Re-enables building yuzu with Vulkan enabled via the CI scripts

This is possible now with the updated Docker images and their updated packages.
Before, there were build errors due to old QT5 packages on Ubuntu, but now since
they have updated packages it is feasible to build with Vulkan enabled once more.

.ci/scripts/linux/docker.sh

@@ -5,7 +5,7 @@ cd /yuzu
 ccache -s
 
 mkdir build || true && cd build
-cmake .. -G Ninja -DDISPLAY_VERSION=$1 -DYUZU_USE_BUNDLED_UNICORN=ON -DYUZU_USE_QT_WEB_ENGINE=ON -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=/usr/lib/ccache/gcc -DCMAKE_CXX_COMPILER=/usr/lib/ccache/g++ -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${ENABLE_COMPATIBILITY_REPORTING:-"OFF"} -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DUSE_DISCORD_PRESENCE=ON -DENABLE_VULKAN=No
+cmake .. -G Ninja -DDISPLAY_VERSION=$1 -DYUZU_USE_BUNDLED_UNICORN=ON -DYUZU_USE_QT_WEB_ENGINE=ON -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=/usr/lib/ccache/gcc -DCMAKE_CXX_COMPILER=/usr/lib/ccache/g++ -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${ENABLE_COMPATIBILITY_REPORTING:-"OFF"} -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DUSE_DISCORD_PRESENCE=ON
 
 ninja
 

.ci/scripts/windows/docker.sh

@@ -13,7 +13,7 @@ echo '' >> /bin/cmd
 chmod +x /bin/cmd
 
 mkdir build || true && cd build
-cmake .. -G Ninja -DDISPLAY_VERSION=$1 -DCMAKE_TOOLCHAIN_FILE="$(pwd)/../CMakeModules/MinGWCross.cmake" -DUSE_CCACHE=ON -DYUZU_USE_BUNDLED_UNICORN=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DCMAKE_BUILD_TYPE=Release -DENABLE_VULKAN=No
+cmake .. -G Ninja -DDISPLAY_VERSION=$1 -DCMAKE_TOOLCHAIN_FILE="$(pwd)/../CMakeModules/MinGWCross.cmake" -DUSE_CCACHE=ON -DYUZU_USE_BUNDLED_UNICORN=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DCMAKE_BUILD_TYPE=Release
 ninja
 
 # Clean up the dirty hacks
@@ -29,7 +29,13 @@ echo 'Prepare binaries...'
 cd ..
 mkdir package
 
-QT_PLATFORM_DLL_PATH='/usr/x86_64-w64-mingw32/lib/qt5/plugins/platforms/'
+if [ -d "/usr/x86_64-w64-mingw32/lib/qt5/plugins/platforms/" ]; then
+  QT_PLATFORM_DLL_PATH='/usr/x86_64-w64-mingw32/lib/qt5/plugins/platforms/'
+else
+  #fallback to qt
+  QT_PLATFORM_DLL_PATH='/usr/x86_64-w64-mingw32/lib/qt/plugins/platforms/'
+fi
+
 find build/ -name "yuzu*.exe" -exec cp {} 'package' \;
 
 # copy Qt plugins

CMakeModules/MinGWCross.cmake

@@ -10,8 +10,8 @@ set(SDL2_PATH                       ${MINGW_PREFIX})
 set(MINGW_TOOL_PREFIX               ${CMAKE_SYSTEM_PROCESSOR}-w64-mingw32-)
 
 # Specify the cross compiler
-set(CMAKE_C_COMPILER            ${MINGW_TOOL_PREFIX}gcc-posix)
-set(CMAKE_CXX_COMPILER          ${MINGW_TOOL_PREFIX}g++-posix)
+set(CMAKE_C_COMPILER            ${MINGW_TOOL_PREFIX}gcc)
+set(CMAKE_CXX_COMPILER          ${MINGW_TOOL_PREFIX}g++)
 set(CMAKE_RC_COMPILER           ${MINGW_TOOL_PREFIX}windres)
 
 # Mingw tools

src/core/arm/dynarmic/arm_dynarmic_64.cpp

@@ -185,10 +185,9 @@ void ARM_Dynarmic_64::Step() {
 
 ARM_Dynarmic_64::ARM_Dynarmic_64(System& system, ExclusiveMonitor& exclusive_monitor,
                                  std::size_t core_index)
-    : ARM_Interface{system},
-      cb(std::make_unique<DynarmicCallbacks64>(*this)), inner_unicorn{system},
-      core_index{core_index}, exclusive_monitor{
-                                  dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {}
+    : ARM_Interface{system}, cb(std::make_unique<DynarmicCallbacks64>(*this)),
+      inner_unicorn{system, ARM_Unicorn::Arch::AArch64}, core_index{core_index},
+      exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {}
 
 ARM_Dynarmic_64::~ARM_Dynarmic_64() = default;
 

src/core/arm/unicorn/arm_unicorn.cpp

@@ -62,8 +62,9 @@ static bool UnmappedMemoryHook(uc_engine* uc, uc_mem_type type, u64 addr, int si
     return false;
 }
 
-ARM_Unicorn::ARM_Unicorn(System& system) : ARM_Interface{system} {
-    CHECKED(uc_open(UC_ARCH_ARM64, UC_MODE_ARM, &uc));
+ARM_Unicorn::ARM_Unicorn(System& system, Arch architecture) : ARM_Interface{system} {
+    const auto arch = architecture == Arch::AArch32 ? UC_ARCH_ARM : UC_ARCH_ARM64;
+    CHECKED(uc_open(arch, UC_MODE_ARM, &uc));
 
     auto fpv = 3 << 20;
     CHECKED(uc_reg_write(uc, UC_ARM64_REG_CPACR_EL1, &fpv));

src/core/arm/unicorn/arm_unicorn.h

@@ -15,7 +15,12 @@ class System;
 
 class ARM_Unicorn final : public ARM_Interface {
 public:
-    explicit ARM_Unicorn(System& system);
+    enum class Arch {
+        AArch32, // 32-bit ARM
+        AArch64, // 64-bit ARM
+    };
+
+    explicit ARM_Unicorn(System& system, Arch architecture);
     ~ARM_Unicorn() override;
 
     void SetPC(u64 pc) override;

src/core/hle/kernel/handle_table.cpp

@@ -30,6 +30,7 @@ HandleTable::~HandleTable() = default;
 
 ResultCode HandleTable::SetSize(s32 handle_table_size) {
     if (static_cast<u32>(handle_table_size) > MAX_COUNT) {
+        LOG_ERROR(Kernel, "Handle table size {} is greater than {}", handle_table_size, MAX_COUNT);
         return ERR_OUT_OF_MEMORY;
     }
 
@@ -80,6 +81,7 @@ ResultVal<Handle> HandleTable::Duplicate(Handle handle) {
 
 ResultCode HandleTable::Close(Handle handle) {
     if (!IsValid(handle)) {
+        LOG_ERROR(Kernel, "Handle is not valid! handle={:08X}", handle);
         return ERR_INVALID_HANDLE;
     }
 

src/core/hle/kernel/mutex.cpp

@@ -7,6 +7,7 @@
 #include <vector>
 
 #include "common/assert.h"
+#include "common/logging/log.h"
 #include "core/core.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/handle_table.h"
@@ -67,6 +68,7 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle,
                              Handle requesting_thread_handle) {
     // The mutex address must be 4-byte aligned
     if ((address % sizeof(u32)) != 0) {
+        LOG_ERROR(Kernel, "Address is not 4-byte aligned! address={:016X}", address);
         return ERR_INVALID_ADDRESS;
     }
 
@@ -88,6 +90,8 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle,
     }
 
     if (holding_thread == nullptr) {
+        LOG_ERROR(Kernel, "Holding thread does not exist! thread_handle={:08X}",
+                  holding_thread_handle);
         return ERR_INVALID_HANDLE;
     }
 
@@ -109,6 +113,7 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle,
 ResultCode Mutex::Release(VAddr address) {
     // The mutex address must be 4-byte aligned
     if ((address % sizeof(u32)) != 0) {
+        LOG_ERROR(Kernel, "Address is not 4-byte aligned! address={:016X}", address);
         return ERR_INVALID_ADDRESS;
     }
 

src/core/hle/kernel/physical_core.cpp

@@ -27,7 +27,9 @@ PhysicalCore::PhysicalCore(Core::System& system, std::size_t id,
         std::make_unique<Core::ARM_Dynarmic_64>(system, exclusive_monitor, core_index);
 
 #else
-    arm_interface = std::make_shared<Core::ARM_Unicorn>(system);
+    using Core::ARM_Unicorn;
+    arm_interface_32 = std::make_unique<ARM_Unicorn>(system, ARM_Unicorn::Arch::AArch32);
+    arm_interface_64 = std::make_unique<ARM_Unicorn>(system, ARM_Unicorn::Arch::AArch64);
     LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
 #endif
 

src/core/hle/kernel/process_capability.cpp

@@ -3,6 +3,7 @@
 // Refer to the license.txt file included.
 
 #include "common/bit_util.h"
+#include "common/logging/log.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/handle_table.h"
 #include "core/hle/kernel/memory/page_table.h"
@@ -119,22 +120,30 @@ ResultCode ProcessCapabilities::ParseCapabilities(const u32* capabilities,
             // The MapPhysical type uses two descriptor flags for its parameters.
             // If there's only one, then there's a problem.
             if (i >= num_capabilities) {
+                LOG_ERROR(Kernel, "Invalid combination! i={}", i);
                 return ERR_INVALID_COMBINATION;
             }
 
             const auto size_flags = capabilities[i];
             if (GetCapabilityType(size_flags) != CapabilityType::MapPhysical) {
+                LOG_ERROR(Kernel, "Invalid capability type! size_flags={}", size_flags);
                 return ERR_INVALID_COMBINATION;
             }
 
             const auto result = HandleMapPhysicalFlags(descriptor, size_flags, page_table);
             if (result.IsError()) {
+                LOG_ERROR(Kernel, "Failed to map physical flags! descriptor={}, size_flags={}",
+                          descriptor, size_flags);
                 return result;
             }
         } else {
             const auto result =
                 ParseSingleFlagCapability(set_flags, set_svc_bits, descriptor, page_table);
             if (result.IsError()) {
+                LOG_ERROR(
+                    Kernel,
+                    "Failed to parse capability flag! set_flags={}, set_svc_bits={}, descriptor={}",
+                    set_flags, set_svc_bits, descriptor);
                 return result;
             }
         }
@@ -162,6 +171,9 @@ ResultCode ProcessCapabilities::ParseSingleFlagCapability(u32& set_flags, u32& s
     const u32 flag_length = GetFlagBitOffset(type);
     const u32 set_flag = 1U << flag_length;
     if ((set_flag & set_flags & InitializeOnceMask) != 0) {
+        LOG_ERROR(Kernel,
+                  "Attempted to initialize flags that may only be initialized once. set_flags={}",
+                  set_flags);
         return ERR_INVALID_COMBINATION;
     }
     set_flags |= set_flag;
@@ -187,6 +199,7 @@ ResultCode ProcessCapabilities::ParseSingleFlagCapability(u32& set_flags, u32& s
         break;
     }
 
+    LOG_ERROR(Kernel, "Invalid capability type! type={}", static_cast<u32>(type));
     return ERR_INVALID_CAPABILITY_DESCRIPTOR;
 }
 
@@ -208,23 +221,31 @@ void ProcessCapabilities::Clear() {
 
 ResultCode ProcessCapabilities::HandlePriorityCoreNumFlags(u32 flags) {
     if (priority_mask != 0 || core_mask != 0) {
+        LOG_ERROR(Kernel, "Core or priority mask are not zero! priority_mask={}, core_mask={}",
+                  priority_mask, core_mask);
         return ERR_INVALID_CAPABILITY_DESCRIPTOR;
     }
 
     const u32 core_num_min = (flags >> 16) & 0xFF;
     const u32 core_num_max = (flags >> 24) & 0xFF;
     if (core_num_min > core_num_max) {
+        LOG_ERROR(Kernel, "Core min is greater than core max! core_num_min={}, core_num_max={}",
+                  core_num_min, core_num_max);
         return ERR_INVALID_COMBINATION;
     }
 
     const u32 priority_min = (flags >> 10) & 0x3F;
     const u32 priority_max = (flags >> 4) & 0x3F;
     if (priority_min > priority_max) {
+        LOG_ERROR(Kernel,
+                  "Priority min is greater than priority max! priority_min={}, priority_max={}",
+                  core_num_min, priority_max);
         return ERR_INVALID_COMBINATION;
     }
 
     // The switch only has 4 usable cores.
     if (core_num_max >= 4) {
+        LOG_ERROR(Kernel, "Invalid max cores specified! core_num_max={}", core_num_max);
         return ERR_INVALID_PROCESSOR_ID;
     }
 
@@ -259,6 +280,7 @@ ResultCode ProcessCapabilities::HandleSyscallFlags(u32& set_svc_bits, u32 flags)
         }
 
         if (svc_number >= svc_capabilities.size()) {
+            LOG_ERROR(Kernel, "Process svc capability is out of range! svc_number={}", svc_number);
             return ERR_OUT_OF_RANGE;
         }
 
@@ -295,6 +317,8 @@ ResultCode ProcessCapabilities::HandleInterruptFlags(u32 flags) {
         // emulate that, it's sufficient to mark every interrupt as defined.
 
         if (interrupt >= interrupt_capabilities.size()) {
+            LOG_ERROR(Kernel, "Process interrupt capability is out of range! svc_number={}",
+                      interrupt);
             return ERR_OUT_OF_RANGE;
         }
 
@@ -307,6 +331,7 @@ ResultCode ProcessCapabilities::HandleInterruptFlags(u32 flags) {
 ResultCode ProcessCapabilities::HandleProgramTypeFlags(u32 flags) {
     const u32 reserved = flags >> 17;
     if (reserved != 0) {
+        LOG_ERROR(Kernel, "Reserved value is non-zero! reserved={}", reserved);
         return ERR_RESERVED_VALUE;
     }
 
@@ -324,6 +349,9 @@ ResultCode ProcessCapabilities::HandleKernelVersionFlags(u32 flags) {
     const u32 major_version = kernel_version >> 19;
 
     if (major_version != 0 || flags < 0x80000) {
+        LOG_ERROR(Kernel,
+                  "Kernel version is non zero or flags are too small! major_version={}, flags={}",
+                  major_version, flags);
         return ERR_INVALID_CAPABILITY_DESCRIPTOR;
     }
 
@@ -334,6 +362,7 @@ ResultCode ProcessCapabilities::HandleKernelVersionFlags(u32 flags) {
 ResultCode ProcessCapabilities::HandleHandleTableFlags(u32 flags) {
     const u32 reserved = flags >> 26;
     if (reserved != 0) {
+        LOG_ERROR(Kernel, "Reserved value is non-zero! reserved={}", reserved);
         return ERR_RESERVED_VALUE;
     }
 
@@ -344,6 +373,7 @@ ResultCode ProcessCapabilities::HandleHandleTableFlags(u32 flags) {
 ResultCode ProcessCapabilities::HandleDebugFlags(u32 flags) {
     const u32 reserved = flags >> 19;
     if (reserved != 0) {
+        LOG_ERROR(Kernel, "Reserved value is non-zero! reserved={}", reserved);
         return ERR_RESERVED_VALUE;
     }
 

src/core/hle/kernel/readable_event.cpp

@@ -4,6 +4,7 @@
 
 #include <algorithm>
 #include "common/assert.h"
+#include "common/logging/log.h"
 #include "core/hle/kernel/errors.h"
 #include "core/hle/kernel/object.h"
 #include "core/hle/kernel/readable_event.h"
@@ -23,10 +24,12 @@ void ReadableEvent::Acquire(Thread* thread) {
 }
 
 void ReadableEvent::Signal() {
-    if (!is_signaled) {
-        is_signaled = true;
-        SynchronizationObject::Signal();
-    };
+    if (is_signaled) {
+        return;
+    }
+
+    is_signaled = true;
+    SynchronizationObject::Signal();
 }
 
 void ReadableEvent::Clear() {
@@ -35,6 +38,8 @@ void ReadableEvent::Clear() {
 
 ResultCode ReadableEvent::Reset() {
     if (!is_signaled) {
+        LOG_ERROR(Kernel, "Handle is not signaled! object_id={}, object_type={}, object_name={}",
+                  GetObjectId(), GetTypeName(), GetName());
         return ERR_INVALID_STATE;
     }
 

src/core/hle/kernel/resource_limit.cpp

@@ -69,6 +69,8 @@ ResultCode ResourceLimit::SetLimitValue(ResourceType resource, s64 value) {
         limit[index] = value;
         return RESULT_SUCCESS;
     } else {
+        LOG_ERROR(Kernel, "Limit value is too large! resource={}, value={}, index={}",
+                  static_cast<u32>(resource), value, index);
         return ERR_INVALID_STATE;
     }
 }

src/core/hle/kernel/svc.cpp

@@ -685,6 +685,8 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha
     case GetInfoType::TotalPhysicalMemoryAvailableWithoutSystemResource:
     case GetInfoType::TotalPhysicalMemoryUsedWithoutSystemResource: {
         if (info_sub_id != 0) {
+            LOG_ERROR(Kernel_SVC, "Info sub id is non zero! info_id={}, info_sub_id={}", info_id,
+                      info_sub_id);
             return ERR_INVALID_ENUM_VALUE;
         }
 
@@ -692,6 +694,8 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha
             system.Kernel().CurrentProcess()->GetHandleTable();
         const auto process = current_process_handle_table.Get<Process>(static_cast<Handle>(handle));
         if (!process) {
+            LOG_ERROR(Kernel_SVC, "Process is not valid! info_id={}, info_sub_id={}, handle={:08X}",
+                      info_id, info_sub_id, handle);
             return ERR_INVALID_HANDLE;
         }
 
@@ -773,7 +777,7 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha
             break;
         }
 
-        LOG_WARNING(Kernel_SVC, "(STUBBED) Unimplemented svcGetInfo id=0x{:016X}", info_id);
+        LOG_ERROR(Kernel_SVC, "Unimplemented svcGetInfo id=0x{:016X}", info_id);
         return ERR_INVALID_ENUM_VALUE;
     }
 
@@ -783,10 +787,13 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha
 
     case GetInfoType::RegisterResourceLimit: {
         if (handle != 0) {
+            LOG_ERROR(Kernel, "Handle is non zero! handle={:08X}", handle);
             return ERR_INVALID_HANDLE;
         }
 
         if (info_sub_id != 0) {
+            LOG_ERROR(Kernel, "Info sub id is non zero! info_id={}, info_sub_id={}", info_id,
+                      info_sub_id);
             return ERR_INVALID_COMBINATION;
         }
 
@@ -866,7 +873,7 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha
     }
 
     default:
-        LOG_WARNING(Kernel_SVC, "(STUBBED) Unimplemented svcGetInfo id=0x{:016X}", info_id);
+        LOG_ERROR(Kernel_SVC, "Unimplemented svcGetInfo id=0x{:016X}", info_id);
         return ERR_INVALID_ENUM_VALUE;
     }
 }

src/core/hle/kernel/thread.cpp

@@ -423,6 +423,8 @@ ResultCode Thread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) {
     if (new_core == THREADPROCESSORID_DONT_UPDATE) {
         new_core = use_override ? ideal_core_override : ideal_core;
         if ((new_affinity_mask & (1ULL << new_core)) == 0) {
+            LOG_ERROR(Kernel, "New affinity mask is incorrect! new_core={}, new_affinity_mask={}",
+                      new_core, new_affinity_mask);
             return ERR_INVALID_COMBINATION;
         }
     }

src/core/hle/service/acc/acc.cpp

@@ -319,46 +319,37 @@ void Module::Interface::IsUserRegistrationRequestPermitted(Kernel::HLERequestCon
 
 void Module::Interface::InitializeApplicationInfo(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
-    auto pid = rp.Pop<u64>();
 
-    LOG_DEBUG(Service_ACC, "called, process_id={}", pid);
+    LOG_DEBUG(Service_ACC, "called");
     IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(InitializeApplicationInfoBase(pid));
+    rb.Push(InitializeApplicationInfoBase());
 }
 
 void Module::Interface::InitializeApplicationInfoRestricted(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
-    auto pid = rp.Pop<u64>();
 
-    LOG_WARNING(Service_ACC, "(Partial implementation) called, process_id={}", pid);
+    LOG_WARNING(Service_ACC, "(Partial implementation) called");
 
     // TODO(ogniK): We require checking if the user actually owns the title and what not. As of
     // currently, we assume the user owns the title. InitializeApplicationInfoBase SHOULD be called
     // first then we do extra checks if the game is a digital copy.
 
     IPC::ResponseBuilder rb{ctx, 2};
-    rb.Push(InitializeApplicationInfoBase(pid));
+    rb.Push(InitializeApplicationInfoBase());
 }
 
-ResultCode Module::Interface::InitializeApplicationInfoBase(u64 process_id) {
+ResultCode Module::Interface::InitializeApplicationInfoBase() {
     if (application_info) {
         LOG_ERROR(Service_ACC, "Application already initialized");
         return ERR_ACCOUNTINFO_ALREADY_INITIALIZED;
     }
 
-    const auto& list = system.Kernel().GetProcessList();
-    const auto iter = std::find_if(list.begin(), list.end(), [&process_id](const auto& process) {
-        return process->GetProcessID() == process_id;
-    });
-
-    if (iter == list.end()) {
-        LOG_ERROR(Service_ACC, "Failed to find process ID");
-        application_info.application_type = ApplicationType::Unknown;
-
-        return ERR_ACCOUNTINFO_BAD_APPLICATION;
-    }
-
-    const auto launch_property = system.GetARPManager().GetLaunchProperty((*iter)->GetTitleID());
+    // TODO(ogniK): This should be changed to reflect the target process for when we have multiple
+    // processes emulated. As we don't actually have pid support we should assume we're just using
+    // our own process
+    const auto& current_process = system.Kernel().CurrentProcess();
+    const auto launch_property =
+        system.GetARPManager().GetLaunchProperty(current_process->GetTitleID());
 
     if (launch_property.Failed()) {
         LOG_ERROR(Service_ACC, "Failed to get launch property");
@@ -372,10 +363,12 @@ ResultCode Module::Interface::InitializeApplicationInfoBase(u64 process_id) {
     case FileSys::StorageId::Host:
     case FileSys::StorageId::NandUser:
     case FileSys::StorageId::SdCard:
+    case FileSys::StorageId::None: // Yuzu specific, differs from hardware
         application_info.application_type = ApplicationType::Digital;
         break;
     default:
-        LOG_ERROR(Service_ACC, "Invalid game storage ID");
+        LOG_ERROR(Service_ACC, "Invalid game storage ID! storage_id={}",
+                  launch_property->base_game_storage_id);
         return ERR_ACCOUNTINFO_BAD_APPLICATION;
     }
 
@@ -428,6 +421,17 @@ void Module::Interface::GetProfileEditor(Kernel::HLERequestContext& ctx) {
     rb.PushIpcInterface<IProfileEditor>(user_id, *profile_manager);
 }
 
+void Module::Interface::ListQualifiedUsers(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_ACC, "called");
+
+    // All users should be qualified. We don't actually have parental control or anything to do with
+    // nintendo online currently. We're just going to assume the user running the game has access to
+    // the game regardless of parental control settings.
+    ctx.WriteBuffer(profile_manager->GetAllUsers());
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
 void Module::Interface::TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx) {
     LOG_DEBUG(Service_ACC, "called");
     // A u8 is passed into this function which we can safely ignore. It's to determine if we have

src/core/hle/service/acc/acc.h

@@ -33,9 +33,10 @@ public:
         void TrySelectUserWithoutInteraction(Kernel::HLERequestContext& ctx);
         void IsUserAccountSwitchLocked(Kernel::HLERequestContext& ctx);
         void GetProfileEditor(Kernel::HLERequestContext& ctx);
+        void ListQualifiedUsers(Kernel::HLERequestContext& ctx);
 
     private:
-        ResultCode InitializeApplicationInfoBase(u64 process_id);
+        ResultCode InitializeApplicationInfoBase();
 
         enum class ApplicationType : u32_le {
             GameCard = 0,

src/core/hle/service/acc/acc_su.cpp

@@ -35,7 +35,7 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
         {113, nullptr, "GetSaveDataThumbnailExistence"},
         {120, nullptr, "ListOpenUsersInApplication"},
         {130, nullptr, "ActivateOpenContextRetention"},
-        {140, nullptr, "ListQualifiedUsers"},
+        {140, &ACC_SU::ListQualifiedUsers, "ListQualifiedUsers"},
         {150, nullptr, "AuthenticateApplicationAsync"},
         {190, nullptr, "GetUserLastOpenedApplication"},
         {191, nullptr, "ActivateOpenContextHolder"},

src/core/hle/service/acc/acc_u0.cpp

@@ -32,7 +32,7 @@ ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
         {130, nullptr, "LoadOpenContext"},
         {131, nullptr, "ListOpenContextStoredUsers"},
         {140, &ACC_U0::InitializeApplicationInfoRestricted, "InitializeApplicationInfoRestricted"},
-        {141, nullptr, "ListQualifiedUsers"},
+        {141, &ACC_U0::ListQualifiedUsers, "ListQualifiedUsers"},
         {150, &ACC_U0::IsUserAccountSwitchLocked, "IsUserAccountSwitchLocked"},
     };
     // clang-format on

src/core/hle/service/acc/acc_u1.cpp

@@ -34,7 +34,7 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p
         {112, nullptr, "LoadSaveDataThumbnail"},
         {113, nullptr, "GetSaveDataThumbnailExistence"},
         {130, nullptr, "ActivateOpenContextRetention"},
-        {140, nullptr, "ListQualifiedUsers"},
+        {140, &ACC_U1::ListQualifiedUsers, "ListQualifiedUsers"},
         {150, nullptr, "AuthenticateApplicationAsync"},
         {190, nullptr, "GetUserLastOpenedApplication"},
         {191, nullptr, "ActivateOpenContextHolder"},

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

@@ -43,20 +43,15 @@
 
 namespace Service::AM {
 
-constexpr ResultCode ERR_NO_DATA_IN_CHANNEL{ErrorModule::AM, 0x2};
-constexpr ResultCode ERR_NO_MESSAGES{ErrorModule::AM, 0x3};
-constexpr ResultCode ERR_SIZE_OUT_OF_BOUNDS{ErrorModule::AM, 0x1F7};
+constexpr ResultCode ERR_NO_DATA_IN_CHANNEL{ErrorModule::AM, 2};
+constexpr ResultCode ERR_NO_MESSAGES{ErrorModule::AM, 3};
+constexpr ResultCode ERR_SIZE_OUT_OF_BOUNDS{ErrorModule::AM, 503};
 
 enum class LaunchParameterKind : u32 {
     ApplicationSpecific = 1,
     AccountPreselectedUser = 2,
 };
 
-enum class VrMode : u8 {
-    Disabled = 0,
-    Enabled = 1,
-};
-
 constexpr u32 LAUNCH_PARAMETER_ACCOUNT_PRESELECTED_USER_MAGIC = 0xC79497CA;
 
 struct LaunchParameterAccountPreselectedUser {
@@ -685,27 +680,21 @@ void ICommonStateGetter::GetCurrentFocusState(Kernel::HLERequestContext& ctx) {
 }
 
 void ICommonStateGetter::IsVrModeEnabled(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_AM, "(STUBBED) called");
+    LOG_DEBUG(Service_AM, "called");
 
     IPC::ResponseBuilder rb{ctx, 3};
     rb.Push(RESULT_SUCCESS);
-    rb.PushEnum(VrMode::Disabled);
+    rb.Push(vr_mode_state);
 }
 
 void ICommonStateGetter::SetVrModeEnabled(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
-    const auto is_vr_mode_enabled = rp.Pop<bool>();
+    vr_mode_state = rp.Pop<bool>();
 
-    LOG_WARNING(Service_AM, "(STUBBED) called. is_vr_mode_enabled={}", is_vr_mode_enabled);
+    LOG_WARNING(Service_AM, "VR Mode is {}", vr_mode_state ? "on" : "off");
 
     IPC::ResponseBuilder rb{ctx, 2};
-    if (!is_vr_mode_enabled) {
-        rb.Push(RESULT_SUCCESS);
-    } else {
-        // TODO: Find better error code for this
-        UNIMPLEMENTED_MSG("is_vr_mode_enabled={}", is_vr_mode_enabled);
-        rb.Push(RESULT_UNKNOWN);
-    }
+    rb.Push(RESULT_SUCCESS);
 }
 
 void ICommonStateGetter::SetLcdBacklighOffEnabled(Kernel::HLERequestContext& ctx) {
@@ -1169,7 +1158,7 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_)
         {121, nullptr, "ClearUserChannel"},
         {122, nullptr, "UnpopToUserChannel"},
         {130, &IApplicationFunctions::GetGpuErrorDetectedSystemEvent, "GetGpuErrorDetectedSystemEvent"},
-        {140, nullptr, "GetFriendInvitationStorageChannelEvent"},
+        {140, &IApplicationFunctions::GetFriendInvitationStorageChannelEvent, "GetFriendInvitationStorageChannelEvent"},
         {141, nullptr, "TryPopFromFriendInvitationStorageChannel"},
         {150, nullptr, "GetNotificationStorageChannelEvent"},
         {151, nullptr, "TryPopFromNotificationStorageChannel"},
@@ -1186,6 +1175,9 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_)
     auto& kernel = system.Kernel();
     gpu_error_detected_event = Kernel::WritableEvent::CreateEventPair(
         kernel, "IApplicationFunctions:GpuErrorDetectedSystemEvent");
+
+    friend_invitation_storage_channel_event = Kernel::WritableEvent::CreateEventPair(
+        kernel, "IApplicationFunctions:FriendInvitationStorageChannelEvent");
 }
 
 IApplicationFunctions::~IApplicationFunctions() = default;
@@ -1343,12 +1335,23 @@ void IApplicationFunctions::SetTerminateResult(Kernel::HLERequestContext& ctx) {
 }
 
 void IApplicationFunctions::GetDisplayVersion(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_AM, "(STUBBED) called");
+    LOG_DEBUG(Service_AM, "called");
+
+    std::array<u8, 0x10> version_string{};
+
+    FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID()};
+    const auto res = pm.GetControlMetadata();
+    if (res.first != nullptr) {
+        const auto& version = res.first->GetVersionString();
+        std::copy(version.begin(), version.end(), version_string.begin());
+    } else {
+        constexpr u128 default_version = {1, 0};
+        std::memcpy(version_string.data(), default_version.data(), sizeof(u128));
+    }
 
     IPC::ResponseBuilder rb{ctx, 6};
     rb.Push(RESULT_SUCCESS);
-    rb.Push<u64>(1);
-    rb.Push<u64>(0);
+    rb.PushRaw(version_string);
 }
 
 void IApplicationFunctions::GetDesiredLanguage(Kernel::HLERequestContext& ctx) {
@@ -1500,6 +1503,14 @@ void IApplicationFunctions::GetGpuErrorDetectedSystemEvent(Kernel::HLERequestCon
     rb.PushCopyObjects(gpu_error_detected_event.readable);
 }
 
+void IApplicationFunctions::GetFriendInvitationStorageChannelEvent(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_AM, "called");
+
+    IPC::ResponseBuilder rb{ctx, 2, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushCopyObjects(friend_invitation_storage_channel_event.readable);
+}
+
 void InstallInterfaces(SM::ServiceManager& service_manager,
                        std::shared_ptr<NVFlinger::NVFlinger> nvflinger, Core::System& system) {
     auto message_queue = std::make_shared<AppletMessageQueue>(system.Kernel());

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

@@ -191,6 +191,7 @@ private:
 
     Core::System& system;
     std::shared_ptr<AppletMessageQueue> msg_queue;
+    bool vr_mode_state{};
 };
 
 class IStorageImpl {
@@ -280,10 +281,12 @@ private:
     void QueryApplicationPlayStatistics(Kernel::HLERequestContext& ctx);
     void QueryApplicationPlayStatisticsByUid(Kernel::HLERequestContext& ctx);
     void GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx);
+    void GetFriendInvitationStorageChannelEvent(Kernel::HLERequestContext& ctx);
 
     bool launch_popped_application_specific = false;
     bool launch_popped_account_preselect = false;
     Kernel::EventPair gpu_error_detected_event;
+    Kernel::EventPair friend_invitation_storage_channel_event;
     Core::System& system;
 };
 

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

@@ -2,6 +2,9 @@
 // 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 {
@@ -36,11 +39,12 @@ public:
 AudInU::AudInU() : ServiceFramework("audin:u") {
     // clang-format off
     static const FunctionInfo functions[] = {
-        {0, nullptr, "ListAudioIns"},
-        {1, nullptr, "OpenAudioIn"},
-        {2, nullptr, "Unknown"},
-        {3, nullptr, "OpenAudioInAuto"},
-        {4, nullptr, "ListAudioInsAuto"},
+        {0, &AudInU::ListAudioIns, "ListAudioIns"},
+        {1, &AudInU::OpenAudioIn, "OpenAudioIn"},
+        {2, &AudInU::ListAudioIns, "ListAudioInsAuto"},
+        {3, &AudInU::OpenAudioIn, "OpenAudioInAuto"},
+        {4, &AudInU::ListAudioInsAutoFiltered, "ListAudioInsAutoFiltered"},
+        {5, &AudInU::OpenAudioInProtocolSpecified, "OpenAudioInProtocolSpecified"},
     };
     // clang-format on
 
@@ -49,4 +53,60 @@ AudInU::AudInU() : ServiceFramework("audin:u") {
 
 AudInU::~AudInU() = default;
 
+void AudInU::ListAudioIns(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_Audio, "called");
+    const std::size_t count = ctx.GetWriteBufferSize() / sizeof(AudioInDeviceName);
+
+    const std::size_t device_count = std::min(count, audio_device_names.size());
+    std::vector<AudioInDeviceName> device_names;
+    device_names.reserve(device_count);
+
+    for (std::size_t i = 0; i < device_count; i++) {
+        const auto& device_name = audio_device_names[i];
+        auto& entry = device_names.emplace_back();
+        device_name.copy(entry.data(), device_name.size());
+    }
+
+    ctx.WriteBuffer(device_names);
+
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push(static_cast<u32>(device_names.size()));
+}
+
+void AudInU::ListAudioInsAutoFiltered(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_Audio, "called");
+    constexpr u32 device_count = 0;
+
+    // Since we don't actually use any other audio input devices, we return 0 devices. Filtered
+    // device listing just omits the default input device
+
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push(static_cast<u32>(device_count));
+}
+
+void AudInU::OpenInOutImpl(Kernel::HLERequestContext& ctx) {
+    AudInOutParams params{};
+    params.channel_count = 2;
+    params.sample_format = SampleFormat::PCM16;
+    params.sample_rate = 48000;
+    params.state = State::Started;
+
+    IPC::ResponseBuilder rb{ctx, 6, 0, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushRaw<AudInOutParams>(params);
+    rb.PushIpcInterface<IAudioIn>();
+}
+
+void AudInU::OpenAudioIn(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_Audio, "(STUBBED) called");
+    OpenInOutImpl(ctx);
+}
+
+void AudInU::OpenAudioInProtocolSpecified(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_Audio, "(STUBBED) called");
+    OpenInOutImpl(ctx);
+}
+
 } // namespace Service::Audio

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

@@ -16,6 +16,35 @@ class AudInU final : public ServiceFramework<AudInU> {
 public:
     explicit AudInU();
     ~AudInU() override;
+
+private:
+    enum class SampleFormat : u32_le {
+        PCM16 = 2,
+    };
+
+    enum class State : u32_le {
+        Started = 0,
+        Stopped = 1,
+    };
+
+    struct AudInOutParams {
+        u32_le sample_rate{};
+        u32_le channel_count{};
+        SampleFormat sample_format{};
+        State state{};
+    };
+    static_assert(sizeof(AudInOutParams) == 0x10, "AudInOutParams is an invalid size");
+
+    using AudioInDeviceName = std::array<char, 256>;
+    static constexpr std::array<std::string_view, 1> audio_device_names{{
+        "BuiltInHeadset",
+    }};
+
+    void ListAudioIns(Kernel::HLERequestContext& ctx);
+    void ListAudioInsAutoFiltered(Kernel::HLERequestContext& ctx);
+    void OpenInOutImpl(Kernel::HLERequestContext& ctx);
+    void OpenAudioIn(Kernel::HLERequestContext& ctx);
+    void OpenAudioInProtocolSpecified(Kernel::HLERequestContext& ctx);
 };
 
 } // namespace Service::Audio

src/core/hle/service/caps/caps_su.cpp

@@ -2,6 +2,8 @@
 // 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/service/caps/caps_su.h"
 
 namespace Service::Capture {
@@ -9,8 +11,11 @@ namespace Service::Capture {
 CAPS_SU::CAPS_SU() : ServiceFramework("caps:su") {
     // clang-format off
     static const FunctionInfo functions[] = {
+        {32, &CAPS_SU::SetShimLibraryVersion, "SetShimLibraryVersion"},
         {201, nullptr, "SaveScreenShot"},
         {203, nullptr, "SaveScreenShotEx0"},
+        {205, nullptr, "SaveScreenShotEx1"},
+        {210, nullptr, "SaveScreenShotEx2"},
     };
     // clang-format on
 
@@ -19,4 +24,11 @@ CAPS_SU::CAPS_SU() : ServiceFramework("caps:su") {
 
 CAPS_SU::~CAPS_SU() = default;
 
+void CAPS_SU::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_Capture, "(STUBBED) called");
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
 } // namespace Service::Capture

src/core/hle/service/caps/caps_su.h

@@ -16,6 +16,9 @@ class CAPS_SU final : public ServiceFramework<CAPS_SU> {
 public:
     explicit CAPS_SU();
     ~CAPS_SU() override;
+
+private:
+    void SetShimLibraryVersion(Kernel::HLERequestContext& ctx);
 };
 
