Experimenting with scissors offsets

This reverts commit 2ea449f71a.

.gitignore

@@ -63,3 +63,6 @@ artifacts
 /install*
 vulkansdk*.exe
 *.tar.zst
+
+# Build-time generated external sources
+externals/generated

CMakeLists.txt

@@ -214,8 +214,6 @@ if(MSVC)
 endif()
 
 # TODO(crueter): Cleanup, each dep that has a bundled option should allow to choose between bundled, external, system
-cmake_dependent_option(YUZU_USE_BUNDLED_SDL3 "Download bundled SDL3 build" "${MSVC}" "NOT ANDROID" OFF)
-
 option(ENABLE_CUBEB "Enables the cubeb audio backend" ON)
 
 set(EXT_DEFAULT OFF)
@@ -223,6 +221,8 @@ if (MSVC OR ANDROID)
     set(EXT_DEFAULT ON)
 endif()
 
+option(YUZU_USE_BUNDLED_SDL3 "Download bundled SDL3 build" ${EXT_DEFAULT})
+
 # ffmpeg
 option(YUZU_USE_BUNDLED_FFMPEG "Download bundled FFmpeg" ${EXT_DEFAULT})
 cmake_dependent_option(YUZU_USE_EXTERNAL_FFMPEG "Build FFmpeg from external source" "${SOLARIS}" "NOT WIN32 AND NOT ANDROID" OFF)

cpmfile.json

@@ -186,12 +186,6 @@
         "repo": "eden-emulator/oaknut",
         "version": "v2.0.3"
     },
-    "oboe": {
-        "bundled": true,
-        "hash": "ce4011afe7345370d4ead3b891cd69a5ef224b129535783586c0ca75051d303ed446e6c7f10bde8da31fff58d6e307f1732a3ffd03b249f9ef1fd48fd4132715",
-        "repo": "google/oboe",
-        "version": "1.10.0"
-    },
     "openssl": {
         "hash": "29002ce50cb95a4f4f1d0e9d3f684401fbd4eac34203dc2eef3b6334af5d44aa46bf788b63a6f5c139c383eafb7269ae87a58a9a3ad5912903b9773e545ccc0a",
         "min_version": "3.0.0",

externals/CMakeLists.txt

@@ -136,48 +136,49 @@ if(ENABLE_CUBEB)
     endif()
 endif()
 
-if (NOT ANDROID)
+if (NOT YUZU_USE_BUNDLED_SDL3)
-    if (NOT YUZU_USE_BUNDLED_SDL3)
+    if (NOT WIN32)
-        if (NOT WIN32)
+        # Yuzu itself needs: Atomic Audio Events Joystick Haptic Sensor Threads Timers
-            # Yuzu itself needs: Atomic Audio Events Joystick Haptic Sensor Threads Timers
+        # Since 2.0.18 Atomic+Threads required for HIDAPI/libusb (see https://github.com/libsdl-org/SDL/issues/5095)
-            # Since 2.0.18 Atomic+Threads required for HIDAPI/libusb (see https://github.com/libsdl-org/SDL/issues/5095)
+        # Yuzu-cmd also needs: Video (depends on Loadso/Dlopen)
-            # Yuzu-cmd also needs: Video (depends on Loadso/Dlopen)
+        # CPUinfo also required for SDL Audio, at least until 2.28.0 (see https://github.com/libsdl-org/SDL/issues/7809)
-            # CPUinfo also required for SDL Audio, at least until 2.28.0 (see https://github.com/libsdl-org/SDL/issues/7809)
+        set(SDL_UNUSED_SUBSYSTEMS
-            set(SDL_UNUSED_SUBSYSTEMS
+            File Filesystem
-                File Filesystem
+            Locale Power Render)
-                Locale Power Render)
+        foreach(_SUB ${SDL_UNUSED_SUBSYSTEMS})
-            foreach(_SUB ${SDL_UNUSED_SUBSYSTEMS})
+          string(TOUPPER ${_SUB} _OPT)
-              string(TOUPPER ${_SUB} _OPT)
+          set(SDL_${_OPT} OFF)
-              set(SDL_${_OPT} OFF)
+        endforeach()
-            endforeach()
 
-            set(HIDAPI ON)
+        set(HIDAPI ON)
-        endif()
-
-        if (APPLE)
-            set(SDL_FILE ON)
-        endif()
-
-        AddJsonPackage(sdl3)
-    else()
-        message(STATUS "Using bundled SDL3")
-        AddJsonPackage(sdl3-ci)
     endif()
 
-    # Normalize SDL3 link target across package variants.
+    if (APPLE)
-    # Some SDL3 packages export only SDL3::SDL3-shared or SDL3::SDL3-static.
+        set(SDL_FILE ON)
-    if (NOT TARGET SDL3::SDL3)
+    endif()
-        if (TARGET SDL3::SDL3-shared)
+
-            add_library(SDL3::SDL3 ALIAS SDL3::SDL3-shared)
+    AddJsonPackage(sdl3)
-        elseif (TARGET SDL3::SDL3-static)
+else()
-            add_library(SDL3::SDL3 ALIAS SDL3::SDL3-static)
+    message(STATUS "Using bundled SDL3")
-        else()
+    AddJsonPackage(sdl3-ci)
-            message(FATAL_ERROR "SDL3 package found, but no usable SDL3 target was exported")
+
-        endif()
+    # copy sdl java sources for consumption by Gradle
+    if (ANDROID)
+        file(COPY ${SDL3_SOURCE_DIR}/java DESTINATION ${CMAKE_CURRENT_SOURCE_DIR}/generated/sdl)
     endif()
 endif()
 
-set(BUILD_SHARED_LIBS OFF)
+# Normalize SDL3 link target across package variants.
+# Some SDL3 packages export only SDL3::SDL3-shared or SDL3::SDL3-static.
+if (NOT TARGET SDL3::SDL3)
+    if (TARGET SDL3::SDL3-shared)
+        add_library(SDL3::SDL3 ALIAS SDL3::SDL3-shared)
+    elseif (TARGET SDL3::SDL3-static)
+        add_library(SDL3::SDL3 ALIAS SDL3::SDL3-static)
+    else()
+        message(FATAL_ERROR "SDL3 package found, but no usable SDL3 target was exported")
+    endif()
+endif()
 
 # SPIRV Headers
 AddJsonPackage(spirv-headers)
@@ -385,13 +386,6 @@ if (YUZU_CRASH_DUMPS AND NOT TARGET libbreakpad_client)
     endif()
 endif()
 
-# oboe
-if (ANDROID)
-    AddJsonPackage(oboe)
-
-    add_library(oboe::oboe ALIAS oboe)
-endif()
-
 if (APPLE)
     # moltenvk
     if (NOT YUZU_USE_BUNDLED_MOLTENVK)

externals/ffmpeg/CMakeLists.txt

@@ -35,10 +35,7 @@ if (NOT YUZU_USE_BUNDLED_FFMPEG)
     elseif (NOT (CMAKE_HOST_SYSTEM_PROCESSOR MATCHES CMAKE_SYSTEM_PROCESSOR
     AND CMAKE_HOST_SYSTEM_NAME MATCHES CMAKE_SYSTEM_NAME))
         string(TOLOWER "${CMAKE_SYSTEM_NAME}" FFmpeg_SYSTEM_NAME)
-        # systems FFmpeg natively supports
+        if (FFmpeg_SYSTEM_NAME STREQUAL "openorbis" OR FFmpeg_SYSTEM_NAME STREQUAL "managarm")
-        if (NOT (NETBSD OR SOLARIS OR FREEBSD
-        OR OPENBSD OR DRAGONFLYBSD OR HAIKU
-        OR LINUX OR MSYS OR WIN32 OR ANDROID OR APPLE))
             set(FFmpeg_SYSTEM_NAME "none")
         endif()
         # TODO: Can we really do better? Auto-detection? Something clever?
@@ -57,24 +54,23 @@ if (NOT YUZU_USE_BUNDLED_FFMPEG)
     endif()
 endif()
 
-if (OPENORBIS OR MANAGARM OR EMSCRIPTEN)
+if (OPENORBIS OR MANAGARM)
     # Doesn't support VA-API, don't go thru the embarrassment of trying to enable it
+    list(APPEND FFmpeg_HWACCEL_FLAGS --disable-vaapi)
+elseif (ANDROID)
     list(APPEND FFmpeg_HWACCEL_FLAGS
-        --disable-vaapi
+        --enable-mediacodec
-        --disnable-cuvid
+        --enable-jni
-        --disnable-ffnvcodec
+    )
-        --disnable-nvdec
+elseif (UNIX AND NOT DEFINED FFmpeg_IS_CROSS_COMPILING AND NOT ANDROID)
-        --disable-vulkan
-        --disable-libdrm)
-elseif (UNIX)
     find_package(PkgConfig REQUIRED)
+    pkg_check_modules(LIBVA libva)
+    pkg_check_modules(CUDA cuda)
+    pkg_check_modules(FFNVCODEC ffnvcodec)
+    pkg_check_modules(VDPAU vdpau)
+
     find_package(X11)
     if(X11_FOUND)
-        # Include X11 if possible, some APIs such as VDPAU heavily depend
-        # upon it's existence!
-        list(APPEND FFmpeg_HWACCEL_LIBRARIES ${X11_LIBRARIES})
-        list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${X11_INCLUDE_DIRS})
-        list(APPEND FFmpeg_HWACCEL_LDFLAGS ${X11_LDFLAGS})
         if (NOT APPLE)
             # In Solaris needs explicit linking for ffmpeg which links to /lib/amd64/libX11.so
             if(SOLARIS)
@@ -83,37 +79,32 @@ elseif (UNIX)
                     "${CMAKE_SYSROOT}/usr/lib/xorg/amd64/libdrm.so")
             else()
                 pkg_check_modules(LIBDRM libdrm REQUIRED)
-                list(APPEND FFmpeg_HWACCEL_LIBRARIES ${LIBDRM_LIBRARIES})
+                list(APPEND FFmpeg_HWACCEL_LIBRARIES
-                list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${LIBDRM_INCLUDE_DIRS})
+                    ${LIBDRM_LIBRARIES})
-                list(APPEND FFmpeg_HWACCEL_LDFLAGS ${LIBDRM_LDFLAGS})
+                list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS
+                    ${LIBDRM_INCLUDE_DIRS})
             endif()
-            list(APPEND FFmpeg_HWACCEL_FLAGS --enable-libdrm)
+            list(APPEND FFmpeg_HWACCEL_FLAGS
+                --enable-libdrm)
         endif()
-
-        pkg_check_modules(LIBVA libva)
         if(LIBVA_FOUND)
-            list(APPEND FFmpeg_HWACCEL_LIBRARIES ${LIBVA_LIBRARIES})
-            list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${LIBVA_INCLUDE_DIRS})
-            list(APPEND FFmpeg_HWACCEL_LDFLAGS ${LIBVA_LDFLAGS})
-
-            # Logically, they can only exist if libva itself exists
             pkg_check_modules(LIBVA-DRM libva-drm REQUIRED)
             pkg_check_modules(LIBVA-X11 libva-x11 REQUIRED)
-            if (LIBVA-DRM_FOUND)
+            list(APPEND FFmpeg_HWACCEL_LIBRARIES
-                list(APPEND FFmpeg_HWACCEL_LIBRARIES ${LIBVA-DRM_LIBRARIES})
+                ${X11_LIBRARIES}
-                list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${LIBVA-DRM_INCLUDE_DIRS})
+                ${LIBVA-DRM_LIBRARIES}
-                list(APPEND FFmpeg_HWACCEL_LDFLAGS ${LIBVA-DRM_LDFLAGS})
+                ${LIBVA-X11_LIBRARIES}
-            endif()
+                ${LIBVA_LIBRARIES})
-            if (LIBVA-X11_FOUND)
-                list(APPEND FFmpeg_HWACCEL_LIBRARIES ${LIBVA-X11_LIBRARIES})
-                list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${LIBVA-X11_INCLUDE_DIRS})
-                list(APPEND FFmpeg_HWACCEL_LDFLAGS ${LIBVA-X11_LDFLAGS})
-            endif()
-
             list(APPEND FFmpeg_HWACCEL_FLAGS
                 --enable-hwaccel=h264_vaapi
                 --enable-hwaccel=vp8_vaapi
                 --enable-hwaccel=vp9_vaapi)
+            list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS
+                ${X11_INCLUDE_DIRS}
+                ${LIBVA-DRM_INCLUDE_DIRS}
+                ${LIBVA-X11_INCLUDE_DIRS}
+                ${LIBVA_INCLUDE_DIRS}
+            )
             message(STATUS "ffmpeg: va-api libraries version ${LIBVA_VERSION} found")
         else()
             list(APPEND FFmpeg_HWACCEL_FLAGS --disable-vaapi)
@@ -124,12 +115,6 @@ elseif (UNIX)
         message(WARNING "ffmpeg: X11 libraries not found, disabling VA-API...")
     endif()
 
-    pkg_check_modules(CUDA cuda)
-    pkg_check_modules(FFNVCODEC ffnvcodec)
-    pkg_check_modules(VDPAU vdpau)
-    pkg_check_modules(VULKAN vulkan)
-    find_package(spirv-headers)
-
     if (FFNVCODEC_FOUND)
         list(APPEND FFmpeg_HWACCEL_FLAGS
             --enable-cuvid
@@ -137,7 +122,8 @@ elseif (UNIX)
             --enable-nvdec
             --enable-hwaccel=h264_nvdec
             --enable-hwaccel=vp8_nvdec
-            --enable-hwaccel=vp9_nvdec)
+            --enable-hwaccel=vp9_nvdec
+        )
         list(APPEND FFmpeg_HWACCEL_LIBRARIES ${FFNVCODEC_LIBRARIES})
         list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${FFNVCODEC_INCLUDE_DIRS})
         list(APPEND FFmpeg_HWACCEL_LDFLAGS ${FFNVCODEC_LDFLAGS})
@@ -158,7 +144,8 @@ elseif (UNIX)
         list(APPEND FFmpeg_HWACCEL_FLAGS
             --enable-vdpau
             --enable-hwaccel=h264_vdpau
-            --enable-hwaccel=vp9_vdpau)
+            --enable-hwaccel=vp9_vdpau
+        )
         list(APPEND FFmpeg_HWACCEL_LIBRARIES ${VDPAU_LIBRARIES})
         list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${VDPAU_INCLUDE_DIRS})
         list(APPEND FFmpeg_HWACCEL_LDFLAGS ${VDPAU_LDFLAGS})
@@ -167,24 +154,6 @@ elseif (UNIX)
         list(APPEND FFmpeg_HWACCEL_FLAGS --disable-vdpau)
         message(WARNING "ffmpeg: libvdpau-dev not found, disabling Video Decode and Presentation API for Unix (VDPAU)...")
     endif()
-
-    if (VULKAN_FOUND AND NOT APPLE)
-        list(APPEND FFmpeg_HWACCEL_FLAGS
-            --enable-vulkan
-            --enable-hwaccel=h264_vulkan
-            --enable-hwaccel=vp9_vulkan)
-        list(APPEND FFmpeg_HWACCEL_LIBRARIES ${VULKAN_LIBRARIES})
-        list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${VULKAN_INCLUDE_DIRS})
-        list(APPEND FFmpeg_HWACCEL_LDFLAGS ${VULKAN_LDFLAGS})
-        message(STATUS "ffmpeg: vulkan libraries version ${VULKAN_VERSION} found")
-    endif()
-
-    if (SPIRV_HEADERS_FOUND)
-        list(APPEND FFmpeg_HWACCEL_LIBRARIES ${SPIRV_HEADERS_LIBRARIES})
-        list(APPEND FFmpeg_HWACCEL_INCLUDE_DIRS ${SPIRV_HEADERS_INCLUDE_DIRS})
-        list(APPEND FFmpeg_HWACCEL_LDFLAGS ${SPIRV_HEADERS_LDFLAGS})
-        message(STATUS "ffmpeg: spirv-headers libraries version ${SPIRV_HEADERS_VERSION} found")
-    endif()
 endif()
 
 if (OPENORBIS)
@@ -193,28 +162,20 @@ if (OPENORBIS)
         -lSceUserService
         -lSceSysmodule
         -lSceNet
-        -lSceLibcInternal)
+        -lSceLibcInternal
+    )
     list(APPEND FFmpeg_CROSS_COMPILE_FLAGS
         --disable-pthreads
         --extra-cflags=${CMAKE_SYSROOT}/usr/include
         --extra-cxxflags=${CMAKE_SYSROOT}/usr/include
-        --extra-libs="${FFmpeg_CROSS_COMPILE_LIBS}")
+        --extra-libs="${FFmpeg_CROSS_COMPILE_LIBS}"
+    )
 elseif (MANAGARM)
     # Required for proper stuff
     list(APPEND FFmpeg_CROSS_COMPILE_FLAGS
         --disable-pthreads
-        --extra-libs="${FFmpeg_CROSS_COMPILE_LIBS}")
+        --extra-libs="${FFmpeg_CROSS_COMPILE_LIBS}"
-endif()
+    )
-
-# Usually used by Emscripten
-if (DEFINED CMAKE_AR)
-    list(APPEND FFmpeg_CROSS_COMPILE_FLAGS --ar=${CMAKE_AR})
-endif()
-if (DEFINED CMAKE_NM)
-    list(APPEND FFmpeg_CROSS_COMPILE_FLAGS --nm=${CMAKE_NM})
-endif()
-if (DEFINED CMAKE_RANLIB)
-    list(APPEND FFmpeg_CROSS_COMPILE_FLAGS --ranlib=${CMAKE_RANLIB})
 endif()
 
 if (YUZU_USE_BUNDLED_FFMPEG)
@@ -289,14 +250,8 @@ else()
 
     # `configure` parameters builds only exactly what yuzu needs from FFmpeg
     # `--disable-vdpau` is needed to avoid linking issues
-    set(FFmpeg_CC ${CMAKE_C_COMPILER_LAUNCHER} ${CMAKE_C_COMPILER})
+    set(FFmpeg_CC ${CMAKE_C_COMPILER_LAUNCHER}  ${CMAKE_C_COMPILER})
-    set(FFmpeg_CXX ${CMAKE_CXX_COMPILER_LAUNCHER} ${CMAKE_CXX_COMPILER})
+    set(FFmpeg_CXX ${CMAKE_CXX_COMPILER_LAUNCHER}  ${CMAKE_CXX_COMPILER})
-    if (FFmpeg_IS_CROSS_COMPILING)
-        list(APPEND FFmpeg_CROSS_COMPILE_FLAGS --ld=${CMAKE_LINKER})
-    endif ()
-    # TODO: This is a bit fragile, it depends on FFmpeg_HWACCEL_LDFLAGS being a list
-    # whereas CMAKE_C_FLAGS (and likewise) is meant to NOT be a list, but rather
-    # an elongated piece of string
     add_custom_command(
         OUTPUT
             ${FFmpeg_MAKEFILE}
@@ -310,7 +265,6 @@ else()
                 --disable-ffprobe
                 --disable-network
                 --disable-swresample
-                --disable-iconv
                 --enable-decoder=h264
                 --enable-decoder=vp8
                 --enable-decoder=vp9
@@ -318,9 +272,10 @@ else()
                 --enable-pic
                 --cc=${FFmpeg_CC}
                 --cxx=${FFmpeg_CXX}
+                --ld=${CMAKE_LINKER}
                 --extra-cflags=${CMAKE_C_FLAGS}
                 --extra-cxxflags=${CMAKE_CXX_FLAGS}
-                --extra-ldflags="${FFmpeg_HWACCEL_LDFLAGS}"
+                --extra-ldflags=${CMAKE_C_LINK_FLAGS}
                 ${FFmpeg_HWACCEL_FLAGS}
                 ${FFmpeg_CROSS_COMPILE_FLAGS}
         WORKING_DIRECTORY
@@ -328,30 +283,31 @@ else()
     )
     unset(FFmpeg_CC)
     unset(FFmpeg_CXX)
-    unset(FFmpeg_LD)
     unset(FFmpeg_HWACCEL_FLAGS)
     unset(FFmpeg_CROSS_COMPILE_FLAGS)
 
