summaryrefslogtreecommitdiff
path: root/src
diff options
context:
space:
mode:
Diffstat (limited to 'src')
-rw-r--r--src/CMakeLists.txt61
-rw-r--r--src/audio_core/stream.cpp2
-rw-r--r--src/common/CMakeLists.txt1
-rw-r--r--src/common/bit_field.h12
-rw-r--r--src/common/bit_util.h39
-rw-r--r--src/common/common_types.h7
-rw-r--r--src/common/multi_level_queue.h337
-rw-r--r--src/common/page_table.cpp2
-rw-r--r--src/common/page_table.h6
-rw-r--r--src/common/swap.h174
-rw-r--r--src/common/thread_queue_list.h6
-rw-r--r--src/common/uint128.cpp4
-rw-r--r--src/common/uint128.h5
-rw-r--r--src/core/CMakeLists.txt10
-rw-r--r--src/core/core.cpp11
-rw-r--r--src/core/core.h4
-rw-r--r--src/core/core_timing.cpp2
-rw-r--r--src/core/core_timing.h2
-rw-r--r--src/core/file_sys/cheat_engine.cpp492
-rw-r--r--src/core/file_sys/cheat_engine.h234
-rw-r--r--src/core/file_sys/content_archive.h15
-rw-r--r--src/core/file_sys/errors.h3
-rw-r--r--src/core/file_sys/patch_manager.cpp79
-rw-r--r--src/core/file_sys/patch_manager.h9
-rw-r--r--src/core/file_sys/registered_cache.cpp2
-rw-r--r--src/core/file_sys/system_archive/system_archive.cpp3
-rw-r--r--src/core/file_sys/system_archive/system_version.cpp52
-rw-r--r--src/core/file_sys/system_archive/system_version.h16
-rw-r--r--src/core/hle/ipc.h44
-rw-r--r--src/core/hle/kernel/code_set.cpp12
-rw-r--r--src/core/hle/kernel/code_set.h90
-rw-r--r--src/core/hle/kernel/kernel.cpp2
-rw-r--r--src/core/hle/kernel/kernel.h3
-rw-r--r--src/core/hle/kernel/mutex.cpp35
-rw-r--r--src/core/hle/kernel/mutex.h20
-rw-r--r--src/core/hle/kernel/object.cpp1
-rw-r--r--src/core/hle/kernel/object.h1
-rw-r--r--src/core/hle/kernel/process.cpp15
-rw-r--r--src/core/hle/kernel/process.h65
-rw-r--r--src/core/hle/kernel/scheduler.cpp42
-rw-r--r--src/core/hle/kernel/scheduler.h4
-rw-r--r--src/core/hle/kernel/svc.cpp146
-rw-r--r--src/core/hle/kernel/thread.cpp3
-rw-r--r--src/core/hle/kernel/transfer_memory.cpp73
-rw-r--r--src/core/hle/kernel/transfer_memory.h91
-rw-r--r--src/core/hle/kernel/vm_manager.cpp96
-rw-r--r--src/core/hle/kernel/vm_manager.h60
-rw-r--r--src/core/hle/service/am/am.cpp16
-rw-r--r--src/core/hle/service/audio/hwopus.cpp82
-rw-r--r--src/core/hle/service/hid/controllers/debug_pad.h30
-rw-r--r--src/core/hle/service/hid/controllers/npad.h102
-rw-r--r--src/core/hle/service/hid/controllers/touchscreen.h4
-rw-r--r--src/core/hle/service/hid/hid.cpp10
-rw-r--r--src/core/hle/service/hid/hid.h5
-rw-r--r--src/core/hle/service/ldr/ldr.cpp8
-rw-r--r--src/core/hle/service/lm/lm.cpp2
-rw-r--r--src/core/hle/service/nvdrv/devices/nvdevice.h10
-rw-r--r--src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp16
-rw-r--r--src/core/hle/service/nvflinger/nvflinger.cpp4
-rw-r--r--src/core/hle/service/set/set_sys.cpp79
-rw-r--r--src/core/hle/service/set/set_sys.h2
-rw-r--r--src/core/loader/elf.cpp1
-rw-r--r--src/core/loader/linker.cpp147
-rw-r--r--src/core/loader/linker.h36
-rw-r--r--src/core/loader/nro.cpp1
-rw-r--r--src/core/loader/nro.h4
-rw-r--r--src/core/loader/nso.cpp114
-rw-r--r--src/core/loader/nso.h43
-rw-r--r--src/core/memory.cpp2
-rw-r--r--src/core/memory.h3
-rw-r--r--src/input_common/sdl/sdl.h12
-rw-r--r--src/input_common/sdl/sdl_impl.cpp9
-rw-r--r--src/input_common/sdl/sdl_impl.h5
-rw-r--r--src/tests/CMakeLists.txt3
-rw-r--r--src/tests/common/bit_field.cpp90
-rw-r--r--src/tests/common/bit_utils.cpp23
-rw-r--r--src/tests/common/multi_level_queue.cpp55
-rw-r--r--src/video_core/dma_pusher.h1
-rw-r--r--src/video_core/engines/kepler_memory.cpp2
-rw-r--r--src/video_core/engines/maxwell_3d.cpp8
-rw-r--r--src/video_core/engines/maxwell_dma.cpp10
-rw-r--r--src/video_core/gpu.cpp10
-rw-r--r--src/video_core/gpu.h12
-rw-r--r--src/video_core/memory_manager.cpp482
-rw-r--r--src/video_core/memory_manager.h170
-rw-r--r--src/video_core/rasterizer_cache.h4
-rw-r--r--src/video_core/rasterizer_interface.h1
-rw-r--r--src/video_core/renderer_opengl/gl_buffer_cache.cpp4
-rw-r--r--src/video_core/renderer_opengl/gl_buffer_cache.h2
-rw-r--r--src/video_core/renderer_opengl/gl_global_cache.cpp6
-rw-r--r--src/video_core/renderer_opengl/gl_global_cache.h2
-rw-r--r--src/video_core/renderer_opengl/gl_primitive_assembler.cpp3
-rw-r--r--src/video_core/renderer_opengl/gl_primitive_assembler.h2
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer.cpp8
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer_cache.cpp43
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer_cache.h19
-rw-r--r--src/video_core/renderer_opengl/gl_shader_cache.cpp4
-rw-r--r--src/video_core/renderer_vulkan/vk_buffer_cache.cpp3
-rw-r--r--src/video_core/renderer_vulkan/vk_buffer_cache.h3
-rw-r--r--src/video_core/renderer_vulkan/vk_resource_manager.cpp2
-rw-r--r--src/video_core/renderer_vulkan/vk_resource_manager.h2
-rw-r--r--src/yuzu/bootmanager.cpp4
-rw-r--r--src/yuzu/debugger/graphics/graphics_surface.cpp2
-rw-r--r--src/yuzu/debugger/graphics/graphics_surface.h2
-rw-r--r--src/yuzu_cmd/yuzu.cpp4
105 files changed, 3196 insertions, 930 deletions
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index f69d00a2b..6c99dd5e2 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -1,18 +1,79 @@
# Enable modules to include each other's files
include_directories(.)
+# CMake seems to only define _DEBUG on Windows
+set_property(DIRECTORY APPEND PROPERTY
+ COMPILE_DEFINITIONS $<$<CONFIG:Debug>:_DEBUG> $<$<NOT:$<CONFIG:Debug>>:NDEBUG>)
+
+# Set compilation flags
+if (MSVC)
+ set(CMAKE_CONFIGURATION_TYPES Debug Release CACHE STRING "" FORCE)
+
+ # Silence "deprecation" warnings
+ add_definitions(-D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS)
+
+ # Avoid windows.h junk
+ add_definitions(-DNOMINMAX)
+
+ # Avoid windows.h from including some usually unused libs like winsocks.h, since this might cause some redefinition errors.
+ add_definitions(-DWIN32_LEAN_AND_MEAN)
+
+ # /W3 - Level 3 warnings
+ # /MP - Multi-threaded compilation
+ # /Zi - Output debugging information
+ # /Zo - enhanced debug info for optimized builds
+ # /permissive- - enables stricter C++ standards conformance checks
+ # /EHsc - C++-only exception handling semantics
+ # /Zc:throwingNew - let codegen assume `operator new` will never return null
+ # /Zc:inline - let codegen omit inline functions in object files
+ add_compile_options(/W3 /MP /Zi /Zo /permissive- /EHsc /std:c++latest /Zc:throwingNew,inline)
+
+ # /GS- - No stack buffer overflow checks
+ add_compile_options("$<$<CONFIG:Release>:/GS->")
+
+ set(CMAKE_EXE_LINKER_FLAGS_DEBUG "/DEBUG /MANIFEST:NO" CACHE STRING "" FORCE)
+ set(CMAKE_EXE_LINKER_FLAGS_RELEASE "/DEBUG /MANIFEST:NO /INCREMENTAL:NO /OPT:REF,ICF" CACHE STRING "" FORCE)
+else()
+ add_compile_options("-Wno-attributes")
+
+ if (APPLE AND CMAKE_CXX_COMPILER_ID STREQUAL Clang)
+ add_compile_options("-stdlib=libc++")
+ endif()
+
+ # Set file offset size to 64 bits.
+ #
+ # On modern Unixes, this is typically already the case. The lone exception is
+ # glibc, which may default to 32 bits. glibc allows this to be configured
+ # by setting _FILE_OFFSET_BITS.
+ if(CMAKE_SYSTEM_NAME STREQUAL "Linux" OR MINGW)
+ add_definitions(-D_FILE_OFFSET_BITS=64)
+ endif()
+
+ if (MINGW)
+ add_definitions(-DMINGW_HAS_SECURE_API)
+
+ if (MINGW_STATIC_BUILD)
+ add_definitions(-DQT_STATICPLUGIN)
+ add_compile_options("-static")
+ endif()
+ endif()
+endif()
+
add_subdirectory(common)
add_subdirectory(core)
add_subdirectory(audio_core)
add_subdirectory(video_core)
add_subdirectory(input_common)
add_subdirectory(tests)
+
if (ENABLE_SDL2)
add_subdirectory(yuzu_cmd)
endif()
+
if (ENABLE_QT)
add_subdirectory(yuzu)
endif()
+
if (ENABLE_WEB_SERVICE)
add_subdirectory(web_service)
endif()
diff --git a/src/audio_core/stream.cpp b/src/audio_core/stream.cpp
index 4b66a6786..22a3f8c84 100644
--- a/src/audio_core/stream.cpp
+++ b/src/audio_core/stream.cpp
@@ -38,7 +38,7 @@ Stream::Stream(Core::Timing::CoreTiming& core_timing, u32 sample_rate, Format fo
sink_stream{sink_stream}, core_timing{core_timing}, name{std::move(name_)} {
release_event = core_timing.RegisterEvent(
- name, [this](u64 userdata, int cycles_late) { ReleaseActiveBuffer(); });
+ name, [this](u64 userdata, s64 cycles_late) { ReleaseActiveBuffer(); });
}
void Stream::Play() {
diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt
index 43ae8a9e7..850ce8006 100644
--- a/src/common/CMakeLists.txt
+++ b/src/common/CMakeLists.txt
@@ -98,6 +98,7 @@ add_library(common STATIC
microprofile.h
microprofileui.h
misc.cpp
+ multi_level_queue.h
page_table.cpp
page_table.h
param_package.cpp
diff --git a/src/common/bit_field.h b/src/common/bit_field.h
index 7433c39ba..902e668e3 100644
--- a/src/common/bit_field.h
+++ b/src/common/bit_field.h
@@ -34,6 +34,7 @@
#include <limits>
#include <type_traits>
#include "common/common_funcs.h"
+#include "common/swap.h"
/*
* Abstract bitfield class
@@ -108,7 +109,7 @@
* symptoms.
*/
#pragma pack(1)
-template <std::size_t Position, std::size_t Bits, typename T>
+template <std::size_t Position, std::size_t Bits, typename T, typename EndianTag = LETag>
struct BitField {
private:
// UnderlyingType is T for non-enum types and the underlying type of T if
@@ -121,6 +122,8 @@ private:
// We store the value as the unsigned type to avoid undefined behaviour on value shifting
using StorageType = std::make_unsigned_t<UnderlyingType>;
+ using StorageTypeWithEndian = typename AddEndian<StorageType, EndianTag>::type;
+
public:
/// Constants to allow limited introspection of fields if needed
static constexpr std::size_t position = Position;
@@ -170,7 +173,7 @@ public:
}
constexpr FORCE_INLINE void Assign(const T& value) {
- storage = (storage & ~mask) | FormatValue(value);
+ storage = (static_cast<StorageType>(storage) & ~mask) | FormatValue(value);
}
constexpr T Value() const {
@@ -182,7 +185,7 @@ public:
}
private:
- StorageType storage;
+ StorageTypeWithEndian storage;
static_assert(bits + position <= 8 * sizeof(T), "Bitfield out of range");
@@ -193,3 +196,6 @@ private:
static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable in a BitField");
};
#pragma pack()
+
+template <std::size_t Position, std::size_t Bits, typename T>
+using BitFieldBE = BitField<Position, Bits, T, BETag>;
diff --git a/src/common/bit_util.h b/src/common/bit_util.h
index 1eea17ba1..a4f9ed4aa 100644
--- a/src/common/bit_util.h
+++ b/src/common/bit_util.h
@@ -58,4 +58,43 @@ inline u64 CountLeadingZeroes64(u64 value) {
return __builtin_clzll(value);
}
#endif
+
+#ifdef _MSC_VER
+inline u32 CountTrailingZeroes32(u32 value) {
+ unsigned long trailing_zero = 0;
+
+ if (_BitScanForward(&trailing_zero, value) != 0) {
+ return trailing_zero;
+ }
+
+ return 32;
+}
+
+inline u64 CountTrailingZeroes64(u64 value) {
+ unsigned long trailing_zero = 0;
+
+ if (_BitScanForward64(&trailing_zero, value) != 0) {
+ return trailing_zero;
+ }
+
+ return 64;
+}
+#else
+inline u32 CountTrailingZeroes32(u32 value) {
+ if (value == 0) {
+ return 32;
+ }
+
+ return __builtin_ctz(value);
+}
+
+inline u64 CountTrailingZeroes64(u64 value) {
+ if (value == 0) {
+ return 64;
+ }
+
+ return __builtin_ctzll(value);
+}
+#endif
+
} // namespace Common
diff --git a/src/common/common_types.h b/src/common/common_types.h
index 6b1766dca..4cec89fbd 100644
--- a/src/common/common_types.h
+++ b/src/common/common_types.h
@@ -40,10 +40,9 @@ using s64 = std::int64_t; ///< 64-bit signed int
using f32 = float; ///< 32-bit floating point
using f64 = double; ///< 64-bit floating point
-// TODO: It would be nice to eventually replace these with strong types that prevent accidental
-// conversion between each other.
-using VAddr = u64; ///< Represents a pointer in the userspace virtual address space.
-using PAddr = u64; ///< Represents a pointer in the ARM11 physical address space.
+using VAddr = u64; ///< Represents a pointer in the userspace virtual address space.
+using PAddr = u64; ///< Represents a pointer in the ARM11 physical address space.
+using GPUVAddr = u64; ///< Represents a pointer in the GPU virtual address space.
using u128 = std::array<std::uint64_t, 2>;
static_assert(sizeof(u128) == 16, "u128 must be 128 bits wide");
diff --git a/src/common/multi_level_queue.h b/src/common/multi_level_queue.h
new file mode 100644
index 000000000..2b61b91e0
--- /dev/null
+++ b/src/common/multi_level_queue.h
@@ -0,0 +1,337 @@
+// Copyright 2019 TuxSH
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <iterator>
+#include <list>
+#include <utility>
+
+#include "common/bit_util.h"
+#include "common/common_types.h"
+
+namespace Common {
+
+/**
+ * A MultiLevelQueue is a type of priority queue which has the following characteristics:
+ * - iteratable through each of its elements.
+ * - back can be obtained.
+ * - O(1) add, lookup (both front and back)
+ * - discrete priorities and a max of 64 priorities (limited domain)
+ * This type of priority queue is normaly used for managing threads within an scheduler
+ */
+template <typename T, std::size_t Depth>
+class MultiLevelQueue {
+public:
+ using value_type = T;
+ using reference = value_type&;
+ using const_reference = const value_type&;
+ using pointer = value_type*;
+ using const_pointer = const value_type*;
+
+ using difference_type = typename std::pointer_traits<pointer>::difference_type;
+ using size_type = std::size_t;
+
+ template <bool is_constant>
+ class iterator_impl {
+ public:
+ using iterator_category = std::bidirectional_iterator_tag;
+ using value_type = T;
+ using pointer = std::conditional_t<is_constant, T*, const T*>;
+ using reference = std::conditional_t<is_constant, const T&, T&>;
+ using difference_type = typename std::pointer_traits<pointer>::difference_type;
+
+ friend bool operator==(const iterator_impl& lhs, const iterator_impl& rhs) {
+ if (lhs.IsEnd() && rhs.IsEnd())
+ return true;
+ return std::tie(lhs.current_priority, lhs.it) == std::tie(rhs.current_priority, rhs.it);
+ }
+
+ friend bool operator!=(const iterator_impl& lhs, const iterator_impl& rhs) {
+ return !operator==(lhs, rhs);
+ }
+
+ reference operator*() const {
+ return *it;
+ }
+
+ pointer operator->() const {
+ return it.operator->();
+ }
+
+ iterator_impl& operator++() {
+ if (IsEnd()) {
+ return *this;
+ }
+
+ ++it;
+
+ if (it == GetEndItForPrio()) {
+ u64 prios = mlq.used_priorities;
+ prios &= ~((1ULL << (current_priority + 1)) - 1);
+ if (prios == 0) {
+ current_priority = mlq.depth();
+ } else {
+ current_priority = CountTrailingZeroes64(prios);
+ it = GetBeginItForPrio();
+ }
+ }
+ return *this;
+ }
+
+ iterator_impl& operator--() {
+ if (IsEnd()) {
+ if (mlq.used_priorities != 0) {
+ current_priority = 63 - CountLeadingZeroes64(mlq.used_priorities);
+ it = GetEndItForPrio();
+ --it;
+ }
+ } else if (it == GetBeginItForPrio()) {
+ u64 prios = mlq.used_priorities;
+ prios &= (1ULL << current_priority) - 1;
+ if (prios != 0) {
+ current_priority = CountTrailingZeroes64(prios);
+ it = GetEndItForPrio();
+ --it;
+ }
+ } else {
+ --it;
+ }
+ return *this;
+ }
+
+ iterator_impl operator++(int) {
+ const iterator_impl v{*this};
+ ++(*this);
+ return v;
+ }
+
+ iterator_impl operator--(int) {
+ const iterator_impl v{*this};
+ --(*this);
+ return v;
+ }
+
+ // allow implicit const->non-const
+ iterator_impl(const iterator_impl<false>& other)
+ : mlq(other.mlq), it(other.it), current_priority(other.current_priority) {}
+
+ iterator_impl(const iterator_impl<true>& other)
+ : mlq(other.mlq), it(other.it), current_priority(other.current_priority) {}
+
+ iterator_impl& operator=(const iterator_impl<false>& other) {
+ mlq = other.mlq;
+ it = other.it;
+ current_priority = other.current_priority;
+ return *this;
+ }
+
+ friend class iterator_impl<true>;
+ iterator_impl() = default;
+
+ private:
+ friend class MultiLevelQueue;
+ using container_ref =
+ std::conditional_t<is_constant, const MultiLevelQueue&, MultiLevelQueue&>;
+ using list_iterator = std::conditional_t<is_constant, typename std::list<T>::const_iterator,
+ typename std::list<T>::iterator>;
+
+ explicit iterator_impl(container_ref mlq, list_iterator it, u32 current_priority)
+ : mlq(mlq), it(it), current_priority(current_priority) {}
+ explicit iterator_impl(container_ref mlq, u32 current_priority)
+ : mlq(mlq), it(), current_priority(current_priority) {}
+
+ bool IsEnd() const {
+ return current_priority == mlq.depth();
+ }
+
+ list_iterator GetBeginItForPrio() const {
+ return mlq.levels[current_priority].begin();
+ }
+
+ list_iterator GetEndItForPrio() const {
+ return mlq.levels[current_priority].end();
+ }
+
+ container_ref mlq;
+ list_iterator it;
+ u32 current_priority;
+ };
+
+ using iterator = iterator_impl<false>;
+ using const_iterator = iterator_impl<true>;
+
+ void add(const T& element, u32 priority, bool send_back = true) {
+ if (send_back)
+ levels[priority].push_back(element);
+ else
+ levels[priority].push_front(element);
+ used_priorities |= 1ULL << priority;
+ }
+
+ void remove(const T& element, u32 priority) {
+ auto it = ListIterateTo(levels[priority], element);
+ if (it == levels[priority].end())
+ return;
+ levels[priority].erase(it);
+ if (levels[priority].empty()) {
+ used_priorities &= ~(1ULL << priority);
+ }
+ }
+
+ void adjust(const T& element, u32 old_priority, u32 new_priority, bool adjust_front = false) {
+ remove(element, old_priority);
+ add(element, new_priority, !adjust_front);
+ }
+ void adjust(const_iterator it, u32 old_priority, u32 new_priority, bool adjust_front = false) {
+ adjust(*it, old_priority, new_priority, adjust_front);
+ }
+
+ void transfer_to_front(const T& element, u32 priority, MultiLevelQueue& other) {
+ ListSplice(other.levels[priority], other.levels[priority].begin(), levels[priority],
+ ListIterateTo(levels[priority], element));
+
+ other.used_priorities |= 1ULL << priority;
+
+ if (levels[priority].empty()) {
+ used_priorities &= ~(1ULL << priority);
+ }
+ }
+
+ void transfer_to_front(const_iterator it, u32 priority, MultiLevelQueue& other) {
+ transfer_to_front(*it, priority, other);
+ }
+
+ void transfer_to_back(const T& element, u32 priority, MultiLevelQueue& other) {
+ ListSplice(other.levels[priority], other.levels[priority].end(), levels[priority],
+ ListIterateTo(levels[priority], element));
+
+ other.used_priorities |= 1ULL << priority;
+
+ if (levels[priority].empty()) {
+ used_priorities &= ~(1ULL << priority);
+ }
+ }
+
+ void transfer_to_back(const_iterator it, u32 priority, MultiLevelQueue& other) {
+ transfer_to_back(*it, priority, other);
+ }
+
+ void yield(u32 priority, std::size_t n = 1) {
+ ListShiftForward(levels[priority], n);
+ }
+
+ std::size_t depth() const {
+ return Depth;
+ }
+
+ std::size_t size(u32 priority) const {
+ return levels[priority].size();
+ }
+
+ std::size_t size() const {
+ u64 priorities = used_priorities;
+ std::size_t size = 0;
+ while (priorities != 0) {
+ const u64 current_priority = CountTrailingZeroes64(priorities);
+ size += levels[current_priority].size();
+ priorities &= ~(1ULL << current_priority);
+ }
+ return size;
+ }
+
+ bool empty() const {
+ return used_priorities == 0;
+ }
+
+ bool empty(u32 priority) const {
+ return (used_priorities & (1ULL << priority)) == 0;
+ }
+
+ u32 highest_priority_set(u32 max_priority = 0) const {
+ const u64 priorities =
+ max_priority == 0 ? used_priorities : (used_priorities & ~((1ULL << max_priority) - 1));
+ return priorities == 0 ? Depth : static_cast<u32>(CountTrailingZeroes64(priorities));
+ }
+
+ u32 lowest_priority_set(u32 min_priority = Depth - 1) const {
+ const u64 priorities = min_priority >= Depth - 1
+ ? used_priorities
+ : (used_priorities & ((1ULL << (min_priority + 1)) - 1));
+ return priorities == 0 ? Depth : 63 - CountLeadingZeroes64(priorities);
+ }
+
+ const_iterator cbegin(u32 max_prio = 0) const {
+ const u32 priority = highest_priority_set(max_prio);
+ return priority == Depth ? cend()
+ : const_iterator{*this, levels[priority].cbegin(), priority};
+ }
+ const_iterator begin(u32 max_prio = 0) const {
+ return cbegin(max_prio);
+ }
+ iterator begin(u32 max_prio = 0) {
+ const u32 priority = highest_priority_set(max_prio);
+ return priority == Depth ? end() : iterator{*this, levels[priority].begin(), priority};
+ }
+
+ const_iterator cend(u32 min_prio = Depth - 1) const {
+ return min_prio == Depth - 1 ? const_iterator{*this, Depth} : cbegin(min_prio + 1);
+ }
+ const_iterator end(u32 min_prio = Depth - 1) const {
+ return cend(min_prio);
+ }
+ iterator end(u32 min_prio = Depth - 1) {
+ return min_prio == Depth - 1 ? iterator{*this, Depth} : begin(min_prio + 1);
+ }
+
+ T& front(u32 max_priority = 0) {
+ const u32 priority = highest_priority_set(max_priority);
+ return levels[priority == Depth ? 0 : priority].front();
+ }
+ const T& front(u32 max_priority = 0) const {
+ const u32 priority = highest_priority_set(max_priority);
+ return levels[priority == Depth ? 0 : priority].front();
+ }
+
+ T back(u32 min_priority = Depth - 1) {
+ const u32 priority = lowest_priority_set(min_priority); // intended
+ return levels[priority == Depth ? 63 : priority].back();
+ }
+ const T& back(u32 min_priority = Depth - 1) const {
+ const u32 priority = lowest_priority_set(min_priority); // intended
+ return levels[priority == Depth ? 63 : priority].back();
+ }
+
+private:
+ using const_list_iterator = typename std::list<T>::const_iterator;
+
+ static void ListShiftForward(std::list<T>& list, const std::size_t shift = 1) {
+ if (shift >= list.size()) {
+ return;
+ }
+
+ const auto begin_range = list.begin();
+ const auto end_range = std::next(begin_range, shift);
+ list.splice(list.end(), list, begin_range, end_range);
+ }
+
+ static void ListSplice(std::list<T>& in_list, const_list_iterator position,
+ std::list<T>& out_list, const_list_iterator element) {
+ in_list.splice(position, out_list, element);
+ }
+
+ static const_list_iterator ListIterateTo(const std::list<T>& list, const T& element) {
+ auto it = list.cbegin();
+ while (it != list.cend() && *it != element) {
+ ++it;
+ }
+ return it;
+ }
+
+ std::array<std::list<T>, Depth> levels;
+ u64 used_priorities = 0;
+};
+
+} // namespace Common
diff --git a/src/common/page_table.cpp b/src/common/page_table.cpp
index 8eba1c3f1..69b7abc54 100644
--- a/src/common/page_table.cpp
+++ b/src/common/page_table.cpp
@@ -16,6 +16,7 @@ void PageTable::Resize(std::size_t address_space_width_in_bits) {
pointers.resize(num_page_table_entries);
attributes.resize(num_page_table_entries);
+ backing_addr.resize(num_page_table_entries);
// The default is a 39-bit address space, which causes an initial 1GB allocation size. If the
// vector size is subsequently decreased (via resize), the vector might not automatically
@@ -24,6 +25,7 @@ void PageTable::Resize(std::size_t address_space_width_in_bits) {
pointers.shrink_to_fit();
attributes.shrink_to_fit();
+ backing_addr.shrink_to_fit();
}
} // namespace Common
diff --git a/src/common/page_table.h b/src/common/page_table.h
index 8339f2890..8b8ff0bb8 100644
--- a/src/common/page_table.h
+++ b/src/common/page_table.h
@@ -21,6 +21,8 @@ enum class PageType : u8 {
RasterizerCachedMemory,
/// Page is mapped to a I/O region. Writing and reading to this page is handled by functions.
Special,
+ /// Page is allocated for use.
+ Allocated,
};
struct SpecialRegion {
@@ -66,7 +68,7 @@ struct PageTable {
* Contains MMIO handlers that back memory regions whose entries in the `attribute` vector is
* of type `Special`.
