Compare commits
20 Commits
android-10
...
android-11
Author | SHA1 | Date | |
---|---|---|---|
f4b79b0982 | |||
76fe4da3fb | |||
9274eaecd0 | |||
c733620024 | |||
b1909b0435 | |||
1cc764988f | |||
da5c49f22d | |||
6b93b0b08c | |||
3d4a064674 | |||
cfe73af6f2 | |||
77fb9d415b | |||
bbdaa62175 | |||
db37e583ff | |||
d28e826e47 | |||
13beb85514 | |||
4b06bcc82c | |||
12ebc8d9d1 | |||
2b85e9e997 | |||
59b62c6507 | |||
e02ee8e59d |
@ -1,3 +1,12 @@
|
||||
| Pull Request | Commit | Title | Author | Merged? |
|
||||
|----|----|----|----|----|
|
||||
| [11827](https://github.com/yuzu-emu/yuzu//pull/11827) | [`689f346e9`](https://github.com/yuzu-emu/yuzu//pull/11827/files) | nvnflinger: fix reporting and freeing of preallocated buffers | [liamwhite](https://github.com/liamwhite/) | Yes |
|
||||
|
||||
|
||||
End of merge log. You can find the original README.md below the break.
|
||||
|
||||
-----
|
||||
|
||||
<!--
|
||||
SPDX-FileCopyrightText: 2018 yuzu Emulator Project
|
||||
SPDX-License-Identifier: GPL-2.0-or-later
|
||||
|
@ -11,6 +11,7 @@
|
||||
#include <mach/mach.h>
|
||||
#elif defined(_WIN32)
|
||||
#include <windows.h>
|
||||
#include "common/string_util.h"
|
||||
#else
|
||||
#if defined(__Bitrig__) || defined(__DragonFly__) || defined(__FreeBSD__) || defined(__OpenBSD__)
|
||||
#include <pthread_np.h>
|
||||
@ -82,29 +83,8 @@ void SetCurrentThreadPriority(ThreadPriority new_priority) {
|
||||
#ifdef _MSC_VER
|
||||
|
||||
// Sets the debugger-visible name of the current thread.
|
||||
// Uses trick documented in:
|
||||
// https://docs.microsoft.com/en-us/visualstudio/debugger/how-to-set-a-thread-name-in-native-code
|
||||
void SetCurrentThreadName(const char* name) {
|
||||
static const DWORD MS_VC_EXCEPTION = 0x406D1388;
|
||||
|
||||
#pragma pack(push, 8)
|
||||
struct THREADNAME_INFO {
|
||||
DWORD dwType; // must be 0x1000
|
||||
LPCSTR szName; // pointer to name (in user addr space)
|
||||
DWORD dwThreadID; // thread ID (-1=caller thread)
|
||||
DWORD dwFlags; // reserved for future use, must be zero
|
||||
} info;
|
||||
#pragma pack(pop)
|
||||
|
||||
info.dwType = 0x1000;
|
||||
info.szName = name;
|
||||
info.dwThreadID = std::numeric_limits<DWORD>::max();
|
||||
info.dwFlags = 0;
|
||||
|
||||
__try {
|
||||
RaiseException(MS_VC_EXCEPTION, 0, sizeof(info) / sizeof(ULONG_PTR), (ULONG_PTR*)&info);
|
||||
} __except (EXCEPTION_CONTINUE_EXECUTION) {
|
||||
}
|
||||
SetThreadDescription(GetCurrentThread(), UTF8ToUTF16W(name).data());
|
||||
}
|
||||
|
||||
#else // !MSVC_VER, so must be POSIX threads
|
||||
|
@ -1,6 +1,7 @@
|
||||
// SPDX-FileCopyrightText: Copyright 2019 yuzu Emulator Project
|
||||
// SPDX-License-Identifier: GPL-2.0-or-later
|
||||
|
||||
#include "common/logging/log.h"
|
||||
#include "core/file_sys/system_archive/system_version.h"
|
||||
#include "core/file_sys/vfs_vector.h"
|
||||
#include "core/hle/api_version.h"
|
||||
@ -12,6 +13,9 @@ std::string GetLongDisplayVersion() {
|
||||
}
|
||||
|
||||
VirtualDir SystemVersion() {
|
||||
LOG_WARNING(Common_Filesystem, "called - Using hardcoded firmware version '{}'",
|
||||
GetLongDisplayVersion());
|
||||
|
||||
VirtualFile file = std::make_shared<VectorVfsFile>(std::vector<u8>(0x100), "file");
|
||||
file->WriteObject(HLE::ApiVersion::HOS_VERSION_MAJOR, 0);
|
||||
file->WriteObject(HLE::ApiVersion::HOS_VERSION_MINOR, 1);
|
||||
|
@ -27,10 +27,12 @@ namespace {
|
||||
static thread_local std::array read_buffer_data_a{
|
||||
Common::ScratchBuffer<u8>(),
|
||||
Common::ScratchBuffer<u8>(),
|
||||
Common::ScratchBuffer<u8>(),
|
||||
};
|
||||
static thread_local std::array read_buffer_data_x{
|
||||
Common::ScratchBuffer<u8>(),
|
||||
Common::ScratchBuffer<u8>(),
|
||||
Common::ScratchBuffer<u8>(),
|
||||
};
|
||||
} // Anonymous namespace
|
||||
|
||||
@ -343,6 +345,7 @@ std::span<const u8> HLERequestContext::ReadBufferA(std::size_t buffer_index) con
|
||||
static thread_local std::array read_buffer_a{
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
};
|
||||
|
||||
ASSERT_OR_EXECUTE_MSG(
|
||||
@ -358,6 +361,7 @@ std::span<const u8> HLERequestContext::ReadBufferX(std::size_t buffer_index) con
|
||||
static thread_local std::array read_buffer_x{
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
};
|
||||
|
||||
ASSERT_OR_EXECUTE_MSG(
|
||||
@ -373,10 +377,12 @@ std::span<const u8> HLERequestContext::ReadBuffer(std::size_t buffer_index) cons
|
||||
static thread_local std::array read_buffer_a{
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
};
|
||||
static thread_local std::array read_buffer_x{
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
Core::Memory::CpuGuestMemory<u8, Core::Memory::GuestMemoryFlags::SafeRead>(memory, 0, 0),
|
||||
};
|
||||
|
||||
const bool is_buffer_a{BufferDescriptorA().size() > buffer_index &&
|
||||
|
@ -41,7 +41,7 @@ bool BufferQueueCore::WaitForDequeueCondition(std::unique_lock<std::mutex>& lk)
|
||||
s32 BufferQueueCore::GetMinUndequeuedBufferCountLocked(bool async) const {
|
||||
// If DequeueBuffer is allowed to error out, we don't have to add an extra buffer.
