Qt: Fix Winconsistent-missing-override warning.

IopBios: remove clang deprecated carveout
we use snprintf now, so this isn't useful anymore.
2026-01-31 01:15:24 +01:00 · 2025-06-11 23:17:53 +02:00 · 2025-06-10 20:59:25 +02:00 · 2025-06-10 20:59:25 +02:00 · 2025-06-10 20:12:41 +02:00 · 2025-06-10 20:12:41 +02:00
42 changed files with 1177 additions and 301 deletions
--- a/.github/workflows/scripts/linux/flatpak/modules/24-plutovg.json
+++ b/.github/workflows/scripts/linux/flatpak/modules/24-plutovg.json
@@ -11,10 +11,9 @@
  },
  "sources": [
    {
-	  "type": "git",
+      "type": "git",
      "url": "https://github.com/sammycage/plutovg.git",
-      "tag": "v0.0.13",
-      "sha256": "5e4712cf873b0c7829a4a6157763e2ad3ac49164"
+      "tag": "v0.0.13"
    }
  ],
  "cleanup": [
--- a/.github/workflows/scripts/linux/flatpak/modules/25-plutosvg.json
+++ b/.github/workflows/scripts/linux/flatpak/modules/25-plutosvg.json
@@ -12,10 +12,9 @@
  },
  "sources": [
    {
-	  "type": "git",
+      "type": "git",
      "url": "https://github.com/sammycage/plutosvg.git",
-	  "tag": "v0.0.6",
-      "sha256": "c5388fa96feca1f1376a3d0485d5e35159452707"
+      "tag": "v0.0.6"
    }
  ],
  "cleanup": [
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -62,10 +62,12 @@ endif()

 # gsrunner
 if(ENABLE_GSRUNNER)
-	if (NOT WIN32)
-		message(WARNING "GSRunner is only supported on Windows and may not build on your system")
+	if (NOT WIN32 AND NOT APPLE)
+		message(WARNING "GSRunner is only supported on Windows and macOS and may not build on your system")
 	endif()
 	add_subdirectory(pcsx2-gsrunner)
+else()
+	add_subdirectory(pcsx2-gsrunner EXCLUDE_FROM_ALL)
 endif()

 #-------------------------------------------------------------------------------
--- a/bin/resources/GameIndex.yaml
+++ b/bin/resources/GameIndex.yaml
@@ -2557,8 +2557,6 @@ SCAJ-25012:
 SCAJ-25026:
  name: "Kunoichi Shinobi"
  region: "NTSC-Unk"
-  gsHWFixes:
-    getSkipCount: "GSC_Kunoichi"
 SCAJ-25034:
  name: "Sakura Taisen Monogatari"
  region: "NTSC-Unk"
@@ -12878,8 +12876,6 @@ SLAJ-25023:
 SLAJ-25026:
  name: "Kunoichi Shinobi"
  region: "NTSC-Unk"
-  gsHWFixes:
-    getSkipCount: "GSC_Kunoichi"
 SLAJ-25027:
  name: "Sonic Heroes"
  region: "NTSC-Unk"
@@ -12894,8 +12890,6 @@ SLAJ-25030:
 SLAJ-25031:
  name: "Kunoichi - Shinobu"
  region: "NTSC-C-J"
-  gsHWFixes:
-    getSkipCount: "GSC_Kunoichi"
 SLAJ-25033:
  name: "Puyo Puyo Fever"
  region: "NTSC-Unk"
@@ -12927,7 +12921,7 @@ SLAJ-25041:
  name: "Harry Potter and the Prisoner of Azkaban"
  region: "NTSC-C"
  gsHWFixes:
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Aligns mirror reflections, removes some bloom-related ghosting.
    recommendedBlendingLevel: 4 # Improves lighting effects.
 SLAJ-25042:
@@ -16829,7 +16823,7 @@ SLES-51192:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51193:
@@ -16839,7 +16833,7 @@ SLES-51193:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51194:
@@ -16849,7 +16843,7 @@ SLES-51194:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51195:
@@ -16859,7 +16853,7 @@ SLES-51195:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51196:
@@ -16869,7 +16863,7 @@ SLES-51196:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51197:
@@ -16926,7 +16920,7 @@ SLES-51214:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51215:
@@ -16936,7 +16930,7 @@ SLES-51215:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51216:
@@ -16946,7 +16940,7 @@ SLES-51216:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51217:
@@ -16956,7 +16950,7 @@ SLES-51217:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51218:
@@ -16966,7 +16960,7 @@ SLES-51218:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51219:
@@ -16976,7 +16970,7 @@ SLES-51219:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
    texturePreloading: 0 # Performs better on certain areas like mid-game in and around the castle as the hash goes easily to 200 MB.
 SLES-51220:
@@ -19388,8 +19382,6 @@ SLES-52238:
  name: "Nightshade"
  region: "PAL-M5"
  compat: 5
-  gsHWFixes:
-    getSkipCount: "GSC_Kunoichi"
 SLES-52240:
  name: "International Pool Championship"
  region: "PAL-M5"
@@ -19896,7 +19888,7 @@ SLES-52440:
  region: "PAL-M7"
  compat: 3
  gsHWFixes:
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Aligns mirror reflections, removes some bloom-related ghosting.
    recommendedBlendingLevel: 4 # Improves lighting effects.
 SLES-52444:
@@ -20123,7 +20115,7 @@ SLES-52527:
  name: "Harry Potter og Fangen fra Azkaban"
  region: "PAL-M4"
  gsHWFixes:
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Aligns mirror reflections, removes some bloom-related ghosting.
    recommendedBlendingLevel: 4 # Improves lighting effects.
 SLES-52531:
@@ -20434,7 +20426,7 @@ SLES-52600:
  name: "Harry Potter and the Prisoner of Azkaban"
  region: "PAL-PL"
  gsHWFixes:
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Aligns mirror reflections, removes some bloom-related ghosting.
    recommendedBlendingLevel: 4 # Improves lighting effects.
 SLES-52601:
@@ -22086,6 +22078,8 @@ SLES-53099:
 SLES-53100:
  name: "Scooby-Doo! Unmasked"
  region: "PAL-M4"
+  gsHWFixes:
+    nativeScaling: 2 # Fixes post processing position.
 SLES-53104:
  name: "Tom Clancy's Rainbow Six - Lockdown"
  region: "PAL-M5"
@@ -30956,8 +30950,6 @@ SLKA-25135:
  name: "Kunoichi"
  region: "NTSC-K"
  compat: 5
-  gsHWFixes:
-    getSkipCount: "GSC_Kunoichi"
 SLKA-25136:
  name: "Need for Speed - Underground"
  region: "NTSC-K"
@@ -31108,7 +31100,7 @@ SLKA-25172:
  name: "Harry Potter and the Prisoner of Azkaban"
  region: "NTSC-K"
  gsHWFixes:
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Aligns mirror reflections, removes some bloom-related ghosting.
    recommendedBlendingLevel: 4 # Improves lighting effects.
 SLKA-25173:
@@ -32522,8 +32514,6 @@ SLKA-35004:
  name: "Sakura Wars 5 So Long My Love"
  region: "NTSC-K"
  compat: 5
-  gsHWFixes:
-    getSkipCount: "GSC_SakuraWarsSoLongMyLove"
 SLKA-35005:
  name: "Jin Samguk Mussang 5 - Special"
  region: "NTSC-K"
@@ -35217,8 +35207,6 @@ SLPM-61059:
  name-sort: "くのいち -しのび- [たいけんばん]"
  name-en: "Kunoichi -Shinobi- [Trial]"
  region: "NTSC-J"
-  gsHWFixes:
-    getSkipCount: "GSC_Kunoichi"
 SLPM-61060:
  name: "BUSIN 0 ～Wizardry Alternative NEO～ [体験版]"
  name-sort: "ぶしん ぜろ ～うぃざーどりぃ おるたなてぃぶ ねお～ [たいけんばん]"
@@ -36957,7 +36945,7 @@ SLPM-62241:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
 SLPM-62242:
  name: "撞球 ビリヤードマスター2"
@@ -38369,7 +38357,7 @@ SLPM-62513:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
 SLPM-62514:
  name: "シムピープル ～お茶の間劇場～ [EA BEST HITS]"
@@ -39800,7 +39788,7 @@ SLPM-64528:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
 SLPM-64532:
  name: "Digital Holmes"
@@ -42473,8 +42461,6 @@ SLPM-65447:
  name-en: "Kunoichi -Shinobi-"
  region: "NTSC-J"
  compat: 5
-  gsHWFixes:
-    getSkipCount: "GSC_Kunoichi"
 SLPM-65448:
  name: "カンブリアンQTS ～化石になっても～"
  name-sort: "かんぶりあんQTS ～かせきになっても～"
@@ -43355,7 +43341,7 @@ SLPM-65612:
  name-en: "Harry Potter and the Prisoner of Azkaban"
  region: "NTSC-J"
  gsHWFixes:
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Aligns mirror reflections, removes some bloom-related ghosting.
    recommendedBlendingLevel: 4 # Improves lighting effects.
 SLPM-65613:
@@ -52157,7 +52143,7 @@ SLPM-68005:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
 SLPM-68007:
  name: "カラオケレボリューション用マイク同梱お試しディスク"
@@ -54308,7 +54294,7 @@ SLPS-20234:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
 SLPS-20237:
  name: "Saikyou Toudai Shougi 3"
@@ -59136,6 +59122,8 @@ SLPS-25581:
  name-sort: "しんぷる2000しりーず Vol.  92 THE のろいのげーむ"
  name-en: "Simple 2000 Series Vol. 92 - The Game of a Curse"
  region: "NTSC-J"
+  gsHWFixes:
+    cpuCLUTRender: 1 # Fixes corrupted and missing graphics.
 SLPS-25582:
  name: "アストロ球団 決戦！！ビクトリー球団編"
  name-sort: "あすとろきゅうだん けっせん！！びくとりーきゅうだんへん"
@@ -65273,7 +65261,7 @@ SLUS-20576:
    - EETimingHack # Fixes flickering textures.
  gsHWFixes:
    recommendedBlendingLevel: 4 # Fixes missing lighting effects.
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Fixes misaligned lighting and other effects.
 SLUS-20577:
  name: "Drome Racers"
@@ -66538,8 +66526,6 @@ SLUS-20810:
  name: "Nightshade"
  region: "NTSC-U"
  compat: 5
-  gsHWFixes:
-    getSkipCount: "GSC_Kunoichi"
 SLUS-20811:
  name: "Need for Speed - Underground"
  region: "NTSC-U"
@@ -67218,7 +67204,7 @@ SLUS-20926:
  name: "Harry Potter and the Prisoner of Azkaban"
  region: "NTSC-U"
  gsHWFixes:
-    cpuFramebufferConversion: 1 # Fixes right side of the screen from garbage textures.
+    textureInsideRT: 1 # Fixes right side shadow lookup.
    halfPixelOffset: 2 # Aligns mirror reflections, removes some bloom-related ghosting.
    recommendedBlendingLevel: 4 # Improves lighting effects.
 SLUS-20927:
@@ -68231,6 +68217,8 @@ SLUS-21091:
  name: "Scooby-Doo! Unmasked"
  region: "NTSC-U"
  compat: 5
+  gsHWFixes:
+    nativeScaling: 2 # Fixes post processing position.
 SLUS-21093:
  name: "Worms Forts - Under Siege"
  region: "NTSC-U"
@@ -73133,8 +73121,6 @@ SLUS-21927:
  name: "Sakura Wars - So Long, My Love [English - Disc 1]"
  region: "NTSC-U"
  compat: 5
-  gsHWFixes:
-    getSkipCount: "GSC_SakuraWarsSoLongMyLove"
 SLUS-21928:
  name: "Scooby-Doo! and the Spooky Swamp"
  region: "NTSC-U"
@@ -73151,8 +73137,6 @@ SLUS-21930:
  name: "Sakura Wars - So Long, My Love [Japanese - Disc 2]"
  region: "NTSC-U"
  compat: 5
-  gsHWFixes:
-    getSkipCount: "GSC_SakuraWarsSoLongMyLove"
 SLUS-21931:
  name: "Disney/Pixar Toy Story 3"
  region: "NTSC-U"
--- a/bin/resources/game_controller_db.txt
+++ b/bin/resources/game_controller_db.txt
@@ -1000,6 +1000,7 @@ xinput,XInput Controller,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,
 03000000d62000007162000001000000,Moga Pro 2,a:b0,b:b1,back:b9,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,leftshoulder:b4,leftstick:b7,lefttrigger:a5,leftx:a0,lefty:a1,rightshoulder:b5,rightstick:b8,righttrigger:a4,rightx:a2,righty:a3,start:b6,x:b2,y:b3,platform:Mac OS X,
 03000000c62400002a89000000010000,MOGA XP5A Plus,a:b0,b:b1,back:b10,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b21,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:Mac OS X,
 03000000c62400002b89000000010000,MOGA XP5A Plus,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:Mac OS X,
+03000000853200008906000000010000,Nacon Revolution X Unlimited,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:Mac OS X,
 03000000632500007505000000020000,NeoGeo mini PAD Controller,a:b1,b:b0,back:b8,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,start:b9,x:b2,y:b3,platform:Mac OS X,
 03000000921200004b46000003020000,NES 2-port Adapter,a:b1,b:b0,back:b10,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,start:b11,platform:Mac OS X,
 030000001008000001e5000006010000,NEXT SNES Controller,a:b2,b:b1,back:b8,dpdown:+a1,dpleft:-a0,dpright:+a0,dpup:-a1,leftshoulder:b4,rightshoulder:b5,righttrigger:b6,start:b9,x:b3,y:b0,platform:Mac OS X,
@@ -1607,6 +1608,7 @@ xinput,XInput Controller,a:b0,b:b1,back:b6,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,
 030000009b2800003200000001010000,Raphnet GC and N64 Adapter,a:b0,b:b7,dpdown:b11,dpleft:b12,dpright:b13,dpup:b10,lefttrigger:b4,leftx:a0,lefty:a1,rightshoulder:b2,righttrigger:b5,rightx:a3,righty:a4,start:b3,x:b1,y:b8,platform:Linux,
 030000009b2800006000000001010000,Raphnet GC and N64 Adapter,a:b0,b:b7,dpdown:b11,dpleft:b12,dpright:b13,dpup:b10,lefttrigger:b4,leftx:a0,lefty:a1,rightshoulder:b2,righttrigger:b5,rightx:a3,righty:a4,start:b3,x:b1,y:b8,platform:Linux,
 030000009b2800006100000001010000,Raphnet N64 Adapter,+rightx:b9,+righty:b7,-rightx:b8,-righty:b6,a:b0,b:b1,dpdown:b11,dpleft:b12,dpright:b13,dpup:b10,leftshoulder:b4,lefttrigger:b2,leftx:a0,lefty:a1,rightshoulder:b5,start:b3,platform:Linux,
+030000009b2800006400000001010000,Raphnet N64 Adapter,+rightx:b9,+righty:b7,-rightx:b8,-righty:b6,a:b0,b:b1,dpdown:b11,dpleft:b12,dpright:b13,dpup:b10,leftshoulder:b4,lefttrigger:b2,leftx:a0,lefty:a1,rightshoulder:b5,start:b3,platform:Linux,
 030000009b2800008000000020020000,Raphnet Wii Classic Adapter,a:b1,b:b4,back:b2,dpdown:b13,dpleft:b14,dpright:b15,dpup:b12,leftshoulder:b6,rightshoulder:b7,start:b3,x:b0,y:b5,platform:Linux,
 030000009b2800008000000001010000,Raphnet Wii Classic Adapter V3,a:b1,b:b4,back:b2,dpdown:b13,dpleft:b14,dpright:b15,dpup:b12,leftshoulder:b6,rightshoulder:b7,start:b3,x:b0,y:b5,platform:Linux,
 03000000f8270000bf0b000011010000,Razer Kishi,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,
--- a/bin/resources/shaders/dx11/convert.fx
+++ b/bin/resources/shaders/dx11/convert.fx
@@ -313,6 +313,127 @@ float ps_convert_rgb5a1_float16_biln(PS_INPUT input) : SV_Depth
 	SAMPLE_RGBA_DEPTH_BILN(rgb5a1_to_depth16);
 }

