early-access version 3347

README.md

@@ -1,7 +1,7 @@
 yuzu emulator early access
 =============
 
-This is the source code for early-access 3345.
+This is the source code for early-access 3347.
 
 ## Legal Notice
 

dist/yuzu.svg

@@ -1 +1 @@
-<svg id="svg815" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" viewBox="0 0 614.4 682.67"><defs><style>.cls-1{fill:none;}.cls-2{clip-path:url(#clip-path);}.cls-3{fill:#ff3c28;}.cls-4{fill:#0ab9e6;}</style><clipPath id="clip-path"><rect class="cls-1" x="-43" y="-46.67" width="699.6" height="777.33"/></clipPath></defs><title>Artboard 1</title><g id="g823"><g id="right"><g class="cls-2"><g id="g827"><g id="g833"><path id="path835" class="cls-3" d="M340.81,138V682.08c150.26,0,272.06-121.81,272.06-272.06S491.07,138,340.81,138M394,197.55a219.06,219.06,0,0,1,0,424.94V197.55"/></g></g></g></g><g id="left"><g class="cls-2"><g id="g839"><g id="g845"><path id="path847" class="cls-4" d="M272.79,1.92C122.53,1.92.73,123.73.73,274s121.8,272.07,272.06,272.07ZM219.65,61.51v425A219,219,0,0,1,118,119.18,217.51,217.51,0,0,1,219.65,61.51"/></g></g></g></g></g></svg>
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 612.15 680.17"><defs><style>.cls-1{fill:#c6c6c6;}.cls-2{fill:#ffdc00;}</style></defs><title>newAsset 7</title><g id="Layer_2" data-name="Layer 2"><g id="Layer_1-2" data-name="Layer 1"><g id="g823"><g id="right"><g id="g827"><g id="g833"><path id="path835" class="cls-1" d="M340.08,136V680.17c150.26,0,272.07-121.81,272.07-272.07S490.34,136,340.08,136m53.14,59.6a219.06,219.06,0,0,1,0,424.94V195.63"/></g></g></g><g id="left"><g id="g839"><g id="g845"><path id="path847" class="cls-2" d="M272.07,0C121.81,0,0,121.81,0,272.07S121.81,544.13,272.07,544.13ZM218.93,59.6V484.54A219,219,0,0,1,117.26,117.26,217.44,217.44,0,0,1,218.93,59.6"/></g></g></g></g></g></g></svg>

src/common/input.h

@@ -130,6 +130,8 @@ struct ButtonStatus {
     bool inverted{};
     // Press once to activate, press again to release
     bool toggle{};
+    // Spams the button when active
+    bool turbo{};
     // Internal lock for the toggle status
     bool locked{};
 };

src/common/polyfill_thread.h

@@ -41,17 +41,18 @@ bool StoppableTimedWait(std::stop_token token, const std::chrono::duration<Rep,
 #include <chrono>
 #include <condition_variable>
 #include <functional>
-#include <list>
+#include <map>
 #include <memory>
 #include <mutex>
 #include <optional>
 #include <thread>
 #include <type_traits>
+#include <utility>
 
 namespace std {
 namespace polyfill {
 
-using stop_state_callbacks = list<function<void()>>;
+using stop_state_callback = size_t;
 
 class stop_state {
 public:
@@ -59,61 +60,69 @@ public:
     ~stop_state() = default;
 
     bool request_stop() {
-        stop_state_callbacks callbacks;
+        unique_lock lk{m_lock};
 
-        {
+        if (m_stop_requested) {
-            scoped_lock lk{m_lock};
+            // Already set, nothing to do.
-
-            if (m_stop_requested.load()) {
-                // Already set, nothing to do
             return false;
         }
 
-            // Set as requested
+        // Mark stop requested.
         m_stop_requested = true;
 
-            // Copy callback list
+        while (!m_callbacks.empty()) {
-            callbacks = m_callbacks;
+            // Get an iterator to the first element.
-        }
+            const auto it = m_callbacks.begin();
 
-        for (auto callback : callbacks) {
+            // Move the callback function out of the map.
-            callback();
+            function<void()> f;
+            swap(it->second, f);
+
+            // Erase the now-empty map element.
+            m_callbacks.erase(it);
+
+            // Run the callback.
+            if (f) {
+                f();
+            }
         }
 
         return true;
     }
 
     bool stop_requested() const {
-        return m_stop_requested.load();
+        unique_lock lk{m_lock};
+        return m_stop_requested;
     }
 
-    stop_state_callbacks::const_iterator insert_callback(function<void()> f) {
+    stop_state_callback insert_callback(function<void()> f) {
-        stop_state_callbacks::const_iterator ret{};
+        unique_lock lk{m_lock};
-        bool should_run{};
 
-        {
+        if (m_stop_requested) {
-            scoped_lock lk{m_lock};
+            // Stop already requested. Don't insert anything,
-            should_run = m_stop_requested.load();
+            // just run the callback synchronously.
-            m_callbacks.push_front(f);
+            if (f) {
-            ret = m_callbacks.begin();
-        }
-
-        if (should_run) {
                 f();
             }
+            return 0;
+        }
 
+        // Insert the callback.
+        stop_state_callback ret = ++m_next_callback;
+        m_callbacks.emplace(ret, move(f));
         return ret;
     }
 
-    void remove_callback(stop_state_callbacks::const_iterator it) {
+    void remove_callback(stop_state_callback cb) {
-        scoped_lock lk{m_lock};
+        unique_lock lk{m_lock};
-        m_callbacks.erase(it);
+        m_callbacks.erase(cb);
     }
 
 private:
-    mutex m_lock;
+    mutable recursive_mutex m_lock;
-    atomic<bool> m_stop_requested;
+    map<stop_state_callback, function<void()>> m_callbacks;
-    stop_state_callbacks m_callbacks;
+    stop_state_callback m_next_callback{0};
+    bool m_stop_requested{false};
 };
 
 } // namespace polyfill
@@ -223,7 +232,7 @@ public:
     }
     ~stop_callback() {
         if (m_stop_state && m_callback) {
-            m_stop_state->remove_callback(*m_callback);
+            m_stop_state->remove_callback(m_callback);
         }
     }
 
@@ -234,7 +243,7 @@ public:
 
 private:
     shared_ptr<polyfill::stop_state> m_stop_state;
-    optional<polyfill::stop_state_callbacks::const_iterator> m_callback;
+    polyfill::stop_state_callback m_callback;
 };
 
 template <typename Callback>

src/common/string_util.cpp

@@ -30,7 +30,7 @@ std::string ToUpper(std::string str) {
     return str;
 }
 
-std::string StringFromBuffer(const std::vector<u8>& data) {
+std::string StringFromBuffer(std::span<const u8> data) {
     return std::string(data.begin(), std::find(data.begin(), data.end(), '\0'));
 }
 

src/common/string_util.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <cstddef>
+#include <span>
 #include <string>
 #include <vector>
 #include "common/common_types.h"
@@ -17,7 +18,7 @@ namespace Common {
 /// Make a string uppercase
 [[nodiscard]] std::string ToUpper(std::string str);
 
-[[nodiscard]] std::string StringFromBuffer(const std::vector<u8>& data);
+[[nodiscard]] std::string StringFromBuffer(std::span<const u8> data);
 
 [[nodiscard]] std::string StripSpaces(const std::string& s);
 [[nodiscard]] std::string StripQuotes(const std::string& s);

src/core/CMakeLists.txt

@@ -182,6 +182,8 @@ add_library(core STATIC
     hle/kernel/k_auto_object_container.cpp
     hle/kernel/k_auto_object_container.h
     hle/kernel/k_affinity_mask.h
+    hle/kernel/k_capabilities.cpp
+    hle/kernel/k_capabilities.h
     hle/kernel/k_class_token.cpp
     hle/kernel/k_class_token.h
     hle/kernel/k_client_port.cpp

src/core/hardware_properties.h

@@ -25,6 +25,26 @@ constexpr std::array<s32, Common::BitSize<u64>()> VirtualToPhysicalCoreMap{
     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3,
 };
 
+static constexpr inline size_t NumVirtualCores = Common::BitSize<u64>();
+
+static constexpr inline u64 VirtualCoreMask = [] {
+    u64 mask = 0;
+    for (size_t i = 0; i < NumVirtualCores; ++i) {
+        mask |= (UINT64_C(1) << i);
+    }
+    return mask;
+}();
+
+static constexpr inline u64 ConvertVirtualCoreMaskToPhysical(u64 v_core_mask) {
+    u64 p_core_mask = 0;
+    while (v_core_mask != 0) {
+        const u64 next = std::countr_zero(v_core_mask);
+        v_core_mask &= ~(static_cast<u64>(1) << next);
+        p_core_mask |= (static_cast<u64>(1) << VirtualToPhysicalCoreMap[next]);
+    }
+    return p_core_mask;
+}
+
 // Cortex-A57 supports 4 memory watchpoints
 constexpr u64 NUM_WATCHPOINTS = 4;
 

src/core/hid/emulated_controller.cpp

@@ -687,6 +687,7 @@ void EmulatedController::SetButton(const Common::Input::CallbackStatus& callback
     }
 
     current_status.toggle = new_status.toggle;
+    current_status.turbo = new_status.turbo;
     current_status.uuid = uuid;
 
     // Update button status with current
@@ -1548,7 +1549,7 @@ NpadButtonState EmulatedController::GetNpadButtons() const {
     if (is_configuring) {
         return {};
     }
-    return controller.npad_button_state;
+    return {controller.npad_button_state.raw & GetTurboButtonMask()};
 }
 
 DebugPadButton EmulatedController::GetDebugPadButtons() const {
@@ -1656,4 +1657,74 @@ void EmulatedController::DeleteCallback(int key) {
     }
     callback_list.erase(iterator);
 }
+
+void EmulatedController::TurboButtonUpdate() {
+    turbo_button_state = !turbo_button_state;
+}
+
+NpadButton EmulatedController::GetTurboButtonMask() const {
+    // Apply no mask when disabled
+    if (!turbo_button_state) {
+        return {NpadButton::All};
+    }
+
+    NpadButtonState button_mask{};
+    for (std::size_t index = 0; index < controller.button_values.size(); ++index) {
+        if (!controller.button_values[index].turbo) {
+            continue;
+        }
+
+        switch (index) {
+        case Settings::NativeButton::A:
+            button_mask.a.Assign(1);
+            break;
+        case Settings::NativeButton::B:
+            button_mask.b.Assign(1);
+            break;
+        case Settings::NativeButton::X:
+            button_mask.x.Assign(1);
+            break;
+        case Settings::NativeButton::Y:
+            button_mask.y.Assign(1);
+            break;
+        case Settings::NativeButton::L:
+            button_mask.l.Assign(1);
+            break;
+        case Settings::NativeButton::R:
+            button_mask.r.Assign(1);
+            break;
+        case Settings::NativeButton::ZL:
+            button_mask.zl.Assign(1);
+            break;
+        case Settings::NativeButton::ZR:
+            button_mask.zr.Assign(1);
+            break;
+        case Settings::NativeButton::DLeft:
+            button_mask.left.Assign(1);
+            break;
+        case Settings::NativeButton::DUp:
+            button_mask.up.Assign(1);
+            break;
+        case Settings::NativeButton::DRight:
+            button_mask.right.Assign(1);
+            break;
+        case Settings::NativeButton::DDown:
+            button_mask.down.Assign(1);
+            break;
+        case Settings::NativeButton::SL:
+            button_mask.left_sl.Assign(1);
+            button_mask.right_sl.Assign(1);
+            break;
+        case Settings::NativeButton::SR:
+            button_mask.left_sr.Assign(1);
+            button_mask.right_sr.Assign(1);
+            break;
+        default:
+            break;
+        }
+    }
+
+    return static_cast<NpadButton>(~button_mask.raw);
+}
+
 } // namespace Core::HID

src/core/hid/emulated_controller.h

@@ -411,6 +411,9 @@ public:
      */
     void DeleteCallback(int key);
 
+    /// Swaps the state of the turbo buttons
+    void TurboButtonUpdate();
+
 private:
     /// creates input devices from params
     void LoadDevices();
@@ -511,6 +514,8 @@ private:
      */
     void TriggerOnChange(ControllerTriggerType type, bool is_service_update);
 
+    NpadButton GetTurboButtonMask() const;
+
     const NpadIdType npad_id_type;
     NpadStyleIndex npad_type{NpadStyleIndex::None};
     NpadStyleIndex original_npad_type{NpadStyleIndex::None};
@@ -520,6 +525,7 @@ private:
     bool system_buttons_enabled{true};
     f32 motion_sensitivity{0.01f};
     bool force_update_motion{false};
+    bool turbo_button_state{false};
 
     // Temporary values to avoid doing changes while the controller is in configuring mode
     NpadStyleIndex tmp_npad_type{NpadStyleIndex::None};

src/core/hle/kernel/hle_ipc.cpp

@@ -11,6 +11,7 @@
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/logging/log.h"
+#include "common/scratch_buffer.h"
 #include "core/hle/ipc_helpers.h"
 #include "core/hle/kernel/hle_ipc.h"
 #include "core/hle/kernel/k_auto_object.h"
@@ -325,7 +326,7 @@ Result HLERequestContext::WriteToOutgoingCommandBuffer(KThread& requesting_threa
     return ResultSuccess;
 }
 
-std::vector<u8> HLERequestContext::ReadBuffer(std::size_t buffer_index) const {
+std::vector<u8> HLERequestContext::ReadBufferCopy(std::size_t buffer_index) const {
     const bool is_buffer_a{BufferDescriptorA().size() > buffer_index &&
                            BufferDescriptorA()[buffer_index].Size()};
     if (is_buffer_a) {
@@ -345,6 +346,33 @@ std::vector<u8> HLERequestContext::ReadBuffer(std::size_t buffer_index) const {
     }
 }
 
+std::span<const u8> HLERequestContext::ReadBuffer(std::size_t buffer_index) const {
+    static thread_local std::array<Common::ScratchBuffer<u8>, 2> read_buffer_a;
+    static thread_local std::array<Common::ScratchBuffer<u8>, 2> read_buffer_x;
+
+    const bool is_buffer_a{BufferDescriptorA().size() > buffer_index &&
+                           BufferDescriptorA()[buffer_index].Size()};
+    if (is_buffer_a) {
+        ASSERT_OR_EXECUTE_MSG(
+            BufferDescriptorA().size() > buffer_index, { return {}; },
+            "BufferDescriptorA invalid buffer_index {}", buffer_index);
+        auto& read_buffer = read_buffer_a[buffer_index];
+        read_buffer.resize_destructive(BufferDescriptorA()[buffer_index].Size());
+        memory.ReadBlock(BufferDescriptorA()[buffer_index].Address(), read_buffer.data(),
+                         read_buffer.size());
+        return read_buffer;
+    } else {
+        ASSERT_OR_EXECUTE_MSG(
+            BufferDescriptorX().size() > buffer_index, { return {}; },
+            "BufferDescriptorX invalid buffer_index {}", buffer_index);
+        auto& read_buffer = read_buffer_x[buffer_index];
+        read_buffer.resize_destructive(BufferDescriptorX()[buffer_index].Size());
+        memory.ReadBlock(BufferDescriptorX()[buffer_index].Address(), read_buffer.data(),
+                         read_buffer.size());
+        return read_buffer;
+    }
+}
+
 std::size_t HLERequestContext::WriteBuffer(const void* buffer, std::size_t size,
                                            std::size_t buffer_index) const {
     if (size == 0) {

src/core/hle/kernel/hle_ipc.h

@@ -7,6 +7,7 @@
 #include <functional>
 #include <memory>
 #include <optional>
+#include <span>
 #include <string>
 #include <type_traits>
 #include <vector>
@@ -270,8 +271,11 @@ public:
         return domain_message_header.has_value();
     }
 
-    /// Helper function to read a buffer using the appropriate buffer descriptor
+    /// Helper function to get a span of a buffer using the appropriate buffer descriptor
-    [[nodiscard]] std::vector<u8> ReadBuffer(std::size_t buffer_index = 0) const;
+    [[nodiscard]] std::span<const u8> ReadBuffer(std::size_t buffer_index = 0) const;
+
+    /// Helper function to read a copy of a buffer using the appropriate buffer descriptor
+    [[nodiscard]] std::vector<u8> ReadBufferCopy(std::size_t buffer_index = 0) const;
 
     /// Helper function to write a buffer using the appropriate buffer descriptor
     std::size_t WriteBuffer(const void* buffer, std::size_t size,

src/core/hle/kernel/k_capabilities.cpp

@@ -0,0 +1,358 @@
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include "core/hardware_properties.h"
+#include "core/hle/kernel/k_capabilities.h"
+#include "core/hle/kernel/k_memory_layout.h"
+#include "core/hle/kernel/k_page_table.h"
+#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/svc_results.h"
+#include "core/hle/kernel/svc_version.h"
+
+namespace Kernel {
+
+Result KCapabilities::InitializeForKIP(std::span<const u32> kern_caps, KPageTable* page_table) {
+    // We're initializing an initial process.
+    m_svc_access_flags.reset();
+    m_irq_access_flags.reset();
+    m_debug_capabilities = 0;
+    m_handle_table_size = 0;
+    m_intended_kernel_version = 0;
+    m_program_type = 0;
+
+    // Initial processes may run on all cores.
+    constexpr u64 VirtMask = Core::Hardware::VirtualCoreMask;
+    constexpr u64 PhysMask = Core::Hardware::ConvertVirtualCoreMaskToPhysical(VirtMask);
+
+    m_core_mask = VirtMask;
+    m_phys_core_mask = PhysMask;
+
+    // Initial processes may use any user priority they like.
+    m_priority_mask = ~0xFULL;
+
+    // Here, Nintendo sets the kernel version to the current kernel version.
+    // We will follow suit and set the version to the highest supported kernel version.
+    KernelVersion intended_kernel_version{};
+    intended_kernel_version.major_version.Assign(Svc::SupportedKernelMajorVersion);
+    intended_kernel_version.minor_version.Assign(Svc::SupportedKernelMinorVersion);
+    m_intended_kernel_version = intended_kernel_version.raw;
+
+    // Parse the capabilities array.
+    R_RETURN(this->SetCapabilities(kern_caps, page_table));
+}
+
+Result KCapabilities::InitializeForUser(std::span<const u32> user_caps, KPageTable* page_table) {
+    // We're initializing a user process.
+    m_svc_access_flags.reset();
+    m_irq_access_flags.reset();
+    m_debug_capabilities = 0;
+    m_handle_table_size = 0;
+    m_intended_kernel_version = 0;
+    m_program_type = 0;
+
+    // User processes must specify what cores/priorities they can use.
+    m_core_mask = 0;
+    m_priority_mask = 0;
+
+    // Parse the user capabilities array.
+    R_RETURN(this->SetCapabilities(user_caps, page_table));
+}
+
+Result KCapabilities::SetCorePriorityCapability(const u32 cap) {
+    // We can't set core/priority if we've already set them.
+    R_UNLESS(m_core_mask == 0, ResultInvalidArgument);
+    R_UNLESS(m_priority_mask == 0, ResultInvalidArgument);
+
+    // Validate the core/priority.
+    CorePriority pack{cap};
+    const u32 min_core = pack.minimum_core_id;
+    const u32 max_core = pack.maximum_core_id;
+    const u32 max_prio = pack.lowest_thread_priority;
+    const u32 min_prio = pack.highest_thread_priority;
+
+    R_UNLESS(min_core <= max_core, ResultInvalidCombination);
+    R_UNLESS(min_prio <= max_prio, ResultInvalidCombination);
+    R_UNLESS(max_core < Core::Hardware::NumVirtualCores, ResultInvalidCoreId);
+
+    ASSERT(max_prio < Common::BitSize<u64>());
+
+    // Set core mask.
+    for (auto core_id = min_core; core_id <= max_core; core_id++) {
+        m_core_mask |= (1ULL << core_id);
+    }
+    ASSERT((m_core_mask & Core::Hardware::VirtualCoreMask) == m_core_mask);
+
+    // Set physical core mask.
+    m_phys_core_mask = Core::Hardware::ConvertVirtualCoreMaskToPhysical(m_core_mask);
+
+    // Set priority mask.
+    for (auto prio = min_prio; prio <= max_prio; prio++) {
+        m_priority_mask |= (1ULL << prio);
+    }
+
+    // We must have some core/priority we can use.
+    R_UNLESS(m_core_mask != 0, ResultInvalidArgument);
+    R_UNLESS(m_priority_mask != 0, ResultInvalidArgument);
+
+    // Processes must not have access to kernel thread priorities.
+    R_UNLESS((m_priority_mask & 0xF) == 0, ResultInvalidArgument);
+
+    R_SUCCEED();
+}
+
+Result KCapabilities::SetSyscallMaskCapability(const u32 cap, u32& set_svc) {
+    // Validate the index.
+    SyscallMask pack{cap};
+    const u32 mask = pack.mask;
+    const u32 index = pack.index;
+
+    const u32 index_flag = (1U << index);
+    R_UNLESS((set_svc & index_flag) == 0, ResultInvalidCombination);
+    set_svc |= index_flag;
+
+    // Set SVCs.
+    for (size_t i = 0; i < decltype(SyscallMask::mask)::bits; i++) {
+        const u32 svc_id = static_cast<u32>(decltype(SyscallMask::mask)::bits* index + i);
+        if (mask & (1U << i)) {
+            R_UNLESS(this->SetSvcAllowed(svc_id), ResultOutOfRange);
+        }
+    }
+
+    R_SUCCEED();
+}
+
+Result KCapabilities::MapRange_(const u32 cap, const u32 size_cap, KPageTable* page_table) {
+    const auto range_pack = MapRange{cap};
+    const auto size_pack = MapRangeSize{size_cap};
+
+    // Get/validate address/size
+    const u64 phys_addr = range_pack.address.Value() * PageSize;
+
+    // Validate reserved bits are unused.
+    R_UNLESS(size_pack.reserved.Value() == 0, ResultOutOfRange);
+
+    const size_t num_pages = size_pack.pages;
+    const size_t size = num_pages * PageSize;
+    R_UNLESS(num_pages != 0, ResultInvalidSize);
+    R_UNLESS(phys_addr < phys_addr + size, ResultInvalidAddress);
+    R_UNLESS(((phys_addr + size - 1) & ~PhysicalMapAllowedMask) == 0, ResultInvalidAddress);
+
+    // Do the mapping.
+    [[maybe_unused]] const KMemoryPermission perm = range_pack.read_only.Value()
+                                                        ? KMemoryPermission::UserRead
+                                                        : KMemoryPermission::UserReadWrite;
+    if (MapRangeSize{size_cap}.normal) {
+        // R_RETURN(page_table->MapStatic(phys_addr, size, perm));
+    } else {
+        // R_RETURN(page_table->MapIo(phys_addr, size, perm));
+    }
+
+    UNIMPLEMENTED();
+    R_SUCCEED();
+}
+
+Result KCapabilities::MapIoPage_(const u32 cap, KPageTable* page_table) {
+    // Get/validate address/size
+    const u64 phys_addr = MapIoPage{cap}.address.Value() * PageSize;
+    const size_t num_pages = 1;
+    const size_t size = num_pages * PageSize;
+    R_UNLESS(num_pages != 0, ResultInvalidSize);
+    R_UNLESS(phys_addr < phys_addr + size, ResultInvalidAddress);
+    R_UNLESS(((phys_addr + size - 1) & ~PhysicalMapAllowedMask) == 0, ResultInvalidAddress);
+
+    // Do the mapping.
+    // R_RETURN(page_table->MapIo(phys_addr, size, KMemoryPermission_UserReadWrite));
+
+    UNIMPLEMENTED();
+    R_SUCCEED();
+}
+
+template <typename F>
+Result KCapabilities::ProcessMapRegionCapability(const u32 cap, F f) {
+    // Define the allowed memory regions.
+    constexpr std::array<KMemoryRegionType, 4> MemoryRegions{
+        KMemoryRegionType_None,
+        KMemoryRegionType_KernelTraceBuffer,
+        KMemoryRegionType_OnMemoryBootImage,
+        KMemoryRegionType_DTB,
+    };
+
+    // Extract regions/read only.
+    const MapRegion pack{cap};
+    const std::array<RegionType, 3> types{pack.region0, pack.region1, pack.region2};
+    const std::array<u32, 3> ro{pack.read_only0, pack.read_only1, pack.read_only2};
+
+    for (size_t i = 0; i < types.size(); i++) {
+        const auto type = types[i];
+        const auto perm = ro[i] ? KMemoryPermission::UserRead : KMemoryPermission::UserReadWrite;
+        switch (type) {
+        case RegionType::NoMapping:
+            break;
+        case RegionType::KernelTraceBuffer:
+        case RegionType::OnMemoryBootImage:
+        case RegionType::DTB:
+            R_TRY(f(MemoryRegions[static_cast<u32>(type)], perm));
+            break;
+        default:
+            R_THROW(ResultNotFound);
+        }
+    }
+
+    R_SUCCEED();
+}
+
+Result KCapabilities::MapRegion_(const u32 cap, KPageTable* page_table) {
+    // Map each region into the process's page table.
+    R_RETURN(ProcessMapRegionCapability(
+        cap, [](KMemoryRegionType region_type, KMemoryPermission perm) -> Result {
+            // R_RETURN(page_table->MapRegion(region_type, perm));
+            UNIMPLEMENTED();
+            R_SUCCEED();
+        }));
+}
+
+Result KCapabilities::CheckMapRegion(KernelCore& kernel, const u32 cap) {
+    // Check that each region has a physical backing store.
+    R_RETURN(ProcessMapRegionCapability(
+        cap, [&](KMemoryRegionType region_type, KMemoryPermission perm) -> Result {
+            R_UNLESS(kernel.MemoryLayout().GetPhysicalMemoryRegionTree().FindFirstDerived(
+                         region_type) != nullptr,
+                     ResultOutOfRange);
+            R_SUCCEED();
+        }));
+}
+
+Result KCapabilities::SetInterruptPairCapability(const u32 cap) {
+    // Extract interrupts.
+    const InterruptPair pack{cap};
+    const std::array<u32, 2> ids{pack.interrupt_id0, pack.interrupt_id1};
+
+    for (size_t i = 0; i < ids.size(); i++) {
+        if (ids[i] != PaddingInterruptId) {
+            UNIMPLEMENTED();
+            // R_UNLESS(Kernel::GetInterruptManager().IsInterruptDefined(ids[i]), ResultOutOfRange);
+            // R_UNLESS(this->SetInterruptPermitted(ids[i]), ResultOutOfRange);
+        }
+    }
+
+    R_SUCCEED();
+}
+
+Result KCapabilities::SetProgramTypeCapability(const u32 cap) {
+    // Validate.
+    const ProgramType pack{cap};
+    R_UNLESS(pack.reserved == 0, ResultReservedUsed);
+
+    m_program_type = pack.type;
+    R_SUCCEED();
+}
+
+Result KCapabilities::SetKernelVersionCapability(const u32 cap) {
+    // Ensure we haven't set our version before.
+    R_UNLESS(KernelVersion{m_intended_kernel_version}.major_version == 0, ResultInvalidArgument);
+
+    // Set, ensure that we set a valid version.
+    m_intended_kernel_version = cap;
+    R_UNLESS(KernelVersion{m_intended_kernel_version}.major_version != 0, ResultInvalidArgument);
+
+    R_SUCCEED();
+}
+
+Result KCapabilities::SetHandleTableCapability(const u32 cap) {
+    // Validate.
+    const HandleTable pack{cap};
+    R_UNLESS(pack.reserved == 0, ResultReservedUsed);
+
+    m_handle_table_size = pack.size;
+    R_SUCCEED();
+}
+
+Result KCapabilities::SetDebugFlagsCapability(const u32 cap) {
+    // Validate.
+    const DebugFlags pack{cap};
+    R_UNLESS(pack.reserved == 0, ResultReservedUsed);
+
+    DebugFlags debug_capabilities{m_debug_capabilities};
+    debug_capabilities.allow_debug.Assign(pack.allow_debug);
+    debug_capabilities.force_debug.Assign(pack.force_debug);
+    m_debug_capabilities = debug_capabilities.raw;
+
+    R_SUCCEED();
+}
+
+Result KCapabilities::SetCapability(const u32 cap, u32& set_flags, u32& set_svc,
+                                    KPageTable* page_table) {
+    // Validate this is a capability we can act on.
+    const auto type = GetCapabilityType(cap);
+    R_UNLESS(type != CapabilityType::Invalid, ResultInvalidArgument);
+
+    // If the type is padding, we have no work to do.
+    R_SUCCEED_IF(type == CapabilityType::Padding);
+
+    // Check that we haven't already processed this capability.
+    const auto flag = GetCapabilityFlag(type);
+    R_UNLESS(((set_flags & InitializeOnceFlags) & flag) == 0, ResultInvalidCombination);
+    set_flags |= flag;
+
+    // Process the capability.
+    switch (type) {
+    case CapabilityType::CorePriority:
+        R_RETURN(this->SetCorePriorityCapability(cap));
+    case CapabilityType::SyscallMask:
+        R_RETURN(this->SetSyscallMaskCapability(cap, set_svc));
+    case CapabilityType::MapIoPage:
+        R_RETURN(this->MapIoPage_(cap, page_table));
+    case CapabilityType::MapRegion:
+        R_RETURN(this->MapRegion_(cap, page_table));
+    case CapabilityType::InterruptPair:
+        R_RETURN(this->SetInterruptPairCapability(cap));
+    case CapabilityType::ProgramType:
+        R_RETURN(this->SetProgramTypeCapability(cap));
+    case CapabilityType::KernelVersion:
+        R_RETURN(this->SetKernelVersionCapability(cap));
+    case CapabilityType::HandleTable:
+        R_RETURN(this->SetHandleTableCapability(cap));
+    case CapabilityType::DebugFlags:
+        R_RETURN(this->SetDebugFlagsCapability(cap));
+    default:
+        R_THROW(ResultInvalidArgument);
+    }
+}
+
+Result KCapabilities::SetCapabilities(std::span<const u32> caps, KPageTable* page_table) {
+    u32 set_flags = 0, set_svc = 0;
+
+    for (size_t i = 0; i < caps.size(); i++) {
+        const u32 cap{caps[i]};
+
+        if (GetCapabilityType(cap) == CapabilityType::MapRange) {
+            // Check that the pair cap exists.
+            R_UNLESS((++i) < caps.size(), ResultInvalidCombination);
+
+            // Check the pair cap is a map range cap.
+            const u32 size_cap{caps[i]};
+            R_UNLESS(GetCapabilityType(size_cap) == CapabilityType::MapRange,
+                     ResultInvalidCombination);
+
+            // Map the range.
+            R_TRY(this->MapRange_(cap, size_cap, page_table));
+        } else {
+            R_TRY(this->SetCapability(cap, set_flags, set_svc, page_table));
+        }
+    }
+
+    R_SUCCEED();
+}
+
+Result KCapabilities::CheckCapabilities(KernelCore& kernel, std::span<const u32> caps) {
+    for (auto cap : caps) {
+        // Check the capability refers to a valid region.
+        if (GetCapabilityType(cap) == CapabilityType::MapRegion) {
+            R_TRY(CheckMapRegion(kernel, cap));
+        }
+    }
+
+    R_SUCCEED();
+}
+
+} // namespace Kernel

