Create python-publish.yml

Replace GLSL as the default OpenGL shader backend

.github/workflows/python-publish.yml

@@ -0,0 +1,39 @@
+# This workflow will upload a Python Package using Twine when a release is created
+# For more information see: https://docs.github.com/en/actions/automating-builds-and-tests/building-and-testing-python#publishing-to-package-registries
+
+# This workflow uses actions that are not certified by GitHub.
+# They are provided by a third-party and are governed by
+# separate terms of service, privacy policy, and support
+# documentation.
+
+name: Upload Python Package
+
+on:
+  release:
+    types: [published]
+
+permissions:
+  contents: read
+
+jobs:
+  deploy:
+
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v3
+    - name: Set up Python
+      uses: actions/setup-python@v3
+      with:
+        python-version: '3.x'
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        pip install build
+    - name: Build package
+      run: python -m build
+    - name: Publish package
+      uses: pypa/gh-action-pypi-publish@27b31702a0e7fc50959f5ad993c78deac1bdfc29
+      with:
+        user: __token__
+        password: ${{ secrets.PYPI_API_TOKEN }}

src/audio_core/renderer/command/command_buffer.cpp

@@ -460,21 +460,23 @@ void CommandBuffer::GenerateDeviceSinkCommand(const s32 node_id, const s16 buffe
 
     cmd.session_id = session_id;
 
+    cmd.input_count = parameter.input_count;
+    s16 max_input{0};
+    for (u32 i = 0; i < parameter.input_count; i++) {
+        cmd.inputs[i] = buffer_offset + parameter.inputs[i];
+        max_input = std::max(max_input, cmd.inputs[i]);
+    }
+
     if (state.upsampler_info != nullptr) {
         const auto size_{state.upsampler_info->sample_count * parameter.input_count};
         const auto size_bytes{size_ * sizeof(s32)};
         const auto addr{memory_pool->Translate(state.upsampler_info->samples_pos, size_bytes)};
         cmd.sample_buffer = {reinterpret_cast<s32*>(addr),
-                             parameter.input_count * state.upsampler_info->sample_count};
+                             (max_input + 1) * state.upsampler_info->sample_count};
     } else {
         cmd.sample_buffer = samples_buffer;
     }
 
-    cmd.input_count = parameter.input_count;
-    for (u32 i = 0; i < parameter.input_count; i++) {
-        cmd.inputs[i] = buffer_offset + parameter.inputs[i];
-    }
-
     GenerateEnd<DeviceSinkCommand>(cmd);
 }
 

src/audio_core/sink/sink_stream.cpp

@@ -266,20 +266,19 @@ void SinkStream::ProcessAudioOutAndRender(std::span<s16> output_buffer, std::siz
 }
 
 void SinkStream::Stall() {
-    std::scoped_lock lk{stall_guard};
-    if (stalled_lock) {
+    if (stalled) {
         return;
     }
-    stalled_lock = system.StallProcesses();
+    stalled = true;
+    system.StallProcesses();
 }
 
 void SinkStream::Unstall() {
-    std::scoped_lock lk{stall_guard};
-    if (!stalled_lock) {
+    if (!stalled) {
         return;
     }
     system.UnstallProcesses();
-    stalled_lock.unlock();
+    stalled = false;
 }
 
 } // namespace AudioCore::Sink

src/audio_core/sink/sink_stream.h

@@ -6,7 +6,6 @@
 #include <array>
 #include <atomic>
 #include <memory>
-#include <mutex>
 #include <span>
 #include <vector>
 
@@ -241,8 +240,8 @@ private:
     f32 system_volume{1.0f};
     /// Set via IAudioDevice service calls
     f32 device_volume{1.0f};
-    std::mutex stall_guard;
-    std::unique_lock<std::mutex> stalled_lock;
+    /// True if coretiming has been stalled
+    bool stalled{false};
 };
 
 using SinkStreamPtr = std::unique_ptr<SinkStream>;

src/common/settings.h

@@ -442,7 +442,7 @@ struct Values {
     SwitchableSetting<NvdecEmulation> nvdec_emulation{NvdecEmulation::GPU, "nvdec_emulation"};
     SwitchableSetting<bool> accelerate_astc{true, "accelerate_astc"};
     SwitchableSetting<bool> use_vsync{true, "use_vsync"};
-    SwitchableSetting<ShaderBackend, true> shader_backend{ShaderBackend::GLASM, ShaderBackend::GLSL,
+    SwitchableSetting<ShaderBackend, true> shader_backend{ShaderBackend::GLSL, ShaderBackend::GLSL,
                                                           ShaderBackend::SPIRV, "shader_backend"};
     SwitchableSetting<bool> use_asynchronous_shaders{false, "use_asynchronous_shaders"};
     SwitchableSetting<bool> use_fast_gpu_time{true, "use_fast_gpu_time"};

