early-access version 1829

This commit is contained in:
pineappleEA 2021-06-26 10:15:46 +02:00
parent 230a0aa491
commit d2ffa9ae83
10 changed files with 93 additions and 67 deletions

View File

@ -1,7 +1,7 @@
yuzu emulator early access yuzu emulator early access
============= =============
This is the source code for early-access 1826. This is the source code for early-access 1829.
## Legal Notice ## Legal Notice

View File

@ -19,29 +19,26 @@
namespace Service::Nvidia::Devices { namespace Service::Nvidia::Devices {
namespace { namespace {
// Copies count amount of type T from the input vector into the dst vector. // Splice vectors will copy count amount of type T from the input vector into the dst vector.
// Returns the number of bytes written into dst.
template <typename T> template <typename T>
std::size_t SpliceVectors(const std::vector<u8>& input, std::vector<T>& dst, std::size_t count, std::size_t SpliceVectors(const std::vector<u8>& input, std::vector<T>& dst, std::size_t count,
std::size_t offset) { std::size_t offset) {
if (dst.empty()) { if (!dst.empty()) {
return 0; std::memcpy(dst.data(), input.data() + offset, count * sizeof(T));
} }
const size_t bytes_copied = count * sizeof(T); return 0;
std::memcpy(dst.data(), input.data() + offset, bytes_copied);
return bytes_copied;
} }
// Writes the data in src to an offset into the dst vector. The offset is specified in bytes // Write vectors will write data to the output buffer
// Returns the number of bytes written into dst.
template <typename T> template <typename T>
std::size_t WriteVectors(std::vector<u8>& dst, const std::vector<T>& src, std::size_t offset) { std::size_t WriteVectors(std::vector<u8>& dst, const std::vector<T>& src, std::size_t offset) {
if (src.empty()) { if (src.empty()) {
return 0; return 0;
} else {
std::memcpy(dst.data() + offset, src.data(), src.size() * sizeof(T));
offset += src.size() * sizeof(T);
return offset;
} }
const size_t bytes_copied = src.size() * sizeof(T);
std::memcpy(dst.data() + offset, src.data(), bytes_copied);
return bytes_copied;
} }
} // Anonymous namespace } // Anonymous namespace
@ -65,6 +62,7 @@ NvResult nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u
LOG_DEBUG(Service_NVDRV, "called NVDEC Submit, cmd_buffer_count={}", params.cmd_buffer_count); LOG_DEBUG(Service_NVDRV, "called NVDEC Submit, cmd_buffer_count={}", params.cmd_buffer_count);
// Instantiate param buffers // Instantiate param buffers
std::size_t offset = sizeof(IoctlSubmit);
std::vector<CommandBuffer> command_buffers(params.cmd_buffer_count); std::vector<CommandBuffer> command_buffers(params.cmd_buffer_count);
std::vector<Reloc> relocs(params.relocation_count); std::vector<Reloc> relocs(params.relocation_count);
std::vector<u32> reloc_shifts(params.relocation_count); std::vector<u32> reloc_shifts(params.relocation_count);
@ -73,13 +71,12 @@ NvResult nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u
std::vector<Fence> fences(params.fence_count); std::vector<Fence> fences(params.fence_count);
// Splice input into their respective buffers // Splice input into their respective buffers
std::size_t offset = sizeof(IoctlSubmit); offset = SpliceVectors(input, command_buffers, params.cmd_buffer_count, offset);
offset += SpliceVectors(input, command_buffers, params.cmd_buffer_count, offset); offset = SpliceVectors(input, relocs, params.relocation_count, offset);
offset += SpliceVectors(input, relocs, params.relocation_count, offset); offset = SpliceVectors(input, reloc_shifts, params.relocation_count, offset);
offset += SpliceVectors(input, reloc_shifts, params.relocation_count, offset); offset = SpliceVectors(input, syncpt_increments, params.syncpoint_count, offset);
offset += SpliceVectors(input, syncpt_increments, params.syncpoint_count, offset); offset = SpliceVectors(input, wait_checks, params.syncpoint_count, offset);
offset += SpliceVectors(input, wait_checks, params.syncpoint_count, offset); offset = SpliceVectors(input, fences, params.fence_count, offset);
offset += SpliceVectors(input, fences, params.fence_count, offset);
auto& gpu = system.GPU(); auto& gpu = system.GPU();
