Compare commits

...

36 Commits

Author SHA1 Message Date
Morph 2c9308954c hid: Stub a series of "SevenSixAxisSensor" service commands
- Used by Captain Toad: Treasure Tracker Update 1.3.0

While we're at it, fix the input parameters for SetIsPalmaAllConnectable and SetPalmaBoostMode
2020-06-24 11:57:39 -04:00
Rodrigo Locatti 406d298457 Merge pull request #4110 from ReinUsesLisp/direct-upload-sets
vk_update_descriptor: Upload descriptor sets data directly
2020-06-22 05:02:13 -03:00
bunnei 14a1181a97 Merge pull request #4122 from lioncash/hide
video_core: Eliminate some variable shadowing
2020-06-21 22:38:04 -04:00
bunnei c27c76ed43 Merge pull request #4126 from lioncash/noexcept
vulkan/wrapper: Remove noexcept from GetSurfaceCapabilitiesKHR()
2020-06-21 22:36:14 -04:00
bunnei e8855ed0fc Merge pull request #4134 from FearlessTobi/port-5322
Port citra-emu/citra#5322: "Fix: fatal error CVT1100 when compiling manifest file"
2020-06-21 22:35:17 -04:00
bunnei f98bf1025f Merge pull request #4120 from lioncash/arb
gl_arb_decompiler: Avoid several string copies
2020-06-20 22:11:49 -04:00
FearlessTobi a8674a7b86 Fix: fatal error CVT1100 when compiling manifest file
Occurs when doing a local compile in MSVC build. The compiler I'm using is as below:
Microsoft Visual Studio Community 2019 Preview
Version 16.6.0 Preview 5.0

Fixes this error:
CVTRES : fatal error CVT1100: duplicate resource. type:MANIFEST, name:1, language:0x0409
LINK : fatal error LNK1123: failure during conversion to COFF: file invalid or corrupt

I have put 0 since previous name was 1. If have other names in mind, please let me know.

Co-Authored-By: dragios <dragios@users.noreply.github.com>
2020-06-21 03:11:23 +02:00
LC c6ba7a228d Merge pull request #4133 from MerryMage/macrojit-shifts
macro_jit_x64: Use ecx for shift register
2020-06-20 19:58:51 -04:00
MerryMage c12eb814b4 macro_jit_x64: Use ecx for shift register
shl/shr only accept cl as their second argument
2020-06-20 22:24:05 +01:00
merry 928e9c09aa Merge pull request #4125 from lioncash/macro-shift
macro_jit_x64: Amend readability of Compile_ExtractShiftLeftRegister()
2020-06-20 16:08:23 +01:00
merry 2bd903e021 Merge pull request #4123 from lioncash/unused-var
macro_jit_x64: Remove unused variable
2020-06-20 16:07:58 +01:00
bunnei 7d1dca4c98 Merge pull request #4099 from MerryMage/macOS-build
Fix compilation on macOS
2020-06-19 23:31:04 -04:00
Lioncash 5865a10885 gl_arb_decompiler: Avoid several string copies
Variables that are marked as const cannot have the move constructor
invoked when returning from a function (the move constructor requires a
non-const variable so it can "steal" the resources from it.
2020-06-19 23:09:16 -04:00
Lioncash a6e5b84d1f vulkan/wrapper: Remove noexcept from GetSurfaceCapabilitiesKHR()
Check() can throw an exception if the Vulkan result isn't successful.

We remove the check so that std::terminate isn't outright called and
allows for better debugging (should it ever actually fail).
2020-06-19 23:01:59 -04:00
Lioncash 5a4e89b901 macro_jit_x64: Correct readability of Compile_ExtractShiftLeftImmediate()
Previously dst wasn't being used.
2020-06-19 22:57:23 -04:00
Lioncash 140f953b6a macro_jit_x64: Correct readability of Compile_ExtractShiftLeftRegister()
Previously dst wasn't being used.
2020-06-19 22:56:55 -04:00
Lioncash 8ea749c1ca macro_jit_x64: Remove unused variable
Removes a completely unused label and marks another variable as unused,
given it seems like it has potential uses in the future.
2020-06-19 22:10:45 -04:00
Lioncash 479605b3e5 memory_manager: Eliminate variable shadowing
Renames some variables to prevent ones in inner scopes from shadowing
outer-scoped variables.

The Copy* functions have no shadowing, but we rename them anyways to
remain consistent with the other functions.
2020-06-19 22:02:58 -04:00
bunnei 9c5ed4408d Merge pull request #4113 from ogniK5377/boxcat-disable
Fix compilation when not building with boxcat
2020-06-19 21:59:59 -04:00
Lioncash 811bff009e macro_jit_x64: Eliminate variable shadowing in Compile_ProcessResult()
We can reduce the capture scope so that it's not possible for both "reg"
variables to clash with one another.

While we're at it, we can prevent unnecessary copies while we're at it.
2020-06-19 21:57:44 -04:00
Lioncash 4514b80b3e buffer_cache: Eliminate local variable shadowing
We can just make use of the instance in the scope above this one.
2020-06-19 21:55:02 -04:00
bunnei 7daea551c0 Merge pull request #4087 from MerryMage/macrojit-inline-Read
macro_jit_x64: Inline Engines::Maxwell3D::GetRegisterValue
2020-06-19 21:32:07 -04:00
LC 8434630dcc Merge pull request #4114 from MerryMage/nrvo
Remove redundant moves
2020-06-19 15:17:04 -04:00
MerryMage c6a963c48e input_common/motion_emu: Remove redundant move
Named return value optimization automatically applies here.
2020-06-19 14:29:59 +01:00
MerryMage 8272f53cf9 input_common/keyboard: Remove redundant move
Named return value optimization automatically applies here.
2020-06-19 14:29:36 +01:00
MerryMage 7236393114 mii_model: Remove redundant std::move
Named return value optimization automatically applies here.
2020-06-19 14:29:09 +01:00
MerryMage 977ceb4056 macro_jit_x64: Remove unused function Read 2020-06-19 11:39:41 +01:00
ReinUsesLisp 7d763f060e vk_update_descriptor: Upload descriptor sets data directly
Instead of copying to a temporary payload before sending the update task
to the worker thread, insert elements to the payload directly.
