | D16 = 0, | ||
| 421 | |||
| 422 | D24 = 2, | ||
| 423 | D24S8 = 3 | ||
| 424 | }; | ||
| 425 | |||
| 426 | /* | ||
| 427 | * Returns the number of bytes in the specified depth format | ||
| 428 | */ | ||
| 429 | static u32 BytesPerDepthPixel(DepthFormat format) { | ||
| 430 | switch (format) { | ||
| 431 | case DepthFormat::D16: | ||
| 432 | return 2; | ||
| 433 | case DepthFormat::D24: | ||
| 434 | return 3; | ||
| 435 | case DepthFormat::D24S8: | ||
| 436 | return 4; | ||
| 437 | default: | ||
| 438 | LOG_CRITICAL(HW_GPU, "Unknown depth format %u", format); | ||
| 439 | UNIMPLEMENTED(); | ||
| 440 | } | ||
| 441 | } | ||
| 442 | |||
| 411 | struct { | 443 | struct { |
| 412 | // Components are laid out in reverse byte order, most significant bits first. | 444 | // Components are laid out in reverse byte order, most significant bits first. |
| 413 | enum ColorFormat : u32 { | 445 | enum ColorFormat : u32 { |
| @@ -418,16 +450,9 @@ struct Regs { | |||
| 418 | RGBA4 = 4, | 450 | RGBA4 = 4, |
| 419 | }; | 451 | }; |
| 420 | 452 | ||
| 421 | enum DepthFormat : u32 { | ||
| 422 | D16 = 0, | ||
| 423 | |||
| 424 | D24 = 2, | ||
| 425 | D24S8 = 3 | ||
| 426 | }; | ||
| 427 | |||
| 428 | INSERT_PADDING_WORDS(0x6); | 453 | INSERT_PADDING_WORDS(0x6); |
| 429 | 454 | ||
| 430 | u32 depth_format; | 455 | DepthFormat depth_format; |
| 431 | BitField<16, 3, u32> color_format; | 456 | BitField<16, 3, u32> color_format; |
| 432 | 457 | ||
| 433 | INSERT_PADDING_WORDS(0x4); | 458 | INSERT_PADDING_WORDS(0x4); |
diff --git a/src/video_core/rasterizer.cpp b/src/video_core/rasterizer.cpp index dc32128c6..dd46f0ec3 100644 --- a/src/video_core/rasterizer.cpp +++ b/src/video_core/rasterizer.cpp | |||
| @@ -100,24 +100,19 @@ static u32 GetDepth(int x, int y) { | |||
| 100 | y = (registers.framebuffer.height - y); | 100 | y = (registers.framebuffer.height - y); |
| 101 | 101 | ||
| 102 | const u32 coarse_y = y & ~7; | 102 | const u32 coarse_y = y & ~7; |
| 103 | u32 bytes_per_pixel = Pica::Regs::BytesPerDepthPixel(registers.framebuffer.depth_format); | ||
| 104 | u32 stride = registers.framebuffer.width * bytes_per_pixel; | ||
| 105 | |||
| 106 | u32 src_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * stride; | ||
| 107 | u8* src_pixel = depth_buffer + src_offset; | ||
| 103 | 108 | ||
| 104 | switch (registers.framebuffer.depth_format) { | 109 | switch (registers.framebuffer.depth_format) { |
| 105 | case registers.framebuffer.D16: | 110 | case Pica::Regs::DepthFormat::D16: |
| 106 | { | 111 | return Color::DecodeD16(src_pixel); |
| 107 | u32 stride = registers.framebuffer.width * 2; | 112 | case Pica::Regs::DepthFormat::D24: |
| 108 | return Color::DecodeD16(depth_buffer + VideoCore::GetMortonOffset(x, y, 2) + coarse_y * stride); | 113 | return Color::DecodeD24(src_pixel); |
| 109 | } | 114 | case Pica::Regs::DepthFormat::D24S8: |
| 110 | case registers.framebuffer.D24: | 115 | return Color::DecodeD24S8(src_pixel).x; |
| 111 | { | ||