 } // namespace Service::Capture

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

@@ -316,8 +316,8 @@ public:
             {8, &IFileSystem::OpenFile, "OpenFile"},
             {9, &IFileSystem::OpenDirectory, "OpenDirectory"},
             {10, &IFileSystem::Commit, "Commit"},
-            {11, nullptr, "GetFreeSpaceSize"},
-            {12, nullptr, "GetTotalSpaceSize"},
+            {11, &IFileSystem::GetFreeSpaceSize, "GetFreeSpaceSize"},
+            {12, &IFileSystem::GetTotalSpaceSize, "GetTotalSpaceSize"},
             {13, &IFileSystem::CleanDirectoryRecursively, "CleanDirectoryRecursively"},
             {14, nullptr, "GetFileTimeStampRaw"},
             {15, nullptr, "QueryEntry"},

src/core/hle/service/glue/errors.h

@@ -8,9 +8,9 @@
 
 namespace Service::Glue {
 
-constexpr ResultCode ERR_INVALID_RESOURCE{ErrorModule::ARP, 0x1E};
-constexpr ResultCode ERR_INVALID_PROCESS_ID{ErrorModule::ARP, 0x1F};
-constexpr ResultCode ERR_INVALID_ACCESS{ErrorModule::ARP, 0x2A};
-constexpr ResultCode ERR_NOT_REGISTERED{ErrorModule::ARP, 0x66};
+constexpr ResultCode ERR_INVALID_RESOURCE{ErrorModule::ARP, 30};
+constexpr ResultCode ERR_INVALID_PROCESS_ID{ErrorModule::ARP, 31};
+constexpr ResultCode ERR_INVALID_ACCESS{ErrorModule::ARP, 42};
+constexpr ResultCode ERR_NOT_REGISTERED{ErrorModule::ARP, 102};
 
 } // namespace Service::Glue

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

@@ -233,7 +233,7 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) {
         {302, nullptr, "StopConsoleSixAxisSensor"},
         {303, nullptr, "ActivateSevenSixAxisSensor"},
         {304, nullptr, "StartSevenSixAxisSensor"},
-        {305, nullptr, "StopSevenSixAxisSensor"},
+        {305, &Hid::StopSevenSixAxisSensor, "StopSevenSixAxisSensor"},
         {306, &Hid::InitializeSevenSixAxisSensor, "InitializeSevenSixAxisSensor"},
         {307, nullptr, "FinalizeSevenSixAxisSensor"},
         {308, nullptr, "SetSevenSixAxisSensorFusionStrength"},
@@ -853,6 +853,17 @@ void Hid::SetPalmaBoostMode(Kernel::HLERequestContext& ctx) {
     rb.Push(RESULT_SUCCESS);
 }
 
+void Hid::StopSevenSixAxisSensor(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}",
+                applet_resource_user_id);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
 void Hid::InitializeSevenSixAxisSensor(Kernel::HLERequestContext& ctx) {
     LOG_WARNING(Service_HID, "(STUBBED) called");
 

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

@@ -128,6 +128,7 @@ private:
     void StopSixAxisSensor(Kernel::HLERequestContext& ctx);
     void SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx);
     void SetPalmaBoostMode(Kernel::HLERequestContext& ctx);
+    void StopSevenSixAxisSensor(Kernel::HLERequestContext& ctx);
     void InitializeSevenSixAxisSensor(Kernel::HLERequestContext& ctx);
 
     std::shared_ptr<IAppletResource> applet_resource;

src/core/hle/service/nim/nim.cpp

@@ -15,6 +15,66 @@
 
 namespace Service::NIM {
 
+class IShopServiceAsync final : public ServiceFramework<IShopServiceAsync> {
+public:
+    IShopServiceAsync() : ServiceFramework("IShopServiceAsync") {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, nullptr, "Cancel"},
+            {1, nullptr, "GetSize"},
+            {2, nullptr, "Read"},
+            {3, nullptr, "GetErrorCode"},
+            {4, nullptr, "Request"},
+            {5, nullptr, "Prepare"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+};
+
+class IShopServiceAccessor final : public ServiceFramework<IShopServiceAccessor> {
+public:
+    IShopServiceAccessor() : ServiceFramework("IShopServiceAccessor") {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, &IShopServiceAccessor::CreateAsyncInterface, "CreateAsyncInterface"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+
+private:
+    void CreateAsyncInterface(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NIM, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IShopServiceAsync>();
+    }
+};
+
+class IShopServiceAccessServer final : public ServiceFramework<IShopServiceAccessServer> {
+public:
+    IShopServiceAccessServer() : ServiceFramework("IShopServiceAccessServer") {
+        // clang-format off
+        static const FunctionInfo functions[] = {
+            {0, &IShopServiceAccessServer::CreateAccessorInterface, "CreateAccessorInterface"},
+        };
+        // clang-format on
+
+        RegisterHandlers(functions);
+    }
+
+private:
+    void CreateAccessorInterface(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NIM, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IShopServiceAccessor>();
+    }
+};
+
 class NIM final : public ServiceFramework<NIM> {
 public:
     explicit NIM() : ServiceFramework{"nim"} {
@@ -78,7 +138,7 @@ public:
     explicit NIM_ECA() : ServiceFramework{"nim:eca"} {
         // clang-format off
         static const FunctionInfo functions[] = {
-            {0, nullptr, "CreateServerInterface"},
+            {0, &NIM_ECA::CreateServerInterface, "CreateServerInterface"},
             {1, nullptr, "RefreshDebugAvailability"},
             {2, nullptr, "ClearDebugResponse"},
             {3, nullptr, "RegisterDebugResponse"},
@@ -87,6 +147,14 @@ public:
 
         RegisterHandlers(functions);
     }
+
+private:
+    void CreateServerInterface(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NIM, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushIpcInterface<IShopServiceAccessServer>();
+    }
 };
 
 class NIM_SHP final : public ServiceFramework<NIM_SHP> {

src/core/hle/service/ns/ns.cpp

@@ -371,10 +371,15 @@ ResultVal<u8> IApplicationManagerInterface::GetApplicationDesiredLanguage(
     // Convert to application language, get priority list
     const auto application_language = ConvertToApplicationLanguage(language_code);
     if (application_language == std::nullopt) {
+        LOG_ERROR(Service_NS, "Could not convert application language! language_code={}",
+                  language_code);
         return ERR_APPLICATION_LANGUAGE_NOT_FOUND;
     }
     const auto priority_list = GetApplicationLanguagePriorityList(*application_language);
     if (!priority_list) {
+        LOG_ERROR(Service_NS,
+                  "Could not find application language priorities! application_language={}",
+                  *application_language);
         return ERR_APPLICATION_LANGUAGE_NOT_FOUND;
     }
 
@@ -386,6 +391,8 @@ ResultVal<u8> IApplicationManagerInterface::GetApplicationDesiredLanguage(
         }
     }
 
+    LOG_ERROR(Service_NS, "Could not find a valid language! supported_languages={:08X}",
+              supported_languages);
     return ERR_APPLICATION_LANGUAGE_NOT_FOUND;
 }
 
@@ -410,6 +417,7 @@ ResultVal<u64> IApplicationManagerInterface::ConvertApplicationLanguageToLanguag
     const auto language_code =
         ConvertToLanguageCode(static_cast<ApplicationLanguage>(application_language));
     if (language_code == std::nullopt) {
+        LOG_ERROR(Service_NS, "Language not found! application_language={}", application_language);
         return ERR_APPLICATION_LANGUAGE_NOT_FOUND;
     }
 

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

@@ -159,9 +159,10 @@ private:
     static_assert(sizeof(IoctlFlushL2) == 8, "IoctlFlushL2 is incorrect size");
 
     struct IoctlGetGpuTime {
-        u64_le gpu_time;
+        u64_le gpu_time{};
+        INSERT_PADDING_WORDS(2);
     };
-    static_assert(sizeof(IoctlGetGpuTime) == 8, "IoctlGetGpuTime is incorrect size");
+    static_assert(sizeof(IoctlGetGpuTime) == 0x10, "IoctlGetGpuTime is incorrect size");
 
     u32 GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output,
                            std::vector<u8>& output2, IoctlVersion version);

src/core/hle/service/ptm/psm.cpp

@@ -12,9 +12,6 @@
 
 namespace Service::PSM {
 
-constexpr u32 BATTERY_FULLY_CHARGED = 100;    // 100% Full
-constexpr u32 BATTERY_CURRENTLY_CHARGING = 1; // Plugged into an official dock
-
 class PSM final : public ServiceFramework<PSM> {
 public:
     explicit PSM() : ServiceFramework{"psm"} {
@@ -48,20 +45,30 @@ public:
 
 private:
     void GetBatteryChargePercentage(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_PSM, "(STUBBED) called");
+        LOG_DEBUG(Service_PSM, "called");
 
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(BATTERY_FULLY_CHARGED);
+        rb.Push<u32>(battery_charge_percentage);
     }
 
     void GetChargerType(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_PSM, "(STUBBED) called");
+        LOG_DEBUG(Service_PSM, "called");
 
         IPC::ResponseBuilder rb{ctx, 3};
         rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(BATTERY_CURRENTLY_CHARGING);
+        rb.PushEnum(charger_type);
     }
+
+    enum class ChargerType : u32 {
+        Unplugged = 0,
+        RegularCharger = 1,
+        LowPowerCharger = 2,
+        Unknown = 3,
+    };
+
+    u32 battery_charge_percentage{100}; // 100%
+    ChargerType charger_type{ChargerType::RegularCharger};
 };
 
 void InstallInterfaces(SM::ServiceManager& sm) {

src/core/hle/service/set/set.cpp

@@ -67,6 +67,7 @@ void SET::MakeLanguageCode(Kernel::HLERequestContext& ctx) {
     const auto index = rp.Pop<u32>();
 
     if (index >= available_language_codes.size()) {
+        LOG_ERROR(Service_SET, "Invalid language code index! index={}", index);
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(ERR_INVALID_LANGUAGE);
         return;

src/core/hle/service/sm/sm.cpp

@@ -28,9 +28,11 @@ void ServiceManager::InvokeControlRequest(Kernel::HLERequestContext& context) {
 
 static ResultCode ValidateServiceName(const std::string& name) {
     if (name.size() <= 0 || name.size() > 8) {
+        LOG_ERROR(Service_SM, "Invalid service name! service={}", name);
         return ERR_INVALID_NAME;
     }
     if (name.find('\0') != std::string::npos) {
+        LOG_ERROR(Service_SM, "A non null terminated service was passed");
         return ERR_INVALID_NAME;
     }
     return RESULT_SUCCESS;
@@ -51,8 +53,10 @@ ResultVal<std::shared_ptr<Kernel::ServerPort>> ServiceManager::RegisterService(
 
     CASCADE_CODE(ValidateServiceName(name));
 
-    if (registered_services.find(name) != registered_services.end())
+    if (registered_services.find(name) != registered_services.end()) {
+        LOG_ERROR(Service_SM, "Service is already registered! service={}", name);
         return ERR_ALREADY_REGISTERED;
+    }
 
     auto& kernel = Core::System::GetInstance().Kernel();
     auto [server_port, client_port] =
@@ -66,9 +70,10 @@ ResultCode ServiceManager::UnregisterService(const std::string& name) {
     CASCADE_CODE(ValidateServiceName(name));
 
     const auto iter = registered_services.find(name);
-    if (iter == registered_services.end())
+    if (iter == registered_services.end()) {
+        LOG_ERROR(Service_SM, "Server is not registered! service={}", name);
         return ERR_SERVICE_NOT_REGISTERED;
-
+    }
     registered_services.erase(iter);
     return RESULT_SUCCESS;
 }
@@ -79,6 +84,7 @@ ResultVal<std::shared_ptr<Kernel::ClientPort>> ServiceManager::GetServicePort(
     CASCADE_CODE(ValidateServiceName(name));
     auto it = registered_services.find(name);
     if (it == registered_services.end()) {
+        LOG_ERROR(Service_SM, "Server is not registered! service={}", name);
         return ERR_SERVICE_NOT_REGISTERED;
     }
 

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

@@ -267,7 +267,7 @@ protected:
 
 private:
     struct Data {
-        u32_le unk_0;
+        u32_le unk_0{};
     };
 
     Data data{};
@@ -614,6 +614,14 @@ private:
             ctx.WriteBuffer(response.Serialize());
             break;
         }
+        case TransactionId::SetBufferCount: {
+            LOG_WARNING(Service_VI, "(STUBBED) called, transaction=SetBufferCount");
+            [[maybe_unused]] const auto buffer = ctx.ReadBuffer();
+
+            IGBPEmptyResponseParcel response{};
+            ctx.WriteBuffer(response.Serialize());
+            break;
+        }
         default:
             ASSERT_MSG(false, "Unimplemented");
         }
@@ -859,6 +867,7 @@ private:
 
         const auto layer_id = nv_flinger->CreateLayer(display);
         if (!layer_id) {
+            LOG_ERROR(Service_VI, "Layer not found! display=0x{:016X}", display);
             IPC::ResponseBuilder rb{ctx, 2};
             rb.Push(ERR_NOT_FOUND);
             return;
@@ -975,6 +984,7 @@ private:
 
         const auto display_id = nv_flinger->OpenDisplay(name);
         if (!display_id) {
+            LOG_ERROR(Service_VI, "Display not found! display_name={}", name);
             IPC::ResponseBuilder rb{ctx, 2};
             rb.Push(ERR_NOT_FOUND);
             return;
@@ -1074,6 +1084,7 @@ private:
 
         const auto display_id = nv_flinger->OpenDisplay(display_name);
         if (!display_id) {
+            LOG_ERROR(Service_VI, "Layer not found! layer_id={}", layer_id);
             IPC::ResponseBuilder rb{ctx, 2};
             rb.Push(ERR_NOT_FOUND);
             return;
@@ -1081,6 +1092,7 @@ private:
 
         const auto buffer_queue_id = nv_flinger->FindBufferQueueId(*display_id, layer_id);
         if (!buffer_queue_id) {
+            LOG_ERROR(Service_VI, "Buffer queue id not found! display_id={}", *display_id);
             IPC::ResponseBuilder rb{ctx, 2};
             rb.Push(ERR_NOT_FOUND);
             return;
@@ -1116,6 +1128,7 @@ private:
 
         const auto layer_id = nv_flinger->CreateLayer(display_id);
         if (!layer_id) {
+            LOG_ERROR(Service_VI, "Layer not found! layer_id={}", *layer_id);
             IPC::ResponseBuilder rb{ctx, 2};
             rb.Push(ERR_NOT_FOUND);
             return;
@@ -1123,6 +1136,7 @@ private:
 
         const auto buffer_queue_id = nv_flinger->FindBufferQueueId(display_id, *layer_id);
         if (!buffer_queue_id) {
+            LOG_ERROR(Service_VI, "Buffer queue id not found! display_id={}", display_id);
             IPC::ResponseBuilder rb{ctx, 2};
             rb.Push(ERR_NOT_FOUND);
             return;
@@ -1153,6 +1167,7 @@ private:
 
         const auto vsync_event = nv_flinger->FindVsyncEvent(display_id);
         if (!vsync_event) {
+            LOG_ERROR(Service_VI, "Vsync event was not found for display_id={}", display_id);
             IPC::ResponseBuilder rb{ctx, 2};
             rb.Push(ERR_NOT_FOUND);
             return;
@@ -1193,6 +1208,7 @@ private:
         case NintendoScaleMode::PreserveAspectRatio:
             return MakeResult(ConvertedScaleMode::PreserveAspectRatio);
         default:
+            LOG_ERROR(Service_VI, "Invalid scaling mode specified, mode={}", mode);
             return ERR_OPERATION_FAILED;
         }
     }
@@ -1249,6 +1265,7 @@ void detail::GetDisplayServiceImpl(Kernel::HLERequestContext& ctx,
     const auto policy = rp.PopEnum<Policy>();
 
     if (!IsValidServiceAccess(permission, policy)) {
+        LOG_ERROR(Service_VI, "Permission denied for policy {}", static_cast<u32>(policy));
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(ERR_PERMISSION_DENIED);
         return;

src/core/settings.h

@@ -446,6 +446,7 @@ struct Values {
     bool use_asynchronous_gpu_emulation;
     bool use_vsync;
     bool force_30fps_mode;
+    bool use_fast_gpu_time;
 
     float bg_red;
     float bg_green;

src/input_common/main.cpp

@@ -18,7 +18,9 @@ namespace InputCommon {
 
 static std::shared_ptr<Keyboard> keyboard;
 static std::shared_ptr<MotionEmu> motion_emu;
+#ifdef HAVE_SDL2
 static std::unique_ptr<SDL::State> sdl;
+#endif
 static std::unique_ptr<CemuhookUDP::State> udp;
 
 void Init() {
@@ -29,7 +31,9 @@ void Init() {
     motion_emu = std::make_shared<MotionEmu>();
     Input::RegisterFactory<Input::MotionDevice>("motion_emu", motion_emu);
 
+#ifdef HAVE_SDL2
     sdl = SDL::Init();
+#endif
 
     udp = CemuhookUDP::Init();
 }
@@ -40,7 +44,9 @@ void Shutdown() {
     Input::UnregisterFactory<Input::AnalogDevice>("analog_from_button");
     Input::UnregisterFactory<Input::MotionDevice>("motion_emu");
     motion_emu.reset();
+#ifdef HAVE_SDL2
     sdl.reset();
+#endif
     udp.reset();
 }
 

src/video_core/CMakeLists.txt

@@ -124,6 +124,8 @@ add_library(video_core STATIC
     shader/decode.cpp
     shader/expr.cpp
     shader/expr.h
+    shader/memory_util.cpp
+    shader/memory_util.h
     shader/node_helper.cpp
     shader/node_helper.h
     shader/node.h

src/video_core/buffer_cache/buffer_cache.h

@@ -4,7 +4,6 @@
 
 #pragma once
 
-#include <array>
 #include <list>
 #include <memory>
 #include <mutex>

src/video_core/dma_pusher.cpp

@@ -71,16 +71,22 @@ bool DmaPusher::Step() {
     gpu.MemoryManager().ReadBlockUnsafe(dma_get, command_headers.data(),
                                         command_list_header.size * sizeof(u32));
 
-    for (const CommandHeader& command_header : command_headers) {
+    for (std::size_t index = 0; index < command_headers.size();) {
+        const CommandHeader& command_header = command_headers[index];
 
-        // now, see if we're in the middle of a command
-        if (dma_state.length_pending) {
-            // Second word of long non-inc methods command - method count
-            dma_state.length_pending = 0;
-            dma_state.method_count = command_header.method_count_;
-        } else if (dma_state.method_count) {
+        if (dma_state.method_count) {
             // Data word of methods command
-            CallMethod(command_header.argument);
+            if (dma_state.non_incrementing) {
+                const u32 max_write = static_cast<u32>(
+                    std::min<std::size_t>(index + dma_state.method_count, command_headers.size()) -
+                    index);
+                CallMultiMethod(&command_header.argument, max_write);
+                dma_state.method_count -= max_write;
+                index += max_write;
+                continue;
+            } else {
+                CallMethod(command_header.argument);
+            }
 
             if (!dma_state.non_incrementing) {
                 dma_state.method++;
@@ -120,6 +126,7 @@ bool DmaPusher::Step() {
                 break;
             }
         }
+        index++;
     }
 
     if (!non_main) {
@@ -140,4 +147,9 @@ void DmaPusher::CallMethod(u32 argument) const {
     gpu.CallMethod({dma_state.method, argument, dma_state.subchannel, dma_state.method_count});
 }
 
+void DmaPusher::CallMultiMethod(const u32* base_start, u32 num_methods) const {
+    gpu.CallMultiMethod(dma_state.method, dma_state.subchannel, base_start, num_methods,
+                        dma_state.method_count);
+}
+
 } // namespace Tegra

src/video_core/dma_pusher.h

@@ -75,6 +75,7 @@ private:
     void SetState(const CommandHeader& command_header);
 
     void CallMethod(u32 argument) const;
+    void CallMultiMethod(const u32* base_start, u32 num_methods) const;
 
     std::vector<CommandHeader> command_headers; ///< Buffer for list of commands fetched at once
 

src/video_core/engines/fermi_2d.cpp

@@ -28,6 +28,12 @@ void Fermi2D::CallMethod(const GPU::MethodCall& method_call) {
     }
 }
 
+void Fermi2D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending) {
+    for (std::size_t i = 0; i < amount; i++) {
+        CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)});
+    }
+}
+
 static std::pair<u32, u32> DelimitLine(u32 src_1, u32 src_2, u32 dst_1, u32 dst_2, u32 src_line) {
     const u32 line_a = src_2 - src_1;
     const u32 line_b = dst_2 - dst_1;

src/video_core/engines/fermi_2d.h

@@ -39,6 +39,9 @@ public:
     /// Write the value to the register identified by method.
     void CallMethod(const GPU::MethodCall& method_call);
 
+    /// Write multiple values to the register identified by method.
+    void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending);
+
     enum class Origin : u32 {
         Center = 0,
         Corner = 1,

src/video_core/engines/kepler_compute.cpp

@@ -51,6 +51,13 @@ void KeplerCompute::CallMethod(const GPU::MethodCall& method_call) {
     }
 }
 
+void KeplerCompute::CallMultiMethod(u32 method, const u32* base_start, u32 amount,
+                                    u32 methods_pending) {
+    for (std::size_t i = 0; i < amount; i++) {
+        CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)});
+    }
+}
+
 Texture::FullTextureInfo KeplerCompute::GetTexture(std::size_t offset) const {
     const std::bitset<8> cbuf_mask = launch_description.const_buffer_enable_mask.Value();
     ASSERT(cbuf_mask[regs.tex_cb_index]);

src/video_core/engines/kepler_compute.h

@@ -202,6 +202,9 @@ public:
     /// Write the value to the register identified by method.
     void CallMethod(const GPU::MethodCall& method_call);
 
+    /// Write multiple values to the register identified by method.
+    void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending);
+
     Texture::FullTextureInfo GetTexture(std::size_t offset) const;
 
     /// Given a texture handle, returns the TSC and TIC entries.

src/video_core/engines/kepler_memory.cpp

@@ -41,4 +41,11 @@ void KeplerMemory::CallMethod(const GPU::MethodCall& method_call) {
     }
 }
 
+void KeplerMemory::CallMultiMethod(u32 method, const u32* base_start, u32 amount,
+                                   u32 methods_pending) {
+    for (std::size_t i = 0; i < amount; i++) {
+        CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)});
+    }
+}
+
 } // namespace Tegra::Engines

src/video_core/engines/kepler_memory.h

@@ -40,6 +40,9 @@ public:
     /// Write the value to the register identified by method.
     void CallMethod(const GPU::MethodCall& method_call);
 
+    /// Write multiple values to the register identified by method.
+    void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending);
+
     struct Regs {
         static constexpr size_t NUM_REGS = 0x7F;
 

src/video_core/engines/maxwell_3d.cpp

@@ -280,6 +280,58 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) {
     }
 }
 
+void Maxwell3D::CallMultiMethod(u32 method, const u32* base_start, u32 amount,
+                                u32 methods_pending) {
+    // Methods after 0xE00 are special, they're actually triggers for some microcode that was
+    // uploaded to the GPU during initialization.
+    if (method >= MacroRegistersStart) {
+        // We're trying to execute a macro
+        if (executing_macro == 0) {
+            // A macro call must begin by writing the macro method's register, not its argument.
+            ASSERT_MSG((method % 2) == 0,
+                       "Can't start macro execution by writing to the ARGS register");
+            executing_macro = method;
+        }
+
+        for (std::size_t i = 0; i < amount; i++) {
+            macro_params.push_back(base_start[i]);
+        }
+
+        // Call the macro when there are no more parameters in the command buffer
+        if (amount == methods_pending) {
+            CallMacroMethod(executing_macro, macro_params.size(), macro_params.data());
+            macro_params.clear();
+        }
+        return;
+    }
+    switch (method) {
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[1]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[2]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[3]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[4]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[5]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[6]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[7]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[8]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[9]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[10]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[11]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[12]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[13]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[14]):
+    case MAXWELL3D_REG_INDEX(const_buffer.cb_data[15]): {
+        ProcessCBMultiData(method, base_start, amount);
+        break;
+    }
+    default: {
+        for (std::size_t i = 0; i < amount; i++) {
+            CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)});
+        }
+    }
+    }
+}
+
 void Maxwell3D::StepInstance(const MMEDrawMode expected_mode, const u32 count) {
     if (mme_draw.current_mode == MMEDrawMode::Undefined) {
         if (mme_draw.gl_begin_consume) {
@@ -570,6 +622,28 @@ void Maxwell3D::StartCBData(u32 method) {
     ProcessCBData(regs.const_buffer.cb_data[cb_data_state.id]);
 }
 
+void Maxwell3D::ProcessCBMultiData(u32 method, const u32* start_base, u32 amount) {
+    if (cb_data_state.current != method) {
+        if (cb_data_state.current != null_cb_data) {
+            FinishCBData();
+        }
+        constexpr u32 first_cb_data = MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]);
+        cb_data_state.start_pos = regs.const_buffer.cb_pos;
+        cb_data_state.id = method - first_cb_data;
+        cb_data_state.current = method;
+        cb_data_state.counter = 0;
+    }
+    const std::size_t id = cb_data_state.id;
+    const std::size_t size = amount;
+    std::size_t i = 0;
+    for (; i < size; i++) {
+        cb_data_state.buffer[id][cb_data_state.counter] = start_base[i];
+        cb_data_state.counter++;
+    }
+    // Increment the current buffer position.
+    regs.const_buffer.cb_pos = regs.const_buffer.cb_pos + 4 * amount;
+}
+
 void Maxwell3D::FinishCBData() {
     // Write the input value to the current const buffer at the current position.
     const GPUVAddr buffer_address = regs.const_buffer.BufferAddress();

src/video_core/engines/maxwell_3d.h

@@ -1179,6 +1179,7 @@ public:
                     BitField<0, 1, u32> depth_range_0_1;
                     BitField<3, 1, u32> depth_clamp_near;
                     BitField<4, 1, u32> depth_clamp_far;
+                    BitField<11, 1, u32> depth_clamp_disabled;
                 } view_volume_clip_control;
 
                 INSERT_UNION_PADDING_WORDS(0x1F);
@@ -1359,6 +1360,9 @@ public:
     /// Write the value to the register identified by method.
     void CallMethod(const GPU::MethodCall& method_call);
 
+    /// Write multiple values to the register identified by method.
+    void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending);
+
     /// Write the value to the register identified by method.
     void CallMethodFromMME(const GPU::MethodCall& method_call);
 
@@ -1512,6 +1516,7 @@ private:
     /// Handles a write to the CB_DATA[i] register.
     void StartCBData(u32 method);
     void ProcessCBData(u32 value);
+    void ProcessCBMultiData(u32 method, const u32* start_base, u32 amount);
     void FinishCBData();
 
     /// Handles a write to the CB_BIND register.

src/video_core/engines/maxwell_dma.cpp

@@ -36,6 +36,13 @@ void MaxwellDMA::CallMethod(const GPU::MethodCall& method_call) {
 #undef MAXWELLDMA_REG_INDEX
 }
 
+void MaxwellDMA::CallMultiMethod(u32 method, const u32* base_start, u32 amount,
+                                 u32 methods_pending) {
+    for (std::size_t i = 0; i < amount; i++) {
+        CallMethod({method, base_start[i], 0, methods_pending - static_cast<u32>(i)});
+    }
+}
+
 void MaxwellDMA::HandleCopy() {
     LOG_TRACE(HW_GPU, "Requested a DMA copy");
 

src/video_core/engines/maxwell_dma.h

@@ -35,6 +35,9 @@ public:
     /// Write the value to the register identified by method.
     void CallMethod(const GPU::MethodCall& method_call);
 
+    /// Write multiple values to the register identified by method.
+    void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending);
+
     struct Regs {
         static constexpr std::size_t NUM_REGS = 0x1D6;
 

src/video_core/engines/shader_bytecode.h

@@ -813,6 +813,10 @@ union Instruction {
         BitField<49, 1, u64> negate_a;
     } alu_integer;
 
+    union {
+        BitField<43, 1, u64> x;
+    } iadd;
+
     union {
         BitField<39, 1, u64> ftz;
         BitField<32, 1, u64> saturate;

src/video_core/fence_manager.h

@@ -88,7 +88,8 @@ public:
             }
             PopAsyncFlushes();
             if (current_fence->IsSemaphore()) {
-                memory_manager.Write<u32>(current_fence->GetAddress(), current_fence->GetPayload());
+                memory_manager.template Write<u32>(current_fence->GetAddress(),
+                                                   current_fence->GetPayload());
             } else {
                 gpu.IncrementSyncPoint(current_fence->GetPayload());
             }
@@ -134,7 +135,8 @@ private:
             }
             PopAsyncFlushes();
             if (current_fence->IsSemaphore()) {
-                memory_manager.Write<u32>(current_fence->GetAddress(), current_fence->GetPayload());
+                memory_manager.template Write<u32>(current_fence->GetAddress(),
+                                                   current_fence->GetPayload());
             } else {
                 gpu.IncrementSyncPoint(current_fence->GetPayload());
             }

src/video_core/gpu.cpp

@@ -9,6 +9,7 @@
 #include "core/core_timing_util.h"
 #include "core/frontend/emu_window.h"
 #include "core/memory.h"
+#include "core/settings.h"
 #include "video_core/engines/fermi_2d.h"
 #include "video_core/engines/kepler_compute.h"
 #include "video_core/engines/kepler_memory.h"
@@ -154,7 +155,10 @@ u64 GPU::GetTicks() const {
     constexpr u64 gpu_ticks_den = 625;
 
     const u64 cpu_ticks = system.CoreTiming().GetTicks();
-    const u64 nanoseconds = Core::Timing::CyclesToNs(cpu_ticks).count();
+    u64 nanoseconds = Core::Timing::CyclesToNs(cpu_ticks).count();
+    if (Settings::values.use_fast_gpu_time) {
+        nanoseconds /= 256;
+    }
     const u64 nanoseconds_num = nanoseconds / gpu_ticks_den;
     const u64 nanoseconds_rem = nanoseconds % gpu_ticks_den;
     return nanoseconds_num * gpu_ticks_num + (nanoseconds_rem * gpu_ticks_num) / gpu_ticks_den;
@@ -209,16 +213,32 @@ void GPU::CallMethod(const MethodCall& method_call) {
 
     ASSERT(method_call.subchannel < bound_engines.size());
 
-    if (ExecuteMethodOnEngine(method_call)) {
+    if (ExecuteMethodOnEngine(method_call.method)) {
         CallEngineMethod(method_call);
     } else {
         CallPullerMethod(method_call);
     }
 }
 
-bool GPU::ExecuteMethodOnEngine(const MethodCall& method_call) {
-    const auto method = static_cast<BufferMethods>(method_call.method);
-    return method >= BufferMethods::NonPullerMethods;
+void GPU::CallMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 amount,
+                          u32 methods_pending) {
+    LOG_TRACE(HW_GPU, "Processing method {:08X} on subchannel {}", method, subchannel);
+
+    ASSERT(subchannel < bound_engines.size());
+
+    if (ExecuteMethodOnEngine(method)) {
+        CallEngineMultiMethod(method, subchannel, base_start, amount, methods_pending);
+    } else {
+        for (std::size_t i = 0; i < amount; i++) {
+            CallPullerMethod(
+                {method, base_start[i], subchannel, methods_pending - static_cast<u32>(i)});
+        }
+    }
+}
+
+bool GPU::ExecuteMethodOnEngine(u32 method) {
+    const auto buffer_method = static_cast<BufferMethods>(method);
+    return buffer_method >= BufferMethods::NonPullerMethods;
 }
 
 void GPU::CallPullerMethod(const MethodCall& method_call) {
@@ -298,6 +318,31 @@ void GPU::CallEngineMethod(const MethodCall& method_call) {
     }
 }
 
+void GPU::CallEngineMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 amount,
+                                u32 methods_pending) {
+    const EngineID engine = bound_engines[subchannel];
+
+    switch (engine) {
+    case EngineID::FERMI_TWOD_A:
+        fermi_2d->CallMultiMethod(method, base_start, amount, methods_pending);
+        break;
+    case EngineID::MAXWELL_B:
+        maxwell_3d->CallMultiMethod(method, base_start, amount, methods_pending);
+        break;
+    case EngineID::KEPLER_COMPUTE_B:
+        kepler_compute->CallMultiMethod(method, base_start, amount, methods_pending);
+        break;
+    case EngineID::MAXWELL_DMA_COPY_A:
+        maxwell_dma->CallMultiMethod(method, base_start, amount, methods_pending);
+        break;
+    case EngineID::KEPLER_INLINE_TO_MEMORY_B:
+        kepler_memory->CallMultiMethod(method, base_start, amount, methods_pending);
+        break;
+    default:
+        UNIMPLEMENTED_MSG("Unimplemented engine");
+    }
+}
+
 void GPU::ProcessBindMethod(const MethodCall& method_call) {
     // Bind the current subchannel to the desired engine id.
     LOG_DEBUG(HW_GPU, "Binding subchannel {} to engine {}", method_call.subchannel,

src/video_core/gpu.h

@@ -155,6 +155,10 @@ public:
     /// Calls a GPU method.
     void CallMethod(const MethodCall& method_call);
 
+    /// Calls a GPU multivalue method.
+    void CallMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 amount,
+                         u32 methods_pending);
+
     /// Flush all current written commands into the host GPU for execution.
     void FlushCommands();
     /// Synchronizes CPU writes with Host GPU memory.
@@ -309,8 +313,12 @@ private:
     /// Calls a GPU engine method.
     void CallEngineMethod(const MethodCall& method_call);
 
+    /// Calls a GPU engine multivalue method.
+    void CallEngineMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 amount,
+                               u32 methods_pending);
+
     /// Determines where the method should be executed.
-    bool ExecuteMethodOnEngine(const MethodCall& method_call);
+    bool ExecuteMethodOnEngine(u32 method);
 
 protected:
     std::unique_ptr<Tegra::DmaPusher> dma_pusher;

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -59,14 +59,12 @@ constexpr std::size_t NumSupportedVertexAttributes = 16;
 template <typename Engine, typename Entry>
 Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry& entry,
                                                ShaderType shader_type, std::size_t index = 0) {
-    if (entry.IsBindless()) {
-        const Tegra::Texture::TextureHandle tex_handle =
-            engine.AccessConstBuffer32(shader_type, entry.GetBuffer(), entry.GetOffset());
+    if (entry.is_bindless) {
+        const auto tex_handle = engine.AccessConstBuffer32(shader_type, entry.buffer, entry.offset);
         return engine.GetTextureInfo(tex_handle);
     }
     const auto& gpu_profile = engine.AccessGuestDriverProfile();
-    const u32 offset =
-        entry.GetOffset() + static_cast<u32>(index * gpu_profile.GetTextureHandlerSize());
+    const u32 offset = entry.offset + static_cast<u32>(index * gpu_profile.GetTextureHandlerSize());
     if constexpr (std::is_same_v<Engine, Tegra::Engines::Maxwell3D>) {
         return engine.GetStageTexture(shader_type, offset);
     } else {
@@ -856,9 +854,9 @@ void RasterizerOpenGL::SetupDrawGlobalMemory(std::size_t stage_index, const Shad
 
     u32 binding = device.GetBaseBindings(stage_index).shader_storage_buffer;
     for (const auto& entry : shader->GetEntries().global_memory_entries) {
-        const auto addr{cbufs.const_buffers[entry.GetCbufIndex()].address + entry.GetCbufOffset()};
-        const auto gpu_addr{memory_manager.Read<u64>(addr)};
-        const auto size{memory_manager.Read<u32>(addr + 8)};
+        const GPUVAddr addr{cbufs.const_buffers[entry.cbuf_index].address + entry.cbuf_offset};
+        const GPUVAddr gpu_addr{memory_manager.Read<u64>(addr)};
+        const u32 size{memory_manager.Read<u32>(addr + 8)};
         SetupGlobalMemory(binding++, entry, gpu_addr, size);
     }
 }
@@ -870,7 +868,7 @@ void RasterizerOpenGL::SetupComputeGlobalMemory(const Shader& kernel) {
 
     u32 binding = 0;
     for (const auto& entry : kernel->GetEntries().global_memory_entries) {
-        const auto addr{cbufs[entry.GetCbufIndex()].Address() + entry.GetCbufOffset()};
+        const auto addr{cbufs[entry.cbuf_index].Address() + entry.cbuf_offset};
         const auto gpu_addr{memory_manager.Read<u64>(addr)};
         const auto size{memory_manager.Read<u32>(addr + 8)};
         SetupGlobalMemory(binding++, entry, gpu_addr, size);
@@ -881,7 +879,7 @@ void RasterizerOpenGL::SetupGlobalMemory(u32 binding, const GlobalMemoryEntry& e
                                          GPUVAddr gpu_addr, std::size_t size) {
     const auto alignment{device.GetShaderStorageBufferAlignment()};
     const auto [ssbo, buffer_offset] =
-        buffer_cache.UploadMemory(gpu_addr, size, alignment, entry.IsWritten());
+        buffer_cache.UploadMemory(gpu_addr, size, alignment, entry.is_written);
     glBindBufferRange(GL_SHADER_STORAGE_BUFFER, binding, ssbo, buffer_offset,
                       static_cast<GLsizeiptr>(size));
 }
@@ -892,7 +890,7 @@ void RasterizerOpenGL::SetupDrawTextures(std::size_t stage_index, const Shader&
     u32 binding = device.GetBaseBindings(stage_index).sampler;
     for (const auto& entry : shader->GetEntries().samplers) {
         const auto shader_type = static_cast<ShaderType>(stage_index);
-        for (std::size_t i = 0; i < entry.Size(); ++i) {
+        for (std::size_t i = 0; i < entry.size; ++i) {
             const auto texture = GetTextureInfo(maxwell3d, entry, shader_type, i);
             SetupTexture(binding++, texture, entry);
         }
@@ -904,7 +902,7 @@ void RasterizerOpenGL::SetupComputeTextures(const Shader& kernel) {
     const auto& compute = system.GPU().KeplerCompute();
     u32 binding = 0;
     for (const auto& entry : kernel->GetEntries().samplers) {
-        for (std::size_t i = 0; i < entry.Size(); ++i) {
+        for (std::size_t i = 0; i < entry.size; ++i) {
             const auto texture = GetTextureInfo(compute, entry, ShaderType::Compute, i);
             SetupTexture(binding++, texture, entry);
         }
@@ -961,7 +959,7 @@ void RasterizerOpenGL::SetupImage(u32 binding, const Tegra::Texture::TICEntry& t
     if (!tic.IsBuffer()) {
         view->ApplySwizzle(tic.x_source, tic.y_source, tic.z_source, tic.w_source);
     }
-    if (entry.IsWritten()) {
+    if (entry.is_written) {
         view->MarkAsModified(texture_cache.Tick());
     }
     glBindImageTexture(binding, view->GetTexture(), 0, GL_TRUE, 0, GL_READ_WRITE,
@@ -1022,11 +1020,7 @@ void RasterizerOpenGL::SyncDepthClamp() {
     }
     flags[Dirty::DepthClampEnabled] = false;
 
