Merge pull request #2830 from FearlessTobi/port-4911

Port citra-emu/citra#4911: "Add cancel option to analog stick configuration"
Merge pull request #2834 from Morph1984/audrenu_QueryAudioDeviceInputEvent
2019-09-05 09:53:44 +10:00 · 2019-09-05 09:51:50 +10:00 · 2019-09-05 09:50:38 +10:00 · 2019-09-05 09:50:17 +10:00 · 2019-09-04 17:42:13 -04:00 · 2019-09-04 15:03:51 -03:00
118 changed files with 2549 additions and 1251 deletions
@@ -10,7 +10,7 @@ jobs:
        BuildSuffix: 'windows-testing'
        ScriptFolder: 'windows'
  steps:
-  - script: sudo apt upgrade python3-pip && pip install requests urllib3
+  - script: sudo apt-get update && sudo apt-get --only-upgrade -y install python3-pip && pip install requests urllib3
    displayName: 'Prepare Environment'
  - task: PythonScript@0
    condition: eq(variables['Build.Reason'], 'PullRequest')
@@ -81,6 +81,7 @@ set(HASH_FILES
    "${VIDEO_CORE}/shader/decode/register_set_predicate.cpp"
    "${VIDEO_CORE}/shader/decode/shift.cpp"
    "${VIDEO_CORE}/shader/decode/video.cpp"
+    "${VIDEO_CORE}/shader/decode/warp.cpp"
    "${VIDEO_CORE}/shader/decode/xmad.cpp"
    "${VIDEO_CORE}/shader/control_flow.cpp"
    "${VIDEO_CORE}/shader/control_flow.h"
@@ -0,0 +1,31 @@
+The icons in this folder and its subfolders have the following licenses:
+
+Icon Name | License | Origin/Author
+--- | --- | ---
+qt_themes/default/icons/16x16/checked.png | Free for non-commercial use
+qt_themes/default/icons/16x16/failed.png | Free for non-commercial use
+qt_themes/default/icons/16x16/lock.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/256x256/plus_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/48x48/bad_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/48x48/chip.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/48x48/folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/48x48/plus.png | CC0 1.0 | Designed by BreadFish64 from the Citra team
+qt_themes/default/icons/48x48/sd_card.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/16x16/checked.png | Free for non-commercial use
+qt_themes/qdarkstyle/icons/16x16/failed.png | Free for non-commercial use
+qt_themes/qdarkstyle/icons/16x16/lock.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/256x256/plus_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/48x48/bad_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/48x48/chip.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/48x48/folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/48x48/plus.png | CC0 1.0 | Designed by BreadFish64 from the Citra team
+qt_themes/qdarkstyle/icons/48x48/sd_card.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/16x16/lock.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/256x256/plus_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/bad_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/chip.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/plus.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/sd_card.png | CC BY-ND 3.0 | https://icons8.com
+
+<!-- TODO: Add the license of the yuzu icon -->
@@ -0,0 +1,14 @@
+[Icon Theme]
+Name=colorful
+Comment=Colorful theme
+Inherits=default
+Directories=16x16,48x48,256x256
+ 
+[16x16]
+Size=16
+
+[48x48]
+Size=48
+
+[256x256]
+Size=256
@@ -0,0 +1,15 @@
+<RCC>
+    <qresource prefix="icons/colorful">
+        <file alias="index.theme">icons/index.theme</file>
+        <file alias="16x16/lock.png">icons/16x16/lock.png</file>
+        <file alias="48x48/bad_folder.png">icons/48x48/bad_folder.png</file>
+        <file alias="48x48/chip.png">icons/48x48/chip.png</file>
+        <file alias="48x48/folder.png">icons/48x48/folder.png</file>
+        <file alias="48x48/plus.png">icons/48x48/plus.png</file>
+        <file alias="48x48/sd_card.png">icons/48x48/sd_card.png</file>
+        <file alias="256x256/plus_folder.png">icons/256x256/plus_folder.png</file>
+    </qresource>
+    <qresource prefix="colorful">
+        <file>style.qss</file>
+    </qresource>
+</RCC>
@@ -0,0 +1,4 @@
+/*
+    This file is intentionally left blank.
+    We do not want to apply any stylesheet for colorful, only icons.
+*/
@@ -0,0 +1,8 @@
+[Icon Theme]
+Name=colorful_dark
+Comment=Colorful theme (Dark style)
+Inherits=default
+Directories=16x16
+ 
+[16x16]
+Size=16
@@ -0,0 +1,57 @@
+<RCC>
+    <qresource prefix="icons/colorful_dark">
+        <file alias="index.theme">icons/index.theme</file>
+        <file alias="16x16/lock.png">icons/16x16/lock.png</file>
+        <file alias="48x48/bad_folder.png">../colorful/icons/48x48/bad_folder.png</file>
+        <file alias="48x48/chip.png">../colorful/icons/48x48/chip.png</file>
+        <file alias="48x48/folder.png">../colorful/icons/48x48/folder.png</file>
+        <file alias="48x48/plus.png">../colorful/icons/48x48/plus.png</file>
+        <file alias="48x48/sd_card.png">../colorful/icons/48x48/sd_card.png</file>
+        <file alias="256x256/plus_folder.png">../colorful/icons/256x256/plus_folder.png</file>
+    </qresource>
+
+    <qresource prefix="qss_icons">
+        <file alias="rc/up_arrow_disabled.png">../qdarkstyle/rc/up_arrow_disabled.png</file>
+        <file alias="rc/Hmovetoolbar.png">../qdarkstyle/rc/Hmovetoolbar.png</file>
+        <file alias="rc/stylesheet-branch-end.png">../qdarkstyle/rc/stylesheet-branch-end.png</file>
+        <file alias="rc/branch_closed-on.png">../qdarkstyle/rc/branch_closed-on.png</file>
+        <file alias="rc/stylesheet-vline.png">../qdarkstyle/rc/stylesheet-vline.png</file>
+        <file alias="rc/branch_closed.png">../qdarkstyle/rc/branch_closed.png</file>
+        <file alias="rc/branch_open-on.png">../qdarkstyle/rc/branch_open-on.png</file>
+        <file alias="rc/transparent.png">../qdarkstyle/rc/transparent.png</file>
+        <file alias="rc/right_arrow_disabled.png">../qdarkstyle/rc/right_arrow_disabled.png</file>
+        <file alias="rc/sizegrip.png">../qdarkstyle/rc/sizegrip.png</file>
+        <file alias="rc/close.png">../qdarkstyle/rc/close.png</file>
+        <file alias="rc/close-hover.png">../qdarkstyle/rc/close-hover.png</file>
+        <file alias="rc/close-pressed.png">../qdarkstyle/rc/close-pressed.png</file>
+        <file alias="rc/down_arrow.png">../qdarkstyle/rc/down_arrow.png</file>
+        <file alias="rc/Vmovetoolbar.png">../qdarkstyle/rc/Vmovetoolbar.png</file>
+        <file alias="rc/left_arrow.png">../qdarkstyle/rc/left_arrow.png</file>
+        <file alias="rc/stylesheet-branch-more.png">../qdarkstyle/rc/stylesheet-branch-more.png</file>
+        <file alias="rc/up_arrow.png">../qdarkstyle/rc/up_arrow.png</file>
+        <file alias="rc/right_arrow.png">../qdarkstyle/rc/right_arrow.png</file>
+        <file alias="rc/left_arrow_disabled.png">../qdarkstyle/rc/left_arrow_disabled.png</file>
+        <file alias="rc/Hsepartoolbar.png">../qdarkstyle/rc/Hsepartoolbar.png</file>
+        <file alias="rc/branch_open.png">../qdarkstyle/rc/branch_open.png</file>
+        <file alias="rc/Vsepartoolbar.png">../qdarkstyle/rc/Vsepartoolbar.png</file>
+        <file alias="rc/down_arrow_disabled.png">../qdarkstyle/rc/down_arrow_disabled.png</file>
+        <file alias="rc/undock.png">../qdarkstyle/rc/undock.png</file>
+        <file alias="rc/checkbox_checked_disabled.png">../qdarkstyle/rc/checkbox_checked_disabled.png</file>
+        <file alias="rc/checkbox_checked_focus.png">../qdarkstyle/rc/checkbox_checked_focus.png</file>
+        <file alias="rc/checkbox_checked.png">../qdarkstyle/rc/checkbox_checked.png</file>
+        <file alias="rc/checkbox_indeterminate.png">../qdarkstyle/rc/checkbox_indeterminate.png</file>
+        <file alias="rc/checkbox_indeterminate_focus.png">../qdarkstyle/rc/checkbox_indeterminate_focus.png</file>
+        <file alias="rc/checkbox_unchecked_disabled.png">../qdarkstyle/rc/checkbox_unchecked_disabled.png</file>
+        <file alias="rc/checkbox_unchecked_focus.png">../qdarkstyle/rc/checkbox_unchecked_focus.png</file>
+        <file alias="rc/checkbox_unchecked.png">../qdarkstyle/rc/checkbox_unchecked.png</file>
+        <file alias="rc/radio_checked_disabled.png">../qdarkstyle/rc/radio_checked_disabled.png</file>
+        <file alias="rc/radio_checked_focus.png">../qdarkstyle/rc/radio_checked_focus.png</file>
+        <file alias="rc/radio_checked.png">../qdarkstyle/rc/radio_checked.png</file>
+        <file alias="rc/radio_unchecked_disabled.png">../qdarkstyle/rc/radio_unchecked_disabled.png</file>
+        <file alias="rc/radio_unchecked_focus.png">../qdarkstyle/rc/radio_unchecked_focus.png</file>
+        <file alias="rc/radio_unchecked.png">../qdarkstyle/rc/radio_unchecked.png</file>
+    </qresource>
+    <qresource prefix="colorful_dark">
+        <file alias="style.qss">../qdarkstyle/style.qss</file>
+    </qresource>
+</RCC>
@@ -5,7 +5,21 @@
        <file alias="16x16/checked.png">icons/16x16/checked.png</file>

        <file alias="16x16/failed.png">icons/16x16/failed.png</file>
+
+        <file alias="16x16/lock.png">icons/16x16/lock.png</file>
+
+        <file alias="48x48/bad_folder.png">icons/48x48/bad_folder.png</file>
+      
+        <file alias="48x48/chip.png">icons/48x48/chip.png</file>
+
+        <file alias="48x48/folder.png">icons/48x48/folder.png</file>
+
+        <file alias="48x48/plus.png">icons/48x48/plus.png</file>
+      
+        <file alias="48x48/sd_card.png">icons/48x48/sd_card.png</file>
      
        <file alias="256x256/yuzu.png">icons/256x256/yuzu.png</file>
+
+        <file alias="256x256/plus_folder.png">icons/256x256/plus_folder.png</file>
    </qresource>
 </RCC>
@@ -1,10 +1,13 @@
 [Icon Theme]
 Name=default
 Comment=default theme
-Directories=16x16,256x256
+Directories=16x16,48x48,256x256
 
 [16x16]
 Size=16
+
+[48x48]
+Size=48
 
 [256x256]
 Size=256
@@ -2,10 +2,13 @@
 Name=qdarkstyle
 Comment=dark theme
 Inherits=default
-Directories=16x16,256x256
+Directories=16x16,48x48,256x256
 
 [16x16]
 Size=16
- 
+
+[48x48]
+Size=48
+
 [256x256]
 Size=256
@@ -1,6 +1,13 @@
 <RCC>
  <qresource prefix="icons/qdarkstyle">
    <file alias="index.theme">icons/index.theme</file>
+    <file alias="16x16/lock.png">icons/16x16/lock.png</file>
+    <file alias="48x48/bad_folder.png">icons/48x48/bad_folder.png</file>
+    <file alias="48x48/chip.png">icons/48x48/chip.png</file>
+    <file alias="48x48/folder.png">icons/48x48/folder.png</file>
+    <file alias="48x48/plus.png">icons/48x48/plus.png</file>
+    <file alias="48x48/sd_card.png">icons/48x48/sd_card.png</file>
+    <file alias="256x256/plus_folder.png">icons/256x256/plus_folder.png</file>
  </qresource>
  <qresource prefix="qss_icons">
    <file>rc/up_arrow_disabled.png</file>
@@ -337,3 +337,19 @@ proprietary programs.  If your program is a subroutine library, you may
 consider it more useful to permit linking proprietary applications with the
 library.  If this is what you want to do, use the GNU Lesser General
 Public License instead of this License.
+
+
+The icons used in this project have the following licenses:
+
+Icon Name                          | License       | Origin/Author
+---                                | ---           | ---
+checked.png                        | Free for non-commercial use
+failed.png                         | Free for non-commercial use
+lock.png                           | CC BY-ND 3.0  | https://icons8.com
+plus_folder.png                    | CC BY-ND 3.0  | https://icons8.com
+bad_folder.png                     | CC BY-ND 3.0  | https://icons8.com
+chip.png                           | CC BY-ND 3.0  | https://icons8.com
+folder.png                         | CC BY-ND 3.0  | https://icons8.com
+plus.png (Default, Dark)           | CC0 1.0       | Designed by BreadFish64 from the Citra team
+plus.png (Colorful, Colorful Dark) | CC BY-ND 3.0  | https://icons8.com
+sd_card.png | CC BY-ND 3.0 | https://icons8.com
@@ -55,6 +55,7 @@ add_custom_command(OUTPUT scm_rev.cpp
      "${VIDEO_CORE}/shader/decode/register_set_predicate.cpp"
      "${VIDEO_CORE}/shader/decode/shift.cpp"
      "${VIDEO_CORE}/shader/decode/video.cpp"
+      "${VIDEO_CORE}/shader/decode/warp.cpp"
      "${VIDEO_CORE}/shader/decode/xmad.cpp"
      "${VIDEO_CORE}/shader/control_flow.cpp"
      "${VIDEO_CORE}/shader/control_flow.h"
@@ -1057,6 +1057,7 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
        {120, nullptr, "ExecuteProgram"},
        {121, nullptr, "ClearUserChannel"},
        {122, nullptr, "UnpopToUserChannel"},
+        {130, &IApplicationFunctions::GetGpuErrorDetectedSystemEvent, "GetGpuErrorDetectedSystemEvent"},
        {500, nullptr, "StartContinuousRecordingFlushForDebug"},
        {1000, nullptr, "CreateMovieMaker"},
        {1001, nullptr, "PrepareForJit"},
@@ -1064,6 +1065,10 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
    // clang-format on

    RegisterHandlers(functions);
+
+    auto& kernel = Core::System::GetInstance().Kernel();
+    gpu_error_detected_event = Kernel::WritableEvent::CreateEventPair(
+        kernel, Kernel::ResetType::Manual, "IApplicationFunctions:GpuErrorDetectedSystemEvent");
 }

 IApplicationFunctions::~IApplicationFunctions() = default;
@@ -1285,6 +1290,14 @@ void IApplicationFunctions::GetSaveDataSize(Kernel::HLERequestContext& ctx) {
    rb.Push(size.journal);
 }

+void IApplicationFunctions::GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+
+    IPC::ResponseBuilder rb{ctx, 2, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushCopyObjects(gpu_error_detected_event.readable);
+}
+
 void InstallInterfaces(SM::ServiceManager& service_manager,
                       std::shared_ptr<NVFlinger::NVFlinger> nvflinger, Core::System& system) {
    auto message_queue = std::make_shared<AppletMessageQueue>();
@@ -242,6 +242,9 @@ private:
    void BeginBlockingHomeButton(Kernel::HLERequestContext& ctx);
    void EndBlockingHomeButton(Kernel::HLERequestContext& ctx);
    void EnableApplicationCrashReport(Kernel::HLERequestContext& ctx);
+    void GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx);
+
+    Kernel::EventPair gpu_error_detected_event;
 };

 class IHomeMenuFunctions final : public ServiceFramework<IHomeMenuFunctions> {
@@ -165,15 +165,15 @@ public:
        static const FunctionInfo functions[] = {
            {0, &IAudioDevice::ListAudioDeviceName, "ListAudioDeviceName"},
            {1, &IAudioDevice::SetAudioDeviceOutputVolume, "SetAudioDeviceOutputVolume"},
-            {2, nullptr, "GetAudioDeviceOutputVolume"},
+            {2, &IAudioDevice::GetAudioDeviceOutputVolume, "GetAudioDeviceOutputVolume"},
            {3, &IAudioDevice::GetActiveAudioDeviceName, "GetActiveAudioDeviceName"},
            {4, &IAudioDevice::QueryAudioDeviceSystemEvent, "QueryAudioDeviceSystemEvent"},
            {5, &IAudioDevice::GetActiveChannelCount, "GetActiveChannelCount"},
            {6, &IAudioDevice::ListAudioDeviceName, "ListAudioDeviceNameAuto"},
            {7, &IAudioDevice::SetAudioDeviceOutputVolume, "SetAudioDeviceOutputVolumeAuto"},
-            {8, nullptr, "GetAudioDeviceOutputVolumeAuto"},
+            {8, &IAudioDevice::GetAudioDeviceOutputVolume, "GetAudioDeviceOutputVolumeAuto"},
            {10, &IAudioDevice::GetActiveAudioDeviceName, "GetActiveAudioDeviceNameAuto"},
-            {11, nullptr, "QueryAudioDeviceInputEvent"},
+            {11, &IAudioDevice::QueryAudioDeviceInputEvent, "QueryAudioDeviceInputEvent"},
            {12, &IAudioDevice::QueryAudioDeviceOutputEvent, "QueryAudioDeviceOutputEvent"},
            {13, nullptr, "GetAudioSystemMasterVolumeSetting"},
        };
@@ -183,6 +183,10 @@ public:
        buffer_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
                                                              "IAudioOutBufferReleasedEvent");

