early-access version 1850

README.md

@@ -1,7 +1,7 @@
 yuzu emulator early access
 =============
 
-This is the source code for early-access 1848.
+This is the source code for early-access 1850.
 
 ## Legal Notice
 

src/CMakeLists.txt

@@ -45,14 +45,23 @@ if (MSVC)
         /Zc:inline
         /Zc:throwingNew
 
+        # External headers diagnostics
+        /experimental:external  # Enables the external headers options. This option isn't required in Visual Studio 2019 version 16.10 and later
+        /external:anglebrackets # Treats all headers included by #include <header>, where the header file is enclosed in angle brackets (< >), as external headers
+        /external:W0            # Sets the default warning level to 0 for external headers, effectively turning off warnings for external headers
+
         # Warnings
         /W3
-        /we4062 # enumerator 'identifier' in a switch of enum 'enumeration' is not handled
+        /we4018 # 'expression': signed/unsigned mismatch
+        /we4062 # Enumerator 'identifier' in a switch of enum 'enumeration' is not handled
         /we4101 # 'identifier': unreferenced local variable
         /we4189 # 'identifier': local variable is initialized but not referenced
         /we4265 # 'class': class has virtual functions, but destructor is not virtual
-        /we4388 # signed/unsigned mismatch
+        /we4267 # 'var': conversion from 'size_t' to 'type', possible loss of data
-        /we4547 # 'operator' : operator before comma has no effect; expected operator with side-effect
+        /we4305 # 'context': truncation from 'type1' to 'type2'
+        /we4388 # 'expression': signed/unsigned mismatch
+        /we4389 # 'operator': signed/unsigned mismatch
+        /we4547 # 'operator': operator before comma has no effect; expected operator with side-effect
         /we4549 # 'operator1': operator before comma has no effect; did you intend 'operator2'?
         /we4555 # Expression has no effect; expected expression with side-effect
         /we4715 # 'function': not all control paths return a value

src/audio_core/command_generator.cpp

@@ -400,7 +400,10 @@ void CommandGenerator::GenerateDataSourceCommand(ServerVoiceInfo& voice_info, Vo
         }
     } else {
         switch (in_params.sample_format) {
+        case SampleFormat::Pcm8:
         case SampleFormat::Pcm16:
+        case SampleFormat::Pcm32:
+        case SampleFormat::PcmFloat:
             DecodeFromWaveBuffers(voice_info, GetChannelMixBuffer(channel), dsp_state, channel,
                                   worker_params.sample_rate, worker_params.sample_count,
                                   in_params.node_id);
@@ -1003,9 +1006,10 @@ void CommandGenerator::GenerateFinalMixCommand() {
     }
 }
 
-s32 CommandGenerator::DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_state,
+template <typename T>
-                                  s32 sample_start_offset, s32 sample_end_offset, s32 sample_count,
+s32 CommandGenerator::DecodePcm(ServerVoiceInfo& voice_info, VoiceState& dsp_state,
-                                  s32 channel, std::size_t mix_offset) {
+                                s32 sample_start_offset, s32 sample_end_offset, s32 sample_count,
+                                s32 channel, std::size_t mix_offset) {
     const auto& in_params = voice_info.GetInParams();
     const auto& wave_buffer = in_params.wave_buffer[dsp_state.wave_buffer_index];
     if (wave_buffer.buffer_address == 0) {
@@ -1019,21 +1023,20 @@ s32 CommandGenerator::DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_s
     }
     const auto samples_remaining = (sample_end_offset - sample_start_offset) - dsp_state.offset;
     const auto start_offset =
-        ((dsp_state.offset + sample_start_offset) * in_params.channel_count) * sizeof(s16);
+        ((dsp_state.offset + sample_start_offset) * in_params.channel_count) * sizeof(T);
     const auto buffer_pos = wave_buffer.buffer_address + start_offset;
     const auto samples_processed = std::min(sample_count, samples_remaining);
 