src/core/hle/kernel/k_capabilities.h

@@ -0,0 +1,295 @@
+
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#pragma once
+
+#include <bitset>
+#include <span>
+
+#include "common/bit_field.h"
+#include "common/common_types.h"
+
+#include "core/hle/kernel/svc_types.h"
+#include "core/hle/result.h"
+
+namespace Kernel {
+
+class KPageTable;
+class KernelCore;
+
+class KCapabilities {
+public:
+    constexpr explicit KCapabilities() = default;
+
+    Result InitializeForKIP(std::span<const u32> kern_caps, KPageTable* page_table);
+    Result InitializeForUser(std::span<const u32> user_caps, KPageTable* page_table);
+
+    static Result CheckCapabilities(KernelCore& kernel, std::span<const u32> user_caps);
+
+    constexpr u64 GetCoreMask() const {
+        return m_core_mask;
+    }
+
+    constexpr u64 GetPhysicalCoreMask() const {
+        return m_phys_core_mask;
+    }
+
+    constexpr u64 GetPriorityMask() const {
+        return m_priority_mask;
+    }
+
+    constexpr s32 GetHandleTableSize() const {
+        return m_handle_table_size;
+    }
+
+    constexpr const Svc::SvcAccessFlagSet& GetSvcPermissions() const {
+        return m_svc_access_flags;
+    }
+
+    constexpr bool IsPermittedSvc(u32 id) const {
+        return (id < m_svc_access_flags.size()) && m_svc_access_flags[id];
+    }
+
+    constexpr bool IsPermittedInterrupt(u32 id) const {
+        return (id < m_irq_access_flags.size()) && m_irq_access_flags[id];
+    }
+
+    constexpr bool IsPermittedDebug() const {
+        return DebugFlags{m_debug_capabilities}.allow_debug.Value() != 0;
+    }
+
+    constexpr bool CanForceDebug() const {
+        return DebugFlags{m_debug_capabilities}.force_debug.Value() != 0;
+    }
+
+    constexpr u32 GetIntendedKernelMajorVersion() const {
+        return KernelVersion{m_intended_kernel_version}.major_version;
+    }
+
+    constexpr u32 GetIntendedKernelMinorVersion() const {
+        return KernelVersion{m_intended_kernel_version}.minor_version;
+    }
+
+private:
+    static constexpr size_t InterruptIdCount = 0x400;
+    using InterruptFlagSet = std::bitset<InterruptIdCount>;
+
+    enum class CapabilityType : u32 {
+        CorePriority = (1U << 3) - 1,
+        SyscallMask = (1U << 4) - 1,
+        MapRange = (1U << 6) - 1,
+        MapIoPage = (1U << 7) - 1,
+        MapRegion = (1U << 10) - 1,
+        InterruptPair = (1U << 11) - 1,
+        ProgramType = (1U << 13) - 1,
+        KernelVersion = (1U << 14) - 1,
+        HandleTable = (1U << 15) - 1,
+        DebugFlags = (1U << 16) - 1,
+
+        Invalid = 0U,
+        Padding = ~0U,
+    };
+
+    using RawCapabilityValue = u32;
+
+    static constexpr CapabilityType GetCapabilityType(const RawCapabilityValue value) {
+        return static_cast<CapabilityType>((~value & (value + 1)) - 1);
+    }
+
+    static constexpr u32 GetCapabilityFlag(CapabilityType type) {
+        return static_cast<u32>(type) + 1;
+    }
+
+    template <CapabilityType Type>
+    static constexpr inline u32 CapabilityFlag = static_cast<u32>(Type) + 1;
+
+    template <CapabilityType Type>
+    static constexpr inline u32 CapabilityId = std::countr_zero(CapabilityFlag<Type>);
+
+    union CorePriority {
+        static_assert(CapabilityId<CapabilityType::CorePriority> + 1 == 4);
+
+        RawCapabilityValue raw;
+        BitField<0, 4, CapabilityType> id;
+        BitField<4, 6, u32> lowest_thread_priority;
+        BitField<10, 6, u32> highest_thread_priority;
+        BitField<16, 8, u32> minimum_core_id;
+        BitField<24, 8, u32> maximum_core_id;
+    };
+
+    union SyscallMask {
+        static_assert(CapabilityId<CapabilityType::SyscallMask> + 1 == 5);
+
+        RawCapabilityValue raw;
+        BitField<0, 5, CapabilityType> id;
+        BitField<5, 24, u32> mask;
+        BitField<29, 3, u32> index;
+    };
+
+    // #undef MESOSPHERE_ENABLE_LARGE_PHYSICAL_ADDRESS_CAPABILITIES
+    static constexpr u64 PhysicalMapAllowedMask = (1ULL << 36) - 1;
+
+    union MapRange {
+        static_assert(CapabilityId<CapabilityType::MapRange> + 1 == 7);
+
+        RawCapabilityValue raw;
+        BitField<0, 7, CapabilityType> id;
+        BitField<7, 24, u32> address;
+        BitField<31, 1, u32> read_only;
+    };
+
+    union MapRangeSize {
+        static_assert(CapabilityId<CapabilityType::MapRange> + 1 == 7);
+
+        RawCapabilityValue raw;
+        BitField<0, 7, CapabilityType> id;
+        BitField<7, 20, u32> pages;
+        BitField<27, 4, u32> reserved;
+        BitField<31, 1, u32> normal;
+    };
+
+    union MapIoPage {
+        static_assert(CapabilityId<CapabilityType::MapIoPage> + 1 == 8);
+
+        RawCapabilityValue raw;
+        BitField<0, 8, CapabilityType> id;
+        BitField<8, 24, u32> address;
+    };
+
+    enum class RegionType : u32 {
+        NoMapping = 0,
+        KernelTraceBuffer = 1,
+        OnMemoryBootImage = 2,
+        DTB = 3,
+    };
+
+    union MapRegion {
+        static_assert(CapabilityId<CapabilityType::MapRegion> + 1 == 11);
+
+        RawCapabilityValue raw;
+        BitField<0, 11, CapabilityType> id;
+        BitField<11, 6, RegionType> region0;
+        BitField<17, 1, u32> read_only0;
+        BitField<18, 6, RegionType> region1;
+        BitField<24, 1, u32> read_only1;
+        BitField<25, 6, RegionType> region2;
+        BitField<31, 1, u32> read_only2;
+    };
+
+    union InterruptPair {
+        static_assert(CapabilityId<CapabilityType::InterruptPair> + 1 == 12);
+
+        RawCapabilityValue raw;
+        BitField<0, 12, CapabilityType> id;
+        BitField<12, 10, u32> interrupt_id0;
+        BitField<22, 10, u32> interrupt_id1;
+    };
+
+    union ProgramType {
+        static_assert(CapabilityId<CapabilityType::ProgramType> + 1 == 14);
+
+        RawCapabilityValue raw;
+        BitField<0, 14, CapabilityType> id;
+        BitField<14, 3, u32> type;
+        BitField<17, 15, u32> reserved;
+    };
+
+    union KernelVersion {
+        static_assert(CapabilityId<CapabilityType::KernelVersion> + 1 == 15);
+
+        RawCapabilityValue raw;
+        BitField<0, 15, CapabilityType> id;
+        BitField<15, 4, u32> major_version;
+        BitField<19, 13, u32> minor_version;
+    };
+
+    union HandleTable {
+        static_assert(CapabilityId<CapabilityType::HandleTable> + 1 == 16);
+
+        RawCapabilityValue raw;
+        BitField<0, 16, CapabilityType> id;
+        BitField<16, 10, u32> size;
+        BitField<26, 6, u32> reserved;
+    };
+
+    union DebugFlags {
+        static_assert(CapabilityId<CapabilityType::DebugFlags> + 1 == 17);
+
+        RawCapabilityValue raw;
+        BitField<0, 17, CapabilityType> id;
+        BitField<17, 1, u32> allow_debug;
+        BitField<18, 1, u32> force_debug;
+        BitField<19, 13, u32> reserved;
+    };
+
+    static_assert(sizeof(CorePriority) == 4);
+    static_assert(sizeof(SyscallMask) == 4);
+    static_assert(sizeof(MapRange) == 4);
+    static_assert(sizeof(MapRangeSize) == 4);
+    static_assert(sizeof(MapIoPage) == 4);
+    static_assert(sizeof(MapRegion) == 4);
+    static_assert(sizeof(InterruptPair) == 4);
+    static_assert(sizeof(ProgramType) == 4);
+    static_assert(sizeof(KernelVersion) == 4);
+    static_assert(sizeof(HandleTable) == 4);
+    static_assert(sizeof(DebugFlags) == 4);
+
+    static constexpr u32 InitializeOnceFlags =
+        CapabilityFlag<CapabilityType::CorePriority> | CapabilityFlag<CapabilityType::ProgramType> |
+        CapabilityFlag<CapabilityType::KernelVersion> |
+        CapabilityFlag<CapabilityType::HandleTable> | CapabilityFlag<CapabilityType::DebugFlags>;
+
+    static const u32 PaddingInterruptId = 0x3FF;
+    static_assert(PaddingInterruptId < InterruptIdCount);
+
+private:
+    constexpr bool SetSvcAllowed(u32 id) {
+        if (id < m_svc_access_flags.size()) [[likely]] {
+            m_svc_access_flags[id] = true;
+            return true;
+        } else {
+            return false;
+        }
+    }
+
+    constexpr bool SetInterruptPermitted(u32 id) {
+        if (id < m_irq_access_flags.size()) [[likely]] {
+            m_irq_access_flags[id] = true;
+            return true;
+        } else {
+            return false;
+        }
+    }
+
+    Result SetCorePriorityCapability(const u32 cap);
+    Result SetSyscallMaskCapability(const u32 cap, u32& set_svc);
+    Result MapRange_(const u32 cap, const u32 size_cap, KPageTable* page_table);
+    Result MapIoPage_(const u32 cap, KPageTable* page_table);
+    Result MapRegion_(const u32 cap, KPageTable* page_table);
+    Result SetInterruptPairCapability(const u32 cap);
+    Result SetProgramTypeCapability(const u32 cap);
+    Result SetKernelVersionCapability(const u32 cap);
+    Result SetHandleTableCapability(const u32 cap);
+    Result SetDebugFlagsCapability(const u32 cap);
+
+    template <typename F>
+    static Result ProcessMapRegionCapability(const u32 cap, F f);
+    static Result CheckMapRegion(KernelCore& kernel, const u32 cap);
+
+    Result SetCapability(const u32 cap, u32& set_flags, u32& set_svc, KPageTable* page_table);
+    Result SetCapabilities(std::span<const u32> caps, KPageTable* page_table);
+
+private:
+    Svc::SvcAccessFlagSet m_svc_access_flags{};
+    InterruptFlagSet m_irq_access_flags{};
+    u64 m_core_mask{};
+    u64 m_phys_core_mask{};
+    u64 m_priority_mask{};
+    u32 m_debug_capabilities{};
+    s32 m_handle_table_size{};
+    u32 m_intended_kernel_version{};
+    u32 m_program_type{};
+};
+
+} // namespace Kernel

src/core/hle/kernel/svc_types.h

@@ -3,6 +3,8 @@
 
 #pragma once
 
+#include <bitset>
+
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 
@@ -592,4 +594,7 @@ struct CreateProcessParameter {
 };
 static_assert(sizeof(CreateProcessParameter) == 0x30);
 
+constexpr size_t NumSupervisorCalls = 0xC0;
+using SvcAccessFlagSet = std::bitset<NumSupervisorCalls>;
+
 } // namespace Kernel::Svc

src/core/hle/kernel/svc_version.h

@@ -0,0 +1,58 @@
+// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#pragma once
+
+#include "common/bit_field.h"
+#include "common/common_types.h"
+#include "common/literals.h"
+
+namespace Kernel::Svc {
+
+constexpr inline u32 ConvertToSvcMajorVersion(u32 sdk) {
+    return sdk + 4;
+}
+constexpr inline u32 ConvertToSdkMajorVersion(u32 svc) {
+    return svc - 4;
+}
+
+constexpr inline u32 ConvertToSvcMinorVersion(u32 sdk) {
+    return sdk;
+}
+constexpr inline u32 ConvertToSdkMinorVersion(u32 svc) {
+    return svc;
+}
+
+union KernelVersion {
+    u32 value;
+    BitField<0, 4, u32> minor_version;
+    BitField<4, 13, u32> major_version;
+};
+
+constexpr inline u32 EncodeKernelVersion(u32 major, u32 minor) {
+    return decltype(KernelVersion::minor_version)::FormatValue(minor) |
+           decltype(KernelVersion::major_version)::FormatValue(major);
+}
+
+constexpr inline u32 GetKernelMajorVersion(u32 encoded) {
+    return std::bit_cast<decltype(KernelVersion::major_version)>(encoded).Value();
+}
+
+constexpr inline u32 GetKernelMinorVersion(u32 encoded) {
+    return std::bit_cast<decltype(KernelVersion::minor_version)>(encoded).Value();
+}
+
+// Nintendo doesn't support programs targeting SVC versions < 3.0.
+constexpr inline u32 RequiredKernelMajorVersion = 3;
+constexpr inline u32 RequiredKernelMinorVersion = 0;
+constexpr inline u32 RequiredKernelVersion =
+    EncodeKernelVersion(RequiredKernelMajorVersion, RequiredKernelMinorVersion);
+
+// This is the highest SVC version supported, to be updated on new kernel releases.
+// NOTE: Official kernel versions have SVC major = SDK major + 4, SVC minor = SDK minor.
+constexpr inline u32 SupportedKernelMajorVersion = ConvertToSvcMajorVersion(15);
+constexpr inline u32 SupportedKernelMinorVersion = ConvertToSvcMinorVersion(3);
+constexpr inline u32 SupportedKernelVersion =
+    EncodeKernelVersion(SupportedKernelMajorVersion, SupportedKernelMinorVersion);
+
+} // namespace Kernel::Svc

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

@@ -1124,7 +1124,7 @@ void IStorageAccessor::Write(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
 
     const u64 offset{rp.Pop<u64>()};
-    const std::vector<u8> data{ctx.ReadBuffer()};
+    const auto data{ctx.ReadBuffer()};
     const std::size_t size{std::min<u64>(data.size(), backing.GetSize() - offset)};
 
     LOG_DEBUG(Service_AM, "called, offset={}, size={}", offset, size);

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

@@ -112,7 +112,7 @@ private:
     void RequestUpdate(Kernel::HLERequestContext& ctx) {
         LOG_TRACE(Service_Audio, "called");
 
-        std::vector<u8> input{ctx.ReadBuffer(0)};
+        const auto input{ctx.ReadBuffer(0)};
 
         // These buffers are written manually to avoid an issue with WriteBuffer throwing errors for
         // checking size 0. Performance size is 0 for most games.

