Update core.cpp

Fix yuzu crash on pause in single-thread mode #544

.travis.yml

@@ -42,7 +42,3 @@ notifications:
   webhooks:
     urls:
       - https://api.yuzu-emu.org/code/travis/notify
-
-cache:
-  directories:
-    - $HOME/.ccache

.travis/linux/build.sh

@@ -1,3 +1,3 @@
 #!/bin/bash -ex
 
-docker run -e CCACHE_DIR=/ccache -v $HOME/.ccache:/ccache -v $(pwd):/yuzu ubuntu:18.04 /bin/bash /yuzu/.travis/linux/docker.sh
+docker run -v $(pwd):/yuzu ubuntu:18.04 /bin/bash /yuzu/.travis/linux/docker.sh

.travis/linux/docker.sh

@@ -1,18 +1,16 @@
 #!/bin/bash -ex
 
 apt-get update
-apt-get install --no-install-recommends -y build-essential git libqt5opengl5-dev libsdl2-dev libssl-dev python qtbase5-dev wget cmake ninja-build ccache
+apt-get install -y build-essential git libqt5opengl5-dev libsdl2-dev libssl-dev python qtbase5-dev wget ninja-build
+
+# Get a recent version of CMake
+wget https://cmake.org/files/v3.10/cmake-3.10.1-Linux-x86_64.sh
+sh cmake-3.10.1-Linux-x86_64.sh --exclude-subdir --prefix=/ --skip-license
 
 cd /yuzu
 
-export PATH=/usr/lib/ccache:$PATH
-ln -sf /usr/bin/ccache /usr/lib/ccache/cc
-ln -sf /usr/bin/ccache /usr/lib/ccache/c++
 mkdir build && cd build
-ccache --show-stats > ccache_before
 cmake .. -DYUZU_BUILD_UNICORN=ON -DCMAKE_BUILD_TYPE=Release -G Ninja
 ninja
-ccache --show-stats > ccache_after
-diff -U100 ccache_before ccache_after || true
 
 ctest -VV -C Release

.travis/macos/build.sh

@@ -7,12 +7,8 @@ export Qt5_DIR=$(brew --prefix)/opt/qt5
 export UNICORNDIR=$(pwd)/externals/unicorn
 
 mkdir build && cd build
-export PATH=/usr/local/opt/ccache/libexec:$PATH
-ccache --show-stats > ccache_before
 cmake --version
 cmake .. -DYUZU_BUILD_UNICORN=ON -DCMAKE_BUILD_TYPE=Release
 make -j4
-ccache --show-stats > ccache_after
-diff -U100 ccache_before ccache_after || true
 
 ctest -VV -C Release

.travis/macos/deps.sh

@@ -1,5 +1,5 @@
 #!/bin/sh -ex
 
 brew update
-brew install dylibbundler p7zip qt5 sdl2 ccache
+brew install dylibbundler p7zip qt5 sdl2
 brew outdated cmake || brew upgrade cmake

src/common/string_util.cpp

@@ -64,10 +64,6 @@ std::string ArrayToString(const u8* data, size_t size, int line_len, bool spaces
     return oss.str();
 }
 
-std::string StringFromBuffer(const std::vector<u8>& data) {
-    return std::string(data.begin(), std::find(data.begin(), data.end(), '\0'));
-}
-
 // Turns "  hej " into "hej". Also handles tabs.
 std::string StripSpaces(const std::string& str) {
     const size_t s = str.find_first_not_of(" \t\r\n");

src/common/string_util.h

@@ -21,8 +21,6 @@ std::string ToUpper(std::string str);
 
 std::string ArrayToString(const u8* data, size_t size, int line_len = 20, bool spaces = true);
 
-std::string StringFromBuffer(const std::vector<u8>& data);
-
 std::string StripSpaces(const std::string& s);
 std::string StripQuotes(const std::string& s);
 

src/core/core.cpp

@@ -44,14 +44,14 @@ Cpu& System::CurrentCpuCore() {
     }
 
     // Otherwise, use single-threaded mode active_core variable
-    return *cpu_cores[active_core];
+     return *cpu_cores[active_core = 4]; 
 }
 
 System::ResultStatus System::RunLoop(bool tight_loop) {
     status = ResultStatus::Success;
 
     // Update thread_to_cpu in case Core 0 is run from a different host thread
-    thread_to_cpu[std::this_thread::get_id()] = cpu_cores[0];
+    thread_to_cpu[std::this_thread::get_id()] = cpu_cores[4];
 
     if (GDBStub::IsServerEnabled()) {
         GDBStub::HandlePacket();
@@ -68,7 +68,7 @@ System::ResultStatus System::RunLoop(bool tight_loop) {
         }
     }
 
-    for (active_core = 0; active_core < NUM_CPU_CORES; ++active_core) {
+    for (active_core = 4; active_core < NUM_CPU_CORES; ++active_core) { 
         cpu_cores[active_core]->RunLoop(tight_loop);
         if (Settings::values.use_multi_core) {
             // Cores 1-3 are run on other threads in this mode

src/core/hle/service/filesystem/fsp_srv.cpp

@@ -4,7 +4,6 @@
 
 #include <cinttypes>
 #include "common/logging/log.h"
-#include "common/string_util.h"
 #include "core/core.h"
 #include "core/file_sys/directory.h"
 #include "core/file_sys/filesystem.h"
@@ -259,7 +258,9 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
+
+        std::string name(file_buffer.begin(), end);
 
         u64 mode = rp.Pop<u64>();
         u32 size = rp.Pop<u32>();
@@ -274,7 +275,9 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
+
+        std::string name(file_buffer.begin(), end);
 
         NGLOG_DEBUG(Service_FS, "called file {}", name);
 
@@ -286,7 +289,9 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
+
+        std::string name(file_buffer.begin(), end);
 
         NGLOG_DEBUG(Service_FS, "called directory {}", name);
 
@@ -300,11 +305,13 @@ public:
         std::vector<u8> buffer;
         buffer.resize(ctx.BufferDescriptorX()[0].Size());
         Memory::ReadBlock(ctx.BufferDescriptorX()[0].Address(), buffer.data(), buffer.size());
-        std::string src_name = Common::StringFromBuffer(buffer);
+        auto end = std::find(buffer.begin(), buffer.end(), '\0');
+        std::string src_name(buffer.begin(), end);
 
         buffer.resize(ctx.BufferDescriptorX()[1].Size());
         Memory::ReadBlock(ctx.BufferDescriptorX()[1].Address(), buffer.data(), buffer.size());
-        std::string dst_name = Common::StringFromBuffer(buffer);
+        end = std::find(buffer.begin(), buffer.end(), '\0');
+        std::string dst_name(buffer.begin(), end);
 
         NGLOG_DEBUG(Service_FS, "called file '{}' to file '{}'", src_name, dst_name);
 
@@ -316,7 +323,9 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
+
+        std::string name(file_buffer.begin(), end);
 
         auto mode = static_cast<FileSys::Mode>(rp.Pop<u32>());
 
@@ -340,7 +349,9 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
+
+        std::string name(file_buffer.begin(), end);
 
         // TODO(Subv): Implement this filter.
         u32 filter_flags = rp.Pop<u32>();
@@ -365,7 +376,9 @@ public:
         IPC::RequestParser rp{ctx};
 
         auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
+
+        std::string name(file_buffer.begin(), end);
 
         NGLOG_DEBUG(Service_FS, "called file {}", name);
 

src/core/hle/service/hid/hid.cpp

@@ -94,6 +94,7 @@ private:
                 layout.header.latest_entry = (layout.header.latest_entry + 1) % HID_NUM_ENTRIES;
 
                 ControllerInputEntry& entry = layout.entries[layout.header.latest_entry];
+                entry.connection_state = ConnectionState_Connected | ConnectionState_Wired;
                 entry.timestamp++;
                 // TODO(shinyquagsire23): Is this always identical to timestamp?
                 entry.timestamp_2++;
@@ -102,8 +103,6 @@ private:
                 if (controller != Controller_Handheld)
                     continue;
 
-                entry.connection_state = ConnectionState_Connected | ConnectionState_Wired;
-
                 // TODO(shinyquagsire23): Set up some LUTs for each layout mapping in the future?
                 // For now everything is just the default handheld layout, but split Joy-Con will
                 // rotate the face buttons and directions for certain layouts.

src/core/hle/service/hid/hid.h

@@ -12,7 +12,7 @@ namespace Service::HID {
 // Begin enums and output structs
 
 constexpr u32 HID_NUM_ENTRIES = 17;
-constexpr u32 HID_NUM_LAYOUTS = 7;
+constexpr u32 HID_NUM_LAYOUTS = 2;
 constexpr s32 HID_JOYSTICK_MAX = 0x8000;
 constexpr s32 HID_JOYSTICK_MIN = -0x8000;
 

src/video_core/engines/maxwell_3d.h

@@ -318,7 +318,6 @@ public:
             Equation equation_a;
             Factor factor_source_a;
             Factor factor_dest_a;
-            INSERT_PADDING_WORDS(1);
         };
 
         union {
@@ -433,27 +432,7 @@ public:
                     };
                 } rt_control;
 
-                INSERT_PADDING_WORDS(0x31);
-
-                u32 independent_blend_enable;
-
-                INSERT_PADDING_WORDS(0x15);
-
-                struct {
-                    u32 separate_alpha;
-                    Blend::Equation equation_rgb;
-                    Blend::Factor factor_source_rgb;
-                    Blend::Factor factor_dest_rgb;
-                    Blend::Equation equation_a;
-                    Blend::Factor factor_source_a;
-                    INSERT_PADDING_WORDS(1);
-                    Blend::Factor factor_dest_a;
-
-                    u32 enable_common;
-                    u32 enable[NumRenderTargets];
-                } blend;
-
-                INSERT_PADDING_WORDS(0x77);
+                INSERT_PADDING_WORDS(0xCF);
 
                 struct {
                     u32 tsc_address_high;
@@ -578,7 +557,9 @@ public:
 
                 } vertex_array[NumVertexArrays];
 
-                Blend independent_blend[NumRenderTargets];
+                Blend blend;
+
+                INSERT_PADDING_WORDS(0x39);
 
                 struct {
                     u32 limit_high;
@@ -741,8 +722,6 @@ ASSERT_REG_POSITION(vertex_buffer, 0x35D);
 ASSERT_REG_POSITION(zeta, 0x3F8);
 ASSERT_REG_POSITION(vertex_attrib_format[0], 0x458);
 ASSERT_REG_POSITION(rt_control, 0x487);
-ASSERT_REG_POSITION(independent_blend_enable, 0x4B9);
-ASSERT_REG_POSITION(blend, 0x4CF);
 ASSERT_REG_POSITION(tsc, 0x557);
 ASSERT_REG_POSITION(tic, 0x55D);
 ASSERT_REG_POSITION(code_address, 0x582);
@@ -750,7 +729,7 @@ ASSERT_REG_POSITION(draw, 0x585);
 ASSERT_REG_POSITION(index_array, 0x5F2);
 ASSERT_REG_POSITION(query, 0x6C0);
 ASSERT_REG_POSITION(vertex_array[0], 0x700);
-ASSERT_REG_POSITION(independent_blend, 0x780);
+ASSERT_REG_POSITION(blend, 0x780);
 ASSERT_REG_POSITION(vertex_array_limit[0], 0x7C0);
 ASSERT_REG_POSITION(shader_config[0], 0x800);
 ASSERT_REG_POSITION(const_buffer, 0x8E0);

src/video_core/engines/shader_bytecode.h

@@ -213,7 +213,6 @@ union Instruction {
     BitField<28, 8, Register> gpr28;
     BitField<39, 8, Register> gpr39;
     BitField<48, 16, u64> opcode;
-    BitField<50, 1, u64> saturate_a;
 
     union {
         BitField<20, 19, u64> imm20_19;
@@ -260,15 +259,11 @@ union Instruction {
         }
     } alu;
 
-    union {
-        BitField<48, 1, u64> is_signed;
-    } shift;
-
     union {
         BitField<39, 5, u64> shift_amount;
         BitField<48, 1, u64> negate_b;
         BitField<49, 1, u64> negate_a;
-    } alu_integer;
+    } iscadd;
 
     union {
         BitField<20, 8, u64> shift_position;
@@ -329,15 +324,6 @@ union Instruction {
         BitField<56, 1, u64> neg_imm;
     } fset;
 
-    union {
-        BitField<39, 3, u64> pred39;
-        BitField<42, 1, u64> neg_pred;
-        BitField<44, 1, u64> bf;
-        BitField<45, 2, PredOperation> op;
-        BitField<48, 1, u64> is_signed;
-        BitField<49, 3, PredCondition> cond;
-    } iset;
-
     union {
         BitField<10, 2, Register::Size> size;
         BitField<12, 1, u64> is_output_signed;
@@ -345,6 +331,7 @@ union Instruction {
         BitField<41, 2, u64> selector;
         BitField<45, 1, u64> negate_a;
         BitField<49, 1, u64> abs_a;
+        BitField<50, 1, u64> saturate_a;
 
         union {
             BitField<39, 2, F2iRoundingOp> rounding;
@@ -423,7 +410,6 @@ class OpCode {
 public:
     enum class Id {
         KIL,
-        SSY,
         BFE_C,
         BFE_R,
         BFE_IMM,
@@ -448,9 +434,6 @@ public:
         FMUL_R,
         FMUL_IMM,
         FMUL32_IMM,
-        IADD_C,
-        IADD_R,
-        IADD_IMM,
         ISCADD_C, // Scale and Add
         ISCADD_R,
         ISCADD_IMM,
@@ -496,9 +479,6 @@ public:
         ISETP_C,
         ISETP_IMM,
         ISETP_R,
-        ISET_R,
-        ISET_C,
-        ISET_IMM,
         PSETP,
         XMAD_IMM,
         XMAD_CR,
@@ -509,16 +489,15 @@ public:
     enum class Type {
         Trivial,
         Arithmetic,
-        ArithmeticInteger,
         Bfe,
         Logic,
         Shift,
+        ScaledAdd,
         Ffma,
         Flow,
         Memory,
         FloatSet,
         FloatSetPredicate,
-        IntegerSet,
         IntegerSetPredicate,
         PredicateSetPredicate,
         Conversion,
@@ -617,7 +596,6 @@ private:
         std::vector<Matcher> table = {
 #define INST(bitstring, op, type, name) Detail::GetMatcher(bitstring, op, type, name)
             INST("111000110011----", Id::KIL, Type::Flow, "KIL"),
-            INST("111000101001----", Id::SSY, Type::Flow, "SSY"),
             INST("111000100100----", Id::BRA, Type::Flow, "BRA"),
             INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"),
             INST("1110111110010---", Id::LD_C, Type::Memory, "LD_C"),
@@ -639,12 +617,9 @@ private:
             INST("0101110001101---", Id::FMUL_R, Type::Arithmetic, "FMUL_R"),
             INST("0011100-01101---", Id::FMUL_IMM, Type::Arithmetic, "FMUL_IMM"),
             INST("00011110--------", Id::FMUL32_IMM, Type::Arithmetic, "FMUL32_IMM"),
-            INST("0100110000010---", Id::IADD_C, Type::ArithmeticInteger, "IADD_C"),
-            INST("0101110000010---", Id::IADD_R, Type::ArithmeticInteger, "IADD_R"),
-            INST("0011100-00010---", Id::IADD_IMM, Type::ArithmeticInteger, "IADD_IMM"),
-            INST("0100110000011---", Id::ISCADD_C, Type::ArithmeticInteger, "ISCADD_C"),
-            INST("0101110000011---", Id::ISCADD_R, Type::ArithmeticInteger, "ISCADD_R"),
-            INST("0011100-00011---", Id::ISCADD_IMM, Type::ArithmeticInteger, "ISCADD_IMM"),
+            INST("0100110000011---", Id::ISCADD_C, Type::ScaledAdd, "ISCADD_C"),
+            INST("0101110000011---", Id::ISCADD_R, Type::ScaledAdd, "ISCADD_R"),
+            INST("0011100-00011---", Id::ISCADD_IMM, Type::ScaledAdd, "ISCADD_IMM"),
             INST("0101000010000---", Id::MUFU, Type::Arithmetic, "MUFU"),
             INST("0100110010010---", Id::RRO_C, Type::Arithmetic, "RRO_C"),
             INST("0101110010010---", Id::RRO_R, Type::Arithmetic, "RRO_R"),
@@ -690,9 +665,6 @@ private:
             INST("010010110110----", Id::ISETP_C, Type::IntegerSetPredicate, "ISETP_C"),
             INST("010110110110----", Id::ISETP_R, Type::IntegerSetPredicate, "ISETP_R"),
             INST("0011011-0110----", Id::ISETP_IMM, Type::IntegerSetPredicate, "ISETP_IMM"),
-            INST("010110110101----", Id::ISET_R, Type::IntegerSet, "ISET_R"),
-            INST("010010110101----", Id::ISET_C, Type::IntegerSet, "ISET_C"),
-            INST("0011011-0101----", Id::ISET_IMM, Type::IntegerSet, "ISET_IMM"),
             INST("0101000010010---", Id::PSETP, Type::PredicateSetPredicate, "PSETP"),
             INST("0011011-00------", Id::XMAD_IMM, Type::Arithmetic, "XMAD_IMM"),
             INST("0100111---------", Id::XMAD_CR, Type::Arithmetic, "XMAD_CR"),

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -218,9 +218,6 @@ void RasterizerOpenGL::SetupShaders(u8* buffer_ptr, GLintptr buffer_offset) {
         ubo.SetFromRegs(gpu.state.shader_stages[stage]);
         std::memcpy(buffer_ptr, &ubo, sizeof(ubo));
 
-        // Flush the buffer so that the GPU can see the data we just wrote.
-        glFlushMappedBufferRange(GL_ARRAY_BUFFER, buffer_offset, sizeof(ubo));
-
         // Upload uniform data as one UBO per stage
         const GLintptr ubo_offset = buffer_offset;
         copy_buffer(uniform_buffers[stage].handle, ubo_offset,
@@ -349,9 +346,6 @@ void RasterizerOpenGL::DrawArrays() {
     // Sync the viewport
     SyncViewport(surfaces_rect, res_scale);
 
-    // Sync the blend state registers
-    SyncBlendState();
-
     // TODO(bunnei): Sync framebuffer_scale uniform here
     // TODO(bunnei): Sync scissorbox uniform(s) here
 
@@ -458,7 +452,32 @@ void RasterizerOpenGL::DrawArrays() {
     }
 }
 
-void RasterizerOpenGL::NotifyMaxwellRegisterChanged(u32 method) {}
+void RasterizerOpenGL::NotifyMaxwellRegisterChanged(u32 method) {
+    const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
+    switch (method) {
+    case MAXWELL3D_REG_INDEX(blend.separate_alpha):
+        ASSERT_MSG(false, "unimplemented");
+        break;
+    case MAXWELL3D_REG_INDEX(blend.equation_rgb):
+        state.blend.rgb_equation = MaxwellToGL::BlendEquation(regs.blend.equation_rgb);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.factor_source_rgb):
+        state.blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_rgb);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.factor_dest_rgb):
+        state.blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_rgb);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.equation_a):
+        state.blend.a_equation = MaxwellToGL::BlendEquation(regs.blend.equation_a);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.factor_source_a):
+        state.blend.src_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_source_a);
+        break;
+    case MAXWELL3D_REG_INDEX(blend.factor_dest_a):
+        state.blend.dst_a_func = MaxwellToGL::BlendFunc(regs.blend.factor_dest_a);
+        break;
+    }
+}
 
 void RasterizerOpenGL::FlushAll() {
     MICROPROFILE_SCOPE(OpenGL_CacheManagement);
@@ -738,21 +757,14 @@ void RasterizerOpenGL::SyncDepthOffset() {
     UNREACHABLE();
 }
 
-void RasterizerOpenGL::SyncBlendState() {
-    const auto& regs = Core::System().GetInstance().GPU().Maxwell3D().regs;
-    ASSERT_MSG(regs.independent_blend_enable == 1, "Only independent blending is implemented");
-
-    // TODO(Subv): Support more than just render target 0.
-    state.blend.enabled = regs.blend.enable[0] != 0;
-
-    if (!state.blend.enabled)
-        return;
-
-    ASSERT_MSG(!regs.independent_blend[0].separate_alpha, "Unimplemented");
-    state.blend.rgb_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_rgb);
-    state.blend.src_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_rgb);
-    state.blend.dst_rgb_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_rgb);
-    state.blend.a_equation = MaxwellToGL::BlendEquation(regs.independent_blend[0].equation_a);
-    state.blend.src_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_source_a);
-    state.blend.dst_a_func = MaxwellToGL::BlendFunc(regs.independent_blend[0].factor_dest_a);
+void RasterizerOpenGL::SyncBlendEnabled() {
+    UNREACHABLE();
+}
+
+void RasterizerOpenGL::SyncBlendFuncs() {
+    UNREACHABLE();
+}
+
+void RasterizerOpenGL::SyncBlendColor() {
+    UNREACHABLE();
 }

src/video_core/renderer_opengl/gl_rasterizer.h

@@ -121,8 +121,14 @@ private:
     /// Syncs the depth offset to match the guest state
     void SyncDepthOffset();
 
-    /// Syncs the blend state to match the guest state
-    void SyncBlendState();
+    /// Syncs the blend enabled status to match the guest state
+    void SyncBlendEnabled();
+
+    /// Syncs the blend functions to match the guest state
+    void SyncBlendFuncs();
+
+    /// Syncs the blend color to match the guest state
+    void SyncBlendColor();
 
     bool has_ARB_buffer_storage;
     bool has_ARB_direct_state_access;

src/video_core/renderer_opengl/gl_rasterizer_cache.cpp

@@ -1033,11 +1033,8 @@ Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextu
     params.addr = config.tic.Address();
     params.is_tiled = config.tic.IsTiled();
     params.pixel_format = SurfaceParams::PixelFormatFromTextureFormat(config.tic.format);
-
-    params.width = Common::AlignUp(config.tic.Width(), params.GetCompresssionFactor()) /
-                   params.GetCompresssionFactor();
-    params.height = Common::AlignUp(config.tic.Height(), params.GetCompresssionFactor()) /
-                    params.GetCompresssionFactor();
+    params.width = config.tic.Width() / params.GetCompresssionFactor();
+    params.height = config.tic.Height() / params.GetCompresssionFactor();
 
     // TODO(Subv): Different types per component are not supported.
     ASSERT(config.tic.r_type.Value() == config.tic.g_type.Value() &&
@@ -1048,8 +1045,6 @@ Surface RasterizerCacheOpenGL::GetTextureSurface(const Tegra::Texture::FullTextu
 
     if (config.tic.IsTiled()) {
         params.block_height = config.tic.BlockHeight();
-        params.width = Common::AlignUp(params.width, params.block_height);
-        params.height = Common::AlignUp(params.height, params.block_height);
     } else {
         // Use the texture-provided stride value if the texture isn't tiled.
         params.stride = static_cast<u32>(params.PixelsInBytes(config.tic.Pitch()));

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -538,7 +538,7 @@ private:
             // vertex shader, and what's the value of the fourth element when inside a Tess Eval
             // shader.
             ASSERT(stage == Maxwell3D::Regs::ShaderStage::Vertex);
-            return "vec4(0, 0, uintBitsToFloat(gl_InstanceID), uintBitsToFloat(gl_VertexID))";
+            return "vec4(0, 0, gl_InstanceID, gl_VertexID)";
         default:
             const u32 index{static_cast<u32>(attribute) -
                             static_cast<u32>(Attribute::Index::Attribute_0)};
@@ -808,8 +808,6 @@ private:
             case OpCode::Id::FMUL_C:
             case OpCode::Id::FMUL_R:
             case OpCode::Id::FMUL_IMM: {
-                ASSERT_MSG(!instr.saturate_a, "Unimplemented");
-
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " * " + op_b, 1, 1, instr.alu.abs_d);
                 break;
             }
@@ -823,14 +821,10 @@ private:
             case OpCode::Id::FADD_C:
             case OpCode::Id::FADD_R:
             case OpCode::Id::FADD_IMM: {
-                ASSERT_MSG(!instr.saturate_a, "Unimplemented");
-
                 regs.SetRegisterToFloat(instr.gpr0, 0, op_a + " + " + op_b, 1, 1, instr.alu.abs_d);
                 break;
             }
             case OpCode::Id::MUFU: {
-                ASSERT_MSG(!instr.saturate_a, "Unimplemented");
-
                 switch (instr.sub_op) {
                 case SubOp::Cos:
                     regs.SetRegisterToFloat(instr.gpr0, 0, "cos(" + op_a + ')', 1, 1,
@@ -979,19 +973,6 @@ private:
             }
 
             switch (opcode->GetId()) {
-            case OpCode::Id::SHR_C:
-            case OpCode::Id::SHR_R:
-            case OpCode::Id::SHR_IMM: {
-                if (!instr.shift.is_signed) {
-                    // Logical shift right
-                    op_a = "uint(" + op_a + ')';
-                }
-
-                // Cast to int is superfluous for arithmetic shift, it's only for a logical shift
-                regs.SetRegisterToInteger(instr.gpr0, true, 0, "int(" + op_a + " >> " + op_b + ')',
-                                          1, 1);
-                break;
-            }
             case OpCode::Id::SHL_C:
             case OpCode::Id::SHL_R:
             case OpCode::Id::SHL_IMM:
@@ -1005,13 +986,13 @@ private:
             break;
         }
 
-        case OpCode::Type::ArithmeticInteger: {
+        case OpCode::Type::ScaledAdd: {
             std::string op_a = regs.GetRegisterAsInteger(instr.gpr8);
 
-            if (instr.alu_integer.negate_a)
+            if (instr.iscadd.negate_a)
                 op_a = '-' + op_a;
 
-            std::string op_b = instr.alu_integer.negate_b ? "-" : "";
+            std::string op_b = instr.iscadd.negate_b ? "-" : "";
 
             if (instr.is_b_imm) {
                 op_b += '(' + std::to_string(instr.alu.GetSignedImm20_20()) + ')';
@@ -1024,35 +1005,13 @@ private:
                 }
             }
 
-            switch (opcode->GetId()) {
-            case OpCode::Id::IADD_C:
-            case OpCode::Id::IADD_R:
-            case OpCode::Id::IADD_IMM: {
-                ASSERT_MSG(!instr.saturate_a, "Unimplemented");
-                regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " + " + op_b, 1, 1);
-                break;
-            }
-            case OpCode::Id::ISCADD_C:
-            case OpCode::Id::ISCADD_R:
-            case OpCode::Id::ISCADD_IMM: {
-                std::string shift = std::to_string(instr.alu_integer.shift_amount.Value());
-
-                regs.SetRegisterToInteger(instr.gpr0, true, 0,
-                                          "((" + op_a + " << " + shift + ") + " + op_b + ')', 1, 1);
-                break;
-            }
-            default: {
-                NGLOG_CRITICAL(HW_GPU, "Unhandled ArithmeticInteger instruction: {}",
-                               opcode->GetName());
-                UNREACHABLE();
-            }
-            }
+            std::string shift = std::to_string(instr.iscadd.shift_amount.Value());
 
+            regs.SetRegisterToInteger(instr.gpr0, true, 0,
+                                      "((" + op_a + " << " + shift + ") + " + op_b + ')', 1, 1);
             break;
         }
         case OpCode::Type::Ffma: {
-            ASSERT_MSG(!instr.saturate_a, "Unimplemented");
-
             std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
             std::string op_b = instr.ffma.negate_b ? "-" : "";
             std::string op_c = instr.ffma.negate_c ? "-" : "";
@@ -1092,7 +1051,7 @@ private:
         case OpCode::Type::Conversion: {
             ASSERT_MSG(instr.conversion.size == Register::Size::Word, "Unimplemented");
             ASSERT_MSG(!instr.conversion.negate_a, "Unimplemented");
-            ASSERT_MSG(!instr.saturate_a, "Unimplemented");
+            ASSERT_MSG(!instr.conversion.saturate_a, "Unimplemented");
 
             switch (opcode->GetId()) {
             case OpCode::Id::I2I_R: {
@@ -1122,8 +1081,6 @@ private:
                 break;
             }
             case OpCode::Id::F2F_R: {
-                ASSERT_MSG(!instr.saturate_a, "Unimplemented");
-
                 std::string op_a = regs.GetRegisterAsFloat(instr.gpr20);
 
                 switch (instr.conversion.f2f.rounding) {
@@ -1241,8 +1198,8 @@ private:
                 const std::string op_b = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
                 const std::string sampler = GetSampler(instr.sampler);
                 const std::string coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
-                // Add an extra scope and declare the texture coords inside to prevent
-                // overwriting them in case they are used as outputs of the texs instruction.
+                // Add an extra scope and declare the texture coords inside to prevent overwriting
+                // them in case they are used as outputs of the texs instruction.
                 shader.AddLine("{");
                 ++shader.scope;
                 shader.AddLine(coord);
@@ -1273,8 +1230,8 @@ private:
                 shader.AddLine(coord);
                 const std::string texture = "texture(" + sampler + ", coords)";
 
-                // TEXS has two destination registers. RG goes into gpr0+0 and gpr0+1, and BA
-                // goes into gpr28+0 and gpr28+1
+                // TEXS has two destination registers. RG goes into gpr0+0 and gpr0+1, and BA goes
+                // into gpr28+0 and gpr28+1
                 size_t offset{};
 
                 for (const auto& dest : {instr.gpr0.Value(), instr.gpr28.Value()}) {
@@ -1423,8 +1380,8 @@ private:
                 op_b = "abs(" + op_b + ')';
             }
 
-            // The fset instruction sets a register to 1.0 or -1 (depending on the bf bit) if the
-            // condition is true, and to 0 otherwise.
+            // The fset instruction sets a register to 1.0 if the condition is true, and to 0
+            // otherwise.
             std::string second_pred =
                 GetPredicateCondition(instr.fset.pred39, instr.fset.neg_pred != 0);
 
@@ -1442,41 +1399,6 @@ private:
             }
             break;
         }
-        case OpCode::Type::IntegerSet: {
-            std::string op_a = regs.GetRegisterAsInteger(instr.gpr8, 0, instr.iset.is_signed);
-
-            std::string op_b;
-
-            if (instr.is_b_imm) {
-                op_b = std::to_string(instr.alu.GetSignedImm20_20());
-            } else {
-                if (instr.is_b_gpr) {
-                    op_b = regs.GetRegisterAsInteger(instr.gpr20, 0, instr.iset.is_signed);
-                } else {
-                    op_b = regs.GetUniform(instr.cbuf34.index, instr.cbuf34.offset,
-                                           GLSLRegister::Type::Integer);
-                }
-            }
-
-            // The iset instruction sets a register to 1.0 or -1 (depending on the bf bit) if the
-            // condition is true, and to 0 otherwise.
-            std::string second_pred =
-                GetPredicateCondition(instr.iset.pred39, instr.iset.neg_pred != 0);
-
-            std::string comparator = GetPredicateComparison(instr.iset.cond);
-            std::string combiner = GetPredicateCombiner(instr.iset.op);
-
-            std::string predicate = "(((" + op_a + ") " + comparator + " (" + op_b + ")) " +
-                                    combiner + " (" + second_pred + "))";
-
-            if (instr.iset.bf) {
-                regs.SetRegisterToFloat(instr.gpr0, 0, predicate + " ? 1.0 : 0.0", 1, 1);
-            } else {
-                regs.SetRegisterToInteger(instr.gpr0, false, 0, predicate + " ? 0xFFFFFFFF : 0", 1,
-                                          1);
-            }
-            break;
-        }
         default: {
             switch (opcode->GetId()) {
             case OpCode::Id::EXIT: {
@@ -1490,8 +1412,8 @@ private:
 
                 shader.AddLine("return true;");
                 if (instr.pred.pred_index == static_cast<u64>(Pred::UnusedIndex)) {
-                    // If this is an unconditional exit then just end processing here, otherwise
-                    // we have to account for the possibility of the condition not being met, so
+                    // If this is an unconditional exit then just end processing here, otherwise we
+                    // have to account for the possibility of the condition not being met, so
                     // continue processing the next instruction.
                     offset = PROGRAM_END - 1;
                 }
@@ -1513,11 +1435,6 @@ private:
                 regs.SetRegisterToInputAttibute(instr.gpr0, attribute.element, attribute.index);
                 break;
             }
-            case OpCode::Id::SSY: {
-                // The SSY opcode tells the GPU where to re-converge divergent execution paths, we
-                // can ignore this when generating GLSL code.
-                break;
-            }
             default: {
                 NGLOG_CRITICAL(HW_GPU, "Unhandled instruction: {}", opcode->GetName());
                 UNREACHABLE();