Compare commits

..

13 Commits

Author SHA1 Message Date
lizzie 16b55f5802 bleh 2026-06-28 06:06:42 +00:00
lizzie 44f6229b26 sync with master (2/2) 2026-06-28 05:43:58 +00:00
lizzie bb66fbea51 wow, this is fuckery of abi breakage (1/2) 2026-06-28 05:41:31 +00:00
CamilleLaVey 97cc97b933 smol change 2026-06-28 00:16:56 -04:00
CamilleLaVey 5d6897edac [texture_cache] Replace LRU index with frame tick in ImageBase + update garbage collection logic 2026-06-27 23:58:02 -04:00
CamilleLaVey 8217304b60 I got meowed by Gidoly 2026-06-27 23:18:19 -04:00
CamilleLaVey 689f4d3214 [buffer_cache] Removal of LRU inside buffer cache and replaced with tick operations inside frames. 2026-06-27 23:16:14 -04:00
CamilleLaVey bdbdfe8e1d [maxwell] Refactor execution mask initialization to use fill() instead of reset() 2026-06-27 23:15:12 -04:00
CamilleLaVey d25d460116 Revert "small test to msaa copy texture behavior" 2026-06-28 05:07:54 +02:00
CamilleLaVey 06c8f9806b small test to msaa copy texture behavior 2026-06-28 05:07:54 +02:00
CamilleLaVey aa012961b8 Another test 2026-06-28 05:07:54 +02:00
CamilleLaVey d628737b48 [TEST] Review the concurrency of resolve compute + barriers operations 2026-06-28 05:07:54 +02:00
CamilleLaVey d5cc29a2f7 [query_cache] Reduction of synchronization within queries 2026-06-28 05:07:54 +02:00
72 changed files with 586 additions and 633 deletions
+5 -1
View File
@@ -4,8 +4,12 @@
// SPDX-FileCopyrightText: 2019 Dolphin Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
#include <cstring>
#include <string>
#ifndef _WIN32
#include <cstring>
#endif
#include <utility>
#include <fmt/ranges.h>
-142
View File
@@ -1,142 +0,0 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2021 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
#pragma once
#include <deque>
#include <memory>
#include <type_traits>
#include "common/common_types.h"
namespace Common {
template <class Traits>
class LeastRecentlyUsedCache {
using ObjectType = typename Traits::ObjectType;
using TickType = typename Traits::TickType;
struct Item {
ObjectType obj;
TickType tick;
Item* next{};
Item* prev{};
};
public:
LeastRecentlyUsedCache() : first_item{}, last_item{} {}
~LeastRecentlyUsedCache() = default;
size_t Insert(ObjectType obj, TickType tick) {
const auto new_id = Build();
auto& item = item_pool[new_id];
item.obj = obj;
item.tick = tick;
Attach(item);
return new_id;
}
void Touch(size_t id, TickType tick) {
auto& item = item_pool[id];
if (item.tick >= tick) {
return;
}
item.tick = tick;
if (&item == last_item) {
return;
}
Detach(item);
Attach(item);
}
void Free(size_t id) {
auto& item = item_pool[id];
Detach(item);
item.prev = nullptr;
item.next = nullptr;
free_items.push_back(id);
}
template <typename Func>
void ForEachItemBelow(TickType tick, Func&& func) {
static constexpr bool RETURNS_BOOL =
std::is_same_v<std::invoke_result_t<Func, ObjectType>, bool>;
Item* iterator = first_item;
while (iterator) {
if (static_cast<s64>(tick) - static_cast<s64>(iterator->tick) < 0) {
return;
}
Item* next = iterator->next;
if constexpr (RETURNS_BOOL) {
if (func(iterator->obj)) {
return;
}
} else {
func(iterator->obj);
}
iterator = next;
}
}
private:
size_t Build() {
if (free_items.empty()) {
const size_t item_id = item_pool.size();
auto& item = item_pool.emplace_back();
item.next = nullptr;
item.prev = nullptr;
return item_id;
}
const size_t item_id = free_items.front();
free_items.pop_front();
auto& item = item_pool[item_id];
item.next = nullptr;
item.prev = nullptr;
return item_id;
}
void Attach(Item& item) {
if (!first_item) {
first_item = &item;
}
if (!last_item) {
last_item = &item;
} else {
item.prev = last_item;
last_item->next = &item;
item.next = nullptr;
last_item = &item;
}
}
void Detach(Item& item) {
if (item.prev) {
item.prev->next = item.next;
}
if (item.next) {
item.next->prev = item.prev;
}
if (&item == first_item) {
first_item = item.next;
if (first_item) {
first_item->prev = nullptr;
}
}
if (&item == last_item) {
last_item = item.prev;
if (last_item) {
last_item->next = nullptr;
}
}
}
std::deque<Item> item_pool;
std::deque<size_t> free_items;
Item* first_item{};
Item* last_item{};
};
} // namespace Common
+6 -6
View File
@@ -729,7 +729,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface(ctx, ensure_token_id);
rb.PushIpcInterface(ensure_token_id);
}
void LoadIdTokenCacheDeprecated(HLERequestContext& ctx) {
@@ -921,7 +921,7 @@ void Module::Interface::GetProfile(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IProfile>(ctx, system, user_id, *profile_manager);
rb.PushIpcInterface<IProfile>(system, user_id, *profile_manager);
}
void Module::Interface::IsUserRegistrationRequestPermitted(HLERequestContext& ctx) {
@@ -993,7 +993,7 @@ void Module::Interface::GetBaasAccountManagerForApplication(HLERequestContext& c
LOG_DEBUG(Service_ACC, "called");
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IManagerForApplication>(ctx, system, profile_manager);
rb.PushIpcInterface<IManagerForApplication>(system, profile_manager);
}
void Module::Interface::IsUserAccountSwitchLocked(HLERequestContext& ctx) {
@@ -1089,7 +1089,7 @@ void Module::Interface::GetProfileEditor(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IProfileEditor>(ctx, system, user_id, *profile_manager);
rb.PushIpcInterface<IProfileEditor>(system, user_id, *profile_manager);
}
void Module::Interface::GetBaasAccountAdministrator(HLERequestContext &ctx) {
@@ -1100,7 +1100,7 @@ void Module::Interface::GetBaasAccountAdministrator(HLERequestContext &ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IAdministrator>(ctx, system, uuid);
rb.PushIpcInterface<IAdministrator>(system, uuid);
}
void Module::Interface::ListQualifiedUsers(HLERequestContext& ctx) {
@@ -1143,7 +1143,7 @@ void Module::Interface::GetBaasAccountManagerForSystemService(HLERequestContext&
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IManagerForSystemService>(ctx, system, uuid);
rb.PushIpcInterface<IManagerForSystemService>(system, uuid);
}
void Module::Interface::StoreSaveDataThumbnailSystem(HLERequestContext& ctx) {
+1 -1
View File
@@ -35,7 +35,7 @@ void IAsyncContext::GetSystemEvent(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, completion_event->GetReadableEvent());
rb.PushCopyObjects(completion_event->GetReadableEvent());
}
void IAsyncContext::Cancel(HLERequestContext& ctx) {
+2 -2
View File
@@ -82,7 +82,7 @@ void APM::OpenSession(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ISession>(ctx, system, controller);
rb.PushIpcInterface<ISession>(system, controller);
}
void APM::GetPerformanceMode(HLERequestContext& ctx) {
@@ -125,7 +125,7 @@ void APM_Sys::GetPerformanceEvent(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ISession>(ctx, system, controller);
rb.PushIpcInterface<ISession>(system, controller);
}
void APM_Sys::SetCpuBoostMode(HLERequestContext& ctx) {
+1 -4
View File
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2018 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -48,7 +45,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IRequest>(ctx, system);
rb.PushIpcInterface<IRequest>(system);
}
};
+7 -7
View File
@@ -184,7 +184,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(readable_event.Signal(system.Kernel()));
rb.PushCopyObjects(ctx, readable_event);
rb.PushCopyObjects(readable_event);
}
void Cancel(HLERequestContext& ctx) {
@@ -400,7 +400,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, notification_event->GetReadableEvent());
rb.PushCopyObjects(notification_event->GetReadableEvent());
}
void Clear(HLERequestContext& ctx) {
@@ -476,7 +476,7 @@ private:
void Module::Interface::CreateFriendService(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IFriendService>(ctx, system);
rb.PushIpcInterface<IFriendService>(system);
LOG_DEBUG(Service_Friend, "called");
}
@@ -488,12 +488,12 @@ void Module::Interface::CreateNotificationService(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<INotificationService>(ctx, system, uuid);
rb.PushIpcInterface<INotificationService>(system, uuid);
}
Module::Interface::Interface(std::shared_ptr<Module> module_, Core::System& system_, const char* name)
: ServiceFramework{system_, name}, module{std::move(module_)}
{}
Module::Interface::Interface(std::shared_ptr<Module> module_, Core::System& system_,
const char* name)
: ServiceFramework{system_, name}, module{std::move(module_)} {}
Module::Interface::~Interface() = default;
+1 -1
View File
@@ -279,7 +279,7 @@ void ARP_W::AcquireRegistrar(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface(ctx, registrar);
rb.PushIpcInterface(registrar);
}
void ARP_W::UnregisterApplicationInstance(HLERequestContext& ctx) {
+1 -1
View File
@@ -28,7 +28,7 @@ void BGTC_T::OpenTaskService(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ITaskService>(ctx, system);
rb.PushIpcInterface<ITaskService>(system);
}
ITaskService::ITaskService(Core::System& system_) : ServiceFramework{system_, "ITaskService"} {
+1 -4
View File
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2021 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -59,7 +56,7 @@ ECTX_AW::~ECTX_AW() = default;
void ECTX_AW::CreateContextRegistrar(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IContextRegistrar>(ctx, system);
rb.PushIpcInterface<IContextRegistrar>(std::make_shared<IContextRegistrar>(system));
}
} // namespace Service::Glue
@@ -726,7 +726,7 @@ void IHidSystemServer::AcquireConnectionTriggerTimeoutEvent(HLERequestContext& c
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, acquire_device_registered_event->GetReadableEvent());
rb.PushCopyObjects(acquire_device_registered_event->GetReadableEvent());
}
void IHidSystemServer::AcquireDeviceRegisteredEventForControllerSupport(HLERequestContext& ctx) {
@@ -734,7 +734,7 @@ void IHidSystemServer::AcquireDeviceRegisteredEventForControllerSupport(HLEReque
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, acquire_device_registered_event->GetReadableEvent());
rb.PushCopyObjects(acquire_device_registered_event->GetReadableEvent());
}
void IHidSystemServer::GetRegisteredDevices(HLERequestContext& ctx) {
@@ -759,7 +759,7 @@ void IHidSystemServer::AcquireUniquePadConnectionEventHandle(HLERequestContext&
LOG_WARNING(Service_HID, "(STUBBED) called");
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.PushCopyObjects(ctx, unique_pad_connection_event->GetReadableEvent());
rb.PushCopyObjects(unique_pad_connection_event->GetReadableEvent());
rb.Push(ResultSuccess);
}
@@ -776,7 +776,7 @@ void IHidSystemServer::AcquireJoyDetachOnBluetoothOffEventHandle(HLERequestConte
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, joy_detach_event->GetReadableEvent());
rb.PushCopyObjects(joy_detach_event->GetReadableEvent());
}
void IHidSystemServer::IsUsbFullKeyControllerEnabled(HLERequestContext& ctx) {
+2 -2
View File
@@ -31,7 +31,7 @@ class Memory;
}
namespace IPC {
struct ResponseBuilder;
class ResponseBuilder;
}
namespace Service {
@@ -392,7 +392,7 @@ public:
}
private:
friend struct IPC::ResponseBuilder;
friend class IPC::ResponseBuilder;
void ParseCommandBuffer(u32_le* src_cmdbuf, bool incoming);
+153 -91
View File
@@ -24,7 +24,45 @@ namespace IPC {
constexpr Result ResultSessionClosed{ErrorModule::HIPC, 301};
struct ResponseBuilder {
class RequestHelperBase {
protected:
Service::HLERequestContext* context = nullptr;
u32* cmdbuf;
u32 index = 0;
public:
explicit RequestHelperBase(u32* command_buffer) : cmdbuf(command_buffer) {}
explicit RequestHelperBase(Service::HLERequestContext& ctx)
: context(&ctx), cmdbuf(ctx.CommandBuffer()) {}
void Skip(u32 size_in_words, bool set_to_null) {
if (set_to_null) {
memset(cmdbuf + index, 0, size_in_words * sizeof(u32));
}
index += size_in_words;
}
/**
* Aligns the current position forward to a 16-byte boundary, padding with zeros.