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

@@ -93,7 +93,7 @@ private:
         ctx.WriteBuffer(samples);
     }
 
-    bool DecodeOpusData(u32& consumed, u32& sample_count, const std::vector<u8>& input,
+    bool DecodeOpusData(u32& consumed, u32& sample_count, std::span<const u8> input,
                         std::vector<opus_int16>& output, u64* out_performance_time) const {
         const auto start_time = std::chrono::steady_clock::now();
         const std::size_t raw_output_sz = output.size() * sizeof(opus_int16);

src/core/hle/service/es/es.cpp

@@ -122,7 +122,7 @@ private:
 
     void ImportTicket(Kernel::HLERequestContext& ctx) {
         const auto ticket = ctx.ReadBuffer();
-        const auto cert = ctx.ReadBuffer(1);
+        [[maybe_unused]] const auto cert = ctx.ReadBuffer(1);
 
         if (ticket.size() < sizeof(Core::Crypto::Ticket)) {
             LOG_ERROR(Service_ETicket, "The input buffer is not large enough!");

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

@@ -190,7 +190,7 @@ private:
             return;
         }
 
-        const std::vector<u8> data = ctx.ReadBuffer();
+        const auto data = ctx.ReadBuffer();
 
         ASSERT_MSG(
             static_cast<s64>(data.size()) <= length,
@@ -401,11 +401,8 @@ public:
     }
 
     void RenameFile(Kernel::HLERequestContext& ctx) {
-        std::vector<u8> buffer = ctx.ReadBuffer(0);
+        const std::string src_name = Common::StringFromBuffer(ctx.ReadBuffer(0));
-        const std::string src_name = Common::StringFromBuffer(buffer);
+        const std::string dst_name = Common::StringFromBuffer(ctx.ReadBuffer(1));
-
-        buffer = ctx.ReadBuffer(1);
-        const std::string dst_name = Common::StringFromBuffer(buffer);
 
         LOG_DEBUG(Service_FS, "called. file '{}' to file '{}'", src_name, dst_name);
 

src/core/hle/service/glue/arp.cpp

@@ -228,7 +228,8 @@ private:
             return;
         }
 
-        control = ctx.ReadBuffer();
+        // TODO: Can this be a span?
+        control = ctx.ReadBufferCopy();
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(ResultSuccess);

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

@@ -428,6 +428,9 @@ void Controller_NPad::RequestPadStateUpdate(Core::HID::NpadIdType npad_id) {
         return;
     }
 
+    // This function is unique to yuzu for the turbo buttons to work properly
+    controller.device->TurboButtonUpdate();
+
     auto& pad_entry = controller.npad_pad_state;
     auto& trigger_entry = controller.npad_trigger_state;
     const auto button_state = controller.device->GetNpadButtons();
@@ -755,11 +758,12 @@ Core::HID::NpadStyleTag Controller_NPad::GetSupportedStyleSet() const {
     return hid_core.GetSupportedStyleTag();
 }
 
-void Controller_NPad::SetSupportedNpadIdTypes(u8* data, std::size_t length) {
+void Controller_NPad::SetSupportedNpadIdTypes(std::span<const u8> data) {
+    const auto length = data.size();
     ASSERT(length > 0 && (length % sizeof(u32)) == 0);
     supported_npad_id_types.clear();
     supported_npad_id_types.resize(length / sizeof(u32));
-    std::memcpy(supported_npad_id_types.data(), data, length);
+    std::memcpy(supported_npad_id_types.data(), data.data(), length);
 }
 
 void Controller_NPad::GetSupportedNpadIdTypes(u32* data, std::size_t max_length) {

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

@@ -6,6 +6,7 @@
 #include <array>
 #include <atomic>
 #include <mutex>
+#include <span>
 
 #include "common/bit_field.h"
 #include "common/common_types.h"
@@ -95,7 +96,7 @@ public:
     void SetSupportedStyleSet(Core::HID::NpadStyleTag style_set);
     Core::HID::NpadStyleTag GetSupportedStyleSet() const;
 
-    void SetSupportedNpadIdTypes(u8* data, std::size_t length);
+    void SetSupportedNpadIdTypes(std::span<const u8> data);
     void GetSupportedNpadIdTypes(u32* data, std::size_t max_length);
     std::size_t GetSupportedNpadIdTypesSize() const;
 

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

@@ -1026,7 +1026,7 @@ void Hid::SetSupportedNpadIdType(Kernel::HLERequestContext& ctx) {
     const auto applet_resource_user_id{rp.Pop<u64>()};
 
     applet_resource->GetController<Controller_NPad>(HidController::NPad)
-        .SetSupportedNpadIdTypes(ctx.ReadBuffer().data(), ctx.GetReadBufferSize());
+        .SetSupportedNpadIdTypes(ctx.ReadBuffer());
 
     LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id);
 
@@ -2104,7 +2104,7 @@ void Hid::WritePalmaRgbLedPatternEntry(Kernel::HLERequestContext& ctx) {
     const auto connection_handle{rp.PopRaw<Controller_Palma::PalmaConnectionHandle>()};
     const auto unknown{rp.Pop<u64>()};
 
-    const auto buffer = ctx.ReadBuffer();
+    [[maybe_unused]] const auto buffer = ctx.ReadBuffer();
 
     LOG_WARNING(Service_HID, "(STUBBED) called, connection_handle={}, unknown={}",
                 connection_handle.npad_id, unknown);

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

@@ -4,6 +4,7 @@
 #pragma once
 
 #include <array>
+#include <span>
 #include "common/common_types.h"
 #include "core/hle/result.h"
 
@@ -150,7 +151,7 @@ public:
     }
 
     // Assigns a command from data
-    virtual bool SetCommand(const std::vector<u8>& data) {
+    virtual bool SetCommand(std::span<const u8> data) {
         return {};
     }
 

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

@@ -116,7 +116,7 @@ std::vector<u8> RingController::GetReply() const {
     }
 }
 
-bool RingController::SetCommand(const std::vector<u8>& data) {
+bool RingController::SetCommand(std::span<const u8> data) {
     if (data.size() < 4) {
         LOG_ERROR(Service_HID, "Command size not supported {}", data.size());
         command = RingConCommands::Error;

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

@@ -4,6 +4,7 @@
 #pragma once
 
 #include <array>
+#include <span>
 
 #include "common/common_types.h"
 #include "core/hle/service/hid/hidbus/hidbus_base.h"
@@ -31,7 +32,7 @@ public:
     u8 GetDeviceId() const override;
 
     // Assigns a command from data
-    bool SetCommand(const std::vector<u8>& data) override;
+    bool SetCommand(std::span<const u8> data) override;
 
     // Returns a reply from a command
     std::vector<u8> GetReply() const override;

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

@@ -42,7 +42,7 @@ std::vector<u8> Starlink::GetReply() const {
     return {};
 }
 
-bool Starlink::SetCommand(const std::vector<u8>& data) {
+bool Starlink::SetCommand(std::span<const u8> data) {
     LOG_ERROR(Service_HID, "Command not implemented");
     return false;
 }

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

@@ -29,7 +29,7 @@ public:
     u8 GetDeviceId() const override;
 
     // Assigns a command from data
-    bool SetCommand(const std::vector<u8>& data) override;
+    bool SetCommand(std::span<const u8> data) override;
 
     // Returns a reply from a command
     std::vector<u8> GetReply() const override;

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

@@ -43,7 +43,7 @@ std::vector<u8> HidbusStubbed::GetReply() const {
     return {};
 }
 
-bool HidbusStubbed::SetCommand(const std::vector<u8>& data) {
+bool HidbusStubbed::SetCommand(std::span<const u8> data) {
     LOG_ERROR(Service_HID, "Command not implemented");
     return false;
 }

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

@@ -29,7 +29,7 @@ public:
     u8 GetDeviceId() const override;
 
     // Assigns a command from data
-    bool SetCommand(const std::vector<u8>& data) override;
+    bool SetCommand(std::span<const u8> data) override;
 
     // Returns a reply from a command
     std::vector<u8> GetReply() const override;

src/core/hle/service/jit/jit.cpp

@@ -62,7 +62,7 @@ public:
         const auto parameters{rp.PopRaw<InputParameters>()};
 
         // Optional input/output buffers
-        std::vector<u8> input_buffer{ctx.CanReadBuffer() ? ctx.ReadBuffer() : std::vector<u8>()};
+        const auto input_buffer{ctx.CanReadBuffer() ? ctx.ReadBuffer() : std::span<const u8>()};
         std::vector<u8> output_buffer(ctx.CanWriteBuffer() ? ctx.GetWriteBufferSize() : 0);
 
         // Function call prototype:
@@ -132,7 +132,7 @@ public:
         const auto command{rp.PopRaw<u64>()};
 
         // Optional input/output buffers
-        std::vector<u8> input_buffer{ctx.CanReadBuffer() ? ctx.ReadBuffer() : std::vector<u8>()};
+        const auto input_buffer{ctx.CanReadBuffer() ? ctx.ReadBuffer() : std::span<const u8>()};
         std::vector<u8> output_buffer(ctx.CanWriteBuffer() ? ctx.GetWriteBufferSize() : 0);
 
         // Function call prototype:

src/core/hle/service/ldn/ldn.cpp

@@ -427,7 +427,7 @@ public:
     }
 
     void SetAdvertiseData(Kernel::HLERequestContext& ctx) {
-        std::vector<u8> read_buffer = ctx.ReadBuffer();
+        const auto read_buffer = ctx.ReadBuffer();
 
         IPC::ResponseBuilder rb{ctx, 2};
         rb.Push(lan_discovery.SetAdvertiseData(read_buffer));
@@ -479,7 +479,7 @@ public:
                  parameters.security_config.passphrase_size,
                  parameters.security_config.security_mode, parameters.local_communication_version);
 
-        const std::vector<u8> read_buffer = ctx.ReadBuffer();
+        const auto read_buffer = ctx.ReadBuffer();
         if (read_buffer.size() != sizeof(NetworkInfo)) {
             LOG_ERROR(Frontend, "NetworkInfo doesn't match read_buffer size!");
             IPC::ResponseBuilder rb{ctx, 2};

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

@@ -3,7 +3,9 @@
 
 #pragma once
 
+#include <span>
 #include <vector>
+
 #include "common/common_types.h"
 #include "core/hle/service/nvdrv/nvdata.h"
 
@@ -31,7 +33,7 @@ public:
      * @param output A buffer where the output data will be written to.
      * @returns The result code of the ioctl.
      */
-    virtual NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    virtual NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                             std::vector<u8>& output) = 0;
 
     /**
@@ -42,8 +44,8 @@ public:
      * @param output A buffer where the output data will be written to.
      * @returns The result code of the ioctl.
      */
-    virtual NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    virtual NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                            const std::vector<u8>& inline_input, std::vector<u8>& output) = 0;
+                            std::span<const u8> inline_input, std::vector<u8>& output) = 0;
 
     /**
      * Handles an ioctl3 request.
@@ -53,7 +55,7 @@ public:
      * @param inline_output A buffer where the inlined output data will be written to.
      * @returns The result code of the ioctl.
      */
-    virtual NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    virtual NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                             std::vector<u8>& output, std::vector<u8>& inline_output) = 0;
 
     /**

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

@@ -17,19 +17,19 @@ nvdisp_disp0::nvdisp_disp0(Core::System& system_, NvCore::Container& core)
     : nvdevice{system_}, container{core}, nvmap{core.GetNvMapFile()} {}
 nvdisp_disp0::~nvdisp_disp0() = default;
 
-NvResult nvdisp_disp0::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvdisp_disp0::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                               std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
 }
 
-NvResult nvdisp_disp0::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvdisp_disp0::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                              const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                              std::span<const u8> inline_input, std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
 }
 
-NvResult nvdisp_disp0::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvdisp_disp0::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                               std::vector<u8>& output, std::vector<u8>& inline_output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;

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

@@ -25,12 +25,12 @@ public:
     explicit nvdisp_disp0(Core::System& system_, NvCore::Container& core);
     ~nvdisp_disp0() override;
 
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                     std::vector<u8>& output) override;
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output) override;
+                    std::span<const u8> inline_input, std::vector<u8>& output) override;
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output) override;
+                    std::vector<u8>& inline_output) override;
 
     void OnOpen(DeviceFD fd) override;
     void OnClose(DeviceFD fd) override;

src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp

@@ -27,7 +27,7 @@ nvhost_as_gpu::nvhost_as_gpu(Core::System& system_, Module& module_, NvCore::Con
 
 nvhost_as_gpu::~nvhost_as_gpu() = default;
 
-NvResult nvhost_as_gpu::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_as_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                                std::vector<u8>& output) {
     switch (command.group) {
     case 'A':
@@ -60,13 +60,13 @@ NvResult nvhost_as_gpu::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_as_gpu::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_as_gpu::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                               const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                               std::span<const u8> inline_input, std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_as_gpu::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_as_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                                std::vector<u8>& output, std::vector<u8>& inline_output) {
     switch (command.group) {
     case 'A':
@@ -87,7 +87,7 @@ NvResult nvhost_as_gpu::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>
 void nvhost_as_gpu::OnOpen(DeviceFD fd) {}
 void nvhost_as_gpu::OnClose(DeviceFD fd) {}
 
-NvResult nvhost_as_gpu::AllocAsEx(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_as_gpu::AllocAsEx(std::span<const u8> input, std::vector<u8>& output) {
     IoctlAllocAsEx params{};
     std::memcpy(&params, input.data(), input.size());
 
@@ -141,7 +141,7 @@ NvResult nvhost_as_gpu::AllocAsEx(const std::vector<u8>& input, std::vector<u8>&
     return NvResult::Success;
 }
 
-NvResult nvhost_as_gpu::AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_as_gpu::AllocateSpace(std::span<const u8> input, std::vector<u8>& output) {
     IoctlAllocSpace params{};
     std::memcpy(&params, input.data(), input.size());
 
@@ -220,7 +220,7 @@ void nvhost_as_gpu::FreeMappingLocked(u64 offset) {
     mapping_map.erase(offset);
 }
 
-NvResult nvhost_as_gpu::FreeSpace(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_as_gpu::FreeSpace(std::span<const u8> input, std::vector<u8>& output) {
     IoctlFreeSpace params{};
     std::memcpy(&params, input.data(), input.size());
 
@@ -266,7 +266,7 @@ NvResult nvhost_as_gpu::FreeSpace(const std::vector<u8>& input, std::vector<u8>&
     return NvResult::Success;
 }
 
-NvResult nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_as_gpu::Remap(std::span<const u8> input, std::vector<u8>& output) {
     const auto num_entries = input.size() / sizeof(IoctlRemapEntry);
 
     LOG_DEBUG(Service_NVDRV, "called, num_entries=0x{:X}", num_entries);
@@ -320,7 +320,7 @@ NvResult nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& out
     return NvResult::Success;
 }
 
-NvResult nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_as_gpu::MapBufferEx(std::span<const u8> input, std::vector<u8>& output) {
     IoctlMapBufferEx params{};
     std::memcpy(&params, input.data(), input.size());
 
@@ -424,7 +424,7 @@ NvResult nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8
     return NvResult::Success;
 }
 
-NvResult nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_as_gpu::UnmapBuffer(std::span<const u8> input, std::vector<u8>& output) {
     IoctlUnmapBuffer params{};
     std::memcpy(&params, input.data(), input.size());
 
@@ -463,7 +463,7 @@ NvResult nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8
     return NvResult::Success;
 }
 
-NvResult nvhost_as_gpu::BindChannel(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_as_gpu::BindChannel(std::span<const u8> input, std::vector<u8>& output) {
     IoctlBindChannel params{};
     std::memcpy(&params, input.data(), input.size());
     LOG_DEBUG(Service_NVDRV, "called, fd={:X}", params.fd);
@@ -492,7 +492,7 @@ void nvhost_as_gpu::GetVARegionsImpl(IoctlGetVaRegions& params) {
     };
 }
 
-NvResult nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_as_gpu::GetVARegions(std::span<const u8> input, std::vector<u8>& output) {
     IoctlGetVaRegions params{};
     std::memcpy(&params, input.data(), input.size());
 
@@ -511,7 +511,7 @@ NvResult nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u
     return NvResult::Success;
 }
 
-NvResult nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& output,
+NvResult nvhost_as_gpu::GetVARegions(std::span<const u8> input, std::vector<u8>& output,
                                      std::vector<u8>& inline_output) {
     IoctlGetVaRegions params{};
     std::memcpy(&params, input.data(), input.size());

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

@@ -47,12 +47,12 @@ public:
     explicit nvhost_as_gpu(Core::System& system_, Module& module, NvCore::Container& core);
     ~nvhost_as_gpu() override;
 
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                     std::vector<u8>& output) override;
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output) override;
+                    std::span<const u8> inline_input, std::vector<u8>& output) override;
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output) override;
+                    std::vector<u8>& inline_output) override;
 
     void OnOpen(DeviceFD fd) override;
     void OnClose(DeviceFD fd) override;
@@ -138,17 +138,17 @@ private:
     static_assert(sizeof(IoctlGetVaRegions) == 16 + sizeof(VaRegion) * 2,
                   "IoctlGetVaRegions is incorrect size");
 
-    NvResult AllocAsEx(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult AllocAsEx(std::span<const u8> input, std::vector<u8>& output);
-    NvResult AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult AllocateSpace(std::span<const u8> input, std::vector<u8>& output);
-    NvResult Remap(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult Remap(std::span<const u8> input, std::vector<u8>& output);
-    NvResult MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult MapBufferEx(std::span<const u8> input, std::vector<u8>& output);
-    NvResult UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult UnmapBuffer(std::span<const u8> input, std::vector<u8>& output);
-    NvResult FreeSpace(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult FreeSpace(std::span<const u8> input, std::vector<u8>& output);
-    NvResult BindChannel(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult BindChannel(std::span<const u8> input, std::vector<u8>& output);
 
     void GetVARegionsImpl(IoctlGetVaRegions& params);
-    NvResult GetVARegions(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult GetVARegions(std::span<const u8> input, std::vector<u8>& output);
-    NvResult GetVARegions(const std::vector<u8>& input, std::vector<u8>& output,
+    NvResult GetVARegions(std::span<const u8> input, std::vector<u8>& output,
                           std::vector<u8>& inline_output);
 
     void FreeMappingLocked(u64 offset);

src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp

@@ -34,7 +34,7 @@ nvhost_ctrl::~nvhost_ctrl() {
     }
 }
 
-NvResult nvhost_ctrl::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_ctrl::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                              std::vector<u8>& output) {
     switch (command.group) {
     case 0x0:
@@ -63,13 +63,13 @@ NvResult nvhost_ctrl::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>&
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_ctrl::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_ctrl::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                             const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                             std::span<const u8> inline_input, std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_ctrl::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_ctrl::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                              std::vector<u8>& output, std::vector<u8>& inline_outpu) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
@@ -79,7 +79,7 @@ void nvhost_ctrl::OnOpen(DeviceFD fd) {}
 
 void nvhost_ctrl::OnClose(DeviceFD fd) {}
 
-NvResult nvhost_ctrl::NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl::NvOsGetConfigU32(std::span<const u8> input, std::vector<u8>& output) {
     IocGetConfigParams params{};
     std::memcpy(&params, input.data(), sizeof(params));
     LOG_TRACE(Service_NVDRV, "called, setting={}!{}", params.domain_str.data(),
@@ -87,7 +87,7 @@ NvResult nvhost_ctrl::NvOsGetConfigU32(const std::vector<u8>& input, std::vector
     return NvResult::ConfigVarNotFound; // Returns error on production mode
 }
 
-NvResult nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output,
+NvResult nvhost_ctrl::IocCtrlEventWait(std::span<const u8> input, std::vector<u8>& output,
                                        bool is_allocation) {
     IocCtrlEventWaitParams params{};
     std::memcpy(&params, input.data(), sizeof(params));
@@ -231,7 +231,7 @@ NvResult nvhost_ctrl::FreeEvent(u32 slot) {
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl::IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl::IocCtrlEventRegister(std::span<const u8> input, std::vector<u8>& output) {
     IocCtrlEventRegisterParams params{};
     std::memcpy(&params, input.data(), sizeof(params));
     const u32 event_id = params.user_event_id;
@@ -252,8 +252,7 @@ NvResult nvhost_ctrl::IocCtrlEventRegister(const std::vector<u8>& input, std::ve
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl::IocCtrlEventUnregister(const std::vector<u8>& input,
+NvResult nvhost_ctrl::IocCtrlEventUnregister(std::span<const u8> input, std::vector<u8>& output) {
-                                             std::vector<u8>& output) {
     IocCtrlEventUnregisterParams params{};
     std::memcpy(&params, input.data(), sizeof(params));
     const u32 event_id = params.user_event_id & 0x00FF;
@@ -263,7 +262,7 @@ NvResult nvhost_ctrl::IocCtrlEventUnregister(const std::vector<u8>& input,
     return FreeEvent(event_id);
 }
 
-NvResult nvhost_ctrl::IocCtrlEventUnregisterBatch(const std::vector<u8>& input,
+NvResult nvhost_ctrl::IocCtrlEventUnregisterBatch(std::span<const u8> input,
                                                   std::vector<u8>& output) {
     IocCtrlEventUnregisterBatchParams params{};
     std::memcpy(&params, input.data(), sizeof(params));
@@ -282,7 +281,7 @@ NvResult nvhost_ctrl::IocCtrlEventUnregisterBatch(const std::vector<u8>& input,
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl::IocCtrlClearEventWait(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl::IocCtrlClearEventWait(std::span<const u8> input, std::vector<u8>& output) {
     IocCtrlEventClearParams params{};
     std::memcpy(&params, input.data(), sizeof(params));
 

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

@@ -25,12 +25,12 @@ public:
                          NvCore::Container& core);
     ~nvhost_ctrl() override;
 
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                     std::vector<u8>& output) override;
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output) override;
+                    std::span<const u8> inline_input, std::vector<u8>& output) override;
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output) override;
+                    std::vector<u8>& inline_output) override;
 
     void OnOpen(DeviceFD fd) override;
     void OnClose(DeviceFD fd) override;
@@ -186,13 +186,13 @@ private:
     static_assert(sizeof(IocCtrlEventUnregisterBatchParams) == 8,
                   "IocCtrlEventKill is incorrect size");
 
-    NvResult NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult NvOsGetConfigU32(std::span<const u8> input, std::vector<u8>& output);
-    NvResult IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output,
+    NvResult IocCtrlEventWait(std::span<const u8> input, std::vector<u8>& output,
                               bool is_allocation);
-    NvResult IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocCtrlEventRegister(std::span<const u8> input, std::vector<u8>& output);
-    NvResult IocCtrlEventUnregister(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocCtrlEventUnregister(std::span<const u8> input, std::vector<u8>& output);
-    NvResult IocCtrlEventUnregisterBatch(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocCtrlEventUnregisterBatch(std::span<const u8> input, std::vector<u8>& output);
-    NvResult IocCtrlClearEventWait(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocCtrlClearEventWait(std::span<const u8> input, std::vector<u8>& output);
 
     NvResult FreeEvent(u32 slot);
 

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

@@ -21,7 +21,7 @@ nvhost_ctrl_gpu::~nvhost_ctrl_gpu() {
     events_interface.FreeEvent(unknown_event);
 }
 
-NvResult nvhost_ctrl_gpu::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_ctrl_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                                  std::vector<u8>& output) {
     switch (command.group) {
     case 'G':
@@ -53,13 +53,13 @@ NvResult nvhost_ctrl_gpu::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_ctrl_gpu::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_ctrl_gpu::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                                 const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                                 std::span<const u8> inline_input, std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_ctrl_gpu::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_ctrl_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                                  std::vector<u8>& output, std::vector<u8>& inline_output) {
     switch (command.group) {
     case 'G':
@@ -82,8 +82,7 @@ NvResult nvhost_ctrl_gpu::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u
 void nvhost_ctrl_gpu::OnOpen(DeviceFD fd) {}
 void nvhost_ctrl_gpu::OnClose(DeviceFD fd) {}
 
-NvResult nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input,
+NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span<const u8> input, std::vector<u8>& output) {
-                                             std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called");
     IoctlCharacteristics params{};
     std::memcpy(&params, input.data(), input.size());
@@ -128,7 +127,7 @@ NvResult nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input,
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output,
+NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span<const u8> input, std::vector<u8>& output,
                                              std::vector<u8>& inline_output) {
     LOG_DEBUG(Service_NVDRV, "called");
     IoctlCharacteristics params{};
@@ -176,7 +175,7 @@ NvResult nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span<const u8> input, std::vector<u8>& output) {
     IoctlGpuGetTpcMasksArgs params{};
     std::memcpy(&params, input.data(), input.size());
     LOG_DEBUG(Service_NVDRV, "called, mask_buffer_size=0x{:X}", params.mask_buffer_size);
@@ -187,7 +186,7 @@ NvResult nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output,
+NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span<const u8> input, std::vector<u8>& output,
                                       std::vector<u8>& inline_output) {
     IoctlGpuGetTpcMasksArgs params{};
     std::memcpy(&params, input.data(), input.size());
@@ -200,7 +199,7 @@ NvResult nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl_gpu::GetActiveSlotMask(std::span<const u8> input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called");
 
     IoctlActiveSlotMask params{};
@@ -213,7 +212,7 @@ NvResult nvhost_ctrl_gpu::GetActiveSlotMask(const std::vector<u8>& input, std::v
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(std::span<const u8> input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called");
 
     IoctlZcullGetCtxSize params{};
@@ -225,7 +224,7 @@ NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(const std::vector<u8>& input, std::vec
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl_gpu::ZCullGetInfo(std::span<const u8> input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called");
 
     IoctlNvgpuGpuZcullGetInfoArgs params{};
@@ -248,7 +247,7 @@ NvResult nvhost_ctrl_gpu::ZCullGetInfo(const std::vector<u8>& input, std::vector
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl_gpu::ZBCSetTable(std::span<const u8> input, std::vector<u8>& output) {
     LOG_WARNING(Service_NVDRV, "(STUBBED) called");
 
     IoctlZbcSetTable params{};
@@ -264,7 +263,7 @@ NvResult nvhost_ctrl_gpu::ZBCSetTable(const std::vector<u8>& input, std::vector<
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl_gpu::ZBCQueryTable(std::span<const u8> input, std::vector<u8>& output) {
     LOG_WARNING(Service_NVDRV, "(STUBBED) called");
 
     IoctlZbcQueryTable params{};
@@ -274,7 +273,7 @@ NvResult nvhost_ctrl_gpu::ZBCQueryTable(const std::vector<u8>& input, std::vecto
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::FlushL2(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl_gpu::FlushL2(std::span<const u8> input, std::vector<u8>& output) {
     LOG_WARNING(Service_NVDRV, "(STUBBED) called");
 