+        // Should be similar to audio_output_device_switch_event
+        audio_input_device_switch_event = Kernel::WritableEvent::CreateEventPair(
+            kernel, Kernel::ResetType::Automatic, "IAudioDevice:AudioInputDeviceSwitchedEvent");
+
        // Should only be signalled when an audio output device has been changed, example: speaker
        // to headset
        audio_output_device_switch_event = Kernel::WritableEvent::CreateEventPair(
@@ -246,6 +250,19 @@ private:
        rb.Push(RESULT_SUCCESS);
    }

+    void GetAudioDeviceOutputVolume(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+
+        const auto device_name_buffer = ctx.ReadBuffer();
+        const std::string name = Common::StringFromBuffer(device_name_buffer);
+
+        LOG_WARNING(Service_Audio, "(STUBBED) called. name={}", name);
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push(1.0f);
+    }
+
    void GetActiveAudioDeviceName(Kernel::HLERequestContext& ctx) {
        LOG_WARNING(Service_Audio, "(STUBBED) called");

@@ -279,6 +296,15 @@ private:
        rb.Push<u32>(1);
    }

+    // Should be similar to QueryAudioDeviceOutputEvent
+    void QueryAudioDeviceInputEvent(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_Audio, "(STUBBED) called");
+
+        IPC::ResponseBuilder rb{ctx, 2, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushCopyObjects(audio_input_device_switch_event.readable);
+    }
+
    void QueryAudioDeviceOutputEvent(Kernel::HLERequestContext& ctx) {
        LOG_DEBUG(Service_Audio, "called");

@@ -289,6 +315,7 @@ private:

    u32_le revision = 0;
    Kernel::EventPair buffer_event;
+    Kernel::EventPair audio_input_device_switch_event;
    Kernel::EventPair audio_output_device_switch_event;

 }; // namespace Audio
@@ -5,7 +5,7 @@
 #include <array>
 #include <cstring>
 #include <ctime>
-#include <fmt/time.h>
+#include <fmt/chrono.h>
 #include "common/file_util.h"
 #include "common/logging/log.h"
 #include "common/scm_rev.h"
@@ -636,10 +636,15 @@ Controller_NPad::LedPattern Controller_NPad::GetLedPattern(u32 npad_id) {
        return LedPattern{0, 0, 0, 0};
    };
 }
+
 void Controller_NPad::SetVibrationEnabled(bool can_vibrate) {
    can_controllers_vibrate = can_vibrate;
 }

+bool Controller_NPad::IsVibrationEnabled() const {
+    return can_controllers_vibrate;
+}
+
 void Controller_NPad::ClearAllConnectedControllers() {
    for (auto& controller : connected_controllers) {
        if (controller.is_connected && controller.type != NPadControllerType::None) {
@@ -648,6 +653,7 @@ void Controller_NPad::ClearAllConnectedControllers() {
        }
    }
 }
+
 void Controller_NPad::DisconnectAllConnectedControllers() {
    std::for_each(connected_controllers.begin(), connected_controllers.end(),
                  [](ControllerHolder& controller) { controller.is_connected = false; });
@@ -119,6 +119,7 @@ public:
    void DisconnectNPad(u32 npad_id);
    LedPattern GetLedPattern(u32 npad_id);
    void SetVibrationEnabled(bool can_vibrate);
+    bool IsVibrationEnabled() const;
    void ClearAllConnectedControllers();
    void DisconnectAllConnectedControllers();
    void ConnectAllDisconnectedControllers();
@@ -216,8 +216,8 @@ Hid::Hid() : ServiceFramework("hid") {
        {201, &Hid::SendVibrationValue, "SendVibrationValue"},
        {202, &Hid::GetActualVibrationValue, "GetActualVibrationValue"},
        {203, &Hid::CreateActiveVibrationDeviceList, "CreateActiveVibrationDeviceList"},
-        {204, nullptr, "PermitVibration"},
-        {205, nullptr, "IsVibrationPermitted"},
+        {204, &Hid::PermitVibration, "PermitVibration"},
+        {205, &Hid::IsVibrationPermitted, "IsVibrationPermitted"},
        {206, &Hid::SendVibrationValues, "SendVibrationValues"},
        {207, nullptr, "SendVibrationGcErmCommand"},
        {208, nullptr, "GetActualVibrationGcErmCommand"},
@@ -679,6 +679,27 @@ void Hid::CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx) {
    rb.PushIpcInterface<IActiveVibrationDeviceList>();
 }

+void Hid::PermitVibration(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto can_vibrate{rp.Pop<bool>()};
+    applet_resource->GetController<Controller_NPad>(HidController::NPad)
+        .SetVibrationEnabled(can_vibrate);
+
+    LOG_DEBUG(Service_HID, "called, can_vibrate={}", can_vibrate);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::IsVibrationPermitted(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_HID, "called");
+
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push(
+        applet_resource->GetController<Controller_NPad>(HidController::NPad).IsVibrationEnabled());
+}
+
 void Hid::ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) {
    IPC::RequestParser rp{ctx};
    const auto applet_resource_user_id{rp.Pop<u64>()};
@@ -114,6 +114,8 @@ private:
    void SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx);
    void GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx);
    void CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx);
+    void PermitVibration(Kernel::HLERequestContext& ctx);
+    void IsVibrationPermitted(Kernel::HLERequestContext& ctx);
    void ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx);
    void StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx);
    void StopSixAxisSensor(Kernel::HLERequestContext& ctx);
@@ -175,6 +175,10 @@ MiiStoreData ConvertInfoToStoreData(const MiiInfo& info) {
 } // namespace

 std::ostream& operator<<(std::ostream& os, Source source) {
+    if (static_cast<std::size_t>(source) >= SOURCE_NAMES.size()) {
+        return os << "[UNKNOWN SOURCE]";
+    }
+
    os << SOURCE_NAMES.at(static_cast<std::size_t>(source));
    return os;
 }
@@ -37,7 +37,7 @@ void nvdisp_disp0::flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u3
        transform, crop_rect};

    system.GetPerfStats().EndGameFrame();
-    system.GPU().SwapBuffers(framebuffer);
+    system.GPU().SwapBuffers(&framebuffer);
 }

 } // namespace Service::Nvidia::Devices
@@ -5,8 +5,8 @@
 #include <ctime>
 #include <fstream>

+#include <fmt/chrono.h>
 #include <fmt/format.h>
-#include <fmt/time.h>
 #include <json.hpp>

