Compare commits

...

82 Commits

Author SHA1 Message Date
David Marcec
8248d76964 Address issues 2020-07-20 11:52:07 +10:00
David Marcec
d7d2c27b48 swkbd: Return result for Calc request for inlined swkbd
Fixes random swkbd popups in monster hunter
2020-07-19 18:56:16 +10:00
Rodrigo Locatti
ad0b295125 Merge pull request #4376 from ogniK5377/dark-wait-tree
frontend: Improve wait tree readability for dark themes
2020-07-19 03:55:36 -03:00
LC
8a8924902b Merge pull request #4379 from jbeich/libc++
alignment: unbreak build with Clang
2020-07-18 19:22:56 -04:00
Jan Beich
db43b37e15 alignment: explicitly include <new> after 723edb4c06
In file included from src/core/hle/kernel/memory/page_table.cpp:5:
src/./common/alignment.h:67:68: error: no member named 'align_val_t' in namespace 'std'
        return static_cast<T*>(::operator new (n * sizeof(T), std::align_val_t{Align}));
                                                              ~~~~~^
src/./common/alignment.h:71:51: error: no member named 'align_val_t' in namespace 'std'
        ::operator delete (p, n * sizeof(T), std::align_val_t{Align});
                                             ~~~~~^
2020-07-18 23:06:16 +00:00
David Marcec
617eeb09e0 Address issues 2020-07-19 01:02:41 +10:00
David
9943a478fe Merge pull request #3349 from FearlessTobi/translationnns
[WIP] yuzu: Port translation support from Citra v2
2020-07-18 22:43:37 +10:00
FearlessTobi
4450a2688a Address trivial review comments. 2020-07-18 14:09:13 +02:00
FearlessTobi
0072003a14 configure_ui: Address some review comments from the previous PR 2020-07-18 14:09:13 +02:00
FearlessTobi
347b50ad43 yuzu: Port translation support from Citra
Co-Authored-By: Weiyi Wang <wwylele@gmail.com>
2020-07-18 14:09:11 +02:00
David Marcec
db4502b7b7 frontend: Improve wait tree readability for dark themes 2020-07-18 20:22:03 +10:00
bunnei
4a8cb9a706 Merge pull request #4348 from lioncash/nano
core_timing: Make usage of nanoseconds more consistent in the interface
2020-07-18 01:45:10 -04:00
bunnei
bbeea7502c Merge pull request #4373 from lioncash/allocator
alignment: Simplify AlignmentAllocator implementation
2020-07-18 01:04:17 -04:00
bunnei
c700079e08 Merge pull request #4345 from Morph1984/fix-createfile
filesystem: Create subdirectories prior to creating a file
2020-07-18 00:59:26 -04:00
bunnei
90cbcaa44a Merge pull request #4273 from ogniK5377/async-shaders-prod
video_core: Add asynchronous shader decompilation and compilation
2020-07-18 00:48:27 -04:00
David Marcec
967307d3be Fix style issues 2020-07-18 14:24:32 +10:00
bunnei
821d295f24 Merge pull request #4364 from lioncash/desig5
vulkan: Make use of designated initializers where applicable
2020-07-18 00:12:43 -04:00
bunnei
af1151b1b0 Merge pull request #4365 from lioncash/mii
mii/manager: Make use of designated initializers
2020-07-17 22:20:15 -04:00
bunnei
0b733903f0 Merge pull request #4374 from ReinUsesLisp/fix-err
vk_device: Fix build error on old MSVC versions
2020-07-17 20:30:24 -04:00
bunnei
47b305387a Merge pull request #4366 from lioncash/mii-sign
mii/manager: Resolve sign mismatch warnings
2020-07-17 20:13:06 -04:00
ReinUsesLisp
81c8f92f2e vk_device: Fix build error on old MSVC versions
Designated initializers on old MSVC versions fail to build when they
take the address of a constant.
2020-07-17 20:27:53 -03:00
bunnei
5d95e62443 Merge pull request #4344 from VolcaEM/c3
dmnt_cheat_vm: Implement opcode 0xC3 (ReadWriteStaticRegister)
2020-07-17 17:11:52 -04:00
bunnei
0f0b756775 Merge pull request #4309 from Morph1984/fix-romfs-bug
fs: Fix RomFS building when zero byte files are present
2020-07-17 17:01:20 -04:00
bunnei
19c6bf72db Merge pull request #4322 from ReinUsesLisp/fix-dynstate
vk_state_tracker: Fix dirty flags for stencil_enable on VK_EXT_extended_dynamic_state
2020-07-17 09:50:45 -04:00
Lioncash
723edb4c06 alignment: Simplify AlignmentAllocator implementation
With C++20, much of the allocator interface has been simplified, so we
can make the same adjustments.
2020-07-17 08:30:12 -04:00
LC
47956a3bbc Merge pull request #4369 from lioncash/hle-macro
macro_hle: Remove unnecessary std::make_pair calls
2020-07-17 05:20:41 -04:00
LC
9d3cbf6a90 Merge pull request #4340 from lioncash/remove
shader_cache: Make use of std::erase_if
2020-07-17 05:19:20 -04:00
David Marcec
4f473cda64 Drop settings namespace 2020-07-17 17:23:24 +10:00
David Marcec
85b591f6f0 Remove duplicate config 2020-07-17 14:26:18 +10:00
David Marcec
f48187449e Use conditional var 2020-07-17 14:26:17 +10:00
David Marcec
2ba195aa0d Drop max workers from 8->2 for testing 2020-07-17 14:26:15 +10:00
David Marcec
85d7a8f466 Rebase for per game settings 2020-07-17 14:26:14 +10:00
David Marcec
468bd9c1b0 async shaders 2020-07-17 14:24:57 +10:00
David
c783cf443e Merge pull request #4347 from lioncash/logging
settings: Make use of std::string_view over std::string for logging
2020-07-17 13:25:06 +10:00
David
92f37a229e Merge pull request #4371 from lioncash/cmake2
core/CMakeLists: Add missing physical_memory.h header file
2020-07-17 13:22:19 +10:00
David
adbf5ca50b Merge pull request #4357 from lioncash/unused4
kernel: Remove unused variables
2020-07-17 13:18:31 +10:00
David
69f8b6a53e Merge pull request #4358 from lioncash/unused5
kernel/thread: Remove unimplemented function prototype
2020-07-17 13:17:52 +10:00
Lioncash
c0650cd82c macro_hle: Remove unnecessary static keywords
These functions are already in an anonymous namespace which makes the
functions internally linked.
2020-07-16 23:17:17 -04:00
David
0d10c863a5 Merge pull request #4367 from lioncash/inc2
constants: Add missing <array> include
2020-07-17 13:14:01 +10:00
David
9cca0c2f83 Merge pull request #4368 from lioncash/macro
macro: Resolve missing parameter in doxygen comment
2020-07-17 13:13:22 +10:00
David
3ce4edba64 Merge pull request #4370 from lioncash/simplify
macro_hle: Simplify shift expression in HLE_771BB18C62444DA0()
2020-07-17 13:13:05 +10:00
Lioncash
311f500753 core/CMakeLists: Add missing physical_memory.h header file
Allows this header file to show up in IDE CMake generators.
2020-07-16 22:56:31 -04:00
Lioncash
be6b7591d9 macro_hle: Simplify shift expression in HLE_771BB18C62444DA0()
Given the expression involves a 32-bit value, this simplifies down to
just: 0x3ffffff. This is likely a remnant from testing that was never
cleaned up.

Resolves a -Wshift-overflow warning.
2020-07-16 22:16:11 -04:00
Lioncash
cc935d997b macro_hle: Remove unnecessary std::make_pair calls
The purpose of make_pair is generally to deduce the types within the
pair without explicitly specifying the types, so these usages were
generally unnecessary, particularly when the type is enforced by the
array declaration.
2020-07-16 21:59:25 -04:00
Lioncash
502dbfb9eb macro: Resolve missing parameter in doxygen comment
Resolves a -Wdocumentation warning.
2020-07-16 21:54:42 -04:00
Lioncash
e07eb5b223 constants: Add missing <array> include
Eliminates reliance on an indirect include.
2020-07-16 21:43:20 -04:00
Rodrigo Locatti
39ae2deb28 Merge pull request #4363 from lioncash/mismatch
vk_texture_cache: Amend mismatched access masks and indices in UploadBuffer
2020-07-16 21:54:40 -03:00
Lioncash
7f989378c0 mii/manager: Make use of designated initializers
Allows returning the structure in a more concise manner.
2020-07-16 20:45:33 -04:00
Lioncash
7785123b1c wrapper: Make use of designated initializers where applicable 2020-07-16 20:01:01 -04:00
Lioncash
01da386617 vk_texture_cache: Make use of designated initializers where applicable 2020-07-16 19:52:38 -04:00
Lioncash
169759e069 vk_texture_cache: Amend mismatched access masks and indices in UploadBuffer
Discovered while converting relevant parts of the codebase over to
designated initializers.
2020-07-16 19:45:46 -04:00
Lioncash
08d36afd40 vk_swapchain: Make use of designated initializers where applicable 2020-07-16 19:27:02 -04:00
Lioncash
3c060503bc vk_stream_buffer: Make use of designated initializers where applicable 2020-07-16 19:22:11 -04:00
Lioncash
70147e913f vk_staging_buffer_pool: Make use of designated initializers where applicable 2020-07-16 19:22:03 -04:00
Lioncash
2025f847bb vk_shader_util: Make use of designated initializers where applicable 2020-07-16 19:17:41 -04:00
Lioncash
97e7663004 vk_scheduler: Make use of designated initializers where applicable 2020-07-16 19:11:43 -04:00
Lioncash
fd7af52ec3 vk_sampler_cache: Make use of designated initializers where applicable 2020-07-16 19:06:40 -04:00
Lioncash
772b6e4d28 vk_resource_manager: Make use of designated initializers where applicable 2020-07-16 19:02:35 -04:00
Lioncash
8ebd6a21c5 vk_renderpass_cache: Make use of designated initializers where applicable 2020-07-16 18:57:23 -04:00
Lioncash
01f297f2e0 vk_rasterizer: Make use of designated initializers where applicable 2020-07-16 18:49:42 -04:00
Lioncash
c07b0ffe47 vk_query_cache: Make use of designated initializers where applicable 2020-07-16 18:34:04 -04:00
Lioncash
d43e923990 vk_pipeline_cache: Make use of designated initializers where applicable 2020-07-16 18:32:29 -04:00
Lioncash
7d5f93832c vk_memory_manager: Make use of designated initializers where applicable 2020-07-16 18:26:30 -04:00
Lioncash
75c00c3cb0 vk_image: Make use of designated initializers where applicable 2020-07-16 18:24:26 -04:00
Lioncash
6d165481ad vk_descriptor_pool: Make use of designated initializers where applicable 2020-07-16 18:19:45 -04:00
Lioncash
5330ca396d vk_compute_pipeline: Make use of designated initializers where applicable 2020-07-16 17:32:12 -04:00
Lioncash
757ddd8158 vk_compute_pass: Make use of designated initializers where applicable
Note: Some barriers can't be converted over yet, as they ICE MSVC.
2020-07-16 17:23:56 -04:00
Lioncash
a66a0a6a53 vk_buffer_cache: Make use of designated initializers where applicable
Note: An array within CopyFrom() cannot be converted over yet, as it
ICEs MSVC when converted over.
2020-07-16 16:59:39 -04:00
Lioncash
8bef49cde5 kernel/thread: Remove unimplemented function prototype
This isn't used, so it can be removed.
2020-07-16 14:32:46 -04:00
Lioncash
2bab07c367 kernel: Remove unused variables
Resolves some compiler warnings in the Linux build.
2020-07-16 14:17:50 -04:00
Morph
5892fc1555 Add comment to clarify the nullptr check 2020-07-16 00:15:27 -04:00
Morph
2079bb4090 filesystem: Create subdirectories prior to creating a file
If subdirectories exist in the given path parameter and don't exist in the real filesystem create them prior to creating the files within.
This fixes the softlocks upon save creation in The Legend of Zelda: Breath of the Wild
2020-07-16 00:15:27 -04:00
Lioncash
0435b7d361 core_timing: Remove unused data member
Shrinks the size of the CoreTiming class by 8 bytes.
2020-07-15 19:41:37 -04:00
Lioncash
bef1844a51 core_timing: Make TimedCallback take std::chrono::nanoseconds
Enforces our desired time units directly with a concrete type.
2020-07-15 19:41:22 -04:00
Lioncash
8b50c660df core_timing: Make use of std::chrono with ScheduleEvent 2020-07-15 18:54:15 -04:00
Lioncash
af5a56ddc4 settings: Resolve a sign conversion warning within GetTimeZoneString()
A sign conversion warning was occurring due to an int < size_t
comparison.
2020-07-15 13:45:22 -04:00
Lioncash
5dbf91d739 settings: Make use of std::string_view over std::string for logging
In all usages of LogSetting(), string literals are provided.
std::string_view is better suited here, as we won't churn a bunch of
string allocations every time the settings are logged out.

While we're at it, we can fold LogSetting() into LogSettings(), given
it's only ever used there.
2020-07-15 13:37:33 -04:00
VolcaEM
e90802e762 clang-format 2020-07-15 01:22:52 +02:00
VolcaEM
b608acd688 dmnt_cheat_vm: Implement opcode 0xC3 (ReadWriteStaticRegister)
This was based on Atmosphére's DMNT Cheat VM:

- https://github.com/Atmosphere-NX/Atmosphere/blob/master/stratosphere/dmnt/source/cheat/impl/dmnt_cheat_vm.hpp
- https://github.com/Atmosphere-NX/Atmosphere/blob/master/stratosphere/dmnt/source/cheat/impl/dmnt_cheat_vm.cpp

From Atmosphére's documentation: "Code type 0xC3 reads or writes a static register with a given register"

There are now only two remaining opcodes to implement (PauseProcess and BreakProcess)

This is untested because I don't have any experience in testing cheats on yuzu
2020-07-15 01:19:22 +02:00
Lioncash
969100d41a shader_cache: Make use of std::erase_if
Now that we use C++20, we can also make use of std::erase_if instead of
needing to do the erase-remove idiom.
2020-07-14 15:49:15 -04:00
ReinUsesLisp
0fe09df386 vk_state_tracker: Fix dirty flags for stencil_enable on VK_EXT_extended_dynamic_state
Fixes a regression on any game using stencil on devices with
VK_EXT_extended_dynamic_state.
2020-07-12 20:43:42 -03:00
Morph
fd1c3aa14e fs: Fix RomFS building when zero byte files are present
When zero byte files are present, the key (offset) for that file is identical to the file right after. A std::map isn't able to fit key-value pairs with identical keys (offsets), therefore, the solution is to use std::multimap which permits multiple entries with the same key.

This most prominently fixes Pokemon Sword and Shield weather with any RomFS mod applied.
2020-07-12 04:33:35 -04:00
96 changed files with 2063 additions and 1117 deletions

View File

@@ -5,7 +5,7 @@ cd /yuzu
ccache -s
mkdir build || true && cd build
cmake .. -G Ninja -DDISPLAY_VERSION=$1 -DYUZU_USE_BUNDLED_UNICORN=ON -DYUZU_USE_QT_WEB_ENGINE=ON -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=/usr/lib/ccache/gcc -DCMAKE_CXX_COMPILER=/usr/lib/ccache/g++ -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${ENABLE_COMPATIBILITY_REPORTING:-"OFF"} -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DUSE_DISCORD_PRESENCE=ON
cmake .. -G Ninja -DDISPLAY_VERSION=$1 -DYUZU_USE_BUNDLED_UNICORN=ON -DYUZU_USE_QT_WEB_ENGINE=ON -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=/usr/lib/ccache/gcc -DCMAKE_CXX_COMPILER=/usr/lib/ccache/g++ -DYUZU_ENABLE_COMPATIBILITY_REPORTING=${ENABLE_COMPATIBILITY_REPORTING:-"OFF"} -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DUSE_DISCORD_PRESENCE=ON -DENABLE_QT_TRANSLATION=ON
ninja

View File

@@ -5,7 +5,7 @@ cd /yuzu
ccache -s
mkdir build || true && cd build
cmake .. -G Ninja -DDISPLAY_VERSION=$1 -DCMAKE_TOOLCHAIN_FILE="$(pwd)/../CMakeModules/MinGWCross.cmake" -DUSE_CCACHE=ON -DYUZU_USE_BUNDLED_UNICORN=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DCMAKE_BUILD_TYPE=Release
cmake .. -G Ninja -DDISPLAY_VERSION=$1 -DCMAKE_TOOLCHAIN_FILE="$(pwd)/../CMakeModules/MinGWCross.cmake" -DUSE_CCACHE=ON -DYUZU_USE_BUNDLED_UNICORN=ON -DENABLE_COMPATIBILITY_LIST_DOWNLOAD=ON -DCMAKE_BUILD_TYPE=Release -DENABLE_QT_TRANSLATION=ON
ninja
ccache -s

View File

@@ -13,6 +13,7 @@ project(yuzu)
option(ENABLE_SDL2 "Enable the SDL2 frontend" ON)
option(ENABLE_QT "Enable the Qt frontend" ON)
option(ENABLE_QT_TRANSLATION "Enable translations for the Qt frontend" OFF)
CMAKE_DEPENDENT_OPTION(YUZU_USE_BUNDLED_QT "Download bundled Qt binaries" ON "ENABLE_QT;MSVC" OFF)
option(ENABLE_WEB_SERVICE "Enable web services (telemetry, etc.)" ON)
@@ -224,6 +225,10 @@ if(ENABLE_QT)
if (YUZU_USE_QT_WEB_ENGINE)
find_package(Qt5 COMPONENTS WebEngineCore WebEngineWidgets)
endif()
if (ENABLE_QT_TRANSLATION)
find_package(Qt5 REQUIRED COMPONENTS LinguistTools ${QT_PREFIX_HINT})
endif()
if (NOT Qt5_FOUND)
list(APPEND CONAN_REQUIRED_LIBS "qt/5.14.1@bincrafters/stable")
endif()

View File

@@ -24,6 +24,8 @@ Most of the development happens on GitHub. It's also where [our central reposito
If you want to contribute please take a look at the [Contributor's Guide](https://github.com/yuzu-emu/yuzu/wiki/Contributing) and [Developer Information](https://github.com/yuzu-emu/yuzu/wiki/Developer-Information). You should also contact any of the developers on Discord in order to know about the current state of the emulator.
If you want to contribute to the user interface translation, please check out the [yuzu project on transifex](https://www.transifex.com/yuzu-emulator/yuzu). We centralize translation work there, and periodically upstream translations.
### Building
* __Windows__: [Windows Build](https://github.com/yuzu-emu/yuzu/wiki/Building-For-Windows)

2
dist/languages/.gitignore vendored Normal file
View File

@@ -0,0 +1,2 @@
# Ignore the source language file
en.ts

8
dist/languages/.tx/config vendored Normal file
View File

@@ -0,0 +1,8 @@
[main]
host = https://www.transifex.com
[yuzu.emulator]
file_filter = <lang>.ts
source_file = en.ts
source_lang = en
type = QT

1
dist/languages/README.md vendored Normal file
View File

@@ -0,0 +1 @@
This directory stores translation patches (TS files) for yuzu Qt frontend. This directory is linked with [yuzu project on transifex](https://www.transifex.com/yuzu-emulator/yuzu), so you can update the translation by executing `tx pull -a`. If you want to contribute to the translation, please go the transifex link and submit your translation there. This directory on the main repo will be synchronized with transifex periodically. Do not directly open PRs on github to modify the translation.

View File

@@ -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::CreateEvent(
name, [this](u64 userdata, s64 cycles_late) { ReleaseActiveBuffer(cycles_late); });
name, [this](u64, std::chrono::nanoseconds ns_late) { ReleaseActiveBuffer(ns_late); });
}
void Stream::Play() {
@@ -59,11 +59,9 @@ Stream::State Stream::GetState() const {
return state;
}
s64 Stream::GetBufferReleaseNS(const Buffer& buffer) const {
std::chrono::nanoseconds Stream::GetBufferReleaseNS(const Buffer& buffer) const {
const std::size_t num_samples{buffer.GetSamples().size() / GetNumChannels()};
const auto ns =
std::chrono::nanoseconds((static_cast<u64>(num_samples) * 1000000000ULL) / sample_rate);
return ns.count();
return std::chrono::nanoseconds((static_cast<u64>(num_samples) * 1000000000ULL) / sample_rate);
}
static void VolumeAdjustSamples(std::vector<s16>& samples, float game_volume) {
@@ -80,7 +78,7 @@ static void VolumeAdjustSamples(std::vector<s16>& samples, float game_volume) {
}
}
void Stream::PlayNextBuffer(s64 cycles_late) {
void Stream::PlayNextBuffer(std::chrono::nanoseconds ns_late) {
if (!IsPlaying()) {
// Ensure we are in playing state before playing the next buffer
sink_stream.Flush();
@@ -105,17 +103,18 @@ void Stream::PlayNextBuffer(s64 cycles_late) {
sink_stream.EnqueueSamples(GetNumChannels(), active_buffer->GetSamples());
core_timing.ScheduleEvent(
GetBufferReleaseNS(*active_buffer) -
(Settings::values.enable_audio_stretching.GetValue() ? 0 : cycles_late),
release_event, {});
const auto time_stretch_delta = Settings::values.enable_audio_stretching.GetValue()
? std::chrono::nanoseconds::zero()
: ns_late;
const auto future_time = GetBufferReleaseNS(*active_buffer) - time_stretch_delta;
core_timing.ScheduleEvent(future_time, release_event, {});
}
void Stream::ReleaseActiveBuffer(s64 cycles_late) {
void Stream::ReleaseActiveBuffer(std::chrono::nanoseconds ns_late) {
ASSERT(active_buffer);
released_buffers.push(std::move(active_buffer));
release_callback();
PlayNextBuffer(cycles_late);
PlayNextBuffer(ns_late);
}
bool Stream::QueueBuffer(BufferPtr&& buffer) {

View File

@@ -4,6 +4,7 @@
#pragma once
#include <chrono>
#include <functional>
#include <memory>
#include <string>
@@ -90,16 +91,13 @@ public:
private:
/// Plays the next queued buffer in the audio stream, starting playback if necessary
void PlayNextBuffer(s64 cycles_late = 0);
void PlayNextBuffer(std::chrono::nanoseconds ns_late = {});
/// Releases the actively playing buffer, signalling that it has been completed
void ReleaseActiveBuffer(s64 cycles_late = 0);
void ReleaseActiveBuffer(std::chrono::nanoseconds ns_late = {});
/// Gets the number of core cycles when the specified buffer will be released
s64 GetBufferReleaseNS(const Buffer& buffer) const;
/// Gets the number of core cycles when the specified buffer will be released
s64 GetBufferReleaseNSHostTiming(const Buffer& buffer) const;
std::chrono::nanoseconds GetBufferReleaseNS(const Buffer& buffer) const;
u32 sample_rate; ///< Sample rate of the stream
Format format; ///< Format of the stream

View File

@@ -3,7 +3,7 @@
#pragma once
#include <cstddef>
#include <memory>
#include <new>
#include <type_traits>
namespace Common {
@@ -54,66 +54,28 @@ public:
using size_type = std::size_t;
using difference_type = std::ptrdiff_t;
using pointer = T*;
using const_pointer = const T*;
using reference = T&;
using const_reference = const T&;
using propagate_on_container_copy_assignment = std::true_type;
using propagate_on_container_move_assignment = std::true_type;
using propagate_on_container_swap = std::true_type;
using is_always_equal = std::true_type;
public:
constexpr AlignmentAllocator() noexcept = default;
template <typename T2>
constexpr AlignmentAllocator(const AlignmentAllocator<T2, Align>&) noexcept {}
pointer address(reference r) noexcept {
return std::addressof(r);
T* allocate(size_type n) {
return static_cast<T*>(::operator new (n * sizeof(T), std::align_val_t{Align}));
}
const_pointer address(const_reference r) const noexcept {
return std::addressof(r);
}
pointer allocate(size_type n) {
return static_cast<pointer>(::operator new (n, std::align_val_t{Align}));
}
void deallocate(pointer p, size_type) {
::operator delete (p, std::align_val_t{Align});
}
void construct(pointer p, const value_type& wert) {
new (p) value_type(wert);
}
void destroy(pointer p) {
p->~value_type();
}
size_type max_size() const noexcept {
return size_type(-1) / sizeof(value_type);
void deallocate(T* p, size_type n) {
::operator delete (p, n * sizeof(T), std::align_val_t{Align});
}
template <typename T2>
struct rebind {
using other = AlignmentAllocator<T2, Align>;
};
bool operator!=(const AlignmentAllocator<T, Align>& other) const noexcept {
return !(*this == other);
}
// Returns true if and only if storage allocated from *this
// can be deallocated from other, and vice versa.
// Always returns true for stateless allocators.
bool operator==(const AlignmentAllocator<T, Align>& other) const noexcept {
return true;
}
};
} // namespace Common

View File

@@ -185,6 +185,7 @@ add_library(core STATIC
hle/kernel/object.h
hle/kernel/physical_core.cpp
hle/kernel/physical_core.h
hle/kernel/physical_memory.h
hle/kernel/process.cpp
hle/kernel/process.h
hle/kernel/process_capability.cpp

View File

@@ -4,6 +4,7 @@
#pragma once
#include <array>
#include "common/common_types.h"
// This is to consolidate system-wide constants that are used by multiple components of yuzu.

