GS/HW: Invalidate cleared area if overlapping existing dirty

Add memcard filters for OutRun2 SP - SPECIAL TOURS.
Add memcard filters for Another Century's Episode 2 Special Vocal Version.
Add memcard filters for Armored Core - Last Raven.

.github/CONTRIBUTING.md

@@ -1,54 +0,0 @@
-# So you want to contribute to PCSX2? Great
-
-As a first step, please review these links as they'll help you understand how the development of PCSX2 works.
-
-*   [Just Starting Out](#just-starting-out)
-*   [Issue Reporting](#issue-reporting)
-*   [Pull Request Guidelines](#pull-request-guidelines)
-*   [General Documentation And Coding Strategies](#general-documentation-and-coding-strategies)
-*   [Tasks](#tasks)
-
-## Just Starting Out
-
-*   If you're unfamilar with git, check out this [brief introduction to Git](https://github.com/PCSX2/pcsx2/wiki/07-Git-survival-guide)
-*   [How to build PCSX2 for Windows](https://github.com/PCSX2/pcsx2/wiki/12-Building-on-Windows)
-*   [How to build PCSX2 for Linux](https://github.com/PCSX2/pcsx2/wiki/10-Building-on-Linux)
-
-## Issue Reporting
-
-*   [How to write a useful issue](https://github.com/PCSX2/pcsx2/wiki/How-to-create-useful-and-valid-issues)
-
-## Pull Request Guidelines
-
-The following is a list of *general* style recommendations that will make reviewing and merging easier:
-
-*   Commit Messages
-    *   Please try to prefix your commit message, indicating what area of the project was modified.
-        *   For example `gs: message...`.
-        *   Looking at the project's commit history will help with keeping prefixes consistent overtime, *there is no strictly enforced list*.
-
-    *   Try to keep messages brief and informative
-
-    *   Remove unnecessary commits and squash commits together when appropriate.
-        *   If you are not familiar with rebasing with git, check out the following resources:
-            *   CLI - https://thoughtbot.com/blog/git-interactive-rebase-squash-amend-rewriting-history
-            *   GUI (SourceTree) - https://www.atlassian.com/blog/sourcetree/interactive-rebase-sourcetree
-
-*   Code Styling and Formatting
-    *   [Consult the style guide](https://github.com/tadanokojin/pcsx2/blob/coding-guide/pcsx2/Docs/Coding_Guidelines.md)
-
-    *   Run `clang-format` using the configuration file in the root of the repository
-        *   Visual Studio Setup - https://devblogs.microsoft.com/cppblog/clangformat-support-in-visual-studio-2017-15-7-preview-1/
-        *   IMPORTANT - if you are running `clang-format` on unrelated changes (ie. formatting an entire file), please do so in a separate commit.
-            *   If you cannot scope your `clang-format` to just your changes and do not want to format unrelated code.  Try your best to stick with the existing formatting already established in the file in question.
-
-## General Documentation And Coding Strategies
-
-*   [Commenting Etiquette](https://github.com/PCSX2/pcsx2/wiki/06-Commenting-Etiquette)
-
-*   [Coding style](https://github.com/PCSX2/pcsx2/wiki/Code-Formatting-Guidelines)
-    *   [More comprehensive style-guide (Currently in Draft)](https://github.com/tadanokojin/pcsx2/blob/coding-guide/pcsx2/Docs/Coding_Guidelines.md)
-
-## Tasks
-
-*   [Issues](https://github.com/PCSX2/pcsx2/issues)

.github/ISSUE_TEMPLATE/emu_bug_report.yaml

@@ -66,6 +66,10 @@ body:
 
         Performance issues as a result of not meeting our hardware requirements are not valid.
 
+        Please read our troubleshooting pages and our issue reporting guide.
+        - [Troubleshooting page](https://pcsx2.net/docs/category/troubleshooting).
+        - [Issue reporting guide](https://pcsx2.net/docs/troubleshooting/identify).
+
         We are **not** accepting issues related to the **libretro** core. The libretro core is being maintained separately at this time
   - type: input
     id: rev

.github/PULL_REQUEST_TEMPLATE.md

@@ -6,3 +6,6 @@
 
 ### Suggested Testing Steps
 <!-- If applicable, including examples you've already tested with / recommendations for how to test further is very helpful! -->
+
+### Did you use AI to help find, test, or implement this issue or feature?
+<!-- Answer yes or no. If you answer yes, please provide a brief explanation how. -->

.github/workflows/scripts/linux/build-dependencies-qt.sh

@@ -18,7 +18,7 @@ LIBBACKTRACE=ad106d5fdd5d960bd33fae1c48a351af567fd075
 LIBJPEGTURBO=3.1.0
 LIBPNG=1.6.48
 LIBWEBP=1.5.0
-SDL=SDL3-3.2.12
+SDL=SDL3-3.2.14
 QT=6.9.0
 LZ4=1.10.0
 ZSTD=1.5.7
@@ -39,7 +39,7 @@ fd6f417fe9e3a071cf1424a5152d926a34c4a3c5070745470be6cf12a404ed79  $LIBBACKTRACE.
 9564c72b1dfd1d6fe6274c5f95a8d989b59854575d4bbee44ade7bc17aa9bc93  libjpeg-turbo-$LIBJPEGTURBO.tar.gz
 46fd06ff37db1db64c0dc288d78a3f5efd23ad9ac41561193f983e20937ece03  libpng-$LIBPNG.tar.xz
 7d6fab70cf844bf6769077bd5d7a74893f8ffd4dfb42861745750c63c2a5c92c  libwebp-$LIBWEBP.tar.gz
-9734f308e130c64a2b4df6bca5884c5aca577ee6c7c77ab3379474ea85e51f96  $SDL.tar.gz
+b7e7dc05011b88c69170fe18935487b2559276955e49113f8c1b6b72c9b79c1f  $SDL.tar.gz
 537512904744b35e232912055ccf8ec66d768639ff3abe5788d90d792ec5f48b  lz4-$LZ4.tar.gz
 eb33e51f49a15e023950cd7825ca74a4a2b43db8354825ac24fc1b7ee09e6fa3  zstd-$ZSTD.tar.gz
 c1800c2ea835801af04a05d4a32321d79a93954ee3ae2172bbeacf13d1f0598c  qtbase-everywhere-src-$QT.tar.xz

.github/workflows/scripts/linux/flatpak/modules/20-sdl3.json

@@ -14,8 +14,8 @@
   "sources": [
     {
       "type": "archive",
-      "url": "https://libsdl.org/release/SDL3-3.2.12.tar.gz",
-      "sha256": "9734f308e130c64a2b4df6bca5884c5aca577ee6c7c77ab3379474ea85e51f96"
+      "url": "https://libsdl.org/release/SDL3-3.2.14.tar.gz",
+      "sha256": "b7e7dc05011b88c69170fe18935487b2559276955e49113f8c1b6b72c9b79c1f"
     }
   ],
   "cleanup": [

.github/workflows/scripts/macos/build-dependencies-universal.sh

@@ -11,7 +11,7 @@ merge_binaries() {
 "
 	pushd "$X86DIR"
 	for X86BIN in $(find . -type f \( -name '*.dylib' -o -name '*.a' -o -perm +111 \)); do
-		if file "$X86DIR/$X86BIN" | grep "Mach-O " >/dev/null; then
+		if file "$X86DIR/$X86BIN" | grep "Mach-O.*x86_64" >/dev/null; then
 			ARMBIN="${ARMDIR}/${X86BIN}"
 			echo "Merge $ARMBIN to $X86BIN..."
 			lipo -create "$X86BIN" "$ARMBIN" -o "$X86BIN"
@@ -40,7 +40,7 @@ fi
 
 FREETYPE=2.13.3
 HARFBUZZ=11.2.0
-SDL=SDL3-3.2.12
+SDL=SDL3-3.2.14
 ZSTD=1.5.7
 LZ4=1.10.0
 LIBPNG=1.6.48
@@ -79,7 +79,7 @@ CMAKE_ARCH_UNIVERSAL=-DCMAKE_OSX_ARCHITECTURES="x86_64;arm64"
 cat > SHASUMS <<EOF
 0550350666d427c74daeb85d5ac7bb353acba5f76956395995311a9c6f063289  freetype-$FREETYPE.tar.xz
 16c0204704f3ebeed057aba100fe7db18d71035505cb10e595ea33d346457fc8  harfbuzz-$HARFBUZZ.tar.gz
-9734f308e130c64a2b4df6bca5884c5aca577ee6c7c77ab3379474ea85e51f96  $SDL.tar.gz
+b7e7dc05011b88c69170fe18935487b2559276955e49113f8c1b6b72c9b79c1f  $SDL.tar.gz
 eb33e51f49a15e023950cd7825ca74a4a2b43db8354825ac24fc1b7ee09e6fa3  zstd-$ZSTD.tar.gz
 537512904744b35e232912055ccf8ec66d768639ff3abe5788d90d792ec5f48b  lz4-$LZ4.tar.gz
 46fd06ff37db1db64c0dc288d78a3f5efd23ad9ac41561193f983e20937ece03  libpng-$LIBPNG.tar.xz
@@ -112,11 +112,11 @@ curl -C - -L \
 	-O "https://storage.googleapis.com/downloads.webmproject.org/releases/webp/libwebp-$LIBWEBP.tar.gz" \
 	-O "https://ffmpeg.org/releases/ffmpeg-$FFMPEG.tar.xz" \
 	-O "https://github.com/KhronosGroup/MoltenVK/archive/refs/tags/v$MOLTENVK.tar.gz" \
-	-O "https://download.qt.io/official_releases/qt/${QT%.*}/$QT/submodules/qtbase-everywhere-src-$QT.tar.xz" \
-	-O "https://download.qt.io/official_releases/qt/${QT%.*}/$QT/submodules/qtimageformats-everywhere-src-$QT.tar.xz" \
-	-O "https://download.qt.io/official_releases/qt/${QT%.*}/$QT/submodules/qtsvg-everywhere-src-$QT.tar.xz" \
-	-O "https://download.qt.io/official_releases/qt/${QT%.*}/$QT/submodules/qttools-everywhere-src-$QT.tar.xz" \
-	-O "https://download.qt.io/official_releases/qt/${QT%.*}/$QT/submodules/qttranslations-everywhere-src-$QT.tar.xz" \
+	-O "https://download.qt.io/archive/qt/${QT%.*}/$QT/submodules/qtbase-everywhere-src-$QT.tar.xz" \
+	-O "https://download.qt.io/archive/qt/${QT%.*}/$QT/submodules/qtimageformats-everywhere-src-$QT.tar.xz" \
+	-O "https://download.qt.io/archive/qt/${QT%.*}/$QT/submodules/qtsvg-everywhere-src-$QT.tar.xz" \
+	-O "https://download.qt.io/archive/qt/${QT%.*}/$QT/submodules/qttools-everywhere-src-$QT.tar.xz" \
+	-O "https://download.qt.io/archive/qt/${QT%.*}/$QT/submodules/qttranslations-everywhere-src-$QT.tar.xz" \
 	-o "shaderc-$SHADERC.tar.gz" "https://github.com/google/shaderc/archive/refs/tags/v$SHADERC.tar.gz" \
 	-o "shaderc-glslang-$SHADERC_GLSLANG.tar.gz" "https://github.com/KhronosGroup/glslang/archive/$SHADERC_GLSLANG.tar.gz" \
 	-o "shaderc-spirv-headers-$SHADERC_SPIRVHEADERS.tar.gz" "https://github.com/KhronosGroup/SPIRV-Headers/archive/$SHADERC_SPIRVHEADERS.tar.gz" \
@@ -213,8 +213,11 @@ echo "Installing libjpegturbo..."
 rm -fr "libjpeg-turbo-$LIBJPEGTURBO"
 tar xf "libjpeg-turbo-$LIBJPEGTURBO.tar.gz"
 cd "libjpeg-turbo-$LIBJPEGTURBO"
-cmake "${CMAKE_COMMON[@]}" "$CMAKE_ARCH_ARM64" -DENABLE_STATIC=OFF -DENABLE_SHARED=ON -B build-arm64
+cmake "${CMAKE_COMMON[@]}" "$CMAKE_ARCH_X64" -DENABLE_STATIC=OFF -DENABLE_SHARED=ON -B build
 make -C build "-j$NPROCS"
+cmake "${CMAKE_COMMON[@]}" "$CMAKE_ARCH_ARM64" -DENABLE_STATIC=OFF -DENABLE_SHARED=ON -B build-arm64
+make -C build-arm64 "-j$NPROCS"
+merge_binaries $(realpath build) $(realpath build-arm64)
 make -C build install
 cd ..
 
@@ -380,7 +383,7 @@ echo "Building PlutoVG..."
 rm -fr "plutovg-$PLUTOVG"
 tar xf "plutovg-$PLUTOVG.tar.gz"
 cd "plutovg-$PLUTOVG"
-cmake "${CMAKE_COMMON[@]}" -DBUILD_SHARED_LIBS=ON -DPLUTOVG_BUILD_EXAMPLES=OFF -B build
+cmake "${CMAKE_COMMON[@]}" "$CMAKE_ARCH_UNIVERSAL" -DBUILD_SHARED_LIBS=ON -DPLUTOVG_BUILD_EXAMPLES=OFF -B build
 make -C build "-j$NPROCS"
 make -C build install
 cd ..
@@ -389,7 +392,7 @@ echo "Building PlutoSVG..."
 rm -fr "plutosvg-$PLUTOSVG"
 tar xf "plutosvg-$PLUTOSVG.tar.gz"
 cd "plutosvg-$PLUTOSVG"
-cmake "${CMAKE_COMMON[@]}" -DBUILD_SHARED_LIBS=ON -DPLUTOSVG_ENABLE_FREETYPE=ON -DPLUTOSVG_BUILD_EXAMPLES=OFF -B build
+cmake "${CMAKE_COMMON[@]}" "$CMAKE_ARCH_UNIVERSAL" -DBUILD_SHARED_LIBS=ON -DPLUTOSVG_ENABLE_FREETYPE=ON -DPLUTOSVG_BUILD_EXAMPLES=OFF -B build
 make -C build "-j$NPROCS"
 make -C build install
 cd ..

.github/workflows/scripts/macos/build-dependencies.sh

@@ -22,7 +22,7 @@ fi
 
 FREETYPE=2.13.3
 HARFBUZZ=11.2.0
-SDL=SDL3-3.2.12
+SDL=SDL3-3.2.14
 ZSTD=1.5.7
 LZ4=1.10.0
 LIBPNG=1.6.48
@@ -59,7 +59,7 @@ CMAKE_COMMON=(
 cat > SHASUMS <<EOF
 0550350666d427c74daeb85d5ac7bb353acba5f76956395995311a9c6f063289  freetype-$FREETYPE.tar.xz
 16c0204704f3ebeed057aba100fe7db18d71035505cb10e595ea33d346457fc8  harfbuzz-$HARFBUZZ.tar.gz
-9734f308e130c64a2b4df6bca5884c5aca577ee6c7c77ab3379474ea85e51f96  $SDL.tar.gz
+b7e7dc05011b88c69170fe18935487b2559276955e49113f8c1b6b72c9b79c1f  $SDL.tar.gz
 eb33e51f49a15e023950cd7825ca74a4a2b43db8354825ac24fc1b7ee09e6fa3  zstd-$ZSTD.tar.gz
 537512904744b35e232912055ccf8ec66d768639ff3abe5788d90d792ec5f48b  lz4-$LZ4.tar.gz
 46fd06ff37db1db64c0dc288d78a3f5efd23ad9ac41561193f983e20937ece03  libpng-$LIBPNG.tar.xz

.github/workflows/scripts/releases/announce-release/index.js

@@ -66,7 +66,7 @@ const embed = new MessageEmbed()
   .setDescription("To download the latest or previous builds, [visit the official downloads page](https://pcsx2.net/downloads/).")
   .addFields(
     { name: 'Version', value: releaseInfo.tag_name, inline: true },
-    { name: 'Installation Steps', value: '[See Here](https://github.com/PCSX2/pcsx2/wiki/Nightly-Build-Usage-Guide)', inline: true },
+    { name: 'Installation Steps', value: '[See Here](https://pcsx2.net/docs/category/setup)', inline: true },
     { name: 'Included Changes', value: releaseInfo.body, inline: false }
   );
 console.log(embed);

.github/workflows/scripts/windows/build-dependencies-arm64.bat

@@ -46,7 +46,7 @@ set FREETYPE=2.13.3
 set HARFBUZZ=11.2.0
 set LIBJPEGTURBO=3.1.0
 set LIBPNG=1648
-set SDL=SDL3-3.2.12
+set SDL=SDL3-3.2.14
 set QT=6.9.0
 set QTMINOR=6.9
 set LZ4=1.10.0
@@ -68,7 +68,7 @@ call :downloadfile "harfbuzz-%HARFBUZZ%.zip" https://github.com/harfbuzz/harfbuz
 call :downloadfile "lpng%LIBPNG%.zip" https://download.sourceforge.net/libpng/lpng1648.zip 2e5f080360f77376eb2bfa9e2ed773b9c7728159aba47b638ad53ca839379040 || goto error
 call :downloadfile "libjpeg-turbo-%LIBJPEGTURBO%.tar.gz" "https://github.com/libjpeg-turbo/libjpeg-turbo/releases/download/%LIBJPEGTURBO%/libjpeg-turbo-%LIBJPEGTURBO%.tar.gz" 9564c72b1dfd1d6fe6274c5f95a8d989b59854575d4bbee44ade7bc17aa9bc93 || goto error
 call :downloadfile "libwebp-%WEBP%.tar.gz" "https://storage.googleapis.com/downloads.webmproject.org/releases/webp/libwebp-%WEBP%.tar.gz" 7d6fab70cf844bf6769077bd5d7a74893f8ffd4dfb42861745750c63c2a5c92c || goto error
-call :downloadfile "%SDL%.zip" "https://libsdl.org/release/%SDL%.zip" ebf1efd71527afbca6bdf2b0310caf726d00cc102c0e59ac86c1f8cd201f9593 || goto error
+call :downloadfile "%SDL%.zip" "https://libsdl.org/release/%SDL%.zip" 46a17d3ea71fe2580a7f43ca7da286c5b9106dd761e2fd5533bb113e5d86b633 || goto error
 call :downloadfile "qtbase-everywhere-src-%QT%.zip" "https://download.qt.io/official_releases/qt/%QTMINOR%/%QT%/submodules/qtbase-everywhere-src-%QT%.zip" 513df15a6365a40f6230ec9463ad8c71b824e181d4b661dac9707e103b24ae0c || goto error
 call :downloadfile "qtimageformats-everywhere-src-%QT%.zip" "https://download.qt.io/official_releases/qt/%QTMINOR%/%QT%/submodules/qtimageformats-everywhere-src-%QT%.zip" d428fd17a0d3f92c48a30f1d23806bf20352fbce2e80e5bbee27fa80576480ee || goto error
 call :downloadfile "qtsvg-everywhere-src-%QT%.zip" "https://download.qt.io/official_releases/qt/%QTMINOR%/%QT%/submodules/qtsvg-everywhere-src-%QT%.zip" 54bf06afeb67035f1c6afcd00beec755c0d776626b4cce9ab56992a55215ba69 || goto error
@@ -246,12 +246,21 @@ cmake --build . --parallel || goto error
 ninja install || goto error
 cd ..\.. || goto error
 
+if %DEBUG%==1 (
+  set KDDOCKWIDGETSBUILDSPEC=-DCMAKE_CONFIGURATION_TYPES="Release;Debug" -DCMAKE_CROSS_CONFIGS=all -DCMAKE_DEFAULT_BUILD_TYPE=Release -DCMAKE_DEFAULT_CONFIGS=all -G "Ninja Multi-Config"
+) else (
+  rem kddockwidgets slightly changes the name of the dll depending on if CMAKE_BUILD_TYPE or CMAKE_CONFIGURATION_TYPES is used
+  rem The dll name being kddockwidgets-qt62.dll or kddockwidgets-qt62.dll respectively
+  rem Always use CMAKE_CONFIGURATION_TYPES to give consistant naming
+  set KDDOCKWIDGETSBUILDSPEC=-DCMAKE_CONFIGURATION_TYPES=Release -DCMAKE_CROSS_CONFIGS=all -DCMAKE_DEFAULT_BUILD_TYPE=Release -DCMAKE_DEFAULT_CONFIGS=Release -G "Ninja Multi-Config"
+)
+
 echo "Building KDDockWidgets..."
 rmdir /S /Q "KDDockWidgets-%KDDOCKWIDGETS%"
 %SEVENZIP% x "KDDockWidgets-%KDDOCKWIDGETS%.zip" || goto error
 cd "KDDockWidgets-%KDDOCKWIDGETS%" || goto error
 %PATCH% -p1 < "%SCRIPTDIR%\..\common\kddockwidgets-dodgy-include.patch" || goto error
-cmake %ARM64TOOLCHAIN% -DCMAKE_BUILD_TYPE=Release -DCMAKE_PREFIX_PATH="%INSTALLDIR%" -DCMAKE_INSTALL_PREFIX="%INSTALLDIR%" -DKDDockWidgets_QT6=true -DKDDockWidgets_EXAMPLES=false -DKDDockWidgets_FRONTENDS=qtwidgets -B build -G Ninja || goto error
+cmake -B build %ARM64TOOLCHAIN% -DCMAKE_PREFIX_PATH="%INSTALLDIR%" -DCMAKE_INSTALL_PREFIX="%INSTALLDIR%" -DKDDockWidgets_QT6=true -DKDDockWidgets_EXAMPLES=false -DKDDockWidgets_FRONTENDS=qtwidgets %KDDOCKWIDGETSBUILDSPEC% || goto error
 cmake --build build --parallel || goto error
 ninja -C build install || goto error
 cd .. || goto error

.github/workflows/scripts/windows/build-dependencies.bat

@@ -44,7 +44,7 @@ set FREETYPE=2.13.3
 set HARFBUZZ=11.2.0
 set LIBJPEGTURBO=3.1.0
 set LIBPNG=1648
-set SDL=SDL3-3.2.12
+set SDL=SDL3-3.2.14
 set QT=6.9.0
 set QTMINOR=6.9
 set LZ4=1.10.0
@@ -66,7 +66,7 @@ call :downloadfile "harfbuzz-%HARFBUZZ%.zip" https://github.com/harfbuzz/harfbuz
 call :downloadfile "lpng%LIBPNG%.zip" https://download.sourceforge.net/libpng/lpng1648.zip 2e5f080360f77376eb2bfa9e2ed773b9c7728159aba47b638ad53ca839379040 || goto error
 call :downloadfile "libjpeg-turbo-%LIBJPEGTURBO%.tar.gz" "https://github.com/libjpeg-turbo/libjpeg-turbo/releases/download/%LIBJPEGTURBO%/libjpeg-turbo-%LIBJPEGTURBO%.tar.gz" 9564c72b1dfd1d6fe6274c5f95a8d989b59854575d4bbee44ade7bc17aa9bc93 || goto error
 call :downloadfile "libwebp-%WEBP%.tar.gz" "https://storage.googleapis.com/downloads.webmproject.org/releases/webp/libwebp-%WEBP%.tar.gz" 7d6fab70cf844bf6769077bd5d7a74893f8ffd4dfb42861745750c63c2a5c92c || goto error
-call :downloadfile "%SDL%.zip" "https://libsdl.org/release/%SDL%.zip" ebf1efd71527afbca6bdf2b0310caf726d00cc102c0e59ac86c1f8cd201f9593 || goto error
+call :downloadfile "%SDL%.zip" "https://libsdl.org/release/%SDL%.zip" 46a17d3ea71fe2580a7f43ca7da286c5b9106dd761e2fd5533bb113e5d86b633 || goto error
 call :downloadfile "qtbase-everywhere-src-%QT%.zip" "https://download.qt.io/official_releases/qt/%QTMINOR%/%QT%/submodules/qtbase-everywhere-src-%QT%.zip" 513df15a6365a40f6230ec9463ad8c71b824e181d4b661dac9707e103b24ae0c || goto error
 call :downloadfile "qtimageformats-everywhere-src-%QT%.zip" "https://download.qt.io/official_releases/qt/%QTMINOR%/%QT%/submodules/qtimageformats-everywhere-src-%QT%.zip" d428fd17a0d3f92c48a30f1d23806bf20352fbce2e80e5bbee27fa80576480ee || goto error
 call :downloadfile "qtsvg-everywhere-src-%QT%.zip" "https://download.qt.io/official_releases/qt/%QTMINOR%/%QT%/submodules/qtsvg-everywhere-src-%QT%.zip" 54bf06afeb67035f1c6afcd00beec755c0d776626b4cce9ab56992a55215ba69 || goto error
@@ -250,12 +250,21 @@ cmake --build . --parallel || goto error
 ninja install || goto error
 cd ..\.. || goto error
 
+if %DEBUG%==1 (
+  set KDDOCKWIDGETSBUILDSPEC=-DCMAKE_CONFIGURATION_TYPES="Release;Debug" -DCMAKE_CROSS_CONFIGS=all -DCMAKE_DEFAULT_BUILD_TYPE=Release -DCMAKE_DEFAULT_CONFIGS=all -G "Ninja Multi-Config"
+) else (
+  rem kddockwidgets slightly changes the name of the dll depending on if CMAKE_BUILD_TYPE or CMAKE_CONFIGURATION_TYPES is used
+  rem The dll name being kddockwidgets-qt62.dll or kddockwidgets-qt62.dll respectively
+  rem Always use CMAKE_CONFIGURATION_TYPES to give consistant naming
+  set KDDOCKWIDGETSBUILDSPEC=-DCMAKE_CONFIGURATION_TYPES=Release -DCMAKE_CROSS_CONFIGS=all -DCMAKE_DEFAULT_BUILD_TYPE=Release -DCMAKE_DEFAULT_CONFIGS=Release -G "Ninja Multi-Config"
+)
+
 echo "Building KDDockWidgets..."
 rmdir /S /Q "KDDockWidgets-%KDDOCKWIDGETS%"
 %SEVENZIP% x "KDDockWidgets-%KDDOCKWIDGETS%.zip" || goto error
 cd "KDDockWidgets-%KDDOCKWIDGETS%" || goto error
 %PATCH% -p1 < "%SCRIPTDIR%\..\common\kddockwidgets-dodgy-include.patch" || goto error
-cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_PREFIX_PATH="%INSTALLDIR%" -DCMAKE_INSTALL_PREFIX="%INSTALLDIR%" -DKDDockWidgets_QT6=true -DKDDockWidgets_EXAMPLES=false -DKDDockWidgets_FRONTENDS=qtwidgets -B build -G Ninja || goto error
+cmake -B build -DCMAKE_PREFIX_PATH="%INSTALLDIR%" -DCMAKE_INSTALL_PREFIX="%INSTALLDIR%" -DKDDockWidgets_QT6=true -DKDDockWidgets_EXAMPLES=false -DKDDockWidgets_FRONTENDS=qtwidgets %KDDOCKWIDGETSBUILDSPEC% || goto error
 cmake --build build --parallel || goto error
 ninja -C build install || goto error
 cd .. || goto error

.github/workflows/triage_pr.yml

@@ -19,7 +19,7 @@ jobs:
         pr-message: |-
           ## Thank you for submitting a contribution to PCSX2
 
-          As this is your first pull request, [please be aware of the contributing guidelines](https://github.com/PCSX2/pcsx2/blob/master/.github/CONTRIBUTING.md).
+          As this is your first pull request, [please be aware of the contributing guidelines](https://pcsx2.net/docs/contributing/).
 
