Change fullscreen logic to attempt desktop resolution first on yuzu-cmd

Changed the order in which we attempt to switch to fullscreen. First try desktop resolution first, if it fails fall back to streched fullscreen using windowed resolution.

CMakeLists.txt

@@ -13,7 +13,7 @@ project(yuzu)
 option(ENABLE_SDL2 "Enable the SDL2 frontend" ON)
 CMAKE_DEPENDENT_OPTION(YUZU_USE_BUNDLED_SDL2 "Download bundled SDL2 binaries" ON "ENABLE_SDL2;MSVC" OFF)
 # On Linux system SDL2 is likely to be lacking HIDAPI support which have drawbacks but is needed for SDL motion
-option(YUZU_ALLOW_SYSTEM_SDL2 "Try using system SDL2 before fallling back to one from externals" OFF)
+CMAKE_DEPENDENT_OPTION(YUZU_USE_EXTERNAL_SDL2 "Compiled external SDL2" ON "ENABLE_SDL2;NOT MSVC" OFF)
 
 option(ENABLE_QT "Enable the Qt frontend" ON)
 option(ENABLE_QT_TRANSLATION "Enable translations for the Qt frontend" OFF)
@@ -47,9 +47,10 @@ if (NOT IS_MULTI_CONFIG AND NOT CMAKE_BUILD_TYPE)
 endif()
 
 if(EXISTS ${PROJECT_SOURCE_DIR}/hooks/pre-commit AND NOT EXISTS ${PROJECT_SOURCE_DIR}/.git/hooks/pre-commit)
-    message(STATUS "Copying pre-commit hook")
-    file(COPY hooks/pre-commit
-        DESTINATION ${PROJECT_SOURCE_DIR}/.git/hooks)
+    if (EXISTS ${PROJECT_SOURCE_DIR}/.git/)
+        message(STATUS "Copying pre-commit hook")
+        file(COPY hooks/pre-commit DESTINATION ${PROJECT_SOURCE_DIR}/.git/hooks)
+    endif()
 endif()
 
 # Sanity check : Check that all submodules are present
@@ -386,26 +387,20 @@ if (ENABLE_SDL2)
         add_library(SDL2 INTERFACE)
         target_link_libraries(SDL2 INTERFACE "${SDL2_LIBRARY}")
         target_include_directories(SDL2 INTERFACE "${SDL2_INCLUDE_DIR}")
+    elseif (YUZU_USE_EXTERNAL_SDL2)
+        message(STATUS "Using SDL2 from externals.")
     else()
-        if (YUZU_ALLOW_SYSTEM_SDL2)
-            find_package(SDL2 2.0.15 QUIET)
+        find_package(SDL2 2.0.15 REQUIRED)
 
-            if (SDL2_FOUND)
-                # Some installations don't set SDL2_LIBRARIES
-                if("${SDL2_LIBRARIES}" STREQUAL "")
-                    message(WARNING "SDL2_LIBRARIES wasn't set, manually setting to SDL2::SDL2")
-                    set(SDL2_LIBRARIES "SDL2::SDL2")
-                endif()
-
-                include_directories(SYSTEM ${SDL2_INCLUDE_DIRS})
-                add_library(SDL2 INTERFACE)
-                target_link_libraries(SDL2 INTERFACE "${SDL2_LIBRARIES}")
-            else()
-                message(STATUS "SDL2 2.0.15 or newer not found, falling back to externals.")
-            endif()
-        else()
-            message(STATUS "Using SDL2 from externals.")
+        # Some installations don't set SDL2_LIBRARIES
+        if("${SDL2_LIBRARIES}" STREQUAL "")
+            message(WARNING "SDL2_LIBRARIES wasn't set, manually setting to SDL2::SDL2")
+            set(SDL2_LIBRARIES "SDL2::SDL2")
         endif()
+
+        include_directories(SYSTEM ${SDL2_INCLUDE_DIRS})
+        add_library(SDL2 INTERFACE)
+        target_link_libraries(SDL2 INTERFACE "${SDL2_LIBRARIES}")
     endif()
 endif()
 

externals/CMakeLists.txt

@@ -51,7 +51,7 @@ if (NOT LIBUSB_FOUND OR YUZU_USE_BUNDLED_LIBUSB)
 endif()
 
 # SDL2
-if (NOT SDL2_FOUND AND ENABLE_SDL2)
+if (YUZU_USE_EXTERNAL_SDL2)
     if (NOT WIN32)
         # Yuzu itself needs: Events Joystick Haptic Sensor Timers Audio
         # Yuzu-cmd also needs: Video (depends on Loadso/Dlopen)

src/audio_core/audio_out.cpp

@@ -30,7 +30,8 @@ StreamPtr AudioOut::OpenStream(Core::Timing::CoreTiming& core_timing, u32 sample
                                u32 num_channels, std::string&& name,
                                Stream::ReleaseCallback&& release_callback) {
     if (!sink) {
-        sink = CreateSinkFromID(Settings::values.sink_id, Settings::values.audio_device_id);
+        sink = CreateSinkFromID(Settings::values.sink_id.GetValue(),
+                                Settings::values.audio_device_id.GetValue());
     }
 
     return std::make_shared<Stream>(

src/audio_core/audio_renderer.cpp

@@ -29,10 +29,9 @@ namespace {
                                        (static_cast<float>(r_channel) * r_mix_amount)));
 }
 
-[[nodiscard]] static constexpr std::tuple<s16, s16> Mix6To2(s16 fl_channel, s16 fr_channel,
-                                                            s16 fc_channel,
-                                                            [[maybe_unused]] s16 lf_channel,
-                                                            s16 bl_channel, s16 br_channel) {
+[[maybe_unused, nodiscard]] static constexpr std::tuple<s16, s16> Mix6To2(
+    s16 fl_channel, s16 fr_channel, s16 fc_channel, [[maybe_unused]] s16 lf_channel, s16 bl_channel,
+    s16 br_channel) {
     // Front channels are mixed 36.94%, Center channels are mixed to be 26.12% & the back channels
     // are mixed to be 36.94%
 
@@ -57,11 +56,11 @@ namespace {
     const std::array<float_le, 4>& coeff) {
     const auto left =
         static_cast<float>(fl_channel) * coeff[0] + static_cast<float>(fc_channel) * coeff[1] +
-        static_cast<float>(lf_channel) * coeff[2] + static_cast<float>(bl_channel) * coeff[0];
+        static_cast<float>(lf_channel) * coeff[2] + static_cast<float>(bl_channel) * coeff[3];
 
     const auto right =
         static_cast<float>(fr_channel) * coeff[0] + static_cast<float>(fc_channel) * coeff[1] +
-        static_cast<float>(lf_channel) * coeff[2] + static_cast<float>(br_channel) * coeff[0];
+        static_cast<float>(lf_channel) * coeff[2] + static_cast<float>(br_channel) * coeff[3];
 
     return {ClampToS16(static_cast<s32>(left)), ClampToS16(static_cast<s32>(right))};
 }
@@ -241,7 +240,7 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) {
         const auto channel_count = buffer_offsets.size();
         const auto& final_mix = mix_context.GetFinalMixInfo();
         const auto& in_params = final_mix.GetInParams();
-        std::vector<s32*> mix_buffers(channel_count);
+        std::vector<std::span<s32>> mix_buffers(channel_count);
         for (std::size_t i = 0; i < channel_count; i++) {
             mix_buffers[i] =
                 command_generator.GetMixBuffer(in_params.buffer_offset + buffer_offsets[i]);
@@ -294,18 +293,11 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) {
                     buffer[i * stream_channel_count + 0] = Mix2To1(fl_sample, fr_sample);
                 } else if (stream_channel_count == 2) {
                     // Mix all channels into 2 channels
-                    if (sink_context.HasDownMixingCoefficients()) {
-                        const auto [left, right] = Mix6To2WithCoefficients(
-                            fl_sample, fr_sample, fc_sample, lf_sample, bl_sample, br_sample,
-                            sink_context.GetDownmixCoefficients());
-                        buffer[i * stream_channel_count + 0] = left;
-                        buffer[i * stream_channel_count + 1] = right;
-                    } else {
-                        const auto [left, right] = Mix6To2(fl_sample, fr_sample, fc_sample,
-                                                           lf_sample, bl_sample, br_sample);
-                        buffer[i * stream_channel_count + 0] = left;
-                        buffer[i * stream_channel_count + 1] = right;
-                    }
+                    const auto [left, right] = Mix6To2WithCoefficients(
+                        fl_sample, fr_sample, fc_sample, lf_sample, bl_sample, br_sample,
+                        sink_context.GetDownmixCoefficients());
+                    buffer[i * stream_channel_count + 0] = left;
+                    buffer[i * stream_channel_count + 1] = right;
                 } else if (stream_channel_count == 6) {
                     // Pass through
                     buffer[i * stream_channel_count + 0] = fl_sample;

src/audio_core/command_generator.cpp

@@ -31,7 +31,7 @@ constexpr std::array<f32, AudioCommon::I3DL2REVERB_TAPS> EARLY_GAIN{
     0.72867f, 0.69794f, 0.5464f,  0.24563f, 0.45214f, 0.44042f};
 
 template <std::size_t N>
-void ApplyMix(s32* output, const s32* input, s32 gain, s32 sample_count) {
+void ApplyMix(std::span<s32> output, std::span<const s32> input, s32 gain, s32 sample_count) {
     for (std::size_t i = 0; i < static_cast<std::size_t>(sample_count); i += N) {
         for (std::size_t j = 0; j < N; j++) {
             output[i + j] +=
@@ -40,7 +40,17 @@ void ApplyMix(s32* output, const s32* input, s32 gain, s32 sample_count) {
     }
 }
 
-s32 ApplyMixRamp(s32* output, const s32* input, float gain, float delta, s32 sample_count) {
+s32 ApplyMixRamp(std::span<s32> output, std::span<const s32> input, float gain, float delta,
+                 s32 sample_count) {
+    // XC2 passes in NaN mix volumes, causing further issues as we handle everything as s32 rather
+    // than float, so the NaN propogation is lost. As the samples get further modified for
+    // volume etc, they can get out of NaN range, so a later heuristic for catching this is
+    // more difficult. Handle it here by setting these samples to silence.
+    if (std::isnan(gain)) {
+        gain = 0.0f;
+        delta = 0.0f;
+    }
+
     s32 x = 0;
     for (s32 i = 0; i < sample_count; i++) {
         x = static_cast<s32>(static_cast<float>(input[i]) * gain);
@@ -50,20 +60,22 @@ s32 ApplyMixRamp(s32* output, const s32* input, float gain, float delta, s32 sam
     return x;
 }
 
-void ApplyGain(s32* output, const s32* input, s32 gain, s32 delta, s32 sample_count) {
+void ApplyGain(std::span<s32> output, std::span<const s32> input, s32 gain, s32 delta,
+               s32 sample_count) {
     for (s32 i = 0; i < sample_count; i++) {
         output[i] = static_cast<s32>((static_cast<s64>(input[i]) * gain + 0x4000) >> 15);
         gain += delta;
     }
 }
 
-void ApplyGainWithoutDelta(s32* output, const s32* input, s32 gain, s32 sample_count) {
+void ApplyGainWithoutDelta(std::span<s32> output, std::span<const s32> input, s32 gain,
+                           s32 sample_count) {
     for (s32 i = 0; i < sample_count; i++) {
         output[i] = static_cast<s32>((static_cast<s64>(input[i]) * gain + 0x4000) >> 15);
     }
 }
 
-s32 ApplyMixDepop(s32* output, s32 first_sample, s32 delta, s32 sample_count) {
+s32 ApplyMixDepop(std::span<s32> output, s32 first_sample, s32 delta, s32 sample_count) {
     const bool positive = first_sample > 0;
     auto final_sample = std::abs(first_sample);
     for (s32 i = 0; i < sample_count; i++) {
@@ -128,10 +140,10 @@ constexpr std::array<std::size_t, 20> REVERB_TAP_INDEX_6CH{4, 0, 0, 1, 1, 1, 1,
                                                            1, 1, 1, 0, 0, 0, 0, 3, 3, 3};
 
 template <std::size_t CHANNEL_COUNT>
-void ApplyReverbGeneric(I3dl2ReverbState& state,
-                        const std::array<const s32*, AudioCommon::MAX_CHANNEL_COUNT>& input,
-                        const std::array<s32*, AudioCommon::MAX_CHANNEL_COUNT>& output,
-                        s32 sample_count) {
+void ApplyReverbGeneric(
+    I3dl2ReverbState& state,
+    const std::array<std::span<const s32>, AudioCommon::MAX_CHANNEL_COUNT>& input,
+    const std::array<std::span<s32>, AudioCommon::MAX_CHANNEL_COUNT>& output, s32 sample_count) {
 
     auto GetTapLookup = []() {
         if constexpr (CHANNEL_COUNT == 1) {
@@ -400,7 +412,10 @@ 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);
@@ -454,8 +469,8 @@ void CommandGenerator::GenerateBiquadFilterCommand([[maybe_unused]] s32 mix_buff
                   "input_mix_buffer={}, output_mix_buffer={}",
                   node_id, input_offset, output_offset);
     }
-    const auto* input = GetMixBuffer(input_offset);
-    auto* output = GetMixBuffer(output_offset);
+    std::span<const s32> input = GetMixBuffer(input_offset);
+    std::span<s32> output = GetMixBuffer(output_offset);
 
     // Biquad filter parameters
     const auto [n0, n1, n2] = params.numerator;
@@ -548,8 +563,8 @@ void CommandGenerator::GenerateI3dl2ReverbEffectCommand(s32 mix_buffer_offset, E
         return;
     }
 
-    std::array<const s32*, AudioCommon::MAX_CHANNEL_COUNT> input{};
-    std::array<s32*, AudioCommon::MAX_CHANNEL_COUNT> output{};
+    std::array<std::span<const s32>, AudioCommon::MAX_CHANNEL_COUNT> input{};
+    std::array<std::span<s32>, AudioCommon::MAX_CHANNEL_COUNT> output{};
 
     const auto status = params.status;
     for (s32 i = 0; i < channel_count; i++) {
@@ -584,7 +599,8 @@ void CommandGenerator::GenerateI3dl2ReverbEffectCommand(s32 mix_buffer_offset, E
         for (s32 i = 0; i < channel_count; i++) {
             // Only copy if the buffer input and output do not match!
             if ((mix_buffer_offset + params.input[i]) != (mix_buffer_offset + params.output[i])) {
-                std::memcpy(output[i], input[i], worker_params.sample_count * sizeof(s32));
+                std::memcpy(output[i].data(), input[i].data(),
+                            worker_params.sample_count * sizeof(s32));
             }
         }
     }
@@ -600,8 +616,8 @@ void CommandGenerator::GenerateBiquadFilterEffectCommand(s32 mix_buffer_offset,
     for (s32 i = 0; i < channel_count; i++) {
         // TODO(ogniK): Actually implement biquad filter
         if (params.input[i] != params.output[i]) {
-            const auto* input = GetMixBuffer(mix_buffer_offset + params.input[i]);
-            auto* output = GetMixBuffer(mix_buffer_offset + params.output[i]);
+            std::span<const s32> input = GetMixBuffer(mix_buffer_offset + params.input[i]);
+            std::span<s32> output = GetMixBuffer(mix_buffer_offset + params.output[i]);
             ApplyMix<1>(output, input, 32768, worker_params.sample_count);
         }
     }
@@ -640,14 +656,15 @@ void CommandGenerator::GenerateAuxCommand(s32 mix_buffer_offset, EffectBase* inf
 
                 if (samples_read != static_cast<int>(worker_params.sample_count) &&
                     samples_read <= params.sample_count) {
-                    std::memset(GetMixBuffer(output_index), 0, params.sample_count - samples_read);
+                    std::memset(GetMixBuffer(output_index).data(), 0,
+                                params.sample_count - samples_read);
                 }
             } else {
                 AuxInfoDSP empty{};
                 memory.WriteBlock(aux->GetSendInfo(), &empty, sizeof(AuxInfoDSP));
                 memory.WriteBlock(aux->GetRecvInfo(), &empty, sizeof(AuxInfoDSP));
                 if (output_index != input_index) {
-                    std::memcpy(GetMixBuffer(output_index), GetMixBuffer(input_index),
+                    std::memcpy(GetMixBuffer(output_index).data(), GetMixBuffer(input_index).data(),
                                 worker_params.sample_count * sizeof(s32));
                 }
             }
@@ -665,7 +682,7 @@ ServerSplitterDestinationData* CommandGenerator::GetDestinationData(s32 splitter
 }
 
 s32 CommandGenerator::WriteAuxBuffer(AuxInfoDSP& dsp_info, VAddr send_buffer, u32 max_samples,
-                                     const s32* data, u32 sample_count, u32 write_offset,
+                                     std::span<const s32> data, u32 sample_count, u32 write_offset,
                                      u32 write_count) {
     if (max_samples == 0) {
         return 0;
@@ -675,14 +692,14 @@ s32 CommandGenerator::WriteAuxBuffer(AuxInfoDSP& dsp_info, VAddr send_buffer, u3
         return 0;
     }
 
