summaryrefslogtreecommitdiffstats
path: root/src
diff options
context:
space:
mode:
Diffstat (limited to 'src')
-rw-r--r--src/audio_core/audio_renderer.cpp14
-rw-r--r--src/audio_core/audio_renderer.h3
-rw-r--r--src/common/CMakeLists.txt2
-rw-r--r--src/core/CMakeLists.txt3
-rw-r--r--src/core/arm/arm_interface.h7
-rw-r--r--src/core/arm/dynarmic/arm_dynarmic.cpp9
-rw-r--r--src/core/arm/dynarmic/arm_dynarmic.h3
-rw-r--r--src/core/arm/unicorn/arm_unicorn.cpp9
-rw-r--r--src/core/arm/unicorn/arm_unicorn.h3
-rw-r--r--src/core/core.cpp14
-rw-r--r--src/core/core.h8
-rw-r--r--src/core/core_cpu.cpp19
-rw-r--r--src/core/hle/kernel/svc.cpp9
-rw-r--r--src/core/hle/kernel/vm_manager.cpp20
-rw-r--r--src/core/hle/service/am/am.cpp58
-rw-r--r--src/core/hle/service/am/am.h8
-rw-r--r--src/core/hle/service/am/applet_ae.cpp4
-rw-r--r--src/core/hle/service/am/applet_oe.cpp2
-rw-r--r--src/core/hle/service/apm/apm.cpp13
-rw-r--r--src/core/hle/service/apm/apm.h7
-rw-r--r--src/core/hle/service/apm/controller.cpp68
-rw-r--r--src/core/hle/service/apm/controller.h70
-rw-r--r--src/core/hle/service/apm/interface.cpp82
-rw-r--r--src/core/hle/service/apm/interface.h14
-rw-r--r--src/core/hle/service/audio/audren_u.cpp9
-rw-r--r--src/core/hle/service/audio/audren_u.h1
-rw-r--r--src/core/hle/service/filesystem/filesystem.cpp10
-rw-r--r--src/core/hle/service/filesystem/filesystem.h2
-rw-r--r--src/core/hle/service/filesystem/fsp_srv.cpp54
-rw-r--r--src/core/hle/service/filesystem/fsp_srv.h26
-rw-r--r--src/core/hle/service/friend/friend.cpp35
-rw-r--r--src/core/hle/service/hid/controllers/npad.cpp31
-rw-r--r--src/core/hle/service/hid/controllers/npad.h5
-rw-r--r--src/core/hle/service/hid/errors.h13
-rw-r--r--src/core/hle/service/hid/hid.cpp50
-rw-r--r--src/core/hle/service/hid/hid.h3
-rw-r--r--src/core/hle/service/ldr/ldr.cpp32
-rw-r--r--src/core/hle/service/mii/mii.cpp16
-rw-r--r--src/core/hle/service/pm/pm.cpp124
-rw-r--r--src/core/hle/service/pm/pm.h6
-rw-r--r--src/core/hle/service/service.cpp9
-rw-r--r--src/core/hle/service/service.h3
-rw-r--r--src/core/reporter.cpp18
-rw-r--r--src/core/reporter.h7
-rw-r--r--src/core/settings.cpp1
-rw-r--r--src/core/settings.h1
-rw-r--r--src/core/telemetry_session.cpp1
-rw-r--r--src/video_core/CMakeLists.txt2
-rw-r--r--src/video_core/engines/shader_bytecode.h19
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer.h16
-rw-r--r--src/video_core/renderer_opengl/gl_sampler_cache.h4
-rw-r--r--src/video_core/renderer_opengl/gl_shader_cache.cpp22
-rw-r--r--src/video_core/renderer_opengl/gl_shader_decompiler.cpp19
-rw-r--r--src/video_core/renderer_opengl/gl_shader_gen.cpp8
-rw-r--r--src/video_core/renderer_opengl/gl_shader_gen.h2
-rw-r--r--src/video_core/renderer_vulkan/vk_buffer_cache.cpp4
-rw-r--r--src/video_core/renderer_vulkan/vk_buffer_cache.h2
-rw-r--r--src/video_core/renderer_vulkan/vk_sampler_cache.h7
-rw-r--r--src/video_core/renderer_vulkan/vk_scheduler.cpp16
-rw-r--r--src/video_core/renderer_vulkan/vk_scheduler.h78
-rw-r--r--src/video_core/renderer_vulkan/vk_shader_decompiler.cpp9
-rw-r--r--src/video_core/renderer_vulkan/vk_stream_buffer.cpp8
-rw-r--r--src/video_core/renderer_vulkan/vk_stream_buffer.h2
-rw-r--r--src/video_core/shader/control_flow.cpp476
-rw-r--r--src/video_core/shader/control_flow.h63
-rw-r--r--src/video_core/shader/decode.cpp177
-rw-r--r--src/video_core/shader/decode/decode_integer_set.cpp0
-rw-r--r--src/video_core/shader/decode/other.cpp58
-rw-r--r--src/video_core/shader/decode/texture.cpp8
-rw-r--r--src/video_core/shader/node.h11
-rw-r--r--src/video_core/shader/shader_ir.cpp4
-rw-r--r--src/video_core/shader/shader_ir.h27
-rw-r--r--src/yuzu/configuration/config.cpp3
-rw-r--r--src/yuzu_cmd/config.cpp1
-rw-r--r--src/yuzu_cmd/default_ini.h4
-rw-r--r--src/yuzu_tester/config.cpp1
-rw-r--r--src/yuzu_tester/default_ini.h4
77 files changed, 1568 insertions, 393 deletions
diff --git a/src/audio_core/audio_renderer.cpp b/src/audio_core/audio_renderer.cpp
index 9a0939883..da50a0bbc 100644
--- a/src/audio_core/audio_renderer.cpp
+++ b/src/audio_core/audio_renderer.cpp
@@ -73,13 +73,15 @@ private:
EffectInStatus info{};
};
AudioRenderer::AudioRenderer(Core::Timing::CoreTiming& core_timing, AudioRendererParameter params,
- Kernel::SharedPtr<Kernel::WritableEvent> buffer_event)
+ Kernel::SharedPtr<Kernel::WritableEvent> buffer_event,
+ std::size_t instance_number)
: worker_params{params}, buffer_event{buffer_event}, voices(params.voice_count),
effects(params.effect_count) {
audio_out = std::make_unique<AudioCore::AudioOut>();
stream = audio_out->OpenStream(core_timing, STREAM_SAMPLE_RATE, STREAM_NUM_CHANNELS,
- "AudioRenderer", [=]() { buffer_event->Signal(); });