2020-06-18 17:47:19 -03:00
MerryMage 778f86989a bootmanager: Remove references to OpenGL for macOS
OpenGL macOS headers definitions clash heavily with each other
2020-06-18 15:47:44 +01:00
MerryMage b19fe55f84 memory_manager: Explicitly specifcy std::min<size_t> 2020-06-18 15:47:44 +01:00
MerryMage 4f09f0aea4 shared_font: Service::NS::EncryptSharedFont takes a size_t& 2020-06-18 15:47:44 +01:00
MerryMage 69f38355ed vk_rasterizer: BindTransformFeedbackBuffersEXT accepts a size of type VkDeviceSize 2020-06-18 15:47:44 +01:00
MerryMage b1eada6079 renderer_vulkan: Fix macOS GetBundleDirectory reference 2020-06-18 15:47:44 +01:00
MerryMage 442e48ef4c memory_util: boost hashes are size_t
* boost::hash_value returns a size_t
* boost::hash_combine takes a size_t& argument
2020-06-18 15:47:43 +01:00
MerryMage 8ae7154541 Rename PAGE_SHIFT to PAGE_BITS
macOS header files #define PAGE_SHIFT
2020-06-18 15:47:43 +01:00
MerryMage 44f10d9b9f macro_jit_x64: Inline Engines::Maxwell3D::GetRegisterValue 2020-06-17 17:17:08 +01:00
27 changed files with 253 additions and 173 deletions
@@ -40,7 +40,7 @@ VirtualDir MiiModel() {
out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::SHAPE_MID.size()>>(
MiiModelData::SHAPE_MID, "ShapeMid.dat"));
return std::move(out);
return out;
}
} // namespace FileSys::SystemArchive
@@ -23,7 +23,7 @@ VirtualFile PackBFTTF(const std::array<u8, Size>& data, const std::string& name)
std::vector<u8> bfttf(Size + sizeof(u64));
u64 offset = 0;
size_t offset = 0;
Service::NS::EncryptSharedFont(vec, bfttf, offset);
return std::make_shared<VectorVfsFile>(std::move(bfttf), name);
}
@@ -104,7 +104,7 @@ ResultCode MemoryManager::Allocate(PageLinkedList& page_list, std::size_t num_pa
// Ensure that we don't leave anything un-freed
auto group_guard = detail::ScopeExit([&] {
for (const auto& it : page_list.Nodes()) {
const auto min_num_pages{std::min(
const auto min_num_pages{std::min<size_t>(
it.GetNumPages(), (chosen_manager.GetEndAddress() - it.GetAddress()) / PageSize)};
chosen_manager.Free(it.GetAddress(), min_num_pages);
}
@@ -165,7 +165,7 @@ ResultCode MemoryManager::Free(PageLinkedList& page_list, std::size_t num_pages,
// Free all of the pages
for (const auto& it : page_list.Nodes()) {
const auto min_num_pages{std::min(
const auto min_num_pages{std::min<size_t>(
it.GetNumPages(), (chosen_manager.GetEndAddress() - it.GetAddress()) / PageSize)};
chosen_manager.Free(it.GetAddress(), min_num_pages);
}
+76 -17
View File
@@ -230,15 +230,15 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) {
{211, nullptr, "IsVibrationDeviceMounted"},
{300, &Hid::ActivateConsoleSixAxisSensor, "ActivateConsoleSixAxisSensor"},
{301, &Hid::StartConsoleSixAxisSensor, "StartConsoleSixAxisSensor"},
{302, nullptr, "StopConsoleSixAxisSensor"},
{303, nullptr, "ActivateSevenSixAxisSensor"},
{304, nullptr, "StartSevenSixAxisSensor"},
{302, &Hid::StopConsoleSixAxisSensor, "StopConsoleSixAxisSensor"},
{303, &Hid::ActivateSevenSixAxisSensor, "ActivateSevenSixAxisSensor"},
{304, &Hid::StartSevenSixAxisSensor, "StartSevenSixAxisSensor"},
{305, &Hid::StopSevenSixAxisSensor, "StopSevenSixAxisSensor"},
{306, &Hid::InitializeSevenSixAxisSensor, "InitializeSevenSixAxisSensor"},
{307, nullptr, "FinalizeSevenSixAxisSensor"},
{307, &Hid::FinalizeSevenSixAxisSensor, "FinalizeSevenSixAxisSensor"},
{308, nullptr, "SetSevenSixAxisSensorFusionStrength"},
{309, nullptr, "GetSevenSixAxisSensorFusionStrength"},
{310, nullptr, "ResetSevenSixAxisSensorTimestamp"},
{310, &Hid::ResetSevenSixAxisSensorTimestamp, "ResetSevenSixAxisSensorTimestamp"},
{400, nullptr, "IsUsbFullKeyControllerEnabled"},
{401, nullptr, "EnableUsbFullKeyController"},
{402, nullptr, "IsUsbFullKeyControllerConnected"},
@@ -374,6 +374,15 @@ void Hid::ActivateKeyboard(Kernel::HLERequestContext& ctx) {
rb.Push(RESULT_SUCCESS);
}
void Hid::SendKeyboardLockKeyEvent(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto flags{rp.Pop<u32>()};
LOG_WARNING(Service_HID, "(STUBBED) called. flags={}", flags);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
}
void Hid::ActivateGesture(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto unknown{rp.Pop<u32>()};
@@ -413,6 +422,18 @@ void Hid::StartSixAxisSensor(Kernel::HLERequestContext& ctx) {
rb.Push(RESULT_SUCCESS);
}
void Hid::StopSixAxisSensor(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto handle{rp.Pop<u32>()};
const auto applet_resource_user_id{rp.Pop<u64>()};
LOG_WARNING(Service_HID, "(STUBBED) called, handle={}, applet_resource_user_id={}", handle,
applet_resource_user_id);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
}
void Hid::SetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto handle{rp.Pop<u32>()};
@@ -832,33 +853,35 @@ void Hid::StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) {
rb.Push(RESULT_SUCCESS);
}
void Hid::StopSixAxisSensor(Kernel::HLERequestContext& ctx) {
void Hid::StopConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto handle{rp.Pop<u32>()};
const auto applet_resource_user_id{rp.Pop<u64>()};
LOG_WARNING(Service_HID, "(STUBBED) called, handle={}", handle);
LOG_WARNING(Service_HID, "(STUBBED) called, handle={}, applet_resource_user_id={}", handle,
applet_resource_user_id);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
}
void Hid::SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx) {
void Hid::ActivateSevenSixAxisSensor(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto applet_resource_user_id{rp.Pop<u64>()};
const auto unknown{rp.Pop<u32>()};
LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}, unknown={}",
applet_resource_user_id, unknown);
LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}",
applet_resource_user_id);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
}
void Hid::SetPalmaBoostMode(Kernel::HLERequestContext& ctx) {
void Hid::StartSevenSixAxisSensor(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto unknown{rp.Pop<u32>()};
const auto applet_resource_user_id{rp.Pop<u64>()};
LOG_WARNING(Service_HID, "(STUBBED) called, unknown={}", unknown);
LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}",
applet_resource_user_id);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
@@ -882,10 +905,46 @@ void Hid::InitializeSevenSixAxisSensor(Kernel::HLERequestContext& ctx) {
rb.Push(RESULT_SUCCESS);
}
void Hid::SendKeyboardLockKeyEvent(Kernel::HLERequestContext& ctx) {
void Hid::FinalizeSevenSixAxisSensor(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto flags{rp.Pop<u32>()};
