Remove LGTM config

LGTM.com is no longer available since it was superseded by CodeQL.

.lgtm.yml

@@ -1,13 +0,0 @@
-# SPDX-FileCopyrightText: 2020 yuzu Emulator Project
-# SPDX-License-Identifier: GPL-2.0-or-later
-
-path_classifiers:
-  library: "externals"
-extraction:
-  cpp:
-    prepare:
-      packages:
-      - "libsdl2-dev"
-      - "qtmultimedia5-dev"
-      - "libtbb-dev"
-      - "libjack-jackd2-dev"

dist/languages/ca.ts

@@ -1372,8 +1372,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Interfície de memòria ampliada (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/cs.ts

@@ -1364,7 +1364,7 @@ Tato možnost zlepšuje rychlost díky závislosti na sémantice cmpxchg pro zaj
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
+        <source>Extended memory layout (8GB DRAM)</source>
         <translation type="unfinished"/>
     </message>
     <message>

dist/languages/da.ts

@@ -1380,8 +1380,8 @@ Dette vil bandlyse både vedkommendes forum-brugernavn og IP-adresse.</translati
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Udvidet hukommelsesopsætning (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/de.ts

@@ -1360,8 +1360,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Erweitertes Speicherlayout (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/el.ts

@@ -1364,8 +1364,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Διάταξη εκτεταμένης μνήμης (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/es.ts

@@ -381,17 +381,17 @@ Esto banearía su nombre del foro y su dirección IP.</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_audio.ui" line="42"/>
         <source>Output Device:</source>
-        <translation type="unfinished"/>
+        <translation>Dispositivo de salida:</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_audio.ui" line="56"/>
         <source>Input Device:</source>
-        <translation type="unfinished"/>
+        <translation>Dispositivo de entrada:</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_audio.ui" line="70"/>
         <source>Sound Output Mode:</source>
-        <translation type="unfinished"/>
+        <translation>Método de salida de sonido:</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_audio.ui" line="78"/>
@@ -1383,8 +1383,8 @@ Esto banearía su nombre del foro y su dirección IP.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Interfaz de memoria extendida (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation>Interfaz de memoria extendida (8GB DRAM)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>
@@ -1638,7 +1638,7 @@ Esto banearía su nombre del foro y su dirección IP.</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_graphics.ui" line="463"/>
         <source>AMD FidelityFX™️ Super Resolution</source>
-        <translation type="unfinished"/>
+        <translation>AMD FidelityFX™️ Super Resolution</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_graphics.ui" line="489"/>
@@ -1753,12 +1753,12 @@ Esto banearía su nombre del foro y su dirección IP.</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_graphics_advanced.ui" line="95"/>
         <source>Enables asynchronous ASTC texture decoding, which may reduce load time stutter. This feature is experimental.</source>
-        <translation type="unfinished"/>
+        <translation>Activa la decodificación de texturas asíncrona de ASTC, lo cuál podría reducir la duración de los parones. Esta función es experimental.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_graphics_advanced.ui" line="98"/>
         <source>Decode ASTC textures asynchronously (Hack)</source>
-        <translation type="unfinished"/>
+        <translation>Decodificar texturas ASTC de manera asíncrona (Hack)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_graphics_advanced.ui" line="105"/>
@@ -2268,12 +2268,12 @@ Esto banearía su nombre del foro y su dirección IP.</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_advanced.ui" line="2710"/>
         <source>Enable direct JoyCon driver</source>
-        <translation type="unfinished"/>
+        <translation>Activar driver directo JoyCon</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_advanced.ui" line="2726"/>
         <source>Enable direct Pro Controller driver [EXPERIMENTAL]</source>
-        <translation type="unfinished"/>
+        <translation>Activar driver directo Pro Controller [EXPERIMENTAL]</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_advanced.ui" line="2739"/>
@@ -2637,7 +2637,7 @@ Esto banearía su nombre del foro y su dirección IP.</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="399"/>
         <source>Turbo button</source>
-        <translation type="unfinished"/>
+        <translation>Botón Turbo</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="407"/>
@@ -3339,7 +3339,7 @@ UUID: %2</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="52"/>
         <source>Virtual Ring Sensor Parameters</source>
-        <translation type="unfinished"/>
+        <translation>Parámetros del sensor Ring virtual</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="84"/>
@@ -3361,29 +3361,29 @@ UUID: %2</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="233"/>
         <source>Direct Joycon Driver</source>
-        <translation type="unfinished"/>
+        <translation>Driver directo del JoyCon</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="293"/>
         <source>Enable Ring Input</source>
-        <translation type="unfinished"/>
+        <translation>Activar entrada del Ring</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="300"/>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="314"/>
         <source>Enable</source>
-        <translation type="unfinished"/>
+        <translation>Activar</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="307"/>
         <source>Ring Sensor Value</source>
-        <translation type="unfinished"/>
+        <translation>Valor del sensor Ring</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="314"/>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="282"/>
         <source>Not connected</source>
-        <translation type="unfinished"/>
+        <translation>No conectado</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="344"/>
@@ -3414,12 +3414,12 @@ UUID: %2</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="279"/>
         <source>Error enabling ring input</source>
-        <translation type="unfinished"/>
+        <translation>Error al activar la entrada del Ring</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="285"/>
         <source>Direct Joycon driver is not enabled</source>
-        <translation type="unfinished"/>
+        <translation>El driver directo JoyCon no está activo.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="290"/>
@@ -3429,17 +3429,17 @@ UUID: %2</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="302"/>
         <source>The current mapped device doesn&apos;t support the ring controller</source>
-        <translation type="unfinished"/>
+        <translation>El dispositivo de entrada actual no soporta el control Ring.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="306"/>
         <source>The current mapped device doesn&apos;t have a ring attached</source>
-        <translation type="unfinished"/>
+        <translation>El dispositivo de entrada actual no tiene puesto el Ring</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="310"/>
         <source>Unexpected driver result %1</source>
-        <translation type="unfinished"/>
+        <translation>Resultado inesperado del driver %1</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="333"/>
@@ -4498,12 +4498,12 @@ Arrastra los puntos para cambiar de posición, o haz doble clic en las celdas de
     <message>
         <location filename="../../src/yuzu/multiplayer/direct_connect.ui" line="47"/>
         <source>Server Address</source>
-        <translation type="unfinished"/>
+        <translation>Dirección del Servidor</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/multiplayer/direct_connect.ui" line="54"/>
         <source>&lt;html&gt;&lt;head/&gt;&lt;body&gt;&lt;p&gt;Server address of the host&lt;/p&gt;&lt;/body&gt;&lt;/html&gt;</source>
-        <translation type="unfinished"/>
+        <translation>&lt;html&gt;&lt;head/&gt;&lt;body&gt;&lt;p&gt;Dirección del servidor del anfitrión&lt;/p&gt;&lt;/body&gt;&lt;/html&gt;</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/multiplayer/direct_connect.ui" line="64"/>
@@ -4617,12 +4617,12 @@ Arrastra los puntos para cambiar de posición, o haz doble clic en las celdas de
     <message>
         <location filename="../../src/yuzu/main.cpp" line="1230"/>
         <source>Emulated mouse is enabled</source>
-        <translation type="unfinished"/>
+        <translation>El ratón emulado está activado</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="1231"/>
         <source>Real mouse input and mouse panning are incompatible. Please disable the emulated mouse in input advanced settings to allow mouse panning.</source>
-        <translation type="unfinished"/>
+        <translation>La entrada de un ratón real y la panoramización del ratón son incompatibles. Por favor, desactive el ratón emulado en la configuración avanzada de entrada para permitir así la panoramización del ratón.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="1453"/>
@@ -5489,13 +5489,13 @@ Por favor, utiliza esta función sólo para instalar actualizaciones y DLCs.</tr
     <message>
         <location filename="../../src/yuzu/main.cpp" line="4069"/>
         <source>VOLUME: MUTE</source>
-        <translation type="unfinished"/>
+        <translation>VOLUMEN: SILENCIO</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="4072"/>
         <source>VOLUME: %1%</source>
         <comment>Volume percentage (e.g. 50%)</comment>
-        <translation type="unfinished"/>
+        <translation>VOLUMEN: %1%</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="4153"/>
@@ -6227,7 +6227,7 @@ Mensaje de depuración: </translation>
     <message>
         <location filename="../../src/yuzu/multiplayer/lobby.ui" line="83"/>
         <source>Hide Empty Rooms</source>
-        <translation type="unfinished"/>
+        <translation>Ocultar salas vacías</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/multiplayer/lobby.ui" line="90"/>
@@ -7152,12 +7152,12 @@ p, li { white-space: pre-wrap; }
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="78"/>
         <source>Stick L</source>
-        <translation type="unfinished"/>
+        <translation>Palanca L</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="80"/>
         <source>Stick R</source>
-        <translation type="unfinished"/>
+        <translation>Palanca R</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="92"/>
@@ -7214,19 +7214,19 @@ p, li { white-space: pre-wrap; }
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="191"/>
         <source>%1%2%3%4</source>
-        <translation type="unfinished"/>
+        <translation>%1%2%3%4</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="205"/>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="229"/>
         <source>%1%2%3Hat %4</source>
-        <translation type="unfinished"/>
+        <translation>%1%2%3Rotación %4</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="223"/>
         <location filename="../../src/yuzu/configuration/configure_input_player.cpp" line="238"/>
         <source>%1%2%3Button %4</source>
-        <translation type="unfinished"/>
+        <translation>%1%2%3Botón %4</translation>
     </message>
 </context>
 <context>
@@ -7647,7 +7647,7 @@ Por favor, inténtalo de nuevo o contacta con el desarrollador del software.</tr
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="166"/>
         <source>Profile Creator</source>
-        <translation type="unfinished"/>
+        <translation>Creador de perfil</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="169"/>
@@ -7658,57 +7658,57 @@ Por favor, inténtalo de nuevo o contacta con el desarrollador del software.</tr
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="172"/>
         <source>Profile Icon Editor</source>
-        <translation type="unfinished"/>
+        <translation>Editor de icono de perfil</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="175"/>
         <source>Profile Nickname Editor</source>
-        <translation type="unfinished"/>
+        <translation>Editor de nombre de perfil</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="197"/>
         <source>Who will receive the points?</source>
-        <translation type="unfinished"/>
+        <translation>¿Quién recibirá los puntos?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="200"/>
         <source>Who is using Nintendo eShop?</source>
-        <translation type="unfinished"/>
+        <translation>¿Quién va a utilizar Nintendo eShop?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="203"/>
         <source>Who is making this purchase?</source>
-        <translation type="unfinished"/>
+        <translation>¿Quién está haciendo la compra?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="206"/>
         <source>Who is posting?</source>
-        <translation type="unfinished"/>
+        <translation>¿Quién está publicando esto?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="209"/>
         <source>Select a user to link to a Nintendo Account.</source>
-        <translation type="unfinished"/>
+        <translation>Elige un usuario para vincularlo a una Cuenta Nintendo.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="212"/>
         <source>Change settings for which user?</source>
-        <translation type="unfinished"/>
+        <translation>¿Para qué usuario desea cambiar la configuración?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="215"/>
         <source>Format data for which user?</source>
-        <translation type="unfinished"/>
+        <translation>¿Para qué usuario se borrarán sus datos?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="218"/>
         <source>Which user will be transferred to another console?</source>
-        <translation type="unfinished"/>
+        <translation>¿Qué usuario será transferido a otra consola?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="221"/>
         <source>Send save data for which user?</source>
-        <translation type="unfinished"/>
+        <translation>¿A qué usuario se le enviarán los datos de guardado?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="225"/>
@@ -7774,7 +7774,7 @@ p, li { white-space: pre-wrap; }
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="158"/>
         <source>[%1] %2</source>
-        <translation type="unfinished"/>
+        <translation>[%1] %2</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="184"/>

dist/languages/fr.ts

@@ -1382,8 +1382,8 @@ Cette option améliore la vitesse en réduisant la précision des instructions f
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Disposition de la mémoire étendue (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation>Disposition de la mémoire étendue (8 Go de DRAM)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>
@@ -7637,7 +7637,7 @@ Veuillez essayer à nouveau ou contactez le développeur du logiciel.</translati
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="166"/>
         <source>Profile Creator</source>
-        <translation type="unfinished"/>
+        <translation>Créateur de profil</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="169"/>
@@ -7648,57 +7648,57 @@ Veuillez essayer à nouveau ou contactez le développeur du logiciel.</translati
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="172"/>
         <source>Profile Icon Editor</source>
-        <translation type="unfinished"/>
+        <translation>Éditeur d&apos;icônes de profil</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="175"/>
         <source>Profile Nickname Editor</source>
-        <translation type="unfinished"/>
+        <translation>Éditeur de surnom de profil</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="197"/>
         <source>Who will receive the points?</source>
-        <translation type="unfinished"/>
+        <translation>Qui recevra les points ?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="200"/>
         <source>Who is using Nintendo eShop?</source>
-        <translation type="unfinished"/>
+        <translation>Qui utilise le Nintendo eShop ?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="203"/>
         <source>Who is making this purchase?</source>
-        <translation type="unfinished"/>
+        <translation>Qui effectue cet achat ?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="206"/>
         <source>Who is posting?</source>
-        <translation type="unfinished"/>
+        <translation>Qui publie ?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="209"/>
         <source>Select a user to link to a Nintendo Account.</source>
-        <translation type="unfinished"/>
+        <translation>Sélectionnez un utilisateur à associer à un compte Nintendo.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="212"/>
         <source>Change settings for which user?</source>
-        <translation type="unfinished"/>
+        <translation>Modifier les paramètres pour quel utilisateur ?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="215"/>
         <source>Format data for which user?</source>
-        <translation type="unfinished"/>
+        <translation>Formater les données pour quel utilisateur ?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="218"/>
         <source>Which user will be transferred to another console?</source>
-        <translation type="unfinished"/>
+        <translation>Quel utilisateur sera transféré sur une autre console ?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="221"/>
         <source>Send save data for which user?</source>
-        <translation type="unfinished"/>
+        <translation>Envoyer les données de sauvegarde pour quel utilisateur ?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="225"/>

dist/languages/id.ts

@@ -1339,8 +1339,8 @@ Memungkinkan berbagai macam optimasi IR.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Tata letak memori yang diperluas (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/it.ts

@@ -1368,8 +1368,8 @@ Questo bannerà sia il suo nome utente del forum che il suo indirizzo IP.</trans
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Layout di memoria esteso (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>
@@ -1846,7 +1846,7 @@ Questo bannerà sia il suo nome utente del forum che il suo indirizzo IP.</trans
     <message>
         <location filename="../../src/yuzu/configuration/configure_hotkeys.ui" line="52"/>
         <source>Restore Defaults</source>
-        <translation>Ripristina predefiniti</translation>
+        <translation>Ripristina predefinite</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_hotkeys.cpp" line="98"/>
@@ -1894,7 +1894,7 @@ Questo bannerà sia il suo nome utente del forum che il suo indirizzo IP.</trans
     <message>
         <location filename="../../src/yuzu/configuration/configure_hotkeys.cpp" line="343"/>
         <source>Restore Default</source>
-        <translation>Ripristina predefinito</translation>
+        <translation>Ripristina predefinita</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_hotkeys.cpp" line="344"/>
@@ -3886,7 +3886,7 @@ UUID: %2</translation>
     <message>
         <location filename="../../src/yuzu/configuration/configure_tas.ui" line="77"/>
         <source>Pause execution during loads</source>
-        <translation type="unfinished"/>
+        <translation>Metti in pausa l&apos;esecuzione durante i caricamenti</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_tas.ui" line="91"/>
@@ -5352,7 +5352,7 @@ Configurazione &amp;gt; Web.</translation>
         <location filename="../../src/yuzu/main.cpp" line="3949"/>
         <source>Scale: %1x</source>
         <comment>%1 is the resolution scaling factor</comment>
-        <translation type="unfinished"/>
+        <translation>Risoluzione: %1x</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3952"/>
@@ -5601,7 +5601,7 @@ Desideri uscire comunque?</translation>
     <message>
         <location filename="../../src/yuzu/bootmanager.cpp" line="992"/>
         <source>yuzu has not been compiled with OpenGL support.</source>
-        <translation>yuzu non è stato compilato con il supporto OpenGL.</translation>
+        <translation>yuzu è stato compilato senza il supporto a OpenGL.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/bootmanager.cpp" line="1016"/>

dist/languages/ja_JP.ts

@@ -1383,8 +1383,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>拡張メモリレイアウト（6GB DRAM）</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/ko_KR.ts

@@ -1384,8 +1384,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>확장 메모리 레이아웃(6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/nb.ts

@@ -1355,8 +1355,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Utvidet minneutforming (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/pl.ts

@@ -1377,8 +1377,8 @@ Gdy ta opcja jest włączona, niedopasowanie jest uruchamiane tylko wtedy, gdy d
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Rozszerzony układ pamięci (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/pt_BR.ts

@@ -1383,8 +1383,8 @@ Isto banirá tanto o nome de usuário do fórum como o endereço IP.</translatio
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Layout de memória extendida (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/pt_PT.ts

@@ -1373,8 +1373,8 @@ Isto banirá tanto o nome de usuário do fórum como o endereço IP.</translatio
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Layout de memória extendida (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/ru_RU.ts

@@ -381,17 +381,17 @@ This would ban both their forum username and their IP address.</source>
     <message>
         <location filename="../../src/yuzu/configuration/configure_audio.ui" line="42"/>
         <source>Output Device:</source>
-        <translation type="unfinished"/>
+        <translation>Устройство вывода:</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_audio.ui" line="56"/>
         <source>Input Device:</source>
-        <translation type="unfinished"/>
+        <translation>Устройство ввода:</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_audio.ui" line="70"/>
         <source>Sound Output Mode:</source>
-        <translation type="unfinished"/>
+        <translation>Режим вывода звука:</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_audio.ui" line="78"/>
@@ -551,13 +551,13 @@ This would ban both their forum username and their IP address.</source>
             &lt;div&gt;This option improves speed by reducing accuracy of fused-multiply-add instructions on CPUs without native FMA support.&lt;/div&gt;
            </source>
         <translation>
-            &lt;div&gt;Эта опция повышает скорость, уменьшая точность сложенных умноженных инструкций на ЦП без поддержки FMA.&lt;/div&gt;
+            &lt;div&gt;Эта опция повышает скорость за счет снижения точности инструкций fused-multiply-add на ЦП без встроенной поддержки FMA.&lt;/div&gt;
            </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu.ui" line="106"/>
         <source>Unfuse FMA (improve performance on CPUs without FMA)</source>
-        <translation>Не использовать FMA (улучшает производительность на ЦП без FMA)</translation>
+        <translation>Отключить FMA (улучшает производительность на ЦП без FMA)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu.ui" line="113"/>
@@ -565,7 +565,7 @@ This would ban both their forum username and their IP address.</source>
             &lt;div&gt;This option improves the speed of some approximate floating-point functions by using less accurate native approximations.&lt;/div&gt;
            </source>
         <translation>
-            &lt;div&gt;Эта опция повышает скорость некоторых аппроксимирующих функций с плавающей точкой за счет использования менее точных нативных приближений.&lt;/div&gt;
+            &lt;div&gt;Эта опция повышает скорость некоторых приближенных функций с плавающей точкой за счет использования менее точных нативных приближений&lt;/div&gt;
            </translation>
     </message>
     <message>
@@ -579,13 +579,13 @@ This would ban both their forum username and their IP address.</source>
             &lt;div&gt;This option improves the speed of 32 bits ASIMD floating-point functions by running with incorrect rounding modes.&lt;/div&gt;
            </source>
         <translation>
-            &lt;div&gt;Эта опция улучшает скорость 32-битных ASIMD-функций с плавающей запятой путём работы с некорректными режимами округления.&lt;/div&gt;
+            &lt;div&gt;Эта опция повышает скорость работы 32-битных ASIMD-функций с плавающей запятой, работая с неправильными режимами округления.&lt;/div&gt;
            </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu.ui" line="130"/>
         <source>Faster ASIMD instructions (32 bits only)</source>
-        <translation>Более быстрые инструкции ASIMD (только 32 бит)</translation>
+        <translation>Ускоренные инструкции ASIMD (только 32 бит)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu.ui" line="137"/>
@@ -593,20 +593,22 @@ This would ban both their forum username and their IP address.</source>
             &lt;div&gt;This option improves speed by removing NaN checking. Please note this also reduces accuracy of certain floating-point instructions.&lt;/div&gt;
            </source>
         <translation>
-            &lt;div&gt;Эта опция повышает скорость, убирая проверку на NaN. Обратите внимание, что это также снижает точность некоторых инструкций с плавающей точкой.&lt;/div&gt;
+            &lt;div&gt;Эта опция повышает скорость, удаляя проверку NaN. Обратите внимание, что это также снижает точность некоторых инструкций с плавающей точкой.&lt;/div&gt;
            </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu.ui" line="142"/>
         <source>Inaccurate NaN handling</source>
-        <translation>Неправильная обработка NaN</translation>
+        <translation>Неточная обработка NaN</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu.ui" line="149"/>
         <source>
             &lt;div&gt;This option improves speed by eliminating a safety check before every memory read/write in guest. Disabling it may allow a game to read/write the emulator's memory.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div&gt;Эта опция повышает скорость за счет исключения проверки безопасности перед каждым чтением/записью памяти в гостевом режиме. Отключение этой опции может позволить игре читать/записывать память эмулятора.
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu.ui" line="154"/>
@@ -618,12 +620,14 @@ This would ban both their forum username and their IP address.</source>
         <source>
             &lt;div&gt;This option improves speed by relying only on the semantics of cmpxchg to ensure safety of exclusive access instructions. Please note this may result in deadlocks and other race conditions.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div&gt;Эта опция повышает скорость, полагаясь только на семантику cmpxchg для обеспечения безопасности инструкций исключительного доступа. Обратите внимание, что это может привести к полным зависаниям и другим условиям гонки.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu.ui" line="166"/>
         <source>Ignore global monitor</source>
-        <translation type="unfinished"/>
+        <translation>Игнорировать глобальный мониторинг</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu.ui" line="191"/>
@@ -660,7 +664,11 @@ This would ban both their forum username and their IP address.</source>
             &lt;div style=&quot;white-space: nowrap&quot;&gt;Enabling it inlines accesses to PageTable::pointers into emitted code.&lt;/div&gt;
             &lt;div style=&quot;white-space: nowrap&quot;&gt;Disabling this forces all memory accesses to go through the Memory::Read/Memory::Write functions.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Эта оптимизация ускоряет доступ гостевой программы к памяти.&lt;/div&gt;
+            &lt;div style=&quot;white-space: nowrap&quot;&gt; Включение этой оптимизации встраивает доступ к указателям PageTable::pointers в эмулируемый код.&lt;/div&gt;
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Отключение этой функции заставляет все обращения к памяти проходить через функции Memory::Read/Memory::Write.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="48"/>
@@ -672,7 +680,9 @@ This would ban both their forum username and their IP address.</source>
         <source>
             &lt;div&gt;This optimization avoids dispatcher lookups by allowing emitted basic blocks to jump directly to other basic blocks if the destination PC is static.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div&gt;Эта оптимизация позволяет избежать поиска диспетчера, позволяя эмитированным базовым блокам переходить непосредственно к другим базовым блокам, если конечный ПК статичен.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="60"/>
@@ -684,7 +694,9 @@ This would ban both their forum username and their IP address.</source>
         <source>
             &lt;div&gt;This optimization avoids dispatcher lookups by keeping track potential return addresses of BL instructions. This approximates what happens with a return stack buffer on a real CPU.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div&gt;Эта оптимизация позволяет избежать поиска диспетчера, отслеживая потенциальные адреса возврата инструкций BL. Это приближено к тому, что происходит с буфером стека возврата на реальном ЦП.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="72"/>
@@ -696,7 +708,9 @@ This would ban both their forum username and their IP address.</source>
         <source>
             &lt;div&gt;Enable a two-tiered dispatch system. A faster dispatcher written in assembly has a small MRU cache of jump destinations is used first. If that fails, dispatch falls back to the slower C++ dispatcher.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div&gt;Включите двухуровневую систему диспетчеризации. Сначала используется более быстрый диспетчер, написанный на ассемблере и имеющий небольшой кэш MRU для мест назначения переходов. Если он не справляется, диспетчеризация возвращается к более медленному диспетчеру на C++.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="84"/>
@@ -708,7 +722,9 @@ This would ban both their forum username and their IP address.</source>
         <source>
             &lt;div&gt;Enables an IR optimization that reduces unnecessary accesses to the CPU context structure.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div&gt;Включает IR-оптимизацию, которая уменьшает ненужные обращения к структуре контекста ЦП.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="96"/>
@@ -720,19 +736,23 @@ This would ban both their forum username and their IP address.</source>
         <source>
             &lt;div&gt;Enables IR optimizations that involve constant propagation.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div&gt;Включает IR-оптимизацию, которая включает распространение констант.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="108"/>
         <source>Enable constant propagation</source>
-        <translation>Включить постоянное распространение</translation>
+        <translation>Включить распространение констант</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="115"/>
         <source>
             &lt;div&gt;Enables miscellaneous IR optimizations.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div&gt;Включает различные IR-оптимизации.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="120"/>
@@ -745,12 +765,15 @@ This would ban both their forum username and their IP address.</source>
             &lt;div style=&quot;white-space: nowrap&quot;&gt;When enabled, a misalignment is only triggered when an access crosses a page boundary.&lt;/div&gt;
             &lt;div style=&quot;white-space: nowrap&quot;&gt;When disabled, a misalignment is triggered on all misaligned accesses.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Если функция включена, смещение срабатывает только тогда, когда доступ пересекает границу страницы.&lt;/div&gt;
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Если отключено, смещение срабатывает при всех смещенных доступах.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="133"/>
         <source>Enable misalignment check reduction</source>
-        <translation type="unfinished"/>
+        <translation>Включить уменьшение проверки несоосности</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="140"/>
@@ -759,12 +782,16 @@ This would ban both their forum username and their IP address.</source>
             &lt;div style=&quot;white-space: nowrap&quot;&gt;Enabling it causes guest memory reads/writes to be done directly into memory and make use of Host's MMU.&lt;/div&gt;
             &lt;div style=&quot;white-space: nowrap&quot;&gt;Disabling this forces all memory accesses to use Software MMU Emulation.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Эта оптимизация ускоряет доступ гостевой программы к памяти.&lt;/div&gt;
+            &lt;div style=&quot;white-space: nowrap&quot;&gt; Включение этой оптимизации приводит к тому, что чтение/запись гостевой памяти производится непосредственно в память и использует MMU хоста.&lt;/div&gt;
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Отключение этой функции заставляет все обращения к памяти использовать программную эмуляцию MMU.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="147"/>
         <source>Enable Host MMU Emulation (general memory instructions)</source>
-        <translation type="unfinished"/>
+        <translation>Включить эмуляцию MMU хоста (инструкции общей памяти)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="154"/>
@@ -773,12 +800,16 @@ This would ban both their forum username and their IP address.</source>
             &lt;div style=&quot;white-space: nowrap&quot;&gt;Enabling it causes guest exclusive memory reads/writes to be done directly into memory and make use of Host's MMU.&lt;/div&gt;
             &lt;div style=&quot;white-space: nowrap&quot;&gt;Disabling this forces all exclusive memory accesses to use Software MMU Emulation.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Эта оптимизация ускоряет доступ гостевой программы к эксклюзивной памяти.&lt;/div&gt;
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Включение этой оптимизации приводит к тому, что чтение/запись в эксклюзивную память гостя выполняется непосредственно в память и использует MMU хоста.&lt;/div&gt;
+            &lt;div style=&quot;white-space: nowrap&quot;&gt; Отключение этой функции заставляет все эксклюзивные доступы к памяти использовать эмуляцию программного MMU.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="161"/>
         <source>Enable Host MMU Emulation (exclusive memory instructions)</source>
-        <translation type="unfinished"/>
+        <translation>Включить эмуляцию MMU хоста (инструкции исключительной памяти)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="168"/>
@@ -786,12 +817,15 @@ This would ban both their forum username and their IP address.</source>
             &lt;div style=&quot;white-space: nowrap&quot;&gt;This optimization speeds up exclusive memory accesses by the guest program.&lt;/div&gt;
             &lt;div style=&quot;white-space: nowrap&quot;&gt;Enabling it reduces the overhead of fastmem failure of exclusive memory accesses.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Эта оптимизация ускоряет обращение гостевой программы к исключительной памяти.&lt;/div&gt;
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Ее включение снижает накладные расходы, связанные с отказом fastmem при доступе к исключительной памяти.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="174"/>
         <source>Enable recompilation of exclusive memory instructions</source>
-        <translation type="unfinished"/>
+        <translation>Разрешить перекомпиляцию инструкций исключительной памяти</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="181"/>
@@ -799,7 +833,10 @@ This would ban both their forum username and their IP address.</source>
             &lt;div style=&quot;white-space: nowrap&quot;&gt;This optimization speeds up memory accesses by allowing invalid memory accesses to succeed.&lt;/div&gt;
             &lt;div style=&quot;white-space: nowrap&quot;&gt;Enabling it reduces the overhead of all memory accesses and has no impact on programs that don't access invalid memory.&lt;/div&gt;
            </source>
-        <translation type="unfinished"/>
+        <translation>
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Эта оптимизация ускоряет обращение к памяти, позволяя успешное обращение к недопустимой памяти.&lt;/div&gt;
+            &lt;div style=&quot;white-space: nowrap&quot;&gt;Включение этой оптимизации снижает накладные расходы на все обращения к памяти и не влияет на программы, которые не обращаются к недопустимой памяти.&lt;/div&gt;
+           </translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_cpu_debug.ui" line="187"/>
@@ -917,17 +954,17 @@ This would ban both their forum username and their IP address.</source>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="198"/>
         <source>When checked, it disables the macro Just In Time compiler. Enabling this makes games run slower</source>
-        <translation>Если включено, отключает компилятор макроса Just In Time. Включение опции делает игры медленнее</translation>
+        <translation>Если включено, отключает компилятор макроса Just In Time. Включение этого параметра замедляет работу игр</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="201"/>
         <source>Disable Macro JIT</source>
-        <translation>Отключить Macro JIT</translation>
+        <translation>Отключить макрос JIT</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="211"/>
         <source>When checked, it disables the macro HLE functions. Enabling this makes games run slower</source>
-        <translation type="unfinished"/>
+        <translation>Если флажок установлен, он отключает функции макроса HLE. Включение этого параметра замедляет работу игр</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="214"/>
@@ -947,12 +984,12 @@ This would ban both their forum username and their IP address.</source>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="231"/>
         <source>When checked, it executes shaders without loop logic changes</source>
-        <translation type="unfinished"/>
+        <translation>Если включено, шейдеры выполняются без изменения логики цикла</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="234"/>
         <source>Disable Loop safety checks</source>
-        <translation type="unfinished"/>
+        <translation>Отключить проверку безопасности цикла</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="244"/>
@@ -967,12 +1004,12 @@ This would ban both their forum username and their IP address.</source>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="257"/>
         <source>Enable FS Access Log</source>
-        <translation type="unfinished"/>
+        <translation>Включить журнал доступа к ФС</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="264"/>
         <source>Enable this to output the latest generated audio command list to the console. Only affects games using the audio renderer.</source>
-        <translation type="unfinished"/>
+        <translation>Включите эту опцию, чтобы вывести на консоль последний сгенерированный список аудиокоманд. Влияет только на игры, использующие аудио рендерер.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="267"/>
@@ -1007,7 +1044,7 @@ This would ban both their forum username and their IP address.</source>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="311"/>
         <source>Enable Auto-Stub**</source>
-        <translation type="unfinished"/>
+        <translation>Включить автоподставку**</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="318"/>
@@ -1346,8 +1383,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Расширенная компоновка памяти (6 ГБ DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation>Расширенная компоновка памяти (8 ГБ DRAM)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>
@@ -1716,12 +1753,12 @@ This would ban both their forum username and their IP address.</source>
     <message>
         <location filename="../../src/yuzu/configuration/configure_graphics_advanced.ui" line="95"/>
         <source>Enables asynchronous ASTC texture decoding, which may reduce load time stutter. This feature is experimental.</source>
-        <translation type="unfinished"/>
+        <translation>Включает асинхронное декодирование текстур ASTC, что может уменьшить фризы при загрузке. Эта функция является экспериментальной.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_graphics_advanced.ui" line="98"/>
         <source>Decode ASTC textures asynchronously (Hack)</source>
-        <translation type="unfinished"/>
+        <translation>Асинхронное декодирование текстур ASTC (Хак)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_graphics_advanced.ui" line="105"/>
@@ -2236,7 +2273,7 @@ This would ban both their forum username and their IP address.</source>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_advanced.ui" line="2726"/>
         <source>Enable direct Pro Controller driver [EXPERIMENTAL]</source>
-        <translation type="unfinished"/>
+        <translation>Включить прямой драйвер Pro Controller [ЭКСПЕРИМЕНТАЛЬНО]</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_input_advanced.ui" line="2739"/>
@@ -4580,12 +4617,12 @@ Drag points to change position, or double-click table cells to edit values.</sou
     <message>
         <location filename="../../src/yuzu/main.cpp" line="1230"/>
         <source>Emulated mouse is enabled</source>
-        <translation type="unfinished"/>
+        <translation>Эмулированная мышь включена</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="1231"/>
         <source>Real mouse input and mouse panning are incompatible. Please disable the emulated mouse in input advanced settings to allow mouse panning.</source>
-        <translation type="unfinished"/>
+        <translation>Ввод реальной мыши и панорамирование мышью несовместимы. Пожалуйста, отключите эмулированную мышь в расширенных настройках ввода, чтобы разрешить панорамирование мышью.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="1453"/>
@@ -7613,7 +7650,7 @@ Please try again or contact the developer of the software.</source>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="166"/>
         <source>Profile Creator</source>
-        <translation type="unfinished"/>
+        <translation>Создатель профиля</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="169"/>
@@ -7624,57 +7661,57 @@ Please try again or contact the developer of the software.</source>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="172"/>
         <source>Profile Icon Editor</source>
-        <translation type="unfinished"/>
+        <translation>Редактор иконки профиля</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="175"/>
         <source>Profile Nickname Editor</source>
-        <translation type="unfinished"/>
+        <translation>Редактор никнейма профиля</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="197"/>
         <source>Who will receive the points?</source>
-        <translation type="unfinished"/>
+        <translation>Кто будет получать очки?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="200"/>
         <source>Who is using Nintendo eShop?</source>
-        <translation type="unfinished"/>
+        <translation>Кто использует Nintendo eShop?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="203"/>
         <source>Who is making this purchase?</source>
-        <translation type="unfinished"/>
+        <translation>Кто совершает эту покупку?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="206"/>
         <source>Who is posting?</source>
-        <translation type="unfinished"/>
+        <translation>Кто публикует?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="209"/>
         <source>Select a user to link to a Nintendo Account.</source>
-        <translation type="unfinished"/>
+        <translation>Выберите пользователя для привязки к учетной записи Nintendo.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="212"/>
         <source>Change settings for which user?</source>
-        <translation type="unfinished"/>
+        <translation>Изменить настройки для какого пользователя?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="215"/>
         <source>Format data for which user?</source>
-        <translation type="unfinished"/>
+        <translation>Форматировать данные для какого пользователя?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="218"/>
         <source>Which user will be transferred to another console?</source>
-        <translation type="unfinished"/>
+        <translation>Какой пользователь будет переходить на другую консоль?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="221"/>
         <source>Send save data for which user?</source>
-        <translation type="unfinished"/>
+        <translation>Отправить сохранение какому пользователю?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/applets/qt_profile_select.cpp" line="225"/>
@@ -7740,7 +7777,7 @@ p, li { white-space: pre-wrap; }
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="158"/>
         <source>[%1] %2</source>
-        <translation type="unfinished"/>
+        <translation>[%1] %2</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="184"/>
@@ -7753,17 +7790,17 @@ p, li { white-space: pre-wrap; }
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="202"/>
         <source>runnable</source>
-        <translation type="unfinished"/>
+        <translation>runnable</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="204"/>
         <source>paused</source>
-        <translation type="unfinished"/>
+        <translation>paused</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="210"/>
         <source>sleeping</source>
-        <translation type="unfinished"/>
+        <translation>sleeping</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="213"/>
@@ -7778,32 +7815,32 @@ p, li { white-space: pre-wrap; }
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="219"/>
         <source>waiting for condition variable</source>
-        <translation type="unfinished"/>
+        <translation>waiting for condition variable</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="222"/>
         <source>waiting for address arbiter</source>
-        <translation type="unfinished"/>
+        <translation>waiting for address arbiter</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="225"/>
         <source>waiting for suspend resume</source>
-        <translation type="unfinished"/>
+        <translation>waiting for suspend resume</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="228"/>
         <source>waiting</source>
-        <translation type="unfinished"/>
+        <translation>waiting</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="233"/>
         <source>initialized</source>
-        <translation type="unfinished"/>
+        <translation>initialized</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="236"/>
         <source>terminated</source>
-        <translation type="unfinished"/>
+        <translation>terminated</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="239"/>
@@ -7818,7 +7855,7 @@ p, li { white-space: pre-wrap; }
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="294"/>
         <source>ideal</source>
-        <translation type="unfinished"/>
+        <translation>ideal</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="297"/>
@@ -7848,7 +7885,7 @@ p, li { white-space: pre-wrap; }
     <message>
         <location filename="../../src/yuzu/debugger/wait_tree.cpp" line="309"/>
         <source>last running ticks = %1</source>
-        <translation type="unfinished"/>
+        <translation>last running ticks = %1</translation>
     </message>
 </context>
 <context>

dist/languages/sv.ts

@@ -122,7 +122,7 @@ p, li { white-space: pre-wrap; }
     <message>
         <location filename="../../src/yuzu/multiplayer/chat_room.cpp" line="330"/>
         <source>%1 has been unbanned</source>
-        <translation type="unfinished"/>
+        <translation>%1 har haft dess bannlysning upphävd.</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/multiplayer/chat_room.cpp" line="446"/>
@@ -242,22 +242,22 @@ Detta kommer bannlysa både dennes användarnamn på forum samt IP-adress.</tran
     <message>
         <location filename="../../src/yuzu/compatdb.ui" line="77"/>
         <source>&lt;html&gt;&lt;head/&gt;&lt;body&gt;&lt;p&gt;Does the game boot?&lt;/p&gt;&lt;/body&gt;&lt;/html&gt;</source>
-        <translation type="unfinished"/>
+        <translation>&lt;html&gt;&lt;head/&gt;&lt;body&gt;&lt;p&gt;Startar Spelet? &lt;/p&gt;&lt;/body&gt;&lt;/html&gt;</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/compatdb.ui" line="100"/>
         <source>Yes   The game starts to output video or audio</source>
-        <translation type="unfinished"/>
+        <translation>Ja   Spelet öppnar till utmatning av video eller audio</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/compatdb.ui" line="107"/>
         <source>No    The game doesn&apos;t get past the &quot;Launching...&quot; screen</source>
-        <translation type="unfinished"/>
+        <translation>Nej Spelet öppnar ej förbi &quot;Startar...&quot; skärmen</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/compatdb.ui" line="124"/>
         <source>Yes   The game gets past the intro/menu and into gameplay</source>
-        <translation type="unfinished"/>
+        <translation>Ja   Spelet öppnar förbi introt/menyn och in i själva spelandet</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/compatdb.ui" line="131"/>
@@ -1031,7 +1031,7 @@ avgjord kod.&lt;/div&gt;
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="325"/>
         <source>Disable Web Applet</source>
-        <translation type="unfinished"/>
+        <translation>Avaktivera Webbappletten</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_debug.ui" line="332"/>
@@ -1360,8 +1360,8 @@ avgjord kod.&lt;/div&gt;
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Utökad minnesöversikt (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>
@@ -3320,13 +3320,13 @@ UUID: %2</source>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="84"/>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="123"/>
         <source>Pull</source>
-        <translation type="unfinished"/>
+        <translation>Dra</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="133"/>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="172"/>
         <source>Push</source>
-        <translation type="unfinished"/>
+        <translation>Knuff</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="206"/>
@@ -3347,7 +3347,7 @@ UUID: %2</source>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="300"/>
         <location filename="../../src/yuzu/configuration/configure_ringcon.cpp" line="314"/>
         <source>Enable</source>
-        <translation type="unfinished"/>
+        <translation>Aktivera</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_ringcon.ui" line="307"/>
@@ -4493,17 +4493,17 @@ Dra punkter för att ändra position, eller dubbelklicka tabellceller för att r
     <message>
         <location filename="../../src/yuzu/multiplayer/direct_connect.ui" line="97"/>
         <source>Nickname</source>
-        <translation type="unfinished"/>
+        <translation>Smeknamn</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/multiplayer/direct_connect.ui" line="111"/>
         <source>Password</source>
-        <translation type="unfinished"/>
+        <translation>Lösenord</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/multiplayer/direct_connect.ui" line="153"/>
         <source>Connect</source>
-        <translation type="unfinished"/>
+        <translation>Anslut</translation>
     </message>
 </context>
 <context>
@@ -4511,12 +4511,12 @@ Dra punkter för att ändra position, eller dubbelklicka tabellceller för att r
     <message>
         <location filename="../../src/yuzu/multiplayer/direct_connect.cpp" line="120"/>
         <source>Connecting</source>
-        <translation type="unfinished"/>
+        <translation>Ansluter</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/multiplayer/direct_connect.cpp" line="125"/>
         <source>Connect</source>
-        <translation type="unfinished"/>
+        <translation>Anslut</translation>
     </message>
 </context>
 <context>
@@ -4534,7 +4534,7 @@ Dra punkter för att ändra position, eller dubbelklicka tabellceller för att r
     <message>
         <location filename="../../src/yuzu/main.cpp" line="432"/>
         <source>Broken Vulkan Installation Detected</source>
-        <translation type="unfinished"/>
+        <translation>Felaktig Vulkaninstallation Upptäckt</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="433"/>
@@ -4550,7 +4550,7 @@ Dra punkter för att ändra position, eller dubbelklicka tabellceller för att r
         <location filename="../../src/yuzu/main.cpp" line="874"/>
         <location filename="../../src/yuzu/main.cpp" line="877"/>
         <source>Disable Web Applet</source>
-        <translation type="unfinished"/>
+        <translation>Avaktivera Webbappletten</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="878"/>
@@ -4591,7 +4591,7 @@ Dra punkter för att ändra position, eller dubbelklicka tabellceller för att r
     <message>
         <location filename="../../src/yuzu/main.cpp" line="1230"/>
         <source>Emulated mouse is enabled</source>
-        <translation type="unfinished"/>
+        <translation>Emulerad datormus är aktiverad</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="1231"/>
@@ -5171,7 +5171,7 @@ Please, only use this feature to install updates and DLC.</source>
         <location filename="../../src/yuzu/main.cpp" line="3182"/>
         <location filename="../../src/yuzu/main.cpp" line="3201"/>
         <source>Hardware requirements not met</source>
-        <translation type="unfinished"/>
+        <translation> Hårdvarukraven uppfylls ej</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3183"/>
@@ -5202,17 +5202,17 @@ Please, only use this feature to install updates and DLC.</source>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3514"/>
         <source>TAS Recording</source>
-        <translation type="unfinished"/>
+        <translation>TAS Inspelning</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3515"/>
         <source>Overwrite file of player 1?</source>
-        <translation type="unfinished"/>
+        <translation>Överskriv spelare 1:s fil?</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3541"/>
         <source>Invalid config detected</source>
-        <translation type="unfinished"/>
+        <translation>Ogiltig konfiguration upptäckt</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3542"/>
@@ -5223,13 +5223,13 @@ Please, only use this feature to install updates and DLC.</source>
         <location filename="../../src/yuzu/main.cpp" line="3712"/>
         <location filename="../../src/yuzu/main.cpp" line="3740"/>
         <source>Amiibo</source>
-        <translation type="unfinished"/>
+        <translation>Amiibo</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3712"/>
         <location filename="../../src/yuzu/main.cpp" line="3740"/>
         <source>The current amiibo has been removed</source>
-        <translation type="unfinished"/>
+        <translation>Den aktuella amiibon har avlägsnats</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3717"/>
@@ -5240,7 +5240,7 @@ Please, only use this feature to install updates and DLC.</source>
         <location filename="../../src/yuzu/main.cpp" line="3717"/>
         <location filename="../../src/yuzu/main.cpp" line="3752"/>
         <source>The current game is not looking for amiibos</source>
-        <translation type="unfinished"/>
+        <translation>Det aktuella spelet letar ej efter amiibos</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3723"/>
@@ -5260,17 +5260,17 @@ Please, only use this feature to install updates and DLC.</source>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3746"/>
         <source>The selected file is not a valid amiibo</source>
-        <translation type="unfinished"/>
+        <translation>Den valda filen är inte en giltig amiibo</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3749"/>
         <source>The selected file is already on use</source>
-        <translation type="unfinished"/>
+        <translation>Den valda filen är redan använd</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3755"/>
         <source>An unknown error occurred</source>
-        <translation type="unfinished"/>
+        <translation>Ett okänt fel har inträffat</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3807"/>
@@ -5285,22 +5285,22 @@ Please, only use this feature to install updates and DLC.</source>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3891"/>
         <source>TAS state: Running %1/%2</source>
-        <translation type="unfinished"/>
+        <translation>TAStillstånd: pågående %1/%2</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3895"/>
         <source>TAS state: Recording %1</source>
-        <translation type="unfinished"/>
+        <translation>TAStillstånd: spelar in %1</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3897"/>
         <source>TAS state: Idle %1/%2</source>
-        <translation type="unfinished"/>
+        <translation>TAStillstånd: inaktiv %1/%2</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3901"/>
         <source>TAS State: Invalid</source>
-        <translation type="unfinished"/>
+        <translation>TAStillstånd: ogiltigt</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/main.cpp" line="3915"/>
@@ -5902,7 +5902,7 @@ Vill du strunta i detta och avsluta ändå?</translation>
     <message>
         <location filename="../../src/yuzu/multiplayer/host_room.ui" line="118"/>
         <source>Password</source>
-        <translation type="unfinished"/>
+        <translation>Lösenord</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/multiplayer/host_room.ui" line="125"/>
@@ -6169,7 +6169,7 @@ Debug Message: </source>
         <location filename="../../src/yuzu/multiplayer/lobby.ui" line="32"/>
         <location filename="../../src/yuzu/multiplayer/lobby.ui" line="39"/>
         <source>Nickname</source>
-        <translation type="unfinished"/>
+        <translation>Smeknamn</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/multiplayer/lobby.ui" line="59"/>

dist/languages/tr_TR.ts

@@ -1368,8 +1368,8 @@ Bu seçenek belleğe yazma/okuma işlemlerindeki güvenlik kontrolünü kaldıra
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Artırılmış hafıza düzeni (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/uk.ts