-    const auto& state = gpu.regs.view_volume_clip_control;
-    UNIMPLEMENTED_IF_MSG(state.depth_clamp_far != state.depth_clamp_near,
-                         "Unimplemented depth clamp separation!");
-
-    oglEnable(GL_DEPTH_CLAMP, state.depth_clamp_far || state.depth_clamp_near);
+    oglEnable(GL_DEPTH_CLAMP, gpu.regs.view_volume_clip_control.depth_clamp_disabled == 0);
 }
 
 void RasterizerOpenGL::SyncClipEnabled(u32 clip_mask) {

src/video_core/renderer_opengl/gl_shader_cache.cpp

@@ -10,8 +10,6 @@
 #include <thread>
 #include <unordered_set>
 
-#include <boost/functional/hash.hpp>
-
 #include "common/alignment.h"
 #include "common/assert.h"
 #include "common/logging/log.h"
@@ -28,76 +26,26 @@
 #include "video_core/renderer_opengl/gl_shader_disk_cache.h"
 #include "video_core/renderer_opengl/gl_state_tracker.h"
 #include "video_core/renderer_opengl/utils.h"
+#include "video_core/shader/memory_util.h"
 #include "video_core/shader/registry.h"
 #include "video_core/shader/shader_ir.h"
 
 namespace OpenGL {
 
 using Tegra::Engines::ShaderType;
+using VideoCommon::Shader::GetShaderAddress;
+using VideoCommon::Shader::GetShaderCode;
+using VideoCommon::Shader::GetUniqueIdentifier;
+using VideoCommon::Shader::KERNEL_MAIN_OFFSET;
 using VideoCommon::Shader::ProgramCode;
 using VideoCommon::Shader::Registry;
 using VideoCommon::Shader::ShaderIR;
+using VideoCommon::Shader::STAGE_MAIN_OFFSET;
 
 namespace {
 
-constexpr u32 STAGE_MAIN_OFFSET = 10;
-constexpr u32 KERNEL_MAIN_OFFSET = 0;
-
 constexpr VideoCommon::Shader::CompilerSettings COMPILER_SETTINGS{};
 
-/// Gets the address for the specified shader stage program
-GPUVAddr GetShaderAddress(Core::System& system, Maxwell::ShaderProgram program) {
-    const auto& gpu{system.GPU().Maxwell3D()};
-    const auto& shader_config{gpu.regs.shader_config[static_cast<std::size_t>(program)]};
-    return gpu.regs.code_address.CodeAddress() + shader_config.offset;
-}
-
-/// Gets if the current instruction offset is a scheduler instruction
-constexpr bool IsSchedInstruction(std::size_t offset, std::size_t main_offset) {
-    // Sched instructions appear once every 4 instructions.
-    constexpr std::size_t SchedPeriod = 4;
-    const std::size_t absolute_offset = offset - main_offset;
-    return (absolute_offset % SchedPeriod) == 0;
-}
-
-/// Calculates the size of a program stream
-std::size_t CalculateProgramSize(const ProgramCode& program) {
-    constexpr std::size_t start_offset = 10;
-    // This is the encoded version of BRA that jumps to itself. All Nvidia
-    // shaders end with one.
-    constexpr u64 self_jumping_branch = 0xE2400FFFFF07000FULL;
-    constexpr u64 mask = 0xFFFFFFFFFF7FFFFFULL;
-    std::size_t offset = start_offset;
-    while (offset < program.size()) {
-        const u64 instruction = program[offset];
-        if (!IsSchedInstruction(offset, start_offset)) {
-            if ((instruction & mask) == self_jumping_branch) {
-                // End on Maxwell's "nop" instruction
-                break;
-            }
-            if (instruction == 0) {
-                break;
-            }
-        }
-        offset++;
-    }
-    // The last instruction is included in the program size
-    return std::min(offset + 1, program.size());
-}
-
-/// Gets the shader program code from memory for the specified address
-ProgramCode GetShaderCode(Tegra::MemoryManager& memory_manager, const GPUVAddr gpu_addr,
-                          const u8* host_ptr) {
-    ProgramCode code(VideoCommon::Shader::MAX_PROGRAM_LENGTH);
-    ASSERT_OR_EXECUTE(host_ptr != nullptr, {
-        std::fill(code.begin(), code.end(), 0);
-        return code;
-    });
-    memory_manager.ReadBlockUnsafe(gpu_addr, code.data(), code.size() * sizeof(u64));
-    code.resize(CalculateProgramSize(code));
-    return code;
-}
-
 /// Gets the shader type from a Maxwell program type
 constexpr GLenum GetGLShaderType(ShaderType shader_type) {
     switch (shader_type) {
@@ -114,17 +62,6 @@ constexpr GLenum GetGLShaderType(ShaderType shader_type) {
     }
 }
 
-/// Hashes one (or two) program streams
-u64 GetUniqueIdentifier(ShaderType shader_type, bool is_a, const ProgramCode& code,
-                        const ProgramCode& code_b = {}) {
-    u64 unique_identifier = boost::hash_value(code);
-    if (is_a) {
-        // VertexA programs include two programs
-        boost::hash_combine(unique_identifier, boost::hash_value(code_b));
-    }
-    return unique_identifier;
-}
-
 constexpr const char* GetShaderTypeName(ShaderType shader_type) {
     switch (shader_type) {
     case ShaderType::Vertex:
@@ -456,11 +393,12 @@ Shader ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program) {
     const auto host_ptr{memory_manager.GetPointer(address)};
 
     // No shader found - create a new one
-    ProgramCode code{GetShaderCode(memory_manager, address, host_ptr)};
+    ProgramCode code{GetShaderCode(memory_manager, address, host_ptr, false)};
     ProgramCode code_b;
     if (program == Maxwell::ShaderProgram::VertexA) {
         const GPUVAddr address_b{GetShaderAddress(system, Maxwell::ShaderProgram::VertexB)};
-        code_b = GetShaderCode(memory_manager, address_b, memory_manager.GetPointer(address_b));
+        const u8* host_ptr_b = memory_manager.GetPointer(address_b);
+        code_b = GetShaderCode(memory_manager, address_b, host_ptr_b, false);
     }
 
     const auto unique_identifier = GetUniqueIdentifier(
@@ -498,7 +436,7 @@ Shader ShaderCacheOpenGL::GetComputeKernel(GPUVAddr code_addr) {
 
     const auto host_ptr{memory_manager.GetPointer(code_addr)};
     // No kernel found, create a new one
-    auto code{GetShaderCode(memory_manager, code_addr, host_ptr)};
+    auto code{GetShaderCode(memory_manager, code_addr, host_ptr, true)};
     const auto unique_identifier{GetUniqueIdentifier(ShaderType::Compute, false, code)};
 
     const ShaderParameters params{system,    disk_cache, device,

src/video_core/renderer_opengl/gl_shader_decompiler.cpp

@@ -870,13 +870,13 @@ private:
         for (const auto& sampler : ir.GetSamplers()) {
             const std::string name = GetSampler(sampler);
             const std::string description = fmt::format("layout (binding = {}) uniform", binding);
-            binding += sampler.IsIndexed() ? sampler.Size() : 1;
+            binding += sampler.is_indexed ? sampler.size : 1;
 
             std::string sampler_type = [&]() {
-                if (sampler.IsBuffer()) {
+                if (sampler.is_buffer) {
                     return "samplerBuffer";
                 }
-                switch (sampler.GetType()) {
+                switch (sampler.type) {
                 case Tegra::Shader::TextureType::Texture1D:
                     return "sampler1D";
                 case Tegra::Shader::TextureType::Texture2D:
@@ -890,17 +890,17 @@ private:
                     return "sampler2D";
                 }
             }();
-            if (sampler.IsArray()) {
+            if (sampler.is_array) {
                 sampler_type += "Array";
             }
-            if (sampler.IsShadow()) {
+            if (sampler.is_shadow) {
                 sampler_type += "Shadow";
             }
 
-            if (!sampler.IsIndexed()) {
+            if (!sampler.is_indexed) {
                 code.AddLine("{} {} {};", description, sampler_type, name);
             } else {
-                code.AddLine("{} {} {}[{}];", description, sampler_type, name, sampler.Size());
+                code.AddLine("{} {} {}[{}];", description, sampler_type, name, sampler.size);
             }
         }
         if (!ir.GetSamplers().empty()) {
@@ -946,14 +946,14 @@ private:
         u32 binding = device.GetBaseBindings(stage).image;
         for (const auto& image : ir.GetImages()) {
             std::string qualifier = "coherent volatile";
-            if (image.IsRead() && !image.IsWritten()) {
+            if (image.is_read && !image.is_written) {
                 qualifier += " readonly";
-            } else if (image.IsWritten() && !image.IsRead()) {
+            } else if (image.is_written && !image.is_read) {
                 qualifier += " writeonly";
             }
 
-            const char* format = image.IsAtomic() ? "r32ui, " : "";
-            const char* type_declaration = GetImageTypeDeclaration(image.GetType());
+            const char* format = image.is_atomic ? "r32ui, " : "";
+            const char* type_declaration = GetImageTypeDeclaration(image.type);
             code.AddLine("layout ({}binding = {}) {} uniform uimage{} {};", format, binding++,
                          qualifier, type_declaration, GetImage(image));
         }
@@ -1337,8 +1337,8 @@ private:
         ASSERT(meta);
 
         const std::size_t count = operation.GetOperandsCount();
-        const bool has_array = meta->sampler.IsArray();
-        const bool has_shadow = meta->sampler.IsShadow();
+        const bool has_array = meta->sampler.is_array;
+        const bool has_shadow = meta->sampler.is_shadow;
 
         std::string expr = "texture" + function_suffix;
         if (!meta->aoffi.empty()) {
@@ -1346,7 +1346,7 @@ private:
         } else if (!meta->ptp.empty()) {
             expr += "Offsets";
         }
-        if (!meta->sampler.IsIndexed()) {
+        if (!meta->sampler.is_indexed) {
             expr += '(' + GetSampler(meta->sampler) + ", ";
         } else {
             expr += '(' + GetSampler(meta->sampler) + '[' + Visit(meta->index).AsUint() + "], ";
@@ -1870,6 +1870,14 @@ private:
         return GenerateBinaryInfix(operation, ">=", Type::Bool, type, type);
     }
 
+    Expression LogicalAddCarry(Operation operation) {
+        const std::string carry = code.GenerateTemporary();
+        code.AddLine("uint {};", carry);
+        code.AddLine("uaddCarry({}, {}, {});", VisitOperand(operation, 0).AsUint(),
+                     VisitOperand(operation, 1).AsUint(), carry);
+        return {fmt::format("({} != 0)", carry), Type::Bool};
+    }
+
     Expression LogicalFIsNan(Operation operation) {
         return GenerateUnary(operation, "isnan", Type::Bool, Type::Float);
     }
@@ -1974,7 +1982,7 @@ private:
 
         std::string expr = GenerateTexture(
             operation, "", {TextureOffset{}, TextureArgument{Type::Float, meta->bias}});
-        if (meta->sampler.IsShadow()) {
+        if (meta->sampler.is_shadow) {
             expr = "vec4(" + expr + ')';
         }
         return {expr + GetSwizzle(meta->element), Type::Float};
@@ -1986,7 +1994,7 @@ private:
 
         std::string expr = GenerateTexture(
             operation, "Lod", {TextureArgument{Type::Float, meta->lod}, TextureOffset{}});
-        if (meta->sampler.IsShadow()) {
+        if (meta->sampler.is_shadow) {
             expr = "vec4(" + expr + ')';
         }
         return {expr + GetSwizzle(meta->element), Type::Float};
@@ -1995,11 +2003,11 @@ private:
     Expression TextureGather(Operation operation) {
         const auto& meta = std::get<MetaTexture>(operation.GetMeta());
 
-        const auto type = meta.sampler.IsShadow() ? Type::Float : Type::Int;
-        const bool separate_dc = meta.sampler.IsShadow();
+        const auto type = meta.sampler.is_shadow ? Type::Float : Type::Int;
+        const bool separate_dc = meta.sampler.is_shadow;
 
         std::vector<TextureIR> ir;
-        if (meta.sampler.IsShadow()) {
+        if (meta.sampler.is_shadow) {
             ir = {TextureOffset{}};
         } else {
             ir = {TextureOffset{}, TextureArgument{type, meta.component}};
@@ -2044,7 +2052,7 @@ private:
         constexpr std::array constructors = {"int", "ivec2", "ivec3", "ivec4"};
         const auto meta = std::get_if<MetaTexture>(&operation.GetMeta());
         ASSERT(meta);
-        UNIMPLEMENTED_IF(meta->sampler.IsArray());
+        UNIMPLEMENTED_IF(meta->sampler.is_array);
         const std::size_t count = operation.GetOperandsCount();
 
         std::string expr = "texelFetch(";
@@ -2065,7 +2073,7 @@ private:
         }
         expr += ')';
 
-        if (meta->lod && !meta->sampler.IsBuffer()) {
+        if (meta->lod && !meta->sampler.is_buffer) {
             expr += ", ";
             expr += Visit(meta->lod).AsInt();
         }
@@ -2076,12 +2084,10 @@ private:
     }
 
     Expression TextureGradient(Operation operation) {
-        const auto meta = std::get_if<MetaTexture>(&operation.GetMeta());
-        ASSERT(meta);
-
+        const auto& meta = std::get<MetaTexture>(operation.GetMeta());
         std::string expr =
             GenerateTexture(operation, "Grad", {TextureDerivates{}, TextureOffset{}});
-        return {std::move(expr) + GetSwizzle(meta->element), Type::Float};
+        return {std::move(expr) + GetSwizzle(meta.element), Type::Float};
     }
 
     Expression ImageLoad(Operation operation) {
@@ -2441,6 +2447,8 @@ private:
         &GLSLDecompiler::LogicalNotEqual<Type::Uint>,
         &GLSLDecompiler::LogicalGreaterEqual<Type::Uint>,
 
+        &GLSLDecompiler::LogicalAddCarry,
+
         &GLSLDecompiler::Logical2HLessThan<false>,
         &GLSLDecompiler::Logical2HEqual<false>,
         &GLSLDecompiler::Logical2HLessEqual<false>,
@@ -2598,11 +2606,11 @@ private:
     }
 
     std::string GetSampler(const Sampler& sampler) const {
-        return AppendSuffix(static_cast<u32>(sampler.GetIndex()), "sampler");
+        return AppendSuffix(sampler.index, "sampler");
     }
 
     std::string GetImage(const Image& image) const {
-        return AppendSuffix(static_cast<u32>(image.GetIndex()), "image");
+        return AppendSuffix(image.index, "image");
     }
 
     std::string AppendSuffix(u32 index, std::string_view name) const {

src/video_core/renderer_opengl/gl_shader_decompiler.h

@@ -33,36 +33,19 @@ public:
     }
 
 private:
-    u32 index{};
+    u32 index = 0;
 };
 
-class GlobalMemoryEntry {
-public:
-    explicit GlobalMemoryEntry(u32 cbuf_index, u32 cbuf_offset, bool is_read, bool is_written)
+struct GlobalMemoryEntry {
+    constexpr explicit GlobalMemoryEntry(u32 cbuf_index, u32 cbuf_offset, bool is_read,
+                                         bool is_written)
         : cbuf_index{cbuf_index}, cbuf_offset{cbuf_offset}, is_read{is_read}, is_written{
                                                                                   is_written} {}
 
-    u32 GetCbufIndex() const {
-        return cbuf_index;
-    }
-
-    u32 GetCbufOffset() const {
-        return cbuf_offset;
-    }
-
-    bool IsRead() const {
-        return is_read;
-    }
-
-    bool IsWritten() const {
-        return is_written;
-    }
-
-private:
-    u32 cbuf_index{};
-    u32 cbuf_offset{};
-    bool is_read{};
-    bool is_written{};
+    u32 cbuf_index = 0;
+    u32 cbuf_offset = 0;
+    bool is_read = false;
+    bool is_written = false;
 };
 
 struct ShaderEntries {

src/video_core/renderer_vulkan/fixed_pipeline_state.cpp

@@ -81,7 +81,7 @@ void FixedPipelineState::Rasterizer::Fill(const Maxwell& regs) noexcept {
     primitive_restart_enable.Assign(regs.primitive_restart.enabled != 0 ? 1 : 0);
     cull_enable.Assign(regs.cull_test_enabled != 0 ? 1 : 0);
     depth_bias_enable.Assign(enabled_lut[POLYGON_OFFSET_ENABLE_LUT[topology_index]] != 0 ? 1 : 0);
-    depth_clamp_enable.Assign(clip.depth_clamp_near == 1 || clip.depth_clamp_far == 1 ? 1 : 0);
+    depth_clamp_disabled.Assign(regs.view_volume_clip_control.depth_clamp_disabled.Value());
     ndc_minus_one_to_one.Assign(regs.depth_mode == Maxwell::DepthMode::MinusOneToOne ? 1 : 0);
     cull_face.Assign(PackCullFace(regs.cull_face));
     front_face.Assign(packed_front_face);
@@ -140,6 +140,12 @@ void FixedPipelineState::BlendingAttachment::Fill(const Maxwell& regs, std::size
     enable.Assign(1);
 }
 
+void FixedPipelineState::Fill(const Maxwell& regs) {
+    rasterizer.Fill(regs);
+    depth_stencil.Fill(regs);
+    color_blending.Fill(regs);
+}
+
 std::size_t FixedPipelineState::Hash() const noexcept {
     const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), sizeof *this);
     return static_cast<std::size_t>(hash);
@@ -149,15 +155,6 @@ bool FixedPipelineState::operator==(const FixedPipelineState& rhs) const noexcep
     return std::memcmp(this, &rhs, sizeof *this) == 0;
 }
 
-FixedPipelineState GetFixedPipelineState(const Maxwell& regs) {
-    FixedPipelineState fixed_state;
-    fixed_state.rasterizer.Fill(regs);
-    fixed_state.depth_stencil.Fill(regs);
-    fixed_state.color_blending.Fill(regs);
-    fixed_state.padding = {};
-    return fixed_state;
-}
-
 u32 FixedPipelineState::PackComparisonOp(Maxwell::ComparisonOp op) noexcept {
     // OpenGL enums go from 0x200 to 0x207 and the others from 1 to 8
     // If we substract 0x200 to OpenGL enums and 1 to the others we get a 0-7 range.

src/video_core/renderer_vulkan/fixed_pipeline_state.h

@@ -17,7 +17,7 @@ namespace Vulkan {
 
 using Maxwell = Tegra::Engines::Maxwell3D::Regs;
 