+PS_OUTPUT ps_convert_rgb5a1_8i(PS_INPUT input)
+{
+	PS_OUTPUT output;
+
+	// Convert a RGB5A1 texture into a 8 bits packed texture
+	// Input column: 16x2 RGB5A1 pixels
+	// 0: 16 RGBA
+	// 1: 16 RGBA
+	// Output column: 16x4 Index pixels
+	// 0: 16 R5G2
+	// 1: 16 R5G2
+	// 2: 16 G2B5A1
+	// 3: 16 G2B5A1
+	uint2 pos = uint2(input.p.xy);
+
+	// Collapse separate R G B A areas into their base pixel
+	uint2 column = (pos & ~uint2(0u, 3u)) / uint2(1,2);
+	uint2 subcolumn = (pos & uint2(0u, 1u));
+	column.x -= (column.x / 128) * 64;
+	column.y += (column.y / 32) * 32;
+	
+	uint PSM = uint(DOFFSET);
+	
+	// Deal with swizzling differences
+	if ((PSM & 0x8) != 0) // PSMCT16S
+	{
+		if ((pos.x & 32) != 0)
+		{
+			column.y += 32; // 4 columns high times 4 to get bottom 4 blocks
+			column.x &= ~32;
+		}
+		
+		if ((pos.x & 64) != 0)
+		{
+			column.x -= 32;
+		}
+		
+		if (((pos.x & 16) != 0) != ((pos.y & 16) != 0))
+		{
+			column.x ^= 16; 
+			column.y ^= 8;
+		}
+		
+		if ((PSM & 0x30) != 0) // PSMZ16S - Untested but hopefully ok if anything uses it.
+		{
+			column.x ^= 32;
+			column.y ^= 16;
+		}
+	}
+	else // PSMCT16
+	{
+		if ((pos.y & 32) != 0)
+		{
+			column.y -= 16;
+			column.x += 32;
+		}
+		
+		if ((pos.x & 96) != 0)
+		{
+			uint multi = (pos.x & 96) / 32;
+			column.y += 16 * multi; // 4 columns high times 4 to get bottom 4 blocks
+			column.x -= (pos.x & 96);
+		}
+		
+		if (((pos.x & 16) != 0) != ((pos.y & 16) != 0))
+		{
+			column.x ^= 16; 
+			column.y ^= 8;
+		}
+		
+		if ((PSM & 0x30) != 0) // PSMZ16 - Untested but hopefully ok if anything uses it.
+		{
+			column.x ^= 32;
+			column.y ^= 32;
+		}
+	}
+	
+	uint2 coord = column | subcolumn;
+
+	// Compensate for potentially differing page pitch.
+	uint SBW = uint(EMODA);
+	uint DBW = uint(EMODC);
+	uint2 block_xy = coord / uint2(64,64);
+	uint block_num = (block_xy.y * (DBW / 128)) + block_xy.x;
+	uint2 block_offset = uint2((block_num % (SBW / 64)) * 64, (block_num / (SBW / 64)) * 64);
+	coord = (coord % uint2(64, 64)) + block_offset;
+
+	// Apply offset to cols 1 and 2
+	uint is_col23 = pos.y & 4u;
+	uint is_col13 = pos.y & 2u;
+	uint is_col12 = is_col23 ^ (is_col13 << 1);
+	coord.x ^= is_col12; // If cols 1 or 2, flip bit 3 of x
+
+	float ScaleFactor = BGColor.x;
+	if (floor(ScaleFactor) != ScaleFactor)
+		coord = uint2(float2(coord) * ScaleFactor);
+	else
+		coord *= uint(ScaleFactor);
+
+	float4 pixel = Texture.Load(int3(int2(coord), 0));
+	uint4 denorm_c = (uint4)(pixel * 255.5f);
+	if ((pos.y & 2u) == 0u)
+	{
+		uint red = (denorm_c.r >> 3) & 0x1Fu;
+		uint green = (denorm_c.g >> 3) & 0x1Fu;
+		float sel0 = (float)(((green << 5) | red) & 0xFF) / 255.0f;
+		
+		output.c = (float4)(sel0);
+	}
+	else
+	{
+		uint green = (denorm_c.g >> 3) & 0x1Fu;
+		uint blue = (denorm_c.b >> 3) & 0x1Fu;
+		uint alpha = denorm_c.a & 0x80u;
+		float sel0 = (float)((alpha | (blue << 2) | (green >> 3)) & 0xFF) / 255.0f;
+
+		output.c = (float4)(sel0);
+	}
+	return output;
+}
+
 PS_OUTPUT ps_convert_rgba_8i(PS_INPUT input)
 {
 	PS_OUTPUT output;
--- a/bin/resources/shaders/opengl/convert.glsl
+++ b/bin/resources/shaders/opengl/convert.glsl
@@ -237,9 +237,131 @@ void ps_convert_rgb5a1_float16_biln()
 }
 #endif

+#ifdef ps_convert_rgb5a1_8i
+uniform uint SBW;
+uniform uint DBW;
+uniform uint PSM;
+uniform float ScaleFactor;
+
+void ps_convert_rgb5a1_8i()
+{
+	// Convert a RGB5A1 texture into a 8 bits packed texture
+	// Input column: 16x2 RGB5A1 pixels
+	// 0: 16 RGBA
+	// 1: 16 RGBA
+	// Output column: 16x4 Index pixels
+	// 0: 16 R5G2
+	// 1: 16 R5G2
+	// 2: 16 G2B5A1
+	// 3: 16 G2B5A1
+	
+	uvec2 pos = uvec2(gl_FragCoord.xy);
+
+	// Collapse separate R G B A areas into their base pixel
+	uvec2 column = (pos & ~uvec2(0u, 3u)) / uvec2(1,2);
+	uvec2 subcolumn = (pos & uvec2(0u, 1u));
+	column.x -= (column.x / 128) * 64;
+	column.y += (column.y / 32) * 32;
+	
+	// Deal with swizzling differences
+	if ((PSM & 0x8) != 0) // PSMCT16S
+	{
+		if ((pos.x & 32) != 0)
+		{
+			column.y += 32; // 4 columns high times 4 to get bottom 4 blocks
+			column.x &= ~32;
+		}
+		
+		if ((pos.x & 64) != 0)
+		{
+			column.x -= 32;
+		}
+		
+		if (((pos.x & 16) != 0) != ((pos.y & 16) != 0))
+		{
+			column.x ^= 16; 
+			column.y ^= 8;
+		}
+		
+		if ((PSM & 0x30) != 0) // PSMZ16S - Untested but hopefully ok if anything uses it.
+		{
+			column.x ^= 32;
+			column.y ^= 16;
+		}
+	}
+	else // PSMCT16
+	{
+		if ((pos.y & 32) != 0)
+		{
+			column.y -= 16;
+			column.x += 32;
+		}
+		
+		if ((pos.x & 96) != 0)
+		{
+			uint multi = (pos.x & 96) / 32;
+			column.y += 16 * multi; // 4 columns high times 4 to get bottom 4 blocks
+			column.x -= (pos.x & 96);
+		}
+		
+		if (((pos.x & 16) != 0) != ((pos.y & 16) != 0))
+		{
+			column.x ^= 16; 
+			column.y ^= 8;
+		}
+		
+		if ((PSM & 0x30) != 0) // PSMZ16 - Untested but hopefully ok if anything uses it.
+		{
+			column.x ^= 32;
+			column.y ^= 32;
+		}
+	}
+	uvec2 coord = column | subcolumn;
+	
+	// Compensate for potentially differing page pitch.
+	uvec2 block_xy = coord / uvec2(64u, 64u);
+	uint block_num = (block_xy.y * (DBW / 128u)) + block_xy.x;
+	uvec2 block_offset = uvec2((block_num % (SBW / 64u)) * 64u, (block_num / (SBW / 64u)) * 64u);
+	coord = (coord % uvec2(64u, 64u)) + block_offset;
+
+	// Apply offset to cols 1 and 2
+	uint is_col23 = pos.y & 4u;
+	uint is_col13 = pos.y & 2u;
+	uint is_col12 = is_col23 ^ (is_col13 << 1);
+	coord.x ^= is_col12; // If cols 1 or 2, flip bit 3 of x
+
+	if (floor(ScaleFactor) != ScaleFactor)
+		coord = uvec2(vec2(coord) * ScaleFactor);
+	else
+		coord *= uvec2(ScaleFactor);
+
+	vec4 pixel = texelFetch(TextureSampler, ivec2(coord), 0);
+	
+	uvec4 denorm_c = uvec4(pixel * 255.5f);
+	if ((pos.y & 2u) == 0u)
+	{
+		uint red = (denorm_c.r >> 3) & 0x1Fu;
+		uint green = (denorm_c.g >> 3) & 0x1Fu;
+		float sel0 = float(((green << 5) | red) & 0xFF) / 255.0f;
+		
+		SV_Target0 = vec4(sel0);
+	}
+	else
+	{
+		uint green = (denorm_c.g >> 3) & 0x1Fu;
+		uint blue = (denorm_c.b >> 3) & 0x1Fu;
+		uint alpha = denorm_c.a & 0x80u;
+		float sel0 = float((alpha | (blue << 2) | (green >> 3)) & 0xFF) / 255.0f;
+
+		SV_Target0 = vec4(sel0);
+	}
+}
+#endif
+
 #ifdef ps_convert_rgba_8i
 uniform uint SBW;
 uniform uint DBW;
+uniform uint PSM;
 uniform float ScaleFactor;

 void ps_convert_rgba_8i()
--- a/bin/resources/shaders/vulkan/convert.glsl
+++ b/bin/resources/shaders/vulkan/convert.glsl
@@ -307,12 +307,138 @@ void ps_convert_rgb5a1_float16_biln()
 }
 #endif

+#ifdef ps_convert_rgb5a1_8i
+layout(push_constant) uniform cb10
+{
+	uint SBW;
+	uint DBW;
+	uint PSM;
+	float cb_pad1;
+	float ScaleFactor;
+	vec3 cb_pad2;
+};
+
+void ps_convert_rgb5a1_8i()
+{
+	// Convert a RGB5A1 texture into a 8 bits packed texture
+	// Input column: 16x2 RGB5A1 pixels
+	// 0: 16 RGBA
+	// 1: 16 RGBA
+	// Output column: 16x4 Index pixels
+	// 0: 16 R5G2
+	// 1: 16 R5G2
+	// 2: 16 G2B5A1
+	// 3: 16 G2B5A1
+	
+	uvec2 pos = uvec2(gl_FragCoord.xy);
+
+	// Collapse separate R G B A areas into their base pixel
+	uvec2 column = (pos & ~uvec2(0u, 3u)) / uvec2(1,2);
+	uvec2 subcolumn = (pos & uvec2(0u, 1u));
+	column.x -= (column.x / 128) * 64;
+	column.y += (column.y / 32) * 32;
+	
+	// Deal with swizzling differences
+	if ((PSM & 0x8) != 0) // PSMCT16S
+	{
+		if ((pos.x & 32) != 0)
+		{
+			column.y += 32; // 4 columns high times 4 to get bottom 4 blocks
+			column.x &= ~32;
+		}
+		
+		if ((pos.x & 64) != 0)
+		{
+			column.x -= 32;
+		}
+		
+		if (((pos.x & 16) != 0) != ((pos.y & 16) != 0))
+		{
+			column.x ^= 16; 
+			column.y ^= 8;
+		}
+		
+		if ((PSM & 0x30) != 0) // PSMZ16S - Untested but hopefully ok if anything uses it.
+		{
+			column.x ^= 32;
+			column.y ^= 16;
+		}
+	}
+	else // PSMCT16
+	{
+		if ((pos.y & 32) != 0)
+		{
+			column.y -= 16;
+			column.x += 32;
+		}
+		
+		if ((pos.x & 96) != 0)
+		{
+			uint multi = (pos.x & 96) / 32;
+			column.y += 16 * multi; // 4 columns high times 4 to get bottom 4 blocks
+			column.x -= (pos.x & 96);
+		}
+		
+		if (((pos.x & 16) != 0) != ((pos.y & 16) != 0))
+		{
+			column.x ^= 16; 
+			column.y ^= 8;
+		}
+		
+		if ((PSM & 0x30) != 0) // PSMZ16 - Untested but hopefully ok if anything uses it.
+		{
+			column.x ^= 32;
+			column.y ^= 32;
+		}
+	}
+	uvec2 coord = column | subcolumn;
+	
+	// Compensate for potentially differing page pitch.
+	uvec2 block_xy = coord / uvec2(64u, 64u);
+	uint block_num = (block_xy.y * (DBW / 128u)) + block_xy.x;
+	uvec2 block_offset = uvec2((block_num % (SBW / 64u)) * 64u, (block_num / (SBW / 64u)) * 64u);
+	coord = (coord % uvec2(64u, 64u)) + block_offset;
+
+	// Apply offset to cols 1 and 2
+	uint is_col23 = pos.y & 4u;
+	uint is_col13 = pos.y & 2u;
+	uint is_col12 = is_col23 ^ (is_col13 << 1);
+	coord.x ^= is_col12; // If cols 1 or 2, flip bit 3 of x
+
+	if (floor(ScaleFactor) != ScaleFactor)
+		coord = uvec2(vec2(coord) * ScaleFactor);
+	else
+		coord *= uvec2(ScaleFactor);
+
+	vec4 pixel = texelFetch(samp0, ivec2(coord), 0);
+	uvec4 denorm_c = uvec4(pixel * 255.5f);
+	if ((pos.y & 2u) == 0u)
+	{
+		uint red = (denorm_c.r >> 3) & 0x1Fu;
+		uint green = (denorm_c.g >> 3) & 0x1Fu;
+		float sel0 = float(((green << 5) | red) & 0xFF) / 255.0f;
+		
+		o_col0 = vec4(sel0);
+	}
+	else
+	{
+		uint green = (denorm_c.g >> 3) & 0x1Fu;
+		uint blue = (denorm_c.b >> 3) & 0x1Fu;
+		uint alpha = denorm_c.a & 0x80u;
+		float sel0 = float((alpha | (blue << 2) | (green >> 3)) & 0xFF) / 255.0f;
+
+		o_col0 = vec4(sel0);
+	}
+}
+#endif
+
 #ifdef ps_convert_rgba_8i
 layout(push_constant) uniform cb10
 {
 	uint SBW;
 	uint DBW;
-	uvec2 cb_pad1;
+	uint PSM;
+	float cb_pad1;
 	float ScaleFactor;
 	vec3 cb_pad2;
 };
--- a/cmake/BuildParameters.cmake
+++ b/cmake/BuildParameters.cmake
@@ -7,7 +7,7 @@ include(GNUInstallDirs)
 # Misc option
 #-------------------------------------------------------------------------------
 option(ENABLE_TESTS "Enables building the unit tests" ON)
-option(ENABLE_GSRUNNER "Enables building the GSRunner" OFF)
+option(ENABLE_GSRUNNER "Enables building the GSRunner by default.  It can still be built with `make pcsx2-gsrunner` otherwise." OFF)
 option(LTO_PCSX2_CORE "Enable LTO/IPO/LTCG on the subset of pcsx2 that benefits most from it but not anything else")
 option(USE_VTUNE "Plug VTUNE to profile GS JIT.")
 option(PACKAGE_MODE "Use this option to ease packaging of PCSX2 (developer/distribution option)")
--- a/cmake/SearchForStuff.cmake
+++ b/cmake/SearchForStuff.cmake
@@ -120,7 +120,7 @@ add_subdirectory(3rdparty/demangler EXCLUDE_FROM_ALL)
 add_subdirectory(3rdparty/ccc EXCLUDE_FROM_ALL)

 # The docking system for the debugger.
-find_package(KDDockWidgets-qt6 REQUIRED)
+find_package(KDDockWidgets-qt6 2.0.0 REQUIRED)
 # Add an extra include path to work around a broken include directive.
 # TODO: Remove this the next time we update KDDockWidgets.
 get_target_property(KDDOCKWIDGETS_INCLUDE_DIRECTORY KDAB::kddockwidgets INTERFACE_INCLUDE_DIRECTORIES)
--- a/common/CocoaTools.h
+++ b/common/CocoaTools.h
@@ -31,6 +31,19 @@ namespace CocoaTools
 	bool DelayedLaunch(std::string_view file);
 	/// Open a Finder window to the given URL
 	bool ShowInFinder(std::string_view file);
+	/// Get the path to the resources directory of the current application
+	std::optional<std::string> GetResourcePath();
+
+	/// Create a window
+	void* CreateWindow(std::string_view title, uint32_t width, uint32_t height);
+	/// Destroy a window
+	void DestroyWindow(void* window);
+	/// Make a WindowInfo from the given window
+	void GetWindowInfoFromWindow(WindowInfo* wi, void* window);
+	/// Run cocoa event loop
+	void RunCocoaEventLoop(bool wait_forever = false);
+	/// Posts an event to the main telling `RunCocoaEventLoop(true)` to exit
+	void StopMainThreadEventLoop();
 }

