Merge branch 'refs/heads/main' into loadaudio

CMakeLists.txt

@@ -222,7 +222,7 @@ if (ENABLE_CUBEB)
 		option(BUILD_TOOLS "" OFF)
 		option(BUNDLE_SPEEX "" OFF)
 		set(USE_WINMM OFF CACHE BOOL "")
-		add_subdirectory("dependencies/cubeb" EXCLUDE_FROM_ALL)
+		add_subdirectory("dependencies/cubeb" EXCLUDE_FROM_ALL SYSTEM)
 		set_property(TARGET cubeb PROPERTY MSVC_RUNTIME_LIBRARY "MultiThreaded$<$<CONFIG:Debug>:Debug>")
 		add_library(cubeb::cubeb ALIAS cubeb)
 	endif()

src/Cafe/HW/Latte/Renderer/Vulkan/VulkanRenderer.cpp

@@ -1892,6 +1892,7 @@ void VulkanRenderer::ProcessFinishedCommandBuffers()
 		if (fenceStatus == VK_SUCCESS)
 		{
 			ProcessDestructionQueue();
+			m_uniformVarBufferReadIndex = m_cmdBufferUniformRingbufIndices[m_commandBufferSyncIndex];
 			m_commandBufferSyncIndex = (m_commandBufferSyncIndex + 1) % m_commandBuffers.size();
 			memoryManager->cleanupBuffers(m_countCommandBufferFinished);
 			m_countCommandBufferFinished++;
@@ -1985,6 +1986,7 @@ void VulkanRenderer::SubmitCommandBuffer(VkSemaphore signalSemaphore, VkSemaphor
 		cemuLog_logDebug(LogType::Force, "Vulkan: Waiting for available command buffer...");
 		WaitForNextFinishedCommandBuffer();
 	}
+	m_cmdBufferUniformRingbufIndices[nextCmdBufferIndex] = m_cmdBufferUniformRingbufIndices[m_commandBufferIndex];
 	m_commandBufferIndex = nextCmdBufferIndex;
 
 
@@ -3562,13 +3564,13 @@ void VulkanRenderer::buffer_bindUniformBuffer(LatteConst::ShaderType shaderType,
 	switch (shaderType)
 	{
 	case LatteConst::ShaderType::Vertex:
-		dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_VERTEX].unformBufferOffset[bufferIndex] = offset;
+		dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_VERTEX].uniformBufferOffset[bufferIndex] = offset;
 		break;
 	case LatteConst::ShaderType::Geometry:
-		dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_GEOMETRY].unformBufferOffset[bufferIndex] = offset;
+		dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_GEOMETRY].uniformBufferOffset[bufferIndex] = offset;
 		break;
 	case LatteConst::ShaderType::Pixel:
-		dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_FRAGMENT].unformBufferOffset[bufferIndex] = offset;
+		dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_FRAGMENT].uniformBufferOffset[bufferIndex] = offset;
 		break;
 	default:
 		cemu_assert_debug(false);

src/Cafe/HW/Latte/Renderer/Vulkan/VulkanRenderer.h

@@ -591,6 +591,7 @@ private:
 	bool m_uniformVarBufferMemoryIsCoherent{false};
 	uint8* m_uniformVarBufferPtr = nullptr;
 	uint32 m_uniformVarBufferWriteIndex = 0;
+	uint32 m_uniformVarBufferReadIndex = 0;
 
 	// transform feedback ringbuffer
 	VkBuffer m_xfbRingBuffer = VK_NULL_HANDLE;
@@ -637,6 +638,7 @@ private:
 	size_t m_commandBufferIndex = 0; // current buffer being filled
 	size_t m_commandBufferSyncIndex = 0; // latest buffer that finished execution (updated on submit)
 	size_t m_commandBufferIDOfPrevFrame = 0;
+	std::array<size_t, kCommandBufferPoolSize> m_cmdBufferUniformRingbufIndices {}; // index in the uniform ringbuffer
 	std::array<VkFence, kCommandBufferPoolSize> m_cmd_buffer_fences;
 	std::array<VkCommandBuffer, kCommandBufferPoolSize> m_commandBuffers;
 	std::array<VkSemaphore, kCommandBufferPoolSize> m_commandBufferSemaphores;
@@ -659,7 +661,7 @@ private:
 		uint32 uniformVarBufferOffset[VulkanRendererConst::SHADER_STAGE_INDEX_COUNT];
 		struct  
 		{
-			uint32 unformBufferOffset[LATTE_NUM_MAX_UNIFORM_BUFFERS];
+			uint32 uniformBufferOffset[LATTE_NUM_MAX_UNIFORM_BUFFERS];
 		}shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_COUNT];
 	}dynamicOffsetInfo{};
 

src/Cafe/HW/Latte/Renderer/Vulkan/VulkanRendererCore.cpp

@@ -375,24 +375,20 @@ float s_vkUniformData[512 * 4];
 
 void VulkanRenderer::uniformData_updateUniformVars(uint32 shaderStageIndex, LatteDecompilerShader* shader)
 {
-	auto GET_UNIFORM_DATA_PTR = [&](size_t index) { return s_vkUniformData + (index / 4); };
+	auto GET_UNIFORM_DATA_PTR = [](size_t index) { return s_vkUniformData + (index / 4); };
 
 	sint32 shaderAluConst;
-	sint32 shaderUniformRegisterOffset;
 