           Additionally, as per recent changes in GitHub Actions, your pull request will need to be approved by a maintainer before GitHub Actions can run against it. [You can find more information about this change here.](https://github.blog/2021-04-22-github-actions-update-helping-maintainers-combat-bad-actors/)
 

README.md

@@ -24,4 +24,4 @@ Please note that a BIOS dump from a legitimately-owned PS2 console is required t
 
 PCSX2 supports translation into other languages using [Crowdin](https://crowdin.com/project/pcsx2-emulator).
 
-See the [Contributing Guidelines](https://github.com/PCSX2/pcsx2/blob/master/.github/CONTRIBUTING.md) or visit the [GitHub Wiki](https://github.com/PCSX2/pcsx2/wiki) for more info on how to contribute.
+See the [Contribution Guide](https://pcsx2.net/docs/contributing/) for more info on how to contribute.

bin/resources/game_controller_db.txt

@@ -1100,6 +1100,7 @@ xinput,XInput Controller,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,
 050000005769696d6f74652028303000,Wii Remote,a:b4,b:b5,back:b7,dpdown:b3,dpleft:b0,dpright:b1,dpup:b2,guide:b8,leftshoulder:b11,lefttrigger:b12,leftx:a0,lefty:a1,start:b6,x:b10,y:b9,platform:Mac OS X,
 050000005769696d6f74652028313800,Wii U Pro Controller,a:b16,b:b15,back:b7,dpdown:b12,dpleft:b13,dpright:b14,dpup:b11,guide:b8,leftshoulder:b19,leftstick:b23,lefttrigger:b21,leftx:a0,lefty:a1,rightshoulder:b20,rightstick:b24,righttrigger:b22,rightx:a2,righty:a3,start:b6,x:b18,y:b17,platform:Mac OS X,
 030000005e0400008e02000000000000,Xbox 360 Controller,a:b0,b:b1,back:b9,dpdown:b12,dpleft:b13,dpright:b14,dpup:b11,guide:b10,leftshoulder:b4,leftstick:b6,lefttrigger:a2,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b7,righttrigger:a5,rightx:a3,righty:a4,start:b8,x:b2,y:b3,platform:Mac OS X,
+030000005e0400008e02000010010000,Xbox 360 Controller,a:b0,b:b1,back:b9,dpdown:b12,dpleft:b13,dpright:b14,dpup:b11,guide:b10,leftshoulder:b4,leftstick:b6,lefttrigger:a2,leftx:a0,lefty:a1~,rightshoulder:b5,rightstick:b7,righttrigger:a5,rightx:a3,righty:a4~,start:b8,x:b2,y:b3,platform:Mac OS X,
 030000006f0e00000104000000000000,Xbox 360 Controller,a:b0,b:b1,back:b9,dpdown:b12,dpleft:b13,dpright:b14,dpup:b11,guide:b10,leftshoulder:b4,leftstick:b6,lefttrigger:a2,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b7,righttrigger:a5,rightx:a3,righty:a4,start:b8,x:b2,y:b3,platform:Mac OS X,
 03000000c6240000045d000000000000,Xbox 360 Controller,a:b0,b:b1,back:b9,dpdown:b12,dpleft:b13,dpright:b14,dpup:b11,guide:b10,leftshoulder:b4,leftstick:b6,lefttrigger:a2,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b7,righttrigger:a5,rightx:a3,righty:a4,start:b8,x:b2,y:b3,platform:Mac OS X,
 030000005e0400000a0b000000000000,Xbox Adaptive Controller,a:b0,b:b1,back:b9,dpdown:b12,dpleft:b13,dpright:b14,dpup:b11,guide:b10,leftshoulder:b4,leftstick:b6,lefttrigger:a2,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b7,righttrigger:a5,rightx:a3,righty:a4,start:b8,x:b2,y:b3,platform:Mac OS X,
@@ -1290,6 +1291,7 @@ xinput,XInput Controller,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,
 03000000558500001b06000010010000,GameSir G4 Pro,a:b0,b:b1,back:b10,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,leftshoulder:b6,leftstick:b13,lefttrigger:b8,leftx:a0,lefty:a1,rightshoulder:b7,rightstick:b14,righttrigger:b9,rightx:a2,righty:a3,start:b11,x:b3,y:b4,platform:Linux,
 05000000ac0500002d0200001b010000,GameSir G4s,a:b0,b:b1,back:b10,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b33,leftshoulder:b6,leftstick:b13,lefttrigger:a5,leftx:a0,lefty:a1,rightshoulder:b7,rightstick:b14,righttrigger:a4,rightx:a2,righty:a3,start:b11,x:b3,y:b4,platform:Linux,
 03000000ac0500007a05000011010000,GameSir G5,a:b0,b:b1,back:b10,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b12,leftshoulder:b6,leftstick:b13,lefttrigger:b8,leftx:a0,lefty:a1,rightshoulder:b7,rightstick:b16,righttrigger:b9,rightx:a2,righty:a3,start:b11,x:b3,y:b4,platform:Linux,
+03000000373500009710000001020000,GameSir Kaleid Flux,a:b0,b:b1,back:b10,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b12,leftshoulder:b6,leftstick:b13,lefttrigger:a5,leftx:a0,lefty:a1,misc1:b15,rightshoulder:b7,rightstick:b14,righttrigger:a4,rightx:a2,righty:a3,start:b11,x:b3,y:b4,platform:Linux,
 03000000bc2000005656000011010000,GameSir T4w,a:b1,b:b2,back:b8,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b12,leftshoulder:b4,leftstick:b10,lefttrigger:b6,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b11,righttrigger:b7,rightx:a2,righty:a3,start:b9,x:b0,y:b3,platform:Linux,
 03000000ac0500001a06000011010000,GameSir-T3 2.02,a:b0,b:b1,back:b10,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b15,leftshoulder:b6,leftstick:b13,lefttrigger:a5,leftx:a0,lefty:a1,rightshoulder:b7,rightstick:b14,righttrigger:a4,rightx:a2,righty:a3,start:b11,x:b3,y:b4,platform:Linux,
 0500000047532047616d657061640000,GameStop Gamepad,a:b0,b:b1,back:b8,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,leftshoulder:b4,leftstick:b10,lefttrigger:b6,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b11,righttrigger:b7,rightx:a2,righty:a3,start:b9,x:b2,y:b3,platform:Linux,
@@ -1765,6 +1767,7 @@ xinput,XInput Controller,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,
 050000005e040000e002000003090000,Xbox One Controller,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b10,leftshoulder:b4,leftstick:b8,lefttrigger:a2,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b9,righttrigger:a5,rightx:a3,righty:a4,start:b7,x:b2,y:b3,platform:Linux,
 050000005e040000fd02000003090000,Xbox One Controller,a:b0,b:b1,back:b15,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b16,leftshoulder:b6,leftstick:b13,lefttrigger:a5,leftx:a0,lefty:a1,rightshoulder:b7,rightstick:b14,righttrigger:a4,rightx:a2,righty:a3,start:b11,x:b3,y:b4,platform:Linux,
 050000005e040000fd02000030110000,Xbox One Controller,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b8,leftshoulder:b4,leftstick:b9,lefttrigger:a2,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b10,righttrigger:a5,rightx:a3,righty:a4,start:b7,x:b2,y:b3,platform:Linux,
+060000005e040000dd02000003020000,Xbox One Controller,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b8,leftshoulder:b4,leftstick:b9,lefttrigger:a2,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b10,righttrigger:a5,rightx:a3,righty:a4,start:b7,x:b2,y:b3,platform:Linux,
 050000005e040000e302000002090000,Xbox One Elite,a:b0,b:b1,back:b136,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,leftshoulder:b6,leftstick:b13,lefttrigger:a6,leftx:a0,lefty:a1,rightshoulder:b7,rightstick:b14,righttrigger:a5,rightx:a2,righty:a3,start:b11,x:b3,y:b4,platform:Linux,
 050000005e040000220b000013050000,Xbox One Elite 2 Controller,a:b0,b:b1,back:b10,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b12,leftshoulder:b6,leftstick:b13,lefttrigger:a5,leftx:a0,lefty:a1,rightshoulder:b7,rightstick:b14,righttrigger:a4,rightx:a2,righty:a3,start:b11,x:b3,y:b4,platform:Linux,
 050000005e040000050b000002090000,Xbox One Elite Series 2,a:b0,b:b1,back:b136,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,leftshoulder:b6,leftstick:b13,lefttrigger:a6,leftx:a0,lefty:a1,rightshoulder:b7,rightstick:b14,righttrigger:a5,rightx:a2,righty:a3,start:b11,x:b3,y:b4,platform:Linux,

bin/resources/shaders/dx11/tfx.fx

@@ -397,7 +397,11 @@ float4 fetch_raw_color(int2 xy)
 
 float4 fetch_c(int2 uv)
 {
+#if PS_TEX_IS_FB == 1
+	return RtTexture.Load(int3(uv, 0));
+#else
 	return Texture.Load(int3(uv, 0));
+#endif
 }
 
 //////////////////////////////////////////////////////////////////////

bin/resources/shaders/opengl/tfx_fs.glsl

@@ -330,7 +330,11 @@ vec4 fetch_raw_color()
 
 vec4 fetch_c(ivec2 uv)
 {
+#if PS_TEX_IS_FB == 1
+	return sample_from_rt();
+#else
 	return texelFetch(TextureSampler, ivec2(uv), 0);
+#endif
 }
 
 //////////////////////////////////////////////////////////////////////

cmake/BuildParameters.cmake

@@ -64,10 +64,13 @@ set(CMAKE_SHARED_LINKER_FLAGS_DEVEL "${CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFO}
 	CACHE STRING "Flags used for linking shared libraries during development builds" FORCE)
 set(CMAKE_EXE_LINKER_FLAGS_DEVEL "${CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO}"
 	CACHE STRING "Flags used for linking executables during development builds" FORCE)
+# Exclude Debug from the configurations we can import from
+set(CMAKE_MAP_IMPORTED_CONFIG_DEVEL "RelWithDebInfo" "Release" "MinSizeRel" "None" "NoConfig" ""
+	CACHE STRING "Configurations used when importing packages for development builds" FORCE)
 if(CMAKE_CONFIGURATION_TYPES)
 	list(INSERT CMAKE_CONFIGURATION_TYPES 0 Devel)
 endif()
-mark_as_advanced(CMAKE_C_FLAGS_DEVEL CMAKE_CXX_FLAGS_DEVEL CMAKE_LINKER_FLAGS_DEVEL CMAKE_SHARED_LINKER_FLAGS_DEVEL CMAKE_EXE_LINKER_FLAGS_DEVEL)
+mark_as_advanced(CMAKE_C_FLAGS_DEVEL CMAKE_CXX_FLAGS_DEVEL CMAKE_LINKER_FLAGS_DEVEL CMAKE_SHARED_LINKER_FLAGS_DEVEL CMAKE_EXE_LINKER_FLAGS_DEVEL CMAKE_MAP_IMPORTED_CONFIG_DEVEL)
 
 #-------------------------------------------------------------------------------
 # Select the architecture

common/Darwin/DarwinMisc.cpp

@@ -437,7 +437,7 @@ std::unique_ptr<SharedMemoryMappingArea> SharedMemoryMappingArea::Create(size_t
 {
 	pxAssertRel(Common::IsAlignedPow2(size, __pagesize), "Size is page aligned");
 
-	mach_vm_address_t alloc;
+	mach_vm_address_t alloc = 0;
 	const kern_return_t res =
 		mach_vm_map(mach_task_self(), &alloc, size, 0, VM_FLAGS_ANYWHERE,
 			MEMORY_OBJECT_NULL, 0, false, VM_PROT_NONE, VM_PROT_NONE, VM_INHERIT_NONE);

common/vsprops/LinkPCSX2Deps.props

@@ -4,8 +4,10 @@
   <ItemDefinitionGroup>
     <Link>
       <AdditionalLibraryDirectories>$(DepsLibDir);%(AdditionalLibraryDirectories)</AdditionalLibraryDirectories>
-      <AdditionalDependencies>%(AdditionalDependencies);freetype.lib;jpeg.lib;libpng16.lib;libwebp.lib;lz4.lib;SDL3.lib;zlib.lib;zstd.lib;kddockwidgets-qt62.lib;plutovg.lib;plutosvg.lib</AdditionalDependencies>
-   </Link>
+      <AdditionalDependencies>%(AdditionalDependencies);freetype.lib;jpeg.lib;libpng16.lib;libwebp.lib;lz4.lib;SDL3.lib;zlib.lib;zstd.lib;plutovg.lib;plutosvg.lib</AdditionalDependencies>
+      <AdditionalDependencies Condition="$(Configuration.Contains(Debug))">%(AdditionalDependencies);kddockwidgets-qt6d.lib;</AdditionalDependencies>
+      <AdditionalDependencies Condition="!$(Configuration.Contains(Debug))">%(AdditionalDependencies);kddockwidgets-qt6.lib;</AdditionalDependencies>
+    </Link>
   </ItemDefinitionGroup>
   <ItemGroup>
     <DepsDLLs Include="$(DepsBinDir)freetype.dll" />
@@ -19,7 +21,8 @@
     <DepsDLLs Include="$(DepsBinDir)shaderc_shared.dll" />
     <DepsDLLs Include="$(DepsBinDir)zlib1.dll" />
     <DepsDLLs Include="$(DepsBinDir)zstd.dll" />
-    <DepsDLLs Include="$(DepsBinDir)kddockwidgets-qt62.dll" />
+    <DepsDLLs Condition="$(Configuration.Contains(Debug))" Include="$(DepsBinDir)kddockwidgets-qt6d.dll" />
+    <DepsDLLs Condition="!$(Configuration.Contains(Debug))" Include="$(DepsBinDir)kddockwidgets-qt6.dll" />
     <DepsDLLs Include="$(DepsBinDir)plutovg.dll" />
     <DepsDLLs Include="$(DepsBinDir)plutosvg.dll" />
   </ItemGroup>

pcsx2-qt/Debugger/ThreadView.cpp

@@ -65,16 +65,17 @@ void ThreadView::openContextMenu(QPoint pos)
 
 void ThreadView::onDoubleClick(const QModelIndex& index)
 {
+	auto real_index = m_proxy_model->mapToSource(index);
 	switch (index.column())
 	{
 		case ThreadModel::ThreadColumns::ENTRY:
 		{
-			goToInMemoryView(m_model->data(index, Qt::UserRole).toUInt(), true);
+			goToInDisassembler(m_model->data(real_index, Qt::UserRole).toUInt(), true);
 			break;
 		}
 		default: // Default to PC
 		{
-			QModelIndex pc_index = m_model->index(index.row(), ThreadModel::ThreadColumns::PC);
+			QModelIndex pc_index = m_model->index(real_index.row(), ThreadModel::ThreadColumns::PC);
 			goToInDisassembler(m_model->data(pc_index, Qt::UserRole).toUInt(), true);
 			break;
 		}

pcsx2/CDVD/CDVD.cpp

@@ -1455,7 +1455,7 @@ static uint cdvdStartSeek(uint newsector, CDVD_MODE_TYPE mode, bool transition_t
 		}
 		isSeeking = true;
 	}
-	else if(!drive_speed_change_cycles)
+	else if (!drive_speed_change_cycles)
 	{
 		CDVD_LOG("CdSeek Begin > Contiguous block without seek - delta=%d sectors", delta);
 

pcsx2/CDVD/Windows/IOCtlSrc.cpp

@@ -222,7 +222,7 @@ bool IOCtlSrc::ReadDVDInfo()
 		{
 			Console.Warning("IOCTL_DVD_READ_STRUCTURE not supported");
 		}
-		else if(GetLastError() != ERROR_UNRECOGNIZED_MEDIA) // ERROR_UNRECOGNIZED_MEDIA means probably a CD or no disc
+		else if (GetLastError() != ERROR_UNRECOGNIZED_MEDIA) // ERROR_UNRECOGNIZED_MEDIA means probably a CD or no disc
 		{
 			Console.Warning("IOCTL Unknown Error %d", GetLastError());
 		}

pcsx2/CMakeLists.txt

@@ -1286,7 +1286,11 @@ function(setup_main_executable target)
 
 		# Copy dependency libraries.
 		set(DEPS_BINDIR "${CMAKE_SOURCE_DIR}/deps/bin")
-		set(DEPS_TO_COPY freetype.dll harfbuzz.dll jpeg62.dll libpng16.dll libsharpyuv.dll libwebp.dll lz4.dll SDL3.dll shaderc_shared.dll zlib1.dll zstd.dll kddockwidgets-qt62.dll plutovg.dll plutosvg.dll)
+		set(DEPS_TO_COPY freetype.dll harfbuzz.dll jpeg62.dll libpng16.dll libsharpyuv.dll libwebp.dll lz4.dll SDL3.dll shaderc_shared.dll zlib1.dll zstd.dll plutovg.dll plutosvg.dll)
+		set(DEPS_TO_COPY
+			$<IF:$<CONFIG:Debug>,kddockwidgets-qt6d.dll,kddockwidgets-qt6.dll>
+			${DEPS_TO_COPY}
+		)
 		foreach(DEP_TO_COPY ${DEPS_TO_COPY})
 			install(FILES "${DEPS_BINDIR}/${DEP_TO_COPY}" DESTINATION "${CMAKE_SOURCE_DIR}/bin")
 		endforeach()

pcsx2/DebugTools/Breakpoints.h

@@ -33,7 +33,7 @@ struct BreakPoint
 	bool temporary = false;
 	bool stepping = false;
 
-	bool hasCond;
+	bool hasCond = false;
 	BreakPointCond cond;
 	BreakPointCpu cpu;
 

pcsx2/Dmac.cpp

@@ -236,7 +236,7 @@ static __ri void DmaExec( void (*func)(), u32 mem, u32 value )
 				cpuClearInt( 10 );
 				QueuedDMA._u16 &= ~(1 << 10); //Clear any queued DMA requests for this channel
 			}
-			else if(channel == 2)
+			else if (channel == 2)
 			{
 				cpuClearInt( 11 );
 				QueuedDMA._u16 &= ~(1 << 11); //Clear any queued DMA requests for this channel
@@ -274,7 +274,7 @@ static __ri void DmaExec( void (*func)(), u32 mem, u32 value )
 	{
 		func();
 	}
-	else if(reg.chcr.STR)
+	else if (reg.chcr.STR)
 	{
 		//DevCon.Warning(L"32bit %s DMA Start while DMAC Disabled\n", ChcrName(mem));
 		QueuedDMA._u16 |= (1 << ChannelNumber(mem)); //Queue the DMA up to be started then the DMA's are Enabled and or the Suspend is lifted

pcsx2/Docs/PCSX2_FAQ.md

@@ -20,7 +20,11 @@ There are a number of ways to help the project, whether it be bug reporting, gam
 
 *   Want to make changes to emulator code? [Check out the issue tracker](https://github.com/PCSX2/pcsx2/issues), or [fork the PCSX2 repository and work on your own ideas](https://github.com/PCSX2/pcsx2).
 *   Want to patch games? [Check out the cheats and patches forum thread for inspiration](https://forums.pcsx2.net/Thread-Post-your-PCSX2-cheats-patches-here). There are other threads to find as well, [such as those dedicated to 60 FPS patches](https://forums.pcsx2.net/Thread-60-fps-codes) or [widescreen patches](https://forums.pcsx2.net/Thread-PCSX2-Widescreen-Game-Patches).
+<<<<<<< HEAD
 *   Want to report bugs you have discovered in your games? Head over to [the Bug Reporting section of the PCSX2 forums](https://forums.pcsx2.net/Forum-Bug-reporting), to [the PCSX2 Discord](https://pcsx2.net/discord), or to [the GitHub issues section](https://github.com/PCSX2/pcsx2/issues).
+=======
+*   Want to report bugs you have discovered in your games? [Head over to the Bug Reporting section of the PCSX2 forums](https://forums.pcsx2.net/Forum-Bug-reporting). Also check out our [Issue reporting guide](https://pcsx2.net/docs/troubleshooting/identify) for info on how to better report to us issues.
+>>>>>>> 091a4ee5a (docs: update in order to redirect from the GH wiki to our website)
 *   Want to update us on the compatibility of your games? [Take a look at the Public Compatibility List on the PCSX2 forums](https://forums.pcsx2.net/Forum-Public-compatibility-list)
 *   Want to improve the PCSX2 wiki? [Here is how to contribute](https://wiki.pcsx2.net/How_to_contribute)
 
@@ -120,9 +124,7 @@ If you are using a disc:
 Fast Boot, enabled by default, will directly mount and launch the game without first launching the PS2 BIOS. You may disable this in Settings > BIOS > Fast Boot if you wish to see the BIOS startup animation or if Fast Boot is causing an issue.
 
 ### Question 18: How do I build the PCSX2 source code?
-*   [Windows build guide](https://github.com/PCSX2/pcsx2/wiki/12-Building-on-Windows)
-*   [Linux build guide](https://github.com/PCSX2/pcsx2/wiki/10-Building-on-Linux)
-*   [macOS build guide](https://github.com/PCSX2/pcsx2/wiki/11-Building-on-MacOS)
+Have a nice [building guide](https://pcsx2.net/docs/advanced/building) :) .
 
 ### Question 19: When will the next version be released?
 It will be released when it is ready. Please don't waste your time and ours asking when.

pcsx2/GS/GSClut.cpp

@@ -339,7 +339,7 @@ void GSClut::Read(const GIFRegTEX0& TEX0)
 				break;
 			}
 		}
-		else if(TEX0.CPSM == PSMCT16 || TEX0.CPSM == PSMCT16S)
+		else if (TEX0.CPSM == PSMCT16 || TEX0.CPSM == PSMCT16S)
 		{
 			switch(TEX0.PSM)
 			{
@@ -462,7 +462,8 @@ void GSClut::GetAlphaMinMax32(int& amin_out, int& amax_out)
 {
 	// call only after Read32
 
-	pxAssert(!m_read.dirty);
+	if (m_read.dirty)
+		GL_INS("GSClut: GetAlphaMinMax32 m_read.dirty");
 
 	if (m_read.adirty)
 	{

pcsx2/GS/GSState.cpp

@@ -2260,7 +2260,7 @@ void GSState::Move()
 				vm[doff] = (vm[doff] & 0xff000000) | (vm[soff] & 0x00ffffff);
 			});
 		}
-		else // if(spsm.trbpp == 16)
+		else // if (spsm.trbpp == 16)
 		{
 			u16* vm = m_mem.vm16();
 			copy(dpo.assertSizesMatch(GSLocalMemory::swizzle16), spo.assertSizesMatch(GSLocalMemory::swizzle16), [vm](u32 doff, u32 soff)
@@ -3996,8 +3996,14 @@ GSState::TextureMinMaxResult GSState::GetTextureMinMax(GIFRegTEX0 TEX0, GIFRegCL
 			// If it's the start of the texture and our little adjustment is all that pushed it over, clamp it to 0.
 			// This stops the border check failing when using repeat but needed less than the full texture
 			// since this was making it take the full texture even though it wasn't needed.
-			if (!clamp_to_tsize && ((m_vt.m_min.t.floor() == GSVector4::zero()).mask() & 0x3) == 0x3)
-				st = st.max(GSVector4::zero());
+			if (!clamp_to_tsize)
+			{
+				const u32 mask = (m_vt.m_min.t.floor() == GSVector4::zero()).mask();
+				if (mask & 1) // X == 0
+					st.x = st.max(GSVector4::zero()).x;
+				if (mask & 2) // Y == 0
+					st.y = st.max(GSVector4::zero()).y;
+			}
 		}
 
 		// draw will get scissored, adjust UVs to suit

pcsx2/GS/Renderers/Common/GSDevice.cpp

@@ -493,6 +493,7 @@ GSTexture* GSDevice::FetchSurface(GSTexture::Type type, int width, int height, i
 			{
 				ERROR_LOG("GS: Memory allocation failure for {}x{} texture. Purging pool and retrying.", size.x, size.y);
 				PurgePool();
+				t = CreateSurface(type, size.x, size.y, levels, format);
 				if (!t)
 				{
 					ERROR_LOG("GS: Memory allocation failure for {}x{} texture after purging pool.", size.x, size.y);

pcsx2/GS/Renderers/Common/GSDevice.h

@@ -372,6 +372,7 @@ struct alignas(16) GSHWDrawConfig
 
 				// Others ways to fetch the texture
 				u32 channel : 3;
+				u32 channel_fb : 1;
 
 				// Dithering
 				u32 dither : 2;
@@ -410,7 +411,7 @@ struct alignas(16) GSHWDrawConfig
 		{
 			const u32 sw_blend_bits = blend_a | blend_b | blend_d;
 			const bool sw_blend_needs_rt = (sw_blend_bits != 0 && ((sw_blend_bits | blend_c) & 1u)) || ((a_masked & blend_c) != 0);
-			return tex_is_fb || fbmask || (date > 0 && date != 3) || sw_blend_needs_rt;
+			return channel_fb || tex_is_fb || fbmask || (date > 0 && date != 3) || sw_blend_needs_rt;
 		}
 
 		/// Disables color output from the pixel shader, this is done when all channels are masked.
@@ -934,7 +935,8 @@ public:
 	__fi bool IsPresentThrottleAllowed() const { return m_allow_present_throttle; }
 
 	__fi GSTexture* GetCurrent() const { return m_current; }
-
+	__fi GSTexture* GetMAD() const { return m_mad; }
+	
 	void Recycle(GSTexture* t);
 
 	/// Returns true if it's an OpenGL-based renderer.