if (gpu.UseNvdec()) { if (gpu.UseNvdec()) {
@ -91,7 +88,7 @@ NvResult nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u
} }
} }
for (const auto& cmd_buffer : command_buffers) { for (const auto& cmd_buffer : command_buffers) {
const auto object = nvmap_dev->GetObject(cmd_buffer.memory_id); auto object = nvmap_dev->GetObject(cmd_buffer.memory_id);
ASSERT_OR_EXECUTE(object, return NvResult::InvalidState;); ASSERT_OR_EXECUTE(object, return NvResult::InvalidState;);
const auto map = FindBufferMap(object->dma_map_addr); const auto map = FindBufferMap(object->dma_map_addr);
if (!map) { if (!map) {
@ -105,19 +102,21 @@ NvResult nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u
gpu.PushCommandBuffer(cmdlist); gpu.PushCommandBuffer(cmdlist);
} }
if (gpu.UseNvdec()) { if (gpu.UseNvdec()) {
fences[0].value = syncpoint_manager.IncreaseSyncpoint(fences[0].id, 1); fences[0].value = syncpoint_manager.IncreaseSyncpoint(fences[0].id, 1);
Tegra::ChCommandHeaderList cmdlist{{(4 << 28) | fences[0].id}}; Tegra::ChCommandHeaderList cmdlist{{(4 << 28) | fences[0].id}};
gpu.PushCommandBuffer(cmdlist); gpu.PushCommandBuffer(cmdlist);
} }
std::memcpy(output.data(), &params, sizeof(IoctlSubmit)); std::memcpy(output.data(), &params, sizeof(IoctlSubmit));
// Some games expect command_buffers to be written back // Some games expect command_buffers to be written back
offset = sizeof(IoctlSubmit); offset = sizeof(IoctlSubmit);
offset += WriteVectors(output, command_buffers, offset); offset = WriteVectors(output, command_buffers, offset);
offset += WriteVectors(output, relocs, offset); offset = WriteVectors(output, relocs, offset);
offset += WriteVectors(output, reloc_shifts, offset); offset = WriteVectors(output, reloc_shifts, offset);
offset += WriteVectors(output, syncpt_increments, offset); offset = WriteVectors(output, syncpt_increments, offset);
offset += WriteVectors(output, wait_checks, offset); offset = WriteVectors(output, wait_checks, offset);
offset += WriteVectors(output, fences, offset); offset = WriteVectors(output, fences, offset);
return NvResult::Success; return NvResult::Success;
} }
@ -153,10 +152,10 @@ NvResult nvhost_nvdec_common::MapBuffer(const std::vector<u8>& input, std::vecto
auto& gpu = system.GPU(); auto& gpu = system.GPU();
for (auto& cmd_buffer : cmd_buffer_handles) { for (auto& cmf_buff : cmd_buffer_handles) {
auto object{nvmap_dev->GetObject(cmd_buffer.map_handle)}; auto object{nvmap_dev->GetObject(cmf_buff.map_handle)};
if (!object) { if (!object) {
LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmd_buffer.map_handle); LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle);
std::memcpy(output.data(), &params, output.size()); std::memcpy(output.data(), &params, output.size());
return NvResult::InvalidState; return NvResult::InvalidState;
} }
@ -171,7 +170,7 @@ NvResult nvhost_nvdec_common::MapBuffer(const std::vector<u8>& input, std::vecto
if (!object->dma_map_addr) { if (!object->dma_map_addr) {
LOG_ERROR(Service_NVDRV, "failed to map size={}", object->size); LOG_ERROR(Service_NVDRV, "failed to map size={}", object->size);
} else { } else {
cmd_buffer.map_address = object->dma_map_addr; cmf_buff.map_address = object->dma_map_addr;
AddBufferMap(object->dma_map_addr, object->size, object->addr, AddBufferMap(object->dma_map_addr, object->size, object->addr,
object->status == nvmap::Object::Status::Allocated); object->status == nvmap::Object::Status::Allocated);
} }
@ -191,10 +190,10 @@ NvResult nvhost_nvdec_common::UnmapBuffer(const std::vector<u8>& input, std::vec
auto& gpu = system.GPU(); auto& gpu = system.GPU();
for (auto& cmd_buffer : cmd_buffer_handles) { for (auto& cmf_buff : cmd_buffer_handles) {
const auto object{nvmap_dev->GetObject(cmd_buffer.map_handle)}; const auto object{nvmap_dev->GetObject(cmf_buff.map_handle)};
if (!object) { if (!object) {
LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmd_buffer.map_handle); LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle);
std::memcpy(output.data(), &params, output.size()); std::memcpy(output.data(), &params, output.size());
return NvResult::InvalidState; return NvResult::InvalidState;
} }