View File

@@ -53,12 +53,12 @@ void CoreTiming::ThreadEntry(CoreTiming& instance) {
instance.ThreadLoop();
}
void CoreTiming::Initialize(std::function<void(void)>&& on_thread_init_) {
void CoreTiming::Initialize(std::function<void()>&& on_thread_init_) {
on_thread_init = std::move(on_thread_init_);
event_fifo_id = 0;
shutting_down = false;
ticks = 0;
const auto empty_timed_callback = [](u64, s64) {};
const auto empty_timed_callback = [](u64, std::chrono::nanoseconds) {};
ev_lost = CreateEvent("_lost_event", empty_timed_callback);
if (is_multicore) {
timer_thread = std::make_unique<std::thread>(ThreadEntry, std::ref(*this));
@@ -106,11 +106,11 @@ bool CoreTiming::HasPendingEvents() const {
return !(wait_set && event_queue.empty());
}
void CoreTiming::ScheduleEvent(s64 ns_into_future, const std::shared_ptr<EventType>& event_type,
u64 userdata) {
void CoreTiming::ScheduleEvent(std::chrono::nanoseconds ns_into_future,
const std::shared_ptr<EventType>& event_type, u64 userdata) {
{
std::scoped_lock scope{basic_lock};
const u64 timeout = static_cast<u64>(GetGlobalTimeNs().count() + ns_into_future);
const u64 timeout = static_cast<u64>((GetGlobalTimeNs() + ns_into_future).count());
event_queue.emplace_back(Event{timeout, event_fifo_id++, userdata, event_type});
@@ -195,8 +195,9 @@ std::optional<s64> CoreTiming::Advance() {
event_queue.pop_back();
basic_lock.unlock();
if (auto event_type{evt.type.lock()}) {
event_type->callback(evt.userdata, global_timer - evt.time);
if (const auto event_type{evt.type.lock()}) {
event_type->callback(
evt.userdata, std::chrono::nanoseconds{static_cast<s64>(global_timer - evt.time)});
}
basic_lock.lock();

View File

@@ -17,14 +17,12 @@
#include "common/common_types.h"
#include "common/spin_lock.h"
#include "common/thread.h"
#include "common/threadsafe_queue.h"
#include "common/wall_clock.h"
#include "core/hardware_properties.h"
namespace Core::Timing {
/// A callback that may be scheduled for a particular core timing event.
using TimedCallback = std::function<void(u64 userdata, s64 cycles_late)>;
using TimedCallback = std::function<void(u64 userdata, std::chrono::nanoseconds ns_late)>;
/// Contains the characteristics of a particular event.
struct EventType {
@@ -42,12 +40,12 @@ struct EventType {
* in main CPU clock cycles.
*
* To schedule an event, you first have to register its type. This is where you pass in the
* callback. You then schedule events using the type id you get back.
* callback. You then schedule events using the type ID you get back.
*
* The int cyclesLate that the callbacks get is how many cycles late it was.
* The s64 ns_late that the callbacks get is how many ns late it was.
* So to schedule a new event on a regular basis:
* inside callback:
* ScheduleEvent(periodInCycles - cyclesLate, callback, "whatever")
* ScheduleEvent(period_in_ns - ns_late, callback, "whatever")
*/
class CoreTiming {
public:
@@ -62,7 +60,7 @@ public:
/// CoreTiming begins at the boundary of timing slice -1. An initial call to Advance() is
/// required to end slice - 1 and start slice 0 before the first cycle of code is executed.
void Initialize(std::function<void(void)>&& on_thread_init_);
void Initialize(std::function<void()>&& on_thread_init_);
/// Tears down all timing related functionality.
void Shutdown();
@@ -95,8 +93,8 @@ public:
bool HasPendingEvents() const;
/// Schedules an event in core timing
void ScheduleEvent(s64 ns_into_future, const std::shared_ptr<EventType>& event_type,
u64 userdata = 0);
void ScheduleEvent(std::chrono::nanoseconds ns_into_future,
const std::shared_ptr<EventType>& event_type, u64 userdata = 0);
void UnscheduleEvent(const std::shared_ptr<EventType>& event_type, u64 userdata);
@@ -141,8 +139,6 @@ private:
u64 global_timer = 0;
std::chrono::nanoseconds start_point;
// The queue is a min-heap using std::make_heap/push_heap/pop_heap.
// We don't use std::priority_queue because we need to be able to serialize, unserialize and
// erase arbitrary events (RemoveEvent()) regardless of the queue order. These aren't
@@ -161,7 +157,7 @@ private:
std::atomic<bool> wait_set{};
std::atomic<bool> shutting_down{};
std::atomic<bool> has_started{};
std::function<void(void)> on_thread_init{};
std::function<void()> on_thread_init{};
bool is_multicore{};

View File

@@ -240,7 +240,7 @@ RomFSBuildContext::RomFSBuildContext(VirtualDir base_, VirtualDir ext_)
RomFSBuildContext::~RomFSBuildContext() = default;
std::map<u64, VirtualFile> RomFSBuildContext::Build() {
std::multimap<u64, VirtualFile> RomFSBuildContext::Build() {
const u64 dir_hash_table_entry_count = romfs_get_hash_table_count(num_dirs);
const u64 file_hash_table_entry_count = romfs_get_hash_table_count(num_files);
dir_hash_table_size = 4 * dir_hash_table_entry_count;
@@ -294,7 +294,7 @@ std::map<u64, VirtualFile> RomFSBuildContext::Build() {
cur_dir->parent->child = cur_dir;
}
std::map<u64, VirtualFile> out;
std::multimap<u64, VirtualFile> out;
// Populate file tables.
for (const auto& it : files) {

View File

@@ -43,7 +43,7 @@ public:
~RomFSBuildContext();
// This finalizes the context.
std::map<u64, VirtualFile> Build();
std::multimap<u64, VirtualFile> Build();
private:
VirtualDir base;

View File

@@ -11,7 +11,7 @@
namespace FileSys {
static bool VerifyConcatenationMapContinuity(const std::map<u64, VirtualFile>& map) {
static bool VerifyConcatenationMapContinuity(const std::multimap<u64, VirtualFile>& map) {
const auto last_valid = --map.end();
for (auto iter = map.begin(); iter != last_valid;) {
const auto old = iter++;
@@ -27,12 +27,12 @@ ConcatenatedVfsFile::ConcatenatedVfsFile(std::vector<VirtualFile> files_, std::s
: name(std::move(name)) {
std::size_t next_offset = 0;
for (const auto& file : files_) {
files[next_offset] = file;
files.emplace(next_offset, file);
next_offset += file->GetSize();
}
}
ConcatenatedVfsFile::ConcatenatedVfsFile(std::map<u64, VirtualFile> files_, std::string name)
ConcatenatedVfsFile::ConcatenatedVfsFile(std::multimap<u64, VirtualFile> files_, std::string name)
: files(std::move(files_)), name(std::move(name)) {
ASSERT(VerifyConcatenationMapContinuity(files));
}
@@ -50,7 +50,7 @@ VirtualFile ConcatenatedVfsFile::MakeConcatenatedFile(std::vector<VirtualFile> f
}
VirtualFile ConcatenatedVfsFile::MakeConcatenatedFile(u8 filler_byte,
std::map<u64, VirtualFile> files,
std::multimap<u64, VirtualFile> files,
std::string name) {
if (files.empty())
return nullptr;

View File

@@ -15,7 +15,7 @@ namespace FileSys {
// read-only.
class ConcatenatedVfsFile : public VfsFile {
ConcatenatedVfsFile(std::vector<VirtualFile> files, std::string name);
ConcatenatedVfsFile(std::map<u64, VirtualFile> files, std::string name);
ConcatenatedVfsFile(std::multimap<u64, VirtualFile> files, std::string name);
public:
~ConcatenatedVfsFile() override;
@@ -25,7 +25,7 @@ public:
/// Convenience function that turns a map of offsets to files into a concatenated file, filling
/// gaps with a given filler byte.
static VirtualFile MakeConcatenatedFile(u8 filler_byte, std::map<u64, VirtualFile> files,
static VirtualFile MakeConcatenatedFile(u8 filler_byte, std::multimap<u64, VirtualFile> files,
std::string name);
std::string GetName() const override;
@@ -40,7 +40,7 @@ public:
private:
// Maps starting offset to file -- more efficient.
std::map<u64, VirtualFile> files;
std::multimap<u64, VirtualFile> files;
std::string name;
};

View File

@@ -11,19 +11,20 @@
namespace Core::Hardware {
InterruptManager::InterruptManager(Core::System& system_in) : system(system_in) {
gpu_interrupt_event = Core::Timing::CreateEvent("GPUInterrupt", [this](u64 message, s64) {
auto nvdrv = system.ServiceManager().GetService<Service::Nvidia::NVDRV>("nvdrv");
const u32 syncpt = static_cast<u32>(message >> 32);
const u32 value = static_cast<u32>(message);
nvdrv->SignalGPUInterruptSyncpt(syncpt, value);
});
gpu_interrupt_event =
Core::Timing::CreateEvent("GPUInterrupt", [this](u64 message, std::chrono::nanoseconds) {
auto nvdrv = system.ServiceManager().GetService<Service::Nvidia::NVDRV>("nvdrv");
const u32 syncpt = static_cast<u32>(message >> 32);
const u32 value = static_cast<u32>(message);
nvdrv->SignalGPUInterruptSyncpt(syncpt, value);
});
}
InterruptManager::~InterruptManager() = default;
void InterruptManager::GPUInterruptSyncpt(const u32 syncpoint_id, const u32 value) {
const u64 msg = (static_cast<u64>(syncpoint_id) << 32ULL) | value;
system.CoreTiming().ScheduleEvent(10, gpu_interrupt_event, msg);
system.CoreTiming().ScheduleEvent(std::chrono::nanoseconds{10}, gpu_interrupt_event, msg);
}
} // namespace Core::Hardware

View File

@@ -24,7 +24,6 @@ namespace Kernel {
// Wake up num_to_wake (or all) threads in a vector.
void AddressArbiter::WakeThreads(const std::vector<std::shared_ptr<Thread>>& waiting_threads,
s32 num_to_wake) {
auto& time_manager = system.Kernel().TimeManager();
// Only process up to 'target' threads, unless 'target' is <= 0, in which case process
// them all.
std::size_t last = waiting_threads.size();

View File

@@ -145,16 +145,18 @@ struct KernelCore::Impl {
void InitializePreemption(KernelCore& kernel) {
preemption_event = Core::Timing::CreateEvent(
"PreemptionCallback", [this, &kernel](u64 userdata, s64 cycles_late) {
"PreemptionCallback", [this, &kernel](u64, std::chrono::nanoseconds) {
{
SchedulerLock lock(kernel);
global_scheduler.PreemptThreads();
}
s64 time_interval = Core::Timing::msToCycles(std::chrono::milliseconds(10));
const auto time_interval = std::chrono::nanoseconds{
Core::Timing::msToCycles(std::chrono::milliseconds(10))};
system.CoreTiming().ScheduleEvent(time_interval, preemption_event);
});
s64 time_interval = Core::Timing::msToCycles(std::chrono::milliseconds(10));
const auto time_interval =
std::chrono::nanoseconds{Core::Timing::msToCycles(std::chrono::milliseconds(10))};
system.CoreTiming().ScheduleEvent(time_interval, preemption_event);
}

View File

@@ -34,7 +34,7 @@ ResultVal<std::shared_ptr<ServerSession>> ServerSession::Create(KernelCore& kern
std::shared_ptr<ServerSession> session{std::make_shared<ServerSession>(kernel)};
session->request_event = Core::Timing::CreateEvent(
name, [session](u64 userdata, s64 cycles_late) { session->CompleteSyncRequest(); });
name, [session](u64, std::chrono::nanoseconds) { session->CompleteSyncRequest(); });
session->name = std::move(name);
session->parent = std::move(parent);
@@ -184,8 +184,8 @@ ResultCode ServerSession::CompleteSyncRequest() {
ResultCode ServerSession::HandleSyncRequest(std::shared_ptr<Thread> thread,
Core::Memory::Memory& memory) {
ResultCode result = QueueSyncRequest(std::move(thread), memory);
const u64 delay = kernel.IsMulticore() ? 0U : 20000U;
const ResultCode result = QueueSyncRequest(std::move(thread), memory);
const auto delay = std::chrono::nanoseconds{kernel.IsMulticore() ? 0 : 20000};
Core::System::GetInstance().CoreTiming().ScheduleEvent(delay, request_event, {});
return result;
}

View File

@@ -458,9 +458,7 @@ static ResultCode WaitSynchronization(Core::System& system, Handle* index, VAddr
return ERR_OUT_OF_RANGE;
}
auto* const thread = system.CurrentScheduler().GetCurrentThread();
auto& kernel = system.Kernel();
using ObjectPtr = Thread::ThreadSynchronizationObjects::value_type;
Thread::ThreadSynchronizationObjects objects(handle_count);
const auto& handle_table = kernel.CurrentProcess()->GetHandleTable();
@@ -1750,9 +1748,9 @@ static void SignalProcessWideKey(Core::System& system, VAddr condition_variable_
// Only process up to 'target' threads, unless 'target' is less equal 0, in which case process
// them all.
std::size_t last = waiting_threads.size();
if (target > 0)
if (target > 0) {
last = std::min(waiting_threads.size(), static_cast<std::size_t>(target));
auto& time_manager = kernel.TimeManager();
}
for (std::size_t index = 0; index < last; ++index) {
auto& thread = waiting_threads[index];
@@ -1763,7 +1761,6 @@ static void SignalProcessWideKey(Core::System& system, VAddr condition_variable_
const std::size_t current_core = system.CurrentCoreIndex();
auto& monitor = system.Monitor();
auto& memory = system.Memory();
// Atomically read the value of the mutex.
u32 mutex_val = 0;

View File

@@ -19,7 +19,6 @@ Synchronization::Synchronization(Core::System& system) : system{system} {}
void Synchronization::SignalObject(SynchronizationObject& obj) const {
auto& kernel = system.Kernel();
SchedulerLock lock(kernel);
auto& time_manager = kernel.TimeManager();
if (obj.IsSignaled()) {
for (auto thread : obj.GetWaitingThreads()) {
if (thread->GetSchedulingStatus() == ThreadSchedStatus::Paused) {

View File

@@ -583,8 +583,6 @@ private:
void SetCurrentPriority(u32 new_priority);
void AdjustSchedulingOnAffinity(u64 old_affinity_mask, s32 old_core);
Common::SpinLock context_guard{};
ThreadContext32 context_32{};
ThreadContext64 context_64{};

View File

@@ -16,7 +16,7 @@ namespace Kernel {
TimeManager::TimeManager(Core::System& system_) : system{system_} {
time_manager_event_type = Core::Timing::CreateEvent(
"Kernel::TimeManagerCallback", [this](u64 thread_handle, [[maybe_unused]] s64 cycles_late) {
"Kernel::TimeManagerCallback", [this](u64 thread_handle, std::chrono::nanoseconds) {
SchedulerLock lock(system.Kernel());
Handle proper_handle = static_cast<Handle>(thread_handle);
if (cancelled_events[proper_handle]) {
@@ -34,7 +34,8 @@ void TimeManager::ScheduleTimeEvent(Handle& event_handle, Thread* timetask, s64
ASSERT(timetask);
ASSERT(timetask->GetStatus() != ThreadStatus::Ready);
ASSERT(timetask->GetStatus() != ThreadStatus::WaitMutex);
system.CoreTiming().ScheduleEvent(nanoseconds, time_manager_event_type, event_handle);
system.CoreTiming().ScheduleEvent(std::chrono::nanoseconds{nanoseconds},
time_manager_event_type, event_handle);
} else {
event_handle = InvalidHandle;
}

View File

@@ -13,11 +13,23 @@
namespace Service::AM::Applets {
namespace {
enum class Request : u32 {
Finalize = 0x4,
SetUserWordInfo = 0x6,
SetCustomizeDic = 0x7,
Calc = 0xa,
SetCustomizedDictionaries = 0xb,
UnsetCustomizedDictionaries = 0xc,
UnknownD = 0xd,
UnknownE = 0xe,
};
constexpr std::size_t SWKBD_INLINE_INIT_SIZE = 0x8;
constexpr std::size_t SWKBD_OUTPUT_BUFFER_SIZE = 0x7D8;
constexpr std::size_t SWKBD_OUTPUT_INTERACTIVE_BUFFER_SIZE = 0x7D4;
constexpr std::size_t DEFAULT_MAX_LENGTH = 500;
constexpr bool INTERACTIVE_STATUS_OK = false;
} // Anonymous namespace
static Core::Frontend::SoftwareKeyboardParameters ConvertToFrontendParameters(
KeyboardConfig config, std::u16string initial_text) {
Core::Frontend::SoftwareKeyboardParameters params{};
@@ -47,6 +59,7 @@ SoftwareKeyboard::~SoftwareKeyboard() = default;
void SoftwareKeyboard::Initialize() {
complete = false;
is_inline = false;
initial_text.clear();
final_data.clear();
@@ -56,6 +69,11 @@ void SoftwareKeyboard::Initialize() {
ASSERT(keyboard_config_storage != nullptr);
const auto& keyboard_config = keyboard_config_storage->GetData();
if (keyboard_config.size() == SWKBD_INLINE_INIT_SIZE) {
is_inline = true;
return;
}
ASSERT(keyboard_config.size() >= sizeof(KeyboardConfig));
std::memcpy(&config, keyboard_config.data(), sizeof(KeyboardConfig));
@@ -87,16 +105,32 @@ void SoftwareKeyboard::ExecuteInteractive() {
const auto storage = broker.PopInteractiveDataToApplet();
ASSERT(storage != nullptr);
const auto data = storage->GetData();
const auto status = static_cast<bool>(data[0]);
if (status == INTERACTIVE_STATUS_OK) {
complete = true;
if (!is_inline) {
const auto status = static_cast<bool>(data[0]);
if (status == INTERACTIVE_STATUS_OK) {
complete = true;
} else {
std::array<char16_t, SWKBD_OUTPUT_INTERACTIVE_BUFFER_SIZE / 2 - 2> string;
std::memcpy(string.data(), data.data() + 4, string.size() * 2);
frontend.SendTextCheckDialog(
Common::UTF16StringFromFixedZeroTerminatedBuffer(string.data(), string.size()),
[this] { broker.SignalStateChanged(); });
}
} else {
std::array<char16_t, SWKBD_OUTPUT_INTERACTIVE_BUFFER_SIZE / 2 - 2> string;
std::memcpy(string.data(), data.data() + 4, string.size() * 2);
frontend.SendTextCheckDialog(
Common::UTF16StringFromFixedZeroTerminatedBuffer(string.data(), string.size()),
[this] { broker.SignalStateChanged(); });
Request request{};
std::memcpy(&request, data.data(), sizeof(Request));
switch (request) {
case Request::Calc: {
broker.PushNormalDataFromApplet(
std::make_shared<IStorage>(std::move(std::vector<u8>{1})));
broker.SignalStateChanged();
break;
}
default:
UNIMPLEMENTED_MSG("Request {:X} is not implemented", request);
break;
}
}
}
@@ -108,9 +142,10 @@ void SoftwareKeyboard::Execute() {
}
const auto parameters = ConvertToFrontendParameters(config, initial_text);
frontend.RequestText([this](std::optional<std::u16string> text) { WriteText(std::move(text)); },
parameters);
if (!is_inline) {
frontend.RequestText(
[this](std::optional<std::u16string> text) { WriteText(std::move(text)); }, parameters);
}
}
void SoftwareKeyboard::WriteText(std::optional<std::u16string> text) {

View File

@@ -78,6 +78,7 @@ private:
KeyboardConfig config;
std::u16string initial_text;
bool complete = false;
bool is_inline = false;
std::vector<u8> final_data;
};

View File

@@ -55,6 +55,10 @@ std::string VfsDirectoryServiceWrapper::GetName() const {
ResultCode VfsDirectoryServiceWrapper::CreateFile(const std::string& path_, u64 size) const {
std::string path(FileUtil::SanitizePath(path_));
auto dir = GetDirectoryRelativeWrapped(backing, FileUtil::GetParentPath(path));
// dir can be nullptr if path contains subdirectories, create those prior to creating the file.
if (dir == nullptr) {
dir = backing->CreateSubdirectory(FileUtil::GetParentPath(path));
}
auto file = dir->CreateFile(FileUtil::GetFilename(path));
if (file == nullptr) {
// TODO(DarkLordZach): Find a better error code for this

View File

@@ -39,9 +39,10 @@ namespace Service::HID {
// Updating period for each HID device.
// TODO(ogniK): Find actual polling rate of hid
constexpr s64 pad_update_ticks = static_cast<s64>(1000000000 / 66);
[[maybe_unused]] constexpr s64 accelerometer_update_ticks = static_cast<s64>(1000000000 / 100);
[[maybe_unused]] constexpr s64 gyroscope_update_ticks = static_cast<s64>(1000000000 / 100);
constexpr auto pad_update_ns = std::chrono::nanoseconds{1000000000 / 66};
[[maybe_unused]] constexpr auto accelerometer_update_ns =
std::chrono::nanoseconds{1000000000 / 100};
[[maybe_unused]] constexpr auto gyroscope_update_ticks = std::chrono::nanoseconds{1000000000 / 100};
constexpr std::size_t SHARED_MEMORY_SIZE = 0x40000;
IAppletResource::IAppletResource(Core::System& system)
@@ -75,14 +76,14 @@ IAppletResource::IAppletResource(Core::System& system)
GetController<Controller_Stubbed>(HidController::Unknown3).SetCommonHeaderOffset(0x5000);
// Register update callbacks
pad_update_event =
Core::Timing::CreateEvent("HID::UpdatePadCallback", [this](u64 userdata, s64 ns_late) {
pad_update_event = Core::Timing::CreateEvent(
"HID::UpdatePadCallback", [this](u64 userdata, std::chrono::nanoseconds ns_late) {
UpdateControllers(userdata, ns_late);
});
// TODO(shinyquagsire23): Other update callbacks? (accel, gyro?)
system.CoreTiming().ScheduleEvent(pad_update_ticks, pad_update_event);
system.CoreTiming().ScheduleEvent(pad_update_ns, pad_update_event);
ReloadInputDevices();
}
@@ -107,7 +108,7 @@ void IAppletResource::GetSharedMemoryHandle(Kernel::HLERequestContext& ctx) {
rb.PushCopyObjects(shared_mem);
}
void IAppletResource::UpdateControllers(u64 userdata, s64 ns_late) {
void IAppletResource::UpdateControllers(u64 userdata, std::chrono::nanoseconds ns_late) {
auto& core_timing = system.CoreTiming();
const bool should_reload = Settings::values.is_device_reload_pending.exchange(false);
@@ -118,7 +119,7 @@ void IAppletResource::UpdateControllers(u64 userdata, s64 ns_late) {
controller->OnUpdate(core_timing, shared_mem->GetPointer(), SHARED_MEMORY_SIZE);
}
core_timing.ScheduleEvent(pad_update_ticks - ns_late, pad_update_event);
core_timing.ScheduleEvent(pad_update_ns - ns_late, pad_update_event);
}
class IActiveVibrationDeviceList final : public ServiceFramework<IActiveVibrationDeviceList> {

View File

@@ -4,10 +4,9 @@
#pragma once
#include "core/hle/service/hid/controllers/controller_base.h"
#include "core/hle/service/service.h"
#include <chrono>
#include "controllers/controller_base.h"
#include "core/hle/service/hid/controllers/controller_base.h"
#include "core/hle/service/service.h"
namespace Core::Timing {
@@ -65,7 +64,7 @@ private:
}
void GetSharedMemoryHandle(Kernel::HLERequestContext& ctx);
void UpdateControllers(u64 userdata, s64 cycles_late);
void UpdateControllers(u64 userdata, std::chrono::nanoseconds ns_late);
std::shared_ptr<Kernel::SharedMemory> shared_mem;

View File

@@ -47,59 +47,60 @@ std::array<T, DestArraySize> ResizeArray(const std::array<T, SourceArraySize>& i
MiiInfo ConvertStoreDataToInfo(const MiiStoreData& data) {
MiiStoreBitFields bf;
std::memcpy(&bf, data.data.data.data(), sizeof(MiiStoreBitFields));
MiiInfo info{};
info.name = ResizeArray<char16_t, 10, 11>(data.data.name);
info.uuid = data.data.uuid;
info.font_region = static_cast<u8>(bf.font_region.Value());
info.favorite_color = static_cast<u8>(bf.favorite_color.Value());
info.gender = static_cast<u8>(bf.gender.Value());
info.height = static_cast<u8>(bf.height.Value());
info.build = static_cast<u8>(bf.build.Value());
info.type = static_cast<u8>(bf.type.Value());
info.region_move = static_cast<u8>(bf.region_move.Value());
info.faceline_type = static_cast<u8>(bf.faceline_type.Value());
info.faceline_color = static_cast<u8>(bf.faceline_color.Value());
info.faceline_wrinkle = static_cast<u8>(bf.faceline_wrinkle.Value());
info.faceline_make = static_cast<u8>(bf.faceline_makeup.Value());
info.hair_type = static_cast<u8>(bf.hair_type.Value());
info.hair_color = static_cast<u8>(bf.hair_color.Value());
info.hair_flip = static_cast<u8>(bf.hair_flip.Value());
info.eye_type = static_cast<u8>(bf.eye_type.Value());
info.eye_color = static_cast<u8>(bf.eye_color.Value());
info.eye_scale = static_cast<u8>(bf.eye_scale.Value());
info.eye_aspect = static_cast<u8>(bf.eye_aspect.Value());
info.eye_rotate = static_cast<u8>(bf.eye_rotate.Value());
info.eye_x = static_cast<u8>(bf.eye_x.Value());
info.eye_y = static_cast<u8>(bf.eye_y.Value());
info.eyebrow_type = static_cast<u8>(bf.eyebrow_type.Value());
info.eyebrow_color = static_cast<u8>(bf.eyebrow_color.Value());
info.eyebrow_scale = static_cast<u8>(bf.eyebrow_scale.Value());
info.eyebrow_aspect = static_cast<u8>(bf.eyebrow_aspect.Value());
info.eyebrow_rotate = static_cast<u8>(bf.eyebrow_rotate.Value());
info.eyebrow_x = static_cast<u8>(bf.eyebrow_x.Value());
info.eyebrow_y = static_cast<u8>(bf.eyebrow_y.Value() + 3);
info.nose_type = static_cast<u8>(bf.nose_type.Value());
info.nose_scale = static_cast<u8>(bf.nose_scale.Value());
info.nose_y = static_cast<u8>(bf.nose_y.Value());
info.mouth_type = static_cast<u8>(bf.mouth_type.Value());
info.mouth_color = static_cast<u8>(bf.mouth_color.Value());
info.mouth_scale = static_cast<u8>(bf.mouth_scale.Value());
info.mouth_aspect = static_cast<u8>(bf.mouth_aspect.Value());
info.mouth_y = static_cast<u8>(bf.mouth_y.Value());
info.beard_color = static_cast<u8>(bf.beard_color.Value());
info.beard_type = static_cast<u8>(bf.beard_type.Value());
info.mustache_type = static_cast<u8>(bf.mustache_type.Value());
info.mustache_scale = static_cast<u8>(bf.mustache_scale.Value());
info.mustache_y = static_cast<u8>(bf.mustache_y.Value());
info.glasses_type = static_cast<u8>(bf.glasses_type.Value());
info.glasses_color = static_cast<u8>(bf.glasses_color.Value());
info.glasses_scale = static_cast<u8>(bf.glasses_scale.Value());
info.glasses_y = static_cast<u8>(bf.glasses_y.Value());
info.mole_type = static_cast<u8>(bf.mole_type.Value());
info.mole_scale = static_cast<u8>(bf.mole_scale.Value());
info.mole_x = static_cast<u8>(bf.mole_x.Value());
info.mole_y = static_cast<u8>(bf.mole_y.Value());
return info;
return {
.uuid = data.data.uuid,
.name = ResizeArray<char16_t, 10, 11>(data.data.name),
.font_region = static_cast<u8>(bf.font_region.Value()),
.favorite_color = static_cast<u8>(bf.favorite_color.Value()),
.gender = static_cast<u8>(bf.gender.Value()),
.height = static_cast<u8>(bf.height.Value()),
.build = static_cast<u8>(bf.build.Value()),
.type = static_cast<u8>(bf.type.Value()),
.region_move = static_cast<u8>(bf.region_move.Value()),
.faceline_type = static_cast<u8>(bf.faceline_type.Value()),
.faceline_color = static_cast<u8>(bf.faceline_color.Value()),
.faceline_wrinkle = static_cast<u8>(bf.faceline_wrinkle.Value()),
.faceline_make = static_cast<u8>(bf.faceline_makeup.Value()),
.hair_type = static_cast<u8>(bf.hair_type.Value()),
.hair_color = static_cast<u8>(bf.hair_color.Value()),
.hair_flip = static_cast<u8>(bf.hair_flip.Value()),
.eye_type = static_cast<u8>(bf.eye_type.Value()),
.eye_color = static_cast<u8>(bf.eye_color.Value()),
.eye_scale = static_cast<u8>(bf.eye_scale.Value()),
.eye_aspect = static_cast<u8>(bf.eye_aspect.Value()),
.eye_rotate = static_cast<u8>(bf.eye_rotate.Value()),
.eye_x = static_cast<u8>(bf.eye_x.Value()),
.eye_y = static_cast<u8>(bf.eye_y.Value()),
.eyebrow_type = static_cast<u8>(bf.eyebrow_type.Value()),
.eyebrow_color = static_cast<u8>(bf.eyebrow_color.Value()),
.eyebrow_scale = static_cast<u8>(bf.eyebrow_scale.Value()),
.eyebrow_aspect = static_cast<u8>(bf.eyebrow_aspect.Value()),
.eyebrow_rotate = static_cast<u8>(bf.eyebrow_rotate.Value()),
.eyebrow_x = static_cast<u8>(bf.eyebrow_x.Value()),
.eyebrow_y = static_cast<u8>(bf.eyebrow_y.Value() + 3),
.nose_type = static_cast<u8>(bf.nose_type.Value()),
.nose_scale = static_cast<u8>(bf.nose_scale.Value()),
.nose_y = static_cast<u8>(bf.nose_y.Value()),
.mouth_type = static_cast<u8>(bf.mouth_type.Value()),
.mouth_color = static_cast<u8>(bf.mouth_color.Value()),
.mouth_scale = static_cast<u8>(bf.mouth_scale.Value()),
.mouth_aspect = static_cast<u8>(bf.mouth_aspect.Value()),
.mouth_y = static_cast<u8>(bf.mouth_y.Value()),
.beard_color = static_cast<u8>(bf.beard_color.Value()),
.beard_type = static_cast<u8>(bf.beard_type.Value()),
.mustache_type = static_cast<u8>(bf.mustache_type.Value()),
.mustache_scale = static_cast<u8>(bf.mustache_scale.Value()),
.mustache_y = static_cast<u8>(bf.mustache_y.Value()),
.glasses_type = static_cast<u8>(bf.glasses_type.Value()),
.glasses_color = static_cast<u8>(bf.glasses_color.Value()),
.glasses_scale = static_cast<u8>(bf.glasses_scale.Value()),
.glasses_y = static_cast<u8>(bf.glasses_y.Value()),
.mole_type = static_cast<u8>(bf.mole_type.Value()),
.mole_scale = static_cast<u8>(bf.mole_scale.Value()),
.mole_x = static_cast<u8>(bf.mole_x.Value()),
.mole_y = static_cast<u8>(bf.mole_y.Value()),
};
}
u16 GenerateCrc16(const void* data, std::size_t size) {

View File

@@ -28,8 +28,7 @@
namespace Service::NVFlinger {
constexpr s64 frame_ticks = static_cast<s64>(1000000000 / 60);
constexpr s64 frame_ticks_30fps = static_cast<s64>(1000000000 / 30);
constexpr auto frame_ns = std::chrono::nanoseconds{1000000000 / 60};
void NVFlinger::VSyncThread(NVFlinger& nv_flinger) {
nv_flinger.SplitVSync();
@@ -67,20 +66,24 @@ NVFlinger::NVFlinger(Core::System& system) : system(system) {
guard = std::make_shared<std::mutex>();
// Schedule the screen composition events
composition_event =
Core::Timing::CreateEvent("ScreenComposition", [this](u64 userdata, s64 ns_late) {
composition_event = Core::Timing::CreateEvent(
"ScreenComposition", [this](u64, std::chrono::nanoseconds ns_late) {
Lock();
Compose();
const auto ticks = GetNextTicks();
this->system.CoreTiming().ScheduleEvent(std::max<s64>(0LL, ticks - ns_late),
composition_event);
const auto ticks = std::chrono::nanoseconds{GetNextTicks()};
const auto ticks_delta = ticks - ns_late;
const auto future_ns = std::max(std::chrono::nanoseconds::zero(), ticks_delta);
this->system.CoreTiming().ScheduleEvent(future_ns, composition_event);
});
if (system.IsMulticore()) {
is_running = true;
wait_event = std::make_unique<Common::Event>();
vsync_thread = std::make_unique<std::thread>(VSyncThread, std::ref(*this));
} else {
system.CoreTiming().ScheduleEvent(frame_ticks, composition_event);
system.CoreTiming().ScheduleEvent(frame_ns, composition_event);
}
}

View File

@@ -20,7 +20,7 @@
namespace Core::Memory {
constexpr s64 CHEAT_ENGINE_TICKS = static_cast<s64>(1000000000 / 12);
constexpr auto CHEAT_ENGINE_NS = std::chrono::nanoseconds{1000000000 / 12};
constexpr u32 KEYPAD_BITMASK = 0x3FFFFFF;
StandardVmCallbacks::StandardVmCallbacks(Core::System& system, const CheatProcessMetadata& metadata)
@@ -188,10 +188,12 @@ CheatEngine::~CheatEngine() {
}
void CheatEngine::Initialize() {
event = Core::Timing::CreateEvent(
"CheatEngine::FrameCallback::" + Common::HexToString(metadata.main_nso_build_id),
[this](u64 userdata, s64 ns_late) { FrameCallback(userdata, ns_late); });
core_timing.ScheduleEvent(CHEAT_ENGINE_TICKS, event);
event = Core::Timing::CreateEvent("CheatEngine::FrameCallback::" +
Common::HexToString(metadata.main_nso_build_id),
[this](u64 userdata, std::chrono::nanoseconds ns_late) {
FrameCallback(userdata, ns_late);
});
core_timing.ScheduleEvent(CHEAT_ENGINE_NS, event);
metadata.process_id = system.CurrentProcess()->GetProcessID();
metadata.title_id = system.CurrentProcess()->GetTitleID();
@@ -217,7 +219,7 @@ void CheatEngine::Reload(std::vector<CheatEntry> cheats) {
MICROPROFILE_DEFINE(Cheat_Engine, "Add-Ons", "Cheat Engine", MP_RGB(70, 200, 70));
void CheatEngine::FrameCallback(u64 userdata, s64 ns_late) {
void CheatEngine::FrameCallback(u64, std::chrono::nanoseconds ns_late) {
if (is_pending_reload.exchange(false)) {
vm.LoadProgram(cheats);
}
@@ -230,7 +232,7 @@ void CheatEngine::FrameCallback(u64 userdata, s64 ns_late) {
vm.Execute(metadata);
core_timing.ScheduleEvent(CHEAT_ENGINE_TICKS - ns_late, event);
core_timing.ScheduleEvent(CHEAT_ENGINE_NS - ns_late, event);
}
} // namespace Core::Memory

View File

@@ -5,6 +5,7 @@
#pragma once
#include <atomic>
#include <chrono>
#include <memory>
#include <vector>
#include "common/common_types.h"
@@ -71,7 +72,7 @@ public:
void Reload(std::vector<CheatEntry> cheats);
private:
void FrameCallback(u64 userdata, s64 cycles_late);
void FrameCallback(u64 userdata, std::chrono::nanoseconds ns_late);
DmntCheatVm vm;
CheatProcessMetadata metadata;

View File

@@ -190,6 +190,15 @@ void DmntCheatVm::LogOpcode(const CheatVmOpcode& opcode) {
callbacks->CommandLog(
fmt::format("Act[{:02X}]: {:d}", i, save_restore_regmask->should_operate[i]));
}
} else if (auto rw_static_reg = std::get_if<ReadWriteStaticRegisterOpcode>(&opcode.opcode)) {
callbacks->CommandLog("Opcode: Read/Write Static Register");
if (rw_static_reg->static_idx < NumReadableStaticRegisters) {
callbacks->CommandLog("Op Type: ReadStaticRegister");
} else {
callbacks->CommandLog("Op Type: WriteStaticRegister");
}
callbacks->CommandLog(fmt::format("Reg Idx {:X}", rw_static_reg->idx));
callbacks->CommandLog(fmt::format("Stc Idx {:X}", rw_static_reg->static_idx));
} else if (auto debug_log = std::get_if<DebugLogOpcode>(&opcode.opcode)) {
callbacks->CommandLog("Opcode: Debug Log");
callbacks->CommandLog(fmt::format("Bit Width: {:X}", debug_log->bit_width));
@@ -544,6 +553,16 @@ bool DmntCheatVm::DecodeNextOpcode(CheatVmOpcode& out) {
}
opcode.opcode = save_restore_regmask;
} break;
case CheatVmOpcodeType::ReadWriteStaticRegister: {
ReadWriteStaticRegisterOpcode rw_static_reg{};
// C3000XXx
// C3 = opcode 0xC3.
// XX = static register index.
// x = register index.
rw_static_reg.static_idx = ((first_dword >> 4) & 0xFF);
rw_static_reg.idx = (first_dword & 0xF);
opcode.opcode = rw_static_reg;
} break;
case CheatVmOpcodeType::DebugLog: {
DebugLogOpcode debug_log{};
// FFFTIX##
@@ -667,6 +686,7 @@ void DmntCheatVm::ResetState() {
registers.fill(0);
saved_values.fill(0);
loop_tops.fill(0);
static_registers.fill(0);
instruction_ptr = 0;
condition_depth = 0;
decode_success = true;
@@ -1153,6 +1173,15 @@ void DmntCheatVm::Execute(const CheatProcessMetadata& metadata) {
}
}
}
} else if (auto rw_static_reg =
std::get_if<ReadWriteStaticRegisterOpcode>(&cur_opcode.opcode)) {
if (rw_static_reg->static_idx < NumReadableStaticRegisters) {
// Load a register with a static register.
registers[rw_static_reg->idx] = static_registers[rw_static_reg->static_idx];
} else {
// Store a register to a static register.
static_registers[rw_static_reg->static_idx] = registers[rw_static_reg->idx];
}
} else if (auto debug_log = std::get_if<DebugLogOpcode>(&cur_opcode.opcode)) {
// Read value from memory.
u64 log_value = 0;

View File

@@ -56,6 +56,7 @@ enum class CheatVmOpcodeType : u32 {
BeginRegisterConditionalBlock = 0xC0,
SaveRestoreRegister = 0xC1,
SaveRestoreRegisterMask = 0xC2,
ReadWriteStaticRegister = 0xC3,
// This is a meta entry, and not a real opcode.
// This is to facilitate multi-nybble instruction decoding.
@@ -237,6 +238,11 @@ struct SaveRestoreRegisterMaskOpcode {
std::array<bool, 0x10> should_operate{};
};
struct ReadWriteStaticRegisterOpcode {
u32 static_idx{};
u32 idx{};
};
struct DebugLogOpcode {
u32 bit_width{};
u32 log_id{};
@@ -259,7 +265,8 @@ struct CheatVmOpcode {
PerformArithmeticStaticOpcode, BeginKeypressConditionalOpcode,
PerformArithmeticRegisterOpcode, StoreRegisterToAddressOpcode,
BeginRegisterConditionalOpcode, SaveRestoreRegisterOpcode,
SaveRestoreRegisterMaskOpcode, DebugLogOpcode, UnrecognizedInstruction>
SaveRestoreRegisterMaskOpcode, ReadWriteStaticRegisterOpcode, DebugLogOpcode,
UnrecognizedInstruction>
opcode{};
};
@@ -281,6 +288,10 @@ public:
static constexpr std::size_t MaximumProgramOpcodeCount = 0x400;
static constexpr std::size_t NumRegisters = 0x10;
static constexpr std::size_t NumReadableStaticRegisters = 0x80;
static constexpr std::size_t NumWritableStaticRegisters = 0x80;
static constexpr std::size_t NumStaticRegisters =
NumReadableStaticRegisters + NumWritableStaticRegisters;
explicit DmntCheatVm(std::unique_ptr<Callbacks> callbacks);
~DmntCheatVm();
@@ -302,6 +313,7 @@ private:
std::array<u32, MaximumProgramOpcodeCount> program{};
std::array<u64, NumRegisters> registers{};
std::array<u64, NumRegisters> saved_values{};
std::array<u64, NumStaticRegisters> static_registers{};
std::array<std::size_t, NumRegisters> loop_tops{};
bool DecodeNextOpcode(CheatVmOpcode& out);

View File

@@ -2,6 +2,8 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <string_view>
#include "common/file_util.h"
#include "core/core.h"
#include "core/gdbstub/gdbstub.h"
@@ -65,18 +67,18 @@ Values values = {};
bool configuring_global = true;
std::string GetTimeZoneString() {
static constexpr std::array<const char*, 46> timezones{{
static constexpr std::array timezones{
"auto", "default", "CET", "CST6CDT", "Cuba", "EET", "Egypt", "Eire",
"EST", "EST5EDT", "GB", "GB-Eire", "GMT", "GMT+0", "GMT-0", "GMT0",
"Greenwich", "Hongkong", "HST", "Iceland", "Iran", "Israel", "Jamaica", "Japan",
"Kwajalein", "Libya", "MET", "MST", "MST7MDT", "Navajo", "NZ", "NZ-CHAT",
"Poland", "Portugal", "PRC", "PST8PDT", "ROC", "ROK", "Singapore", "Turkey",
"UCT", "Universal", "UTC", "W-SU", "WET", "Zulu",
}};
};
ASSERT(Settings::values.time_zone_index.GetValue() < timezones.size());
return timezones[Settings::values.time_zone_index.GetValue()];
const auto time_zone_index = static_cast<std::size_t>(values.time_zone_index.GetValue());
ASSERT(time_zone_index < timezones.size());
return timezones[time_zone_index];
}
void Apply() {
@@ -91,41 +93,41 @@ void Apply() {
Service::HID::ReloadInputDevices();
}
template <typename T>
void LogSetting(const std::string& name, const T& value) {
LOG_INFO(Config, "{}: {}", name, value);
}
void LogSettings() {
const auto log_setting = [](std::string_view name, const auto& value) {
LOG_INFO(Config, "{}: {}", name, value);
};
LOG_INFO(Config, "yuzu Configuration:");
LogSetting("Controls_UseDockedMode", Settings::values.use_docked_mode);
LogSetting("System_RngSeed", Settings::values.rng_seed.GetValue().value_or(0));
LogSetting("System_CurrentUser", Settings::values.current_user);
LogSetting("System_LanguageIndex", Settings::values.language_index.GetValue());
LogSetting("System_RegionIndex", Settings::values.region_index.GetValue());
LogSetting("System_TimeZoneIndex", Settings::values.time_zone_index.GetValue());
LogSetting("Core_UseMultiCore", Settings::values.use_multi_core.GetValue());
LogSetting("Renderer_UseResolutionFactor", Settings::values.resolution_factor.GetValue());
LogSetting("Renderer_UseFrameLimit", Settings::values.use_frame_limit.GetValue());
LogSetting("Renderer_FrameLimit", Settings::values.frame_limit.GetValue());
LogSetting("Renderer_UseDiskShaderCache", Settings::values.use_disk_shader_cache.GetValue());
LogSetting("Renderer_GPUAccuracyLevel", Settings::values.gpu_accuracy.GetValue());
LogSetting("Renderer_UseAsynchronousGpuEmulation",
Settings::values.use_asynchronous_gpu_emulation.GetValue());
LogSetting("Renderer_UseVsync", Settings::values.use_vsync.GetValue());
LogSetting("Renderer_UseAssemblyShaders", Settings::values.use_assembly_shaders.GetValue());
LogSetting("Renderer_AnisotropicFilteringLevel", Settings::values.max_anisotropy.GetValue());
LogSetting("Audio_OutputEngine", Settings::values.sink_id);
LogSetting("Audio_EnableAudioStretching", Settings::values.enable_audio_stretching.GetValue());
LogSetting("Audio_OutputDevice", Settings::values.audio_device_id);
LogSetting("DataStorage_UseVirtualSd", Settings::values.use_virtual_sd);
LogSetting("DataStorage_NandDir", FileUtil::GetUserPath(FileUtil::UserPath::NANDDir));
LogSetting("DataStorage_SdmcDir", FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir));
LogSetting("Debugging_UseGdbstub", Settings::values.use_gdbstub);
LogSetting("Debugging_GdbstubPort", Settings::values.gdbstub_port);
LogSetting("Debugging_ProgramArgs", Settings::values.program_args);
LogSetting("Services_BCATBackend", Settings::values.bcat_backend);
LogSetting("Services_BCATBoxcatLocal", Settings::values.bcat_boxcat_local);
log_setting("Controls_UseDockedMode", values.use_docked_mode);
log_setting("System_RngSeed", values.rng_seed.GetValue().value_or(0));
log_setting("System_CurrentUser", values.current_user);
log_setting("System_LanguageIndex", values.language_index.GetValue());
log_setting("System_RegionIndex", values.region_index.GetValue());
log_setting("System_TimeZoneIndex", values.time_zone_index.GetValue());
log_setting("Core_UseMultiCore", values.use_multi_core.GetValue());
log_setting("Renderer_UseResolutionFactor", values.resolution_factor.GetValue());
log_setting("Renderer_UseFrameLimit", values.use_frame_limit.GetValue());
log_setting("Renderer_FrameLimit", values.frame_limit.GetValue());
log_setting("Renderer_UseDiskShaderCache", values.use_disk_shader_cache.GetValue());
log_setting("Renderer_GPUAccuracyLevel", values.gpu_accuracy.GetValue());
log_setting("Renderer_UseAsynchronousGpuEmulation",
values.use_asynchronous_gpu_emulation.GetValue());
log_setting("Renderer_UseVsync", values.use_vsync.GetValue());
log_setting("Renderer_UseAssemblyShaders", values.use_assembly_shaders.GetValue());
log_setting("Renderer_UseAsynchronousShaders", values.use_asynchronous_shaders.GetValue());
log_setting("Renderer_AnisotropicFilteringLevel", values.max_anisotropy.GetValue());
log_setting("Audio_OutputEngine", values.sink_id);
log_setting("Audio_EnableAudioStretching", values.enable_audio_stretching.GetValue());
log_setting("Audio_OutputDevice", values.audio_device_id);
log_setting("DataStorage_UseVirtualSd", values.use_virtual_sd);
log_setting("DataStorage_NandDir", FileUtil::GetUserPath(FileUtil::UserPath::NANDDir));
log_setting("DataStorage_SdmcDir", FileUtil::GetUserPath(FileUtil::UserPath::SDMCDir));
log_setting("Debugging_UseGdbstub", values.use_gdbstub);
log_setting("Debugging_GdbstubPort", values.gdbstub_port);
log_setting("Debugging_ProgramArgs", values.program_args);
log_setting("Services_BCATBackend", values.bcat_backend);
log_setting("Services_BCATBoxcatLocal", values.bcat_boxcat_local);
}
float Volume() {
@@ -169,6 +171,7 @@ void RestoreGlobalState() {
values.use_asynchronous_gpu_emulation.SetGlobal(true);
values.use_vsync.SetGlobal(true);
values.use_assembly_shaders.SetGlobal(true);
values.use_asynchronous_shaders.SetGlobal(true);
values.use_fast_gpu_time.SetGlobal(true);
values.force_30fps_mode.SetGlobal(true);
values.bg_red.SetGlobal(true);

View File

@@ -434,6 +434,7 @@ struct Values {
Setting<bool> use_asynchronous_gpu_emulation;
Setting<bool> use_vsync;
Setting<bool> use_assembly_shaders;
Setting<bool> use_asynchronous_shaders;
Setting<bool> force_30fps_mode;
Setting<bool> use_fast_gpu_time;

View File

@@ -207,6 +207,8 @@ void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader) {
AddField(field_type, "Renderer_UseVsync", Settings::values.use_vsync.GetValue());
AddField(field_type, "Renderer_UseAssemblyShaders",
Settings::values.use_assembly_shaders.GetValue());
AddField(field_type, "Renderer_UseAsynchronousShaders",
Settings::values.use_asynchronous_shaders.GetValue());
AddField(field_type, "System_UseDockedMode", Settings::values.use_docked_mode);
}

View File

@@ -14,7 +14,7 @@
namespace Tools {
namespace {
constexpr s64 MEMORY_FREEZER_TICKS = static_cast<s64>(1000000000 / 60);
constexpr auto memory_freezer_ns = std::chrono::nanoseconds{1000000000 / 60};
u64 MemoryReadWidth(Core::Memory::Memory& memory, u32 width, VAddr addr) {
switch (width) {
@@ -55,10 +55,11 @@ void MemoryWriteWidth(Core::Memory::Memory& memory, u32 width, VAddr addr, u64 v
Freezer::Freezer(Core::Timing::CoreTiming& core_timing_, Core::Memory::Memory& memory_)
: core_timing{core_timing_}, memory{memory_} {
event = Core::Timing::CreateEvent(
"MemoryFreezer::FrameCallback",
[this](u64 userdata, s64 ns_late) { FrameCallback(userdata, ns_late); });
core_timing.ScheduleEvent(MEMORY_FREEZER_TICKS, event);
event = Core::Timing::CreateEvent("MemoryFreezer::FrameCallback",
[this](u64 userdata, std::chrono::nanoseconds ns_late) {
FrameCallback(userdata, ns_late);
});
core_timing.ScheduleEvent(memory_freezer_ns, event);
}
Freezer::~Freezer() {
@@ -68,7 +69,7 @@ Freezer::~Freezer() {
void Freezer::SetActive(bool active) {
if (!this->active.exchange(active)) {
FillEntryReads();
core_timing.ScheduleEvent(MEMORY_FREEZER_TICKS, event);
core_timing.ScheduleEvent(memory_freezer_ns, event);
LOG_DEBUG(Common_Memory, "Memory freezer activated!");
} else {
LOG_DEBUG(Common_Memory, "Memory freezer deactivated!");
@@ -158,7 +159,7 @@ std::vector<Freezer::Entry> Freezer::GetEntries() const {
return entries;
}
void Freezer::FrameCallback(u64 userdata, s64 ns_late) {
void Freezer::FrameCallback(u64, std::chrono::nanoseconds ns_late) {
if (!IsActive()) {
LOG_DEBUG(Common_Memory, "Memory freezer has been deactivated, ending callback events.");
return;
@@ -173,7 +174,7 @@ void Freezer::FrameCallback(u64 userdata, s64 ns_late) {
MemoryWriteWidth(memory, entry.width, entry.address, entry.value);
}
core_timing.ScheduleEvent(MEMORY_FREEZER_TICKS - ns_late, event);
core_timing.ScheduleEvent(memory_freezer_ns - ns_late, event);
}
void Freezer::FillEntryReads() {

View File

@@ -5,6 +5,7 @@
#pragma once
#include <atomic>
#include <chrono>
#include <memory>
#include <mutex>
#include <optional>
@@ -72,7 +73,7 @@ public:
std::vector<Entry> GetEntries() const;
private:
void FrameCallback(u64 userdata, s64 cycles_late);
void FrameCallback(u64 userdata, std::chrono::nanoseconds ns_late);
void FillEntryReads();
std::atomic_bool active{false};

View File

@@ -6,6 +6,7 @@
#include <array>
#include <bitset>
#include <chrono>
#include <cstdlib>
#include <memory>
#include <string>
@@ -17,7 +18,6 @@
namespace {
// Numbers are chosen randomly to make sure the correct one is given.
constexpr std::array<u64, 5> CB_IDS{{42, 144, 93, 1026, UINT64_C(0xFFFF7FFFF7FFFF)}};
constexpr int MAX_SLICE_LENGTH = 10000; // Copied from CoreTiming internals
constexpr std::array<u64, 5> calls_order{{2, 0, 1, 4, 3}};
std::array<s64, 5> delays{};
@@ -25,12 +25,12 @@ std::bitset<CB_IDS.size()> callbacks_ran_flags;
u64 expected_callback = 0;
template <unsigned int IDX>
void HostCallbackTemplate(u64 userdata, s64 nanoseconds_late) {
void HostCallbackTemplate(u64 userdata, std::chrono::nanoseconds ns_late) {
static_assert(IDX < CB_IDS.size(), "IDX out of range");
callbacks_ran_flags.set(IDX);
REQUIRE(CB_IDS[IDX] == userdata);
REQUIRE(CB_IDS[IDX] == CB_IDS[calls_order[expected_callback]]);
delays[IDX] = nanoseconds_late;
delays[IDX] = ns_late.count();
++expected_callback;
}
@@ -77,10 +77,12 @@ TEST_CASE("CoreTiming[BasicOrder]", "[core]") {
core_timing.SyncPause(true);
u64 one_micro = 1000U;
const u64 one_micro = 1000U;
for (std::size_t i = 0; i < events.size(); i++) {
u64 order = calls_order[i];
core_timing.ScheduleEvent(i * one_micro + 100U, events[order], CB_IDS[order]);
const u64 order = calls_order[i];
const auto future_ns = std::chrono::nanoseconds{static_cast<s64>(i * one_micro + 100)};
core_timing.ScheduleEvent(future_ns, events[order], CB_IDS[order]);
}
/// test pause
REQUIRE(callbacks_ran_flags.none());
@@ -116,13 +118,16 @@ TEST_CASE("CoreTiming[BasicOrderNoPausing]", "[core]") {
expected_callback = 0;
u64 start = core_timing.GetGlobalTimeNs().count();
u64 one_micro = 1000U;
const u64 start = core_timing.GetGlobalTimeNs().count();
const u64 one_micro = 1000U;
for (std::size_t i = 0; i < events.size(); i++) {
u64 order = calls_order[i];
core_timing.ScheduleEvent(i * one_micro + 100U, events[order], CB_IDS[order]);
const u64 order = calls_order[i];
const auto future_ns = std::chrono::nanoseconds{static_cast<s64>(i * one_micro + 100)};
core_timing.ScheduleEvent(future_ns, events[order], CB_IDS[order]);
}
u64 end = core_timing.GetGlobalTimeNs().count();
const u64 end = core_timing.GetGlobalTimeNs().count();
const double scheduling_time = static_cast<double>(end - start);
const double timer_time = static_cast<double>(TestTimerSpeed(core_timing));

View File

@@ -98,6 +98,8 @@ add_library(video_core STATIC
sampler_cache.cpp
sampler_cache.h
shader_cache.h
shader_notify.cpp
shader_notify.h
shader/decode/arithmetic.cpp
shader/decode/arithmetic_immediate.cpp
shader/decode/bfe.cpp
@@ -128,6 +130,8 @@ add_library(video_core STATIC
shader/decode/other.cpp
shader/ast.cpp
shader/ast.h
shader/async_shaders.cpp
shader/async_shaders.h
shader/compiler_settings.cpp
shader/compiler_settings.h
shader/control_flow.cpp

View File

@@ -20,6 +20,7 @@
#include "video_core/gpu.h"
#include "video_core/memory_manager.h"
#include "video_core/renderer_base.h"
#include "video_core/shader_notify.h"
#include "video_core/video_core.h"
namespace Tegra {
@@ -36,6 +37,7 @@ GPU::GPU(Core::System& system, std::unique_ptr<VideoCore::RendererBase>&& render
kepler_compute = std::make_unique<Engines::KeplerCompute>(system, rasterizer, *memory_manager);
maxwell_dma = std::make_unique<Engines::MaxwellDMA>(system, *memory_manager);
kepler_memory = std::make_unique<Engines::KeplerMemory>(system, *memory_manager);
shader_notify = std::make_unique<VideoCore::ShaderNotify>();
}
GPU::~GPU() = default;

View File

@@ -33,6 +33,7 @@ class System;
namespace VideoCore {
class RendererBase;
class ShaderNotify;
} // namespace VideoCore
namespace Tegra {
@@ -207,6 +208,14 @@ public:
return *renderer;
}
VideoCore::ShaderNotify& ShaderNotify() {
return *shader_notify;
}
const VideoCore::ShaderNotify& ShaderNotify() const {
return *shader_notify;
}
// Waits for the GPU to finish working
virtual void WaitIdle() const = 0;
@@ -347,6 +356,8 @@ private:
std::unique_ptr<Engines::MaxwellDMA> maxwell_dma;
/// Inline memory engine
std::unique_ptr<Engines::KeplerMemory> kepler_memory;
/// Shader build notifier
std::unique_ptr<VideoCore::ShaderNotify> shader_notify;
std::array<std::atomic<u32>, Service::Nvidia::MaxSyncPoints> syncpoints{};