+ fmt::format("AudioRenderer-Instance{}", instance_number),
+ [=]() { buffer_event->Signal(); });
audio_out->StartStream(stream);
QueueMixedBuffer(0);
@@ -217,13 +219,15 @@ std::vector<s16> AudioRenderer::VoiceState::DequeueSamples(std::size_t sample_co
if (offset == samples.size()) {
offset = 0;
- if (!wave_buffer.is_looping) {
+ if (!wave_buffer.is_looping && wave_buffer.buffer_sz) {
SetWaveIndex(wave_index + 1);
}
- out_status.wave_buffer_consumed++;
+ if (wave_buffer.buffer_sz) {
+ out_status.wave_buffer_consumed++;
+ }
- if (wave_buffer.end_of_stream) {
+ if (wave_buffer.end_of_stream || wave_buffer.buffer_sz == 0) {
info.play_state = PlayState::Paused;
}
}
diff --git a/src/audio_core/audio_renderer.h b/src/audio_core/audio_renderer.h
index b2e5d336c..45afbe759 100644
--- a/src/audio_core/audio_renderer.h
+++ b/src/audio_core/audio_renderer.h
@@ -215,7 +215,8 @@ static_assert(sizeof(UpdateDataHeader) == 0x40, "UpdateDataHeader has wrong size
class AudioRenderer {
public:
AudioRenderer(Core::Timing::CoreTiming& core_timing, AudioRendererParameter params,
- Kernel::SharedPtr<Kernel::WritableEvent> buffer_event);
+ Kernel::SharedPtr<Kernel::WritableEvent> buffer_event,
+ std::size_t instance_number);
~AudioRenderer();
std::vector<u8> UpdateAudioRenderer(const std::vector<u8>& input_params);
diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt
index 2554add28..2b4266f29 100644
--- a/src/common/CMakeLists.txt
+++ b/src/common/CMakeLists.txt
@@ -56,6 +56,8 @@ add_custom_command(OUTPUT scm_rev.cpp
"${VIDEO_CORE}/shader/decode/shift.cpp"
"${VIDEO_CORE}/shader/decode/video.cpp"
"${VIDEO_CORE}/shader/decode/xmad.cpp"
+ "${VIDEO_CORE}/shader/control_flow.cpp"
+ "${VIDEO_CORE}/shader/control_flow.h"
"${VIDEO_CORE}/shader/decode.cpp"
"${VIDEO_CORE}/shader/node.h"
"${VIDEO_CORE}/shader/node_helper.cpp"
diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt
index 30eb9d82e..f4325f0f8 100644
--- a/src/core/CMakeLists.txt
+++ b/src/core/CMakeLists.txt
@@ -208,6 +208,8 @@ add_library(core STATIC
hle/service/aoc/aoc_u.h
hle/service/apm/apm.cpp
hle/service/apm/apm.h
+ hle/service/apm/controller.cpp
+ hle/service/apm/controller.h
hle/service/apm/interface.cpp
hle/service/apm/interface.h
hle/service/audio/audctl.cpp
@@ -293,6 +295,7 @@ add_library(core STATIC
hle/service/hid/irs.h
hle/service/hid/xcd.cpp
hle/service/hid/xcd.h
+ hle/service/hid/errors.h
hle/service/hid/controllers/controller_base.cpp
hle/service/hid/controllers/controller_base.h
hle/service/hid/controllers/debug_pad.cpp
diff --git a/src/core/arm/arm_interface.h b/src/core/arm/arm_interface.h
index c6691a8e1..45e94e625 100644
--- a/src/core/arm/arm_interface.h
+++ b/src/core/arm/arm_interface.h
@@ -44,13 +44,6 @@ public:
/// Step CPU by one instruction
virtual void Step() = 0;
- /// Maps a backing memory region for the CPU
- virtual void MapBackingMemory(VAddr address, std::size_t size, u8* memory,
- Kernel::VMAPermission perms) = 0;
-
- /// Unmaps a region of memory that was previously mapped using MapBackingMemory
- virtual void UnmapMemory(VAddr address, std::size_t size) = 0;
-
/// Clear all instruction cache
virtual void ClearInstructionCache() = 0;
diff --git a/src/core/arm/dynarmic/arm_dynarmic.cpp b/src/core/arm/dynarmic/arm_dynarmic.cpp
index 44307fa19..f1506b372 100644
--- a/src/core/arm/dynarmic/arm_dynarmic.cpp
+++ b/src/core/arm/dynarmic/arm_dynarmic.cpp
@@ -177,15 +177,6 @@ ARM_Dynarmic::ARM_Dynarmic(System& system, ExclusiveMonitor& exclusive_monitor,
ARM_Dynarmic::~ARM_Dynarmic() = default;
-void ARM_Dynarmic::MapBackingMemory(u64 address, std::size_t size, u8* memory,
- Kernel::VMAPermission perms) {
- inner_unicorn.MapBackingMemory(address, size, memory, perms);
-}
-
-void ARM_Dynarmic::UnmapMemory(u64 address, std::size_t size) {
- inner_unicorn.UnmapMemory(address, size);
-}
-
void ARM_Dynarmic::SetPC(u64 pc) {
jit->SetPC(pc);
}
diff --git a/src/core/arm/dynarmic/arm_dynarmic.h b/src/core/arm/dynarmic/arm_dynarmic.h
index b701e97a3..504d46c68 100644
--- a/src/core/arm/dynarmic/arm_dynarmic.h
+++ b/src/core/arm/dynarmic/arm_dynarmic.h
@@ -23,9 +23,6 @@ public:
ARM_Dynarmic(System& system, ExclusiveMonitor& exclusive_monitor, std::size_t core_index);
~ARM_Dynarmic() override;
- void MapBackingMemory(VAddr address, std::size_t size, u8* memory,
- Kernel::VMAPermission perms) override;
- void UnmapMemory(u64 address, std::size_t size) override;
void SetPC(u64 pc) override;
u64 GetPC() const override;
u64 GetReg(int index) const override;
diff --git a/src/core/arm/unicorn/arm_unicorn.cpp b/src/core/arm/unicorn/arm_unicorn.cpp
index 4e07fe8b5..b0ee7821a 100644
--- a/src/core/arm/unicorn/arm_unicorn.cpp
+++ b/src/core/arm/unicorn/arm_unicorn.cpp
