audio_core: Accept Audren REV8

According to Ryujinx, REV8 only added changes on Performance buffer and Wavebuffer DSP command generation.

As we don't support any of those, we can just increment the revision number for now.

externals/microprofile/microprofile.h

@@ -828,7 +828,7 @@ inline MicroProfileLogEntry MicroProfileMakeLogIndex(uint64_t nBegin, MicroProfi
     MicroProfileLogEntry Entry =  (nBegin<<62) | ((0x3fff&nToken)<<48) | (MP_LOG_TICK_MASK&nTick);
     int t = MicroProfileLogType(Entry);
     uint64_t nTimerIndex = MicroProfileLogTimerIndex(Entry);
-    MP_ASSERT(t == nBegin);
+    MP_ASSERT((uint64_t)t == nBegin);
     MP_ASSERT(nTimerIndex == (nToken&0x3fff));
     return Entry;
 
@@ -1556,10 +1556,10 @@ void MicroProfileFlip()
 
         pFramePut->nFrameStartCpu = MP_TICK();
         pFramePut->nFrameStartGpu = (uint32_t)MicroProfileGpuInsertTimeStamp();
-        if(pFrameNext->nFrameStartGpu != (uint64_t)-1)
+        if(pFrameNext->nFrameStartGpu != -1)
             pFrameNext->nFrameStartGpu = MicroProfileGpuGetTimeStamp((uint32_t)pFrameNext->nFrameStartGpu);
 
-        if(pFrameCurrent->nFrameStartGpu == (uint64_t)-1)
+        if(pFrameCurrent->nFrameStartGpu == -1)
             pFrameCurrent->nFrameStartGpu = pFrameNext->nFrameStartGpu + 1;
 
         uint64_t nFrameStartCpu = pFrameCurrent->nFrameStartCpu;

src/audio_core/audio_renderer.h

@@ -189,7 +189,7 @@ struct UpdateDataHeader {
     UpdateDataHeader() {}
 
     explicit UpdateDataHeader(const AudioRendererParameter& config) {
-        revision = Common::MakeMagic('R', 'E', 'V', '4'); // 5.1.0 Revision
+        revision = Common::MakeMagic('R', 'E', 'V', '8'); // 9.2.0 Revision
         behavior_size = 0xb0;
         memory_pools_size = (config.effect_count + (config.voice_count * 4)) * 0x10;
         voices_size = config.voice_count * 0x10;

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

@@ -111,6 +111,14 @@ void SET::GetLanguageCode(Kernel::HLERequestContext& ctx) {
     rb.PushEnum(available_language_codes[Settings::values.language_index]);
 }
 
+void SET::GetRegionCode(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_SET, "called");
+
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push(Settings::values.region_index);
+}
+
 SET::SET() : ServiceFramework("set") {
     // clang-format off
     static const FunctionInfo functions[] = {
@@ -118,7 +126,7 @@ SET::SET() : ServiceFramework("set") {
         {1, &SET::GetAvailableLanguageCodes, "GetAvailableLanguageCodes"},
         {2, &SET::MakeLanguageCode, "MakeLanguageCode"},
         {3, &SET::GetAvailableLanguageCodeCount, "GetAvailableLanguageCodeCount"},
-        {4, nullptr, "GetRegionCode"},
+        {4, &SET::GetRegionCode, "GetRegionCode"},
         {5, &SET::GetAvailableLanguageCodes2, "GetAvailableLanguageCodes2"},
         {6, &SET::GetAvailableLanguageCodeCount2, "GetAvailableLanguageCodeCount2"},
         {7, nullptr, "GetKeyCodeMap"},

src/core/hle/service/set/set.h

@@ -43,6 +43,7 @@ private:
     void GetAvailableLanguageCodeCount(Kernel::HLERequestContext& ctx);
     void GetAvailableLanguageCodeCount2(Kernel::HLERequestContext& ctx);
     void GetQuestFlag(Kernel::HLERequestContext& ctx);
+    void GetRegionCode(Kernel::HLERequestContext& ctx);
 };
 
 } // namespace Service::Set

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

@@ -53,7 +53,7 @@ static std::vector<std::string> BuildLocationNameCache(Core::System& system) {
         return {};
     }
 
-    std::vector<char> raw_data(binary_list->GetSize());
+    std::vector<char> raw_data(binary_list->GetSize() + 1);
     binary_list->ReadBytes<char>(raw_data.data(), binary_list->GetSize());
 
     std::stringstream data_stream{raw_data.data()};

src/core/settings.cpp

@@ -86,6 +86,7 @@ void LogSettings() {
     LogSetting("System_RngSeed", Settings::values.rng_seed.value_or(0));
     LogSetting("System_CurrentUser", Settings::values.current_user);
     LogSetting("System_LanguageIndex", Settings::values.language_index);
+    LogSetting("System_RegionIndex", Settings::values.region_index);
     LogSetting("Core_UseMultiCore", Settings::values.use_multi_core);
     LogSetting("Renderer_UseResolutionFactor", Settings::values.resolution_factor);
     LogSetting("Renderer_UseFrameLimit", Settings::values.use_frame_limit);

src/core/settings.h

@@ -387,6 +387,8 @@ struct Values {
 
     s32 current_user;
     s32 language_index;
+    s32 region_index;
+    s32 sound_index;
 
     // Controls
     std::array<PlayerInput, 10> players;

src/input_common/udp/client.cpp

@@ -35,7 +35,7 @@ public:
           pad_index(pad_index) {
         boost::system::error_code ec{};
         auto ipv4 = boost::asio::ip::make_address_v4(host, ec);
-        if (ec.failed()) {
+        if (ec.value() != boost::system::errc::success) {
             LOG_ERROR(Input, "Invalid IPv4 address \"{}\" provided to socket", host);
             ipv4 = boost::asio::ip::address_v4{};
         }

src/input_common/udp/udp.cpp

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include <mutex>
+#include <optional>
 #include <tuple>
 
 #include "common/param_package.h"
@@ -44,7 +45,7 @@ public:
     std::unique_ptr<Input::TouchDevice> Create(const Common::ParamPackage& params) override {
         {
             std::lock_guard guard(status->update_mutex);
-            status->touch_calibration.emplace();
+            status->touch_calibration = DeviceStatus::CalibrationData{};
             // These default values work well for DS4 but probably not other touch inputs
             status->touch_calibration->min_x = params.Get("min_x", 100);
             status->touch_calibration->min_y = params.Get("min_y", 50);

src/video_core/dirty_flags.cpp

@@ -15,14 +15,6 @@ namespace VideoCommon::Dirty {
 
 using Tegra::Engines::Maxwell3D;
 
-void SetupCommonOnWriteStores(Tegra::Engines::Maxwell3D::DirtyState::Flags& store) {
-    store[RenderTargets] = true;
-    store[ZetaBuffer] = true;
-    for (std::size_t i = 0; i < Maxwell3D::Regs::NumRenderTargets; ++i) {
-        store[ColorBuffer0 + i] = true;
-    }
-}
-
 void SetupDirtyRenderTargets(Tegra::Engines::Maxwell3D::DirtyState::Tables& tables) {
     static constexpr std::size_t num_per_rt = NUM(rt[0]);
     static constexpr std::size_t begin = OFF(rt);

src/video_core/dirty_flags.h

@@ -44,8 +44,6 @@ void FillBlock(Tegra::Engines::Maxwell3D::DirtyState::Tables& tables, std::size_
     FillBlock(tables[1], begin, num, index_b);
 }
 
-void SetupCommonOnWriteStores(Tegra::Engines::Maxwell3D::DirtyState::Flags& store);
-
 void SetupDirtyRenderTargets(Tegra::Engines::Maxwell3D::DirtyState::Tables& tables);
 
 } // namespace VideoCommon::Dirty

src/video_core/engines/kepler_compute.cpp

@@ -119,14 +119,6 @@ Texture::TICEntry KeplerCompute::GetTICEntry(u32 tic_index) const {
     Texture::TICEntry tic_entry;
     memory_manager.ReadBlockUnsafe(tic_address_gpu, &tic_entry, sizeof(Texture::TICEntry));
 