View File

@@ -103,8 +103,9 @@ public:
virtual ~CachedMacro() = default;
/**
* Executes the macro code with the specified input parameters.
* @param code The macro byte code to execute
*
* @param parameters The parameters of the macro
* @param method The method to execute
*/
virtual void Execute(const std::vector<u32>& parameters, u32 method) = 0;
};

View File

@@ -12,13 +12,11 @@ namespace Tegra {
namespace {
// HLE'd functions
static void HLE_771BB18C62444DA0(Engines::Maxwell3D& maxwell3d,
const std::vector<u32>& parameters) {
void HLE_771BB18C62444DA0(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters) {
const u32 instance_count = parameters[2] & maxwell3d.GetRegisterValue(0xD1B);
maxwell3d.regs.draw.topology.Assign(
static_cast<Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology>(parameters[0] &
~(0x3ffffff << 26)));
static_cast<Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology>(parameters[0] & 0x3ffffff));
maxwell3d.regs.vb_base_instance = parameters[5];
maxwell3d.mme_draw.instance_count = instance_count;
maxwell3d.regs.vb_element_base = parameters[3];
@@ -33,8 +31,7 @@ static void HLE_771BB18C62444DA0(Engines::Maxwell3D& maxwell3d,
maxwell3d.mme_draw.current_mode = Engines::Maxwell3D::MMEDrawMode::Undefined;
}
static void HLE_0D61FC9FAAC9FCAD(Engines::Maxwell3D& maxwell3d,
const std::vector<u32>& parameters) {
void HLE_0D61FC9FAAC9FCAD(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters) {
const u32 count = (maxwell3d.GetRegisterValue(0xD1B) & parameters[2]);
maxwell3d.regs.vertex_buffer.first = parameters[3];
@@ -52,8 +49,7 @@ static void HLE_0D61FC9FAAC9FCAD(Engines::Maxwell3D& maxwell3d,
maxwell3d.mme_draw.current_mode = Engines::Maxwell3D::MMEDrawMode::Undefined;
}
static void HLE_0217920100488FF7(Engines::Maxwell3D& maxwell3d,
const std::vector<u32>& parameters) {
void HLE_0217920100488FF7(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters) {
const u32 instance_count = (maxwell3d.GetRegisterValue(0xD1B) & parameters[2]);
const u32 element_base = parameters[4];
const u32 base_instance = parameters[5];
@@ -81,12 +77,12 @@ static void HLE_0217920100488FF7(Engines::Maxwell3D& maxwell3d,
maxwell3d.CallMethodFromMME(0x8e5, 0x0);
maxwell3d.mme_draw.current_mode = Engines::Maxwell3D::MMEDrawMode::Undefined;
}
} // namespace
} // Anonymous namespace
constexpr std::array<std::pair<u64, HLEFunction>, 3> hle_funcs{{
std::make_pair<u64, HLEFunction>(0x771BB18C62444DA0, &HLE_771BB18C62444DA0),
std::make_pair<u64, HLEFunction>(0x0D61FC9FAAC9FCAD, &HLE_0D61FC9FAAC9FCAD),
std::make_pair<u64, HLEFunction>(0x0217920100488FF7, &HLE_0217920100488FF7),
{0x771BB18C62444DA0, &HLE_771BB18C62444DA0},
{0x0D61FC9FAAC9FCAD, &HLE_0D61FC9FAAC9FCAD},
{0x0217920100488FF7, &HLE_0217920100488FF7},
}};
HLEMacro::HLEMacro(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {}

View File

@@ -233,6 +233,8 @@ Device::Device()
GLAD_GL_NV_gpu_program5 && GLAD_GL_NV_compute_program5 &&
GLAD_GL_NV_transform_feedback && GLAD_GL_NV_transform_feedback2;
use_asynchronous_shaders = Settings::values.use_asynchronous_shaders.GetValue();
LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", has_variable_aoffi);
LOG_INFO(Render_OpenGL, "Renderer_ComponentIndexingBug: {}", has_component_indexing_bug);
LOG_INFO(Render_OpenGL, "Renderer_PreciseBug: {}", has_precise_bug);

View File

@@ -104,6 +104,10 @@ public:
return use_assembly_shaders;
}
bool UseAsynchronousShaders() const {
return use_asynchronous_shaders;
}
private:
static bool TestVariableAoffi();
static bool TestPreciseBug();
@@ -127,6 +131,7 @@ private:
bool has_fast_buffer_sub_data{};
bool has_nv_viewport_array2{};
bool use_assembly_shaders{};
bool use_asynchronous_shaders{};
};
} // namespace OpenGL

View File

@@ -149,7 +149,8 @@ RasterizerOpenGL::RasterizerOpenGL(Core::System& system, Core::Frontend::EmuWind
shader_cache{*this, system, emu_window, device}, query_cache{system, *this},
buffer_cache{*this, system, device, STREAM_BUFFER_SIZE},
fence_manager{system, *this, texture_cache, buffer_cache, query_cache}, system{system},
screen_info{info}, program_manager{program_manager}, state_tracker{state_tracker} {
screen_info{info}, program_manager{program_manager}, state_tracker{state_tracker},
async_shaders{emu_window} {
CheckExtensions();
unified_uniform_buffer.Create();
@@ -162,6 +163,23 @@ RasterizerOpenGL::RasterizerOpenGL(Core::System& system, Core::Frontend::EmuWind
nullptr, 0);
}
}
if (device.UseAsynchronousShaders()) {
// Max worker threads we should allow
constexpr auto MAX_THREADS = 2u;
// Amount of threads we should reserve for other parts of yuzu
constexpr auto RESERVED_THREADS = 6u;
// Get the amount of threads we can use(this can return zero)
const auto cpu_thread_count =
std::max(RESERVED_THREADS, std::thread::hardware_concurrency());
// Deduce how many "extra" threads we have to use.
const auto max_threads_unused = cpu_thread_count - RESERVED_THREADS;
// Always allow at least 1 thread regardless of our settings
const auto max_worker_count = std::max(1u, max_threads_unused);
// Don't use more than MAX_THREADS
const auto worker_count = std::min(max_worker_count, MAX_THREADS);
async_shaders.AllocateWorkers(worker_count);
}
}
RasterizerOpenGL::~RasterizerOpenGL() {
@@ -336,7 +354,7 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) {
continue;
}
Shader* const shader = shader_cache.GetStageProgram(program);
Shader* shader = shader_cache.GetStageProgram(program, async_shaders);
if (device.UseAssemblyShaders()) {
// Check for ARB limitation. We only have 16 SSBOs per context state. To workaround this
@@ -353,7 +371,7 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) {
SetupDrawTextures(stage, shader);
SetupDrawImages(stage, shader);
const GLuint program_handle = shader->GetHandle();
const GLuint program_handle = shader->IsBuilt() ? shader->GetHandle() : 0;
switch (program) {
case Maxwell::ShaderProgram::VertexA:
case Maxwell::ShaderProgram::VertexB:

View File

@@ -33,6 +33,7 @@
#include "video_core/renderer_opengl/gl_state_tracker.h"
#include "video_core/renderer_opengl/gl_texture_cache.h"
#include "video_core/renderer_opengl/utils.h"
#include "video_core/shader/async_shaders.h"
#include "video_core/textures/texture.h"
namespace Core {
@@ -91,6 +92,14 @@ public:
return num_queued_commands > 0;
}
VideoCommon::Shader::AsyncShaders& GetAsyncShaders() {
return async_shaders;
}
const VideoCommon::Shader::AsyncShaders& GetAsyncShaders() const {
return async_shaders;
}
private:
/// Configures the color and depth framebuffer states.
void ConfigureFramebuffers();
@@ -242,6 +251,7 @@ private:
ScreenInfo& screen_info;
ProgramManager& program_manager;
StateTracker& state_tracker;
VideoCommon::Shader::AsyncShaders async_shaders;
static constexpr std::size_t STREAM_BUFFER_SIZE = 128 * 1024 * 1024;

View File

@@ -177,6 +177,12 @@ public:
Release();
}
OGLAssemblyProgram& operator=(OGLAssemblyProgram&& o) noexcept {
Release();
handle = std::exchange(o.handle, 0);
return *this;
}
/// Deletes the internal OpenGL resource
void Release();

View File

@@ -31,6 +31,7 @@
#include "video_core/shader/registry.h"
#include "video_core/shader/shader_ir.h"
#include "video_core/shader_cache.h"
#include "video_core/shader_notify.h"
namespace OpenGL {
@@ -140,9 +141,24 @@ std::shared_ptr<Registry> MakeRegistry(const ShaderDiskCacheEntry& entry) {
return registry;
}
std::unordered_set<GLenum> GetSupportedFormats() {
GLint num_formats;
glGetIntegerv(GL_NUM_PROGRAM_BINARY_FORMATS, &num_formats);
std::vector<GLint> formats(num_formats);
glGetIntegerv(GL_PROGRAM_BINARY_FORMATS, formats.data());
std::unordered_set<GLenum> supported_formats;
for (const GLint format : formats) {
supported_formats.insert(static_cast<GLenum>(format));
}
return supported_formats;
}
} // Anonymous namespace
ProgramSharedPtr BuildShader(const Device& device, ShaderType shader_type, u64 unique_identifier,
const ShaderIR& ir, const Registry& registry,
bool hint_retrievable = false) {
const ShaderIR& ir, const Registry& registry, bool hint_retrievable) {
const std::string shader_id = MakeShaderID(unique_identifier, shader_type);
LOG_INFO(Render_OpenGL, "{}", shader_id);
@@ -181,30 +197,17 @@ ProgramSharedPtr BuildShader(const Device& device, ShaderType shader_type, u64 u
return program;
}
std::unordered_set<GLenum> GetSupportedFormats() {
GLint num_formats;
glGetIntegerv(GL_NUM_PROGRAM_BINARY_FORMATS, &num_formats);
std::vector<GLint> formats(num_formats);
glGetIntegerv(GL_PROGRAM_BINARY_FORMATS, formats.data());
std::unordered_set<GLenum> supported_formats;
for (const GLint format : formats) {
supported_formats.insert(static_cast<GLenum>(format));
}
return supported_formats;
}
} // Anonymous namespace
Shader::Shader(std::shared_ptr<VideoCommon::Shader::Registry> registry_, ShaderEntries entries_,
ProgramSharedPtr program_)
: registry{std::move(registry_)}, entries{std::move(entries_)}, program{std::move(program_)} {
ProgramSharedPtr program_, bool is_built)
: registry{std::move(registry_)}, entries{std::move(entries_)}, program{std::move(program_)},
is_built(is_built) {
handle = program->assembly_program.handle;
if (handle == 0) {
handle = program->source_program.handle;
}
ASSERT(handle != 0);
if (is_built) {
ASSERT(handle != 0);
}
}
Shader::~Shader() = default;
@@ -214,42 +217,82 @@ GLuint Shader::GetHandle() const {
return handle;
}
std::unique_ptr<Shader> Shader::CreateStageFromMemory(const ShaderParameters& params,
Maxwell::ShaderProgram program_type,
ProgramCode code, ProgramCode code_b) {
bool Shader::IsBuilt() const {
return is_built;
}
void Shader::AsyncOpenGLBuilt(OGLProgram new_program) {
program->source_program = std::move(new_program);
handle = program->source_program.handle;
is_built = true;
}
void Shader::AsyncGLASMBuilt(OGLAssemblyProgram new_program) {
program->assembly_program = std::move(new_program);
handle = program->assembly_program.handle;
is_built = true;
}
std::unique_ptr<Shader> Shader::CreateStageFromMemory(
const ShaderParameters& params, Maxwell::ShaderProgram program_type, ProgramCode code,
ProgramCode code_b, VideoCommon::Shader::AsyncShaders& async_shaders, VAddr cpu_addr) {
const auto shader_type = GetShaderType(program_type);
const std::size_t size_in_bytes = code.size() * sizeof(u64);
auto registry = std::make_shared<Registry>(shader_type, params.system.GPU().Maxwell3D());
const ShaderIR ir(code, STAGE_MAIN_OFFSET, COMPILER_SETTINGS, *registry);
// TODO(Rodrigo): Handle VertexA shaders
// std::optional<ShaderIR> ir_b;
// if (!code_b.empty()) {
// ir_b.emplace(code_b, STAGE_MAIN_OFFSET);
// }
auto program = BuildShader(params.device, shader_type, params.unique_identifier, ir, *registry);
auto& gpu = params.system.GPU();
gpu.ShaderNotify().MarkSharderBuilding();
ShaderDiskCacheEntry entry;
entry.type = shader_type;
entry.code = std::move(code);
entry.code_b = std::move(code_b);
entry.unique_identifier = params.unique_identifier;
entry.bound_buffer = registry->GetBoundBuffer();
entry.graphics_info = registry->GetGraphicsInfo();
entry.keys = registry->GetKeys();
entry.bound_samplers = registry->GetBoundSamplers();
entry.bindless_samplers = registry->GetBindlessSamplers();
params.disk_cache.SaveEntry(std::move(entry));
auto registry = std::make_shared<Registry>(shader_type, gpu.Maxwell3D());
if (!async_shaders.IsShaderAsync(params.system.GPU()) ||
!params.device.UseAsynchronousShaders()) {
const ShaderIR ir(code, STAGE_MAIN_OFFSET, COMPILER_SETTINGS, *registry);
// TODO(Rodrigo): Handle VertexA shaders
// std::optional<ShaderIR> ir_b;
// if (!code_b.empty()) {
// ir_b.emplace(code_b, STAGE_MAIN_OFFSET);
// }
auto program =
BuildShader(params.device, shader_type, params.unique_identifier, ir, *registry);
ShaderDiskCacheEntry entry;
entry.type = shader_type;
entry.code = std::move(code);
entry.code_b = std::move(code_b);
entry.unique_identifier = params.unique_identifier;
entry.bound_buffer = registry->GetBoundBuffer();
entry.graphics_info = registry->GetGraphicsInfo();
entry.keys = registry->GetKeys();
entry.bound_samplers = registry->GetBoundSamplers();
entry.bindless_samplers = registry->GetBindlessSamplers();
params.disk_cache.SaveEntry(std::move(entry));
return std::unique_ptr<Shader>(new Shader(
std::move(registry), MakeEntries(params.device, ir, shader_type), std::move(program)));
gpu.ShaderNotify().MarkShaderComplete();
return std::unique_ptr<Shader>(new Shader(std::move(registry),
MakeEntries(params.device, ir, shader_type),
std::move(program), true));
} else {
// Required for entries
const ShaderIR ir(code, STAGE_MAIN_OFFSET, COMPILER_SETTINGS, *registry);
auto entries = MakeEntries(params.device, ir, shader_type);
async_shaders.QueueOpenGLShader(params.device, shader_type, params.unique_identifier,
std::move(code), std::move(code_b), STAGE_MAIN_OFFSET,
COMPILER_SETTINGS, *registry, cpu_addr);
auto program = std::make_shared<ProgramHandle>();
return std::unique_ptr<Shader>(
new Shader(std::move(registry), std::move(entries), std::move(program), false));
}
}
std::unique_ptr<Shader> Shader::CreateKernelFromMemory(const ShaderParameters& params,
ProgramCode code) {
const std::size_t size_in_bytes = code.size() * sizeof(u64);
auto& engine = params.system.GPU().KeplerCompute();
auto& gpu = params.system.GPU();
gpu.ShaderNotify().MarkSharderBuilding();
auto& engine = gpu.KeplerCompute();
auto registry = std::make_shared<Registry>(ShaderType::Compute, engine);
const ShaderIR ir(code, KERNEL_MAIN_OFFSET, COMPILER_SETTINGS, *registry);
const u64 uid = params.unique_identifier;
@@ -266,6 +309,8 @@ std::unique_ptr<Shader> Shader::CreateKernelFromMemory(const ShaderParameters& p
entry.bindless_samplers = registry->GetBindlessSamplers();
params.disk_cache.SaveEntry(std::move(entry));
gpu.ShaderNotify().MarkShaderComplete();
return std::unique_ptr<Shader>(new Shader(std::move(registry),
MakeEntries(params.device, ir, ShaderType::Compute),
std::move(program)));
@@ -436,14 +481,51 @@ ProgramSharedPtr ShaderCacheOpenGL::GeneratePrecompiledProgram(
return program;
}
Shader* ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program) {
Shader* ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program,
VideoCommon::Shader::AsyncShaders& async_shaders) {
if (!system.GPU().Maxwell3D().dirty.flags[Dirty::Shaders]) {
return last_shaders[static_cast<std::size_t>(program)];
auto* last_shader = last_shaders[static_cast<std::size_t>(program)];
if (last_shader->IsBuilt()) {
return last_shader;
}
}
auto& memory_manager{system.GPU().MemoryManager()};
const GPUVAddr address{GetShaderAddress(system, program)};
if (device.UseAsynchronousShaders() && async_shaders.HasCompletedWork()) {
auto completed_work = async_shaders.GetCompletedWork();
for (auto& work : completed_work) {
Shader* shader = TryGet(work.cpu_address);
auto& gpu = system.GPU();
gpu.ShaderNotify().MarkShaderComplete();
if (shader == nullptr) {
continue;
}
using namespace VideoCommon::Shader;
if (work.backend == AsyncShaders::Backend::OpenGL) {
shader->AsyncOpenGLBuilt(std::move(work.program.opengl));
} else if (work.backend == AsyncShaders::Backend::GLASM) {
shader->AsyncGLASMBuilt(std::move(work.program.glasm));
}
ShaderDiskCacheEntry entry;
entry.type = work.shader_type;
entry.code = std::move(work.code);
entry.code_b = std::move(work.code_b);
entry.unique_identifier = work.uid;
auto& registry = shader->GetRegistry();
entry.bound_buffer = registry.GetBoundBuffer();
entry.graphics_info = registry.GetGraphicsInfo();
entry.keys = registry.GetKeys();
entry.bound_samplers = registry.GetBoundSamplers();
entry.bindless_samplers = registry.GetBindlessSamplers();
disk_cache.SaveEntry(std::move(entry));
}
}
// Look up shader in the cache based on address
const auto cpu_addr{memory_manager.GpuToCpuAddress(address)};
if (Shader* const shader{cpu_addr ? TryGet(*cpu_addr) : null_shader.get()}) {
@@ -471,7 +553,8 @@ Shader* ShaderCacheOpenGL::GetStageProgram(Maxwell::ShaderProgram program) {
std::unique_ptr<Shader> shader;
const auto found = runtime_cache.find(unique_identifier);
if (found == runtime_cache.end()) {
shader = Shader::CreateStageFromMemory(params, program, std::move(code), std::move(code_b));
shader = Shader::CreateStageFromMemory(params, program, std::move(code), std::move(code_b),
async_shaders, cpu_addr.value_or(0));
} else {
shader = Shader::CreateFromCache(params, found->second);
}

View File

@@ -33,6 +33,10 @@ namespace Core::Frontend {
class EmuWindow;
}
namespace VideoCommon::Shader {
class AsyncShaders;
}
namespace OpenGL {
class Device;
@@ -61,6 +65,11 @@ struct ShaderParameters {
u64 unique_identifier;
};
ProgramSharedPtr BuildShader(const Device& device, Tegra::Engines::ShaderType shader_type,
u64 unique_identifier, const VideoCommon::Shader::ShaderIR& ir,
const VideoCommon::Shader::Registry& registry,
bool hint_retrievable = false);
class Shader final {
public:
~Shader();
@@ -68,15 +77,28 @@ public:
/// Gets the GL program handle for the shader
GLuint GetHandle() const;
bool IsBuilt() const;
/// Gets the shader entries for the shader
const ShaderEntries& GetEntries() const {
return entries;
}
static std::unique_ptr<Shader> CreateStageFromMemory(const ShaderParameters& params,
Maxwell::ShaderProgram program_type,
ProgramCode program_code,
ProgramCode program_code_b);
const VideoCommon::Shader::Registry& GetRegistry() const {
return *registry;
}
/// Mark a OpenGL shader as built
void AsyncOpenGLBuilt(OGLProgram new_program);
/// Mark a GLASM shader as built
void AsyncGLASMBuilt(OGLAssemblyProgram new_program);
static std::unique_ptr<Shader> CreateStageFromMemory(
const ShaderParameters& params, Maxwell::ShaderProgram program_type,
ProgramCode program_code, ProgramCode program_code_b,
VideoCommon::Shader::AsyncShaders& async_shaders, VAddr cpu_addr);
static std::unique_ptr<Shader> CreateKernelFromMemory(const ShaderParameters& params,
ProgramCode code);
@@ -85,12 +107,13 @@ public:
private:
explicit Shader(std::shared_ptr<VideoCommon::Shader::Registry> registry, ShaderEntries entries,
ProgramSharedPtr program);
ProgramSharedPtr program, bool is_built = true);
std::shared_ptr<VideoCommon::Shader::Registry> registry;
ShaderEntries entries;
ProgramSharedPtr program;
GLuint handle = 0;
bool is_built{};
};
class ShaderCacheOpenGL final : public VideoCommon::ShaderCache<Shader> {
@@ -104,7 +127,8 @@ public:
const VideoCore::DiskResourceLoadCallback& callback);
/// Gets the current specified shader stage program
Shader* GetStageProgram(Maxwell::ShaderProgram program);
Shader* GetStageProgram(Maxwell::ShaderProgram program,
VideoCommon::Shader::AsyncShaders& async_shaders);
/// Gets a compute kernel in the passed address
Shader* GetComputeKernel(GPUVAddr code_addr);

View File

@@ -39,16 +39,17 @@ std::unique_ptr<VKStreamBuffer> CreateStreamBuffer(const VKDevice& device, VKSch
Buffer::Buffer(const VKDevice& device, VKMemoryManager& memory_manager, VKScheduler& scheduler_,
VKStagingBufferPool& staging_pool_, VAddr cpu_addr, std::size_t size)
: VideoCommon::BufferBlock{cpu_addr, size}, scheduler{scheduler_}, staging_pool{staging_pool_} {
VkBufferCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.size = static_cast<VkDeviceSize>(size);
ci.usage = BUFFER_USAGE | VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT;
ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE;
ci.queueFamilyIndexCount = 0;
ci.pQueueFamilyIndices = nullptr;
: BufferBlock{cpu_addr, size}, scheduler{scheduler_}, staging_pool{staging_pool_} {
const VkBufferCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.size = static_cast<VkDeviceSize>(size),
.usage = BUFFER_USAGE | VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT,
.sharingMode = VK_SHARING_MODE_EXCLUSIVE,
.queueFamilyIndexCount = 0,
.pQueueFamilyIndices = nullptr,
};
buffer.handle = device.GetLogical().CreateBuffer(ci);
buffer.commit = memory_manager.Commit(buffer.handle, false);
@@ -66,16 +67,17 @@ void Buffer::Upload(std::size_t offset, std::size_t size, const u8* data) {
scheduler.Record([staging = *staging.handle, handle, offset, size](vk::CommandBuffer cmdbuf) {
cmdbuf.CopyBuffer(staging, handle, VkBufferCopy{0, offset, size});
VkBufferMemoryBarrier barrier;
barrier.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER;
barrier.pNext = nullptr;
barrier.srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT;
barrier.dstAccessMask = UPLOAD_ACCESS_BARRIERS;
barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED;
barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED;
barrier.buffer = handle;
barrier.offset = offset;
barrier.size = size;
const VkBufferMemoryBarrier barrier{
.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER,
.pNext = nullptr,
.srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT,
.dstAccessMask = UPLOAD_ACCESS_BARRIERS,
.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
.buffer = handle,
.offset = offset,
.size = size,
};
cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, UPLOAD_PIPELINE_STAGE, 0, {},
barrier, {});
});
@@ -87,16 +89,17 @@ void Buffer::Download(std::size_t offset, std::size_t size, u8* data) {
const VkBuffer handle = Handle();
scheduler.Record([staging = *staging.handle, handle, offset, size](vk::CommandBuffer cmdbuf) {
VkBufferMemoryBarrier barrier;
barrier.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER;
barrier.pNext = nullptr;
barrier.srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT;
barrier.dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT;
barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED;
barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED;
barrier.buffer = handle;
barrier.offset = offset;
barrier.size = size;
const VkBufferMemoryBarrier barrier{
.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER,
.pNext = nullptr,
.srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT,
.dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT,
.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
.buffer = handle,
.offset = offset,
.size = size,
};
cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_VERTEX_SHADER_BIT |
VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT |

View File

@@ -115,32 +115,32 @@ constexpr u8 quad_array[] = {
0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00};
VkDescriptorSetLayoutBinding BuildQuadArrayPassDescriptorSetLayoutBinding() {
VkDescriptorSetLayoutBinding binding;
binding.binding = 0;
binding.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER;
binding.descriptorCount = 1;
binding.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT;
binding.pImmutableSamplers = nullptr;
return binding;
return {
.binding = 0,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.descriptorCount = 1,
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.pImmutableSamplers = nullptr,
};
}
VkDescriptorUpdateTemplateEntryKHR BuildQuadArrayPassDescriptorUpdateTemplateEntry() {
VkDescriptorUpdateTemplateEntryKHR entry;
entry.dstBinding = 0;
entry.dstArrayElement = 0;
entry.descriptorCount = 1;
entry.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER;
entry.offset = 0;
entry.stride = sizeof(DescriptorUpdateEntry);
return entry;
return {
.dstBinding = 0,
.dstArrayElement = 0,
.descriptorCount = 1,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.offset = 0,
.stride = sizeof(DescriptorUpdateEntry),
};
}
VkPushConstantRange BuildComputePushConstantRange(std::size_t size) {
VkPushConstantRange range;
range.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT;
range.offset = 0;
range.size = static_cast<u32>(size);
return range;
return {
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.offset = 0,
.size = static_cast<u32>(size),
};
}
// Uint8 SPIR-V module. Generated from the "shaders/" directory.
@@ -344,29 +344,33 @@ constexpr u8 QUAD_INDEXED_SPV[] = {
0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00};
std::array<VkDescriptorSetLayoutBinding, 2> BuildInputOutputDescriptorSetBindings() {
std::array<VkDescriptorSetLayoutBinding, 2> bindings;
bindings[0].binding = 0;
bindings[0].descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER;
bindings[0].descriptorCount = 1;
bindings[0].stageFlags = VK_SHADER_STAGE_COMPUTE_BIT;
bindings[0].pImmutableSamplers = nullptr;
bindings[1].binding = 1;
bindings[1].descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER;
bindings[1].descriptorCount = 1;
bindings[1].stageFlags = VK_SHADER_STAGE_COMPUTE_BIT;
bindings[1].pImmutableSamplers = nullptr;
return bindings;
return {{
{
.binding = 0,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.descriptorCount = 1,
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.pImmutableSamplers = nullptr,
},
{
.binding = 1,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.descriptorCount = 1,
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.pImmutableSamplers = nullptr,
},
}};
}
VkDescriptorUpdateTemplateEntryKHR BuildInputOutputDescriptorUpdateTemplate() {
VkDescriptorUpdateTemplateEntryKHR entry;
entry.dstBinding = 0;
entry.dstArrayElement = 0;
entry.descriptorCount = 2;
entry.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER;
entry.offset = 0;
entry.stride = sizeof(DescriptorUpdateEntry);
return entry;
return {
.dstBinding = 0,
.dstArrayElement = 0,
.descriptorCount = 2,
.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
.offset = 0,
.stride = sizeof(DescriptorUpdateEntry),
};
}
} // Anonymous namespace
@@ -376,37 +380,37 @@ VKComputePass::VKComputePass(const VKDevice& device, VKDescriptorPool& descripto
vk::Span<VkDescriptorUpdateTemplateEntryKHR> templates,
vk::Span<VkPushConstantRange> push_constants, std::size_t code_size,
const u8* code) {
VkDescriptorSetLayoutCreateInfo descriptor_layout_ci;
descriptor_layout_ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO;
descriptor_layout_ci.pNext = nullptr;
descriptor_layout_ci.flags = 0;
descriptor_layout_ci.bindingCount = bindings.size();
descriptor_layout_ci.pBindings = bindings.data();
descriptor_set_layout = device.GetLogical().CreateDescriptorSetLayout(descriptor_layout_ci);
descriptor_set_layout = device.GetLogical().CreateDescriptorSetLayout({
.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.bindingCount = bindings.size(),
.pBindings = bindings.data(),
});
VkPipelineLayoutCreateInfo pipeline_layout_ci;
pipeline_layout_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO;
pipeline_layout_ci.pNext = nullptr;
pipeline_layout_ci.flags = 0;
pipeline_layout_ci.setLayoutCount = 1;
pipeline_layout_ci.pSetLayouts = descriptor_set_layout.address();
pipeline_layout_ci.pushConstantRangeCount = push_constants.size();
pipeline_layout_ci.pPushConstantRanges = push_constants.data();
layout = device.GetLogical().CreatePipelineLayout(pipeline_layout_ci);
layout = device.GetLogical().CreatePipelineLayout({
.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.setLayoutCount = 1,
.pSetLayouts = descriptor_set_layout.address(),
.pushConstantRangeCount = push_constants.size(),
.pPushConstantRanges = push_constants.data(),
});
if (!templates.empty()) {
VkDescriptorUpdateTemplateCreateInfoKHR template_ci;
template_ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR;
template_ci.pNext = nullptr;
template_ci.flags = 0;
template_ci.descriptorUpdateEntryCount = templates.size();
template_ci.pDescriptorUpdateEntries = templates.data();
template_ci.templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR;
template_ci.descriptorSetLayout = *descriptor_set_layout;
template_ci.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS;
template_ci.pipelineLayout = *layout;
template_ci.set = 0;
descriptor_template = device.GetLogical().CreateDescriptorUpdateTemplateKHR(template_ci);
descriptor_template = device.GetLogical().CreateDescriptorUpdateTemplateKHR({
.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR,
.pNext = nullptr,
.flags = 0,
.descriptorUpdateEntryCount = templates.size(),
.pDescriptorUpdateEntries = templates.data(),
.templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR,
.descriptorSetLayout = *descriptor_set_layout,
.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS,
.pipelineLayout = *layout,
.set = 0,
});
descriptor_allocator.emplace(descriptor_pool, *descriptor_set_layout);
}
@@ -414,32 +418,32 @@ VKComputePass::VKComputePass(const VKDevice& device, VKDescriptorPool& descripto
auto code_copy = std::make_unique<u32[]>(code_size / sizeof(u32) + 1);
std::memcpy(code_copy.get(), code, code_size);
VkShaderModuleCreateInfo module_ci;
module_ci.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO;
module_ci.pNext = nullptr;
module_ci.flags = 0;
module_ci.codeSize = code_size;
module_ci.pCode = code_copy.get();
module = device.GetLogical().CreateShaderModule(module_ci);
module = device.GetLogical().CreateShaderModule({
.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.codeSize = code_size,
.pCode = code_copy.get(),
});
VkComputePipelineCreateInfo pipeline_ci;
pipeline_ci.sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO;
pipeline_ci.pNext = nullptr;
pipeline_ci.flags = 0;
pipeline_ci.layout = *layout;
pipeline_ci.basePipelineHandle = nullptr;
pipeline_ci.basePipelineIndex = 0;
VkPipelineShaderStageCreateInfo& stage_ci = pipeline_ci.stage;
stage_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
stage_ci.pNext = nullptr;
stage_ci.flags = 0;
stage_ci.stage = VK_SHADER_STAGE_COMPUTE_BIT;
stage_ci.module = *module;
stage_ci.pName = "main";
stage_ci.pSpecializationInfo = nullptr;
pipeline = device.GetLogical().CreateComputePipeline(pipeline_ci);
pipeline = device.GetLogical().CreateComputePipeline({
.sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.stage =
{
.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.stage = VK_SHADER_STAGE_COMPUTE_BIT,
.module = *module,
.pName = "main",
.pSpecializationInfo = nullptr,
},
.layout = *layout,
.basePipelineHandle = nullptr,
.basePipelineIndex = 0,
});
}
VKComputePass::~VKComputePass() = default;

View File

@@ -43,12 +43,13 @@ vk::DescriptorSetLayout VKComputePipeline::CreateDescriptorSetLayout() const {
const auto add_bindings = [&](VkDescriptorType descriptor_type, std::size_t num_entries) {
// TODO(Rodrigo): Maybe make individual bindings here?
for (u32 bindpoint = 0; bindpoint < static_cast<u32>(num_entries); ++bindpoint) {
VkDescriptorSetLayoutBinding& entry = bindings.emplace_back();
entry.binding = binding++;
entry.descriptorType = descriptor_type;
entry.descriptorCount = 1;
entry.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT;
entry.pImmutableSamplers = nullptr;
bindings.push_back({
.binding = binding++,
.descriptorType = descriptor_type,
.descriptorCount = 1,
.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
.pImmutableSamplers = nullptr,
});
}
};
add_bindings(VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER, entries.const_buffers.size());
@@ -58,25 +59,25 @@ vk::DescriptorSetLayout VKComputePipeline::CreateDescriptorSetLayout() const {
add_bindings(VK_DESCRIPTOR_TYPE_STORAGE_TEXEL_BUFFER, entries.storage_texels.size());
add_bindings(VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, entries.images.size());
VkDescriptorSetLayoutCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.bindingCount = static_cast<u32>(bindings.size());
ci.pBindings = bindings.data();
return device.GetLogical().CreateDescriptorSetLayout(ci);
return device.GetLogical().CreateDescriptorSetLayout({
.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.bindingCount = static_cast<u32>(bindings.size()),
.pBindings = bindings.data(),
});
}
vk::PipelineLayout VKComputePipeline::CreatePipelineLayout() const {
VkPipelineLayoutCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.setLayoutCount = 1;
ci.pSetLayouts = descriptor_set_layout.address();
ci.pushConstantRangeCount = 0;
ci.pPushConstantRanges = nullptr;
return device.GetLogical().CreatePipelineLayout(ci);
return device.GetLogical().CreatePipelineLayout({
.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.setLayoutCount = 1,
.pSetLayouts = descriptor_set_layout.address(),
.pushConstantRangeCount = 0,
.pPushConstantRanges = nullptr,
});
}
vk::DescriptorUpdateTemplateKHR VKComputePipeline::CreateDescriptorUpdateTemplate() const {
@@ -89,59 +90,63 @@ vk::DescriptorUpdateTemplateKHR VKComputePipeline::CreateDescriptorUpdateTemplat
return {};
}
VkDescriptorUpdateTemplateCreateInfoKHR ci;
ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR;
ci.pNext = nullptr;
ci.flags = 0;
ci.descriptorUpdateEntryCount = static_cast<u32>(template_entries.size());
ci.pDescriptorUpdateEntries = template_entries.data();
ci.templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR;
ci.descriptorSetLayout = *descriptor_set_layout;
ci.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS;
ci.pipelineLayout = *layout;
ci.set = DESCRIPTOR_SET;
return device.GetLogical().CreateDescriptorUpdateTemplateKHR(ci);
return device.GetLogical().CreateDescriptorUpdateTemplateKHR({
.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR,
.pNext = nullptr,
.flags = 0,
.descriptorUpdateEntryCount = static_cast<u32>(template_entries.size()),
.pDescriptorUpdateEntries = template_entries.data(),
.templateType = VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET_KHR,
.descriptorSetLayout = *descriptor_set_layout,
.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS,
.pipelineLayout = *layout,
.set = DESCRIPTOR_SET,
});
}
vk::ShaderModule VKComputePipeline::CreateShaderModule(const std::vector<u32>& code) const {
device.SaveShader(code);
VkShaderModuleCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.codeSize = code.size() * sizeof(u32);
ci.pCode = code.data();
return device.GetLogical().CreateShaderModule(ci);
return device.GetLogical().CreateShaderModule({
.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.codeSize = code.size() * sizeof(u32),
.pCode = code.data(),
});
}
vk::Pipeline VKComputePipeline::CreatePipeline() const {
VkComputePipelineCreateInfo ci;
VkPipelineShaderStageCreateInfo& stage_ci = ci.stage;
stage_ci.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
stage_ci.pNext = nullptr;
stage_ci.flags = 0;
stage_ci.stage = VK_SHADER_STAGE_COMPUTE_BIT;
stage_ci.module = *shader_module;
stage_ci.pName = "main";
stage_ci.pSpecializationInfo = nullptr;
VkPipelineShaderStageRequiredSubgroupSizeCreateInfoEXT subgroup_size_ci;
subgroup_size_ci.sType =
VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_REQUIRED_SUBGROUP_SIZE_CREATE_INFO_EXT;
subgroup_size_ci.pNext = nullptr;
subgroup_size_ci.requiredSubgroupSize = GuestWarpSize;
VkComputePipelineCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.stage =
{
.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.stage = VK_SHADER_STAGE_COMPUTE_BIT,
.module = *shader_module,
.pName = "main",
.pSpecializationInfo = nullptr,
},
.layout = *layout,
.basePipelineHandle = nullptr,
.basePipelineIndex = 0,
};
const VkPipelineShaderStageRequiredSubgroupSizeCreateInfoEXT subgroup_size_ci{
.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_REQUIRED_SUBGROUP_SIZE_CREATE_INFO_EXT,
.pNext = nullptr,
.requiredSubgroupSize = GuestWarpSize,
};
if (entries.uses_warps && device.IsGuestWarpSizeSupported(VK_SHADER_STAGE_COMPUTE_BIT)) {
stage_ci.pNext = &subgroup_size_ci;
ci.stage.pNext = &subgroup_size_ci;
}
ci.sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.layout = *layout;
ci.basePipelineHandle = nullptr;
ci.basePipelineIndex = 0;
return device.GetLogical().CreateComputePipeline(ci);
}

View File

@@ -43,27 +43,30 @@ vk::DescriptorPool* VKDescriptorPool::AllocateNewPool() {
{VK_DESCRIPTOR_TYPE_UNIFORM_TEXEL_BUFFER, num_sets * 64},
{VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, num_sets * 64},
{VK_DESCRIPTOR_TYPE_STORAGE_TEXEL_BUFFER, num_sets * 64},
{VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, num_sets * 40}};
{VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, num_sets * 40},
};
VkDescriptorPoolCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = VK_DESCRIPTOR_POOL_CREATE_FREE_DESCRIPTOR_SET_BIT;
ci.maxSets = num_sets;
ci.poolSizeCount = static_cast<u32>(std::size(pool_sizes));
ci.pPoolSizes = std::data(pool_sizes);
const VkDescriptorPoolCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO,
.pNext = nullptr,
.flags = VK_DESCRIPTOR_POOL_CREATE_FREE_DESCRIPTOR_SET_BIT,
.maxSets = num_sets,
.poolSizeCount = static_cast<u32>(std::size(pool_sizes)),
.pPoolSizes = std::data(pool_sizes),
};
return &pools.emplace_back(device.GetLogical().CreateDescriptorPool(ci));
}
vk::DescriptorSets VKDescriptorPool::AllocateDescriptors(VkDescriptorSetLayout layout,
std::size_t count) {
const std::vector layout_copies(count, layout);
VkDescriptorSetAllocateInfo ai;
ai.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO;
ai.pNext = nullptr;
ai.descriptorPool = **active_pool;
ai.descriptorSetCount = static_cast<u32>(count);
ai.pSetLayouts = layout_copies.data();
VkDescriptorSetAllocateInfo ai{
.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO,
.pNext = nullptr,
.descriptorPool = **active_pool,
.descriptorSetCount = static_cast<u32>(count),
.pSetLayouts = layout_copies.data(),
};
vk::DescriptorSets sets = active_pool->Allocate(ai);
if (!sets.IsOutOfPoolMemory()) {

View File

@@ -757,14 +757,14 @@ std::vector<VkDeviceQueueCreateInfo> VKDevice::GetDeviceQueueCreateInfos() const
queue_cis.reserve(unique_queue_families.size());
for (const u32 queue_family : unique_queue_families) {
queue_cis.push_back({
auto& ci = queue_cis.emplace_back(VkDeviceQueueCreateInfo{
.sType = VK_STRUCTURE_TYPE_DEVICE_QUEUE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.queueFamilyIndex = queue_family,
.queueCount = 1,
.pQueuePriorities = &QUEUE_PRIORITY,
});
ci.queueCount = 1;
ci.pQueuePriorities = &QUEUE_PRIORITY;
}
return queue_cis;

View File

@@ -102,21 +102,29 @@ bool VKImage::HasChanged(u32 base_layer, u32 num_layers, u32 base_level, u32 num
void VKImage::CreatePresentView() {
// Image type has to be 2D to be presented.
VkImageViewCreateInfo image_view_ci;
image_view_ci.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO;
image_view_ci.pNext = nullptr;
image_view_ci.flags = 0;
image_view_ci.image = *image;
image_view_ci.viewType = VK_IMAGE_VIEW_TYPE_2D;
image_view_ci.format = format;
image_view_ci.components = {VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY,
VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY};
image_view_ci.subresourceRange.aspectMask = aspect_mask;
image_view_ci.subresourceRange.baseMipLevel = 0;
image_view_ci.subresourceRange.levelCount = 1;
image_view_ci.subresourceRange.baseArrayLayer = 0;
image_view_ci.subresourceRange.layerCount = 1;
present_view = device.GetLogical().CreateImageView(image_view_ci);
present_view = device.GetLogical().CreateImageView({
.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.image = *image,
.viewType = VK_IMAGE_VIEW_TYPE_2D,
.format = format,
.components =
{
.r = VK_COMPONENT_SWIZZLE_IDENTITY,
.g = VK_COMPONENT_SWIZZLE_IDENTITY,
.b = VK_COMPONENT_SWIZZLE_IDENTITY,
.a = VK_COMPONENT_SWIZZLE_IDENTITY,
},
.subresourceRange =
{
.aspectMask = aspect_mask,
.baseMipLevel = 0,
.levelCount = 1,
.baseArrayLayer = 0,
.layerCount = 1,
},
});
}
VKImage::SubrangeState& VKImage::GetSubrangeState(u32 layer, u32 level) noexcept {

View File

@@ -178,13 +178,12 @@ bool VKMemoryManager::AllocMemory(VkMemoryPropertyFlags wanted_properties, u32 t
}();
// Try to allocate found type.
VkMemoryAllocateInfo memory_ai;
memory_ai.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO;
memory_ai.pNext = nullptr;
memory_ai.allocationSize = size;
memory_ai.memoryTypeIndex = type;
vk::DeviceMemory memory = device.GetLogical().TryAllocateMemory(memory_ai);
vk::DeviceMemory memory = device.GetLogical().TryAllocateMemory({
.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO,
.pNext = nullptr,
.allocationSize = size,
.memoryTypeIndex = type,
});
if (!memory) {
LOG_CRITICAL(Render_Vulkan, "Device allocation failed!");
return false;

View File

@@ -88,12 +88,13 @@ void AddBindings(std::vector<VkDescriptorSetLayoutBinding>& bindings, u32& bindi
// Combined image samplers can be arrayed.
count = container[i].size;
}
VkDescriptorSetLayoutBinding& entry = bindings.emplace_back();
entry.binding = binding++;
entry.descriptorType = descriptor_type;
entry.descriptorCount = count;
entry.stageFlags = stage_flags;
entry.pImmutableSamplers = nullptr;
bindings.push_back({
.binding = binding++,
.descriptorType = descriptor_type,
.descriptorCount = count,
.stageFlags = stage_flags,
.pImmutableSamplers = nullptr,
});
}
}
@@ -259,10 +260,10 @@ VKComputePipeline& VKPipelineCache::GetComputePipeline(const ComputePipelineCach
}
}
Specialization specialization;
specialization.workgroup_size = key.workgroup_size;
specialization.shared_memory_size = key.shared_memory_size;
const Specialization specialization{
.workgroup_size = key.workgroup_size,
.shared_memory_size = key.shared_memory_size,
};
const SPIRVShader spirv_shader{Decompile(device, shader->GetIR(), ShaderType::Compute,
shader->GetRegistry(), specialization),
shader->GetEntries()};
@@ -370,13 +371,14 @@ void AddEntry(std::vector<VkDescriptorUpdateTemplateEntry>& template_entries, u3
if constexpr (descriptor_type == COMBINED_IMAGE_SAMPLER) {
for (u32 i = 0; i < count; ++i) {
const u32 num_samplers = container[i].size;
VkDescriptorUpdateTemplateEntry& entry = template_entries.emplace_back();
entry.dstBinding = binding;
entry.dstArrayElement = 0;
entry.descriptorCount = num_samplers;
entry.descriptorType = descriptor_type;
entry.offset = offset;
entry.stride = entry_size;
template_entries.push_back({
.dstBinding = binding,
.dstArrayElement = 0,
.descriptorCount = num_samplers,
.descriptorType = descriptor_type,
.offset = offset,
.stride = entry_size,
});
++binding;
offset += num_samplers * entry_size;
@@ -389,22 +391,24 @@ void AddEntry(std::vector<VkDescriptorUpdateTemplateEntry>& template_entries, u3
// Nvidia has a bug where updating multiple texels at once causes the driver to crash.
// Note: Fixed in driver Windows 443.24, Linux 440.66.15
for (u32 i = 0; i < count; ++i) {
VkDescriptorUpdateTemplateEntry& entry = template_entries.emplace_back();
entry.dstBinding = binding + i;
entry.dstArrayElement = 0;
entry.descriptorCount = 1;
entry.descriptorType = descriptor_type;
entry.offset = static_cast<std::size_t>(offset + i * entry_size);
entry.stride = entry_size;
template_entries.push_back({
.dstBinding = binding + i,
.dstArrayElement = 0,
.descriptorCount = 1,
.descriptorType = descriptor_type,
.offset = static_cast<std::size_t>(offset + i * entry_size),
.stride = entry_size,
});
}
} else if (count > 0) {
VkDescriptorUpdateTemplateEntry& entry = template_entries.emplace_back();
entry.dstBinding = binding;
entry.dstArrayElement = 0;
entry.descriptorCount = count;
entry.descriptorType = descriptor_type;
entry.offset = offset;
entry.stride = entry_size;
template_entries.push_back({
.dstBinding = binding,
.dstArrayElement = 0,
.descriptorCount = count,
.descriptorType = descriptor_type,
.offset = offset,
.stride = entry_size,
});
}
offset += count * entry_size;
binding += count;

View File

@@ -47,14 +47,14 @@ std::pair<VkQueryPool, u32> QueryPool::Commit(VKFence& fence) {
void QueryPool::Allocate(std::size_t begin, std::size_t end) {
usage.resize(end);
VkQueryPoolCreateInfo query_pool_ci;
query_pool_ci.sType = VK_STRUCTURE_TYPE_QUERY_POOL_CREATE_INFO;
query_pool_ci.pNext = nullptr;
query_pool_ci.flags = 0;
query_pool_ci.queryType = GetTarget(type);
query_pool_ci.queryCount = static_cast<u32>(end - begin);
query_pool_ci.pipelineStatistics = 0;
pools.push_back(device->GetLogical().CreateQueryPool(query_pool_ci));
pools.push_back(device->GetLogical().CreateQueryPool({
.sType = VK_STRUCTURE_TYPE_QUERY_POOL_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.queryType = GetTarget(type),
.queryCount = static_cast<u32>(end - begin),
.pipelineStatistics = 0,
}));
}
void QueryPool::Reserve(std::pair<VkQueryPool, u32> query) {

View File

@@ -64,20 +64,22 @@ VkViewport GetViewportState(const VKDevice& device, const Maxwell& regs, std::si
const auto& src = regs.viewport_transform[index];
const float width = src.scale_x * 2.0f;
const float height = src.scale_y * 2.0f;
VkViewport viewport;
viewport.x = src.translate_x - src.scale_x;
viewport.y = src.translate_y - src.scale_y;
viewport.width = width != 0.0f ? width : 1.0f;
viewport.height = height != 0.0f ? height : 1.0f;
const float reduce_z = regs.depth_mode == Maxwell::DepthMode::MinusOneToOne ? 1.0f : 0.0f;
viewport.minDepth = src.translate_z - src.scale_z * reduce_z;
viewport.maxDepth = src.translate_z + src.scale_z;
VkViewport viewport{
.x = src.translate_x - src.scale_x,
.y = src.translate_y - src.scale_y,
.width = width != 0.0f ? width : 1.0f,
.height = height != 0.0f ? height : 1.0f,
.minDepth = src.translate_z - src.scale_z * reduce_z,
.maxDepth = src.translate_z + src.scale_z,
};
if (!device.IsExtDepthRangeUnrestrictedSupported()) {
viewport.minDepth = std::clamp(viewport.minDepth, 0.0f, 1.0f);
viewport.maxDepth = std::clamp(viewport.maxDepth, 0.0f, 1.0f);
}
return viewport;
}
@@ -508,10 +510,11 @@ void RasterizerVulkan::Clear() {
const u32 color_attachment = regs.clear_buffers.RT;
scheduler.Record([color_attachment, clear_value, clear_rect](vk::CommandBuffer cmdbuf) {
VkClearAttachment attachment;
attachment.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT;
attachment.colorAttachment = color_attachment;
attachment.clearValue = clear_value;
const VkClearAttachment attachment{
.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
.colorAttachment = color_attachment,
.clearValue = clear_value,
};
cmdbuf.ClearAttachments(attachment, clear_rect);
});
}
@@ -551,13 +554,16 @@ void RasterizerVulkan::DispatchCompute(GPUVAddr code_addr) {
query_cache.UpdateCounters();
const auto& launch_desc = system.GPU().KeplerCompute().launch_description;
ComputePipelineCacheKey key;
key.shader = code_addr;
key.shared_memory_size = launch_desc.shared_alloc;
key.workgroup_size = {launch_desc.block_dim_x, launch_desc.block_dim_y,
launch_desc.block_dim_z};
auto& pipeline = pipeline_cache.GetComputePipeline(key);
auto& pipeline = pipeline_cache.GetComputePipeline({
.shader = code_addr,
.shared_memory_size = launch_desc.shared_alloc,
.workgroup_size =
{
launch_desc.block_dim_x,
launch_desc.block_dim_y,
launch_desc.block_dim_z,
},
});
// Compute dispatches can't be executed inside a renderpass
scheduler.RequestOutsideRenderPassOperationContext();
@@ -841,17 +847,17 @@ std::tuple<VkFramebuffer, VkExtent2D> RasterizerVulkan::ConfigureFramebuffers(
const auto [fbentry, is_cache_miss] = framebuffer_cache.try_emplace(key);
auto& framebuffer = fbentry->second;
if (is_cache_miss) {
VkFramebufferCreateInfo framebuffer_ci;
framebuffer_ci.sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO;
framebuffer_ci.pNext = nullptr;
framebuffer_ci.flags = 0;
framebuffer_ci.renderPass = key.renderpass;
framebuffer_ci.attachmentCount = static_cast<u32>(key.views.size());
framebuffer_ci.pAttachments = key.views.data();
framebuffer_ci.width = key.width;
framebuffer_ci.height = key.height;
framebuffer_ci.layers = key.layers;
framebuffer = device.GetLogical().CreateFramebuffer(framebuffer_ci);
framebuffer = device.GetLogical().CreateFramebuffer({
.sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.renderPass = key.renderpass,
.attachmentCount = static_cast<u32>(key.views.size()),
.pAttachments = key.views.data(),
.width = key.width,
.height = key.height,
.layers = key.layers,
});
}
return {*framebuffer, VkExtent2D{key.width, key.height}};
@@ -1553,17 +1559,17 @@ VkBuffer RasterizerVulkan::DefaultBuffer() {
return *default_buffer;
}
VkBufferCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.size = DEFAULT_BUFFER_SIZE;
ci.usage = VK_BUFFER_USAGE_TRANSFER_DST_BIT | VK_BUFFER_USAGE_VERTEX_BUFFER_BIT |
VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT;
ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE;
ci.queueFamilyIndexCount = 0;
ci.pQueueFamilyIndices = nullptr;
default_buffer = device.GetLogical().CreateBuffer(ci);
default_buffer = device.GetLogical().CreateBuffer({
.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.size = DEFAULT_BUFFER_SIZE,
.usage = VK_BUFFER_USAGE_TRANSFER_DST_BIT | VK_BUFFER_USAGE_VERTEX_BUFFER_BIT |
VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT,
.sharingMode = VK_SHARING_MODE_EXCLUSIVE,
.queueFamilyIndexCount = 0,
.pQueueFamilyIndices = nullptr,
});
default_buffer_commit = memory_manager.Commit(default_buffer, false);
scheduler.RequestOutsideRenderPassOperationContext();

View File

@@ -39,10 +39,14 @@ VkRenderPass VKRenderPassCache::GetRenderPass(const RenderPassParams& params) {
vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& params) const {
using namespace VideoCore::Surface;
std::vector<VkAttachmentDescription> descriptors;
std::vector<VkAttachmentReference> color_references;
const std::size_t num_attachments = static_cast<std::size_t>(params.num_color_attachments);
std::vector<VkAttachmentDescription> descriptors;
descriptors.reserve(num_attachments);
std::vector<VkAttachmentReference> color_references;
color_references.reserve(num_attachments);
for (std::size_t rt = 0; rt < num_attachments; ++rt) {
const auto guest_format = static_cast<Tegra::RenderTargetFormat>(params.color_formats[rt]);
const PixelFormat pixel_format = PixelFormatFromRenderTargetFormat(guest_format);
@@ -54,20 +58,22 @@ vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& param
const VkImageLayout color_layout = ((params.texceptions >> rt) & 1) != 0
? VK_IMAGE_LAYOUT_GENERAL
: VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL;
VkAttachmentDescription& descriptor = descriptors.emplace_back();
descriptor.flags = VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT;
descriptor.format = format.format;
descriptor.samples = VK_SAMPLE_COUNT_1_BIT;
descriptor.loadOp = VK_ATTACHMENT_LOAD_OP_LOAD;
descriptor.storeOp = VK_ATTACHMENT_STORE_OP_STORE;
descriptor.stencilLoadOp = VK_ATTACHMENT_LOAD_OP_DONT_CARE;
descriptor.stencilStoreOp = VK_ATTACHMENT_STORE_OP_DONT_CARE;
descriptor.initialLayout = color_layout;
descriptor.finalLayout = color_layout;
descriptors.push_back({
.flags = VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT,
.format = format.format,
.samples = VK_SAMPLE_COUNT_1_BIT,
.loadOp = VK_ATTACHMENT_LOAD_OP_LOAD,
.storeOp = VK_ATTACHMENT_STORE_OP_STORE,
.stencilLoadOp = VK_ATTACHMENT_LOAD_OP_DONT_CARE,
.stencilStoreOp = VK_ATTACHMENT_STORE_OP_DONT_CARE,
.initialLayout = color_layout,
.finalLayout = color_layout,
});
VkAttachmentReference& reference = color_references.emplace_back();
reference.attachment = static_cast<u32>(rt);
reference.layout = color_layout;
color_references.push_back({
.attachment = static_cast<u32>(rt),
.layout = color_layout,
});
}
VkAttachmentReference zeta_attachment_ref;
@@ -82,32 +88,36 @@ vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& param
const VkImageLayout zeta_layout = params.zeta_texception != 0
? VK_IMAGE_LAYOUT_GENERAL
: VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL;
VkAttachmentDescription& descriptor = descriptors.emplace_back();
descriptor.flags = 0;
descriptor.format = format.format;
descriptor.samples = VK_SAMPLE_COUNT_1_BIT;
descriptor.loadOp = VK_ATTACHMENT_LOAD_OP_LOAD;
descriptor.storeOp = VK_ATTACHMENT_STORE_OP_STORE;
descriptor.stencilLoadOp = VK_ATTACHMENT_LOAD_OP_LOAD;
descriptor.stencilStoreOp = VK_ATTACHMENT_STORE_OP_STORE;
descriptor.initialLayout = zeta_layout;
descriptor.finalLayout = zeta_layout;
descriptors.push_back({
.flags = 0,
.format = format.format,
.samples = VK_SAMPLE_COUNT_1_BIT,
.loadOp = VK_ATTACHMENT_LOAD_OP_LOAD,
.storeOp = VK_ATTACHMENT_STORE_OP_STORE,
.stencilLoadOp = VK_ATTACHMENT_LOAD_OP_LOAD,
.stencilStoreOp = VK_ATTACHMENT_STORE_OP_STORE,
.initialLayout = zeta_layout,
.finalLayout = zeta_layout,
});
zeta_attachment_ref.attachment = static_cast<u32>(num_attachments);
zeta_attachment_ref.layout = zeta_layout;
zeta_attachment_ref = {
.attachment = static_cast<u32>(num_attachments),
.layout = zeta_layout,
};
}
VkSubpassDescription subpass_description;
subpass_description.flags = 0;
subpass_description.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS;
subpass_description.inputAttachmentCount = 0;
subpass_description.pInputAttachments = nullptr;
subpass_description.colorAttachmentCount = static_cast<u32>(color_references.size());
subpass_description.pColorAttachments = color_references.data();
subpass_description.pResolveAttachments = nullptr;
subpass_description.pDepthStencilAttachment = has_zeta ? &zeta_attachment_ref : nullptr;
subpass_description.preserveAttachmentCount = 0;
subpass_description.pPreserveAttachments = nullptr;
const VkSubpassDescription subpass_description{
.flags = 0,
.pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS,
.inputAttachmentCount = 0,
.pInputAttachments = nullptr,
.colorAttachmentCount = static_cast<u32>(color_references.size()),
.pColorAttachments = color_references.data(),
.pResolveAttachments = nullptr,
.pDepthStencilAttachment = has_zeta ? &zeta_attachment_ref : nullptr,
.preserveAttachmentCount = 0,
.pPreserveAttachments = nullptr,
};
VkAccessFlags access = 0;
VkPipelineStageFlags stage = 0;
@@ -122,26 +132,27 @@ vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& param
stage |= VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT;
}
VkSubpassDependency subpass_dependency;
subpass_dependency.srcSubpass = VK_SUBPASS_EXTERNAL;
subpass_dependency.dstSubpass = 0;
subpass_dependency.srcStageMask = stage;
subpass_dependency.dstStageMask = stage;
subpass_dependency.srcAccessMask = 0;
subpass_dependency.dstAccessMask = access;
subpass_dependency.dependencyFlags = 0;
const VkSubpassDependency subpass_dependency{
.srcSubpass = VK_SUBPASS_EXTERNAL,
.dstSubpass = 0,
.srcStageMask = stage,
.dstStageMask = stage,
.srcAccessMask = 0,
.dstAccessMask = access,
.dependencyFlags = 0,
};
VkRenderPassCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.attachmentCount = static_cast<u32>(descriptors.size());
ci.pAttachments = descriptors.data();
ci.subpassCount = 1;
ci.pSubpasses = &subpass_description;
ci.dependencyCount = 1;
ci.pDependencies = &subpass_dependency;
return device.GetLogical().CreateRenderPass(ci);
return device.GetLogical().CreateRenderPass({
.sType = VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.attachmentCount = static_cast<u32>(descriptors.size()),
.pAttachments = descriptors.data(),
.subpassCount = 1,
.pSubpasses = &subpass_description,
.dependencyCount = 1,
.pDependencies = &subpass_dependency,
});
}
} // namespace Vulkan

View File

@@ -18,33 +18,32 @@ namespace {
constexpr std::size_t COMMAND_BUFFER_POOL_SIZE = 0x1000;
constexpr std::size_t FENCES_GROW_STEP = 0x40;
VkFenceCreateInfo BuildFenceCreateInfo() {
VkFenceCreateInfo fence_ci;
fence_ci.sType = VK_STRUCTURE_TYPE_FENCE_CREATE_INFO;
fence_ci.pNext = nullptr;
fence_ci.flags = 0;
return fence_ci;
constexpr VkFenceCreateInfo BuildFenceCreateInfo() {
return {
.sType = VK_STRUCTURE_TYPE_FENCE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
};
}
} // Anonymous namespace
class CommandBufferPool final : public VKFencedPool {
public:
CommandBufferPool(const VKDevice& device)
explicit CommandBufferPool(const VKDevice& device)
: VKFencedPool(COMMAND_BUFFER_POOL_SIZE), device{device} {}
void Allocate(std::size_t begin, std::size_t end) override {
// Command buffers are going to be commited, recorded, executed every single usage cycle.
// They are also going to be reseted when commited.
VkCommandPoolCreateInfo command_pool_ci;
command_pool_ci.sType = VK_STRUCTURE_TYPE_COMMAND_POOL_CREATE_INFO;
command_pool_ci.pNext = nullptr;
command_pool_ci.flags =
VK_COMMAND_POOL_CREATE_TRANSIENT_BIT | VK_COMMAND_POOL_CREATE_RESET_COMMAND_BUFFER_BIT;
command_pool_ci.queueFamilyIndex = device.GetGraphicsFamily();
Pool& pool = pools.emplace_back();
pool.handle = device.GetLogical().CreateCommandPool(command_pool_ci);
pool.handle = device.GetLogical().CreateCommandPool({
.sType = VK_STRUCTURE_TYPE_COMMAND_POOL_CREATE_INFO,
.pNext = nullptr,
.flags = VK_COMMAND_POOL_CREATE_TRANSIENT_BIT |
VK_COMMAND_POOL_CREATE_RESET_COMMAND_BUFFER_BIT,
.queueFamilyIndex = device.GetGraphicsFamily(),
});
pool.cmdbufs = pool.handle.Allocate(COMMAND_BUFFER_POOL_SIZE);
}

View File

@@ -44,32 +44,35 @@ vk::Sampler VKSamplerCache::CreateSampler(const Tegra::Texture::TSCEntry& tsc) c
const bool arbitrary_borders = device.IsExtCustomBorderColorSupported();
const std::array color = tsc.GetBorderColor();
VkSamplerCustomBorderColorCreateInfoEXT border;
border.sType = VK_STRUCTURE_TYPE_SAMPLER_CUSTOM_BORDER_COLOR_CREATE_INFO_EXT;
border.pNext = nullptr;
border.format = VK_FORMAT_UNDEFINED;
VkSamplerCustomBorderColorCreateInfoEXT border{
.sType = VK_STRUCTURE_TYPE_SAMPLER_CUSTOM_BORDER_COLOR_CREATE_INFO_EXT,
.pNext = nullptr,
.format = VK_FORMAT_UNDEFINED,
};
std::memcpy(&border.customBorderColor, color.data(), sizeof(color));
VkSamplerCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO;
ci.pNext = arbitrary_borders ? &border : nullptr;
ci.flags = 0;
ci.magFilter = MaxwellToVK::Sampler::Filter(tsc.mag_filter);
ci.minFilter = MaxwellToVK::Sampler::Filter(tsc.min_filter);
ci.mipmapMode = MaxwellToVK::Sampler::MipmapMode(tsc.mipmap_filter);
ci.addressModeU = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_u, tsc.mag_filter);
ci.addressModeV = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_v, tsc.mag_filter);
ci.addressModeW = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_p, tsc.mag_filter);
ci.mipLodBias = tsc.GetLodBias();
ci.anisotropyEnable = tsc.GetMaxAnisotropy() > 1.0f ? VK_TRUE : VK_FALSE;
ci.maxAnisotropy = tsc.GetMaxAnisotropy();
ci.compareEnable = tsc.depth_compare_enabled;
ci.compareOp = MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func);
ci.minLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.0f : tsc.GetMinLod();
ci.maxLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.25f : tsc.GetMaxLod();
ci.borderColor = arbitrary_borders ? VK_BORDER_COLOR_INT_CUSTOM_EXT : ConvertBorderColor(color);
ci.unnormalizedCoordinates = VK_FALSE;
return device.GetLogical().CreateSampler(ci);
return device.GetLogical().CreateSampler({
.sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO,
.pNext = arbitrary_borders ? &border : nullptr,
.flags = 0,
.magFilter = MaxwellToVK::Sampler::Filter(tsc.mag_filter),
.minFilter = MaxwellToVK::Sampler::Filter(tsc.min_filter),
.mipmapMode = MaxwellToVK::Sampler::MipmapMode(tsc.mipmap_filter),
.addressModeU = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_u, tsc.mag_filter),
.addressModeV = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_v, tsc.mag_filter),
.addressModeW = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_p, tsc.mag_filter),
.mipLodBias = tsc.GetLodBias(),
.anisotropyEnable =
static_cast<VkBool32>(tsc.GetMaxAnisotropy() > 1.0f ? VK_TRUE : VK_FALSE),
.maxAnisotropy = tsc.GetMaxAnisotropy(),
.compareEnable = tsc.depth_compare_enabled,
.compareOp = MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func),
.minLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.0f : tsc.GetMinLod(),
.maxLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.25f : tsc.GetMaxLod(),
.borderColor =
arbitrary_borders ? VK_BORDER_COLOR_INT_CUSTOM_EXT : ConvertBorderColor(color),
.unnormalizedCoordinates = VK_FALSE,
});
}
VkSampler VKSamplerCache::ToSamplerType(const vk::Sampler& sampler) const {

View File

@@ -100,16 +100,19 @@ void VKScheduler::RequestRenderpass(VkRenderPass renderpass, VkFramebuffer frame
state.framebuffer = framebuffer;
state.render_area = render_area;
VkRenderPassBeginInfo renderpass_bi;
renderpass_bi.sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO;
renderpass_bi.pNext = nullptr;
renderpass_bi.renderPass = renderpass;
renderpass_bi.framebuffer = framebuffer;
renderpass_bi.renderArea.offset.x = 0;
renderpass_bi.renderArea.offset.y = 0;
renderpass_bi.renderArea.extent = render_area;
renderpass_bi.clearValueCount = 0;
renderpass_bi.pClearValues = nullptr;
const VkRenderPassBeginInfo renderpass_bi{
.sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO,
.pNext = nullptr,
.renderPass = renderpass,
.framebuffer = framebuffer,
.renderArea =
{
.offset = {.x = 0, .y = 0},
.extent = render_area,
},
.clearValueCount = 0,
.pClearValues = nullptr,
};
Record([renderpass_bi, end_renderpass](vk::CommandBuffer cmdbuf) {
if (end_renderpass) {
@@ -157,16 +160,17 @@ void VKScheduler::SubmitExecution(VkSemaphore semaphore) {
current_cmdbuf.End();
VkSubmitInfo submit_info;
submit_info.sType = VK_STRUCTURE_TYPE_SUBMIT_INFO;
submit_info.pNext = nullptr;
submit_info.waitSemaphoreCount = 0;
submit_info.pWaitSemaphores = nullptr;
submit_info.pWaitDstStageMask = nullptr;
submit_info.commandBufferCount = 1;
submit_info.pCommandBuffers = current_cmdbuf.address();
submit_info.signalSemaphoreCount = semaphore ? 1 : 0;
submit_info.pSignalSemaphores = &semaphore;
const VkSubmitInfo submit_info{
.sType = VK_STRUCTURE_TYPE_SUBMIT_INFO,
.pNext = nullptr,
.waitSemaphoreCount = 0,
.pWaitSemaphores = nullptr,
.pWaitDstStageMask = nullptr,
.commandBufferCount = 1,
.pCommandBuffers = current_cmdbuf.address(),
.signalSemaphoreCount = semaphore ? 1U : 0U,
.pSignalSemaphores = &semaphore,
};
switch (const VkResult result = device.GetGraphicsQueue().Submit(submit_info, *current_fence)) {
case VK_SUCCESS:
break;
@@ -181,19 +185,18 @@ void VKScheduler::SubmitExecution(VkSemaphore semaphore) {
void VKScheduler::AllocateNewContext() {
++ticks;
VkCommandBufferBeginInfo cmdbuf_bi;
cmdbuf_bi.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_BEGIN_INFO;
cmdbuf_bi.pNext = nullptr;
cmdbuf_bi.flags = VK_COMMAND_BUFFER_USAGE_ONE_TIME_SUBMIT_BIT;
cmdbuf_bi.pInheritanceInfo = nullptr;
std::unique_lock lock{mutex};
current_fence = next_fence;
next_fence = &resource_manager.CommitFence();
current_cmdbuf = vk::CommandBuffer(resource_manager.CommitCommandBuffer(*current_fence),
device.GetDispatchLoader());
current_cmdbuf.Begin(cmdbuf_bi);
current_cmdbuf.Begin({
.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_BEGIN_INFO,
.pNext = nullptr,
.flags = VK_COMMAND_BUFFER_USAGE_ONE_TIME_SUBMIT_BIT,
.pInheritanceInfo = nullptr,
});
// Enable counters once again. These are disabled when a command buffer is finished.
if (query_cache) {

View File

@@ -19,13 +19,13 @@ vk::ShaderModule BuildShader(const VKDevice& device, std::size_t code_size, cons
const auto data = std::make_unique<u32[]>(code_size / sizeof(u32));
std::memcpy(data.get(), code_data, code_size);
VkShaderModuleCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.codeSize = code_size;
ci.pCode = data.get();
return device.GetLogical().CreateShaderModule(ci);
return device.GetLogical().CreateShaderModule({
.sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.codeSize = code_size,
.pCode = data.get(),
});
}
} // namespace Vulkan

View File

@@ -71,20 +71,19 @@ VKBuffer* VKStagingBufferPool::TryGetReservedBuffer(std::size_t size, bool host_
VKBuffer& VKStagingBufferPool::CreateStagingBuffer(std::size_t size, bool host_visible) {
const u32 log2 = Common::Log2Ceil64(size);
VkBufferCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.size = 1ULL << log2;
ci.usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT |
VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT | VK_BUFFER_USAGE_STORAGE_BUFFER_BIT |
VK_BUFFER_USAGE_INDEX_BUFFER_BIT | VK_BUFFER_USAGE_VERTEX_BUFFER_BIT;
ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE;
ci.queueFamilyIndexCount = 0;
ci.pQueueFamilyIndices = nullptr;
auto buffer = std::make_unique<VKBuffer>();
buffer->handle = device.GetLogical().CreateBuffer(ci);
buffer->handle = device.GetLogical().CreateBuffer({
.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.size = 1ULL << log2,
.usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT |
VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT | VK_BUFFER_USAGE_STORAGE_BUFFER_BIT |
VK_BUFFER_USAGE_INDEX_BUFFER_BIT | VK_BUFFER_USAGE_VERTEX_BUFFER_BIT,
.sharingMode = VK_SHARING_MODE_EXCLUSIVE,
.queueFamilyIndexCount = 0,
.pQueueFamilyIndices = nullptr,
});
buffer->commit = memory_manager.Commit(buffer->handle, host_visible);
auto& entries = GetCache(host_visible)[log2].entries;

View File

@@ -158,6 +158,7 @@ void StateTracker::Initialize() {
SetupDirtyFrontFace(tables);
SetupDirtyPrimitiveTopology(tables);
SetupDirtyStencilOp(tables);
SetupDirtyStencilTestEnable(tables);
}
void StateTracker::InvalidateCommandBufferState() {

View File

@@ -122,30 +122,27 @@ void VKStreamBuffer::CreateBuffers(VkBufferUsageFlags usage) {
// Substract from the preferred heap size some bytes to avoid getting out of memory.
const VkDeviceSize heap_size = memory_properties.memoryHeaps[preferred_heap].size;
const VkDeviceSize allocable_size = heap_size - 9 * 1024 * 1024;
VkBufferCreateInfo buffer_ci;
buffer_ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO;
buffer_ci.pNext = nullptr;
buffer_ci.flags = 0;
buffer_ci.size = std::min(PREFERRED_STREAM_BUFFER_SIZE, allocable_size);
buffer_ci.usage = usage;
buffer_ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE;
buffer_ci.queueFamilyIndexCount = 0;
buffer_ci.pQueueFamilyIndices = nullptr;
buffer = device.GetLogical().CreateBuffer(buffer_ci);
buffer = device.GetLogical().CreateBuffer({
.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.size = std::min(PREFERRED_STREAM_BUFFER_SIZE, allocable_size),
.usage = usage,
.sharingMode = VK_SHARING_MODE_EXCLUSIVE,
.queueFamilyIndexCount = 0,
.pQueueFamilyIndices = nullptr,
});
const auto requirements = device.GetLogical().GetBufferMemoryRequirements(*buffer);
const u32 required_flags = requirements.memoryTypeBits;
stream_buffer_size = static_cast<u64>(requirements.size);
VkMemoryAllocateInfo memory_ai;
memory_ai.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO;
memory_ai.pNext = nullptr;
memory_ai.allocationSize = requirements.size;
memory_ai.memoryTypeIndex = GetMemoryType(memory_properties, required_flags);
memory = device.GetLogical().AllocateMemory(memory_ai);
memory = device.GetLogical().AllocateMemory({
.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO,
.pNext = nullptr,
.allocationSize = requirements.size,
.memoryTypeIndex = GetMemoryType(memory_properties, required_flags),
});
buffer.BindMemory(*memory, 0);
}

View File

@@ -95,15 +95,16 @@ bool VKSwapchain::Present(VkSemaphore render_semaphore, VKFence& fence) {
const auto present_queue{device.GetPresentQueue()};
bool recreated = false;
VkPresentInfoKHR present_info;
present_info.sType = VK_STRUCTURE_TYPE_PRESENT_INFO_KHR;
present_info.pNext = nullptr;
present_info.waitSemaphoreCount = render_semaphore ? 2U : 1U;
present_info.pWaitSemaphores = semaphores.data();
present_info.swapchainCount = 1;
present_info.pSwapchains = swapchain.address();
present_info.pImageIndices = &image_index;
present_info.pResults = nullptr;
const VkPresentInfoKHR present_info{
.sType = VK_STRUCTURE_TYPE_PRESENT_INFO_KHR,
.pNext = nullptr,
.waitSemaphoreCount = render_semaphore ? 2U : 1U,
.pWaitSemaphores = semaphores.data(),
.swapchainCount = 1,
.pSwapchains = swapchain.address(),
.pImageIndices = &image_index,
.pResults = nullptr,
};
switch (const VkResult result = present_queue.Present(present_info)) {
case VK_SUCCESS:
@@ -147,24 +148,25 @@ void VKSwapchain::CreateSwapchain(const VkSurfaceCapabilitiesKHR& capabilities,
requested_image_count = capabilities.maxImageCount;
}
VkSwapchainCreateInfoKHR swapchain_ci;
swapchain_ci.sType = VK_STRUCTURE_TYPE_SWAPCHAIN_CREATE_INFO_KHR;
swapchain_ci.pNext = nullptr;
swapchain_ci.flags = 0;
swapchain_ci.surface = surface;
swapchain_ci.minImageCount = requested_image_count;
swapchain_ci.imageFormat = surface_format.format;
swapchain_ci.imageColorSpace = surface_format.colorSpace;
swapchain_ci.imageArrayLayers = 1;
swapchain_ci.imageUsage = VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT;
swapchain_ci.imageSharingMode = VK_SHARING_MODE_EXCLUSIVE;
swapchain_ci.queueFamilyIndexCount = 0;
swapchain_ci.pQueueFamilyIndices = nullptr;
swapchain_ci.preTransform = capabilities.currentTransform;
swapchain_ci.compositeAlpha = VK_COMPOSITE_ALPHA_OPAQUE_BIT_KHR;
swapchain_ci.presentMode = present_mode;
swapchain_ci.clipped = VK_FALSE;
swapchain_ci.oldSwapchain = nullptr;
VkSwapchainCreateInfoKHR swapchain_ci{
.sType = VK_STRUCTURE_TYPE_SWAPCHAIN_CREATE_INFO_KHR,
.pNext = nullptr,
.flags = 0,
.surface = surface,
.minImageCount = requested_image_count,
.imageFormat = surface_format.format,
.imageColorSpace = surface_format.colorSpace,
.imageArrayLayers = 1,
.imageUsage = VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT,
.imageSharingMode = VK_SHARING_MODE_EXCLUSIVE,
.queueFamilyIndexCount = 0,
.pQueueFamilyIndices = nullptr,
.preTransform = capabilities.currentTransform,
.compositeAlpha = VK_COMPOSITE_ALPHA_OPAQUE_BIT_KHR,
.presentMode = present_mode,
.clipped = VK_FALSE,
.oldSwapchain = nullptr,
};
const u32 graphics_family{device.GetGraphicsFamily()};
const u32 present_family{device.GetPresentFamily()};
@@ -173,8 +175,6 @@ void VKSwapchain::CreateSwapchain(const VkSurfaceCapabilitiesKHR& capabilities,
swapchain_ci.imageSharingMode = VK_SHARING_MODE_CONCURRENT;
swapchain_ci.queueFamilyIndexCount = static_cast<u32>(queue_indices.size());
swapchain_ci.pQueueFamilyIndices = queue_indices.data();
} else {
swapchain_ci.imageSharingMode = VK_SHARING_MODE_EXCLUSIVE;
}
// Request the size again to reduce the possibility of a TOCTOU race condition.
@@ -200,20 +200,28 @@ void VKSwapchain::CreateSemaphores() {
}
void VKSwapchain::CreateImageViews() {
VkImageViewCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
// ci.image
ci.viewType = VK_IMAGE_VIEW_TYPE_2D;
ci.format = image_format;
ci.components = {VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY,
VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY};
ci.subresourceRange.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT;
ci.subresourceRange.baseMipLevel = 0;
ci.subresourceRange.levelCount = 1;
ci.subresourceRange.baseArrayLayer = 0;
ci.subresourceRange.layerCount = 1;
VkImageViewCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.viewType = VK_IMAGE_VIEW_TYPE_2D,
.format = image_format,
.components =
{
.r = VK_COMPONENT_SWIZZLE_IDENTITY,
.g = VK_COMPONENT_SWIZZLE_IDENTITY,
.b = VK_COMPONENT_SWIZZLE_IDENTITY,
.a = VK_COMPONENT_SWIZZLE_IDENTITY,
},
.subresourceRange =
{
.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
.baseMipLevel = 0,
.levelCount = 1,
.baseArrayLayer = 0,
.layerCount = 1,
},
};
image_views.resize(image_count);
for (std::size_t i = 0; i < image_count; i++) {