*/
void AlignWithPadding() {
if (index & 3) {
Skip(static_cast<u32>(4 - (index & 3)), true);
}
}
u32 GetCurrentOffset() const {
return index;
}
void SetCurrentOffset(u32 offset) {
index = offset;
}
};
class ResponseBuilder : public RequestHelperBase {
public:
/// Flags used for customizing the behavior of ResponseBuilder
enum class Flags : u32 {
None = 0,
@@ -33,13 +71,14 @@ struct ResponseBuilder {
AlwaysMoveHandles = 1,
};
inline explicit ResponseBuilder(Service::HLERequestContext& ctx, u32 normal_params_size_, u32 num_handles_to_copy_ = 0, u32 num_objects_to_move_ = 0, Flags flags = Flags::None)
: cmdbuf(ctx.CommandBuffer())
, normal_params_size(normal_params_size_)
, num_handles_to_copy(num_handles_to_copy_)
, num_objects_to_move(num_objects_to_move_)
{
std::memset(cmdbuf, 0, sizeof(u32) * IPC::COMMAND_BUFFER_LENGTH);
explicit ResponseBuilder(Service::HLERequestContext& ctx, u32 normal_params_size_,
u32 num_handles_to_copy_ = 0, u32 num_objects_to_move_ = 0,
Flags flags = Flags::None)
: RequestHelperBase(ctx), normal_params_size(normal_params_size_),
num_handles_to_copy(num_handles_to_copy_),
num_objects_to_move(num_objects_to_move_), kernel{ctx.kernel} {
memset(cmdbuf, 0, sizeof(u32) * IPC::COMMAND_BUFFER_LENGTH);
IPC::CommandHeader header{};
auto const mgr = ctx.GetManager().get();
@@ -78,7 +117,9 @@ struct ResponseBuilder {
handle_descriptor_header.num_handles_to_copy.Assign(num_handles_to_copy_);
handle_descriptor_header.num_handles_to_move.Assign(num_handles_to_move);
PushRaw(handle_descriptor_header);
ctx.handles_offset = index;
Skip(num_handles_to_copy + num_handles_to_move, true);
}
@@ -90,6 +131,7 @@ struct ResponseBuilder {
domain_header.num_objects = num_domain_objects;
PushRaw(domain_header);
}
IPC::DataPayloadHeader data_payload_header{};
data_payload_header.magic = Common::MakeMagic('S', 'F', 'C', 'O');
PushRaw(data_payload_header);
@@ -99,39 +141,34 @@ struct ResponseBuilder {
ctx.data_payload_offset = index;
ctx.write_size += index;
ctx.domain_offset = u32(index + raw_data_size / sizeof(u32));
ctx.domain_offset = static_cast<u32>(index + raw_data_size / sizeof(u32));
}
inline void Skip(u32 size_in_words, bool set_to_null) {
if (set_to_null) std::memset(cmdbuf + index, 0, size_in_words * sizeof(u32));
index += size_in_words;
}
/// @brief Aligns the current position forward to a 16-byte boundary, padding with zeros.
inline void AlignWithPadding() { if (index & 3) Skip(u32(4 - (index & 3)), true); }
inline u32 GetCurrentOffset() const { return index; }
inline void SetCurrentOffset(u32 offset) { index = offset; }
template <class T>
void PushIpcInterface(std::shared_ptr<T> iface) {
auto manager{context->GetManager()};
template <class T> inline void PushIpcInterface(Service::HLERequestContext& ctx, std::shared_ptr<T> iface) {
auto manager = ctx.GetManager();
if (manager->IsDomain()) {
ctx.AddDomainObject(std::move(iface));
context->AddDomainObject(std::move(iface));
} else {
ASSERT(Kernel::GetCurrentProcess(ctx.kernel).GetResourceLimit()->Reserve(ctx.kernel, Kernel::LimitableResource::SessionCountMax, 1));
ASSERT(Kernel::GetCurrentProcess(kernel).GetResourceLimit()->Reserve(kernel, Kernel::LimitableResource::SessionCountMax, 1));
auto* session = Kernel::KSession::Create(ctx.kernel);
session->Initialize(ctx.kernel, nullptr, 0);
Kernel::KSession::Register(ctx.kernel, session);
auto* session = Kernel::KSession::Create(kernel);
session->Initialize(kernel, nullptr, 0);
Kernel::KSession::Register(kernel, session);
auto next_manager = std::make_shared<Service::SessionRequestManager>(ctx.kernel, manager->GetServerManager());
auto next_manager = std::make_shared<Service::SessionRequestManager>(
kernel, manager->GetServerManager());
next_manager->SetSessionHandler(iface);
manager->GetServerManager().RegisterSession(&session->GetServerSession(), next_manager);
ctx.AddMoveObject(&session->GetClientSession());
context->AddMoveObject(&session->GetClientSession());
}
}
template <class T, class... Args> inline void PushIpcInterface(Service::HLERequestContext& ctx, Args&&... args) {
PushIpcInterface<T>(ctx, std::make_shared<T>(std::forward<Args>(args)...));
template <class T, class... Args>
void PushIpcInterface(Args&&... args) {
PushIpcInterface<T>(std::make_shared<T>(std::forward<Args>(args)...));
}
void PushImpl(s8 value);
@@ -147,38 +184,59 @@ struct ResponseBuilder {
void PushImpl(bool value);
void PushImpl(Result value);
template <typename T> inline void Push(T value) {
template <typename T>
void Push(T value) {
return PushImpl(value);
}
template <typename First, typename... Other>
void Push(const First& first_value, const Other&... other_values);
/// @brief Helper function for pushing strongly-typed enumeration values.
/// @tparam Enum The enumeration type to be pushed
/// @param value The value to push.
/// @note The underlying size of the enumeration type is the size of the data that gets pushed.
/// e.g. "enum class SomeEnum : u16" will push a u16-sized amount of data.
template <typename Enum> inline void PushEnum(Enum value) {
/**
* Helper function for pushing strongly-typed enumeration values.
*
* @tparam Enum The enumeration type to be pushed
*
* @param value The value to push.
*
* @note The underlying size of the enumeration type is the size of the
* data that gets pushed. e.g. "enum class SomeEnum : u16" will
* push a u16-sized amount of data.
*/
template <typename Enum>
void PushEnum(Enum value) {
static_assert(std::is_enum_v<Enum>, "T must be an enum type within a PushEnum call.");
static_assert(!std::is_convertible_v<Enum, int>, "enum type in PushEnum must be a strongly typed enum.");
static_assert(!std::is_convertible_v<Enum, int>,
"enum type in PushEnum must be a strongly typed enum.");
Push(static_cast<std::underlying_type_t<Enum>>(value));
}
/// @brief Copies the content of the given trivially copyable class to the buffer as a normal param
/// @note: The input class must be correctly packed/padded to fit hardware layout.
template <typename T> void PushRaw(const T& value);
template <typename... O> void PushMoveObjects(Service::HLERequestContext& ctx, O*... pointers);
template <typename... O> void PushMoveObjects(Service::HLERequestContext& ctx, O&... pointers);
template <typename... O> void PushCopyObjects(Service::HLERequestContext& ctx, O*... pointers);
template <typename... O> void PushCopyObjects(Service::HLERequestContext& ctx, O&... pointers);
/**
* @brief Copies the content of the given trivially copyable class to the buffer as a normal
* param
* @note: The input class must be correctly packed/padded to fit hardware layout.
*/
template <typename T>
void PushRaw(const T& value);
u32* cmdbuf;
u32 index = 0;
template <typename... O>
void PushMoveObjects(O*... pointers);
template <typename... O>
void PushMoveObjects(O&... pointers);
template <typename... O>
void PushCopyObjects(O*... pointers);
template <typename... O>
void PushCopyObjects(O&... pointers);
private:
u32 normal_params_size{};
u32 num_handles_to_copy{};
u32 num_objects_to_move{}; ///< Domain objects or move handles, context dependent
u32 data_payload_index{};
Kernel::KernelCore& kernel;
};
/// Push ///
@@ -193,7 +251,8 @@ inline void ResponseBuilder::PushImpl(u32 value) {
template <typename T>
void ResponseBuilder::PushRaw(const T& value) {
static_assert(std::is_trivially_copyable_v<T>, "It's undefined behavior to use memcpy with non-trivially copyable objects");
static_assert(std::is_trivially_copyable_v<T>,
"It's undefined behavior to use memcpy with non-trivially copyable objects");
std::memcpy(cmdbuf + index, &value, sizeof(T));
index += (sizeof(T) + 3) / 4; // round up to word length
}
@@ -213,8 +272,8 @@ inline void ResponseBuilder::PushImpl(s16 value) {
}
inline void ResponseBuilder::PushImpl(s64 value) {
PushImpl(u32(value));
PushImpl(u32(value >> 32));
PushImpl(static_cast<u32>(value));
PushImpl(static_cast<u32>(value >> 32));
}
inline void ResponseBuilder::PushImpl(u8 value) {
@@ -226,8 +285,8 @@ inline void ResponseBuilder::PushImpl(u16 value) {
}
inline void ResponseBuilder::PushImpl(u64 value) {
PushImpl(u32(value));
PushImpl(u32(value >> 32));
PushImpl(static_cast<u32>(value));
PushImpl(static_cast<u32>(value >> 32));
}
inline void ResponseBuilder::PushImpl(float value) {
@@ -253,88 +312,90 @@ void ResponseBuilder::Push(const First& first_value, const Other&... other_value
}
template <typename... O>
inline void ResponseBuilder::PushCopyObjects(Service::HLERequestContext& ctx, O*... pointers) {
inline void ResponseBuilder::PushCopyObjects(O*... pointers) {
auto objects = {pointers...};
for (auto& object : objects) {
ctx.AddCopyObject(object);
context->AddCopyObject(object);
}
}
template <typename... O>
inline void ResponseBuilder::PushCopyObjects(Service::HLERequestContext& ctx, O&... pointers) {
inline void ResponseBuilder::PushCopyObjects(O&... pointers) {
auto objects = {&pointers...};
for (auto& object : objects) {
ctx.AddCopyObject(object);
context->AddCopyObject(object);
}
}
template <typename... O>
inline void ResponseBuilder::PushMoveObjects(Service::HLERequestContext& ctx, O*... pointers) {
inline void ResponseBuilder::PushMoveObjects(O*... pointers) {
auto objects = {pointers...};
for (auto& object : objects) {
ctx.AddMoveObject(object);
context->AddMoveObject(object);
}
}
template <typename... O>
inline void ResponseBuilder::PushMoveObjects(Service::HLERequestContext& ctx, O&... pointers) {
inline void ResponseBuilder::PushMoveObjects(O&... pointers) {
auto objects = {&pointers...};
for (auto& object : objects) {
ctx.AddMoveObject(object);
context->AddMoveObject(object);
}
}
struct RequestParser {
inline explicit RequestParser(u32* command_buffer) : cmdbuf(command_buffer) {}
inline explicit RequestParser(Service::HLERequestContext& ctx)
: cmdbuf(ctx.CommandBuffer())
{
class RequestParser : public RequestHelperBase {
public:
explicit RequestParser(u32* command_buffer) : RequestHelperBase(command_buffer) {}
explicit RequestParser(Service::HLERequestContext& ctx) : RequestHelperBase(ctx) {
// TIPC does not have data payload offset
if (!ctx.IsTipc()) {
ASSERT_MSG(ctx.GetDataPayloadOffset(), "context is incomplete");
Skip(ctx.GetDataPayloadOffset(), false);
}
// Skip the u64 command id, it's already stored in the context
static constexpr u32 CommandIdSize = 2;
Skip(CommandIdSize, false);
}
inline void Skip(u32 size_in_words, bool set_to_null) {
if (set_to_null) std::memset(cmdbuf + index, 0, size_in_words * sizeof(u32));
index += size_in_words;
}
/// @brief Aligns the current position forward to a 16-byte boundary, padding with zeros.
inline void AlignWithPadding() { if (index & 3) Skip(u32(4 - (index & 3)), true); }
inline u32 GetCurrentOffset() const { return index; }
inline void SetCurrentOffset(u32 offset) { index = offset; }
template <typename T>
T Pop();
template <typename T> T Pop();
template <typename T> void Pop(T& value);
template <typename First, typename... Other> void Pop(First& first_value, Other&... other_values);
template <typename T>
void Pop(T& value);
template <typename First, typename... Other>
void Pop(First& first_value, Other&... other_values);
template <typename T>
T PopEnum() {
static_assert(std::is_enum_v<T>, "T must be an enum type within a PopEnum call.");
static_assert(!std::is_convertible_v<T, int>, "enum type in PopEnum must be a strongly typed enum.");
return T(Pop<std::underlying_type_t<T>>());
static_assert(!std::is_convertible_v<T, int>,
"enum type in PopEnum must be a strongly typed enum.");
return static_cast<T>(Pop<std::underlying_type_t<T>>());
}
/// @brief Reads the next normal parameters as a struct, by copying it
/// @note: The output class must be correctly packed/padded to fit hardware layout.
template <typename T> void PopRaw(T& value);
/**
* @brief Reads the next normal parameters as a struct, by copying it
* @note: The output class must be correctly packed/padded to fit hardware layout.
*/
template <typename T>
void PopRaw(T& value);
/// @brief Reads the next normal parameters as a struct, by copying it into a new value
/// @note: The output class must be correctly packed/padded to fit hardware layout.
template <typename T> T PopRaw();
/**
* @brief Reads the next normal parameters as a struct, by copying it into a new value
* @note: The output class must be correctly packed/padded to fit hardware layout.
*/
template <typename T>
T PopRaw();
template <class T> [[nodiscard]] std::weak_ptr<T> PopIpcInterface(Service::HLERequestContext& ctx) {
ASSERT(ctx.GetManager()->IsDomain());
ASSERT(ctx.GetDomainMessageHeader().input_object_count > 0);
return ctx.GetDomainHandler<T>(Pop<u32>() - 1);
template <class T>
std::weak_ptr<T> PopIpcInterface() {
ASSERT(context->GetManager()->IsDomain());
ASSERT(context->GetDomainMessageHeader().input_object_count > 0);
return context->GetDomainHandler<T>(Pop<u32>() - 1);
}
u32* cmdbuf;
u32 index = 0;
};
/// Pop ///
@@ -346,7 +407,7 @@ inline u32 RequestParser::Pop() {
template <>
inline s32 RequestParser::Pop() {
return s32(Pop<u32>());
return static_cast<s32>(Pop<u32>());
}
// Ignore the -Wclass-memaccess warning on memcpy for non-trivially default constructible objects.