pcsx2/GS/Renderers/Common/GSRenderer.cpp

@@ -118,6 +118,18 @@ bool GSRenderer::Merge(int field)
 	if (!tex[0] && !tex[1])
 	{
 		m_real_size = GSVector2i(0, 0);
+
+		// Clear out the MAD buffer as some remnants of the previously shown frame came be left over, causing a flash for one frame.
+		if (GSConfig.InterlaceMode == GSInterlaceMode::Automatic || GSConfig.InterlaceMode >= GSInterlaceMode::AdaptiveTFF)
+		{
+			GSTexture* mad_tex = g_gs_device->GetMAD();
+
+			if (mad_tex)
+			{
+				g_gs_device->ClearRenderTarget(mad_tex, 0);
+				mad_tex = nullptr;
+			}
+		}
 		return false;
 	}
 

pcsx2/GS/Renderers/DX11/GSDevice11.cpp

@@ -1236,36 +1236,18 @@ void GSDevice11::CopyRect(GSTexture* sTex, GSTexture* dTex, const GSVector4i& r,
 
 	g_perfmon.Put(GSPerfMon::TextureCopies, 1);
 
-	D3D11_BOX box = {(UINT)r.left, (UINT)r.top, 0U, (UINT)r.right, (UINT)r.bottom, 1U};
+	D3D11_BOX box = {static_cast<UINT>(r.left), static_cast<UINT>(r.top), 0U, static_cast<UINT>(r.right), static_cast<UINT>(r.bottom), 1U};
 
-	// DX api isn't happy if we pass a box for depth copy
-	// It complains that depth/multisample must be a full copy
-	// and asks us to use a NULL for the box
+	// DX11 doesn't support partial depth copy so we need to
+	// either pass a nullptr D3D11_BOX for a full depth copy or use CopyResource instead.
+	// Alternatively use shader copy StretchRect, or full depth copy with
+	// adjusting the scissor and UVs in the shader.
 	const bool depth = (sTex->GetType() == GSTexture::Type::DepthStencil);
 	auto pBox = depth ? nullptr : &box;
+	const u32 x = depth ? 0 : destX;
+	const u32 y = depth ? 0 : destY;
 
-	m_ctx->CopySubresourceRegion(*(GSTexture11*)dTex, 0, destX, destY, 0, *(GSTexture11*)sTex, 0, pBox);
-}
-
-void GSDevice11::CloneTexture(GSTexture* src, GSTexture** dest, const GSVector4i& rect)
-{
-	pxAssertMsg(src->GetType() == GSTexture::Type::DepthStencil || src->GetType() == GSTexture::Type::RenderTarget, "Source is RT or DS.");
-	CommitClear(src);
-
-	const int w = src->GetWidth();
-	const int h = src->GetHeight();
-
-	if (src->GetType() == GSTexture::Type::DepthStencil)
-	{
-		// DX11 requires that you copy the entire depth buffer.
-		*dest = CreateDepthStencil(w, h, src->GetFormat(), false);
-		CopyRect(src, *dest, GSVector4i(0, 0, w, h), 0, 0);
-	}
-	else
-	{
-		*dest = CreateRenderTarget(w, h, src->GetFormat(), false);
-		CopyRect(src, *dest, rect, rect.left, rect.top);
-	}
+	m_ctx->CopySubresourceRegion(*(GSTexture11*)dTex, 0, x, y, 0, *(GSTexture11*)sTex, 0, pBox);
 }
 
 void GSDevice11::StretchRect(GSTexture* sTex, const GSVector4& sRect, GSTexture* dTex, const GSVector4& dRect, ShaderConvert shader, bool linear)
@@ -2495,20 +2477,9 @@ D3D_SHADER_MACRO* GSDevice11::ShaderMacro::GetPtr()
 	return (D3D_SHADER_MACRO*)mout.data();
 }
 
-/// Checks that we weren't sent things we declared we don't support
-/// Clears things we don't support that can be quietly disabled
-static void preprocessSel(GSDevice11::PSSelector& sel)
-{
-	pxAssert(sel.write_rg  == 0); // Not supported, shouldn't be sent
-}
-
 void GSDevice11::RenderHW(GSHWDrawConfig& config)
 {
-	pxAssert(!config.require_full_barrier); // We always specify no support so it shouldn't request this
-	preprocessSel(config.ps);
-
-	GSVector2i rtsize = (config.rt ? config.rt : config.ds)->GetSize();
-
+	const GSVector2i rtsize = (config.rt ? config.rt : config.ds)->GetSize();
 	GSTexture* colclip_rt = g_gs_device->GetColorClipTexture();
 
 	if (colclip_rt)
@@ -2560,7 +2531,8 @@ void GSDevice11::RenderHW(GSHWDrawConfig& config)
 		StretchRect(colclip_rt ? colclip_rt : config.rt, GSVector4(config.drawarea) / GSVector4(rtsize).xyxy(),
 			primid_tex, GSVector4(config.drawarea), m_date.primid_init_ps[static_cast<u8>(config.datm)].get(), nullptr, false);
 	}
-	else if (config.destination_alpha != GSHWDrawConfig::DestinationAlphaMode::Off)
+	else if (config.destination_alpha == GSHWDrawConfig::DestinationAlphaMode::Stencil ||
+			 config.destination_alpha == GSHWDrawConfig::DestinationAlphaMode::StencilOne)
 	{
 		const GSVector4 src = GSVector4(config.drawarea) / GSVector4(config.ds->GetSize()).xyxy();
 		const GSVector4 dst = src * 2.0f - 1.0f;
@@ -2630,20 +2602,33 @@ void GSDevice11::RenderHW(GSHWDrawConfig& config)
 		PSSetShaderResource(1, config.pal);
 	}
 
-	GSTexture* rt_copy = nullptr;
-	if (config.require_one_barrier || (config.tex && config.tex == config.rt)) // Used as "bind rt" flag when texture barrier is unsupported.
+	GSTexture* draw_rt_clone = nullptr;
+
+	if (config.require_one_barrier || (config.tex && config.tex == config.rt))
 	{
-		// Bind the RT.This way special effect can use it.
-		// Do not always bind the rt when it's not needed,
-		// only bind it when effects use it such as fbmask emulation currently
-		// because we copy the frame buffer and it is quite slow.
-		CloneTexture(colclip_rt ? colclip_rt : config.rt, &rt_copy, config.drawarea);
-		if (rt_copy)
+		// Requires a copy of the RT.
+		// Used as "bind rt" flag when texture barrier is unsupported for tex is fb.
+		draw_rt_clone = CreateTexture(rtsize.x, rtsize.y, 1, colclip_rt ? GSTexture::Format::ColorClip : GSTexture::Format::Color, true);
+		if (draw_rt_clone)
 		{
+			CopyRect(colclip_rt ? colclip_rt : config.rt, draw_rt_clone, config.drawarea, config.drawarea.left, config.drawarea.top);
 			if (config.require_one_barrier)
-				PSSetShaderResource(2, rt_copy);
+				PSSetShaderResource(2, draw_rt_clone);
 			if (config.tex && config.tex == config.rt)
-				PSSetShaderResource(0, rt_copy);
+				PSSetShaderResource(0, draw_rt_clone);
+		}
+	}
+
+	GSTexture* draw_ds_clone = nullptr;
+
+	if (config.tex && config.tex == config.ds)
+	{
+		// DX requires a copy when sampling the depth buffer.
+		draw_ds_clone = CreateDepthStencil(rtsize.x, rtsize.y, config.ds->GetFormat(), false);
+		if (draw_ds_clone)
+		{
+			CopyRect(config.ds, draw_ds_clone, config.drawarea, config.drawarea.left, config.drawarea.top);
+			PSSetShaderResource(0, draw_ds_clone);
 		}
 	}
 
@@ -2681,7 +2666,6 @@ void GSDevice11::RenderHW(GSHWDrawConfig& config)
 
 	if (config.alpha_second_pass.enable)
 	{
-		preprocessSel(config.alpha_second_pass.ps);
 		if (config.cb_ps.FogColor_AREF.a != config.alpha_second_pass.ps_aref)
 		{
 			config.cb_ps.FogColor_AREF.a = config.alpha_second_pass.ps_aref;
@@ -2697,8 +2681,12 @@ void GSDevice11::RenderHW(GSHWDrawConfig& config)
 		DrawIndexedPrimitive();
 	}
 
-	if (rt_copy)
-		Recycle(rt_copy);
+	if (draw_rt_clone)
+		Recycle(draw_rt_clone);
+
+	if (draw_ds_clone)
+		Recycle(draw_ds_clone);
+
 	if (primid_tex)
 		Recycle(primid_tex);
 

pcsx2/GS/Renderers/DX11/GSDevice11.h

@@ -291,7 +291,6 @@ public:
 	std::unique_ptr<GSDownloadTexture> CreateDownloadTexture(u32 width, u32 height, GSTexture::Format format) override;
 
 	void CommitClear(GSTexture* t);
-	void CloneTexture(GSTexture* src, GSTexture** dest, const GSVector4i& rect);
 
 	void CopyRect(GSTexture* sTex, GSTexture* dTex, const GSVector4i& r, u32 destX, u32 destY) override;
 

pcsx2/GS/Renderers/DX12/GSDevice12.cpp

@@ -1087,7 +1087,14 @@ GSDevice::PresentResult GSDevice12::BeginPresent(bool frame_skip)
 		return PresentResult::DeviceLost;
 
 	if (frame_skip || !m_swap_chain)
+	{
+		if (!m_swap_chain)
+		{
+			ExecuteCommandList(WaitType::None);
+			InvalidateCachedState();
+		}
 		return PresentResult::FrameSkipped;
+	}
 
 	// Check if we lost exclusive fullscreen. If so, notify the host, so it can switch to windowed mode.
 	// This might get called repeatedly if it takes a while to switch back, that's the host's problem.
@@ -3821,6 +3828,7 @@ void GSDevice12::RenderHW(GSHWDrawConfig& config)
 	GSTexture12* draw_rt = static_cast<GSTexture12*>(config.rt);
 	GSTexture12* draw_ds = static_cast<GSTexture12*>(config.ds);
 	GSTexture12* draw_rt_clone = nullptr;
+	GSTexture12* draw_ds_clone = nullptr;
 
 	// Align the render area to 128x128, hopefully avoiding render pass restarts for small render area changes (e.g. Ratchet and Clank).
 	const GSVector2i rtsize(config.rt ? config.rt->GetSize() : config.ds->GetSize());
@@ -3876,7 +3884,7 @@ void GSDevice12::RenderHW(GSHWDrawConfig& config)
 	// bind textures before checking the render pass, in case we need to transition them
 	if (config.tex)
 	{
-		PSSetShaderResource(0, config.tex, config.tex != config.rt);
+		PSSetShaderResource(0, config.tex, config.tex != config.rt && config.tex != config.ds);
 		PSSetSampler(config.sampler);
 	}
 	if (config.pal)
@@ -3900,15 +3908,16 @@ void GSDevice12::RenderHW(GSHWDrawConfig& config)
 		}
 	}
 
-	if (config.require_one_barrier || (config.tex && config.tex == config.rt)) // Used as "bind rt" flag when texture barrier is unsupported.
+	if (config.require_one_barrier || (config.tex && config.tex == config.rt))
 	{
-		// requires a copy of the RT
+		// Requires a copy of the RT.
+		// Used as "bind rt" flag when texture barrier is unsupported for tex is fb.
 		draw_rt_clone = static_cast<GSTexture12*>(CreateTexture(rtsize.x, rtsize.y, 1, colclip_rt ? GSTexture::Format::ColorClip : GSTexture::Format::Color, true));
 		if (draw_rt_clone)
 		{
 			EndRenderPass();
 
-			GL_PUSH("Copy RT to temp texture for fbmask {%d,%d %dx%d}", config.drawarea.left, config.drawarea.top,
+			GL_PUSH("Copy RT to temp texture {%d,%d %dx%d}", config.drawarea.left, config.drawarea.top,
 				config.drawarea.width(), config.drawarea.height());
 
 			draw_rt_clone->SetState(GSTexture::State::Invalidated);
@@ -3920,6 +3929,23 @@ void GSDevice12::RenderHW(GSHWDrawConfig& config)
 		}
 	}
 
+	if (config.tex && config.tex == config.ds)
+	{
+		// DX requires a copy when sampling the depth buffer.
+		draw_ds_clone = static_cast<GSTexture12*>(CreateDepthStencil(rtsize.x, rtsize.y, config.ds->GetFormat(), false));
+		if (draw_ds_clone)
+		{
+			EndRenderPass();
+
+			GL_PUSH("Copy RT to temp texture {%d,%d %dx%d}", config.drawarea.left, config.drawarea.top,
+				config.drawarea.width(), config.drawarea.height());
+
+			draw_ds_clone->SetState(GSTexture::State::Invalidated);
+			CopyRect(config.ds, draw_ds_clone, config.drawarea, config.drawarea.left, config.drawarea.top);
+			PSSetShaderResource(0, draw_ds_clone, true);
+		}
+	}
+
 	// Switch to colclip target for colclip hw rendering
 	if (pipe.ps.colclip_hw)
 	{
@@ -4066,6 +4092,9 @@ void GSDevice12::RenderHW(GSHWDrawConfig& config)
 	if (draw_rt_clone)
 		Recycle(draw_rt_clone);
 
+	if (draw_ds_clone)
+		Recycle(draw_ds_clone);
+
 	if (date_image)
 		Recycle(date_image);
 

pcsx2/GS/Renderers/HW/GSRendererHW.cpp

@@ -1037,7 +1037,7 @@ float GSRendererHW::GetTextureScaleFactor()
 	return GetUpscaleMultiplier();
 }
 
-GSVector2i GSRendererHW::GetValidSize(const GSTextureCache::Source* tex)
+GSVector2i GSRendererHW::GetValidSize(const GSTextureCache::Source* tex, const bool is_shuffle)
 {
 	// Don't blindly expand out to the scissor size if we're not drawing to it.
 	// e.g. Burnout 3, God of War II, etc.
@@ -1088,10 +1088,9 @@ GSVector2i GSRendererHW::GetValidSize(const GSTextureCache::Source* tex)
 	// Early detection of texture shuffles. These double the input height because they're interpreting 64x32 C32 pages as 64x64 C16.
 	// Why? Well, we don't want to be doubling the heights of targets, but also we don't want to align C32 targets to 64 instead of 32.
 	// Yumeria's text breaks, and GOW goes to 512x448 instead of 512x416 if we don't.
-	const bool possible_texture_shuffle =
-		(tex && m_vt.m_primclass == GS_SPRITE_CLASS && frame_psm.bpp == 16 &&
+	const bool possible_texture_shuffle = tex && m_vt.m_primclass == GS_SPRITE_CLASS && frame_psm.bpp == 16 &&
 			GSLocalMemory::m_psm[m_cached_ctx.TEX0.PSM].bpp == 16 &&
-			(tex->m_32_bits_fmt ||
+			(is_shuffle || (tex->m_32_bits_fmt ||
 				(m_cached_ctx.TEX0.TBP0 != m_cached_ctx.FRAME.Block() && IsOpaque() && !(m_context->TEX1.MMIN & 1) &&
 					m_cached_ctx.FRAME.FBMSK && g_texture_cache->Has32BitTarget(m_cached_ctx.FRAME.Block()))));
 	if (possible_texture_shuffle)
@@ -1128,9 +1127,9 @@ GSVector2i GSRendererHW::GetValidSize(const GSTextureCache::Source* tex)
 	return  GSVector2i(width, height);
 }
 
-GSVector2i GSRendererHW::GetTargetSize(const GSTextureCache::Source* tex, const bool can_expand)
+GSVector2i GSRendererHW::GetTargetSize(const GSTextureCache::Source* tex, const bool can_expand, const bool is_shuffle)
 {
-	const GSVector2i valid_size = GetValidSize(tex);
+	const GSVector2i valid_size = GetValidSize(tex, is_shuffle);
 
 	return g_texture_cache->GetTargetSize(m_cached_ctx.FRAME.Block(), m_cached_ctx.FRAME.FBW, m_cached_ctx.FRAME.PSM, valid_size.x, valid_size.y, can_expand);
 }
@@ -2859,7 +2858,9 @@ void GSRendererHW::Draw()
 		if (!no_rt && GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].bpp == 16 && GSLocalMemory::m_psm[m_cached_ctx.TEX0.PSM].bpp >= 16 &&
 			(m_vt.m_primclass == GS_SPRITE_CLASS || (m_vt.m_primclass == GS_TRIANGLE_CLASS && (m_index.tail % 6) == 0 && TrianglesAreQuads(true) && m_index.tail > 6)))
 		{
-			if (GSLocalMemory::m_psm[m_cached_ctx.TEX0.PSM].bpp == 16)
+			// Tail check is to make sure we have enough strips to go all the way across the page, or if it's using a region clamp could be used to draw strips.
+			if (GSLocalMemory::m_psm[m_cached_ctx.TEX0.PSM].bpp == 16 && 
+				(m_index.tail >= (m_cached_ctx.TEX0.TBW * 2) || m_cached_ctx.TEX0.TBP0 == m_cached_ctx.FRAME.Block() || m_cached_ctx.CLAMP.WMS > CLAMP_CLAMP || m_cached_ctx.CLAMP.WMT > CLAMP_CLAMP))
 			{
 				const GSVertex* v = &m_vertex.buff[0];
 
@@ -2883,21 +2884,45 @@ void GSRendererHW::Draw()
 			}
 
 			// It's possible it's writing to an old 32bit target, but is actually just a 16bit copy, so let's make sure it's actually using a mask.
-			if (!shuffle_target && m_cached_ctx.FRAME.FBMSK)
+			if (!shuffle_target)
 			{
-				// FBW is going to be wrong for channel shuffling into a new target, so take it from the source.
-				FRAME_TEX0.U64 = 0;
-				FRAME_TEX0.TBP0 = m_cached_ctx.FRAME.Block();
-				FRAME_TEX0.TBW = m_cached_ctx.FRAME.FBW;
-				FRAME_TEX0.PSM = m_cached_ctx.FRAME.PSM;
+				bool shuffle_channel_reads = !m_cached_ctx.FRAME.FBMSK;
+				const u32 increment = (m_vt.m_primclass == GS_TRIANGLE_CLASS) ? 3 : 2;
+				const GSVertex* v = &m_vertex.buff[0];
 
-				GSTextureCache::Target* tgt = g_texture_cache->LookupTarget(FRAME_TEX0, GSVector2i(m_vt.m_max.p.x, m_vt.m_max.p.y), GetTextureScaleFactor(), GSTextureCache::RenderTarget, false,
-					fm, false, false, false, false, GSVector4i::zero(), true);
+				if (shuffle_channel_reads)
+				{
+					for (u32 i = 0; i < m_index.tail; i += increment)
+					{
+						const int first_u = (PRIM->FST ? v[i].U : static_cast<int>(v[i].ST.S / v[(increment == 2) ? i + 1 : i].RGBAQ.Q)) >> 4;
+						const int second_u = (PRIM->FST ? v[i + 1].U : static_cast<int>(v[i + 1].ST.S / v[i + 1].RGBAQ.Q)) >> 4;
+						const int vector_width = std::abs(v[i + 1].XYZ.X - v[i].XYZ.X) / 16;
+						const int tex_width = std::abs(second_u - first_u);
+						// & 7 just a quicker way of doing % 8
+						if ((vector_width & 7) != 0 || (tex_width & 7) != 0 || tex_width != vector_width)
+						{
+							shuffle_channel_reads = false;
+							break;
+						}
+					}
+				}
+				if (m_cached_ctx.FRAME.FBMSK || shuffle_channel_reads)
+				{
+					// FBW is going to be wrong for channel shuffling into a new target, so take it from the source.
+					FRAME_TEX0.U64 = 0;
+					FRAME_TEX0.TBP0 = m_cached_ctx.FRAME.Block();
+					FRAME_TEX0.TBW = m_cached_ctx.FRAME.FBW;
+					FRAME_TEX0.PSM = m_cached_ctx.FRAME.PSM;
 
-				if (tgt)
-					shuffle_target = tgt->m_32_bits_fmt;
+					GSTextureCache::Target* tgt = g_texture_cache->FindOverlappingTarget(FRAME_TEX0.TBP0, GSLocalMemory::GetEndBlockAddress(FRAME_TEX0.TBP0, FRAME_TEX0.TBW, FRAME_TEX0.PSM, m_r));
 
-				tgt = nullptr;
+					if (tgt)
+						shuffle_target = tgt->m_32_bits_fmt;
+					else
+						shuffle_target = shuffle_channel_reads;
+
+					tgt = nullptr;
+				}
 			}
 		}
 
@@ -2919,9 +2944,9 @@ void GSRendererHW::Draw()
 		}
 		else
 		{
-			src = tex_psm.depth ? g_texture_cache->LookupDepthSource(true, TEX0, env.TEXA, MIP_CLAMP, tmm.coverage, possible_shuffle, m_vt.IsLinear(), m_cached_ctx.FRAME.Block(), req_color, req_alpha) :
+			src = tex_psm.depth ? g_texture_cache->LookupDepthSource(true, TEX0, env.TEXA, MIP_CLAMP, tmm.coverage, possible_shuffle, m_vt.IsLinear(), m_cached_ctx.FRAME, req_color, req_alpha) :
 								  g_texture_cache->LookupSource(true, TEX0, env.TEXA, MIP_CLAMP, tmm.coverage, (GSConfig.HWMipmap || GSConfig.TriFilter == TriFiltering::Forced) ? &hash_lod_range : nullptr,
-					possible_shuffle, m_vt.IsLinear(), m_cached_ctx.FRAME.Block(), req_color, req_alpha);
+					possible_shuffle, m_vt.IsLinear(), m_cached_ctx.FRAME, req_color, req_alpha);
 
 			if (!src) [[unlikely]]
 			{
@@ -2941,7 +2966,13 @@ void GSRendererHW::Draw()
 			if (possible_shuffle && GSLocalMemory::m_psm[m_cached_ctx.TEX0.PSM].bpp != 16)
 				possible_shuffle &= draw_uses_target;
 
-			possible_shuffle &= src && (src->m_from_target != nullptr || (m_skip && possible_shuffle));
+			const bool shuffle_source = possible_shuffle && src && ((src->m_from_target != nullptr && GSLocalMemory::m_psm[src->m_from_target->m_TEX0.PSM].bpp != 16) || m_skip);
+
+			if (!shuffle_source && possible_shuffle)
+			{
+				const bool is_16bit_copy = m_cached_ctx.TEX0.TBP0 != m_cached_ctx.FRAME.Block() && shuffle_target && IsOpaque() && !(context->TEX1.MMIN & 1) && !src->m_32_bits_fmt && m_cached_ctx.FRAME.FBMSK;
+				possible_shuffle &= is_16bit_copy || (m_cached_ctx.TEX0.TBP0 == m_cached_ctx.FRAME.Block() && shuffle_target);
+			}
 			// We don't know the alpha range of direct sources when we first tried to optimize the alpha test.
 			// Moving the texture lookup before the ATST optimization complicates things a lot, so instead,
 			// recompute it, and everything derived from it again if it changes.
@@ -2987,7 +3018,7 @@ void GSRendererHW::Draw()
 	const bool can_expand = !(m_cached_ctx.ZBUF.ZMSK && output_black);
 
 	// Estimate size based on the scissor rectangle and height cache.
-	GSVector2i t_size = GetTargetSize(src, can_expand);
+	GSVector2i t_size = GetTargetSize(src, can_expand, possible_shuffle);
 	const GSVector4i t_size_rect = GSVector4i::loadh(t_size);
 
 	// Ensure draw rect is clamped to framebuffer size. Necessary for updating valid area.
@@ -3076,7 +3107,7 @@ void GSRendererHW::Draw()
 
 		if (!ds && m_cached_ctx.FRAME.FBP != m_cached_ctx.ZBUF.ZBP)
 		{
-			ds = g_texture_cache->CreateTarget(ZBUF_TEX0, t_size, GetValidSize(src), target_scale, GSTextureCache::DepthStencil,
+			ds = g_texture_cache->CreateTarget(ZBUF_TEX0, t_size, GetValidSize(src, possible_shuffle), target_scale, GSTextureCache::DepthStencil,
 				true, 0, false, force_preload, preserve_depth, m_r, src);
 			if (!ds) [[unlikely]]
 			{
@@ -3211,7 +3242,7 @@ void GSRendererHW::Draw()
 				CleanupDraw(true);
 				return;
 			}
-			else if (IsPageCopy() && src->m_from_target && m_cached_ctx.TEX0.TBP0 >= src->m_from_target->m_TEX0.TBP0)
+			else if (IsPageCopy() && src->m_from_target && m_cached_ctx.TEX0.TBP0 >= src->m_from_target->m_TEX0.TBP0 && m_cached_ctx.FRAME.FBW < ((src->m_from_target->m_TEX0.TBW + 1) >> 1))
 			{
 				FRAME_TEX0.TBW = src->m_from_target->m_TEX0.TBW;
 			}
@@ -3227,7 +3258,7 @@ void GSRendererHW::Draw()
 				return;
 			}
 
-			rt = g_texture_cache->CreateTarget(FRAME_TEX0, t_size, GetValidSize(src), (GSConfig.UserHacks_NativeScaling != GSNativeScaling::Off && scale_draw < 0 && is_possible_mem_clear != ClearType::NormalClear) ? src->m_from_target->GetScale() : target_scale, 
+			rt = g_texture_cache->CreateTarget(FRAME_TEX0, t_size, GetValidSize(src, possible_shuffle), (GSConfig.UserHacks_NativeScaling != GSNativeScaling::Off && scale_draw < 0 && is_possible_mem_clear != ClearType::NormalClear) ? src->m_from_target->GetScale() : target_scale, 
 												GSTextureCache::RenderTarget, true, fm, false, force_preload, preserve_rt_color || possible_shuffle, lookup_rect, src);
 
 			if (!rt) [[unlikely]]
@@ -3241,6 +3272,49 @@ void GSRendererHW::Draw()
 			{
 				rt->UpdateValidity(GSVector4i::loadh(GSVector2i(GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.x, GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.y)), true);
 			}
+
+			if (src && !src->m_from_target && GSLocalMemory::m_psm[src->m_TEX0.PSM].bpp == GSLocalMemory::m_psm[m_context->FRAME.PSM].bpp && 
+											(GSUtil::GetChannelMask(src->m_TEX0.PSM) & GSUtil::GetChannelMask(m_context->FRAME.PSM)) != 0)
+			{
+				const u32 draw_end = GSLocalMemory::GetEndBlockAddress(m_cached_ctx.FRAME.Block(), m_cached_ctx.FRAME.FBW, m_cached_ctx.FRAME.PSM, m_r) + 1;
+				const u32 draw_start = GSLocalMemory::GetStartBlockAddress(m_cached_ctx.FRAME.Block(), m_cached_ctx.FRAME.FBW, m_cached_ctx.FRAME.PSM, m_r);
+
+				if (draw_start <= src->m_TEX0.TBP0 && draw_end > src->m_TEX0.TBP0)
+				{
+					g_texture_cache->ReplaceSourceTexture(src, rt->GetTexture(), rt->GetScale(), rt->GetUnscaledSize(), nullptr, true);
+
+					src->m_from_target = rt;
+					src->m_from_target_TEX0 = rt->m_TEX0;
+					src->m_target_direct = true;
+					src->m_shared_texture = true;
+					src->m_target = true;
+					src->m_texture = rt->m_texture;
+					src->m_32_bits_fmt = rt->m_32_bits_fmt;
+					src->m_valid_rect = rt->m_valid;
+					src->m_alpha_minmax.first = rt->m_alpha_min;
+					src->m_alpha_minmax.second = rt->m_alpha_max;
+					
+					const int target_width = std::max(FRAME_TEX0.TBW, 1U);
+					const int page_offset = (src->m_TEX0.TBP0 - rt->m_TEX0.TBP0) >> 5;
+					const int vertical_page_offset = page_offset / target_width;
+					const int horizontal_page_offset = page_offset - (vertical_page_offset * target_width);
+
+					if (vertical_page_offset)
+					{
+						const int height = std::max(rt->m_valid.w, possible_shuffle ? (m_r.w / 2) : m_r.w);
+						src->m_region.SetY(vertical_page_offset * GSLocalMemory::m_psm[rt->m_TEX0.PSM].pgs.y, height);
+					}
+					if (horizontal_page_offset)
+						src->m_region.SetX(horizontal_page_offset * GSLocalMemory::m_psm[rt->m_TEX0.PSM].pgs.x, target_width * GSLocalMemory::m_psm[rt->m_TEX0.PSM].pgs.x);
+
+					if (rt->m_dirty.empty())
+					{
+						RGBAMask rgba_mask;
+						rgba_mask._u32 = GSUtil::GetChannelMask(rt->m_TEX0.PSM);
+						g_texture_cache->AddDirtyRectTarget(rt, m_r, FRAME_TEX0.PSM, FRAME_TEX0.TBW, rgba_mask, GSLocalMemory::m_psm[FRAME_TEX0.PSM].trbpp >= 16);
+					}
+				}
+			}
 		}
 		else if (rt->m_TEX0.TBP0 != m_cached_ctx.FRAME.Block())
 		{
@@ -3263,8 +3337,6 @@ void GSRendererHW::Draw()
 				const GSVector4i new_drect = GSVector4i(0, new_offset * rt->m_scale, new_size.x * rt->m_scale, new_size.y * rt->m_scale);
 				rt->ResizeTexture(new_size.x, new_size.y, true, true, new_drect);
 
