13 files changed, 118 insertions, 12 deletions

diff --git a/.ci/scripts/clang/docker.sh b/.ci/scripts/clang/docker.sh
new file mode 100755
index 000000000..885d74e97
--- /dev/null
+++ b/.ci/scripts/clang/docker.sh
@@ -0,0 +1,18 @@
+#!/bin/bash -ex
+
+# Exit on error, rather than continuing with the rest of the script.
+set -e
+
+cd /yuzu
+
+ccache -s
+
+mkdir build || true && cd build
+cmake .. -DDISPLAY_VERSION=$1 -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=/usr/lib/ccache/clang -DCMAKE_CXX_COMPILER=/usr/lib/ccache/clang++ -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${ENABLE_COMPATIBILITY_REPORTING:-"OFF"} -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DUSE_DISCORD_PRESENCE=ON -DENABLE_QT_TRANSLATION=ON -DCMAKE_INSTALL_PREFIX="/usr"
+
+make -j$(nproc)
+
+ccache -s
+
+ctest -VV -C Release
+
diff --git a/.ci/scripts/clang/exec.sh b/.ci/scripts/clang/exec.sh
new file mode 100644
index 000000000..e56cd4325
--- /dev/null
+++ b/.ci/scripts/clang/exec.sh
@@ -0,0 +1,8 @@
+#!/bin/bash -ex
+
+mkdir -p "ccache"  || true
+chmod a+x ./.ci/scripts/clang/docker.sh
+# the UID for the container yuzu user is 1027
+sudo chown -R 1027 ./
+docker run -e ENABLE_COMPATIBILITY_REPORTING -e CCACHE_DIR=/yuzu/ccache -v $(pwd):/yuzu yuzuemu/build-environments:linux-fresh /bin/bash /yuzu/.ci/scripts/clang/docker.sh $1
+sudo chown -R $UID ./
diff --git a/.ci/scripts/clang/upload.sh b/.ci/scripts/clang/upload.sh
new file mode 100644
index 000000000..fe4e6b2ac
--- /dev/null
+++ b/.ci/scripts/clang/upload.sh
@@ -0,0 +1,20 @@
+#!/bin/bash -ex
+
+. .ci/scripts/common/pre-upload.sh
+
+REV_NAME="yuzu-linux-${GITDATE}-${GITREV}"
+ARCHIVE_NAME="${REV_NAME}.tar.xz"
+COMPRESSION_FLAGS="-cJvf"
+
+if [ "${RELEASE_NAME}" = "mainline" ]; then
+    DIR_NAME="${REV_NAME}"
+else
+    DIR_NAME="${REV_NAME}_${RELEASE_NAME}"
+fi
+
+mkdir "$DIR_NAME"
+
+cp build/bin/yuzu-cmd "$DIR_NAME"
+cp build/bin/yuzu "$DIR_NAME"
+
+. .ci/scripts/common/post-upload.sh
diff --git a/.ci/templates/build-standard.yml b/.ci/templates/build-standard.yml
index 7422c8346..57d36f813 100644
--- a/.ci/templates/build-standard.yml
+++ b/.ci/templates/build-standard.yml
@@ -12,6 +12,9 @@ jobs:
       windows:
         BuildSuffix: 'windows-mingw'
         ScriptFolder: 'windows'
+      clang:
+        BuildSuffix: 'clang'
+        ScriptFolder: 'clang'
       linux:
         BuildSuffix: 'linux'
         ScriptFolder: 'linux'
@@ -24,4 +27,4 @@ jobs:
     parameters:
       artifactSource: 'false'
       cache: $(parameters.cache)
-      version: $(parameters.version)
\ No newline at end of file
+      version: $(parameters.version)
diff --git a/externals/glad/include/glad/glad.h b/externals/glad/include/glad/glad.h
index 6e16358ea..191bb9fcb 100644
--- a/externals/glad/include/glad/glad.h
+++ b/externals/glad/include/glad/glad.h
@@ -5156,6 +5156,9 @@ GLAPI PFNGLDEPTHRANGEARRAYVPROC glad_glDepthRangeArrayv;
 typedef void (APIENTRYP PFNGLDEPTHRANGEINDEXEDPROC)(GLuint index, GLdouble n, GLdouble f);
 GLAPI PFNGLDEPTHRANGEINDEXEDPROC glad_glDepthRangeIndexed;
 #define glDepthRangeIndexed glad_glDepthRangeIndexed
+typedef void (APIENTRYP PFNGLDEPTHRANGEINDEXEDDNVPROC)(GLuint index, GLdouble n, GLdouble f);
+GLAPI PFNGLDEPTHRANGEINDEXEDDNVPROC glad_glDepthRangeIndexeddNV;
+#define glDepthRangeIndexeddNV glad_glDepthRangeIndexeddNV
 typedef void (APIENTRYP PFNGLGETFLOATI_VPROC)(GLenum target, GLuint index, GLfloat *data);
 GLAPI PFNGLGETFLOATI_VPROC glad_glGetFloati_v;
 #define glGetFloati_v glad_glGetFloati_v
diff --git a/externals/glad/src/glad.c b/externals/glad/src/glad.c
index d3e13163f..7b24cd68d 100644
--- a/externals/glad/src/glad.c
+++ b/externals/glad/src/glad.c
@@ -1044,6 +1044,7 @@ PFNGLDEPTHMASKPROC glad_glDepthMask = NULL;
 PFNGLDEPTHRANGEPROC glad_glDepthRange = NULL;
 PFNGLDEPTHRANGEARRAYVPROC glad_glDepthRangeArrayv = NULL;
 PFNGLDEPTHRANGEINDEXEDPROC glad_glDepthRangeIndexed = NULL;
+PFNGLDEPTHRANGEINDEXEDDNVPROC glad_glDepthRangeIndexeddNV = NULL;
 PFNGLDEPTHRANGEFPROC glad_glDepthRangef = NULL;
 PFNGLDETACHSHADERPROC glad_glDetachShader = NULL;
 PFNGLDISABLEPROC glad_glDisable = NULL;
@@ -7971,6 +7972,7 @@ static void load_GL_NV_depth_buffer_float(GLADloadproc load) {
         glad_glDepthRangedNV = (PFNGLDEPTHRANGEDNVPROC)load("glDepthRangedNV");
         glad_glClearDepthdNV = (PFNGLCLEARDEPTHDNVPROC)load("glClearDepthdNV");
         glad_glDepthBoundsdNV = (PFNGLDEPTHBOUNDSDNVPROC)load("glDepthBoundsdNV");
+        glad_glDepthRangeIndexeddNV = (PFNGLDEPTHRANGEINDEXEDDNVPROC)load("glDepthRangeIndexeddNV");
 }
 static void load_GL_NV_draw_texture(GLADloadproc load) {
         if(!GLAD_GL_NV_draw_texture) return;
diff --git a/src/core/hle/service/acc/acc.cpp b/src/core/hle/service/acc/acc.cpp
index 3ec0e1eca..615e20a54 100644
--- a/src/core/hle/service/acc/acc.cpp
+++ b/src/core/hle/service/acc/acc.cpp
@@ -508,7 +508,7 @@ public:
             {1, &IManagerForApplication::GetAccountId, "GetAccountId"},
             {2, nullptr, "EnsureIdTokenCacheAsync"},
             {3, nullptr, "LoadIdTokenCache"},
-            {130, nullptr, "GetNintendoAccountUserResourceCacheForApplication"},
+            {130, &IManagerForApplication::GetNintendoAccountUserResourceCacheForApplication, "GetNintendoAccountUserResourceCacheForApplication"},
             {150, nullptr, "CreateAuthorizationRequest"},
             {160, &IManagerForApplication::StoreOpenContext, "StoreOpenContext"},
             {170, nullptr, "LoadNetworkServiceLicenseKindAsync"},
@@ -534,6 +534,22 @@ private:
         rb.PushRaw<u64>(user_id.GetNintendoID());
     }
 