@@ -1346,8 +1346,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>Розширене компонування пам&apos;яті (6 ГБ DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation type="unfinished"/>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/zh_CN.ts

@@ -1380,8 +1380,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>扩展的内存布局 (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation>扩展的内存布局 (8GB DRAM)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

dist/languages/zh_TW.ts

@@ -1382,8 +1382,8 @@ This would ban both their forum username and their IP address.</source>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="67"/>
-        <source>Extended memory layout (6GB DRAM)</source>
-        <translation>扩展的内存布局 (6GB DRAM)</translation>
+        <source>Extended memory layout (8GB DRAM)</source>
+        <translation>扩展的内存布局 (8GB DRAM)</translation>
     </message>
     <message>
         <location filename="../../src/yuzu/configuration/configure_general.ui" line="74"/>

src/CMakeLists.txt

@@ -113,6 +113,9 @@ else()
 
         $<$<CXX_COMPILER_ID:Clang>:-Wno-braced-scalar-init>
         $<$<CXX_COMPILER_ID:Clang>:-Wno-unused-private-field>
+        $<$<CXX_COMPILER_ID:Clang>:-Werror=shadow-uncaptured-local>
+        $<$<CXX_COMPILER_ID:Clang>:-Werror=implicit-fallthrough>
+        $<$<CXX_COMPILER_ID:Clang>:-Werror=type-limits>
         $<$<CXX_COMPILER_ID:AppleClang>:-Wno-braced-scalar-init>
         $<$<CXX_COMPILER_ID:AppleClang>:-Wno-unused-private-field>
     )

src/common/intrusive_list.h

@@ -0,0 +1,631 @@
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#pragma once
+
+#include "common/common_funcs.h"
+#include "common/parent_of_member.h"
+
+namespace Common {
+
+// Forward declare implementation class for Node.
+namespace impl {
+
+class IntrusiveListImpl;
+
+}
+
+class IntrusiveListNode {
+    YUZU_NON_COPYABLE(IntrusiveListNode);
+
+private:
+    friend class impl::IntrusiveListImpl;
+
+    IntrusiveListNode* m_prev;
+    IntrusiveListNode* m_next;
+
+public:
+    constexpr IntrusiveListNode() : m_prev(this), m_next(this) {}
+
+    constexpr bool IsLinked() const {
+        return m_next != this;
+    }
+
+private:
+    constexpr void LinkPrev(IntrusiveListNode* node) {
+        // We can't link an already linked node.
+        ASSERT(!node->IsLinked());
+        this->SplicePrev(node, node);
+    }
+
+    constexpr void SplicePrev(IntrusiveListNode* first, IntrusiveListNode* last) {
+        // Splice a range into the list.
+        auto last_prev = last->m_prev;
+        first->m_prev = m_prev;
+        last_prev->m_next = this;
+        m_prev->m_next = first;
+        m_prev = last_prev;
+    }
+
+    constexpr void LinkNext(IntrusiveListNode* node) {
+        // We can't link an already linked node.
+        ASSERT(!node->IsLinked());
+        return this->SpliceNext(node, node);
+    }
+
+    constexpr void SpliceNext(IntrusiveListNode* first, IntrusiveListNode* last) {
+        // Splice a range into the list.
+        auto last_prev = last->m_prev;
+        first->m_prev = this;
+        last_prev->m_next = m_next;
+        m_next->m_prev = last_prev;
+        m_next = first;
+    }
+
+    constexpr void Unlink() {
+        this->Unlink(m_next);
+    }
+
+    constexpr void Unlink(IntrusiveListNode* last) {
+        // Unlink a node from a next node.
+        auto last_prev = last->m_prev;
+        m_prev->m_next = last;
+        last->m_prev = m_prev;
+        last_prev->m_next = this;
+        m_prev = last_prev;
+    }
+
+    constexpr IntrusiveListNode* GetPrev() {
+        return m_prev;
+    }
+
+    constexpr const IntrusiveListNode* GetPrev() const {
+        return m_prev;
+    }
+
+    constexpr IntrusiveListNode* GetNext() {
+        return m_next;
+    }
+
+    constexpr const IntrusiveListNode* GetNext() const {
+        return m_next;
+    }
+};
+// DEPRECATED: static_assert(std::is_literal_type<IntrusiveListNode>::value);
+
+namespace impl {
+
+class IntrusiveListImpl {
+    YUZU_NON_COPYABLE(IntrusiveListImpl);
+
+private:
+    IntrusiveListNode m_root_node;
+
+public:
+    template <bool Const>
+    class Iterator;
+
+    using value_type = IntrusiveListNode;
+    using size_type = size_t;
+    using difference_type = ptrdiff_t;
+    using pointer = value_type*;
+    using const_pointer = const value_type*;
+    using reference = value_type&;
+    using const_reference = const value_type&;
+    using iterator = Iterator<false>;
+    using const_iterator = Iterator<true>;
+    using reverse_iterator = std::reverse_iterator<iterator>;
+    using const_reverse_iterator = std::reverse_iterator<const_iterator>;
+
+    template <bool Const>
+    class Iterator {
+    public:
+        using iterator_category = std::bidirectional_iterator_tag;
+        using value_type = typename IntrusiveListImpl::value_type;
+        using difference_type = typename IntrusiveListImpl::difference_type;
+        using pointer =
+            std::conditional_t<Const, IntrusiveListImpl::const_pointer, IntrusiveListImpl::pointer>;
+        using reference = std::conditional_t<Const, IntrusiveListImpl::const_reference,
+                                             IntrusiveListImpl::reference>;
+
+    private:
+        pointer m_node;
+
+    public:
+        constexpr explicit Iterator(pointer n) : m_node(n) {}
+
+        constexpr bool operator==(const Iterator& rhs) const {
+            return m_node == rhs.m_node;
+        }
+
+        constexpr pointer operator->() const {
+            return m_node;
+        }
+
+        constexpr reference operator*() const {
+            return *m_node;
+        }
+
+        constexpr Iterator& operator++() {
+            m_node = m_node->m_next;
+            return *this;
+        }
+
+        constexpr Iterator& operator--() {
+            m_node = m_node->m_prev;
+            return *this;
+        }
+
+        constexpr Iterator operator++(int) {
+            const Iterator it{*this};
+            ++(*this);
+            return it;
+        }
+
+        constexpr Iterator operator--(int) {
+            const Iterator it{*this};
+            --(*this);
+            return it;
+        }
+
+        constexpr operator Iterator<true>() const {
+            return Iterator<true>(m_node);
+        }
+
+        constexpr Iterator<false> GetNonConstIterator() const {
+            return Iterator<false>(const_cast<IntrusiveListImpl::pointer>(m_node));
+        }
+    };
+
+public:
+    constexpr IntrusiveListImpl() : m_root_node() {}
+
+    // Iterator accessors.
+    constexpr iterator begin() {
+        return iterator(m_root_node.GetNext());
+    }
+
+    constexpr const_iterator begin() const {
+        return const_iterator(m_root_node.GetNext());
+    }
+
+    constexpr iterator end() {
+        return iterator(std::addressof(m_root_node));
+    }
+
+    constexpr const_iterator end() const {
+        return const_iterator(std::addressof(m_root_node));
+    }
+
+    constexpr iterator iterator_to(reference v) {
+        // Only allow iterator_to for values in lists.
+        ASSERT(v.IsLinked());
+        return iterator(std::addressof(v));
+    }
+
+    constexpr const_iterator iterator_to(const_reference v) const {
+        // Only allow iterator_to for values in lists.
+        ASSERT(v.IsLinked());
+        return const_iterator(std::addressof(v));
+    }
+
+    // Content management.
+    constexpr bool empty() const {
+        return !m_root_node.IsLinked();
+    }
+
+    constexpr size_type size() const {
+        return static_cast<size_type>(std::distance(this->begin(), this->end()));
+    }
+
+    constexpr reference back() {
+        return *m_root_node.GetPrev();
+    }
+
+    constexpr const_reference back() const {
+        return *m_root_node.GetPrev();
+    }
+
+    constexpr reference front() {
+        return *m_root_node.GetNext();
+    }
+
+    constexpr const_reference front() const {
+        return *m_root_node.GetNext();
+    }
+
+    constexpr void push_back(reference node) {
+        m_root_node.LinkPrev(std::addressof(node));
+    }
+
+    constexpr void push_front(reference node) {
+        m_root_node.LinkNext(std::addressof(node));
+    }
+
+    constexpr void pop_back() {
+        m_root_node.GetPrev()->Unlink();
+    }
+
+    constexpr void pop_front() {
+        m_root_node.GetNext()->Unlink();
+    }
+
+    constexpr iterator insert(const_iterator pos, reference node) {
+        pos.GetNonConstIterator()->LinkPrev(std::addressof(node));
+        return iterator(std::addressof(node));
+    }
+
+    constexpr void splice(const_iterator pos, IntrusiveListImpl& o) {
+        splice_impl(pos, o.begin(), o.end());
+    }
+
+    constexpr void splice(const_iterator pos, IntrusiveListImpl& o, const_iterator first) {
+        const_iterator last(first);
+        std::advance(last, 1);
+        splice_impl(pos, first, last);
+    }
+
+    constexpr void splice(const_iterator pos, IntrusiveListImpl& o, const_iterator first,
+                          const_iterator last) {
+        splice_impl(pos, first, last);
+    }
+
+    constexpr iterator erase(const_iterator pos) {
+        if (pos == this->end()) {
+            return this->end();
+        }
+        iterator it(pos.GetNonConstIterator());
+        (it++)->Unlink();
+        return it;
+    }
+
+    constexpr void clear() {
+        while (!this->empty()) {
+            this->pop_front();
+        }
+    }
+
+private:
+    constexpr void splice_impl(const_iterator _pos, const_iterator _first, const_iterator _last) {
+        if (_first == _last) {
+            return;
+        }
+        iterator pos(_pos.GetNonConstIterator());
+        iterator first(_first.GetNonConstIterator());
+        iterator last(_last.GetNonConstIterator());
+        first->Unlink(std::addressof(*last));
+        pos->SplicePrev(std::addressof(*first), std::addressof(*first));
+    }
+};
+
+} // namespace impl
+
+template <class T, class Traits>
+class IntrusiveList {
+    YUZU_NON_COPYABLE(IntrusiveList);
+
+private:
+    impl::IntrusiveListImpl m_impl;
+
+public:
+    template <bool Const>
+    class Iterator;
+
+    using value_type = T;
+    using size_type = size_t;
+    using difference_type = ptrdiff_t;
+    using pointer = value_type*;
+    using const_pointer = const value_type*;
+    using reference = value_type&;
+    using const_reference = const value_type&;
+    using iterator = Iterator<false>;
+    using const_iterator = Iterator<true>;
+    using reverse_iterator = std::reverse_iterator<iterator>;
+    using const_reverse_iterator = std::reverse_iterator<const_iterator>;
+
+    template <bool Const>
+    class Iterator {
+    public:
+        friend class Common::IntrusiveList<T, Traits>;
+
+        using ImplIterator =
+            std::conditional_t<Const, Common::impl::IntrusiveListImpl::const_iterator,
+                               Common::impl::IntrusiveListImpl::iterator>;
+
+        using iterator_category = std::bidirectional_iterator_tag;
+        using value_type = typename IntrusiveList::value_type;
+        using difference_type = typename IntrusiveList::difference_type;
+        using pointer =
+            std::conditional_t<Const, IntrusiveList::const_pointer, IntrusiveList::pointer>;
+        using reference =
+            std::conditional_t<Const, IntrusiveList::const_reference, IntrusiveList::reference>;
+
+    private:
+        ImplIterator m_iterator;
+
+    private:
+        constexpr explicit Iterator(ImplIterator it) : m_iterator(it) {}
+
+        constexpr ImplIterator GetImplIterator() const {
+            return m_iterator;
+        }
+
+    public:
+        constexpr bool operator==(const Iterator& rhs) const {
+            return m_iterator == rhs.m_iterator;
+        }
+
+        constexpr pointer operator->() const {
+            return std::addressof(Traits::GetParent(*m_iterator));
+        }
+
+        constexpr reference operator*() const {
+            return Traits::GetParent(*m_iterator);
+        }
+
+        constexpr Iterator& operator++() {
+            ++m_iterator;
+            return *this;
+        }
+
+        constexpr Iterator& operator--() {
+            --m_iterator;
+            return *this;
+        }
+
+        constexpr Iterator operator++(int) {
+            const Iterator it{*this};
+            ++m_iterator;
+            return it;
+        }
+
+        constexpr Iterator operator--(int) {
+            const Iterator it{*this};
+            --m_iterator;
+            return it;
+        }
+
+        constexpr operator Iterator<true>() const {
+            return Iterator<true>(m_iterator);
+        }
+    };
+
+private:
+    static constexpr IntrusiveListNode& GetNode(reference ref) {
+        return Traits::GetNode(ref);
+    }
+
+    static constexpr IntrusiveListNode const& GetNode(const_reference ref) {
+        return Traits::GetNode(ref);
+    }
+
+    static constexpr reference GetParent(IntrusiveListNode& node) {
+        return Traits::GetParent(node);
+    }
+
+    static constexpr const_reference GetParent(IntrusiveListNode const& node) {
+        return Traits::GetParent(node);
+    }
+
+public:
+    constexpr IntrusiveList() : m_impl() {}
+
+    // Iterator accessors.
+    constexpr iterator begin() {
+        return iterator(m_impl.begin());
+    }
+
+    constexpr const_iterator begin() const {
+        return const_iterator(m_impl.begin());
+    }
+
+    constexpr iterator end() {
+        return iterator(m_impl.end());
+    }
+
+    constexpr const_iterator end() const {
+        return const_iterator(m_impl.end());
+    }
+
+    constexpr const_iterator cbegin() const {
+        return this->begin();
+    }
+
+    constexpr const_iterator cend() const {
+        return this->end();
+    }
+
+    constexpr reverse_iterator rbegin() {
+        return reverse_iterator(this->end());
+    }
+
+    constexpr const_reverse_iterator rbegin() const {
+        return const_reverse_iterator(this->end());
+    }
+
+    constexpr reverse_iterator rend() {
+        return reverse_iterator(this->begin());
+    }
+
+    constexpr const_reverse_iterator rend() const {
+        return const_reverse_iterator(this->begin());
+    }
+
+    constexpr const_reverse_iterator crbegin() const {
+        return this->rbegin();
+    }
+
+    constexpr const_reverse_iterator crend() const {
+        return this->rend();
+    }
+
+    constexpr iterator iterator_to(reference v) {
+        return iterator(m_impl.iterator_to(GetNode(v)));
+    }
+
+    constexpr const_iterator iterator_to(const_reference v) const {
+        return const_iterator(m_impl.iterator_to(GetNode(v)));
+    }
+
+    // Content management.
+    constexpr bool empty() const {
+        return m_impl.empty();
+    }
+
+    constexpr size_type size() const {
+        return m_impl.size();
+    }
+
+    constexpr reference back() {
+        return GetParent(m_impl.back());
+    }
+
+    constexpr const_reference back() const {
+        return GetParent(m_impl.back());
+    }
+
+    constexpr reference front() {
+        return GetParent(m_impl.front());
+    }
+
+    constexpr const_reference front() const {
+        return GetParent(m_impl.front());
+    }
+
+    constexpr void push_back(reference ref) {
+        m_impl.push_back(GetNode(ref));
+    }
+
+    constexpr void push_front(reference ref) {
+        m_impl.push_front(GetNode(ref));
+    }
+
+    constexpr void pop_back() {
+        m_impl.pop_back();
+    }
+
+    constexpr void pop_front() {
+        m_impl.pop_front();
+    }
+
+    constexpr iterator insert(const_iterator pos, reference ref) {
+        return iterator(m_impl.insert(pos.GetImplIterator(), GetNode(ref)));
+    }
+
+    constexpr void splice(const_iterator pos, IntrusiveList& o) {
+        m_impl.splice(pos.GetImplIterator(), o.m_impl);
+    }
+
+    constexpr void splice(const_iterator pos, IntrusiveList& o, const_iterator first) {
+        m_impl.splice(pos.GetImplIterator(), o.m_impl, first.GetImplIterator());
+    }
+
+    constexpr void splice(const_iterator pos, IntrusiveList& o, const_iterator first,
+                          const_iterator last) {
+        m_impl.splice(pos.GetImplIterator(), o.m_impl, first.GetImplIterator(),
+                      last.GetImplIterator());
+    }
+
+    constexpr iterator erase(const_iterator pos) {
+        return iterator(m_impl.erase(pos.GetImplIterator()));
+    }
+
+    constexpr void clear() {
+        m_impl.clear();
+    }
+};
+
+template <auto T, class Derived = Common::impl::GetParentType<T>>
+class IntrusiveListMemberTraits;
+
+template <class Parent, IntrusiveListNode Parent::*Member, class Derived>
+class IntrusiveListMemberTraits<Member, Derived> {
+public:
+    using ListType = IntrusiveList<Derived, IntrusiveListMemberTraits>;
+
+private:
+    friend class IntrusiveList<Derived, IntrusiveListMemberTraits>;
+
+    static constexpr IntrusiveListNode& GetNode(Derived& parent) {
+        return parent.*Member;
+    }
+
+    static constexpr IntrusiveListNode const& GetNode(Derived const& parent) {
+        return parent.*Member;
+    }
+
+    static Derived& GetParent(IntrusiveListNode& node) {
+        return Common::GetParentReference<Member, Derived>(std::addressof(node));
+    }
+
+    static Derived const& GetParent(IntrusiveListNode const& node) {
+        return Common::GetParentReference<Member, Derived>(std::addressof(node));
+    }
+};
+
+template <auto T, class Derived = Common::impl::GetParentType<T>>
+class IntrusiveListMemberTraitsByNonConstexprOffsetOf;
+
+template <class Parent, IntrusiveListNode Parent::*Member, class Derived>
+class IntrusiveListMemberTraitsByNonConstexprOffsetOf<Member, Derived> {
+public:
+    using ListType = IntrusiveList<Derived, IntrusiveListMemberTraitsByNonConstexprOffsetOf>;
+
+private:
+    friend class IntrusiveList<Derived, IntrusiveListMemberTraitsByNonConstexprOffsetOf>;
+
+    static constexpr IntrusiveListNode& GetNode(Derived& parent) {
+        return parent.*Member;
+    }
+
+    static constexpr IntrusiveListNode const& GetNode(Derived const& parent) {
+        return parent.*Member;
+    }
+
+    static Derived& GetParent(IntrusiveListNode& node) {
+        return *reinterpret_cast<Derived*>(reinterpret_cast<char*>(std::addressof(node)) -
+                                           GetOffset());
+    }
+
+    static Derived const& GetParent(IntrusiveListNode const& node) {
+        return *reinterpret_cast<const Derived*>(
+            reinterpret_cast<const char*>(std::addressof(node)) - GetOffset());
+    }
+
+    static uintptr_t GetOffset() {
+        return reinterpret_cast<uintptr_t>(std::addressof(reinterpret_cast<Derived*>(0)->*Member));
+    }
+};
+
+template <class Derived>
+class IntrusiveListBaseNode : public IntrusiveListNode {};
+
+template <class Derived>
+class IntrusiveListBaseTraits {
+public:
+    using ListType = IntrusiveList<Derived, IntrusiveListBaseTraits>;
+
+private:
+    friend class IntrusiveList<Derived, IntrusiveListBaseTraits>;
+
+    static constexpr IntrusiveListNode& GetNode(Derived& parent) {
+        return static_cast<IntrusiveListNode&>(
+            static_cast<IntrusiveListBaseNode<Derived>&>(parent));
+    }
+
+    static constexpr IntrusiveListNode const& GetNode(Derived const& parent) {
+        return static_cast<const IntrusiveListNode&>(
+            static_cast<const IntrusiveListBaseNode<Derived>&>(parent));
+    }
+
+    static constexpr Derived& GetParent(IntrusiveListNode& node) {
+        return static_cast<Derived&>(static_cast<IntrusiveListBaseNode<Derived>&>(node));
+    }
+
+    static constexpr Derived const& GetParent(IntrusiveListNode const& node) {
+        return static_cast<const Derived&>(
+            static_cast<const IntrusiveListBaseNode<Derived>&>(node));
+    }
+};
+
+} // namespace Common

src/common/settings.cpp

@@ -45,6 +45,7 @@ void LogSettings() {
     log_setting("System_LanguageIndex", values.language_index.GetValue());
     log_setting("System_RegionIndex", values.region_index.GetValue());
     log_setting("System_TimeZoneIndex", values.time_zone_index.GetValue());
+    log_setting("System_UnsafeMemoryLayout", values.use_unsafe_extended_memory_layout.GetValue());
     log_setting("Core_UseMultiCore", values.use_multi_core.GetValue());
     log_setting("CPU_Accuracy", values.cpu_accuracy.GetValue());
     log_setting("Renderer_UseResolutionScaling", values.resolution_setup.GetValue());
@@ -191,7 +192,7 @@ void RestoreGlobalState(bool is_powered_on) {
 
     // Core
     values.use_multi_core.SetGlobal(true);
-    values.use_extended_memory_layout.SetGlobal(true);
+    values.use_unsafe_extended_memory_layout.SetGlobal(true);
 
     // CPU
     values.cpu_accuracy.SetGlobal(true);
@@ -205,6 +206,7 @@ void RestoreGlobalState(bool is_powered_on) {
     // Renderer
     values.fsr_sharpening_slider.SetGlobal(true);
     values.renderer_backend.SetGlobal(true);
+    values.async_presentation.SetGlobal(true);
     values.renderer_force_max_clock.SetGlobal(true);
     values.vulkan_device.SetGlobal(true);
     values.fullscreen_mode.SetGlobal(true);
@@ -225,7 +227,6 @@ void RestoreGlobalState(bool is_powered_on) {
     values.shader_backend.SetGlobal(true);
     values.use_asynchronous_shaders.SetGlobal(true);
     values.use_fast_gpu_time.SetGlobal(true);
-    values.use_pessimistic_flushes.SetGlobal(true);
     values.use_vulkan_driver_pipeline_cache.SetGlobal(true);
     values.bg_red.SetGlobal(true);
     values.bg_green.SetGlobal(true);

src/common/settings.h

@@ -388,7 +388,8 @@ struct Values {
 
     // Core
     SwitchableSetting<bool> use_multi_core{true, "use_multi_core"};
-    SwitchableSetting<bool> use_extended_memory_layout{false, "use_extended_memory_layout"};
+    SwitchableSetting<bool> use_unsafe_extended_memory_layout{false,
+                                                              "use_unsafe_extended_memory_layout"};
 
     // Cpu
     SwitchableSetting<CPUAccuracy, true> cpu_accuracy{CPUAccuracy::Auto, CPUAccuracy::Auto,
@@ -422,6 +423,7 @@ struct Values {
     // Renderer
     SwitchableSetting<RendererBackend, true> renderer_backend{
         RendererBackend::Vulkan, RendererBackend::OpenGL, RendererBackend::Null, "backend"};
+    SwitchableSetting<bool> async_presentation{false, "async_presentation"};
     SwitchableSetting<bool> renderer_force_max_clock{false, "force_max_clock"};
     Setting<bool> renderer_debug{false, "debug"};
     Setting<bool> renderer_shader_feedback{false, "shader_feedback"};
@@ -459,7 +461,6 @@ struct Values {
                                                           ShaderBackend::SPIRV, "shader_backend"};
     SwitchableSetting<bool> use_asynchronous_shaders{false, "use_asynchronous_shaders"};
     SwitchableSetting<bool> use_fast_gpu_time{true, "use_fast_gpu_time"};
-    SwitchableSetting<bool> use_pessimistic_flushes{false, "use_pessimistic_flushes"};
     SwitchableSetting<bool> use_vulkan_driver_pipeline_cache{true,
                                                              "use_vulkan_driver_pipeline_cache"};
 

src/core/CMakeLists.txt

@@ -570,10 +570,10 @@ add_library(core STATIC
     hle/service/nfp/nfp.h
     hle/service/nfp/nfp_device.cpp
     hle/service/nfp/nfp_device.h
+    hle/service/nfp/nfp_interface.cpp
+    hle/service/nfp/nfp_interface.h
     hle/service/nfp/nfp_result.h
     hle/service/nfp/nfp_types.h
-    hle/service/nfp/nfp_user.cpp
-    hle/service/nfp/nfp_user.h
     hle/service/ngct/ngct.cpp
     hle/service/ngct/ngct.h
     hle/service/nifm/nifm.cpp

src/core/core.cpp

@@ -137,7 +137,7 @@ struct System::Impl {
         device_memory = std::make_unique<Core::DeviceMemory>();
 
         is_multicore = Settings::values.use_multi_core.GetValue();
-        extended_memory_layout = Settings::values.use_extended_memory_layout.GetValue();
+        extended_memory_layout = Settings::values.use_unsafe_extended_memory_layout.GetValue();
 
         core_timing.SetMulticore(is_multicore);
         core_timing.Initialize([&system]() { system.RegisterHostThread(); });
@@ -169,7 +169,7 @@ struct System::Impl {
     void ReinitializeIfNecessary(System& system) {
         const bool must_reinitialize =
             is_multicore != Settings::values.use_multi_core.GetValue() ||
-            extended_memory_layout != Settings::values.use_extended_memory_layout.GetValue();
+            extended_memory_layout != Settings::values.use_unsafe_extended_memory_layout.GetValue();
 
         if (!must_reinitialize) {
             return;
@@ -178,7 +178,7 @@ struct System::Impl {
         LOG_DEBUG(Kernel, "Re-initializing");
 
         is_multicore = Settings::values.use_multi_core.GetValue();
-        extended_memory_layout = Settings::values.use_extended_memory_layout.GetValue();
+        extended_memory_layout = Settings::values.use_unsafe_extended_memory_layout.GetValue();
 
         Initialize(system);
     }
@@ -293,6 +293,7 @@ struct System::Impl {
         ASSERT(Kernel::KProcess::Initialize(main_process, system, "main",
                                             Kernel::KProcess::ProcessType::Userland, resource_limit)
                    .IsSuccess());
+        Kernel::KProcess::Register(system.Kernel(), main_process);
         kernel.MakeApplicationProcess(main_process);
         const auto [load_result, load_parameters] = app_loader->Load(*main_process, system);
         if (load_result != Loader::ResultStatus::Success) {

src/core/hid/emulated_controller.cpp

@@ -280,6 +280,10 @@ void EmulatedController::LoadVirtualGamepadParams() {
     virtual_stick_params[Settings::NativeAnalog::LStick].Set("axis_y", 1);
     virtual_stick_params[Settings::NativeAnalog::RStick].Set("axis_x", 2);
     virtual_stick_params[Settings::NativeAnalog::RStick].Set("axis_y", 3);
+    virtual_stick_params[Settings::NativeAnalog::LStick].Set("deadzone", 0.0f);
+    virtual_stick_params[Settings::NativeAnalog::LStick].Set("range", 1.0f);
+    virtual_stick_params[Settings::NativeAnalog::RStick].Set("deadzone", 0.0f);
+    virtual_stick_params[Settings::NativeAnalog::RStick].Set("range", 1.0f);
 }
 
 void EmulatedController::ReloadInput() {

src/core/hle/kernel/board/nintendo/nx/k_system_control.cpp

@@ -35,12 +35,13 @@ namespace {
 using namespace Common::Literals;
 
 u32 GetMemorySizeForInit() {
-    return Settings::values.use_extended_memory_layout ? Smc::MemorySize_8GB : Smc::MemorySize_4GB;
+    return Settings::values.use_unsafe_extended_memory_layout ? Smc::MemorySize_8GB
+                                                              : Smc::MemorySize_4GB;
 }
 
 Smc::MemoryArrangement GetMemoryArrangeForInit() {
-    return Settings::values.use_extended_memory_layout ? Smc::MemoryArrangement_8GB
-                                                       : Smc::MemoryArrangement_4GB;
+    return Settings::values.use_unsafe_extended_memory_layout ? Smc::MemoryArrangement_8GB
+                                                              : Smc::MemoryArrangement_4GB;
 }
 } // namespace
 

src/core/hle/kernel/k_auto_object.h

@@ -182,8 +182,8 @@ public:
     explicit KAutoObjectWithList(KernelCore& kernel) : KAutoObject(kernel) {}
 
     static int Compare(const KAutoObjectWithList& lhs, const KAutoObjectWithList& rhs) {
-        const u64 lid = lhs.GetId();
-        const u64 rid = rhs.GetId();
+        const uintptr_t lid = reinterpret_cast<uintptr_t>(std::addressof(lhs));
+        const uintptr_t rid = reinterpret_cast<uintptr_t>(std::addressof(rhs));
 
         if (lid < rid) {
             return -1;

src/core/hle/kernel/k_event_info.h

@@ -5,14 +5,15 @@
 
 #include <array>
 
-#include <boost/intrusive/list.hpp>
+#include "common/intrusive_list.h"
 
 #include "core/hle/kernel/slab_helpers.h"
 #include "core/hle/kernel/svc_types.h"
 
 namespace Kernel {
 
-class KEventInfo : public KSlabAllocated<KEventInfo>, public boost::intrusive::list_base_hook<> {
+class KEventInfo : public KSlabAllocated<KEventInfo>,
+                   public Common::IntrusiveListBaseNode<KEventInfo> {
 public:
     struct InfoCreateThread {
         u32 thread_id{};

src/core/hle/kernel/k_object_name.h

@@ -5,7 +5,8 @@
 
 #include <array>
 #include <memory>
-#include <boost/intrusive/list.hpp>
+
+#include "common/intrusive_list.h"
 
 #include "core/hle/kernel/k_light_lock.h"
 #include "core/hle/kernel/slab_helpers.h"
@@ -15,13 +16,14 @@ namespace Kernel {
 
 class KObjectNameGlobalData;
 
-class KObjectName : public KSlabAllocated<KObjectName>, public boost::intrusive::list_base_hook<> {
+class KObjectName : public KSlabAllocated<KObjectName>,
+                    public Common::IntrusiveListBaseNode<KObjectName> {
 public:
     explicit KObjectName(KernelCore&) {}
     virtual ~KObjectName() = default;
 
     static constexpr size_t NameLengthMax = 12;
-    using List = boost::intrusive::list<KObjectName>;
+    using List = Common::IntrusiveListBaseTraits<KObjectName>::ListType;
 
     static Result NewFromName(KernelCore& kernel, KAutoObject* obj, const char* name);
     static Result Delete(KernelCore& kernel, KAutoObject* obj, const char* name);

src/core/hle/kernel/k_server_port.h

@@ -7,7 +7,7 @@
 #include <string>
 #include <utility>
 
-#include <boost/intrusive/list.hpp>
+#include "common/intrusive_list.h"
 
 #include "core/hle/kernel/k_server_session.h"
 #include "core/hle/kernel/k_synchronization_object.h"
@@ -42,7 +42,7 @@ public:
     bool IsSignaled() const override;
 
 private:
-    using SessionList = boost::intrusive::list<KServerSession>;
+    using SessionList = Common::IntrusiveListBaseTraits<KServerSession>::ListType;
 
     void CleanupSessions();
 

src/core/hle/kernel/k_server_session.h

@@ -8,7 +8,7 @@
 #include <string>
 #include <utility>
 
-#include <boost/intrusive/list.hpp>
+#include "common/intrusive_list.h"
 
 #include "core/hle/kernel/k_light_lock.h"
 #include "core/hle/kernel/k_session_request.h"
@@ -27,7 +27,7 @@ class KSession;
 class KThread;
 
 class KServerSession final : public KSynchronizationObject,
-                             public boost::intrusive::list_base_hook<> {
+                             public Common::IntrusiveListBaseNode<KServerSession> {
     KERNEL_AUTOOBJECT_TRAITS(KServerSession, KSynchronizationObject);
 
     friend class ServiceThread;
@@ -67,7 +67,8 @@ private:
     KSession* m_parent{};
 
     /// List of threads which are pending a reply.
-    boost::intrusive::list<KSessionRequest> m_request_list{};
+    using RequestList = Common::IntrusiveListBaseTraits<KSessionRequest>::ListType;
+    RequestList m_request_list{};
     KSessionRequest* m_current_request{};
 
     KLightLock m_lock;

src/core/hle/kernel/k_session_request.h

@@ -5,6 +5,8 @@
 
 #include <array>
 
+#include "common/intrusive_list.h"
+
 #include "core/hle/kernel/k_auto_object.h"
 #include "core/hle/kernel/k_event.h"
 #include "core/hle/kernel/k_memory_block.h"
@@ -16,7 +18,7 @@ namespace Kernel {
 
 class KSessionRequest final : public KSlabAllocated<KSessionRequest>,
                               public KAutoObject,
-                              public boost::intrusive::list_base_hook<> {
+                              public Common::IntrusiveListBaseNode<KSessionRequest> {
     KERNEL_AUTOOBJECT_TRAITS(KSessionRequest, KAutoObject);
 
 public:

src/core/hle/kernel/k_shared_memory_info.h

@@ -3,7 +3,7 @@
 
 #pragma once
 
-#include <boost/intrusive/list.hpp>
+#include "common/intrusive_list.h"
 
 #include "core/hle/kernel/slab_helpers.h"
 
@@ -12,7 +12,7 @@ namespace Kernel {
 class KSharedMemory;
 
 class KSharedMemoryInfo final : public KSlabAllocated<KSharedMemoryInfo>,
-                                public boost::intrusive::list_base_hook<> {
+                                public Common::IntrusiveListBaseNode<KSharedMemoryInfo> {
 
 public:
     explicit KSharedMemoryInfo(KernelCore&) {}

src/core/hle/kernel/k_thread.h

@@ -12,7 +12,7 @@
 #include <utility>
 #include <vector>
 
-#include <boost/intrusive/list.hpp>
+#include "common/intrusive_list.h"
 
 #include "common/intrusive_red_black_tree.h"
 #include "common/spin_lock.h"
@@ -119,7 +119,7 @@ s32 GetCurrentCoreId(KernelCore& kernel);
 Core::Memory::Memory& GetCurrentMemory(KernelCore& kernel);
 
 class KThread final : public KAutoObjectWithSlabHeapAndContainer<KThread, KWorkerTask>,
-                      public boost::intrusive::list_base_hook<>,
+                      public Common::IntrusiveListBaseNode<KThread>,
                       public KTimerTask {
     KERNEL_AUTOOBJECT_TRAITS(KThread, KSynchronizationObject);
 
@@ -138,7 +138,7 @@ public:
 public:
     using ThreadContext32 = Core::ARM_Interface::ThreadContext32;
     using ThreadContext64 = Core::ARM_Interface::ThreadContext64;
-    using WaiterList = boost::intrusive::list<KThread>;
+    using WaiterList = Common::IntrusiveListBaseTraits<KThread>::ListType;
 
     /**
      * Gets the thread's current priority
@@ -750,8 +750,9 @@ private:
         ConditionVariableThreadTreeTraits::TreeType<LockWithPriorityInheritanceComparator>;
 
 public:
-    class LockWithPriorityInheritanceInfo : public KSlabAllocated<LockWithPriorityInheritanceInfo>,
-                                            public boost::intrusive::list_base_hook<> {
+    class LockWithPriorityInheritanceInfo
+        : public KSlabAllocated<LockWithPriorityInheritanceInfo>,
+          public Common::IntrusiveListBaseNode<LockWithPriorityInheritanceInfo> {
     public:
         explicit LockWithPriorityInheritanceInfo(KernelCore&) {}
 
@@ -839,7 +840,7 @@ public:
 
 private:
     using LockWithPriorityInheritanceInfoList =
-        boost::intrusive::list<LockWithPriorityInheritanceInfo>;
+        Common::IntrusiveListBaseTraits<LockWithPriorityInheritanceInfo>::ListType;
 
     ConditionVariableThreadTree* m_condvar_tree{};
     u64 m_condvar_key{};

src/core/hle/kernel/kernel.cpp

@@ -95,7 +95,7 @@ struct KernelCore::Impl {
                                        pt_heap_region.GetSize());
         }
 
-        InitializeHackSharedMemory();
+        InitializeHackSharedMemory(kernel);
         RegisterHostThread(nullptr);
     }
 
@@ -216,10 +216,12 @@ struct KernelCore::Impl {
             auto* main_thread{Kernel::KThread::Create(system.Kernel())};
             main_thread->SetCurrentCore(core);
             ASSERT(Kernel::KThread::InitializeMainThread(system, main_thread, core).IsSuccess());
+            KThread::Register(system.Kernel(), main_thread);
 
             auto* idle_thread{Kernel::KThread::Create(system.Kernel())};
             idle_thread->SetCurrentCore(core);
             ASSERT(Kernel::KThread::InitializeIdleThread(system, idle_thread, core).IsSuccess());
+            KThread::Register(system.Kernel(), idle_thread);
 
             schedulers[i]->Initialize(main_thread, idle_thread, core);
         }
@@ -230,6 +232,7 @@ struct KernelCore::Impl {
                                        const Core::Timing::CoreTiming& core_timing) {
         system_resource_limit = KResourceLimit::Create(system.Kernel());
         system_resource_limit->Initialize(&core_timing);
+        KResourceLimit::Register(kernel, system_resource_limit);
 
         const auto sizes{memory_layout->GetTotalAndKernelMemorySizes()};
         const auto total_size{sizes.first};
@@ -355,6 +358,7 @@ struct KernelCore::Impl {
             ASSERT(KThread::InitializeHighPriorityThread(system, shutdown_threads[core_id], {}, {},
                                                          core_id)
                        .IsSuccess());
+            KThread::Register(system.Kernel(), shutdown_threads[core_id]);
         }
     }
 
@@ -729,7 +733,7 @@ struct KernelCore::Impl {
         memory_manager->Initialize(management_region.GetAddress(), management_region.GetSize());
     }
 
-    void InitializeHackSharedMemory() {
+    void InitializeHackSharedMemory(KernelCore& kernel) {
         // Setup memory regions for emulated processes
         // TODO(bunnei): These should not be hardcoded regions initialized within the kernel
         constexpr std::size_t hid_size{0x40000};
@@ -746,14 +750,23 @@ struct KernelCore::Impl {
 
         hid_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None,
                                    Svc::MemoryPermission::Read, hid_size);
+        KSharedMemory::Register(kernel, hid_shared_mem);
+
         font_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None,
                                     Svc::MemoryPermission::Read, font_size);
+        KSharedMemory::Register(kernel, font_shared_mem);
+
         irs_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None,
                                    Svc::MemoryPermission::Read, irs_size);
+        KSharedMemory::Register(kernel, irs_shared_mem);
+
         time_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None,
                                     Svc::MemoryPermission::Read, time_size);
+        KSharedMemory::Register(kernel, time_shared_mem);
+
         hidbus_shared_mem->Initialize(system.DeviceMemory(), nullptr, Svc::MemoryPermission::None,
                                       Svc::MemoryPermission::Read, hidbus_size);
+        KSharedMemory::Register(kernel, hidbus_shared_mem);
     }
 
     std::mutex registered_objects_lock;
@@ -1072,12 +1085,15 @@ static std::jthread RunHostThreadFunc(KernelCore& kernel, KProcess* process,
     // Commit the thread reservation.
     thread_reservation.Commit();
 
+    // Register the thread.
+    KThread::Register(kernel, thread);
+
     return std::jthread(
         [&kernel, thread, thread_name{std::move(thread_name)}, func{std::move(func)}] {
             // Set the thread name.
             Common::SetCurrentThreadName(thread_name.c_str());
 
-            // Register the thread.
+            // Set the thread as current.
             kernel.RegisterHostThread(thread);
 
             // Run the callback.
@@ -1099,6 +1115,9 @@ std::jthread KernelCore::RunOnHostCoreProcess(std::string&& process_name,
     // Ensure that we don't hold onto any extra references.
     SCOPE_EXIT({ process->Close(); });
 
+    // Register the new process.
+    KProcess::Register(*this, process);
+
     // Run the host thread.
     return RunHostThreadFunc(*this, process, std::move(process_name), std::move(func));
 }
@@ -1124,6 +1143,9 @@ void KernelCore::RunOnGuestCoreProcess(std::string&& process_name, std::function
     // Ensure that we don't hold onto any extra references.
     SCOPE_EXIT({ process->Close(); });
 
+    // Register the new process.
+    KProcess::Register(*this, process);
+
     // Reserve a new thread from the process resource limit.
     KScopedResourceReservation thread_reservation(process, LimitableResource::ThreadCountMax);
     ASSERT(thread_reservation.Succeeded());
@@ -1136,6 +1158,9 @@ void KernelCore::RunOnGuestCoreProcess(std::string&& process_name, std::function
     // Commit the thread reservation.
     thread_reservation.Commit();
 
+    // Register the new thread.
+    KThread::Register(*this, thread);
+
     // Begin running the thread.
     ASSERT(R_SUCCEEDED(thread->Run()));
 }

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

@@ -1807,7 +1807,7 @@ void IApplicationFunctions::GetFriendInvitationStorageChannelEvent(HLERequestCon
 }
 
 void IApplicationFunctions::TryPopFromFriendInvitationStorageChannel(HLERequestContext& ctx) {
-    LOG_WARNING(Service_AM, "(STUBBED) called");
+    LOG_DEBUG(Service_AM, "(STUBBED) called");
 
     IPC::ResponseBuilder rb{ctx, 2};
     rb.Push(AM::ResultNoDataInChannel);

src/core/hle/service/ipc_helpers.h

@@ -156,6 +156,7 @@ public:
 
             auto* session = Kernel::KSession::Create(kernel);
             session->Initialize(nullptr, 0);
+            Kernel::KSession::Register(kernel, session);
 
             auto next_manager = std::make_shared<Service::SessionRequestManager>(
                 kernel, manager->GetServerManager());

src/core/hle/service/kernel_helpers.cpp

@@ -25,6 +25,9 @@ ServiceContext::ServiceContext(Core::System& system_, std::string name_)
                                         Kernel::KProcess::ProcessType::KernelInternal,
                                         kernel.GetSystemResourceLimit())
                .IsSuccess());
+
+    // Register the process.
+    Kernel::KProcess::Register(kernel, process);
     process_created = true;
 }
 

src/core/hle/service/mutex.cpp

@@ -12,6 +12,9 @@ Mutex::Mutex(Core::System& system) : m_system(system) {
     m_event = Kernel::KEvent::Create(system.Kernel());
     m_event->Initialize(nullptr);
 
