GPU: Implement additional render target formats.

7 files changed, 126 insertions, 12 deletions

diff --git a/src/video_core/engines/sw_blitter/converter.cpp b/src/video_core/engines/sw_blitter/converter.cpp
index 408d87944..37c5eff69 100644
--- a/src/video_core/engines/sw_blitter/converter.cpp
+++ b/src/video_core/engines/sw_blitter/converter.cpp
@@ -133,13 +133,13 @@ constexpr std::array<f32, 256> RGB_TO_SRGB_LUT = {
 
 } // namespace
 
-struct R32B32G32A32_FLOATTraits {
+struct R32G32B32A32_FLOATTraits {
     static constexpr size_t num_components = 4;
     static constexpr std::array<ComponentType, num_components> component_types = {
         ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT};
     static constexpr std::array<size_t, num_components> component_sizes = {32, 32, 32, 32};
     static constexpr std::array<Swizzle, num_components> component_swizzle = {
-        Swizzle::R, Swizzle::B, Swizzle::G, Swizzle::A};
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::A};
 };
 
 struct R32G32B32A32_SINTTraits {
@@ -160,6 +160,33 @@ struct R32G32B32A32_UINTTraits {
         Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::A};
 };
 
+struct R32G32B32X32_FLOATTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32, 32, 32, 32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::None};
+};
+
+struct R32G32B32X32_SINTTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT, ComponentType::SINT, ComponentType::SINT, ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32, 32, 32, 32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::None};
+};
+
+struct R32G32B32X32_UINTTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT, ComponentType::UINT, ComponentType::UINT, ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32, 32, 32, 32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::None};
+};
+
 struct R16G16B16A16_UNORMTraits {
     static constexpr size_t num_components = 4;
     static constexpr std::array<ComponentType, num_components> component_types = {
@@ -277,6 +304,15 @@ struct A2B10G10R10_UINTTraits {
         Swizzle::A, Swizzle::B, Swizzle::G, Swizzle::R};
 };
 
+struct A2R10G10B10_UNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {2, 10, 10, 10};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::A, Swizzle::R, Swizzle::G, Swizzle::B};
+};
+
 struct A8B8G8R8_UNORMTraits {
     static constexpr size_t num_components = 4;
     static constexpr std::array<ComponentType, num_components> component_types = {
@@ -544,6 +580,33 @@ struct R8_UINTTraits {
     static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
 };
 
+struct X1R5G5B5_UNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {1, 5, 5, 5};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::None, Swizzle::R, Swizzle::G, Swizzle::B};
+};
+
+struct X8B8G8R8_UNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8, 8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::None, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct X8B8G8R8_SRGBTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SRGB, ComponentType::SRGB, ComponentType::SRGB, ComponentType::SRGB};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8, 8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::None, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
 template <class ConverterTraits>
 class ConverterImpl : public Converter {
 private:
@@ -884,9 +947,9 @@ public:
 
 Converter* ConverterFactory::BuildConverter(RenderTargetFormat format) {
     switch (format) {
-    case RenderTargetFormat::R32B32G32A32_FLOAT:
+    case RenderTargetFormat::R32G32B32A32_FLOAT:
         return impl->converters_cache
-            .emplace(format, std::make_unique<ConverterImpl<R32B32G32A32_FLOATTraits>>())
+            .emplace(format, std::make_unique<ConverterImpl<R32G32B32A32_FLOATTraits>>())
             .first->second.get();
         break;
     case RenderTargetFormat::R32G32B32A32_SINT:
@@ -899,6 +962,21 @@ Converter* ConverterFactory::BuildConverter(RenderTargetFormat format) {
             .emplace(format, std::make_unique<ConverterImpl<R32G32B32A32_UINTTraits>>())
             .first->second.get();
         break;
+    case RenderTargetFormat::R32G32B32X32_FLOAT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32G32B32X32_FLOATTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32G32B32X32_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32G32B32X32_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32G32B32X32_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32G32B32X32_UINTTraits>>())
+            .first->second.get();
+        break;
     case RenderTargetFormat::R16G16B16A16_UNORM:
         return impl->converters_cache
             .emplace(format, std::make_unique<ConverterImpl<R16G16B16A16_UNORMTraits>>())
@@ -964,6 +1042,11 @@ Converter* ConverterFactory::BuildConverter(RenderTargetFormat format) {
             .emplace(format, std::make_unique<ConverterImpl<A2B10G10R10_UINTTraits>>())
             .first->second.get();
         break;
+    case RenderTargetFormat::A2R10G10B10_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A2R10G10B10_UNORMTraits>>())
+            .first->second.get();
+        break;
     case RenderTargetFormat::A8B8G8R8_UNORM:
         return impl->converters_cache
             .emplace(format, std::make_unique<ConverterImpl<A8B8G8R8_UNORMTraits>>())
@@ -1119,6 +1202,21 @@ Converter* ConverterFactory::BuildConverter(RenderTargetFormat format) {
             .emplace(format, std::make_unique<ConverterImpl<R8_UINTTraits>>())
             .first->second.get();
         break;
+    case RenderTargetFormat::X1R5G5B5_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<X1R5G5B5_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::X8B8G8R8_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<X8B8G8R8_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::X8B8G8R8_SRGB:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<X8B8G8R8_SRGBTraits>>())
+            .first->second.get();
+        break;
     default: {
         UNIMPLEMENTED_MSG("This format {} converter is not implemented", format);
         return impl->converters_cache.emplace(format, std::make_unique<NullConverter>())
diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h
index 87ebf2054..8a871593a 100644
--- a/src/video_core/gpu.h
+++ b/src/video_core/gpu.h
@@ -27,12 +27,12 @@ struct CommandList;
 // TODO: Implement the commented ones
 enum class RenderTargetFormat : u32 {
     NONE = 0x0,
-    R32B32G32A32_FLOAT = 0xC0,
+    R32G32B32A32_FLOAT = 0xC0,
     R32G32B32A32_SINT = 0xC1,
     R32G32B32A32_UINT = 0xC2,
-    // R32G32B32X32_FLOAT = 0xC3,
-    // R32G32B32X32_SINT = 0xC4,
-    // R32G32B32X32_UINT = 0xC5,
+    R32G32B32X32_FLOAT = 0xC3,
+    R32G32B32X32_SINT = 0xC4,
+    R32G32B32X32_UINT = 0xC5,
     R16G16B16A16_UNORM = 0xC6,
     R16G16B16A16_SNORM = 0xC7,
     R16G16B16A16_SINT = 0xC8,
@@ -56,7 +56,7 @@ enum class RenderTargetFormat : u32 {
     R16G16_SINT = 0xDC,
     R16G16_UINT = 0xDD,
     R16G16_FLOAT = 0xDE,
-    // A2R10G10B10_UNORM = 0xDF,
+    A2R10G10B10_UNORM = 0xDF,
     B10G11R11_FLOAT = 0xE0,
     R32_SINT = 0xE3,
     R32_UINT = 0xE4,
@@ -79,11 +79,11 @@ enum class RenderTargetFormat : u32 {
     R8_SINT = 0xF5,
     R8_UINT = 0xF6,
 