-    if (in_params.channel_count == 1) {
+    const auto channel_count = in_params.channel_count;
-        std::vector<s16> buffer(samples_processed);
+    std::vector<T> buffer(samples_processed * channel_count);
-        memory.ReadBlock(buffer_pos, buffer.data(), buffer.size() * sizeof(s16));
+    memory.ReadBlock(buffer_pos, buffer.data(), buffer.size() * sizeof(T));
-        for (std::size_t i = 0; i < buffer.size(); i++) {
+
-            sample_buffer[mix_offset + i] = buffer[i];
+    if constexpr (std::is_floating_point_v<T>) {
+        for (std::size_t i = 0; i < static_cast<std::size_t>(samples_processed); i++) {
+            sample_buffer[mix_offset + i] = static_cast<s32>(buffer[i * channel_count + channel] *
+                                                             std::numeric_limits<s16>::max());
         }
     } else {
-        const auto channel_count = in_params.channel_count;
-        std::vector<s16> buffer(samples_processed * channel_count);
-        memory.ReadBlock(buffer_pos, buffer.data(), buffer.size() * sizeof(s16));
-
         for (std::size_t i = 0; i < static_cast<std::size_t>(samples_processed); i++) {
             sample_buffer[mix_offset + i] = buffer[i * channel_count + channel];
         }
@@ -1249,10 +1252,25 @@ void CommandGenerator::DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* o
 
             s32 samples_decoded{0};
             switch (in_params.sample_format) {
+            case SampleFormat::Pcm8:
+                samples_decoded =
+                    DecodePcm<s8>(voice_info, dsp_state, samples_offset_start, samples_offset_end,
+                                  samples_to_read - samples_read, channel, temp_mix_offset);
+                break;
             case SampleFormat::Pcm16:
                 samples_decoded =
-                    DecodePcm16(voice_info, dsp_state, samples_offset_start, samples_offset_end,
+                    DecodePcm<s16>(voice_info, dsp_state, samples_offset_start, samples_offset_end,
-                                samples_to_read - samples_read, channel, temp_mix_offset);
+                                   samples_to_read - samples_read, channel, temp_mix_offset);
+                break;
+            case SampleFormat::Pcm32:
+                samples_decoded =
+                    DecodePcm<s32>(voice_info, dsp_state, samples_offset_start, samples_offset_end,
+                                   samples_to_read - samples_read, channel, temp_mix_offset);
+                break;
+            case SampleFormat::PcmFloat:
+                samples_decoded =
+                    DecodePcm<f32>(voice_info, dsp_state, samples_offset_start, samples_offset_end,
+                                   samples_to_read - samples_read, channel, temp_mix_offset);
                 break;
             case SampleFormat::Adpcm:
                 samples_decoded =

src/audio_core/command_generator.h

@@ -86,8 +86,9 @@ private:
                                std::vector<u8>& work_buffer);
     void UpdateI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state, bool should_clear);
     // DSP Code
-    s32 DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 sample_start_offset,
+    template <typename T>
-                    s32 sample_end_offset, s32 sample_count, s32 channel, std::size_t mix_offset);
+    s32 DecodePcm(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 sample_start_offset,
+                  s32 sample_end_offset, s32 sample_count, s32 channel, std::size_t mix_offset);
     s32 DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 sample_start_offset,
                     s32 sample_end_offset, s32 sample_count, s32 channel, std::size_t mix_offset);
     void DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* output, VoiceState& dsp_state,

src/core/CMakeLists.txt

@@ -656,17 +656,14 @@ endif()
 