@@ -356,7 +417,8 @@ inline s32 RequestParser::Pop() {
#endif
template <typename T>
void RequestParser::PopRaw(T& value) {
static_assert(std::is_trivially_copyable_v<T>, "It's undefined behavior to use memcpy with non-trivially copyable objects");
static_assert(std::is_trivially_copyable_v<T>,
"It's undefined behavior to use memcpy with non-trivially copyable objects");
std::memcpy(&value, cmdbuf + index, sizeof(T));
index += (sizeof(T) + 3) / 4; // round up to word length
}
+1 -1
View File
@@ -353,7 +353,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ILogger>(ctx, system);
rb.PushIpcInterface<ILogger>(system);
}
};
+4 -4
View File
@@ -151,7 +151,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IAm>(ctx, system);
rb.PushIpcInterface<IAm>(system);
}
};
@@ -173,7 +173,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<MFIUser>(ctx, system);
rb.PushIpcInterface<MFIUser>(system);
}
};
@@ -195,7 +195,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IUser>(ctx, system);
rb.PushIpcInterface<IUser>(system);
}
};
@@ -217,7 +217,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ISystem>(ctx, system);
rb.PushIpcInterface<ISystem>(system);
}
};
+3 -3
View File
@@ -143,7 +143,7 @@ void NfcInterface::AttachAvailabilityChangeEvent(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, GetManager()->AttachAvailabilityChangeEvent());
rb.PushCopyObjects(GetManager()->AttachAvailabilityChangeEvent());
}
void NfcInterface::StartDetection(HLERequestContext& ctx) {
@@ -203,7 +203,7 @@ void NfcInterface::AttachActivateEvent(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(result);
rb.PushCopyObjects(ctx, out_event);
rb.PushCopyObjects(out_event);
}
void NfcInterface::AttachDeactivateEvent(HLERequestContext& ctx) {
@@ -217,7 +217,7 @@ void NfcInterface::AttachDeactivateEvent(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(result);
rb.PushCopyObjects(ctx, out_event);
rb.PushCopyObjects(out_event);
}
void NfcInterface::SetNfcEnabled(HLERequestContext& ctx) {
+3 -3
View File
@@ -157,7 +157,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IUser>(ctx, system);
rb.PushIpcInterface<IUser>(system);
}
};
@@ -179,7 +179,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ISystem>(ctx, system);
rb.PushIpcInterface<ISystem>(system);
}
};
@@ -201,7 +201,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IDebug>(ctx, system);
rb.PushIpcInterface<IDebug>(system);
}
};
+8 -8
View File
@@ -307,7 +307,7 @@ private:
void GetSystemEventReadableHandle(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 2};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, evt_scan_complete->GetReadableEvent(),
rb.PushCopyObjects(evt_scan_complete->GetReadableEvent(),
evt_processing->GetReadableEvent());
}
@@ -452,7 +452,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 2};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, event1->GetReadableEvent(), event2->GetReadableEvent());
rb.PushCopyObjects(event1->GetReadableEvent(), event2->GetReadableEvent());
}
void Cancel(HLERequestContext& ctx) {
@@ -528,7 +528,7 @@ void IGeneralService::CreateScanRequest(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IScanRequest>(ctx, system);
rb.PushIpcInterface<IScanRequest>(system);
}
void IGeneralService::CreateRequest(HLERequestContext& ctx) {
@@ -537,7 +537,7 @@ void IGeneralService::CreateRequest(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IRequest>(ctx, system);
rb.PushIpcInterface<IRequest>(system);
}
void IGeneralService::GetCurrentNetworkProfile(HLERequestContext& ctx) {
@@ -716,7 +716,7 @@ void IGeneralService::GetNetworkProfile(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(ResultSuccess);
rb.PushIpcInterface<INetworkProfile>(ctx, system);
rb.PushIpcInterface<INetworkProfile>(system);
}
void IGeneralService::SetNetworkProfile(HLERequestContext& ctx) {
@@ -869,7 +869,7 @@ void IGeneralService::CreateTemporaryNetworkProfile(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 6, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<INetworkProfile>(ctx, system);
rb.PushIpcInterface<INetworkProfile>(system);
rb.PushRaw<u128>(uuid);
}
@@ -1124,7 +1124,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IGeneralService>(ctx, system);
rb.PushIpcInterface<IGeneralService>(system);
}
void CreateGeneralService(HLERequestContext& ctx) {
@@ -1132,7 +1132,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IGeneralService>(ctx, system);
rb.PushIpcInterface<IGeneralService>(system);
}
};
+6 -6
View File
@@ -53,7 +53,7 @@ private:
LOG_WARNING(Service_NIM, "(STUBBED) called");
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IShopServiceAsync>(ctx, system);
rb.PushIpcInterface<IShopServiceAsync>(system);
}
};
@@ -75,7 +75,7 @@ private:
LOG_WARNING(Service_NIM, "(STUBBED) called");
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IShopServiceAccessor>(ctx, system);
rb.PushIpcInterface<IShopServiceAccessor>(system);
}
};
@@ -336,7 +336,7 @@ private:
LOG_DEBUG(Service_NIM, "(STUBBED) called");
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IShopServiceAccessServer>(ctx, system);
rb.PushIpcInterface<IShopServiceAccessServer>(system);
}
void IsLargeResourceAvailable(HLERequestContext& ctx) {
@@ -356,7 +356,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IShopServiceAccessServer>(ctx, system);
rb.PushIpcInterface<IShopServiceAccessServer>(system);
}
};
@@ -439,7 +439,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, finished_event->GetReadableEvent());
rb.PushCopyObjects(finished_event->GetReadableEvent());
}
void GetResult(HLERequestContext& ctx) {
@@ -500,7 +500,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IEnsureNetworkClockAvailabilityService>(ctx, system);
rb.PushIpcInterface<IEnsureNetworkClockAvailabilityService>(system);
}
// TODO(ogniK): Do we need these?
@@ -359,8 +359,8 @@ void IReadOnlyApplicationControlDataInterface::ListApplicationTitle(HLERequestCo
IPC::ResponseBuilder rb{ctx, 2, 1, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, async_value->ReadableEvent());
rb.PushIpcInterface(ctx, std::move(async_value));
rb.PushCopyObjects(async_value->ReadableEvent());
rb.PushIpcInterface(std::move(async_value));
}
Result IReadOnlyApplicationControlDataInterface::GetApplicationControlData3(
@@ -199,7 +199,7 @@ void NVDRV::QueryEvent(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 3, 1};
rb.Push(ResultSuccess);
auto& readable_event = event->GetReadableEvent();
rb.PushCopyObjects(ctx, readable_event);
rb.PushCopyObjects(readable_event);
rb.PushEnum(NvResult::Success);
} else {
LOG_ERROR(Service_NVDRV, "Invalid event request!");
-11
View File
@@ -6,8 +6,6 @@
#pragma once
#include <utility>
#include "common/common_types.h"
namespace Core {
@@ -30,15 +28,6 @@ public:
inline explicit Process(Core::System& system) noexcept : m_system(system) {}
inline ~Process() { this->Finalize(); }
Process(const Process&) = delete;
Process& operator=(const Process&) = delete;
Process& operator=(Process&&) = delete;
inline Process(Process&& other) noexcept
: m_system(other.m_system), m_process(std::exchange(other.m_process, nullptr)),
m_main_thread_stack_size(std::exchange(other.m_main_thread_stack_size, 0)),
m_main_thread_priority(std::exchange(other.m_main_thread_priority, 0)),
m_process_started(std::exchange(other.m_process_started, false)) {}
bool Initialize(Loader::AppLoader& loader, Loader::ResultStatus& out_load_result);
void Finalize();
+1 -4
View File
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2018 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -127,7 +124,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IClkrstSession>(ctx, system, device_code);
rb.PushIpcInterface<IClkrstSession>(system, device_code);
}
};
+1 -1
View File
@@ -161,7 +161,7 @@ private:
IPC::ResponseBuilder rb{ctx, 10, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, *process);
rb.PushCopyObjects(*process);
rb.PushRaw(program_location);
rb.PushRaw(override_status);
}
+3 -3
View File
@@ -147,7 +147,7 @@ void IAlarmService::CreateWakeupAlarm(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ISteadyClockAlarm>(ctx, system, m_alarms, AlarmType::WakeupAlarm);
rb.PushIpcInterface<ISteadyClockAlarm>(system, m_alarms, AlarmType::WakeupAlarm);
}
void IAlarmService::CreateBackgroundTaskAlarm(HLERequestContext& ctx) {
@@ -155,7 +155,7 @@ void IAlarmService::CreateBackgroundTaskAlarm(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ISteadyClockAlarm>(ctx, system, m_alarms, AlarmType::BackgroundTaskAlarm);
rb.PushIpcInterface<ISteadyClockAlarm>(system, m_alarms, AlarmType::BackgroundTaskAlarm);
}
ISteadyClockAlarm::ISteadyClockAlarm(Core::System& system_, Alarms& alarms, AlarmType type)
@@ -179,7 +179,7 @@ void ISteadyClockAlarm::GetAlarmEvent(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, m_alarm.GetEventHandle());
rb.PushCopyObjects(m_alarm.GetEventHandle());
}
void ISteadyClockAlarm::Enable(HLERequestContext& ctx) {
+2 -2
View File
@@ -67,7 +67,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(ResultSuccess);
rb.PushCopyObjects(ctx, state_change_event->GetReadableEvent());
rb.PushCopyObjects(state_change_event->GetReadableEvent());
}
void UnbindStateChangeEvent(HLERequestContext& ctx) {
@@ -186,7 +186,7 @@ void PSM::OpenSession(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IPsmSession>(ctx, system);
rb.PushIpcInterface<IPsmSession>(system);
}
void PSM::GetBatteryVoltageState(HLERequestContext& ctx) {
+1 -4
View File
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
@@ -85,7 +82,7 @@ void TS::OpenSession(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ISession>(ctx, system);
rb.PushIpcInterface<ISession>(system);
}
} // namespace Service::PTM
+5 -4
View File
@@ -140,7 +140,7 @@ void SM::GetServiceCmif(HLERequestContext& ctx) {
if (result == ResultSuccess) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1, IPC::ResponseBuilder::Flags::AlwaysMoveHandles};
rb.Push(result);
rb.PushMoveObjects(ctx, client_session);
rb.PushMoveObjects(client_session);
} else {
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(result);
@@ -157,7 +157,7 @@ void SM::GetServiceTipc(HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1, IPC::ResponseBuilder::Flags::AlwaysMoveHandles};
rb.Push(result);
rb.PushMoveObjects(ctx, result == ResultSuccess ? client_session : nullptr);
rb.PushMoveObjects(result == ResultSuccess ? client_session : nullptr);
}
static std::string PopServiceName(IPC::RequestParser& rp) {
@@ -230,7 +230,8 @@ void SM::RegisterServiceImpl(HLERequestContext& ctx, std::string name, u32 max_s
max_session_count, is_light);
Kernel::KServerPort* server_port{};
if (const auto result = service_manager.RegisterService(std::addressof(server_port), name, max_session_count, nullptr);
if (const auto result = service_manager.RegisterService(std::addressof(server_port), name,
max_session_count, nullptr);
result.IsError()) {
LOG_ERROR(Service_SM, "failed to register service with error_code={:08X}", result.raw);
IPC::ResponseBuilder rb{ctx, 2};
@@ -240,7 +241,7 @@ void SM::RegisterServiceImpl(HLERequestContext& ctx, std::string name, u32 max_s
IPC::ResponseBuilder rb{ctx, 2, 0, 1, IPC::ResponseBuilder::Flags::AlwaysMoveHandles};
rb.Push(ResultSuccess);
rb.PushMoveObjects(ctx, server_port);
rb.PushMoveObjects(server_port);
}
void SM::UnregisterService(HLERequestContext& ctx) {
+1 -1
View File
@@ -60,7 +60,7 @@ void Controller::CloneCurrentObject(HLERequestContext& ctx) {
// We succeeded.
