"Merge PR 1012"

"Merge PR 1340"
"Merge PR 1703"
2019-09-25 12:01:32 +00:00 · 2019-09-25 12:01:31 +00:00 · 2019-09-25 12:01:30 +00:00 · 2019-09-25 12:01:29 +00:00 · 2019-09-25 12:01:27 +00:00 · 2019-09-25 12:01:26 +00:00
351 changed files with 11691 additions and 4285 deletions
@@ -0,0 +1,5 @@
+steps:
+  - script: mkdir artifacts || echo 'X' > artifacts/T1.txt
+  - publish: artifacts
+    artifact: 'yuzu-$(BuildName)-$(BuildSuffix)'
+    displayName: 'Upload Artifacts'
@@ -3,17 +3,18 @@ parameters:
  cache: 'false'

 steps:
+- script: export DATE=`date '+%Y.%m.%d'` && export CI=true && AZURE_REPO_NAME=yuzu-emu/yuzu-$(BuildName) && AZURE_REPO_TAG=$(BuildName)-$DATE
+  displayName: 'Determine Build Name'
 - task: DockerInstaller@0
  displayName: 'Prepare Environment'
  inputs:
    dockerVersion: '17.09.0-ce'
- ${{ if eq(parameters.cache, 'true') }}:
-  - task: CacheBeta@0
-    displayName: 'Cache Build System'
-    inputs:
-      key: yuzu-v1-$(BuildName)-$(BuildSuffix)-$(CacheSuffix)
-      path: $(System.DefaultWorkingDirectory)/ccache
-      cacheHitVar: CACHE_RESTORED
+- task: CacheBeta@0
+  displayName: 'Cache Build System'
+  inputs:
+    key: yuzu-v1-$(BuildName)-$(BuildSuffix)-$(CacheSuffix)
+    path: $(System.DefaultWorkingDirectory)/ccache
+    cacheHitVar: CACHE_RESTORED
 - script: chmod a+x ./.ci/scripts/$(ScriptFolder)/exec.sh && ./.ci/scripts/$(ScriptFolder)/exec.sh
  displayName: 'Build'
 - script: chmod a+x ./.ci/scripts/$(ScriptFolder)/upload.sh && RELEASE_NAME=$(BuildName) ./.ci/scripts/$(ScriptFolder)/upload.sh
@@ -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')
@@ -0,0 +1,13 @@
+steps:
+  - task: DownloadPipelineArtifact@2
+    displayName: 'Download Windows Release'
+    inputs:
+      artifactName: 'yuzu-$(BuildName)-windows-mingw'
+      buildType: 'current'
+      targetPath: '$(Build.ArtifactStagingDirectory)'
+  - task: DownloadPipelineArtifact@2
+    displayName: 'Download Linux Release'
+    inputs:
+      artifactName: 'yuzu-$(BuildName)-linux'
+      buildType: 'current'
+      targetPath: '$(Build.ArtifactStagingDirectory)'
@@ -0,0 +1,11 @@
+steps:
+  - template: ./release-download.yml
+  - task: GitHubRelease@0
+    inputs:
+      action: 'create'
+      title: 'yuzu $(BuildName) #$(Build.BuildId)'
+      assets: '$(Build.ArtifactStagingDirectory)/*'
+      gitHubConnection: $(GitHubReleaseConnectionName)
+      repositoryName: '$(Build.Repository.Name)'
+      target: '$(Build.SourceVersion)'
+      tagSource: 'auto'
@@ -0,0 +1,10 @@
+steps:
+  - template: ./release-download.yml
+  - task: UniversalPackages@0
+    displayName: Publish Artifacts
+    inputs:
+      command: publish
+      publishDirectory: '$(Build.ArtifactStagingDirectory)'
+      vstsFeedPublish: 'yuzu-$(BuildName)'
+      vstsFeedPackagePublish: 'main'
+      packagePublishDescription: 'Yuzu Windows and Linux Executable Packages'
@@ -0,0 +1,8 @@
+trigger:
+- master
+
+stages:
+- stage: merge
+  displayName: 'merge'
+  jobs:
+  - template: ./templates/merge.yml
@@ -2,12 +2,7 @@ trigger:
 - master

 stages:
- stage: merge
-  displayName: 'merge'
-  jobs:
-  - template: ./templates/merge.yml
 - stage: format
-  dependsOn: merge
  displayName: 'format'
  jobs:
  - job: format
@@ -17,9 +12,17 @@ stages:
    steps:
    - template: ./templates/format-check.yml
 - stage: build
-  displayName: 'build'
  dependsOn: format
+  displayName: 'build'
  jobs:
  - template: ./templates/build-standard.yml
    parameters:
      cache: 'true'
+- stage: release
+  displayName: 'Release'
+  dependsOn: build
+  jobs:
+  - job: github
+    displayName: 'GitHub Release'
+    steps:
+    - template: ./templates/release-github.yml
@@ -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"
@@ -6,9 +6,11 @@ yuzu emulator

 yuzu is an experimental open-source emulator for the Nintendo Switch from the creators of [Citra](https://citra-emu.org/).

-It is written in C++ with portability in mind, with builds actively maintained for Windows, Linux and macOS. The emulator is currently only useful for homebrew development and research purposes.
+It is written in C++ with portability in mind, with builds actively maintained for Windows and Linux. The emulator is capable of running several commercial games.

-yuzu only emulates a subset of Switch hardware and therefore is generally only useful for running/debugging homebrew applications. yuzu can boot some games, to varying degrees of success.
+yuzu only emulates a subset of Switch hardware and therefore most commercial games **do not** run at full speed or are not fully functional.
+
+Do you want to check which games are compatible and which ones are not? Please visit our [Compatibility page](https://yuzu-emu.org/game/)!

 yuzu is licensed under the GPLv2 (or any later version). Refer to the license.txt file included.

@@ -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
@@ -107,6 +107,11 @@ Stream::State AudioRenderer::GetStreamState() const {
    return stream->GetState();
 }

+static constexpr u32 VersionFromRevision(u32_le rev) {
+    // "REV7" -> 7
+    return ((rev >> 24) & 0xff) - 0x30;
+}
+
 std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_params) {
    // Copy UpdateDataHeader struct
    UpdateDataHeader config{};
@@ -166,6 +171,11 @@ std::vector<u8> AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_
    // Copy output header
    UpdateDataHeader response_data{worker_params};
    std::vector<u8> output_params(response_data.total_size);
+    const auto audren_revision = VersionFromRevision(config.revision);
+    if (audren_revision >= 5) {
+        response_data.frame_count = 0x10;
+        response_data.total_size += 0x10;
+    }
    std::memcpy(output_params.data(), &response_data, sizeof(UpdateDataHeader));

    // Copy output memory pool entries
@@ -194,21 +194,24 @@ struct UpdateDataHeader {
        mixes_size = 0x0;
        sinks_size = config.sink_count * 0x20;
        performance_manager_size = 0x10;
+        frame_count = 0;
        total_size = sizeof(UpdateDataHeader) + behavior_size + memory_pools_size + voices_size +
                     effects_size + sinks_size + performance_manager_size;
    }

-    u32_le revision;
-    u32_le behavior_size;
-    u32_le memory_pools_size;
-    u32_le voices_size;
-    u32_le voice_resource_size;
-    u32_le effects_size;
-    u32_le mixes_size;
-    u32_le sinks_size;
-    u32_le performance_manager_size;
-    INSERT_PADDING_WORDS(6);
-    u32_le total_size;
+    u32_le revision{};
+    u32_le behavior_size{};
+    u32_le memory_pools_size{};
+    u32_le voices_size{};
+    u32_le voice_resource_size{};
+    u32_le effects_size{};
+    u32_le mixes_size{};
+    u32_le sinks_size{};
+    u32_le performance_manager_size{};
+    INSERT_PADDING_WORDS(1);
+    u32_le frame_count{};
+    INSERT_PADDING_WORDS(4);
+    u32_le total_size{};
 };
 static_assert(sizeof(UpdateDataHeader) == 0x40, "UpdateDataHeader has wrong size");

@@ -10,6 +10,9 @@ if (DEFINED ENV{CI})
  elseif(DEFINED ENV{APPVEYOR})
    set(BUILD_REPOSITORY $ENV{APPVEYOR_REPO_NAME})
    set(BUILD_TAG $ENV{APPVEYOR_REPO_TAG_NAME})
+  elseif(DEFINED ENV{AZURE})
+    set(BUILD_REPOSITORY $ENV{AZURE_REPO_NAME})
+    set(BUILD_TAG $ENV{AZURE_REPO_TAG})
  endif()
 endif()
 add_custom_command(OUTPUT scm_rev.cpp
@@ -55,6 +58,7 @@ add_custom_command(OUTPUT scm_rev.cpp
      "${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"
@@ -28,18 +28,14 @@ __declspec(noinline, noreturn)
 }

 #define ASSERT(_a_)                                                                                \
-    do                                                                                             \
-        if (!(_a_)) {                                                                              \
-            assert_noinline_call([] { LOG_CRITICAL(Debug, "Assertion Failed!"); });                \
-        }                                                                                          \
-    while (0)
+    if (!(_a_)) {                                                                                  \
+        LOG_CRITICAL(Debug, "Assertion Failed!");                                                  \
+    }

 #define ASSERT_MSG(_a_, ...)                                                                       \
-    do                                                                                             \
-        if (!(_a_)) {                                                                              \
-            assert_noinline_call([&] { LOG_CRITICAL(Debug, "Assertion Failed!\n" __VA_ARGS__); }); \
-        }                                                                                          \
-    while (0)
+    if (!(_a_)) {                                                                                  \
+        LOG_CRITICAL(Debug, "Assertion Failed! " __VA_ARGS__);                                     \
+    }

 #define UNREACHABLE() ASSERT_MSG(false, "Unreachable code!")
 #define UNREACHABLE_MSG(...) ASSERT_MSG(false, __VA_ARGS__)
@@ -255,6 +255,7 @@ void DebuggerBackend::Write(const Entry& entry) {
    CLS(Input)                                                                                     \
    CLS(Network)                                                                                   \
    CLS(Loader)                                                                                    \
+    CLS(CheatEngine)                                                                               \
    CLS(Crypto)                                                                                    \
    CLS(WebService)

@@ -117,6 +117,7 @@ enum class Class : ClassType {
    Audio_DSP,         ///< The HLE implementation of the DSP
    Audio_Sink,        ///< Emulator audio output backend
    Loader,            ///< ROM loader
+    CheatEngine,       ///< Memory manipulation and engine VM functions
    Crypto,            ///< Cryptographic engine/functions
    Input,             ///< Input emulation
    Network,           ///< Network emulation
@@ -33,8 +33,6 @@ add_library(core STATIC
    file_sys/bis_factory.h
    file_sys/card_image.cpp
    file_sys/card_image.h
-    file_sys/cheat_engine.cpp
-    file_sys/cheat_engine.h
    file_sys/content_archive.cpp
    file_sys/content_archive.h
    file_sys/control_metadata.cpp
@@ -70,6 +68,8 @@ add_library(core STATIC
    file_sys/sdmc_factory.h
    file_sys/submission_package.cpp
    file_sys/submission_package.h
+    file_sys/system_archive/mii_model.cpp
+    file_sys/system_archive/mii_model.h
    file_sys/system_archive/ng_word.cpp
    file_sys/system_archive/ng_word.h
    file_sys/system_archive/system_archive.cpp
@@ -475,6 +475,11 @@ add_library(core STATIC
    loader/nsp.h
    loader/xci.cpp
    loader/xci.h
+    memory/cheat_engine.cpp
+    memory/cheat_engine.h
+    memory/dmnt_cheat_types.h
+    memory/dmnt_cheat_vm.cpp
+    memory/dmnt_cheat_vm.h
    memory.cpp
    memory.h
    memory_setup.h
@@ -248,6 +248,9 @@ void ARM_Dynarmic::LoadContext(const ThreadContext& ctx) {
 }

 void ARM_Dynarmic::PrepareReschedule() {
+    if (jit == nullptr)
+        return;
+
    jit->HaltExecution();
 }

@@ -14,8 +14,14 @@
 #include "core/core_cpu.h"
 #include "core/core_timing.h"
 #include "core/cpu_core_manager.h"
+#include "core/file_sys/bis_factory.h"
+#include "core/file_sys/card_image.h"
 #include "core/file_sys/mode.h"
+#include "core/file_sys/patch_manager.h"
 #include "core/file_sys/registered_cache.h"
+#include "core/file_sys/romfs_factory.h"
+#include "core/file_sys/savedata_factory.h"
+#include "core/file_sys/sdmc_factory.h"
 #include "core/file_sys/vfs_concat.h"
 #include "core/file_sys/vfs_real.h"
 #include "core/gdbstub/gdbstub.h"
@@ -27,17 +33,17 @@
 #include "core/hle/kernel/thread.h"
 #include "core/hle/service/am/applets/applets.h"
 #include "core/hle/service/apm/controller.h"
+#include "core/hle/service/filesystem/filesystem.h"
 #include "core/hle/service/glue/manager.h"
 #include "core/hle/service/service.h"
 #include "core/hle/service/sm/sm.h"
 #include "core/loader/loader.h"
+#include "core/memory/cheat_engine.h"
 #include "core/perf_stats.h"
 #include "core/reporter.h"
 #include "core/settings.h"
 #include "core/telemetry_session.h"
 #include "core/tools/freezer.h"
-#include "file_sys/cheat_engine.h"
-#include "file_sys/patch_manager.h"
 #include "video_core/debug_utils/debug_utils.h"
 #include "video_core/renderer_base.h"
 #include "video_core/video_core.h"
@@ -104,7 +110,8 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
    return vfs->OpenFile(path, FileSys::Mode::Read);
 }
 struct System::Impl {
-    explicit Impl(System& system) : kernel{system}, cpu_core_manager{system}, reporter{system} {}
+    explicit Impl(System& system)
+        : kernel{system}, cpu_core_manager{system}, applet_manager{system}, reporter{system} {}

    Cpu& CurrentCpuCore() {
        return cpu_core_manager.GetCurrentCore();
@@ -156,13 +163,10 @@ struct System::Impl {
        gpu_core = VideoCore::CreateGPU(system);

        is_powered_on = true;
+        exit_lock = false;

        LOG_DEBUG(Core, "Initialized OK");

-        // Reset counters and set time origin to current frame
-        GetAndResetPerfStats();
-        perf_stats.BeginSystemFrame();
-
        return ResultStatus::Success;
    }

@@ -201,10 +205,34 @@ struct System::Impl {
        gpu_core->Start();
        cpu_core_manager.StartThreads();

+        // Initialize cheat engine
+        if (cheat_engine) {
+            cheat_engine->Initialize();
+        }
+
        // All threads are started, begin main process execution, now that we're in the clear.
        main_process->Run(load_parameters->main_thread_priority,
                          load_parameters->main_thread_stack_size);

+        if (Settings::values.gamecard_inserted) {
+            if (Settings::values.gamecard_current_game) {
+                fs_controller.SetGameCard(GetGameFileFromPath(virtual_filesystem, filepath));
+            } else if (!Settings::values.gamecard_path.empty()) {
+                fs_controller.SetGameCard(
+                    GetGameFileFromPath(virtual_filesystem, Settings::values.gamecard_path));
+            }
+        }
+
+        u64 title_id{0};
+        if (app_loader->ReadProgramId(title_id) != Loader::ResultStatus::Success) {
+            LOG_ERROR(Core, "Failed to find title id for ROM (Error {})",
+                      static_cast<u32>(load_result));
+        }
+        perf_stats = std::make_unique<PerfStats>(title_id);
+        // Reset counters and set time origin to current frame
+        GetAndResetPerfStats();
+        perf_stats->BeginSystemFrame();
+
        status = ResultStatus::Success;
        return status;
    }
@@ -218,8 +246,11 @@ struct System::Impl {
                                    perf_results.game_fps);
        telemetry_session->AddField(Telemetry::FieldType::Performance, "Shutdown_Frametime",
                                    perf_results.frametime * 1000.0);
+        telemetry_session->AddField(Telemetry::FieldType::Performance, "Mean_Frametime_MS",
+                                    perf_stats->GetMeanFrametime());

        is_powered_on = false;
+        exit_lock = false;

        // Shutdown emulation session
        renderer.reset();
@@ -228,6 +259,7 @@ struct System::Impl {
        service_manager.reset();
        cheat_engine.reset();
        telemetry_session.reset();
+        perf_stats.reset();
        gpu_core.reset();

        // Close all CPU/threading state
@@ -285,7 +317,7 @@ struct System::Impl {
    }

    PerfStatsResults GetAndResetPerfStats() {
-        return perf_stats.GetAndResetStats(core_timing.GetGlobalTimeUs());
+        return perf_stats->GetAndResetStats(core_timing.GetGlobalTimeUs());
    }

    Timing::CoreTiming core_timing;
@@ -294,6 +326,7 @@ struct System::Impl {
    FileSys::VirtualFilesystem virtual_filesystem;
    /// ContentProviderUnion instance
    std::unique_ptr<FileSys::ContentProviderUnion> content_provider;
+    Service::FileSystem::FileSystemController fs_controller;
    /// AppLoader used to load the current executing application
    std::unique_ptr<Loader::AppLoader> app_loader;
    std::unique_ptr<VideoCore::RendererBase> renderer;
@@ -302,8 +335,9 @@ struct System::Impl {
    std::unique_ptr<Core::Hardware::InterruptManager> interrupt_manager;
    CpuCoreManager cpu_core_manager;
    bool is_powered_on = false;
+    bool exit_lock = false;

-    std::unique_ptr<FileSys::CheatEngine> cheat_engine;
+    std::unique_ptr<Memory::CheatEngine> cheat_engine;
    std::unique_ptr<Tools::Freezer> memory_freezer;

    /// Frontend applets
@@ -326,7 +360,7 @@ struct System::Impl {
    ResultStatus status = ResultStatus::Success;
    std::string status_details = "";

-    Core::PerfStats perf_stats;
+    std::unique_ptr<Core::PerfStats> perf_stats;
    Core::FrameLimiter frame_limiter;
 };

@@ -365,6 +399,12 @@ void System::PrepareReschedule() {
    CurrentCpuCore().PrepareReschedule();
 }

+void System::PrepareReschedule(const u32 core_index) {
+    if (core_index < GlobalScheduler().CpuCoresCount()) {
+        CpuCore(core_index).PrepareReschedule();
+    }
+}
+
 PerfStatsResults System::GetAndResetPerfStats() {
    return impl->GetAndResetPerfStats();
 }
@@ -405,6 +445,16 @@ const Kernel::Scheduler& System::Scheduler(std::size_t core_index) const {
    return CpuCore(core_index).Scheduler();
 }

+/// Gets the global scheduler
+Kernel::GlobalScheduler& System::GlobalScheduler() {
+    return impl->kernel.GlobalScheduler();
+}
+
+/// Gets the global scheduler
+const Kernel::GlobalScheduler& System::GlobalScheduler() const {
+    return impl->kernel.GlobalScheduler();
+}
+
 Kernel::Process* System::CurrentProcess() {
    return impl->kernel.CurrentProcess();
 }
@@ -479,11 +529,11 @@ const Timing::CoreTiming& System::CoreTiming() const {
 }

 Core::PerfStats& System::GetPerfStats() {
-    return impl->perf_stats;
+    return *impl->perf_stats;
 }

 const Core::PerfStats& System::GetPerfStats() const {
-    return impl->perf_stats;
+    return *impl->perf_stats;
 }

 Core::FrameLimiter& System::FrameLimiter() {
@@ -518,13 +568,6 @@ Tegra::DebugContext* System::GetGPUDebugContext() const {
    return impl->debug_context.get();
 }

-void System::RegisterCheatList(const std::vector<FileSys::CheatList>& list,
-                               const std::string& build_id, VAddr code_region_start,
-                               VAddr code_region_end) {
-    impl->cheat_engine = std::make_unique<FileSys::CheatEngine>(*this, list, build_id,
-                                                                code_region_start, code_region_end);
-}
-
 void System::SetFilesystem(std::shared_ptr<FileSys::VfsFilesystem> vfs) {
    impl->virtual_filesystem = std::move(vfs);
 }
@@ -533,6 +576,13 @@ std::shared_ptr<FileSys::VfsFilesystem> System::GetFilesystem() const {
    return impl->virtual_filesystem;
 }

+void System::RegisterCheatList(const std::vector<Memory::CheatEntry>& list,
+                               const std::array<u8, 32>& build_id, VAddr main_region_begin,
+                               u64 main_region_size) {
+    impl->cheat_engine = std::make_unique<Memory::CheatEngine>(*this, list, build_id);
+    impl->cheat_engine->SetMainMemoryParameters(main_region_begin, main_region_size);
+}
+
 void System::SetAppletFrontendSet(Service::AM::Applets::AppletFrontendSet&& set) {
    impl->applet_manager.SetAppletFrontendSet(std::move(set));
 }
@@ -561,6 +611,14 @@ const FileSys::ContentProvider& System::GetContentProvider() const {
    return *impl->content_provider;
 }

+Service::FileSystem::FileSystemController& System::GetFileSystemController() {
+    return impl->fs_controller;
+}
+
+const Service::FileSystem::FileSystemController& System::GetFileSystemController() const {
+    return impl->fs_controller;
+}
+
 void System::RegisterContentProvider(FileSys::ContentProviderUnionSlot slot,
                                     FileSys::ContentProvider* provider) {
    impl->content_provider->SetSlot(slot, provider);
@@ -590,6 +648,14 @@ const Service::APM::Controller& System::GetAPMController() const {
    return impl->apm_controller;
 }

+void System::SetExitLock(bool locked) {
+    impl->exit_lock = locked;
+}
+
+bool System::GetExitLock() const {
+    return impl->exit_lock;
+}
+
 System::ResultStatus System::Init(Frontend::EmuWindow& emu_window) {
    return impl->Init(*this, emu_window);
 }
@@ -18,7 +18,6 @@ class EmuWindow;
 } // namespace Core::Frontend

 namespace FileSys {
-class CheatList;
 class ContentProvider;
 class ContentProviderUnion;
 enum class ContentProviderUnionSlot;
@@ -26,6 +25,7 @@ class VfsFilesystem;
 } // namespace FileSys

 namespace Kernel {
+class GlobalScheduler;
 class KernelCore;
 class Process;
 class Scheduler;
@@ -36,6 +36,10 @@ class AppLoader;
 enum class ResultStatus : u16;
 } // namespace Loader

+namespace Memory {
+struct CheatEntry;
+} // namespace Memory
+
 namespace Service {

 namespace AM::Applets {
@@ -47,6 +51,10 @@ namespace APM {
 class Controller;
 }

+namespace FileSystem {
+class FileSystemController;
+} // namespace FileSystem
+
 namespace Glue {
 class ARPManager;
 }
@@ -172,6 +180,9 @@ public:
    /// Prepare the core emulation for a reschedule
    void PrepareReschedule();

+    /// Prepare the core emulation for a reschedule
+    void PrepareReschedule(u32 core_index);
+
    /// Gets and resets core performance statistics
    PerfStatsResults GetAndResetPerfStats();

@@ -226,6 +237,12 @@ public:
    /// Gets the scheduler for the CPU core with the specified index
    const Kernel::Scheduler& Scheduler(std::size_t core_index) const;

+    /// Gets the global scheduler
+    Kernel::GlobalScheduler& GlobalScheduler();
+
+    /// Gets the global scheduler
+    const Kernel::GlobalScheduler& GlobalScheduler() const;
+
    /// Provides a pointer to the current process
    Kernel::Process* CurrentProcess();

@@ -282,8 +299,9 @@ public:

    std::shared_ptr<FileSys::VfsFilesystem> GetFilesystem() const;

-    void RegisterCheatList(const std::vector<FileSys::CheatList>& list, const std::string& build_id,
-                           VAddr code_region_start, VAddr code_region_end);
+    void RegisterCheatList(const std::vector<Memory::CheatEntry>& list,
+                           const std::array<u8, 0x20>& build_id, VAddr main_region_begin,
+                           u64 main_region_size);

    void SetAppletFrontendSet(Service::AM::Applets::AppletFrontendSet&& set);

@@ -299,6 +317,10 @@ public:

    const FileSys::ContentProvider& GetContentProvider() const;

+    Service::FileSystem::FileSystemController& GetFileSystemController();
+
+    const Service::FileSystem::FileSystemController& GetFileSystemController() const;
+
    void RegisterContentProvider(FileSys::ContentProviderUnionSlot slot,
                                 FileSys::ContentProvider* provider);

@@ -314,6 +336,10 @@ public:

    const Service::APM::Controller& GetAPMController() const;

+    void SetExitLock(bool locked);
+
+    bool GetExitLock() const;
+
 private:
    System();