@@ -76,15 +76,6 @@ ARM_Unicorn::~ARM_Unicorn() {
CHECKED(uc_close(uc));
}
-void ARM_Unicorn::MapBackingMemory(VAddr address, std::size_t size, u8* memory,
- Kernel::VMAPermission perms) {
- CHECKED(uc_mem_map_ptr(uc, address, size, static_cast<u32>(perms), memory));
-}
-
-void ARM_Unicorn::UnmapMemory(VAddr address, std::size_t size) {
- CHECKED(uc_mem_unmap(uc, address, size));
-}
-
void ARM_Unicorn::SetPC(u64 pc) {
CHECKED(uc_reg_write(uc, UC_ARM64_REG_PC, &pc));
}
diff --git a/src/core/arm/unicorn/arm_unicorn.h b/src/core/arm/unicorn/arm_unicorn.h
index 34e974b4d..fe2ffd70c 100644
--- a/src/core/arm/unicorn/arm_unicorn.h
+++ b/src/core/arm/unicorn/arm_unicorn.h
@@ -18,9 +18,6 @@ public:
explicit ARM_Unicorn(System& system);
~ARM_Unicorn() override;
- void MapBackingMemory(VAddr address, std::size_t size, u8* memory,
- Kernel::VMAPermission perms) override;
- void UnmapMemory(VAddr address, std::size_t size) override;
void SetPC(u64 pc) override;
u64 GetPC() const override;
u64 GetReg(int index) const override;
diff --git a/src/core/core.cpp b/src/core/core.cpp
index 262411db8..4aceee785 100644
--- a/src/core/core.cpp
+++ b/src/core/core.cpp
@@ -25,6 +25,7 @@
#include "core/hle/kernel/scheduler.h"
#include "core/hle/kernel/thread.h"
#include "core/hle/service/am/applets/applets.h"
+#include "core/hle/service/apm/controller.h"
#include "core/hle/service/glue/manager.h"
#include "core/hle/service/service.h"
#include "core/hle/service/sm/sm.h"
@@ -143,7 +144,7 @@ struct System::Impl {
telemetry_session = std::make_unique<Core::TelemetrySession>();
service_manager = std::make_shared<Service::SM::ServiceManager>();
- Service::Init(service_manager, system, *virtual_filesystem);
+ Service::Init(service_manager, system);
GDBStub::Init();
renderer = VideoCore::CreateRenderer(emu_window, system);
@@ -306,6 +307,9 @@ struct System::Impl {
/// Frontend applets
Service::AM::Applets::AppletManager applet_manager;
+ /// APM (Performance) services
+ Service::APM::Controller apm_controller{core_timing};
+
/// Glue services
Service::Glue::ARPManager arp_manager;
@@ -568,6 +572,14 @@ const Service::Glue::ARPManager& System::GetARPManager() const {
return impl->arp_manager;
}
+Service::APM::Controller& System::GetAPMController() {
+ return impl->apm_controller;
+}
+
+const Service::APM::Controller& System::GetAPMController() const {
+ return impl->apm_controller;
+}
+
System::ResultStatus System::Init(Frontend::EmuWindow& emu_window) {
return impl->Init(*this, emu_window);
}
diff --git a/src/core/core.h b/src/core/core.h
index 70adb7af9..11e73278e 100644
--- a/src/core/core.h
+++ b/src/core/core.h
@@ -43,6 +43,10 @@ struct AppletFrontendSet;
class AppletManager;
} // namespace AM::Applets
+namespace APM {
+class Controller;
+}
+
namespace Glue {
class ARPManager;
}
@@ -296,6 +300,10 @@ public:
const Service::Glue::ARPManager& GetARPManager() const;
+ Service::APM::Controller& GetAPMController();
+
+ const Service::APM::Controller& GetAPMController() const;
+
private:
System();
diff --git a/src/core/core_cpu.cpp b/src/core/core_cpu.cpp
index 99b7d387d..21c410e34 100644
--- a/src/core/core_cpu.cpp
+++ b/src/core/core_cpu.cpp
@@ -53,16 +53,12 @@ bool CpuBarrier::Rendezvous() {
Cpu::Cpu(System& system, ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_barrier,
std::size_t core_index)
: cpu_barrier{cpu_barrier}, core_timing{system.CoreTiming()}, core_index{core_index} {
- if (Settings::values.cpu_jit_enabled) {
#ifdef ARCHITECTURE_x86_64
- arm_interface = std::make_unique<ARM_Dynarmic>(system, exclusive_monitor, core_index);
+ arm_interface = std::make_unique<ARM_Dynarmic>(system, exclusive_monitor, core_index);
#else
- arm_interface = std::make_unique<ARM_Unicorn>(system);
- LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
+ arm_interface = std::make_unique<ARM_Unicorn>(system);
+ LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available");
#endif
- } else {
- arm_interface = std::make_unique<ARM_Unicorn>(system);
- }
scheduler = std::make_unique<Kernel::Scheduler>(system, *arm_interface);
}
@@ -70,15 +66,12 @@ Cpu::Cpu(System& system, ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_ba
Cpu::~Cpu() = default;
std::unique_ptr<ExclusiveMonitor> Cpu::MakeExclusiveMonitor(std::size_t num_cores) {
- if (Settings::values.cpu_jit_enabled) {
#ifdef ARCHITECTURE_x86_64
- return std::make_unique<DynarmicExclusiveMonitor>(num_cores);
+ return std::make_unique<DynarmicExclusiveMonitor>(num_cores);
#else
- return nullptr; // TODO(merry): Passthrough exclusive monitor
+ // TODO(merry): Passthrough exclusive monitor
+ return nullptr;
#endif
- } else {
- return nullptr; // TODO(merry): Passthrough exclusive monitor