LOG_WARNING(Service_HID, "(STUBBED) called. flags={}", flags);
const auto applet_resource_user_id{rp.Pop<u64>()};
LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}",
applet_resource_user_id);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
}
void Hid::ResetSevenSixAxisSensorTimestamp(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto applet_resource_user_id{rp.Pop<u64>()};
LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}",
applet_resource_user_id);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
}
void Hid::SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto applet_resource_user_id{rp.Pop<u64>()};
const auto is_palma_all_connectable{rp.Pop<bool>()};
LOG_WARNING(Service_HID,
"(STUBBED) called, applet_resource_user_id={}, is_palma_all_connectable={}",
applet_resource_user_id, is_palma_all_connectable);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
}
void Hid::SetPalmaBoostMode(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
const auto palma_boost_mode{rp.Pop<bool>()};
LOG_WARNING(Service_HID, "(STUBBED) called, palma_boost_mode={}", palma_boost_mode);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
+9 -4
View File
@@ -91,9 +91,11 @@ private:
void ActivateTouchScreen(Kernel::HLERequestContext& ctx);
void ActivateMouse(Kernel::HLERequestContext& ctx);
void ActivateKeyboard(Kernel::HLERequestContext& ctx);
void SendKeyboardLockKeyEvent(Kernel::HLERequestContext& ctx);
void ActivateGesture(Kernel::HLERequestContext& ctx);
void ActivateNpadWithRevision(Kernel::HLERequestContext& ctx);
void StartSixAxisSensor(Kernel::HLERequestContext& ctx);
void StopSixAxisSensor(Kernel::HLERequestContext& ctx);
void SetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx);
void IsSixAxisSensorAtRest(Kernel::HLERequestContext& ctx);
void SetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx);
@@ -126,12 +128,15 @@ private:
void IsVibrationPermitted(Kernel::HLERequestContext& ctx);
void ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx);
void StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx);
void StopSixAxisSensor(Kernel::HLERequestContext& ctx);
void SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx);
void SetPalmaBoostMode(Kernel::HLERequestContext& ctx);
void StopConsoleSixAxisSensor(Kernel::HLERequestContext& ctx);
void ActivateSevenSixAxisSensor(Kernel::HLERequestContext& ctx);
void StartSevenSixAxisSensor(Kernel::HLERequestContext& ctx);
void StopSevenSixAxisSensor(Kernel::HLERequestContext& ctx);
void InitializeSevenSixAxisSensor(Kernel::HLERequestContext& ctx);
void SendKeyboardLockKeyEvent(Kernel::HLERequestContext& ctx);
void FinalizeSevenSixAxisSensor(Kernel::HLERequestContext& ctx);
void ResetSevenSixAxisSensorTimestamp(Kernel::HLERequestContext& ctx);
void SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx);
void SetPalmaBoostMode(Kernel::HLERequestContext& ctx);
std::shared_ptr<IAppletResource> applet_resource;
Core::System& system;
+1 -1
View File
@@ -76,7 +76,7 @@ std::unique_ptr<Input::ButtonDevice> Keyboard::Create(const Common::ParamPackage
int key_code = params.Get("code", 0);
std::unique_ptr<KeyButton> button = std::make_unique<KeyButton>(key_button_list);
key_button_list->AddKeyButton(key_code, button.get());
return std::move(button);
return button;
}
void Keyboard::PressKey(int key_code) {
+1 -1
View File
@@ -145,7 +145,7 @@ std::unique_ptr<Input::MotionDevice> MotionEmu::Create(const Common::ParamPackag
// Previously created device is disconnected here. Having two motion devices for 3DS is not
// expected.
current_device = device_wrapper->device;
return std::move(device_wrapper);
return device_wrapper;
}
void MotionEmu::BeginTilt(int x, int y) {
+1 -2
View File
@@ -47,7 +47,7 @@ public:
bool is_written = false, bool use_fast_cbuf = false) {
std::lock_guard lock{mutex};
const auto& memory_manager = system.GPU().MemoryManager();
auto& memory_manager = system.GPU().MemoryManager();
const std::optional<VAddr> cpu_addr_opt = memory_manager.GpuToCpuAddress(gpu_addr);
if (!cpu_addr_opt) {
return {GetEmptyBuffer(size), 0};
@@ -59,7 +59,6 @@ public:
constexpr std::size_t max_stream_size = 0x800;
if (use_fast_cbuf || size < max_stream_size) {
if (!is_written && !IsRegionWritten(cpu_addr, cpu_addr + size - 1)) {
auto& memory_manager = system.GPU().MemoryManager();
const bool is_granular = memory_manager.IsGranularRange(gpu_addr, size);
if (use_fast_cbuf) {
u8* dest;
+31 -24
View File
@@ -54,7 +54,7 @@ void MacroJITx64Impl::Compile_ALU(Macro::Opcode opcode) {
const bool is_a_zero = opcode.src_a == 0;
const bool is_b_zero = opcode.src_b == 0;
const bool valid_operation = !is_a_zero && !is_b_zero;
const bool is_move_operation = !is_a_zero && is_b_zero;
[[maybe_unused]] const bool is_move_operation = !is_a_zero && is_b_zero;
const bool has_zero_register = is_a_zero || is_b_zero;
const bool no_zero_reg_skip = opcode.alu_operation == Macro::ALUOperation::AddWithCarry ||
opcode.alu_operation == Macro::ALUOperation::SubtractWithBorrow;
@@ -73,7 +73,6 @@ void MacroJITx64Impl::Compile_ALU(Macro::Opcode opcode) {
src_b = Compile_GetRegister(opcode.src_b, eax);
}
}
Xbyak::Label skip_carry{};
bool has_emitted = false;
@@ -240,10 +239,10 @@ void MacroJITx64Impl::Compile_ExtractInsert(Macro::Opcode opcode) {
}
void MacroJITx64Impl::Compile_ExtractShiftLeftImmediate(Macro::Opcode opcode) {
auto dst = Compile_GetRegister(opcode.src_a, eax);
auto src = Compile_GetRegister(opcode.src_b, RESULT);
const auto dst = Compile_GetRegister(opcode.src_a, ecx);
const auto src = Compile_GetRegister(opcode.src_b, RESULT);
shr(src, al);
shr(src, dst.cvt8());
if (opcode.bf_size != 0 && opcode.bf_size != 31) {
and_(src, opcode.GetBitfieldMask());
} else if (opcode.bf_size == 0) {
@@ -259,8 +258,8 @@ void MacroJITx64Impl::Compile_ExtractShiftLeftImmediate(Macro::Opcode opcode) {
}
void MacroJITx64Impl::Compile_ExtractShiftLeftRegister(Macro::Opcode opcode) {
auto dst = Compile_GetRegister(opcode.src_a, eax);
auto src = Compile_GetRegister(opcode.src_b, RESULT);
const auto dst = Compile_GetRegister(opcode.src_a, ecx);
const auto src = Compile_GetRegister(opcode.src_b, RESULT);
if (opcode.bf_src_bit != 0) {
shr(src, opcode.bf_src_bit);
@@ -269,18 +268,11 @@ void MacroJITx64Impl::Compile_ExtractShiftLeftRegister(Macro::Opcode opcode) {
if (opcode.bf_size != 31) {
and_(src, opcode.GetBitfieldMask());
}
shl(src, al);
shl(src, dst.cvt8());
Compile_ProcessResult(opcode.result_operation, opcode.dst);
}
static u32 Read(Engines::Maxwell3D* maxwell3d, u32 method) {