-struct alignas(32) FixedPipelineState {
+struct FixedPipelineState {
     static u32 PackComparisonOp(Maxwell::ComparisonOp op) noexcept;
     static Maxwell::ComparisonOp UnpackComparisonOp(u32 packed) noexcept;
 
@@ -153,7 +153,7 @@ struct alignas(32) FixedPipelineState {
             BitField<4, 1, u32> primitive_restart_enable;
             BitField<5, 1, u32> cull_enable;
             BitField<6, 1, u32> depth_bias_enable;
-            BitField<7, 1, u32> depth_clamp_enable;
+            BitField<7, 1, u32> depth_clamp_disabled;
             BitField<8, 1, u32> ndc_minus_one_to_one;
             BitField<9, 2, u32> cull_face;
             BitField<11, 1, u32> front_face;
@@ -237,7 +237,8 @@ struct alignas(32) FixedPipelineState {
     Rasterizer rasterizer;
     DepthStencil depth_stencil;
     ColorBlending color_blending;
-    std::array<u8, 20> padding;
+
+    void Fill(const Maxwell& regs);
 
     std::size_t Hash() const noexcept;
 
@@ -250,9 +251,6 @@ struct alignas(32) FixedPipelineState {
 static_assert(std::has_unique_object_representations_v<FixedPipelineState>);
 static_assert(std::is_trivially_copyable_v<FixedPipelineState>);
 static_assert(std::is_trivially_constructible_v<FixedPipelineState>);
-static_assert(sizeof(FixedPipelineState) % 32 == 0, "Size is not aligned");
-
-FixedPipelineState GetFixedPipelineState(const Maxwell& regs);
 
 } // namespace Vulkan
 

src/video_core/renderer_vulkan/renderer_vulkan.cpp

@@ -12,15 +12,12 @@
 
 #include <fmt/format.h>
 
-#include "common/assert.h"
 #include "common/dynamic_library.h"
 #include "common/logging/log.h"
 #include "common/telemetry.h"
 #include "core/core.h"
 #include "core/core_timing.h"
 #include "core/frontend/emu_window.h"
-#include "core/memory.h"
-#include "core/perf_stats.h"
 #include "core/settings.h"
 #include "core/telemetry_session.h"
 #include "video_core/gpu.h"

src/video_core/renderer_vulkan/renderer_vulkan.h

@@ -5,7 +5,6 @@
 #pragma once
 
 #include <memory>
-#include <optional>
 #include <string>
 #include <vector>
 

src/video_core/renderer_vulkan/vk_blit_screen.h

@@ -4,7 +4,6 @@
 
 #pragma once
 
-#include <array>
 #include <memory>
 #include <tuple>
 

src/video_core/renderer_vulkan/vk_buffer_cache.cpp

@@ -5,11 +5,7 @@
 #include <algorithm>
 #include <cstring>
 #include <memory>
-#include <optional>
-#include <tuple>
 
-#include "common/assert.h"
-#include "common/bit_util.h"
 #include "core/core.h"
 #include "video_core/renderer_vulkan/vk_buffer_cache.h"
 #include "video_core/renderer_vulkan/vk_device.h"

src/video_core/renderer_vulkan/vk_buffer_cache.h

@@ -5,14 +5,11 @@
 #pragma once
 
 #include <memory>
-#include <unordered_map>
-#include <vector>
 
 #include "common/common_types.h"
 #include "video_core/buffer_cache/buffer_cache.h"
 #include "video_core/rasterizer_cache.h"
 #include "video_core/renderer_vulkan/vk_memory_manager.h"
-#include "video_core/renderer_vulkan/vk_resource_manager.h"
 #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h"
 #include "video_core/renderer_vulkan/vk_stream_buffer.h"
 #include "video_core/renderer_vulkan/wrapper.h"

src/video_core/renderer_vulkan/vk_compute_pass.cpp

@@ -6,7 +6,7 @@
 #include <memory>
 #include <optional>
 #include <utility>
-#include <vector>
+
 #include "common/alignment.h"
 #include "common/assert.h"
 #include "common/common_types.h"

src/video_core/renderer_vulkan/vk_compute_pass.h

@@ -6,7 +6,7 @@
 
 #include <optional>
 #include <utility>
-#include <vector>
+
 #include "common/common_types.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_vulkan/vk_descriptor_pool.h"

src/video_core/renderer_vulkan/vk_compute_pipeline.cpp

@@ -2,14 +2,12 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <memory>
 #include <vector>
 
 #include "video_core/renderer_vulkan/vk_compute_pipeline.h"
 #include "video_core/renderer_vulkan/vk_descriptor_pool.h"
 #include "video_core/renderer_vulkan/vk_device.h"
 #include "video_core/renderer_vulkan/vk_pipeline_cache.h"
-#include "video_core/renderer_vulkan/vk_resource_manager.h"
 #include "video_core/renderer_vulkan/vk_scheduler.h"
 #include "video_core/renderer_vulkan/vk_shader_decompiler.h"
 #include "video_core/renderer_vulkan/vk_update_descriptor.h"

src/video_core/renderer_vulkan/vk_compute_pipeline.h

@@ -4,8 +4,6 @@
 
 #pragma once
 
-#include <memory>
-
 #include "common/common_types.h"
 #include "video_core/renderer_vulkan/vk_descriptor_pool.h"
 #include "video_core/renderer_vulkan/vk_shader_decompiler.h"

src/video_core/renderer_vulkan/vk_descriptor_pool.cpp

@@ -2,7 +2,6 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <memory>
 #include <vector>
 
 #include "common/common_types.h"

src/video_core/renderer_vulkan/vk_descriptor_pool.h

@@ -4,10 +4,8 @@
 
 #pragma once
 
-#include <memory>
 #include <vector>
 
-#include "common/common_types.h"
 #include "video_core/renderer_vulkan/vk_resource_manager.h"
 #include "video_core/renderer_vulkan/wrapper.h"
 

src/video_core/renderer_vulkan/vk_device.cpp

@@ -4,7 +4,6 @@
 
 #include <bitset>
 #include <chrono>
-#include <cstdlib>
 #include <optional>
 #include <string_view>
 #include <thread>

src/video_core/renderer_vulkan/vk_device.h

@@ -82,11 +82,6 @@ public:
         return present_family;
     }
 
-    /// Returns true if the device is integrated with the host CPU.
-    bool IsIntegrated() const {
-        return properties.deviceType == VK_PHYSICAL_DEVICE_TYPE_INTEGRATED_GPU;
-    }
-
     /// Returns the current Vulkan API version provided in Vulkan-formatted version numbers.
     u32 GetApiVersion() const {
         return properties.apiVersion;

src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp

@@ -6,7 +6,6 @@
 #include <cstring>
 #include <vector>
 
-#include "common/assert.h"
 #include "common/common_types.h"
 #include "common/microprofile.h"
 #include "video_core/renderer_vulkan/fixed_pipeline_state.h"
@@ -249,7 +248,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa
     rasterization_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_STATE_CREATE_INFO;
     rasterization_ci.pNext = nullptr;
     rasterization_ci.flags = 0;
-    rasterization_ci.depthClampEnable = rs.depth_clamp_enable;
+    rasterization_ci.depthClampEnable = rs.depth_clamp_disabled == 0 ? VK_TRUE : VK_FALSE;
     rasterization_ci.rasterizerDiscardEnable = VK_FALSE;
     rasterization_ci.polygonMode = VK_POLYGON_MODE_FILL;
     rasterization_ci.cullMode =
@@ -288,7 +287,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa
     depth_stencil_ci.maxDepthBounds = 0.0f;
 
     std::array<VkPipelineColorBlendAttachmentState, Maxwell::NumRenderTargets> cb_attachments;
-    const std::size_t num_attachments = renderpass_params.color_attachments.size();
+    const auto num_attachments = static_cast<std::size_t>(renderpass_params.num_color_attachments);
     for (std::size_t index = 0; index < num_attachments; ++index) {
         static constexpr std::array COMPONENT_TABLE = {
             VK_COLOR_COMPONENT_R_BIT, VK_COLOR_COMPONENT_G_BIT, VK_COLOR_COMPONENT_B_BIT,

src/video_core/renderer_vulkan/vk_graphics_pipeline.h

@@ -5,16 +5,13 @@
 #pragma once
 
 #include <array>
-#include <memory>
 #include <optional>
-#include <unordered_map>
 #include <vector>
 
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_vulkan/fixed_pipeline_state.h"
 #include "video_core/renderer_vulkan/vk_descriptor_pool.h"
 #include "video_core/renderer_vulkan/vk_renderpass_cache.h"
-#include "video_core/renderer_vulkan/vk_resource_manager.h"
 #include "video_core/renderer_vulkan/vk_shader_decompiler.h"
 #include "video_core/renderer_vulkan/wrapper.h"
 

src/video_core/renderer_vulkan/vk_memory_manager.cpp

@@ -118,8 +118,7 @@ private:
 };
 
 VKMemoryManager::VKMemoryManager(const VKDevice& device)
-    : device{device}, properties{device.GetPhysical().GetMemoryProperties()},
-      is_memory_unified{GetMemoryUnified(properties)} {}
+    : device{device}, properties{device.GetPhysical().GetMemoryProperties()} {}
 
 VKMemoryManager::~VKMemoryManager() = default;
 
@@ -209,16 +208,6 @@ VKMemoryCommit VKMemoryManager::TryAllocCommit(const VkMemoryRequirements& requi
     return {};
 }
 
-bool VKMemoryManager::GetMemoryUnified(const VkPhysicalDeviceMemoryProperties& properties) {
-    for (u32 heap_index = 0; heap_index < properties.memoryHeapCount; ++heap_index) {
-        if (!(properties.memoryHeaps[heap_index].flags & VK_MEMORY_HEAP_DEVICE_LOCAL_BIT)) {
-            // Memory is considered unified when heaps are device local only.
-            return false;
-        }
-    }
-    return true;
-}
-
 VKMemoryCommitImpl::VKMemoryCommitImpl(const VKDevice& device, VKMemoryAllocation* allocation,
                                        const vk::DeviceMemory& memory, u64 begin, u64 end)
     : device{device}, memory{memory}, interval{begin, end}, allocation{allocation} {}

src/video_core/renderer_vulkan/vk_memory_manager.h

@@ -40,11 +40,6 @@ public:
     /// Commits memory required by the image and binds it.
     VKMemoryCommit Commit(const vk::Image& image, bool host_visible);
 
-    /// Returns true if the memory allocations are done always in host visible and coherent memory.
-    bool IsMemoryUnified() const {
-        return is_memory_unified;
-    }
-
 private:
     /// Allocates a chunk of memory.
     bool AllocMemory(VkMemoryPropertyFlags wanted_properties, u32 type_mask, u64 size);
@@ -53,12 +48,8 @@ private:
     VKMemoryCommit TryAllocCommit(const VkMemoryRequirements& requirements,
                                   VkMemoryPropertyFlags wanted_properties);
 
-    /// Returns true if the device uses an unified memory model.
-    static bool GetMemoryUnified(const VkPhysicalDeviceMemoryProperties& properties);
-
-    const VKDevice& device;                            ///< Device handler.
-    const VkPhysicalDeviceMemoryProperties properties; ///< Physical device properties.
-    const bool is_memory_unified;                      ///< True if memory model is unified.
+    const VKDevice& device;                                       ///< Device handler.
+    const VkPhysicalDeviceMemoryProperties properties;            ///< Physical device properties.
     std::vector<std::unique_ptr<VKMemoryAllocation>> allocations; ///< Current allocations.
 };
 

src/video_core/renderer_vulkan/vk_pipeline_cache.cpp

@@ -22,17 +22,22 @@
 #include "video_core/renderer_vulkan/vk_pipeline_cache.h"
 #include "video_core/renderer_vulkan/vk_rasterizer.h"
 #include "video_core/renderer_vulkan/vk_renderpass_cache.h"
-#include "video_core/renderer_vulkan/vk_resource_manager.h"
 #include "video_core/renderer_vulkan/vk_scheduler.h"
 #include "video_core/renderer_vulkan/vk_update_descriptor.h"
 #include "video_core/renderer_vulkan/wrapper.h"
 #include "video_core/shader/compiler_settings.h"
+#include "video_core/shader/memory_util.h"
 
 namespace Vulkan {
 
 MICROPROFILE_DECLARE(Vulkan_PipelineCache);
 
 using Tegra::Engines::ShaderType;
+using VideoCommon::Shader::GetShaderAddress;
+using VideoCommon::Shader::GetShaderCode;
+using VideoCommon::Shader::KERNEL_MAIN_OFFSET;
+using VideoCommon::Shader::ProgramCode;
+using VideoCommon::Shader::STAGE_MAIN_OFFSET;
 
 namespace {
 
@@ -45,60 +50,6 @@ constexpr VkDescriptorType STORAGE_IMAGE = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE;
 constexpr VideoCommon::Shader::CompilerSettings compiler_settings{
     VideoCommon::Shader::CompileDepth::FullDecompile};
 
-/// Gets the address for the specified shader stage program
-GPUVAddr GetShaderAddress(Core::System& system, Maxwell::ShaderProgram program) {
-    const auto& gpu{system.GPU().Maxwell3D()};
-    const auto& shader_config{gpu.regs.shader_config[static_cast<std::size_t>(program)]};
-    return gpu.regs.code_address.CodeAddress() + shader_config.offset;
-}
-
-/// Gets if the current instruction offset is a scheduler instruction
-constexpr bool IsSchedInstruction(std::size_t offset, std::size_t main_offset) {
-    // Sched instructions appear once every 4 instructions.
-    constexpr std::size_t SchedPeriod = 4;
-    const std::size_t absolute_offset = offset - main_offset;
-    return (absolute_offset % SchedPeriod) == 0;
-}
-
-/// Calculates the size of a program stream
-std::size_t CalculateProgramSize(const ProgramCode& program, bool is_compute) {
-    const std::size_t start_offset = is_compute ? 0 : 10;
-    // This is the encoded version of BRA that jumps to itself. All Nvidia
-    // shaders end with one.
-    constexpr u64 self_jumping_branch = 0xE2400FFFFF07000FULL;
-    constexpr u64 mask = 0xFFFFFFFFFF7FFFFFULL;
-    std::size_t offset = start_offset;
-    while (offset < program.size()) {
-        const u64 instruction = program[offset];
-        if (!IsSchedInstruction(offset, start_offset)) {
-            if ((instruction & mask) == self_jumping_branch) {
-                // End on Maxwell's "nop" instruction
-                break;
-            }
-            if (instruction == 0) {
-                break;
-            }
-        }
-        ++offset;
-    }
-    // The last instruction is included in the program size
-    return std::min(offset + 1, program.size());
-}
-
-/// Gets the shader program code from memory for the specified address
-ProgramCode GetShaderCode(Tegra::MemoryManager& memory_manager, const GPUVAddr gpu_addr,
-                          const u8* host_ptr, bool is_compute) {
-    ProgramCode program_code(VideoCommon::Shader::MAX_PROGRAM_LENGTH);
-    ASSERT_OR_EXECUTE(host_ptr != nullptr, {
-        std::fill(program_code.begin(), program_code.end(), 0);
-        return program_code;
-    });
-    memory_manager.ReadBlockUnsafe(gpu_addr, program_code.data(),
-                                   program_code.size() * sizeof(u64));
-    program_code.resize(CalculateProgramSize(program_code, is_compute));
-    return program_code;
-}
-
 constexpr std::size_t GetStageFromProgram(std::size_t program) {
     return program == 0 ? 0 : program - 1;
 }
@@ -133,7 +84,7 @@ void AddBindings(std::vector<VkDescriptorSetLayoutBinding>& bindings, u32& bindi
         u32 count = 1;
         if constexpr (descriptor_type == VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER) {
             // Combined image samplers can be arrayed.
-            count = container[i].Size();
+            count = container[i].size;
         }
         VkDescriptorSetLayoutBinding& entry = bindings.emplace_back();
         entry.binding = binding++;
@@ -161,6 +112,24 @@ u32 FillDescriptorLayout(const ShaderEntries& entries,
 
 } // Anonymous namespace
 
+std::size_t GraphicsPipelineCacheKey::Hash() const noexcept {
+    const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), sizeof *this);
+    return static_cast<std::size_t>(hash);
+}
+
+bool GraphicsPipelineCacheKey::operator==(const GraphicsPipelineCacheKey& rhs) const noexcept {
+    return std::memcmp(&rhs, this, sizeof *this) == 0;
+}
+
+std::size_t ComputePipelineCacheKey::Hash() const noexcept {
+    const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), sizeof *this);
+    return static_cast<std::size_t>(hash);
+}
+
+bool ComputePipelineCacheKey::operator==(const ComputePipelineCacheKey& rhs) const noexcept {
+    return std::memcmp(&rhs, this, sizeof *this) == 0;
+}
+
 CachedShader::CachedShader(Core::System& system, Tegra::Engines::ShaderType stage,
                            GPUVAddr gpu_addr, VAddr cpu_addr, ProgramCode program_code,
                            u32 main_offset)
@@ -212,9 +181,9 @@ std::array<Shader, Maxwell::MaxShaderProgram> VKPipelineCache::GetShaders() {
             const auto host_ptr{memory_manager.GetPointer(program_addr)};
 
             // No shader found - create a new one
-            constexpr u32 stage_offset = 10;
+            constexpr u32 stage_offset = STAGE_MAIN_OFFSET;
             const auto stage = static_cast<Tegra::Engines::ShaderType>(index == 0 ? 0 : index - 1);
-            auto code = GetShaderCode(memory_manager, program_addr, host_ptr, false);
+            ProgramCode code = GetShaderCode(memory_manager, program_addr, host_ptr, false);
 
             shader = std::make_shared<CachedShader>(system, stage, program_addr, *cpu_addr,
                                                     std::move(code), stage_offset);
@@ -270,11 +239,10 @@ VKComputePipeline& VKPipelineCache::GetComputePipeline(const ComputePipelineCach
         // No shader found - create a new one
         const auto host_ptr = memory_manager.GetPointer(program_addr);
 
-        auto code = GetShaderCode(memory_manager, program_addr, host_ptr, true);
-        constexpr u32 kernel_main_offset = 0;
+        ProgramCode code = GetShaderCode(memory_manager, program_addr, host_ptr, true);
         shader = std::make_shared<CachedShader>(system, Tegra::Engines::ShaderType::Compute,
                                                 program_addr, *cpu_addr, std::move(code),
-                                                kernel_main_offset);
+                                                KERNEL_MAIN_OFFSET);
         if (cpu_addr) {
             Register(shader);
         } else {
@@ -393,7 +361,7 @@ void AddEntry(std::vector<VkDescriptorUpdateTemplateEntry>& template_entries, u3
 
     if constexpr (descriptor_type == COMBINED_IMAGE_SAMPLER) {
         for (u32 i = 0; i < count; ++i) {
-            const u32 num_samplers = container[i].Size();
+            const u32 num_samplers = container[i].size;
             VkDescriptorUpdateTemplateEntry& entry = template_entries.emplace_back();
             entry.dstBinding = binding;
             entry.dstArrayElement = 0;

src/video_core/renderer_vulkan/vk_pipeline_cache.h

@@ -7,7 +7,6 @@
 #include <array>
 #include <cstddef>
 #include <memory>
-#include <tuple>
 #include <type_traits>
 #include <unordered_map>
 #include <utility>
@@ -22,12 +21,11 @@
 #include "video_core/renderer_vulkan/fixed_pipeline_state.h"
 #include "video_core/renderer_vulkan/vk_graphics_pipeline.h"
 #include "video_core/renderer_vulkan/vk_renderpass_cache.h"
-#include "video_core/renderer_vulkan/vk_resource_manager.h"
 #include "video_core/renderer_vulkan/vk_shader_decompiler.h"
 #include "video_core/renderer_vulkan/wrapper.h"
+#include "video_core/shader/memory_util.h"
 #include "video_core/shader/registry.h"
 #include "video_core/shader/shader_ir.h"
-#include "video_core/surface.h"
 
 namespace Core {
 class System;
@@ -47,46 +45,40 @@ class CachedShader;
 using Shader = std::shared_ptr<CachedShader>;
 using Maxwell = Tegra::Engines::Maxwell3D::Regs;
 