| 112 | u32 stride = registers.framebuffer.width * 3; | ||
| 113 | u8* address = depth_buffer + VideoCore::GetMortonOffset(x, y, 3) + coarse_y * stride; | ||
| 114 | return Color::DecodeD24(address); | ||
| 115 | } | ||
| 116 | case registers.framebuffer.D24S8: | ||
| 117 | { | ||
| 118 | u32 stride = registers.framebuffer.width * 4; | ||
| 119 | return Color::DecodeD24S8(depth_buffer + VideoCore::GetMortonOffset(x, y, 4) + coarse_y * stride).x; | ||
| 120 | } | ||
| 121 | default: | 116 | default: |
| 122 | LOG_CRITICAL(HW_GPU, "Unimplemented depth format %u", registers.framebuffer.depth_format); | 117 | LOG_CRITICAL(HW_GPU, "Unimplemented depth format %u", registers.framebuffer.depth_format); |
| 123 | UNIMPLEMENTED(); | 118 | UNIMPLEMENTED(); |
| @@ -132,28 +127,23 @@ static void SetDepth(int x, int y, u32 value) { | |||
| 132 | y = (registers.framebuffer.height - y); | 127 | y = (registers.framebuffer.height - y); |
| 133 | 128 | ||
| 134 | const u32 coarse_y = y & ~7; | 129 | const u32 coarse_y = y & ~7; |
| 130 | u32 bytes_per_pixel = Pica::Regs::BytesPerDepthPixel(registers.framebuffer.depth_format); | ||
| 131 | u32 stride = registers.framebuffer.width * bytes_per_pixel; | ||
| 132 | |||
| 133 | u32 dst_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * stride; | ||
| 134 | u8* dst_pixel = depth_buffer + dst_offset; | ||
| 135 | 135 | ||
| 136 | switch (registers.framebuffer.depth_format) { | 136 | switch (registers.framebuffer.depth_format) { |
| 137 | case registers.framebuffer.D16: | 137 | case Pica::Regs::DepthFormat::D16: |
| 138 | { | 138 | Color::EncodeD16(value, dst_pixel); |
| 139 | u32 stride = registers.framebuffer.width * 2; | ||
| 140 | Color::EncodeD16(value, depth_buffer + VideoCore::GetMortonOffset(x, y, 2) + coarse_y * stride); | ||
| 141 | break; | 139 | break; |
| 142 | } | 140 | case Pica::Regs::DepthFormat::D24: |
| 143 | case registers.framebuffer.D24: | 141 | Color::EncodeD24(value, dst_pixel); |
| 144 | { | ||
| 145 | u32 stride = registers.framebuffer.width * 3; | ||
| 146 | u8* address = depth_buffer + VideoCore::GetMortonOffset(x, y, 3) + coarse_y * stride; | ||
| 147 | Color::EncodeD24(value, address); | ||
| 148 | break; | 142 | break; |
| 149 | } | 143 | case Pica::Regs::DepthFormat::D24S8: |
| 150 | case registers.framebuffer.D24S8: | ||
| 151 | { | ||
| 152 | u32 stride = registers.framebuffer.width * 4; | ||
| 153 | // TODO(Subv): Implement the stencil buffer | 144 | // TODO(Subv): Implement the stencil buffer |
| 154 | Color::EncodeD24S8(value, 0, depth_buffer + VideoCore::GetMortonOffset(x, y, 4) + coarse_y * stride); | 145 | Color::EncodeD24S8(value, 0, dst_pixel); |
| 155 | break; | 146 | break; |
| 156 | } | ||
| 157 | default: | 147 | default: |
| 158 | LOG_CRITICAL(HW_GPU, "Unimplemented depth format %u", registers.framebuffer.depth_format); | 148 | LOG_CRITICAL(HW_GPU, "Unimplemented depth format %u", registers.framebuffer.depth_format); |
| 159 | UNIMPLEMENTED(); | 149 | UNIMPLEMENTED(); |