 #include "common/file_util.h"
@@ -1,5 +1,7 @@
 add_library(video_core STATIC
-    buffer_cache.h
+    buffer_cache/buffer_block.h
+    buffer_cache/buffer_cache.h
+    buffer_cache/map_interval.h
    dma_pusher.cpp
    dma_pusher.h
    debug_utils/debug_utils.cpp
@@ -100,6 +102,7 @@ add_library(video_core STATIC
    shader/decode/integer_set.cpp
    shader/decode/half_set.cpp
    shader/decode/video.cpp
+    shader/decode/warp.cpp
    shader/decode/xmad.cpp
    shader/decode/other.cpp
    shader/control_flow.cpp
@@ -1,299 +0,0 @@
-// Copyright 2019 yuzu Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <array>
-#include <memory>
-#include <mutex>
-#include <unordered_map>
-#include <unordered_set>
-#include <utility>
-#include <vector>
-
-#include "common/alignment.h"
-#include "common/common_types.h"
-#include "core/core.h"
-#include "video_core/memory_manager.h"
-#include "video_core/rasterizer_cache.h"
-
-namespace VideoCore {
-class RasterizerInterface;
-}
-
-namespace VideoCommon {
-
-template <typename BufferStorageType>
-class CachedBuffer final : public RasterizerCacheObject {
-public:
-    explicit CachedBuffer(VAddr cpu_addr, u8* host_ptr)
-        : RasterizerCacheObject{host_ptr}, host_ptr{host_ptr}, cpu_addr{cpu_addr} {}
-    ~CachedBuffer() override = default;
-
-    VAddr GetCpuAddr() const override {
-        return cpu_addr;
-    }
-
-    std::size_t GetSizeInBytes() const override {
-        return size;
-    }
-
-    u8* GetWritableHostPtr() const {
-        return host_ptr;
-    }
-
-    std::size_t GetSize() const {
-        return size;
-    }
-
-    std::size_t GetCapacity() const {
-        return capacity;
-    }
-
-    bool IsInternalized() const {
-        return is_internal;
-    }
-
-    const BufferStorageType& GetBuffer() const {
-        return buffer;
-    }
-
-    void SetSize(std::size_t new_size) {
-        size = new_size;
-    }
-
-    void SetInternalState(bool is_internal_) {
-        is_internal = is_internal_;
-    }
-
-    BufferStorageType ExchangeBuffer(BufferStorageType buffer_, std::size_t new_capacity) {
-        capacity = new_capacity;
-        std::swap(buffer, buffer_);
-        return buffer_;
-    }
-
-private:
-    u8* host_ptr{};
-    VAddr cpu_addr{};
-    std::size_t size{};
-    std::size_t capacity{};
-    bool is_internal{};
-    BufferStorageType buffer;
-};
-
-template <typename BufferStorageType, typename BufferType, typename StreamBuffer>
-class BufferCache : public RasterizerCache<std::shared_ptr<CachedBuffer<BufferStorageType>>> {
-public:
-    using Buffer = std::shared_ptr<CachedBuffer<BufferStorageType>>;
-    using BufferInfo = std::pair<const BufferType*, u64>;
-
-    explicit BufferCache(VideoCore::RasterizerInterface& rasterizer, Core::System& system,
-                         std::unique_ptr<StreamBuffer> stream_buffer)
-        : RasterizerCache<Buffer>{rasterizer}, system{system},
-          stream_buffer{std::move(stream_buffer)}, stream_buffer_handle{
-                                                       this->stream_buffer->GetHandle()} {}
-    ~BufferCache() = default;
-
-    void Unregister(const Buffer& entry) override {
-        std::lock_guard lock{RasterizerCache<Buffer>::mutex};
-        if (entry->IsInternalized()) {
-            internalized_entries.erase(entry->GetCacheAddr());
-        }
-        ReserveBuffer(entry);
-        RasterizerCache<Buffer>::Unregister(entry);
-    }
-
-    void TickFrame() {
-        marked_for_destruction_index =
-            (marked_for_destruction_index + 1) % marked_for_destruction_ring_buffer.size();
-        MarkedForDestruction().clear();
-    }
-
-    BufferInfo UploadMemory(GPUVAddr gpu_addr, std::size_t size, std::size_t alignment = 4,
-                            bool internalize = false, bool is_written = false) {
-        std::lock_guard lock{RasterizerCache<Buffer>::mutex};
-
-        auto& memory_manager = system.GPU().MemoryManager();
-        const auto host_ptr = memory_manager.GetPointer(gpu_addr);
-        if (!host_ptr) {
-            return {GetEmptyBuffer(size), 0};
-        }
-        const auto cache_addr = ToCacheAddr(host_ptr);
-
-        // Cache management is a big overhead, so only cache entries with a given size.
-        // TODO: Figure out which size is the best for given games.
-        constexpr std::size_t max_stream_size = 0x800;
-        if (!internalize && size < max_stream_size &&
-            internalized_entries.find(cache_addr) == internalized_entries.end()) {
-            return StreamBufferUpload(host_ptr, size, alignment);
-        }
-
-        auto entry = RasterizerCache<Buffer>::TryGet(cache_addr);
-        if (!entry) {
-            return FixedBufferUpload(gpu_addr, host_ptr, size, internalize, is_written);
-        }
-
-        if (entry->GetSize() < size) {
-            IncreaseBufferSize(entry, size);
-        }
-        if (is_written) {
-            entry->MarkAsModified(true, *this);
-        }
-        return {ToHandle(entry->GetBuffer()), 0};
-    }
-
-    /// Uploads from a host memory. Returns the OpenGL buffer where it's located and its offset.
-    BufferInfo UploadHostMemory(const void* raw_pointer, std::size_t size,
-                                std::size_t alignment = 4) {
-        std::lock_guard lock{RasterizerCache<Buffer>::mutex};
-        return StreamBufferUpload(raw_pointer, size, alignment);
-    }
-
-    void Map(std::size_t max_size) {
-        std::tie(buffer_ptr, buffer_offset_base, invalidated) = stream_buffer->Map(max_size, 4);
-        buffer_offset = buffer_offset_base;
-    }
-
-    /// Finishes the upload stream, returns true on bindings invalidation.
-    bool Unmap() {
-        stream_buffer->Unmap(buffer_offset - buffer_offset_base);
-        return std::exchange(invalidated, false);
-    }
-
-    virtual const BufferType* GetEmptyBuffer(std::size_t size) = 0;
-
-protected:
-    void FlushObjectInner(const Buffer& entry) override {
-        DownloadBufferData(entry->GetBuffer(), 0, entry->GetSize(), entry->GetWritableHostPtr());
-    }
-
-    virtual BufferStorageType CreateBuffer(std::size_t size) = 0;
-
-    virtual const BufferType* ToHandle(const BufferStorageType& storage) = 0;
-
-    virtual void UploadBufferData(const BufferStorageType& buffer, std::size_t offset,
-                                  std::size_t size, const u8* data) = 0;
-
-    virtual void DownloadBufferData(const BufferStorageType& buffer, std::size_t offset,
-                                    std::size_t size, u8* data) = 0;
-
-    virtual void CopyBufferData(const BufferStorageType& src, const BufferStorageType& dst,
-                                std::size_t src_offset, std::size_t dst_offset,
-                                std::size_t size) = 0;
-
-private:
-    BufferInfo StreamBufferUpload(const void* raw_pointer, std::size_t size,
-                                  std::size_t alignment) {
-        AlignBuffer(alignment);
-        const std::size_t uploaded_offset = buffer_offset;
-        std::memcpy(buffer_ptr, raw_pointer, size);
-
-        buffer_ptr += size;
-        buffer_offset += size;
-        return {&stream_buffer_handle, uploaded_offset};
-    }
-
-    BufferInfo FixedBufferUpload(GPUVAddr gpu_addr, u8* host_ptr, std::size_t size,
-                                 bool internalize, bool is_written) {
-        auto& memory_manager = Core::System::GetInstance().GPU().MemoryManager();
-        const auto cpu_addr = memory_manager.GpuToCpuAddress(gpu_addr);
-        ASSERT(cpu_addr);
-
-        auto entry = GetUncachedBuffer(*cpu_addr, host_ptr);
-        entry->SetSize(size);
-        entry->SetInternalState(internalize);
-        RasterizerCache<Buffer>::Register(entry);
-
-        if (internalize) {
-            internalized_entries.emplace(ToCacheAddr(host_ptr));
-        }
-        if (is_written) {
-            entry->MarkAsModified(true, *this);
-        }
-
-        if (entry->GetCapacity() < size) {
-            MarkedForDestruction().push_back(entry->ExchangeBuffer(CreateBuffer(size), size));
-        }
-
-        UploadBufferData(entry->GetBuffer(), 0, size, host_ptr);
-        return {ToHandle(entry->GetBuffer()), 0};
-    }
-
-    void IncreaseBufferSize(Buffer& entry, std::size_t new_size) {
-        const std::size_t old_size = entry->GetSize();
-        if (entry->GetCapacity() < new_size) {
-            const auto& old_buffer = entry->GetBuffer();
-            auto new_buffer = CreateBuffer(new_size);
-
-            // Copy bits from the old buffer to the new buffer.
-            CopyBufferData(old_buffer, new_buffer, 0, 0, old_size);
-            MarkedForDestruction().push_back(
-                entry->ExchangeBuffer(std::move(new_buffer), new_size));
-
-            // This buffer could have been used
-            invalidated = true;
-        }
-        // Upload the new bits.
-        const std::size_t size_diff = new_size - old_size;
-        UploadBufferData(entry->GetBuffer(), old_size, size_diff, entry->GetHostPtr() + old_size);
-
-        // Update entry's size in the object and in the cache.
-        Unregister(entry);
-
-        entry->SetSize(new_size);
-        RasterizerCache<Buffer>::Register(entry);
-    }
-
-    Buffer GetUncachedBuffer(VAddr cpu_addr, u8* host_ptr) {
-        if (auto entry = TryGetReservedBuffer(host_ptr)) {
-            return entry;
-        }
-        return std::make_shared<CachedBuffer<BufferStorageType>>(cpu_addr, host_ptr);
-    }
-
-    Buffer TryGetReservedBuffer(u8* host_ptr) {
-        const auto it = buffer_reserve.find(ToCacheAddr(host_ptr));
-        if (it == buffer_reserve.end()) {
-            return {};
-        }
-        auto& reserve = it->second;
-        auto entry = reserve.back();
-        reserve.pop_back();
-        return entry;
-    }
-
-    void ReserveBuffer(Buffer entry) {
-        buffer_reserve[entry->GetCacheAddr()].push_back(std::move(entry));
-    }
-
-    void AlignBuffer(std::size_t alignment) {
-        // Align the offset, not the mapped pointer
-        const std::size_t offset_aligned = Common::AlignUp(buffer_offset, alignment);
-        buffer_ptr += offset_aligned - buffer_offset;
-        buffer_offset = offset_aligned;
-    }
-
-    std::vector<BufferStorageType>& MarkedForDestruction() {
-        return marked_for_destruction_ring_buffer[marked_for_destruction_index];
-    }
-
-    Core::System& system;
-
-    std::unique_ptr<StreamBuffer> stream_buffer;
-    BufferType stream_buffer_handle{};
-
-    bool invalidated = false;
-
-    u8* buffer_ptr = nullptr;
-    u64 buffer_offset = 0;
-    u64 buffer_offset_base = 0;
-
-    std::size_t marked_for_destruction_index = 0;
-    std::array<std::vector<BufferStorageType>, 4> marked_for_destruction_ring_buffer;
-
-    std::unordered_set<CacheAddr> internalized_entries;
-    std::unordered_map<CacheAddr, std::vector<Buffer>> buffer_reserve;
-};
-
-} // namespace VideoCommon
@@ -0,0 +1,76 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <unordered_set>
+#include <utility>
+
+#include "common/alignment.h"
+#include "common/common_types.h"
+#include "video_core/gpu.h"
+
+namespace VideoCommon {
+
+class BufferBlock {
+public:
+    bool Overlaps(const CacheAddr start, const CacheAddr end) const {
+        return (cache_addr < end) && (cache_addr_end > start);
+    }
+
+    bool IsInside(const CacheAddr other_start, const CacheAddr other_end) const {
+        return cache_addr <= other_start && other_end <= cache_addr_end;
+    }
+
+    u8* GetWritableHostPtr() const {
+        return FromCacheAddr(cache_addr);
+    }
+
+    u8* GetWritableHostPtr(std::size_t offset) const {
+        return FromCacheAddr(cache_addr + offset);
+    }
+
+    std::size_t GetOffset(const CacheAddr in_addr) {
+        return static_cast<std::size_t>(in_addr - cache_addr);
+    }
+
+    CacheAddr GetCacheAddr() const {
+        return cache_addr;
+    }
+
+    CacheAddr GetCacheAddrEnd() const {
+        return cache_addr_end;
+    }
+
+    void SetCacheAddr(const CacheAddr new_addr) {
+        cache_addr = new_addr;
+        cache_addr_end = new_addr + size;
+    }
+
+    std::size_t GetSize() const {
+        return size;
+    }
+
+    void SetEpoch(u64 new_epoch) {
+        epoch = new_epoch;
+    }
+
+    u64 GetEpoch() {
+        return epoch;
+    }
+
+protected:
+    explicit BufferBlock(CacheAddr cache_addr, const std::size_t size) : size{size} {
+        SetCacheAddr(cache_addr);
+    }
+    ~BufferBlock() = default;
+
+private:
+    CacheAddr cache_addr{};
+    CacheAddr cache_addr_end{};
+    std::size_t size{};
+    u64 epoch{};
+};
+
+} // namespace VideoCommon
@@ -0,0 +1,447 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <memory>
+#include <mutex>
+#include <unordered_map>
+#include <unordered_set>
+#include <utility>
+#include <vector>
+
+#include "common/alignment.h"
+#include "common/common_types.h"
+#include "core/core.h"
+#include "video_core/buffer_cache/buffer_block.h"
+#include "video_core/buffer_cache/map_interval.h"
+#include "video_core/memory_manager.h"
+#include "video_core/rasterizer_interface.h"
+
+namespace VideoCommon {
+
+using MapInterval = std::shared_ptr<MapIntervalBase>;
+
+template <typename TBuffer, typename TBufferType, typename StreamBuffer>
+class BufferCache {
+public:
+    using BufferInfo = std::pair<const TBufferType*, u64>;
+
+    BufferInfo UploadMemory(GPUVAddr gpu_addr, std::size_t size, std::size_t alignment = 4,
+                            bool is_written = false) {
+        std::lock_guard lock{mutex};
+
+        auto& memory_manager = system.GPU().MemoryManager();
+        const auto host_ptr = memory_manager.GetPointer(gpu_addr);
+        if (!host_ptr) {
+            return {GetEmptyBuffer(size), 0};
+        }
+        const auto cache_addr = ToCacheAddr(host_ptr);
+
+        // Cache management is a big overhead, so only cache entries with a given size.
+        // TODO: Figure out which size is the best for given games.
+        constexpr std::size_t max_stream_size = 0x800;
+        if (size < max_stream_size) {
+            if (!is_written && !IsRegionWritten(cache_addr, cache_addr + size - 1)) {
+                return StreamBufferUpload(host_ptr, size, alignment);
+            }
+        }
+
+        auto block = GetBlock(cache_addr, size);
+        auto map = MapAddress(block, gpu_addr, cache_addr, size);
+        if (is_written) {
+            map->MarkAsModified(true, GetModifiedTicks());
+            if (!map->IsWritten()) {
+                map->MarkAsWritten(true);
+                MarkRegionAsWritten(map->GetStart(), map->GetEnd() - 1);
+            }
+        } else {
+            if (map->IsWritten()) {
+                WriteBarrier();
+            }
+        }
+
+        const u64 offset = static_cast<u64>(block->GetOffset(cache_addr));
+
+        return {ToHandle(block), offset};
+    }
+
+    /// Uploads from a host memory. Returns the OpenGL buffer where it's located and its offset.
+    BufferInfo UploadHostMemory(const void* raw_pointer, std::size_t size,
+                                std::size_t alignment = 4) {
+        std::lock_guard lock{mutex};
+        return StreamBufferUpload(raw_pointer, size, alignment);
+    }
+
+    void Map(std::size_t max_size) {
+        std::lock_guard lock{mutex};
+
+        std::tie(buffer_ptr, buffer_offset_base, invalidated) = stream_buffer->Map(max_size, 4);
+        buffer_offset = buffer_offset_base;
+    }
+
+    /// Finishes the upload stream, returns true on bindings invalidation.
+    bool Unmap() {
+        std::lock_guard lock{mutex};
+
+        stream_buffer->Unmap(buffer_offset - buffer_offset_base);
+        return std::exchange(invalidated, false);
+    }
+
+    void TickFrame() {
+        ++epoch;
+        while (!pending_destruction.empty()) {
+            if (pending_destruction.front()->GetEpoch() + 1 > epoch) {
+                break;
+            }
+            pending_destruction.pop_front();
+        }
+    }
+
+    /// Write any cached resources overlapping the specified region back to memory
+    void FlushRegion(CacheAddr addr, std::size_t size) {
+        std::lock_guard lock{mutex};
+
+        std::vector<MapInterval> objects = GetMapsInRange(addr, size);
+        std::sort(objects.begin(), objects.end(), [](const MapInterval& a, const MapInterval& b) {
+            return a->GetModificationTick() < b->GetModificationTick();
+        });
+        for (auto& object : objects) {
+            if (object->IsModified() && object->IsRegistered()) {
+                FlushMap(object);
+            }
+        }
+    }
+
+    /// Mark the specified region as being invalidated
+    void InvalidateRegion(CacheAddr addr, u64 size) {
+        std::lock_guard lock{mutex};
+
+        std::vector<MapInterval> objects = GetMapsInRange(addr, size);
+        for (auto& object : objects) {
+            if (object->IsRegistered()) {
+                Unregister(object);
+            }
+        }
+    }
+
+    virtual const TBufferType* GetEmptyBuffer(std::size_t size) = 0;
+
+protected:
+    explicit BufferCache(VideoCore::RasterizerInterface& rasterizer, Core::System& system,
+                         std::unique_ptr<StreamBuffer> stream_buffer)
+        : rasterizer{rasterizer}, system{system}, stream_buffer{std::move(stream_buffer)},
+          stream_buffer_handle{this->stream_buffer->GetHandle()} {}
+
+    ~BufferCache() = default;
+
+    virtual const TBufferType* ToHandle(const TBuffer& storage) = 0;
+
+    virtual void WriteBarrier() = 0;
+
+    virtual TBuffer CreateBlock(CacheAddr cache_addr, std::size_t size) = 0;
+
+    virtual void UploadBlockData(const TBuffer& buffer, std::size_t offset, std::size_t size,
+                                 const u8* data) = 0;
+
+    virtual void DownloadBlockData(const TBuffer& buffer, std::size_t offset, std::size_t size,
+                                   u8* data) = 0;
+
+    virtual void CopyBlock(const TBuffer& src, const TBuffer& dst, std::size_t src_offset,
+                           std::size_t dst_offset, std::size_t size) = 0;
+
+    /// Register an object into the cache
+    void Register(const MapInterval& new_map, bool inherit_written = false) {
+        const CacheAddr cache_ptr = new_map->GetStart();
+        const std::optional<VAddr> cpu_addr =
+            system.GPU().MemoryManager().GpuToCpuAddress(new_map->GetGpuAddress());
+        if (!cache_ptr || !cpu_addr) {
+            LOG_CRITICAL(HW_GPU, "Failed to register buffer with unmapped gpu_address 0x{:016x}",
+                         new_map->GetGpuAddress());
+            return;
+        }
+        const std::size_t size = new_map->GetEnd() - new_map->GetStart();
+        new_map->SetCpuAddress(*cpu_addr);
+        new_map->MarkAsRegistered(true);
+        const IntervalType interval{new_map->GetStart(), new_map->GetEnd()};
+        mapped_addresses.insert({interval, new_map});
+        rasterizer.UpdatePagesCachedCount(*cpu_addr, size, 1);
+        if (inherit_written) {
+            MarkRegionAsWritten(new_map->GetStart(), new_map->GetEnd() - 1);
+            new_map->MarkAsWritten(true);
+        }
+    }
+
+    /// Unregisters an object from the cache
+    void Unregister(MapInterval& map) {
+        const std::size_t size = map->GetEnd() - map->GetStart();
+        rasterizer.UpdatePagesCachedCount(map->GetCpuAddress(), size, -1);
+        map->MarkAsRegistered(false);
+        if (map->IsWritten()) {
+            UnmarkRegionAsWritten(map->GetStart(), map->GetEnd() - 1);
+        }
+        const IntervalType delete_interval{map->GetStart(), map->GetEnd()};
+        mapped_addresses.erase(delete_interval);
+    }
+
+private:
+    MapInterval CreateMap(const CacheAddr start, const CacheAddr end, const GPUVAddr gpu_addr) {
+        return std::make_shared<MapIntervalBase>(start, end, gpu_addr);
+    }
+
+    MapInterval MapAddress(const TBuffer& block, const GPUVAddr gpu_addr,
+                           const CacheAddr cache_addr, const std::size_t size) {
+
+        std::vector<MapInterval> overlaps = GetMapsInRange(cache_addr, size);
+        if (overlaps.empty()) {
+            const CacheAddr cache_addr_end = cache_addr + size;
+            MapInterval new_map = CreateMap(cache_addr, cache_addr_end, gpu_addr);
+            u8* host_ptr = FromCacheAddr(cache_addr);
+            UploadBlockData(block, block->GetOffset(cache_addr), size, host_ptr);
+            Register(new_map);
+            return new_map;
+        }
+
+        const CacheAddr cache_addr_end = cache_addr + size;
+        if (overlaps.size() == 1) {
+            MapInterval& current_map = overlaps[0];
+            if (current_map->IsInside(cache_addr, cache_addr_end)) {
+                return current_map;
+            }
+        }
+        CacheAddr new_start = cache_addr;
+        CacheAddr new_end = cache_addr_end;
+        bool write_inheritance = false;
+        bool modified_inheritance = false;
+        // Calculate new buffer parameters
+        for (auto& overlap : overlaps) {
+            new_start = std::min(overlap->GetStart(), new_start);
+            new_end = std::max(overlap->GetEnd(), new_end);
+            write_inheritance |= overlap->IsWritten();
+            modified_inheritance |= overlap->IsModified();
+        }
+        GPUVAddr new_gpu_addr = gpu_addr + new_start - cache_addr;
+        for (auto& overlap : overlaps) {
+            Unregister(overlap);
+        }
+        UpdateBlock(block, new_start, new_end, overlaps);
+        MapInterval new_map = CreateMap(new_start, new_end, new_gpu_addr);
+        if (modified_inheritance) {
+            new_map->MarkAsModified(true, GetModifiedTicks());
+        }
+        Register(new_map, write_inheritance);
+        return new_map;
+    }
+
+    void UpdateBlock(const TBuffer& block, CacheAddr start, CacheAddr end,
+                     std::vector<MapInterval>& overlaps) {
+        const IntervalType base_interval{start, end};
+        IntervalSet interval_set{};
+        interval_set.add(base_interval);
+        for (auto& overlap : overlaps) {
+            const IntervalType subtract{overlap->GetStart(), overlap->GetEnd()};
+            interval_set.subtract(subtract);
+        }
+        for (auto& interval : interval_set) {
+            std::size_t size = interval.upper() - interval.lower();
+            if (size > 0) {
+                u8* host_ptr = FromCacheAddr(interval.lower());
+                UploadBlockData(block, block->GetOffset(interval.lower()), size, host_ptr);
+            }
+        }
+    }
+
+    std::vector<MapInterval> GetMapsInRange(CacheAddr addr, std::size_t size) {
+        if (size == 0) {
+            return {};
+        }
+
+        std::vector<MapInterval> objects{};
+        const IntervalType interval{addr, addr + size};
+        for (auto& pair : boost::make_iterator_range(mapped_addresses.equal_range(interval))) {
+            objects.push_back(pair.second);
+        }
+
+        return objects;
+    }
+
+    /// Returns a ticks counter used for tracking when cached objects were last modified
+    u64 GetModifiedTicks() {
+        return ++modified_ticks;
+    }
+
+    void FlushMap(MapInterval map) {
+        std::size_t size = map->GetEnd() - map->GetStart();
+        TBuffer block = blocks[map->GetStart() >> block_page_bits];
+        u8* host_ptr = FromCacheAddr(map->GetStart());
+        DownloadBlockData(block, block->GetOffset(map->GetStart()), size, host_ptr);
+        map->MarkAsModified(false, 0);
+    }
+
+    BufferInfo StreamBufferUpload(const void* raw_pointer, std::size_t size,
+                                  std::size_t alignment) {
+        AlignBuffer(alignment);
+        const std::size_t uploaded_offset = buffer_offset;
+        std::memcpy(buffer_ptr, raw_pointer, size);
+
+        buffer_ptr += size;
+        buffer_offset += size;
+        return {&stream_buffer_handle, uploaded_offset};
+    }
+
+    void AlignBuffer(std::size_t alignment) {
+        // Align the offset, not the mapped pointer
+        const std::size_t offset_aligned = Common::AlignUp(buffer_offset, alignment);
+        buffer_ptr += offset_aligned - buffer_offset;
+        buffer_offset = offset_aligned;
+    }
+
+    TBuffer EnlargeBlock(TBuffer buffer) {
+        const std::size_t old_size = buffer->GetSize();
+        const std::size_t new_size = old_size + block_page_size;
+        const CacheAddr cache_addr = buffer->GetCacheAddr();
+        TBuffer new_buffer = CreateBlock(cache_addr, new_size);
+        CopyBlock(buffer, new_buffer, 0, 0, old_size);
+        buffer->SetEpoch(epoch);
+        pending_destruction.push_back(buffer);
+        const CacheAddr cache_addr_end = cache_addr + new_size - 1;
+        u64 page_start = cache_addr >> block_page_bits;
+        const u64 page_end = cache_addr_end >> block_page_bits;
+        while (page_start <= page_end) {
+            blocks[page_start] = new_buffer;
+            ++page_start;
+        }
+        return new_buffer;
+    }
+
+    TBuffer MergeBlocks(TBuffer first, TBuffer second) {
+        const std::size_t size_1 = first->GetSize();
+        const std::size_t size_2 = second->GetSize();
+        const CacheAddr first_addr = first->GetCacheAddr();
+        const CacheAddr second_addr = second->GetCacheAddr();
+        const CacheAddr new_addr = std::min(first_addr, second_addr);
+        const std::size_t new_size = size_1 + size_2;
+        TBuffer new_buffer = CreateBlock(new_addr, new_size);
+        CopyBlock(first, new_buffer, 0, new_buffer->GetOffset(first_addr), size_1);
+        CopyBlock(second, new_buffer, 0, new_buffer->GetOffset(second_addr), size_2);
+        first->SetEpoch(epoch);
+        second->SetEpoch(epoch);
+        pending_destruction.push_back(first);
+        pending_destruction.push_back(second);
+        const CacheAddr cache_addr_end = new_addr + new_size - 1;
+        u64 page_start = new_addr >> block_page_bits;
+        const u64 page_end = cache_addr_end >> block_page_bits;
+        while (page_start <= page_end) {
+            blocks[page_start] = new_buffer;
+            ++page_start;
+        }
+        return new_buffer;
+    }
+
+    TBuffer GetBlock(const CacheAddr cache_addr, const std::size_t size) {
+        TBuffer found{};
+        const CacheAddr cache_addr_end = cache_addr + size - 1;
+        u64 page_start = cache_addr >> block_page_bits;
+        const u64 page_end = cache_addr_end >> block_page_bits;
+        while (page_start <= page_end) {
+            auto it = blocks.find(page_start);
+            if (it == blocks.end()) {
+                if (found) {
+                    found = EnlargeBlock(found);
+                } else {
+                    const CacheAddr start_addr = (page_start << block_page_bits);
+                    found = CreateBlock(start_addr, block_page_size);
+                    blocks[page_start] = found;
+                }
+            } else {
+                if (found) {
+                    if (found == it->second) {
+                        ++page_start;
+                        continue;
+                    }
+                    found = MergeBlocks(found, it->second);
+                } else {
+                    found = it->second;
+                }
+            }
+            ++page_start;
+        }
+        return found;
+    }
+
+    void MarkRegionAsWritten(const CacheAddr start, const CacheAddr end) {
+        u64 page_start = start >> write_page_bit;
+        const u64 page_end = end >> write_page_bit;
+        while (page_start <= page_end) {
+            auto it = written_pages.find(page_start);
+            if (it != written_pages.end()) {
+                it->second = it->second + 1;
+            } else {
+                written_pages[page_start] = 1;
+            }
+            page_start++;
+        }
+    }
+
+    void UnmarkRegionAsWritten(const CacheAddr start, const CacheAddr end) {
+        u64 page_start = start >> write_page_bit;
+        const u64 page_end = end >> write_page_bit;
+        while (page_start <= page_end) {
+            auto it = written_pages.find(page_start);
+            if (it != written_pages.end()) {
+                if (it->second > 1) {
+                    it->second = it->second - 1;
+                } else {
+                    written_pages.erase(it);
+                }
+            }
+            page_start++;
+        }
+    }
+
+    bool IsRegionWritten(const CacheAddr start, const CacheAddr end) const {
+        u64 page_start = start >> write_page_bit;
+        const u64 page_end = end >> write_page_bit;
+        while (page_start <= page_end) {
+            if (written_pages.count(page_start) > 0) {
+                return true;
+            }
+            page_start++;
+        }
+        return false;
+    }
+
+    VideoCore::RasterizerInterface& rasterizer;
+    Core::System& system;
+    std::unique_ptr<StreamBuffer> stream_buffer;
+
+    TBufferType stream_buffer_handle{};
+
+    bool invalidated = false;
+
+    u8* buffer_ptr = nullptr;
+    u64 buffer_offset = 0;
+    u64 buffer_offset_base = 0;
+
+    using IntervalSet = boost::icl::interval_set<CacheAddr>;
+    using IntervalCache = boost::icl::interval_map<CacheAddr, MapInterval>;
+    using IntervalType = typename IntervalCache::interval_type;
+    IntervalCache mapped_addresses{};
+
+    static constexpr u64 write_page_bit{11};
+    std::unordered_map<u64, u32> written_pages{};
+
+    static constexpr u64 block_page_bits{21};
+    static constexpr u64 block_page_size{1 << block_page_bits};
+    std::unordered_map<u64, TBuffer> blocks{};
+
+    std::list<TBuffer> pending_destruction{};
+    u64 epoch{};
+    u64 modified_ticks{};
+
+    std::recursive_mutex mutex;
+};
+
+} // namespace VideoCommon
@@ -0,0 +1,89 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+#include "video_core/gpu.h"
+
+namespace VideoCommon {
+
+class MapIntervalBase {
+public:
+    MapIntervalBase(const CacheAddr start, const CacheAddr end, const GPUVAddr gpu_addr)
+        : start{start}, end{end}, gpu_addr{gpu_addr} {}
+
+    void SetCpuAddress(VAddr new_cpu_addr) {
+        cpu_addr = new_cpu_addr;
+    }
+
+    VAddr GetCpuAddress() const {
+        return cpu_addr;
+    }
+
+    GPUVAddr GetGpuAddress() const {
+        return gpu_addr;
+    }
+
+    bool IsInside(const CacheAddr other_start, const CacheAddr other_end) const {
+        return (start <= other_start && other_end <= end);
+    }
+
+    bool operator==(const MapIntervalBase& rhs) const {
+        return std::tie(start, end) == std::tie(rhs.start, rhs.end);
+    }
+
+    bool operator!=(const MapIntervalBase& rhs) const {
+        return !operator==(rhs);
+    }
+
+    void MarkAsRegistered(const bool registered) {
+        is_registered = registered;
+    }
+
+    bool IsRegistered() const {
+        return is_registered;
+    }
+
+    CacheAddr GetStart() const {
+        return start;
+    }
+
+    CacheAddr GetEnd() const {
+        return end;
+    }
+
+    void MarkAsModified(const bool is_modified_, const u64 tick) {
+        is_modified = is_modified_;
+        ticks = tick;
+    }
+
+    bool IsModified() const {
+        return is_modified;
+    }
+
+    u64 GetModificationTick() const {
+        return ticks;
+    }
+
+    void MarkAsWritten(const bool is_written_) {
+        is_written = is_written_;
+    }
+
+    bool IsWritten() const {
+        return is_written;
+    }
+
+private:
+    CacheAddr start;
+    CacheAddr end;
+    GPUVAddr gpu_addr;
+    VAddr cpu_addr{};
+    bool is_written{};
+    bool is_modified{};
+    bool is_registered{};
+    u64 ticks{};
+};
+
+} // namespace VideoCommon
@@ -10,8 +10,7 @@