-    std::size_t data_offset{};
+    s32 data_offset{};
     u32 remaining = sample_count;
     while (remaining > 0) {
         // Get position in buffer
         const auto base = send_buffer + (offset * sizeof(u32));
         const auto samples_to_grab = std::min(max_samples - offset, remaining);
         // Write to output
-        memory.WriteBlock(base, (data + data_offset), samples_to_grab * sizeof(u32));
+        memory.WriteBlock(base, (data.data() + data_offset), samples_to_grab * sizeof(u32));
         offset = (offset + samples_to_grab) % max_samples;
         remaining -= samples_to_grab;
         data_offset += samples_to_grab;
@@ -695,7 +712,7 @@ s32 CommandGenerator::WriteAuxBuffer(AuxInfoDSP& dsp_info, VAddr send_buffer, u3
 }
 
 s32 CommandGenerator::ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u32 max_samples,
-                                    s32* out_data, u32 sample_count, u32 read_offset,
+                                    std::span<s32> out_data, u32 sample_count, u32 read_offset,
                                     u32 read_count) {
     if (max_samples == 0) {
         return 0;
@@ -707,15 +724,16 @@ s32 CommandGenerator::ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u3
     }
 
     u32 remaining = sample_count;
+    s32 data_offset{};
     while (remaining > 0) {
         const auto base = recv_buffer + (offset * sizeof(u32));
         const auto samples_to_grab = std::min(max_samples - offset, remaining);
         std::vector<s32> buffer(samples_to_grab);
         memory.ReadBlock(base, buffer.data(), buffer.size() * sizeof(u32));
-        std::memcpy(out_data, buffer.data(), buffer.size() * sizeof(u32));
-        out_data += samples_to_grab;
+        std::memcpy(out_data.data() + data_offset, buffer.data(), buffer.size() * sizeof(u32));
         offset = (offset + samples_to_grab) % max_samples;
         remaining -= samples_to_grab;
+        data_offset += samples_to_grab;
     }
 
     if (read_count != 0) {
@@ -962,8 +980,8 @@ void CommandGenerator::GenerateMixCommand(std::size_t output_offset, std::size_t
                   node_id, input_offset, output_offset, volume);
     }
 
-    auto* output = GetMixBuffer(output_offset);
-    const auto* input = GetMixBuffer(input_offset);
+    std::span<s32> output = GetMixBuffer(output_offset);
+    std::span<const s32> input = GetMixBuffer(input_offset);
 
     const s32 gain = static_cast<s32>(volume * 32768.0f);
     // Mix with loop unrolling
@@ -1003,9 +1021,10 @@ void CommandGenerator::GenerateFinalMixCommand() {
     }
 }
 
-s32 CommandGenerator::DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_state,
-                                  s32 sample_start_offset, s32 sample_end_offset, s32 sample_count,
-                                  s32 channel, std::size_t mix_offset) {
+template <typename T>
+s32 CommandGenerator::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) {
     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) {
@@ -1019,24 +1038,37 @@ 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) {
-        std::vector<s16> buffer(samples_processed);
-        memory.ReadBlock(buffer_pos, buffer.data(), buffer.size() * sizeof(s16));
-        for (std::size_t i = 0; i < buffer.size(); i++) {
-            sample_buffer[mix_offset + i] = buffer[i];
-        }
-    } 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));
+    const auto channel_count = in_params.channel_count;
+    std::vector<T> buffer(samples_processed * channel_count);
+    memory.ReadBlock(buffer_pos, buffer.data(), buffer.size() * sizeof(T));
 
+    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 if constexpr (sizeof(T) == 1) {
+        for (std::size_t i = 0; i < static_cast<std::size_t>(samples_processed); i++) {
+            sample_buffer[mix_offset + i] =
+                static_cast<s32>(static_cast<f32>(buffer[i * channel_count + channel] /
+                                                  std::numeric_limits<s8>::max()) *
+                                 std::numeric_limits<s16>::max());
+        }
+    } else if constexpr (sizeof(T) == 2) {
         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];
         }
+    } else {
+        for (std::size_t i = 0; i < static_cast<std::size_t>(samples_processed); i++) {
+            sample_buffer[mix_offset + i] =
+                static_cast<s32>(static_cast<f32>(buffer[i * channel_count + channel] /
+                                                  std::numeric_limits<s32>::max()) *
+                                 std::numeric_limits<s16>::max());
+        }
     }
 
     return samples_processed;
@@ -1155,12 +1187,14 @@ s32 CommandGenerator::DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_s
     return samples_processed;
 }
 
-s32* CommandGenerator::GetMixBuffer(std::size_t index) {
-    return mix_buffer.data() + (index * worker_params.sample_count);
+std::span<s32> CommandGenerator::GetMixBuffer(std::size_t index) {
+    return std::span<s32>(mix_buffer.data() + (index * worker_params.sample_count),
+                          worker_params.sample_count);
 }
 
-const s32* CommandGenerator::GetMixBuffer(std::size_t index) const {
-    return mix_buffer.data() + (index * worker_params.sample_count);
+std::span<const s32> CommandGenerator::GetMixBuffer(std::size_t index) const {
+    return std::span<const s32>(mix_buffer.data() + (index * worker_params.sample_count),
+                                worker_params.sample_count);
 }
 
 std::size_t CommandGenerator::GetMixChannelBufferOffset(s32 channel) const {
@@ -1171,15 +1205,15 @@ std::size_t CommandGenerator::GetTotalMixBufferCount() const {
     return worker_params.mix_buffer_count + AudioCommon::MAX_CHANNEL_COUNT;
 }
 
-s32* CommandGenerator::GetChannelMixBuffer(s32 channel) {
+std::span<s32> CommandGenerator::GetChannelMixBuffer(s32 channel) {
     return GetMixBuffer(worker_params.mix_buffer_count + channel);
 }
 
-const s32* CommandGenerator::GetChannelMixBuffer(s32 channel) const {
+std::span<const s32> CommandGenerator::GetChannelMixBuffer(s32 channel) const {
     return GetMixBuffer(worker_params.mix_buffer_count + channel);
 }
 
-void CommandGenerator::DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* output,
+void CommandGenerator::DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, std::span<s32> output,
                                              VoiceState& dsp_state, s32 channel,
                                              s32 target_sample_rate, s32 sample_count,
                                              s32 node_id) {
@@ -1191,7 +1225,7 @@ void CommandGenerator::DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* o
                   node_id, channel, in_params.sample_format, sample_count, in_params.sample_rate,
                   in_params.mix_id, in_params.splitter_info_id);
     }
-    ASSERT_OR_EXECUTE(output != nullptr, { return; });
+    ASSERT_OR_EXECUTE(output.data() != nullptr, { return; });
 
     const auto resample_rate = static_cast<s32>(
         static_cast<float>(in_params.sample_rate) / static_cast<float>(target_sample_rate) *
@@ -1208,6 +1242,7 @@ void CommandGenerator::DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* o
     }
 
     std::size_t temp_mix_offset{};
+    s32 samples_output{};
     auto samples_remaining = sample_count;
     while (samples_remaining > 0) {
         const auto samples_to_output = std::min(samples_remaining, min_required_samples);
@@ -1249,10 +1284,25 @@ 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,
-                                samples_to_read - samples_read, channel, temp_mix_offset);
+                    DecodePcm<s16>(voice_info, dsp_state, samples_offset_start, samples_offset_end,
+                                   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 =
@@ -1296,20 +1346,21 @@ void CommandGenerator::DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* o
 
         if (in_params.behavior_flags.is_pitch_and_src_skipped.Value()) {
             // No need to resample
-            std::memcpy(output, sample_buffer.data(), samples_read * sizeof(s32));
+            std::memcpy(output.data() + samples_output, sample_buffer.data(),
+                        samples_read * sizeof(s32));
         } else {
             std::fill(sample_buffer.begin() + temp_mix_offset,
                       sample_buffer.begin() + temp_mix_offset + (samples_to_read - samples_read),
                       0);
-            AudioCore::Resample(output, sample_buffer.data(), resample_rate, dsp_state.fraction,
-                                samples_to_output);
+            AudioCore::Resample(output.data() + samples_output, sample_buffer.data(), resample_rate,
+                                dsp_state.fraction, samples_to_output);
             // Resample
             for (std::size_t i = 0; i < AudioCommon::MAX_SAMPLE_HISTORY; i++) {
                 dsp_state.sample_history[i] = sample_buffer[samples_to_read + i];
             }
         }
-        output += samples_to_output;
         samples_remaining -= samples_to_output;
+        samples_output += samples_to_output;
     }
 }
 

src/audio_core/command_generator.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <array>
+#include <span>
 #include "audio_core/common.h"
 #include "audio_core/voice_context.h"
 #include "common/common_types.h"
@@ -41,10 +42,10 @@ public:
     void PreCommand();
     void PostCommand();
 
-    [[nodiscard]] s32* GetChannelMixBuffer(s32 channel);
-    [[nodiscard]] const s32* GetChannelMixBuffer(s32 channel) const;
-    [[nodiscard]] s32* GetMixBuffer(std::size_t index);
-    [[nodiscard]] const s32* GetMixBuffer(std::size_t index) const;
+    [[nodiscard]] std::span<s32> GetChannelMixBuffer(s32 channel);
+    [[nodiscard]] std::span<const s32> GetChannelMixBuffer(s32 channel) const;
+    [[nodiscard]] std::span<s32> GetMixBuffer(std::size_t index);
+    [[nodiscard]] std::span<const s32> GetMixBuffer(std::size_t index) const;
     [[nodiscard]] std::size_t GetMixChannelBufferOffset(s32 channel) const;
 
     [[nodiscard]] std::size_t GetTotalMixBufferCount() const;
@@ -77,21 +78,24 @@ private:
     void GenerateAuxCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled);
     [[nodiscard]] ServerSplitterDestinationData* GetDestinationData(s32 splitter_id, s32 index);
 
-    s32 WriteAuxBuffer(AuxInfoDSP& dsp_info, VAddr send_buffer, u32 max_samples, const s32* data,
-                       u32 sample_count, u32 write_offset, u32 write_count);
-    s32 ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u32 max_samples, s32* out_data,
-                      u32 sample_count, u32 read_offset, u32 read_count);
+    s32 WriteAuxBuffer(AuxInfoDSP& dsp_info, VAddr send_buffer, u32 max_samples,
+                       std::span<const s32> data, u32 sample_count, u32 write_offset,
+                       u32 write_count);
+    s32 ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u32 max_samples,
+                      std::span<s32> out_data, u32 sample_count, u32 read_offset, u32 read_count);
 
     void InitializeI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state,
                                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,
-                    s32 sample_end_offset, s32 sample_count, s32 channel, std::size_t mix_offset);
+    template <typename T>
+    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,
-                               s32 channel, s32 target_sample_rate, s32 sample_count, s32 node_id);
+    void DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, std::span<s32> output,
+                               VoiceState& dsp_state, s32 channel, s32 target_sample_rate,
+                               s32 sample_count, s32 node_id);
 
     AudioCommon::AudioRendererParameter& worker_params;
     VoiceContext& voice_context;

src/audio_core/sink_context.cpp

@@ -15,10 +15,17 @@ std::size_t SinkContext::GetCount() const {
 void SinkContext::UpdateMainSink(const SinkInfo::InParams& in) {
     ASSERT(in.type == SinkTypes::Device);
 
-    has_downmix_coefs = in.device.down_matrix_enabled;
-    if (has_downmix_coefs) {
+    if (in.device.down_matrix_enabled) {
         downmix_coefficients = in.device.down_matrix_coef;
+    } else {
+        downmix_coefficients = {
+            1.0f,   // front
+            0.707f, // center
+            0.0f,   // lfe
+            0.707f, // back
+        };
     }
+
     in_use = in.in_use;
     use_count = in.device.input_count;
     buffers = in.device.input;
@@ -34,10 +41,6 @@ std::vector<u8> SinkContext::OutputBuffers() const {
     return buffer_ret;
 }
 
-bool SinkContext::HasDownMixingCoefficients() const {
-    return has_downmix_coefs;
-}
-
 const DownmixCoefficients& SinkContext::GetDownmixCoefficients() const {
     return downmix_coefficients;
 }

src/audio_core/sink_context.h

@@ -84,7 +84,6 @@ public:
     [[nodiscard]] bool InUse() const;
     [[nodiscard]] std::vector<u8> OutputBuffers() const;
 
-    [[nodiscard]] bool HasDownMixingCoefficients() const;
     [[nodiscard]] const DownmixCoefficients& GetDownmixCoefficients() const;
 
 private:
@@ -92,7 +91,6 @@ private:
     s32 use_count{};
     std::array<u8, AudioCommon::MAX_CHANNEL_COUNT> buffers{};
     std::size_t sink_count{};
-    bool has_downmix_coefs{false};
     DownmixCoefficients downmix_coefficients{};
 };
 } // namespace AudioCore

src/common/CMakeLists.txt

@@ -180,7 +180,6 @@ add_library(common STATIC
     thread.cpp
     thread.h
     thread_queue_list.h
-    thread_worker.cpp
     thread_worker.h
     threadsafe_queue.h
     time_zone.cpp
@@ -188,6 +187,7 @@ add_library(common STATIC
     tiny_mt.h
     tree.h
     uint128.h
+    unique_function.h
     uuid.cpp
     uuid.h
     vector_math.h

src/common/fs/file.cpp

@@ -306,9 +306,9 @@ bool IOFile::Flush() const {
     errno = 0;
 
 #ifdef _WIN32
-    const auto flush_result = std::fflush(file) == 0 && _commit(fileno(file)) == 0;
+    const auto flush_result = std::fflush(file) == 0;
 #else
-    const auto flush_result = std::fflush(file) == 0 && fsync(fileno(file)) == 0;
+    const auto flush_result = std::fflush(file) == 0;
 #endif
 
     if (!flush_result) {
@@ -320,6 +320,28 @@ bool IOFile::Flush() const {
     return flush_result;
 }
 
+bool IOFile::Commit() const {
+    if (!IsOpen()) {
+        return false;
+    }
+
+    errno = 0;
+
+#ifdef _WIN32
+    const auto commit_result = std::fflush(file) == 0 && _commit(fileno(file)) == 0;
+#else
+    const auto commit_result = std::fflush(file) == 0 && fsync(fileno(file)) == 0;
+#endif
+
+    if (!commit_result) {
+        const auto ec = std::error_code{errno, std::generic_category()};
+        LOG_ERROR(Common_Filesystem, "Failed to commit the file at path={}, ec_message={}",
+                  PathToUTF8String(file_path), ec.message());
+    }
+
+    return commit_result;
+}
+
 bool IOFile::SetSize(u64 size) const {
     if (!IsOpen()) {
         return false;
@@ -347,6 +369,9 @@ u64 IOFile::GetSize() const {
         return 0;
     }
 
+    // Flush any unwritten buffered data into the file prior to retrieving the file size.
+    std::fflush(file);
+
     std::error_code ec;
 
     const auto file_size = fs::file_size(file_path, ec);

src/common/fs/file.h

@@ -396,12 +396,21 @@ public:
     [[nodiscard]] size_t WriteString(std::span<const char> string) const;
 
     /**
-     * Attempts to flush any unwritten buffered data into the file and flush the file into the disk.
+     * Attempts to flush any unwritten buffered data into the file.
      *
      * @returns True if the flush was successful, false otherwise.
      */
     bool Flush() const;
 
+    /**
+     * Attempts to commit the file into the disk.
+     * Note that this is an expensive operation as this forces the operating system to write
+     * the contents of the file associated with the file descriptor into the disk.
+     *
+     * @returns True if the commit was successful, false otherwise.
+     */
+    bool Commit() const;
+
     /**
      * Resizes the file to a given size.
      * If the file is resized to a smaller size, the remainder of the file is discarded.

src/common/logging/backend.cpp

@@ -171,19 +171,22 @@ FileBackend::FileBackend(const std::filesystem::path& filename) {
 FileBackend::~FileBackend() = default;
 
 void FileBackend::Write(const Entry& entry) {
-    using namespace Common::Literals;
-    // prevent logs from going over the maximum size (in case its spamming and the user doesn't
-    // know)
-    constexpr std::size_t MAX_BYTES_WRITTEN = 100_MiB;
-    constexpr std::size_t MAX_BYTES_WRITTEN_EXTENDED = 1_GiB;
-
     if (!file->IsOpen()) {
         return;
     }
 
-    if (Settings::values.extended_logging && bytes_written > MAX_BYTES_WRITTEN_EXTENDED) {
-        return;
-    } else if (!Settings::values.extended_logging && bytes_written > MAX_BYTES_WRITTEN) {
+    using namespace Common::Literals;
+    // Prevent logs from exceeding a set maximum size in the event that log entries are spammed.
+    constexpr std::size_t MAX_BYTES_WRITTEN = 100_MiB;
+    constexpr std::size_t MAX_BYTES_WRITTEN_EXTENDED = 1_GiB;
+
+    const bool write_limit_exceeded =
+        bytes_written > MAX_BYTES_WRITTEN_EXTENDED ||
+        (bytes_written > MAX_BYTES_WRITTEN && !Settings::values.extended_logging);
+
+    // Close the file after the write limit is exceeded.
+    if (write_limit_exceeded) {
+        file->Close();
         return;
     }
 

src/common/settings.cpp

@@ -41,7 +41,7 @@ void LogSettings() {
     LOG_INFO(Config, "yuzu Configuration:");
     log_setting("Controls_UseDockedMode", values.use_docked_mode.GetValue());
     log_setting("System_RngSeed", values.rng_seed.GetValue().value_or(0));
-    log_setting("System_CurrentUser", values.current_user);
+    log_setting("System_CurrentUser", values.current_user.GetValue());
     log_setting("System_LanguageIndex", values.language_index.GetValue());
     log_setting("System_RegionIndex", values.region_index.GetValue());
     log_setting("System_TimeZoneIndex", values.time_zone_index.GetValue());
@@ -61,18 +61,18 @@ void LogSettings() {
     log_setting("Renderer_UseAsynchronousShaders", values.use_asynchronous_shaders.GetValue());
     log_setting("Renderer_UseGarbageCollection", values.use_caches_gc.GetValue());
     log_setting("Renderer_AnisotropicFilteringLevel", values.max_anisotropy.GetValue());
-    log_setting("Audio_OutputEngine", values.sink_id);
+    log_setting("Audio_OutputEngine", values.sink_id.GetValue());
     log_setting("Audio_EnableAudioStretching", values.enable_audio_stretching.GetValue());
-    log_setting("Audio_OutputDevice", values.audio_device_id);
-    log_setting("DataStorage_UseVirtualSd", values.use_virtual_sd);
+    log_setting("Audio_OutputDevice", values.audio_device_id.GetValue());
+    log_setting("DataStorage_UseVirtualSd", values.use_virtual_sd.GetValue());
     log_path("DataStorage_CacheDir", Common::FS::GetYuzuPath(Common::FS::YuzuPath::CacheDir));
     log_path("DataStorage_ConfigDir", Common::FS::GetYuzuPath(Common::FS::YuzuPath::ConfigDir));
     log_path("DataStorage_LoadDir", Common::FS::GetYuzuPath(Common::FS::YuzuPath::LoadDir));
     log_path("DataStorage_NANDDir", Common::FS::GetYuzuPath(Common::FS::YuzuPath::NANDDir));
     log_path("DataStorage_SDMCDir", Common::FS::GetYuzuPath(Common::FS::YuzuPath::SDMCDir));
-    log_setting("Debugging_ProgramArgs", values.program_args);
-    log_setting("Services_BCATBackend", values.bcat_backend);
-    log_setting("Services_BCATBoxcatLocal", values.bcat_boxcat_local);
+    log_setting("Debugging_ProgramArgs", values.program_args.GetValue());
+    log_setting("Services_BCATBackend", values.bcat_backend.GetValue());
+    log_setting("Services_BCATBoxcatLocal", values.bcat_boxcat_local.GetValue());
 }
 
 bool IsConfiguringGlobal() {
@@ -93,8 +93,8 @@ bool IsGPULevelHigh() {
 }
 
 bool IsFastmemEnabled() {
-    if (values.cpu_accuracy.GetValue() == CPUAccuracy::DebugMode) {
-        return values.cpuopt_fastmem;
+    if (values.cpu_debug_mode) {
+        return static_cast<bool>(values.cpuopt_fastmem);
     }
     return true;
 }
@@ -103,7 +103,7 @@ float Volume() {
     if (values.audio_muted) {
         return 0.0f;
     }
-    return values.volume.GetValue();
+    return values.volume.GetValue() / 100.0f;
 }
 
 void RestoreGlobalState(bool is_powered_on) {

src/common/settings.h

@@ -10,10 +10,12 @@
 #include <map>
 #include <optional>
 #include <string>
+#include <utility>
 #include <vector>
 