- }
}
void Cpu::RunLoop(bool tight_loop) {
diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp
index f9fe7dce0..a46eed3da 100644
--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -318,7 +318,14 @@ static ResultCode UnmapMemory(Core::System& system, VAddr dst_addr, VAddr src_ad
return result;
}
- return vm_manager.UnmapRange(dst_addr, size);
+ const auto unmap_res = vm_manager.UnmapRange(dst_addr, size);
+
+ // Reprotect the source mapping on success
+ if (unmap_res.IsSuccess()) {
+ ASSERT(vm_manager.ReprotectRange(src_addr, size, VMAPermission::ReadWrite).IsSuccess());
+ }
+
+ return unmap_res;
}
/// Connect to an OS service given the port name, returns the handle to the port to out
diff --git a/src/core/hle/kernel/vm_manager.cpp b/src/core/hle/kernel/vm_manager.cpp
index 02170d548..4f45fb03b 100644
--- a/src/core/hle/kernel/vm_manager.cpp
+++ b/src/core/hle/kernel/vm_manager.cpp
@@ -8,7 +8,6 @@
#include "common/assert.h"
#include "common/logging/log.h"
#include "common/memory_hook.h"
-#include "core/arm/arm_interface.h"
#include "core/core.h"
#include "core/file_sys/program_metadata.h"
#include "core/hle/kernel/errors.h"
@@ -115,15 +114,6 @@ ResultVal<VMManager::VMAHandle> VMManager::MapMemoryBlock(VAddr target,
VirtualMemoryArea& final_vma = vma_handle->second;
ASSERT(final_vma.size == size);
- system.ArmInterface(0).MapBackingMemory(target, size, block->data() + offset,
- VMAPermission::ReadWriteExecute);
- system.ArmInterface(1).MapBackingMemory(target, size, block->data() + offset,
- VMAPermission::ReadWriteExecute);
- system.ArmInterface(2).MapBackingMemory(target, size, block->data() + offset,
- VMAPermission::ReadWriteExecute);
- system.ArmInterface(3).MapBackingMemory(target, size, block->data() + offset,
- VMAPermission::ReadWriteExecute);
-
final_vma.type = VMAType::AllocatedMemoryBlock;
final_vma.permissions = perm;
final_vma.state = state;
@@ -143,11 +133,6 @@ ResultVal<VMManager::VMAHandle> VMManager::MapBackingMemory(VAddr target, u8* me
VirtualMemoryArea& final_vma = vma_handle->second;
ASSERT(final_vma.size == size);
- system.ArmInterface(0).MapBackingMemory(target, size, memory, VMAPermission::ReadWriteExecute);
- system.ArmInterface(1).MapBackingMemory(target, size, memory, VMAPermission::ReadWriteExecute);
- system.ArmInterface(2).MapBackingMemory(target, size, memory, VMAPermission::ReadWriteExecute);
- system.ArmInterface(3).MapBackingMemory(target, size, memory, VMAPermission::ReadWriteExecute);
-
final_vma.type = VMAType::BackingMemory;
final_vma.permissions = VMAPermission::ReadWrite;
final_vma.state = state;
@@ -236,11 +221,6 @@ ResultCode VMManager::UnmapRange(VAddr target, u64 size) {
ASSERT(FindVMA(target)->second.size >= size);
- system.ArmInterface(0).UnmapMemory(target, size);
- system.ArmInterface(1).UnmapMemory(target, size);
- system.ArmInterface(2).UnmapMemory(target, size);
- system.ArmInterface(3).UnmapMemory(target, size);
-
return RESULT_SUCCESS;
}
diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp
index eced38001..a192a1f5f 100644
--- a/src/core/hle/service/am/am.cpp
+++ b/src/core/hle/service/am/am.cpp
@@ -29,7 +29,8 @@
#include "core/hle/service/am/omm.h"
#include "core/hle/service/am/spsm.h"
#include "core/hle/service/am/tcap.h"
-#include "core/hle/service/apm/apm.h"
+#include "core/hle/service/apm/controller.h"
+#include "core/hle/service/apm/interface.h"
#include "core/hle/service/filesystem/filesystem.h"
#include "core/hle/service/ns/ns.h"
#include "core/hle/service/nvflinger/nvflinger.h"
@@ -265,8 +266,8 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger
{65, nullptr, "ReportUserIsActive"},
{66, nullptr, "GetCurrentIlluminance"},
{67, nullptr, "IsIlluminanceAvailable"},
- {68, nullptr, "SetAutoSleepDisabled"},
- {69, nullptr, "IsAutoSleepDisabled"},
+ {68, &ISelfController::SetAutoSleepDisabled, "SetAutoSleepDisabled"},
+ {69, &ISelfController::IsAutoSleepDisabled, "IsAutoSleepDisabled"},
{70, nullptr, "ReportMultimediaError"},
{71, nullptr, "GetCurrentIlluminanceEx"},
{80, nullptr, "SetWirelessPriorityMode"},
@@ -453,6 +454,34 @@ void ISelfController::GetIdleTimeDetectionExtension(Kernel::HLERequestContext& c
rb.Push<u32>(idle_time_detection_extension);
}
+void ISelfController::SetAutoSleepDisabled(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ is_auto_sleep_disabled = rp.Pop<bool>();
+
+ // On the system itself, if the previous state of is_auto_sleep_disabled
+ // differed from the current value passed in, it'd signify the internal
+ // window manager to update (and also increment some statistics like update counts)
+ //
+ // It'd also indicate this change to an idle handling context.