 #endif // __APPLE__
--- a/common/CocoaTools.mm
+++ b/common/CocoaTools.mm
@@ -224,5 +224,103 @@ bool CocoaTools::DelayedLaunch(std::string_view file)
 bool CocoaTools::ShowInFinder(std::string_view file)
 {
 	return [[NSWorkspace sharedWorkspace] selectFile:NSStringFromStringView(file)
-	                        inFileViewerRootedAtPath:nil];
+	                        inFileViewerRootedAtPath:@""];
 }
+
+std::optional<std::string> CocoaTools::GetResourcePath()
+{ @autoreleasepool {
+	if (NSBundle* bundle = [NSBundle mainBundle])
+	{
+		NSString* rsrc = [bundle resourcePath];
+		NSString* root = [bundle bundlePath];
+		if ([rsrc isEqualToString:root])
+			rsrc = [rsrc stringByAppendingString:@"/resources"];
+		return [rsrc UTF8String];
+	}
+	return std::nullopt;
+}}
+
+// MARK: - GSRunner
+
+void* CocoaTools::CreateWindow(std::string_view title, u32 width, u32 height)
+{
+	if (!NSApp)
+	{
+		[NSApplication sharedApplication];
+		[NSApp setActivationPolicy:NSApplicationActivationPolicyRegular];
+		[NSApp finishLaunching];
+	}
+	constexpr NSWindowStyleMask style = NSWindowStyleMaskTitled | NSWindowStyleMaskClosable | NSWindowStyleMaskMiniaturizable | NSWindowStyleMaskResizable;
+	NSScreen* mainScreen = [NSScreen mainScreen];
+	// Center the window on the screen, because why not
+	NSRect screenFrame = [mainScreen frame];
+	NSRect viewFrame = screenFrame;
+	viewFrame.size = NSMakeSize(width, height);
+	viewFrame.origin.x += (screenFrame.size.width - viewFrame.size.width) / 2;
+	viewFrame.origin.y += (screenFrame.size.height - viewFrame.size.height) / 2;
+	NSWindow* window = [[NSWindow alloc]
+		initWithContentRect:viewFrame
+		          styleMask:style
+		            backing:NSBackingStoreBuffered
+		              defer:NO];
+	[window setTitle:NSStringFromStringView(title)];
+	[window makeKeyAndOrderFront:window];
+	return (__bridge_retained void*)window;
+}
+
+void CocoaTools::DestroyWindow(void* window)
+{
+	(void)(__bridge_transfer NSWindow*)window;
+}
+
+void CocoaTools::GetWindowInfoFromWindow(WindowInfo* wi, void* cf_window)
+{
+	if (cf_window)
+	{
+		NSWindow* window = (__bridge NSWindow*)cf_window;
+		float scale = [window backingScaleFactor];
+		NSView* view = [window contentView];
+		NSRect dims = [view frame];
+		wi->type = WindowInfo::Type::MacOS;
+		wi->window_handle = (__bridge void*)view;
+		wi->surface_width = dims.size.width * scale;
+		wi->surface_height = dims.size.height * scale;
+		wi->surface_scale = scale;
+	}
+	else
+	{
+		wi->type = WindowInfo::Type::Surfaceless;
+	}
+}
+
+static constexpr short STOP_EVENT_LOOP = 0x100;
+
+void CocoaTools::RunCocoaEventLoop(bool forever)
+{
+	NSDate* end = forever ? [NSDate distantFuture] : [NSDate distantPast];
+	[NSApplication sharedApplication]; // Ensure NSApp is initialized
+	while (true)
+	{ @autoreleasepool {
+		NSEvent* ev = [NSApp nextEventMatchingMask:NSEventMaskAny
+		                                 untilDate:end
+		                                    inMode:NSDefaultRunLoopMode
+		                                   dequeue:YES];
+		if (!ev || ([ev type] == NSEventTypeApplicationDefined && [ev subtype] == STOP_EVENT_LOOP))
+			break;
+		[NSApp sendEvent:ev];
+	}}
+}
+
+void CocoaTools::StopMainThreadEventLoop()
+{ @autoreleasepool {
+	NSEvent* ev = [NSEvent otherEventWithType:NSEventTypeApplicationDefined
+	                                 location:{}
+	                            modifierFlags:0
+	                                timestamp:0
+	                             windowNumber:0
+	                                  context:nil
+	                                  subtype:STOP_EVENT_LOOP
+	                                    data1:0
+	                                    data2:0];
+	[NSApp postEvent:ev atStart:NO];
+}}
--- a/pcsx2-gsrunner/Main.cpp
+++ b/pcsx2-gsrunner/Main.cpp
@@ -16,6 +16,7 @@
 #include "fmt/format.h"

 #include "common/Assertions.h"
+#include "common/CocoaTools.h"
 #include "common/Console.h"
 #include "common/CrashHandler.h"
 #include "common/FileSystem.h"
@@ -56,7 +57,8 @@ namespace GSRunner
 	static bool CreatePlatformWindow();
 	static void DestroyPlatformWindow();
 	static std::optional<WindowInfo> GetPlatformWindowInfo();
-	static void PumpPlatformMessages();
+	static void PumpPlatformMessages(bool forever = false);
+	static void StopPlatformMessagePump();
 } // namespace GSRunner

 static constexpr u32 WINDOW_WIDTH = 640;
@@ -532,7 +534,7 @@ bool GSRunner::ParseCommandLineArgs(int argc, char* argv[], VMBootParameters& pa
 			{
 				std::string str(argv[++i]);

-				s_settings_interface.SetBoolValue("EmuCore/GS", "dump", true);
+				s_settings_interface.SetBoolValue("EmuCore/GS", "DumpGSData", true);

 				if (str.find("rt") != std::string::npos)
 					s_settings_interface.SetBoolValue("EmuCore/GS", "SaveRT", true);
@@ -764,7 +766,7 @@ bool GSRunner::ParseCommandLineArgs(int argc, char* argv[], VMBootParameters& pa
 		return false;
 	}

-	if (s_settings_interface.GetBoolValue("EmuCore/GS", "dump") && !dumpdir.empty())
+	if (s_settings_interface.GetBoolValue("EmuCore/GS", "DumpGSData") && !dumpdir.empty())
 	{
 		if (s_settings_interface.GetStringValue("EmuCore/GS", "HWDumpDirectory").empty())
 			s_settings_interface.SetStringValue("EmuCore/GS", "HWDumpDirectory", dumpdir.c_str());
@@ -809,6 +811,22 @@ void GSRunner::DumpStats()
 #define main real_main
 #endif

+static void CPUThreadMain(VMBootParameters* params) {
+	if (VMManager::Initialize(*params))
+	{
+		// run until end
+		GSDumpReplayer::SetLoopCount(s_loop_count);
+		VMManager::SetState(VMState::Running);
+		while (VMManager::GetState() == VMState::Running)
+			VMManager::Execute();
+		VMManager::Shutdown(false);
+		GSRunner::DumpStats();
+	}
+
+	VMManager::Internal::CPUThreadShutdown();
+	GSRunner::StopPlatformMessagePump();
+}
+
 int main(int argc, char* argv[])
 {
 	CrashHandler::Install();
@@ -837,16 +855,9 @@ int main(int argc, char* argv[])
 	VMManager::ApplySettings();
 	GSDumpReplayer::SetIsDumpRunner(true);

-	if (VMManager::Initialize(params))
-	{
-		// run until end
-		GSDumpReplayer::SetLoopCount(s_loop_count);
-		VMManager::SetState(VMState::Running);
-		while (VMManager::GetState() == VMState::Running)
-			VMManager::Execute();
-		VMManager::Shutdown(false);
-		GSRunner::DumpStats();
-	}
+	std::thread cputhread(CPUThreadMain, &params);
+	GSRunner::PumpPlatformMessages(/*forever=*/true);
+	cputhread.join();

 	VMManager::Internal::CPUThreadShutdown();
 	GSRunner::DestroyPlatformWindow();
@@ -859,9 +870,6 @@ void Host::PumpMessagesOnCPUThread()
 	// update GS thread copy of frame number
 	MTGS::RunOnGSThread([frame_number = GSDumpReplayer::GetFrameNumber()]() { s_dump_frame_number = frame_number; });
 	MTGS::RunOnGSThread([loop_number = GSDumpReplayer::GetLoopCount()]() { s_loop_number = loop_number; });
-
-	// process any window messages (but we shouldn't really have any)
-	GSRunner::PumpPlatformMessages();
 }

 s32 Host::Internal::GetTranslatedStringImpl(
@@ -975,16 +983,32 @@ std::optional<WindowInfo> GSRunner::GetPlatformWindowInfo()
 	return wi;
 }

-void GSRunner::PumpPlatformMessages()
+static constexpr int SHUTDOWN_MSG = WM_APP + 0x100;
+static DWORD MainThreadID;
+
+void GSRunner::PumpPlatformMessages(bool forever)
 {
 	MSG msg;
-	while (PeekMessageW(&msg, NULL, 0, 0, PM_REMOVE))
+	while (true)
 	{
-		TranslateMessage(&msg);
-		DispatchMessageW(&msg);
+		while (PeekMessageW(&msg, NULL, 0, 0, PM_REMOVE))
+		{
+			if (msg.message == SHUTDOWN_MSG)
+				return;
+			TranslateMessage(&msg);
+			DispatchMessageW(&msg);
+		}
+		if (!forever)
+			return;
+		WaitMessage();
 	}
 }

+void GSRunner::StopPlatformMessagePump()
+{
+	PostThreadMessageW(MainThreadID, SHUTDOWN_MSG, 0, 0);
+}
+
 LRESULT CALLBACK WndProc(HWND hwnd, UINT msg, WPARAM wParam, LPARAM lParam)
 {
 	return DefWindowProcW(hwnd, msg, wParam, lParam);
@@ -1003,7 +1027,51 @@ int wmain(int argc, wchar_t** argv)
 		u8_argptrs.push_back(u8_args[i].data());
 	u8_argptrs.push_back(nullptr);

+	MainThreadID = GetCurrentThreadId();
+
 	return real_main(argc, u8_argptrs.data());
 }

-#endif // _WIN32
+#elif defined(__APPLE__)
+
+static void* s_window;
+static WindowInfo s_wi;
+
+bool GSRunner::CreatePlatformWindow()
+{
+	pxAssertRel(!s_window, "Tried to create window when there already was one!");
+	s_window = CocoaTools::CreateWindow("PCSX2 GS Runner", WINDOW_WIDTH, WINDOW_HEIGHT);
+	CocoaTools::GetWindowInfoFromWindow(&s_wi, s_window);
+	PumpPlatformMessages();
+	return s_window;
+}
+
+void GSRunner::DestroyPlatformWindow()
+{
+	if (s_window) {
+		CocoaTools::DestroyWindow(s_window);
+		s_window = nullptr;
+	}
+}
+
+std::optional<WindowInfo> GSRunner::GetPlatformWindowInfo()
+{
+	WindowInfo wi;
+	if (s_window)
+		wi = s_wi;
+	else
+		wi.type = WindowInfo::Type::Surfaceless;
+	return wi;
+}
+
+void GSRunner::PumpPlatformMessages(bool forever)
+{
+	CocoaTools::RunCocoaEventLoop(forever);
+}
+
+void GSRunner::StopPlatformMessagePump()
+{
+	CocoaTools::StopMainThreadEventLoop();
+}
+
+#endif // _WIN32 / __APPLE__
--- a/pcsx2-qt/Debugger/Docking/DropIndicators.cpp
+++ b/pcsx2-qt/Debugger/Docking/DropIndicators.cpp
@@ -10,6 +10,7 @@

 #include <kddockwidgets/Config.h>
 #include <kddockwidgets/core/Group.h>
+#include <kddockwidgets/core/Platform.h>
 #include <kddockwidgets/core/indicators/SegmentedDropIndicatorOverlay.h>
 #include <kddockwidgets/qtwidgets/ViewFactory.h>

--- a/pcsx2-qt/Debugger/Memory/MemorySearchView.cpp
+++ b/pcsx2-qt/Debugger/Memory/MemorySearchView.cpp
@@ -311,7 +311,7 @@ void searchWorker(DebugInterface* cpu, std::vector<SearchResult>& searchResults,
 			const auto readValue = readValueAtAddress<T>(cpu, addr);

 			const bool doesMatch = handleSearchComparison(searchComparison, addr, &searchResult, searchValue, readValue);
-			if (!doesMatch)
+			if (doesMatch)
 				searchResult = MemorySearchView::SearchResult(addr, QVariant::fromValue(readValue), searchType);

 			return !doesMatch;
--- a/pcsx2-qt/GameList/GameListWidget.cpp
+++ b/pcsx2-qt/GameList/GameListWidget.cpp
@@ -138,16 +138,15 @@ namespace
 			if (option.state & QStyle::State_Selected)
 			{
 				// See QItemDelegate::selectedPixmap()
-				QString key = QString::fromStdString(fmt::format("{:016X}-{:d}", pix.cacheKey(), enabled));
+				QColor color = option.palette.color(enabled ? QPalette::Normal : QPalette::Disabled, QPalette::Highlight);
+				color.setAlphaF(0.3f);
+
+				QString key = QString::fromStdString(fmt::format("{:016X}-{:d}-{:08X}", pix.cacheKey(), enabled, color.rgba()));
 				QPixmap pm;
 				if (!QPixmapCache::find(key, &pm))
 				{
 					QImage img = pix.toImage().convertToFormat(QImage::Format_ARGB32_Premultiplied);

-					QColor color = option.palette.color(enabled ? QPalette::Normal : QPalette::Disabled,
-						QPalette::Highlight);
-					color.setAlphaF(0.3f);
-
 					QPainter tinted_painter(&img);
 					tinted_painter.setCompositionMode(QPainter::CompositionMode_SourceAtop);
 					tinted_painter.fillRect(0, 0, img.width(), img.height(), color);
--- a/pcsx2-qt/GameList/GameListWidget.h
+++ b/pcsx2-qt/GameList/GameListWidget.h
@@ -92,7 +92,7 @@ public Q_SLOTS:
 	void refreshGridCovers();

 protected:
-	void resizeEvent(QResizeEvent* event);
+	void resizeEvent(QResizeEvent* event) override;
 	bool event(QEvent* event) override;

 private:
--- a/pcsx2-qt/Themes.cpp
+++ b/pcsx2-qt/Themes.cpp
@@ -10,6 +10,7 @@

 #include <QtCore/QFile>
 #include <QtGui/QPalette>
+#include <QtGui/QPixmapCache>
 #include <QtWidgets/QApplication>
 #include <QtWidgets/QStyle>
 #include <QtWidgets/QStyleFactory>
@@ -43,6 +44,12 @@ void QtHost::UpdateApplicationTheme()