 namespace Tegra::Engines {

-Fermi2D::Fermi2D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager)
-    : rasterizer{rasterizer}, memory_manager{memory_manager} {}
+Fermi2D::Fermi2D(VideoCore::RasterizerInterface& rasterizer) : rasterizer{rasterizer} {}

 void Fermi2D::CallMethod(const GPU::MethodCall& method_call) {
    ASSERT_MSG(method_call.method < Regs::NUM_REGS,
@@ -33,7 +33,7 @@ namespace Tegra::Engines {

 class Fermi2D final {
 public:
-    explicit Fermi2D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager);
+    explicit Fermi2D(VideoCore::RasterizerInterface& rasterizer);
    ~Fermi2D() = default;

    /// Write the value to the register identified by method.
@@ -145,7 +145,6 @@ public:

 private:
    VideoCore::RasterizerInterface& rasterizer;
-    MemoryManager& memory_manager;

    /// Performs the copy from the source surface to the destination surface as configured in the
    /// registers.
@@ -15,7 +15,7 @@
 namespace Tegra::Engines {

 KeplerMemory::KeplerMemory(Core::System& system, MemoryManager& memory_manager)
-    : system{system}, memory_manager{memory_manager}, upload_state{memory_manager, regs.upload} {}
+    : system{system}, upload_state{memory_manager, regs.upload} {}

 KeplerMemory::~KeplerMemory() = default;

@@ -65,7 +65,6 @@ public:

 private:
    Core::System& system;
-    MemoryManager& memory_manager;
    Upload::State upload_state;
 };

@@ -249,16 +249,10 @@ void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
    executing_macro = 0;

    // Lookup the macro offset
-    const u32 entry{(method - MacroRegistersStart) >> 1};
-    const auto& search{macro_offsets.find(entry)};
-    if (search == macro_offsets.end()) {
-        LOG_CRITICAL(HW_GPU, "macro not found for method 0x{:X}!", method);
-        UNREACHABLE();
-        return;
-    }
+    const u32 entry = ((method - MacroRegistersStart) >> 1) % macro_positions.size();

    // Execute the current macro.
-    macro_interpreter.Execute(search->second, std::move(parameters));
+    macro_interpreter.Execute(macro_positions[entry], std::move(parameters));
 }

 void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) {
@@ -421,7 +415,7 @@ void Maxwell3D::ProcessMacroUpload(u32 data) {
 }

 void Maxwell3D::ProcessMacroBind(u32 data) {
-    macro_offsets[regs.macros.entry] = data;
+    macro_positions[regs.macros.entry++] = data;
 }

 void Maxwell3D::ProcessQueryGet() {
@@ -524,7 +518,7 @@ void Maxwell3D::ProcessQueryCondition() {
 void Maxwell3D::ProcessSyncPoint() {
    const u32 sync_point = regs.sync_info.sync_point.Value();
    const u32 increment = regs.sync_info.increment.Value();
-    const u32 cache_flush = regs.sync_info.unknown.Value();
+    [[maybe_unused]] const u32 cache_flush = regs.sync_info.unknown.Value();
    if (increment) {
        system.GPU().IncrementSyncPoint(sync_point);
    }
@@ -626,10 +620,10 @@ Texture::TICEntry Maxwell3D::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()};
+    [[maybe_unused]] const auto r_type{tic_entry.r_type.Value()};
+    [[maybe_unused]] const auto g_type{tic_entry.g_type.Value()};
+    [[maybe_unused]] const auto b_type{tic_entry.b_type.Value()};
+    [[maybe_unused]] 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);
@@ -1270,7 +1270,7 @@ private:
    MemoryManager& memory_manager;

    /// Start offsets of each macro in macro_memory
-    std::unordered_map<u32, u32> macro_offsets;
+    std::array<u32, 0x80> macro_positions = {};

    /// Memory for macro code
    MacroMemory macro_memory;
@@ -5,18 +5,17 @@
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "core/core.h"
+#include "core/settings.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/engines/maxwell_dma.h"
 #include "video_core/memory_manager.h"
-#include "video_core/rasterizer_interface.h"
 #include "video_core/renderer_base.h"
 #include "video_core/textures/decoders.h"

 namespace Tegra::Engines {

-MaxwellDMA::MaxwellDMA(Core::System& system, VideoCore::RasterizerInterface& rasterizer,
-                       MemoryManager& memory_manager)
-    : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager} {}
+MaxwellDMA::MaxwellDMA(Core::System& system, MemoryManager& memory_manager)
+    : system{system}, memory_manager{memory_manager} {}

 void MaxwellDMA::CallMethod(const GPU::MethodCall& method_call) {
    ASSERT_MSG(method_call.method < Regs::NUM_REGS,
@@ -84,13 +83,17 @@ void MaxwellDMA::HandleCopy() {
    ASSERT(regs.exec.enable_2d == 1);

    if (regs.exec.is_dst_linear && !regs.exec.is_src_linear) {
-        ASSERT(regs.src_params.size_z == 1);
+        ASSERT(regs.src_params.BlockDepth() == 0);
        // If the input is tiled and the output is linear, deswizzle the input and copy it over.
-        const u32 src_bytes_per_pixel = regs.src_pitch / regs.src_params.size_x;
+        const u32 bytes_per_pixel = regs.dst_pitch / regs.x_count;
        const std::size_t src_size = Texture::CalculateSize(
-            true, src_bytes_per_pixel, regs.src_params.size_x, regs.src_params.size_y,
+            true, bytes_per_pixel, regs.src_params.size_x, regs.src_params.size_y,
            regs.src_params.size_z, regs.src_params.BlockHeight(), regs.src_params.BlockDepth());

+        const std::size_t src_layer_size = Texture::CalculateSize(
+            true, bytes_per_pixel, regs.src_params.size_x, regs.src_params.size_y, 1,
+            regs.src_params.BlockHeight(), regs.src_params.BlockDepth());
+
        const std::size_t dst_size = regs.dst_pitch * regs.y_count;

        if (read_buffer.size() < src_size) {
@@ -104,23 +107,23 @@ void MaxwellDMA::HandleCopy() {
        memory_manager.ReadBlock(source, read_buffer.data(), src_size);
        memory_manager.ReadBlock(dest, write_buffer.data(), dst_size);

-        Texture::UnswizzleSubrect(regs.x_count, regs.y_count, regs.dst_pitch,
-                                  regs.src_params.size_x, src_bytes_per_pixel, read_buffer.data(),
-                                  write_buffer.data(), regs.src_params.BlockHeight(),
-                                  regs.src_params.pos_x, regs.src_params.pos_y);
+        Texture::UnswizzleSubrect(
+            regs.x_count, regs.y_count, regs.dst_pitch, regs.src_params.size_x, bytes_per_pixel,
+            read_buffer.data() + src_layer_size * regs.src_params.pos_z, write_buffer.data(),
+            regs.src_params.BlockHeight(), regs.src_params.pos_x, regs.src_params.pos_y);

        memory_manager.WriteBlock(dest, write_buffer.data(), dst_size);
    } else {
        ASSERT(regs.dst_params.BlockDepth() == 0);

-        const u32 src_bytes_per_pixel = regs.src_pitch / regs.x_count;
+        const u32 bytes_per_pixel = regs.src_pitch / regs.x_count;

        const std::size_t dst_size = Texture::CalculateSize(
-            true, src_bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y,
+            true, bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y,
            regs.dst_params.size_z, regs.dst_params.BlockHeight(), regs.dst_params.BlockDepth());

        const std::size_t dst_layer_size = Texture::CalculateSize(
-            true, src_bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y, 1,
+            true, bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y, 1,
            regs.dst_params.BlockHeight(), regs.dst_params.BlockDepth());

        const std::size_t src_size = regs.src_pitch * regs.y_count;
@@ -133,14 +136,19 @@ void MaxwellDMA::HandleCopy() {
            write_buffer.resize(dst_size);
        }

-        memory_manager.ReadBlock(source, read_buffer.data(), src_size);
-        memory_manager.ReadBlock(dest, write_buffer.data(), dst_size);
+        if (Settings::values.use_accurate_gpu_emulation) {
+            memory_manager.ReadBlock(source, read_buffer.data(), src_size);
+            memory_manager.ReadBlock(dest, write_buffer.data(), dst_size);
+        } else {
+            memory_manager.ReadBlockUnsafe(source, read_buffer.data(), src_size);
+            memory_manager.ReadBlockUnsafe(dest, write_buffer.data(), dst_size);
+        }

        // If the input is linear and the output is tiled, swizzle the input and copy it over.
-        Texture::SwizzleSubrect(regs.x_count, regs.y_count, regs.src_pitch, regs.dst_params.size_x,
-                                src_bytes_per_pixel,
-                                write_buffer.data() + dst_layer_size * regs.dst_params.pos_z,
-                                read_buffer.data(), regs.dst_params.BlockHeight());
+        Texture::SwizzleSubrect(
+            regs.x_count, regs.y_count, regs.src_pitch, regs.dst_params.size_x, bytes_per_pixel,
+            write_buffer.data() + dst_layer_size * regs.dst_params.pos_z, read_buffer.data(),
+            regs.dst_params.BlockHeight(), regs.dst_params.pos_x, regs.dst_params.pos_y);

        memory_manager.WriteBlock(dest, write_buffer.data(), dst_size);
    }
@@ -20,10 +20,6 @@ namespace Tegra {
 class MemoryManager;
 }

-namespace VideoCore {
-class RasterizerInterface;
-}
-
 namespace Tegra::Engines {

 /**
@@ -33,8 +29,7 @@ namespace Tegra::Engines {

 class MaxwellDMA final {
 public:
-    explicit MaxwellDMA(Core::System& system, VideoCore::RasterizerInterface& rasterizer,
-                        MemoryManager& memory_manager);
+    explicit MaxwellDMA(Core::System& system, MemoryManager& memory_manager);
    ~MaxwellDMA() = default;

    /// Write the value to the register identified by method.
@@ -180,8 +175,6 @@ public:
 private:
    Core::System& system;

-    VideoCore::RasterizerInterface& rasterizer;
-
    MemoryManager& memory_manager;

    std::vector<u8> read_buffer;
@@ -538,6 +538,12 @@ enum class PhysicalAttributeDirection : u64 {
    Output = 1,
 };