IPC::ResponseBuilder rb{ctx, 2, 0, 1, IPC::ResponseBuilder::Flags::AlwaysMoveHandles};
rb.Push(ResultSuccess);
rb.PushMoveObjects(ctx, session->GetClientSession());
rb.PushMoveObjects(session->GetClientSession());
}
void Controller::CloneCurrentObjectEx(HLERequestContext& ctx) {
+5 -3
View File
@@ -543,7 +543,8 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(res);
if (res == ResultSuccess) {
rb.PushIpcInterface<ISslConnection>(ctx, system, ssl_version, shared_data, std::move(backend));
rb.PushIpcInterface<ISslConnection>(system, ssl_version, shared_data,
std::move(backend));
}
}
@@ -623,11 +624,12 @@ private:
IPC::RequestParser rp{ctx};
const auto parameters = rp.PopRaw<Parameters>();
LOG_WARNING(Service_SSL, "(STUBBED) called, api_version={}, pid_placeholder={}", parameters.ssl_version.api_version, parameters.pid_placeholder);
LOG_WARNING(Service_SSL, "(STUBBED) called, api_version={}, pid_placeholder={}",
parameters.ssl_version.api_version, parameters.pid_placeholder);
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<ISslContext>(ctx, system, parameters.ssl_version);
rb.PushIpcInterface<ISslContext>(system, parameters.ssl_version);
}
void SetInterfaceVersion(HLERequestContext& ctx) {
+2 -2
View File
@@ -155,7 +155,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IPdSession>(ctx, system);
rb.PushIpcInterface<IPdSession>(system);
}
};
@@ -199,7 +199,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(ResultSuccess);
rb.PushIpcInterface<IPdCradleSession>(ctx, system);
rb.PushIpcInterface<IPdCradleSession>(system);
}
};
+2 -2
View File
@@ -50,8 +50,8 @@ NPad::NPad(Core::HID::HIDCore& hid_core_, KernelHelpers::ServiceContext& service
auto& controller = controller_data[aruid_index][i];
controller.device = hid_core.GetEmulatedControllerByIndex(i);
Core::HID::ControllerUpdateCallback engine_callback{
.on_change = [this, i, kernel = &hid_core.kernel](Core::HID::ControllerTriggerType type) {
ControllerUpdate(*kernel, type, i);
.on_change = [this, i](Core::HID::ControllerTriggerType type) {
ControllerUpdate(hid_core.kernel, type, i);
},
.is_npad_service = true,
};
@@ -236,11 +236,8 @@ void LowerGeometryPassthrough(const IR::Program& program, const HostTranslateInf
IR::Program TranslateProgram(ObjectPool<IR::Inst>& inst_pool, ObjectPool<IR::Block>& block_pool,
Environment& env, Flow::CFG& cfg, const HostTranslateInfo& host_info) {
HostTranslateInfo normalized_host_info{host_info};
normalized_host_info.ApplyDescriptorLimitPolicy();
IR::Program program;
program.syntax_list = BuildASL(inst_pool, block_pool, env, cfg, normalized_host_info);
program.syntax_list = BuildASL(inst_pool, block_pool, env, cfg, host_info);
program.blocks = GenerateBlocks(program.syntax_list);
program.post_order_blocks = PostOrder(program.syntax_list.front());
program.stage = env.ShaderStage();
@@ -263,9 +260,9 @@ IR::Program TranslateProgram(ObjectPool<IR::Inst>& inst_pool, ObjectPool<IR::Blo
program.info.passthrough.mask[i] = ((mask[i / 32] >> (i % 32)) & 1) == 0;
}
if (!normalized_host_info.support_geometry_shader_passthrough) {
if (!host_info.support_geometry_shader_passthrough) {
program.output_vertices = GetOutputTopologyVertices(program.output_topology);
LowerGeometryPassthrough(program, normalized_host_info);
LowerGeometryPassthrough(program, host_info);
}
}
break;
@@ -280,16 +277,16 @@ IR::Program TranslateProgram(ObjectPool<IR::Inst>& inst_pool, ObjectPool<IR::Blo
RemoveUnreachableBlocks(program);
// Replace instructions before the SSA rewrite
if (!normalized_host_info.support_float64) {
if (!host_info.support_float64) {
Optimization::LowerFp64ToFp32(program);
}
if (!normalized_host_info.support_float16) {
if (!host_info.support_float16) {
Optimization::LowerFp16ToFp32(program);
}
if (!normalized_host_info.support_int64) {
if (!host_info.support_int64) {
Optimization::LowerInt64ToInt32(program);
}
if (!normalized_host_info.support_conditional_barrier) {
if (!host_info.support_conditional_barrier) {
Optimization::ConditionalBarrierPass(program);
}
Optimization::SsaRewritePass(program);
@@ -298,8 +295,8 @@ IR::Program TranslateProgram(ObjectPool<IR::Inst>& inst_pool, ObjectPool<IR::Blo
Optimization::PositionPass(env, program);
Optimization::GlobalMemoryToStorageBufferPass(program, normalized_host_info);
Optimization::TexturePass(env, program, normalized_host_info);
Optimization::GlobalMemoryToStorageBufferPass(program, host_info);
Optimization::TexturePass(env, program, host_info);
if (Settings::values.resolution_info.active || Settings::values.rescale_hack.GetValue()) {
Optimization::RescalingPass(program);
@@ -309,7 +306,7 @@ IR::Program TranslateProgram(ObjectPool<IR::Inst>& inst_pool, ObjectPool<IR::Blo
Optimization::VerificationPass(program);
}
Optimization::CollectShaderInfoPass(env, program);
Optimization::LayerPass(program, normalized_host_info);
Optimization::LayerPass(program, host_info);
Optimization::VendorWorkaroundPass(program);
CollectInterpolationInfo(env, program);
+4 -38
View File
@@ -6,8 +6,6 @@
#pragma once
#include "common/common_types.h"
namespace Shader {
// Try to keep entries here to a minimum
@@ -15,52 +13,20 @@ namespace Shader {
/// Misc information about the host
struct HostTranslateInfo {
static constexpr u32 DEFAULT_DESCRIPTOR_LIMIT = 1024;
u64 min_ssbo_alignment{}; ///< Minimum alignment supported by the device for SSBOs
u32 max_per_stage_descriptor_sampled_images{}; ///< maximum sampled descriptors per stage
u32 max_per_stage_resources{}; ///< maximum resources per stage
u32 max_descriptor_set_samplers{};
u32 max_descriptor_set_uniform_buffers{};
u32 max_descriptor_set_uniform_buffers_dynamic{};
u32 max_descriptor_set_storage_buffers{};
u32 max_descriptor_set_storage_buffers_dynamic{};
u32 max_descriptor_set_sampled_images{};
u32 max_descriptor_set_storage_images{};
u32 max_descriptor_set_input_attachements{};
bool support_float64{}; ///< True when the device supports 64-bit floats
bool support_float16{}; ///< True when the device supports 16-bit floats
bool support_int64{}; ///< True when the device supports 64-bit integers
bool needs_demote_reorder{}; ///< True when the device needs DemoteToHelperInvocation reordered
bool support_snorm_render_buffer{}; ///< True when the device supports SNORM render buffers
bool support_viewport_index_layer{}; ///< True when the device supports gl_Layer in VS
u32 min_ssbo_alignment{}; ///< Minimum alignment supported by the device for SSBOs
u32 max_per_stage_descriptor_sampled_images{1024}; ///< maximum sampled descriptors per stage
u32 max_per_stage_resources{4096}; ///< maximum resources per stage
u32 max_descriptor_set_sampled_images{1024}; ///< maximum sampled descriptors per set
bool support_geometry_shader_passthrough{}; ///< True when the device supports geometry
///< passthrough shaders
bool support_conditional_barrier{}; ///< True when the device supports barriers in conditional
///< control flow
void ApplyDescriptorLimitPolicy() noexcept {
if (min_ssbo_alignment == 0) {
min_ssbo_alignment = 1;
}
ApplyDescriptorLimitFallback(max_per_stage_descriptor_sampled_images);
ApplyDescriptorLimitFallback(max_per_stage_resources);
ApplyDescriptorLimitFallback(max_descriptor_set_samplers);
ApplyDescriptorLimitFallback(max_descriptor_set_uniform_buffers);
ApplyDescriptorLimitFallback(max_descriptor_set_uniform_buffers_dynamic);
ApplyDescriptorLimitFallback(max_descriptor_set_storage_buffers);
ApplyDescriptorLimitFallback(max_descriptor_set_storage_buffers_dynamic);
ApplyDescriptorLimitFallback(max_descriptor_set_sampled_images);
ApplyDescriptorLimitFallback(max_descriptor_set_storage_images);
ApplyDescriptorLimitFallback(max_descriptor_set_input_attachements);
}
private:
static void ApplyDescriptorLimitFallback(u32& limit) noexcept {
if (limit == 0) {
limit = DEFAULT_DESCRIPTOR_LIMIT;
}
}
};
} // namespace Shader
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2021 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -548,7 +545,7 @@ void GlobalMemoryToStorageBufferPass(IR::Program& program, const HostTranslateIn
IR::Block* const block{storage_inst.block};
IR::Inst* const inst{storage_inst.inst};
const IR::U32 offset{
StorageOffset(*block, *inst, storage_buffer, u32(host_info.min_ssbo_alignment))};
StorageOffset(*block, *inst, storage_buffer, host_info.min_ssbo_alignment)};
Replace(*block, *inst, index, offset);
}
}
+71 -65
View File
@@ -32,62 +32,71 @@ struct TextureInst {
using TextureInstVector = boost::container::small_vector<TextureInst, 24>;
constexpr u32 DESCRIPTOR_SIZE = 8;
constexpr u32 DESCRIPTOR_SIZE_SHIFT = u32(std::countr_zero(DESCRIPTOR_SIZE));
constexpr u32 DESCRIPTOR_MAX_COUNT = 1024;
constexpr u32 DESCRIPTOR_SIZE_SHIFT = static_cast<u32>(std::countr_zero(DESCRIPTOR_SIZE));
constexpr u32 DYNAMIC_DESCRIPTOR_CBUF_BYTES = 16 * 1024;
constexpr u32 MAX_DYNAMIC_DESCRIPTOR_COUNT = 1024;
u32 DynamicDescriptorSizeShift(const IR::U32& dynamic_offset) {
const IR::Inst* const inst = dynamic_offset.InstRecursive();
if (!inst || inst->GetOpcode() != IR::Opcode::ShiftLeftLogical32)
const IR::Inst* const inst{dynamic_offset.InstRecursive()};
if (!inst || inst->GetOpcode() != IR::Opcode::ShiftLeftLogical32) {
return DESCRIPTOR_SIZE_SHIFT;
const IR::Value shift = inst->Arg(1);
if (!shift.IsImmediate())
}
const IR::Value shift{inst->Arg(1)};
if (!shift.IsImmediate()) {
return DESCRIPTOR_SIZE_SHIFT;
const u32 size_shift = shift.U32();
return size_shift >= DESCRIPTOR_SIZE_SHIFT && size_shift < 31 ? size_shift : DESCRIPTOR_SIZE_SHIFT;
}
const u32 size_shift{shift.U32()};
return size_shift >= DESCRIPTOR_SIZE_SHIFT && size_shift < 31 ? size_shift
: DESCRIPTOR_SIZE_SHIFT;
}
u32 DynamicDescriptorCount(u32 base_offset, u32 size_shift, u32 max_descriptors) {
auto const descriptor_limit = (std::max)(1U, max_descriptors);
auto const max_cbuf_bytes = 16 * descriptor_limit;
if (size_shift >= 31 || base_offset >= max_cbuf_bytes)
u32 DynamicDescriptorCount(u32 base_offset, u32 size_shift) {
if (size_shift >= 31 || base_offset >= DYNAMIC_DESCRIPTOR_CBUF_BYTES) {
return 1;
auto const stride = 1U << size_shift;
auto const available = max_cbuf_bytes - base_offset;
if (available < DESCRIPTOR_SIZE)
}
const u32 stride{1U << size_shift};
const u32 available{DYNAMIC_DESCRIPTOR_CBUF_BYTES - base_offset};
if (available < DESCRIPTOR_SIZE) {
return 1;
auto const available_count = 1U + (available - DESCRIPTOR_SIZE) / stride;
return std::min(descriptor_limit, available_count);
}
const u32 available_count{1U + (available - DESCRIPTOR_SIZE) / stride};
return std::min(MAX_DYNAMIC_DESCRIPTOR_COUNT, available_count);
}
u32 SaturatingSub(u32 lhs, u32 rhs) {
return lhs > rhs ? lhs - rhs : 0;
}
template <typename T>
[[nodiscard]] u32 StaticDescriptorCount(T const& descriptors) noexcept {
return std::accumulate(descriptors.cbegin(), descriptors.cend(), 0U, [](auto const& acc, auto const& e) {
return acc + (e.count <= 1 ? e.count : 0);
});
template <typename Descriptors>
u32 StaticDescriptorCount(const Descriptors& descriptors) {
u32 count{};
for (const auto& desc : descriptors) {
if (desc.count <= 1) {
count += desc.count;
}
}
return count;
}
u32 DynamicSampledTextureCap(const Info& info, const HostTranslateInfo& host_info, u32 dynamic_arrays) {
auto const sampled_limit = (std::max)(1U, std::min(host_info.max_per_stage_descriptor_sampled_images,
host_info.max_descriptor_set_sampled_images));
auto const resource_limit = (std::max)(1U, host_info.max_per_stage_resources);
if (dynamic_arrays > 0) {
auto const sampled_static_count = StaticDescriptorCount(info.texture_buffer_descriptors) + StaticDescriptorCount(info.texture_descriptors);
auto const resource_static_count =
NumDescriptors(info.constant_buffer_descriptors)
+ NumDescriptors(info.storage_buffers_descriptors)
+ sampled_static_count + NumDescriptors(info.image_buffer_descriptors)
+ NumDescriptors(info.image_descriptors);
auto const sampled_budget = SaturatingSub(sampled_limit, sampled_static_count);
auto const resource_budget = SaturatingSub(resource_limit, resource_static_count);
auto const sampled_cap = sampled_budget / dynamic_arrays;
auto const resource_cap = resource_budget / dynamic_arrays;
return (std::max)(1U, (std::min)(sampled_cap, resource_cap));
u32 DynamicSampledTextureCap(const Info& info, const HostTranslateInfo& host_info,
u32 dynamic_arrays) {
if (dynamic_arrays == 0) {
return MAX_DYNAMIC_DESCRIPTOR_COUNT;
}
return (std::min)({DESCRIPTOR_MAX_COUNT, sampled_limit, resource_limit});
const u32 sampled_static_count{StaticDescriptorCount(info.texture_buffer_descriptors) +
StaticDescriptorCount(info.texture_descriptors)};
const u32 resource_static_count{
NumDescriptors(info.constant_buffer_descriptors) +