 if (MSVC)
     target_compile_options(core PRIVATE
-        /we4018 # 'expression' : signed/unsigned mismatch
+        /we4242 # 'identifier': conversion from 'type1' to 'type2', possible loss of data
-        /we4244 # 'argument' : conversion from 'type1' to 'type2', possible loss of data (floating-point)
+        /we4244 # 'conversion': conversion from 'type1' to 'type2', possible loss of data
-        /we4245 # 'conversion' : conversion from 'type1' to 'type2', signed/unsigned mismatch
+        /we4245 # 'conversion': conversion from 'type1' to 'type2', signed/unsigned mismatch
         /we4254 # 'operator': conversion from 'type1:field_bits' to 'type2:field_bits', possible loss of data
-        /we4267 # 'var' : conversion from 'size_t' to 'type', possible loss of data
-        /we4305 # 'context' : truncation from 'type1' to 'type2'
         /we4456 # Declaration of 'identifier' hides previous local declaration
         /we4457 # Declaration of 'identifier' hides function parameter
         /we4458 # Declaration of 'identifier' hides class member
         /we4459 # Declaration of 'identifier' hides global declaration
-        /we4715 # 'function' : not all control paths return a value
     )
 else()
     target_compile_options(core PRIVATE

src/input_common/CMakeLists.txt

@@ -34,18 +34,10 @@ if (MSVC)
         /W4
         /WX
 
-        # 'expression' : signed/unsigned mismatch
+        /we4242 # 'identifier': conversion from 'type1' to 'type2', possible loss of data
-        /we4018
+        /we4244 # 'conversion': conversion from 'type1' to 'type2', possible loss of data
-        # 'argument' : conversion from 'type1' to 'type2', possible loss of data (floating-point)
+        /we4245 # 'conversion': conversion from 'type1' to 'type2', signed/unsigned mismatch
-        /we4244
+        /we4254 # 'operator': conversion from 'type1:field_bits' to 'type2:field_bits', possible loss of data
-        # 'conversion' : conversion from 'type1' to 'type2', signed/unsigned mismatch
-        /we4245
-        # 'operator': conversion from 'type1:field_bits' to 'type2:field_bits', possible loss of data
-        /we4254
-        # 'var' : conversion from 'size_t' to 'type', possible loss of data
-        /we4267
-        # 'context' : truncation from 'type1' to 'type2'
-        /we4305
     )
 else()
     target_compile_options(input_common PRIVATE

src/input_common/gcadapter/gc_adapter.cpp

@@ -5,14 +5,7 @@
 #include <chrono>
 #include <thread>
 
-#ifdef _MSC_VER
-#pragma warning(push)
-#pragma warning(disable : 4200) // nonstandard extension used : zero-sized array in struct/union
-#endif
 #include <libusb.h>
-#ifdef _MSC_VER
-#pragma warning(pop)
-#endif
 
 #include "common/logging/log.h"
 #include "common/param_package.h"

src/input_common/udp/protocol.h

@@ -8,14 +8,7 @@
 #include <optional>
 #include <type_traits>
 
-#ifdef _MSC_VER
-#pragma warning(push)
-#pragma warning(disable : 4701)
-#endif
 #include <boost/crc.hpp>
-#ifdef _MSC_VER
-#pragma warning(pop)
-#endif
 
 #include "common/bit_field.h"
 #include "common/swap.h"

src/video_core/CMakeLists.txt

@@ -292,13 +292,12 @@ endif()
 
 if (MSVC)
     target_compile_options(video_core PRIVATE
-        /we4267 # 'var' : conversion from 'size_t' to 'type', possible loss of data
+        /we4242 # 'identifier': conversion from 'type1' to 'type2', possible loss of data
-        /we4244 # 'var' : conversion from integer to 'type', possible loss of data
+        /we4244 # 'conversion': conversion from 'type1' to 'type2', possible loss of data
         /we4456 # Declaration of 'identifier' hides previous local declaration
         /we4457 # Declaration of 'identifier' hides function parameter
         /we4458 # Declaration of 'identifier' hides class member
         /we4459 # Declaration of 'identifier' hides global declaration
-        /we4715 # 'function' : not all control paths return a value
     )
 else()
     target_compile_options(video_core PRIVATE

src/video_core/command_classes/codecs/codec.h

@@ -14,18 +14,10 @@ extern "C" {
 #pragma GCC diagnostic push
 #pragma GCC diagnostic ignored "-Wconversion"
 #endif
-#ifdef _MSC_VER
-#pragma warning(push)
-#pragma warning(disable : 4242) // conversion from 'type' to 'type', possible loss of data
-#pragma warning(disable : 4244) // conversion from 'type' to 'type', possible loss of data
-#endif
 #include <libavcodec/avcodec.h>
 #if defined(__GNUC__) || defined(__clang__)
 #pragma GCC diagnostic pop
 #endif
-#ifdef _MSC_VER
-#pragma warning(pop)
-#endif
 }
 