*/
- boost::icl::interval_map<VAddr, std::set<SpecialRegion>> special_regions;
+ boost::icl::interval_map<u64, std::set<SpecialRegion>> special_regions;
/**
* Vector of fine grained page attributes. If it is set to any value other than `Memory`, then
@@ -74,6 +76,8 @@ struct PageTable {
*/
std::vector<PageType> attributes;
+ std::vector<u64> backing_addr;
+
const std::size_t page_size_in_bits{};
};
diff --git a/src/common/swap.h b/src/common/swap.h
index 0e219747f..b3eab1324 100644
--- a/src/common/swap.h
+++ b/src/common/swap.h
@@ -17,6 +17,8 @@
#pragma once
+#include <type_traits>
+
#if defined(_MSC_VER)
#include <cstdlib>
#elif defined(__linux__)
@@ -170,7 +172,7 @@ struct swap_struct_t {
using swapped_t = swap_struct_t;
protected:
- T value = T();
+ T value;
static T swap(T v) {
return F::swap(v);
@@ -605,52 +607,154 @@ struct swap_double_t {
}
};
-#if COMMON_LITTLE_ENDIAN
-using u16_le = u16;
-using u32_le = u32;
-using u64_le = u64;
+template <typename T>
+struct swap_enum_t {
+ static_assert(std::is_enum_v<T>);
+ using base = std::underlying_type_t<T>;
+
+public:
+ swap_enum_t() = default;
+ swap_enum_t(const T& v) : value(swap(v)) {}
+
+ swap_enum_t& operator=(const T& v) {
+ value = swap(v);
+ return *this;
+ }
+
+ operator T() const {
+ return swap(value);
+ }
+
+ explicit operator base() const {
+ return static_cast<base>(swap(value));
+ }
-using s16_le = s16;
-using s32_le = s32;
-using s64_le = s64;
+protected:
+ T value{};
+ // clang-format off
+ using swap_t = std::conditional_t<
+ std::is_same_v<base, u16>, swap_16_t<u16>, std::conditional_t<
+ std::is_same_v<base, s16>, swap_16_t<s16>, std::conditional_t<
+ std::is_same_v<base, u32>, swap_32_t<u32>, std::conditional_t<
+ std::is_same_v<base, s32>, swap_32_t<s32>, std::conditional_t<
+ std::is_same_v<base, u64>, swap_64_t<u64>, std::conditional_t<
+ std::is_same_v<base, s64>, swap_64_t<s64>, void>>>>>>;
+ // clang-format on
+ static T swap(T x) {
+ return static_cast<T>(swap_t::swap(static_cast<base>(x)));
+ }
+};
-using float_le = float;
-using double_le = double;
+struct SwapTag {}; // Use the different endianness from the system
+struct KeepTag {}; // Use the same endianness as the system
-using u64_be = swap_struct_t<u64, swap_64_t<u64>>;
-using s64_be = swap_struct_t<s64, swap_64_t<s64>>;
+template <typename T, typename Tag>
+struct AddEndian;
-using u32_be = swap_struct_t<u32, swap_32_t<u32>>;
-using s32_be = swap_struct_t<s32, swap_32_t<s32>>;
+// KeepTag specializations
-using u16_be = swap_struct_t<u16, swap_16_t<u16>>;
-using s16_be = swap_struct_t<s16, swap_16_t<s16>>;
+template <typename T>
+struct AddEndian<T, KeepTag> {
+ using type = T;
+};
-using float_be = swap_struct_t<float, swap_float_t<float>>;
-using double_be = swap_struct_t<double, swap_double_t<double>>;
-#else
+// SwapTag specializations
+
+template <>
+struct AddEndian<u8, SwapTag> {
+ using type = u8;
+};
+
+template <>
+struct AddEndian<u16, SwapTag> {
+ using type = swap_struct_t<u16, swap_16_t<u16>>;
+};
+
+template <>
+struct AddEndian<u32, SwapTag> {
+ using type = swap_struct_t<u32, swap_32_t<u32>>;
+};
-using u64_le = swap_struct_t<u64, swap_64_t<u64>>;
-using s64_le = swap_struct_t<s64, swap_64_t<s64>>;
+template <>
+struct AddEndian<u64, SwapTag> {
+ using type = swap_struct_t<u64, swap_64_t<u64>>;
+};
+
+template <>
+struct AddEndian<s8, SwapTag> {
+ using type = s8;
+};
-using u32_le = swap_struct_t<u32, swap_32_t<u32>>;
-using s32_le = swap_struct_t<s32, swap_32_t<s32>>;
+template <>
+struct AddEndian<s16, SwapTag> {
+ using type = swap_struct_t<s16, swap_16_t<s16>>;
+};
-using u16_le = swap_struct_t<u16, swap_16_t<u16>>;
-using s16_le = swap_struct_t<s16, swap_16_t<s16>>;
+template <>
+struct AddEndian<s32, SwapTag> {
+ using type = swap_struct_t<s32, swap_32_t<s32>>;
+};
+
+template <>
+struct AddEndian<s64, SwapTag> {
+ using type = swap_struct_t<s64, swap_64_t<s64>>;
+};
+
+template <>
+struct AddEndian<float, SwapTag> {
+ using type = swap_struct_t<float, swap_float_t<float>>;
+};
+
+template <>
+struct AddEndian<double, SwapTag> {
+ using type = swap_struct_t<double, swap_double_t<double>>;
+};
+
+template <typename T>
+struct AddEndian<T, SwapTag> {
+ static_assert(std::is_enum_v<T>);
+ using type = swap_enum_t<T>;
+};
-using float_le = swap_struct_t<float, swap_float_t<float>>;
-using double_le = swap_struct_t<double, swap_double_t<double>>;
+// Alias LETag/BETag as KeepTag/SwapTag depending on the system
+#if COMMON_LITTLE_ENDIAN
-using u16_be = u16;
-using u32_be = u32;
-using u64_be = u64;
+using LETag = KeepTag;
+using BETag = SwapTag;
-using s16_be = s16;
-using s32_be = s32;
-using s64_be = s64;
+#else
-using float_be = float;
-using double_be = double;
+using BETag = KeepTag;
+using LETag = SwapTag;
#endif
+
+// Aliases for LE types
+using u16_le = AddEndian<u16, LETag>::type;
+using u32_le = AddEndian<u32, LETag>::type;
+using u64_le = AddEndian<u64, LETag>::type;
+
+using s16_le = AddEndian<s16, LETag>::type;
+using s32_le = AddEndian<s32, LETag>::type;
+using s64_le = AddEndian<s64, LETag>::type;
+
+template <typename T>
+using enum_le = std::enable_if_t<std::is_enum_v<T>, typename AddEndian<T, LETag>::type>;
+
+using float_le = AddEndian<float, LETag>::type;
+using double_le = AddEndian<double, LETag>::type;
+
+// Aliases for BE types
+using u16_be = AddEndian<u16, BETag>::type;
+using u32_be = AddEndian<u32, BETag>::type;
+using u64_be = AddEndian<u64, BETag>::type;
+
+using s16_be = AddEndian<s16, BETag>::type;
+using s32_be = AddEndian<s32, BETag>::type;
+using s64_be = AddEndian<s64, BETag>::type;
+
+template <typename T>
+using enum_be = std::enable_if_t<std::is_enum_v<T>, typename AddEndian<T, BETag>::type>;
+
+using float_be = AddEndian<float, BETag>::type;
+using double_be = AddEndian<double, BETag>::type;
diff --git a/src/common/thread_queue_list.h b/src/common/thread_queue_list.h
index e7594db68..791f99a8c 100644
--- a/src/common/thread_queue_list.h
+++ b/src/common/thread_queue_list.h
@@ -6,7 +6,6 @@
#include <array>
#include <deque>
-#include <boost/range/algorithm_ext/erase.hpp>
namespace Common {
@@ -111,8 +110,9 @@ struct ThreadQueueList {
}
void remove(Priority priority, const T& thread_id) {
- Queue* cur = &queues[priority];
- boost::remove_erase(cur->data, thread_id);
+ Queue* const cur = &queues[priority];
+ const auto iter = std::remove(cur->data.begin(), cur->data.end(), thread_id);
+ cur->data.erase(iter, cur->data.end());
}
void rotate(Priority priority) {
diff --git a/src/common/uint128.cpp b/src/common/uint128.cpp
index 2238a52c5..32bf56730 100644
--- a/src/common/uint128.cpp
+++ b/src/common/uint128.cpp
@@ -1,3 +1,7 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
#ifdef _MSC_VER
#include <intrin.h>
diff --git a/src/common/uint128.h b/src/common/uint128.h
index 52e6b46eb..a3be2a2cb 100644
--- a/src/common/uint128.h
+++ b/src/common/uint128.h
@@ -1,3 +1,8 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
#include <utility>
#include "common/common_types.h"
diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt
index aee8bc27d..9e23afe85 100644
--- a/src/core/CMakeLists.txt
+++ b/src/core/CMakeLists.txt
@@ -31,6 +31,8 @@ add_library(core STATIC
file_sys/bis_factory.h
file_sys/card_image.cpp
file_sys/card_image.h
+ file_sys/cheat_engine.cpp
+ file_sys/cheat_engine.h
file_sys/content_archive.cpp
file_sys/content_archive.h
file_sys/control_metadata.cpp
@@ -68,6 +70,8 @@ add_library(core STATIC
file_sys/system_archive/ng_word.h
file_sys/system_archive/system_archive.cpp
file_sys/system_archive/system_archive.h
+ file_sys/system_archive/system_version.cpp
+ file_sys/system_archive/system_version.h
file_sys/vfs.cpp
file_sys/vfs.h
file_sys/vfs_concat.cpp
@@ -107,6 +111,8 @@ add_library(core STATIC
hle/kernel/client_port.h
hle/kernel/client_session.cpp
hle/kernel/client_session.h
+ hle/kernel/code_set.cpp
+ hle/kernel/code_set.h
hle/kernel/errors.h
hle/kernel/handle_table.cpp
hle/kernel/handle_table.h
@@ -140,6 +146,8 @@ add_library(core STATIC
hle/kernel/svc_wrap.h
hle/kernel/thread.cpp
hle/kernel/thread.h
+ hle/kernel/transfer_memory.cpp
+ hle/kernel/transfer_memory.h
hle/kernel/vm_manager.cpp
hle/kernel/vm_manager.h
hle/kernel/wait_object.cpp
@@ -419,8 +427,6 @@ add_library(core STATIC
loader/deconstructed_rom_directory.h
loader/elf.cpp
loader/elf.h
- loader/linker.cpp
- loader/linker.h
loader/loader.cpp
loader/loader.h
loader/nax.cpp
diff --git a/src/core/core.cpp b/src/core/core.cpp
index 89b3fb418..4fe77c25b 100644
--- a/src/core/core.cpp
+++ b/src/core/core.cpp
@@ -32,6 +32,7 @@
#include "core/perf_stats.h"
#include "core/settings.h"
#include "core/telemetry_session.h"
+#include "file_sys/cheat_engine.h"
#include "frontend/applets/profile_select.h"
#include "frontend/applets/software_keyboard.h"
#include "frontend/applets/web_browser.h"
@@ -205,6 +206,7 @@ struct System::Impl {
GDBStub::Shutdown();
Service::Shutdown();
service_manager.reset();
+ cheat_engine.reset();
telemetry_session.reset();
gpu_core.reset();
@@ -255,6 +257,8 @@ struct System::Impl {
CpuCoreManager cpu_core_manager;
bool is_powered_on = false;
+ std::unique_ptr<FileSys::CheatEngine> cheat_engine;
+
/// Frontend applets
std::unique_ptr<Core::Frontend::ProfileSelectApplet> profile_selector;
std::unique_ptr<Core::Frontend::SoftwareKeyboardApplet> software_keyboard;
@@ -453,6 +457,13 @@ Tegra::DebugContext* System::GetGPUDebugContext() const {
return impl->debug_context.get();
}
+void System::RegisterCheatList(const std::vector<FileSys::CheatList>& list,
+ const std::string& build_id, VAddr code_region_start,
+ VAddr code_region_end) {
+ impl->cheat_engine = std::make_unique<FileSys::CheatEngine>(*this, list, build_id,
+ code_region_start, code_region_end);
+}
+
void System::SetFilesystem(std::shared_ptr<FileSys::VfsFilesystem> vfs) {
impl->virtual_filesystem = std::move(vfs);
}
diff --git a/src/core/core.h b/src/core/core.h
index ba76a41d8..4d83b93cc 100644
--- a/src/core/core.h
+++ b/src/core/core.h
@@ -20,6 +20,7 @@ class WebBrowserApplet;
} // namespace Core::Frontend
namespace FileSys {
+class CheatList;
class VfsFilesystem;
} // namespace FileSys
@@ -253,6 +254,9 @@ public:
std::shared_ptr<FileSys::VfsFilesystem> GetFilesystem() const;
+ void RegisterCheatList(const std::vector<FileSys::CheatList>& list, const std::string& build_id,
+ VAddr code_region_start, VAddr code_region_end);
+
void SetProfileSelector(std::unique_ptr<Frontend::ProfileSelectApplet> applet);
const Frontend::ProfileSelectApplet& GetProfileSelector() const;
diff --git a/src/core/core_timing.cpp b/src/core/core_timing.cpp
index a0dd5db24..41adb2302 100644
--- a/src/core/core_timing.cpp
+++ b/src/core/core_timing.cpp
@@ -186,7 +186,7 @@ void CoreTiming::Advance() {
Event evt = std::move(event_queue.front());
std::pop_heap(event_queue.begin(), event_queue.end(), std::greater<>());
event_queue.pop_back();
- evt.type->callback(evt.userdata, static_cast<int>(global_timer - evt.time));
+ evt.type->callback(evt.userdata, global_timer - evt.time);
}
is_global_timer_sane = false;
diff --git a/src/core/core_timing.h b/src/core/core_timing.h
index 59163bae1..9d2efde37 100644
--- a/src/core/core_timing.h
+++ b/src/core/core_timing.h
@@ -15,7 +15,7 @@
namespace Core::Timing {
/// A callback that may be scheduled for a particular core timing event.
-using TimedCallback = std::function<void(u64 userdata, int cycles_late)>;
+using TimedCallback = std::function<void(u64 userdata, s64 cycles_late)>;
/// Contains the characteristics of a particular event.
struct EventType {
diff --git a/src/core/file_sys/cheat_engine.cpp b/src/core/file_sys/cheat_engine.cpp
new file mode 100644
index 000000000..b06c2f20a
--- /dev/null
+++ b/src/core/file_sys/cheat_engine.cpp
@@ -0,0 +1,492 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <locale>
+#include "common/hex_util.h"
+#include "common/microprofile.h"
+#include "common/swap.h"
+#include "core/core.h"
+#include "core/core_timing.h"
+#include "core/core_timing_util.h"
+#include "core/file_sys/cheat_engine.h"
+#include "core/hle/kernel/process.h"
+#include "core/hle/service/hid/controllers/npad.h"
+#include "core/hle/service/hid/hid.h"
+#include "core/hle/service/sm/sm.h"
+
+namespace FileSys {
+
+constexpr s64 CHEAT_ENGINE_TICKS = static_cast<s64>(Core::Timing::BASE_CLOCK_RATE / 60);
+constexpr u32 KEYPAD_BITMASK = 0x3FFFFFF;
+
+u64 Cheat::Address() const {
+ u64 out;
+ std::memcpy(&out, raw.data(), sizeof(u64));
+ return Common::swap64(out) & 0xFFFFFFFFFF;
+}
+
+u64 Cheat::ValueWidth(u64 offset) const {
+ return Value(offset, width);
+}
+
+u64 Cheat::Value(u64 offset, u64 width) const {
+ u64 out;
+ std::memcpy(&out, raw.data() + offset, sizeof(u64));
+ out = Common::swap64(out);
+ if (width == 8)
+ return out;
+ return out & ((1ull << (width * CHAR_BIT)) - 1);
+}
+
+u32 Cheat::KeypadValue() const {
+ u32 out;
+ std::memcpy(&out, raw.data(), sizeof(u32));
+ return Common::swap32(out) & 0x0FFFFFFF;
+}
+
+void CheatList::SetMemoryParameters(VAddr main_begin, VAddr heap_begin, VAddr main_end,
+ VAddr heap_end, MemoryWriter writer, MemoryReader reader) {
+ this->main_region_begin = main_begin;
+ this->main_region_end = main_end;
+ this->heap_region_begin = heap_begin;
+ this->heap_region_end = heap_end;
+ this->writer = writer;
+ this->reader = reader;
+}
+
+MICROPROFILE_DEFINE(Cheat_Engine, "Add-Ons", "Cheat Engine", MP_RGB(70, 200, 70));
+
+void CheatList::Execute() {
+ MICROPROFILE_SCOPE(Cheat_Engine);
+
+ std::fill(scratch.begin(), scratch.end(), 0);
+ in_standard = false;
+ for (std::size_t i = 0; i < master_list.size(); ++i) {
+ LOG_DEBUG(Common_Filesystem, "Executing block #{:08X} ({})", i, master_list[i].first);
+ current_block = i;
+ ExecuteBlock(master_list[i].second);
+ }
+
+ in_standard = true;
+ for (std::size_t i = 0; i < standard_list.size(); ++i) {
+ LOG_DEBUG(Common_Filesystem, "Executing block #{:08X} ({})", i, standard_list[i].first);
+ current_block = i;
+ ExecuteBlock(standard_list[i].second);
+ }
+}
+
+CheatList::CheatList(const Core::System& system_, ProgramSegment master, ProgramSegment standard)
+ : master_list{std::move(master)}, standard_list{std::move(standard)}, system{&system_} {}
+
+bool CheatList::EvaluateConditional(const Cheat& cheat) const {
+ using ComparisonFunction = bool (*)(u64, u64);
+ constexpr std::array<ComparisonFunction, 6> comparison_functions{
+ [](u64 a, u64 b) { return a > b; }, [](u64 a, u64 b) { return a >= b; },
+ [](u64 a, u64 b) { return a < b; }, [](u64 a, u64 b) { return a <= b; },
+ [](u64 a, u64 b) { return a == b; }, [](u64 a, u64 b) { return a != b; },
+ };
+
+ if (cheat.type == CodeType::ConditionalInput) {
+ const auto applet_resource =
+ system->ServiceManager().GetService<Service::HID::Hid>("hid")->GetAppletResource();
+ if (applet_resource == nullptr) {
+ LOG_WARNING(
+ Common_Filesystem,
+ "Attempted to evaluate input conditional, but applet resource is not initialized!");
+ return false;
+ }
+
+ const auto press_state =
+ applet_resource
+ ->GetController<Service::HID::Controller_NPad>(Service::HID::HidController::NPad)
+ .GetAndResetPressState();
+ return ((press_state & cheat.KeypadValue()) & KEYPAD_BITMASK) != 0;
+ }
+
+ ASSERT(cheat.type == CodeType::Conditional);
+
+ const auto offset =
+ cheat.memory_type == MemoryType::MainNSO ? main_region_begin : heap_region_begin;
+ ASSERT(static_cast<u8>(cheat.comparison_op.Value()) < 6);
+ auto* function = comparison_functions[static_cast<u8>(cheat.comparison_op.Value())];
+ const auto addr = cheat.Address() + offset;
+
+ return function(reader(cheat.width, SanitizeAddress(addr)), cheat.ValueWidth(8));
+}
+
+void CheatList::ProcessBlockPairs(const Block& block) {
+ block_pairs.clear();
+
+ u64 scope = 0;
+ std::map<u64, u64> pairs;
+
+ for (std::size_t i = 0; i < block.size(); ++i) {
+ const auto& cheat = block[i];
+
+ switch (cheat.type) {
+ case CodeType::Conditional:
+ case CodeType::ConditionalInput:
+ pairs.insert_or_assign(scope, i);
+ ++scope;
+ break;
+ case CodeType::EndConditional: {
+ --scope;
+ const auto idx = pairs.at(scope);
+ block_pairs.insert_or_assign(idx, i);
+ break;
+ }
+ case CodeType::Loop: {
+ if (cheat.end_of_loop) {
+ --scope;
+ const auto idx = pairs.at(scope);
+ block_pairs.insert_or_assign(idx, i);
+ } else {
+ pairs.insert_or_assign(scope, i);
+ ++scope;
+ }
+ break;
+ }
+ }
+ }
+}
+
+void CheatList::WriteImmediate(const Cheat& cheat) {
+ const auto offset =
+ cheat.memory_type == MemoryType::MainNSO ? main_region_begin : heap_region_begin;
+ const auto& register_3 = scratch.at(cheat.register_3);
+
+ const auto addr = cheat.Address() + offset + register_3;
+ LOG_DEBUG(Common_Filesystem, "writing value={:016X} to addr={:016X}", addr,
+ cheat.Value(8, cheat.width));
+ writer(cheat.width, SanitizeAddress(addr), cheat.ValueWidth(8));
+}
+
+void CheatList::BeginConditional(const Cheat& cheat) {
+ if (EvaluateConditional(cheat)) {
+ return;
+ }
+
+ const auto iter = block_pairs.find(current_index);
+ ASSERT(iter != block_pairs.end());
+ current_index = iter->second - 1;
+}
+
+void CheatList::EndConditional(const Cheat& cheat) {
+ LOG_DEBUG(Common_Filesystem, "Ending conditional block.");
+}
+
+void CheatList::Loop(const Cheat& cheat) {
+ if (cheat.end_of_loop.Value())
+ ASSERT(!cheat.end_of_loop.Value());
+
+ auto& register_3 = scratch.at(cheat.register_3);
+ const auto iter = block_pairs.find(current_index);
+ ASSERT(iter != block_pairs.end());
+ ASSERT(iter->first < iter->second);
+
+ const s32 initial_value = static_cast<s32>(cheat.Value(4, sizeof(s32)));
+ for (s32 i = initial_value; i >= 0; --i) {
+ register_3 = static_cast<u64>(i);
+ for (std::size_t c = iter->first + 1; c < iter->second; ++c) {
+ current_index = c;
+ ExecuteSingleCheat(
+ (in_standard ? standard_list : master_list)[current_block].second[c]);
+ }
+ }
+
+ current_index = iter->second;
+}
+
+void CheatList::LoadImmediate(const Cheat& cheat) {
+ auto& register_3 = scratch.at(cheat.register_3);
+
+ LOG_DEBUG(Common_Filesystem, "setting register={:01X} equal to value={:016X}", cheat.register_3,
+ cheat.Value(4, 8));
+ register_3 = cheat.Value(4, 8);
+}
+
+void CheatList::LoadIndexed(const Cheat& cheat) {
+ const auto offset =
+ cheat.memory_type == MemoryType::MainNSO ? main_region_begin : heap_region_begin;
+ auto& register_3 = scratch.at(cheat.register_3);
+
+ const auto addr = (cheat.load_from_register.Value() ? register_3 : offset) + cheat.Address();
+ LOG_DEBUG(Common_Filesystem, "writing indexed value to register={:01X}, addr={:016X}",
+ cheat.register_3, addr);
+ register_3 = reader(cheat.width, SanitizeAddress(addr));
+}
+
+void CheatList::StoreIndexed(const Cheat& cheat) {
+ const auto& register_3 = scratch.at(cheat.register_3);
+
+ const auto addr =
+ register_3 + (cheat.add_additional_register.Value() ? scratch.at(cheat.register_6) : 0);
+ LOG_DEBUG(Common_Filesystem, "writing value={:016X} to addr={:016X}",
+ cheat.Value(4, cheat.width), addr);
+ writer(cheat.width, SanitizeAddress(addr), cheat.ValueWidth(4));
+}
+
+void CheatList::RegisterArithmetic(const Cheat& cheat) {
+ using ArithmeticFunction = u64 (*)(u64, u64);
+ constexpr std::array<ArithmeticFunction, 5> arithmetic_functions{
+ [](u64 a, u64 b) { return a + b; }, [](u64 a, u64 b) { return a - b; },
+ [](u64 a, u64 b) { return a * b; }, [](u64 a, u64 b) { return a << b; },
+ [](u64 a, u64 b) { return a >> b; },
+ };
+
+ using ArithmeticOverflowCheck = bool (*)(u64, u64);
+ constexpr std::array<ArithmeticOverflowCheck, 5> arithmetic_overflow_checks{
+ [](u64 a, u64 b) { return a > (std::numeric_limits<u64>::max() - b); }, // a + b
+ [](u64 a, u64 b) { return a > (std::numeric_limits<u64>::max() + b); }, // a - b
+ [](u64 a, u64 b) { return a > (std::numeric_limits<u64>::max() / b); }, // a * b
+ [](u64 a, u64 b) { return b >= 64 || (a & ~((1ull << (64 - b)) - 1)) != 0; }, // a << b
+ [](u64 a, u64 b) { return b >= 64 || (a & ((1ull << b) - 1)) != 0; }, // a >> b
+ };
+
+ static_assert(sizeof(arithmetic_functions) == sizeof(arithmetic_overflow_checks),
+ "Missing or have extra arithmetic overflow checks compared to functions!");
+
+ auto& register_3 = scratch.at(cheat.register_3);
+
+ ASSERT(static_cast<u8>(cheat.arithmetic_op.Value()) < 5);
+ auto* function = arithmetic_functions[static_cast<u8>(cheat.arithmetic_op.Value())];
+ auto* overflow_function =
+ arithmetic_overflow_checks[static_cast<u8>(cheat.arithmetic_op.Value())];
+ LOG_DEBUG(Common_Filesystem, "performing arithmetic with register={:01X}, value={:016X}",
+ cheat.register_3, cheat.ValueWidth(4));
+
+ if (overflow_function(register_3, cheat.ValueWidth(4))) {
+ LOG_WARNING(Common_Filesystem,
+ "overflow will occur when performing arithmetic operation={:02X} with operands "
+ "a={:016X}, b={:016X}!",
+ static_cast<u8>(cheat.arithmetic_op.Value()), register_3, cheat.ValueWidth(4));
+ }
+
+ register_3 = function(register_3, cheat.ValueWidth(4));
+}
+
+void CheatList::BeginConditionalInput(const Cheat& cheat) {
+ if (EvaluateConditional(cheat))
+ return;
+
+ const auto iter = block_pairs.find(current_index);
+ ASSERT(iter != block_pairs.end());
+ current_index = iter->second - 1;
+}
+
+VAddr CheatList::SanitizeAddress(VAddr in) const {
+ if ((in < main_region_begin || in >= main_region_end) &&
+ (in < heap_region_begin || in >= heap_region_end)) {
+ LOG_ERROR(Common_Filesystem,
+ "Cheat attempting to access memory at invalid address={:016X}, if this persists, "
+ "the cheat may be incorrect. However, this may be normal early in execution if "
+ "the game has not properly set up yet.",
+ in);
+ return 0; ///< Invalid addresses will hard crash
+ }
+
+ return in;
+}
+
+void CheatList::ExecuteSingleCheat(const Cheat& cheat) {
+ using CheatOperationFunction = void (CheatList::*)(const Cheat&);
+ constexpr std::array<CheatOperationFunction, 9> cheat_operation_functions{
+ &CheatList::WriteImmediate, &CheatList::BeginConditional,
+ &CheatList::EndConditional, &CheatList::Loop,
+ &CheatList::LoadImmediate, &CheatList::LoadIndexed,
+ &CheatList::StoreIndexed, &CheatList::RegisterArithmetic,
+ &CheatList::BeginConditionalInput,
+ };
+
+ const auto index = static_cast<u8>(cheat.type.Value());
+ ASSERT(index < sizeof(cheat_operation_functions));
+ const auto op = cheat_operation_functions[index];
+ (this->*op)(cheat);
+}
+
+void CheatList::ExecuteBlock(const Block& block) {
+ encountered_loops.clear();
+
+ ProcessBlockPairs(block);
+ for (std::size_t i = 0; i < block.size(); ++i) {
+ current_index = i;
+ ExecuteSingleCheat(block[i]);
+ i = current_index;
+ }
+}
+
+CheatParser::~CheatParser() = default;
+
+CheatList CheatParser::MakeCheatList(const Core::System& system, CheatList::ProgramSegment master,
+ CheatList::ProgramSegment standard) const {
+ return {system, std::move(master), std::move(standard)};
+}
+
+TextCheatParser::~TextCheatParser() = default;
+
+CheatList TextCheatParser::Parse(const Core::System& system, const std::vector<u8>& data) const {
+ std::stringstream ss;
+ ss.write(reinterpret_cast<const char*>(data.data()), data.size());
+
+ std::vector<std::string> lines;
+ std::string stream_line;
+ while (std::getline(ss, stream_line)) {
+ // Remove a trailing \r
+ if (!stream_line.empty() && stream_line.back() == '\r')
+ stream_line.pop_back();
+ lines.push_back(std::move(stream_line));
+ }
+
+ CheatList::ProgramSegment master_list;
+ CheatList::ProgramSegment standard_list;
+
+ for (std::size_t i = 0; i < lines.size(); ++i) {
+ auto line = lines[i];
+
+ if (!line.empty() && (line[0] == '[' || line[0] == '{')) {
+ const auto master = line[0] == '{';
+ const auto begin = master ? line.find('{') : line.find('[');
+ const auto end = master ? line.rfind('}') : line.rfind(']');
+
+ ASSERT(begin != std::string::npos && end != std::string::npos);
+
+ const std::string patch_name{line.begin() + begin + 1, line.begin() + end};
+ CheatList::Block block{};
+
+ while (i < lines.size() - 1) {
+ line = lines[++i];
+ if (!line.empty() && (line[0] == '[' || line[0] == '{')) {
+ --i;
+ break;
+ }
+
+ if (line.size() < 8)
+ continue;
+
+ Cheat out{};
+ out.raw = ParseSingleLineCheat(line);
+ block.push_back(out);
+ }
+
+ (master ? master_list : standard_list).emplace_back(patch_name, block);
+ }
+ }
+
+ return MakeCheatList(system, master_list, standard_list);
+}
+
+std::array<u8, 16> TextCheatParser::ParseSingleLineCheat(const std::string& line) const {
+ std::array<u8, 16> out{};
+
+ if (line.size() < 8)
+ return out;
+
+ const auto word1 = Common::HexStringToArray<sizeof(u32)>(std::string_view{line.data(), 8});
+ std::memcpy(out.data(), word1.data(), sizeof(u32));
+
+ if (line.size() < 17 || line[8] != ' ')
+ return out;
+
+ const auto word2 = Common::HexStringToArray<sizeof(u32)>(std::string_view{line.data() + 9, 8});
+ std::memcpy(out.data() + sizeof(u32), word2.data(), sizeof(u32));
+
+ if (line.size() < 26 || line[17] != ' ') {
+ // Perform shifting in case value is truncated early.
+ const auto type = static_cast<CodeType>((out[0] & 0xF0) >> 4);
+ if (type == CodeType::Loop || type == CodeType::LoadImmediate ||
+ type == CodeType::StoreIndexed || type == CodeType::RegisterArithmetic) {
+ std::memcpy(out.data() + 8, out.data() + 4, sizeof(u32));
+ std::memset(out.data() + 4, 0, sizeof(u32));
+ }
+
+ return out;
+ }
+
+ const auto word3 = Common::HexStringToArray<sizeof(u32)>(std::string_view{line.data() + 18, 8});
+ std::memcpy(out.data() + 2 * sizeof(u32), word3.data(), sizeof(u32));
+
+ if (line.size() < 35 || line[26] != ' ') {
+ // Perform shifting in case value is truncated early.
+ const auto type = static_cast<CodeType>((out[0] & 0xF0) >> 4);
+ if (type == CodeType::WriteImmediate || type == CodeType::Conditional) {
+ std::memcpy(out.data() + 12, out.data() + 8, sizeof(u32));
+ std::memset(out.data() + 8, 0, sizeof(u32));
+ }
+
+ return out;
+ }
+
+ const auto word4 = Common::HexStringToArray<sizeof(u32)>(std::string_view{line.data() + 27, 8});
+ std::memcpy(out.data() + 3 * sizeof(u32), word4.data(), sizeof(u32));
+
+ return out;
+}
+
+namespace {
+u64 MemoryReadImpl(u32 width, VAddr addr) {
+ switch (width) {
+ case 1:
+ return Memory::Read8(addr);
+ case 2:
+ return Memory::Read16(addr);
+ case 4:
+ return Memory::Read32(addr);
+ case 8:
+ return Memory::Read64(addr);
+ default:
+ UNREACHABLE();
+ return 0;
+ }
+}
+
+void MemoryWriteImpl(u32 width, VAddr addr, u64 value) {
+ switch (width) {
+ case 1:
+ Memory::Write8(addr, static_cast<u8>(value));
+ break;
+ case 2:
+ Memory::Write16(addr, static_cast<u16>(value));
+ break;
+ case 4:
+ Memory::Write32(addr, static_cast<u32>(value));
+ break;
+ case 8:
+ Memory::Write64(addr, value);
+ break;
+ default:
+ UNREACHABLE();
+ }
+}
+} // Anonymous namespace
+
+CheatEngine::CheatEngine(Core::System& system, std::vector<CheatList> cheats_,
+ const std::string& build_id, VAddr code_region_start,
+ VAddr code_region_end)
+ : cheats{std::move(cheats_)}, core_timing{system.CoreTiming()} {
+ event = core_timing.RegisterEvent(
+ "CheatEngine::FrameCallback::" + build_id,
+ [this](u64 userdata, s64 cycles_late) { FrameCallback(userdata, cycles_late); });
+ core_timing.ScheduleEvent(CHEAT_ENGINE_TICKS, event);
+
+ const auto& vm_manager = system.CurrentProcess()->VMManager();
+ for (auto& list : this->cheats) {
+ list.SetMemoryParameters(code_region_start, vm_manager.GetHeapRegionBaseAddress(),
+ code_region_end, vm_manager.GetHeapRegionEndAddress(),
+ &MemoryWriteImpl, &MemoryReadImpl);
+ }
+}
+
+CheatEngine::~CheatEngine() {
+ core_timing.UnscheduleEvent(event, 0);
+}
+
+void CheatEngine::FrameCallback(u64 userdata, s64 cycles_late) {
+ for (auto& list : cheats) {
+ list.Execute();
+ }
+
+ core_timing.ScheduleEvent(CHEAT_ENGINE_TICKS - cycles_late, event);
+}
+
+} // namespace FileSys
diff --git a/src/core/file_sys/cheat_engine.h b/src/core/file_sys/cheat_engine.h
new file mode 100644
index 000000000..ac22a82cb
--- /dev/null
+++ b/src/core/file_sys/cheat_engine.h
@@ -0,0 +1,234 @@
+// Copyright 2018 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <map>
+#include <set>
+#include <vector>
+#include "common/bit_field.h"
+#include "common/common_types.h"
+
+namespace Core {
+class System;
+}
+
+namespace Core::Timing {
+class CoreTiming;
+struct EventType;
+} // namespace Core::Timing
+
+namespace FileSys {
+
+enum class CodeType : u32 {
+ // 0TMR00AA AAAAAAAA YYYYYYYY YYYYYYYY
+ // Writes a T sized value Y to the address A added to the value of register R in memory domain M
+ WriteImmediate = 0,
+
+ // 1TMC00AA AAAAAAAA YYYYYYYY YYYYYYYY
+ // Compares the T sized value Y to the value at address A in memory domain M using the
+ // conditional function C. If success, continues execution. If failure, jumps to the matching
+ // EndConditional statement.
+ Conditional = 1,
+
+ // 20000000
+ // Terminates a Conditional or ConditionalInput block.
+ EndConditional = 2,
+
+ // 300R0000 VVVVVVVV
+ // Starts looping V times, storing the current count in register R.
+ // Loop block is terminated with a matching 310R0000.
+ Loop = 3,
+
+ // 400R0000 VVVVVVVV VVVVVVVV
+ // Sets the value of register R to the value V.
+ LoadImmediate = 4,
+
+ // 5TMRI0AA AAAAAAAA
+ // Sets the value of register R to the value of width T at address A in memory domain M, with
+ // the current value of R added to the address if I == 1.
+ LoadIndexed = 5,
+
+ // 6T0RIFG0 VVVVVVVV VVVVVVVV
+ // Writes the value V of width T to the memory address stored in register R. Adds the value of
+ // register G to the final calculation if F is nonzero. Increments the value of register R by T
+ // after operation if I is nonzero.
+ StoreIndexed = 6,
+
+ // 7T0RA000 VVVVVVVV
+ // Performs the arithmetic operation A on the value in register R and the value V of width T,
+ // storing the result in register R.
+ RegisterArithmetic = 7,
+
+ // 8KKKKKKK
+ // Checks to see if any of the buttons defined by the bitmask K are pressed. If any are,
+ // execution continues. If none are, execution skips to the next EndConditional command.
+ ConditionalInput = 8,
+};
+
+enum class MemoryType : u32 {
+ // Addressed relative to start of main NSO
+ MainNSO = 0,
+
+ // Addressed relative to start of heap
+ Heap = 1,
+};
+
+enum class ArithmeticOp : u32 {
+ Add = 0,
+ Sub = 1,
+ Mult = 2,
+ LShift = 3,
+ RShift = 4,
+};
+
+enum class ComparisonOp : u32 {
+ GreaterThan = 1,
+ GreaterThanEqual = 2,
+ LessThan = 3,
+ LessThanEqual = 4,
+ Equal = 5,
+ Inequal = 6,
+};
+
+union Cheat {
+ std::array<u8, 16> raw;
+
+ BitField<4, 4, CodeType> type;
+ BitField<0, 4, u32> width; // Can be 1, 2, 4, or 8. Measured in bytes.