 #include "common/common_types.h"
 #include "common/settings_input.h"
+#include "input_common/udp/client.h"
 
 namespace Settings {
 
@@ -29,73 +31,240 @@ enum class GPUAccuracy : u32 {
 };
 
 enum class CPUAccuracy : u32 {
-    Accurate = 0,
-    Unsafe = 1,
-    DebugMode = 2,
+    Auto = 0,
+    Accurate = 1,
+    Unsafe = 2,
 };
 
+/** The BasicSetting class is a simple resource manager. It defines a label and default value
+ * alongside the actual value of the setting for simpler and less-error prone use with frontend
+ * configurations. Setting a default value and label is required, though subclasses may deviate from
+ * this requirement.
+ */
 template <typename Type>
-class Setting final {
+class BasicSetting {
+protected:
+    BasicSetting() = default;
+
+    /**
+     * Only sets the setting to the given initializer, leaving the other members to their default
+     * initializers.
+     *
+     * @param global_val Initial value of the setting
+     */
+    explicit BasicSetting(const Type& global_val) : global{global_val} {}
+
 public:
-    Setting() = default;
-    explicit Setting(Type val) : global{val} {}
-    ~Setting() = default;
-    void SetGlobal(bool to_global) {
-        use_global = to_global;
-    }
-    bool UsingGlobal() const {
-        return use_global;
-    }
-    Type GetValue(bool need_global = false) const {
-        if (use_global || need_global) {
-            return global;
-        }
-        return local;
-    }
-    void SetValue(const Type& value) {
-        if (use_global) {
-            global = value;
-        } else {
-            local = value;
-        }
+    /**
+     * Sets a default value, label, and setting value.
+     *
+     * @param default_val Intial value of the setting, and default value of the setting
+     * @param name Label for the setting
+     */
+    explicit BasicSetting(const Type& default_val, const std::string& name)
+        : default_value{default_val}, global{default_val}, label{name} {}
+    ~BasicSetting() = default;
+
+    /**
+     *  Returns a reference to the setting's value.
+     *
+     * @returns A reference to the setting
+     */
+    [[nodiscard]] const Type& GetValue() const {
+        return global;
     }
 
-private:
-    bool use_global = true;
-    Type global{};
-    Type local{};
+    /**
+     * Sets the setting to the given value.
+     *
+     * @param value The desired value
+     */
+    void SetValue(const Type& value) {
+        Type temp{value};
+        std::swap(global, temp);
+    }
+
+    /**
+     * Returns the value that this setting was created with.
+     *
+     * @returns A reference to the default value
+     */
+    [[nodiscard]] const Type& GetDefault() const {
+        return default_value;
+    }
+
+    /**
+     * Returns the label this setting was created with.
+     *
+     * @returns A reference to the label
+     */
+    [[nodiscard]] const std::string& GetLabel() const {
+        return label;
+    }
+
+    /**
+     * Assigns a value to the setting.
+     *
+     * @param value The desired setting value
+     *
+     * @returns A reference to the setting
+     */
+    const Type& operator=(const Type& value) {
+        Type temp{value};
+        std::swap(global, temp);
+        return global;
+    }
+
+    /**
+     * Returns a reference to the setting.
+     *
+     * @returns A reference to the setting
+     */
+    explicit operator const Type&() const {
+        return global;
+    }
+
+protected:
+    const Type default_value{}; ///< The default value
+    Type global{};              ///< The setting
+    const std::string label{};  ///< The setting's label
 };
 
 /**
- * The InputSetting class allows for getting a reference to either the global or local members.
+ * The Setting class is a slightly more complex version of the BasicSetting class. This adds a
+ * custom setting to switch to when a guest application specifically requires it. The effect is that
+ * other components of the emulator can access the setting's intended value without any need for the
+ * component to ask whether the custom or global setting is needed at the moment.
+ *
+ * By default, the global setting is used.
+ *
+ * Like the BasicSetting, this requires setting a default value and label to use.
+ */
+template <typename Type>
+class Setting final : public BasicSetting<Type> {
+public:
+    /**
+     * Sets a default value, label, and setting value.
+     *
+     * @param default_val Intial value of the setting, and default value of the setting
+     * @param name Label for the setting
+     */
+    explicit Setting(const Type& default_val, const std::string& name)
+        : BasicSetting<Type>(default_val, name) {}
+    ~Setting() = default;
+
+    /**
+     * Tells this setting to represent either the global or custom setting when other member
+     * functions are used.
+     *
+     * @param to_global Whether to use the global or custom setting.
+     */
+    void SetGlobal(bool to_global) {
+        use_global = to_global;
+    }
+
+    /**
+     * Returns whether this setting is using the global setting or not.
+     *
+     * @returns The global state
+     */
+    [[nodiscard]] bool UsingGlobal() const {
+        return use_global;
+    }
+
+    /**
+     * Returns either the global or custom setting depending on the values of this setting's global
+     * state or if the global value was specifically requested.
+     *
+     * @param need_global Request global value regardless of setting's state; defaults to false
+     *
+     * @returns The required value of the setting
+     */
+    [[nodiscard]] const Type& GetValue(bool need_global = false) const {
+        if (use_global || need_global) {
+            return this->global;
+        }
+        return custom;
+    }
+
+    /**
+     * Sets the current setting value depending on the global state.
+     *
+     * @param value The new value
+     */
+    void SetValue(const Type& value) {
+        Type temp{value};
+        if (use_global) {
+            std::swap(this->global, temp);
+        } else {
+            std::swap(custom, temp);
+        }
+    }
+
+    /**
+     * Assigns the current setting value depending on the global state.
+     *
+     * @param value The new value
+     *
+     * @returns A reference to the current setting value
+     */
+    const Type& operator=(const Type& value) {
+        Type temp{value};
+        if (use_global) {
+            std::swap(this->global, temp);
+            return this->global;
+        }
+        std::swap(custom, temp);
+        return custom;
+    }
+
+    /**
+     * Returns the current setting value depending on the global state.
+     *
+     * @returns A reference to the current setting value
+     */
+    explicit operator const Type&() const {
+        if (use_global) {
+            return this->global;
+        }
+        return custom;
+    }
+
+private:
+    bool use_global{true}; ///< The setting's global state
+    Type custom{};         ///< The custom value of the setting
+};
+
+/**
+ * The InputSetting class allows for getting a reference to either the global or custom members.
  * This is required as we cannot easily modify the values of user-defined types within containers
  * using the SetValue() member function found in the Setting class. The primary purpose of this
- * class is to store an array of 10 PlayerInput structs for both the global and local (per-game)
- * setting and allows for easily accessing and modifying both settings.
+ * class is to store an array of 10 PlayerInput structs for both the global and custom setting and
+ * allows for easily accessing and modifying both settings.
  */
 template <typename Type>
 class InputSetting final {
 public:
     InputSetting() = default;
-    explicit InputSetting(Type val) : global{val} {}
+    explicit InputSetting(Type val) : BasicSetting<Type>(val) {}
     ~InputSetting() = default;
     void SetGlobal(bool to_global) {
         use_global = to_global;
     }
-    bool UsingGlobal() const {
+    [[nodiscard]] bool UsingGlobal() const {
         return use_global;
     }
-    Type& GetValue(bool need_global = false) {
+    [[nodiscard]] Type& GetValue(bool need_global = false) {
         if (use_global || need_global) {
             return global;
         }
-        return local;
+        return custom;
     }
 
 private:
-    bool use_global = true;
-    Type global{};
-    Type local{};
+    bool use_global{true}; ///< The setting's global state
+    Type global{};         ///< The setting
+    Type custom{};         ///< The custom setting value
 };
 
 struct TouchFromButtonMap {
@@ -105,144 +274,158 @@ struct TouchFromButtonMap {
 
 struct Values {
     // Audio
-    std::string audio_device_id;
-    std::string sink_id;
-    bool audio_muted;
-    Setting<bool> enable_audio_stretching;
-    Setting<float> volume;
+    BasicSetting<std::string> audio_device_id{"auto", "output_device"};
+    BasicSetting<std::string> sink_id{"auto", "output_engine"};
+    BasicSetting<bool> audio_muted{false, "audio_muted"};
+    Setting<bool> enable_audio_stretching{true, "enable_audio_stretching"};
+    Setting<u8> volume{100, "volume"};
 
     // Core
-    Setting<bool> use_multi_core;
+    Setting<bool> use_multi_core{true, "use_multi_core"};
 
     // Cpu
-    Setting<CPUAccuracy> cpu_accuracy;
+    Setting<CPUAccuracy> cpu_accuracy{CPUAccuracy::Auto, "cpu_accuracy"};
+    // TODO: remove cpu_accuracy_first_time, migration setting added 8 July 2021
+    BasicSetting<bool> cpu_accuracy_first_time{true, "cpu_accuracy_first_time"};
+    BasicSetting<bool> cpu_debug_mode{false, "cpu_debug_mode"};
 
-    bool cpuopt_page_tables;
-    bool cpuopt_block_linking;
-    bool cpuopt_return_stack_buffer;
-    bool cpuopt_fast_dispatcher;
-    bool cpuopt_context_elimination;
-    bool cpuopt_const_prop;
-    bool cpuopt_misc_ir;
-    bool cpuopt_reduce_misalign_checks;
-    bool cpuopt_fastmem;
+    BasicSetting<bool> cpuopt_page_tables{true, "cpuopt_page_tables"};
+    BasicSetting<bool> cpuopt_block_linking{true, "cpuopt_block_linking"};
+    BasicSetting<bool> cpuopt_return_stack_buffer{true, "cpuopt_return_stack_buffer"};
+    BasicSetting<bool> cpuopt_fast_dispatcher{true, "cpuopt_fast_dispatcher"};
+    BasicSetting<bool> cpuopt_context_elimination{true, "cpuopt_context_elimination"};
+    BasicSetting<bool> cpuopt_const_prop{true, "cpuopt_const_prop"};
+    BasicSetting<bool> cpuopt_misc_ir{true, "cpuopt_misc_ir"};
+    BasicSetting<bool> cpuopt_reduce_misalign_checks{true, "cpuopt_reduce_misalign_checks"};
+    BasicSetting<bool> cpuopt_fastmem{true, "cpuopt_fastmem"};
 
-    Setting<bool> cpuopt_unsafe_unfuse_fma;
-    Setting<bool> cpuopt_unsafe_reduce_fp_error;
-    Setting<bool> cpuopt_unsafe_ignore_standard_fpcr;
-    Setting<bool> cpuopt_unsafe_inaccurate_nan;
-    Setting<bool> cpuopt_unsafe_fastmem_check;
+    Setting<bool> cpuopt_unsafe_unfuse_fma{true, "cpuopt_unsafe_unfuse_fma"};
+    Setting<bool> cpuopt_unsafe_reduce_fp_error{true, "cpuopt_unsafe_reduce_fp_error"};
+    Setting<bool> cpuopt_unsafe_ignore_standard_fpcr{true, "cpuopt_unsafe_ignore_standard_fpcr"};
+    Setting<bool> cpuopt_unsafe_inaccurate_nan{true, "cpuopt_unsafe_inaccurate_nan"};
+    Setting<bool> cpuopt_unsafe_fastmem_check{true, "cpuopt_unsafe_fastmem_check"};
 
     // Renderer
-    Setting<RendererBackend> renderer_backend;
-    bool renderer_debug;
-    Setting<int> vulkan_device;
+    Setting<RendererBackend> renderer_backend{RendererBackend::OpenGL, "backend"};
+    BasicSetting<bool> renderer_debug{false, "debug"};
+    Setting<int> vulkan_device{0, "vulkan_device"};
 
-    Setting<u16> resolution_factor{1};
-    Setting<int> fullscreen_mode;
-    Setting<int> aspect_ratio;
-    Setting<int> max_anisotropy;
-    Setting<bool> use_frame_limit;
-    Setting<u16> frame_limit;
-    Setting<bool> use_disk_shader_cache;
-    Setting<GPUAccuracy> gpu_accuracy;
-    Setting<bool> use_asynchronous_gpu_emulation;
-    Setting<bool> use_nvdec_emulation;
-    Setting<bool> accelerate_astc;
-    Setting<bool> use_vsync;
-    Setting<bool> disable_fps_limit;
-    Setting<bool> use_assembly_shaders;
-    Setting<bool> use_asynchronous_shaders;
-    Setting<bool> use_fast_gpu_time;
-    Setting<bool> use_caches_gc;
+    Setting<u16> resolution_factor{1, "resolution_factor"};
+    // *nix platforms may have issues with the borderless windowed fullscreen mode.
+    // Default to exclusive fullscreen on these platforms for now.
+    Setting<int> fullscreen_mode{
+#ifdef _WIN32
+        0,
+#else
+        1,
+#endif
+        "fullscreen_mode"};
+    Setting<int> aspect_ratio{0, "aspect_ratio"};
+    Setting<int> max_anisotropy{0, "max_anisotropy"};
+    Setting<bool> use_frame_limit{true, "use_frame_limit"};
+    Setting<u16> frame_limit{100, "frame_limit"};
+    Setting<bool> use_disk_shader_cache{true, "use_disk_shader_cache"};
+    Setting<GPUAccuracy> gpu_accuracy{GPUAccuracy::High, "gpu_accuracy"};
+    Setting<bool> use_asynchronous_gpu_emulation{true, "use_asynchronous_gpu_emulation"};
+    Setting<bool> use_nvdec_emulation{true, "use_nvdec_emulation"};
+    Setting<bool> accelerate_astc{true, "accelerate_astc"};
+    Setting<bool> use_vsync{true, "use_vsync"};
+    BasicSetting<bool> disable_fps_limit{false, "disable_fps_limit"};
+    Setting<bool> use_assembly_shaders{false, "use_assembly_shaders"};
+    Setting<bool> use_asynchronous_shaders{false, "use_asynchronous_shaders"};
+    Setting<bool> use_fast_gpu_time{true, "use_fast_gpu_time"};
+    Setting<bool> use_caches_gc{false, "use_caches_gc"};
 
-    Setting<float> bg_red;
-    Setting<float> bg_green;
-    Setting<float> bg_blue;
+    Setting<u8> bg_red{0, "bg_red"};
+    Setting<u8> bg_green{0, "bg_green"};
+    Setting<u8> bg_blue{0, "bg_blue"};
 
     // System
-    Setting<std::optional<u32>> rng_seed;
+    Setting<std::optional<u32>> rng_seed{std::optional<u32>(), "rng_seed"};
     // Measured in seconds since epoch
     std::optional<std::chrono::seconds> custom_rtc;
     // Set on game boot, reset on stop. Seconds difference between current time and `custom_rtc`
     std::chrono::seconds custom_rtc_differential;
 
-    s32 current_user;
-    Setting<s32> language_index;
-    Setting<s32> region_index;
-    Setting<s32> time_zone_index;
-    Setting<s32> sound_index;
+    BasicSetting<s32> current_user{0, "current_user"};
+    Setting<s32> language_index{1, "language_index"};
+    Setting<s32> region_index{1, "region_index"};
+    Setting<s32> time_zone_index{0, "time_zone_index"};
+    Setting<s32> sound_index{1, "sound_index"};
 
     // Controls
     InputSetting<std::array<PlayerInput, 10>> players;
 
-    Setting<bool> use_docked_mode;
+    Setting<bool> use_docked_mode{true, "use_docked_mode"};
 
-    Setting<bool> vibration_enabled;
-    Setting<bool> enable_accurate_vibrations;
+    Setting<bool> vibration_enabled{true, "vibration_enabled"};
+    Setting<bool> enable_accurate_vibrations{false, "enable_accurate_vibrations"};
 
-    Setting<bool> motion_enabled;
-    std::string motion_device;
-    std::string udp_input_servers;
+    Setting<bool> motion_enabled{true, "motion_enabled"};
+    BasicSetting<std::string> motion_device{"engine:motion_emu,update_period:100,sensitivity:0.01",
+                                            "motion_device"};
+    BasicSetting<std::string> udp_input_servers{InputCommon::CemuhookUDP::DEFAULT_SRV,
+                                                "udp_input_servers"};
 