     IoctlFlushL2 params{};
@@ -284,7 +283,7 @@ NvResult nvhost_ctrl_gpu::FlushL2(const std::vector<u8>& input, std::vector<u8>&
     return NvResult::Success;
 }
 
-NvResult nvhost_ctrl_gpu::GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_ctrl_gpu::GetGpuTime(std::span<const u8> input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called");
 
     IoctlGetGpuTime params{};

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

@@ -21,12 +21,12 @@ public:
     explicit nvhost_ctrl_gpu(Core::System& system_, EventInterface& events_interface_);
     ~nvhost_ctrl_gpu() override;
 
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                     std::vector<u8>& output) override;
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output) override;
+                    std::span<const u8> inline_input, std::vector<u8>& output) override;
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output) override;
+                    std::vector<u8>& inline_output) override;
 
     void OnOpen(DeviceFD fd) override;
     void OnClose(DeviceFD fd) override;
@@ -151,21 +151,21 @@ private:
     };
     static_assert(sizeof(IoctlGetGpuTime) == 0x10, "IoctlGetGpuTime is incorrect size");
 
-    NvResult GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult GetCharacteristics(std::span<const u8> input, std::vector<u8>& output);
-    NvResult GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output,
+    NvResult GetCharacteristics(std::span<const u8> input, std::vector<u8>& output,
                                 std::vector<u8>& inline_output);
 
-    NvResult GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult GetTPCMasks(std::span<const u8> input, std::vector<u8>& output);
-    NvResult GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output,
+    NvResult GetTPCMasks(std::span<const u8> input, std::vector<u8>& output,
                          std::vector<u8>& inline_output);
 
-    NvResult GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult GetActiveSlotMask(std::span<const u8> input, std::vector<u8>& output);
-    NvResult ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult ZCullGetCtxSize(std::span<const u8> input, std::vector<u8>& output);
-    NvResult ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult ZCullGetInfo(std::span<const u8> input, std::vector<u8>& output);
-    NvResult ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult ZBCSetTable(std::span<const u8> input, std::vector<u8>& output);
-    NvResult ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult ZBCQueryTable(std::span<const u8> input, std::vector<u8>& output);
-    NvResult FlushL2(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult FlushL2(std::span<const u8> input, std::vector<u8>& output);
-    NvResult GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult GetGpuTime(std::span<const u8> input, std::vector<u8>& output);
 
     EventInterface& events_interface;
 

src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp

@@ -46,7 +46,7 @@ nvhost_gpu::~nvhost_gpu() {
     syncpoint_manager.FreeSyncpoint(channel_syncpoint);
 }
 
-NvResult nvhost_gpu::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                             std::vector<u8>& output) {
     switch (command.group) {
     case 0x0:
@@ -98,8 +98,8 @@ NvResult nvhost_gpu::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& i
     return NvResult::NotImplemented;
 };
 
-NvResult nvhost_gpu::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_gpu::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                            const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                            std::span<const u8> inline_input, std::vector<u8>& output) {
     switch (command.group) {
     case 'H':
         switch (command.cmd) {
@@ -112,7 +112,7 @@ NvResult nvhost_gpu::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& i
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_gpu::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                             std::vector<u8>& output, std::vector<u8>& inline_output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
@@ -121,7 +121,7 @@ NvResult nvhost_gpu::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& i
 void nvhost_gpu::OnOpen(DeviceFD fd) {}
 void nvhost_gpu::OnClose(DeviceFD fd) {}
 
-NvResult nvhost_gpu::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::SetNVMAPfd(std::span<const u8> input, std::vector<u8>& output) {
     IoctlSetNvmapFD params{};
     std::memcpy(&params, input.data(), input.size());
     LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd);
@@ -130,7 +130,7 @@ NvResult nvhost_gpu::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& o
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::SetClientData(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::SetClientData(std::span<const u8> input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called");
 
     IoctlClientData params{};
@@ -139,7 +139,7 @@ NvResult nvhost_gpu::SetClientData(const std::vector<u8>& input, std::vector<u8>
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::GetClientData(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::GetClientData(std::span<const u8> input, std::vector<u8>& output) {
     LOG_DEBUG(Service_NVDRV, "called");
 
     IoctlClientData params{};
@@ -149,7 +149,7 @@ NvResult nvhost_gpu::GetClientData(const std::vector<u8>& input, std::vector<u8>
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::ZCullBind(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::ZCullBind(std::span<const u8> input, std::vector<u8>& output) {
     std::memcpy(&zcull_params, input.data(), input.size());
     LOG_DEBUG(Service_NVDRV, "called, gpu_va={:X}, mode={:X}", zcull_params.gpu_va,
               zcull_params.mode);
@@ -158,7 +158,7 @@ NvResult nvhost_gpu::ZCullBind(const std::vector<u8>& input, std::vector<u8>& ou
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::SetErrorNotifier(std::span<const u8> input, std::vector<u8>& output) {
     IoctlSetErrorNotifier params{};
     std::memcpy(&params, input.data(), input.size());
     LOG_WARNING(Service_NVDRV, "(STUBBED) called, offset={:X}, size={:X}, mem={:X}", params.offset,
@@ -168,14 +168,14 @@ NvResult nvhost_gpu::SetErrorNotifier(const std::vector<u8>& input, std::vector<
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::SetChannelPriority(std::span<const u8> input, std::vector<u8>& output) {
     std::memcpy(&channel_priority, input.data(), input.size());
     LOG_DEBUG(Service_NVDRV, "(STUBBED) called, priority={:X}", channel_priority);
 
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::AllocGPFIFOEx2(std::span<const u8> input, std::vector<u8>& output) {
     IoctlAllocGpfifoEx2 params{};
     std::memcpy(&params, input.data(), input.size());
     LOG_WARNING(Service_NVDRV,
@@ -197,7 +197,7 @@ NvResult nvhost_gpu::AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::AllocateObjectContext(std::span<const u8> input, std::vector<u8>& output) {
     IoctlAllocObjCtx params{};
     std::memcpy(&params, input.data(), input.size());
     LOG_WARNING(Service_NVDRV, "(STUBBED) called, class_num={:X}, flags={:X}", params.class_num,
@@ -293,7 +293,7 @@ NvResult nvhost_gpu::SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector<u8>
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::SubmitGPFIFOBase(const std::vector<u8>& input, std::vector<u8>& output,
+NvResult nvhost_gpu::SubmitGPFIFOBase(std::span<const u8> input, std::vector<u8>& output,
                                       bool kickoff) {
     if (input.size() < sizeof(IoctlSubmitGpfifo)) {
         UNIMPLEMENTED();
@@ -314,8 +314,7 @@ NvResult nvhost_gpu::SubmitGPFIFOBase(const std::vector<u8>& input, std::vector<
     return SubmitGPFIFOImpl(params, output, std::move(entries));
 }
 
-NvResult nvhost_gpu::SubmitGPFIFOBase(const std::vector<u8>& input,
+NvResult nvhost_gpu::SubmitGPFIFOBase(std::span<const u8> input, std::span<const u8> input_inline,
-                                      const std::vector<u8>& input_inline,
                                       std::vector<u8>& output) {
     if (input.size() < sizeof(IoctlSubmitGpfifo)) {
         UNIMPLEMENTED();
@@ -328,7 +327,7 @@ NvResult nvhost_gpu::SubmitGPFIFOBase(const std::vector<u8>& input,
     return SubmitGPFIFOImpl(params, output, std::move(entries));
 }
 
-NvResult nvhost_gpu::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::GetWaitbase(std::span<const u8> input, std::vector<u8>& output) {
     IoctlGetWaitbase params{};
     std::memcpy(&params, input.data(), sizeof(IoctlGetWaitbase));
     LOG_INFO(Service_NVDRV, "called, unknown=0x{:X}", params.unknown);
@@ -338,7 +337,7 @@ NvResult nvhost_gpu::GetWaitbase(const std::vector<u8>& input, std::vector<u8>&
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::ChannelSetTimeout(std::span<const u8> input, std::vector<u8>& output) {
     IoctlChannelSetTimeout params{};
     std::memcpy(&params, input.data(), sizeof(IoctlChannelSetTimeout));
     LOG_INFO(Service_NVDRV, "called, timeout=0x{:X}", params.timeout);
@@ -346,7 +345,7 @@ NvResult nvhost_gpu::ChannelSetTimeout(const std::vector<u8>& input, std::vector
     return NvResult::Success;
 }
 
-NvResult nvhost_gpu::ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_gpu::ChannelSetTimeslice(std::span<const u8> input, std::vector<u8>& output) {
     IoctlSetTimeslice params{};
     std::memcpy(&params, input.data(), sizeof(IoctlSetTimeslice));
     LOG_INFO(Service_NVDRV, "called, timeslice=0x{:X}", params.timeslice);

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

@@ -40,12 +40,12 @@ public:
                         NvCore::Container& core);
     ~nvhost_gpu() override;
 
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                     std::vector<u8>& output) override;
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output) override;
+                    std::span<const u8> inline_input, std::vector<u8>& output) override;
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output) override;
+                    std::vector<u8>& inline_output) override;
 
     void OnOpen(DeviceFD fd) override;
     void OnClose(DeviceFD fd) override;
@@ -186,23 +186,23 @@ private:
     u32_le channel_priority{};
     u32_le channel_timeslice{};
 
-    NvResult SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult SetNVMAPfd(std::span<const u8> input, std::vector<u8>& output);
-    NvResult SetClientData(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult SetClientData(std::span<const u8> input, std::vector<u8>& output);
-    NvResult GetClientData(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult GetClientData(std::span<const u8> input, std::vector<u8>& output);
-    NvResult ZCullBind(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult ZCullBind(std::span<const u8> input, std::vector<u8>& output);
-    NvResult SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult SetErrorNotifier(std::span<const u8> input, std::vector<u8>& output);
-    NvResult SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult SetChannelPriority(std::span<const u8> input, std::vector<u8>& output);
-    NvResult AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult AllocGPFIFOEx2(std::span<const u8> input, std::vector<u8>& output);
-    NvResult AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult AllocateObjectContext(std::span<const u8> input, std::vector<u8>& output);
     NvResult SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector<u8>& output,
                               Tegra::CommandList&& entries);
-    NvResult SubmitGPFIFOBase(const std::vector<u8>& input, std::vector<u8>& output,
+    NvResult SubmitGPFIFOBase(std::span<const u8> input, std::vector<u8>& output,
                               bool kickoff = false);
-    NvResult SubmitGPFIFOBase(const std::vector<u8>& input, const std::vector<u8>& input_inline,
+    NvResult SubmitGPFIFOBase(std::span<const u8> input, std::span<const u8> input_inline,
                               std::vector<u8>& output);
-    NvResult GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult GetWaitbase(std::span<const u8> input, std::vector<u8>& output);
-    NvResult ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult ChannelSetTimeout(std::span<const u8> input, std::vector<u8>& output);
-    NvResult ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult ChannelSetTimeslice(std::span<const u8> input, std::vector<u8>& output);
 
     EventInterface& events_interface;
     NvCore::Container& core;

src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp

@@ -15,7 +15,7 @@ nvhost_nvdec::nvhost_nvdec(Core::System& system_, NvCore::Container& core_)
     : nvhost_nvdec_common{system_, core_, NvCore::ChannelType::NvDec} {}
 nvhost_nvdec::~nvhost_nvdec() = default;
 
-NvResult nvhost_nvdec::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_nvdec::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                               std::vector<u8>& output) {
     switch (command.group) {
     case 0x0:
@@ -55,13 +55,13 @@ NvResult nvhost_nvdec::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>&
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_nvdec::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_nvdec::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                              const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                              std::span<const u8> inline_input, std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_nvdec::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_nvdec::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                               std::vector<u8>& output, std::vector<u8>& inline_output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;

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

@@ -13,12 +13,12 @@ public:
     explicit nvhost_nvdec(Core::System& system_, NvCore::Container& core);
     ~nvhost_nvdec() override;
 
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                     std::vector<u8>& output) override;
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output) override;
+                    std::span<const u8> inline_input, std::vector<u8>& output) override;
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output) override;
+                    std::vector<u8>& inline_output) override;
 
     void OnOpen(DeviceFD fd) override;
     void OnClose(DeviceFD fd) override;

src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp

@@ -23,7 +23,7 @@ namespace {
 // Copies count amount of type T from the input vector into the dst vector.
 // Returns the number of bytes written into dst.
 template <typename T>
-std::size_t SliceVectors(const std::vector<u8>& input, std::vector<T>& dst, std::size_t count,
+std::size_t SliceVectors(std::span<const u8> input, std::vector<T>& dst, std::size_t count,
                          std::size_t offset) {
     if (dst.empty()) {
         return 0;
@@ -63,7 +63,7 @@ nvhost_nvdec_common::~nvhost_nvdec_common() {
     core.Host1xDeviceFile().syncpts_accumulated.push_back(channel_syncpoint);
 }
 
-NvResult nvhost_nvdec_common::SetNVMAPfd(const std::vector<u8>& input) {
+NvResult nvhost_nvdec_common::SetNVMAPfd(std::span<const u8> input) {
     IoctlSetNvmapFD params{};
     std::memcpy(&params, input.data(), sizeof(IoctlSetNvmapFD));
     LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd);
@@ -72,7 +72,7 @@ NvResult nvhost_nvdec_common::SetNVMAPfd(const std::vector<u8>& input) {
     return NvResult::Success;
 }
 
-NvResult nvhost_nvdec_common::Submit(DeviceFD fd, const std::vector<u8>& input,
+NvResult nvhost_nvdec_common::Submit(DeviceFD fd, std::span<const u8> input,
                                      std::vector<u8>& output) {
     IoctlSubmit params{};
     std::memcpy(&params, input.data(), sizeof(IoctlSubmit));
@@ -121,7 +121,7 @@ NvResult nvhost_nvdec_common::Submit(DeviceFD fd, const std::vector<u8>& input,
     return NvResult::Success;
 }
 
-NvResult nvhost_nvdec_common::GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_nvdec_common::GetSyncpoint(std::span<const u8> input, std::vector<u8>& output) {
     IoctlGetSyncpoint params{};
     std::memcpy(&params, input.data(), sizeof(IoctlGetSyncpoint));
     LOG_DEBUG(Service_NVDRV, "called GetSyncpoint, id={}", params.param);
@@ -133,7 +133,7 @@ NvResult nvhost_nvdec_common::GetSyncpoint(const std::vector<u8>& input, std::ve
     return NvResult::Success;
 }
 
-NvResult nvhost_nvdec_common::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_nvdec_common::GetWaitbase(std::span<const u8> input, std::vector<u8>& output) {
     IoctlGetWaitbase params{};
     LOG_CRITICAL(Service_NVDRV, "called WAITBASE");
     std::memcpy(&params, input.data(), sizeof(IoctlGetWaitbase));
@@ -142,7 +142,7 @@ NvResult nvhost_nvdec_common::GetWaitbase(const std::vector<u8>& input, std::vec
     return NvResult::Success;
 }
 
-NvResult nvhost_nvdec_common::MapBuffer(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_nvdec_common::MapBuffer(std::span<const u8> input, std::vector<u8>& output) {
     IoctlMapBuffer params{};
     std::memcpy(&params, input.data(), sizeof(IoctlMapBuffer));
     std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries);
@@ -159,7 +159,7 @@ NvResult nvhost_nvdec_common::MapBuffer(const std::vector<u8>& input, std::vecto
     return NvResult::Success;
 }
 
-NvResult nvhost_nvdec_common::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_nvdec_common::UnmapBuffer(std::span<const u8> input, std::vector<u8>& output) {
     IoctlMapBuffer params{};
     std::memcpy(&params, input.data(), sizeof(IoctlMapBuffer));
     std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries);
@@ -173,8 +173,7 @@ NvResult nvhost_nvdec_common::UnmapBuffer(const std::vector<u8>& input, std::vec
     return NvResult::Success;
 }
 
-NvResult nvhost_nvdec_common::SetSubmitTimeout(const std::vector<u8>& input,
+NvResult nvhost_nvdec_common::SetSubmitTimeout(std::span<const u8> input, std::vector<u8>& output) {
-                                               std::vector<u8>& output) {
     std::memcpy(&submit_timeout, input.data(), input.size());
     LOG_WARNING(Service_NVDRV, "(STUBBED) called");
     return NvResult::Success;

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

@@ -107,13 +107,13 @@ protected:
     static_assert(sizeof(IoctlMapBuffer) == 0x0C, "IoctlMapBuffer is incorrect size");
 
     /// Ioctl command implementations
-    NvResult SetNVMAPfd(const std::vector<u8>& input);
+    NvResult SetNVMAPfd(std::span<const u8> input);
-    NvResult Submit(DeviceFD fd, const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult Submit(DeviceFD fd, std::span<const u8> input, std::vector<u8>& output);
-    NvResult GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult GetSyncpoint(std::span<const u8> input, std::vector<u8>& output);
-    NvResult GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult GetWaitbase(std::span<const u8> input, std::vector<u8>& output);
-    NvResult MapBuffer(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult MapBuffer(std::span<const u8> input, std::vector<u8>& output);
-    NvResult UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult UnmapBuffer(std::span<const u8> input, std::vector<u8>& output);
-    NvResult SetSubmitTimeout(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult SetSubmitTimeout(std::span<const u8> input, std::vector<u8>& output);
 
     Kernel::KEvent* QueryEvent(u32 event_id) override;
 

src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp

@@ -12,7 +12,7 @@ namespace Service::Nvidia::Devices {
 nvhost_nvjpg::nvhost_nvjpg(Core::System& system_) : nvdevice{system_} {}
 nvhost_nvjpg::~nvhost_nvjpg() = default;
 
-NvResult nvhost_nvjpg::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_nvjpg::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                               std::vector<u8>& output) {
     switch (command.group) {
     case 'H':
@@ -31,13 +31,13 @@ NvResult nvhost_nvjpg::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>&
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_nvjpg::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_nvjpg::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                              const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                              std::span<const u8> inline_input, std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_nvjpg::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_nvjpg::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                               std::vector<u8>& output, std::vector<u8>& inline_output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
@@ -46,7 +46,7 @@ NvResult nvhost_nvjpg::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>&
 void nvhost_nvjpg::OnOpen(DeviceFD fd) {}
 void nvhost_nvjpg::OnClose(DeviceFD fd) {}
 
-NvResult nvhost_nvjpg::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvhost_nvjpg::SetNVMAPfd(std::span<const u8> input, std::vector<u8>& output) {
     IoctlSetNvmapFD params{};
     std::memcpy(&params, input.data(), input.size());
     LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd);

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

@@ -15,12 +15,12 @@ public:
     explicit nvhost_nvjpg(Core::System& system_);
     ~nvhost_nvjpg() override;
 
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                     std::vector<u8>& output) override;
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output) override;
+                    std::span<const u8> inline_input, std::vector<u8>& output) override;
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output) override;
+                    std::vector<u8>& inline_output) override;
 
     void OnOpen(DeviceFD fd) override;
     void OnClose(DeviceFD fd) override;
@@ -33,7 +33,7 @@ private:
 
     s32_le nvmap_fd{};
 
-    NvResult SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult SetNVMAPfd(std::span<const u8> input, std::vector<u8>& output);
 };
 
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_vic.cpp

@@ -15,7 +15,7 @@ nvhost_vic::nvhost_vic(Core::System& system_, NvCore::Container& core_)
 
 nvhost_vic::~nvhost_vic() = default;
 
-NvResult nvhost_vic::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_vic::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                             std::vector<u8>& output) {
     switch (command.group) {
     case 0x0:
@@ -55,13 +55,13 @@ NvResult nvhost_vic::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& i
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_vic::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_vic::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                            const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                            std::span<const u8> inline_input, std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
 }
 
-NvResult nvhost_vic::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvhost_vic::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                             std::vector<u8>& output, std::vector<u8>& inline_output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;

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

@@ -12,12 +12,12 @@ public:
     explicit nvhost_vic(Core::System& system_, NvCore::Container& core);
     ~nvhost_vic();
 
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                     std::vector<u8>& output) override;
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output) override;
+                    std::span<const u8> inline_input, std::vector<u8>& output) override;
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output) override;
+                    std::vector<u8>& inline_output) override;
 
     void OnOpen(DeviceFD fd) override;
     void OnClose(DeviceFD fd) override;

src/core/hle/service/nvdrv/devices/nvmap.cpp

@@ -25,7 +25,7 @@ nvmap::nvmap(Core::System& system_, NvCore::Container& container_)
 
 nvmap::~nvmap() = default;
 
-NvResult nvmap::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvmap::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                        std::vector<u8>& output) {
     switch (command.group) {
     case 0x1:
@@ -54,13 +54,13 @@ NvResult nvmap::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
     return NvResult::NotImplemented;
 }
 
-NvResult nvmap::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvmap::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                       const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                       std::span<const u8> inline_input, std::vector<u8>& output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
 }
 
-NvResult nvmap::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult nvmap::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                        std::vector<u8>& output, std::vector<u8>& inline_output) {
     UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw);
     return NvResult::NotImplemented;
@@ -69,7 +69,7 @@ NvResult nvmap::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
 void nvmap::OnOpen(DeviceFD fd) {}
 void nvmap::OnClose(DeviceFD fd) {}
 
-NvResult nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvmap::IocCreate(std::span<const u8> input, std::vector<u8>& output) {
     IocCreateParams params;
     std::memcpy(&params, input.data(), sizeof(params));
     LOG_DEBUG(Service_NVDRV, "called, size=0x{:08X}", params.size);
@@ -89,7 +89,7 @@ NvResult nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output)
     return NvResult::Success;
 }
 
-NvResult nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvmap::IocAlloc(std::span<const u8> input, std::vector<u8>& output) {
     IocAllocParams params;
     std::memcpy(&params, input.data(), sizeof(params));
     LOG_DEBUG(Service_NVDRV, "called, addr={:X}", params.address);
@@ -137,7 +137,7 @@ NvResult nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output)
     return result;
 }
 
-NvResult nvmap::IocGetId(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvmap::IocGetId(std::span<const u8> input, std::vector<u8>& output) {
     IocGetIdParams params;
     std::memcpy(&params, input.data(), sizeof(params));
 
@@ -161,7 +161,7 @@ NvResult nvmap::IocGetId(const std::vector<u8>& input, std::vector<u8>& output)
     return NvResult::Success;
 }
 
-NvResult nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvmap::IocFromId(std::span<const u8> input, std::vector<u8>& output) {
     IocFromIdParams params;
     std::memcpy(&params, input.data(), sizeof(params));
 
@@ -192,7 +192,7 @@ NvResult nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output)
     return NvResult::Success;
 }
 
-NvResult nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvmap::IocParam(std::span<const u8> input, std::vector<u8>& output) {
     enum class ParamTypes { Size = 1, Alignment = 2, Base = 3, Heap = 4, Kind = 5, Compr = 6 };
 
     IocParamParams params;
@@ -241,7 +241,7 @@ NvResult nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output)
     return NvResult::Success;
 }
 
-NvResult nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) {
+NvResult nvmap::IocFree(std::span<const u8> input, std::vector<u8>& output) {
     IocFreeParams params;
     std::memcpy(&params, input.data(), sizeof(params));
 

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

@@ -26,12 +26,12 @@ public:
     nvmap(const nvmap&) = delete;
     nvmap& operator=(const nvmap&) = delete;
 
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                     std::vector<u8>& output) override;
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output) override;
+                    std::span<const u8> inline_input, std::vector<u8>& output) override;
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output) override;
+                    std::vector<u8>& inline_output) override;
 
     void OnOpen(DeviceFD fd) override;
     void OnClose(DeviceFD fd) override;
@@ -106,12 +106,12 @@ private:
     };
     static_assert(sizeof(IocGetIdParams) == 8, "IocGetIdParams has wrong size");
 
-    NvResult IocCreate(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocCreate(std::span<const u8> input, std::vector<u8>& output);
-    NvResult IocAlloc(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocAlloc(std::span<const u8> input, std::vector<u8>& output);
-    NvResult IocGetId(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocGetId(std::span<const u8> input, std::vector<u8>& output);
-    NvResult IocFromId(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocFromId(std::span<const u8> input, std::vector<u8>& output);
-    NvResult IocParam(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocParam(std::span<const u8> input, std::vector<u8>& output);
-    NvResult IocFree(const std::vector<u8>& input, std::vector<u8>& output);
+    NvResult IocFree(std::span<const u8> input, std::vector<u8>& output);
 