-    const auto r_type{tic_entry.r_type.Value()};
-    const auto g_type{tic_entry.g_type.Value()};
-    const auto b_type{tic_entry.b_type.Value()};
-    const auto a_type{tic_entry.a_type.Value()};
-
-    // TODO(Subv): Different data types for separate components are not supported
-    DEBUG_ASSERT(r_type == g_type && r_type == b_type && r_type == a_type);
-
     return tic_entry;
 }
 

src/video_core/engines/shader_bytecode.h

@@ -911,14 +911,9 @@ union Instruction {
     } fadd32i;
 
     union {
-        BitField<20, 8, u64> shift_position;
-        BitField<28, 8, u64> shift_length;
-        BitField<48, 1, u64> negate_b;
-        BitField<49, 1, u64> negate_a;
-
-        u64 GetLeftShiftValue() const {
-            return 32 - (shift_position + shift_length);
-        }
+        BitField<40, 1, u64> brev;
+        BitField<47, 1, u64> rd_cc;
+        BitField<48, 1, u64> is_signed;
     } bfe;
 
     union {

src/video_core/gpu.h

@@ -39,6 +39,7 @@ enum class RenderTargetFormat : u32 {
     RGBA32_FLOAT = 0xC0,
     RGBA32_UINT = 0xC2,
     RGBA16_UNORM = 0xC6,
+    RGBA16_SNORM = 0xC7,
     RGBA16_UINT = 0xC9,
     RGBA16_FLOAT = 0xCA,
     RG32_FLOAT = 0xCB,

src/video_core/morton.cpp

@@ -51,6 +51,7 @@ static constexpr ConversionArray morton_to_linear_fns = {
     MortonCopy<true, PixelFormat::R8UI>,
     MortonCopy<true, PixelFormat::RGBA16F>,
     MortonCopy<true, PixelFormat::RGBA16U>,
+    MortonCopy<true, PixelFormat::RGBA16S>,
     MortonCopy<true, PixelFormat::RGBA16UI>,
     MortonCopy<true, PixelFormat::R11FG11FB10F>,
     MortonCopy<true, PixelFormat::RGBA32UI>,
@@ -131,6 +132,7 @@ static constexpr ConversionArray linear_to_morton_fns = {
     MortonCopy<false, PixelFormat::R8U>,
     MortonCopy<false, PixelFormat::R8UI>,
     MortonCopy<false, PixelFormat::RGBA16F>,
+    MortonCopy<false, PixelFormat::RGBA16S>,
     MortonCopy<false, PixelFormat::RGBA16U>,
     MortonCopy<false, PixelFormat::RGBA16UI>,
     MortonCopy<false, PixelFormat::R11FG11FB10F>,

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -93,10 +93,6 @@ void oglEnable(GLenum cap, bool state) {
     (state ? glEnable : glDisable)(cap);
 }
 
-void oglEnablei(GLenum cap, bool state, GLuint index) {
-    (state ? glEnablei : glDisablei)(cap, index);
-}
-
 } // Anonymous namespace
 
 RasterizerOpenGL::RasterizerOpenGL(Core::System& system, Core::Frontend::EmuWindow& emu_window,
@@ -478,7 +474,6 @@ void RasterizerOpenGL::Clear() {
 void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) {
     MICROPROFILE_SCOPE(OpenGL_Drawing);
     auto& gpu = system.GPU().Maxwell3D();
-    const auto& regs = gpu.regs;
 
     query_cache.UpdateCounters();
 
@@ -529,7 +524,7 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) {
     // Upload vertex and index data.
     SetupVertexBuffer();
     SetupVertexInstances();
-    GLintptr index_buffer_offset;
+    GLintptr index_buffer_offset = 0;
     if (is_indexed) {
         index_buffer_offset = SetupIndexBuffer();
     }
@@ -555,7 +550,7 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) {
     ConfigureFramebuffers();
 
     // Signal the buffer cache that we are not going to upload more things.
-    const bool invalidate = buffer_cache.Unmap();
+    buffer_cache.Unmap();
 
     // Now that we are no longer uploading data, we can safely bind the buffers to OpenGL.
     vertex_array_pushbuffer.Bind();

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -393,10 +393,6 @@ std::string FlowStackTopName(MetaStackClass stack) {
     return fmt::format("{}_flow_stack_top", GetFlowStackPrefix(stack));
 }
 
-[[deprecated]] constexpr bool IsVertexShader(ShaderType stage) {
-    return stage == ShaderType::Vertex;
-}
-
 struct GenericVaryingDescription {
     std::string name;
     u8 first_element = 0;
@@ -529,8 +525,9 @@ private:
     }
 
     void DeclareVertex() {
-        if (!IsVertexShader(stage))
+        if (stage != ShaderType::Vertex) {
             return;
+        }
 
         DeclareVertexRedeclarations();
     }
@@ -602,14 +599,14 @@ private:
                 break;
             }
         }
-        if (!IsVertexShader(stage) || device.HasVertexViewportLayer()) {
+        if (stage != ShaderType::Vertex || device.HasVertexViewportLayer()) {
             if (ir.UsesLayer()) {
                 code.AddLine("int gl_Layer;");
             }
             if (ir.UsesViewportIndex()) {
                 code.AddLine("int gl_ViewportIndex;");
             }
-        } else if ((ir.UsesLayer() || ir.UsesViewportIndex()) && IsVertexShader(stage) &&
+        } else if ((ir.UsesLayer() || ir.UsesViewportIndex()) && stage == ShaderType::Vertex &&
                    !device.HasVertexViewportLayer()) {
             LOG_ERROR(
                 Render_OpenGL,
@@ -1147,7 +1144,7 @@ private:
             // TODO(Subv): Find out what the values are for the first two elements when inside a
             // vertex shader, and what's the value of the fourth element when inside a Tess Eval
             // shader.
-            ASSERT(IsVertexShader(stage));
+            ASSERT(stage == ShaderType::Vertex);
             switch (element) {
             case 2:
                 // Config pack's first value is instance_id.
@@ -1218,12 +1215,12 @@ private:
                 UNIMPLEMENTED();
                 return {};
             case 1:
-                if (IsVertexShader(stage) && !device.HasVertexViewportLayer()) {
+                if (stage == ShaderType::Vertex && !device.HasVertexViewportLayer()) {
                     return {};
                 }
                 return {{"gl_Layer", Type::Int}};
             case 2:
-                if (IsVertexShader(stage) && !device.HasVertexViewportLayer()) {
+                if (stage == ShaderType::Vertex && !device.HasVertexViewportLayer()) {
                     return {};
                 }
                 return {{"gl_ViewportIndex", Type::Int}};
@@ -2009,16 +2006,19 @@ private:
         expr += GetSampler(meta->sampler);
         expr += ", ";
 
-        expr += constructors.at(operation.GetOperandsCount() - 1);
+        expr += constructors.at(operation.GetOperandsCount() + (meta->array ? 1 : 0) - 1);
         expr += '(';
         for (std::size_t i = 0; i < count; ++i) {
-            expr += VisitOperand(operation, i).AsInt();
-            const std::size_t next = i + 1;
-            if (next == count)
-                expr += ')';
-            else if (next < count)
+            if (i > 0) {
                 expr += ", ";
+            }
+            expr += VisitOperand(operation, i).AsInt();
         }
+        if (meta->array) {
+            expr += ", ";
+            expr += Visit(meta->array).AsInt();
+        }
+        expr += ')';
 
         if (meta->lod && !meta->sampler.IsBuffer()) {
             expr += ", ";
@@ -2529,7 +2529,7 @@ private:
     }
 
     u32 GetNumPhysicalInputAttributes() const {
-        return IsVertexShader(stage) ? GetNumPhysicalAttributes() : GetNumPhysicalVaryings();
+        return stage == ShaderType::Vertex ? GetNumPhysicalAttributes() : GetNumPhysicalVaryings();
     }
 
     u32 GetNumPhysicalAttributes() const {

src/video_core/renderer_opengl/gl_state_tracker.cpp

@@ -238,7 +238,6 @@ void StateTracker::Initialize() {
     SetupDirtyMisc(tables);
 
     auto& store = dirty.on_write_stores;
-    SetupCommonOnWriteStores(store);
     store[VertexBuffers] = true;
     for (std::size_t i = 0; i < Regs::NumVertexArrays; ++i) {
         store[VertexBuffer0 + i] = true;

src/video_core/renderer_opengl/gl_texture_cache.cpp

@@ -53,6 +53,7 @@ constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format
     {GL_R8UI, GL_RED_INTEGER, GL_UNSIGNED_BYTE, false},                             // R8UI
     {GL_RGBA16F, GL_RGBA, GL_HALF_FLOAT, false},                                    // RGBA16F
     {GL_RGBA16, GL_RGBA, GL_UNSIGNED_SHORT, false},                                 // RGBA16U
+    {GL_RGBA16_SNORM, GL_RGBA, GL_SHORT, false},                                    // RGBA16S
     {GL_RGBA16UI, GL_RGBA_INTEGER, GL_UNSIGNED_SHORT, false},                       // RGBA16UI
     {GL_R11F_G11F_B10F, GL_RGB, GL_UNSIGNED_INT_10F_11F_11F_REV, false},            // R11FG11FB10F
     {GL_RGBA32UI, GL_RGBA_INTEGER, GL_UNSIGNED_INT, false},                         // RGBA32UI

src/video_core/renderer_opengl/renderer_opengl.cpp

@@ -5,8 +5,11 @@
 #include <algorithm>
 #include <cstddef>
 #include <cstdlib>
+#include <cstring>
 #include <memory>
+
 #include <glad/glad.h>
+
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "common/microprofile.h"
@@ -25,6 +28,8 @@
 
 namespace OpenGL {
 
+namespace {
+
 // If the size of this is too small, it ends up creating a soft cap on FPS as the renderer will have
 // to wait on available presentation frames.
 constexpr std::size_t SWAP_CHAIN_SIZE = 3;
@@ -41,124 +46,6 @@ struct Frame {
     bool is_srgb{};                   /// Framebuffer is sRGB or RGB
 };
 