     # Workaround for Ubuntu 18.04's older version of make not being able to call make as a child
     # with context of the jobserver. Also helps ninja users.
-    cmake_host_system_information(RESULT SYSTEM_THREADS QUERY NUMBER_OF_LOGICAL_CORES)
+    execute_process(
+        COMMAND
+            nproc
+        OUTPUT_VARIABLE
+            SYSTEM_THREADS)
 
     set(FFmpeg_BUILD_LIBRARIES ${FFmpeg_LIBRARIES})
 
     # BSD make or Solaris make don't support ffmpeg make-j8
-    if (DEFINED SYSTEM_THREADS AND (LINUX OR ANDROID OR APPLE OR WIN32))
+    if (LINUX OR ANDROID OR APPLE OR WIN32 OR FREEBSD)
         set(FFmpeg_MAKE_ARGS -j${SYSTEM_THREADS})
     else()
         set(FFmpeg_MAKE_ARGS "")
     endif()
-    # Most distros use 'gmake', Arch uses 'make'
-    find_program(MAKE gmake NAMES gmake make REQUIRED)
 
     add_custom_command(
         OUTPUT
             ${FFmpeg_BUILD_LIBRARIES}
         COMMAND
-            ${MAKE} ${FFmpeg_MAKE_ARGS}
+            gmake ${FFmpeg_MAKE_ARGS}
         WORKING_DIRECTORY
             ${FFmpeg_BUILD_DIR}
     )

src/android/app/build.gradle.kts

@@ -272,6 +272,13 @@ android {
         resValue("string", "app_name_suffixed", "$currentName$suffix")
         resValue("string", "app_name", "Eden$suffix")
     }
+
+    sourceSets {
+        named("main") {
+            java.srcDir("${edenDir}/externals/generated/sdl/java")
+            kotlin.srcDir("${edenDir}/externals/generated/sdl/java")
+        }
+    }
 }
 
 idea {

src/android/app/proguard-rules.pro

@@ -9,6 +9,9 @@
 -keep class org.ini4j.spi.IniBuilder
 -keep class org.ini4j.spi.IniFormatter
 
+-keep class org.libsdl.app.** { *; }
+-keepclassmembers class org.libsdl.app.** { *; }
+
 # Suppress warnings for R8
 -dontwarn org.bouncycastle.jsse.BCSSLParameters
 -dontwarn org.bouncycastle.jsse.BCSSLSocket

src/android/app/src/main/java/org/yuzu/yuzu_emu/NativeLibrary.kt

@@ -16,6 +16,7 @@ import android.widget.TextView
 import androidx.annotation.Keep
 import androidx.core.net.toUri
 import com.google.android.material.dialog.MaterialAlertDialogBuilder
+import org.libsdl.app.SDL
 import java.lang.ref.WeakReference
 import org.yuzu.yuzu_emu.activities.EmulationActivity
 import org.yuzu.yuzu_emu.fragments.CoreErrorDialogFragment
@@ -52,6 +53,8 @@ object NativeLibrary {
     init {
         try {
             System.loadLibrary("yuzu-android")
+            SDL.setupJNI()
+            initJvm()
         } catch (ex: UnsatisfiedLinkError) {
             error("[NativeLibrary] $ex")
         }
@@ -367,6 +370,7 @@ object NativeLibrary {
         NetPlayManager.clearChat()
     }
 
+    external fun initJvm()
     external fun initMultiplayer()
 
     @Keep

src/android/app/src/main/jni/CMakeLists.txt

@@ -22,7 +22,7 @@ add_library(yuzu-android SHARED
 set_property(TARGET yuzu-android PROPERTY IMPORTED_LOCATION ${FFmpeg_LIBRARY_DIR})
 
 target_link_libraries(yuzu-android PRIVATE audio_core common core input_common frontend_common video_core)
-target_link_libraries(yuzu-android PRIVATE android camera2ndk EGL glad jnigraphics log)
+target_link_libraries(yuzu-android PRIVATE android camera2ndk EGL glad jnigraphics log GLESv2)
 if (ARCHITECTURE_arm64)
     target_link_libraries(yuzu-android PRIVATE adrenotools)
 endif()

src/android/app/src/main/jni/native.cpp

@@ -1730,6 +1730,15 @@ jint Java_org_yuzu_yuzu_1emu_NativeLibrary_loadAmiibo(JNIEnv* env, jobject jobj,
     return static_cast<jint>(info);
 }
 
+extern "C" JNIEXPORT void JNICALL
+Java_org_yuzu_yuzu_1emu_NativeLibrary_initJvm(JNIEnv *env, jclass clazz) {
+    JavaVM *vm;
+    if (env->GetJavaVM(&vm) != JNI_OK)
+        return;
+
+    Common::Android::Initialize(vm, env);
+}
+
 JNIEXPORT void JNICALL
 Java_org_yuzu_yuzu_1emu_NativeLibrary_initMultiplayer(
         JNIEnv* env, [[maybe_unused]] jobject obj) {

src/android/app/src/main/res/values-fa/strings.xml

@@ -800,7 +800,6 @@
     <string name="theme_mode_dark">تاریک</string>
 
     <!-- Audio output engines -->
-    <string name="oboe">oboe</string>
     <string name="cubeb">cubeb</string>
 
     <!-- Anisotropic filtering options -->

src/android/app/src/main/res/values/arrays.xml

@@ -465,13 +465,13 @@
 
     <string-array name="outputEngineEntries">
         <item>@string/auto</item>
-        <item>@string/oboe</item>
+        <item>@string/sdl3</item>
         <item>@string/cubeb</item>
         <item>@string/string_null</item>
     </string-array>
     <integer-array name="outputEngineValues">
         <item>0</item>
-        <item>4</item>
+        <item>2</item>
         <item>1</item>
         <item>3</item>
     </integer-array>

src/android/app/src/main/res/values/strings.xml

@@ -1240,7 +1240,7 @@
     <string name="theme_mode_dark">Dark</string>
 
     <!-- Audio output engines -->
-    <string name="oboe" translatable="false">oboe</string>
+    <string name="sdl3" translatable="false">SDL3</string>
     <string name="cubeb" translatable="false">cubeb</string>
 
     <!-- Anisotropic filtering options -->

src/audio_core/CMakeLists.txt

@@ -238,20 +238,11 @@ if (ENABLE_CUBEB)
     target_compile_definitions(audio_core PRIVATE HAVE_CUBEB=1)
 endif()
 
-if(ANDROID)
+target_sources(audio_core PRIVATE
-    target_sources(audio_core PRIVATE
+    sink/sdl3_sink.cpp
-        sink/oboe_sink.cpp
+    sink/sdl3_sink.h)
-        sink/oboe_sink.h)
 
-    target_link_libraries(audio_core PRIVATE oboe)
+target_link_libraries(audio_core PRIVATE SDL3::SDL3)
-    target_compile_definitions(audio_core PUBLIC HAVE_OBOE)
+target_compile_definitions(audio_core PRIVATE HAVE_SDL3)
-else()
-    target_sources(audio_core PRIVATE
-        sink/sdl3_sink.cpp
-        sink/sdl3_sink.h)
-
-    target_link_libraries(audio_core PRIVATE SDL3::SDL3)
-    target_compile_definitions(audio_core PRIVATE HAVE_SDL3)
-endif()
 
 create_target_directory_groups(audio_core)

src/audio_core/sink/oboe_sink.cpp

@@ -1,230 +0,0 @@
-// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
-// SPDX-License-Identifier: GPL-3.0-or-later
-
-// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
-// SPDX-License-Identifier: GPL-2.0-or-later
-
-#include <span>
-#include <vector>
-
-#include <oboe/Oboe.h>
-
-#include "audio_core/common/common.h"
-#include "audio_core/sink/oboe_sink.h"
-#include "audio_core/sink/sink_stream.h"
-#include "common/logging.h"
-#include "common/scope_exit.h"
-#include "core/core.h"
-
-namespace AudioCore::Sink {
-
-class OboeSinkStream final : public SinkStream,
-                             public oboe::AudioStreamDataCallback,
-                             public oboe::AudioStreamErrorCallback {
-public:
-    explicit OboeSinkStream(Core::System& system_, StreamType type_, const std::string& name_,
-                            u32 system_channels_)
-        : SinkStream(system_, type_) {
-        name = name_;
-        system_channels = system_channels_;
-
-        this->OpenStream();
-    }
-
-    ~OboeSinkStream() override {
-        LOG_INFO(Audio_Sink, "Destroyed Oboe stream");
-    }
-
-    void Finalize() override {
-        this->Stop();
-        m_stream.reset();
-    }
-
-    void Start(bool resume = false) override {
-        if (!m_stream || !paused) {
-            return;
-        }
-
-        paused = false;
-
-        if (m_stream->start() != oboe::Result::OK) {
-            LOG_CRITICAL(Audio_Sink, "Error starting Oboe stream");
-        }
-    }
-
-    void Stop() override {
-        if (!m_stream || paused) {
-            return;
-        }
-
-        this->SignalPause();
-
-        if (m_stream->stop() != oboe::Result::OK) {
-            LOG_CRITICAL(Audio_Sink, "Error stopping Oboe stream");
-        }
-    }
-
-public:
-    static s32 QueryChannelCount(oboe::Direction direction) {
-        std::shared_ptr<oboe::AudioStream> temp_stream;
-        oboe::AudioStreamBuilder builder;
-
-        const auto result = ConfigureBuilder(builder, direction)->openStream(temp_stream);
-        if (result == oboe::Result::OK) {
-            return temp_stream->getChannelCount() >= 6 ? 6 : 2;
-        }
-
-        LOG_ERROR(Audio_Sink, "Failed to open {} stream. Using default channel count 2",
-                  direction == oboe::Direction::Output ? "output" : "input");
-        return 2;
-    }
-
-protected:
-    oboe::DataCallbackResult onAudioReady(oboe::AudioStream*, void* audio_data,
-                                          s32 num_buffer_frames) override {
-        const size_t num_channels = this->GetDeviceChannels();
-        const size_t frame_size = num_channels;
-        const size_t num_frames = static_cast<size_t>(num_buffer_frames);
-
-        if (type == StreamType::In) {
-            std::span<const s16> input_buffer{reinterpret_cast<const s16*>(audio_data),
-                                              num_frames * frame_size};
-            this->ProcessAudioIn(input_buffer, num_frames);
-        } else {
-            std::span<s16> output_buffer{reinterpret_cast<s16*>(audio_data),
-                                         num_frames * frame_size};
-            this->ProcessAudioOutAndRender(output_buffer, num_frames);
-        }
-
-        return oboe::DataCallbackResult::Continue;
-    }
-
-    void onErrorAfterClose(oboe::AudioStream*, oboe::Result) override {
-        LOG_INFO(Audio_Sink, "Audio stream closed, reinitializing");
-
-        if (this->OpenStream()) {
-            m_stream->start();
-        }
-    }
-
-private:
-    static oboe::AudioStreamBuilder* ConfigureBuilder(oboe::AudioStreamBuilder& builder,
-                                                      oboe::Direction direction) {
-        // TODO: investigate callback delay issues when using AAudio
-        return builder.setPerformanceMode(oboe::PerformanceMode::LowLatency)
-            ->setAudioApi(oboe::AudioApi::OpenSLES)
-            ->setDirection(direction)
-            ->setSampleRate(TargetSampleRate)
-            ->setSampleRateConversionQuality(oboe::SampleRateConversionQuality::High)
-            ->setFormat(oboe::AudioFormat::I16)
-            ->setFormatConversionAllowed(true)
-            ->setUsage(oboe::Usage::Game)
-            ->setBufferCapacityInFrames(TargetSampleCount * 2);
-    }
-
-    bool OpenStream() {
-        const auto direction = [&]() {
-            switch (type) {
-            case StreamType::In:
-                return oboe::Direction::Input;
-            case StreamType::Out:
-            case StreamType::Render:
-                return oboe::Direction::Output;
-            default:
-                ASSERT(false);
-                return oboe::Direction::Output;
-            }
-        }();
-
-        const auto expected_channels = QueryChannelCount(direction);
-        const auto expected_mask = [&]() {
-            switch (expected_channels) {
-            case 1:
-                return oboe::ChannelMask::Mono;
-            case 2:
-                return oboe::ChannelMask::Stereo;
-            case 6:
-                return oboe::ChannelMask::CM5Point1;
-            default:
-                ASSERT(false);
-                return oboe::ChannelMask::Unspecified;
-            }
-        }();
-
-        oboe::AudioStreamBuilder builder;
-        const auto result = ConfigureBuilder(builder, direction)
-                                ->setChannelCount(expected_channels)
-                                ->setChannelMask(expected_mask)
-                                ->setChannelConversionAllowed(true)
-                                ->setDataCallback(this)
-                                ->setErrorCallback(this)
-                                ->openStream(m_stream);
-        ASSERT(result == oboe::Result::OK);
-        return result == oboe::Result::OK && this->SetStreamProperties();
-    }
-
-    bool SetStreamProperties() {
-        ASSERT(m_stream);
-
-        m_stream->setBufferSizeInFrames(TargetSampleCount * 2);
-        device_channels = m_stream->getChannelCount();
-
-        const auto sample_rate = m_stream->getSampleRate();
-        const auto buffer_capacity = m_stream->getBufferCapacityInFrames();
-        const auto stream_backend =
-            m_stream->getAudioApi() == oboe::AudioApi::AAudio ? "AAudio" : "OpenSLES";
-
-        LOG_INFO(Audio_Sink, "Opened Oboe {} stream with {} channels sample rate {} capacity {}",
-                 stream_backend, device_channels, sample_rate, buffer_capacity);
-
-        return true;
-    }
-
-    std::shared_ptr<oboe::AudioStream> m_stream{};
-};
-
-OboeSink::OboeSink() {
-    // TODO: This is not generally knowable
-    // The channel count is distinct based on direction and can change
-    device_channels = OboeSinkStream::QueryChannelCount(oboe::Direction::Output);
-}
-
-OboeSink::~OboeSink() = default;
-
-SinkStream* OboeSink::AcquireSinkStream(Core::System& system, u32 system_channels,
-                                        const std::string& name, StreamType type) {
-    SinkStreamPtr& stream = sink_streams.emplace_back(
-        std::make_unique<OboeSinkStream>(system, type, name, system_channels));
-
-    return stream.get();
-}
-
-void OboeSink::CloseStream(SinkStream* to_remove) {
-    sink_streams.remove_if([&](auto& stream) { return stream.get() == to_remove; });
-}
-
-void OboeSink::CloseStreams() {
-    sink_streams.clear();
-}
-
-f32 OboeSink::GetDeviceVolume() const {
-    if (sink_streams.empty()) {
-        return 1.0f;
-    }
-
-    return sink_streams.front()->GetDeviceVolume();
-}
-
-void OboeSink::SetDeviceVolume(f32 volume) {
-    for (auto& stream : sink_streams) {
-        stream->SetDeviceVolume(volume);
-    }
-}
-
-void OboeSink::SetSystemVolume(f32 volume) {
-    for (auto& stream : sink_streams) {
-        stream->SetSystemVolume(volume);
-    }
-}
-
-} // namespace AudioCore::Sink

src/audio_core/sink/oboe_sink.h

@@ -1,75 +0,0 @@
-// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
-// SPDX-License-Identifier: GPL-2.0-or-later
-
-#pragma once
-
-#include <list>
-#include <string>
-
-#include "audio_core/sink/sink.h"
-
-namespace Core {
-class System;
-}
-
-namespace AudioCore::Sink {
-class SinkStream;
-
-class OboeSink final : public Sink {
-public:
-    explicit OboeSink();
-    ~OboeSink() override;
-
-    /**
-     * Create a new sink stream.
-     *
-     * @param system          - Core system.
-     * @param system_channels - Number of channels the audio system expects.
-     *                          May differ from the device's channel count.
-     * @param name            - Name of this stream.
-     * @param type            - Type of this stream, render/in/out.
-     *
-     * @return A pointer to the created SinkStream
-     */
-    SinkStream* AcquireSinkStream(Core::System& system, u32 system_channels,
-                                  const std::string& name, StreamType type) override;
-
-    /**
-     * Close a given stream.
-     *
-     * @param stream - The stream to close.
-     */
-    void CloseStream(SinkStream* stream) override;
-
-    /**
-     * Close all streams.
-     */
-    void CloseStreams() override;
-
-    /**
-     * Get the device volume. Set from calls to the IAudioDevice service.
-     *
-     * @return Volume of the device.
-     */
-    f32 GetDeviceVolume() const override;
-
-    /**
-     * Set the device volume. Set from calls to the IAudioDevice service.
-     *
-     * @param volume - New volume of the device.
-     */
-    void SetDeviceVolume(f32 volume) override;
-
-    /**
-     * Set the system volume. Comes from the audio system using this stream.
-     *
-     * @param volume - New volume of the system.
-     */
-    void SetSystemVolume(f32 volume) override;
-
-private:
-    /// List of streams managed by this sink
-    std::list<SinkStreamPtr> sink_streams{};
-};
-
-} // namespace AudioCore::Sink

src/audio_core/sink/sink_details.cpp

@@ -10,9 +10,6 @@
 #include <vector>
 
 #include "audio_core/sink/sink_details.h"
-#ifdef HAVE_OBOE
-#include "audio_core/sink/oboe_sink.h"
-#endif
 #ifdef HAVE_CUBEB
 #include "audio_core/sink/cubeb_sink.h"
 #endif
@@ -43,24 +40,8 @@ struct SinkDetails {
     /// SuitableFn is_suitable; // REVERTED FOR LatencyFn latency ABOVE - DIABLO 3 FIX
 };
 
-// NOTE TO PROBABLY FIX LATER FOR ANDROID - the return value "0u" for the first HAVE_OBOE
-// section below was just copied from the null section so there's a somewhat valid value
-// being returned, since the previous "true" value probably isn't compatible with the
-// previous EA-3833 code. (HAVE_OBOE was introduced in a later release.) Eventually need
-// to change "0u" for something else directly from the oboe_sink.cpp functions.
-
 // sink_details is ordered in terms of desirability, with the best choice at the top.
 constexpr SinkDetails sink_details[] = {
-#ifdef HAVE_OBOE
-    SinkDetails{
-        Settings::AudioEngine::Oboe,
-        [](std::string_view device_id) -> std::unique_ptr<Sink> {
-            return std::make_unique<OboeSink>();
-        },
-        [](bool capture) { return std::vector<std::string>{"Default"}; },
-        []() { return 0u; },
-    },
-#endif
 #ifdef HAVE_CUBEB
     SinkDetails{
         Settings::AudioEngine::Cubeb,

src/common/android/android_common.cpp

@@ -1,7 +1,7 @@
 // SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
 // SPDX-License-Identifier: GPL-2.0-or-later
 