     NvCore::Container& container;
     NvCore::NvMap& file;

src/core/hle/service/nvdrv/nvdrv.cpp

@@ -124,7 +124,7 @@ DeviceFD Module::Open(const std::string& device_name) {
     return fd;
 }
 
-NvResult Module::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult Module::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input,
                         std::vector<u8>& output) {
     if (fd < 0) {
         LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd);
@@ -141,8 +141,8 @@ NvResult Module::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input
     return itr->second->Ioctl1(fd, command, input, output);
 }
 
-NvResult Module::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult Module::Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                        const std::vector<u8>& inline_input, std::vector<u8>& output) {
+                        std::span<const u8> inline_input, std::vector<u8>& output) {
     if (fd < 0) {
         LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd);
         return NvResult::InvalidState;
@@ -158,7 +158,7 @@ NvResult Module::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input
     return itr->second->Ioctl2(fd, command, input, inline_input, output);
 }
 
-NvResult Module::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+NvResult Module::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input,
                         std::vector<u8>& output, std::vector<u8>& inline_output) {
     if (fd < 0) {
         LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd);

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

@@ -7,6 +7,7 @@
 #include <functional>
 #include <list>
 #include <memory>
+#include <span>
 #include <string>
 #include <unordered_map>
 #include <vector>
@@ -79,14 +80,13 @@ public:
     DeviceFD Open(const std::string& device_name);
 
     /// Sends an ioctl command to the specified file descriptor.
-    NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output);
-                    std::vector<u8>& output);
 
-    NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span<const u8> input,
-                    const std::vector<u8>& inline_input, std::vector<u8>& output);
+                    std::span<const u8> inline_input, std::vector<u8>& output);
 
-    NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input,
+    NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> input, std::vector<u8>& output,
-                    std::vector<u8>& output, std::vector<u8>& inline_output);
+                    std::vector<u8>& inline_output);
 
     /// Closes a device file descriptor and returns operation success.
     NvResult Close(DeviceFD fd);

src/core/hle/service/nvflinger/buffer_queue_producer.cpp

@@ -815,8 +815,8 @@ Status BufferQueueProducer::SetPreallocatedBuffer(s32 slot,
 
 void BufferQueueProducer::Transact(Kernel::HLERequestContext& ctx, TransactionId code, u32 flags) {
     Status status{Status::NoError};
-    Parcel parcel_in{ctx.ReadBuffer()};
+    InputParcel parcel_in{ctx.ReadBuffer()};
-    Parcel parcel_out{};
+    OutputParcel parcel_out{};
 
     switch (code) {
     case TransactionId::Connect: {

src/core/hle/service/nvflinger/graphic_buffer_producer.cpp

@@ -9,7 +9,7 @@
 
 namespace Service::android {
 
-QueueBufferInput::QueueBufferInput(Parcel& parcel) {
+QueueBufferInput::QueueBufferInput(InputParcel& parcel) {
     parcel.ReadFlattened(*this);
 }
 

src/core/hle/service/nvflinger/graphic_buffer_producer.h

@@ -14,11 +14,11 @@
 
 namespace Service::android {
 
-class Parcel;
+class InputParcel;
 
 #pragma pack(push, 1)
 struct QueueBufferInput final {
-    explicit QueueBufferInput(Parcel& parcel);
+    explicit QueueBufferInput(InputParcel& parcel);
 
     void Deflate(s64* timestamp_, bool* is_auto_timestamp_, Common::Rectangle<s32>* crop_,
                  NativeWindowScalingMode* scaling_mode_, NativeWindowTransform* transform_,

src/core/hle/service/nvflinger/parcel.h

@@ -4,6 +4,7 @@
 #pragma once
 
 #include <memory>
+#include <span>
 #include <vector>
 
 #include "common/alignment.h"
@@ -12,18 +13,17 @@
 
 namespace Service::android {
 
-class Parcel final {
+struct ParcelHeader {
+    u32 data_size;
+    u32 data_offset;
+    u32 objects_size;
+    u32 objects_offset;
+};
+static_assert(sizeof(ParcelHeader) == 16, "ParcelHeader has wrong size");
+
+class InputParcel final {
 public:
-    static constexpr std::size_t DefaultBufferSize = 0x40;
+    explicit InputParcel(std::span<const u8> in_data) : read_buffer(std::move(in_data)) {
-
-    Parcel() : buffer(DefaultBufferSize) {}
-
-    template <typename T>
-    explicit Parcel(const T& out_data) : buffer(DefaultBufferSize) {
-        Write(out_data);
-    }
-
-    explicit Parcel(std::vector<u8> in_data) : buffer(std::move(in_data)) {
         DeserializeHeader();
         [[maybe_unused]] const std::u16string token = ReadInterfaceToken();
     }
@@ -31,9 +31,9 @@ public:
     template <typename T>
     void Read(T& val) {
         static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable.");
-        ASSERT(read_index + sizeof(T) <= buffer.size());
+        ASSERT(read_index + sizeof(T) <= read_buffer.size());
 
-        std::memcpy(&val, buffer.data() + read_index, sizeof(T));
+        std::memcpy(&val, read_buffer.data() + read_index, sizeof(T));
         read_index += sizeof(T);
         read_index = Common::AlignUp(read_index, 4);
     }
@@ -62,10 +62,10 @@ public:
     template <typename T>
     T ReadUnaligned() {
         static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable.");
-        ASSERT(read_index + sizeof(T) <= buffer.size());
+        ASSERT(read_index + sizeof(T) <= read_buffer.size());
 
         T val;
-        std::memcpy(&val, buffer.data() + read_index, sizeof(T));
+        std::memcpy(&val, read_buffer.data() + read_index, sizeof(T));
         read_index += sizeof(T);
         return val;
     }
@@ -101,6 +101,31 @@ public:
         return token;
     }
 
+    void DeserializeHeader() {
+        ASSERT(read_buffer.size() > sizeof(ParcelHeader));
+
+        ParcelHeader header{};
+        std::memcpy(&header, read_buffer.data(), sizeof(ParcelHeader));
+
+        read_index = header.data_offset;
+    }
+
+private:
+    std::span<const u8> read_buffer;
+    std::size_t read_index = 0;
+};
+
+class OutputParcel final {
+public:
+    static constexpr std::size_t DefaultBufferSize = 0x40;
+
+    OutputParcel() : buffer(DefaultBufferSize) {}
+
+    template <typename T>
+    explicit OutputParcel(const T& out_data) : buffer(DefaultBufferSize) {
+        Write(out_data);
+    }
+
     template <typename T>
     void Write(const T& val) {
         static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable.");
@@ -133,40 +158,20 @@ public:
         WriteObject(ptr.get());
     }
 
-    void DeserializeHeader() {
-        ASSERT(buffer.size() > sizeof(Header));
-
-        Header header{};
-        std::memcpy(&header, buffer.data(), sizeof(Header));
-
-        read_index = header.data_offset;
-    }
-
     std::vector<u8> Serialize() const {
-        ASSERT(read_index == 0);
+        ParcelHeader header{};
-
+        header.data_size = static_cast<u32>(write_index - sizeof(ParcelHeader));
-        Header header{};
+        header.data_offset = sizeof(ParcelHeader);
-        header.data_size = static_cast<u32>(write_index - sizeof(Header));
-        header.data_offset = sizeof(Header);
         header.objects_size = 4;
-        header.objects_offset = static_cast<u32>(sizeof(Header) + header.data_size);
+        header.objects_offset = static_cast<u32>(sizeof(ParcelHeader) + header.data_size);
-        std::memcpy(buffer.data(), &header, sizeof(Header));
+        std::memcpy(buffer.data(), &header, sizeof(ParcelHeader));
 
         return buffer;
     }
 
 private:
-    struct Header {
-        u32 data_size;
-        u32 data_offset;
-        u32 objects_size;
-        u32 objects_offset;
-    };
-    static_assert(sizeof(Header) == 16, "ParcelHeader has wrong size");
-
     mutable std::vector<u8> buffer;
-    std::size_t read_index = 0;
+    std::size_t write_index = sizeof(ParcelHeader);
-    std::size_t write_index = sizeof(Header);
 };
 
 } // namespace Service::android

src/core/hle/service/prepo/prepo.cpp

@@ -63,7 +63,7 @@ private:
                 return ctx.ReadBuffer(1);
             }
 
-            return std::vector<u8>{};
+            return std::span<const u8>{};
         }();
 
         LOG_DEBUG(Service_PREPO,
@@ -90,7 +90,7 @@ private:
                 return ctx.ReadBuffer(1);
             }
 
-            return std::vector<u8>{};
+            return std::span<const u8>{};
         }();
 
         LOG_DEBUG(Service_PREPO,
@@ -142,7 +142,7 @@ private:
                 return ctx.ReadBuffer(1);
             }
 
-            return std::vector<u8>{};
+            return std::span<const u8>{};
         }();
 
         LOG_DEBUG(Service_PREPO, "called, title_id={:016X}, data1_size={:016X}, data2_size={:016X}",
@@ -166,7 +166,7 @@ private:
                 return ctx.ReadBuffer(1);
             }
 
-            return std::vector<u8>{};
+            return std::span<const u8>{};
         }();
 
         LOG_DEBUG(Service_PREPO,

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

@@ -208,7 +208,6 @@ void BSD::Bind(Kernel::HLERequestContext& ctx) {
     const s32 fd = rp.Pop<s32>();
 
     LOG_DEBUG(Service, "called. fd={} addrlen={}", fd, ctx.GetReadBufferSize());
-
     BuildErrnoResponse(ctx, BindImpl(fd, ctx.ReadBuffer()));
 }
 
@@ -312,7 +311,7 @@ void BSD::SetSockOpt(Kernel::HLERequestContext& ctx) {
     const u32 level = rp.Pop<u32>();
     const OptName optname = static_cast<OptName>(rp.Pop<u32>());
 
-    const std::vector<u8> buffer = ctx.ReadBuffer();
+    const auto buffer = ctx.ReadBuffer();
     const u8* optval = buffer.empty() ? nullptr : buffer.data();
     size_t optlen = buffer.size();
 
@@ -489,7 +488,7 @@ std::pair<s32, Errno> BSD::SocketImpl(Domain domain, Type type, Protocol protoco
     return {fd, Errno::SUCCESS};
 }
 
-std::pair<s32, Errno> BSD::PollImpl(std::vector<u8>& write_buffer, std::vector<u8> read_buffer,
+std::pair<s32, Errno> BSD::PollImpl(std::vector<u8>& write_buffer, std::span<const u8> read_buffer,
                                     s32 nfds, s32 timeout) {
     if (write_buffer.size() < nfds * sizeof(PollFD)) {
         return {-1, Errno::INVAL};
@@ -584,7 +583,7 @@ std::pair<s32, Errno> BSD::AcceptImpl(s32 fd, std::vector<u8>& write_buffer) {
     return {new_fd, Errno::SUCCESS};
 }
 
-Errno BSD::BindImpl(s32 fd, const std::vector<u8>& addr) {
+Errno BSD::BindImpl(s32 fd, std::span<const u8> addr) {
     if (!IsFileDescriptorValid(fd)) {
         return Errno::BADF;
     }
@@ -595,7 +594,7 @@ Errno BSD::BindImpl(s32 fd, const std::vector<u8>& addr) {
     return Translate(file_descriptors[fd]->socket->Bind(Translate(addr_in)));
 }
 
-Errno BSD::ConnectImpl(s32 fd, const std::vector<u8>& addr) {
+Errno BSD::ConnectImpl(s32 fd, std::span<const u8> addr) {
     if (!IsFileDescriptorValid(fd)) {
         return Errno::BADF;
     }
@@ -800,15 +799,15 @@ std::pair<s32, Errno> BSD::RecvFromImpl(s32 fd, u32 flags, std::vector<u8>& mess
     return {ret, bsd_errno};
 }
 
-std::pair<s32, Errno> BSD::SendImpl(s32 fd, u32 flags, const std::vector<u8>& message) {
+std::pair<s32, Errno> BSD::SendImpl(s32 fd, u32 flags, std::span<const u8> message) {
     if (!IsFileDescriptorValid(fd)) {
         return {-1, Errno::BADF};
     }
     return Translate(file_descriptors[fd]->socket->Send(message, flags));
 }
 
-std::pair<s32, Errno> BSD::SendToImpl(s32 fd, u32 flags, const std::vector<u8>& message,
+std::pair<s32, Errno> BSD::SendToImpl(s32 fd, u32 flags, std::span<const u8> message,
-                                      const std::vector<u8>& addr) {
+                                      std::span<const u8> addr) {
     if (!IsFileDescriptorValid(fd)) {
         return {-1, Errno::BADF};
     }

src/core/hle/service/sockets/bsd.h

@@ -4,6 +4,7 @@
 #pragma once
 
 #include <memory>
+#include <span>
 #include <vector>
 
 #include "common/common_types.h"
@@ -44,7 +45,7 @@ private:
 
         s32 nfds;
         s32 timeout;
-        std::vector<u8> read_buffer;
+        std::span<const u8> read_buffer;
         std::vector<u8> write_buffer;
         s32 ret{};
         Errno bsd_errno{};
@@ -65,7 +66,7 @@ private:
         void Response(Kernel::HLERequestContext& ctx);
 
         s32 fd;
-        std::vector<u8> addr;
+        std::span<const u8> addr;
         Errno bsd_errno{};
     };
 
@@ -98,7 +99,7 @@ private:
 
         s32 fd;
         u32 flags;
-        std::vector<u8> message;
+        std::span<const u8> message;
         s32 ret{};
         Errno bsd_errno{};
     };
@@ -109,8 +110,8 @@ private:
 
         s32 fd;
         u32 flags;
-        std::vector<u8> message;
+        std::span<const u8> message;
-        std::vector<u8> addr;
+        std::span<const u8> addr;
         s32 ret{};
         Errno bsd_errno{};
     };
@@ -143,11 +144,11 @@ private:
     void ExecuteWork(Kernel::HLERequestContext& ctx, Work work);
 
     std::pair<s32, Errno> SocketImpl(Domain domain, Type type, Protocol protocol);
-    std::pair<s32, Errno> PollImpl(std::vector<u8>& write_buffer, std::vector<u8> read_buffer,
+    std::pair<s32, Errno> PollImpl(std::vector<u8>& write_buffer, std::span<const u8> read_buffer,
                                    s32 nfds, s32 timeout);
     std::pair<s32, Errno> AcceptImpl(s32 fd, std::vector<u8>& write_buffer);
-    Errno BindImpl(s32 fd, const std::vector<u8>& addr);
+    Errno BindImpl(s32 fd, std::span<const u8> addr);
-    Errno ConnectImpl(s32 fd, const std::vector<u8>& addr);
+    Errno ConnectImpl(s32 fd, std::span<const u8> addr);
     Errno GetPeerNameImpl(s32 fd, std::vector<u8>& write_buffer);
     Errno GetSockNameImpl(s32 fd, std::vector<u8>& write_buffer);
     Errno ListenImpl(s32 fd, s32 backlog);
@@ -157,9 +158,9 @@ private:
     std::pair<s32, Errno> RecvImpl(s32 fd, u32 flags, std::vector<u8>& message);
     std::pair<s32, Errno> RecvFromImpl(s32 fd, u32 flags, std::vector<u8>& message,
                                        std::vector<u8>& addr);
-    std::pair<s32, Errno> SendImpl(s32 fd, u32 flags, const std::vector<u8>& message);
+    std::pair<s32, Errno> SendImpl(s32 fd, u32 flags, std::span<const u8> message);
-    std::pair<s32, Errno> SendToImpl(s32 fd, u32 flags, const std::vector<u8>& message,
+    std::pair<s32, Errno> SendToImpl(s32 fd, u32 flags, std::span<const u8> message,
-                                     const std::vector<u8>& addr);
+                                     std::span<const u8> addr);
     Errno CloseImpl(s32 fd);
 
     s32 FindFreeFileDescriptorHandle() noexcept;

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

@@ -101,7 +101,7 @@ private:
     void ImportServerPki(Kernel::HLERequestContext& ctx) {
         IPC::RequestParser rp{ctx};
         const auto certificate_format = rp.PopEnum<CertificateFormat>();
-        const auto pkcs_12_certificates = ctx.ReadBuffer(0);
+        [[maybe_unused]] const auto pkcs_12_certificates = ctx.ReadBuffer(0);
 
         constexpr u64 server_id = 0;
 
@@ -113,13 +113,13 @@ private:
     }
 
     void ImportClientPki(Kernel::HLERequestContext& ctx) {
-        const auto pkcs_12_certificate = ctx.ReadBuffer(0);
+        [[maybe_unused]] const auto pkcs_12_certificate = ctx.ReadBuffer(0);
-        const auto ascii_password = [&ctx] {
+        [[maybe_unused]] const auto ascii_password = [&ctx] {
             if (ctx.CanReadBuffer(1)) {
                 return ctx.ReadBuffer(1);
             }
 
-            return std::vector<u8>{};
+            return std::span<const u8>{};
         }();
 
         constexpr u64 client_id = 0;

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

@@ -603,7 +603,7 @@ private:
             return;
         }
 
-        const auto parcel = android::Parcel{NativeWindow{*buffer_queue_id}};
+        const auto parcel = android::OutputParcel{NativeWindow{*buffer_queue_id}};
         const auto buffer_size = ctx.WriteBuffer(parcel.Serialize());
 
         IPC::ResponseBuilder rb{ctx, 4};
@@ -649,7 +649,7 @@ private:
             return;
         }
 
-        const auto parcel = android::Parcel{NativeWindow{*buffer_queue_id}};
+        const auto parcel = android::OutputParcel{NativeWindow{*buffer_queue_id}};
         const auto buffer_size = ctx.WriteBuffer(parcel.Serialize());
 
         IPC::ResponseBuilder rb{ctx, 6};

src/core/internal_network/network.cpp

@@ -550,7 +550,7 @@ std::pair<s32, Errno> Socket::RecvFrom(int flags, std::vector<u8>& message, Sock
     return {-1, GetAndLogLastError()};
 }
 
-std::pair<s32, Errno> Socket::Send(const std::vector<u8>& message, int flags) {
+std::pair<s32, Errno> Socket::Send(std::span<const u8> message, int flags) {
     ASSERT(message.size() < static_cast<size_t>(std::numeric_limits<int>::max()));
     ASSERT(flags == 0);
 
@@ -563,7 +563,7 @@ std::pair<s32, Errno> Socket::Send(const std::vector<u8>& message, int flags) {
     return {-1, GetAndLogLastError()};
 }
 
-std::pair<s32, Errno> Socket::SendTo(u32 flags, const std::vector<u8>& message,
+std::pair<s32, Errno> Socket::SendTo(u32 flags, std::span<const u8> message,
                                      const SockAddrIn* addr) {
     ASSERT(flags == 0);
 

src/core/internal_network/socket_proxy.cpp

@@ -182,7 +182,7 @@ std::pair<s32, Errno> ProxySocket::ReceivePacket(int flags, std::vector<u8>& mes
     return {static_cast<u32>(read_bytes), Errno::SUCCESS};
 }
 
-std::pair<s32, Errno> ProxySocket::Send(const std::vector<u8>& message, int flags) {
+std::pair<s32, Errno> ProxySocket::Send(std::span<const u8> message, int flags) {
     LOG_WARNING(Network, "(STUBBED) called");
     ASSERT(message.size() < static_cast<size_t>(std::numeric_limits<int>::max()));
     ASSERT(flags == 0);
@@ -200,7 +200,7 @@ void ProxySocket::SendPacket(ProxyPacket& packet) {
     }
 }
 
-std::pair<s32, Errno> ProxySocket::SendTo(u32 flags, const std::vector<u8>& message,
+std::pair<s32, Errno> ProxySocket::SendTo(u32 flags, std::span<const u8> message,
                                           const SockAddrIn* addr) {
     ASSERT(flags == 0);
 

src/core/internal_network/socket_proxy.h

@@ -4,6 +4,7 @@
 #pragma once
 
 #include <mutex>
+#include <span>
 #include <vector>
 #include <queue>
 
@@ -48,11 +49,11 @@ public:
     std::pair<s32, Errno> ReceivePacket(int flags, std::vector<u8>& message, SockAddrIn* addr,
                                         std::size_t max_length);
 
-    std::pair<s32, Errno> Send(const std::vector<u8>& message, int flags) override;
+    std::pair<s32, Errno> Send(std::span<const u8> message, int flags) override;
 
     void SendPacket(ProxyPacket& packet);
 
-    std::pair<s32, Errno> SendTo(u32 flags, const std::vector<u8>& message,
+    std::pair<s32, Errno> SendTo(u32 flags, std::span<const u8> message,
                                  const SockAddrIn* addr) override;
 
     Errno SetLinger(bool enable, u32 linger) override;

src/core/internal_network/sockets.h

@@ -5,6 +5,7 @@
 
 #include <map>
 #include <memory>
+#include <span>
 #include <utility>
 
 #if defined(_WIN32)
@@ -66,9 +67,9 @@ public:
     virtual std::pair<s32, Errno> RecvFrom(int flags, std::vector<u8>& message,
                                            SockAddrIn* addr) = 0;
 
-    virtual std::pair<s32, Errno> Send(const std::vector<u8>& message, int flags) = 0;
+    virtual std::pair<s32, Errno> Send(std::span<const u8> message, int flags) = 0;
 
-    virtual std::pair<s32, Errno> SendTo(u32 flags, const std::vector<u8>& message,
+    virtual std::pair<s32, Errno> SendTo(u32 flags, std::span<const u8> message,
                                          const SockAddrIn* addr) = 0;
 
     virtual Errno SetLinger(bool enable, u32 linger) = 0;
@@ -138,9 +139,9 @@ public:
 
     std::pair<s32, Errno> RecvFrom(int flags, std::vector<u8>& message, SockAddrIn* addr) override;
 
-    std::pair<s32, Errno> Send(const std::vector<u8>& message, int flags) override;
+    std::pair<s32, Errno> Send(std::span<const u8> message, int flags) override;
 
-    std::pair<s32, Errno> SendTo(u32 flags, const std::vector<u8>& message,
+    std::pair<s32, Errno> SendTo(u32 flags, std::span<const u8> message,
                                  const SockAddrIn* addr) override;
 
     Errno SetLinger(bool enable, u32 linger) override;

src/core/reporter.cpp

@@ -312,7 +312,7 @@ void Reporter::SaveUnimplementedAppletReport(
 }
 
 void Reporter::SavePlayReport(PlayReportType type, u64 title_id,
-                              const std::vector<std::vector<u8>>& data,
+                              const std::vector<std::span<const u8>>& data,
                               std::optional<u64> process_id, std::optional<u128> user_id) const {
     if (!IsReportingEnabled()) {
         return;

src/core/reporter.h

@@ -5,6 +5,7 @@
 
 #include <array>
 #include <optional>
+#include <span>
 #include <string>
 #include <vector>
 #include "common/common_types.h"
@@ -56,7 +57,8 @@ public:
         System,
     };
 
-    void SavePlayReport(PlayReportType type, u64 title_id, const std::vector<std::vector<u8>>& data,
+    void SavePlayReport(PlayReportType type, u64 title_id,
+                        const std::vector<std::span<const u8>>& data,
                         std::optional<u64> process_id = {}, std::optional<u128> user_id = {}) const;
 
     // Used by error applet

src/input_common/helpers/joycon_protocol/calibration.cpp

@@ -13,34 +13,34 @@ CalibrationProtocol::CalibrationProtocol(std::shared_ptr<JoyconHandle> handle)
 
 DriverResult CalibrationProtocol::GetLeftJoyStickCalibration(JoyStickCalibration& calibration) {
     ScopedSetBlocking sb(this);
-    std::vector<u8> buffer;
     DriverResult result{DriverResult::Success};
+    JoystickLeftSpiCalibration spi_calibration{};
+    bool has_user_calibration = false;
     calibration = {};
 
