kernel/svc: Deglobalize the supervisor call handlers

Adjusts the interface of the wrappers to take a system reference, which
allows accessing a system instance without using the global accessors.

This also allows getting rid of all global accessors within the
supervisor call handling code. While this does make the wrappers
themselves slightly more noisy, this will be further cleaned up in a
follow-up. This eliminates the global system accessors in the current
code while preserving the existing interface.

.gitmodules

@@ -40,3 +40,6 @@
 [submodule "Vulkan-Headers"]
     path = externals/Vulkan-Headers
     url = https://github.com/KhronosGroup/Vulkan-Headers.git
+[submodule "externals/zstd"]
+    path = externals/zstd
+    url = https://github.com/facebook/zstd

CMakeLists.txt

@@ -309,7 +309,7 @@ if (CLANG_FORMAT)
     set(CCOMMENT "Running clang format against all the .h and .cpp files in src/")
     if (WIN32)
         add_custom_target(clang-format
-            COMMAND powershell.exe -Command "Get-ChildItem ${SRCS}/* -Include *.cpp,*.h -Recurse | Foreach {${CLANG_FORMAT} -i $_.fullname}"
+            COMMAND powershell.exe -Command "Get-ChildItem '${SRCS}/*' -Include *.cpp,*.h -Recurse | Foreach {&'${CLANG_FORMAT}' -i $_.fullname}"
             COMMENT ${CCOMMENT})
     elseif(MINGW)
         add_custom_target(clang-format

externals/CMakeLists.txt

@@ -49,6 +49,10 @@ add_subdirectory(open_source_archives EXCLUDE_FROM_ALL)
 add_library(unicorn-headers INTERFACE)
 target_include_directories(unicorn-headers INTERFACE ./unicorn/include)
 
+# Zstandard
+add_subdirectory(zstd/build/cmake EXCLUDE_FROM_ALL)
+target_include_directories(libzstd_static INTERFACE ./zstd/lib)
+
 # SoundTouch
 add_subdirectory(soundtouch)
 

src/common/CMakeLists.txt

@@ -91,6 +91,8 @@ add_library(common STATIC
     logging/log.h
     logging/text_formatter.cpp
     logging/text_formatter.h
+    lz4_compression.cpp
+    lz4_compression.h
     math_util.h
     memory_hook.cpp
     memory_hook.h
@@ -123,6 +125,8 @@ add_library(common STATIC
     uint128.h
     vector_math.h
     web_result.h
+    zstd_compression.cpp
+    zstd_compression.h
 )
 
 if(ARCHITECTURE_x86_64)
@@ -136,3 +140,4 @@ endif()
 create_target_directory_groups(common)
 
 target_link_libraries(common PUBLIC Boost::boost fmt microprofile)
+target_link_libraries(common PRIVATE lz4_static libzstd_static)

src/common/assert.h

@@ -57,3 +57,21 @@ __declspec(noinline, noreturn)
 
 #define UNIMPLEMENTED_IF(cond) ASSERT_MSG(!(cond), "Unimplemented code!")
 #define UNIMPLEMENTED_IF_MSG(cond, ...) ASSERT_MSG(!(cond), __VA_ARGS__)
+
+// If the assert is ignored, execute _b_
+#define ASSERT_OR_EXECUTE(_a_, _b_)                                                                \
+    do {                                                                                           \
+        ASSERT(_a_);                                                                               \
+        if (!(_a_)) {                                                                              \
+            _b_                                                                                    \
+        }                                                                                          \
+    } while (0)
+
+// If the assert is ignored, execute _b_
+#define ASSERT_OR_EXECUTE_MSG(_a_, _b_, ...)                                                       \
+    do {                                                                                           \
+        ASSERT_MSG(_a_, __VA_ARGS__);                                                              \
+        if (!(_a_)) {                                                                              \
+            _b_                                                                                    \
+        }                                                                                          \
+    } while (0)

src/common/bit_util.h

@@ -32,7 +32,7 @@ inline u32 CountLeadingZeroes32(u32 value) {
     return 32;
 }
 
-inline u64 CountLeadingZeroes64(u64 value) {
+inline u32 CountLeadingZeroes64(u64 value) {
     unsigned long leading_zero = 0;
 
     if (_BitScanReverse64(&leading_zero, value) != 0) {
@@ -47,15 +47,15 @@ inline u32 CountLeadingZeroes32(u32 value) {
         return 32;
     }
 
-    return __builtin_clz(value);
+    return static_cast<u32>(__builtin_clz(value));
 }
 
-inline u64 CountLeadingZeroes64(u64 value) {
+inline u32 CountLeadingZeroes64(u64 value) {
     if (value == 0) {
         return 64;
     }
 
-    return __builtin_clzll(value);
+    return static_cast<u32>(__builtin_clzll(value));
 }
 #endif
 
@@ -70,7 +70,7 @@ inline u32 CountTrailingZeroes32(u32 value) {
     return 32;
 }
 
-inline u64 CountTrailingZeroes64(u64 value) {
+inline u32 CountTrailingZeroes64(u64 value) {
     unsigned long trailing_zero = 0;
 
     if (_BitScanForward64(&trailing_zero, value) != 0) {
@@ -85,15 +85,15 @@ inline u32 CountTrailingZeroes32(u32 value) {
         return 32;
     }
 
-    return __builtin_ctz(value);
+    return static_cast<u32>(__builtin_ctz(value));
 }
 
-inline u64 CountTrailingZeroes64(u64 value) {
+inline u32 CountTrailingZeroes64(u64 value) {
     if (value == 0) {
         return 64;
     }
 
-    return __builtin_ctzll(value);
+    return static_cast<u32>(__builtin_ctzll(value));
 }
 #endif
 

src/common/lz4_compression.cpp

@@ -0,0 +1,76 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <algorithm>
+#include <lz4hc.h>
+
+#include "common/assert.h"
+#include "common/lz4_compression.h"
+
+namespace Common::Compression {
+
+std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size) {
+    ASSERT_MSG(source_size <= LZ4_MAX_INPUT_SIZE, "Source size exceeds LZ4 maximum input size");
+
+    const auto source_size_int = static_cast<int>(source_size);
+    const int max_compressed_size = LZ4_compressBound(source_size_int);
+    std::vector<u8> compressed(max_compressed_size);
+
+    const int compressed_size = LZ4_compress_default(reinterpret_cast<const char*>(source),
+                                                     reinterpret_cast<char*>(compressed.data()),
+                                                     source_size_int, max_compressed_size);
+
+    if (compressed_size <= 0) {
+        // Compression failed
+        return {};
+    }
+
+    compressed.resize(compressed_size);
+
+    return compressed;
+}
+
+std::vector<u8> CompressDataLZ4HC(const u8* source, std::size_t source_size,
+                                  s32 compression_level) {
+    ASSERT_MSG(source_size <= LZ4_MAX_INPUT_SIZE, "Source size exceeds LZ4 maximum input size");
+
+    compression_level = std::clamp(compression_level, LZ4HC_CLEVEL_MIN, LZ4HC_CLEVEL_MAX);
+
+    const auto source_size_int = static_cast<int>(source_size);
+    const int max_compressed_size = LZ4_compressBound(source_size_int);
+    std::vector<u8> compressed(max_compressed_size);
+
+    const int compressed_size = LZ4_compress_HC(
+        reinterpret_cast<const char*>(source), reinterpret_cast<char*>(compressed.data()),
+        source_size_int, max_compressed_size, compression_level);
+
+    if (compressed_size <= 0) {
+        // Compression failed
+        return {};
+    }
+
+    compressed.resize(compressed_size);
+
+    return compressed;
+}
+
+std::vector<u8> CompressDataLZ4HCMax(const u8* source, std::size_t source_size) {
+    return CompressDataLZ4HC(source, source_size, LZ4HC_CLEVEL_MAX);
+}
+
+std::vector<u8> DecompressDataLZ4(const std::vector<u8>& compressed,
+                                  std::size_t uncompressed_size) {
+    std::vector<u8> uncompressed(uncompressed_size);
+    const int size_check = LZ4_decompress_safe(reinterpret_cast<const char*>(compressed.data()),
+                                               reinterpret_cast<char*>(uncompressed.data()),
+                                               static_cast<int>(compressed.size()),
+                                               static_cast<int>(uncompressed.size()));
+    if (static_cast<int>(uncompressed_size) != size_check) {
+        // Decompression failed
+        return {};
+    }
+    return uncompressed;
+}
+
+} // namespace Common::Compression

src/common/lz4_compression.h

@@ -0,0 +1,55 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <vector>
+
+#include "common/common_types.h"
+
+namespace Common::Compression {
+
+/**
+ * Compresses a source memory region with LZ4 and returns the compressed data in a vector.
+ *
+ * @param source the uncompressed source memory region.
+ * @param source_size the size in bytes of the uncompressed source memory region.
+ *
+ * @return the compressed data.
+ */
+std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size);
+
+/**
+ * Utilizes the LZ4 subalgorithm LZ4HC with the specified compression level. Higher compression
+ * levels result in a smaller compressed size, but require more CPU time for compression. The
+ * compression level has almost no impact on decompression speed. Data compressed with LZ4HC can
+ * also be decompressed with the default LZ4 decompression.
+ *
+ * @param source the uncompressed source memory region.
+ * @param source_size the size in bytes of the uncompressed source memory region.
+ * @param compression_level the used compression level. Should be between 3 and 12.
+ *
+ * @return the compressed data.
+ */
+std::vector<u8> CompressDataLZ4HC(const u8* source, std::size_t source_size, s32 compression_level);
+
+/**
+ * Utilizes the LZ4 subalgorithm LZ4HC with the highest possible compression level.
+ *
+ * @param source the uncompressed source memory region.
+ * @param source_size the size in bytes of the uncompressed source memory region.
+ *
+ * @return the compressed data.
+ */
+std::vector<u8> CompressDataLZ4HCMax(const u8* source, std::size_t source_size);
+
+/**
+ * Decompresses a source memory region with LZ4 and returns the uncompressed data in a vector.
+ *
+ * @param compressed the compressed source memory region.
+ * @param uncompressed_size the size in bytes of the uncompressed data.
+ *
+ * @return the decompressed data.
+ */
+std::vector<u8> DecompressDataLZ4(const std::vector<u8>& compressed, std::size_t uncompressed_size);
+
+} // namespace Common::Compression

src/common/multi_level_queue.h

@@ -72,7 +72,7 @@ public:
                 u64 prios = mlq.used_priorities;
                 prios &= ~((1ULL << (current_priority + 1)) - 1);
                 if (prios == 0) {
-                    current_priority = mlq.depth();
+                    current_priority = static_cast<u32>(mlq.depth());
                 } else {
                     current_priority = CountTrailingZeroes64(prios);
                     it = GetBeginItForPrio();

src/common/zstd_compression.cpp

@@ -0,0 +1,53 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <algorithm>
+#include <zstd.h>
+
+#include "common/assert.h"
+#include "common/zstd_compression.h"
+
+namespace Common::Compression {
+
+std::vector<u8> CompressDataZSTD(const u8* source, std::size_t source_size, s32 compression_level) {
+    compression_level = std::clamp(compression_level, 1, ZSTD_maxCLevel());
+
+    const std::size_t max_compressed_size = ZSTD_compressBound(source_size);
+    std::vector<u8> compressed(max_compressed_size);
+
+    const std::size_t compressed_size =
+        ZSTD_compress(compressed.data(), compressed.size(), source, source_size, compression_level);
+
+    if (ZSTD_isError(compressed_size)) {
+        // Compression failed
+        return {};
+    }
+
+    compressed.resize(compressed_size);
+
+    return compressed;
+}
+
+std::vector<u8> CompressDataZSTDDefault(const u8* source, std::size_t source_size) {
+    return CompressDataZSTD(source, source_size, ZSTD_CLEVEL_DEFAULT);
+}
+
+std::vector<u8> DecompressDataZSTD(const std::vector<u8>& compressed) {
+    const std::size_t decompressed_size =
+        ZSTD_getDecompressedSize(compressed.data(), compressed.size());
+    std::vector<u8> decompressed(decompressed_size);
+
+    const std::size_t uncompressed_result_size = ZSTD_decompress(
+        decompressed.data(), decompressed.size(), compressed.data(), compressed.size());
+
+    if (decompressed_size != uncompressed_result_size || ZSTD_isError(uncompressed_result_size)) {
+        // Decompression failed
+        return {};
+    }
+    return decompressed;
+}
+
+} // namespace Common::Compression

src/common/zstd_compression.h

@@ -0,0 +1,42 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <vector>
+
+#include "common/common_types.h"
+
+namespace Common::Compression {
+
+/**
+ * Compresses a source memory region with Zstandard and returns the compressed data in a vector.
+ *
+ * @param source the uncompressed source memory region.
+ * @param source_size the size in bytes of the uncompressed source memory region.
+ * @param compression_level the used compression level. Should be between 1 and 22.
+ *
+ * @return the compressed data.
+ */
+std::vector<u8> CompressDataZSTD(const u8* source, std::size_t source_size, s32 compression_level);
+
+/**
+ * Compresses a source memory region with Zstandard with the default compression level and returns
+ * the compressed data in a vector.
+ *
+ * @param source the uncompressed source memory region.
+ * @param source_size the size in bytes of the uncompressed source memory region.
+ *
+ * @return the compressed data.
+ */
+std::vector<u8> CompressDataZSTDDefault(const u8* source, std::size_t source_size);
+
+/**
+ * Decompresses a source memory region with Zstandard and returns the uncompressed data in a vector.
+ *
+ * @param compressed the compressed source memory region.
+ *
+ * @return the decompressed data.
+ */
+std::vector<u8> DecompressDataZSTD(const std::vector<u8>& compressed);
+
+} // namespace Common::Compression

src/core/CMakeLists.txt

@@ -458,7 +458,7 @@ add_library(core STATIC
 create_target_directory_groups(core)
 
 target_link_libraries(core PUBLIC common PRIVATE audio_core video_core)
-target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt lz4_static mbedtls opus unicorn open_source_archives)
+target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt mbedtls opus unicorn open_source_archives)
 if (ENABLE_WEB_SERVICE)
     target_compile_definitions(core PRIVATE -DENABLE_WEB_SERVICE)
     target_link_libraries(core PRIVATE web_service)

src/core/arm/dynarmic/arm_dynarmic.cpp

@@ -14,6 +14,7 @@
 #include "core/core_timing.h"
 #include "core/core_timing_util.h"
 #include "core/gdbstub/gdbstub.h"
+#include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/svc.h"
 #include "core/hle/kernel/vm_manager.h"
@@ -26,7 +27,6 @@ using Vector = Dynarmic::A64::Vector;
 class ARM_Dynarmic_Callbacks : public Dynarmic::A64::UserCallbacks {
 public:
     explicit ARM_Dynarmic_Callbacks(ARM_Dynarmic& parent) : parent(parent) {}
-    ~ARM_Dynarmic_Callbacks() = default;
 
     u8 MemoryRead8(u64 vaddr) override {
         return Memory::Read8(vaddr);
@@ -100,7 +100,7 @@ public:
     }
 
     void CallSVC(u32 swi) override {
-        Kernel::CallSVC(swi);
+        Kernel::CallSVC(parent.system, swi);
     }
 
     void AddTicks(u64 ticks) override {
@@ -113,14 +113,14 @@ public:
         // Always execute at least one tick.
         amortized_ticks = std::max<u64>(amortized_ticks, 1);
 
-        parent.core_timing.AddTicks(amortized_ticks);
+        parent.system.CoreTiming().AddTicks(amortized_ticks);
         num_interpreted_instructions = 0;
     }
     u64 GetTicksRemaining() override {
-        return std::max(parent.core_timing.GetDowncount(), 0);
+        return std::max(parent.system.CoreTiming().GetDowncount(), 0);
     }
     u64 GetCNTPCT() override {
-        return Timing::CpuCyclesToClockCycles(parent.core_timing.GetTicks());
+        return Timing::CpuCyclesToClockCycles(parent.system.CoreTiming().GetTicks());
     }
 
     ARM_Dynarmic& parent;
@@ -130,7 +130,7 @@ public:
 };
 
 std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const {
-    auto* current_process = Core::CurrentProcess();
+    auto* current_process = system.Kernel().CurrentProcess();
     auto** const page_table = current_process->VMManager().page_table.pointers.data();
 
     Dynarmic::A64::UserConfig config;
@@ -164,7 +164,6 @@ MICROPROFILE_DEFINE(ARM_Jit_Dynarmic, "ARM JIT", "Dynarmic", MP_RGB(255, 64, 64)
 
 void ARM_Dynarmic::Run() {
     MICROPROFILE_SCOPE(ARM_Jit_Dynarmic);
-    ASSERT(Memory::GetCurrentPageTable() == current_page_table);
 
     jit->Run();
 }
@@ -173,10 +172,10 @@ void ARM_Dynarmic::Step() {
     cb->InterpreterFallback(jit->GetPC(), 1);
 }
 
-ARM_Dynarmic::ARM_Dynarmic(Timing::CoreTiming& core_timing, ExclusiveMonitor& exclusive_monitor,
+ARM_Dynarmic::ARM_Dynarmic(System& system, ExclusiveMonitor& exclusive_monitor,
                            std::size_t core_index)
-    : cb(std::make_unique<ARM_Dynarmic_Callbacks>(*this)), inner_unicorn{core_timing},
-      core_index{core_index}, core_timing{core_timing},
+    : cb(std::make_unique<ARM_Dynarmic_Callbacks>(*this)), inner_unicorn{system},
+      core_index{core_index}, system{system},
       exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {
     ThreadContext ctx{};
     inner_unicorn.SaveContext(ctx);
@@ -279,7 +278,6 @@ void ARM_Dynarmic::ClearExclusiveState() {
 
 void ARM_Dynarmic::PageTableChanged() {
     jit = MakeJit();
-    current_page_table = Memory::GetCurrentPageTable();
 }
 