return maxwell3d->GetRegisterValue(method);
}
static void Send(Engines::Maxwell3D* maxwell3d, Macro::MethodAddress method_address, u32 value) {
maxwell3d->CallMethodFromMME(method_address.address, value);
}
void MacroJITx64Impl::Compile_Read(Macro::Opcode opcode) {
if (optimizer.zero_reg_skip && opcode.src_a == 0) {
if (opcode.immediate == 0) {
@@ -298,15 +290,27 @@ void MacroJITx64Impl::Compile_Read(Macro::Opcode opcode) {
sub(result, opcode.immediate * -1);
}
}
Common::X64::ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
mov(Common::X64::ABI_PARAM1, qword[STATE]);
mov(Common::X64::ABI_PARAM2, RESULT);
Common::X64::CallFarFunction(*this, &Read);
Common::X64::ABI_PopRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
mov(RESULT, Common::X64::ABI_RETURN.cvt32());
// Equivalent to Engines::Maxwell3D::GetRegisterValue:
if (optimizer.enable_asserts) {
Xbyak::Label pass_range_check;
cmp(RESULT, static_cast<u32>(Engines::Maxwell3D::Regs::NUM_REGS));
jb(pass_range_check);
int3();
L(pass_range_check);
}
mov(rax, qword[STATE]);
mov(RESULT,
dword[rax + offsetof(Engines::Maxwell3D, regs) +
offsetof(Engines::Maxwell3D::Regs, reg_array) + RESULT.cvt64() * sizeof(u32)]);
Compile_ProcessResult(opcode.result_operation, opcode.dst);
}
static void Send(Engines::Maxwell3D* maxwell3d, Macro::MethodAddress method_address, u32 value) {
maxwell3d->CallMethodFromMME(method_address.address, value);
}
void Tegra::MacroJITx64Impl::Compile_Send(Xbyak::Reg32 value) {
Common::X64::ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
mov(Common::X64::ABI_PARAM1, qword[STATE]);
@@ -438,6 +442,9 @@ void MacroJITx64Impl::Compile() {
// one if our register isn't "dirty"
optimizer.optimize_for_method_move = true;
// Enable run-time assertions in JITted code
optimizer.enable_asserts = false;
// Check to see if we can skip emitting certain instructions
Optimizer_ScanFlags();
@@ -546,7 +553,7 @@ Xbyak::Reg32 MacroJITx64Impl::Compile_GetRegister(u32 index, Xbyak::Reg32 dst) {
}
void MacroJITx64Impl::Compile_ProcessResult(Macro::ResultOperation operation, u32 reg) {
auto SetRegister = [=](u32 reg, Xbyak::Reg32 result) {
const auto SetRegister = [this](u32 reg, const Xbyak::Reg32& result) {
// Register 0 is supposed to always return 0. NOP is implemented as a store to the zero
// register.
if (reg == 0) {
@@ -554,7 +561,7 @@ void MacroJITx64Impl::Compile_ProcessResult(Macro::ResultOperation operation, u3
}
mov(dword[STATE + offsetof(JITState, registers) + reg * sizeof(u32)], result);
};
auto SetMethodAddress = [=](Xbyak::Reg32 reg) { mov(METHOD_ADDRESS, reg); };
const auto SetMethodAddress = [this](const Xbyak::Reg32& reg) { mov(METHOD_ADDRESS, reg); };
switch (operation) {
case Macro::ResultOperation::IgnoreAndFetch:
+1
View File
@@ -76,6 +76,7 @@ private:
bool zero_reg_skip{};
bool skip_dummy_addimmediate{};
bool optimize_for_method_move{};
bool enable_asserts{};
};
OptimizerState optimizer{};
+22 -18
View File
@@ -210,10 +210,11 @@ bool MemoryManager::IsBlockContinuous(const GPUVAddr start, const std::size_t si
return range == inner_size;
}
void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, const std::size_t size) const {
void MemoryManager::ReadBlock(GPUVAddr gpu_src_addr, void* dest_buffer,
const std::size_t size) const {
std::size_t remaining_size{size};
std::size_t page_index{src_addr >> page_bits};
std::size_t page_offset{src_addr & page_mask};
std::size_t page_index{gpu_src_addr >> page_bits};
std::size_t page_offset{gpu_src_addr & page_mask};
auto& memory = system.Memory();
@@ -234,11 +235,11 @@ void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, const std::s
}
}
void MemoryManager::ReadBlockUnsafe(GPUVAddr src_addr, void* dest_buffer,
void MemoryManager::ReadBlockUnsafe(GPUVAddr gpu_src_addr, void* dest_buffer,
const std::size_t size) const {
std::size_t remaining_size{size};
std::size_t page_index{src_addr >> page_bits};
std::size_t page_offset{src_addr & page_mask};
std::size_t page_index{gpu_src_addr >> page_bits};
std::size_t page_offset{gpu_src_addr & page_mask};
auto& memory = system.Memory();
@@ -259,10 +260,11 @@ void MemoryManager::ReadBlockUnsafe(GPUVAddr src_addr, void* dest_buffer,
}
}
void MemoryManager::WriteBlock(GPUVAddr dest_addr, const void* src_buffer, const std::size_t size) {
void MemoryManager::WriteBlock(GPUVAddr gpu_dest_addr, const void* src_buffer,
const std::size_t size) {
std::size_t remaining_size{size};
std::size_t page_index{dest_addr >> page_bits};
std::size_t page_offset{dest_addr & page_mask};
std::size_t page_index{gpu_dest_addr >> page_bits};
std::size_t page_offset{gpu_dest_addr & page_mask};
auto& memory = system.Memory();
@@ -283,11 +285,11 @@ void MemoryManager::WriteBlock(GPUVAddr dest_addr, const void* src_buffer, const
}
}
void MemoryManager::WriteBlockUnsafe(GPUVAddr dest_addr, const void* src_buffer,
void MemoryManager::WriteBlockUnsafe(GPUVAddr gpu_dest_addr, const void* src_buffer,
const std::size_t size) {
std::size_t remaining_size{size};
std::size_t page_index{dest_addr >> page_bits};
std::size_t page_offset{dest_addr & page_mask};
std::size_t page_index{gpu_dest_addr >> page_bits};
std::size_t page_offset{gpu_dest_addr & page_mask};
auto& memory = system.Memory();
@@ -306,16 +308,18 @@ void MemoryManager::WriteBlockUnsafe(GPUVAddr dest_addr, const void* src_buffer,
}
}
void MemoryManager::CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, const std::size_t size) {
void MemoryManager::CopyBlock(GPUVAddr gpu_dest_addr, GPUVAddr gpu_src_addr,
const std::size_t size) {
std::vector<u8> tmp_buffer(size);
ReadBlock(src_addr, tmp_buffer.data(), size);
WriteBlock(dest_addr, tmp_buffer.data(), size);
ReadBlock(gpu_src_addr, tmp_buffer.data(), size);
WriteBlock(gpu_dest_addr, tmp_buffer.data(), size);
}
void MemoryManager::CopyBlockUnsafe(GPUVAddr dest_addr, GPUVAddr src_addr, const std::size_t size) {
void MemoryManager::CopyBlockUnsafe(GPUVAddr gpu_dest_addr, GPUVAddr gpu_src_addr,
const std::size_t size) {
std::vector<u8> tmp_buffer(size);
ReadBlockUnsafe(src_addr, tmp_buffer.data(), size);
WriteBlockUnsafe(dest_addr, tmp_buffer.data(), size);
ReadBlockUnsafe(gpu_src_addr, tmp_buffer.data(), size);
WriteBlockUnsafe(gpu_dest_addr, tmp_buffer.data(), size);
}
bool MemoryManager::IsGranularRange(GPUVAddr gpu_addr, std::size_t size) {
+6 -6
View File
@@ -79,9 +79,9 @@ public:
* in the Host Memory counterpart. Note: This functions cause Host GPU Memory
* Flushes and Invalidations, respectively to each operation.
*/
void ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) const;
void WriteBlock(GPUVAddr dest_addr, const void* src_buffer, std::size_t size);
void CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size);
void ReadBlock(GPUVAddr gpu_src_addr, void* dest_buffer, std::size_t size) const;
void WriteBlock(GPUVAddr gpu_dest_addr, const void* src_buffer, std::size_t size);
void CopyBlock(GPUVAddr gpu_dest_addr, GPUVAddr gpu_src_addr, std::size_t size);
/**
* ReadBlockUnsafe and WriteBlockUnsafe are special versions of ReadBlock and
@@ -93,9 +93,9 @@ public:
* WriteBlockUnsafe instead of WriteBlock since it shouldn't invalidate the texture
* being flushed.
*/
void ReadBlockUnsafe(GPUVAddr src_addr, void* dest_buffer, std::size_t size) const;
void WriteBlockUnsafe(GPUVAddr dest_addr, const void* src_buffer, std::size_t size);
void CopyBlockUnsafe(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size);
void ReadBlockUnsafe(GPUVAddr gpu_src_addr, void* dest_buffer, std::size_t size) const;
void WriteBlockUnsafe(GPUVAddr gpu_dest_addr, const void* src_buffer, std::size_t size);
void CopyBlockUnsafe(GPUVAddr gpu_dest_addr, GPUVAddr gpu_src_addr, std::size_t size);
/**
* IsGranularRange checks if a gpu region can be simply read with a pointer
+5 -5
View File
@@ -220,8 +220,8 @@ private:
return cache_begin < addr_end && addr_begin < cache_end;
};
const u64 page_end = addr_end >> PAGE_SHIFT;
for (u64 page = addr_begin >> PAGE_SHIFT; page <= page_end; ++page) {
const u64 page_end = addr_end >> PAGE_BITS;
for (u64 page = addr_begin >> PAGE_BITS; page <= page_end; ++page) {
const auto& it = cached_queries.find(page);
if (it == std::end(cached_queries)) {
continue;
@@ -242,14 +242,14 @@ private:
/// Registers the passed parameters as cached and returns a pointer to the stored cached query.
CachedQuery* Register(VideoCore::QueryType type, VAddr cpu_addr, u8* host_ptr, bool timestamp) {
rasterizer.UpdatePagesCachedCount(cpu_addr, CachedQuery::SizeInBytes(timestamp), 1);
const u64 page = static_cast<u64>(cpu_addr) >> PAGE_SHIFT;
const u64 page = static_cast<u64>(cpu_addr) >> PAGE_BITS;
return &cached_queries[page].emplace_back(static_cast<QueryCache&>(*this), type, cpu_addr,
host_ptr);
}
/// Tries to a get a cached query. Returns nullptr on failure.
CachedQuery* TryGet(VAddr addr) {
const u64 page = static_cast<u64>(addr) >> PAGE_SHIFT;
const u64 page = static_cast<u64>(addr) >> PAGE_BITS;
const auto it = cached_queries.find(page);
if (it == std::end(cached_queries)) {
return nullptr;
@@ -268,7 +268,7 @@ private:
}
static constexpr std::uintptr_t PAGE_SIZE = 4096;
static constexpr unsigned PAGE_SHIFT = 12;
static constexpr unsigned PAGE_BITS = 12;
Core::System& system;
VideoCore::RasterizerInterface& rasterizer;
@@ -281,14 +281,14 @@ private:
template <const std::string_view& op>
std::string Unary(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("{}{} {}, {};", op, Modifiers(operation), temporary, Visit(operation[0]));
return temporary;
}
template <const std::string_view& op>
std::string Binary(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("{}{} {}, {}, {};", op, Modifiers(operation), temporary, Visit(operation[0]),
Visit(operation[1]));
return temporary;
@@ -296,7 +296,7 @@ private:
template <const std::string_view& op>
std::string Trinary(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("{}{} {}, {}, {}, {};", op, Modifiers(operation), temporary, Visit(operation[0]),
Visit(operation[1]), Visit(operation[2]));
return temporary;
@@ -304,7 +304,7 @@ private:
template <const std::string_view& op, bool unordered>
std::string FloatComparison(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("TRUNC.U.CC RC.x, {};", Binary<op>(operation));
AddLine("MOV.S {}, 0;", temporary);
AddLine("MOV.S {} (NE.x), -1;", temporary);
@@ -331,7 +331,7 @@ private:
template <const std::string_view& op, bool is_nan>
std::string HalfComparison(Operation operation) {
const std::string tmp1 = AllocVectorTemporary();
std::string tmp1 = AllocVectorTemporary();
const std::string tmp2 = AllocVectorTemporary();
const std::string op_a = Visit(operation[0]);
const std::string op_b = Visit(operation[1]);
@@ -367,15 +367,14 @@ private:
AddLine("MOV.F {}.{}, {};", value, Swizzle(i), Visit(meta.values[i]));
}
const std::string result = coord;
AddLine("ATOMIM.{}.{} {}.x, {}, {}, image[{}], {};", op, type, result, value, coord,
AddLine("ATOMIM.{}.{} {}.x, {}, {}, image[{}], {};", op, type, coord, value, coord,
image_id, ImageType(meta.image.type));
return fmt::format("{}.x", result);
return fmt::format("{}.x", coord);
}
template <const std::string_view& op, const std::string_view& type>
std::string Atomic(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
std::string address;
std::string_view opname;
if (const auto gmem = std::get_if<GmemNode>(&*operation[0])) {
@@ -396,7 +395,7 @@ private:
template <char type>
std::string Negate(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
if constexpr (type == 'F') {
AddLine("MOV.F32 {}, -{};", temporary, Visit(operation[0]));
} else {
@@ -407,7 +406,7 @@ private:
template <char type>
std::string Absolute(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("MOV.{} {}, |{}|;", type, temporary, Visit(operation[0]));
return temporary;
}
@@ -1156,20 +1155,20 @@ void ARBDecompiler::VisitAST(const ASTNode& node) {
}
std::string ARBDecompiler::VisitExpression(const Expr& node) {
const std::string result = AllocTemporary();
if (const auto expr = std::get_if<ExprAnd>(&*node)) {
std::string result = AllocTemporary();
AddLine("AND.U {}, {}, {};", result, VisitExpression(expr->operand1),
VisitExpression(expr->operand2));
return result;
}
if (const auto expr = std::get_if<ExprOr>(&*node)) {
const std::string result = AllocTemporary();
std::string result = AllocTemporary();
AddLine("OR.U {}, {}, {};", result, VisitExpression(expr->operand1),
VisitExpression(expr->operand2));
return result;
}
if (const auto expr = std::get_if<ExprNot>(&*node)) {
const std::string result = AllocTemporary();
std::string result = AllocTemporary();
AddLine("CMP.S {}, {}, 0, -1;", result, VisitExpression(expr->operand1));
return result;
}
@@ -1186,7 +1185,7 @@ std::string ARBDecompiler::VisitExpression(const Expr& node) {
return expr->value ? "0xffffffff" : "0";
}
if (const auto expr = std::get_if<ExprGprEqual>(&*node)) {
const std::string result = AllocTemporary();
std::string result = AllocTemporary();
AddLine("SEQ.U {}, R{}.x, {};", result, expr->gpr, expr->value);
return result;
}
@@ -1231,13 +1230,13 @@ std::string ARBDecompiler::Visit(const Node& node) {
}
if (const auto immediate = std::get_if<ImmediateNode>(&*node)) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("MOV.U {}, {};", temporary, immediate->GetValue());
return temporary;
}
if (const auto predicate = std::get_if<PredicateNode>(&*node)) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
switch (const auto index = predicate->GetIndex(); index) {
case Tegra::Shader::Pred::UnusedIndex:
AddLine("MOV.S {}, -1;", temporary);
@@ -1333,13 +1332,13 @@ std::string ARBDecompiler::Visit(const Node& node) {
} else {
offset_string = Visit(offset);
}
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("LDC.F32 {}, cbuf{}[{}];", temporary, cbuf->GetIndex(), offset_string);
return temporary;
}
if (const auto gmem = std::get_if<GmemNode>(&*node)) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("SUB.U {}, {}, {};", temporary, Visit(gmem->GetRealAddress()),
Visit(gmem->GetBaseAddress()));
AddLine("LDB.U32 {}, {}[{}];", temporary, GlobalMemoryName(gmem->GetDescriptor()),
@@ -1348,14 +1347,14 @@ std::string ARBDecompiler::Visit(const Node& node) {
}
if (const auto lmem = std::get_if<LmemNode>(&*node)) {
const std::string temporary = Visit(lmem->GetAddress());
std::string temporary = Visit(lmem->GetAddress());
AddLine("SHR.U {}, {}, 2;", temporary, temporary);
AddLine("MOV.U {}, lmem[{}].x;", temporary, temporary);
return temporary;
}
if (const auto smem = std::get_if<SmemNode>(&*node)) {
const std::string temporary = Visit(smem->GetAddress());
std::string temporary = Visit(smem->GetAddress());
AddLine("LDS.U32 {}, shared_mem[{}];", temporary, temporary);
return temporary;
}
@@ -1535,7 +1534,7 @@ std::string ARBDecompiler::Assign(Operation operation) {
}
std::string ARBDecompiler::Select(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("CMP.S {}, {}, {}, {};", temporary, Visit(operation[0]), Visit(operation[1]),
Visit(operation[2]));
return temporary;
@@ -1545,12 +1544,12 @@ std::string ARBDecompiler::FClamp(Operation operation) {
// 1.0f in hex, replace with std::bit_cast on C++20
static constexpr u32 POSITIVE_ONE = 0x3f800000;
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
const Node& value = operation[0];
const Node& low = operation[1];
const Node& high = operation[2];
const auto imm_low = std::get_if<ImmediateNode>(&*low);
const auto imm_high = std::get_if<ImmediateNode>(&*high);
const auto* const imm_low = std::get_if<ImmediateNode>(&*low);
const auto* const imm_high = std::get_if<ImmediateNode>(&*high);
if (imm_low && imm_high && imm_low->GetValue() == 0 && imm_high->GetValue() == POSITIVE_ONE) {
AddLine("MOV.F32.SAT {}, {};", temporary, Visit(value));
} else {
@@ -1574,7 +1573,7 @@ std::string ARBDecompiler::FCastHalf1(Operation operation) {
}
std::string ARBDecompiler::FSqrt(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("RSQ.F32 {}, {};", temporary, Visit(operation[0]));
AddLine("RCP.F32 {}, {};", temporary, temporary);
return temporary;
@@ -1588,7 +1587,7 @@ std::string ARBDecompiler::FSwizzleAdd(Operation operation) {
AddLine("ADD.F {}.x, {}, {};", temporary, Visit(operation[0]), Visit(operation[1]));
return fmt::format("{}.x", temporary);
}
const std::string lut = AllocVectorTemporary();
AddLine("AND.U {}.z, {}.threadid, 3;", temporary, StageInputName(stage));
AddLine("SHL.U {}.z, {}.z, 1;", temporary, temporary);
AddLine("SHR.U {}.z, {}, {}.z;", temporary, Visit(operation[2]), temporary);
@@ -1766,21 +1765,21 @@ std::string ARBDecompiler::LogicalAssign(Operation operation) {
}
std::string ARBDecompiler::LogicalPick2(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
const u32 index = std::get<ImmediateNode>(*operation[1]).GetValue();
AddLine("MOV.U {}, {}.{};", temporary, Visit(operation[0]), Swizzle(index));
return temporary;
}
std::string ARBDecompiler::LogicalAnd2(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
const std::string op = Visit(operation[0]);
AddLine("AND.U {}, {}.x, {}.y;", temporary, op, op);
return temporary;
}
std::string ARBDecompiler::FloatOrdered(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("MOVC.F32 RC.x, {};", Visit(operation[0]));
AddLine("MOVC.F32 RC.y, {};", Visit(operation[1]));
AddLine("MOV.S {}, -1;", temporary);
@@ -1790,7 +1789,7 @@ std::string ARBDecompiler::FloatOrdered(Operation operation) {
}
std::string ARBDecompiler::FloatUnordered(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("MOVC.F32 RC.x, {};", Visit(operation[0]));
AddLine("MOVC.F32 RC.y, {};", Visit(operation[1]));
AddLine("MOV.S {}, 0;", temporary);
@@ -1800,7 +1799,7 @@ std::string ARBDecompiler::FloatUnordered(Operation operation) {
}
std::string ARBDecompiler::LogicalAddCarry(Operation operation) {
const std::string temporary = AllocTemporary();
std::string temporary = AllocTemporary();
AddLine("ADDC.U RC, {}, {};", Visit(operation[0]), Visit(operation[1]));
AddLine("MOV.S {}, 0;", temporary);
AddLine("IF CF.x;");
@@ -13,6 +13,7 @@
#include <fmt/format.h>
#include "common/dynamic_library.h"
#include "common/file_util.h"
#include "common/logging/log.h"
#include "common/telemetry.h"
#include "core/core.h"
@@ -76,7 +77,8 @@ Common::DynamicLibrary OpenVulkanLibrary() {
char* libvulkan_env = getenv("LIBVULKAN_PATH");
if (!libvulkan_env || !library.Open(libvulkan_env)) {
// Use the libvulkan.dylib from the application bundle.