+    // Register the event.
+    Kernel::KEvent::Register(system.Kernel(), m_event);
+
     ASSERT(R_SUCCEEDED(m_event->Signal()));
 }
 

src/core/hle/service/nfp/nfp.cpp

@@ -4,11 +4,138 @@
 #include "common/logging/log.h"
 #include "core/hle/service/ipc_helpers.h"
 #include "core/hle/service/nfp/nfp.h"
-#include "core/hle/service/nfp/nfp_user.h"
+#include "core/hle/service/nfp/nfp_interface.h"
 #include "core/hle/service/server_manager.h"
 
 namespace Service::NFP {
 
+class IUser final : public Interface {
+public:
+    explicit IUser(Core::System& system_) : Interface(system_, "NFP:IUser") {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, &IUser::Initialize, "Initialize"},
+            {1, &IUser::Finalize, "Finalize"},
+            {2, &IUser::ListDevices, "ListDevices"},
+            {3, &IUser::StartDetection, "StartDetection"},
+            {4, &IUser::StopDetection, "StopDetection"},
+            {5, &IUser::Mount, "Mount"},
+            {6, &IUser::Unmount, "Unmount"},
+            {7, &IUser::OpenApplicationArea, "OpenApplicationArea"},
+            {8, &IUser::GetApplicationArea, "GetApplicationArea"},
+            {9, &IUser::SetApplicationArea, "SetApplicationArea"},
+            {10, &IUser::Flush, "Flush"},
+            {11, &IUser::Restore, "Restore"},
+            {12, &IUser::CreateApplicationArea, "CreateApplicationArea"},
+            {13, &IUser::GetTagInfo, "GetTagInfo"},
+            {14, &IUser::GetRegisterInfo, "GetRegisterInfo"},
+            {15, &IUser::GetCommonInfo, "GetCommonInfo"},
+            {16, &IUser::GetModelInfo, "GetModelInfo"},
+            {17, &IUser::AttachActivateEvent, "AttachActivateEvent"},
+            {18, &IUser::AttachDeactivateEvent, "AttachDeactivateEvent"},
+            {19, &IUser::GetState, "GetState"},
+            {20, &IUser::GetDeviceState, "GetDeviceState"},
+            {21, &IUser::GetNpadId, "GetNpadId"},
+            {22, &IUser::GetApplicationAreaSize, "GetApplicationAreaSize"},
+            {23, &IUser::AttachAvailabilityChangeEvent, "AttachAvailabilityChangeEvent"},
+            {24, &IUser::RecreateApplicationArea, "RecreateApplicationArea"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
+class ISystem final : public Interface {
+public:
+    explicit ISystem(Core::System& system_) : Interface(system_, "NFP:ISystem") {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, &ISystem::InitializeSystem, "InitializeSystem"},
+            {1, &ISystem::FinalizeSystem, "FinalizeSystem"},
+            {2, &ISystem::ListDevices, "ListDevices"},
+            {3, &ISystem::StartDetection, "StartDetection"},
+            {4, &ISystem::StopDetection, "StopDetection"},
+            {5, &ISystem::Mount, "Mount"},
+            {6, &ISystem::Unmount, "Unmount"},
+            {10, &ISystem::Flush, "Flush"},
+            {11, &ISystem::Restore, "Restore"},
+            {12, &ISystem::CreateApplicationArea, "CreateApplicationArea"},
+            {13, &ISystem::GetTagInfo, "GetTagInfo"},
+            {14, &ISystem::GetRegisterInfo, "GetRegisterInfo"},
+            {15, &ISystem::GetCommonInfo, "GetCommonInfo"},
+            {16, &ISystem::GetModelInfo, "GetModelInfo"},
+            {17, &ISystem::AttachActivateEvent, "AttachActivateEvent"},
+            {18, &ISystem::AttachDeactivateEvent, "AttachDeactivateEvent"},
+            {19, &ISystem::GetState, "GetState"},
+            {20, &ISystem::GetDeviceState, "GetDeviceState"},
+            {21, &ISystem::GetNpadId, "GetNpadId"},
+            {23, &ISystem::AttachAvailabilityChangeEvent, "AttachAvailabilityChangeEvent"},
+            {100, &ISystem::Format, "Format"},
+            {101, &ISystem::GetAdminInfo, "GetAdminInfo"},
+            {102, &ISystem::GetRegisterInfoPrivate, "GetRegisterInfoPrivate"},
+            {103, &ISystem::SetRegisterInfoPrivate, "SetRegisterInfoPrivate"},
+            {104, &ISystem::DeleteRegisterInfo, "DeleteRegisterInfo"},
+            {105, &ISystem::DeleteApplicationArea, "DeleteApplicationArea"},
+            {106, &ISystem::ExistsApplicationArea, "ExistsApplicationArea"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
+class IDebug final : public Interface {
+public:
+    explicit IDebug(Core::System& system_) : Interface(system_, "NFP:IDebug") {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, &IDebug::InitializeDebug, "InitializeDebug"},
+            {1, &IDebug::FinalizeDebug, "FinalizeDebug"},
+            {2, &IDebug::ListDevices, "ListDevices"},
+            {3, &IDebug::StartDetection, "StartDetection"},
+            {4, &IDebug::StopDetection, "StopDetection"},
+            {5, &IDebug::Mount, "Mount"},
+            {6, &IDebug::Unmount, "Unmount"},
+            {7, &IDebug::OpenApplicationArea, "OpenApplicationArea"},
+            {8, &IDebug::GetApplicationArea, "GetApplicationArea"},
+            {9, &IDebug::SetApplicationArea, "SetApplicationArea"},
+            {10, &IDebug::Flush, "Flush"},
+            {11, &IDebug::Restore, "Restore"},
+            {12, &IDebug::CreateApplicationArea, "CreateApplicationArea"},
+            {13, &IDebug::GetTagInfo, "GetTagInfo"},
+            {14, &IDebug::GetRegisterInfo, "GetRegisterInfo"},
+            {15, &IDebug::GetCommonInfo, "GetCommonInfo"},
+            {16, &IDebug::GetModelInfo, "GetModelInfo"},
+            {17, &IDebug::AttachActivateEvent, "AttachActivateEvent"},
+            {18, &IDebug::AttachDeactivateEvent, "AttachDeactivateEvent"},
+            {19, &IDebug::GetState, "GetState"},
+            {20, &IDebug::GetDeviceState, "GetDeviceState"},
+            {21, &IDebug::GetNpadId, "GetNpadId"},
+            {22, &IDebug::GetApplicationAreaSize, "GetApplicationAreaSize"},
+            {23, &IDebug::AttachAvailabilityChangeEvent, "AttachAvailabilityChangeEvent"},
+            {24, &IDebug::RecreateApplicationArea, "RecreateApplicationArea"},
+            {100, &IDebug::Format, "Format"},
+            {101, &IDebug::GetAdminInfo, "GetAdminInfo"},
+            {102, &IDebug::GetRegisterInfoPrivate, "GetRegisterInfoPrivate"},
+            {103, &IDebug::SetRegisterInfoPrivate, "SetRegisterInfoPrivate"},
+            {104, &IDebug::DeleteRegisterInfo, "DeleteRegisterInfo"},
+            {105, &IDebug::DeleteApplicationArea, "DeleteApplicationArea"},
+            {106, &IDebug::ExistsApplicationArea, "ExistsApplicationArea"},
+            {200, &IDebug::GetAll, "GetAll"},
+            {201, &IDebug::SetAll, "SetAll"},
+            {202, &IDebug::FlushDebug, "FlushDebug"},
+            {203, &IDebug::BreakTag, "BreakTag"},
+            {204, nullptr, "ReadBackupData"},
+            {205, nullptr, "WriteBackupData"},
+            {206, nullptr, "WriteNtf"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
 class IUserManager final : public ServiceFramework<IUserManager> {
 public:
     explicit IUserManager(Core::System& system_) : ServiceFramework{system_, "nfp:user"} {
@@ -37,10 +164,68 @@ private:
     std::shared_ptr<IUser> user_interface;
 };
 
+class ISystemManager final : public ServiceFramework<ISystemManager> {
+public:
+    explicit ISystemManager(Core::System& system_) : ServiceFramework{system_, "nfp:sys"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, &ISystemManager::CreateSystemInterface, "CreateSystemInterface"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+
+private:
+    void CreateSystemInterface(HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        if (system_interface == nullptr) {
+            system_interface = std::make_shared<ISystem>(system);
+        }
+
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(ResultSuccess);
+        rb.PushIpcInterface<ISystem>(system_interface);
+    }
+
+    std::shared_ptr<ISystem> system_interface;
+};
+
+class IDebugManager final : public ServiceFramework<IDebugManager> {
+public:
+    explicit IDebugManager(Core::System& system_) : ServiceFramework{system_, "nfp:dbg"} {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, &IDebugManager::CreateDebugInterface, "CreateDebugInterface"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+
+private:
+    void CreateDebugInterface(HLERequestContext& ctx) {
+        LOG_DEBUG(Service_NFP, "called");
+
+        if (system_interface == nullptr) {
+            system_interface = std::make_shared<IDebug>(system);
+        }
+
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(ResultSuccess);
+        rb.PushIpcInterface<IDebug>(system_interface);
+    }
+
+    std::shared_ptr<IDebug> system_interface;
+};
+
 void LoopProcess(Core::System& system) {
     auto server_manager = std::make_unique<ServerManager>(system);
 
     server_manager->RegisterNamedService("nfp:user", std::make_shared<IUserManager>(system));
+    server_manager->RegisterNamedService("nfp:sys", std::make_shared<ISystemManager>(system));
+    server_manager->RegisterNamedService("nfp:dbg", std::make_shared<IDebugManager>(system));
     ServerManager::RunServer(std::move(server_manager));
 }
 

src/core/hle/service/nfp/nfp_device.cpp

@@ -29,7 +29,6 @@
 #include "core/hle/service/nfp/amiibo_crypto.h"
 #include "core/hle/service/nfp/nfp_device.h"
 #include "core/hle/service/nfp/nfp_result.h"
-#include "core/hle/service/nfp/nfp_user.h"
 #include "core/hle/service/time/time_manager.h"
 #include "core/hle/service/time/time_zone_content_manager.h"
 #include "core/hle/service/time/time_zone_types.h"
@@ -241,6 +240,42 @@ Result NfpDevice::Flush() {
 
     tag_data.write_counter++;
 
+    FlushWithBreak(BreakType::Normal);
+
+    is_data_moddified = false;
+
+    return ResultSuccess;
+}
+
+Result NfpDevice::FlushDebug() {
+    if (device_state != DeviceState::TagMounted) {
+        LOG_ERROR(Service_NFC, "Wrong device state {}", device_state);
+        if (device_state == DeviceState::TagRemoved) {
+            return TagRemoved;
+        }
+        return WrongDeviceState;
+    }
+
+    if (mount_target == MountTarget::None || mount_target == MountTarget::Rom) {
+        LOG_ERROR(Service_NFC, "Amiibo is read only", device_state);
+        return WrongDeviceState;
+    }
+
+    tag_data.write_counter++;
+
+    FlushWithBreak(BreakType::Normal);
+
+    is_data_moddified = false;
+
+    return ResultSuccess;
+}
+
+Result NfpDevice::FlushWithBreak(BreakType break_type) {
+    if (break_type != BreakType::Normal) {
+        LOG_ERROR(Service_NFC, "Break type not implemented {}", break_type);
+        return WrongDeviceState;
+    }
+
     std::vector<u8> data(sizeof(EncryptedNTAG215File));
     if (is_plain_amiibo) {
         memcpy(data.data(), &tag_data, sizeof(tag_data));
@@ -258,8 +293,6 @@ Result NfpDevice::Flush() {
         return WriteAmiiboFailed;
     }
 
-    is_data_moddified = false;
-
     return ResultSuccess;
 }
 
@@ -417,6 +450,38 @@ Result NfpDevice::GetRegisterInfo(RegisterInfo& register_info) const {
     return ResultSuccess;
 }
 
+Result NfpDevice::GetRegisterInfoPrivate(RegisterInfoPrivate& register_info) const {
+    if (device_state != DeviceState::TagMounted) {
+        LOG_ERROR(Service_NFP, "Wrong device state {}", device_state);
+        if (device_state == DeviceState::TagRemoved) {
+            return TagRemoved;
+        }
+        return WrongDeviceState;
+    }
+
+    if (mount_target == MountTarget::None || mount_target == MountTarget::Rom) {
+        LOG_ERROR(Service_NFP, "Amiibo is read only", device_state);
+        return WrongDeviceState;
+    }
+
+    if (tag_data.settings.settings.amiibo_initialized == 0) {
+        return RegistrationIsNotInitialized;
+    }
+
+    Service::Mii::MiiManager manager;
+    const auto& settings = tag_data.settings;
+
+    // TODO: Validate and complete this data
+    register_info = {
+        .mii_store_data = {},
+        .creation_date = settings.init_date.GetWriteDate(),
+        .amiibo_name = GetAmiiboName(settings),
+        .font_region = settings.settings.font_region,
+    };
+
+    return ResultSuccess;
+}
+
 Result NfpDevice::GetAdminInfo(AdminInfo& admin_info) const {
     if (device_state != DeviceState::TagMounted) {
         LOG_ERROR(Service_NFC, "Wrong device state {}", device_state);
@@ -807,6 +872,159 @@ Result NfpDevice::DeleteApplicationArea() {
     return Flush();
 }
 
+Result NfpDevice::ExistApplicationArea(bool& has_application_area) {
+    if (device_state != DeviceState::TagMounted) {
+        LOG_ERROR(Service_NFC, "Wrong device state {}", device_state);
+        if (device_state == DeviceState::TagRemoved) {
+            return TagRemoved;
+        }
+        return WrongDeviceState;
+    }
+
+    if (mount_target == MountTarget::None || mount_target == MountTarget::Rom) {
+        LOG_ERROR(Service_NFC, "Amiibo is read only", device_state);
+        return WrongDeviceState;
+    }
+
+    has_application_area = tag_data.settings.settings.appdata_initialized.Value() != 0;
+
+    return ResultSuccess;
+}
+
+Result NfpDevice::GetAll(NfpData& data) const {
+    if (device_state != DeviceState::TagMounted) {
+        LOG_ERROR(Service_NFC, "Wrong device state {}", device_state);
+        if (device_state == DeviceState::TagRemoved) {
+            return TagRemoved;
+        }
+        return WrongDeviceState;
+    }
+
+    if (mount_target == MountTarget::None || mount_target == MountTarget::Rom) {
+        LOG_ERROR(Service_NFC, "Amiibo is read only", device_state);
+        return WrongDeviceState;
+    }
+
+    CommonInfo common_info{};
+    Service::Mii::MiiManager manager;
+    const u64 application_id = tag_data.application_id;
+
+    GetCommonInfo(common_info);
+
+    data = {
+        .magic = tag_data.constant_value,
+        .write_counter = tag_data.write_counter,
+        .settings_crc = tag_data.settings.crc,
+        .common_info = common_info,
+        .mii_char_info = tag_data.owner_mii,
+        .mii_store_data_extension = tag_data.mii_extension,
+        .creation_date = tag_data.settings.init_date.GetWriteDate(),
+        .amiibo_name = tag_data.settings.amiibo_name,
+        .amiibo_name_null_terminated = 0,
+        .settings = tag_data.settings.settings,
+        .unknown1 = tag_data.unknown,
+        .register_info_crc = tag_data.register_info_crc,
+        .unknown2 = tag_data.unknown2,
+        .application_id = application_id,
+        .access_id = tag_data.application_area_id,
+        .settings_crc_counter = tag_data.settings.crc_counter,
+        .font_region = tag_data.settings.settings.font_region,
+        .tag_type = PackedTagType::Type2,
+        .console_type =
+            static_cast<AppAreaVersion>(application_id >> application_id_version_offset & 0xf),
+        .application_id_byte = tag_data.application_id_byte,
+        .application_area = tag_data.application_area,
+    };
+
+    return ResultSuccess;
+}
+
+Result NfpDevice::SetAll(const NfpData& data) {
+    if (device_state != DeviceState::TagMounted) {
+        LOG_ERROR(Service_NFC, "Wrong device state {}", device_state);
+        if (device_state == DeviceState::TagRemoved) {
+            return TagRemoved;
+        }
+        return WrongDeviceState;
+    }
+
+    if (mount_target == MountTarget::None || mount_target == MountTarget::Rom) {
+        LOG_ERROR(Service_NFC, "Amiibo is read only", device_state);
+        return WrongDeviceState;
+    }
+
+    tag_data.constant_value = data.magic;
+    tag_data.write_counter = data.write_counter;
+    tag_data.settings.crc = data.settings_crc;
+    tag_data.settings.write_date.SetWriteDate(data.common_info.last_write_date);
+    tag_data.write_counter = data.common_info.write_counter;
+    tag_data.amiibo_version = data.common_info.version;
+    tag_data.owner_mii = data.mii_char_info;
+    tag_data.mii_extension = data.mii_store_data_extension;
+    tag_data.settings.init_date.SetWriteDate(data.creation_date);
+    tag_data.settings.amiibo_name = data.amiibo_name;
+    tag_data.settings.settings = data.settings;
+    tag_data.unknown = data.unknown1;
+    tag_data.register_info_crc = data.register_info_crc;
+    tag_data.unknown2 = data.unknown2;
+    tag_data.application_id = data.application_id;
+    tag_data.application_area_id = data.access_id;
+    tag_data.settings.crc_counter = data.settings_crc_counter;
+    tag_data.settings.settings.font_region.Assign(data.font_region);
+    tag_data.application_id_byte = data.application_id_byte;
+    tag_data.application_area = data.application_area;
+
+    return ResultSuccess;
+}
+
+Result NfpDevice::BreakTag(BreakType break_type) {
+    if (device_state != DeviceState::TagMounted) {
+        LOG_ERROR(Service_NFC, "Wrong device state {}", device_state);
+        if (device_state == DeviceState::TagRemoved) {
+            return TagRemoved;
+        }
+        return WrongDeviceState;
+    }
+
+    if (mount_target == MountTarget::None || mount_target == MountTarget::Rom) {
+        LOG_ERROR(Service_NFC, "Amiibo is read only", device_state);
+        return WrongDeviceState;
+    }
+
+    // TODO: Complete this implementation
+
+    return FlushWithBreak(break_type);
+}
+
+Result NfpDevice::ReadBackupData() {
+    // Not implemented
+    return ResultSuccess;
+}
+
+Result NfpDevice::WriteBackupData() {
+    // Not implemented
+    return ResultSuccess;
+}
+
+Result NfpDevice::WriteNtf() {
+    if (device_state != DeviceState::TagMounted) {
+        LOG_ERROR(Service_NFC, "Wrong device state {}", device_state);
+        if (device_state == DeviceState::TagRemoved) {
+            return TagRemoved;
+        }
+        return WrongDeviceState;
+    }
+
+    if (mount_target == MountTarget::None || mount_target == MountTarget::Rom) {
+        LOG_ERROR(Service_NFC, "Amiibo is read only", device_state);
+        return WrongDeviceState;
+    }
+
+    // Not implemented
+
+    return ResultSuccess;
+}
+
 u64 NfpDevice::GetHandle() const {
     // Generate a handle based of the npad id
     return static_cast<u64>(npad_id);

src/core/hle/service/nfp/nfp_device.h

@@ -41,12 +41,16 @@ public:
     Result StopDetection();
     Result Mount(MountTarget mount_target);
     Result Unmount();
+
     Result Flush();
+    Result FlushDebug();
+    Result FlushWithBreak(BreakType break_type);
 
     Result GetTagInfo(TagInfo& tag_info) const;
     Result GetCommonInfo(CommonInfo& common_info) const;
     Result GetModelInfo(ModelInfo& model_info) const;
     Result GetRegisterInfo(RegisterInfo& register_info) const;
+    Result GetRegisterInfoPrivate(RegisterInfoPrivate& register_info) const;
     Result GetAdminInfo(AdminInfo& admin_info) const;
 
     Result DeleteRegisterInfo();
@@ -61,6 +65,14 @@ public:
     Result CreateApplicationArea(u32 access_id, std::span<const u8> data);
     Result RecreateApplicationArea(u32 access_id, std::span<const u8> data);
     Result DeleteApplicationArea();
+    Result ExistApplicationArea(bool& has_application_area);
+
+    Result GetAll(NfpData& data) const;
+    Result SetAll(const NfpData& data);
+    Result BreakTag(BreakType break_type);
+    Result ReadBackupData();
+    Result WriteBackupData();
+    Result WriteNtf();
 
     u64 GetHandle() const;
     u32 GetApplicationAreaSize() const;

src/core/hle/service/nfp/nfp_interface.cpp

@@ -7,42 +7,13 @@
 #include "core/hle/kernel/k_event.h"
 #include "core/hle/service/ipc_helpers.h"
 #include "core/hle/service/nfp/nfp_device.h"
+#include "core/hle/service/nfp/nfp_interface.h"
 #include "core/hle/service/nfp/nfp_result.h"
-#include "core/hle/service/nfp/nfp_user.h"
 
 namespace Service::NFP {
 
-IUser::IUser(Core::System& system_)
-    : ServiceFramework{system_, "NFP::IUser"}, service_context{system_, service_name} {
-    static const FunctionInfo functions[] = {
-        {0, &IUser::Initialize, "Initialize"},
-        {1, &IUser::Finalize, "Finalize"},
-        {2, &IUser::ListDevices, "ListDevices"},
-        {3, &IUser::StartDetection, "StartDetection"},
-        {4, &IUser::StopDetection, "StopDetection"},
-        {5, &IUser::Mount, "Mount"},
-        {6, &IUser::Unmount, "Unmount"},
-        {7, &IUser::OpenApplicationArea, "OpenApplicationArea"},
-        {8, &IUser::GetApplicationArea, "GetApplicationArea"},
-        {9, &IUser::SetApplicationArea, "SetApplicationArea"},
-        {10, &IUser::Flush, "Flush"},
-        {11, &IUser::Restore, "Restore"},
-        {12, &IUser::CreateApplicationArea, "CreateApplicationArea"},
-        {13, &IUser::GetTagInfo, "GetTagInfo"},
-        {14, &IUser::GetRegisterInfo, "GetRegisterInfo"},
-        {15, &IUser::GetCommonInfo, "GetCommonInfo"},
-        {16, &IUser::GetModelInfo, "GetModelInfo"},
-        {17, &IUser::AttachActivateEvent, "AttachActivateEvent"},
-        {18, &IUser::AttachDeactivateEvent, "AttachDeactivateEvent"},
-        {19, &IUser::GetState, "GetState"},
-        {20, &IUser::GetDeviceState, "GetDeviceState"},
-        {21, &IUser::GetNpadId, "GetNpadId"},
-        {22, &IUser::GetApplicationAreaSize, "GetApplicationAreaSize"},
-        {23, &IUser::AttachAvailabilityChangeEvent, "AttachAvailabilityChangeEvent"},
-        {24, &IUser::RecreateApplicationArea, "RecreateApplicationArea"},
-    };
-    RegisterHandlers(functions);
-
+Interface::Interface(Core::System& system_, const char* name)
+    : ServiceFramework{system_, name}, service_context{system_, service_name} {
     availability_change_event = service_context.CreateEvent("IUser:AvailabilityChangeEvent");
 
     for (u32 device_index = 0; device_index < 10; device_index++) {
@@ -52,11 +23,11 @@ IUser::IUser(Core::System& system_)
     }
 }
 
-IUser ::~IUser() {
+Interface::~Interface() {
     availability_change_event->Close();
 }
 
-void IUser::Initialize(HLERequestContext& ctx) {
+void Interface::Initialize(HLERequestContext& ctx) {
     LOG_INFO(Service_NFP, "called");
 
     state = State::Initialized;
@@ -69,7 +40,33 @@ void IUser::Initialize(HLERequestContext& ctx) {
     rb.Push(ResultSuccess);
 }
 
-void IUser::Finalize(HLERequestContext& ctx) {
+void Interface::InitializeSystem(HLERequestContext& ctx) {
+    LOG_INFO(Service_NFP, "called");
+
+    state = State::Initialized;
+
+    for (auto& device : devices) {
+        device->Initialize();
+    }
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(ResultSuccess);
+}
+
+void Interface::InitializeDebug(HLERequestContext& ctx) {
+    LOG_INFO(Service_NFP, "called");
+
+    state = State::Initialized;
+
+    for (auto& device : devices) {
+        device->Initialize();
+    }
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(ResultSuccess);
+}
+
+void Interface::Finalize(HLERequestContext& ctx) {
     LOG_INFO(Service_NFP, "called");
 
     state = State::NonInitialized;
@@ -82,7 +79,33 @@ void IUser::Finalize(HLERequestContext& ctx) {
     rb.Push(ResultSuccess);
 }
 
-void IUser::ListDevices(HLERequestContext& ctx) {
+void Interface::FinalizeSystem(HLERequestContext& ctx) {
+    LOG_INFO(Service_NFP, "called");
+
+    state = State::NonInitialized;
+
+    for (auto& device : devices) {
+        device->Finalize();
+    }
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(ResultSuccess);
+}
+
+void Interface::FinalizeDebug(HLERequestContext& ctx) {
+    LOG_INFO(Service_NFP, "called");
+
+    state = State::NonInitialized;
+
+    for (auto& device : devices) {
+        device->Finalize();
+    }
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(ResultSuccess);
+}
+
+void Interface::ListDevices(HLERequestContext& ctx) {
     LOG_DEBUG(Service_NFP, "called");
 
     if (state == State::NonInitialized) {
@@ -128,7 +151,7 @@ void IUser::ListDevices(HLERequestContext& ctx) {
     rb.Push(static_cast<s32>(nfp_devices.size()));
 }
 
-void IUser::StartDetection(HLERequestContext& ctx) {
+void Interface::StartDetection(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     const auto nfp_protocol{rp.PopEnum<TagProtocol>()};
@@ -153,7 +176,7 @@ void IUser::StartDetection(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::StopDetection(HLERequestContext& ctx) {
+void Interface::StopDetection(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_INFO(Service_NFP, "called, device_handle={}", device_handle);
@@ -177,7 +200,7 @@ void IUser::StopDetection(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::Mount(HLERequestContext& ctx) {
+void Interface::Mount(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     const auto model_type{rp.PopEnum<ModelType>()};
@@ -204,7 +227,7 @@ void IUser::Mount(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::Unmount(HLERequestContext& ctx) {
+void Interface::Unmount(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_INFO(Service_NFP, "called, device_handle={}", device_handle);
@@ -228,7 +251,7 @@ void IUser::Unmount(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::OpenApplicationArea(HLERequestContext& ctx) {
+void Interface::OpenApplicationArea(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     const auto access_id{rp.Pop<u32>()};
@@ -253,7 +276,7 @@ void IUser::OpenApplicationArea(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::GetApplicationArea(HLERequestContext& ctx) {
+void Interface::GetApplicationArea(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     const auto data_size = ctx.GetWriteBufferSize();
@@ -287,7 +310,7 @@ void IUser::GetApplicationArea(HLERequestContext& ctx) {
     rb.Push(static_cast<u32>(data_size));
 }
 
-void IUser::SetApplicationArea(HLERequestContext& ctx) {
+void Interface::SetApplicationArea(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     const auto data{ctx.ReadBuffer()};
@@ -318,7 +341,7 @@ void IUser::SetApplicationArea(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::Flush(HLERequestContext& ctx) {
+void Interface::Flush(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_INFO(Service_NFP, "called, device_handle={}", device_handle);
@@ -342,7 +365,7 @@ void IUser::Flush(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::Restore(HLERequestContext& ctx) {
+void Interface::Restore(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_WARNING(Service_NFP, "(STUBBED) called, device_handle={}", device_handle);
@@ -366,7 +389,7 @@ void IUser::Restore(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::CreateApplicationArea(HLERequestContext& ctx) {
+void Interface::CreateApplicationArea(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     const auto access_id{rp.Pop<u32>()};
@@ -399,7 +422,7 @@ void IUser::CreateApplicationArea(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::GetTagInfo(HLERequestContext& ctx) {
+void Interface::GetTagInfo(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_INFO(Service_NFP, "called, device_handle={}", device_handle);
@@ -425,7 +448,7 @@ void IUser::GetTagInfo(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::GetRegisterInfo(HLERequestContext& ctx) {
+void Interface::GetRegisterInfo(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_INFO(Service_NFP, "called, device_handle={}", device_handle);
@@ -451,7 +474,7 @@ void IUser::GetRegisterInfo(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::GetCommonInfo(HLERequestContext& ctx) {
+void Interface::GetCommonInfo(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_INFO(Service_NFP, "called, device_handle={}", device_handle);
@@ -477,7 +500,7 @@ void IUser::GetCommonInfo(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::GetModelInfo(HLERequestContext& ctx) {
+void Interface::GetModelInfo(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_INFO(Service_NFP, "called, device_handle={}", device_handle);
@@ -503,7 +526,7 @@ void IUser::GetModelInfo(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-void IUser::AttachActivateEvent(HLERequestContext& ctx) {
+void Interface::AttachActivateEvent(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
@@ -527,7 +550,7 @@ void IUser::AttachActivateEvent(HLERequestContext& ctx) {
     rb.PushCopyObjects(device.value()->GetActivateEvent());
 }
 
-void IUser::AttachDeactivateEvent(HLERequestContext& ctx) {
+void Interface::AttachDeactivateEvent(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
@@ -551,7 +574,7 @@ void IUser::AttachDeactivateEvent(HLERequestContext& ctx) {
     rb.PushCopyObjects(device.value()->GetDeactivateEvent());
 }
 
-void IUser::GetState(HLERequestContext& ctx) {
+void Interface::GetState(HLERequestContext& ctx) {
     LOG_DEBUG(Service_NFP, "called");
 
     IPC::ResponseBuilder rb{ctx, 3};
@@ -559,7 +582,7 @@ void IUser::GetState(HLERequestContext& ctx) {
     rb.PushEnum(state);
 }
 
-void IUser::GetDeviceState(HLERequestContext& ctx) {
+void Interface::GetDeviceState(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
@@ -577,7 +600,7 @@ void IUser::GetDeviceState(HLERequestContext& ctx) {
     rb.PushEnum(device.value()->GetCurrentState());
 }
 
-void IUser::GetNpadId(HLERequestContext& ctx) {
+void Interface::GetNpadId(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
@@ -601,7 +624,7 @@ void IUser::GetNpadId(HLERequestContext& ctx) {
     rb.PushEnum(device.value()->GetNpadId());
 }
 
-void IUser::GetApplicationAreaSize(HLERequestContext& ctx) {
+void Interface::GetApplicationAreaSize(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
@@ -619,7 +642,7 @@ void IUser::GetApplicationAreaSize(HLERequestContext& ctx) {
     rb.Push(device.value()->GetApplicationAreaSize());
 }
 
-void IUser::AttachAvailabilityChangeEvent(HLERequestContext& ctx) {
+void Interface::AttachAvailabilityChangeEvent(HLERequestContext& ctx) {
     LOG_INFO(Service_NFP, "called");
 
     if (state == State::NonInitialized) {
@@ -633,7 +656,7 @@ void IUser::AttachAvailabilityChangeEvent(HLERequestContext& ctx) {
     rb.PushCopyObjects(availability_change_event->GetReadableEvent());
 }
 
-void IUser::RecreateApplicationArea(HLERequestContext& ctx) {
+void Interface::RecreateApplicationArea(HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto device_handle{rp.Pop<u64>()};
     const auto access_id{rp.Pop<u32>()};
@@ -660,7 +683,361 @@ void IUser::RecreateApplicationArea(HLERequestContext& ctx) {
     rb.Push(result);
 }
 
-std::optional<std::shared_ptr<NfpDevice>> IUser::GetNfpDevice(u64 handle) {
+void Interface::Format(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    const auto result = device.value()->Format();
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::GetAdminInfo(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_INFO(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    AdminInfo admin_info{};
+    const auto result = device.value()->GetAdminInfo(admin_info);
+    ctx.WriteBuffer(admin_info);
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::GetRegisterInfoPrivate(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_INFO(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    RegisterInfoPrivate register_info{};
+    const auto result = device.value()->GetRegisterInfoPrivate(register_info);
+    ctx.WriteBuffer(register_info);
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::SetRegisterInfoPrivate(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    const auto buffer{ctx.ReadBuffer()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}, buffer_size={}", device_handle,
+              buffer.size());
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    const auto result = device.value()->SetRegisterInfoPrivate({});
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::DeleteRegisterInfo(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    const auto result = device.value()->DeleteRegisterInfo();
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::DeleteApplicationArea(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    const auto result = device.value()->DeleteApplicationArea();
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::ExistsApplicationArea(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    bool has_application_area = false;
+    const auto result = device.value()->ExistApplicationArea(has_application_area);
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(result);
+    rb.Push(has_application_area);
+}
+
+void Interface::GetAll(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    NfpData data{};
+    const auto result = device.value()->GetAll(data);
+
+    ctx.WriteBuffer(data);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::SetAll(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    const auto nfp_data{ctx.ReadBuffer()};
+
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    NfpData data{};
+    memcpy(&data, nfp_data.data(), sizeof(NfpData));
+
+    const auto result = device.value()->SetAll(data);
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::FlushDebug(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    const auto result = device.value()->FlushDebug();
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::BreakTag(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    const auto break_type{rp.PopEnum<BreakType>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}, break_type={}", device_handle, break_type);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    const auto result = device.value()->BreakTag(break_type);
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::ReadBackupData(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    const auto result = device.value()->ReadBackupData();
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::WriteBackupData(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    const auto result = device.value()->WriteBackupData();
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+void Interface::WriteNtf(HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto device_handle{rp.Pop<u64>()};
+    LOG_DEBUG(Service_NFP, "called, device_handle={}", device_handle);
+
+    if (state == State::NonInitialized) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(NfcDisabled);
+        return;
+    }
+
+    auto device = GetNfpDevice(device_handle);
+
+    if (!device.has_value()) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(DeviceNotFound);
+        return;
+    }
+
+    const auto result = device.value()->WriteNtf();
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(result);
+}
+
+std::optional<std::shared_ptr<NfpDevice>> Interface::GetNfpDevice(u64 handle) {
     for (auto& device : devices) {
         if (device->GetHandle() == handle) {
             return device;

src/core/hle/service/nfp/nfp_interface.h

@@ -13,19 +13,17 @@
 namespace Service::NFP {
 class NfpDevice;
 
-class IUser final : public ServiceFramework<IUser> {
+class Interface : public ServiceFramework<Interface> {
 public:
-    explicit IUser(Core::System& system_);
-    ~IUser();
-
-private:
-    enum class State : u32 {
-        NonInitialized,
-        Initialized,
-    };
+    explicit Interface(Core::System& system_, const char* name);
+    ~Interface() override;
 
     void Initialize(HLERequestContext& ctx);
+    void InitializeSystem(HLERequestContext& ctx);
+    void InitializeDebug(HLERequestContext& ctx);
     void Finalize(HLERequestContext& ctx);
+    void FinalizeSystem(HLERequestContext& ctx);
+    void FinalizeDebug(HLERequestContext& ctx);
     void ListDevices(HLERequestContext& ctx);
     void StartDetection(HLERequestContext& ctx);
     void StopDetection(HLERequestContext& ctx);
@@ -49,6 +47,26 @@ private:
     void GetApplicationAreaSize(HLERequestContext& ctx);
     void AttachAvailabilityChangeEvent(HLERequestContext& ctx);
     void RecreateApplicationArea(HLERequestContext& ctx);
+    void Format(HLERequestContext& ctx);
+    void GetAdminInfo(HLERequestContext& ctx);
+    void GetRegisterInfoPrivate(HLERequestContext& ctx);
+    void SetRegisterInfoPrivate(HLERequestContext& ctx);
+    void DeleteRegisterInfo(HLERequestContext& ctx);
+    void DeleteApplicationArea(HLERequestContext& ctx);
+    void ExistsApplicationArea(HLERequestContext& ctx);
+    void GetAll(HLERequestContext& ctx);
+    void SetAll(HLERequestContext& ctx);
+    void FlushDebug(HLERequestContext& ctx);
+    void BreakTag(HLERequestContext& ctx);
+    void ReadBackupData(HLERequestContext& ctx);
+    void WriteBackupData(HLERequestContext& ctx);
+    void WriteNtf(HLERequestContext& ctx);
+
+private:
+    enum class State : u32 {
+        NonInitialized,
+        Initialized,
+    };
 
     std::optional<std::shared_ptr<NfpDevice>> GetNfpDevice(u64 handle);
 

src/core/hle/service/nfp/nfp_types.h

@@ -109,6 +109,12 @@ enum class AppAreaVersion : u8 {
     NotSet = 0xFF,
 };
 
+enum class BreakType : u32 {
+    Normal,
+    Unknown1,
+    Unknown2,
+};
+
 enum class CabinetMode : u8 {
     StartNicknameAndOwnerSettings,
     StartGameDataEraser,
@@ -181,6 +187,12 @@ struct AmiiboDate {
         };
     }
 
+    void SetWriteDate(const WriteDate& write_date) {
+        SetYear(write_date.year);
+        SetMonth(write_date.month);
+        SetDay(write_date.day);
+    }
+
     void SetYear(u16 year) {
         const u16 year_converted = static_cast<u16>((year - 2000) << 9);
         raw_date = Common::swap16((GetValue() & ~0xFE00) | year_converted);
@@ -354,6 +366,15 @@ struct RegisterInfo {
 };
 static_assert(sizeof(RegisterInfo) == 0x100, "RegisterInfo is an invalid size");
 
+struct RegisterInfoPrivate {
+    Service::Mii::MiiStoreData mii_store_data;
+    WriteDate creation_date;
+    AmiiboName amiibo_name;
+    u8 font_region;
+    INSERT_PADDING_BYTES(0x8E);
+};
+static_assert(sizeof(RegisterInfoPrivate) == 0x100, "RegisterInfoPrivate is an invalid size");
+
 struct AdminInfo {
     u64 application_id;
     u32 application_area_id;
@@ -366,6 +387,39 @@ struct AdminInfo {
 };
 static_assert(sizeof(AdminInfo) == 0x40, "AdminInfo is an invalid size");
 
+#pragma pack(1)
+// This is nn::nfp::NfpData
+struct NfpData {
+    u8 magic;
+    INSERT_PADDING_BYTES(0x1);
+    u8 write_counter;
+    INSERT_PADDING_BYTES(0x1);
+    u32 settings_crc;
+    INSERT_PADDING_BYTES(0x38);
+    CommonInfo common_info;
+    Service::Mii::Ver3StoreData mii_char_info;
+    Service::Mii::NfpStoreDataExtension mii_store_data_extension;
+    WriteDate creation_date;
+    std::array<u16_be, amiibo_name_length> amiibo_name;
+    u16 amiibo_name_null_terminated;
+    Settings settings;
+    u8 unknown1;
+    u32 register_info_crc;
+    std::array<u32, 5> unknown2;
+    INSERT_PADDING_BYTES(0x64);
+    u64 application_id;
+    u32 access_id;
+    u16 settings_crc_counter;
+    u8 font_region;
+    PackedTagType tag_type;
+    AppAreaVersion console_type;
+    u8 application_id_byte;
+    INSERT_PADDING_BYTES(0x2E);
+    ApplicationArea application_area;
+};
+static_assert(sizeof(NfpData) == 0x298, "NfpData is an invalid size");
+#pragma pack()
+
 struct SectorKey {
     MifareCmd command;
     u8 unknown; // Usually 1

src/core/hle/service/server_manager.cpp

@@ -33,6 +33,9 @@ ServerManager::ServerManager(Core::System& system) : m_system{system}, m_serve_m
     // Initialize event.
     m_event = Kernel::KEvent::Create(system.Kernel());
     m_event->Initialize(nullptr);
+
+    // Register event.
+    Kernel::KEvent::Register(system.Kernel(), m_event);
 }
 
 ServerManager::~ServerManager() {
@@ -160,6 +163,9 @@ Result ServerManager::ManageDeferral(Kernel::KEvent** out_event) {
     // Initialize the event.
     m_deferral_event->Initialize(nullptr);
 
+    // Register the event.
+    Kernel::KEvent::Register(m_system.Kernel(), m_deferral_event);
+
     // Set the output.
     *out_event = m_deferral_event;
 

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

@@ -64,6 +64,9 @@ Result ServiceManager::RegisterService(std::string name, u32 max_sessions,
     auto* port = Kernel::KPort::Create(kernel);
     port->Initialize(ServerSessionCountMax, false, 0);
 
+    // Register the port.
+    Kernel::KPort::Register(kernel, port);
+
     service_ports.emplace(name, port);
     registered_services.emplace(name, handler);
     if (deferral_event) {

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

@@ -49,6 +49,9 @@ void Controller::CloneCurrentObject(HLERequestContext& ctx) {
     // Commit the session reservation.
     session_reservation.Commit();
 
+    // Register the session.
+    Kernel::KSession::Register(system.Kernel(), session);
+
     // Register with server manager.
     session_manager->GetServerManager().RegisterSession(&session->GetServerSession(),
                                                         session_manager);

src/core/memory.cpp

@@ -462,7 +462,7 @@ struct Memory::Impl {
         }
 
         if (Settings::IsFastmemEnabled()) {
-            const bool is_read_enable = Settings::IsGPULevelHigh() || !cached;
+            const bool is_read_enable = !Settings::IsGPULevelExtreme() || !cached;
             system.DeviceMemory().buffer.Protect(vaddr, size, is_read_enable, !cached);
         }
 

src/dedicated_room/yuzu_room.cpp

@@ -49,6 +49,7 @@ static void PrintHelp(const char* argv0) {
              " [options] <filename>\n"
              "--room-name         The name of the room\n"
              "--room-description  The room description\n"
+             "--bind-address      The bind address for the room\n"
              "--port              The port used for the room\n"
              "--max_members       The maximum number of players for this room\n"
              "--password          The password for the room\n"
@@ -195,6 +196,7 @@ int main(int argc, char** argv) {
     std::string web_api_url;
     std::string ban_list_file;
     std::string log_file = "yuzu-room.log";
+    std::string bind_address;
     u64 preferred_game_id = 0;
     u32 port = Network::DefaultRoomPort;
     u32 max_members = 16;
@@ -203,6 +205,7 @@ int main(int argc, char** argv) {
     static struct option long_options[] = {
         {"room-name", required_argument, 0, 'n'},
         {"room-description", required_argument, 0, 'd'},
+        {"bind-address", required_argument, 0, 's'},
         {"port", required_argument, 0, 'p'},
         {"max_members", required_argument, 0, 'm'},
         {"password", required_argument, 0, 'w'},
@@ -222,7 +225,8 @@ int main(int argc, char** argv) {
     InitializeLogging(log_file);
 
     while (optind < argc) {
-        int arg = getopt_long(argc, argv, "n:d:p:m:w:g:u:t:a:i:l:hv", long_options, &option_index);
+        int arg =
+            getopt_long(argc, argv, "n:d:s:p:m:w:g:u:t:a:i:l:hv", long_options, &option_index);
         if (arg != -1) {
             switch (static_cast<char>(arg)) {
             case 'n':
@@ -231,6 +235,9 @@ int main(int argc, char** argv) {
             case 'd':
                 room_description.assign(optarg);
                 break;
+            case 's':
+                bind_address.assign(optarg);
+                break;
             case 'p':
                 port = strtoul(optarg, &endarg, 0);
                 break;
@@ -295,6 +302,9 @@ int main(int argc, char** argv) {
         PrintHelp(argv[0]);
         return -1;
     }
+    if (bind_address.empty()) {
+        LOG_INFO(Network, "Bind address is empty: defaulting to 0.0.0.0");
+    }
     if (port > UINT16_MAX) {
         LOG_ERROR(Network, "Port needs to be in the range 0 - 65535!");
         PrintHelp(argv[0]);
@@ -358,8 +368,8 @@ int main(int argc, char** argv) {
     if (auto room = network.GetRoom().lock()) {
         AnnounceMultiplayerRoom::GameInfo preferred_game_info{.name = preferred_game,
                                                               .id = preferred_game_id};
-        if (!room->Create(room_name, room_description, "", port, password, max_members, username,
-                          preferred_game_info, std::move(verify_backend), ban_list,
+        if (!room->Create(room_name, room_description, bind_address, port, password, max_members,
+                          username, preferred_game_info, std::move(verify_backend), ban_list,
                           enable_yuzu_mods)) {
             LOG_INFO(Network, "Failed to create room: ");
             return -1;

src/shader_recompiler/backend/spirv/emit_spirv_context_get_set.cpp

@@ -473,7 +473,8 @@ void EmitSetFragColor(EmitContext& ctx, u32 index, u32 component, Id value) {
 }
 
 void EmitSetSampleMask(EmitContext& ctx, Id value) {
-    ctx.OpStore(ctx.sample_mask, value);
+    const Id pointer{ctx.OpAccessChain(ctx.output_u32, ctx.sample_mask, ctx.u32_zero_value)};
+    ctx.OpStore(pointer, value);
 }
 
 void EmitSetFragDepth(EmitContext& ctx, Id value) {

src/shader_recompiler/backend/spirv/spirv_emit_context.cpp

@@ -1572,7 +1572,8 @@ void EmitContext::DefineOutputs(const IR::Program& program) {
             Decorate(frag_depth, spv::Decoration::BuiltIn, spv::BuiltIn::FragDepth);
         }
         if (info.stores_sample_mask) {
-            sample_mask = DefineOutput(*this, U32[1], std::nullopt);
+            const Id array_type{TypeArray(U32[1], Const(1U))};
+            sample_mask = DefineOutput(*this, array_type, std::nullopt);
             Decorate(sample_mask, spv::Decoration::BuiltIn, spv::BuiltIn::SampleMask);
         }
         break;

src/tests/CMakeLists.txt

@@ -15,7 +15,7 @@ add_executable(tests
     core/core_timing.cpp
     core/internal_network/network.cpp
     precompiled_headers.h
-    video_core/buffer_base.cpp
+    video_core/memory_tracker.cpp
     input_common/calibration_configuration_job.cpp
 )
 