-// SPDX-FileCopyrightText: 2025 Eden Emulator Project
+// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
 // SPDX-License-Identifier: GPL-3.0-or-later
 
 #include "android_common.h"
@@ -51,7 +51,7 @@ jstring ToJString(JNIEnv* env, std::u16string_view str) {
 }
 
 double GetJDouble(JNIEnv* env, jobject jdouble) {
-    return env->GetDoubleField(jdouble, GetDoubleValueField());
+    return env->CallDoubleMethod(jdouble, GetDoubleValueMethod());
 }
 
 jobject ToJDouble(JNIEnv* env, double value) {
@@ -59,7 +59,7 @@ jobject ToJDouble(JNIEnv* env, double value) {
 }
 
 s32 GetJInteger(JNIEnv* env, jobject jinteger) {
-    return env->GetIntField(jinteger, GetIntegerValueField());
+    return env->CallIntMethod(jinteger, GetIntegerValueMethod());
 }
 
 jobject ToJInteger(JNIEnv* env, s32 value) {
@@ -67,7 +67,7 @@ jobject ToJInteger(JNIEnv* env, s32 value) {
 }
 
 bool GetJBoolean(JNIEnv* env, jobject jboolean) {
-    return env->GetBooleanField(jboolean, GetBooleanValueField());
+    return env->CallBooleanMethod(jboolean, GetBooleanValueMethod());
 }
 
 jobject ToJBoolean(JNIEnv* env, bool value) {

src/common/android/id_cache.cpp

@@ -63,15 +63,15 @@ static jfieldID s_patch_title_id_field;
 
 static jclass s_double_class;
 static jmethodID s_double_constructor;
-static jfieldID s_double_value_field;
+static jmethodID s_double_value_method;
 
 static jclass s_integer_class;
 static jmethodID s_integer_constructor;
-static jfieldID s_integer_value_field;
+static jmethodID s_integer_value_method;
 
 static jclass s_boolean_class;
 static jmethodID s_boolean_constructor;
-static jfieldID s_boolean_value_field;
+static jmethodID s_boolean_value_method;
 
 static jclass s_player_input_class;
 static jmethodID s_player_input_constructor;
@@ -103,7 +103,6 @@ static jmethodID s_clear_chat;
 static constexpr jint JNI_VERSION = JNI_VERSION_1_6;
 
 namespace Common::Android {
-
     JNIEnv *GetEnvForThread() {
         thread_local static struct OwnedEnv {
             OwnedEnv() {
@@ -299,8 +298,8 @@ namespace Common::Android {
         return s_double_constructor;
     }
 
-    jfieldID GetDoubleValueField() {
+    jmethodID GetDoubleValueMethod() {
-        return s_double_value_field;
+        return s_double_value_method;
     }
 
     jclass GetIntegerClass() {
@@ -311,8 +310,8 @@ namespace Common::Android {
         return s_integer_constructor;
     }
 
-    jfieldID GetIntegerValueField() {
+    jmethodID GetIntegerValueMethod() {
-        return s_integer_value_field;
+        return s_integer_value_method;
     }
 
     jclass GetBooleanClass() {
@@ -323,8 +322,8 @@ namespace Common::Android {
         return s_boolean_constructor;
     }
 
-    jfieldID GetBooleanValueField() {
+    jmethodID GetBooleanValueMethod() {
-        return s_boolean_value_field;
+        return s_boolean_value_method;
     }
 
     jclass GetPlayerInputClass() {
@@ -429,188 +428,6 @@ namespace Common::Android {
 
     jint InitFFmpegOnLoad(JavaVM *vm);
 
-    jint JNI_OnLoad(JavaVM *vm, void *reserved) {
-        s_java_vm = vm;
-        InitFFmpegOnLoad(vm);
-
-        JNIEnv *env;
-        if (vm->GetEnv(reinterpret_cast<void **>(&env), JNI_VERSION) != JNI_OK)
-            return JNI_ERR;
-
-        // Initialize Java classes
-        const jclass native_library_class = env->FindClass("org/yuzu/yuzu_emu/NativeLibrary");
-        s_native_library_class = reinterpret_cast<jclass>(env->NewGlobalRef(native_library_class));
-        s_disk_cache_progress_class = reinterpret_cast<jclass>(env->NewGlobalRef(
-                env->FindClass("org/yuzu/yuzu_emu/disk_shader_cache/DiskShaderCacheProgress")));
-        s_load_callback_stage_class = reinterpret_cast<jclass>(env->NewGlobalRef(env->FindClass(
-                "org/yuzu/yuzu_emu/disk_shader_cache/DiskShaderCacheProgress$LoadCallbackStage")));
-
-        const jclass game_dir_class = env->FindClass("org/yuzu/yuzu_emu/model/GameDir");
-        s_game_dir_class = reinterpret_cast<jclass>(env->NewGlobalRef(game_dir_class));
-        s_game_dir_constructor = env->GetMethodID(game_dir_class, "<init>",
-                                                  "(Ljava/lang/String;Z)V");
-        env->DeleteLocalRef(game_dir_class);
-
-        // Initialize methods
-        s_exit_emulation_activity =
-                env->GetStaticMethodID(s_native_library_class, "exitEmulationActivity", "(I)V");
-        s_disk_cache_load_progress =
-                env->GetStaticMethodID(s_disk_cache_progress_class, "loadProgress", "(III)V");
-        s_copy_to_storage = env->GetStaticMethodID(s_native_library_class, "copyFileToStorage",
-                                                   "(Ljava/lang/String;Ljava/lang/String;)Z");
-        s_file_exists = env->GetStaticMethodID(s_native_library_class, "exists",
-                                               "(Ljava/lang/String;)Z");
-        s_file_extension = env->GetStaticMethodID(s_native_library_class, "getFileExtension",
-                                               "(Ljava/lang/String;)Ljava/lang/String;");
-        s_on_emulation_started =
-                env->GetStaticMethodID(s_native_library_class, "onEmulationStarted", "()V");
-        s_on_emulation_stopped =
-                env->GetStaticMethodID(s_native_library_class, "onEmulationStopped", "(I)V");
-        s_on_program_changed =
-                env->GetStaticMethodID(s_native_library_class, "onProgramChanged", "(I)V");
-
-        const jclass game_class = env->FindClass("org/yuzu/yuzu_emu/model/Game");
-        s_game_class = reinterpret_cast<jclass>(env->NewGlobalRef(game_class));
-        s_game_constructor = env->GetMethodID(game_class, "<init>",
-                                              "(Ljava/lang/String;Ljava/lang/String;Ljava/lang/"
-                                              "String;Ljava/lang/String;Ljava/lang/String;Z)V");
-        s_game_title_field = env->GetFieldID(game_class, "title", "Ljava/lang/String;");
-        s_game_path_field = env->GetFieldID(game_class, "path", "Ljava/lang/String;");
-        s_game_program_id_field = env->GetFieldID(game_class, "programId", "Ljava/lang/String;");
-        s_game_developer_field = env->GetFieldID(game_class, "developer", "Ljava/lang/String;");
-        s_game_version_field = env->GetFieldID(game_class, "version", "Ljava/lang/String;");
-        s_game_is_homebrew_field = env->GetFieldID(game_class, "isHomebrew", "Z");
-        env->DeleteLocalRef(game_class);
-
-        const jclass string_class = env->FindClass("java/lang/String");
-        s_string_class = reinterpret_cast<jclass>(env->NewGlobalRef(string_class));
-        env->DeleteLocalRef(string_class);
-
-        const jclass pair_class = env->FindClass("kotlin/Pair");
-        s_pair_class = reinterpret_cast<jclass>(env->NewGlobalRef(pair_class));
-        s_pair_constructor =
-                env->GetMethodID(pair_class, "<init>", "(Ljava/lang/Object;Ljava/lang/Object;)V");
-        s_pair_first_field = env->GetFieldID(pair_class, "first", "Ljava/lang/Object;");
-        s_pair_second_field = env->GetFieldID(pair_class, "second", "Ljava/lang/Object;");
-        env->DeleteLocalRef(pair_class);
-
-        const jclass overlay_control_data_class =
-                env->FindClass("org/yuzu/yuzu_emu/overlay/model/OverlayControlData");
-        s_overlay_control_data_class =
-                reinterpret_cast<jclass>(env->NewGlobalRef(overlay_control_data_class));
-        s_overlay_control_data_constructor =
-                env->GetMethodID(overlay_control_data_class, "<init>",
-                                 "(Ljava/lang/String;ZLkotlin/Pair;Lkotlin/Pair;Lkotlin/Pair;F)V");
-        s_overlay_control_data_id_field =
-                env->GetFieldID(overlay_control_data_class, "id", "Ljava/lang/String;");
-        s_overlay_control_data_enabled_field =
-                env->GetFieldID(overlay_control_data_class, "enabled", "Z");
-        s_overlay_control_data_landscape_position_field =
-                env->GetFieldID(overlay_control_data_class, "landscapePosition", "Lkotlin/Pair;");
-        s_overlay_control_data_portrait_position_field =
-                env->GetFieldID(overlay_control_data_class, "portraitPosition", "Lkotlin/Pair;");
-        s_overlay_control_data_foldable_position_field =
-                env->GetFieldID(overlay_control_data_class, "foldablePosition", "Lkotlin/Pair;");
-        s_overlay_control_data_individual_scale_field =
-                env->GetFieldID(overlay_control_data_class, "individualScale", "F");
-        env->DeleteLocalRef(overlay_control_data_class);
-
-        const jclass patch_class = env->FindClass("org/yuzu/yuzu_emu/model/Patch");
-        s_patch_class = reinterpret_cast<jclass>(env->NewGlobalRef(patch_class));
-        s_patch_constructor = env->GetMethodID(
-                patch_class, "<init>",
-                "(ZLjava/lang/String;Ljava/lang/String;ILjava/lang/String;Ljava/lang/String;JI)V");
-        s_patch_enabled_field = env->GetFieldID(patch_class, "enabled", "Z");
-        s_patch_name_field = env->GetFieldID(patch_class, "name", "Ljava/lang/String;");
-        s_patch_version_field = env->GetFieldID(patch_class, "version", "Ljava/lang/String;");
-        s_patch_type_field = env->GetFieldID(patch_class, "type", "I");
-        s_patch_program_id_field = env->GetFieldID(patch_class, "programId", "Ljava/lang/String;");
-        s_patch_title_id_field = env->GetFieldID(patch_class, "titleId", "Ljava/lang/String;");
-        env->DeleteLocalRef(patch_class);
-
-        const jclass double_class = env->FindClass("java/lang/Double");
-        s_double_class = reinterpret_cast<jclass>(env->NewGlobalRef(double_class));
-        s_double_constructor = env->GetMethodID(double_class, "<init>", "(D)V");
-        s_double_value_field = env->GetFieldID(double_class, "value", "D");
-        env->DeleteLocalRef(double_class);
-
-        const jclass int_class = env->FindClass("java/lang/Integer");
-        s_integer_class = reinterpret_cast<jclass>(env->NewGlobalRef(int_class));
-        s_integer_constructor = env->GetMethodID(int_class, "<init>", "(I)V");
-        s_integer_value_field = env->GetFieldID(int_class, "value", "I");
-        env->DeleteLocalRef(int_class);
-
-        const jclass boolean_class = env->FindClass("java/lang/Boolean");
-        s_boolean_class = reinterpret_cast<jclass>(env->NewGlobalRef(boolean_class));
-        s_boolean_constructor = env->GetMethodID(boolean_class, "<init>", "(Z)V");
-        s_boolean_value_field = env->GetFieldID(boolean_class, "value", "Z");
-        env->DeleteLocalRef(boolean_class);
-
-        const jclass player_input_class =
-                env->FindClass("org/yuzu/yuzu_emu/features/input/model/PlayerInput");
-        s_player_input_class = reinterpret_cast<jclass>(env->NewGlobalRef(player_input_class));
-        s_player_input_constructor = env->GetMethodID(
-                player_input_class, "<init>",
-                "(Z[Ljava/lang/String;[Ljava/lang/String;[Ljava/lang/String;ZIJJJJLjava/lang/String;Z)V");
-        s_player_input_connected_field = env->GetFieldID(player_input_class, "connected", "Z");
-        s_player_input_buttons_field =
-                env->GetFieldID(player_input_class, "buttons", "[Ljava/lang/String;");
-        s_player_input_analogs_field =
-                env->GetFieldID(player_input_class, "analogs", "[Ljava/lang/String;");
-        s_player_input_motions_field =
-                env->GetFieldID(player_input_class, "motions", "[Ljava/lang/String;");
-        s_player_input_vibration_enabled_field =
-                env->GetFieldID(player_input_class, "vibrationEnabled", "Z");
-        s_player_input_vibration_strength_field =
-                env->GetFieldID(player_input_class, "vibrationStrength", "I");
-        s_player_input_body_color_left_field =
-                env->GetFieldID(player_input_class, "bodyColorLeft", "J");
-        s_player_input_body_color_right_field =
-                env->GetFieldID(player_input_class, "bodyColorRight", "J");
-        s_player_input_button_color_left_field =
-                env->GetFieldID(player_input_class, "buttonColorLeft", "J");
-        s_player_input_button_color_right_field =
-                env->GetFieldID(player_input_class, "buttonColorRight", "J");
-        s_player_input_profile_name_field =
-                env->GetFieldID(player_input_class, "profileName", "Ljava/lang/String;");
-        s_player_input_use_system_vibrator_field =
-                env->GetFieldID(player_input_class, "useSystemVibrator", "Z");
-        env->DeleteLocalRef(player_input_class);
-
-        const jclass yuzu_input_device_interface =
-                env->FindClass("org/yuzu/yuzu_emu/features/input/YuzuInputDevice");
-        s_yuzu_input_device_interface =
-                reinterpret_cast<jclass>(env->NewGlobalRef(yuzu_input_device_interface));
-        s_yuzu_input_device_get_name =
-                env->GetMethodID(yuzu_input_device_interface, "getName", "()Ljava/lang/String;");
-        s_yuzu_input_device_get_guid =
-                env->GetMethodID(yuzu_input_device_interface, "getGUID", "()Ljava/lang/String;");
-        s_yuzu_input_device_get_port = env->GetMethodID(yuzu_input_device_interface, "getPort",
-                                                        "()I");
-        s_yuzu_input_device_get_supports_vibration =
-                env->GetMethodID(yuzu_input_device_interface, "getSupportsVibration", "()Z");
-        s_yuzu_input_device_vibrate = env->GetMethodID(yuzu_input_device_interface, "vibrate",
-                                                       "(F)V");
-        s_yuzu_input_device_get_axes =
-                env->GetMethodID(yuzu_input_device_interface, "getAxes", "()[Ljava/lang/Integer;");
-        s_yuzu_input_device_has_keys =
-                env->GetMethodID(yuzu_input_device_interface, "hasKeys", "([I)[Z");
-        env->DeleteLocalRef(yuzu_input_device_interface);
-        s_add_netplay_message = env->GetStaticMethodID(s_native_library_class, "addNetPlayMessage",
-                                                       "(ILjava/lang/String;)V");
-        s_clear_chat = env->GetStaticMethodID(s_native_library_class, "clearChat", "()V");
-
-
-        // Initialize Android Storage
-        Common::FS::Android::RegisterCallbacks(env, s_native_library_class);
-
-        // Initialize applets
-        Common::Android::SoftwareKeyboard::InitJNI(env);
-        Common::Android::WebBrowser::InitJNI(env);
-
-        return JNI_VERSION;
-    }
-
     void JNI_OnUnload(JavaVM *vm, void *reserved) {
         JNIEnv *env;
         if (vm->GetEnv(reinterpret_cast<void **>(&env), JNI_VERSION) != JNI_OK) {
@@ -645,4 +462,183 @@ namespace Common::Android {
     }
 #endif
 