@@ -52,7 +52,8 @@ bool CpuBarrier::Rendezvous() {

 Cpu::Cpu(System& system, ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_barrier,
         std::size_t core_index)
-    : cpu_barrier{cpu_barrier}, core_timing{system.CoreTiming()}, core_index{core_index} {
+    : cpu_barrier{cpu_barrier}, global_scheduler{system.GlobalScheduler()},
+      core_timing{system.CoreTiming()}, core_index{core_index} {
 #ifdef ARCHITECTURE_x86_64
    arm_interface = std::make_unique<ARM_Dynarmic>(system, exclusive_monitor, core_index);
 #else
@@ -60,7 +61,7 @@ Cpu::Cpu(System& system, ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_ba
    LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
 #endif

-    scheduler = std::make_unique<Kernel::Scheduler>(system, *arm_interface);
+    scheduler = std::make_unique<Kernel::Scheduler>(system, *arm_interface, core_index);
 }

 Cpu::~Cpu() = default;
@@ -81,6 +82,8 @@ void Cpu::RunLoop(bool tight_loop) {
        return;
    }

+    Reschedule();
+
    // If we don't have a currently active thread then don't execute instructions,
    // instead advance to the next event and try to yield to the next thread
    if (Kernel::GetCurrentThread() == nullptr) {
@@ -92,7 +95,6 @@ void Cpu::RunLoop(bool tight_loop) {
            core_timing.Advance();
        }

-        PrepareReschedule();
    } else {
        if (IsMainCore()) {
            core_timing.Advance();
@@ -114,18 +116,14 @@ void Cpu::SingleStep() {

 void Cpu::PrepareReschedule() {
    arm_interface->PrepareReschedule();
-    reschedule_pending = true;
 }

 void Cpu::Reschedule() {
-    if (!reschedule_pending) {
-        return;
-    }
-
-    reschedule_pending = false;
    // Lock the global kernel mutex when we manipulate the HLE state
-    std::lock_guard lock{HLE::g_hle_lock};
-    scheduler->Reschedule();
+    std::lock_guard lock(HLE::g_hle_lock);
+
+    global_scheduler.SelectThread(core_index);
+    scheduler->TryDoContextSwitch();
 }

 } // namespace Core
@@ -12,8 +12,9 @@
 #include "common/common_types.h"

 namespace Kernel {
+class GlobalScheduler;
 class Scheduler;
-}
+} // namespace Kernel

 namespace Core {
 class System;
@@ -90,6 +91,7 @@ private:

    std::unique_ptr<ARM_Interface> arm_interface;
    CpuBarrier& cpu_barrier;
+    Kernel::GlobalScheduler& global_scheduler;
    std::unique_ptr<Kernel::Scheduler> scheduler;
    Timing::CoreTiming& core_timing;

@@ -37,6 +37,7 @@
 namespace Core::Crypto {

 constexpr u64 CURRENT_CRYPTO_REVISION = 0x5;
+constexpr u64 FULL_TICKET_SIZE = 0x400;

 using namespace Common;

@@ -55,6 +56,99 @@ const std::map<std::pair<S128KeyType, u64>, std::string> KEYS_VARIABLE_LENGTH{
    {{S128KeyType::KeyblobMAC, 0}, "keyblob_mac_key_"},
 };

+namespace {
+template <std::size_t Size>
+bool IsAllZeroArray(const std::array<u8, Size>& array) {
+    return std::all_of(array.begin(), array.end(), [](const auto& elem) { return elem == 0; });
+}
+} // namespace
+
+u64 GetSignatureTypeDataSize(SignatureType type) {
+    switch (type) {
+    case SignatureType::RSA_4096_SHA1:
+    case SignatureType::RSA_4096_SHA256:
+        return 0x200;
+    case SignatureType::RSA_2048_SHA1:
+    case SignatureType::RSA_2048_SHA256:
+        return 0x100;
+    case SignatureType::ECDSA_SHA1:
+    case SignatureType::ECDSA_SHA256:
+        return 0x3C;
+    }
+    UNREACHABLE();
+}
+
+u64 GetSignatureTypePaddingSize(SignatureType type) {
+    switch (type) {
+    case SignatureType::RSA_4096_SHA1:
+    case SignatureType::RSA_4096_SHA256:
+    case SignatureType::RSA_2048_SHA1:
+    case SignatureType::RSA_2048_SHA256:
+        return 0x3C;
+    case SignatureType::ECDSA_SHA1:
+    case SignatureType::ECDSA_SHA256:
+        return 0x40;
+    }
+    UNREACHABLE();
+}
+
+SignatureType Ticket::GetSignatureType() const {
+    if (auto ticket = std::get_if<RSA4096Ticket>(&data)) {
+        return ticket->sig_type;
+    }
+    if (auto ticket = std::get_if<RSA2048Ticket>(&data)) {
+        return ticket->sig_type;
+    }
+    if (auto ticket = std::get_if<ECDSATicket>(&data)) {
+        return ticket->sig_type;
+    }
+
+    UNREACHABLE();
+}
+
+TicketData& Ticket::GetData() {
+    if (auto ticket = std::get_if<RSA4096Ticket>(&data)) {
+        return ticket->data;
+    }
+    if (auto ticket = std::get_if<RSA2048Ticket>(&data)) {
+        return ticket->data;
+    }
+    if (auto ticket = std::get_if<ECDSATicket>(&data)) {
+        return ticket->data;
+    }
+
+    UNREACHABLE();
+}
+
+const TicketData& Ticket::GetData() const {
+    if (auto ticket = std::get_if<RSA4096Ticket>(&data)) {
+        return ticket->data;
+    }
+    if (auto ticket = std::get_if<RSA2048Ticket>(&data)) {
+        return ticket->data;
+    }
+    if (auto ticket = std::get_if<ECDSATicket>(&data)) {
+        return ticket->data;
+    }
+
+    UNREACHABLE();
+}
+
+u64 Ticket::GetSize() const {
+    const auto sig_type = GetSignatureType();
+
+    return sizeof(SignatureType) + GetSignatureTypeDataSize(sig_type) +
+           GetSignatureTypePaddingSize(sig_type) + sizeof(TicketData);
+}
+
+Ticket Ticket::SynthesizeCommon(Key128 title_key, const std::array<u8, 16>& rights_id) {
+    RSA2048Ticket out{};
+    out.sig_type = SignatureType::RSA_2048_SHA256;
+    out.data.rights_id = rights_id;
+    out.data.title_key_common = title_key;
+    return Ticket{out};
+}
+
 Key128 GenerateKeyEncryptionKey(Key128 source, Key128 master, Key128 kek_seed, Key128 key_seed) {
    Key128 out{};

@@ -135,6 +229,27 @@ void KeyManager::DeriveGeneralPurposeKeys(std::size_t crypto_revision) {
    }
 }

+RSAKeyPair<2048> KeyManager::GetETicketRSAKey() const {
+    if (IsAllZeroArray(eticket_extended_kek) || !HasKey(S128KeyType::ETicketRSAKek))
+        return {};
+
+    const auto eticket_final = GetKey(S128KeyType::ETicketRSAKek);
+
+    std::vector<u8> extended_iv(eticket_extended_kek.begin(), eticket_extended_kek.begin() + 0x10);
+    std::array<u8, 0x230> extended_dec{};
+    AESCipher<Key128> rsa_1(eticket_final, Mode::CTR);
+    rsa_1.SetIV(extended_iv);
+    rsa_1.Transcode(eticket_extended_kek.data() + 0x10, eticket_extended_kek.size() - 0x10,
+                    extended_dec.data(), Op::Decrypt);
+
+    RSAKeyPair<2048> rsa_key{};
+    std::memcpy(rsa_key.decryption_key.data(), extended_dec.data(), rsa_key.decryption_key.size());
+    std::memcpy(rsa_key.modulus.data(), extended_dec.data() + 0x100, rsa_key.modulus.size());
+    std::memcpy(rsa_key.exponent.data(), extended_dec.data() + 0x200, rsa_key.exponent.size());
+
+    return rsa_key;
+}
+
 Key128 DeriveKeyblobMACKey(const Key128& keyblob_key, const Key128& mac_source) {
    AESCipher<Key128> mac_cipher(keyblob_key, Mode::ECB);
    Key128 mac_key{};
@@ -237,7 +352,7 @@ Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& ke
    return Loader::ResultStatus::Success;
 }

-std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save) {
+std::vector<Ticket> GetTicketblob(const FileUtil::IOFile& ticket_save) {
    if (!ticket_save.IsOpen())
        return {};

@@ -246,14 +361,14 @@ std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save) {
        return {};
    }

-    std::vector<TicketRaw> out;
+    std::vector<Ticket> out;
    for (std::size_t offset = 0; offset + 0x4 < buffer.size(); ++offset) {
        if (buffer[offset] == 0x4 && buffer[offset + 1] == 0x0 && buffer[offset + 2] == 0x1 &&
            buffer[offset + 3] == 0x0) {
            out.emplace_back();
            auto& next = out.back();
-            std::memcpy(&next, buffer.data() + offset, sizeof(TicketRaw));
-            offset += next.size();
+            std::memcpy(&next, buffer.data() + offset, sizeof(Ticket));
+            offset += FULL_TICKET_SIZE;
        }
    }

@@ -305,29 +420,23 @@ static std::optional<u64> FindTicketOffset(const std::array<u8, size>& data) {
    return offset;
 }

-std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
+std::optional<std::pair<Key128, Key128>> ParseTicket(const Ticket& ticket,
                                                     const RSAKeyPair<2048>& key) {
-    u32 cert_authority;
-    std::memcpy(&cert_authority, ticket.data() + 0x140, sizeof(cert_authority));
-    if (cert_authority == 0)
+    const auto issuer = ticket.GetData().issuer;
+    if (issuer == std::array<u8, 0x40>{})
        return {};
-    if (cert_authority != Common::MakeMagic('R', 'o', 'o', 't')) {
-        LOG_INFO(Crypto,
-                 "Attempting to parse ticket with non-standard certificate authority {:08X}.",
-                 cert_authority);
+    if (issuer[0] != 'R' || issuer[1] != 'o' || issuer[2] != 'o' || issuer[3] != 't') {
+        LOG_INFO(Crypto, "Attempting to parse ticket with non-standard certificate authority.");
    }

-    Key128 rights_id;
-    std::memcpy(rights_id.data(), ticket.data() + 0x2A0, sizeof(Key128));
+    Key128 rights_id = ticket.GetData().rights_id;

    if (rights_id == Key128{})
        return {};

-    Key128 key_temp{};
-
-    if (!std::any_of(ticket.begin() + 0x190, ticket.begin() + 0x280, [](u8 b) { return b != 0; })) {
-        std::memcpy(key_temp.data(), ticket.data() + 0x180, key_temp.size());
-        return std::make_pair(rights_id, key_temp);
+    if (!std::any_of(ticket.GetData().title_key_common_pad.begin(),
+                     ticket.GetData().title_key_common_pad.end(), [](u8 b) { return b != 0; })) {
+        return std::make_pair(rights_id, ticket.GetData().title_key_common);
    }

    mbedtls_mpi D; // RSA Private Exponent
@@ -342,7 +451,7 @@ std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,

    mbedtls_mpi_read_binary(&D, key.decryption_key.data(), key.decryption_key.size());
    mbedtls_mpi_read_binary(&N, key.modulus.data(), key.modulus.size());
-    mbedtls_mpi_read_binary(&S, ticket.data() + 0x180, 0x100);
+    mbedtls_mpi_read_binary(&S, ticket.GetData().title_key_block.data(), 0x100);

    mbedtls_mpi_exp_mod(&M, &S, &D, &N, nullptr);

@@ -366,6 +475,7 @@ std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
        return {};
    ASSERT(*offset > 0);

+    Key128 key_temp{};
    std::memcpy(key_temp.data(), m_2.data() + *offset, key_temp.size());

    return std::make_pair(rights_id, key_temp);
@@ -450,6 +560,8 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) {

                const auto index = std::stoul(out[0].substr(18, 2), nullptr, 16);
                encrypted_keyblobs[index] = Common::HexStringToArray<0xB0>(out[1]);
+            } else if (out[0].compare(0, 20, "eticket_extended_kek") == 0) {
+                eticket_extended_kek = Common::HexStringToArray<576>(out[1]);
            } else {
                for (const auto& kv : KEYS_VARIABLE_LENGTH) {
                    if (!ValidCryptoRevisionString(out[0], kv.second.size(), 2))
@@ -862,20 +974,19 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) {
    // Titlekeys
    data.DecryptProdInfo(GetBISKey(0));

-    const auto eticket_extended_kek = data.GetETicketExtendedKek();
+    eticket_extended_kek = data.GetETicketExtendedKek();
+    WriteKeyToFile(KeyCategory::Console, "eticket_extended_kek", eticket_extended_kek);
+    PopulateTickets();
+}

-    std::vector<u8> extended_iv(0x10);
-    std::memcpy(extended_iv.data(), eticket_extended_kek.data(), extended_iv.size());
-    std::array<u8, 0x230> extended_dec{};
-    AESCipher<Key128> rsa_1(eticket_final, Mode::CTR);
-    rsa_1.SetIV(extended_iv);
-    rsa_1.Transcode(eticket_extended_kek.data() + 0x10, eticket_extended_kek.size() - 0x10,
-                    extended_dec.data(), Op::Decrypt);
+void KeyManager::PopulateTickets() {
+    const auto rsa_key = GetETicketRSAKey();

-    RSAKeyPair<2048> rsa_key{};
-    std::memcpy(rsa_key.decryption_key.data(), extended_dec.data(), rsa_key.decryption_key.size());
-    std::memcpy(rsa_key.modulus.data(), extended_dec.data() + 0x100, rsa_key.modulus.size());
-    std::memcpy(rsa_key.exponent.data(), extended_dec.data() + 0x200, rsa_key.exponent.size());
+    if (rsa_key == RSAKeyPair<2048>{})
+        return;
+
+    if (!common_tickets.empty() && !personal_tickets.empty())
+        return;

    const FileUtil::IOFile save1(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
                                     "/system/save/80000000000000e1",
@@ -886,19 +997,41 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) {

    const auto blob2 = GetTicketblob(save2);
    auto res = GetTicketblob(save1);
+    const auto idx = res.size();
    res.insert(res.end(), blob2.begin(), blob2.end());

-    for (const auto& raw : res) {
-        const auto pair = ParseTicket(raw, rsa_key);
+    for (std::size_t i = 0; i < res.size(); ++i) {
+        const auto common = i < idx;
+        const auto pair = ParseTicket(res[i], rsa_key);
        if (!pair)
            continue;
        const auto& [rid, key] = *pair;
        u128 rights_id;
        std::memcpy(rights_id.data(), rid.data(), rid.size());
+
+        if (common) {
+            common_tickets[rights_id] = res[i];
+        } else {
+            personal_tickets[rights_id] = res[i];
+        }
+
        SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
    }
 }

+void KeyManager::SynthesizeTickets() {
+    for (const auto& key : s128_keys) {
+        if (key.first.type != S128KeyType::Titlekey) {
+            continue;
+        }
+        u128 rights_id{key.first.field1, key.first.field2};
+        Key128 rights_id_2;
+        std::memcpy(rights_id_2.data(), rights_id.data(), rights_id_2.size());
+        const auto ticket = Ticket::SynthesizeCommon(key.second, rights_id_2);
+        common_tickets.insert_or_assign(rights_id, ticket);
+    }
+}
+
 void KeyManager::SetKeyWrapped(S128KeyType id, Key128 key, u64 field1, u64 field2) {
    if (key == Key128{})
        return;
@@ -997,6 +1130,46 @@ void KeyManager::PopulateFromPartitionData(PartitionDataManager& data) {
    DeriveBase();
 }

+const std::map<u128, Ticket>& KeyManager::GetCommonTickets() const {
+    return common_tickets;
+}
+
+const std::map<u128, Ticket>& KeyManager::GetPersonalizedTickets() const {
+    return personal_tickets;
+}
+
+bool KeyManager::AddTicketCommon(Ticket raw) {
+    const auto rsa_key = GetETicketRSAKey();
+    if (rsa_key == RSAKeyPair<2048>{})
+        return false;
+
+    const auto pair = ParseTicket(raw, rsa_key);
+    if (!pair)
+        return false;
+    const auto& [rid, key] = *pair;
+    u128 rights_id;
+    std::memcpy(rights_id.data(), rid.data(), rid.size());
+    common_tickets[rights_id] = raw;
+    SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
+    return true;
+}
+
+bool KeyManager::AddTicketPersonalized(Ticket raw) {
+    const auto rsa_key = GetETicketRSAKey();
+    if (rsa_key == RSAKeyPair<2048>{})
+        return false;
+
+    const auto pair = ParseTicket(raw, rsa_key);
+    if (!pair)
+        return false;
+    const auto& [rid, key] = *pair;
+    u128 rights_id;
+    std::memcpy(rights_id.data(), rid.data(), rid.size());
+    common_tickets[rights_id] = raw;
+    SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
+    return true;
+}
+
 const boost::container::flat_map<std::string, KeyIndex<S128KeyType>> KeyManager::s128_file_id = {
    {"eticket_rsa_kek", {S128KeyType::ETicketRSAKek, 0, 0}},
    {"eticket_rsa_kek_source",
@@ -9,8 +9,10 @@
 #include <optional>
 #include <string>

+#include <variant>
 #include <boost/container/flat_map.hpp>
 #include <fmt/format.h>
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "core/crypto/partition_data_manager.h"
 #include "core/file_sys/vfs_types.h"
@@ -30,7 +32,79 @@ constexpr u64 TICKET_FILE_TITLEKEY_OFFSET = 0x180;
 using Key128 = std::array<u8, 0x10>;
 using Key256 = std::array<u8, 0x20>;
 using SHA256Hash = std::array<u8, 0x20>;
-using TicketRaw = std::array<u8, 0x400>;
+
+enum class SignatureType {
+    RSA_4096_SHA1 = 0x10000,
+    RSA_2048_SHA1 = 0x10001,
+    ECDSA_SHA1 = 0x10002,
+    RSA_4096_SHA256 = 0x10003,
+    RSA_2048_SHA256 = 0x10004,
+    ECDSA_SHA256 = 0x10005,
+};
+
+u64 GetSignatureTypeDataSize(SignatureType type);
+u64 GetSignatureTypePaddingSize(SignatureType type);
+
+enum class TitleKeyType : u8 {
+    Common = 0,
+    Personalized = 1,
+};
+
+struct TicketData {
+    std::array<u8, 0x40> issuer;
+    union {
+        std::array<u8, 0x100> title_key_block;
+
+        struct {
+            Key128 title_key_common;
+            std::array<u8, 0xF0> title_key_common_pad;
+        };
+    };
+
+    INSERT_PADDING_BYTES(0x1);
+    TitleKeyType type;
+    INSERT_PADDING_BYTES(0x3);
+    u8 revision;
+    INSERT_PADDING_BYTES(0xA);
+    u64 ticket_id;
+    u64 device_id;
+    std::array<u8, 0x10> rights_id;
+    u32 account_id;
+    INSERT_PADDING_BYTES(0x14C);
+};
+static_assert(sizeof(TicketData) == 0x2C0, "TicketData has incorrect size.");
+
+struct RSA4096Ticket {
+    SignatureType sig_type;
+    std::array<u8, 0x200> sig_data;
+    INSERT_PADDING_BYTES(0x3C);
+    TicketData data;
+};
+
+struct RSA2048Ticket {
+    SignatureType sig_type;
+    std::array<u8, 0x100> sig_data;
+    INSERT_PADDING_BYTES(0x3C);
+    TicketData data;
+};
+
+struct ECDSATicket {
+    SignatureType sig_type;
+    std::array<u8, 0x3C> sig_data;
+    INSERT_PADDING_BYTES(0x40);
+    TicketData data;
+};
+
+struct Ticket {
+    std::variant<RSA4096Ticket, RSA2048Ticket, ECDSATicket> data;
+
+    SignatureType GetSignatureType() const;
+    TicketData& GetData();
+    const TicketData& GetData() const;
+    u64 GetSize() const;
+
+    static Ticket SynthesizeCommon(Key128 title_key, const std::array<u8, 0x10>& rights_id);
+};

 static_assert(sizeof(Key128) == 16, "Key128 must be 128 bytes big.");
 static_assert(sizeof(Key256) == 32, "Key256 must be 256 bytes big.");
@@ -43,6 +117,19 @@ struct RSAKeyPair {
    std::array<u8, 4> exponent;
 };

+template <size_t bit_size, size_t byte_size>
+bool operator==(const RSAKeyPair<bit_size, byte_size>& lhs,
+                const RSAKeyPair<bit_size, byte_size>& rhs) {
+    return std::tie(lhs.encryption_key, lhs.decryption_key, lhs.modulus, lhs.exponent) ==
+           std::tie(rhs.encryption_key, rhs.decryption_key, rhs.modulus, rhs.exponent);
+}
+
+template <size_t bit_size, size_t byte_size>
+bool operator!=(const RSAKeyPair<bit_size, byte_size>& lhs,
+                const RSAKeyPair<bit_size, byte_size>& rhs) {
+    return !(lhs == rhs);
+}
+
 enum class KeyCategory : u8 {
    Standard,
    Title,
@@ -151,22 +238,35 @@ public:

    static bool KeyFileExists(bool title);

-    // Call before using the sd seed to attempt to derive it if it dosen't exist. Needs system save
-    // 8*43 and the private file to exist.
+    // Call before using the sd seed to attempt to derive it if it dosen't exist. Needs system
+    // save 8*43 and the private file to exist.
    void DeriveSDSeedLazy();

    bool BaseDeriveNecessary() const;
    void DeriveBase();
    void DeriveETicket(PartitionDataManager& data);
+    void PopulateTickets();
+    void SynthesizeTickets();

    void PopulateFromPartitionData(PartitionDataManager& data);

+    const std::map<u128, Ticket>& GetCommonTickets() const;
+    const std::map<u128, Ticket>& GetPersonalizedTickets() const;
+
+    bool AddTicketCommon(Ticket raw);
+    bool AddTicketPersonalized(Ticket raw);
+
 private:
    std::map<KeyIndex<S128KeyType>, Key128> s128_keys;
    std::map<KeyIndex<S256KeyType>, Key256> s256_keys;

+    // Map from rights ID to ticket
+    std::map<u128, Ticket> common_tickets;
+    std::map<u128, Ticket> personal_tickets;
+
    std::array<std::array<u8, 0xB0>, 0x20> encrypted_keyblobs{};
    std::array<std::array<u8, 0x90>, 0x20> keyblobs{};
+    std::array<u8, 576> eticket_extended_kek{};

    bool dev_mode;
    void LoadFromFile(const std::string& filename, bool is_title_keys);
@@ -178,6 +278,8 @@ private:

    void DeriveGeneralPurposeKeys(std::size_t crypto_revision);

+    RSAKeyPair<2048> GetETicketRSAKey() const;
+
    void SetKeyWrapped(S128KeyType id, Key128 key, u64 field1 = 0, u64 field2 = 0);
    void SetKeyWrapped(S256KeyType id, Key256 key, u64 field1 = 0, u64 field2 = 0);

@@ -195,11 +297,11 @@ std::array<u8, 0x90> DecryptKeyblob(const std::array<u8, 0xB0>& encrypted_keyblo
 std::optional<Key128> DeriveSDSeed();
 Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& keys);

-std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save);
+std::vector<Ticket> GetTicketblob(const FileUtil::IOFile& ticket_save);

-// Returns a pair of {rights_id, titlekey}. Fails if the ticket has no certificate authority (offset
-// 0x140-0x144 is zero)
-std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
+// Returns a pair of {rights_id, titlekey}. Fails if the ticket has no certificate authority
+// (offset 0x140-0x144 is zero)
+std::optional<std::pair<Key128, Key128>> ParseTicket(const Ticket& ticket,
                                                     const RSAKeyPair<2048>& eticket_extended_key);

 } // namespace Core::Crypto
@@ -480,6 +480,10 @@ void PartitionDataManager::DecryptProdInfo(std::array<u8, 0x20> bis_key) {
    prodinfo_decrypted = std::make_shared<XTSEncryptionLayer>(prodinfo, bis_key);
 }

+FileSys::VirtualFile PartitionDataManager::GetDecryptedProdInfo() const {
+    return prodinfo_decrypted;
+}
+
 std::array<u8, 576> PartitionDataManager::GetETicketExtendedKek() const {
    std::array<u8, 0x240> out{};
    if (prodinfo_decrypted != nullptr)
@@ -84,6 +84,7 @@ public:
    bool HasProdInfo() const;
    FileSys::VirtualFile GetProdInfoRaw() const;
    void DecryptProdInfo(std::array<u8, 0x20> bis_key);
+    FileSys::VirtualFile GetDecryptedProdInfo() const;
    std::array<u8, 0x240> GetETicketExtendedKek() const;

 private:
@@ -3,8 +3,12 @@
 // Refer to the license.txt file included.