 	SetStyleFromSettings();
 	SetIconThemeFromStyle();
+
+	// Qt generates tinted versions of icons and stores them in QPixmapCache
+	// The key used does not seem to include the theme (or tint colour).
+	// This can cause icons tinted for wrong theme to be used for selected/disabled.
+	// As a workaround, reset the pixmap cache to clear icons tinted for the old theme.
+	QPixmapCache::clear();
 }

 bool QtHost::IsDarkApplicationTheme()
--- a/pcsx2-qt/Translations/pcsx2-qt_en.ts
+++ b/pcsx2-qt/Translations/pcsx2-qt_en.ts
@@ -10856,7 +10856,7 @@ Do you want to load this save and continue?</source>
        <translation type="unfinished"></translation>
    </message>
    <message>
-        <location filename="../../pcsx2/GS/Renderers/HW/GSTextureReplacements.cpp" line="699"/>
+        <location filename="../../pcsx2/GS/Renderers/HW/GSTextureReplacements.cpp" line="740"/>
        <source>Disabling autogenerated mipmaps on one or more compressed replacement textures. Please generate mipmaps when compressing your textures.</source>
        <translation type="unfinished"></translation>
    </message>
@@ -10866,7 +10866,7 @@ Do you want to load this save and continue?</source>
        <translation type="unfinished"></translation>
    </message>
    <message>
-        <location filename="../../pcsx2/GS/Renderers/Vulkan/GSDeviceVK.cpp" line="5060"/>
+        <location filename="../../pcsx2/GS/Renderers/Vulkan/GSDeviceVK.cpp" line="5061"/>
        <source>Spin GPU During Readbacks is enabled, but calibrated timestamps are unavailable.  This might be really slow.</source>
        <translation type="unfinished"></translation>
    </message>
@@ -10959,7 +10959,7 @@ Please see our official documentation for more information.</source>
 <context>
    <name>GSDeviceOGL</name>
    <message>
-        <location filename="../../pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.cpp" line="641"/>
+        <location filename="../../pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.cpp" line="642"/>
        <source>OpenGL renderer is not supported. Only OpenGL {}.{}
 was found</source>
        <translation type="unfinished"></translation>
@@ -20316,6 +20316,15 @@ Scanning recursively takes more time, but will identify files in subdirectories.
        <translation type="unfinished"></translation>
    </message>
 </context>
+<context>
+    <name>TextureReplacement</name>
+    <message>
+        <location filename="../../pcsx2/GS/Renderers/HW/GSTextureReplacements.cpp" line="409"/>
+        <source>Texture replacement directory {} will not work on case sensitive filesystems.
+Rename it to {} to remove this warning.</source>
+        <translation type="unfinished"></translation>
+    </message>
+</context>
 <context>
    <name>ThreadModel</name>
    <message>
--- a/pcsx2/CMakeLists.txt
+++ b/pcsx2/CMakeLists.txt
@@ -1237,9 +1237,16 @@ fixup_file_properties(PCSX2)
 force_include_last(PCSX2_FLAGS "/(usr|local)/include/?$")

 if (APPLE)
+	find_library(APPKIT_LIBRARY AppKit)
+	find_library(IOKIT_LIBRARY IOKit)
 	find_library(METAL_LIBRARY Metal)
 	find_library(QUARTZCORE_LIBRARY QuartzCore)
-	target_link_libraries(PCSX2_FLAGS INTERFACE ${METAL_LIBRARY} ${QUARTZCORE_LIBRARY})
+	target_link_libraries(PCSX2_FLAGS INTERFACE
+		${APPKIT_LIBRARY}
+		${IOKIT_LIBRARY}
+		${METAL_LIBRARY}
+		${QUARTZCORE_LIBRARY}
+	)
 endif()

 set_property(GLOBAL PROPERTY PCSX2_SOURCE_DIR ${CMAKE_CURRENT_SOURCE_DIR})
--- a/pcsx2/GS/GSLocalMemory.cpp
+++ b/pcsx2/GS/GSLocalMemory.cpp
@@ -446,6 +446,13 @@ std::vector<GSVector2i>* GSLocalMemory::GetPage2TileMap(const GIFRegTEX0& TEX0)
 	return p2t;
 }

+u32 GSLocalMemory::IsPageAlignedMasked(u32 psm, const GSVector4i& rc)
+{
+	const psm_t& psm_s = m_psm[psm];
+	const GSVector4i pgmsk = GSVector4i(psm_s.pgs).xyxy() - GSVector4i(1);
+	return ((rc & pgmsk) == GSVector4i::zero()).mask();
+}
+
 bool GSLocalMemory::IsPageAligned(u32 psm, const GSVector4i& rc)
 {
 	const psm_t& psm_s = m_psm[psm];
--- a/pcsx2/GS/GSLocalMemory.h
+++ b/pcsx2/GS/GSLocalMemory.h
@@ -524,6 +524,7 @@ public:
 	GSPixelOffset4* GetPixelOffset4(const GIFRegFRAME& FRAME, const GIFRegZBUF& ZBUF);
 	std::vector<GSVector2i>* GetPage2TileMap(const GIFRegTEX0& TEX0);
 	static bool HasOverlap(u32 src_bp, u32 src_bw, u32 src_psm, GSVector4i src_rect, u32 dst_bp, u32 dst_bw, u32 dst_psm, GSVector4i dst_rect);
+	static u32 IsPageAlignedMasked(u32 psm, const GSVector4i& rc);
 	static bool IsPageAligned(u32 psm, const GSVector4i& rc);
 	static u32 GetStartBlockAddress(u32 bp, u32 bw, u32 psm, GSVector4i rect);
 	static u32 GetEndBlockAddress(u32 bp, u32 bw, u32 psm, GSVector4i rect);
--- a/pcsx2/GS/Renderers/Common/GSDevice.cpp
+++ b/pcsx2/GS/Renderers/Common/GSDevice.cpp
@@ -68,6 +68,7 @@ const char* shaderName(ShaderConvert value)
 		case ShaderConvert::DEPTH_COPY:             return "ps_depth_copy";
 		case ShaderConvert::DOWNSAMPLE_COPY:        return "ps_downsample_copy";
 		case ShaderConvert::RGBA_TO_8I:             return "ps_convert_rgba_8i";
+		case ShaderConvert::RGB5A1_TO_8I:           return "ps_convert_rgb5a1_8i";
 		case ShaderConvert::CLUT_4:                 return "ps_convert_clut_4";
 		case ShaderConvert::CLUT_8:                 return "ps_convert_clut_8";
 		case ShaderConvert::YUV:                    return "ps_yuv";
--- a/pcsx2/GS/Renderers/Common/GSDevice.h
+++ b/pcsx2/GS/Renderers/Common/GSDevice.h
@@ -44,6 +44,7 @@ enum class ShaderConvert
 	DEPTH_COPY,
 	DOWNSAMPLE_COPY,
 	RGBA_TO_8I,
+	RGB5A1_TO_8I,
 	CLUT_4,
 	CLUT_8,
 	YUV,
--- a/pcsx2/GS/Renderers/DX11/GSDevice11.cpp
+++ b/pcsx2/GS/Renderers/DX11/GSDevice11.cpp
@@ -1417,14 +1417,14 @@ void GSDevice11::ConvertToIndexedTexture(GSTexture* sTex, float sScale, u32 offs
 	{
 		float scale;
 		float pad1[3];
-		u32 SBW, DBW, pad3;
+		u32 SBW, DBW, SPSM;
 	};

-	const Uniforms cb = {sScale, {}, SBW, DBW};
+	const Uniforms cb = {sScale, {}, SBW, DBW, SPSM};
 	m_ctx->UpdateSubresource(m_merge.cb.get(), 0, nullptr, &cb, 0, 0);

 	const GSVector4 dRect(0, 0, dTex->GetWidth(), dTex->GetHeight());
-	const ShaderConvert shader = ShaderConvert::RGBA_TO_8I;
+	const ShaderConvert shader = ((SPSM & 0xE) == 0) ? ShaderConvert::RGBA_TO_8I : ShaderConvert::RGB5A1_TO_8I;
 	StretchRect(sTex, GSVector4::zero(), dTex, dRect, m_convert.ps[static_cast<int>(shader)].get(), m_merge.cb.get(), nullptr, false);
 }

--- a/pcsx2/GS/Renderers/DX12/GSDevice12.cpp
+++ b/pcsx2/GS/Renderers/DX12/GSDevice12.cpp
@@ -1486,15 +1486,15 @@ void GSDevice12::ConvertToIndexedTexture(
 	{
 		float scale;
 		float pad1[3];
-		u32 SBW, DBW, pad2;
+		u32 SBW, DBW, SPSM;
 	};

-	const Uniforms cb = {sScale, {}, SBW, DBW};
+	const Uniforms cb = {sScale, {}, SBW, DBW, SPSM};
 	SetUtilityRootSignature();
 	SetUtilityPushConstants(&cb, sizeof(cb));

 	const GSVector4 dRect(0, 0, dTex->GetWidth(), dTex->GetHeight());
-	const ShaderConvert shader = ShaderConvert::RGBA_TO_8I;
+	const ShaderConvert shader = ((SPSM & 0xE) == 0) ? ShaderConvert::RGBA_TO_8I : ShaderConvert::RGB5A1_TO_8I;
 	DoStretchRect(static_cast<GSTexture12*>(sTex), GSVector4::zero(), static_cast<GSTexture12*>(dTex), dRect,
 		m_convert[static_cast<int>(shader)].get(), false, true);
 }
--- a/pcsx2/GS/Renderers/HW/GSHwHack.cpp
+++ b/pcsx2/GS/Renderers/HW/GSHwHack.cpp
@@ -184,40 +184,6 @@ bool GSHwHack::GSC_NamcoGames(GSRendererHW& r, int& skip)
 {
 	if (skip == 0)
 	{
-		if (r.IsPossibleChannelShuffle() && !(RTBP0 & 31))
-		{
-			GSVertex* v = &r.m_vertex.buff[0];
-
-			// Make sure we're detecting the right effect.
-			if (((v[1].XYZ.X - v[0].XYZ.X) >> 4) != 8 || ((v[1].XYZ.Y - v[0].XYZ.Y) >> 4) != 14)
-				return false;
-
-			GSTextureCache::Target* rt = g_texture_cache->LookupTarget(GIFRegTEX0::Create(RTBP0, RFBW, RFPSM),
-				GSVector2i(1, 1), r.GetTextureScaleFactor(), GSTextureCache::RenderTarget);
-			if (!rt)
-				return false;
-
-			GL_INS("GSC_NamcoGames(): HLE channel shuffle");
-
-			// have to set up the palette ourselves too, since GSC executes before it does
-			r.m_mem.m_clut.Read32(RTEX0, r.m_draw_env->TEXA);
-			std::shared_ptr<GSTextureCache::Palette> palette =
-				g_texture_cache->LookupPaletteObject(r.m_mem.m_clut, GSLocalMemory::m_psm[RTEX0.PSM].pal, true);
-			if (!palette)
-				return false;
-
-			GSHWDrawConfig& conf = r.BeginHLEHardwareDraw(
-				rt->GetTexture(), nullptr, rt->GetScale(), rt->GetTexture(), rt->GetScale(), rt->GetUnscaledRect());
-			conf.pal = palette->GetPaletteGSTexture();
-			conf.ps.channel = ChannelFetch_RGB;
-			conf.colormask.wa = false;
-			r.EndHLEHardwareDraw(false);
-
-			// 12 pages: 2 calls by channel, 3 channels, 1 blit
-			skip = 12 * (3 + 3 + 1);
-			return true;
-		}
-
 		if (!s_nativeres && r.PRIM->PRIM == GS_SPRITE && RTME && RTEX0.TFX == 1 && RFPSM == RTPSM && RTPSM == PSMCT32 && RFBMSK == 0xFF000000 && r.m_index.tail > 2)
 		{
 			GSVertex* v = &r.m_vertex.buff[0];
@@ -431,31 +397,6 @@ bool GSHwHack::GSC_TalesOfLegendia(GSRendererHW& r, int& skip)
 	return true;
 }

-bool GSHwHack::GSC_Kunoichi(GSRendererHW& r, int& skip)
-{
-	if (skip == 0)
-	{
-		if (!RTME && (RFBP == 0x0 || RFBP == 0x00700 || RFBP == 0x00800) && RFPSM == PSMCT32 && RFBMSK == 0x00FFFFFF)
-		{
-			// Removes depth effects(shadows) not rendered correctly on all renders.
-			skip = 3;
-		}
-		if (RTME && (RFBP == 0x0700 || RFBP == 0) && RTBP0 == 0x0e00 && RTPSM == 0 && RFBMSK == 0)
-		{
-			skip = 1; // Removes black screen (not needed anymore maybe)?
-		}
-	}
-	else
-	{
-		if (RTME && (RFBP == 0x0e00) && RFPSM == PSMCT32 && RFBMSK == 0xFF000000)
-		{
-			skip = 0;
-		}
-	}
-
-	return true;
-}
-
 bool GSHwHack::GSC_ZettaiZetsumeiToshi2(GSRendererHW& r, int& skip)
 {
 	if (skip == 0)
@@ -505,27 +446,6 @@ bool GSHwHack::GSC_ZettaiZetsumeiToshi2(GSRendererHW& r, int& skip)
 	return true;
 }

-bool GSHwHack::GSC_SakuraWarsSoLongMyLove(GSRendererHW& r, int& skip)
-{
-	if (skip == 0)
-	{
-		if (RTME == 0 && RFBP != RTBP0 && RTBP0 && RFBMSK == 0x00FFFFFF)
-		{
-			skip = 3; // Remove darkness
-		}
-		else if (RTME == 0 && RFBP == RTBP0 && (RTBP0 == 0x1200 || RTBP0 == 0x1180 || RTBP0 == 0) && RFBMSK == 0x00FFFFFF)
-		{
-			skip = 3; // Remove darkness
-		}
-		else if (RTME && (RFBP == 0 || RFBP == 0x1180) && RFPSM == PSMCT32 && RTBP0 == 0x3F3F && RTPSM == PSMT8)
-		{
-			skip = 1; // Floodlight
-		}
-	}
-
-	return true;
-}
-
 bool GSHwHack::GSC_UltramanFightingEvolution(GSRendererHW& r, int& skip)
 {
 	if (skip == 0)
@@ -1393,12 +1313,10 @@ bool GSHwHack::MV_Ico(GSRendererHW& r)
 #define CRC_F(name) { #name, &GSHwHack::name }

 const GSHwHack::Entry<GSRendererHW::GSC_Ptr> GSHwHack::s_get_skip_count_functions[] = {
-	CRC_F(GSC_Kunoichi),
 	CRC_F(GSC_Manhunt2),
 	CRC_F(GSC_MidnightClub3),
 	CRC_F(GSC_SacredBlaze),
 	CRC_F(GSC_GuitarHero),
-	CRC_F(GSC_SakuraWarsSoLongMyLove),
 	CRC_F(GSC_SFEX3),
 	CRC_F(GSC_DTGames),
 	CRC_F(GSC_TalesOfLegendia),
--- a/pcsx2/GS/Renderers/HW/GSHwHack.h
+++ b/pcsx2/GS/Renderers/HW/GSHwHack.h
@@ -16,9 +16,7 @@ public:
 	static bool GSC_BlackAndBurnoutSky(GSRendererHW& r, int& skip);
 	static bool GSC_MidnightClub3(GSRendererHW& r, int& skip);
 	static bool GSC_TalesOfLegendia(GSRendererHW& r, int& skip);
-	static bool GSC_Kunoichi(GSRendererHW& r, int& skip);
 	static bool GSC_ZettaiZetsumeiToshi2(GSRendererHW& r, int& skip);
-	static bool GSC_SakuraWarsSoLongMyLove(GSRendererHW& r, int& skip);
 	static bool GSC_UltramanFightingEvolution(GSRendererHW& r, int& skip);
 	static bool GSC_TalesofSymphonia(GSRendererHW& r, int& skip);
 	static bool GSC_UrbanReign(GSRendererHW& r, int& skip);
--- a/pcsx2/GS/Renderers/HW/GSRendererHW.cpp
+++ b/pcsx2/GS/Renderers/HW/GSRendererHW.cpp
@@ -2862,7 +2862,9 @@ void GSRendererHW::Draw()