+enum class VoteOperation : u64 {
+    All = 0, // allThreadsNV
+    Any = 1, // anyThreadNV
+    Eq = 2,  // allThreadsEqualNV
+};
+
 union Instruction {
    Instruction& operator=(const Instruction& instr) {
        value = instr.value;
@@ -564,6 +570,13 @@ union Instruction {
        BitField<13, 1, u64> trigger;
    } nop;

+    union {
+        BitField<48, 2, VoteOperation> operation;
+        BitField<45, 3, u64> dest_pred;
+        BitField<39, 3, u64> value;
+        BitField<42, 1, u64> negate_value;
+    } vote;
+
    union {
        BitField<8, 8, Register> gpr;
        BitField<20, 24, s64> offset;
@@ -873,6 +886,7 @@ union Instruction {
    union {
        BitField<0, 3, u64> pred0;
        BitField<3, 3, u64> pred3;
+        BitField<6, 1, u64> neg_b;
        BitField<7, 1, u64> abs_a;
        BitField<39, 3, u64> pred39;
        BitField<42, 1, u64> neg_pred;
@@ -1006,7 +1020,6 @@ union Instruction {
    } iset;

    union {
-        BitField<41, 2, u64> selector; // i2i and i2f only
        BitField<45, 1, u64> negate_a;
        BitField<49, 1, u64> abs_a;
        BitField<10, 2, Register::Size> src_size;
@@ -1032,6 +1045,13 @@ union Instruction {
            }
        } f2f;

+        union {
+            BitField<41, 2, u64> selector;
+        } int_src;
+
+        union {
+            BitField<41, 1, u64> selector;
+        } float_src;
    } conversion;

    union {
@@ -1487,6 +1507,7 @@ public:
        SYNC,
        BRK,
        DEPBAR,
+        VOTE,
        BFE_C,
        BFE_R,
        BFE_IMM,
@@ -1649,6 +1670,7 @@ public:
        Hfma2,
        Flow,
        Synch,
+        Warp,
        Memory,
        Texture,
        Image,
@@ -1775,6 +1797,7 @@ private:
            INST("111000110100---", Id::BRK, Type::Flow, "BRK"),
            INST("111000110000----", Id::EXIT, Type::Flow, "EXIT"),
            INST("1111000011110---", Id::DEPBAR, Type::Synch, "DEPBAR"),
+            INST("0101000011011---", Id::VOTE, Type::Warp, "VOTE"),
            INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"),
            INST("1110111101001---", Id::LD_S, Type::Memory, "LD_S"),
            INST("1110111101000---", Id::LD_L, Type::Memory, "LD_L"),
@@ -17,27 +17,15 @@

 namespace Tegra {

-u32 FramebufferConfig::BytesPerPixel(PixelFormat format) {
-    switch (format) {
-    case PixelFormat::ABGR8:
-    case PixelFormat::BGRA8:
-        return 4;
-    default:
-        return 4;
-    }
-
-    UNREACHABLE();
-}
-
 GPU::GPU(Core::System& system, VideoCore::RendererBase& renderer, bool is_async)
    : system{system}, renderer{renderer}, is_async{is_async} {
    auto& rasterizer{renderer.Rasterizer()};
    memory_manager = std::make_unique<Tegra::MemoryManager>(system, rasterizer);
    dma_pusher = std::make_unique<Tegra::DmaPusher>(*this);
    maxwell_3d = std::make_unique<Engines::Maxwell3D>(system, rasterizer, *memory_manager);
-    fermi_2d = std::make_unique<Engines::Fermi2D>(rasterizer, *memory_manager);
+    fermi_2d = std::make_unique<Engines::Fermi2D>(rasterizer);
    kepler_compute = std::make_unique<Engines::KeplerCompute>(system, rasterizer, *memory_manager);
-    maxwell_dma = std::make_unique<Engines::MaxwellDMA>(system, rasterizer, *memory_manager);
+    maxwell_dma = std::make_unique<Engines::MaxwellDMA>(system, *memory_manager);
    kepler_memory = std::make_unique<Engines::KeplerMemory>(system, *memory_manager);
 }

@@ -19,6 +19,10 @@ inline CacheAddr ToCacheAddr(const void* host_ptr) {
    return reinterpret_cast<CacheAddr>(host_ptr);
 }

+inline u8* FromCacheAddr(CacheAddr cache_addr) {
+    return reinterpret_cast<u8*>(cache_addr);
+}
+
 namespace Core {
 class System;
 }
@@ -91,14 +95,10 @@ class DebugContext;
 struct FramebufferConfig {
    enum class PixelFormat : u32 {
        ABGR8 = 1,
+        RGB565 = 4,
        BGRA8 = 5,
    };

-    /**
-     * Returns the number of bytes per pixel.
-     */
-    static u32 BytesPerPixel(PixelFormat format);
-
    VAddr address;
    u32 offset;
    u32 width;
@@ -249,8 +249,7 @@ public:
    virtual void PushGPUEntries(Tegra::CommandList&& entries) = 0;

    /// Swap buffers (render frame)
-    virtual void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) = 0;
+    virtual void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) = 0;

    /// Notify rasterizer that any caches of the specified region should be flushed to Switch memory
    virtual void FlushRegion(CacheAddr addr, u64 size) = 0;
@@ -281,8 +280,8 @@ private:

 protected:
    std::unique_ptr<Tegra::DmaPusher> dma_pusher;
-    VideoCore::RendererBase& renderer;
    Core::System& system;
+    VideoCore::RendererBase& renderer;

 private:
    std::unique_ptr<Tegra::MemoryManager> memory_manager;
@@ -23,9 +23,8 @@ void GPUAsynch::PushGPUEntries(Tegra::CommandList&& entries) {
    gpu_thread.SubmitList(std::move(entries));
 }

-void GPUAsynch::SwapBuffers(
-    std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
-    gpu_thread.SwapBuffers(std::move(framebuffer));
+void GPUAsynch::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
+    gpu_thread.SwapBuffers(framebuffer);
 }

 void GPUAsynch::FlushRegion(CacheAddr addr, u64 size) {
@@ -14,15 +14,14 @@ class RendererBase;
 namespace VideoCommon {

 /// Implementation of GPU interface that runs the GPU asynchronously
-class GPUAsynch : public Tegra::GPU {
+class GPUAsynch final : public Tegra::GPU {
 public:
    explicit GPUAsynch(Core::System& system, VideoCore::RendererBase& renderer);
    ~GPUAsynch() override;

    void Start() override;
    void PushGPUEntries(Tegra::CommandList&& entries) override;
-    void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) override;
+    void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override;
    void FlushRegion(CacheAddr addr, u64 size) override;
    void InvalidateRegion(CacheAddr addr, u64 size) override;
    void FlushAndInvalidateRegion(CacheAddr addr, u64 size) override;
@@ -19,9 +19,8 @@ void GPUSynch::PushGPUEntries(Tegra::CommandList&& entries) {
    dma_pusher->DispatchCalls();
 }

-void GPUSynch::SwapBuffers(
-    std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
-    renderer.SwapBuffers(std::move(framebuffer));
+void GPUSynch::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
+    renderer.SwapBuffers(framebuffer);
 }

 void GPUSynch::FlushRegion(CacheAddr addr, u64 size) {
@@ -13,15 +13,14 @@ class RendererBase;
 namespace VideoCommon {

 /// Implementation of GPU interface that runs the GPU synchronously
-class GPUSynch : public Tegra::GPU {
+class GPUSynch final : public Tegra::GPU {
 public:
    explicit GPUSynch(Core::System& system, VideoCore::RendererBase& renderer);
    ~GPUSynch() override;

    void Start() override;
    void PushGPUEntries(Tegra::CommandList&& entries) override;
-    void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) override;
+    void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override;
    void FlushRegion(CacheAddr addr, u64 size) override;
    void InvalidateRegion(CacheAddr addr, u64 size) override;
    void FlushAndInvalidateRegion(CacheAddr addr, u64 size) override;
@@ -39,7 +39,7 @@ static void RunThread(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_p
                dma_pusher.Push(std::move(submit_list->entries));
                dma_pusher.DispatchCalls();
            } else if (const auto data = std::get_if<SwapBuffersCommand>(&next.data)) {
-                renderer.SwapBuffers(std::move(data->framebuffer));
+                renderer.SwapBuffers(data->framebuffer ? &*data->framebuffer : nullptr);
            } else if (const auto data = std::get_if<FlushRegionCommand>(&next.data)) {
                renderer.Rasterizer().FlushRegion(data->addr, data->size);
            } else if (const auto data = std::get_if<InvalidateRegionCommand>(&next.data)) {
@@ -78,9 +78,9 @@ void ThreadManager::SubmitList(Tegra::CommandList&& entries) {
    system.CoreTiming().ScheduleEvent(synchronization_ticks, synchronization_event, fence);
 }

-void ThreadManager::SwapBuffers(
-    std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
-    PushCommand(SwapBuffersCommand(std::move(framebuffer)));
+void ThreadManager::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
+    PushCommand(SwapBuffersCommand(framebuffer ? *framebuffer
+                                               : std::optional<const Tegra::FramebufferConfig>{}));
 }

 void ThreadManager::FlushRegion(CacheAddr addr, u64 size) {
@@ -110,8 +110,7 @@ public:
    void SubmitList(Tegra::CommandList&& entries);

    /// Swap buffers (render frame)
-    void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer);
+    void SwapBuffers(const Tegra::FramebufferConfig* framebuffer);

    /// Notify rasterizer that any caches of the specified region should be flushed to Switch memory
    void FlushRegion(CacheAddr addr, u64 size);
@@ -25,8 +25,8 @@ static void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth

    // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
    // pixel values.
-    const u32 tile_size_x{GetDefaultBlockWidth(format)};
-    const u32 tile_size_y{GetDefaultBlockHeight(format)};
+    constexpr u32 tile_size_x{GetDefaultBlockWidth(format)};
+    constexpr u32 tile_size_y{GetDefaultBlockHeight(format)};

    if constexpr (morton_to_linear) {
        Tegra::Texture::UnswizzleTexture(buffer, addr, tile_size_x, tile_size_y, bytes_per_pixel,
@@ -186,99 +186,6 @@ static MortonCopyFn GetSwizzleFunction(MortonSwizzleMode mode, Surface::PixelFor
    return morton_to_linear_fns[static_cast<std::size_t>(format)];
 }

-static u32 MortonInterleave128(u32 x, u32 y) {
-    // 128x128 Z-Order coordinate from 2D coordinates
-    static constexpr u32 xlut[] = {
-        0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042,
-        0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809,
-        0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000,
-        0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043,
-        0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a,
-        0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001,
-        0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048,
-        0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b,
-        0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002,
-        0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049,
-        0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840,
-        0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003,
-        0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a,
-        0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841,
-        0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008,
-        0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b,
-        0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842,
-        0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009,
-        0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800,
-        0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843,
-        0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a,
-        0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801,
-        0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848,
-        0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b,
-        0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802,
-        0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849,
-        0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040,
-        0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803,
-        0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a,
-        0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041,
-        0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808,
-        0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b,
-        0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042,
-        0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809,
-        0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b,
-    };
-    static constexpr u32 ylut[] = {
-        0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090,
-        0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124,
-        0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200,
-        0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294,
-        0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330,
-        0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404,
-        0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0,
-        0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534,
-        0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610,
-        0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4,
-        0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780,
-        0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014,
-        0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0,
-        0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184,
-        0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220,
-        0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4,
-        0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390,
-        0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424,
-        0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500,
-        0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594,
-        0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630,
-        0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704,
-        0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0,
-        0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034,
-        0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110,
-        0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4,
-        0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280,
-        0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314,
-        0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0,
-        0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484,
-        0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520,
-        0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4,
-        0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690,
-        0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724,
-        0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4,
-    };
-    return xlut[x % 128] + ylut[y % 128];
-}
-
-static u32 GetMortonOffset128(u32 x, u32 y, u32 bytes_per_pixel) {
-    // Calculates the offset of the position of the pixel in Morton order
-    // Framebuffer images are split into 128x128 tiles.
-
-    constexpr u32 block_height = 128;
-    const u32 coarse_x = x & ~127;
-
-    const u32 i = MortonInterleave128(x, y);
-
-    const u32 offset = coarse_x * block_height;
-
-    return (i + offset) * bytes_per_pixel;
-}
-
 void MortonSwizzle(MortonSwizzleMode mode, Surface::PixelFormat format, u32 stride,
                   u32 block_height, u32 height, u32 block_depth, u32 depth, u32 tile_width_spacing,
                   u8* buffer, u8* addr) {
@@ -286,23 +193,4 @@ void MortonSwizzle(MortonSwizzleMode mode, Surface::PixelFormat format, u32 stri
                                     tile_width_spacing, buffer, addr);
 }

-void MortonCopyPixels128(MortonSwizzleMode mode, u32 width, u32 height, u32 bytes_per_pixel,
-                         u32 linear_bytes_per_pixel, u8* morton_data, u8* linear_data) {
-    const bool morton_to_linear = mode == MortonSwizzleMode::MortonToLinear;
-    u8* data_ptrs[2];
-    for (u32 y = 0; y < height; ++y) {
-        for (u32 x = 0; x < width; ++x) {
-            const u32 coarse_y = y & ~127;
-            const u32 morton_offset =
-                GetMortonOffset128(x, y, bytes_per_pixel) + coarse_y * width * bytes_per_pixel;
-            const u32 linear_pixel_index = (x + y * width) * linear_bytes_per_pixel;
-
-            data_ptrs[morton_to_linear ? 1 : 0] = morton_data + morton_offset;
-            data_ptrs[morton_to_linear ? 0 : 1] = &linear_data[linear_pixel_index];
-
-            std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
-        }
-    }
-}
-
 } // namespace VideoCore
@@ -15,7 +15,4 @@ void MortonSwizzle(MortonSwizzleMode mode, VideoCore::Surface::PixelFormat forma
                   u32 block_height, u32 height, u32 block_depth, u32 depth, u32 tile_width_spacing,
                   u8* buffer, u8* addr);

-void MortonCopyPixels128(MortonSwizzleMode mode, u32 width, u32 height, u32 bytes_per_pixel,
-                         u32 linear_bytes_per_pixel, u8* morton_data, u8* linear_data);
-
 } // namespace VideoCore
@@ -50,7 +50,7 @@ public:
    /// and invalidated
    virtual void FlushAndInvalidateRegion(CacheAddr addr, u64 size) = 0;

-    // Notify the rasterizer to send all written commands to the host GPU.
+    /// Notify the rasterizer to send all written commands to the host GPU.
    virtual void FlushCommands() = 0;

    /// Notify rasterizer that a frame is about to finish
@@ -36,8 +36,7 @@ public:
    virtual ~RendererBase();

    /// Swap buffers (render frame)
-    virtual void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) = 0;
+    virtual void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) = 0;

    /// Initialize the renderer
    virtual bool Init() = 0;
@@ -7,28 +7,41 @@
 #include <glad/glad.h>

 #include "common/assert.h"
+#include "common/microprofile.h"
+#include "video_core/rasterizer_interface.h"
 #include "video_core/renderer_opengl/gl_buffer_cache.h"
 #include "video_core/renderer_opengl/gl_rasterizer.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"

 namespace OpenGL {

+MICROPROFILE_DEFINE(OpenGL_Buffer_Download, "OpenGL", "Buffer Download", MP_RGB(192, 192, 128));
+
+CachedBufferBlock::CachedBufferBlock(CacheAddr cache_addr, const std::size_t size)
+    : VideoCommon::BufferBlock{cache_addr, size} {
+    gl_buffer.Create();
+    glNamedBufferData(gl_buffer.handle, static_cast<GLsizeiptr>(size), nullptr, GL_DYNAMIC_DRAW);
+}
+
+CachedBufferBlock::~CachedBufferBlock() = default;
+
 OGLBufferCache::OGLBufferCache(RasterizerOpenGL& rasterizer, Core::System& system,
                               std::size_t stream_size)
-    : VideoCommon::BufferCache<OGLBuffer, GLuint, OGLStreamBuffer>{
+    : VideoCommon::BufferCache<Buffer, GLuint, OGLStreamBuffer>{
          rasterizer, system, std::make_unique<OGLStreamBuffer>(stream_size, true)} {}

 OGLBufferCache::~OGLBufferCache() = default;

-OGLBuffer OGLBufferCache::CreateBuffer(std::size_t size) {
-    OGLBuffer buffer;
-    buffer.Create();
-    glNamedBufferData(buffer.handle, static_cast<GLsizeiptr>(size), nullptr, GL_DYNAMIC_DRAW);
-    return buffer;
+Buffer OGLBufferCache::CreateBlock(CacheAddr cache_addr, std::size_t size) {
+    return std::make_shared<CachedBufferBlock>(cache_addr, size);
 }