|
||||
if (!use_async_buffer) {
|
||||
return max_acquired_buffer_count;
|
||||
return 0;
|
||||
}
|
||||
|
||||
if (dequeue_buffer_cannot_block || async) {
|
||||
@ -52,7 +52,7 @@ s32 BufferQueueCore::GetMinUndequeuedBufferCountLocked(bool async) const {
|
||||
}
|
||||
|
||||
s32 BufferQueueCore::GetMinMaxBufferCountLocked(bool async) const {
|
||||
return GetMinUndequeuedBufferCountLocked(async) + 1;
|
||||
return GetMinUndequeuedBufferCountLocked(async);
|
||||
}
|
||||
|
||||
s32 BufferQueueCore::GetMaxBufferCountLocked(bool async) const {
|
||||
@ -61,7 +61,7 @@ s32 BufferQueueCore::GetMaxBufferCountLocked(bool async) const {
|
||||
|
||||
if (override_max_buffer_count != 0) {
|
||||
ASSERT(override_max_buffer_count >= min_buffer_count);
|
||||
max_buffer_count = override_max_buffer_count;
|
||||
return override_max_buffer_count;
|
||||
}
|
||||
|
||||
// Any buffers that are dequeued by the producer or sitting in the queue waiting to be consumed
|
||||
|
@ -134,7 +134,7 @@ Status BufferQueueProducer::WaitForFreeSlotThenRelock(bool async, s32* found, St
|
||||
const s32 max_buffer_count = core->GetMaxBufferCountLocked(async);
|
||||
if (async && core->override_max_buffer_count) {
|
||||
if (core->override_max_buffer_count < max_buffer_count) {
|
||||
LOG_ERROR(Service_Nvnflinger, "async mode is invalid with buffer count override");
|
||||
*found = BufferQueueCore::INVALID_BUFFER_SLOT;
|
||||
return Status::BadValue;
|
||||
}
|
||||
}
|
||||
@ -142,7 +142,8 @@ Status BufferQueueProducer::WaitForFreeSlotThenRelock(bool async, s32* found, St
|
||||
// Free up any buffers that are in slots beyond the max buffer count
|
||||
for (s32 s = max_buffer_count; s < BufferQueueDefs::NUM_BUFFER_SLOTS; ++s) {
|
||||
ASSERT(slots[s].buffer_state == BufferState::Free);
|
||||
if (slots[s].graphic_buffer != nullptr) {
|
||||
if (slots[s].graphic_buffer != nullptr && slots[s].buffer_state == BufferState::Free &&
|
||||
!slots[s].is_preallocated) {
|
||||
core->FreeBufferLocked(s);
|
||||
*return_flags |= Status::ReleaseAllBuffers;
|
||||
}
|
||||
|
@ -9,6 +9,35 @@
|
||||
|
||||
namespace Service::PTM {
|
||||
|
||||
enum class Location : u8 {
|
||||
Internal,
|
||||
External,
|
||||
};
|
||||
|
||||
class ISession : public ServiceFramework<ISession> {
|
||||
public:
|
||||
explicit ISession(Core::System& system_) : ServiceFramework{system_, "ISession"} {
|
||||
// clang-format off
|
||||
static const FunctionInfo functions[] = {
|
||||
{0, nullptr, "GetTemperatureRange"},
|
||||
{2, nullptr, "SetMeasurementMode"},
|
||||
{4, &ISession::GetTemperature, "GetTemperature"},
|
||||
};
|
||||
// clang-format on
|
||||
|
||||
RegisterHandlers(functions);
|
||||
}
|
||||
|
||||
private:
|
||||
void GetTemperature(HLERequestContext& ctx) {
|
||||
constexpr f32 temperature = 35;
|
||||
|
||||
IPC::ResponseBuilder rb{ctx, 3};
|
||||
rb.Push(ResultSuccess);
|
||||
rb.Push(temperature);
|
||||
}
|
||||
};
|
||||
|
||||
TS::TS(Core::System& system_) : ServiceFramework{system_, "ts"} {
|
||||
// clang-format off
|
||||
static const FunctionInfo functions[] = {
|
||||
@ -16,7 +45,7 @@ TS::TS(Core::System& system_) : ServiceFramework{system_, "ts"} {
|
||||
{1, &TS::GetTemperature, "GetTemperature"},
|
||||
{2, nullptr, "SetMeasurementMode"},
|
||||
{3, &TS::GetTemperatureMilliC, "GetTemperatureMilliC"},
|
||||
{4, nullptr, "OpenSession"},
|
||||
{4, &TS::OpenSession, "OpenSession"},
|
||||
};
|
||||
// clang-format on
|
||||
|
||||
@ -47,4 +76,13 @@ void TS::GetTemperatureMilliC(HLERequestContext& ctx) {
|
||||
rb.Push(temperature);
|
||||
}
|
||||
|
||||
void TS::OpenSession(HLERequestContext& ctx) {
|
||||
IPC::RequestParser rp{ctx};
|
||||
[[maybe_unused]] const u32 device_code = rp.Pop<u32>();
|
||||
|
||||
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
|
||||
rb.Push(ResultSuccess);
|
||||
rb.PushIpcInterface<ISession>(system);
|
||||
}
|
||||
|
||||
} // namespace Service::PTM
|
||||
|
@ -14,13 +14,9 @@ public:
|
||||
~TS() override;
|