 namespace Tegra {

src/video_core/command_classes/vic.cpp

@@ -9,17 +9,10 @@ extern "C" {
 #pragma GCC diagnostic push
 #pragma GCC diagnostic ignored "-Wconversion"
 #endif
-#ifdef _MSC_VER
-#pragma warning(disable : 4244) // conversion from 'type' to 'type', possible loss of data
-#pragma warning(push)
-#endif
 #include <libswscale/swscale.h>
 #if defined(__GNUC__) || defined(__clang__)
 #pragma GCC diagnostic pop
 #endif
-#ifdef _MSC_VER
-#pragma warning(pop)
-#endif
 }
 
 #include "common/assert.h"

src/video_core/engines/fermi_2d.cpp

@@ -7,10 +7,6 @@
 #include "video_core/engines/fermi_2d.h"
 #include "video_core/memory_manager.h"
 #include "video_core/rasterizer_interface.h"
-#include "video_core/surface.h"
-
-using VideoCore::Surface::BytesPerBlock;
-using VideoCore::Surface::PixelFormatFromRenderTargetFormat;
 
 namespace Tegra::Engines {
 
@@ -53,7 +49,7 @@ void Fermi2D::Blit() {
     UNIMPLEMENTED_IF_MSG(regs.clip_enable != 0, "Clipped blit enabled");
 
     const auto& args = regs.pixels_from_memory;
-    Config config{
+    const Config config{
         .operation = regs.operation,
         .filter = args.sample_mode.filter,
         .dst_x0 = args.dst_x0,
@@ -65,20 +61,7 @@ void Fermi2D::Blit() {
         .src_x1 = static_cast<s32>((args.du_dx * args.dst_width + args.src_x0) >> 32),
         .src_y1 = static_cast<s32>((args.dv_dy * args.dst_height + args.src_y0) >> 32),
     };
-    Surface src = regs.src;
+    if (!rasterizer->AccelerateSurfaceCopy(regs.src, regs.dst, config)) {
-    const auto bytes_per_pixel = BytesPerBlock(PixelFormatFromRenderTargetFormat(src.format));
-    const auto need_align_to_pitch =
-        src.linear == Tegra::Engines::Fermi2D::MemoryLayout::Pitch && src.width == config.src_x1 &&
-        config.src_x1 > static_cast<s32>(src.pitch / bytes_per_pixel) && config.src_x0 > 0;
-    if (need_align_to_pitch) {
-        auto address = src.Address() + config.src_x0 * bytes_per_pixel;
-        src.addr_upper = static_cast<u32>(address >> 32);
-        src.addr_lower = static_cast<u32>(address);
-        src.width -= config.src_x0;
-        config.src_x1 -= config.src_x0;
-        config.src_x0 = 0;
-    }
-    if (!rasterizer->AccelerateSurfaceCopy(src, regs.dst, config)) {
         UNIMPLEMENTED();
     }
 }

src/video_core/renderer_opengl/gl_texture_cache.cpp

@@ -327,7 +327,8 @@ void ApplySwizzle(GLuint handle, PixelFormat format, std::array<SwizzleSource, 4
     if (format_info.is_compressed) {
         return false;
     }
-    if (std::ranges::find(ACCELERATED_FORMATS, internal_format) == ACCELERATED_FORMATS.end()) {
+    if (std::ranges::find(ACCELERATED_FORMATS, static_cast<int>(internal_format)) ==
+        ACCELERATED_FORMATS.end()) {
         return false;
     }
     if (format_info.compatibility_by_size) {