 	switch (shader->shaderType)
 	{
 	case LatteConst::ShaderType::Vertex:
 		shaderAluConst = 0x400;
-		shaderUniformRegisterOffset = mmSQ_VTX_UNIFORM_BLOCK_START;
 		break;
 	case LatteConst::ShaderType::Pixel:
 		shaderAluConst = 0;
-		shaderUniformRegisterOffset = mmSQ_PS_UNIFORM_BLOCK_START;
 		break;
 	case LatteConst::ShaderType::Geometry:
 		shaderAluConst = 0; // geometry shader has no ALU const
-		shaderUniformRegisterOffset = mmSQ_GS_UNIFORM_BLOCK_START;
 		break;
 	default:
 		UNREACHABLE;
@@ -445,7 +441,7 @@ void VulkanRenderer::uniformData_updateUniformVars(uint32 shaderStageIndex, Latt
 		}
 		if (shader->uniform.loc_verticesPerInstance >= 0)
 		{
-			*(int*)(s_vkUniformData + ((size_t)shader->uniform.loc_verticesPerInstance / 4)) = m_streamoutState.verticesPerInstance;
+			*(int*)GET_UNIFORM_DATA_PTR(shader->uniform.loc_verticesPerInstance) = m_streamoutState.verticesPerInstance;
 			for (sint32 b = 0; b < LATTE_NUM_STREAMOUT_BUFFER; b++)
 			{
 				if (shader->uniform.loc_streamoutBufferBase[b] >= 0)
@@ -455,26 +451,63 @@ void VulkanRenderer::uniformData_updateUniformVars(uint32 shaderStageIndex, Latt
 			}
 		}
 		// upload
-		if ((m_uniformVarBufferWriteIndex + shader->uniform.uniformRangeSize + 1024) > UNIFORMVAR_RINGBUFFER_SIZE)
+		const uint32 bufferAlignmentM1 = std::max(m_featureControl.limits.minUniformBufferOffsetAlignment, m_featureControl.limits.nonCoherentAtomSize) - 1;
+		const uint32 uniformSize = (shader->uniform.uniformRangeSize + bufferAlignmentM1) & ~bufferAlignmentM1;
+
+		auto waitWhileCondition = [&](std::function<bool()> condition) {
+			while (condition())
+			{
+				if (m_commandBufferSyncIndex == m_commandBufferIndex)
+				{
+					if (m_cmdBufferUniformRingbufIndices[m_commandBufferIndex] != m_uniformVarBufferReadIndex)
+					{
+						draw_endRenderPass();
+						SubmitCommandBuffer();
+					}
+					else
+					{
+						// submitting work would not change readIndex, so there's no way for conditions based on it to change
+						cemuLog_log(LogType::Force, "draw call overflowed and corrupted uniform ringbuffer. expect visual corruption");
+						cemu_assert_suspicious();
+						break;
+					}
+				}
+				WaitForNextFinishedCommandBuffer();
+			}
+		};
+
+		// wrap around if it doesnt fit consecutively
+		if (m_uniformVarBufferWriteIndex + uniformSize > UNIFORMVAR_RINGBUFFER_SIZE)
 		{
+			waitWhileCondition([&]() {
+				return m_uniformVarBufferReadIndex > m_uniformVarBufferWriteIndex || m_uniformVarBufferReadIndex == 0;
+			});
 			m_uniformVarBufferWriteIndex = 0;
 		}
-		uint32 bufferAlignmentM1 = std::max(m_featureControl.limits.minUniformBufferOffsetAlignment, m_featureControl.limits.nonCoherentAtomSize) - 1;
+
+		auto ringBufRemaining = [&]() {
+			ssize_t ringBufferUsedBytes = (ssize_t)m_uniformVarBufferWriteIndex - m_uniformVarBufferReadIndex;
+			if (ringBufferUsedBytes < 0)
+				ringBufferUsedBytes += UNIFORMVAR_RINGBUFFER_SIZE;
+			return UNIFORMVAR_RINGBUFFER_SIZE - 1 - ringBufferUsedBytes;
+		};
+		waitWhileCondition([&]() {
+			return ringBufRemaining() < uniformSize;
+		});
+
 		const uint32 uniformOffset = m_uniformVarBufferWriteIndex;
 		memcpy(m_uniformVarBufferPtr + uniformOffset, s_vkUniformData, shader->uniform.uniformRangeSize);
-		m_uniformVarBufferWriteIndex += shader->uniform.uniformRangeSize;
+		m_uniformVarBufferWriteIndex += uniformSize;
-		m_uniformVarBufferWriteIndex = (m_uniformVarBufferWriteIndex + bufferAlignmentM1) & ~bufferAlignmentM1;
 		// update dynamic offset
 		dynamicOffsetInfo.uniformVarBufferOffset[shaderStageIndex] = uniformOffset;
 		// flush if not coherent
 		if (!m_uniformVarBufferMemoryIsCoherent)
 		{
-			uint32 nonCoherentAtomSizeM1 = m_featureControl.limits.nonCoherentAtomSize - 1;
 			VkMappedMemoryRange flushedRange{};
 			flushedRange.sType = VK_STRUCTURE_TYPE_MAPPED_MEMORY_RANGE;
 			flushedRange.memory = m_uniformVarBufferMemory;
 			flushedRange.offset = uniformOffset;
-			flushedRange.size = (shader->uniform.uniformRangeSize + nonCoherentAtomSizeM1) & ~nonCoherentAtomSizeM1;
+			flushedRange.size = uniformSize;
 			vkFlushMappedMemoryRanges(m_logicalDevice, 1, &flushedRange);
 		}
 	}
@@ -494,7 +527,7 @@ void VulkanRenderer::draw_prepareDynamicOffsetsForDescriptorSet(uint32 shaderSta
 	{
 		for (auto& itr : pipeline_info->dynamicOffsetInfo.list_uniformBuffers[shaderStageIndex])
 		{
-			dynamicOffsets[numDynOffsets] = dynamicOffsetInfo.shaderUB[shaderStageIndex].unformBufferOffset[itr];
+			dynamicOffsets[numDynOffsets] = dynamicOffsetInfo.shaderUB[shaderStageIndex].uniformBufferOffset[itr];
 			numDynOffsets++;
 		}
 	}
@@ -1357,6 +1390,24 @@ void VulkanRenderer::draw_execute(uint32 baseVertex, uint32 baseInstance, uint32
 		return;
 	}
 