-/**
- * For smooth Vsync rendering, we want to always present the latest frame that the core generates,
- * but also make sure that rendering happens at the pace that the frontend dictates. This is a
- * helper class that the renderer uses to sync frames between the render thread and the presentation
- * thread
- */
-class FrameMailbox {
-public:
-    std::mutex swap_chain_lock;
-    std::condition_variable present_cv;
-    std::array<Frame, SWAP_CHAIN_SIZE> swap_chain{};
-    std::queue<Frame*> free_queue;
-    std::deque<Frame*> present_queue;
-    Frame* previous_frame{};
-
-    FrameMailbox() {
-        for (auto& frame : swap_chain) {
-            free_queue.push(&frame);
-        }
-    }
-
-    ~FrameMailbox() {
-        // lock the mutex and clear out the present and free_queues and notify any people who are
-        // blocked to prevent deadlock on shutdown
-        std::scoped_lock lock{swap_chain_lock};
-        std::queue<Frame*>().swap(free_queue);
-        present_queue.clear();
-        present_cv.notify_all();
-    }
-
-    void ReloadPresentFrame(Frame* frame, u32 height, u32 width) {
-        frame->present.Release();
-        frame->present.Create();
-        GLint previous_draw_fbo{};
-        glGetIntegerv(GL_DRAW_FRAMEBUFFER_BINDING, &previous_draw_fbo);
-        glBindFramebuffer(GL_FRAMEBUFFER, frame->present.handle);
-        glFramebufferRenderbuffer(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_RENDERBUFFER,
-                                  frame->color.handle);
-        if (glCheckFramebufferStatus(GL_FRAMEBUFFER) != GL_FRAMEBUFFER_COMPLETE) {
-            LOG_CRITICAL(Render_OpenGL, "Failed to recreate present FBO!");
-        }
-        glBindFramebuffer(GL_DRAW_FRAMEBUFFER, previous_draw_fbo);
-        frame->color_reloaded = false;
-    }
-
-    void ReloadRenderFrame(Frame* frame, u32 width, u32 height) {
-        // Recreate the color texture attachment
-        frame->color.Release();
-        frame->color.Create();
-        const GLenum internal_format = frame->is_srgb ? GL_SRGB8 : GL_RGB8;
-        glNamedRenderbufferStorage(frame->color.handle, internal_format, width, height);
-
-        // Recreate the FBO for the render target
-        frame->render.Release();
-        frame->render.Create();
-        glBindFramebuffer(GL_FRAMEBUFFER, frame->render.handle);
-        glFramebufferRenderbuffer(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_RENDERBUFFER,
-                                  frame->color.handle);
-        if (glCheckFramebufferStatus(GL_FRAMEBUFFER) != GL_FRAMEBUFFER_COMPLETE) {
-            LOG_CRITICAL(Render_OpenGL, "Failed to recreate render FBO!");
-        }
-
-        frame->width = width;
-        frame->height = height;
-        frame->color_reloaded = true;
-    }
-
-    Frame* GetRenderFrame() {
-        std::unique_lock lock{swap_chain_lock};
-
-        // If theres no free frames, we will reuse the oldest render frame
-        if (free_queue.empty()) {
-            auto frame = present_queue.back();
-            present_queue.pop_back();
-            return frame;
-        }
-
-        Frame* frame = free_queue.front();
-        free_queue.pop();
-        return frame;
-    }
-
-    void ReleaseRenderFrame(Frame* frame) {
-        std::unique_lock lock{swap_chain_lock};
-        present_queue.push_front(frame);
-        present_cv.notify_one();
-    }
-
-    Frame* TryGetPresentFrame(int timeout_ms) {
-        std::unique_lock lock{swap_chain_lock};
-        // wait for new entries in the present_queue
-        present_cv.wait_for(lock, std::chrono::milliseconds(timeout_ms),
-                            [&] { return !present_queue.empty(); });
-        if (present_queue.empty()) {
-            // timed out waiting for a frame to draw so return the previous frame
-            return previous_frame;
-        }
-
-        // free the previous frame and add it back to the free queue
-        if (previous_frame) {
-            free_queue.push(previous_frame);
-        }
-
-        // the newest entries are pushed to the front of the queue
-        Frame* frame = present_queue.front();
-        present_queue.pop_front();
-        // remove all old entries from the present queue and move them back to the free_queue
-        for (auto f : present_queue) {
-            free_queue.push(f);
-        }
-        present_queue.clear();
-        previous_frame = frame;
-        return frame;
-    }
-};
-
-namespace {
-
 constexpr char VERTEX_SHADER[] = R"(
 #version 430 core
 
@@ -211,6 +98,24 @@ struct ScreenRectVertex {
     std::array<GLfloat, 2> tex_coord;
 };
 
+/// Returns true if any debug tool is attached
+bool HasDebugTool() {
+    const bool nsight = std::getenv("NVTX_INJECTION64_PATH") || std::getenv("NSIGHT_LAUNCHED");
+    if (nsight) {
+        return true;
+    }
+
+    GLint num_extensions;
+    glGetIntegerv(GL_NUM_EXTENSIONS, &num_extensions);
+    for (GLuint index = 0; index < static_cast<GLuint>(num_extensions); ++index) {
+        const auto name = reinterpret_cast<const char*>(glGetStringi(GL_EXTENSIONS, index));
+        if (!std::strcmp(name, "GL_EXT_debug_tool")) {
+            return true;
+        }
+    }
+    return false;
+}
+
 /**
  * Defines a 1:1 pixel ortographic projection matrix with (0,0) on the top-left
  * corner and (width, height) on the lower-bottom.
@@ -294,6 +199,153 @@ void APIENTRY DebugHandler(GLenum source, GLenum type, GLuint id, GLenum severit
 
 } // Anonymous namespace
 
+/**
+ * For smooth Vsync rendering, we want to always present the latest frame that the core generates,
+ * but also make sure that rendering happens at the pace that the frontend dictates. This is a
+ * helper class that the renderer uses to sync frames between the render thread and the presentation
+ * thread
+ */
+class FrameMailbox {
+public:
+    std::mutex swap_chain_lock;
+    std::condition_variable present_cv;
+    std::array<Frame, SWAP_CHAIN_SIZE> swap_chain{};
+    std::queue<Frame*> free_queue;
+    std::deque<Frame*> present_queue;
+    Frame* previous_frame{};
+
+    FrameMailbox() : has_debug_tool{HasDebugTool()} {
+        for (auto& frame : swap_chain) {
+            free_queue.push(&frame);
+        }
+    }
+
+    ~FrameMailbox() {
+        // lock the mutex and clear out the present and free_queues and notify any people who are
+        // blocked to prevent deadlock on shutdown
+        std::scoped_lock lock{swap_chain_lock};
+        std::queue<Frame*>().swap(free_queue);
+        present_queue.clear();
+        present_cv.notify_all();
+    }
+
+    void ReloadPresentFrame(Frame* frame, u32 height, u32 width) {
+        frame->present.Release();
+        frame->present.Create();
+        GLint previous_draw_fbo{};
+        glGetIntegerv(GL_DRAW_FRAMEBUFFER_BINDING, &previous_draw_fbo);
+        glBindFramebuffer(GL_FRAMEBUFFER, frame->present.handle);
+        glFramebufferRenderbuffer(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_RENDERBUFFER,
+                                  frame->color.handle);
+        if (glCheckFramebufferStatus(GL_FRAMEBUFFER) != GL_FRAMEBUFFER_COMPLETE) {
+            LOG_CRITICAL(Render_OpenGL, "Failed to recreate present FBO!");
+        }
+        glBindFramebuffer(GL_DRAW_FRAMEBUFFER, previous_draw_fbo);
+        frame->color_reloaded = false;
+    }
+
+    void ReloadRenderFrame(Frame* frame, u32 width, u32 height) {
+        // Recreate the color texture attachment
+        frame->color.Release();
+        frame->color.Create();
+        const GLenum internal_format = frame->is_srgb ? GL_SRGB8 : GL_RGB8;
+        glNamedRenderbufferStorage(frame->color.handle, internal_format, width, height);
+
+        // Recreate the FBO for the render target
+        frame->render.Release();
+        frame->render.Create();
+        glBindFramebuffer(GL_FRAMEBUFFER, frame->render.handle);
+        glFramebufferRenderbuffer(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_RENDERBUFFER,
+                                  frame->color.handle);
+        if (glCheckFramebufferStatus(GL_FRAMEBUFFER) != GL_FRAMEBUFFER_COMPLETE) {
+            LOG_CRITICAL(Render_OpenGL, "Failed to recreate render FBO!");
+        }
+
+        frame->width = width;
+        frame->height = height;
+        frame->color_reloaded = true;
+    }
+
+    Frame* GetRenderFrame() {
+        std::unique_lock lock{swap_chain_lock};
+
+        // If theres no free frames, we will reuse the oldest render frame
+        if (free_queue.empty()) {
+            auto frame = present_queue.back();
+            present_queue.pop_back();
+            return frame;
+        }
+
+        Frame* frame = free_queue.front();
+        free_queue.pop();
+        return frame;
+    }
+
+    void ReleaseRenderFrame(Frame* frame) {
+        std::unique_lock lock{swap_chain_lock};
+        present_queue.push_front(frame);
+        present_cv.notify_one();
+
+        DebugNotifyNextFrame();
+    }
+
+    Frame* TryGetPresentFrame(int timeout_ms) {
+        DebugWaitForNextFrame();
+
+        std::unique_lock lock{swap_chain_lock};
+        // wait for new entries in the present_queue
+        present_cv.wait_for(lock, std::chrono::milliseconds(timeout_ms),
+                            [&] { return !present_queue.empty(); });
+        if (present_queue.empty()) {
+            // timed out waiting for a frame to draw so return the previous frame
+            return previous_frame;
+        }
+
+        // free the previous frame and add it back to the free queue
+        if (previous_frame) {
+            free_queue.push(previous_frame);
+        }
+
+        // the newest entries are pushed to the front of the queue
+        Frame* frame = present_queue.front();
+        present_queue.pop_front();
+        // remove all old entries from the present queue and move them back to the free_queue
+        for (auto f : present_queue) {
+            free_queue.push(f);
+        }
+        present_queue.clear();
+        previous_frame = frame;
+        return frame;
+    }
+
+private:
+    std::mutex debug_synch_mutex;
+    std::condition_variable debug_synch_condition;
+    std::atomic_int frame_for_debug{};
+    const bool has_debug_tool; // When true, using a GPU debugger, so keep frames in lock-step
+
+    /// Signal that a new frame is available (called from GPU thread)
+    void DebugNotifyNextFrame() {
+        if (!has_debug_tool) {
+            return;
+        }
+        frame_for_debug++;
+        std::lock_guard lock{debug_synch_mutex};
+        debug_synch_condition.notify_one();
+    }
+
+    /// Wait for a new frame to be available (called from presentation thread)
+    void DebugWaitForNextFrame() {
+        if (!has_debug_tool) {
+            return;
+        }
+        const int last_frame = frame_for_debug;
+        std::unique_lock lock{debug_synch_mutex};
+        debug_synch_condition.wait(lock,
+                                   [this, last_frame] { return frame_for_debug > last_frame; });
+    }
+};
+
 RendererOpenGL::RendererOpenGL(Core::Frontend::EmuWindow& emu_window, Core::System& system)
     : VideoCore::RendererBase{emu_window}, emu_window{emu_window}, system{system},
       frame_mailbox{std::make_unique<FrameMailbox>()} {}