||||
|
||||
private:
|
||||
enum class Location : u8 {
|
||||
Internal,
|
||||
External,
|
||||
};
|
||||
|
||||
void GetTemperature(HLERequestContext& ctx);
|
||||
void GetTemperatureMilliC(HLERequestContext& ctx);
|
||||
void OpenSession(HLERequestContext& ctx);
|
||||
};
|
||||
|
||||
} // namespace Service::PTM
|
||||
|
@ -5,8 +5,13 @@
|
||||
#include "common/logging/log.h"
|
||||
#include "common/settings.h"
|
||||
#include "common/string_util.h"
|
||||
#include "core/core.h"
|
||||
#include "core/file_sys/content_archive.h"
|
||||
#include "core/file_sys/errors.h"
|
||||
#include "core/file_sys/system_archive/system_version.h"
|
||||
#include "core/file_sys/nca_metadata.h"
|
||||
#include "core/file_sys/registered_cache.h"
|
||||
#include "core/file_sys/romfs.h"
|
||||
#include "core/file_sys/system_archive/system_archive.h"
|
||||
#include "core/hle/service/filesystem/filesystem.h"
|
||||
#include "core/hle/service/ipc_helpers.h"
|
||||
#include "core/hle/service/set/set.h"
|
||||
@ -22,18 +27,30 @@ enum class GetFirmwareVersionType {
|
||||
Version2,
|
||||
};
|
||||
|
||||
void GetFirmwareVersionImpl(HLERequestContext& ctx, GetFirmwareVersionType type) {
|
||||
LOG_WARNING(Service_SET, "called - Using hardcoded firmware version '{}'",
|
||||
FileSys::SystemArchive::GetLongDisplayVersion());
|
||||
|
||||
void GetFirmwareVersionImpl(Core::System& system, HLERequestContext& ctx,
|
||||
GetFirmwareVersionType type) {
|
||||
ASSERT_MSG(ctx.GetWriteBufferSize() == 0x100,
|
||||
"FirmwareVersion output buffer must be 0x100 bytes in size!");
|
||||
|
||||
// Instead of using the normal procedure of checking for the real system archive and if it
|
||||
// doesn't exist, synthesizing one, I feel that that would lead to strange bugs because a
|
||||
// used is using a really old or really new SystemVersion title. The synthesized one ensures
|
||||
// consistence (currently reports as 5.1.0-0.0)
|
||||
const auto archive = FileSys::SystemArchive::SystemVersion();
|
||||
constexpr u64 FirmwareVersionSystemDataId = 0x0100000000000809;
|
||||
auto& fsc = system.GetFileSystemController();
|
||||
|
||||
// Attempt to load version data from disk
|
||||
const FileSys::RegisteredCache* bis_system{};
|
||||
std::unique_ptr<FileSys::NCA> nca{};
|
||||
FileSys::VirtualDir romfs{};
|
||||
|
||||
bis_system = fsc.GetSystemNANDContents();
|
||||
if (bis_system) {
|
||||
nca = bis_system->GetEntry(FirmwareVersionSystemDataId, FileSys::ContentRecordType::Data);
|
||||
}
|
||||
if (nca) {
|
||||
romfs = FileSys::ExtractRomFS(nca->GetRomFS());
|
||||
}
|
||||
if (!romfs) {
|
||||
romfs = FileSys::ExtractRomFS(
|
||||
FileSys::SystemArchive::SynthesizeSystemArchive(FirmwareVersionSystemDataId));
|
||||
}
|
||||
|
||||
const auto early_exit_failure = [&ctx](std::string_view desc, Result code) {
|
||||
LOG_ERROR(Service_SET, "General failure while attempting to resolve firmware version ({}).",
|
||||
@ -42,13 +59,7 @@ void GetFirmwareVersionImpl(HLERequestContext& ctx, GetFirmwareVersionType type)
|
||||
rb.Push(code);
|
||||
};
|
||||
|
||||
if (archive == nullptr) {
|
||||
early_exit_failure("The system version archive couldn't be synthesized.",
|
||||
FileSys::ERROR_FAILED_MOUNT_ARCHIVE);
|
||||
return;
|
||||
}
|
||||
|
||||
const auto ver_file = archive->GetFile("file");
|
||||
const auto ver_file = romfs->GetFile("file");
|
||||
if (ver_file == nullptr) {
|
||||
early_exit_failure("The system version archive didn't contain the file 'file'.",
|
||||
FileSys::ERROR_INVALID_ARGUMENT);
|
||||
@ -87,12 +98,12 @@ void SET_SYS::SetLanguageCode(HLERequestContext& ctx) {
|
||||
|
||||
void SET_SYS::GetFirmwareVersion(HLERequestContext& ctx) {
|
||||
LOG_DEBUG(Service_SET, "called");
|
||||
GetFirmwareVersionImpl(ctx, GetFirmwareVersionType::Version1);
|
||||
GetFirmwareVersionImpl(system, ctx, GetFirmwareVersionType::Version1);
|
||||
}
|
||||
|
||||
void SET_SYS::GetFirmwareVersion2(HLERequestContext& ctx) {
|
||||
LOG_DEBUG(Service_SET, "called");