+ //
+ // However, given we're emulating this behavior, most of this can be ignored
+ // and it's sufficient to simply set the member variable for querying via
+ // IsAutoSleepDisabled().
+
+ LOG_DEBUG(Service_AM, "called. is_auto_sleep_disabled={}", is_auto_sleep_disabled);
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(RESULT_SUCCESS);
+}
+
+void ISelfController::IsAutoSleepDisabled(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_AM, "called.");
+
+ IPC::ResponseBuilder rb{ctx, 3};
+ rb.Push(RESULT_SUCCESS);
+ rb.Push(is_auto_sleep_disabled);
+}
+
void ISelfController::GetAccumulatedSuspendedTickValue(Kernel::HLERequestContext& ctx) {
LOG_DEBUG(Service_AM, "called.");
@@ -520,8 +549,9 @@ void AppletMessageQueue::OperationModeChanged() {
on_operation_mode_changed.writable->Signal();
}
-ICommonStateGetter::ICommonStateGetter(std::shared_ptr<AppletMessageQueue> msg_queue)
- : ServiceFramework("ICommonStateGetter"), msg_queue(std::move(msg_queue)) {
+ICommonStateGetter::ICommonStateGetter(Core::System& system,
+ std::shared_ptr<AppletMessageQueue> msg_queue)
+ : ServiceFramework("ICommonStateGetter"), system(system), msg_queue(std::move(msg_queue)) {
// clang-format off
static const FunctionInfo functions[] = {
{0, &ICommonStateGetter::GetEventHandle, "GetEventHandle"},
@@ -554,7 +584,7 @@ ICommonStateGetter::ICommonStateGetter(std::shared_ptr<AppletMessageQueue> msg_q
{63, nullptr, "GetHdcpAuthenticationStateChangeEvent"},
{64, nullptr, "SetTvPowerStateMatchingMode"},
{65, nullptr, "GetApplicationIdByContentActionName"},
- {66, nullptr, "SetCpuBoostMode"},
+ {66, &ICommonStateGetter::SetCpuBoostMode, "SetCpuBoostMode"},
{80, nullptr, "PerformSystemButtonPressingIfInFocus"},
{90, nullptr, "SetPerformanceConfigurationChangedNotification"},
{91, nullptr, "GetCurrentPerformanceConfiguration"},
@@ -635,6 +665,16 @@ void ICommonStateGetter::GetDefaultDisplayResolution(Kernel::HLERequestContext&
}
}
+void ICommonStateGetter::SetCpuBoostMode(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_AM, "called, forwarding to APM:SYS");
+
+ const auto& sm = system.ServiceManager();
+ const auto apm_sys = sm.GetService<APM::APM_Sys>("apm:sys");
+ ASSERT(apm_sys != nullptr);
+
+ apm_sys->SetCpuBoostMode(ctx);
+}
+
IStorage::IStorage(std::vector<u8> buffer)
: ServiceFramework("IStorage"), buffer(std::move(buffer)) {
// clang-format off
@@ -663,13 +703,11 @@ void ICommonStateGetter::GetOperationMode(Kernel::HLERequestContext& ctx) {
}
void ICommonStateGetter::GetPerformanceMode(Kernel::HLERequestContext& ctx) {
- const bool use_docked_mode{Settings::values.use_docked_mode};
- LOG_DEBUG(Service_AM, "called, use_docked_mode={}", use_docked_mode);
+ LOG_DEBUG(Service_AM, "called");
IPC::ResponseBuilder rb{ctx, 3};
rb.Push(RESULT_SUCCESS);
- rb.Push(static_cast<u32>(use_docked_mode ? APM::PerformanceMode::Docked
- : APM::PerformanceMode::Handheld));
+ rb.PushEnum(system.GetAPMController().GetCurrentPerformanceMode());
}
class ILibraryAppletAccessor final : public ServiceFramework<ILibraryAppletAccessor> {
diff --git a/src/core/hle/service/am/am.h b/src/core/hle/service/am/am.h
index 49ff20959..6cb582483 100644
--- a/src/core/hle/service/am/am.h
+++ b/src/core/hle/service/am/am.h
@@ -133,6 +133,8 @@ private:
void SetHandlesRequestToDisplay(Kernel::HLERequestContext& ctx);
void SetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx);
void GetIdleTimeDetectionExtension(Kernel::HLERequestContext& ctx);
+ void SetAutoSleepDisabled(Kernel::HLERequestContext& ctx);
+ void IsAutoSleepDisabled(Kernel::HLERequestContext& ctx);
void GetAccumulatedSuspendedTickValue(Kernel::HLERequestContext& ctx);
void GetAccumulatedSuspendedTickChangedEvent(Kernel::HLERequestContext& ctx);
@@ -142,11 +144,13 @@ private:
u32 idle_time_detection_extension = 0;
u64 num_fatal_sections_entered = 0;
+ bool is_auto_sleep_disabled = false;
};
class ICommonStateGetter final : public ServiceFramework<ICommonStateGetter> {
public:
- explicit ICommonStateGetter(std::shared_ptr<AppletMessageQueue> msg_queue);
+ explicit ICommonStateGetter(Core::System& system,
+ std::shared_ptr<AppletMessageQueue> msg_queue);
~ICommonStateGetter() override;
private:
@@ -168,7 +172,9 @@ private:
void GetPerformanceMode(Kernel::HLERequestContext& ctx);
void GetBootMode(Kernel::HLERequestContext& ctx);
void GetDefaultDisplayResolution(Kernel::HLERequestContext& ctx);
+ void SetCpuBoostMode(Kernel::HLERequestContext& ctx);
+ Core::System& system;
std::shared_ptr<AppletMessageQueue> msg_queue;
};
diff --git a/src/core/hle/service/am/applet_ae.cpp b/src/core/hle/service/am/applet_ae.cpp
index fe5beb8f9..a34368c8b 100644
--- a/src/core/hle/service/am/applet_ae.cpp
+++ b/src/core/hle/service/am/applet_ae.cpp
@@ -42,7 +42,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(RESULT_SUCCESS);
- rb.PushIpcInterface<ICommonStateGetter>(msg_queue);
+ rb.PushIpcInterface<ICommonStateGetter>(system, msg_queue);
}
void GetSelfController(Kernel::HLERequestContext& ctx) {
@@ -146,7 +146,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(RESULT_SUCCESS);
- rb.PushIpcInterface<ICommonStateGetter>(msg_queue);
+ rb.PushIpcInterface<ICommonStateGetter>(system, msg_queue);
}
void GetSelfController(Kernel::HLERequestContext& ctx) {
diff --git a/src/core/hle/service/am/applet_oe.cpp b/src/core/hle/service/am/applet_oe.cpp
index 6e255fe95..5d53ef113 100644
--- a/src/core/hle/service/am/applet_oe.cpp
+++ b/src/core/hle/service/am/applet_oe.cpp
@@ -80,7 +80,7 @@ private:
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(RESULT_SUCCESS);
- rb.PushIpcInterface<ICommonStateGetter>(msg_queue);
+ rb.PushIpcInterface<ICommonStateGetter>(system, msg_queue);
}
void GetLibraryAppletCreator(Kernel::HLERequestContext& ctx) {
diff --git a/src/core/hle/service/apm/apm.cpp b/src/core/hle/service/apm/apm.cpp
index f3c09bbb1..85bbf5988 100644
--- a/src/core/hle/service/apm/apm.cpp
+++ b/src/core/hle/service/apm/apm.cpp
@@ -2,7 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include "common/logging/log.h"
#include "core/hle/ipc_helpers.h"
#include "core/hle/service/apm/apm.h"
#include "core/hle/service/apm/interface.h"
@@ -12,11 +11,15 @@ namespace Service::APM {
Module::Module() = default;
Module::~Module() = default;
-void InstallInterfaces(SM::ServiceManager& service_manager) {
+void InstallInterfaces(Core::System& system) {
auto module_ = std::make_shared<Module>();
- std::make_shared<APM>(module_, "apm")->InstallAsService(service_manager);
- std::make_shared<APM>(module_, "apm:p")->InstallAsService(service_manager);
- std::make_shared<APM_Sys>()->InstallAsService(service_manager);
+ std::make_shared<APM>(module_, system.GetAPMController(), "apm")
+ ->InstallAsService(system.ServiceManager());
+ std::make_shared<APM>(module_, system.GetAPMController(), "apm:p")
+ ->InstallAsService(system.ServiceManager());
+ std::make_shared<APM>(module_, system.GetAPMController(), "apm:am")
+ ->InstallAsService(system.ServiceManager());
+ std::make_shared<APM_Sys>(system.GetAPMController())->InstallAsService(system.ServiceManager());
}
} // namespace Service::APM
diff --git a/src/core/hle/service/apm/apm.h b/src/core/hle/service/apm/apm.h
index 4d7d5bb7c..cf4c2bb11 100644
--- a/src/core/hle/service/apm/apm.h
+++ b/src/core/hle/service/apm/apm.h
@@ -8,11 +8,6 @@
namespace Service::APM {
-enum class PerformanceMode : u8 {
- Handheld = 0,
- Docked = 1,
-};
-
class Module final {
public:
Module();
@@ -20,6 +15,6 @@ public:
};
/// Registers all AM services with the specified service manager.