-				g_texture_cache->CombineAlignedInsideTargets(rt, src);
-
 				if (src && src->m_from_target && src->m_from_target == rt && src->m_target_direct)
 				{
 					src->m_texture = rt->m_texture;
@@ -3282,6 +3354,8 @@ void GSRendererHW::Draw()
 				rt->m_drawn_since_read.y += new_offset;
 				rt->m_drawn_since_read.w += new_offset;
 
+				g_texture_cache->CombineAlignedInsideTargets(rt, src);
+
 				if (rt->m_dirty.size())
 				{
 					for (int i = 0; i < static_cast<int>(rt->m_dirty.size()); i++)
@@ -3347,79 +3421,134 @@ void GSRendererHW::Draw()
 
 				t_size.x = rt->m_unscaled_size.x - horizontal_offset;
 				t_size.y = rt->m_unscaled_size.y - vertical_offset;
+			}
 
-				// Z isn't offset but RT is, so we need a temp Z to align it, hopefully nothing will ever write to the Z too, right??
-				if (ds && vertical_offset && (m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0) != (m_cached_ctx.FRAME.Block() - rt->m_TEX0.TBP0))
+			// Don't resize if the BPP don't match.
+			GSVector2i new_size = GetValidSize(src, possible_shuffle);
+			if (new_size.x > rt->m_unscaled_size.x || new_size.y > rt->m_unscaled_size.y)
+			{
+				const u32 new_width = std::max(new_size.x, rt->m_unscaled_size.x);
+				const u32 new_height = std::max(new_size.y, rt->m_unscaled_size.y);
+
+				//DevCon.Warning("HW: Resizing texture %d x %d draw %d", rt->m_unscaled_size.x, new_height, s_n);
+				rt->ResizeTexture(new_width, new_height);
+			}
+			else if ((IsPageCopy() || is_possible_mem_clear) && m_r.width() <= frame_psm.pgs.x && m_r.height() <= frame_psm.pgs.y)
+			{
+				const int get_next_ctx = m_env.PRIM.CTXT;
+				const GSDrawingContext& next_ctx = m_env.CTXT[get_next_ctx];
+				GSVector4i update_valid = GSVector4i::loadh(GSVector2i(horizontal_offset + GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.x, GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.y + vertical_offset));
+				rt->UpdateValidity(update_valid, true);
+				if (is_possible_mem_clear)
 				{
-					const int z_vertical_offset = ((static_cast<int>(m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0) / 32) / std::max(rt->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.y;
-					if (g_texture_cache->GetTemporaryZ() != nullptr)
+					if ((horizontal_offset + GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.x) >= static_cast<int>(rt->m_TEX0.TBW * 64) && next_ctx.FRAME.Block() == (m_cached_ctx.FRAME.Block() + 0x20))
 					{
-						GSTextureCache::TempZAddress z_address_info = g_texture_cache->GetTemporaryZInfo();
-
-						if (ds->m_TEX0.TBP0 != z_address_info.ZBP || z_address_info.offset != static_cast<u32>(vertical_offset - z_vertical_offset))
-							g_texture_cache->InvalidateTemporaryZ();
-						else if (!m_r.rintersect(z_address_info.rect_since + GSVector4i(0, z_address_info.offset, 0, z_address_info.offset)).rempty() && m_cached_ctx.TEST.ZTST > ZTST_ALWAYS)
-						{
-							GL_CACHE("HW: RT in RT Updating Z copy on draw %d z_offset %d", s_n, z_address_info.offset);
-							GSVector4i dRect = GSVector4i(z_address_info.rect_since.x * ds->m_scale, (z_address_info.offset + z_address_info.rect_since.y) * ds->m_scale, (z_address_info.rect_since.z + (1.0f / ds->m_scale)) * ds->m_scale, (z_address_info.offset + z_address_info.rect_since.w + (1.0f / ds->m_scale)) * ds->m_scale);
-							g_gs_device->StretchRect(ds->m_texture, GSVector4(z_address_info.rect_since.x / static_cast<float>(ds->m_unscaled_size.x), z_address_info.rect_since.y / static_cast<float>(ds->m_unscaled_size.y), (z_address_info.rect_since.z + (1.0f / ds->m_scale)) / static_cast<float>(ds->m_unscaled_size.x), (z_address_info.rect_since.w + (1.0f / ds->m_scale)) / static_cast<float>(ds->m_unscaled_size.y)), g_texture_cache->GetTemporaryZ(), GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
-							g_perfmon.Put(GSPerfMon::TextureCopies, 1);
-							z_address_info.rect_since = GSVector4i::zero();
-							g_texture_cache->SetTemporaryZInfo(z_address_info);
-						}
+						update_valid.x = 0;
+						update_valid.z = GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.x;
+						update_valid.y += GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.y;
+						update_valid.w += GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.y;
+						rt->UpdateValidity(update_valid, true);
 					}
-
-					if (g_texture_cache->GetTemporaryZ() == nullptr)
-					{
-						m_temp_z_full_copy = false;
-						u32 vertical_size = std::max(rt->m_unscaled_size.y, ds->m_unscaled_size.y);
-						GL_CACHE("HW: RT in RT Z copy on draw %d z_vert_offset %d z_offset %d", s_n, z_vertical_offset, vertical_offset);
-						GSVector4i dRect = GSVector4i(0, vertical_offset * ds->m_scale, ds->m_unscaled_size.x * ds->m_scale, (vertical_offset + ds->m_unscaled_size.y - z_vertical_offset) * ds->m_scale);
-						const int new_height = std::max(static_cast<int>(vertical_size * ds->m_scale), dRect.w);
-						GSTexture* tex = g_gs_device->CreateDepthStencil(ds->m_unscaled_size.x * ds->m_scale, new_height, GSTexture::Format::DepthStencil, true);
-						g_gs_device->StretchRect(ds->m_texture, GSVector4(0.0f, z_vertical_offset / static_cast<float>(ds->m_unscaled_size.y), 1.0f, (ds->m_unscaled_size.y - z_vertical_offset) / static_cast<float>(ds->m_unscaled_size.y)), tex, GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
-						g_perfmon.Put(GSPerfMon::TextureCopies, 1);
-						g_texture_cache->SetTemporaryZ(tex);
-						g_texture_cache->SetTemporaryZInfo(ds->m_TEX0.TBP0, vertical_offset - z_vertical_offset);
-						t_size.y = std::max(new_height, t_size.y);
-					}
-					m_using_temp_z = true;
-
 				}
 			}
-			// Don't resize if the BPP don't match.
-			if (frame_psm.bpp == GSLocalMemory::m_psm[rt->m_TEX0.PSM].bpp)
+		}	
+		// Z or RT are offset from each other, so we need a temp Z to align it
+		if (ds && rt && (m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0) != (m_cached_ctx.FRAME.Block() - rt->m_TEX0.TBP0))
+		{
+			m_using_temp_z = true;
+			const int page_offset = (static_cast<int>(m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0) / 32);
+			const int z_vertical_offset = (page_offset / std::max(ds->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.y;
+			const int z_horizontal_offset = (page_offset % std::max(ds->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.x;
+			if (g_texture_cache->GetTemporaryZ() != nullptr)
 			{
-				if (m_r.w > rt->m_unscaled_size.y || m_r.z > rt->m_unscaled_size.x)
+				GSTextureCache::TempZAddress z_address_info = g_texture_cache->GetTemporaryZInfo();
+
+				const int old_z_vertical_offset = (page_offset / std::max(ds->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.y;
+				const int old_z_horizontal_offset = (page_offset % std::max(ds->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.x;
+
+				if (ds->m_TEX0.TBP0 != z_address_info.ZBP || z_address_info.offset != page_offset)
+					g_texture_cache->InvalidateTemporaryZ();
+				else if (!m_r.rintersect(z_address_info.rect_since).rempty() && m_cached_ctx.TEST.ZTST > ZTST_ALWAYS)
 				{
-					const u32 new_height = std::max(m_r.w, rt->m_unscaled_size.y);
-					const u32 new_width = std::max(m_r.z, rt->m_unscaled_size.x);
+					GL_CACHE("HW: RT in RT Updating Z copy on draw %d z_offset %d", s_n, z_address_info.offset);
+					GSVector4 sRect = GSVector4(z_address_info.rect_since.x / static_cast<float>(ds->m_unscaled_size.x), z_address_info.rect_since.y / static_cast<float>(ds->m_unscaled_size.y), (z_address_info.rect_since.z + (1.0f / ds->m_scale)) / static_cast<float>(ds->m_unscaled_size.x), (z_address_info.rect_since.w + (1.0f / ds->m_scale)) / static_cast<float>(ds->m_unscaled_size.y));
+					GSVector4i dRect = GSVector4i((old_z_horizontal_offset + z_address_info.rect_since.x) * ds->m_scale, (old_z_vertical_offset + z_address_info.rect_since.y) * ds->m_scale, (old_z_horizontal_offset + z_address_info.rect_since.z + (1.0f / ds->m_scale)) * ds->m_scale, (old_z_vertical_offset + z_address_info.rect_since.w + (1.0f / ds->m_scale)) * ds->m_scale);
 
-					//DevCon.Warning("HW: Resizing texture %d x %d draw %d", rt->m_unscaled_size.x, new_height, s_n);
-					rt->ResizeTexture(new_width, new_height);
+					sRect = sRect.min(GSVector4(1.0f));
+					dRect = dRect.min_u32(GSVector4i(ds->m_unscaled_size.x * ds->m_scale, ds->m_unscaled_size.y * ds->m_scale).xyxy());
 
-					const bool frame_masked = ((m_cached_ctx.FRAME.FBMSK & frame_psm.fmsk) == frame_psm.fmsk) || (m_cached_ctx.TEST.ATE && m_cached_ctx.TEST.ATST == ATST_NEVER && !(m_cached_ctx.TEST.AFAIL & AFAIL_FB_ONLY));
-
-					rt->UpdateValidity(m_r, !frame_masked);
-					rt->UpdateDrawn(m_r, !frame_masked);
+					g_gs_device->StretchRect(ds->m_texture, sRect, g_texture_cache->GetTemporaryZ(), GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
+					g_perfmon.Put(GSPerfMon::TextureCopies, 1);
+					z_address_info.rect_since = GSVector4i::zero();
+					g_texture_cache->SetTemporaryZInfo(z_address_info);
 				}
-				else if ((IsPageCopy() || is_possible_mem_clear) && m_r.width() <= frame_psm.pgs.x && m_r.height() <= frame_psm.pgs.y)
+			}
+
+			if (g_texture_cache->GetTemporaryZ() == nullptr)
+			{
+				m_temp_z_full_copy = false;
+				const int get_next_ctx = m_env.PRIM.CTXT;
+				const GSDrawingContext& next_ctx = m_env.CTXT[get_next_ctx];
+				const int rt_page_offset = (static_cast<int>(m_cached_ctx.FRAME.Block() - rt->m_TEX0.TBP0) / 32);
+				const int vertical_page_offset = (rt_page_offset / std::max(static_cast<int>(rt->m_TEX0.TBW), 1));
+				const int vertical_offset = vertical_page_offset * frame_psm.pgs.y;
+				const int horizontal_offset = (rt_page_offset - (vertical_page_offset * std::max(static_cast<int>(rt->m_TEX0.TBW), 1))) * frame_psm.pgs.x; 
+
+				const u32 horizontal_size = std::max(rt->m_unscaled_size.x, ds->m_unscaled_size.x);
+				const u32 vertical_size = std::max(rt->m_unscaled_size.y, ds->m_unscaled_size.y);
+
+				GSVector4i dRect = GSVector4i(horizontal_offset * ds->m_scale, vertical_offset * ds->m_scale, ds->m_unscaled_size.x * ds->m_scale, ds->m_unscaled_size.y * ds->m_scale);
+
+				
+				const int new_height = std::max(static_cast<int>(vertical_size * ds->m_scale), dRect.w);
+				const int new_width = std::max(static_cast<int>(horizontal_size * ds->m_scale), dRect.z);
+				const int height_diff = new_height - (ds->m_unscaled_size.y * ds->m_scale);
+				const int width_diff = new_width - (ds->m_unscaled_size.x * ds->m_scale);
+
+				if (GSTexture* tex = g_gs_device->CreateDepthStencil(new_width, new_height, GSTexture::Format::DepthStencil, true))
 				{
-					const int get_next_ctx = m_env.PRIM.CTXT;
-					const GSDrawingContext& next_ctx = m_env.CTXT[get_next_ctx];
-					GSVector4i update_valid = GSVector4i::loadh(GSVector2i(horizontal_offset + GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.x, GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.y + vertical_offset));
-					rt->UpdateValidity(update_valid, true);
-					if (is_possible_mem_clear)
+					GSVector4 sRect = GSVector4(static_cast<float>(z_horizontal_offset) / static_cast<float>(ds->m_unscaled_size.x), static_cast<float>(z_vertical_offset) / static_cast<float>(ds->m_unscaled_size.y), 1.0f - (static_cast<float>(horizontal_offset - z_horizontal_offset) / static_cast<float>(ds->m_unscaled_size.x)), 1.0f - (static_cast<float>(vertical_offset - z_vertical_offset) / static_cast<float>(ds->m_unscaled_size.y)));
+
+					const bool restricted_copy = !(((next_ctx.ZBUF.ZBP == m_context->ZBUF.ZBP && next_ctx.FRAME.FBP == m_context->FRAME.FBP)) && !(IsPossibleChannelShuffle() && !IsPageCopy()));
+
+					if (!restricted_copy)
 					{
-						if ((horizontal_offset + GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.x) >= static_cast<int>(rt->m_TEX0.TBW * 64) && next_ctx.FRAME.Block() == (m_cached_ctx.FRAME.Block() + 0x20))
+						if (height_diff)
 						{
-							update_valid.x = 0;
-							update_valid.z = GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.x;
-							update_valid.y += GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.y;
-							update_valid.w += GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].pgs.y;
-							rt->UpdateValidity(update_valid, true);
+							const int adjust = std::min(height_diff, (vertical_offset - z_vertical_offset));
+							sRect.w += static_cast<float>(adjust) / static_cast<float>(ds->m_unscaled_size.y);
+							dRect.w += adjust;
+						}
+						if (width_diff)
+						{
+							const int adjust = std::min(width_diff, (horizontal_offset - z_horizontal_offset));
+							sRect.z += static_cast<float>(adjust) / static_cast<float>(ds->m_unscaled_size.x);
+							dRect.z += adjust;
 						}
 					}
+					else
+					{
+						// m_r already has horizontal_offset (rt offset) applied)
+						dRect = GSVector4i(m_r.x * ds->m_scale, m_r.y * ds->m_scale, ((1 + m_r.z) * ds->m_scale), ((1 + m_r.w) * ds->m_scale));
+						sRect = GSVector4(static_cast<float>((m_r.x - horizontal_offset) + z_horizontal_offset) / static_cast<float>(ds->m_unscaled_size.x), static_cast<float>((m_r.y - vertical_offset) + z_vertical_offset) / static_cast<float>(ds->m_unscaled_size.y), (static_cast<float>((m_r.z - horizontal_offset) + z_horizontal_offset) + 1.0f) / static_cast<float>(ds->m_unscaled_size.x), (static_cast<float>((m_r.w - vertical_offset) + z_vertical_offset) + 1.0f) / static_cast<float>(ds->m_unscaled_size.y));
+					}
+
+					// No point in copying more width than the width of the draw, it's going to be wasted (could still be tall, though).
+					sRect.z = std::min(sRect.z, sRect.x + ((1.0f * ds->m_scale) + (static_cast<float>(m_cached_ctx.FRAME.FBW * 64)) / static_cast<float>(ds->m_unscaled_size.x)));
+					dRect.z = std::min(dRect.z, dRect.x + static_cast<int>(1 * ds->m_scale) + static_cast<int>(static_cast<float>(m_cached_ctx.FRAME.FBW * 64) * ds->m_scale));
+
+					GL_CACHE("HW: RT in RT Z copy on draw %d z_vert_offset %d", s_n, page_offset);
+
+					g_gs_device->StretchRect(ds->m_texture, sRect, tex, GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
+					g_perfmon.Put(GSPerfMon::TextureCopies, 1);
+					g_texture_cache->SetTemporaryZ(tex);
+					g_texture_cache->SetTemporaryZInfo(ds->m_TEX0.TBP0, page_offset);
+					t_size.y = std::max(static_cast<int>(new_height / ds->m_scale), t_size.y);
+				}
+				else
+				{
+					DevCon.Warning("HW: Temporary depth buffer creation failed.");
+					m_using_temp_z = false;
 				}
 			}
 		}
@@ -3477,7 +3606,7 @@ void GSRendererHW::Draw()
 		// This should never happen, but just to be safe..
 		if (!ds)
 		{
-			ds = g_texture_cache->CreateTarget(ZBUF_TEX0, t_size, GetValidSize(src), target_scale, GSTextureCache::DepthStencil,
+			ds = g_texture_cache->CreateTarget(ZBUF_TEX0, t_size, GetValidSize(src, possible_shuffle), target_scale, GSTextureCache::DepthStencil,
 				true, 0, false, force_preload, preserve_depth, m_r, src);
 			if (!ds) [[unlikely]]
 			{
@@ -3686,7 +3815,7 @@ void GSRendererHW::Draw()
 				m_vt.m_min.t *= 0.5f;
 				m_vt.m_max.t *= 0.5f;
 
-				tmm = GetTextureMinMax(MIP_TEX0, MIP_CLAMP, m_vt.IsLinear(), false);
+				tmm = GetTextureMinMax(MIP_TEX0, MIP_CLAMP, m_vt.IsLinear(), true);
 
 				src->UpdateLayer(MIP_TEX0, tmm.coverage, layer - m_lod.x);
 			}
@@ -3753,9 +3882,11 @@ void GSRendererHW::Draw()
 			}
 		}
 		const bool blending_cd = PRIM->ABE && !m_context->ALPHA.IsOpaque();
+		bool valid_width_change = false;
 		if (rt && ((!is_possible_mem_clear || blending_cd) || rt->m_TEX0.PSM != FRAME_TEX0.PSM) && !m_in_target_draw)
 		{
-			if (rt->m_TEX0.TBW != FRAME_TEX0.TBW && !m_cached_ctx.ZBUF.ZMSK && (m_cached_ctx.FRAME.FBMSK & 0xFF000000))
+			valid_width_change = rt->m_TEX0.TBW != FRAME_TEX0.TBW;
+			if (valid_width_change && !m_cached_ctx.ZBUF.ZMSK && (m_cached_ctx.FRAME.FBMSK & 0xFF000000))
 			{
 				// Alpha could be a font, and since the width is changing it's no longer valid.
 				// Be careful of downsize copies or other effects, checking Z MSK should hopefully be enough.. (Okami).
@@ -3769,6 +3900,13 @@ void GSRendererHW::Draw()
 				FRAME_TEX0.TBP0 = rt->m_TEX0.TBP0;
 				rt->m_TEX0 = FRAME_TEX0;
 			}
+
+			if (valid_width_change)
+			{
+				GSVector4i new_valid_width = rt->m_valid;
+				new_valid_width.z = std::min(new_valid_width.z, static_cast<int>(rt->m_TEX0.TBW) * 64);
+				rt->ResizeValidity(new_valid_width);
+			}
 		}
 
 		if (ds && (!is_possible_mem_clear || ds->m_TEX0.PSM != ZBUF_TEX0.PSM || (rt && ds->m_TEX0.TBW != rt->m_TEX0.TBW)) && !m_in_target_draw)
@@ -3778,7 +3916,18 @@ void GSRendererHW::Draw()
 				ZBUF_TEX0.TBP0 = ds->m_TEX0.TBP0;
 				ds->m_TEX0 = ZBUF_TEX0;
 			}
+			if (valid_width_change)
+			{
+				GSVector4i new_valid_width = ds->m_valid;
+				new_valid_width.z = std::min(new_valid_width.z, static_cast<int>(ds->m_TEX0.TBW) * 64);
+				ds->ResizeValidity(new_valid_width);
+			}
 		}
+
+		if (rt)
+			g_texture_cache->CombineAlignedInsideTargets(rt, src);
+		if (ds)
+			g_texture_cache->CombineAlignedInsideTargets(ds, src);
 	}
 	else if (!m_texture_shuffle)
 	{
@@ -3855,7 +4004,7 @@ void GSRendererHW::Draw()
 		}
 		// NFS Undercover does a draw with double width of the actual width 1280x240, which functions the same as doubling the height.
 		// Ignore single page/0 page stuff, that's just gonna get silly
-		else if (buffer_width > 64 && update_rect.z > buffer_width)
+		else if (m_texture_shuffle && buffer_width > 64 && update_rect.z > buffer_width)
 		{
 			update_rect.w *= static_cast<float>(update_rect.z) / static_cast<float>(buffer_width);
 			update_rect.z = buffer_width;
@@ -3881,7 +4030,7 @@ void GSRendererHW::Draw()
 			const bool new_rect = rt->m_valid.rempty();
 			const bool new_height = new_h > rt->GetUnscaledHeight();
 			const int old_height = rt->m_texture->GetHeight();
-
+			bool merge_targets = false;
 			pxAssert(rt->GetScale() == target_scale);
 			if (rt->GetUnscaledWidth() != new_w || rt->GetUnscaledHeight() != new_h)
 				GL_INS("HW: Resize RT from %dx%d to %dx%d", rt->GetUnscaledWidth(), rt->GetUnscaledHeight(), new_w, new_h);
@@ -3904,7 +4053,10 @@ void GSRendererHW::Draw()
 					g_texture_cache->AddDirtyRectTarget(rt, height_dirty_rect, rt->m_TEX0.PSM, rt->m_TEX0.TBW, mask);
 				}
 			}*/
-			
+
+			if ((new_w > rt->m_unscaled_size.x || new_h > rt->m_unscaled_size.y) && GSConfig.UserHacks_TextureInsideRt >= GSTextureInRtMode::InsideTargets)
+				merge_targets = true;
+
 			rt->ResizeTexture(new_w, new_h);
 
 			if (!m_texture_shuffle && !m_channel_shuffle)
@@ -3935,6 +4087,8 @@ void GSRendererHW::Draw()
 			rt->UpdateValidity(update_rect, !frame_masked && (rt_update || (m_r.w <= (resolution.y * 2) && !m_texture_shuffle)));
 			rt->UpdateDrawn(update_rect, !frame_masked && (rt_update || (m_r.w <= (resolution.y * 2) && !m_texture_shuffle)));
 
+			if (merge_targets)
+				g_texture_cache->CombineAlignedInsideTargets(rt, src);
 			// Probably changing to double buffering, so invalidate any old target that was next to it.
 			// This resolves an issue where the PCRTC will find the old target in FMV's causing flashing.
 			// Grandia Xtreme, Onimusha Warlord.
@@ -3979,13 +4133,16 @@ void GSRendererHW::Draw()
 
 				if (z_width != new_w || z_height != new_h)
 				{
-					GSVector4i dRect = GSVector4i(0, 0, g_texture_cache->GetTemporaryZ()->GetWidth(), g_texture_cache->GetTemporaryZ()->GetHeight());
-
-					GSTexture* tex = g_gs_device->CreateDepthStencil(new_w * ds->m_scale, new_h * ds->m_scale, GSTexture::Format::DepthStencil, true);
-					g_gs_device->StretchRect(g_texture_cache->GetTemporaryZ(), GSVector4(0.0f, 0.0f, 1.0f, 1.0f), tex, GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
-					g_perfmon.Put(GSPerfMon::TextureCopies, 1);
-					g_texture_cache->InvalidateTemporaryZ();
-					g_texture_cache->SetTemporaryZ(tex);
+					if (GSTexture* tex = g_gs_device->CreateDepthStencil(new_w * ds->m_scale, new_h * ds->m_scale, GSTexture::Format::DepthStencil, true))
+					{
+						const GSVector4i dRect = GSVector4i(0, 0, g_texture_cache->GetTemporaryZ()->GetWidth(), g_texture_cache->GetTemporaryZ()->GetHeight());
+						g_gs_device->StretchRect(g_texture_cache->GetTemporaryZ(), GSVector4(0.0f, 0.0f, 1.0f, 1.0f), tex, GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
+						g_perfmon.Put(GSPerfMon::TextureCopies, 1);
+						g_texture_cache->InvalidateTemporaryZ();
+						g_texture_cache->SetTemporaryZ(tex);
+					}
+					else
+						DevCon.Warning("HW: Temporary depth buffer creation failed.");
 				}
 			}
 			if (!m_texture_shuffle && !m_channel_shuffle)
@@ -3998,9 +4155,21 @@ void GSRendererHW::Draw()
 			// Dark cloud writes to 424 when the buffer is only 416 high, but masks the Z.
 			// Updating the valid causes the Z to overlap the framebuffer, which is obviously incorrect.
 			const bool z_masked = m_cached_ctx.ZBUF.ZMSK;
+			const bool z_update = can_update_size && !z_masked;
 