-using ProgramCode = std::vector<u64>;
-
 struct GraphicsPipelineCacheKey {
     FixedPipelineState fixed_state;
-    std::array<GPUVAddr, Maxwell::MaxShaderProgram> shaders;
     RenderPassParams renderpass_params;
+    std::array<GPUVAddr, Maxwell::MaxShaderProgram> shaders;
+    u64 padding; // This is necessary for unique object representations
 
-    std::size_t Hash() const noexcept {
-        std::size_t hash = fixed_state.Hash();
-        for (const auto& shader : shaders) {
-            boost::hash_combine(hash, shader);
-        }
-        boost::hash_combine(hash, renderpass_params.Hash());
-        return hash;
-    }
+    std::size_t Hash() const noexcept;
 
-    bool operator==(const GraphicsPipelineCacheKey& rhs) const noexcept {
-        return std::tie(fixed_state, shaders, renderpass_params) ==
-               std::tie(rhs.fixed_state, rhs.shaders, rhs.renderpass_params);
+    bool operator==(const GraphicsPipelineCacheKey& rhs) const noexcept;
+
+    bool operator!=(const GraphicsPipelineCacheKey& rhs) const noexcept {
+        return !operator==(rhs);
     }
 };
+static_assert(std::has_unique_object_representations_v<GraphicsPipelineCacheKey>);
+static_assert(std::is_trivially_copyable_v<GraphicsPipelineCacheKey>);
+static_assert(std::is_trivially_constructible_v<GraphicsPipelineCacheKey>);
 
 struct ComputePipelineCacheKey {
-    GPUVAddr shader{};
-    u32 shared_memory_size{};
-    std::array<u32, 3> workgroup_size{};
+    GPUVAddr shader;
+    u32 shared_memory_size;
+    std::array<u32, 3> workgroup_size;
 
-    std::size_t Hash() const noexcept {
-        return static_cast<std::size_t>(shader) ^
-               ((static_cast<std::size_t>(shared_memory_size) >> 7) << 40) ^
-               static_cast<std::size_t>(workgroup_size[0]) ^
-               (static_cast<std::size_t>(workgroup_size[1]) << 16) ^
-               (static_cast<std::size_t>(workgroup_size[2]) << 24);
-    }
+    std::size_t Hash() const noexcept;
 
-    bool operator==(const ComputePipelineCacheKey& rhs) const noexcept {
-        return std::tie(shader, shared_memory_size, workgroup_size) ==
-               std::tie(rhs.shader, rhs.shared_memory_size, rhs.workgroup_size);
+    bool operator==(const ComputePipelineCacheKey& rhs) const noexcept;
+
+    bool operator!=(const ComputePipelineCacheKey& rhs) const noexcept {
+        return !operator==(rhs);
     }
 };
+static_assert(std::has_unique_object_representations_v<ComputePipelineCacheKey>);
+static_assert(std::is_trivially_copyable_v<ComputePipelineCacheKey>);
+static_assert(std::is_trivially_constructible_v<ComputePipelineCacheKey>);
 
 } // namespace Vulkan
 
@@ -113,7 +105,8 @@ namespace Vulkan {
 class CachedShader final : public RasterizerCacheObject {
 public:
     explicit CachedShader(Core::System& system, Tegra::Engines::ShaderType stage, GPUVAddr gpu_addr,
-                          VAddr cpu_addr, ProgramCode program_code, u32 main_offset);
+                          VAddr cpu_addr, VideoCommon::Shader::ProgramCode program_code,
+                          u32 main_offset);
     ~CachedShader();
 
     GPUVAddr GetGpuAddr() const {
@@ -145,7 +138,7 @@ private:
                                                                  Tegra::Engines::ShaderType stage);
 
     GPUVAddr gpu_addr{};
-    ProgramCode program_code;
+    VideoCommon::Shader::ProgramCode program_code;
     VideoCommon::Shader::Registry registry;
     VideoCommon::Shader::ShaderIR shader_ir;
     ShaderEntries entries;

src/video_core/renderer_vulkan/vk_query_cache.cpp

@@ -4,7 +4,6 @@
 
 #include <algorithm>
 #include <cstddef>
-#include <cstdint>
 #include <utility>
 #include <vector>
 

src/video_core/renderer_vulkan/vk_query_cache.h

@@ -5,7 +5,6 @@
 #pragma once
 
 #include <cstddef>
-#include <cstdint>
 #include <memory>
 #include <utility>
 #include <vector>

src/video_core/renderer_vulkan/vk_rasterizer.cpp

@@ -9,14 +9,12 @@
 #include <vector>
 
 #include <boost/container/static_vector.hpp>
-#include <boost/functional/hash.hpp>
 
 #include "common/alignment.h"
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "common/microprofile.h"
 #include "core/core.h"
-#include "core/memory.h"
 #include "core/settings.h"
 #include "video_core/engines/kepler_compute.h"
 #include "video_core/engines/maxwell_3d.h"
@@ -119,14 +117,13 @@ template <typename Engine, typename Entry>
 Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry& entry,
                                                std::size_t stage, std::size_t index = 0) {
     const auto stage_type = static_cast<Tegra::Engines::ShaderType>(stage);
-    if (entry.IsBindless()) {
-        const Tegra::Texture::TextureHandle tex_handle =
-            engine.AccessConstBuffer32(stage_type, entry.GetBuffer(), entry.GetOffset());
+    if (entry.is_bindless) {
+        const auto tex_handle = engine.AccessConstBuffer32(stage_type, entry.buffer, entry.offset);
         return engine.GetTextureInfo(tex_handle);
     }
     const auto& gpu_profile = engine.AccessGuestDriverProfile();
     const u32 entry_offset = static_cast<u32>(index * gpu_profile.GetTextureHandlerSize());
-    const u32 offset = entry.GetOffset() + entry_offset;
+    const u32 offset = entry.offset + entry_offset;
     if constexpr (std::is_same_v<Engine, Tegra::Engines::Maxwell3D>) {
         return engine.GetStageTexture(stage_type, offset);
     } else {
@@ -316,7 +313,8 @@ void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) {
     query_cache.UpdateCounters();
 
     const auto& gpu = system.GPU().Maxwell3D();
-    GraphicsPipelineCacheKey key{GetFixedPipelineState(gpu.regs)};
+    GraphicsPipelineCacheKey key;
+    key.fixed_state.Fill(gpu.regs);
 
     buffer_cache.Map(CalculateGraphicsStreamBufferSize(is_indexed));
 
@@ -334,10 +332,11 @@ void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) {
 
     buffer_cache.Unmap();
 
-    const auto texceptions = UpdateAttachments();
+    const Texceptions texceptions = UpdateAttachments();
     SetupImageTransitions(texceptions, color_attachments, zeta_attachment);
 
     key.renderpass_params = GetRenderPassParams(texceptions);
+    key.padding = 0;
 
     auto& pipeline = pipeline_cache.GetGraphicsPipeline(key);
     scheduler.BindGraphicsPipeline(pipeline.GetHandle());
@@ -453,10 +452,12 @@ void RasterizerVulkan::DispatchCompute(GPUVAddr code_addr) {
     query_cache.UpdateCounters();
 
     const auto& launch_desc = system.GPU().KeplerCompute().launch_description;
-    const ComputePipelineCacheKey key{
-        code_addr,
-        launch_desc.shared_alloc,
-        {launch_desc.block_dim_x, launch_desc.block_dim_y, launch_desc.block_dim_z}};
+    ComputePipelineCacheKey key;
+    key.shader = code_addr;
+    key.shared_memory_size = launch_desc.shared_alloc;
+    key.workgroup_size = {launch_desc.block_dim_x, launch_desc.block_dim_y,
+                          launch_desc.block_dim_z};
+
     auto& pipeline = pipeline_cache.GetComputePipeline(key);
 
     // Compute dispatches can't be executed inside a renderpass
@@ -688,7 +689,7 @@ std::tuple<VkFramebuffer, VkExtent2D> RasterizerVulkan::ConfigureFramebuffers(
     FramebufferCacheKey key{renderpass, std::numeric_limits<u32>::max(),
                             std::numeric_limits<u32>::max(), std::numeric_limits<u32>::max()};
 
-    const auto try_push = [&](const View& view) {
+    const auto try_push = [&key](const View& view) {
         if (!view) {
             return false;
         }
@@ -699,7 +700,9 @@ std::tuple<VkFramebuffer, VkExtent2D> RasterizerVulkan::ConfigureFramebuffers(
         return true;
     };
 
-    for (std::size_t index = 0; index < std::size(color_attachments); ++index) {
+    const auto& regs = system.GPU().Maxwell3D().regs;
+    const std::size_t num_attachments = static_cast<std::size_t>(regs.rt_control.count);
+    for (std::size_t index = 0; index < num_attachments; ++index) {
         if (try_push(color_attachments[index])) {
             texture_cache.MarkColorBufferInUse(index);
         }
@@ -890,6 +893,9 @@ void RasterizerVulkan::SetupVertexArrays(FixedPipelineState::VertexInput& vertex
 
 void RasterizerVulkan::SetupIndexBuffer(BufferBindings& buffer_bindings, DrawParameters& params,
                                         bool is_indexed) {
+    if (params.num_vertices == 0) {
+        return;
+    }
     const auto& regs = system.GPU().Maxwell3D().regs;
     switch (regs.draw.topology) {
     case Maxwell::PrimitiveTopology::Quads: {
@@ -965,7 +971,7 @@ void RasterizerVulkan::SetupGraphicsTextures(const ShaderEntries& entries, std::
     MICROPROFILE_SCOPE(Vulkan_Textures);
     const auto& gpu = system.GPU().Maxwell3D();
     for (const auto& entry : entries.samplers) {
-        for (std::size_t i = 0; i < entry.Size(); ++i) {
+        for (std::size_t i = 0; i < entry.size; ++i) {
             const auto texture = GetTextureInfo(gpu, entry, stage, i);
             SetupTexture(texture, entry);
         }
@@ -1017,7 +1023,7 @@ void RasterizerVulkan::SetupComputeTextures(const ShaderEntries& entries) {
     MICROPROFILE_SCOPE(Vulkan_Textures);
     const auto& gpu = system.GPU().KeplerCompute();
     for (const auto& entry : entries.samplers) {
-        for (std::size_t i = 0; i < entry.Size(); ++i) {
+        for (std::size_t i = 0; i < entry.size; ++i) {
             const auto texture = GetTextureInfo(gpu, entry, ComputeShaderIndex, i);
             SetupTexture(texture, entry);
         }
@@ -1099,7 +1105,7 @@ void RasterizerVulkan::SetupTexture(const Tegra::Texture::FullTextureInfo& textu
 void RasterizerVulkan::SetupImage(const Tegra::Texture::TICEntry& tic, const ImageEntry& entry) {
     auto view = texture_cache.GetImageSurface(tic, entry);
 
-    if (entry.IsWritten()) {
+    if (entry.is_written) {
         view->MarkAsModified(texture_cache.Tick());
     }
 
@@ -1250,28 +1256,29 @@ std::size_t RasterizerVulkan::CalculateConstBufferSize(
 }
 
 RenderPassParams RasterizerVulkan::GetRenderPassParams(Texceptions texceptions) const {
-    using namespace VideoCore::Surface;
-
     const auto& regs = system.GPU().Maxwell3D().regs;
-    RenderPassParams renderpass_params;
+    const std::size_t num_attachments = static_cast<std::size_t>(regs.rt_control.count);
 
-    for (std::size_t rt = 0; rt < static_cast<std::size_t>(regs.rt_control.count); ++rt) {
+    RenderPassParams params;
+    params.color_formats = {};
+    std::size_t color_texceptions = 0;
+
+    std::size_t index = 0;
+    for (std::size_t rt = 0; rt < num_attachments; ++rt) {
         const auto& rendertarget = regs.rt[rt];
         if (rendertarget.Address() == 0 || rendertarget.format == Tegra::RenderTargetFormat::NONE) {
             continue;
         }
-        renderpass_params.color_attachments.push_back(RenderPassParams::ColorAttachment{
-            static_cast<u32>(rt), PixelFormatFromRenderTargetFormat(rendertarget.format),
-            texceptions[rt]});
+        params.color_formats[index] = static_cast<u8>(rendertarget.format);
+        color_texceptions |= (texceptions[rt] ? 1ULL : 0ULL) << index;
+        ++index;
     }
+    params.num_color_attachments = static_cast<u8>(index);
+    params.texceptions = static_cast<u8>(color_texceptions);
 
-    renderpass_params.has_zeta = regs.zeta_enable;
-    if (renderpass_params.has_zeta) {
-        renderpass_params.zeta_pixel_format = PixelFormatFromDepthFormat(regs.zeta.format);
-        renderpass_params.zeta_texception = texceptions[ZETA_TEXCEPTION_INDEX];
-    }
-
-    return renderpass_params;
+    params.zeta_format = regs.zeta_enable ? static_cast<u8>(regs.zeta.format) : 0;
+    params.zeta_texception = texceptions[ZETA_TEXCEPTION_INDEX];
+    return params;
 }
 
 VkBuffer RasterizerVulkan::DefaultBuffer() {

src/video_core/renderer_vulkan/vk_rasterizer.h

@@ -14,7 +14,6 @@
 #include <boost/functional/hash.hpp>
 
 #include "common/common_types.h"
-#include "video_core/memory_manager.h"
 #include "video_core/rasterizer_accelerated.h"
 #include "video_core/rasterizer_interface.h"
 #include "video_core/renderer_vulkan/fixed_pipeline_state.h"

src/video_core/renderer_vulkan/vk_renderpass_cache.cpp

@@ -2,9 +2,11 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cstring>
 #include <memory>
 #include <vector>
 
+#include "common/cityhash.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_vulkan/maxwell_to_vk.h"
 #include "video_core/renderer_vulkan/vk_device.h"
@@ -13,6 +15,15 @@
 
 namespace Vulkan {
 
+std::size_t RenderPassParams::Hash() const noexcept {
+    const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), sizeof *this);
+    return static_cast<std::size_t>(hash);
+}
+
+bool RenderPassParams::operator==(const RenderPassParams& rhs) const noexcept {
+    return std::memcmp(&rhs, this, sizeof *this) == 0;
+}
+
 VKRenderPassCache::VKRenderPassCache(const VKDevice& device) : device{device} {}
 
 VKRenderPassCache::~VKRenderPassCache() = default;
@@ -27,20 +38,22 @@ VkRenderPass VKRenderPassCache::GetRenderPass(const RenderPassParams& params) {
 }
 
 vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& params) const {
+    using namespace VideoCore::Surface;
     std::vector<VkAttachmentDescription> descriptors;
     std::vector<VkAttachmentReference> color_references;
 
-    for (std::size_t rt = 0; rt < params.color_attachments.size(); ++rt) {
-        const auto attachment = params.color_attachments[rt];
-        const auto format =
-            MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, attachment.pixel_format);
+    const std::size_t num_attachments = static_cast<std::size_t>(params.num_color_attachments);
+    for (std::size_t rt = 0; rt < num_attachments; ++rt) {
+        const auto guest_format = static_cast<Tegra::RenderTargetFormat>(params.color_formats[rt]);
+        const PixelFormat pixel_format = PixelFormatFromRenderTargetFormat(guest_format);
+        const auto format = MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, pixel_format);
         ASSERT_MSG(format.attachable, "Trying to attach a non-attachable format with format={}",
-                   static_cast<u32>(attachment.pixel_format));
+                   static_cast<int>(pixel_format));
 
-        // TODO(Rodrigo): Add eMayAlias when it's needed.
-        const auto color_layout = attachment.is_texception
-                                      ? VK_IMAGE_LAYOUT_GENERAL
-                                      : VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL;
+        // TODO(Rodrigo): Add MAY_ALIAS_BIT when it's needed.
+        const VkImageLayout color_layout = ((params.texceptions >> rt) & 1) != 0
+                                               ? VK_IMAGE_LAYOUT_GENERAL
+                                               : VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL;
         VkAttachmentDescription& descriptor = descriptors.emplace_back();
         descriptor.flags = VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT;
         descriptor.format = format.format;
@@ -58,15 +71,17 @@ vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& param
     }
 
     VkAttachmentReference zeta_attachment_ref;
-    if (params.has_zeta) {
-        const auto format =
-            MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, params.zeta_pixel_format);
+    const bool has_zeta = params.zeta_format != 0;
+    if (has_zeta) {
+        const auto guest_format = static_cast<Tegra::DepthFormat>(params.zeta_format);
+        const PixelFormat pixel_format = PixelFormatFromDepthFormat(guest_format);
+        const auto format = MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, pixel_format);
         ASSERT_MSG(format.attachable, "Trying to attach a non-attachable format with format={}",
-                   static_cast<u32>(params.zeta_pixel_format));
+                   static_cast<int>(pixel_format));
 
-        const auto zeta_layout = params.zeta_texception
-                                     ? VK_IMAGE_LAYOUT_GENERAL
-                                     : VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL;
+        const VkImageLayout zeta_layout = params.zeta_texception != 0
+                                              ? VK_IMAGE_LAYOUT_GENERAL
+                                              : VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL;
         VkAttachmentDescription& descriptor = descriptors.emplace_back();
         descriptor.flags = 0;
         descriptor.format = format.format;
@@ -78,7 +93,7 @@ vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& param
         descriptor.initialLayout = zeta_layout;
         descriptor.finalLayout = zeta_layout;
 
-        zeta_attachment_ref.attachment = static_cast<u32>(params.color_attachments.size());
+        zeta_attachment_ref.attachment = static_cast<u32>(num_attachments);
         zeta_attachment_ref.layout = zeta_layout;
     }
 
@@ -90,7 +105,7 @@ vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& param
     subpass_description.colorAttachmentCount = static_cast<u32>(color_references.size());
     subpass_description.pColorAttachments = color_references.data();
     subpass_description.pResolveAttachments = nullptr;
-    subpass_description.pDepthStencilAttachment = params.has_zeta ? &zeta_attachment_ref : nullptr;
+    subpass_description.pDepthStencilAttachment = has_zeta ? &zeta_attachment_ref : nullptr;
     subpass_description.preserveAttachmentCount = 0;
     subpass_description.pPreserveAttachments = nullptr;
 