src/core/core.cpp

@@ -189,7 +189,7 @@ struct System::Impl {
 
         kernel.Suspend(false);
         core_timing.SyncPause(false);
-        is_paused.store(false, std::memory_order_relaxed);
+        is_paused = false;
 
         return status;
     }
@@ -200,13 +200,14 @@ struct System::Impl {
 
         core_timing.SyncPause(true);
         kernel.Suspend(true);
-        is_paused.store(true, std::memory_order_relaxed);
+        is_paused = true;
 
         return status;
     }
 
     bool IsPaused() const {
-        return is_paused.load(std::memory_order_relaxed);
+        std::unique_lock lk(suspend_guard);
+        return is_paused;
     }
 
     std::unique_lock<std::mutex> StallProcesses() {
@@ -217,7 +218,7 @@ struct System::Impl {
     }
 
     void UnstallProcesses() {
-        if (!IsPaused()) {
+        if (!is_paused) {
             core_timing.SyncPause(false);
             kernel.Suspend(false);
         }
@@ -464,7 +465,7 @@ struct System::Impl {
     }
 
     mutable std::mutex suspend_guard;
-    std::atomic_bool is_paused{};
+    bool is_paused{};
     std::atomic<bool> is_shutting_down{};
 
     Timing::CoreTiming core_timing;

src/video_core/engines/maxwell_3d.cpp

@@ -126,6 +126,7 @@ void Maxwell3D::InitializeRegisterDefaults() {
     draw_command[MAXWELL3D_REG_INDEX(draw_inline_index)] = true;
     draw_command[MAXWELL3D_REG_INDEX(inline_index_2x16.even)] = true;
     draw_command[MAXWELL3D_REG_INDEX(inline_index_4x8.index0)] = true;
+    draw_command[MAXWELL3D_REG_INDEX(draw.instance_id)] = true;
 }
 
 void Maxwell3D::ProcessMacro(u32 method, const u32* base_start, u32 amount, bool is_last_call) {
@@ -285,31 +286,58 @@ void Maxwell3D::CallMethod(u32 method, u32 method_argument, bool is_last_call) {
     ASSERT_MSG(method < Regs::NUM_REGS,
                "Invalid Maxwell3D register, increase the size of the Regs structure");
 