-			ds->UpdateValidity(m_r, !z_masked && (can_update_size || m_r.w <= (resolution.y * 2)));
-			ds->UpdateDrawn(m_r, !z_masked && (can_update_size || m_r.w <= (resolution.y * 2)));
+			if (rt && m_using_temp_z)
+			{
+				const GSLocalMemory::psm_t& z_psm = GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM];
+				const int vertical_offset = ((static_cast<int>(m_cached_ctx.FRAME.Block() - rt->m_TEX0.TBP0) / 32) / std::max(static_cast<int>(rt->m_TEX0.TBW), 1)) * frame_psm.pgs.y;
+				const int z_vertical_offset = ((static_cast<int>(m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0) / 32) / std::max(static_cast<int>(ds->m_TEX0.TBW), 1)) * z_psm.pgs.y;
+				const GSVector4i ds_rect = m_r - GSVector4i(vertical_offset - z_vertical_offset);
+				ds->UpdateValidity(ds_rect, z_update && (can_update_size || (ds_rect.w <= (resolution.y * 2) && !m_texture_shuffle)));
+			}
+			else
+			{
+				ds->UpdateValidity(m_r, z_update && (can_update_size || m_r.w <= (resolution.y * 2)));
+				ds->UpdateDrawn(m_r, z_update && (can_update_size || m_r.w <= (resolution.y * 2)));
+			}
 
 			if (!new_rect && new_height && old_end_block != ds->m_end_block)
 			{
@@ -4225,54 +4394,63 @@ void GSRendererHW::Draw()
 
 	}
 
-	if (zm != 0xffffffff && ds)
+	if (ds)
 	{
 		const bool z_masked = m_cached_ctx.ZBUF.ZMSK;
+		const bool was_written = zm != 0xffffffff && m_cached_ctx.DepthWrite();
 
 		//ds->m_valid = ds->m_valid.runion(r);
 		// Limit to 2x the vertical height of the resolution (for double buffering)
-		ds->UpdateValidity(real_rect, !z_masked && (can_update_size || (real_rect.w <= (resolution.y * 2) && !m_texture_shuffle)));
 
 		if (m_using_temp_z)
 		{
-			if (m_cached_ctx.DepthWrite())
-			{
-				const int get_next_ctx = m_env.PRIM.CTXT;
-				const GSDrawingContext& next_ctx = m_env.CTXT[get_next_ctx];
-				if ((m_state_flush_reason != CONTEXTCHANGE) || (next_ctx.ZBUF.ZBP == m_context->ZBUF.ZBP && next_ctx.FRAME.FBP == m_context->FRAME.FBP))
-				{
-					m_temp_z_full_copy = true;
-				}
-				else
-				{
-					const int vertical_offset = ((static_cast<int>(m_cached_ctx.FRAME.Block() - rt->m_TEX0.TBP0) / 32) / std::max(static_cast<int>(rt->m_TEX0.TBW), 1)) * frame_psm.pgs.y;
-					const int z_vertical_offset = ((static_cast<int>(m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0) / 32) / std::max(rt->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.y;
-					
-					if (!m_temp_z_full_copy)
-					{
-						const GSVector4i dRect = GSVector4i(real_rect.x * ds->m_scale, (z_vertical_offset + (real_rect.y - vertical_offset)) * ds->m_scale, (real_rect.z + (1.0f / ds->m_scale)) * ds->m_scale, (z_vertical_offset + (real_rect.w + (1.0f / ds->m_scale) - vertical_offset)) * ds->m_scale);
-						const GSVector4 sRect = GSVector4((real_rect.x * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetWidth()), static_cast<float>(real_rect.y * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetHeight()), ((real_rect.z + (1.0f / ds->m_scale)) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetWidth()),
-							static_cast<float>((real_rect.w + (1.0f / ds->m_scale)) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetHeight()));
-						GL_CACHE("HW: RT in RT Z copy back draw %d z_vert_offset %d z_offset %d", s_n, z_vertical_offset, vertical_offset);
-						g_gs_device->StretchRect(g_texture_cache->GetTemporaryZ(), sRect, ds->m_texture, GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
-						g_perfmon.Put(GSPerfMon::TextureCopies, 1);
-					}
-					else
-					{
-						const GSVector4i dRect = GSVector4i(0, ds->m_valid.y * ds->m_scale, ds->m_valid.z * ds->m_scale, ds->m_valid.w * ds->m_scale);
-						const GSVector4 sRect = GSVector4((ds->m_valid.x * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetWidth()), static_cast<float>((ds->m_valid.y + vertical_offset) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetHeight()), ((ds->m_valid.z + (1.0f / ds->m_scale)) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetWidth()),
-							static_cast<float>(((ds->m_valid.w + vertical_offset) + (1.0f / ds->m_scale)) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetHeight()));
-						GL_CACHE("HW: RT in RT Z copy back draw %d z_vert_offset %d z_offset %d", s_n, z_vertical_offset, vertical_offset);
-						g_gs_device->StretchRect(g_texture_cache->GetTemporaryZ(), sRect, ds->m_texture, GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
-						g_perfmon.Put(GSPerfMon::TextureCopies, 1);
-					}
+			const int get_next_ctx = m_env.PRIM.CTXT;
+			const GSDrawingContext& next_ctx = m_env.CTXT[get_next_ctx];
+			const int z_vertical_offset = ((static_cast<int>(m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0) / 32) / std::max(rt->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.y;
+			const int z_horizontal_offset = ((static_cast<int>(m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0) / 32) % std::max(rt->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.x;
+			const int vertical_offset = ((static_cast<int>(m_cached_ctx.FRAME.Block() - rt->m_TEX0.TBP0) / 32) / std::max(static_cast<int>(rt->m_TEX0.TBW), 1)) * frame_psm.pgs.y;
+			const int horizontal_offset = ((static_cast<int>(m_cached_ctx.FRAME.Block() - rt->m_TEX0.TBP0) / 32) % std::max(static_cast<int>(rt->m_TEX0.TBW), 1)) * frame_psm.pgs.x;
 
-					m_temp_z_full_copy = false;
+			if (was_written)
+			{
+				const GSVector4i ds_real_rect = real_rect - GSVector4i(vertical_offset - z_vertical_offset);
+				ds->UpdateValidity(ds_real_rect, !z_masked && (can_update_size || (ds_real_rect.w <= (resolution.y * 2) && !m_texture_shuffle)));
+			}
+
+			if (((m_state_flush_reason != CONTEXTCHANGE) || (next_ctx.ZBUF.ZBP == m_context->ZBUF.ZBP && next_ctx.FRAME.FBP == m_context->FRAME.FBP)) && !(IsPossibleChannelShuffle() && !IsPageCopy()))
+			{
+				m_temp_z_full_copy |= was_written;
+			}
+			else
+			{
+				if (!m_temp_z_full_copy && was_written)
+				{
+					GSVector4i dRect = GSVector4i((z_horizontal_offset + (real_rect.x - horizontal_offset)) * ds->m_scale, (z_vertical_offset + (real_rect.y - vertical_offset)) * ds->m_scale, ((z_horizontal_offset + real_rect.z + (1.0f / ds->m_scale)) - horizontal_offset) * ds->m_scale, (z_vertical_offset + (real_rect.w + (1.0f / ds->m_scale) - vertical_offset)) * ds->m_scale);
+					GSVector4 sRect = GSVector4((real_rect.x * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetWidth()), static_cast<float>(real_rect.y * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetHeight()), ((real_rect.z + (1.0f / ds->m_scale)) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetWidth()),
+												static_cast<float>((real_rect.w + (1.0f / ds->m_scale)) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetHeight()));
+
+					GL_CACHE("HW: RT in RT Z copy back draw %d z_vert_offset %d rt_vert_offset %d z_horz_offset %d rt_horz_offset %d", s_n, z_vertical_offset, vertical_offset, z_horizontal_offset, horizontal_offset);
+					g_gs_device->StretchRect(g_texture_cache->GetTemporaryZ(), sRect, ds->m_texture, GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
+					g_perfmon.Put(GSPerfMon::TextureCopies, 1);
 				}
+				else if (m_temp_z_full_copy)
+				{
+					GSVector4i dRect = GSVector4i((ds->m_valid.x + z_horizontal_offset) * ds->m_scale, (ds->m_valid.y + z_vertical_offset) * ds->m_scale, (ds->m_valid.z + z_horizontal_offset + (1.0f / ds->m_scale)) * ds->m_scale, (ds->m_valid.w + z_vertical_offset + (1.0f / ds->m_scale)) * ds->m_scale);
+					GSVector4 sRect = GSVector4(((ds->m_valid.x + horizontal_offset) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetWidth()), static_cast<float>((ds->m_valid.y + vertical_offset) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetHeight()), (((ds->m_valid.z + horizontal_offset) + (1.0f / ds->m_scale)) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetWidth()),
+						static_cast<float>((ds->m_valid.w + vertical_offset + (1.0f / ds->m_scale)) * ds->m_scale) / static_cast<float>(g_texture_cache->GetTemporaryZ()->GetHeight()));
+
+					GL_CACHE("HW: RT in RT Z copy back draw %d z_vert_offset %d z_offset %d", s_n, z_vertical_offset, vertical_offset);
+					g_gs_device->StretchRect(g_texture_cache->GetTemporaryZ(), sRect, ds->m_texture, GSVector4(dRect), ShaderConvert::DEPTH_COPY, false);
+					g_perfmon.Put(GSPerfMon::TextureCopies, 1);
+				}
+
+				m_temp_z_full_copy = false;
 			}
 		}
-		else if (m_cached_ctx.DepthWrite() && g_texture_cache->GetTemporaryZ() != nullptr)
+		else if (was_written && g_texture_cache->GetTemporaryZ() != nullptr)
 		{
+			ds->UpdateValidity(real_rect, !z_masked && (can_update_size || (real_rect.w <= (resolution.y * 2) && !m_texture_shuffle)));
+
 			GSTextureCache::TempZAddress z_address_info = g_texture_cache->GetTemporaryZInfo();
 			if (ds->m_TEX0.TBP0 == z_address_info.ZBP)
 			{
@@ -5015,6 +5193,8 @@ __ri bool GSRendererHW::EmulateChannelShuffle(GSTextureCache::Target* src, bool
 		// Hitman suffers from this, not sure on the exact scenario at the moment, but we need the barrier.
 		if (PRIM->ABE && m_context->ALPHA.IsCdInBlend())
 		{
+			// Needed to enable IsFeedbackLoop.
+			m_conf.ps.channel_fb = 1;
 			// Assume no overlap when it's a channel shuffle, no need for full barriers.
 			m_conf.require_one_barrier = true;
 		}
@@ -5873,10 +6053,9 @@ void GSRendererHW::EmulateBlending(int rt_alpha_min, int rt_alpha_max, const boo
 	// Switch DATE_PRIMID with DATE_BARRIER in such cases to ensure accuracy.
 	// No mix of COLCLIP + sw blend + DATE_PRIMID, neither sw fbmask + DATE_PRIMID.
 	// Note: Do the swap in the end, saves the expensive draw splitting/barriers when mixed software blending is used.
-	if (sw_blending && DATE_PRIMID && m_conf.require_full_barrier)
+	if (sw_blending && DATE_PRIMID && features.texture_barrier && m_conf.require_full_barrier)
 	{
 		GL_PERF("DATE: Swap DATE_PRIMID with DATE_BARRIER");
-		m_conf.require_full_barrier = true;
 		DATE_PRIMID = false;
 		DATE_BARRIER = true;
 	}
@@ -5923,9 +6102,6 @@ __ri void GSRendererHW::EmulateTextureSampler(const GSTextureCache::Target* rt,
 
 	m_conf.cb_ps.ScaleFactor = GSVector4(scale_factor * (1.0f / 16.0f), 1.0f / scale_factor, scale_rt, 0.0f);
 
-	if ((m_conf.ps.tex_is_fb && rt->m_rt_alpha_scale) || (tex->m_target && tex->m_from_target && tex->m_target_direct && tex->m_from_target->m_rt_alpha_scale))
-		m_conf.ps.rta_source_correction = 1;
-
 	// Warning fetch the texture PSM format rather than the context format. The latter could have been corrected in the texture cache for depth.
 	//const GSLocalMemory::psm_t &psm = GSLocalMemory::m_psm[m_cached_ctx.TEX0.PSM];
 	const GSLocalMemory::psm_t& psm = GSLocalMemory::m_psm[tex->m_TEX0.PSM];
@@ -6296,56 +6472,58 @@ __ri void GSRendererHW::HandleTextureHazards(const GSTextureCache::Target* rt, c
 
 	// Detect framebuffer read that will need special handling
 	const GSTextureCache::Target* src_target = nullptr;
-
-	if (rt && m_conf.tex == m_conf.rt && !(m_channel_shuffle && tex && (tex_diff != frame_diff || target_region)))
+	if (!m_downscale_source || !tex->m_from_target)
 	{
-		// Can we read the framebuffer directly? (i.e. sample location matches up).
-		if (CanUseTexIsFB(rt, tex, tmm))
+		if (rt && m_conf.tex == m_conf.rt && !(m_channel_shuffle && tex && (tex_diff != frame_diff || target_region)))
 		{
-			m_conf.tex = nullptr;
-			m_conf.ps.tex_is_fb = true;
-			if (m_prim_overlap == PRIM_OVERLAP_NO || !g_gs_device->Features().texture_barrier)
-				m_conf.require_one_barrier = true;
-			else
-				m_conf.require_full_barrier = true;
+			// Can we read the framebuffer directly? (i.e. sample location matches up).
+			if (CanUseTexIsFB(rt, tex, tmm))
+			{
+				m_conf.tex = nullptr;
+				m_conf.ps.tex_is_fb = true;
+				if (m_prim_overlap == PRIM_OVERLAP_NO || !g_gs_device->Features().texture_barrier)
+					m_conf.require_one_barrier = true;
+				else
+					m_conf.require_full_barrier = true;
 
-			unscaled_size = rt->GetUnscaledSize();
-			scale = rt->GetScale();
+				unscaled_size = rt->GetUnscaledSize();
+				scale = rt->GetScale();
+				return;
+			}
+
+			GL_CACHE("HW: Source is render target, taking copy.");
+			src_target = rt;
+		}
+		// Be careful of single page channel shuffles where depth is the source but it's not going to the same place, we can't read this directly.
+		else if (ds && m_conf.tex == m_conf.ds && (!m_channel_shuffle || (rt && static_cast<int>(m_cached_ctx.FRAME.Block() - rt->m_TEX0.TBP0) == static_cast<int>(m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0))))
+		{
+			// GL, Vulkan (in General layout), not DirectX!
+			const bool can_read_current_depth_buffer = g_gs_device->Features().test_and_sample_depth;
+
+			// If this is our current Z buffer, we might not be able to read it directly if it's being written to.
+			// Rather than leaving the backend to do it, we'll check it here.
+			if (can_read_current_depth_buffer && (m_cached_ctx.ZBUF.ZMSK || m_cached_ctx.TEST.ZTST == ZTST_NEVER))
+			{
+				// Safe to read!
+				GL_CACHE("HW: Source is depth buffer, not writing, safe to read.");
+				unscaled_size = ds->GetUnscaledSize();
+				scale = ds->GetScale();
+				return;
+			}
+
+			// Can't safely read the depth buffer, so we need to take a copy of it.
+			GL_CACHE("HW: Source is depth buffer, unsafe to read, taking copy.");
+			src_target = ds;
+		}
+		else if (m_channel_shuffle && tex->m_from_target && tex_diff != frame_diff)
+		{
+			src_target = tex->m_from_target;
+		}
+		else
+		{
+			// No match.
 			return;
 		}
-
-		GL_CACHE("HW: Source is render target, taking copy.");
-		src_target = rt;
-	}
-	// Be careful of single page channel shuffles where depth is the source but it's not going to the same place, we can't read this directly.
-	else if (ds && m_conf.tex == m_conf.ds && (!m_channel_shuffle || static_cast<int>(m_cached_ctx.FRAME.Block() - rt->m_TEX0.TBP0) ==  static_cast<int>(m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0)))
-	{
-		// GL, Vulkan (in General layout), not DirectX!
-		const bool can_read_current_depth_buffer = g_gs_device->Features().test_and_sample_depth;
-
-		// If this is our current Z buffer, we might not be able to read it directly if it's being written to.
-		// Rather than leaving the backend to do it, we'll check it here.
-		if (can_read_current_depth_buffer && (m_cached_ctx.ZBUF.ZMSK || m_cached_ctx.TEST.ZTST == ZTST_NEVER))
-		{
-			// Safe to read!
-			GL_CACHE("HW: Source is depth buffer, not writing, safe to read.");
-			unscaled_size = ds->GetUnscaledSize();
-			scale = ds->GetScale();
-			return;
-		}
-
-		// Can't safely read the depth buffer, so we need to take a copy of it.
-		GL_CACHE("HW: Source is depth buffer, unsafe to read, taking copy.");
-		src_target = ds;
-	}
-	else if (m_channel_shuffle && tex->m_from_target && tex_diff != frame_diff)
-	{
-		src_target = tex->m_from_target;
-	}
-	else if (!m_downscale_source || !tex->m_from_target)
-	{
-		// No match.
-		return;
 	}
 	else
 		src_target = tex->m_from_target;
@@ -6353,10 +6531,9 @@ __ri void GSRendererHW::HandleTextureHazards(const GSTextureCache::Target* rt, c
 	// We need to copy. Try to cut down the source range as much as possible so we don't copy texels we're not reading.
 	const GSVector2i& src_unscaled_size = src_target->GetUnscaledSize();
 	const GSVector4i src_bounds = src_target->GetUnscaledRect();
-	GSVector4i copy_range;
-	GSVector2i copy_size;
-	GSVector2i copy_dst_offset;
-	const bool copied_rt = src_target && !tex->m_shared_texture;
+	GSVector4i copy_range = GSVector4i::zero();
+	GSVector2i copy_size = GSVector2i(0);
+	GSVector2i copy_dst_offset = GSVector2i(0);
 	// Shuffles take the whole target. This should've already been halved.
 	// We can't partially copy depth targets in DirectX, and GL/Vulkan should use the direct read above.
 	// Restricting it also breaks Tom and Jerry...
@@ -6480,21 +6657,15 @@ __ri void GSRendererHW::HandleTextureHazards(const GSTextureCache::Target* rt, c
 
 	unscaled_size = copy_size;
 	scale = m_downscale_source ? 1.0f : src_target->GetScale();
-	const float src_scale = src_target->GetScale();
 	GL_CACHE("HW: Copy size: %dx%d, range: %d,%d -> %d,%d (%dx%d) @ %.1f", copy_size.x, copy_size.y, copy_range.x,
 		copy_range.y, copy_range.z, copy_range.w, copy_range.width(), copy_range.height(), scale);
 
 	const GSVector2i scaled_copy_size = GSVector2i(static_cast<int>(std::ceil(static_cast<float>(copy_size.x) * scale)),
 		static_cast<int>(std::ceil(static_cast<float>(copy_size.y) * scale)));
-	const GSVector4i scaled_copy_range = GSVector4i((GSVector4(copy_range) * GSVector4(src_scale)).ceil());
-	const GSVector2i scaled_copy_dst_offset =
-		GSVector2i(static_cast<int>(std::ceil(static_cast<float>(copy_dst_offset.x) * scale)),
-			static_cast<int>(std::ceil(static_cast<float>(copy_dst_offset.y) * scale)));
 
 	src_copy.reset(src_target->m_texture->IsDepthStencil() ?
 				   g_gs_device->CreateDepthStencil(scaled_copy_size.x, scaled_copy_size.y, src_target->m_texture->GetFormat(), false) :
-					   (m_downscale_source || copied_rt) ? g_gs_device->CreateRenderTarget(scaled_copy_size.x, scaled_copy_size.y, src_target->m_texture->GetFormat(), true, true) : 
-															g_gs_device->CreateTexture(scaled_copy_size.x, scaled_copy_size.y, 1, src_target->m_texture->GetFormat(), true));
+					   g_gs_device->CreateRenderTarget(scaled_copy_size.x, scaled_copy_size.y, src_target->m_texture->GetFormat(), true, true));
 	if (!src_copy) [[unlikely]]
 	{
 		Console.Error("HW: Failed to allocate %dx%d texture for hazard copy", scaled_copy_size.x, scaled_copy_size.y);
@@ -6534,8 +6705,21 @@ __ri void GSRendererHW::HandleTextureHazards(const GSTextureCache::Target* rt, c
 	}
 	else
 	{
-		g_gs_device->CopyRect(
-			src_target->m_texture, src_copy.get(), scaled_copy_range, scaled_copy_dst_offset.x, scaled_copy_dst_offset.y);
+		g_perfmon.Put(GSPerfMon::TextureCopies, 1);
+		const GSVector4i offset = copy_range - GSVector4i(copy_dst_offset).xyxy();
+
+		// Adjust for bilinear, must be done after calculating offset.
+		copy_range.x -= 1;
+		copy_range.y -= 1;
+		copy_range.z += 1;
+		copy_range.w += 1;
+		copy_range = copy_range.rintersect(src_bounds);
+
+		const GSVector4 src_rect = GSVector4(copy_range) / GSVector4(src_unscaled_size).xyxy();
+		const GSVector4 dst_rect = (GSVector4(copy_range) - GSVector4(offset).xyxy()) * scale;
+
+		g_gs_device->StretchRect(src_target->m_texture, src_rect, src_copy.get(), dst_rect,
+			src_target->m_texture->IsDepthStencil() ? ShaderConvert::DEPTH_COPY : ShaderConvert::COPY, false);
 	}
 	m_conf.tex = src_copy.get();
 }
@@ -6630,7 +6814,7 @@ bool GSRendererHW::CanUseTexIsFB(const GSTextureCache::Target* rt, const GSTextu
 		const GSLocalMemory::psm_t& rt_psm = GSLocalMemory::m_psm[rt->m_TEX0.PSM];
 		if (tex_psm.pal > 0 && tex_psm.bpp < rt_psm.bpp)
 		{
-			GL_CACHE("HW: Disabling tex-is-fb due to palette conversion.");
+			GL_CACHE("HW: Enabling tex-is-fb for palette conversion.");
 			return true;
 		}
 
@@ -7115,6 +7299,19 @@ __ri void GSRendererHW::DrawPrims(GSTextureCache::Target* rt, GSTextureCache::Ta
 		new_scale_rt_alpha = rt->m_rt_alpha_scale;
 	}
 
+	GSDevice::RecycledTexture tex_copy;
+	if (tex)
+	{
+		EmulateTextureSampler(rt, ds, tex, tmm, tex_copy);
+	}
+	else
+	{
+		const float scale_factor = rt ? rt->GetScale() : ds->GetScale();
+		m_conf.cb_ps.ScaleFactor = GSVector4(scale_factor * (1.0f / 16.0f), 1.0f / scale_factor, scale_factor, 0.0f);
+
+		m_conf.ps.tfx = 4;
+	}
+
 	// AA1: Set alpha source to coverage 128 when there is no alpha blending.
 	m_conf.ps.fixed_one_a = IsCoverageAlpha();
 
@@ -7138,6 +7335,9 @@ __ri void GSRendererHW::DrawPrims(GSTextureCache::Target* rt, GSTextureCache::Ta
 		}
 	}
 
+	if ((m_conf.ps.tex_is_fb && rt && rt->m_rt_alpha_scale) || (tex && tex->m_from_target && tex->m_target_direct && tex->m_from_target->m_rt_alpha_scale))
+		m_conf.ps.rta_source_correction = 1;
+
 	if (req_src_update && tex->m_texture != rt->m_texture)
 		tex->m_texture = rt->m_texture;
 
@@ -7270,8 +7470,25 @@ __ri void GSRendererHW::DrawPrims(GSTextureCache::Target* rt, GSTextureCache::Ta
 	if (m_conf.destination_alpha >= GSHWDrawConfig::DestinationAlphaMode::Stencil &&
 		m_conf.destination_alpha <= GSHWDrawConfig::DestinationAlphaMode::StencilOne && !m_conf.ds)
 	{
-		temp_ds.reset(g_gs_device->CreateDepthStencil(m_conf.rt->GetWidth(), m_conf.rt->GetHeight(), GSTexture::Format::DepthStencil, false));
-		m_conf.ds = temp_ds.get();
+		const bool is_one_barrier = (features.texture_barrier && m_conf.require_full_barrier && (m_prim_overlap == PRIM_OVERLAP_NO || m_conf.ps.shuffle || m_channel_shuffle));
+		if ((temp_ds.reset(g_gs_device->CreateDepthStencil(m_conf.rt->GetWidth(), m_conf.rt->GetHeight(), GSTexture::Format::DepthStencil, false)), temp_ds))
+		{
+			m_conf.ds = temp_ds.get();
+		}
+		else if (features.primitive_id && !(m_conf.ps.scanmsk & 2) && (!m_conf.require_full_barrier || is_one_barrier))
+		{
+			DATE_one = false;
+			DATE_PRIMID = true;
+			m_conf.destination_alpha = GSHWDrawConfig::DestinationAlphaMode::PrimIDTracking;
+			DevCon.Warning("HW: Depth buffer creation failed for Stencil Date. Fallback to PrimIDTracking.");
+		}
+		else
+		{
+			DATE = false;
+			DATE_one = false;
+			m_conf.destination_alpha = GSHWDrawConfig::DestinationAlphaMode::Off;
+			DevCon.Warning("HW: Depth buffer creation failed for Stencil Date.");
+		}
 	}
 
 	// vs
@@ -7404,18 +7621,7 @@ __ri void GSRendererHW::DrawPrims(GSTextureCache::Target* rt, GSTextureCache::Ta
 		m_conf.cb_ps.FogColor_AREF = fc.blend32<8>(m_conf.cb_ps.FogColor_AREF);
 	}
 
-	GSDevice::RecycledTexture tex_copy;
-	if (tex)
-	{
-		EmulateTextureSampler(rt, ds, tex, tmm, tex_copy);
-	}
-	else
-	{
-		const float scale_factor = rt ? rt->GetScale() : ds->GetScale();
-		m_conf.cb_ps.ScaleFactor = GSVector4(scale_factor * (1.0f / 16.0f), 1.0f / scale_factor, scale_factor, 0.0f);
 
-		m_conf.ps.tfx = 4;
-	}
 
 	// Update RT scaled alpha flag, nothing's going to read it anymore.
 	if (rt)
@@ -7440,11 +7646,17 @@ __ri void GSRendererHW::DrawPrims(GSTextureCache::Target* rt, GSTextureCache::Ta
 	pxAssert(!m_conf.require_full_barrier || !m_conf.ps.colclip_hw);
 