-    /*
-    A8_UNORM = 0xF7,
+    // A8_UNORM = 0xF7,
     X1R5G5B5_UNORM = 0xF8,
     X8B8G8R8_UNORM = 0xF9,
     X8B8G8R8_SRGB = 0xFA,
+    /*
     Z1R5G5B5_UNORM = 0xFB,
     O1R5G5B5_UNORM = 0xFC,
     Z8R8G8B8_UNORM = 0xFD,
diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h
index e14f9b2db..ef1190e1f 100644
--- a/src/video_core/renderer_opengl/maxwell_to_gl.h
+++ b/src/video_core/renderer_opengl/maxwell_to_gl.h
@@ -28,6 +28,7 @@ constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> FORMAT_TAB
     {GL_RGB5_A1, GL_BGRA, GL_UNSIGNED_SHORT_1_5_5_5_REV},             // A1R5G5B5_UNORM
     {GL_RGB10_A2, GL_RGBA, GL_UNSIGNED_INT_2_10_10_10_REV},           // A2B10G10R10_UNORM
     {GL_RGB10_A2UI, GL_RGBA_INTEGER, GL_UNSIGNED_INT_2_10_10_10_REV}, // A2B10G10R10_UINT
+    {GL_RGB10_A2, GL_BGRA, GL_UNSIGNED_INT_2_10_10_10_REV},           // A2R10G10B10_UNORM
     {GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_1_5_5_5_REV},             // A1B5G5R5_UNORM
     {GL_RGB5_A1, GL_RGBA, GL_UNSIGNED_SHORT_5_5_5_1},                 // A5B5G5R1_UNORM
     {GL_R8, GL_RED, GL_UNSIGNED_BYTE},                                // R8_UNORM
diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp
index 5c156087b..1da53f203 100644
--- a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp
+++ b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp
@@ -125,6 +125,7 @@ struct FormatTuple {
     {VK_FORMAT_A1R5G5B5_UNORM_PACK16, Attachable},              // A1R5G5B5_UNORM
     {VK_FORMAT_A2B10G10R10_UNORM_PACK32, Attachable | Storage}, // A2B10G10R10_UNORM
     {VK_FORMAT_A2B10G10R10_UINT_PACK32, Attachable | Storage},  // A2B10G10R10_UINT
+    {VK_FORMAT_A2R10G10B10_UNORM_PACK32, Attachable | Storage}, // A2R10G10B10_UNORM
     {VK_FORMAT_A1R5G5B5_UNORM_PACK16, Attachable},         // A1B5G5R5_UNORM (flipped with swizzle)
     {VK_FORMAT_R5G5B5A1_UNORM_PACK16},                     // A5B5G5R1_UNORM (specially swizzled)
     {VK_FORMAT_R8_UNORM, Attachable | Storage},            // R8_UNORM
diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp
index 80a7d908f..b618e1a25 100644
--- a/src/video_core/surface.cpp
+++ b/src/video_core/surface.cpp
@@ -93,11 +93,14 @@ PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
 
 PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
     switch (format) {
-    case Tegra::RenderTargetFormat::R32B32G32A32_FLOAT:
+    case Tegra::RenderTargetFormat::R32G32B32A32_FLOAT:
+    case Tegra::RenderTargetFormat::R32G32B32X32_FLOAT:
         return PixelFormat::R32G32B32A32_FLOAT;
     case Tegra::RenderTargetFormat::R32G32B32A32_SINT:
+    case Tegra::RenderTargetFormat::R32G32B32X32_SINT:
         return PixelFormat::R32G32B32A32_SINT;
     case Tegra::RenderTargetFormat::R32G32B32A32_UINT:
+    case Tegra::RenderTargetFormat::R32G32B32X32_UINT:
         return PixelFormat::R32G32B32A32_UINT;
     case Tegra::RenderTargetFormat::R16G16B16A16_UNORM:
         return PixelFormat::R16G16B16A16_UNORM;
@@ -127,9 +130,13 @@ PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format)
         return PixelFormat::A2B10G10R10_UNORM;
     case Tegra::RenderTargetFormat::A2B10G10R10_UINT:
         return PixelFormat::A2B10G10R10_UINT;
+    case Tegra::RenderTargetFormat::A2R10G10B10_UNORM:
+        return PixelFormat::A2R10G10B10_UNORM;
     case Tegra::RenderTargetFormat::A8B8G8R8_UNORM:
+    case Tegra::RenderTargetFormat::X8B8G8R8_UNORM:
         return PixelFormat::A8B8G8R8_UNORM;
     case Tegra::RenderTargetFormat::A8B8G8R8_SRGB:
+    case Tegra::RenderTargetFormat::X8B8G8R8_SRGB:
         return PixelFormat::A8B8G8R8_SRGB;
     case Tegra::RenderTargetFormat::A8B8G8R8_SNORM:
         return PixelFormat::A8B8G8R8_SNORM;
@@ -158,6 +165,7 @@ PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format)
     case Tegra::RenderTargetFormat::R5G6B5_UNORM:
         return PixelFormat::R5G6B5_UNORM;
     case Tegra::RenderTargetFormat::A1R5G5B5_UNORM:
+    case Tegra::RenderTargetFormat::X1R5G5B5_UNORM:
         return PixelFormat::A1R5G5B5_UNORM;
     case Tegra::RenderTargetFormat::R8G8_UNORM:
         return PixelFormat::R8G8_UNORM;
diff --git a/src/video_core/surface.h b/src/video_core/surface.h
index 57ca7f597..44b79af20 100644
--- a/src/video_core/surface.h
+++ b/src/video_core/surface.h
@@ -23,6 +23,7 @@ enum class PixelFormat {
     A1R5G5B5_UNORM,
     A2B10G10R10_UNORM,
     A2B10G10R10_UINT,
+    A2R10G10B10_UNORM,
     A1B5G5R5_UNORM,
     A5B5G5R1_UNORM,
     R8_UNORM,
@@ -159,6 +160,7 @@ constexpr std::array<u8, MaxPixelFormat> BLOCK_WIDTH_TABLE = {{
     1,  // A1R5G5B5_UNORM
     1,  // A2B10G10R10_UNORM
     1,  // A2B10G10R10_UINT
+    1,  // A2R10G10B10_UNORM
     1,  // A1B5G5R5_UNORM
     1,  // A5B5G5R1_UNORM
     1,  // R8_UNORM
@@ -264,6 +266,7 @@ constexpr std::array<u8, MaxPixelFormat> BLOCK_HEIGHT_TABLE = {{
     1,  // A1R5G5B5_UNORM
     1,  // A2B10G10R10_UNORM
     1,  // A2B10G10R10_UINT
+    1,  // A2R10G10B10_UNORM
     1,  // A1B5G5R5_UNORM
     1,  // A5B5G5R1_UNORM
     1,  // R8_UNORM
@@ -369,6 +372,7 @@ constexpr std::array<u8, MaxPixelFormat> BITS_PER_BLOCK_TABLE = {{
     16,  // A1R5G5B5_UNORM
     32,  // A2B10G10R10_UNORM
     32,  // A2B10G10R10_UINT
+    32,  // A2R10G10B10_UNORM
     16,  // A1B5G5R5_UNORM
     16,  // A5B5G5R1_UNORM
     8,   // R8_UNORM
diff --git a/src/video_core/texture_cache/formatter.h b/src/video_core/texture_cache/formatter.h
index acc854715..f1f0a057b 100644
--- a/src/video_core/texture_cache/formatter.h
+++ b/src/video_core/texture_cache/formatter.h
@@ -35,6 +35,8 @@ struct fmt::formatter<VideoCore::Surface::PixelFormat> : fmt::formatter<fmt::str
                 return "A2B10G10R10_UNORM";
             case PixelFormat::A2B10G10R10_UINT:
                 return "A2B10G10R10_UINT";
+            case PixelFormat::A2R10G10B10_UNORM:
+                return "A2R10G10B10_UNORM";
             case PixelFormat::A1B5G5R5_UNORM:
                 return "A1B5G5R5_UNORM";
             case PixelFormat::A5B5G5R1_UNORM: