diff options
31 files changed, 765 insertions, 78 deletions
diff --git a/.ci/scripts/merge/apply-patches-by-label-private.py b/.ci/scripts/merge/apply-patches-by-label-private.py index fe0acd510..16b45043e 100644 --- a/.ci/scripts/merge/apply-patches-by-label-private.py +++ b/.ci/scripts/merge/apply-patches-by-label-private.py @@ -25,7 +25,7 @@ def check_individual(repo_id, pr_id): def merge_pr(pn, ref): print("Matched PR# %s" % pn) - print(subprocess.check_output(["git", "fetch", "https://%sdev.azure.com/%s/_git/%s" % (user, org, repo), ref, "-f"])) + print(subprocess.check_output(["git", "fetch", "https://%sdev.azure.com/%s/_git/%s" % (user, org, repo), ref, "-f", "--no-recurse-submodules"])) print(subprocess.check_output(["git", "merge", "--squash", 'origin/' + ref.replace('refs/heads/','')])) print(subprocess.check_output(["git", "commit", "-m\"Merge %s PR %s\"" % (tagline, pn)])) diff --git a/.ci/scripts/merge/apply-patches-by-label.py b/.ci/scripts/merge/apply-patches-by-label.py index b2e430ac7..c288a70a1 100644 --- a/.ci/scripts/merge/apply-patches-by-label.py +++ b/.ci/scripts/merge/apply-patches-by-label.py @@ -25,7 +25,7 @@ def do_page(page): if (check_individual(pr["labels"])): pn = pr["number"] print("Matched PR# %s" % pn) - print(subprocess.check_output(["git", "fetch", "https://github.com/yuzu-emu/yuzu.git", "pull/%s/head:pr-%s" % (pn, pn), "-f"])) + print(subprocess.check_output(["git", "fetch", "https://github.com/yuzu-emu/yuzu.git", "pull/%s/head:pr-%s" % (pn, pn), "-f", "--no-recurse-submodules"])) print(subprocess.check_output(["git", "merge", "--squash", "pr-%s" % pn])) print(subprocess.check_output(["git", "commit", "-m\"Merge %s PR %s\"" % (tagline, pn)])) diff --git a/CMakeLists.txt b/CMakeLists.txt index 6044e311a..2c044c1ea 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -57,7 +57,7 @@ function(check_submodules_present) string(REGEX REPLACE "path *= *" "" module ${module}) if (NOT EXISTS "${PROJECT_SOURCE_DIR}/${module}/.git") message(FATAL_ERROR "Git submodule ${module} not found. " - "Please run: git submodule update --init --recursive") + "Please run: \ngit submodule update --init --recursive") endif() endforeach() endfunction() diff --git a/src/common/bit_util.h b/src/common/bit_util.h index 64520ca4e..eef8c1c5a 100644 --- a/src/common/bit_util.h +++ b/src/common/bit_util.h @@ -7,6 +7,7 @@ #include <bit> #include <climits> #include <cstddef> +#include <type_traits> #include "common/common_types.h" @@ -44,4 +45,10 @@ template <typename T> return static_cast<u32>(log2_f + static_cast<u64>((value ^ (1ULL << log2_f)) != 0ULL)); } +template <typename T> +requires std::is_integral_v<T> +[[nodiscard]] T NextPow2(T value) { + return static_cast<T>(1ULL << ((8U * sizeof(T)) - std::countl_zero(value - 1U))); +} + } // namespace Common diff --git a/src/core/hle/service/pm/pm.cpp b/src/core/hle/service/pm/pm.cpp index 88fc5b5cc..277abc17a 100644 --- a/src/core/hle/service/pm/pm.cpp +++ b/src/core/hle/service/pm/pm.cpp @@ -13,7 +13,12 @@ namespace Service::PM { namespace { -constexpr ResultCode ERROR_PROCESS_NOT_FOUND{ErrorModule::PM, 1}; +constexpr ResultCode ResultProcessNotFound{ErrorModule::PM, 1}; +[[maybe_unused]] constexpr ResultCode ResultAlreadyStarted{ErrorModule::PM, 2}; +[[maybe_unused]] constexpr ResultCode ResultNotTerminated{ErrorModule::PM, 3}; +[[maybe_unused]] constexpr ResultCode ResultDebugHookInUse{ErrorModule::PM, 4}; +[[maybe_unused]] constexpr ResultCode ResultApplicationRunning{ErrorModule::PM, 5}; +[[maybe_unused]] constexpr ResultCode ResultInvalidSize{ErrorModule::PM, 6}; constexpr u64 NO_PROCESS_FOUND_PID{0}; @@ -95,18 +100,18 @@ public: private: void GetProcessId(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto title_id = rp.PopRaw<u64>(); + const auto program_id = rp.PopRaw<u64>(); - LOG_DEBUG(Service_PM, "called, title_id={:016X}", title_id); + LOG_DEBUG(Service_PM, "called, program_id={:016X}", program_id); const auto process = - SearchProcessList(kernel.GetProcessList(), [title_id](const auto& proc) { - return proc->GetProgramID() == title_id; + SearchProcessList(kernel.GetProcessList(), [program_id](const auto& proc) { + return proc->GetProgramID() == program_id; }); if (!process.has_value()) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_PROCESS_NOT_FOUND); + rb.Push(ResultProcessNotFound); return; } @@ -128,13 +133,16 @@ public: explicit Info(Core::System& system_, const std::vector<Kernel::KProcess*>& process_list_) : ServiceFramework{system_, "pm:info"}, process_list{process_list_} { static const FunctionInfo functions[] = { - {0, &Info::GetTitleId, "GetTitleId"}, + {0, &Info::GetProgramId, "GetProgramId"}, + {65000, &Info::AtmosphereGetProcessId, "AtmosphereGetProcessId"}, + {65001, nullptr, "AtmosphereHasLaunchedProgram"}, + {65002, nullptr, "AtmosphereGetProcessInfo"}, }; RegisterHandlers(functions); } private: - void GetTitleId(Kernel::HLERequestContext& ctx) { + void GetProgramId(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto process_id = rp.PopRaw<u64>(); @@ -146,7 +154,7 @@ private: if (!process.has_value()) { IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(ERROR_PROCESS_NOT_FOUND); + rb.Push(ResultProcessNotFound); return; } @@ -155,6 +163,27 @@ private: rb.Push((*process)->GetProgramID()); } + void AtmosphereGetProcessId(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto program_id = rp.PopRaw<u64>(); + + LOG_DEBUG(Service_PM, "called, program_id={:016X}", program_id); + + const auto process = SearchProcessList(process_list, [program_id](const auto& proc) { + return proc->GetProgramID() == program_id; + }); + + if (!process.has_value()) { + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(ResultProcessNotFound); + return; + } + + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(ResultSuccess); + rb.Push((*process)->GetProcessID()); + } + const std::vector<Kernel::KProcess*>& process_list; }; diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h index 05e5c94f3..c89a5d693 100644 --- a/src/video_core/gpu.h +++ b/src/video_core/gpu.h @@ -83,6 +83,7 @@ enum class DepthFormat : u32 { S8_UINT_Z24_UNORM = 0x14, D24X8_UNORM = 0x15, D24S8_UNORM = 0x16, + S8_UINT = 0x17, D24C8_UNORM = 0x18, D32_FLOAT_S8X24_UINT = 0x19, }; diff --git a/src/video_core/host_shaders/CMakeLists.txt b/src/video_core/host_shaders/CMakeLists.txt index d779a967a..1c91999d7 100644 --- a/src/video_core/host_shaders/CMakeLists.txt +++ b/src/video_core/host_shaders/CMakeLists.txt @@ -10,8 +10,14 @@ set(SHADER_FILES astc_decoder.comp block_linear_unswizzle_2d.comp block_linear_unswizzle_3d.comp + convert_abgr8_to_d24s8.frag + convert_b10g11r11_to_d24s8.frag + convert_d24s8_to_abgr8.frag + convert_d24s8_to_b10g11r11.frag + convert_d24s8_to_r16g16.frag convert_depth_to_float.frag convert_float_to_depth.frag + convert_r16g16_to_d24s8.frag full_screen_triangle.vert fxaa.frag fxaa.vert diff --git a/src/video_core/host_shaders/convert_abgr8_to_d24s8.frag b/src/video_core/host_shaders/convert_abgr8_to_d24s8.frag new file mode 100644 index 000000000..4e4ab6a26 --- /dev/null +++ b/src/video_core/host_shaders/convert_abgr8_to_d24s8.frag @@ -0,0 +1,17 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 +#extension GL_ARB_shader_stencil_export : require + +layout(binding = 0) uniform sampler2D color_texture; + +void main() { + ivec2 coord = ivec2(gl_FragCoord.xy); + uvec4 color = uvec4(texelFetch(color_texture, coord, 0).rgba * (exp2(8) - 1.0f)); + uint depth_unorm = (color.r << 16) | (color.g << 8) | color.b; + + gl_FragDepth = float(depth_unorm) / (exp2(24.0) - 1.0f); + gl_FragStencilRefARB = int(color.a); +} diff --git a/src/video_core/host_shaders/convert_b10g11r11_to_d24s8.frag b/src/video_core/host_shaders/convert_b10g11r11_to_d24s8.frag new file mode 100644 index 000000000..2999a84cf --- /dev/null +++ b/src/video_core/host_shaders/convert_b10g11r11_to_d24s8.frag @@ -0,0 +1,19 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 +#extension GL_ARB_shader_stencil_export : require + +layout(binding = 0) uniform sampler2D color_texture; + +void main() { + ivec2 coord = ivec2(gl_FragCoord.xy); + vec4 color = texelFetch(color_texture, coord, 0).rgba; + uint depth_stencil_unorm = (uint(color.b * (exp2(10) - 1.0f)) << 22) + | (uint(color.g * (exp2(11) - 1.0f)) << 11) + | (uint(color.r * (exp2(11) - 1.0f))); + + gl_FragDepth = float(depth_stencil_unorm >> 8) / (exp2(24.0) - 1.0f); + gl_FragStencilRefARB = int(depth_stencil_unorm & 0x00FF); +} diff --git a/src/video_core/host_shaders/convert_d24s8_to_abgr8.frag b/src/video_core/host_shaders/convert_d24s8_to_abgr8.frag new file mode 100644 index 000000000..ff3bf8209 --- /dev/null +++ b/src/video_core/host_shaders/convert_d24s8_to_abgr8.frag @@ -0,0 +1,21 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 + +layout(binding = 0) uniform sampler2D depth_tex; +layout(binding = 1) uniform isampler2D stencil_tex; + +layout(location = 0) out vec4 color; + +void main() { + ivec2 coord = ivec2(gl_FragCoord.xy); + uint depth = uint(textureLod(depth_tex, coord, 0).r * (exp2(24.0) - 1.0f)); + uint stencil = uint(textureLod(stencil_tex, coord, 0).r); + + color.r = float(depth >> 16) / (exp2(8) - 1.0); + color.g = float((depth >> 8) & 0x00FF) / (exp2(8) - 1.0); + color.b = float(depth & 0x00FF) / (exp2(8) - 1.0); + color.a = float(stencil) / (exp2(8) - 1.0); +} diff --git a/src/video_core/host_shaders/convert_d24s8_to_b10g11r11.frag b/src/video_core/host_shaders/convert_d24s8_to_b10g11r11.frag new file mode 100644 index 000000000..c743d3a13 --- /dev/null +++ b/src/video_core/host_shaders/convert_d24s8_to_b10g11r11.frag @@ -0,0 +1,21 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 + +layout(binding = 0) uniform sampler2D depth_tex; +layout(binding = 1) uniform isampler2D stencil_tex; + +layout(location = 0) out vec4 color; + +void main() { + ivec2 coord = ivec2(gl_FragCoord.xy); + uint depth = uint(textureLod(depth_tex, coord, 0).r * (exp2(24.0) - 1.0f)); + uint stencil = uint(textureLod(stencil_tex, coord, 0).r); + + color.b = float(depth >> 22) / (exp2(10) - 1.0); + color.g = float((depth >> 11) & 0x00FF) / (exp2(11) - 1.0); + color.r = float(depth & 0x00FF) / (exp2(11) - 1.0); + color.a = 1.0f; +} diff --git a/src/video_core/host_shaders/convert_d24s8_to_r16g16.frag b/src/video_core/host_shaders/convert_d24s8_to_r16g16.frag new file mode 100644 index 000000000..2a9443d3d --- /dev/null +++ b/src/video_core/host_shaders/convert_d24s8_to_r16g16.frag @@ -0,0 +1,21 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 + +layout(binding = 0) uniform sampler2D depth_tex; +layout(binding = 1) uniform isampler2D stencil_tex; + +layout(location = 0) out vec4 color; + +void main() { + ivec2 coord = ivec2(gl_FragCoord.xy); + uint depth = uint(textureLod(depth_tex, coord, 0).r * (exp2(24.0) - 1.0f)); + uint stencil = uint(textureLod(stencil_tex, coord, 0).r); + + color.r = float(depth >> 16) / (exp2(16) - 1.0); + color.g = float((depth >> 16) & 0x00FF) / (exp2(16) - 1.0); + color.b = 0.0f; + color.a = 1.0f; +} diff --git a/src/video_core/host_shaders/convert_r16g16_to_d24s8.frag b/src/video_core/host_shaders/convert_r16g16_to_d24s8.frag new file mode 100644 index 000000000..3df70575e --- /dev/null +++ b/src/video_core/host_shaders/convert_r16g16_to_d24s8.frag @@ -0,0 +1,18 @@ +// Copyright 2021 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 +#extension GL_ARB_shader_stencil_export : require + +layout(binding = 0) uniform sampler2D color_texture; + +void main() { + ivec2 coord = ivec2(gl_FragCoord.xy); + vec4 color = texelFetch(color_texture, coord, 0).rgba; + uint depth_stencil_unorm = (uint(color.r * (exp2(16) - 1.0f)) << 16) + | (uint(color.g * (exp2(16) - 1.0f)) << 16); + + gl_FragDepth = float(depth_stencil_unorm >> 8) / (exp2(24.0) - 1.0f); + gl_FragStencilRefARB = int(depth_stencil_unorm & 0x00FF); +} diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index 2f7d98d8b..14e6522f2 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -9,6 +9,7 @@ #include <glad/glad.h> +#include "common/bit_util.h" #include "common/literals.h" #include "common/settings.h" #include "video_core/renderer_opengl/gl_device.h" @@ -148,6 +149,8 @@ GLenum AttachmentType(PixelFormat format) { switch (const SurfaceType type = VideoCore::Surface::GetFormatType(format); type) { case SurfaceType::Depth: return GL_DEPTH_ATTACHMENT; + case SurfaceType::Stencil: + return GL_STENCIL_ATTACHMENT; case SurfaceType::DepthStencil: return GL_DEPTH_STENCIL_ATTACHMENT; default: @@ -317,13 +320,12 @@ void AttachTexture(GLuint fbo, GLenum attachment, const ImageView* image_view) { } } -OGLTexture MakeImage(const VideoCommon::ImageInfo& info, GLenum gl_internal_format) { +OGLTexture MakeImage(const VideoCommon::ImageInfo& info, GLenum gl_internal_format, + GLsizei gl_num_levels) { const GLenum target = ImageTarget(info); const GLsizei width = info.size.width; const GLsizei height = info.size.height; const GLsizei depth = info.size.depth; - const int max_host_mip_levels = std::bit_width(info.size.width); - const GLsizei num_levels = std::min(info.resources.levels, max_host_mip_levels); const GLsizei num_layers = info.resources.layers; const GLsizei num_samples = info.num_samples; @@ -335,10 +337,10 @@ OGLTexture MakeImage(const VideoCommon::ImageInfo& info, GLenum gl_internal_form } switch (target) { case GL_TEXTURE_1D_ARRAY: - glTextureStorage2D(handle, num_levels, gl_internal_format, width, num_layers); + glTextureStorage2D(handle, gl_num_levels, gl_internal_format, width, num_layers); break; case GL_TEXTURE_2D_ARRAY: - glTextureStorage3D(handle, num_levels, gl_internal_format, width, height, num_layers); + glTextureStorage3D(handle, gl_num_levels, gl_internal_format, width, height, num_layers); break; case GL_TEXTURE_2D_MULTISAMPLE_ARRAY: { // TODO: Where should 'fixedsamplelocations' come from? @@ -348,10 +350,10 @@ OGLTexture MakeImage(const VideoCommon::ImageInfo& info, GLenum gl_internal_form break; } case GL_TEXTURE_RECTANGLE: - glTextureStorage2D(handle, num_levels, gl_internal_format, width, height); + glTextureStorage2D(handle, gl_num_levels, gl_internal_format, width, height); break; case GL_TEXTURE_3D: - glTextureStorage3D(handle, num_levels, gl_internal_format, width, height, depth); + glTextureStorage3D(handle, gl_num_levels, gl_internal_format, width, height, depth); break; case GL_TEXTURE_BUFFER: UNREACHABLE(); @@ -396,6 +398,7 @@ OGLTexture MakeImage(const VideoCommon::ImageInfo& info, GLenum gl_internal_form UNREACHABLE_MSG("Invalid image format={}", format); return GL_R32UI; } + } // Anonymous namespace ImageBufferMap::~ImageBufferMap() { @@ -522,6 +525,12 @@ void TextureCacheRuntime::CopyImage(Image& dst_image, Image& src_image, } } +void TextureCacheRuntime::ReinterpretImage(Image& dst, Image& src, + std::span<const VideoCommon::ImageCopy> copies) { + LOG_DEBUG(Render_OpenGL, "Converting {} to {}", src.info.format, dst.info.format); + format_conversion_pass.ConvertImage(dst, src, copies); +} + bool TextureCacheRuntime::CanImageBeCopied(const Image& dst, const Image& src) { if (dst.info.type == ImageType::e3D && dst.info.format == PixelFormat::BC4_UNORM) { return false; @@ -538,7 +547,7 @@ void TextureCacheRuntime::EmulateCopyImage(Image& dst, Image& src, ASSERT(src.info.type == ImageType::e3D); util_shaders.CopyBC4(dst, src, copies); } else if (IsPixelFormatBGR(dst.info.format) || IsPixelFormatBGR(src.info.format)) { - bgr_copy_pass.CopyBGR(dst, src, copies); + format_conversion_pass.ConvertImage(dst, src, copies); } else { UNREACHABLE(); } @@ -686,7 +695,9 @@ Image::Image(TextureCacheRuntime& runtime_, const VideoCommon::ImageInfo& info_, gl_format = tuple.format; gl_type = tuple.type; } - texture = MakeImage(info, gl_internal_format); + const int max_host_mip_levels = std::bit_width(info.size.width); + gl_num_levels = std::min(info.resources.levels, max_host_mip_levels); + texture = MakeImage(info, gl_internal_format, gl_num_levels); current_texture = texture.handle; if (runtime->device.HasDebuggingToolAttached()) { const std::string name = VideoCommon::Name(*this); @@ -714,6 +725,9 @@ void Image::UploadMemory(const ImageBufferMap& map, u32 current_image_height = std::numeric_limits<u32>::max(); for (const VideoCommon::BufferImageCopy& copy : copies) { + if (copy.image_subresource.base_level >= gl_num_levels) { + continue; + } if (current_row_length != copy.buffer_row_length) { current_row_length = copy.buffer_row_length; glPixelStorei(GL_UNPACK_ROW_LENGTH, current_row_length); @@ -743,6 +757,9 @@ void Image::DownloadMemory(ImageBufferMap& map, u32 current_image_height = std::numeric_limits<u32>::max(); for (const VideoCommon::BufferImageCopy& copy : copies) { + if (copy.image_subresource.base_level >= gl_num_levels) { + continue; + } if (current_row_length != copy.buffer_row_length) { current_row_length = copy.buffer_row_length; glPixelStorei(GL_PACK_ROW_LENGTH, current_row_length); @@ -782,7 +799,7 @@ GLuint Image::StorageHandle() noexcept { } store_view.Create(); glTextureView(store_view.handle, ImageTarget(info), current_texture, GL_RGBA8, 0, - info.resources.levels, 0, info.resources.layers); + gl_num_levels, 0, info.resources.layers); return store_view.handle; default: return current_texture; @@ -897,6 +914,8 @@ void Image::Scale(bool up_scale) { return GL_COLOR_ATTACHMENT0; case SurfaceType::Depth: return GL_DEPTH_ATTACHMENT; + case SurfaceType::Stencil: + return GL_STENCIL_ATTACHMENT; case SurfaceType::DepthStencil: return GL_DEPTH_STENCIL_ATTACHMENT; default: @@ -910,8 +929,10 @@ void Image::Scale(bool up_scale) { return GL_COLOR_BUFFER_BIT; case SurfaceType::Depth: return GL_DEPTH_BUFFER_BIT; + case SurfaceType::Stencil: + return GL_STENCIL_BUFFER_BIT; case SurfaceType::DepthStencil: - return GL_STENCIL_BUFFER_BIT | GL_DEPTH_BUFFER_BIT; + return GL_DEPTH_BUFFER_BIT | GL_STENCIL_BUFFER_BIT; default: UNREACHABLE(); return GL_COLOR_BUFFER_BIT; @@ -923,8 +944,10 @@ void Image::Scale(bool up_scale) { return 0; case SurfaceType::Depth: return 1; - case SurfaceType::DepthStencil: + case SurfaceType::Stencil: return 2; + case SurfaceType::DepthStencil: + return 3; default: UNREACHABLE(); return 0; @@ -946,7 +969,7 @@ void Image::Scale(bool up_scale) { auto dst_info = info; dst_info.size.width = scaled_width; dst_info.size.height = scaled_height; - upscaled_backup = MakeImage(dst_info, gl_internal_format); + upscaled_backup = MakeImage(dst_info, gl_internal_format, gl_num_levels); } const u32 src_width = up_scale ? original_width : scaled_width; const u32 src_height = up_scale ? original_height : scaled_height; @@ -1254,10 +1277,20 @@ Framebuffer::Framebuffer(TextureCacheRuntime& runtime, std::span<ImageView*, NUM } if (const ImageView* const image_view = depth_buffer; image_view) { - if (GetFormatType(image_view->format) == SurfaceType::DepthStencil) { + switch (GetFormatType(image_view->format)) { + case SurfaceType::Depth: + buffer_bits |= GL_DEPTH_BUFFER_BIT; + break; + case SurfaceType::Stencil: + buffer_bits |= GL_STENCIL_BUFFER_BIT; + break; + case SurfaceType::DepthStencil: buffer_bits |= GL_DEPTH_BUFFER_BIT | GL_STENCIL_BUFFER_BIT; - } else { + break; + default: + UNREACHABLE(); buffer_bits |= GL_DEPTH_BUFFER_BIT; + break; } const GLenum attachment = AttachmentType(image_view->format); AttachTexture(handle, attachment, image_view); @@ -1286,35 +1319,37 @@ Framebuffer::Framebuffer(TextureCacheRuntime& runtime, std::span<ImageView*, NUM Framebuffer::~Framebuffer() = default; -void BGRCopyPass::CopyBGR(Image& dst_image, Image& src_image, - std::span<const VideoCommon::ImageCopy> copies) { - static constexpr VideoCommon::Offset3D zero_offset{0, 0, 0}; +void FormatConversionPass::ConvertImage(Image& dst_image, Image& src_image, + std::span<const VideoCommon::ImageCopy> copies) { + const GLenum dst_target = ImageTarget(dst_image.info); + const GLenum src_target = ImageTarget(src_image.info); const u32 img_bpp = BytesPerBlock(src_image.info.format); for (const ImageCopy& copy : copies) { - ASSERT(copy.src_offset == zero_offset); - ASSERT(copy.dst_offset == zero_offset); - const u32 num_src_layers = static_cast<u32>(copy.src_subresource.num_layers); - const u32 copy_size = copy.extent.width * copy.extent.height * num_src_layers * img_bpp; - if (bgr_pbo_size < copy_size) { - bgr_pbo.Create(); - bgr_pbo_size = copy_size; - glNamedBufferData(bgr_pbo.handle, bgr_pbo_size, nullptr, GL_STREAM_COPY); + const auto src_origin = MakeCopyOrigin(copy.src_offset, copy.src_subresource, src_target); + const auto dst_origin = MakeCopyOrigin(copy.dst_offset, copy.dst_subresource, dst_target); + const auto region = MakeCopyRegion(copy.extent, copy.dst_subresource, dst_target); + const u32 copy_size = region.width * region.height * region.depth * img_bpp; + if (pbo_size < copy_size) { + intermediate_pbo.Create(); + pbo_size = Common::NextPow2(copy_size); + glNamedBufferData(intermediate_pbo.handle, pbo_size, nullptr, GL_STREAM_COPY); } // Copy from source to PBO glPixelStorei(GL_PACK_ALIGNMENT, 1); glPixelStorei(GL_PACK_ROW_LENGTH, copy.extent.width); - glBindBuffer(GL_PIXEL_PACK_BUFFER, bgr_pbo.handle); - glGetTextureSubImage(src_image.Handle(), 0, 0, 0, 0, copy.extent.width, copy.extent.height, - num_src_layers, src_image.GlFormat(), src_image.GlType(), - static_cast<GLsizei>(bgr_pbo_size), nullptr); + glBindBuffer(GL_PIXEL_PACK_BUFFER, intermediate_pbo.handle); + glGetTextureSubImage(src_image.Handle(), src_origin.level, src_origin.x, src_origin.y, + src_origin.z, region.width, region.height, region.depth, + src_image.GlFormat(), src_image.GlType(), + static_cast<GLsizei>(pbo_size), nullptr); // Copy from PBO to destination in desired GL format glPixelStorei(GL_UNPACK_ALIGNMENT, 1); glPixelStorei(GL_UNPACK_ROW_LENGTH, copy.extent.width); - glBindBuffer(GL_PIXEL_UNPACK_BUFFER, bgr_pbo.handle); - glTextureSubImage3D(dst_image.Handle(), 0, 0, 0, 0, copy.extent.width, copy.extent.height, - copy.dst_subresource.num_layers, dst_image.GlFormat(), - dst_image.GlType(), nullptr); + glBindBuffer(GL_PIXEL_UNPACK_BUFFER, intermediate_pbo.handle); + glTextureSubImage3D(dst_image.Handle(), dst_origin.level, dst_origin.x, dst_origin.y, + dst_origin.z, region.width, region.height, region.depth, + dst_image.GlFormat(), dst_image.GlType(), nullptr); } } diff --git a/src/video_core/renderer_opengl/gl_texture_cache.h b/src/video_core/renderer_opengl/gl_texture_cache.h index 1bb762568..37d5e6a6b 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.h +++ b/src/video_core/renderer_opengl/gl_texture_cache.h @@ -52,17 +52,17 @@ struct FormatProperties { bool is_compressed; }; -class BGRCopyPass { +class FormatConversionPass { public: - BGRCopyPass() = default; - ~BGRCopyPass() = default; + FormatConversionPass() = default; + ~FormatConversionPass() = default; - void CopyBGR(Image& dst_image, Image& src_image, - std::span<const VideoCommon::ImageCopy> copies); + void ConvertImage(Image& dst_image, Image& src_image, + std::span<const VideoCommon::ImageCopy> copies); private: - OGLBuffer bgr_pbo; - size_t bgr_pbo_size{}; + OGLBuffer intermediate_pbo; + size_t pbo_size{}; }; class TextureCacheRuntime { @@ -84,8 +84,14 @@ public: u64 GetDeviceLocalMemory() const; + bool ShouldReinterpret([[maybe_unused]] Image& dst, [[maybe_unused]] Image& src) { + return true; + } + void CopyImage(Image& dst, Image& src, std::span<const VideoCommon::ImageCopy> copies); + void ReinterpretImage(Image& dst, Image& src, std::span<const VideoCommon::ImageCopy> copies); + void ConvertImage(Framebuffer* dst, ImageView& dst_view, ImageView& src_view, bool rescaled) { UNIMPLEMENTED(); } @@ -144,7 +150,7 @@ private: const Device& device; StateTracker& state_tracker; UtilShaders util_shaders; - BGRCopyPass bgr_copy_pass; + FormatConversionPass format_conversion_pass; std::array<std::unordered_map<GLenum, FormatProperties>, 3> format_properties; bool has_broken_texture_view_formats = false; @@ -162,8 +168,8 @@ private: std::array<GLuint, Shader::NUM_TEXTURE_TYPES> null_image_views{}; - std::array<OGLFramebuffer, 3> rescale_draw_fbos; - std::array<OGLFramebuffer, 3> rescale_read_fbos; + std::array<OGLFramebuffer, 4> rescale_draw_fbos; + std::array<OGLFramebuffer, 4> rescale_read_fbos; const Settings::ResolutionScalingInfo& resolution; }; @@ -219,6 +225,7 @@ private: GLenum gl_internal_format = GL_NONE; GLenum gl_format = GL_NONE; GLenum gl_type = GL_NONE; + GLsizei gl_num_levels{}; TextureCacheRuntime* runtime{}; GLuint current_texture{}; }; diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index 39158aa3e..daba42ed9 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -108,6 +108,7 @@ constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> FORMAT_TAB {GL_RGB9_E5, GL_RGB, GL_UNSIGNED_INT_5_9_9_9_REV}, // E5B9G9R9_FLOAT {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT}, // D32_FLOAT {GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT}, // D16_UNORM + {GL_STENCIL_INDEX8, GL_STENCIL, GL_UNSIGNED_BYTE}, // S8_UINT {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // D24_UNORM_S8_UINT {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // S8_UINT_D24_UNORM {GL_DEPTH32F_STENCIL8, GL_DEPTH_STENCIL, diff --git a/src/video_core/renderer_vulkan/blit_image.cpp b/src/video_core/renderer_vulkan/blit_image.cpp index b3884a4f5..28b631f73 100644 --- a/src/video_core/renderer_vulkan/blit_image.cpp +++ b/src/video_core/renderer_vulkan/blit_image.cpp @@ -4,8 +4,14 @@ #include <algorithm> +#include "video_core/host_shaders/convert_abgr8_to_d24s8_frag_spv.h" +#include "video_core/host_shaders/convert_b10g11r11_to_d24s8_frag_spv.h" +#include "video_core/host_shaders/convert_d24s8_to_abgr8_frag_spv.h" +#include "video_core/host_shaders/convert_d24s8_to_b10g11r11_frag_spv.h" +#include "video_core/host_shaders/convert_d24s8_to_r16g16_frag_spv.h" #include "video_core/host_shaders/convert_depth_to_float_frag_spv.h" #include "video_core/host_shaders/convert_float_to_depth_frag_spv.h" +#include "video_core/host_shaders/convert_r16g16_to_d24s8_frag_spv.h" #include "video_core/host_shaders/full_screen_triangle_vert_spv.h" #include "video_core/host_shaders/vulkan_blit_color_float_frag_spv.h" #include "video_core/host_shaders/vulkan_blit_depth_stencil_frag_spv.h" @@ -354,6 +360,12 @@ BlitImageHelper::BlitImageHelper(const Device& device_, VKScheduler& scheduler_, blit_color_to_color_frag(BuildShader(device, VULKAN_BLIT_COLOR_FLOAT_FRAG_SPV)), convert_depth_to_float_frag(BuildShader(device, CONVERT_DEPTH_TO_FLOAT_FRAG_SPV)), convert_float_to_depth_frag(BuildShader(device, CONVERT_FLOAT_TO_DEPTH_FRAG_SPV)), + convert_abgr8_to_d24s8_frag(BuildShader(device, CONVERT_ABGR8_TO_D24S8_FRAG_SPV)), + convert_b10g11r11_to_d24s8_frag(BuildShader(device, CONVERT_B10G11R11_TO_D24S8_FRAG_SPV)), + convert_r16g16_to_d24s8_frag(BuildShader(device, CONVERT_R16G16_TO_D24S8_FRAG_SPV)), + convert_d24s8_to_abgr8_frag(BuildShader(device, CONVERT_D24S8_TO_ABGR8_FRAG_SPV)), + convert_d24s8_to_b10g11r11_frag(BuildShader(device, CONVERT_D24S8_TO_B10G11R11_FRAG_SPV)), + convert_d24s8_to_r16g16_frag(BuildShader(device, CONVERT_D24S8_TO_R16G16_FRAG_SPV)), linear_sampler(device.GetLogical().CreateSampler(SAMPLER_CREATE_INFO<VK_FILTER_LINEAR>)), nearest_sampler(device.GetLogical().CreateSampler(SAMPLER_CREATE_INFO<VK_FILTER_NEAREST>)) { if (device.IsExtShaderStencilExportSupported()) { @@ -448,6 +460,59 @@ void BlitImageHelper::ConvertR16ToD16(const Framebuffer* dst_framebuffer, Convert(*convert_r16_to_d16_pipeline, dst_framebuffer, src_image_view, up_scale, down_shift); } +void BlitImageHelper::ConvertABGR8ToD24S8(const Framebuffer* dst_framebuffer, + const ImageView& src_image_view, u32 up_scale, + u32 down_shift) { + ConvertPipelineDepthTargetEx(convert_abgr8_to_d24s8_pipeline, dst_framebuffer->RenderPass(), + convert_abgr8_to_d24s8_frag, true); + Convert(*convert_abgr8_to_d24s8_pipeline, dst_framebuffer, src_image_view, up_scale, + down_shift); +} + +void BlitImageHelper::ConvertB10G11R11ToD24S8(const Framebuffer* dst_framebuffer, + const ImageView& src_image_view, u32 up_scale, + u32 down_shift) { + ConvertPipelineDepthTargetEx(convert_b10g11r11_to_d24s8_pipeline, dst_framebuffer->RenderPass(), + convert_b10g11r11_to_d24s8_frag, true); + Convert(*convert_b10g11r11_to_d24s8_pipeline, dst_framebuffer, src_image_view, up_scale, + down_shift); +} + +void BlitImageHelper::ConvertR16G16ToD24S8(const Framebuffer* dst_framebuffer, + const ImageView& src_image_view, u32 up_scale, + u32 down_shift) { + ConvertPipelineDepthTargetEx(convert_r16g16_to_d24s8_pipeline, dst_framebuffer->RenderPass(), + convert_r16g16_to_d24s8_frag, true); + Convert(*convert_r16g16_to_d24s8_pipeline, dst_framebuffer, src_image_view, up_scale, + down_shift); +} + +void BlitImageHelper::ConvertD24S8ToABGR8(const Framebuffer* dst_framebuffer, + ImageView& src_image_view, u32 up_scale, u32 down_shift) { + ConvertPipelineColorTargetEx(convert_d24s8_to_abgr8_pipeline, dst_framebuffer->RenderPass(), + convert_d24s8_to_abgr8_frag, false); + ConvertDepthStencil(*convert_d24s8_to_abgr8_pipeline, dst_framebuffer, src_image_view, up_scale, + down_shift); +} + +void BlitImageHelper::ConvertD24S8ToB10G11R11(const Framebuffer* dst_framebuffer, + ImageView& src_image_view, u32 up_scale, + u32 down_shift) { + ConvertPipelineColorTargetEx(convert_d24s8_to_b10g11r11_pipeline, dst_framebuffer->RenderPass(), + convert_d24s8_to_b10g11r11_frag, false); + ConvertDepthStencil(*convert_d24s8_to_b10g11r11_pipeline, dst_framebuffer, src_image_view, + up_scale, down_shift); +} + +void BlitImageHelper::ConvertD24S8ToR16G16(const Framebuffer* dst_framebuffer, + ImageView& src_image_view, u32 up_scale, + u32 down_shift) { + ConvertPipelineColorTargetEx(convert_d24s8_to_r16g16_pipeline, dst_framebuffer->RenderPass(), + convert_d24s8_to_r16g16_frag, false); + ConvertDepthStencil(*convert_d24s8_to_r16g16_pipeline, dst_framebuffer, src_image_view, + up_scale, down_shift); +} + void BlitImageHelper::Convert(VkPipeline pipeline, const Framebuffer* dst_framebuffer, const ImageView& src_image_view, u32 up_scale, u32 down_shift) { const VkPipelineLayout layout = *one_texture_pipeline_layout; @@ -495,6 +560,54 @@ void BlitImageHelper::Convert(VkPipeline pipeline, const Framebuffer* dst_frameb scheduler.InvalidateState(); } +void BlitImageHelper::ConvertDepthStencil(VkPipeline pipeline, const Framebuffer* dst_framebuffer, + ImageView& src_image_view, u32 up_scale, u32 down_shift) { + const VkPipelineLayout layout = *two_textures_pipeline_layout; + const VkImageView src_depth_view = src_image_view.DepthView(); + const VkImageView src_stencil_view = src_image_view.StencilView(); + const VkSampler sampler = *nearest_sampler; + const VkExtent2D extent{ + .width = std::max((src_image_view.size.width * up_scale) >> down_shift, 1U), + .height = std::max((src_image_view.size.height * up_scale) >> down_shift, 1U), + }; + scheduler.RequestRenderpass(dst_framebuffer); + scheduler.Record([pipeline, layout, sampler, src_depth_view, src_stencil_view, extent, up_scale, + down_shift, this](vk::CommandBuffer cmdbuf) { + const VkOffset2D offset{ + .x = 0, + .y = 0, + }; + const VkViewport viewport{ + .x = 0.0f, + .y = 0.0f, + .width = static_cast<float>(extent.width), + .height = static_cast<float>(extent.height), + .minDepth = 0.0f, + .maxDepth = 0.0f, + }; + const VkRect2D scissor{ + .offset = offset, + .extent = extent, + }; + const PushConstants push_constants{ + .tex_scale = {viewport.width, viewport.height}, + .tex_offset = {0.0f, 0.0f}, + }; + const VkDescriptorSet descriptor_set = two_textures_descriptor_allocator.Commit(); + UpdateTwoTexturesDescriptorSet(device, descriptor_set, sampler, src_depth_view, + src_stencil_view); + // TODO: Barriers + cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline); + cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS, layout, 0, descriptor_set, + nullptr); + cmdbuf.SetViewport(0, viewport); + cmdbuf.SetScissor(0, scissor); + cmdbuf.PushConstants(layout, VK_SHADER_STAGE_VERTEX_BIT, push_constants); + cmdbuf.Draw(3, 1, 0, 0); + }); + scheduler.InvalidateState(); +} + VkPipeline BlitImageHelper::FindOrEmplaceColorPipeline(const BlitImagePipelineKey& key) { const auto it = std::ranges::find(blit_color_keys, key); if (it != blit_color_keys.end()) { @@ -636,4 +749,62 @@ void BlitImageHelper::ConvertColorToDepthPipeline(vk::Pipeline& pipeline, VkRend }); } +void BlitImageHelper::ConvertPipelineColorTargetEx(vk::Pipeline& pipeline, VkRenderPass renderpass, + vk::ShaderModule& module, bool single_texture) { + if (pipeline) { + return; + } + const std::array stages = MakeStages(*full_screen_vert, *module); + pipeline = device.GetLogical().CreateGraphicsPipeline({ + .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stageCount = static_cast<u32>(stages.size()), + .pStages = stages.data(), + .pVertexInputState = &PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO, + .pInputAssemblyState = &PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, + .pTessellationState = nullptr, + .pViewportState = &PIPELINE_VIEWPORT_STATE_CREATE_INFO, + .pRasterizationState = &PIPELINE_RASTERIZATION_STATE_CREATE_INFO, + .pMultisampleState = &PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, + .pDepthStencilState = nullptr, + .pColorBlendState = &PIPELINE_COLOR_BLEND_STATE_GENERIC_CREATE_INFO, + .pDynamicState = &PIPELINE_DYNAMIC_STATE_CREATE_INFO, + .layout = single_texture ? *one_texture_pipeline_layout : *two_textures_pipeline_layout, + .renderPass = renderpass, + .subpass = 0, + .basePipelineHandle = VK_NULL_HANDLE, + .basePipelineIndex = 0, + }); +} + +void BlitImageHelper::ConvertPipelineDepthTargetEx(vk::Pipeline& pipeline, VkRenderPass renderpass, + vk::ShaderModule& module, bool single_texture) { + if (pipeline) { + return; + } + const std::array stages = MakeStages(*full_screen_vert, *module); + pipeline = device.GetLogical().CreateGraphicsPipeline({ + .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stageCount = static_cast<u32>(stages.size()), + .pStages = stages.data(), + .pVertexInputState = &PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO, + .pInputAssemblyState = &PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, + .pTessellationState = nullptr, + .pViewportState = &PIPELINE_VIEWPORT_STATE_CREATE_INFO, + .pRasterizationState = &PIPELINE_RASTERIZATION_STATE_CREATE_INFO, + .pMultisampleState = &PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, + .pDepthStencilState = &PIPELINE_DEPTH_STENCIL_STATE_CREATE_INFO, + .pColorBlendState = &PIPELINE_COLOR_BLEND_STATE_EMPTY_CREATE_INFO, + .pDynamicState = &PIPELINE_DYNAMIC_STATE_CREATE_INFO, + .layout = single_texture ? *one_texture_pipeline_layout : *two_textures_pipeline_layout, + .renderPass = renderpass, + .subpass = 0, + .basePipelineHandle = VK_NULL_HANDLE, + .basePipelineIndex = 0, + }); +} + } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/blit_image.h b/src/video_core/renderer_vulkan/blit_image.h index d77f76678..cec095341 100644 --- a/src/video_core/renderer_vulkan/blit_image.h +++ b/src/video_core/renderer_vulkan/blit_image.h @@ -56,10 +56,31 @@ public: void ConvertR16ToD16(const Framebuffer* dst_framebuffer, const ImageView& src_image_view, u32 up_scale, u32 down_shift); + void ConvertABGR8ToD24S8(const Framebuffer* dst_framebuffer, const ImageView& src_image_view, + u32 up_scale, u32 down_shift); + + void ConvertB10G11R11ToD24S8(const Framebuffer* dst_framebuffer, + const ImageView& src_image_view, u32 up_scale, u32 down_shift); + + void ConvertR16G16ToD24S8(const Framebuffer* dst_framebuffer, const ImageView& src_image_view, + u32 up_scale, u32 down_shift); + + void ConvertD24S8ToABGR8(const Framebuffer* dst_framebuffer, ImageView& src_image_view, + u32 up_scale, u32 down_shift); + + void ConvertD24S8ToB10G11R11(const Framebuffer* dst_framebuffer, ImageView& src_image_view, + u32 up_scale, u32 down_shift); + + void ConvertD24S8ToR16G16(const Framebuffer* dst_framebuffer, ImageView& src_image_view, + u32 up_scale, u32 down_shift); + private: void Convert(VkPipeline pipeline, const Framebuffer* dst_framebuffer, const ImageView& src_image_view, u32 up_scale, u32 down_shift); + void ConvertDepthStencil(VkPipeline pipeline, const Framebuffer* dst_framebuffer, + ImageView& src_image_view, u32 up_scale, u32 down_shift); + [[nodiscard]] VkPipeline FindOrEmplaceColorPipeline(const BlitImagePipelineKey& key); [[nodiscard]] VkPipeline FindOrEmplaceDepthStencilPipeline(const BlitImagePipelineKey& key); @@ -68,6 +89,12 @@ private: void ConvertColorToDepthPipeline(vk::Pipeline& pipeline, VkRenderPass renderpass); + void ConvertPipelineColorTargetEx(vk::Pipeline& pipeline, VkRenderPass renderpass, + vk::ShaderModule& module, bool single_texture); + + void ConvertPipelineDepthTargetEx(vk::Pipeline& pipeline, VkRenderPass renderpass, + vk::ShaderModule& module, bool single_texture); + const Device& device; VKScheduler& scheduler; StateTracker& state_tracker; @@ -83,6 +110,12 @@ private: vk::ShaderModule blit_depth_stencil_frag; vk::ShaderModule convert_depth_to_float_frag; vk::ShaderModule convert_float_to_depth_frag; + vk::ShaderModule convert_abgr8_to_d24s8_frag; + vk::ShaderModule convert_b10g11r11_to_d24s8_frag; + vk::ShaderModule convert_r16g16_to_d24s8_frag; + vk::ShaderModule convert_d24s8_to_abgr8_frag; + vk::ShaderModule convert_d24s8_to_b10g11r11_frag; + vk::ShaderModule convert_d24s8_to_r16g16_frag; vk::Sampler linear_sampler; vk::Sampler nearest_sampler; @@ -94,6 +127,12 @@ private: vk::Pipeline convert_r32_to_d32_pipeline; vk::Pipeline convert_d16_to_r16_pipeline; vk::Pipeline convert_r16_to_d16_pipeline; + vk::Pipeline