 #include <fmt/format.h>
+#include "common/file_util.h"
+#include "core/core.h"
 #include "core/file_sys/bis_factory.h"
+#include "core/file_sys/mode.h"
 #include "core/file_sys/registered_cache.h"
+#include "core/settings.h"

 namespace FileSys {

@@ -14,10 +18,22 @@ BISFactory::BISFactory(VirtualDir nand_root_, VirtualDir load_root_, VirtualDir
      sysnand_cache(std::make_unique<RegisteredCache>(
          GetOrCreateDirectoryRelative(nand_root, "/system/Contents/registered"))),
      usrnand_cache(std::make_unique<RegisteredCache>(
-          GetOrCreateDirectoryRelative(nand_root, "/user/Contents/registered"))) {}
+          GetOrCreateDirectoryRelative(nand_root, "/user/Contents/registered"))),
+      sysnand_placeholder(std::make_unique<PlaceholderCache>(
+          GetOrCreateDirectoryRelative(nand_root, "/system/Contents/placehld"))),
+      usrnand_placeholder(std::make_unique<PlaceholderCache>(
+          GetOrCreateDirectoryRelative(nand_root, "/user/Contents/placehld"))) {}

 BISFactory::~BISFactory() = default;

+VirtualDir BISFactory::GetSystemNANDContentDirectory() const {
+    return GetOrCreateDirectoryRelative(nand_root, "/system/Contents");
+}
+
+VirtualDir BISFactory::GetUserNANDContentDirectory() const {
+    return GetOrCreateDirectoryRelative(nand_root, "/user/Contents");
+}
+
 RegisteredCache* BISFactory::GetSystemNANDContents() const {
    return sysnand_cache.get();
 }
@@ -26,9 +42,17 @@ RegisteredCache* BISFactory::GetUserNANDContents() const {
    return usrnand_cache.get();
 }

+PlaceholderCache* BISFactory::GetSystemNANDPlaceholder() const {
+    return sysnand_placeholder.get();
+}
+
+PlaceholderCache* BISFactory::GetUserNANDPlaceholder() const {
+    return usrnand_placeholder.get();
+}
+
 VirtualDir BISFactory::GetModificationLoadRoot(u64 title_id) const {
    // LayeredFS doesn't work on updates and title id-less homebrew
-    if (title_id == 0 || (title_id & 0x800) > 0)
+    if (title_id == 0 || (title_id & 0xFFF) == 0x800)
        return nullptr;
    return GetOrCreateDirectoryRelative(load_root, fmt::format("/{:016X}", title_id));
 }
@@ -39,4 +63,77 @@ VirtualDir BISFactory::GetModificationDumpRoot(u64 title_id) const {
    return GetOrCreateDirectoryRelative(dump_root, fmt::format("/{:016X}", title_id));
 }

+VirtualDir BISFactory::OpenPartition(BisPartitionId id) const {
+    switch (id) {
+    case BisPartitionId::CalibrationFile:
+        return GetOrCreateDirectoryRelative(nand_root, "/prodinfof");
+    case BisPartitionId::SafeMode:
+        return GetOrCreateDirectoryRelative(nand_root, "/safe");
+    case BisPartitionId::System:
+        return GetOrCreateDirectoryRelative(nand_root, "/system");
+    case BisPartitionId::User:
+        return GetOrCreateDirectoryRelative(nand_root, "/user");
+    default:
+        return nullptr;
+    }
+}
+
+VirtualFile BISFactory::OpenPartitionStorage(BisPartitionId id) const {
+    Core::Crypto::KeyManager keys;
+    Core::Crypto::PartitionDataManager pdm{
+        Core::System::GetInstance().GetFilesystem()->OpenDirectory(
+            FileUtil::GetUserPath(FileUtil::UserPath::SysDataDir), Mode::Read)};
+    keys.PopulateFromPartitionData(pdm);
+
+    switch (id) {
+    case BisPartitionId::CalibrationBinary:
+        return pdm.GetDecryptedProdInfo();
+    case BisPartitionId::BootConfigAndPackage2Part1:
+    case BisPartitionId::BootConfigAndPackage2Part2:
+    case BisPartitionId::BootConfigAndPackage2Part3:
+    case BisPartitionId::BootConfigAndPackage2Part4:
+    case BisPartitionId::BootConfigAndPackage2Part5:
+    case BisPartitionId::BootConfigAndPackage2Part6: {
+        const auto new_id = static_cast<u8>(id) -
+                            static_cast<u8>(BisPartitionId::BootConfigAndPackage2Part1) +
+                            static_cast<u8>(Core::Crypto::Package2Type::NormalMain);
+        return pdm.GetPackage2Raw(static_cast<Core::Crypto::Package2Type>(new_id));
+    }
+    default:
+        return nullptr;
+    }
+}
+
+VirtualDir BISFactory::GetImageDirectory() const {
+    return GetOrCreateDirectoryRelative(nand_root, "/user/Album");
+}
+
+u64 BISFactory::GetSystemNANDFreeSpace() const {
+    const auto sys_dir = GetOrCreateDirectoryRelative(nand_root, "/system");
+    if (sys_dir == nullptr)
+        return 0;
+
+    return GetSystemNANDTotalSpace() - sys_dir->GetSize();
+}
+
+u64 BISFactory::GetSystemNANDTotalSpace() const {
+    return static_cast<u64>(Settings::values.nand_system_size);
+}
+
+u64 BISFactory::GetUserNANDFreeSpace() const {
+    const auto usr_dir = GetOrCreateDirectoryRelative(nand_root, "/user");
+    if (usr_dir == nullptr)
+        return 0;
+
+    return GetUserNANDTotalSpace() - usr_dir->GetSize();
+}
+
+u64 BISFactory::GetUserNANDTotalSpace() const {
+    return static_cast<u64>(Settings::values.nand_user_size);
+}
+
+u64 BISFactory::GetFullNANDTotalSpace() const {
+    return static_cast<u64>(Settings::values.nand_total_size);
+}
+
 } // namespace FileSys
@@ -10,7 +10,25 @@

 namespace FileSys {

+enum class BisPartitionId : u32 {
+    UserDataRoot = 20,
+    CalibrationBinary = 27,
+    CalibrationFile = 28,
+    BootConfigAndPackage2Part1 = 21,
+    BootConfigAndPackage2Part2 = 22,
+    BootConfigAndPackage2Part3 = 23,
+    BootConfigAndPackage2Part4 = 24,
+    BootConfigAndPackage2Part5 = 25,
+    BootConfigAndPackage2Part6 = 26,
+    SafeMode = 29,
+    System = 31,
+    SystemProperEncryption = 32,
+    SystemProperPartition = 33,
+    User = 30,
+};
+
 class RegisteredCache;
+class PlaceholderCache;

 /// File system interface to the Built-In Storage
 /// This is currently missing accessors to BIS partitions, but seemed like a good place for the NAND
@@ -20,12 +38,29 @@ public:
    explicit BISFactory(VirtualDir nand_root, VirtualDir load_root, VirtualDir dump_root);
    ~BISFactory();

+    VirtualDir GetSystemNANDContentDirectory() const;
+    VirtualDir GetUserNANDContentDirectory() const;
+
    RegisteredCache* GetSystemNANDContents() const;
    RegisteredCache* GetUserNANDContents() const;

+    PlaceholderCache* GetSystemNANDPlaceholder() const;
+    PlaceholderCache* GetUserNANDPlaceholder() const;
+
    VirtualDir GetModificationLoadRoot(u64 title_id) const;
    VirtualDir GetModificationDumpRoot(u64 title_id) const;

+    VirtualDir OpenPartition(BisPartitionId id) const;
+    VirtualFile OpenPartitionStorage(BisPartitionId id) const;
+
+    VirtualDir GetImageDirectory() const;
+
+    u64 GetSystemNANDFreeSpace() const;
+    u64 GetSystemNANDTotalSpace() const;
+    u64 GetUserNANDFreeSpace() const;
+    u64 GetUserNANDTotalSpace() const;
+    u64 GetFullNANDTotalSpace() const;
+
 private:
    VirtualDir nand_root;
    VirtualDir load_root;
@@ -33,6 +68,9 @@ private:

    std::unique_ptr<RegisteredCache> sysnand_cache;
    std::unique_ptr<RegisteredCache> usrnand_cache;
+
+    std::unique_ptr<PlaceholderCache> sysnand_placeholder;
+    std::unique_ptr<PlaceholderCache> usrnand_placeholder;
 };

 } // namespace FileSys
@@ -12,12 +12,16 @@
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/nca_metadata.h"
 #include "core/file_sys/partition_filesystem.h"
+#include "core/file_sys/romfs.h"
 #include "core/file_sys/submission_package.h"
+#include "core/file_sys/vfs_concat.h"
 #include "core/file_sys/vfs_offset.h"
+#include "core/file_sys/vfs_vector.h"
 #include "core/loader/loader.h"

 namespace FileSys {

+constexpr u64 GAMECARD_CERTIFICATE_OFFSET = 0x7000;
 constexpr std::array partition_names{
    "update",
    "normal",
@@ -175,6 +179,26 @@ VirtualDir XCI::GetParentDirectory() const {
    return file->GetContainingDirectory();
 }

+VirtualDir XCI::ConcatenatedPseudoDirectory() {
+    const auto out = std::make_shared<VectorVfsDirectory>();
+    for (const auto& part_id : {XCIPartition::Normal, XCIPartition::Logo, XCIPartition::Secure}) {
+        const auto& part = GetPartition(part_id);
+        if (part == nullptr)
+            continue;
+
+        for (const auto& file : part->GetFiles())
+            out->AddFile(file);
+    }
+
+    return out;
+}
+
+std::array<u8, 0x200> XCI::GetCertificate() const {
+    std::array<u8, 0x200> out;
+    file->Read(out.data(), out.size(), GAMECARD_CERTIFICATE_OFFSET);
+    return out;
+}
+
 Loader::ResultStatus XCI::AddNCAFromPartition(XCIPartition part) {
    const auto partition_index = static_cast<std::size_t>(part);
    const auto& partition = partitions[partition_index];
@@ -91,6 +91,8 @@ public:
    VirtualDir GetLogoPartition() const;

    u64 GetProgramTitleID() const;
+    u32 GetSystemUpdateVersion();
+    u64 GetSystemUpdateTitleID() const;

    bool HasProgramNCA() const;
    VirtualFile GetProgramNCAFile() const;
@@ -106,6 +108,11 @@ public:

    VirtualDir GetParentDirectory() const override;

+    // Creates a directory that contains all the NCAs in the gamecard
+    VirtualDir ConcatenatedPseudoDirectory();
+
+    std::array<u8, 0x200> GetCertificate() const;
+
 private:
    Loader::ResultStatus AddNCAFromPartition(XCIPartition part);

@@ -120,6 +127,8 @@ private:
    std::shared_ptr<NCA> program;
    std::vector<std::shared_ptr<NCA>> ncas;