 		GIFRegTEX0 FRAME_TEX0;
 		bool shuffle_target = false;
-		if (!no_rt && GSLocalMemory::m_psm[m_cached_ctx.FRAME.PSM].bpp == 16 && GSLocalMemory::m_psm[m_cached_ctx.TEX0.PSM].bpp >= 16 &&
+		const u32 page_alignment = GSLocalMemory::IsPageAlignedMasked(m_cached_ctx.TEX0.PSM, m_r);
+		const bool page_aligned = (page_alignment & 0xF0F0) != 0; // Make sure Y is page aligned.
+		if (!no_rt && page_aligned && m_cached_ctx.ZBUF.ZMSK && 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)))
 		{
 			// 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.
@@ -3164,6 +3166,144 @@ void GSRendererHW::Draw()
 				}
 			}
 		}
+
+		if (no_rt && ds->m_TEX0.TBP0 != m_cached_ctx.ZBUF.Block())
+		{
+			const GSLocalMemory::psm_t& zbuf_psm = GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM];
+			int 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)) * zbuf_psm.pgs.y; // I know I could just not shift it..
+			int texture_offset = 0;
+			int horizontal_offset = ((static_cast<int>((m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0)) / 32) % static_cast<int>(std::max(ds->m_TEX0.TBW, 1U))) * zbuf_psm.pgs.x;
+			// Used to reduce the offset made later in channel shuffles
+			m_target_offset = std::abs(static_cast<int>((m_cached_ctx.ZBUF.Block() - ds->m_TEX0.TBP0)) >> 5);
+
+			if (vertical_offset < 0)
+			{
+				ds->m_TEX0.TBP0 = m_cached_ctx.ZBUF.Block();
+				GSVector2i new_size = ds->m_unscaled_size;
+				// Make sure to use the original format for the offset.
+				const int new_offset = std::abs((vertical_offset / zbuf_psm.pgs.y) * GSLocalMemory::m_psm[ds->m_TEX0.PSM].pgs.y);
+				texture_offset = new_offset;
+
+				new_size.y += new_offset;
+
+				const GSVector4i new_drect = GSVector4i(0, new_offset * ds->m_scale, new_size.x * ds->m_scale, new_size.y * ds->m_scale);
+				ds->ResizeTexture(new_size.x, new_size.y, true, true, new_drect);
+
+				if (src && src->m_from_target && src->m_from_target == ds && src->m_target_direct)
+				{
+					src->m_texture = ds->m_texture;
+
+					// If we've moved it and the source is expecting to be inside this target, we need to update the region to point to it.
+					int max_region_y = src->m_region.GetMaxY() + new_offset;
+					if (max_region_y == new_offset)
+						max_region_y = new_size.y;
+
+					src->m_region.SetY(src->m_region.GetMinY() + new_offset, max_region_y);
+				}
+
+				ds->m_valid.y += new_offset;
+				ds->m_valid.w += new_offset;
+				ds->m_drawn_since_read.y += new_offset;
+				ds->m_drawn_since_read.w += new_offset;
+
+				g_texture_cache->CombineAlignedInsideTargets(ds, src);
+
+				if (ds->m_dirty.size())
+				{
+					for (int i = 0; i < static_cast<int>(ds->m_dirty.size()); i++)
+					{
+						ds->m_dirty[i].r.y += new_offset;
+						ds->m_dirty[i].r.w += new_offset;
+					}
+				}
+
+				t_size.y += std::abs(vertical_offset);
+				vertical_offset = 0;
+			}
+
+			if (horizontal_offset < 0)
+			{
+				// Thankfully this doesn't really happen, but catwoman moves the framebuffer backwards 1 page with a channel shuffle, which is really messy and not easy to deal with.
+				// Hopefully the quick channel shuffle will just guess this and run with it.
+				ds->m_TEX0.TBP0 += horizontal_offset;
+				horizontal_offset = 0;
+			}
+
+			if (vertical_offset || horizontal_offset)
+			{
+				GSVertex* v = &m_vertex.buff[0];
+
+				for (u32 i = 0; i < m_vertex.tail; i++)
+				{
+					v[i].XYZ.X += horizontal_offset << 4;
+					v[i].XYZ.Y += vertical_offset << 4;
+				}
+
+				if (texture_offset && src && src->m_from_target && src->m_target_direct && src->m_from_target == ds)
+				{
+					GSVector4i src_region = src->GetRegionRect();
+
+					if (src_region.rempty())
+					{
+						src_region = GSVector4i::loadh(ds->m_unscaled_size);
+						src_region.y += texture_offset;
+					}
+					else
+					{
+						src_region.y += texture_offset;
+						src_region.w += texture_offset;
+					}
+					src->m_region.SetX(src_region.x, src_region.z);
+					src->m_region.SetY(src_region.y, src_region.w);
+				}
+
+				m_context->scissor.in.x += horizontal_offset;
+				m_context->scissor.in.z += horizontal_offset;
+				m_context->scissor.in.y += vertical_offset;
+				m_context->scissor.in.w += vertical_offset;
+				m_r.y += vertical_offset;
+				m_r.w += vertical_offset;
+				m_r.x += horizontal_offset;
+				m_r.z += horizontal_offset;
+				m_in_target_draw = ds->m_TEX0.TBP0 != m_cached_ctx.ZBUF.Block();
+				m_vt.m_min.p.x += horizontal_offset;
+				m_vt.m_max.p.x += horizontal_offset;
+				m_vt.m_min.p.y += vertical_offset;
+				m_vt.m_max.p.y += vertical_offset;
+
+				t_size.y = ds->m_unscaled_size.y - vertical_offset;
+				t_size.x = ds->m_unscaled_size.x - horizontal_offset;
+			}
+
+			// Don't resize if the BPP don't match.
+			GSVector2i new_size = GetValidSize(src, possible_shuffle);
+			if (new_size.x > ds->m_unscaled_size.x || new_size.y > ds->m_unscaled_size.y)
+			{
+				const u32 new_width = std::max(new_size.x, ds->m_unscaled_size.x);
+				const u32 new_height = std::max(new_size.y, ds->m_unscaled_size.y);
+
+				//DevCon.Warning("HW: Resizing texture %d x %d draw %d", ds->m_unscaled_size.x, new_height, s_n);
+				ds->ResizeTexture(new_width, new_height);
+			}
+			else if ((IsPageCopy() || is_possible_mem_clear) && m_r.width() <= zbuf_psm.pgs.x && m_r.height() <= zbuf_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.ZBUF.PSM].pgs.x, GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.y + vertical_offset));
+				ds->UpdateValidity(update_valid, true);
+				if (is_possible_mem_clear)
+				{
+					if ((horizontal_offset + GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.x) >= static_cast<int>(ds->m_TEX0.TBW * 64) && next_ctx.ZBUF.Block() == (m_cached_ctx.ZBUF.Block() + 0x20))
+					{
+						update_valid.x = 0;
+						update_valid.z = GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.x;
+						update_valid.y += GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.y;
+						update_valid.w += GSLocalMemory::m_psm[m_cached_ctx.ZBUF.PSM].pgs.y;
+						ds->UpdateValidity(update_valid, true);
+					}
+				}
+			}
+		}
 	}

 	if (!no_rt)
--- a/pcsx2/GS/Renderers/HW/GSTextureCache.cpp
+++ b/pcsx2/GS/Renderers/HW/GSTextureCache.cpp
@@ -228,8 +228,8 @@ bool GSTextureCache::CanTranslate(u32 bp, u32 bw, u32 spsm, GSVector4i r, u32 db
 {
 	const GSVector2i src_page_size = GSLocalMemory::m_psm[spsm].pgs;
 	const GSVector2i dst_page_size = GSLocalMemory::m_psm[dpsm].pgs;
-	const bool bp_page_aligned_bp = ((bp & ~((1 << 5) - 1)) == bp) || bp == dbp;
 	const bool block_layout_match = GSLocalMemory::m_psm[spsm].bpp == GSLocalMemory::m_psm[dpsm].bpp;
+	const bool bp_page_aligned_bp = ((bp & ~((1 << 5) - 1)) == bp) || bp == dbp || (block_layout_match && bw == dbw);
 	const GSVector4i page_mask(GSVector4i((src_page_size.x - 1), (src_page_size.y - 1)).xyxy());
 	const GSVector4i masked_rect(r & ~page_mask);
 	const int src_pixel_width = static_cast<int>(bw * 64);
@@ -276,7 +276,7 @@ GSVector4i GSTextureCache::TranslateAlignedRectByPage(u32 tbp, u32 tebp, u32 tbw
 	int page_offset = (static_cast<int>(sbp) - static_cast<int>(tbp)) >> 5;
 	int block_offset = (static_cast<int>(sbp) - static_cast<int>(tbp)) & 0x1F;

-	if (!(s_psm.bpp == t_psm.bpp))
+	if (!(s_psm.bpp == t_psm.bpp) || block_offset)
 	{
 		if (block_offset)
 			in_rect = in_rect.ralign<Align_Outside>(s_psm.bs);
@@ -299,6 +299,33 @@ GSVector4i GSTextureCache::TranslateAlignedRectByPage(u32 tbp, u32 tebp, u32 tbw
 			DevCon.Warning("Error translating rect");
 			return GSVector4i::zero();
 		}
+
+		const bool block_matched_format = s_psm.bpp == t_psm.bpp && (sbw == tbw || sbw == 1);
+		// If there is block offset left over, try to adjust to that.
+		if (block_matched_format)
+		{
+			GSVector4i b2a_offset = GSVector4i::zero();
+			const GSVector4i target_rect = GSVector4i(0, 0, src_bw, 2048);
+			bool offset_found = false;
+
+			for (b2a_offset.x = target_rect.x; b2a_offset.x < target_rect.z; b2a_offset.x += s_psm.bs.x)
+			{
+				for (b2a_offset.y = target_rect.y; b2a_offset.y < target_rect.w; b2a_offset.y += s_psm.bs.y)
+				{
+					const u32 a_candidate_bp = s_psm.info.bn(b2a_offset.x, b2a_offset.y, tbp, sbw);
+					if (sbp == a_candidate_bp)
+					{
+						offset_found = true;
+						break;
+					}
+				}
+				if (offset_found)
+					break;
+			}
+
+			if (offset_found)
+				in_rect = (in_rect + b2a_offset.xyxy()).max_i32(GSVector4i(0));
+		}
 	}

 	GSVector4i new_rect = GSVector4i::zero();
@@ -618,7 +645,7 @@ void GSTextureCache::DirtyRectByPage(u32 sbp, u32 spsm, u32 sbw, Target* t, GSVe
 	int block_offset = static_cast<int>(sbp) - static_cast<int>(target_bp);
 	// Different format needs to be page aligned, unless the block layout matches, then we can block align
 	// Might be able to translate the original rect.
-	if (!(src_info->bpp == dst_info->bpp))
+	if (!(src_info->bpp == dst_info->bpp) || block_offset)
 	{
 		const int src_bpp = src_info->bpp;
 		const bool column_align = !block_offset && src_r.z <= src_info->cs.x && src_r.w <= src_info->cs.y && src_info->depth == dst_info->depth;
@@ -741,7 +768,7 @@ void GSTextureCache::DirtyRectByPage(u32 sbp, u32 spsm, u32 sbw, Target* t, GSVe
 			const int yblocks = in_rect.height() / src_info->bs.y;
 			// if !(block_offset & 0x7) is false, this is technically incorrect, but FFX hates it and starts making a mess, so it's better this way without adding complexity.
 			// TODO maybe: Add per block invalidation? ugh, would have to keep that to small blocks. 2 blocks in the case of FFX.
-			if ((xblocks <= 4 && yblocks <= 2) || req_depth_offset)
+			if ((xblocks <= (src_info->pgs.x / src_info->bs.x) && yblocks <= (src_info->pgs.y / src_info->bs.y)) || req_depth_offset)
 			{
 				GSVector4i b2a_offset = GSVector4i::zero();
 				const GSVector4i target_rect = GSVector4i(0, 0, src_width, 2048);
@@ -1165,8 +1192,8 @@ GSTextureCache::Source* GSTextureCache::LookupDepthSource(const bool is_depth, c
 		if (inside_target)
 		{
 			// Need to set it up as a region target.
-			src->m_region.SetX(block_boundary_rect.x, src->m_from_target->m_valid.z);
-			src->m_region.SetY(block_boundary_rect.y, src->m_from_target->m_valid.w);
+			src->m_region.SetX(block_boundary_rect.x, std::max(src->m_from_target->m_valid.z, block_boundary_rect.z));
+			src->m_region.SetY(block_boundary_rect.y, std::max(src->m_from_target->m_valid.w, block_boundary_rect.w));
 		}

 		if (GSRendererHW::GetInstance()->IsTBPFrameOrZ(dst->m_TEX0.TBP0))
@@ -1279,11 +1306,6 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 			// Make sure it is page aligned, otherwise things get messy with the pixel order (Tomb Raider Legend).
 			if (t->m_used)
 			{
-				// If the BP is block offset it's unlikely to be a target, but a target can be made by a HW move, so we need to check for a match.
-				// Good for Baldurs Gate Dark Alliance (HW Move), bad for Tomb Raider Legends (just offset).
-				if (((bp & (BLOCKS_PER_PAGE - 1)) != (t->m_TEX0.TBP0 & (BLOCKS_PER_PAGE - 1))) && (bp & (BLOCKS_PER_PAGE - 1)))
-					continue;
-
 				//const bool overlaps = t->Inside(bp, bw, psm, block_boundary_rect);
 				const bool overlaps = t->Overlaps(bp, bw, psm, block_boundary_rect);
 				// Try to make sure the target has available what we need, be careful of self referencing frames with font in the alpha.
@@ -1293,6 +1315,12 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 				if (!overlaps || (found_t && (GSState::s_n - dst->m_last_draw) < (GSState::s_n - t->m_last_draw)))
 					continue;

+				// If the BP is offset in to a page and the format does not match, trying to match up the correct position is very difficult since we don't swizzle.
+				// Tomb Raider Legends does a block level BP in PSMT8 over a C16 target, which is just a nightmare to get right.
+				// Baldurs Gate used to have this too, but now we can translate HW moves inside targets when the format matches.
+				if (((bp & (BLOCKS_PER_PAGE - 1)) != (t->m_TEX0.TBP0 & (BLOCKS_PER_PAGE - 1))) && (bp & (BLOCKS_PER_PAGE - 1)) && !CanTranslate(bp, bw, psm, block_boundary_rect, t->m_TEX0.TBP0, t->m_TEX0.PSM, t->m_TEX0.TBW))
+					continue;
+
 				const bool width_match = (std::max(64U, bw * 64U) >> GSLocalMemory::m_psm[psm].info.pageShiftX()) ==
 										 (std::max(64U, t->m_TEX0.TBW * 64U) >> GSLocalMemory::m_psm[t->m_TEX0.PSM].info.pageShiftX());

@@ -1617,8 +1645,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 						{
 							const bool outside_target = !t->Overlaps(bp, bw, psm, r);

-							// We don't have a shader for this.
-							if (!possible_shuffle && TEX0.PSM == PSMT8 && ((GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp != 32) || outside_target))
+							if (!possible_shuffle && TEX0.PSM == PSMT8 && outside_target)
 							{
 								continue;
 							}
@@ -1647,7 +1674,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 				// Make sure the texture actually is INSIDE the RT, it's possibly not valid if it isn't.
 				// Also check BP >= TBP, create source isn't equpped to expand it backwards and all data comes from the target. (GH3)
 				else if (GSConfig.UserHacks_TextureInsideRt >= GSTextureInRtMode::InsideTargets &&
-						 (GSLocalMemory::m_psm[color_psm].bpp >= 16 || (/*possible_shuffle &&*/ GSLocalMemory::m_psm[color_psm].bpp == 8 && GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp == 32)) && // Channel shuffles or non indexed lookups.
+						 (GSLocalMemory::m_psm[color_psm].bpp >= 16 || (/*possible_shuffle &&*/ GSLocalMemory::m_psm[color_psm].bpp == 8 && GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp >= 16)) && // Channel shuffles or non indexed lookups.
 					t->m_age <= 1 && (!found_t || t->m_last_draw > dst->m_last_draw) /*&& CanTranslate(bp, bw, psm, block_boundary_rect, t->m_TEX0.TBP0, t->m_TEX0.PSM, t->m_TEX0.TBW)*/)
 				{
 					u32 rt_tbw = std::max(1U, t->m_TEX0.TBW);
@@ -1666,7 +1693,7 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 					// Keep note that 2 bw is basically 1 normal page, as bw is in 64 pixels, and 8bit pages are 128 pixels wide, aka 2 bw.
 					// Also check for 4HH/HL and 8H which use the alpha channel, if the page order is wrong this can cause problems as well (Jak X font).
 					else if (!possible_shuffle && GSLocalMemory::m_psm[psm].trbpp <= 8 && (GSUtil::GetChannelMask(t->m_TEX0.PSM) != 0xF || 
-							(GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp == 32 && (!(block_boundary_rect.w <= GSLocalMemory::m_psm[psm].pgs.y &&
+							((GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp != 16 || GSLocalMemory::m_psm[psm].bpp < 16) && (!(block_boundary_rect.w <= GSLocalMemory::m_psm[psm].pgs.y &&
 							((GSLocalMemory::m_psm[psm].bpp == 32) ? bw : ((bw + 1) / 2)) <= t->m_TEX0.TBW) &&
 							!(((GSLocalMemory::m_psm[psm].bpp == 32) ? bw : ((bw + 1) / 2)) == rt_tbw)))))
 					{
@@ -1773,10 +1800,6 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 							if (!t->HasValidBitsForFormat(psm, req_color, req_alpha, rt_tbw == TEX0.TBW) && !(possible_shuffle && GSLocalMemory::m_psm[psm].bpp == 16 && GSLocalMemory::m_psm[t->m_TEX0.PSM].bpp == 32))
 								continue;

-							// Be careful of shuffles where it can shuffle the width of the target, even though it may not have all been drawn to.
-							if (!possible_shuffle && frame.Block() != TEX0.TBP0 && !t->Inside(bp, bw, psm, block_boundary_rect))
-								continue;
-
 							x_offset = rect.x;
 							y_offset = rect.y;
 							dst = t;
@@ -2007,7 +2030,6 @@ GSTextureCache::Source* GSTextureCache::LookupSource(const bool is_color, const
 			GL_CACHE("TC: src miss (0x%x, 0x%x, %s)", TEX0.TBP0, psm_s.pal > 0 ? TEX0.CBP : 0, psm_str(TEX0.PSM));
 		}
 #endif
-
 		// This is for the condition where the target doesn't exist on a shuffle and it needs to load from memory.
 		// The Godfather clears the depth buffer with a normal clear, so our depth target gets deleted, then because it finds no target
 		// it assumes it really is 16bits, causing the texture to be full of garbage, and our shuffle handling becomes a mess.
@@ -3545,11 +3567,11 @@ bool GSTextureCache::PreloadTarget(GIFRegTEX0 TEX0, const GSVector2i& size, cons
 							continue;
 						}

-						const int height_adjust = (((dst_end_block - t->m_TEX0.TBP0) >> 5) / std::max(t->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[t->m_TEX0.PSM].pgs.y;
+						const int height_adjust = ((((dst_end_block + 31) - t->m_TEX0.TBP0) >> 5) / std::max(t->m_TEX0.TBW, 1U)) * GSLocalMemory::m_psm[t->m_TEX0.PSM].pgs.y;

 						if (height_adjust < t->m_unscaled_size.y)
 						{
-							t->m_TEX0.TBP0 = dst_end_block;
+							t->m_TEX0.TBP0 = GSLocalMemory::GetStartBlockAddress(t->m_TEX0.TBP0, t->m_TEX0.TBW, t->m_TEX0.PSM, GSVector4i(0, height_adjust, t->m_valid.z, t->m_valid.w));
 							t->m_valid.w -= height_adjust;
 							t->ResizeValidity(t->m_valid);

@@ -4735,7 +4757,7 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 	// Not even going to go down the rabbit hole of palette formats on the GPU.. We shouldn't have any targets with P4/P8 anyway.
 	const GSLocalMemory::psm_t& spsm_s = GSLocalMemory::m_psm[SPSM];
 	const GSLocalMemory::psm_t& dpsm_s = GSLocalMemory::m_psm[DPSM];
-	if (SPSM != DPSM || ((spsm_s.pal + dpsm_s.pal) != 0 && !alpha_only))
+	if (GSLocalMemory::m_psm[SPSM].bpp != GSLocalMemory::m_psm[DPSM].bpp || ((spsm_s.pal + dpsm_s.pal) != 0 && !alpha_only))
 	{
 		GL_CACHE("TC: Skipping HW move from 0x%X to 0x%X with SPSM=%s DPSM=%s", SBP, DBP, psm_str(SPSM), psm_str(DPSM));
 		return false;
@@ -4751,6 +4773,7 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 			return false;
 	}

+	bool req_resize = false;
 	if (m_expected_src_bp == static_cast<int>(SBP) && m_expected_dst_bp == static_cast<int>(DBP))
 	{
 		// Get the new position so we can work out the offset.
@@ -4761,7 +4784,7 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 		sy += rect_offset.y;
 		dx += rect_offset.x;
 		dy += rect_offset.y;
-
+		req_resize = true;
 		GL_INS("TC: Detected striped move, realigning from SBP %x->%x DBP %x->%x", SBP, m_remembered_src_bp, DBP, m_remembered_dst_bp);

 		SBP = m_remembered_src_bp;
@@ -4806,15 +4829,18 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 		const GSVector2i target_size = GetTargetSize(DBP, DBW, DPSM, Common::AlignUpPow2(w, 64), h);
 		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);
+		{
+			dst = Target::Create(new_TEX0, target_size.x, target_size.y, src->m_scale, GSLocalMemory::m_psm[DPSM].depth, true);
+			if (!dst) [[unlikely]]
+				return false;
+		}
 		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));
+			req_resize = true;

 			// If it was matched to an old target, make sure to clear the other type and update its information.
 			if (dst->m_was_dst_matched)
 			{
-				g_texture_cache->InvalidateVideoMemType(GSTextureCache::DepthStencil - dst->m_type, dst->m_TEX0.TBP0);
 				dst->m_TEX0 = new_TEX0;
 			}
 		}
@@ -4829,6 +4855,24 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 	if (!src || !dst || src->m_scale != dst->m_scale)
 		return false;

+	// If we have an offset, adjust the base positions
+	if (src->m_TEX0.TBP0 != SBP)
+	{
+		GSVector4i offset = TranslateAlignedRectByPage(dst, SBP, SPSM, SBW, GSVector4i(0, 1).xxyy(), false);
+
+		sx += offset.x;
+		sy += offset.y;
+	}
+
+	if (dst->m_TEX0.TBP0 != DBP)
+	{
+		GSVector4i offset = TranslateAlignedRectByPage(dst, DBP, DPSM, DBW, GSVector4i(0, 1).xxyy(), false);
+
+		dx += offset.x;
+		dy += offset.y;
+		req_resize = true;
+	}
+
 	// Scale coordinates.
 	const float scale = src->m_scale;
 	const int scaled_sx = static_cast<int>(sx * scale);
@@ -4842,6 +4886,11 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 	if ((scaled_sx + scaled_w) > src->m_texture->GetWidth() || (scaled_sy + scaled_h) > src->m_texture->GetHeight())
 		return false;

+	if (req_resize)
+	{
+		const GSVector2i target_size = GetTargetSize(DBP, DBW, DPSM, Common::AlignUpPow2(dx + w, 64), dx + h);
+		dst->ResizeTexture(std::max(dst->m_unscaled_size.x, target_size.x), std::max(dst->m_unscaled_size.y, target_size.y));
+	}
 	// We don't want to copy "old" data that the game has overwritten with writes,
 	// so flush any overlapping dirty area.
 	src->UpdateIfDirtyIntersects(GSVector4i(sx, sy, sx + w, sy + h));
@@ -4850,6 +4899,9 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 	// we're not going to be able to use it as a source.
 	dst->Update();

+	// Invalidate any opposite targets.
+	g_texture_cache->InvalidateVideoMemType(GSTextureCache::DepthStencil - dst->m_type, dst->m_TEX0.TBP0);
+
 	// Expand the target when we used a more conservative size.
 	const int required_dh = scaled_dy + scaled_h;
 	if ((scaled_dx + scaled_w) <= dst->m_texture->GetWidth() && required_dh > dst->m_texture->GetHeight())
@@ -4938,9 +4990,30 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 	{
 		if (SPSM == PSMT8H && SPSM == DPSM)
 		{
+			ShaderConvert shader = ShaderConvert::COPY;
+
 			const GSVector4 src_rect = GSVector4(scaled_sx, scaled_sy, scaled_sx + scaled_w, scaled_sy + scaled_h) / (GSVector4(src->m_texture->GetSize()).xyxy());
 			const GSVector4 dst_rect = GSVector4(scaled_dx, scaled_dy, (scaled_dx + scaled_w), (scaled_dy + scaled_h));
-			g_gs_device->StretchRect(src->m_texture, src_rect, dst->m_texture, dst_rect, false, false, false, true);
+			g_gs_device->StretchRect(src->m_texture, src_rect, dst->m_texture, dst_rect, false, false, false, true, shader);
+		}
+		else if (src->m_type != dst->m_type)
+		{
+			ShaderConvert shader = dst->m_type ? ShaderConvert::RGBA8_TO_FLOAT32 : ShaderConvert::FLOAT32_TO_RGBA8;
+
+			switch (dpsm_s.trbpp)
+			{
+				case 24:
+					shader = dst->m_type ? ShaderConvert::RGBA8_TO_FLOAT24 : ShaderConvert::FLOAT32_TO_RGB8;
+					break;
+				case 16:
+					shader = dst->m_type ? ShaderConvert::RGB5A1_TO_FLOAT16 : ShaderConvert::FLOAT16_TO_RGB5A1;
+					break;
+				default:
+					break;
+			}
+			const GSVector4 src_rect = GSVector4(scaled_sx, scaled_sy, scaled_sx + scaled_w, scaled_sy + scaled_h) / (GSVector4(src->m_texture->GetSize()).xyxy());
+			g_gs_device->StretchRect(src->m_texture, src_rect, dst->m_texture, GSVector4(scaled_sx, scaled_sy, scaled_sx + scaled_w, scaled_sy + scaled_h), shader, false);
+
 		}
 		else
 		{
@@ -4968,7 +5041,8 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u
 		dst->m_alpha_range |= src->m_alpha_range;
 	}

-	if (GSLocalMemory::IsPageAligned(src->m_TEX0.PSM, GSVector4i(sx, sy, sx + w, sy + h)) && (w == GSLocalMemory::m_psm[src->m_TEX0.PSM].pgs.x || h == GSLocalMemory::m_psm[src->m_TEX0.PSM].pgs.y))
+	u32 page_mask = GSLocalMemory::IsPageAlignedMasked(src->m_TEX0.PSM, GSVector4i(sx, sy, sx + w, sy + h));
+	if (((page_mask & 0x0f0f) == 0x0f0f || (page_mask & 0xf0f0) == 0xf0f0) && (w == GSLocalMemory::m_psm[src->m_TEX0.PSM].pgs.x || h == GSLocalMemory::m_psm[src->m_TEX0.PSM].pgs.y))
 	{
 		// Vertical Strips
 		if (w == GSLocalMemory::m_psm[src->m_TEX0.PSM].pgs.x)
@@ -5003,7 +5077,7 @@ bool GSTextureCache::Move(u32 SBP, u32 SBW, u32 SPSM, int sx, int sy, u32 DBP, u

 	// Invalidate any sources that overlap with the target (since they're now stale).
 	InvalidateVideoMem(g_gs_renderer->m_mem.GetOffset(DBP, DBW, DPSM), GSVector4i(dx, dy, dx + w, dy + h), false);
-
+	CombineAlignedInsideTargets(dst);
 	return true;
 }

@@ -5190,7 +5264,7 @@ GSTextureCache::Target* GSTextureCache::GetExactTarget(u32 BP, u32 BW, int type,
 	{
 		Target* t = *it;
 		const u32 tgt_bw = std::max(t->m_TEX0.TBW, 1U);
-		if ((t->m_TEX0.TBP0 == BP || (GSConfig.UserHacks_TextureInsideRt >= GSTextureInRtMode::InsideTargets && t->m_TEX0.TBP0 < BP && (((BP - t->m_TEX0.TBP0) >> 5) % tgt_bw) == 0)) && tgt_bw == BW && t->UnwrappedEndBlock() >= end_bp)
+		if ((t->m_TEX0.TBP0 == BP || (GSConfig.UserHacks_TextureInsideRt >= GSTextureInRtMode::InsideTargets && t->m_TEX0.TBP0 < BP && !(BP & 0x1f) && (((BP - t->m_TEX0.TBP0) >> 5) % tgt_bw) == 0)) && tgt_bw == BW && t->UnwrappedEndBlock() >= end_bp)
 		{
 			rts.MoveFront(it.Index());
 			return t;
@@ -5667,7 +5741,7 @@ GSTextureCache::Source* GSTextureCache::CreateSource(const GIFRegTEX0& TEX0, con
 		if (is_8bits)
 		{
 			GL_INS("TC: Reading RT as a packed-indexed 8 bits format");
-			shader = ShaderConvert::RGBA_TO_8I;
+			shader = GSLocalMemory::m_psm[dst->m_TEX0.PSM].bpp == 16 ? ShaderConvert::RGB5A1_TO_8I : ShaderConvert::RGBA_TO_8I;
 		}

 #ifdef ENABLE_OGL_DEBUG
@@ -5737,7 +5811,11 @@ GSTextureCache::Source* GSTextureCache::CreateSource(const GIFRegTEX0& TEX0, con
 			{
 				// We're inside the target, so conversion needs to happen on the entire target so we can offset properly.
 				src->m_unscaled_size.x = dst->m_unscaled_size.x * 2;
-				src->m_unscaled_size.y = dst->m_unscaled_size.y * 2;
+				if (GSLocalMemory::m_psm[dst->m_TEX0.PSM].bpp == 32)
+					src->m_unscaled_size.y = dst->m_unscaled_size.y * 2;
+				else
+					src->m_unscaled_size.y = dst->m_unscaled_size.y;
+
 				new_size.x = src->m_unscaled_size.x;
 				new_size.y = src->m_unscaled_size.y;
 			}
@@ -5908,7 +5986,8 @@ GSTextureCache::Source* GSTextureCache::CreateSource(const GIFRegTEX0& TEX0, con

 					// Adjust to match a PSMT8 texture (coordinates are double C32, we shouldn't be converting from anything else).
 					x_offset *= 2;
-					y_offset *= 2;
+					if (GSLocalMemory::m_psm[dst->m_TEX0.PSM].bpp == 32) 
+						y_offset *= 2;

 					src->m_region.SetX(x_offset, x_offset + tw);
 					src->m_region.SetY(y_offset, y_offset + th);
@@ -6695,6 +6774,8 @@ void GSTextureCache::Read(Target* t, const GSVector4i& r)
 	{
 		case PSMCT32:
 		case PSMCT24:
+		case PSMZ32:
+		case PSMZ24:
 		{
 			// If we're downloading a depth buffer that's been reinterpreted as a color
 			// format, convert it to integer. The format/swizzle is likely wrong, but it's
@@ -6720,27 +6801,11 @@ void GSTextureCache::Read(Target* t, const GSVector4i& r)

 		case PSMCT16:
 		case PSMCT16S:
-		{
-			fmt = GSTexture::Format::UInt16;
-			ps_shader = is_depth ? ShaderConvert::FLOAT32_TO_16_BITS : ShaderConvert::RGBA8_TO_16_BITS;
-			dltex = &m_uint16_download_texture;
-		}
-		break;
-
-		case PSMZ32:
-		case PSMZ24:
-		{
-			fmt = GSTexture::Format::UInt32;
-			ps_shader = ShaderConvert::FLOAT32_TO_32_BITS;
-			dltex = &m_uint32_download_texture;
-		}
-		break;
-
 		case PSMZ16:
 		case PSMZ16S:
 		{
 			fmt = GSTexture::Format::UInt16;
-			ps_shader = ShaderConvert::FLOAT32_TO_16_BITS;
+			ps_shader = is_depth ? ShaderConvert::FLOAT32_TO_16_BITS : ShaderConvert::RGBA8_TO_16_BITS;
 			dltex = &m_uint16_download_texture;
 		}
 		break;
--- a/pcsx2/GS/Renderers/HW/GSTextureReplacements.cpp
+++ b/pcsx2/GS/Renderers/HW/GSTextureReplacements.cpp
@@ -270,56 +270,57 @@ std::string GSTextureReplacements::GetDumpFilename(const TextureName& name, u32
 		return ret;

 	const std::string game_dir(GetGameTextureDirectory());
-	if (!FileSystem::DirectoryExists(game_dir.c_str()))
+	const std::string game_subdir(Path::Combine(game_dir, TEXTURE_DUMP_SUBDIRECTORY_NAME));
+
+	if (!FileSystem::DirectoryExists(game_subdir.c_str()))
 	{
 		// create both dumps and replacements
 		if (!FileSystem::CreateDirectoryPath(game_dir.c_str(), false) ||
-			!FileSystem::EnsureDirectoryExists(Path::Combine(game_dir, "dumps").c_str(), false) ||
-			!FileSystem::EnsureDirectoryExists(Path::Combine(game_dir, "replacements").c_str(), false))
+			!FileSystem::EnsureDirectoryExists(game_subdir.c_str(), false) ||
+			!FileSystem::EnsureDirectoryExists(Path::Combine(game_dir, TEXTURE_REPLACEMENT_SUBDIRECTORY_NAME).c_str(), false))
 		{
 			// if it fails to create, we're not going to be able to use it anyway
 			return ret;
 		}
 	}