src/video_core/texture_cache/image_base.cpp

@@ -85,7 +85,7 @@ std::optional<SubresourceBase> ImageBase::TryFindBase(GPUVAddr other_addr) const
     if (info.type != ImageType::e3D) {
         const auto [layer, mip_offset] = LayerMipOffset(diff, info.layer_stride);
         const auto end = mip_level_offsets.begin() + info.resources.levels;
-        const auto it = std::find(mip_level_offsets.begin(), end, mip_offset);
+        const auto it = std::find(mip_level_offsets.begin(), end, static_cast<u32>(mip_offset));
         if (layer > info.resources.layers || it == end) {
             return std::nullopt;
         }

src/video_core/texture_cache/texture_cache.h

@@ -159,7 +159,9 @@ public:
     /// Blit an image with the given parameters
     void BlitImage(const Tegra::Engines::Fermi2D::Surface& dst,
                    const Tegra::Engines::Fermi2D::Surface& src,
-                   const Tegra::Engines::Fermi2D::Config& copy);
+                   const Tegra::Engines::Fermi2D::Config& copy,
+                   std::optional<Region2D> src_region_override = {},
+                   std::optional<Region2D> dst_region_override = {});
 
     /// Invalidate the contents of the color buffer index
     /// These contents become unspecified, the cache can assume aggressive optimizations.
@@ -758,7 +760,9 @@ void TextureCache<P>::UnmapGPUMemory(GPUVAddr gpu_addr, size_t size) {
 template <class P>
 void TextureCache<P>::BlitImage(const Tegra::Engines::Fermi2D::Surface& dst,
                                 const Tegra::Engines::Fermi2D::Surface& src,
-                                const Tegra::Engines::Fermi2D::Config& copy) {
+                                const Tegra::Engines::Fermi2D::Config& copy,
+                                std::optional<Region2D> src_override,
+                                std::optional<Region2D> dst_override) {
     const BlitImages images = GetBlitImages(dst, src);
     const ImageId dst_id = images.dst_id;
     const ImageId src_id = images.src_id;
@@ -769,25 +773,47 @@ void TextureCache<P>::BlitImage(const Tegra::Engines::Fermi2D::Surface& dst,
     const ImageBase& src_image = slot_images[src_id];
 
     // TODO: Deduplicate
-    const std::optional src_base = src_image.TryFindBase(src.Address());
-    const SubresourceRange src_range{.base = src_base.value(), .extent = {1, 1}};
-    const ImageViewInfo src_view_info(ImageViewType::e2D, images.src_format, src_range);
-    const auto [src_framebuffer_id, src_view_id] = RenderTargetFromImage(src_id, src_view_info);
-    const auto [src_samples_x, src_samples_y] = SamplesLog2(src_image.info.num_samples);
-    const Region2D src_region{
-        Offset2D{.x = copy.src_x0 >> src_samples_x, .y = copy.src_y0 >> src_samples_y},
-        Offset2D{.x = copy.src_x1 >> src_samples_x, .y = copy.src_y1 >> src_samples_y},
-    };
-
     const std::optional dst_base = dst_image.TryFindBase(dst.Address());
     const SubresourceRange dst_range{.base = dst_base.value(), .extent = {1, 1}};
     const ImageViewInfo dst_view_info(ImageViewType::e2D, images.dst_format, dst_range);
     const auto [dst_framebuffer_id, dst_view_id] = RenderTargetFromImage(dst_id, dst_view_info);
-    const auto [dst_samples_x, dst_samples_y] = SamplesLog2(dst_image.info.num_samples);
+    const auto [src_samples_x, src_samples_y] = SamplesLog2(src_image.info.num_samples);
-    const Region2D dst_region{
+
-        Offset2D{.x = copy.dst_x0 >> dst_samples_x, .y = copy.dst_y0 >> dst_samples_y},
+    // out of bounds texture blit checking
-        Offset2D{.x = copy.dst_x1 >> dst_samples_x, .y = copy.dst_y1 >> dst_samples_y},
+    const bool use_override = src_override.has_value();
+    const s32 src_x0 = copy.src_x0 >> src_samples_x;
+    s32 src_x1 = use_override ? src_override->end.x : copy.src_x1 >> src_samples_x;
+    const s32 src_y0 = copy.src_y0 >> src_samples_y;
+    const s32 src_y1 = copy.src_y1 >> src_samples_y;
+
+    const auto src_width = static_cast<s32>(src_image.info.size.width);
+    const bool width_oob = src_x1 > src_width;
+    const auto width_diff = width_oob ? src_x1 - src_width : 0;
+    if (width_oob) {
+        src_x1 = src_width;
+    }
+
+    const Region2D src_dimensions{
+        Offset2D{.x = src_x0, .y = src_y0},
+        Offset2D{.x = src_x1, .y = src_y1},
     };
+    const auto src_region = use_override ? *src_override : src_dimensions;
+
+    const std::optional src_base = src_image.TryFindBase(src.Address());
+    const SubresourceRange src_range{.base = src_base.value(), .extent = {1, 1}};
+    const ImageViewInfo src_view_info(ImageViewType::e2D, images.src_format, src_range);
+    const auto [src_framebuffer_id, src_view_id] = RenderTargetFromImage(src_id, src_view_info);
+    const auto [dst_samples_x, dst_samples_y] = SamplesLog2(dst_image.info.num_samples);
+
+    const s32 dst_x0 = copy.dst_x0 >> dst_samples_x;
+    const s32 dst_x1 = copy.dst_x1 >> dst_samples_x;
+    const s32 dst_y0 = copy.dst_y0 >> dst_samples_y;
+    const s32 dst_y1 = copy.dst_y1 >> dst_samples_y;
+    const Region2D dst_dimensions{
+        Offset2D{.x = dst_x0, .y = dst_y0},
+        Offset2D{.x = dst_x1 - width_diff, .y = dst_y1},
+    };
+    const auto dst_region = use_override ? *dst_override : dst_dimensions;
 