src/video_core/renderer_vulkan/maxwell_to_vk.cpp

@@ -125,6 +125,7 @@ struct FormatTuple {
     {vk::Format::eR8Uint, Attachable | Storage},                 // R8UI
     {vk::Format::eR16G16B16A16Sfloat, Attachable | Storage},     // RGBA16F
     {vk::Format::eR16G16B16A16Unorm, Attachable | Storage},      // RGBA16U
+    {vk::Format::eR16G16B16A16Snorm, Attachable | Storage},      // RGBA16S
     {vk::Format::eR16G16B16A16Uint, Attachable | Storage},       // RGBA16UI
     {vk::Format::eB10G11R11UfloatPack32, Attachable | Storage},  // R11FG11FB10F
     {vk::Format::eR32G32B32A32Uint, Attachable | Storage},       // RGBA32UI
@@ -256,6 +257,8 @@ vk::ShaderStageFlagBits ShaderStage(Tegra::Engines::ShaderType stage) {
         return vk::ShaderStageFlagBits::eGeometry;
     case Tegra::Engines::ShaderType::Fragment:
         return vk::ShaderStageFlagBits::eFragment;
+    case Tegra::Engines::ShaderType::Compute:
+        return vk::ShaderStageFlagBits::eCompute;
     }
     UNIMPLEMENTED_MSG("Unimplemented shader stage={}", static_cast<u32>(stage));
     return {};
@@ -331,6 +334,8 @@ vk::Format VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttr
             return vk::Format::eR16G16B16Unorm;
         case Maxwell::VertexAttribute::Size::Size_16_16_16_16:
             return vk::Format::eR16G16B16A16Unorm;
+        case Maxwell::VertexAttribute::Size::Size_10_10_10_2:
+            return vk::Format::eA2B10G10R10UnormPack32;
         default:
             break;
         }
@@ -364,6 +369,10 @@ vk::Format VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttr
             return vk::Format::eR8G8B8A8Uint;
         case Maxwell::VertexAttribute::Size::Size_32:
             return vk::Format::eR32Uint;
+        case Maxwell::VertexAttribute::Size::Size_32_32:
+            return vk::Format::eR32G32Uint;
+        case Maxwell::VertexAttribute::Size::Size_32_32_32:
+            return vk::Format::eR32G32B32Uint;
         case Maxwell::VertexAttribute::Size::Size_32_32_32_32:
             return vk::Format::eR32G32B32A32Uint;
         default:

src/video_core/renderer_vulkan/vk_device.cpp

@@ -535,6 +535,7 @@ std::unordered_map<vk::Format, vk::FormatProperties> VKDevice::GetFormatProperti
                                         vk::Format::eR32G32Sfloat,
                                         vk::Format::eR32G32Uint,
                                         vk::Format::eR16G16B16A16Uint,
+                                        vk::Format::eR16G16B16A16Snorm,
                                         vk::Format::eR16G16B16A16Unorm,
                                         vk::Format::eR16G16Unorm,
                                         vk::Format::eR16G16Snorm,

src/video_core/renderer_vulkan/vk_pipeline_cache.cpp

@@ -179,10 +179,11 @@ Tegra::Engines::ConstBufferEngineInterface& CachedShader::GetEngine(
 VKPipelineCache::VKPipelineCache(Core::System& system, RasterizerVulkan& rasterizer,
                                  const VKDevice& device, VKScheduler& scheduler,
                                  VKDescriptorPool& descriptor_pool,
-                                 VKUpdateDescriptorQueue& update_descriptor_queue)
+                                 VKUpdateDescriptorQueue& update_descriptor_queue,
+                                 VKRenderPassCache& renderpass_cache)
     : RasterizerCache{rasterizer}, system{system}, device{device}, scheduler{scheduler},
       descriptor_pool{descriptor_pool}, update_descriptor_queue{update_descriptor_queue},
-      renderpass_cache(device) {}
+      renderpass_cache{renderpass_cache} {}
 
 VKPipelineCache::~VKPipelineCache() = default;
 
@@ -191,7 +192,6 @@ std::array<Shader, Maxwell::MaxShaderProgram> VKPipelineCache::GetShaders() {
 
     std::array<Shader, Maxwell::MaxShaderProgram> shaders;
     for (std::size_t index = 0; index < Maxwell::MaxShaderProgram; ++index) {
-        const auto& shader_config = gpu.regs.shader_config[index];
         const auto program{static_cast<Maxwell::ShaderProgram>(index)};
 
         // Skip stages that are not enabled

src/video_core/renderer_vulkan/vk_pipeline_cache.h

@@ -161,7 +161,8 @@ public:
     explicit VKPipelineCache(Core::System& system, RasterizerVulkan& rasterizer,
                              const VKDevice& device, VKScheduler& scheduler,
                              VKDescriptorPool& descriptor_pool,
-                             VKUpdateDescriptorQueue& update_descriptor_queue);
+                             VKUpdateDescriptorQueue& update_descriptor_queue,
+                             VKRenderPassCache& renderpass_cache);
     ~VKPipelineCache();
 
     std::array<Shader, Maxwell::MaxShaderProgram> GetShaders();
@@ -184,8 +185,7 @@ private:
     VKScheduler& scheduler;
     VKDescriptorPool& descriptor_pool;
     VKUpdateDescriptorQueue& update_descriptor_queue;
-
-    VKRenderPassCache renderpass_cache;
+    VKRenderPassCache& renderpass_cache;
 
     std::array<Shader, Maxwell::MaxShaderProgram> last_shaders;
 

src/video_core/renderer_vulkan/vk_rasterizer.cpp

@@ -287,12 +287,13 @@ RasterizerVulkan::RasterizerVulkan(Core::System& system, Core::Frontend::EmuWind
       screen_info{screen_info}, device{device}, resource_manager{resource_manager},
       memory_manager{memory_manager}, state_tracker{state_tracker}, scheduler{scheduler},
       staging_pool(device, memory_manager, scheduler), descriptor_pool(device),
-      update_descriptor_queue(device, scheduler),
+      update_descriptor_queue(device, scheduler), renderpass_cache(device),
       quad_array_pass(device, scheduler, descriptor_pool, staging_pool, update_descriptor_queue),
       uint8_pass(device, scheduler, descriptor_pool, staging_pool, update_descriptor_queue),
       texture_cache(system, *this, device, resource_manager, memory_manager, scheduler,
                     staging_pool),
-      pipeline_cache(system, *this, device, scheduler, descriptor_pool, update_descriptor_queue),
+      pipeline_cache(system, *this, device, scheduler, descriptor_pool, update_descriptor_queue,
+                     renderpass_cache),
       buffer_cache(*this, system, device, memory_manager, scheduler, staging_pool),
       sampler_cache(device), query_cache(system, *this, device, scheduler) {
     scheduler.SetQueryCache(query_cache);
@@ -365,13 +366,16 @@ void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) {
 void RasterizerVulkan::Clear() {
     MICROPROFILE_SCOPE(Vulkan_Clearing);
 