-	const std::string game_subdir(Path::Combine(game_dir, TEXTURE_DUMP_SUBDIRECTORY_NAME));
-
 	std::string filename;
 	if (name.HasRegion())
 	{
 		if (name.HasPalette())
 		{
-			filename = (level > 0) ?
-						   StringUtil::StdStringFromFormat(TEXTURE_FILENAME_REGION_CLUT_FORMAT_STRING "-mip%u.png",
-							   name.TEX0Hash, name.CLUTHash, name.region_width, name.region_height, name.bits, level) :
-						   StringUtil::StdStringFromFormat(TEXTURE_FILENAME_REGION_CLUT_FORMAT_STRING ".png",
-							   name.TEX0Hash, name.CLUTHash, name.region_width, name.region_height, name.bits);
+			filename = (level > 0)
+				? StringUtil::StdStringFromFormat(TEXTURE_FILENAME_REGION_CLUT_FORMAT_STRING "-mip%u.png",
+					name.TEX0Hash, name.CLUTHash, name.region_width, name.region_height, name.bits, level)
+				: StringUtil::StdStringFromFormat(TEXTURE_FILENAME_REGION_CLUT_FORMAT_STRING ".png",
+					name.TEX0Hash, name.CLUTHash, name.region_width, name.region_height, name.bits);
 		}
 		else
 		{
-			filename = (level > 0) ? StringUtil::StdStringFromFormat(
-										 TEXTURE_FILENAME_REGION_FORMAT_STRING "-mip%u.png", name.TEX0Hash,
-										 name.region_width, name.region_height, name.bits, level) :
-									 StringUtil::StdStringFromFormat(
-										 TEXTURE_FILENAME_REGION_FORMAT_STRING ".png", name.TEX0Hash,
-										 name.region_width, name.region_height, name.bits);
+			filename = (level > 0)
+				? StringUtil::StdStringFromFormat(TEXTURE_FILENAME_REGION_FORMAT_STRING "-mip%u.png",
+					name.TEX0Hash, name.region_width, name.region_height, name.bits, level)
+				: StringUtil::StdStringFromFormat(TEXTURE_FILENAME_REGION_FORMAT_STRING ".png",
+					name.TEX0Hash, name.region_width, name.region_height, name.bits);
 		}
 	}
 	else
 	{
 		if (name.HasPalette())
 		{
-			filename = (level > 0) ? StringUtil::StdStringFromFormat(TEXTURE_FILENAME_CLUT_FORMAT_STRING "-mip%u.png",
-										 name.TEX0Hash, name.CLUTHash, name.bits, level) :
-									 StringUtil::StdStringFromFormat(TEXTURE_FILENAME_CLUT_FORMAT_STRING ".png",
-										 name.TEX0Hash, name.CLUTHash, name.bits);
+			filename = (level > 0)
+				? StringUtil::StdStringFromFormat(TEXTURE_FILENAME_CLUT_FORMAT_STRING "-mip%u.png",
+				                                  name.TEX0Hash, name.CLUTHash, name.bits, level)
+				: StringUtil::StdStringFromFormat(TEXTURE_FILENAME_CLUT_FORMAT_STRING ".png",
+				                                  name.TEX0Hash, name.CLUTHash, name.bits);
 		}
 		else
 		{
-			filename = (level > 0) ? StringUtil::StdStringFromFormat(
-										 TEXTURE_FILENAME_FORMAT_STRING "-mip%u.png", name.TEX0Hash, name.bits, level) :
-									 StringUtil::StdStringFromFormat(
-										 TEXTURE_FILENAME_FORMAT_STRING ".png", name.TEX0Hash, name.bits);
+			filename = (level > 0)
+				? StringUtil::StdStringFromFormat(TEXTURE_FILENAME_FORMAT_STRING "-mip%u.png",
+				                                  name.TEX0Hash, name.bits, level)
+				: StringUtil::StdStringFromFormat(TEXTURE_FILENAME_FORMAT_STRING ".png",
+				                                  name.TEX0Hash, name.bits);
 		}
 	}

@@ -349,6 +350,28 @@ void GSTextureReplacements::GameChanged()
 	ClearDumpedTextureList();
 }

+/// If the given file exists in the given directory, but with a different case than the original file, write its path to `*output` and return true.
+static bool GetWrongCasePath(std::string* output, const char* dir, std::string_view file, FileSystem::FindResultsArray* reuseme)
+{
+	if (FileSystem::FindFiles(dir, "*", FILESYSTEM_FIND_FOLDERS | FILESYSTEM_FIND_HIDDEN_FILES, reuseme))
+	{
+		for (const FILESYSTEM_FIND_DATA& fd : *reuseme)
+		{
+			std::string_view name = Path::GetFileName(fd.FileName);
+			if (name.size() != file.size())
+				continue;
+			if (0 == strncmp(name.data(), file.data(), name.size()))
+				continue;
+			if (0 == StringUtil::Strncasecmp(name.data(), file.data(), name.size()))
+			{
+				*output = fd.FileName;
+				return true;
+			}
+		}
+	}
+	return false;
+}
+
 void GSTextureReplacements::ReloadReplacementMap()
 {
 	SyncWorkerThread();
@@ -368,9 +391,27 @@ void GSTextureReplacements::ReloadReplacementMap()
 	if (s_current_serial.empty() || !GSConfig.LoadTextureReplacements)
 		return;

-	const std::string replacement_dir(Path::Combine(GetGameTextureDirectory(), TEXTURE_REPLACEMENT_SUBDIRECTORY_NAME));
+	const std::string texture_dir = GetGameTextureDirectory();
+	const std::string replacement_dir(Path::Combine(texture_dir, TEXTURE_REPLACEMENT_SUBDIRECTORY_NAME));

 	FileSystem::FindResultsArray files;
+
+	// For some reason texture pack authors think it's a good idea to rename the replacements directory to something with the wrong case...
+	std::string wrong_case_path;
+	const std::string* right_case_path = nullptr;
+	if (GetWrongCasePath(&wrong_case_path, EmuFolders::Textures.c_str(), s_current_serial, &files))
+		right_case_path = &texture_dir;
+	else if (GetWrongCasePath(&wrong_case_path, texture_dir.c_str(), TEXTURE_REPLACEMENT_SUBDIRECTORY_NAME, &files))
+		right_case_path = &replacement_dir;
+	if (right_case_path)
+	{
+		Host::AddKeyedOSDMessage("TextureReplacementDirCaseMismatch",
+			fmt::format(TRANSLATE_FS("TextureReplacement", "Texture replacement directory {} will not work on case sensitive filesystems.\n"
+			                                               "Rename it to {} to remove this warning."),
+			            wrong_case_path, *right_case_path),
+			Host::OSD_WARNING_DURATION);
+	}
+
 	if (!FileSystem::FindFiles(replacement_dir.c_str(), "*", FILESYSTEM_FIND_FILES | FILESYSTEM_FIND_HIDDEN_FILES | FILESYSTEM_FIND_RECURSIVE, &files))
 		return;

--- a/pcsx2/GS/Renderers/Metal/GSDeviceMTL.mm
+++ b/pcsx2/GS/Renderers/Metal/GSDeviceMTL.mm
@@ -1713,12 +1713,12 @@ void GSDeviceMTL::UpdateCLUTTexture(GSTexture* sTex, float sScale, u32 offsetX,

 void GSDeviceMTL::ConvertToIndexedTexture(GSTexture* sTex, float sScale, u32 offsetX, u32 offsetY, u32 SBW, u32 SPSM, GSTexture* dTex, u32 DBW, u32 DPSM)
 { @autoreleasepool {
-	const ShaderConvert shader = ShaderConvert::RGBA_TO_8I;
+	const ShaderConvert shader = ((SPSM & 0xE) == 0) ? ShaderConvert::RGBA_TO_8I : ShaderConvert::RGB5A1_TO_8I;
 	id<MTLRenderPipelineState> pipeline = m_convert_pipeline[static_cast<int>(shader)];
 	if (!pipeline)
 		[NSException raise:@"StretchRect Missing Pipeline" format:@"No pipeline for %d", static_cast<int>(shader)];

-	GSMTLIndexedConvertPSUniform uniform = { sScale, SBW, DBW };
+	GSMTLIndexedConvertPSUniform uniform = { sScale, SBW, DBW, SPSM };

 	const GSVector4 dRect(0, 0, dTex->GetWidth(), dTex->GetHeight());
 	DoStretchRect(sTex, GSVector4::zero(), dTex, dRect, pipeline, false, LoadAction::DontCareIfFull, &uniform, sizeof(uniform));
--- a/pcsx2/GS/Renderers/Metal/GSMTLDeviceInfo.mm
+++ b/pcsx2/GS/Renderers/Metal/GSMTLDeviceInfo.mm
@@ -4,12 +4,19 @@
 #include "GSMTLDeviceInfo.h"
 #include "GS/GS.h"
 #include "common/Console.h"
+#include "common/Path.h"

 #ifdef __APPLE__

 static id<MTLLibrary> loadMainLibrary(id<MTLDevice> dev, NSString* name)
 {
 	NSString* path = [[NSBundle mainBundle] pathForResource:name ofType:@"metallib"];
+	if (!path)
+	{
+		std::string ssname = std::string([name UTF8String]) + ".metallib";
+		std::string sspath = Path::Combine(EmuFolders::Resources, ssname);
+		path = [[NSString alloc] initWithBytes:sspath.data() length:sspath.length() encoding:NSUTF8StringEncoding];
+	}
 	return path ? [dev newLibraryWithFile:path error:nullptr] : nullptr;
 }

@@ -24,6 +31,8 @@ static MRCOwned<id<MTLLibrary>> loadMainLibrary(id<MTLDevice> dev)
 	if (@available(macOS 10.14, iOS 12.0, *))
 		if (id<MTLLibrary> lib = loadMainLibrary(dev, @"Metal21"))
 			return MRCTransfer(lib);
+	if (id<MTLLibrary> lib = loadMainLibrary(dev, @"default"))
+		return MRCTransfer(lib);
 	return MRCTransfer([dev newDefaultLibrary]);
 }

--- a/pcsx2/GS/Renderers/Metal/GSMTLSharedHeader.h
+++ b/pcsx2/GS/Renderers/Metal/GSMTLSharedHeader.h
@@ -64,6 +64,7 @@ struct GSMTLIndexedConvertPSUniform
 	float scale;
 	uint sbw;
 	uint dbw;
+	uint psm;
 };

 struct GSMTLDownsamplePSUniform
--- a/pcsx2/GS/Renderers/Metal/convert.metal
+++ b/pcsx2/GS/Renderers/Metal/convert.metal
@@ -296,6 +296,121 @@ fragment DepthOut ps_convert_rgb5a1_float16_biln(ConvertShaderData data [[stage_
 	return res.sample_biln<rgb5a1_to_depth16>(data.t);
 }

+fragment float4 ps_convert_rgb5a1_8i(ConvertShaderData data [[stage_in]], DirectReadTextureIn<float> res,
+	constant GSMTLIndexedConvertPSUniform& uniform [[buffer(GSMTLBufferIndexUniforms)]])
+{
+	// Convert a RGB5A1 texture into a 8 bits packed texture
+	// Input column: 16x2 RGB5A1 pixels
+	// 0: 16 RGBA
+	// 1: 16 RGBA
+	// Output column: 16x4 Index pixels
+	// 0: 16 R5G2
+	// 1: 16 R5G2
+	// 2: 16 G2B5A1
+	// 3: 16 G2B5A1
+	uint2 pos = uint2(data.p.xy);
+
+	// Collapse separate R G B A areas into their base pixel
+	uint2 column = (pos & ~uint2(0u, 3u)) / uint2(1,2);
+	uint2 subcolumn = (pos & uint2(0u, 1u));
+	column.x -= (column.x / 128) * 64;
+	column.y += (column.y / 32) * 32;
+
+	// Deal with swizzling differences
+	if ((uniform.psm & 0x8) != 0) // PSMCT16S
+	{
+		if ((pos.x & 32) != 0)
+		{
+			column.y += 32; // 4 columns high times 4 to get bottom 4 blocks
+			column.x &= ~32;
+		}
+		
+		if ((pos.x & 64) != 0)
+		{
+			column.x -= 32;
+		}
+		
+		if (((pos.x & 16) != 0) != ((pos.y & 16) != 0))
+		{
+			column.x ^= 16; 
+			column.y ^= 8;
+		}
+		
+		if ((uniform.psm & 0x30) != 0) // PSMZ16S - Untested but hopefully ok if anything uses it.
+		{
+			column.x ^= 32;
+			column.y ^= 16;
+		}
+	}
+	else // PSMCT16
+	{
+		if ((pos.y & 32) != 0)
+		{
+			column.y -= 16;
+			column.x += 32;
+		}
+		
+		if ((pos.x & 96) != 0)
+		{
+			uint multi = (pos.x & 96) / 32;
+			column.y += 16 * multi; // 4 columns high times 4 to get bottom 4 blocks
+			column.x -= (pos.x & 96);
+		}
+		
+		if (((pos.x & 16) != 0) != ((pos.y & 16) != 0))
+		{
+			column.x ^= 16; 
+			column.y ^= 8;
+		}
+		
+		if ((uniform.psm & 0x30) != 0) // PSMZ16 - Untested but hopefully ok if anything uses it.
+		{
+			column.x ^= 32;
+			column.y ^= 32;
+		}
+	}
+	
+	uint2 coord = column | subcolumn;
+
+	// Compensate for potentially differing page pitch.
+	uint2 block_xy = coord / uint2(64, 64);
+	uint block_num = (block_xy.y * (uniform.dbw / 128)) + block_xy.x;
+	uint2 block_offset = uint2((block_num % (uniform.sbw / 64)) * 64, (block_num / (uniform.sbw / 64)) * 64);
+	coord = (coord % uint2(64, 64)) + block_offset;
+
+	// Apply offset to cols 1 and 2
+	uint is_col23 = pos.y & 4;
+	uint is_col13 = pos.y & 2;
+	uint is_col12 = is_col23 ^ (is_col13 << 1);
+	coord.x ^= is_col12; // If cols 1 or 2, flip bit 3 of x
+
+	if (any(floor(uniform.scale) != uniform.scale))
+		coord = uint2(float2(coord) * uniform.scale);
+	else
+		coord = mul24(coord, uint2(uniform.scale));
+
+	float4 pixel = res.tex.read(coord);
+	
+	uint4 denorm_c = (uint4)(pixel * 255.5f);
+	if ((pos.y & 2u) == 0u)
+	{
+		uint red = (denorm_c.r >> 3) & 0x1Fu;
+		uint green = (denorm_c.g >> 3) & 0x1Fu;
+		float sel0 = (float)(((green << 5) | red) & 0xFF) / 255.0f;
+		
+		return float4(sel0);
+	}
+	else
+	{
+		uint green = (denorm_c.g >> 3) & 0x1Fu;
+		uint blue = (denorm_c.b >> 3) & 0x1Fu;
+		uint alpha = denorm_c.a & 0x80u;
+		float sel0 = (float)((alpha | (blue << 2) | (green >> 3)) & 0xFF) / 255.0f;
+
+		return float4(sel0);
+	}
+}
+
 fragment float4 ps_convert_rgba_8i(ConvertShaderData data [[stage_in]], DirectReadTextureIn<float> res,
 	constant GSMTLIndexedConvertPSUniform& uniform [[buffer(GSMTLBufferIndexUniforms)]])
 {
--- a/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.cpp
+++ b/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.cpp
@@ -350,10 +350,11 @@ bool GSDeviceOGL::Create(GSVSyncMode vsync_mode, bool allow_present_throttle)
 				return false;
 			m_convert.ps[i].SetFormattedName("Convert pipe %s", name);

-			if (static_cast<ShaderConvert>(i) == ShaderConvert::RGBA_TO_8I)
+			if (static_cast<ShaderConvert>(i) == ShaderConvert::RGBA_TO_8I || static_cast<ShaderConvert>(i) == ShaderConvert::RGB5A1_TO_8I)
 			{
 				m_convert.ps[i].RegisterUniform("SBW");
 				m_convert.ps[i].RegisterUniform("DBW");
+				m_convert.ps[i].RegisterUniform("PSM");
 				m_convert.ps[i].RegisterUniform("ScaleFactor");
 			}
 			else if (static_cast<ShaderConvert>(i) == ShaderConvert::YUV)
@@ -1594,12 +1595,13 @@ void GSDeviceOGL::ConvertToIndexedTexture(GSTexture* sTex, float sScale, u32 off
 {
 	CommitClear(sTex, false);

-	const ShaderConvert shader = ShaderConvert::RGBA_TO_8I;
+	const ShaderConvert shader = ((SPSM & 0xE) == 0) ? ShaderConvert::RGBA_TO_8I : ShaderConvert::RGB5A1_TO_8I;
 	GLProgram& prog = m_convert.ps[static_cast<int>(shader)];
 	prog.Bind();
 	prog.Uniform1ui(0, SBW);
 	prog.Uniform1ui(1, DBW);
-	prog.Uniform1f(2, sScale);
+	prog.Uniform1ui(2, SPSM);
+	prog.Uniform1f(3, sScale);

 	OMSetDepthStencilState(m_convert.dss);
 	OMSetBlendState(false);
--- a/pcsx2/GS/Renderers/Vulkan/GSDeviceVK.cpp
+++ b/pcsx2/GS/Renderers/Vulkan/GSDeviceVK.cpp
@@ -3158,15 +3158,16 @@ void GSDeviceVK::ConvertToIndexedTexture(
 	{
 		u32 SBW;
 		u32 DBW;
-		u32 pad1[2];
+		u32 PSM;
+		u32 pad1[1];
 		float ScaleFactor;
 		float pad2[3];
 	};

-	const Uniforms uniforms = {SBW, DBW, {}, sScale, {}};
+	const Uniforms uniforms = {SBW, DBW, SPSM, {}, sScale, {}};
 	SetUtilityPushConstants(&uniforms, sizeof(uniforms));