|
||||
GetFirmwareVersionImpl(ctx, GetFirmwareVersionType::Version2);
|
||||
GetFirmwareVersionImpl(system, ctx, GetFirmwareVersionType::Version2);
|
||||
}
|
||||
|
||||
void SET_SYS::GetAccountSettings(HLERequestContext& ctx) {
|
||||
|
@ -19,16 +19,23 @@ namespace Core::Memory {
|
||||
namespace {
|
||||
constexpr auto CHEAT_ENGINE_NS = std::chrono::nanoseconds{1000000000 / 12};
|
||||
|
||||
std::string_view ExtractName(std::string_view data, std::size_t start_index, char match) {
|
||||
std::string_view ExtractName(std::size_t& out_name_size, std::string_view data,
|
||||
std::size_t start_index, char match) {
|
||||
auto end_index = start_index;
|
||||
while (data[end_index] != match) {
|
||||
++end_index;
|
||||
if (end_index > data.size() ||
|
||||
(end_index - start_index - 1) > sizeof(CheatDefinition::readable_name)) {
|
||||
if (end_index > data.size()) {
|
||||
return {};
|
||||
}
|
||||
}
|
||||
|
||||
out_name_size = end_index - start_index;
|
||||
|
||||
// Clamp name if it's too big
|
||||
if (out_name_size > sizeof(CheatDefinition::readable_name)) {
|
||||
end_index = start_index + sizeof(CheatDefinition::readable_name);
|
||||
}
|
||||
|
||||
return data.substr(start_index, end_index - start_index);
|
||||
}
|
||||
} // Anonymous namespace
|
||||
@ -113,7 +120,8 @@ std::vector<CheatEntry> TextCheatParser::Parse(std::string_view data) const {
|
||||
return {};
|
||||
}
|
||||
|
||||
const auto name = ExtractName(data, i + 1, '}');
|
||||
std::size_t name_size{};
|
||||
const auto name = ExtractName(name_size, data, i + 1, '}');
|
||||
if (name.empty()) {
|
||||
return {};
|
||||
}
|
||||
@ -125,12 +133,13 @@ std::vector<CheatEntry> TextCheatParser::Parse(std::string_view data) const {
|
||||
.definition.readable_name[out[*current_entry].definition.readable_name.size() - 1] =
|
||||
'\0';
|
||||
|
||||
i += name.length() + 1;
|
||||
i += name_size + 1;
|
||||
} else if (data[i] == '[') {
|
||||
current_entry = out.size();
|
||||
out.emplace_back();
|
||||
|
||||
const auto name = ExtractName(data, i + 1, ']');
|
||||
std::size_t name_size{};
|
||||
const auto name = ExtractName(name_size, data, i + 1, ']');
|
||||
if (name.empty()) {
|
||||
return {};
|
||||
}
|
||||
@ -142,7 +151,7 @@ std::vector<CheatEntry> TextCheatParser::Parse(std::string_view data) const {
|
||||
.definition.readable_name[out[*current_entry].definition.readable_name.size() - 1] =
|
||||
'\0';
|
||||
|
||||
i += name.length() + 1;
|
||||
i += name_size + 1;
|
||||
} else if (::isxdigit(data[i])) {
|
||||
if (!current_entry || out[*current_entry].definition.num_opcodes >=
|
||||
out[*current_entry].definition.opcodes.size()) {
|
||||
|
@ -242,6 +242,7 @@ std::string EmitGLSL(const Profile& profile, const RuntimeInfo& runtime_info, IR
|
||||
}
|
||||
if (program.info.uses_subgroup_shuffles) {
|
||||
ctx.header += "bool shfl_in_bounds;";
|
||||
ctx.header += "uint shfl_result;";
|
||||
}
|
||||
ctx.code.insert(0, ctx.header);
|
||||
ctx.code += '}';
|
||||
|
@ -141,7 +141,8 @@ void EmitShuffleIndex(EmitContext& ctx, IR::Inst& inst, std::string_view value,
|
||||
const auto src_thread_id{fmt::format("({})|({})", lhs, min_thread_id)};
|
||||
ctx.Add("shfl_in_bounds=int({})<=int({});", src_thread_id, max_thread_id);
|
||||
SetInBoundsFlag(ctx, inst);
|
||||
ctx.AddU32("{}=shfl_in_bounds?readInvocationARB({},{}):{};", inst, value, src_thread_id, value);
|
||||
ctx.Add("shfl_result=readInvocationARB({},{});", value, src_thread_id);
|
||||
ctx.AddU32("{}=shfl_in_bounds?shfl_result:{};", inst, value);
|
||||
}
|
||||
|
||||
void EmitShuffleUp(EmitContext& ctx, IR::Inst& inst, std::string_view value, std::string_view index,
|
||||
@ -158,7 +159,8 @@ void EmitShuffleUp(EmitContext& ctx, IR::Inst& inst, std::string_view value, std
|
||||
const auto src_thread_id{fmt::format("({}-{})", THREAD_ID, index)};
|
||||
ctx.Add("shfl_in_bounds=int({})>=int({});", src_thread_id, max_thread_id);
|
||||
SetInBoundsFlag(ctx, inst);
|
||||