-    bool mouse_panning;
-    float mouse_panning_sensitivity;
-    bool mouse_enabled;
+    BasicSetting<bool> mouse_panning{false, "mouse_panning"};
+    BasicSetting<u8> mouse_panning_sensitivity{1, "mouse_panning_sensitivity"};
+    BasicSetting<bool> mouse_enabled{false, "mouse_enabled"};
     std::string mouse_device;
     MouseButtonsRaw mouse_buttons;
 
-    bool emulate_analog_keyboard;
-    bool keyboard_enabled;
+    BasicSetting<bool> emulate_analog_keyboard{false, "emulate_analog_keyboard"};
+    BasicSetting<bool> keyboard_enabled{false, "keyboard_enabled"};
     KeyboardKeysRaw keyboard_keys;
     KeyboardModsRaw keyboard_mods;
 
-    bool debug_pad_enabled;
+    BasicSetting<bool> debug_pad_enabled{false, "debug_pad_enabled"};
     ButtonsRaw debug_pad_buttons;
     AnalogsRaw debug_pad_analogs;
 
     TouchscreenInput touchscreen;
 
-    bool use_touch_from_button;
-    std::string touch_device;
-    int touch_from_button_map_index;
+    BasicSetting<bool> use_touch_from_button{false, "use_touch_from_button"};
+    BasicSetting<std::string> touch_device{"min_x:100,min_y:50,max_x:1800,max_y:850",
+                                           "touch_device"};
+    BasicSetting<int> touch_from_button_map_index{0, "touch_from_button_map"};
     std::vector<TouchFromButtonMap> touch_from_button_maps;
 
     std::atomic_bool is_device_reload_pending{true};
 
     // Data Storage
-    bool use_virtual_sd;
-    bool gamecard_inserted;
-    bool gamecard_current_game;
-    std::string gamecard_path;
+    BasicSetting<bool> use_virtual_sd{true, "use_virtual_sd"};
+    BasicSetting<bool> gamecard_inserted{false, "gamecard_inserted"};
+    BasicSetting<bool> gamecard_current_game{false, "gamecard_current_game"};
+    BasicSetting<std::string> gamecard_path{std::string(), "gamecard_path"};
 
     // Debugging
     bool record_frame_times;
-    bool use_gdbstub;
-    u16 gdbstub_port;
-    std::string program_args;
-    bool dump_exefs;
-    bool dump_nso;
-    bool enable_fs_access_log;
-    bool reporting_services;
-    bool quest_flag;
-    bool disable_macro_jit;
-    bool extended_logging;
-    bool use_debug_asserts;
-    bool use_auto_stub;
+    BasicSetting<bool> use_gdbstub{false, "use_gdbstub"};
+    BasicSetting<u16> gdbstub_port{0, "gdbstub_port"};
+    BasicSetting<std::string> program_args{std::string(), "program_args"};
+    BasicSetting<bool> dump_exefs{false, "dump_exefs"};
+    BasicSetting<bool> dump_nso{false, "dump_nso"};
+    BasicSetting<bool> enable_fs_access_log{false, "enable_fs_access_log"};
+    BasicSetting<bool> reporting_services{false, "reporting_services"};
+    BasicSetting<bool> quest_flag{false, "quest_flag"};
+    BasicSetting<bool> disable_macro_jit{false, "disable_macro_jit"};
+    BasicSetting<bool> extended_logging{false, "extended_logging"};
+    BasicSetting<bool> use_debug_asserts{false, "use_debug_asserts"};
+    BasicSetting<bool> use_auto_stub{false, "use_auto_stub"};
 
     // Miscellaneous
-    std::string log_filter;
-    bool use_dev_keys;
+    BasicSetting<std::string> log_filter{"*:Info", "log_filter"};
+    BasicSetting<bool> use_dev_keys{false, "use_dev_keys"};
 
     // Services
-    std::string bcat_backend;
-    bool bcat_boxcat_local;
+    BasicSetting<std::string> bcat_backend{"none", "bcat_backend"};
+    BasicSetting<bool> bcat_boxcat_local{false, "bcat_boxcat_local"};
 
     // WebService
-    bool enable_telemetry;
-    std::string web_api_url;
-    std::string yuzu_username;
-    std::string yuzu_token;
+    BasicSetting<bool> enable_telemetry{true, "enable_telemetry"};
+    BasicSetting<std::string> web_api_url{"https://api.yuzu-emu.org", "web_api_url"};
+    BasicSetting<std::string> yuzu_username{std::string(), "yuzu_username"};
+    BasicSetting<std::string> yuzu_token{std::string(), "yuzu_token"};
 
     // Add-Ons
     std::map<u64, std::vector<std::string>> disabled_addons;

src/common/thread_worker.cpp

@@ -1,58 +0,0 @@
-// Copyright 2020 yuzu emulator team
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#include "common/thread.h"
-#include "common/thread_worker.h"
-
-namespace Common {
-
-ThreadWorker::ThreadWorker(std::size_t num_workers, const std::string& name) {
-    for (std::size_t i = 0; i < num_workers; ++i)
-        threads.emplace_back([this, thread_name{std::string{name}}] {
-            Common::SetCurrentThreadName(thread_name.c_str());
-
-            // Wait for first request
-            {
-                std::unique_lock lock{queue_mutex};
-                condition.wait(lock, [this] { return stop || !requests.empty(); });
-            }
-
-            while (true) {
-                std::function<void()> task;
-
-                {
-                    std::unique_lock lock{queue_mutex};
-                    condition.wait(lock, [this] { return stop || !requests.empty(); });
-                    if (stop || requests.empty()) {
-                        return;
-                    }
-                    task = std::move(requests.front());
-                    requests.pop();
-                }
-
-                task();
-            }
-        });
-}
-
-ThreadWorker::~ThreadWorker() {
-    {
-        std::unique_lock lock{queue_mutex};
-        stop = true;
-    }
-    condition.notify_all();
-    for (std::thread& thread : threads) {
-        thread.join();
-    }
-}
-
-void ThreadWorker::QueueWork(std::function<void()>&& work) {
-    {
-        std::unique_lock lock{queue_mutex};
-        requests.emplace(work);
-    }
-    condition.notify_one();
-}
-
-} // namespace Common

src/common/thread_worker.h

@@ -7,24 +7,110 @@
 #include <atomic>
 #include <functional>
 #include <mutex>
+#include <stop_token>
 #include <string>
+#include <thread>
+#include <type_traits>
 #include <vector>
 #include <queue>
 
+#include "common/thread.h"
+#include "common/unique_function.h"
+
 namespace Common {
 
-class ThreadWorker final {
+template <class StateType = void>
+class StatefulThreadWorker {
+    static constexpr bool with_state = !std::is_same_v<StateType, void>;
+
+    struct DummyCallable {
+        int operator()() const noexcept {
+            return 0;
+        }
+    };
+
+    using Task =
+        std::conditional_t<with_state, UniqueFunction<void, StateType*>, UniqueFunction<void>>;
+    using StateMaker = std::conditional_t<with_state, std::function<StateType()>, DummyCallable>;
+
 public:
-    explicit ThreadWorker(std::size_t num_workers, const std::string& name);
-    ~ThreadWorker();
-    void QueueWork(std::function<void()>&& work);
+    explicit StatefulThreadWorker(size_t num_workers, std::string name, StateMaker func = {})
+        : workers_queued{num_workers}, thread_name{std::move(name)} {
+        const auto lambda = [this, func](std::stop_token stop_token) {
+            Common::SetCurrentThreadName(thread_name.c_str());
+            {
+                std::conditional_t<with_state, StateType, int> state{func()};
+                while (!stop_token.stop_requested()) {
+                    Task task;
+                    {
+                        std::unique_lock lock{queue_mutex};
+                        if (requests.empty()) {
+                            wait_condition.notify_all();
+                        }
+                        condition.wait(lock, stop_token, [this] { return !requests.empty(); });
+                        if (stop_token.stop_requested()) {
+                            break;
+                        }
+                        task = std::move(requests.front());
+                        requests.pop();
+                    }
+                    if constexpr (with_state) {
+                        task(&state);
+                    } else {
+                        task();
+                    }
+                    ++work_done;
+                }
+            }
+            ++workers_stopped;
+            wait_condition.notify_all();
+        };
+        threads.reserve(num_workers);
+        for (size_t i = 0; i < num_workers; ++i) {
+            threads.emplace_back(lambda);
+        }
+    }
+
+    StatefulThreadWorker& operator=(const StatefulThreadWorker&) = delete;
+    StatefulThreadWorker(const StatefulThreadWorker&) = delete;
+
+    StatefulThreadWorker& operator=(StatefulThreadWorker&&) = delete;
+    StatefulThreadWorker(StatefulThreadWorker&&) = delete;
+
+    void QueueWork(Task work) {
+        {
+            std::unique_lock lock{queue_mutex};
+            requests.emplace(std::move(work));
+            ++work_scheduled;
+        }
+        condition.notify_one();
+    }
+
+    void WaitForRequests(std::stop_token stop_token = {}) {
+        std::stop_callback callback(stop_token, [this] {
+            for (auto& thread : threads) {
+                thread.request_stop();
+            }
+        });
+        std::unique_lock lock{queue_mutex};
+        wait_condition.wait(lock, [this] {
+            return workers_stopped >= workers_queued || work_done >= work_scheduled;
+        });
+    }
 
 private:
-    std::vector<std::thread> threads;
-    std::queue<std::function<void()>> requests;
+    std::queue<Task> requests;
     std::mutex queue_mutex;
-    std::condition_variable condition;
-    std::atomic_bool stop{};
+    std::condition_variable_any condition;
+    std::condition_variable wait_condition;
+    std::atomic<size_t> work_scheduled{};
+    std::atomic<size_t> work_done{};
+    std::atomic<size_t> workers_stopped{};
+    std::atomic<size_t> workers_queued{};
+    std::string thread_name;
+    std::vector<std::jthread> threads;
 };
 
+using ThreadWorker = StatefulThreadWorker<>;
+
 } // namespace Common

src/common/unique_function.h

@@ -0,0 +1,62 @@
+// Copyright 2021 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <memory>
+#include <utility>
+
+namespace Common {
+
+/// General purpose function wrapper similar to std::function.
+/// Unlike std::function, the captured values don't have to be copyable.
+/// This class can be moved but not copied.
+template <typename ResultType, typename... Args>
+class UniqueFunction {
+    class CallableBase {
+    public:
+        virtual ~CallableBase() = default;
+        virtual ResultType operator()(Args&&...) = 0;
+    };
+
+    template <typename Functor>
+    class Callable final : public CallableBase {
+    public:
+        Callable(Functor&& functor_) : functor{std::move(functor_)} {}
+        ~Callable() override = default;
+
+        ResultType operator()(Args&&... args) override {
+            return functor(std::forward<Args>(args)...);
+        }
+
+    private:
+        Functor functor;
+    };
+
+public:
+    UniqueFunction() = default;
+
+    template <typename Functor>
+    UniqueFunction(Functor&& functor)
+        : callable{std::make_unique<Callable<Functor>>(std::move(functor))} {}
+
+    UniqueFunction& operator=(UniqueFunction&& rhs) noexcept = default;
+    UniqueFunction(UniqueFunction&& rhs) noexcept = default;
+
+    UniqueFunction& operator=(const UniqueFunction&) = delete;
+    UniqueFunction(const UniqueFunction&) = delete;
+
+    ResultType operator()(Args&&... args) const {
+        return (*callable)(std::forward<Args>(args)...);
+    }
+
+    explicit operator bool() const noexcept {
+        return static_cast<bool>(callable);
+    }
+
+private:
+    std::unique_ptr<CallableBase> callable;
+};
+
+} // namespace Common

src/core/CMakeLists.txt

@@ -272,22 +272,22 @@ add_library(core STATIC
     hle/service/am/applet_ae.h
     hle/service/am/applet_oe.cpp
     hle/service/am/applet_oe.h
+    hle/service/am/applets/applet_controller.cpp
+    hle/service/am/applets/applet_controller.h
+    hle/service/am/applets/applet_error.cpp
+    hle/service/am/applets/applet_error.h
+    hle/service/am/applets/applet_general_backend.cpp
+    hle/service/am/applets/applet_general_backend.h
+    hle/service/am/applets/applet_profile_select.cpp
+    hle/service/am/applets/applet_profile_select.h
+    hle/service/am/applets/applet_software_keyboard.cpp
+    hle/service/am/applets/applet_software_keyboard.h
+    hle/service/am/applets/applet_software_keyboard_types.h
+    hle/service/am/applets/applet_web_browser.cpp
+    hle/service/am/applets/applet_web_browser.h
+    hle/service/am/applets/applet_web_browser_types.h
     hle/service/am/applets/applets.cpp
     hle/service/am/applets/applets.h
-    hle/service/am/applets/controller.cpp
-    hle/service/am/applets/controller.h
-    hle/service/am/applets/error.cpp
-    hle/service/am/applets/error.h
-    hle/service/am/applets/general_backend.cpp
-    hle/service/am/applets/general_backend.h
-    hle/service/am/applets/profile_select.cpp
-    hle/service/am/applets/profile_select.h
-    hle/service/am/applets/software_keyboard.cpp
-    hle/service/am/applets/software_keyboard.h
-    hle/service/am/applets/software_keyboard_types.h
-    hle/service/am/applets/web_browser.cpp
-    hle/service/am/applets/web_browser.h
-    hle/service/am/applets/web_types.h
     hle/service/am/idle.cpp
     hle/service/am/idle.h
     hle/service/am/omm.cpp
@@ -300,10 +300,10 @@ add_library(core STATIC
     hle/service/aoc/aoc_u.h
     hle/service/apm/apm.cpp
     hle/service/apm/apm.h
-    hle/service/apm/controller.cpp
-    hle/service/apm/controller.h
-    hle/service/apm/interface.cpp
-    hle/service/apm/interface.h
+    hle/service/apm/apm_controller.cpp
+    hle/service/apm/apm_controller.h
+    hle/service/apm/apm_interface.cpp
+    hle/service/apm/apm_interface.h
     hle/service/audio/audctl.cpp
     hle/service/audio/audctl.h
     hle/service/audio/auddbg.cpp
@@ -335,8 +335,8 @@ add_library(core STATIC
     hle/service/bcat/backend/backend.h
     hle/service/bcat/bcat.cpp
     hle/service/bcat/bcat.h
-    hle/service/bcat/module.cpp
-    hle/service/bcat/module.h
+    hle/service/bcat/bcat_module.cpp
+    hle/service/bcat/bcat_module.h
     hle/service/bpc/bpc.cpp
     hle/service/bpc/bpc.h
     hle/service/btdrv/btdrv.cpp
@@ -382,8 +382,8 @@ add_library(core STATIC
     hle/service/friend/errors.h
     hle/service/friend/friend.cpp
     hle/service/friend/friend.h
-    hle/service/friend/interface.cpp
-    hle/service/friend/interface.h
+    hle/service/friend/friend_interface.cpp
+    hle/service/friend/friend_interface.h
     hle/service/glue/arp.cpp
     hle/service/glue/arp.h
     hle/service/glue/bgtc.cpp
@@ -393,8 +393,8 @@ add_library(core STATIC
     hle/service/glue/errors.h
     hle/service/glue/glue.cpp
     hle/service/glue/glue.h
-    hle/service/glue/manager.cpp
-    hle/service/glue/manager.h
+    hle/service/glue/glue_manager.cpp
+    hle/service/glue/glue_manager.h
     hle/service/grc/grc.cpp
     hle/service/grc/grc.h
     hle/service/hid/hid.cpp
@@ -435,10 +435,10 @@ add_library(core STATIC
     hle/service/lm/lm.h
     hle/service/mig/mig.cpp
     hle/service/mig/mig.h
-    hle/service/mii/manager.cpp
-    hle/service/mii/manager.h
     hle/service/mii/mii.cpp
     hle/service/mii/mii.h
+    hle/service/mii/mii_manager.cpp
+    hle/service/mii/mii_manager.h
     hle/service/mii/raw_data.cpp
     hle/service/mii/raw_data.h
     hle/service/mii/types.h
@@ -486,11 +486,11 @@ add_library(core STATIC
     hle/service/nvdrv/devices/nvhost_vic.h
     hle/service/nvdrv/devices/nvmap.cpp
     hle/service/nvdrv/devices/nvmap.h
-    hle/service/nvdrv/interface.cpp
-    hle/service/nvdrv/interface.h
     hle/service/nvdrv/nvdata.h
     hle/service/nvdrv/nvdrv.cpp
     hle/service/nvdrv/nvdrv.h
+    hle/service/nvdrv/nvdrv_interface.cpp
+    hle/service/nvdrv/nvdrv_interface.h
     hle/service/nvdrv/nvmemp.cpp
     hle/service/nvdrv/nvmemp.h
     hle/service/nvdrv/syncpoint_manager.cpp
@@ -503,10 +503,10 @@ add_library(core STATIC
     hle/service/olsc/olsc.h
     hle/service/pcie/pcie.cpp
     hle/service/pcie/pcie.h
-    hle/service/pctl/module.cpp
-    hle/service/pctl/module.h
     hle/service/pctl/pctl.cpp
     hle/service/pctl/pctl.h
+    hle/service/pctl/pctl_module.cpp
+    hle/service/pctl/pctl_module.h
     hle/service/pcv/pcv.cpp
     hle/service/pcv/pcv.h
     hle/service/pm/pm.cpp
@@ -529,10 +529,10 @@ add_library(core STATIC
     hle/service/set/set_sys.h
     hle/service/set/settings.cpp
     hle/service/set/settings.h
-    hle/service/sm/controller.cpp
-    hle/service/sm/controller.h
     hle/service/sm/sm.cpp
     hle/service/sm/sm.h
+    hle/service/sm/sm_controller.cpp
+    hle/service/sm/sm_controller.h
     hle/service/sockets/bsd.cpp
     hle/service/sockets/bsd.h
     hle/service/sockets/ethc.cpp
@@ -547,10 +547,10 @@ add_library(core STATIC
     hle/service/sockets/sockets_translate.h
     hle/service/spl/csrng.cpp
     hle/service/spl/csrng.h
-    hle/service/spl/module.cpp
-    hle/service/spl/module.h
     hle/service/spl/spl.cpp
     hle/service/spl/spl.h
+    hle/service/spl/spl_module.cpp
+    hle/service/spl/spl_module.h
     hle/service/spl/spl_results.h
     hle/service/spl/spl_types.h
     hle/service/ssl/ssl.cpp
@@ -559,8 +559,6 @@ add_library(core STATIC
     hle/service/time/ephemeral_network_system_clock_context_writer.h
     hle/service/time/ephemeral_network_system_clock_core.h
     hle/service/time/errors.h
-    hle/service/time/interface.cpp
-    hle/service/time/interface.h
     hle/service/time/local_system_clock_context_writer.h
     hle/service/time/network_system_clock_context_writer.h
     hle/service/time/standard_local_system_clock_core.h
@@ -578,6 +576,8 @@ add_library(core STATIC
     hle/service/time/tick_based_steady_clock_core.h
     hle/service/time/time.cpp
     hle/service/time/time.h
+    hle/service/time/time_interface.cpp
+    hle/service/time/time_interface.h
     hle/service/time/time_manager.cpp
     hle/service/time/time_manager.h
     hle/service/time/time_sharedmemory.cpp

src/core/arm/dynarmic/arm_dynarmic_32.cpp

@@ -150,7 +150,7 @@ std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable*
     config.far_code_offset = 400_MiB;
 