-	const ShaderConvert shader = ShaderConvert::RGBA_TO_8I;
+	const ShaderConvert shader = ((SPSM & 0xE) == 0) ? ShaderConvert::RGBA_TO_8I : ShaderConvert::RGB5A1_TO_8I;
 	const GSVector4 dRect(0, 0, dTex->GetWidth(), dTex->GetHeight());
 	DoStretchRect(static_cast<GSTextureVK*>(sTex), GSVector4::zero(), static_cast<GSTextureVK*>(dTex), dRect,
 		m_convert[static_cast<int>(shader)], false, true);
--- a/pcsx2/Host/AudioStream.cpp
+++ b/pcsx2/Host/AudioStream.cpp
@@ -857,7 +857,7 @@ void AudioStreamParameters::LoadSave(SettingsWrapper& wrap, const char* section)
 	wrap.EnumEntry(section, "ExpansionMode", expansion_mode, &AudioStream::ParseExpansionMode, &AudioStream::GetExpansionModeName, DEFAULT_EXPANSION_MODE);
 	minimal_output_latency = wrap.EntryBitBool(section, "OutputLatencyMinimal", DEFAULT_OUTPUT_LATENCY_MINIMAL);
 	buffer_ms = static_cast<u16>(std::clamp<int>(wrap.EntryBitfield(section, "BufferMS", buffer_ms, DEFAULT_BUFFER_MS), 0, std::numeric_limits<u16>::max()));
-	output_latency_ms = static_cast<u16>(std::clamp<int>(wrap.EntryBitfield(section, "OutputLatencyMS", buffer_ms, DEFAULT_OUTPUT_LATENCY_MS), 0, std::numeric_limits<u16>::max()));
+	output_latency_ms = static_cast<u16>(std::clamp<int>(wrap.EntryBitfield(section, "OutputLatencyMS", output_latency_ms, DEFAULT_OUTPUT_LATENCY_MS), 0, std::numeric_limits<u16>::max()));

 	stretch_sequence_length_ms = static_cast<u16>(std::clamp<int>(wrap.EntryBitfield(section, "StretchSequenceLengthMS", DEFAULT_STRETCH_SEQUENCE_LENGTH), 0, std::numeric_limits<u16>::max()));
 	stretch_seekwindow_ms = static_cast<u16>(std::clamp<int>(wrap.EntryBitfield(section, "StretchSeekWindowMS", DEFAULT_STRETCH_SEEKWINDOW), 0, std::numeric_limits<u16>::max()));
--- a/pcsx2/Input/InputManager.cpp
+++ b/pcsx2/Input/InputManager.cpp
@@ -1615,7 +1615,7 @@ void InputManager::CloseSources()
 {
 	for (u32 i = FIRST_EXTERNAL_INPUT_SOURCE; i < LAST_EXTERNAL_INPUT_SOURCE; i++)
 	{
-		if (s_input_sources[i]->IsInitialized())
+		if (s_input_sources[i] && s_input_sources[i]->IsInitialized())
 		{
 			s_input_sources[i]->Shutdown();
 		}
--- a/pcsx2/IopBios.cpp
+++ b/pcsx2/IopBios.cpp
@@ -21,6 +21,7 @@
 #include <cstring>
 #include <fmt/format.h>
 #include <sys/stat.h>
+#include <algorithm>

 #include <fcntl.h>

@@ -641,7 +642,7 @@ namespace R3000A

 				for (size_t i = 0; i < handles.size(); i++)
 				{
-					if (handles[i].fd_index == (u32) fd - firstfd)
+					if (handles[i].fd_index == (u32)fd - firstfd)
 					{
 						handles.erase(handles.begin() + i);
 						break;
@@ -928,12 +929,6 @@ namespace R3000A
 	{
 		int Kprintf_HLE()
 		{
-			// Using sprintf here is a bit nasty, but it has a large buffer..
-			// Don't feel like rewriting it.
-#ifdef __clang__
-#pragma clang diagnostic push
-#pragma clang diagnostic ignored "-Wdeprecated-declarations"
-#endif

 			// Emulate the expected Kprintf functionality:
 			iopMemWrite32(sp, a0);
@@ -951,8 +946,12 @@ namespace R3000A
 			if (!ConsoleLogging.iopConsole.IsActive())
 				return 1;

-			char tmp[1024], tmp2[1024];
+			// maximum allowed size for our buffer before we truncate
+			const unsigned int max_len = 4096;
+			char tmp[max_len], tmp2[max_len];
 			char* ptmp = tmp;
+			unsigned int printed_bytes = 0;
+			int remaining_buf = max_len - 1;
 			int n = 1, i = 0, j = 0;

 			while (fmt[i])
@@ -963,21 +962,32 @@ namespace R3000A
 						j = 0;
 						tmp2[j++] = '%';
 					_start:
-						switch (fmt[++i])
+						// let's check whether this is our null terminator
+						// before allowing the parser to proceed
+						if (fmt[i + 1])
 						{
-							case '.':
-							case 'l':
-								tmp2[j++] = fmt[i];
-								goto _start;
-							default:
-								if (fmt[i] >= '0' && fmt[i] <= '9')
-								{
+							switch (fmt[++i])
+							{
+								case '.':
+								case 'l':
+									if (j >= max_len)
+										break;
 									tmp2[j++] = fmt[i];
 									goto _start;
-								}
-								break;
+								default:
+									if (fmt[i] >= '0' && fmt[i] <= '9')
+									{
+										if (j >= max_len)
+											break;
+										tmp2[j++] = fmt[i];
+										goto _start;
+									}
+									break;
+							}
 						}

+						if (j >= max_len)
+							break;
 						tmp2[j++] = fmt[i];
 						tmp2[j] = 0;

@@ -985,7 +995,9 @@ namespace R3000A
 						{
 							case 'f':
 							case 'F':
-								ptmp += sprintf(ptmp, tmp2, (float)iopMemRead32(sp + n * 4));
+								printed_bytes = std::min(remaining_buf, snprintf(ptmp, remaining_buf, tmp2, (float)iopMemRead32(sp + n * 4)));
+								remaining_buf -= printed_bytes;
+								ptmp += printed_bytes;
 								n++;
 								break;

@@ -995,7 +1007,9 @@ namespace R3000A
 							case 'E':
 							case 'g':
 							case 'G':
-								ptmp += sprintf(ptmp, tmp2, (double)iopMemRead32(sp + n * 4));
+								printed_bytes = std::min(remaining_buf, snprintf(ptmp, remaining_buf, tmp2, (double)iopMemRead32(sp + n * 4)));
+								remaining_buf -= printed_bytes;
+								ptmp += printed_bytes;
 								n++;
 								break;

@@ -1007,19 +1021,27 @@ namespace R3000A
 							case 'O':
 							case 'x':
 							case 'X':
-								ptmp += sprintf(ptmp, tmp2, (u32)iopMemRead32(sp + n * 4));
+							case 'u':
+							case 'U':
+								printed_bytes = std::min(remaining_buf, snprintf(ptmp, remaining_buf, tmp2, (u32)iopMemRead32(sp + n * 4)));
+								remaining_buf -= printed_bytes;
+								ptmp += printed_bytes;
 								n++;
 								break;

 							case 'c':
-								ptmp += sprintf(ptmp, tmp2, (u8)iopMemRead32(sp + n * 4));
+								printed_bytes = std::min(remaining_buf, snprintf(ptmp, remaining_buf, tmp2, (u8)iopMemRead32(sp + n * 4)));
+								remaining_buf -= printed_bytes;
+								ptmp += printed_bytes;
 								n++;
 								break;

 							case 's':
 							{
 								std::string s = iopMemReadString(iopMemRead32(sp + n * 4));
-								ptmp += sprintf(ptmp, tmp2, s.data());
+								printed_bytes = std::min(remaining_buf, snprintf(ptmp, remaining_buf, tmp2, s.data()));
+								remaining_buf -= printed_bytes;
+								ptmp += printed_bytes;
 								n++;
 							}
 							break;
@@ -1043,10 +1065,6 @@ namespace R3000A
 			iopConLog(ShiftJIS_ConvertString(tmp, 1023));

 			return 1;
-
-#ifdef __clang__
-#pragma clang diagnostic pop
-#endif
 		}
 	} // namespace sysmem

--- a/pcsx2/Pcsx2Config.cpp
+++ b/pcsx2/Pcsx2Config.cpp
@@ -2084,17 +2084,16 @@ void Pcsx2Config::ClearInvalidPerGameConfiguration(SettingsInterface* si)
 void EmuFolders::SetAppRoot()
 {
 	std::string program_path = FileSystem::GetProgramPath();
+	Console.WriteLnFmt("Program Path: {}", program_path);
+	AppRoot = Path::Canonicalize(Path::GetDirectory(program_path));
 #ifdef __APPLE__
 	const auto bundle_path = CocoaTools::GetNonTranslocatedBundlePath();
 	if (bundle_path.has_value())
 	{
 		// On macOS, override with the bundle path if launched from a bundle.
-		program_path = bundle_path.value();
+		AppRoot = StringUtil::EndsWithNoCase(*bundle_path, ".app") ? Path::GetDirectory(*bundle_path) : *bundle_path;
 	}
 #endif
-	Console.WriteLnFmt("Program Path: {}", program_path);
-
-	AppRoot = Path::Canonicalize(Path::GetDirectory(program_path));

 	// logging of directories in case something goes wrong super early
 	Console.WriteLnFmt("AppRoot Directory: {}", AppRoot);
@@ -2111,8 +2110,10 @@ bool EmuFolders::SetResourcesDirectory()
 #endif
 #else
 	// On macOS, this is in the bundle resources directory.
-	const std::string program_path = FileSystem::GetProgramPath();
-	Resources = Path::Canonicalize(Path::Combine(Path::GetDirectory(program_path), "../Resources"));
+	if (auto resources = CocoaTools::GetResourcePath())
+		Resources = *resources;
+	else
+		Resources = Path::Combine(AppRoot, "resources");
 #endif

 	Console.WriteLnFmt("Resources Directory: {}", Resources);
--- a/pcsx2/ShaderCacheVersion.h
+++ b/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 = 65;
+static constexpr u32 SHADER_CACHE_VERSION = 66;
Author	SHA1	Message	Date
lightningterror	c4a8fc5b71	Qt: Fix Winconsistent-missing-override warning.	2025-06-11 23:17:53 +02:00
GovanifY	8eb46b5a4c	IopBios: remove clang deprecated carveout we use snprintf now, so this isn't useful anymore.	2025-06-10 20:59:25 +02:00
GovanifY	8be16d1039	IopBios: do not overflow snprintf tmp buffer We could otherwise overflow as snprintf does not return the number of written bytes but the number of written bytes assuming an infinite buffer.	2025-06-10 20:59:25 +02:00
$refractionpcsx2$ refractionpcsx2	76e6208d1b	GS/TC: Fix region for tex in rt depth	2025-06-10 20:12:41 +02:00
$refractionpcsx2$ refractionpcsx2	bc09080ba5	GS/TC: Remove old inside check from source lookup	2025-06-10 20:12:41 +02:00
$refractionpcsx2$ refractionpcsx2	a6eb257a3a	GS/HW: Allow RT in RT offset on Z buffer if FRAME is disabled	2025-06-10 20:11:32 +02:00
PCSX2 Bot	e62450d255	[ci skip] PAD: Update to latest controller database.	2025-06-10 20:10:22 +02:00
GovanifY	1aa922f700	IopBios: allow %u string formatting in IOP kprintf	2025-06-10 14:47:27 +02:00
GovanifY	4c9d2f99b1	IopBios: truncate printf output if bigger than our buffer I also increased the buffer size while we are at it to avoid breakage, despite the obvious unintended allowed situations that we allowed up until now Reported-By: Michael Lappas	2025-06-10 14:47:27 +02:00
Ty Lamontagne	f6e899b570	Debugger: Update search results when we are _not_ going to remove them	2025-06-08 15:44:56 -04:00
$refractionpcsx2$ refractionpcsx2	695c39fba2	GS/HW: Remove channel shuffle override from Namco CRC hack	2025-06-08 16:05:24 +02:00
$refractionpcsx2$ refractionpcsx2	e5616cff98	GS/HW: Improve shuffle detection robustness	2025-06-08 13:51:19 +02:00
$refractionpcsx2$ refractionpcsx2	76d5994c1e	GS/TC: Improve rect block offset calculation for translation and invalidation	2025-06-08 13:51:19 +02:00
$refractionpcsx2$ refractionpcsx2	cbb40832a1	GS/HW: Correct block offset target usage	2025-06-08 13:51:19 +02:00
$refractionpcsx2$ refractionpcsx2	e4bdcde1ca	GS/HW: Remove CRC hacks for Kunoichi and Sakura Wars	2025-06-08 13:51:19 +02:00
$refractionpcsx2$ refractionpcsx2	863e119ff4	GS/HW: Allow conversion of colour to Z formats during HW move	2025-06-08 13:51:19 +02:00
$refractionpcsx2$ refractionpcsx2	2ccf6dc872	GS: Add IsPageAlignedMasked to return an alignment mask	2025-06-08 13:51:19 +02:00
$refractionpcsx2$ refractionpcsx2	24ebf1b4f1	GS/HW: Fix bug in target preloading causing misaligned base addresses	2025-06-08 13:51:19 +02:00
$refractionpcsx2$ refractionpcsx2	ed2832434c	GS/HW: Don't look up block offset targets on Exact target lookup	2025-06-08 13:51:19 +02:00
TJnotJT	5e160fca8f	Tools: Small fix to gsrunner args.	2025-06-06 18:05:37 -04:00
JordanTheToaster	a0ef82e221	AudioStream: Fix typo with default output latency	2025-06-06 18:03:29 -04:00
icup321	3e7ac3d66c	GameDB: Add native scaling to Scooby-Doo! Unmasked	2025-06-06 23:09:37 +02:00
JordanTheToaster	730e6fa737	GameDB: Simple 2000 Vol 92 Fixes	2025-06-06 23:09:26 +02:00
PCSX2 Bot	cdfcd9fddd	[ci skip] Qt: Update Base Translation.	2025-06-06 02:31:08 +02:00
$refractionpcsx2$ refractionpcsx2	b6930c10b9	GS/HW: Clean up target download formats	2025-06-05 13:51:21 +02:00
$refractionpcsx2$ refractionpcsx2	852734580d	GameDB: Update Harry Potter fixes.	2025-06-05 13:51:21 +02:00
$refractionpcsx2$ refractionpcsx2	d1dc6a9c1d	GS/HW: Add 16bit to 8bit conversion shader	2025-06-05 13:51:21 +02:00
Ty	4fa6d3ed3f	GSRunner: Add type to shutdown message code	2025-06-04 20:25:32 -04:00
TellowKrinkle	92e190ad6c	GSRunner: Fix surfaceless run on macOS	2025-06-04 20:25:32 -04:00
TellowKrinkle	da4fcffef4	Input: Fix crash when shutting down without initializing input	2025-06-04 20:25:32 -04:00
TellowKrinkle	1a5731dd8e	MacOS: Better handle directories of non-bundle applications	2025-06-04 20:25:32 -04:00
TellowKrinkle	e764c5cd4e	GSRunner: macOS support	2025-06-04 20:25:32 -04:00
TellowKrinkle	e23b247947	GSRunner: Use separate CPU thread	2025-06-04 20:25:32 -04:00
PCSX2 Bot	3d7792436f	[ci skip] Qt: Update Base Translation.	2025-06-04 20:03:38 -04:00
chaoticgd	d8187fbea4	Deps: Specify minimum version of KDDockWidgets	2025-06-04 20:02:02 -04:00
chaoticgd	02259ad0a5	Debugger: Add include required for older versions of KDDockWidgets	2025-06-04 20:02:02 -04:00
TellowKrinkle	220a68df9a	GS: Warn on texture replacement folder with wrong case	2025-06-04 19:58:41 -04:00
TellowKrinkle	2ced24f69e	GS: Create texture dump directory if it doesn't exist	2025-06-04 19:58:41 -04:00
TellowKrinkle	ec91d0dc74	GS: Formatting	2025-06-04 19:58:41 -04:00
TheLastRar	46874f4673	Qt: Add workaround for incorrectly tinted icons after theme switch	2025-06-04 19:39:18 -04:00
TheLastRar	9eac47dc6c	Qt: Fix selected gamelist icons being wrong colour after theme switch	2025-06-04 19:39:18 -04:00
RedDevilus	9e3fd5c2e0	CI: Fix flatpak Try to fix build failing	2025-06-04 20:19:46 +01:00