View File

@@ -95,17 +95,18 @@ VkImageViewType GetImageViewType(SurfaceTarget target) {
vk::Buffer CreateBuffer(const VKDevice& device, const SurfaceParams& params,
std::size_t host_memory_size) {
// TODO(Rodrigo): Move texture buffer creation to the buffer cache
VkBufferCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.size = static_cast<VkDeviceSize>(host_memory_size);
ci.usage = VK_BUFFER_USAGE_UNIFORM_TEXEL_BUFFER_BIT | VK_BUFFER_USAGE_STORAGE_TEXEL_BUFFER_BIT |
VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT;
ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE;
ci.queueFamilyIndexCount = 0;
ci.pQueueFamilyIndices = nullptr;
return device.GetLogical().CreateBuffer(ci);
return device.GetLogical().CreateBuffer({
.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.size = static_cast<VkDeviceSize>(host_memory_size),
.usage = VK_BUFFER_USAGE_UNIFORM_TEXEL_BUFFER_BIT |
VK_BUFFER_USAGE_STORAGE_TEXEL_BUFFER_BIT | VK_BUFFER_USAGE_TRANSFER_SRC_BIT |
VK_BUFFER_USAGE_TRANSFER_DST_BIT,
.sharingMode = VK_SHARING_MODE_EXCLUSIVE,
.queueFamilyIndexCount = 0,
.pQueueFamilyIndices = nullptr,
});
}
VkBufferViewCreateInfo GenerateBufferViewCreateInfo(const VKDevice& device,
@@ -113,15 +114,16 @@ VkBufferViewCreateInfo GenerateBufferViewCreateInfo(const VKDevice& device,
std::size_t host_memory_size) {
ASSERT(params.IsBuffer());
VkBufferViewCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_BUFFER_VIEW_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.buffer = buffer;
ci.format = MaxwellToVK::SurfaceFormat(device, FormatType::Buffer, params.pixel_format).format;
ci.offset = 0;
ci.range = static_cast<VkDeviceSize>(host_memory_size);
return ci;
return {
.sType = VK_STRUCTURE_TYPE_BUFFER_VIEW_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.buffer = buffer,
.format =
MaxwellToVK::SurfaceFormat(device, FormatType::Buffer, params.pixel_format).format,
.offset = 0,
.range = static_cast<VkDeviceSize>(host_memory_size),
};
}
VkImageCreateInfo GenerateImageCreateInfo(const VKDevice& device, const SurfaceParams& params) {
@@ -130,23 +132,23 @@ VkImageCreateInfo GenerateImageCreateInfo(const VKDevice& device, const SurfaceP
const auto [format, attachable, storage] =
MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, params.pixel_format);
VkImageCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.imageType = SurfaceTargetToImage(params.target);
ci.format = format;
ci.mipLevels = params.num_levels;
ci.arrayLayers = static_cast<u32>(params.GetNumLayers());
ci.samples = VK_SAMPLE_COUNT_1_BIT;
ci.tiling = VK_IMAGE_TILING_OPTIMAL;
ci.sharingMode = VK_SHARING_MODE_EXCLUSIVE;
ci.queueFamilyIndexCount = 0;
ci.pQueueFamilyIndices = nullptr;
ci.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED;
ci.usage = VK_IMAGE_USAGE_SAMPLED_BIT | VK_IMAGE_USAGE_TRANSFER_DST_BIT |
VK_IMAGE_USAGE_TRANSFER_SRC_BIT;
VkImageCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.imageType = SurfaceTargetToImage(params.target),
.format = format,
.mipLevels = params.num_levels,
.arrayLayers = static_cast<u32>(params.GetNumLayers()),
.samples = VK_SAMPLE_COUNT_1_BIT,
.tiling = VK_IMAGE_TILING_OPTIMAL,
.usage = VK_IMAGE_USAGE_SAMPLED_BIT | VK_IMAGE_USAGE_TRANSFER_DST_BIT |
VK_IMAGE_USAGE_TRANSFER_SRC_BIT,
.sharingMode = VK_SHARING_MODE_EXCLUSIVE,
.queueFamilyIndexCount = 0,
.pQueueFamilyIndices = nullptr,
.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED,
};
if (attachable) {
ci.usage |= params.IsPixelFormatZeta() ? VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT
: VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT;
@@ -281,12 +283,10 @@ void CachedSurface::UploadBuffer(const std::vector<u8>& staging_buffer) {
VkBufferMemoryBarrier barrier;
barrier.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER;
barrier.pNext = nullptr;
barrier.srcAccessMask = VK_PIPELINE_STAGE_TRANSFER_BIT;
barrier.dstAccessMask = VK_PIPELINE_STAGE_VERTEX_SHADER_BIT;
barrier.srcQueueFamilyIndex = VK_ACCESS_TRANSFER_WRITE_BIT;
barrier.dstQueueFamilyIndex = VK_ACCESS_SHADER_READ_BIT;
barrier.srcQueueFamilyIndex = 0;
barrier.dstQueueFamilyIndex = 0;
barrier.srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT;
barrier.dstAccessMask = VK_ACCESS_SHADER_READ_BIT;
barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; // They'll be ignored anyway
barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED;
barrier.buffer = dst_buffer;
barrier.offset = 0;
barrier.size = size;
@@ -323,22 +323,25 @@ void CachedSurface::UploadImage(const std::vector<u8>& staging_buffer) {
}
VkBufferImageCopy CachedSurface::GetBufferImageCopy(u32 level) const {
VkBufferImageCopy copy;
copy.bufferOffset = params.GetHostMipmapLevelOffset(level, is_converted);
copy.bufferRowLength = 0;
copy.bufferImageHeight = 0;
copy.imageSubresource.aspectMask = image->GetAspectMask();
copy.imageSubresource.mipLevel = level;
copy.imageSubresource.baseArrayLayer = 0;
copy.imageSubresource.layerCount = static_cast<u32>(params.GetNumLayers());
copy.imageOffset.x = 0;
copy.imageOffset.y = 0;
copy.imageOffset.z = 0;
copy.imageExtent.width = params.GetMipWidth(level);
copy.imageExtent.height = params.GetMipHeight(level);
copy.imageExtent.depth =
params.target == SurfaceTarget::Texture3D ? params.GetMipDepth(level) : 1;
return copy;
return {
.bufferOffset = params.GetHostMipmapLevelOffset(level, is_converted),
.bufferRowLength = 0,
.bufferImageHeight = 0,
.imageSubresource =
{
.aspectMask = image->GetAspectMask(),
.mipLevel = level,
.baseArrayLayer = 0,
.layerCount = static_cast<u32>(params.GetNumLayers()),
},
.imageOffset = {.x = 0, .y = 0, .z = 0},
.imageExtent =
{
.width = params.GetMipWidth(level),
.height = params.GetMipHeight(level),
.depth = params.target == SurfaceTarget::Texture3D ? params.GetMipDepth(level) : 1U,
},
};
}
VkImageSubresourceRange CachedSurface::GetImageSubresourceRange() const {
@@ -418,20 +421,29 @@ VkImageView CachedSurfaceView::GetImageView(SwizzleSource x_source, SwizzleSourc
ASSERT(num_slices == params.depth);
}
VkImageViewCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.image = surface.GetImageHandle();
ci.viewType = image_view_type;
ci.format = surface.GetImage().GetFormat();
ci.components = {swizzle[0], swizzle[1], swizzle[2], swizzle[3]};
ci.subresourceRange.aspectMask = aspect;
ci.subresourceRange.baseMipLevel = base_level;
ci.subresourceRange.levelCount = num_levels;
ci.subresourceRange.baseArrayLayer = base_layer;
ci.subresourceRange.layerCount = num_layers;
image_view = device.GetLogical().CreateImageView(ci);
image_view = device.GetLogical().CreateImageView({
.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.image = surface.GetImageHandle(),
.viewType = image_view_type,
.format = surface.GetImage().GetFormat(),
.components =
{
.r = swizzle[0],
.g = swizzle[1],
.b = swizzle[2],
.a = swizzle[3],
},
.subresourceRange =
{
.aspectMask = aspect,
.baseMipLevel = base_level,
.levelCount = num_levels,
.baseArrayLayer = base_layer,
.layerCount = num_layers,
},
});
return last_image_view = *image_view;
}
@@ -441,17 +453,26 @@ VkImageView CachedSurfaceView::GetAttachment() {
return *render_target;
}
VkImageViewCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.image = surface.GetImageHandle();
ci.format = surface.GetImage().GetFormat();
ci.components = {VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY,
VK_COMPONENT_SWIZZLE_IDENTITY, VK_COMPONENT_SWIZZLE_IDENTITY};
ci.subresourceRange.aspectMask = aspect_mask;
ci.subresourceRange.baseMipLevel = base_level;
ci.subresourceRange.levelCount = num_levels;
VkImageViewCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.image = surface.GetImageHandle(),
.format = surface.GetImage().GetFormat(),
.components =
{
.r = VK_COMPONENT_SWIZZLE_IDENTITY,
.g = VK_COMPONENT_SWIZZLE_IDENTITY,
.b = VK_COMPONENT_SWIZZLE_IDENTITY,
.a = VK_COMPONENT_SWIZZLE_IDENTITY,
},
.subresourceRange =
{
.aspectMask = aspect_mask,
.baseMipLevel = base_level,
.levelCount = num_levels,
},
};
if (image_view_type == VK_IMAGE_VIEW_TYPE_3D) {
ci.viewType = num_slices > 1 ? VK_IMAGE_VIEW_TYPE_2D_ARRAY : VK_IMAGE_VIEW_TYPE_2D;
ci.subresourceRange.baseArrayLayer = base_slice;
@@ -504,24 +525,40 @@ void VKTextureCache::ImageCopy(Surface& src_surface, Surface& dst_surface,
VK_PIPELINE_STAGE_TRANSFER_BIT, VK_ACCESS_TRANSFER_WRITE_BIT,
VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL);
VkImageCopy copy;
copy.srcSubresource.aspectMask = src_surface->GetAspectMask();
copy.srcSubresource.mipLevel = copy_params.source_level;
copy.srcSubresource.baseArrayLayer = copy_params.source_z;
copy.srcSubresource.layerCount = num_layers;
copy.srcOffset.x = copy_params.source_x;
copy.srcOffset.y = copy_params.source_y;
copy.srcOffset.z = 0;
copy.dstSubresource.aspectMask = dst_surface->GetAspectMask();
copy.dstSubresource.mipLevel = copy_params.dest_level;
copy.dstSubresource.baseArrayLayer = dst_base_layer;
copy.dstSubresource.layerCount = num_layers;
copy.dstOffset.x = copy_params.dest_x;
copy.dstOffset.y = copy_params.dest_y;
copy.dstOffset.z = dst_offset_z;
copy.extent.width = copy_params.width;
copy.extent.height = copy_params.height;
copy.extent.depth = extent_z;
const VkImageCopy copy{
.srcSubresource =
{
.aspectMask = src_surface->GetAspectMask(),
.mipLevel = copy_params.source_level,
.baseArrayLayer = copy_params.source_z,
.layerCount = num_layers,
},
.srcOffset =
{
.x = static_cast<s32>(copy_params.source_x),
.y = static_cast<s32>(copy_params.source_y),
.z = 0,
},
.dstSubresource =
{
.aspectMask = dst_surface->GetAspectMask(),
.mipLevel = copy_params.dest_level,
.baseArrayLayer = dst_base_layer,
.layerCount = num_layers,
},
.dstOffset =
{
.x = static_cast<s32>(copy_params.dest_x),
.y = static_cast<s32>(copy_params.dest_y),
.z = static_cast<s32>(dst_offset_z),
},
.extent =
{
.width = copy_params.width,
.height = copy_params.height,
.depth = extent_z,
},
};
const VkImage src_image = src_surface->GetImageHandle();
const VkImage dst_image = dst_surface->GetImageHandle();

View File

@@ -377,24 +377,26 @@ VkResult Free(VkDevice device, VkCommandPool handle, Span<VkCommandBuffer> buffe
Instance Instance::Create(Span<const char*> layers, Span<const char*> extensions,
InstanceDispatch& dld) noexcept {
VkApplicationInfo application_info;
application_info.sType = VK_STRUCTURE_TYPE_APPLICATION_INFO;
application_info.pNext = nullptr;
application_info.pApplicationName = "yuzu Emulator";
application_info.applicationVersion = VK_MAKE_VERSION(0, 1, 0);
application_info.pEngineName = "yuzu Emulator";
application_info.engineVersion = VK_MAKE_VERSION(0, 1, 0);
application_info.apiVersion = VK_API_VERSION_1_1;
static constexpr VkApplicationInfo application_info{
.sType = VK_STRUCTURE_TYPE_APPLICATION_INFO,
.pNext = nullptr,
.pApplicationName = "yuzu Emulator",
.applicationVersion = VK_MAKE_VERSION(0, 1, 0),
.pEngineName = "yuzu Emulator",
.engineVersion = VK_MAKE_VERSION(0, 1, 0),
.apiVersion = VK_API_VERSION_1_1,
};
VkInstanceCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_INSTANCE_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
ci.pApplicationInfo = &application_info;
ci.enabledLayerCount = layers.size();
ci.ppEnabledLayerNames = layers.data();
ci.enabledExtensionCount = extensions.size();
ci.ppEnabledExtensionNames = extensions.data();
const VkInstanceCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_INSTANCE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
.pApplicationInfo = &application_info,
.enabledLayerCount = layers.size(),
.ppEnabledLayerNames = layers.data(),
.enabledExtensionCount = extensions.size(),
.ppEnabledExtensionNames = extensions.data(),
};
VkInstance instance;
if (dld.vkCreateInstance(&ci, nullptr, &instance) != VK_SUCCESS) {
@@ -425,19 +427,20 @@ std::optional<std::vector<VkPhysicalDevice>> Instance::EnumeratePhysicalDevices(
DebugCallback Instance::TryCreateDebugCallback(
PFN_vkDebugUtilsMessengerCallbackEXT callback) noexcept {
VkDebugUtilsMessengerCreateInfoEXT ci;
ci.sType = VK_STRUCTURE_TYPE_DEBUG_UTILS_MESSENGER_CREATE_INFO_EXT;
ci.pNext = nullptr;
ci.flags = 0;
ci.messageSeverity = VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT;
ci.messageType = VK_DEBUG_UTILS_MESSAGE_TYPE_GENERAL_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_TYPE_VALIDATION_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_TYPE_PERFORMANCE_BIT_EXT;
ci.pfnUserCallback = callback;
ci.pUserData = nullptr;
const VkDebugUtilsMessengerCreateInfoEXT ci{
.sType = VK_STRUCTURE_TYPE_DEBUG_UTILS_MESSENGER_CREATE_INFO_EXT,
.pNext = nullptr,
.flags = 0,
.messageSeverity = VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT,
.messageType = VK_DEBUG_UTILS_MESSAGE_TYPE_GENERAL_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_TYPE_VALIDATION_BIT_EXT |
VK_DEBUG_UTILS_MESSAGE_TYPE_PERFORMANCE_BIT_EXT,
.pfnUserCallback = callback,
.pUserData = nullptr,
};
VkDebugUtilsMessengerEXT messenger;
if (dld->vkCreateDebugUtilsMessengerEXT(handle, &ci, nullptr, &messenger) != VK_SUCCESS) {
@@ -468,12 +471,13 @@ DescriptorSets DescriptorPool::Allocate(const VkDescriptorSetAllocateInfo& ai) c
}
CommandBuffers CommandPool::Allocate(std::size_t num_buffers, VkCommandBufferLevel level) const {
VkCommandBufferAllocateInfo ai;
ai.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO;
ai.pNext = nullptr;
ai.commandPool = handle;
ai.level = level;
ai.commandBufferCount = static_cast<u32>(num_buffers);
const VkCommandBufferAllocateInfo ai{
.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO,
.pNext = nullptr,
.commandPool = handle,
.level = level,
.commandBufferCount = static_cast<u32>(num_buffers),
};
std::unique_ptr buffers = std::make_unique<VkCommandBuffer[]>(num_buffers);
switch (const VkResult result = dld->vkAllocateCommandBuffers(owner, &ai, buffers.get())) {
@@ -497,17 +501,18 @@ std::vector<VkImage> SwapchainKHR::GetImages() const {
Device Device::Create(VkPhysicalDevice physical_device, Span<VkDeviceQueueCreateInfo> queues_ci,
Span<const char*> enabled_extensions, const void* next,
DeviceDispatch& dld) noexcept {
VkDeviceCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO;
ci.pNext = next;
ci.flags = 0;
ci.queueCreateInfoCount = queues_ci.size();
ci.pQueueCreateInfos = queues_ci.data();
ci.enabledLayerCount = 0;
ci.ppEnabledLayerNames = nullptr;
ci.enabledExtensionCount = enabled_extensions.size();
ci.ppEnabledExtensionNames = enabled_extensions.data();
ci.pEnabledFeatures = nullptr;
const VkDeviceCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO,
.pNext = next,
.flags = 0,
.queueCreateInfoCount = queues_ci.size(),
.pQueueCreateInfos = queues_ci.data(),
.enabledLayerCount = 0,
.ppEnabledLayerNames = nullptr,
.enabledExtensionCount = enabled_extensions.size(),
.ppEnabledExtensionNames = enabled_extensions.data(),
.pEnabledFeatures = nullptr,
};
VkDevice device;
if (dld.vkCreateDevice(physical_device, &ci, nullptr, &device) != VK_SUCCESS) {
@@ -548,10 +553,11 @@ ImageView Device::CreateImageView(const VkImageViewCreateInfo& ci) const {
}
Semaphore Device::CreateSemaphore() const {
VkSemaphoreCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
static constexpr VkSemaphoreCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
};
VkSemaphore object;
Check(dld->vkCreateSemaphore(handle, &ci, nullptr, &object));
@@ -639,10 +645,12 @@ ShaderModule Device::CreateShaderModule(const VkShaderModuleCreateInfo& ci) cons
}
Event Device::CreateEvent() const {
VkEventCreateInfo ci;
ci.sType = VK_STRUCTURE_TYPE_EVENT_CREATE_INFO;
ci.pNext = nullptr;
ci.flags = 0;
static constexpr VkEventCreateInfo ci{
.sType = VK_STRUCTURE_TYPE_EVENT_CREATE_INFO,
.pNext = nullptr,
.flags = 0,
};
VkEvent object;
Check(dld->vkCreateEvent(handle, &ci, nullptr, &object));
return Event(object, handle, *dld);

View File

@@ -0,0 +1,181 @@
// Copyright 2020 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <chrono>
#include <condition_variable>
#include <mutex>
#include <thread>
#include <vector>
#include "video_core/engines/maxwell_3d.h"
#include "video_core/renderer_base.h"
#include "video_core/renderer_opengl/gl_shader_cache.h"
#include "video_core/shader/async_shaders.h"
namespace VideoCommon::Shader {
AsyncShaders::AsyncShaders(Core::Frontend::EmuWindow& emu_window) : emu_window(emu_window) {}
AsyncShaders::~AsyncShaders() {
KillWorkers();
}
void AsyncShaders::AllocateWorkers(std::size_t num_workers) {
// If we're already have workers queued or don't want to queue workers, ignore
if (num_workers == worker_threads.size() || num_workers == 0) {
return;
}
// If workers already exist, clear them
if (!worker_threads.empty()) {
FreeWorkers();
}
// Create workers
for (std::size_t i = 0; i < num_workers; i++) {
context_list.push_back(emu_window.CreateSharedContext());
worker_threads.push_back(std::move(
std::thread(&AsyncShaders::ShaderCompilerThread, this, context_list[i].get())));
}
}
void AsyncShaders::FreeWorkers() {
// Mark all threads to quit
is_thread_exiting.store(true);
cv.notify_all();
for (auto& thread : worker_threads) {
thread.join();
}
// Clear our shared contexts
context_list.clear();
// Clear our worker threads
worker_threads.clear();
}
void AsyncShaders::KillWorkers() {
is_thread_exiting.store(true);
for (auto& thread : worker_threads) {
thread.detach();
}
// Clear our shared contexts
context_list.clear();
// Clear our worker threads
worker_threads.clear();
}
bool AsyncShaders::HasWorkQueued() {
return !pending_queue.empty();
}
bool AsyncShaders::HasCompletedWork() {
std::shared_lock lock{completed_mutex};
return !finished_work.empty();
}
bool AsyncShaders::IsShaderAsync(const Tegra::GPU& gpu) const {
const auto& regs = gpu.Maxwell3D().regs;
// If something is using depth, we can assume that games are not rendering anything which will
// be used one time.
if (regs.zeta_enable) {
return true;
}
// If games are using a small index count, we can assume these are full screen quads. Usually
// these shaders are only used once for building textures so we can assume they can't be built
// async
if (regs.index_array.count <= 6 || regs.vertex_buffer.count <= 6) {
return false;
}
return true;
}
std::vector<AsyncShaders::Result> AsyncShaders::GetCompletedWork() {
std::vector<AsyncShaders::Result> results;
{
std::unique_lock lock{completed_mutex};
results.assign(std::make_move_iterator(finished_work.begin()),
std::make_move_iterator(finished_work.end()));
finished_work.clear();
}
return results;
}
void AsyncShaders::QueueOpenGLShader(const OpenGL::Device& device,
Tegra::Engines::ShaderType shader_type, u64 uid,
std::vector<u64> code, std::vector<u64> code_b,
u32 main_offset,
VideoCommon::Shader::CompilerSettings compiler_settings,
const VideoCommon::Shader::Registry& registry,
VAddr cpu_addr) {
WorkerParams params{device.UseAssemblyShaders() ? AsyncShaders::Backend::GLASM
: AsyncShaders::Backend::OpenGL,
device,
shader_type,
uid,
std::move(code),
std::move(code_b),
main_offset,
compiler_settings,
registry,
cpu_addr};
std::unique_lock lock(queue_mutex);
pending_queue.push_back(std::move(params));
cv.notify_one();
}
void AsyncShaders::ShaderCompilerThread(Core::Frontend::GraphicsContext* context) {
using namespace std::chrono_literals;
while (!is_thread_exiting.load(std::memory_order_relaxed)) {
std::unique_lock lock{queue_mutex};
cv.wait(lock, [this] { return HasWorkQueued() || is_thread_exiting; });
if (is_thread_exiting) {
return;
}
// Partial lock to allow all threads to read at the same time
if (!HasWorkQueued()) {
continue;
}
// Another thread beat us, just unlock and wait for the next load
if (pending_queue.empty()) {
continue;
}
// Pull work from queue
WorkerParams work = std::move(pending_queue.front());
pending_queue.pop_front();
lock.unlock();
if (work.backend == AsyncShaders::Backend::OpenGL ||
work.backend == AsyncShaders::Backend::GLASM) {
const ShaderIR ir(work.code, work.main_offset, work.compiler_settings, work.registry);
const auto scope = context->Acquire();
auto program =
OpenGL::BuildShader(work.device, work.shader_type, work.uid, ir, work.registry);
Result result{};
result.backend = work.backend;
result.cpu_address = work.cpu_address;
result.uid = work.uid;
result.code = std::move(work.code);
result.code_b = std::move(work.code_b);
result.shader_type = work.shader_type;
if (work.backend == AsyncShaders::Backend::OpenGL) {
result.program.opengl = std::move(program->source_program);
} else if (work.backend == AsyncShaders::Backend::GLASM) {
result.program.glasm = std::move(program->assembly_program);
}
{
std::unique_lock complete_lock(completed_mutex);
finished_work.push_back(std::move(result));
}
}
}
}
} // namespace VideoCommon::Shader

View File

@@ -0,0 +1,109 @@
// Copyright 2020 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <condition_variable>
#include <deque>
#include <memory>
#include <shared_mutex>
#include <thread>
#include "common/bit_field.h"
#include "common/common_types.h"
#include "video_core/renderer_opengl/gl_device.h"
#include "video_core/renderer_opengl/gl_resource_manager.h"
#include "video_core/renderer_opengl/gl_shader_decompiler.h"
namespace Core::Frontend {
class EmuWindow;
class GraphicsContext;
} // namespace Core::Frontend
namespace Tegra {
class GPU;
}
namespace VideoCommon::Shader {
class AsyncShaders {
public:
enum class Backend {
OpenGL,
GLASM,
};
struct ResultPrograms {
OpenGL::OGLProgram opengl;
OpenGL::OGLAssemblyProgram glasm;
};
struct Result {
u64 uid;
VAddr cpu_address;
Backend backend;
ResultPrograms program;
std::vector<u64> code;
std::vector<u64> code_b;
Tegra::Engines::ShaderType shader_type;
};
explicit AsyncShaders(Core::Frontend::EmuWindow& emu_window);
~AsyncShaders();
/// Start up shader worker threads
void AllocateWorkers(std::size_t num_workers);
/// Clear the shader queue and kill all worker threads
void FreeWorkers();
// Force end all threads
void KillWorkers();
/// Check to see if any shaders have actually been compiled
bool HasCompletedWork();
/// Deduce if a shader can be build on another thread of MUST be built in sync. We cannot build
/// every shader async as some shaders are only built and executed once. We try to "guess" which
/// shader would be used only once
bool IsShaderAsync(const Tegra::GPU& gpu) const;
/// Pulls completed compiled shaders
std::vector<Result> GetCompletedWork();
void QueueOpenGLShader(const OpenGL::Device& device, Tegra::Engines::ShaderType shader_type,
u64 uid, std::vector<u64> code, std::vector<u64> code_b, u32 main_offset,
VideoCommon::Shader::CompilerSettings compiler_settings,
const VideoCommon::Shader::Registry& registry, VAddr cpu_addr);
private:
void ShaderCompilerThread(Core::Frontend::GraphicsContext* context);
/// Check our worker queue to see if we have any work queued already
bool HasWorkQueued();
struct WorkerParams {
AsyncShaders::Backend backend;
OpenGL::Device device;
Tegra::Engines::ShaderType shader_type;
u64 uid;
std::vector<u64> code;
std::vector<u64> code_b;
u32 main_offset;
VideoCommon::Shader::CompilerSettings compiler_settings;
VideoCommon::Shader::Registry registry;
VAddr cpu_address;
};
std::condition_variable cv;
std::mutex queue_mutex;
std::shared_mutex completed_mutex;
std::atomic<bool> is_thread_exiting{};
std::vector<std::unique_ptr<Core::Frontend::GraphicsContext>> context_list;
std::vector<std::thread> worker_threads;
std::deque<WorkerParams> pending_queue;
std::vector<AsyncShaders::Result> finished_work;
Core::Frontend::EmuWindow& emu_window;
};
} // namespace VideoCommon::Shader

View File

@@ -209,11 +209,11 @@ private:
}
// Remove them from the cache
const auto is_removed = [&removed_shaders](std::unique_ptr<T>& shader) {
const auto is_removed = [&removed_shaders](const std::unique_ptr<T>& shader) {
return std::find(removed_shaders.begin(), removed_shaders.end(), shader.get()) !=
removed_shaders.end();
};
storage.erase(std::remove_if(storage.begin(), storage.end(), is_removed), storage.end());
std::erase_if(storage, is_removed);
}
/// @brief Creates a new entry in the lookup cache and returns its pointer

View File

@@ -0,0 +1,42 @@
// Copyright 2020 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include "video_core/shader_notify.h"
using namespace std::chrono_literals;
namespace VideoCore {
namespace {
constexpr auto UPDATE_TICK = 32ms;
}
ShaderNotify::ShaderNotify() = default;
ShaderNotify::~ShaderNotify() = default;
std::size_t ShaderNotify::GetShadersBuilding() {
const auto now = std::chrono::high_resolution_clock::now();
const auto diff = now - last_update;
if (diff > UPDATE_TICK) {
std::shared_lock lock(mutex);
last_updated_count = accurate_count;
}
return last_updated_count;
}
std::size_t ShaderNotify::GetShadersBuildingAccurate() {
std::shared_lock lock{mutex};
return accurate_count;
}
void ShaderNotify::MarkShaderComplete() {
std::unique_lock lock{mutex};
accurate_count--;
}
void ShaderNotify::MarkSharderBuilding() {
std::unique_lock lock{mutex};
accurate_count++;
}
} // namespace VideoCore

View File

@@ -0,0 +1,29 @@
// Copyright 2020 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <chrono>
#include <shared_mutex>
#include "common/common_types.h"
namespace VideoCore {
class ShaderNotify {
public:
ShaderNotify();
~ShaderNotify();
std::size_t GetShadersBuilding();
std::size_t GetShadersBuildingAccurate();
void MarkShaderComplete();
void MarkSharderBuilding();
private:
std::size_t last_updated_count{};
std::size_t accurate_count{};
std::shared_mutex mutex;
std::chrono::high_resolution_clock::time_point last_update{};
};
} // namespace VideoCore

View File

@@ -133,11 +133,44 @@ file(GLOB COMPAT_LIST
file(GLOB_RECURSE ICONS ${PROJECT_SOURCE_DIR}/dist/icons/*)
file(GLOB_RECURSE THEMES ${PROJECT_SOURCE_DIR}/dist/qt_themes/*)
if (ENABLE_QT_TRANSLATION)
set(YUZU_QT_LANGUAGES "${PROJECT_SOURCE_DIR}/dist/languages" CACHE PATH "Path to the translation bundle for the Qt frontend")
option(GENERATE_QT_TRANSLATION "Generate en.ts as the translation source file" OFF)
# Update source TS file if enabled
if (GENERATE_QT_TRANSLATION)
get_target_property(SRCS yuzu SOURCES)
qt5_create_translation(QM_FILES ${SRCS} ${UIS} ${YUZU_QT_LANGUAGES}/en.ts)
add_custom_target(translation ALL DEPENDS ${YUZU_QT_LANGUAGES}/en.ts)
endif()
# Find all TS files except en.ts
file(GLOB_RECURSE LANGUAGES_TS ${YUZU_QT_LANGUAGES}/*.ts)
list(REMOVE_ITEM LANGUAGES_TS ${YUZU_QT_LANGUAGES}/en.ts)
# Compile TS files to QM files
qt5_add_translation(LANGUAGES_QM ${LANGUAGES_TS})
# Build a QRC file from the QM file list
set(LANGUAGES_QRC ${CMAKE_CURRENT_BINARY_DIR}/languages.qrc)
file(WRITE ${LANGUAGES_QRC} "<RCC><qresource prefix=\"languages\">\n")
foreach (QM ${LANGUAGES_QM})
get_filename_component(QM_FILE ${QM} NAME)
file(APPEND ${LANGUAGES_QRC} "<file>${QM_FILE}</file>\n")
endforeach (QM)
file(APPEND ${LANGUAGES_QRC} "</qresource></RCC>")
# Add the QRC file to package in all QM files
qt5_add_resources(LANGUAGES ${LANGUAGES_QRC})
else()
set(LANGUAGES)
endif()
target_sources(yuzu
PRIVATE
${COMPAT_LIST}
${ICONS}
${LANGUAGES}
${THEMES}
)

View File

@@ -611,6 +611,7 @@ void Config::ReadPathValues() {
}
}
UISettings::values.recent_files = ReadSetting(QStringLiteral("recentFiles")).toStringList();
UISettings::values.language = ReadSetting(QStringLiteral("language"), QString{}).toString();
qt_config->endGroup();
}
@@ -661,6 +662,8 @@ void Config::ReadRendererValues() {
ReadSettingGlobal(Settings::values.use_vsync, QStringLiteral("use_vsync"), true);
ReadSettingGlobal(Settings::values.use_assembly_shaders, QStringLiteral("use_assembly_shaders"),
false);
ReadSettingGlobal(Settings::values.use_asynchronous_shaders,
QStringLiteral("use_asynchronous_shaders"), false);
ReadSettingGlobal(Settings::values.use_fast_gpu_time, QStringLiteral("use_fast_gpu_time"),
true);
ReadSettingGlobal(Settings::values.force_30fps_mode, QStringLiteral("force_30fps_mode"), false);
@@ -1093,6 +1096,7 @@ void Config::SavePathValues() {
}
qt_config->endArray();
WriteSetting(QStringLiteral("recentFiles"), UISettings::values.recent_files);
WriteSetting(QStringLiteral("language"), UISettings::values.language, QString{});
qt_config->endGroup();
}
@@ -1145,6 +1149,8 @@ void Config::SaveRendererValues() {
WriteSettingGlobal(QStringLiteral("use_vsync"), Settings::values.use_vsync, true);
WriteSettingGlobal(QStringLiteral("use_assembly_shaders"),
Settings::values.use_assembly_shaders, false);
WriteSettingGlobal(QStringLiteral("use_asynchronous_shaders"),
Settings::values.use_asynchronous_shaders, false);
WriteSettingGlobal(QStringLiteral("use_fast_gpu_time"), Settings::values.use_fast_gpu_time,
true);
WriteSettingGlobal(QStringLiteral("force_30fps_mode"), Settings::values.force_30fps_mode,