+	// prepare streamout
+	m_streamoutState.verticesPerInstance = count;
+	LatteStreamout_PrepareDrawcall(count, instanceCount);
+
+	// update uniform vars
+	LatteDecompilerShader* vertexShader = LatteSHRC_GetActiveVertexShader();
+	LatteDecompilerShader* pixelShader = LatteSHRC_GetActivePixelShader();
+	LatteDecompilerShader* geometryShader = LatteSHRC_GetActiveGeometryShader();
+
+	if (vertexShader)
+		uniformData_updateUniformVars(VulkanRendererConst::SHADER_STAGE_INDEX_VERTEX, vertexShader);
+	if (pixelShader)
+		uniformData_updateUniformVars(VulkanRendererConst::SHADER_STAGE_INDEX_FRAGMENT, pixelShader);
+	if (geometryShader)
+		uniformData_updateUniformVars(VulkanRendererConst::SHADER_STAGE_INDEX_GEOMETRY, geometryShader);
+	// store where the read pointer should go after command buffer execution
+	m_cmdBufferUniformRingbufIndices[m_commandBufferIndex] = m_uniformVarBufferWriteIndex;
+
 	// process index data
 	const LattePrimitiveMode primitiveMode = static_cast<LattePrimitiveMode>(LatteGPUState.contextRegister[mmVGT_PRIMITIVE_TYPE]);
 
@@ -1410,22 +1461,6 @@ void VulkanRenderer::draw_execute(uint32 baseVertex, uint32 baseInstance, uint32
 		LatteBufferCache_Sync(indexMin + baseVertex, indexMax + baseVertex, baseInstance, instanceCount);
 	}
 
-	// prepare streamout
-	m_streamoutState.verticesPerInstance = count;
-	LatteStreamout_PrepareDrawcall(count, instanceCount);
-
-	// update uniform vars
-	LatteDecompilerShader* vertexShader = LatteSHRC_GetActiveVertexShader();
-	LatteDecompilerShader* pixelShader = LatteSHRC_GetActivePixelShader();
-	LatteDecompilerShader* geometryShader = LatteSHRC_GetActiveGeometryShader();
-
-	if (vertexShader)
-		uniformData_updateUniformVars(VulkanRendererConst::SHADER_STAGE_INDEX_VERTEX, vertexShader);
-	if (pixelShader)
-		uniformData_updateUniformVars(VulkanRendererConst::SHADER_STAGE_INDEX_FRAGMENT, pixelShader);
-	if (geometryShader)
-		uniformData_updateUniformVars(VulkanRendererConst::SHADER_STAGE_INDEX_GEOMETRY, geometryShader);
-
 	PipelineInfo* pipeline_info;
 
 	if (!isFirst)
@@ -1613,13 +1648,13 @@ void VulkanRenderer::draw_updateUniformBuffersDirectAccess(LatteDecompilerShader
 			switch (shaderType)
 			{
 			case LatteConst::ShaderType::Vertex:
-				dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_VERTEX].unformBufferOffset[bufferIndex] = physicalAddr - m_importedMemBaseAddress;
+				dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_VERTEX].uniformBufferOffset[bufferIndex] = physicalAddr - m_importedMemBaseAddress;
 				break;
 			case LatteConst::ShaderType::Geometry:
-				dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_GEOMETRY].unformBufferOffset[bufferIndex] = physicalAddr - m_importedMemBaseAddress;
+				dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_GEOMETRY].uniformBufferOffset[bufferIndex] = physicalAddr - m_importedMemBaseAddress;
 				break;
 			case LatteConst::ShaderType::Pixel:
-				dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_FRAGMENT].unformBufferOffset[bufferIndex] = physicalAddr - m_importedMemBaseAddress;
+				dynamicOffsetInfo.shaderUB[VulkanRendererConst::SHADER_STAGE_INDEX_FRAGMENT].uniformBufferOffset[bufferIndex] = physicalAddr - m_importedMemBaseAddress;
 				break;
 			default:
 				UNREACHABLE;

src/Cafe/IOSU/legacy/iosu_boss.cpp

@@ -137,6 +137,10 @@ namespace iosu
 			this->task_settings.taskType = settings->taskType;
 