NumDescriptors(info.storage_buffers_descriptors) + sampled_static_count +
NumDescriptors(info.image_buffer_descriptors) + NumDescriptors(info.image_descriptors)};
const u32 sampled_limit{std::min(host_info.max_per_stage_descriptor_sampled_images,
host_info.max_descriptor_set_sampled_images)};
const u32 sampled_budget{SaturatingSub(sampled_limit, sampled_static_count)};
const u32 resource_budget{SaturatingSub(host_info.max_per_stage_resources,
resource_static_count)};
const u32 sampled_cap{sampled_budget / dynamic_arrays};
const u32 resource_cap{resource_budget / dynamic_arrays};
return std::max(1U, std::min({MAX_DYNAMIC_DESCRIPTOR_COUNT, sampled_cap, resource_cap}));
}
IR::Opcode IndexedInstruction(const IR::Inst& inst) {
@@ -295,23 +304,21 @@ static inline bool IsTexturePixelFormatIntegerCached(Environment& env,
}
std::optional<ConstBufferAddr> Track(const IR::Value& value, Environment& env, const HostTranslateInfo& host_info);
static inline std::optional<ConstBufferAddr> TrackCached(const IR::Value& v, Environment& env, const HostTranslateInfo& host_info) {
std::optional<ConstBufferAddr> Track(const IR::Value& value, Environment& env);
static inline std::optional<ConstBufferAddr> TrackCached(const IR::Value& v, Environment& env) {
if (const IR::Inst* key = v.InstRecursive()) {
if (auto it = env.track_cache.find(key); it != env.track_cache.end()) return it->second;
auto found = Track(v, env, host_info);
auto found = Track(v, env);
if (found) env.track_cache.emplace(key, *found);
return found;
}
return Track(v, env, host_info);
return Track(v, env);
}
std::optional<ConstBufferAddr> TryGetConstBuffer(const IR::Inst* inst, Environment& env, const HostTranslateInfo& host_info);
std::optional<ConstBufferAddr> TryGetConstBuffer(const IR::Inst* inst, Environment& env);
std::optional<ConstBufferAddr> Track(const IR::Value& value, Environment& env, const HostTranslateInfo& host_info) {
return IR::BreadthFirstSearch(value, [&env, &host_info](const IR::Inst* inst) {
return TryGetConstBuffer(inst, env, host_info);
});
std::optional<ConstBufferAddr> Track(const IR::Value& value, Environment& env) {
return IR::BreadthFirstSearch(value, [&env](const IR::Inst* inst) { return TryGetConstBuffer(inst, env); });
}
std::optional<u32> TryGetConstant(IR::Value& value, Environment& env) {
@@ -335,13 +342,13 @@ std::optional<u32> TryGetConstant(IR::Value& value, Environment& env) {
return ReadCbufCached(env, index_number, offset_number);
}
std::optional<ConstBufferAddr> TryGetConstBuffer(const IR::Inst* inst, Environment& env, const HostTranslateInfo& host_info) {
std::optional<ConstBufferAddr> TryGetConstBuffer(const IR::Inst* inst, Environment& env) {
switch (inst->GetOpcode()) {
default:
return std::nullopt;
case IR::Opcode::BitwiseOr32: {
std::optional lhs{TrackCached(inst->Arg(0), env, host_info)};
std::optional rhs{TrackCached(inst->Arg(1), env, host_info)};
std::optional lhs{TrackCached(inst->Arg(0), env)};
std::optional rhs{TrackCached(inst->Arg(1), env)};
if (!lhs || !rhs) {
return std::nullopt;
}
@@ -371,11 +378,12 @@ std::optional<ConstBufferAddr> TryGetConstBuffer(const IR::Inst* inst, Environme
if (!shift.IsImmediate()) {
return std::nullopt;
}
std::optional lhs{TrackCached(inst->Arg(0), env, host_info)};
std::optional lhs{TrackCached(inst->Arg(0), env)};
if (lhs) {
lhs->shift_left = shift.U32();
}
return lhs;
break;
}
case IR::Opcode::BitwiseAnd32: {
IR::Value op1{inst->Arg(0)};
@@ -399,7 +407,7 @@ std::optional<ConstBufferAddr> TryGetConstBuffer(const IR::Inst* inst, Environme
return std::nullopt;
} while (false);
}
std::optional lhs{TrackCached(op1, env, host_info)};
std::optional lhs{TrackCached(op1, env)};
if (lhs) {
lhs->shift_left = static_cast<u32>(std::countr_zero(op2.U32()));
}
@@ -445,10 +453,7 @@ std::optional<ConstBufferAddr> TryGetConstBuffer(const IR::Inst* inst, Environme
} else {
return std::nullopt;
}
auto const size_shift = DynamicDescriptorSizeShift(dynamic_offset);
auto const sampled_limit = (std::max)(1U, (std::min)(host_info.max_per_stage_descriptor_sampled_images,
host_info.max_descriptor_set_sampled_images));
auto const resource_limit = (std::max)(1U, host_info.max_per_stage_resources);
const u32 size_shift{DynamicDescriptorSizeShift(dynamic_offset)};
return ConstBufferAddr{
.index = index.U32(),
.offset = base_offset,
@@ -457,15 +462,15 @@ std::optional<ConstBufferAddr> TryGetConstBuffer(const IR::Inst* inst, Environme
.secondary_offset = 0,
.secondary_shift_left = 0,
.dynamic_offset = dynamic_offset,
.count = DynamicDescriptorCount(base_offset, size_shift, (std::min)({DESCRIPTOR_MAX_COUNT, sampled_limit, resource_limit})),
.count = DynamicDescriptorCount(base_offset, size_shift),
.has_secondary = false,
};
}
TextureInst MakeInst(Environment& env, IR::Block* block, IR::Inst& inst, const HostTranslateInfo& host_info) {
TextureInst MakeInst(Environment& env, IR::Block* block, IR::Inst& inst) {
ConstBufferAddr addr;
if (IsBindless(inst)) {
const std::optional<ConstBufferAddr> track_addr{TrackCached(inst.Arg(0), env, host_info)};
const std::optional<ConstBufferAddr> track_addr{TrackCached(inst.Arg(0), env)};
if (!track_addr) {
throw NotImplementedException("Failed to track bindless texture constant buffer");
@@ -501,15 +506,15 @@ u32 GetTextureHandle(Environment& env, const ConstBufferAddr& cbuf) {
return lhs_raw | rhs_raw;
}
[[maybe_unused]] TextureType ReadTextureType(Environment& env, const ConstBufferAddr& cbuf) {
[[maybe_unused]]TextureType ReadTextureType(Environment& env, const ConstBufferAddr& cbuf) {
return env.ReadTextureType(GetTextureHandle(env, cbuf));
}
[[maybe_unused]] TexturePixelFormat ReadTexturePixelFormat(Environment& env, const ConstBufferAddr& cbuf) {
[[maybe_unused]]TexturePixelFormat ReadTexturePixelFormat(Environment& env, const ConstBufferAddr& cbuf) {
return env.ReadTexturePixelFormat(GetTextureHandle(env, cbuf));
}
[[maybe_unused]] bool IsTexturePixelFormatInteger(Environment& env, const ConstBufferAddr& cbuf) {
[[maybe_unused]]bool IsTexturePixelFormatInteger(Environment& env, const ConstBufferAddr& cbuf) {
return env.IsTexturePixelFormatInteger(GetTextureHandle(env, cbuf));
}
@@ -670,7 +675,7 @@ void TexturePass(Environment& env, IR::Program& program, const HostTranslateInfo
if (!IsTextureInstruction(inst)) {
continue;
}
to_replace.push_back(MakeInst(env, block, inst, host_info));
to_replace.push_back(MakeInst(env, block, inst));
}
}
// Sort instructions to visit textures by constant buffer index, then by offset
@@ -684,7 +689,8 @@ void TexturePass(Environment& env, IR::Program& program, const HostTranslateInfo
program.info.texture_descriptors,
program.info.image_descriptors,
};
const u32 sampled_dynamic_cap = DynamicSampledTextureCap(program.info, host_info, DynamicSampledTextureArrayCount(to_replace));
const u32 sampled_dynamic_cap{
DynamicSampledTextureCap(program.info, host_info, DynamicSampledTextureArrayCount(to_replace))};
for (TextureInst& texture_inst : to_replace) {
// TODO: Handle arrays
IR::Inst* const inst{texture_inst.inst};
+1
View File
@@ -92,6 +92,7 @@ struct Profile {
bool has_broken_robust{};
u64 min_ssbo_alignment{};
u32 max_user_clip_distances{};
};
+2 -27
View File
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2021 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -18,19 +15,12 @@ static constexpr auto PERMS = Common::MemoryPermission::ReadWrite;
static constexpr auto HEAP = false;
TEST_CASE("HostMemory: Initialize and deinitialize", "[common]") {
{
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
}
{
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
}
{ HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE); }
{ HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE); }
}
TEST_CASE("HostMemory: Simple map", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x5000, 0x8000, 0x1000, PERMS, HEAP);
volatile u8* const data = mem.VirtualBasePointer() + 0x5000;
@@ -40,7 +30,6 @@ TEST_CASE("HostMemory: Simple map", "[common]") {
TEST_CASE("HostMemory: Simple mirror map", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x5000, 0x3000, 0x2000, PERMS, HEAP);
mem.Map(0x8000, 0x4000, 0x1000, PERMS, HEAP);
@@ -52,7 +41,6 @@ TEST_CASE("HostMemory: Simple mirror map", "[common]") {
TEST_CASE("HostMemory: Simple unmap", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x5000, 0x3000, 0x2000, PERMS, HEAP);
volatile u8* const data = mem.VirtualBasePointer() + 0x5000;
@@ -64,7 +52,6 @@ TEST_CASE("HostMemory: Simple unmap", "[common]") {
TEST_CASE("HostMemory: Simple unmap and remap", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x5000, 0x3000, 0x2000, PERMS, HEAP);
volatile u8* const data = mem.VirtualBasePointer() + 0x5000;
@@ -82,7 +69,6 @@ TEST_CASE("HostMemory: Simple unmap and remap", "[common]") {
TEST_CASE("HostMemory: Nieche allocation", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x0000, 0, 0x20000, PERMS, HEAP);
mem.Unmap(0x0000, 0x4000, HEAP);
mem.Map(0x1000, 0, 0x2000, PERMS, HEAP);
@@ -92,7 +78,6 @@ TEST_CASE("HostMemory: Nieche allocation", "[common]") {
TEST_CASE("HostMemory: Full unmap", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x8000, 0, 0x4000, PERMS, HEAP);
mem.Unmap(0x8000, 0x4000, HEAP);
mem.Map(0x6000, 0, 0x16000, PERMS, HEAP);
@@ -100,7 +85,6 @@ TEST_CASE("HostMemory: Full unmap", "[common]") {
TEST_CASE("HostMemory: Right out of bounds unmap", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x0000, 0, 0x4000, PERMS, HEAP);
mem.Unmap(0x2000, 0x4000, HEAP);
mem.Map(0x2000, 0x80000, 0x4000, PERMS, HEAP);
@@ -108,7 +92,6 @@ TEST_CASE("HostMemory: Right out of bounds unmap", "[common]") {
TEST_CASE("HostMemory: Left out of bounds unmap", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
mem.Map(0x8000, 0, 0x4000, PERMS, HEAP);
mem.Unmap(0x6000, 0x4000, HEAP);
mem.Map(0x8000, 0, 0x2000, PERMS, HEAP);
@@ -116,7 +99,6 @@ TEST_CASE("HostMemory: Left out of bounds unmap", "[common]") {
TEST_CASE("HostMemory: Multiple placeholder unmap", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x0000, 0, 0x4000, PERMS, HEAP);
mem.Map(0x4000, 0, 0x1b000, PERMS, HEAP);
mem.Unmap(0x3000, 0x1c000, HEAP);
@@ -125,7 +107,6 @@ TEST_CASE("HostMemory: Multiple placeholder unmap", "[common]") {
TEST_CASE("HostMemory: Unmap between placeholders", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x0000, 0, 0x4000, PERMS, HEAP);
mem.Map(0x4000, 0, 0x4000, PERMS, HEAP);
mem.Unmap(0x2000, 0x4000, HEAP);
@@ -134,7 +115,6 @@ TEST_CASE("HostMemory: Unmap between placeholders", "[common]") {
TEST_CASE("HostMemory: Unmap to origin", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x4000, 0, 0x4000, PERMS, HEAP);
mem.Map(0x8000, 0, 0x4000, PERMS, HEAP);
mem.Unmap(0x4000, 0x4000, HEAP);
@@ -144,7 +124,6 @@ TEST_CASE("HostMemory: Unmap to origin", "[common]") {
TEST_CASE("HostMemory: Unmap to right", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x4000, 0, 0x4000, PERMS, HEAP);
mem.Map(0x8000, 0, 0x4000, PERMS, HEAP);
mem.Unmap(0x8000, 0x4000, HEAP);
@@ -153,7 +132,6 @@ TEST_CASE("HostMemory: Unmap to right", "[common]") {
TEST_CASE("HostMemory: Partial right unmap check bindings", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x4000, 0x10000, 0x4000, PERMS, HEAP);
volatile u8* const ptr = mem.VirtualBasePointer() + 0x4000;
@@ -166,7 +144,6 @@ TEST_CASE("HostMemory: Partial right unmap check bindings", "[common]") {
TEST_CASE("HostMemory: Partial left unmap check bindings", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x4000, 0x10000, 0x4000, PERMS, HEAP);
volatile u8* const ptr = mem.VirtualBasePointer() + 0x4000;
@@ -181,7 +158,6 @@ TEST_CASE("HostMemory: Partial left unmap check bindings", "[common]") {
TEST_CASE("HostMemory: Partial middle unmap check bindings", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x4000, 0x10000, 0x4000, PERMS, HEAP);
volatile u8* const ptr = mem.VirtualBasePointer() + 0x4000;
@@ -196,7 +172,6 @@ TEST_CASE("HostMemory: Partial middle unmap check bindings", "[common]") {
TEST_CASE("HostMemory: Partial sparse middle unmap and check bindings", "[common]") {
HostMemory mem(BACKING_SIZE, VIRTUAL_SIZE);
REQUIRE(mem.BackingBasePointer() != nullptr);
mem.Map(0x4000, 0x10000, 0x2000, PERMS, HEAP);
mem.Map(0x6000, 0x20000, 0x2000, PERMS, HEAP);
+5 -5
View File
@@ -109,12 +109,12 @@ public:
return static_cast<u32>(other_cpu_addr - cpu_addr);
}
size_t getLRUID() const noexcept {
return lru_id;
u64 GetFrameTick() const noexcept {
return frame_tick;
}
void setLRUID(size_t lru_id_) {
lru_id = lru_id_;
void SetFrameTick(u64 tick) noexcept {
frame_tick = tick;
}
size_t SizeBytes() const {
@@ -125,7 +125,7 @@ private:
VAddr cpu_addr = 0;
BufferFlagBits flags{};
int stream_score = 0;
size_t lru_id = SIZE_MAX;