+void Initialize(JavaVM* vm, JNIEnv *env) {
+    s_java_vm = vm;
+    InitFFmpegOnLoad(vm);
+
+    if (env->ExceptionCheck()) {
+        env->ExceptionClear();
+    }
+
+    // Initialize Java classes
+    const jclass native_library_class = env->FindClass("org/yuzu/yuzu_emu/NativeLibrary");
+    s_native_library_class = reinterpret_cast<jclass>(env->NewGlobalRef(native_library_class));
+    s_disk_cache_progress_class = reinterpret_cast<jclass>(env->NewGlobalRef(
+        env->FindClass("org/yuzu/yuzu_emu/disk_shader_cache/DiskShaderCacheProgress")));
+    s_load_callback_stage_class = reinterpret_cast<jclass>(env->NewGlobalRef(env->FindClass(
+        "org/yuzu/yuzu_emu/disk_shader_cache/DiskShaderCacheProgress$LoadCallbackStage")));
+
+    const jclass game_dir_class = env->FindClass("org/yuzu/yuzu_emu/model/GameDir");
+    s_game_dir_class = reinterpret_cast<jclass>(env->NewGlobalRef(game_dir_class));
+    s_game_dir_constructor = env->GetMethodID(game_dir_class, "<init>",
+                                              "(Ljava/lang/String;Z)V");
+    env->DeleteLocalRef(game_dir_class);
+
+    // Initialize methods
+    s_exit_emulation_activity =
+        env->GetStaticMethodID(s_native_library_class, "exitEmulationActivity", "(I)V");
+    s_disk_cache_load_progress =
+        env->GetStaticMethodID(s_disk_cache_progress_class, "loadProgress", "(III)V");
+    s_copy_to_storage = env->GetStaticMethodID(s_native_library_class, "copyFileToStorage",
+                                               "(Ljava/lang/String;Ljava/lang/String;)Z");
+    s_file_exists = env->GetStaticMethodID(s_native_library_class, "exists",
+                                           "(Ljava/lang/String;)Z");
+    s_file_extension = env->GetStaticMethodID(s_native_library_class, "getFileExtension",
+                                              "(Ljava/lang/String;)Ljava/lang/String;");
+    s_on_emulation_started =
+        env->GetStaticMethodID(s_native_library_class, "onEmulationStarted", "()V");
+    s_on_emulation_stopped =
+        env->GetStaticMethodID(s_native_library_class, "onEmulationStopped", "(I)V");
+    s_on_program_changed =
+        env->GetStaticMethodID(s_native_library_class, "onProgramChanged", "(I)V");
+
+    const jclass game_class = env->FindClass("org/yuzu/yuzu_emu/model/Game");
+    s_game_class = reinterpret_cast<jclass>(env->NewGlobalRef(game_class));
+    s_game_constructor = env->GetMethodID(game_class, "<init>",
+                                          "(Ljava/lang/String;Ljava/lang/String;Ljava/lang/"
+                                          "String;Ljava/lang/String;Ljava/lang/String;Z)V");
+    s_game_title_field = env->GetFieldID(game_class, "title", "Ljava/lang/String;");
+    s_game_path_field = env->GetFieldID(game_class, "path", "Ljava/lang/String;");
+    s_game_program_id_field = env->GetFieldID(game_class, "programId", "Ljava/lang/String;");
+    s_game_developer_field = env->GetFieldID(game_class, "developer", "Ljava/lang/String;");
+    s_game_version_field = env->GetFieldID(game_class, "version", "Ljava/lang/String;");
+    s_game_is_homebrew_field = env->GetFieldID(game_class, "isHomebrew", "Z");
+    env->DeleteLocalRef(game_class);
+
+    const jclass string_class = env->FindClass("java/lang/String");
+    s_string_class = reinterpret_cast<jclass>(env->NewGlobalRef(string_class));
+    env->DeleteLocalRef(string_class);
+
+    const jclass pair_class = env->FindClass("kotlin/Pair");
+    s_pair_class = reinterpret_cast<jclass>(env->NewGlobalRef(pair_class));
+    s_pair_constructor =
+        env->GetMethodID(pair_class, "<init>", "(Ljava/lang/Object;Ljava/lang/Object;)V");
+    s_pair_first_field = env->GetFieldID(pair_class, "first", "Ljava/lang/Object;");
+    s_pair_second_field = env->GetFieldID(pair_class, "second", "Ljava/lang/Object;");
+    env->DeleteLocalRef(pair_class);
+
+    const jclass overlay_control_data_class =
+        env->FindClass("org/yuzu/yuzu_emu/overlay/model/OverlayControlData");
+    s_overlay_control_data_class =
+        reinterpret_cast<jclass>(env->NewGlobalRef(overlay_control_data_class));
+    s_overlay_control_data_constructor =
+        env->GetMethodID(overlay_control_data_class, "<init>",
+                         "(Ljava/lang/String;ZLkotlin/Pair;Lkotlin/Pair;Lkotlin/Pair;F)V");
+    s_overlay_control_data_id_field =
+        env->GetFieldID(overlay_control_data_class, "id", "Ljava/lang/String;");
+    s_overlay_control_data_enabled_field =
+        env->GetFieldID(overlay_control_data_class, "enabled", "Z");
+    s_overlay_control_data_landscape_position_field =
+        env->GetFieldID(overlay_control_data_class, "landscapePosition", "Lkotlin/Pair;");
+    s_overlay_control_data_portrait_position_field =
+        env->GetFieldID(overlay_control_data_class, "portraitPosition", "Lkotlin/Pair;");
+    s_overlay_control_data_foldable_position_field =
+        env->GetFieldID(overlay_control_data_class, "foldablePosition", "Lkotlin/Pair;");
+    s_overlay_control_data_individual_scale_field =
+        env->GetFieldID(overlay_control_data_class, "individualScale", "F");
+    env->DeleteLocalRef(overlay_control_data_class);
+
+    const jclass patch_class = env->FindClass("org/yuzu/yuzu_emu/model/Patch");
+    s_patch_class = reinterpret_cast<jclass>(env->NewGlobalRef(patch_class));
+    s_patch_constructor = env->GetMethodID(
+        patch_class, "<init>",
+        "(ZLjava/lang/String;Ljava/lang/String;ILjava/lang/String;Ljava/lang/String;JI)V");
+    s_patch_enabled_field = env->GetFieldID(patch_class, "enabled", "Z");
+    s_patch_name_field = env->GetFieldID(patch_class, "name", "Ljava/lang/String;");
+    s_patch_version_field = env->GetFieldID(patch_class, "version", "Ljava/lang/String;");
+    s_patch_type_field = env->GetFieldID(patch_class, "type", "I");
+    s_patch_program_id_field = env->GetFieldID(patch_class, "programId", "Ljava/lang/String;");
+    s_patch_title_id_field = env->GetFieldID(patch_class, "titleId", "Ljava/lang/String;");
+    env->DeleteLocalRef(patch_class);
+
+    const jclass double_class = env->FindClass("java/lang/Double");
+    s_double_class = reinterpret_cast<jclass>(env->NewGlobalRef(double_class));
+    s_double_constructor = env->GetMethodID(double_class, "<init>", "(D)V");
+    s_double_value_method = env->GetMethodID(double_class, "doubleValue", "()D");
+    env->DeleteLocalRef(double_class);
+
+    const jclass int_class = env->FindClass("java/lang/Integer");
+    s_integer_class = reinterpret_cast<jclass>(env->NewGlobalRef(int_class));
+    s_integer_constructor = env->GetMethodID(int_class, "<init>", "(I)V");
+    s_integer_value_method = env->GetMethodID(int_class, "intValue", "()I");
+    env->DeleteLocalRef(int_class);
+
+    const jclass boolean_class = env->FindClass("java/lang/Boolean");
+    s_boolean_class = reinterpret_cast<jclass>(env->NewGlobalRef(boolean_class));
+    s_boolean_constructor = env->GetMethodID(boolean_class, "<init>", "(Z)V");
+    s_boolean_value_method = env->GetMethodID(boolean_class, "booleanValue", "()Z");
+    env->DeleteLocalRef(boolean_class);
+
+    const jclass player_input_class =
+        env->FindClass("org/yuzu/yuzu_emu/features/input/model/PlayerInput");
+    s_player_input_class = reinterpret_cast<jclass>(env->NewGlobalRef(player_input_class));
+    s_player_input_constructor = env->GetMethodID(
+        player_input_class, "<init>",
+        "(Z[Ljava/lang/String;[Ljava/lang/String;[Ljava/lang/String;ZIJJJJLjava/lang/String;Z)V");
+    s_player_input_connected_field = env->GetFieldID(player_input_class, "connected", "Z");
+    s_player_input_buttons_field =
+        env->GetFieldID(player_input_class, "buttons", "[Ljava/lang/String;");
+    s_player_input_analogs_field =
+        env->GetFieldID(player_input_class, "analogs", "[Ljava/lang/String;");
+    s_player_input_motions_field =
+        env->GetFieldID(player_input_class, "motions", "[Ljava/lang/String;");
+    s_player_input_vibration_enabled_field =
+        env->GetFieldID(player_input_class, "vibrationEnabled", "Z");
+    s_player_input_vibration_strength_field =
+        env->GetFieldID(player_input_class, "vibrationStrength", "I");
+    s_player_input_body_color_left_field =
+        env->GetFieldID(player_input_class, "bodyColorLeft", "J");
+    s_player_input_body_color_right_field =
+        env->GetFieldID(player_input_class, "bodyColorRight", "J");
+    s_player_input_button_color_left_field =
+        env->GetFieldID(player_input_class, "buttonColorLeft", "J");
+    s_player_input_button_color_right_field =
+        env->GetFieldID(player_input_class, "buttonColorRight", "J");
+    s_player_input_profile_name_field =
+        env->GetFieldID(player_input_class, "profileName", "Ljava/lang/String;");
+    s_player_input_use_system_vibrator_field =
+        env->GetFieldID(player_input_class, "useSystemVibrator", "Z");
+    env->DeleteLocalRef(player_input_class);
+
+    const jclass yuzu_input_device_interface =
+        env->FindClass("org/yuzu/yuzu_emu/features/input/YuzuInputDevice");
+    s_yuzu_input_device_interface =
+        reinterpret_cast<jclass>(env->NewGlobalRef(yuzu_input_device_interface));
+    s_yuzu_input_device_get_name =
+        env->GetMethodID(yuzu_input_device_interface, "getName", "()Ljava/lang/String;");
+    s_yuzu_input_device_get_guid =
+        env->GetMethodID(yuzu_input_device_interface, "getGUID", "()Ljava/lang/String;");
+    s_yuzu_input_device_get_port = env->GetMethodID(yuzu_input_device_interface, "getPort",
+                                                    "()I");
+    s_yuzu_input_device_get_supports_vibration =
+        env->GetMethodID(yuzu_input_device_interface, "getSupportsVibration", "()Z");
+    s_yuzu_input_device_vibrate = env->GetMethodID(yuzu_input_device_interface, "vibrate",
+                                                   "(F)V");
+    s_yuzu_input_device_get_axes =
+        env->GetMethodID(yuzu_input_device_interface, "getAxes", "()[Ljava/lang/Integer;");
+    s_yuzu_input_device_has_keys =
+        env->GetMethodID(yuzu_input_device_interface, "hasKeys", "([I)[Z");
+    env->DeleteLocalRef(yuzu_input_device_interface);
+    s_add_netplay_message = env->GetStaticMethodID(s_native_library_class, "addNetPlayMessage",
+                                                   "(ILjava/lang/String;)V");
+    s_clear_chat = env->GetStaticMethodID(s_native_library_class, "clearChat", "()V");
+
+    // Initialize Android Storage
+    Common::FS::Android::RegisterCallbacks(env, s_native_library_class);
+
+    // Initialize applets
+    Common::Android::SoftwareKeyboard::InitJNI(env);
+    Common::Android::WebBrowser::InitJNI(env);
+}
+
 } // namespace Common::Android

src/common/android/id_cache.h

@@ -11,6 +11,7 @@
 
 namespace Common::Android {
 
+void Initialize(JavaVM* vm, JNIEnv *env);
 JNIEnv* GetEnvForThread();
 
 /**
@@ -80,15 +81,15 @@ jfieldID GetPatchTitleIdField();
 
 jclass GetDoubleClass();
 jmethodID GetDoubleConstructor();
-jfieldID GetDoubleValueField();
+jmethodID GetDoubleValueMethod();
 
 jclass GetIntegerClass();
 jmethodID GetIntegerConstructor();
-jfieldID GetIntegerValueField();
+jmethodID GetIntegerValueMethod();
 
 jclass GetBooleanClass();
 jmethodID GetBooleanConstructor();
-jfieldID GetBooleanValueField();
+jmethodID GetBooleanValueMethod();
 
 jclass GetPlayerInputClass();
 jmethodID GetPlayerInputConstructor();

src/common/settings_enums.h

@@ -92,14 +92,14 @@ struct EnumMetadata {
 // AudioEngine must be specified discretely due to having existing but slightly different
 // canonicalizations
 // TODO (lat9nq): Remove explicit definition of AudioEngine/sink_id
-enum class AudioEngine : u32 { Auto, Cubeb, Sdl3, Null, Oboe, };
+enum class AudioEngine : u32 { Auto, Cubeb, Sdl3, Null, };
 template<>
 inline std::vector<std::pair<std::string_view, AudioEngine>> EnumMetadata<AudioEngine>::Canonicalizations() {
     return {
         {"auto", AudioEngine::Auto},
         {"cubeb", AudioEngine::Cubeb},
         {"sdl3", AudioEngine::Sdl3},
-        {"null", AudioEngine::Null}, {"oboe", AudioEngine::Oboe},
+        {"null", AudioEngine::Null},
     };
 }
 /// @brief This is just a sufficiently large number that is more than the number of other enums declared here
@@ -113,7 +113,7 @@ inline AudioEngine EnumMetadata<AudioEngine>::GetFirst() {
 }
 template<>
 inline AudioEngine EnumMetadata<AudioEngine>::GetLast() {
-    return AudioEngine::Oboe;
+    return AudioEngine::Null;
 }
 
 ENUM(AudioMode, Mono, Stereo, Surround);

src/input_common/drivers/android.cpp

@@ -114,7 +114,7 @@ std::set<s32> Android::GetDeviceAxes(JNIEnv* env, jobject& j_device) const {
     std::set<s32> axes;
     for (int i = 0; i < env->GetArrayLength(j_axes); ++i) {
         jobject axis = env->GetObjectArrayElement(j_axes, i);
-        axes.insert(env->GetIntField(axis, Common::Android::GetIntegerValueField()));
+        axes.insert(env->CallIntMethod(axis, Common::Android::GetIntegerValueMethod()));
     }
     return axes;
 }

src/video_core/buffer_cache/buffer_base.h

@@ -20,8 +20,6 @@ namespace VideoCommon {
 
 enum class BufferFlagBits {
     Picked = 1 << 0,
-    CachedWrites = 1 << 1,
-    PreemtiveDownload = 1 << 2,
 };
 DECLARE_ENUM_FLAG_OPERATORS(BufferFlagBits)
 
@@ -58,15 +56,6 @@ public:
         flags |= BufferFlagBits::Picked;
     }
 
-    void MarkPreemtiveDownload() noexcept {
-        flags |= BufferFlagBits::PreemtiveDownload;
-    }
-
-    /// Unmark buffer as picked
-    void Unpick() noexcept {
-        flags &= ~BufferFlagBits::Picked;
-    }
-
     /// Increases the likeliness of this being a stream buffer
     void IncreaseStreamScore(int score) noexcept {
         stream_score += score;
@@ -87,15 +76,6 @@ public:
         return True(flags & BufferFlagBits::Picked);
     }
 
-    /// Returns true when the buffer has pending cached writes
-    [[nodiscard]] bool HasCachedWrites() const noexcept {
-        return True(flags & BufferFlagBits::CachedWrites);
-    }
-
-    bool IsPreemtiveDownload() const noexcept {
-        return True(flags & BufferFlagBits::PreemtiveDownload);
-    }
-
     /// Returns the base CPU address of the buffer
     [[nodiscard]] VAddr CpuAddr() const noexcept {
         return cpu_addr;

src/video_core/buffer_cache/buffer_cache.h

@@ -7,6 +7,7 @@
 #pragma once
 
 #include <algorithm>
+#include <limits>
 #include <memory>
 #include <numeric>
 
@@ -121,25 +122,6 @@ void BufferCache<P>::WriteMemory(DAddr device_addr, u64 size) {
     memory_tracker.MarkRegionAsCpuModified(device_addr, size);
 }
 
-template <class P>
-void BufferCache<P>::CachedWriteMemory(DAddr device_addr, u64 size) {
-    const bool is_dirty = IsRegionRegistered(device_addr, size);
-    if (!is_dirty) {
-        return;
-    }
-    DAddr aligned_start = Common::AlignDown(device_addr, DEVICE_PAGESIZE);
-    DAddr aligned_end = Common::AlignUp(device_addr + size, DEVICE_PAGESIZE);
-    if (!IsRegionGpuModified(aligned_start, aligned_end - aligned_start)) {
-        WriteMemory(device_addr, size);
-        return;
-    }
-
-    tmp_buffer.resize_destructive(size);
-    device_memory.ReadBlockUnsafe(device_addr, tmp_buffer.data(), size);
-
-    InlineMemoryImplementation(device_addr, size, tmp_buffer);
-}
-
 template <class P>
 bool BufferCache<P>::OnCPUWrite(DAddr device_addr, u64 size) {
     const bool is_dirty = IsRegionRegistered(device_addr, size);
@@ -422,7 +404,7 @@ void BufferCache<P>::UnbindGraphicsStorageBuffers(size_t stage) {
 }
 
 template <class P>
-bool BufferCache<P>::BindGraphicsStorageBuffer(size_t stage, size_t ssbo_index, u32 cbuf_index,
+void BufferCache<P>::BindGraphicsStorageBuffer(size_t stage, size_t ssbo_index, u32 cbuf_index,
                                                u32 cbuf_offset, bool is_written) {
     const bool already_enabled =
         ((channel_state->enabled_storage_buffers[stage] >> ssbo_index) & 1U) != 0;
@@ -433,7 +415,7 @@ bool BufferCache<P>::BindGraphicsStorageBuffer(size_t stage, size_t ssbo_index,
                 LOG_WARNING(HW_GPU,
                             "Skipping graphics storage buffer {} due to driver limit {}",
                             ssbo_index, max_bindings);
-                return false;
+                return;
             }
         }
     }
@@ -449,7 +431,6 @@ bool BufferCache<P>::BindGraphicsStorageBuffer(size_t stage, size_t ssbo_index,
     const GPUVAddr ssbo_addr = cbufs.const_buffers[cbuf_index].address + cbuf_offset;
     channel_state->storage_buffers[stage][ssbo_index] =
         StorageBufferBinding(ssbo_addr, cbuf_index, is_written);
-    return (channel_state->storage_buffers[stage][ssbo_index].buffer_id != NULL_BUFFER_ID);
 }
 
 template <class P>
@@ -762,16 +743,6 @@ void BufferCache<P>::BindHostIndexBuffer() {
     }
 }
 
-template <class P>
-void BufferCache<P>::BindHostVertexBuffer(u32 index, Buffer& buffer, u32 offset, u32 size,
-                                          u32 stride) {
-    if constexpr (IS_OPENGL) {
-        runtime.BindVertexBuffer(index, buffer, offset, size, stride);
-    } else {
-        runtime.BindVertexBuffer(index, buffer.Handle(), offset, size, stride);
-    }
-}
-
 template <class P>
 Binding& BufferCache<P>::VertexBufferSlot(u32 index) {
     ASSERT(index < NUM_VERTEX_BUFFERS);
@@ -1197,7 +1168,7 @@ void BufferCache<P>::DoUpdateGraphicsBuffers(bool is_indexed) {
         if (is_indexed) {
             UpdateIndexBuffer();
         }
-        UpdateVertexBuffers();
+        UpdateVertexBuffers(is_indexed);
         UpdateTransformFeedbackBuffers();
         for (size_t stage = 0; stage < NUM_STAGES; ++stage) {
             UpdateUniformBuffers(stage);
@@ -1251,9 +1222,14 @@ void BufferCache<P>::UpdateIndexBuffer() {
     const GPUVAddr gpu_addr_begin = index_buffer_ref.StartAddress();
     const GPUVAddr gpu_addr_end = index_buffer_ref.EndAddress();
     const std::optional<DAddr> device_addr = gpu_memory->GpuToCpuAddress(gpu_addr_begin);
-    const u32 address_size = static_cast<u32>(gpu_addr_end - gpu_addr_begin);
+    u64 address_size = 0;
-    const u32 draw_size = (index_buffer_ref.count + index_buffer_ref.first) * u32(index_buffer_ref.FormatSizeInBytes());
+    if (gpu_addr_end > gpu_addr_begin) {
-    const u32 size = (std::min)(address_size, draw_size);
+        address_size = (std::min)(gpu_addr_end - gpu_addr_begin,
+                                  u64{(std::numeric_limits<u32>::max)()});
+    }
+    const u64 draw_size = (u64{index_buffer_ref.count} + u64{index_buffer_ref.first}) *
+                          u64{index_buffer_ref.FormatSizeInBytes()};
+    const u32 size = static_cast<u32>((std::min)(address_size, draw_size));
     if (size == 0 || !device_addr) {
         channel_state->index_buffer = NULL_BINDING;
         return;
@@ -1266,20 +1242,142 @@ void BufferCache<P>::UpdateIndexBuffer() {
 }
 