 			curl = std::shared_ptr<CURL>(curl_easy_init(), curl_easy_cleanup);
+			if(GetConfig().proxy_server.GetValue() != "")
+			{
+			  curl_easy_setopt(curl.get(), CURLOPT_PROXY, GetConfig().proxy_server.GetValue().c_str());
+			}
 		}
 	};
 

src/Cafe/OS/libs/nn_olv/nn_olv_OfflineDB.cpp

@@ -112,7 +112,7 @@ namespace nn
 
 		nnResult _Async_OfflineDB_DownloadPostDataListParam_DownloadPostDataList(coreinit::OSEvent* event, DownloadedTopicData* downloadedTopicData, DownloadedPostData* downloadedPostData, uint32be* postCountOut, uint32 maxCount, DownloadPostDataListParam* param)
 		{
-			scope_exit _se([&](){coreinit::OSSignalEvent(event);});
+			stdx::scope_exit _se([&](){coreinit::OSSignalEvent(event);});
 
 			uint64 titleId = CafeSystem::GetForegroundTitleId();
 
@@ -184,7 +184,7 @@ namespace nn
 
 		nnResult _Async_OfflineDB_DownloadPostDataListParam_DownloadExternalImageData(coreinit::OSEvent* event, DownloadedDataBase* _this, void* imageDataOut, uint32be* imageSizeOut, uint32 maxSize)
 		{
-			scope_exit _se([&](){coreinit::OSSignalEvent(event);});
+			stdx::scope_exit _se([&](){coreinit::OSSignalEvent(event);});
 
 			if (!_this->TestFlags(_this, DownloadedDataBase::FLAGS::HAS_EXTERNAL_IMAGE))
 				return OLV_RESULT_MISSING_DATA;

src/Cafe/OS/libs/nsyshid/Infinity.cpp

@@ -1017,11 +1017,7 @@ namespace nsyshid
 	std::array<uint8, 16> InfinityUSB::GenerateInfinityFigureKey(const std::vector<uint8>& sha1Data)
 	{
 		std::array<uint8, 20> digest = {};
-		SHA_CTX ctx;
+		SHA1(sha1Data.data(), sha1Data.size(), digest.data());
-		SHA1_Init(&ctx);
-		SHA1_Update(&ctx, sha1Data.data(), sha1Data.size());
-		SHA1_Final(digest.data(), &ctx);
-		OPENSSL_cleanse(&ctx, sizeof(ctx));
 		// Infinity AES keys are the first 16 bytes of the SHA1 Digest, every set of 4 bytes need to be
 		// reversed due to endianness
 		std::array<uint8, 16> key = {};

src/Cafe/OS/libs/ntag/ntag.cpp

@@ -509,7 +509,7 @@ namespace ntag
 			noftHeader->writeCount = _swapEndianU16(_swapEndianU16(noftHeader->writeCount) + 1);
 		}
 
-		memcpy(decryptedBuffer + 0x20, noftHeader, sizeof(noftHeader));
+		memcpy(decryptedBuffer + 0x20, noftHeader, sizeof(NTAGNoftHeader));
 		memcpy(decryptedBuffer + _swapEndianU16(rwHeader->offset), data, dataSize);
 
 		// Encrypt

src/Cafe/OS/libs/snd_core/ax_out.cpp

@@ -467,10 +467,10 @@ namespace snd_core
 	// called periodically to check for AX updates
 	void AXOut_update()
 	{
-		constexpr auto kTimeout = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::milliseconds(((IAudioAPI::kBlockCount * 3) / 4) * (AX_FRAMES_PER_GROUP * 3)));
+		constexpr static auto kTimeout = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::milliseconds(((IAudioAPI::kBlockCount * 3) / 4) * (AX_FRAMES_PER_GROUP * 3)));
-		constexpr auto kWaitDuration = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::milliseconds(3));
+		constexpr static auto kWaitDuration = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::milliseconds(3));
-		constexpr auto kWaitDurationFast = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::microseconds(2900));
+		constexpr static auto kWaitDurationFast = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::microseconds(2900));
-		constexpr auto kWaitDurationMinimum = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::microseconds(1700));
+		constexpr static auto kWaitDurationMinimum = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::microseconds(1700));
 
 		// if we haven't buffered any blocks, we will wait less time than usual
 		bool additional_blocks_required = false;

src/Cemu/Logging/CemuLogging.h

@@ -91,7 +91,11 @@ bool cemuLog_log(LogType type, std::basic_string<T> formatStr, TArgs&&... args)
 	else
 	{
 		const auto format_view = fmt::basic_string_view<T>(formatStr);
+#if FMT_VERSION >= 110000
+		const auto text = fmt::vformat(format_view, fmt::make_format_args<fmt::buffered_context<T>>(args...));
+#else
 		const auto text = fmt::vformat(format_view, fmt::make_format_args<fmt::buffer_context<T>>(args...));
+#endif
 		cemuLog_log(type, std::basic_string_view(text.data(), text.size()));
 	}
 	return true;

src/Common/MemPtr.h

@@ -4,7 +4,7 @@
 
 using MPTR = uint32; // generic address in PowerPC memory space
 
-#define MPTR_NULL	(0)
+#define MPTR_NULL (0)
 
 using VAddr = uint32; // virtual address
 using PAddr = uint32; // physical address
@@ -14,137 +14,175 @@ extern uint8* PPCInterpreterGetStackPointer();
 extern uint8* PPCInterpreter_PushAndReturnStackPointer(sint32 offset);
 extern void PPCInterpreterModifyStackPointer(sint32 offset);
 
-class MEMPTRBase {};
+class MEMPTRBase
+{
+};
 
-template <typename T>
+template<typename T>
 class MEMPTR : MEMPTRBase
 {
-public:
+  public:
-	constexpr MEMPTR()
+	constexpr MEMPTR() noexcept
-		: m_value(0) { }
+		: m_value(0) {}
 
-	explicit constexpr MEMPTR(uint32 offset)
+	explicit constexpr MEMPTR(uint32 offset) noexcept
-		: m_value(offset) { }
+		: m_value(offset) {}
 
-	explicit constexpr MEMPTR(const uint32be& offset)
+	explicit constexpr MEMPTR(const uint32be& offset) noexcept
-		: m_value(offset) { }
+		: m_value(offset) {}
 
-	constexpr MEMPTR(std::nullptr_t)
+	constexpr MEMPTR(std::nullptr_t) noexcept
-		: m_value(0) { }
+		: m_value(0) {}
 
-	MEMPTR(T* ptr)
+	MEMPTR(T* ptr) noexcept
 	{
 		if (ptr == nullptr)
 			m_value = 0;
 		else
-        {
+		{
-            cemu_assert_debug((uint8*)ptr >= memory_base && (uint8*)ptr <= memory_base + 0x100000000);
+			cemu_assert_debug((uint8*)ptr >= memory_base && (uint8*)ptr <= memory_base + 0x100000000);
-            m_value = (uint32)((uintptr_t)ptr - (uintptr_t)memory_base);
+			m_value = (uint32)((uintptr_t)ptr - (uintptr_t)memory_base);
-        }
+		}
 	}
 
-	constexpr MEMPTR(const MEMPTR& memptr)
+	constexpr MEMPTR(const MEMPTR&) noexcept = default;
-		: m_value(memptr.m_value) { }
 
-	constexpr MEMPTR& operator=(const MEMPTR& memptr)
+	constexpr MEMPTR& operator=(const MEMPTR&) noexcept = default;
-	{
-		m_value = memptr.m_value;
-		return *this;
-	}
 
-	constexpr MEMPTR& operator=(const uint32& offset)
+	constexpr MEMPTR& operator=(const uint32& offset) noexcept
 	{
 		m_value = offset;
 		return *this;
 	}
 
-	constexpr MEMPTR& operator=(const std::nullptr_t rhs)
+	constexpr MEMPTR& operator=(std::nullptr_t) noexcept
 	{
 		m_value = 0;
 		return *this;
 	}
 
-	MEMPTR& operator=(T* ptr)
+	MEMPTR& operator=(T* ptr) noexcept
 	{
-        if (ptr == nullptr)
+		if (ptr == nullptr)
 			m_value = 0;
 		else
-        {
+		{
-            cemu_assert_debug((uint8*)ptr >= memory_base && (uint8*)ptr <= memory_base + 0x100000000);
+			cemu_assert_debug((uint8*)ptr >= memory_base && (uint8*)ptr <= memory_base + 0x100000000);
-            m_value = (uint32)((uintptr_t)ptr - (uintptr_t)memory_base);
+			m_value = (uint32)((uintptr_t)ptr - (uintptr_t)memory_base);
-        }
+		}
 		return *this;
 	}
 
-	bool atomic_compare_exchange(T* comparePtr, T* newPtr)
+	bool atomic_compare_exchange(T* comparePtr, T* newPtr) noexcept
 	{
 		MEMPTR<T> mp_compare = comparePtr;
 		MEMPTR<T> mp_new = newPtr;
-		std::atomic<uint32be>* thisValueAtomic = (std::atomic<uint32be>*)&m_value;
+		auto* thisValueAtomic = reinterpret_cast<std::atomic<uint32be>*>(&m_value);
 		return thisValueAtomic->compare_exchange_strong(mp_compare.m_value, mp_new.m_value);
 	}
 
-	explicit constexpr operator bool() const noexcept { return m_value != 0; }
+	explicit constexpr operator bool() const noexcept
-	
+	{
-	constexpr operator T*() const noexcept { return GetPtr(); } // allow implicit cast to wrapped pointer type
+		return m_value != 0;
+	}
 
+	// allow implicit cast to wrapped pointer type
+	constexpr operator T*() const noexcept
+	{
+		return GetPtr();
+	}
 
-	template <typename X>
+	template<typename X>
-	explicit operator MEMPTR<X>() const { return MEMPTR<X>(this->m_value); }
+	explicit operator MEMPTR<X>() const noexcept
+	{
+		return MEMPTR<X>(this->m_value);
+	}
 
-	MEMPTR operator+(const MEMPTR& ptr) { return MEMPTR(this->GetMPTR() + ptr.GetMPTR()); }
+	MEMPTR operator+(const MEMPTR& ptr) noexcept
-	MEMPTR operator-(const MEMPTR& ptr) { return MEMPTR(this->GetMPTR() - ptr.GetMPTR()); }
+	{
+		return MEMPTR(this->GetMPTR() + ptr.GetMPTR());
+	}
+	MEMPTR operator-(const MEMPTR& ptr) noexcept
+	{
+		return MEMPTR(this->GetMPTR() - ptr.GetMPTR());
+	}
 