     // Safe optimizations
-    if (Settings::values.cpu_accuracy.GetValue() == Settings::CPUAccuracy::DebugMode) {
+    if (Settings::values.cpu_debug_mode) {
         if (!Settings::values.cpuopt_page_tables) {
             config.page_table = nullptr;
         }
@@ -183,20 +183,28 @@ std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable*
     // Unsafe optimizations
     if (Settings::values.cpu_accuracy.GetValue() == Settings::CPUAccuracy::Unsafe) {
         config.unsafe_optimizations = true;
-        if (Settings::values.cpuopt_unsafe_unfuse_fma.GetValue()) {
+        if (Settings::values.cpuopt_unsafe_unfuse_fma) {
             config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_UnfuseFMA;
         }
-        if (Settings::values.cpuopt_unsafe_reduce_fp_error.GetValue()) {
+        if (Settings::values.cpuopt_unsafe_reduce_fp_error) {
             config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_ReducedErrorFP;
         }
-        if (Settings::values.cpuopt_unsafe_ignore_standard_fpcr.GetValue()) {
+        if (Settings::values.cpuopt_unsafe_ignore_standard_fpcr) {
             config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_IgnoreStandardFPCRValue;
         }
-        if (Settings::values.cpuopt_unsafe_inaccurate_nan.GetValue()) {
+        if (Settings::values.cpuopt_unsafe_inaccurate_nan) {
             config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_InaccurateNaN;
         }
     }
 
+    // Curated optimizations
+    if (Settings::values.cpu_accuracy.GetValue() == Settings::CPUAccuracy::Auto) {
+        config.unsafe_optimizations = true;
+        config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_UnfuseFMA;
+        config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_IgnoreStandardFPCRValue;
+        config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_InaccurateNaN;
+    }
+
     return std::make_unique<Dynarmic::A32::Jit>(config);
 }
 

src/core/arm/dynarmic/arm_dynarmic_64.cpp

@@ -190,7 +190,7 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable*
     config.far_code_offset = 400_MiB;
 
     // Safe optimizations
-    if (Settings::values.cpu_accuracy.GetValue() == Settings::CPUAccuracy::DebugMode) {
+    if (Settings::values.cpu_debug_mode) {
         if (!Settings::values.cpuopt_page_tables) {
             config.page_table = nullptr;
         }
@@ -223,20 +223,28 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable*
     // Unsafe optimizations
     if (Settings::values.cpu_accuracy.GetValue() == Settings::CPUAccuracy::Unsafe) {
         config.unsafe_optimizations = true;
-        if (Settings::values.cpuopt_unsafe_unfuse_fma.GetValue()) {
+        if (Settings::values.cpuopt_unsafe_unfuse_fma) {
             config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_UnfuseFMA;
         }
-        if (Settings::values.cpuopt_unsafe_reduce_fp_error.GetValue()) {
+        if (Settings::values.cpuopt_unsafe_reduce_fp_error) {
             config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_ReducedErrorFP;
         }
-        if (Settings::values.cpuopt_unsafe_inaccurate_nan.GetValue()) {
+        if (Settings::values.cpuopt_unsafe_inaccurate_nan) {
             config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_InaccurateNaN;
         }
-        if (Settings::values.cpuopt_unsafe_fastmem_check.GetValue()) {
+        if (Settings::values.cpuopt_unsafe_fastmem_check) {
             config.fastmem_address_space_bits = 64;
         }
     }
 
+    // Curated optimizations
+    if (Settings::values.cpu_accuracy.GetValue() == Settings::CPUAccuracy::Auto) {
+        config.unsafe_optimizations = true;
+        config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_UnfuseFMA;
+        config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_InaccurateNaN;
+        config.fastmem_address_space_bits = 64;
+    }
+
     return std::make_shared<Dynarmic::A64::Jit>(config);
 }
 

src/core/core.cpp

@@ -35,9 +35,9 @@
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/physical_core.h"
 #include "core/hle/service/am/applets/applets.h"
-#include "core/hle/service/apm/controller.h"
+#include "core/hle/service/apm/apm_controller.h"
 #include "core/hle/service/filesystem/filesystem.h"
-#include "core/hle/service/glue/manager.h"
+#include "core/hle/service/glue/glue_manager.h"
 #include "core/hle/service/hid/hid.h"
 #include "core/hle/service/service.h"
 #include "core/hle/service/sm/sm.h"
@@ -263,9 +263,9 @@ struct System::Impl {
         if (Settings::values.gamecard_inserted) {
             if (Settings::values.gamecard_current_game) {
                 fs_controller.SetGameCard(GetGameFileFromPath(virtual_filesystem, filepath));
-            } else if (!Settings::values.gamecard_path.empty()) {
-                fs_controller.SetGameCard(
-                    GetGameFileFromPath(virtual_filesystem, Settings::values.gamecard_path));
+            } else if (!Settings::values.gamecard_path.GetValue().empty()) {
+                const auto gamecard_path = Settings::values.gamecard_path.GetValue();
+                fs_controller.SetGameCard(GetGameFileFromPath(virtual_filesystem, gamecard_path));
             }
         }
 

src/core/file_sys/content_archive.cpp

@@ -5,7 +5,6 @@
 #include <algorithm>
 #include <cstring>
 #include <optional>
-#include <ranges>
 #include <utility>
 
 #include "common/logging/log.h"

src/core/frontend/applets/profile_select.cpp

@@ -13,7 +13,7 @@ ProfileSelectApplet::~ProfileSelectApplet() = default;
 void DefaultProfileSelectApplet::SelectProfile(
     std::function<void(std::optional<Common::UUID>)> callback) const {
     Service::Account::ProfileManager manager;
-    callback(manager.GetUser(Settings::values.current_user).value_or(Common::UUID{}));
+    callback(manager.GetUser(Settings::values.current_user.GetValue()).value_or(Common::UUID{}));
     LOG_INFO(Service_ACC, "called, selecting current user instead of prompting...");
 }
 

src/core/frontend/applets/software_keyboard.h

@@ -9,7 +9,7 @@
 
 #include "common/common_types.h"
 
-#include "core/hle/service/am/applets/software_keyboard_types.h"
+#include "core/hle/service/am/applets/applet_software_keyboard_types.h"
 
 namespace Core::Frontend {
 

src/core/frontend/applets/web_browser.h

@@ -7,7 +7,7 @@
 #include <functional>
 #include <string_view>
 
-#include "core/hle/service/am/applets/web_types.h"
+#include "core/hle/service/am/applets/applet_web_browser_types.h"
 
 namespace Core::Frontend {
 

src/core/hardware_interrupt_manager.cpp

@@ -5,7 +5,7 @@
 #include "core/core.h"
 #include "core/core_timing.h"
 #include "core/hardware_interrupt_manager.h"
-#include "core/hle/service/nvdrv/interface.h"
+#include "core/hle/service/nvdrv/nvdrv_interface.h"
 #include "core/hle/service/sm/sm.h"
 
 namespace Core::Hardware {

src/core/hle/service/acc/acc.cpp

@@ -26,7 +26,7 @@
 #include "core/hle/service/acc/errors.h"
 #include "core/hle/service/acc/profile_manager.h"
 #include "core/hle/service/glue/arp.h"
-#include "core/hle/service/glue/manager.h"
+#include "core/hle/service/glue/glue_manager.h"
 #include "core/hle/service/sm/sm.h"
 #include "core/loader/loader.h"
 

src/core/hle/service/acc/acc.h

@@ -5,7 +5,7 @@
 #pragma once
 
 #include "common/uuid.h"
-#include "core/hle/service/glue/manager.h"
+#include "core/hle/service/glue/glue_manager.h"
 #include "core/hle/service/service.h"
 
 namespace Service::Account {

src/core/hle/service/acc/profile_manager.cpp

@@ -48,7 +48,8 @@ ProfileManager::ProfileManager() {
         CreateNewUser(UUID::Generate(), "yuzu");
     }
 
-    auto current = std::clamp<int>(Settings::values.current_user, 0, MAX_USERS - 1);
+    auto current =
+        std::clamp<int>(static_cast<s32>(Settings::values.current_user), 0, MAX_USERS - 1);
 
     // If user index don't exist. Load the first user and change the active user
     if (!UserExistsIndex(current)) {

src/core/hle/service/am/am.cpp

@@ -24,16 +24,16 @@
 #include "core/hle/service/am/am.h"
 #include "core/hle/service/am/applet_ae.h"
 #include "core/hle/service/am/applet_oe.h"
+#include "core/hle/service/am/applets/applet_profile_select.h"
+#include "core/hle/service/am/applets/applet_software_keyboard.h"
+#include "core/hle/service/am/applets/applet_web_browser.h"
 #include "core/hle/service/am/applets/applets.h"
-#include "core/hle/service/am/applets/profile_select.h"
-#include "core/hle/service/am/applets/software_keyboard.h"
-#include "core/hle/service/am/applets/web_browser.h"
 #include "core/hle/service/am/idle.h"
 #include "core/hle/service/am/omm.h"
 #include "core/hle/service/am/spsm.h"
 #include "core/hle/service/am/tcap.h"
-#include "core/hle/service/apm/controller.h"
-#include "core/hle/service/apm/interface.h"
+#include "core/hle/service/apm/apm_controller.h"
+#include "core/hle/service/apm/apm_interface.h"
 #include "core/hle/service/bcat/backend/backend.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/hle/service/ns/ns.h"
@@ -1443,7 +1443,7 @@ void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) {
         params.is_account_selected = 1;
 
         Account::ProfileManager profile_manager{};
-        const auto uuid = profile_manager.GetUser(Settings::values.current_user);
+        const auto uuid = profile_manager.GetUser(static_cast<s32>(Settings::values.current_user));
         ASSERT(uuid);
         params.current_user = uuid->uuid;
 

src/core/hle/service/am/applets/applet_controller.cpp

@@ -12,7 +12,7 @@
 #include "core/frontend/applets/controller.h"
 #include "core/hle/result.h"
 #include "core/hle/service/am/am.h"
-#include "core/hle/service/am/applets/controller.h"
+#include "core/hle/service/am/applets/applet_controller.h"
 #include "core/hle/service/hid/controllers/npad.h"
 
 namespace Service::AM::Applets {

src/core/hle/service/am/applets/applet_error.cpp

@@ -11,7 +11,7 @@
 #include "core/frontend/applets/error.h"
 #include "core/hle/kernel/k_process.h"
 #include "core/hle/service/am/am.h"
-#include "core/hle/service/am/applets/error.h"
+#include "core/hle/service/am/applets/applet_error.h"
 #include "core/reporter.h"
 
 namespace Service::AM::Applets {

src/core/hle/service/am/applets/applet_general_backend.cpp

@@ -12,7 +12,7 @@
 #include "core/hle/kernel/k_process.h"
 #include "core/hle/result.h"
 #include "core/hle/service/am/am.h"
-#include "core/hle/service/am/applets/general_backend.h"
+#include "core/hle/service/am/applets/applet_general_backend.h"
 #include "core/reporter.h"
 
 namespace Service::AM::Applets {

src/core/hle/service/am/applets/applet_profile_select.cpp

@@ -9,7 +9,7 @@
 #include "core/core.h"
 #include "core/frontend/applets/profile_select.h"
 #include "core/hle/service/am/am.h"
-#include "core/hle/service/am/applets/profile_select.h"
+#include "core/hle/service/am/applets/applet_profile_select.h"
 
 namespace Service::AM::Applets {
 

src/core/hle/service/am/applets/applet_software_keyboard.cpp

@@ -6,7 +6,7 @@
 #include "core/core.h"
 #include "core/frontend/applets/software_keyboard.h"
 #include "core/hle/service/am/am.h"
-#include "core/hle/service/am/applets/software_keyboard.h"
+#include "core/hle/service/am/applets/applet_software_keyboard.h"
 
 namespace Service::AM::Applets {
 

src/core/hle/service/am/applets/applet_software_keyboard.h

@@ -7,8 +7,8 @@
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "core/hle/result.h"
+#include "core/hle/service/am/applets/applet_software_keyboard_types.h"
 #include "core/hle/service/am/applets/applets.h"
-#include "core/hle/service/am/applets/software_keyboard_types.h"
 
 namespace Core {
 class System;

src/core/hle/service/am/applets/applet_web_browser.cpp

@@ -21,7 +21,7 @@
 #include "core/hle/kernel/k_process.h"
 #include "core/hle/result.h"
 #include "core/hle/service/am/am.h"
-#include "core/hle/service/am/applets/web_browser.h"
+#include "core/hle/service/am/applets/applet_web_browser.h"
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/hle/service/ns/pl_u.h"
 
@@ -58,6 +58,16 @@ std::string GetMainURL(const std::string& url) {
     return url.substr(0, index);
 }
 
+std::string ResolveURL(const std::string& url) {
+    const auto index = url.find_first_of('%');
+
+    if (index == std::string::npos) {
+        return url;
+    }
+
+    return url.substr(0, index) + "lp1" + url.substr(index + 1);
+}
+
 WebArgInputTLVMap ReadWebArgs(const std::vector<u8>& web_arg, WebArgHeader& web_arg_header) {
     std::memcpy(&web_arg_header, web_arg.data(), sizeof(WebArgHeader));
 
@@ -407,6 +417,9 @@ void WebBrowser::InitializeShare() {}
 
 void WebBrowser::InitializeWeb() {
     external_url = ParseStringValue(GetInputTLVData(WebArgInputTLVType::InitialURL).value());
+
+    // Resolve Nintendo CDN URLs.
+    external_url = ResolveURL(external_url);
 }
 
 void WebBrowser::InitializeWifi() {}

src/core/hle/service/am/applets/applet_web_browser.h

@@ -11,8 +11,8 @@
 #include "common/common_types.h"
 #include "core/file_sys/vfs_types.h"
 #include "core/hle/result.h"
+#include "core/hle/service/am/applets/applet_web_browser_types.h"
 #include "core/hle/service/am/applets/applets.h"
-#include "core/hle/service/am/applets/web_types.h"
 
 namespace Core {
 class System;

src/core/hle/service/am/applets/applets.cpp

@@ -17,13 +17,13 @@
 #include "core/hle/service/am/am.h"
 #include "core/hle/service/am/applet_ae.h"
 #include "core/hle/service/am/applet_oe.h"
+#include "core/hle/service/am/applets/applet_controller.h"
+#include "core/hle/service/am/applets/applet_error.h"
+#include "core/hle/service/am/applets/applet_general_backend.h"
+#include "core/hle/service/am/applets/applet_profile_select.h"
+#include "core/hle/service/am/applets/applet_software_keyboard.h"
+#include "core/hle/service/am/applets/applet_web_browser.h"
 #include "core/hle/service/am/applets/applets.h"
-#include "core/hle/service/am/applets/controller.h"
-#include "core/hle/service/am/applets/error.h"
-#include "core/hle/service/am/applets/general_backend.h"
-#include "core/hle/service/am/applets/profile_select.h"
-#include "core/hle/service/am/applets/software_keyboard.h"
-#include "core/hle/service/am/applets/web_browser.h"
 #include "core/hle/service/sm/sm.h"
 
 namespace Service::AM::Applets {

src/core/hle/service/apm/apm.cpp

@@ -5,7 +5,7 @@
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/service/apm/apm.h"
-#include "core/hle/service/apm/interface.h"
+#include "core/hle/service/apm/apm_interface.h"
 
 namespace Service::APM {
 

src/core/hle/service/apm/apm_controller.cpp

@@ -9,7 +9,7 @@
 #include "common/logging/log.h"
 #include "common/settings.h"
 #include "core/core_timing.h"
-#include "core/hle/service/apm/controller.h"
+#include "core/hle/service/apm/apm_controller.h"
 
 namespace Service::APM {
 

src/core/hle/service/apm/apm_interface.cpp

@@ -5,8 +5,8 @@
 #include "common/logging/log.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/service/apm/apm.h"
-#include "core/hle/service/apm/controller.h"
-#include "core/hle/service/apm/interface.h"
+#include "core/hle/service/apm/apm_controller.h"
+#include "core/hle/service/apm/apm_interface.h"
 
 namespace Service::APM {
 

src/core/hle/service/audio/audren_u.cpp

@@ -290,7 +290,7 @@ private:
 
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(ResultSuccess);
-        rb.Push<u32>(1);
+        rb.Push<u32>(2);
     }
 