+    u64 update_normal_partition_end;
+
    Core::Crypto::KeyManager keys;
 };
 } // namespace FileSys
@@ -1,492 +0,0 @@
-// Copyright 2018 yuzu emulator team
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#include <locale>
-#include "common/hex_util.h"
-#include "common/microprofile.h"
-#include "common/swap.h"
-#include "core/core.h"
-#include "core/core_timing.h"
-#include "core/core_timing_util.h"
-#include "core/file_sys/cheat_engine.h"
-#include "core/hle/kernel/process.h"
-#include "core/hle/service/hid/controllers/npad.h"
-#include "core/hle/service/hid/hid.h"
-#include "core/hle/service/sm/sm.h"
-
-namespace FileSys {
-
-constexpr s64 CHEAT_ENGINE_TICKS = static_cast<s64>(Core::Timing::BASE_CLOCK_RATE / 60);
-constexpr u32 KEYPAD_BITMASK = 0x3FFFFFF;
-
-u64 Cheat::Address() const {
-    u64 out;
-    std::memcpy(&out, raw.data(), sizeof(u64));
-    return Common::swap64(out) & 0xFFFFFFFFFF;
-}
-
-u64 Cheat::ValueWidth(u64 offset) const {
-    return Value(offset, width);
-}
-
-u64 Cheat::Value(u64 offset, u64 width) const {
-    u64 out;
-    std::memcpy(&out, raw.data() + offset, sizeof(u64));
-    out = Common::swap64(out);
-    if (width == 8)
-        return out;
-    return out & ((1ull << (width * CHAR_BIT)) - 1);
-}
-
-u32 Cheat::KeypadValue() const {
-    u32 out;
-    std::memcpy(&out, raw.data(), sizeof(u32));
-    return Common::swap32(out) & 0x0FFFFFFF;
-}
-
-void CheatList::SetMemoryParameters(VAddr main_begin, VAddr heap_begin, VAddr main_end,
-                                    VAddr heap_end, MemoryWriter writer, MemoryReader reader) {
-    this->main_region_begin = main_begin;
-    this->main_region_end = main_end;
-    this->heap_region_begin = heap_begin;
-    this->heap_region_end = heap_end;
-    this->writer = writer;
-    this->reader = reader;
-}
-
-MICROPROFILE_DEFINE(Cheat_Engine, "Add-Ons", "Cheat Engine", MP_RGB(70, 200, 70));
-
-void CheatList::Execute() {
-    MICROPROFILE_SCOPE(Cheat_Engine);
-
-    std::fill(scratch.begin(), scratch.end(), 0);
-    in_standard = false;
-    for (std::size_t i = 0; i < master_list.size(); ++i) {
-        LOG_DEBUG(Common_Filesystem, "Executing block #{:08X} ({})", i, master_list[i].first);
-        current_block = i;
-        ExecuteBlock(master_list[i].second);
-    }
-
-    in_standard = true;
-    for (std::size_t i = 0; i < standard_list.size(); ++i) {
-        LOG_DEBUG(Common_Filesystem, "Executing block #{:08X} ({})", i, standard_list[i].first);
-        current_block = i;
-        ExecuteBlock(standard_list[i].second);
-    }
-}
-
-CheatList::CheatList(const Core::System& system_, ProgramSegment master, ProgramSegment standard)
-    : master_list{std::move(master)}, standard_list{std::move(standard)}, system{&system_} {}
-
-bool CheatList::EvaluateConditional(const Cheat& cheat) const {
-    using ComparisonFunction = bool (*)(u64, u64);
-    constexpr std::array<ComparisonFunction, 6> comparison_functions{
-        [](u64 a, u64 b) { return a > b; },  [](u64 a, u64 b) { return a >= b; },
-        [](u64 a, u64 b) { return a < b; },  [](u64 a, u64 b) { return a <= b; },
-        [](u64 a, u64 b) { return a == b; }, [](u64 a, u64 b) { return a != b; },
-    };
-
-    if (cheat.type == CodeType::ConditionalInput) {
-        const auto applet_resource =
-            system->ServiceManager().GetService<Service::HID::Hid>("hid")->GetAppletResource();
-        if (applet_resource == nullptr) {
-            LOG_WARNING(
-                Common_Filesystem,
-                "Attempted to evaluate input conditional, but applet resource is not initialized!");
-            return false;
-        }
-
-        const auto press_state =
-            applet_resource
-                ->GetController<Service::HID::Controller_NPad>(Service::HID::HidController::NPad)
-                .GetAndResetPressState();
-        return ((press_state & cheat.KeypadValue()) & KEYPAD_BITMASK) != 0;
-    }
-
-    ASSERT(cheat.type == CodeType::Conditional);
-
-    const auto offset =
-        cheat.memory_type == MemoryType::MainNSO ? main_region_begin : heap_region_begin;
-    ASSERT(static_cast<u8>(cheat.comparison_op.Value()) < 6);
-    auto* function = comparison_functions[static_cast<u8>(cheat.comparison_op.Value())];
-    const auto addr = cheat.Address() + offset;
-
-    return function(reader(cheat.width, SanitizeAddress(addr)), cheat.ValueWidth(8));
-}
-
-void CheatList::ProcessBlockPairs(const Block& block) {
-    block_pairs.clear();
-
-    u64 scope = 0;
-    std::map<u64, u64> pairs;
-
-    for (std::size_t i = 0; i < block.size(); ++i) {
-        const auto& cheat = block[i];
-
-        switch (cheat.type) {
-        case CodeType::Conditional:
-        case CodeType::ConditionalInput:
-            pairs.insert_or_assign(scope, i);
-            ++scope;
-            break;
-        case CodeType::EndConditional: {
-            --scope;
-            const auto idx = pairs.at(scope);
-            block_pairs.insert_or_assign(idx, i);
-            break;
-        }
-        case CodeType::Loop: {
-            if (cheat.end_of_loop) {
-                --scope;
-                const auto idx = pairs.at(scope);
-                block_pairs.insert_or_assign(idx, i);
-            } else {
-                pairs.insert_or_assign(scope, i);
-                ++scope;
-            }
-            break;
-        }
-        }
-    }
-}
-
-void CheatList::WriteImmediate(const Cheat& cheat) {
-    const auto offset =
-        cheat.memory_type == MemoryType::MainNSO ? main_region_begin : heap_region_begin;
-    const auto& register_3 = scratch.at(cheat.register_3);
-
-    const auto addr = cheat.Address() + offset + register_3;
-    LOG_DEBUG(Common_Filesystem, "writing value={:016X} to addr={:016X}", addr,
-              cheat.Value(8, cheat.width));
-    writer(cheat.width, SanitizeAddress(addr), cheat.ValueWidth(8));
-}
-
-void CheatList::BeginConditional(const Cheat& cheat) {
-    if (EvaluateConditional(cheat)) {
-        return;
-    }
-
-    const auto iter = block_pairs.find(current_index);
-    ASSERT(iter != block_pairs.end());
-    current_index = iter->second - 1;
-}
-
-void CheatList::EndConditional(const Cheat& cheat) {
-    LOG_DEBUG(Common_Filesystem, "Ending conditional block.");
-}
-
-void CheatList::Loop(const Cheat& cheat) {
-    if (cheat.end_of_loop.Value())
-        ASSERT(!cheat.end_of_loop.Value());
-
-    auto& register_3 = scratch.at(cheat.register_3);
-    const auto iter = block_pairs.find(current_index);
-    ASSERT(iter != block_pairs.end());
-    ASSERT(iter->first < iter->second);
-
-    const s32 initial_value = static_cast<s32>(cheat.Value(4, sizeof(s32)));
-    for (s32 i = initial_value; i >= 0; --i) {
-        register_3 = static_cast<u64>(i);
-        for (std::size_t c = iter->first + 1; c < iter->second; ++c) {
-            current_index = c;
-            ExecuteSingleCheat(
-                (in_standard ? standard_list : master_list)[current_block].second[c]);
-        }
-    }
-
-    current_index = iter->second;
-}
-
-void CheatList::LoadImmediate(const Cheat& cheat) {
-    auto& register_3 = scratch.at(cheat.register_3);
-
-    LOG_DEBUG(Common_Filesystem, "setting register={:01X} equal to value={:016X}", cheat.register_3,
-              cheat.Value(4, 8));
-    register_3 = cheat.Value(4, 8);
-}
-
-void CheatList::LoadIndexed(const Cheat& cheat) {
-    const auto offset =
-        cheat.memory_type == MemoryType::MainNSO ? main_region_begin : heap_region_begin;
-    auto& register_3 = scratch.at(cheat.register_3);
-
-    const auto addr = (cheat.load_from_register.Value() ? register_3 : offset) + cheat.Address();
-    LOG_DEBUG(Common_Filesystem, "writing indexed value to register={:01X}, addr={:016X}",
-              cheat.register_3, addr);
-    register_3 = reader(cheat.width, SanitizeAddress(addr));
-}
-
-void CheatList::StoreIndexed(const Cheat& cheat) {
-    const auto& register_3 = scratch.at(cheat.register_3);
-
-    const auto addr =
-        register_3 + (cheat.add_additional_register.Value() ? scratch.at(cheat.register_6) : 0);
-    LOG_DEBUG(Common_Filesystem, "writing value={:016X} to addr={:016X}",
-              cheat.Value(4, cheat.width), addr);
-    writer(cheat.width, SanitizeAddress(addr), cheat.ValueWidth(4));
-}
-
-void CheatList::RegisterArithmetic(const Cheat& cheat) {
-    using ArithmeticFunction = u64 (*)(u64, u64);
-    constexpr std::array<ArithmeticFunction, 5> arithmetic_functions{
-        [](u64 a, u64 b) { return a + b; },  [](u64 a, u64 b) { return a - b; },
-        [](u64 a, u64 b) { return a * b; },  [](u64 a, u64 b) { return a << b; },
-        [](u64 a, u64 b) { return a >> b; },
-    };
-
-    using ArithmeticOverflowCheck = bool (*)(u64, u64);
-    constexpr std::array<ArithmeticOverflowCheck, 5> arithmetic_overflow_checks{
-        [](u64 a, u64 b) { return a > (std::numeric_limits<u64>::max() - b); },       // a + b
-        [](u64 a, u64 b) { return a > (std::numeric_limits<u64>::max() + b); },       // a - b
-        [](u64 a, u64 b) { return a > (std::numeric_limits<u64>::max() / b); },       // a * b
-        [](u64 a, u64 b) { return b >= 64 || (a & ~((1ull << (64 - b)) - 1)) != 0; }, // a << b
-        [](u64 a, u64 b) { return b >= 64 || (a & ((1ull << b) - 1)) != 0; },         // a >> b
-    };
-
-    static_assert(sizeof(arithmetic_functions) == sizeof(arithmetic_overflow_checks),
-                  "Missing or have extra arithmetic overflow checks compared to functions!");
-
-    auto& register_3 = scratch.at(cheat.register_3);
-
-    ASSERT(static_cast<u8>(cheat.arithmetic_op.Value()) < 5);
-    auto* function = arithmetic_functions[static_cast<u8>(cheat.arithmetic_op.Value())];
-    auto* overflow_function =
-        arithmetic_overflow_checks[static_cast<u8>(cheat.arithmetic_op.Value())];
-    LOG_DEBUG(Common_Filesystem, "performing arithmetic with register={:01X}, value={:016X}",
-              cheat.register_3, cheat.ValueWidth(4));
-
-    if (overflow_function(register_3, cheat.ValueWidth(4))) {
-        LOG_WARNING(Common_Filesystem,
-                    "overflow will occur when performing arithmetic operation={:02X} with operands "
-                    "a={:016X}, b={:016X}!",
-                    static_cast<u8>(cheat.arithmetic_op.Value()), register_3, cheat.ValueWidth(4));
-    }
-
-    register_3 = function(register_3, cheat.ValueWidth(4));
-}
-
-void CheatList::BeginConditionalInput(const Cheat& cheat) {
-    if (EvaluateConditional(cheat))
-        return;
-
-    const auto iter = block_pairs.find(current_index);
-    ASSERT(iter != block_pairs.end());
-    current_index = iter->second - 1;
-}
-
-VAddr CheatList::SanitizeAddress(VAddr in) const {
-    if ((in < main_region_begin || in >= main_region_end) &&
-        (in < heap_region_begin || in >= heap_region_end)) {
-        LOG_ERROR(Common_Filesystem,
-                  "Cheat attempting to access memory at invalid address={:016X}, if this persists, "
-                  "the cheat may be incorrect. However, this may be normal early in execution if "
-                  "the game has not properly set up yet.",
-                  in);
-        return 0; ///< Invalid addresses will hard crash
-    }
-
-    return in;
-}
-
-void CheatList::ExecuteSingleCheat(const Cheat& cheat) {
-    using CheatOperationFunction = void (CheatList::*)(const Cheat&);
-    constexpr std::array<CheatOperationFunction, 9> cheat_operation_functions{
-        &CheatList::WriteImmediate,        &CheatList::BeginConditional,
-        &CheatList::EndConditional,        &CheatList::Loop,
-        &CheatList::LoadImmediate,         &CheatList::LoadIndexed,
-        &CheatList::StoreIndexed,          &CheatList::RegisterArithmetic,
-        &CheatList::BeginConditionalInput,
-    };
-
-    const auto index = static_cast<u8>(cheat.type.Value());
-    ASSERT(index < sizeof(cheat_operation_functions));
-    const auto op = cheat_operation_functions[index];
-    (this->*op)(cheat);
-}
-
-void CheatList::ExecuteBlock(const Block& block) {
-    encountered_loops.clear();
-
-    ProcessBlockPairs(block);
-    for (std::size_t i = 0; i < block.size(); ++i) {
-        current_index = i;
-        ExecuteSingleCheat(block[i]);
-        i = current_index;
-    }
-}
-
-CheatParser::~CheatParser() = default;
-
-CheatList CheatParser::MakeCheatList(const Core::System& system, CheatList::ProgramSegment master,
-                                     CheatList::ProgramSegment standard) const {
-    return {system, std::move(master), std::move(standard)};
-}
-
-TextCheatParser::~TextCheatParser() = default;
-
-CheatList TextCheatParser::Parse(const Core::System& system, const std::vector<u8>& data) const {
-    std::stringstream ss;
-    ss.write(reinterpret_cast<const char*>(data.data()), data.size());
-
-    std::vector<std::string> lines;
-    std::string stream_line;
-    while (std::getline(ss, stream_line)) {
-        // Remove a trailing \r
-        if (!stream_line.empty() && stream_line.back() == '\r')
-            stream_line.pop_back();
-        lines.push_back(std::move(stream_line));
-    }
-
-    CheatList::ProgramSegment master_list;
-    CheatList::ProgramSegment standard_list;
-
-    for (std::size_t i = 0; i < lines.size(); ++i) {
-        auto line = lines[i];
-
-        if (!line.empty() && (line[0] == '[' || line[0] == '{')) {
-            const auto master = line[0] == '{';
-            const auto begin = master ? line.find('{') : line.find('[');
-            const auto end = master ? line.rfind('}') : line.rfind(']');
-
-            ASSERT(begin != std::string::npos && end != std::string::npos);
-
-            const std::string patch_name{line.begin() + begin + 1, line.begin() + end};
-            CheatList::Block block{};
-
-            while (i < lines.size() - 1) {
-                line = lines[++i];
-                if (!line.empty() && (line[0] == '[' || line[0] == '{')) {
-                    --i;
-                    break;
-                }
-
-                if (line.size() < 8)
-                    continue;
-
-                Cheat out{};
-                out.raw = ParseSingleLineCheat(line);
-                block.push_back(out);
-            }
-
-            (master ? master_list : standard_list).emplace_back(patch_name, block);
-        }
-    }
-
-    return MakeCheatList(system, master_list, standard_list);
-}
-
-std::array<u8, 16> TextCheatParser::ParseSingleLineCheat(const std::string& line) const {
-    std::array<u8, 16> out{};
-
-    if (line.size() < 8)
-        return out;
-
-    const auto word1 = Common::HexStringToArray<sizeof(u32)>(std::string_view{line.data(), 8});
-    std::memcpy(out.data(), word1.data(), sizeof(u32));
-
-    if (line.size() < 17 || line[8] != ' ')
-        return out;
-
-    const auto word2 = Common::HexStringToArray<sizeof(u32)>(std::string_view{line.data() + 9, 8});
-    std::memcpy(out.data() + sizeof(u32), word2.data(), sizeof(u32));
-
-    if (line.size() < 26 || line[17] != ' ') {
-        // Perform shifting in case value is truncated early.
-        const auto type = static_cast<CodeType>((out[0] & 0xF0) >> 4);
-        if (type == CodeType::Loop || type == CodeType::LoadImmediate ||
-            type == CodeType::StoreIndexed || type == CodeType::RegisterArithmetic) {
-            std::memcpy(out.data() + 8, out.data() + 4, sizeof(u32));
-            std::memset(out.data() + 4, 0, sizeof(u32));
-        }
-
-        return out;
-    }
-
-    const auto word3 = Common::HexStringToArray<sizeof(u32)>(std::string_view{line.data() + 18, 8});
-    std::memcpy(out.data() + 2 * sizeof(u32), word3.data(), sizeof(u32));
-
-    if (line.size() < 35 || line[26] != ' ') {
-        // Perform shifting in case value is truncated early.
-        const auto type = static_cast<CodeType>((out[0] & 0xF0) >> 4);
-        if (type == CodeType::WriteImmediate || type == CodeType::Conditional) {
-            std::memcpy(out.data() + 12, out.data() + 8, sizeof(u32));
-            std::memset(out.data() + 8, 0, sizeof(u32));
-        }
-
-        return out;
-    }
-
-    const auto word4 = Common::HexStringToArray<sizeof(u32)>(std::string_view{line.data() + 27, 8});
-    std::memcpy(out.data() + 3 * sizeof(u32), word4.data(), sizeof(u32));
-
-    return out;
-}
-
-namespace {
-u64 MemoryReadImpl(u32 width, VAddr addr) {
-    switch (width) {
-    case 1:
-        return Memory::Read8(addr);
-    case 2:
-        return Memory::Read16(addr);
-    case 4:
-        return Memory::Read32(addr);
-    case 8:
-        return Memory::Read64(addr);
-    default:
-        UNREACHABLE();
-        return 0;
-    }
-}
-
-void MemoryWriteImpl(u32 width, VAddr addr, u64 value) {
-    switch (width) {
-    case 1:
-        Memory::Write8(addr, static_cast<u8>(value));
-        break;
-    case 2:
-        Memory::Write16(addr, static_cast<u16>(value));
-        break;
-    case 4:
-        Memory::Write32(addr, static_cast<u32>(value));
-        break;
-    case 8:
-        Memory::Write64(addr, value);
-        break;
-    default:
-        UNREACHABLE();
-    }
-}
-} // Anonymous namespace
-
-CheatEngine::CheatEngine(Core::System& system, std::vector<CheatList> cheats_,
-                         const std::string& build_id, VAddr code_region_start,
-                         VAddr code_region_end)
-    : cheats{std::move(cheats_)}, core_timing{system.CoreTiming()} {
-    event = core_timing.RegisterEvent(
-        "CheatEngine::FrameCallback::" + build_id,
-        [this](u64 userdata, s64 cycles_late) { FrameCallback(userdata, cycles_late); });
-    core_timing.ScheduleEvent(CHEAT_ENGINE_TICKS, event);
-
-    const auto& vm_manager = system.CurrentProcess()->VMManager();
-    for (auto& list : this->cheats) {
-        list.SetMemoryParameters(code_region_start, vm_manager.GetHeapRegionBaseAddress(),
-                                 code_region_end, vm_manager.GetHeapRegionEndAddress(),
-                                 &MemoryWriteImpl, &MemoryReadImpl);
-    }
-}
-
-CheatEngine::~CheatEngine() {
-    core_timing.UnscheduleEvent(event, 0);
-}
-
-void CheatEngine::FrameCallback(u64 userdata, s64 cycles_late) {
-    for (auto& list : cheats) {
-        list.Execute();
-    }
-
-    core_timing.ScheduleEvent(CHEAT_ENGINE_TICKS - cycles_late, event);
-}
-
-} // namespace FileSys
@@ -1,234 +0,0 @@
-// Copyright 2018 yuzu emulator team
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <map>
-#include <set>
-#include <vector>
-#include "common/bit_field.h"
-#include "common/common_types.h"
-
-namespace Core {
-class System;
-}
-
-namespace Core::Timing {
-class CoreTiming;
-struct EventType;
-} // namespace Core::Timing
-
-namespace FileSys {
-
-enum class CodeType : u32 {
-    // 0TMR00AA AAAAAAAA YYYYYYYY YYYYYYYY
-    // Writes a T sized value Y to the address A added to the value of register R in memory domain M
-    WriteImmediate = 0,
-
-    // 1TMC00AA AAAAAAAA YYYYYYYY YYYYYYYY
-    // Compares the T sized value Y to the value at address A in memory domain M using the
-    // conditional function C. If success, continues execution. If failure, jumps to the matching
-    // EndConditional statement.
-    Conditional = 1,
-
-    // 20000000
-    // Terminates a Conditional or ConditionalInput block.
-    EndConditional = 2,
-
-    // 300R0000 VVVVVVVV
-    // Starts looping V times, storing the current count in register R.
-    // Loop block is terminated with a matching 310R0000.
-    Loop = 3,
-
-    // 400R0000 VVVVVVVV VVVVVVVV
-    // Sets the value of register R to the value V.
-    LoadImmediate = 4,
-
-    // 5TMRI0AA AAAAAAAA
-    // Sets the value of register R to the value of width T at address A in memory domain M, with
-    // the current value of R added to the address if I == 1.
-    LoadIndexed = 5,
-
-    // 6T0RIFG0 VVVVVVVV VVVVVVVV
-    // Writes the value V of width T to the memory address stored in register R. Adds the value of
-    // register G to the final calculation if F is nonzero. Increments the value of register R by T
-    // after operation if I is nonzero.
-    StoreIndexed = 6,
-
-    // 7T0RA000 VVVVVVVV
-    // Performs the arithmetic operation A on the value in register R and the value V of width T,
-    // storing the result in register R.
-    RegisterArithmetic = 7,
-
-    // 8KKKKKKK
-    // Checks to see if any of the buttons defined by the bitmask K are pressed. If any are,
-    // execution continues. If none are, execution skips to the next EndConditional command.
-    ConditionalInput = 8,
-};
-
-enum class MemoryType : u32 {
-    // Addressed relative to start of main NSO
-    MainNSO = 0,
-
-    // Addressed relative to start of heap
-    Heap = 1,
-};
-
-enum class ArithmeticOp : u32 {
-    Add = 0,
-    Sub = 1,
-    Mult = 2,
-    LShift = 3,
-    RShift = 4,
-};
-
-enum class ComparisonOp : u32 {
-    GreaterThan = 1,
-    GreaterThanEqual = 2,
-    LessThan = 3,
-    LessThanEqual = 4,
-    Equal = 5,
-    Inequal = 6,
-};
-
-union Cheat {
-    std::array<u8, 16> raw;
-
-    BitField<4, 4, CodeType> type;
-    BitField<0, 4, u32> width; // Can be 1, 2, 4, or 8. Measured in bytes.
-    BitField<0, 4, u32> end_of_loop;
-    BitField<12, 4, MemoryType> memory_type;
-    BitField<8, 4, u32> register_3;
-    BitField<8, 4, ComparisonOp> comparison_op;
-    BitField<20, 4, u32> load_from_register;
-    BitField<20, 4, u32> increment_register;
-    BitField<20, 4, ArithmeticOp> arithmetic_op;
-    BitField<16, 4, u32> add_additional_register;
-    BitField<28, 4, u32> register_6;
-
-    u64 Address() const;
-    u64 ValueWidth(u64 offset) const;
-    u64 Value(u64 offset, u64 width) const;
-    u32 KeypadValue() const;
-};
-
-class CheatParser;
-
-// Represents a full collection of cheats for a game. The Execute function should be called every
-// interval that all cheats should be executed. Clients should not directly instantiate this class
-// (hence private constructor), they should instead receive an instance from CheatParser, which
-// guarantees the list is always in an acceptable state.
-class CheatList {
-public:
-    friend class CheatParser;
-
-    using Block = std::vector<Cheat>;
-    using ProgramSegment = std::vector<std::pair<std::string, Block>>;
-
-    // (width in bytes, address, value)
-    using MemoryWriter = void (*)(u32, VAddr, u64);
-    // (width in bytes, address) -> value
-    using MemoryReader = u64 (*)(u32, VAddr);
-
-    void SetMemoryParameters(VAddr main_begin, VAddr heap_begin, VAddr main_end, VAddr heap_end,
-                             MemoryWriter writer, MemoryReader reader);
-
-    void Execute();
-
-private:
-    CheatList(const Core::System& system_, ProgramSegment master, ProgramSegment standard);
-
-    void ProcessBlockPairs(const Block& block);
-    void ExecuteSingleCheat(const Cheat& cheat);
-
-    void ExecuteBlock(const Block& block);
-
-    bool EvaluateConditional(const Cheat& cheat) const;
-
-    // Individual cheat operations
-    void WriteImmediate(const Cheat& cheat);
-    void BeginConditional(const Cheat& cheat);
-    void EndConditional(const Cheat& cheat);
-    void Loop(const Cheat& cheat);
-    void LoadImmediate(const Cheat& cheat);
-    void LoadIndexed(const Cheat& cheat);
-    void StoreIndexed(const Cheat& cheat);
-    void RegisterArithmetic(const Cheat& cheat);
-    void BeginConditionalInput(const Cheat& cheat);
-
-    VAddr SanitizeAddress(VAddr in) const;
-
-    // Master Codes are defined as codes that cannot be disabled and are run prior to all
-    // others.
-    ProgramSegment master_list;
-    // All other codes
-    ProgramSegment standard_list;
-
-    bool in_standard = false;
-
-    // 16 (0x0-0xF) scratch registers that can be used by cheats
-    std::array<u64, 16> scratch{};
-
-    MemoryWriter writer = nullptr;
-    MemoryReader reader = nullptr;
-
-    u64 main_region_begin{};
-    u64 heap_region_begin{};
-    u64 main_region_end{};
-    u64 heap_region_end{};
-
-    u64 current_block{};
-    // The current index of the cheat within the current Block
-    u64 current_index{};
-
-    // The 'stack' of the program. When a conditional or loop statement is encountered, its index is
-    // pushed onto this queue. When a end block is encountered, the condition is checked.
-    std::map<u64, u64> block_pairs;
-
-    std::set<u64> encountered_loops;
-
-    const Core::System* system;
-};
-
-// Intermediary class that parses a text file or other disk format for storing cheats into a
-// CheatList object, that can be used for execution.
-class CheatParser {
-public:
-    virtual ~CheatParser();
-
-    virtual CheatList Parse(const Core::System& system, const std::vector<u8>& data) const = 0;
-
-protected:
-    CheatList MakeCheatList(const Core::System& system_, CheatList::ProgramSegment master,
-                            CheatList::ProgramSegment standard) const;
-};
-
-// CheatParser implementation that parses text files
-class TextCheatParser final : public CheatParser {
-public:
-    ~TextCheatParser() override;
-
-    CheatList Parse(const Core::System& system, const std::vector<u8>& data) const override;
-
-private:
-    std::array<u8, 16> ParseSingleLineCheat(const std::string& line) const;
-};
-
-// Class that encapsulates a CheatList and manages its interaction with memory and CoreTiming
-class CheatEngine final {
-public:
-    CheatEngine(Core::System& system_, std::vector<CheatList> cheats_, const std::string& build_id,
-                VAddr code_region_start, VAddr code_region_end);
-    ~CheatEngine();
-
-private:
-    void FrameCallback(u64 userdata, s64 cycles_late);
-
-    std::vector<CheatList> cheats;
-
-    Core::Timing::EventType* event;
-    Core::Timing::CoreTiming& core_timing;
-};
-
-} // namespace FileSys
@@ -528,6 +528,14 @@ u64 NCA::GetTitleId() const {
    return header.title_id;
 }

+std::array<u8, 16> NCA::GetRightsId() const {
+    return header.rights_id;
+}
+
+u32 NCA::GetSDKVersion() const {
+    return header.sdk_version;
+}
+
 bool NCA::IsUpdate() const {
    return is_update;
 }
@@ -112,6 +112,8 @@ public:

    NCAContentType GetType() const;
    u64 GetTitleId() const;
+    std::array<u8, 0x10> GetRightsId() const;
+    u32 GetSDKVersion() const;
    bool IsUpdate() const;

    VirtualFile GetRomFS() const;
@@ -22,6 +22,7 @@
 #include "core/hle/service/filesystem/filesystem.h"
 #include "core/loader/loader.h"
 #include "core/loader/nso.h"
+#include "core/memory/cheat_engine.h"
 #include "core/settings.h"

 namespace FileSys {
@@ -63,7 +64,8 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {

    if (Settings::values.dump_exefs) {
        LOG_INFO(Loader, "Dumping ExeFS for title_id={:016X}", title_id);
-        const auto dump_dir = Service::FileSystem::GetModificationDumpRoot(title_id);
+        const auto dump_dir =
+            Core::System::GetInstance().GetFileSystemController().GetModificationDumpRoot(title_id);
        if (dump_dir != nullptr) {
            const auto exefs_dir = GetOrCreateDirectoryRelative(dump_dir, "/exefs");
            VfsRawCopyD(exefs, exefs_dir);
@@ -88,7 +90,8 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const {
    }

    // LayeredExeFS
-    const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
+    const auto load_dir =
+        Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id);
    if (load_dir != nullptr && load_dir->GetSize() > 0) {
        auto patch_dirs = load_dir->GetSubdirectories();
        std::sort(
@@ -174,7 +177,8 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso, const std::st
    if (Settings::values.dump_nso) {
        LOG_INFO(Loader, "Dumping NSO for name={}, build_id={}, title_id={:016X}", name, build_id,
                 title_id);
-        const auto dump_dir = Service::FileSystem::GetModificationDumpRoot(title_id);
+        const auto dump_dir =
+            Core::System::GetInstance().GetFileSystemController().GetModificationDumpRoot(title_id);
        if (dump_dir != nullptr) {
            const auto nso_dir = GetOrCreateDirectoryRelative(dump_dir, "/nso");
            const auto file = nso_dir->CreateFile(fmt::format("{}-{}.nso", name, build_id));
@@ -186,7 +190,13 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso, const std::st

    LOG_INFO(Loader, "Patching NSO for name={}, build_id={}", name, build_id);

-    const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
+    const auto load_dir =
+        Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id);
+    if (load_dir == nullptr) {
+        LOG_ERROR(Loader, "Cannot load mods for invalid title_id={:016X}", title_id);
+        return nso;
+    }
+
    auto patch_dirs = load_dir->GetSubdirectories();
    std::sort(patch_dirs.begin(), patch_dirs.end(),
              [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });
@@ -224,7 +234,13 @@ bool PatchManager::HasNSOPatch(const std::array<u8, 32>& build_id_) const {

    LOG_INFO(Loader, "Querying NSO patch existence for build_id={}", build_id);

-    const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
+    const auto load_dir =
+        Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id);
+    if (load_dir == nullptr) {
+        LOG_ERROR(Loader, "Cannot load mods for invalid title_id={:016X}", title_id);
+        return false;
+    }
+
    auto patch_dirs = load_dir->GetSubdirectories();
    std::sort(patch_dirs.begin(), patch_dirs.end(),
              [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });
@@ -232,9 +248,10 @@ bool PatchManager::HasNSOPatch(const std::array<u8, 32>& build_id_) const {
    return !CollectPatches(patch_dirs, build_id).empty();
 }

-static std::optional<CheatList> ReadCheatFileFromFolder(const Core::System& system, u64 title_id,
-                                                        const std::array<u8, 0x20>& build_id_,
-                                                        const VirtualDir& base_path, bool upper) {
+namespace {
+std::optional<std::vector<Memory::CheatEntry>> ReadCheatFileFromFolder(
+    const Core::System& system, u64 title_id, const std::array<u8, 0x20>& build_id_,
+    const VirtualDir& base_path, bool upper) {
    const auto build_id_raw = Common::HexToString(build_id_, upper);
    const auto build_id = build_id_raw.substr(0, sizeof(u64) * 2);
    const auto file = base_path->GetFile(fmt::format("{}.txt", build_id));
@@ -252,31 +269,39 @@ static std::optional<CheatList> ReadCheatFileFromFolder(const Core::System& syst
        return std::nullopt;
    }

-    TextCheatParser parser;
-    return parser.Parse(system, data);
+    Memory::TextCheatParser parser;
+    return parser.Parse(
+        system, std::string_view(reinterpret_cast<const char* const>(data.data()), data.size()));
 }

-std::vector<CheatList> PatchManager::CreateCheatList(const Core::System& system,
-                                                     const std::array<u8, 32>& build_id_) const {
-    const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
+} // Anonymous namespace
+
+std::vector<Memory::CheatEntry> PatchManager::CreateCheatList(
+    const Core::System& system, const std::array<u8, 32>& build_id_) const {
+    const auto load_dir = system.GetFileSystemController().GetModificationLoadRoot(title_id);
+    if (load_dir == nullptr) {
+        LOG_ERROR(Loader, "Cannot load mods for invalid title_id={:016X}", title_id);
+        return {};
+    }
+
    auto patch_dirs = load_dir->GetSubdirectories();
    std::sort(patch_dirs.begin(), patch_dirs.end(),
              [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });

-    std::vector<CheatList> out;
-    out.reserve(patch_dirs.size());
+    std::vector<Memory::CheatEntry> out;
    for (const auto& subdir : patch_dirs) {
        auto cheats_dir = subdir->GetSubdirectory("cheats");
        if (cheats_dir != nullptr) {
            auto res = ReadCheatFileFromFolder(system, title_id, build_id_, cheats_dir, true);
            if (res.has_value()) {
-                out.push_back(std::move(*res));
+                std::copy(res->begin(), res->end(), std::back_inserter(out));
                continue;
            }

            res = ReadCheatFileFromFolder(system, title_id, build_id_, cheats_dir, false);
-            if (res.has_value())
-                out.push_back(std::move(*res));
+            if (res.has_value()) {
+                std::copy(res->begin(), res->end(), std::back_inserter(out));
+            }
        }
    }

@@ -284,7 +309,8 @@ std::vector<CheatList> PatchManager::CreateCheatList(const Core::System& system,
 }

 static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType type) {
-    const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
+    const auto load_dir =
+        Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id);
    if ((type != ContentRecordType::Program && type != ContentRecordType::Data) ||
        load_dir == nullptr || load_dir->GetSize() <= 0) {
        return;
@@ -393,6 +419,8 @@ static bool IsDirValidAndNonEmpty(const VirtualDir& dir) {

 std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNames(
    VirtualFile update_raw) const {
+    if (title_id == 0)
+        return {};
    std::map<std::string, std::string, std::less<>> out;
    const auto& installed = Core::System::GetInstance().GetContentProvider();
    const auto& disabled = Settings::values.disabled_addons[title_id];
@@ -423,7 +451,8 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
    }

    // General Mods (LayeredFS and IPS)
-    const auto mod_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
+    const auto mod_dir =
+        Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id);
    if (mod_dir != nullptr && mod_dir->GetSize() > 0) {
        for (const auto& mod : mod_dir->GetSubdirectories()) {
            std::string types;
@@ -8,9 +8,9 @@
 #include <memory>
 #include <string>
 #include "common/common_types.h"
-#include "core/file_sys/cheat_engine.h"
 #include "core/file_sys/nca_metadata.h"
 #include "core/file_sys/vfs.h"
+#include "core/memory/dmnt_cheat_types.h"

 namespace Core {
 class System;
@@ -51,8 +51,8 @@ public:
    bool HasNSOPatch(const std::array<u8, 0x20>& build_id) const;

    // Creates a CheatList object with all
-    std::vector<CheatList> CreateCheatList(const Core::System& system,
-                                           const std::array<u8, 0x20>& build_id) const;
+    std::vector<Memory::CheatEntry> CreateCheatList(const Core::System& system,
+                                                    const std::array<u8, 0x20>& build_id) const;

    // Currently tracked RomFS patches:
    // - Game Updates
@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.