View File

@ -293,6 +293,7 @@ endif()
if (MSVC) if (MSVC)
target_compile_options(video_core PRIVATE target_compile_options(video_core PRIVATE
/we4267 # 'var' : conversion from 'size_t' to 'type', possible loss of data /we4267 # 'var' : conversion from 'size_t' to 'type', possible loss of data
/we4244 # 'var' : conversion from integer to 'type', possible loss of data
/we4456 # Declaration of 'identifier' hides previous local declaration /we4456 # Declaration of 'identifier' hides previous local declaration
/we4457 # Declaration of 'identifier' hides function parameter /we4457 # Declaration of 'identifier' hides function parameter
/we4458 # Declaration of 'identifier' hides class member /we4458 # Declaration of 'identifier' hides class member

View File

@ -99,7 +99,7 @@ class BufferCache {
}; };
public: public:
static constexpr u32 DEFAULT_SKIP_CACHE_SIZE = 4_KiB; static constexpr u32 DEFAULT_SKIP_CACHE_SIZE = static_cast<u32>(4_KiB);
explicit BufferCache(VideoCore::RasterizerInterface& rasterizer_, explicit BufferCache(VideoCore::RasterizerInterface& rasterizer_,
Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::Engines::Maxwell3D& maxwell3d_,
@ -109,8 +109,6 @@ public:
void TickFrame(); void TickFrame();
void RunGarbageCollector();
void WriteMemory(VAddr cpu_addr, u64 size); void WriteMemory(VAddr cpu_addr, u64 size);
void CachedWriteMemory(VAddr cpu_addr, u64 size); void CachedWriteMemory(VAddr cpu_addr, u64 size);
@ -197,6 +195,8 @@ private:
((cpu_addr + size) & ~Core::Memory::PAGE_MASK); ((cpu_addr + size) & ~Core::Memory::PAGE_MASK);
} }
void RunGarbageCollector();
void BindHostIndexBuffer(); void BindHostIndexBuffer();
void BindHostVertexBuffers(); void BindHostVertexBuffers();
@ -416,8 +416,9 @@ void BufferCache<P>::CachedWriteMemory(VAddr cpu_addr, u64 size) {
template <class P> template <class P>
void BufferCache<P>::DownloadMemory(VAddr cpu_addr, u64 size) { void BufferCache<P>::DownloadMemory(VAddr cpu_addr, u64 size) {
ForEachBufferInRange(cpu_addr, size, ForEachBufferInRange(cpu_addr, size, [&](BufferId, Buffer& buffer) {
[&](BufferId, Buffer& buffer) { DownloadBufferMemory(buffer); }); DownloadBufferMemory(buffer, cpu_addr, size);
});
} }
template <class P> template <class P>