 template <class P>
-void BufferCache<P>::UpdateVertexBuffers() {
+u64 BufferCache<P>::DrawMaxIndex() {
+    if (max_index_scanned) {
+        return cached_max_index;
+    }
+    max_index_scanned = true;
+    cached_max_index = 0;
+    const auto& index_buffer_ref = maxwell3d->draw_manager.draw_state.index_buffer;
+    const u32 count = index_buffer_ref.count;
+    if (count == 0) {
+        return 0;
+    }
+    index_scan_buffer.resize_destructive(count);
+    gpu_memory->ReadBlockUnsafe(index_buffer_ref.IndexStart(), index_scan_buffer.data(),
+                                size_t{count} * sizeof(u32));
+    u32 restart_index = (std::numeric_limits<u32>::max)();
+    if (maxwell3d->regs.primitive_restart.enabled != 0) {
+        restart_index = maxwell3d->regs.primitive_restart.index;
+    }
+    u32 max_index = 0;
+    for (u32 i = 0; i < count; ++i) {
+        const u32 value = index_scan_buffer[i];
+        if (value == restart_index) {
+            continue;
+        }
+        max_index = (std::max)(max_index, value);
+    }
+    cached_max_index = max_index;
+    return cached_max_index;
+}
+
+template <class P>
+u64 BufferCache<P>::StreamAttributeExtent(u32 index) {
+    using VertexAttribute = typename Maxwell::VertexAttribute;
+    if (stream_extents_valid) {
+        return stream_extents[index];
+    }
+    stream_extents_valid = true;
+    stream_extents.fill(0);
+    for (size_t i = 0; i < Maxwell::NumVertexAttributes; ++i) {
+        const auto& attribute = maxwell3d->regs.vertex_attrib_format[i];
+        if (attribute.constant != 0 || attribute.size == VertexAttribute::Size::Invalid) {
+            continue;
+        }
+        const u32 buffer = attribute.buffer.Value();
+        if (buffer >= NUM_VERTEX_BUFFERS) {
+            continue;
+        }
+        const u64 end = static_cast<u64>(attribute.offset.Value()) +
+                        static_cast<u64>(attribute.SizeInBytes());
+        stream_extents[buffer] = (std::max)(stream_extents[buffer], end);
+    }
+    return stream_extents[index];
+}
+
+template <class P>
+u64 BufferCache<P>::DrawVertexBound(u32 index, bool is_indexed) {
+    const auto& array = maxwell3d->regs.vertex_streams[index];
+    if (array.enable == 0) {
+        return 0;
+    }
+    const u64 extent = StreamAttributeExtent(index);
+    if (extent == 0) {
+        return 0;
+    }
+    const u64 stride = static_cast<u64>(array.stride);
+    if (stride == 0) {
+        return extent;
+    }
+    const auto& draw_state = maxwell3d->draw_manager.draw_state;
+    u64 elements = 0;
+    if (maxwell3d->regs.vertex_stream_instances.IsInstancingEnabled(index)) {
+        if (draw_instance_count == 0) {
+            return 0;
+        }
+        const u64 base_instance = static_cast<u64>(draw_state.base_instance);
+        elements = base_instance + 1;
+        if (array.frequency != 0) {
+            elements = (base_instance + static_cast<u64>(draw_instance_count) - 1) /
+                           static_cast<u64>(array.frequency) +
+                       1;
+        }
+    } else if (!is_indexed) {
+        elements = static_cast<u64>(draw_state.vertex_buffer.first) +
+                   static_cast<u64>(draw_state.vertex_buffer.count);
+    } else {
+        const auto format = draw_state.index_buffer.format;
+        u64 max_index = 0xFF;
+        if (format == Maxwell::IndexFormat::UnsignedShort) {
+            max_index = 0xFFFF;
+        } else if (format != Maxwell::IndexFormat::UnsignedByte) {
+            const auto& limit = maxwell3d->regs.vertex_stream_limits[index];
+            const GPUVAddr gpu_addr_begin = array.Address();
+            const GPUVAddr gpu_addr_end = limit.Address() + 1;
+            if (gpu_addr_end <= gpu_addr_begin) {
+                return 0;
+            }
+            const bool walks = gpu_addr_end - gpu_addr_begin >= IMPLAUSIBLE_VERTEX_SIZE ||
+                               !gpu_memory->IsWithinGPUAddressRange(gpu_addr_end);
+            if (!walks) {
+                return 0;
+            }
+            max_index = DrawMaxIndex();
+        }
+        elements = static_cast<u64>(draw_state.base_index) + max_index + 1;
+    }
+    if (elements == 0) {
+        return extent;
+    }
+    return (elements - 1) * stride + extent;
+}
+
+template <class P>
+void BufferCache<P>::UpdateVertexBuffers(bool is_indexed) {
     auto& flags = maxwell3d->dirty.flags;
+    max_index_scanned = false;
+    stream_extents_valid = false;
+    for (u32 index = 0; index < NUM_VERTEX_BUFFERS; ++index) {
+        const u64 bound = DrawVertexBound(index, is_indexed);
+        if (bound <= last_draw_bounds[index]) {
+            continue;
+        }
+        flags[Dirty::VertexBuffer0 + index] = true;
+        flags[Dirty::VertexBuffers] = true;
+    }
     if (!maxwell3d->dirty.flags[Dirty::VertexBuffers]) {
         return;
     }
     flags[Dirty::VertexBuffers] = false;
 
     for (u32 index = 0; index < NUM_VERTEX_BUFFERS; ++index) {
-        UpdateVertexBuffer(index);
+        UpdateVertexBuffer(index, is_indexed);
     }
 }
 
 template <class P>
-void BufferCache<P>::UpdateVertexBuffer(u32 index) {
+void BufferCache<P>::UpdateVertexBuffer(u32 index, bool is_indexed) {
     if (!maxwell3d->dirty.flags[Dirty::VertexBuffer0 + index]) {
         return;
     }
@@ -1288,15 +1386,31 @@ void BufferCache<P>::UpdateVertexBuffer(u32 index) {
     const GPUVAddr gpu_addr_begin = array.Address();
     const GPUVAddr gpu_addr_end = limit.Address() + 1;
     const std::optional<DAddr> device_addr = gpu_memory->GpuToCpuAddress(gpu_addr_begin);
-    const u32 address_size = static_cast<u32>(gpu_addr_end - gpu_addr_begin);
+    if (array.enable == 0 || !device_addr || gpu_addr_end <= gpu_addr_begin) {
-    u32 size = address_size; // TODO: Analyze stride and number of vertices
-    if (array.enable == 0 || size == 0 || !device_addr) {
         channel_state->vertex_buffers[index] = NULL_BINDING;
         UpdateVertexBufferSlot(index, NULL_BINDING);
         return;
     }
-    if (!gpu_memory->IsWithinGPUAddressRange(gpu_addr_end) || size >= 64_MiB) {
+    // TODO: Analyze stride and number of vertices
-        size = static_cast<u32>(gpu_memory->MaxContinuousRange(gpu_addr_begin, size));
+    constexpr u64 implausible_size = IMPLAUSIBLE_VERTEX_SIZE;
+    u64 address_size = gpu_addr_end - gpu_addr_begin;
+    if (address_size > u64{(std::numeric_limits<u32>::max)()}) {
+        address_size = implausible_size;
+    }
+    const u64 draw_bound = DrawVertexBound(index, is_indexed);
+    last_draw_bounds[index] = (std::numeric_limits<u64>::max)();
+    if (draw_bound != 0) {
+        last_draw_bounds[index] = draw_bound;
+        address_size = (std::min)(address_size, draw_bound);
+    }
+    if (!gpu_memory->IsWithinGPUAddressRange(gpu_addr_end) || address_size >= implausible_size) {
+        address_size = gpu_memory->MaxContinuousRange(gpu_addr_begin, address_size);
+    }
+    const u32 size = static_cast<u32>(address_size);
+    if (size == 0) {
+        channel_state->vertex_buffers[index] = NULL_BINDING;
+        UpdateVertexBufferSlot(index, NULL_BINDING);
+        return;
     }
     const BufferId buffer_id = FindBuffer(*device_addr, size);
     const Binding binding{
@@ -1578,9 +1692,10 @@ template <class P>
 BufferId BufferCache<P>::CreateBuffer(DAddr device_addr, u32 wanted_size) {
     DAddr device_addr_end = Common::AlignUp(device_addr + wanted_size, CACHING_PAGESIZE);
     device_addr = Common::AlignDown(device_addr, CACHING_PAGESIZE);
-    wanted_size = static_cast<u32>(device_addr_end - device_addr);
+    constexpr u64 max_buffer_size = u64{(std::numeric_limits<u32>::max)()};
+    wanted_size = static_cast<u32>((std::min)(device_addr_end - device_addr, max_buffer_size));
     const OverlapResult overlap = ResolveOverlaps(device_addr, wanted_size);
-    const u32 size = static_cast<u32>(overlap.end - overlap.begin);
+    const u32 size = static_cast<u32>((std::min)(overlap.end - overlap.begin, max_buffer_size));
     const BufferId new_buffer_id = slot_buffers.insert(runtime, overlap.begin, size);
     auto& new_buffer = slot_buffers[new_buffer_id];
     const size_t size_bytes = new_buffer.SizeBytes();

src/video_core/buffer_cache/buffer_cache_base.h

@@ -51,6 +51,7 @@ constexpr u32 NUM_VERTEX_BUFFERS = 16;
 #else
 constexpr u32 NUM_VERTEX_BUFFERS = 32;
 #endif
+constexpr u64 IMPLAUSIBLE_VERTEX_SIZE = 64_MiB;
 constexpr u32 NUM_TRANSFORM_FEEDBACK_BUFFERS = 4;
 constexpr u32 NUM_GRAPHICS_UNIFORM_BUFFERS = 18;
 constexpr u32 NUM_COMPUTE_UNIFORM_BUFFERS = 8;
@@ -217,8 +218,6 @@ public:
 
     void WriteMemory(DAddr device_addr, u64 size);
 
-    void CachedWriteMemory(DAddr device_addr, u64 size);
-
     bool OnCPUWrite(DAddr device_addr, u64 size);
 
     void DownloadMemory(DAddr device_addr, u64 size);
@@ -248,7 +247,7 @@ public:
 
     void UnbindGraphicsStorageBuffers(size_t stage);
 
-    bool BindGraphicsStorageBuffer(size_t stage, size_t ssbo_index, u32 cbuf_index, u32 cbuf_offset,
+    void BindGraphicsStorageBuffer(size_t stage, size_t ssbo_index, u32 cbuf_index, u32 cbuf_offset,
                                    bool is_written);
 
     void UnbindGraphicsTextureBuffers(size_t stage);
@@ -309,6 +308,10 @@ public:
         current_draw_indirect = current_draw_indirect_;
     }
 
+    void SetDrawInstanceCount(u32 draw_instance_count_) {
+        draw_instance_count = draw_instance_count_;
+    }
+
     [[nodiscard]] std::pair<Buffer*, u32> GetDrawIndirectCount();
 
     [[nodiscard]] std::pair<Buffer*, u32> GetDrawIndirectBuffer();
@@ -376,8 +379,6 @@ private:
 
     void BindHostTransformFeedbackBuffers();
 
-    void BindHostVertexBuffer(u32 index, Buffer& buffer, u32 offset, u32 size, u32 stride);
-
     void BindHostComputeUniformBuffers();
 
     void BindHostComputeStorageBuffers();
@@ -390,9 +391,15 @@ private:
 
     void UpdateIndexBuffer();
 
-    void UpdateVertexBuffers();
+    void UpdateVertexBuffers(bool is_indexed);
 
-    void UpdateVertexBuffer(u32 index);
+    void UpdateVertexBuffer(u32 index, bool is_indexed);
+
+    [[nodiscard]] u64 DrawVertexBound(u32 index, bool is_indexed);
+
+    [[nodiscard]] u64 DrawMaxIndex();
+
+    [[nodiscard]] u64 StreamAttributeExtent(u32 index);
 
     void UpdateDrawIndirect();
 
@@ -484,6 +491,14 @@ private:
 
     const Tegra::Engines::Maxwell3D::DrawManager::IndirectParams* current_draw_indirect{};
 
+    u32 draw_instance_count = 0;
+    std::array<u64, NUM_VERTEX_BUFFERS> last_draw_bounds{};
+    Common::ScratchBuffer<u32> index_scan_buffer;
+    u64 cached_max_index = 0;
+    bool max_index_scanned = false;
+    std::array<u64, NUM_VERTEX_BUFFERS> stream_extents{};
+    bool stream_extents_valid = false;
+
     u32 last_index_count = 0;
 
     u32 enabled_vertex_buffers_mask = 0;

src/video_core/host_shaders/CMakeLists.txt

@@ -206,6 +206,40 @@ foreach(VARIANT IN ITEMS ${SHADER_TYPE_VARIANTS})
     set(SHADER_HEADERS ${SHADER_HEADERS} ${VARIANT_HEADER_FILE})
 endforeach()
 
+set(SHADER_DEFINE_VARIANTS
+    "block_linear_unswizzle_2d.comp|nonarrow|HAS_EXTENDED_TYPES=0"
+    "pitch_unswizzle.comp|nonarrow|HAS_EXTENDED_TYPES=0"
+    "block_linear_unswizzle_3d.comp|nonarrow|HAS_EXTENDED_TYPES=0"
+)
+
+foreach(VARIANT IN ITEMS ${SHADER_DEFINE_VARIANTS})
+    string(REPLACE "|" ";" VARIANT_PARTS ${VARIANT})
+    list(GET VARIANT_PARTS 0 VARIANT_FILENAME)
+    list(GET VARIANT_PARTS 1 VARIANT_SUFFIX)
+    list(GET VARIANT_PARTS 2 VARIANT_DEFINE)
+
+    set(VARIANT_SOURCE ${CMAKE_CURRENT_SOURCE_DIR}/${VARIANT_FILENAME})
+    get_filename_component(VARIANT_STEM ${VARIANT_FILENAME} NAME_WE)
+    get_filename_component(VARIANT_EXT ${VARIANT_FILENAME} EXT)
+    string(REPLACE "." "" VARIANT_EXT ${VARIANT_EXT})
+    set(VARIANT_NAME ${VARIANT_STEM}_${VARIANT_SUFFIX}_${VARIANT_EXT})
+
+    string(TOUPPER ${VARIANT_NAME}_SPV VARIANT_VARIABLE_NAME)
+    set(VARIANT_HEADER_FILE ${SHADER_DIR}/${VARIANT_NAME}_spv.h)
+    add_custom_command(
+        OUTPUT
+            ${VARIANT_HEADER_FILE}
+        COMMAND
+            ${GLSLANGVALIDATOR} -V ${QUIET_FLAG} -I"${FIDELITYFX_INCLUDE_DIR}" ${GLSL_FLAGS}
+            -D${VARIANT_DEFINE}
+            --variable-name ${VARIANT_VARIABLE_NAME} -o ${VARIANT_HEADER_FILE} ${VARIANT_SOURCE}
+            --target-env ${SPIR_V_VERSION}
+        MAIN_DEPENDENCY
+            ${VARIANT_SOURCE}
+    )
+    set(SHADER_HEADERS ${SHADER_HEADERS} ${VARIANT_HEADER_FILE})
+endforeach()
+
 foreach(FILEPATH IN ITEMS ${FIDELITYFX_FILES})
     get_filename_component(FILENAME ${FILEPATH} NAME)
     string(REPLACE "." "_" HEADER_NAME ${FILENAME})

src/video_core/host_shaders/block_linear_unswizzle_2d.comp

@@ -5,9 +5,13 @@
 
 #ifdef VULKAN
 
+#ifndef HAS_EXTENDED_TYPES
+#define HAS_EXTENDED_TYPES 1
+#endif
+#if HAS_EXTENDED_TYPES
 #extension GL_EXT_shader_16bit_storage : require
 #extension GL_EXT_shader_8bit_storage : require
-#define HAS_EXTENDED_TYPES 1
+#endif
 #define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants {
 #define END_PUSH_CONSTANTS };
 #define UNIFORM(n)

src/video_core/host_shaders/block_linear_unswizzle_3d.comp

@@ -5,9 +5,13 @@
 
 #ifdef VULKAN
 
+#ifndef HAS_EXTENDED_TYPES
+#define HAS_EXTENDED_TYPES 1
+#endif
+#if HAS_EXTENDED_TYPES
 #extension GL_EXT_shader_16bit_storage : require
 #extension GL_EXT_shader_8bit_storage : require
-#define HAS_EXTENDED_TYPES 1
+#endif
 #define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants {
 #define END_PUSH_CONSTANTS };
 #define UNIFORM(n)

src/video_core/host_shaders/pitch_unswizzle.comp

@@ -5,9 +5,13 @@
 
 #ifdef VULKAN
 
+#ifndef HAS_EXTENDED_TYPES
+#define HAS_EXTENDED_TYPES 1
+#endif
+#if HAS_EXTENDED_TYPES
 #extension GL_EXT_shader_16bit_storage : require
 #extension GL_EXT_shader_8bit_storage : require
-#define HAS_EXTENDED_TYPES 1
+#endif
 #define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants {
 #define END_PUSH_CONSTANTS };
 #define UNIFORM(n)

src/video_core/renderer_opengl/gl_buffer_cache.cpp

@@ -226,22 +226,6 @@ void BufferCacheRuntime::BindIndexBuffer(Buffer& buffer, u32 offset, u32 size) {
     }
 }
 
-void BufferCacheRuntime::BindVertexBuffer(u32 index, Buffer& buffer, u32 offset, u32 size,
-                                          u32 stride) {
-    if (index >= max_attributes) {
-        return;
-    }
-    if (has_unified_vertex_buffers) {
-        buffer.MakeResident(GL_READ_ONLY);
-        glBindVertexBuffer(index, 0, 0, static_cast<GLsizei>(stride));
-        glBufferAddressRangeNV(GL_VERTEX_ATTRIB_ARRAY_ADDRESS_NV, index,
-                               buffer.HostGpuAddr() + offset, static_cast<GLsizeiptr>(size));
-    } else {
-        glBindVertexBuffer(index, buffer.Handle(), static_cast<GLintptr>(offset),
-                           static_cast<GLsizei>(stride));
-    }
-}
-
 void BufferCacheRuntime::BindVertexBuffers(VideoCommon::HostBindings<Buffer>& bindings) {
     // TODO: Should HostBindings provide the correct runtime types to avoid these transforms?
     std::array<GLuint, 32> buffer_handles;

src/video_core/renderer_opengl/gl_buffer_cache.h

@@ -99,8 +99,6 @@ public:
 
     void BindIndexBuffer(Buffer& buffer, u32 offset, u32 size);
 
-    void BindVertexBuffer(u32 index, Buffer& buffer, u32 offset, u32 size, u32 stride);
-
     void BindVertexBuffers(VideoCommon::HostBindings<Buffer>& bindings);
 
     void BindUniformBuffer(size_t stage, u32 binding_index, Buffer& buffer, u32 offset, u32 size);

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -259,6 +259,7 @@ void RasterizerOpenGL::PrepareDraw(bool is_indexed, Func&& draw_func) {
 }
 
 void RasterizerOpenGL::Draw(bool is_indexed, u32 instance_count) {
+    buffer_cache.SetDrawInstanceCount(instance_count);
     PrepareDraw(is_indexed, [this, is_indexed, instance_count](GLenum primitive_mode) {
         const auto& draw_state = maxwell3d->draw_manager.draw_state;
         const GLuint base_instance = GLuint(draw_state.base_instance);
@@ -304,6 +305,7 @@ void RasterizerOpenGL::Draw(bool is_indexed, u32 instance_count) {
 void RasterizerOpenGL::DrawIndirect() {
     const auto& params = maxwell3d->draw_manager.indirect_state;
     buffer_cache.SetDrawIndirect(&params);
+    buffer_cache.SetDrawInstanceCount(0);
     PrepareDraw(params.is_indexed, [this, &params](GLenum primitive_mode) {
         if (params.is_byte_count) {
             const GPUVAddr tfb_object_base_addr = params.indirect_start_address - 4U;

src/video_core/renderer_vulkan/vk_buffer_cache.cpp

@@ -585,29 +585,6 @@ void BufferCacheRuntime::BindQuadIndexBuffer(PrimitiveTopology topology, u32 fir
     }
 }
 