-    result = ReadSPI(CalAddr::USER_LEFT_MAGIC, sizeof(u16), buffer);
-
     if (result == DriverResult::Success) {
-        const bool has_user_calibration = buffer[0] == 0xB2 && buffer[1] == 0xA1;
+        result = HasUserCalibration(SpiAddress::USER_LEFT_MAGIC, has_user_calibration);
-        if (has_user_calibration) {
-            result = ReadSPI(CalAddr::USER_LEFT_DATA, 9, buffer);
-        } else {
-            result = ReadSPI(CalAddr::FACT_LEFT_DATA, 9, buffer);
     }
+
+    // Read User defined calibration
+    if (result == DriverResult::Success && has_user_calibration) {
+        result = ReadSPI(SpiAddress::USER_LEFT_DATA, spi_calibration);
+    }
+
+    // Read Factory calibration
+    if (result == DriverResult::Success && !has_user_calibration) {
+        result = ReadSPI(SpiAddress::FACT_LEFT_DATA, spi_calibration);
     }
 
     if (result == DriverResult::Success) {
-        calibration.x.max = static_cast<u16>(((buffer[1] & 0x0F) << 8) | buffer[0]);
+        calibration.x.center = GetXAxisCalibrationValue(spi_calibration.center);
-        calibration.y.max = static_cast<u16>((buffer[2] << 4) | (buffer[1] >> 4));
+        calibration.y.center = GetYAxisCalibrationValue(spi_calibration.center);
-        calibration.x.center = static_cast<u16>(((buffer[4] & 0x0F) << 8) | buffer[3]);
+        calibration.x.min = GetXAxisCalibrationValue(spi_calibration.min);
-        calibration.y.center = static_cast<u16>((buffer[5] << 4) | (buffer[4] >> 4));
+        calibration.y.min = GetYAxisCalibrationValue(spi_calibration.min);
-        calibration.x.min = static_cast<u16>(((buffer[7] & 0x0F) << 8) | buffer[6]);
+        calibration.x.max = GetXAxisCalibrationValue(spi_calibration.max);
-        calibration.y.min = static_cast<u16>((buffer[8] << 4) | (buffer[7] >> 4));
+        calibration.y.max = GetYAxisCalibrationValue(spi_calibration.max);
     }
 
-    // Nintendo fix for drifting stick
-    // result = ReadSPI(0x60, 0x86 ,buffer, 16);
-    // calibration.deadzone = (u16)((buffer[4] << 8) & 0xF00 | buffer[3]);
-
     // Set a valid default calibration if data is missing
     ValidateCalibration(calibration);
 
@@ -49,34 +49,34 @@ DriverResult CalibrationProtocol::GetLeftJoyStickCalibration(JoyStickCalibration
 
 DriverResult CalibrationProtocol::GetRightJoyStickCalibration(JoyStickCalibration& calibration) {
     ScopedSetBlocking sb(this);
-    std::vector<u8> buffer;
     DriverResult result{DriverResult::Success};
+    JoystickRightSpiCalibration spi_calibration{};
+    bool has_user_calibration = false;
     calibration = {};
 
-    result = ReadSPI(CalAddr::USER_RIGHT_MAGIC, sizeof(u16), buffer);
-
     if (result == DriverResult::Success) {
-        const bool has_user_calibration = buffer[0] == 0xB2 && buffer[1] == 0xA1;
+        result = HasUserCalibration(SpiAddress::USER_RIGHT_MAGIC, has_user_calibration);
-        if (has_user_calibration) {
-            result = ReadSPI(CalAddr::USER_RIGHT_DATA, 9, buffer);
-        } else {
-            result = ReadSPI(CalAddr::FACT_RIGHT_DATA, 9, buffer);
     }
+
+    // Read User defined calibration
+    if (result == DriverResult::Success && has_user_calibration) {
+        result = ReadSPI(SpiAddress::USER_RIGHT_DATA, spi_calibration);
+    }
+
+    // Read Factory calibration
+    if (result == DriverResult::Success && !has_user_calibration) {
+        result = ReadSPI(SpiAddress::FACT_RIGHT_DATA, spi_calibration);
     }
 
     if (result == DriverResult::Success) {
-        calibration.x.center = static_cast<u16>(((buffer[1] & 0x0F) << 8) | buffer[0]);
+        calibration.x.center = GetXAxisCalibrationValue(spi_calibration.center);
-        calibration.y.center = static_cast<u16>((buffer[2] << 4) | (buffer[1] >> 4));
+        calibration.y.center = GetYAxisCalibrationValue(spi_calibration.center);
-        calibration.x.min = static_cast<u16>(((buffer[4] & 0x0F) << 8) | buffer[3]);
+        calibration.x.min = GetXAxisCalibrationValue(spi_calibration.min);
-        calibration.y.min = static_cast<u16>((buffer[5] << 4) | (buffer[4] >> 4));
+        calibration.y.min = GetYAxisCalibrationValue(spi_calibration.min);
-        calibration.x.max = static_cast<u16>(((buffer[7] & 0x0F) << 8) | buffer[6]);
+        calibration.x.max = GetXAxisCalibrationValue(spi_calibration.max);
-        calibration.y.max = static_cast<u16>((buffer[8] << 4) | (buffer[7] >> 4));
+        calibration.y.max = GetYAxisCalibrationValue(spi_calibration.max);
     }
 
-    // Nintendo fix for drifting stick
-    // buffer = ReadSPI(0x60, 0x98 , 16);
-    // joystick.deadzone = (u16)((buffer[4] << 8) & 0xF00 | buffer[3]);
-
     // Set a valid default calibration if data is missing
     ValidateCalibration(calibration);
 
@@ -85,39 +85,41 @@ DriverResult CalibrationProtocol::GetRightJoyStickCalibration(JoyStickCalibratio
 
 DriverResult CalibrationProtocol::GetImuCalibration(MotionCalibration& calibration) {
     ScopedSetBlocking sb(this);
-    std::vector<u8> buffer;
     DriverResult result{DriverResult::Success};
+    ImuSpiCalibration spi_calibration{};
+    bool has_user_calibration = false;
     calibration = {};
 
-    result = ReadSPI(CalAddr::USER_IMU_MAGIC, sizeof(u16), buffer);
-
     if (result == DriverResult::Success) {
-        const bool has_user_calibration = buffer[0] == 0xB2 && buffer[1] == 0xA1;
+        result = HasUserCalibration(SpiAddress::USER_IMU_MAGIC, has_user_calibration);
-        if (has_user_calibration) {
-            result = ReadSPI(CalAddr::USER_IMU_DATA, sizeof(IMUCalibration), buffer);
-        } else {
-            result = ReadSPI(CalAddr::FACT_IMU_DATA, sizeof(IMUCalibration), buffer);
     }
+
+    // Read User defined calibration
+    if (result == DriverResult::Success && has_user_calibration) {
+        result = ReadSPI(SpiAddress::USER_IMU_DATA, spi_calibration);
+    }
+
+    // Read Factory calibration
+    if (result == DriverResult::Success && !has_user_calibration) {
+        result = ReadSPI(SpiAddress::FACT_IMU_DATA, spi_calibration);
     }
 
     if (result == DriverResult::Success) {
-        IMUCalibration device_calibration{};
+        calibration.accelerometer[0].offset = spi_calibration.accelerometer_offset[0];
-        memcpy(&device_calibration, buffer.data(), sizeof(IMUCalibration));
+        calibration.accelerometer[1].offset = spi_calibration.accelerometer_offset[1];
-        calibration.accelerometer[0].offset = device_calibration.accelerometer_offset[0];
+        calibration.accelerometer[2].offset = spi_calibration.accelerometer_offset[2];
-        calibration.accelerometer[1].offset = device_calibration.accelerometer_offset[1];
-        calibration.accelerometer[2].offset = device_calibration.accelerometer_offset[2];
 
-        calibration.accelerometer[0].scale = device_calibration.accelerometer_scale[0];
+        calibration.accelerometer[0].scale = spi_calibration.accelerometer_scale[0];
-        calibration.accelerometer[1].scale = device_calibration.accelerometer_scale[1];
+        calibration.accelerometer[1].scale = spi_calibration.accelerometer_scale[1];
-        calibration.accelerometer[2].scale = device_calibration.accelerometer_scale[2];
+        calibration.accelerometer[2].scale = spi_calibration.accelerometer_scale[2];
 
-        calibration.gyro[0].offset = device_calibration.gyroscope_offset[0];
+        calibration.gyro[0].offset = spi_calibration.gyroscope_offset[0];
-        calibration.gyro[1].offset = device_calibration.gyroscope_offset[1];
+        calibration.gyro[1].offset = spi_calibration.gyroscope_offset[1];
-        calibration.gyro[2].offset = device_calibration.gyroscope_offset[2];
+        calibration.gyro[2].offset = spi_calibration.gyroscope_offset[2];
 
-        calibration.gyro[0].scale = device_calibration.gyroscope_scale[0];
+        calibration.gyro[0].scale = spi_calibration.gyroscope_scale[0];
-        calibration.gyro[1].scale = device_calibration.gyroscope_scale[1];
+        calibration.gyro[1].scale = spi_calibration.gyroscope_scale[1];
-        calibration.gyro[2].scale = device_calibration.gyroscope_scale[2];
+        calibration.gyro[2].scale = spi_calibration.gyroscope_scale[2];
     }
 
     ValidateCalibration(calibration);
@@ -127,10 +129,12 @@ DriverResult CalibrationProtocol::GetImuCalibration(MotionCalibration& calibrati
 
 DriverResult CalibrationProtocol::GetRingCalibration(RingCalibration& calibration,
                                                      s16 current_value) {
+    constexpr s16 DefaultRingRange{800};
+
     // TODO: Get default calibration form ring itself
     if (ring_data_max == 0 && ring_data_min == 0) {
-        ring_data_max = current_value + 800;
+        ring_data_max = current_value + DefaultRingRange;
-        ring_data_min = current_value - 800;
+        ring_data_min = current_value - DefaultRingRange;
         ring_data_default = current_value;
     }
     ring_data_max = std::max(ring_data_max, current_value);
@@ -143,42 +147,72 @@ DriverResult CalibrationProtocol::GetRingCalibration(RingCalibration& calibratio
     return DriverResult::Success;
 }
 
+DriverResult CalibrationProtocol::HasUserCalibration(SpiAddress address,
+                                                     bool& has_user_calibration) {
+    MagicSpiCalibration spi_magic{};
+    const DriverResult result{ReadSPI(address, spi_magic)};
+    has_user_calibration = false;
+    if (result == DriverResult::Success) {
+        has_user_calibration = spi_magic.first == CalibrationMagic::USR_MAGIC_0 &&
+                               spi_magic.second == CalibrationMagic::USR_MAGIC_1;
+    }
+    return result;
+}
+
+u16 CalibrationProtocol::GetXAxisCalibrationValue(std::span<u8> block) const {
+    return static_cast<u16>(((block[1] & 0x0F) << 8) | block[0]);
+}
+
+u16 CalibrationProtocol::GetYAxisCalibrationValue(std::span<u8> block) const {
+    return static_cast<u16>((block[2] << 4) | (block[1] >> 4));
+}
+
 void CalibrationProtocol::ValidateCalibration(JoyStickCalibration& calibration) {
-    constexpr u16 DefaultStickCenter{2048};
+    constexpr u16 DefaultStickCenter{0x800};
-    constexpr u16 DefaultStickRange{1740};
+    constexpr u16 DefaultStickRange{0x6cc};
 
-    if (calibration.x.center == 0xFFF || calibration.x.center == 0) {
+    calibration.x.center = ValidateValue(calibration.x.center, DefaultStickCenter);
-        calibration.x.center = DefaultStickCenter;
+    calibration.x.max = ValidateValue(calibration.x.max, DefaultStickRange);
-    }
+    calibration.x.min = ValidateValue(calibration.x.min, DefaultStickRange);
-    if (calibration.x.max == 0xFFF || calibration.x.max == 0) {
-        calibration.x.max = DefaultStickRange;
-    }
-    if (calibration.x.min == 0xFFF || calibration.x.min == 0) {
-        calibration.x.min = DefaultStickRange;
-    }
 
-    if (calibration.y.center == 0xFFF || calibration.y.center == 0) {
+    calibration.y.center = ValidateValue(calibration.y.center, DefaultStickCenter);
-        calibration.y.center = DefaultStickCenter;
+    calibration.y.max = ValidateValue(calibration.y.max, DefaultStickRange);
-    }
+    calibration.y.min = ValidateValue(calibration.y.min, DefaultStickRange);
-    if (calibration.y.max == 0xFFF || calibration.y.max == 0) {
-        calibration.y.max = DefaultStickRange;
-    }
-    if (calibration.y.min == 0xFFF || calibration.y.min == 0) {
-        calibration.y.min = DefaultStickRange;
-    }
 }
 
 void CalibrationProtocol::ValidateCalibration(MotionCalibration& calibration) {
+    constexpr s16 DefaultAccelerometerScale{0x4000};
+    constexpr s16 DefaultGyroScale{0x3be7};
+    constexpr s16 DefaultOffset{0};
+
     for (auto& sensor : calibration.accelerometer) {
-        if (sensor.scale == 0) {
+        sensor.scale = ValidateValue(sensor.scale, DefaultAccelerometerScale);
-            sensor.scale = 0x4000;
+        sensor.offset = ValidateValue(sensor.offset, DefaultOffset);
-        }
     }
     for (auto& sensor : calibration.gyro) {
-        if (sensor.scale == 0) {
+        sensor.scale = ValidateValue(sensor.scale, DefaultGyroScale);
-            sensor.scale = 0x3be7;
+        sensor.offset = ValidateValue(sensor.offset, DefaultOffset);
-        }
     }
 }
 
+u16 CalibrationProtocol::ValidateValue(u16 value, u16 default_value) const {
+    if (value == 0) {
+        return default_value;
+    }
+    if (value == 0xFFF) {
+        return default_value;
+    }
+    return value;
+}
+
+s16 CalibrationProtocol::ValidateValue(s16 value, s16 default_value) const {
+    if (value == 0) {
+        return default_value;
+    }
+    if (value == 0xFFF) {
+        return default_value;
+    }
+    return value;
+}
+
 } // namespace InputCommon::Joycon

src/input_common/helpers/joycon_protocol/calibration.h

@@ -53,9 +53,27 @@ public:
     DriverResult GetRingCalibration(RingCalibration& calibration, s16 current_value);
 
 private:
+    /// Returns true if the specified address corresponds to the magic value of user calibration
+    DriverResult HasUserCalibration(SpiAddress address, bool& has_user_calibration);
+
+    /// Converts a raw calibration block to an u16 value containing the x axis value
+    u16 GetXAxisCalibrationValue(std::span<u8> block) const;
+
+    /// Converts a raw calibration block to an u16 value containing the y axis value
+    u16 GetYAxisCalibrationValue(std::span<u8> block) const;
+
+    /// Ensures that all joystick calibration values are set
     void ValidateCalibration(JoyStickCalibration& calibration);
+
+    /// Ensures that all motion calibration values are set
     void ValidateCalibration(MotionCalibration& calibration);
 
+    /// Returns the default value if the value is either zero or 0xFFF
+    u16 ValidateValue(u16 value, u16 default_value) const;
+
+    /// Returns the default value if the value is either zero or 0xFFF
+    s16 ValidateValue(s16 value, s16 default_value) const;
+
     s16 ring_data_max = 0;
     s16 ring_data_default = 0;
     s16 ring_data_min = 0;

src/input_common/helpers/joycon_protocol/common_protocol.cpp

@@ -22,8 +22,8 @@ void JoyconCommonProtocol::SetNonBlocking() {
 }
 
 DriverResult JoyconCommonProtocol::GetDeviceType(ControllerType& controller_type) {
-    std::vector<u8> buffer;
+    std::array<u8, 1> buffer{};
-    const auto result = ReadSPI(CalAddr::DEVICE_TYPE, 1, buffer);
+    const auto result = ReadRawSPI(SpiAddress::DEVICE_TYPE, buffer);
     controller_type = ControllerType::None;
 
     if (result == DriverResult::Success) {
@@ -148,11 +148,13 @@ DriverResult JoyconCommonProtocol::SendVibrationReport(std::span<const u8> buffe
     return SendData(local_buffer);
 }
 
-DriverResult JoyconCommonProtocol::ReadSPI(CalAddr addr, u8 size, std::vector<u8>& output) {
+DriverResult JoyconCommonProtocol::ReadRawSPI(SpiAddress addr, std::span<u8> output) {
+    constexpr std::size_t HeaderSize = 20;
     constexpr std::size_t MaxTries = 10;
+    const auto size = output.size();
     std::size_t tries = 0;
-    std::array<u8, 5> buffer = {0x00, 0x00, 0x00, 0x00, size};
+    std::array<u8, 5> buffer = {0x00, 0x00, 0x00, 0x00, static_cast<u8>(size)};
-    std::vector<u8> local_buffer(size + 20);
+    std::vector<u8> local_buffer{};
 
     buffer[0] = static_cast<u8>(static_cast<u16>(addr) & 0x00FF);
     buffer[1] = static_cast<u8>((static_cast<u16>(addr) & 0xFF00) >> 8);
@@ -167,8 +169,12 @@ DriverResult JoyconCommonProtocol::ReadSPI(CalAddr addr, u8 size, std::vector<u8
         }
     } while (local_buffer[15] != buffer[0] || local_buffer[16] != buffer[1]);
 
+    if (local_buffer.size() < size + HeaderSize) {
+        return DriverResult::WrongReply;
+    }
+
     // Remove header from output
-    output = std::vector<u8>(local_buffer.begin() + 20, local_buffer.begin() + 20 + size);
+    memcpy(output.data(), local_buffer.data() + HeaderSize, size);
     return DriverResult::Success;
 }
 

src/input_common/helpers/joycon_protocol/common_protocol.h

@@ -97,10 +97,29 @@ public:
     /**
      * Reads the SPI memory stored on the joycon
      * @param Initial address location
-     * @param size in bytes to be read
      * @returns output buffer containing the responce
      */
-    DriverResult ReadSPI(CalAddr addr, u8 size, std::vector<u8>& output);
+    DriverResult ReadRawSPI(SpiAddress addr, std::span<u8> output);
+
+    /**
+     * Reads the SPI memory stored on the joycon
+     * @param Initial address location
+     * @returns output object containing the responce
+     */
+    template <typename Output>
+    requires std::is_trivially_copyable_v<Output> DriverResult ReadSPI(SpiAddress addr,
+                                                                       Output& output) {
+        std::array<u8, sizeof(Output)> buffer;
+        output = {};
+
+        const auto result = ReadRawSPI(addr, buffer);
+        if (result != DriverResult::Success) {
+            return result;
+        }
+
+        std::memcpy(&output, buffer.data(), sizeof(Output));
+        return DriverResult::Success;
+    }
 
     /**
      * Enables MCU chip on the joycon

src/input_common/helpers/joycon_protocol/generic_functions.cpp

@@ -71,8 +71,8 @@ DriverResult GenericProtocol::GetBattery(u32& battery_level) {
 
 DriverResult GenericProtocol::GetColor(Color& color) {
     ScopedSetBlocking sb(this);
-    std::vector<u8> buffer;
+    std::array<u8, 12> buffer{};
-    const auto result = ReadSPI(CalAddr::COLOR_DATA, 12, buffer);
+    const auto result = ReadRawSPI(SpiAddress::COLOR_DATA, buffer);
 
     color = {};
     if (result == DriverResult::Success) {
@@ -87,8 +87,8 @@ DriverResult GenericProtocol::GetColor(Color& color) {
 
 DriverResult GenericProtocol::GetSerialNumber(SerialNumber& serial_number) {
     ScopedSetBlocking sb(this);
-    std::vector<u8> buffer;
+    std::array<u8, 16> buffer{};
-    const auto result = ReadSPI(CalAddr::SERIAL_NUMBER, 16, buffer);
+    const auto result = ReadRawSPI(SpiAddress::SERIAL_NUMBER, buffer);
 
     serial_number = {};
     if (result == DriverResult::Success) {

src/input_common/helpers/joycon_protocol/joycon_types.h

@@ -159,13 +159,12 @@ enum class UsbSubCommand : u8 {
     SEND_UART = 0x92,
 };
 
-enum class CalMagic : u8 {
+enum class CalibrationMagic : u8 {
     USR_MAGIC_0 = 0xB2,
     USR_MAGIC_1 = 0xA1,
-    USRR_MAGI_SIZE = 2,
 };
 
-enum class CalAddr {
+enum class SpiAddress {
     SERIAL_NUMBER = 0X6000,
     DEVICE_TYPE = 0X6012,
     COLOR_EXIST = 0X601B,
@@ -396,10 +395,35 @@ struct MotionData {
     u64 delta_timestamp{};
 };
 
+// Output from SPI read command containing user calibration magic
+struct MagicSpiCalibration {
+    CalibrationMagic first;
+    CalibrationMagic second;
+};
+static_assert(sizeof(MagicSpiCalibration) == 0x2, "MagicSpiCalibration is an invalid size");
+
+// Output from SPI read command containing left joystick calibration
+struct JoystickLeftSpiCalibration {
+    std::array<u8, 3> max;
+    std::array<u8, 3> center;
+    std::array<u8, 3> min;
+};
+static_assert(sizeof(JoystickLeftSpiCalibration) == 0x9,
+              "JoystickLeftSpiCalibration is an invalid size");
+
+// Output from SPI read command containing right joystick calibration
+struct JoystickRightSpiCalibration {
+    std::array<u8, 3> center;
+    std::array<u8, 3> min;
+    std::array<u8, 3> max;
+};
+static_assert(sizeof(JoystickRightSpiCalibration) == 0x9,
+              "JoystickRightSpiCalibration is an invalid size");
+
 struct JoyStickAxisCalibration {
-    u16 max{1};
+    u16 max;
-    u16 min{1};
+    u16 min;
-    u16 center{0};
+    u16 center;
 };
 
 struct JoyStickCalibration {
@@ -407,6 +431,14 @@ struct JoyStickCalibration {
     JoyStickAxisCalibration y;
 };
 
+struct ImuSpiCalibration {
+    std::array<s16, 3> accelerometer_offset;
+    std::array<s16, 3> accelerometer_scale;
+    std::array<s16, 3> gyroscope_offset;
+    std::array<s16, 3> gyroscope_scale;
+};
+static_assert(sizeof(ImuSpiCalibration) == 0x18, "ImuSpiCalibration is an invalid size");
+
 struct RingCalibration {
     s16 default_value;
     s16 max_value;
@@ -488,14 +520,6 @@ struct InputReportNfcIr {
 static_assert(sizeof(InputReportNfcIr) == 0x29, "InputReportNfcIr is an invalid size");
 #pragma pack(pop)
 
-struct IMUCalibration {
-    std::array<s16, 3> accelerometer_offset;
-    std::array<s16, 3> accelerometer_scale;
-    std::array<s16, 3> gyroscope_offset;
-    std::array<s16, 3> gyroscope_scale;
-};
-static_assert(sizeof(IMUCalibration) == 0x18, "IMUCalibration is an invalid size");
-
 struct NFCReadBlock {
     u8 start;
     u8 end;

src/input_common/input_poller.cpp

@@ -16,10 +16,10 @@ public:
 
 class InputFromButton final : public Common::Input::InputDevice {
 public:
-    explicit InputFromButton(PadIdentifier identifier_, int button_, bool toggle_, bool inverted_,
+    explicit InputFromButton(PadIdentifier identifier_, int button_, bool turbo_, bool toggle_,
-                             InputEngine* input_engine_)
+                             bool inverted_, InputEngine* input_engine_)
-        : identifier(identifier_), button(button_), toggle(toggle_), inverted(inverted_),
+        : identifier(identifier_), button(button_), turbo(turbo_), toggle(toggle_),
-          input_engine(input_engine_) {
+          inverted(inverted_), input_engine(input_engine_) {
         UpdateCallback engine_callback{[this]() { OnChange(); }};
         const InputIdentifier input_identifier{
             .identifier = identifier,
@@ -40,6 +40,7 @@ public:
             .value = input_engine->GetButton(identifier, button),
             .inverted = inverted,
             .toggle = toggle,
+            .turbo = turbo,
         };
     }
 