View File

@ -14,10 +14,18 @@ extern "C" {
#pragma GCC diagnostic push #pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wconversion" #pragma GCC diagnostic ignored "-Wconversion"
#endif #endif
#ifdef _MSC_VER
#pragma warning(push)
#pragma warning(disable : 4242) // conversion from 'type' to 'type', possible loss of data
#pragma warning(disable : 4244) // conversion from 'type' to 'type', possible loss of data
#endif
#include <libavcodec/avcodec.h> #include <libavcodec/avcodec.h>
#if defined(__GNUC__) || defined(__clang__) #if defined(__GNUC__) || defined(__clang__)
#pragma GCC diagnostic pop #pragma GCC diagnostic pop
#endif #endif
#ifdef _MSC_VER
#pragma warning(pop)
#endif
} }
namespace Tegra { namespace Tegra {

View File

@ -3,7 +3,28 @@
// Refer to the license.txt file included. // Refer to the license.txt file included.
#include <array> #include <array>
extern "C" {
#if defined(__GNUC__) || defined(__clang__)
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wconversion"
#endif
#ifdef _MSC_VER
#pragma warning(disable : 4244) // conversion from 'type' to 'type', possible loss of data
#pragma warning(push)
#endif
#include <libswscale/swscale.h>
#if defined(__GNUC__) || defined(__clang__)
#pragma GCC diagnostic pop
#endif
#ifdef _MSC_VER
#pragma warning(pop)
#endif
}
#include "common/assert.h" #include "common/assert.h"
#include "common/logging/log.h"
#include "video_core/command_classes/nvdec.h" #include "video_core/command_classes/nvdec.h"
#include "video_core/command_classes/vic.h" #include "video_core/command_classes/vic.h"
#include "video_core/engines/maxwell_3d.h" #include "video_core/engines/maxwell_3d.h"
@ -11,10 +32,6 @@
#include "video_core/memory_manager.h" #include "video_core/memory_manager.h"
#include "video_core/textures/decoders.h" #include "video_core/textures/decoders.h"
extern "C" {
#include <libswscale/swscale.h>
}
namespace Tegra { namespace Tegra {
Vic::Vic(GPU& gpu_, std::shared_ptr<Nvdec> nvdec_processor_) Vic::Vic(GPU& gpu_, std::shared_ptr<Nvdec> nvdec_processor_)

View File

@ -55,8 +55,9 @@ size_t BytesPerIndex(VkIndexType index_type) {
template <typename T> template <typename T>
std::array<T, 6> MakeQuadIndices(u32 quad, u32 first) { std::array<T, 6> MakeQuadIndices(u32 quad, u32 first) {
std::array<T, 6> indices{0, 1, 2, 0, 2, 3}; std::array<T, 6> indices{0, 1, 2, 0, 2, 3};
std::ranges::transform(indices, indices.begin(), for (T& index : indices) {
[quad, first](u32 index) { return first + index + quad * 4; }); index = static_cast<T>(first + index + quad * 4);
}
return indices; return indices;
} }
} // Anonymous namespace } // Anonymous namespace

View File

@ -111,9 +111,6 @@ public:
/// Notify the cache that a new frame has been queued /// Notify the cache that a new frame has been queued
void TickFrame(); void TickFrame();
/// Runs the Garbage Collector.
void RunGarbageCollector();
/// Return a constant reference to the given image view id /// Return a constant reference to the given image view id
[[nodiscard]] const ImageView& GetImageView(ImageViewId id) const noexcept; [[nodiscard]] const ImageView& GetImageView(ImageViewId id) const noexcept;
@ -226,6 +223,9 @@ private:
} }
} }
/// Runs the Garbage Collector.
void RunGarbageCollector();
/// Fills image_view_ids in the image views in indices /// Fills image_view_ids in the image views in indices
void FillImageViews(DescriptorTable<TICEntry>& table, void FillImageViews(DescriptorTable<TICEntry>& table,
std::span<ImageViewId> cached_image_view_ids, std::span<const u32> indices, std::span<ImageViewId> cached_image_view_ids, std::span<const u32> indices,