-void BufferCacheRuntime::BindVertexBuffer(u32 index, VkBuffer buffer, u32 offset, u32 size, u32 stride) {
-    if (index >= device.GetMaxVertexInputBindings()) {
-        return;
-    }
-    if (device.IsExtExtendedDynamicStateSupported()) {
-        scheduler.Record([index, buffer, offset, size, stride](vk::CommandBuffer cmdbuf) {
-            const VkDeviceSize vk_offset = buffer != VK_NULL_HANDLE ? offset : 0;
-            const VkDeviceSize vk_size = buffer != VK_NULL_HANDLE ? size : VK_WHOLE_SIZE;
-            const VkDeviceSize vk_stride = stride;
-            cmdbuf.BindVertexBuffers2EXT(index, 1, &buffer, &vk_offset, &vk_size, &vk_stride);
-        });
-    } else {
-        if (!device.HasNullDescriptor() && buffer == VK_NULL_HANDLE) {
-            ReserveNullBuffer();
-            buffer = *null_buffer;
-            offset = 0;
-        }
-        scheduler.Record([index, buffer, offset](vk::CommandBuffer cmdbuf) {
-            cmdbuf.BindVertexBuffer(index, buffer, offset);
-        });
-    }
-}
-
 void BufferCacheRuntime::BindVertexBuffers(VideoCommon::HostBindings<Buffer>& bindings) {
     boost::container::static_vector<VkBuffer, VideoCommon::NUM_VERTEX_BUFFERS> buffer_handles(bindings.buffers.size());
     for (u32 i = 0; i < bindings.buffers.size(); ++i) {

src/video_core/renderer_vulkan/vk_buffer_cache.h

@@ -138,8 +138,6 @@ public:
 
     void BindQuadIndexBuffer(PrimitiveTopology topology, u32 first, u32 count);
 
-    void BindVertexBuffer(u32 index, VkBuffer buffer, u32 offset, u32 size, u32 stride);
-
     void BindVertexBuffers(VideoCommon::HostBindings<Buffer>& bindings);
 
     void BindTransformFeedbackBuffer(u32 index, VkBuffer buffer, u32 offset, u32 size);

src/video_core/renderer_vulkan/vk_compute_pass.cpp

@@ -22,7 +22,13 @@
 #include "video_core/host_shaders/resolve_conditional_render_comp_spv.h"
 #include "video_core/host_shaders/vulkan_quad_indexed_comp_spv.h"
 #include "video_core/host_shaders/vulkan_uint8_comp_spv.h"
+#include "video_core/host_shaders/block_linear_unswizzle_2d_comp_spv.h"
+#include "video_core/host_shaders/block_linear_unswizzle_2d_nonarrow_comp_spv.h"
 #include "video_core/host_shaders/block_linear_unswizzle_3d_bcn_comp_spv.h"
+#include "video_core/host_shaders/block_linear_unswizzle_3d_comp_spv.h"
+#include "video_core/host_shaders/block_linear_unswizzle_3d_nonarrow_comp_spv.h"
+#include "video_core/host_shaders/pitch_unswizzle_comp_spv.h"
+#include "video_core/host_shaders/pitch_unswizzle_nonarrow_comp_spv.h"
 #include "video_core/renderer_vulkan/vk_compute_pass.h"
 #include "video_core/surface.h"
 #include "video_core/renderer_vulkan/vk_descriptor_pool.h"
@@ -872,4 +878,291 @@ void BlockLinearUnswizzle3DPass::UnswizzleChunk(
     });
 }
 
+namespace {
+
+constexpr u32 UNSWIZZLE_BINDING_INPUT_BUFFER = 0;
+constexpr u32 UNSWIZZLE_BINDING_OUTPUT_IMAGE = 1;
+constexpr size_t UNSWIZZLE_NUM_BINDINGS = 2;
+
+constexpr std::array<VkDescriptorSetLayoutBinding, UNSWIZZLE_NUM_BINDINGS>
+    UNSWIZZLE_DESCRIPTOR_SET_BINDINGS{{
+        {
+            .binding = UNSWIZZLE_BINDING_INPUT_BUFFER,
+            .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
+            .descriptorCount = 1,
+            .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
+            .pImmutableSamplers = nullptr,
+        },
+        {
+            .binding = UNSWIZZLE_BINDING_OUTPUT_IMAGE,
+            .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE,
+            .descriptorCount = 1,
+            .stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
+            .pImmutableSamplers = nullptr,
+        },
+    }};
+
+constexpr std::array<VkDescriptorUpdateTemplateEntry, UNSWIZZLE_NUM_BINDINGS>
+    UNSWIZZLE_DESCRIPTOR_UPDATE_TEMPLATE{{
+        {
+            .dstBinding = UNSWIZZLE_BINDING_INPUT_BUFFER,
+            .dstArrayElement = 0,
+            .descriptorCount = 1,
+            .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
+            .offset = UNSWIZZLE_BINDING_INPUT_BUFFER * sizeof(DescriptorUpdateEntry),
+            .stride = sizeof(DescriptorUpdateEntry),
+        },
+        {
+            .dstBinding = UNSWIZZLE_BINDING_OUTPUT_IMAGE,
+            .dstArrayElement = 0,
+            .descriptorCount = 1,
+            .descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE,
+            .offset = UNSWIZZLE_BINDING_OUTPUT_IMAGE * sizeof(DescriptorUpdateEntry),
+            .stride = sizeof(DescriptorUpdateEntry),
+        },
+    }};
+
+constexpr DescriptorBankInfo UNSWIZZLE_BANK_INFO{
+    .uniform_buffers = 0,
+    .storage_buffers = 1,
+    .texture_buffers = 0,
+    .image_buffers = 0,
+    .textures = 0,
+    .images = 1,
+    .score = 2,
+};
+
+[[nodiscard]] std::span<const u32> UnswizzleSpv(const Device& device,
+                                                std::span<const u32> extended,
+                                                std::span<const u32> narrow) {
+    if (device.IsStorageBuffer8BitAccessSupported() &&
+        device.IsStorageBuffer16BitAccessSupported()) {
+        return extended;
+    }
+    return narrow;
+}
+
+struct PitchUnswizzlePushConstants {
+    alignas(8) std::array<u32, 2> origin;
+    alignas(8) std::array<s32, 2> destination;
+    u32 bytes_per_block;
+    u32 pitch;
+};
+
+void RecordUnswizzleEntryBarrier(Scheduler& scheduler, VkPipeline vk_pipeline, VkImage vk_image,
+                                 VkImageAspectFlags aspect_mask, bool is_initialized) {
+    scheduler.Record([vk_pipeline, vk_image, aspect_mask,
+                      is_initialized](vk::CommandBuffer cmdbuf) {
+        VkAccessFlags src_access = VK_ACCESS_NONE;
+        VkImageLayout old_layout = VK_IMAGE_LAYOUT_UNDEFINED;
+        if (is_initialized) {
+            src_access = VK_ACCESS_SHADER_WRITE_BIT | VK_ACCESS_TRANSFER_WRITE_BIT |
+                         VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT;
+            old_layout = VK_IMAGE_LAYOUT_GENERAL;
+        }
+        const VkImageMemoryBarrier image_barrier{
+            .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
+            .pNext = nullptr,
+            .srcAccessMask = src_access,
+            .dstAccessMask = VK_ACCESS_SHADER_WRITE_BIT,
+            .oldLayout = old_layout,
+            .newLayout = VK_IMAGE_LAYOUT_GENERAL,
+            .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .image = vk_image,
+            .subresourceRange{
+                .aspectMask = aspect_mask,
+                .baseMipLevel = 0,
+                .levelCount = VK_REMAINING_MIP_LEVELS,
+                .baseArrayLayer = 0,
+                .layerCount = VK_REMAINING_ARRAY_LAYERS,
+            },
+        };
+        VkPipelineStageFlags src_stage = VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT;
+        if (is_initialized) {
+            src_stage = vk::PIPELINE_STAGE_GRAPHICS_COMPUTE_TRANSFER;
+        }
+        cmdbuf.PipelineBarrier(src_stage, VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, 0, image_barrier);
+        cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_COMPUTE, vk_pipeline);
+    });
+}
+
+void RecordUnswizzleExitBarrier(Scheduler& scheduler, VkImage vk_image,
+                                VkImageAspectFlags aspect_mask) {
+    scheduler.Record([vk_image, aspect_mask](vk::CommandBuffer cmdbuf) {
+        const VkImageMemoryBarrier image_barrier{
+            .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
+            .pNext = nullptr,
+            .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT,
+            .dstAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_TRANSFER_READ_BIT |
+                             VK_ACCESS_COLOR_ATTACHMENT_READ_BIT,
+            .oldLayout = VK_IMAGE_LAYOUT_GENERAL,
+            .newLayout = VK_IMAGE_LAYOUT_GENERAL,
+            .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
+            .image = vk_image,
+            .subresourceRange{
+                .aspectMask = aspect_mask,
+                .baseMipLevel = 0,
+                .levelCount = VK_REMAINING_MIP_LEVELS,
+                .baseArrayLayer = 0,
+                .layerCount = VK_REMAINING_ARRAY_LAYERS,
+            },
+        };
+        cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT,
+                               vk::PIPELINE_STAGE_GRAPHICS_COMPUTE_TRANSFER, 0, image_barrier);
+    });
+}
+
+} // Anonymous namespace
+
+BlockLinearUnswizzle2DPass::BlockLinearUnswizzle2DPass(
+    const Device& device_, Scheduler& scheduler_, DescriptorPool& descriptor_pool_,
+    ComputePassDescriptorQueue& compute_pass_descriptor_queue_)
+    : ComputePass(device_, scheduler_, descriptor_pool_, UNSWIZZLE_DESCRIPTOR_SET_BINDINGS,
+                  UNSWIZZLE_DESCRIPTOR_UPDATE_TEMPLATE, UNSWIZZLE_BANK_INFO,
+                  COMPUTE_PUSH_CONSTANT_RANGE<sizeof(
+                      VideoCommon::Accelerated::BlockLinearSwizzle2DParams)>,
+                  UnswizzleSpv(device_, BLOCK_LINEAR_UNSWIZZLE_2D_COMP_SPV,
+                               BLOCK_LINEAR_UNSWIZZLE_2D_NONARROW_COMP_SPV)),
+      scheduler{scheduler_}, compute_pass_descriptor_queue{compute_pass_descriptor_queue_} {}
+
+BlockLinearUnswizzle2DPass::~BlockLinearUnswizzle2DPass() = default;
+
+void BlockLinearUnswizzle2DPass::Unswizzle(
+    Image& image, const StagingBufferRef& map,
+    std::span<const VideoCommon::SwizzleParameters> swizzles) {
+    using namespace VideoCommon::Accelerated;
+    scheduler.RequestOutsideRenderPassOperationContext();
+    const VkPipeline vk_pipeline = *pipeline;
+    const VkImageAspectFlags aspect_mask = image.AspectMask();
+    const VkImage vk_image = image.Handle();
+    const bool is_initialized = image.ExchangeInitialization();
+    RecordUnswizzleEntryBarrier(scheduler, vk_pipeline, vk_image, aspect_mask, is_initialized);
+
+    const u32 num_layers = static_cast<u32>(image.info.resources.layers);
+    for (const VideoCommon::SwizzleParameters& swizzle : swizzles) {
+        const size_t input_offset = swizzle.buffer_offset + map.offset;
+        const u32 num_dispatches_x = Common::DivCeil(swizzle.num_tiles.width, 32U);
+        const u32 num_dispatches_y = Common::DivCeil(swizzle.num_tiles.height, 32U);
+
+        compute_pass_descriptor_queue.Acquire(scheduler, 2);
+        compute_pass_descriptor_queue.AddBuffer(map.buffer, input_offset,
+                                                image.guest_size_bytes - swizzle.buffer_offset);
+        compute_pass_descriptor_queue.AddImage(image.StorageImageView(swizzle.level));
+        const void* const descriptor_data{compute_pass_descriptor_queue.UpdateData()};
+
+        const auto params = MakeBlockLinearSwizzle2DParams(swizzle, image.info);
+        scheduler.Record([this, num_dispatches_x, num_dispatches_y, num_layers, params,
+                          descriptor_data](vk::CommandBuffer cmdbuf) {
+            const VkDescriptorSet set = descriptor_allocator.Commit();
+            device.GetLogical().UpdateDescriptorSet(set, *descriptor_template, descriptor_data);
+            cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_COMPUTE, *layout, 0, set, {});
+            cmdbuf.PushConstants(*layout, VK_SHADER_STAGE_COMPUTE_BIT, params);
+            cmdbuf.Dispatch(num_dispatches_x, num_dispatches_y, num_layers);
+        });
+    }
+    RecordUnswizzleExitBarrier(scheduler, vk_image, aspect_mask);
+}
+
+BlockLinearUnswizzleImage3DPass::BlockLinearUnswizzleImage3DPass(
+    const Device& device_, Scheduler& scheduler_, DescriptorPool& descriptor_pool_,
+    ComputePassDescriptorQueue& compute_pass_descriptor_queue_)
+    : ComputePass(device_, scheduler_, descriptor_pool_, UNSWIZZLE_DESCRIPTOR_SET_BINDINGS,
+                  UNSWIZZLE_DESCRIPTOR_UPDATE_TEMPLATE, UNSWIZZLE_BANK_INFO,
+                  COMPUTE_PUSH_CONSTANT_RANGE<sizeof(BlockLinearSwizzle3DParams)>,
+                  UnswizzleSpv(device_, BLOCK_LINEAR_UNSWIZZLE_3D_COMP_SPV,
+                               BLOCK_LINEAR_UNSWIZZLE_3D_NONARROW_COMP_SPV)),
+      scheduler{scheduler_}, compute_pass_descriptor_queue{compute_pass_descriptor_queue_} {}
+
+BlockLinearUnswizzleImage3DPass::~BlockLinearUnswizzleImage3DPass() = default;
+
+void BlockLinearUnswizzleImage3DPass::Unswizzle(
+    Image& image, const StagingBufferRef& map,
+    std::span<const VideoCommon::SwizzleParameters> swizzles) {
+    using namespace VideoCommon::Accelerated;
+    scheduler.RequestOutsideRenderPassOperationContext();
+    const VkPipeline vk_pipeline = *pipeline;
+    const VkImageAspectFlags aspect_mask = image.AspectMask();
+    const VkImage vk_image = image.Handle();
+    const bool is_initialized = image.ExchangeInitialization();
+    RecordUnswizzleEntryBarrier(scheduler, vk_pipeline, vk_image, aspect_mask, is_initialized);
+
+    for (const VideoCommon::SwizzleParameters& swizzle : swizzles) {
+        const size_t input_offset = swizzle.buffer_offset + map.offset;
+        const u32 num_dispatches_x = Common::DivCeil(swizzle.num_tiles.width, 16U);
+        const u32 num_dispatches_y = Common::DivCeil(swizzle.num_tiles.height, 8U);
+        const u32 num_dispatches_z = Common::DivCeil(swizzle.num_tiles.depth, 8U);
+
+        compute_pass_descriptor_queue.Acquire(scheduler, 2);
+        compute_pass_descriptor_queue.AddBuffer(map.buffer, input_offset,
+                                                image.guest_size_bytes - swizzle.buffer_offset);
+        compute_pass_descriptor_queue.AddImage(image.StorageImageView(swizzle.level));
+        const void* const descriptor_data{compute_pass_descriptor_queue.UpdateData()};
+
+        const auto params = MakeBlockLinearSwizzle3DParams(swizzle, image.info);
+        scheduler.Record([this, num_dispatches_x, num_dispatches_y, num_dispatches_z, params,
+                          descriptor_data](vk::CommandBuffer cmdbuf) {
+            const VkDescriptorSet set = descriptor_allocator.Commit();
+            device.GetLogical().UpdateDescriptorSet(set, *descriptor_template, descriptor_data);
+            cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_COMPUTE, *layout, 0, set, {});
+            cmdbuf.PushConstants(*layout, VK_SHADER_STAGE_COMPUTE_BIT, params);
+            cmdbuf.Dispatch(num_dispatches_x, num_dispatches_y, num_dispatches_z);
+        });
+    }
+    RecordUnswizzleExitBarrier(scheduler, vk_image, aspect_mask);
+}
+
+PitchUnswizzlePass::PitchUnswizzlePass(
+    const Device& device_, Scheduler& scheduler_, DescriptorPool& descriptor_pool_,
+    ComputePassDescriptorQueue& compute_pass_descriptor_queue_)
+    : ComputePass(device_, scheduler_, descriptor_pool_, UNSWIZZLE_DESCRIPTOR_SET_BINDINGS,
+                  UNSWIZZLE_DESCRIPTOR_UPDATE_TEMPLATE, UNSWIZZLE_BANK_INFO,
+                  COMPUTE_PUSH_CONSTANT_RANGE<sizeof(PitchUnswizzlePushConstants)>,
+                  UnswizzleSpv(device_, PITCH_UNSWIZZLE_COMP_SPV,
+                               PITCH_UNSWIZZLE_NONARROW_COMP_SPV)),
+      scheduler{scheduler_}, compute_pass_descriptor_queue{compute_pass_descriptor_queue_} {}
+
+PitchUnswizzlePass::~PitchUnswizzlePass() = default;
+
+void PitchUnswizzlePass::Unswizzle(Image& image, const StagingBufferRef& map,
+                                   std::span<const VideoCommon::SwizzleParameters> swizzles) {
+    scheduler.RequestOutsideRenderPassOperationContext();
+    const VkPipeline vk_pipeline = *pipeline;
+    const VkImageAspectFlags aspect_mask = image.AspectMask();
+    const VkImage vk_image = image.Handle();
+    const bool is_initialized = image.ExchangeInitialization();
+    RecordUnswizzleEntryBarrier(scheduler, vk_pipeline, vk_image, aspect_mask, is_initialized);
+
+    const u32 bytes_per_block = VideoCore::Surface::BytesPerBlock(image.info.format);
+    const u32 pitch = image.info.pitch;
+    for (const VideoCommon::SwizzleParameters& swizzle : swizzles) {
+        const size_t input_offset = swizzle.buffer_offset + map.offset;
+        const u32 num_dispatches_x = Common::DivCeil(swizzle.num_tiles.width, 32U);
+        const u32 num_dispatches_y = Common::DivCeil(swizzle.num_tiles.height, 32U);
+
+        compute_pass_descriptor_queue.Acquire(scheduler, 2);
+        compute_pass_descriptor_queue.AddBuffer(map.buffer, input_offset,
+                                                image.guest_size_bytes - swizzle.buffer_offset);
+        compute_pass_descriptor_queue.AddImage(image.StorageImageView(swizzle.level));
+        const void* const descriptor_data{compute_pass_descriptor_queue.UpdateData()};
+
+        const PitchUnswizzlePushConstants params{
+            .origin{0, 0},
+            .destination{0, 0},
+            .bytes_per_block = bytes_per_block,
+            .pitch = pitch,
+        };
+        scheduler.Record([this, num_dispatches_x, num_dispatches_y, params,
+                          descriptor_data](vk::CommandBuffer cmdbuf) {
+            const VkDescriptorSet set = descriptor_allocator.Commit();
+            device.GetLogical().UpdateDescriptorSet(set, *descriptor_template, descriptor_data);
+            cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_COMPUTE, *layout, 0, set, {});
+            cmdbuf.PushConstants(*layout, VK_SHADER_STAGE_COMPUTE_BIT, params);
+            cmdbuf.Dispatch(num_dispatches_x, num_dispatches_y, 1);
+        });
+    }
+    RecordUnswizzleExitBarrier(scheduler, vk_image, aspect_mask);
+}
+
 } // namespace Vulkan

src/video_core/renderer_vulkan/vk_compute_pass.h

@@ -164,4 +164,49 @@ private:
     ComputePassDescriptorQueue& compute_pass_descriptor_queue;
 };
 