-const GLuint* OGLBufferCache::ToHandle(const OGLBuffer& buffer) {
-    return &buffer.handle;
+void OGLBufferCache::WriteBarrier() {
+    glMemoryBarrier(GL_ALL_BARRIER_BITS);
+}
+
+const GLuint* OGLBufferCache::ToHandle(const Buffer& buffer) {
+    return buffer->GetHandle();
 }

 const GLuint* OGLBufferCache::GetEmptyBuffer(std::size_t) {
@@ -36,23 +49,24 @@ const GLuint* OGLBufferCache::GetEmptyBuffer(std::size_t) {
    return &null_buffer;
 }

-void OGLBufferCache::UploadBufferData(const OGLBuffer& buffer, std::size_t offset, std::size_t size,
-                                      const u8* data) {
-    glNamedBufferSubData(buffer.handle, static_cast<GLintptr>(offset),
+void OGLBufferCache::UploadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size,
+                                     const u8* data) {
+    glNamedBufferSubData(*buffer->GetHandle(), static_cast<GLintptr>(offset),
                         static_cast<GLsizeiptr>(size), data);
 }

-void OGLBufferCache::DownloadBufferData(const OGLBuffer& buffer, std::size_t offset,
-                                        std::size_t size, u8* data) {
-    glGetNamedBufferSubData(buffer.handle, static_cast<GLintptr>(offset),
+void OGLBufferCache::DownloadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size,
+                                       u8* data) {
+    MICROPROFILE_SCOPE(OpenGL_Buffer_Download);
+    glGetNamedBufferSubData(*buffer->GetHandle(), static_cast<GLintptr>(offset),
                            static_cast<GLsizeiptr>(size), data);
 }

-void OGLBufferCache::CopyBufferData(const OGLBuffer& src, const OGLBuffer& dst,
-                                    std::size_t src_offset, std::size_t dst_offset,
-                                    std::size_t size) {
-    glCopyNamedBufferSubData(src.handle, dst.handle, static_cast<GLintptr>(src_offset),
-                             static_cast<GLintptr>(dst_offset), static_cast<GLsizeiptr>(size));
+void OGLBufferCache::CopyBlock(const Buffer& src, const Buffer& dst, std::size_t src_offset,
+                               std::size_t dst_offset, std::size_t size) {
+    glCopyNamedBufferSubData(*src->GetHandle(), *dst->GetHandle(),
+                             static_cast<GLintptr>(src_offset), static_cast<GLintptr>(dst_offset),
+                             static_cast<GLsizeiptr>(size));
 }

 } // namespace OpenGL
@@ -7,7 +7,7 @@
 #include <memory>

 #include "common/common_types.h"
-#include "video_core/buffer_cache.h"
+#include "video_core/buffer_cache/buffer_cache.h"
 #include "video_core/rasterizer_cache.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
 #include "video_core/renderer_opengl/gl_stream_buffer.h"
@@ -21,7 +21,24 @@ namespace OpenGL {
 class OGLStreamBuffer;
 class RasterizerOpenGL;

-class OGLBufferCache final : public VideoCommon::BufferCache<OGLBuffer, GLuint, OGLStreamBuffer> {
+class CachedBufferBlock;
+
+using Buffer = std::shared_ptr<CachedBufferBlock>;
+
+class CachedBufferBlock : public VideoCommon::BufferBlock {
+public:
+    explicit CachedBufferBlock(CacheAddr cache_addr, const std::size_t size);
+    ~CachedBufferBlock();
+
+    const GLuint* GetHandle() const {
+        return &gl_buffer.handle;
+    }
+
+private:
+    OGLBuffer gl_buffer{};
+};
+
+class OGLBufferCache final : public VideoCommon::BufferCache<Buffer, GLuint, OGLStreamBuffer> {
 public:
    explicit OGLBufferCache(RasterizerOpenGL& rasterizer, Core::System& system,
                            std::size_t stream_size);
@@ -30,18 +47,20 @@ public:
    const GLuint* GetEmptyBuffer(std::size_t) override;

 protected:
-    OGLBuffer CreateBuffer(std::size_t size) override;
+    Buffer CreateBlock(CacheAddr cache_addr, std::size_t size) override;

-    const GLuint* ToHandle(const OGLBuffer& buffer) override;
+    void WriteBarrier() override;

-    void UploadBufferData(const OGLBuffer& buffer, std::size_t offset, std::size_t size,
-                          const u8* data) override;
+    const GLuint* ToHandle(const Buffer& buffer) override;

-    void DownloadBufferData(const OGLBuffer& buffer, std::size_t offset, std::size_t size,
-                            u8* data) override;
+    void UploadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size,
+                         const u8* data) override;

-    void CopyBufferData(const OGLBuffer& src, const OGLBuffer& dst, std::size_t src_offset,
-                        std::size_t dst_offset, std::size_t size) override;
+    void DownloadBlockData(const Buffer& buffer, std::size_t offset, std::size_t size,
+                           u8* data) override;
+
+    void CopyBlock(const Buffer& src, const Buffer& dst, std::size_t src_offset,
+                   std::size_t dst_offset, std::size_t size) override;
 };

 } // namespace OpenGL
@@ -14,12 +14,22 @@
 namespace OpenGL {

 namespace {
+
 template <typename T>
 T GetInteger(GLenum pname) {
    GLint temporary;
    glGetIntegerv(pname, &temporary);
    return static_cast<T>(temporary);
 }
+
+bool TestProgram(const GLchar* glsl) {
+    const GLuint shader{glCreateShaderProgramv(GL_VERTEX_SHADER, 1, &glsl)};
+    GLint link_status;
+    glGetProgramiv(shader, GL_LINK_STATUS, &link_status);
+    glDeleteProgram(shader);
+    return link_status == GL_TRUE;
+}
+
 } // Anonymous namespace

 Device::Device() {
@@ -27,42 +37,41 @@ Device::Device() {
    shader_storage_alignment = GetInteger<std::size_t>(GL_SHADER_STORAGE_BUFFER_OFFSET_ALIGNMENT);
    max_vertex_attributes = GetInteger<u32>(GL_MAX_VERTEX_ATTRIBS);
    max_varyings = GetInteger<u32>(GL_MAX_VARYING_VECTORS);
+    has_warp_intrinsics = GLAD_GL_NV_gpu_shader5 && GLAD_GL_NV_shader_thread_group &&
+                          GLAD_GL_NV_shader_thread_shuffle;
    has_vertex_viewport_layer = GLAD_GL_ARB_shader_viewport_layer_array;
    has_variable_aoffi = TestVariableAoffi();
    has_component_indexing_bug = TestComponentIndexingBug();
+    has_precise_bug = TestPreciseBug();
+
+    LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", has_variable_aoffi);
+    LOG_INFO(Render_OpenGL, "Renderer_ComponentIndexingBug: {}", has_component_indexing_bug);
+    LOG_INFO(Render_OpenGL, "Renderer_PreciseBug: {}", has_precise_bug);
 }

 Device::Device(std::nullptr_t) {
    uniform_buffer_alignment = 0;
    max_vertex_attributes = 16;
    max_varyings = 15;
+    has_warp_intrinsics = true;
    has_vertex_viewport_layer = true;
    has_variable_aoffi = true;
    has_component_indexing_bug = false;
+    has_precise_bug = false;
 }

 bool Device::TestVariableAoffi() {
-    const GLchar* AOFFI_TEST = R"(#version 430 core
+    return TestProgram(R"(#version 430 core
 // This is a unit test, please ignore me on apitrace bug reports.
 uniform sampler2D tex;
 uniform ivec2 variable_offset;
 out vec4 output_attribute;
 void main() {
    output_attribute = textureOffset(tex, vec2(0), variable_offset);
-}
-)";
-    const GLuint shader{glCreateShaderProgramv(GL_VERTEX_SHADER, 1, &AOFFI_TEST)};
-    GLint link_status{};
-    glGetProgramiv(shader, GL_LINK_STATUS, &link_status);
-    glDeleteProgram(shader);
-
-    const bool supported{link_status == GL_TRUE};
-    LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", supported);
-    return supported;
+})");
 }

 bool Device::TestComponentIndexingBug() {
-    constexpr char log_message[] = "Renderer_ComponentIndexingBug: {}";
    const GLchar* COMPONENT_TEST = R"(#version 430 core
 layout (std430, binding = 0) buffer OutputBuffer {
    uint output_value;
@@ -102,12 +111,21 @@ void main() {
        GLuint result;
        glGetNamedBufferSubData(ssbo.handle, 0, sizeof(result), &result);
        if (result != values.at(index)) {
-            LOG_INFO(Render_OpenGL, log_message, true);
            return true;
        }
    }
-    LOG_INFO(Render_OpenGL, log_message, false);
    return false;
 }

+bool Device::TestPreciseBug() {
+    return !TestProgram(R"(#version 430 core
+in vec3 coords;
+out float out_value;
+uniform sampler2DShadow tex;
+void main() {
+    precise float tmp_value = vec4(texture(tex, coords)).x;
+    out_value = tmp_value;
+})");
+}
+
 } // namespace OpenGL
@@ -30,6 +30,10 @@ public:
        return max_varyings;
    }

+    bool HasWarpIntrinsics() const {
+        return has_warp_intrinsics;
+    }
+
    bool HasVertexViewportLayer() const {
        return has_vertex_viewport_layer;
    }
@@ -42,17 +46,24 @@ public:
        return has_component_indexing_bug;
    }

+    bool HasPreciseBug() const {
+        return has_precise_bug;
+    }
+
 private:
    static bool TestVariableAoffi();
    static bool TestComponentIndexingBug();
+    static bool TestPreciseBug();

    std::size_t uniform_buffer_alignment{};
    std::size_t shader_storage_alignment{};
    u32 max_vertex_attributes{};
    u32 max_varyings{};
+    bool has_warp_intrinsics{};
    bool has_vertex_viewport_layer{};
    bool has_variable_aoffi{};
    bool has_component_indexing_bug{};
+    bool has_precise_bug{};
 };

 } // namespace OpenGL
@@ -708,8 +708,6 @@ void RasterizerOpenGL::DrawArrays() {
        return;
    }

-    const auto& regs = gpu.regs;
-
    SyncColorMask();
    SyncFragmentColorClampState();
    SyncMultiSampleState();
@@ -980,7 +978,7 @@ void RasterizerOpenGL::SetupGlobalMemory(const GLShader::GlobalMemoryEntry& entr
                                         GPUVAddr gpu_addr, std::size_t size) {
    const auto alignment{device.GetShaderStorageBufferAlignment()};
    const auto [ssbo, buffer_offset] =
-        buffer_cache.UploadMemory(gpu_addr, size, alignment, true, entry.IsWritten());
+        buffer_cache.UploadMemory(gpu_addr, size, alignment, entry.IsWritten());
    bind_ssbo_pushbuffer.Push(ssbo, buffer_offset, static_cast<GLsizeiptr>(size));
 }

@@ -212,7 +212,9 @@ CachedProgram SpecializeShader(const std::string& code, const GLShader::ShaderEn
    const auto texture_buffer_usage{variant.texture_buffer_usage};

    std::string source = "#version 430 core\n"
-                         "#extension GL_ARB_separate_shader_objects : enable\n";
+                         "#extension GL_ARB_separate_shader_objects : enable\n"
+                         "#extension GL_NV_gpu_shader5 : enable\n"
+                         "#extension GL_NV_shader_thread_group : enable\n";
    if (entries.shader_viewport_layer_array) {
        source += "#extension GL_ARB_shader_viewport_layer_array : enable\n";
    }
@@ -247,20 +249,24 @@ CachedProgram SpecializeShader(const std::string& code, const GLShader::ShaderEn
        if (!texture_buffer_usage.test(i)) {
            continue;
        }
-        source += fmt::format("#define SAMPLER_{}_IS_BUFFER", i);
+        source += fmt::format("#define SAMPLER_{}_IS_BUFFER\n", i);
+    }
+    if (texture_buffer_usage.any()) {
+        source += '\n';
    }

    if (program_type == ProgramType::Geometry) {
        const auto [glsl_topology, debug_name, max_vertices] =
            GetPrimitiveDescription(primitive_mode);

-        source += "layout (" + std::string(glsl_topology) + ") in;\n";
+        source += "layout (" + std::string(glsl_topology) + ") in;\n\n";
        source += "#define MAX_VERTEX_INPUT " + std::to_string(max_vertices) + '\n';
    }
    if (program_type == ProgramType::Compute) {
        source += "layout (local_size_variable) in;\n";
    }

+    source += '\n';
    source += code;

    OGLShader shader;
@@ -289,7 +295,7 @@ std::set<GLenum> GetSupportedFormats() {

 CachedShader::CachedShader(const ShaderParameters& params, ProgramType program_type,
                           GLShader::ProgramResult result)
-    : RasterizerCacheObject{params.host_ptr}, host_ptr{params.host_ptr}, cpu_addr{params.cpu_addr},
+    : RasterizerCacheObject{params.host_ptr}, cpu_addr{params.cpu_addr},
      unique_identifier{params.unique_identifier}, program_type{program_type},
      disk_cache{params.disk_cache}, precompiled_programs{params.precompiled_programs},
      entries{result.second}, code{std::move(result.first)}, shader_length{entries.shader_length} {}
@@ -106,7 +106,6 @@ private:

    ShaderDiskCacheUsage GetUsage(const ProgramVariant& variant) const;

-    u8* host_ptr{};
    VAddr cpu_addr{};
    u64 unique_identifier{};
    ProgramType program_type{};
@@ -184,6 +184,9 @@ GLint GetSwizzleSource(SwizzleSource source) {
 }

 void ApplyTextureDefaults(const SurfaceParams& params, GLuint texture) {
+    if (params.IsBuffer()) {
+        return;
+    }
    glTextureParameteri(texture, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
    glTextureParameteri(texture, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
    glTextureParameteri(texture, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
@@ -208,6 +211,7 @@ OGLTexture CreateTexture(const SurfaceParams& params, GLenum target, GLenum inte
        glNamedBufferStorage(texture_buffer.handle, params.width * params.GetBytesPerPixel(),
                             nullptr, GL_DYNAMIC_STORAGE_BIT);
        glTextureBuffer(texture.handle, internal_format, texture_buffer.handle);
+        break;
    case SurfaceTarget::Texture2D:
    case SurfaceTarget::TextureCubemap:
        glTextureStorage2D(texture.handle, params.emulated_levels, internal_format, params.width,
@@ -51,7 +51,7 @@ public:
    }

 protected:
-    void DecorateSurfaceName();
+    void DecorateSurfaceName() override;

    View CreateView(const ViewParams& view_key) override;
    View CreateViewInner(const ViewParams& view_key, bool is_proxy);
@@ -101,9 +101,7 @@ RendererOpenGL::RendererOpenGL(Core::Frontend::EmuWindow& emu_window, Core::Syst

 RendererOpenGL::~RendererOpenGL() = default;

-void RendererOpenGL::SwapBuffers(
-    std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
-
+void RendererOpenGL::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
    system.GetPerfStats().EndSystemFrame();

    // Maintain the rasterizer's state as a priority
@@ -113,9 +111,9 @@ void RendererOpenGL::SwapBuffers(

    if (framebuffer) {
        // If framebuffer is provided, reload it from memory to a texture
-        if (screen_info.texture.width != (GLsizei)framebuffer->get().width ||
-            screen_info.texture.height != (GLsizei)framebuffer->get().height ||
-            screen_info.texture.pixel_format != framebuffer->get().pixel_format) {
+        if (screen_info.texture.width != static_cast<GLsizei>(framebuffer->width) ||
+            screen_info.texture.height != static_cast<GLsizei>(framebuffer->height) ||
+            screen_info.texture.pixel_format != framebuffer->pixel_format) {
            // Reallocate texture if the framebuffer size has changed.
            // This is expected to not happen very often and hence should not be a
            // performance problem.
@@ -149,43 +147,43 @@ void RendererOpenGL::SwapBuffers(
 * Loads framebuffer from emulated memory into the active OpenGL texture.
 */
 void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuffer) {
-    const u32 bytes_per_pixel{Tegra::FramebufferConfig::BytesPerPixel(framebuffer.pixel_format)};
-    const u64 size_in_bytes{framebuffer.stride * framebuffer.height * bytes_per_pixel};
-    const VAddr framebuffer_addr{framebuffer.address + framebuffer.offset};
-
    // Framebuffer orientation handling
    framebuffer_transform_flags = framebuffer.transform_flags;
    framebuffer_crop_rect = framebuffer.crop_rect;