+ BitField<0, 4, u32> end_of_loop;
+ BitField<12, 4, MemoryType> memory_type;
+ BitField<8, 4, u32> register_3;
+ BitField<8, 4, ComparisonOp> comparison_op;
+ BitField<20, 4, u32> load_from_register;
+ BitField<20, 4, u32> increment_register;
+ BitField<20, 4, ArithmeticOp> arithmetic_op;
+ BitField<16, 4, u32> add_additional_register;
+ BitField<28, 4, u32> register_6;
+
+ u64 Address() const;
+ u64 ValueWidth(u64 offset) const;
+ u64 Value(u64 offset, u64 width) const;
+ u32 KeypadValue() const;
+};
+
+class CheatParser;
+
+// Represents a full collection of cheats for a game. The Execute function should be called every
+// interval that all cheats should be executed. Clients should not directly instantiate this class
+// (hence private constructor), they should instead receive an instance from CheatParser, which
+// guarantees the list is always in an acceptable state.
+class CheatList {
+public:
+ friend class CheatParser;
+
+ using Block = std::vector<Cheat>;
+ using ProgramSegment = std::vector<std::pair<std::string, Block>>;
+
+ // (width in bytes, address, value)
+ using MemoryWriter = void (*)(u32, VAddr, u64);
+ // (width in bytes, address) -> value
+ using MemoryReader = u64 (*)(u32, VAddr);
+
+ void SetMemoryParameters(VAddr main_begin, VAddr heap_begin, VAddr main_end, VAddr heap_end,
+ MemoryWriter writer, MemoryReader reader);
+
+ void Execute();
+
+private:
+ CheatList(const Core::System& system_, ProgramSegment master, ProgramSegment standard);
+
+ void ProcessBlockPairs(const Block& block);
+ void ExecuteSingleCheat(const Cheat& cheat);
+
+ void ExecuteBlock(const Block& block);
+
+ bool EvaluateConditional(const Cheat& cheat) const;
+
+ // Individual cheat operations
+ void WriteImmediate(const Cheat& cheat);
+ void BeginConditional(const Cheat& cheat);
+ void EndConditional(const Cheat& cheat);
+ void Loop(const Cheat& cheat);
+ void LoadImmediate(const Cheat& cheat);
+ void LoadIndexed(const Cheat& cheat);
+ void StoreIndexed(const Cheat& cheat);
+ void RegisterArithmetic(const Cheat& cheat);
+ void BeginConditionalInput(const Cheat& cheat);
+
+ VAddr SanitizeAddress(VAddr in) const;
+
+ // Master Codes are defined as codes that cannot be disabled and are run prior to all
+ // others.
+ ProgramSegment master_list;
+ // All other codes
+ ProgramSegment standard_list;
+
+ bool in_standard = false;
+
+ // 16 (0x0-0xF) scratch registers that can be used by cheats
+ std::array<u64, 16> scratch{};
+
+ MemoryWriter writer = nullptr;
+ MemoryReader reader = nullptr;
+
+ u64 main_region_begin{};
+ u64 heap_region_begin{};
+ u64 main_region_end{};
+ u64 heap_region_end{};
+
+ u64 current_block{};
+ // The current index of the cheat within the current Block
+ u64 current_index{};
+
+ // The 'stack' of the program. When a conditional or loop statement is encountered, its index is
+ // pushed onto this queue. When a end block is encountered, the condition is checked.
+ std::map<u64, u64> block_pairs;
+
+ std::set<u64> encountered_loops;
+
+ const Core::System* system;
+};
+
+// Intermediary class that parses a text file or other disk format for storing cheats into a
+// CheatList object, that can be used for execution.
+class CheatParser {
+public:
+ virtual ~CheatParser();
+
+ virtual CheatList Parse(const Core::System& system, const std::vector<u8>& data) const = 0;
+
+protected:
+ CheatList MakeCheatList(const Core::System& system_, CheatList::ProgramSegment master,
+ CheatList::ProgramSegment standard) const;
+};
+
+// CheatParser implementation that parses text files
+class TextCheatParser final : public CheatParser {
+public:
+ ~TextCheatParser() override;
+
+ CheatList Parse(const Core::System& system, const std::vector<u8>& data) const override;
+
+private:
+ std::array<u8, 16> ParseSingleLineCheat(const std::string& line) const;
+};
+
+// Class that encapsulates a CheatList and manages its interaction with memory and CoreTiming
+class CheatEngine final {
+public:
+ CheatEngine(Core::System& system_, std::vector<CheatList> cheats_, const std::string& build_id,
+ VAddr code_region_start, VAddr code_region_end);
+ ~CheatEngine();
+
+private:
+ void FrameCallback(u64 userdata, s64 cycles_late);
+
+ std::vector<CheatList> cheats;
+
+ Core::Timing::EventType* event;
+ Core::Timing::CoreTiming& core_timing;
+};
+
+} // namespace FileSys
diff --git a/src/core/file_sys/content_archive.h b/src/core/file_sys/content_archive.h
index 5d4d05c82..15b9e6624 100644
--- a/src/core/file_sys/content_archive.h
+++ b/src/core/file_sys/content_archive.h
@@ -24,13 +24,26 @@ namespace FileSys {
union NCASectionHeader;
+/// Describes the type of content within an NCA archive.
enum class NCAContentType : u8 {
+ /// Executable-related data
Program = 0,
+
+ /// Metadata.
Meta = 1,
+
+ /// Access control data.
Control = 2,
+
+ /// Information related to the game manual
+ /// e.g. Legal information, etc.
Manual = 3,
+
+ /// System data.
Data = 4,
- Data_Unknown5 = 5, ///< Seems to be used on some system archives
+
+ /// Data that can be accessed by applications.
+ PublicData = 5,
};
enum class NCASectionCryptoType : u8 {
diff --git a/src/core/file_sys/errors.h b/src/core/file_sys/errors.h
index e4a4ee4ab..bb4654366 100644
--- a/src/core/file_sys/errors.h
+++ b/src/core/file_sys/errors.h
@@ -11,6 +11,9 @@ namespace FileSys {
constexpr ResultCode ERROR_PATH_NOT_FOUND{ErrorModule::FS, 1};
constexpr ResultCode ERROR_ENTITY_NOT_FOUND{ErrorModule::FS, 1002};
constexpr ResultCode ERROR_SD_CARD_NOT_FOUND{ErrorModule::FS, 2001};
+constexpr ResultCode ERROR_OUT_OF_BOUNDS{ErrorModule::FS, 3005};
+constexpr ResultCode ERROR_FAILED_MOUNT_ARCHIVE{ErrorModule::FS, 3223};
+constexpr ResultCode ERROR_INVALID_ARGUMENT{ErrorModule::FS, 6001};
constexpr ResultCode ERROR_INVALID_OFFSET{ErrorModule::FS, 6061};
constexpr ResultCode ERROR_INVALID_SIZE{ErrorModule::FS, 6062};
diff --git a/src/core/file_sys/patch_manager.cpp b/src/core/file_sys/patch_manager.cpp
index 61706966e..e11217708 100644
--- a/src/core/file_sys/patch_manager.cpp
+++ b/src/core/file_sys/patch_manager.cpp
@@ -7,6 +7,7 @@
#include <cstddef>
#include <cstring>
+#include "common/file_util.h"
#include "common/hex_util.h"
#include "common/logging/log.h"
#include "core/file_sys/content_archive.h"
@@ -19,6 +20,7 @@
#include "core/file_sys/vfs_vector.h"
#include "core/hle/service/filesystem/filesystem.h"
#include "core/loader/loader.h"
+#include "core/loader/nso.h"
#include "core/settings.h"
namespace FileSys {
@@ -31,14 +33,6 @@ constexpr std::array<const char*, 14> EXEFS_FILE_NAMES{
"subsdk3", "subsdk4", "subsdk5", "subsdk6", "subsdk7", "subsdk8", "subsdk9",
};
-struct NSOBuildHeader {
- u32_le magic;
- INSERT_PADDING_BYTES(0x3C);
- std::array<u8, 0x20> build_id;
- INSERT_PADDING_BYTES(0xA0);
-};
-static_assert(sizeof(NSOBuildHeader) == 0x100, "NSOBuildHeader has incorrect size.");
-
std::string FormatTitleVersion(u32 version, TitleVersionFormat format) {
std::array<u8, sizeof(u32)> bytes{};
bytes[0] = version % SINGLE_BYTE_MODULUS;
@@ -162,14 +156,16 @@ std::vector<VirtualFile> PatchManager::CollectPatches(const std::vector<VirtualD
}
std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso) const {
- if (nso.size() < 0x100)
+ if (nso.size() < sizeof(Loader::NSOHeader)) {
return nso;
+ }
- NSOBuildHeader header;
- std::memcpy(&header, nso.data(), sizeof(NSOBuildHeader));
+ Loader::NSOHeader header;
+ std::memcpy(&header, nso.data(), sizeof(header));
- if (header.magic != Common::MakeMagic('N', 'S', 'O', '0'))
+ if (header.magic != Common::MakeMagic('N', 'S', 'O', '0')) {
return nso;
+ }
const auto build_id_raw = Common::HexArrayToString(header.build_id);
const auto build_id = build_id_raw.substr(0, build_id_raw.find_last_not_of('0') + 1);
@@ -212,9 +208,11 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso) const {
}
}
- if (out.size() < 0x100)
+ if (out.size() < sizeof(Loader::NSOHeader)) {
return nso;
- std::memcpy(out.data(), &header, sizeof(NSOBuildHeader));
+ }
+
+ std::memcpy(out.data(), &header, sizeof(header));
return out;
}
@@ -232,6 +230,57 @@ bool PatchManager::HasNSOPatch(const std::array<u8, 32>& build_id_) const {
return !CollectPatches(patch_dirs, build_id).empty();
}
+static std::optional<CheatList> ReadCheatFileFromFolder(const Core::System& system, u64 title_id,
+ const std::array<u8, 0x20>& build_id_,
+ const VirtualDir& base_path, bool upper) {
+ const auto build_id_raw = Common::HexArrayToString(build_id_, upper);
+ const auto build_id = build_id_raw.substr(0, sizeof(u64) * 2);
+ const auto file = base_path->GetFile(fmt::format("{}.txt", build_id));
+
+ if (file == nullptr) {
+ LOG_INFO(Common_Filesystem, "No cheats file found for title_id={:016X}, build_id={}",
+ title_id, build_id);
+ return std::nullopt;
+ }
+
+ std::vector<u8> data(file->GetSize());
+ if (file->Read(data.data(), data.size()) != data.size()) {
+ LOG_INFO(Common_Filesystem, "Failed to read cheats file for title_id={:016X}, build_id={}",
+ title_id, build_id);
+ return std::nullopt;
+ }
+
+ TextCheatParser parser;
+ return parser.Parse(system, data);
+}
+
+std::vector<CheatList> PatchManager::CreateCheatList(const Core::System& system,
+ const std::array<u8, 32>& build_id_) const {
+ const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
+ auto patch_dirs = load_dir->GetSubdirectories();
+ std::sort(patch_dirs.begin(), patch_dirs.end(),
+ [](const VirtualDir& l, const VirtualDir& r) { return l->GetName() < r->GetName(); });
+
+ std::vector<CheatList> out;
+ out.reserve(patch_dirs.size());
+ for (const auto& subdir : patch_dirs) {
+ auto cheats_dir = subdir->GetSubdirectory("cheats");
+ if (cheats_dir != nullptr) {
+ auto res = ReadCheatFileFromFolder(system, title_id, build_id_, cheats_dir, true);
+ if (res.has_value()) {
+ out.push_back(std::move(*res));
+ continue;
+ }
+
+ res = ReadCheatFileFromFolder(system, title_id, build_id_, cheats_dir, false);
+ if (res.has_value())
+ out.push_back(std::move(*res));
+ }
+ }
+
+ return out;
+}
+
static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType type) {
const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id);
if ((type != ContentRecordType::Program && type != ContentRecordType::Data) ||
@@ -403,6 +452,8 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam
}
if (IsDirValidAndNonEmpty(mod->GetSubdirectory("romfs")))
AppendCommaIfNotEmpty(types, "LayeredFS");
+ if (IsDirValidAndNonEmpty(mod->GetSubdirectory("cheats")))
+ AppendCommaIfNotEmpty(types, "Cheats");
if (types.empty())
continue;
diff --git a/src/core/file_sys/patch_manager.h b/src/core/file_sys/patch_manager.h
index b8a1652fd..de2672c76 100644
--- a/src/core/file_sys/patch_manager.h
+++ b/src/core/file_sys/patch_manager.h
@@ -8,9 +8,14 @@
#include <memory>
#include <string>
#include "common/common_types.h"
+#include "core/file_sys/cheat_engine.h"
#include "core/file_sys/nca_metadata.h"
#include "core/file_sys/vfs.h"
+namespace Core {
+class System;
+}
+
namespace FileSys {
class NCA;
@@ -45,6 +50,10 @@ public:
// Used to prevent expensive copies in NSO loader.
bool HasNSOPatch(const std::array<u8, 0x20>& build_id) const;
+ // Creates a CheatList object with all
+ std::vector<CheatList> CreateCheatList(const Core::System& system,
+ const std::array<u8, 0x20>& build_id) const;
+
// Currently tracked RomFS patches:
// - Game Updates
// - LayeredFS
diff --git a/src/core/file_sys/registered_cache.cpp b/src/core/file_sys/registered_cache.cpp
index 128199063..1c6bacace 100644
--- a/src/core/file_sys/registered_cache.cpp
+++ b/src/core/file_sys/registered_cache.cpp
@@ -94,7 +94,7 @@ static ContentRecordType GetCRTypeFromNCAType(NCAContentType type) {
case NCAContentType::Control:
return ContentRecordType::Control;
case NCAContentType::Data:
- case NCAContentType::Data_Unknown5:
+ case NCAContentType::PublicData:
return ContentRecordType::Data;
case NCAContentType::Manual:
// TODO(DarkLordZach): Peek at NCA contents to differentiate Manual and Legal.
diff --git a/src/core/file_sys/system_archive/system_archive.cpp b/src/core/file_sys/system_archive/system_archive.cpp
index e3e79f40a..c9722ed77 100644
--- a/src/core/file_sys/system_archive/system_archive.cpp
+++ b/src/core/file_sys/system_archive/system_archive.cpp
@@ -6,6 +6,7 @@
#include "core/file_sys/romfs.h"
#include "core/file_sys/system_archive/ng_word.h"
#include "core/file_sys/system_archive/system_archive.h"
+#include "core/file_sys/system_archive/system_version.h"
namespace FileSys::SystemArchive {
@@ -30,7 +31,7 @@ constexpr std::array<SystemArchiveDescriptor, SYSTEM_ARCHIVE_COUNT> SYSTEM_ARCHI
{0x0100000000000806, "NgWord", &NgWord1},
{0x0100000000000807, "SsidList", nullptr},
{0x0100000000000808, "Dictionary", nullptr},
- {0x0100000000000809, "SystemVersion", nullptr},
+ {0x0100000000000809, "SystemVersion", &SystemVersion},
{0x010000000000080A, "AvatarImage", nullptr},
{0x010000000000080B, "LocalNews", nullptr},
{0x010000000000080C, "Eula", nullptr},
diff --git a/src/core/file_sys/system_archive/system_version.cpp b/src/core/file_sys/system_archive/system_version.cpp
new file mode 100644
index 000000000..6e22f97b0
--- /dev/null
+++ b/src/core/file_sys/system_archive/system_version.cpp
@@ -0,0 +1,52 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/file_sys/system_archive/system_version.h"
+#include "core/file_sys/vfs_vector.h"
+
+namespace FileSys::SystemArchive {
+
+namespace SystemVersionData {
+
+// This section should reflect the best system version to describe yuzu's HLE api.
+// TODO(DarkLordZach): Update when HLE gets better.
+
+constexpr u8 VERSION_MAJOR = 5;
+constexpr u8 VERSION_MINOR = 1;
+constexpr u8 VERSION_MICRO = 0;
+
+constexpr u8 REVISION_MAJOR = 3;
+constexpr u8 REVISION_MINOR = 0;
+
+constexpr char PLATFORM_STRING[] = "NX";
+constexpr char VERSION_HASH[] = "23f9df53e25709d756e0c76effcb2473bd3447dd";
+constexpr char DISPLAY_VERSION[] = "5.1.0";
+constexpr char DISPLAY_TITLE[] = "NintendoSDK Firmware for NX 5.1.0-3.0";
+
+} // namespace SystemVersionData
+
+std::string GetLongDisplayVersion() {
+ return SystemVersionData::DISPLAY_TITLE;
+}
+
+VirtualDir SystemVersion() {
+ VirtualFile file = std::make_shared<VectorVfsFile>(std::vector<u8>(0x100), "file");
+ file->WriteObject(SystemVersionData::VERSION_MAJOR, 0);
+ file->WriteObject(SystemVersionData::VERSION_MINOR, 1);
+ file->WriteObject(SystemVersionData::VERSION_MICRO, 2);
+ file->WriteObject(SystemVersionData::REVISION_MAJOR, 4);
+ file->WriteObject(SystemVersionData::REVISION_MINOR, 5);
+ file->WriteArray(SystemVersionData::PLATFORM_STRING,
+ std::min<u64>(sizeof(SystemVersionData::PLATFORM_STRING), 0x20ULL), 0x8);
+ file->WriteArray(SystemVersionData::VERSION_HASH,
+ std::min<u64>(sizeof(SystemVersionData::VERSION_HASH), 0x40ULL), 0x28);
+ file->WriteArray(SystemVersionData::DISPLAY_VERSION,
+ std::min<u64>(sizeof(SystemVersionData::DISPLAY_VERSION), 0x18ULL), 0x68);
+ file->WriteArray(SystemVersionData::DISPLAY_TITLE,
+ std::min<u64>(sizeof(SystemVersionData::DISPLAY_TITLE), 0x80ULL), 0x80);
+ return std::make_shared<VectorVfsDirectory>(std::vector<VirtualFile>{file},
+ std::vector<VirtualDir>{}, "data");
+}
+
+} // namespace FileSys::SystemArchive
diff --git a/src/core/file_sys/system_archive/system_version.h b/src/core/file_sys/system_archive/system_version.h
new file mode 100644
index 000000000..deed79b26
--- /dev/null
+++ b/src/core/file_sys/system_archive/system_version.h
@@ -0,0 +1,16 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <string>
+#include "core/file_sys/vfs_types.h"
+
+namespace FileSys::SystemArchive {
+
+std::string GetLongDisplayVersion();
+
+VirtualDir SystemVersion();
+
+} // namespace FileSys::SystemArchive
diff --git a/src/core/hle/ipc.h b/src/core/hle/ipc.h
index 455d1f346..fae54bcc7 100644
--- a/src/core/hle/ipc.h
+++ b/src/core/hle/ipc.h
@@ -39,10 +39,10 @@ struct CommandHeader {
union {
u32_le raw_low;
BitField<0, 16, CommandType> type;
- BitField<16, 4, u32_le> num_buf_x_descriptors;
- BitField<20, 4, u32_le> num_buf_a_descriptors;
- BitField<24, 4, u32_le> num_buf_b_descriptors;
- BitField<28, 4, u32_le> num_buf_w_descriptors;
+ BitField<16, 4, u32> num_buf_x_descriptors;
+ BitField<20, 4, u32> num_buf_a_descriptors;
+ BitField<24, 4, u32> num_buf_b_descriptors;
+ BitField<28, 4, u32> num_buf_w_descriptors;
};
enum class BufferDescriptorCFlag : u32 {
@@ -53,28 +53,28 @@ struct CommandHeader {
union {
u32_le raw_high;
- BitField<0, 10, u32_le> data_size;
+ BitField<0, 10, u32> data_size;
BitField<10, 4, BufferDescriptorCFlag> buf_c_descriptor_flags;
- BitField<31, 1, u32_le> enable_handle_descriptor;
+ BitField<31, 1, u32> enable_handle_descriptor;
};
};
static_assert(sizeof(CommandHeader) == 8, "CommandHeader size is incorrect");
union HandleDescriptorHeader {
u32_le raw_high;
- BitField<0, 1, u32_le> send_current_pid;
- BitField<1, 4, u32_le> num_handles_to_copy;
- BitField<5, 4, u32_le> num_handles_to_move;
+ BitField<0, 1, u32> send_current_pid;
+ BitField<1, 4, u32> num_handles_to_copy;
+ BitField<5, 4, u32> num_handles_to_move;
};
static_assert(sizeof(HandleDescriptorHeader) == 4, "HandleDescriptorHeader size is incorrect");
struct BufferDescriptorX {
union {
- BitField<0, 6, u32_le> counter_bits_0_5;
- BitField<6, 3, u32_le> address_bits_36_38;
- BitField<9, 3, u32_le> counter_bits_9_11;
- BitField<12, 4, u32_le> address_bits_32_35;
- BitField<16, 16, u32_le> size;
+ BitField<0, 6, u32> counter_bits_0_5;
+ BitField<6, 3, u32> address_bits_36_38;
+ BitField<9, 3, u32> counter_bits_9_11;
+ BitField<12, 4, u32> address_bits_32_35;
+ BitField<16, 16, u32> size;
};
u32_le address_bits_0_31;
@@ -103,10 +103,10 @@ struct BufferDescriptorABW {
u32_le address_bits_0_31;
union {
- BitField<0, 2, u32_le> flags;
- BitField<2, 3, u32_le> address_bits_36_38;
- BitField<24, 4, u32_le> size_bits_32_35;
- BitField<28, 4, u32_le> address_bits_32_35;
+ BitField<0, 2, u32> flags;
+ BitField<2, 3, u32> address_bits_36_38;
+ BitField<24, 4, u32> size_bits_32_35;
+ BitField<28, 4, u32> address_bits_32_35;
};
VAddr Address() const {
@@ -128,8 +128,8 @@ struct BufferDescriptorC {
u32_le address_bits_0_31;
union {
- BitField<0, 16, u32_le> address_bits_32_47;
- BitField<16, 16, u32_le> size;
+ BitField<0, 16, u32> address_bits_32_47;
+ BitField<16, 16, u32> size;
};
VAddr Address() const {
@@ -167,8 +167,8 @@ struct DomainMessageHeader {
struct {
union {
BitField<0, 8, CommandType> command;
- BitField<8, 8, u32_le> input_object_count;
- BitField<16, 16, u32_le> size;
+ BitField<8, 8, u32> input_object_count;
+ BitField<16, 16, u32> size;
};
u32_le object_id;
INSERT_PADDING_WORDS(2);
diff --git a/src/core/hle/kernel/code_set.cpp b/src/core/hle/kernel/code_set.cpp
new file mode 100644
index 000000000..1f434e9af
--- /dev/null
+++ b/src/core/hle/kernel/code_set.cpp
@@ -0,0 +1,12 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/hle/kernel/code_set.h"
+
+namespace Kernel {
+
+CodeSet::CodeSet() = default;
+CodeSet::~CodeSet() = default;
+
+} // namespace Kernel
diff --git a/src/core/hle/kernel/code_set.h b/src/core/hle/kernel/code_set.h
new file mode 100644
index 000000000..834fd23d2
--- /dev/null
+++ b/src/core/hle/kernel/code_set.h
@@ -0,0 +1,90 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <cstddef>
+#include <memory>
+#include <vector>
+
+#include "common/common_types.h"
+
+namespace Kernel {
+
+/**
+ * Represents executable data that may be loaded into a kernel process.
+ *
+ * A code set consists of three basic segments:
+ * - A code (AKA text) segment,
+ * - A read-only data segment (rodata)
+ * - A data segment
+ *
+ * The code segment is the portion of the object file that contains
+ * executable instructions.
+ *
+ * The read-only data segment in the portion of the object file that
+ * contains (as one would expect) read-only data, such as fixed constant
+ * values and data structures.
+ *
+ * The data segment is similar to the read-only data segment -- it contains
+ * variables and data structures that have predefined values, however,
+ * entities within this segment can be modified.
+ */
+struct CodeSet final {
+ /// A single segment within a code set.
+ struct Segment final {
+ /// The byte offset that this segment is located at.
+ std::size_t offset = 0;
+
+ /// The address to map this segment to.
+ VAddr addr = 0;
+
+ /// The size of this segment in bytes.
+ u32 size = 0;
+ };
+
+ explicit CodeSet();
+ ~CodeSet();
+
+ CodeSet(const CodeSet&) = delete;
+ CodeSet& operator=(const CodeSet&) = delete;
+
+ CodeSet(CodeSet&&) = default;
+ CodeSet& operator=(CodeSet&&) = default;
+
+ Segment& CodeSegment() {
+ return segments[0];
+ }
+
+ const Segment& CodeSegment() const {
+ return segments[0];
+ }
+
+ Segment& RODataSegment() {
+ return segments[1];
+ }
+
+ const Segment& RODataSegment() const {
+ return segments[1];
+ }
+
+ Segment& DataSegment() {
+ return segments[2];
+ }
+
+ const Segment& DataSegment() const {
+ return segments[2];
+ }
+
+ /// The overall data that backs this code set.
+ std::shared_ptr<std::vector<u8>> memory;
+
+ /// The segments that comprise this code set.
+ std::array<Segment, 3> segments;
+
+ /// The entry point address for this code set.
+ VAddr entrypoint = 0;
+};
+
+} // namespace Kernel
diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp
index 8de62d073..3b73be67b 100644
--- a/src/core/hle/kernel/kernel.cpp
+++ b/src/core/hle/kernel/kernel.cpp
@@ -29,7 +29,7 @@ namespace Kernel {
* @param thread_handle The handle of the thread that's been awoken
* @param cycles_late The number of CPU cycles that have passed since the desired wakeup time
*/
-static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] int cycles_late) {
+static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] s64 cycles_late) {
const auto proper_handle = static_cast<Handle>(thread_handle);
const auto& system = Core::System::GetInstance();
diff --git a/src/core/hle/kernel/kernel.h b/src/core/hle/kernel/kernel.h
index ff17ff865..03ea5b659 100644
--- a/src/core/hle/kernel/kernel.h
+++ b/src/core/hle/kernel/kernel.h
@@ -8,9 +8,6 @@
#include <unordered_map>
#include "core/hle/kernel/object.h"
-template <typename T>
-class ResultVal;
-
namespace Core {
class System;
}
diff --git a/src/core/hle/kernel/mutex.cpp b/src/core/hle/kernel/mutex.cpp
index 0743670ad..98e87313b 100644
--- a/src/core/hle/kernel/mutex.cpp
+++ b/src/core/hle/kernel/mutex.cpp
@@ -2,7 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <map>
#include <utility>
#include <vector>
@@ -10,8 +9,11 @@
#include "core/core.h"
#include "core/hle/kernel/errors.h"
#include "core/hle/kernel/handle_table.h"
+#include "core/hle/kernel/kernel.h"
#include "core/hle/kernel/mutex.h"
#include "core/hle/kernel/object.h"
+#include "core/hle/kernel/process.h"
+#include "core/hle/kernel/scheduler.h"
#include "core/hle/kernel/thread.h"
#include "core/hle/result.h"
#include "core/memory.h"
@@ -57,41 +59,47 @@ static void TransferMutexOwnership(VAddr mutex_addr, SharedPtr<Thread> current_t
}
}
-ResultCode Mutex::TryAcquire(HandleTable& handle_table, VAddr address, Handle holding_thread_handle,
+Mutex::Mutex(Core::System& system) : system{system} {}
+Mutex::~Mutex() = default;
+
+ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle,
Handle requesting_thread_handle) {
// The mutex address must be 4-byte aligned
if ((address % sizeof(u32)) != 0) {
return ERR_INVALID_ADDRESS;
}
+ const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable();
+ Thread* const current_thread = system.CurrentScheduler().GetCurrentThread();
SharedPtr<Thread> holding_thread = handle_table.Get<Thread>(holding_thread_handle);
SharedPtr<Thread> requesting_thread = handle_table.Get<Thread>(requesting_thread_handle);
// TODO(Subv): It is currently unknown if it is possible to lock a mutex in behalf of another
// thread.
- ASSERT(requesting_thread == GetCurrentThread());
+ ASSERT(requesting_thread == current_thread);
- u32 addr_value = Memory::Read32(address);
+ const u32 addr_value = Memory::Read32(address);
// If the mutex isn't being held, just return success.
if (addr_value != (holding_thread_handle | Mutex::MutexHasWaitersFlag)) {
return RESULT_SUCCESS;
}
- if (holding_thread == nullptr)
+ if (holding_thread == nullptr) {
return ERR_INVALID_HANDLE;
+ }
// Wait until the mutex is released
- GetCurrentThread()->SetMutexWaitAddress(address);
- GetCurrentThread()->SetWaitHandle(requesting_thread_handle);
+ current_thread->SetMutexWaitAddress(address);
+ current_thread->SetWaitHandle(requesting_thread_handle);
- GetCurrentThread()->SetStatus(ThreadStatus::WaitMutex);
- GetCurrentThread()->InvalidateWakeupCallback();
+ current_thread->SetStatus(ThreadStatus::WaitMutex);
+ current_thread->InvalidateWakeupCallback();
// Update the lock holder thread's priority to prevent priority inversion.
- holding_thread->AddMutexWaiter(GetCurrentThread());
+ holding_thread->AddMutexWaiter(current_thread);
- Core::System::GetInstance().PrepareReschedule();
+ system.PrepareReschedule();
return RESULT_SUCCESS;
}
@@ -102,7 +110,8 @@ ResultCode Mutex::Release(VAddr address) {
return ERR_INVALID_ADDRESS;
}
- auto [thread, num_waiters] = GetHighestPriorityMutexWaitingThread(GetCurrentThread(), address);
+ auto* const current_thread = system.CurrentScheduler().GetCurrentThread();
+ auto [thread, num_waiters] = GetHighestPriorityMutexWaitingThread(current_thread, address);
// There are no more threads waiting for the mutex, release it completely.
if (thread == nullptr) {
@@ -111,7 +120,7 @@ ResultCode Mutex::Release(VAddr address) {
}
// Transfer the ownership of the mutex from the previous owner to the new one.
- TransferMutexOwnership(address, GetCurrentThread(), thread);
+ TransferMutexOwnership(address, current_thread, thread);
u32 mutex_value = thread->GetWaitHandle();
diff --git a/src/core/hle/kernel/mutex.h b/src/core/hle/kernel/mutex.h
index 81e62d497..b904de2e8 100644
--- a/src/core/hle/kernel/mutex.h
+++ b/src/core/hle/kernel/mutex.h
@@ -5,32 +5,34 @@
#pragma once
#include "common/common_types.h"
-#include "core/hle/kernel/object.h"
union ResultCode;
-namespace Kernel {
+namespace Core {
+class System;
+}
-class HandleTable;
-class Thread;
+namespace Kernel {
class Mutex final {
public:
+ explicit Mutex(Core::System& system);
+ ~Mutex();
+
/// Flag that indicates that a mutex still has threads waiting for it.
static constexpr u32 MutexHasWaitersFlag = 0x40000000;
/// Mask of the bits in a mutex address value that contain the mutex owner.
static constexpr u32 MutexOwnerMask = 0xBFFFFFFF;
/// Attempts to acquire a mutex at the specified address.
- static ResultCode TryAcquire(HandleTable& handle_table, VAddr address,
- Handle holding_thread_handle, Handle requesting_thread_handle);
+ ResultCode TryAcquire(VAddr address, Handle holding_thread_handle,
+ Handle requesting_thread_handle);
/// Releases the mutex at the specified address.
- static ResultCode Release(VAddr address);
+ ResultCode Release(VAddr address);
private:
- Mutex() = default;
- ~Mutex() = default;
+ Core::System& system;
};
} // namespace Kernel
diff --git a/src/core/hle/kernel/object.cpp b/src/core/hle/kernel/object.cpp
index 8870463d0..217144efc 100644
--- a/src/core/hle/kernel/object.cpp
+++ b/src/core/hle/kernel/object.cpp
@@ -23,6 +23,7 @@ bool Object::IsWaitable() const {
case HandleType::Unknown:
case HandleType::WritableEvent:
case HandleType::SharedMemory:
+ case HandleType::TransferMemory:
case HandleType::AddressArbiter:
case HandleType::ResourceLimit:
case HandleType::ClientPort:
diff --git a/src/core/hle/kernel/object.h b/src/core/hle/kernel/object.h
index 4c2505908..3f6baa094 100644
--- a/src/core/hle/kernel/object.h
+++ b/src/core/hle/kernel/object.h
@@ -22,6 +22,7 @@ enum class HandleType : u32 {
WritableEvent,
ReadableEvent,
SharedMemory,
+ TransferMemory,
Thread,
Process,
AddressArbiter,
diff --git a/src/core/hle/kernel/process.cpp b/src/core/hle/kernel/process.cpp
index 65c51003d..0d782e4ba 100644
--- a/src/core/hle/kernel/process.cpp
+++ b/src/core/hle/kernel/process.cpp
@@ -9,6 +9,7 @@
#include "common/logging/log.h"
#include "core/core.h"
#include "core/file_sys/program_metadata.h"
+#include "core/hle/kernel/code_set.h"
#include "core/hle/kernel/errors.h"
#include "core/hle/kernel/kernel.h"
#include "core/hle/kernel/process.h"
@@ -50,9 +51,6 @@ void SetupMainThread(Process& owner_process, KernelCore& kernel, VAddr entry_poi
}
} // Anonymous namespace
-CodeSet::CodeSet() = default;
-CodeSet::~CodeSet() = default;
-
SharedPtr<Process> Process::Create(Core::System& system, std::string&& name) {
auto& kernel = system.Kernel();
@@ -212,7 +210,7 @@ void Process::FreeTLSSlot(VAddr tls_address) {
}
void Process::LoadModule(CodeSet module_, VAddr base_addr) {
- const auto MapSegment = [&](CodeSet::Segment& segment, VMAPermission permissions,
+ const auto MapSegment = [&](const CodeSet::Segment& segment, VMAPermission permissions,
MemoryState memory_state) {
const auto vma = vm_manager
.MapMemoryBlock(segment.addr + base_addr, module_.memory,
@@ -222,16 +220,17 @@ void Process::LoadModule(CodeSet module_, VAddr base_addr) {
};
// Map CodeSet segments
- MapSegment(module_.CodeSegment(), VMAPermission::ReadExecute, MemoryState::CodeStatic);
- MapSegment(module_.RODataSegment(), VMAPermission::Read, MemoryState::CodeMutable);
- MapSegment(module_.DataSegment(), VMAPermission::ReadWrite, MemoryState::CodeMutable);
+ MapSegment(module_.CodeSegment(), VMAPermission::ReadExecute, MemoryState::Code);
+ MapSegment(module_.RODataSegment(), VMAPermission::Read, MemoryState::CodeData);
+ MapSegment(module_.DataSegment(), VMAPermission::ReadWrite, MemoryState::CodeData);
// Clear instruction cache in CPU JIT
system.InvalidateCpuInstructionCaches();
}
Process::Process(Core::System& system)
- : WaitObject{system.Kernel()}, address_arbiter{system}, system{system} {}
+ : WaitObject{system.Kernel()}, address_arbiter{system}, mutex{system}, system{system} {}
+
Process::~Process() = default;
void Process::Acquire(Thread* thread) {
diff --git a/src/core/hle/kernel/process.h b/src/core/hle/kernel/process.h
index 47ffd4ad3..a0217d3d8 100644
--- a/src/core/hle/kernel/process.h
+++ b/src/core/hle/kernel/process.h
@@ -7,13 +7,13 @@
#include <array>
#include <bitset>
#include <cstddef>
-#include <memory>
#include <string>
#include <vector>
#include <boost/container/static_vector.hpp>
#include "common/common_types.h"
#include "core/hle/kernel/address_arbiter.h"
#include "core/hle/kernel/handle_table.h"
+#include "core/hle/kernel/mutex.h"
#include "core/hle/kernel/process_capability.h"
#include "core/hle/kernel/vm_manager.h"
#include "core/hle/kernel/wait_object.h"
@@ -33,13 +33,7 @@ class KernelCore;
class ResourceLimit;
class Thread;
-struct AddressMapping {
- // Address and size must be page-aligned
- VAddr address;
- u64 size;
- bool read_only;
- bool unk_flag;
-};
+struct CodeSet;
enum class MemoryRegion : u16 {
APPLICATION = 1,
@@ -65,46 +59,6 @@ enum class ProcessStatus {
DebugBreak,
};
-struct CodeSet final {
- struct Segment {
- std::size_t offset = 0;
- VAddr addr = 0;
- u32 size = 0;
- };
-
- explicit CodeSet();
- ~CodeSet();
-
- Segment& CodeSegment() {
- return segments[0];
- }
-
- const Segment& CodeSegment() const {
- return segments[0];
- }
-
- Segment& RODataSegment() {
- return segments[1];
- }
-
- const Segment& RODataSegment() const {
- return segments[1];
- }
-
- Segment& DataSegment() {
- return segments[2];
- }
-
- const Segment& DataSegment() const {
- return segments[2];
- }
-
- std::shared_ptr<std::vector<u8>> memory;
-
- std::array<Segment, 3> segments;
- VAddr entrypoint = 0;
-};
-
class Process final : public WaitObject {
public:
enum : u64 {
@@ -165,6 +119,16 @@ public:
return address_arbiter;
}
+ /// Gets a reference to the process' mutex lock.
+ Mutex& GetMutex() {
+ return mutex;
+ }
+
+ /// Gets a const reference to the process' mutex lock
+ const Mutex& GetMutex() const {
+ return mutex;
+ }
+
/// Gets the current status of the process
ProcessStatus GetStatus() const {
return status;
@@ -327,6 +291,11 @@ private:
/// Per-process address arbiter.
AddressArbiter address_arbiter;
+ /// The per-process mutex lock instance used for handling various
+ /// forms of services, such as lock arbitration, and condition
+ /// variable related facilities.
+ Mutex mutex;
+
/// Random values for svcGetInfo RandomEntropy
std::array<u64, RANDOM_ENTROPY_SIZE> random_entropy;
diff --git a/src/core/hle/kernel/scheduler.cpp b/src/core/hle/kernel/scheduler.cpp
index cc189cc64..6d0f13ecf 100644
--- a/src/core/hle/kernel/scheduler.cpp
+++ b/src/core/hle/kernel/scheduler.cpp
@@ -30,7 +30,7 @@ Scheduler::~Scheduler() {
bool Scheduler::HaveReadyThreads() const {
std::lock_guard<std::mutex> lock(scheduler_mutex);
- return ready_queue.get_first() != nullptr;
+ return !ready_queue.empty();
}
Thread* Scheduler::GetCurrentThread() const {
@@ -46,22 +46,27 @@ Thread* Scheduler::PopNextReadyThread() {
Thread* thread = GetCurrentThread();
if (thread && thread->GetStatus() == ThreadStatus::Running) {
+ if (ready_queue.empty()) {
+ return thread;
+ }
// We have to do better than the current thread.
// This call returns null when that's not possible.
- next = ready_queue.pop_first_better(thread->GetPriority());
- if (!next) {
- // Otherwise just keep going with the current thread
+ next = ready_queue.front();
+ if (next == nullptr || next->GetPriority() >= thread->GetPriority()) {
next = thread;
}
} else {
- next = ready_queue.pop_first();
+ if (ready_queue.empty()) {
+ return nullptr;
+ }
+ next = ready_queue.front();
}
return next;
}
void Scheduler::SwitchContext(Thread* new_thread) {
- Thread* const previous_thread = GetCurrentThread();
+ Thread* previous_thread = GetCurrentThread();
Process* const previous_process = system.Kernel().CurrentProcess();
UpdateLastContextSwitchTime(previous_thread, previous_process);
@@ -75,7 +80,7 @@ void Scheduler::SwitchContext(Thread* new_thread) {
if (previous_thread->GetStatus() == ThreadStatus::Running) {
// This is only the case when a reschedule is triggered without the current thread
// yielding execution (i.e. an event triggered, system core time-sliced, etc)
- ready_queue.push_front(previous_thread->GetPriority(), previous_thread);
+ ready_queue.add(previous_thread, previous_thread->GetPriority(), false);
previous_thread->SetStatus(ThreadStatus::Ready);
}
}
@@ -90,7 +95,7 @@ void Scheduler::SwitchContext(Thread* new_thread) {
current_thread = new_thread;
- ready_queue.remove(new_thread->GetPriority(), new_thread);
+ ready_queue.remove(new_thread, new_thread->GetPriority());
new_thread->SetStatus(ThreadStatus::Running);
auto* const thread_owner_process = current_thread->GetOwnerProcess();
@@ -147,7 +152,6 @@ void Scheduler::AddThread(SharedPtr<Thread> thread, u32 priority) {
std::lock_guard<std::mutex> lock(scheduler_mutex);
thread_list.push_back(std::move(thread));
- ready_queue.prepare(priority);
}
void Scheduler::RemoveThread(Thread* thread) {
@@ -161,33 +165,37 @@ void Scheduler::ScheduleThread(Thread* thread, u32 priority) {
std::lock_guard<std::mutex> lock(scheduler_mutex);
ASSERT(thread->GetStatus() == ThreadStatus::Ready);
- ready_queue.push_back(priority, thread);
+ ready_queue.add(thread, priority);
}
void Scheduler::UnscheduleThread(Thread* thread, u32 priority) {
std::lock_guard<std::mutex> lock(scheduler_mutex);
ASSERT(thread->GetStatus() == ThreadStatus::Ready);
- ready_queue.remove(priority, thread);
+ ready_queue.remove(thread, priority);
}
void Scheduler::SetThreadPriority(Thread* thread, u32 priority) {
std::lock_guard<std::mutex> lock(scheduler_mutex);
+ if (thread->GetPriority() == priority) {
+ return;
+ }
// If thread was ready, adjust queues
if (thread->GetStatus() == ThreadStatus::Ready)
- ready_queue.move(thread, thread->GetPriority(), priority);
- else
- ready_queue.prepare(priority);
+ ready_queue.adjust(thread, thread->GetPriority(), priority);
}
Thread* Scheduler::GetNextSuggestedThread(u32 core, u32 maximum_priority) const {
std::lock_guard<std::mutex> lock(scheduler_mutex);
const u32 mask = 1U << core;
- return ready_queue.get_first_filter([mask, maximum_priority](Thread const* thread) {
- return (thread->GetAffinityMask() & mask) != 0 && thread->GetPriority() < maximum_priority;
- });
+ for (auto* thread : ready_queue) {
+ if ((thread->GetAffinityMask() & mask) != 0 && thread->GetPriority() < maximum_priority) {
+ return thread;
+ }
+ }
+ return nullptr;
}
void Scheduler::YieldWithoutLoadBalancing(Thread* thread) {
diff --git a/src/core/hle/kernel/scheduler.h b/src/core/hle/kernel/scheduler.h
index 1c5bf57d9..44baeb713 100644
--- a/src/core/hle/kernel/scheduler.h
+++ b/src/core/hle/kernel/scheduler.h
@@ -7,7 +7,7 @@
#include <mutex>
#include <vector>
#include "common/common_types.h"
-#include "common/thread_queue_list.h"
+#include "common/multi_level_queue.h"
#include "core/hle/kernel/object.h"
#include "core/hle/kernel/thread.h"
@@ -156,7 +156,7 @@ private:
std::vector<SharedPtr<Thread>> thread_list;
/// Lists only ready thread ids.
- Common::ThreadQueueList<Thread*, THREADPRIO_LOWEST + 1> ready_queue;
+ Common::MultiLevelQueue<Thread*, THREADPRIO_LOWEST + 1> ready_queue;
SharedPtr<Thread> current_thread = nullptr;
diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp
index d40a2226b..11796e5e5 100644
--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -32,6 +32,7 @@
#include "core/hle/kernel/svc.h"
#include "core/hle/kernel/svc_wrap.h"
#include "core/hle/kernel/thread.h"
+#include "core/hle/kernel/transfer_memory.h"
#include "core/hle/kernel/writable_event.h"
#include "core/hle/lock.h"
#include "core/hle/result.h"
@@ -174,11 +175,8 @@ static ResultCode SetHeapSize(VAddr* heap_addr, u64 heap_size) {
return ERR_INVALID_SIZE;
}
- auto& vm_manager = Core::CurrentProcess()->VMManager();
- const VAddr heap_base = vm_manager.GetHeapRegionBaseAddress();
- const auto alloc_result =
- vm_manager.HeapAllocate(heap_base, heap_size, VMAPermission::ReadWrite);
-
+ auto& vm_manager = Core::System::GetInstance().Kernel().CurrentProcess()->VMManager();
+ const auto alloc_result = vm_manager.SetHeapSize(heap_size);
if (alloc_result.Failed()) {
return alloc_result.Code();
}
@@ -551,9 +549,9 @@ static ResultCode ArbitrateLock(Handle holding_thread_handle, VAddr mutex_addr,
return ERR_INVALID_ADDRESS;
}
- auto& handle_table = Core::CurrentProcess()->GetHandleTable();
- return Mutex::TryAcquire(handle_table, mutex_addr, holding_thread_handle,
- requesting_thread_handle);
+ auto* const current_process = Core::System::GetInstance().Kernel().CurrentProcess();
+ return current_process->GetMutex().TryAcquire(mutex_addr, holding_thread_handle,
+ requesting_thread_handle);
}
/// Unlock a mutex
@@ -571,7 +569,8 @@ static ResultCode ArbitrateUnlock(VAddr mutex_addr) {
return ERR_INVALID_ADDRESS;
}
- return Mutex::Release(mutex_addr);
+ auto* const current_process = Core::System::GetInstance().Kernel().CurrentProcess();
+ return current_process->GetMutex().Release(mutex_addr);
}
enum class BreakType : u32 {
@@ -807,7 +806,7 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id)
return RESULT_SUCCESS;
case GetInfoType::TotalHeapUsage:
- *result = process->VMManager().GetTotalHeapUsage();
+ *result = process->VMManager().GetCurrentHeapSize();
return RESULT_SUCCESS;
case GetInfoType::IsVirtualAddressMemoryEnabled:
@@ -1340,11 +1339,15 @@ static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_var
"called mutex_addr={:X}, condition_variable_addr={:X}, thread_handle=0x{:08X}, timeout={}",
mutex_addr, condition_variable_addr, thread_handle, nano_seconds);
- const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
+ auto* const current_process = Core::System::GetInstance().Kernel().CurrentProcess();
+ const auto& handle_table = current_process->GetHandleTable();
SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle);
ASSERT(thread);
- CASCADE_CODE(Mutex::Release(mutex_addr));
+ const auto release_result = current_process->GetMutex().Release(mutex_addr);
+ if (release_result.IsError()) {
+ return release_result;
+ }
SharedPtr<Thread> current_thread = GetCurrentThread();
current_thread->SetCondVarWaitAddress(condition_variable_addr);
@@ -1582,14 +1585,121 @@ static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32
}
auto& kernel = Core::System::GetInstance().Kernel();
- auto process = kernel.CurrentProcess();
- auto& handle_table = process->GetHandleTable();
- const auto shared_mem_handle = SharedMemory::Create(kernel, process, size, perms, perms, addr);
+ auto transfer_mem_handle = TransferMemory::Create(kernel, addr, size, perms);
- CASCADE_RESULT(*handle, handle_table.Create(shared_mem_handle));
+ auto& handle_table = kernel.CurrentProcess()->GetHandleTable();
+ const auto result = handle_table.Create(std::move(transfer_mem_handle));
+ if (result.Failed()) {
+ return result.Code();
+ }
+
+ *handle = *result;
return RESULT_SUCCESS;
}
+static ResultCode MapTransferMemory(Handle handle, VAddr address, u64 size, u32 permission_raw) {
+ LOG_DEBUG(Kernel_SVC,
+ "called. handle=0x{:08X}, address=0x{:016X}, size=0x{:016X}, permissions=0x{:08X}",
+ handle, address, size, permission_raw);
+
+ if (!Common::Is4KBAligned(address)) {
+ LOG_ERROR(Kernel_SVC, "Transfer memory addresses must be 4KB aligned (size=0x{:016X}).",
+ address);
+ return ERR_INVALID_ADDRESS;
+ }
+
+ if (size == 0 || !Common::Is4KBAligned(size)) {
+ LOG_ERROR(Kernel_SVC,
+ "Transfer memory sizes must be 4KB aligned and not be zero (size=0x{:016X}).",
+ size);
+ return ERR_INVALID_SIZE;
+ }
+
+ if (!IsValidAddressRange(address, size)) {
+ LOG_ERROR(Kernel_SVC,
+ "Given address and size overflows the 64-bit range (address=0x{:016X}, "
+ "size=0x{:016X}).",
+ address, size);
+ return ERR_INVALID_ADDRESS_STATE;
+ }
+
+ const auto permissions = static_cast<MemoryPermission>(permission_raw);
+ if (permissions != MemoryPermission::None && permissions != MemoryPermission::Read &&
+ permissions != MemoryPermission::ReadWrite) {
+ LOG_ERROR(Kernel_SVC, "Invalid transfer memory permissions given (permissions=0x{:08X}).",
+ permission_raw);
+ return ERR_INVALID_STATE;
+ }
+
+ const auto& kernel = Core::System::GetInstance().Kernel();
+ const auto* const current_process = kernel.CurrentProcess();
+ const auto& handle_table = current_process->GetHandleTable();
+
+ auto transfer_memory = handle_table.Get<TransferMemory>(handle);
+ if (!transfer_memory) {
+ LOG_ERROR(Kernel_SVC, "Nonexistent transfer memory handle given (handle=0x{:08X}).",
+ handle);
+ return ERR_INVALID_HANDLE;
+ }
+
+ if (!current_process->VMManager().IsWithinASLRRegion(address, size)) {
+ LOG_ERROR(Kernel_SVC,
+ "Given address and size don't fully fit within the ASLR region "
+ "(address=0x{:016X}, size=0x{:016X}).",
+ address, size);
+ return ERR_INVALID_MEMORY_RANGE;
+ }
+
+ return transfer_memory->MapMemory(address, size, permissions);
+}
+
+static ResultCode UnmapTransferMemory(Handle handle, VAddr address, u64 size) {
+ LOG_DEBUG(Kernel_SVC, "called. handle=0x{:08X}, address=0x{:016X}, size=0x{:016X}", handle,
+ address, size);
+
+ if (!Common::Is4KBAligned(address)) {
+ LOG_ERROR(Kernel_SVC, "Transfer memory addresses must be 4KB aligned (size=0x{:016X}).",
+ address);
+ return ERR_INVALID_ADDRESS;
+ }
+
+ if (size == 0 || !Common::Is4KBAligned(size)) {
+ LOG_ERROR(Kernel_SVC,
+ "Transfer memory sizes must be 4KB aligned and not be zero (size=0x{:016X}).",
+ size);
+ return ERR_INVALID_SIZE;
+ }
+
+ if (!IsValidAddressRange(address, size)) {
+ LOG_ERROR(Kernel_SVC,
+ "Given address and size overflows the 64-bit range (address=0x{:016X}, "
+ "size=0x{:016X}).",
+ address, size);
+ return ERR_INVALID_ADDRESS_STATE;
+ }
+
+ const auto& kernel = Core::System::GetInstance().Kernel();
+ const auto* const current_process = kernel.CurrentProcess();
+ const auto& handle_table = current_process->GetHandleTable();
+
+ auto transfer_memory = handle_table.Get<TransferMemory>(handle);
+ if (!transfer_memory) {
+ LOG_ERROR(Kernel_SVC, "Nonexistent transfer memory handle given (handle=0x{:08X}).",
+ handle);
+ return ERR_INVALID_HANDLE;
+ }
+
+ if (!current_process->VMManager().IsWithinASLRRegion(address, size)) {
+ LOG_ERROR(Kernel_SVC,
+ "Given address and size don't fully fit within the ASLR region "
+ "(address=0x{:016X}, size=0x{:016X}).",
+ address, size);
+ return ERR_INVALID_MEMORY_RANGE;
+ }
+
+ return transfer_memory->UnmapMemory(address, size);
+}
+
static ResultCode GetThreadCoreMask(Handle thread_handle, u32* core, u64* mask) {
LOG_TRACE(Kernel_SVC, "called, handle=0x{:08X}", thread_handle);
@@ -1965,8 +2075,8 @@ static const FunctionDef SVC_Table[] = {
{0x4E, nullptr, "ReadWriteRegister"},
{0x4F, nullptr, "SetProcessActivity"},
{0x50, SvcWrap<CreateSharedMemory>, "CreateSharedMemory"},
- {0x51, nullptr, "MapTransferMemory"},
- {0x52, nullptr, "UnmapTransferMemory"},
+ {0x51, SvcWrap<MapTransferMemory>, "MapTransferMemory"},
+ {0x52, SvcWrap<UnmapTransferMemory>, "UnmapTransferMemory"},
{0x53, nullptr, "CreateInterruptEvent"},
{0x54, nullptr, "QueryPhysicalAddress"},
{0x55, nullptr, "QueryIoMapping"},
diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp
index 79c80bb01..e5853c46f 100644
--- a/src/core/hle/kernel/thread.cpp
+++ b/src/core/hle/kernel/thread.cpp
@@ -315,8 +315,9 @@ void Thread::UpdatePriority() {
}
// Ensure that the thread is within the correct location in the waiting list.
+ auto old_owner = lock_owner;
lock_owner->RemoveMutexWaiter(this);
- lock_owner->AddMutexWaiter(this);
+ old_owner->AddMutexWaiter(this);
// Recursively update the priority of the thread that depends on the priority of this one.
lock_owner->UpdatePriority();
diff --git a/src/core/hle/kernel/transfer_memory.cpp b/src/core/hle/kernel/transfer_memory.cpp
new file mode 100644
index 000000000..23228e1b5
--- /dev/null
+++ b/src/core/hle/kernel/transfer_memory.cpp
@@ -0,0 +1,73 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/hle/kernel/errors.h"
+#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/process.h"
+#include "core/hle/kernel/shared_memory.h"
+#include "core/hle/kernel/transfer_memory.h"
+#include "core/hle/result.h"
+
+namespace Kernel {
+
+TransferMemory::TransferMemory(KernelCore& kernel) : Object{kernel} {}
+TransferMemory::~TransferMemory() = default;
+
+SharedPtr<TransferMemory> TransferMemory::Create(KernelCore& kernel, VAddr base_address,
+ size_t size, MemoryPermission permissions) {
+ SharedPtr<TransferMemory> transfer_memory{new TransferMemory(kernel)};
+
+ transfer_memory->base_address = base_address;
+ transfer_memory->memory_size = size;
+ transfer_memory->owner_permissions = permissions;
+ transfer_memory->owner_process = kernel.CurrentProcess();
+
+ return transfer_memory;
+}
+
+ResultCode TransferMemory::MapMemory(VAddr address, size_t size, MemoryPermission permissions) {
+ if (memory_size != size) {
+ return ERR_INVALID_SIZE;
+ }
+
+ if (owner_permissions != permissions) {
+ return ERR_INVALID_STATE;
+ }
+
+ if (is_mapped) {
+ return ERR_INVALID_STATE;
+ }
+
+ const auto map_state = owner_permissions == MemoryPermission::None
+ ? MemoryState::TransferMemoryIsolated
+ : MemoryState::TransferMemory;
+ auto& vm_manager = owner_process->VMManager();
+ const auto map_result = vm_manager.MapMemoryBlock(
+ address, std::make_shared<std::vector<u8>>(size), 0, size, map_state);
+
+ if (map_result.Failed()) {
+ return map_result.Code();
+ }
+
+ is_mapped = true;
+ return RESULT_SUCCESS;
+}
+
+ResultCode TransferMemory::UnmapMemory(VAddr address, size_t size) {
+ if (memory_size != size) {
+ return ERR_INVALID_SIZE;
+ }
+
+ auto& vm_manager = owner_process->VMManager();
+ const auto result = vm_manager.UnmapRange(address, size);
+
+ if (result.IsError()) {
+ return result;
+ }
+
+ is_mapped = false;
+ return RESULT_SUCCESS;
+}
+
+} // namespace Kernel
diff --git a/src/core/hle/kernel/transfer_memory.h b/src/core/hle/kernel/transfer_memory.h
new file mode 100644
index 000000000..ec294951e
--- /dev/null
+++ b/src/core/hle/kernel/transfer_memory.h
@@ -0,0 +1,91 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/kernel/object.h"
+
+union ResultCode;
+
+namespace Kernel {
+
+class KernelCore;
+class Process;
+
+enum class MemoryPermission : u32;
+
+/// Defines the interface for transfer memory objects.
+///
+/// Transfer memory is typically used for the purpose of
+/// transferring memory between separate process instances,
+/// thus the name.
+///
+class TransferMemory final : public Object {
+public:
+ static constexpr HandleType HANDLE_TYPE = HandleType::TransferMemory;
+
+ static SharedPtr<TransferMemory> Create(KernelCore& kernel, VAddr base_address, size_t size,
+ MemoryPermission permissions);
+
+ TransferMemory(const TransferMemory&) = delete;
+ TransferMemory& operator=(const TransferMemory&) = delete;
+
+ TransferMemory(TransferMemory&&) = delete;
+ TransferMemory& operator=(TransferMemory&&) = delete;
+
+ std::string GetTypeName() const override {
+ return "TransferMemory";
+ }
+
+ std::string GetName() const override {
+ return GetTypeName();
+ }
+
+ HandleType GetHandleType() const override {
+ return HANDLE_TYPE;
+ }
+
+ /// Attempts to map transfer memory with the given range and memory permissions.
+ ///
+ /// @param address The base address to being mapping memory at.
+ /// @param size The size of the memory to map, in bytes.
+ /// @param permissions The memory permissions to check against when mapping memory.
+ ///
+ /// @pre The given address, size, and memory permissions must all match
+ /// the same values that were given when creating the transfer memory
+ /// instance.
+ ///
+ ResultCode MapMemory(VAddr address, size_t size, MemoryPermission permissions);
+
+ /// Unmaps the transfer memory with the given range
+ ///
+ /// @param address The base address to begin unmapping memory at.
+ /// @param size The size of the memory to unmap, in bytes.
+ ///
+ /// @pre The given address and size must be the same as the ones used
+ /// to create the transfer memory instance.
+ ///
+ ResultCode UnmapMemory(VAddr address, size_t size);
+
+private:
+ explicit TransferMemory(KernelCore& kernel);
+ ~TransferMemory() override;
+
+ /// The base address for the memory managed by this instance.
+ VAddr base_address = 0;
+
+ /// Size of the memory, in bytes, that this instance manages.
+ size_t memory_size = 0;
+
+ /// The memory permissions that are applied to this instance.
+ MemoryPermission owner_permissions{};
+
+ /// The process that this transfer memory instance was created under.
+ Process* owner_process = nullptr;
+
+ /// Whether or not this transfer memory instance has mapped memory.
+ bool is_mapped = false;
+};
+
+} // namespace Kernel
diff --git a/src/core/hle/kernel/vm_manager.cpp b/src/core/hle/kernel/vm_manager.cpp
index 3def3e52c..ec0a480ce 100644
--- a/src/core/hle/kernel/vm_manager.cpp
+++ b/src/core/hle/kernel/vm_manager.cpp
@@ -20,16 +20,16 @@ namespace Kernel {
namespace {
const char* GetMemoryStateName(MemoryState state) {
static constexpr const char* names[] = {
- "Unmapped", "Io",
- "Normal", "CodeStatic",
- "CodeMutable", "Heap",
- "Shared", "Unknown1",
- "ModuleCodeStatic", "ModuleCodeMutable",
- "IpcBuffer0", "Stack",
- "ThreadLocal", "TransferMemoryIsolated",
- "TransferMemory", "ProcessMemory",
- "Inaccessible", "IpcBuffer1",
- "IpcBuffer3", "KernelStack",
+ "Unmapped", "Io",
+ "Normal", "Code",
+ "CodeData", "Heap",
+ "Shared", "Unknown1",
+ "ModuleCode", "ModuleCodeData",
+ "IpcBuffer0", "Stack",
+ "ThreadLocal", "TransferMemoryIsolated",
+ "TransferMemory", "ProcessMemory",
+ "Inaccessible", "IpcBuffer1",
+ "IpcBuffer3", "KernelStack",
};
return names[ToSvcMemoryState(state)];
@@ -256,57 +256,50 @@ ResultCode VMManager::ReprotectRange(VAddr target, u64 size, VMAPermission new_p
return RESULT_SUCCESS;
}
-ResultVal<VAddr> VMManager::HeapAllocate(VAddr target, u64 size, VMAPermission perms) {
- if (!IsWithinHeapRegion(target, size)) {
- return ERR_INVALID_ADDRESS;
+ResultVal<VAddr> VMManager::SetHeapSize(u64 size) {
+ if (size > GetHeapRegionSize()) {
+ return ERR_OUT_OF_MEMORY;
+ }
+
+ // No need to do any additional work if the heap is already the given size.
+ if (size == GetCurrentHeapSize()) {
+ return MakeResult(heap_region_base);
}
if (heap_memory == nullptr) {
// Initialize heap
- heap_memory = std::make_shared<std::vector<u8>>();
- heap_start = heap_end = target;
+ heap_memory = std::make_shared<std::vector<u8>>(size);
+ heap_end = heap_region_base + size;
} else {
- UnmapRange(heap_start, heap_end - heap_start);
- }
-
- // If necessary, expand backing vector to cover new heap extents.
- if (target < heap_start) {
- heap_memory->insert(begin(*heap_memory), heap_start - target, 0);
- heap_start = target;
- RefreshMemoryBlockMappings(heap_memory.get());
- }
- if (target + size > heap_end) {
- heap_memory->insert(end(*heap_memory), (target + size) - heap_end, 0);
- heap_end = target + size;
- RefreshMemoryBlockMappings(heap_memory.get());
+ UnmapRange(heap_region_base, GetCurrentHeapSize());
}
- ASSERT(heap_end - heap_start == heap_memory->size());
- CASCADE_RESULT(auto vma, MapMemoryBlock(target, heap_memory, target - heap_start, size,
- MemoryState::Heap));
- Reprotect(vma, perms);
+ // If necessary, expand backing vector to cover new heap extents in
+ // the case of allocating. Otherwise, shrink the backing memory,
+ // if a smaller heap has been requested.
+ const u64 old_heap_size = GetCurrentHeapSize();
+ if (size > old_heap_size) {
+ const u64 alloc_size = size - old_heap_size;
- heap_used = size;
-
- return MakeResult<VAddr>(heap_end - size);
-}
+ heap_memory->insert(heap_memory->end(), alloc_size, 0);
+ RefreshMemoryBlockMappings(heap_memory.get());
+ } else if (size < old_heap_size) {
+ heap_memory->resize(size);
+ heap_memory->shrink_to_fit();
-ResultCode VMManager::HeapFree(VAddr target, u64 size) {
- if (!IsWithinHeapRegion(target, size)) {
- return ERR_INVALID_ADDRESS;
+ RefreshMemoryBlockMappings(heap_memory.get());
}
- if (size == 0) {
- return RESULT_SUCCESS;
- }
+ heap_end = heap_region_base + size;
+ ASSERT(GetCurrentHeapSize() == heap_memory->size());
- const ResultCode result = UnmapRange(target, size);
- if (result.IsError()) {
- return result;
+ const auto mapping_result =
+ MapMemoryBlock(heap_region_base, heap_memory, 0, size, MemoryState::Heap);
+ if (mapping_result.Failed()) {
+ return mapping_result.Code();
}
- heap_used -= size;
- return RESULT_SUCCESS;
+ return MakeResult<VAddr>(heap_region_base);
}
MemoryInfo VMManager::QueryMemory(VAddr address) const {
@@ -598,6 +591,7 @@ void VMManager::InitializeMemoryRegionRanges(FileSys::ProgramAddressSpaceType ty
heap_region_base = map_region_end;
heap_region_end = heap_region_base + heap_region_size;
+ heap_end = heap_region_base;
new_map_region_base = heap_region_end;
new_map_region_end = new_map_region_base + new_map_region_size;
@@ -692,10 +686,6 @@ u64 VMManager::GetTotalMemoryUsage() const {
return 0xF8000000;
}
-u64 VMManager::GetTotalHeapUsage() const {
- return heap_used;
-}
-
VAddr VMManager::GetAddressSpaceBaseAddress() const {
return address_space_base;
}
@@ -778,6 +768,10 @@ u64 VMManager::GetHeapRegionSize() const {
return heap_region_end - heap_region_base;
}
+u64 VMManager::GetCurrentHeapSize() const {
+ return heap_end - heap_region_base;
+}
+
bool VMManager::IsWithinHeapRegion(VAddr address, u64 size) const {
return IsInsideAddressRange(address, size, GetHeapRegionBaseAddress(),
GetHeapRegionEndAddress());
diff --git a/src/core/hle/kernel/vm_manager.h b/src/core/hle/kernel/vm_manager.h
index b96980f8f..6f484b7bf 100644
--- a/src/core/hle/kernel/vm_manager.h
+++ b/src/core/hle/kernel/vm_manager.h
@@ -165,12 +165,12 @@ enum class MemoryState : u32 {
Unmapped = 0x00,
Io = 0x01 | FlagMapped,
Normal = 0x02 | FlagMapped | FlagQueryPhysicalAddressAllowed,
- CodeStatic = 0x03 | CodeFlags | FlagMapProcess,
- CodeMutable = 0x04 | CodeFlags | FlagMapProcess | FlagCodeMemory,
+ Code = 0x03 | CodeFlags | FlagMapProcess,
+ CodeData = 0x04 | DataFlags | FlagMapProcess | FlagCodeMemory,
Heap = 0x05 | DataFlags | FlagCodeMemory,
Shared = 0x06 | FlagMapped | FlagMemoryPoolAllocated,
- ModuleCodeStatic = 0x08 | CodeFlags | FlagModule | FlagMapProcess,
- ModuleCodeMutable = 0x09 | DataFlags | FlagModule | FlagMapProcess | FlagCodeMemory,
+ ModuleCode = 0x08 | CodeFlags | FlagModule | FlagMapProcess,
+ ModuleCodeData = 0x09 | DataFlags | FlagModule | FlagMapProcess | FlagCodeMemory,
IpcBuffer0 = 0x0A | FlagMapped | FlagQueryPhysicalAddressAllowed | FlagMemoryPoolAllocated |
IPCFlags | FlagSharedDevice | FlagSharedDeviceAligned,
@@ -380,11 +380,41 @@ public:
/// Changes the permissions of a range of addresses, splitting VMAs as necessary.
ResultCode ReprotectRange(VAddr target, u64 size, VMAPermission new_perms);
- ResultVal<VAddr> HeapAllocate(VAddr target, u64 size, VMAPermission perms);
- ResultCode HeapFree(VAddr target, u64 size);
-
ResultCode MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size, MemoryState state);
+ /// Attempts to allocate a heap with the given size.
+ ///
+ /// @param size The size of the heap to allocate in bytes.
+ ///
+ /// @note If a heap is currently allocated, and this is called
+ /// with a size that is equal to the size of the current heap,
+ /// then this function will do nothing and return the current
+ /// heap's starting address, as there's no need to perform
+ /// any additional heap allocation work.
+ ///
+ /// @note If a heap is currently allocated, and this is called
+ /// with a size less than the current heap's size, then
+ /// this function will attempt to shrink the heap.
+ ///
+ /// @note If a heap is currently allocated, and this is called
+ /// with a size larger than the current heap's size, then
+ /// this function will attempt to extend the size of the heap.
+ ///
+ /// @returns A result indicating either success or failure.
+ /// <p>
+ /// If successful, this function will return a result
+ /// containing the starting address to the allocated heap.
+ /// <p>
+ /// If unsuccessful, this function will return a result
+ /// containing an error code.
+ ///
+ /// @pre The given size must lie within the allowable heap
+ /// memory region managed by this VMManager instance.
+ /// Failure to abide by this will result in ERR_OUT_OF_MEMORY
+ /// being returned as the result.
+ ///
+ ResultVal<VAddr> SetHeapSize(u64 size);
+
/// Queries the memory manager for information about the given address.
///
/// @param address The address to query the memory manager about for information.
@@ -418,9 +448,6 @@ public:
/// Gets the total memory usage, used by svcGetInfo
u64 GetTotalMemoryUsage() const;
- /// Gets the total heap usage, used by svcGetInfo
- u64 GetTotalHeapUsage() const;
-
/// Gets the address space base address
VAddr GetAddressSpaceBaseAddress() const;
@@ -469,6 +496,13 @@ public:
/// Gets the total size of the heap region in bytes.
u64 GetHeapRegionSize() const;
+ /// Gets the total size of the current heap in bytes.
+ ///
+ /// @note This is the current allocated heap size, not the size
+ /// of the region it's allowed to exist within.
+ ///
+ u64 GetCurrentHeapSize() const;
+
/// Determines whether or not the specified range is within the heap region.
bool IsWithinHeapRegion(VAddr address, u64 size) const;
@@ -625,9 +659,9 @@ private:
// This makes deallocation and reallocation of holes fast and keeps process memory contiguous
// in the emulator address space, allowing Memory::GetPointer to be reasonably safe.
std::shared_ptr<std::vector<u8>> heap_memory;
- // The left/right bounds of the address space covered by heap_memory.
- VAddr heap_start = 0;
+
+ // The end of the currently allocated heap. This is not an inclusive
+ // end of the range. This is essentially 'base_address + current_size'.
VAddr heap_end = 0;
- u64 heap_used = 0;
};
} // namespace Kernel
diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp
index c750d70ac..9c44e27c6 100644
--- a/src/core/hle/service/am/am.cpp
+++ b/src/core/hle/service/am/am.cpp
@@ -215,7 +215,21 @@ IDisplayController::IDisplayController() : ServiceFramework("IDisplayController"
IDisplayController::~IDisplayController() = default;
-IDebugFunctions::IDebugFunctions() : ServiceFramework("IDebugFunctions") {}
+IDebugFunctions::IDebugFunctions() : ServiceFramework{"IDebugFunctions"} {
+ // clang-format off
+ static const FunctionInfo functions[] = {
+ {0, nullptr, "NotifyMessageToHomeMenuForDebug"},
+ {1, nullptr, "OpenMainApplication"},
+ {10, nullptr, "EmulateButtonEvent"},
+ {20, nullptr, "InvalidateTransitionLayer"},
+ {30, nullptr, "RequestLaunchApplicationWithUserAndArgumentForDebug"},
+ {40, nullptr, "GetAppletResourceUsageInfo"},
+ };
+ // clang-format on
+
+ RegisterHandlers(functions);
+}
+
IDebugFunctions::~IDebugFunctions() = default;
ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger)
diff --git a/src/core/hle/service/audio/hwopus.cpp b/src/core/hle/service/audio/hwopus.cpp
index 377e12cfa..cb4a1160d 100644
--- a/src/core/hle/service/audio/hwopus.cpp
+++ b/src/core/hle/service/audio/hwopus.cpp
@@ -8,6 +8,7 @@
#include <vector>
#include <opus.h>
+#include <opus_multistream.h>
#include "common/assert.h"
#include "common/logging/log.h"
@@ -18,12 +19,12 @@
namespace Service::Audio {
namespace {
struct OpusDeleter {
- void operator()(void* ptr) const {
- operator delete(ptr);
+ void operator()(OpusMSDecoder* ptr) const {
+ opus_multistream_decoder_destroy(ptr);
}
};
-using OpusDecoderPtr = std::unique_ptr<OpusDecoder, OpusDeleter>;
+using OpusDecoderPtr = std::unique_ptr<OpusMSDecoder, OpusDeleter>;
struct OpusPacketHeader {
// Packet size in bytes.
@@ -33,7 +34,7 @@ struct OpusPacketHeader {
};
static_assert(sizeof(OpusPacketHeader) == 0x8, "OpusHeader is an invalid size");
-class OpusDecoderStateBase {
+class OpusDecoderState {
public:
/// Describes extra behavior that may be asked of the decoding context.
enum class ExtraBehavior {
@@ -49,22 +50,13 @@ public:
Enabled,
};
- virtual ~OpusDecoderStateBase() = default;
-
- // Decodes interleaved Opus packets. Optionally allows reporting time taken to
- // perform the decoding, as well as any relevant extra behavior.
- virtual void DecodeInterleaved(Kernel::HLERequestContext& ctx, PerfTime perf_time,
- ExtraBehavior extra_behavior) = 0;
-};
-
-// Represents the decoder state for a non-multistream decoder.
-class OpusDecoderState final : public OpusDecoderStateBase {
-public:
explicit OpusDecoderState(OpusDecoderPtr decoder, u32 sample_rate, u32 channel_count)
: decoder{std::move(decoder)}, sample_rate{sample_rate}, channel_count{channel_count} {}
+ // Decodes interleaved Opus packets. Optionally allows reporting time taken to
+ // perform the decoding, as well as any relevant extra behavior.
void DecodeInterleaved(Kernel::HLERequestContext& ctx, PerfTime perf_time,
- ExtraBehavior extra_behavior) override {
+ ExtraBehavior extra_behavior) {
if (perf_time == PerfTime::Disabled) {
DecodeInterleavedHelper(ctx, nullptr, extra_behavior);
} else {
@@ -135,7 +127,7 @@ private:
const int frame_size = (static_cast<int>(raw_output_sz / sizeof(s16) / channel_count));
const auto out_sample_count =
- opus_decode(decoder.get(), frame, hdr.size, output.data(), frame_size, 0);
+ opus_multistream_decode(decoder.get(), frame, hdr.size, output.data(), frame_size, 0);
if (out_sample_count < 0) {
LOG_ERROR(Audio,
"Incorrect sample count received from opus_decode, "
@@ -158,7 +150,7 @@ private:
void ResetDecoderContext() {
ASSERT(decoder != nullptr);
- opus_decoder_ctl(decoder.get(), OPUS_RESET_STATE);
+ opus_multistream_decoder_ctl(decoder.get(), OPUS_RESET_STATE);
}
OpusDecoderPtr decoder;
@@ -168,7 +160,7 @@ private:
class IHardwareOpusDecoderManager final : public ServiceFramework<IHardwareOpusDecoderManager> {
public:
- explicit IHardwareOpusDecoderManager(std::unique_ptr<OpusDecoderStateBase> decoder_state)
+ explicit IHardwareOpusDecoderManager(OpusDecoderState decoder_state)
: ServiceFramework("IHardwareOpusDecoderManager"), decoder_state{std::move(decoder_state)} {
// clang-format off
static const FunctionInfo functions[] = {
@@ -190,35 +182,51 @@ private:
void DecodeInterleavedOld(Kernel::HLERequestContext& ctx) {
LOG_DEBUG(Audio, "called");
- decoder_state->DecodeInterleaved(ctx, OpusDecoderStateBase::PerfTime::Disabled,
- OpusDecoderStateBase::ExtraBehavior::None);
+ decoder_state.DecodeInterleaved(ctx, OpusDecoderState::PerfTime::Disabled,
+ OpusDecoderState::ExtraBehavior::None);
}
void DecodeInterleavedWithPerfOld(Kernel::HLERequestContext& ctx) {
LOG_DEBUG(Audio, "called");
- decoder_state->DecodeInterleaved(ctx, OpusDecoderStateBase::PerfTime::Enabled,
- OpusDecoderStateBase::ExtraBehavior::None);
+ decoder_state.DecodeInterleaved(ctx, OpusDecoderState::PerfTime::Enabled,
+ OpusDecoderState::ExtraBehavior::None);
}
void DecodeInterleaved(Kernel::HLERequestContext& ctx) {
LOG_DEBUG(Audio, "called");
IPC::RequestParser rp{ctx};
- const auto extra_behavior = rp.Pop<bool>()
- ? OpusDecoderStateBase::ExtraBehavior::ResetContext
- : OpusDecoderStateBase::ExtraBehavior::None;
+ const auto extra_behavior = rp.Pop<bool>() ? OpusDecoderState::ExtraBehavior::ResetContext
+ : OpusDecoderState::ExtraBehavior::None;
- decoder_state->DecodeInterleaved(ctx, OpusDecoderStateBase::PerfTime::Enabled,
- extra_behavior);
+ decoder_state.DecodeInterleaved(ctx, OpusDecoderState::PerfTime::Enabled, extra_behavior);
}
- std::unique_ptr<OpusDecoderStateBase> decoder_state;
+ OpusDecoderState decoder_state;
};
std::size_t WorkerBufferSize(u32 channel_count) {
ASSERT_MSG(channel_count == 1 || channel_count == 2, "Invalid channel count");
- return opus_decoder_get_size(static_cast<int>(channel_count));
+ constexpr int num_streams = 1;
+ const int num_stereo_streams = channel_count == 2 ? 1 : 0;
+ return opus_multistream_decoder_get_size(num_streams, num_stereo_streams);
+}
+
+// Creates the mapping table that maps the input channels to the particular
+// output channels. In the stereo case, we map the left and right input channels
+// to the left and right output channels respectively.
+//
+// However, in the monophonic case, we only map the one available channel
+// to the sole output channel. We specify 255 for the would-be right channel
+// as this is a special value defined by Opus to indicate to the decoder to
+// ignore that channel.
+std::array<u8, 2> CreateMappingTable(u32 channel_count) {
+ if (channel_count == 2) {
+ return {{0, 1}};
+ }
+
+ return {{0, 255}};
}
} // Anonymous namespace
@@ -259,9 +267,15 @@ void HwOpus::OpenOpusDecoder(Kernel::HLERequestContext& ctx) {
const std::size_t worker_sz = WorkerBufferSize(channel_count);
ASSERT_MSG(buffer_sz >= worker_sz, "Worker buffer too large");
- OpusDecoderPtr decoder{static_cast<OpusDecoder*>(operator new(worker_sz))};
- if (const int err = opus_decoder_init(decoder.get(), sample_rate, channel_count)) {
- LOG_ERROR(Audio, "Failed to init opus decoder with error={}", err);
+ const int num_stereo_streams = channel_count == 2 ? 1 : 0;
+ const auto mapping_table = CreateMappingTable(channel_count);
+
+ int error = 0;
+ OpusDecoderPtr decoder{
+ opus_multistream_decoder_create(sample_rate, static_cast<int>(channel_count), 1,
+ num_stereo_streams, mapping_table.data(), &error)};
+ if (error != OPUS_OK || decoder == nullptr) {
+ LOG_ERROR(Audio, "Failed to create Opus decoder (error={}).", error);
IPC::ResponseBuilder rb{ctx, 2};
// TODO(ogniK): Use correct error code
rb.Push(ResultCode(-1));
@@ -271,7 +285,7 @@ void HwOpus::OpenOpusDecoder(Kernel::HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(RESULT_SUCCESS);
rb.PushIpcInterface<IHardwareOpusDecoderManager>(
- std::make_unique<OpusDecoderState>(std::move(decoder), sample_rate, channel_count));
+ OpusDecoderState{std::move(decoder), sample_rate, channel_count});
}
HwOpus::HwOpus() : ServiceFramework("hwopus") {
diff --git a/src/core/hle/service/hid/controllers/debug_pad.h b/src/core/hle/service/hid/controllers/debug_pad.h
index 929035034..e584b92ec 100644
--- a/src/core/hle/service/hid/controllers/debug_pad.h
+++ b/src/core/hle/service/hid/controllers/debug_pad.h
@@ -41,20 +41,20 @@ private:
struct PadState {
union {
u32_le raw{};
- BitField<0, 1, u32_le> a;
- BitField<1, 1, u32_le> b;
- BitField<2, 1, u32_le> x;
- BitField<3, 1, u32_le> y;
- BitField<4, 1, u32_le> l;
- BitField<5, 1, u32_le> r;
- BitField<6, 1, u32_le> zl;
- BitField<7, 1, u32_le> zr;
- BitField<8, 1, u32_le> plus;
- BitField<9, 1, u32_le> minus;
- BitField<10, 1, u32_le> d_left;
- BitField<11, 1, u32_le> d_up;
- BitField<12, 1, u32_le> d_right;
- BitField<13, 1, u32_le> d_down;
+ BitField<0, 1, u32> a;
+ BitField<1, 1, u32> b;
+ BitField<2, 1, u32> x;
+ BitField<3, 1, u32> y;
+ BitField<4, 1, u32> l;
+ BitField<5, 1, u32> r;
+ BitField<6, 1, u32> zl;
+ BitField<7, 1, u32> zr;
+ BitField<8, 1, u32> plus;
+ BitField<9, 1, u32> minus;
+ BitField<10, 1, u32> d_left;
+ BitField<11, 1, u32> d_up;
+ BitField<12, 1, u32> d_right;
+ BitField<13, 1, u32> d_down;
};
};
static_assert(sizeof(PadState) == 0x4, "PadState is an invalid size");
@@ -62,7 +62,7 @@ private:
struct Attributes {
union {
u32_le raw{};
- BitField<0, 1, u32_le> connected;
+ BitField<0, 1, u32> connected;
};
};
static_assert(sizeof(Attributes) == 0x4, "Attributes is an invalid size");
diff --git a/src/core/hle/service/hid/controllers/npad.h b/src/core/hle/service/hid/controllers/npad.h
index 18c7a94e6..4ff50b3cd 100644
--- a/src/core/hle/service/hid/controllers/npad.h
+++ b/src/core/hle/service/hid/controllers/npad.h
@@ -39,13 +39,13 @@ public:
union {
u32_le raw{};
- BitField<0, 1, u32_le> pro_controller;
- BitField<1, 1, u32_le> handheld;
- BitField<2, 1, u32_le> joycon_dual;
- BitField<3, 1, u32_le> joycon_left;
- BitField<4, 1, u32_le> joycon_right;
+ BitField<0, 1, u32> pro_controller;
+ BitField<1, 1, u32> handheld;
+ BitField<2, 1, u32> joycon_dual;
+ BitField<3, 1, u32> joycon_left;
+ BitField<4, 1, u32> joycon_right;
- BitField<6, 1, u32_le> pokeball; // TODO(ogniK): Confirm when possible
+ BitField<6, 1, u32> pokeball; // TODO(ogniK): Confirm when possible
};
};
static_assert(sizeof(NPadType) == 4, "NPadType is an invalid size");
@@ -150,43 +150,43 @@ private:
union {
u64_le raw{};
// Button states
- BitField<0, 1, u64_le> a;
- BitField<1, 1, u64_le> b;
- BitField<2, 1, u64_le> x;
- BitField<3, 1, u64_le> y;
- BitField<4, 1, u64_le> l_stick;
- BitField<5, 1, u64_le> r_stick;
- BitField<6, 1, u64_le> l;
- BitField<7, 1, u64_le> r;
- BitField<8, 1, u64_le> zl;
- BitField<9, 1, u64_le> zr;
- BitField<10, 1, u64_le> plus;
- BitField<11, 1, u64_le> minus;
+ BitField<0, 1, u64> a;
+ BitField<1, 1, u64> b;
+ BitField<2, 1, u64> x;
+ BitField<3, 1, u64> y;
+ BitField<4, 1, u64> l_stick;
+ BitField<5, 1, u64> r_stick;
+ BitField<6, 1, u64> l;
+ BitField<7, 1, u64> r;
+ BitField<8, 1, u64> zl;
+ BitField<9, 1, u64> zr;
+ BitField<10, 1, u64> plus;
+ BitField<11, 1, u64> minus;
// D-Pad
- BitField<12, 1, u64_le> d_left;
- BitField<13, 1, u64_le> d_up;
- BitField<14, 1, u64_le> d_right;
- BitField<15, 1, u64_le> d_down;
+ BitField<12, 1, u64> d_left;
+ BitField<13, 1, u64> d_up;
+ BitField<14, 1, u64> d_right;
+ BitField<15, 1, u64> d_down;
// Left JoyStick
- BitField<16, 1, u64_le> l_stick_left;
- BitField<17, 1, u64_le> l_stick_up;
- BitField<18, 1, u64_le> l_stick_right;
- BitField<19, 1, u64_le> l_stick_down;
+ BitField<16, 1, u64> l_stick_left;
+ BitField<17, 1, u64> l_stick_up;
+ BitField<18, 1, u64> l_stick_right;
+ BitField<19, 1, u64> l_stick_down;
// Right JoyStick
- BitField<20, 1, u64_le> r_stick_left;
- BitField<21, 1, u64_le> r_stick_up;
- BitField<22, 1, u64_le> r_stick_right;
- BitField<23, 1, u64_le> r_stick_down;
+ BitField<20, 1, u64> r_stick_left;
+ BitField<21, 1, u64> r_stick_up;
+ BitField<22, 1, u64> r_stick_right;
+ BitField<23, 1, u64> r_stick_down;
// Not always active?
- BitField<24, 1, u64_le> left_sl;
- BitField<25, 1, u64_le> left_sr;
+ BitField<24, 1, u64> left_sl;
+ BitField<25, 1, u64> left_sr;
- BitField<26, 1, u64_le> right_sl;
- BitField<27, 1, u64_le> right_sr;
+ BitField<26, 1, u64> right_sl;
+ BitField<27, 1, u64> right_sr;
};
};
static_assert(sizeof(ControllerPadState) == 8, "ControllerPadState is an invalid size");
@@ -200,12 +200,12 @@ private:
struct ConnectionState {
union {
u32_le raw{};
- BitField<0, 1, u32_le> IsConnected;
- BitField<1, 1, u32_le> IsWired;
- BitField<2, 1, u32_le> IsLeftJoyConnected;
- BitField<3, 1, u32_le> IsLeftJoyWired;
- BitField<4, 1, u32_le> IsRightJoyConnected;
- BitField<5, 1, u32_le> IsRightJoyWired;
+ BitField<0, 1, u32> IsConnected;
+ BitField<1, 1, u32> IsWired;
+ BitField<2, 1, u32> IsLeftJoyConnected;
+ BitField<3, 1, u32> IsLeftJoyWired;
+ BitField<4, 1, u32> IsRightJoyConnected;
+ BitField<5, 1, u32> IsRightJoyWired;
};
};
static_assert(sizeof(ConnectionState) == 4, "ConnectionState is an invalid size");
@@ -240,23 +240,23 @@ private:
struct NPadProperties {
union {
s64_le raw{};
- BitField<11, 1, s64_le> is_vertical;
- BitField<12, 1, s64_le> is_horizontal;
- BitField<13, 1, s64_le> use_plus;
- BitField<14, 1, s64_le> use_minus;
+ BitField<11, 1, s64> is_vertical;
+ BitField<12, 1, s64> is_horizontal;
+ BitField<13, 1, s64> use_plus;
+ BitField<14, 1, s64> use_minus;
};
};
struct NPadDevice {
union {
u32_le raw{};
- BitField<0, 1, s32_le> pro_controller;
- BitField<1, 1, s32_le> handheld;
- BitField<2, 1, s32_le> handheld_left;
- BitField<3, 1, s32_le> handheld_right;
- BitField<4, 1, s32_le> joycon_left;
- BitField<5, 1, s32_le> joycon_right;
- BitField<6, 1, s32_le> pokeball;
+ BitField<0, 1, s32> pro_controller;
+ BitField<1, 1, s32> handheld;
+ BitField<2, 1, s32> handheld_left;
+ BitField<3, 1, s32> handheld_right;
+ BitField<4, 1, s32> joycon_left;
+ BitField<5, 1, s32> joycon_right;
+ BitField<6, 1, s32> pokeball;
};
};
diff --git a/src/core/hle/service/hid/controllers/touchscreen.h b/src/core/hle/service/hid/controllers/touchscreen.h
index 012b6e0dd..76fc340e9 100644
--- a/src/core/hle/service/hid/controllers/touchscreen.h
+++ b/src/core/hle/service/hid/controllers/touchscreen.h
@@ -33,8 +33,8 @@ private:
struct Attributes {
union {
u32 raw{};
- BitField<0, 1, u32_le> start_touch;
- BitField<1, 1, u32_le> end_touch;
+ BitField<0, 1, u32> start_touch;
+ BitField<1, 1, u32> end_touch;
};
};
static_assert(sizeof(Attributes) == 0x4, "Attributes is an invalid size");
diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp
index 8a6de83a2..63b55758b 100644
--- a/src/core/hle/service/hid/hid.cpp
+++ b/src/core/hle/service/hid/hid.cpp
@@ -36,9 +36,9 @@ namespace Service::HID {
// Updating period for each HID device.
// TODO(ogniK): Find actual polling rate of hid
-constexpr u64 pad_update_ticks = Core::Timing::BASE_CLOCK_RATE / 66;
-constexpr u64 accelerometer_update_ticks = Core::Timing::BASE_CLOCK_RATE / 100;
-constexpr u64 gyroscope_update_ticks = Core::Timing::BASE_CLOCK_RATE / 100;
+constexpr s64 pad_update_ticks = static_cast<s64>(Core::Timing::BASE_CLOCK_RATE / 66);
+constexpr s64 accelerometer_update_ticks = static_cast<s64>(Core::Timing::BASE_CLOCK_RATE / 100);
+constexpr s64 gyroscope_update_ticks = static_cast<s64>(Core::Timing::BASE_CLOCK_RATE / 100);
constexpr std::size_t SHARED_MEMORY_SIZE = 0x40000;
IAppletResource::IAppletResource() : ServiceFramework("IAppletResource") {
@@ -75,7 +75,7 @@ IAppletResource::IAppletResource() : ServiceFramework("IAppletResource") {
// Register update callbacks
auto& core_timing = Core::System::GetInstance().CoreTiming();
pad_update_event =
- core_timing.RegisterEvent("HID::UpdatePadCallback", [this](u64 userdata, int cycles_late) {
+ core_timing.RegisterEvent("HID::UpdatePadCallback", [this](u64 userdata, s64 cycles_late) {
UpdateControllers(userdata, cycles_late);
});
@@ -106,7 +106,7 @@ void IAppletResource::GetSharedMemoryHandle(Kernel::HLERequestContext& ctx) {
rb.PushCopyObjects(shared_mem);
}
-void IAppletResource::UpdateControllers(u64 userdata, int cycles_late) {
+void IAppletResource::UpdateControllers(u64 userdata, s64 cycles_late) {
auto& core_timing = Core::System::GetInstance().CoreTiming();
const bool should_reload = Settings::values.is_device_reload_pending.exchange(false);
diff --git a/src/core/hle/service/hid/hid.h b/src/core/hle/service/hid/hid.h
index 7cc58db4c..d3660cad2 100644
--- a/src/core/hle/service/hid/hid.h
+++ b/src/core/hle/service/hid/hid.h
@@ -4,6 +4,9 @@
#pragma once
+#include "core/hle/service/hid/controllers/controller_base.h"
+#include "core/hle/service/service.h"
+
#include "controllers/controller_base.h"
#include "core/hle/service/service.h"
@@ -62,7 +65,7 @@ private:
}
void GetSharedMemoryHandle(Kernel::HLERequestContext& ctx);
- void UpdateControllers(u64 userdata, int cycles_late);
+ void UpdateControllers(u64 userdata, s64 cycles_late);
Kernel::SharedPtr<Kernel::SharedMemory> shared_mem;
diff --git a/src/core/hle/service/ldr/ldr.cpp b/src/core/hle/service/ldr/ldr.cpp
index 9df7ac50f..d65693fc7 100644
--- a/src/core/hle/service/ldr/ldr.cpp
+++ b/src/core/hle/service/ldr/ldr.cpp
@@ -319,15 +319,14 @@ public:
}
ASSERT(vm_manager
- .MirrorMemory(*map_address, nro_addr, nro_size,
- Kernel::MemoryState::ModuleCodeStatic)
+ .MirrorMemory(*map_address, nro_addr, nro_size, Kernel::MemoryState::ModuleCode)
.IsSuccess());
ASSERT(vm_manager.UnmapRange(nro_addr, nro_size).IsSuccess());
if (bss_size > 0) {
ASSERT(vm_manager
.MirrorMemory(*map_address + nro_size, bss_addr, bss_size,
- Kernel::MemoryState::ModuleCodeStatic)
+ Kernel::MemoryState::ModuleCode)
.IsSuccess());
ASSERT(vm_manager.UnmapRange(bss_addr, bss_size).IsSuccess());
}
@@ -388,8 +387,7 @@ public:
const auto& nro_size = iter->second.size;
ASSERT(vm_manager
- .MirrorMemory(heap_addr, mapped_addr, nro_size,
- Kernel::MemoryState::ModuleCodeStatic)
+ .MirrorMemory(heap_addr, mapped_addr, nro_size, Kernel::MemoryState::ModuleCode)
.IsSuccess());
ASSERT(vm_manager.UnmapRange(mapped_addr, nro_size).IsSuccess());
diff --git a/src/core/hle/service/lm/lm.cpp b/src/core/hle/service/lm/lm.cpp
index 1f462e087..2a61593e2 100644
--- a/src/core/hle/service/lm/lm.cpp
+++ b/src/core/hle/service/lm/lm.cpp
@@ -42,7 +42,7 @@ private:
union {
BitField<0, 16, Flags> flags;
BitField<16, 8, Severity> severity;
- BitField<24, 8, u32_le> verbosity;
+ BitField<24, 8, u32> verbosity;
};
u32_le payload_size;
diff --git a/src/core/hle/service/nvdrv/devices/nvdevice.h b/src/core/hle/service/nvdrv/devices/nvdevice.h
index 0f02a1a18..4f6042b00 100644
--- a/src/core/hle/service/nvdrv/devices/nvdevice.h
+++ b/src/core/hle/service/nvdrv/devices/nvdevice.h
@@ -19,11 +19,11 @@ public:
virtual ~nvdevice() = default;
union Ioctl {
u32_le raw;
- BitField<0, 8, u32_le> cmd;
- BitField<8, 8, u32_le> group;
- BitField<16, 14, u32_le> length;
- BitField<30, 1, u32_le> is_in;
- BitField<31, 1, u32_le> is_out;
+ BitField<0, 8, u32> cmd;
+ BitField<8, 8, u32> group;
+ BitField<16, 14, u32> length;
+ BitField<30, 1, u32> is_in;
+ BitField<31, 1, u32> is_out;
};
/**
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
index b031ebc66..af62d33d2 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp
@@ -89,7 +89,7 @@ u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output)
for (const auto& entry : entries) {
LOG_WARNING(Service_NVDRV, "remap entry, offset=0x{:X} handle=0x{:X} pages=0x{:X}",
entry.offset, entry.nvmap_handle, entry.pages);
- Tegra::GPUVAddr offset = static_cast<Tegra::GPUVAddr>(entry.offset) << 0x10;
+ GPUVAddr offset = static_cast<GPUVAddr>(entry.offset) << 0x10;
auto object = nvmap_dev->GetObject(entry.nvmap_handle);
if (!object) {
LOG_CRITICAL(Service_NVDRV, "nvmap {} is an invalid handle!", entry.nvmap_handle);
@@ -102,7 +102,7 @@ u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output)
u64 size = static_cast<u64>(entry.pages) << 0x10;
ASSERT(size <= object->size);
- Tegra::GPUVAddr returned = gpu.MemoryManager().MapBufferEx(object->addr, offset, size);
+ GPUVAddr returned = gpu.MemoryManager().MapBufferEx(object->addr, offset, size);
ASSERT(returned == offset);
}
std::memcpy(output.data(), entries.data(), output.size());
@@ -173,16 +173,8 @@ u32 nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& ou
return 0;
}
- auto& system_instance = Core::System::GetInstance();
-
- // Remove this memory region from the rasterizer cache.
- auto& gpu = system_instance.GPU();
- auto cpu_addr = gpu.MemoryManager().GpuToCpuAddress(params.offset);
- ASSERT(cpu_addr);
- gpu.FlushAndInvalidateRegion(ToCacheAddr(Memory::GetPointer(*cpu_addr)), itr->second.size);
-
- params.offset = gpu.MemoryManager().UnmapBuffer(params.offset, itr->second.size);
-
+ params.offset = Core::System::GetInstance().GPU().MemoryManager().UnmapBuffer(params.offset,
+ itr->second.size);
buffer_mappings.erase(itr->second.offset);
std::memcpy(output.data(), &params, output.size());
diff --git a/src/core/hle/service/nvflinger/nvflinger.cpp b/src/core/hle/service/nvflinger/nvflinger.cpp
index fc496b654..c7f5bbf28 100644
--- a/src/core/hle/service/nvflinger/nvflinger.cpp
+++ b/src/core/hle/service/nvflinger/nvflinger.cpp
@@ -26,7 +26,7 @@
namespace Service::NVFlinger {
constexpr std::size_t SCREEN_REFRESH_RATE = 60;
-constexpr u64 frame_ticks = static_cast<u64>(Core::Timing::BASE_CLOCK_RATE / SCREEN_REFRESH_RATE);
+constexpr s64 frame_ticks = static_cast<s64>(Core::Timing::BASE_CLOCK_RATE / SCREEN_REFRESH_RATE);
NVFlinger::NVFlinger(Core::Timing::CoreTiming& core_timing) : core_timing{core_timing} {
displays.emplace_back(0, "Default");
@@ -37,7 +37,7 @@ NVFlinger::NVFlinger(Core::Timing::CoreTiming& core_timing) : core_timing{core_t
// Schedule the screen composition events
composition_event =
- core_timing.RegisterEvent("ScreenComposition", [this](u64 userdata, int cycles_late) {
+ core_timing.RegisterEvent("ScreenComposition", [this](u64 userdata, s64 cycles_late) {
Compose();
this->core_timing.ScheduleEvent(frame_ticks - cycles_late, composition_event);
});
diff --git a/src/core/hle/service/set/set_sys.cpp b/src/core/hle/service/set/set_sys.cpp
index c9b4da5b0..ecee554bf 100644
--- a/src/core/hle/service/set/set_sys.cpp
+++ b/src/core/hle/service/set/set_sys.cpp
@@ -2,13 +2,88 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
+#include "common/assert.h"
#include "common/logging/log.h"
+#include "core/file_sys/errors.h"
+#include "core/file_sys/system_archive/system_version.h"
#include "core/hle/ipc_helpers.h"
#include "core/hle/kernel/client_port.h"
+#include "core/hle/service/filesystem/filesystem.h"
#include "core/hle/service/set/set_sys.h"
namespace Service::Set {
+namespace {
+constexpr u64 SYSTEM_VERSION_FILE_MINOR_REVISION_OFFSET = 0x05;
+
+enum class GetFirmwareVersionType {
+ Version1,
+ Version2,
+};
+
+void GetFirmwareVersionImpl(Kernel::HLERequestContext& ctx, GetFirmwareVersionType type) {
+ LOG_WARNING(Service_SET, "called - Using hardcoded firmware version '{}'",
+ FileSys::SystemArchive::GetLongDisplayVersion());
+
+ ASSERT_MSG(ctx.GetWriteBufferSize() == 0x100,
+ "FirmwareVersion output buffer must be 0x100 bytes in size!");
+
+ // Instead of using the normal procedure of checking for the real system archive and if it
+ // doesn't exist, synthesizing one, I feel that that would lead to strange bugs because a
+ // used is using a really old or really new SystemVersion title. The synthesized one ensures
+ // consistence (currently reports as 5.1.0-0.0)
+ const auto archive = FileSys::SystemArchive::SystemVersion();
+
+ const auto early_exit_failure = [&ctx](const std::string& desc, ResultCode code) {
+ LOG_ERROR(Service_SET, "General failure while attempting to resolve firmware version ({}).",
+ desc.c_str());
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(code);
+ };
+
+ if (archive == nullptr) {
+ early_exit_failure("The system version archive couldn't be synthesized.",
+ FileSys::ERROR_FAILED_MOUNT_ARCHIVE);
+ return;
+ }
+
+ const auto ver_file = archive->GetFile("file");
+ if (ver_file == nullptr) {
+ early_exit_failure("The system version archive didn't contain the file 'file'.",
+ FileSys::ERROR_INVALID_ARGUMENT);
+ return;
+ }
+
+ auto data = ver_file->ReadAllBytes();
+ if (data.size() != 0x100) {
+ early_exit_failure("The system version file 'file' was not the correct size.",
+ FileSys::ERROR_OUT_OF_BOUNDS);
+ return;
+ }
+
+ // If the command is GetFirmwareVersion (as opposed to GetFirmwareVersion2), hardware will
+ // zero out the REVISION_MINOR field.
+ if (type == GetFirmwareVersionType::Version1) {
+ data[SYSTEM_VERSION_FILE_MINOR_REVISION_OFFSET] = 0;
+ }
+
+ ctx.WriteBuffer(data);
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(RESULT_SUCCESS);
+}
+} // Anonymous namespace
+
+void SET_SYS::GetFirmwareVersion(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_SET, "called");
+ GetFirmwareVersionImpl(ctx, GetFirmwareVersionType::Version1);
+}
+
+void SET_SYS::GetFirmwareVersion2(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_SET, "called");
+ GetFirmwareVersionImpl(ctx, GetFirmwareVersionType::Version2);
+}
+
void SET_SYS::GetColorSetId(Kernel::HLERequestContext& ctx) {
LOG_DEBUG(Service_SET, "called");
@@ -33,8 +108,8 @@ SET_SYS::SET_SYS() : ServiceFramework("set:sys") {
{0, nullptr, "SetLanguageCode"},
{1, nullptr, "SetNetworkSettings"},
{2, nullptr, "GetNetworkSettings"},
- {3, nullptr, "GetFirmwareVersion"},
- {4, nullptr, "GetFirmwareVersion2"},
+ {3, &SET_SYS::GetFirmwareVersion, "GetFirmwareVersion"},
+ {4, &SET_SYS::GetFirmwareVersion2, "GetFirmwareVersion2"},
{5, nullptr, "GetFirmwareVersionDigest"},
{7, nullptr, "GetLockScreenFlag"},
{8, nullptr, "SetLockScreenFlag"},
diff --git a/src/core/hle/service/set/set_sys.h b/src/core/hle/service/set/set_sys.h
index f602f3c77..13ee2cf46 100644
--- a/src/core/hle/service/set/set_sys.h
+++ b/src/core/hle/service/set/set_sys.h
@@ -20,6 +20,8 @@ private:
BasicBlack = 1,
};
+ void GetFirmwareVersion(Kernel::HLERequestContext& ctx);
+ void GetFirmwareVersion2(Kernel::HLERequestContext& ctx);
void GetColorSetId(Kernel::HLERequestContext& ctx);
void SetColorSetId(Kernel::HLERequestContext& ctx);
diff --git a/src/core/loader/elf.cpp b/src/core/loader/elf.cpp
index 6057c7f26..8b1920f22 100644
--- a/src/core/loader/elf.cpp
+++ b/src/core/loader/elf.cpp
@@ -9,6 +9,7 @@
#include "common/common_types.h"
#include "common/file_util.h"
#include "common/logging/log.h"
+#include "core/hle/kernel/code_set.h"
#include "core/hle/kernel/process.h"
#include "core/hle/kernel/vm_manager.h"
#include "core/loader/elf.h"
diff --git a/src/core/loader/linker.cpp b/src/core/loader/linker.cpp
deleted file mode 100644
index 57ca8c3ee..000000000
--- a/src/core/loader/linker.cpp
+++ /dev/null
@@ -1,147 +0,0 @@
-// Copyright 2018 yuzu emulator team
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#include <vector>
-
-#include "common/common_funcs.h"
-#include "common/logging/log.h"
-#include "common/swap.h"
-#include "core/loader/linker.h"
-#include "core/memory.h"
-
-namespace Loader {
-
-enum class RelocationType : u32 { ABS64 = 257, GLOB_DAT = 1025, JUMP_SLOT = 1026, RELATIVE = 1027 };
-
-enum DynamicType : u32 {
- DT_NULL = 0,
- DT_PLTRELSZ = 2,
- DT_STRTAB = 5,
- DT_SYMTAB = 6,
- DT_RELA = 7,
- DT_RELASZ = 8,
- DT_STRSZ = 10,
- DT_JMPREL = 23,
-};
-
-struct Elf64_Rela {
- u64_le offset;
- RelocationType type;
- u32_le symbol;
- s64_le addend;
-};
-static_assert(sizeof(Elf64_Rela) == 0x18, "Elf64_Rela has incorrect size.");
-
-struct Elf64_Dyn {
- u64_le tag;
- u64_le value;
-};
-static_assert(sizeof(Elf64_Dyn) == 0x10, "Elf64_Dyn has incorrect size.");
-
-struct Elf64_Sym {
- u32_le name;
- INSERT_PADDING_BYTES(0x2);
- u16_le shndx;
- u64_le value;
- u64_le size;
-};
-static_assert(sizeof(Elf64_Sym) == 0x18, "Elf64_Sym has incorrect size.");
-
-void Linker::WriteRelocations(std::vector<u8>& program_image, const std::vector<Symbol>& symbols,
- u64 relocation_offset, u64 size, VAddr load_base) {
- for (u64 i = 0; i < size; i += sizeof(Elf64_Rela)) {
- Elf64_Rela rela;
- std::memcpy(&rela, &program_image[relocation_offset + i], sizeof(Elf64_Rela));
-
- const Symbol& symbol = symbols[rela.symbol];
- switch (rela.type) {
- case RelocationType::RELATIVE: {
- const u64 value = load_base + rela.addend;
- if (!symbol.name.empty()) {
- exports[symbol.name] = value;
- }
- std::memcpy(&program_image[rela.offset], &value, sizeof(u64));
- break;
- }
- case RelocationType::JUMP_SLOT:
- case RelocationType::GLOB_DAT:
- if (!symbol.value) {
- imports[symbol.name] = {rela.offset + load_base, 0};
- } else {
- exports[symbol.name] = symbol.value;
- std::memcpy(&program_image[rela.offset], &symbol.value, sizeof(u64));
- }
- break;
- case RelocationType::ABS64:
- if (!symbol.value) {
- imports[symbol.name] = {rela.offset + load_base, rela.addend};
- } else {
- const u64 value = symbol.value + rela.addend;
- exports[symbol.name] = value;
- std::memcpy(&program_image[rela.offset], &value, sizeof(u64));
- }
- break;
- default:
- LOG_CRITICAL(Loader, "Unknown relocation type: {}", static_cast<int>(rela.type));
- break;
- }
- }
-}
-
-void Linker::Relocate(std::vector<u8>& program_image, u32 dynamic_section_offset, VAddr load_base) {
- std::map<u64, u64> dynamic;
- while (dynamic_section_offset < program_image.size()) {
- Elf64_Dyn dyn;
- std::memcpy(&dyn, &program_image[dynamic_section_offset], sizeof(Elf64_Dyn));
- dynamic_section_offset += sizeof(Elf64_Dyn);
-
- if (dyn.tag == DT_NULL) {
- break;
- }
- dynamic[dyn.tag] = dyn.value;
- }
-
- u64 offset = dynamic[DT_SYMTAB];
- std::vector<Symbol> symbols;
- while (offset < program_image.size()) {
- Elf64_Sym sym;
- std::memcpy(&sym, &program_image[offset], sizeof(Elf64_Sym));
- offset += sizeof(Elf64_Sym);
-
- if (sym.name >= dynamic[DT_STRSZ]) {
- break;
- }
-
- std::string name = reinterpret_cast<char*>(&program_image[dynamic[DT_STRTAB] + sym.name]);
- if (sym.value) {
- exports[name] = load_base + sym.value;
- symbols.emplace_back(std::move(name), load_base + sym.value);
- } else {
- symbols.emplace_back(std::move(name), 0);
- }
- }
-
- if (dynamic.find(DT_RELA) != dynamic.end()) {
- WriteRelocations(program_image, symbols, dynamic[DT_RELA], dynamic[DT_RELASZ], load_base);
- }
-
- if (dynamic.find(DT_JMPREL) != dynamic.end()) {
- WriteRelocations(program_image, symbols, dynamic[DT_JMPREL], dynamic[DT_PLTRELSZ],
- load_base);
- }
-}
-
-void Linker::ResolveImports() {
- // Resolve imports
- for (const auto& import : imports) {
- const auto& search = exports.find(import.first);
- if (search != exports.end()) {
- Memory::Write64(import.second.ea, search->second + import.second.addend);
- } else {
- LOG_ERROR(Loader, "Unresolved import: {}", import.first);
- }
- }
-}
-
-} // namespace Loader
diff --git a/src/core/loader/linker.h b/src/core/loader/linker.h
deleted file mode 100644
index 107625837..000000000
--- a/src/core/loader/linker.h
+++ /dev/null
@@ -1,36 +0,0 @@
-// Copyright 2018 yuzu emulator team
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <map>
-#include <string>
-#include "common/common_types.h"
-
-namespace Loader {
-
-class Linker {
-protected:
- struct Symbol {
- Symbol(std::string&& name, u64 value) : name(std::move(name)), value(value) {}
- std::string name;
- u64 value;
- };
-
- struct Import {
- VAddr ea;
- s64 addend;
- };
-
- void WriteRelocations(std::vector<u8>& program_image, const std::vector<Symbol>& symbols,
- u64 relocation_offset, u64 size, VAddr load_base);
- void Relocate(std::vector<u8>& program_image, u32 dynamic_section_offset, VAddr load_base);
-
- void ResolveImports();
-
- std::map<std::string, Import> imports;
- std::map<std::string, VAddr> exports;
-};
-
-} // namespace Loader
diff --git a/src/core/loader/nro.cpp b/src/core/loader/nro.cpp
index 4fad0c0dd..5de02a94b 100644
--- a/src/core/loader/nro.cpp
+++ b/src/core/loader/nro.cpp
@@ -14,6 +14,7 @@
#include "core/file_sys/romfs_factory.h"
#include "core/file_sys/vfs_offset.h"
#include "core/gdbstub/gdbstub.h"
+#include "core/hle/kernel/code_set.h"
#include "core/hle/kernel/process.h"
#include "core/hle/kernel/vm_manager.h"
#include "core/hle/service/filesystem/filesystem.h"
diff --git a/src/core/loader/nro.h b/src/core/loader/nro.h
index 013d629c0..85b0ed644 100644
--- a/src/core/loader/nro.h
+++ b/src/core/loader/nro.h
@@ -4,10 +4,10 @@
#pragma once
+#include <memory>
#include <string>
#include <vector>
#include "common/common_types.h"
-#include "core/loader/linker.h"
#include "core/loader/loader.h"
namespace FileSys {
@@ -21,7 +21,7 @@ class Process;
namespace Loader {
/// Loads an NRO file
-class AppLoader_NRO final : public AppLoader, Linker {
+class AppLoader_NRO final : public AppLoader {
public:
explicit AppLoader_NRO(FileSys::VirtualFile file);
~AppLoader_NRO() override;
diff --git a/src/core/loader/nso.cpp b/src/core/loader/nso.cpp
index 6ded0b707..714d85a59 100644
--- a/src/core/loader/nso.cpp
+++ b/src/core/loader/nso.cpp
@@ -7,10 +7,13 @@
#include <lz4.h>
#include "common/common_funcs.h"
#include "common/file_util.h"
+#include "common/hex_util.h"
#include "common/logging/log.h"
#include "common/swap.h"
+#include "core/core.h"
#include "core/file_sys/patch_manager.h"
#include "core/gdbstub/gdbstub.h"
+#include "core/hle/kernel/code_set.h"
#include "core/hle/kernel/process.h"
#include "core/hle/kernel/vm_manager.h"
#include "core/loader/nso.h"
@@ -18,36 +21,8 @@
#include "core/settings.h"
namespace Loader {
-
-struct NsoSegmentHeader {
- u32_le offset;
- u32_le location;
- u32_le size;
- union {
- u32_le alignment;
- u32_le bss_size;
- };
-};
-static_assert(sizeof(NsoSegmentHeader) == 0x10, "NsoSegmentHeader has incorrect size.");
-
-struct NsoHeader {
- u32_le magic;
- u32_le version;
- INSERT_PADDING_WORDS(1);
- u8 flags;
- std::array<NsoSegmentHeader, 3> segments; // Text, RoData, Data (in that order)
- std::array<u8, 0x20> build_id;
- std::array<u32_le, 3> segments_compressed_size;
-
- bool IsSegmentCompressed(size_t segment_num) const {
- ASSERT_MSG(segment_num < 3, "Invalid segment {}", segment_num);
- return ((flags >> segment_num) & 1);
- }
-};
-static_assert(sizeof(NsoHeader) == 0x6c, "NsoHeader has incorrect size.");
-static_assert(std::is_trivially_copyable_v<NsoHeader>, "NsoHeader isn't trivially copyable.");
-
-struct ModHeader {
+namespace {
+struct MODHeader {
u32_le magic;
u32_le dynamic_offset;
u32_le bss_start_offset;
@@ -56,25 +31,10 @@ struct ModHeader {
u32_le eh_frame_hdr_end_offset;
u32_le module_offset; // Offset to runtime-generated module object. typically equal to .bss base
};
-static_assert(sizeof(ModHeader) == 0x1c, "ModHeader has incorrect size.");
-
-AppLoader_NSO::AppLoader_NSO(FileSys::VirtualFile file) : AppLoader(std::move(file)) {}
+static_assert(sizeof(MODHeader) == 0x1c, "MODHeader has incorrect size.");
-FileType AppLoader_NSO::IdentifyType(const FileSys::VirtualFile& file) {
- u32 magic = 0;
- if (file->ReadObject(&magic) != sizeof(magic)) {
- return FileType::Error;
- }
-
- if (Common::MakeMagic('N', 'S', 'O', '0') != magic) {
- return FileType::Error;
- }
-
- return FileType::NSO;
-}
-
-static std::vector<u8> DecompressSegment(const std::vector<u8>& compressed_data,
- const NsoSegmentHeader& header) {
+std::vector<u8> DecompressSegment(const std::vector<u8>& compressed_data,
+ const NSOSegmentHeader& header) {
std::vector<u8> uncompressed_data(header.size);
const int bytes_uncompressed =
LZ4_decompress_safe(reinterpret_cast<const char*>(compressed_data.data()),
@@ -88,23 +48,47 @@ static std::vector<u8> DecompressSegment(const std::vector<u8>& compressed_data,
return uncompressed_data;
}
-static constexpr u32 PageAlignSize(u32 size) {
+constexpr u32 PageAlignSize(u32 size) {
return (size + Memory::PAGE_MASK) & ~Memory::PAGE_MASK;
}
+} // Anonymous namespace
+
+bool NSOHeader::IsSegmentCompressed(size_t segment_num) const {
+ ASSERT_MSG(segment_num < 3, "Invalid segment {}", segment_num);
+ return ((flags >> segment_num) & 1) != 0;
+}
+
+AppLoader_NSO::AppLoader_NSO(FileSys::VirtualFile file) : AppLoader(std::move(file)) {}
+
+FileType AppLoader_NSO::IdentifyType(const FileSys::VirtualFile& file) {
+ u32 magic = 0;
+ if (file->ReadObject(&magic) != sizeof(magic)) {
+ return FileType::Error;
+ }
+
+ if (Common::MakeMagic('N', 'S', 'O', '0') != magic) {
+ return FileType::Error;
+ }
+
+ return FileType::NSO;
+}
std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process,
const FileSys::VfsFile& file, VAddr load_base,
bool should_pass_arguments,
std::optional<FileSys::PatchManager> pm) {
- if (file.GetSize() < sizeof(NsoHeader))
+ if (file.GetSize() < sizeof(NSOHeader)) {
return {};
+ }
- NsoHeader nso_header{};
- if (sizeof(NsoHeader) != file.ReadObject(&nso_header))
+ NSOHeader nso_header{};
+ if (sizeof(NSOHeader) != file.ReadObject(&nso_header)) {
return {};
+ }
- if (nso_header.magic != Common::MakeMagic('N', 'S', 'O', '0'))
+ if (nso_header.magic != Common::MakeMagic('N', 'S', 'O', '0')) {
return {};
+ }
// Build program image
Kernel::CodeSet codeset;
@@ -140,10 +124,10 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process,
std::memcpy(&module_offset, program_image.data() + 4, sizeof(u32));
// Read MOD header
- ModHeader mod_header{};
+ MODHeader mod_header{};
// Default .bss to size in segment header if MOD0 section doesn't exist
u32 bss_size{PageAlignSize(nso_header.segments[2].bss_size)};
- std::memcpy(&mod_header, program_image.data() + module_offset, sizeof(ModHeader));
+ std::memcpy(&mod_header, program_image.data() + module_offset, sizeof(MODHeader));
const bool has_mod_header{mod_header.magic == Common::MakeMagic('M', 'O', 'D', '0')};
if (has_mod_header) {
// Resize program image to include .bss section and page align each section
@@ -155,13 +139,25 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process,
// Apply patches if necessary
if (pm && (pm->HasNSOPatch(nso_header.build_id) || Settings::values.dump_nso)) {
- std::vector<u8> pi_header(program_image.size() + 0x100);
- std::memcpy(pi_header.data(), &nso_header, sizeof(NsoHeader));
- std::memcpy(pi_header.data() + 0x100, program_image.data(), program_image.size());
+ std::vector<u8> pi_header(sizeof(NSOHeader) + program_image.size());
+ pi_header.insert(pi_header.begin(), reinterpret_cast<u8*>(&nso_header),
+ reinterpret_cast<u8*>(&nso_header) + sizeof(NSOHeader));
+ pi_header.insert(pi_header.begin() + sizeof(NSOHeader), program_image.begin(),
+ program_image.end());
pi_header = pm->PatchNSO(pi_header);
- std::memcpy(program_image.data(), pi_header.data() + 0x100, program_image.size());
+ std::copy(pi_header.begin() + sizeof(NSOHeader), pi_header.end(), program_image.begin());
+ }
+
+ // Apply cheats if they exist and the program has a valid title ID
+ if (pm) {
+ auto& system = Core::System::GetInstance();
+ const auto cheats = pm->CreateCheatList(system, nso_header.build_id);
+ if (!cheats.empty()) {
+ system.RegisterCheatList(cheats, Common::HexArrayToString(nso_header.build_id),
+ load_base, load_base + program_image.size());
+ }
}
// Load codeset for current process
diff --git a/src/core/loader/nso.h b/src/core/loader/nso.h
index 135b6ea5a..4674c3724 100644
--- a/src/core/loader/nso.h
+++ b/src/core/loader/nso.h
@@ -4,10 +4,12 @@
#pragma once
+#include <array>
#include <optional>
+#include <type_traits>
#include "common/common_types.h"
+#include "common/swap.h"
#include "core/file_sys/patch_manager.h"
-#include "core/loader/linker.h"
#include "core/loader/loader.h"
namespace Kernel {
@@ -16,6 +18,43 @@ class Process;
namespace Loader {
+struct NSOSegmentHeader {
+ u32_le offset;
+ u32_le location;
+ u32_le size;
+ union {
+ u32_le alignment;
+ u32_le bss_size;
+ };
+};
+static_assert(sizeof(NSOSegmentHeader) == 0x10, "NsoSegmentHeader has incorrect size.");
+
+struct NSOHeader {
+ using SHA256Hash = std::array<u8, 0x20>;
+
+ struct RODataRelativeExtent {
+ u32_le data_offset;
+ u32_le size;
+ };
+
+ u32_le magic;
+ u32_le version;
+ u32 reserved;
+ u32_le flags;
+ std::array<NSOSegmentHeader, 3> segments; // Text, RoData, Data (in that order)
+ std::array<u8, 0x20> build_id;
+ std::array<u32_le, 3> segments_compressed_size;
+ std::array<u8, 0x1C> padding;
+ RODataRelativeExtent api_info_extent;
+ RODataRelativeExtent dynstr_extent;
+ RODataRelativeExtent dynsyn_extent;
+ std::array<SHA256Hash, 3> segment_hashes;
+
+ bool IsSegmentCompressed(size_t segment_num) const;
+};
+static_assert(sizeof(NSOHeader) == 0x100, "NSOHeader has incorrect size.");
+static_assert(std::is_trivially_copyable_v<NSOHeader>, "NSOHeader must be trivially copyable.");
+
constexpr u64 NSO_ARGUMENT_DATA_ALLOCATION_SIZE = 0x9000;
struct NSOArgumentHeader {
@@ -26,7 +65,7 @@ struct NSOArgumentHeader {
static_assert(sizeof(NSOArgumentHeader) == 0x20, "NSOArgumentHeader has incorrect size.");
/// Loads an NSO file
-class AppLoader_NSO final : public AppLoader, Linker {
+class AppLoader_NSO final : public AppLoader {
public:
explicit AppLoader_NSO(FileSys::VirtualFile file);
diff --git a/src/core/memory.cpp b/src/core/memory.cpp
index 365ac82b4..332c1037c 100644
--- a/src/core/memory.cpp
+++ b/src/core/memory.cpp
@@ -48,7 +48,7 @@ static void MapPages(Common::PageTable& page_table, VAddr base, u64 size, u8* me
(base + size) * PAGE_SIZE);
// During boot, current_page_table might not be set yet, in which case we need not flush
- if (current_page_table) {
+ if (Core::System::GetInstance().IsPoweredOn()) {
Core::System::GetInstance().GPU().FlushAndInvalidateRegion(base << PAGE_BITS,
size * PAGE_SIZE);
}
diff --git a/src/core/memory.h b/src/core/memory.h
index 3f60d868c..1d38cdca8 100644
--- a/src/core/memory.h
+++ b/src/core/memory.h
@@ -6,9 +6,6 @@
#include <cstddef>
#include <string>
-#include <tuple>
-#include <vector>
-#include <boost/icl/interval_map.hpp>
#include "common/common_types.h"
namespace Common {
diff --git a/src/input_common/sdl/sdl.h b/src/input_common/sdl/sdl.h
index 02a8d2e2c..d7f24c68a 100644
--- a/src/input_common/sdl/sdl.h
+++ b/src/input_common/sdl/sdl.h
@@ -24,17 +24,19 @@ namespace InputCommon::SDL {
class State {
public:
- /// Unresisters SDL device factories and shut them down.
+ using Pollers = std::vector<std::unique_ptr<Polling::DevicePoller>>;
+
+ /// Unregisters SDL device factories and shut them down.
virtual ~State() = default;
- virtual std::vector<std::unique_ptr<InputCommon::Polling::DevicePoller>> GetPollers(
- InputCommon::Polling::DeviceType type) = 0;
+ virtual Pollers GetPollers(Polling::DeviceType type) = 0;
};
class NullState : public State {
public:
- std::vector<std::unique_ptr<InputCommon::Polling::DevicePoller>> GetPollers(
- InputCommon::Polling::DeviceType type) override {}
+ Pollers GetPollers(Polling::DeviceType type) override {
+ return {};
+ }
};
std::unique_ptr<State> Init();
diff --git a/src/input_common/sdl/sdl_impl.cpp b/src/input_common/sdl/sdl_impl.cpp
index 6e8376549..b132d77f5 100644
--- a/src/input_common/sdl/sdl_impl.cpp
+++ b/src/input_common/sdl/sdl_impl.cpp
@@ -650,9 +650,9 @@ private:
};
} // namespace Polling
-std::vector<std::unique_ptr<InputCommon::Polling::DevicePoller>> SDLState::GetPollers(
- InputCommon::Polling::DeviceType type) {
- std::vector<std::unique_ptr<InputCommon::Polling::DevicePoller>> pollers;
+SDLState::Pollers SDLState::GetPollers(InputCommon::Polling::DeviceType type) {
+ Pollers pollers;
+
switch (type) {
case InputCommon::Polling::DeviceType::Analog:
pollers.emplace_back(std::make_unique<Polling::SDLAnalogPoller>(*this));
@@ -660,8 +660,9 @@ std::vector<std::unique_ptr<InputCommon::Polling::DevicePoller>> SDLState::GetPo
case InputCommon::Polling::DeviceType::Button:
pollers.emplace_back(std::make_unique<Polling::SDLButtonPoller>(*this));
break;
- return pollers;
}
+
+ return pollers;
}
} // namespace SDL
diff --git a/src/input_common/sdl/sdl_impl.h b/src/input_common/sdl/sdl_impl.h
index fec82fbe6..2579741d6 100644
--- a/src/input_common/sdl/sdl_impl.h
+++ b/src/input_common/sdl/sdl_impl.h
@@ -25,7 +25,7 @@ public:
/// Initializes and registers SDL device factories
SDLState();
- /// Unresisters SDL device factories and shut them down.
+ /// Unregisters SDL device factories and shut them down.
~SDLState() override;
/// Handle SDL_Events for joysticks from SDL_PollEvent
@@ -35,8 +35,7 @@ public:
std::shared_ptr<SDLJoystick> GetSDLJoystickByGUID(const std::string& guid, int port);
/// Get all DevicePoller that use the SDL backend for a specific device type
- std::vector<std::unique_ptr<InputCommon::Polling::DevicePoller>> GetPollers(
- InputCommon::Polling::DeviceType type) override;
+ Pollers GetPollers(Polling::DeviceType type) override;
/// Used by the Pollers during config
std::atomic<bool> polling = false;
diff --git a/src/tests/CMakeLists.txt b/src/tests/CMakeLists.txt
index 37f09ce5f..c7038b217 100644
--- a/src/tests/CMakeLists.txt
+++ b/src/tests/CMakeLists.txt
@@ -1,4 +1,7 @@
add_executable(tests
+ common/bit_field.cpp
+ common/bit_utils.cpp
+ common/multi_level_queue.cpp
common/param_package.cpp
common/ring_buffer.cpp
core/arm/arm_test_common.cpp
diff --git a/src/tests/common/bit_field.cpp b/src/tests/common/bit_field.cpp
new file mode 100644
index 000000000..8ca1889f9
--- /dev/null
+++ b/src/tests/common/bit_field.cpp
@@ -0,0 +1,90 @@
+// Copyright 2019 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <array>
+#include <cstring>
+#include <type_traits>
+#include <catch2/catch.hpp>
+#include "common/bit_field.h"
+
+TEST_CASE("BitField", "[common]") {
+ enum class TestEnum : u32 {
+ A = 0b10111101,
+ B = 0b10101110,
+ C = 0b00001111,
+ };
+
+ union LEBitField {
+ u32_le raw;
+ BitField<0, 6, u32> a;
+ BitField<6, 4, s32> b;
+ BitField<10, 8, TestEnum> c;
+ BitField<18, 14, u32> d;
+ } le_bitfield;
+
+ union BEBitField {
+ u32_be raw;
+ BitFieldBE<0, 6, u32> a;
+ BitFieldBE<6, 4, s32> b;
+ BitFieldBE<10, 8, TestEnum> c;
+ BitFieldBE<18, 14, u32> d;
+ } be_bitfield;
+
+ static_assert(sizeof(LEBitField) == sizeof(u32));
+ static_assert(sizeof(BEBitField) == sizeof(u32));
+ static_assert(std::is_trivially_copyable_v<LEBitField>);
+ static_assert(std::is_trivially_copyable_v<BEBitField>);
+
+ std::array<u8, 4> raw{{
+ 0b01101100,
+ 0b11110110,
+ 0b10111010,
+ 0b11101100,
+ }};
+
+ std::memcpy(&le_bitfield, &raw, sizeof(raw));
+ std::memcpy(&be_bitfield, &raw, sizeof(raw));
+
+ // bit fields: 11101100101110'10111101'1001'101100
+ REQUIRE(le_bitfield.raw == 0b11101100'10111010'11110110'01101100);
+ REQUIRE(le_bitfield.a == 0b101100);
+ REQUIRE(le_bitfield.b == -7); // 1001 as two's complement
+ REQUIRE(le_bitfield.c == TestEnum::A);
+ REQUIRE(le_bitfield.d == 0b11101100101110);
+
+ le_bitfield.a.Assign(0b000111);
+ le_bitfield.b.Assign(-1);
+ le_bitfield.c.Assign(TestEnum::C);
+ le_bitfield.d.Assign(0b01010101010101);
+ std::memcpy(&raw, &le_bitfield, sizeof(raw));
+ // bit fields: 01010101010101'00001111'1111'000111
+ REQUIRE(le_bitfield.raw == 0b01010101'01010100'00111111'11000111);
+ REQUIRE(raw == std::array<u8, 4>{{
+ 0b11000111,
+ 0b00111111,
+ 0b01010100,
+ 0b01010101,
+ }});
+
+ // bit fields: 01101100111101'10101110'1011'101100
+ REQUIRE(be_bitfield.raw == 0b01101100'11110110'10111010'11101100);
+ REQUIRE(be_bitfield.a == 0b101100);
+ REQUIRE(be_bitfield.b == -5); // 1011 as two's complement
+ REQUIRE(be_bitfield.c == TestEnum::B);
+ REQUIRE(be_bitfield.d == 0b01101100111101);
+
+ be_bitfield.a.Assign(0b000111);
+ be_bitfield.b.Assign(-1);
+ be_bitfield.c.Assign(TestEnum::C);
+ be_bitfield.d.Assign(0b01010101010101);
+ std::memcpy(&raw, &be_bitfield, sizeof(raw));
+ // bit fields: 01010101010101'00001111'1111'000111
+ REQUIRE(be_bitfield.raw == 0b01010101'01010100'00111111'11000111);
+ REQUIRE(raw == std::array<u8, 4>{{
+ 0b01010101,
+ 0b01010100,
+ 0b00111111,
+ 0b11000111,
+ }});
+}
diff --git a/src/tests/common/bit_utils.cpp b/src/tests/common/bit_utils.cpp
new file mode 100644
index 000000000..479b5995a
--- /dev/null
+++ b/src/tests/common/bit_utils.cpp
@@ -0,0 +1,23 @@
+// Copyright 2017 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <catch2/catch.hpp>
+#include <math.h>
+#include "common/bit_util.h"
+
+namespace Common {
+
+TEST_CASE("BitUtils::CountTrailingZeroes", "[common]") {
+ REQUIRE(Common::CountTrailingZeroes32(0) == 32);
+ REQUIRE(Common::CountTrailingZeroes64(0) == 64);
+ REQUIRE(Common::CountTrailingZeroes32(9) == 0);
+ REQUIRE(Common::CountTrailingZeroes32(8) == 3);
+ REQUIRE(Common::CountTrailingZeroes32(0x801000) == 12);
+ REQUIRE(Common::CountTrailingZeroes64(9) == 0);
+ REQUIRE(Common::CountTrailingZeroes64(8) == 3);
+ REQUIRE(Common::CountTrailingZeroes64(0x801000) == 12);
+ REQUIRE(Common::CountTrailingZeroes64(0x801000000000UL) == 36);
+}
+
+} // namespace Common
diff --git a/src/tests/common/multi_level_queue.cpp b/src/tests/common/multi_level_queue.cpp
new file mode 100644
index 000000000..cca7ec7da
--- /dev/null
+++ b/src/tests/common/multi_level_queue.cpp
@@ -0,0 +1,55 @@
+// Copyright 2019 Yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <catch2/catch.hpp>
+#include <math.h>
+#include "common/common_types.h"
+#include "common/multi_level_queue.h"
+
+namespace Common {
+
+TEST_CASE("MultiLevelQueue", "[common]") {
+ std::array<f32, 8> values = {0.0, 5.0, 1.0, 9.0, 8.0, 2.0, 6.0, 7.0};
+ Common::MultiLevelQueue<f32, 64> mlq;
+ REQUIRE(mlq.empty());
+ mlq.add(values[2], 2);
+ mlq.add(values[7], 7);
+ mlq.add(values[3], 3);
+ mlq.add(values[4], 4);
+ mlq.add(values[0], 0);
+ mlq.add(values[5], 5);
+ mlq.add(values[6], 6);
+ mlq.add(values[1], 1);
+ u32 index = 0;
+ bool all_set = true;
+ for (auto& f : mlq) {
+ all_set &= (f == values[index]);
+ index++;
+ }
+ REQUIRE(all_set);
+ REQUIRE(!mlq.empty());
+ f32 v = 8.0;
+ mlq.add(v, 2);
+ v = -7.0;
+ mlq.add(v, 2, false);
+ REQUIRE(mlq.front(2) == -7.0);
+ mlq.yield(2);
+ REQUIRE(mlq.front(2) == values[2]);
+ REQUIRE(mlq.back(2) == -7.0);
+ REQUIRE(mlq.empty(8));
+ v = 10.0;
+ mlq.add(v, 8);
+ mlq.adjust(v, 8, 9);
+ REQUIRE(mlq.front(9) == v);
+ REQUIRE(mlq.empty(8));
+ REQUIRE(!mlq.empty(9));
+ mlq.adjust(values[0], 0, 9);
+ REQUIRE(mlq.highest_priority_set() == 1);
+ REQUIRE(mlq.lowest_priority_set() == 9);
+ mlq.remove(values[1], 1);
+ REQUIRE(mlq.highest_priority_set() == 2);
+ REQUIRE(mlq.empty(1));
+}
+
+} // namespace Common
diff --git a/src/video_core/dma_pusher.h b/src/video_core/dma_pusher.h
index 27a36348c..6ab06518f 100644
--- a/src/video_core/dma_pusher.h
+++ b/src/video_core/dma_pusher.h
@@ -9,7 +9,6 @@
#include "common/bit_field.h"
#include "common/common_types.h"
-#include "video_core/memory_manager.h"
namespace Tegra {
diff --git a/src/video_core/engines/kepler_memory.cpp b/src/video_core/engines/kepler_memory.cpp
index 0931b9626..e259bf46b 100644
--- a/src/video_core/engines/kepler_memory.cpp
+++ b/src/video_core/engines/kepler_memory.cpp
@@ -46,7 +46,7 @@ void KeplerMemory::ProcessData(u32 data) {
// contain a dirty surface that will have to be written back to memory.
const GPUVAddr address{regs.dest.Address() + state.write_offset * sizeof(u32)};
rasterizer.InvalidateRegion(ToCacheAddr(memory_manager.GetPointer(address)), sizeof(u32));
- memory_manager.Write32(address, data);
+ memory_manager.Write<u32>(address, data);
system.GPU().Maxwell3D().dirty_flags.OnMemoryWrite();
diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp
index c5d5be4ef..defcfbd3f 100644
--- a/src/video_core/engines/maxwell_3d.cpp
+++ b/src/video_core/engines/maxwell_3d.cpp
@@ -307,7 +307,7 @@ void Maxwell3D::ProcessQueryGet() {
// Write the current query sequence to the sequence address.