ctx.AddU32("{}=shfl_in_bounds?readInvocationARB({},{}):{};", inst, value, src_thread_id, value);
|
||||
ctx.Add("shfl_result=readInvocationARB({},{});", value, src_thread_id);
|
||||
ctx.AddU32("{}=shfl_in_bounds?shfl_result:{};", inst, value);
|
||||
}
|
||||
|
||||
void EmitShuffleDown(EmitContext& ctx, IR::Inst& inst, std::string_view value,
|
||||
@ -175,7 +177,8 @@ void EmitShuffleDown(EmitContext& ctx, IR::Inst& inst, std::string_view value,
|
||||
const auto src_thread_id{fmt::format("({}+{})", THREAD_ID, index)};
|
||||
ctx.Add("shfl_in_bounds=int({})<=int({});", src_thread_id, max_thread_id);
|
||||
SetInBoundsFlag(ctx, inst);
|
||||
ctx.AddU32("{}=shfl_in_bounds?readInvocationARB({},{}):{};", inst, value, src_thread_id, value);
|
||||
ctx.Add("shfl_result=readInvocationARB({},{});", value, src_thread_id);
|
||||
ctx.AddU32("{}=shfl_in_bounds?shfl_result:{};", inst, value);
|
||||
}
|
||||
|
||||
void EmitShuffleButterfly(EmitContext& ctx, IR::Inst& inst, std::string_view value,
|
||||
@ -193,7 +196,8 @@ void EmitShuffleButterfly(EmitContext& ctx, IR::Inst& inst, std::string_view val
|
||||
const auto src_thread_id{fmt::format("({}^{})", THREAD_ID, index)};
|
||||
ctx.Add("shfl_in_bounds=int({})<=int({});", src_thread_id, max_thread_id);
|
||||
SetInBoundsFlag(ctx, inst);
|
||||
ctx.AddU32("{}=shfl_in_bounds?readInvocationARB({},{}):{};", inst, value, src_thread_id, value);
|
||||
ctx.Add("shfl_result=readInvocationARB({},{});", value, src_thread_id);
|
||||
ctx.AddU32("{}=shfl_in_bounds?shfl_result:{};", inst, value);
|
||||
}
|
||||
|
||||
void EmitFSwizzleAdd(EmitContext& ctx, IR::Inst& inst, std::string_view op_a, std::string_view op_b,
|
||||
|
@ -111,16 +111,33 @@ Id GetCbuf(EmitContext& ctx, Id result_type, Id UniformDefinitions::*member_ptr,
|
||||
} else if (element_size > 1) {
|
||||
const u32 log2_element_size{static_cast<u32>(std::countr_zero(element_size))};
|
||||
const Id shift{ctx.Const(log2_element_size)};
|
||||
buffer_offset = ctx.OpShiftRightArithmetic(ctx.U32[1], ctx.Def(offset), shift);
|
||||
buffer_offset = ctx.OpShiftRightLogical(ctx.U32[1], ctx.Def(offset), shift);
|
||||
} else {
|
||||
buffer_offset = ctx.Def(offset);
|
||||
}
|
||||
if (!binding.IsImmediate()) {
|
||||
return ctx.OpFunctionCall(result_type, indirect_func, ctx.Def(binding), buffer_offset);
|
||||
}
|
||||
|
||||
const Id cbuf{ctx.cbufs[binding.U32()].*member_ptr};
|
||||
const Id access_chain{ctx.OpAccessChain(uniform_type, cbuf, ctx.u32_zero_value, buffer_offset)};
|
||||
return ctx.OpLoad(result_type, access_chain);
|
||||
const Id val = ctx.OpLoad(result_type, access_chain);
|
||||
|
||||
if (offset.IsImmediate() || !ctx.profile.has_broken_robust) {
|
||||
return val;
|
||||
}
|
||||
|
||||
const auto is_float = UniformDefinitions::IsFloat(member_ptr);
|
||||
const auto num_elements = UniformDefinitions::NumElements(member_ptr);
|
||||
const std::array zero_vec{
|
||||
is_float ? ctx.Const(0.0f) : ctx.Const(0u),
|
||||
is_float ? ctx.Const(0.0f) : ctx.Const(0u),
|
||||
is_float ? ctx.Const(0.0f) : ctx.Const(0u),
|
||||
is_float ? ctx.Const(0.0f) : ctx.Const(0u),
|
||||
};
|
||||
const Id cond = ctx.OpULessThanEqual(ctx.TypeBool(), buffer_offset, ctx.Const(0xFFFFu));
|
||||
const Id zero = ctx.OpCompositeConstruct(result_type, std::span(zero_vec.data(), num_elements));
|
||||
return ctx.OpSelect(result_type, cond, val, zero);
|
||||
}
|
||||
|
||||
Id GetCbufU32(EmitContext& ctx, const IR::Value& binding, const IR::Value& offset) {
|
||||
@ -138,7 +155,7 @@ Id GetCbufElement(EmitContext& ctx, Id vector, const IR::Value& offset, u32 inde
|
||||
const u32 element{(offset.U32() / 4) % 4 + index_offset};
|
||||
return ctx.OpCompositeExtract(ctx.U32[1], vector, element);
|
||||
}
|
||||
const Id shift{ctx.OpShiftRightArithmetic(ctx.U32[1], ctx.Def(offset), ctx.Const(2u))};
|
||||
const Id shift{ctx.OpShiftRightLogical(ctx.U32[1], ctx.Def(offset), ctx.Const(2u))};
|
||||
Id element{ctx.OpBitwiseAnd(ctx.U32[1], shift, ctx.Const(3u))};
|