@@ -68,6 +69,7 @@ public:
 private:
     const PadIdentifier identifier;
     const int button;
+    const bool turbo;
     const bool toggle;
     const bool inverted;
     int callback_key;
@@ -77,10 +79,10 @@ private:
 
 class InputFromHatButton final : public Common::Input::InputDevice {
 public:
-    explicit InputFromHatButton(PadIdentifier identifier_, int button_, u8 direction_, bool toggle_,
+    explicit InputFromHatButton(PadIdentifier identifier_, int button_, u8 direction_, bool turbo_,
-                                bool inverted_, InputEngine* input_engine_)
+                                bool toggle_, bool inverted_, InputEngine* input_engine_)
-        : identifier(identifier_), button(button_), direction(direction_), toggle(toggle_),
+        : identifier(identifier_), button(button_), direction(direction_), turbo(turbo_),
-          inverted(inverted_), input_engine(input_engine_) {
+          toggle(toggle_), inverted(inverted_), input_engine(input_engine_) {
         UpdateCallback engine_callback{[this]() { OnChange(); }};
         const InputIdentifier input_identifier{
             .identifier = identifier,
@@ -101,6 +103,7 @@ public:
             .value = input_engine->GetHatButton(identifier, button, direction),
             .inverted = inverted,
             .toggle = toggle,
+            .turbo = turbo,
         };
     }
 
@@ -130,6 +133,7 @@ private:
     const PadIdentifier identifier;
     const int button;
     const u8 direction;
+    const bool turbo;
     const bool toggle;
     const bool inverted;
     int callback_key;
@@ -853,14 +857,15 @@ std::unique_ptr<Common::Input::InputDevice> InputFactory::CreateButtonDevice(
     const auto keyboard_key = params.Get("code", 0);
     const auto toggle = params.Get("toggle", false) != 0;
     const auto inverted = params.Get("inverted", false) != 0;
+    const auto turbo = params.Get("turbo", false) != 0;
     input_engine->PreSetController(identifier);
     input_engine->PreSetButton(identifier, button_id);
     input_engine->PreSetButton(identifier, keyboard_key);
     if (keyboard_key != 0) {
-        return std::make_unique<InputFromButton>(identifier, keyboard_key, toggle, inverted,
+        return std::make_unique<InputFromButton>(identifier, keyboard_key, turbo, toggle, inverted,
                                                  input_engine.get());
     }
-    return std::make_unique<InputFromButton>(identifier, button_id, toggle, inverted,
+    return std::make_unique<InputFromButton>(identifier, button_id, turbo, toggle, inverted,
                                              input_engine.get());
 }
 
@@ -876,11 +881,12 @@ std::unique_ptr<Common::Input::InputDevice> InputFactory::CreateHatButtonDevice(
     const auto direction = input_engine->GetHatButtonId(params.Get("direction", ""));
     const auto toggle = params.Get("toggle", false) != 0;
     const auto inverted = params.Get("inverted", false) != 0;
+    const auto turbo = params.Get("turbo", false) != 0;
 
     input_engine->PreSetController(identifier);
     input_engine->PreSetHatButton(identifier, button_id);
-    return std::make_unique<InputFromHatButton>(identifier, button_id, direction, toggle, inverted,
+    return std::make_unique<InputFromHatButton>(identifier, button_id, direction, turbo, toggle,
-                                                input_engine.get());
+                                                inverted, input_engine.get());
 }
 
 std::unique_ptr<Common::Input::InputDevice> InputFactory::CreateStickDevice(

src/shader_recompiler/backend/glasm/emit_glasm_image.cpp

@@ -532,7 +532,7 @@ void EmitImageFetch(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
 }
 
 void EmitImageQueryDimensions(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
-                              ScalarS32 lod) {
+                              ScalarS32 lod, [[maybe_unused]] const IR::Value& skip_mips) {
     const auto info{inst.Flags<IR::TextureInstInfo>()};
     const std::string texture{Texture(ctx, info, index)};
     const std::string_view type{TextureType(info)};

src/shader_recompiler/backend/glasm/emit_glasm_instructions.h

@@ -581,7 +581,7 @@ void EmitImageGatherDref(EmitContext& ctx, IR::Inst& inst, const IR::Value& inde
 void EmitImageFetch(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
                     const IR::Value& coord, const IR::Value& offset, ScalarS32 lod, ScalarS32 ms);
 void EmitImageQueryDimensions(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
-                              ScalarS32 lod);
+                              ScalarS32 lod, const IR::Value& skip_mips);
 void EmitImageQueryLod(EmitContext& ctx, IR::Inst& inst, const IR::Value& index, Register coord);
 void EmitImageGradient(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
                        const IR::Value& coord, const IR::Value& derivatives,

src/shader_recompiler/backend/glsl/emit_glsl_image.cpp

@@ -460,27 +460,27 @@ void EmitImageFetch(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
 }
 
 void EmitImageQueryDimensions(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
-                              std::string_view lod) {
+                              std::string_view lod, const IR::Value& skip_mips_val) {
     const auto info{inst.Flags<IR::TextureInstInfo>()};
     const auto texture{Texture(ctx, info, index)};
+    const bool skip_mips{skip_mips_val.U1()};
+    const auto mips{
+        [&] { return skip_mips ? "0u" : fmt::format("uint(textureQueryLevels({}))", texture); }};
     switch (info.type) {
     case TextureType::Color1D:
-        return ctx.AddU32x4(
+        return ctx.AddU32x4("{}=uvec4(uint(textureSize({},int({}))),0u,0u,{});", inst, texture, lod,
-            "{}=uvec4(uint(textureSize({},int({}))),0u,0u,uint(textureQueryLevels({})));", inst,
+                            mips());
-            texture, lod, texture);
     case TextureType::ColorArray1D:
     case TextureType::Color2D:
     case TextureType::ColorCube:
     case TextureType::Color2DRect:
-        return ctx.AddU32x4(
+        return ctx.AddU32x4("{}=uvec4(uvec2(textureSize({},int({}))),0u,{});", inst, texture, lod,
-            "{}=uvec4(uvec2(textureSize({},int({}))),0u,uint(textureQueryLevels({})));", inst,
+                            mips());
-            texture, lod, texture);
     case TextureType::ColorArray2D:
     case TextureType::Color3D:
     case TextureType::ColorArrayCube:
-        return ctx.AddU32x4(
+        return ctx.AddU32x4("{}=uvec4(uvec3(textureSize({},int({}))),{});", inst, texture, lod,
-            "{}=uvec4(uvec3(textureSize({},int({}))),uint(textureQueryLevels({})));", inst, texture,
+                            mips());
-            lod, texture);
     case TextureType::Buffer:
         throw NotImplementedException("EmitImageQueryDimensions Texture buffers");
     }

src/shader_recompiler/backend/glsl/emit_glsl_instructions.h

@@ -654,7 +654,7 @@ void EmitImageFetch(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
                     std::string_view coords, std::string_view offset, std::string_view lod,
                     std::string_view ms);
 void EmitImageQueryDimensions(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
-                              std::string_view lod);
+                              std::string_view lod, const IR::Value& skip_mips);
 void EmitImageQueryLod(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,
                        std::string_view coords);
 void EmitImageGradient(EmitContext& ctx, IR::Inst& inst, const IR::Value& index,

src/shader_recompiler/backend/spirv/emit_spirv_image.cpp

@@ -445,11 +445,13 @@ Id EmitImageFetch(EmitContext& ctx, IR::Inst* inst, const IR::Value& index, Id c
                 TextureImage(ctx, info, index), coords, operands.MaskOptional(), operands.Span());
 }
 
-Id EmitImageQueryDimensions(EmitContext& ctx, IR::Inst* inst, const IR::Value& index, Id lod) {
+Id EmitImageQueryDimensions(EmitContext& ctx, IR::Inst* inst, const IR::Value& index, Id lod,
+                            const IR::Value& skip_mips_val) {
     const auto info{inst->Flags<IR::TextureInstInfo>()};
     const Id image{TextureImage(ctx, info, index)};
     const Id zero{ctx.u32_zero_value};
-    const auto mips{[&] { return ctx.OpImageQueryLevels(ctx.U32[1], image); }};
+    const bool skip_mips{skip_mips_val.U1()};
+    const auto mips{[&] { return skip_mips ? zero : ctx.OpImageQueryLevels(ctx.U32[1], image); }};
     switch (info.type) {
     case TextureType::Color1D:
         return ctx.OpCompositeConstruct(ctx.U32[4], ctx.OpImageQuerySizeLod(ctx.U32[1], image, lod),

src/shader_recompiler/backend/spirv/emit_spirv_instructions.h

@@ -539,7 +539,8 @@ Id EmitImageGatherDref(EmitContext& ctx, IR::Inst* inst, const IR::Value& index,
                        const IR::Value& offset, const IR::Value& offset2, Id dref);
 Id EmitImageFetch(EmitContext& ctx, IR::Inst* inst, const IR::Value& index, Id coords, Id offset,
                   Id lod, Id ms);
-Id EmitImageQueryDimensions(EmitContext& ctx, IR::Inst* inst, const IR::Value& index, Id lod);
+Id EmitImageQueryDimensions(EmitContext& ctx, IR::Inst* inst, const IR::Value& index, Id lod,
+                            const IR::Value& skip_mips);
 Id EmitImageQueryLod(EmitContext& ctx, IR::Inst* inst, const IR::Value& index, Id coords);
 Id EmitImageGradient(EmitContext& ctx, IR::Inst* inst, const IR::Value& index, Id coords,
                      Id derivates, Id offset, Id lod_clamp);

src/shader_recompiler/frontend/ir/ir_emitter.cpp

@@ -1846,15 +1846,16 @@ Value IREmitter::ImageFetch(const Value& handle, const Value& coords, const Valu
     return Inst(op, Flags{info}, handle, coords, offset, lod, multisampling);
 }
 
-Value IREmitter::ImageQueryDimension(const Value& handle, const IR::U32& lod) {
+Value IREmitter::ImageQueryDimension(const Value& handle, const IR::U32& lod,
+                                     const IR::U1& skip_mips) {
     const Opcode op{handle.IsImmediate() ? Opcode::BoundImageQueryDimensions
                                          : Opcode::BindlessImageQueryDimensions};
-    return Inst(op, handle, lod);
+    return Inst(op, handle, lod, skip_mips);
 }
 
 Value IREmitter::ImageQueryDimension(const Value& handle, const IR::U32& lod,
-                                     TextureInstInfo info) {
+                                     const IR::U1& skip_mips, TextureInstInfo info) {
-    return Inst(Opcode::ImageQueryDimensions, Flags{info}, handle, lod);
+    return Inst(Opcode::ImageQueryDimensions, Flags{info}, handle, lod, skip_mips);
 }
 
 Value IREmitter::ImageQueryLod(const Value& handle, const Value& coords, TextureInstInfo info) {

src/shader_recompiler/frontend/ir/ir_emitter.h

@@ -320,9 +320,10 @@ public:
     [[nodiscard]] F32 ImageSampleDrefExplicitLod(const Value& handle, const Value& coords,
                                                  const F32& dref, const F32& lod,
                                                  const Value& offset, TextureInstInfo info);
-    [[nodiscard]] Value ImageQueryDimension(const Value& handle, const IR::U32& lod);
     [[nodiscard]] Value ImageQueryDimension(const Value& handle, const IR::U32& lod,
-                                            TextureInstInfo info);
+                                            const IR::U1& skip_mips);
+    [[nodiscard]] Value ImageQueryDimension(const Value& handle, const IR::U32& lod,
+                                            const IR::U1& skip_mips, TextureInstInfo info);
 
     [[nodiscard]] Value ImageQueryLod(const Value& handle, const Value& coords,
                                       TextureInstInfo info);

src/shader_recompiler/frontend/ir/opcodes.inc

@@ -482,7 +482,7 @@ OPCODE(BindlessImageSampleDrefExplicitLod,                  F32,            U32,
 OPCODE(BindlessImageGather,                                 F32x4,          U32,            Opaque,         Opaque,         Opaque,                         )
 OPCODE(BindlessImageGatherDref,                             F32x4,          U32,            Opaque,         Opaque,         Opaque,         F32,            )
 OPCODE(BindlessImageFetch,                                  F32x4,          U32,            Opaque,         Opaque,         U32,            Opaque,         )
-OPCODE(BindlessImageQueryDimensions,                        U32x4,          U32,            U32,                                                            )
+OPCODE(BindlessImageQueryDimensions,                        U32x4,          U32,            U32,            U1,                                             )
 OPCODE(BindlessImageQueryLod,                               F32x4,          U32,            Opaque,                                                         )
 OPCODE(BindlessImageGradient,                               F32x4,          U32,            Opaque,         Opaque,         Opaque,         Opaque,         )
 OPCODE(BindlessImageRead,                                   U32x4,          U32,            Opaque,                                                         )
@@ -495,7 +495,7 @@ OPCODE(BoundImageSampleDrefExplicitLod,                     F32,            U32,
 OPCODE(BoundImageGather,                                    F32x4,          U32,            Opaque,         Opaque,         Opaque,                         )
 OPCODE(BoundImageGatherDref,                                F32x4,          U32,            Opaque,         Opaque,         Opaque,         F32,            )
 OPCODE(BoundImageFetch,                                     F32x4,          U32,            Opaque,         Opaque,         U32,            Opaque,         )
-OPCODE(BoundImageQueryDimensions,                           U32x4,          U32,            U32,                                                            )
+OPCODE(BoundImageQueryDimensions,                           U32x4,          U32,            U32,            U1,                                             )
 OPCODE(BoundImageQueryLod,                                  F32x4,          U32,            Opaque,                                                         )
 OPCODE(BoundImageGradient,                                  F32x4,          U32,            Opaque,         Opaque,         Opaque,         Opaque,         )
 OPCODE(BoundImageRead,                                      U32x4,          U32,            Opaque,                                                         )
@@ -508,7 +508,7 @@ OPCODE(ImageSampleDrefExplicitLod,                          F32,            Opaq
 OPCODE(ImageGather,                                         F32x4,          Opaque,         Opaque,         Opaque,         Opaque,                         )
 OPCODE(ImageGatherDref,                                     F32x4,          Opaque,         Opaque,         Opaque,         Opaque,         F32,            )
 OPCODE(ImageFetch,                                          F32x4,          Opaque,         Opaque,         Opaque,         U32,            Opaque,         )
-OPCODE(ImageQueryDimensions,                                U32x4,          Opaque,         U32,                                                            )
+OPCODE(ImageQueryDimensions,                                U32x4,          Opaque,         U32,            U1,                                             )
 OPCODE(ImageQueryLod,                                       F32x4,          Opaque,         Opaque,                                                         )
 OPCODE(ImageGradient,                                       F32x4,          Opaque,         Opaque,         Opaque,         Opaque,         Opaque,         )
 OPCODE(ImageRead,                                           U32x4,          Opaque,         Opaque,                                                         )

src/shader_recompiler/frontend/maxwell/translate/impl/texture_query.cpp

@@ -15,11 +15,13 @@ enum class Mode : u64 {
     SamplePos = 5,
 };
 
-IR::Value Query(TranslatorVisitor& v, const IR::U32& handle, Mode mode, IR::Reg src_reg) {
+IR::Value Query(TranslatorVisitor& v, const IR::U32& handle, Mode mode, IR::Reg src_reg, u64 mask) {
     switch (mode) {
     case Mode::Dimension: {
+        const bool needs_num_mips{((mask >> 3) & 1) != 0};
+        const IR::U1 skip_mips{v.ir.Imm1(!needs_num_mips)};
         const IR::U32 lod{v.X(src_reg)};
-        return v.ir.ImageQueryDimension(handle, lod);
+        return v.ir.ImageQueryDimension(handle, lod, skip_mips);
     }
     case Mode::TextureType:
     case Mode::SamplePos:
@@ -46,7 +48,7 @@ void Impl(TranslatorVisitor& v, u64 insn, std::optional<u32> cbuf_offset) {
         handle = v.X(src_reg);
         ++src_reg;
     }
-    const IR::Value query{Query(v, handle, txq.mode, src_reg)};
+    const IR::Value query{Query(v, handle, txq.mode, src_reg, txq.mask)};
     IR::Reg dest_reg{txq.dest_reg};
     for (int element = 0; element < 4; ++element) {
         if (((txq.mask >> element) & 1) == 0) {

src/shader_recompiler/ir_opt/texture_pass.cpp

@@ -355,21 +355,21 @@ TextureInst MakeInst(Environment& env, IR::Block* block, IR::Inst& inst) {
     };
 }
 
-TextureType ReadTextureType(Environment& env, const ConstBufferAddr& cbuf) {
+u32 GetTextureHandle(Environment& env, const ConstBufferAddr& cbuf) {
     const u32 secondary_index{cbuf.has_secondary ? cbuf.secondary_index : cbuf.index};
     const u32 secondary_offset{cbuf.has_secondary ? cbuf.secondary_offset : cbuf.offset};
     const u32 lhs_raw{env.ReadCbufValue(cbuf.index, cbuf.offset) << cbuf.shift_left};
     const u32 rhs_raw{env.ReadCbufValue(secondary_index, secondary_offset)
                       << cbuf.secondary_shift_left};
-    return env.ReadTextureType(lhs_raw | rhs_raw);
+    return lhs_raw | rhs_raw;
+}
+
+TextureType ReadTextureType(Environment& env, const ConstBufferAddr& cbuf) {
+    return env.ReadTextureType(GetTextureHandle(env, cbuf));
 }
 
 TexturePixelFormat ReadTexturePixelFormat(Environment& env, const ConstBufferAddr& cbuf) {
-    const u32 secondary_index{cbuf.has_secondary ? cbuf.secondary_index : cbuf.index};
+    return env.ReadTexturePixelFormat(GetTextureHandle(env, cbuf));
-    const u32 secondary_offset{cbuf.has_secondary ? cbuf.secondary_offset : cbuf.offset};
-    const u32 lhs_raw{env.ReadCbufValue(cbuf.index, cbuf.offset)};
-    const u32 rhs_raw{env.ReadCbufValue(secondary_index, secondary_offset)};
-    return env.ReadTexturePixelFormat(lhs_raw | rhs_raw);
 }
 
 class Descriptors {
@@ -386,8 +386,10 @@ public:
         return Add(texture_buffer_descriptors, desc, [&desc](const auto& existing) {
             return desc.cbuf_index == existing.cbuf_index &&
                    desc.cbuf_offset == existing.cbuf_offset &&
+                   desc.shift_left == existing.shift_left &&
                    desc.secondary_cbuf_index == existing.secondary_cbuf_index &&
                    desc.secondary_cbuf_offset == existing.secondary_cbuf_offset &&
+                   desc.secondary_shift_left == existing.secondary_shift_left &&
                    desc.count == existing.count && desc.size_shift == existing.size_shift &&
                    desc.has_secondary == existing.has_secondary;
         });
@@ -405,15 +407,20 @@ public:
     }
 
     u32 Add(const TextureDescriptor& desc) {
-        return Add(texture_descriptors, desc, [&desc](const auto& existing) {
+        const u32 index{Add(texture_descriptors, desc, [&desc](const auto& existing) {
             return desc.type == existing.type && desc.is_depth == existing.is_depth &&
                    desc.has_secondary == existing.has_secondary &&
                    desc.cbuf_index == existing.cbuf_index &&
                    desc.cbuf_offset == existing.cbuf_offset &&
+                   desc.shift_left == existing.shift_left &&
                    desc.secondary_cbuf_index == existing.secondary_cbuf_index &&
                    desc.secondary_cbuf_offset == existing.secondary_cbuf_offset &&
+                   desc.secondary_shift_left == existing.secondary_shift_left &&
                    desc.count == existing.count && desc.size_shift == existing.size_shift;
-        });
+        })};
+        // TODO: Read this from TIC
+        texture_descriptors[index].is_multisample |= desc.is_multisample;
+        return index;
     }
 
     u32 Add(const ImageDescriptor& desc) {
@@ -452,7 +459,8 @@ void PatchImageSampleImplicitLod(IR::Block& block, IR::Inst& inst) {
     const IR::Value coord(inst.Arg(1));
     const IR::Value handle(ir.Imm32(0));
     const IR::U32 lod{ir.Imm32(0)};
-    const IR::Value texture_size = ir.ImageQueryDimension(handle, lod, info);
+    const IR::U1 skip_mips{ir.Imm1(true)};
+    const IR::Value texture_size = ir.ImageQueryDimension(handle, lod, skip_mips, info);
     inst.SetArg(
         1, ir.CompositeConstruct(
                ir.FPMul(IR::F32(ir.CompositeExtract(coord, 0)),

src/video_core/host_shaders/CMakeLists.txt

@@ -45,6 +45,8 @@ set(SHADER_FILES
     smaa_neighborhood_blending.vert
     smaa_neighborhood_blending.frag
     vulkan_blit_depth_stencil.frag
+    vulkan_color_clear.frag
+    vulkan_color_clear.vert
     vulkan_fidelityfx_fsr_easu_fp16.comp
     vulkan_fidelityfx_fsr_easu_fp32.comp
     vulkan_fidelityfx_fsr_rcas_fp16.comp

src/video_core/renderer_vulkan/blit_image.cpp

@@ -12,6 +12,8 @@
 #include "video_core/host_shaders/convert_s8d24_to_abgr8_frag_spv.h"
 #include "video_core/host_shaders/full_screen_triangle_vert_spv.h"
 #include "video_core/host_shaders/vulkan_blit_depth_stencil_frag_spv.h"
+#include "video_core/host_shaders/vulkan_color_clear_frag_spv.h"
+#include "video_core/host_shaders/vulkan_color_clear_vert_spv.h"
 #include "video_core/renderer_vulkan/blit_image.h"
 #include "video_core/renderer_vulkan/maxwell_to_vk.h"
 #include "video_core/renderer_vulkan/vk_scheduler.h"
@@ -69,10 +71,11 @@ constexpr VkDescriptorSetLayoutCreateInfo TWO_TEXTURES_DESCRIPTOR_SET_LAYOUT_CRE
     .bindingCount = static_cast<u32>(TWO_TEXTURES_DESCRIPTOR_SET_LAYOUT_BINDINGS.size()),
     .pBindings = TWO_TEXTURES_DESCRIPTOR_SET_LAYOUT_BINDINGS.data(),
 };
-constexpr VkPushConstantRange PUSH_CONSTANT_RANGE{
+template <VkShaderStageFlags stageFlags, size_t size>
-    .stageFlags = VK_SHADER_STAGE_VERTEX_BIT,
+inline constexpr VkPushConstantRange PUSH_CONSTANT_RANGE{
+    .stageFlags = stageFlags,
     .offset = 0,
-    .size = sizeof(PushConstants),
+    .size = static_cast<u32>(size),
 };
 constexpr VkPipelineVertexInputStateCreateInfo PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO{
     .sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO,
@@ -125,10 +128,8 @@ constexpr VkPipelineMultisampleStateCreateInfo PIPELINE_MULTISAMPLE_STATE_CREATE
     .alphaToCoverageEnable = VK_FALSE,
     .alphaToOneEnable = VK_FALSE,
 };
-constexpr std::array DYNAMIC_STATES{
+constexpr std::array DYNAMIC_STATES{VK_DYNAMIC_STATE_VIEWPORT, VK_DYNAMIC_STATE_SCISSOR,
-    VK_DYNAMIC_STATE_VIEWPORT,
+                                    VK_DYNAMIC_STATE_BLEND_CONSTANTS};
-    VK_DYNAMIC_STATE_SCISSOR,
-};
 constexpr VkPipelineDynamicStateCreateInfo PIPELINE_DYNAMIC_STATE_CREATE_INFO{
     .sType = VK_STRUCTURE_TYPE_PIPELINE_DYNAMIC_STATE_CREATE_INFO,
     .pNext = nullptr,
@@ -205,15 +206,15 @@ inline constexpr VkSamplerCreateInfo SAMPLER_CREATE_INFO{
 };
 
 constexpr VkPipelineLayoutCreateInfo PipelineLayoutCreateInfo(
-    const VkDescriptorSetLayout* set_layout) {
+    const VkDescriptorSetLayout* set_layout, vk::Span<VkPushConstantRange> push_constants) {
     return VkPipelineLayoutCreateInfo{
         .sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO,
         .pNext = nullptr,
         .flags = 0,
-        .setLayoutCount = 1,
+        .setLayoutCount = (set_layout != nullptr ? 1u : 0u),
         .pSetLayouts = set_layout,
-        .pushConstantRangeCount = 1,
+        .pushConstantRangeCount = push_constants.size(),
-        .pPushConstantRanges = &PUSH_CONSTANT_RANGE,
+        .pPushConstantRanges = push_constants.data(),
     };
 }
 