-    // Ensure no bad interactions with GL_UNPACK_ALIGNMENT, which by default
-    // only allows rows to have a memory alignement of 4.
-    ASSERT(framebuffer.stride % 4 == 0);
-
-    if (!rasterizer->AccelerateDisplay(framebuffer, framebuffer_addr, framebuffer.stride)) {
-        // Reset the screen info's display texture to its own permanent texture
-        screen_info.display_texture = screen_info.texture.resource.handle;
-
-        rasterizer->FlushRegion(ToCacheAddr(Memory::GetPointer(framebuffer_addr)), size_in_bytes);
-
-        constexpr u32 linear_bpp = 4;
-        VideoCore::MortonCopyPixels128(VideoCore::MortonSwizzleMode::MortonToLinear,
-                                       framebuffer.width, framebuffer.height, bytes_per_pixel,
-                                       linear_bpp, Memory::GetPointer(framebuffer_addr),
-                                       gl_framebuffer_data.data());
-
-        glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(framebuffer.stride));
-
-        // Update existing texture
-        // TODO: Test what happens on hardware when you change the framebuffer dimensions so that
-        //       they differ from the LCD resolution.
-        // TODO: Applications could theoretically crash yuzu here by specifying too large
-        //       framebuffer sizes. We should make sure that this cannot happen.
-        glTextureSubImage2D(screen_info.texture.resource.handle, 0, 0, 0, framebuffer.width,
-                            framebuffer.height, screen_info.texture.gl_format,
-                            screen_info.texture.gl_type, gl_framebuffer_data.data());
-
-        glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
+    const VAddr framebuffer_addr{framebuffer.address + framebuffer.offset};
+    if (rasterizer->AccelerateDisplay(framebuffer, framebuffer_addr, framebuffer.stride)) {
+        return;
    }
+
+    // Reset the screen info's display texture to its own permanent texture
+    screen_info.display_texture = screen_info.texture.resource.handle;
+
+    const auto pixel_format{
+        VideoCore::Surface::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format)};
+    const u32 bytes_per_pixel{VideoCore::Surface::GetBytesPerPixel(pixel_format)};
+    const u64 size_in_bytes{framebuffer.stride * framebuffer.height * bytes_per_pixel};
+    const auto host_ptr{Memory::GetPointer(framebuffer_addr)};
+    rasterizer->FlushRegion(ToCacheAddr(host_ptr), size_in_bytes);
+
+    // TODO(Rodrigo): Read this from HLE
+    constexpr u32 block_height_log2 = 4;
+    VideoCore::MortonSwizzle(VideoCore::MortonSwizzleMode::MortonToLinear, pixel_format,
+                             framebuffer.stride, block_height_log2, framebuffer.height, 0, 1, 1,
+                             gl_framebuffer_data.data(), host_ptr);
+
+    glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(framebuffer.stride));
+
+    // Update existing texture
+    // TODO: Test what happens on hardware when you change the framebuffer dimensions so that
+    //       they differ from the LCD resolution.
+    // TODO: Applications could theoretically crash yuzu here by specifying too large
+    //       framebuffer sizes. We should make sure that this cannot happen.
+    glTextureSubImage2D(screen_info.texture.resource.handle, 0, 0, 0, framebuffer.width,
+                        framebuffer.height, screen_info.texture.gl_format,
+                        screen_info.texture.gl_type, gl_framebuffer_data.data());
+
+    glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
 }

 /**
@@ -276,22 +274,29 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture,
    texture.height = framebuffer.height;
    texture.pixel_format = framebuffer.pixel_format;

+    const auto pixel_format{
+        VideoCore::Surface::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format)};
+    const u32 bytes_per_pixel{VideoCore::Surface::GetBytesPerPixel(pixel_format)};
+    gl_framebuffer_data.resize(texture.width * texture.height * bytes_per_pixel);
+
    GLint internal_format;
    switch (framebuffer.pixel_format) {
    case Tegra::FramebufferConfig::PixelFormat::ABGR8:
        internal_format = GL_RGBA8;
        texture.gl_format = GL_RGBA;
        texture.gl_type = GL_UNSIGNED_INT_8_8_8_8_REV;
-        gl_framebuffer_data.resize(texture.width * texture.height * 4);
+        break;
+    case Tegra::FramebufferConfig::PixelFormat::RGB565:
+        internal_format = GL_RGB565;
+        texture.gl_format = GL_RGB;
+        texture.gl_type = GL_UNSIGNED_SHORT_5_6_5;
        break;
    default:
        internal_format = GL_RGBA8;
        texture.gl_format = GL_RGBA;
        texture.gl_type = GL_UNSIGNED_INT_8_8_8_8_REV;
-        gl_framebuffer_data.resize(texture.width * texture.height * 4);
-        LOG_CRITICAL(Render_OpenGL, "Unknown framebuffer pixel format: {}",
-                     static_cast<u32>(framebuffer.pixel_format));
-        UNREACHABLE();
+        UNIMPLEMENTED_MSG("Unknown framebuffer pixel format: {}",
+                          static_cast<u32>(framebuffer.pixel_format));
    }

    texture.resource.Release();
@@ -43,14 +43,13 @@ struct ScreenInfo {
    TextureInfo texture;
 };

-class RendererOpenGL : public VideoCore::RendererBase {
+class RendererOpenGL final : public VideoCore::RendererBase {
 public:
    explicit RendererOpenGL(Core::Frontend::EmuWindow& emu_window, Core::System& system);
    ~RendererOpenGL() override;

    /// Swap buffers (render frame)
-    void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) override;
+    void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override;

    /// Initialize the renderer
    bool Init() override;
@@ -1072,6 +1072,26 @@ private:
        return {};
    }

+    Id BallotThread(Operation) {
+        UNIMPLEMENTED();
+        return {};
+    }
+
+    Id VoteAll(Operation) {
+        UNIMPLEMENTED();
+        return {};
+    }
+
+    Id VoteAny(Operation) {
+        UNIMPLEMENTED();
+        return {};
+    }
+
+    Id VoteEqual(Operation) {
+        UNIMPLEMENTED();
+        return {};
+    }
+
    Id DeclareBuiltIn(spv::BuiltIn builtin, spv::StorageClass storage, Id type,
                      const std::string& name) {
        const Id id = OpVariable(type, storage);
@@ -1364,6 +1384,11 @@ private:
        &SPIRVDecompiler::WorkGroupId<0>,
        &SPIRVDecompiler::WorkGroupId<1>,
        &SPIRVDecompiler::WorkGroupId<2>,
+
+        &SPIRVDecompiler::BallotThread,
+        &SPIRVDecompiler::VoteAll,
+        &SPIRVDecompiler::VoteAny,
+        &SPIRVDecompiler::VoteEqual,
    };
    static_assert(operation_decompilers.size() == static_cast<std::size_t>(OperationCode::Amount));

@@ -176,6 +176,7 @@ u32 ShaderIR::DecodeInstr(NodeBlock& bb, u32 pc) {
        {OpCode::Type::Ffma, &ShaderIR::DecodeFfma},
        {OpCode::Type::Hfma2, &ShaderIR::DecodeHfma2},
        {OpCode::Type::Conversion, &ShaderIR::DecodeConversion},
+        {OpCode::Type::Warp, &ShaderIR::DecodeWarp},
        {OpCode::Type::Memory, &ShaderIR::DecodeMemory},
        {OpCode::Type::Texture, &ShaderIR::DecodeTexture},
        {OpCode::Type::Image, &ShaderIR::DecodeImage},
@@ -14,6 +14,12 @@ using Tegra::Shader::Instruction;
 using Tegra::Shader::OpCode;
 using Tegra::Shader::Register;

+namespace {
+constexpr OperationCode GetFloatSelector(u64 selector) {
+    return selector == 0 ? OperationCode::FCastHalf0 : OperationCode::FCastHalf1;
+}
+} // Anonymous namespace
+
 u32 ShaderIR::DecodeConversion(NodeBlock& bb, u32 pc) {
    const Instruction instr = {program_code[pc]};
    const auto opcode = OpCode::Decode(instr);
@@ -22,7 +28,7 @@ u32 ShaderIR::DecodeConversion(NodeBlock& bb, u32 pc) {
    case OpCode::Id::I2I_R:
    case OpCode::Id::I2I_C:
    case OpCode::Id::I2I_IMM: {
-        UNIMPLEMENTED_IF(instr.conversion.selector);
+        UNIMPLEMENTED_IF(instr.conversion.int_src.selector != 0);
        UNIMPLEMENTED_IF(instr.conversion.dst_size != Register::Size::Word);
        UNIMPLEMENTED_IF(instr.alu.saturate_d);

@@ -57,8 +63,8 @@ u32 ShaderIR::DecodeConversion(NodeBlock& bb, u32 pc) {
    case OpCode::Id::I2F_R:
    case OpCode::Id::I2F_C:
    case OpCode::Id::I2F_IMM: {
+        UNIMPLEMENTED_IF(instr.conversion.int_src.selector != 0);
        UNIMPLEMENTED_IF(instr.conversion.dst_size == Register::Size::Long);
-        UNIMPLEMENTED_IF(instr.conversion.selector);
        UNIMPLEMENTED_IF_MSG(instr.generates_cc,
                             "Condition codes generation in I2F is not implemented");

@@ -113,8 +119,10 @@ u32 ShaderIR::DecodeConversion(NodeBlock& bb, u32 pc) {
        }();

        if (instr.conversion.src_size == Register::Size::Short) {
-            // TODO: figure where extract is sey in the encoding
-            value = Operation(OperationCode::FCastHalf0, PRECISE, value);
+            value = Operation(GetFloatSelector(instr.conversion.float_src.selector), NO_PRECISE,
+                              std::move(value));
+        } else {
+            ASSERT(instr.conversion.float_src.selector == 0);
        }

        value = GetOperandAbsNegFloat(value, instr.conversion.abs_a, instr.conversion.negate_a);
@@ -169,8 +177,10 @@ u32 ShaderIR::DecodeConversion(NodeBlock& bb, u32 pc) {
        }();

        if (instr.conversion.src_size == Register::Size::Short) {
-            // TODO: figure where extract is sey in the encoding
-            value = Operation(OperationCode::FCastHalf0, PRECISE, value);
+            value = Operation(GetFloatSelector(instr.conversion.float_src.selector), NO_PRECISE,
+                              std::move(value));
+        } else {
+            ASSERT(instr.conversion.float_src.selector == 0);
        }

        value = GetOperandAbsNegFloat(value, instr.conversion.abs_a, instr.conversion.negate_a);
@@ -15,7 +15,6 @@ using Tegra::Shader::OpCode;

 u32 ShaderIR::DecodeFloatSet(NodeBlock& bb, u32 pc) {
    const Instruction instr = {program_code[pc]};
-    const auto opcode = OpCode::Decode(instr);

    const Node op_a = GetOperandAbsNegFloat(GetRegister(instr.gpr8), instr.fset.abs_a != 0,
                                            instr.fset.neg_a != 0);
@@ -16,10 +16,9 @@ using Tegra::Shader::Pred;

 u32 ShaderIR::DecodeFloatSetPredicate(NodeBlock& bb, u32 pc) {
    const Instruction instr = {program_code[pc]};
-    const auto opcode = OpCode::Decode(instr);

-    const Node op_a = GetOperandAbsNegFloat(GetRegister(instr.gpr8), instr.fsetp.abs_a != 0,
-                                            instr.fsetp.neg_a != 0);
+    Node op_a = GetOperandAbsNegFloat(GetRegister(instr.gpr8), instr.fsetp.abs_a != 0,
+                                      instr.fsetp.neg_a != 0);
    Node op_b = [&]() {
        if (instr.is_b_imm) {
            return GetImmediate19(instr);
@@ -29,12 +28,13 @@ u32 ShaderIR::DecodeFloatSetPredicate(NodeBlock& bb, u32 pc) {
            return GetConstBuffer(instr.cbuf34.index, instr.cbuf34.GetOffset());
        }
    }();
-    op_b = GetOperandAbsNegFloat(op_b, instr.fsetp.abs_b, false);
+    op_b = GetOperandAbsNegFloat(std::move(op_b), instr.fsetp.abs_b, instr.fsetp.neg_b);

    // We can't use the constant predicate as destination.
    ASSERT(instr.fsetp.pred3 != static_cast<u64>(Pred::UnusedIndex));

-    const Node predicate = GetPredicateComparisonFloat(instr.fsetp.cond, op_a, op_b);
+    const Node predicate =
+        GetPredicateComparisonFloat(instr.fsetp.cond, std::move(op_a), std::move(op_b));
    const Node second_pred = GetPredicate(instr.fsetp.pred39, instr.fsetp.neg_pred != 0);

    const OperationCode combiner = GetPredicateCombiner(instr.fsetp.op);
@@ -14,7 +14,6 @@ using Tegra::Shader::OpCode;

 u32 ShaderIR::DecodeIntegerSet(NodeBlock& bb, u32 pc) {
    const Instruction instr = {program_code[pc]};
-    const auto opcode = OpCode::Decode(instr);

    const Node op_a = GetRegister(instr.gpr8);
    const Node op_b = [&]() {
@@ -16,7 +16,6 @@ using Tegra::Shader::Pred;

 u32 ShaderIR::DecodeIntegerSetPredicate(NodeBlock& bb, u32 pc) {
    const Instruction instr = {program_code[pc]};
-    const auto opcode = OpCode::Decode(instr);

    const Node op_a = GetRegister(instr.gpr8);

@@ -74,6 +74,13 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
            case SystemVariable::InvocationInfo:
                LOG_WARNING(HW_GPU, "MOV_SYS instruction with InvocationInfo is incomplete");
                return Immediate(0u);
+            case SystemVariable::Tid: {
+                Node value = Immediate(0);
+                value = BitfieldInsert(value, Operation(OperationCode::LocalInvocationIdX), 0, 9);
+                value = BitfieldInsert(value, Operation(OperationCode::LocalInvocationIdY), 16, 9);
+                value = BitfieldInsert(value, Operation(OperationCode::LocalInvocationIdZ), 26, 5);
+                return value;
+            }
            case SystemVariable::TidX:
                return Operation(OperationCode::LocalInvocationIdX);
            case SystemVariable::TidY:
@@ -15,7 +15,6 @@ using Tegra::Shader::OpCode;

 u32 ShaderIR::DecodePredicateSetRegister(NodeBlock& bb, u32 pc) {
    const Instruction instr = {program_code[pc]};
-    const auto opcode = OpCode::Decode(instr);

    UNIMPLEMENTED_IF_MSG(instr.generates_cc,
                         "Condition codes generation in PSET is not implemented");
@@ -0,0 +1,55 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/assert.h"
+#include "common/common_types.h"
+#include "video_core/engines/shader_bytecode.h"
+#include "video_core/shader/node_helper.h"
+#include "video_core/shader/shader_ir.h"
+
+namespace VideoCommon::Shader {
+
+using Tegra::Shader::Instruction;
+using Tegra::Shader::OpCode;
+using Tegra::Shader::Pred;
+using Tegra::Shader::VoteOperation;
+
+namespace {
+OperationCode GetOperationCode(VoteOperation vote_op) {
+    switch (vote_op) {
+    case VoteOperation::All:
+        return OperationCode::VoteAll;
+    case VoteOperation::Any:
+        return OperationCode::VoteAny;
+    case VoteOperation::Eq:
+        return OperationCode::VoteEqual;
+    default:
+        UNREACHABLE_MSG("Invalid vote operation={}", static_cast<u64>(vote_op));
+        return OperationCode::VoteAll;
+    }
+}
+} // Anonymous namespace
+
+u32 ShaderIR::DecodeWarp(NodeBlock& bb, u32 pc) {
+    const Instruction instr = {program_code[pc]};
+    const auto opcode = OpCode::Decode(instr);
+
+    switch (opcode->get().GetId()) {
+    case OpCode::Id::VOTE: {
+        const Node value = GetPredicate(instr.vote.value, instr.vote.negate_value != 0);
+        const Node active = Operation(OperationCode::BallotThread, value);
+        const Node vote = Operation(GetOperationCode(instr.vote.operation), value);
+        SetRegister(bb, instr.gpr0, active);
+        SetPredicate(bb, instr.vote.dest_pred, vote);
+        break;
+    }
+    default:
+        UNIMPLEMENTED_MSG("Unhandled warp instruction: {}", opcode->get().GetName());
+        break;
+    }
+
+    return pc;
+}
+
+} // namespace VideoCommon::Shader
@@ -168,6 +168,11 @@ enum class OperationCode {
    WorkGroupIdY,       /// () -> uint
    WorkGroupIdZ,       /// () -> uint

+    BallotThread, /// (bool) -> uint
+    VoteAll,      /// (bool) -> bool
+    VoteAny,      /// (bool) -> bool
+    VoteEqual,    /// (bool) -> bool
+
    Amount,
 };

@@ -405,4 +405,9 @@ Node ShaderIR::BitfieldExtract(Node value, u32 offset, u32 bits) {
                     Immediate(offset), Immediate(bits));
 }