+    const u32 argument = ProcessShadowRam(method, method_argument);
+    ProcessDirtyRegisters(method, argument);
+
     if (draw_command[method]) {
         regs.reg_array[method] = method_argument;
         deferred_draw_method.push_back(method);
-        auto u32_to_u8 = [&](const u32 argument) {
-            inline_index_draw_indexes.push_back(static_cast<u8>(argument & 0x000000ff));
-            inline_index_draw_indexes.push_back(static_cast<u8>((argument & 0x0000ff00) >> 8));
-            inline_index_draw_indexes.push_back(static_cast<u8>((argument & 0x00ff0000) >> 16));
-            inline_index_draw_indexes.push_back(static_cast<u8>((argument & 0xff000000) >> 24));
+        auto update_inline_index = [&](const u32 index) {
+            inline_index_draw_indexes.push_back(static_cast<u8>(index & 0x000000ff));
+            inline_index_draw_indexes.push_back(static_cast<u8>((index & 0x0000ff00) >> 8));
+            inline_index_draw_indexes.push_back(static_cast<u8>((index & 0x00ff0000) >> 16));
+            inline_index_draw_indexes.push_back(static_cast<u8>((index & 0xff000000) >> 24));
+            draw_mode = DrawMode::InlineIndex;
         };
-        if (MAXWELL3D_REG_INDEX(draw_inline_index) == method) {
-            u32_to_u8(method_argument);
-        } else if (MAXWELL3D_REG_INDEX(inline_index_2x16.even) == method) {
-            u32_to_u8(regs.inline_index_2x16.even);
-            u32_to_u8(regs.inline_index_2x16.odd);
-        } else if (MAXWELL3D_REG_INDEX(inline_index_4x8.index0) == method) {
-            u32_to_u8(regs.inline_index_4x8.index0);
-            u32_to_u8(regs.inline_index_4x8.index1);
-            u32_to_u8(regs.inline_index_4x8.index2);
-            u32_to_u8(regs.inline_index_4x8.index3);
+        switch (method) {
+        case MAXWELL3D_REG_INDEX(draw.end):
+            switch (draw_mode) {
+            case DrawMode::General:
+                ProcessDraw(1);
+                break;
+            case DrawMode::InlineIndex:
+                regs.index_buffer.count = static_cast<u32>(inline_index_draw_indexes.size() / 4);
+                regs.index_buffer.format = Regs::IndexFormat::UnsignedInt;
+                ProcessDraw(1);
+                inline_index_draw_indexes.clear();
+                break;
+            case DrawMode::Instance:
+                break;
+            }
+            break;
+        case MAXWELL3D_REG_INDEX(draw_inline_index):
+            update_inline_index(method_argument);
+            break;
+        case MAXWELL3D_REG_INDEX(inline_index_2x16.even):
+            update_inline_index(regs.inline_index_2x16.even);
+            update_inline_index(regs.inline_index_2x16.odd);
+            break;
+        case MAXWELL3D_REG_INDEX(inline_index_4x8.index0):
+            update_inline_index(regs.inline_index_4x8.index0);
+            update_inline_index(regs.inline_index_4x8.index1);
+            update_inline_index(regs.inline_index_4x8.index2);
+            update_inline_index(regs.inline_index_4x8.index3);
+            break;
+        case MAXWELL3D_REG_INDEX(draw.instance_id):
+            draw_mode =
+                (regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::Subsequent) ||
+                        (regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::Unchanged)
+                    ? DrawMode::Instance
+                    : DrawMode::General;
+            break;
         }
     } else {
         ProcessDeferredDraw();
-
-        const u32 argument = ProcessShadowRam(method, method_argument);
-        ProcessDirtyRegisters(method, argument);
         ProcessMethodCall(method, argument, method_argument, is_last_call);
     }
 }
@@ -620,57 +648,27 @@ void Maxwell3D::ProcessDraw(u32 instance_count) {
 }
 
 void Maxwell3D::ProcessDeferredDraw() {
-    if (deferred_draw_method.empty()) {
+    if (draw_mode != DrawMode::Instance || deferred_draw_method.empty()) {
         return;
     }
 
-    enum class DrawMode {
-        Undefined,
-        General,
-        Instance,
-    };
-    DrawMode draw_mode{DrawMode::Undefined};
     u32 method_count = static_cast<u32>(deferred_draw_method.size());
-    u32 method = deferred_draw_method[method_count - 1];
-    if (MAXWELL3D_REG_INDEX(draw.end) != method) {
-        return;
-    }
-    draw_mode = (regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::Subsequent) ||
-                        (regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::Unchanged)
-                    ? DrawMode::Instance
-                    : DrawMode::General;
-    u32 instance_count = 0;
-    if (draw_mode == DrawMode::Instance) {
-        u32 vertex_buffer_count = 0;
-        u32 index_buffer_count = 0;
-        for (u32 index = 0; index < method_count; ++index) {
-            method = deferred_draw_method[index];
-            if (method == MAXWELL3D_REG_INDEX(vertex_buffer.count)) {
-                instance_count = ++vertex_buffer_count;
-            } else if (method == MAXWELL3D_REG_INDEX(index_buffer.count)) {
-                instance_count = ++index_buffer_count;
-            }
-        }
-        ASSERT_MSG(!(vertex_buffer_count && index_buffer_count),
-                   "Instance both indexed and direct?");
-    } else {
-        instance_count = 1;
-        for (u32 index = 0; index < method_count; ++index) {
-            method = deferred_draw_method[index];
-            if (MAXWELL3D_REG_INDEX(draw_inline_index) == method ||
-                MAXWELL3D_REG_INDEX(inline_index_2x16.even) == method ||
-                MAXWELL3D_REG_INDEX(inline_index_4x8.index0) == method) {
-                regs.index_buffer.count = static_cast<u32>(inline_index_draw_indexes.size() / 4);
-                regs.index_buffer.format = Regs::IndexFormat::UnsignedInt;
-                break;
-            }
+    u32 instance_count = 1;
+    u32 vertex_buffer_count = 0;
+    u32 index_buffer_count = 0;
+    for (u32 index = 0; index < method_count; ++index) {
+        u32 method = deferred_draw_method[index];
+        if (method == MAXWELL3D_REG_INDEX(vertex_buffer.count)) {
+            instance_count = ++vertex_buffer_count;
+        } else if (method == MAXWELL3D_REG_INDEX(index_buffer.count)) {
+            instance_count = ++index_buffer_count;
         }
     }
+    ASSERT_MSG(!(vertex_buffer_count && index_buffer_count), "Instance both indexed and direct?");
 