-    query_cache.UpdateCounters();
-
     const auto& gpu = system.GPU().Maxwell3D();
     if (!system.GPU().Maxwell3D().ShouldExecute()) {
         return;
     }
 
+    sampled_views.clear();
+    image_views.clear();
+
+    query_cache.UpdateCounters();
+
     const auto& regs = gpu.regs;
     const bool use_color = regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B ||
                            regs.clear_buffers.A;
@@ -380,52 +384,54 @@ void RasterizerVulkan::Clear() {
     if (!use_color && !use_depth && !use_stencil) {
         return;
     }
-    // Clearing images requires to be out of a renderpass
-    scheduler.RequestOutsideRenderPassOperationContext();
 
-    // TODO(Rodrigo): Implement clears rendering a quad or using beginning a renderpass.
+    [[maybe_unused]] const auto texceptions = UpdateAttachments();
+    DEBUG_ASSERT(texceptions.none());
+    SetupImageTransitions(0, color_attachments, zeta_attachment);
+
+    const vk::RenderPass renderpass = renderpass_cache.GetRenderPass(GetRenderPassParams(0));
+    const auto [framebuffer, render_area] = ConfigureFramebuffers(renderpass);
+    scheduler.RequestRenderpass({renderpass, framebuffer, {{0, 0}, render_area}, 0, nullptr});
+
+    const auto& scissor = regs.scissor_test[0];
+    const vk::Offset2D scissor_offset(scissor.min_x, scissor.min_y);
+    vk::Extent2D scissor_extent{scissor.max_x - scissor.min_x, scissor.max_y - scissor.min_y};
+    scissor_extent.width = std::min(scissor_extent.width, render_area.width);
+    scissor_extent.height = std::min(scissor_extent.height, render_area.height);
+
+    const u32 layer = regs.clear_buffers.layer;
+    const vk::ClearRect clear_rect({scissor_offset, scissor_extent}, layer, 1);
 
     if (use_color) {
-        View color_view;
-        {
-            MICROPROFILE_SCOPE(Vulkan_RenderTargets);
-            color_view = texture_cache.GetColorBufferSurface(regs.clear_buffers.RT.Value(), false);
-        }
-
-        color_view->Transition(vk::ImageLayout::eTransferDstOptimal,
-                               vk::PipelineStageFlagBits::eTransfer,
-                               vk::AccessFlagBits::eTransferWrite);
-
         const std::array clear_color = {regs.clear_color[0], regs.clear_color[1],
                                         regs.clear_color[2], regs.clear_color[3]};
-        const vk::ClearColorValue clear(clear_color);
-        scheduler.Record([image = color_view->GetImage(),
-                          subresource = color_view->GetImageSubresourceRange(),
-                          clear](auto cmdbuf, auto& dld) {
-            cmdbuf.clearColorImage(image, vk::ImageLayout::eTransferDstOptimal, clear, subresource,
-                                   dld);
+        const vk::ClearValue clear_value{clear_color};
+        const u32 color_attachment = regs.clear_buffers.RT;
+        scheduler.Record([color_attachment, clear_value, clear_rect](auto cmdbuf, auto& dld) {
+            const vk::ClearAttachment attachment(vk::ImageAspectFlagBits::eColor, color_attachment,
+                                                 clear_value);
+            cmdbuf.clearAttachments(1, &attachment, 1, &clear_rect, dld);
         });
     }
-    if (use_depth || use_stencil) {
-        View zeta_surface;
-        {
-            MICROPROFILE_SCOPE(Vulkan_RenderTargets);
-            zeta_surface = texture_cache.GetDepthBufferSurface(false);
-        }
 
-        zeta_surface->Transition(vk::ImageLayout::eTransferDstOptimal,
-                                 vk::PipelineStageFlagBits::eTransfer,
-                                 vk::AccessFlagBits::eTransferWrite);
-
-        const vk::ClearDepthStencilValue clear(regs.clear_depth,
-                                               static_cast<u32>(regs.clear_stencil));
-        scheduler.Record([image = zeta_surface->GetImage(),
-                          subresource = zeta_surface->GetImageSubresourceRange(),
-                          clear](auto cmdbuf, auto& dld) {
-            cmdbuf.clearDepthStencilImage(image, vk::ImageLayout::eTransferDstOptimal, clear,
-                                          subresource, dld);
-        });
+    if (!use_depth && !use_stencil) {
+        return;
     }
+    vk::ImageAspectFlags aspect_flags;
+    if (use_depth) {
+        aspect_flags |= vk::ImageAspectFlagBits::eDepth;
+    }
+    if (use_stencil) {
+        aspect_flags |= vk::ImageAspectFlagBits::eStencil;
+    }
+
+    scheduler.Record([clear_depth = regs.clear_depth, clear_stencil = regs.clear_stencil,
+                      clear_rect, aspect_flags](auto cmdbuf, auto& dld) {
+        const vk::ClearDepthStencilValue clear_zeta(clear_depth, clear_stencil);
+        const vk::ClearValue clear_value{clear_zeta};
+        const vk::ClearAttachment attachment(aspect_flags, 0, clear_value);
+        cmdbuf.clearAttachments(1, &attachment, 1, &clear_rect, dld);
+    });
 }
 
 void RasterizerVulkan::DispatchCompute(GPUVAddr code_addr) {
@@ -542,8 +548,6 @@ bool RasterizerVulkan::AccelerateDisplay(const Tegra::FramebufferConfig& config,
 
     // Verify that the cached surface is the same size and format as the requested framebuffer
     const auto& params{surface->GetSurfaceParams()};
-    const auto& pixel_format{
-        VideoCore::Surface::PixelFormatFromGPUPixelFormat(config.pixel_format)};
     ASSERT_MSG(params.width == config.width, "Framebuffer width is different");
     ASSERT_MSG(params.height == config.height, "Framebuffer height is different");
 
@@ -1151,7 +1155,7 @@ std::size_t RasterizerVulkan::CalculateVertexArraysSize() const {
         // This implementation assumes that all attributes are used in the shader.
         const GPUVAddr start{regs.vertex_array[index].StartAddress()};
         const GPUVAddr end{regs.vertex_array_limit[index].LimitAddress()};
-        DEBUG_ASSERT(end > start);
+        DEBUG_ASSERT(end >= start);
 
         size += (end - start + 1) * regs.vertex_array[index].enable;
     }

src/video_core/renderer_vulkan/vk_rasterizer.h

@@ -253,6 +253,7 @@ private:
     VKStagingBufferPool staging_pool;
     VKDescriptorPool descriptor_pool;
     VKUpdateDescriptorQueue update_descriptor_queue;
+    VKRenderPassCache renderpass_cache;
     QuadArrayPass quad_array_pass;
     Uint8Pass uint8_pass;
 

src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp

@@ -100,7 +100,6 @@ void VKStagingBufferPool::ReleaseCache(bool host_visible) {
 }
 
 u64 VKStagingBufferPool::ReleaseLevel(StagingBuffersCache& cache, std::size_t log2) {
-    static constexpr u64 epochs_to_destroy = 180;
     static constexpr std::size_t deletions_per_tick = 16;
 
     auto& staging = cache[log2];
@@ -108,6 +107,7 @@ u64 VKStagingBufferPool::ReleaseLevel(StagingBuffersCache& cache, std::size_t lo
     const std::size_t old_size = entries.size();
 
     const auto is_deleteable = [this](const auto& entry) {
+        static constexpr u64 epochs_to_destroy = 180;
         return entry.last_epoch + epochs_to_destroy < epoch && !entry.watch.IsUsed();
     };
     const std::size_t begin_offset = staging.delete_index;

src/video_core/renderer_vulkan/vk_state_tracker.cpp

@@ -90,8 +90,6 @@ void StateTracker::Initialize() {
     SetupDirtyBlendConstants(tables);
     SetupDirtyDepthBounds(tables);
     SetupDirtyStencilProperties(tables);
-
-    SetupCommonOnWriteStores(dirty.on_write_stores);
 }
 
 void StateTracker::InvalidateCommandBufferState() {

src/video_core/renderer_vulkan/vk_texture_cache.cpp

@@ -52,6 +52,9 @@ vk::ImageType SurfaceTargetToImage(SurfaceTarget target) {
         return vk::ImageType::e2D;
     case SurfaceTarget::Texture3D:
         return vk::ImageType::e3D;
+    case SurfaceTarget::TextureBuffer:
+        UNREACHABLE();
+        return {};
     }
     UNREACHABLE_MSG("Unknown texture target={}", static_cast<u32>(target));
     return {};
@@ -273,7 +276,6 @@ void CachedSurface::UploadImage(const std::vector<u8>& staging_buffer) {
 
     for (u32 level = 0; level < params.num_levels; ++level) {
         vk::BufferImageCopy copy = GetBufferImageCopy(level);
-        const auto& dld = device.GetDispatchLoader();
         if (image->GetAspectMask() ==
             (vk::ImageAspectFlagBits::eDepth | vk::ImageAspectFlagBits::eStencil)) {
             vk::BufferImageCopy depth = copy;
@@ -422,7 +424,6 @@ void VKTextureCache::ImageCopy(Surface& src_surface, Surface& dst_surface,
         dst_base_layer, num_layers, copy_params.dest_level, 1, vk::PipelineStageFlagBits::eTransfer,
         vk::AccessFlagBits::eTransferWrite, vk::ImageLayout::eTransferDstOptimal);
 