 	// Swap full barrier for one barrier when there's no overlap, or a shuffle.
-	if (m_conf.require_full_barrier && (m_prim_overlap == PRIM_OVERLAP_NO || m_conf.ps.shuffle || m_channel_shuffle))
+	if (features.texture_barrier && m_conf.require_full_barrier && (m_prim_overlap == PRIM_OVERLAP_NO || m_conf.ps.shuffle || m_channel_shuffle))
 	{
 		m_conf.require_full_barrier = false;
 		m_conf.require_one_barrier = true;
 	}
+	else if (!features.texture_barrier)
+	{
+		// These shouldn't be enabled if texture barriers aren't supported, make sure they are off.
+		m_conf.ps.write_rg = 0;
+		m_conf.require_full_barrier = false;
+	}
 
 	// rs
 	const GSVector4i hacked_scissor = m_channel_shuffle ? GSVector4i::cxpr(0, 0, 1024, 1024) : m_context->scissor.in;
@@ -8909,10 +9121,7 @@ void GSRendererHW::EndHLEHardwareDraw(bool force_copy_on_hazard /* = false */)
 				return;
 			}
 
-			// DX11 can't partial copy depth textures.
-			const GSVector4i copy_rect = (src->IsDepthStencil() && !features.test_and_sample_depth) ?
-											 src->GetRect() :
-											 config.drawarea.rintersect(src->GetRect());
+			const GSVector4i copy_rect =  config.drawarea.rintersect(src->GetRect());
 			g_gs_device->CopyRect(src, copy, copy_rect - copy_rect.xyxy(), copy_rect.x, copy_rect.y);
 			config.tex = copy;
 		}

pcsx2/GS/Renderers/HW/GSRendererHW.h

@@ -222,8 +222,8 @@ public:
 	GSVector4i ComputeBoundingBox(const GSVector2i& rtsize, float rtscale);
 	void MergeSprite(GSTextureCache::Source* tex);
 	float GetTextureScaleFactor() override;
-	GSVector2i GetValidSize(const GSTextureCache::Source* tex = nullptr);
-	GSVector2i GetTargetSize(const GSTextureCache::Source* tex = nullptr, const bool can_expand = true);
+	GSVector2i GetValidSize(const GSTextureCache::Source* tex = nullptr, const bool is_shuffle = false);
+	GSVector2i GetTargetSize(const GSTextureCache::Source* tex = nullptr, const bool can_expand = true, const bool is_shuffle = false);
 
 	void Reset(bool hardware_reset) override;
 	void UpdateSettings(const Pcsx2Config::GSOptions& old_config) override;

pcsx2/GS/Renderers/HW/GSTextureCache.cpp

@@ -240,7 +240,7 @@ bool GSTextureCache::CanTranslate(u32 bp, u32 bw, u32 spsm, GSVector4i r, u32 db
 	const bool page_aligned_rect = masked_rect.xyxy().eq(r.xyxy());
 	const bool width_match = ((bw * 64) / src_page_size.x) == ((dbw * 64) / dst_page_size.x);
 	const bool sequential_pages = page_aligned_rect && r.x == 0 && r.z == src_pixel_width;
-	const bool single_row = (((bw * 64) / src_page_size.x) <= ((dbw * 64) / dst_page_size.x)) && r.z <= src_pixel_width && r.w <= src_page_size.y;
+	const bool single_row = (((bw * 64) / src_page_size.x) <= ((dbw * 64) / dst_page_size.x)) && r.width() <= src_pixel_width && r.height() <= src_page_size.y;
 	const bool single_page_aligned = page_aligned_rect && r.z <= src_page_size.x && r.w <= src_page_size.y;
 	if (block_layout_match)
 	{
@@ -998,7 +998,7 @@ __ri static GSTextureCache::Source* FindSourceInMap(const GIFRegTEX0& TEX0, cons
 	return nullptr;
 }
 
-GSTextureCache::Source* GSTextureCache::LookupDepthSource(const bool is_depth, const GIFRegTEX0& TEX0, const GIFRegTEXA& TEXA, const GIFRegCLAMP& CLAMP, const GSVector4i& r, const bool possible_shuffle, const bool linear, const u32 frame_fbp, bool req_color, bool req_alpha, bool palette)
+GSTextureCache::Source* GSTextureCache::LookupDepthSource(const bool is_depth, const GIFRegTEX0& TEX0, const GIFRegTEXA& TEXA, const GIFRegCLAMP& CLAMP, const GSVector4i& r, const bool possible_shuffle, const bool linear, const GIFRegFRAME& frame, bool req_color, bool req_alpha, bool palette)
 {
 	if (GSConfig.UserHacks_DisableDepthSupport)
 	{
@@ -1107,17 +1107,15 @@ GSTextureCache::Source* GSTextureCache::LookupDepthSource(const bool is_depth, c
 					if (dst->m_unscaled_size != t->m_unscaled_size)
 					{
 						t->ResizeTexture(t->m_unscaled_size.x, t->m_unscaled_size.y);
-						t->m_valid = dst->m_valid;
 					}
 
 					CopyRGBFromDepthToColor(t, dst);
 				}
 
+				t->m_valid = t->m_valid.runion(dst->m_valid);
 				dst = t;
 
-				if (GSUtil::GetChannelMask(TEX0.PSM) & 0x8)
-					t->UnscaleRTAlpha();
-
+				// Don't need to de-RTA here as we were actually copying the RGB over, preserving the existing alpha.
 				inside_target = false;
 				break;
 			}
@@ -1183,7 +1181,7 @@ GSTextureCache::Source* GSTextureCache::LookupDepthSource(const bool is_depth, c
 	else
 	{
 		// This is a bit of a worry, since it could load junk from local memory... but it's better than skipping the draw.
-		return is_depth ? LookupSource(false, TEX0, TEXA, CLAMP, r, nullptr, possible_shuffle, linear, frame_fbp, req_color, req_alpha) : nullptr;
+		return is_depth ? LookupSource(false, TEX0, TEXA, CLAMP, r, nullptr, possible_shuffle, linear, frame, req_color, req_alpha) : nullptr;
 	}
 
 	pxAssert(src->m_texture);
@@ -1192,7 +1190,7 @@ GSTextureCache::Source* GSTextureCache::LookupDepthSource(const bool is_depth, c
 	return src;
 }
 
-GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const GIFRegTEX0& TEX0, const GIFRegTEXA& TEXA, const GIFRegCLAMP& CLAMP, const GSVector4i& r, const GSVector2i* lod, const bool possible_shuffle, const bool linear, const u32 frame_fbp, bool req_color, bool req_alpha)
+GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const GIFRegTEX0& TEX0, const GIFRegTEXA& TEXA, const GIFRegCLAMP& CLAMP, const GSVector4i& r, const GSVector2i* lod, const bool possible_shuffle, const bool linear, const GIFRegFRAME& frame, bool req_color, bool req_alpha)
 {
 	GL_CACHE("TC: Lookup Source <%d,%d => %d,%d> (0x%x, %s, BW: %u, CBP: 0x%x, TW: %d, TH: %d)", r.x, r.y, r.z, r.w, TEX0.TBP0, psm_str(TEX0.PSM), TEX0.TBW, TEX0.CBP, 1 << TEX0.TW, 1 << TEX0.TH);
 
@@ -1410,26 +1408,50 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 
 					if (rect_clean)
 					{
+						bool can_use = true;
 						for (auto& dirty : t->m_dirty)
 						{
-							GSVector4i dirty_rect = dirty.GetDirtyRect(t->m_TEX0, t->m_TEX0.PSM != dirty.psm);
+							const GSVector4i dirty_rect = dirty.GetDirtyRect(t->m_TEX0, t->m_TEX0.PSM != dirty.psm);
 							if (!dirty_rect.rintersect(new_rect).rempty())
 							{
 								rect_clean = false;
-								partial |= !new_rect.rintersect(dirty_rect).eq(new_rect) || dirty_rect.eq(new_rect);
+
+								if(!dirty_rect.rintersect(t->m_valid).eq(t->m_valid) || GSUtil::GetChannelMask(t->m_TEX0.PSM) != t->m_dirty.GetDirtyChannels())
+									partial |= !new_rect.rintersect(dirty_rect).eq(new_rect) || dirty_rect.eq(new_rect);
+								else // Nothing is valid anymore, kill it.
+								{
+									can_use = false;
+								}
 								break;
 							}
 						}
+
+						if (!can_use)
+						{
+							InvalidateSourcesFromTarget(t);
+							i = list.erase(i);
+							delete t;
+							continue;
+						}
 					}
 
 					const u32 channel_mask = GSUtil::GetChannelMask(psm);
 					const u32 channels = t->m_dirty.GetDirtyChannels() & channel_mask;
 
 					// If the source is reading the rt, make sure it's big enough.
-					if (!possible_shuffle && t && GSUtil::HasCompatibleBits(psm, t->m_TEX0.PSM) && real_fmt_match)
+					if (!possible_shuffle && t && GSUtil::HasCompatibleBits(psm, t->m_TEX0.PSM)&& real_fmt_match)
 					{
+						// Be careful if a new texture has been uploaded that expands the current one (Valkyrie Profile 2 does this)
+						GSVector4i dirty_rect = (t->m_dirty.size() > 0 && bw == t->m_TEX0.TBW) ? t->m_dirty.GetTotalRect(t->m_TEX0, GSVector2i(new_rect.z, new_rect.w)) : GSVector4i(GSVector4(t->m_valid) * GSVector4(2));
+						// Try to clamp the size of the target when using repeat, we don't want it getting too huge.
+						GSVector4i resize_rect = new_rect;
+						if (CLAMP.WMS == 0 || CLAMP.WMS == 3)
+							resize_rect.z = std::min(resize_rect.z, static_cast<int>(t->m_TEX0.TBW) * 64);
+						if ((CLAMP.WMT == 0 || CLAMP.WMT == 3) && resize_rect.w > (t->m_valid.w * 2))
+							resize_rect.w = std::min(resize_rect.w, std::max(t->m_valid.w * 2, dirty_rect.w));
+
 						if (t->Overlaps(bp, bw, psm, new_rect))
-							ResizeTarget(t, new_rect, bp, psm, bw);
+							ResizeTarget(t, resize_rect, bp, psm, bw);
 					}
 					// If not all channels are clean/dirty or only part of the rect is dirty, we need to update the target.
 					if (((channels & channel_mask) != channel_mask || partial))
@@ -1447,7 +1469,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 				else
 				{
 					rect_clean = t->m_dirty.empty();
-					if (!possible_shuffle && frame_fbp != t->m_TEX0.TBP0 && rect_clean && bp == t->m_TEX0.TBP0 && t && GSUtil::HasCompatibleBits(psm, t->m_TEX0.PSM) && width_match && real_fmt_match)
+					if (!possible_shuffle && frame.Block() != t->m_TEX0.TBP0 && rect_clean && bp == t->m_TEX0.TBP0 && t && GSUtil::HasCompatibleBits(psm, t->m_TEX0.PSM) && width_match && real_fmt_match)
 					{
 						if (!tex_merge_rt && t->Overlaps(bp, bw, psm, req_rect))
 						{
@@ -1462,13 +1484,21 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 								if (block_boundary_rect.w > t->m_valid.w)
 									AddDirtyRectTarget(t, GSVector4i(t->m_valid.x, t->m_valid.w, std::max(block_boundary_rect.z, t->m_valid.z), block_boundary_rect.w), t->m_TEX0.PSM, t->m_TEX0.TBW, rgba_mask);
 							}
+
+							// Try to clamp the size of the target when using repeat, we don't want it getting too huge.
+							GSVector4i resize_rect = req_rect;
+							if (CLAMP.WMS == 0 || CLAMP.WMS == 3)
+								resize_rect.z = std::min(resize_rect.z, static_cast<int>(t->m_TEX0.TBW) * 64);
+							if ((CLAMP.WMT == 0 || CLAMP.WMT == 3) && resize_rect.w > (t->m_valid.w * 2))
+								resize_rect.w = std::min(resize_rect.w, t->m_valid.w * 2);
+
 							// Resize including the extra pixel for bilinear.
-							ResizeTarget(t, req_rect, bp, psm, bw);
+							ResizeTarget(t, resize_rect, bp, psm, bw);
 						}
 					}
 				}
 
-				if (t->m_TEX0.TBP0 != frame_fbp && !possible_shuffle && bp > t->m_TEX0.TBP0 && t->Overlaps(bp, bw, psm, req_rect) && GSUtil::GetChannelMask(psm) == GSUtil::GetChannelMask(t->m_TEX0.PSM) && !width_match)
+				if (t->m_TEX0.TBP0 != frame.Block() && !possible_shuffle && bp > t->m_TEX0.TBP0 && t->Overlaps(bp, bw, psm, req_rect) && GSUtil::GetChannelMask(psm) == GSUtil::GetChannelMask(t->m_TEX0.PSM) && !width_match)
 				{
 					GSVector4i new_rect = req_rect;
 
@@ -1633,6 +1663,10 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 				{
 					u32 rt_tbw = std::max(1U, t->m_TEX0.TBW);
 					u32 horz_page_offset = ((bp - t->m_TEX0.TBP0) >> 5) % rt_tbw;
+
+					if (GSLocalMemory::m_psm[psm].bpp == GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp && bw != rt_tbw && block_boundary_rect.height() > GSLocalMemory::m_psm[psm].pgs.y)
+						continue;
+
 					if (GSLocalMemory::m_psm[color_psm].bpp == 16 && GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp == 32 && bw != 1 && 
 						((t->m_TEX0.TBW < (horz_page_offset + ((block_boundary_rect.z + GSLocalMemory::m_psm[psm].pgs.x - 1) / GSLocalMemory::m_psm[psm].pgs.x)) || 
 						(t->m_TEX0.TBW != bw && block_boundary_rect.w > GSLocalMemory::m_psm[psm].pgs.y))))
@@ -1662,22 +1696,27 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 					u32 src_psm = psm;
 
 					// If the input is C16 and it's actually a shuffle of 32bits we need to correct the size.
-					if ((tex_color_psm & 0xF) <= PSMCT24 && (psm & 0x7) == PSMCT16 && possible_shuffle)
+					if ((tex_color_psm & 0xF) <= PSMCT24 && (psm & 0x7) == PSMCT16)
 					{
-						src_psm = t->m_TEX0.PSM;
-						// If it's taking double width for the shuffle, half that.
-						if (src_bw == (rt_tbw * 2))
+						if (possible_shuffle)
 						{
-							src_bw = rt_tbw;
+							src_psm = t->m_TEX0.PSM;
+							// If it's taking double width for the shuffle, half that.
+							if (src_bw == (rt_tbw * 2))
+							{
+								src_bw = rt_tbw;
 
-							rect.x /= 2;
-							rect.z /= 2;
-						}
-						else
-						{
-							rect.y /= 2;
-							rect.w /= 2;
+								rect.x /= 2;
+								rect.z /= 2;
+							}
+							else
+							{
+								rect.y /= 2;
+								rect.w /= 2;
+							}
 						}
+						else // Formats are not compatible for normal draws, only shuffles.
+							continue;
 					}
 					if (bp > t->m_TEX0.TBP0)
 					{
@@ -1689,11 +1728,12 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 
 						GSVector4i new_rect = (GSLocalMemory::m_psm[color_psm].bpp != GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp && (psm & 0x7) != PSMCT16) ? block_boundary_rect : rect;
 
-						// Check if it is possible to hit with valid <x,y> offset on the given Target.
-						// Fixes Jak eyes rendering.
-						// Fixes Xenosaga 3 last dungeon graphic bug.
-						// Fixes Pause menu in The Getaway.
-						const bool can_translate = CanTranslate(bp, bw, src_psm, new_rect, t->m_TEX0.TBP0, t->m_TEX0.PSM, t->m_TEX0.TBW);
+						// If the sizing is completely wrong on the frame vs the source when reading from alpha then it's likely the target has 2 different sizes for rgb and alpha.
+						// This is just changing the target width for the rect translation, it has no bearing on the actual source read or the target itself.
+						// Hitman Blood Money is an example of this in the theatre.
+						const u32 rt_tbw = (possible_shuffle || bw == 1 || GSUtil::GetChannelMask(psm) != 0x8 || frame.FBW <= bw || frame.FBW == t->m_TEX0.TBW || bw == t->m_TEX0.TBW) ? t->m_TEX0.TBW : frame.FBW;
+
+						const bool can_translate = CanTranslate(bp, bw, src_psm, new_rect, t->m_TEX0.TBP0, t->m_TEX0.PSM, rt_tbw);
 						if (can_translate)
 						{
 							const bool swizzle_match = GSLocalMemory::m_psm[src_psm].depth == GSLocalMemory::m_psm[t->m_TEX0.PSM].depth;
@@ -1703,7 +1743,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 
 							if (swizzle_match)
 							{
-								rect = TranslateAlignedRectByPage(t, bp, src_psm, bw, new_rect);
+								rect = TranslateAlignedRectByPage(t->m_TEX0.TBP0, t->m_end_block, rt_tbw, t->m_TEX0.PSM, bp, src_psm, bw, new_rect);
 								rect.x -= new_rect.x;
 								rect.y -= new_rect.y;
 							}
@@ -1752,7 +1792,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 							dst = t;
 							tex_merge_rt = false;
 							found_t = true;
-							if (dst->m_TEX0.TBP0 == frame_fbp && possible_shuffle)
+							if (dst->m_TEX0.TBP0 == frame.Block() && possible_shuffle)
 								break;
 							else
 								continue;
@@ -1778,7 +1818,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 								tex_merge_rt = false;
 								found_t = true;
 								// Keep looking, just in case there is an exact match (Situation: Target frame drawn inside target frame, current makes a separate texture)
-								if (dst->m_TEX0.TBP0 == frame_fbp && possible_shuffle)
+								if (dst->m_TEX0.TBP0 == frame.Block() && possible_shuffle)
 									break;
 								else
 									continue;
@@ -1835,7 +1875,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 								break;
 							}
 
-							if (dst->m_TEX0.TBP0 == frame_fbp && possible_shuffle)
+							if (dst->m_TEX0.TBP0 == frame.Block() && possible_shuffle)
 								break;
 							else
 								continue;
@@ -1852,7 +1892,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 
 							// Prefer a target inside over a target outside.
 							found_t = false;
-							if (dst->m_TEX0.TBP0 == frame_fbp && possible_shuffle)
+							if (dst->m_TEX0.TBP0 == frame.Block() && possible_shuffle)
 								break;
 							else
 								continue;
@@ -1913,7 +1953,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 							GIFRegTEX0 depth_TEX0;
 							depth_TEX0.U32[0] = TEX0.U32[0] | (0x30u << 20u);
 							depth_TEX0.U32[1] = TEX0.U32[1];
-							src = LookupDepthSource(false, depth_TEX0, TEXA, CLAMP, block_boundary_rect, possible_shuffle, linear, frame_fbp, req_color, req_alpha);
+							src = LookupDepthSource(false, depth_TEX0, TEXA, CLAMP, block_boundary_rect, possible_shuffle, linear, frame, req_color, req_alpha);
 
 							if (src != nullptr)
 							{
@@ -1935,7 +1975,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 							}
 							else
 							{
-								src = LookupDepthSource(false, TEX0, TEXA, CLAMP, block_boundary_rect, possible_shuffle, linear, frame_fbp, req_color, req_alpha, true);
+								src = LookupDepthSource(false, TEX0, TEXA, CLAMP, block_boundary_rect, possible_shuffle, linear, frame, req_color, req_alpha, true);
 
 								if (src != nullptr)
 								{
@@ -2066,6 +2106,10 @@ GSVector2i GSTextureCache::ScaleRenderTargetSize(const GSVector2i& sz, float sca
 
 void GSTextureCache::CombineAlignedInsideTargets(Target* target, GSTextureCache::Source* src)
 {
+	// Don't combine targets if Tex in RT is off, it will just fail to find them and make a new one, causing a loop of copies.
+	if (GSConfig.UserHacks_TextureInsideRt < GSTextureInRtMode::InsideTargets)
+		return;
+
 	auto& list = m_dst[target->m_type];
 
 	for (auto i = list.begin(); i != list.end();)
@@ -2075,7 +2119,7 @@ void GSTextureCache::CombineAlignedInsideTargets(Target* target, GSTextureCache:
 		if (t != target)
 		{
 			// Target not contained, skip it.
-			if (t->m_TEX0.TBP0 < target->m_TEX0.TBP0 || t->m_end_block > target->m_end_block)
+			if (t->m_TEX0.TBP0 < target->m_TEX0.TBP0 || t->UnwrappedEndBlock() > target->UnwrappedEndBlock())
 			{
 				i++;
 				continue;
@@ -2089,14 +2133,32 @@ void GSTextureCache::CombineAlignedInsideTargets(Target* target, GSTextureCache:
 
 				if ((page_offset + page_width) <= target->m_TEX0.TBW)
 				{
-					const u32 vertical_offset = (((t->m_TEX0.TBP0 - target->m_TEX0.TBP0) >> 5) / std::max(1U, t->m_TEX0.TBW)) * t_psm.pgs.y;
-					const u32 horizontal_offset = page_offset * t_psm.pgs.x;
-					const GSVector4i target_drect_unscaled = t->m_valid + GSVector4i(horizontal_offset, vertical_offset).xyxy();
-					const GSVector4 target_drect = GSVector4(target_drect_unscaled) * target->m_scale;
+					if (t->m_last_draw > target->m_last_draw || t->m_valid.rintersect(target->m_valid).rempty())
+					{
+						if (!t->m_drawn_since_read.rempty())
+						{
+							t->Update();
 
-					g_gs_device->StretchRect(t->m_texture, GSVector4(0, 0, 1, 1), target->m_texture, target_drect, (target->m_type == RenderTarget) ? ShaderConvert::COPY : ShaderConvert::DEPTH_COPY, t->m_scale < target->m_scale);
+							const u32 vertical_offset = (((t->m_TEX0.TBP0 - target->m_TEX0.TBP0) >> 5) / std::max(1U, t->m_TEX0.TBW)) * t_psm.pgs.y;
+							const u32 horizontal_offset = page_offset * t_psm.pgs.x;
+							const GSVector4i target_drect_unscaled = t->m_drawn_since_read + GSVector4i(horizontal_offset, vertical_offset).xyxy();
 
-					target->UpdateValidity(target_drect_unscaled);
+							const GSVector4 source_rect = GSVector4(t->m_drawn_since_read) / (GSVector4(t->m_unscaled_size).xyxy() * t->GetScale());
+							const GSVector4 target_drect = GSVector4(target_drect_unscaled) * target->m_scale;
+
+							const bool valid_color = t->m_valid_rgb;
+							const bool valid_alpha = (t->m_valid_alpha_high | t->m_valid_alpha_low) && (GSUtil::GetChannelMask(t->m_TEX0.PSM) & 0x8);
+
+							target->m_valid_alpha_high |= t->m_valid_alpha_high;
+							target->m_valid_alpha_low |= t->m_valid_alpha_low;
+
+							GL_CACHE("Combining %x-%x in to %x-%x draw %d", t->m_TEX0.TBP0, t->m_end_block, target->m_TEX0.TBP0, target->m_end_block, GSState::s_n);
+
+							g_gs_device->StretchRect(t->m_texture, source_rect, target->m_texture, target_drect, valid_color, valid_color, valid_color, valid_alpha, (target->m_type == RenderTarget) ? ShaderConvert::COPY : ShaderConvert::DEPTH_COPY);
+
+							target->UpdateValidity(target_drect_unscaled);
+						}
+					}
 
 					if (src && src->m_from_target == t)
 					{
@@ -2235,7 +2297,7 @@ GSTextureCache::Target* GSTextureCache::LookupTarget(GIFRegTEX0 TEX0, const GSVe
 					dst->m_32_bits_fmt |= (psm_s.bpp != 16);
 					break;
 				}
-				else if(!(src && src->m_from_target == t))
+				else if (!(src && src->m_from_target == t))
 				{
 					GL_INS("TC: Deleting RT BP 0x%x BW %d PSM %s due to change in target", t->m_TEX0.TBP0, t->m_TEX0.TBW, psm_str(t->m_TEX0.PSM));
 					InvalidateSourcesFromTarget(t);
@@ -2252,10 +2314,12 @@ GSTextureCache::Target* GSTextureCache::LookupTarget(GIFRegTEX0 TEX0, const GSVe
 				const bool is_aligned_ok = widthpage_offset == 0 || ((min_rect.width() <= static_cast<int>((t->m_TEX0.TBW - widthpage_offset) * 64) && (t->m_TEX0.TBW == TEX0.TBW || TEX0.TBW == 1)) && bp >= t->m_TEX0.TBP0);
 				const bool no_target_or_newer = (!dst || ((GSState::s_n - dst->m_last_draw) < (GSState::s_n - t->m_last_draw)));
 				const bool width_match = (t->m_TEX0.TBW == TEX0.TBW || (TEX0.TBW == 1 && draw_rect.w <= GSLocalMemory::m_psm[t->m_TEX0.PSM].pgs.y));
+				const bool ds_offset = !ds || offset != 0;
+				const bool is_double_buffer = TEX0.TBP0 == ((((t->m_end_block + 1) - t->m_TEX0.TBP0) / 2) + t->m_TEX0.TBP0);
 				// if it's a shuffle, some games tend to offset back by a page, such as Tomb Raider, for no disernable reason, but it then causes problems.
 				// This can also happen horizontally (Catwoman moves everything one page left with shuffles), but this is too messy to deal with right now.
 				const bool overlaps = t->Overlaps(bp, TEX0.TBW, TEX0.PSM, min_rect) || (is_shuffle && src && GSLocalMemory::m_psm[src->m_TEX0.PSM].bpp == 8 && t->Overlaps(bp, TEX0.TBW, TEX0.PSM, min_rect + GSVector4i(0, 0, 0, 32)));
-				if (no_target_or_newer && is_aligned_ok && width_match && overlaps)
+				if (no_target_or_newer && is_aligned_ok && width_match && overlaps && (is_shuffle || ds_offset || is_double_buffer))
 				{
 					const GSLocalMemory::psm_t& s_psm = GSLocalMemory::m_psm[TEX0.PSM];
 