+Node ShaderIR::BitfieldInsert(Node base, Node insert, u32 offset, u32 bits) {
+    return Operation(OperationCode::UBitfieldInsert, NO_PRECISE, base, insert, Immediate(offset),
+                     Immediate(bits));
+}
+
 } // namespace VideoCommon::Shader
@@ -167,6 +167,7 @@ private:
    u32 DecodeFfma(NodeBlock& bb, u32 pc);
    u32 DecodeHfma2(NodeBlock& bb, u32 pc);
    u32 DecodeConversion(NodeBlock& bb, u32 pc);
+    u32 DecodeWarp(NodeBlock& bb, u32 pc);
    u32 DecodeMemory(NodeBlock& bb, u32 pc);
    u32 DecodeTexture(NodeBlock& bb, u32 pc);
    u32 DecodeImage(NodeBlock& bb, u32 pc);
@@ -279,6 +280,9 @@ private:
    /// Extracts a sequence of bits from a node
    Node BitfieldExtract(Node value, u32 offset, u32 bits);

+    /// Inserts a sequence of bits from a node
+    Node BitfieldInsert(Node base, Node insert, u32 offset, u32 bits);
+
    void WriteTexInstructionFloat(NodeBlock& bb, Tegra::Shader::Instruction instr,
                                  const Node4& components);

@@ -445,11 +445,12 @@ PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat
    switch (format) {
    case Tegra::FramebufferConfig::PixelFormat::ABGR8:
        return PixelFormat::ABGR8U;
+    case Tegra::FramebufferConfig::PixelFormat::RGB565:
+        return PixelFormat::B5G6R5U;
    case Tegra::FramebufferConfig::PixelFormat::BGRA8:
        return PixelFormat::BGRA8;
    default:
-        LOG_CRITICAL(HW_GPU, "Unimplemented format={}", static_cast<u32>(format));
-        UNREACHABLE();
+        UNIMPLEMENTED_MSG("Unimplemented format={}", static_cast<u32>(format));
        return PixelFormat::ABGR8U;
    }
 }
Author	SHA1	Message	Date
David	a956d0b0eb	Merge pull request #2830 from FearlessTobi/port-4911 Port citra-emu/citra#4911: "Add cancel option to analog stick configuration"	2019-09-05 09:53:44 +10:00
David	d8e59a28ea	Merge pull request #2834 from Morph1984/audrenu_QueryAudioDeviceInputEvent audren_u: Stub IAudioDevice::QueryAudioDeviceInputEvent	2019-09-05 09:51:50 +10:00
David	67bdd8ed58	Merge pull request #2836 from Morph1984/hid_vibration hid: Implement PermitVibration and IsVibrationPermitted	2019-09-05 09:50:38 +10:00
David	0a4f0b6a5d	Merge pull request #2818 from MysticExile/fmt externals: update fmt to 6.0.0	2019-09-05 09:50:17 +10:00
bunnei	e77d2b2103	Merge pull request #2801 from ReinUsesLisp/typed-decompiler gl_shader_decompiler: Rework GLSL decompiler type system	2019-09-04 17:42:13 -04:00
ReinUsesLisp	6177cbdbe1	gl_shader_decompiler: Fixup slow path	2019-09-04 15:03:51 -03:00
mailwl	a1d48b5f52	AM: Stub IApplicationFunctions::GetGpuErrorDetectedSystemEvent (#2827 ) * AM: Implement IApplicationFunctions::GetGpuErrorDetectedSystemEvent * Remove unneeded event clear * Fix event name	2019-09-04 11:43:04 -04:00
bunnei	fe83ee102b	Merge pull request #2829 from Morph1984/audio service/audio/audren_u: Stub IAudioDevice::GetAudioDeviceOutputVolume	2019-09-04 11:42:23 -04:00
bunnei	a139fdf4ac	Merge pull request #2444 from FearlessTobi/port-3617-new Port citra-emu/citra#3617: "QT: Add support for multiple game directories"	2019-09-04 11:40:35 -04:00
Ethan	30448641f2	Fix clang-format	2019-09-04 16:53:29 +02:00
fearlessTobi	c49c3e9f27	Fix uisettings include	2019-09-04 16:47:33 +02:00
fearlessTobi	053da44ecd	Limit the size of directory icons, fix text when icon size is none	2019-09-04 16:47:33 +02:00
fearlessTobi	13891fd62d	Change QList to QVector	2019-09-04 16:47:33 +02:00
fearlessTobi	5aaafa6a56	Separate UserNand and Sdmc directories	2019-09-04 16:47:32 +02:00
fearlessTobi	dfec9c9a43	Address more trivial review comments	2019-09-04 16:47:32 +02:00
fearlessTobi	7a8f484020	Address trivial review comments	2019-09-04 16:47:32 +02:00
fearlessTobi	2d8eba5baf	yuzu: Add support for multiple game directories Ported from https://github.com/citra-emu/citra/pull/3617.	2019-09-04 16:47:32 +02:00
fearlessTobi	7fc5af3622	Add assets and licenses	2019-09-04 16:46:39 +02:00
Morph1984andDavid	cd81194fc0	ditto Co-Authored-By: David <[email protected]>	2019-09-04 10:16:22 -04:00
Morph1984	ef98828d40	IsVibrationEnabled() as a const member func	2019-09-04 10:15:49 -04:00
Morph1984	d6969fa7d4	clang-format	2019-09-04 10:06:38 -04:00
Morph1984	ee35f7adf7	Update npad.h	2019-09-04 02:43:17 -04:00
Morph1984	5130b8a6a9	Update npad.cpp	2019-09-04 02:42:58 -04:00
Morph1984	1559477740	Update hid.h	2019-09-04 02:42:42 -04:00
Morph1984	a83eb90a78	Update hid.cpp	2019-09-04 02:42:22 -04:00
ReinUsesLisp	9cf52d027d	gl_device: Disable precise in fragment shaders on bugged drivers	2019-09-04 01:54:00 -03:00
ReinUsesLisp	03276e7490	gl_shader_decompiler: Fixup AMD's slow path type	2019-09-04 01:54:00 -03:00
ReinUsesLisp	6c449793b8	gl_shader_decompiler: Rework GLSL decompiler type system GLSL decompiler type system was broken. We converted all return values to float except for some cases where returning we couldn't and implicitly broke the rule of returning floats (e.g. for bools or bool pairs). Instead of doing this introduce class Expression that knows what type a return value has and when a consumer wants to use the string it asks for it with a required type, emitting a runtime error if types are incompatible. This has the disadvantage that there's more C++ code, but we can emit better GLSL code that's easier to read.	2019-09-04 01:54:00 -03:00
David	922c7f4e51	Merge pull request #2835 from chris062689/master CI Fix (Azure/Testing) - apt update before upgrade. Use apt-get.	2019-09-04 14:37:29 +10:00
Flame Sage	ce3edaad26	Changed apt-get upgrade for specific package.	2019-09-04 00:15:51 -04:00
Morph1984	84815fa879	Add Kernel::EventPair audio_input_device_switch_event;	2019-09-03 23:50:56 -04:00
Flame Sage	573a1e7662	apt update before upgrade. Use apt-get. Identified a bug in the script which uses the azure image when attempting to upgrade python3-pip. Package index was out of date because apt-get update was not ran before attempting the upgrade.	2019-09-04 03:49:55 +00:00
Morph1984	ec95c73a12	remove <f32> We can remove this since its already a f32 value	2019-09-03 23:20:19 -04:00
Morph1984	1449ed9dbf	audren_u: Stub IAudioDevice::QueryAudioDeviceInputEvent	2019-09-03 23:13:32 -04:00
David	e1981b8b8d	Merge pull request #2708 from DarkLordZach/mii-db-source-crash mii: Handle logging of unknown database source	2019-09-04 13:07:10 +10:00
Morph1984	58783b8a46	explicitly represent 1 as a float (1.0f instead of 1)	2019-09-03 23:06:32 -04:00
bunnei	19af91434e	Merge pull request #2793 from ReinUsesLisp/bgr565 renderer_opengl: Implement RGB565 framebuffer format	2019-09-03 22:36:32 -04:00
bunnei	81fbc5370d	Merge pull request #2812 from ReinUsesLisp/f2i-selector shader_ir/conversion: Implement F2I and F2F F16 selector	2019-09-03 22:35:33 -04:00
bunnei	d4f33b822b	Merge pull request #2811 from ReinUsesLisp/fsetp-fix float_set_predicate: Add missing negation bit for the second operand	2019-09-03 22:34:34 -04:00
bunnei	137d165672	Merge pull request #2826 from ReinUsesLisp/macro-binding maxwell_3d: Fix macro binding cursor	2019-09-03 22:32:42 -04:00
bunnei	86b39e0677	Merge pull request #2831 from FearlessTobi/port-4914 Port citra-emu/citra#4914: "Fix to Windows sleep issues"	2019-09-03 22:32:09 -04:00
bunnei	7397289266	Merge pull request #2832 from FearlessTobi/im-an-idiot configuration/config: Add missing screenshot path read	2019-09-03 22:31:32 -04:00
Morph1984	b1ca56bed2	Change u32 -> f32 Volume is a f32 value. (SwIPC describes it as a u32, but it is actually f32 as corroborated by switchbrew docs and SetAudioDeviceOutputVolume) ```cpp const f32 volume = rp.Pop<f32>(); ```	2019-09-03 22:30:20 -04:00
fearlessTobi	952f010c2c	configuration/config: Add missing screenshot path read I missed this in my original PR (https://github.com/yuzu-emu/yuzu/pull/1886).	2019-09-04 03:08:15 +02:00
fearlessTobiandVitor K	4ea572791b	Fix to Windows sleep issues Co-Authored-By: Vitor K <[email protected]>	2019-09-03 23:00:34 +02:00
fearlessTobiandVitor K	22fd208e8d	Add cancel option to analog stick configuration Co-Authored-By: Vitor K <[email protected]>	2019-09-03 22:53:59 +02:00
Morph1984	ba661c8d9a	service/audio/audren_u: Stub IAudioDevice::GetAudioDeviceOutputVolume	2019-09-03 16:05:33 -04:00
bunnei	50b5bb44a0	Merge pull request #2765 from FernandoS27/dma-fix MaxwellDMA: Fixes, corrections and relaxations.	2019-09-01 13:13:05 -04:00
ReinUsesLisp	52a41f482f	maxwell_3d: Fix macro binding cursor	2019-09-01 05:01:11 -03:00
Rodrigo Locatti	4d4f9cc104	video_core: Silent miscellaneous warnings (#2820 ) * texture_cache/surface_params: Remove unused local variable * rasterizer_interface: Add missing documentation commentary * maxwell_dma: Remove unused rasterizer reference * video_core/gpu: Sort member declaration order to silent -Wreorder warning * fermi_2d: Remove unused MemoryManager reference * video_core: Silent unused variable warnings * buffer_cache: Silent -Wreorder warnings * kepler_memory: Remove unused MemoryManager reference * gl_texture_cache: Add missing override * buffer_cache: Add missing include * shader/decode: Remove unused variables	2019-08-30 14:08:00 -04:00
Fernando Sahmkow	67cc2d5046	Merge pull request #2819 from ReinUsesLisp/fixup-clang gl_buffer_cache: Add missing include	2019-08-29 18:13:35 -04:00
ReinUsesLisp	878adee0a3	gl_buffer_cache: Add missing include RasterizerInterface was considered an incomplete object by clang.	2019-08-29 22:02:52 +00:00
bunnei	a67c4e6e02	Merge pull request #2742 from ReinUsesLisp/fix-texture-buffers gl_texture_cache: Miscellaneous texture buffer fixes	2019-08-29 15:59:17 -04:00
Ethan	5adbe66ae8	accommodate for fmt update	2019-08-29 21:34:51 +02:00
Ethan	46a962eb5d	externals: update fmt to 6.0.0	2019-08-29 21:32:58 +02:00
James Rowe	f0c75573b1	Revert "externals: Update FMT to 6.0.0" This reverts commit `ca4ca8a6dc`.	2019-08-29 12:23:34 -06:00
Ethan	ca4ca8a6dc	externals: Update FMT to 6.0.0	2019-08-29 19:37:46 +02:00
bunnei	e424615839	Merge pull request #2783 from FernandoS27/new-buffer-cache Implement a New LLE Buffer Cache	2019-08-29 13:07:01 -04:00
bunnei	f8cc5668f8	Merge pull request #2758 from ReinUsesLisp/packed-tid shader/decode: Implement S2R Tic	2019-08-29 12:58:43 -04:00
bunnei	680ab61327	Merge pull request #2786 from ReinUsesLisp/vote shader_ir: Implement VOTE on Nvidia drivers	2019-08-29 12:58:10 -04:00
ReinUsesLisp	e3534700d7	shader_ir/conversion: Split int and float selector and implement F2F H1	2019-08-28 16:09:33 -03:00
ReinUsesLisp	b13fbc25b8	shader_ir/conversion: Implement F2I F16 Ra.H1	2019-08-27 23:40:40 -03:00
ReinUsesLisp	6207751b00	float_set_predicate: Add missing negation bit for the second operand	2019-08-27 21:57:43 -03:00
ReinUsesLisp	4e35177e23	shader_ir: Implement VOTE Implement VOTE using Nvidia's intrinsics. Documentation about these can be found here https://developer.nvidia.com/reading-between-threads-shader-intrinsics Instead of using portable ARB instructions I opted to use Nvidia intrinsics because these are the closest we have to how Tegra X1 hardware renders. To stub VOTE on non-Nvidia drivers (including nouveau) this commit simulates a GPU with a warp size of one, returning what is meaningful for the instruction being emulated: * anyThreadNV(value) -> value * allThreadsNV(value) -> value * allThreadsEqualNV(value) -> true ballotARB, also known as "uint64_t(activeThreadsNV())", emits VOTE.ANY Rd, PT, PT; on nouveau's compiler. This doesn't match exactly to Nvidia's code VOTE.ALL Rd, PT, PT; Which is emulated with activeThreadsNV() by this commit. In theory this shouldn't really matter since .ANY, .ALL and .EQ affect the predicates (set to PT on those cases) and not the registers.	2019-08-21 14:50:38 -03:00
Fernando Sahmkow	83ec2091c1	Buffer Cache: Adress Feedback.	2019-08-21 12:14:27 -04:00
Fernando Sahmkow	6ce2c85047	Buffer_Cache: Implement flushing.	2019-08-21 12:14:26 -04:00
Fernando Sahmkow	de8ff8a1c6	Buffer_Cache: Implement barriers.	2019-08-21 12:14:25 -04:00
Fernando Sahmkow	286f4c446a	Buffer_Cache: Optimize and track written areas.	2019-08-21 12:14:25 -04:00
Fernando Sahmkow	5f4b746a1e	BufferCache: Rework mapping caching.	2019-08-21 12:14:24 -04:00
Fernando Sahmkow	86d8563314	Buffer_Cache: Fixes and optimizations.	2019-08-21 12:14:23 -04:00
Fernando Sahmkow	862bec001b	Video_Core: Implement a new Buffer Cache	2019-08-21 12:14:22 -04:00
ReinUsesLisp	80702aa88f	renderer_opengl: Implement RGB565 framebuffer format	2019-08-21 02:28:31 -03:00
ReinUsesLisp	9cdf5c6c31	renderer_opengl: Use block linear swizzling for CPU framebuffers	2019-08-21 02:17:14 -03:00
ReinUsesLisp	8ad7268c75	renderer_opengl: Use VideoCore pixel format	2019-08-21 02:16:40 -03:00
ReinUsesLisp	9a76e94b3d	gpu: Change optional<reference_wrapper<T>> to T* for FramebufferConfig	2019-08-21 01:55:25 -03:00
Fernando Sahmkow	a452ff983d	MaxwellDMA: Fixes, corrections and relaxations. This commit fixes offsets on Linear -> Tiled copies, corrects z pos fortiled->linear copies, corrects bytes_per_pixel calculation in tiled -> linear copies and relaxes some limitations set by latest dma fixes refactors.	2019-07-25 20:41:42 -04:00
ReinUsesLisp	104641db07	shader/decode: Implement S2R Tic	2019-07-22 16:16:10 -03:00
ReinUsesLisp	74632c76ce	gl_shader_decompiler: Rename bufferImage to imageBuffer The online OpenGL documentation is wrong. The type definition is imageBuffer.	2019-07-18 01:16:44 -03:00
ReinUsesLisp	87909d327f	gl_shader_cache: Fix newline on buffer preprocessor definitions	2019-07-18 01:16:15 -03:00
ReinUsesLisp	e7bdf8b22a	textures: Fix texture buffer size calculation	2019-07-18 01:07:08 -03:00
ReinUsesLisp	84027f4808	gl_texture_cache: Do not set texture parameters to buffers	2019-07-18 01:06:26 -03:00
ReinUsesLisp	73b2dc6d4f	gl_texture_cache: Add missing break in CreateTexture	2019-07-18 01:04:18 -03:00
Zach Hilman	37a352e9d3	mii: Handle logging of unknown database source	2019-07-10 07:07:24 -04:00