-	MEMPTR operator+(sint32 v)
+	MEMPTR operator+(sint32 v) noexcept
 	{
 		// pointer arithmetic
 		return MEMPTR(this->GetMPTR() + v * 4);
 	}
 
-	MEMPTR operator-(sint32 v)
+	MEMPTR operator-(sint32 v) noexcept
 	{
 		// pointer arithmetic
 		return MEMPTR(this->GetMPTR() - v * 4);
 	}
 
-	MEMPTR& operator+=(sint32 v)
+	MEMPTR& operator+=(sint32 v) noexcept
 	{
 		m_value += v * sizeof(T);
 		return *this;
 	}
 
-	template <class Q = T>
+	template<typename Q = T>
-	typename std::enable_if<!std::is_same<Q, void>::value, Q>::type&
+	std::enable_if_t<!std::is_same_v<Q, void>, Q>& operator*() const noexcept
-	operator*() const { return *GetPtr(); }
+	{
+		return *GetPtr();
+	}
 
-	T* operator->() const { return GetPtr(); }
+	constexpr T* operator->() const noexcept
+	{
+		return GetPtr();
+	}
 
-	template <class Q = T>
+	template<typename Q = T>
-	typename std::enable_if<!std::is_same<Q, void>::value, Q>::type&
+	std::enable_if_t<!std::is_same_v<Q, void>, Q>& operator[](int index) noexcept
-	operator[](int index) { return GetPtr()[index]; }
+	{
+		return GetPtr()[index];
+	}
 
-	T* GetPtr() const { return (T*)(m_value == 0 ? nullptr : memory_base + (uint32)m_value); }
+	T* GetPtr() const noexcept
+	{
+		return (T*)(m_value == 0 ? nullptr : memory_base + (uint32)m_value);
+	}
 
-	template <typename C>
+	template<typename C>
-	C* GetPtr() const { return (C*)(GetPtr()); }
+	C* GetPtr() const noexcept
+	{
+		return static_cast<C*>(GetPtr());
+	}
 
-	constexpr uint32 GetMPTR() const { return m_value.value(); }
+	[[nodiscard]] constexpr uint32 GetMPTR() const noexcept
-	constexpr const uint32be& GetBEValue() const { return m_value; }
+	{
+		return m_value.value();
+	}
+	[[nodiscard]] constexpr const uint32be& GetBEValue() const noexcept
+	{
+		return m_value;
+	}
 
-	constexpr bool IsNull() const { return m_value == 0; }
+	[[nodiscard]] constexpr bool IsNull() const noexcept
+	{
+		return m_value == 0;
+	}
 
-private:
+  private:
 	uint32be m_value;
 };
 
 static_assert(sizeof(MEMPTR<void*>) == sizeof(uint32be));
+static_assert(std::is_trivially_copyable_v<MEMPTR<void*>>);
 
 #include "StackAllocator.h"
 #include "SysAllocator.h"
 
-template <typename T>
+template<typename T>
 struct fmt::formatter<MEMPTR<T>> : formatter<string_view>
 {
-	template <typename FormatContext>
+	template<typename FormatContext>
-	auto format(const MEMPTR<T>& v, FormatContext& ctx) const -> format_context::iterator { return fmt::format_to(ctx.out(), "{:#x}", v.GetMPTR()); }
+	auto format(const MEMPTR<T>& v, FormatContext& ctx) const -> format_context::iterator
+	{
+		return fmt::format_to(ctx.out(), "{:#x}", v.GetMPTR());
+	}
 };

src/Common/precompiled.h

@@ -394,16 +394,10 @@ void vectorRemoveByIndex(std::vector<T>& vec, const size_t index)
     vec.erase(vec.begin() + index);
 }
 
-template<typename T1, typename T2>
+template<typename T1, typename... Types>
-int match_any_of(T1 value, T2 compareTo)
+bool match_any_of(T1&& value, Types&&... others)
 {
-    return value == compareTo;
+    return ((value == others) || ...);
-}
-
-template<typename T1, typename T2, typename... Types>
-bool match_any_of(T1 value, T2 compareTo, Types&&... others)
-{
-    return value == compareTo || match_any_of(value, others...);
 }
 
 // we cache the frequency in a static variable
@@ -501,13 +495,6 @@ bool future_is_ready(std::future<T>& f)
 #endif
 }
 
-// replace with std::scope_exit once available
-struct scope_exit
-{
-	std::function<void()> f_;
-	explicit scope_exit(std::function<void()> f) noexcept : f_(std::move(f)) {}
-	~scope_exit() { if (f_) f_(); }
-};
 
 // helper function to cast raw pointers to std::atomic
 // this is technically not legal but works on most platforms as long as alignment restrictions are met and the implementation of atomic doesnt come with additional members
@@ -515,6 +502,8 @@ struct scope_exit
 template<typename T>
 std::atomic<T>* _rawPtrToAtomic(T* ptr)
 {
+	static_assert(sizeof(T) == sizeof(std::atomic<T>));
+	cemu_assert_debug((reinterpret_cast<std::uintptr_t>(ptr) % alignof(std::atomic<T>)) == 0);
     return reinterpret_cast<std::atomic<T>*>(ptr);
 }
 
@@ -578,13 +567,34 @@ struct fmt::formatter<betype<T>> : fmt::formatter<T>
 	}
 };
 