std::string filename = File::GetBundleDirectory() + "/Contents/Frameworks/libvulkan.dylib";
const std::string filename =
FileUtil::GetBundleDirectory() + "/Contents/Frameworks/libvulkan.dylib";
library.Open(filename.c_str());
}
#else
@@ -870,7 +870,7 @@ void RasterizerVulkan::BeginTransformFeedback() {
UNIMPLEMENTED_IF(binding.buffer_offset != 0);
const GPUVAddr gpu_addr = binding.Address();
const std::size_t size = binding.buffer_size;
const auto size = static_cast<VkDeviceSize>(binding.buffer_size);
const auto [buffer, offset] = buffer_cache.UploadMemory(gpu_addr, size, 4, true);
scheduler.Record([buffer = buffer, offset = offset, size](vk::CommandBuffer cmdbuf) {
@@ -1154,7 +1154,7 @@ void RasterizerVulkan::SetupTexture(const Tegra::Texture::FullTextureInfo& textu
const auto sampler = sampler_cache.GetSampler(texture.tsc);
update_descriptor_queue.AddSampledImage(sampler, image_view);
const auto image_layout = update_descriptor_queue.GetLastImageLayout();
VkImageLayout* const image_layout = update_descriptor_queue.LastImageLayout();
*image_layout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL;
sampled_views.push_back(ImageView{std::move(view), image_layout});
}
@@ -1180,7 +1180,7 @@ void RasterizerVulkan::SetupImage(const Tegra::Texture::TICEntry& tic, const Ima
view->GetImageView(tic.x_source, tic.y_source, tic.z_source, tic.w_source);
update_descriptor_queue.AddImage(image_view);
const auto image_layout = update_descriptor_queue.GetLastImageLayout();
VkImageLayout* const image_layout = update_descriptor_queue.LastImageLayout();
*image_layout = VK_IMAGE_LAYOUT_GENERAL;
image_views.push_back(ImageView{std::move(view), image_layout});
}
@@ -24,35 +24,25 @@ void VKUpdateDescriptorQueue::TickFrame() {
}
void VKUpdateDescriptorQueue::Acquire() {
entries.clear();
}
// Minimum number of entries required.
// This is the maximum number of entries a single draw call migth use.
static constexpr std::size_t MIN_ENTRIES = 0x400;
void VKUpdateDescriptorQueue::Send(VkDescriptorUpdateTemplateKHR update_template,
VkDescriptorSet set) {
if (payload.size() + entries.size() >= payload.max_size()) {
if (payload.size() + MIN_ENTRIES >= payload.max_size()) {
LOG_WARNING(Render_Vulkan, "Payload overflow, waiting for worker thread");
scheduler.WaitWorker();
payload.clear();
}
upload_start = &*payload.end();
}
// TODO(Rodrigo): Rework to write the payload directly
const auto payload_start = payload.data() + payload.size();
for (const auto& entry : entries) {
if (const auto image = std::get_if<VkDescriptorImageInfo>(&entry)) {
payload.push_back(*image);
} else if (const auto buffer = std::get_if<VkDescriptorBufferInfo>(&entry)) {
payload.push_back(*buffer);
} else if (const auto texel = std::get_if<VkBufferView>(&entry)) {
payload.push_back(*texel);
} else {
UNREACHABLE();
}
}
scheduler.Record(
[payload_start, set, update_template, logical = &device.GetLogical()](vk::CommandBuffer) {
logical->UpdateDescriptorSet(set, update_template, payload_start);
});
void VKUpdateDescriptorQueue::Send(VkDescriptorUpdateTemplateKHR update_template,
VkDescriptorSet set) {
const void* const data = upload_start;
const vk::Device* const logical = &device.GetLogical();
scheduler.Record([data, logical, set, update_template](vk::CommandBuffer) {
logical->UpdateDescriptorSet(set, update_template, data);
});
}
} // namespace Vulkan
@@ -15,17 +15,13 @@ namespace Vulkan {
class VKDevice;
class VKScheduler;
class DescriptorUpdateEntry {
public:
explicit DescriptorUpdateEntry() {}
struct DescriptorUpdateEntry {
DescriptorUpdateEntry(VkDescriptorImageInfo image_) : image{image_} {}
DescriptorUpdateEntry(VkDescriptorImageInfo image) : image{image} {}
DescriptorUpdateEntry(VkDescriptorBufferInfo buffer_) : buffer{buffer_} {}
DescriptorUpdateEntry(VkDescriptorBufferInfo buffer) : buffer{buffer} {}
DescriptorUpdateEntry(VkBufferView texel_buffer_) : texel_buffer{texel_buffer_} {}
DescriptorUpdateEntry(VkBufferView texel_buffer) : texel_buffer{texel_buffer} {}
private:
union {
VkDescriptorImageInfo image;
VkDescriptorBufferInfo buffer;
@@ -45,32 +41,34 @@ public:
void Send(VkDescriptorUpdateTemplateKHR update_template, VkDescriptorSet set);
void AddSampledImage(VkSampler sampler, VkImageView image_view) {
entries.emplace_back(VkDescriptorImageInfo{sampler, image_view, {}});
payload.emplace_back(VkDescriptorImageInfo{sampler, image_view, {}});
}
void AddImage(VkImageView image_view) {
entries.emplace_back(VkDescriptorImageInfo{{}, image_view, {}});
payload.emplace_back(VkDescriptorImageInfo{{}, image_view, {}});
}
void AddBuffer(VkBuffer buffer, u64 offset, std::size_t size) {
entries.emplace_back(VkDescriptorBufferInfo{buffer, offset, size});
payload.emplace_back(VkDescriptorBufferInfo{buffer, offset, size});
}
void AddTexelBuffer(VkBufferView texel_buffer) {
entries.emplace_back(texel_buffer);
payload.emplace_back(texel_buffer);
}
VkImageLayout* GetLastImageLayout() {
return &std::get<VkDescriptorImageInfo>(entries.back()).imageLayout;
VkImageLayout* LastImageLayout() {
return &payload.back().image.imageLayout;
}
const VkImageLayout* LastImageLayout() const {
return &payload.back().image.imageLayout;
}
private:
using Variant = std::variant<VkDescriptorImageInfo, VkDescriptorBufferInfo, VkBufferView>;
const VKDevice& device;
VKScheduler& scheduler;
boost::container::static_vector<Variant, 0x400> entries;
const DescriptorUpdateEntry* upload_start = nullptr;
boost::container::static_vector<DescriptorUpdateEntry, 0x10000> payload;