u64 frame_tick = 0;
size_t size_bytes = 0;
};
+12 -8
View File
@@ -58,17 +58,22 @@ void BufferCache<P>::RunGarbageCollector() {
const bool aggressive_gc = total_used_memory >= critical_memory;
const u64 ticks_to_destroy = aggressive_gc ? 60 : 120;
int num_iterations = aggressive_gc ? 64 : 32;
const auto clean_up = [this, &num_iterations](BufferId buffer_id) {
const u64 threshold = frame_tick - ticks_to_destroy;
boost::container::small_vector<BufferId, 64> expired;
for (auto [id, buffer] : slot_buffers) {
if (buffer->GetFrameTick() < threshold) {
expired.push_back(id);
}
}
for (const auto buffer_id : expired) {
if (num_iterations == 0) {
return true;
break;
}
--num_iterations;
auto& buffer = slot_buffers[buffer_id];
DownloadBufferMemory(buffer);
DeleteBuffer(buffer_id);
return false;
};
lru_cache.ForEachItemBelow(frame_tick - ticks_to_destroy, clean_up);
}
}
template <class P>
@@ -1591,10 +1596,9 @@ void BufferCache<P>::ChangeRegister(BufferId buffer_id) {
const auto size = buffer.SizeBytes();
if (insert) {
total_used_memory += Common::AlignUp(size, 1024);
buffer.setLRUID(lru_cache.Insert(buffer_id, frame_tick));
buffer.SetFrameTick(frame_tick);
} else {
total_used_memory -= Common::AlignUp(size, 1024);
lru_cache.Free(buffer.getLRUID());
}
const DAddr device_addr_begin = buffer.CpuAddr();
const DAddr device_addr_end = device_addr_begin + size;
@@ -1612,7 +1616,7 @@ void BufferCache<P>::ChangeRegister(BufferId buffer_id) {
template <class P>
void BufferCache<P>::TouchBuffer(Buffer& buffer, BufferId buffer_id) noexcept {
if (buffer_id != NULL_BUFFER_ID) {
lru_cache.Touch(buffer.getLRUID(), frame_tick);
buffer.SetFrameTick(frame_tick);
}
}
@@ -23,7 +23,6 @@
#include "common/common_types.h"
#include "common/div_ceil.h"
#include "common/literals.h"
#include "common/lru_cache.h"
#include "common/range_sets.h"
#include "common/scope_exit.h"
#include "common/settings.h"
@@ -506,11 +505,6 @@ private:
size_t immediate_buffer_capacity = 0;
Common::ScratchBuffer<u8> immediate_buffer_alloc;
struct LRUItemParams {
using ObjectType = BufferId;
using TickType = u64;
};
Common::LeastRecentlyUsedCache<LRUItemParams> lru_cache;
u64 frame_tick = 0;
u64 total_used_memory = 0;
u64 minimum_memory = 0;
+34 -2
View File
@@ -117,7 +117,35 @@ void DmaPusher::ProcessCommands(std::span<const CommandHeader> commands) {
dma_state.is_last_call = true;
index += max_write;
} else if (dma_state.method_count) {
auto const command_header = commands[index]; //can copy
if (!dma_state.non_incrementing && !dma_increment_once &&
dma_state.method >= non_puller_methods) {
auto subchannel = subchannels[dma_state.subchannel];
const u32 available = u32(std::min<size_t>(
index + dma_state.method_count, commands.size()) - index);
u32 batch = 0;
u32 method = dma_state.method;
while (batch < available) {
const bool needs_exec =
(method < Engines::EngineInterface::EXECUTION_MASK_TABLE_SIZE)
? subchannel->execution_mask[method]
: subchannel->execution_mask_default;
if (needs_exec) break;
batch++;
method++;
}
if (batch > 0) {
auto& sink = subchannel->method_sink;
sink.reserve(sink.size() + batch);
for (u32 j = 0; j < batch; j++) {
sink.emplace_back(dma_state.method + j, commands[index + j].argument);
}
dma_state.method += batch;
dma_state.method_count -= batch;
index += batch;
continue;
}
}
auto const command_header = commands[index];
dma_state.dma_word_offset = u32(index * sizeof(u32));
dma_state.is_last_call = dma_state.method_count <= 1;
CallMethod(command_header.argument);
@@ -176,7 +204,11 @@ void DmaPusher::CallMethod(u32 argument) {
});
} else {
auto subchannel = subchannels[dma_state.subchannel];
if (!subchannel->execution_mask[dma_state.method]) {
const bool needs_execution =
(dma_state.method < Engines::EngineInterface::EXECUTION_MASK_TABLE_SIZE)
? subchannel->execution_mask[dma_state.method]
: subchannel->execution_mask_default;
if (!needs_execution) {
subchannel->method_sink.emplace_back(dma_state.method, argument);
} else {
subchannel->ConsumeSink(system);
+9 -5
View File
@@ -6,9 +6,8 @@
#pragma once
#include <bitset>
#include <limits>
#include <vector>
#include <array>
#include <boost/container/small_vector.hpp>
#include "common/common_types.h"
@@ -43,10 +42,15 @@ public:
}
}
std::bitset<(std::numeric_limits<u16>::max)()> execution_mask{};
std::vector<std::pair<u32, u32>> method_sink{};
static constexpr size_t EXECUTION_MASK_TABLE_SIZE = 0xE00;
std::array<u8, EXECUTION_MASK_TABLE_SIZE> execution_mask{};
bool execution_mask_default{};
boost::container::small_vector<std::pair<u32, u32>, 64> method_sink{};
GPUVAddr current_dma_segment;
/// @brief Indicates whether the current DMA segment is dirty.
bool current_dirty{};
protected:
virtual void ConsumeSinkImpl(Core::System& system) {
for (auto [method, value] : method_sink) {
+1 -1
View File
@@ -26,7 +26,7 @@ Fermi2D::Fermi2D(MemoryManager& memory_manager_) : memory_manager{memory_manager
regs.src.depth = 1;
regs.dst.depth = 1;
execution_mask.reset();
execution_mask.fill(0);
execution_mask[FERMI2D_REG_INDEX(pixels_from_memory.src_y0) + 1] = true;
}
+1 -1
View File
@@ -20,7 +20,7 @@ KeplerCompute::KeplerCompute(MemoryManager& memory_manager_)
: memory_manager{memory_manager_}
, upload_state{memory_manager, regs.upload}
{
execution_mask.reset();
execution_mask.fill(0);
execution_mask[KEPLER_COMPUTE_REG_INDEX(exec_upload)] = true;
execution_mask[KEPLER_COMPUTE_REG_INDEX(data_upload)] = true;
execution_mask[KEPLER_COMPUTE_REG_INDEX(launch)] = true;
+1 -1
View File
@@ -23,7 +23,7 @@ KeplerMemory::~KeplerMemory() = default;
void KeplerMemory::BindRasterizer(VideoCore::RasterizerInterface* rasterizer_) {
upload_state.BindRasterizer(rasterizer_);
execution_mask.reset();
execution_mask.fill(0);
execution_mask[KEPLERMEMORY_REG_INDEX(exec)] = true;
execution_mask[KEPLERMEMORY_REG_INDEX(data)] = true;
}
+48 -5
View File
@@ -4,8 +4,14 @@
// SPDX-FileCopyrightText: Copyright 2018 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
#include <algorithm>
#include <cstring>
#include <optional>
#if defined(_MSC_VER) && !defined(__clang__)
#include <intrin.h>
#endif
#include "common/assert.h"
#include "common/bit_util.h"
#include "common/scope_exit.h"
@@ -22,6 +28,16 @@
namespace Tegra::Engines {
namespace {
inline void PrefetchLine(const void* addr) {
#if defined(_MSC_VER) && !defined(__clang__)
_mm_prefetch(static_cast<const char*>(addr), _MM_HINT_T0);
#else
__builtin_prefetch(addr, 0, 1);
#endif
}
} // namespace
/// First register id that is actually a Macro call.
constexpr u32 MacroRegistersStart = 0xE00;
@@ -37,9 +53,10 @@ Maxwell3D::Maxwell3D(MemoryManager& memory_manager_)
{
dirty.flags.flip();
InitializeRegisterDefaults();
execution_mask.reset();
for (size_t i = 0; i < execution_mask.size(); i++)
execution_mask.fill(0);
for (size_t i = 0; i < EXECUTION_MASK_TABLE_SIZE; i++)
execution_mask[i] = IsMethodExecutable(u32(i));
execution_mask_default = true;
}
Maxwell3D::~Maxwell3D() = default;
@@ -282,18 +299,44 @@ u32 Maxwell3D::ProcessShadowRam(u32 method, u32 argument) {
}
void Maxwell3D::ConsumeSinkImpl(Core::System& system) {
std::stable_sort(method_sink.begin(), method_sink.end(),
[](const auto& a, const auto& b) { return a.first < b.first; });
const auto sink_size = method_sink.size();
const auto control = shadow_state.shadow_ram_control;
if (control == Regs::ShadowRamControl::Track || control == Regs::ShadowRamControl::TrackWithFilter) {
for (auto [method, value] : method_sink) {
for (size_t i = 0; i < sink_size; ++i) {
const auto [method, value] = method_sink[i];
if (i + 1 < sink_size) {
const u32 next = method_sink[i + 1].first;
PrefetchLine(&regs.reg_array[next]);
PrefetchLine(&shadow_state.reg_array[next]);
PrefetchLine(&dirty.tables[0][next]);
}
shadow_state.reg_array[method] = value;
ProcessDirtyRegisters(method, value);
}
} else if (control == Regs::ShadowRamControl::Replay) {
for (auto [method, value] : method_sink)
for (size_t i = 0; i < sink_size; ++i) {
const auto [method, value] = method_sink[i];
if (i + 1 < sink_size) {
const u32 next = method_sink[i + 1].first;
PrefetchLine(&regs.reg_array[next]);
PrefetchLine(&shadow_state.reg_array[next]);
PrefetchLine(&dirty.tables[0][next]);
}
ProcessDirtyRegisters(method, shadow_state.reg_array[method]);
}
} else {
for (auto [method, value] : method_sink)
for (size_t i = 0; i < sink_size; ++i) {
const auto [method, value] = method_sink[i];
if (i + 1 < sink_size) {
const u32 next = method_sink[i + 1].first;
PrefetchLine(&regs.reg_array[next]);
PrefetchLine(&dirty.tables[0][next]);
}
ProcessDirtyRegisters(method, value);
}
}
method_sink.clear();
}
+1 -1
View File
@@ -24,7 +24,7 @@ using namespace Texture;
MaxwellDMA::MaxwellDMA(MemoryManager& memory_manager_)
: memory_manager{memory_manager_}
{
execution_mask.reset();
execution_mask.fill(0);
execution_mask[offsetof(Regs, launch_dma) / sizeof(u32)] = true;
}
+3
View File
@@ -287,6 +287,7 @@ void QueryCacheBase<Traits>::CounterReport(GPUVAddr addr, QueryType counter_type
u32 value = static_cast<u32>(query_base->value);
std::memcpy(pointer, &value, sizeof(value));
}
query_base->flags |= QueryFlagBits::IsGuestSynced;
if (!is_synced) [[likely]] {
impl->pending_unregister.push_back(query_location);
}
@@ -569,10 +570,12 @@ bool QueryCacheBase<Traits>::SemiFlushQueryDirty(QueryCacheBase<Traits>::QueryLo
auto* ptr = impl->device_memory.template GetPointer<u8>(query_base->guest_address);
if (True(query_base->flags & QueryFlagBits::HasTimestamp)) {
std::memcpy(ptr, &query_base->value, sizeof(query_base->value));
query_base->flags |= QueryFlagBits::IsGuestSynced;
return false;
}
u32 value_l = static_cast<u32>(query_base->value);
std::memcpy(ptr, &value_l, sizeof(value_l));
query_base->flags |= QueryFlagBits::IsGuestSynced;
return false;
}
return True(query_base->flags & QueryFlagBits::IsHostManaged) &&
@@ -245,31 +245,16 @@ ShaderCache::ShaderCache(Tegra::MaxwellDeviceMemoryManager& device_memory_,
std::min<u32>(device.GetMaxUserClipDistances(), Maxwell::Regs::NumClipDistances),
},
host_info{
.min_ssbo_alignment = static_cast<u32>(device.GetShaderStorageBufferAlignment()),
.max_per_stage_descriptor_sampled_images =
Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.max_per_stage_resources = Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.max_descriptor_set_samplers = Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.max_descriptor_set_uniform_buffers = Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.max_descriptor_set_uniform_buffers_dynamic =
Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.max_descriptor_set_storage_buffers = Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.max_descriptor_set_storage_buffers_dynamic =
Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.max_descriptor_set_sampled_images = Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.max_descriptor_set_storage_images = Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.max_descriptor_set_input_attachements =
Shader::HostTranslateInfo::DEFAULT_DESCRIPTOR_LIMIT,
.support_float64 = true,
.support_float16 = false,
.support_int64 = device.HasShaderInt64(),
.needs_demote_reorder = device.IsAmd(),
.support_snorm_render_buffer = false,
.support_viewport_index_layer = device.HasVertexViewportLayer(),
.support_geometry_shader_passthrough = device.HasGeometryShaderPassthrough(),
.support_conditional_barrier = device.SupportsConditionalBarriers(),
.support_float64 = true,
.support_float16 = false,
.support_int64 = device.HasShaderInt64(),
.needs_demote_reorder = device.IsAmd(),
.support_snorm_render_buffer = false,
.support_viewport_index_layer = device.HasVertexViewportLayer(),
.min_ssbo_alignment = static_cast<u32>(device.GetShaderStorageBufferAlignment()),
.support_geometry_shader_passthrough = device.HasGeometryShaderPassthrough(),
.support_conditional_barrier = device.SupportsConditionalBarriers(),
} {
host_info.ApplyDescriptorLimitPolicy();
if (use_asynchronous_shaders) {
workers = CreateWorkers();
}
@@ -62,6 +62,7 @@ void FixedPipelineState::Refresh(Tegra::Engines::Maxwell3D& maxwell3d, DynamicFe
extended_dynamic_state_2_logic_op.Assign(features.has_extended_dynamic_state_2_logic_op ? 1 : 0);
extended_dynamic_state_3_blend.Assign(features.has_extended_dynamic_state_3_blend ? 1 : 0);
extended_dynamic_state_3_enables.Assign(features.has_extended_dynamic_state_3_enables ? 1 : 0);
dynamic_state3_depth_clamp_enable.Assign(features.has_dynamic_state3_depth_clamp_enable ? 1 : 0);
dynamic_vertex_input.Assign(features.has_dynamic_vertex_input ? 1 : 0);
xfb_enabled.Assign(regs.transform_feedback_enabled != 0);
ndc_minus_one_to_one.Assign(regs.depth_mode == Maxwell::DepthMode::MinusOneToOne ? 1 : 0);