-    const auto& dld{device.GetDispatchLoader()};
     const vk::ImageSubresourceLayers src_subresource(
         src_surface->GetAspectMask(), copy_params.source_level, copy_params.source_z, num_layers);
     const vk::ImageSubresourceLayers dst_subresource(
@@ -458,7 +459,6 @@ void VKTextureCache::ImageBlit(View& src_view, View& dst_view,
                              dst_view->GetImageSubresourceLayers(), {dst_top_left, dst_bot_right});
     const bool is_linear = copy_config.filter == Tegra::Engines::Fermi2D::Filter::Linear;
 
-    const auto& dld{device.GetDispatchLoader()};
     scheduler.Record([src_image = src_view->GetImage(), dst_image = dst_view->GetImage(), blit,
                       is_linear](auto cmdbuf, auto& dld) {
         cmdbuf.blitImage(src_image, vk::ImageLayout::eTransferSrcOptimal, dst_image,

src/video_core/shader/decode/bfe.cpp

@@ -17,33 +17,60 @@ u32 ShaderIR::DecodeBfe(NodeBlock& bb, u32 pc) {
     const Instruction instr = {program_code[pc]};
     const auto opcode = OpCode::Decode(instr);
 
-    UNIMPLEMENTED_IF(instr.bfe.negate_b);
-
     Node op_a = GetRegister(instr.gpr8);
-    op_a = GetOperandAbsNegInteger(op_a, false, instr.bfe.negate_a, false);
+    Node op_b = [&] {
+        switch (opcode->get().GetId()) {
+        case OpCode::Id::BFE_R:
+            return GetRegister(instr.gpr20);
+        case OpCode::Id::BFE_C:
+            return GetConstBuffer(instr.cbuf34.index, instr.cbuf34.GetOffset());
+        case OpCode::Id::BFE_IMM:
+            return Immediate(instr.alu.GetSignedImm20_20());
+        default:
+            UNREACHABLE();
+            return Immediate(0);
+        }
+    }();
 
-    switch (opcode->get().GetId()) {
-    case OpCode::Id::BFE_IMM: {
-        UNIMPLEMENTED_IF_MSG(instr.generates_cc,
-                             "Condition codes generation in BFE is not implemented");
+    UNIMPLEMENTED_IF_MSG(instr.bfe.rd_cc, "Condition codes in BFE is not implemented");
 
-        const Node inner_shift_imm = Immediate(static_cast<u32>(instr.bfe.GetLeftShiftValue()));
-        const Node outer_shift_imm =
-            Immediate(static_cast<u32>(instr.bfe.GetLeftShiftValue() + instr.bfe.shift_position));
+    const bool is_signed = instr.bfe.is_signed;
 
-        const Node inner_shift =
-            Operation(OperationCode::ILogicalShiftLeft, NO_PRECISE, op_a, inner_shift_imm);
-        const Node outer_shift =
-            Operation(OperationCode::ILogicalShiftRight, NO_PRECISE, inner_shift, outer_shift_imm);
-
-        SetInternalFlagsFromInteger(bb, outer_shift, instr.generates_cc);
-        SetRegister(bb, instr.gpr0, outer_shift);
-        break;
-    }
-    default:
-        UNIMPLEMENTED_MSG("Unhandled BFE instruction: {}", opcode->get().GetName());
+    // using reverse parallel method in
+    // https://graphics.stanford.edu/~seander/bithacks.html#ReverseParallel
+    // note for later if possible to implement faster method.
+    if (instr.bfe.brev) {
+        const auto swap = [&](u32 s, u32 mask) {
+            Node v1 =
+                SignedOperation(OperationCode::ILogicalShiftRight, is_signed, op_a, Immediate(s));
+            if (mask != 0) {
+                v1 = SignedOperation(OperationCode::IBitwiseAnd, is_signed, std::move(v1),
+                                     Immediate(mask));
+            }
+            Node v2 = op_a;
+            if (mask != 0) {
+                v2 = SignedOperation(OperationCode::IBitwiseAnd, is_signed, std::move(v2),
+                                     Immediate(mask));
+            }
+            v2 = SignedOperation(OperationCode::ILogicalShiftLeft, is_signed, std::move(v2),
+                                 Immediate(s));
+            return SignedOperation(OperationCode::IBitwiseOr, is_signed, std::move(v1),
+                                   std::move(v2));
+        };
+        op_a = swap(1, 0x55555555U);
+        op_a = swap(2, 0x33333333U);
+        op_a = swap(4, 0x0F0F0F0FU);
+        op_a = swap(8, 0x00FF00FFU);
+        op_a = swap(16, 0);
     }
 
+    const auto offset = SignedOperation(OperationCode::IBitfieldExtract, is_signed, op_b,
+                                        Immediate(0), Immediate(8));
+    const auto bits = SignedOperation(OperationCode::IBitfieldExtract, is_signed, op_b,
+                                      Immediate(8), Immediate(8));
+    auto result = SignedOperation(OperationCode::IBitfieldExtract, is_signed, op_a, offset, bits);
+    SetRegister(bb, instr.gpr0, std::move(result));
+
     return pc;
 }
 

src/video_core/shader/decode/xmad.cpp

@@ -12,6 +12,7 @@ namespace VideoCommon::Shader {
 
 using Tegra::Shader::Instruction;
 using Tegra::Shader::OpCode;
+using Tegra::Shader::PredCondition;
 
 u32 ShaderIR::DecodeXmad(NodeBlock& bb, u32 pc) {
     const Instruction instr = {program_code[pc]};
@@ -63,15 +64,18 @@ u32 ShaderIR::DecodeXmad(NodeBlock& bb, u32 pc) {
         }
     }();
 
-    op_a = BitfieldExtract(op_a, instr.xmad.high_a ? 16 : 0, 16);
+    op_a = SignedOperation(OperationCode::IBitfieldExtract, is_signed_a, std::move(op_a),
+                           instr.xmad.high_a ? Immediate(16) : Immediate(0), Immediate(16));
 
     const Node original_b = op_b;
-    op_b = BitfieldExtract(op_b, is_high_b ? 16 : 0, 16);
+    op_b = SignedOperation(OperationCode::IBitfieldExtract, is_signed_b, std::move(op_b),
+                           is_high_b ? Immediate(16) : Immediate(0), Immediate(16));
 
-    // TODO(Rodrigo): Use an appropiate sign for this operation
-    Node product = Operation(OperationCode::IMul, NO_PRECISE, op_a, op_b);
+    // we already check sign_a and sign_b is difference or not before so just use one in here.
+    Node product = SignedOperation(OperationCode::IMul, is_signed_a, op_a, op_b);
     if (is_psl) {
-        product = Operation(OperationCode::ILogicalShiftLeft, NO_PRECISE, product, Immediate(16));
+        product =
+            SignedOperation(OperationCode::ILogicalShiftLeft, is_signed_a, product, Immediate(16));
     }
     SetTemporary(bb, 0, product);
     product = GetTemporary(0);
@@ -88,12 +92,40 @@ u32 ShaderIR::DecodeXmad(NodeBlock& bb, u32 pc) {
             return BitfieldExtract(original_c, 16, 16);
         case Tegra::Shader::XmadMode::CBcc: {
             const Node shifted_b = SignedOperation(OperationCode::ILogicalShiftLeft, is_signed_b,
-                                                   NO_PRECISE, original_b, Immediate(16));
-            return SignedOperation(OperationCode::IAdd, is_signed_c, NO_PRECISE, original_c,
-                                   shifted_b);
+                                                   original_b, Immediate(16));
+            return SignedOperation(OperationCode::IAdd, is_signed_c, original_c, shifted_b);
+        }
+        case Tegra::Shader::XmadMode::CSfu: {
+            const Node comp_a = GetPredicateComparisonInteger(PredCondition::Equal, is_signed_a,
+                                                              op_a, Immediate(0));
+            const Node comp_b = GetPredicateComparisonInteger(PredCondition::Equal, is_signed_b,
+                                                              op_b, Immediate(0));
+            const Node comp = Operation(OperationCode::LogicalOr, comp_a, comp_b);
+
+            const Node comp_minus_a = GetPredicateComparisonInteger(
+                PredCondition::NotEqual, is_signed_a,
+                SignedOperation(OperationCode::IBitwiseAnd, is_signed_a, op_a,
+                                Immediate(0x80000000)),
+                Immediate(0));
+            const Node comp_minus_b = GetPredicateComparisonInteger(
+                PredCondition::NotEqual, is_signed_b,
+                SignedOperation(OperationCode::IBitwiseAnd, is_signed_b, op_b,
+                                Immediate(0x80000000)),
+                Immediate(0));
+
+            Node new_c = Operation(
+                OperationCode::Select, comp_minus_a,
+                SignedOperation(OperationCode::IAdd, is_signed_c, original_c, Immediate(-65536)),
+                original_c);
+            new_c = Operation(
+                OperationCode::Select, comp_minus_b,
+                SignedOperation(OperationCode::IAdd, is_signed_c, new_c, Immediate(-65536)),
+                std::move(new_c));
+
+            return Operation(OperationCode::Select, comp, original_c, std::move(new_c));
         }
         default:
-            UNIMPLEMENTED_MSG("Unhandled XMAD mode: {}", static_cast<u32>(instr.xmad.mode.Value()));
+            UNREACHABLE();
             return Immediate(0);
         }
     }();
@@ -102,18 +134,19 @@ u32 ShaderIR::DecodeXmad(NodeBlock& bb, u32 pc) {
     op_c = GetTemporary(1);
 