src/tests/video_core/buffer_base.cpp

@@ -1,549 +0,0 @@
-// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
-// SPDX-License-Identifier: GPL-2.0-or-later
-
-#include <stdexcept>
-#include <unordered_map>
-
-#include <catch2/catch_test_macros.hpp>
-
-#include "common/alignment.h"
-#include "common/common_types.h"
-#include "video_core/buffer_cache/buffer_base.h"
-
-namespace {
-using VideoCommon::BufferBase;
-using Range = std::pair<u64, u64>;
-
-constexpr u64 PAGE = 4096;
-constexpr u64 WORD = 4096 * 64;
-
-constexpr VAddr c = 0x1328914000;
-
-class RasterizerInterface {
-public:
-    void UpdatePagesCachedCount(VAddr addr, u64 size, int delta) {
-        const u64 page_start{addr >> Core::Memory::YUZU_PAGEBITS};
-        const u64 page_end{(addr + size + Core::Memory::YUZU_PAGESIZE - 1) >>
-                           Core::Memory::YUZU_PAGEBITS};
-        for (u64 page = page_start; page < page_end; ++page) {
-            int& value = page_table[page];
-            value += delta;
-            if (value < 0) {
-                throw std::logic_error{"negative page"};
-            }
-            if (value == 0) {
-                page_table.erase(page);
-            }
-        }
-    }
-
-    [[nodiscard]] int Count(VAddr addr) const noexcept {
-        const auto it = page_table.find(addr >> Core::Memory::YUZU_PAGEBITS);
-        return it == page_table.end() ? 0 : it->second;
-    }
-
-    [[nodiscard]] unsigned Count() const noexcept {
-        unsigned count = 0;
-        for (const auto& [index, value] : page_table) {
-            count += value;
-        }
-        return count;
-    }
-
-private:
-    std::unordered_map<u64, int> page_table;
-};
-} // Anonymous namespace
-
-TEST_CASE("BufferBase: Small buffer", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    REQUIRE(rasterizer.Count() == 0);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    REQUIRE(rasterizer.Count() == WORD / PAGE);
-    REQUIRE(buffer.ModifiedCpuRegion(c, WORD) == Range{0, 0});
-
-    buffer.MarkRegionAsCpuModified(c + PAGE, 1);
-    REQUIRE(buffer.ModifiedCpuRegion(c, WORD) == Range{PAGE * 1, PAGE * 2});
-}
-
-TEST_CASE("BufferBase: Large buffer", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 32);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 32);
-    buffer.MarkRegionAsCpuModified(c + 4096, WORD * 4);
-    REQUIRE(buffer.ModifiedCpuRegion(c, WORD + PAGE * 2) == Range{PAGE, WORD + PAGE * 2});
-    REQUIRE(buffer.ModifiedCpuRegion(c + PAGE * 2, PAGE * 6) == Range{PAGE * 2, PAGE * 8});
-    REQUIRE(buffer.ModifiedCpuRegion(c, WORD * 32) == Range{PAGE, WORD * 4 + PAGE});
-    REQUIRE(buffer.ModifiedCpuRegion(c + WORD * 4, PAGE) == Range{WORD * 4, WORD * 4 + PAGE});
-    REQUIRE(buffer.ModifiedCpuRegion(c + WORD * 3 + PAGE * 63, PAGE) ==
-            Range{WORD * 3 + PAGE * 63, WORD * 4});
-
-    buffer.MarkRegionAsCpuModified(c + WORD * 5 + PAGE * 6, PAGE);
-    buffer.MarkRegionAsCpuModified(c + WORD * 5 + PAGE * 8, PAGE);
-    REQUIRE(buffer.ModifiedCpuRegion(c + WORD * 5, WORD) ==
-            Range{WORD * 5 + PAGE * 6, WORD * 5 + PAGE * 9});
-
-    buffer.UnmarkRegionAsCpuModified(c + WORD * 5 + PAGE * 8, PAGE);
-    REQUIRE(buffer.ModifiedCpuRegion(c + WORD * 5, WORD) ==
-            Range{WORD * 5 + PAGE * 6, WORD * 5 + PAGE * 7});
-
-    buffer.MarkRegionAsCpuModified(c + PAGE, WORD * 31 + PAGE * 63);
-    REQUIRE(buffer.ModifiedCpuRegion(c, WORD * 32) == Range{PAGE, WORD * 32});
-
-    buffer.UnmarkRegionAsCpuModified(c + PAGE * 4, PAGE);
-    buffer.UnmarkRegionAsCpuModified(c + PAGE * 6, PAGE);
-
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 32);
-    REQUIRE(buffer.ModifiedCpuRegion(c, WORD * 32) == Range{0, 0});
-}
-
-TEST_CASE("BufferBase: Rasterizer counting", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, PAGE * 2);
-    REQUIRE(rasterizer.Count() == 0);
-    buffer.UnmarkRegionAsCpuModified(c, PAGE);
-    REQUIRE(rasterizer.Count() == 1);
-    buffer.MarkRegionAsCpuModified(c, PAGE * 2);
-    REQUIRE(rasterizer.Count() == 0);
-    buffer.UnmarkRegionAsCpuModified(c, PAGE);
-    buffer.UnmarkRegionAsCpuModified(c + PAGE, PAGE);
-    REQUIRE(rasterizer.Count() == 2);
-    buffer.MarkRegionAsCpuModified(c, PAGE * 2);
-    REQUIRE(rasterizer.Count() == 0);
-}
-
-TEST_CASE("BufferBase: Basic range", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    buffer.MarkRegionAsCpuModified(c, PAGE);
-    int num = 0;
-    buffer.ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) {
-        REQUIRE(offset == 0U);
-        REQUIRE(size == PAGE);
-        ++num;
-    });
-    REQUIRE(num == 1U);
-}
-
-TEST_CASE("BufferBase: Border upload", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 2);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 2);
-    buffer.MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
-    buffer.ForEachUploadRange(c, WORD * 2, [](u64 offset, u64 size) {
-        REQUIRE(offset == WORD - PAGE);
-        REQUIRE(size == PAGE * 2);
-    });
-}
-
-TEST_CASE("BufferBase: Border upload range", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 2);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 2);
-    buffer.MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
-    buffer.ForEachUploadRange(c + WORD - PAGE, PAGE * 2, [](u64 offset, u64 size) {
-        REQUIRE(offset == WORD - PAGE);
-        REQUIRE(size == PAGE * 2);
-    });
-    buffer.MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
-    buffer.ForEachUploadRange(c + WORD - PAGE, PAGE, [](u64 offset, u64 size) {
-        REQUIRE(offset == WORD - PAGE);
-        REQUIRE(size == PAGE);
-    });
-    buffer.ForEachUploadRange(c + WORD, PAGE, [](u64 offset, u64 size) {
-        REQUIRE(offset == WORD);
-        REQUIRE(size == PAGE);
-    });
-}
-
-TEST_CASE("BufferBase: Border upload partial range", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 2);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 2);
-    buffer.MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
-    buffer.ForEachUploadRange(c + WORD - 1, 2, [](u64 offset, u64 size) {
-        REQUIRE(offset == WORD - PAGE);
-        REQUIRE(size == PAGE * 2);
-    });
-    buffer.MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
-    buffer.ForEachUploadRange(c + WORD - 1, 1, [](u64 offset, u64 size) {
-        REQUIRE(offset == WORD - PAGE);
-        REQUIRE(size == PAGE);
-    });
-    buffer.ForEachUploadRange(c + WORD + 50, 1, [](u64 offset, u64 size) {
-        REQUIRE(offset == WORD);
-        REQUIRE(size == PAGE);
-    });
-}
-
-TEST_CASE("BufferBase: Partial word uploads", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, 0x9d000);
-    int num = 0;
-    buffer.ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) {
-        REQUIRE(offset == 0U);
-        REQUIRE(size == WORD);
-        ++num;
-    });
-    REQUIRE(num == 1);
-    buffer.ForEachUploadRange(c + WORD, WORD, [&](u64 offset, u64 size) {
-        REQUIRE(offset == WORD);
-        REQUIRE(size == WORD);
-        ++num;
-    });
-    REQUIRE(num == 2);
-    buffer.ForEachUploadRange(c + 0x79000, 0x24000, [&](u64 offset, u64 size) {
-        REQUIRE(offset == WORD * 2);
-        REQUIRE(size == PAGE * 0x1d);
-        ++num;
-    });
-    REQUIRE(num == 3);
-}
-
-TEST_CASE("BufferBase: Partial page upload", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    int num = 0;
-    buffer.MarkRegionAsCpuModified(c + PAGE * 2, PAGE);
-    buffer.MarkRegionAsCpuModified(c + PAGE * 9, PAGE);
-    buffer.ForEachUploadRange(c, PAGE * 3, [&](u64 offset, u64 size) {
-        REQUIRE(offset == PAGE * 2);
-        REQUIRE(size == PAGE);
-        ++num;
-    });
-    REQUIRE(num == 1);
-    buffer.ForEachUploadRange(c + PAGE * 7, PAGE * 3, [&](u64 offset, u64 size) {
-        REQUIRE(offset == PAGE * 9);
-        REQUIRE(size == PAGE);
-        ++num;
-    });
-    REQUIRE(num == 2);
-}
-
-TEST_CASE("BufferBase: Partial page upload with multiple words on the right") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 8);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 8);
-    buffer.MarkRegionAsCpuModified(c + PAGE * 13, WORD * 7);
-    int num = 0;
-    buffer.ForEachUploadRange(c + PAGE * 10, WORD * 7, [&](u64 offset, u64 size) {
-        REQUIRE(offset == PAGE * 13);
-        REQUIRE(size == WORD * 7 - PAGE * 3);
-        ++num;
-    });
-    REQUIRE(num == 1);
-    buffer.ForEachUploadRange(c + PAGE, WORD * 8, [&](u64 offset, u64 size) {
-        REQUIRE(offset == WORD * 7 + PAGE * 10);
-        REQUIRE(size == PAGE * 3);
-        ++num;
-    });
-    REQUIRE(num == 2);
-}
-
-TEST_CASE("BufferBase: Partial page upload with multiple words on the left", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 8);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 8);
-    buffer.MarkRegionAsCpuModified(c + PAGE * 13, WORD * 7);
-    int num = 0;
-    buffer.ForEachUploadRange(c + PAGE * 16, WORD * 7, [&](u64 offset, u64 size) {
-        REQUIRE(offset == PAGE * 16);
-        REQUIRE(size == WORD * 7 - PAGE * 3);
-        ++num;
-    });
-    REQUIRE(num == 1);
-    buffer.ForEachUploadRange(c + PAGE, WORD, [&](u64 offset, u64 size) {
-        REQUIRE(offset == PAGE * 13);
-        REQUIRE(size == PAGE * 3);
-        ++num;
-    });
-    REQUIRE(num == 2);
-}
-
-TEST_CASE("BufferBase: Partial page upload with multiple words in the middle", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 8);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 8);
-    buffer.MarkRegionAsCpuModified(c + PAGE * 13, PAGE * 140);
-    int num = 0;
-    buffer.ForEachUploadRange(c + PAGE * 16, WORD, [&](u64 offset, u64 size) {
-        REQUIRE(offset == PAGE * 16);
-        REQUIRE(size == WORD);
-        ++num;
-    });
-    REQUIRE(num == 1);
-    buffer.ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) {
-        REQUIRE(offset == PAGE * 13);
-        REQUIRE(size == PAGE * 3);
-        ++num;
-    });
-    REQUIRE(num == 2);
-    buffer.ForEachUploadRange(c, WORD * 8, [&](u64 offset, u64 size) {
-        REQUIRE(offset == WORD + PAGE * 16);
-        REQUIRE(size == PAGE * 73);
-        ++num;
-    });
-    REQUIRE(num == 3);
-}
-
-TEST_CASE("BufferBase: Empty right bits", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 2048);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 2048);
-    buffer.MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
-    buffer.ForEachUploadRange(c, WORD * 2048, [](u64 offset, u64 size) {
-        REQUIRE(offset == WORD - PAGE);
-        REQUIRE(size == PAGE * 2);
-    });
-}
-
-TEST_CASE("BufferBase: Out of bound ranges 1", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    buffer.MarkRegionAsCpuModified(c, PAGE);
-    int num = 0;
-    buffer.ForEachUploadRange(c - WORD, WORD, [&](u64 offset, u64 size) { ++num; });
-    buffer.ForEachUploadRange(c + WORD, WORD, [&](u64 offset, u64 size) { ++num; });
-    buffer.ForEachUploadRange(c - PAGE, PAGE, [&](u64 offset, u64 size) { ++num; });
-    REQUIRE(num == 0);
-    buffer.ForEachUploadRange(c - PAGE, PAGE * 2, [&](u64 offset, u64 size) { ++num; });
-    REQUIRE(num == 1);
-    buffer.MarkRegionAsCpuModified(c, WORD);
-    REQUIRE(rasterizer.Count() == 0);
-}
-
-TEST_CASE("BufferBase: Out of bound ranges 2", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, 0x22000);
-    REQUIRE_NOTHROW(buffer.UnmarkRegionAsCpuModified(c + 0x22000, PAGE));
-    REQUIRE_NOTHROW(buffer.UnmarkRegionAsCpuModified(c + 0x28000, PAGE));
-    REQUIRE(rasterizer.Count() == 0);
-    REQUIRE_NOTHROW(buffer.UnmarkRegionAsCpuModified(c + 0x21100, PAGE - 0x100));
-    REQUIRE(rasterizer.Count() == 1);
-    REQUIRE_NOTHROW(buffer.UnmarkRegionAsCpuModified(c - 0x1000, PAGE * 2));
-    buffer.UnmarkRegionAsCpuModified(c - 0x3000, PAGE * 2);
-    buffer.UnmarkRegionAsCpuModified(c - 0x2000, PAGE * 2);
-    REQUIRE(rasterizer.Count() == 2);
-}
-
-TEST_CASE("BufferBase: Out of bound ranges 3", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, 0x310720);
-    buffer.UnmarkRegionAsCpuModified(c, 0x310720);
-    REQUIRE(rasterizer.Count(c) == 1);
-    REQUIRE(rasterizer.Count(c + PAGE) == 1);
-    REQUIRE(rasterizer.Count(c + WORD) == 1);
-    REQUIRE(rasterizer.Count(c + WORD + PAGE) == 1);
-}
-
-TEST_CASE("BufferBase: Sparse regions 1", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    buffer.MarkRegionAsCpuModified(c + PAGE * 1, PAGE);
-    buffer.MarkRegionAsCpuModified(c + PAGE * 3, PAGE * 4);
-    buffer.ForEachUploadRange(c, WORD, [i = 0](u64 offset, u64 size) mutable {
-        static constexpr std::array<u64, 2> offsets{PAGE, PAGE * 3};
-        static constexpr std::array<u64, 2> sizes{PAGE, PAGE * 4};
-        REQUIRE(offset == offsets.at(i));
-        REQUIRE(size == sizes.at(i));
-        ++i;
-    });
-}
-
-TEST_CASE("BufferBase: Sparse regions 2", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, 0x22000);
-    buffer.UnmarkRegionAsCpuModified(c, 0x22000);
-    REQUIRE(rasterizer.Count() == 0x22);
-    buffer.MarkRegionAsCpuModified(c + PAGE * 0x1B, PAGE);
-    buffer.MarkRegionAsCpuModified(c + PAGE * 0x21, PAGE);
-    buffer.ForEachUploadRange(c, WORD, [i = 0](u64 offset, u64 size) mutable {
-        static constexpr std::array<u64, 2> offsets{PAGE * 0x1B, PAGE * 0x21};
-        static constexpr std::array<u64, 2> sizes{PAGE, PAGE};
-        REQUIRE(offset == offsets.at(i));
-        REQUIRE(size == sizes.at(i));
-        ++i;
-    });
-}
-
-TEST_CASE("BufferBase: Single page modified range", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, PAGE);
-    REQUIRE(buffer.IsRegionCpuModified(c, PAGE));
-    buffer.UnmarkRegionAsCpuModified(c, PAGE);
-    REQUIRE(!buffer.IsRegionCpuModified(c, PAGE));
-}
-
-TEST_CASE("BufferBase: Two page modified range", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, PAGE * 2);
-    REQUIRE(buffer.IsRegionCpuModified(c, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c, PAGE * 2));
-    buffer.UnmarkRegionAsCpuModified(c, PAGE);
-    REQUIRE(!buffer.IsRegionCpuModified(c, PAGE));
-}
-
-TEST_CASE("BufferBase: Multi word modified ranges", "[video_core]") {
-    for (int offset = 0; offset < 4; ++offset) {
-        const VAddr address = c + WORD * offset;
-        RasterizerInterface rasterizer;
-        BufferBase buffer(rasterizer, address, WORD * 4);
-        REQUIRE(buffer.IsRegionCpuModified(address, PAGE));
-        REQUIRE(buffer.IsRegionCpuModified(address + PAGE * 48, PAGE));
-        REQUIRE(buffer.IsRegionCpuModified(address + PAGE * 56, PAGE));
-
-        buffer.UnmarkRegionAsCpuModified(address + PAGE * 32, PAGE);
-        REQUIRE(buffer.IsRegionCpuModified(address + PAGE, WORD));
-        REQUIRE(buffer.IsRegionCpuModified(address + PAGE * 31, PAGE));
-        REQUIRE(!buffer.IsRegionCpuModified(address + PAGE * 32, PAGE));
-        REQUIRE(buffer.IsRegionCpuModified(address + PAGE * 33, PAGE));
-        REQUIRE(buffer.IsRegionCpuModified(address + PAGE * 31, PAGE * 2));
-        REQUIRE(buffer.IsRegionCpuModified(address + PAGE * 32, PAGE * 2));
-
-        buffer.UnmarkRegionAsCpuModified(address + PAGE * 33, PAGE);
-        REQUIRE(!buffer.IsRegionCpuModified(address + PAGE * 32, PAGE * 2));
-    }
-}
-
-TEST_CASE("BufferBase: Single page in large buffer", "[video_core]") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 16);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 16);
-    REQUIRE(!buffer.IsRegionCpuModified(c, WORD * 16));
-
-    buffer.MarkRegionAsCpuModified(c + WORD * 12 + PAGE * 8, PAGE);
-    REQUIRE(buffer.IsRegionCpuModified(c, WORD * 16));
-    REQUIRE(buffer.IsRegionCpuModified(c + WORD * 10, WORD * 2));
-    REQUIRE(buffer.IsRegionCpuModified(c + WORD * 11, WORD * 2));
-    REQUIRE(buffer.IsRegionCpuModified(c + WORD * 12, WORD * 2));
-    REQUIRE(buffer.IsRegionCpuModified(c + WORD * 12 + PAGE * 4, PAGE * 8));
-    REQUIRE(buffer.IsRegionCpuModified(c + WORD * 12 + PAGE * 6, PAGE * 8));
-    REQUIRE(!buffer.IsRegionCpuModified(c + WORD * 12 + PAGE * 6, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + WORD * 12 + PAGE * 7, PAGE * 2));
-    REQUIRE(buffer.IsRegionCpuModified(c + WORD * 12 + PAGE * 8, PAGE * 2));
-}
-
-TEST_CASE("BufferBase: Out of bounds region query") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 16);
-    REQUIRE(!buffer.IsRegionCpuModified(c - PAGE, PAGE));
-    REQUIRE(!buffer.IsRegionCpuModified(c - PAGE * 2, PAGE));
-    REQUIRE(!buffer.IsRegionCpuModified(c + WORD * 16, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + WORD * 16 - PAGE, WORD * 64));
-    REQUIRE(!buffer.IsRegionCpuModified(c + WORD * 16, WORD * 64));
-}
-
-TEST_CASE("BufferBase: Wrap word regions") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD * 2);
-    buffer.UnmarkRegionAsCpuModified(c, WORD * 2);
-    buffer.MarkRegionAsCpuModified(c + PAGE * 63, PAGE * 2);
-    REQUIRE(buffer.IsRegionCpuModified(c, WORD * 2));
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE * 62, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE * 63, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE * 64, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE * 63, PAGE * 2));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE * 63, PAGE * 8));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE * 60, PAGE * 8));
-
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE * 127, WORD * 16));
-    buffer.MarkRegionAsCpuModified(c + PAGE * 127, PAGE);
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE * 127, WORD * 16));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE * 127, PAGE));
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE * 126, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE * 126, PAGE * 2));
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE * 128, WORD * 16));
-}
-
-TEST_CASE("BufferBase: Unaligned page region query") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    buffer.MarkRegionAsCpuModified(c + 4000, 1000);
-    REQUIRE(buffer.IsRegionCpuModified(c, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + 4000, 1000));
-    REQUIRE(buffer.IsRegionCpuModified(c + 4000, 1));
-}
-
-TEST_CASE("BufferBase: Cached write") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    buffer.CachedCpuWrite(c + PAGE, PAGE);
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    buffer.FlushCachedWrites();
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    buffer.MarkRegionAsCpuModified(c, WORD);
-    REQUIRE(rasterizer.Count() == 0);
-}
-
-TEST_CASE("BufferBase: Multiple cached write") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    buffer.CachedCpuWrite(c + PAGE, PAGE);
-    buffer.CachedCpuWrite(c + PAGE * 3, PAGE);
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE * 3, PAGE));
-    buffer.FlushCachedWrites();
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE * 3, PAGE));
-    buffer.MarkRegionAsCpuModified(c, WORD);
-    REQUIRE(rasterizer.Count() == 0);
-}
-
-TEST_CASE("BufferBase: Cached write unmarked") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    buffer.CachedCpuWrite(c + PAGE, PAGE);
-    buffer.UnmarkRegionAsCpuModified(c + PAGE, PAGE);
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    buffer.FlushCachedWrites();
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    buffer.MarkRegionAsCpuModified(c, WORD);
-    REQUIRE(rasterizer.Count() == 0);
-}
-
-TEST_CASE("BufferBase: Cached write iterated") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    buffer.CachedCpuWrite(c + PAGE, PAGE);
-    int num = 0;
-    buffer.ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) { ++num; });
-    REQUIRE(num == 0);
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    buffer.FlushCachedWrites();
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    buffer.MarkRegionAsCpuModified(c, WORD);
-    REQUIRE(rasterizer.Count() == 0);
-}
-
-TEST_CASE("BufferBase: Cached write downloads") {
-    RasterizerInterface rasterizer;
-    BufferBase buffer(rasterizer, c, WORD);
-    buffer.UnmarkRegionAsCpuModified(c, WORD);
-    REQUIRE(rasterizer.Count() == 64);
-    buffer.CachedCpuWrite(c + PAGE, PAGE);
-    REQUIRE(rasterizer.Count() == 63);
-    buffer.MarkRegionAsGpuModified(c + PAGE, PAGE);
-    int num = 0;
-    buffer.ForEachDownloadRangeAndClear(c, WORD, [&](u64 offset, u64 size) { ++num; });
-    buffer.ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) { ++num; });
-    REQUIRE(num == 0);
-    REQUIRE(!buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    REQUIRE(!buffer.IsRegionGpuModified(c + PAGE, PAGE));
-    buffer.FlushCachedWrites();
-    REQUIRE(buffer.IsRegionCpuModified(c + PAGE, PAGE));
-    REQUIRE(!buffer.IsRegionGpuModified(c + PAGE, PAGE));
-    buffer.MarkRegionAsCpuModified(c, WORD);
-    REQUIRE(rasterizer.Count() == 0);
-}

src/tests/video_core/memory_tracker.cpp

@@ -0,0 +1,549 @@
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#include <memory>
+#include <stdexcept>
+#include <unordered_map>
+
+#include <catch2/catch_test_macros.hpp>
+
+#include "common/alignment.h"
+#include "common/common_types.h"
+#include "video_core/buffer_cache/memory_tracker_base.h"
+
+namespace {
+using Range = std::pair<u64, u64>;
+
+constexpr u64 PAGE = 4096;
+constexpr u64 WORD = 4096 * 64;
+constexpr u64 HIGH_PAGE_BITS = 22;
+constexpr u64 HIGH_PAGE_SIZE = 1ULL << HIGH_PAGE_BITS;
+
+constexpr VAddr c = 16 * HIGH_PAGE_SIZE;
+
+class RasterizerInterface {
+public:
+    void UpdatePagesCachedCount(VAddr addr, u64 size, int delta) {
+        const u64 page_start{addr >> Core::Memory::YUZU_PAGEBITS};
+        const u64 page_end{(addr + size + Core::Memory::YUZU_PAGESIZE - 1) >>
+                           Core::Memory::YUZU_PAGEBITS};
+        for (u64 page = page_start; page < page_end; ++page) {
+            int& value = page_table[page];
+            value += delta;
+            if (value < 0) {
+                throw std::logic_error{"negative page"};
+            }
+            if (value == 0) {
+                page_table.erase(page);
+            }
+        }
+    }
+
+    [[nodiscard]] int Count(VAddr addr) const noexcept {
+        const auto it = page_table.find(addr >> Core::Memory::YUZU_PAGEBITS);
+        return it == page_table.end() ? 0 : it->second;
+    }
+
+    [[nodiscard]] unsigned Count() const noexcept {
+        unsigned count = 0;
+        for (const auto& [index, value] : page_table) {
+            count += value;
+        }
+        return count;
+    }
+
+private:
+    std::unordered_map<u64, int> page_table;
+};
+} // Anonymous namespace
+
+using MemoryTracker = VideoCommon::MemoryTrackerBase<RasterizerInterface>;
+
+TEST_CASE("MemoryTracker: Small region", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    REQUIRE(rasterizer.Count() == 0);
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    REQUIRE(rasterizer.Count() == WORD / PAGE);
+    REQUIRE(memory_track->ModifiedCpuRegion(c, WORD) == Range{0, 0});
+
+    memory_track->MarkRegionAsCpuModified(c + PAGE, 1);
+    REQUIRE(memory_track->ModifiedCpuRegion(c, WORD) == Range{c + PAGE * 1, c + PAGE * 2});
+}
+
+TEST_CASE("MemoryTracker: Large region", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 32);
+    memory_track->MarkRegionAsCpuModified(c + 4096, WORD * 4);
+    REQUIRE(memory_track->ModifiedCpuRegion(c, WORD + PAGE * 2) ==
+            Range{c + PAGE, c + WORD + PAGE * 2});
+    REQUIRE(memory_track->ModifiedCpuRegion(c + PAGE * 2, PAGE * 6) ==
+            Range{c + PAGE * 2, c + PAGE * 8});
+    REQUIRE(memory_track->ModifiedCpuRegion(c, WORD * 32) == Range{c + PAGE, c + WORD * 4 + PAGE});
+    REQUIRE(memory_track->ModifiedCpuRegion(c + WORD * 4, PAGE) ==
+            Range{c + WORD * 4, c + WORD * 4 + PAGE});
+    REQUIRE(memory_track->ModifiedCpuRegion(c + WORD * 3 + PAGE * 63, PAGE) ==
+            Range{c + WORD * 3 + PAGE * 63, c + WORD * 4});
+
+    memory_track->MarkRegionAsCpuModified(c + WORD * 5 + PAGE * 6, PAGE);
+    memory_track->MarkRegionAsCpuModified(c + WORD * 5 + PAGE * 8, PAGE);
+    REQUIRE(memory_track->ModifiedCpuRegion(c + WORD * 5, WORD) ==
+            Range{c + WORD * 5 + PAGE * 6, c + WORD * 5 + PAGE * 9});
+
+    memory_track->UnmarkRegionAsCpuModified(c + WORD * 5 + PAGE * 8, PAGE);
+    REQUIRE(memory_track->ModifiedCpuRegion(c + WORD * 5, WORD) ==
+            Range{c + WORD * 5 + PAGE * 6, c + WORD * 5 + PAGE * 7});
+
+    memory_track->MarkRegionAsCpuModified(c + PAGE, WORD * 31 + PAGE * 63);
+    REQUIRE(memory_track->ModifiedCpuRegion(c, WORD * 32) == Range{c + PAGE, c + WORD * 32});
+
+    memory_track->UnmarkRegionAsCpuModified(c + PAGE * 4, PAGE);
+    memory_track->UnmarkRegionAsCpuModified(c + PAGE * 6, PAGE);
+
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 32);
+    REQUIRE(memory_track->ModifiedCpuRegion(c, WORD * 32) == Range{0, 0});
+}
+
+TEST_CASE("MemoryTracker: Rasterizer counting", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    REQUIRE(rasterizer.Count() == 0);
+    memory_track->UnmarkRegionAsCpuModified(c, PAGE);
+    REQUIRE(rasterizer.Count() == 1);
+    memory_track->MarkRegionAsCpuModified(c, PAGE * 2);
+    REQUIRE(rasterizer.Count() == 0);
+    memory_track->UnmarkRegionAsCpuModified(c, PAGE);
+    memory_track->UnmarkRegionAsCpuModified(c + PAGE, PAGE);
+    REQUIRE(rasterizer.Count() == 2);
+    memory_track->MarkRegionAsCpuModified(c, PAGE * 2);
+    REQUIRE(rasterizer.Count() == 0);
+}
+
+TEST_CASE("MemoryTracker: Basic range", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    memory_track->MarkRegionAsCpuModified(c, PAGE);
+    int num = 0;
+    memory_track->ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c);
+        REQUIRE(size == PAGE);
+        ++num;
+    });
+    REQUIRE(num == 1U);
+}
+
+TEST_CASE("MemoryTracker: Border upload", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 2);
+    memory_track->MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
+    memory_track->ForEachUploadRange(c, WORD * 2, [](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD - PAGE);
+        REQUIRE(size == PAGE * 2);
+    });
+}
+
+TEST_CASE("MemoryTracker: Border upload range", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 2);
+    memory_track->MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
+    memory_track->ForEachUploadRange(c + WORD - PAGE, PAGE * 2, [](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD - PAGE);
+        REQUIRE(size == PAGE * 2);
+    });
+    memory_track->MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
+    memory_track->ForEachUploadRange(c + WORD - PAGE, PAGE, [](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD - PAGE);
+        REQUIRE(size == PAGE);
+    });
+    memory_track->ForEachUploadRange(c + WORD, PAGE, [](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD);
+        REQUIRE(size == PAGE);
+    });
+}
+
+TEST_CASE("MemoryTracker: Border upload partial range", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 2);
+    memory_track->MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
+    memory_track->ForEachUploadRange(c + WORD - 1, 2, [](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD - PAGE);
+        REQUIRE(size == PAGE * 2);
+    });
+    memory_track->MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
+    memory_track->ForEachUploadRange(c + WORD - 1, 1, [](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD - PAGE);
+        REQUIRE(size == PAGE);
+    });
+    memory_track->ForEachUploadRange(c + WORD + 50, 1, [](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD);
+        REQUIRE(size == PAGE);
+    });
+}
+
+TEST_CASE("MemoryTracker: Partial word uploads", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    int num = 0;
+    memory_track->ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c);
+        REQUIRE(size == WORD);
+        ++num;
+    });
+    REQUIRE(num == 1);
+    memory_track->ForEachUploadRange(c + WORD, WORD, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD);
+        REQUIRE(size == WORD);
+        ++num;
+    });
+    REQUIRE(num == 2);
+    memory_track->ForEachUploadRange(c + 0x79000, 0x24000, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD * 2);
+        REQUIRE(size == PAGE * 0x1d);
+        ++num;
+    });
+    REQUIRE(num == 3);
+}
+
+TEST_CASE("MemoryTracker: Partial page upload", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    int num = 0;
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 2, PAGE);
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 9, PAGE);
+    memory_track->ForEachUploadRange(c, PAGE * 3, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + PAGE * 2);
+        REQUIRE(size == PAGE);
+        ++num;
+    });
+    REQUIRE(num == 1);
+    memory_track->ForEachUploadRange(c + PAGE * 7, PAGE * 3, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + PAGE * 9);
+        REQUIRE(size == PAGE);
+        ++num;
+    });
+    REQUIRE(num == 2);
+}
+
+TEST_CASE("MemoryTracker: Partial page upload with multiple words on the right") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 9);
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 13, WORD * 7);
+    int num = 0;
+    memory_track->ForEachUploadRange(c + PAGE * 10, WORD * 7, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + PAGE * 13);
+        REQUIRE(size == WORD * 7 - PAGE * 3);
+        ++num;
+    });
+    REQUIRE(num == 1);
+    memory_track->ForEachUploadRange(c + PAGE, WORD * 8, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD * 7 + PAGE * 10);
+        REQUIRE(size == PAGE * 3);
+        ++num;
+    });
+    REQUIRE(num == 2);
+}
+
+TEST_CASE("MemoryTracker: Partial page upload with multiple words on the left", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 8);
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 13, WORD * 7);
+    int num = 0;
+    memory_track->ForEachUploadRange(c + PAGE * 16, WORD * 7, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + PAGE * 16);
+        REQUIRE(size == WORD * 7 - PAGE * 3);
+        ++num;
+    });
+    REQUIRE(num == 1);
+    memory_track->ForEachUploadRange(c + PAGE, WORD, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + PAGE * 13);
+        REQUIRE(size == PAGE * 3);
+        ++num;
+    });
+    REQUIRE(num == 2);
+}
+
+TEST_CASE("MemoryTracker: Partial page upload with multiple words in the middle", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 8);
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 13, PAGE * 140);
+    int num = 0;
+    memory_track->ForEachUploadRange(c + PAGE * 16, WORD, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + PAGE * 16);
+        REQUIRE(size == WORD);
+        ++num;
+    });
+    REQUIRE(num == 1);
+    memory_track->ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + PAGE * 13);
+        REQUIRE(size == PAGE * 3);
+        ++num;
+    });
+    REQUIRE(num == 2);
+    memory_track->ForEachUploadRange(c, WORD * 8, [&](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD + PAGE * 16);
+        REQUIRE(size == PAGE * 73);
+        ++num;
+    });
+    REQUIRE(num == 3);
+}
+
+TEST_CASE("MemoryTracker: Empty right bits", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 2048);
+    memory_track->MarkRegionAsCpuModified(c + WORD - PAGE, PAGE * 2);
+    memory_track->ForEachUploadRange(c, WORD * 2048, [](u64 offset, u64 size) {
+        REQUIRE(offset == c + WORD - PAGE);
+        REQUIRE(size == PAGE * 2);
+    });
+}
+
+TEST_CASE("MemoryTracker: Out of bound ranges 1", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c - WORD, 3 * WORD);
+    memory_track->MarkRegionAsCpuModified(c, PAGE);
+    REQUIRE(rasterizer.Count() == (3 * WORD - PAGE) / PAGE);
+    int num = 0;
+    memory_track->ForEachUploadRange(c - WORD, WORD, [&](u64 offset, u64 size) { ++num; });
+    memory_track->ForEachUploadRange(c + WORD, WORD, [&](u64 offset, u64 size) { ++num; });
+    memory_track->ForEachUploadRange(c - PAGE, PAGE, [&](u64 offset, u64 size) { ++num; });
+    REQUIRE(num == 0);
+    memory_track->ForEachUploadRange(c - PAGE, PAGE * 2, [&](u64 offset, u64 size) { ++num; });
+    REQUIRE(num == 1);
+    memory_track->MarkRegionAsCpuModified(c, WORD);
+    REQUIRE(rasterizer.Count() == 2 * WORD / PAGE);
+}
+
+TEST_CASE("MemoryTracker: Out of bound ranges 2", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    REQUIRE_NOTHROW(memory_track->UnmarkRegionAsCpuModified(c + 0x22000, PAGE));
+    REQUIRE_NOTHROW(memory_track->UnmarkRegionAsCpuModified(c + 0x28000, PAGE));
+    REQUIRE(rasterizer.Count() == 2);
+    REQUIRE_NOTHROW(memory_track->UnmarkRegionAsCpuModified(c + 0x21100, PAGE - 0x100));
+    REQUIRE(rasterizer.Count() == 3);
+    REQUIRE_NOTHROW(memory_track->UnmarkRegionAsCpuModified(c - PAGE, PAGE * 2));
+    memory_track->UnmarkRegionAsCpuModified(c - PAGE * 3, PAGE * 2);
+    memory_track->UnmarkRegionAsCpuModified(c - PAGE * 2, PAGE * 2);
+    REQUIRE(rasterizer.Count() == 7);
+}
+
+TEST_CASE("MemoryTracker: Out of bound ranges 3", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, 0x310720);
+    REQUIRE(rasterizer.Count(c) == 1);
+    REQUIRE(rasterizer.Count(c + PAGE) == 1);
+    REQUIRE(rasterizer.Count(c + WORD) == 1);
+    REQUIRE(rasterizer.Count(c + WORD + PAGE) == 1);
+}
+
+TEST_CASE("MemoryTracker: Sparse regions 1", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 1, PAGE);
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 3, PAGE * 4);
+    memory_track->ForEachUploadRange(c, WORD, [i = 0](u64 offset, u64 size) mutable {
+        static constexpr std::array<u64, 2> offsets{c + PAGE, c + PAGE * 3};
+        static constexpr std::array<u64, 2> sizes{PAGE, PAGE * 4};
+        REQUIRE(offset == offsets.at(i));
+        REQUIRE(size == sizes.at(i));
+        ++i;
+    });
+}
+
+TEST_CASE("MemoryTracker: Sparse regions 2", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, PAGE * 0x23);
+    REQUIRE(rasterizer.Count() == 0x23);
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 0x1B, PAGE);
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 0x21, PAGE);
+    memory_track->ForEachUploadRange(c, PAGE * 0x23, [i = 0](u64 offset, u64 size) mutable {
+        static constexpr std::array<u64, 3> offsets{c + PAGE * 0x1B, c + PAGE * 0x21};
+        static constexpr std::array<u64, 3> sizes{PAGE, PAGE};
+        REQUIRE(offset == offsets.at(i));
+        REQUIRE(size == sizes.at(i));
+        ++i;
+    });
+}
+
+TEST_CASE("MemoryTracker: Single page modified range", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    REQUIRE(memory_track->IsRegionCpuModified(c, PAGE));
+    memory_track->UnmarkRegionAsCpuModified(c, PAGE);
+    REQUIRE(!memory_track->IsRegionCpuModified(c, PAGE));
+}
+
+TEST_CASE("MemoryTracker: Two page modified range", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    REQUIRE(memory_track->IsRegionCpuModified(c, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c, PAGE * 2));
+    memory_track->UnmarkRegionAsCpuModified(c, PAGE);
+    REQUIRE(!memory_track->IsRegionCpuModified(c, PAGE));
+}
+
+TEST_CASE("MemoryTracker: Multi word modified ranges", "[video_core]") {
+    for (int offset = 0; offset < 4; ++offset) {
+        const VAddr address = c + WORD * offset;
+        RasterizerInterface rasterizer;
+        std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+        REQUIRE(memory_track->IsRegionCpuModified(address, PAGE));
+        REQUIRE(memory_track->IsRegionCpuModified(address + PAGE * 48, PAGE));
+        REQUIRE(memory_track->IsRegionCpuModified(address + PAGE * 56, PAGE));
+
+        memory_track->UnmarkRegionAsCpuModified(address + PAGE * 32, PAGE);
+        REQUIRE(memory_track->IsRegionCpuModified(address + PAGE, WORD));
+        REQUIRE(memory_track->IsRegionCpuModified(address + PAGE * 31, PAGE));
+        REQUIRE(!memory_track->IsRegionCpuModified(address + PAGE * 32, PAGE));
+        REQUIRE(memory_track->IsRegionCpuModified(address + PAGE * 33, PAGE));
+        REQUIRE(memory_track->IsRegionCpuModified(address + PAGE * 31, PAGE * 2));
+        REQUIRE(memory_track->IsRegionCpuModified(address + PAGE * 32, PAGE * 2));
+
+        memory_track->UnmarkRegionAsCpuModified(address + PAGE * 33, PAGE);
+        REQUIRE(!memory_track->IsRegionCpuModified(address + PAGE * 32, PAGE * 2));
+    }
+}
+
+TEST_CASE("MemoryTracker: Single page in large region", "[video_core]") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 16);
+    REQUIRE(!memory_track->IsRegionCpuModified(c, WORD * 16));
+
+    memory_track->MarkRegionAsCpuModified(c + WORD * 12 + PAGE * 8, PAGE);
+    REQUIRE(memory_track->IsRegionCpuModified(c, WORD * 16));
+    REQUIRE(!memory_track->IsRegionCpuModified(c + WORD * 10, WORD * 2));
+    REQUIRE(memory_track->IsRegionCpuModified(c + WORD * 11, WORD * 2));
+    REQUIRE(memory_track->IsRegionCpuModified(c + WORD * 12, WORD * 2));
+    REQUIRE(memory_track->IsRegionCpuModified(c + WORD * 12 + PAGE * 4, PAGE * 8));
+    REQUIRE(memory_track->IsRegionCpuModified(c + WORD * 12 + PAGE * 6, PAGE * 8));
+    REQUIRE(!memory_track->IsRegionCpuModified(c + WORD * 12 + PAGE * 6, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c + WORD * 12 + PAGE * 7, PAGE * 2));
+    REQUIRE(memory_track->IsRegionCpuModified(c + WORD * 12 + PAGE * 8, PAGE * 2));
+}
+
+TEST_CASE("MemoryTracker: Wrap word regions") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD * 32);
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 63, PAGE * 2);
+    REQUIRE(memory_track->IsRegionCpuModified(c, WORD * 2));
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE * 62, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE * 63, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE * 64, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE * 63, PAGE * 2));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE * 63, PAGE * 8));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE * 60, PAGE * 8));
+
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE * 127, WORD * 16));
+    memory_track->MarkRegionAsCpuModified(c + PAGE * 127, PAGE);
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE * 127, WORD * 16));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE * 127, PAGE));
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE * 126, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE * 126, PAGE * 2));
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE * 128, WORD * 16));
+}
+
+TEST_CASE("MemoryTracker: Unaligned page region query") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    memory_track->MarkRegionAsCpuModified(c + 4000, 1000);
+    REQUIRE(memory_track->IsRegionCpuModified(c, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c + 4000, 1000));
+    REQUIRE(memory_track->IsRegionCpuModified(c + 4000, 1));
+}
+
+TEST_CASE("MemoryTracker: Cached write") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    memory_track->CachedCpuWrite(c + PAGE, c + PAGE);
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    memory_track->FlushCachedWrites();
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    memory_track->MarkRegionAsCpuModified(c, WORD);
+    REQUIRE(rasterizer.Count() == 0);
+}
+
+TEST_CASE("MemoryTracker: Multiple cached write") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    memory_track->CachedCpuWrite(c + PAGE, PAGE);
+    memory_track->CachedCpuWrite(c + PAGE * 3, PAGE);
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE * 3, PAGE));
+    memory_track->FlushCachedWrites();
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE * 3, PAGE));
+    memory_track->MarkRegionAsCpuModified(c, WORD);
+    REQUIRE(rasterizer.Count() == 0);
+}
+
+TEST_CASE("MemoryTracker: Cached write unmarked") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    memory_track->CachedCpuWrite(c + PAGE, PAGE);
+    memory_track->UnmarkRegionAsCpuModified(c + PAGE, PAGE);
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    memory_track->FlushCachedWrites();
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    memory_track->MarkRegionAsCpuModified(c, WORD);
+    REQUIRE(rasterizer.Count() == 0);
+}
+
+TEST_CASE("MemoryTracker: Cached write iterated") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    memory_track->CachedCpuWrite(c + PAGE, PAGE);
+    int num = 0;
+    memory_track->ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) { ++num; });
+    REQUIRE(num == 0);
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    memory_track->FlushCachedWrites();
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    memory_track->MarkRegionAsCpuModified(c, WORD);
+    REQUIRE(rasterizer.Count() == 0);
+}
+
+TEST_CASE("MemoryTracker: Cached write downloads") {
+    RasterizerInterface rasterizer;
+    std::unique_ptr<MemoryTracker> memory_track(std::make_unique<MemoryTracker>(rasterizer));
+    memory_track->UnmarkRegionAsCpuModified(c, WORD);
+    REQUIRE(rasterizer.Count() == 64);
+    memory_track->CachedCpuWrite(c + PAGE, PAGE);
+    REQUIRE(rasterizer.Count() == 63);
+    memory_track->MarkRegionAsGpuModified(c + PAGE, PAGE);
+    int num = 0;
+    memory_track->ForEachDownloadRangeAndClear(c, WORD, [&](u64 offset, u64 size) { ++num; });
+    REQUIRE(num == 1);
+    num = 0;
+    memory_track->ForEachUploadRange(c, WORD, [&](u64 offset, u64 size) { ++num; });
+    REQUIRE(num == 0);
+    REQUIRE(!memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    REQUIRE(!memory_track->IsRegionGpuModified(c + PAGE, PAGE));
+    memory_track->FlushCachedWrites();
+    REQUIRE(memory_track->IsRegionCpuModified(c + PAGE, PAGE));
+    REQUIRE(!memory_track->IsRegionGpuModified(c + PAGE, PAGE));
+    memory_track->MarkRegionAsCpuModified(c, WORD);
+    REQUIRE(rasterizer.Count() == 0);
+}

src/video_core/CMakeLists.txt

@@ -11,8 +11,11 @@ endif()
 
 add_library(video_core STATIC
     buffer_cache/buffer_base.h
+    buffer_cache/buffer_cache_base.h
     buffer_cache/buffer_cache.cpp
     buffer_cache/buffer_cache.h
+    buffer_cache/memory_tracker_base.h
+    buffer_cache/word_manager.h
     cache_types.h
     cdma_pusher.cpp
     cdma_pusher.h
@@ -104,6 +107,7 @@ add_library(video_core STATIC
     renderer_null/renderer_null.h
     renderer_opengl/blit_image.cpp
     renderer_opengl/blit_image.h
+    renderer_opengl/gl_buffer_cache_base.cpp
     renderer_opengl/gl_buffer_cache.cpp
     renderer_opengl/gl_buffer_cache.h
     renderer_opengl/gl_compute_pipeline.cpp
@@ -154,6 +158,7 @@ add_library(video_core STATIC
     renderer_vulkan/renderer_vulkan.cpp
     renderer_vulkan/vk_blit_screen.cpp
     renderer_vulkan/vk_blit_screen.h
+    renderer_vulkan/vk_buffer_cache_base.cpp
     renderer_vulkan/vk_buffer_cache.cpp
     renderer_vulkan/vk_buffer_cache.h
     renderer_vulkan/vk_command_pool.cpp
@@ -174,6 +179,8 @@ add_library(video_core STATIC
     renderer_vulkan/vk_master_semaphore.h
     renderer_vulkan/vk_pipeline_cache.cpp
     renderer_vulkan/vk_pipeline_cache.h
+    renderer_vulkan/vk_present_manager.cpp
+    renderer_vulkan/vk_present_manager.h
     renderer_vulkan/vk_query_cache.cpp
     renderer_vulkan/vk_query_cache.h
     renderer_vulkan/vk_rasterizer.cpp

src/video_core/buffer_cache/buffer_base.h

@@ -1,5 +1,5 @@
-// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
-// SPDX-License-Identifier: GPL-2.0-or-later
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
 