+class BlockLinearUnswizzle2DPass final : public ComputePass {
+public:
+    explicit BlockLinearUnswizzle2DPass(
+        const Device& device_, Scheduler& scheduler_, DescriptorPool& descriptor_pool_,
+        ComputePassDescriptorQueue& compute_pass_descriptor_queue_);
+    ~BlockLinearUnswizzle2DPass();
+
+    void Unswizzle(Image& image, const StagingBufferRef& map,
+                   std::span<const VideoCommon::SwizzleParameters> swizzles);
+
+private:
+    Scheduler& scheduler;
+    ComputePassDescriptorQueue& compute_pass_descriptor_queue;
+};
+
+class BlockLinearUnswizzleImage3DPass final : public ComputePass {
+public:
+    explicit BlockLinearUnswizzleImage3DPass(
+        const Device& device_, Scheduler& scheduler_, DescriptorPool& descriptor_pool_,
+        ComputePassDescriptorQueue& compute_pass_descriptor_queue_);
+    ~BlockLinearUnswizzleImage3DPass();
+
+    void Unswizzle(Image& image, const StagingBufferRef& map,
+                   std::span<const VideoCommon::SwizzleParameters> swizzles);
+
+private:
+    Scheduler& scheduler;
+    ComputePassDescriptorQueue& compute_pass_descriptor_queue;
+};
+
+class PitchUnswizzlePass final : public ComputePass {
+public:
+    explicit PitchUnswizzlePass(const Device& device_, Scheduler& scheduler_,
+                                DescriptorPool& descriptor_pool_,
+                                ComputePassDescriptorQueue& compute_pass_descriptor_queue_);
+    ~PitchUnswizzlePass();
+
+    void Unswizzle(Image& image, const StagingBufferRef& map,
+                   std::span<const VideoCommon::SwizzleParameters> swizzles);
+
+private:
+    Scheduler& scheduler;
+    ComputePassDescriptorQueue& compute_pass_descriptor_queue;
+};
+
 } // namespace Vulkan

src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp

@@ -694,9 +694,11 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) {
         const size_t num_vertex_arrays = (std::min)(
             Maxwell::NumVertexArrays, static_cast<size_t>(device.GetMaxVertexInputBindings()));
         for (size_t index = 0; index < num_vertex_arrays; ++index) {
-            const bool instanced = key.state.binding_divisors[index] != 0;
+            const bool instanced = ((key.state.enabled_divisors >> index) & 1) != 0;
-            const auto rate =
+            auto rate = VK_VERTEX_INPUT_RATE_VERTEX;
-                instanced ? VK_VERTEX_INPUT_RATE_INSTANCE : VK_VERTEX_INPUT_RATE_VERTEX;
+            if (instanced) {
+                rate = VK_VERTEX_INPUT_RATE_INSTANCE;
+            }
             vertex_bindings.push_back({
                 .binding = static_cast<u32>(index),
                 .stride = key.state.vertex_strides[index],
@@ -705,7 +707,7 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) {
             if (instanced) {
                 vertex_binding_divisors.push_back({
                     .binding = static_cast<u32>(index),
-                    .divisor = key.state.binding_divisors[index],
+                    .divisor = device.GetVertexAttribDivisor(key.state.binding_divisors[index]),
                 });
             }
         }

src/video_core/renderer_vulkan/vk_rasterizer.cpp

@@ -137,10 +137,12 @@ VkRect2D GetScissorState(const Maxwell& regs, size_t index, u32 up_scale = 1, u3
     max_y = (std::max)(max_y, 0);
 
     if (src.enable) {
-        scissor.offset.x = scale_up(src.min_x);
+        const s32 min_x = static_cast<s32>(src.min_x.Value());
+        const s32 max_x = static_cast<s32>(src.max_x.Value());
+        scissor.offset.x = scale_up(min_x);
         scissor.offset.y = scale_up(min_y);
-        scissor.extent.width = scale_up(src.max_x - src.min_x);
+        scissor.extent.width = scale_up((std::max)(max_x - min_x, 0));
-        scissor.extent.height = scale_up(max_y - min_y);
+        scissor.extent.height = scale_up((std::max)(max_y - min_y, 0));
     } else {
         scissor.offset.x = 0;
         scissor.offset.y = 0;
@@ -260,6 +262,7 @@ void RasterizerVulkan::PrepareDraw(bool is_indexed, Func&& draw_func) {
 }
 
 void RasterizerVulkan::Draw(bool is_indexed, u32 instance_count) {
+    buffer_cache.SetDrawInstanceCount(instance_count);
     PrepareDraw(is_indexed, [this, is_indexed, instance_count] {
         const auto& draw_state = maxwell3d->draw_manager.draw_state;
         const u32 num_instances{instance_count};
@@ -295,6 +298,7 @@ void RasterizerVulkan::Draw(bool is_indexed, u32 instance_count) {
 void RasterizerVulkan::DrawIndirect() {
     const auto& params = maxwell3d->draw_manager.indirect_state;
     buffer_cache.SetDrawIndirect(&params);
+    buffer_cache.SetDrawInstanceCount(0);
     PrepareDraw(params.is_indexed, [this, &params] {
         const auto indirect_buffer = buffer_cache.GetDrawIndirectBuffer();
         const auto& buffer = indirect_buffer.first;
@@ -1917,13 +1921,19 @@ void RasterizerVulkan::UpdateVertexInput(Tegra::Engines::Maxwell3D::Regs& regs)
     for (u32 binding = 0; binding < max_bindings; ++binding) {
         const auto& input_binding{regs.vertex_streams[binding]};
         const bool is_instanced{regs.vertex_stream_instances.IsInstancingEnabled(binding)};
+        auto input_rate = VK_VERTEX_INPUT_RATE_VERTEX;
+        u32 divisor = 1;
+        if (is_instanced) {
+            input_rate = VK_VERTEX_INPUT_RATE_INSTANCE;
+            divisor = device.GetVertexAttribDivisor(input_binding.frequency);
+        }
         bindings.push_back({
             .sType = VK_STRUCTURE_TYPE_VERTEX_INPUT_BINDING_DESCRIPTION_2_EXT,
             .pNext = nullptr,
             .binding = binding,
             .stride = input_binding.stride,
-            .inputRate = is_instanced ? VK_VERTEX_INPUT_RATE_INSTANCE : VK_VERTEX_INPUT_RATE_VERTEX,
+            .inputRate = input_rate,
-            .divisor = is_instanced ? input_binding.frequency : 1,
+            .divisor = divisor,
         });
     }
 

src/video_core/renderer_vulkan/vk_scheduler.cpp

@@ -449,7 +449,9 @@ void Scheduler::EndRenderPass()
                                          | VK_ACCESS_COLOR_ATTACHMENT_READ_BIT
                                          | VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT
                                          | VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT
-                                         | VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT,
+                                         | VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT
+                                         | VK_ACCESS_TRANSFER_READ_BIT
+                                         | VK_ACCESS_TRANSFER_WRITE_BIT,
                         .oldLayout = VK_IMAGE_LAYOUT_GENERAL,
                         .newLayout = VK_IMAGE_LAYOUT_GENERAL,
                         .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
@@ -460,7 +462,7 @@ void Scheduler::EndRenderPass()
             }
             cmdbuf.EndRenderPass();
             cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT | VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT |
-                                   VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT, vk::PIPELINE_STAGE_GRAPHICS_COMPUTE,
+                                   VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT, vk::PIPELINE_STAGE_GRAPHICS_COMPUTE_TRANSFER,
                                    0, nullptr, nullptr, vk::Span(barriers.data(), num_images));
             if (has_transform_feedback) {
                 static constexpr VkMemoryBarrier XFB_OUTPUT_BARRIER{

src/video_core/renderer_vulkan/vk_texture_cache.cpp

@@ -160,6 +160,55 @@ constexpr VkBorderColor ConvertBorderColor(const std::array<float, 4>& color) {
           info.size.depth == 1;
 }
 
+[[nodiscard]] PixelFormat UnswizzleViewFormat(u32 bytes_per_block) {
+    switch (bytes_per_block) {
+    case 1:
+        return PixelFormat::R8_UINT;
+    case 2:
+        return PixelFormat::R16_UINT;
+    case 4:
+        return PixelFormat::R32_UINT;
+    case 8:
+        return PixelFormat::R32G32_UINT;
+    case 16:
+        return PixelFormat::R32G32B32A32_UINT;
+    default:
+        return PixelFormat::Invalid;
+    }
+}
+
+constexpr u32 UNSWIZZLE_WORKGROUP_INVOCATIONS = 32 * 32;
+
+[[nodiscard]] bool SupportsAcceleratedUnswizzleDevice(const Device& device) {
+    return device.IsKhrImageFormatListSupported() &&
+           device.GetMaxComputeWorkGroupInvocations() >= UNSWIZZLE_WORKGROUP_INVOCATIONS;
+}
+
+[[nodiscard]] bool SupportsAcceleratedUnswizzle(const Device& device, const ImageInfo& info) {
+    if (!SupportsAcceleratedUnswizzleDevice(device)) {
+        return false;
+    }
+    if (info.num_samples > 1) {
+        return false;
+    }
+    if (info.type != ImageType::e2D && info.type != ImageType::e3D &&
+        info.type != ImageType::Linear) {
+        return false;
+    }
+    const PixelFormat view_format =
+        UnswizzleViewFormat(VideoCore::Surface::BytesPerBlock(info.format));
+    if (view_format == PixelFormat::Invalid) {
+        return false;
+    }
+    if (!VideoCore::Surface::IsViewCompatible(info.format, view_format, false, true)) {
+        return false;
+    }
+    const auto host_format =
+        MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, false, view_format);
+    return device.IsFormatSupported(host_format.format, VK_FORMAT_FEATURE_STORAGE_IMAGE_BIT,
+                                    FormatType::Optimal);
+}
+
 [[nodiscard]] VkImageCreateInfo MakeImageCreateInfo(const Device& device, const ImageInfo& info,
                                                     std::optional<VkFormat> format_override = {}) {
     auto format_info =
@@ -248,8 +297,18 @@ constexpr VkBorderColor ConvertBorderColor(const std::array<float, 4>& color) {
     return allocator.CreateImage(image_ci);
 }
 
+[[nodiscard]] VkImageViewType StorageViewType(ImageType type) {
+    if (type == ImageType::e3D) {
+        return VK_IMAGE_VIEW_TYPE_3D;
+    }
+    if (type == ImageType::Linear) {
+        return VK_IMAGE_VIEW_TYPE_2D;
+    }
+    return VK_IMAGE_VIEW_TYPE_2D_ARRAY;
+}
+
 [[nodiscard]] vk::ImageView MakeStorageView(const vk::Device& device, u32 level, VkImage image,
-                                            VkFormat format) {
+                                            VkFormat format, VkImageViewType view_type) {
     static constexpr VkImageViewUsageCreateInfo storage_image_view_usage_create_info{
         .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_USAGE_CREATE_INFO,
         .pNext = nullptr,
@@ -260,7 +319,7 @@ constexpr VkBorderColor ConvertBorderColor(const std::array<float, 4>& color) {
         .pNext = &storage_image_view_usage_create_info,
         .flags = 0,
         .image = image,
-        .viewType = VK_IMAGE_VIEW_TYPE_2D_ARRAY,
+        .viewType = view_type,
         .format = format,
         .components{
             .r = VK_COMPONENT_SWIZZLE_IDENTITY,
@@ -658,18 +717,11 @@ void CopyBufferToImage(vk::CommandBuffer cmdbuf, VkBuffer src_buffer, VkImage im
             .subresourceRange = subresource_range,
     };
 
-    cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT |
+    cmdbuf.PipelineBarrier(vk::PIPELINE_STAGE_GRAPHICS_COMPUTE, VK_PIPELINE_STAGE_TRANSFER_BIT, 0,
-                           VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT |
-                           VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, 0,
                            read_barrier);
     cmdbuf.CopyBufferToImage(src_buffer, image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, copies);
-    // TODO: Move this to another API
+    cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, vk::PIPELINE_STAGE_GRAPHICS_COMPUTE, 0,
-    cmdbuf.PipelineBarrier(
+                           nullptr, nullptr, write_barrier);
-            VK_PIPELINE_STAGE_TRANSFER_BIT,
-            VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT |
-            VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT |
-            VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT,
-            0, nullptr, nullptr, write_barrier);
 }
 
 [[nodiscard]] VkImageBlit MakeImageBlit(const Region2D& dst_region, const Region2D& src_region,
@@ -968,6 +1020,14 @@ TextureCacheRuntime::TextureCacheRuntime(const Device& device_, Scheduler& sched
         bl3d_unswizzle_pass.emplace(device, scheduler, descriptor_pool,
                                    staging_buffer_pool, compute_pass_descriptor_queue);
     }
+    if (SupportsAcceleratedUnswizzleDevice(device)) {
+        bl_unswizzle_2d_pass.emplace(device, scheduler, descriptor_pool,
+                                     compute_pass_descriptor_queue);
+        bl_unswizzle_image_3d_pass.emplace(device, scheduler, descriptor_pool,
+                                           compute_pass_descriptor_queue);
+        pitch_unswizzle_pass.emplace(device, scheduler, descriptor_pool,
+                                     compute_pass_descriptor_queue);
+    }
 }
 
 void TextureCacheRuntime::Finish() {
@@ -1902,16 +1962,12 @@ Image::Image(TextureCacheRuntime& runtime_, const ImageInfo& info_, GPUVAddr gpu
     if (runtime->device.HasDebuggingToolAttached()) {
         original_image.SetObjectNameEXT(VideoCommon::Name(*this).c_str());
     }
+    if (False(flags & VideoCommon::ImageFlagBits::Converted) &&
+        SupportsAcceleratedUnswizzle(runtime->device, info)) {
+        flags |= VideoCommon::ImageFlagBits::AcceleratedUpload;
+    }
     current_image = &Image::original_image;
     storage_image_views.resize(info.resources.levels);
-    if (WillUseAcceleratedAstcDecode(runtime->device, info)) {
-        const auto& device = runtime->device.GetLogical();
-        const VkFormat storage_format = VK_FORMAT_A8B8G8R8_UNORM_PACK32;
-        for (s32 level = 0; level < info.resources.levels; ++level) {
-            storage_image_views[level] =
-                MakeStorageView(device, level, *original_image, storage_format);
-        }
-    }
 }
 
 Image::Image(const VideoCommon::NullImageParams& params) : VideoCommon::ImageBase{params} {}
@@ -2035,7 +2091,7 @@ void Image::UploadMemory(VkBuffer buffer, VkDeviceSize offset,
                                                 temp_vk_image, info.format, info.num_samples,
                                                 {image_copies.data(), image_copies.size()}, false);
         }
-        initialized = true;
+        InitializationFor(current_image) = true;
         runtime->ReleaseMsaaScratchImage(temp_vk_image);
 
         if (is_rescaled) {
@@ -2056,7 +2112,7 @@ void Image::UploadMemory(VkBuffer buffer, VkDeviceSize offset,
     const VkBuffer src_buffer = buffer;
     const VkImage vk_image = *original_image;
     const VkImageAspectFlags vk_aspect_mask = aspect_mask;
-    const bool was_initialized = std::exchange(initialized, true);
+    const bool was_initialized = std::exchange(InitializationFor(&Image::original_image), true);
 
     scheduler->Record([src_buffer, vk_image, vk_aspect_mask, was_initialized,
                        vk_copies](vk::CommandBuffer cmdbuf) {
@@ -2321,16 +2377,46 @@ void Image::DownloadMemory(const StagingBufferRef& map, std::span<const BufferIm
     DownloadMemory(buffers, offsets, copies);
 }
 
+std::vector<vk::ImageView>& Image::StorageViewsFor(vk::Image Image::*image) {
+    if (image == &Image::scaled_image) {
+        if (scaled_storage_image_views.empty()) {
+            scaled_storage_image_views.resize(info.resources.levels);
+        }
+        return scaled_storage_image_views;
+    }
+    return storage_image_views;
+}
+
+bool& Image::InitializationFor(vk::Image Image::*image) noexcept {
+    if (image == &Image::scaled_image) {
+        return scaled_initialized;
+    }
+    return original_initialized;
+}
+
 VkImageView Image::StorageImageView(s32 level) noexcept {
-    auto& view = storage_image_views[level];
+    const bool astc_decode = WillUseAcceleratedAstcDecode(runtime->device, info);
+    const bool unswizzle_upload =
+        !astc_decode && True(flags & ImageFlagBits::AcceleratedUpload);
+    vk::Image Image::*target = current_image;
+    if (astc_decode || unswizzle_upload) {
+        target = &Image::original_image;
+    }
+    auto& view = StorageViewsFor(target)[level];
     if (!view) {
         auto format_info =
             MaxwellToVK::SurfaceFormat(runtime->device, FormatType::Optimal, true, info.format);
-        if (WillUseAcceleratedAstcDecode(runtime->device, info)) {
+        if (astc_decode) {
             format_info.format = VK_FORMAT_A8B8G8R8_UNORM_PACK32;
         }
-        view = MakeStorageView(runtime->device.GetLogical(), level, *(this->*current_image),
+        if (unswizzle_upload) {
-                               format_info.format);
+            const PixelFormat view_format =
+                UnswizzleViewFormat(VideoCore::Surface::BytesPerBlock(info.format));
+            format_info = MaxwellToVK::SurfaceFormat(runtime->device, FormatType::Optimal, false,
+                                                     view_format);
+        }
+        view = MakeStorageView(runtime->device.GetLogical(), level, *(this->*target),
+                               format_info.format, StorageViewType(info.type));
     }
     return *view;
 }
@@ -2370,6 +2456,7 @@ bool Image::ScaleUp(bool ignore) {
     }
     if (NeedsScaleHelper()) {
         if (!BlitScaleHelper(true)) {
+            flags &= ~ImageFlagBits::Rescaled;
             current_image = &Image::original_image;
             return false;
         }
@@ -2559,6 +2646,10 @@ ImageView::ImageView(TextureCacheRuntime& runtime, const VideoCommon::ImageViewI
     if (device->IsExtAstcDecodeModeSupported() && IsLdrAstcFormat(format_info.format)) {
         view_next = &astc_decode_mode;
     }
+    auto subresource_range = MakeSubresourceRange(aspect_mask, info.range);
+    if (True(flags & VideoCommon::ImageViewFlagBits::Slice)) {
+        subresource_range.levelCount = 1;
+    }
     const VkImageViewCreateInfo create_info{
         .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO,
         .pNext = view_next,
@@ -2567,7 +2658,7 @@ ImageView::ImageView(TextureCacheRuntime& runtime, const VideoCommon::ImageViewI
         .viewType = VkImageViewType{},
         .format = format_info.format,
         .components = swizzle_mapping,
-        .subresourceRange = MakeSubresourceRange(aspect_mask, info.range),
+        .subresourceRange = subresource_range,
     };
     const auto create = [&](TextureType tex_type, std::optional<u32> num_layers) {
         VkImageViewCreateInfo ci{create_info};
@@ -3141,6 +3232,19 @@ void TextureCacheRuntime::AccelerateImageUpload(
         return astc_decoder_pass->Assemble(image, map, swizzles);
     }
 