-void InstallInterfaces(SM::ServiceManager& service_manager);
+void InstallInterfaces(Core::System& system);
} // namespace Service::APM
diff --git a/src/core/hle/service/apm/controller.cpp b/src/core/hle/service/apm/controller.cpp
new file mode 100644
index 000000000..4376612eb
--- /dev/null
+++ b/src/core/hle/service/apm/controller.cpp
@@ -0,0 +1,68 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/logging/log.h"
+#include "core/core_timing.h"
+#include "core/hle/service/apm/controller.h"
+#include "core/settings.h"
+
+namespace Service::APM {
+
+constexpr PerformanceConfiguration DEFAULT_PERFORMANCE_CONFIGURATION =
+ PerformanceConfiguration::Config7;
+
+Controller::Controller(Core::Timing::CoreTiming& core_timing)
+ : core_timing(core_timing), configs{
+ {PerformanceMode::Handheld, DEFAULT_PERFORMANCE_CONFIGURATION},
+ {PerformanceMode::Docked, DEFAULT_PERFORMANCE_CONFIGURATION},
+ } {}
+
+Controller::~Controller() = default;
+
+void Controller::SetPerformanceConfiguration(PerformanceMode mode,
+ PerformanceConfiguration config) {
+ static const std::map<PerformanceConfiguration, u32> PCONFIG_TO_SPEED_MAP{
+ {PerformanceConfiguration::Config1, 1020}, {PerformanceConfiguration::Config2, 1020},
+ {PerformanceConfiguration::Config3, 1224}, {PerformanceConfiguration::Config4, 1020},
+ {PerformanceConfiguration::Config5, 1020}, {PerformanceConfiguration::Config6, 1224},
+ {PerformanceConfiguration::Config7, 1020}, {PerformanceConfiguration::Config8, 1020},
+ {PerformanceConfiguration::Config9, 1020}, {PerformanceConfiguration::Config10, 1020},
+ {PerformanceConfiguration::Config11, 1020}, {PerformanceConfiguration::Config12, 1020},
+ {PerformanceConfiguration::Config13, 1785}, {PerformanceConfiguration::Config14, 1785},
+ {PerformanceConfiguration::Config15, 1020}, {PerformanceConfiguration::Config16, 1020},
+ };
+
+ SetClockSpeed(PCONFIG_TO_SPEED_MAP.find(config)->second);
+ configs.insert_or_assign(mode, config);
+}
+
+void Controller::SetFromCpuBoostMode(CpuBoostMode mode) {
+ constexpr std::array<PerformanceConfiguration, 3> BOOST_MODE_TO_CONFIG_MAP{{
+ PerformanceConfiguration::Config7,
+ PerformanceConfiguration::Config13,
+ PerformanceConfiguration::Config15,
+ }};
+
+ SetPerformanceConfiguration(PerformanceMode::Docked,
+ BOOST_MODE_TO_CONFIG_MAP.at(static_cast<u32>(mode)));
+}
+
+PerformanceMode Controller::GetCurrentPerformanceMode() {
+ return Settings::values.use_docked_mode ? PerformanceMode::Docked : PerformanceMode::Handheld;
+}
+
+PerformanceConfiguration Controller::GetCurrentPerformanceConfiguration(PerformanceMode mode) {
+ if (configs.find(mode) == configs.end()) {
+ configs.insert_or_assign(mode, DEFAULT_PERFORMANCE_CONFIGURATION);
+ }
+
+ return configs[mode];
+}
+
+void Controller::SetClockSpeed(u32 mhz) {
+ LOG_INFO(Service_APM, "called, mhz={:08X}", mhz);
+ // TODO(DarkLordZach): Actually signal core_timing to change clock speed.
+}
+
+} // namespace Service::APM
diff --git a/src/core/hle/service/apm/controller.h b/src/core/hle/service/apm/controller.h
new file mode 100644
index 000000000..8ac80eaea
--- /dev/null
+++ b/src/core/hle/service/apm/controller.h
@@ -0,0 +1,70 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <map>
+#include "common/common_types.h"
+
+namespace Core::Timing {
+class CoreTiming;
+}
+
+namespace Service::APM {
+
+enum class PerformanceConfiguration : u32 {
+ Config1 = 0x00010000,
+ Config2 = 0x00010001,
+ Config3 = 0x00010002,
+ Config4 = 0x00020000,
+ Config5 = 0x00020001,
+ Config6 = 0x00020002,
+ Config7 = 0x00020003,
+ Config8 = 0x00020004,
+ Config9 = 0x00020005,
+ Config10 = 0x00020006,
+ Config11 = 0x92220007,
+ Config12 = 0x92220008,
+ Config13 = 0x92220009,
+ Config14 = 0x9222000A,
+ Config15 = 0x9222000B,
+ Config16 = 0x9222000C,
+};
+
+enum class CpuBoostMode : u32 {
+ Disabled = 0,
+ Full = 1, // CPU + GPU -> Config 13, 14, 15, or 16
+ Partial = 2, // GPU Only -> Config 15 or 16
+};
+
+enum class PerformanceMode : u8 {
+ Handheld = 0,
+ Docked = 1,
+};
+
+// Class to manage the state and change of the emulated system performance.
+// Specifically, this deals with PerformanceMode, which corresponds to the system being docked or
+// undocked, and PerformanceConfig which specifies the exact CPU, GPU, and Memory clocks to operate
+// at. Additionally, this manages 'Boost Mode', which allows games to temporarily overclock the
+// system during times of high load -- this simply maps to different PerformanceConfigs to use.