+    void GetNintendoAccountUserResourceCacheForApplication(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_ACC, "(STUBBED) called");
+
+        std::vector<u8> nas_user_base_for_application(0x68);
+        ctx.WriteBuffer(nas_user_base_for_application, 0);
+
+        if (ctx.CanWriteBuffer(1)) {
+            std::vector<u8> unknown_out_buffer(ctx.GetWriteBufferSize(1));
+            ctx.WriteBuffer(unknown_out_buffer, 1);
+        }
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushRaw<u64>(user_id.GetNintendoID());
+    }
+
     void StoreOpenContext(Kernel::HLERequestContext& ctx) {
         LOG_WARNING(Service_ACC, "(STUBBED) called");
         IPC::ResponseBuilder rb{ctx, 2};
diff --git a/src/video_core/renderer_opengl/gl_device.cpp b/src/video_core/renderer_opengl/gl_device.cpp
index 48d5c4a5e..1ae5f1d62 100644
--- a/src/video_core/renderer_opengl/gl_device.cpp
+++ b/src/video_core/renderer_opengl/gl_device.cpp
@@ -239,6 +239,7 @@ Device::Device() {
     has_nv_viewport_array2 = GLAD_GL_NV_viewport_array2;
     has_vertex_buffer_unified_memory = GLAD_GL_NV_vertex_buffer_unified_memory;
     has_debugging_tool_attached = IsDebugToolAttached(extensions);
+    has_depth_buffer_float = HasExtension(extensions, "GL_NV_depth_buffer_float");
 
     // At the moment of writing this, only Nvidia's driver optimizes BufferSubData on exclusive
     // uniform buffers as "push constants"
@@ -275,6 +276,7 @@ Device::Device(std::nullptr_t) {
     has_image_load_formatted = true;
     has_texture_shadow_lod = true;
     has_variable_aoffi = true;
+    has_depth_buffer_float = true;
 }
 
 bool Device::TestVariableAoffi() {
diff --git a/src/video_core/renderer_opengl/gl_device.h b/src/video_core/renderer_opengl/gl_device.h
index ee053776d..f24bd0c7b 100644
--- a/src/video_core/renderer_opengl/gl_device.h
+++ b/src/video_core/renderer_opengl/gl_device.h
@@ -122,6 +122,10 @@ public:
         return use_driver_cache;
     }
 
+    bool HasDepthBufferFloat() const {
+        return has_depth_buffer_float;
+    }
+
 private:
     static bool TestVariableAoffi();
     static bool TestPreciseBug();
@@ -150,6 +154,7 @@ private:
     bool use_assembly_shaders{};
     bool use_asynchronous_shaders{};
     bool use_driver_cache{};
+    bool has_depth_buffer_float{};
 };
 
 } // namespace OpenGL
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 418644108..4610fd160 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -889,7 +889,11 @@ void RasterizerOpenGL::SyncViewport() {
             const GLdouble reduce_z = regs.depth_mode == Maxwell::DepthMode::MinusOneToOne;
             const GLdouble near_depth = src.translate_z - src.scale_z * reduce_z;
             const GLdouble far_depth = src.translate_z + src.scale_z;
-            glDepthRangeIndexed(static_cast<GLuint>(i), near_depth, far_depth);
+            if (device.HasDepthBufferFloat()) {
+                glDepthRangeIndexeddNV(static_cast<GLuint>(i), near_depth, far_depth);
+            } else {
+                glDepthRangeIndexed(static_cast<GLuint>(i), near_depth, far_depth);
+            }
 
             if (!GLAD_GL_NV_viewport_swizzle) {
                 continue;
diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
index 40e2e0d38..c6846d886 100644
--- a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
+++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
@@ -1845,13 +1845,21 @@ private:
 
     Expression TextureGather(Operation operation) {
         const auto& meta = std::get<MetaTexture>(operation.GetMeta());
-        UNIMPLEMENTED_IF(!meta.aoffi.empty());
 
         const Id coords = GetCoordinates(operation, Type::Float);
+
+        spv::ImageOperandsMask mask = spv::ImageOperandsMask::MaskNone;
+        std::vector<Id> operands;
         Id texture{};
+
+        if (!meta.aoffi.empty()) {
+            mask = mask | spv::ImageOperandsMask::Offset;
+            operands.push_back(GetOffsetCoordinates(operation));
+        }
+
         if (meta.sampler.is_shadow) {
             texture = OpImageDrefGather(t_float4, GetTextureSampler(operation), coords,
-                                        AsFloat(Visit(meta.depth_compare)));
+                                        AsFloat(Visit(meta.depth_compare)), mask, operands);
         } else {
             u32 component_value = 0;
             if (meta.component) {
@@ -1860,7 +1868,7 @@ private:
                 component_value = component->GetValue();
             }
             texture = OpImageGather(t_float4, GetTextureSampler(operation), coords,
-                                    Constant(t_uint, component_value));
+                                    Constant(t_uint, component_value), mask, operands);
         }
         return GetTextureElement(operation, texture, Type::Float);
     }
@@ -1928,13 +1936,22 @@ private:
 
         const Id image = GetTextureImage(operation);
         const Id coords = GetCoordinates(operation, Type::Int);
+
+        spv::ImageOperandsMask mask = spv::ImageOperandsMask::MaskNone;
+        std::vector<Id> operands;
         Id fetch;
+
         if (meta.lod && !meta.sampler.is_buffer) {
-            fetch = OpImageFetch(t_float4, image, coords, spv::ImageOperandsMask::Lod,
-                                 AsInt(Visit(meta.lod)));
-        } else {
-            fetch = OpImageFetch(t_float4, image, coords);
+            mask = mask | spv::ImageOperandsMask::Lod;
+            operands.push_back(AsInt(Visit(meta.lod)));
+        }
+
+        if (!meta.aoffi.empty()) {
+            mask = mask | spv::ImageOperandsMask::Offset;
+            operands.push_back(GetOffsetCoordinates(operation));
         }
+
+        fetch = OpImageFetch(t_float4, image, coords, mask, operands);
         return GetTextureElement(operation, fetch, Type::Float);
     }
 
diff --git a/src/video_core/shader/decode/memory.cpp b/src/video_core/shader/decode/memory.cpp
index 50f4e7d35..7728f600e 100644
--- a/src/video_core/shader/decode/memory.cpp
+++ b/src/video_core/shader/decode/memory.cpp
@@ -330,6 +330,7 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) {
         case StoreType::Bits32:
             (this->*set_memory)(bb, GetAddress(0), GetRegister(instr.gpr0));
             break;
+        case StoreType::Unsigned16:
         case StoreType::Signed16: {
             Node address = GetAddress(0);
             Node memory = (this->*get_memory)(address);
diff --git a/src/video_core/shader/decode/texture.cpp b/src/video_core/shader/decode/texture.cpp
index 833fa2a39..c69681e8d 100644
--- a/src/video_core/shader/decode/texture.cpp
+++ b/src/video_core/shader/decode/texture.cpp
@@ -806,6 +806,7 @@ Node4 ShaderIR::GetTldsCode(Instruction instr, TextureType texture_type, bool is
 
     const std::size_t type_coord_count = GetCoordCount(texture_type);
     const bool lod_enabled = instr.tlds.GetTextureProcessMode() == TextureProcessMode::LL;
+    const bool aoffi_enabled = instr.tlds.UsesMiscMode(TextureMiscMode::AOFFI);
 
     // If enabled arrays index is always stored in the gpr8 field
     const u64 array_register = instr.gpr8.Value();
@@ -820,17 +821,23 @@ Node4 ShaderIR::GetTldsCode(Instruction instr, TextureType texture_type, bool is
     std::vector<Node> coords;
     for (std::size_t i = 0; i < type_coord_count; ++i) {
         const bool last = (i == (type_coord_count - 1)) && (type_coord_count > 1);
-        coords.push_back(GetRegister(last ? last_coord_register : coord_register + i));
+        coords.push_back(
+            GetRegister(last && !aoffi_enabled ? last_coord_register : coord_register + i));
     }
 
     const Node array = is_array ? GetRegister(array_register) : nullptr;
     // When lod is used always is in gpr20
     const Node lod = lod_enabled ? GetRegister(instr.gpr20) : Immediate(0);
 
+    std::vector<Node> aoffi;
+    if (aoffi_enabled) {
+        aoffi = GetAoffiCoordinates(GetRegister(instr.gpr20), type_coord_count, false);
+    }
+
     Node4 values;
     for (u32 element = 0; element < values.size(); ++element) {
         auto coords_copy = coords;
-        MetaTexture meta{*sampler, array, {}, {}, {}, {}, {}, lod, {}, element, {}};
+        MetaTexture meta{*sampler, array, {}, aoffi, {}, {}, {}, lod, {}, element, {}};
         values[element] = Operation(OperationCode::TexelFetch, meta, std::move(coords_copy));
     }
     return values;