     // TODO(Rodrigo): Use an appropiate sign for this operation
-    Node sum = Operation(OperationCode::IAdd, product, op_c);
+    Node sum = SignedOperation(OperationCode::IAdd, is_signed_a, product, std::move(op_c));
     SetTemporary(bb, 2, sum);
     sum = GetTemporary(2);
     if (is_merge) {
-        const Node a = BitfieldExtract(sum, 0, 16);
-        const Node b =
-            Operation(OperationCode::ILogicalShiftLeft, NO_PRECISE, original_b, Immediate(16));
-        sum = Operation(OperationCode::IBitwiseOr, NO_PRECISE, a, b);
+        const Node a = SignedOperation(OperationCode::IBitfieldExtract, is_signed_a, std::move(sum),
+                                       Immediate(0), Immediate(16));
+        const Node b = SignedOperation(OperationCode::ILogicalShiftLeft, is_signed_b, original_b,
+                                       Immediate(16));
+        sum = SignedOperation(OperationCode::IBitwiseOr, is_signed_a, a, b);
     }
 
     SetInternalFlagsFromInteger(bb, sum, instr.generates_cc);
-    SetRegister(bb, instr.gpr0, sum);
+    SetRegister(bb, instr.gpr0, std::move(sum));
 
     return pc;
 }

src/video_core/shader/node_helper.cpp

@@ -68,6 +68,8 @@ OperationCode SignedToUnsignedCode(OperationCode operation_code, bool is_signed)
         return OperationCode::UBitwiseXor;
     case OperationCode::IBitwiseNot:
         return OperationCode::UBitwiseNot;
+    case OperationCode::IBitfieldExtract:
+        return OperationCode::UBitfieldExtract;
     case OperationCode::IBitfieldInsert:
         return OperationCode::UBitfieldInsert;
     case OperationCode::IBitCount:

src/video_core/surface.cpp

@@ -111,6 +111,8 @@ PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format)
         return PixelFormat::RGBA16F;
     case Tegra::RenderTargetFormat::RGBA16_UNORM:
         return PixelFormat::RGBA16U;
+    case Tegra::RenderTargetFormat::RGBA16_SNORM:
+        return PixelFormat::RGBA16S;
     case Tegra::RenderTargetFormat::RGBA16_UINT:
         return PixelFormat::RGBA16UI;
     case Tegra::RenderTargetFormat::RGBA32_FLOAT:

src/video_core/surface.h

@@ -25,82 +25,83 @@ enum class PixelFormat {
     R8UI = 7,
     RGBA16F = 8,
     RGBA16U = 9,
-    RGBA16UI = 10,
-    R11FG11FB10F = 11,
-    RGBA32UI = 12,
-    DXT1 = 13,
-    DXT23 = 14,
-    DXT45 = 15,
-    DXN1 = 16, // This is also known as BC4
-    DXN2UNORM = 17,
-    DXN2SNORM = 18,
-    BC7U = 19,
-    BC6H_UF16 = 20,
-    BC6H_SF16 = 21,
-    ASTC_2D_4X4 = 22,
-    BGRA8 = 23,
-    RGBA32F = 24,
-    RG32F = 25,
-    R32F = 26,
-    R16F = 27,
-    R16U = 28,
-    R16S = 29,
-    R16UI = 30,
-    R16I = 31,
-    RG16 = 32,
-    RG16F = 33,
-    RG16UI = 34,
-    RG16I = 35,
-    RG16S = 36,
-    RGB32F = 37,
-    RGBA8_SRGB = 38,
-    RG8U = 39,
-    RG8S = 40,
-    RG32UI = 41,
-    RGBX16F = 42,
-    R32UI = 43,
-    R32I = 44,
-    ASTC_2D_8X8 = 45,
-    ASTC_2D_8X5 = 46,
-    ASTC_2D_5X4 = 47,
-    BGRA8_SRGB = 48,
-    DXT1_SRGB = 49,
-    DXT23_SRGB = 50,
-    DXT45_SRGB = 51,
-    BC7U_SRGB = 52,
-    R4G4B4A4U = 53,
-    ASTC_2D_4X4_SRGB = 54,
-    ASTC_2D_8X8_SRGB = 55,
-    ASTC_2D_8X5_SRGB = 56,
-    ASTC_2D_5X4_SRGB = 57,
-    ASTC_2D_5X5 = 58,
-    ASTC_2D_5X5_SRGB = 59,
-    ASTC_2D_10X8 = 60,
-    ASTC_2D_10X8_SRGB = 61,
-    ASTC_2D_6X6 = 62,
-    ASTC_2D_6X6_SRGB = 63,
-    ASTC_2D_10X10 = 64,
-    ASTC_2D_10X10_SRGB = 65,
-    ASTC_2D_12X12 = 66,
-    ASTC_2D_12X12_SRGB = 67,
-    ASTC_2D_8X6 = 68,
-    ASTC_2D_8X6_SRGB = 69,
-    ASTC_2D_6X5 = 70,
-    ASTC_2D_6X5_SRGB = 71,
-    E5B9G9R9F = 72,
+    RGBA16S = 10,
+    RGBA16UI = 11,
+    R11FG11FB10F = 12,
+    RGBA32UI = 13,
+    DXT1 = 14,
+    DXT23 = 15,
+    DXT45 = 16,
+    DXN1 = 17, // This is also known as BC4
+    DXN2UNORM = 18,
+    DXN2SNORM = 19,
+    BC7U = 20,
+    BC6H_UF16 = 21,
+    BC6H_SF16 = 22,
+    ASTC_2D_4X4 = 23,
+    BGRA8 = 24,
+    RGBA32F = 25,
+    RG32F = 26,
+    R32F = 27,
+    R16F = 28,
+    R16U = 29,
+    R16S = 30,
+    R16UI = 31,
+    R16I = 32,
+    RG16 = 33,
+    RG16F = 34,
+    RG16UI = 35,
+    RG16I = 36,
+    RG16S = 37,
+    RGB32F = 38,
+    RGBA8_SRGB = 39,
+    RG8U = 40,
+    RG8S = 41,
+    RG32UI = 42,
+    RGBX16F = 43,
+    R32UI = 44,
+    R32I = 45,
+    ASTC_2D_8X8 = 46,
+    ASTC_2D_8X5 = 47,
+    ASTC_2D_5X4 = 48,
+    BGRA8_SRGB = 49,
+    DXT1_SRGB = 50,
+    DXT23_SRGB = 51,
+    DXT45_SRGB = 52,
+    BC7U_SRGB = 53,
+    R4G4B4A4U = 54,
+    ASTC_2D_4X4_SRGB = 55,
+    ASTC_2D_8X8_SRGB = 56,
+    ASTC_2D_8X5_SRGB = 57,
+    ASTC_2D_5X4_SRGB = 58,
+    ASTC_2D_5X5 = 59,
+    ASTC_2D_5X5_SRGB = 60,
+    ASTC_2D_10X8 = 61,
+    ASTC_2D_10X8_SRGB = 62,
+    ASTC_2D_6X6 = 63,
+    ASTC_2D_6X6_SRGB = 64,
+    ASTC_2D_10X10 = 65,
+    ASTC_2D_10X10_SRGB = 66,
+    ASTC_2D_12X12 = 67,
+    ASTC_2D_12X12_SRGB = 68,
+    ASTC_2D_8X6 = 69,
+    ASTC_2D_8X6_SRGB = 70,
+    ASTC_2D_6X5 = 71,
+    ASTC_2D_6X5_SRGB = 72,
+    E5B9G9R9F = 73,
 
     MaxColorFormat,
 
     // Depth formats
-    Z32F = 73,
-    Z16 = 74,
+    Z32F = 74,
+    Z16 = 75,
 
     MaxDepthFormat,
 
     // DepthStencil formats
-    Z24S8 = 75,
-    S8Z24 = 76,
-    Z32FS8 = 77,
+    Z24S8 = 76,
+    S8Z24 = 77,
+    Z32FS8 = 78,
 