 #pragma once
 
@@ -11,9 +11,7 @@
 #include "common/alignment.h"
 #include "common/common_funcs.h"
 #include "common/common_types.h"
-#include "common/div_ceil.h"
-#include "common/settings.h"
-#include "core/memory.h"
+#include "video_core/buffer_cache/word_manager.h"
 
 namespace VideoCommon {
 
@@ -36,116 +34,12 @@ struct NullBufferParams {};
  */
 template <class RasterizerInterface>
 class BufferBase {
-    static constexpr u64 PAGES_PER_WORD = 64;
-    static constexpr u64 BYTES_PER_PAGE = Core::Memory::YUZU_PAGESIZE;
-    static constexpr u64 BYTES_PER_WORD = PAGES_PER_WORD * BYTES_PER_PAGE;
-
-    /// Vector tracking modified pages tightly packed with small vector optimization
-    union WordsArray {
-        /// Returns the pointer to the words state
-        [[nodiscard]] const u64* Pointer(bool is_short) const noexcept {
-            return is_short ? &stack : heap;
-        }
-
-        /// Returns the pointer to the words state
-        [[nodiscard]] u64* Pointer(bool is_short) noexcept {
-            return is_short ? &stack : heap;
-        }
-
-        u64 stack = 0; ///< Small buffers storage
-        u64* heap;     ///< Not-small buffers pointer to the storage
-    };
-
-    struct Words {
-        explicit Words() = default;
-        explicit Words(u64 size_bytes_) : size_bytes{size_bytes_} {
-            if (IsShort()) {
-                cpu.stack = ~u64{0};
-                gpu.stack = 0;
-                cached_cpu.stack = 0;
-                untracked.stack = ~u64{0};
-            } else {
-                // Share allocation between CPU and GPU pages and set their default values
-                const size_t num_words = NumWords();
-                u64* const alloc = new u64[num_words * 4];
-                cpu.heap = alloc;
-                gpu.heap = alloc + num_words;
-                cached_cpu.heap = alloc + num_words * 2;
-                untracked.heap = alloc + num_words * 3;
-                std::fill_n(cpu.heap, num_words, ~u64{0});
-                std::fill_n(gpu.heap, num_words, 0);
-                std::fill_n(cached_cpu.heap, num_words, 0);
-                std::fill_n(untracked.heap, num_words, ~u64{0});
-            }
-            // Clean up tailing bits
-            const u64 last_word_size = size_bytes % BYTES_PER_WORD;
-            const u64 last_local_page = Common::DivCeil(last_word_size, BYTES_PER_PAGE);
-            const u64 shift = (PAGES_PER_WORD - last_local_page) % PAGES_PER_WORD;
-            const u64 last_word = (~u64{0} << shift) >> shift;
-            cpu.Pointer(IsShort())[NumWords() - 1] = last_word;
-            untracked.Pointer(IsShort())[NumWords() - 1] = last_word;
-        }
-
-        ~Words() {
-            Release();
-        }
-
-        Words& operator=(Words&& rhs) noexcept {
-            Release();
-            size_bytes = rhs.size_bytes;
-            cpu = rhs.cpu;
-            gpu = rhs.gpu;
-            cached_cpu = rhs.cached_cpu;
-            untracked = rhs.untracked;
-            rhs.cpu.heap = nullptr;
-            return *this;
-        }
-
-        Words(Words&& rhs) noexcept
-            : size_bytes{rhs.size_bytes}, cpu{rhs.cpu}, gpu{rhs.gpu},
-              cached_cpu{rhs.cached_cpu}, untracked{rhs.untracked} {
-            rhs.cpu.heap = nullptr;
-        }
-
-        Words& operator=(const Words&) = delete;
-        Words(const Words&) = delete;
-
-        /// Returns true when the buffer fits in the small vector optimization
-        [[nodiscard]] bool IsShort() const noexcept {
-            return size_bytes <= BYTES_PER_WORD;
-        }
-
-        /// Returns the number of words of the buffer
-        [[nodiscard]] size_t NumWords() const noexcept {
-            return Common::DivCeil(size_bytes, BYTES_PER_WORD);
-        }
-
-        /// Release buffer resources
-        void Release() {
-            if (!IsShort()) {
-                // CPU written words is the base for the heap allocation
-                delete[] cpu.heap;
-            }
-        }
-
-        u64 size_bytes = 0;
-        WordsArray cpu;
-        WordsArray gpu;
-        WordsArray cached_cpu;
-        WordsArray untracked;
-    };
-
-    enum class Type {
-        CPU,
-        GPU,
-        CachedCPU,
-        Untracked,
-    };
-
 public:
-    explicit BufferBase(RasterizerInterface& rasterizer_, VAddr cpu_addr_, u64 size_bytes)
-        : rasterizer{&rasterizer_}, cpu_addr{Common::AlignDown(cpu_addr_, BYTES_PER_PAGE)},
-          words(Common::AlignUp(size_bytes + (cpu_addr_ - cpu_addr), BYTES_PER_PAGE)) {}
+    static constexpr u64 BASE_PAGE_BITS = 16;
+    static constexpr u64 BASE_PAGE_SIZE = 1ULL << BASE_PAGE_BITS;
+
+    explicit BufferBase(RasterizerInterface& rasterizer_, VAddr cpu_addr_, u64 size_bytes_)
+        : cpu_addr{cpu_addr_}, size_bytes{size_bytes_} {}
 
     explicit BufferBase(NullBufferParams) {}
 
@@ -155,100 +49,6 @@ public:
     BufferBase& operator=(BufferBase&&) = default;
     BufferBase(BufferBase&&) = default;
 
-    /// Returns the inclusive CPU modified range in a begin end pair
-    [[nodiscard]] std::pair<u64, u64> ModifiedCpuRegion(VAddr query_cpu_addr,
-                                                        u64 query_size) const noexcept {
-        const u64 offset = query_cpu_addr - cpu_addr;
-        return ModifiedRegion<Type::CPU>(offset, query_size);
-    }
-
-    /// Returns the inclusive GPU modified range in a begin end pair
-    [[nodiscard]] std::pair<u64, u64> ModifiedGpuRegion(VAddr query_cpu_addr,
-                                                        u64 query_size) const noexcept {
-        const u64 offset = query_cpu_addr - cpu_addr;
-        return ModifiedRegion<Type::GPU>(offset, query_size);
-    }
-
-    /// Returns true if a region has been modified from the CPU
-    [[nodiscard]] bool IsRegionCpuModified(VAddr query_cpu_addr, u64 query_size) const noexcept {
-        const u64 offset = query_cpu_addr - cpu_addr;
-        return IsRegionModified<Type::CPU>(offset, query_size);
-    }
-
-    /// Returns true if a region has been modified from the GPU
-    [[nodiscard]] bool IsRegionGpuModified(VAddr query_cpu_addr, u64 query_size) const noexcept {
-        const u64 offset = query_cpu_addr - cpu_addr;
-        return IsRegionModified<Type::GPU>(offset, query_size);
-    }
-
-    /// Mark region as CPU modified, notifying the rasterizer about this change
-    void MarkRegionAsCpuModified(VAddr dirty_cpu_addr, u64 size) {
-        ChangeRegionState<Type::CPU, true>(dirty_cpu_addr, size);
-    }
-
-    /// Unmark region as CPU modified, notifying the rasterizer about this change
-    void UnmarkRegionAsCpuModified(VAddr dirty_cpu_addr, u64 size) {
-        ChangeRegionState<Type::CPU, false>(dirty_cpu_addr, size);
-    }
-
-    /// Mark region as modified from the host GPU
-    void MarkRegionAsGpuModified(VAddr dirty_cpu_addr, u64 size) noexcept {
-        ChangeRegionState<Type::GPU, true>(dirty_cpu_addr, size);
-    }
-
-    /// Unmark region as modified from the host GPU
-    void UnmarkRegionAsGpuModified(VAddr dirty_cpu_addr, u64 size) noexcept {
-        ChangeRegionState<Type::GPU, false>(dirty_cpu_addr, size);
-    }
-
-    /// Mark region as modified from the CPU
-    /// but don't mark it as modified until FlusHCachedWrites is called.
-    void CachedCpuWrite(VAddr dirty_cpu_addr, u64 size) {
-        flags |= BufferFlagBits::CachedWrites;
-        ChangeRegionState<Type::CachedCPU, true>(dirty_cpu_addr, size);
-    }
-
-    /// Flushes cached CPU writes, and notify the rasterizer about the deltas
-    void FlushCachedWrites() noexcept {
-        flags &= ~BufferFlagBits::CachedWrites;
-        const u64 num_words = NumWords();
-        u64* const cached_words = Array<Type::CachedCPU>();
-        u64* const untracked_words = Array<Type::Untracked>();
-        u64* const cpu_words = Array<Type::CPU>();
-        for (u64 word_index = 0; word_index < num_words; ++word_index) {
-            const u64 cached_bits = cached_words[word_index];
-            NotifyRasterizer<false>(word_index, untracked_words[word_index], cached_bits);
-            untracked_words[word_index] |= cached_bits;
-            cpu_words[word_index] |= cached_bits;
-            if (!Settings::values.use_pessimistic_flushes) {
-                cached_words[word_index] = 0;
-            }
-        }
-    }
-
-    /// Call 'func' for each CPU modified range and unmark those pages as CPU modified
-    template <typename Func>
-    void ForEachUploadRange(VAddr query_cpu_range, u64 size, Func&& func) {
-        ForEachModifiedRange<Type::CPU>(query_cpu_range, size, true, func);
-    }
-
-    /// Call 'func' for each GPU modified range and unmark those pages as GPU modified
-    template <typename Func>
-    void ForEachDownloadRange(VAddr query_cpu_range, u64 size, bool clear, Func&& func) {
-        ForEachModifiedRange<Type::GPU>(query_cpu_range, size, clear, func);
-    }
-
-    template <typename Func>
-    void ForEachDownloadRangeAndClear(VAddr query_cpu_range, u64 size, Func&& func) {
-        ForEachModifiedRange<Type::GPU>(query_cpu_range, size, true, func);
-    }
-
-    /// Call 'func' for each GPU modified range and unmark those pages as GPU modified
-    template <typename Func>
-    void ForEachDownloadRange(Func&& func) {
-        ForEachModifiedRange<Type::GPU>(cpu_addr, SizeBytes(), true, func);
-    }
-
     /// Mark buffer as picked
     void Pick() noexcept {
         flags |= BufferFlagBits::Picked;
@@ -295,11 +95,6 @@ public:
         return static_cast<u32>(other_cpu_addr - cpu_addr);
     }
 
-    /// Returns the size in bytes of the buffer
-    [[nodiscard]] u64 SizeBytes() const noexcept {
-        return words.size_bytes;
-    }
-
     size_t getLRUID() const noexcept {
         return lru_id;
     }
@@ -308,305 +103,16 @@ public:
         lru_id = lru_id_;
     }
 
+    size_t SizeBytes() const {
+        return size_bytes;
+    }
+
 private:
-    template <Type type>
-    u64* Array() noexcept {
-        if constexpr (type == Type::CPU) {
-            return words.cpu.Pointer(IsShort());
-        } else if constexpr (type == Type::GPU) {
-            return words.gpu.Pointer(IsShort());
-        } else if constexpr (type == Type::CachedCPU) {
-            return words.cached_cpu.Pointer(IsShort());
-        } else if constexpr (type == Type::Untracked) {
-            return words.untracked.Pointer(IsShort());
-        }
-    }
-
-    template <Type type>
-    const u64* Array() const noexcept {
-        if constexpr (type == Type::CPU) {
-            return words.cpu.Pointer(IsShort());
-        } else if constexpr (type == Type::GPU) {
-            return words.gpu.Pointer(IsShort());
-        } else if constexpr (type == Type::CachedCPU) {
-            return words.cached_cpu.Pointer(IsShort());
-        } else if constexpr (type == Type::Untracked) {
-            return words.untracked.Pointer(IsShort());
-        }
-    }
-
-    /**
-     * Change the state of a range of pages
-     *
-     * @param dirty_addr    Base address to mark or unmark as modified
-     * @param size          Size in bytes to mark or unmark as modified
-     */
-    template <Type type, bool enable>
-    void ChangeRegionState(u64 dirty_addr, s64 size) noexcept(type == Type::GPU) {
-        const s64 difference = dirty_addr - cpu_addr;
-        const u64 offset = std::max<s64>(difference, 0);
-        size += std::min<s64>(difference, 0);
-        if (offset >= SizeBytes() || size < 0) {
-            return;
-        }
-        u64* const untracked_words = Array<Type::Untracked>();
-        u64* const state_words = Array<type>();
-        const u64 offset_end = std::min(offset + size, SizeBytes());
-        const u64 begin_page_index = offset / BYTES_PER_PAGE;
-        const u64 begin_word_index = begin_page_index / PAGES_PER_WORD;
-        const u64 end_page_index = Common::DivCeil(offset_end, BYTES_PER_PAGE);
-        const u64 end_word_index = Common::DivCeil(end_page_index, PAGES_PER_WORD);
-        u64 page_index = begin_page_index % PAGES_PER_WORD;
-        u64 word_index = begin_word_index;
-        while (word_index < end_word_index) {
-            const u64 next_word_first_page = (word_index + 1) * PAGES_PER_WORD;
-            const u64 left_offset =
-                std::min(next_word_first_page - end_page_index, PAGES_PER_WORD) % PAGES_PER_WORD;
-            const u64 right_offset = page_index;
-            u64 bits = ~u64{0};
-            bits = (bits >> right_offset) << right_offset;
-            bits = (bits << left_offset) >> left_offset;
-            if constexpr (type == Type::CPU || type == Type::CachedCPU) {
-                NotifyRasterizer<!enable>(word_index, untracked_words[word_index], bits);
-            }
-            if constexpr (enable) {
-                state_words[word_index] |= bits;
-                if constexpr (type == Type::CPU || type == Type::CachedCPU) {
-                    untracked_words[word_index] |= bits;
-                }
-            } else {
-                state_words[word_index] &= ~bits;
-                if constexpr (type == Type::CPU || type == Type::CachedCPU) {
-                    untracked_words[word_index] &= ~bits;
-                }
-            }
-            page_index = 0;
-            ++word_index;
-        }
-    }
-
-    /**
-     * Notify rasterizer about changes in the CPU tracking state of a word in the buffer
-     *
-     * @param word_index   Index to the word to notify to the rasterizer
-     * @param current_bits Current state of the word
-     * @param new_bits     New state of the word
-     *
-     * @tparam add_to_rasterizer True when the rasterizer should start tracking the new pages
-     */
-    template <bool add_to_rasterizer>
-    void NotifyRasterizer(u64 word_index, u64 current_bits, u64 new_bits) const {
-        u64 changed_bits = (add_to_rasterizer ? current_bits : ~current_bits) & new_bits;
-        VAddr addr = cpu_addr + word_index * BYTES_PER_WORD;
-        while (changed_bits != 0) {
-            const int empty_bits = std::countr_zero(changed_bits);
-            addr += empty_bits * BYTES_PER_PAGE;
-            changed_bits >>= empty_bits;
-
-            const u32 continuous_bits = std::countr_one(changed_bits);
-            const u64 size = continuous_bits * BYTES_PER_PAGE;
-            const VAddr begin_addr = addr;
-            addr += size;
-            changed_bits = continuous_bits < PAGES_PER_WORD ? (changed_bits >> continuous_bits) : 0;
-            rasterizer->UpdatePagesCachedCount(begin_addr, size, add_to_rasterizer ? 1 : -1);
-        }
-    }
-
-    /**
-     * Loop over each page in the given range, turn off those bits and notify the rasterizer if
-     * needed. Call the given function on each turned off range.
-     *
-     * @param query_cpu_range Base CPU address to loop over
-     * @param size            Size in bytes of the CPU range to loop over
-     * @param func            Function to call for each turned off region
-     */
-    template <Type type, typename Func>
-    void ForEachModifiedRange(VAddr query_cpu_range, s64 size, bool clear, Func&& func) {
-        static_assert(type != Type::Untracked);
-
-        const s64 difference = query_cpu_range - cpu_addr;
-        const u64 query_begin = std::max<s64>(difference, 0);
-        size += std::min<s64>(difference, 0);
-        if (query_begin >= SizeBytes() || size < 0) {
-            return;
-        }
-        u64* const untracked_words = Array<Type::Untracked>();
-        u64* const state_words = Array<type>();
-        const u64 query_end = query_begin + std::min(static_cast<u64>(size), SizeBytes());
-        u64* const words_begin = state_words + query_begin / BYTES_PER_WORD;
-        u64* const words_end = state_words + Common::DivCeil(query_end, BYTES_PER_WORD);
-
-        const auto modified = [](u64 word) { return word != 0; };
-        const auto first_modified_word = std::find_if(words_begin, words_end, modified);
-        if (first_modified_word == words_end) {
-            // Exit early when the buffer is not modified
-            return;
-        }
-        const auto last_modified_word = std::find_if_not(first_modified_word, words_end, modified);
-
-        const u64 word_index_begin = std::distance(state_words, first_modified_word);
-        const u64 word_index_end = std::distance(state_words, last_modified_word);
-
-        const unsigned local_page_begin = std::countr_zero(*first_modified_word);
-        const unsigned local_page_end =
-            static_cast<unsigned>(PAGES_PER_WORD) - std::countl_zero(last_modified_word[-1]);
-        const u64 word_page_begin = word_index_begin * PAGES_PER_WORD;
-        const u64 word_page_end = (word_index_end - 1) * PAGES_PER_WORD;
-        const u64 query_page_begin = query_begin / BYTES_PER_PAGE;
-        const u64 query_page_end = Common::DivCeil(query_end, BYTES_PER_PAGE);
-        const u64 page_index_begin = std::max(word_page_begin + local_page_begin, query_page_begin);
-        const u64 page_index_end = std::min(word_page_end + local_page_end, query_page_end);
-        const u64 first_word_page_begin = page_index_begin % PAGES_PER_WORD;
-        const u64 last_word_page_end = (page_index_end - 1) % PAGES_PER_WORD + 1;
-
-        u64 page_begin = first_word_page_begin;
-        u64 current_base = 0;
-        u64 current_size = 0;
-        bool on_going = false;
-        for (u64 word_index = word_index_begin; word_index < word_index_end; ++word_index) {
-            const bool is_last_word = word_index + 1 == word_index_end;
-            const u64 page_end = is_last_word ? last_word_page_end : PAGES_PER_WORD;
-            const u64 right_offset = page_begin;
-            const u64 left_offset = PAGES_PER_WORD - page_end;
-            u64 bits = ~u64{0};
-            bits = (bits >> right_offset) << right_offset;
-            bits = (bits << left_offset) >> left_offset;
-
-            const u64 current_word = state_words[word_index] & bits;
-            if (clear) {
-                state_words[word_index] &= ~bits;
-            }
-
-            if constexpr (type == Type::CPU) {
-                const u64 current_bits = untracked_words[word_index] & bits;
-                untracked_words[word_index] &= ~bits;
-                NotifyRasterizer<true>(word_index, current_bits, ~u64{0});
-            }
-            // Exclude CPU modified pages when visiting GPU pages
-            const u64 word = current_word & ~(type == Type::GPU ? untracked_words[word_index] : 0);
-            u64 page = page_begin;
-            page_begin = 0;
-
-            while (page < page_end) {
-                const int empty_bits = std::countr_zero(word >> page);
-                if (on_going && empty_bits != 0) {
-                    InvokeModifiedRange(func, current_size, current_base);
-                    current_size = 0;
-                    on_going = false;
-                }
-                if (empty_bits == PAGES_PER_WORD) {
-                    break;
-                }
-                page += empty_bits;
-
-                const int continuous_bits = std::countr_one(word >> page);
-                if (!on_going && continuous_bits != 0) {
-                    current_base = word_index * PAGES_PER_WORD + page;
-                    on_going = true;
-                }
-                current_size += continuous_bits;
-                page += continuous_bits;
-            }
-        }
-        if (on_going && current_size > 0) {
-            InvokeModifiedRange(func, current_size, current_base);
-        }
-    }
-
-    template <typename Func>
-    void InvokeModifiedRange(Func&& func, u64 current_size, u64 current_base) {
-        const u64 current_size_bytes = current_size * BYTES_PER_PAGE;
-        const u64 offset_begin = current_base * BYTES_PER_PAGE;
-        const u64 offset_end = std::min(offset_begin + current_size_bytes, SizeBytes());
-        func(offset_begin, offset_end - offset_begin);
-    }
-
-    /**
-     * Returns true when a region has been modified
-     *
-     * @param offset Offset in bytes from the start of the buffer
-     * @param size   Size in bytes of the region to query for modifications
-     */
-    template <Type type>
-    [[nodiscard]] bool IsRegionModified(u64 offset, u64 size) const noexcept {
-        static_assert(type != Type::Untracked);
-
-        const u64* const untracked_words = Array<Type::Untracked>();
-        const u64* const state_words = Array<type>();
-        const u64 num_query_words = size / BYTES_PER_WORD + 1;
-        const u64 word_begin = offset / BYTES_PER_WORD;
-        const u64 word_end = std::min<u64>(word_begin + num_query_words, NumWords());
-        const u64 page_limit = Common::DivCeil(offset + size, BYTES_PER_PAGE);
-        u64 page_index = (offset / BYTES_PER_PAGE) % PAGES_PER_WORD;
-        for (u64 word_index = word_begin; word_index < word_end; ++word_index, page_index = 0) {
-            const u64 off_word = type == Type::GPU ? untracked_words[word_index] : 0;
-            const u64 word = state_words[word_index] & ~off_word;
-            if (word == 0) {
-                continue;
-            }
-            const u64 page_end = std::min((word_index + 1) * PAGES_PER_WORD, page_limit);
-            const u64 local_page_end = page_end % PAGES_PER_WORD;
-            const u64 page_end_shift = (PAGES_PER_WORD - local_page_end) % PAGES_PER_WORD;
-            if (((word >> page_index) << page_index) << page_end_shift != 0) {
-                return true;
-            }
-        }
-        return false;
-    }
-
-    /**
-     * Returns a begin end pair with the inclusive modified region
-     *
-     * @param offset Offset in bytes from the start of the buffer
-     * @param size   Size in bytes of the region to query for modifications
-     */
-    template <Type type>
-    [[nodiscard]] std::pair<u64, u64> ModifiedRegion(u64 offset, u64 size) const noexcept {
-        static_assert(type != Type::Untracked);
-
-        const u64* const untracked_words = Array<Type::Untracked>();
-        const u64* const state_words = Array<type>();
-        const u64 num_query_words = size / BYTES_PER_WORD + 1;
-        const u64 word_begin = offset / BYTES_PER_WORD;
-        const u64 word_end = std::min<u64>(word_begin + num_query_words, NumWords());
-        const u64 page_base = offset / BYTES_PER_PAGE;
-        const u64 page_limit = Common::DivCeil(offset + size, BYTES_PER_PAGE);
-        u64 begin = std::numeric_limits<u64>::max();
-        u64 end = 0;
-        for (u64 word_index = word_begin; word_index < word_end; ++word_index) {
-            const u64 off_word = type == Type::GPU ? untracked_words[word_index] : 0;
-            const u64 word = state_words[word_index] & ~off_word;
-            if (word == 0) {
-                continue;
-            }
-            const u64 local_page_begin = std::countr_zero(word);
-            const u64 local_page_end = PAGES_PER_WORD - std::countl_zero(word);
-            const u64 page_index = word_index * PAGES_PER_WORD;
-            const u64 page_begin = std::max(page_index + local_page_begin, page_base);
-            const u64 page_end = std::min(page_index + local_page_end, page_limit);
-            begin = std::min(begin, page_begin);
-            end = std::max(end, page_end);
-        }
-        static constexpr std::pair<u64, u64> EMPTY{0, 0};
-        return begin < end ? std::make_pair(begin * BYTES_PER_PAGE, end * BYTES_PER_PAGE) : EMPTY;
-    }
-
-    /// Returns the number of words of the buffer
-    [[nodiscard]] size_t NumWords() const noexcept {
-        return words.NumWords();
-    }
-
-    /// Returns true when the buffer fits in the small vector optimization
-    [[nodiscard]] bool IsShort() const noexcept {
-        return words.IsShort();
-    }
-
-    RasterizerInterface* rasterizer = nullptr;
     VAddr cpu_addr = 0;
-    Words words;
     BufferFlagBits flags{};
     int stream_score = 0;
     size_t lru_id = SIZE_MAX;
+    size_t size_bytes = 0;
 };
 
 } // namespace VideoCommon

src/video_core/buffer_cache/buffer_cache.cpp

@@ -1,5 +1,5 @@
-// SPDX-FileCopyrightText: Copyright 2021 yuzu Emulator Project
-// SPDX-License-Identifier: GPL-2.0-or-later
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
 
 #include "common/microprofile.h"
 

src/video_core/buffer_cache/buffer_cache_base.h

@@ -0,0 +1,580 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#pragma once
+
+#include <algorithm>
+#include <array>
+#include <functional>
+#include <memory>
+#include <mutex>
+#include <numeric>
+#include <span>
+#include <unordered_map>
+#include <vector>
+
+#include <boost/container/small_vector.hpp>
+#define BOOST_NO_MT
+#include <boost/pool/detail/mutex.hpp>
+#undef BOOST_NO_MT
+#include <boost/icl/interval.hpp>
+#include <boost/icl/interval_base_set.hpp>
+#include <boost/icl/interval_set.hpp>
+#include <boost/icl/split_interval_map.hpp>
+#include <boost/pool/pool.hpp>
+#include <boost/pool/pool_alloc.hpp>
+#include <boost/pool/poolfwd.hpp>
+
+#include "common/common_types.h"
+#include "common/div_ceil.h"
+#include "common/literals.h"
+#include "common/lru_cache.h"
+#include "common/microprofile.h"
+#include "common/scope_exit.h"
+#include "common/settings.h"
+#include "core/memory.h"
+#include "video_core/buffer_cache/buffer_base.h"
+#include "video_core/control/channel_state_cache.h"
+#include "video_core/delayed_destruction_ring.h"
+#include "video_core/dirty_flags.h"
+#include "video_core/engines/draw_manager.h"
+#include "video_core/engines/kepler_compute.h"
+#include "video_core/engines/maxwell_3d.h"
+#include "video_core/memory_manager.h"
+#include "video_core/rasterizer_interface.h"
+#include "video_core/surface.h"
+#include "video_core/texture_cache/slot_vector.h"
+#include "video_core/texture_cache/types.h"
+
+namespace boost {
+template <typename T>
+class fast_pool_allocator<T, default_user_allocator_new_delete, details::pool::null_mutex, 4096, 0>;
+}
+
+namespace VideoCommon {
+
+MICROPROFILE_DECLARE(GPU_PrepareBuffers);
+MICROPROFILE_DECLARE(GPU_BindUploadBuffers);
+MICROPROFILE_DECLARE(GPU_DownloadMemory);
+
+using BufferId = SlotId;
+
+using VideoCore::Surface::PixelFormat;
+using namespace Common::Literals;
+
+constexpr u32 NUM_VERTEX_BUFFERS = 32;
+constexpr u32 NUM_TRANSFORM_FEEDBACK_BUFFERS = 4;
+constexpr u32 NUM_GRAPHICS_UNIFORM_BUFFERS = 18;
+constexpr u32 NUM_COMPUTE_UNIFORM_BUFFERS = 8;
+constexpr u32 NUM_STORAGE_BUFFERS = 16;
+constexpr u32 NUM_TEXTURE_BUFFERS = 16;
+constexpr u32 NUM_STAGES = 5;
+
+using UniformBufferSizes = std::array<std::array<u32, NUM_GRAPHICS_UNIFORM_BUFFERS>, NUM_STAGES>;
+using ComputeUniformBufferSizes = std::array<u32, NUM_COMPUTE_UNIFORM_BUFFERS>;
+
+enum class ObtainBufferSynchronize : u32 {
+    NoSynchronize = 0,
+    FullSynchronize = 1,
+    SynchronizeNoDirty = 2,
+};
+
+enum class ObtainBufferOperation : u32 {
+    DoNothing = 0,
+    MarkAsWritten = 1,
+    DiscardWrite = 2,
+    MarkQuery = 3,
+};
+
+template <typename P>
+class BufferCache : public VideoCommon::ChannelSetupCaches<VideoCommon::ChannelInfo> {
+    // Page size for caching purposes.
+    // This is unrelated to the CPU page size and it can be changed as it seems optimal.
+    static constexpr u32 CACHING_PAGEBITS = 16;
+    static constexpr u64 CACHING_PAGESIZE = u64{1} << CACHING_PAGEBITS;
+
+    static constexpr bool IS_OPENGL = P::IS_OPENGL;
+    static constexpr bool HAS_PERSISTENT_UNIFORM_BUFFER_BINDINGS =
+        P::HAS_PERSISTENT_UNIFORM_BUFFER_BINDINGS;
+    static constexpr bool HAS_FULL_INDEX_AND_PRIMITIVE_SUPPORT =
+        P::HAS_FULL_INDEX_AND_PRIMITIVE_SUPPORT;
+    static constexpr bool NEEDS_BIND_UNIFORM_INDEX = P::NEEDS_BIND_UNIFORM_INDEX;
+    static constexpr bool NEEDS_BIND_STORAGE_INDEX = P::NEEDS_BIND_STORAGE_INDEX;
+    static constexpr bool USE_MEMORY_MAPS = P::USE_MEMORY_MAPS;
+    static constexpr bool SEPARATE_IMAGE_BUFFERS_BINDINGS = P::SEPARATE_IMAGE_BUFFER_BINDINGS;
+    static constexpr bool IMPLEMENTS_ASYNC_DOWNLOADS = P::IMPLEMENTS_ASYNC_DOWNLOADS;
+
+    static constexpr BufferId NULL_BUFFER_ID{0};
+
+    static constexpr s64 DEFAULT_EXPECTED_MEMORY = 512_MiB;
+    static constexpr s64 DEFAULT_CRITICAL_MEMORY = 1_GiB;
+    static constexpr s64 TARGET_THRESHOLD = 4_GiB;
+
+    // Debug Flags.
+
+    static constexpr bool DISABLE_DOWNLOADS = true;
+
+    using Maxwell = Tegra::Engines::Maxwell3D::Regs;
+
+    using Runtime = typename P::Runtime;
+    using Buffer = typename P::Buffer;
+    using Async_Buffer = typename P::Async_Buffer;
+    using MemoryTracker = typename P::MemoryTracker;
+
+    using IntervalCompare = std::less<VAddr>;
+    using IntervalInstance = boost::icl::interval_type_default<VAddr, std::less>;
+    using IntervalAllocator = boost::fast_pool_allocator<VAddr>;
+    using IntervalSet = boost::icl::interval_set<VAddr>;
+    using IntervalType = typename IntervalSet::interval_type;
+
+    template <typename Type>
+    struct counter_add_functor : public boost::icl::identity_based_inplace_combine<Type> {
+        // types
+        typedef counter_add_functor<Type> type;
+        typedef boost::icl::identity_based_inplace_combine<Type> base_type;
+
+        // public member functions
+        void operator()(Type& current, const Type& added) const {
+            current += added;
+            if (current < base_type::identity_element()) {
+                current = base_type::identity_element();
+            }
+        }
+
+        // public static functions
+        static void version(Type&){};
+    };
+
+    using OverlapCombine = counter_add_functor<int>;
+    using OverlapSection = boost::icl::inter_section<int>;
+    using OverlapCounter = boost::icl::split_interval_map<VAddr, int>;
+
+    struct Empty {};
+
+    struct OverlapResult {
+        std::vector<BufferId> ids;
+        VAddr begin;
+        VAddr end;
+        bool has_stream_leap = false;
+    };
+
+    struct Binding {
+        VAddr cpu_addr{};
+        u32 size{};
+        BufferId buffer_id;
+    };
+
+    struct TextureBufferBinding : Binding {
+        PixelFormat format;
+    };
+
+    static constexpr Binding NULL_BINDING{
+        .cpu_addr = 0,
+        .size = 0,
+        .buffer_id = NULL_BUFFER_ID,
+    };
+
+public:
+    static constexpr u32 DEFAULT_SKIP_CACHE_SIZE = static_cast<u32>(4_KiB);
+
+    explicit BufferCache(VideoCore::RasterizerInterface& rasterizer_,
+                         Core::Memory::Memory& cpu_memory_, Runtime& runtime_);
+
+    void TickFrame();
+
+    void WriteMemory(VAddr cpu_addr, u64 size);
+
+    void CachedWriteMemory(VAddr cpu_addr, u64 size);
+
+    void DownloadMemory(VAddr cpu_addr, u64 size);
+
+    bool InlineMemory(VAddr dest_address, size_t copy_size, std::span<const u8> inlined_buffer);
+
+    void BindGraphicsUniformBuffer(size_t stage, u32 index, GPUVAddr gpu_addr, u32 size);
+
+    void DisableGraphicsUniformBuffer(size_t stage, u32 index);
+
+    void UpdateGraphicsBuffers(bool is_indexed);
+
+    void UpdateComputeBuffers();
+
+    void BindHostGeometryBuffers(bool is_indexed);
+
+    void BindHostStageBuffers(size_t stage);
+
+    void BindHostComputeBuffers();
+
+    void SetUniformBuffersState(const std::array<u32, NUM_STAGES>& mask,
+                                const UniformBufferSizes* sizes);
+
+    void SetComputeUniformBufferState(u32 mask, const ComputeUniformBufferSizes* sizes);
+
+    void UnbindGraphicsStorageBuffers(size_t stage);
+
+    void BindGraphicsStorageBuffer(size_t stage, size_t ssbo_index, u32 cbuf_index, u32 cbuf_offset,
+                                   bool is_written);
+
+    void UnbindGraphicsTextureBuffers(size_t stage);
+
+    void BindGraphicsTextureBuffer(size_t stage, size_t tbo_index, GPUVAddr gpu_addr, u32 size,
+                                   PixelFormat format, bool is_written, bool is_image);
+
+    void UnbindComputeStorageBuffers();
+
+    void BindComputeStorageBuffer(size_t ssbo_index, u32 cbuf_index, u32 cbuf_offset,
+                                  bool is_written);
+
+    void UnbindComputeTextureBuffers();
+
+    void BindComputeTextureBuffer(size_t tbo_index, GPUVAddr gpu_addr, u32 size, PixelFormat format,
+                                  bool is_written, bool is_image);
+
+    [[nodiscard]] std::pair<Buffer*, u32> ObtainBuffer(GPUVAddr gpu_addr, u32 size,
+                                                       ObtainBufferSynchronize sync_info,
+                                                       ObtainBufferOperation post_op);
+    void FlushCachedWrites();
+
+    /// Return true when there are uncommitted buffers to be downloaded
+    [[nodiscard]] bool HasUncommittedFlushes() const noexcept;
+
+    void AccumulateFlushes();
+
+    /// Return true when the caller should wait for async downloads
+    [[nodiscard]] bool ShouldWaitAsyncFlushes() const noexcept;
+
+    /// Commit asynchronous downloads
+    void CommitAsyncFlushes();
+    void CommitAsyncFlushesHigh();
+
+    /// Pop asynchronous downloads
+    void PopAsyncFlushes();
+    void PopAsyncBuffers();
+
+    bool DMACopy(GPUVAddr src_address, GPUVAddr dest_address, u64 amount);
+
+    bool DMAClear(GPUVAddr src_address, u64 amount, u32 value);
+
+    /// Return true when a CPU region is modified from the GPU
+    [[nodiscard]] bool IsRegionGpuModified(VAddr addr, size_t size);
+
+    /// Return true when a region is registered on the cache
+    [[nodiscard]] bool IsRegionRegistered(VAddr addr, size_t size);
+
+    /// Return true when a CPU region is modified from the CPU
+    [[nodiscard]] bool IsRegionCpuModified(VAddr addr, size_t size);
+
+    void SetDrawIndirect(
+        const Tegra::Engines::DrawManager::IndirectParams* current_draw_indirect_) {
+        current_draw_indirect = current_draw_indirect_;
+    }
+
+    [[nodiscard]] std::pair<Buffer*, u32> GetDrawIndirectCount();
+
+    [[nodiscard]] std::pair<Buffer*, u32> GetDrawIndirectBuffer();
+
+    std::recursive_mutex mutex;
+    Runtime& runtime;
+
+private:
+    template <typename Func>
+    static void ForEachEnabledBit(u32 enabled_mask, Func&& func) {
+        for (u32 index = 0; enabled_mask != 0; ++index, enabled_mask >>= 1) {
+            const int disabled_bits = std::countr_zero(enabled_mask);
+            index += disabled_bits;
+            enabled_mask >>= disabled_bits;
+            func(index);
+        }
+    }
+
+    template <typename Func>
+    void ForEachBufferInRange(VAddr cpu_addr, u64 size, Func&& func) {
+        const u64 page_end = Common::DivCeil(cpu_addr + size, CACHING_PAGESIZE);
+        for (u64 page = cpu_addr >> CACHING_PAGEBITS; page < page_end;) {
+            const BufferId buffer_id = page_table[page];
+            if (!buffer_id) {
+                ++page;
+                continue;
+            }
+            Buffer& buffer = slot_buffers[buffer_id];
+            func(buffer_id, buffer);
+
+            const VAddr end_addr = buffer.CpuAddr() + buffer.SizeBytes();
+            page = Common::DivCeil(end_addr, CACHING_PAGESIZE);
+        }
+    }
+
+    template <typename Func>
+    void ForEachInRangeSet(IntervalSet& current_range, VAddr cpu_addr, u64 size, Func&& func) {
+        const VAddr start_address = cpu_addr;
+        const VAddr end_address = start_address + size;
+        const IntervalType search_interval{start_address, end_address};
+        auto it = current_range.lower_bound(search_interval);
+        if (it == current_range.end()) {
+            return;
+        }
+        auto end_it = current_range.upper_bound(search_interval);
+        for (; it != end_it; it++) {
+            VAddr inter_addr_end = it->upper();
+            VAddr inter_addr = it->lower();
+            if (inter_addr_end > end_address) {
+                inter_addr_end = end_address;
+            }
+            if (inter_addr < start_address) {
+                inter_addr = start_address;
+            }
+            func(inter_addr, inter_addr_end);
+        }
+    }
+
+    template <typename Func>
+    void ForEachInOverlapCounter(OverlapCounter& current_range, VAddr cpu_addr, u64 size,
+                                 Func&& func) {
+        const VAddr start_address = cpu_addr;
+        const VAddr end_address = start_address + size;
+        const IntervalType search_interval{start_address, end_address};
+        auto it = current_range.lower_bound(search_interval);
+        if (it == current_range.end()) {
+            return;
+        }
+        auto end_it = current_range.upper_bound(search_interval);
+        for (; it != end_it; it++) {
+            auto& inter = it->first;
+            VAddr inter_addr_end = inter.upper();
+            VAddr inter_addr = inter.lower();
+            if (inter_addr_end > end_address) {
+                inter_addr_end = end_address;
+            }
+            if (inter_addr < start_address) {
+                inter_addr = start_address;
+            }
+            func(inter_addr, inter_addr_end, it->second);
+        }
+    }
+
+    void RemoveEachInOverlapCounter(OverlapCounter& current_range,
+                                    const IntervalType search_interval, int subtract_value) {
+        bool any_removals = false;
+        current_range.add(std::make_pair(search_interval, subtract_value));
+        do {
+            any_removals = false;
+            auto it = current_range.lower_bound(search_interval);
+            if (it == current_range.end()) {
+                return;
+            }
+            auto end_it = current_range.upper_bound(search_interval);
+            for (; it != end_it; it++) {
+                if (it->second <= 0) {
+                    any_removals = true;
+                    current_range.erase(it);
+                    break;
+                }
+            }
+        } while (any_removals);
+    }
+
+    static bool IsRangeGranular(VAddr cpu_addr, size_t size) {
+        return (cpu_addr & ~Core::Memory::YUZU_PAGEMASK) ==
+               ((cpu_addr + size) & ~Core::Memory::YUZU_PAGEMASK);
+    }
+
+    void RunGarbageCollector();
+
+    void WaitOnAsyncFlushes(VAddr cpu_addr, u64 size);
+
+    void BindHostIndexBuffer();
+
+    void BindHostVertexBuffers();
+
+    void BindHostDrawIndirectBuffers();
+
+    void BindHostGraphicsUniformBuffers(size_t stage);
+
+    void BindHostGraphicsUniformBuffer(size_t stage, u32 index, u32 binding_index, bool needs_bind);
+
+    void BindHostGraphicsStorageBuffers(size_t stage);
+
+    void BindHostGraphicsTextureBuffers(size_t stage);
+
+    void BindHostTransformFeedbackBuffers();
+
+    void BindHostComputeUniformBuffers();
+
+    void BindHostComputeStorageBuffers();
+
+    void BindHostComputeTextureBuffers();
+
+    void DoUpdateGraphicsBuffers(bool is_indexed);
+
+    void DoUpdateComputeBuffers();
+
+    void UpdateIndexBuffer();
+
+    void UpdateVertexBuffers();
+
+    void UpdateVertexBuffer(u32 index);
+
+    void UpdateDrawIndirect();
+
+    void UpdateUniformBuffers(size_t stage);
+
+    void UpdateStorageBuffers(size_t stage);
+
+    void UpdateTextureBuffers(size_t stage);
+
+    void UpdateTransformFeedbackBuffers();
+
+    void UpdateTransformFeedbackBuffer(u32 index);
+
+    void UpdateComputeUniformBuffers();
+
+    void UpdateComputeStorageBuffers();
+
+    void UpdateComputeTextureBuffers();
+
+    void MarkWrittenBuffer(BufferId buffer_id, VAddr cpu_addr, u32 size);
+
+    [[nodiscard]] BufferId FindBuffer(VAddr cpu_addr, u32 size);
+
+    [[nodiscard]] OverlapResult ResolveOverlaps(VAddr cpu_addr, u32 wanted_size);
+
+    void JoinOverlap(BufferId new_buffer_id, BufferId overlap_id, bool accumulate_stream_score);
+
+    [[nodiscard]] BufferId CreateBuffer(VAddr cpu_addr, u32 wanted_size);
+
+    void Register(BufferId buffer_id);
+
+    void Unregister(BufferId buffer_id);
+
+    template <bool insert>
+    void ChangeRegister(BufferId buffer_id);
+
+    void TouchBuffer(Buffer& buffer, BufferId buffer_id) noexcept;
+
+    bool SynchronizeBuffer(Buffer& buffer, VAddr cpu_addr, u32 size);
+
+    bool SynchronizeBufferImpl(Buffer& buffer, VAddr cpu_addr, u32 size);
+
+    bool SynchronizeBufferNoModified(Buffer& buffer, VAddr cpu_addr, u32 size);
+
+    void UploadMemory(Buffer& buffer, u64 total_size_bytes, u64 largest_copy,
+                      std::span<BufferCopy> copies);
+
+    void ImmediateUploadMemory(Buffer& buffer, u64 largest_copy,
+                               std::span<const BufferCopy> copies);
+
+    void MappedUploadMemory(Buffer& buffer, u64 total_size_bytes, std::span<BufferCopy> copies);
+
+    void DownloadBufferMemory(Buffer& buffer_id);
+
+    void DownloadBufferMemory(Buffer& buffer_id, VAddr cpu_addr, u64 size);
+
+    void DeleteBuffer(BufferId buffer_id, bool do_not_mark = false);
+
+    void NotifyBufferDeletion();
+
+    [[nodiscard]] Binding StorageBufferBinding(GPUVAddr ssbo_addr, u32 cbuf_index,
+                                               bool is_written) const;
+
+    [[nodiscard]] TextureBufferBinding GetTextureBufferBinding(GPUVAddr gpu_addr, u32 size,
+                                                               PixelFormat format);
+
+    [[nodiscard]] std::span<const u8> ImmediateBufferWithData(VAddr cpu_addr, size_t size);
+
+    [[nodiscard]] std::span<u8> ImmediateBuffer(size_t wanted_capacity);
+
+    [[nodiscard]] bool HasFastUniformBufferBound(size_t stage, u32 binding_index) const noexcept;
+
+    void ClearDownload(IntervalType subtract_interval);
+
+    VideoCore::RasterizerInterface& rasterizer;
+    Core::Memory::Memory& cpu_memory;
+
+    SlotVector<Buffer> slot_buffers;
+    DelayedDestructionRing<Buffer, 8> delayed_destruction_ring;
+
+    const Tegra::Engines::DrawManager::IndirectParams* current_draw_indirect{};
+
+    u32 last_index_count = 0;
+
+    Binding index_buffer;
+    std::array<Binding, NUM_VERTEX_BUFFERS> vertex_buffers;
+    std::array<std::array<Binding, NUM_GRAPHICS_UNIFORM_BUFFERS>, NUM_STAGES> uniform_buffers;
+    std::array<std::array<Binding, NUM_STORAGE_BUFFERS>, NUM_STAGES> storage_buffers;
+    std::array<std::array<TextureBufferBinding, NUM_TEXTURE_BUFFERS>, NUM_STAGES> texture_buffers;
+    std::array<Binding, NUM_TRANSFORM_FEEDBACK_BUFFERS> transform_feedback_buffers;
+    Binding count_buffer_binding;
+    Binding indirect_buffer_binding;
+
+    std::array<Binding, NUM_COMPUTE_UNIFORM_BUFFERS> compute_uniform_buffers;
+    std::array<Binding, NUM_STORAGE_BUFFERS> compute_storage_buffers;
+    std::array<TextureBufferBinding, NUM_TEXTURE_BUFFERS> compute_texture_buffers;
+
+    std::array<u32, NUM_STAGES> enabled_uniform_buffer_masks{};
+    u32 enabled_compute_uniform_buffer_mask = 0;
+
+    const UniformBufferSizes* uniform_buffer_sizes{};
+    const ComputeUniformBufferSizes* compute_uniform_buffer_sizes{};
+
+    std::array<u32, NUM_STAGES> enabled_storage_buffers{};
+    std::array<u32, NUM_STAGES> written_storage_buffers{};
+    u32 enabled_compute_storage_buffers = 0;
+    u32 written_compute_storage_buffers = 0;
+
+    std::array<u32, NUM_STAGES> enabled_texture_buffers{};
+    std::array<u32, NUM_STAGES> written_texture_buffers{};
+    std::array<u32, NUM_STAGES> image_texture_buffers{};
+    u32 enabled_compute_texture_buffers = 0;
+    u32 written_compute_texture_buffers = 0;
+    u32 image_compute_texture_buffers = 0;
+
+    std::array<u32, 16> uniform_cache_hits{};
+    std::array<u32, 16> uniform_cache_shots{};
+
+    u32 uniform_buffer_skip_cache_size = DEFAULT_SKIP_CACHE_SIZE;
+
+    bool has_deleted_buffers = false;
+
+    std::conditional_t<HAS_PERSISTENT_UNIFORM_BUFFER_BINDINGS, std::array<u32, NUM_STAGES>, Empty>
+        dirty_uniform_buffers{};
+    std::conditional_t<IS_OPENGL, std::array<u32, NUM_STAGES>, Empty> fast_bound_uniform_buffers{};
+    std::conditional_t<HAS_PERSISTENT_UNIFORM_BUFFER_BINDINGS,
+                       std::array<std::array<u32, NUM_GRAPHICS_UNIFORM_BUFFERS>, NUM_STAGES>, Empty>
+        uniform_buffer_binding_sizes{};
+
+    std::vector<BufferId> cached_write_buffer_ids;
+
+    MemoryTracker memory_tracker;
+    IntervalSet uncommitted_ranges;
+    IntervalSet common_ranges;
+    IntervalSet cached_ranges;
+    IntervalSet pending_ranges;
+    std::deque<IntervalSet> committed_ranges;
+
+    // Async Buffers
+    OverlapCounter async_downloads;
+    std::deque<std::optional<Async_Buffer>> async_buffers;
+    std::deque<boost::container::small_vector<BufferCopy, 4>> pending_downloads;
+    std::optional<Async_Buffer> current_buffer;
+
+    std::deque<Async_Buffer> async_buffers_death_ring;
+
+    size_t immediate_buffer_capacity = 0;
+    Common::ScratchBuffer<u8> immediate_buffer_alloc;
+
+    struct LRUItemParams {
+        using ObjectType = BufferId;
+        using TickType = u64;
+    };
+    Common::LeastRecentlyUsedCache<LRUItemParams> lru_cache;
+    u64 frame_tick = 0;
+    u64 total_used_memory = 0;
+    u64 minimum_memory = 0;
+    u64 critical_memory = 0;
+    BufferId inline_buffer_id;
+
+    bool active_async_buffers = false;
+
+    std::array<BufferId, ((1ULL << 39) >> CACHING_PAGEBITS)> page_table;
+};
+
+} // namespace VideoCommon