||||
if (index_offset > 0) {
|
||||
element = ctx.OpIAdd(ctx.U32[1], element, ctx.Const(index_offset));
|
||||
|
@ -64,6 +64,42 @@ struct UniformDefinitions {
|
||||
Id F32{};
|
||||
Id U32x2{};
|
||||
Id U32x4{};
|
||||
|
||||
constexpr static size_t NumElements(Id UniformDefinitions::*member_ptr) {
|
||||
if (member_ptr == &UniformDefinitions::U8) {
|
||||
return 1;
|
||||
}
|
||||
if (member_ptr == &UniformDefinitions::S8) {
|
||||
return 1;
|
||||
}
|
||||
if (member_ptr == &UniformDefinitions::U16) {
|
||||
return 1;
|
||||
}
|
||||
if (member_ptr == &UniformDefinitions::S16) {
|
||||
return 1;
|
||||
}
|
||||
if (member_ptr == &UniformDefinitions::U32) {
|
||||
return 1;
|
||||
}
|
||||
if (member_ptr == &UniformDefinitions::F32) {
|
||||
return 1;
|
||||
}
|
||||
if (member_ptr == &UniformDefinitions::U32x2) {
|
||||
return 2;
|
||||
}
|
||||
if (member_ptr == &UniformDefinitions::U32x4) {
|
||||
return 4;
|
||||
}
|
||||
ASSERT(false);
|
||||
return 1;
|
||||
}
|
||||
|
||||
constexpr static bool IsFloat(Id UniformDefinitions::*member_ptr) {
|
||||
if (member_ptr == &UniformDefinitions::F32) {
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
}
|
||||
};
|
||||
|
||||
struct StorageTypeDefinition {
|
||||
|
@ -9,7 +9,6 @@ namespace Shader {
|
||||
|
||||
struct Profile {
|
||||
u32 supported_spirv{0x00010000};
|
||||
|
||||
bool unified_descriptor_binding{};
|
||||
bool support_descriptor_aliasing{};
|
||||
bool support_int8{};
|
||||
@ -82,6 +81,9 @@ struct Profile {
|
||||
bool has_broken_spirv_subgroup_mask_vector_extract_dynamic{};
|
||||
|
||||
u32 gl_max_compute_smem_size{};
|
||||
|
||||
/// Maxwell and earlier nVidia architectures have broken robust support
|
||||
bool has_broken_robust{};
|
||||
};
|
||||
|
||||
} // namespace Shader
|
||||
|
@ -356,7 +356,11 @@ PipelineCache::PipelineCache(RasterizerVulkan& rasterizer_, const Device& device
|
||||
.has_broken_fp16_float_controls = driver_id == VK_DRIVER_ID_NVIDIA_PROPRIETARY,
|
||||
.ignore_nan_fp_comparisons = false,
|
||||
.has_broken_spirv_subgroup_mask_vector_extract_dynamic =
|
||||
driver_id == VK_DRIVER_ID_QUALCOMM_PROPRIETARY};
|
||||
driver_id == VK_DRIVER_ID_QUALCOMM_PROPRIETARY,
|
||||
.has_broken_robust =
|
||||
device.IsNvidia() && device.GetNvidiaArch() <= NvidiaArchitecture::Arch_Maxwell,
|
||||
};
|
||||
|
||||
host_info = Shader::HostTranslateInfo{
|
||||
.support_float64 = device.IsFloat64Supported(),
|
||||
.support_float16 = device.IsFloat16Supported(),
|
||||
|
@ -83,15 +83,6 @@ constexpr std::array VK_FORMAT_A4B4G4R4_UNORM_PACK16{
|
||||
|
||||
} // namespace Alternatives
|
||||
|
||||
enum class NvidiaArchitecture {
|
||||
KeplerOrOlder,
|
||||
Maxwell,
|
||||
Pascal,
|
||||
Volta,
|
||||
Turing,
|
||||
AmpereOrNewer,
|
||||
};
|
||||
|
||||
template <typename T>
|
||||
void SetNext(void**& next, T& data) {
|
||||
*next = &data;
|
||||
@ -326,9 +317,9 @@ NvidiaArchitecture GetNvidiaArchitecture(vk::PhysicalDevice physical,
|
||||
if (shading_rate_props.primitiveFragmentShadingRateWithMultipleViewports) {
|
||||
// Only Ampere and newer support this feature
|
||||
// TODO: Find a way to differentiate Ampere and Ada
|
||||
return NvidiaArchitecture::AmpereOrNewer;
|
||||
return NvidiaArchitecture::Arch_AmpereOrNewer;
|
||||
}
|
||||
return NvidiaArchitecture::Turing;
|
||||
return NvidiaArchitecture::Arch_Turing;
|
||||
}
|
||||
|
||||
if (exts.contains(VK_EXT_BLEND_OPERATION_ADVANCED_EXTENSION_NAME)) {
|
||||
@ -340,7 +331,7 @@ NvidiaArchitecture GetNvidiaArchitecture(vk::PhysicalDevice physical,
|
||||
physical_properties.pNext = &advanced_blending_props;
|
||||
physical.GetProperties2(physical_properties);
|
||||
if (advanced_blending_props.advancedBlendMaxColorAttachments == 1) {
|
||||
return NvidiaArchitecture::Maxwell;
|
||||
return NvidiaArchitecture::Arch_Maxwell;
|
||||
}
|
||||
|
||||
if (exts.contains(VK_EXT_CONSERVATIVE_RASTERIZATION_EXTENSION_NAME)) {
|
||||
@ -350,13 +341,13 @@ NvidiaArchitecture GetNvidiaArchitecture(vk::PhysicalDevice physical,