-// useful C++23 stuff that isn't yet widely supported
+// useful future C++ stuff
-
-// std::to_underlying
 namespace stdx
 {
+	// std::to_underlying
     template <typename EnumT, typename = std::enable_if_t < std::is_enum<EnumT>{} >>
         constexpr std::underlying_type_t<EnumT> to_underlying(EnumT e) noexcept {
         return static_cast<std::underlying_type_t<EnumT>>(e);
     };
+
+	// std::scope_exit
+	template <typename Fn>
+	class scope_exit
+	{
+		Fn m_func;
+		bool m_released = false;
+	public:
+		explicit scope_exit(Fn&& f) noexcept
+			: m_func(std::forward<Fn>(f))
+		{}
+		~scope_exit()
+		{
+			if (!m_released) m_func();
+		}
+		scope_exit(scope_exit&& other) noexcept
+			: m_func(std::move(other.m_func)), m_released(std::exchange(other.m_released, true))
+		{}
+		scope_exit(const scope_exit&) = delete;
+		scope_exit& operator=(scope_exit) = delete;
+		void release() { m_released = true;}
+	};
 }

src/config/CemuConfig.h

@@ -194,7 +194,7 @@ ENABLE_ENUM_ITERATORS(CrashDump, CrashDump::Disabled, CrashDump::Enabled);
 template <>
 struct fmt::formatter<PrecompiledShaderOption> : formatter<string_view> {
 	template <typename FormatContext>
-	auto format(const PrecompiledShaderOption c, FormatContext &ctx) {
+	auto format(const PrecompiledShaderOption c, FormatContext &ctx) const {
 		string_view name;
 		switch (c)
 		{
@@ -209,7 +209,7 @@ struct fmt::formatter<PrecompiledShaderOption> : formatter<string_view> {
 template <>
 struct fmt::formatter<AccurateShaderMulOption> : formatter<string_view> {
 	template <typename FormatContext>
-	auto format(const AccurateShaderMulOption c, FormatContext &ctx) {
+	auto format(const AccurateShaderMulOption c, FormatContext &ctx) const {
 		string_view name;
 		switch (c)
 		{
@@ -223,7 +223,7 @@ struct fmt::formatter<AccurateShaderMulOption> : formatter<string_view> {
 template <>
 struct fmt::formatter<CPUMode> : formatter<string_view> {
 	template <typename FormatContext>
-	auto format(const CPUMode c, FormatContext &ctx) {
+	auto format(const CPUMode c, FormatContext &ctx) const {
 		string_view name;
 		switch (c)
 		{
@@ -240,7 +240,7 @@ struct fmt::formatter<CPUMode> : formatter<string_view> {
 template <>
 struct fmt::formatter<CPUModeLegacy> : formatter<string_view> {
 	template <typename FormatContext>
-	auto format(const CPUModeLegacy c, FormatContext &ctx) {
+	auto format(const CPUModeLegacy c, FormatContext &ctx) const {
 		string_view name;
 		switch (c)
 		{
@@ -257,7 +257,7 @@ struct fmt::formatter<CPUModeLegacy> : formatter<string_view> {
 template <>
 struct fmt::formatter<CafeConsoleRegion> : formatter<string_view> {
 	template <typename FormatContext>
-	auto format(const CafeConsoleRegion v, FormatContext &ctx) {
+	auto format(const CafeConsoleRegion v, FormatContext &ctx) const {
 		string_view name;
 		switch (v)
 		{

src/gui/CemuApp.cpp

@@ -15,6 +15,9 @@
 #if BOOST_OS_LINUX && HAS_WAYLAND
 #include "gui/helpers/wxWayland.h"
 #endif
+#if __WXGTK__
+#include <glib.h>
+#endif
 
 #include <wx/image.h>
 #include <wx/filename.h>

src/gui/helpers/wxHelpers.h

@@ -8,7 +8,7 @@ template <>
 struct fmt::formatter<wxString> : formatter<string_view>
 {
 	template <typename FormatContext>
-	auto format(const wxString& str, FormatContext& ctx)
+	auto format(const wxString& str, FormatContext& ctx) const
 	{
 		return formatter<string_view>::format(str.c_str().AsChar(), ctx);
 	}

src/input/emulated/EmulatedController.h

@@ -127,7 +127,7 @@ using EmulatedControllerPtr = std::shared_ptr<EmulatedController>;
 template <>
 struct fmt::formatter<EmulatedController::Type> : formatter<string_view> {
 	template <typename FormatContext>
-	auto format(EmulatedController::Type v, FormatContext& ctx) {
+	auto format(EmulatedController::Type v, FormatContext& ctx) const {
 		switch (v)
 		{
 		case EmulatedController::Type::VPAD: return formatter<string_view>::format("Wii U Gamepad", ctx);

src/util/CMakeLists.txt

@@ -52,7 +52,6 @@ add_library(CemuUtil
   MemMapper/MemMapper.h
   SystemInfo/SystemInfo.cpp
   SystemInfo/SystemInfo.h
-  ThreadPool/ThreadPool.cpp
   ThreadPool/ThreadPool.h
   tinyxml2/tinyxml2.cpp
   tinyxml2/tinyxml2.h

src/util/containers/robin_hood.h

@@ -194,7 +194,7 @@ namespace robin_hood {
 
 // workaround missing "is_trivially_copyable" in g++ < 5.0
 // See https://stackoverflow.com/a/31798726/48181
-#if defined(__GNUC__) && __GNUC__ < 5
+#if defined(__GNUC__) && __GNUC__ < 5 && !defined(__clang__)
 #    define ROBIN_HOOD_IS_TRIVIALLY_COPYABLE(...) __has_trivial_copy(__VA_ARGS__)
 #else
 #    define ROBIN_HOOD_IS_TRIVIALLY_COPYABLE(...) std::is_trivially_copyable<__VA_ARGS__>::value

src/util/crypto/crc32.cpp

@@ -1,29 +1,6 @@
 #include "crc32.h"
 