     // Should be similar to QueryAudioDeviceOutputEvent

src/core/hle/service/bcat/backend/boxcat.cpp

@@ -7,6 +7,9 @@
 #ifdef __GNUC__
 #pragma GCC diagnostic push
 #pragma GCC diagnostic ignored "-Wshadow"
+#ifndef __clang__
+#pragma GCC diagnostic ignored "-Wmaybe-uninitialized"
+#endif
 #endif
 #include <httplib.h>
 #include <mbedtls/sha256.h>

src/core/hle/service/bcat/bcat.h

@@ -4,7 +4,7 @@
 
 #pragma once
 
-#include "core/hle/service/bcat/module.h"
+#include "core/hle/service/bcat/bcat_module.h"
 
 namespace Core {
 class System;

src/core/hle/service/bcat/bcat_module.cpp

@@ -17,7 +17,7 @@
 #include "core/hle/kernel/k_writable_event.h"
 #include "core/hle/service/bcat/backend/backend.h"
 #include "core/hle/service/bcat/bcat.h"
-#include "core/hle/service/bcat/module.h"
+#include "core/hle/service/bcat/bcat_module.h"
 #include "core/hle/service/filesystem/filesystem.h"
 
 namespace Service::BCAT {
@@ -579,7 +579,7 @@ void Module::Interface::CreateDeliveryCacheStorageServiceWithApplicationId(
 std::unique_ptr<Backend> CreateBackendFromSettings([[maybe_unused]] Core::System& system,
                                                    DirectoryGetter getter) {
 #ifdef YUZU_ENABLE_BOXCAT
-    if (Settings::values.bcat_backend == "boxcat") {
+    if (Settings::values.bcat_backend.GetValue() == "boxcat") {
         return std::make_unique<Boxcat>(system.GetAppletManager(), std::move(getter));
     }
 #endif

src/core/hle/service/friend/friend.cpp

@@ -12,7 +12,7 @@
 #include "core/hle/kernel/k_writable_event.h"
 #include "core/hle/service/friend/errors.h"
 #include "core/hle/service/friend/friend.h"
-#include "core/hle/service/friend/interface.h"
+#include "core/hle/service/friend/friend_interface.h"
 
 namespace Service::Friend {
 

src/core/hle/service/friend/friend_interface.cpp

@@ -2,7 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include "core/hle/service/friend/interface.h"
+#include "core/hle/service/friend/friend_interface.h"
 
 namespace Service::Friend {
 

src/core/hle/service/glue/arp.cpp

@@ -13,7 +13,7 @@
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/service/glue/arp.h"
 #include "core/hle/service/glue/errors.h"
-#include "core/hle/service/glue/manager.h"
+#include "core/hle/service/glue/glue_manager.h"
 #include "core/hle/service/service.h"
 
 namespace Service::Glue {

src/core/hle/service/glue/glue_manager.cpp

@@ -3,7 +3,7 @@
 // Refer to the license.txt file included.
 
 #include "core/hle/service/glue/errors.h"
-#include "core/hle/service/glue/manager.h"
+#include "core/hle/service/glue/glue_manager.h"
 
 namespace Service::Glue {
 

src/core/hle/service/hid/controllers/npad.cpp

@@ -941,6 +941,11 @@ void Controller_NPad::InitializeVibrationDevice(const DeviceHandle& vibration_de
 
 void Controller_NPad::InitializeVibrationDeviceAtIndex(std::size_t npad_index,
                                                        std::size_t device_index) {
+    if (!Settings::values.vibration_enabled.GetValue()) {
+        vibration_devices_mounted[npad_index][device_index] = false;
+        return;
+    }
+
     if (vibrations[npad_index][device_index]) {
         vibration_devices_mounted[npad_index][device_index] =
             vibrations[npad_index][device_index]->GetStatus() == 1;

src/core/hle/service/mii/mii.cpp

@@ -7,8 +7,8 @@
 #include "common/logging/log.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/hle_ipc.h"
-#include "core/hle/service/mii/manager.h"
 #include "core/hle/service/mii/mii.h"
+#include "core/hle/service/mii/mii_manager.h"
 #include "core/hle/service/service.h"
 #include "core/hle/service/sm/sm.h"
 

src/core/hle/service/mii/mii_manager.cpp

@@ -10,7 +10,7 @@
 #include "common/string_util.h"
 
 #include "core/hle/service/acc/profile_manager.h"
-#include "core/hle/service/mii/manager.h"
+#include "core/hle/service/mii/mii_manager.h"
 #include "core/hle/service/mii/raw_data.h"
 #include "core/hle/service/mii/types.h"
 

src/core/hle/service/mii/raw_data.h

@@ -7,7 +7,7 @@
 #include <array>
 
 #include "common/common_types.h"
-#include "core/hle/service/mii/manager.h"
+#include "core/hle/service/mii/mii_manager.h"
 
 namespace Service::Mii::RawData {
 

src/core/hle/service/nifm/nifm.cpp

@@ -179,7 +179,7 @@ private:
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(ResultSuccess);
 
-        if (Settings::values.bcat_backend == "none") {
+        if (Settings::values.bcat_backend.GetValue() == "none") {
             rb.PushEnum(RequestState::NotSubmitted);
         } else {
             rb.PushEnum(RequestState::Connected);
@@ -384,7 +384,7 @@ private:
 
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(ResultSuccess);
-        if (Settings::values.bcat_backend == "none") {
+        if (Settings::values.bcat_backend.GetValue() == "none") {
             rb.Push<u8>(0);
         } else {
             rb.Push<u8>(1);
@@ -395,7 +395,7 @@ private:
 
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(ResultSuccess);
-        if (Settings::values.bcat_backend == "none") {
+        if (Settings::values.bcat_backend.GetValue() == "none") {
             rb.Push<u8>(0);
         } else {
             rb.Push<u8>(1);

src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp

@@ -19,26 +19,29 @@
 namespace Service::Nvidia::Devices {
 
 namespace {
-// Splice vectors will copy count amount of type T from the input vector into the dst vector.
+// Copies count amount of type T from the input vector into the dst vector.
+// Returns the number of bytes written into dst.
 template <typename T>
-std::size_t SpliceVectors(const std::vector<u8>& input, std::vector<T>& dst, std::size_t count,
-                          std::size_t offset) {
-    if (!dst.empty()) {
-        std::memcpy(dst.data(), input.data() + offset, count * sizeof(T));
+std::size_t SliceVectors(const std::vector<u8>& input, std::vector<T>& dst, std::size_t count,
+                         std::size_t offset) {
+    if (dst.empty()) {
+        return 0;
     }
-    return 0;
+    const size_t bytes_copied = count * sizeof(T);
+    std::memcpy(dst.data(), input.data() + offset, bytes_copied);
+    return bytes_copied;
 }
 
-// Write vectors will write data to the output buffer
+// Writes the data in src to an offset into the dst vector. The offset is specified in bytes
+// Returns the number of bytes written into dst.
 template <typename T>
 std::size_t WriteVectors(std::vector<u8>& dst, const std::vector<T>& src, std::size_t offset) {
     if (src.empty()) {
         return 0;
-    } else {
-        std::memcpy(dst.data() + offset, src.data(), src.size() * sizeof(T));
-        offset += src.size() * sizeof(T);
-        return offset;
     }
+    const size_t bytes_copied = src.size() * sizeof(T);
+    std::memcpy(dst.data() + offset, src.data(), bytes_copied);
+    return bytes_copied;
 }
 } // Anonymous namespace
 
@@ -62,7 +65,6 @@ NvResult nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u
     LOG_DEBUG(Service_NVDRV, "called NVDEC Submit, cmd_buffer_count={}", params.cmd_buffer_count);
 
     // Instantiate param buffers
-    std::size_t offset = sizeof(IoctlSubmit);
     std::vector<CommandBuffer> command_buffers(params.cmd_buffer_count);
     std::vector<Reloc> relocs(params.relocation_count);
     std::vector<u32> reloc_shifts(params.relocation_count);
@@ -70,13 +72,14 @@ NvResult nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u
     std::vector<SyncptIncr> wait_checks(params.syncpoint_count);
     std::vector<Fence> fences(params.fence_count);
 
-    // Splice input into their respective buffers
-    offset = SpliceVectors(input, command_buffers, params.cmd_buffer_count, offset);
-    offset = SpliceVectors(input, relocs, params.relocation_count, offset);
-    offset = SpliceVectors(input, reloc_shifts, params.relocation_count, offset);
-    offset = SpliceVectors(input, syncpt_increments, params.syncpoint_count, offset);
-    offset = SpliceVectors(input, wait_checks, params.syncpoint_count, offset);
-    offset = SpliceVectors(input, fences, params.fence_count, offset);
+    // Slice input into their respective buffers
+    std::size_t offset = sizeof(IoctlSubmit);
+    offset += SliceVectors(input, command_buffers, params.cmd_buffer_count, offset);
+    offset += SliceVectors(input, relocs, params.relocation_count, offset);
+    offset += SliceVectors(input, reloc_shifts, params.relocation_count, offset);
+    offset += SliceVectors(input, syncpt_increments, params.syncpoint_count, offset);
+    offset += SliceVectors(input, wait_checks, params.syncpoint_count, offset);
+    offset += SliceVectors(input, fences, params.fence_count, offset);
 
     auto& gpu = system.GPU();
     if (gpu.UseNvdec()) {
@@ -88,35 +91,27 @@ NvResult nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u
         }
     }
     for (const auto& cmd_buffer : command_buffers) {
-        auto object = nvmap_dev->GetObject(cmd_buffer.memory_id);
+        const auto object = nvmap_dev->GetObject(cmd_buffer.memory_id);
         ASSERT_OR_EXECUTE(object, return NvResult::InvalidState;);
-        const auto map = FindBufferMap(object->dma_map_addr);
-        if (!map) {
-            LOG_ERROR(Service_NVDRV, "Tried to submit an invalid offset 0x{:X} dma 0x{:X}",
-                      object->addr, object->dma_map_addr);
-            return NvResult::Success;
-        }
         Tegra::ChCommandHeaderList cmdlist(cmd_buffer.word_count);
-        gpu.MemoryManager().ReadBlock(map->StartAddr() + cmd_buffer.offset, cmdlist.data(),
-                                      cmdlist.size() * sizeof(u32));
+        system.Memory().ReadBlock(object->addr + cmd_buffer.offset, cmdlist.data(),
+                                  cmdlist.size() * sizeof(u32));
         gpu.PushCommandBuffer(cmdlist);
     }
     if (gpu.UseNvdec()) {
-
         fences[0].value = syncpoint_manager.IncreaseSyncpoint(fences[0].id, 1);
-
         Tegra::ChCommandHeaderList cmdlist{{(4 << 28) | fences[0].id}};
         gpu.PushCommandBuffer(cmdlist);
     }
     std::memcpy(output.data(), &params, sizeof(IoctlSubmit));
     // Some games expect command_buffers to be written back
     offset = sizeof(IoctlSubmit);
-    offset = WriteVectors(output, command_buffers, offset);
-    offset = WriteVectors(output, relocs, offset);
-    offset = WriteVectors(output, reloc_shifts, offset);
-    offset = WriteVectors(output, syncpt_increments, offset);
-    offset = WriteVectors(output, wait_checks, offset);
-    offset = WriteVectors(output, fences, offset);
+    offset += WriteVectors(output, command_buffers, offset);
+    offset += WriteVectors(output, relocs, offset);
+    offset += WriteVectors(output, reloc_shifts, offset);
+    offset += WriteVectors(output, syncpt_increments, offset);
+    offset += WriteVectors(output, wait_checks, offset);
+    offset += WriteVectors(output, fences, offset);
 
     return NvResult::Success;
 }
@@ -148,14 +143,14 @@ NvResult nvhost_nvdec_common::MapBuffer(const std::vector<u8>& input, std::vecto
     std::memcpy(&params, input.data(), sizeof(IoctlMapBuffer));
     std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries);
 
-    SpliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer));
+    SliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer));
 
     auto& gpu = system.GPU();
 
-    for (auto& cmf_buff : cmd_buffer_handles) {
-        auto object{nvmap_dev->GetObject(cmf_buff.map_handle)};
+    for (auto& cmd_buffer : cmd_buffer_handles) {
+        auto object{nvmap_dev->GetObject(cmd_buffer.map_handle)};
         if (!object) {
-            LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle);
+            LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmd_buffer.map_handle);
             std::memcpy(output.data(), &params, output.size());
             return NvResult::InvalidState;
         }
@@ -170,7 +165,7 @@ NvResult nvhost_nvdec_common::MapBuffer(const std::vector<u8>& input, std::vecto
         if (!object->dma_map_addr) {
             LOG_ERROR(Service_NVDRV, "failed to map size={}", object->size);
         } else {
-            cmf_buff.map_address = object->dma_map_addr;
+            cmd_buffer.map_address = object->dma_map_addr;
             AddBufferMap(object->dma_map_addr, object->size, object->addr,
                          object->status == nvmap::Object::Status::Allocated);
         }
@@ -186,14 +181,14 @@ NvResult nvhost_nvdec_common::UnmapBuffer(const std::vector<u8>& input, std::vec
     IoctlMapBuffer params{};
     std::memcpy(&params, input.data(), sizeof(IoctlMapBuffer));
     std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries);
-    SpliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer));
+    SliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer));
 
     auto& gpu = system.GPU();
 
-    for (auto& cmf_buff : cmd_buffer_handles) {
-        const auto object{nvmap_dev->GetObject(cmf_buff.map_handle)};
+    for (auto& cmd_buffer : cmd_buffer_handles) {
+        const auto object{nvmap_dev->GetObject(cmd_buffer.map_handle)};
         if (!object) {
-            LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle);
+            LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmd_buffer.map_handle);
             std::memcpy(output.data(), &params, output.size());
             return NvResult::InvalidState;
         }

src/core/hle/service/nvdrv/nvdrv.cpp

@@ -20,8 +20,8 @@
 #include "core/hle/service/nvdrv/devices/nvhost_nvjpg.h"
 #include "core/hle/service/nvdrv/devices/nvhost_vic.h"
 #include "core/hle/service/nvdrv/devices/nvmap.h"
-#include "core/hle/service/nvdrv/interface.h"
 #include "core/hle/service/nvdrv/nvdrv.h"
+#include "core/hle/service/nvdrv/nvdrv_interface.h"
 #include "core/hle/service/nvdrv/nvmemp.h"
 #include "core/hle/service/nvdrv/syncpoint_manager.h"
 #include "core/hle/service/nvflinger/nvflinger.h"

src/core/hle/service/nvdrv/nvdrv_interface.cpp

@@ -10,9 +10,9 @@
 #include "core/hle/kernel/k_thread.h"
 #include "core/hle/kernel/k_writable_event.h"
 #include "core/hle/kernel/kernel.h"
-#include "core/hle/service/nvdrv/interface.h"
 #include "core/hle/service/nvdrv/nvdata.h"
 #include "core/hle/service/nvdrv/nvdrv.h"
+#include "core/hle/service/nvdrv/nvdrv_interface.h"
 
 namespace Service::Nvidia {
 

src/core/hle/service/pctl/pctl.h

@@ -4,7 +4,7 @@
 
 #pragma once
 
-#include "core/hle/service/pctl/module.h"
+#include "core/hle/service/pctl/pctl_module.h"
 
 namespace Core {
 class System;

src/core/hle/service/pctl/pctl_module.cpp

@@ -8,8 +8,8 @@
 #include "core/file_sys/patch_manager.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/k_process.h"
-#include "core/hle/service/pctl/module.h"
 #include "core/hle/service/pctl/pctl.h"
+#include "core/hle/service/pctl/pctl_module.h"
 
 namespace Service::PCTL {
 

src/core/hle/service/service.cpp

@@ -21,7 +21,7 @@
 #include "core/hle/service/aoc/aoc_u.h"
 #include "core/hle/service/apm/apm.h"
 #include "core/hle/service/audio/audio.h"
-#include "core/hle/service/bcat/module.h"
+#include "core/hle/service/bcat/bcat_module.h"
 #include "core/hle/service/bpc/bpc.h"
 #include "core/hle/service/btdrv/btdrv.h"
 #include "core/hle/service/btm/btm.h"
@@ -54,7 +54,7 @@
 #include "core/hle/service/nvflinger/nvflinger.h"
 #include "core/hle/service/olsc/olsc.h"
 #include "core/hle/service/pcie/pcie.h"
-#include "core/hle/service/pctl/module.h"
+#include "core/hle/service/pctl/pctl_module.h"
 #include "core/hle/service/pcv/pcv.h"
 #include "core/hle/service/pm/pm.h"
 #include "core/hle/service/prepo/prepo.h"
@@ -64,7 +64,7 @@
 #include "core/hle/service/set/settings.h"
 #include "core/hle/service/sm/sm.h"
 #include "core/hle/service/sockets/sockets.h"
-#include "core/hle/service/spl/module.h"
+#include "core/hle/service/spl/spl_module.h"
 #include "core/hle/service/ssl/ssl.h"
 #include "core/hle/service/time/time.h"
 #include "core/hle/service/usb/usb.h"

src/core/hle/service/set/set.cpp

@@ -160,7 +160,7 @@ void SET::GetQuestFlag(Kernel::HLERequestContext& ctx) {
 
     IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(ResultSuccess);
-    rb.Push(static_cast<u32>(Settings::values.quest_flag));
+    rb.Push(static_cast<u32>(Settings::values.quest_flag.GetValue()));
 }
 
 void SET::GetLanguageCode(Kernel::HLERequestContext& ctx) {

src/core/hle/service/sm/sm.cpp

@@ -14,8 +14,8 @@
 #include "core/hle/kernel/k_server_session.h"
 #include "core/hle/kernel/k_session.h"
 #include "core/hle/result.h"
-#include "core/hle/service/sm/controller.h"
 #include "core/hle/service/sm/sm.h"
+#include "core/hle/service/sm/sm_controller.h"
 
 namespace Service::SM {
 

src/core/hle/service/sm/sm_controller.cpp

@@ -13,7 +13,7 @@
 #include "core/hle/kernel/k_server_port.h"
 #include "core/hle/kernel/k_server_session.h"
 #include "core/hle/kernel/k_session.h"
-#include "core/hle/service/sm/controller.h"
+#include "core/hle/service/sm/sm_controller.h"
 
 namespace Service::SM {
 

src/core/hle/service/spl/csrng.h

@@ -4,7 +4,7 @@
 
 #pragma once
 
-#include "core/hle/service/spl/module.h"
+#include "core/hle/service/spl/spl_module.h"
 
 namespace Core {
 class System;

src/core/hle/service/spl/spl.h

@@ -4,7 +4,7 @@
 
 #pragma once
 
-#include "core/hle/service/spl/module.h"
+#include "core/hle/service/spl/spl_module.h"
 
 namespace Core {
 class System;

src/core/hle/service/spl/spl_module.cpp

@@ -13,8 +13,8 @@
 #include "core/hle/api_version.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/service/spl/csrng.h"
-#include "core/hle/service/spl/module.h"
 #include "core/hle/service/spl/spl.h"
+#include "core/hle/service/spl/spl_module.h"
 
 namespace Service::SPL {
 

src/core/hle/service/time/time.cpp

@@ -11,8 +11,8 @@
 #include "core/hle/kernel/k_client_port.h"
 #include "core/hle/kernel/k_scheduler.h"
 #include "core/hle/kernel/kernel.h"
-#include "core/hle/service/time/interface.h"
 #include "core/hle/service/time/time.h"
+#include "core/hle/service/time/time_interface.h"
 #include "core/hle/service/time/time_sharedmemory.h"
 #include "core/hle/service/time/time_zone_service.h"
 

src/core/hle/service/time/time_interface.cpp

@@ -2,7 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include "core/hle/service/time/interface.h"
+#include "core/hle/service/time/time_interface.h"
 
 namespace Service::Time {
 

src/core/loader/nro.cpp

@@ -155,8 +155,8 @@ static bool LoadNroImpl(Kernel::KProcess& process, const std::vector<u8>& data)
         codeset.segments[i].size = PageAlignSize(nro_header.segments[i].size);
     }
 