 #include <algorithm>
+#include <random>
 #include <regex>
 #include <mbedtls/sha256.h>
 #include "common/assert.h"
@@ -48,18 +49,21 @@ static bool FollowsTwoDigitDirFormat(std::string_view name) {
 static bool FollowsNcaIdFormat(std::string_view name) {
    static const std::regex nca_id_regex("[0-9A-F]{32}\\.nca", std::regex_constants::ECMAScript |
                                                                   std::regex_constants::icase);
-    return name.size() == 36 && std::regex_match(name.begin(), name.end(), nca_id_regex);
+    static const std::regex nca_id_cnmt_regex(
+        "[0-9A-F]{32}\\.cnmt.nca", std::regex_constants::ECMAScript | std::regex_constants::icase);
+    return (name.size() == 36 && std::regex_match(name.begin(), name.end(), nca_id_regex)) ||
+           (name.size() == 41 && std::regex_match(name.begin(), name.end(), nca_id_cnmt_regex));
 }

 static std::string GetRelativePathFromNcaID(const std::array<u8, 16>& nca_id, bool second_hex_upper,
-                                            bool within_two_digit) {
-    if (!within_two_digit) {
-        return fmt::format("/{}.nca", Common::HexToString(nca_id, second_hex_upper));
-    }
+                                            bool within_two_digit, bool cnmt_suffix) {
+    if (!within_two_digit)
+        return fmt::format(cnmt_suffix ? "{}.cnmt.nca" : "/{}.nca",
+                           Common::HexToString(nca_id, second_hex_upper));

    Core::Crypto::SHA256Hash hash{};
    mbedtls_sha256(nca_id.data(), nca_id.size(), hash.data(), 0);
-    return fmt::format("/000000{:02X}/{}.nca", hash[0],
+    return fmt::format(cnmt_suffix ? "/000000{:02X}/{}.cnmt.nca" : "/000000{:02X}/{}.nca", hash[0],
                       Common::HexToString(nca_id, second_hex_upper));
 }

@@ -127,6 +131,156 @@ std::vector<ContentProviderEntry> ContentProvider::ListEntries() const {
    return ListEntriesFilter(std::nullopt, std::nullopt, std::nullopt);
 }

+PlaceholderCache::PlaceholderCache(VirtualDir dir_) : dir(std::move(dir_)) {}
+
+bool PlaceholderCache::Create(const NcaID& id, u64 size) const {
+    const auto path = GetRelativePathFromNcaID(id, false, true, false);
+
+    if (dir->GetFileRelative(path) != nullptr) {
+        return false;
+    }
+
+    Core::Crypto::SHA256Hash hash{};
+    mbedtls_sha256(id.data(), id.size(), hash.data(), 0);
+    const auto dirname = fmt::format("000000{:02X}", hash[0]);
+
+    const auto dir2 = GetOrCreateDirectoryRelative(dir, dirname);
+
+    if (dir2 == nullptr)
+        return false;
+
+    const auto file = dir2->CreateFile(fmt::format("{}.nca", Common::HexToString(id, false)));
+
+    if (file == nullptr)
+        return false;
+
+    return file->Resize(size);
+}
+
+bool PlaceholderCache::Delete(const NcaID& id) const {
+    const auto path = GetRelativePathFromNcaID(id, false, true, false);
+
+    if (dir->GetFileRelative(path) == nullptr) {
+        return false;
+    }
+
+    Core::Crypto::SHA256Hash hash{};
+    mbedtls_sha256(id.data(), id.size(), hash.data(), 0);
+    const auto dirname = fmt::format("000000{:02X}", hash[0]);
+
+    const auto dir2 = GetOrCreateDirectoryRelative(dir, dirname);
+
+    const auto res = dir2->DeleteFile(fmt::format("{}.nca", Common::HexToString(id, false)));
+
+    return res;
+}
+
+bool PlaceholderCache::Exists(const NcaID& id) const {
+    const auto path = GetRelativePathFromNcaID(id, false, true, false);
+
+    return dir->GetFileRelative(path) != nullptr;
+}
+
+bool PlaceholderCache::Write(const NcaID& id, u64 offset, const std::vector<u8>& data) const {
+    const auto path = GetRelativePathFromNcaID(id, false, true, false);
+    const auto file = dir->GetFileRelative(path);
+
+    if (file == nullptr)
+        return false;
+
+    return file->WriteBytes(data, offset) == data.size();
+}
+
+bool PlaceholderCache::Register(RegisteredCache* cache, const NcaID& placeholder,
+                                const NcaID& install) const {
+    const auto path = GetRelativePathFromNcaID(placeholder, false, true, false);
+    const auto file = dir->GetFileRelative(path);
+
+    if (file == nullptr)
+        return false;
+
+    const auto res = cache->RawInstallNCA(NCA{file}, &VfsRawCopy, false, install);
+
+    if (res != InstallResult::Success)
+        return false;
+
+    return Delete(placeholder);
+}
+
+bool PlaceholderCache::CleanAll() const {
+    return dir->GetParentDirectory()->CleanSubdirectoryRecursive(dir->GetName());
+}
+
+std::optional<std::array<u8, 0x10>> PlaceholderCache::GetRightsID(const NcaID& id) const {
+    const auto path = GetRelativePathFromNcaID(id, false, true, false);
+    const auto file = dir->GetFileRelative(path);
+
+    if (file == nullptr)
+        return std::nullopt;
+
+    NCA nca{file};
+
+    if (nca.GetStatus() != Loader::ResultStatus::Success &&
+        nca.GetStatus() != Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) {
+        return std::nullopt;
+    }
+
+    const auto rights_id = nca.GetRightsId();
+    if (rights_id == NcaID{})
+        return std::nullopt;
+
+    return rights_id;
+}
+
+u64 PlaceholderCache::Size(const NcaID& id) const {
+    const auto path = GetRelativePathFromNcaID(id, false, true, false);
+    const auto file = dir->GetFileRelative(path);
+
+    if (file == nullptr)
+        return 0;
+
+    return file->GetSize();
+}
+
+bool PlaceholderCache::SetSize(const NcaID& id, u64 new_size) const {
+    const auto path = GetRelativePathFromNcaID(id, false, true, false);
+    const auto file = dir->GetFileRelative(path);
+
+    if (file == nullptr)
+        return false;
+
+    return file->Resize(new_size);
+}
+
+std::vector<NcaID> PlaceholderCache::List() const {
+    std::vector<NcaID> out;
+    for (const auto& sdir : dir->GetSubdirectories()) {
+        for (const auto& file : sdir->GetFiles()) {
+            const auto name = file->GetName();
+            if (name.length() == 36 && name[32] == '.' && name[33] == 'n' && name[34] == 'c' &&
+                name[35] == 'a') {
+                out.push_back(Common::HexStringToArray<0x10>(name.substr(0, 32)));
+            }
+        }
+    }
+    return out;
+}
+
+NcaID PlaceholderCache::Generate() {
+    std::random_device device;
+    std::mt19937 gen(device());
+    std::uniform_int_distribution<u64> distribution(1, std::numeric_limits<u64>::max());
+
+    NcaID out{};
+
+    const auto v1 = distribution(gen);
+    const auto v2 = distribution(gen);
+    std::memcpy(out.data(), &v1, sizeof(u64));
+    std::memcpy(out.data() + sizeof(u64), &v2, sizeof(u64));
+
+    return out;
+}
+
 VirtualFile RegisteredCache::OpenFileOrDirectoryConcat(const VirtualDir& dir,
                                                       std::string_view path) const {
    const auto file = dir->GetFileRelative(path);
@@ -169,14 +323,18 @@ VirtualFile RegisteredCache::OpenFileOrDirectoryConcat(const VirtualDir& dir,

 VirtualFile RegisteredCache::GetFileAtID(NcaID id) const {
    VirtualFile file;
-    // Try all four modes of file storage:
-    // (bit 1 = uppercase/lower, bit 0 = within a two-digit dir)
-    // 00: /000000**/{:032X}.nca
-    // 01: /{:032X}.nca
-    // 10: /000000**/{:032x}.nca
-    // 11: /{:032x}.nca
-    for (u8 i = 0; i < 4; ++i) {
-        const auto path = GetRelativePathFromNcaID(id, (i & 0b10) == 0, (i & 0b01) == 0);
+    // Try all five relevant modes of file storage:
+    // (bit 2 = uppercase/lower, bit 1 = within a two-digit dir, bit 0 = .cnmt suffix)
+    // 000: /000000**/{:032X}.nca
+    // 010: /{:032X}.nca
+    // 100: /000000**/{:032x}.nca
+    // 110: /{:032x}.nca
+    // 111: /{:032x}.cnmt.nca
+    for (u8 i = 0; i < 8; ++i) {
+        if ((i % 2) == 1 && i != 7)
+            continue;
+        const auto path =
+            GetRelativePathFromNcaID(id, (i & 0b100) == 0, (i & 0b010) == 0, (i & 0b001) == 0b001);
        file = OpenFileOrDirectoryConcat(dir, path);
        if (file != nullptr)
            return file;
@@ -472,7 +630,7 @@ InstallResult RegisteredCache::RawInstallNCA(const NCA& nca, const VfsCopyFuncti
        memcpy(id.data(), hash.data(), 16);
    }

-    std::string path = GetRelativePathFromNcaID(id, false, true);
+    std::string path = GetRelativePathFromNcaID(id, false, true, false);

    if (GetFileAtID(id) != nullptr && !overwrite_if_exists) {
        LOG_WARNING(Loader, "Attempting to overwrite existing NCA. Skipping...");
@@ -25,6 +25,8 @@ enum class NCAContentType : u8;
 enum class TitleType : u8;

 struct ContentRecord;
+struct MetaRecord;
+class RegisteredCache;

 using NcaID = std::array<u8, 0x10>;
 using ContentProviderParsingFunction = std::function<VirtualFile(const VirtualFile&, const NcaID&)>;
@@ -89,6 +91,27 @@ protected:
    Core::Crypto::KeyManager keys;
 };

+class PlaceholderCache {
+public:
+    explicit PlaceholderCache(VirtualDir dir);
+
+    bool Create(const NcaID& id, u64 size) const;
+    bool Delete(const NcaID& id) const;
+    bool Exists(const NcaID& id) const;
+    bool Write(const NcaID& id, u64 offset, const std::vector<u8>& data) const;
+    bool Register(RegisteredCache* cache, const NcaID& placeholder, const NcaID& install) const;
+    bool CleanAll() const;
+    std::optional<std::array<u8, 0x10>> GetRightsID(const NcaID& id) const;
+    u64 Size(const NcaID& id) const;
+    bool SetSize(const NcaID& id, u64 new_size) const;
+    std::vector<NcaID> List() const;
+
+    static NcaID Generate();
+
+private:
+    VirtualDir dir;
+};
+
 /*
 * A class that catalogues NCAs in the registered directory structure.
 * Nintendo's registered format follows this structure:
@@ -103,6 +126,8 @@ protected:
 * when 4GB splitting can be ignored.)
 */
 class RegisteredCache : public ContentProvider {
+    friend class PlaceholderCache;
+
 public:
    // Parsing function defines the conversion from raw file to NCA. If there are other steps
    // besides creating the NCA from the file (e.g. NAX0 on SD Card), that should go in a custom
@@ -7,6 +7,7 @@
 #include "common/common_types.h"
 #include "common/logging/log.h"
 #include "core/core.h"
+#include "core/file_sys/card_image.h"
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/nca_metadata.h"
 #include "core/file_sys/patch_manager.h"
@@ -34,7 +35,7 @@ void RomFSFactory::SetPackedUpdate(VirtualFile update_raw) {
    this->update_raw = std::move(update_raw);
 }

-ResultVal<VirtualFile> RomFSFactory::OpenCurrentProcess() {
+ResultVal<VirtualFile> RomFSFactory::OpenCurrentProcess() const {
    if (!updatable)
        return MakeResult<VirtualFile>(file);

@@ -43,7 +44,8 @@ ResultVal<VirtualFile> RomFSFactory::OpenCurrentProcess() {
        patch_manager.PatchRomFS(file, ivfc_offset, ContentRecordType::Program, update_raw));
 }

-ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id, StorageId storage, ContentRecordType type) {
+ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id, StorageId storage,
+                                          ContentRecordType type) const {
    std::shared_ptr<NCA> res;

    switch (storage) {
@@ -51,13 +53,17 @@ ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id, StorageId storage, Conte
        res = Core::System::GetInstance().GetContentProvider().GetEntry(title_id, type);
        break;
    case StorageId::NandSystem:
-        res = Service::FileSystem::GetSystemNANDContents()->GetEntry(title_id, type);
+        res =
+            Core::System::GetInstance().GetFileSystemController().GetSystemNANDContents()->GetEntry(
+                title_id, type);
        break;
    case StorageId::NandUser:
-        res = Service::FileSystem::GetUserNANDContents()->GetEntry(title_id, type);
+        res = Core::System::GetInstance().GetFileSystemController().GetUserNANDContents()->GetEntry(
+            title_id, type);
        break;
    case StorageId::SdCard:
-        res = Service::FileSystem::GetSDMCContents()->GetEntry(title_id, type);
+        res = Core::System::GetInstance().GetFileSystemController().GetSDMCContents()->GetEntry(
+            title_id, type);
        break;
    default:
        UNIMPLEMENTED_MSG("Unimplemented storage_id={:02X}", static_cast<u8>(storage));
@@ -33,8 +33,8 @@ public:
    ~RomFSFactory();

    void SetPackedUpdate(VirtualFile update_raw);
-    ResultVal<VirtualFile> OpenCurrentProcess();
-    ResultVal<VirtualFile> Open(u64 title_id, StorageId storage, ContentRecordType type);
+    ResultVal<VirtualFile> OpenCurrentProcess() const;
+    ResultVal<VirtualFile> Open(u64 title_id, StorageId storage, ContentRecordType type) const;

 private:
    VirtualFile file;
@@ -15,22 +15,8 @@ namespace FileSys {

 constexpr char SAVE_DATA_SIZE_FILENAME[] = ".yuzu_save_size";

-std::string SaveDataDescriptor::DebugInfo() const {
-    return fmt::format("[type={:02X}, title_id={:016X}, user_id={:016X}{:016X}, save_id={:016X}, "
-                       "rank={}, index={}]",
-                       static_cast<u8>(type), title_id, user_id[1], user_id[0], save_id,
-                       static_cast<u8>(rank), index);
-}
-
-SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save_directory)) {
-    // Delete all temporary storages
-    // On hardware, it is expected that temporary storage be empty at first use.
-    dir->DeleteSubdirectoryRecursive("temp");
-}
-
-SaveDataFactory::~SaveDataFactory() = default;
-
-ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, const SaveDataDescriptor& meta) {
+namespace {
+void PrintSaveDataDescriptorWarnings(SaveDataDescriptor meta) {
    if (meta.type == SaveDataType::SystemSaveData || meta.type == SaveDataType::SaveData) {
        if (meta.zero_1 != 0) {
            LOG_WARNING(Service_FS,
@@ -65,23 +51,51 @@ ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, const SaveDat
                    "non-zero ({:016X}{:016X})",
                    meta.user_id[1], meta.user_id[0]);
    }
+}
+} // Anonymous namespace

-    std::string save_directory =
+std::string SaveDataDescriptor::DebugInfo() const {
+    return fmt::format("[type={:02X}, title_id={:016X}, user_id={:016X}{:016X}, "
+                       "save_id={:016X}, "
+                       "rank={}, index={}]",
+                       static_cast<u8>(type), title_id, user_id[1], user_id[0], save_id,
+                       static_cast<u8>(rank), index);
+}
+
+SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save_directory)) {
+    // Delete all temporary storages
+    // On hardware, it is expected that temporary storage be empty at first use.
+    dir->DeleteSubdirectoryRecursive("temp");
+}
+
+SaveDataFactory::~SaveDataFactory() = default;
+
+ResultVal<VirtualDir> SaveDataFactory::Create(SaveDataSpaceId space,
+                                              const SaveDataDescriptor& meta) const {
+    PrintSaveDataDescriptorWarnings(meta);
+
+    const auto save_directory =
        GetFullPath(space, meta.type, meta.title_id, meta.user_id, meta.save_id);

-    // TODO(DarkLordZach): Try to not create when opening, there are dedicated create save methods.
-    // But, user_ids don't match so this works for now.
+    auto out = dir->CreateDirectoryRelative(save_directory);
+
+    // Return an error if the save data doesn't actually exist.
+    if (out == nullptr) {
+        // TODO(DarkLordZach): Find out correct error code.
+        return ResultCode(-1);
+    }
+
+    return MakeResult<VirtualDir>(std::move(out));
+}
+
+ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space,
+                                            const SaveDataDescriptor& meta) const {
+
+    const auto save_directory =
+        GetFullPath(space, meta.type, meta.title_id, meta.user_id, meta.save_id);

    auto out = dir->GetDirectoryRelative(save_directory);

-    if (out == nullptr) {
-        // TODO(bunnei): This is a work-around to always create a save data directory if it does not
-        // already exist. This is a hack, as we do not understand yet how this works on hardware.
-        // Without a save data directory, many games will assert on boot. This should not have any
-        // bad side-effects.
-        out = dir->CreateDirectoryRelative(save_directory);
-    }
-
    // Return an error if the save data doesn't actually exist.
    if (out == nullptr) {
        // TODO(Subv): Find out correct error code.
@@ -152,7 +166,7 @@ SaveDataSize SaveDataFactory::ReadSaveDataSize(SaveDataType type, u64 title_id,
 }

 void SaveDataFactory::WriteSaveDataSize(SaveDataType type, u64 title_id, u128 user_id,
-                                        SaveDataSize new_value) {
+                                        SaveDataSize new_value) const {
    const auto path = GetFullPath(SaveDataSpaceId::NandUser, type, title_id, user_id, 0);
    const auto dir = GetOrCreateDirectoryRelative(this->dir, path);

@@ -64,7 +64,8 @@ public:
    explicit SaveDataFactory(VirtualDir dir);
    ~SaveDataFactory();

-    ResultVal<VirtualDir> Open(SaveDataSpaceId space, const SaveDataDescriptor& meta);
+    ResultVal<VirtualDir> Create(SaveDataSpaceId space, const SaveDataDescriptor& meta) const;
+    ResultVal<VirtualDir> Open(SaveDataSpaceId space, const SaveDataDescriptor& meta) const;

    VirtualDir GetSaveDataSpaceDirectory(SaveDataSpaceId space) const;

@@ -73,7 +74,8 @@ public:
                                   u128 user_id, u64 save_id);

    SaveDataSize ReadSaveDataSize(SaveDataType type, u64 title_id, u128 user_id) const;
-    void WriteSaveDataSize(SaveDataType type, u64 title_id, u128 user_id, SaveDataSize new_value);
+    void WriteSaveDataSize(SaveDataType type, u64 title_id, u128 user_id,
+                           SaveDataSize new_value) const;

 private:
    VirtualDir dir;
@@ -6,6 +6,7 @@
 #include "core/file_sys/registered_cache.h"
 #include "core/file_sys/sdmc_factory.h"
 #include "core/file_sys/xts_archive.h"
+#include "core/settings.h"

 namespace FileSys {

@@ -14,16 +15,38 @@ SDMCFactory::SDMCFactory(VirtualDir dir_)
                                GetOrCreateDirectoryRelative(dir, "/Nintendo/Contents/registered"),
                                [](const VirtualFile& file, const NcaID& id) {
                                    return NAX{file, id}.GetDecrypted();
-                                })) {}
+                                })),
+      placeholder(std::make_unique<PlaceholderCache>(
+          GetOrCreateDirectoryRelative(dir, "/Nintendo/Contents/placehld"))) {}

 SDMCFactory::~SDMCFactory() = default;

-ResultVal<VirtualDir> SDMCFactory::Open() {
+ResultVal<VirtualDir> SDMCFactory::Open() const {
    return MakeResult<VirtualDir>(dir);
 }

+VirtualDir SDMCFactory::GetSDMCContentDirectory() const {
+    return GetOrCreateDirectoryRelative(dir, "/Nintendo/Contents");
+}
+
 RegisteredCache* SDMCFactory::GetSDMCContents() const {
    return contents.get();
 }

+PlaceholderCache* SDMCFactory::GetSDMCPlaceholder() const {
+    return placeholder.get();
+}
+
+VirtualDir SDMCFactory::GetImageDirectory() const {
+    return GetOrCreateDirectoryRelative(dir, "/Nintendo/Album");
+}
+
+u64 SDMCFactory::GetSDMCFreeSpace() const {
+    return GetSDMCTotalSpace() - dir->GetSize();
+}
+
+u64 SDMCFactory::GetSDMCTotalSpace() const {
+    return static_cast<u64>(Settings::values.sdmc_size);
+}
+
 } // namespace FileSys
@@ -11,6 +11,7 @@
 namespace FileSys {

 class RegisteredCache;
+class PlaceholderCache;

 /// File system interface to the SDCard archive
 class SDMCFactory {
@@ -18,13 +19,23 @@ public:
    explicit SDMCFactory(VirtualDir dir);
    ~SDMCFactory();

-    ResultVal<VirtualDir> Open();
+    ResultVal<VirtualDir> Open() const;
+
+    VirtualDir GetSDMCContentDirectory() const;
+
    RegisteredCache* GetSDMCContents() const;
+    PlaceholderCache* GetSDMCPlaceholder() const;
+
+    VirtualDir GetImageDirectory() const;
+
+    u64 GetSDMCFreeSpace() const;
+    u64 GetSDMCTotalSpace() const;

 private:
    VirtualDir dir;

    std::unique_ptr<RegisteredCache> contents;
+    std::unique_ptr<PlaceholderCache> placeholder;
 };

 } // namespace FileSys
@@ -14,6 +14,7 @@
 #include "core/file_sys/content_archive.h"
 #include "core/file_sys/nca_metadata.h"
 #include "core/file_sys/partition_filesystem.h"
+#include "core/file_sys/program_metadata.h"
 #include "core/file_sys/submission_package.h"
 #include "core/loader/loader.h"

@@ -78,6 +79,10 @@ Loader::ResultStatus NSP::GetStatus() const {
 }

 Loader::ResultStatus NSP::GetProgramStatus(u64 title_id) const {
+    if (IsExtractedType() && GetExeFS() != nullptr && FileSys::IsDirectoryExeFS(GetExeFS())) {
+        return Loader::ResultStatus::Success;
+    }
+
    const auto iter = program_status.find(title_id);
    if (iter == program_status.end())
        return Loader::ResultStatus::ErrorNSPMissingProgramNCA;
@@ -85,12 +90,29 @@ Loader::ResultStatus NSP::GetProgramStatus(u64 title_id) const {
 }

 u64 NSP::GetFirstTitleID() const {
+    if (IsExtractedType()) {
+        return GetProgramTitleID();
+    }
+
    if (program_status.empty())
        return 0;
    return program_status.begin()->first;
 }

 u64 NSP::GetProgramTitleID() const {
+    if (IsExtractedType()) {
+        if (GetExeFS() == nullptr || !IsDirectoryExeFS(GetExeFS())) {
+            return 0;
+        }
+
+        ProgramMetadata meta;
+        if (meta.Load(GetExeFS()->GetFile("main.npdm")) == Loader::ResultStatus::Success) {
+            return meta.GetTitleID();
+        } else {
+            return 0;
+        }
+    }
+
    const auto out = GetFirstTitleID();
    if ((out & 0x800) == 0)
        return out;
@@ -102,6 +124,10 @@ u64 NSP::GetProgramTitleID() const {
 }

 std::vector<u64> NSP::GetTitleIDs() const {
+    if (IsExtractedType()) {
+        return {GetProgramTitleID()};
+    }
+
    std::vector<u64> out;
    out.reserve(ncas.size());
    for (const auto& kv : ncas)
@@ -222,7 +248,8 @@ void NSP::InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files) {

 void NSP::ReadNCAs(const std::vector<VirtualFile>& files) {
    for (const auto& outer_file : files) {
-        if (outer_file->GetName().substr(outer_file->GetName().size() - 9) != ".cnmt.nca") {
+        if (outer_file->GetName().size() < 9 ||
+            outer_file->GetName().substr(outer_file->GetName().size() - 9) != ".cnmt.nca") {
            continue;
        }

@@ -0,0 +1,46 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/file_sys/system_archive/mii_model.h"
+#include "core/file_sys/vfs_vector.h"
+
+namespace FileSys::SystemArchive {
+
+namespace MiiModelData {
+
+constexpr std::array<u8, 0x10> NFTR_STANDARD{'N',  'F',  'T',  'R',  0x01, 0x00, 0x00, 0x00,
+                                             0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00};
+constexpr std::array<u8, 0x10> NFSR_STANDARD{'N',  'F',  'S',  'R',  0x01, 0x00, 0x00, 0x00,
+                                             0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00};
+
+constexpr auto TEXTURE_LOW_LINEAR = NFTR_STANDARD;
+constexpr auto TEXTURE_LOW_SRGB = NFTR_STANDARD;
+constexpr auto TEXTURE_MID_LINEAR = NFTR_STANDARD;
+constexpr auto TEXTURE_MID_SRGB = NFTR_STANDARD;
+constexpr auto SHAPE_HIGH = NFSR_STANDARD;
+constexpr auto SHAPE_MID = NFSR_STANDARD;
+
+} // namespace MiiModelData
+
+VirtualDir MiiModel() {
+    auto out = std::make_shared<VectorVfsDirectory>(std::vector<VirtualFile>{},
+                                                    std::vector<VirtualDir>{}, "data");
+
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_LOW_LINEAR.size()>>(
+        MiiModelData::TEXTURE_LOW_LINEAR, "NXTextureLowLinear.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_LOW_SRGB.size()>>(
+        MiiModelData::TEXTURE_LOW_SRGB, "NXTextureLowSRGB.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_MID_LINEAR.size()>>(
+        MiiModelData::TEXTURE_MID_LINEAR, "NXTextureMidLinear.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_MID_SRGB.size()>>(
+        MiiModelData::TEXTURE_MID_SRGB, "NXTextureMidSRGB.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::SHAPE_HIGH.size()>>(
+        MiiModelData::SHAPE_HIGH, "ShapeHigh.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::SHAPE_MID.size()>>(
+        MiiModelData::SHAPE_MID, "ShapeMid.dat"));
+
+    return std::move(out);
+}
+
+} // namespace FileSys::SystemArchive
@@ -0,0 +1,13 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/file_sys/vfs_types.h"
+
+namespace FileSys::SystemArchive {
+
+VirtualDir MiiModel();
+
+} // namespace FileSys::SystemArchive
@@ -4,6 +4,7 @@

