summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat (limited to '')
m---------externals/boost0
-rw-r--r--src/core/hle/kernel/svc.cpp5
-rw-r--r--src/core/hle/kernel/svc_wrap.h10
-rw-r--r--src/video_core/rasterizer_accelerated.cpp1
-rw-r--r--src/video_core/renderer_opengl/gl_device.cpp5
-rw-r--r--src/video_core/renderer_opengl/gl_shader_cache.cpp4
-rw-r--r--src/video_core/renderer_vulkan/maxwell_to_vk.cpp329
-rw-r--r--src/video_core/renderer_vulkan/maxwell_to_vk.h18
-rw-r--r--src/video_core/renderer_vulkan/vk_sampler_cache.cpp7
9 files changed, 256 insertions, 123 deletions
diff --git a/externals/boost b/externals/boost
-Subproject 0b920df1c90f4eb84e2f2b10bee8804b0f6472b
+Subproject 5e8300b76a627f3a1ba215304e04ead33b5bc23
diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp
index 35ff26c39..dbcdb0b88 100644
--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -1650,8 +1650,7 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add
}
/// Signal process wide key
-static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_variable_addr,
- s32 target) {
+static void SignalProcessWideKey(Core::System& system, VAddr condition_variable_addr, s32 target) {
LOG_TRACE(Kernel_SVC, "called, condition_variable_addr=0x{:X}, target=0x{:08X}",
condition_variable_addr, target);
@@ -1726,8 +1725,6 @@ static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_var
system.PrepareReschedule(thread->GetProcessorID());
}
}
-
- return RESULT_SUCCESS;
}
// Wait for an address (via Address Arbiter)
diff --git a/src/core/hle/kernel/svc_wrap.h b/src/core/hle/kernel/svc_wrap.h
index 9452e3b6f..29a2cfa9d 100644
--- a/src/core/hle/kernel/svc_wrap.h
+++ b/src/core/hle/kernel/svc_wrap.h
@@ -112,11 +112,6 @@ void SvcWrap(Core::System& system) {
FuncReturn(system, retval);
}
-template <ResultCode func(Core::System&, u64, s32)>
-void SvcWrap(Core::System& system) {
- FuncReturn(system, func(system, Param(system, 0), static_cast<s32>(Param(system, 1))).raw);
-}
-
template <ResultCode func(Core::System&, u64, u32)>
void SvcWrap(Core::System& system) {
FuncReturn(system, func(system, Param(system, 0), static_cast<u32>(Param(system, 1))).raw);
@@ -327,6 +322,11 @@ void SvcWrap(Core::System& system) {
func(system, static_cast<s64>(Param(system, 0)));
}
+template <void func(Core::System&, u64, s32)>
+void SvcWrap(Core::System& system) {
+ func(system, Param(system, 0), static_cast<s32>(Param(system, 1)));
+}
+
template <void func(Core::System&, u64, u64)>
void SvcWrap(Core::System& system) {
func(system, Param(system, 0), Param(system, 1));
diff --git a/src/video_core/rasterizer_accelerated.cpp b/src/video_core/rasterizer_accelerated.cpp
index fc6ecb899..d01db97da 100644
--- a/src/video_core/rasterizer_accelerated.cpp
+++ b/src/video_core/rasterizer_accelerated.cpp
@@ -5,6 +5,7 @@
#include <mutex>
#include <boost/icl/interval_map.hpp>
+#include <boost/range/iterator_range.hpp>
#include "common/assert.h"
#include "common/common_types.h"
diff --git a/src/video_core/renderer_opengl/gl_device.cpp b/src/video_core/renderer_opengl/gl_device.cpp
index 413d8546b..1a2e2a9f7 100644
--- a/src/video_core/renderer_opengl/gl_device.cpp
+++ b/src/video_core/renderer_opengl/gl_device.cpp
@@ -5,6 +5,7 @@
#include <algorithm>
#include <array>
#include <cstddef>
+#include <cstring>
#include <optional>
#include <vector>
@@ -134,11 +135,13 @@ std::array<Device::BaseBindings, Tegra::Engines::MaxShaderTypes> BuildBaseBindin
Device::Device() : base_bindings{BuildBaseBindings()} {
const std::string_view vendor = reinterpret_cast<const char*>(glGetString(GL_VENDOR));
+ const auto renderer = reinterpret_cast<const char*>(glGetString(GL_RENDERER));
const std::vector extensions = GetExtensions();
const bool is_nvidia = vendor == "NVIDIA Corporation";
const bool is_amd = vendor == "ATI Technologies Inc.";
const bool is_intel = vendor == "Intel";
+ const bool is_intel_proprietary = is_intel && std::strstr(renderer, "Mesa") == nullptr;
uniform_buffer_alignment = GetInteger<std::size_t>(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT);
shader_storage_alignment = GetInteger<std::size_t>(GL_SHADER_STORAGE_BUFFER_OFFSET_ALIGNMENT);
@@ -152,7 +155,7 @@ Device::Device() : base_bindings{BuildBaseBindings()} {
has_variable_aoffi = TestVariableAoffi();
has_component_indexing_bug = is_amd;
has_precise_bug = TestPreciseBug();
- has_broken_compute = is_intel;
+ has_broken_compute = is_intel_proprietary;
has_fast_buffer_sub_data = is_nvidia;
LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", has_variable_aoffi);
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index 370bdf052..270a9dc2b 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -281,11 +281,11 @@ CachedProgram BuildShader(const Device& device, u64 unique_identifier, ShaderTyp
if (variant.shared_memory_size > 0) {
// TODO(Rodrigo): We should divide by four here, but having a larger shared memory pool
// avoids out of bound stores. Find out why shared memory size is being invalid.
- source += fmt::format("shared uint smem[{}];", variant.shared_memory_size);
+ source += fmt::format("shared uint smem[{}];\n", variant.shared_memory_size);
}
if (variant.local_memory_size > 0) {
- source += fmt::format("#define LOCAL_MEMORY_SIZE {}",
+ source += fmt::format("#define LOCAL_MEMORY_SIZE {}\n",
Common::AlignUp(variant.local_memory_size, 4) / 4);
}
}
diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp
index 7f0eb6b74..000e3616d 100644
--- a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp
+++ b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp
@@ -44,7 +44,8 @@ vk::SamplerMipmapMode MipmapMode(Tegra::Texture::TextureMipmapFilter mipmap_filt
return {};
}
-vk::SamplerAddressMode WrapMode(Tegra::Texture::WrapMode wrap_mode) {
+vk::SamplerAddressMode WrapMode(Tegra::Texture::WrapMode wrap_mode,
+ Tegra::Texture::TextureFilter filter) {
switch (wrap_mode) {
case Tegra::Texture::WrapMode::Wrap:
return vk::SamplerAddressMode::eRepeat;
@@ -55,10 +56,15 @@ vk::SamplerAddressMode WrapMode(Tegra::Texture::WrapMode wrap_mode) {
case Tegra::Texture::WrapMode::Border:
return vk::SamplerAddressMode::eClampToBorder;
case Tegra::Texture::WrapMode::Clamp:
- // TODO(Rodrigo): GL_CLAMP was removed as of OpenGL 3.1, to implement GL_CLAMP, we can use
- // eClampToBorder to get the border color of the texture, and then sample the edge to
- // manually mix them. However the shader part of this is not yet implemented.
- return vk::SamplerAddressMode::eClampToBorder;
+ // TODO(Rodrigo): Emulate GL_CLAMP properly
+ switch (filter) {
+ case Tegra::Texture::TextureFilter::Nearest:
+ return vk::SamplerAddressMode::eClampToEdge;
+ case Tegra::Texture::TextureFilter::Linear:
+ return vk::SamplerAddressMode::eClampToBorder;
+ }
+ UNREACHABLE();
+ return vk::SamplerAddressMode::eClampToEdge;
case Tegra::Texture::WrapMode::MirrorOnceClampToEdge:
return vk::SamplerAddressMode::eMirrorClampToEdge;
case Tegra::Texture::WrapMode::MirrorOnceBorder:
@@ -96,106 +102,140 @@ vk::CompareOp DepthCompareFunction(Tegra::Texture::DepthCompareFunc depth_compar
} // namespace Sampler
+namespace {
+
+enum : u32 { Attachable = 1, Storage = 2 };
+
struct FormatTuple {
vk::Format format; ///< Vulkan format
- bool attachable; ///< True when this format can be used as an attachment
-};
-
-static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format_tuples = {{
- {vk::Format::eA8B8G8R8UnormPack32, true}, // ABGR8U
- {vk::Format::eUndefined, false}, // ABGR8S
- {vk::Format::eUndefined, false}, // ABGR8UI
- {vk::Format::eB5G6R5UnormPack16, false}, // B5G6R5U
- {vk::Format::eA2B10G10R10UnormPack32, true}, // A2B10G10R10U
- {vk::Format::eUndefined, false}, // A1B5G5R5U
- {vk::Format::eR8Unorm, true}, // R8U
- {vk::Format::eUndefined, false}, // R8UI
- {vk::Format::eUndefined, false}, // RGBA16F
- {vk::Format::eUndefined, false}, // RGBA16U
- {vk::Format::eUndefined, false}, // RGBA16UI
- {vk::Format::eUndefined, false}, // R11FG11FB10F
- {vk::Format::eUndefined, false}, // RGBA32UI
- {vk::Format::eBc1RgbaUnormBlock, false}, // DXT1
- {vk::Format::eBc2UnormBlock, false}, // DXT23
- {vk::Format::eBc3UnormBlock, false}, // DXT45
- {vk::Format::eBc4UnormBlock, false}, // DXN1
- {vk::Format::eUndefined, false}, // DXN2UNORM
- {vk::Format::eUndefined, false}, // DXN2SNORM
- {vk::Format::eUndefined, false}, // BC7U
- {vk::Format::eUndefined, false}, // BC6H_UF16
- {vk::Format::eUndefined, false}, // BC6H_SF16
- {vk::Format::eUndefined, false}, // ASTC_2D_4X4
- {vk::Format::eUndefined, false}, // BGRA8
- {vk::Format::eUndefined, false}, // RGBA32F
- {vk::Format::eUndefined, false}, // RG32F
- {vk::Format::eUndefined, false}, // R32F
- {vk::Format::eUndefined, false}, // R16F
- {vk::Format::eUndefined, false}, // R16U
- {vk::Format::eUndefined, false}, // R16S
- {vk::Format::eUndefined, false}, // R16UI
- {vk::Format::eUndefined, false}, // R16I
- {vk::Format::eUndefined, false}, // RG16
- {vk::Format::eUndefined, false}, // RG16F
- {vk::Format::eUndefined, false}, // RG16UI
- {vk::Format::eUndefined, false}, // RG16I
- {vk::Format::eUndefined, false}, // RG16S
- {vk::Format::eUndefined, false}, // RGB32F
- {vk::Format::eA8B8G8R8SrgbPack32, true}, // RGBA8_SRGB
- {vk::Format::eUndefined, false}, // RG8U
- {vk::Format::eUndefined, false}, // RG8S
- {vk::Format::eUndefined, false}, // RG32UI
- {vk::Format::eUndefined, false}, // RGBX16F
- {vk::Format::eUndefined, false}, // R32UI
- {vk::Format::eUndefined, false}, // ASTC_2D_8X8
- {vk::Format::eUndefined, false}, // ASTC_2D_8X5
- {vk::Format::eUndefined, false}, // ASTC_2D_5X4
-
- // Compressed sRGB formats
- {vk::Format::eUndefined, false}, // BGRA8_SRGB
- {vk::Format::eUndefined, false}, // DXT1_SRGB
- {vk::Format::eUndefined, false}, // DXT23_SRGB
- {vk::Format::eUndefined, false}, // DXT45_SRGB
- {vk::Format::eUndefined, false}, // BC7U_SRGB
- {vk::Format::eUndefined, false}, // ASTC_2D_4X4_SRGB
- {vk::Format::eUndefined, false}, // ASTC_2D_8X8_SRGB
- {vk::Format::eUndefined, false}, // ASTC_2D_8X5_SRGB
- {vk::Format::eUndefined, false}, // ASTC_2D_5X4_SRGB
- {vk::Format::eUndefined, false}, // ASTC_2D_5X5
- {vk::Format::eUndefined, false}, // ASTC_2D_5X5_SRGB
- {vk::Format::eUndefined, false}, // ASTC_2D_10X8
- {vk::Format::eUndefined, false}, // ASTC_2D_10X8_SRGB
+ int usage; ///< Describes image format usage
+} constexpr tex_format_tuples[] = {
+ {vk::Format::eA8B8G8R8UnormPack32, Attachable | Storage}, // ABGR8U
+ {vk::Format::eA8B8G8R8SnormPack32, Attachable | Storage}, // ABGR8S
+ {vk::Format::eA8B8G8R8UintPack32, Attachable | Storage}, // ABGR8UI
+ {vk::Format::eB5G6R5UnormPack16, {}}, // B5G6R5U
+ {vk::Format::eA2B10G10R10UnormPack32, Attachable | Storage}, // A2B10G10R10U
+ {vk::Format::eA1R5G5B5UnormPack16, Attachable | Storage}, // A1B5G5R5U (flipped with swizzle)
+ {vk::Format::eR8Unorm, Attachable | Storage}, // R8U
+ {vk::Format::eR8Uint, Attachable | Storage}, // R8UI
+ {vk::Format::eR16G16B16A16Sfloat, Attachable | Storage}, // RGBA16F
+ {vk::Format::eR16G16B16A16Unorm, Attachable | Storage}, // RGBA16U
+ {vk::Format::eR16G16B16A16Uint, Attachable | Storage}, // RGBA16UI
+ {vk::Format::eB10G11R11UfloatPack32, Attachable | Storage}, // R11FG11FB10F
+ {vk::Format::eR32G32B32A32Uint, Attachable | Storage}, // RGBA32UI
+ {vk::Format::eBc1RgbaUnormBlock, {}}, // DXT1
+ {vk::Format::eBc2UnormBlock, {}}, // DXT23
+ {vk::Format::eBc3UnormBlock, {}}, // DXT45
+ {vk::Format::eBc4UnormBlock, {}}, // DXN1
+ {vk::Format::eBc5UnormBlock, {}}, // DXN2UNORM
+ {vk::Format::eBc5SnormBlock, {}}, // DXN2SNORM
+ {vk::Format::eBc7UnormBlock, {}}, // BC7U
+ {vk::Format::eBc6HUfloatBlock, {}}, // BC6H_UF16
+ {vk::Format::eBc6HSfloatBlock, {}}, // BC6H_SF16
+ {vk::Format::eAstc4x4UnormBlock, {}}, // ASTC_2D_4X4
+ {vk::Format::eB8G8R8A8Unorm, {}}, // BGRA8
+ {vk::Format::eR32G32B32A32Sfloat, Attachable | Storage}, // RGBA32F
+ {vk::Format::eR32G32Sfloat, Attachable | Storage}, // RG32F
+ {vk::Format::eR32Sfloat, Attachable | Storage}, // R32F
+ {vk::Format::eR16Sfloat, Attachable | Storage}, // R16F
+ {vk::Format::eR16Unorm, Attachable | Storage}, // R16U
+ {vk::Format::eUndefined, {}}, // R16S
+ {vk::Format::eUndefined, {}}, // R16UI
+ {vk::Format::eUndefined, {}}, // R16I
+ {vk::Format::eR16G16Unorm, Attachable | Storage}, // RG16
+ {vk::Format::eR16G16Sfloat, Attachable | Storage}, // RG16F
+ {vk::Format::eUndefined, {}}, // RG16UI
+ {vk::Format::eUndefined, {}}, // RG16I
+ {vk::Format::eR16G16Snorm, Attachable | Storage}, // RG16S
+ {vk::Format::eUndefined, {}}, // RGB32F
+ {vk::Format::eR8G8B8A8Srgb, Attachable}, // RGBA8_SRGB
+ {vk::Format::eR8G8Unorm, Attachable | Storage}, // RG8U
+ {vk::Format::eR8G8Snorm, Attachable | Storage}, // RG8S
+ {vk::Format::eR32G32Uint, Attachable | Storage}, // RG32UI
+ {vk::Format::eUndefined, {}}, // RGBX16F
+ {vk::Format::eR32Uint, Attachable | Storage}, // R32UI
+ {vk::Format::eAstc8x8UnormBlock, {}}, // ASTC_2D_8X8
+ {vk::Format::eUndefined, {}}, // ASTC_2D_8X5
+ {vk::Format::eUndefined, {}}, // ASTC_2D_5X4
+ {vk::Format::eUndefined, {}}, // BGRA8_SRGB
+ {vk::Format::eBc1RgbaSrgbBlock, {}}, // DXT1_SRGB
+ {vk::Format::eUndefined, {}}, // DXT23_SRGB
+ {vk::Format::eBc3SrgbBlock, {}}, // DXT45_SRGB
+ {vk::Format::eBc7SrgbBlock, {}}, // BC7U_SRGB
+ {vk::Format::eR4G4B4A4UnormPack16, Attachable}, // R4G4B4A4U
+ {vk::Format::eAstc4x4SrgbBlock, {}}, // ASTC_2D_4X4_SRGB
+ {vk::Format::eAstc8x8SrgbBlock, {}}, // ASTC_2D_8X8_SRGB
+ {vk::Format::eAstc8x5SrgbBlock, {}}, // ASTC_2D_8X5_SRGB
+ {vk::Format::eAstc5x4SrgbBlock, {}}, // ASTC_2D_5X4_SRGB
+ {vk::Format::eAstc5x5UnormBlock, {}}, // ASTC_2D_5X5
+ {vk::Format::eAstc5x5SrgbBlock, {}}, // ASTC_2D_5X5_SRGB
+ {vk::Format::eAstc10x8UnormBlock, {}}, // ASTC_2D_10X8
+ {vk::Format::eAstc10x8SrgbBlock, {}}, // ASTC_2D_10X8_SRGB
+ {vk::Format::eAstc6x6UnormBlock, {}}, // ASTC_2D_6X6
+ {vk::Format::eAstc6x6SrgbBlock, {}}, // ASTC_2D_6X6_SRGB
+ {vk::Format::eAstc10x10UnormBlock, {}}, // ASTC_2D_10X10
+ {vk::Format::eAstc10x10SrgbBlock, {}}, // ASTC_2D_10X10_SRGB
+ {vk::Format::eAstc12x12UnormBlock, {}}, // ASTC_2D_12X12
+ {vk::Format::eAstc12x12SrgbBlock, {}}, // ASTC_2D_12X12_SRGB
+ {vk::Format::eAstc8x6UnormBlock, {}}, // ASTC_2D_8X6
+ {vk::Format::eAstc8x6SrgbBlock, {}}, // ASTC_2D_8X6_SRGB
+ {vk::Format::eAstc6x5UnormBlock, {}}, // ASTC_2D_6X5
+ {vk::Format::eAstc6x5SrgbBlock, {}}, // ASTC_2D_6X5_SRGB
+ {vk::Format::eE5B9G9R9UfloatPack32, {}}, // E5B9G9R9F
// Depth formats
- {vk::Format::eD32Sfloat, true}, // Z32F
- {vk::Format::eD16Unorm, true}, // Z16
+ {vk::Format::eD32Sfloat, Attachable}, // Z32F
+ {vk::Format::eD16Unorm, Attachable}, // Z16
// DepthStencil formats
- {vk::Format::eD24UnormS8Uint, true}, // Z24S8
- {vk::Format::eD24UnormS8Uint, true}, // S8Z24 (emulated)
- {vk::Format::eUndefined, false}, // Z32FS8
-}};
+ {vk::Format::eD24UnormS8Uint, Attachable}, // Z24S8
+ {vk::Format::eD24UnormS8Uint, Attachable}, // S8Z24 (emulated)
+ {vk::Format::eD32SfloatS8Uint, Attachable}, // Z32FS8
+};
+static_assert(std::size(tex_format_tuples) == VideoCore::Surface::MaxPixelFormat);
-static constexpr bool IsZetaFormat(PixelFormat pixel_format) {
+constexpr bool IsZetaFormat(PixelFormat pixel_format) {
return pixel_format >= PixelFormat::MaxColorFormat &&
pixel_format < PixelFormat::MaxDepthStencilFormat;
}
-std::pair<vk::Format, bool> SurfaceFormat(const VKDevice& device, FormatType format_type,
- PixelFormat pixel_format) {
- ASSERT(static_cast<std::size_t>(pixel_format) < tex_format_tuples.size());
+} // Anonymous namespace
+
+FormatInfo SurfaceFormat(const VKDevice& device, FormatType format_type, PixelFormat pixel_format) {
+ ASSERT(static_cast<std::size_t>(pixel_format) < std::size(tex_format_tuples));
- const auto tuple = tex_format_tuples[static_cast<u32>(pixel_format)];
- UNIMPLEMENTED_IF_MSG(tuple.format == vk::Format::eUndefined,
- "Unimplemented texture format with pixel format={}",
- static_cast<u32>(pixel_format));
+ auto tuple = tex_format_tuples[static_cast<std::size_t>(pixel_format)];
+ if (tuple.format == vk::Format::eUndefined) {
+ UNIMPLEMENTED_MSG("Unimplemented texture format with pixel format={}",
+ static_cast<u32>(pixel_format));
+ return {vk::Format::eA8B8G8R8UnormPack32, true, true};
+ }
+
+ // Use ABGR8 on hardware that doesn't support ASTC natively
+ if (!device.IsOptimalAstcSupported() && VideoCore::Surface::IsPixelFormatASTC(pixel_format)) {
+ tuple.format = VideoCore::Surface::IsPixelFormatSRGB(pixel_format)
+ ? vk::Format::eA8B8G8R8SrgbPack32
+ : vk::Format::eA8B8G8R8UnormPack32;
+ }
+ const bool attachable = tuple.usage & Attachable;
+ const bool storage = tuple.usage & Storage;
- auto usage = vk::FormatFeatureFlagBits::eSampledImage |
- vk::FormatFeatureFlagBits::eTransferDst | vk::FormatFeatureFlagBits::eTransferSrc;
- if (tuple.attachable) {
- usage |= IsZetaFormat(pixel_format) ? vk::FormatFeatureFlagBits::eDepthStencilAttachment
- : vk::FormatFeatureFlagBits::eColorAttachment;
+ vk::FormatFeatureFlags usage;
+ if (format_type == FormatType::Buffer) {
+ usage = vk::FormatFeatureFlagBits::eStorageTexelBuffer |
+ vk::FormatFeatureFlagBits::eUniformTexelBuffer;
+ } else {
+ usage = vk::FormatFeatureFlagBits::eSampledImage | vk::FormatFeatureFlagBits::eTransferDst |
+ vk::FormatFeatureFlagBits::eTransferSrc;
+ if (attachable) {
+ usage |= IsZetaFormat(pixel_format) ? vk::FormatFeatureFlagBits::eDepthStencilAttachment
+ : vk::FormatFeatureFlagBits::eColorAttachment;
+ }
+ if (storage) {
+ usage |= vk::FormatFeatureFlagBits::eStorageImage;
+ }
}
- return {device.GetSupportedFormat(tuple.format, usage, format_type), tuple.attachable};
+ return {device.GetSupportedFormat(tuple.format, usage, format_type), attachable, storage};
}
vk::ShaderStageFlagBits ShaderStage(Tegra::Engines::ShaderType stage) {
@@ -215,7 +255,8 @@ vk::ShaderStageFlagBits ShaderStage(Tegra::Engines::ShaderType stage) {
return {};
}
-vk::PrimitiveTopology PrimitiveTopology(Maxwell::PrimitiveTopology topology) {
+vk::PrimitiveTopology PrimitiveTopology([[maybe_unused]] const VKDevice& device,
+ Maxwell::PrimitiveTopology topology) {
switch (topology) {
case Maxwell::PrimitiveTopology::Points:
return vk::PrimitiveTopology::ePointList;
@@ -227,6 +268,13 @@ vk::PrimitiveTopology PrimitiveTopology(Maxwell::PrimitiveTopology topology) {
return vk::PrimitiveTopology::eTriangleList;
case Maxwell::PrimitiveTopology::TriangleStrip:
return vk::PrimitiveTopology::eTriangleStrip;
+ case Maxwell::PrimitiveTopology::TriangleFan:
+ return vk::PrimitiveTopology::eTriangleFan;
+ case Maxwell::PrimitiveTopology::Quads:
+ // TODO(Rodrigo): Use VK_PRIMITIVE_TOPOLOGY_QUAD_LIST_EXT whenever it releases
+ return vk::PrimitiveTopology::eTriangleList;
+ case Maxwell::PrimitiveTopology::Patches:
+ return vk::PrimitiveTopology::ePatchList;
default:
UNIMPLEMENTED_MSG("Unimplemented topology={}", static_cast<u32>(topology));
return {};
@@ -236,37 +284,111 @@ vk::PrimitiveTopology PrimitiveTopology(Maxwell::PrimitiveTopology topology) {
vk::Format VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttribute::Size size) {
switch (type) {
case Maxwell::VertexAttribute::Type::SignedNorm:
+ switch (size) {
+ case Maxwell::VertexAttribute::Size::Size_8:
+ return vk::Format::eR8Snorm;
+ case Maxwell::VertexAttribute::Size::Size_8_8:
+ return vk::Format::eR8G8Snorm;
+ case Maxwell::VertexAttribute::Size::Size_8_8_8:
+ return vk::Format::eR8G8B8Snorm;
+ case Maxwell::VertexAttribute::Size::Size_8_8_8_8:
+ return vk::Format::eR8G8B8A8Snorm;
+ case Maxwell::VertexAttribute::Size::Size_16:
+ return vk::Format::eR16Snorm;
+ case Maxwell::VertexAttribute::Size::Size_16_16:
+ return vk::Format::eR16G16Snorm;
+ case Maxwell::VertexAttribute::Size::Size_16_16_16:
+ return vk::Format::eR16G16B16Snorm;
+ case Maxwell::VertexAttribute::Size::Size_16_16_16_16:
+ return vk::Format::eR16G16B16A16Snorm;
+ case Maxwell::VertexAttribute::Size::Size_10_10_10_2:
+ return vk::Format::eA2B10G10R10SnormPack32;
+ default:
+ break;
+ }
break;
case Maxwell::VertexAttribute::Type::UnsignedNorm:
switch (size) {
+ case Maxwell::VertexAttribute::Size::Size_8:
+ return vk::Format::eR8Unorm;
+ case Maxwell::VertexAttribute::Size::Size_8_8:
+ return vk::Format::eR8G8Unorm;
+ case Maxwell::VertexAttribute::Size::Size_8_8_8:
+ return vk::Format::eR8G8B8Unorm;
case Maxwell::VertexAttribute::Size::Size_8_8_8_8:
return vk::Format::eR8G8B8A8Unorm;
+ case Maxwell::VertexAttribute::Size::Size_16:
+ return vk::Format::eR16Unorm;
+ case Maxwell::VertexAttribute::Size::Size_16_16:
+ return vk::Format::eR16G16Unorm;
+ case Maxwell::VertexAttribute::Size::Size_16_16_16:
+ return vk::Format::eR16G16B16Unorm;
+ case Maxwell::VertexAttribute::Size::Size_16_16_16_16:
+ return vk::Format::eR16G16B16A16Unorm;
default:
break;
}
break;
case Maxwell::VertexAttribute::Type::SignedInt:
- break;
+ switch (size) {
+ case Maxwell::VertexAttribute::Size::Size_16_16_16_16:
+ return vk::Format::eR16G16B16A16Sint;
+ case Maxwell::VertexAttribute::Size::Size_8:
+ return vk::Format::eR8Sint;
+ case Maxwell::VertexAttribute::Size::Size_8_8:
+ return vk::Format::eR8G8Sint;
+ case Maxwell::VertexAttribute::Size::Size_8_8_8:
+ return vk::Format::eR8G8B8Sint;
+ case Maxwell::VertexAttribute::Size::Size_8_8_8_8:
+ return vk::Format::eR8G8B8A8Sint;
+ case Maxwell::VertexAttribute::Size::Size_32:
+ return vk::Format::eR32Sint;
+ default:
+ break;
+ }
case Maxwell::VertexAttribute::Type::UnsignedInt:
switch (size) {
+ case Maxwell::VertexAttribute::Size::Size_8:
+ return vk::Format::eR8Uint;
+ case Maxwell::VertexAttribute::Size::Size_8_8:
+ return vk::Format::eR8G8Uint;
+ case Maxwell::VertexAttribute::Size::Size_8_8_8:
+ return vk::Format::eR8G8B8Uint;
+ case Maxwell::VertexAttribute::Size::Size_8_8_8_8:
+ return vk::Format::eR8G8B8A8Uint;
case Maxwell::VertexAttribute::Size::Size_32:
return vk::Format::eR32Uint;
default:
break;
}
case Maxwell::VertexAttribute::Type::UnsignedScaled:
+ switch (size) {
+ case Maxwell::VertexAttribute::Size::Size_8_8:
+ return vk::Format::eR8G8Uscaled;
+ default:
+ break;
+ }
+ break;
case Maxwell::VertexAttribute::Type::SignedScaled:
break;
case Maxwell::VertexAttribute::Type::Float:
switch (size) {
- case Maxwell::VertexAttribute::Size::Size_32_32_32_32:
- return vk::Format::eR32G32B32A32Sfloat;
- case Maxwell::VertexAttribute::Size::Size_32_32_32:
- return vk::Format::eR32G32B32Sfloat;
- case Maxwell::VertexAttribute::Size::Size_32_32:
- return vk::Format::eR32G32Sfloat;
case Maxwell::VertexAttribute::Size::Size_32:
return vk::Format::eR32Sfloat;
+ case Maxwell::VertexAttribute::Size::Size_32_32:
+ return vk::Format::eR32G32Sfloat;
+ case Maxwell::VertexAttribute::Size::Size_32_32_32:
+ return vk::Format::eR32G32B32Sfloat;
+ case Maxwell::VertexAttribute::Size::Size_32_32_32_32:
+ return vk::Format::eR32G32B32A32Sfloat;
+ case Maxwell::VertexAttribute::Size::Size_16:
+ return vk::Format::eR16Sfloat;
+ case Maxwell::VertexAttribute::Size::Size_16_16:
+ return vk::Format::eR16G16Sfloat;
+ case Maxwell::VertexAttribute::Size::Size_16_16_16:
+ return vk::Format::eR16G16B16Sfloat;
+ case Maxwell::VertexAttribute::Size::Size_16_16_16_16:
+ return vk::Format::eR16G16B16A16Sfloat;
default:
break;
}
@@ -308,11 +430,14 @@ vk::CompareOp ComparisonOp(Maxwell::ComparisonOp comparison) {
return {};
}
-vk::IndexType IndexFormat(Maxwell::IndexFormat index_format) {
+vk::IndexType IndexFormat(const VKDevice& device, Maxwell::IndexFormat index_format) {
switch (index_format) {
case Maxwell::IndexFormat::UnsignedByte:
- UNIMPLEMENTED_MSG("Vulkan does not support native u8 index format");
- return vk::IndexType::eUint16;
+ if (!device.IsExtIndexTypeUint8Supported()) {
+ UNIMPLEMENTED_MSG("Native uint8 indices are not supported on this device");
+ return vk::IndexType::eUint16;
+ }
+ return vk::IndexType::eUint8EXT;