@@ -208,6 +208,7 @@ struct FixedPipelineState {
BitField<12, 2, u32> tessellation_spacing;
BitField<14, 1, u32> tessellation_clockwise;
BitField<15, 5, u32> patch_control_points_minus_one;
BitField<20, 1, u32> dynamic_state3_depth_clamp_enable;
BitField<24, 4, Maxwell::PrimitiveTopology> topology;
BitField<28, 4, Tegra::Texture::MsaaMode> msaa_mode;
@@ -22,15 +22,6 @@ namespace Vulkan {
using Shader::Backend::SPIRV::NUM_TEXTURE_AND_IMAGE_SCALING_WORDS;
[[nodiscard]] inline u32 NumDescriptorEntries(const Shader::Info& info) {
return Shader::NumDescriptors(info.constant_buffer_descriptors) +
Shader::NumDescriptors(info.storage_buffers_descriptors) +
Shader::NumDescriptors(info.texture_buffer_descriptors) +
Shader::NumDescriptors(info.image_buffer_descriptors) +
Shader::NumDescriptors(info.texture_descriptors) +
Shader::NumDescriptors(info.image_descriptors);
}
class DescriptorLayoutBuilder {
public:
DescriptorLayoutBuilder(const Device& device_) : device{&device_} {}
@@ -326,7 +326,7 @@ std::pair<VkBuffer, VkDeviceSize> Uint8Pass::Assemble(u32 num_vertices, VkBuffer
const u32 staging_size = static_cast<u32>(num_vertices * sizeof(u16));
const auto staging = staging_buffer_pool.Request(staging_size, MemoryUsage::DeviceLocal);
compute_pass_descriptor_queue.Acquire(scheduler, 2);
compute_pass_descriptor_queue.Acquire();
compute_pass_descriptor_queue.AddBuffer(src_buffer, src_offset, num_vertices);
compute_pass_descriptor_queue.AddBuffer(staging.buffer, staging.offset, staging_size);
const void* const descriptor_data{compute_pass_descriptor_queue.UpdateData()};
@@ -384,7 +384,7 @@ std::pair<VkBuffer, VkDeviceSize> QuadIndexedPass::Assemble(
const std::size_t staging_size = num_tri_vertices * sizeof(u32);
const auto staging = staging_buffer_pool.Request(staging_size, MemoryUsage::DeviceLocal);
compute_pass_descriptor_queue.Acquire(scheduler, 2);
compute_pass_descriptor_queue.Acquire();
compute_pass_descriptor_queue.AddBuffer(src_buffer, src_offset, input_size);
compute_pass_descriptor_queue.AddBuffer(staging.buffer, staging.offset, staging_size);
const void* const descriptor_data{compute_pass_descriptor_queue.UpdateData()};
@@ -429,7 +429,7 @@ void ConditionalRenderingResolvePass::Resolve(VkBuffer dst_buffer, VkBuffer src_
}
const size_t compare_size = compare_to_zero ? 8 : 24;
compute_pass_descriptor_queue.Acquire(scheduler, 2);
compute_pass_descriptor_queue.Acquire();
compute_pass_descriptor_queue.AddBuffer(src_buffer, src_offset, compare_size);
compute_pass_descriptor_queue.AddBuffer(dst_buffer, 0, sizeof(u32));
const void* const descriptor_data{compute_pass_descriptor_queue.UpdateData()};
@@ -498,7 +498,7 @@ void QueriesPrefixScanPass::Run(VkBuffer accumulation_buffer, VkBuffer dst_buffe
static constexpr size_t DISPATCH_SIZE = 2048U;
size_t runs_to_do = std::min<size_t>(current_runs, DISPATCH_SIZE);
current_runs -= runs_to_do;
compute_pass_descriptor_queue.Acquire(scheduler, 3);
compute_pass_descriptor_queue.Acquire();
compute_pass_descriptor_queue.AddBuffer(src_buffer, 0, number_of_sums * sizeof(u64));
compute_pass_descriptor_queue.AddBuffer(dst_buffer, 0, number_of_sums * sizeof(u64));
compute_pass_descriptor_queue.AddBuffer(accumulation_buffer, 0, sizeof(u64));
@@ -600,7 +600,7 @@ void ASTCDecoderPass::Assemble(Image& image, const StagingBufferRef& map,
const u32 num_dispatches_y = Common::DivCeil(swizzle.num_tiles.height, 8U);
const u32 num_dispatches_z = image.info.resources.layers;
compute_pass_descriptor_queue.Acquire(scheduler, 2);
compute_pass_descriptor_queue.Acquire();
compute_pass_descriptor_queue.AddBuffer(map.buffer, input_offset,
image.guest_size_bytes - swizzle.buffer_offset);
compute_pass_descriptor_queue.AddImage(image.StorageImageView(swizzle.level));
@@ -821,7 +821,7 @@ void BlockLinearUnswizzle3DPass::UnswizzleChunk(
pc.blocks_dim[1] = blocks_y;
pc.blocks_dim[2] = z_count; // Only process the count
compute_pass_descriptor_queue.Acquire(scheduler, 3);
compute_pass_descriptor_queue.Acquire();
compute_pass_descriptor_queue.AddBuffer(*image.runtime->swizzle_table_buffer, 0,
image.runtime->swizzle_table_size);
compute_pass_descriptor_queue.AddBuffer(swizzled.buffer,
@@ -989,7 +989,7 @@ void MSAACopyPass::CopyImage(Image& dst_image, Image& src_image,
ASSERT(copy.dst_subresource.base_layer == 0);
ASSERT(copy.dst_subresource.num_layers == 1);
compute_pass_descriptor_queue.Acquire(scheduler, 2);
compute_pass_descriptor_queue.Acquire();
compute_pass_descriptor_queue.AddImage(
src_image.StorageImageView(copy.src_subresource.base_level));
compute_pass_descriptor_queue.AddImage(
@@ -45,7 +45,6 @@ ComputePipeline::ComputePipeline(const Device& device_, Scheduler& scheduler, vk
}
std::copy_n(info.constant_buffer_used_sizes.begin(), uniform_buffer_sizes.size(),
uniform_buffer_sizes.begin());
num_descriptor_entries = NumDescriptorEntries(info);
auto func{[this, &scheduler, &descriptor_pool, shader_notify, pipeline_statistics] {
DescriptorLayoutBuilder builder{device};
@@ -114,7 +113,7 @@ ComputePipeline::ComputePipeline(const Device& device_, Scheduler& scheduler, vk
void ComputePipeline::Configure(Tegra::Engines::KeplerCompute& kepler_compute,
Tegra::MemoryManager& gpu_memory, Scheduler& scheduler,
BufferCache& buffer_cache, TextureCache& texture_cache) {
guest_descriptor_queue.Acquire(scheduler, num_descriptor_entries);
guest_descriptor_queue.Acquire();
buffer_cache.SetComputeUniformBufferState(info.constant_buffer_mask, &uniform_buffer_sizes);
buffer_cache.UnbindComputeStorageBuffers();
@@ -53,7 +53,6 @@ private:
vk::PipelineCache& pipeline_cache;
GuestDescriptorQueue& guest_descriptor_queue;
Shader::Info info;
u32 num_descriptor_entries{};
VideoCommon::ComputeUniformBufferSizes uniform_buffer_sizes{};
@@ -268,7 +268,6 @@ GraphicsPipeline::GraphicsPipeline(
num_textures += Shader::NumDescriptors(info->texture_descriptors);
num_image_elements += Shader::NumDescriptors(info->texture_descriptors);
num_image_elements += Shader::NumDescriptors(info->image_descriptors);
num_descriptor_entries += NumDescriptorEntries(*info);
}
fragment_has_color0_output = stage_infos[NUM_STAGES - 1].stores_frag_color[0];
auto func{[this, shader_notify, &render_pass_cache, &descriptor_pool, pipeline_statistics] {
@@ -474,7 +473,7 @@ bool GraphicsPipeline::ConfigureImpl(bool is_indexed) {
buffer_cache.UpdateGraphicsBuffers(is_indexed);
buffer_cache.BindHostGeometryBuffers(is_indexed);
guest_descriptor_queue.Acquire(scheduler, num_descriptor_entries);
guest_descriptor_queue.Acquire();
RescalingPushConstant rescaling;
RenderAreaPushConstant render_area;
@@ -908,7 +907,7 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) {
// EDS3 - Enables (composite: per-feature)
if (key.state.extended_dynamic_state_3_enables) {
if (device.SupportsDynamicState3DepthClampEnable()) {
if (key.state.dynamic_state3_depth_clamp_enable != 0) {
dynamic_states.push_back(VK_DYNAMIC_STATE_DEPTH_CLAMP_ENABLE_EXT);
}
if (device.SupportsDynamicState3LogicOpEnable()) {
@@ -159,7 +159,6 @@ private:
std::array<Shader::Info, NUM_STAGES> stage_infos;
std::array<u32, 5> enabled_uniform_buffer_masks{};
VideoCommon::UniformBufferSizes uniform_buffer_sizes{};
u32 num_descriptor_entries{};
size_t num_image_elements{};
u32 num_textures{};
bool fragment_has_color0_output{};
@@ -439,21 +439,10 @@ PipelineCache::PipelineCache(Tegra::MaxwellDeviceMemoryManager& device_memory_,
.has_broken_robust =
device.IsNvidia() && device.GetNvidiaArch() <= NvidiaArchitecture::Arch_Pascal,
.min_ssbo_alignment = device.GetStorageBufferAlignment(),
.max_user_clip_distances = device.GetMaxUserClipDistances()
.max_user_clip_distances = device.GetMaxUserClipDistances(),
};
host_info = Shader::HostTranslateInfo{
.min_ssbo_alignment = device.GetStorageBufferAlignment(),
.max_per_stage_descriptor_sampled_images = device.GetMaxPerStageDescriptorSampledImages(),
.max_per_stage_resources = device.GetMaxPerStageResources(),
.max_descriptor_set_samplers = device.GetMaxDescriptorSetSamplers(),
.max_descriptor_set_uniform_buffers = device.GetMaxDescriptorSetUniformBuffers(),
.max_descriptor_set_uniform_buffers_dynamic = device.GetMaxDescriptorSetUniformBuffersDynamic(),
.max_descriptor_set_storage_buffers = device.GetMaxDescriptorSetStorageBuffers(),
.max_descriptor_set_storage_buffers_dynamic = device.GetMaxDescriptorSetStorageBuffersDynamic(),
.max_descriptor_set_sampled_images = device.GetMaxDescriptorSetSampledImages(),
.max_descriptor_set_storage_images = device.GetMaxDescriptorSetStorageImages(),
.max_descriptor_set_input_attachements = device.GetMaxDescriptorSetInputAttachments(),
.support_float64 = device.IsFloat64Supported(),
.support_float16 = device.IsFloat16Supported(),
.support_int64 = device.IsShaderInt64Supported(),
@@ -462,10 +451,13 @@ PipelineCache::PipelineCache(Tegra::MaxwellDeviceMemoryManager& device_memory_,
driver_id == VK_DRIVER_ID_SAMSUNG_PROPRIETARY,
.support_snorm_render_buffer = true,
.support_viewport_index_layer = device.IsExtShaderViewportIndexLayerSupported(),
.min_ssbo_alignment = static_cast<u32>(device.GetStorageBufferAlignment()),
.max_per_stage_descriptor_sampled_images = device.GetMaxPerStageDescriptorSampledImages(),
.max_per_stage_resources = device.GetMaxPerStageResources(),
.max_descriptor_set_sampled_images = device.GetMaxDescriptorSetSampledImages(),
.support_geometry_shader_passthrough = device.IsNvGeometryShaderPassthroughSupported(),
.support_conditional_barrier = device.SupportsConditionalBarriers(),
};
host_info.ApplyDescriptorLimitPolicy();
if (device.GetMaxVertexInputAttributes() < Maxwell::NumVertexAttributes) {
LOG_WARNING(Render_Vulkan, "maxVertexInputAttributes is too low: {} < {}",
@@ -504,10 +496,8 @@ PipelineCache::PipelineCache(Tegra::MaxwellDeviceMemoryManager& device_memory_,
dynamic_features.has_extended_dynamic_state_3_enables &&
device.SupportsDynamicState3DepthClampEnable();
dynamic_features.has_dynamic_state3_logic_op_enable =
dynamic_features.has_extended_dynamic_state_3_enables &&
device.SupportsDynamicState3LogicOpEnable();
dynamic_features.has_dynamic_state3_line_stipple_enable =
dynamic_features.has_extended_dynamic_state_3_enables &&
device.SupportsDynamicState3LineStippleEnable();
// VIDS: Independent toggle (not affected by dyna_state levels)
@@ -113,6 +113,10 @@ public:
[[nodiscard]] ComputePipeline* CurrentComputePipeline();
[[nodiscard]] bool SupportsDynamicState3DepthClampEnable() const {
return dynamic_features.has_dynamic_state3_depth_clamp_enable;
}
void LoadDiskResources(u64 title_id, std::stop_token stop_loading,
const VideoCore::DiskResourceLoadCallback& callback);
@@ -203,7 +203,7 @@ RasterizerVulkan::RasterizerVulkan(Core::Frontend::EmuWindow& emu_window_, Tegra
: gpu{gpu_}, device_memory{device_memory_}, device{device_},
memory_allocator{memory_allocator_}, state_tracker{state_tracker_}, scheduler{scheduler_},
staging_pool(device, memory_allocator, scheduler), descriptor_pool(device, scheduler),
guest_descriptor_queue(device), compute_pass_descriptor_queue(device),
guest_descriptor_queue(device, scheduler), compute_pass_descriptor_queue(device, scheduler),
blit_image(device, scheduler, state_tracker, descriptor_pool), render_pass_cache(device),
texture_cache_runtime{
device, scheduler, memory_allocator, staging_pool,
@@ -1578,7 +1578,7 @@ void RasterizerVulkan::UpdateDepthClampEnable(Tegra::Engines::Maxwell3D::Regs& r
if (!state_tracker.TouchDepthClampEnable()) {
return;
}
if (!device.SupportsDynamicState3DepthClampEnable()) {
if (!pipeline_cache.SupportsDynamicState3DepthClampEnable()) {
return;
}
bool is_enabled = !(regs.viewport_clip_control.geometry_clip ==
@@ -155,14 +155,15 @@ void Scheduler::WaitWorker() {
}
void Scheduler::DispatchWork() {
if (chunk && !chunk->Empty()) {
{
std::scoped_lock ql{queue_mutex};
work_queue.push(std::move(chunk));
}
event_cv.notify_all();
AcquireNewChunk();
if (chunk->Empty()) {
return;
}
{
std::scoped_lock ql{queue_mutex};
work_queue.push(std::move(chunk));
}
event_cv.notify_all();
AcquireNewChunk();
}
void Scheduler::RequestRenderpass(const Framebuffer* framebuffer) {
@@ -6,6 +6,7 @@
#pragma once
#include <atomic>
#include <condition_variable>
#include <cstddef>
#include <functional>
@@ -92,10 +93,12 @@ public:
requires std::is_invocable_v<T, vk::CommandBuffer, vk::CommandBuffer>
void RecordWithUploadBuffer(T&& command) {
if (chunk->Record(command)) {
record_serial.fetch_add(1, std::memory_order_relaxed);
return;
}
DispatchWork();
(void)chunk->Record(command);
record_serial.fetch_add(1, std::memory_order_relaxed);
}
template <typename T>
@@ -117,6 +120,11 @@ public:
return master_semaphore->IsFree(tick);
}
/// Returns a monotonic serial incremented for every recorded command callback.