View File

@@ -23,6 +23,7 @@ ConfigureDialog::ConfigureDialog(QWidget* parent, HotkeyRegistry& registry)
SetConfiguration();
PopulateSelectionList();
connect(ui->uiTab, &ConfigureUi::LanguageChanged, this, &ConfigureDialog::OnLanguageChanged);
connect(ui->selectorList, &QListWidget::itemSelectionChanged, this,
&ConfigureDialog::UpdateVisibleTabs);
@@ -98,6 +99,14 @@ void ConfigureDialog::PopulateSelectionList() {
}
}
void ConfigureDialog::OnLanguageChanged(const QString& locale) {
emit LanguageChanged(locale);
// first apply the configuration, and then restore the display
ApplyConfiguration();
RetranslateUI();
SetConfiguration();
}
void ConfigureDialog::UpdateVisibleTabs() {
const auto items = ui->selectorList->selectedItems();
if (items.isEmpty()) {

View File

@@ -22,6 +22,12 @@ public:
void ApplyConfiguration();
private slots:
void OnLanguageChanged(const QString& locale);
signals:
void LanguageChanged(const QString& locale);
private:
void changeEvent(QEvent* event) override;

View File

@@ -24,6 +24,7 @@ void ConfigureGraphicsAdvanced::SetConfiguration() {
const bool runtime_lock = !Core::System::GetInstance().IsPoweredOn();
ui->use_vsync->setEnabled(runtime_lock);
ui->use_assembly_shaders->setEnabled(runtime_lock);
ui->use_asynchronous_shaders->setEnabled(runtime_lock);
ui->force_30fps_mode->setEnabled(runtime_lock);
ui->anisotropic_filtering_combobox->setEnabled(runtime_lock);
@@ -32,6 +33,8 @@ void ConfigureGraphicsAdvanced::SetConfiguration() {
static_cast<int>(Settings::values.gpu_accuracy.GetValue()));
ui->use_vsync->setChecked(Settings::values.use_vsync.GetValue());
ui->use_assembly_shaders->setChecked(Settings::values.use_assembly_shaders.GetValue());
ui->use_asynchronous_shaders->setChecked(
Settings::values.use_asynchronous_shaders.GetValue());
ui->use_fast_gpu_time->setChecked(Settings::values.use_fast_gpu_time.GetValue());
ui->force_30fps_mode->setChecked(Settings::values.force_30fps_mode.GetValue());
ui->anisotropic_filtering_combobox->setCurrentIndex(
@@ -41,6 +44,10 @@ void ConfigureGraphicsAdvanced::SetConfiguration() {
ConfigurationShared::SetPerGameSetting(ui->use_vsync, &Settings::values.use_vsync);
ConfigurationShared::SetPerGameSetting(ui->use_assembly_shaders,
&Settings::values.use_assembly_shaders);
ConfigurationShared::SetPerGameSetting(ui->use_asynchronous_shaders,
&Settings::values.use_asynchronous_shaders);
ConfigurationShared::SetPerGameSetting(ui->use_asynchronous_shaders,
&Settings::values.use_asynchronous_shaders);
ConfigurationShared::SetPerGameSetting(ui->use_fast_gpu_time,
&Settings::values.use_fast_gpu_time);
ConfigurationShared::SetPerGameSetting(ui->force_30fps_mode,
@@ -67,6 +74,14 @@ void ConfigureGraphicsAdvanced::ApplyConfiguration() {
if (Settings::values.use_assembly_shaders.UsingGlobal()) {
Settings::values.use_assembly_shaders.SetValue(ui->use_assembly_shaders->isChecked());
}
if (Settings::values.use_asynchronous_shaders.UsingGlobal()) {
Settings::values.use_asynchronous_shaders.SetValue(
ui->use_asynchronous_shaders->isChecked());
}
if (Settings::values.use_asynchronous_shaders.UsingGlobal()) {
Settings::values.use_asynchronous_shaders.SetValue(
ui->use_asynchronous_shaders->isChecked());
}
if (Settings::values.use_fast_gpu_time.UsingGlobal()) {
Settings::values.use_fast_gpu_time.SetValue(ui->use_fast_gpu_time->isChecked());
}
@@ -83,6 +98,10 @@ void ConfigureGraphicsAdvanced::ApplyConfiguration() {
ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_vsync, ui->use_vsync);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_assembly_shaders,
ui->use_assembly_shaders);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_asynchronous_shaders,
ui->use_asynchronous_shaders);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_asynchronous_shaders,
ui->use_asynchronous_shaders);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_fast_gpu_time,
ui->use_fast_gpu_time);
ConfigurationShared::ApplyPerGameSetting(&Settings::values.force_30fps_mode,
@@ -117,6 +136,8 @@ void ConfigureGraphicsAdvanced::SetupPerGameUI() {
ui->gpu_accuracy->setEnabled(Settings::values.gpu_accuracy.UsingGlobal());
ui->use_vsync->setEnabled(Settings::values.use_vsync.UsingGlobal());
ui->use_assembly_shaders->setEnabled(Settings::values.use_assembly_shaders.UsingGlobal());
ui->use_asynchronous_shaders->setEnabled(
Settings::values.use_asynchronous_shaders.UsingGlobal());
ui->use_fast_gpu_time->setEnabled(Settings::values.use_fast_gpu_time.UsingGlobal());
ui->force_30fps_mode->setEnabled(Settings::values.force_30fps_mode.UsingGlobal());
ui->anisotropic_filtering_combobox->setEnabled(
@@ -128,6 +149,7 @@ void ConfigureGraphicsAdvanced::SetupPerGameUI() {
ConfigurationShared::InsertGlobalItem(ui->gpu_accuracy);
ui->use_vsync->setTristate(true);
ui->use_assembly_shaders->setTristate(true);
ui->use_asynchronous_shaders->setTristate(true);
ui->use_fast_gpu_time->setTristate(true);
ui->force_30fps_mode->setTristate(true);
ConfigurationShared::InsertGlobalItem(ui->anisotropic_filtering_combobox);

View File

@@ -72,6 +72,16 @@
</property>
</widget>
</item>
<item>
<widget class="QCheckBox" name="use_asynchronous_shaders">
<property name="toolTip">
<string>Enables asynchronous shader compilation, which may reduce shader stutter. This feature is experimental.</string>
</property>
<property name="text">
<string>Use asynchronous shader building (experimental, OpenGL or Assembly shaders only)</string>
</property>
</widget>
</item>
<item>
<widget class="QCheckBox" name="force_30fps_mode">
<property name="text">

View File

@@ -5,6 +5,7 @@
#include <array>
#include <utility>
#include <QDirIterator>
#include "common/common_types.h"
#include "core/settings.h"
#include "ui_configure_ui.h"
@@ -29,6 +30,8 @@ constexpr std::array row_text_names{
ConfigureUi::ConfigureUi(QWidget* parent) : QWidget(parent), ui(new Ui::ConfigureUi) {
ui->setupUi(this);
InitializeLanguageComboBox();
for (const auto& theme : UISettings::themes) {
ui->theme_combobox->addItem(QString::fromUtf8(theme.first),
QString::fromUtf8(theme.second));
@@ -72,6 +75,8 @@ void ConfigureUi::RequestGameListUpdate() {
void ConfigureUi::SetConfiguration() {
ui->theme_combobox->setCurrentIndex(ui->theme_combobox->findData(UISettings::values.theme));
ui->language_combobox->setCurrentIndex(
ui->language_combobox->findData(UISettings::values.language));
ui->show_add_ons->setChecked(UISettings::values.show_add_ons);
ui->icon_size_combobox->setCurrentIndex(
ui->icon_size_combobox->findData(UISettings::values.icon_size));
@@ -100,6 +105,25 @@ void ConfigureUi::RetranslateUI() {
}
}
void ConfigureUi::InitializeLanguageComboBox() {
ui->language_combobox->addItem(tr("<System>"), QString{});
ui->language_combobox->addItem(tr("English"), QStringLiteral("en"));
QDirIterator it(QStringLiteral(":/languages"), QDirIterator::NoIteratorFlags);
while (it.hasNext()) {
QString locale = it.next();
locale.truncate(locale.lastIndexOf(QLatin1Char{'.'}));
locale.remove(0, locale.lastIndexOf(QLatin1Char{'/'}) + 1);
const QString lang = QLocale::languageToString(QLocale(locale).language());
ui->language_combobox->addItem(lang, locale);
}
// Unlike other configuration changes, interface language changes need to be reflected on the
// interface immediately. This is done by passing a signal to the main window, and then
// retranslating when passing back.
connect(ui->language_combobox, QOverload<int>::of(&QComboBox::currentIndexChanged), this,
&ConfigureUi::OnLanguageChanged);
}
void ConfigureUi::InitializeIconSizeComboBox() {
for (const auto& size : default_icon_sizes) {
ui->icon_size_combobox->addItem(QString::fromUtf8(size.second), size.first);
@@ -147,3 +171,10 @@ void ConfigureUi::UpdateSecondRowComboBox(bool init) {
ui->row_2_text_combobox->removeItem(
ui->row_2_text_combobox->findData(ui->row_1_text_combobox->currentData()));
}
void ConfigureUi::OnLanguageChanged(int index) {
if (index == -1)
return;
emit LanguageChanged(ui->language_combobox->itemData(index).toString());
}

View File

@@ -20,6 +20,12 @@ public:
void ApplyConfiguration();
private slots:
void OnLanguageChanged(int index);
signals:
void LanguageChanged(const QString& locale);
private:
void RequestGameListUpdate();
@@ -28,6 +34,7 @@ private:
void changeEvent(QEvent*) override;
void RetranslateUI();
void InitializeLanguageComboBox();
void InitializeIconSizeComboBox();
void InitializeRowComboBoxes();

View File

@@ -13,112 +13,132 @@
<property name="windowTitle">
<string>Form</string>
</property>
<layout class="QHBoxLayout" name="HorizontalLayout">
<layout class="QVBoxLayout" name="verticalLayout">
<item>
<layout class="QVBoxLayout" name="VerticalLayout">
<item>
<widget class="QGroupBox" name="GeneralGroupBox">
<property name="title">
<string>General</string>
</property>
<layout class="QHBoxLayout" name="horizontalLayout">
<widget class="QGroupBox" name="general_groupBox">
<property name="title">
<string>General</string>
</property>
<layout class="QHBoxLayout" name="horizontalLayout">
<item>
<layout class="QVBoxLayout" name="verticalLayout_2">
<item>
<layout class="QVBoxLayout" name="verticalLayout">
<item>
<layout class="QHBoxLayout" name="horizontalLayout_3">
<item>
<widget class="QLabel" name="theme_label">
<property name="text">
<string>Theme:</string>
</property>
</widget>
</item>
<item>
<widget class="QComboBox" name="theme_combobox"/>
</item>
</layout>
</item>
</layout>
<widget class="QLabel" name="label_change_language_info">
<property name="text">
<string>Note: Changing language will apply your configuration.</string>
</property>
<property name="wordWrap">
<bool>true</bool>
</property>
</widget>
</item>
</layout>
</widget>
</item>
<item>
<widget class="QGroupBox" name="GameListGroupBox">
<property name="title">
<string>Game List</string>
</property>
<layout class="QHBoxLayout" name="GameListHorizontalLayout">
<item>
<layout class="QVBoxLayout" name="GeneralVerticalLayout">
<layout class="QHBoxLayout" name="horizontalLayout_2">
<item>
<widget class="QCheckBox" name="show_add_ons">
<widget class="QLabel" name="language_label">
<property name="text">
<string>Show Add-Ons Column</string>
<string>Interface language:</string>
</property>
</widget>
</item>
<item>
<layout class="QHBoxLayout" name="icon_size_qhbox_layout_2">
<item>
<widget class="QLabel" name="icon_size_label">
<property name="text">
<string>Icon Size:</string>
</property>
</widget>
</item>
<item>
<widget class="QComboBox" name="icon_size_combobox"/>
</item>
</layout>
<widget class="QComboBox" name="language_combobox"/>
</item>
</layout>
</item>
<item>
<layout class="QHBoxLayout" name="horizontalLayout_3">
<item>
<widget class="QLabel" name="theme_label">
<property name="text">
<string>Theme:</string>
</property>
</widget>
</item>
<item>
<layout class="QHBoxLayout" name="row_1_qhbox_layout">
<item>
<widget class="QLabel" name="row_1_label">
<property name="text">
<string>Row 1 Text:</string>
</property>
</widget>
</item>
<item>
<widget class="QComboBox" name="row_1_text_combobox"/>
</item>
</layout>
</item>
<item>
<layout class="QHBoxLayout" name="row_2_qhbox_layout">
<item>
<widget class="QLabel" name="row_2_label">
<property name="text">
<string>Row 2 Text:</string>
</property>
</widget>
</item>
<item>
<widget class="QComboBox" name="row_2_text_combobox"/>
</item>
</layout>
<widget class="QComboBox" name="theme_combobox"/>
</item>
</layout>
</item>
</layout>
</widget>
</item>
<item>
<spacer name="verticalSpacer">
<property name="orientation">
<enum>Qt::Vertical</enum>
</property>
<property name="sizeHint" stdset="0">
<size>
<width>20</width>
<height>40</height>
</size>
</property>
</spacer>
</item>
</layout>
</item>
</layout>
</widget>
</item>
<item>
<widget class="QGroupBox" name="GameListGroupBox">
<property name="title">
<string>Game List</string>
</property>
<layout class="QHBoxLayout" name="GameListHorizontalLayout">
<item>
<layout class="QVBoxLayout" name="GeneralVerticalLayout">
<item>
<widget class="QCheckBox" name="show_add_ons">
<property name="text">
<string>Show Add-Ons Column</string>
</property>
</widget>
</item>
<item>
<layout class="QHBoxLayout" name="icon_size_qhbox_layout_2">
<item>
<widget class="QLabel" name="icon_size_label">
<property name="text">
<string>Icon Size:</string>
</property>
</widget>
</item>
<item>
<widget class="QComboBox" name="icon_size_combobox"/>
</item>
</layout>
</item>
<item>
<layout class="QHBoxLayout" name="row_1_qhbox_layout">
<item>
<widget class="QLabel" name="row_1_label">
<property name="text">
<string>Row 1 Text:</string>
</property>
</widget>
</item>
<item>
<widget class="QComboBox" name="row_1_text_combobox"/>
</item>
</layout>
</item>
<item>
<layout class="QHBoxLayout" name="row_2_qhbox_layout">
<item>
<widget class="QLabel" name="row_2_label">
<property name="text">
<string>Row 2 Text:</string>
</property>
</widget>
</item>
<item>
<widget class="QComboBox" name="row_2_text_combobox"/>
</item>
</layout>
</item>
</layout>
</item>
</layout>
</widget>
</item>
<item>
<spacer name="verticalSpacer">
<property name="orientation">
<enum>Qt::Vertical</enum>
</property>
<property name="sizeHint" stdset="0">
<size>
<width>20</width>
<height>40</height>
</size>
</property>
</spacer>
</item>
</layout>
</widget>

View File

@@ -2,9 +2,11 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <array>
#include <fmt/format.h>
#include "yuzu/debugger/wait_tree.h"
#include "yuzu/uisettings.h"
#include "yuzu/util/util.h"
#include "common/assert.h"
@@ -19,11 +21,37 @@
#include "core/hle/kernel/thread.h"
#include "core/memory.h"
namespace {
constexpr std::array<std::array<Qt::GlobalColor, 2>, 10> WaitTreeColors{{
{Qt::GlobalColor::darkGreen, Qt::GlobalColor::green},
{Qt::GlobalColor::darkGreen, Qt::GlobalColor::green},
{Qt::GlobalColor::darkBlue, Qt::GlobalColor::cyan},
{Qt::GlobalColor::lightGray, Qt::GlobalColor::lightGray},
{Qt::GlobalColor::lightGray, Qt::GlobalColor::lightGray},
{Qt::GlobalColor::darkRed, Qt::GlobalColor::red},
{Qt::GlobalColor::darkYellow, Qt::GlobalColor::yellow},
{Qt::GlobalColor::red, Qt::GlobalColor::red},
{Qt::GlobalColor::darkCyan, Qt::GlobalColor::cyan},
{Qt::GlobalColor::gray, Qt::GlobalColor::gray},
}};
bool IsDarkTheme() {
const auto& theme = UISettings::values.theme;
return theme == QStringLiteral("qdarkstyle") || theme == QStringLiteral("colorful_dark");
}
} // namespace
WaitTreeItem::WaitTreeItem() = default;
WaitTreeItem::~WaitTreeItem() = default;
QColor WaitTreeItem::GetColor() const {
return QColor(Qt::GlobalColor::black);
if (IsDarkTheme()) {
return QColor(Qt::GlobalColor::white);
} else {
return QColor(Qt::GlobalColor::black);
}
}
std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeItem::GetChildren() const {
@@ -263,36 +291,38 @@ QString WaitTreeThread::GetText() const {
}
QColor WaitTreeThread::GetColor() const {
const std::size_t color_index = IsDarkTheme() ? 1 : 0;
const auto& thread = static_cast<const Kernel::Thread&>(object);
switch (thread.GetStatus()) {
case Kernel::ThreadStatus::Running:
return QColor(Qt::GlobalColor::darkGreen);
return QColor(WaitTreeColors[0][color_index]);
case Kernel::ThreadStatus::Ready:
if (!thread.IsPaused()) {
if (thread.WasRunning()) {
return QColor(Qt::GlobalColor::darkGreen);
return QColor(WaitTreeColors[1][color_index]);
} else {
return QColor(Qt::GlobalColor::darkBlue);
return QColor(WaitTreeColors[2][color_index]);
}
} else {
return QColor(Qt::GlobalColor::lightGray);
return QColor(WaitTreeColors[3][color_index]);
}
case Kernel::ThreadStatus::Paused:
return QColor(Qt::GlobalColor::lightGray);
return QColor(WaitTreeColors[4][color_index]);
case Kernel::ThreadStatus::WaitHLEEvent:
case Kernel::ThreadStatus::WaitIPC:
return QColor(Qt::GlobalColor::darkRed);
return QColor(WaitTreeColors[5][color_index]);
case Kernel::ThreadStatus::WaitSleep:
return QColor(Qt::GlobalColor::darkYellow);
return QColor(WaitTreeColors[6][color_index]);
case Kernel::ThreadStatus::WaitSynch:
case Kernel::ThreadStatus::WaitMutex:
case Kernel::ThreadStatus::WaitCondVar:
case Kernel::ThreadStatus::WaitArb:
return QColor(Qt::GlobalColor::red);
return QColor(WaitTreeColors[7][color_index]);
case Kernel::ThreadStatus::Dormant:
return QColor(Qt::GlobalColor::darkCyan);
return QColor(WaitTreeColors[8][color_index]);
case Kernel::ThreadStatus::Dead:
return QColor(Qt::GlobalColor::gray);
return QColor(WaitTreeColors[9][color_index]);
default:
return WaitTreeItem::GetColor();
}

View File

@@ -94,6 +94,8 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual
#include "core/perf_stats.h"
#include "core/settings.h"
#include "core/telemetry_session.h"
#include "video_core/gpu.h"
#include "video_core/shader_notify.h"
#include "yuzu/about_dialog.h"
#include "yuzu/bootmanager.h"
#include "yuzu/compatdb.h"
@@ -189,6 +191,8 @@ GMainWindow::GMainWindow()
provider(std::make_unique<FileSys::ManualContentProvider>()) {
InitializeLogging();
LoadTranslation();
setAcceptDrops(true);
ui.setupUi(this);
statusBar()->hide();
@@ -498,6 +502,8 @@ void GMainWindow::InitializeWidgets() {
message_label->setAlignment(Qt::AlignLeft);
statusBar()->addPermanentWidget(message_label, 1);
shader_building_label = new QLabel();
shader_building_label->setToolTip(tr("The amount of shaders currently being built"));
emu_speed_label = new QLabel();
emu_speed_label->setToolTip(
tr("Current emulation speed. Values higher or lower than 100% "
@@ -510,7 +516,8 @@ void GMainWindow::InitializeWidgets() {
tr("Time taken to emulate a Switch frame, not counting framelimiting or v-sync. For "
"full-speed emulation this should be at most 16.67 ms."));
for (auto& label : {emu_speed_label, game_fps_label, emu_frametime_label}) {
for (auto& label :
{shader_building_label, emu_speed_label, game_fps_label, emu_frametime_label}) {
label->setVisible(false);
label->setFrameStyle(QFrame::NoFrame);
label->setContentsMargins(4, 0, 4, 0);
@@ -1176,6 +1183,7 @@ void GMainWindow::ShutdownGame() {
// Disable status bar updates
status_bar_update_timer.stop();
shader_building_label->setVisible(false);
emu_speed_label->setVisible(false);
game_fps_label->setVisible(false);
emu_frametime_label->setVisible(false);
@@ -2042,6 +2050,9 @@ void GMainWindow::OnConfigure() {
const bool old_discord_presence = UISettings::values.enable_discord_presence;
ConfigureDialog configure_dialog(this, hotkey_registry);
connect(&configure_dialog, &ConfigureDialog::LanguageChanged, this,
&GMainWindow::OnLanguageChanged);
const auto result = configure_dialog.exec();
if (result != QDialog::Accepted) {
return;
@@ -2186,6 +2197,17 @@ void GMainWindow::UpdateStatusBar() {
}
auto results = Core::System::GetInstance().GetAndResetPerfStats();
auto& shader_notify = Core::System::GetInstance().GPU().ShaderNotify();
const auto shaders_building = shader_notify.GetShadersBuilding();
if (shaders_building != 0) {
shader_building_label->setText(
tr("Building: %1 shader").arg(shaders_building) +
(shaders_building != 1 ? QString::fromStdString("s") : QString::fromStdString("")));
shader_building_label->setVisible(true);
} else {
shader_building_label->setVisible(false);
}
if (Settings::values.use_frame_limit.GetValue()) {
emu_speed_label->setText(tr("Speed: %1% / %2%")
@@ -2315,9 +2337,12 @@ void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) {
if (behavior == ReinitializeKeyBehavior::Warning) {
const auto res = QMessageBox::information(
this, tr("Confirm Key Rederivation"),
tr("You are about to force rederive all of your keys. \nIf you do not know what this "
"means or what you are doing, \nthis is a potentially destructive action. \nPlease "
"make sure this is what you want \nand optionally make backups.\n\nThis will delete "
tr("You are about to force rederive all of your keys. \nIf you do not know what "
"this "
"means or what you are doing, \nthis is a potentially destructive action. "
"\nPlease "
"make sure this is what you want \nand optionally make backups.\n\nThis will "
"delete "
"your autogenerated key files and re-run the key derivation module."),
QMessageBox::StandardButtons{QMessageBox::Ok, QMessageBox::Cancel});
@@ -2600,6 +2625,43 @@ void GMainWindow::UpdateUITheme() {
QIcon::setThemeSearchPaths(theme_paths);
}
void GMainWindow::LoadTranslation() {
// If the selected language is English, no need to install any translation
if (UISettings::values.language == QStringLiteral("en")) {
return;
}
bool loaded;
if (UISettings::values.language.isEmpty()) {
// If the selected language is empty, use system locale
loaded = translator.load(QLocale(), {}, {}, QStringLiteral(":/languages/"));
} else {
// Otherwise load from the specified file
loaded = translator.load(UISettings::values.language, QStringLiteral(":/languages/"));
}
if (loaded) {
qApp->installTranslator(&translator);
} else {
UISettings::values.language = QStringLiteral("en");
}
}
void GMainWindow::OnLanguageChanged(const QString& locale) {
if (UISettings::values.language != QStringLiteral("en")) {
qApp->removeTranslator(&translator);
}
UISettings::values.language = locale;
LoadTranslation();
ui.retranslateUi(this);
UpdateWindowTitle();
if (emulation_running)
ui.action_Start->setText(tr("Continue"));
}
void GMainWindow::SetDiscordEnabled([[maybe_unused]] bool state) {
#ifdef USE_DISCORD_PRESENCE
if (state) {
@@ -2628,8 +2690,8 @@ int main(int argc, char* argv[]) {
#ifdef __APPLE__
// If you start a bundle (binary) on OSX without the Terminal, the working directory is "/".
// But since we require the working directory to be the executable path for the location of the
// user folder in the Qt Frontend, we need to cd into that working directory
// But since we require the working directory to be the executable path for the location of
// the user folder in the Qt Frontend, we need to cd into that working directory
const std::string bin_path = FileUtil::GetBundleDirectory() + DIR_SEP + "..";
chdir(bin_path.c_str());
#endif

View File

@@ -10,6 +10,7 @@
#include <QMainWindow>
#include <QTimer>
#include <QTranslator>
#include "common/common_types.h"
#include "core/core.h"
@@ -225,6 +226,7 @@ private slots:
void OnCaptureScreenshot();
void OnCoreError(Core::System::ResultStatus, std::string);
void OnReinitializeKeys(ReinitializeKeyBehavior behavior);
void OnLanguageChanged(const QString& locale);
private:
std::optional<u64> SelectRomFSDumpTarget(const FileSys::ContentProvider&, u64 program_id);
@@ -237,6 +239,7 @@ private:
void HideMouseCursor();
void ShowMouseCursor();
void OpenURL(const QUrl& url);
void LoadTranslation();
Ui::MainWindow ui;
@@ -248,6 +251,7 @@ private:
// Status bar elements
QLabel* message_label = nullptr;
QLabel* shader_building_label = nullptr;
QLabel* emu_speed_label = nullptr;
QLabel* game_fps_label = nullptr;
QLabel* emu_frametime_label = nullptr;
@@ -284,6 +288,8 @@ private:
HotkeyRegistry hotkey_registry;
QTranslator translator;
// Install progress dialog
QProgressDialog* install_progress;

View File

@@ -75,6 +75,7 @@ struct Values {
bool game_dir_deprecated_deepscan;
QVector<UISettings::GameDir> game_dirs;
QStringList recent_files;
QString language;
QString theme;

View File

@@ -394,6 +394,10 @@ void Config::ReadValues() {
static_cast<u16>(sdl2_config->GetInteger("Renderer", "use_vsync", 1)));
Settings::values.use_assembly_shaders.SetValue(
sdl2_config->GetBoolean("Renderer", "use_assembly_shaders", false));
Settings::values.use_asynchronous_shaders.SetValue(
sdl2_config->GetBoolean("Renderer", "use_asynchronous_shaders", false));
Settings::values.use_asynchronous_shaders.SetValue(
sdl2_config->GetBoolean("Renderer", "use_asynchronous_shaders", false));
Settings::values.use_fast_gpu_time.SetValue(
sdl2_config->GetBoolean("Renderer", "use_fast_gpu_time", true));

View File

@@ -166,6 +166,10 @@ use_vsync =
# 0 (default): Off, 1: On
use_assembly_shaders =
# Whether to allow asynchronous shader building.
# 0 (default): Off, 1: On
use_asynchronous_shaders =
# Turns on the frame limiter, which will limit frames output to the target game speed
# 0: Off, 1: On (default)
use_frame_limit =