     ProcessDraw(instance_count);
 
     deferred_draw_method.clear();
-    inline_index_draw_indexes.clear();
 }
 
 } // namespace Tegra::Engines

src/video_core/engines/maxwell_3d.h

@@ -3148,10 +3148,12 @@ private:
     /// Handles use of topology overrides (e.g., to avoid using a topology assigned from a macro)
     void ProcessTopologyOverride();
 
-    void ProcessDraw(u32 instance_count = 1);
-
+    /// Handles deferred draw(e.g., instance draw).
     void ProcessDeferredDraw();
 
+    /// Handles a draw.
+    void ProcessDraw(u32 instance_count = 1);
+
     /// Returns a query's value or an empty object if the value will be deferred through a cache.
     std::optional<u64> GetQueryResult();
 
@@ -3178,6 +3180,8 @@ private:
 
     std::array<bool, Regs::NUM_REGS> draw_command{};
     std::vector<u32> deferred_draw_method;
+    enum class DrawMode : u32 { General = 0, Instance, InlineIndex };
+    DrawMode draw_mode{DrawMode::General};
 };
 
 #define ASSERT_REG_POSITION(field_name, position)                                                  \

src/video_core/renderer_vulkan/maxwell_to_vk.cpp

@@ -150,7 +150,7 @@ struct FormatTuple {
     {VK_FORMAT_BC6H_UFLOAT_BLOCK},                             // BC6H_UFLOAT
     {VK_FORMAT_BC6H_SFLOAT_BLOCK},                             // BC6H_SFLOAT
     {VK_FORMAT_ASTC_4x4_UNORM_BLOCK},                          // ASTC_2D_4X4_UNORM
-    {VK_FORMAT_B8G8R8A8_UNORM, Attachable},                    // B8G8R8A8_UNORM
+    {VK_FORMAT_B8G8R8A8_UNORM, Attachable | Storage},          // B8G8R8A8_UNORM
     {VK_FORMAT_R32G32B32A32_SFLOAT, Attachable | Storage},     // R32G32B32A32_FLOAT
     {VK_FORMAT_R32G32B32A32_SINT, Attachable | Storage},       // R32G32B32A32_SINT
     {VK_FORMAT_R32G32_SFLOAT, Attachable | Storage},           // R32G32_FLOAT
@@ -160,7 +160,7 @@ struct FormatTuple {
     {VK_FORMAT_R16_UNORM, Attachable | Storage},               // R16_UNORM
     {VK_FORMAT_R16_SNORM, Attachable | Storage},               // R16_SNORM
     {VK_FORMAT_R16_UINT, Attachable | Storage},                // R16_UINT
-    {VK_FORMAT_UNDEFINED},                                     // R16_SINT
+    {VK_FORMAT_R16_SINT, Attachable | Storage},                // R16_SINT
     {VK_FORMAT_R16G16_UNORM, Attachable | Storage},            // R16G16_UNORM
     {VK_FORMAT_R16G16_SFLOAT, Attachable | Storage},           // R16G16_FLOAT
     {VK_FORMAT_R16G16_UINT, Attachable | Storage},             // R16G16_UINT
@@ -184,7 +184,7 @@ struct FormatTuple {
     {VK_FORMAT_BC2_SRGB_BLOCK},                                // BC2_SRGB
     {VK_FORMAT_BC3_SRGB_BLOCK},                                // BC3_SRGB
     {VK_FORMAT_BC7_SRGB_BLOCK},                                // BC7_SRGB
-    {VK_FORMAT_R4G4B4A4_UNORM_PACK16, Attachable},             // A4B4G4R4_UNORM
+    {VK_FORMAT_R4G4B4A4_UNORM_PACK16},                         // A4B4G4R4_UNORM
     {VK_FORMAT_R4G4_UNORM_PACK8},                              // G4R4_UNORM
     {VK_FORMAT_ASTC_4x4_SRGB_BLOCK},                           // ASTC_2D_4X4_SRGB
     {VK_FORMAT_ASTC_8x8_SRGB_BLOCK},                           // ASTC_2D_8X8_SRGB