-    if (!Settings::values.program_args.empty()) {
-        const auto arg_data = Settings::values.program_args;
+    if (!Settings::values.program_args.GetValue().empty()) {
+        const auto arg_data = Settings::values.program_args.GetValue();
         codeset.DataSegment().size += NSO_ARGUMENT_DATA_ALLOCATION_SIZE;
         NSOArgumentHeader args_header{
             NSO_ARGUMENT_DATA_ALLOCATION_SIZE, static_cast<u32_le>(arg_data.size()), {}};

src/core/loader/nso.cpp

@@ -104,8 +104,8 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::KProcess& process, Core::
         codeset.segments[i].size = nso_header.segments[i].size;
     }
 
-    if (should_pass_arguments && !Settings::values.program_args.empty()) {
-        const auto arg_data{Settings::values.program_args};
+    if (should_pass_arguments && !Settings::values.program_args.GetValue().empty()) {
+        const auto arg_data{Settings::values.program_args.GetValue()};
 
         codeset.DataSegment().size += NSO_ARGUMENT_DATA_ALLOCATION_SIZE;
         NSOArgumentHeader args_header{

src/core/reporter.cpp

@@ -397,7 +397,7 @@ void Reporter::ClearFSAccessLog() const {
 }
 
 bool Reporter::IsReportingEnabled() const {
-    return Settings::values.reporting_services;
+    return Settings::values.reporting_services.GetValue();
 }
 
 } // namespace Core

src/core/telemetry_session.cpp

@@ -135,7 +135,7 @@ u64 RegenerateTelemetryId() {
 
 bool VerifyLogin(const std::string& username, const std::string& token) {
 #ifdef ENABLE_WEB_SERVICE
-    return WebService::VerifyLogin(Settings::values.web_api_url, username, token);
+    return WebService::VerifyLogin(Settings::values.web_api_url.GetValue(), username, token);
 #else
     return false;
 #endif
@@ -152,7 +152,8 @@ TelemetrySession::~TelemetrySession() {
 
 #ifdef ENABLE_WEB_SERVICE
     auto backend = std::make_unique<WebService::TelemetryJson>(
-        Settings::values.web_api_url, Settings::values.yuzu_username, Settings::values.yuzu_token);
+        Settings::values.web_api_url.GetValue(), Settings::values.yuzu_username.GetValue(),
+        Settings::values.yuzu_token.GetValue());
 #else
     auto backend = std::make_unique<Telemetry::NullVisitor>();
 #endif
@@ -212,7 +213,7 @@ void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader,
 
     // Log user configuration information
     constexpr auto field_type = Telemetry::FieldType::UserConfig;
-    AddField(field_type, "Audio_SinkId", Settings::values.sink_id);
+    AddField(field_type, "Audio_SinkId", Settings::values.sink_id.GetValue());
     AddField(field_type, "Audio_EnableAudioStretching",
              Settings::values.enable_audio_stretching.GetValue());
     AddField(field_type, "Core_UseMultiCore", Settings::values.use_multi_core.GetValue());
@@ -242,7 +243,8 @@ void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader,
 bool TelemetrySession::SubmitTestcase() {
 #ifdef ENABLE_WEB_SERVICE
     auto backend = std::make_unique<WebService::TelemetryJson>(
-        Settings::values.web_api_url, Settings::values.yuzu_username, Settings::values.yuzu_token);
+        Settings::values.web_api_url.GetValue(), Settings::values.yuzu_username.GetValue(),
+        Settings::values.yuzu_token.GetValue());
     field_collection.Accept(*backend);
     return backend->SubmitTestcase();
 #else

src/input_common/main.cpp

@@ -294,8 +294,8 @@ void InputSubsystem::ReloadInputDevices() {
     impl->udp->ReloadSockets();
 }
 
-std::vector<std::unique_ptr<Polling::DevicePoller>> InputSubsystem::GetPollers(
-    Polling::DeviceType type) const {
+std::vector<std::unique_ptr<Polling::DevicePoller>> InputSubsystem::GetPollers([
+    [maybe_unused]] Polling::DeviceType type) const {
 #ifdef HAVE_SDL2
     return impl->sdl->GetPollers(type);
 #else

src/input_common/mouse/mouse_poller.cpp

@@ -84,7 +84,8 @@ public:
         std::lock_guard lock{mutex};
         const auto axis_value =
             static_cast<float>(mouse_input->GetMouseState(button).axis.at(axis));
-        return axis_value * Settings::values.mouse_panning_sensitivity / (100.0f * range);
+        const float sensitivity = Settings::values.mouse_panning_sensitivity.GetValue() * 0.15f;
+        return axis_value * sensitivity / (100.0f * range);
     }
 
     std::pair<float, float> GetAnalog(u32 analog_axis_x, u32 analog_axis_y) const {

src/input_common/sdl/sdl_impl.h

@@ -10,15 +10,7 @@
 #include <thread>
 #include <unordered_map>
 
-// Ignore -Wimplicit-fallthrough due to https://github.com/libsdl-org/SDL/issues/4307
-#ifdef __GNUC__
-#pragma GCC diagnostic push
-#pragma GCC diagnostic ignored "-Wimplicit-fallthrough"
-#endif
 #include <SDL.h>
-#ifdef __GNUC__
-#pragma GCC diagnostic pop
-#endif
 
 #include "common/common_types.h"
 #include "common/threadsafe_queue.h"

src/input_common/touch_from_button.cpp

@@ -13,7 +13,7 @@ class TouchFromButtonDevice final : public Input::TouchDevice {
 public:
     TouchFromButtonDevice() {
         const auto button_index =
-            static_cast<std::size_t>(Settings::values.touch_from_button_map_index);
+            static_cast<u64>(Settings::values.touch_from_button_map_index.GetValue());
         const auto& buttons = Settings::values.touch_from_button_maps[button_index].buttons;
 
         for (const auto& config_entry : buttons) {

src/input_common/udp/client.cpp

@@ -201,7 +201,7 @@ bool Client::DeviceConnected(std::size_t pad) const {
 void Client::ReloadSockets() {
     Reset();
 
-    std::stringstream servers_ss(Settings::values.udp_input_servers);
+    std::stringstream servers_ss(static_cast<std::string>(Settings::values.udp_input_servers));
     std::string server_token;
     std::size_t client = 0;
     while (std::getline(servers_ss, server_token, ',')) {
@@ -370,7 +370,7 @@ std::optional<std::size_t> Client::GetUnusedFingerID() const {
 
 void Client::UpdateTouchInput(Response::TouchPad& touch_pad, std::size_t client, std::size_t id) {
     // TODO: Use custom calibration per device
-    const Common::ParamPackage touch_param(Settings::values.touch_device);
+    const Common::ParamPackage touch_param(Settings::values.touch_device.GetValue());
     const u16 min_x = static_cast<u16>(touch_param.Get("min_x", 100));
     const u16 min_y = static_cast<u16>(touch_param.Get("min_y", 50));
     const u16 max_x = static_cast<u16>(touch_param.Get("max_x", 1800));

src/tests/CMakeLists.txt

@@ -5,6 +5,7 @@ add_executable(tests
     common/host_memory.cpp
     common/param_package.cpp
     common/ring_buffer.cpp
+    common/unique_function.cpp
     core/core_timing.cpp
     core/network/network.cpp
     tests.cpp

src/tests/common/unique_function.cpp

@@ -0,0 +1,108 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <string>
+
+#include <catch2/catch.hpp>
+
+#include "common/unique_function.h"
+
+namespace {
+struct Noisy {
+    Noisy() : state{"Default constructed"} {}
+    Noisy(Noisy&& rhs) noexcept : state{"Move constructed"} {
+        rhs.state = "Moved away";
+    }
+    Noisy& operator=(Noisy&& rhs) noexcept {
+        state = "Move assigned";
+        rhs.state = "Moved away";
+    }
+    Noisy(const Noisy&) : state{"Copied constructed"} {}
+    Noisy& operator=(const Noisy&) {
+        state = "Copied assigned";
+    }
+
+    std::string state;
+};
+} // Anonymous namespace
+
+TEST_CASE("UniqueFunction", "[common]") {
+    SECTION("Capture reference") {
+        int value = 0;
+        Common::UniqueFunction<void> func = [&value] { value = 5; };
+        func();
+        REQUIRE(value == 5);
+    }
+    SECTION("Capture pointer") {
+        int value = 0;
+        int* pointer = &value;
+        Common::UniqueFunction<void> func = [pointer] { *pointer = 5; };
+        func();
+        REQUIRE(value == 5);
+    }
+    SECTION("Move object") {
+        Noisy noisy;
+        REQUIRE(noisy.state == "Default constructed");
+
+        Common::UniqueFunction<void> func = [noisy = std::move(noisy)] {
+            REQUIRE(noisy.state == "Move constructed");
+        };
+        REQUIRE(noisy.state == "Moved away");
+        func();
+    }
+    SECTION("Move construct function") {
+        int value = 0;
+        Common::UniqueFunction<void> func = [&value] { value = 5; };
+        Common::UniqueFunction<void> new_func = std::move(func);
+        new_func();
+        REQUIRE(value == 5);
+    }
+    SECTION("Move assign function") {
+        int value = 0;
+        Common::UniqueFunction<void> func = [&value] { value = 5; };
+        Common::UniqueFunction<void> new_func;
+        new_func = std::move(func);
+        new_func();
+        REQUIRE(value == 5);
+    }
+    SECTION("Default construct then assign function") {
+        int value = 0;
+        Common::UniqueFunction<void> func;
+        func = [&value] { value = 5; };
+        func();
+        REQUIRE(value == 5);
+    }
+    SECTION("Pass arguments") {
+        int result = 0;
+        Common::UniqueFunction<void, int, int> func = [&result](int a, int b) { result = a + b; };
+        func(5, 4);
+        REQUIRE(result == 9);
+    }
+    SECTION("Pass arguments and return value") {
+        Common::UniqueFunction<int, int, int> func = [](int a, int b) { return a + b; };
+        REQUIRE(func(5, 4) == 9);
+    }
+    SECTION("Destructor") {
+        int num_destroyed = 0;
+        struct Foo {
+            Foo(int* num_) : num{num_} {}
+            Foo(Foo&& rhs) : num{std::exchange(rhs.num, nullptr)} {}
+            Foo(const Foo&) = delete;
+
+            ~Foo() {
+                if (num) {
+                    ++*num;
+                }
+            }
+
+            int* num = nullptr;
+        };
+        Foo object{&num_destroyed};
+        {
+            Common::UniqueFunction<void> func = [object = std::move(object)] {};
+            REQUIRE(num_destroyed == 0);
+        }
+        REQUIRE(num_destroyed == 1);
+    }
+}

src/tests/video_core/buffer_base.cpp

@@ -536,7 +536,7 @@ TEST_CASE("BufferBase: Cached write downloads") {
     REQUIRE(rasterizer.Count() == 63);
     buffer.MarkRegionAsGpuModified(c + PAGE, PAGE);
     int num = 0;
-    buffer.ForEachDownloadRange(c, WORD, [&](u64 offset, u64 size) { ++num; });
+    buffer.ForEachDownloadRangeAndClear(c, WORD, [&](u64 offset, u64 size) { ++num; });
     buffer.ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) { ++num; });
     REQUIRE(num == 0);
     REQUIRE(!buffer.IsRegionCpuModified(c + PAGE, PAGE));

src/video_core/buffer_cache/buffer_base.h

@@ -226,19 +226,24 @@ public:
     /// Call 'func' for each CPU modified range and unmark those pages as CPU modified
     template <typename Func>
     void ForEachUploadRange(VAddr query_cpu_range, u64 size, Func&& func) {
-        ForEachModifiedRange<Type::CPU>(query_cpu_range, size, func);
+        ForEachModifiedRange<Type::CPU>(query_cpu_range, size, true, func);
     }
 
     /// Call 'func' for each GPU modified range and unmark those pages as GPU modified
     template <typename Func>
-    void ForEachDownloadRange(VAddr query_cpu_range, u64 size, Func&& func) {
-        ForEachModifiedRange<Type::GPU>(query_cpu_range, size, func);
+    void ForEachDownloadRange(VAddr query_cpu_range, u64 size, bool clear, Func&& func) {
+        ForEachModifiedRange<Type::GPU>(query_cpu_range, size, clear, func);
+    }
+
+    template <typename Func>
+    void ForEachDownloadRangeAndClear(VAddr query_cpu_range, u64 size, Func&& func) {
+        ForEachModifiedRange<Type::GPU>(query_cpu_range, size, true, func);
     }
 
     /// Call 'func' for each GPU modified range and unmark those pages as GPU modified
     template <typename Func>
     void ForEachDownloadRange(Func&& func) {
-        ForEachModifiedRange<Type::GPU>(cpu_addr, SizeBytes(), func);
+        ForEachModifiedRange<Type::GPU>(cpu_addr, SizeBytes(), true, func);
     }
 
     /// Mark buffer as picked
@@ -415,7 +420,7 @@ private:
      * @param func            Function to call for each turned off region
      */
     template <Type type, typename Func>
-    void ForEachModifiedRange(VAddr query_cpu_range, s64 size, Func&& func) {
+    void ForEachModifiedRange(VAddr query_cpu_range, s64 size, bool clear, Func&& func) {
         static_assert(type != Type::Untracked);
 
         const s64 difference = query_cpu_range - cpu_addr;
@@ -467,7 +472,9 @@ private:
             bits = (bits << left_offset) >> left_offset;
 
             const u64 current_word = state_words[word_index] & bits;
-            state_words[word_index] &= ~bits;
+            if (clear) {
+                state_words[word_index] &= ~bits;
+            }
 
             if constexpr (type == Type::CPU) {
                 const u64 current_bits = untracked_words[word_index] & bits;

src/video_core/buffer_cache/buffer_cache.h

@@ -15,6 +15,7 @@
 #include <vector>
 
 #include <boost/container/small_vector.hpp>
+#include <boost/icl/interval_set.hpp>
 
 #include "common/common_types.h"
 #include "common/div_ceil.h"
@@ -77,6 +78,9 @@ class BufferCache {
     using Runtime = typename P::Runtime;
     using Buffer = typename P::Buffer;
 
+    using IntervalSet = boost::icl::interval_set<VAddr>;
+    using IntervalType = typename IntervalSet::interval_type;
+
     struct Empty {};
 
     struct OverlapResult {
@@ -148,18 +152,26 @@ public:
     /// Return true when there are uncommitted buffers to be downloaded
     [[nodiscard]] bool HasUncommittedFlushes() const noexcept;
 
+    void AccumulateFlushes();
+
     /// Return true when the caller should wait for async downloads
     [[nodiscard]] bool ShouldWaitAsyncFlushes() const noexcept;
 
     /// Commit asynchronous downloads
     void CommitAsyncFlushes();
+    void CommitAsyncFlushesHigh();
 
     /// Pop asynchronous downloads
     void PopAsyncFlushes();
 
+    [[nodiscard]] bool DMACopy(GPUVAddr src_address, GPUVAddr dest_address, u64 amount);
+
     /// Return true when a CPU region is modified from the GPU
     [[nodiscard]] bool IsRegionGpuModified(VAddr addr, size_t size);
 