-#if defined(_MSC_VER) || defined(__MINGW32__)
+constexpr uint32 Crc32Lookup[8][256] =
-#define __LITTLE_ENDIAN 1234
-#define __BIG_ENDIAN    4321
-#define __BYTE_ORDER    __LITTLE_ENDIAN
-
-#include <xmmintrin.h>
-#ifdef __MINGW32__
-#define PREFETCH(location) __builtin_prefetch(location)
-#else
-#define PREFETCH(location) _mm_prefetch(location, _MM_HINT_T0)
-#endif
-#else
-// defines __BYTE_ORDER as __LITTLE_ENDIAN or __BIG_ENDIAN
-#include <sys/param.h>
-
-#ifdef __GNUC__
-#define PREFETCH(location) __builtin_prefetch(location)
-#else
-  // no prefetching
-#define PREFETCH(location) ;
-#endif
-#endif
-
-unsigned int Crc32Lookup[8][256] =
 {
 	{
 		0x00000000,0x77073096,0xEE0E612C,0x990951BA,0x076DC419,0x706AF48F,0xE963A535,0x9E6495A3,
@@ -301,20 +278,7 @@ unsigned int Crc32Lookup[8][256] =
 	  }
 };
 
-/// swap endianess
+uint32 crc32_calc_slice_by_8(uint32 previousCrc32, const void* data, size_t length)
-static inline uint32_t swap(uint32_t x)
-{
-#if defined(__GNUC__) || defined(__clang__)
-	return __builtin_bswap32(x);
-#else
-	return (x >> 24) |
-		((x >> 8) & 0x0000FF00) |
-		((x << 8) & 0x00FF0000) |
-		(x << 24);
-#endif
-}
-
-unsigned int crc32_calc_slice_by_8(unsigned int previousCrc32, const void* data, int length)
 {
 	uint32_t crc = ~previousCrc32; // same as previousCrc32 ^ 0xFFFFFFFF
 	const uint32_t* current = (const uint32_t*)data;
@@ -323,7 +287,7 @@ unsigned int crc32_calc_slice_by_8(unsigned int previousCrc32, const void* data,
 	while (length >= 8)
 	{
 		if constexpr (std::endian::native == std::endian::big){
-			uint32_t one = *current++ ^ swap(crc);
+			uint32_t one = *current++ ^ _swapEndianU32(crc);
 			uint32_t two = *current++;
 			crc = Crc32Lookup[0][two & 0xFF] ^
 				  Crc32Lookup[1][(two >> 8) & 0xFF] ^
@@ -348,13 +312,14 @@ unsigned int crc32_calc_slice_by_8(unsigned int previousCrc32, const void* data,
 				  Crc32Lookup[7][one & 0xFF];
 		}
 		else {
-			cemu_assert(false);
+			static_assert(std::endian::native == std::endian::big || std::endian::native == std::endian::little,
+						  "Platform byte-order is unsupported");
 		}
 
 		length -= 8;
 	}
 
-	const uint8_t* currentChar = (const uint8_t*)current;
+	const uint8* currentChar = (const uint8*)current;
 	// remaining 1 to 7 bytes (standard algorithm)
 	while (length-- != 0)
 		crc = (crc >> 8) ^ Crc32Lookup[0][(crc & 0xFF) ^ *currentChar++];
@@ -362,20 +327,20 @@ unsigned int crc32_calc_slice_by_8(unsigned int previousCrc32, const void* data,
 	return ~crc; // same as crc ^ 0xFFFFFFFF
 }
 
-unsigned int crc32_calc(unsigned int c, const void* data, int length)
+uint32 crc32_calc(uint32 c, const void* data, size_t length)
 {
 	if (length >= 16)
 	{
 		return crc32_calc_slice_by_8(c, data, length);
 	}
-	unsigned char* p = (unsigned char*)data;
+	const uint8* p = (const uint8*)data;
 	if (length == 0)
 		return c;
 	c ^= 0xFFFFFFFF;
 	while (length)
 	{
-		unsigned char temp = *p;
+		uint8 temp = *p;
-		temp ^= (unsigned char)c;
+		temp ^= (uint8)c;
 		c = (c >> 8) ^ Crc32Lookup[0][temp];
 		// next
 		length--;

src/util/crypto/crc32.h

@@ -1,8 +1,8 @@
 #pragma once
 
-unsigned int crc32_calc(unsigned int c, const void* data, int length);
+uint32 crc32_calc(uint32 c, const void* data, size_t length);
 
-inline unsigned int crc32_calc(const void* data, int length)
+inline uint32 crc32_calc(const void* data, size_t length)
 {
 	return crc32_calc(0, data, length);
 }