};
+1 -2
View File
@@ -725,8 +725,7 @@ bool PhysicalDevice::GetSurfaceSupportKHR(u32 queue_family_index, VkSurfaceKHR s
return supported == VK_TRUE;
}
VkSurfaceCapabilitiesKHR PhysicalDevice::GetSurfaceCapabilitiesKHR(VkSurfaceKHR surface) const
noexcept {
VkSurfaceCapabilitiesKHR PhysicalDevice::GetSurfaceCapabilitiesKHR(VkSurfaceKHR surface) const {
VkSurfaceCapabilitiesKHR capabilities;
Check(dld->vkGetPhysicalDeviceSurfaceCapabilitiesKHR(physical_device, surface, &capabilities));
return capabilities;
+1 -1
View File
@@ -779,7 +779,7 @@ public:
bool GetSurfaceSupportKHR(u32 queue_family_index, VkSurfaceKHR) const;
VkSurfaceCapabilitiesKHR GetSurfaceCapabilitiesKHR(VkSurfaceKHR) const noexcept;
VkSurfaceCapabilitiesKHR GetSurfaceCapabilitiesKHR(VkSurfaceKHR) const;
std::vector<VkSurfaceFormatKHR> GetSurfaceFormatsKHR(VkSurfaceKHR) const;
+2 -2
View File
@@ -66,12 +66,12 @@ ProgramCode GetShaderCode(Tegra::MemoryManager& memory_manager, GPUVAddr gpu_add
u64 GetUniqueIdentifier(Tegra::Engines::ShaderType shader_type, bool is_a, const ProgramCode& code,
const ProgramCode& code_b) {
u64 unique_identifier = boost::hash_value(code);
size_t unique_identifier = boost::hash_value(code);
if (is_a) {
// VertexA programs include two programs
boost::hash_combine(unique_identifier, boost::hash_value(code_b));
}
return unique_identifier;
return static_cast<u64>(unique_identifier);
}
} // namespace VideoCommon::Shader
+5 -5
View File
@@ -19,7 +19,7 @@ namespace VideoCommon {
template <class T>
class ShaderCache {
static constexpr u64 PAGE_SHIFT = 14;
static constexpr u64 PAGE_BITS = 14;
struct Entry {
VAddr addr_start;
@@ -87,8 +87,8 @@ protected:
const VAddr addr_end = addr + size;
Entry* const entry = NewEntry(addr, addr_end, data.get());
const u64 page_end = addr_end >> PAGE_SHIFT;
for (u64 page = addr >> PAGE_SHIFT; page <= page_end; ++page) {
const u64 page_end = addr_end >> PAGE_BITS;
for (u64 page = addr >> PAGE_BITS; page <= page_end; ++page) {
invalidation_cache[page].push_back(entry);
}
@@ -108,8 +108,8 @@ private:
/// @pre invalidation_mutex is locked
void InvalidatePagesInRegion(VAddr addr, std::size_t size) {
const VAddr addr_end = addr + size;
const u64 page_end = addr_end >> PAGE_SHIFT;
for (u64 page = addr >> PAGE_SHIFT; page <= page_end; ++page) {
const u64 page_end = addr_end >> PAGE_BITS;
for (u64 page = addr >> PAGE_BITS; page <= page_end; ++page) {
const auto it = invalidation_cache.find(page);
if (it == invalidation_cache.end()) {
continue;
+4
View File
@@ -208,6 +208,10 @@ if (MSVC)
copy_yuzu_unicorn_deps(yuzu)
endif()
if (NOT APPLE)
target_compile_definitions(yuzu PRIVATE HAS_OPENGL)
endif()
if (ENABLE_VULKAN)
target_include_directories(yuzu PRIVATE ../../externals/Vulkan-Headers/include)
target_compile_definitions(yuzu PRIVATE HAS_VULKAN)
+15 -2
View File
@@ -8,13 +8,16 @@
#include <QHBoxLayout>
#include <QKeyEvent>
#include <QMessageBox>
#include <QOffscreenSurface>
#include <QOpenGLContext>
#include <QPainter>
#include <QScreen>
#include <QStringList>
#include <QWindow>
#ifdef HAS_OPENGL
#include <QOffscreenSurface>
#include <QOpenGLContext>
#endif
#if !defined(WIN32) && HAS_VULKAN
#include <qpa/qplatformnativeinterface.h>
#endif
@@ -98,6 +101,7 @@ void EmuThread::run() {
#endif
}
#ifdef HAS_OPENGL
class OpenGLSharedContext : public Core::Frontend::GraphicsContext {
public:
/// Create the original context that should be shared from
@@ -183,6 +187,7 @@ private:
std::unique_ptr<QOffscreenSurface> offscreen_surface{};
QSurface* surface;
};
#endif
class DummyContext : public Core::Frontend::GraphicsContext {};
@@ -473,6 +478,7 @@ void GRenderWindow::resizeEvent(QResizeEvent* event) {
}
std::unique_ptr<Core::Frontend::GraphicsContext> GRenderWindow::CreateSharedContext() const {
#ifdef HAS_OPENGL
if (Settings::values.renderer_backend == Settings::RendererBackend::OpenGL) {
auto c = static_cast<OpenGLSharedContext*>(main_context.get());
// Bind the shared contexts to the main surface in case the backend wants to take over
@@ -480,6 +486,7 @@ std::unique_ptr<Core::Frontend::GraphicsContext> GRenderWindow::CreateSharedCont
return std::make_unique<OpenGLSharedContext>(c->GetShareContext(),
child_widget->windowHandle());
}
#endif
return std::make_unique<DummyContext>();
}
@@ -560,6 +567,7 @@ void GRenderWindow::OnMinimalClientAreaChangeRequest(std::pair<u32, u32> minimal
}
bool GRenderWindow::InitializeOpenGL() {
#ifdef HAS_OPENGL
// TODO: One of these flags might be interesting: WA_OpaquePaintEvent, WA_NoBackground,
// WA_DontShowOnScreen, WA_DeleteOnClose
auto child = new OpenGLRenderWidget(this);
@@ -571,6 +579,11 @@ bool GRenderWindow::InitializeOpenGL() {
std::make_unique<OpenGLSharedContext>(context->GetShareContext(), child->windowHandle()));
return true;
#else
QMessageBox::warning(this, tr("OpenGL not available!"),
tr("yuzu has not been compiled with OpenGL support."));
return false;
#endif
}
bool GRenderWindow::InitializeVulkan() {
+1 -1
View File
@@ -16,4 +16,4 @@ IDI_ICON1 ICON "../../dist/yuzu.ico"
// RT_MANIFEST
//
1 RT_MANIFEST "../../dist/yuzu.manifest"
0 RT_MANIFEST "../../dist/yuzu.manifest"
+1 -1
View File
@@ -14,4 +14,4 @@ YUZU_ICON ICON "../../dist/yuzu.ico"
// RT_MANIFEST
//
1 RT_MANIFEST "../../dist/yuzu.manifest"
0 RT_MANIFEST "../../dist/yuzu.manifest"
+1 -1
View File
@@ -14,4 +14,4 @@ YUZU_ICON ICON "../../dist/yuzu.ico"
// RT_MANIFEST
//
1 RT_MANIFEST "../../dist/yuzu.manifest"
0 RT_MANIFEST "../../dist/yuzu.manifest"