// TODO(Subv): Find out what happens if you use a long query type but mark it as a short
// query.
- memory_manager.Write32(sequence_address, sequence);
+ memory_manager.Write<u32>(sequence_address, sequence);
} else {
// Write the 128-bit result structure in long mode. Note: We emulate an infinitely fast
// GPU, this command may actually take a while to complete in real hardware due to GPU
@@ -395,7 +395,7 @@ void Maxwell3D::ProcessCBData(u32 value) {
u8* ptr{memory_manager.GetPointer(address)};
rasterizer.InvalidateRegion(ToCacheAddr(ptr), sizeof(u32));
- memory_manager.Write32(address, value);
+ memory_manager.Write<u32>(address, value);
dirty_flags.OnMemoryWrite();
@@ -447,7 +447,7 @@ std::vector<Texture::FullTextureInfo> Maxwell3D::GetStageTextures(Regs::ShaderSt
for (GPUVAddr current_texture = tex_info_buffer.address + TextureInfoOffset;
current_texture < tex_info_buffer_end; current_texture += sizeof(Texture::TextureHandle)) {
- const Texture::TextureHandle tex_handle{memory_manager.Read32(current_texture)};
+ const Texture::TextureHandle tex_handle{memory_manager.Read<u32>(current_texture)};
Texture::FullTextureInfo tex_info{};
// TODO(Subv): Use the shader to determine which textures are actually accessed.
@@ -482,7 +482,7 @@ Texture::FullTextureInfo Maxwell3D::GetStageTexture(Regs::ShaderStage stage,
ASSERT(tex_info_address < tex_info_buffer.address + tex_info_buffer.size);
- const Texture::TextureHandle tex_handle{memory_manager.Read32(tex_info_address)};
+ const Texture::TextureHandle tex_handle{memory_manager.Read<u32>(tex_info_address)};
Texture::FullTextureInfo tex_info{};
tex_info.index = static_cast<u32>(offset);
diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp
index a0ded4c25..5cca5c29a 100644
--- a/src/video_core/engines/maxwell_dma.cpp
+++ b/src/video_core/engines/maxwell_dma.cpp
@@ -88,6 +88,16 @@ void MaxwellDMA::HandleCopy() {
auto source_ptr{memory_manager.GetPointer(source)};
auto dst_ptr{memory_manager.GetPointer(dest)};
+ if (!source_ptr) {
+ LOG_ERROR(HW_GPU, "source_ptr is invalid");
+ return;
+ }
+
+ if (!dst_ptr) {
+ LOG_ERROR(HW_GPU, "dst_ptr is invalid");
+ return;
+ }
+
const auto FlushAndInvalidate = [&](u32 src_size, u64 dst_size) {
// TODO(Subv): For now, manually flush the regions until we implement GPU-accelerated
// copying.
diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp
index 66c690494..30b29e14d 100644
--- a/src/video_core/gpu.cpp
+++ b/src/video_core/gpu.cpp
@@ -12,6 +12,7 @@
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/maxwell_dma.h"
#include "video_core/gpu.h"
+#include "video_core/memory_manager.h"
#include "video_core/renderer_base.h"
namespace Tegra {
@@ -285,9 +286,10 @@ void GPU::ProcessSemaphoreTriggerMethod() {
// TODO(Kmather73): Generate a real GPU timestamp and write it here instead of
// CoreTiming
block.timestamp = Core::System::GetInstance().CoreTiming().GetTicks();
- memory_manager->WriteBlock(regs.smaphore_address.SmaphoreAddress(), &block, sizeof(block));
+ memory_manager->WriteBlock(regs.semaphore_address.SemaphoreAddress(), &block,
+ sizeof(block));
} else {
- const u32 word{memory_manager->Read32(regs.smaphore_address.SmaphoreAddress())};
+ const u32 word{memory_manager->Read<u32>(regs.semaphore_address.SemaphoreAddress())};
if ((op == GpuSemaphoreOperation::AcquireEqual && word == regs.semaphore_sequence) ||
(op == GpuSemaphoreOperation::AcquireGequal &&
static_cast<s32>(word - regs.semaphore_sequence) > 0) ||
@@ -314,11 +316,11 @@ void GPU::ProcessSemaphoreTriggerMethod() {
}
void GPU::ProcessSemaphoreRelease() {
- memory_manager->Write32(regs.smaphore_address.SmaphoreAddress(), regs.semaphore_release);
+ memory_manager->Write<u32>(regs.semaphore_address.SemaphoreAddress(), regs.semaphore_release);
}
void GPU::ProcessSemaphoreAcquire() {
- const u32 word = memory_manager->Read32(regs.smaphore_address.SmaphoreAddress());
+ const u32 word = memory_manager->Read<u32>(regs.semaphore_address.SemaphoreAddress());
const auto value = regs.semaphore_acquire;
if (word != value) {
regs.acquire_active = true;
diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h
index a14b95c30..de30ea354 100644
--- a/src/video_core/gpu.h
+++ b/src/video_core/gpu.h
@@ -9,7 +9,6 @@
#include "common/common_types.h"
#include "core/hle/service/nvflinger/buffer_queue.h"
#include "video_core/dma_pusher.h"
-#include "video_core/memory_manager.h"
using CacheAddr = std::uintptr_t;
inline CacheAddr ToCacheAddr(const void* host_ptr) {
@@ -124,6 +123,8 @@ enum class EngineID {
MAXWELL_DMA_COPY_A = 0xB0B5,
};
+class MemoryManager;
+
class GPU {
public:
explicit GPU(Core::System& system, VideoCore::RendererBase& renderer);
@@ -176,11 +177,11 @@ public:
u32 address_high;
u32 address_low;
- GPUVAddr SmaphoreAddress() const {
+ GPUVAddr SemaphoreAddress() const {
return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) |
address_low);
}
- } smaphore_address;
+ } semaphore_address;
u32 semaphore_sequence;
u32 semaphore_trigger;
@@ -244,9 +245,8 @@ protected:
private:
std::unique_ptr<Tegra::MemoryManager> memory_manager;
- /// Mapping of command subchannels to their bound engine ids.
+ /// Mapping of command subchannels to their bound engine ids
std::array<EngineID, 8> bound_engines = {};
-
/// 3D engine
std::unique_ptr<Engines::Maxwell3D> maxwell_3d;
/// 2D engine
@@ -263,7 +263,7 @@ private:
static_assert(offsetof(GPU::Regs, field_name) == position * 4, \
"Field " #field_name " has invalid position")
-ASSERT_REG_POSITION(smaphore_address, 0x4);
+ASSERT_REG_POSITION(semaphore_address, 0x4);
ASSERT_REG_POSITION(semaphore_sequence, 0x6);
ASSERT_REG_POSITION(semaphore_trigger, 0x7);
ASSERT_REG_POSITION(reference_count, 0x14);
diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp
index 8e8f36f28..e76b59842 100644
--- a/src/video_core/memory_manager.cpp
+++ b/src/video_core/memory_manager.cpp
@@ -5,218 +5,446 @@
#include "common/alignment.h"
#include "common/assert.h"
#include "common/logging/log.h"
+#include "core/core.h"
#include "core/memory.h"
+#include "video_core/gpu.h"
#include "video_core/memory_manager.h"
+#include "video_core/rasterizer_interface.h"
+#include "video_core/renderer_base.h"
namespace Tegra {
MemoryManager::MemoryManager() {
- // Mark the first page as reserved, so that 0 is not a valid GPUVAddr. Otherwise, games might
- // try to use 0 as a valid address, which is also used to mean nullptr. This fixes a bug with
- // Undertale using 0 for a render target.
- PageSlot(0) = static_cast<u64>(PageStatus::Reserved);
-}
-
-GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
- const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, align, PageStatus::Unmapped)};
+ std::fill(page_table.pointers.begin(), page_table.pointers.end(), nullptr);
+ std::fill(page_table.attributes.begin(), page_table.attributes.end(),
+ Common::PageType::Unmapped);
+ page_table.Resize(address_space_width);
- ASSERT_MSG(gpu_addr, "unable to find available GPU memory");
+ // Initialize the map with a single free region covering the entire managed space.
+ VirtualMemoryArea initial_vma;
+ initial_vma.size = address_space_end;
+ vma_map.emplace(initial_vma.base, initial_vma);
- for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
- VAddr& slot{PageSlot(*gpu_addr + offset)};
+ UpdatePageTableForVMA(initial_vma);
+}
- ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
+GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
+ const u64 aligned_size{Common::AlignUp(size, page_size)};
+ const GPUVAddr gpu_addr{FindFreeRegion(address_space_base, aligned_size)};
- slot = static_cast<u64>(PageStatus::Allocated);
- }
+ AllocateMemory(gpu_addr, 0, aligned_size);
- return *gpu_addr;
+ return gpu_addr;
}
GPUVAddr MemoryManager::AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align) {
- for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
- VAddr& slot{PageSlot(gpu_addr + offset)};
+ const u64 aligned_size{Common::AlignUp(size, page_size)};
- ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
-
- slot = static_cast<u64>(PageStatus::Allocated);
- }
+ AllocateMemory(gpu_addr, 0, aligned_size);
return gpu_addr;
}
GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, u64 size) {
- const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, PAGE_SIZE, PageStatus::Unmapped)};
+ const u64 aligned_size{Common::AlignUp(size, page_size)};
+ const GPUVAddr gpu_addr{FindFreeRegion(address_space_base, aligned_size)};
- ASSERT_MSG(gpu_addr, "unable to find available GPU memory");
+ MapBackingMemory(gpu_addr, Memory::GetPointer(cpu_addr), aligned_size, cpu_addr);
- for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
- VAddr& slot{PageSlot(*gpu_addr + offset)};
+ return gpu_addr;
+}
- ASSERT(slot == static_cast<u64>(PageStatus::Unmapped));
+GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size) {
+ ASSERT((gpu_addr & page_mask) == 0);
- slot = cpu_addr + offset;
- }
+ const u64 aligned_size{Common::AlignUp(size, page_size)};
- const MappedRegion region{cpu_addr, *gpu_addr, size};
- mapped_regions.push_back(region);
+ MapBackingMemory(gpu_addr, Memory::GetPointer(cpu_addr), aligned_size, cpu_addr);
- return *gpu_addr;
+ return gpu_addr;
}
-GPUVAddr MemoryManager::MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size) {
- ASSERT((gpu_addr & PAGE_MASK) == 0);
+GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) {
+ ASSERT((gpu_addr & page_mask) == 0);
- if (PageSlot(gpu_addr) != static_cast<u64>(PageStatus::Allocated)) {
- // Page has been already mapped. In this case, we must find a new area of memory to use that
- // is different than the specified one. Super Mario Odyssey hits this scenario when changing
- // areas, but we do not want to overwrite the old pages.
- // TODO(bunnei): We need to write a hardware test to confirm this behavior.
+ const u64 aligned_size{Common::AlignUp(size, page_size)};
+ const CacheAddr cache_addr{ToCacheAddr(GetPointer(gpu_addr))};
- LOG_ERROR(HW_GPU, "attempting to map addr 0x{:016X}, which is not available!", gpu_addr);
+ Core::System::GetInstance().Renderer().Rasterizer().FlushAndInvalidateRegion(cache_addr,
+ aligned_size);
+ UnmapRange(gpu_addr, aligned_size);
- const std::optional<GPUVAddr> new_gpu_addr{
- FindFreeBlock(gpu_addr, size, PAGE_SIZE, PageStatus::Allocated)};
+ return gpu_addr;
+}
- ASSERT_MSG(new_gpu_addr, "unable to find available GPU memory");
+GPUVAddr MemoryManager::FindFreeRegion(GPUVAddr region_start, u64 size) {
+ // Find the first Free VMA.
+ const VMAHandle vma_handle{std::find_if(vma_map.begin(), vma_map.end(), [&](const auto& vma) {
+ if (vma.second.type != VirtualMemoryArea::Type::Unmapped) {
+ return false;
+ }
- gpu_addr = *new_gpu_addr;
+ const VAddr vma_end{vma.second.base + vma.second.size};
+ return vma_end > region_start && vma_end >= region_start + size;
+ })};
+
+ if (vma_handle == vma_map.end()) {
+ return {};
}
- for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
- VAddr& slot{PageSlot(gpu_addr + offset)};
+ return std::max(region_start, vma_handle->second.base);
+}
- ASSERT(slot == static_cast<u64>(PageStatus::Allocated));
+bool MemoryManager::IsAddressValid(GPUVAddr addr) const {
+ return (addr >> page_bits) < page_table.pointers.size();
+}
- slot = cpu_addr + offset;
+std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr addr) {
+ if (!IsAddressValid(addr)) {
+ return {};
}
- const MappedRegion region{cpu_addr, gpu_addr, size};
- mapped_regions.push_back(region);
+ VAddr cpu_addr{page_table.backing_addr[addr >> page_bits]};
+ if (cpu_addr) {
+ return cpu_addr + (addr & page_mask);
+ }
- return gpu_addr;
+ return {};
}
-GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) {
- ASSERT((gpu_addr & PAGE_MASK) == 0);
+template <typename T>
+T MemoryManager::Read(GPUVAddr addr) {
+ if (!IsAddressValid(addr)) {
+ return {};
+ }
- for (u64 offset{}; offset < size; offset += PAGE_SIZE) {
- VAddr& slot{PageSlot(gpu_addr + offset)};
+ const u8* page_pointer{page_table.pointers[addr >> page_bits]};
+ if (page_pointer) {
+ // NOTE: Avoid adding any extra logic to this fast-path block
+ T value;
+ std::memcpy(&value, &page_pointer[addr & page_mask], sizeof(T));
+ return value;
+ }
- ASSERT(slot != static_cast<u64>(PageStatus::Allocated) &&
- slot != static_cast<u64>(PageStatus::Unmapped));
+ switch (page_table.attributes[addr >> page_bits]) {
+ case Common::PageType::Unmapped:
+ LOG_ERROR(HW_GPU, "Unmapped Read{} @ 0x{:08X}", sizeof(T) * 8, addr);
+ return 0;
+ case Common::PageType::Memory:
+ ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", addr);
+ break;
+ default:
+ UNREACHABLE();
+ }
+ return {};
+}
- slot = static_cast<u64>(PageStatus::Unmapped);
+template <typename T>
+void MemoryManager::Write(GPUVAddr addr, T data) {
+ if (!IsAddressValid(addr)) {
+ return;
}
- // Delete the region mappings that are contained within the unmapped region
- mapped_regions.erase(std::remove_if(mapped_regions.begin(), mapped_regions.end(),
- [&](const MappedRegion& region) {
- return region.gpu_addr <= gpu_addr &&
- region.gpu_addr + region.size < gpu_addr + size;
- }),
- mapped_regions.end());
- return gpu_addr;
-}
+ u8* page_pointer{page_table.pointers[addr >> page_bits]};
+ if (page_pointer) {
+ // NOTE: Avoid adding any extra logic to this fast-path block
+ std::memcpy(&page_pointer[addr & page_mask], &data, sizeof(T));
+ return;
+ }
-GPUVAddr MemoryManager::GetRegionEnd(GPUVAddr region_start) const {
- for (const auto& region : mapped_regions) {
- const GPUVAddr region_end{region.gpu_addr + region.size};
- if (region_start >= region.gpu_addr && region_start < region_end) {
- return region_end;
- }
+ switch (page_table.attributes[addr >> page_bits]) {
+ case Common::PageType::Unmapped:
+ LOG_ERROR(HW_GPU, "Unmapped Write{} 0x{:08X} @ 0x{:016X}", sizeof(data) * 8,
+ static_cast<u32>(data), addr);
+ return;
+ case Common::PageType::Memory:
+ ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", addr);
+ break;
+ default:
+ UNREACHABLE();
}
- return {};
}
-std::optional<GPUVAddr> MemoryManager::FindFreeBlock(GPUVAddr region_start, u64 size, u64 align,
- PageStatus status) {
- GPUVAddr gpu_addr{region_start};
- u64 free_space{};
- align = (align + PAGE_MASK) & ~PAGE_MASK;
-
- while (gpu_addr + free_space < MAX_ADDRESS) {
- if (PageSlot(gpu_addr + free_space) == static_cast<u64>(status)) {
- free_space += PAGE_SIZE;
- if (free_space >= size) {
- return gpu_addr;
- }
- } else {
- gpu_addr += free_space + PAGE_SIZE;
- free_space = 0;
- gpu_addr = Common::AlignUp(gpu_addr, align);
- }
+template u8 MemoryManager::Read<u8>(GPUVAddr addr);
+template u16 MemoryManager::Read<u16>(GPUVAddr addr);
+template u32 MemoryManager::Read<u32>(GPUVAddr addr);
+template u64 MemoryManager::Read<u64>(GPUVAddr addr);
+template void MemoryManager::Write<u8>(GPUVAddr addr, u8 data);
+template void MemoryManager::Write<u16>(GPUVAddr addr, u16 data);
+template void MemoryManager::Write<u32>(GPUVAddr addr, u32 data);
+template void MemoryManager::Write<u64>(GPUVAddr addr, u64 data);
+
+u8* MemoryManager::GetPointer(GPUVAddr addr) {
+ if (!IsAddressValid(addr)) {
+ return {};
}
+ u8* page_pointer{page_table.pointers[addr >> page_bits]};
+ if (page_pointer) {
+ return page_pointer + (addr & page_mask);
+ }
+
+ LOG_ERROR(HW_GPU, "Unknown GetPointer @ 0x{:016X}", addr);
return {};
}
-std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr gpu_addr) {
- const VAddr base_addr{PageSlot(gpu_addr)};
+void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) {
+ std::memcpy(dest_buffer, GetPointer(src_addr), size);
+}
+void MemoryManager::WriteBlock(GPUVAddr dest_addr, const void* src_buffer, std::size_t size) {
+ std::memcpy(GetPointer(dest_addr), src_buffer, size);
+}
+
+void MemoryManager::CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size) {
+ std::memcpy(GetPointer(dest_addr), GetPointer(src_addr), size);
+}
- if (base_addr == static_cast<u64>(PageStatus::Allocated) ||
- base_addr == static_cast<u64>(PageStatus::Unmapped) ||
- base_addr == static_cast<u64>(PageStatus::Reserved)) {
- return {};
+void MemoryManager::MapPages(GPUVAddr base, u64 size, u8* memory, Common::PageType type,
+ VAddr backing_addr) {
+ LOG_DEBUG(HW_GPU, "Mapping {} onto {:016X}-{:016X}", fmt::ptr(memory), base * page_size,
+ (base + size) * page_size);
+
+ const VAddr end{base + size};
+ ASSERT_MSG(end <= page_table.pointers.size(), "out of range mapping at {:016X}",
+ base + page_table.pointers.size());
+
+ std::fill(page_table.attributes.begin() + base, page_table.attributes.begin() + end, type);
+
+ if (memory == nullptr) {
+ std::fill(page_table.pointers.begin() + base, page_table.pointers.begin() + end, memory);
+ std::fill(page_table.backing_addr.begin() + base, page_table.backing_addr.begin() + end,
+ backing_addr);
+ } else {
+ while (base != end) {
+ page_table.pointers[base] = memory;
+ page_table.backing_addr[base] = backing_addr;
+
+ base += 1;
+ memory += page_size;
+ backing_addr += page_size;
+ }
}
+}
- return base_addr + (gpu_addr & PAGE_MASK);
+void MemoryManager::MapMemoryRegion(GPUVAddr base, u64 size, u8* target, VAddr backing_addr) {
+ ASSERT_MSG((size & page_mask) == 0, "non-page aligned size: {:016X}", size);
+ ASSERT_MSG((base & page_mask) == 0, "non-page aligned base: {:016X}", base);
+ MapPages(base / page_size, size / page_size, target, Common::PageType::Memory, backing_addr);
}
-u8 MemoryManager::Read8(GPUVAddr addr) {
- return Memory::Read8(*GpuToCpuAddress(addr));
+void MemoryManager::UnmapRegion(GPUVAddr base, u64 size) {
+ ASSERT_MSG((size & page_mask) == 0, "non-page aligned size: {:016X}", size);
+ ASSERT_MSG((base & page_mask) == 0, "non-page aligned base: {:016X}", base);
+ MapPages(base / page_size, size / page_size, nullptr, Common::PageType::Unmapped);
}
-u16 MemoryManager::Read16(GPUVAddr addr) {
- return Memory::Read16(*GpuToCpuAddress(addr));
+bool VirtualMemoryArea::CanBeMergedWith(const VirtualMemoryArea& next) const {
+ ASSERT(base + size == next.base);
+ if (type != next.type) {
+ return {};
+ }
+ if (type == VirtualMemoryArea::Type::Allocated && (offset + size != next.offset)) {
+ return {};
+ }
+ if (type == VirtualMemoryArea::Type::Mapped && backing_memory + size != next.backing_memory) {
+ return {};
+ }
+ return true;
}
-u32 MemoryManager::Read32(GPUVAddr addr) {
- return Memory::Read32(*GpuToCpuAddress(addr));
+MemoryManager::VMAHandle MemoryManager::FindVMA(GPUVAddr target) const {
+ if (target >= address_space_end) {
+ return vma_map.end();
+ } else {
+ return std::prev(vma_map.upper_bound(target));
+ }
}
-u64 MemoryManager::Read64(GPUVAddr addr) {
- return Memory::Read64(*GpuToCpuAddress(addr));
+MemoryManager::VMAIter MemoryManager::Allocate(VMAIter vma_handle) {
+ VirtualMemoryArea& vma{vma_handle->second};
+
+ vma.type = VirtualMemoryArea::Type::Allocated;
+ vma.backing_addr = 0;
+ vma.backing_memory = {};
+ UpdatePageTableForVMA(vma);
+
+ return MergeAdjacent(vma_handle);
}
-void MemoryManager::Write8(GPUVAddr addr, u8 data) {
- Memory::Write8(*GpuToCpuAddress(addr), data);
+MemoryManager::VMAHandle MemoryManager::AllocateMemory(GPUVAddr target, std::size_t offset,
+ u64 size) {
+
+ // This is the appropriately sized VMA that will turn into our allocation.
+ VMAIter vma_handle{CarveVMA(target, size)};
+ VirtualMemoryArea& vma{vma_handle->second};
+
+ ASSERT(vma.size == size);
+
+ vma.offset = offset;
+
+ return Allocate(vma_handle);
}
-void MemoryManager::Write16(GPUVAddr addr, u16 data) {
- Memory::Write16(*GpuToCpuAddress(addr), data);
+MemoryManager::VMAHandle MemoryManager::MapBackingMemory(GPUVAddr target, u8* memory, u64 size,
+ VAddr backing_addr) {
+ // This is the appropriately sized VMA that will turn into our allocation.
+ VMAIter vma_handle{CarveVMA(target, size)};
+ VirtualMemoryArea& vma{vma_handle->second};
+
+ ASSERT(vma.size == size);
+
+ vma.type = VirtualMemoryArea::Type::Mapped;
+ vma.backing_memory = memory;
+ vma.backing_addr = backing_addr;
+ UpdatePageTableForVMA(vma);
+
+ return MergeAdjacent(vma_handle);
}
-void MemoryManager::Write32(GPUVAddr addr, u32 data) {
- Memory::Write32(*GpuToCpuAddress(addr), data);
+void MemoryManager::UnmapRange(GPUVAddr target, u64 size) {
+ VMAIter vma{CarveVMARange(target, size)};
+ const VAddr target_end{target + size};
+ const VMAIter end{vma_map.end()};
+
+ // The comparison against the end of the range must be done using addresses since VMAs can be
+ // merged during this process, causing invalidation of the iterators.
+ while (vma != end && vma->second.base < target_end) {
+ // Unmapped ranges return to allocated state and can be reused
+ // This behavior is used by Super Mario Odyssey, Sonic Forces, and likely other games
+ vma = std::next(Allocate(vma));
+ }
+
+ ASSERT(FindVMA(target)->second.size >= size);
}
-void MemoryManager::Write64(GPUVAddr addr, u64 data) {
- Memory::Write64(*GpuToCpuAddress(addr), data);
+MemoryManager::VMAIter MemoryManager::StripIterConstness(const VMAHandle& iter) {
+ // This uses a neat C++ trick to convert a const_iterator to a regular iterator, given
+ // non-const access to its container.
+ return vma_map.erase(iter, iter); // Erases an empty range of elements
}
-u8* MemoryManager::GetPointer(GPUVAddr addr) {
- return Memory::GetPointer(*GpuToCpuAddress(addr));
+MemoryManager::VMAIter MemoryManager::CarveVMA(GPUVAddr base, u64 size) {
+ ASSERT_MSG((size & page_mask) == 0, "non-page aligned size: 0x{:016X}", size);
+ ASSERT_MSG((base & page_mask) == 0, "non-page aligned base: 0x{:016X}", base);
+
+ VMAIter vma_handle{StripIterConstness(FindVMA(base))};
+ if (vma_handle == vma_map.end()) {
+ // Target address is outside the managed range
+ return {};
+ }
+
+ const VirtualMemoryArea& vma{vma_handle->second};
+ if (vma.type == VirtualMemoryArea::Type::Mapped) {
+ // Region is already allocated
+ return {};
+ }
+
+ const VAddr start_in_vma{base - vma.base};
+ const VAddr end_in_vma{start_in_vma + size};
+
+ ASSERT_MSG(end_in_vma <= vma.size, "region size 0x{:016X} is less than required size 0x{:016X}",
+ vma.size, end_in_vma);
+
+ if (end_in_vma < vma.size) {
+ // Split VMA at the end of the allocated region
+ SplitVMA(vma_handle, end_in_vma);
+ }
+ if (start_in_vma != 0) {
+ // Split VMA at the start of the allocated region
+ vma_handle = SplitVMA(vma_handle, start_in_vma);
+ }
+
+ return vma_handle;
}
-void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) {
- std::memcpy(dest_buffer, GetPointer(src_addr), size);
+MemoryManager::VMAIter MemoryManager::CarveVMARange(GPUVAddr target, u64 size) {
+ ASSERT_MSG((size & page_mask) == 0, "non-page aligned size: 0x{:016X}", size);
+ ASSERT_MSG((target & page_mask) == 0, "non-page aligned base: 0x{:016X}", target);
+
+ const VAddr target_end{target + size};
+ ASSERT(target_end >= target);
+ ASSERT(size > 0);
+
+ VMAIter begin_vma{StripIterConstness(FindVMA(target))};
+ const VMAIter i_end{vma_map.lower_bound(target_end)};
+ if (std::any_of(begin_vma, i_end, [](const auto& entry) {
+ return entry.second.type == VirtualMemoryArea::Type::Unmapped;
+ })) {
+ return {};
+ }
+
+ if (target != begin_vma->second.base) {
+ begin_vma = SplitVMA(begin_vma, target - begin_vma->second.base);
+ }
+
+ VMAIter end_vma{StripIterConstness(FindVMA(target_end))};
+ if (end_vma != vma_map.end() && target_end != end_vma->second.base) {
+ end_vma = SplitVMA(end_vma, target_end - end_vma->second.base);
+ }
+
+ return begin_vma;
}
-void MemoryManager::WriteBlock(GPUVAddr dest_addr, const void* src_buffer, std::size_t size) {
- std::memcpy(GetPointer(dest_addr), src_buffer, size);
+
+MemoryManager::VMAIter MemoryManager::SplitVMA(VMAIter vma_handle, u64 offset_in_vma) {
+ VirtualMemoryArea& old_vma{vma_handle->second};
+ VirtualMemoryArea new_vma{old_vma}; // Make a copy of the VMA
+
+ // For now, don't allow no-op VMA splits (trying to split at a boundary) because it's probably
+ // a bug. This restriction might be removed later.