@@ -101,7 +116,7 @@ vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& param
         stage |= VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT;
     }
 
-    if (params.has_zeta) {
+    if (has_zeta) {
         access |= VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT |
                   VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT;
         stage |= VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT;

src/video_core/renderer_vulkan/vk_renderpass_cache.h

@@ -4,8 +4,7 @@
 
 #pragma once
 
-#include <memory>
-#include <tuple>
+#include <type_traits>
 #include <unordered_map>
 
 #include <boost/container/static_vector.hpp>
@@ -19,51 +18,25 @@ namespace Vulkan {
 
 class VKDevice;
 
-// TODO(Rodrigo): Optimize this structure for faster hashing
-
 struct RenderPassParams {
-    struct ColorAttachment {
-        u32 index = 0;
-        VideoCore::Surface::PixelFormat pixel_format = VideoCore::Surface::PixelFormat::Invalid;
-        bool is_texception = false;
+    std::array<u8, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> color_formats;
+    u8 num_color_attachments;
+    u8 texceptions;
 
-        std::size_t Hash() const noexcept {
-            return static_cast<std::size_t>(pixel_format) |
-                   static_cast<std::size_t>(is_texception) << 6 |
-                   static_cast<std::size_t>(index) << 7;
-        }
+    u8 zeta_format;
+    u8 zeta_texception;
 
-        bool operator==(const ColorAttachment& rhs) const noexcept {
-            return std::tie(index, pixel_format, is_texception) ==
-                   std::tie(rhs.index, rhs.pixel_format, rhs.is_texception);
-        }
-    };
+    std::size_t Hash() const noexcept;
 
-    boost::container::static_vector<ColorAttachment,
-                                    Tegra::Engines::Maxwell3D::Regs::NumRenderTargets>
-        color_attachments{};
-    // TODO(Rodrigo): Unify has_zeta into zeta_pixel_format and zeta_component_type.
-    VideoCore::Surface::PixelFormat zeta_pixel_format = VideoCore::Surface::PixelFormat::Invalid;
-    bool has_zeta = false;
-    bool zeta_texception = false;
+    bool operator==(const RenderPassParams& rhs) const noexcept;
 
-    std::size_t Hash() const noexcept {
-        std::size_t hash = 0;
-        for (const auto& rt : color_attachments) {
-            boost::hash_combine(hash, rt.Hash());
-        }
-        boost::hash_combine(hash, zeta_pixel_format);
-        boost::hash_combine(hash, has_zeta);
-        boost::hash_combine(hash, zeta_texception);
-        return hash;
-    }
-
-    bool operator==(const RenderPassParams& rhs) const {
-        return std::tie(color_attachments, zeta_pixel_format, has_zeta, zeta_texception) ==
-               std::tie(rhs.color_attachments, rhs.zeta_pixel_format, rhs.has_zeta,
-                        rhs.zeta_texception);
+    bool operator!=(const RenderPassParams& rhs) const noexcept {
+        return !operator==(rhs);
     }
 };
+static_assert(std::has_unique_object_representations_v<RenderPassParams>);
+static_assert(std::is_trivially_copyable_v<RenderPassParams>);
+static_assert(std::is_trivially_constructible_v<RenderPassParams>);
 
 } // namespace Vulkan
 

src/video_core/renderer_vulkan/vk_sampler_cache.cpp

@@ -2,11 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <cstring>
-#include <optional>
 #include <unordered_map>
 
-#include "common/assert.h"
 #include "video_core/renderer_vulkan/maxwell_to_vk.h"
 #include "video_core/renderer_vulkan/vk_sampler_cache.h"
 #include "video_core/renderer_vulkan/wrapper.h"

src/video_core/renderer_vulkan/vk_scheduler.cpp

@@ -8,7 +8,6 @@
 #include <thread>
 #include <utility>
 
-#include "common/assert.h"
 #include "common/microprofile.h"
 #include "video_core/renderer_vulkan/vk_device.h"
 #include "video_core/renderer_vulkan/vk_query_cache.h"

src/video_core/renderer_vulkan/vk_scheduler.h

@@ -7,7 +7,6 @@
 #include <atomic>
 #include <condition_variable>
 #include <memory>
-#include <optional>
 #include <stack>
 #include <thread>
 #include <utility>

src/video_core/renderer_vulkan/vk_shader_decompiler.cpp

@@ -103,8 +103,8 @@ struct GenericVaryingDescription {
 };
 
 spv::Dim GetSamplerDim(const Sampler& sampler) {
-    ASSERT(!sampler.IsBuffer());
-    switch (sampler.GetType()) {
+    ASSERT(!sampler.is_buffer);
+    switch (sampler.type) {
     case Tegra::Shader::TextureType::Texture1D:
         return spv::Dim::Dim1D;
     case Tegra::Shader::TextureType::Texture2D:
@@ -114,13 +114,13 @@ spv::Dim GetSamplerDim(const Sampler& sampler) {
     case Tegra::Shader::TextureType::TextureCube:
         return spv::Dim::Cube;
     default:
-        UNIMPLEMENTED_MSG("Unimplemented sampler type={}", static_cast<u32>(sampler.GetType()));
+        UNIMPLEMENTED_MSG("Unimplemented sampler type={}", static_cast<int>(sampler.type));
         return spv::Dim::Dim2D;
     }
 }
 
 std::pair<spv::Dim, bool> GetImageDim(const Image& image) {
-    switch (image.GetType()) {
+    switch (image.type) {
     case Tegra::Shader::ImageType::Texture1D:
         return {spv::Dim::Dim1D, false};
     case Tegra::Shader::ImageType::TextureBuffer:
@@ -134,7 +134,7 @@ std::pair<spv::Dim, bool> GetImageDim(const Image& image) {
     case Tegra::Shader::ImageType::Texture3D:
         return {spv::Dim::Dim3D, false};
     default:
-        UNIMPLEMENTED_MSG("Unimplemented image type={}", static_cast<u32>(image.GetType()));
+        UNIMPLEMENTED_MSG("Unimplemented image type={}", static_cast<int>(image.type));
         return {spv::Dim::Dim2D, false};
     }
 }
@@ -879,11 +879,11 @@ private:
 
     u32 DeclareTexelBuffers(u32 binding) {
         for (const auto& sampler : ir.GetSamplers()) {
-            if (!sampler.IsBuffer()) {
+            if (!sampler.is_buffer) {
                 continue;
             }
-            ASSERT(!sampler.IsArray());
-            ASSERT(!sampler.IsShadow());
+            ASSERT(!sampler.is_array);
+            ASSERT(!sampler.is_shadow);
 
             constexpr auto dim = spv::Dim::Buffer;
             constexpr int depth = 0;
@@ -894,23 +894,23 @@ private:
             const Id image_type = TypeImage(t_float, dim, depth, arrayed, ms, sampled, format);
             const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, image_type);
             const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant);
-            AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.GetIndex())));
+            AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.index)));
             Decorate(id, spv::Decoration::Binding, binding++);
             Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET);
 
-            texel_buffers.emplace(sampler.GetIndex(), TexelBuffer{image_type, id});
+            texel_buffers.emplace(sampler.index, TexelBuffer{image_type, id});
         }
         return binding;
     }
 
     u32 DeclareSamplers(u32 binding) {
         for (const auto& sampler : ir.GetSamplers()) {
-            if (sampler.IsBuffer()) {
+            if (sampler.is_buffer) {
                 continue;
             }
             const auto dim = GetSamplerDim(sampler);
-            const int depth = sampler.IsShadow() ? 1 : 0;
-            const int arrayed = sampler.IsArray() ? 1 : 0;
+            const int depth = sampler.is_shadow ? 1 : 0;
+            const int arrayed = sampler.is_array ? 1 : 0;
             constexpr bool ms = false;
             constexpr int sampled = 1;
             constexpr auto format = spv::ImageFormat::Unknown;
@@ -918,17 +918,17 @@ private:
             const Id sampler_type = TypeSampledImage(image_type);
             const Id sampler_pointer_type =
                 TypePointer(spv::StorageClass::UniformConstant, sampler_type);
-            const Id type = sampler.IsIndexed()
-                                ? TypeArray(sampler_type, Constant(t_uint, sampler.Size()))
+            const Id type = sampler.is_indexed
+                                ? TypeArray(sampler_type, Constant(t_uint, sampler.size))
                                 : sampler_type;
             const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, type);
             const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant);
-            AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.GetIndex())));
+            AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.index)));
             Decorate(id, spv::Decoration::Binding, binding++);
             Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET);
 
-            sampled_images.emplace(sampler.GetIndex(), SampledImage{image_type, sampler_type,
-                                                                    sampler_pointer_type, id});
+            sampled_images.emplace(
+                sampler.index, SampledImage{image_type, sampler_type, sampler_pointer_type, id});
         }
         return binding;
     }
@@ -943,17 +943,17 @@ private:
             const Id image_type = TypeImage(t_uint, dim, depth, arrayed, ms, sampled, format, {});
             const Id pointer_type = TypePointer(spv::StorageClass::UniformConstant, image_type);
             const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant);
-            AddGlobalVariable(Name(id, fmt::format("image_{}", image.GetIndex())));
+            AddGlobalVariable(Name(id, fmt::format("image_{}", image.index)));
 
             Decorate(id, spv::Decoration::Binding, binding++);
             Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET);
-            if (image.IsRead() && !image.IsWritten()) {
+            if (image.is_read && !image.is_written) {
                 Decorate(id, spv::Decoration::NonWritable);
-            } else if (image.IsWritten() && !image.IsRead()) {
+            } else if (image.is_written && !image.is_read) {
                 Decorate(id, spv::Decoration::NonReadable);
             }
 
-            images.emplace(static_cast<u32>(image.GetIndex()), StorageImage{image_type, id});
+            images.emplace(image.index, StorageImage{image_type, id});
         }
         return binding;
     }
@@ -1584,6 +1584,15 @@ private:
         return {OpCompositeConstruct(t_half, low, high), Type::HalfFloat};
     }
 
+    Expression LogicalAddCarry(Operation operation) {
+        const Id op_a = AsUint(Visit(operation[0]));
+        const Id op_b = AsUint(Visit(operation[1]));
+
+        const Id result = OpIAddCarry(TypeStruct({t_uint, t_uint}), op_a, op_b);
+        const Id carry = OpCompositeExtract(t_uint, result, 1);
+        return {OpINotEqual(t_bool, carry, Constant(t_uint, 0)), Type::Bool};
+    }
+
     Expression LogicalAssign(Operation operation) {
         const Node& dest = operation[0];
         const Node& src = operation[1];
@@ -1611,11 +1620,11 @@ private:
 
     Id GetTextureSampler(Operation operation) {
         const auto& meta = std::get<MetaTexture>(operation.GetMeta());
-        ASSERT(!meta.sampler.IsBuffer());
+        ASSERT(!meta.sampler.is_buffer);
 
-        const auto& entry = sampled_images.at(meta.sampler.GetIndex());
+        const auto& entry = sampled_images.at(meta.sampler.index);
         Id sampler = entry.variable;
-        if (meta.sampler.IsIndexed()) {
+        if (meta.sampler.is_indexed) {
             const Id index = AsInt(Visit(meta.index));
             sampler = OpAccessChain(entry.sampler_pointer_type, sampler, index);
         }
@@ -1624,8 +1633,8 @@ private:
 
     Id GetTextureImage(Operation operation) {
         const auto& meta = std::get<MetaTexture>(operation.GetMeta());
-        const u32 index = meta.sampler.GetIndex();
-        if (meta.sampler.IsBuffer()) {
+        const u32 index = meta.sampler.index;
+        if (meta.sampler.is_buffer) {
             const auto& entry = texel_buffers.at(index);
             return OpLoad(entry.image_type, entry.image);
         } else {
@@ -1636,7 +1645,7 @@ private:
 
     Id GetImage(Operation operation) {
         const auto& meta = std::get<MetaImage>(operation.GetMeta());
-        const auto entry = images.at(meta.image.GetIndex());
+        const auto entry = images.at(meta.image.index);
         return OpLoad(entry.image_type, entry.image);
     }
 
@@ -1652,7 +1661,7 @@ private:
         }
         if (const auto meta = std::get_if<MetaTexture>(&operation.GetMeta())) {
             // Add array coordinate for textures
-            if (meta->sampler.IsArray()) {
+            if (meta->sampler.is_array) {
                 Id array = AsInt(Visit(meta->array));
                 if (type == Type::Float) {
                     array = OpConvertSToF(t_float, array);
@@ -1758,7 +1767,7 @@ private:
             operands.push_back(GetOffsetCoordinates(operation));
         }
 
-        if (meta.sampler.IsShadow()) {
+        if (meta.sampler.is_shadow) {
             const Id dref = AsFloat(Visit(meta.depth_compare));
             return {OpImageSampleDrefExplicitLod(t_float, sampler, coords, dref, mask, operands),
                     Type::Float};
@@ -1773,7 +1782,7 @@ private:
 
         const Id coords = GetCoordinates(operation, Type::Float);
         Id texture{};
-        if (meta.sampler.IsShadow()) {
+        if (meta.sampler.is_shadow) {
             texture = OpImageDrefGather(t_float4, GetTextureSampler(operation), coords,
                                         AsFloat(Visit(meta.depth_compare)));
         } else {
@@ -1800,8 +1809,8 @@ private:
         }
 
         const Id lod = AsUint(Visit(operation[0]));
-        const std::size_t coords_count = [&]() {
-            switch (const auto type = meta.sampler.GetType(); type) {
+        const std::size_t coords_count = [&meta] {
+            switch (const auto type = meta.sampler.type) {
             case Tegra::Shader::TextureType::Texture1D:
                 return 1;
             case Tegra::Shader::TextureType::Texture2D:
@@ -1810,7 +1819,7 @@ private:
             case Tegra::Shader::TextureType::Texture3D:
                 return 3;
             default:
-                UNREACHABLE_MSG("Invalid texture type={}", static_cast<u32>(type));
+                UNREACHABLE_MSG("Invalid texture type={}", static_cast<int>(type));
                 return 2;
             }
         }();
@@ -1853,7 +1862,7 @@ private:
         const Id image = GetTextureImage(operation);
         const Id coords = GetCoordinates(operation, Type::Int);
         Id fetch;
-        if (meta.lod && !meta.sampler.IsBuffer()) {
+        if (meta.lod && !meta.sampler.is_buffer) {
             fetch = OpImageFetch(t_float4, image, coords, spv::ImageOperandsMask::Lod,
                                  AsInt(Visit(meta.lod)));
         } else {
@@ -2518,6 +2527,8 @@ private:
         &SPIRVDecompiler::Binary<&Module::OpINotEqual, Type::Bool, Type::Uint>,
         &SPIRVDecompiler::Binary<&Module::OpUGreaterThanEqual, Type::Bool, Type::Uint>,
 
+        &SPIRVDecompiler::LogicalAddCarry,
+
         &SPIRVDecompiler::Binary<&Module::OpFOrdLessThan, Type::Bool2, Type::HalfFloat>,
         &SPIRVDecompiler::Binary<&Module::OpFOrdEqual, Type::Bool2, Type::HalfFloat>,
         &SPIRVDecompiler::Binary<&Module::OpFOrdLessThanEqual, Type::Bool2, Type::HalfFloat>,
@@ -2969,7 +2980,7 @@ ShaderEntries GenerateShaderEntries(const VideoCommon::Shader::ShaderIR& ir) {
         entries.global_buffers.emplace_back(base.cbuf_index, base.cbuf_offset, usage.is_written);
     }
     for (const auto& sampler : ir.GetSamplers()) {
-        if (sampler.IsBuffer()) {
+        if (sampler.is_buffer) {
             entries.texel_buffers.emplace_back(sampler);
         } else {
             entries.samplers.emplace_back(sampler);

src/video_core/renderer_vulkan/vk_shader_decompiler.h

@@ -5,11 +5,7 @@
 #pragma once
 
 #include <array>
-#include <bitset>
-#include <memory>
 #include <set>
-#include <type_traits>
-#include <utility>
 #include <vector>
 
 #include "common/common_types.h"

src/video_core/renderer_vulkan/vk_shader_util.cpp

@@ -4,8 +4,7 @@
 
 #include <cstring>
 #include <memory>
-#include <vector>
-#include "common/alignment.h"
+
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "video_core/renderer_vulkan/vk_device.h"

src/video_core/renderer_vulkan/vk_shader_util.h

@@ -4,7 +4,6 @@
 
 #pragma once
 
-#include <vector>
 #include "common/common_types.h"
 #include "video_core/renderer_vulkan/wrapper.h"
 

src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp

@@ -39,8 +39,7 @@ VKStagingBufferPool::StagingBuffer& VKStagingBufferPool::StagingBuffer::operator
 
 VKStagingBufferPool::VKStagingBufferPool(const VKDevice& device, VKMemoryManager& memory_manager,
                                          VKScheduler& scheduler)
-    : device{device}, memory_manager{memory_manager}, scheduler{scheduler},
-      is_device_integrated{device.IsIntegrated()} {}
+    : device{device}, memory_manager{memory_manager}, scheduler{scheduler} {}
 
 VKStagingBufferPool::~VKStagingBufferPool() = default;
 
@@ -56,9 +55,7 @@ void VKStagingBufferPool::TickFrame() {
     current_delete_level = (current_delete_level + 1) % NumLevels;
 
     ReleaseCache(true);
-    if (!is_device_integrated) {
-        ReleaseCache(false);
-    }
+    ReleaseCache(false);
 }
 
 VKBuffer* VKStagingBufferPool::TryGetReservedBuffer(std::size_t size, bool host_visible) {
@@ -95,7 +92,7 @@ VKBuffer& VKStagingBufferPool::CreateStagingBuffer(std::size_t size, bool host_v
 }
 
 VKStagingBufferPool::StagingBuffersCache& VKStagingBufferPool::GetCache(bool host_visible) {
-    return is_device_integrated || host_visible ? host_staging_buffers : device_staging_buffers;
+    return host_visible ? host_staging_buffers : device_staging_buffers;
 }
 
 void VKStagingBufferPool::ReleaseCache(bool host_visible) {

src/video_core/renderer_vulkan/vk_staging_buffer_pool.h

@@ -5,8 +5,6 @@
 #pragma once
 
 #include <climits>
-#include <unordered_map>
-#include <utility>
 #include <vector>
 
 #include "common/common_types.h"
@@ -71,7 +69,6 @@ private:
     const VKDevice& device;
     VKMemoryManager& memory_manager;
     VKScheduler& scheduler;
-    const bool is_device_integrated;
 
     StagingBuffersCache host_staging_buffers;
     StagingBuffersCache device_staging_buffers;

src/video_core/renderer_vulkan/vk_texture_cache.cpp

@@ -10,11 +10,9 @@
 #include <variant>
 #include <vector>
 
-#include "common/alignment.h"
 #include "common/assert.h"
 #include "common/common_types.h"
 #include "core/core.h"
-#include "core/memory.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/morton.h"
 #include "video_core/renderer_vulkan/maxwell_to_vk.h"
@@ -26,7 +24,6 @@
 #include "video_core/renderer_vulkan/vk_texture_cache.h"
 #include "video_core/renderer_vulkan/wrapper.h"
 #include "video_core/surface.h"
-#include "video_core/textures/convert.h"
 
 namespace Vulkan {