 #include "common/logging/log.h"
 #include "core/file_sys/romfs.h"
+#include "core/file_sys/system_archive/mii_model.h"
 #include "core/file_sys/system_archive/ng_word.h"
 #include "core/file_sys/system_archive/system_archive.h"
 #include "core/file_sys/system_archive/system_version.h"
@@ -24,7 +25,7 @@ struct SystemArchiveDescriptor {
 constexpr std::array<SystemArchiveDescriptor, SYSTEM_ARCHIVE_COUNT> SYSTEM_ARCHIVES{{
    {0x0100000000000800, "CertStore", nullptr},
    {0x0100000000000801, "ErrorMessage", nullptr},
-    {0x0100000000000802, "MiiModel", nullptr},
+    {0x0100000000000802, "MiiModel", &MiiModel},
    {0x0100000000000803, "BrowserDll", nullptr},
    {0x0100000000000804, "Help", nullptr},
    {0x0100000000000805, "SharedFont", nullptr},
@@ -202,13 +202,11 @@ void RegisterModule(std::string name, VAddr beg, VAddr end, bool add_elf_ext) {
 }

 static Kernel::Thread* FindThreadById(s64 id) {
-    for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
-        const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
-        for (auto& thread : threads) {
-            if (thread->GetThreadID() == static_cast<u64>(id)) {
-                current_core = core;
-                return thread.get();
-            }
+    const auto& threads = Core::System::GetInstance().GlobalScheduler().GetThreadList();
+    for (auto& thread : threads) {
+        if (thread->GetThreadID() == static_cast<u64>(id)) {
+            current_core = thread->GetProcessorID();
+            return thread.get();
        }
    }
    return nullptr;
@@ -646,11 +644,9 @@ static void HandleQuery() {
        SendReply(buffer.c_str());
    } else if (strncmp(query, "fThreadInfo", strlen("fThreadInfo")) == 0) {
        std::string val = "m";
-        for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
-            const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
-            for (const auto& thread : threads) {
-                val += fmt::format("{:x},", thread->GetThreadID());
-            }
+        const auto& threads = Core::System::GetInstance().GlobalScheduler().GetThreadList();
+        for (const auto& thread : threads) {
+            val += fmt::format("{:x},", thread->GetThreadID());
        }
        val.pop_back();
        SendReply(val.c_str());
@@ -660,13 +656,11 @@ static void HandleQuery() {
        std::string buffer;
        buffer += "l<?xml version=\"1.0\"?>";
        buffer += "<threads>";
-        for (u32 core = 0; core < Core::NUM_CPU_CORES; core++) {
-            const auto& threads = Core::System::GetInstance().Scheduler(core).GetThreadList();
-            for (const auto& thread : threads) {
-                buffer +=
-                    fmt::format(R"*(<thread id="{:x}" core="{:d}" name="Thread {:x}"></thread>)*",
-                                thread->GetThreadID(), core, thread->GetThreadID());
-            }
+        const auto& threads = Core::System::GetInstance().GlobalScheduler().GetThreadList();
+        for (const auto& thread : threads) {
+            buffer +=
+                fmt::format(R"*(<thread id="{:x}" core="{:d}" name="Thread {:x}"></thread>)*",
+                            thread->GetThreadID(), thread->GetProcessorID(), thread->GetThreadID());
        }
        buffer += "</threads>";
        SendReply(buffer.c_str());
@@ -22,6 +22,7 @@ namespace Kernel {
 namespace {
 // Wake up num_to_wake (or all) threads in a vector.
 void WakeThreads(const std::vector<SharedPtr<Thread>>& waiting_threads, s32 num_to_wake) {
+    auto& system = Core::System::GetInstance();
    // Only process up to 'target' threads, unless 'target' is <= 0, in which case process
    // them all.
    std::size_t last = waiting_threads.size();
@@ -35,6 +36,7 @@ void WakeThreads(const std::vector<SharedPtr<Thread>>& waiting_threads, s32 num_
        waiting_threads[i]->SetWaitSynchronizationResult(RESULT_SUCCESS);
        waiting_threads[i]->SetArbiterWaitAddress(0);
        waiting_threads[i]->ResumeFromWait();
+        system.PrepareReschedule(waiting_threads[i]->GetProcessorID());
    }
 }
 } // Anonymous namespace
@@ -169,30 +171,22 @@ ResultCode AddressArbiter::WaitForAddressImpl(VAddr address, s64 timeout) {

    current_thread->WakeAfterDelay(timeout);

-    system.CpuCore(current_thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(current_thread->GetProcessorID());
    return RESULT_TIMEOUT;
 }

 std::vector<SharedPtr<Thread>> AddressArbiter::GetThreadsWaitingOnAddress(VAddr address) const {
-    const auto RetrieveWaitingThreads = [this](std::size_t core_index,
-                                               std::vector<SharedPtr<Thread>>& waiting_threads,
-                                               VAddr arb_addr) {
-        const auto& scheduler = system.Scheduler(core_index);
-        const auto& thread_list = scheduler.GetThreadList();
-
-        for (const auto& thread : thread_list) {
-            if (thread->GetArbiterWaitAddress() == arb_addr) {
-                waiting_threads.push_back(thread);
-            }
-        }
-    };

    // Retrieve all threads that are waiting for this address.
    std::vector<SharedPtr<Thread>> threads;
-    RetrieveWaitingThreads(0, threads, address);
-    RetrieveWaitingThreads(1, threads, address);
-    RetrieveWaitingThreads(2, threads, address);
-    RetrieveWaitingThreads(3, threads, address);
+    const auto& scheduler = system.GlobalScheduler();
+    const auto& thread_list = scheduler.GetThreadList();
+
+    for (const auto& thread : thread_list) {
+        if (thread->GetArbiterWaitAddress() == address) {
+            threads.push_back(thread);
+        }
+    }

    // Sort them by priority, such that the highest priority ones come first.
    std::sort(threads.begin(), threads.end(),
@@ -18,6 +18,7 @@
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/resource_limit.h"
+#include "core/hle/kernel/scheduler.h"
 #include "core/hle/kernel/thread.h"
 #include "core/hle/lock.h"
 #include "core/hle/result.h"
@@ -88,7 +89,7 @@ static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] s64 cycles_
 }

 struct KernelCore::Impl {
-    explicit Impl(Core::System& system) : system{system} {}
+    explicit Impl(Core::System& system) : system{system}, global_scheduler{system} {}

    void Initialize(KernelCore& kernel) {
        Shutdown();
@@ -140,6 +141,7 @@ struct KernelCore::Impl {
    // Lists all processes that exist in the current session.
    std::vector<SharedPtr<Process>> process_list;
    Process* current_process = nullptr;
+    Kernel::GlobalScheduler global_scheduler;

    SharedPtr<ResourceLimit> system_resource_limit;

@@ -203,6 +205,14 @@ const std::vector<SharedPtr<Process>>& KernelCore::GetProcessList() const {
    return impl->process_list;
 }

+Kernel::GlobalScheduler& KernelCore::GlobalScheduler() {
+    return impl->global_scheduler;
+}
+
+const Kernel::GlobalScheduler& KernelCore::GlobalScheduler() const {
+    return impl->global_scheduler;
+}
+
 void KernelCore::AddNamedPort(std::string name, SharedPtr<ClientPort> port) {
    impl->named_ports.emplace(std::move(name), std::move(port));
 }
@@ -25,6 +25,7 @@ class HandleTable;
 class Process;
 class ResourceLimit;
 class Thread;
+class GlobalScheduler;

 /// Represents a single instance of the kernel.
 class KernelCore {
@@ -75,6 +76,12 @@ public:
    /// Retrieves the list of processes.
    const std::vector<SharedPtr<Process>>& GetProcessList() const;

+    /// Gets the sole instance of the global scheduler
+    Kernel::GlobalScheduler& GlobalScheduler();
+
+    /// Gets the sole instance of the global scheduler
+    const Kernel::GlobalScheduler& GlobalScheduler() const;
+
    /// Adds a port to the named port table
    void AddNamedPort(std::string name, SharedPtr<ClientPort> port);

@@ -7,6 +7,7 @@

 #include "common/assert.h"
 #include "core/core.h"
+#include "core/core_cpu.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/kernel.h"
@@ -78,7 +79,7 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle,
    // thread.
    ASSERT(requesting_thread == current_thread);

-    const u32 addr_value = Memory::Read32(address);
+    u32 addr_value = Memory::Read32(address);

    // If the mutex isn't being held, just return success.
    if (addr_value != (holding_thread_handle | Mutex::MutexHasWaitersFlag)) {
@@ -89,6 +90,20 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle,
        return ERR_INVALID_HANDLE;
    }

+    // This a workaround where an unknown bug writes the mutex value to give ownership to a cond var
+    // waiting thread.
+    if (holding_thread->GetStatus() == ThreadStatus::WaitCondVar) {
+        if (holding_thread->GetMutexWaitAddress() == address) {
+            Release(address, holding_thread.get());
+            addr_value = Memory::Read32(address);
+            if (addr_value == 0)
+                return RESULT_SUCCESS;
+            else {
+                holding_thread = handle_table.Get<Thread>(addr_value & Mutex::MutexOwnerMask);
+            }
+        }
+    }
+
    // Wait until the mutex is released
    current_thread->SetMutexWaitAddress(address);
    current_thread->SetWaitHandle(requesting_thread_handle);
@@ -104,14 +119,13 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle,
    return RESULT_SUCCESS;
 }

-ResultCode Mutex::Release(VAddr address) {
+ResultCode Mutex::Release(VAddr address, Thread* holding_thread) {
    // The mutex address must be 4-byte aligned
    if ((address % sizeof(u32)) != 0) {
        return ERR_INVALID_ADDRESS;
    }

-    auto* const current_thread = system.CurrentScheduler().GetCurrentThread();
-    auto [thread, num_waiters] = GetHighestPriorityMutexWaitingThread(current_thread, address);
+    auto [thread, num_waiters] = GetHighestPriorityMutexWaitingThread(holding_thread, address);

    // There are no more threads waiting for the mutex, release it completely.
    if (thread == nullptr) {
@@ -120,7 +134,7 @@ ResultCode Mutex::Release(VAddr address) {
    }

    // Transfer the ownership of the mutex from the previous owner to the new one.
-    TransferMutexOwnership(address, current_thread, thread);
+    TransferMutexOwnership(address, holding_thread, thread);

    u32 mutex_value = thread->GetWaitHandle();

@@ -140,6 +154,11 @@ ResultCode Mutex::Release(VAddr address) {
    thread->SetMutexWaitAddress(0);
    thread->SetWaitHandle(0);

+    if (thread->GetProcessorID() >= 0)
+        system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+    if (holding_thread->GetProcessorID() >= 0)
+        system.CpuCore(holding_thread->GetProcessorID()).PrepareReschedule();
+
    return RESULT_SUCCESS;
 }
 } // namespace Kernel
@@ -29,7 +29,7 @@ public:
                          Handle requesting_thread_handle);

    /// Releases the mutex at the specified address.
-    ResultCode Release(VAddr address);
+    ResultCode Release(VAddr address, Thread* holding_thread);

 private:
    Core::System& system;
@@ -213,10 +213,7 @@ void Process::PrepareForTermination() {
        }
    };

-    stop_threads(system.Scheduler(0).GetThreadList());
-    stop_threads(system.Scheduler(1).GetThreadList());
-    stop_threads(system.Scheduler(2).GetThreadList());
-    stop_threads(system.Scheduler(3).GetThreadList());
+    stop_threads(system.GlobalScheduler().GetThreadList());

    FreeTLSRegion(tls_region_address);
    tls_region_address = 0;
@@ -1,8 +1,13 @@
 // Copyright 2018 yuzu emulator team
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
+//
+// SelectThreads, Yield functions originally by TuxSH.
+// licensed under GPLv2 or later under exception provided by the author.

 #include <algorithm>
+#include <set>
+#include <unordered_set>
 #include <utility>

 #include "common/assert.h"
@@ -17,56 +22,277 @@

 namespace Kernel {

-std::mutex Scheduler::scheduler_mutex;
+GlobalScheduler::GlobalScheduler(Core::System& system) : system{system} {
+    reselection_pending = false;
+}

-Scheduler::Scheduler(Core::System& system, Core::ARM_Interface& cpu_core)
-    : cpu_core{cpu_core}, system{system} {}
+void GlobalScheduler::AddThread(SharedPtr<Thread> thread) {
+    thread_list.push_back(std::move(thread));
+}

-Scheduler::~Scheduler() {
-    for (auto& thread : thread_list) {
-        thread->Stop();
+void GlobalScheduler::RemoveThread(const Thread* thread) {
+    thread_list.erase(std::remove(thread_list.begin(), thread_list.end(), thread),
+                      thread_list.end());
+}
+
+/*
+ * UnloadThread selects a core and forces it to unload its current thread's context
+ */
+void GlobalScheduler::UnloadThread(s32 core) {
+    Scheduler& sched = system.Scheduler(core);
+    sched.UnloadThread();
+}
+
+/*
+ * SelectThread takes care of selecting the new scheduled thread.
+ * It does it in 3 steps:
+ * - First a thread is selected from the top of the priority queue. If no thread
+ * is obtained then we move to step two, else we are done.
+ * - Second we try to get a suggested thread that's not assigned to any core or
+ * that is not the top thread in that core.
+ * - Third is no suggested thread is found, we do a second pass and pick a running
+ * thread in another core and swap it with its current thread.
+ */
+void GlobalScheduler::SelectThread(u32 core) {
+    const auto update_thread = [](Thread* thread, Scheduler& sched) {
+        if (thread != sched.selected_thread) {
+            if (thread == nullptr) {
+                ++sched.idle_selection_count;
+            }
+            sched.selected_thread = thread;
+        }
+        sched.context_switch_pending = sched.selected_thread != sched.current_thread;
+        std::atomic_thread_fence(std::memory_order_seq_cst);
+    };
+    Scheduler& sched = system.Scheduler(core);
+    Thread* current_thread = nullptr;
+    // Step 1: Get top thread in schedule queue.
+    current_thread = scheduled_queue[core].empty() ? nullptr : scheduled_queue[core].front();
+    if (current_thread) {
+        update_thread(current_thread, sched);
+        return;
+    }
+    // Step 2: Try selecting a suggested thread.
+    Thread* winner = nullptr;
+    std::set<s32> sug_cores;
+    for (auto thread : suggested_queue[core]) {
+        s32 this_core = thread->GetProcessorID();
+        Thread* thread_on_core = nullptr;
+        if (this_core >= 0) {
+            thread_on_core = scheduled_queue[this_core].front();
+        }
+        if (this_core < 0 || thread != thread_on_core) {
+            winner = thread;
+            break;
+        }
+        sug_cores.insert(this_core);
+    }
+    // if we got a suggested thread, select it, else do a second pass.
+    if (winner && winner->GetPriority() > 2) {
+        if (winner->IsRunning()) {
+            UnloadThread(winner->GetProcessorID());
+        }
+        TransferToCore(winner->GetPriority(), core, winner);
+        update_thread(winner, sched);
+        return;
+    }
+    // Step 3: Select a suggested thread from another core
+    for (auto& src_core : sug_cores) {
+        auto it = scheduled_queue[src_core].begin();
+        it++;
+        if (it != scheduled_queue[src_core].end()) {
+            Thread* thread_on_core = scheduled_queue[src_core].front();
+            Thread* to_change = *it;
+            if (thread_on_core->IsRunning() || to_change->IsRunning()) {
+                UnloadThread(src_core);
+            }
+            TransferToCore(thread_on_core->GetPriority(), core, thread_on_core);
+            current_thread = thread_on_core;
+            break;
+        }
+    }
+    update_thread(current_thread, sched);
+}
+
+/*
+ * YieldThread takes a thread and moves it to the back of the it's priority list
+ * This operation can be redundant and no scheduling is changed if marked as so.
+ */
+void GlobalScheduler::YieldThread(Thread* yielding_thread) {
+    // Note: caller should use critical section, etc.
+    const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID());
+    const u32 priority = yielding_thread->GetPriority();
+
+    // Yield the thread
+    ASSERT_MSG(yielding_thread == scheduled_queue[core_id].front(priority),
+               "Thread yielding without being in front");
+    scheduled_queue[core_id].yield(priority);
+
+    Thread* winner = scheduled_queue[core_id].front(priority);
+    AskForReselectionOrMarkRedundant(yielding_thread, winner);
+}
+
+/*
+ * YieldThreadAndBalanceLoad takes a thread and moves it to the back of the it's priority list.
+ * Afterwards, tries to pick a suggested thread from the suggested queue that has worse time or
+ * a better priority than the next thread in the core.
+ * This operation can be redundant and no scheduling is changed if marked as so.
+ */
+void GlobalScheduler::YieldThreadAndBalanceLoad(Thread* yielding_thread) {
+    // Note: caller should check if !thread.IsSchedulerOperationRedundant and use critical section,
+    // etc.
+    const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID());
+    const u32 priority = yielding_thread->GetPriority();
+
+    // Yield the thread
+    ASSERT_MSG(yielding_thread == scheduled_queue[core_id].front(priority),
+               "Thread yielding without being in front");
+    scheduled_queue[core_id].yield(priority);
+
+    std::array<Thread*, NUM_CPU_CORES> current_threads;
+    for (u32 i = 0; i < NUM_CPU_CORES; i++) {
+        current_threads[i] = scheduled_queue[i].empty() ? nullptr : scheduled_queue[i].front();
+    }
+
+    Thread* next_thread = scheduled_queue[core_id].front(priority);
+    Thread* winner = nullptr;
+    for (auto& thread : suggested_queue[core_id]) {
+        const s32 source_core = thread->GetProcessorID();
+        if (source_core >= 0) {
+            if (current_threads[source_core] != nullptr) {
+                if (thread == current_threads[source_core] ||
+                    current_threads[source_core]->GetPriority() < min_regular_priority) {
+                    continue;
+                }
+            }
+            if (next_thread->GetLastRunningTicks() >= thread->GetLastRunningTicks() ||
+                next_thread->GetPriority() < thread->GetPriority()) {
+                if (thread->GetPriority() <= priority) {
+                    winner = thread;
+                    break;
+                }
+            }
+        }
+    }
+
+    if (winner != nullptr) {
+        if (winner != yielding_thread) {
+            if (winner->IsRunning()) {
+                UnloadThread(winner->GetProcessorID());
+            }
+            TransferToCore(winner->GetPriority(), core_id, winner);
+        }
+    } else {
+        winner = next_thread;
+    }
+
+    AskForReselectionOrMarkRedundant(yielding_thread, winner);
+}
+
+/*
+ * YieldThreadAndWaitForLoadBalancing takes a thread and moves it out of the scheduling queue
+ * and into the suggested queue. If no thread can be squeduled afterwards in that core,
+ * a suggested thread is obtained instead.
+ * This operation can be redundant and no scheduling is changed if marked as so.
+ */
+void GlobalScheduler::YieldThreadAndWaitForLoadBalancing(Thread* yielding_thread) {
+    // Note: caller should check if !thread.IsSchedulerOperationRedundant and use critical section,
+    // etc.
+    Thread* winner = nullptr;
+    const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID());
+
+    // Remove the thread from its scheduled mlq, put it on the corresponding "suggested" one instead
+    TransferToCore(yielding_thread->GetPriority(), -1, yielding_thread);
+
+    // If the core is idle, perform load balancing, excluding the threads that have just used this
+    // function...
+    if (scheduled_queue[core_id].empty()) {
+        // Here, "current_threads" is calculated after the ""yield"", unlike yield -1
+        std::array<Thread*, NUM_CPU_CORES> current_threads;
+        for (u32 i = 0; i < NUM_CPU_CORES; i++) {
+            current_threads[i] = scheduled_queue[i].empty() ? nullptr : scheduled_queue[i].front();
+        }
+        for (auto& thread : suggested_queue[core_id]) {
+            const s32 source_core = thread->GetProcessorID();
+            if (source_core < 0 || thread == current_threads[source_core]) {
+                continue;
+            }
+            if (current_threads[source_core] == nullptr ||
+                current_threads[source_core]->GetPriority() >= min_regular_priority) {
+                winner = thread;
+            }
+            break;
+        }
+        if (winner != nullptr) {
+            if (winner != yielding_thread) {
+                if (winner->IsRunning()) {
+                    UnloadThread(winner->GetProcessorID());
+                }
+                TransferToCore(winner->GetPriority(), core_id, winner);
+            }
+        } else {
+            winner = yielding_thread;
+        }
+    }
+
+    AskForReselectionOrMarkRedundant(yielding_thread, winner);
+}
+
+void GlobalScheduler::Schedule(u32 priority, u32 core, Thread* thread) {
+    ASSERT_MSG(thread->GetProcessorID() == core, "Thread must be assigned to this core.");
+    scheduled_queue[core].add(thread, priority);
+}
+
+void GlobalScheduler::SchedulePrepend(u32 priority, u32 core, Thread* thread) {
+    ASSERT_MSG(thread->GetProcessorID() == core, "Thread must be assigned to this core.");
+    scheduled_queue[core].add(thread, priority, false);
+}
+
+void GlobalScheduler::AskForReselectionOrMarkRedundant(Thread* current_thread, Thread* winner) {
+    if (current_thread == winner) {
+        // TODO(blinkhawk): manage redundant operations, this is not implemented.
+        // as its mostly an optimization.
+        // current_thread->SetRedundantSchedulerOperation();
+    } else {
+        reselection_pending.store(true, std::memory_order_release);
    }
 }

+GlobalScheduler::~GlobalScheduler() = default;
+
+Scheduler::Scheduler(Core::System& system, Core::ARM_Interface& cpu_core, u32 core_id)
+    : system(system), cpu_core(cpu_core), core_id(core_id) {}
+
+Scheduler::~Scheduler() = default;
+
 bool Scheduler::HaveReadyThreads() const {
-    std::lock_guard lock{scheduler_mutex};
-    return !ready_queue.empty();
+    return system.GlobalScheduler().HaveReadyThreads(core_id);
 }

 Thread* Scheduler::GetCurrentThread() const {
    return current_thread.get();
 }

+Thread* Scheduler::GetSelectedThread() const {
+    return selected_thread.get();
+}
+
+void Scheduler::SelectThreads() {
+    system.GlobalScheduler().SelectThread(core_id);
+}
+
 u64 Scheduler::GetLastContextSwitchTicks() const {
    return last_context_switch_time;
 }

-Thread* Scheduler::PopNextReadyThread() {
-    Thread* next = nullptr;
-    Thread* thread = GetCurrentThread();
-
-    if (thread && thread->GetStatus() == ThreadStatus::Running) {
-        if (ready_queue.empty()) {
-            return thread;
-        }
-        // We have to do better than the current thread.
-        // This call returns null when that's not possible.
-        next = ready_queue.front();
-        if (next == nullptr || next->GetPriority() >= thread->GetPriority()) {
-            next = thread;
-        }
-    } else {
-        if (ready_queue.empty()) {
-            return nullptr;
-        }
-        next = ready_queue.front();
+void Scheduler::TryDoContextSwitch() {
+    if (context_switch_pending) {
+        SwitchContext();
    }
-
-    return next;
 }

-void Scheduler::SwitchContext(Thread* new_thread) {
-    Thread* previous_thread = GetCurrentThread();
+void Scheduler::UnloadThread() {
+    Thread* const previous_thread = GetCurrentThread();
    Process* const previous_process = system.Kernel().CurrentProcess();

    UpdateLastContextSwitchTime(previous_thread, previous_process);
@@ -80,23 +306,52 @@ void Scheduler::SwitchContext(Thread* new_thread) {
        if (previous_thread->GetStatus() == ThreadStatus::Running) {
            // This is only the case when a reschedule is triggered without the current thread
            // yielding execution (i.e. an event triggered, system core time-sliced, etc)
-            ready_queue.add(previous_thread, previous_thread->GetPriority(), false);
            previous_thread->SetStatus(ThreadStatus::Ready);
        }
+        previous_thread->SetIsRunning(false);
+    }
+    current_thread = nullptr;
+}
+
+void Scheduler::SwitchContext() {
+    Thread* const previous_thread = GetCurrentThread();
+    Thread* const new_thread = GetSelectedThread();
+
+    context_switch_pending = false;
+    if (new_thread == previous_thread) {
+        return;
+    }
+
+    Process* const previous_process = system.Kernel().CurrentProcess();
+
+    UpdateLastContextSwitchTime(previous_thread, previous_process);
+
+    // Save context for previous thread
+    if (previous_thread) {
+        cpu_core.SaveContext(previous_thread->GetContext());
+        // Save the TPIDR_EL0 system register in case it was modified.
+        previous_thread->SetTPIDR_EL0(cpu_core.GetTPIDR_EL0());
+
+        if (previous_thread->GetStatus() == ThreadStatus::Running) {
+            // This is only the case when a reschedule is triggered without the current thread
+            // yielding execution (i.e. an event triggered, system core time-sliced, etc)
+            previous_thread->SetStatus(ThreadStatus::Ready);
+        }
+        previous_thread->SetIsRunning(false);
    }

    // Load context of new thread
    if (new_thread) {
+        ASSERT_MSG(new_thread->GetProcessorID() == this->core_id,
+                   "Thread must be assigned to this core.");
        ASSERT_MSG(new_thread->GetStatus() == ThreadStatus::Ready,
                   "Thread must be ready to become running.");