+ ASSERT(offset_in_vma < old_vma.size);
+ ASSERT(offset_in_vma > 0);
+
+ old_vma.size = offset_in_vma;
+ new_vma.base += offset_in_vma;
+ new_vma.size -= offset_in_vma;
+
+ switch (new_vma.type) {
+ case VirtualMemoryArea::Type::Unmapped:
+ break;
+ case VirtualMemoryArea::Type::Allocated:
+ new_vma.offset += offset_in_vma;
+ break;
+ case VirtualMemoryArea::Type::Mapped:
+ new_vma.backing_memory += offset_in_vma;
+ break;
+ }
+
+ ASSERT(old_vma.CanBeMergedWith(new_vma));
+
+ return vma_map.emplace_hint(std::next(vma_handle), new_vma.base, new_vma);
}
-void MemoryManager::CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size) {
- std::memcpy(GetPointer(dest_addr), GetPointer(src_addr), size);
+MemoryManager::VMAIter MemoryManager::MergeAdjacent(VMAIter iter) {
+ const VMAIter next_vma{std::next(iter)};
+ if (next_vma != vma_map.end() && iter->second.CanBeMergedWith(next_vma->second)) {
+ iter->second.size += next_vma->second.size;
+ vma_map.erase(next_vma);
+ }
+
+ if (iter != vma_map.begin()) {
+ VMAIter prev_vma{std::prev(iter)};
+ if (prev_vma->second.CanBeMergedWith(iter->second)) {
+ prev_vma->second.size += iter->second.size;
+ vma_map.erase(iter);
+ iter = prev_vma;
+ }
+ }
+
+ return iter;
}
-VAddr& MemoryManager::PageSlot(GPUVAddr gpu_addr) {
- auto& block{page_table[(gpu_addr >> (PAGE_BITS + PAGE_TABLE_BITS)) & PAGE_TABLE_MASK]};
- if (!block) {
- block = std::make_unique<PageBlock>();
- block->fill(static_cast<VAddr>(PageStatus::Unmapped));
+void MemoryManager::UpdatePageTableForVMA(const VirtualMemoryArea& vma) {
+ switch (vma.type) {
+ case VirtualMemoryArea::Type::Unmapped:
+ UnmapRegion(vma.base, vma.size);
+ break;
+ case VirtualMemoryArea::Type::Allocated:
+ MapMemoryRegion(vma.base, vma.size, nullptr, vma.backing_addr);
+ break;
+ case VirtualMemoryArea::Type::Mapped:
+ MapMemoryRegion(vma.base, vma.size, vma.backing_memory, vma.backing_addr);
+ break;
}
- return (*block)[(gpu_addr >> PAGE_BITS) & PAGE_BLOCK_MASK];
}
} // namespace Tegra
diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h
index 425e2f31c..34744bb27 100644
--- a/src/video_core/memory_manager.h
+++ b/src/video_core/memory_manager.h
@@ -1,82 +1,148 @@
-// Copyright 2018 yuzu emulator team
+// Copyright 2018 yuzu emulator team
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
-#include <array>
-#include <memory>
+#include <map>
#include <optional>
-#include <vector>
#include "common/common_types.h"
+#include "common/page_table.h"
namespace Tegra {
-/// Virtual addresses in the GPU's memory map are 64 bit.
-using GPUVAddr = u64;
+/**
+ * Represents a VMA in an address space. A VMA is a contiguous region of virtual addressing space
+ * with homogeneous attributes across its extents. In this particular implementation each VMA is
+ * also backed by a single host memory allocation.
+ */
+struct VirtualMemoryArea {
+ enum class Type : u8 {
+ Unmapped,
+ Allocated,
+ Mapped,
+ };
+
+ /// Virtual base address of the region.
+ GPUVAddr base{};
+ /// Size of the region.
+ u64 size{};
+ /// Memory area mapping type.
+ Type type{Type::Unmapped};
+ /// CPU memory mapped address corresponding to this memory area.
+ VAddr backing_addr{};
+ /// Offset into the backing_memory the mapping starts from.
+ std::size_t offset{};
+ /// Pointer backing this VMA.
+ u8* backing_memory{};
+
+ /// Tests if this area can be merged to the right with `next`.
+ bool CanBeMergedWith(const VirtualMemoryArea& next) const;
+};
class MemoryManager final {
public:
MemoryManager();
GPUVAddr AllocateSpace(u64 size, u64 align);
- GPUVAddr AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align);
+ GPUVAddr AllocateSpace(GPUVAddr addr, u64 size, u64 align);
GPUVAddr MapBufferEx(VAddr cpu_addr, u64 size);
- GPUVAddr MapBufferEx(VAddr cpu_addr, GPUVAddr gpu_addr, u64 size);
- GPUVAddr UnmapBuffer(GPUVAddr gpu_addr, u64 size);
- GPUVAddr GetRegionEnd(GPUVAddr region_start) const;
- std::optional<VAddr> GpuToCpuAddress(GPUVAddr gpu_addr);
-
- static constexpr u64 PAGE_BITS = 16;
- static constexpr u64 PAGE_SIZE = 1 << PAGE_BITS;
- static constexpr u64 PAGE_MASK = PAGE_SIZE - 1;
+ GPUVAddr MapBufferEx(VAddr cpu_addr, GPUVAddr addr, u64 size);
+ GPUVAddr UnmapBuffer(GPUVAddr addr, u64 size);
+ std::optional<VAddr> GpuToCpuAddress(GPUVAddr addr);
- u8 Read8(GPUVAddr addr);
- u16 Read16(GPUVAddr addr);
- u32 Read32(GPUVAddr addr);
- u64 Read64(GPUVAddr addr);
+ template <typename T>
+ T Read(GPUVAddr addr);
- void Write8(GPUVAddr addr, u8 data);
- void Write16(GPUVAddr addr, u16 data);
- void Write32(GPUVAddr addr, u32 data);
- void Write64(GPUVAddr addr, u64 data);
+ template <typename T>
+ void Write(GPUVAddr addr, T data);
- u8* GetPointer(GPUVAddr vaddr);
+ u8* GetPointer(GPUVAddr addr);
void ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size);
void WriteBlock(GPUVAddr dest_addr, const void* src_buffer, std::size_t size);
- void CopyBlock(VAddr dest_addr, VAddr src_addr, std::size_t size);
+ void CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size);
private:
- enum class PageStatus : u64 {
- Unmapped = 0xFFFFFFFFFFFFFFFFULL,
- Allocated = 0xFFFFFFFFFFFFFFFEULL,
- Reserved = 0xFFFFFFFFFFFFFFFDULL,
- };
-
- std::optional<GPUVAddr> FindFreeBlock(GPUVAddr region_start, u64 size, u64 align,
- PageStatus status);
- VAddr& PageSlot(GPUVAddr gpu_addr);
-
- static constexpr u64 MAX_ADDRESS{0x10000000000ULL};
- static constexpr u64 PAGE_TABLE_BITS{10};
- static constexpr u64 PAGE_TABLE_SIZE{1 << PAGE_TABLE_BITS};
- static constexpr u64 PAGE_TABLE_MASK{PAGE_TABLE_SIZE - 1};
- static constexpr u64 PAGE_BLOCK_BITS{14};
- static constexpr u64 PAGE_BLOCK_SIZE{1 << PAGE_BLOCK_BITS};
- static constexpr u64 PAGE_BLOCK_MASK{PAGE_BLOCK_SIZE - 1};
-
- using PageBlock = std::array<VAddr, PAGE_BLOCK_SIZE>;
- std::array<std::unique_ptr<PageBlock>, PAGE_TABLE_SIZE> page_table{};
-
- struct MappedRegion {
- VAddr cpu_addr;
- GPUVAddr gpu_addr;
- u64 size;
- };
+ using VMAMap = std::map<GPUVAddr, VirtualMemoryArea>;
+ using VMAHandle = VMAMap::const_iterator;
+ using VMAIter = VMAMap::iterator;
+
+ bool IsAddressValid(GPUVAddr addr) const;
+ void MapPages(GPUVAddr base, u64 size, u8* memory, Common::PageType type,
+ VAddr backing_addr = 0);
+ void MapMemoryRegion(GPUVAddr base, u64 size, u8* target, VAddr backing_addr);
+ void UnmapRegion(GPUVAddr base, u64 size);
+
+ /// Finds the VMA in which the given address is included in, or `vma_map.end()`.
+ VMAHandle FindVMA(GPUVAddr target) const;
+
+ VMAHandle AllocateMemory(GPUVAddr target, std::size_t offset, u64 size);
+
+ /**
+ * Maps an unmanaged host memory pointer at a given address.
+ *
+ * @param target The guest address to start the mapping at.
+ * @param memory The memory to be mapped.
+ * @param size Size of the mapping.
+ * @param state MemoryState tag to attach to the VMA.
+ */
+ VMAHandle MapBackingMemory(GPUVAddr target, u8* memory, u64 size, VAddr backing_addr);
+
+ /// Unmaps a range of addresses, splitting VMAs as necessary.
+ void UnmapRange(GPUVAddr target, u64 size);
+
+ /// Converts a VMAHandle to a mutable VMAIter.
+ VMAIter StripIterConstness(const VMAHandle& iter);
+
+ /// Marks as the specfied VMA as allocated.
+ VMAIter Allocate(VMAIter vma);
+
+ /**
+ * Carves a VMA of a specific size at the specified address by splitting Free VMAs while doing
+ * the appropriate error checking.
+ */
+ VMAIter CarveVMA(GPUVAddr base, u64 size);
+
+ /**
+ * Splits the edges of the given range of non-Free VMAs so that there is a VMA split at each
+ * end of the range.
+ */
+ VMAIter CarveVMARange(GPUVAddr base, u64 size);
+
+ /**
+ * Splits a VMA in two, at the specified offset.
+ * @returns the right side of the split, with the original iterator becoming the left side.
+ */
+ VMAIter SplitVMA(VMAIter vma, u64 offset_in_vma);
+
+ /**
+ * Checks for and merges the specified VMA with adjacent ones if possible.
+ * @returns the merged VMA or the original if no merging was possible.
+ */
+ VMAIter MergeAdjacent(VMAIter vma);
+
+ /// Updates the pages corresponding to this VMA so they match the VMA's attributes.
+ void UpdatePageTableForVMA(const VirtualMemoryArea& vma);
+
+ /// Finds a free (unmapped region) of the specified size starting at the specified address.
+ GPUVAddr FindFreeRegion(GPUVAddr region_start, u64 size);
- std::vector<MappedRegion> mapped_regions;
+private:
+ static constexpr u64 page_bits{16};
+ static constexpr u64 page_size{1 << page_bits};
+ static constexpr u64 page_mask{page_size - 1};
+
+ /// Address space in bits, this is fairly arbitrary but sufficiently large.
+ static constexpr u32 address_space_width{39};
+ /// Start address for mapping, this is fairly arbitrary but must be non-zero.
+ static constexpr GPUVAddr address_space_base{0x100000};
+ /// End of address space, based on address space in bits.
+ static constexpr GPUVAddr address_space_end{1ULL << address_space_width};
+
+ Common::PageTable page_table{page_bits};
+ VMAMap vma_map;
};
} // namespace Tegra
diff --git a/src/video_core/rasterizer_cache.h b/src/video_core/rasterizer_cache.h
index ecd9986a0..9fc9f3056 100644
--- a/src/video_core/rasterizer_cache.h
+++ b/src/video_core/rasterizer_cache.h
@@ -132,7 +132,7 @@ protected:
}
/// Register an object into the cache
- void Register(const T& object) {
+ virtual void Register(const T& object) {
std::lock_guard<std::recursive_mutex> lock{mutex};
object->SetIsRegistered(true);
@@ -142,7 +142,7 @@ protected:
}
/// Unregisters an object from the cache
- void Unregister(const T& object) {
+ virtual void Unregister(const T& object) {
std::lock_guard<std::recursive_mutex> lock{mutex};
object->SetIsRegistered(false);
diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h
index 76e292e87..d7b86df38 100644
--- a/src/video_core/rasterizer_interface.h
+++ b/src/video_core/rasterizer_interface.h
@@ -9,7 +9,6 @@
#include "common/common_types.h"
#include "video_core/engines/fermi_2d.h"
#include "video_core/gpu.h"
-#include "video_core/memory_manager.h"
namespace VideoCore {
diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp
index 5048ed6ce..f75c65825 100644
--- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp
@@ -21,8 +21,8 @@ CachedBufferEntry::CachedBufferEntry(VAddr cpu_addr, std::size_t size, GLintptr
OGLBufferCache::OGLBufferCache(RasterizerOpenGL& rasterizer, std::size_t size)
: RasterizerCache{rasterizer}, stream_buffer(size, true) {}
-GLintptr OGLBufferCache::UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size,
- std::size_t alignment, bool cache) {
+GLintptr OGLBufferCache::UploadMemory(GPUVAddr gpu_addr, std::size_t size, std::size_t alignment,
+ bool cache) {
auto& memory_manager = Core::System::GetInstance().GPU().MemoryManager();
// Cache management is a big overhead, so only cache entries with a given size.
diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h
index 1de1f84ae..fc33aa433 100644
--- a/src/video_core/renderer_opengl/gl_buffer_cache.h
+++ b/src/video_core/renderer_opengl/gl_buffer_cache.h
@@ -58,7 +58,7 @@ public:
/// Uploads data from a guest GPU address. Returns host's buffer offset where it's been
/// allocated.
- GLintptr UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size, std::size_t alignment = 4,
+ GLintptr UploadMemory(GPUVAddr gpu_addr, std::size_t size, std::size_t alignment = 4,
bool cache = true);
/// Uploads from a host memory. Returns host's buffer offset where it's been allocated.
diff --git a/src/video_core/renderer_opengl/gl_global_cache.cpp b/src/video_core/renderer_opengl/gl_global_cache.cpp
index c8dbcacbd..0fbfbad55 100644
--- a/src/video_core/renderer_opengl/gl_global_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_global_cache.cpp
@@ -46,7 +46,7 @@ GlobalRegion GlobalRegionCacheOpenGL::TryGetReservedGlobalRegion(CacheAddr addr,
return search->second;
}
-GlobalRegion GlobalRegionCacheOpenGL::GetUncachedGlobalRegion(Tegra::GPUVAddr addr, u32 size,
+GlobalRegion GlobalRegionCacheOpenGL::GetUncachedGlobalRegion(GPUVAddr addr, u32 size,
u8* host_ptr) {
GlobalRegion region{TryGetReservedGlobalRegion(ToCacheAddr(host_ptr), size)};
if (!region) {
@@ -76,8 +76,8 @@ GlobalRegion GlobalRegionCacheOpenGL::GetGlobalRegion(
const auto cbufs{gpu.Maxwell3D().state.shader_stages[static_cast<u64>(stage)]};
const auto addr{cbufs.const_buffers[global_region.GetCbufIndex()].address +
global_region.GetCbufOffset()};
- const auto actual_addr{memory_manager.Read64(addr)};
- const auto size{memory_manager.Read32(addr + 8)};
+ const auto actual_addr{memory_manager.Read<u64>(addr)};
+ const auto size{memory_manager.Read<u32>(addr + 8)};
// Look up global region in the cache based on address
const auto& host_ptr{memory_manager.GetPointer(actual_addr)};
diff --git a/src/video_core/renderer_opengl/gl_global_cache.h b/src/video_core/renderer_opengl/gl_global_cache.h
index a840491f7..5a21ab66f 100644
--- a/src/video_core/renderer_opengl/gl_global_cache.h
+++ b/src/video_core/renderer_opengl/gl_global_cache.h
@@ -66,7 +66,7 @@ public:
private:
GlobalRegion TryGetReservedGlobalRegion(CacheAddr addr, u32 size) const;
- GlobalRegion GetUncachedGlobalRegion(Tegra::GPUVAddr addr, u32 size, u8* host_ptr);
+ GlobalRegion GetUncachedGlobalRegion(GPUVAddr addr, u32 size, u8* host_ptr);
void ReserveGlobalRegion(GlobalRegion region);
std::unordered_map<CacheAddr, GlobalRegion> reserve;
diff --git a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp
index 75d816795..2bcbd3da2 100644
--- a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp
+++ b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp
@@ -40,8 +40,7 @@ GLintptr PrimitiveAssembler::MakeQuadArray(u32 first, u32 count) {
return index_offset;
}
-GLintptr PrimitiveAssembler::MakeQuadIndexed(Tegra::GPUVAddr gpu_addr, std::size_t index_size,
- u32 count) {
+GLintptr PrimitiveAssembler::MakeQuadIndexed(GPUVAddr gpu_addr, std::size_t index_size, u32 count) {
const std::size_t map_size{CalculateQuadSize(count)};
auto [dst_pointer, index_offset] = buffer_cache.ReserveMemory(map_size);
diff --git a/src/video_core/renderer_opengl/gl_primitive_assembler.h b/src/video_core/renderer_opengl/gl_primitive_assembler.h
index a8cb88eb5..0e2e7dc36 100644
--- a/src/video_core/renderer_opengl/gl_primitive_assembler.h
+++ b/src/video_core/renderer_opengl/gl_primitive_assembler.h
@@ -24,7 +24,7 @@ public:
GLintptr MakeQuadArray(u32 first, u32 count);
- GLintptr MakeQuadIndexed(Tegra::GPUVAddr gpu_addr, std::size_t index_size, u32 count);
+ GLintptr MakeQuadIndexed(GPUVAddr gpu_addr, std::size_t index_size, u32 count);
private:
OGLBufferCache& buffer_cache;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 198c54872..e06dfe43f 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -225,8 +225,8 @@ void RasterizerOpenGL::SetupVertexBuffer(GLuint vao) {
if (!vertex_array.IsEnabled())
continue;
- const Tegra::GPUVAddr start = vertex_array.StartAddress();
- const Tegra::GPUVAddr end = regs.vertex_array_limit[index].LimitAddress();
+ const GPUVAddr start = vertex_array.StartAddress();
+ const GPUVAddr end = regs.vertex_array_limit[index].LimitAddress();
ASSERT(end > start);
const u64 size = end - start + 1;
@@ -421,8 +421,8 @@ std::size_t RasterizerOpenGL::CalculateVertexArraysSize() const {
if (!regs.vertex_array[index].IsEnabled())
continue;
- const Tegra::GPUVAddr start = regs.vertex_array[index].StartAddress();
- const Tegra::GPUVAddr end = regs.vertex_array_limit[index].LimitAddress();
+ const GPUVAddr start = regs.vertex_array[index].StartAddress();
+ const GPUVAddr end = regs.vertex_array_limit[index].LimitAddress();
ASSERT(end > start);
size += end - start + 1;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
index 57329cd61..0235317c0 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
@@ -55,7 +55,7 @@ static void ApplyTextureDefaults(GLuint texture, u32 max_mip_level) {
}
}
-void SurfaceParams::InitCacheParameters(Tegra::GPUVAddr gpu_addr_) {
+void SurfaceParams::InitCacheParameters(GPUVAddr gpu_addr_) {
auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()};
gpu_addr = gpu_addr_;
@@ -222,7 +222,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only,
}
/*static*/ SurfaceParams SurfaceParams::CreateForDepthBuffer(
- u32 zeta_width, u32 zeta_height, Tegra::GPUVAddr zeta_address, Tegra::DepthFormat format,
+ u32 zeta_width, u32 zeta_height, GPUVAddr zeta_address, Tegra::DepthFormat format,
u32 block_width, u32 block_height, u32 block_depth,
Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout type) {
SurfaceParams params{};
@@ -564,6 +564,12 @@ void RasterizerCacheOpenGL::CopySurface(const Surface& src_surface, const Surfac
CachedSurface::CachedSurface(const SurfaceParams& params)
: params{params}, gl_target{SurfaceTargetToGL(params.target)},
cached_size_in_bytes{params.size_in_bytes}, RasterizerCacheObject{params.host_ptr} {
+
+ const auto optional_cpu_addr{
+ Core::System::GetInstance().GPU().MemoryManager().GpuToCpuAddress(params.gpu_addr)};
+ ASSERT_MSG(optional_cpu_addr, "optional_cpu_addr is invalid");
+ cpu_addr = *optional_cpu_addr;
+
texture.Create(gl_target);
// TODO(Rodrigo): Using params.GetRect() returns a different size than using its Mip*(0)
@@ -603,20 +609,6 @@ CachedSurface::CachedSurface(const SurfaceParams& params)
ApplyTextureDefaults(texture.handle, params.max_mip_level);
OpenGL::LabelGLObject(GL_TEXTURE, texture.handle, params.gpu_addr, params.IdentityString());
-
- // Clamp size to mapped GPU memory region
- // TODO(bunnei): Super Mario Odyssey maps a 0x40000 byte region and then uses it for a 0x80000
- // R32F render buffer. We do not yet know if this is a game bug or something else, but this
- // check is necessary to prevent flushing from overwriting unmapped memory.
-
- auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()};
- const u64 max_size{memory_manager.GetRegionEnd(params.gpu_addr) - params.gpu_addr};
- if (cached_size_in_bytes > max_size) {
- LOG_ERROR(HW_GPU, "Surface size {} exceeds region size {}", params.size_in_bytes, max_size);
- cached_size_in_bytes = max_size;
- }
-
- cpu_addr = *memory_manager.GpuToCpuAddress(params.gpu_addr);
}
MICROPROFILE_DEFINE(OpenGL_SurfaceLoad, "OpenGL", "Surface Load", MP_RGB(128, 192, 64));
@@ -925,7 +917,7 @@ void RasterizerCacheOpenGL::LoadSurface(const Surface& surface) {
}
Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool preserve_contents) {
- if (params.gpu_addr == 0 || params.height * params.width == 0) {
+ if (!params.IsValid()) {
return {};
}
@@ -941,7 +933,7 @@ Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool pres
// If surface parameters changed and we care about keeping the previous data, recreate
// the surface from the old one
Surface new_surface{RecreateSurface(surface, params)};
- UnregisterSurface(surface);
+ Unregister(surface);
Register(new_surface);
if (new_surface->IsUploaded()) {
RegisterReinterpretSurface(new_surface);
@@ -949,7 +941,7 @@ Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool pres
return new_surface;
} else {
// Delete the old surface before creating a new one to prevent collisions.
- UnregisterSurface(surface);
+ Unregister(surface);
}
}
@@ -980,11 +972,11 @@ void RasterizerCacheOpenGL::FastLayeredCopySurface(const Surface& src_surface,
const auto& init_params{src_surface->GetSurfaceParams()};
const auto& dst_params{dst_surface->GetSurfaceParams()};
auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()};
- Tegra::GPUVAddr address{init_params.gpu_addr};
+ GPUVAddr address{init_params.gpu_addr};
const std::size_t layer_size{dst_params.LayerMemorySize()};
for (u32 layer = 0; layer < dst_params.depth; layer++) {
for (u32 mipmap = 0; mipmap < dst_params.max_mip_level; mipmap++) {
- const Tegra::GPUVAddr sub_address{address + dst_params.GetMipmapLevelOffset(mipmap)};
+ const GPUVAddr sub_address{address + dst_params.GetMipmapLevelOffset(mipmap)};
const Surface& copy{TryGet(memory_manager.GetPointer(sub_address))};
if (!copy) {
continue;
@@ -1244,10 +1236,9 @@ static std::optional<u32> TryFindBestMipMap(std::size_t memory, const SurfacePar
return {};
}
-static std::optional<u32> TryFindBestLayer(Tegra::GPUVAddr addr, const SurfaceParams params,
- u32 mipmap) {
+static std::optional<u32> TryFindBestLayer(GPUVAddr addr, const SurfaceParams params, u32 mipmap) {
const std::size_t size{params.LayerMemorySize()};
- Tegra::GPUVAddr start{params.gpu_addr + params.GetMipmapLevelOffset(mipmap)};
+ GPUVAddr start{params.gpu_addr + params.GetMipmapLevelOffset(mipmap)};
for (u32 i = 0; i < params.depth; i++) {
if (start == addr) {
return {i};
@@ -1304,12 +1295,12 @@ static bool IsReinterpretInvalidSecond(const Surface render_surface,
bool RasterizerCacheOpenGL::PartialReinterpretSurface(Surface triggering_surface,
Surface intersect) {
if (IsReinterpretInvalid(triggering_surface, intersect)) {
- UnregisterSurface(intersect);
+ Unregister(intersect);
return false;
}
if (!LayerFitReinterpretSurface(*this, triggering_surface, intersect)) {
if (IsReinterpretInvalidSecond(triggering_surface, intersect)) {
- UnregisterSurface(intersect);
+ Unregister(intersect);
return false;
}
FlushObject(intersect);
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
index 9366f47f2..e8073579f 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
@@ -109,6 +109,11 @@ struct SurfaceParams {
return size;
}
+ /// Returns true if the parameters constitute a valid rasterizer surface.
+ bool IsValid() const {
+ return gpu_addr && host_ptr && height && width;
+ }
+
/// Returns the exact size of the memory occupied by a layer in a texture in VRAM, including
/// mipmaps.
std::size_t LayerMemorySize() const {
@@ -210,7 +215,7 @@ struct SurfaceParams {
/// Creates SurfaceParams for a depth buffer configuration
static SurfaceParams CreateForDepthBuffer(
- u32 zeta_width, u32 zeta_height, Tegra::GPUVAddr zeta_address, Tegra::DepthFormat format,
+ u32 zeta_width, u32 zeta_height, GPUVAddr zeta_address, Tegra::DepthFormat format,
u32 block_width, u32 block_height, u32 block_depth,
Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout type);
@@ -232,7 +237,7 @@ struct SurfaceParams {
}
/// Initializes parameters for caching, should be called after everything has been initialized
- void InitCacheParameters(Tegra::GPUVAddr gpu_addr);
+ void InitCacheParameters(GPUVAddr gpu_addr);
std::string TargetName() const {
switch (target) {
@@ -297,7 +302,7 @@ struct SurfaceParams {
bool srgb_conversion;
// Parameters used for caching
u8* host_ptr;
- Tegra::GPUVAddr gpu_addr;
+ GPUVAddr gpu_addr;
std::size_t size_in_bytes;
std::size_t size_in_bytes_gl;
@@ -533,13 +538,17 @@ private:
return nullptr;
}
+ void Register(const Surface& object) override {
+ RasterizerCache<Surface>::Register(object);
+ }
+
/// Unregisters an object from the cache
- void UnregisterSurface(const Surface& object) {
+ void Unregister(const Surface& object) override {
if (object->IsReinterpreted()) {
auto interval = GetReinterpretInterval(object);
reinterpreted_surfaces.erase(interval);
}
- Unregister(object);
+ RasterizerCache<Surface>::Unregister(object);
}
};
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index 1ed740877..1f8eca6f0 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -32,7 +32,7 @@ struct UnspecializedShader {
namespace {
/// Gets the address for the specified shader stage program
-Tegra::GPUVAddr GetShaderAddress(Maxwell::ShaderProgram program) {
+GPUVAddr GetShaderAddress(Maxwell::ShaderProgram program) {
const auto& gpu{Core::System::GetInstance().GPU().Maxwell3D()};
const auto& shader_config{gpu.regs.shader_config[static_cast<std::size_t>(program)]};
return gpu.regs.code_address.CodeAddress() + shader_config.offset;
@@ -486,7 +486,7 @@ Shader ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program) {
}
auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()};
- const Tegra::GPUVAddr program_addr{GetShaderAddress(program)};
+ const GPUVAddr program_addr{GetShaderAddress(program)};
// Look up shader in the cache based on address
const auto& host_ptr{memory_manager.GetPointer(program_addr)};
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
index 95eab3fec..eac51ecb3 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
@@ -39,8 +39,7 @@ VKBufferCache::VKBufferCache(Tegra::MemoryManager& tegra_memory_manager,
VKBufferCache::~VKBufferCache() = default;
-u64 VKBufferCache::UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size, u64 alignment,
- bool cache) {
+u64 VKBufferCache::UploadMemory(GPUVAddr gpu_addr, std::size_t size, u64 alignment, bool cache) {
const auto cpu_addr{tegra_memory_manager.GpuToCpuAddress(gpu_addr)};
ASSERT_MSG(cpu_addr, "Invalid GPU address");
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h
index 8b415744b..08b786aad 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.h
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h
@@ -68,8 +68,7 @@ public:
/// Uploads data from a guest GPU address. Returns host's buffer offset where it's been
/// allocated.
- u64 UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size, u64 alignment = 4,
- bool cache = true);
+ u64 UploadMemory(GPUVAddr gpu_addr, std::size_t size, u64 alignment = 4, bool cache = true);
/// Uploads from a host memory. Returns host's buffer offset where it's been allocated.
u64 UploadHostMemory(const u8* raw_pointer, std::size_t size, u64 alignment = 4);
diff --git a/src/video_core/renderer_vulkan/vk_resource_manager.cpp b/src/video_core/renderer_vulkan/vk_resource_manager.cpp
index a1e117443..13c46e5b8 100644
--- a/src/video_core/renderer_vulkan/vk_resource_manager.cpp
+++ b/src/video_core/renderer_vulkan/vk_resource_manager.cpp
@@ -21,7 +21,7 @@ public:
CommandBufferPool(const VKDevice& device)
: VKFencedPool(COMMAND_BUFFER_POOL_SIZE), device{device} {}
- void Allocate(std::size_t begin, std::size_t end) {
+ void Allocate(std::size_t begin, std::size_t end) override {
const auto dev = device.GetLogical();
const auto& dld = device.GetDispatchLoader();
const u32 graphics_family = device.GetGraphicsFamily();
diff --git a/src/video_core/renderer_vulkan/vk_resource_manager.h b/src/video_core/renderer_vulkan/vk_resource_manager.h
index 5bfe4cead..08ee86fa6 100644
--- a/src/video_core/renderer_vulkan/vk_resource_manager.h
+++ b/src/video_core/renderer_vulkan/vk_resource_manager.h
@@ -97,7 +97,7 @@ private:
class VKFenceWatch final : public VKResource {
public:
explicit VKFenceWatch();
- ~VKFenceWatch();
+ ~VKFenceWatch() override;
/// Waits for the fence to be released.
void Wait();
diff --git a/src/yuzu/bootmanager.cpp b/src/yuzu/bootmanager.cpp
index d2c97b1f8..05ad19e1d 100644
--- a/src/yuzu/bootmanager.cpp
+++ b/src/yuzu/bootmanager.cpp
@@ -24,8 +24,6 @@ void EmuThread::run() {
MicroProfileOnThreadCreate("EmuThread");
- stop_run = false;
-
emit LoadProgress(VideoCore::LoadCallbackStage::Prepare, 0, 0);
Core::System::GetInstance().Renderer().Rasterizer().LoadDiskResources(
@@ -40,7 +38,7 @@ void EmuThread::run() {
render_window->DoneCurrent();
}
- // holds whether the cpu was running during the last iteration,
+ // Holds whether the cpu was running during the last iteration,
// so that the DebugModeLeft signal can be emitted before the
// next execution step
bool was_active = false;
diff --git a/src/yuzu/debugger/graphics/graphics_surface.cpp b/src/yuzu/debugger/graphics/graphics_surface.cpp
index 29f01dfb2..11023ed63 100644
--- a/src/yuzu/debugger/graphics/graphics_surface.cpp
+++ b/src/yuzu/debugger/graphics/graphics_surface.cpp
@@ -261,7 +261,7 @@ void GraphicsSurfaceWidget::OnSurfaceSourceChanged(int new_value) {
void GraphicsSurfaceWidget::OnSurfaceAddressChanged(qint64 new_value) {
if (surface_address != new_value) {
- surface_address = static_cast<Tegra::GPUVAddr>(new_value);
+ surface_address = static_cast<GPUVAddr>(new_value);
surface_source_list->setCurrentIndex(static_cast<int>(Source::Custom));
emit Update();
diff --git a/src/yuzu/debugger/graphics/graphics_surface.h b/src/yuzu/debugger/graphics/graphics_surface.h
index 323e39d94..89445b18f 100644
--- a/src/yuzu/debugger/graphics/graphics_surface.h
+++ b/src/yuzu/debugger/graphics/graphics_surface.h
@@ -87,7 +87,7 @@ private:
QPushButton* save_surface;
Source surface_source;
- Tegra::GPUVAddr surface_address;
+ GPUVAddr surface_address;
unsigned surface_width;
unsigned surface_height;
Tegra::Texture::TextureFormat surface_format;
diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp
index c6c66a787..245f25847 100644
--- a/src/yuzu_cmd/yuzu.cpp
+++ b/src/yuzu_cmd/yuzu.cpp
@@ -114,9 +114,9 @@ int main(int argc, char** argv) {
};
while (optind < argc) {
- char arg = getopt_long(argc, argv, "g:fhvp::", long_options, &option_index);
+ int arg = getopt_long(argc, argv, "g:fhvp::", long_options, &option_index);
if (arg != -1) {
- switch (arg) {
+ switch (static_cast<char>(arg)) {
case 'g':
errno = 0;
gdb_port = strtoul(optarg, &endarg, 0);