 DynarmicExclusiveMonitor::DynarmicExclusiveMonitor(std::size_t core_count) : monitor(core_count) {}

src/core/arm/dynarmic/arm_dynarmic.h

@@ -12,24 +12,16 @@
 #include "core/arm/exclusive_monitor.h"
 #include "core/arm/unicorn/arm_unicorn.h"
 
-namespace Common {
-struct PageTable;
-}
-
-namespace Core::Timing {
-class CoreTiming;
-}
-
 namespace Core {
 
 class ARM_Dynarmic_Callbacks;
 class DynarmicExclusiveMonitor;
+class System;
 
 class ARM_Dynarmic final : public ARM_Interface {
 public:
-    ARM_Dynarmic(Timing::CoreTiming& core_timing, ExclusiveMonitor& exclusive_monitor,
-                 std::size_t core_index);
-    ~ARM_Dynarmic();
+    ARM_Dynarmic(System& system, ExclusiveMonitor& exclusive_monitor, std::size_t core_index);
+    ~ARM_Dynarmic() override;
 
     void MapBackingMemory(VAddr address, std::size_t size, u8* memory,
                           Kernel::VMAPermission perms) override;
@@ -67,16 +59,14 @@ private:
     ARM_Unicorn inner_unicorn;
 
     std::size_t core_index;
-    Timing::CoreTiming& core_timing;
+    System& system;
     DynarmicExclusiveMonitor& exclusive_monitor;
-
-    Common::PageTable* current_page_table = nullptr;
 };
 
 class DynarmicExclusiveMonitor final : public ExclusiveMonitor {
 public:
     explicit DynarmicExclusiveMonitor(std::size_t core_count);
-    ~DynarmicExclusiveMonitor();
+    ~DynarmicExclusiveMonitor() override;
 
     void SetExclusive(std::size_t core_index, VAddr addr) override;
     void ClearExclusive() override;

src/core/arm/unicorn/arm_unicorn.cpp

@@ -10,7 +10,6 @@
 #include "core/core.h"
 #include "core/core_timing.h"
 #include "core/hle/kernel/svc.h"
-#include "core/memory.h"
 
 namespace Core {
 
@@ -49,20 +48,6 @@ static void CodeHook(uc_engine* uc, uint64_t address, uint32_t size, void* user_
     }
 }
 
-static void InterruptHook(uc_engine* uc, u32 intNo, void* user_data) {
-    u32 esr{};
-    CHECKED(uc_reg_read(uc, UC_ARM64_REG_ESR, &esr));
-
-    auto ec = esr >> 26;
-    auto iss = esr & 0xFFFFFF;
-
-    switch (ec) {
-    case 0x15: // SVC
-        Kernel::CallSVC(iss);
-        break;
-    }
-}
-
 static bool UnmappedMemoryHook(uc_engine* uc, uc_mem_type type, u64 addr, int size, u64 value,
                                void* user_data) {
     ARM_Interface::ThreadContext ctx{};
@@ -72,7 +57,7 @@ static bool UnmappedMemoryHook(uc_engine* uc, uc_mem_type type, u64 addr, int si
     return {};
 }
 
-ARM_Unicorn::ARM_Unicorn(Timing::CoreTiming& core_timing) : core_timing{core_timing} {
+ARM_Unicorn::ARM_Unicorn(System& system) : system{system} {
     CHECKED(uc_open(UC_ARCH_ARM64, UC_MODE_ARM, &uc));
 
     auto fpv = 3 << 20;
@@ -177,7 +162,7 @@ void ARM_Unicorn::Run() {
     if (GDBStub::IsServerEnabled()) {
         ExecuteInstructions(std::max(4000000, 0));
     } else {
-        ExecuteInstructions(std::max(core_timing.GetDowncount(), 0));
+        ExecuteInstructions(std::max(system.CoreTiming().GetDowncount(), 0));
     }
 }
 
@@ -190,14 +175,15 @@ MICROPROFILE_DEFINE(ARM_Jit_Unicorn, "ARM JIT", "Unicorn", MP_RGB(255, 64, 64));
 void ARM_Unicorn::ExecuteInstructions(int num_instructions) {
     MICROPROFILE_SCOPE(ARM_Jit_Unicorn);
     CHECKED(uc_emu_start(uc, GetPC(), 1ULL << 63, 0, num_instructions));
-    core_timing.AddTicks(num_instructions);
+    system.CoreTiming().AddTicks(num_instructions);
     if (GDBStub::IsServerEnabled()) {
-        if (last_bkpt_hit) {
+        if (last_bkpt_hit && last_bkpt.type == GDBStub::BreakpointType::Execute) {
             uc_reg_write(uc, UC_ARM64_REG_PC, &last_bkpt.address);
         }
+
         Kernel::Thread* thread = Kernel::GetCurrentThread();
         SaveContext(thread->GetContext());
-        if (last_bkpt_hit || GDBStub::GetCpuStepFlag()) {
+        if (last_bkpt_hit || GDBStub::IsMemoryBreak() || GDBStub::GetCpuStepFlag()) {
             last_bkpt_hit = false;
             GDBStub::Break();
             GDBStub::SendTrap(thread, 5);
@@ -272,4 +258,20 @@ void ARM_Unicorn::RecordBreak(GDBStub::BreakpointAddress bkpt) {
     last_bkpt_hit = true;
 }
 
+void ARM_Unicorn::InterruptHook(uc_engine* uc, u32 int_no, void* user_data) {
+    u32 esr{};
+    CHECKED(uc_reg_read(uc, UC_ARM64_REG_ESR, &esr));
+
+    const auto ec = esr >> 26;
+    const auto iss = esr & 0xFFFFFF;
+
+    auto* const arm_instance = static_cast<ARM_Unicorn*>(user_data);
+
+    switch (ec) {
+    case 0x15: // SVC
+        Kernel::CallSVC(arm_instance->system, iss);
+        break;
+    }
+}
+
 } // namespace Core

src/core/arm/unicorn/arm_unicorn.h

@@ -9,16 +9,14 @@
 #include "core/arm/arm_interface.h"
 #include "core/gdbstub/gdbstub.h"
 
-namespace Core::Timing {
-class CoreTiming;
-}
-
 namespace Core {
 
+class System;
+
 class ARM_Unicorn final : public ARM_Interface {
 public:
-    explicit ARM_Unicorn(Timing::CoreTiming& core_timing);
-    ~ARM_Unicorn();
+    explicit ARM_Unicorn(System& system);
+    ~ARM_Unicorn() override;
 
     void MapBackingMemory(VAddr address, std::size_t size, u8* memory,
                           Kernel::VMAPermission perms) override;
@@ -47,10 +45,12 @@ public:
     void RecordBreak(GDBStub::BreakpointAddress bkpt);
 
 private:
+    static void InterruptHook(uc_engine* uc, u32 int_no, void* user_data);
+
     uc_engine* uc{};
-    Timing::CoreTiming& core_timing;
+    System& system;
     GDBStub::BreakpointAddress last_bkpt{};
-    bool last_bkpt_hit;
+    bool last_bkpt_hit = false;
 };
 
 } // namespace Core

src/core/core_cpu.cpp

@@ -55,13 +55,13 @@ Cpu::Cpu(System& system, ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_ba
     : cpu_barrier{cpu_barrier}, core_timing{system.CoreTiming()}, core_index{core_index} {
     if (Settings::values.use_cpu_jit) {
 #ifdef ARCHITECTURE_x86_64
-        arm_interface = std::make_unique<ARM_Dynarmic>(core_timing, exclusive_monitor, core_index);
+        arm_interface = std::make_unique<ARM_Dynarmic>(system, exclusive_monitor, core_index);
 #else
-        arm_interface = std::make_unique<ARM_Unicorn>();
+        arm_interface = std::make_unique<ARM_Unicorn>(system);
         LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
 #endif
     } else {
-        arm_interface = std::make_unique<ARM_Unicorn>(core_timing);
+        arm_interface = std::make_unique<ARM_Unicorn>(system);
     }
 
     scheduler = std::make_unique<Kernel::Scheduler>(system, *arm_interface);

src/core/file_sys/control_metadata.cpp

@@ -67,7 +67,7 @@ std::string NACP::GetDeveloperName(Language language) const {
 }
 
 u64 NACP::GetTitleId() const {
-    return raw.title_id;
+    return raw.save_data_owner_id;
 }
 
 u64 NACP::GetDLCBaseTitleId() const {
@@ -80,11 +80,11 @@ std::string NACP::GetVersionString() const {
 }
 
 u64 NACP::GetDefaultNormalSaveSize() const {
-    return raw.normal_save_data_size;
+    return raw.user_account_save_data_size;
 }
 
 u64 NACP::GetDefaultJournalSaveSize() const {
-    return raw.journal_sava_data_size;
+    return raw.user_account_save_data_journal_size;
 }
 
 std::vector<u8> NACP::GetRawBytes() const {

src/core/file_sys/control_metadata.h

@@ -38,23 +38,35 @@ struct RawNACP {
     u8 video_capture_mode;
     bool data_loss_confirmation;
     INSERT_PADDING_BYTES(1);
-    u64_le title_id;
+    u64_le presence_group_id;
     std::array<u8, 0x20> rating_age;
     std::array<char, 0x10> version_string;
     u64_le dlc_base_title_id;
-    u64_le title_id_2;
-    u64_le normal_save_data_size;
-    u64_le journal_sava_data_size;
-    INSERT_PADDING_BYTES(0x18);
-    u64_le product_code;
+    u64_le save_data_owner_id;
+    u64_le user_account_save_data_size;
+    u64_le user_account_save_data_journal_size;
+    u64_le device_save_data_size;
+    u64_le device_save_data_journal_size;
+    u64_le bcat_delivery_cache_storage_size;
+    char application_error_code_category[8];
     std::array<u64_le, 0x8> local_communication;
     u8 logo_type;
     u8 logo_handling;
     bool runtime_add_on_content_install;
     INSERT_PADDING_BYTES(5);
-    u64_le title_id_update;
-    std::array<u8, 0x40> bcat_passphrase;
-    INSERT_PADDING_BYTES(0xEC0);
+    u64_le seed_for_pseudo_device_id;
+    std::array<u8, 0x41> bcat_passphrase;
+    INSERT_PADDING_BYTES(7);
+    u64_le user_account_save_data_max_size;
+    u64_le user_account_save_data_max_journal_size;
+    u64_le device_save_data_max_size;
+    u64_le device_save_data_max_journal_size;
+    u64_le temporary_storage_size;
+    u64_le cache_storage_size;
+    u64_le cache_storage_journal_size;
+    u64_le cache_storage_data_and_journal_max_size;
+    u64_le cache_storage_max_index;
+    INSERT_PADDING_BYTES(0xE70);
 };
 static_assert(sizeof(RawNACP) == 0x4000, "RawNACP has incorrect size.");
 

src/core/file_sys/fsmitm_romfsbuild.cpp

@@ -23,6 +23,7 @@
  */
 
 #include <cstring>
+#include <string_view>
 #include "common/alignment.h"
 #include "common/assert.h"
 #include "core/file_sys/fsmitm_romfsbuild.h"
@@ -97,7 +98,8 @@ struct RomFSBuildFileContext {
     VirtualFile source;
 };
 
-static u32 romfs_calc_path_hash(u32 parent, std::string path, u32 start, std::size_t path_len) {
+static u32 romfs_calc_path_hash(u32 parent, std::string_view path, u32 start,
+                                std::size_t path_len) {
     u32 hash = parent ^ 123456789;
     for (u32 i = 0; i < path_len; i++) {
         hash = (hash >> 5) | (hash << 27);

src/core/file_sys/nca_metadata.cpp

@@ -10,14 +10,6 @@
 
 namespace FileSys {
 
-bool operator>=(TitleType lhs, TitleType rhs) {
-    return static_cast<std::size_t>(lhs) >= static_cast<std::size_t>(rhs);
-}
-
-bool operator<=(TitleType lhs, TitleType rhs) {
-    return static_cast<std::size_t>(lhs) <= static_cast<std::size_t>(rhs);
-}
-
 CNMT::CNMT(VirtualFile file) {
     if (file->ReadObject(&header) != sizeof(CNMTHeader))
         return;

src/core/file_sys/nca_metadata.h

@@ -29,9 +29,6 @@ enum class TitleType : u8 {
     DeltaTitle = 0x83,
 };
 
-bool operator>=(TitleType lhs, TitleType rhs);
-bool operator<=(TitleType lhs, TitleType rhs);
-
 enum class ContentRecordType : u8 {
     Meta = 0,
     Program = 1,

src/core/file_sys/program_metadata.cpp

@@ -3,7 +3,6 @@
 // Refer to the license.txt file included.
 
 #include <cstddef>
-#include <cstring>
 #include <vector>
 
 #include "common/logging/log.h"
@@ -17,28 +16,30 @@ ProgramMetadata::ProgramMetadata() = default;
 ProgramMetadata::~ProgramMetadata() = default;
 
 Loader::ResultStatus ProgramMetadata::Load(VirtualFile file) {
-    std::size_t total_size = static_cast<std::size_t>(file->GetSize());
-    if (total_size < sizeof(Header))
+    const std::size_t total_size = file->GetSize();
+    if (total_size < sizeof(Header)) {
         return Loader::ResultStatus::ErrorBadNPDMHeader;
+    }
 
-    // TODO(DarkLordZach): Use ReadObject when Header/AcidHeader becomes trivially copyable.
-    std::vector<u8> npdm_header_data = file->ReadBytes(sizeof(Header));
-    if (sizeof(Header) != npdm_header_data.size())
+    if (sizeof(Header) != file->ReadObject(&npdm_header)) {
         return Loader::ResultStatus::ErrorBadNPDMHeader;
-    std::memcpy(&npdm_header, npdm_header_data.data(), sizeof(Header));
+    }
 
-    std::vector<u8> acid_header_data = file->ReadBytes(sizeof(AcidHeader), npdm_header.acid_offset);
-    if (sizeof(AcidHeader) != acid_header_data.size())
+    if (sizeof(AcidHeader) != file->ReadObject(&acid_header, npdm_header.acid_offset)) {
         return Loader::ResultStatus::ErrorBadACIDHeader;
-    std::memcpy(&acid_header, acid_header_data.data(), sizeof(AcidHeader));
+    }
 
-    if (sizeof(AciHeader) != file->ReadObject(&aci_header, npdm_header.aci_offset))
+    if (sizeof(AciHeader) != file->ReadObject(&aci_header, npdm_header.aci_offset)) {
         return Loader::ResultStatus::ErrorBadACIHeader;
+    }
 
-    if (sizeof(FileAccessControl) != file->ReadObject(&acid_file_access, acid_header.fac_offset))
+    if (sizeof(FileAccessControl) != file->ReadObject(&acid_file_access, acid_header.fac_offset)) {
         return Loader::ResultStatus::ErrorBadFileAccessControl;
-    if (sizeof(FileAccessHeader) != file->ReadObject(&aci_file_access, aci_header.fah_offset))
+    }
+
+    if (sizeof(FileAccessHeader) != file->ReadObject(&aci_file_access, aci_header.fah_offset)) {
         return Loader::ResultStatus::ErrorBadFileAccessHeader;
+    }
 
     aci_kernel_capabilities.resize(aci_header.kac_size / sizeof(u32));
     const u64 read_size = aci_header.kac_size;

src/core/file_sys/program_metadata.h

@@ -58,7 +58,6 @@ public:
     void Print() const;
 
 private:
-    // TODO(DarkLordZach): BitField is not trivially copyable.
     struct Header {
         std::array<char, 4> magic;
         std::array<u8, 8> reserved;
@@ -85,7 +84,6 @@ private:
 
     static_assert(sizeof(Header) == 0x80, "NPDM header structure size is wrong");
 
-    // TODO(DarkLordZach): BitField is not trivially copyable.
     struct AcidHeader {
         std::array<u8, 0x100> signature;
         std::array<u8, 0x100> nca_modulus;

src/core/file_sys/savedata_factory.cpp

@@ -16,8 +16,10 @@ namespace FileSys {
 constexpr char SAVE_DATA_SIZE_FILENAME[] = ".yuzu_save_size";
 
 std::string SaveDataDescriptor::DebugInfo() const {
-    return fmt::format("[type={:02X}, title_id={:016X}, user_id={:016X}{:016X}, save_id={:016X}]",
-                       static_cast<u8>(type), title_id, user_id[1], user_id[0], save_id);
+    return fmt::format("[type={:02X}, title_id={:016X}, user_id={:016X}{:016X}, save_id={:016X}, "
+                       "rank={}, index={}]",
+                       static_cast<u8>(type), title_id, user_id[1], user_id[0], save_id,
+                       static_cast<u8>(rank), index);
 }
 
 SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save_directory)) {
@@ -28,7 +30,7 @@ SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save
 
 SaveDataFactory::~SaveDataFactory() = default;
 
-ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, SaveDataDescriptor meta) {
+ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, const SaveDataDescriptor& meta) {
     if (meta.type == SaveDataType::SystemSaveData || meta.type == SaveDataType::SaveData) {
         if (meta.zero_1 != 0) {
             LOG_WARNING(Service_FS,

src/core/file_sys/savedata_factory.h

@@ -32,12 +32,19 @@ enum class SaveDataType : u8 {
     CacheStorage = 5,
 };
 
+enum class SaveDataRank : u8 {
+    Primary,
+    Secondary,
+};
+
 struct SaveDataDescriptor {
     u64_le title_id;
     u128 user_id;
     u64_le save_id;
     SaveDataType type;
-    INSERT_PADDING_BYTES(7);
+    SaveDataRank rank;
+    u16_le index;
+    INSERT_PADDING_BYTES(4);
     u64_le zero_1;
     u64_le zero_2;
     u64_le zero_3;
@@ -57,7 +64,7 @@ public:
     explicit SaveDataFactory(VirtualDir dir);
     ~SaveDataFactory();
 
-    ResultVal<VirtualDir> Open(SaveDataSpaceId space, SaveDataDescriptor meta);
+    ResultVal<VirtualDir> Open(SaveDataSpaceId space, const SaveDataDescriptor& meta);
 
     VirtualDir GetSaveDataSpaceDirectory(SaveDataSpaceId space) const;
 

src/core/gdbstub/gdbstub.cpp

@@ -1030,7 +1030,7 @@ static void Step() {
 
 /// Tell the CPU if we hit a memory breakpoint.
 bool IsMemoryBreak() {
-    if (IsConnected()) {
+    if (!IsConnected()) {
         return false;
     }
 

src/core/hle/ipc_helpers.h

@@ -139,10 +139,8 @@ public:
             context->AddDomainObject(std::move(iface));
         } else {
             auto& kernel = Core::System::GetInstance().Kernel();
-            auto sessions =
+            auto [server, client] =
                 Kernel::ServerSession::CreateSessionPair(kernel, iface->GetServiceName());
-            auto server = std::get<Kernel::SharedPtr<Kernel::ServerSession>>(sessions);
-            auto client = std::get<Kernel::SharedPtr<Kernel::ClientSession>>(sessions);
             iface->ClientConnected(server);
             context->AddMoveObject(std::move(client));
         }

src/core/hle/kernel/client_port.cpp

@@ -2,8 +2,6 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <tuple>
-
 #include "core/hle/kernel/client_port.h"
 #include "core/hle/kernel/client_session.h"
 #include "core/hle/kernel/errors.h"
@@ -31,18 +29,18 @@ ResultVal<SharedPtr<ClientSession>> ClientPort::Connect() {
     active_sessions++;
 
     // Create a new session pair, let the created sessions inherit the parent port's HLE handler.
-    auto sessions = ServerSession::CreateSessionPair(kernel, server_port->GetName(), this);
+    auto [server, client] = ServerSession::CreateSessionPair(kernel, server_port->GetName(), this);
 
     if (server_port->HasHLEHandler()) {
-        server_port->GetHLEHandler()->ClientConnected(std::get<SharedPtr<ServerSession>>(sessions));
+        server_port->GetHLEHandler()->ClientConnected(server);
     } else {
-        server_port->AppendPendingSession(std::get<SharedPtr<ServerSession>>(sessions));
+        server_port->AppendPendingSession(server);
     }
 
     // Wake the threads waiting on the ServerPort
     server_port->WakeupAllWaitingThreads();
 
-    return MakeResult(std::get<SharedPtr<ClientSession>>(sessions));
+    return MakeResult(client);
 }
 
 void ClientPort::ConnectionClosed() {

src/core/hle/kernel/kernel.cpp

@@ -21,6 +21,7 @@
 #include "core/hle/kernel/thread.h"
 #include "core/hle/lock.h"
 #include "core/hle/result.h"
+#include "core/memory.h"
 
 namespace Kernel {
 
@@ -115,7 +116,7 @@ struct KernelCore::Impl {
 
     // Creates the default system resource limit
     void InitializeSystemResourceLimit(KernelCore& kernel) {
-        system_resource_limit = ResourceLimit::Create(kernel, "System");
+        system_resource_limit = ResourceLimit::Create(kernel);
 
         // If setting the default system values fails, then something seriously wrong has occurred.
         ASSERT(system_resource_limit->SetLimitValue(ResourceType::PhysicalMemory, 0x200000000)
@@ -181,6 +182,7 @@ void KernelCore::AppendNewProcess(SharedPtr<Process> process) {
 
 void KernelCore::MakeCurrentProcess(Process* process) {
     impl->current_process = process;
+    Memory::SetCurrentPageTable(&process->VMManager().page_table);
 }
 
 Process* KernelCore::CurrentProcess() {
@@ -191,6 +193,10 @@ const Process* KernelCore::CurrentProcess() const {
     return impl->current_process;
 }
 
+const std::vector<SharedPtr<Process>>& KernelCore::GetProcessList() const {
+    return impl->process_list;
+}
+
 void KernelCore::AddNamedPort(std::string name, SharedPtr<ClientPort> port) {
     impl->named_ports.emplace(std::move(name), std::move(port));
 }

src/core/hle/kernel/kernel.h

@@ -72,6 +72,9 @@ public:
     /// Retrieves a const pointer to the current process.
     const Process* CurrentProcess() const;
 
+    /// Retrieves the list of processes.
+    const std::vector<SharedPtr<Process>>& GetProcessList() const;
+
     /// Adds a port to the named port table
     void AddNamedPort(std::string name, SharedPtr<ClientPort> port);
 

src/core/hle/kernel/object.cpp

@@ -24,7 +24,6 @@ bool Object::IsWaitable() const {
     case HandleType::WritableEvent:
     case HandleType::SharedMemory:
     case HandleType::TransferMemory:
-    case HandleType::AddressArbiter:
     case HandleType::ResourceLimit:
     case HandleType::ClientPort:
     case HandleType::ClientSession:

src/core/hle/kernel/object.h

@@ -25,7 +25,6 @@ enum class HandleType : u32 {
     TransferMemory,
     Thread,
     Process,
-    AddressArbiter,
     ResourceLimit,
     ClientPort,
     ServerPort,

src/core/hle/kernel/process.cpp

@@ -32,9 +32,6 @@ namespace {
  * @param priority The priority to give the main thread
  */
 void SetupMainThread(Process& owner_process, KernelCore& kernel, VAddr entry_point, u32 priority) {
-    // Setup page table so we can write to memory
-    Memory::SetCurrentPageTable(&owner_process.VMManager().page_table);
-
     // Initialize new "main" thread
     const VAddr stack_top = owner_process.VMManager().GetTLSIORegionEndAddress();
     auto thread_res = Thread::Create(kernel, "main", entry_point, priority, 0,
@@ -80,6 +77,14 @@ u64 Process::GetTotalPhysicalMemoryUsed() const {
     return vm_manager.GetCurrentHeapSize() + main_thread_stack_size + code_memory_size;
 }
 
+void Process::RegisterThread(const Thread* thread) {
+    thread_list.push_back(thread);
+}
+
+void Process::UnregisterThread(const Thread* thread) {
+    thread_list.remove(thread);
+}
+
 ResultCode Process::ClearSignalState() {
     if (status == ProcessStatus::Exited) {
         LOG_ERROR(Kernel, "called on a terminated process instance.");
@@ -249,7 +254,7 @@ void Process::Acquire(Thread* thread) {
     ASSERT_MSG(!ShouldWait(thread), "Object unavailable!");
 }
 
-bool Process::ShouldWait(Thread* thread) const {
+bool Process::ShouldWait(const Thread* thread) const {
     return !is_signaled;
 }
 

src/core/hle/kernel/process.h

@@ -7,6 +7,7 @@
 #include <array>
 #include <bitset>
 #include <cstddef>
+#include <list>
 #include <string>
 #include <vector>
 #include <boost/container/static_vector.hpp>
@@ -189,6 +190,19 @@ public:
     /// Retrieves the total physical memory used by this process in bytes.
     u64 GetTotalPhysicalMemoryUsed() const;
 
+    /// Gets the list of all threads created with this process as their owner.
+    const std::list<const Thread*>& GetThreadList() const {
+        return thread_list;
+    }
+
+    /// Registers a thread as being created under this process,
+    /// adding it to this process' thread list.
+    void RegisterThread(const Thread* thread);
+
+    /// Unregisters a thread from this process, removing it
+    /// from this process' thread list.
+    void UnregisterThread(const Thread* thread);
+
     /// Clears the signaled state of the process if and only if it's signaled.
     ///
     /// @pre The process must not be already terminated. If this is called on a
@@ -237,7 +251,7 @@ private:
     ~Process() override;
 
     /// Checks if the specified thread should wait until this process is available.
-    bool ShouldWait(Thread* thread) const override;
+    bool ShouldWait(const Thread* thread) const override;
 
     /// Acquires/locks this process for the specified thread if it's available.
     void Acquire(Thread* thread) override;
@@ -308,6 +322,9 @@ private:
     /// Random values for svcGetInfo RandomEntropy
     std::array<u64, RANDOM_ENTROPY_SIZE> random_entropy;
 
+    /// List of threads that are running with this process as their owner.
+    std::list<const Thread*> thread_list;
+
     /// System context
     Core::System& system;
 

src/core/hle/kernel/readable_event.cpp

@@ -14,7 +14,7 @@ namespace Kernel {
 ReadableEvent::ReadableEvent(KernelCore& kernel) : WaitObject{kernel} {}
 ReadableEvent::~ReadableEvent() = default;
 
-bool ReadableEvent::ShouldWait(Thread* thread) const {
+bool ReadableEvent::ShouldWait(const Thread* thread) const {
     return !signaled;
 }
 

src/core/hle/kernel/readable_event.h

@@ -36,7 +36,7 @@ public:
         return HANDLE_TYPE;
     }
 
-    bool ShouldWait(Thread* thread) const override;
+    bool ShouldWait(const Thread* thread) const override;
     void Acquire(Thread* thread) override;
 
     /// Unconditionally clears the readable event's state.

src/core/hle/kernel/resource_limit.cpp

@@ -16,11 +16,8 @@ constexpr std::size_t ResourceTypeToIndex(ResourceType type) {
 ResourceLimit::ResourceLimit(KernelCore& kernel) : Object{kernel} {}
 ResourceLimit::~ResourceLimit() = default;
 
-SharedPtr<ResourceLimit> ResourceLimit::Create(KernelCore& kernel, std::string name) {
-    SharedPtr<ResourceLimit> resource_limit(new ResourceLimit(kernel));
-
-    resource_limit->name = std::move(name);
-    return resource_limit;
+SharedPtr<ResourceLimit> ResourceLimit::Create(KernelCore& kernel) {
+    return new ResourceLimit(kernel);
 }
 
 s64 ResourceLimit::GetCurrentResourceValue(ResourceType resource) const {

src/core/hle/kernel/resource_limit.h

@@ -31,16 +31,14 @@ constexpr bool IsValidResourceType(ResourceType type) {
 
 class ResourceLimit final : public Object {
 public:
-    /**
-     * Creates a resource limit object.
-     */
-    static SharedPtr<ResourceLimit> Create(KernelCore& kernel, std::string name = "Unknown");
+    /// Creates a resource limit object.
+    static SharedPtr<ResourceLimit> Create(KernelCore& kernel);
 
     std::string GetTypeName() const override {
         return "ResourceLimit";
     }
     std::string GetName() const override {
-        return name;
+        return GetTypeName();
     }
 
     static const HandleType HANDLE_TYPE = HandleType::ResourceLimit;
@@ -95,9 +93,6 @@ private:
     ResourceArray limits{};
     /// Current resource limit values.
     ResourceArray values{};
-
-    /// Name of resource limit object.
-    std::string name;
 };
 
 } // namespace Kernel

src/core/hle/kernel/scheduler.cpp

@@ -101,7 +101,6 @@ void Scheduler::SwitchContext(Thread* new_thread) {
         auto* const thread_owner_process = current_thread->GetOwnerProcess();
         if (previous_process != thread_owner_process) {
             system.Kernel().MakeCurrentProcess(thread_owner_process);
-            Memory::SetCurrentPageTable(&thread_owner_process->VMManager().page_table);
         }
 
         cpu_core.LoadContext(new_thread->GetContext());

src/core/hle/kernel/server_port.cpp

@@ -30,7 +30,7 @@ void ServerPort::AppendPendingSession(SharedPtr<ServerSession> pending_session)
     pending_sessions.push_back(std::move(pending_session));
 }
 
-bool ServerPort::ShouldWait(Thread* thread) const {
+bool ServerPort::ShouldWait(const Thread* thread) const {
     // If there are no pending sessions, we wait until a new one is added.
     return pending_sessions.empty();
 }
@@ -39,9 +39,8 @@ void ServerPort::Acquire(Thread* thread) {
     ASSERT_MSG(!ShouldWait(thread), "object unavailable!");
 }
 
-std::tuple<SharedPtr<ServerPort>, SharedPtr<ClientPort>> ServerPort::CreatePortPair(
-    KernelCore& kernel, u32 max_sessions, std::string name) {
-
+ServerPort::PortPair ServerPort::CreatePortPair(KernelCore& kernel, u32 max_sessions,
+                                                std::string name) {
     SharedPtr<ServerPort> server_port(new ServerPort(kernel));
     SharedPtr<ClientPort> client_port(new ClientPort(kernel));
 
@@ -51,7 +50,7 @@ std::tuple<SharedPtr<ServerPort>, SharedPtr<ClientPort>> ServerPort::CreatePortP
     client_port->max_sessions = max_sessions;
     client_port->active_sessions = 0;
 
-    return std::make_tuple(std::move(server_port), std::move(client_port));
+    return std::make_pair(std::move(server_port), std::move(client_port));
 }
 
 } // namespace Kernel

src/core/hle/kernel/server_port.h

@@ -6,7 +6,7 @@
 
 #include <memory>
 #include <string>
-#include <tuple>
+#include <utility>
 #include <vector>
 #include "common/common_types.h"
 #include "core/hle/kernel/object.h"
@@ -23,6 +23,7 @@ class SessionRequestHandler;
 class ServerPort final : public WaitObject {
 public:
     using HLEHandler = std::shared_ptr<SessionRequestHandler>;
+    using PortPair = std::pair<SharedPtr<ServerPort>, SharedPtr<ClientPort>>;
 
     /**
      * Creates a pair of ServerPort and an associated ClientPort.
@@ -32,8 +33,8 @@ public:
      * @param name Optional name of the ports
      * @return The created port tuple
      */
-    static std::tuple<SharedPtr<ServerPort>, SharedPtr<ClientPort>> CreatePortPair(
-        KernelCore& kernel, u32 max_sessions, std::string name = "UnknownPort");
+    static PortPair CreatePortPair(KernelCore& kernel, u32 max_sessions,
+                                   std::string name = "UnknownPort");
 
     std::string GetTypeName() const override {
         return "ServerPort";
@@ -75,7 +76,7 @@ public:
     /// waiting to be accepted by this port.
     void AppendPendingSession(SharedPtr<ServerSession> pending_session);
 
-    bool ShouldWait(Thread* thread) const override;
+    bool ShouldWait(const Thread* thread) const override;
     void Acquire(Thread* thread) override;
 
 private:

src/core/hle/kernel/server_session.cpp

@@ -46,7 +46,7 @@ ResultVal<SharedPtr<ServerSession>> ServerSession::Create(KernelCore& kernel, st
     return MakeResult(std::move(server_session));
 }
 
-bool ServerSession::ShouldWait(Thread* thread) const {
+bool ServerSession::ShouldWait(const Thread* thread) const {
     // Closed sessions should never wait, an error will be returned from svcReplyAndReceive.
     if (parent->client == nullptr)
         return false;
@@ -204,6 +204,6 @@ ServerSession::SessionPair ServerSession::CreateSessionPair(KernelCore& kernel,
     client_session->parent = parent;
     server_session->parent = parent;
 
-    return std::make_tuple(std::move(server_session), std::move(client_session));
+    return std::make_pair(std::move(server_session), std::move(client_session));
 }
 } // namespace Kernel

src/core/hle/kernel/server_session.h

@@ -6,6 +6,7 @@
 
 #include <memory>
 #include <string>
+#include <utility>
 #include <vector>
 
 #include "core/hle/kernel/object.h"
@@ -41,6 +42,10 @@ public:
         return "ServerSession";
     }
 
+    std::string GetName() const override {
+        return name;
+    }
+
     static const HandleType HANDLE_TYPE = HandleType::ServerSession;
     HandleType GetHandleType() const override {
         return HANDLE_TYPE;
@@ -54,7 +59,7 @@ public:
         return parent.get();
     }
 
-    using SessionPair = std::tuple<SharedPtr<ServerSession>, SharedPtr<ClientSession>>;
+    using SessionPair = std::pair<SharedPtr<ServerSession>, SharedPtr<ClientSession>>;
 
     /**
      * Creates a pair of ServerSession and an associated ClientSession.
@@ -82,7 +87,7 @@ public:
      */
     ResultCode HandleSyncRequest(SharedPtr<Thread> thread);
 
-    bool ShouldWait(Thread* thread) const override;
+    bool ShouldWait(const Thread* thread) const override;
 
     void Acquire(Thread* thread) override;
 

src/core/hle/kernel/shared_memory.cpp

@@ -9,7 +9,6 @@
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/shared_memory.h"
-#include "core/memory.h"
 
 namespace Kernel {
 
@@ -119,7 +118,15 @@ ResultCode SharedMemory::Map(Process& target_process, VAddr address, MemoryPermi
                                                      ConvertPermissions(permissions));
 }
 
-ResultCode SharedMemory::Unmap(Process& target_process, VAddr address) {
+ResultCode SharedMemory::Unmap(Process& target_process, VAddr address, u64 unmap_size) {
+    if (unmap_size != size) {
+        LOG_ERROR(Kernel,
+                  "Invalid size passed to Unmap. Size must be equal to the size of the "
+                  "memory managed. Shared memory size=0x{:016X}, Unmap size=0x{:016X}",
+                  size, unmap_size);
+        return ERR_INVALID_SIZE;
+    }
+
     // TODO(Subv): Verify what happens if the application tries to unmap an address that is not
     // mapped to a SharedMemory.
     return target_process.VMManager().UnmapRange(address, size);

src/core/hle/kernel/shared_memory.h

@@ -104,11 +104,17 @@ public:
 
     /**
      * Unmaps a shared memory block from the specified address in system memory
+     *
      * @param target_process Process from which to unmap the memory block.
-     * @param address Address in system memory where the shared memory block is mapped
+     * @param address        Address in system memory where the shared memory block is mapped.
+     * @param unmap_size     The amount of bytes to unmap from this shared memory instance.
+     *
      * @return Result code of the unmap operation
+     *
+     * @pre The given size to unmap must be the same size as the amount of memory managed by
+     *      the SharedMemory instance itself, otherwise ERR_INVALID_SIZE will be returned.
      */
-    ResultCode Unmap(Process& target_process, VAddr address);
+    ResultCode Unmap(Process& target_process, VAddr address, u64 unmap_size);
 
     /**
      * Gets a pointer to the shared memory block

src/core/hle/kernel/svc.h

@@ -6,8 +6,12 @@
 
 #include "common/common_types.h"
 
+namespace Core {
+class System;
+}
+
 namespace Kernel {
 
-void CallSVC(u32 immediate);
+void CallSVC(Core::System& system, u32 immediate);
 
 } // namespace Kernel

src/core/hle/kernel/svc_wrap.h

@@ -11,270 +11,312 @@
 
 namespace Kernel {
 
-static inline u64 Param(int n) {
-    return Core::CurrentArmInterface().GetReg(n);
+static inline u64 Param(const Core::System& system, int n) {
+    return system.CurrentArmInterface().GetReg(n);
 }
 
 /**
  * HLE a function return from the current ARM userland process
- * @param res Result to return
+ * @param system System context
+ * @param result Result to return
  */
-static inline void FuncReturn(u64 res) {
-    Core::CurrentArmInterface().SetReg(0, res);
+static inline void FuncReturn(Core::System& system, u64 result) {
+    system.CurrentArmInterface().SetReg(0, result);
 }
 
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 // Function wrappers that return type ResultCode
 
-template <ResultCode func(u64)>
-void SvcWrap() {
-    FuncReturn(func(Param(0)).raw);
+template <ResultCode func(Core::System&, u64)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, Param(system, 0)).raw);
 }
 
-template <ResultCode func(u32)>
-void SvcWrap() {
-    FuncReturn(func(static_cast<u32>(Param(0))).raw);
+template <ResultCode func(Core::System&, u32)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, static_cast<u32>(Param(system, 0))).raw);
 }
 
-template <ResultCode func(u32, u32)>
-void SvcWrap() {
-    FuncReturn(func(static_cast<u32>(Param(0)), static_cast<u32>(Param(1))).raw);
+template <ResultCode func(Core::System&, u32, u32)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(
+        system,
+        func(system, static_cast<u32>(Param(system, 0)), static_cast<u32>(Param(system, 1))).raw);
 }
 
-template <ResultCode func(u32*)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32*)>
+void SvcWrap(Core::System& system) {
     u32 param = 0;
-    const u32 retval = func(&param).raw;
-    Core::CurrentArmInterface().SetReg(1, param);
-    FuncReturn(retval);
+    const u32 retval = func(system, &param).raw;
+    system.CurrentArmInterface().SetReg(1, param);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u32*, u32)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32*, u32)>
+void SvcWrap(Core::System& system) {
     u32 param_1 = 0;
-    u32 retval = func(&param_1, static_cast<u32>(Param(1))).raw;
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval);
+    const u32 retval = func(system, &param_1, static_cast<u32>(Param(system, 1))).raw;
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u32*, u32*)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32*, u32*)>
+void SvcWrap(Core::System& system) {
     u32 param_1 = 0;
     u32 param_2 = 0;
-    const u32 retval = func(&param_1, &param_2).raw;
+    const u32 retval = func(system, &param_1, &param_2).raw;
 
-    auto& arm_interface = Core::CurrentArmInterface();
+    auto& arm_interface = system.CurrentArmInterface();
     arm_interface.SetReg(1, param_1);
     arm_interface.SetReg(2, param_2);
 
-    FuncReturn(retval);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u32*, u64)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32*, u64)>
+void SvcWrap(Core::System& system) {
     u32 param_1 = 0;
-    const u32 retval = func(&param_1, Param(1)).raw;
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval);
+    const u32 retval = func(system, &param_1, Param(system, 1)).raw;
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u64*, u32)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32*, u64, u32)>
+void SvcWrap(Core::System& system) {
+    u32 param_1 = 0;
+    const u32 retval =
+        func(system, &param_1, Param(system, 1), static_cast<u32>(Param(system, 2))).raw;
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
+}
+
+template <ResultCode func(Core::System&, u64*, u32)>
+void SvcWrap(Core::System& system) {
     u64 param_1 = 0;
-    const u32 retval = func(&param_1, static_cast<u32>(Param(1))).raw;
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval);
+    const u32 retval = func(system, &param_1, static_cast<u32>(Param(system, 1))).raw;
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u64, s32)>
-void SvcWrap() {
-    FuncReturn(func(Param(0), static_cast<s32>(Param(1))).raw);
+template <ResultCode func(Core::System&, u64, s32)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, Param(system, 0), static_cast<s32>(Param(system, 1))).raw);
 }
 
-template <ResultCode func(u64, u32)>
-void SvcWrap() {
-    FuncReturn(func(Param(0), static_cast<u32>(Param(1))).raw);
+template <ResultCode func(Core::System&, u64, u32)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, Param(system, 0), static_cast<u32>(Param(system, 1))).raw);
 }
 
-template <ResultCode func(u64*, u64)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u64*, u64)>
+void SvcWrap(Core::System& system) {
     u64 param_1 = 0;
-    u32 retval = func(&param_1, Param(1)).raw;
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval);
+    const u32 retval = func(system, &param_1, Param(system, 1)).raw;
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u64*, u32, u32)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u64*, u32, u32)>
+void SvcWrap(Core::System& system) {
     u64 param_1 = 0;
-    u32 retval = func(&param_1, static_cast<u32>(Param(1)), static_cast<u32>(Param(2))).raw;
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval);
+    const u32 retval = func(system, &param_1, static_cast<u32>(Param(system, 1)),
+                            static_cast<u32>(Param(system, 2)))
+                           .raw;
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u32, u64)>
-void SvcWrap() {
-    FuncReturn(func(static_cast<u32>(Param(0)), Param(1)).raw);
+template <ResultCode func(Core::System&, u32, u64)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)), Param(system, 1)).raw);
 }
 
-template <ResultCode func(u32, u32, u64)>
-void SvcWrap() {
-    FuncReturn(func(static_cast<u32>(Param(0)), static_cast<u32>(Param(1)), Param(2)).raw);
+template <ResultCode func(Core::System&, u32, u32, u64)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)),
+                            static_cast<u32>(Param(system, 1)), Param(system, 2))
+                           .raw);
 }
 
-template <ResultCode func(u32, u32*, u64*)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32, u32*, u64*)>
+void SvcWrap(Core::System& system) {
     u32 param_1 = 0;
     u64 param_2 = 0;
-    ResultCode retval = func(static_cast<u32>(Param(2)), &param_1, &param_2);
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    Core::CurrentArmInterface().SetReg(2, param_2);
-    FuncReturn(retval.raw);
+    const ResultCode retval = func(system, static_cast<u32>(Param(system, 2)), &param_1, &param_2);
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    system.CurrentArmInterface().SetReg(2, param_2);
+    FuncReturn(system, retval.raw);
 }
 
-template <ResultCode func(u64, u64, u32, u32)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u64, u64, u32, u32)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, Param(system, 0), Param(system, 1),
+                            static_cast<u32>(Param(system, 2)), static_cast<u32>(Param(system, 3)))
+                           .raw);
+}
+
+template <ResultCode func(Core::System&, u64, u64, u32, u64)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, Param(system, 0), Param(system, 1),
+                            static_cast<u32>(Param(system, 2)), Param(system, 3))
+                           .raw);
+}
+
+template <ResultCode func(Core::System&, u32, u64, u32)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)), Param(system, 1),
+                            static_cast<u32>(Param(system, 2)))
+                           .raw);
+}
+
+template <ResultCode func(Core::System&, u64, u64, u64)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, Param(system, 0), Param(system, 1), Param(system, 2)).raw);
+}
+
+template <ResultCode func(Core::System&, u64, u64, u32)>
+void SvcWrap(Core::System& system) {
     FuncReturn(
-        func(Param(0), Param(1), static_cast<u32>(Param(2)), static_cast<u32>(Param(3))).raw);
+        system,
+        func(system, Param(system, 0), Param(system, 1), static_cast<u32>(Param(system, 2))).raw);
 }
 
-template <ResultCode func(u64, u64, u32, u64)>
-void SvcWrap() {
-    FuncReturn(func(Param(0), Param(1), static_cast<u32>(Param(2)), Param(3)).raw);
+template <ResultCode func(Core::System&, u32, u64, u64, u32)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)), Param(system, 1),
+                            Param(system, 2), static_cast<u32>(Param(system, 3)))
+                           .raw);
 }
 
-template <ResultCode func(u32, u64, u32)>
-void SvcWrap() {
-    FuncReturn(func(static_cast<u32>(Param(0)), Param(1), static_cast<u32>(Param(2))).raw);
-}
-
-template <ResultCode func(u64, u64, u64)>
-void SvcWrap() {
-    FuncReturn(func(Param(0), Param(1), Param(2)).raw);
-}
-
-template <ResultCode func(u64, u64, u32)>
-void SvcWrap() {
-    FuncReturn(func(Param(0), Param(1), static_cast<u32>(Param(2))).raw);
-}
-
-template <ResultCode func(u32, u64, u64, u32)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32, u64, u64)>
+void SvcWrap(Core::System& system) {
     FuncReturn(
-        func(static_cast<u32>(Param(0)), Param(1), Param(2), static_cast<u32>(Param(3))).raw);
+        system,
+        func(system, static_cast<u32>(Param(system, 0)), Param(system, 1), Param(system, 2)).raw);
 }
 
-template <ResultCode func(u32, u64, u64)>
-void SvcWrap() {
-    FuncReturn(func(static_cast<u32>(Param(0)), Param(1), Param(2)).raw);
-}
-
-template <ResultCode func(u32*, u64, u64, s64)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32*, u64, u64, s64)>
+void SvcWrap(Core::System& system) {
     u32 param_1 = 0;
-    ResultCode retval =
-        func(&param_1, Param(1), static_cast<u32>(Param(2)), static_cast<s64>(Param(3)));
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval.raw);
+    const u32 retval = func(system, &param_1, Param(system, 1), static_cast<u32>(Param(system, 2)),
+                            static_cast<s64>(Param(system, 3)))
+                           .raw;
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u64, u64, u32, s64)>
-void SvcWrap() {
-    FuncReturn(
-        func(Param(0), Param(1), static_cast<u32>(Param(2)), static_cast<s64>(Param(3))).raw);
+template <ResultCode func(Core::System&, u64, u64, u32, s64)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, Param(system, 0), Param(system, 1),
+                            static_cast<u32>(Param(system, 2)), static_cast<s64>(Param(system, 3)))
+                           .raw);
 }
 
-template <ResultCode func(u64*, u64, u64, u64)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u64*, u64, u64, u64)>
+void SvcWrap(Core::System& system) {
     u64 param_1 = 0;
-    u32 retval = func(&param_1, Param(1), Param(2), Param(3)).raw;
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval);
+    const u32 retval =
+        func(system, &param_1, Param(system, 1), Param(system, 2), Param(system, 3)).raw;
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u32*, u64, u64, u64, u32, s32)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32*, u64, u64, u64, u32, s32)>
+void SvcWrap(Core::System& system) {
     u32 param_1 = 0;
-    u32 retval = func(&param_1, Param(1), Param(2), Param(3), static_cast<u32>(Param(4)),
-                      static_cast<s32>(Param(5)))
-                     .raw;
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval);
+    const u32 retval = func(system, &param_1, Param(system, 1), Param(system, 2), Param(system, 3),
+                            static_cast<u32>(Param(system, 4)), static_cast<s32>(Param(system, 5)))
+                           .raw;
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u32*, u64, u64, u32)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, u32*, u64, u64, u32)>
+void SvcWrap(Core::System& system) {
     u32 param_1 = 0;
-    u32 retval = func(&param_1, Param(1), Param(2), static_cast<u32>(Param(3))).raw;
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval);
+    const u32 retval = func(system, &param_1, Param(system, 1), Param(system, 2),
+                            static_cast<u32>(Param(system, 3)))
+                           .raw;
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(Handle*, u64, u32, u32)>
-void SvcWrap() {
+template <ResultCode func(Core::System&, Handle*, u64, u32, u32)>
+void SvcWrap(Core::System& system) {
     u32 param_1 = 0;
-    u32 retval =
-        func(&param_1, Param(1), static_cast<u32>(Param(2)), static_cast<u32>(Param(3))).raw;
-    Core::CurrentArmInterface().SetReg(1, param_1);
-    FuncReturn(retval);
+    const u32 retval = func(system, &param_1, Param(system, 1), static_cast<u32>(Param(system, 2)),
+                            static_cast<u32>(Param(system, 3)))
+                           .raw;
+
+    system.CurrentArmInterface().SetReg(1, param_1);
+    FuncReturn(system, retval);
 }
 
-template <ResultCode func(u64, u32, s32, s64)>
-void SvcWrap() {
-    FuncReturn(func(Param(0), static_cast<u32>(Param(1)), static_cast<s32>(Param(2)),
-                    static_cast<s64>(Param(3)))
-                   .raw);
+template <ResultCode func(Core::System&, u64, u32, s32, s64)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, Param(system, 0), static_cast<u32>(Param(system, 1)),
+                            static_cast<s32>(Param(system, 2)), static_cast<s64>(Param(system, 3)))
+                           .raw);
 }
 
-template <ResultCode func(u64, u32, s32, s32)>
-void SvcWrap() {
-    FuncReturn(func(Param(0), static_cast<u32>(Param(1)), static_cast<s32>(Param(2)),
-                    static_cast<s32>(Param(3)))
-                   .raw);
+template <ResultCode func(Core::System&, u64, u32, s32, s32)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system, Param(system, 0), static_cast<u32>(Param(system, 1)),
+                            static_cast<s32>(Param(system, 2)), static_cast<s32>(Param(system, 3)))
+                           .raw);
 }
 
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 // Function wrappers that return type u32
 
-template <u32 func()>
-void SvcWrap() {
-    FuncReturn(func());
+template <u32 func(Core::System&)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system));
 }
 
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 // Function wrappers that return type u64
 
-template <u64 func()>
-void SvcWrap() {
-    FuncReturn(func());
+template <u64 func(Core::System&)>
+void SvcWrap(Core::System& system) {
+    FuncReturn(system, func(system));
 }
 
 ////////////////////////////////////////////////////////////////////////////////////////////////////
 /// Function wrappers that return type void
 
-template <void func()>
-void SvcWrap() {
-    func();
+template <void func(Core::System&)>
+void SvcWrap(Core::System& system) {
+    func(system);
 }
 
-template <void func(s64)>
-void SvcWrap() {
-    func(static_cast<s64>(Param(0)));
+template <void func(Core::System&, s64)>
+void SvcWrap(Core::System& system) {
+    func(system, static_cast<s64>(Param(system, 0)));
 }
 
-template <void func(u64, u64 len)>
-void SvcWrap() {
-    func(Param(0), Param(1));
+template <void func(Core::System&, u64, u64)>
+void SvcWrap(Core::System& system) {
+    func(system, Param(system, 0), Param(system, 1));
 }
 
-template <void func(u64, u64, u64)>
-void SvcWrap() {
-    func(Param(0), Param(1), Param(2));
+template <void func(Core::System&, u64, u64, u64)>
+void SvcWrap(Core::System& system) {
+    func(system, Param(system, 0), Param(system, 1), Param(system, 2));
 }
 
-template <void func(u32, u64, u64)>
-void SvcWrap() {
-    func(static_cast<u32>(Param(0)), Param(1), Param(2));
+template <void func(Core::System&, u32, u64, u64)>
+void SvcWrap(Core::System& system) {
+    func(system, static_cast<u32>(Param(system, 0)), Param(system, 1), Param(system, 2));
 }
 
 } // namespace Kernel

src/core/hle/kernel/thread.cpp

@@ -28,7 +28,7 @@
 
 namespace Kernel {
 
-bool Thread::ShouldWait(Thread* thread) const {
+bool Thread::ShouldWait(const Thread* thread) const {
     return status != ThreadStatus::Dead;
 }
 
@@ -62,6 +62,8 @@ void Thread::Stop() {
     }
     wait_objects.clear();
 
+    owner_process->UnregisterThread(this);
+
     // Mark the TLS slot in the thread's page as free.
     owner_process->FreeTLSSlot(tls_address);
 }
@@ -202,6 +204,8 @@ ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name
     thread->scheduler->AddThread(thread);
     thread->tls_address = thread->owner_process->MarkNextAvailableTLSSlotAsUsed(*thread);
 
+    thread->owner_process->RegisterThread(thread.get());
+
     // TODO(peachum): move to ScheduleThread() when scheduler is added so selected core is used
     // to initialize the context
     ResetThreadContext(thread->context, stack_top, entry_point, arg);
@@ -229,16 +233,16 @@ void Thread::SetWaitSynchronizationOutput(s32 output) {
     context.cpu_registers[1] = output;
 }
 
-s32 Thread::GetWaitObjectIndex(WaitObject* object) const {
+s32 Thread::GetWaitObjectIndex(const WaitObject* object) const {
     ASSERT_MSG(!wait_objects.empty(), "Thread is not waiting for anything");
-    auto match = std::find(wait_objects.rbegin(), wait_objects.rend(), object);
+    const auto match = std::find(wait_objects.rbegin(), wait_objects.rend(), object);
     return static_cast<s32>(std::distance(match, wait_objects.rend()) - 1);
 }
 
 VAddr Thread::GetCommandBufferAddress() const {
     // Offset from the start of TLS at which the IPC command buffer begins.
-    static constexpr int CommandHeaderOffset = 0x80;
-    return GetTLSAddress() + CommandHeaderOffset;
+    constexpr u64 command_header_offset = 0x80;
+    return GetTLSAddress() + command_header_offset;
 }
 
 void Thread::SetStatus(ThreadStatus new_status) {
@@ -367,7 +371,7 @@ void Thread::ChangeScheduler() {
     system.CpuCore(processor_id).PrepareReschedule();
 }
 
-bool Thread::AllWaitObjectsReady() {
+bool Thread::AllWaitObjectsReady() const {
     return std::none_of(
         wait_objects.begin(), wait_objects.end(),
         [this](const SharedPtr<WaitObject>& object) { return object->ShouldWait(this); });

src/core/hle/kernel/thread.h

@@ -111,7 +111,7 @@ public:
         return HANDLE_TYPE;
     }
 
-    bool ShouldWait(Thread* thread) const override;
+    bool ShouldWait(const Thread* thread) const override;
     void Acquire(Thread* thread) override;
 
     /**
@@ -205,7 +205,7 @@ public:
      * object in the list.
      * @param object Object to query the index of.
      */
-    s32 GetWaitObjectIndex(WaitObject* object) const;
+    s32 GetWaitObjectIndex(const WaitObject* object) const;
 
     /**
      * Stops a thread, invalidating it from further use
@@ -299,7 +299,7 @@ public:
     }
 
     /// Determines whether all the objects this thread is waiting on are ready.
-    bool AllWaitObjectsReady();
+    bool AllWaitObjectsReady() const;
 
     const MutexWaitingThreads& GetMutexWaitingThreads() const {
         return wait_mutex_threads;

src/core/hle/kernel/transfer_memory.cpp

@@ -14,8 +14,8 @@ namespace Kernel {
 TransferMemory::TransferMemory(KernelCore& kernel) : Object{kernel} {}
 TransferMemory::~TransferMemory() = default;
 
-SharedPtr<TransferMemory> TransferMemory::Create(KernelCore& kernel, VAddr base_address,
-                                                 size_t size, MemoryPermission permissions) {
+SharedPtr<TransferMemory> TransferMemory::Create(KernelCore& kernel, VAddr base_address, u64 size,
+                                                 MemoryPermission permissions) {
     SharedPtr<TransferMemory> transfer_memory{new TransferMemory(kernel)};
 
     transfer_memory->base_address = base_address;
@@ -26,7 +26,15 @@ SharedPtr<TransferMemory> TransferMemory::Create(KernelCore& kernel, VAddr base_
     return transfer_memory;
 }
 
-ResultCode TransferMemory::MapMemory(VAddr address, size_t size, MemoryPermission permissions) {
+const u8* TransferMemory::GetPointer() const {
+    return backing_block.get()->data();
+}
+
+u64 TransferMemory::GetSize() const {
+    return memory_size;
+}
+
+ResultCode TransferMemory::MapMemory(VAddr address, u64 size, MemoryPermission permissions) {
     if (memory_size != size) {
         return ERR_INVALID_SIZE;
     }
@@ -39,13 +47,13 @@ ResultCode TransferMemory::MapMemory(VAddr address, size_t size, MemoryPermissio
         return ERR_INVALID_STATE;
     }
 
+    backing_block = std::make_shared<std::vector<u8>>(size);
+
     const auto map_state = owner_permissions == MemoryPermission::None
                                ? MemoryState::TransferMemoryIsolated
                                : MemoryState::TransferMemory;
     auto& vm_manager = owner_process->VMManager();
-    const auto map_result = vm_manager.MapMemoryBlock(
-        address, std::make_shared<std::vector<u8>>(size), 0, size, map_state);
-
+    const auto map_result = vm_manager.MapMemoryBlock(address, backing_block, 0, size, map_state);
     if (map_result.Failed()) {
         return map_result.Code();
     }
@@ -54,7 +62,7 @@ ResultCode TransferMemory::MapMemory(VAddr address, size_t size, MemoryPermissio
     return RESULT_SUCCESS;
 }
 
-ResultCode TransferMemory::UnmapMemory(VAddr address, size_t size) {
+ResultCode TransferMemory::UnmapMemory(VAddr address, u64 size) {
     if (memory_size != size) {
         return ERR_INVALID_SIZE;
     }

src/core/hle/kernel/transfer_memory.h

@@ -4,6 +4,9 @@
 
 #pragma once
 
+#include <memory>
+#include <vector>
+
 #include "core/hle/kernel/object.h"
 
 union ResultCode;
@@ -25,7 +28,7 @@ class TransferMemory final : public Object {
 public:
     static constexpr HandleType HANDLE_TYPE = HandleType::TransferMemory;
 
-    static SharedPtr<TransferMemory> Create(KernelCore& kernel, VAddr base_address, size_t size,
+    static SharedPtr<TransferMemory> Create(KernelCore& kernel, VAddr base_address, u64 size,
                                             MemoryPermission permissions);
 
     TransferMemory(const TransferMemory&) = delete;
@@ -46,6 +49,12 @@ public:
         return HANDLE_TYPE;
     }
 
+    /// Gets a pointer to the backing block of this instance.
+    const u8* GetPointer() const;
+
+    /// Gets the size of the memory backing this instance in bytes.
+    u64 GetSize() const;
+
     /// Attempts to map transfer memory with the given range and memory permissions.
     ///
     /// @param address     The base address to being mapping memory at.
@@ -56,7 +65,7 @@ public:
     ///      the same values that were given when creating the transfer memory
     ///      instance.
     ///
-    ResultCode MapMemory(VAddr address, size_t size, MemoryPermission permissions);
+    ResultCode MapMemory(VAddr address, u64 size, MemoryPermission permissions);
 
     /// Unmaps the transfer memory with the given range
     ///
@@ -66,17 +75,20 @@ public:
     /// @pre The given address and size must be the same as the ones used
     ///      to create the transfer memory instance.
     ///
-    ResultCode UnmapMemory(VAddr address, size_t size);
+    ResultCode UnmapMemory(VAddr address, u64 size);
 
 private:
     explicit TransferMemory(KernelCore& kernel);
     ~TransferMemory() override;
 
+    /// Memory block backing this instance.
+    std::shared_ptr<std::vector<u8>> backing_block;
+
     /// The base address for the memory managed by this instance.
     VAddr base_address = 0;
 
     /// Size of the memory, in bytes, that this instance manages.
-    size_t memory_size = 0;
+    u64 memory_size = 0;
 
     /// The memory permissions that are applied to this instance.
     MemoryPermission owner_permissions{};

src/core/hle/kernel/wait_object.h

@@ -24,7 +24,7 @@ public:
      * @param thread The thread about which we're deciding.
      * @return True if the current thread should wait due to this object being unavailable
      */
-    virtual bool ShouldWait(Thread* thread) const = 0;
+    virtual bool ShouldWait(const Thread* thread) const = 0;
 
     /// Acquire/lock the object for the specified thread if it is available
     virtual void Acquire(Thread* thread) = 0;

src/core/hle/result.h

@@ -119,10 +119,6 @@ union ResultCode {
     BitField<0, 9, ErrorModule> module;
     BitField<9, 13, u32> description;
 
-    // The last bit of `level` is checked by apps and the kernel to determine if a result code is an
-    // error
-    BitField<31, 1, u32> is_error;
-
     constexpr explicit ResultCode(u32 raw) : raw(raw) {}
 
     constexpr ResultCode(ErrorModule module_, u32 description_)

src/core/hle/service/am/am.cpp

@@ -13,7 +13,7 @@
 #include "core/hle/kernel/kernel.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/kernel/readable_event.h"
-#include "core/hle/kernel/shared_memory.h"
+#include "core/hle/kernel/transfer_memory.h"
 #include "core/hle/kernel/writable_event.h"
 #include "core/hle/service/acc/profile_manager.h"
 #include "core/hle/service/am/am.h"
@@ -239,8 +239,8 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
         {0, nullptr, "Exit"},
         {1, &ISelfController::LockExit, "LockExit"},
         {2, &ISelfController::UnlockExit, "UnlockExit"},
-        {3, nullptr, "EnterFatalSection"},
-        {4, nullptr, "LeaveFatalSection"},
+        {3, &ISelfController::EnterFatalSection, "EnterFatalSection"},
+        {4, &ISelfController::LeaveFatalSection, "LeaveFatalSection"},
         {9, &ISelfController::GetLibraryAppletLaunchableEvent, "GetLibraryAppletLaunchableEvent"},
         {10, &ISelfController::SetScreenShotPermission, "SetScreenShotPermission"},
         {11, &ISelfController::SetOperationModeChangedNotification, "SetOperationModeChangedNotification"},
@@ -285,41 +285,54 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
 
 ISelfController::~ISelfController() = default;
 
-void ISelfController::SetFocusHandlingMode(Kernel::HLERequestContext& ctx) {
-    // Takes 3 input u8s with each field located immediately after the previous
-    // u8, these are bool flags. No output.
-    LOG_WARNING(Service_AM, "(STUBBED) called");
-
-    IPC::RequestParser rp{ctx};
-
-    struct FocusHandlingModeParams {
-        u8 unknown0;
-        u8 unknown1;
-        u8 unknown2;
-    };
-    auto flags = rp.PopRaw<FocusHandlingModeParams>();
-
-    IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(RESULT_SUCCESS);
-}
-
-void ISelfController::SetRestartMessageEnabled(Kernel::HLERequestContext& ctx) {
+void ISelfController::LockExit(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_AM, "(STUBBED) called");
 
     IPC::ResponseBuilder rb{ctx, 2};
     rb.Push(RESULT_SUCCESS);
 }
 
-void ISelfController::SetPerformanceModeChangedNotification(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp{ctx};
-
-    bool flag = rp.Pop<bool>();
-    LOG_WARNING(Service_AM, "(STUBBED) called flag={}", flag);
+void ISelfController::UnlockExit(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_AM, "(STUBBED) called");
 
     IPC::ResponseBuilder rb{ctx, 2};
     rb.Push(RESULT_SUCCESS);
 }
 
+void ISelfController::EnterFatalSection(Kernel::HLERequestContext& ctx) {
+    ++num_fatal_sections_entered;
+    LOG_DEBUG(Service_AM, "called. Num fatal sections entered: {}", num_fatal_sections_entered);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void ISelfController::LeaveFatalSection(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_AM, "called.");
+
+    // Entry and exit of fatal sections must be balanced.
+    if (num_fatal_sections_entered == 0) {
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(ResultCode{ErrorModule::AM, 512});
+        return;
+    }
+
+    --num_fatal_sections_entered;
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void ISelfController::GetLibraryAppletLaunchableEvent(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+
+    launchable_event.writable->Signal();
+
+    IPC::ResponseBuilder rb{ctx, 2, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushCopyObjects(launchable_event.readable);
+}
+
 void ISelfController::SetScreenShotPermission(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_AM, "(STUBBED) called");
 
@@ -337,6 +350,42 @@ void ISelfController::SetOperationModeChangedNotification(Kernel::HLERequestCont
     rb.Push(RESULT_SUCCESS);
 }
 
+void ISelfController::SetPerformanceModeChangedNotification(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+
+    bool flag = rp.Pop<bool>();
+    LOG_WARNING(Service_AM, "(STUBBED) called flag={}", flag);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void ISelfController::SetFocusHandlingMode(Kernel::HLERequestContext& ctx) {
+    // Takes 3 input u8s with each field located immediately after the previous
+    // u8, these are bool flags. No output.
+    IPC::RequestParser rp{ctx};
+
+    struct FocusHandlingModeParams {
+        u8 unknown0;
+        u8 unknown1;
+        u8 unknown2;
+    };
+    const auto flags = rp.PopRaw<FocusHandlingModeParams>();
+
+    LOG_WARNING(Service_AM, "(STUBBED) called. unknown0={}, unknown1={}, unknown2={}",
+                flags.unknown0, flags.unknown1, flags.unknown2);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void ISelfController::SetRestartMessageEnabled(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
 void ISelfController::SetOutOfFocusSuspendingEnabled(Kernel::HLERequestContext& ctx) {
     // Takes 3 input u8s with each field located immediately after the previous
     // u8, these are bool flags. No output.
@@ -349,30 +398,6 @@ void ISelfController::SetOutOfFocusSuspendingEnabled(Kernel::HLERequestContext&
     rb.Push(RESULT_SUCCESS);
 }
 
-void ISelfController::LockExit(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_AM, "(STUBBED) called");
-
-    IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(RESULT_SUCCESS);
-}
-
-void ISelfController::UnlockExit(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_AM, "(STUBBED) called");
-
-    IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(RESULT_SUCCESS);
-}
-
-void ISelfController::GetLibraryAppletLaunchableEvent(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_AM, "(STUBBED) called");
-
-    launchable_event.writable->Signal();
-
-    IPC::ResponseBuilder rb{ctx, 2, 1};
-    rb.Push(RESULT_SUCCESS);
-    rb.PushCopyObjects(launchable_event.readable);
-}
-
 void ISelfController::SetScreenShotImageOrientation(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_AM, "(STUBBED) called");
 
@@ -907,19 +932,19 @@ void ILibraryAppletCreator::CreateTransferMemoryStorage(Kernel::HLERequestContex
     rp.SetCurrentOffset(3);
     const auto handle{rp.Pop<Kernel::Handle>()};
 
-    const auto shared_mem =
-        Core::System::GetInstance().CurrentProcess()->GetHandleTable().Get<Kernel::SharedMemory>(
+    const auto transfer_mem =
+        Core::System::GetInstance().CurrentProcess()->GetHandleTable().Get<Kernel::TransferMemory>(
             handle);
 
-    if (shared_mem == nullptr) {
+    if (transfer_mem == nullptr) {
         LOG_ERROR(Service_AM, "shared_mem is a nullpr for handle={:08X}", handle);
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(ResultCode(-1));
         return;
     }
 
-    const u8* mem_begin = shared_mem->GetPointer();
-    const u8* mem_end = mem_begin + shared_mem->GetSize();
+    const u8* const mem_begin = transfer_mem->GetPointer();
+    const u8* const mem_end = mem_begin + transfer_mem->GetSize();
     std::vector<u8> memory{mem_begin, mem_end};
 
     IPC::ResponseBuilder rb{ctx, 2, 0, 1};

src/core/hle/service/am/am.h

@@ -117,17 +117,19 @@ public:
     ~ISelfController() override;
 
 private:
-    void SetFocusHandlingMode(Kernel::HLERequestContext& ctx);
-    void SetRestartMessageEnabled(Kernel::HLERequestContext& ctx);
-    void SetPerformanceModeChangedNotification(Kernel::HLERequestContext& ctx);
-    void SetOperationModeChangedNotification(Kernel::HLERequestContext& ctx);
-    void SetOutOfFocusSuspendingEnabled(Kernel::HLERequestContext& ctx);
     void LockExit(Kernel::HLERequestContext& ctx);
     void UnlockExit(Kernel::HLERequestContext& ctx);
+    void EnterFatalSection(Kernel::HLERequestContext& ctx);
+    void LeaveFatalSection(Kernel::HLERequestContext& ctx);
     void GetLibraryAppletLaunchableEvent(Kernel::HLERequestContext& ctx);
+    void SetScreenShotPermission(Kernel::HLERequestContext& ctx);
+    void SetOperationModeChangedNotification(Kernel::HLERequestContext& ctx);
+    void SetPerformanceModeChangedNotification(Kernel::HLERequestContext& ctx);
+    void SetFocusHandlingMode(Kernel::HLERequestContext& ctx);
+    void SetRestartMessageEnabled(Kernel::HLERequestContext& ctx);
+    void SetOutOfFocusSuspendingEnabled(Kernel::HLERequestContext& ctx);
     void SetScreenShotImageOrientation(Kernel::HLERequestContext& ctx);
     void CreateManagedDisplayLayer(Kernel::HLERequestContext& ctx);
-    void SetScreenShotPermission(Kernel::HLERequestContext& ctx);
     void SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx);
     void SetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx);
     void GetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx);
@@ -135,6 +137,7 @@ private:
     std::shared_ptr<NVFlinger::NVFlinger> nvflinger;
     Kernel::EventPair launchable_event;
     u32 idle_time_detection_extension = 0;
+    u64 num_fatal_sections_entered = 0;
 };
 
 class ICommonStateGetter final : public ServiceFramework<ICommonStateGetter> {

src/core/hle/service/audio/audin_u.cpp

@@ -2,9 +2,6 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include "common/logging/log.h"
-#include "core/hle/ipc_helpers.h"
-#include "core/hle/kernel/hle_ipc.h"
 #include "core/hle/service/audio/audin_u.h"
 
 namespace Service::Audio {
@@ -33,7 +30,6 @@ public:
 
         RegisterHandlers(functions);
     }
-    ~IAudioIn() = default;
 };
 
 AudInU::AudInU() : ServiceFramework("audin:u") {

src/core/hle/service/audio/audout_u.cpp

@@ -150,7 +150,6 @@ private:
     void GetReleasedAudioOutBufferImpl(Kernel::HLERequestContext& ctx) {
         LOG_DEBUG(Service_Audio, "called {}", ctx.Description());
 
-        IPC::RequestParser rp{ctx};
         const u64 max_count{ctx.GetWriteBufferSize() / sizeof(u64)};
         const auto released_buffers{audio_core.GetTagsAndReleaseBuffers(stream, max_count)};
 
@@ -194,12 +193,9 @@ private:
 void AudOutU::ListAudioOutsImpl(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_Audio, "called");
 
-    IPC::RequestParser rp{ctx};
-
     ctx.WriteBuffer(DefaultDevice);
 
     IPC::ResponseBuilder rb{ctx, 3};
-
     rb.Push(RESULT_SUCCESS);
     rb.Push<u32>(1); // Amount of audio devices
 }

src/core/hle/service/audio/audrec_u.cpp

@@ -2,9 +2,6 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include "common/logging/log.h"
-#include "core/hle/ipc_helpers.h"
-#include "core/hle/kernel/hle_ipc.h"
 #include "core/hle/service/audio/audrec_u.h"
 
 namespace Service::Audio {
@@ -30,7 +27,6 @@ public:
 
         RegisterHandlers(functions);
     }
-    ~IFinalOutputRecorder() = default;
 };
 
 AudRecU::AudRecU() : ServiceFramework("audrec:u") {

src/core/hle/service/audio/audren_u.cpp

@@ -10,6 +10,7 @@
 #include "common/alignment.h"
 #include "common/common_funcs.h"
 #include "common/logging/log.h"
+#include "common/string_util.h"
 #include "core/core.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/hle_ipc.h"
@@ -184,7 +185,6 @@ public:
 private:
     void ListAudioDeviceName(Kernel::HLERequestContext& ctx) {
         LOG_WARNING(Service_Audio, "(STUBBED) called");
-        IPC::RequestParser rp{ctx};
 
         constexpr std::array<char, 15> audio_interface{{"AudioInterface"}};
         ctx.WriteBuffer(audio_interface);
@@ -195,13 +195,13 @@ private:
     }
 
     void SetAudioDeviceOutputVolume(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_Audio, "(STUBBED) called");
-
         IPC::RequestParser rp{ctx};
-        f32 volume = static_cast<f32>(rp.Pop<u32>());
+        const f32 volume = rp.Pop<f32>();
 
-        auto file_buffer = ctx.ReadBuffer();
-        auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0');
+        const auto device_name_buffer = ctx.ReadBuffer();
+        const std::string name = Common::StringFromBuffer(device_name_buffer);
+
+        LOG_WARNING(Service_Audio, "(STUBBED) called. name={}, volume={}", name, volume);
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);
@@ -209,7 +209,6 @@ private:
 
     void GetActiveAudioDeviceName(Kernel::HLERequestContext& ctx) {
         LOG_WARNING(Service_Audio, "(STUBBED) called");
-        IPC::RequestParser rp{ctx};
 
         constexpr std::array<char, 12> audio_interface{{"AudioDevice"}};
         ctx.WriteBuffer(audio_interface);

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

@@ -197,13 +197,16 @@ ResultCode VfsDirectoryServiceWrapper::RenameDirectory(const std::string& src_pa
 
 ResultVal<FileSys::VirtualFile> VfsDirectoryServiceWrapper::OpenFile(const std::string& path_,
                                                                      FileSys::Mode mode) const {
-    std::string path(FileUtil::SanitizePath(path_));
-    auto npath = path;
-    while (npath.size() > 0 && (npath[0] == '/' || npath[0] == '\\'))
-        npath = npath.substr(1);
+    const std::string path(FileUtil::SanitizePath(path_));
+    std::string_view npath = path;
+    while (!npath.empty() && (npath[0] == '/' || npath[0] == '\\')) {
+        npath.remove_prefix(1);
+    }
+
     auto file = backing->GetFileRelative(npath);
-    if (file == nullptr)
+    if (file == nullptr) {
         return FileSys::ERROR_PATH_NOT_FOUND;
+    }
 
     if (mode == FileSys::Mode::Append) {
         return MakeResult<FileSys::VirtualFile>(
@@ -319,15 +322,15 @@ ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId stora
 }
 
 ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space,
-                                            FileSys::SaveDataDescriptor save_struct) {
+                                            const FileSys::SaveDataDescriptor& descriptor) {
     LOG_TRACE(Service_FS, "Opening Save Data for space_id={:01X}, save_struct={}",
-              static_cast<u8>(space), save_struct.DebugInfo());
+              static_cast<u8>(space), descriptor.DebugInfo());
 
     if (save_data_factory == nullptr) {
         return FileSys::ERROR_ENTITY_NOT_FOUND;
     }
 
-    return save_data_factory->Open(space, save_struct);
+    return save_data_factory->Open(space, descriptor);
 }
 
 ResultVal<FileSys::VirtualDir> OpenSaveDataSpace(FileSys::SaveDataSpaceId space) {

src/core/hle/service/filesystem/filesystem.h

@@ -46,7 +46,7 @@ ResultVal<FileSys::VirtualFile> OpenRomFSCurrentProcess();
 ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId storage_id,
                                           FileSys::ContentRecordType type);
 ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space,
-                                            FileSys::SaveDataDescriptor save_struct);
+                                            const FileSys::SaveDataDescriptor& descriptor);
 ResultVal<FileSys::VirtualDir> OpenSaveDataSpace(FileSys::SaveDataSpaceId space);
 ResultVal<FileSys::VirtualDir> OpenSDMC();
 

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

@@ -315,61 +315,53 @@ public:
     void CreateFile(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
 
-        auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        const auto file_buffer = ctx.ReadBuffer();
+        const std::string name = Common::StringFromBuffer(file_buffer);
 
-        u64 mode = rp.Pop<u64>();
-        u32 size = rp.Pop<u32>();
+        const u64 mode = rp.Pop<u64>();
+        const u32 size = rp.Pop<u32>();
 
-        LOG_DEBUG(Service_FS, "called file {} mode 0x{:X} size 0x{:08X}", name, mode, size);
+        LOG_DEBUG(Service_FS, "called. file={}, mode=0x{:X}, size=0x{:08X}", name, mode, size);
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(backend.CreateFile(name, size));
     }
 
     void DeleteFile(Kernel::HLERequestContext& ctx) {
-        IPC::RequestParser rp{ctx};
+        const auto file_buffer = ctx.ReadBuffer();
+        const std::string name = Common::StringFromBuffer(file_buffer);
 
-        auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
-
-        LOG_DEBUG(Service_FS, "called file {}", name);
+        LOG_DEBUG(Service_FS, "called. file={}", name);
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(backend.DeleteFile(name));
     }
 
     void CreateDirectory(Kernel::HLERequestContext& ctx) {
-        IPC::RequestParser rp{ctx};
+        const auto file_buffer = ctx.ReadBuffer();
+        const std::string name = Common::StringFromBuffer(file_buffer);
 
-        auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
-
-        LOG_DEBUG(Service_FS, "called directory {}", name);
+        LOG_DEBUG(Service_FS, "called. directory={}", name);
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(backend.CreateDirectory(name));
     }
 
     void DeleteDirectory(Kernel::HLERequestContext& ctx) {
-        const IPC::RequestParser rp{ctx};
-
         const auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        const std::string name = Common::StringFromBuffer(file_buffer);
 
-        LOG_DEBUG(Service_FS, "called directory {}", name);
+        LOG_DEBUG(Service_FS, "called. directory={}", name);
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(backend.DeleteDirectory(name));
     }
 
     void DeleteDirectoryRecursively(Kernel::HLERequestContext& ctx) {
-        const IPC::RequestParser rp{ctx};
-
         const auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        const std::string name = Common::StringFromBuffer(file_buffer);
 
-        LOG_DEBUG(Service_FS, "called directory {}", name);
+        LOG_DEBUG(Service_FS, "called. directory={}", name);
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(backend.DeleteDirectoryRecursively(name));
@@ -386,18 +378,16 @@ public:
     }
 
     void RenameFile(Kernel::HLERequestContext& ctx) {
-        IPC::RequestParser rp{ctx};
-
         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);
+        const std::string src_name = Common::StringFromBuffer(buffer);
 
         buffer.resize(ctx.BufferDescriptorX()[1].Size());
         Memory::ReadBlock(ctx.BufferDescriptorX()[1].Address(), buffer.data(), buffer.size());
-        std::string dst_name = Common::StringFromBuffer(buffer);
+        const std::string dst_name = Common::StringFromBuffer(buffer);
 
-        LOG_DEBUG(Service_FS, "called file '{}' to file '{}'", src_name, dst_name);
+        LOG_DEBUG(Service_FS, "called. file '{}' to file '{}'", src_name, dst_name);
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(backend.RenameFile(src_name, dst_name));
@@ -406,12 +396,12 @@ public:
     void OpenFile(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
 
-        auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        const auto file_buffer = ctx.ReadBuffer();
+        const std::string name = Common::StringFromBuffer(file_buffer);
 
-        auto mode = static_cast<FileSys::Mode>(rp.Pop<u32>());
+        const auto mode = static_cast<FileSys::Mode>(rp.Pop<u32>());
 
-        LOG_DEBUG(Service_FS, "called file {} mode {}", name, static_cast<u32>(mode));
+        LOG_DEBUG(Service_FS, "called. file={}, mode={}", name, static_cast<u32>(mode));
 
         auto result = backend.OpenFile(name, mode);
         if (result.Failed()) {
@@ -430,13 +420,13 @@ public:
     void OpenDirectory(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
 
-        auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
+        const auto file_buffer = ctx.ReadBuffer();
+        const std::string name = Common::StringFromBuffer(file_buffer);
 
         // TODO(Subv): Implement this filter.
-        u32 filter_flags = rp.Pop<u32>();
+        const u32 filter_flags = rp.Pop<u32>();
 
-        LOG_DEBUG(Service_FS, "called directory {} filter {}", name, filter_flags);
+        LOG_DEBUG(Service_FS, "called. directory={}, filter={}", name, filter_flags);
 
         auto result = backend.OpenDirectory(name);
         if (result.Failed()) {
@@ -453,12 +443,10 @@ public:
     }
 
     void GetEntryType(Kernel::HLERequestContext& ctx) {
-        IPC::RequestParser rp{ctx};
+        const auto file_buffer = ctx.ReadBuffer();
+        const std::string name = Common::StringFromBuffer(file_buffer);
 
-        auto file_buffer = ctx.ReadBuffer();
-        std::string name = Common::StringFromBuffer(file_buffer);
-
-        LOG_DEBUG(Service_FS, "called file {}", name);
+        LOG_DEBUG(Service_FS, "called. file={}", name);
 
         auto result = backend.GetEntryType(name);
         if (result.Failed()) {
@@ -616,7 +604,9 @@ private:
         u64_le save_id;
         u64_le title_id;
         u64_le save_image_size;
-        INSERT_PADDING_BYTES(0x28);
+        u16_le index;
+        FileSys::SaveDataRank rank;
+        INSERT_PADDING_BYTES(0x25);
     };
     static_assert(sizeof(SaveDataInfo) == 0x60, "SaveDataInfo has incorrect size.");
 
@@ -779,16 +769,17 @@ void FSP_SRV::CreateSaveDataFileSystem(Kernel::HLERequestContext& ctx) {
 }
 
 void FSP_SRV::OpenSaveDataFileSystem(Kernel::HLERequestContext& ctx) {
+    LOG_INFO(Service_FS, "called.");
+
+    struct Parameters {
+        FileSys::SaveDataSpaceId save_data_space_id;
+        FileSys::SaveDataDescriptor descriptor;
+    };
+
     IPC::RequestParser rp{ctx};
+    const auto parameters = rp.PopRaw<Parameters>();
 
-    auto space_id = rp.PopRaw<FileSys::SaveDataSpaceId>();
-    auto unk = rp.Pop<u32>();
-    LOG_INFO(Service_FS, "called with unknown={:08X}", unk);
-
-    auto save_struct = rp.PopRaw<FileSys::SaveDataDescriptor>();
-
-    auto dir = OpenSaveData(space_id, save_struct);
-
+    auto dir = OpenSaveData(parameters.save_data_space_id, parameters.descriptor);
     if (dir.Failed()) {
         IPC::ResponseBuilder rb{ctx, 2, 0, 0};
         rb.Push(FileSys::ERROR_ENTITY_NOT_FOUND);

src/core/hle/service/nvdrv/devices/nvdisp_disp0.h

@@ -18,7 +18,7 @@ class nvmap;
 class nvdisp_disp0 final : public nvdevice {
 public:
     explicit nvdisp_disp0(std::shared_ptr<nvmap> nvmap_dev);
-    ~nvdisp_disp0();
+    ~nvdisp_disp0() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override;
 

src/core/hle/service/nvdrv/interface.h

@@ -17,7 +17,7 @@ namespace Service::Nvidia {
 class NVDRV final : public ServiceFramework<NVDRV> {
 public:
     NVDRV(std::shared_ptr<Module> nvdrv, const char* name);
-    ~NVDRV();
+    ~NVDRV() override;
 
 private:
     void Open(Kernel::HLERequestContext& ctx);

src/core/hle/service/nvdrv/nvmemp.h

@@ -11,7 +11,7 @@ namespace Service::Nvidia {
 class NVMEMP final : public ServiceFramework<NVMEMP> {
 public:
     NVMEMP();
-    ~NVMEMP();
+    ~NVMEMP() override;
 
 private:
     void Cmd0(Kernel::HLERequestContext& ctx);

src/core/hle/service/service.h

@@ -90,7 +90,7 @@ private:
                            Kernel::HLERequestContext& ctx);
 
     ServiceFrameworkBase(const char* service_name, u32 max_sessions, InvokerFn* handler_invoker);
-    ~ServiceFrameworkBase();
+    ~ServiceFrameworkBase() override;
 
     void RegisterHandlersBase(const FunctionInfoBase* functions, std::size_t n);
     void ReportUnimplementedFunction(Kernel::HLERequestContext& ctx, const FunctionInfoBase* info);

src/core/hle/service/set/set_cal.h

@@ -11,7 +11,7 @@ namespace Service::Set {
 class SET_CAL final : public ServiceFramework<SET_CAL> {
 public:
     explicit SET_CAL();
-    ~SET_CAL();
+    ~SET_CAL() override;
 };
 
 } // namespace Service::Set

src/core/hle/service/sockets/sfdnsres.cpp

@@ -8,12 +8,20 @@
 namespace Service::Sockets {
 
 void SFDNSRES::GetAddrInfo(Kernel::HLERequestContext& ctx) {
-    IPC::RequestParser rp{ctx};
+    struct Parameters {
+        u8 use_nsd_resolve;
+        u32 unknown;
+        u64 process_id;
+    };
 
-    LOG_WARNING(Service, "(STUBBED) called");
+    IPC::RequestParser rp{ctx};
+    const auto parameters = rp.PopRaw<Parameters>();
+
+    LOG_WARNING(Service,
+                "(STUBBED) called. use_nsd_resolve={}, unknown=0x{:08X}, process_id=0x{:016X}",
+                parameters.use_nsd_resolve, parameters.unknown, parameters.process_id);
 
     IPC::ResponseBuilder rb{ctx, 2};
-
     rb.Push(RESULT_SUCCESS);
 }
 

src/core/hle/service/spl/module.cpp

@@ -26,9 +26,7 @@ Module::Interface::~Interface() = default;
 void Module::Interface::GetRandomBytes(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_SPL, "called");
 
-    IPC::RequestParser rp{ctx};
-
-    std::size_t size = ctx.GetWriteBufferSize();
+    const std::size_t size = ctx.GetWriteBufferSize();
 
     std::uniform_int_distribution<u16> distribution(0, std::numeric_limits<u8>::max());
     std::vector<u8> data(size);

src/core/hle/service/ssl/ssl.cpp

@@ -64,13 +64,19 @@ public:
         };
         RegisterHandlers(functions);
     }
-    ~ISslContext() = default;
 
 private:
     void SetOption(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_SSL, "(STUBBED) called");
+        struct Parameters {
+            u8 enable;
+            u32 option;
+        };
 
         IPC::RequestParser rp{ctx};
+        const auto parameters = rp.PopRaw<Parameters>();
+
+        LOG_WARNING(Service_SSL, "(STUBBED) called. enable={}, option={}", parameters.enable,
+                    parameters.option);
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(RESULT_SUCCESS);

src/core/hle/service/vi/vi.cpp

@@ -498,7 +498,6 @@ public:
         };
         RegisterHandlers(functions);
     }
-    ~IHOSBinderDriver() = default;
 
 private:
     enum class TransactionId {
@@ -692,7 +691,6 @@ public:
         };
         RegisterHandlers(functions);
     }
-    ~ISystemDisplayService() = default;
 
 private:
     void SetLayerZ(Kernel::HLERequestContext& ctx) {
@@ -818,7 +816,6 @@ public:
         };
         RegisterHandlers(functions);
     }
-    ~IManagerDisplayService() = default;
 
 private:
     void CloseDisplay(Kernel::HLERequestContext& ctx) {
@@ -884,7 +881,6 @@ private:
 class IApplicationDisplayService final : public ServiceFramework<IApplicationDisplayService> {
 public:
     explicit IApplicationDisplayService(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger);
-    ~IApplicationDisplayService() = default;
 
 private:
     enum class ConvertedScaleMode : u64 {
@@ -1037,7 +1033,6 @@ private:
     void ListDisplays(Kernel::HLERequestContext& ctx) {
         LOG_WARNING(Service_VI, "(STUBBED) called");
 
-        IPC::RequestParser rp{ctx};
         DisplayInfo display_info;
         display_info.width *= static_cast<u64>(Settings::values.resolution_factor);
         display_info.height *= static_cast<u64>(Settings::values.resolution_factor);

src/core/loader/nso.cpp

@@ -4,11 +4,12 @@
 
 #include <cinttypes>
 #include <vector>
-#include <lz4.h>
+
 #include "common/common_funcs.h"
 #include "common/file_util.h"
 #include "common/hex_util.h"
 #include "common/logging/log.h"
+#include "common/lz4_compression.h"
 #include "common/swap.h"
 #include "core/core.h"
 #include "core/file_sys/patch_manager.h"
@@ -35,15 +36,11 @@ static_assert(sizeof(MODHeader) == 0x1c, "MODHeader has incorrect size.");
 
 std::vector<u8> DecompressSegment(const std::vector<u8>& compressed_data,
                                   const NSOSegmentHeader& header) {
-    std::vector<u8> uncompressed_data(header.size);
-    const int bytes_uncompressed =
-        LZ4_decompress_safe(reinterpret_cast<const char*>(compressed_data.data()),
-                            reinterpret_cast<char*>(uncompressed_data.data()),
-                            static_cast<int>(compressed_data.size()), header.size);
+    const std::vector<u8> uncompressed_data =
+        Common::Compression::DecompressDataLZ4(compressed_data, header.size);
 
-    ASSERT_MSG(bytes_uncompressed == static_cast<int>(header.size) &&
-                   bytes_uncompressed == static_cast<int>(uncompressed_data.size()),
-               "{} != {} != {}", bytes_uncompressed, header.size, uncompressed_data.size());
+    ASSERT_MSG(uncompressed_data.size() == static_cast<int>(header.size), "{} != {}", header.size,
+               uncompressed_data.size());
 
     return uncompressed_data;
 }

src/core/loader/xci.h

@@ -22,7 +22,7 @@ class AppLoader_NCA;
 class AppLoader_XCI final : public AppLoader {
 public:
     explicit AppLoader_XCI(FileSys::VirtualFile file);
-    ~AppLoader_XCI();
+    ~AppLoader_XCI() override;
 
     /**
      * Returns the type of the file

src/core/memory.cpp

@@ -38,10 +38,6 @@ void SetCurrentPageTable(Common::PageTable* page_table) {
     }
 }
 
-Common::PageTable* GetCurrentPageTable() {
-    return current_page_table;
-}
-
 static void MapPages(Common::PageTable& page_table, VAddr base, u64 size, u8* memory,
                      Common::PageType type) {
     LOG_DEBUG(HW_Memory, "Mapping {} onto {:016X}-{:016X}", fmt::ptr(memory), base * PAGE_SIZE,

src/core/memory.h

@@ -28,16 +28,6 @@ constexpr u64 PAGE_MASK = PAGE_SIZE - 1;
 
 /// Virtual user-space memory regions
 enum : VAddr {
-    /// Read-only page containing kernel and system configuration values.
-    CONFIG_MEMORY_VADDR = 0x1FF80000,
-    CONFIG_MEMORY_SIZE = 0x00001000,
-    CONFIG_MEMORY_VADDR_END = CONFIG_MEMORY_VADDR + CONFIG_MEMORY_SIZE,
-
-    /// Usually read-only page containing mostly values read from hardware.
-    SHARED_PAGE_VADDR = 0x1FF81000,
-    SHARED_PAGE_SIZE = 0x00001000,
-    SHARED_PAGE_VADDR_END = SHARED_PAGE_VADDR + SHARED_PAGE_SIZE,
-
     /// TLS (Thread-Local Storage) related.
     TLS_ENTRY_SIZE = 0x200,
 
@@ -50,9 +40,8 @@ enum : VAddr {
     KERNEL_REGION_END = KERNEL_REGION_VADDR + KERNEL_REGION_SIZE,
 };
 
-/// Currently active page table
+/// Changes the currently active page table.
 void SetCurrentPageTable(Common::PageTable* page_table);
-Common::PageTable* GetCurrentPageTable();
 
 /// Determines if the given VAddr is valid for the specified process.
 bool IsValidVirtualAddress(const Kernel::Process& process, VAddr vaddr);

src/tests/core/arm/arm_test_common.cpp

@@ -17,7 +17,6 @@ TestEnvironment::TestEnvironment(bool mutable_memory_)
     : mutable_memory(mutable_memory_),
       test_memory(std::make_shared<TestMemory>(this)), kernel{Core::System::GetInstance()} {
     auto process = Kernel::Process::Create(Core::System::GetInstance(), "");
-    kernel.MakeCurrentProcess(process.get());
     page_table = &process->VMManager().page_table;
 
     std::fill(page_table->pointers.begin(), page_table->pointers.end(), nullptr);
@@ -28,7 +27,7 @@ TestEnvironment::TestEnvironment(bool mutable_memory_)
     Memory::MapIoRegion(*page_table, 0x00000000, 0x80000000, test_memory);
     Memory::MapIoRegion(*page_table, 0x80000000, 0x80000000, test_memory);
 
-    Memory::SetCurrentPageTable(page_table);
+    kernel.MakeCurrentProcess(process.get());
 }
 
 TestEnvironment::~TestEnvironment() {

src/video_core/CMakeLists.txt

@@ -128,7 +128,9 @@ if (ENABLE_VULKAN)
         renderer_vulkan/vk_scheduler.cpp
         renderer_vulkan/vk_scheduler.h
         renderer_vulkan/vk_stream_buffer.cpp
-        renderer_vulkan/vk_stream_buffer.h)
+        renderer_vulkan/vk_stream_buffer.h
+        renderer_vulkan/vk_swapchain.cpp
+        renderer_vulkan/vk_swapchain.h)
 
     target_include_directories(video_core PRIVATE ../../externals/Vulkan-Headers/include)
     target_compile_definitions(video_core PRIVATE HAS_VULKAN)
@@ -137,4 +139,4 @@ endif()
 create_target_directory_groups(video_core)
 
 target_link_libraries(video_core PUBLIC common core)
-target_link_libraries(video_core PRIVATE glad lz4_static)
+target_link_libraries(video_core PRIVATE glad)

src/video_core/engines/fermi_2d.cpp

@@ -6,12 +6,13 @@
 #include "common/logging/log.h"
 #include "common/math_util.h"
 #include "video_core/engines/fermi_2d.h"
+#include "video_core/memory_manager.h"
 #include "video_core/rasterizer_interface.h"
 
 namespace Tegra::Engines {
 
 Fermi2D::Fermi2D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager)
-    : memory_manager(memory_manager), rasterizer{rasterizer} {}
+    : rasterizer{rasterizer}, memory_manager{memory_manager} {}
 
 void Fermi2D::CallMethod(const GPU::MethodCall& method_call) {
     ASSERT_MSG(method_call.method < Regs::NUM_REGS,

src/video_core/engines/fermi_2d.h

@@ -10,7 +10,10 @@
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "video_core/gpu.h"
-#include "video_core/memory_manager.h"
+
+namespace Tegra {
+class MemoryManager;
+}
 
 namespace VideoCore {
 class RasterizerInterface;
@@ -115,10 +118,9 @@ public:
         };
     } regs{};
 
-    MemoryManager& memory_manager;
-
 private:
     VideoCore::RasterizerInterface& rasterizer;
+    MemoryManager& memory_manager;
 
     /// Performs the copy from the source surface to the destination surface as configured in the
     /// registers.

src/video_core/engines/kepler_compute.h

@@ -9,7 +9,10 @@
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "video_core/gpu.h"
-#include "video_core/memory_manager.h"
+
+namespace Tegra {
+class MemoryManager;
+}
 
 namespace Tegra::Engines {
 
@@ -40,10 +43,11 @@ public:
     static_assert(sizeof(Regs) == Regs::NUM_REGS * sizeof(u32),
                   "KeplerCompute Regs has wrong size");
 
-    MemoryManager& memory_manager;
-
     /// Write the value to the register identified by method.
     void CallMethod(const GPU::MethodCall& method_call);
+
+private:
+    MemoryManager& memory_manager;
 };
 
 #define ASSERT_REG_POSITION(field_name, position)                                                  \

src/video_core/engines/kepler_memory.cpp

@@ -5,9 +5,9 @@
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "core/core.h"
-#include "core/memory.h"
 #include "video_core/engines/kepler_memory.h"
 #include "video_core/engines/maxwell_3d.h"
+#include "video_core/memory_manager.h"
 #include "video_core/rasterizer_interface.h"
 #include "video_core/renderer_base.h"
 
@@ -15,7 +15,7 @@ namespace Tegra::Engines {
 
 KeplerMemory::KeplerMemory(Core::System& system, VideoCore::RasterizerInterface& rasterizer,
                            MemoryManager& memory_manager)
-    : system{system}, memory_manager(memory_manager), rasterizer{rasterizer} {}
+    : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager} {}
 
 KeplerMemory::~KeplerMemory() = default;
 

src/video_core/engines/kepler_memory.h

@@ -10,12 +10,15 @@
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "video_core/gpu.h"
-#include "video_core/memory_manager.h"
 
 namespace Core {
 class System;
 }
 
+namespace Tegra {
+class MemoryManager;
+}
+
 namespace VideoCore {
 class RasterizerInterface;
 }
@@ -82,8 +85,8 @@ public:
 
 private:
     Core::System& system;
-    MemoryManager& memory_manager;
     VideoCore::RasterizerInterface& rasterizer;
+    MemoryManager& memory_manager;
 
     void ProcessData(u32 data);
 };

src/video_core/engines/maxwell_3d.cpp

@@ -7,11 +7,10 @@
 #include "common/assert.h"
 #include "core/core.h"
 #include "core/core_timing.h"
-#include "core/memory.h"
 #include "video_core/debug_utils/debug_utils.h"
 #include "video_core/engines/maxwell_3d.h"
+#include "video_core/memory_manager.h"
 #include "video_core/rasterizer_interface.h"
-#include "video_core/renderer_base.h"
 #include "video_core/textures/texture.h"
 
 namespace Tegra::Engines {
@@ -21,8 +20,8 @@ constexpr u32 MacroRegistersStart = 0xE00;
 
 Maxwell3D::Maxwell3D(Core::System& system, VideoCore::RasterizerInterface& rasterizer,
                      MemoryManager& memory_manager)
-    : memory_manager(memory_manager), system{system}, rasterizer{rasterizer},
-      macro_interpreter(*this) {
+    : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager}, macro_interpreter{
+                                                                                  *this} {
     InitializeRegisterDefaults();
 }
 
@@ -250,6 +249,10 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) {
         ProcessQueryGet();
         break;
     }
+    case MAXWELL3D_REG_INDEX(sync_info): {
+        ProcessSyncPoint();
+        break;
+    }
     default:
         break;
     }
@@ -327,6 +330,14 @@ void Maxwell3D::ProcessQueryGet() {
     }
 }
 
+void Maxwell3D::ProcessSyncPoint() {
+    const u32 sync_point = regs.sync_info.sync_point.Value();
+    const u32 increment = regs.sync_info.increment.Value();
+    const u32 cache_flush = regs.sync_info.unknown.Value();
+    LOG_DEBUG(HW_GPU, "Syncpoint set {}, increment: {}, unk: {}", sync_point, increment,
+              cache_flush);
+}
+
 void Maxwell3D::DrawArrays() {
     LOG_DEBUG(HW_GPU, "called, topology={}, count={}", static_cast<u32>(regs.draw.topology.Value()),
               regs.vertex_buffer.count);

src/video_core/engines/maxwell_3d.h

@@ -16,13 +16,16 @@
 #include "common/math_util.h"
 #include "video_core/gpu.h"
 #include "video_core/macro_interpreter.h"
-#include "video_core/memory_manager.h"
 #include "video_core/textures/texture.h"
 
 namespace Core {
 class System;
 }
 
+namespace Tegra {
+class MemoryManager;
+}
+
 namespace VideoCore {
 class RasterizerInterface;
 }
@@ -576,7 +579,17 @@ public:
                     u32 bind;
                 } macros;
 
-                INSERT_PADDING_WORDS(0x188);
+                INSERT_PADDING_WORDS(0x69);
+
+                struct {
+                    union {
+                        BitField<0, 16, u32> sync_point;
+                        BitField<16, 1, u32> unknown;
+                        BitField<20, 1, u32> increment;
+                    };
+                } sync_info;
+
+                INSERT_PADDING_WORDS(0x11E);
 
                 u32 tfb_enabled;
 
@@ -1093,7 +1106,6 @@ public:
     };
 
     State state{};
-    MemoryManager& memory_manager;
 
     struct DirtyFlags {
         std::bitset<8> color_buffer{0xFF};
@@ -1141,6 +1153,8 @@ private:
 
     VideoCore::RasterizerInterface& rasterizer;
 
+    MemoryManager& memory_manager;
+
     /// Start offsets of each macro in macro_memory
     std::unordered_map<u32, u32> macro_offsets;
 
@@ -1180,6 +1194,9 @@ private:
     /// Handles a write to the QUERY_GET register.
     void ProcessQueryGet();
 
+    /// Handles writes to syncing register.
+    void ProcessSyncPoint();
+
     /// Handles a write to the CB_DATA[i] register.
     void ProcessCBData(u32 value);
 
@@ -1195,6 +1212,7 @@ private:
                   "Field " #field_name " has invalid position")
 
 ASSERT_REG_POSITION(macros, 0x45);
+ASSERT_REG_POSITION(sync_info, 0xB2);
 ASSERT_REG_POSITION(tfb_enabled, 0x1D1);
 ASSERT_REG_POSITION(rt, 0x200);
 ASSERT_REG_POSITION(viewport_transform, 0x280);

src/video_core/engines/maxwell_dma.cpp

@@ -5,9 +5,9 @@
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "core/core.h"
-#include "core/memory.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/engines/maxwell_dma.h"
+#include "video_core/memory_manager.h"
 #include "video_core/rasterizer_interface.h"
 #include "video_core/renderer_base.h"
 #include "video_core/textures/decoders.h"
@@ -16,7 +16,7 @@ namespace Tegra::Engines {
 
 MaxwellDMA::MaxwellDMA(Core::System& system, VideoCore::RasterizerInterface& rasterizer,
                        MemoryManager& memory_manager)
-    : memory_manager(memory_manager), system{system}, rasterizer{rasterizer} {}
+    : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager} {}
 
 void MaxwellDMA::CallMethod(const GPU::MethodCall& method_call) {
     ASSERT_MSG(method_call.method < Regs::NUM_REGS,

src/video_core/engines/maxwell_dma.h

@@ -10,12 +10,15 @@
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "video_core/gpu.h"
-#include "video_core/memory_manager.h"
 
 namespace Core {
 class System;
 }
 
+namespace Tegra {
+class MemoryManager;
+}
+
 namespace VideoCore {
 class RasterizerInterface;
 }
@@ -139,13 +142,13 @@ public:
         };
     } regs{};
 
-    MemoryManager& memory_manager;
-
 private:
     Core::System& system;
 
     VideoCore::RasterizerInterface& rasterizer;
 
+    MemoryManager& memory_manager;
+
     /// Performs the copy from the source buffer to the destination buffer as configured in the
     /// registers.
     void HandleCopy();

src/video_core/gpu.cpp

@@ -31,7 +31,7 @@ u32 FramebufferConfig::BytesPerPixel(PixelFormat format) {
 
 GPU::GPU(Core::System& system, VideoCore::RendererBase& renderer) : renderer{renderer} {
     auto& rasterizer{renderer.Rasterizer()};
-    memory_manager = std::make_unique<Tegra::MemoryManager>();
+    memory_manager = std::make_unique<Tegra::MemoryManager>(rasterizer);
     dma_pusher = std::make_unique<Tegra::DmaPusher>(*this);
     maxwell_3d = std::make_unique<Engines::Maxwell3D>(system, rasterizer, *memory_manager);
     fermi_2d = std::make_unique<Engines::Fermi2D>(rasterizer, *memory_manager);

src/video_core/gpu_asynch.cpp

@@ -9,7 +9,7 @@
 namespace VideoCommon {
 
 GPUAsynch::GPUAsynch(Core::System& system, VideoCore::RendererBase& renderer)
-    : Tegra::GPU(system, renderer), gpu_thread{renderer, *dma_pusher} {}
+    : Tegra::GPU(system, renderer), gpu_thread{system, renderer, *dma_pusher} {}
 
 GPUAsynch::~GPUAsynch() = default;
 

src/video_core/gpu_thread.cpp

@@ -4,6 +4,9 @@
 
 #include "common/assert.h"
 #include "common/microprofile.h"
+#include "core/core.h"
+#include "core/core_timing.h"
+#include "core/core_timing_util.h"
 #include "core/frontend/scope_acquire_window_context.h"
 #include "video_core/dma_pusher.h"
 #include "video_core/gpu.h"
@@ -36,7 +39,6 @@ static void RunThread(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_p
                 dma_pusher.Push(std::move(submit_list->entries));
                 dma_pusher.DispatchCalls();
             } else if (const auto data = std::get_if<SwapBuffersCommand>(&next.data)) {
-                state.DecrementFramesCounter();
                 renderer.SwapBuffers(std::move(data->framebuffer));
             } else if (const auto data = std::get_if<FlushRegionCommand>(&next.data)) {
                 renderer.Rasterizer().FlushRegion(data->addr, data->size);
@@ -47,13 +49,18 @@ static void RunThread(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_p
             } else {
                 UNREACHABLE();
             }
+            state.signaled_fence = next.fence;
+            state.TrySynchronize();
         }
     }
 }
 
-ThreadManager::ThreadManager(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_pusher)
-    : renderer{renderer}, thread{RunThread, std::ref(renderer), std::ref(dma_pusher),
-                                 std::ref(state)} {}
+ThreadManager::ThreadManager(Core::System& system, VideoCore::RendererBase& renderer,
+                             Tegra::DmaPusher& dma_pusher)
+    : system{system}, thread{RunThread, std::ref(renderer), std::ref(dma_pusher), std::ref(state)} {
+    synchronization_event = system.CoreTiming().RegisterEvent(
+        "GPUThreadSynch", [this](u64 fence, s64) { state.WaitForSynchronization(fence); });
+}
 
 ThreadManager::~ThreadManager() {
     // Notify GPU thread that a shutdown is pending
@@ -62,14 +69,14 @@ ThreadManager::~ThreadManager() {
 }
 
 void ThreadManager::SubmitList(Tegra::CommandList&& entries) {
-    PushCommand(SubmitListCommand(std::move(entries)));
+    const u64 fence{PushCommand(SubmitListCommand(std::move(entries)))};
+    const s64 synchronization_ticks{Core::Timing::usToCycles(9000)};
+    system.CoreTiming().ScheduleEvent(synchronization_ticks, synchronization_event, fence);
 }
 
 void ThreadManager::SwapBuffers(
     std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
-    state.IncrementFramesCounter();
     PushCommand(SwapBuffersCommand(std::move(framebuffer)));
-    state.WaitForFrames();
 }
 
 void ThreadManager::FlushRegion(CacheAddr addr, u64 size) {
@@ -79,7 +86,7 @@ void ThreadManager::FlushRegion(CacheAddr addr, u64 size) {
 void ThreadManager::InvalidateRegion(CacheAddr addr, u64 size) {
     if (state.queue.Empty()) {
         // It's quicker to invalidate a single region on the CPU if the queue is already empty
-        renderer.Rasterizer().InvalidateRegion(addr, size);
+        system.Renderer().Rasterizer().InvalidateRegion(addr, size);
     } else {
         PushCommand(InvalidateRegionCommand(addr, size));
     }
@@ -90,9 +97,25 @@ void ThreadManager::FlushAndInvalidateRegion(CacheAddr addr, u64 size) {
     InvalidateRegion(addr, size);
 }
 