     MaxDepthStencilFormat,
 
@@ -138,6 +139,7 @@ constexpr std::array<u32, MaxPixelFormat> compression_factor_shift_table = {{
     0, // R8UI
     0, // RGBA16F
     0, // RGBA16U
+    0, // RGBA16S
     0, // RGBA16UI
     0, // R11FG11FB10F
     0, // RGBA32UI
@@ -235,6 +237,7 @@ constexpr std::array<u32, MaxPixelFormat> block_width_table = {{
     1,  // R8UI
     1,  // RGBA16F
     1,  // RGBA16U
+    1,  // RGBA16S
     1,  // RGBA16UI
     1,  // R11FG11FB10F
     1,  // RGBA32UI
@@ -324,6 +327,7 @@ constexpr std::array<u32, MaxPixelFormat> block_height_table = {{
     1,  // R8UI
     1,  // RGBA16F
     1,  // RGBA16U
+    1,  // RGBA16S
     1,  // RGBA16UI
     1,  // R11FG11FB10F
     1,  // RGBA32UI
@@ -413,6 +417,7 @@ constexpr std::array<u32, MaxPixelFormat> bpp_table = {{
     8,   // R8UI
     64,  // RGBA16F
     64,  // RGBA16U
+    64,  // RGBA16S
     64,  // RGBA16UI
     32,  // R11FG11FB10F
     128, // RGBA32UI
@@ -517,6 +522,7 @@ constexpr std::array<SurfaceCompression, MaxPixelFormat> compression_type_table
     SurfaceCompression::None,       // R8UI
     SurfaceCompression::None,       // RGBA16F
     SurfaceCompression::None,       // RGBA16U
+    SurfaceCompression::None,       // RGBA16S
     SurfaceCompression::None,       // RGBA16UI
     SurfaceCompression::None,       // R11FG11FB10F
     SurfaceCompression::None,       // RGBA32UI

src/video_core/texture_cache/format_lookup_table.cpp

@@ -41,7 +41,7 @@ struct Table {
     ComponentType alpha_component;
     bool is_srgb;
 };
-constexpr std::array<Table, 75> DefinitionTable = {{
+constexpr std::array<Table, 76> DefinitionTable = {{
     {TextureFormat::A8R8G8B8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ABGR8U},
     {TextureFormat::A8R8G8B8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::ABGR8S},
     {TextureFormat::A8R8G8B8, C, UINT, UINT, UINT, UINT, PixelFormat::ABGR8UI},
@@ -61,6 +61,7 @@ constexpr std::array<Table, 75> DefinitionTable = {{
     {TextureFormat::G8R8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::RG8U},
     {TextureFormat::G8R8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::RG8S},
 
+    {TextureFormat::R16_G16_B16_A16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::RGBA16S},
     {TextureFormat::R16_G16_B16_A16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::RGBA16U},
     {TextureFormat::R16_G16_B16_A16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::RGBA16F},
     {TextureFormat::R16_G16_B16_A16, C, UINT, UINT, UINT, UINT, PixelFormat::RGBA16UI},

src/yuzu/configuration/config.cpp

@@ -539,7 +539,7 @@ void Config::ReadDebuggingValues() {
 void Config::ReadServiceValues() {
     qt_config->beginGroup(QStringLiteral("Services"));
     Settings::values.bcat_backend =
-        ReadSetting(QStringLiteral("bcat_backend"), QStringLiteral("boxcat"))
+        ReadSetting(QStringLiteral("bcat_backend"), QStringLiteral("null"))
             .toString()
             .toStdString();
     Settings::values.bcat_boxcat_local =
@@ -682,6 +682,8 @@ void Config::ReadSystemValues() {
 
     Settings::values.language_index = ReadSetting(QStringLiteral("language_index"), 1).toInt();
 
+    Settings::values.region_index = ReadSetting(QStringLiteral("region_index"), 1).toInt();
+
     const auto rng_seed_enabled = ReadSetting(QStringLiteral("rng_seed_enabled"), false).toBool();
     if (rng_seed_enabled) {
         Settings::values.rng_seed = ReadSetting(QStringLiteral("rng_seed"), 0).toULongLong();
@@ -698,6 +700,8 @@ void Config::ReadSystemValues() {
         Settings::values.custom_rtc = std::nullopt;
     }
 
+    Settings::values.sound_index = ReadSetting(QStringLiteral("sound_index"), 1).toInt();
+
     qt_config->endGroup();
 }
 
@@ -1114,6 +1118,7 @@ void Config::SaveSystemValues() {
     WriteSetting(QStringLiteral("use_docked_mode"), Settings::values.use_docked_mode, false);
     WriteSetting(QStringLiteral("current_user"), Settings::values.current_user, 0);
     WriteSetting(QStringLiteral("language_index"), Settings::values.language_index, 1);
+    WriteSetting(QStringLiteral("region_index"), Settings::values.region_index, 1);
 
     WriteSetting(QStringLiteral("rng_seed_enabled"), Settings::values.rng_seed.has_value(), false);
     WriteSetting(QStringLiteral("rng_seed"), Settings::values.rng_seed.value_or(0), 0);
@@ -1125,6 +1130,8 @@ void Config::SaveSystemValues() {
                      Settings::values.custom_rtc.value_or(std::chrono::seconds{}).count()),
                  0);
 
+    WriteSetting(QStringLiteral("sound_index"), Settings::values.sound_index, 1);
+
     qt_config->endGroup();
 }
 

src/yuzu/configuration/configure_system.cpp

@@ -56,6 +56,8 @@ void ConfigureSystem::SetConfiguration() {
     enabled = !Core::System::GetInstance().IsPoweredOn();
 
     ui->combo_language->setCurrentIndex(Settings::values.language_index);
+    ui->combo_region->setCurrentIndex(Settings::values.region_index);
+    ui->combo_sound->setCurrentIndex(Settings::values.sound_index);
 
     ui->rng_seed_checkbox->setChecked(Settings::values.rng_seed.has_value());
     ui->rng_seed_edit->setEnabled(Settings::values.rng_seed.has_value());
@@ -81,6 +83,8 @@ void ConfigureSystem::ApplyConfiguration() {
     }
 
     Settings::values.language_index = ui->combo_language->currentIndex();
+    Settings::values.region_index = ui->combo_region->currentIndex();
+    Settings::values.sound_index = ui->combo_sound->currentIndex();
 
     if (ui->rng_seed_checkbox->isChecked()) {
         Settings::values.rng_seed = ui->rng_seed_edit->text().toULongLong(nullptr, 16);

src/yuzu/configuration/configure_system.h

@@ -36,5 +36,6 @@ private:
     bool enabled = false;
 
     int language_index = 0;
+    int region_index = 0;
     int sound_index = 0;
 };

src/yuzu/configuration/configure_system.ui

@@ -22,14 +22,14 @@
         <string>System Settings</string>
        </property>
        <layout class="QGridLayout" name="gridLayout">
-        <item row="1" column="0">
+        <item row="2" column="0">
          <widget class="QLabel" name="label_sound">
           <property name="text">
            <string>Sound output mode</string>
           </property>
          </widget>
         </item>
-        <item row="2" column="0">
+        <item row="3" column="0">
          <widget class="QLabel" name="label_console_id">
           <property name="text">
            <string>Console ID:</string>
@@ -128,14 +128,60 @@
           </item>
          </widget>
         </item>
-        <item row="4" column="0">
+        <item row="1" column="0">
+         <widget class="QLabel" name="label_region">
+          <property name="text">
+           <string>Region:</string>
+          </property>
+         </widget>
+        </item>
+        <item row="1" column="1">
+         <widget class="QComboBox" name="combo_region">
+          <item>
+           <property name="text">
+            <string>Japan</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>USA</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>Europe</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>Australia</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>China</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>Korea</string>
+           </property>
+          </item>
+          <item>
+           <property name="text">
+            <string>Taiwan</string>
+           </property>
+          </item>
+         </widget>
+        </item>
+        <item row="5" column="0">
          <widget class="QCheckBox" name="rng_seed_checkbox">
           <property name="text">
            <string>RNG Seed</string>
           </property>
          </widget>
         </item>
-        <item row="1" column="1">
+        <item row="2" column="1">
          <widget class="QComboBox" name="combo_sound">
           <item>
            <property name="text">
@@ -161,7 +207,7 @@
           </property>
          </widget>
         </item>
-        <item row="2" column="1">
+        <item row="3" column="1">
          <widget class="QPushButton" name="button_regenerate_console_id">
           <property name="sizePolicy">
            <sizepolicy hsizetype="Fixed" vsizetype="Fixed">
@@ -177,14 +223,14 @@
           </property>
          </widget>
         </item>
-        <item row="3" column="0">
+        <item row="4" column="0">
          <widget class="QCheckBox" name="custom_rtc_checkbox">
           <property name="text">
            <string>Custom RTC</string>
           </property>
          </widget>
         </item>
-        <item row="3" column="1">
+        <item row="4" column="1">
          <widget class="QDateTimeEdit" name="custom_rtc_edit">
           <property name="minimumDate">
            <date>
@@ -198,7 +244,7 @@
           </property>
          </widget>
         </item>
-        <item row="4" column="1">
+        <item row="5" column="1">
          <widget class="QLineEdit" name="rng_seed_edit">
           <property name="sizePolicy">
            <sizepolicy hsizetype="Minimum" vsizetype="Fixed">

src/yuzu_cmd/config.cpp

@@ -452,7 +452,7 @@ void Config::ReadValues() {
     Settings::values.yuzu_token = sdl2_config->Get("WebService", "yuzu_token", "");
 
     // Services
-    Settings::values.bcat_backend = sdl2_config->Get("Services", "bcat_backend", "boxcat");
+    Settings::values.bcat_backend = sdl2_config->Get("Services", "bcat_backend", "null");
     Settings::values.bcat_boxcat_local =
         sdl2_config->GetBoolean("Services", "bcat_boxcat_local", false);
 }