|
||||
physical_properties.pNext = &conservative_raster_props;
|
||||
physical.GetProperties2(physical_properties);
|
||||
if (conservative_raster_props.degenerateLinesRasterized) {
|
||||
return NvidiaArchitecture::Volta;
|
||||
return NvidiaArchitecture::Arch_Volta;
|
||||
}
|
||||
return NvidiaArchitecture::Pascal;
|
||||
return NvidiaArchitecture::Arch_Pascal;
|
||||
}
|
||||
}
|
||||
|
||||
return NvidiaArchitecture::KeplerOrOlder;
|
||||
return NvidiaArchitecture::Arch_KeplerOrOlder;
|
||||
}
|
||||
|
||||
std::vector<const char*> ExtensionListForVulkan(
|
||||
@ -436,6 +427,10 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR
|
||||
throw vk::Exception(VK_ERROR_INCOMPATIBLE_DRIVER);
|
||||
}
|
||||
|
||||
if (is_nvidia) {
|
||||
nvidia_arch = GetNvidiaArchitecture(physical, supported_extensions);
|
||||
}
|
||||
|
||||
SetupFamilies(surface);
|
||||
const auto queue_cis = GetDeviceQueueCreateInfos();
|
||||
|
||||
@ -532,11 +527,11 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR
|
||||
|
||||
if (is_nvidia) {
|
||||
const u32 nv_major_version = (properties.properties.driverVersion >> 22) & 0x3ff;
|
||||
const auto arch = GetNvidiaArchitecture(physical, supported_extensions);
|
||||
if (arch >= NvidiaArchitecture::AmpereOrNewer) {
|
||||
const auto arch = GetNvidiaArch();
|
||||
if (arch >= NvidiaArchitecture::Arch_AmpereOrNewer) {
|
||||
LOG_WARNING(Render_Vulkan, "Ampere and newer have broken float16 math");
|
||||
features.shader_float16_int8.shaderFloat16 = false;
|
||||
} else if (arch <= NvidiaArchitecture::Volta) {
|
||||
} else if (arch <= NvidiaArchitecture::Arch_Volta) {
|
||||
if (nv_major_version < 527) {
|
||||
LOG_WARNING(Render_Vulkan, "Volta and older have broken VK_KHR_push_descriptor");
|
||||
RemoveExtension(extensions.push_descriptor, VK_KHR_PUSH_DESCRIPTOR_EXTENSION_NAME);
|
||||
@ -686,8 +681,8 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR
|
||||
RemoveExtension(extensions.push_descriptor, VK_KHR_PUSH_DESCRIPTOR_EXTENSION_NAME);
|
||||
}
|
||||
} else if (extensions.push_descriptor && is_nvidia) {
|
||||
const auto arch = GetNvidiaArchitecture(physical, supported_extensions);
|
||||
if (arch <= NvidiaArchitecture::Pascal) {
|
||||
const auto arch = GetNvidiaArch();
|
||||
if (arch <= NvidiaArchitecture::Arch_Pascal) {
|
||||
LOG_WARNING(Render_Vulkan,
|
||||
"Pascal and older architectures have broken VK_KHR_push_descriptor");
|
||||
RemoveExtension(extensions.push_descriptor, VK_KHR_PUSH_DESCRIPTOR_EXTENSION_NAME);
|
||||
|
@ -177,6 +177,15 @@ enum class FormatType { Linear, Optimal, Buffer };
|
||||
/// Subgroup size of the guest emulated hardware (Nvidia has 32 threads per subgroup).
|
||||
const u32 GuestWarpSize = 32;
|
||||
|
||||
enum class NvidiaArchitecture {
|
||||
Arch_KeplerOrOlder,
|
||||
Arch_Maxwell,
|
||||
Arch_Pascal,
|
||||
Arch_Volta,
|
||||
Arch_Turing,
|
||||
Arch_AmpereOrNewer,
|
||||
};
|
||||
|
||||
/// Handles data specific to a physical device.
|
||||
class Device {
|
||||
public:
|
||||
@ -670,6 +679,14 @@ public:
|
||||
return false;
|
||||
}
|
||||
|
||||
bool IsNvidia() const noexcept {
|
||||
return properties.driver.driverID == VK_DRIVER_ID_NVIDIA_PROPRIETARY;
|
||||
}
|
||||
|
||||
NvidiaArchitecture GetNvidiaArch() const noexcept {
|
||||
return nvidia_arch;
|
||||
}
|
||||
|
||||
private:
|
||||
/// Checks if the physical device is suitable and configures the object state
|
||||
/// with all necessary info about its properties.
|
||||
@ -788,6 +805,7 @@ private:
|
||||
bool supports_conditional_barriers{}; ///< Allows barriers in conditional control flow.
|
||||
u64 device_access_memory{}; ///< Total size of device local memory in bytes.
|
||||
u32 sets_per_pool{}; ///< Sets per Description Pool
|
||||
NvidiaArchitecture nvidia_arch{NvidiaArchitecture::Arch_AmpereOrNewer};
|
||||
|
||||
// Telemetry parameters
|
||||
std::set<std::string, std::less<>> supported_extensions; ///< Reported Vulkan extensions.