        // Cancel any outstanding wakeup events for this thread
        new_thread->CancelWakeupTimer();
-
        current_thread = new_thread;
-
-        ready_queue.remove(new_thread, new_thread->GetPriority());
        new_thread->SetStatus(ThreadStatus::Running);
+        new_thread->SetIsRunning(true);

        auto* const thread_owner_process = current_thread->GetOwnerProcess();
        if (previous_process != thread_owner_process) {
@@ -130,124 +385,4 @@ void Scheduler::UpdateLastContextSwitchTime(Thread* thread, Process* process) {
    last_context_switch_time = most_recent_switch_ticks;
 }

-void Scheduler::Reschedule() {
-    std::lock_guard lock{scheduler_mutex};
-
-    Thread* cur = GetCurrentThread();
-    Thread* next = PopNextReadyThread();
-
-    if (cur && next) {
-        LOG_TRACE(Kernel, "context switch {} -> {}", cur->GetObjectId(), next->GetObjectId());
-    } else if (cur) {
-        LOG_TRACE(Kernel, "context switch {} -> idle", cur->GetObjectId());
-    } else if (next) {
-        LOG_TRACE(Kernel, "context switch idle -> {}", next->GetObjectId());
-    }
-
-    SwitchContext(next);
-}
-
-void Scheduler::AddThread(SharedPtr<Thread> thread) {
-    std::lock_guard lock{scheduler_mutex};
-
-    thread_list.push_back(std::move(thread));
-}
-
-void Scheduler::RemoveThread(Thread* thread) {
-    std::lock_guard lock{scheduler_mutex};
-
-    thread_list.erase(std::remove(thread_list.begin(), thread_list.end(), thread),
-                      thread_list.end());
-}
-
-void Scheduler::ScheduleThread(Thread* thread, u32 priority) {
-    std::lock_guard lock{scheduler_mutex};
-
-    ASSERT(thread->GetStatus() == ThreadStatus::Ready);
-    ready_queue.add(thread, priority);
-}
-
-void Scheduler::UnscheduleThread(Thread* thread, u32 priority) {
-    std::lock_guard lock{scheduler_mutex};
-
-    ASSERT(thread->GetStatus() == ThreadStatus::Ready);
-    ready_queue.remove(thread, priority);
-}
-
-void Scheduler::SetThreadPriority(Thread* thread, u32 priority) {
-    std::lock_guard lock{scheduler_mutex};
-    if (thread->GetPriority() == priority) {
-        return;
-    }
-
-    // If thread was ready, adjust queues
-    if (thread->GetStatus() == ThreadStatus::Ready)
-        ready_queue.adjust(thread, thread->GetPriority(), priority);
-}
-
-Thread* Scheduler::GetNextSuggestedThread(u32 core, u32 maximum_priority) const {
-    std::lock_guard lock{scheduler_mutex};
-
-    const u32 mask = 1U << core;
-    for (auto* thread : ready_queue) {
-        if ((thread->GetAffinityMask() & mask) != 0 && thread->GetPriority() < maximum_priority) {
-            return thread;
-        }
-    }
-    return nullptr;
-}
-
-void Scheduler::YieldWithoutLoadBalancing(Thread* thread) {
-    ASSERT(thread != nullptr);
-    // Avoid yielding if the thread isn't even running.
-    ASSERT(thread->GetStatus() == ThreadStatus::Running);
-
-    // Sanity check that the priority is valid
-    ASSERT(thread->GetPriority() < THREADPRIO_COUNT);
-
-    // Yield this thread -- sleep for zero time and force reschedule to different thread
-    GetCurrentThread()->Sleep(0);
-}
-
-void Scheduler::YieldWithLoadBalancing(Thread* thread) {
-    ASSERT(thread != nullptr);
-    const auto priority = thread->GetPriority();
-    const auto core = static_cast<u32>(thread->GetProcessorID());
-
-    // Avoid yielding if the thread isn't even running.
-    ASSERT(thread->GetStatus() == ThreadStatus::Running);
-
-    // Sanity check that the priority is valid
-    ASSERT(priority < THREADPRIO_COUNT);
-
-    // Sleep for zero time to be able to force reschedule to different thread
-    GetCurrentThread()->Sleep(0);
-
-    Thread* suggested_thread = nullptr;
-
-    // Search through all of the cpu cores (except this one) for a suggested thread.
-    // Take the first non-nullptr one
-    for (unsigned cur_core = 0; cur_core < Core::NUM_CPU_CORES; ++cur_core) {
-        const auto res =
-            system.CpuCore(cur_core).Scheduler().GetNextSuggestedThread(core, priority);
-
-        // If scheduler provides a suggested thread
-        if (res != nullptr) {
-            // And its better than the current suggested thread (or is the first valid one)
-            if (suggested_thread == nullptr ||
-                suggested_thread->GetPriority() > res->GetPriority()) {
-                suggested_thread = res;
-            }
-        }
-    }
-
-    // If a suggested thread was found, queue that for this core
-    if (suggested_thread != nullptr)
-        suggested_thread->ChangeCore(core, suggested_thread->GetAffinityMask());
-}
-
-void Scheduler::YieldAndWaitForLoadBalancing(Thread* thread) {
-    UNIMPLEMENTED_MSG("Wait for load balancing thread yield type is not implemented!");
-}
-
 } // namespace Kernel
@@ -20,124 +20,178 @@ namespace Kernel {

 class Process;

-class Scheduler final {
+class GlobalScheduler final {
 public:
-    explicit Scheduler(Core::System& system, Core::ARM_Interface& cpu_core);
-    ~Scheduler();
-
-    /// Returns whether there are any threads that are ready to run.
-    bool HaveReadyThreads() const;
-
-    /// Reschedules to the next available thread (call after current thread is suspended)
-    void Reschedule();
-
-    /// Gets the current running thread
-    Thread* GetCurrentThread() const;
-
-    /// Gets the timestamp for the last context switch in ticks.
-    u64 GetLastContextSwitchTicks() const;
+    static constexpr u32 NUM_CPU_CORES = 4;

+    explicit GlobalScheduler(Core::System& system);
+    ~GlobalScheduler();
    /// Adds a new thread to the scheduler
    void AddThread(SharedPtr<Thread> thread);

    /// Removes a thread from the scheduler
-    void RemoveThread(Thread* thread);
-
-    /// Schedules a thread that has become "ready"
-    void ScheduleThread(Thread* thread, u32 priority);
-
-    /// Unschedules a thread that was already scheduled
-    void UnscheduleThread(Thread* thread, u32 priority);
-
-    /// Sets the priority of a thread in the scheduler
-    void SetThreadPriority(Thread* thread, u32 priority);
-
-    /// Gets the next suggested thread for load balancing
-    Thread* GetNextSuggestedThread(u32 core, u32 minimum_priority) const;
-
-    /**
-     * YieldWithoutLoadBalancing -- analogous to normal yield on a system
-     * Moves the thread to the end of the ready queue for its priority, and then reschedules the
-     * system to the new head of the queue.
-     *
-     * Example (Single Core -- but can be extrapolated to multi):
-     * ready_queue[prio=0]: ThreadA, ThreadB, ThreadC (->exec order->)
-     * Currently Running: ThreadR
-     *
-     * ThreadR calls YieldWithoutLoadBalancing
-     *
-     * ThreadR is moved to the end of ready_queue[prio=0]:
-     * ready_queue[prio=0]: ThreadA, ThreadB, ThreadC, ThreadR (->exec order->)
-     * Currently Running: Nothing
-     *
-     * System is rescheduled (ThreadA is popped off of queue):
-     * ready_queue[prio=0]: ThreadB, ThreadC, ThreadR (->exec order->)
-     * Currently Running: ThreadA
-     *
-     * If the queue is empty at time of call, no yielding occurs. This does not cross between cores
-     * or priorities at all.
-     */
-    void YieldWithoutLoadBalancing(Thread* thread);
-
-    /**
-     * YieldWithLoadBalancing -- yield but with better selection of the new running thread
-     * Moves the current thread to the end of the ready queue for its priority, then selects a
-     * 'suggested thread' (a thread on a different core that could run on this core) from the
-     * scheduler, changes its core, and reschedules the current core to that thread.
-     *
-     * Example (Dual Core -- can be extrapolated to Quad Core, this is just normal yield if it were
-     * single core):
-     * ready_queue[core=0][prio=0]: ThreadA, ThreadB (affinities not pictured as irrelevant
-     * ready_queue[core=1][prio=0]: ThreadC[affinity=both], ThreadD[affinity=core1only]
-     * Currently Running: ThreadQ on Core 0 || ThreadP on Core 1
-     *
-     * ThreadQ calls YieldWithLoadBalancing
-     *
-     * ThreadQ is moved to the end of ready_queue[core=0][prio=0]:
-     * ready_queue[core=0][prio=0]: ThreadA, ThreadB
-     * ready_queue[core=1][prio=0]: ThreadC[affinity=both], ThreadD[affinity=core1only]
-     * Currently Running: ThreadQ on Core 0 || ThreadP on Core 1
-     *
-     * A list of suggested threads for each core is compiled
-     * Suggested Threads: {ThreadC on Core 1}
-     * If this were quad core (as the switch is), there could be between 0 and 3 threads in this
-     * list. If there are more than one, the thread is selected by highest prio.
-     *
-     * ThreadC is core changed to Core 0:
-     * ready_queue[core=0][prio=0]: ThreadC, ThreadA, ThreadB, ThreadQ
-     * ready_queue[core=1][prio=0]: ThreadD
-     * Currently Running: None on Core 0 || ThreadP on Core 1
-     *
-     * System is rescheduled (ThreadC is popped off of queue):
-     * ready_queue[core=0][prio=0]: ThreadA, ThreadB, ThreadQ
-     * ready_queue[core=1][prio=0]: ThreadD
-     * Currently Running: ThreadC on Core 0 || ThreadP on Core 1
-     *
-     * If no suggested threads can be found this will behave just as normal yield. If there are
-     * multiple candidates for the suggested thread on a core, the highest prio is taken.
-     */
-    void YieldWithLoadBalancing(Thread* thread);
-
-    /// Currently unknown -- asserts as unimplemented on call
-    void YieldAndWaitForLoadBalancing(Thread* thread);
+    void RemoveThread(const Thread* thread);

    /// Returns a list of all threads managed by the scheduler
    const std::vector<SharedPtr<Thread>>& GetThreadList() const {
        return thread_list;
    }

-private:
-    /**
-     * Pops and returns the next thread from the thread queue
-     * @return A pointer to the next ready thread
-     */
-    Thread* PopNextReadyThread();
+    // Add a thread to the suggested queue of a cpu core. Suggested threads may be
+    // picked if no thread is scheduled to run on the core.
+    void Suggest(u32 priority, u32 core, Thread* thread) {
+        suggested_queue[core].add(thread, priority);
+    }

+    // Remove a thread to the suggested queue of a cpu core. Suggested threads may be
+    // picked if no thread is scheduled to run on the core.
+    void Unsuggest(u32 priority, u32 core, Thread* thread) {
+        suggested_queue[core].remove(thread, priority);
+    }
+
+    // Add a thread to the scheduling queue of a cpu core. The thread is added at the
+    // back the queue in its priority level
+    void Schedule(u32 priority, u32 core, Thread* thread);
+
+    // Add a thread to the scheduling queue of a cpu core. The thread is added at the
+    // front the queue in its priority level
+    void SchedulePrepend(u32 priority, u32 core, Thread* thread);
+
+    // Reschedule an already scheduled thread based on a new priority
+    void Reschedule(u32 priority, u32 core, Thread* thread) {
+        scheduled_queue[core].remove(thread, priority);
+        scheduled_queue[core].add(thread, priority);
+    }
+
+    // Unschedule a thread.
+    void Unschedule(u32 priority, u32 core, Thread* thread) {
+        scheduled_queue[core].remove(thread, priority);
+    }
+
+    // Transfers a thread into an specific core. If the destination_core is -1
+    // it will be unscheduled from its source code and added into its suggested
+    // queue.
+    void TransferToCore(u32 priority, s32 destination_core, Thread* thread) {
+        const bool schedulable = thread->GetPriority() < THREADPRIO_COUNT;
+        const s32 source_core = thread->GetProcessorID();
+        if (source_core == destination_core || !schedulable) {
+            return;
+        }
+        thread->SetProcessorID(destination_core);
+        if (source_core >= 0) {
+            Unschedule(priority, source_core, thread);
+        }
+        if (destination_core >= 0) {
+            Unsuggest(priority, destination_core, thread);
+            Schedule(priority, destination_core, thread);
+        }
+        if (source_core >= 0) {
+            Suggest(priority, source_core, thread);
+        }
+    }
+
+    /*
+     * UnloadThread selects a core and forces it to unload its current thread's context
+     */
+    void UnloadThread(s32 core);
+
+    /*
+     * SelectThread takes care of selecting the new scheduled thread.
+     * It does it in 3 steps:
+     * - First a thread is selected from the top of the priority queue. If no thread
+     * is obtained then we move to step two, else we are done.
+     * - Second we try to get a suggested thread that's not assigned to any core or
+     * that is not the top thread in that core.
+     * - Third is no suggested thread is found, we do a second pass and pick a running
+     * thread in another core and swap it with its current thread.
+     */
+    void SelectThread(u32 core);
+
+    bool HaveReadyThreads(u32 core_id) const {
+        return !scheduled_queue[core_id].empty();
+    }
+
+    /*
+     * YieldThread takes a thread and moves it to the back of the it's priority list
+     * This operation can be redundant and no scheduling is changed if marked as so.
+     */
+    void YieldThread(Thread* thread);
+
+    /*
+     * YieldThreadAndBalanceLoad takes a thread and moves it to the back of the it's priority list.
+     * Afterwards, tries to pick a suggested thread from the suggested queue that has worse time or
+     * a better priority than the next thread in the core.
+     * This operation can be redundant and no scheduling is changed if marked as so.
+     */
+    void YieldThreadAndBalanceLoad(Thread* thread);
+
+    /*
+     * YieldThreadAndWaitForLoadBalancing takes a thread and moves it out of the scheduling queue
+     * and into the suggested queue. If no thread can be squeduled afterwards in that core,
+     * a suggested thread is obtained instead.
+     * This operation can be redundant and no scheduling is changed if marked as so.
+     */
+    void YieldThreadAndWaitForLoadBalancing(Thread* thread);
+
+    u32 CpuCoresCount() const {
+        return NUM_CPU_CORES;
+    }
+
+    void SetReselectionPending() {
+        reselection_pending.store(true, std::memory_order_release);
+    }
+
+    bool IsReselectionPending() const {
+        return reselection_pending.load();
+    }
+
+private:
+    void AskForReselectionOrMarkRedundant(Thread* current_thread, Thread* winner);
+
+    static constexpr u32 min_regular_priority = 2;
+    std::array<Common::MultiLevelQueue<Thread*, THREADPRIO_COUNT>, NUM_CPU_CORES> scheduled_queue;
+    std::array<Common::MultiLevelQueue<Thread*, THREADPRIO_COUNT>, NUM_CPU_CORES> suggested_queue;
+    std::atomic<bool> reselection_pending;
+
+    /// Lists all thread ids that aren't deleted/etc.
+    std::vector<SharedPtr<Thread>> thread_list;
+    Core::System& system;
+};
+
+class Scheduler final {
+public:
+    explicit Scheduler(Core::System& system, Core::ARM_Interface& cpu_core, const u32 core_id);
+    ~Scheduler();
+
+    /// Returns whether there are any threads that are ready to run.
+    bool HaveReadyThreads() const;
+
+    /// Reschedules to the next available thread (call after current thread is suspended)
+    void TryDoContextSwitch();
+
+    void UnloadThread();
+
+    void SelectThreads();
+
+    /// Gets the current running thread
+    Thread* GetCurrentThread() const;
+
+    Thread* GetSelectedThread() const;
+
+    /// Gets the timestamp for the last context switch in ticks.
+    u64 GetLastContextSwitchTicks() const;
+
+    bool ContextSwitchPending() const {
+        return context_switch_pending;
+    }
+
+private:
+    friend class GlobalScheduler;
    /**
     * Switches the CPU's active thread context to that of the specified thread
     * @param new_thread The thread to switch to
     */
-    void SwitchContext(Thread* new_thread);
+    void SwitchContext();

    /**
     * Called on every context switch to update the internal timestamp
@@ -152,19 +206,16 @@ private:
     */
    void UpdateLastContextSwitchTime(Thread* thread, Process* process);

-    /// Lists all thread ids that aren't deleted/etc.
-    std::vector<SharedPtr<Thread>> thread_list;
-
-    /// Lists only ready thread ids.
-    Common::MultiLevelQueue<Thread*, THREADPRIO_LOWEST + 1> ready_queue;
-
    SharedPtr<Thread> current_thread = nullptr;
-
-    Core::ARM_Interface& cpu_core;
-    u64 last_context_switch_time = 0;
+    SharedPtr<Thread> selected_thread = nullptr;

    Core::System& system;
-    static std::mutex scheduler_mutex;
+    Core::ARM_Interface& cpu_core;
+    u64 last_context_switch_time = 0;
+    u64 idle_selection_count = 0;
+    const u32 core_id;
+
+    bool context_switch_pending = false;
 };

 } // namespace Kernel
@@ -516,7 +516,7 @@ static ResultCode WaitSynchronization(Core::System& system, Handle* index, VAddr
    thread->WakeAfterDelay(nano_seconds);
    thread->SetWakeupCallback(DefaultThreadWakeupCallback);

-    system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(thread->GetProcessorID());

    return RESULT_TIMEOUT;
 }
@@ -534,6 +534,7 @@ static ResultCode CancelSynchronization(Core::System& system, Handle thread_hand
    }

    thread->CancelWait();
+    system.PrepareReschedule(thread->GetProcessorID());
    return RESULT_SUCCESS;
 }

@@ -577,7 +578,8 @@ static ResultCode ArbitrateUnlock(Core::System& system, VAddr mutex_addr) {
    }

    auto* const current_process = system.Kernel().CurrentProcess();
-    return current_process->GetMutex().Release(mutex_addr);
+    return current_process->GetMutex().Release(mutex_addr,
+                                               system.CurrentScheduler().GetCurrentThread());
 }

 enum class BreakType : u32 {
@@ -1066,6 +1068,8 @@ static ResultCode SetThreadActivity(Core::System& system, Handle handle, u32 act
    }

    thread->SetActivity(static_cast<ThreadActivity>(activity));
+
+    system.PrepareReschedule(thread->GetProcessorID());
    return RESULT_SUCCESS;
 }

@@ -1147,7 +1151,7 @@ static ResultCode SetThreadPriority(Core::System& system, Handle handle, u32 pri

    thread->SetPriority(priority);

-    system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(thread->GetProcessorID());
    return RESULT_SUCCESS;
 }

@@ -1503,7 +1507,7 @@ static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr e
    thread->SetName(
        fmt::format("thread[entry_point={:X}, handle={:X}]", entry_point, *new_thread_handle));

-    system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(thread->GetProcessorID());

    return RESULT_SUCCESS;
 }
@@ -1525,7 +1529,7 @@ static ResultCode StartThread(Core::System& system, Handle thread_handle) {
    thread->ResumeFromWait();

    if (thread->GetStatus() == ThreadStatus::Ready) {
-        system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+        system.PrepareReschedule(thread->GetProcessorID());
    }

    return RESULT_SUCCESS;
@@ -1537,7 +1541,7 @@ static void ExitThread(Core::System& system) {

    auto* const current_thread = system.CurrentScheduler().GetCurrentThread();
    current_thread->Stop();
-    system.CurrentScheduler().RemoveThread(current_thread);
+    system.GlobalScheduler().RemoveThread(current_thread);
    system.PrepareReschedule();
 }

@@ -1557,13 +1561,13 @@ static void SleepThread(Core::System& system, s64 nanoseconds) {
    if (nanoseconds <= 0) {
        switch (static_cast<SleepType>(nanoseconds)) {
        case SleepType::YieldWithoutLoadBalancing:
-            scheduler.YieldWithoutLoadBalancing(current_thread);
+            current_thread->YieldSimple();
            break;
        case SleepType::YieldWithLoadBalancing:
-            scheduler.YieldWithLoadBalancing(current_thread);
+            current_thread->YieldAndBalanceLoad();
            break;
        case SleepType::YieldAndWaitForLoadBalancing:
-            scheduler.YieldAndWaitForLoadBalancing(current_thread);
+            current_thread->YieldAndWaitForLoadBalancing();
            break;
        default:
            UNREACHABLE_MSG("Unimplemented sleep yield type '{:016X}'!", nanoseconds);
@@ -1572,10 +1576,7 @@ static void SleepThread(Core::System& system, s64 nanoseconds) {
        current_thread->Sleep(nanoseconds);
    }

-    // Reschedule all CPU cores
-    for (std::size_t i = 0; i < Core::NUM_CPU_CORES; ++i) {
-        system.CpuCore(i).PrepareReschedule();
-    }
+    system.PrepareReschedule(current_thread->GetProcessorID());
 }

 /// Wait process wide key atomic
@@ -1606,12 +1607,14 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add
    SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
    ASSERT(thread);

-    const auto release_result = current_process->GetMutex().Release(mutex_addr);
+    SharedPtr<Thread> current_thread = system.CurrentScheduler().GetCurrentThread();
+
+    const auto release_result =
+        current_process->GetMutex().Release(mutex_addr, current_thread.get());
    if (release_result.IsError()) {
        return release_result;
    }

-    SharedPtr<Thread> current_thread = system.CurrentScheduler().GetCurrentThread();
    current_thread->SetCondVarWaitAddress(condition_variable_addr);
    current_thread->SetMutexWaitAddress(mutex_addr);
    current_thread->SetWaitHandle(thread_handle);
@@ -1622,7 +1625,7 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add

    // Note: Deliberately don't attempt to inherit the lock owner's priority.