+    /// Return true when a CPU region is modified from the CPU
+    [[nodiscard]] bool IsRegionCpuModified(VAddr addr, size_t size);
+
     std::mutex mutex;
 
 private:
@@ -190,6 +202,36 @@ private:
         }
     }
 
+    template <typename Func>
+    void ForEachWrittenRange(VAddr cpu_addr, u64 size, Func&& func) {
+        const VAddr start_address = cpu_addr;
+        const VAddr end_address = start_address + size;
+        const VAddr search_base =
+            static_cast<VAddr>(std::min<s64>(0LL, static_cast<s64>(start_address - size)));
+        const IntervalType search_interval{search_base, search_base + 1};
+        auto it = common_ranges.lower_bound(search_interval);
+        if (it == common_ranges.end()) {
+            it = common_ranges.begin();
+        }
+        for (; it != common_ranges.end(); it++) {
+            VAddr inter_addr_end = it->upper();
+            VAddr inter_addr = it->lower();
+            if (inter_addr >= end_address) {
+                break;
+            }
+            if (inter_addr_end <= start_address) {
+                continue;
+            }
+            if (inter_addr_end > end_address) {
+                inter_addr_end = end_address;
+            }
+            if (inter_addr < start_address) {
+                inter_addr = start_address;
+            }
+            func(inter_addr, inter_addr_end);
+        }
+    }
+
     static bool IsRangeGranular(VAddr cpu_addr, size_t size) {
         return (cpu_addr & ~Core::Memory::PAGE_MASK) ==
                ((cpu_addr + size) & ~Core::Memory::PAGE_MASK);
@@ -272,8 +314,6 @@ private:
 
     void DeleteBuffer(BufferId buffer_id);
 
-    void ReplaceBufferDownloads(BufferId old_buffer_id, BufferId new_buffer_id);
-
     void NotifyBufferDeletion();
 
     [[nodiscard]] Binding StorageBufferBinding(GPUVAddr ssbo_addr) const;
@@ -327,9 +367,9 @@ private:
 
     std::vector<BufferId> cached_write_buffer_ids;
 
-    // TODO: This data structure is not optimal and it should be reworked
-    std::vector<BufferId> uncommitted_downloads;
-    std::deque<std::vector<BufferId>> committed_downloads;
+    IntervalSet uncommitted_ranges;
+    IntervalSet common_ranges;
+    std::deque<IntervalSet> committed_ranges;
 
     size_t immediate_buffer_capacity = 0;
     std::unique_ptr<u8[]> immediate_buffer_alloc;
@@ -352,6 +392,7 @@ BufferCache<P>::BufferCache(VideoCore::RasterizerInterface& rasterizer_,
     // Ensure the first slot is used for the null buffer
     void(slot_buffers.insert(runtime, NullBufferParams{}));
     deletion_iterator = slot_buffers.end();
+    common_ranges.clear();
 }
 
 template <class P>
@@ -421,6 +462,68 @@ void BufferCache<P>::DownloadMemory(VAddr cpu_addr, u64 size) {
     });
 }
 
+template <class P>
+bool BufferCache<P>::DMACopy(GPUVAddr src_address, GPUVAddr dest_address, u64 amount) {
+    const std::optional<VAddr> cpu_src_address = gpu_memory.GpuToCpuAddress(src_address);
+    const std::optional<VAddr> cpu_dest_address = gpu_memory.GpuToCpuAddress(dest_address);
+    if (!cpu_src_address || !cpu_dest_address) {
+        return false;
+    }
+    const bool source_dirty = IsRegionGpuModified(*cpu_src_address, amount);
+    const bool dest_dirty = IsRegionGpuModified(*cpu_dest_address, amount);
+    if (!source_dirty && !dest_dirty) {
+        return false;
+    }
+
+    const IntervalType subtract_interval{*cpu_dest_address, *cpu_dest_address + amount};
+    uncommitted_ranges.subtract(subtract_interval);
+    for (auto& interval_set : committed_ranges) {
+        interval_set.subtract(subtract_interval);
+    }
+
+    BufferId buffer_a;
+    BufferId buffer_b;
+    do {
+        has_deleted_buffers = false;
+        buffer_a = FindBuffer(*cpu_src_address, static_cast<u32>(amount));
+        buffer_b = FindBuffer(*cpu_dest_address, static_cast<u32>(amount));
+    } while (has_deleted_buffers);
+    auto& src_buffer = slot_buffers[buffer_a];
+    auto& dest_buffer = slot_buffers[buffer_b];
+    SynchronizeBuffer(src_buffer, *cpu_src_address, static_cast<u32>(amount));
+    SynchronizeBuffer(dest_buffer, *cpu_dest_address, static_cast<u32>(amount));
+    std::array copies{BufferCopy{
+        .src_offset = src_buffer.Offset(*cpu_src_address),
+        .dst_offset = dest_buffer.Offset(*cpu_dest_address),
+        .size = amount,
+    }};
+
+    boost::container::small_vector<IntervalType, 4> tmp_intervals;
+    auto mirror = [&](VAddr base_address, VAddr base_address_end) {
+        const u64 size = base_address_end - base_address;
+        const VAddr diff = base_address - *cpu_src_address;
+        const VAddr new_base_address = *cpu_dest_address + diff;
+        const IntervalType add_interval{new_base_address, new_base_address + size};
+        uncommitted_ranges.add(add_interval);
+        tmp_intervals.push_back(add_interval);
+    };
+    ForEachWrittenRange(*cpu_src_address, amount, mirror);
+    // This subtraction in this order is important for overlapping copies.
+    common_ranges.subtract(subtract_interval);
+    for (const IntervalType add_interval : tmp_intervals) {
+        common_ranges.add(add_interval);
+    }
+
+    runtime.CopyBuffer(dest_buffer, src_buffer, copies);
+    if (source_dirty) {
+        dest_buffer.MarkRegionAsGpuModified(*cpu_dest_address, amount);
+    }
+    std::vector<u8> tmp_buffer(amount);
+    cpu_memory.ReadBlockUnsafe(*cpu_src_address, tmp_buffer.data(), amount);
+    cpu_memory.WriteBlockUnsafe(*cpu_dest_address, tmp_buffer.data(), amount);
+    return true;
+}
+
 template <class P>
 void BufferCache<P>::BindGraphicsUniformBuffer(size_t stage, u32 index, GPUVAddr gpu_addr,
                                                u32 size) {
@@ -547,29 +650,30 @@ void BufferCache<P>::FlushCachedWrites() {
 
 template <class P>
 bool BufferCache<P>::HasUncommittedFlushes() const noexcept {
-    return !uncommitted_downloads.empty();
+    return !uncommitted_ranges.empty() || !committed_ranges.empty();
+}
+
+template <class P>
+void BufferCache<P>::AccumulateFlushes() {
+    if (Settings::values.gpu_accuracy.GetValue() != Settings::GPUAccuracy::High) {
+        uncommitted_ranges.clear();
+        return;
+    }
+    if (uncommitted_ranges.empty()) {
+        return;
+    }
+    committed_ranges.emplace_back(std::move(uncommitted_ranges));
 }
 
 template <class P>
 bool BufferCache<P>::ShouldWaitAsyncFlushes() const noexcept {
-    return !committed_downloads.empty() && !committed_downloads.front().empty();
+    return false;
 }
 
 template <class P>
-void BufferCache<P>::CommitAsyncFlushes() {
-    // This is intentionally passing the value by copy
-    committed_downloads.push_front(uncommitted_downloads);
-    uncommitted_downloads.clear();
-}
-
-template <class P>
-void BufferCache<P>::PopAsyncFlushes() {
-    if (committed_downloads.empty()) {
-        return;
-    }
-    auto scope_exit_pop_download = detail::ScopeExit([this] { committed_downloads.pop_back(); });
-    const std::span<const BufferId> download_ids = committed_downloads.back();
-    if (download_ids.empty()) {
+void BufferCache<P>::CommitAsyncFlushesHigh() {
+    AccumulateFlushes();
+    if (committed_ranges.empty()) {
         return;
     }
     MICROPROFILE_SCOPE(GPU_DownloadMemory);
@@ -577,20 +681,43 @@ void BufferCache<P>::PopAsyncFlushes() {
     boost::container::small_vector<std::pair<BufferCopy, BufferId>, 1> downloads;
     u64 total_size_bytes = 0;
     u64 largest_copy = 0;
-    for (const BufferId buffer_id : download_ids) {
-        slot_buffers[buffer_id].ForEachDownloadRange([&](u64 range_offset, u64 range_size) {
-            downloads.push_back({
-                BufferCopy{
-                    .src_offset = range_offset,
-                    .dst_offset = total_size_bytes,
-                    .size = range_size,
-                },
-                buffer_id,
+    for (const IntervalSet& intervals : committed_ranges) {
+        for (auto& interval : intervals) {
+            const std::size_t size = interval.upper() - interval.lower();
+            const VAddr cpu_addr = interval.lower();
+            ForEachBufferInRange(cpu_addr, size, [&](BufferId buffer_id, Buffer& buffer) {
+                boost::container::small_vector<BufferCopy, 1> copies;
+                buffer.ForEachDownloadRangeAndClear(
+                    cpu_addr, size, [&](u64 range_offset, u64 range_size) {
+                        const VAddr buffer_addr = buffer.CpuAddr();
+                        const auto add_download = [&](VAddr start, VAddr end) {
+                            const u64 new_offset = start - buffer_addr;
+                            const u64 new_size = end - start;
+                            downloads.push_back({
+                                BufferCopy{
+                                    .src_offset = new_offset,
+                                    .dst_offset = total_size_bytes,
+                                    .size = new_size,
+                                },
+                                buffer_id,
+                            });
+                            // Align up to avoid cache conflicts
+                            constexpr u64 align = 256ULL;
+                            constexpr u64 mask = ~(align - 1ULL);
+                            total_size_bytes += (new_size + align - 1) & mask;
+                            largest_copy = std::max(largest_copy, new_size);
+                        };
+
+                        const VAddr start_address = buffer_addr + range_offset;
+                        const VAddr end_address = start_address + range_size;
+                        ForEachWrittenRange(start_address, range_size, add_download);
+                        const IntervalType subtract_interval{start_address, end_address};
+                        common_ranges.subtract(subtract_interval);
+                    });
             });
-            total_size_bytes += range_size;
-            largest_copy = std::max(largest_copy, range_size);
-        });
+        }
     }
+    committed_ranges.clear();
     if (downloads.empty()) {
         return;
     }
@@ -622,6 +749,19 @@ void BufferCache<P>::PopAsyncFlushes() {
     }
 }
 
+template <class P>
+void BufferCache<P>::CommitAsyncFlushes() {
+    if (Settings::values.gpu_accuracy.GetValue() == Settings::GPUAccuracy::High) {
+        CommitAsyncFlushesHigh();
+    } else {
+        uncommitted_ranges.clear();
+        committed_ranges.clear();
+    }
+}
+
+template <class P>
+void BufferCache<P>::PopAsyncFlushes() {}
+
 template <class P>
 bool BufferCache<P>::IsRegionGpuModified(VAddr addr, size_t size) {
     const u64 page_end = Common::DivCeil(addr + size, PAGE_SIZE);
@@ -641,6 +781,25 @@ bool BufferCache<P>::IsRegionGpuModified(VAddr addr, size_t size) {
     return false;
 }
 
+template <class P>
+bool BufferCache<P>::IsRegionCpuModified(VAddr addr, size_t size) {
+    const u64 page_end = Common::DivCeil(addr + size, PAGE_SIZE);
+    for (u64 page = addr >> PAGE_BITS; page < page_end;) {
+        const BufferId image_id = page_table[page];
+        if (!image_id) {
+            ++page;
+            continue;
+        }
+        Buffer& buffer = slot_buffers[image_id];
+        if (buffer.IsRegionCpuModified(addr, size)) {
+            return true;
+        }
+        const VAddr end_addr = buffer.CpuAddr() + buffer.SizeBytes();
+        page = Common::DivCeil(end_addr, PAGE_SIZE);
+    }
+    return false;
+}
+
 template <class P>
 void BufferCache<P>::BindHostIndexBuffer() {
     Buffer& buffer = slot_buffers[index_buffer.buffer_id];
@@ -649,7 +808,9 @@ void BufferCache<P>::BindHostIndexBuffer() {
     const u32 size = index_buffer.size;
     SynchronizeBuffer(buffer, index_buffer.cpu_addr, size);
     if constexpr (HAS_FULL_INDEX_AND_PRIMITIVE_SUPPORT) {
-        runtime.BindIndexBuffer(buffer, offset, size);
+        const u32 new_offset = offset + maxwell3d.regs.index_array.first *
+                                            maxwell3d.regs.index_array.FormatSizeInBytes();
+        runtime.BindIndexBuffer(buffer, new_offset, size);
     } else {
         runtime.BindIndexBuffer(maxwell3d.regs.draw.topology, maxwell3d.regs.index_array.format,
                                 maxwell3d.regs.index_array.first, maxwell3d.regs.index_array.count,
@@ -863,7 +1024,7 @@ void BufferCache<P>::UpdateIndexBuffer() {
     const GPUVAddr gpu_addr_end = index_array.EndAddress();
     const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr_begin);
     const u32 address_size = static_cast<u32>(gpu_addr_end - gpu_addr_begin);
-    const u32 draw_size = index_array.count * index_array.FormatSizeInBytes();
+    const u32 draw_size = (index_array.count + index_array.first) * index_array.FormatSizeInBytes();
     const u32 size = std::min(address_size, draw_size);
     if (size == 0 || !cpu_addr) {
         index_buffer = NULL_BINDING;
@@ -1010,16 +1171,16 @@ void BufferCache<P>::MarkWrittenBuffer(BufferId buffer_id, VAddr cpu_addr, u32 s
     Buffer& buffer = slot_buffers[buffer_id];
     buffer.MarkRegionAsGpuModified(cpu_addr, size);
 
-    const bool is_accuracy_high = Settings::IsGPULevelHigh();
+    const IntervalType base_interval{cpu_addr, cpu_addr + size};
+    common_ranges.add(base_interval);
+
+    const bool is_accuracy_high =
+        Settings::values.gpu_accuracy.GetValue() == Settings::GPUAccuracy::High;
     const bool is_async = Settings::values.use_asynchronous_gpu_emulation.GetValue();
-    if (!is_accuracy_high || !is_async) {
+    if (!is_async && !is_accuracy_high) {
         return;
     }
-    if (std::ranges::find(uncommitted_downloads, buffer_id) != uncommitted_downloads.end()) {
-        // Already inserted
-        return;
-    }
-    uncommitted_downloads.push_back(buffer_id);
+    uncommitted_ranges.add(base_interval);
 }
 
 template <class P>
@@ -1103,7 +1264,6 @@ void BufferCache<P>::JoinOverlap(BufferId new_buffer_id, BufferId overlap_id,
     if (!copies.empty()) {
         runtime.CopyBuffer(slot_buffers[new_buffer_id], overlap, copies);
     }
-    ReplaceBufferDownloads(overlap_id, new_buffer_id);
     DeleteBuffer(overlap_id);
 }
 
@@ -1244,14 +1404,28 @@ void BufferCache<P>::DownloadBufferMemory(Buffer& buffer, VAddr cpu_addr, u64 si
     boost::container::small_vector<BufferCopy, 1> copies;
     u64 total_size_bytes = 0;
     u64 largest_copy = 0;
-    buffer.ForEachDownloadRange(cpu_addr, size, [&](u64 range_offset, u64 range_size) {
-        copies.push_back(BufferCopy{
-            .src_offset = range_offset,
-            .dst_offset = total_size_bytes,
-            .size = range_size,
-        });
-        total_size_bytes += range_size;
-        largest_copy = std::max(largest_copy, range_size);
+    buffer.ForEachDownloadRangeAndClear(cpu_addr, size, [&](u64 range_offset, u64 range_size) {
+        const VAddr buffer_addr = buffer.CpuAddr();
+        const auto add_download = [&](VAddr start, VAddr end) {
+            const u64 new_offset = start - buffer_addr;
+            const u64 new_size = end - start;
+            copies.push_back(BufferCopy{
+                .src_offset = new_offset,
+                .dst_offset = total_size_bytes,
+                .size = new_size,
+            });
+            // Align up to avoid cache conflicts
+            constexpr u64 align = 256ULL;
+            constexpr u64 mask = ~(align - 1ULL);
+            total_size_bytes += (new_size + align - 1) & mask;
+            largest_copy = std::max(largest_copy, new_size);
+        };
+
+        const VAddr start_address = buffer_addr + range_offset;
+        const VAddr end_address = start_address + range_size;
+        ForEachWrittenRange(start_address, range_size, add_download);
+        const IntervalType subtract_interval{start_address, end_address};
+        common_ranges.subtract(subtract_interval);
     });
     if (total_size_bytes == 0) {
         return;
@@ -1315,18 +1489,6 @@ void BufferCache<P>::DeleteBuffer(BufferId buffer_id) {
     NotifyBufferDeletion();
 }
 
-template <class P>
-void BufferCache<P>::ReplaceBufferDownloads(BufferId old_buffer_id, BufferId new_buffer_id) {
-    const auto replace = [old_buffer_id, new_buffer_id](std::vector<BufferId>& buffers) {
-        std::ranges::replace(buffers, old_buffer_id, new_buffer_id);
-        if (auto it = std::ranges::find(buffers, new_buffer_id); it != buffers.end()) {
-            buffers.erase(std::remove(it + 1, buffers.end(), new_buffer_id), buffers.end());
-        }
-    };
-    replace(uncommitted_downloads);
-    std::ranges::for_each(committed_downloads, replace);
-}
-
 template <class P>
 void BufferCache<P>::NotifyBufferDeletion() {
     if constexpr (HAS_PERSISTENT_UNIFORM_BUFFER_BINDINGS) {
@@ -1349,15 +1511,9 @@ typename BufferCache<P>::Binding BufferCache<P>::StorageBufferBinding(GPUVAddr s
     if (!cpu_addr || size == 0) {
         return NULL_BINDING;
     }
-    // HACK(Rodrigo): This is the number of bytes bound in host beyond the guest API's range.
-    // It exists due to some games like Astral Chain operate out of bounds.
-    // Binding the whole map range would be technically correct, but games have large maps that make
-    // this approach unaffordable for now.
-    static constexpr u32 arbitrary_extra_bytes = 0xc000;
-    const u32 bytes_to_map_end = static_cast<u32>(gpu_memory.BytesToMapEnd(gpu_addr));
     const Binding binding{
         .cpu_addr = *cpu_addr,
-        .size = std::min(size + arbitrary_extra_bytes, bytes_to_map_end),
+        .size = size,
         .buffer_id = BufferId{},
     };
     return binding;