View File

@ -842,9 +842,9 @@ void Device::CollectTelemetryParameters() {
void Device::CollectPhysicalMemoryInfo() { void Device::CollectPhysicalMemoryInfo() {
const auto mem_properties = physical.GetMemoryProperties(); const auto mem_properties = physical.GetMemoryProperties();
const std::size_t num_properties = mem_properties.memoryHeapCount; const size_t num_properties = mem_properties.memoryHeapCount;
device_access_memory = 0; device_access_memory = 0;
for (std::size_t element = 0; element < num_properties; element++) { for (size_t element = 0; element < num_properties; ++element) {
if ((mem_properties.memoryHeaps[element].flags & VK_MEMORY_HEAP_DEVICE_LOCAL_BIT) != 0) { if ((mem_properties.memoryHeaps[element].flags & VK_MEMORY_HEAP_DEVICE_LOCAL_BIT) != 0) {
device_access_memory += mem_properties.memoryHeaps[element].size; device_access_memory += mem_properties.memoryHeaps[element].size;
} }

View File

@ -267,21 +267,22 @@ private:
bool IsFormatSupported(VkFormat wanted_format, VkFormatFeatureFlags wanted_usage, bool IsFormatSupported(VkFormat wanted_format, VkFormatFeatureFlags wanted_usage,
FormatType format_type) const; FormatType format_type) const;
VkInstance instance; ///< Vulkan instance. VkInstance instance; ///< Vulkan instance.
vk::DeviceDispatch dld; ///< Device function pointers. vk::DeviceDispatch dld; ///< Device function pointers.
vk::PhysicalDevice physical; ///< Physical device. vk::PhysicalDevice physical; ///< Physical device.
VkPhysicalDeviceProperties properties; ///< Device properties. VkPhysicalDeviceProperties properties; ///< Device properties.
vk::Device logical; ///< Logical device. vk::Device logical; ///< Logical device.
vk::Queue graphics_queue; ///< Main graphics queue. vk::Queue graphics_queue; ///< Main graphics queue.
vk::Queue present_queue; ///< Main present queue. vk::Queue present_queue; ///< Main present queue.
u32 instance_version{}; ///< Vulkan onstance version. u32 instance_version{}; ///< Vulkan onstance version.
u32 graphics_family{}; ///< Main graphics queue family index. u32 graphics_family{}; ///< Main graphics queue family index.
u32 present_family{}; ///< Main present queue family index. u32 present_family{}; ///< Main present queue family index.
VkDriverIdKHR driver_id{}; ///< Driver ID. VkDriverIdKHR driver_id{}; ///< Driver ID.
VkShaderStageFlags guest_warp_stages{}; ///< Stages where the guest warp size can be forced.ed VkShaderStageFlags guest_warp_stages{}; ///< Stages where the guest warp size can be forced.
bool is_optimal_astc_supported{}; ///< Support for native ASTC. u64 device_access_memory{}; ///< Total size of device local memory in bytes.
bool is_float16_supported{}; ///< Support for float16 arithmetics. bool is_optimal_astc_supported{}; ///< Support for native ASTC.
bool is_warp_potentially_bigger{}; ///< Host warp size can be bigger than guest. bool is_float16_supported{}; ///< Support for float16 arithmetics.
bool is_warp_potentially_bigger{}; ///< Host warp size can be bigger than guest.
bool is_formatless_image_load_supported{}; ///< Support for shader image read without format. bool is_formatless_image_load_supported{}; ///< Support for shader image read without format.
bool is_shader_storage_image_multisample{}; ///< Support for image operations on MSAA images. bool is_shader_storage_image_multisample{}; ///< Support for image operations on MSAA images.
bool is_blit_depth_stencil_supported{}; ///< Support for blitting from and to depth stencil. bool is_blit_depth_stencil_supported{}; ///< Support for blitting from and to depth stencil.
@ -312,8 +313,6 @@ private:
/// Nsight Aftermath GPU crash tracker /// Nsight Aftermath GPU crash tracker
std::unique_ptr<NsightAftermathTracker> nsight_aftermath_tracker; std::unique_ptr<NsightAftermathTracker> nsight_aftermath_tracker;
u64 device_access_memory;
}; };
} // namespace Vulkan } // namespace Vulkan