convert_abgr8_to_d24s8_pipeline; + vk::Pipeline convert_b10g11r11_to_d24s8_pipeline; + vk::Pipeline convert_r16g16_to_d24s8_pipeline; + vk::Pipeline convert_d24s8_to_abgr8_pipeline; + vk::Pipeline convert_d24s8_to_b10g11r11_pipeline; + vk::Pipeline convert_d24s8_to_r16g16_pipeline; }; } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp index 68a23b602..31adada56 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp @@ -208,6 +208,9 @@ struct FormatTuple { {VK_FORMAT_D32_SFLOAT, Attachable}, // D32_FLOAT {VK_FORMAT_D16_UNORM, Attachable}, // D16_UNORM + // Stencil formats + {VK_FORMAT_S8_UINT, Attachable}, // S8_UINT + // DepthStencil formats {VK_FORMAT_D24_UNORM_S8_UINT, Attachable}, // D24_UNORM_S8_UINT {VK_FORMAT_D24_UNORM_S8_UINT, Attachable}, // S8_UINT_D24_UNORM (emulated) diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index 407fd2a15..3964424af 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -8,6 +8,7 @@ #include <vector> #include "common/bit_cast.h" +#include "common/bit_util.h" #include "common/settings.h" #include "video_core/engines/fermi_2d.h" @@ -102,6 +103,7 @@ constexpr VkBorderColor ConvertBorderColor(const std::array<float, 4>& color) { usage |= VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT; break; case VideoCore::Surface::SurfaceType::Depth: + case VideoCore::Surface::SurfaceType::Stencil: case VideoCore::Surface::SurfaceType::DepthStencil: usage |= VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT; break; @@ -173,6 +175,8 @@ constexpr VkBorderColor ConvertBorderColor(const std::array<float, 4>& color) { return VK_IMAGE_ASPECT_COLOR_BIT; case VideoCore::Surface::SurfaceType::Depth: return VK_IMAGE_ASPECT_DEPTH_BIT; + case VideoCore::Surface::SurfaceType::Stencil: + return VK_IMAGE_ASPECT_STENCIL_BIT; case VideoCore::Surface::SurfaceType::DepthStencil: return VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT; default: @@ -195,6 +199,8 @@ constexpr VkBorderColor ConvertBorderColor(const std::array<float, 4>& color) { case PixelFormat::D16_UNORM: case PixelFormat::D32_FLOAT: return VK_IMAGE_ASPECT_DEPTH_BIT; + case PixelFormat::S8_UINT: + return VK_IMAGE_ASPECT_STENCIL_BIT; default: return VK_IMAGE_ASPECT_COLOR_BIT; } @@ -308,6 +314,19 @@ constexpr VkBorderColor ConvertBorderColor(const std::array<float, 4>& color) { }; } +[[nodiscard]] VkBufferImageCopy MakeBufferImageCopy(const VideoCommon::ImageCopy& copy, bool is_src, + VkImageAspectFlags aspect_mask) noexcept { + return VkBufferImageCopy{ + .bufferOffset = 0, + .bufferRowLength = 0, + .bufferImageHeight = 0, + .imageSubresource = MakeImageSubresourceLayers( + is_src ? copy.src_subresource : copy.dst_subresource, aspect_mask), + .imageOffset = MakeOffset3D(is_src ? copy.src_offset : copy.dst_offset), + .imageExtent = MakeExtent3D(copy.extent), + }; +} + [[maybe_unused]] [[nodiscard]] std::vector<VkBufferCopy> TransformBufferCopies( std::span<const VideoCommon::BufferCopy> copies, size_t buffer_offset) { std::vector<VkBufferCopy> result(copies.size()); @@ -754,6 +773,163 @@ StagingBufferRef TextureCacheRuntime::DownloadStagingBuffer(size_t size) { return staging_buffer_pool.Request(size, MemoryUsage::Download); } +bool TextureCacheRuntime::ShouldReinterpret(Image& dst, Image& src) { + if (VideoCore::Surface::GetFormatType(dst.info.format) == + VideoCore::Surface::SurfaceType::DepthStencil) { + return !device.IsExtShaderStencilExportSupported(); + } + return false; +} + +VkBuffer TextureCacheRuntime::GetTemporaryBuffer(size_t needed_size) { + const auto level = (8 * sizeof(size_t)) - std::countl_zero(needed_size - 1ULL); + if (buffer_commits[level]) { + return *buffers[level]; + } + const auto new_size = Common::NextPow2(needed_size); + VkBufferUsageFlags flags = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT | + VK_BUFFER_USAGE_UNIFORM_TEXEL_BUFFER_BIT | + VK_BUFFER_USAGE_STORAGE_TEXEL_BUFFER_BIT; + buffers[level] = device.GetLogical().CreateBuffer({ + .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .size = new_size, + .usage = flags, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + }); + buffer_commits[level] = std::make_unique<MemoryCommit>( + memory_allocator.Commit(buffers[level], MemoryUsage::DeviceLocal)); + return *buffers[level]; +} + +void TextureCacheRuntime::ReinterpretImage(Image& dst, Image& src, + std::span<const VideoCommon::ImageCopy> copies) { + std::vector<VkBufferImageCopy> vk_in_copies(copies.size()); + std::vector<VkBufferImageCopy> vk_out_copies(copies.size()); + const VkImageAspectFlags src_aspect_mask = src.AspectMask(); + const VkImageAspectFlags dst_aspect_mask = dst.AspectMask(); + + std::ranges::transform(copies, vk_in_copies.begin(), [src_aspect_mask](const auto& copy) { + return MakeBufferImageCopy(copy, true, src_aspect_mask); + }); + std::ranges::transform(copies, vk_out_copies.begin(), [dst_aspect_mask](const auto& copy) { + return MakeBufferImageCopy(copy, false, dst_aspect_mask); + }); + const u32 img_bpp = BytesPerBlock(src.info.format); + size_t total_size = 0; + for (const auto& copy : copies) { + total_size += copy.extent.width * copy.extent.height * copy.extent.depth * img_bpp; + } + const VkBuffer copy_buffer = GetTemporaryBuffer(total_size); + const VkImage dst_image = dst.Handle(); + const VkImage src_image = src.Handle(); + scheduler.RequestOutsideRenderPassOperationContext(); + scheduler.Record([dst_image, src_image, copy_buffer, src_aspect_mask, dst_aspect_mask, + vk_in_copies, vk_out_copies](vk::CommandBuffer cmdbuf) { + RangedBarrierRange dst_range; + RangedBarrierRange src_range; + for (const VkBufferImageCopy& copy : vk_in_copies) { + src_range.AddLayers(copy.imageSubresource); + } + for (const VkBufferImageCopy& copy : vk_out_copies) { + dst_range.AddLayers(copy.imageSubresource); + } + static constexpr VkMemoryBarrier READ_BARRIER{ + .sType = VK_STRUCTURE_TYPE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_MEMORY_WRITE_BIT, + .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_TRANSFER_WRITE_BIT, + }; + static constexpr VkMemoryBarrier WRITE_BARRIER{ + .sType = VK_STRUCTURE_TYPE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_MEMORY_READ_BIT | VK_ACCESS_MEMORY_WRITE_BIT, + }; + const std::array pre_barriers{ + VkImageMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT | VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | + VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT, + .oldLayout = VK_IMAGE_LAYOUT_GENERAL, + .newLayout = VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = src_image, + .subresourceRange = src_range.SubresourceRange(src_aspect_mask), + }, + }; + const std::array middle_in_barrier{ + VkImageMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = 0, + .dstAccessMask = 0, + .oldLayout = VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = src_image, + .subresourceRange = src_range.SubresourceRange(src_aspect_mask), + }, + }; + const std::array middle_out_barrier{ + VkImageMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT | VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | + VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .oldLayout = VK_IMAGE_LAYOUT_GENERAL, + .newLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = dst_image, + .subresourceRange = dst_range.SubresourceRange(dst_aspect_mask), + }, + }; + const std::array post_barriers{ + VkImageMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT | + VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | + VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | + VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_TRANSFER_WRITE_BIT, + .oldLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = dst_image, + .subresourceRange = dst_range.SubresourceRange(dst_aspect_mask), + }, + }; + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, + 0, {}, {}, pre_barriers); + + cmdbuf.CopyImageToBuffer(src_image, VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, copy_buffer, + vk_in_copies); + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, + 0, WRITE_BARRIER, nullptr, middle_in_barrier); + + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, + 0, READ_BARRIER, {}, middle_out_barrier); + cmdbuf.CopyBufferToImage(copy_buffer, dst_image, VK_IMAGE_LAYOUT_GENERAL, vk_out_copies); + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, + 0, {}, {}, post_barriers); + }); +} + void TextureCacheRuntime::BlitImage(Framebuffer* dst_framebuffer, ImageView& dst, ImageView& src, const Region2D& dst_region, const Region2D& src_region, Tegra::Engines::Fermi2D::Filter filter, @@ -881,6 +1057,22 @@ void TextureCacheRuntime::ConvertImage(Framebuffer* dst, ImageView& dst_view, Im return blit_image_helper.ConvertD16ToR16(dst, src_view, up_scale, down_shift); } break; + case PixelFormat::A8B8G8R8_UNORM: + case PixelFormat::B8G8R8A8_UNORM: + if (src_view.format == PixelFormat::S8_UINT_D24_UNORM) { + return blit_image_helper.ConvertD24S8ToABGR8(dst, src_view, up_scale, down_shift); + } + break; + case PixelFormat::B10G11R11_FLOAT: + if (src_view.format == PixelFormat::S8_UINT_D24_UNORM) { + return blit_image_helper.ConvertD24S8ToB10G11R11(dst, src_view, up_scale, down_shift); + } + break; + case PixelFormat::R16G16_UNORM: + if (src_view.format == PixelFormat::S8_UINT_D24_UNORM) { + return blit_image_helper.ConvertD24S8ToR16G16(dst, src_view, up_scale, down_shift); + } + break; case PixelFormat::R32_FLOAT: if (src_view.format == PixelFormat::D32_FLOAT) { return blit_image_helper.ConvertD32ToR32(dst, src_view, up_scale, down_shift); @@ -891,6 +1083,18 @@ void TextureCacheRuntime::ConvertImage(Framebuffer* dst, ImageView& dst_view, Im return blit_image_helper.ConvertR16ToD16(dst, src_view, up_scale, down_shift); } break; + case PixelFormat::S8_UINT_D24_UNORM: + if (src_view.format == PixelFormat::A8B8G8R8_UNORM || + src_view.format == PixelFormat::B8G8R8A8_UNORM) { + return blit_image_helper.ConvertABGR8ToD24S8(dst, src_view, up_scale, down_shift); + } + if (src_view.format == PixelFormat::B10G11R11_FLOAT) { + return blit_image_helper.ConvertB10G11R11ToD24S8(dst, src_view, up_scale, down_shift); + } + if (src_view.format == PixelFormat::R16G16_UNORM) { + return blit_image_helper.ConvertR16G16ToD24S8(dst, src_view, up_scale, down_shift); + } + break; case PixelFormat::D32_FLOAT: if (src_view.format == PixelFormat::R32_FLOAT) { return blit_image_helper.ConvertR32ToD32(dst, src_view, up_scale, down_shift); diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.h b/src/video_core/renderer_vulkan/vk_texture_cache.h index ff28b4e96..44e9dcee4 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.h +++ b/src/video_core/renderer_vulkan/vk_texture_cache.h @@ -61,6 +61,10 @@ public: void CopyImage(Image& dst, Image& src, std::span<const VideoCommon::ImageCopy> copies); + bool ShouldReinterpret(Image& dst, Image& src); + + void ReinterpretImage(Image& dst, Image& src, std::span<const VideoCommon::ImageCopy> copies); + void ConvertImage(Framebuffer* dst, ImageView& dst_view, ImageView& src_view, bool rescaled); bool CanAccelerateImageUpload(Image&) const noexcept { @@ -82,6 +86,8 @@ public: return true; } + [[nodiscard]] VkBuffer GetTemporaryBuffer(size_t needed_size); + const Device& device; VKScheduler& scheduler; MemoryAllocator& memory_allocator; @@ -90,6 +96,10 @@ public: ASTCDecoderPass& astc_decoder_pass; RenderPassCache& render_pass_cache; const Settings::ResolutionScalingInfo& resolution; + + constexpr static size_t indexing_slots = 8 * sizeof(size_t); + std::array<vk::Buffer, indexing_slots> buffers{}; + std::array<std::unique_ptr<MemoryCommit>, indexing_slots> buffer_commits{}; }; class Image : public VideoCommon::ImageBase { diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp index 58d262446..a36015c8c 100644 --- a/src/video_core/surface.cpp +++ b/src/video_core/surface.cpp @@ -82,6 +82,8 @@ PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) { return PixelFormat::D32_FLOAT; case Tegra::DepthFormat::D16_UNORM: return PixelFormat::D16_UNORM; + case Tegra::DepthFormat::S8_UINT: + return PixelFormat::S8_UINT; case Tegra::DepthFormat::D32_FLOAT_S8X24_UINT: return PixelFormat::D32_FLOAT_S8_UINT; default: @@ -214,6 +216,11 @@ SurfaceType GetFormatType(PixelFormat pixel_format) { } if (static_cast<std::size_t>(pixel_format) < + static_cast<std::size_t>(PixelFormat::MaxStencilFormat)) { + return SurfaceType::Stencil; + } + + if (static_cast<std::size_t>(pixel_format) < static_cast<std::size_t>(PixelFormat::MaxDepthStencilFormat)) { return SurfaceType::DepthStencil; } diff --git a/src/video_core/surface.h b/src/video_core/surface.h index 2ce7c7d33..33e8d24ab 100644 --- a/src/video_core/surface.h +++ b/src/video_core/surface.h @@ -110,8 +110,12 @@ enum class PixelFormat { MaxDepthFormat, + // Stencil formats + S8_UINT = MaxDepthFormat, + MaxStencilFormat, + // DepthStencil formats - D24_UNORM_S8_UINT = MaxDepthFormat, + D24_UNORM_S8_UINT = MaxStencilFormat, S8_UINT_D24_UNORM, D32_FLOAT_S8_UINT, @@ -125,8 +129,9 @@ constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max enum class SurfaceType { ColorTexture = 0, Depth = 1, - DepthStencil = 2, - Invalid = 3, + Stencil = 2, + DepthStencil = 3, + Invalid = 4, }; enum class SurfaceTarget { @@ -229,6 +234,7 @@ constexpr std::array<u32, MaxPixelFormat> BLOCK_WIDTH_TABLE = {{ 1, // E5B9G9R9_FLOAT 1, // D32_FLOAT 1, // D16_UNORM + 1, // S8_UINT 1, // D24_UNORM_S8_UINT 1, // S8_UINT_D24_UNORM 1, // D32_FLOAT_S8_UINT @@ -328,6 +334,7 @@ constexpr std::array<u32, MaxPixelFormat> BLOCK_HEIGHT_TABLE = {{ 1, // E5B9G9R9_FLOAT 1, // D32_FLOAT 1, // D16_UNORM + 1, // S8_UINT 1, // D24_UNORM_S8_UINT 1, // S8_UINT_D24_UNORM 1, // D32_FLOAT_S8_UINT @@ -427,6 +434,7 @@ constexpr std::array<u32, MaxPixelFormat> BITS_PER_BLOCK_TABLE = {{ 32, // E5B9G9R9_FLOAT 32, // D32_FLOAT 16, // D16_UNORM + 8, // S8_UINT 32, // D24_UNORM_S8_UINT 32, // S8_UINT_D24_UNORM 64, // D32_FLOAT_S8_UINT diff --git a/src/video_core/texture_cache/formatter.h b/src/video_core/texture_cache/formatter.h index c6cf0583f..b2c81057b 100644 --- a/src/video_core/texture_cache/formatter.h +++ b/src/video_core/texture_cache/formatter.h @@ -194,6 +194,8 @@ struct fmt::formatter<VideoCore::Surface::PixelFormat> : fmt::formatter<fmt::str return "D32_FLOAT"; case PixelFormat::D16_UNORM: return "D16_UNORM"; + case PixelFormat::S8_UINT: + return "S8_UINT"; case PixelFormat::D24_UNORM_S8_UINT: return "D24_UNORM_S8_UINT"; case PixelFormat::S8_UINT_D24_UNORM: diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index 4d2874bf2..44a0d42ba 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -475,6 +475,7 @@ void TextureCache<P>::BlitImage(const Tegra::Engines::Fermi2D::Surface& dst, const BlitImages images = GetBlitImages(dst, src); const ImageId dst_id = images.dst_id; const ImageId src_id = images.src_id; + PrepareImage(src_id, false, false); PrepareImage(dst_id, true, false); @@ -758,7 +759,8 @@ ImageId TextureCache<P>::FindImage(const ImageInfo& info, GPUVAddr gpu_addr, return ImageId{}; } } - const bool broken_views = runtime.HasBrokenTextureViewFormats(); + const bool broken_views = + runtime.HasBrokenTextureViewFormats() || True(options & RelaxedOptions::ForceBrokenViews); const bool native_bgr = runtime.HasNativeBgr(); ImageId image_id; const auto lambda = [&](ImageId existing_image_id, ImageBase& existing_image) { @@ -1094,12 +1096,13 @@ typename TextureCache<P>::BlitImages TextureCache<P>::GetBlitImages( if (GetFormatType(dst_info.format) != GetFormatType(src_info.format)) { continue; } - if (!dst_id) { - dst_id = InsertImage(dst_info, dst_addr, RelaxedOptions{}); - } - if (!src_id) { - src_id = InsertImage(src_info, src_addr, RelaxedOptions{}); + RelaxedOptions find_options{}; + if (src_info.num_samples > 1) { + // it's a resolve, we must enforce the same format. + find_options = RelaxedOptions::ForceBrokenViews; } + src_id = FindOrInsertImage(src_info, src_addr, find_options); + dst_id = FindOrInsertImage(dst_info, dst_addr, find_options); } while (has_deleted_images); return BlitImages{ .dst_id = dst_id, @@ -1759,6 +1762,9 @@ void TextureCache<P>::CopyImage(ImageId dst_id, ImageId src_id, std::vector<Imag } UNIMPLEMENTED_IF(dst.info.type != ImageType::e2D); UNIMPLEMENTED_IF(src.info.type != ImageType::e2D); + if (runtime.ShouldReinterpret(dst, src)) { + return runtime.ReinterpretImage(dst, src, copies); + } for (const ImageCopy& copy : copies) { UNIMPLEMENTED_IF(copy.dst_subresource.num_layers != 1); UNIMPLEMENTED_IF(copy.src_subresource.num_layers != 1); diff --git a/src/video_core/texture_cache/types.h b/src/video_core/texture_cache/types.h index 5c274abdf..5ac27b3a7 100644 --- a/src/video_core/texture_cache/types.h +++ b/src/video_core/texture_cache/types.h @@ -54,6 +54,7 @@ enum class RelaxedOptions : u32 { Size = 1 << 0, Format = 1 << 1, Samples = 1 << 2, + ForceBrokenViews = 1 << 3, }; DECLARE_ENUM_FLAG_OPERATORS(RelaxedOptions) diff --git a/src/video_core/texture_cache/util.cpp b/src/video_core/texture_cache/util.cpp index ddc9fb13a..e4d82631e 100644 --- a/src/video_core/texture_cache/util.cpp +++ b/src/video_core/texture_cache/util.cpp @@ -1151,18 +1151,39 @@ bool IsSubresource(const ImageInfo& candidate, const ImageBase& image, GPUVAddr void DeduceBlitImages(ImageInfo& dst_info, ImageInfo& src_info, const ImageBase* dst, const ImageBase* src) { - if (src && GetFormatType(src->info.format) != SurfaceType::ColorTexture) { - src_info.format = src->info.format; + bool is_resolve = false; + const auto original_src_format = src_info.format; + const auto original_dst_format = dst_info.format; + if (src) { + if (GetFormatType(src->info.format) != SurfaceType::ColorTexture) { + src_info.format = src->info.format; + } + is_resolve = src->info.num_samples > 1; + src_info.num_samples = src->info.num_samples; + src_info.size = src->info.size; } if (dst && GetFormatType(dst->info.format) != SurfaceType::ColorTexture) { dst_info.format = dst->info.format; } if (src && GetFormatType(src->info.format) != SurfaceType::ColorTexture) { - dst_info.format = src->info.format; + if (dst) { + if (GetFormatType(dst->info.format) == SurfaceType::ColorTexture) { + src_info.format = original_src_format; + } + } else { + dst_info.format = src->info.format; + } } if (dst && GetFormatType(dst->info.format) != SurfaceType::ColorTexture) { - src_info.format = dst->info.format; + if (src) { + if (GetFormatType(src->info.format) == SurfaceType::ColorTexture) { + dst_info.format = original_dst_format; + } + } else { + src_info.format = dst->info.format; + } } + ASSERT(!is_resolve || dst_info.format == src_info.format); } u32 MapSizeBytes(const ImageBase& image) { diff --git a/src/video_core/vulkan_common/vulkan_device.cpp b/src/video_core/vulkan_common/vulkan_device.cpp index 95106f88f..70c52aaac 100644 --- a/src/video_core/vulkan_common/vulkan_device.cpp +++ b/src/video_core/vulkan_common/vulkan_device.cpp @@ -21,6 +21,13 @@ namespace Vulkan { namespace { namespace Alternatives { +constexpr std::array STENCIL8_UINT{ + VK_FORMAT_D16_UNORM_S8_UINT, + VK_FORMAT_D24_UNORM_S8_UINT, + VK_FORMAT_D32_SFLOAT_S8_UINT, + VK_FORMAT_UNDEFINED, +}; + constexpr std::array DEPTH24_UNORM_STENCIL8_UINT{ VK_FORMAT_D32_SFLOAT_S8_UINT, VK_FORMAT_D16_UNORM_S8_UINT, @@ -74,6 +81,8 @@ void SetNext(void**& next, T& data) { constexpr const VkFormat* GetFormatAlternatives(VkFormat format) { switch (format) { + case VK_FORMAT_S8_UINT: + return Alternatives::STENCIL8_UINT.data(); case VK_FORMAT_D24_UNORM_S8_UINT: return Alternatives::DEPTH24_UNORM_STENCIL8_UINT.data(); case VK_FORMAT_D16_UNORM_S8_UINT: @@ -145,6 +154,7 @@ std::unordered_map<VkFormat, VkFormatProperties> GetFormatProperties(vk::Physica VK_FORMAT_R4G4B4A4_UNORM_PACK16, VK_FORMAT_D32_SFLOAT, VK_FORMAT_D16_UNORM, + VK_FORMAT_S8_UINT, VK_FORMAT_D16_UNORM_S8_UINT, VK_FORMAT_D24_UNORM_S8_UINT, VK_FORMAT_D32_SFLOAT_S8_UINT, diff --git a/src/yuzu/configuration/configure_graphics.ui b/src/yuzu/configuration/configure_graphics.ui index 660b68c1c..9241678e4 100644 --- a/src/yuzu/configuration/configure_graphics.ui +++ b/src/yuzu/configuration/configure_graphics.ui @@ -429,7 +429,7 @@ </item> <item> <property name="text"> - <string>AMD's FidelityFX™️ Super Resolution [Vulkan Only]</string> + <string>AMD FidelityFX™️ Super Resolution [Vulkan Only]</string> </property> </item> </widget> diff --git a/src/yuzu/hotkeys.cpp b/src/yuzu/hotkeys.cpp index d4e97fa16..e7e58f314 100644 --- a/src/yuzu/hotkeys.cpp +++ b/src/yuzu/hotkeys.cpp @@ -46,6 +46,8 @@ QShortcut* HotkeyRegistry::GetHotkey(const QString& group, const QString& action if (!hk.shortcut) hk.shortcut = new QShortcut(hk.keyseq, widget, nullptr, nullptr, hk.context); + hk.shortcut->setAutoRepeat(false); + return hk.shortcut; } diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index d057dc889..c4c76b094 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -1031,7 +1031,7 @@ void GMainWindow::InitializeHotkeys() { &QShortcut::activatedAmbiguously, ui->action_Fullscreen, &QAction::trigger); connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Exit Fullscreen"), this), &QShortcut::activated, this, [&] { - if (emulation_running) { + if (emulation_running && ui->action_Fullscreen->isChecked()) { ui->action_Fullscreen->setChecked(false); ToggleFullscreen(); } @@ -3106,7 +3106,7 @@ void GMainWindow::UpdateFilterText() { filter_status_button->setText(tr("SCALEFORCE")); break; case Settings::ScalingFilter::Fsr: - filter_status_button->setText(tr("AMD'S FIDELITYFX SR")); + filter_status_button->setText(tr("FSR")); break; default: filter_status_button->setText(tr("BILINEAR")); @@ -3117,15 +3117,15 @@ void GMainWindow::UpdateFilterText() { void GMainWindow::UpdateAAText() { const auto aa_mode = Settings::values.anti_aliasing.GetValue(); switch (aa_mode) { - case Settings::AntiAliasing::Fxaa: - aa_status_button->setText(tr("FXAA")); - break; case Settings::AntiAliasing::None: aa_status_button->setText(tr("NO AA")); break; - default: + case Settings::AntiAliasing::Fxaa: aa_status_button->setText(tr("FXAA")); break; + default: + aa_status_button->setText(tr("NO AA")); + break; } } @@ -3300,9 +3300,9 @@ void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) { if (!errors.isEmpty()) { QMessageBox::warning( this, tr("Derivation Components Missing"), - tr("Components are missing that may hinder key derivation from completing. " + tr("Encryption keys are missing. " "<br>Please follow <a href='https://yuzu-emu.org/help/quickstart/'>the yuzu " - "quickstart guide</a> to get all your keys and " + "quickstart guide</a> to get all your keys, firmware and " "games.<br><br><small>(%1)</small>") .arg(errors)); } |