@@ -302,8 +303,7 @@ void UpdateTwoTexturesDescriptorSet(const Device& device, VkDescriptorSet descri
     device.GetLogical().UpdateDescriptorSets(write_descriptor_sets, nullptr);
 }
 
-void BindBlitState(vk::CommandBuffer cmdbuf, VkPipelineLayout layout, const Region2D& dst_region,
+void BindBlitState(vk::CommandBuffer cmdbuf, const Region2D& dst_region) {
-                   const Region2D& src_region, const Extent3D& src_size = {1, 1, 1}) {
     const VkOffset2D offset{
         .x = std::min(dst_region.start.x, dst_region.end.x),
         .y = std::min(dst_region.start.y, dst_region.end.y),
@@ -325,6 +325,13 @@ void BindBlitState(vk::CommandBuffer cmdbuf, VkPipelineLayout layout, const Regi
         .offset = offset,
         .extent = extent,
     };
+    cmdbuf.SetViewport(0, viewport);
+    cmdbuf.SetScissor(0, scissor);
+}
+
+void BindBlitState(vk::CommandBuffer cmdbuf, VkPipelineLayout layout, const Region2D& dst_region,
+                   const Region2D& src_region, const Extent3D& src_size = {1, 1, 1}) {
+    BindBlitState(cmdbuf, dst_region);
     const float scale_x = static_cast<float>(src_region.end.x - src_region.start.x) /
                           static_cast<float>(src_size.width);
     const float scale_y = static_cast<float>(src_region.end.y - src_region.start.y) /
@@ -335,8 +342,6 @@ void BindBlitState(vk::CommandBuffer cmdbuf, VkPipelineLayout layout, const Regi
                        static_cast<float>(src_region.start.y) /
                            static_cast<float>(src_size.height)},
     };
-    cmdbuf.SetViewport(0, viewport);
-    cmdbuf.SetScissor(0, scissor);
     cmdbuf.PushConstants(layout, VK_SHADER_STAGE_VERTEX_BIT, push_constants);
 }
 
@@ -408,13 +413,20 @@ BlitImageHelper::BlitImageHelper(const Device& device_, Scheduler& scheduler_,
           descriptor_pool.Allocator(*one_texture_set_layout, TEXTURE_DESCRIPTOR_BANK_INFO<1>)},
       two_textures_descriptor_allocator{
           descriptor_pool.Allocator(*two_textures_set_layout, TEXTURE_DESCRIPTOR_BANK_INFO<2>)},
-      one_texture_pipeline_layout(device.GetLogical().CreatePipelineLayout(
+      one_texture_pipeline_layout(device.GetLogical().CreatePipelineLayout(PipelineLayoutCreateInfo(
-          PipelineLayoutCreateInfo(one_texture_set_layout.address()))),
+          one_texture_set_layout.address(),
-      two_textures_pipeline_layout(device.GetLogical().CreatePipelineLayout(
+          PUSH_CONSTANT_RANGE<VK_SHADER_STAGE_VERTEX_BIT, sizeof(PushConstants)>))),
-          PipelineLayoutCreateInfo(two_textures_set_layout.address()))),
+      two_textures_pipeline_layout(
+          device.GetLogical().CreatePipelineLayout(PipelineLayoutCreateInfo(
+              two_textures_set_layout.address(),
+              PUSH_CONSTANT_RANGE<VK_SHADER_STAGE_VERTEX_BIT, sizeof(PushConstants)>))),
+      clear_color_pipeline_layout(device.GetLogical().CreatePipelineLayout(PipelineLayoutCreateInfo(
+          nullptr, PUSH_CONSTANT_RANGE<VK_SHADER_STAGE_FRAGMENT_BIT, sizeof(float) * 4>))),
       full_screen_vert(BuildShader(device, FULL_SCREEN_TRIANGLE_VERT_SPV)),
       blit_color_to_color_frag(BuildShader(device, BLIT_COLOR_FLOAT_FRAG_SPV)),
       blit_depth_stencil_frag(BuildShader(device, VULKAN_BLIT_DEPTH_STENCIL_FRAG_SPV)),
+      clear_color_vert(BuildShader(device, VULKAN_COLOR_CLEAR_VERT_SPV)),
+      clear_color_frag(BuildShader(device, VULKAN_COLOR_CLEAR_FRAG_SPV)),
       convert_depth_to_float_frag(BuildShader(device, CONVERT_DEPTH_TO_FLOAT_FRAG_SPV)),
       convert_float_to_depth_frag(BuildShader(device, CONVERT_FLOAT_TO_DEPTH_FRAG_SPV)),
       convert_abgr8_to_d24s8_frag(BuildShader(device, CONVERT_ABGR8_TO_D24S8_FRAG_SPV)),
@@ -553,6 +565,30 @@ void BlitImageHelper::ConvertS8D24ToABGR8(const Framebuffer* dst_framebuffer,
     ConvertDepthStencil(*convert_s8d24_to_abgr8_pipeline, dst_framebuffer, src_image_view);
 }
 
+void BlitImageHelper::ClearColor(const Framebuffer* dst_framebuffer, u8 color_mask,
+                                 const std::array<f32, 4>& clear_color,
+                                 const Region2D& dst_region) {
+    const BlitImagePipelineKey key{
+        .renderpass = dst_framebuffer->RenderPass(),
+        .operation = Tegra::Engines::Fermi2D::Operation::BlendPremult,
+    };
+    const VkPipeline pipeline = FindOrEmplaceClearColorPipeline(key);
+    const VkPipelineLayout layout = *clear_color_pipeline_layout;
+    scheduler.RequestRenderpass(dst_framebuffer);
+    scheduler.Record(
+        [pipeline, layout, color_mask, clear_color, dst_region](vk::CommandBuffer cmdbuf) {
+            cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline);
+            const std::array blend_color = {
+                (color_mask & 0x1) ? 1.0f : 0.0f, (color_mask & 0x2) ? 1.0f : 0.0f,
+                (color_mask & 0x4) ? 1.0f : 0.0f, (color_mask & 0x8) ? 1.0f : 0.0f};
+            cmdbuf.SetBlendConstants(blend_color.data());
+            BindBlitState(cmdbuf, dst_region);
+            cmdbuf.PushConstants(layout, VK_SHADER_STAGE_FRAGMENT_BIT, clear_color);
+            cmdbuf.Draw(3, 1, 0, 0);
+        });
+    scheduler.InvalidateState();
+}
+
 void BlitImageHelper::Convert(VkPipeline pipeline, const Framebuffer* dst_framebuffer,
                               const ImageView& src_image_view) {
     const VkPipelineLayout layout = *one_texture_pipeline_layout;
@@ -728,6 +764,58 @@ VkPipeline BlitImageHelper::FindOrEmplaceDepthStencilPipeline(const BlitImagePip
     return *blit_depth_stencil_pipelines.back();
 }
 
+VkPipeline BlitImageHelper::FindOrEmplaceClearColorPipeline(const BlitImagePipelineKey& key) {
+    const auto it = std::ranges::find(clear_color_keys, key);
+    if (it != clear_color_keys.end()) {
+        return *clear_color_pipelines[std::distance(clear_color_keys.begin(), it)];
+    }
+    clear_color_keys.push_back(key);
+    const std::array stages = MakeStages(*clear_color_vert, *clear_color_frag);
+    const VkPipelineColorBlendAttachmentState color_blend_attachment_state{
+        .blendEnable = VK_TRUE,
+        .srcColorBlendFactor = VK_BLEND_FACTOR_CONSTANT_COLOR,
+        .dstColorBlendFactor = VK_BLEND_FACTOR_ONE_MINUS_CONSTANT_COLOR,
+        .colorBlendOp = VK_BLEND_OP_ADD,
+        .srcAlphaBlendFactor = VK_BLEND_FACTOR_CONSTANT_ALPHA,
+        .dstAlphaBlendFactor = VK_BLEND_FACTOR_ONE_MINUS_CONSTANT_ALPHA,
+        .alphaBlendOp = VK_BLEND_OP_ADD,
+        .colorWriteMask = VK_COLOR_COMPONENT_R_BIT | VK_COLOR_COMPONENT_G_BIT |
+                          VK_COLOR_COMPONENT_B_BIT | VK_COLOR_COMPONENT_A_BIT,
+    };
+    const VkPipelineColorBlendStateCreateInfo color_blend_state_generic_create_info{
+        .sType = VK_STRUCTURE_TYPE_PIPELINE_COLOR_BLEND_STATE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .logicOpEnable = VK_FALSE,
+        .logicOp = VK_LOGIC_OP_CLEAR,
+        .attachmentCount = 1,
+        .pAttachments = &color_blend_attachment_state,
+        .blendConstants = {0.0f, 0.0f, 0.0f, 0.0f},
+    };
+    clear_color_pipelines.push_back(device.GetLogical().CreateGraphicsPipeline({
+        .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO,
+        .pNext = nullptr,
+        .flags = 0,
+        .stageCount = static_cast<u32>(stages.size()),
+        .pStages = stages.data(),
+        .pVertexInputState = &PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO,
+        .pInputAssemblyState = &PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO,
+        .pTessellationState = nullptr,
+        .pViewportState = &PIPELINE_VIEWPORT_STATE_CREATE_INFO,
+        .pRasterizationState = &PIPELINE_RASTERIZATION_STATE_CREATE_INFO,
+        .pMultisampleState = &PIPELINE_MULTISAMPLE_STATE_CREATE_INFO,
+        .pDepthStencilState = &PIPELINE_DEPTH_STENCIL_STATE_CREATE_INFO,
+        .pColorBlendState = &color_blend_state_generic_create_info,
+        .pDynamicState = &PIPELINE_DYNAMIC_STATE_CREATE_INFO,
+        .layout = *clear_color_pipeline_layout,
+        .renderPass = key.renderpass,
+        .subpass = 0,
+        .basePipelineHandle = VK_NULL_HANDLE,
+        .basePipelineIndex = 0,
+    }));
+    return *clear_color_pipelines.back();
+}
+
 void BlitImageHelper::ConvertPipeline(vk::Pipeline& pipeline, VkRenderPass renderpass,
                                       bool is_target_depth) {
     if (pipeline) {

src/video_core/renderer_vulkan/blit_image.h

@@ -61,6 +61,9 @@ public:
 
     void ConvertS8D24ToABGR8(const Framebuffer* dst_framebuffer, ImageView& src_image_view);
 
+    void ClearColor(const Framebuffer* dst_framebuffer, u8 color_mask,
+                    const std::array<f32, 4>& clear_color, const Region2D& dst_region);
+
 private:
     void Convert(VkPipeline pipeline, const Framebuffer* dst_framebuffer,
                  const ImageView& src_image_view);
@@ -72,6 +75,8 @@ private:
 
     [[nodiscard]] VkPipeline FindOrEmplaceDepthStencilPipeline(const BlitImagePipelineKey& key);
 
+    [[nodiscard]] VkPipeline FindOrEmplaceClearColorPipeline(const BlitImagePipelineKey& key);
+
     void ConvertPipeline(vk::Pipeline& pipeline, VkRenderPass renderpass, bool is_target_depth);
 
     void ConvertDepthToColorPipeline(vk::Pipeline& pipeline, VkRenderPass renderpass);
@@ -97,9 +102,12 @@ private:
     DescriptorAllocator two_textures_descriptor_allocator;
     vk::PipelineLayout one_texture_pipeline_layout;
     vk::PipelineLayout two_textures_pipeline_layout;
+    vk::PipelineLayout clear_color_pipeline_layout;
     vk::ShaderModule full_screen_vert;
     vk::ShaderModule blit_color_to_color_frag;
     vk::ShaderModule blit_depth_stencil_frag;
+    vk::ShaderModule clear_color_vert;
+    vk::ShaderModule clear_color_frag;
     vk::ShaderModule convert_depth_to_float_frag;
     vk::ShaderModule convert_float_to_depth_frag;
     vk::ShaderModule convert_abgr8_to_d24s8_frag;
@@ -112,6 +120,8 @@ private:
     std::vector<vk::Pipeline> blit_color_pipelines;
     std::vector<BlitImagePipelineKey> blit_depth_stencil_keys;
     std::vector<vk::Pipeline> blit_depth_stencil_pipelines;
+    std::vector<BlitImagePipelineKey> clear_color_keys;
+    std::vector<vk::Pipeline> clear_color_pipelines;
     vk::Pipeline convert_d32_to_r32_pipeline;
     vk::Pipeline convert_r32_to_d32_pipeline;
     vk::Pipeline convert_d16_to_r16_pipeline;

src/video_core/renderer_vulkan/vk_rasterizer.cpp

@@ -394,7 +394,15 @@ void RasterizerVulkan::Clear(u32 layer_count) {
                 cmdbuf.ClearAttachments(attachment, clear_rect);
             });
         } else {
-            UNIMPLEMENTED_MSG("Unimplemented Clear only the specified channel");
+            u8 color_mask = static_cast<u8>(regs.clear_surface.R | regs.clear_surface.G << 1 |
+                                            regs.clear_surface.B << 2 | regs.clear_surface.A << 3);
+            Region2D dst_region = {
+                Offset2D{.x = clear_rect.rect.offset.x, .y = clear_rect.rect.offset.y},
+                Offset2D{.x = clear_rect.rect.offset.x +
+                              static_cast<s32>(clear_rect.rect.extent.width),
+                         .y = clear_rect.rect.offset.y +
+                              static_cast<s32>(clear_rect.rect.extent.height)}};
+            blit_image.ClearColor(framebuffer, color_mask, regs.clear_color, dst_region);
         }
     }
 

src/yuzu/applets/qt_software_keyboard.cpp

@@ -575,7 +575,7 @@ void QtSoftwareKeyboardDialog::MoveAndResizeWindow(QPoint pos, QSize size) {
     QDialog::resize(size);
 
     // High DPI
-    const float dpi_scale = qApp->screenAt(pos)->logicalDotsPerInch() / 96.0f;
+    const float dpi_scale = screen()->logicalDotsPerInch() / 96.0f;
 
     RescaleKeyboardElements(size.width(), size.height(), dpi_scale);
 }

src/yuzu/configuration/configure_input_player.cpp

@@ -182,12 +182,13 @@ QString ConfigureInputPlayer::ButtonToText(const Common::ParamPackage& param) {
     const QString toggle = QString::fromStdString(param.Get("toggle", false) ? "~" : "");
     const QString inverted = QString::fromStdString(param.Get("inverted", false) ? "!" : "");
     const QString invert = QString::fromStdString(param.Get("invert", "+") == "-" ? "-" : "");
+    const QString turbo = QString::fromStdString(param.Get("turbo", false) ? "$" : "");
     const auto common_button_name = input_subsystem->GetButtonName(param);
 
     // Retrieve the names from Qt
     if (param.Get("engine", "") == "keyboard") {
         const QString button_str = GetKeyName(param.Get("code", 0));
-        return QObject::tr("%1%2%3").arg(toggle, inverted, button_str);
+        return QObject::tr("%1%2%3%4").arg(turbo, toggle, inverted, button_str);
     }
 
     if (common_button_name == Common::Input::ButtonNames::Invalid) {
@@ -201,7 +202,7 @@ QString ConfigureInputPlayer::ButtonToText(const Common::ParamPackage& param) {
     if (common_button_name == Common::Input::ButtonNames::Value) {
         if (param.Has("hat")) {
             const QString hat = GetDirectionName(param.Get("direction", ""));
-            return QObject::tr("%1%2Hat %3").arg(toggle, inverted, hat);
+            return QObject::tr("%1%2%3Hat %4").arg(turbo, toggle, inverted, hat);
         }
         if (param.Has("axis")) {
             const QString axis = QString::fromStdString(param.Get("axis", ""));
@@ -219,13 +220,13 @@ QString ConfigureInputPlayer::ButtonToText(const Common::ParamPackage& param) {
         }
         if (param.Has("button")) {
             const QString button = QString::fromStdString(param.Get("button", ""));
-            return QObject::tr("%1%2Button %3").arg(toggle, inverted, button);
+            return QObject::tr("%1%2%3Button %4").arg(turbo, toggle, inverted, button);
         }
     }
 
     QString button_name = GetButtonName(common_button_name);
     if (param.Has("hat")) {
-        return QObject::tr("%1%2Hat %3").arg(toggle, inverted, button_name);
+        return QObject::tr("%1%2%3Hat %4").arg(turbo, toggle, inverted, button_name);
     }
     if (param.Has("axis")) {
         return QObject::tr("%1%2Axis %3").arg(toggle, inverted, button_name);
@@ -234,7 +235,7 @@ QString ConfigureInputPlayer::ButtonToText(const Common::ParamPackage& param) {
         return QObject::tr("%1%2Axis %3").arg(toggle, inverted, button_name);
     }
     if (param.Has("button")) {
-        return QObject::tr("%1%2Button %3").arg(toggle, inverted, button_name);
+        return QObject::tr("%1%2%3Button %4").arg(turbo, toggle, inverted, button_name);
     }
 
     return QObject::tr("[unknown]");
@@ -395,6 +396,12 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
                             button_map[button_id]->setText(ButtonToText(param));
                             emulated_controller->SetButtonParam(button_id, param);
                         });
+                        context_menu.addAction(tr("Turbo button"), [&] {
+                            const bool turbo_value = !param.Get("turbo", false);
+                            param.Set("turbo", turbo_value);
+                            button_map[button_id]->setText(ButtonToText(param));
+                            emulated_controller->SetButtonParam(button_id, param);
+                        });
                     }
                     if (param.Has("axis")) {
                         context_menu.addAction(tr("Invert axis"), [&] {

src/yuzu/discord_impl.cpp

@@ -38,7 +38,7 @@ void DiscordImpl::Update() {
         system.GetAppLoader().ReadTitle(title);
     }
     DiscordRichPresence presence{};
-    presence.largeImageKey = "yuzu_logo";
+    presence.largeImageKey = "yuzu_logo_ea";
     presence.largeImageText = "yuzu is an emulator for the Nintendo Switch";
     if (system.IsPoweredOn()) {
         presence.state = title.c_str();

src/yuzu/main.cpp

@@ -680,8 +680,10 @@ void GMainWindow::SoftwareKeyboardShowNormal() {
     const auto y = layout.screen.top;
     const auto w = layout.screen.GetWidth();
     const auto h = layout.screen.GetHeight();
+    const auto scale_ratio = devicePixelRatioF();
 
-    software_keyboard->ShowNormalKeyboard(render_window->mapToGlobal(QPoint(x, y)), QSize(w, h));
+    software_keyboard->ShowNormalKeyboard(render_window->mapToGlobal(QPoint(x, y) / scale_ratio),
+                                          QSize(w, h) / scale_ratio);
 }
 
 void GMainWindow::SoftwareKeyboardShowTextCheck(
@@ -714,9 +716,11 @@ void GMainWindow::SoftwareKeyboardShowInline(
                                                (1.0f - appear_parameters.key_top_scale_y))));
     const auto w = static_cast<int>(layout.screen.GetWidth() * appear_parameters.key_top_scale_x);
     const auto h = static_cast<int>(layout.screen.GetHeight() * appear_parameters.key_top_scale_y);
+    const auto scale_ratio = devicePixelRatioF();
 
     software_keyboard->ShowInlineKeyboard(std::move(appear_parameters),
-                                          render_window->mapToGlobal(QPoint(x, y)), QSize(w, h));
+                                          render_window->mapToGlobal(QPoint(x, y) / scale_ratio),
+                                          QSize(w, h) / scale_ratio);
 }
 
 void GMainWindow::SoftwareKeyboardHideInline() {
@@ -796,10 +800,11 @@ void GMainWindow::WebBrowserOpenWebPage(const std::string& main_url,
         }
 
         const auto& layout = render_window->GetFramebufferLayout();
-        web_browser_view.resize(layout.screen.GetWidth(), layout.screen.GetHeight());
+        const auto scale_ratio = devicePixelRatioF();
-        web_browser_view.move(layout.screen.left, layout.screen.top + menuBar()->height());
+        web_browser_view.resize(layout.screen.GetWidth() / scale_ratio,
-        web_browser_view.setZoomFactor(static_cast<qreal>(layout.screen.GetWidth()) /
+                                layout.screen.GetHeight() / scale_ratio);
-                                       static_cast<qreal>(Layout::ScreenUndocked::Width));
+        web_browser_view.move(layout.screen.left / scale_ratio,
+                              (layout.screen.top / scale_ratio) + menuBar()->height());
 
         web_browser_view.setFocus();
         web_browser_view.show();
@@ -4390,6 +4395,55 @@ void GMainWindow::changeEvent(QEvent* event) {
 #undef main
 #endif
 
+static void SetHighDPIAttributes() {
+#ifdef _WIN32
+    // For Windows, we want to avoid scaling artifacts on fractional scaling ratios.
+    // This is done by setting the optimal scaling policy for the primary screen.
+
+    // Create a temporary QApplication.
+    int temp_argc = 0;
+    char** temp_argv = nullptr;
+    QApplication temp{temp_argc, temp_argv};
+
+    // Get the current screen geometry.
+    const QScreen* primary_screen = QGuiApplication::primaryScreen();
+    if (primary_screen == nullptr) {
+        return;
+    }
+
+    const QRect screen_rect = primary_screen->geometry();
+    const int real_width = screen_rect.width();
+    const int real_height = screen_rect.height();
+    const float real_ratio = primary_screen->logicalDotsPerInch() / 96.0f;
+
+    // Recommended minimum width and height for proper window fit.
+    // Any screen with a lower resolution than this will still have a scale of 1.
+    constexpr float minimum_width = 1350.0f;
+    constexpr float minimum_height = 900.0f;
+
+    const float width_ratio = std::max(1.0f, real_width / minimum_width);
+    const float height_ratio = std::max(1.0f, real_height / minimum_height);
+
+    // Get the lower of the 2 ratios and truncate, this is the maximum integer scale.
+    const float max_ratio = std::trunc(std::min(width_ratio, height_ratio));
+
+    if (max_ratio > real_ratio) {
+        QApplication::setHighDpiScaleFactorRoundingPolicy(
+            Qt::HighDpiScaleFactorRoundingPolicy::Round);
+    } else {
+        QApplication::setHighDpiScaleFactorRoundingPolicy(
+            Qt::HighDpiScaleFactorRoundingPolicy::Floor);
+    }
+#else
+    // Other OSes should be better than Windows at fractional scaling.
+    QApplication::setHighDpiScaleFactorRoundingPolicy(
+        Qt::HighDpiScaleFactorRoundingPolicy::PassThrough);
+#endif
+
+    QApplication::setAttribute(Qt::AA_EnableHighDpiScaling);
+    QApplication::setAttribute(Qt::AA_UseHighDpiPixmaps);
+}
+
 int main(int argc, char* argv[]) {
     std::unique_ptr<Config> config = std::make_unique<Config>();
     bool has_broken_vulkan = false;
@@ -4445,6 +4499,8 @@ int main(int argc, char* argv[]) {
     }
 #endif
 
+    SetHighDPIAttributes();
+
 #if QT_VERSION < QT_VERSION_CHECK(6, 0, 0)
     // Disables the "?" button on all dialogs. Disabled by default on Qt6.
     QCoreApplication::setAttribute(Qt::AA_DisableWindowContextHelpButton);
@@ -4452,6 +4508,7 @@ int main(int argc, char* argv[]) {
 
     // Enables the core to make the qt created contexts current on std::threads
     QCoreApplication::setAttribute(Qt::AA_DontCheckOpenGLContextThreadAffinity);
+
     QApplication app(argc, argv);
 
 #ifdef _WIN32

src/yuzu/util/overlay_dialog.cpp

@@ -163,7 +163,7 @@ void OverlayDialog::MoveAndResizeWindow() {
     const auto height = static_cast<float>(parentWidget()->height());
 
     // High DPI
-    const float dpi_scale = parentWidget()->windowHandle()->screen()->logicalDotsPerInch() / 96.0f;
+    const float dpi_scale = screen()->logicalDotsPerInch() / 96.0f;
 
     const auto title_text_font_size = BASE_TITLE_FONT_SIZE * (height / BASE_HEIGHT) / dpi_scale;
     const auto body_text_font_size =