src/video_core/buffer_cache/memory_tracker_base.h

@@ -0,0 +1,273 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#pragma once
+
+#include <algorithm>
+#include <bit>
+#include <deque>
+#include <limits>
+#include <type_traits>
+#include <unordered_set>
+#include <utility>
+
+#include "common/alignment.h"
+#include "common/common_types.h"
+#include "video_core/buffer_cache/word_manager.h"
+
+namespace VideoCommon {
+
+template <class RasterizerInterface>
+class MemoryTrackerBase {
+    static constexpr size_t MAX_CPU_PAGE_BITS = 39;
+    static constexpr size_t HIGHER_PAGE_BITS = 22;
+    static constexpr size_t HIGHER_PAGE_SIZE = 1ULL << HIGHER_PAGE_BITS;
+    static constexpr size_t HIGHER_PAGE_MASK = HIGHER_PAGE_SIZE - 1ULL;
+    static constexpr size_t NUM_HIGH_PAGES = 1ULL << (MAX_CPU_PAGE_BITS - HIGHER_PAGE_BITS);
+    static constexpr size_t MANAGER_POOL_SIZE = 32;
+    static constexpr size_t WORDS_STACK_NEEDED = HIGHER_PAGE_SIZE / BYTES_PER_WORD;
+    using Manager = WordManager<RasterizerInterface, WORDS_STACK_NEEDED>;
+
+public:
+    MemoryTrackerBase(RasterizerInterface& rasterizer_) : rasterizer{&rasterizer_} {}
+    ~MemoryTrackerBase() = default;
+
+    /// Returns the inclusive CPU modified range in a begin end pair
+    [[nodiscard]] std::pair<u64, u64> ModifiedCpuRegion(VAddr query_cpu_addr,
+                                                        u64 query_size) noexcept {
+        return IteratePairs<true>(
+            query_cpu_addr, query_size, [](Manager* manager, u64 offset, size_t size) {
+                return manager->template ModifiedRegion<Type::CPU>(offset, size);
+            });
+    }
+
+    /// Returns the inclusive GPU modified range in a begin end pair
+    [[nodiscard]] std::pair<u64, u64> ModifiedGpuRegion(VAddr query_cpu_addr,
+                                                        u64 query_size) noexcept {
+        return IteratePairs<false>(
+            query_cpu_addr, query_size, [](Manager* manager, u64 offset, size_t size) {
+                return manager->template ModifiedRegion<Type::GPU>(offset, size);
+            });
+    }
+
+    /// Returns true if a region has been modified from the CPU
+    [[nodiscard]] bool IsRegionCpuModified(VAddr query_cpu_addr, u64 query_size) noexcept {
+        return IteratePages<true>(
+            query_cpu_addr, query_size, [](Manager* manager, u64 offset, size_t size) {
+                return manager->template IsRegionModified<Type::CPU>(offset, size);
+            });
+    }
+
+    /// Returns true if a region has been modified from the GPU
+    [[nodiscard]] bool IsRegionGpuModified(VAddr query_cpu_addr, u64 query_size) noexcept {
+        return IteratePages<false>(
+            query_cpu_addr, query_size, [](Manager* manager, u64 offset, size_t size) {
+                return manager->template IsRegionModified<Type::GPU>(offset, size);
+            });
+    }
+
+    /// Mark region as CPU modified, notifying the rasterizer about this change
+    void MarkRegionAsCpuModified(VAddr dirty_cpu_addr, u64 query_size) {
+        IteratePages<true>(dirty_cpu_addr, query_size,
+                           [](Manager* manager, u64 offset, size_t size) {
+                               manager->template ChangeRegionState<Type::CPU, true>(
+                                   manager->GetCpuAddr() + offset, size);
+                           });
+    }
+
+    /// Unmark region as CPU modified, notifying the rasterizer about this change
+    void UnmarkRegionAsCpuModified(VAddr dirty_cpu_addr, u64 query_size) {
+        IteratePages<true>(dirty_cpu_addr, query_size,
+                           [](Manager* manager, u64 offset, size_t size) {
+                               manager->template ChangeRegionState<Type::CPU, false>(
+                                   manager->GetCpuAddr() + offset, size);
+                           });
+    }
+
+    /// Mark region as modified from the host GPU
+    void MarkRegionAsGpuModified(VAddr dirty_cpu_addr, u64 query_size) noexcept {
+        IteratePages<true>(dirty_cpu_addr, query_size,
+                           [](Manager* manager, u64 offset, size_t size) {
+                               manager->template ChangeRegionState<Type::GPU, true>(
+                                   manager->GetCpuAddr() + offset, size);
+                           });
+    }
+
+    /// Unmark region as modified from the host GPU
+    void UnmarkRegionAsGpuModified(VAddr dirty_cpu_addr, u64 query_size) noexcept {
+        IteratePages<true>(dirty_cpu_addr, query_size,
+                           [](Manager* manager, u64 offset, size_t size) {
+                               manager->template ChangeRegionState<Type::GPU, false>(
+                                   manager->GetCpuAddr() + offset, size);
+                           });
+    }
+
+    /// Mark region as modified from the CPU
+    /// but don't mark it as modified until FlusHCachedWrites is called.
+    void CachedCpuWrite(VAddr dirty_cpu_addr, u64 query_size) {
+        IteratePages<true>(
+            dirty_cpu_addr, query_size, [this](Manager* manager, u64 offset, size_t size) {
+                const VAddr cpu_address = manager->GetCpuAddr() + offset;
+                manager->template ChangeRegionState<Type::CachedCPU, true>(cpu_address, size);
+                cached_pages.insert(static_cast<u32>(cpu_address >> HIGHER_PAGE_BITS));
+            });
+    }
+
+    /// Flushes cached CPU writes, and notify the rasterizer about the deltas
+    void FlushCachedWrites(VAddr query_cpu_addr, u64 query_size) noexcept {
+        IteratePages<false>(query_cpu_addr, query_size,
+                            [](Manager* manager, [[maybe_unused]] u64 offset,
+                               [[maybe_unused]] size_t size) { manager->FlushCachedWrites(); });
+    }
+
+    void FlushCachedWrites() noexcept {
+        for (auto id : cached_pages) {
+            top_tier[id]->FlushCachedWrites();
+        }
+        cached_pages.clear();
+    }
+
+    /// Call 'func' for each CPU modified range and unmark those pages as CPU modified
+    template <typename Func>
+    void ForEachUploadRange(VAddr query_cpu_range, u64 query_size, Func&& func) {
+        IteratePages<true>(query_cpu_range, query_size,
+                           [&func](Manager* manager, u64 offset, size_t size) {
+                               manager->template ForEachModifiedRange<Type::CPU, true>(
+                                   manager->GetCpuAddr() + offset, size, func);
+                           });
+    }
+
+    /// Call 'func' for each GPU modified range and unmark those pages as GPU modified
+    template <typename Func>
+    void ForEachDownloadRange(VAddr query_cpu_range, u64 query_size, bool clear, Func&& func) {
+        IteratePages<false>(query_cpu_range, query_size,
+                            [&func, clear](Manager* manager, u64 offset, size_t size) {
+                                if (clear) {
+                                    manager->template ForEachModifiedRange<Type::GPU, true>(
+                                        manager->GetCpuAddr() + offset, size, func);
+                                } else {
+                                    manager->template ForEachModifiedRange<Type::GPU, false>(
+                                        manager->GetCpuAddr() + offset, size, func);
+                                }
+                            });
+    }
+
+    template <typename Func>
+    void ForEachDownloadRangeAndClear(VAddr query_cpu_range, u64 query_size, Func&& func) {
+        IteratePages<false>(query_cpu_range, query_size,
+                            [&func](Manager* manager, u64 offset, size_t size) {
+                                manager->template ForEachModifiedRange<Type::GPU, true>(
+                                    manager->GetCpuAddr() + offset, size, func);
+                            });
+    }
+
+private:
+    template <bool create_region_on_fail, typename Func>
+    bool IteratePages(VAddr cpu_address, size_t size, Func&& func) {
+        using FuncReturn = typename std::invoke_result<Func, Manager*, u64, size_t>::type;
+        static constexpr bool BOOL_BREAK = std::is_same_v<FuncReturn, bool>;
+        std::size_t remaining_size{size};
+        std::size_t page_index{cpu_address >> HIGHER_PAGE_BITS};
+        u64 page_offset{cpu_address & HIGHER_PAGE_MASK};
+        while (remaining_size > 0) {
+            const std::size_t copy_amount{
+                std::min<std::size_t>(HIGHER_PAGE_SIZE - page_offset, remaining_size)};
+            auto* manager{top_tier[page_index]};
+            if (manager) {
+                if constexpr (BOOL_BREAK) {
+                    if (func(manager, page_offset, copy_amount)) {
+                        return true;
+                    }
+                } else {
+                    func(manager, page_offset, copy_amount);
+                }
+            } else if constexpr (create_region_on_fail) {
+                CreateRegion(page_index);
+                manager = top_tier[page_index];
+                if constexpr (BOOL_BREAK) {
+                    if (func(manager, page_offset, copy_amount)) {
+                        return true;
+                    }
+                } else {
+                    func(manager, page_offset, copy_amount);
+                }
+            }
+            page_index++;
+            page_offset = 0;
+            remaining_size -= copy_amount;
+        }
+        return false;
+    }
+
+    template <bool create_region_on_fail, typename Func>
+    std::pair<u64, u64> IteratePairs(VAddr cpu_address, size_t size, Func&& func) {
+        std::size_t remaining_size{size};
+        std::size_t page_index{cpu_address >> HIGHER_PAGE_BITS};
+        u64 page_offset{cpu_address & HIGHER_PAGE_MASK};
+        u64 begin = std::numeric_limits<u64>::max();
+        u64 end = 0;
+        while (remaining_size > 0) {
+            const std::size_t copy_amount{
+                std::min<std::size_t>(HIGHER_PAGE_SIZE - page_offset, remaining_size)};
+            auto* manager{top_tier[page_index]};
+            const auto execute = [&] {
+                auto [new_begin, new_end] = func(manager, page_offset, copy_amount);
+                if (new_begin != 0 || new_end != 0) {
+                    const u64 base_address = page_index << HIGHER_PAGE_BITS;
+                    begin = std::min(new_begin + base_address, begin);
+                    end = std::max(new_end + base_address, end);
+                }
+            };
+            if (manager) {
+                execute();
+            } else if constexpr (create_region_on_fail) {
+                CreateRegion(page_index);
+                manager = top_tier[page_index];
+                execute();
+            }
+            page_index++;
+            page_offset = 0;
+            remaining_size -= copy_amount;
+        }
+        if (begin < end) {
+            return std::make_pair(begin, end);
+        } else {
+            return std::make_pair(0ULL, 0ULL);
+        }
+    }
+
+    void CreateRegion(std::size_t page_index) {
+        const VAddr base_cpu_addr = page_index << HIGHER_PAGE_BITS;
+        top_tier[page_index] = GetNewManager(base_cpu_addr);
+    }
+
+    Manager* GetNewManager(VAddr base_cpu_addess) {
+        const auto on_return = [&] {
+            auto* new_manager = free_managers.front();
+            new_manager->SetCpuAddress(base_cpu_addess);
+            free_managers.pop_front();
+            return new_manager;
+        };
+        if (!free_managers.empty()) {
+            return on_return();
+        }
+        manager_pool.emplace_back();
+        auto& last_pool = manager_pool.back();
+        for (size_t i = 0; i < MANAGER_POOL_SIZE; i++) {
+            new (&last_pool[i]) Manager(0, *rasterizer, HIGHER_PAGE_SIZE);
+            free_managers.push_back(&last_pool[i]);
+        }
+        return on_return();
+    }
+
+    std::deque<std::array<Manager, MANAGER_POOL_SIZE>> manager_pool;
+    std::deque<Manager*> free_managers;
+
+    std::array<Manager*, NUM_HIGH_PAGES> top_tier{};
+
+    std::unordered_set<u32> cached_pages;
+
+    RasterizerInterface* rasterizer = nullptr;
+};
+
+} // namespace VideoCommon

src/video_core/buffer_cache/word_manager.h

@@ -0,0 +1,462 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#pragma once
+
+#include <algorithm>
+#include <bit>
+#include <limits>
+#include <span>
+#include <utility>
+
+#include "common/alignment.h"
+#include "common/common_funcs.h"
+#include "common/common_types.h"
+#include "common/div_ceil.h"
+#include "core/memory.h"
+
+namespace VideoCommon {
+
+constexpr u64 PAGES_PER_WORD = 64;
+constexpr u64 BYTES_PER_PAGE = Core::Memory::YUZU_PAGESIZE;
+constexpr u64 BYTES_PER_WORD = PAGES_PER_WORD * BYTES_PER_PAGE;
+
+enum class Type {
+    CPU,
+    GPU,
+    CachedCPU,
+    Untracked,
+};
+
+/// Vector tracking modified pages tightly packed with small vector optimization
+template <size_t stack_words = 1>
+struct WordsArray {
+    /// Returns the pointer to the words state
+    [[nodiscard]] const u64* Pointer(bool is_short) const noexcept {
+        return is_short ? stack.data() : heap;
+    }
+
+    /// Returns the pointer to the words state
+    [[nodiscard]] u64* Pointer(bool is_short) noexcept {
+        return is_short ? stack.data() : heap;
+    }
+
+    std::array<u64, stack_words> stack{}; ///< Small buffers storage
+    u64* heap;                            ///< Not-small buffers pointer to the storage
+};
+
+template <size_t stack_words = 1>
+struct Words {
+    explicit Words() = default;
+    explicit Words(u64 size_bytes_) : size_bytes{size_bytes_} {
+        num_words = Common::DivCeil(size_bytes, BYTES_PER_WORD);
+        if (IsShort()) {
+            cpu.stack.fill(~u64{0});
+            gpu.stack.fill(0);
+            cached_cpu.stack.fill(0);
+            untracked.stack.fill(~u64{0});
+        } else {
+            // Share allocation between CPU and GPU pages and set their default values
+            u64* const alloc = new u64[num_words * 4];
+            cpu.heap = alloc;
+            gpu.heap = alloc + num_words;
+            cached_cpu.heap = alloc + num_words * 2;
+            untracked.heap = alloc + num_words * 3;
+            std::fill_n(cpu.heap, num_words, ~u64{0});
+            std::fill_n(gpu.heap, num_words, 0);
+            std::fill_n(cached_cpu.heap, num_words, 0);
+            std::fill_n(untracked.heap, num_words, ~u64{0});
+        }
+        // Clean up tailing bits
+        const u64 last_word_size = size_bytes % BYTES_PER_WORD;
+        const u64 last_local_page = Common::DivCeil(last_word_size, BYTES_PER_PAGE);
+        const u64 shift = (PAGES_PER_WORD - last_local_page) % PAGES_PER_WORD;
+        const u64 last_word = (~u64{0} << shift) >> shift;
+        cpu.Pointer(IsShort())[NumWords() - 1] = last_word;
+        untracked.Pointer(IsShort())[NumWords() - 1] = last_word;
+    }
+
+    ~Words() {
+        Release();
+    }
+
+    Words& operator=(Words&& rhs) noexcept {
+        Release();
+        size_bytes = rhs.size_bytes;
+        num_words = rhs.num_words;
+        cpu = rhs.cpu;
+        gpu = rhs.gpu;
+        cached_cpu = rhs.cached_cpu;
+        untracked = rhs.untracked;
+        rhs.cpu.heap = nullptr;
+        return *this;
+    }
+
+    Words(Words&& rhs) noexcept
+        : size_bytes{rhs.size_bytes}, num_words{rhs.num_words}, cpu{rhs.cpu}, gpu{rhs.gpu},
+          cached_cpu{rhs.cached_cpu}, untracked{rhs.untracked} {
+        rhs.cpu.heap = nullptr;
+    }
+
+    Words& operator=(const Words&) = delete;
+    Words(const Words&) = delete;
+
+    /// Returns true when the buffer fits in the small vector optimization
+    [[nodiscard]] bool IsShort() const noexcept {
+        return num_words <= stack_words;
+    }
+
+    /// Returns the number of words of the buffer
+    [[nodiscard]] size_t NumWords() const noexcept {
+        return num_words;
+    }
+
+    /// Release buffer resources
+    void Release() {
+        if (!IsShort()) {
+            // CPU written words is the base for the heap allocation
+            delete[] cpu.heap;
+        }
+    }
+
+    template <Type type>
+    std::span<u64> Span() noexcept {
+        if constexpr (type == Type::CPU) {
+            return std::span<u64>(cpu.Pointer(IsShort()), num_words);
+        } else if constexpr (type == Type::GPU) {
+            return std::span<u64>(gpu.Pointer(IsShort()), num_words);
+        } else if constexpr (type == Type::CachedCPU) {
+            return std::span<u64>(cached_cpu.Pointer(IsShort()), num_words);
+        } else if constexpr (type == Type::Untracked) {
+            return std::span<u64>(untracked.Pointer(IsShort()), num_words);
+        }
+    }
+
+    template <Type type>
+    std::span<const u64> Span() const noexcept {
+        if constexpr (type == Type::CPU) {
+            return std::span<const u64>(cpu.Pointer(IsShort()), num_words);
+        } else if constexpr (type == Type::GPU) {
+            return std::span<const u64>(gpu.Pointer(IsShort()), num_words);
+        } else if constexpr (type == Type::CachedCPU) {
+            return std::span<const u64>(cached_cpu.Pointer(IsShort()), num_words);
+        } else if constexpr (type == Type::Untracked) {
+            return std::span<const u64>(untracked.Pointer(IsShort()), num_words);
+        }
+    }
+
+    u64 size_bytes = 0;
+    size_t num_words = 0;
+    WordsArray<stack_words> cpu;
+    WordsArray<stack_words> gpu;
+    WordsArray<stack_words> cached_cpu;
+    WordsArray<stack_words> untracked;
+};
+
+template <class RasterizerInterface, size_t stack_words = 1>
+class WordManager {
+public:
+    explicit WordManager(VAddr cpu_addr_, RasterizerInterface& rasterizer_, u64 size_bytes)
+        : cpu_addr{cpu_addr_}, rasterizer{&rasterizer_}, words{size_bytes} {}
+
+    explicit WordManager() = default;
+
+    void SetCpuAddress(VAddr new_cpu_addr) {
+        cpu_addr = new_cpu_addr;
+    }
+
+    VAddr GetCpuAddr() const {
+        return cpu_addr;
+    }
+
+    static u64 ExtractBits(u64 word, size_t page_start, size_t page_end) {
+        constexpr size_t number_bits = sizeof(u64) * 8;
+        const size_t limit_page_end = number_bits - std::min(page_end, number_bits);
+        u64 bits = (word >> page_start) << page_start;
+        bits = (bits << limit_page_end) >> limit_page_end;
+        return bits;
+    }
+
+    static std::pair<size_t, size_t> GetWordPage(VAddr address) {
+        const size_t converted_address = static_cast<size_t>(address);
+        const size_t word_number = converted_address / BYTES_PER_WORD;
+        const size_t amount_pages = converted_address % BYTES_PER_WORD;
+        return std::make_pair(word_number, amount_pages / BYTES_PER_PAGE);
+    }
+
+    template <typename Func>
+    void IterateWords(size_t offset, size_t size, Func&& func) const {
+        using FuncReturn = std::invoke_result_t<Func, std::size_t, u64>;
+        static constexpr bool BOOL_BREAK = std::is_same_v<FuncReturn, bool>;
+        const size_t start = static_cast<size_t>(std::max<s64>(static_cast<s64>(offset), 0LL));
+        const size_t end = static_cast<size_t>(std::max<s64>(static_cast<s64>(offset + size), 0LL));
+        if (start >= SizeBytes() || end <= start) {
+            return;
+        }
+        auto [start_word, start_page] = GetWordPage(start);
+        auto [end_word, end_page] = GetWordPage(end + BYTES_PER_PAGE - 1ULL);
+        const size_t num_words = NumWords();
+        start_word = std::min(start_word, num_words);
+        end_word = std::min(end_word, num_words);
+        const size_t diff = end_word - start_word;
+        end_word += (end_page + PAGES_PER_WORD - 1ULL) / PAGES_PER_WORD;
+        end_word = std::min(end_word, num_words);
+        end_page += diff * PAGES_PER_WORD;
+        constexpr u64 base_mask{~0ULL};
+        for (size_t word_index = start_word; word_index < end_word; word_index++) {
+            const u64 mask = ExtractBits(base_mask, start_page, end_page);
+            start_page = 0;
+            end_page -= PAGES_PER_WORD;
+            if constexpr (BOOL_BREAK) {
+                if (func(word_index, mask)) {
+                    return;
+                }
+            } else {
+                func(word_index, mask);
+            }
+        }
+    }
+
+    template <typename Func>
+    void IteratePages(u64 mask, Func&& func) const {
+        size_t offset = 0;
+        while (mask != 0) {
+            const size_t empty_bits = std::countr_zero(mask);
+            offset += empty_bits;
+            mask = mask >> empty_bits;
+
+            const size_t continuous_bits = std::countr_one(mask);
+            func(offset, continuous_bits);
+            mask = continuous_bits < PAGES_PER_WORD ? (mask >> continuous_bits) : 0;
+            offset += continuous_bits;
+        }
+    }
+
+    /**
+     * Change the state of a range of pages
+     *
+     * @param dirty_addr    Base address to mark or unmark as modified
+     * @param size          Size in bytes to mark or unmark as modified
+     */
+    template <Type type, bool enable>
+    void ChangeRegionState(u64 dirty_addr, u64 size) noexcept(type == Type::GPU) {
+        std::span<u64> state_words = words.template Span<type>();
+        [[maybe_unused]] std::span<u64> untracked_words = words.template Span<Type::Untracked>();
+        [[maybe_unused]] std::span<u64> cached_words = words.template Span<Type::CachedCPU>();
+        IterateWords(dirty_addr - cpu_addr, size, [&](size_t index, u64 mask) {
+            if constexpr (type == Type::CPU || type == Type::CachedCPU) {
+                NotifyRasterizer<!enable>(index, untracked_words[index], mask);
+            }
+            if constexpr (enable) {
+                state_words[index] |= mask;
+                if constexpr (type == Type::CPU || type == Type::CachedCPU) {
+                    untracked_words[index] |= mask;
+                }
+                if constexpr (type == Type::CPU) {
+                    cached_words[index] &= ~mask;
+                }
+            } else {
+                if constexpr (type == Type::CPU) {
+                    const u64 word = state_words[index] & mask;
+                    cached_words[index] &= ~word;
+                }
+                state_words[index] &= ~mask;
+                if constexpr (type == Type::CPU || type == Type::CachedCPU) {
+                    untracked_words[index] &= ~mask;
+                }
+            }
+        });
+    }
+
+    /**
+     * Loop over each page in the given range, turn off those bits and notify the rasterizer if
+     * needed. Call the given function on each turned off range.
+     *
+     * @param query_cpu_range Base CPU address to loop over
+     * @param size            Size in bytes of the CPU range to loop over
+     * @param func            Function to call for each turned off region
+     */
+    template <Type type, bool clear, typename Func>
+    void ForEachModifiedRange(VAddr query_cpu_range, s64 size, Func&& func) {
+        static_assert(type != Type::Untracked);
+
+        std::span<u64> state_words = words.template Span<type>();
+        [[maybe_unused]] std::span<u64> untracked_words = words.template Span<Type::Untracked>();
+        [[maybe_unused]] std::span<u64> cached_words = words.template Span<Type::CachedCPU>();
+        const size_t offset = query_cpu_range - cpu_addr;
+        bool pending = false;
+        size_t pending_offset{};
+        size_t pending_pointer{};
+        const auto release = [&]() {
+            func(cpu_addr + pending_offset * BYTES_PER_PAGE,
+                 (pending_pointer - pending_offset) * BYTES_PER_PAGE);
+        };
+        IterateWords(offset, size, [&](size_t index, u64 mask) {
+            const u64 word = state_words[index] & mask;
+            if constexpr (clear) {
+                if constexpr (type == Type::CPU || type == Type::CachedCPU) {
+                    NotifyRasterizer<true>(index, untracked_words[index], mask);
+                }
+                state_words[index] &= ~mask;
+                if constexpr (type == Type::CPU || type == Type::CachedCPU) {
+                    untracked_words[index] &= ~mask;
+                }
+                if constexpr (type == Type::CPU) {
+                    cached_words[index] &= ~word;
+                }
+            }
+            const size_t base_offset = index * PAGES_PER_WORD;
+            IteratePages(word, [&](size_t pages_offset, size_t pages_size) {
+                const auto reset = [&]() {
+                    pending_offset = base_offset + pages_offset;
+                    pending_pointer = base_offset + pages_offset + pages_size;
+                };
+                if (!pending) {
+                    reset();
+                    pending = true;
+                    return;
+                }
+                if (pending_pointer == base_offset + pages_offset) {
+                    pending_pointer += pages_size;
+                    return;
+                }
+                release();
+                reset();
+            });
+        });
+        if (pending) {
+            release();
+        }
+    }
+
+    /**
+     * Returns true when a region has been modified
+     *
+     * @param offset Offset in bytes from the start of the buffer
+     * @param size   Size in bytes of the region to query for modifications
+     */
+    template <Type type>
+    [[nodiscard]] bool IsRegionModified(u64 offset, u64 size) const noexcept {
+        static_assert(type != Type::Untracked);
+
+        const std::span<const u64> state_words = words.template Span<type>();
+        bool result = false;
+        IterateWords(offset, size, [&](size_t index, u64 mask) {
+            const u64 word = state_words[index] & mask;
+            if (word != 0) {
+                result = true;
+                return true;
+            }
+            return false;
+        });
+        return result;
+    }
+
+    /**
+     * Returns a begin end pair with the inclusive modified region
+     *
+     * @param offset Offset in bytes from the start of the buffer
+     * @param size   Size in bytes of the region to query for modifications
+     */
+    template <Type type>
+    [[nodiscard]] std::pair<u64, u64> ModifiedRegion(u64 offset, u64 size) const noexcept {
+        static_assert(type != Type::Untracked);
+        const std::span<const u64> state_words = words.template Span<type>();
+        u64 begin = std::numeric_limits<u64>::max();
+        u64 end = 0;
+        IterateWords(offset, size, [&](size_t index, u64 mask) {
+            const u64 word = state_words[index] & mask;
+            if (word == 0) {
+                return;
+            }
+            const u64 local_page_begin = std::countr_zero(word);
+            const u64 local_page_end = PAGES_PER_WORD - std::countl_zero(word);
+            const u64 page_index = index * PAGES_PER_WORD;
+            begin = std::min(begin, page_index + local_page_begin);
+            end = page_index + local_page_end;
+        });
+        static constexpr std::pair<u64, u64> EMPTY{0, 0};
+        return begin < end ? std::make_pair(begin * BYTES_PER_PAGE, end * BYTES_PER_PAGE) : EMPTY;
+    }
+
+    /// Returns the number of words of the manager
+    [[nodiscard]] size_t NumWords() const noexcept {
+        return words.NumWords();
+    }
+
+    /// Returns the size in bytes of the manager
+    [[nodiscard]] u64 SizeBytes() const noexcept {
+        return words.size_bytes;
+    }
+
+    /// Returns true when the buffer fits in the small vector optimization
+    [[nodiscard]] bool IsShort() const noexcept {
+        return words.IsShort();
+    }
+
+    void FlushCachedWrites() noexcept {
+        const u64 num_words = NumWords();
+        u64* const cached_words = Array<Type::CachedCPU>();
+        u64* const untracked_words = Array<Type::Untracked>();
+        u64* const cpu_words = Array<Type::CPU>();
+        for (u64 word_index = 0; word_index < num_words; ++word_index) {
+            const u64 cached_bits = cached_words[word_index];
+            NotifyRasterizer<false>(word_index, untracked_words[word_index], cached_bits);
+            untracked_words[word_index] |= cached_bits;
+            cpu_words[word_index] |= cached_bits;
+            cached_words[word_index] = 0;
+        }
+    }
+
+private:
+    template <Type type>
+    u64* Array() noexcept {
+        if constexpr (type == Type::CPU) {
+            return words.cpu.Pointer(IsShort());
+        } else if constexpr (type == Type::GPU) {
+            return words.gpu.Pointer(IsShort());
+        } else if constexpr (type == Type::CachedCPU) {
+            return words.cached_cpu.Pointer(IsShort());
+        } else if constexpr (type == Type::Untracked) {
+            return words.untracked.Pointer(IsShort());
+        }
+    }
+
+    template <Type type>
+    const u64* Array() const noexcept {
+        if constexpr (type == Type::CPU) {
+            return words.cpu.Pointer(IsShort());
+        } else if constexpr (type == Type::GPU) {
+            return words.gpu.Pointer(IsShort());
+        } else if constexpr (type == Type::CachedCPU) {
+            return words.cached_cpu.Pointer(IsShort());
+        } else if constexpr (type == Type::Untracked) {
+            return words.untracked.Pointer(IsShort());
+        }
+    }
+
+    /**
+     * Notify rasterizer about changes in the CPU tracking state of a word in the buffer
+     *
+     * @param word_index   Index to the word to notify to the rasterizer
+     * @param current_bits Current state of the word
+     * @param new_bits     New state of the word
+     *
+     * @tparam add_to_rasterizer True when the rasterizer should start tracking the new pages
+     */
+    template <bool add_to_rasterizer>
+    void NotifyRasterizer(u64 word_index, u64 current_bits, u64 new_bits) const {
+        u64 changed_bits = (add_to_rasterizer ? current_bits : ~current_bits) & new_bits;
+        VAddr addr = cpu_addr + word_index * BYTES_PER_WORD;
+        IteratePages(changed_bits, [&](size_t offset, size_t size) {
+            rasterizer->UpdatePagesCachedCount(addr + offset * BYTES_PER_PAGE,
+                                               size * BYTES_PER_PAGE, add_to_rasterizer ? 1 : -1);
+        });
+    }
+
+    VAddr cpu_addr = 0;
+    RasterizerInterface* rasterizer = nullptr;
+    Words<stack_words> words;
+};
+
+} // namespace VideoCommon

src/video_core/compatible_formats.cpp

@@ -126,15 +126,14 @@ constexpr std::array VIEW_CLASS_ASTC_8x8_RGBA{
     PixelFormat::ASTC_2D_8X8_SRGB,
 };
 
-// Missing formats:
-// PixelFormat::ASTC_2D_10X5_UNORM
-// PixelFormat::ASTC_2D_10X5_SRGB
-
-// Missing formats:
-// PixelFormat::ASTC_2D_10X6_SRGB
+constexpr std::array VIEW_CLASS_ASTC_10x5_RGBA{
+    PixelFormat::ASTC_2D_10X5_UNORM,
+    PixelFormat::ASTC_2D_10X5_SRGB,
+};
 
 constexpr std::array VIEW_CLASS_ASTC_10x6_RGBA{
     PixelFormat::ASTC_2D_10X6_UNORM,
+    PixelFormat::ASTC_2D_10X6_SRGB,
 };
 
 constexpr std::array VIEW_CLASS_ASTC_10x8_RGBA{
@@ -147,9 +146,10 @@ constexpr std::array VIEW_CLASS_ASTC_10x10_RGBA{
     PixelFormat::ASTC_2D_10X10_SRGB,
 };
 
-// Missing formats
-// ASTC_2D_12X10_UNORM,
-// ASTC_2D_12X10_SRGB,
+constexpr std::array VIEW_CLASS_ASTC_12x10_RGBA{
+    PixelFormat::ASTC_2D_12X10_UNORM,
+    PixelFormat::ASTC_2D_12X10_SRGB,
+};
 
 constexpr std::array VIEW_CLASS_ASTC_12x12_RGBA{
     PixelFormat::ASTC_2D_12X12_UNORM,
@@ -229,9 +229,11 @@ constexpr Table MakeViewTable() {
     EnableRange(view, VIEW_CLASS_ASTC_6x6_RGBA);
     EnableRange(view, VIEW_CLASS_ASTC_8x5_RGBA);
     EnableRange(view, VIEW_CLASS_ASTC_8x8_RGBA);
+    EnableRange(view, VIEW_CLASS_ASTC_10x5_RGBA);
     EnableRange(view, VIEW_CLASS_ASTC_10x6_RGBA);
     EnableRange(view, VIEW_CLASS_ASTC_10x8_RGBA);
     EnableRange(view, VIEW_CLASS_ASTC_10x10_RGBA);
+    EnableRange(view, VIEW_CLASS_ASTC_12x10_RGBA);
     EnableRange(view, VIEW_CLASS_ASTC_12x12_RGBA);
     return view;
 }

src/video_core/engines/fermi_2d.cpp

@@ -77,6 +77,14 @@ void Fermi2D::Blit() {
     const auto bytes_per_pixel = BytesPerBlock(PixelFormatFromRenderTargetFormat(src.format));
     const bool delegate_to_gpu = src.width > 512 && src.height > 512 && bytes_per_pixel <= 8 &&
                                  src.format != regs.dst.format;
+
+    auto srcX = args.src_x0;
+    auto srcY = args.src_y0;
+    if (args.sample_mode.origin == Origin::Corner) {
+        srcX -= (args.du_dx >> 33) << 32;
+        srcY -= (args.dv_dy >> 33) << 32;
+    }
+
     Config config{
         .operation = regs.operation,
         .filter = args.sample_mode.filter,
@@ -86,10 +94,10 @@ void Fermi2D::Blit() {
         .dst_y0 = args.dst_y0,
         .dst_x1 = args.dst_x0 + args.dst_width,
         .dst_y1 = args.dst_y0 + args.dst_height,
-        .src_x0 = static_cast<s32>(args.src_x0 >> 32),
-        .src_y0 = static_cast<s32>(args.src_y0 >> 32),
-        .src_x1 = static_cast<s32>((args.du_dx * args.dst_width + args.src_x0) >> 32),
-        .src_y1 = static_cast<s32>((args.dv_dy * args.dst_height + args.src_y0) >> 32),
+        .src_x0 = static_cast<s32>(srcX >> 32),
+        .src_y0 = static_cast<s32>(srcY >> 32),
+        .src_x1 = static_cast<s32>((srcX + args.du_dx * args.dst_width) >> 32),
+        .src_y1 = static_cast<s32>((srcY + args.dv_dy * args.dst_height) >> 32),
     };
 
     const auto need_align_to_pitch =

src/video_core/engines/maxwell_3d.cpp

@@ -4,6 +4,7 @@
 #include <cstring>
 #include <optional>
 #include "common/assert.h"
+#include "common/bit_util.h"
 #include "common/scope_exit.h"
 #include "common/settings.h"
 #include "core/core.h"
@@ -222,6 +223,9 @@ void Maxwell3D::ProcessMacro(u32 method, const u32* base_start, u32 amount, bool
 }
 
 void Maxwell3D::RefreshParametersImpl() {
+    if (!Settings::IsGPULevelHigh()) {
+        return;
+    }
     size_t current_index = 0;
     for (auto& segment : macro_segments) {
         if (segment.first == 0) {
@@ -259,12 +263,13 @@ u32 Maxwell3D::GetMaxCurrentVertices() {
 size_t Maxwell3D::EstimateIndexBufferSize() {
     GPUVAddr start_address = regs.index_buffer.StartAddress();
     GPUVAddr end_address = regs.index_buffer.EndAddress();
-    static constexpr std::array<size_t, 4> max_sizes = {
-        std::numeric_limits<u8>::max(), std::numeric_limits<u16>::max(),
-        std::numeric_limits<u32>::max(), std::numeric_limits<u32>::max()};
+    static constexpr std::array<size_t, 3> max_sizes = {std::numeric_limits<u8>::max(),
+                                                        std::numeric_limits<u16>::max(),
+                                                        std::numeric_limits<u32>::max()};
     const size_t byte_size = regs.index_buffer.FormatSizeInBytes();
+    const size_t log2_byte_size = Common::Log2Ceil64(byte_size);
     return std::min<size_t>(
-        memory_manager.GetMemoryLayoutSize(start_address, byte_size * max_sizes[byte_size]) /
+        memory_manager.GetMemoryLayoutSize(start_address, byte_size * max_sizes[log2_byte_size]) /
             byte_size,
         static_cast<size_t>(end_address - start_address));
 }

src/video_core/fence_manager.h

@@ -4,13 +4,20 @@
 #pragma once
 
 #include <algorithm>
+#include <condition_variable>
 #include <cstring>
 #include <deque>
 #include <functional>
 #include <memory>
+#include <mutex>
+#include <thread>
 #include <queue>
 
 #include "common/common_types.h"
+#include "common/microprofile.h"
+#include "common/scope_exit.h"
+#include "common/settings.h"
+#include "common/thread.h"
 #include "video_core/delayed_destruction_ring.h"
 #include "video_core/gpu.h"
 #include "video_core/host1x/host1x.h"
@@ -23,15 +30,26 @@ class FenceBase {
 public:
     explicit FenceBase(bool is_stubbed_) : is_stubbed{is_stubbed_} {}
 
+    bool IsStubbed() const {
+        return is_stubbed;
+    }
+
 protected:
     bool is_stubbed;
 };
 
-template <typename TFence, typename TTextureCache, typename TTBufferCache, typename TQueryCache>
+template <typename Traits>
 class FenceManager {
+    using TFence = typename Traits::FenceType;
+    using TTextureCache = typename Traits::TextureCacheType;
+    using TBufferCache = typename Traits::BufferCacheType;
+    using TQueryCache = typename Traits::QueryCacheType;
+    static constexpr bool can_async_check = Traits::HAS_ASYNC_CHECK;
+
 public:
     /// Notify the fence manager about a new frame
     void TickFrame() {
+        std::unique_lock lock(ring_guard);
         delayed_destruction_ring.Tick();
     }
 
@@ -46,17 +64,33 @@ public:
     }
 
     void SignalFence(std::function<void()>&& func) {
-        TryReleasePendingFences();
+        rasterizer.InvalidateGPUCache();
+        bool delay_fence = Settings::IsGPULevelHigh();
+        if constexpr (!can_async_check) {
+            TryReleasePendingFences<false>();
+        }
         const bool should_flush = ShouldFlush();
         CommitAsyncFlushes();
-        uncommitted_operations.emplace_back(std::move(func));
-        CommitOperations();
         TFence new_fence = CreateFence(!should_flush);
-        fences.push(new_fence);
+        if constexpr (can_async_check) {
+            guard.lock();
+        }
+        if (delay_fence) {
+            uncommitted_operations.emplace_back(std::move(func));
+        }
+        pending_operations.emplace_back(std::move(uncommitted_operations));
         QueueFence(new_fence);
+        if (!delay_fence) {
+            func();
+        }
+        fences.push(std::move(new_fence));
         if (should_flush) {
             rasterizer.FlushCommands();
         }
+        if constexpr (can_async_check) {
+            guard.unlock();
+            cv.notify_all();
+        }
     }
 
     void SignalSyncPoint(u32 value) {
@@ -66,29 +100,30 @@ public:
     }
 
     void WaitPendingFences() {
-        while (!fences.empty()) {
-            TFence& current_fence = fences.front();
-            if (ShouldWait()) {
-                WaitFence(current_fence);
-            }
-            PopAsyncFlushes();
-            auto operations = std::move(pending_operations.front());
-            pending_operations.pop_front();
-            for (auto& operation : operations) {
-                operation();
-            }
-            PopFence();
+        if constexpr (!can_async_check) {
+            TryReleasePendingFences<true>();
         }
     }
 
 protected:
     explicit FenceManager(VideoCore::RasterizerInterface& rasterizer_, Tegra::GPU& gpu_,
-                          TTextureCache& texture_cache_, TTBufferCache& buffer_cache_,
+                          TTextureCache& texture_cache_, TBufferCache& buffer_cache_,
                           TQueryCache& query_cache_)
         : rasterizer{rasterizer_}, gpu{gpu_}, syncpoint_manager{gpu.Host1x().GetSyncpointManager()},
-          texture_cache{texture_cache_}, buffer_cache{buffer_cache_}, query_cache{query_cache_} {}
+          texture_cache{texture_cache_}, buffer_cache{buffer_cache_}, query_cache{query_cache_} {
+        if constexpr (can_async_check) {
+            fence_thread =
+                std::jthread([this](std::stop_token token) { ReleaseThreadFunc(token); });
+        }
+    }
 
-    virtual ~FenceManager() = default;
+    virtual ~FenceManager() {
+        if constexpr (can_async_check) {
+            fence_thread.request_stop();
+            cv.notify_all();
+            fence_thread.join();
+        }
+    }
 
     /// Creates a Fence Interface, does not create a backend fence if 'is_stubbed' is
     /// true
@@ -104,15 +139,20 @@ protected:
     Tegra::GPU& gpu;
     Tegra::Host1x::SyncpointManager& syncpoint_manager;
     TTextureCache& texture_cache;
-    TTBufferCache& buffer_cache;
+    TBufferCache& buffer_cache;
     TQueryCache& query_cache;
 
 private:
+    template <bool force_wait>
     void TryReleasePendingFences() {
         while (!fences.empty()) {
             TFence& current_fence = fences.front();
             if (ShouldWait() && !IsFenceSignaled(current_fence)) {
-                return;
+                if constexpr (force_wait) {
+                    WaitFence(current_fence);
+                } else {
+                    return;
+                }
             }
             PopAsyncFlushes();
             auto operations = std::move(pending_operations.front());
@@ -120,7 +160,49 @@ private:
             for (auto& operation : operations) {
                 operation();
             }
-            PopFence();
+            {
+                std::unique_lock lock(ring_guard);
+                delayed_destruction_ring.Push(std::move(current_fence));
+            }
+            fences.pop();
+        }
+    }
+
+    void ReleaseThreadFunc(std::stop_token stop_token) {
+        std::string name = "GPUFencingThread";
+        MicroProfileOnThreadCreate(name.c_str());
+
+        // Cleanup
+        SCOPE_EXIT({ MicroProfileOnThreadExit(); });
+
+        Common::SetCurrentThreadName(name.c_str());
+        Common::SetCurrentThreadPriority(Common::ThreadPriority::High);
+
+        TFence current_fence;
+        std::deque<std::function<void()>> current_operations;
+        while (!stop_token.stop_requested()) {
+            {
+                std::unique_lock lock(guard);
+                cv.wait(lock, [&] { return stop_token.stop_requested() || !fences.empty(); });
+                if (stop_token.stop_requested()) [[unlikely]] {
+                    return;
+                }
+                current_fence = std::move(fences.front());
+                current_operations = std::move(pending_operations.front());
+                fences.pop();
+                pending_operations.pop_front();
+            }
+            if (!current_fence->IsStubbed()) {
+                WaitFence(current_fence);
+            }
+            PopAsyncFlushes();
+            for (auto& operation : current_operations) {
+                operation();
+            }
+            {
+                std::unique_lock lock(ring_guard);
+                delayed_destruction_ring.Push(std::move(current_fence));
+            }
         }
     }
 
@@ -154,19 +236,16 @@ private:
         query_cache.CommitAsyncFlushes();
     }
 
-    void PopFence() {
-        delayed_destruction_ring.Push(std::move(fences.front()));
-        fences.pop();
-    }
-
-    void CommitOperations() {
-        pending_operations.emplace_back(std::move(uncommitted_operations));
-    }
-
     std::queue<TFence> fences;
     std::deque<std::function<void()>> uncommitted_operations;
     std::deque<std::deque<std::function<void()>>> pending_operations;
 
+    std::mutex guard;
+    std::mutex ring_guard;
+    std::condition_variable cv;
+
+    std::jthread fence_thread;
+
     DelayedDestructionRing<TFence, 6> delayed_destruction_ring;
 };
 

src/video_core/memory_manager.cpp

@@ -82,6 +82,7 @@ void MemoryManager::SetEntry(size_t position, MemoryManager::EntryType entry) {
 }
 
 PTEKind MemoryManager::GetPageKind(GPUVAddr gpu_addr) const {
+    std::unique_lock<std::mutex> lock(guard);
     return kind_map.GetValueAt(gpu_addr);
 }
 
@@ -160,7 +161,10 @@ GPUVAddr MemoryManager::BigPageTableOp(GPUVAddr gpu_addr, [[maybe_unused]] VAddr
         }
         remaining_size -= big_page_size;
     }
-    kind_map.Map(gpu_addr, gpu_addr + size, kind);
+    {
+        std::unique_lock<std::mutex> lock(guard);
+        kind_map.Map(gpu_addr, gpu_addr + size, kind);
+    }
     return gpu_addr;
 }
 