@@ -2286,13 +2350,33 @@ GSTextureCache::Target* GSTextureCache::LookupTarget(GIFRegTEX0 TEX0, const GSVe
 						InvalidateSourcesFromTarget(t);
 						i = list.erase(i);
 						delete t;
-						
+
 						continue;
 					}
-					else if (t->m_dirty.empty() || (t->m_TEX0.TBP0 <= bp && t->m_last_draw >= (GSState::s_n - 1) && 
-													t->m_dirty.GetTotalRect(t->m_TEX0, t->m_unscaled_size).rintersect(GSVector4i(0, 0, 0, 0).max_i32(TranslateAlignedRectByPage(t, TEX0.TBP0, TEX0.PSM, TEX0.TBW, min_rect))).rempty()))
+
+					GSVector4i lookup_rect = min_rect;
+
+					if (is_shuffle)
+						lookup_rect = lookup_rect & GSVector4i(~8);
+
+					const GSVector4i translated_rect = GSVector4i(0, 0, 0, 0).max_i32(TranslateAlignedRectByPage(t, TEX0.TBP0, TEX0.PSM, TEX0.TBW, lookup_rect));
+					const GSVector4i dirty_rect = t->m_dirty.empty() ? GSVector4i::zero() : t->m_dirty.GetTotalRect(t->m_TEX0, t->m_unscaled_size);
+					const bool all_dirty = dirty_rect.eq(t->m_valid);
+
+					
+					if (!is_shuffle && !dirty_rect.rempty() && (!preserve_alpha && !preserve_rgb) && (GSState::s_n - 3) > t->m_last_draw)
 					{
-						if (TEX0.TBW == t->m_TEX0.TBW && !is_shuffle && widthpage_offset == 0 && ((min_rect.w + 63)/ 64) > 1)
+						GL_INS("TC: Deleting RT BP 0x%x BW %d PSM %s due to dirty areas not preserved (Likely change in target)", t->m_TEX0.TBP0, t->m_TEX0.TBW, psm_str(t->m_TEX0.PSM));
+						InvalidateSourcesFromTarget(t);
+						i = list.erase(i);
+						delete t;
+
+						continue;
+					}
+
+					if (!all_dirty && ((translated_rect.w <= t->m_valid.w) || widthpage_offset == 0 || (GSState::s_n - 3) <= t->m_last_draw))
+					{
+						if (TEX0.TBW == t->m_TEX0.TBW && !is_shuffle && widthpage_offset == 0 && ((min_rect.w + 63) / 64) > 1)
 						{
 							// Beyond Good and Evil does this awful thing where it puts one framebuffer at 0xf00, with the first row of pages blanked out, and the whole thing goes down to 0x2080
 							// which is a problem, because it then puts the Z buffer at 0x1fc0, then offsets THAT by 1 row of pages, so it starts at, you guessed it, 2080.
@@ -2536,7 +2620,7 @@ GSTextureCache::Target* GSTextureCache::LookupTarget(GIFRegTEX0 TEX0, const GSVe
 					src->m_target_direct = false;
 					src->m_shared_texture = false;
 
-					if(!req_copy)
+					if (!req_copy)
 						dst->ResizeTexture(new_size.x, new_size.y, true, true, GSVector4i(dRect), true);
 					else
 					{
@@ -2586,7 +2670,7 @@ GSTextureCache::Target* GSTextureCache::LookupTarget(GIFRegTEX0 TEX0, const GSVe
 			if ((!is_shuffle && (GSLocalMemory::m_psm[dst->m_TEX0.PSM].bpp != GSLocalMemory::m_psm[TEX0.PSM].bpp || GSLocalMemory::m_psm[dst->m_TEX0.PSM].depth != GSLocalMemory::m_psm[TEX0.PSM].depth)) || 
 				(is_shuffle && GSLocalMemory::m_psm[dst->m_TEX0.PSM].bpp == 16))
 			{
-				if(GSLocalMemory::m_psm[dst->m_TEX0.PSM].depth != GSLocalMemory::m_psm[TEX0.PSM].depth || dst->m_TEX0.TBW != TEX0.TBW)
+				if (GSLocalMemory::m_psm[dst->m_TEX0.PSM].depth != GSLocalMemory::m_psm[TEX0.PSM].depth || dst->m_TEX0.TBW != TEX0.TBW)
 					dst->m_32_bits_fmt = GSLocalMemory::m_psm[TEX0.PSM].bpp != 16;
 
 				if (!is_shuffle || (is_shuffle && GSLocalMemory::m_psm[dst->m_TEX0.PSM].bpp == 16))
@@ -2636,7 +2720,8 @@ GSTextureCache::Target* GSTextureCache::LookupTarget(GIFRegTEX0 TEX0, const GSVe
 				}
 			}
 
-			if (!dst->m_valid_rgb)
+			const u32 mask = GSLocalMemory::m_psm[TEX0.PSM].fmsk;
+			if (!dst->m_valid_rgb && ((fbmask & 0x00FFFFFF) & mask) != (mask & 0x00FFFFFF))
 			{
 				GL_CACHE("TC: Cannot find RGB target for %s[%x], clearing.", to_string(type), dst->m_TEX0.TBP0);
 
@@ -2718,11 +2803,11 @@ GSTextureCache::Target* GSTextureCache::LookupTarget(GIFRegTEX0 TEX0, const GSVe
 		for (auto i = rev_list.begin(); i != rev_list.end(); ++i)
 		{
 			Target* t = *i;
-			// Don't pull in targets without valid lower 24 bits, it makes no sense to convert them.
+			// Don't pull in targets without valid lower 24 bits unless the Z is 32bits and the alpha is valid, it makes no sense to convert them otherwise.
 			// FIXME: Technically the difference in size is fine, but if the target gets reinterpreted, the hw renderer doesn't rearrange the target.
 			// This does cause some extra uploads in some games (like Burnout), but without this, bad data gets displayed in games like Transformers.
-			if (bp != t->m_TEX0.TBP0 || !t->m_valid_rgb || (!is_shuffle && t->m_TEX0.TBW != TEX0.TBW && 
-				(possible_clear || ((~GSLocalMemory::m_psm[t->m_TEX0.PSM].fmsk | fbmask) == 0xffffffff))))
+			if (bp != t->m_TEX0.TBP0 || (!t->m_valid_rgb && (!(GSUtil::GetChannelMask(TEX0.PSM) & 0x8) || !(t->m_valid_alpha_low || t->m_valid_alpha_high))) ||
+				(!is_shuffle && t->m_TEX0.TBW != TEX0.TBW && (possible_clear || ((~GSLocalMemory::m_psm[t->m_TEX0.PSM].fmsk | fbmask) == 0xffffffff))))
 			{
 				continue;
 			}
@@ -2957,9 +3042,8 @@ GSTextureCache::Target* GSTextureCache::CreateTarget(GIFRegTEX0 TEX0, const GSVe
 	if (!is_frame)
 	{
 		// Not having this valid could make things explode, but I do enjoy watching the world burn (and this is actually more correct).
-		dst->m_valid_rgb =true;
-
 		const u32 mask = GSLocalMemory::m_psm[TEX0.PSM].fmsk;
+		dst->m_valid_rgb = GSLocalMemory::m_psm[TEX0.PSM].depth || ((fbmask & 0x00FFFFFF) & mask) != (mask & 0x00FFFFFF) || (dst->m_dirty.GetDirtyChannels() & 0x7);
 
 		// If there is an opposite target without valid RGB, we need to match them up
 		auto& rev_list = m_dst[1 - type];
@@ -2968,9 +3052,6 @@ GSTextureCache::Target* GSTextureCache::CreateTarget(GIFRegTEX0 TEX0, const GSVe
 			Target* const rev_t = *j;
 			if (rev_t->m_TEX0.TBP0 == dst->m_TEX0.TBP0 && GSLocalMemory::m_psm[rev_t->m_TEX0.PSM].bpp == GSLocalMemory::m_psm[dst->m_TEX0.PSM].bpp)
 			{
-				if (GSLocalMemory::m_psm[rev_t->m_TEX0.PSM].trbpp == 24 && ((fbmask & 0x00FFFFFF) & mask) == (mask & 0x00FFFFFF))
-					dst->m_valid_rgb = false;
-
 				if (!rev_t->m_valid_rgb && dst->m_valid_rgb)
 					rev_t->m_was_dst_matched = true;
 
@@ -3036,6 +3117,8 @@ GSTextureCache::Target* GSTextureCache::CreateTarget(GIFRegTEX0 TEX0, const GSVe
 	else
 		dst->m_last_draw += 1; // If we preload and it needs to decorrect and we couldn't catch it early, we need to make sure it decorrects the data.
 
+	CombineAlignedInsideTargets(dst, src);
+
 	pxAssert(dst && dst->m_texture && dst->m_scale == scale);
 	return dst;
 }
@@ -3207,23 +3290,19 @@ bool GSTextureCache::PreloadTarget(GIFRegTEX0 TEX0, const GSVector2i& size, cons
 
 					if (buffer_width != std::max(1U, t->m_TEX0.TBW))
 					{
+						i++;
 						// Check if this got messed with at some point, if it did just nuke it.
-						if (t->m_valid.width() == dst->m_valid.width())
+						if (!preserve_target && t->m_age > 0)
 						{
-							// Not correct, but it's better than a null reference.
+							// Probably best we don't poke the beast if it's being used as the current source.
 							if (src && src->m_target_direct && src->m_from_target == t)
-							{
-								DevCon.Warning("Replacing source target, texture may be invalid");
-								src->m_texture = dst->m_texture;
-								src->m_from_target = dst;
-							}
+								continue;
 							
 							InvalidateSourcesFromTarget(t);
 							i = list.erase(j);
 							delete t;
 						}
-						else
-							i++;
+
 						continue;
 					}
 					// If the two targets are misaligned, it's likely a relocation, so we can just kill the old target.
@@ -3859,23 +3938,30 @@ void GSTextureCache::InvalidateContainedTargets(u32 start_bp, u32 end_bp, u32 wr
 				continue;
 			}
 
-			// If not fully contained, just dirty the area.
-			if (start_bp != t->m_TEX0.TBP0 && (t->m_TEX0.TBP0 < start_bp || t->UnwrappedEndBlock() > end_bp))
+			// If not fully contained but they are aligned and or clean, just dirty the area.
+			if (type != DepthStencil && start_bp != t->m_TEX0.TBP0 && (t->m_TEX0.TBP0 < start_bp || t->UnwrappedEndBlock() > end_bp))
 			{
-				if (write_bw == t->m_TEX0.TBW && GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp == GSLocalMemory::m_psm[write_psm].bpp)
-				{
-					const u32 page_offset = ((end_bp - start_bp) >> 5);
-					const u32 end_width = write_bw * 64;
-					const u32 end_height = ((page_offset / std::max(write_bw, 1U)) * GSLocalMemory::m_psm[write_psm].pgs.y) + GSLocalMemory::m_psm[write_psm].pgs.y;
-					const GSVector4i r = GSVector4i(0, 0, end_width, end_height);
-					const GSVector4i invalidate_r = TranslateAlignedRectByPage(t, start_bp, write_psm, write_bw, r, false).rintersect(t->m_valid); // it is invalidation but we need a real rect.
-					RGBAMask mask;
-					mask._u32 = GSUtil::GetChannelMask(write_psm);
-					AddDirtyRectTarget(t, invalidate_r, t->m_TEX0.PSM, t->m_TEX0.TBW, mask, false);
-				}
+				const u32 offset = (std::abs(static_cast<int>(start_bp - t->m_TEX0.TBP0)) >> 5) % std::max(1U, t->m_TEX0.TBW);
+				const GSVector4i dirty_rect = t->m_dirty.GetTotalRect(t->m_TEX0, t->m_unscaled_size).rintersect(t->m_valid);
+				const u32 end_page_offset = ((end_bp - start_bp) >> 5);
+				const u32 end_width = write_bw * 64;
+				const u32 end_height = ((end_page_offset / std::max(write_bw, 1U)) * GSLocalMemory::m_psm[write_psm].pgs.y) + GSLocalMemory::m_psm[write_psm].pgs.y;
+				const GSVector4i r = GSVector4i(0, 0, end_width, end_height);
+				const GSVector4i invalidate_r = TranslateAlignedRectByPage(t, start_bp, write_psm, write_bw, r, false).rintersect(t->m_valid); // it is invalidation but we need a real rect.
 
-				++i;
-				continue;
+				if (offset == 0 || dirty_rect.rempty() || !dirty_rect.rintersect(invalidate_r).rempty())
+				{
+					if (write_bw == t->m_TEX0.TBW && GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp == GSLocalMemory::m_psm[write_psm].bpp)
+					{
+						
+						RGBAMask mask;
+						mask._u32 = GSUtil::GetChannelMask(write_psm);
+						AddDirtyRectTarget(t, invalidate_r, t->m_TEX0.PSM, t->m_TEX0.TBW, mask, false);
+					}
+
+					++i;
+					continue;
+				}
 			}
 
 			InvalidateSourcesFromTarget(t);
@@ -4599,6 +4685,9 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 		dst = LookupTarget(new_TEX0, target_size, src->m_scale, src->m_type);
 		if (!dst)
 			dst = CreateTarget(new_TEX0, target_size, target_size, src->m_scale, src->m_type);
+		else // Expand if necessary (Silent hill 4 takes an old target which is smaller).
+			dst->ResizeTexture(std::max(dst->m_unscaled_size.x, target_size.x), std::max(dst->m_unscaled_size.y, target_size.y));
+
 		if (!dst)
 			return false;
 
@@ -5374,7 +5463,7 @@ GSTextureCache::Source* GSTextureCache::CreateSource(const GIFRegTEX0& TEX0, con
 			src->m_unscaled_size = dst->m_unscaled_size;
 			src->m_shared_texture = true;
 
-			if(channel_shuffle)
+			if (channel_shuffle)
 				m_temporary_source = src;
 		}
 
@@ -5724,7 +5813,7 @@ GSTextureCache::Source* GSTextureCache::CreateSource(const GIFRegTEX0& TEX0, con
 					src->m_region.SetX(x_offset, x_offset + tw);
 					src->m_region.SetY(y_offset, y_offset + th);
 
-					if(!GSConfig.UserHacks_NativePaletteDraw)
+					if (!GSConfig.UserHacks_NativePaletteDraw)
 						m_temporary_source = src;
 				}
 				else
@@ -6714,7 +6803,7 @@ GSTextureCache::Source::~Source()
 	// to recycle.
 	if (!m_shared_texture && !m_from_hash_cache && m_texture)
 	{
-		if(m_from_target)
+		if (m_from_target)
 			g_texture_cache->m_target_memory_usage -= m_texture->GetMemUsage();
 		else
 			g_texture_cache->m_source_memory_usage -= m_texture->GetMemUsage();

pcsx2/GS/Renderers/HW/GSTextureCache.h

@@ -209,7 +209,7 @@ public:
 	struct TempZAddress
 	{
 		u32 ZBP;
-		u32 offset;
+		int offset;
 		GSVector4i rect_since;
 	};
 
@@ -494,8 +494,8 @@ public:
 	GSTexture* LookupPaletteSource(u32 CBP, u32 CPSM, u32 CBW, GSVector2i& offset, float* scale, const GSVector2i& size);
 	std::shared_ptr<Palette> LookupPaletteObject(const u32* clut, u16 pal, bool need_gs_texture);
 
-	Source* LookupSource(const bool is_color, const GIFRegTEX0& TEX0, const GIFRegTEXA& TEXA, const GIFRegCLAMP& CLAMP, const GSVector4i& r, const GSVector2i* lod, const bool possible_shuffle, const bool linear, const u32 frame_fbp = 0xFFFFFFFF, bool req_color = true, bool req_alpha = true);
-	Source* LookupDepthSource(const bool is_depth, const GIFRegTEX0& TEX0, const GIFRegTEXA& TEXA, const GIFRegCLAMP& CLAMP, const GSVector4i& r, const bool possible_shuffle, const bool linear, const u32 frame_fbp = 0xFFFFFFFF, bool req_color = true, bool req_alpha = true, bool palette = false);
+	Source* LookupSource(const bool is_color, const GIFRegTEX0& TEX0, const GIFRegTEXA& TEXA, const GIFRegCLAMP& CLAMP, const GSVector4i& r, const GSVector2i* lod, const bool possible_shuffle, const bool linear, const GIFRegFRAME& frame, bool req_color = true, bool req_alpha = true);
+	Source* LookupDepthSource(const bool is_depth, const GIFRegTEX0& TEX0, const GIFRegTEXA& TEXA, const GIFRegCLAMP& CLAMP, const GSVector4i& r, const bool possible_shuffle, const bool linear, const GIFRegFRAME& frame, bool req_color = true, bool req_alpha = true, bool palette = false);
 
 	Target* FindTargetOverlap(Target* target, int type, int psm);
 	void CombineAlignedInsideTargets(Target* target, GSTextureCache::Source* src = nullptr);

pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.cpp

@@ -2489,12 +2489,15 @@ void GSDeviceOGL::RenderHW(GSHWDrawConfig& config)
 
 	if (config.require_one_barrier && !m_features.texture_barrier)
 	{
-		// Requires a copy of the RT
+		// Requires a copy of the RT.
 		draw_rt_clone = CreateTexture(rtsize.x, rtsize.y, 1, colclip_rt ? GSTexture::Format::ColorClip : GSTexture::Format::Color, true);
-		GL_PUSH("Copy RT to temp texture for fbmask {%d,%d %dx%d}",
-			config.drawarea.left, config.drawarea.top,
-			config.drawarea.width(), config.drawarea.height());
-		CopyRect(colclip_rt ? colclip_rt : config.rt, draw_rt_clone, config.drawarea, config.drawarea.left, config.drawarea.top);
+		if (draw_rt_clone)
+		{
+			GL_PUSH("GL: Copy RT to temp texture {%d,%d %dx%d}",
+				config.drawarea.left, config.drawarea.top,
+				config.drawarea.width(), config.drawarea.height());
+			CopyRect(colclip_rt ? colclip_rt : config.rt, draw_rt_clone, config.drawarea, config.drawarea.left, config.drawarea.top);
+		}
 	}
 
 	IASetVertexBuffer(config.verts, config.nverts);
@@ -2563,7 +2566,7 @@ void GSDeviceOGL::RenderHW(GSHWDrawConfig& config)
 	if (check_barrier && ((config.tex && (config.tex == config.ds || config.tex == config.rt)) || ((psel.ps.IsFeedbackLoop() || psel.ps.blend_c == 1) && GLState::rt == config.rt)))
 	{
 		// Ensure all depth writes are finished before sampling
-		GL_INS("Texture barrier to flush depth or rt before reading");
+		GL_INS("GL: Texture barrier to flush depth or rt before reading");
 		glTextureBarrier();
 	}
 	// additional non-pipeline config stuff

pcsx2/GS/Renderers/SW/GSDrawScanline.cpp

@@ -1491,7 +1491,7 @@ __ri void GSDrawScanline::CDrawScanline(int pixels, int left, int top, const GSV
 						case 2: break;
 					}
 
-					if(!(sel.fpsm == 1 && sel.abc == 1))
+					if (!(sel.fpsm == 1 && sel.abc == 1))
 					{
 						switch(sel.abc)
 						{

pcsx2/GS/Renderers/SW/GSDrawScanlineCodeGenerator.all.cpp

@@ -562,7 +562,7 @@ L("loop");
 
 L("step");
 
-	// if(steps <= 0) break;
+	// if (steps <= 0) break;
 
 	if (!m_sel.edge)
 	{
@@ -768,7 +768,7 @@ void GSDrawScanlineCodeGenerator::Init()
 				cvttps2dq(xym6, vt);
 
 				// s = vti.xxxx() + m_local.d[skip].s;
-				// t = vti.yyyy(); if(!sprite) t += m_local.d[skip].t;
+				// t = vti.yyyy(); if (!sprite) t += m_local.d[skip].t;
 
 				pshufd(s, xym6, _MM_SHUFFLE(0, 0, 0, 0));
 				pshufd(t, xym6, _MM_SHUFFLE(1, 1, 1, 1));
@@ -925,7 +925,7 @@ void GSDrawScanlineCodeGenerator::Step()
 				// GSVector4i stq = m_local.d4.stq;
 
 				// s += stq.xxxx();
-				// if(!sprite) t += st.yyyy();
+				// if (!sprite) t += st.yyyy();
 
 				broadcasti128(stq, _rip_local_d(stq));
 
@@ -2096,7 +2096,7 @@ void GSDrawScanlineCodeGenerator::AlphaTFX()
 
 			clamp16(_ga, tmp);
 
-			// if(!tcc) gat = gat.mix16(ga.srl16(7));
+			// if (!tcc) gat = gat.mix16(ga.srl16(7));
 
 			if (!m_sel.tcc)
 			{
@@ -2109,7 +2109,7 @@ void GSDrawScanlineCodeGenerator::AlphaTFX()
 
 		case TFX_DECAL:
 
-			// if(!tcc) gat = gat.mix16(ga.srl16(7));
+			// if (!tcc) gat = gat.mix16(ga.srl16(7));
 			if (!m_sel.tcc)
 			{
 				// GSVector4i ga = iip ? gaf : m_local.c.ga;
@@ -2139,7 +2139,7 @@ void GSDrawScanlineCodeGenerator::AlphaTFX()
 
 		case TFX_HIGHLIGHT2:
 
-			// if(!tcc) gat = gat.mix16(ga.srl16(7));
+			// if (!tcc) gat = gat.mix16(ga.srl16(7));
 
 			if (!m_sel.tcc)
 			{
@@ -2990,8 +2990,8 @@ void GSDrawScanlineCodeGenerator::WritePixel(const XYm& src_, const AddressReg&
 	{
 		if (fast)
 		{
-			// if(fzm & 0x0f) GSVector4i::storel(&vm16[addr + 0], fs);
-			// if(fzm & 0xf0) GSVector4i::storeh(&vm16[addr + 8], fs);
+			// if (fzm & 0x0f) GSVector4i::storel(&vm16[addr + 0], fs);
+			// if (fzm & 0xf0) GSVector4i::storeh(&vm16[addr + 8], fs);
 
 			test(mask, 0x0000000f << shift);
 			je("@f");
@@ -3020,10 +3020,10 @@ void GSDrawScanlineCodeGenerator::WritePixel(const XYm& src_, const AddressReg&
 		}
 		else
 		{
-			// if(fzm & 0x03) WritePixel(fpsm, &vm16[addr + 0], fs.extract32<0>());
-			// if(fzm & 0x0c) WritePixel(fpsm, &vm16[addr + 2], fs.extract32<1>());
-			// if(fzm & 0x30) WritePixel(fpsm, &vm16[addr + 8], fs.extract32<2>());
-			// if(fzm & 0xc0) WritePixel(fpsm, &vm16[addr + 10], fs.extract32<3>());
+			// if (fzm & 0x03) WritePixel(fpsm, &vm16[addr + 0], fs.extract32<0>());
+			// if (fzm & 0x0c) WritePixel(fpsm, &vm16[addr + 2], fs.extract32<1>());
+			// if (fzm & 0x30) WritePixel(fpsm, &vm16[addr + 8], fs.extract32<2>());
+			// if (fzm & 0xc0) WritePixel(fpsm, &vm16[addr + 10], fs.extract32<3>());
 
 			test(mask, 0x00000003 << shift);
 			je("@f");

pcsx2/GS/Renderers/SW/GSDrawScanlineCodeGenerator.arm64.cpp

@@ -158,7 +158,7 @@ void GSDrawScanlineCodeGenerator::Generate()
 	Label exit;
 	armAsm->Bind(&m_step_label);
 
-	// if(steps <= 0) break;
+	// if (steps <= 0) break;
 
 	if (!m_sel.edge)
 	{
@@ -332,7 +332,7 @@ void GSDrawScanlineCodeGenerator::Init()
 				armAsm->Fcvtzs(v6.V4S(), v4.V4S());
 
 				// s = vti.xxxx() + m_local.d[skip].s;
-				// t = vti.yyyy(); if(!sprite) t += m_local.d[skip].t;
+				// t = vti.yyyy(); if (!sprite) t += m_local.d[skip].t;
 
 				armAsm->Dup(_temp_s.V4S(), v6.V4S(), 0);
 				armAsm->Dup(_temp_t.V4S(), v6.V4S(), 1);
@@ -485,7 +485,7 @@ void GSDrawScanlineCodeGenerator::Step()
 				// GSVector4i stq = m_local.d4.stq;
 
 				// s += stq.xxxx();
-				// if(!sprite) t += stq.yyyy();
+				// if (!sprite) t += stq.yyyy();
 
 				armAsm->Dup(_vscratch.V4S(), _d4_stq.V4S(), 0);
 				if (m_sel.prim != GS_SPRITE_CLASS || m_sel.mmin)
@@ -1394,7 +1394,7 @@ void GSDrawScanlineCodeGenerator::AlphaTFX()
 			modulate16(v6, _temp_ga, 1);
 			clamp16(v6, v3);
 
-			// if(!tcc) gat = gat.mix16(ga.srl16(7));
+			// if (!tcc) gat = gat.mix16(ga.srl16(7));
 
 			if (!m_sel.tcc)
 			{
@@ -1407,7 +1407,7 @@ void GSDrawScanlineCodeGenerator::AlphaTFX()
 
 		case TFX_DECAL:
 
-			// if(!tcc) gat = gat.mix16(ga.srl16(7));
+			// if (!tcc) gat = gat.mix16(ga.srl16(7));
 
 			if (!m_sel.tcc)
 			{
@@ -1435,7 +1435,7 @@ void GSDrawScanlineCodeGenerator::AlphaTFX()
 
 		case TFX_HIGHLIGHT2:
 
-			// if(!tcc) gat = gat.mix16(ga.srl16(7));
+			// if (!tcc) gat = gat.mix16(ga.srl16(7));
 
 			if (!m_sel.tcc)
 			{
@@ -2158,8 +2158,8 @@ void GSDrawScanlineCodeGenerator::WritePixel(const VRegister& src, const Registe
 	{
 		if (fast)
 		{
-			// if(fzm & 0x0f) GSVector4i::storel(&vm16[addr + 0], fs);
-			// if(fzm & 0xf0) GSVector4i::storeh(&vm16[addr + 8], fs);
+			// if (fzm & 0x0f) GSVector4i::storel(&vm16[addr + 0], fs);
+			// if (fzm & 0xf0) GSVector4i::storeh(&vm16[addr + 8], fs);
 
 			Label skip_low, skip_high;
 			armAsm->Lsl(_wscratch, addr, 1); // *2
@@ -2177,10 +2177,10 @@ void GSDrawScanlineCodeGenerator::WritePixel(const VRegister& src, const Registe
 		}
 		else
 		{
-			// if(fzm & 0x03) WritePixel(fpsm, &vm16[addr + 0], fs.extract32<0>());
-			// if(fzm & 0x0c) WritePixel(fpsm, &vm16[addr + 2], fs.extract32<1>());
-			// if(fzm & 0x30) WritePixel(fpsm, &vm16[addr + 8], fs.extract32<2>());
-			// if(fzm & 0xc0) WritePixel(fpsm, &vm16[addr + 10], fs.extract32<3>());
+			// if (fzm & 0x03) WritePixel(fpsm, &vm16[addr + 0], fs.extract32<0>());
+			// if (fzm & 0x0c) WritePixel(fpsm, &vm16[addr + 2], fs.extract32<1>());
+			// if (fzm & 0x30) WritePixel(fpsm, &vm16[addr + 8], fs.extract32<2>());
+			// if (fzm & 0xc0) WritePixel(fpsm, &vm16[addr + 10], fs.extract32<3>());
 