     // Always call this after src_framebuffer_id was queried, as the address might be invalidated.
     Framebuffer* const dst_framebuffer = &slot_framebuffers[dst_framebuffer_id];
@@ -804,6 +830,21 @@ void TextureCache<P>::BlitImage(const Tegra::Engines::Fermi2D::Surface& dst,
         runtime.BlitImage(dst_framebuffer, dst_view, src_view, dst_region, src_region, copy.filter,
                           copy.operation);
     }
+
+    if (width_oob) {
+        // Continue copy of the oob region of the texture on the next row
+        auto oob_src = src;
+        oob_src.height++;
+        const Region2D src_region_override{
+            Offset2D{.x = 0, .y = src_y0 + 1},
+            Offset2D{.x = width_diff, .y = src_y1 + 1},
+        };
+        const Region2D dst_region_override{
+            Offset2D{.x = dst_x1 - width_diff, .y = dst_y0},
+            Offset2D{.x = dst_x1, .y = dst_y1},
+        };
+        BlitImage(dst, oob_src, copy, src_region_override, dst_region_override);
+    }
 }
 
 template <class P>

src/video_core/texture_cache/util.cpp

@@ -394,7 +394,7 @@ template <u32 GOB_EXTENT>
     const s32 mip_offset = diff % layer_stride;
     const std::array offsets = CalculateMipLevelOffsets(new_info);
     const auto end = offsets.begin() + new_info.resources.levels;
-    const auto it = std::find(offsets.begin(), end, mip_offset);
+    const auto it = std::find(offsets.begin(), end, static_cast<u32>(mip_offset));
     if (it == end) {
         // Mipmap is not aligned to any valid size
         return std::nullopt;

src/video_core/textures/astc.cpp

@@ -1369,8 +1369,8 @@ static void DecompressBlock(std::span<const u8, 16> inBuf, const u32 blockWidth,
     // each partition.
 
     // Determine partitions, partition index, and color endpoint modes
-    s32 planeIdx = -1;
+    u32 planeIdx{UINT32_MAX};
-    u32 partitionIndex;
+    u32 partitionIndex{};
     u32 colorEndpointMode[4] = {0, 0, 0, 0};
 
     // Define color data.