-void ThreadManager::PushCommand(CommandData&& command_data) {
-    state.queue.Push(CommandDataContainer(std::move(command_data)));
+u64 ThreadManager::PushCommand(CommandData&& command_data) {
+    const u64 fence{++state.last_fence};
+    state.queue.Push(CommandDataContainer(std::move(command_data), fence));
     state.SignalCommands();
+    return fence;
+}
+
+MICROPROFILE_DEFINE(GPU_wait, "GPU", "Wait for the GPU", MP_RGB(128, 128, 192));
+void SynchState::WaitForSynchronization(u64 fence) {
+    if (signaled_fence >= fence) {
+        return;
+    }
+
+    // Wait for the GPU to be idle (all commands to be executed)
+    {
+        MICROPROFILE_SCOPE(GPU_wait);
+        std::unique_lock<std::mutex> lock{synchronization_mutex};
+        synchronization_condition.wait(lock, [this, fence] { return signaled_fence >= fence; });
+    }
 }
 
 } // namespace VideoCommon::GPUThread

src/video_core/gpu_thread.h

@@ -19,9 +19,12 @@ struct FramebufferConfig;
 class DmaPusher;
 } // namespace Tegra
 
-namespace VideoCore {
-class RendererBase;
-} // namespace VideoCore
+namespace Core {
+class System;
+namespace Timing {
+struct EventType;
+} // namespace Timing
+} // namespace Core
 
 namespace VideoCommon::GPUThread {
 
@@ -75,63 +78,47 @@ using CommandData =
 struct CommandDataContainer {
     CommandDataContainer() = default;
 
-    CommandDataContainer(CommandData&& data) : data{std::move(data)} {}
+    CommandDataContainer(CommandData&& data, u64 next_fence)
+        : data{std::move(data)}, fence{next_fence} {}
 
     CommandDataContainer& operator=(const CommandDataContainer& t) {
         data = std::move(t.data);
+        fence = t.fence;
         return *this;
     }
 
     CommandData data;
+    u64 fence{};
 };
 
 /// Struct used to synchronize the GPU thread
 struct SynchState final {
     std::atomic_bool is_running{true};
     std::atomic_int queued_frame_count{};
-    std::mutex frames_mutex;
+    std::mutex synchronization_mutex;
     std::mutex commands_mutex;
     std::condition_variable commands_condition;
-    std::condition_variable frames_condition;
+    std::condition_variable synchronization_condition;
 
-    void IncrementFramesCounter() {
-        std::lock_guard lock{frames_mutex};
-        ++queued_frame_count;
+    /// Returns true if the gap in GPU commands is small enough that we can consider the CPU and GPU
+    /// synchronized. This is entirely empirical.
+    bool IsSynchronized() const {
+        constexpr std::size_t max_queue_gap{5};
+        return queue.Size() <= max_queue_gap;
     }
 
-    void DecrementFramesCounter() {
-        {
-            std::lock_guard lock{frames_mutex};
-            --queued_frame_count;
-
-            if (queued_frame_count) {
-                return;
-            }
-        }
-        frames_condition.notify_one();
-    }
-
-    void WaitForFrames() {
-        {
-            std::lock_guard lock{frames_mutex};
-            if (!queued_frame_count) {
-                return;
-            }
-        }
-
-        // Wait for the GPU to be idle (all commands to be executed)
-        {
-            std::unique_lock lock{frames_mutex};
-            frames_condition.wait(lock, [this] { return !queued_frame_count; });
+    void TrySynchronize() {
+        if (IsSynchronized()) {
+            std::lock_guard<std::mutex> lock{synchronization_mutex};
+            synchronization_condition.notify_one();
         }
     }
 
+    void WaitForSynchronization(u64 fence);
+
     void SignalCommands() {
-        {
-            std::unique_lock lock{commands_mutex};
-            if (queue.Empty()) {
-                return;
-            }
+        if (queue.Empty()) {
+            return;
         }
 
         commands_condition.notify_one();
@@ -144,12 +131,15 @@ struct SynchState final {
 
     using CommandQueue = Common::SPSCQueue<CommandDataContainer>;
     CommandQueue queue;
+    u64 last_fence{};
+    std::atomic<u64> signaled_fence{};
 };
 
 /// Class used to manage the GPU thread
 class ThreadManager final {
 public:
-    explicit ThreadManager(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_pusher);
+    explicit ThreadManager(Core::System& system, VideoCore::RendererBase& renderer,
+                           Tegra::DmaPusher& dma_pusher);
     ~ThreadManager();
 
     /// Push GPU command entries to be processed
@@ -170,11 +160,12 @@ public:
 
 private:
     /// Pushes a command to be executed by the GPU thread
-    void PushCommand(CommandData&& command_data);
+    u64 PushCommand(CommandData&& command_data);
 
 private:
     SynchState state;
-    VideoCore::RendererBase& renderer;
+    Core::System& system;
+    Core::Timing::EventType* synchronization_event{};
     std::thread thread;
     std::thread::id thread_id;
 };

src/video_core/macro_interpreter.cpp

@@ -223,27 +223,21 @@ void MacroInterpreter::ProcessResult(ResultOperation operation, u32 reg, u32 res
 }
 
 u32 MacroInterpreter::FetchParameter() {
-    ASSERT(next_parameter_index < parameters.size());
-    return parameters[next_parameter_index++];
+    return parameters.at(next_parameter_index++);
 }
 
 u32 MacroInterpreter::GetRegister(u32 register_id) const {
-    // Register 0 is supposed to always return 0.
-    if (register_id == 0)
-        return 0;
-
-    ASSERT(register_id < registers.size());
-    return registers[register_id];
+    return registers.at(register_id);
 }
 
 void MacroInterpreter::SetRegister(u32 register_id, u32 value) {
-    // Register 0 is supposed to always return 0. NOP is implemented as a store to the zero
-    // register.
-    if (register_id == 0)
+    // Register 0 is hardwired as the zero register.
+    // Ensure no writes to it actually occur.
+    if (register_id == 0) {
         return;
+    }
 
-    ASSERT(register_id < registers.size());
-    registers[register_id] = value;
+    registers.at(register_id) = value;
 }
 
 void MacroInterpreter::SetMethodAddress(u32 address) {

src/video_core/memory_manager.cpp

@@ -5,16 +5,13 @@
 #include "common/alignment.h"
 #include "common/assert.h"
 #include "common/logging/log.h"
-#include "core/core.h"
 #include "core/memory.h"
-#include "video_core/gpu.h"
 #include "video_core/memory_manager.h"
 #include "video_core/rasterizer_interface.h"
-#include "video_core/renderer_base.h"
 
 namespace Tegra {
 
-MemoryManager::MemoryManager() {
+MemoryManager::MemoryManager(VideoCore::RasterizerInterface& rasterizer) : rasterizer{rasterizer} {
     std::fill(page_table.pointers.begin(), page_table.pointers.end(), nullptr);
     std::fill(page_table.attributes.begin(), page_table.attributes.end(),
               Common::PageType::Unmapped);
@@ -70,23 +67,23 @@ GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) {
     const u64 aligned_size{Common::AlignUp(size, page_size)};
     const CacheAddr cache_addr{ToCacheAddr(GetPointer(gpu_addr))};
 
-    Core::System::GetInstance().Renderer().Rasterizer().FlushAndInvalidateRegion(cache_addr,
-                                                                                 aligned_size);
+    rasterizer.FlushAndInvalidateRegion(cache_addr, aligned_size);
     UnmapRange(gpu_addr, aligned_size);
 
     return gpu_addr;
 }
 
-GPUVAddr MemoryManager::FindFreeRegion(GPUVAddr region_start, u64 size) {
+GPUVAddr MemoryManager::FindFreeRegion(GPUVAddr region_start, u64 size) const {
     // Find the first Free VMA.
-    const VMAHandle vma_handle{std::find_if(vma_map.begin(), vma_map.end(), [&](const auto& vma) {
-        if (vma.second.type != VirtualMemoryArea::Type::Unmapped) {
-            return false;
-        }
+    const VMAHandle vma_handle{
+        std::find_if(vma_map.begin(), vma_map.end(), [region_start, size](const auto& vma) {
+            if (vma.second.type != VirtualMemoryArea::Type::Unmapped) {
+                return false;
+            }
 
-        const VAddr vma_end{vma.second.base + vma.second.size};
-        return vma_end > region_start && vma_end >= region_start + size;
-    })};
+            const VAddr vma_end{vma.second.base + vma.second.size};
+            return vma_end > region_start && vma_end >= region_start + size;
+        })};
 
     if (vma_handle == vma_map.end()) {
         return {};
@@ -99,12 +96,12 @@ bool MemoryManager::IsAddressValid(GPUVAddr addr) const {
     return (addr >> page_bits) < page_table.pointers.size();
 }
 
-std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr addr) {
+std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr addr) const {
     if (!IsAddressValid(addr)) {
         return {};
     }
 
-    VAddr cpu_addr{page_table.backing_addr[addr >> page_bits]};
+    const VAddr cpu_addr{page_table.backing_addr[addr >> page_bits]};
     if (cpu_addr) {
         return cpu_addr + (addr & page_mask);
     }
@@ -113,7 +110,7 @@ std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr addr) {
 }
 
 template <typename T>
-T MemoryManager::Read(GPUVAddr addr) {
+T MemoryManager::Read(GPUVAddr addr) const {
     if (!IsAddressValid(addr)) {
         return {};
     }
@@ -165,10 +162,10 @@ void MemoryManager::Write(GPUVAddr addr, T data) {
     }
 }
 
-template u8 MemoryManager::Read<u8>(GPUVAddr addr);
-template u16 MemoryManager::Read<u16>(GPUVAddr addr);
-template u32 MemoryManager::Read<u32>(GPUVAddr addr);
-template u64 MemoryManager::Read<u64>(GPUVAddr addr);
+template u8 MemoryManager::Read<u8>(GPUVAddr addr) const;
+template u16 MemoryManager::Read<u16>(GPUVAddr addr) const;
+template u32 MemoryManager::Read<u32>(GPUVAddr addr) const;
+template u64 MemoryManager::Read<u64>(GPUVAddr addr) const;
 template void MemoryManager::Write<u8>(GPUVAddr addr, u8 data);
 template void MemoryManager::Write<u16>(GPUVAddr addr, u16 data);
 template void MemoryManager::Write<u32>(GPUVAddr addr, u32 data);
@@ -179,8 +176,8 @@ u8* MemoryManager::GetPointer(GPUVAddr addr) {
         return {};
     }
 
-    u8* page_pointer{page_table.pointers[addr >> page_bits]};
-    if (page_pointer) {
+    u8* const page_pointer{page_table.pointers[addr >> page_bits]};
+    if (page_pointer != nullptr) {
         return page_pointer + (addr & page_mask);
     }
 
@@ -188,15 +185,100 @@ u8* MemoryManager::GetPointer(GPUVAddr addr) {
     return {};
 }
 
-void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) {
-    std::memcpy(dest_buffer, GetPointer(src_addr), size);
+const u8* MemoryManager::GetPointer(GPUVAddr addr) const {
+    if (!IsAddressValid(addr)) {
+        return {};
+    }
+
+    const u8* const page_pointer{page_table.pointers[addr >> page_bits]};
+    if (page_pointer != nullptr) {
+        return page_pointer + (addr & page_mask);
+    }
+
+    LOG_ERROR(HW_GPU, "Unknown GetPointer @ 0x{:016X}", addr);
+    return {};
 }
+
+void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) const {
+    std::size_t remaining_size{size};
+    std::size_t page_index{src_addr >> page_bits};
+    std::size_t page_offset{src_addr & page_mask};
+
+    while (remaining_size > 0) {
+        const std::size_t copy_amount{
+            std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)};
+
+        switch (page_table.attributes[page_index]) {
+        case Common::PageType::Memory: {
+            const u8* src_ptr{page_table.pointers[page_index] + page_offset};
+            rasterizer.FlushRegion(ToCacheAddr(src_ptr), copy_amount);
+            std::memcpy(dest_buffer, src_ptr, copy_amount);
+            break;
+        }
+        default:
+            UNREACHABLE();
+        }
+
+        page_index++;
+        page_offset = 0;
+        dest_buffer = static_cast<u8*>(dest_buffer) + copy_amount;
+        remaining_size -= copy_amount;
+    }
+}
+
 void MemoryManager::WriteBlock(GPUVAddr dest_addr, const void* src_buffer, std::size_t size) {
-    std::memcpy(GetPointer(dest_addr), src_buffer, size);
+    std::size_t remaining_size{size};
+    std::size_t page_index{dest_addr >> page_bits};
+    std::size_t page_offset{dest_addr & page_mask};
+
+    while (remaining_size > 0) {
+        const std::size_t copy_amount{
+            std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)};
+
+        switch (page_table.attributes[page_index]) {
+        case Common::PageType::Memory: {
+            u8* dest_ptr{page_table.pointers[page_index] + page_offset};
+            rasterizer.InvalidateRegion(ToCacheAddr(dest_ptr), copy_amount);
+            std::memcpy(dest_ptr, src_buffer, copy_amount);
+            break;
+        }
+        default:
+            UNREACHABLE();
+        }
+
+        page_index++;
+        page_offset = 0;
+        src_buffer = static_cast<const u8*>(src_buffer) + copy_amount;
+        remaining_size -= copy_amount;
+    }
 }
 
 void MemoryManager::CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size) {
-    std::memcpy(GetPointer(dest_addr), GetPointer(src_addr), size);
+    std::size_t remaining_size{size};
+    std::size_t page_index{src_addr >> page_bits};
+    std::size_t page_offset{src_addr & page_mask};
+
+    while (remaining_size > 0) {
+        const std::size_t copy_amount{
+            std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)};
+
+        switch (page_table.attributes[page_index]) {
+        case Common::PageType::Memory: {
+            const u8* src_ptr{page_table.pointers[page_index] + page_offset};
+            rasterizer.FlushRegion(ToCacheAddr(src_ptr), copy_amount);
+            WriteBlock(dest_addr, src_ptr, copy_amount);
+            break;
+        }
+        default:
+            UNREACHABLE();
+        }
+
+        page_index++;
+        page_offset = 0;
+        dest_addr += static_cast<VAddr>(copy_amount);
+        src_addr += static_cast<VAddr>(copy_amount);
+        remaining_size -= copy_amount;
+    }
 }
 
 void MemoryManager::MapPages(GPUVAddr base, u64 size, u8* memory, Common::PageType type,
@@ -336,7 +418,7 @@ MemoryManager::VMAIter MemoryManager::CarveVMA(GPUVAddr base, u64 size) {
     const VirtualMemoryArea& vma{vma_handle->second};
     if (vma.type == VirtualMemoryArea::Type::Mapped) {
         // Region is already allocated
-        return {};
+        return vma_handle;
     }
 
     const VAddr start_in_vma{base - vma.base};

src/video_core/memory_manager.h

@@ -10,6 +10,10 @@
 #include "common/common_types.h"
 #include "common/page_table.h"
 
+namespace VideoCore {
+class RasterizerInterface;
+}
+
 namespace Tegra {
 
 /**
@@ -43,24 +47,25 @@ struct VirtualMemoryArea {
 
 class MemoryManager final {
 public:
-    MemoryManager();
+    MemoryManager(VideoCore::RasterizerInterface& rasterizer);
 
     GPUVAddr AllocateSpace(u64 size, u64 align);
     GPUVAddr AllocateSpace(GPUVAddr addr, u64 size, u64 align);
     GPUVAddr MapBufferEx(VAddr cpu_addr, u64 size);
     GPUVAddr MapBufferEx(VAddr cpu_addr, GPUVAddr addr, u64 size);
     GPUVAddr UnmapBuffer(GPUVAddr addr, u64 size);
-    std::optional<VAddr> GpuToCpuAddress(GPUVAddr addr);
+    std::optional<VAddr> GpuToCpuAddress(GPUVAddr addr) const;
 
     template <typename T>
-    T Read(GPUVAddr addr);
+    T Read(GPUVAddr addr) const;
 
     template <typename T>
     void Write(GPUVAddr addr, T data);
 
     u8* GetPointer(GPUVAddr addr);
+    const u8* GetPointer(GPUVAddr addr) const;
 
-    void ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size);
+    void ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) const;
     void WriteBlock(GPUVAddr dest_addr, const void* src_buffer, std::size_t size);
     void CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size);
 
@@ -127,7 +132,7 @@ private:
     void UpdatePageTableForVMA(const VirtualMemoryArea& vma);
 
     /// Finds a free (unmapped region) of the specified size starting at the specified address.
-    GPUVAddr FindFreeRegion(GPUVAddr region_start, u64 size);
+    GPUVAddr FindFreeRegion(GPUVAddr region_start, u64 size) const;
 
 private:
     static constexpr u64 page_bits{16};
@@ -143,6 +148,7 @@ private:
 
     Common::PageTable page_table{page_bits};
     VMAMap vma_map;
+    VideoCore::RasterizerInterface& rasterizer;
 };
 
 } // namespace Tegra

src/video_core/renderer_opengl/gl_buffer_cache.cpp

@@ -7,7 +7,6 @@
 
 #include "common/alignment.h"
 #include "core/core.h"
-#include "core/memory.h"
 #include "video_core/renderer_opengl/gl_buffer_cache.h"
 #include "video_core/renderer_opengl/gl_rasterizer.h"
 

src/video_core/renderer_opengl/gl_global_cache.cpp

@@ -4,7 +4,6 @@
 
 #include <glad/glad.h>
 
-#include "common/assert.h"
 #include "common/logging/log.h"
 #include "core/core.h"
 #include "video_core/renderer_opengl/gl_global_cache.h"

src/video_core/renderer_opengl/gl_primitive_assembler.cpp

@@ -7,7 +7,7 @@
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "core/core.h"
-#include "core/memory.h"
+#include "video_core/memory_manager.h"
 #include "video_core/renderer_opengl/gl_buffer_cache.h"
 #include "video_core/renderer_opengl/gl_primitive_assembler.h"
 

src/video_core/renderer_opengl/gl_primitive_assembler.h

@@ -4,11 +4,9 @@
 
 #pragma once
 
-#include <vector>
 #include <glad/glad.h>
 
 #include "common/common_types.h"
-#include "video_core/memory_manager.h"
 
 namespace OpenGL {