[[nodiscard]] u64 CurrentRecordSerial() const noexcept {
return record_serial.load(std::memory_order_relaxed);
}
/// Waits for the given GPU tick, optionally pacing frames.
void Wait(u64 tick, double target_fps = 0.0) {
if (tick > 0) {
@@ -298,6 +306,7 @@ private:
u64 frame_counter{};
u64 last_submitted_tick = 0;
std::atomic<u64> record_serial{0};
};
} // namespace Vulkan
@@ -7,7 +7,6 @@
#include <variant>
#include <boost/container/static_vector.hpp>
#include "common/assert.h"
#include "common/logging.h"
#include "video_core/renderer_vulkan/vk_scheduler.h"
#include "video_core/renderer_vulkan/vk_update_descriptor.h"
@@ -16,9 +15,8 @@
namespace Vulkan {
UpdateDescriptorQueue::UpdateDescriptorQueue(const Device& device_)
: device{device_}
{
UpdateDescriptorQueue::UpdateDescriptorQueue(const Device& device_, Scheduler& scheduler_)
: device{device_}, scheduler{scheduler_} {
payload_start = payload.data();
payload_cursor = payload.data();
}
@@ -33,15 +31,13 @@ void UpdateDescriptorQueue::TickFrame() {
payload_cursor = payload_start;
}
void UpdateDescriptorQueue::Acquire(Scheduler& scheduler, size_t required_entries) {
static constexpr size_t DEFAULT_REQUIRED_ENTRIES = 0x400;
const size_t reserve = required_entries > 0 ? required_entries : DEFAULT_REQUIRED_ENTRIES;
ASSERT_MSG(reserve < FRAME_PAYLOAD_SIZE, "Descriptor reservation {} >= frame capacity {}",
reserve, FRAME_PAYLOAD_SIZE);
const size_t used = static_cast<size_t>(std::distance(payload_start, payload_cursor));
if (used + reserve >= FRAME_PAYLOAD_SIZE) {
LOG_WARNING(Render_Vulkan, "Payload overflow (used={}, reserve={}, capacity={})",
used, reserve, FRAME_PAYLOAD_SIZE);
void UpdateDescriptorQueue::Acquire() {
// Minimum number of entries required.
// This is the maximum number of entries a single draw call might use.
static constexpr size_t MIN_ENTRIES = 0x400;
if (std::distance(payload_start, payload_cursor) + MIN_ENTRIES >= FRAME_PAYLOAD_SIZE) {
LOG_WARNING(Render_Vulkan, "Payload overflow, waiting for worker thread");
scheduler.WaitWorker();
payload_cursor = payload_start;
}
@@ -34,11 +34,12 @@ class UpdateDescriptorQueue final {
static constexpr size_t PAYLOAD_SIZE = FRAME_PAYLOAD_SIZE * FRAMES_IN_FLIGHT;
public:
explicit UpdateDescriptorQueue(const Device& device_);
explicit UpdateDescriptorQueue(const Device& device_, Scheduler& scheduler_);
~UpdateDescriptorQueue();
void TickFrame();
void Acquire(Scheduler& scheduler, size_t required_entries = 0);
void Acquire();
const DescriptorUpdateEntry* UpdateData() const noexcept {
return upload_start;
@@ -74,6 +75,8 @@ public:
private:
const Device& device;
Scheduler& scheduler;
size_t frame_index{0};
DescriptorUpdateEntry* payload_cursor = nullptr;
DescriptorUpdateEntry* payload_start = nullptr;
+3 -2
View File
@@ -255,7 +255,8 @@ std::optional<u64> GenericEnvironment::TryFindSize() {
static constexpr u64 SELF_BRANCH_A = 0xE2400FFFFF87000FULL;
static constexpr u64 SELF_BRANCH_B = 0xE2400FFFFF07000FULL;
static constexpr u64 EXIT_VALUE = 0xE30000000007000FULL;
static constexpr u64 MESA_EXIT_MASK = 0xFFF00000000F001FULL;
static constexpr u64 MESA_EXIT_VALUE = (0xE30ULL << 52) | (0x7ULL << 16) | 0xFULL;
code.resize(MAXIMUM_SIZE / INST_SIZE);
@@ -270,7 +271,7 @@ std::optional<u64> GenericEnvironment::TryFindSize() {
if (inst == SELF_BRANCH_A || inst == SELF_BRANCH_B) {
return offset + index;
}
if (!is_proprietary_driver && inst == EXIT_VALUE) {
if ((inst & MESA_EXIT_MASK) == MESA_EXIT_VALUE) {
return offset + index + INST_SIZE;
}
}
+4 -1
View File
@@ -1,3 +1,6 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -102,7 +105,7 @@ struct ImageBase {
VAddr cpu_addr_end = 0;
u64 modification_tick = 0;
size_t lru_index = SIZE_MAX;
u64 last_use_tick = 0;
std::array<u32, MAX_MIP_LEVELS> mip_level_offsets{};
+49 -6
View File
@@ -159,11 +159,54 @@ void TextureCache<P>::RunGarbageCollector() {
}
return false;
};
const auto CollectBelow = [this](u64 threshold) {
boost::container::small_vector<ImageId, 64> expired;
for (auto [id, image] : slot_images) {
if (image->last_use_tick < threshold) {
expired.push_back(id);
}
}
return expired;
};
// Aggressively clear massive sparse textures
if (total_used_memory >= expected_memory) {
auto candidates = CollectBelow(frame_tick);
for (const auto image_id : candidates) {
auto& image = slot_images[image_id];
if (image.info.is_sparse &&
image.guest_size_bytes >= 256_MiB &&
image.allocation_tick < frame_tick - 3) {
LOG_DEBUG(HW_GPU, "GC targeting old sparse texture at 0x{:X} ({} MiB, age: {} frames)",
image.gpu_addr, image.guest_size_bytes / (1024 * 1024),
frame_tick - image.allocation_tick);
if (Cleanup(image_id)) {
break;
}
}
}
}
Configure(false);
lru_cache.ForEachItemBelow(frame_tick - ticks_to_destroy, Cleanup);
{
auto expired = CollectBelow(frame_tick - ticks_to_destroy);
for (const auto image_id : expired) {
if (Cleanup(image_id)) {
break;
}
}
}
// If pressure is still too high, prune aggressively.
if (total_used_memory >= critical_memory) {
Configure(true);
lru_cache.ForEachItemBelow(frame_tick - ticks_to_destroy, Cleanup);
auto expired = CollectBelow(frame_tick - ticks_to_destroy);
for (const auto image_id : expired) {
if (Cleanup(image_id)) {
break;
}
}
}
}
@@ -1865,7 +1908,7 @@ std::pair<u32, u32> TextureCache<P>::PrepareDmaImage(ImageId dst_id, GPUVAddr ba
const auto base = image.TryFindBase(base_addr);
PrepareImage(dst_id, mark_as_modified, false);
const auto& new_image = slot_images[dst_id];
lru_cache.Touch(new_image.lru_index, frame_tick);
new_image.last_use_tick = frame_tick;
return std::make_pair(base->level, base->layer);
}
@@ -2192,7 +2235,7 @@ void TextureCache<P>::RegisterImage(ImageId image_id) {
tentative_size = TranscodedAstcSize(tentative_size, image.info.format);
}
total_used_memory += Common::AlignUp(tentative_size, 1024);
image.lru_index = lru_cache.Insert(image_id, frame_tick);
image.last_use_tick = frame_tick;
ForEachGPUPage(image.gpu_addr, image.guest_size_bytes, [this, image_id](u64 page) {
(*channel_state->gpu_page_table)[page].push_back(image_id);
@@ -2226,7 +2269,7 @@ void TextureCache<P>::UnregisterImage(ImageId image_id) {
"Trying to unregister an already registered image");
image.flags &= ~ImageFlagBits::Registered;
image.flags &= ~ImageFlagBits::BadOverlap;
lru_cache.Free(image.lru_index);
const auto& clear_page_table =
[image_id](u64 page, ankerl::unordered_dense::map<u64, std::vector<ImageId>, Common::IdentityHash<u64>>& selected_page_table) {
const auto page_it = selected_page_table.find(page);
@@ -2554,7 +2597,7 @@ void TextureCache<P>::PrepareImage(ImageId image_id, bool is_modification, bool
if (is_modification) {
MarkModification(image);
}
lru_cache.Touch(image.lru_index, frame_tick);
image.last_use_tick = frame_tick;
}
template <class P>
@@ -23,7 +23,7 @@
#include "common/common_types.h"
#include "common/hash.h"
#include "common/literals.h"
#include "common/lru_cache.h"
#include <ranges>
#include "common/scratch_buffer.h"
#include "common/slot_vector.h"
@@ -485,11 +485,7 @@ private:
std::deque<std::vector<AsyncBuffer>> async_buffers;
std::deque<AsyncBuffer> async_buffers_death_ring;
struct LRUItemParams {
using ObjectType = ImageId;
using TickType = u64;
};
Common::LeastRecentlyUsedCache<LRUItemParams> lru_cache;
#ifdef YUZU_LEGACY
static constexpr size_t TICKS_TO_DESTROY = 6;
+26 -17
View File
@@ -321,23 +321,32 @@ public:
return properties.properties.limits.maxPushConstantsSize;
}
#define FN_MAX_LIMIT_LIST \
FN_MAX_LIMIT_ELEM(ComputeSharedMemorySize) \
FN_MAX_LIMIT_ELEM(PerStageDescriptorSampledImages) \
FN_MAX_LIMIT_ELEM(PerStageResources) \
FN_MAX_LIMIT_ELEM(DescriptorSetSamplers) \
FN_MAX_LIMIT_ELEM(DescriptorSetUniformBuffers) \
FN_MAX_LIMIT_ELEM(DescriptorSetUniformBuffersDynamic) \
FN_MAX_LIMIT_ELEM(DescriptorSetStorageBuffers) \
FN_MAX_LIMIT_ELEM(DescriptorSetStorageBuffersDynamic) \
FN_MAX_LIMIT_ELEM(DescriptorSetSampledImages) \
FN_MAX_LIMIT_ELEM(DescriptorSetStorageImages) \
FN_MAX_LIMIT_ELEM(DescriptorSetInputAttachments)
#define FN_MAX_LIMIT_ELEM(name) \
u32 GetMax##name() const { return properties.properties.limits.max##name; }
FN_MAX_LIMIT_LIST
#undef FN_MAX_LIMIT_ELEM
#undef FN_MAX_LIMIT_LIST
/// Returns the maximum size for shared memory.
u32 GetMaxComputeSharedMemorySize() const {
return properties.properties.limits.maxComputeSharedMemorySize;
}
/// Returns the maximum number of dynamic storage buffer descriptors per set.
u32 GetMaxDescriptorSetStorageBuffersDynamic() const {
return properties.properties.limits.maxDescriptorSetStorageBuffersDynamic;
}
/// Returns the maximum number of dynamic uniform buffer descriptors per set.
u32 GetMaxDescriptorSetUniformBuffersDynamic() const {
return properties.properties.limits.maxDescriptorSetUniformBuffersDynamic;
}
u32 GetMaxPerStageDescriptorSampledImages() const {
return properties.properties.limits.maxPerStageDescriptorSampledImages;
}
u32 GetMaxPerStageResources() const {
return properties.properties.limits.maxPerStageResources;
}
u32 GetMaxDescriptorSetSampledImages() const {
return properties.properties.limits.maxDescriptorSetSampledImages;
}
/// Returns float control properties of the device.
const VkPhysicalDeviceFloatControlsPropertiesKHR& FloatControlProperties() const {
-1
View File
@@ -44,7 +44,6 @@ int IrShaderRecompilerImpl(int argc, char *argv[]) {
host_info.support_geometry_shader_passthrough = true;
host_info.support_conditional_barrier = true;
host_info.min_ssbo_alignment = 0;
host_info.ApplyDescriptorLimitPolicy();
auto program = Shader::Maxwell::TranslateProgram(inst_pool, block_pool, env, cfg, host_info);
auto const dumped_ir = Shader::IR::DumpProgram(program);
std::printf("%s\n", dumped_ir.c_str());
@@ -52,7 +52,6 @@ int SpirvShaderRecompilerImpl(int argc, char *argv[]) {
host_info.support_geometry_shader_passthrough = true;
host_info.support_conditional_barrier = true;
host_info.min_ssbo_alignment = 0;
host_info.ApplyDescriptorLimitPolicy();
auto program = Shader::Maxwell::TranslateProgram(inst_pool, block_pool, env, cfg, host_info);
// IR::Program TranslateProgram(ObjectPool<IR::Inst>& inst_pool, ObjectPool<IR::Block>& block_pool,