-    system.CpuCore(current_thread->GetProcessorID()).PrepareReschedule();
+    system.PrepareReschedule(current_thread->GetProcessorID());
    return RESULT_SUCCESS;
 }

@@ -1632,24 +1635,17 @@ static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_var
    LOG_TRACE(Kernel_SVC, "called, condition_variable_addr=0x{:X}, target=0x{:08X}",
              condition_variable_addr, target);

-    const auto RetrieveWaitingThreads = [&system](std::size_t core_index,
-                                                  std::vector<SharedPtr<Thread>>& waiting_threads,
-                                                  VAddr condvar_addr) {
-        const auto& scheduler = system.Scheduler(core_index);
-        const auto& thread_list = scheduler.GetThreadList();
-
-        for (const auto& thread : thread_list) {
-            if (thread->GetCondVarWaitAddress() == condvar_addr)
-                waiting_threads.push_back(thread);
-        }
-    };
-
    // Retrieve a list of all threads that are waiting for this condition variable.
    std::vector<SharedPtr<Thread>> waiting_threads;
-    RetrieveWaitingThreads(0, waiting_threads, condition_variable_addr);
-    RetrieveWaitingThreads(1, waiting_threads, condition_variable_addr);
-    RetrieveWaitingThreads(2, waiting_threads, condition_variable_addr);
-    RetrieveWaitingThreads(3, waiting_threads, condition_variable_addr);
+    const auto& scheduler = system.GlobalScheduler();
+    const auto& thread_list = scheduler.GetThreadList();
+
+    for (const auto& thread : thread_list) {
+        if (thread->GetCondVarWaitAddress() == condition_variable_addr) {
+            waiting_threads.push_back(thread);
+        }
+    }
+
    // Sort them by priority, such that the highest priority ones come first.
    std::sort(waiting_threads.begin(), waiting_threads.end(),
              [](const SharedPtr<Thread>& lhs, const SharedPtr<Thread>& rhs) {
@@ -1704,7 +1700,7 @@ static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_var
            thread->SetLockOwner(nullptr);
            thread->SetMutexWaitAddress(0);
            thread->SetWaitHandle(0);
-            system.CpuCore(thread->GetProcessorID()).PrepareReschedule();
+            system.PrepareReschedule(thread->GetProcessorID());
        } else {
            // Atomically signal that the mutex now has a waiting thread.
            do {
@@ -1728,6 +1724,7 @@ static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_var
            thread->SetStatus(ThreadStatus::WaitMutex);

            owner->AddMutexWaiter(thread);
+            system.PrepareReschedule(thread->GetProcessorID());
        }
    }

@@ -1754,7 +1751,12 @@ static ResultCode WaitForAddress(Core::System& system, VAddr address, u32 type,

    const auto arbitration_type = static_cast<AddressArbiter::ArbitrationType>(type);
    auto& address_arbiter = system.Kernel().CurrentProcess()->GetAddressArbiter();
-    return address_arbiter.WaitForAddress(address, arbitration_type, value, timeout);
+    const ResultCode result =
+        address_arbiter.WaitForAddress(address, arbitration_type, value, timeout);
+    if (result == RESULT_SUCCESS) {
+        system.PrepareReschedule();
+    }
+    return result;
 }

 // Signals to an address (via Address Arbiter)
@@ -2040,7 +2042,10 @@ static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle,
        return ERR_INVALID_HANDLE;
    }

+    system.PrepareReschedule(thread->GetProcessorID());
    thread->ChangeCore(core, affinity_mask);
+    system.PrepareReschedule(thread->GetProcessorID());
+
    return RESULT_SUCCESS;
 }

@@ -2151,6 +2156,7 @@ static ResultCode SignalEvent(Core::System& system, Handle handle) {
    }

    writable_event->Signal();
+    system.PrepareReschedule();
    return RESULT_SUCCESS;
 }

@@ -45,15 +45,7 @@ void Thread::Stop() {
                                                             callback_handle);
    kernel.ThreadWakeupCallbackHandleTable().Close(callback_handle);
    callback_handle = 0;
-
-    // Clean up thread from ready queue
-    // This is only needed when the thread is terminated forcefully (SVC TerminateProcess)
-    if (status == ThreadStatus::Ready || status == ThreadStatus::Paused) {
-        scheduler->UnscheduleThread(this, current_priority);
-    }
-
-    status = ThreadStatus::Dead;
-
+    SetStatus(ThreadStatus::Dead);
    WakeupAllWaitingThreads();

    // Clean up any dangling references in objects that this thread was waiting for
@@ -132,13 +124,11 @@ void Thread::ResumeFromWait() {
    wakeup_callback = nullptr;

    if (activity == ThreadActivity::Paused) {
-        status = ThreadStatus::Paused;
+        SetStatus(ThreadStatus::Paused);
        return;
    }

-    status = ThreadStatus::Ready;
-
-    ChangeScheduler();
+    SetStatus(ThreadStatus::Ready);
 }

 void Thread::CancelWait() {
@@ -205,9 +195,9 @@ ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name
    thread->name = std::move(name);
    thread->callback_handle = kernel.ThreadWakeupCallbackHandleTable().Create(thread).Unwrap();
    thread->owner_process = &owner_process;
+    auto& scheduler = kernel.GlobalScheduler();
+    scheduler.AddThread(thread);
    thread->tls_address = thread->owner_process->CreateTLSRegion();
-    thread->scheduler = &system.Scheduler(processor_id);
-    thread->scheduler->AddThread(thread);

    thread->owner_process->RegisterThread(thread.get());

@@ -250,6 +240,22 @@ void Thread::SetStatus(ThreadStatus new_status) {
        return;
    }

+    switch (new_status) {
+    case ThreadStatus::Ready:
+    case ThreadStatus::Running:
+        SetSchedulingStatus(ThreadSchedStatus::Runnable);
+        break;
+    case ThreadStatus::Dormant:
+        SetSchedulingStatus(ThreadSchedStatus::None);
+        break;
+    case ThreadStatus::Dead:
+        SetSchedulingStatus(ThreadSchedStatus::Exited);
+        break;
+    default:
+        SetSchedulingStatus(ThreadSchedStatus::Paused);
+        break;
+    }
+
    if (status == ThreadStatus::Running) {
        last_running_ticks = Core::System::GetInstance().CoreTiming().GetTicks();
    }
@@ -311,8 +317,7 @@ void Thread::UpdatePriority() {
        return;
    }

-    scheduler->SetThreadPriority(this, new_priority);
-    current_priority = new_priority;
+    SetCurrentPriority(new_priority);

    if (!lock_owner) {
        return;
@@ -328,47 +333,7 @@ void Thread::UpdatePriority() {
 }

 void Thread::ChangeCore(u32 core, u64 mask) {
-    ideal_core = core;
-    affinity_mask = mask;
-    ChangeScheduler();
-}
-
-void Thread::ChangeScheduler() {
-    if (status != ThreadStatus::Ready) {
-        return;
-    }
-
-    auto& system = Core::System::GetInstance();
-    std::optional<s32> new_processor_id{GetNextProcessorId(affinity_mask)};
-
-    if (!new_processor_id) {
-        new_processor_id = processor_id;
-    }
-    if (ideal_core != -1 && system.Scheduler(ideal_core).GetCurrentThread() == nullptr) {
-        new_processor_id = ideal_core;
-    }
-
-    ASSERT(*new_processor_id < 4);
-
-    // Add thread to new core's scheduler
-    auto& next_scheduler = system.Scheduler(*new_processor_id);
-
-    if (*new_processor_id != processor_id) {
-        // Remove thread from previous core's scheduler
-        scheduler->RemoveThread(this);
-        next_scheduler.AddThread(this);
-    }
-
-    processor_id = *new_processor_id;
-
-    // If the thread was ready, unschedule from the previous core and schedule on the new core
-    scheduler->UnscheduleThread(this, current_priority);
-    next_scheduler.ScheduleThread(this, current_priority);
-
-    // Change thread's scheduler
-    scheduler = &next_scheduler;
-
-    system.CpuCore(processor_id).PrepareReschedule();
+    SetCoreAndAffinityMask(core, mask);
 }

 bool Thread::AllWaitObjectsReady() const {
@@ -391,7 +356,7 @@ void Thread::SetActivity(ThreadActivity value) {
        if (status == ThreadStatus::Ready) {
            status = ThreadStatus::Paused;
        } else if (status == ThreadStatus::Running) {
-            status = ThreadStatus::Paused;
+            SetStatus(ThreadStatus::Paused);
            Core::System::GetInstance().CpuCore(processor_id).PrepareReschedule();
        }
    } else if (status == ThreadStatus::Paused) {
@@ -408,6 +373,171 @@ void Thread::Sleep(s64 nanoseconds) {
    WakeAfterDelay(nanoseconds);
 }

+void Thread::YieldSimple() {
+    auto& scheduler = kernel.GlobalScheduler();
+    scheduler.YieldThread(this);
+}
+
+void Thread::YieldAndBalanceLoad() {
+    auto& scheduler = kernel.GlobalScheduler();
+    scheduler.YieldThreadAndBalanceLoad(this);
+}
+
+void Thread::YieldAndWaitForLoadBalancing() {
+    auto& scheduler = kernel.GlobalScheduler();
+    scheduler.YieldThreadAndWaitForLoadBalancing(this);
+}
+
+void Thread::SetSchedulingStatus(ThreadSchedStatus new_status) {
+    const u32 old_flags = scheduling_state;
+    scheduling_state =
+        (scheduling_state & ThreadSchedMasks::HighMask) | static_cast<u32>(new_status);
+    AdjustSchedulingOnStatus(old_flags);
+}
+
+void Thread::SetCurrentPriority(u32 new_priority) {
+    u32 old_priority = std::exchange(current_priority, new_priority);
+    AdjustSchedulingOnPriority(old_priority);
+}
+
+ResultCode Thread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) {
+    const auto HighestSetCore = [](u64 mask, u32 max_cores) {
+        for (s32 core = max_cores - 1; core >= 0; core--) {
+            if (((mask >> core) & 1) != 0) {
+                return core;
+            }
+        }
+        return -1;
+    };
+
+    const bool use_override = affinity_override_count != 0;
+    // The value -3 is "do not change the ideal core".
+    if (new_core == -3) {
+        new_core = use_override ? ideal_core_override : ideal_core;
+        if ((new_affinity_mask & (1 << new_core)) == 0) {
+            return ERR_INVALID_COMBINATION;
+        }
+    }
+    if (use_override) {
+        ideal_core_override = new_core;
+        affinity_mask_override = new_affinity_mask;
+    } else {
+        const u64 old_affinity_mask = std::exchange(affinity_mask, new_affinity_mask);
+        ideal_core = new_core;
+        if (old_affinity_mask != new_affinity_mask) {
+            const s32 old_core = processor_id;
+            if (processor_id >= 0 && ((affinity_mask >> processor_id) & 1) == 0) {
+                if (ideal_core < 0) {
+                    processor_id = HighestSetCore(affinity_mask, GlobalScheduler::NUM_CPU_CORES);
+                } else {
+                    processor_id = ideal_core;
+                }
+            }
+            AdjustSchedulingOnAffinity(old_affinity_mask, old_core);
+        }
+    }
+    return RESULT_SUCCESS;
+}
+
+void Thread::AdjustSchedulingOnStatus(u32 old_flags) {
+    if (old_flags == scheduling_state) {
+        return;
+    }
+
+    auto& scheduler = kernel.GlobalScheduler();
+    if (static_cast<ThreadSchedStatus>(old_flags & ThreadSchedMasks::LowMask) ==
+        ThreadSchedStatus::Runnable) {
+        // In this case the thread was running, now it's pausing/exitting
+        if (processor_id >= 0) {
+            scheduler.Unschedule(current_priority, processor_id, this);
+        }
+
+        for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+            if (core != processor_id && ((affinity_mask >> core) & 1) != 0) {
+                scheduler.Unsuggest(current_priority, core, this);
+            }
+        }
+    } else if (GetSchedulingStatus() == ThreadSchedStatus::Runnable) {
+        // The thread is now set to running from being stopped
+        if (processor_id >= 0) {
+            scheduler.Schedule(current_priority, processor_id, this);
+        }
+
+        for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+            if (core != processor_id && ((affinity_mask >> core) & 1) != 0) {
+                scheduler.Suggest(current_priority, core, this);
+            }
+        }
+    }
+
+    scheduler.SetReselectionPending();
+}
+
+void Thread::AdjustSchedulingOnPriority(u32 old_priority) {
+    if (GetSchedulingStatus() != ThreadSchedStatus::Runnable) {
+        return;
+    }
+    auto& scheduler = Core::System::GetInstance().GlobalScheduler();
+    if (processor_id >= 0) {
+        scheduler.Unschedule(old_priority, processor_id, this);
+    }
+
+    for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+        if (core != processor_id && ((affinity_mask >> core) & 1) != 0) {
+            scheduler.Unsuggest(old_priority, core, this);
+        }
+    }
+
+    // Add thread to the new priority queues.
+    Thread* current_thread = GetCurrentThread();
+
+    if (processor_id >= 0) {
+        if (current_thread == this) {
+            scheduler.SchedulePrepend(current_priority, processor_id, this);
+        } else {
+            scheduler.Schedule(current_priority, processor_id, this);
+        }
+    }
+
+    for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+        if (core != processor_id && ((affinity_mask >> core) & 1) != 0) {
+            scheduler.Suggest(current_priority, core, this);
+        }
+    }
+
+    scheduler.SetReselectionPending();
+}
+
+void Thread::AdjustSchedulingOnAffinity(u64 old_affinity_mask, s32 old_core) {
+    auto& scheduler = Core::System::GetInstance().GlobalScheduler();
+    if (GetSchedulingStatus() != ThreadSchedStatus::Runnable ||
+        current_priority >= THREADPRIO_COUNT) {
+        return;
+    }
+
+    for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+        if (((old_affinity_mask >> core) & 1) != 0) {
+            if (core == old_core) {
+                scheduler.Unschedule(current_priority, core, this);
+            } else {
+                scheduler.Unsuggest(current_priority, core, this);
+            }
+        }
+    }
+
+    for (u32 core = 0; core < GlobalScheduler::NUM_CPU_CORES; core++) {
+        if (((affinity_mask >> core) & 1) != 0) {
+            if (core == processor_id) {
+                scheduler.Schedule(current_priority, core, this);
+            } else {
+                scheduler.Suggest(current_priority, core, this);
+            }
+        }
+    }
+
+    scheduler.SetReselectionPending();
+}
+
 ////////////////////////////////////////////////////////////////////////////////////////////////////

 /**
@@ -75,6 +75,26 @@ enum class ThreadActivity : u32 {
    Paused = 1,
 };

+enum class ThreadSchedStatus : u32 {
+    None = 0,
+    Paused = 1,
+    Runnable = 2,
+    Exited = 3,
+};
+
+enum ThreadSchedFlags : u32 {
+    ProcessPauseFlag = 1 << 4,
+    ThreadPauseFlag = 1 << 5,
+    ProcessDebugPauseFlag = 1 << 6,
+    KernelInitPauseFlag = 1 << 8,
+};
+
+enum ThreadSchedMasks : u32 {
+    LowMask = 0x000f,
+    HighMask = 0xfff0,
+    ForcePauseMask = 0x0070,
+};
+
 class Thread final : public WaitObject {
 public:
    using MutexWaitingThreads = std::vector<SharedPtr<Thread>>;
@@ -278,6 +298,10 @@ public:
        return processor_id;
    }

+    void SetProcessorID(s32 new_core) {
+        processor_id = new_core;
+    }
+
    Process* GetOwnerProcess() {
        return owner_process;
    }
@@ -383,11 +407,38 @@ public:
    /// Sleeps this thread for the given amount of nanoseconds.
    void Sleep(s64 nanoseconds);

+    /// Yields this thread without rebalancing loads.
+    void YieldSimple();
+
+    /// Yields this thread and does a load rebalancing.
+    void YieldAndBalanceLoad();
+
+    /// Yields this thread and if the core is left idle, loads are rebalanced
+    void YieldAndWaitForLoadBalancing();
+
+    ThreadSchedStatus GetSchedulingStatus() const {
+        return static_cast<ThreadSchedStatus>(scheduling_state & ThreadSchedMasks::LowMask);
+    }
+
+    bool IsRunning() const {
+        return is_running;
+    }
+
+    void SetIsRunning(bool value) {
+        is_running = value;
+    }
+
 private:
    explicit Thread(KernelCore& kernel);
    ~Thread() override;

-    void ChangeScheduler();
+    void SetSchedulingStatus(ThreadSchedStatus new_status);
+    void SetCurrentPriority(u32 new_priority);
+    ResultCode SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask);
+
+    void AdjustSchedulingOnStatus(u32 old_flags);
+    void AdjustSchedulingOnPriority(u32 old_priority);
+    void AdjustSchedulingOnAffinity(u64 old_affinity_mask, s32 old_core);

    Core::ARM_Interface::ThreadContext context{};

@@ -453,6 +504,13 @@ private:

    ThreadActivity activity = ThreadActivity::Normal;

+    s32 ideal_core_override = -1;
+    u64 affinity_mask_override = 0x1;
+    u32 affinity_override_count = 0;
+
+    u32 scheduling_state = 0;
+    bool is_running = false;
+
    std::string name;
 };

@@ -23,6 +23,8 @@ SharedPtr<TransferMemory> TransferMemory::Create(KernelCore& kernel, VAddr base_
    transfer_memory->owner_permissions = permissions;
    transfer_memory->owner_process = kernel.CurrentProcess();

+    transfer_memory->MapMemory(base_address, size, permissions);
+
    return transfer_memory;
 }

@@ -296,12 +296,6 @@ ResultVal<VAddr> VMManager::SetHeapSize(u64 size) {
 }

 ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
-    const auto end_addr = target + size;
-    const auto last_addr = end_addr - 1;
-    VAddr cur_addr = target;
-
-    ResultCode result = RESULT_SUCCESS;
-
    // Check how much memory we've already mapped.
    const auto mapped_size_result = SizeOfAllocatedVMAsInRange(target, size);
    if (mapped_size_result.Failed()) {
@@ -324,13 +318,16 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {

    // Keep track of the memory regions we unmap.
    std::vector<std::pair<u64, u64>> mapped_regions;
+    ResultCode result = RESULT_SUCCESS;

    // Iterate, trying to map memory.
    {
-        cur_addr = target;
+        const auto end_addr = target + size;
+        const auto last_addr = end_addr - 1;
+        VAddr cur_addr = target;

        auto iter = FindVMA(target);
-        ASSERT_MSG(iter != vma_map.end(), "MapPhysicalMemory iter != end");
+        ASSERT(iter != vma_map.end());

        while (true) {
            const auto& vma = iter->second;
@@ -342,7 +339,7 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
            const auto map_size = std::min(end_addr - cur_addr, vma_end - cur_addr);
            if (vma.state == MemoryState::Unmapped) {
                const auto map_res =
-                    MapMemoryBlock(cur_addr, std::make_shared<PhysicalMemory>(map_size, 0), 0,
+                    MapMemoryBlock(cur_addr, std::make_shared<PhysicalMemory>(map_size), 0,
                                   map_size, MemoryState::Heap, VMAPermission::ReadWrite);
                result = map_res.Code();
                if (result.IsError()) {
@@ -360,7 +357,7 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
            // Advance to the next block.
            cur_addr = vma_end;
            iter = FindVMA(cur_addr);
-            ASSERT_MSG(iter != vma_map.end(), "MapPhysicalMemory iter != end");
+            ASSERT(iter != vma_map.end());
        }
    }

@@ -368,7 +365,7 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
    if (result.IsError()) {
        for (const auto [unmap_address, unmap_size] : mapped_regions) {
            ASSERT_MSG(UnmapRange(unmap_address, unmap_size).IsSuccess(),
-                       "MapPhysicalMemory un-map on error");
+                       "Failed to unmap memory range.");
        }

        return result;
@@ -381,12 +378,6 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
 }

 ResultCode VMManager::UnmapPhysicalMemory(VAddr target, u64 size) {
-    const auto end_addr = target + size;
-    const auto last_addr = end_addr - 1;
-    VAddr cur_addr = target;
-
-    ResultCode result = RESULT_SUCCESS;
-
    // Check how much memory is currently mapped.
    const auto mapped_size_result = SizeOfUnmappablePhysicalMemoryInRange(target, size);
    if (mapped_size_result.Failed()) {
@@ -401,13 +392,16 @@ ResultCode VMManager::UnmapPhysicalMemory(VAddr target, u64 size) {

    // Keep track of the memory regions we unmap.
    std::vector<std::pair<u64, u64>> unmapped_regions;
+    ResultCode result = RESULT_SUCCESS;

    // Try to unmap regions.
    {
-        cur_addr = target;
+        const auto end_addr = target + size;
+        const auto last_addr = end_addr - 1;
+        VAddr cur_addr = target;

        auto iter = FindVMA(target);
-        ASSERT_MSG(iter != vma_map.end(), "UnmapPhysicalMemory iter != end");
+        ASSERT(iter != vma_map.end());

        while (true) {
            const auto& vma = iter->second;
@@ -434,7 +428,7 @@ ResultCode VMManager::UnmapPhysicalMemory(VAddr target, u64 size) {
            // Advance to the next block.
            cur_addr = vma_end;
            iter = FindVMA(cur_addr);
-            ASSERT_MSG(iter != vma_map.end(), "UnmapPhysicalMemory iter != end");
+            ASSERT(iter != vma_map.end());
        }
    }

@@ -443,10 +437,12 @@ ResultCode VMManager::UnmapPhysicalMemory(VAddr target, u64 size) {
    if (result.IsError()) {
        for (const auto [map_address, map_size] : unmapped_regions) {
            const auto remap_res =
-                MapMemoryBlock(map_address, std::make_shared<PhysicalMemory>(map_size, 0), 0,
-                               map_size, MemoryState::Heap, VMAPermission::None);
-            ASSERT_MSG(remap_res.Succeeded(), "UnmapPhysicalMemory re-map on error");
+                MapMemoryBlock(map_address, std::make_shared<PhysicalMemory>(map_size), 0, map_size,
+                               MemoryState::Heap, VMAPermission::None);
+            ASSERT_MSG(remap_res.Succeeded(), "Failed to remap a memory block.");
        }
+
+        return result;
    }

    // Update mapped amount
@@ -757,20 +753,26 @@ void VMManager::MergeAdjacentVMA(VirtualMemoryArea& left, const VirtualMemoryAre
    // Always merge allocated memory blocks, even when they don't share the same backing block.
    if (left.type == VMAType::AllocatedMemoryBlock &&
        (left.backing_block != right.backing_block || left.offset + left.size != right.offset)) {
+        const auto right_begin = right.backing_block->begin() + right.offset;
+        const auto right_end = right_begin + right.size;
+
        // Check if we can save work.
        if (left.offset == 0 && left.size == left.backing_block->size()) {
            // Fast case: left is an entire backing block.
-            left.backing_block->insert(left.backing_block->end(),
-                                       right.backing_block->begin() + right.offset,
-                                       right.backing_block->begin() + right.offset + right.size);
+            left.backing_block->insert(left.backing_block->end(), right_begin, right_end);
        } else {
            // Slow case: make a new memory block for left and right.
+            const auto left_begin = left.backing_block->begin() + left.offset;
+            const auto left_end = left_begin + left.size;
+            const auto left_size = static_cast<std::size_t>(std::distance(left_begin, left_end));
+            const auto right_size = static_cast<std::size_t>(std::distance(right_begin, right_end));
+
            auto new_memory = std::make_shared<PhysicalMemory>();
-            new_memory->insert(new_memory->end(), left.backing_block->begin() + left.offset,
-                               left.backing_block->begin() + left.offset + left.size);
-            new_memory->insert(new_memory->end(), right.backing_block->begin() + right.offset,
-                               right.backing_block->begin() + right.offset + right.size);
-            left.backing_block = new_memory;
+            new_memory->reserve(left_size + right_size);
+            new_memory->insert(new_memory->end(), left_begin, left_end);
+            new_memory->insert(new_memory->end(), right_begin, right_end);
+
+            left.backing_block = std::move(new_memory);
            left.offset = 0;
        }

@@ -965,7 +967,7 @@ ResultVal<std::size_t> VMManager::SizeOfAllocatedVMAsInRange(VAddr address,

    VAddr cur_addr = address;
    auto iter = FindVMA(cur_addr);
-    ASSERT_MSG(iter != vma_map.end(), "SizeOfAllocatedVMAsInRange iter != end");
+    ASSERT(iter != vma_map.end());

    while (true) {
        const auto& vma = iter->second;
@@ -986,7 +988,7 @@ ResultVal<std::size_t> VMManager::SizeOfAllocatedVMAsInRange(VAddr address,
        // Advance to the next block.
        cur_addr = vma_end;
        iter = std::next(iter);
-        ASSERT_MSG(iter != vma_map.end(), "SizeOfAllocatedVMAsInRange iter != end");
+        ASSERT(iter != vma_map.end());
    }

    return MakeResult(mapped_size);
@@ -1000,7 +1002,7 @@ ResultVal<std::size_t> VMManager::SizeOfUnmappablePhysicalMemoryInRange(VAddr ad

    VAddr cur_addr = address;
    auto iter = FindVMA(cur_addr);
-    ASSERT_MSG(iter != vma_map.end(), "SizeOfUnmappablePhysicalMemoryInRange iter != end");
+    ASSERT(iter != vma_map.end());

    while (true) {
        const auto& vma = iter->second;
@@ -1029,7 +1031,7 @@ ResultVal<std::size_t> VMManager::SizeOfUnmappablePhysicalMemoryInRange(VAddr ad
        // Advance to the next block.
        cur_addr = vma_end;
        iter = std::next(iter);
-        ASSERT_MSG(iter != vma_map.end(), "SizeOfUnmappablePhysicalMemoryInRange iter != end");
+        ASSERT(iter != vma_map.end());
    }

    return MakeResult(mapped_size);
@@ -454,8 +454,8 @@ public:

    /// Maps memory at a given address.
    ///
-    /// @param addr The virtual address to map memory at.
-    /// @param size The amount of memory to map.
+    /// @param target The virtual address to map memory at.
+    /// @param size   The amount of memory to map.
    ///
    /// @note The destination address must lie within the Map region.
    ///
@@ -468,8 +468,8 @@ public:

    /// Unmaps memory at a given address.
    ///
-    /// @param addr The virtual address to unmap memory at.
-    /// @param size The amount of memory to unmap.
+    /// @param target The virtual address to unmap memory at.
+    /// @param size   The amount of memory to unmap.
    ///
    /// @note The destination address must lie within the Map region.
    ///
@@ -6,6 +6,8 @@
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "common/logging/log.h"
+#include "core/core.h"
+#include "core/core_cpu.h"
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/thread.h"
@@ -48,17 +50,8 @@ SharedPtr<Thread> WaitObject::GetHighestPriorityReadyThread() const {
        if (ShouldWait(thread.get()))
            continue;

-        // A thread is ready to run if it's either in ThreadStatus::WaitSynch
-        // and the rest of the objects it is waiting on are ready.
-        bool ready_to_run = true;
-        if (thread_status == ThreadStatus::WaitSynch) {
-            ready_to_run = thread->AllWaitObjectsReady();
-        }
-
-        if (ready_to_run) {
-            candidate = thread.get();
-            candidate_priority = thread->GetPriority();
-        }
+        candidate = thread.get();
+        candidate_priority = thread->GetPriority();
    }

    return candidate;
@@ -95,6 +88,7 @@ void WaitObject::WakeupWaitingThread(SharedPtr<Thread> thread) {
    }
    if (resume) {
        thread->ResumeFromWait();
+        Core::System::GetInstance().PrepareReschedule(thread->GetProcessorID());
    }
 }