+    if (bl_unswizzle_2d_pass && image.info.type == ImageType::e2D) {
+        return bl_unswizzle_2d_pass->Unswizzle(image, map, swizzles);
+    }
+
+    if (bl_unswizzle_image_3d_pass && image.info.type == ImageType::e3D &&
+        !IsPixelFormatBCn(image.info.format)) {
+        return bl_unswizzle_image_3d_pass->Unswizzle(image, map, swizzles);
+    }
+
+    if (pitch_unswizzle_pass && image.info.type == ImageType::Linear) {
+        return pitch_unswizzle_pass->Unswizzle(image, map, swizzles);
+    }
+
     if (!Settings::values.gpu_unswizzle_enabled.GetValue() || !bl3d_unswizzle_pass) {
         if (IsPixelFormatBCn(image.info.format) && image.info.type == ImageType::e3D) {
             ASSERT(false && "GPU unswizzle is disabled for BCn 3D texture");

src/video_core/renderer_vulkan/vk_texture_cache.h

@@ -159,6 +159,9 @@ public:
     std::optional<ASTCDecoderPass> astc_decoder_pass;
 
     std::optional<BlockLinearUnswizzle3DPass> bl3d_unswizzle_pass;
+    std::optional<BlockLinearUnswizzle2DPass> bl_unswizzle_2d_pass;
+    std::optional<BlockLinearUnswizzleImage3DPass> bl_unswizzle_image_3d_pass;
+    std::optional<PitchUnswizzlePass> pitch_unswizzle_pass;
     const Settings::ResolutionScalingInfo& resolution;
     std::array<std::vector<VkFormat>, VideoCore::Surface::MaxPixelFormat> view_formats;
 
@@ -350,7 +353,7 @@ public:
 
     /// Returns true when the image is already initialized and mark it as initialized
     [[nodiscard]] bool ExchangeInitialization() noexcept {
-        return std::exchange(initialized, true);
+        return std::exchange(InitializationFor(current_image), true);
     }
 
     VkImageView StorageImageView(s32 level) noexcept;
@@ -370,6 +373,10 @@ private:
 
     bool NeedsScaleHelper() const;
 
+    std::vector<vk::ImageView>& StorageViewsFor(vk::Image Image::*image);
+
+    bool& InitializationFor(vk::Image Image::*image) noexcept;
+
     Scheduler* scheduler{};
     TextureCacheRuntime* runtime{};
 
@@ -387,8 +394,10 @@ private:
     vk::Image Image::*current_image{};
 
     std::vector<vk::ImageView> storage_image_views;
+    std::vector<vk::ImageView> scaled_storage_image_views;
     VkImageAspectFlags aspect_mask = 0;
-    bool initialized = false;
+    bool original_initialized = false;
+    bool scaled_initialized = false;
 
     std::optional<Framebuffer> scale_framebuffer;
     std::optional<Framebuffer> normal_framebuffer;

src/video_core/texture_cache/accelerated_swizzle.h

@@ -1,3 +1,6 @@
+// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
 // SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
 // SPDX-License-Identifier: GPL-2.0-or-later
 
@@ -23,8 +26,8 @@ struct BlockLinearSwizzle2DParams {
 };
 
 struct BlockLinearSwizzle3DParams {
-    std::array<u32, 3> origin;
+    alignas(16) std::array<u32, 3> origin;
-    std::array<s32, 3> destination;
+    alignas(16) std::array<s32, 3> destination;
     u32 bytes_per_block_log2;
     u32 slice_size;
     u32 block_size;

src/video_core/texture_cache/texture_cache.h

@@ -20,6 +20,7 @@
 #include "video_core/engines/kepler_compute.h"
 #include "video_core/guest_memory.h"
 #include "video_core/host1x/gpu_device_memory_manager.h"
+#include "video_core/texture_cache/accelerated_swizzle.h"
 #include "video_core/texture_cache/image_view_base.h"
 #include "video_core/texture_cache/samples_helper.h"
 #include "video_core/texture_cache/texture_cache_base.h"
@@ -279,11 +280,11 @@ void TextureCache<P>::CheckFeedbackLoop(std::span<const ImageViewInOut> views) {
 
             const ImageId view_image_id = slot_image_views[view.id].image_id;
             {
-                bool is_continue = false;
+                bool is_feedback = false;
                 for (size_t i = 0; i < 8; ++i)
-                    is_continue |= (rt_active_mask & (1u << i)) && view_image_id == rt_image_id[i];
+                    is_feedback |= (rt_active_mask & (1u << i)) && view_image_id == rt_image_id[i];
-                if (is_continue)
+                if (is_feedback)
-                    continue;
+                    return true;
             }
             if (depth_active && view_image_id == rt_depth_image_id) {
                 return true;
@@ -626,14 +627,25 @@ void TextureCache<P>::DownloadMemory(DAddr cpu_addr, size_t size) {
     std::ranges::sort(images, [this](ImageId lhs, ImageId rhs) {
         return slot_images[lhs].modification_tick < slot_images[rhs].modification_tick;
     });
+    size_t total_size_bytes = 0;
+    for (const ImageId image_id : images) {
+        total_size_bytes += slot_images[image_id].unswizzled_size_bytes;
+    }
+    auto download_map = runtime.DownloadStagingBuffer(total_size_bytes);
     for (const ImageId image_id : images) {
         Image& image = slot_images[image_id];
-        auto map = runtime.DownloadStagingBuffer(image.unswizzled_size_bytes);
         const auto copies = FixSmallVectorADL(FullDownloadCopies(image.info));
-        image.DownloadMemory(map, copies);
+        image.DownloadMemory(download_map, copies);
-        runtime.Finish();
+        download_map.offset += image.unswizzled_size_bytes;
-        SwizzleImage(*gpu_memory, image.gpu_addr, image.info, copies, map.mapped_span,
+    }
+    runtime.Finish();
+    std::span<u8> download_span = download_map.mapped_span;
+    for (const ImageId image_id : images) {
+        const ImageBase& image = slot_images[image_id];
+        const auto copies = FixSmallVectorADL(FullDownloadCopies(image.info));
+        SwizzleImage(*gpu_memory, image.gpu_addr, image.info, copies, download_span,
                      swizzle_data_buffer);
+        download_span = download_span.subspan(image.unswizzled_size_bytes);
     }
 }
 
@@ -1122,6 +1134,12 @@ void TextureCache<P>::RefreshContents(Image& image, ImageId image_id) {
 
     TrackImage(image, image_id);
 
+    if (image.info.rescaleable &&
+        IsRegionGpuModified(image.cpu_addr, image.guest_size_bytes)) {
+        runtime.TransitionImageLayout(image);
+        return;
+    }
+
     if (image.info.num_samples > 1 && !runtime.CanUploadMSAA()) {
         LOG_WARNING(HW_GPU, "MSAA image uploads are not implemented");
         runtime.TransitionImageLayout(image);
@@ -1157,8 +1175,7 @@ void TextureCache<P>::UploadImageContents(Image& image, StagingBuffer& staging)
     const GPUVAddr gpu_addr = image.gpu_addr;
 
     if (True(image.flags & ImageFlagBits::AcceleratedUpload)) {
-        gpu_memory->ReadBlock(gpu_addr, mapped_span.data(), mapped_span.size_bytes(),
+        gpu_memory->ReadBlockUnsafe(gpu_addr, mapped_span.data(), image.guest_size_bytes);
-                              VideoCommon::CacheType::NoTextureCache);
         const auto uploads = FullUploadSwizzles(image.info);
         runtime.AccelerateImageUpload(image, staging, FixSmallVectorADL(uploads), 0, 0);
         return;
@@ -1265,6 +1282,9 @@ ImageId TextureCache<P>::FindImage(const ImageInfo& info, GPUVAddr gpu_addr,
 
 template <class P>
 bool TextureCache<P>::ImageCanRescale(ImageBase& image) {
+    if (!Settings::values.resolution_info.active) {
+        return false;
+    }
     if (!image.info.rescaleable) {
         return false;
     }
@@ -1352,12 +1372,12 @@ void TextureCache<P>::QueueAsyncDecode(Image& image, ImageId image_id) {
     decode->image_id = image_id;
     async_decodes.push_back(std::move(decode));
 
-    std::vector<u8> local_unswizzle_data_buffer(image.unswizzled_size_bytes, 0);
+    Common::ScratchBuffer<u8> local_unswizzle_data_buffer(image.unswizzled_size_bytes);
     Tegra::Memory::GpuGuestMemory<u8, Tegra::Memory::GuestMemoryFlags::UnsafeRead> swizzle_data(*gpu_memory, image.gpu_addr, image.guest_size_bytes, &swizzle_data_buffer);
     auto copies = UnswizzleImage(*gpu_memory, image.gpu_addr, image.info, swizzle_data, local_unswizzle_data_buffer);
     const size_t out_size = MapSizeBytes(image);
 
-    auto func = [out_size, copies, info = image.info,
+    auto func = [out_size, copies = std::move(copies), info = image.info,
                  input = std::move(local_unswizzle_data_buffer),
                  async_decode = decode_ptr]() mutable {
         async_decode->decoded_data.resize_destructive(out_size);
@@ -1426,18 +1446,16 @@ void TextureCache<P>::TickAsyncUnswizzle() {
     Image& image = slot_images[task.image_id];
 
     if (!task.initialized) {
-        task.total_size = MapSizeBytes(image);
+        task.total_size = image.guest_size_bytes;
         task.staging_buffer = runtime.UploadStagingBuffer(task.total_size, true);
 
-        const auto& info = image.info;
+        const auto layout = FullUploadSwizzles(task.info);
-        const u32 bytes_per_block = BytesPerBlock(info.format);
+        const auto params =
-        const u32 width_blocks = Common::DivCeil(info.size.width, 4u);
+            VideoCommon::Accelerated::MakeBlockLinearSwizzle3DParams(layout.front(), task.info);
-        const u32 height_blocks = Common::DivCeil(info.size.height, 4u);
+        task.bytes_per_slice = params.slice_size;
-
+        task.chunked = task.info.block.depth == 0;
-        const u32 stride = width_blocks * bytes_per_block;
-        const u32 aligned_height = height_blocks;
-        task.bytes_per_slice = static_cast<size_t>(stride) * aligned_height;
         task.last_submitted_offset = 0;
+        task.slices_submitted = 0;
         task.initialized = true;
     }
 
@@ -1452,31 +1470,39 @@ void TextureCache<P>::TickAsyncUnswizzle() {
             if (copy_amount == 0) copy_amount = task.bytes_per_slice;
         }
 
-        gpu_memory->ReadBlock(image.gpu_addr + task.current_offset,
+        gpu_memory->ReadBlockUnsafe(image.gpu_addr + task.current_offset,
-                              task.staging_buffer.mapped_span.data() + task.current_offset,
+                                    task.staging_buffer.mapped_span.data() + task.current_offset,
-                              copy_amount);
+                                    copy_amount);
         task.current_offset += copy_amount;
     }
 
     const bool is_final_batch = task.current_offset >= task.total_size;
-    const size_t bytes_ready = task.current_offset - task.last_submitted_offset;
-    const u32 complete_slices = static_cast<u32>(bytes_ready / task.bytes_per_slice);
 
-    if (complete_slices >= swizzle_slices_per_batch || (is_final_batch && complete_slices > 0)) {
+    if (task.chunked) {
-        const u32 z_start = static_cast<u32>(task.last_submitted_offset / task.bytes_per_slice);
+        const size_t bytes_ready = task.current_offset - task.last_submitted_offset;
-        const u32 slices_to_process = (std::min)(complete_slices, swizzle_slices_per_batch);
+        const u32 complete_slices = static_cast<u32>(bytes_ready / task.bytes_per_slice);
-        const u32 z_count = (std::min)(slices_to_process, image.info.size.depth - z_start);
 
-        if (z_count > 0) {
+        if (complete_slices >= swizzle_slices_per_batch || (is_final_batch && complete_slices > 0)) {
-            const auto uploads = FullUploadSwizzles(task.info);
+            const u32 z_start = task.slices_submitted;
-            runtime.AccelerateImageUpload(image, task.staging_buffer, FixSmallVectorADL(uploads), z_start, z_count);
+            const u32 slices_to_process = (std::min)(complete_slices, swizzle_slices_per_batch);
-            task.last_submitted_offset += (static_cast<size_t>(z_count) * task.bytes_per_slice);
+            const u32 z_count = (std::min)(slices_to_process, image.info.size.depth - z_start);
+
+            if (z_count > 0) {
+                const auto uploads = FullUploadSwizzles(task.info);
+                runtime.AccelerateImageUpload(image, task.staging_buffer,
+                                              FixSmallVectorADL(uploads), z_start, z_count);
+                task.last_submitted_offset += static_cast<size_t>(z_count) * task.bytes_per_slice;
+                task.slices_submitted += z_count;
+            }
         }
+    } else if (is_final_batch && task.slices_submitted == 0) {
+        const auto uploads = FullUploadSwizzles(task.info);
+        runtime.AccelerateImageUpload(image, task.staging_buffer, FixSmallVectorADL(uploads), 0,
+                                      image.info.size.depth);
+        task.slices_submitted = image.info.size.depth;
     }
 
-    // Check if complete
+    const bool all_slices_submitted = task.slices_submitted >= image.info.size.depth;
-    const u32 slices_submitted = static_cast<u32>(task.last_submitted_offset / task.bytes_per_slice);
-    const bool all_slices_submitted = slices_submitted >= image.info.size.depth;
 
     if (is_final_batch && all_slices_submitted) {
         runtime.FreeDeferredStagingBuffer(task.staging_buffer);

src/video_core/texture_cache/texture_cache_base.h

@@ -139,6 +139,8 @@ class TextureCache : public VideoCommon::ChannelSetupCaches<TextureCacheChannelI
         AsyncBuffer staging_buffer;
         size_t last_submitted_offset = 0;
         size_t bytes_per_slice;
+        u32 slices_submitted = 0;
+        bool chunked = false;
         bool initialized = false;
     };
 

src/video_core/vulkan_common/vulkan_device.cpp

@@ -1219,6 +1219,11 @@ bool Device::GetSuitability(bool requires_swapchain) {
             VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TRANSFORM_FEEDBACK_PROPERTIES_EXT;
         SetNext(next, properties.transform_feedback);
     }
+    if (extensions.vertex_attribute_divisor) {
+        properties.vertex_attribute_divisor.sType =
+            VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_VERTEX_ATTRIBUTE_DIVISOR_PROPERTIES_EXT;
+        SetNext(next, properties.vertex_attribute_divisor);
+    }
     if (extensions.maintenance5) {
         properties.maintenance5.sType =
             VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_MAINTENANCE_5_PROPERTIES_KHR;

src/video_core/vulkan_common/vulkan_device.h

@@ -70,6 +70,7 @@ VK_DEFINE_HANDLE(VmaAllocator)
     FEATURE(EXT, ProvokingVertex, PROVOKING_VERTEX, provoking_vertex)                              \
     FEATURE(EXT, Robustness2, ROBUSTNESS_2, robustness2)                                           \
     FEATURE(EXT, TransformFeedback, TRANSFORM_FEEDBACK, transform_feedback)                        \
+    FEATURE(EXT, VertexAttributeDivisor, VERTEX_ATTRIBUTE_DIVISOR, vertex_attribute_divisor)       \
     FEATURE(EXT, VertexInputDynamicState, VERTEX_INPUT_DYNAMIC_STATE, vertex_input_dynamic_state)  \
     FEATURE(KHR, Maintenance5, MAINTENANCE_5, maintenance5)                                        \
     FEATURE(KHR, Maintenance6, MAINTENANCE_6, maintenance6)                                        \
@@ -92,7 +93,6 @@ VK_DEFINE_HANDLE(VmaAllocator)
     EXTENSION(EXT, SHADER_STENCIL_EXPORT, shader_stencil_export)                                   \
     EXTENSION(EXT, SHADER_VIEWPORT_INDEX_LAYER, shader_viewport_index_layer)                       \
     EXTENSION(EXT, TOOLING_INFO, tooling_info)                                                     \
-    EXTENSION(EXT, VERTEX_ATTRIBUTE_DIVISOR, vertex_attribute_divisor)                             \
     EXTENSION(KHR, CREATE_RENDERPASS_2, create_renderpass2)                                        \
     EXTENSION(KHR, DEPTH_STENCIL_RESOLVE, depth_stencil_resolve)                                   \
     EXTENSION(KHR, DRAW_INDIRECT_COUNT, draw_indirect_count)                                       \
@@ -359,6 +359,7 @@ public:
 
 #define FN_MAX_LIMIT_LIST \
     FN_MAX_LIMIT_ELEM(ComputeSharedMemorySize) \
+    FN_MAX_LIMIT_ELEM(ComputeWorkGroupInvocations) \
     FN_MAX_LIMIT_ELEM(PerStageDescriptorSampledImages) \
     FN_MAX_LIMIT_ELEM(PerStageResources) \
     FN_MAX_LIMIT_ELEM(DescriptorSetSamplers) \
@@ -689,6 +690,32 @@ FN_MAX_LIMIT_LIST
         return features.host_query_reset.hostQueryReset != VK_FALSE;
     }
 
+    u32 GetMaxVertexAttribDivisor() const {
+        const u32 reported = properties.vertex_attribute_divisor.maxVertexAttribDivisor;
+        if (reported == 0) {
+            return 1;
+        }
+        return reported;
+    }
+
+    bool IsVertexAttributeInstanceRateZeroDivisorSupported() const {
+        return features.vertex_attribute_divisor.vertexAttributeInstanceRateZeroDivisor == VK_TRUE;
+    }
+
+    u32 GetVertexAttribDivisor(u32 frequency) const {
+        const u32 max_divisor = GetMaxVertexAttribDivisor();
+        if (frequency == 0) {
+            if (IsVertexAttributeInstanceRateZeroDivisorSupported()) {
+                return 0;
+            }
+            return max_divisor;
+        }
+        if (frequency > max_divisor) {
+            return max_divisor;
+        }
+        return frequency;
+    }
+
     /// Returns true if the device supports VK_EXT_transform_feedback.
     bool IsExtTransformFeedbackSupported() const {
         return extensions.transform_feedback;
@@ -1189,6 +1216,7 @@ private:
         VkPhysicalDeviceDescriptorBufferPropertiesEXT descriptor_buffer{};
         VkPhysicalDeviceSubgroupSizeControlProperties subgroup_size_control{};
         VkPhysicalDeviceTransformFeedbackPropertiesEXT transform_feedback{};
+        VkPhysicalDeviceVertexAttributeDivisorPropertiesEXT vertex_attribute_divisor{};
         VkPhysicalDeviceMaintenance5PropertiesKHR maintenance5{};
         VkPhysicalDeviceDepthStencilResolveProperties depth_stencil_resolve{};
         VkPhysicalDeviceCustomBorderColorPropertiesEXT custom_border_color{};