 			Label skip_0, skip_1, skip_2, skip_3;
 

pcsx2/GS/Renderers/SW/GSRasterizer.cpp

@@ -444,9 +444,9 @@ void GSRasterizer::DrawTriangle(const GSVertexSW* vertex, const u16* index)
 
 	m1 = (y0011 == y1221).mask() & 7;
 
-	// if(i == 0) => y0 < y1 < y2
-	// if(i == 1) => y0 == y1 < y2
-	// if(i == 4) => y0 < y1 == y2
+	// if (i == 0) => y0 < y1 < y2
+	// if (i == 1) => y0 == y1 < y2
+	// if (i == 4) => y0 < y1 == y2
 
 	if (m1 == 7) // y0 == y1 == y2
 		return;
@@ -623,9 +623,9 @@ void GSRasterizer::DrawTriangle(const GSVertexSW* vertex, const u16* index)
 
 	m1 = (y0011 == y1221).mask() & 7;
 
-	// if(i == 0) => y0 < y1 < y2
-	// if(i == 1) => y0 == y1 < y2
-	// if(i == 4) => y0 < y1 == y2
+	// if (i == 0) => y0 < y1 < y2
+	// if (i == 1) => y0 == y1 < y2
+	// if (i == 4) => y0 < y1 == y2
 
 	if (m1 == 7)
 		return; // y0 == y1 == y2

pcsx2/GS/Renderers/SW/GSRendererSW.cpp

@@ -94,7 +94,7 @@ void GSRendererSW::VSync(u32 field, bool registers_written, bool idle_frame)
 	m_tc->IncAge();
 
 	m_draw_transfers.clear();
-	// if((m_perfmon.GetFrame() & 255) == 0) m_rl->PrintStats();
+	// if ((m_perfmon.GetFrame() & 255) == 0) m_rl->PrintStats();
 }
 
 GSTexture* GSRendererSW::GetOutput(int i, float& scale, int& y_offset)

pcsx2/GS/Renderers/SW/GSSetupPrimCodeGenerator.all.cpp

@@ -515,7 +515,7 @@ void GSSetupPrimCodeGenerator::Color()
 		pshufd(xym1, xym0, _MM_SHUFFLE(1, 0, 3, 2));
 		punpcklwd(xym0, xym1);
 
-		// if(!tme) c = c.srl16(7);
+		// if (!tme) c = c.srl16(7);
 
 		if (m_sel.tfx == TFX_NONE)
 		{

pcsx2/GS/Renderers/SW/GSSetupPrimCodeGenerator.arm64.cpp

@@ -322,7 +322,7 @@ void GSSetupPrimCodeGenerator::Color()
 		armAsm->Ext(v1.V16B(), v0.V16B(), v0.V16B(), 8);
 		armAsm->Zip1(v0.V8H(), v0.V8H(), v1.V8H());
 
-		// if(!tme) c = c.srl16(7);
+		// if (!tme) c = c.srl16(7);
 
 		if (m_sel.tfx == TFX_NONE)
 			armAsm->Ushr(v0.V8H(), v0.V8H(), 7);

pcsx2/GSDumpReplayer.cpp

@@ -379,8 +379,8 @@ void GSDumpReplayer::RenderUI()
 	do \
 	{ \
 		text_size = font->CalcTextSizeA(font->FontSize, std::numeric_limits<float>::max(), -1.0f, (text), nullptr, nullptr); \
-		dl->AddText(font, font->FontSize, ImVec2(margin + shadow_offset, position_y + shadow_offset), IM_COL32(0, 0, 0, 100), (text)); \
-		dl->AddText(font, font->FontSize, ImVec2(margin, position_y), color, (text)); \
+		dl->AddText(font, font->FontSize, ImVec2(GSConfig.OsdPerformancePos == OsdOverlayPos::TopLeft ? ImGuiManager::GetWindowWidth() - margin - text_size.x + shadow_offset : margin + shadow_offset, position_y + shadow_offset), IM_COL32(0, 0, 0, 100), (text)); \
+		dl->AddText(font, font->FontSize, ImVec2(GSConfig.OsdPerformancePos == OsdOverlayPos::TopLeft ? ImGuiManager::GetWindowWidth() - margin - text_size.x : margin, position_y), color, (text)); \
 		position_y += text_size.y + spacing; \
 	} while (0)
 

pcsx2/IPU/IPU.h

@@ -10,9 +10,9 @@
 #define ipumsk( src ) ( (src) & 0xff )
 #define ipucase( src ) case ipumsk(src)
 
-#define IPU_INT_TO( cycles )  if(!(cpuRegs.interrupt & (1<<4))) CPU_INT( DMAC_TO_IPU, cycles )
+#define IPU_INT_TO( cycles )  if (!(cpuRegs.interrupt & (1<<4))) CPU_INT( DMAC_TO_IPU, cycles )
 #define IPU_INT_FROM( cycles )  CPU_INT( DMAC_FROM_IPU, cycles )
-#define IPU_INT_PROCESS( cycles ) if(!(cpuRegs.interrupt & (1 << IPU_PROCESS))) CPU_INT( IPU_PROCESS, cycles )
+#define IPU_INT_PROCESS( cycles ) if (!(cpuRegs.interrupt & (1 << IPU_PROCESS))) CPU_INT( IPU_PROCESS, cycles )
 //
 // Bitfield Structures
 //

pcsx2/ImGui/FullscreenUI.cpp

@@ -613,7 +613,7 @@ void FullscreenUI::ApplyLayoutSettings(const SettingsInterface* bsi)
 
 	const InputLayout layout = ImGuiFullscreen::GetGamepadLayout();
 
-	if (sdl2_nintendo_mode == "true" || (sdl2_nintendo_mode == "auto") && layout == InputLayout::Nintendo)
+	if ((sdl2_nintendo_mode == "true" || sdl2_nintendo_mode == "auto") && layout == InputLayout::Nintendo)
 	{
 		// Apply
 		ImGuiManager::SwapGamepadNorthWest(true);
@@ -5738,9 +5738,12 @@ u32 FullscreenUI::PopulateSaveStateListEntries(const std::string& title, const s
 		if (InitializeSaveStateListEntry(&li, title, serial, crc, i) || !s_save_state_selector_loading)
 			s_save_state_selector_slots.push_back(std::move(li));
 
-		SaveStateListEntry bli;
-		if (InitializeSaveStateListEntry(&bli, title, serial, crc, i, true) || !s_save_state_selector_loading)
-			s_save_state_selector_slots.push_back(std::move(bli));
+		if (s_save_state_selector_loading)
+		{
+			SaveStateListEntry bli;
+			if (InitializeSaveStateListEntry(&bli, title, serial, crc, i, true))
+				s_save_state_selector_slots.push_back(std::move(bli));
+		}
 	}
 
 	return static_cast<u32>(s_save_state_selector_slots.size());

pcsx2/ImGui/ImGuiOverlays.cpp

@@ -394,6 +394,8 @@ __ri void ImGuiManager::DrawSettingsOverlay(float scale, float margin, float spa
 		APPEND("IVU ");
 	if (EmuConfig.Speedhacks.vuThread)
 		APPEND("MTVU ");
+	if (EmuConfig.GS.VsyncEnable)
+		APPEND("VSYNC ");
 
 	APPEND("EER={} EEC={} VUR={} VUC={} VQS={} ", static_cast<unsigned>(EmuConfig.Cpu.FPUFPCR.GetRoundMode()),
 		EmuConfig.Cpu.Recompiler.GetEEClampMode(), static_cast<unsigned>(EmuConfig.Cpu.VU0FPCR.GetRoundMode()),

pcsx2/Interpreter.cpp

@@ -190,7 +190,7 @@ static void execI()
 #if 0
 	static long int print_me = 0;
 	// Based on cycle
-	// if( cpuRegs.cycle > 0x4f24d714 )
+	// if ( cpuRegs.cycle > 0x4f24d714 )
 	// Or dump from a particular PC (useful to debug handler/syscall)
 	if (pc == 0x80000000)
 	{

pcsx2/IopGte.cpp

@@ -20,7 +20,7 @@
 #define G_GC(reg)
 #endif
 
-#define SUM_FLAG if(gteFLAG & 0x7F87E000) gteFLAG |= 0x80000000;
+#define SUM_FLAG if (gteFLAG & 0x7F87E000) gteFLAG |= 0x80000000;
 
 #ifdef _MSC_VER_
 #pragma warning(disable:4244)

pcsx2/Memory.cpp

@@ -356,7 +356,7 @@ void ba0W16(u32 mem, u16 value)
 				s_ba_command_executing = true;
 				s_ba_error_detected = false;
 			}
-			else if(s_ba[0x2] == 0x42) // Read Mode
+			else if (s_ba[0x2] == 0x42) // Read Mode
 			{
 				int size = (s_ba[masked_mem] & 0xF);
 
@@ -406,7 +406,7 @@ u16 ba0R16(u32 mem)
 
 void MyMemCheck(u32 mem)
 {
-    if( mem == 0x1c02f2a0 )
+    if ( mem == 0x1c02f2a0 )
         Console.WriteLn("yo; (mem == 0x1c02f2a0) in MyMemCheck...");
 }
 

pcsx2/R3000A.cpp

@@ -256,7 +256,7 @@ void iopTestIntc()
 		// Note: No need to set the iop's branch delta here, since the EE
 		// will run an IOP branch test regardless.
 	}
-	else if( !iopEventTestIsActive )
+	else if ( !iopEventTestIsActive )
 		psxSetNextBranchDelta( 2 );
 }
 

pcsx2/R5900.cpp

@@ -127,9 +127,9 @@ __ri void cpuException(u32 code, u32 bd)
 			cpuUpdateOperationMode();
 			return;
 		}
-		else if((code & 0x38000) == 0x10000)
+		else if ((code & 0x38000) == 0x10000)
 			offset = 0x80; //Performance Counter
-		else if((code & 0x38000) == 0x18000)
+		else if ((code & 0x38000) == 0x18000)
 			offset = 0x100; //Debug
 		else
 			Console.Error("Unknown Level 2 Exception!! Cause %x", code);

pcsx2/SPU2/spu2sys.cpp

@@ -1117,7 +1117,7 @@ static void RegWrite_Core(u16 value)
 
 			if (!thiscore.DmaMode && !(thiscore.Regs.STATX & 0x400))
 				thiscore.Regs.STATX &= ~0x80;
-			else if(!oldDmaMode && thiscore.DmaMode)
+			else if (!oldDmaMode && thiscore.DmaMode)
 				thiscore.Regs.STATX |= 0x80;
 
 			thiscore.ActiveTSA = thiscore.TSA;

pcsx2/ShaderCacheVersion.h

@@ -3,4 +3,4 @@
 
 /// Version number for GS and other shaders. Increment whenever any of the contents of the
 /// shaders change, to invalidate the cache.
-static constexpr u32 SHADER_CACHE_VERSION = 64;
+static constexpr u32 SHADER_CACHE_VERSION = 65;

pcsx2/USB/usb-eyetoy/jo_mpeg.cpp

@@ -164,7 +164,7 @@ static int jo_processDU(jo_bits_t *bits, float A[64], const unsigned char htdc[9
 			jo_writeBits(bits, run, 6);
 			if (AC < -127) {
 				jo_writeBits(bits, 128, 12);
-			} else if(AC > 127) {
+			} else if (AC > 127) {
 				jo_writeBits(bits, 0, 12);
 			}
 			code = AC & 0xFFF;

pcsx2/Vif_Transfer.cpp

@@ -93,7 +93,7 @@ _vifT static __fi bool vifTransfer(u32 *data, int size, bool TTE) {
 
 		if(!vifXch.qwc)
 			vifX.inprogress &= ~0x1;
-		else if(vifX.irqoffset.value != 0)
+		else if (vifX.irqoffset.value != 0)
 			vifX.irqoffset.enabled = true;
 	}
 	else

pcsx2/ps2/Iop/IopHwRead.cpp

@@ -70,12 +70,12 @@ mem8_t iopHwRead8_Page1( u32 addr )
 				DevCon.Warning( "HwRead8 from Counter16 [ignored] @ 0x%08x = 0x%02x", addr, psxHu8(addr) );
 				ret = psxHu8( addr );
 			}
-			else if( masked_addr >= 0x480 && masked_addr < 0x4a0 )
+			else if ( masked_addr >= 0x480 && masked_addr < 0x4a0 )
 			{
 				DevCon.Warning( "HwRead8 from Counter32 [ignored] @ 0x%08x = 0x%02x", addr, psxHu8(addr) );
 				ret = psxHu8( addr );
 			}
-			else if( (masked_addr >= pgmsk(HW_USB_START)) && (masked_addr < pgmsk(HW_USB_END)) )
+			else if ( (masked_addr >= pgmsk(HW_USB_START)) && (masked_addr < pgmsk(HW_USB_END)) )
 			{
 				ret = USBread8( addr );
 				PSXHW_LOG( "HwRead8 from USB @ 0x%08x = 0x%02x", addr, ret );
@@ -179,7 +179,7 @@ static __fi T _HwRead_16or32_Page1( u32 addr )
 	// ------------------------------------------------------------------------
 	// Counters, 32-bit varieties!
 	//
-	else if( masked_addr >= 0x480 && masked_addr < 0x4b0 )
+	else if ( masked_addr >= 0x480 && masked_addr < 0x4b0 )
 	{
 		int cntidx = (( masked_addr >> 4 ) & 0xf) - 5;
 		switch( masked_addr & 0xf )
@@ -218,14 +218,14 @@ static __fi T _HwRead_16or32_Page1( u32 addr )
 	// ------------------------------------------------------------------------
 	// USB, with both 16 and 32 bit interfaces
 	//
-	else if( (masked_addr >= pgmsk(HW_USB_START)) && (masked_addr < pgmsk(HW_USB_END)) )
+	else if ( (masked_addr >= pgmsk(HW_USB_START)) && (masked_addr < pgmsk(HW_USB_END)) )
 	{
 		ret = (sizeof(T) == 2) ? USBread16( addr ) : USBread32( addr );
 	}
 	// ------------------------------------------------------------------------
 	// SPU2, accessible in 16 bit mode only!
 	//
-	else if( masked_addr >= pgmsk(HW_SPU2_START) && masked_addr < pgmsk(HW_SPU2_END) )
+	else if ( masked_addr >= pgmsk(HW_SPU2_START) && masked_addr < pgmsk(HW_SPU2_END) )
 	{
 		if( sizeof(T) == 2 )
 			ret = SPU2read( addr );
@@ -238,7 +238,7 @@ static __fi T _HwRead_16or32_Page1( u32 addr )
 	// ------------------------------------------------------------------------
 	// PS1 GPU access
 	//
-	else if( (masked_addr >= pgmsk(HW_PS1_GPU_START)) && (masked_addr < pgmsk(HW_PS1_GPU_END)) )
+	else if ( (masked_addr >= pgmsk(HW_PS1_GPU_START)) && (masked_addr < pgmsk(HW_PS1_GPU_END)) )
 	{
 		// todo: psx mode: this is new
 		if( sizeof(T) == 2 )
@@ -418,7 +418,7 @@ mem32_t iopHwRead32_Page8( u32 addr )
 			ret = g_Sio2.send3[parm];
 			Sio2Log.WriteLn("%s(%08X) SIO2 SEND3 Read (%08X)", __FUNCTION__, addr, ret);
 		}
-		else if( masked_addr < 0x260 )
+		else if ( masked_addr < 0x260 )
 		{
 			// SIO2 Send commands alternate registers.  First reg maps to Send1, second
 			// to Send2, third to Send1, etc.  And the following clever code does this:
@@ -426,7 +426,7 @@ mem32_t iopHwRead32_Page8( u32 addr )
 			ret = (masked_addr & 4) ? g_Sio2.send2[parm] : g_Sio2.send1[parm];
 			Sio2Log.WriteLn("%s(%08X) SIO2 SEND1/2 Read (%08X)", __FUNCTION__, addr, ret);
 		}
-		else if( masked_addr <= 0x280 )
+		else if ( masked_addr <= 0x280 )
 		{
 			switch( masked_addr )
 			{
@@ -471,7 +471,7 @@ mem32_t iopHwRead32_Page8( u32 addr )
 					break;
 			}
 		}
-		else if( masked_addr >= pgmsk(HW_FW_START) && masked_addr <= pgmsk(HW_FW_END) )
+		else if ( masked_addr >= pgmsk(HW_FW_START) && masked_addr <= pgmsk(HW_FW_END) )
 		{
 			ret = FWread32( addr );
 		} else {

pcsx2/ps2/Iop/IopHwWrite.cpp

@@ -104,12 +104,12 @@ void iopHwWrite8_Page1( u32 addr, mem8_t val )
 				DbgCon.Warning( "HwWrite8 to Counter16 [ignored] @ addr 0x%08x = 0x%02x", addr, psxHu8(addr) );
 				psxHu8( addr ) = val;
 			}
-			else if( masked_addr >= 0x480 && masked_addr < 0x4a0 )
+			else if ( masked_addr >= 0x480 && masked_addr < 0x4a0 )
 			{
 				DbgCon.Warning( "HwWrite8 to Counter32 [ignored] @ addr 0x%08x = 0x%02x", addr, psxHu8(addr) );
 				psxHu8( addr ) = val;
 			}
-			else if( (masked_addr >= pgmsk(HW_USB_START)) && (masked_addr < pgmsk(HW_USB_END)) )
+			else if ( (masked_addr >= pgmsk(HW_USB_START)) && (masked_addr < pgmsk(HW_USB_END)) )
 			{
 				USBwrite8( addr, val );
 			}
@@ -219,7 +219,7 @@ static __fi void _HwWrite_16or32_Page1( u32 addr, T val )
 	// ------------------------------------------------------------------------
 	// Counters, 32-bit varieties!
 	//
-	else if( masked_addr >= 0x480 && masked_addr < 0x4b0 )
+	else if ( masked_addr >= 0x480 && masked_addr < 0x4b0 )
 	{
 		int cntidx = (( masked_addr >> 4 ) & 0xf) - 5;
 		switch( masked_addr & 0xf )
@@ -252,14 +252,14 @@ static __fi void _HwWrite_16or32_Page1( u32 addr, T val )
 	// ------------------------------------------------------------------------
 	// USB, with both 16 and 32 bit interfaces
 	//
-	else if( (masked_addr >= pgmsk(HW_USB_START)) && (masked_addr < pgmsk(HW_USB_END)) )
+	else if ( (masked_addr >= pgmsk(HW_USB_START)) && (masked_addr < pgmsk(HW_USB_END)) )
 	{
 		if( sizeof(T) == 2 ) USBwrite16( addr, val ); else USBwrite32( addr, val );
 	}
 	// ------------------------------------------------------------------------
 	// SPU2, accessible in 16 bit mode only!
 	//
-	else if( (masked_addr >= pgmsk(HW_SPU2_START)) && (masked_addr < pgmsk(HW_SPU2_END)) )
+	else if ( (masked_addr >= pgmsk(HW_SPU2_START)) && (masked_addr < pgmsk(HW_SPU2_END)) )
 	{
 		if( sizeof(T) == 2 )
 			SPU2write( addr, val );
@@ -272,7 +272,7 @@ static __fi void _HwWrite_16or32_Page1( u32 addr, T val )
 	// ------------------------------------------------------------------------
 	// PS1 GPU access
 	//
-	else if( (masked_addr >= pgmsk(HW_PS1_GPU_START)) && (masked_addr < pgmsk(HW_PS1_GPU_END)) )
+	else if ( (masked_addr >= pgmsk(HW_PS1_GPU_START)) && (masked_addr < pgmsk(HW_PS1_GPU_END)) )
 	{
 		// todo: psx mode: this is new
 		if( sizeof(T) == 2 )
@@ -599,7 +599,7 @@ void iopHwWrite32_Page8( u32 addr, mem32_t val )
 			const int parm = (masked_addr - 0x200) / 4;
 			g_Sio2.SetSend3(parm, val);
 		}
-		else if( masked_addr < 0x260 )
+		else if ( masked_addr < 0x260 )
 		{
 			// SIO2 Send commands alternate registers.  First reg maps to Send1, second
 			// to Send2, third to Send1, etc.  And the following clever code does this:
@@ -617,7 +617,7 @@ void iopHwWrite32_Page8( u32 addr, mem32_t val )
 				g_Sio2.send1[parm] = val;
 			}
 		}
-		else if( masked_addr <= 0x280 )
+		else if ( masked_addr <= 0x280 )
 		{
 			switch( masked_addr )
 			{
@@ -661,7 +661,7 @@ void iopHwWrite32_Page8( u32 addr, mem32_t val )
 				break;
 			}
 		}
-		else if( masked_addr >= pgmsk(HW_FW_START) && masked_addr <= pgmsk(HW_FW_END) )
+		else if ( masked_addr >= pgmsk(HW_FW_START) && masked_addr <= pgmsk(HW_FW_END) )
 		{
 			FWwrite32( addr, val );
 		}

pcsx2/ps2/Iop/IopHw_Internal.h

@@ -161,7 +161,7 @@ static __ri const char* _ioplog_GetHwName( u32 addr, T val )
 					default: return "Invalid Counter";
 				}
 			}
-			else if( addr >= 0x1f801480 && addr < 0x1f8014b0 )
+			else if ( addr >= 0x1f801480 && addr < 0x1f8014b0 )
 			{
 				switch( addr & 0xf )
 				{
@@ -174,20 +174,20 @@ static __ri const char* _ioplog_GetHwName( u32 addr, T val )
 					default: return "Invalid Counter";
 				}
 			}
-			else if( (addr >= HW_USB_START) && (addr < HW_USB_END) )
+			else if ( (addr >= HW_USB_START) && (addr < HW_USB_END) )
 			{
 				return "USB";
 			}
-			else if( (addr >= HW_SPU2_START) && (addr < HW_SPU2_END) )
+			else if ( (addr >= HW_SPU2_START) && (addr < HW_SPU2_END) )
 			{
 				return "SPU2";
 			}
-			else if( addr >= pgmsk(HW_FW_START) && addr <= pgmsk(HW_FW_END) )
+			else if ( addr >= pgmsk(HW_FW_START) && addr <= pgmsk(HW_FW_END) )
 			{
 				return "FW";
 			}
-			else if( addr >= 0x1f808200 && addr < 0x1f808240 ) { return "SIO2 param"; }
-			else if( addr >= 0x1f808240 && addr < 0x1f808260 ) { return "SIO2 send"; }
+			else if ( addr >= 0x1f808200 && addr < 0x1f808240 ) { return "SIO2 param"; }
+			else if ( addr >= 0x1f808240 && addr < 0x1f808260 ) { return "SIO2 send"; }
 
 		return NULL; //"Unknown";
 	}

pcsx2/vtlb.cpp

@@ -648,7 +648,7 @@ static OperandType vtlbUnmappedPReadSm(u32 addr) {
 	}
 	return 0;
 }
-static RETURNS_R128 vtlbUnmappedPReadLg(u32 addr) { vtlb_BusError(addr, 0); if(!CHECK_EEREC && CHECK_CACHE && CheckCache(addr)){ return readCache128(addr, false); } return r128_zero(); }
+static RETURNS_R128 vtlbUnmappedPReadLg(u32 addr) { vtlb_BusError(addr, 0); if (!CHECK_EEREC && CHECK_CACHE && CheckCache(addr)){ return readCache128(addr, false); } return r128_zero(); }
 
 template <typename OperandType>
 static void vtlbUnmappedPWriteSm(u32 addr, OperandType data) {
@@ -663,7 +663,7 @@ static void vtlbUnmappedPWriteSm(u32 addr, OperandType data) {
 		}
 	}
 }
-static void TAKES_R128 vtlbUnmappedPWriteLg(u32 addr, r128 data) { vtlb_BusError(addr, 1); if(!CHECK_EEREC && CHECK_CACHE && CheckCache(addr)) { writeCache128(addr, reinterpret_cast<mem128_t*>(&data) /*Safe??*/, false); }}
+static void TAKES_R128 vtlbUnmappedPWriteLg(u32 addr, r128 data) { vtlb_BusError(addr, 1); if (!CHECK_EEREC && CHECK_CACHE && CheckCache(addr)) { writeCache128(addr, reinterpret_cast<mem128_t*>(&data) /*Safe??*/, false); }}
 // clang-format on
 
 // --------------------------------------------------------------------------------------

pcsx2/x86/ix86-32/iR5900.cpp

@@ -1356,7 +1356,7 @@ static void iBranchTest(u32 newpc)
 	// Check the Event scheduler if our "cycle target" has been reached.
 	// Equiv code to:
 	//    cpuRegs.cycle += blockcycles;
-	//    if( cpuRegs.cycle > g_nextEventCycle ) { DoEvents(); }
+	//    if ( cpuRegs.cycle > g_nextEventCycle ) { DoEvents(); }
 
 	if (EmuConfig.Speedhacks.WaitLoop && s_nBlockFF && newpc == s_branchTo)
 	{

pcsx2/x86/microVU_Branch.inl

@@ -247,7 +247,7 @@ void mVUendProgram(mV, microFlagCycles* mFC, int isEbit)
 			xAND(ptr32[&VU0.VI[REG_VPU_STAT].UL], (isVU1 ? ~0x100 : ~0x001)); // VBS0/VBS1 flag
 		}
 	}
-	else if(isEbit)
+	else if (isEbit)
 	{
 		if (EmuConfig.Gamefixes.VUSyncHack || EmuConfig.Gamefixes.FullVU0SyncHack)
 			xMOV(ptr32[&mVU.regs().nextBlockCycles], 0);

pcsx2/x86/microVU_Lower.inl

@@ -1949,7 +1949,7 @@ mVUop(mVU_B)
 	pass2
 	{
 		if (mVUlow.badBranch)  { xMOV(ptr32[&mVU.badBranch],  branchAddr(mVU)); }
-		if (mVUlow.evilBranch) { if(isEvilBlock) xMOV(ptr32[&mVU.evilevilBranch], branchAddr(mVU)); else xMOV(ptr32[&mVU.evilBranch], branchAddr(mVU)); }
+		if (mVUlow.evilBranch) { if (isEvilBlock) xMOV(ptr32[&mVU.evilevilBranch], branchAddr(mVU)); else xMOV(ptr32[&mVU.evilBranch], branchAddr(mVU)); }
 		mVU.profiler.EmitOp(opB);
 	}
 	pass3 { mVUlog("B [<a href=\"#addr%04x\">%04x</a>]", branchAddr(mVU), branchAddr(mVU)); }