@@ -553,6 +557,7 @@ size_t MemoryManager::MaxContinuousRange(GPUVAddr gpu_addr, size_t size) const {
 }
 
 size_t MemoryManager::GetMemoryLayoutSize(GPUVAddr gpu_addr, size_t max_size) const {
+    std::unique_lock<std::mutex> lock(guard);
     return kind_map.GetContinuousSizeFrom(gpu_addr);
 }
 
@@ -745,10 +750,10 @@ void MemoryManager::FlushCaching() {
         return;
     }
     accumulator->Callback([this](GPUVAddr addr, size_t size) {
-        GetSubmappedRangeImpl<false>(addr, size, page_stash);
+        GetSubmappedRangeImpl<false>(addr, size, page_stash2);
     });
-    rasterizer->InnerInvalidation(page_stash);
-    page_stash.clear();
+    rasterizer->InnerInvalidation(page_stash2);
+    page_stash2.clear();
     accumulator->Clear();
 }
 

src/video_core/memory_manager.h

@@ -5,6 +5,7 @@
 
 #include <atomic>
 #include <map>
+#include <mutex>
 #include <optional>
 #include <vector>
 
@@ -215,6 +216,9 @@ private:
 
     std::vector<u64> big_page_continuous;
     std::vector<std::pair<VAddr, std::size_t>> page_stash{};
+    std::vector<std::pair<VAddr, std::size_t>> page_stash2{};
+
+    mutable std::mutex guard;
 
     static constexpr size_t continuous_bits = 64;
 

src/video_core/query_cache.h

@@ -6,6 +6,7 @@
 #include <algorithm>
 #include <array>
 #include <cstring>
+#include <functional>
 #include <iterator>
 #include <list>
 #include <memory>
@@ -17,13 +18,19 @@
 
 #include "common/assert.h"
 #include "common/settings.h"
+#include "core/memory.h"
 #include "video_core/control/channel_state_cache.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/memory_manager.h"
 #include "video_core/rasterizer_interface.h"
+#include "video_core/texture_cache/slot_vector.h"
 
 namespace VideoCommon {
 
+using AsyncJobId = SlotId;
+
+static constexpr AsyncJobId NULL_ASYNC_JOB_ID{0};
+
 template <class QueryCache, class HostCounter>
 class CounterStreamBase {
 public:
@@ -93,9 +100,13 @@ private:
 template <class QueryCache, class CachedQuery, class CounterStream, class HostCounter>
 class QueryCacheBase : public VideoCommon::ChannelSetupCaches<VideoCommon::ChannelInfo> {
 public:
-    explicit QueryCacheBase(VideoCore::RasterizerInterface& rasterizer_)
-        : rasterizer{rasterizer_}, streams{{CounterStream{static_cast<QueryCache&>(*this),
-                                                          VideoCore::QueryType::SamplesPassed}}} {}
+    explicit QueryCacheBase(VideoCore::RasterizerInterface& rasterizer_,
+                            Core::Memory::Memory& cpu_memory_)
+        : rasterizer{rasterizer_},
+          cpu_memory{cpu_memory_}, streams{{CounterStream{static_cast<QueryCache&>(*this),
+                                                          VideoCore::QueryType::SamplesPassed}}} {
+        (void)slot_async_jobs.insert(); // Null value
+    }
 
     void InvalidateRegion(VAddr addr, std::size_t size) {
         std::unique_lock lock{mutex};
@@ -126,10 +137,15 @@ public:
             query = Register(type, *cpu_addr, host_ptr, timestamp.has_value());
         }
 
-        query->BindCounter(Stream(type).Current(), timestamp);
-        if (Settings::values.use_asynchronous_gpu_emulation.GetValue()) {
-            AsyncFlushQuery(*cpu_addr);
+        auto result = query->BindCounter(Stream(type).Current(), timestamp);
+        if (result) {
+            auto async_job_id = query->GetAsyncJob();
+            auto& async_job = slot_async_jobs[async_job_id];
+            async_job.collected = true;
+            async_job.value = *result;
+            query->SetAsyncJob(NULL_ASYNC_JOB_ID);
         }
+        AsyncFlushQuery(query, timestamp, lock);
     }
 
     /// Updates counters from GPU state. Expected to be called once per draw, clear or dispatch.
@@ -173,15 +189,18 @@ public:
     }
 
     void CommitAsyncFlushes() {
+        std::unique_lock lock{mutex};
         committed_flushes.push_back(uncommitted_flushes);
         uncommitted_flushes.reset();
     }
 
     bool HasUncommittedFlushes() const {
+        std::unique_lock lock{mutex};
         return uncommitted_flushes != nullptr;
     }
 
     bool ShouldWaitAsyncFlushes() const {
+        std::unique_lock lock{mutex};
         if (committed_flushes.empty()) {
             return false;
         }
@@ -189,6 +208,7 @@ public:
     }
 
     void PopAsyncFlushes() {
+        std::unique_lock lock{mutex};
         if (committed_flushes.empty()) {
             return;
         }
@@ -197,15 +217,25 @@ public:
             committed_flushes.pop_front();
             return;
         }
-        for (VAddr query_address : *flush_list) {
-            FlushAndRemoveRegion(query_address, 4);
+        for (AsyncJobId async_job_id : *flush_list) {
+            AsyncJob& async_job = slot_async_jobs[async_job_id];
+            if (!async_job.collected) {
+                FlushAndRemoveRegion(async_job.query_location, 2, true);
+            }
         }
         committed_flushes.pop_front();
     }
 
 private:
+    struct AsyncJob {
+        bool collected = false;
+        u64 value = 0;
+        VAddr query_location = 0;
+        std::optional<u64> timestamp{};
+    };
+
     /// Flushes a memory range to guest memory and removes it from the cache.
-    void FlushAndRemoveRegion(VAddr addr, std::size_t size) {
+    void FlushAndRemoveRegion(VAddr addr, std::size_t size, bool async = false) {
         const u64 addr_begin = addr;
         const u64 addr_end = addr_begin + size;
         const auto in_range = [addr_begin, addr_end](const CachedQuery& query) {
@@ -226,7 +256,16 @@ private:
                     continue;
                 }
                 rasterizer.UpdatePagesCachedCount(query.GetCpuAddr(), query.SizeInBytes(), -1);
-                query.Flush();
+                AsyncJobId async_job_id = query.GetAsyncJob();
+                auto flush_result = query.Flush(async);
+                if (async_job_id == NULL_ASYNC_JOB_ID) {
+                    ASSERT_MSG(false, "This should not be reachable at all");
+                    continue;
+                }
+                AsyncJob& async_job = slot_async_jobs[async_job_id];
+                async_job.collected = true;
+                async_job.value = flush_result;
+                query.SetAsyncJob(NULL_ASYNC_JOB_ID);
             }
             std::erase_if(contents, in_range);
         }
@@ -253,26 +292,60 @@ private:
         return found != std::end(contents) ? &*found : nullptr;
     }
 
-    void AsyncFlushQuery(VAddr addr) {
-        if (!uncommitted_flushes) {
-            uncommitted_flushes = std::make_shared<std::vector<VAddr>>();
+    void AsyncFlushQuery(CachedQuery* query, std::optional<u64> timestamp,
+                         std::unique_lock<std::recursive_mutex>& lock) {
+        const AsyncJobId new_async_job_id = slot_async_jobs.insert();
+        {
+            AsyncJob& async_job = slot_async_jobs[new_async_job_id];
+            query->SetAsyncJob(new_async_job_id);
+            async_job.query_location = query->GetCpuAddr();
+            async_job.collected = false;
+
+            if (!uncommitted_flushes) {
+                uncommitted_flushes = std::make_shared<std::vector<AsyncJobId>>();
+            }
+            uncommitted_flushes->push_back(new_async_job_id);
         }
-        uncommitted_flushes->push_back(addr);
+        lock.unlock();
+        std::function<void()> operation([this, new_async_job_id, timestamp] {
+            std::unique_lock local_lock{mutex};
+            AsyncJob& async_job = slot_async_jobs[new_async_job_id];
+            u64 value = async_job.value;
+            VAddr address = async_job.query_location;
+            slot_async_jobs.erase(new_async_job_id);
+            local_lock.unlock();
+            if (timestamp) {
+                u64 timestamp_value = *timestamp;
+                cpu_memory.WriteBlockUnsafe(address + sizeof(u64), &timestamp_value, sizeof(u64));
+                cpu_memory.WriteBlockUnsafe(address, &value, sizeof(u64));
+                rasterizer.InvalidateRegion(address, sizeof(u64) * 2,
+                                            VideoCommon::CacheType::NoQueryCache);
+            } else {
+                u32 small_value = static_cast<u32>(value);
+                cpu_memory.WriteBlockUnsafe(address, &small_value, sizeof(u32));
+                rasterizer.InvalidateRegion(address, sizeof(u32),
+                                            VideoCommon::CacheType::NoQueryCache);
+            }
+        });
+        rasterizer.SyncOperation(std::move(operation));
     }
 
     static constexpr std::uintptr_t YUZU_PAGESIZE = 4096;
     static constexpr unsigned YUZU_PAGEBITS = 12;
 
-    VideoCore::RasterizerInterface& rasterizer;
+    SlotVector<AsyncJob> slot_async_jobs;
 
-    std::recursive_mutex mutex;
+    VideoCore::RasterizerInterface& rasterizer;
+    Core::Memory::Memory& cpu_memory;
+
+    mutable std::recursive_mutex mutex;
 
     std::unordered_map<u64, std::vector<CachedQuery>> cached_queries;
 
     std::array<CounterStream, VideoCore::NumQueryTypes> streams;
 
-    std::shared_ptr<std::vector<VAddr>> uncommitted_flushes{};
-    std::list<std::shared_ptr<std::vector<VAddr>>> committed_flushes;
+    std::shared_ptr<std::vector<AsyncJobId>> uncommitted_flushes{};
+    std::list<std::shared_ptr<std::vector<AsyncJobId>>> committed_flushes;
 };
 
 template <class QueryCache, class HostCounter>
@@ -291,12 +364,12 @@ public:
     virtual ~HostCounterBase() = default;
 
     /// Returns the current value of the query.
-    u64 Query() {
+    u64 Query(bool async = false) {
         if (result) {
             return *result;
         }
 
-        u64 value = BlockingQuery() + base_result;
+        u64 value = BlockingQuery(async) + base_result;
         if (dependency) {
             value += dependency->Query();
             dependency = nullptr;
@@ -317,7 +390,7 @@ public:
 
 protected:
     /// Returns the value of query from the backend API blocking as needed.
-    virtual u64 BlockingQuery() const = 0;
+    virtual u64 BlockingQuery(bool async = false) const = 0;
 
 private:
     std::shared_ptr<HostCounter> dependency; ///< Counter to add to this value.
@@ -340,26 +413,33 @@ public:
     CachedQueryBase& operator=(const CachedQueryBase&) = delete;
 
     /// Flushes the query to guest memory.
-    virtual void Flush() {
+    virtual u64 Flush(bool async = false) {
         // When counter is nullptr it means that it's just been reset. We are supposed to write a
         // zero in these cases.
-        const u64 value = counter ? counter->Query() : 0;
+        const u64 value = counter ? counter->Query(async) : 0;
+        if (async) {
+            return value;
+        }
         std::memcpy(host_ptr, &value, sizeof(u64));
 
         if (timestamp) {
             std::memcpy(host_ptr + TIMESTAMP_OFFSET, &*timestamp, sizeof(u64));
         }
+        return value;
     }
 
     /// Binds a counter to this query.
-    void BindCounter(std::shared_ptr<HostCounter> counter_, std::optional<u64> timestamp_) {
+    std::optional<u64> BindCounter(std::shared_ptr<HostCounter> counter_,
+                                   std::optional<u64> timestamp_) {
+        std::optional<u64> result{};
         if (counter) {
             // If there's an old counter set it means the query is being rewritten by the game.
             // To avoid losing the data forever, flush here.
-            Flush();
+            result = std::make_optional(Flush());
         }
         counter = std::move(counter_);
         timestamp = timestamp_;
+        return result;
     }
 
     VAddr GetCpuAddr() const noexcept {
@@ -374,6 +454,14 @@ public:
         return with_timestamp ? LARGE_QUERY_SIZE : SMALL_QUERY_SIZE;
     }
 
+    void SetAsyncJob(AsyncJobId assigned_async_job_) {
+        assigned_async_job = assigned_async_job_;
+    }
+
+    AsyncJobId GetAsyncJob() const {
+        return assigned_async_job;
+    }
+
 protected:
     /// Returns true when querying the counter may potentially block.
     bool WaitPending() const noexcept {
@@ -389,6 +477,7 @@ private:
     u8* host_ptr;                         ///< Writable host pointer.
     std::shared_ptr<HostCounter> counter; ///< Host counter to query, owns the dependency tree.
     std::optional<u64> timestamp;         ///< Timestamp to flush to guest memory.
+    AsyncJobId assigned_async_job;
 };
 
 } // namespace VideoCommon

src/video_core/renderer_opengl/gl_buffer_cache.h

@@ -8,6 +8,7 @@
 
 #include "common/common_types.h"
 #include "video_core/buffer_cache/buffer_cache.h"
+#include "video_core/buffer_cache/memory_tracker_base.h"
 #include "video_core/rasterizer_interface.h"
 #include "video_core/renderer_opengl/gl_device.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
@@ -200,6 +201,8 @@ private:
 struct BufferCacheParams {
     using Runtime = OpenGL::BufferCacheRuntime;
     using Buffer = OpenGL::Buffer;
+    using Async_Buffer = u32;
+    using MemoryTracker = VideoCommon::MemoryTrackerBase<VideoCore::RasterizerInterface>;
 
     static constexpr bool IS_OPENGL = true;
     static constexpr bool HAS_PERSISTENT_UNIFORM_BUFFER_BINDINGS = true;
@@ -208,6 +211,7 @@ struct BufferCacheParams {
     static constexpr bool NEEDS_BIND_STORAGE_INDEX = true;
     static constexpr bool USE_MEMORY_MAPS = false;
     static constexpr bool SEPARATE_IMAGE_BUFFER_BINDINGS = true;
+    static constexpr bool IMPLEMENTS_ASYNC_DOWNLOADS = false;
 };
 
 using BufferCache = VideoCommon::BufferCache<BufferCacheParams>;

src/video_core/renderer_opengl/gl_buffer_cache_base.cpp

@@ -0,0 +1,9 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#include "video_core/buffer_cache/buffer_cache.h"
+#include "video_core/renderer_opengl/gl_buffer_cache.h"
+
+namespace VideoCommon {
+template class VideoCommon::BufferCache<OpenGL::BufferCacheParams>;
+}

src/video_core/renderer_opengl/gl_fence_manager.h

@@ -30,7 +30,17 @@ private:
 };
 
 using Fence = std::shared_ptr<GLInnerFence>;
-using GenericFenceManager = VideoCommon::FenceManager<Fence, TextureCache, BufferCache, QueryCache>;
+
+struct FenceManagerParams {
+    using FenceType = Fence;
+    using BufferCacheType = BufferCache;
+    using TextureCacheType = TextureCache;
+    using QueryCacheType = QueryCache;
+
+    static constexpr bool HAS_ASYNC_CHECK = false;
+};
+
+using GenericFenceManager = VideoCommon::FenceManager<FenceManagerParams>;
 
 class FenceManagerOpenGL final : public GenericFenceManager {
 public:

src/video_core/renderer_opengl/gl_query_cache.cpp

@@ -26,8 +26,8 @@ constexpr GLenum GetTarget(VideoCore::QueryType type) {
 
 } // Anonymous namespace
 
-QueryCache::QueryCache(RasterizerOpenGL& rasterizer_)
-    : QueryCacheBase(rasterizer_), gl_rasterizer{rasterizer_} {}
+QueryCache::QueryCache(RasterizerOpenGL& rasterizer_, Core::Memory::Memory& cpu_memory_)
+    : QueryCacheBase(rasterizer_, cpu_memory_), gl_rasterizer{rasterizer_} {}
 
 QueryCache::~QueryCache() = default;
 
@@ -74,7 +74,7 @@ void HostCounter::EndQuery() {
     glEndQuery(GetTarget(type));
 }
 
-u64 HostCounter::BlockingQuery() const {
+u64 HostCounter::BlockingQuery([[maybe_unused]] bool async) const {
     GLint64 value;
     glGetQueryObjecti64v(query.handle, GL_QUERY_RESULT, &value);
     return static_cast<u64>(value);
@@ -96,7 +96,7 @@ CachedQuery& CachedQuery::operator=(CachedQuery&& rhs) noexcept {
     return *this;
 }
 
-void CachedQuery::Flush() {
+u64 CachedQuery::Flush([[maybe_unused]] bool async) {
     // Waiting for a query while another query of the same target is enabled locks Nvidia's driver.
     // To avoid this disable and re-enable keeping the dependency stream.
     // But we only have to do this if we have pending waits to be done.
@@ -106,11 +106,13 @@ void CachedQuery::Flush() {
         stream.Update(false);
     }
 
-    VideoCommon::CachedQueryBase<HostCounter>::Flush();
+    auto result = VideoCommon::CachedQueryBase<HostCounter>::Flush();
 
     if (slice_counter) {
         stream.Update(true);
     }
+
+    return result;
 }
 
 } // namespace OpenGL

src/video_core/renderer_opengl/gl_query_cache.h

@@ -28,7 +28,7 @@ using CounterStream = VideoCommon::CounterStreamBase<QueryCache, HostCounter>;
 class QueryCache final
     : public VideoCommon::QueryCacheBase<QueryCache, CachedQuery, CounterStream, HostCounter> {
 public:
-    explicit QueryCache(RasterizerOpenGL& rasterizer_);
+    explicit QueryCache(RasterizerOpenGL& rasterizer_, Core::Memory::Memory& cpu_memory_);
     ~QueryCache();
 
     OGLQuery AllocateQuery(VideoCore::QueryType type);
@@ -51,7 +51,7 @@ public:
     void EndQuery();
 
 private:
-    u64 BlockingQuery() const override;
+    u64 BlockingQuery(bool async = false) const override;
 
     QueryCache& cache;
     const VideoCore::QueryType type;
@@ -70,7 +70,7 @@ public:
     CachedQuery(const CachedQuery&) = delete;
     CachedQuery& operator=(const CachedQuery&) = delete;
 
-    void Flush() override;
+    u64 Flush(bool async = false) override;
 
 private:
     QueryCache* cache;

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -63,7 +63,7 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window_, Tegra
       buffer_cache(*this, cpu_memory_, buffer_cache_runtime),
       shader_cache(*this, emu_window_, device, texture_cache, buffer_cache, program_manager,
                    state_tracker, gpu.ShaderNotify()),
-      query_cache(*this), accelerate_dma(buffer_cache, texture_cache),
+      query_cache(*this, cpu_memory_), accelerate_dma(buffer_cache, texture_cache),
       fence_manager(*this, gpu, texture_cache, buffer_cache, query_cache),
       blit_image(program_manager_) {}
 
@@ -1287,8 +1287,7 @@ bool AccelerateDMA::DmaBufferImageCopy(const Tegra::DMA::ImageCopy& copy_info,
     }
     const u32 buffer_size = static_cast<u32>(buffer_operand.pitch * buffer_operand.height);
     static constexpr auto sync_info = VideoCommon::ObtainBufferSynchronize::FullSynchronize;
-    const auto post_op = IS_IMAGE_UPLOAD ? VideoCommon::ObtainBufferOperation::DoNothing
-                                         : VideoCommon::ObtainBufferOperation::MarkAsWritten;
+    const auto post_op = VideoCommon::ObtainBufferOperation::DoNothing;
     const auto [buffer, offset] =
         buffer_cache.ObtainBuffer(buffer_operand.address, buffer_size, sync_info, post_op);
 
@@ -1299,7 +1298,8 @@ bool AccelerateDMA::DmaBufferImageCopy(const Tegra::DMA::ImageCopy& copy_info,
     if constexpr (IS_IMAGE_UPLOAD) {
         image->UploadMemory(buffer->Handle(), offset, copy_span);
     } else {
-        image->DownloadMemory(buffer->Handle(), offset, copy_span);
+        texture_cache.DownloadImageIntoBuffer(image, buffer->Handle(), offset, copy_span,
+                                              buffer_operand.address, buffer_size);
     }
     return true;
 }

src/video_core/renderer_opengl/gl_texture_cache.cpp

@@ -803,30 +803,40 @@ void Image::UploadMemory(const ImageBufferMap& map,
 
 void Image::DownloadMemory(GLuint buffer_handle, size_t buffer_offset,
                            std::span<const VideoCommon::BufferImageCopy> copies) {
+    std::array buffer_handles{buffer_handle};
+    std::array buffer_offsets{buffer_offset};
+    DownloadMemory(buffer_handles, buffer_offsets, copies);
+}
+
+void Image::DownloadMemory(std::span<GLuint> buffer_handles, std::span<size_t> buffer_offsets,
+                           std::span<const VideoCommon::BufferImageCopy> copies) {
     const bool is_rescaled = True(flags & ImageFlagBits::Rescaled);
     if (is_rescaled) {
         ScaleDown();
     }
     glMemoryBarrier(GL_PIXEL_BUFFER_BARRIER_BIT); // TODO: Move this to its own API
-    glBindBuffer(GL_PIXEL_PACK_BUFFER, buffer_handle);
-    glPixelStorei(GL_PACK_ALIGNMENT, 1);
+    for (size_t i = 0; i < buffer_handles.size(); i++) {
+        auto& buffer_handle = buffer_handles[i];
+        glBindBuffer(GL_PIXEL_PACK_BUFFER, buffer_handle);
+        glPixelStorei(GL_PACK_ALIGNMENT, 1);
 
-    u32 current_row_length = std::numeric_limits<u32>::max();
-    u32 current_image_height = std::numeric_limits<u32>::max();
+        u32 current_row_length = std::numeric_limits<u32>::max();
+        u32 current_image_height = std::numeric_limits<u32>::max();
 
-    for (const VideoCommon::BufferImageCopy& copy : copies) {
-        if (copy.image_subresource.base_level >= gl_num_levels) {
-            continue;
+        for (const VideoCommon::BufferImageCopy& copy : copies) {
+            if (copy.image_subresource.base_level >= gl_num_levels) {
+                continue;
+            }
+            if (current_row_length != copy.buffer_row_length) {
+                current_row_length = copy.buffer_row_length;
+                glPixelStorei(GL_PACK_ROW_LENGTH, current_row_length);
+            }
+            if (current_image_height != copy.buffer_image_height) {
+                current_image_height = copy.buffer_image_height;
+                glPixelStorei(GL_PACK_IMAGE_HEIGHT, current_image_height);
+            }
+            CopyImageToBuffer(copy, buffer_offsets[i]);
         }
-        if (current_row_length != copy.buffer_row_length) {
-            current_row_length = copy.buffer_row_length;
-            glPixelStorei(GL_PACK_ROW_LENGTH, current_row_length);
-        }
-        if (current_image_height != copy.buffer_image_height) {
-            current_image_height = copy.buffer_image_height;
-            glPixelStorei(GL_PACK_IMAGE_HEIGHT, current_image_height);
-        }
-        CopyImageToBuffer(copy, buffer_offset);
     }
     if (is_rescaled) {
         ScaleUp(true);
@@ -851,9 +861,12 @@ GLuint Image::StorageHandle() noexcept {
     case PixelFormat::ASTC_2D_8X5_SRGB:
     case PixelFormat::ASTC_2D_5X4_SRGB:
     case PixelFormat::ASTC_2D_5X5_SRGB:
+    case PixelFormat::ASTC_2D_10X5_SRGB:
+    case PixelFormat::ASTC_2D_10X6_SRGB:
     case PixelFormat::ASTC_2D_10X8_SRGB:
     case PixelFormat::ASTC_2D_6X6_SRGB:
     case PixelFormat::ASTC_2D_10X10_SRGB:
+    case PixelFormat::ASTC_2D_12X10_SRGB:
     case PixelFormat::ASTC_2D_12X12_SRGB:
     case PixelFormat::ASTC_2D_8X6_SRGB:
     case PixelFormat::ASTC_2D_6X5_SRGB:

src/video_core/renderer_opengl/gl_texture_cache.h

@@ -215,6 +215,9 @@ public:
     void DownloadMemory(GLuint buffer_handle, size_t buffer_offset,
                         std::span<const VideoCommon::BufferImageCopy> copies);
 
+    void DownloadMemory(std::span<GLuint> buffer_handle, std::span<size_t> buffer_offset,
+                        std::span<const VideoCommon::BufferImageCopy> copies);
+
     void DownloadMemory(ImageBufferMap& map, std::span<const VideoCommon::BufferImageCopy> copies);
 
     GLuint StorageHandle() noexcept;
@@ -376,6 +379,7 @@ struct TextureCacheParams {
     using Sampler = OpenGL::Sampler;
     using Framebuffer = OpenGL::Framebuffer;
     using AsyncBuffer = u32;
+    using BufferType = GLuint;
 };
 
 using TextureCache = VideoCommon::TextureCache<TextureCacheParams>;

src/video_core/renderer_opengl/maxwell_to_gl.h

@@ -100,10 +100,13 @@ constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> FORMAT_TAB
     {GL_COMPRESSED_RGBA_ASTC_6x6_KHR},                                // ASTC_2D_6X6_UNORM
     {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x6_KHR},                        // ASTC_2D_6X6_SRGB
     {GL_COMPRESSED_RGBA_ASTC_10x6_KHR},                               // ASTC_2D_10X6_UNORM
+    {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x6_KHR},                       // ASTC_2D_10X6_SRGB
     {GL_COMPRESSED_RGBA_ASTC_10x5_KHR},                               // ASTC_2D_10X5_UNORM
     {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x5_KHR},                       // ASTC_2D_10X5_SRGB
     {GL_COMPRESSED_RGBA_ASTC_10x10_KHR},                              // ASTC_2D_10X10_UNORM
     {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x10_KHR},                      // ASTC_2D_10X10_SRGB
+    {GL_COMPRESSED_RGBA_ASTC_12x10_KHR},                              // ASTC_2D_12X10_UNORM
+    {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x10_KHR},                      // ASTC_2D_12X10_SRGB
     {GL_COMPRESSED_RGBA_ASTC_12x12_KHR},                              // ASTC_2D_12X12_UNORM
     {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x12_KHR},                      // ASTC_2D_12X12_SRGB
     {GL_COMPRESSED_RGBA_ASTC_8x6_KHR},                                // ASTC_2D_8X6_UNORM

src/video_core/renderer_vulkan/maxwell_to_vk.cpp

@@ -197,10 +197,13 @@ struct FormatTuple {
     {VK_FORMAT_ASTC_6x6_UNORM_BLOCK},                          // ASTC_2D_6X6_UNORM
     {VK_FORMAT_ASTC_6x6_SRGB_BLOCK},                           // ASTC_2D_6X6_SRGB
     {VK_FORMAT_ASTC_10x6_UNORM_BLOCK},                         // ASTC_2D_10X6_UNORM
+    {VK_FORMAT_ASTC_10x6_SRGB_BLOCK},                          // ASTC_2D_10X6_SRGB
     {VK_FORMAT_ASTC_10x5_UNORM_BLOCK},                         // ASTC_2D_10X5_UNORM
     {VK_FORMAT_ASTC_10x5_SRGB_BLOCK},                          // ASTC_2D_10X5_SRGB
     {VK_FORMAT_ASTC_10x10_UNORM_BLOCK},                        // ASTC_2D_10X10_UNORM
     {VK_FORMAT_ASTC_10x10_SRGB_BLOCK},                         // ASTC_2D_10X10_SRGB
+    {VK_FORMAT_ASTC_12x10_UNORM_BLOCK},                        // ASTC_2D_12X10_UNORM
+    {VK_FORMAT_ASTC_12x10_SRGB_BLOCK},                         // ASTC_2D_12X10_SRGB
     {VK_FORMAT_ASTC_12x12_UNORM_BLOCK},                        // ASTC_2D_12X12_UNORM
     {VK_FORMAT_ASTC_12x12_SRGB_BLOCK},                         // ASTC_2D_12X12_SRGB
     {VK_FORMAT_ASTC_8x6_UNORM_BLOCK},                          // ASTC_2D_8X6_UNORM

src/video_core/renderer_vulkan/renderer_vulkan.cpp

@@ -93,8 +93,9 @@ RendererVulkan::RendererVulkan(Core::TelemetrySession& telemetry_session_,
       state_tracker(), scheduler(device, state_tracker),
       swapchain(*surface, device, scheduler, render_window.GetFramebufferLayout().width,
                 render_window.GetFramebufferLayout().height, false),
-      blit_screen(cpu_memory, render_window, device, memory_allocator, swapchain, scheduler,
-                  screen_info),
+      present_manager(render_window, device, memory_allocator, scheduler, swapchain),
+      blit_screen(cpu_memory, render_window, device, memory_allocator, swapchain, present_manager,
+                  scheduler, screen_info),
       rasterizer(render_window, gpu, cpu_memory, screen_info, device, memory_allocator,
                  state_tracker, scheduler) {
     if (Settings::values.renderer_force_max_clock.GetValue() && device.ShouldBoostClocks()) {
@@ -121,46 +122,19 @@ void RendererVulkan::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
         return;
     }
     // Update screen info if the framebuffer size has changed.
-    if (screen_info.width != framebuffer->width || screen_info.height != framebuffer->height) {
-        screen_info.width = framebuffer->width;
-        screen_info.height = framebuffer->height;
-    }
+    screen_info.width = framebuffer->width;
+    screen_info.height = framebuffer->height;
+
     const VAddr framebuffer_addr = framebuffer->address + framebuffer->offset;
     const bool use_accelerated =
         rasterizer.AccelerateDisplay(*framebuffer, framebuffer_addr, framebuffer->stride);
     const bool is_srgb = use_accelerated && screen_info.is_srgb;
     RenderScreenshot(*framebuffer, use_accelerated);
 
-    bool has_been_recreated = false;
-    const auto recreate_swapchain = [&](u32 width, u32 height) {
-        if (!has_been_recreated) {
-            has_been_recreated = true;
-            scheduler.Finish();
-        }
-        swapchain.Create(width, height, is_srgb);
-    };
-
-    const Layout::FramebufferLayout layout = render_window.GetFramebufferLayout();
-    if (swapchain.NeedsRecreation(is_srgb) || swapchain.GetWidth() != layout.width ||
-        swapchain.GetHeight() != layout.height) {
-        recreate_swapchain(layout.width, layout.height);
-    }
-    bool is_outdated;
-    do {
-        swapchain.AcquireNextImage();
-        is_outdated = swapchain.IsOutDated();
-        if (is_outdated) {
-            recreate_swapchain(layout.width, layout.height);
-        }
-    } while (is_outdated);
-    if (has_been_recreated) {
-        blit_screen.Recreate();
-    }
-    const VkSemaphore render_semaphore = blit_screen.DrawToSwapchain(*framebuffer, use_accelerated);
-    const VkSemaphore present_semaphore = swapchain.CurrentPresentSemaphore();
-    scheduler.Flush(render_semaphore, present_semaphore);
-    scheduler.WaitWorker();
-    swapchain.Present(render_semaphore);
+    Frame* frame = present_manager.GetRenderFrame();
+    blit_screen.DrawToSwapchain(frame, *framebuffer, use_accelerated, is_srgb);
+    scheduler.Flush(*frame->render_ready);
+    present_manager.Present(frame);
 
     gpu.RendererFrameEndNotify();
     rasterizer.TickFrame();
@@ -246,8 +220,7 @@ void Vulkan::RendererVulkan::RenderScreenshot(const Tegra::FramebufferConfig& fr
     });
     const VkExtent2D render_area{.width = layout.width, .height = layout.height};
     const vk::Framebuffer screenshot_fb = blit_screen.CreateFramebuffer(*dst_view, render_area);
-    // Since we're not rendering to the screen, ignore the render semaphore.
-    void(blit_screen.Draw(framebuffer, *screenshot_fb, layout, render_area, use_accelerated));
+    blit_screen.Draw(framebuffer, *screenshot_fb, layout, render_area, use_accelerated);
 
     const auto buffer_size = static_cast<VkDeviceSize>(layout.width * layout.height * 4);
     const VkBufferCreateInfo dst_buffer_info{
@@ -270,7 +243,7 @@ void Vulkan::RendererVulkan::RenderScreenshot(const Tegra::FramebufferConfig& fr
             .pNext = nullptr,
             .srcAccessMask = VK_ACCESS_MEMORY_WRITE_BIT,
             .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT,
-            .oldLayout = VK_IMAGE_LAYOUT_PRESENT_SRC_KHR,
+            .oldLayout = VK_IMAGE_LAYOUT_GENERAL,
             .newLayout = VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL,
             .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
             .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,

src/video_core/renderer_vulkan/renderer_vulkan.h

@@ -9,6 +9,7 @@
 #include "common/dynamic_library.h"
 #include "video_core/renderer_base.h"
 #include "video_core/renderer_vulkan/vk_blit_screen.h"
+#include "video_core/renderer_vulkan/vk_present_manager.h"
 #include "video_core/renderer_vulkan/vk_rasterizer.h"
 #include "video_core/renderer_vulkan/vk_scheduler.h"
 #include "video_core/renderer_vulkan/vk_state_tracker.h"
@@ -76,6 +77,7 @@ private:
     StateTracker state_tracker;
     Scheduler scheduler;
     Swapchain swapchain;
+    PresentManager present_manager;
     BlitScreen blit_screen;
     RasterizerVulkan rasterizer;
     std::optional<TurboMode> turbo_mode;

src/video_core/renderer_vulkan/vk_blit_screen.cpp

@@ -122,10 +122,12 @@ struct BlitScreen::BufferData {
 
 BlitScreen::BlitScreen(Core::Memory::Memory& cpu_memory_, Core::Frontend::EmuWindow& render_window_,
                        const Device& device_, MemoryAllocator& memory_allocator_,
-                       Swapchain& swapchain_, Scheduler& scheduler_, const ScreenInfo& screen_info_)
+                       Swapchain& swapchain_, PresentManager& present_manager_,
+                       Scheduler& scheduler_, const ScreenInfo& screen_info_)
     : cpu_memory{cpu_memory_}, render_window{render_window_}, device{device_},
-      memory_allocator{memory_allocator_}, swapchain{swapchain_}, scheduler{scheduler_},
-      image_count{swapchain.GetImageCount()}, screen_info{screen_info_} {
+      memory_allocator{memory_allocator_}, swapchain{swapchain_}, present_manager{present_manager_},
+      scheduler{scheduler_}, image_count{swapchain.GetImageCount()}, screen_info{screen_info_},
+      current_srgb{swapchain.IsSrgb()}, image_view_format{swapchain.GetImageViewFormat()} {
     resource_ticks.resize(image_count);
 
     CreateStaticResources();
@@ -135,25 +137,20 @@ BlitScreen::BlitScreen(Core::Memory::Memory& cpu_memory_, Core::Frontend::EmuWin
 BlitScreen::~BlitScreen() = default;
 
 void BlitScreen::Recreate() {
+    present_manager.WaitPresent();
+    scheduler.Finish();
+    device.GetLogical().WaitIdle();
     CreateDynamicResources();
 }
 
-VkSemaphore BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
-                             const VkFramebuffer& host_framebuffer,
-                             const Layout::FramebufferLayout layout, VkExtent2D render_area,
-                             bool use_accelerated) {
+void BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
+                      const VkFramebuffer& host_framebuffer, const Layout::FramebufferLayout layout,
+                      VkExtent2D render_area, bool use_accelerated) {
     RefreshResources(framebuffer);
 
     // Finish any pending renderpass
     scheduler.RequestOutsideRenderPassOperationContext();
 
-    if (const auto swapchain_images = swapchain.GetImageCount(); swapchain_images != image_count) {
-        image_count = swapchain_images;
-        Recreate();
-    }
-
-    const std::size_t image_index = swapchain.GetImageIndex();
-
     scheduler.Wait(resource_ticks[image_index]);
     resource_ticks[image_index] = scheduler.CurrentTick();
 
@@ -169,7 +166,7 @@ VkSemaphore BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
     std::memcpy(mapped_span.data(), &data, sizeof(data));
 
     if (!use_accelerated) {
-        const u64 image_offset = GetRawImageOffset(framebuffer, image_index);
+        const u64 image_offset = GetRawImageOffset(framebuffer);
 
         const VAddr framebuffer_addr = framebuffer.address + framebuffer.offset;
         const u8* const host_ptr = cpu_memory.GetPointer(framebuffer_addr);
@@ -204,8 +201,8 @@ VkSemaphore BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
                     .depth = 1,
                 },
         };
-        scheduler.Record([this, copy, image_index](vk::CommandBuffer cmdbuf) {
-            const VkImage image = *raw_images[image_index];
+        scheduler.Record([this, copy, index = image_index](vk::CommandBuffer cmdbuf) {
+            const VkImage image = *raw_images[index];
             const VkImageMemoryBarrier base_barrier{
                 .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
                 .pNext = nullptr,
@@ -245,14 +242,15 @@ VkSemaphore BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
 
     const auto anti_alias_pass = Settings::values.anti_aliasing.GetValue();
     if (use_accelerated && anti_alias_pass == Settings::AntiAliasing::Fxaa) {
-        UpdateAADescriptorSet(image_index, source_image_view, false);
+        UpdateAADescriptorSet(source_image_view, false);
         const u32 up_scale = Settings::values.resolution_info.up_scale;
         const u32 down_shift = Settings::values.resolution_info.down_shift;
         VkExtent2D size{
             .width = (up_scale * framebuffer.width) >> down_shift,
             .height = (up_scale * framebuffer.height) >> down_shift,
         };
-        scheduler.Record([this, image_index, size, anti_alias_pass](vk::CommandBuffer cmdbuf) {
+        scheduler.Record([this, index = image_index, size,
+                          anti_alias_pass](vk::CommandBuffer cmdbuf) {
             const VkImageMemoryBarrier base_barrier{
                 .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
                 .pNext = nullptr,
@@ -326,7 +324,7 @@ VkSemaphore BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
 
             cmdbuf.BindVertexBuffer(0, *buffer, offsetof(BufferData, vertices));
             cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS, *aa_pipeline_layout, 0,
-                                      aa_descriptor_sets[image_index], {});
+                                      aa_descriptor_sets[index], {});
             cmdbuf.Draw(4, 1, 0, 0);
             cmdbuf.EndRenderPass();
 
@@ -369,81 +367,99 @@ VkSemaphore BlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer,
         };
         VkImageView fsr_image_view =
             fsr->Draw(scheduler, image_index, source_image_view, fsr_input_size, crop_rect);
-        UpdateDescriptorSet(image_index, fsr_image_view, true);
+        UpdateDescriptorSet(fsr_image_view, true);
     } else {
         const bool is_nn =
             Settings::values.scaling_filter.GetValue() == Settings::ScalingFilter::NearestNeighbor;
-        UpdateDescriptorSet(image_index, source_image_view, is_nn);
+        UpdateDescriptorSet(source_image_view, is_nn);
     }
 
-    scheduler.Record(
-        [this, host_framebuffer, image_index, size = render_area](vk::CommandBuffer cmdbuf) {
-            const f32 bg_red = Settings::values.bg_red.GetValue() / 255.0f;
-            const f32 bg_green = Settings::values.bg_green.GetValue() / 255.0f;
-            const f32 bg_blue = Settings::values.bg_blue.GetValue() / 255.0f;
-            const VkClearValue clear_color{
-                .color = {.float32 = {bg_red, bg_green, bg_blue, 1.0f}},
-            };
-            const VkRenderPassBeginInfo renderpass_bi{
-                .sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO,
-                .pNext = nullptr,
-                .renderPass = *renderpass,
-                .framebuffer = host_framebuffer,
-                .renderArea =
-                    {
-                        .offset = {0, 0},
-                        .extent = size,
-                    },
-                .clearValueCount = 1,
-                .pClearValues = &clear_color,
-            };
-            const VkViewport viewport{
-                .x = 0.0f,
-                .y = 0.0f,
-                .width = static_cast<float>(size.width),
-                .height = static_cast<float>(size.height),
-                .minDepth = 0.0f,
-                .maxDepth = 1.0f,
-            };
-            const VkRect2D scissor{
-                .offset = {0, 0},
-                .extent = size,
-            };
-            cmdbuf.BeginRenderPass(renderpass_bi, VK_SUBPASS_CONTENTS_INLINE);
-            auto graphics_pipeline = [this]() {
-                switch (Settings::values.scaling_filter.GetValue()) {
-                case Settings::ScalingFilter::NearestNeighbor:
-                case Settings::ScalingFilter::Bilinear:
-                    return *bilinear_pipeline;
-                case Settings::ScalingFilter::Bicubic:
-                    return *bicubic_pipeline;
-                case Settings::ScalingFilter::Gaussian:
-                    return *gaussian_pipeline;
-                case Settings::ScalingFilter::ScaleForce:
-                    return *scaleforce_pipeline;
-                default:
-                    return *bilinear_pipeline;
-                }
-            }();
-            cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, graphics_pipeline);
-            cmdbuf.SetViewport(0, viewport);
-            cmdbuf.SetScissor(0, scissor);
+    scheduler.Record([this, host_framebuffer, index = image_index,
+                      size = render_area](vk::CommandBuffer cmdbuf) {
+        const f32 bg_red = Settings::values.bg_red.GetValue() / 255.0f;
+        const f32 bg_green = Settings::values.bg_green.GetValue() / 255.0f;
+        const f32 bg_blue = Settings::values.bg_blue.GetValue() / 255.0f;
+        const VkClearValue clear_color{
+            .color = {.float32 = {bg_red, bg_green, bg_blue, 1.0f}},
+        };
+        const VkRenderPassBeginInfo renderpass_bi{
+            .sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO,
+            .pNext = nullptr,
+            .renderPass = *renderpass,
+            .framebuffer = host_framebuffer,
+            .renderArea =
+                {
+                    .offset = {0, 0},
+                    .extent = size,
+                },
+            .clearValueCount = 1,
+            .pClearValues = &clear_color,
+        };
+        const VkViewport viewport{
+            .x = 0.0f,
+            .y = 0.0f,
+            .width = static_cast<float>(size.width),
+            .height = static_cast<float>(size.height),
+            .minDepth = 0.0f,
+            .maxDepth = 1.0f,
+        };
+        const VkRect2D scissor{
+            .offset = {0, 0},
+            .extent = size,
+        };
+        cmdbuf.BeginRenderPass(renderpass_bi, VK_SUBPASS_CONTENTS_INLINE);
+        auto graphics_pipeline = [this]() {
+            switch (Settings::values.scaling_filter.GetValue()) {
+            case Settings::ScalingFilter::NearestNeighbor:
+            case Settings::ScalingFilter::Bilinear:
+                return *bilinear_pipeline;
+            case Settings::ScalingFilter::Bicubic:
+                return *bicubic_pipeline;
+            case Settings::ScalingFilter::Gaussian:
+                return *gaussian_pipeline;
+            case Settings::ScalingFilter::ScaleForce:
+                return *scaleforce_pipeline;
+            default:
+                return *bilinear_pipeline;
+            }
+        }();
+        cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, graphics_pipeline);
+        cmdbuf.SetViewport(0, viewport);
+        cmdbuf.SetScissor(0, scissor);
 
-            cmdbuf.BindVertexBuffer(0, *buffer, offsetof(BufferData, vertices));
-            cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS, *pipeline_layout, 0,
-                                      descriptor_sets[image_index], {});
-            cmdbuf.Draw(4, 1, 0, 0);
-            cmdbuf.EndRenderPass();
-        });
-    return *semaphores[image_index];
+        cmdbuf.BindVertexBuffer(0, *buffer, offsetof(BufferData, vertices));
+        cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS, *pipeline_layout, 0,
+                                  descriptor_sets[index], {});
+        cmdbuf.Draw(4, 1, 0, 0);
+        cmdbuf.EndRenderPass();
+    });
 }
 
-VkSemaphore BlitScreen::DrawToSwapchain(const Tegra::FramebufferConfig& framebuffer,
-                                        bool use_accelerated) {
-    const std::size_t image_index = swapchain.GetImageIndex();
-    const VkExtent2D render_area = swapchain.GetSize();
+void BlitScreen::DrawToSwapchain(Frame* frame, const Tegra::FramebufferConfig& framebuffer,
+                                 bool use_accelerated, bool is_srgb) {
+    // Recreate dynamic resources if the the image count or colorspace changed
+    if (const std::size_t swapchain_images = swapchain.GetImageCount();
+        swapchain_images != image_count || current_srgb != is_srgb) {
+        current_srgb = is_srgb;
+        image_view_format = current_srgb ? VK_FORMAT_B8G8R8A8_SRGB : VK_FORMAT_B8G8R8A8_UNORM;
+        image_count = swapchain_images;
+        Recreate();
+    }
+
+    // Recreate the presentation frame if the dimensions of the window changed
     const Layout::FramebufferLayout layout = render_window.GetFramebufferLayout();
-    return Draw(framebuffer, *framebuffers[image_index], layout, render_area, use_accelerated);
+    if (layout.width != frame->width || layout.height != frame->height ||
+        is_srgb != frame->is_srgb) {
+        Recreate();
+        present_manager.RecreateFrame(frame, layout.width, layout.height, is_srgb,
+                                      image_view_format, *renderpass);
+    }
+
+    const VkExtent2D render_area{frame->width, frame->height};
+    Draw(framebuffer, *frame->framebuffer, layout, render_area, use_accelerated);
+    if (++image_index >= image_count) {
+        image_index = 0;
+    }
 }
 
 vk::Framebuffer BlitScreen::CreateFramebuffer(const VkImageView& image_view, VkExtent2D extent) {
@@ -471,13 +487,11 @@ void BlitScreen::CreateStaticResources() {
 }
 
 void BlitScreen::CreateDynamicResources() {
-    CreateSemaphores();
     CreateDescriptorPool();
     CreateDescriptorSetLayout();
     CreateDescriptorSets();
     CreatePipelineLayout();
     CreateRenderPass();
-    CreateFramebuffers();
     CreateGraphicsPipeline();
     fsr.reset();
     smaa.reset();
@@ -525,11 +539,6 @@ void BlitScreen::CreateShaders() {
     }
 }
 