case Maxwell::IndexFormat::UnsignedShort:
return vk::IndexType::eUint16;
case Maxwell::IndexFormat::UnsignedInt:
diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.h b/src/video_core/renderer_vulkan/maxwell_to_vk.h
index 904a32e01..1534b738b 100644
--- a/src/video_core/renderer_vulkan/maxwell_to_vk.h
+++ b/src/video_core/renderer_vulkan/maxwell_to_vk.h
@@ -4,7 +4,6 @@
#pragma once
-#include <utility>
#include "common/common_types.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/renderer_vulkan/declarations.h"
@@ -23,24 +22,31 @@ vk::Filter Filter(Tegra::Texture::TextureFilter filter);
vk::SamplerMipmapMode MipmapMode(Tegra::Texture::TextureMipmapFilter mipmap_filter);
-vk::SamplerAddressMode WrapMode(Tegra::Texture::WrapMode wrap_mode);
+vk::SamplerAddressMode WrapMode(Tegra::Texture::WrapMode wrap_mode,
+ Tegra::Texture::TextureFilter filter);
vk::CompareOp DepthCompareFunction(Tegra::Texture::DepthCompareFunc depth_compare_func);
} // namespace Sampler
-std::pair<vk::Format, bool> SurfaceFormat(const VKDevice& device, FormatType format_type,
- PixelFormat pixel_format);
+struct FormatInfo {
+ vk::Format format;
+ bool attachable;
+ bool storage;
+};
+
+FormatInfo SurfaceFormat(const VKDevice& device, FormatType format_type, PixelFormat pixel_format);
vk::ShaderStageFlagBits ShaderStage(Tegra::Engines::ShaderType stage);
-vk::PrimitiveTopology PrimitiveTopology(Maxwell::PrimitiveTopology topology);
+vk::PrimitiveTopology PrimitiveTopology(const VKDevice& device,
+ Maxwell::PrimitiveTopology topology);
vk::Format VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttribute::Size size);
vk::CompareOp ComparisonOp(Maxwell::ComparisonOp comparison);
-vk::IndexType IndexFormat(Maxwell::IndexFormat index_format);
+vk::IndexType IndexFormat(const VKDevice& device, Maxwell::IndexFormat index_format);
vk::StencilOp StencilOp(Maxwell::StencilOp stencil_op);
diff --git a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp b/src/video_core/renderer_vulkan/vk_sampler_cache.cpp
index 801826d3d..1ce583f75 100644
--- a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp
+++ b/src/video_core/renderer_vulkan/vk_sampler_cache.cpp
@@ -46,9 +46,10 @@ UniqueSampler VKSamplerCache::CreateSampler(const Tegra::Texture::TSCEntry& tsc)
{}, MaxwellToVK::Sampler::Filter(tsc.mag_filter),
MaxwellToVK::Sampler::Filter(tsc.min_filter),
MaxwellToVK::Sampler::MipmapMode(tsc.mipmap_filter),
- MaxwellToVK::Sampler::WrapMode(tsc.wrap_u), MaxwellToVK::Sampler::WrapMode(tsc.wrap_v),
- MaxwellToVK::Sampler::WrapMode(tsc.wrap_p), tsc.GetLodBias(), has_anisotropy,
- max_anisotropy, tsc.depth_compare_enabled,
+ MaxwellToVK::Sampler::WrapMode(tsc.wrap_u, tsc.mag_filter),
+ MaxwellToVK::Sampler::WrapMode(tsc.wrap_v, tsc.mag_filter),
+ MaxwellToVK::Sampler::WrapMode(tsc.wrap_p, tsc.mag_filter), tsc.GetLodBias(),
+ has_anisotropy, max_anisotropy, tsc.depth_compare_enabled,
MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func), tsc.GetMinLod(),
tsc.GetMaxLod(), vk_border_color.value_or(vk::BorderColor::eFloatTransparentBlack),
unnormalized_coords);
generated by cgit v1.2.3 (git 2.25.1) at 2024-10-11 12:20:31 +0200