|
||||
|
@ -114,7 +114,7 @@ const std::map<Settings::ShaderBackend, QString> Config::shader_backend_texts_ma
|
||||
// This must be in alphabetical order according to action name as it must have the same order as
|
||||
// UISetting::values.shortcuts, which is alphabetically ordered.
|
||||
// clang-format off
|
||||
const std::array<UISettings::Shortcut, 22> Config::default_hotkeys{{
|
||||
const std::array<UISettings::Shortcut, 23> Config::default_hotkeys{{
|
||||
{QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Audio Mute/Unmute")), QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Main Window")), {QStringLiteral("Ctrl+M"), QStringLiteral("Home+Dpad_Right"), Qt::WindowShortcut, false}},
|
||||
{QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Audio Volume Down")), QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Main Window")), {QStringLiteral("-"), QStringLiteral("Home+Dpad_Down"), Qt::ApplicationShortcut, true}},
|
||||
{QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Audio Volume Up")), QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Main Window")), {QStringLiteral("="), QStringLiteral("Home+Dpad_Up"), Qt::ApplicationShortcut, true}},
|
||||
@ -136,6 +136,7 @@ const std::array<UISettings::Shortcut, 22> Config::default_hotkeys{{
|
||||
{QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Toggle Filter Bar")), QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Main Window")), {QStringLiteral("Ctrl+F"), QStringLiteral(""), Qt::WindowShortcut, false}},
|
||||
{QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Toggle Framerate Limit")), QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Main Window")), {QStringLiteral("Ctrl+U"), QStringLiteral("Home+Y"), Qt::ApplicationShortcut, false}},
|
||||
{QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Toggle Mouse Panning")), QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Main Window")), {QStringLiteral("Ctrl+F9"), QStringLiteral(""), Qt::ApplicationShortcut, false}},
|
||||
{QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Toggle Renderdoc Capture")), QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Main Window")), {QStringLiteral(""), QStringLiteral(""), Qt::ApplicationShortcut, false}},
|
||||
{QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Toggle Status Bar")), QStringLiteral(QT_TRANSLATE_NOOP("Hotkeys", "Main Window")), {QStringLiteral("Ctrl+S"), QStringLiteral(""), Qt::WindowShortcut, false}},
|
||||
}};
|
||||
// clang-format on
|
||||
|
@ -48,7 +48,7 @@ public:
|
||||
default_mouse_buttons;
|
||||
static const std::array<int, Settings::NativeKeyboard::NumKeyboardKeys> default_keyboard_keys;
|
||||
static const std::array<int, Settings::NativeKeyboard::NumKeyboardMods> default_keyboard_mods;
|
||||
static const std::array<UISettings::Shortcut, 22> default_hotkeys;
|
||||
static const std::array<UISettings::Shortcut, 23> default_hotkeys;
|
||||
|
||||
static const std::map<Settings::AntiAliasing, QString> anti_aliasing_texts_map;
|
||||
static const std::map<Settings::ScalingFilter, QString> scaling_filter_texts_map;
|
||||
|
@ -319,6 +319,13 @@ void ConfigureHotkeys::ApplyConfiguration(HotkeyRegistry& registry) {
|
||||
void ConfigureHotkeys::RestoreDefaults() {
|
||||
for (int r = 0; r < model->rowCount(); ++r) {
|
||||
const QStandardItem* parent = model->item(r, 0);
|
||||
const int hotkey_size = static_cast<int>(Config::default_hotkeys.size());
|
||||
|
||||
if (hotkey_size != parent->rowCount()) {
|
||||
QMessageBox::warning(this, tr("Invalid hotkey settings"),
|
||||
tr("An error occurred. Please report this issue on github."));
|
||||
return;
|
||||
}
|
||||
|
||||
for (int r2 = 0; r2 < parent->rowCount(); ++r2) {
|
||||
model->item(r, 0)
|
||||
|
@ -89,7 +89,7 @@ void ConfigureVibration::VibrateController(Core::HID::ControllerTriggerType type
|
||||
|
||||
auto& player = Settings::values.players.GetValue()[player_index];
|
||||
auto controller = hid_core.GetEmulatedControllerByIndex(player_index);
|
||||
const int vibration_strenght = vibration_spinboxes[player_index]->value();
|
||||
const int vibration_strength = vibration_spinboxes[player_index]->value();
|
||||
const auto& buttons = controller->GetButtonsValues();
|
||||
|
||||
bool button_is_pressed = false;
|
||||
@ -105,10 +105,10 @@ void ConfigureVibration::VibrateController(Core::HID::ControllerTriggerType type
|
||||
return;
|
||||
}
|
||||
|
||||
const int old_vibration_enabled = player.vibration_enabled;
|
||||
const bool old_vibration_strenght = player.vibration_strength;
|
||||
const bool old_vibration_enabled = player.vibration_enabled;
|
||||
const int old_vibration_strength = player.vibration_strength;
|
||||
player.vibration_enabled = true;
|
||||
player.vibration_strength = vibration_strenght;
|
||||
player.vibration_strength = vibration_strength;
|
||||
|
||||
const Core::HID::VibrationValue vibration{
|
||||
.low_amplitude = 1.0f,
|
||||
@ -121,7 +121,7 @@ void ConfigureVibration::VibrateController(Core::HID::ControllerTriggerType type
|
||||
|
||||
// Restore previous values
|
||||
player.vibration_enabled = old_vibration_enabled;
|
||||
player.vibration_strength = old_vibration_strenght;
|
||||
player.vibration_strength = old_vibration_strength;
|
||||
}
|
||||
|
||||
void ConfigureVibration::StopVibrations() {
|
||||
|
Reference in New Issue
Block a user