-void BlitScreen::CreateSemaphores() {
-    semaphores.resize(image_count);
-    std::ranges::generate(semaphores, [this] { return device.GetLogical().CreateSemaphore(); });
-}
-
 void BlitScreen::CreateDescriptorPool() {
     const std::array<VkDescriptorPoolSize, 2> pool_sizes{{
         {
@@ -571,10 +580,10 @@ void BlitScreen::CreateDescriptorPool() {
 }
 
 void BlitScreen::CreateRenderPass() {
-    renderpass = CreateRenderPassImpl(swapchain.GetImageViewFormat());
+    renderpass = CreateRenderPassImpl(image_view_format);
 }
 
-vk::RenderPass BlitScreen::CreateRenderPassImpl(VkFormat format, bool is_present) {
+vk::RenderPass BlitScreen::CreateRenderPassImpl(VkFormat format) {
     const VkAttachmentDescription color_attachment{
         .flags = 0,
         .format = format,
@@ -584,7 +593,7 @@ vk::RenderPass BlitScreen::CreateRenderPassImpl(VkFormat format, bool is_present
         .stencilLoadOp = VK_ATTACHMENT_LOAD_OP_DONT_CARE,
         .stencilStoreOp = VK_ATTACHMENT_STORE_OP_DONT_CARE,
         .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED,
-        .finalLayout = is_present ? VK_IMAGE_LAYOUT_PRESENT_SRC_KHR : VK_IMAGE_LAYOUT_GENERAL,
+        .finalLayout = VK_IMAGE_LAYOUT_GENERAL,
     };
 
     const VkAttachmentReference color_attachment_ref{
@@ -1052,16 +1061,6 @@ void BlitScreen::CreateSampler() {
     nn_sampler = device.GetLogical().CreateSampler(ci_nn);
 }
 
-void BlitScreen::CreateFramebuffers() {
-    const VkExtent2D size{swapchain.GetSize()};
-    framebuffers.resize(image_count);
-
-    for (std::size_t i = 0; i < image_count; ++i) {
-        const VkImageView image_view{swapchain.GetImageViewIndex(i)};
-        framebuffers[i] = CreateFramebuffer(image_view, size, renderpass);
-    }
-}
-
 void BlitScreen::ReleaseRawImages() {
     for (const u64 tick : resource_ticks) {
         scheduler.Wait(tick);
@@ -1175,7 +1174,7 @@ void BlitScreen::CreateRawImages(const Tegra::FramebufferConfig& framebuffer) {
         aa_framebuffer = CreateFramebuffer(*aa_image_view, size, aa_renderpass);
         return;
     }
-    aa_renderpass = CreateRenderPassImpl(GetFormat(framebuffer), false);
+    aa_renderpass = CreateRenderPassImpl(GetFormat(framebuffer));
     aa_framebuffer = CreateFramebuffer(*aa_image_view, size, aa_renderpass);
 
     const std::array<VkPipelineShaderStageCreateInfo, 2> fxaa_shader_stages{{
@@ -1319,8 +1318,7 @@ void BlitScreen::CreateRawImages(const Tegra::FramebufferConfig& framebuffer) {
     aa_pipeline = device.GetLogical().CreateGraphicsPipeline(fxaa_pipeline_ci);
 }
 
-void BlitScreen::UpdateAADescriptorSet(std::size_t image_index, VkImageView image_view,
-                                       bool nn) const {
+void BlitScreen::UpdateAADescriptorSet(VkImageView image_view, bool nn) const {
     const VkDescriptorImageInfo image_info{
         .sampler = nn ? *nn_sampler : *sampler,
         .imageView = image_view,
@@ -1356,8 +1354,7 @@ void BlitScreen::UpdateAADescriptorSet(std::size_t image_index, VkImageView imag
     device.GetLogical().UpdateDescriptorSets(std::array{sampler_write, sampler_write_2}, {});
 }
 
-void BlitScreen::UpdateDescriptorSet(std::size_t image_index, VkImageView image_view,
-                                     bool nn) const {
+void BlitScreen::UpdateDescriptorSet(VkImageView image_view, bool nn) const {
     const VkDescriptorBufferInfo buffer_info{
         .buffer = *buffer,
         .offset = offsetof(BufferData, uniform),
@@ -1480,8 +1477,7 @@ u64 BlitScreen::CalculateBufferSize(const Tegra::FramebufferConfig& framebuffer)
     return sizeof(BufferData) + GetSizeInBytes(framebuffer) * image_count;
 }
 
-u64 BlitScreen::GetRawImageOffset(const Tegra::FramebufferConfig& framebuffer,
-                                  std::size_t image_index) const {
+u64 BlitScreen::GetRawImageOffset(const Tegra::FramebufferConfig& framebuffer) const {
     constexpr auto first_image_offset = static_cast<u64>(sizeof(BufferData));
     return first_image_offset + GetSizeInBytes(framebuffer) * image_index;
 }

src/video_core/renderer_vulkan/vk_blit_screen.h

@@ -5,6 +5,7 @@
 
 #include <memory>
 
+#include "core/frontend/framebuffer_layout.h"
 #include "video_core/vulkan_common/vulkan_memory_allocator.h"
 #include "video_core/vulkan_common/vulkan_wrapper.h"
 
@@ -42,6 +43,9 @@ class RasterizerVulkan;
 class Scheduler;
 class SMAA;
 class Swapchain;
+class PresentManager;
+
+struct Frame;
 
 struct ScreenInfo {
     VkImage image{};
@@ -55,18 +59,17 @@ class BlitScreen {
 public:
     explicit BlitScreen(Core::Memory::Memory& cpu_memory, Core::Frontend::EmuWindow& render_window,
                         const Device& device, MemoryAllocator& memory_manager, Swapchain& swapchain,
-                        Scheduler& scheduler, const ScreenInfo& screen_info);
+                        PresentManager& present_manager, Scheduler& scheduler,
+                        const ScreenInfo& screen_info);
     ~BlitScreen();
 
     void Recreate();
 
-    [[nodiscard]] VkSemaphore Draw(const Tegra::FramebufferConfig& framebuffer,
-                                   const VkFramebuffer& host_framebuffer,
-                                   const Layout::FramebufferLayout layout, VkExtent2D render_area,
-                                   bool use_accelerated);
+    void Draw(const Tegra::FramebufferConfig& framebuffer, const VkFramebuffer& host_framebuffer,
+              const Layout::FramebufferLayout layout, VkExtent2D render_area, bool use_accelerated);
 
-    [[nodiscard]] VkSemaphore DrawToSwapchain(const Tegra::FramebufferConfig& framebuffer,
-                                              bool use_accelerated);
+    void DrawToSwapchain(Frame* frame, const Tegra::FramebufferConfig& framebuffer,
+                         bool use_accelerated, bool is_srgb);
 
     [[nodiscard]] vk::Framebuffer CreateFramebuffer(const VkImageView& image_view,
                                                     VkExtent2D extent);
@@ -79,10 +82,9 @@ private:
 
     void CreateStaticResources();
     void CreateShaders();
-    void CreateSemaphores();
     void CreateDescriptorPool();
     void CreateRenderPass();
-    vk::RenderPass CreateRenderPassImpl(VkFormat, bool is_present = true);
+    vk::RenderPass CreateRenderPassImpl(VkFormat format);
     void CreateDescriptorSetLayout();
     void CreateDescriptorSets();
     void CreatePipelineLayout();
@@ -90,15 +92,14 @@ private:
     void CreateSampler();
 
     void CreateDynamicResources();
-    void CreateFramebuffers();
 
     void RefreshResources(const Tegra::FramebufferConfig& framebuffer);
     void ReleaseRawImages();
     void CreateStagingBuffer(const Tegra::FramebufferConfig& framebuffer);
     void CreateRawImages(const Tegra::FramebufferConfig& framebuffer);
 
-    void UpdateDescriptorSet(std::size_t image_index, VkImageView image_view, bool nn) const;
-    void UpdateAADescriptorSet(std::size_t image_index, VkImageView image_view, bool nn) const;
+    void UpdateDescriptorSet(VkImageView image_view, bool nn) const;
+    void UpdateAADescriptorSet(VkImageView image_view, bool nn) const;
     void SetUniformData(BufferData& data, const Layout::FramebufferLayout layout) const;
     void SetVertexData(BufferData& data, const Tegra::FramebufferConfig& framebuffer,
                        const Layout::FramebufferLayout layout) const;
@@ -107,16 +108,17 @@ private:
     void CreateFSR();
 
     u64 CalculateBufferSize(const Tegra::FramebufferConfig& framebuffer) const;
-    u64 GetRawImageOffset(const Tegra::FramebufferConfig& framebuffer,
-                          std::size_t image_index) const;
+    u64 GetRawImageOffset(const Tegra::FramebufferConfig& framebuffer) const;
 
     Core::Memory::Memory& cpu_memory;
     Core::Frontend::EmuWindow& render_window;
     const Device& device;
     MemoryAllocator& memory_allocator;
     Swapchain& swapchain;
+    PresentManager& present_manager;
     Scheduler& scheduler;
     std::size_t image_count;
+    std::size_t image_index{};
     const ScreenInfo& screen_info;
 
     vk::ShaderModule vertex_shader;
@@ -135,7 +137,6 @@ private:
     vk::Pipeline gaussian_pipeline;
     vk::Pipeline scaleforce_pipeline;
     vk::RenderPass renderpass;
-    std::vector<vk::Framebuffer> framebuffers;
     vk::DescriptorSets descriptor_sets;
     vk::Sampler nn_sampler;
     vk::Sampler sampler;
@@ -145,7 +146,6 @@ private:
 
     std::vector<u64> resource_ticks;
 
-    std::vector<vk::Semaphore> semaphores;
     std::vector<vk::Image> raw_images;
     std::vector<vk::ImageView> raw_image_views;
     std::vector<MemoryCommit> raw_buffer_commits;
@@ -164,6 +164,8 @@ private:
     u32 raw_width = 0;
     u32 raw_height = 0;
     Service::android::PixelFormat pixel_format{};
+    bool current_srgb;
+    VkFormat image_view_format;
 
     std::unique_ptr<FSR> fsr;
     std::unique_ptr<SMAA> smaa;

src/video_core/renderer_vulkan/vk_buffer_cache.cpp

@@ -314,8 +314,12 @@ StagingBufferRef BufferCacheRuntime::UploadStagingBuffer(size_t size) {
     return staging_pool.Request(size, MemoryUsage::Upload);
 }
 
-StagingBufferRef BufferCacheRuntime::DownloadStagingBuffer(size_t size) {
-    return staging_pool.Request(size, MemoryUsage::Download);
+StagingBufferRef BufferCacheRuntime::DownloadStagingBuffer(size_t size, bool deferred) {
+    return staging_pool.Request(size, MemoryUsage::Download, deferred);
+}
+
+void BufferCacheRuntime::FreeDeferredStagingBuffer(StagingBufferRef& ref) {
+    staging_pool.FreeDeferred(ref);
 }
 
 u64 BufferCacheRuntime::GetDeviceLocalMemory() const {

src/video_core/renderer_vulkan/vk_buffer_cache.h

@@ -3,7 +3,8 @@
 
 #pragma once
 
-#include "video_core/buffer_cache/buffer_cache.h"
+#include "video_core/buffer_cache/buffer_cache_base.h"
+#include "video_core/buffer_cache/memory_tracker_base.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_vulkan/vk_compute_pass.h"
 #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h"
@@ -75,7 +76,9 @@ public:
 
     [[nodiscard]] StagingBufferRef UploadStagingBuffer(size_t size);
 
-    [[nodiscard]] StagingBufferRef DownloadStagingBuffer(size_t size);
+    [[nodiscard]] StagingBufferRef DownloadStagingBuffer(size_t size, bool deferred = false);
+
+    void FreeDeferredStagingBuffer(StagingBufferRef& ref);
 
     void PreCopyBarrier();
 
@@ -142,6 +145,8 @@ private:
 struct BufferCacheParams {
     using Runtime = Vulkan::BufferCacheRuntime;
     using Buffer = Vulkan::Buffer;
+    using Async_Buffer = Vulkan::StagingBufferRef;
+    using MemoryTracker = VideoCommon::MemoryTrackerBase<VideoCore::RasterizerInterface>;
 
     static constexpr bool IS_OPENGL = false;
     static constexpr bool HAS_PERSISTENT_UNIFORM_BUFFER_BINDINGS = false;
@@ -150,6 +155,7 @@ struct BufferCacheParams {
     static constexpr bool NEEDS_BIND_STORAGE_INDEX = false;
     static constexpr bool USE_MEMORY_MAPS = true;
     static constexpr bool SEPARATE_IMAGE_BUFFER_BINDINGS = false;
+    static constexpr bool IMPLEMENTS_ASYNC_DOWNLOADS = true;
 };
 
 using BufferCache = VideoCommon::BufferCache<BufferCacheParams>;

src/video_core/renderer_vulkan/vk_buffer_cache_base.cpp

@@ -0,0 +1,9 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include "video_core/buffer_cache/buffer_cache.h"
+#include "video_core/renderer_vulkan/vk_buffer_cache.h"
+
+namespace VideoCommon {
+template class VideoCommon::BufferCache<Vulkan::BufferCacheParams>;
+}

src/video_core/renderer_vulkan/vk_fence_manager.cpp

@@ -5,6 +5,7 @@
 
 #include "video_core/renderer_vulkan/vk_buffer_cache.h"
 #include "video_core/renderer_vulkan/vk_fence_manager.h"
+#include "video_core/renderer_vulkan/vk_query_cache.h"
 #include "video_core/renderer_vulkan/vk_scheduler.h"
 #include "video_core/renderer_vulkan/vk_texture_cache.h"
 #include "video_core/vulkan_common/vulkan_device.h"

src/video_core/renderer_vulkan/vk_fence_manager.h

@@ -40,7 +40,16 @@ private:
 };
 using Fence = std::shared_ptr<InnerFence>;
 
-using GenericFenceManager = VideoCommon::FenceManager<Fence, TextureCache, BufferCache, QueryCache>;
+struct FenceManagerParams {
+    using FenceType = Fence;
+    using BufferCacheType = BufferCache;
+    using TextureCacheType = TextureCache;
+    using QueryCacheType = QueryCache;
+
+    static constexpr bool HAS_ASYNC_CHECK = true;
+};
+
+using GenericFenceManager = VideoCommon::FenceManager<FenceManagerParams>;
 
 class FenceManager final : public GenericFenceManager {
 public:

src/video_core/renderer_vulkan/vk_master_semaphore.cpp

@@ -10,7 +10,14 @@
 
 namespace Vulkan {
 
-MasterSemaphore::MasterSemaphore(const Device& device) {
+MasterSemaphore::MasterSemaphore(const Device& device_) : device(device_) {
+    if (!device.HasTimelineSemaphore()) {
+        static constexpr VkFenceCreateInfo fence_ci{
+            .sType = VK_STRUCTURE_TYPE_FENCE_CREATE_INFO, .pNext = nullptr, .flags = 0};
+        fence = device.GetLogical().CreateFence(fence_ci);
+        return;
+    }
+
     static constexpr VkSemaphoreTypeCreateInfo semaphore_type_ci{
         .sType = VK_STRUCTURE_TYPE_SEMAPHORE_TYPE_CREATE_INFO,
         .pNext = nullptr,
@@ -42,4 +49,134 @@ MasterSemaphore::MasterSemaphore(const Device& device) {
 
 MasterSemaphore::~MasterSemaphore() = default;
 
+void MasterSemaphore::Refresh() {
+    if (!semaphore) {
+        // If we don't support timeline semaphores, there's nothing to refresh
+        return;
+    }
+
+    u64 this_tick{};
+    u64 counter{};
+    do {
+        this_tick = gpu_tick.load(std::memory_order_acquire);
+        counter = semaphore.GetCounter();
+        if (counter < this_tick) {
+            return;
+        }
+    } while (!gpu_tick.compare_exchange_weak(this_tick, counter, std::memory_order_release,
+                                             std::memory_order_relaxed));
+}
+
+void MasterSemaphore::Wait(u64 tick) {
+    if (!semaphore) {
+        // If we don't support timeline semaphores, use an atomic wait
+        while (true) {
+            u64 current_value = gpu_tick.load(std::memory_order_relaxed);
+            if (current_value >= tick) {
+                return;
+            }
+            gpu_tick.wait(current_value);
+        }
+
+        return;
+    }
+
+    // No need to wait if the GPU is ahead of the tick
+    if (IsFree(tick)) {
+        return;
+    }
+
+    // Update the GPU tick and try again
+    Refresh();
+
+    if (IsFree(tick)) {
+        return;
+    }
+
+    // If none of the above is hit, fallback to a regular wait
+    while (!semaphore.Wait(tick)) {
+    }
+
+    Refresh();
+}
+
+VkResult MasterSemaphore::SubmitQueue(vk::CommandBuffer& cmdbuf, VkSemaphore signal_semaphore,
+                                      VkSemaphore wait_semaphore, u64 host_tick) {
+    if (semaphore) {
+        return SubmitQueueTimeline(cmdbuf, signal_semaphore, wait_semaphore, host_tick);
+    } else {
+        return SubmitQueueFence(cmdbuf, signal_semaphore, wait_semaphore, host_tick);
+    }
+}
+
+static constexpr std::array<VkPipelineStageFlags, 2> wait_stage_masks{
+    VK_PIPELINE_STAGE_ALL_COMMANDS_BIT,
+    VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT,
+};
+
+VkResult MasterSemaphore::SubmitQueueTimeline(vk::CommandBuffer& cmdbuf,
+                                              VkSemaphore signal_semaphore,
+                                              VkSemaphore wait_semaphore, u64 host_tick) {
+    const VkSemaphore timeline_semaphore = *semaphore;
+
+    const u32 num_signal_semaphores = signal_semaphore ? 2 : 1;
+    const std::array signal_values{host_tick, u64(0)};
+    const std::array signal_semaphores{timeline_semaphore, signal_semaphore};
+
+    const u32 num_wait_semaphores = wait_semaphore ? 2 : 1;
+    const std::array wait_values{host_tick - 1, u64(1)};
+    const std::array wait_semaphores{timeline_semaphore, wait_semaphore};
+
+    const VkTimelineSemaphoreSubmitInfo timeline_si{
+        .sType = VK_STRUCTURE_TYPE_TIMELINE_SEMAPHORE_SUBMIT_INFO,
+        .pNext = nullptr,
+        .waitSemaphoreValueCount = num_wait_semaphores,
+        .pWaitSemaphoreValues = wait_values.data(),
+        .signalSemaphoreValueCount = num_signal_semaphores,
+        .pSignalSemaphoreValues = signal_values.data(),
+    };
+    const VkSubmitInfo submit_info{
+        .sType = VK_STRUCTURE_TYPE_SUBMIT_INFO,
+        .pNext = &timeline_si,
+        .waitSemaphoreCount = num_wait_semaphores,
+        .pWaitSemaphores = wait_semaphores.data(),
+        .pWaitDstStageMask = wait_stage_masks.data(),
+        .commandBufferCount = 1,
+        .pCommandBuffers = cmdbuf.address(),
+        .signalSemaphoreCount = num_signal_semaphores,
+        .pSignalSemaphores = signal_semaphores.data(),
+    };
+
+    return device.GetGraphicsQueue().Submit(submit_info);
+}
+
+VkResult MasterSemaphore::SubmitQueueFence(vk::CommandBuffer& cmdbuf, VkSemaphore signal_semaphore,
+                                           VkSemaphore wait_semaphore, u64 host_tick) {
+    const u32 num_signal_semaphores = signal_semaphore ? 1 : 0;
+    const u32 num_wait_semaphores = wait_semaphore ? 1 : 0;
+
+    const VkSubmitInfo submit_info{
+        .sType = VK_STRUCTURE_TYPE_SUBMIT_INFO,
+        .pNext = nullptr,
+        .waitSemaphoreCount = num_wait_semaphores,
+        .pWaitSemaphores = &wait_semaphore,
+        .pWaitDstStageMask = wait_stage_masks.data(),
+        .commandBufferCount = 1,
+        .pCommandBuffers = cmdbuf.address(),
+        .signalSemaphoreCount = num_signal_semaphores,
+        .pSignalSemaphores = &signal_semaphore,
+    };
+
+    auto result = device.GetGraphicsQueue().Submit(submit_info, *fence);
+
+    if (result == VK_SUCCESS) {
+        fence.Wait();
+        fence.Reset();
+        gpu_tick.store(host_tick);
+        gpu_tick.notify_all();
+    }
+
+    return result;
+}
+
 } // namespace Vulkan

src/video_core/renderer_vulkan/vk_master_semaphore.h

@@ -4,6 +4,8 @@
 #pragma once
 
 #include <atomic>
+#include <condition_variable>
+#include <mutex>
 #include <thread>
 
 #include "common/common_types.h"
@@ -29,11 +31,6 @@ public:
         return gpu_tick.load(std::memory_order_acquire);
     }
 
-    /// Returns the timeline semaphore handle.
-    [[nodiscard]] VkSemaphore Handle() const noexcept {
-        return *semaphore;
-    }
-
     /// Returns true when a tick has been hit by the GPU.
     [[nodiscard]] bool IsFree(u64 tick) const noexcept {
         return KnownGpuTick() >= tick;
@@ -45,37 +42,24 @@ public:
     }
 
     /// Refresh the known GPU tick
-    void Refresh() {
-        u64 this_tick{};
-        u64 counter{};
-        do {
-            this_tick = gpu_tick.load(std::memory_order_acquire);
-            counter = semaphore.GetCounter();
-            if (counter < this_tick) {
-                return;
-            }
-        } while (!gpu_tick.compare_exchange_weak(this_tick, counter, std::memory_order_release,
-                                                 std::memory_order_relaxed));
-    }
+    void Refresh();
 
     /// Waits for a tick to be hit on the GPU
-    void Wait(u64 tick) {
-        // No need to wait if the GPU is ahead of the tick
-        if (IsFree(tick)) {
-            return;
-        }
-        // Update the GPU tick and try again
-        Refresh();
-        if (IsFree(tick)) {
-            return;
-        }
-        // If none of the above is hit, fallback to a regular wait
-        while (!semaphore.Wait(tick)) {
-        }
-        Refresh();
-    }
+    void Wait(u64 tick);
+
+    /// Submits the device graphics queue, updating the tick as necessary
+    VkResult SubmitQueue(vk::CommandBuffer& cmdbuf, VkSemaphore signal_semaphore,
+                         VkSemaphore wait_semaphore, u64 host_tick);
 
 private:
+    VkResult SubmitQueueTimeline(vk::CommandBuffer& cmdbuf, VkSemaphore signal_semaphore,
+                                 VkSemaphore wait_semaphore, u64 host_tick);
+    VkResult SubmitQueueFence(vk::CommandBuffer& cmdbuf, VkSemaphore signal_semaphore,
+                              VkSemaphore wait_semaphore, u64 host_tick);
+
+private:
+    const Device& device;             ///< Device.
+    vk::Fence fence;                  ///< Fence.
     vk::Semaphore semaphore;          ///< Timeline semaphore.
     std::atomic<u64> gpu_tick{0};     ///< Current known GPU tick.
     std::atomic<u64> current_tick{1}; ///< Current logical tick.

src/video_core/renderer_vulkan/vk_pipeline_cache.cpp

@@ -696,6 +696,13 @@ std::unique_ptr<ComputePipeline> PipelineCache::CreateComputePipeline(
 std::unique_ptr<ComputePipeline> PipelineCache::CreateComputePipeline(
     ShaderPools& pools, const ComputePipelineCacheKey& key, Shader::Environment& env,
     PipelineStatistics* statistics, bool build_in_parallel) try {
+    // TODO: Remove this when Intel fixes their shader compiler.
+    //       https://github.com/IGCIT/Intel-GPU-Community-Issue-Tracker-IGCIT/issues/159
+    if (device.GetDriverID() == VK_DRIVER_ID_INTEL_PROPRIETARY_WINDOWS) {
+        LOG_ERROR(Render_Vulkan, "Skipping 0x{:016x}", key.Hash());
+        return nullptr;
+    }
+
     LOG_INFO(Render_Vulkan, "0x{:016x}", key.Hash());
 
     Shader::Maxwell::Flow::CFG cfg{env, pools.flow_block, env.StartAddress()};

src/video_core/renderer_vulkan/vk_present_manager.cpp

@@ -0,0 +1,457 @@
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include "common/microprofile.h"
+#include "common/settings.h"
+#include "common/thread.h"
+#include "video_core/renderer_vulkan/vk_present_manager.h"
+#include "video_core/renderer_vulkan/vk_scheduler.h"
+#include "video_core/renderer_vulkan/vk_swapchain.h"
+#include "video_core/vulkan_common/vulkan_device.h"
+
+namespace Vulkan {
+
+MICROPROFILE_DEFINE(Vulkan_WaitPresent, "Vulkan", "Wait For Present", MP_RGB(128, 128, 128));
+MICROPROFILE_DEFINE(Vulkan_CopyToSwapchain, "Vulkan", "Copy to swapchain", MP_RGB(192, 255, 192));
+
+namespace {
+
+bool CanBlitToSwapchain(const vk::PhysicalDevice& physical_device, VkFormat format) {
+    const VkFormatProperties props{physical_device.GetFormatProperties(format)};
+    return (props.optimalTilingFeatures & VK_FORMAT_FEATURE_BLIT_DST_BIT);
+}
+
+[[nodiscard]] VkImageSubresourceLayers MakeImageSubresourceLayers() {
+    return VkImageSubresourceLayers{
+        .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+        .mipLevel = 0,
+        .baseArrayLayer = 0,
+        .layerCount = 1,
+    };
+}
+
+[[nodiscard]] VkImageBlit MakeImageBlit(s32 frame_width, s32 frame_height, s32 swapchain_width,
+                                        s32 swapchain_height) {
+    return VkImageBlit{
+        .srcSubresource = MakeImageSubresourceLayers(),
+        .srcOffsets =
+            {
+                {
+                    .x = 0,
+                    .y = 0,
+                    .z = 0,
+                },
+                {
+                    .x = frame_width,
+                    .y = frame_height,
+                    .z = 1,
+                },
+            },
+        .dstSubresource = MakeImageSubresourceLayers(),
+        .dstOffsets =
+            {
+                {
+                    .x = 0,
+                    .y = 0,
+                    .z = 0,
+                },
+                {
+                    .x = swapchain_width,
+                    .y = swapchain_height,
+                    .z = 1,
+                },
+            },
+    };
+}
+
+[[nodiscard]] VkImageCopy MakeImageCopy(u32 frame_width, u32 frame_height, u32 swapchain_width,
+                                        u32 swapchain_height) {
+    return VkImageCopy{
+        .srcSubresource = MakeImageSubresourceLayers(),
+        .srcOffset =
+            {
+                .x = 0,
+                .y = 0,
+                .z = 0,
+            },
+        .dstSubresource = MakeImageSubresourceLayers(),
+        .dstOffset =
+            {
+                .x = 0,
+                .y = 0,
+                .z = 0,
+            },
+        .extent =
+            {
+                .width = std::min(frame_width, swapchain_width),
+                .height = std::min(frame_height, swapchain_height),
+                .depth = 1,
+            },
+    };
+}
+
+} // Anonymous namespace
+
+PresentManager::PresentManager(Core::Frontend::EmuWindow& render_window_, const Device& device_,
+                               MemoryAllocator& memory_allocator_, Scheduler& scheduler_,
+                               Swapchain& swapchain_)
+    : render_window{render_window_}, device{device_},
+      memory_allocator{memory_allocator_}, scheduler{scheduler_}, swapchain{swapchain_},
+      blit_supported{CanBlitToSwapchain(device.GetPhysical(), swapchain.GetImageViewFormat())},
+      use_present_thread{Settings::values.async_presentation.GetValue()},
+      image_count{swapchain.GetImageCount()} {
+
+    auto& dld = device.GetLogical();
+    cmdpool = dld.CreateCommandPool({
+        .sType = VK_STRUCTURE_TYPE_COMMAND_POOL_CREATE_INFO,
+        .pNext = nullptr,
+        .flags =
+            VK_COMMAND_POOL_CREATE_TRANSIENT_BIT | VK_COMMAND_POOL_CREATE_RESET_COMMAND_BUFFER_BIT,
+        .queueFamilyIndex = device.GetGraphicsFamily(),
+    });
+    auto cmdbuffers = cmdpool.Allocate(image_count);
+
+    frames.resize(image_count);
+    for (u32 i = 0; i < frames.size(); i++) {
+        Frame& frame = frames[i];
+        frame.cmdbuf = vk::CommandBuffer{cmdbuffers[i], device.GetDispatchLoader()};
+        frame.render_ready = dld.CreateSemaphore({
+            .sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO,
+            .pNext = nullptr,
+            .flags = 0,
+        });
+        frame.present_done = dld.CreateFence({
+            .sType = VK_STRUCTURE_TYPE_FENCE_CREATE_INFO,
+            .pNext = nullptr,
+            .flags = VK_FENCE_CREATE_SIGNALED_BIT,
+        });
+        free_queue.push(&frame);
+    }
+
+    if (use_present_thread) {
+        present_thread = std::jthread([this](std::stop_token token) { PresentThread(token); });
+    }
+}
+
+PresentManager::~PresentManager() = default;
+
+Frame* PresentManager::GetRenderFrame() {
+    MICROPROFILE_SCOPE(Vulkan_WaitPresent);
+
+    // Wait for free presentation frames
+    std::unique_lock lock{free_mutex};
+    free_cv.wait(lock, [this] { return !free_queue.empty(); });
+
+    // Take the frame from the queue
+    Frame* frame = free_queue.front();
+    free_queue.pop();
+
+    // Wait for the presentation to be finished so all frame resources are free
+    frame->present_done.Wait();
+    frame->present_done.Reset();
+
+    return frame;
+}
+
+void PresentManager::Present(Frame* frame) {
+    if (!use_present_thread) {
+        scheduler.WaitWorker();
+        CopyToSwapchain(frame);
+        free_queue.push(frame);
+        return;
+    }
+
+    scheduler.Record([this, frame](vk::CommandBuffer) {
+        std::unique_lock lock{queue_mutex};
+        present_queue.push(frame);
+        frame_cv.notify_one();
+    });
+}
+
+void PresentManager::RecreateFrame(Frame* frame, u32 width, u32 height, bool is_srgb,
+                                   VkFormat image_view_format, VkRenderPass rd) {
+    auto& dld = device.GetLogical();
+
+    frame->width = width;
+    frame->height = height;
+    frame->is_srgb = is_srgb;
+
+    frame->image = dld.CreateImage({
+        .sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = VK_IMAGE_CREATE_MUTABLE_FORMAT_BIT,
+        .imageType = VK_IMAGE_TYPE_2D,
+        .format = swapchain.GetImageFormat(),
+        .extent =
+            {
+                .width = width,
+                .height = height,
+                .depth = 1,
+            },
+        .mipLevels = 1,
+        .arrayLayers = 1,
+        .samples = VK_SAMPLE_COUNT_1_BIT,
+        .tiling = VK_IMAGE_TILING_OPTIMAL,
+        .usage = VK_IMAGE_USAGE_TRANSFER_SRC_BIT | VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT,
+        .sharingMode = VK_SHARING_MODE_EXCLUSIVE,
+        .queueFamilyIndexCount = 0,
+        .pQueueFamilyIndices = nullptr,
+        .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED,
+    });
+
+    frame->image_commit = memory_allocator.Commit(frame->image, MemoryUsage::DeviceLocal);
+
+    frame->image_view = dld.CreateImageView({
+        .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .image = *frame->image,
+        .viewType = VK_IMAGE_VIEW_TYPE_2D,
+        .format = image_view_format,
+        .components =
+            {
+                .r = VK_COMPONENT_SWIZZLE_IDENTITY,
+                .g = VK_COMPONENT_SWIZZLE_IDENTITY,
+                .b = VK_COMPONENT_SWIZZLE_IDENTITY,
+                .a = VK_COMPONENT_SWIZZLE_IDENTITY,
+            },
+        .subresourceRange =
+            {
+                .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+                .baseMipLevel = 0,
+                .levelCount = 1,
+                .baseArrayLayer = 0,
+                .layerCount = 1,
+            },
+    });
+
+    const VkImageView image_view{*frame->image_view};
+    frame->framebuffer = dld.CreateFramebuffer({
+        .sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .renderPass = rd,
+        .attachmentCount = 1,
+        .pAttachments = &image_view,
+        .width = width,
+        .height = height,
+        .layers = 1,
+    });
+}
+
+void PresentManager::WaitPresent() {
+    if (!use_present_thread) {
+        return;
+    }
+
+    // Wait for the present queue to be empty
+    {
+        std::unique_lock queue_lock{queue_mutex};
+        frame_cv.wait(queue_lock, [this] { return present_queue.empty(); });
+    }
+
+    // The above condition will be satisfied when the last frame is taken from the queue.
+    // To ensure that frame has been presented as well take hold of the swapchain
+    // mutex.
+    std::scoped_lock swapchain_lock{swapchain_mutex};
+}
+
+void PresentManager::PresentThread(std::stop_token token) {
+    Common::SetCurrentThreadName("VulkanPresent");
+    while (!token.stop_requested()) {
+        std::unique_lock lock{queue_mutex};
+
+        // Wait for presentation frames
+        Common::CondvarWait(frame_cv, lock, token, [this] { return !present_queue.empty(); });
+        if (token.stop_requested()) {
+            return;
+        }
+
+        // Take the frame and notify anyone waiting
+        Frame* frame = present_queue.front();
+        present_queue.pop();
+        frame_cv.notify_one();
+
+        // By exchanging the lock ownership we take the swapchain lock
+        // before the queue lock goes out of scope. This way the swapchain
+        // lock in WaitPresent is guaranteed to occur after here.
+        std::exchange(lock, std::unique_lock{swapchain_mutex});
+
+        CopyToSwapchain(frame);
+
+        // Free the frame for reuse
+        std::scoped_lock fl{free_mutex};
+        free_queue.push(frame);
+        free_cv.notify_one();
+    }
+}
+
+void PresentManager::CopyToSwapchain(Frame* frame) {
+    MICROPROFILE_SCOPE(Vulkan_CopyToSwapchain);
+
+    const auto recreate_swapchain = [&] {
+        swapchain.Create(frame->width, frame->height, frame->is_srgb);
+        image_count = swapchain.GetImageCount();
+    };
+
+    // If the size or colorspace of the incoming frames has changed, recreate the swapchain
+    // to account for that.
+    const bool srgb_changed = swapchain.NeedsRecreation(frame->is_srgb);
+    const bool size_changed =
+        swapchain.GetWidth() != frame->width || swapchain.GetHeight() != frame->height;
+    if (srgb_changed || size_changed) {
+        recreate_swapchain();
+    }
+
+    while (swapchain.AcquireNextImage()) {
+        recreate_swapchain();
+    }
+
+    const vk::CommandBuffer cmdbuf{frame->cmdbuf};
+    cmdbuf.Begin({
+        .sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_BEGIN_INFO,
+        .pNext = nullptr,
+        .flags = VK_COMMAND_BUFFER_USAGE_ONE_TIME_SUBMIT_BIT,
+        .pInheritanceInfo = nullptr,
+    });
+
+    const VkImage image{swapchain.CurrentImage()};
+    const VkExtent2D extent = swapchain.GetExtent();
+    const std::array pre_barriers{
+        VkImageMemoryBarrier{
+            .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
+            .pNext = nullptr,
+            .srcAccessMask = 0,
+            .dstAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT,
+            .oldLayout = VK_IMAGE_LAYOUT_UNDEFINED,
+            .newLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL,
+            .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .image = image,
+            .subresourceRange{
+                .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+                .baseMipLevel = 0,
+                .levelCount = 1,
+                .baseArrayLayer = 0,
+                .layerCount = VK_REMAINING_ARRAY_LAYERS,
+            },
+        },
+        VkImageMemoryBarrier{
+            .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
+            .pNext = nullptr,
+            .srcAccessMask = VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT,
+            .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT,
+            .oldLayout = VK_IMAGE_LAYOUT_GENERAL,
+            .newLayout = VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL,
+            .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .image = *frame->image,
+            .subresourceRange{
+                .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+                .baseMipLevel = 0,
+                .levelCount = 1,
+                .baseArrayLayer = 0,
+                .layerCount = VK_REMAINING_ARRAY_LAYERS,
+            },
+        },
+    };
+    const std::array post_barriers{
+        VkImageMemoryBarrier{
+            .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
+            .pNext = nullptr,
+            .srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT,
+            .dstAccessMask = VK_ACCESS_MEMORY_READ_BIT,
+            .oldLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL,
+            .newLayout = VK_IMAGE_LAYOUT_PRESENT_SRC_KHR,
+            .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .image = image,
+            .subresourceRange{
+                .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+                .baseMipLevel = 0,
+                .levelCount = 1,
+                .baseArrayLayer = 0,
+                .layerCount = VK_REMAINING_ARRAY_LAYERS,
+            },
+        },
+        VkImageMemoryBarrier{
+            .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
+            .pNext = nullptr,
+            .srcAccessMask = VK_ACCESS_TRANSFER_READ_BIT,
+            .dstAccessMask = VK_ACCESS_MEMORY_WRITE_BIT,
+            .oldLayout = VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL,
+            .newLayout = VK_IMAGE_LAYOUT_GENERAL,
+            .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .image = *frame->image,
+            .subresourceRange{
+                .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
+                .baseMipLevel = 0,
+                .levelCount = 1,
+                .baseArrayLayer = 0,
+                .layerCount = VK_REMAINING_ARRAY_LAYERS,
+            },
+        },
+    };
+
+    cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, {},
+                           {}, {}, pre_barriers);
+
+    if (blit_supported) {
+        cmdbuf.BlitImage(*frame->image, VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, image,
+                         VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL,
+                         MakeImageBlit(frame->width, frame->height, extent.width, extent.height),
+                         VK_FILTER_LINEAR);
+    } else {
+        cmdbuf.CopyImage(*frame->image, VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, image,
+                         VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL,
+                         MakeImageCopy(frame->width, frame->height, extent.width, extent.height));
+    }
+
+    cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_PIPELINE_STAGE_ALL_GRAPHICS_BIT, {},
+                           {}, {}, post_barriers);
+
+    cmdbuf.End();
+
+    const VkSemaphore present_semaphore = swapchain.CurrentPresentSemaphore();
+    const VkSemaphore render_semaphore = swapchain.CurrentRenderSemaphore();
+    const std::array wait_semaphores = {present_semaphore, *frame->render_ready};
+
+    static constexpr std::array<VkPipelineStageFlags, 2> wait_stage_masks{
+        VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT,
+        VK_PIPELINE_STAGE_ALL_COMMANDS_BIT,
+    };
+
+    const VkSubmitInfo submit_info{
+        .sType = VK_STRUCTURE_TYPE_SUBMIT_INFO,
+        .pNext = nullptr,
+        .waitSemaphoreCount = 2U,
+        .pWaitSemaphores = wait_semaphores.data(),
+        .pWaitDstStageMask = wait_stage_masks.data(),
+        .commandBufferCount = 1,
+        .pCommandBuffers = cmdbuf.address(),
+        .signalSemaphoreCount = 1U,
+        .pSignalSemaphores = &render_semaphore,
+    };
+
+    // Submit the image copy/blit to the swapchain
+    {
+        std::scoped_lock lock{scheduler.submit_mutex};
+        switch (const VkResult result =
+                    device.GetGraphicsQueue().Submit(submit_info, *frame->present_done)) {
+        case VK_SUCCESS:
+            break;
+        case VK_ERROR_DEVICE_LOST:
+            device.ReportLoss();
+            [[fallthrough]];
+        default:
+            vk::Check(result);
+            break;
+        }
+    }
+
+    // Present
+    swapchain.Present(render_semaphore);
+}
+
+} // namespace Vulkan

src/video_core/renderer_vulkan/vk_present_manager.h

@@ -0,0 +1,83 @@
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#pragma once
+
+#include <condition_variable>
+#include <mutex>
+#include <queue>
+
+#include "common/common_types.h"
+#include "common/polyfill_thread.h"
+#include "video_core/vulkan_common/vulkan_memory_allocator.h"
+#include "video_core/vulkan_common/vulkan_wrapper.h"
+
+namespace Core::Frontend {
+class EmuWindow;
+} // namespace Core::Frontend
+
+namespace Vulkan {
+
+class Device;
+class Scheduler;
+class Swapchain;
+
+struct Frame {
+    u32 width;
+    u32 height;
+    bool is_srgb;
+    vk::Image image;
+    vk::ImageView image_view;
+    vk::Framebuffer framebuffer;
+    MemoryCommit image_commit;
+    vk::CommandBuffer cmdbuf;
+    vk::Semaphore render_ready;
+    vk::Fence present_done;
+};
+
+class PresentManager {
+public:
+    PresentManager(Core::Frontend::EmuWindow& render_window, const Device& device,
+                   MemoryAllocator& memory_allocator, Scheduler& scheduler, Swapchain& swapchain);
+    ~PresentManager();
+
+    /// Returns the last used presentation frame
+    Frame* GetRenderFrame();
+
+    /// Pushes a frame for presentation
+    void Present(Frame* frame);
+
+    /// Recreates the present frame to match the provided parameters
+    void RecreateFrame(Frame* frame, u32 width, u32 height, bool is_srgb,
+                       VkFormat image_view_format, VkRenderPass rd);
+
+    /// Waits for the present thread to finish presenting all queued frames.
+    void WaitPresent();
+
+private:
+    void PresentThread(std::stop_token token);
+
+    void CopyToSwapchain(Frame* frame);
+
+private:
+    Core::Frontend::EmuWindow& render_window;
+    const Device& device;
+    MemoryAllocator& memory_allocator;
+    Scheduler& scheduler;
+    Swapchain& swapchain;
+    vk::CommandPool cmdpool;
+    std::vector<Frame> frames;
+    std::queue<Frame*> present_queue;
+    std::queue<Frame*> free_queue;
+    std::condition_variable_any frame_cv;
+    std::condition_variable free_cv;
+    std::mutex swapchain_mutex;
+    std::mutex queue_mutex;
+    std::mutex free_mutex;
+    std::jthread present_thread;
+    bool blit_supported;
+    bool use_present_thread;
+    std::size_t image_count;
+};
+
+} // namespace Vulkan

src/video_core/renderer_vulkan/vk_query_cache.cpp

@@ -66,9 +66,10 @@ void QueryPool::Reserve(std::pair<VkQueryPool, u32> query) {
     }
 }
 
-QueryCache::QueryCache(VideoCore::RasterizerInterface& rasterizer_, const Device& device_,
+QueryCache::QueryCache(VideoCore::RasterizerInterface& rasterizer_,
+                       Core::Memory::Memory& cpu_memory_, const Device& device_,
                        Scheduler& scheduler_)
-    : QueryCacheBase{rasterizer_}, device{device_}, scheduler{scheduler_},
+    : QueryCacheBase{rasterizer_, cpu_memory_}, device{device_}, scheduler{scheduler_},
       query_pools{
           QueryPool{device_, scheduler_, QueryType::SamplesPassed},
       } {}
@@ -98,8 +99,10 @@ HostCounter::HostCounter(QueryCache& cache_, std::shared_ptr<HostCounter> depend
       query{cache_.AllocateQuery(type_)}, tick{cache_.GetScheduler().CurrentTick()} {
     const vk::Device* logical = &cache.GetDevice().GetLogical();
     cache.GetScheduler().Record([logical, query = query](vk::CommandBuffer cmdbuf) {
+        const bool use_precise = Settings::IsGPULevelHigh();
         logical->ResetQueryPool(query.first, query.second, 1);
-        cmdbuf.BeginQuery(query.first, query.second, VK_QUERY_CONTROL_PRECISE_BIT);
+        cmdbuf.BeginQuery(query.first, query.second,
+                          use_precise ? VK_QUERY_CONTROL_PRECISE_BIT : 0);
     });
 }
 
@@ -112,8 +115,10 @@ void HostCounter::EndQuery() {
         [query = query](vk::CommandBuffer cmdbuf) { cmdbuf.EndQuery(query.first, query.second); });
 }
 
-u64 HostCounter::BlockingQuery() const {
-    cache.GetScheduler().Wait(tick);
+u64 HostCounter::BlockingQuery(bool async) const {
+    if (!async) {
+        cache.GetScheduler().Wait(tick);
+    }
     u64 data;
     const VkResult query_result = cache.GetDevice().GetLogical().GetQueryResults(
         query.first, query.second, 1, sizeof(data), &data, sizeof(data),

src/video_core/renderer_vulkan/vk_query_cache.h

@@ -52,7 +52,8 @@ private:
 class QueryCache final
     : public VideoCommon::QueryCacheBase<QueryCache, CachedQuery, CounterStream, HostCounter> {
 public:
-    explicit QueryCache(VideoCore::RasterizerInterface& rasterizer_, const Device& device_,
+    explicit QueryCache(VideoCore::RasterizerInterface& rasterizer_,
+                        Core::Memory::Memory& cpu_memory_, const Device& device_,
                         Scheduler& scheduler_);
     ~QueryCache();
 
@@ -83,7 +84,7 @@ public:
     void EndQuery();
 
 private:
-    u64 BlockingQuery() const override;
+    u64 BlockingQuery(bool async = false) const override;
 
     QueryCache& cache;
     const VideoCore::QueryType type;