+class Controller {
+public:
+ Controller(Core::Timing::CoreTiming& core_timing);
+ ~Controller();
+
+ void SetPerformanceConfiguration(PerformanceMode mode, PerformanceConfiguration config);
+ void SetFromCpuBoostMode(CpuBoostMode mode);
+
+ PerformanceMode GetCurrentPerformanceMode();
+ PerformanceConfiguration GetCurrentPerformanceConfiguration(PerformanceMode mode);
+
+private:
+ void SetClockSpeed(u32 mhz);
+
+ std::map<PerformanceMode, PerformanceConfiguration> configs;
+
+ Core::Timing::CoreTiming& core_timing;
+};
+
+} // namespace Service::APM
diff --git a/src/core/hle/service/apm/interface.cpp b/src/core/hle/service/apm/interface.cpp
index d058c0245..06f0f8edd 100644
--- a/src/core/hle/service/apm/interface.cpp
+++ b/src/core/hle/service/apm/interface.cpp
@@ -5,43 +5,32 @@
#include "common/logging/log.h"
#include "core/hle/ipc_helpers.h"
#include "core/hle/service/apm/apm.h"
+#include "core/hle/service/apm/controller.h"
#include "core/hle/service/apm/interface.h"
namespace Service::APM {
class ISession final : public ServiceFramework<ISession> {
public:
- ISession() : ServiceFramework("ISession") {
+ ISession(Controller& controller) : ServiceFramework("ISession"), controller(controller) {
static const FunctionInfo functions[] = {
{0, &ISession::SetPerformanceConfiguration, "SetPerformanceConfiguration"},
{1, &ISession::GetPerformanceConfiguration, "GetPerformanceConfiguration"},
+ {2, nullptr, "SetCpuOverclockEnabled"},
};
RegisterHandlers(functions);
}
private:
- enum class PerformanceConfiguration : u32 {
- Config1 = 0x00010000,
- Config2 = 0x00010001,
- Config3 = 0x00010002,
- Config4 = 0x00020000,
- Config5 = 0x00020001,
- Config6 = 0x00020002,
- Config7 = 0x00020003,
- Config8 = 0x00020004,
- Config9 = 0x00020005,
- Config10 = 0x00020006,
- Config11 = 0x92220007,
- Config12 = 0x92220008,
- };
-
void SetPerformanceConfiguration(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
- auto mode = static_cast<PerformanceMode>(rp.Pop<u32>());
- u32 config = rp.Pop<u32>();
- LOG_WARNING(Service_APM, "(STUBBED) called mode={} config={}", static_cast<u32>(mode),
- config);
+ const auto mode = rp.PopEnum<PerformanceMode>();
+ const auto config = rp.PopEnum<PerformanceConfiguration>();
+ LOG_DEBUG(Service_APM, "called mode={} config={}", static_cast<u32>(mode),
+ static_cast<u32>(config));
+
+ controller.SetPerformanceConfiguration(mode, config);
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
@@ -50,20 +39,23 @@ private:
void GetPerformanceConfiguration(Kernel::HLERequestContext& ctx) {
IPC::RequestParser rp{ctx};
- auto mode = static_cast<PerformanceMode>(rp.Pop<u32>());
- LOG_WARNING(Service_APM, "(STUBBED) called mode={}", static_cast<u32>(mode));
+ const auto mode = rp.PopEnum<PerformanceMode>();
+ LOG_DEBUG(Service_APM, "called mode={}", static_cast<u32>(mode));
IPC::ResponseBuilder rb{ctx, 3};
rb.Push(RESULT_SUCCESS);
- rb.Push<u32>(static_cast<u32>(PerformanceConfiguration::Config1));
+ rb.PushEnum(controller.GetCurrentPerformanceConfiguration(mode));
}
+
+ Controller& controller;
};
-APM::APM(std::shared_ptr<Module> apm, const char* name)
- : ServiceFramework(name), apm(std::move(apm)) {
+APM::APM(std::shared_ptr<Module> apm, Controller& controller, const char* name)
+ : ServiceFramework(name), apm(std::move(apm)), controller(controller) {
static const FunctionInfo functions[] = {
{0, &APM::OpenSession, "OpenSession"},
- {1, nullptr, "GetPerformanceMode"},
+ {1, &APM::GetPerformanceMode, "GetPerformanceMode"},
+ {6, nullptr, "IsCpuOverclockEnabled"},
};
RegisterHandlers(functions);
}
@@ -75,10 +67,17 @@ void APM::OpenSession(Kernel::HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(RESULT_SUCCESS);
- rb.PushIpcInterface<ISession>();
+ rb.PushIpcInterface<ISession>(controller);
+}
+
+void APM::GetPerformanceMode(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_APM, "called");
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.PushEnum(controller.GetCurrentPerformanceMode());
}
-APM_Sys::APM_Sys() : ServiceFramework{"apm:sys"} {
+APM_Sys::APM_Sys(Controller& controller) : ServiceFramework{"apm:sys"}, controller(controller) {
// clang-format off
static const FunctionInfo functions[] = {
{0, nullptr, "RequestPerformanceMode"},
@@ -87,8 +86,8 @@ APM_Sys::APM_Sys() : ServiceFramework{"apm:sys"} {
{3, nullptr, "GetLastThrottlingState"},
{4, nullptr, "ClearLastThrottlingState"},
{5, nullptr, "LoadAndApplySettings"},
- {6, nullptr, "SetCpuBoostMode"},
- {7, nullptr, "GetCurrentPerformanceConfiguration"},
+ {6, &APM_Sys::SetCpuBoostMode, "SetCpuBoostMode"},
+ {7, &APM_Sys::GetCurrentPerformanceConfiguration, "GetCurrentPerformanceConfiguration"},
};
// clang-format on
@@ -102,7 +101,28 @@ void APM_Sys::GetPerformanceEvent(Kernel::HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(RESULT_SUCCESS);
- rb.PushIpcInterface<ISession>();
+ rb.PushIpcInterface<ISession>(controller);
+}
+
+void APM_Sys::SetCpuBoostMode(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ const auto mode = rp.PopEnum<CpuBoostMode>();
+
+ LOG_DEBUG(Service_APM, "called, mode={:08X}", static_cast<u32>(mode));
+
+ controller.SetFromCpuBoostMode(mode);
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(RESULT_SUCCESS);
+}
+
+void APM_Sys::GetCurrentPerformanceConfiguration(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_APM, "called");
+
+ IPC::ResponseBuilder rb{ctx, 3};
+ rb.Push(RESULT_SUCCESS);
+ rb.PushEnum(
+ controller.GetCurrentPerformanceConfiguration(controller.GetCurrentPerformanceMode()));
}
} // namespace Service::APM
diff --git a/src/core/hle/service/apm/interface.h b/src/core/hle/service/apm/interface.h
index 773541aa4..de1b89437 100644
--- a/src/core/hle/service/apm/interface.h
+++ b/src/core/hle/service/apm/interface.h
@@ -8,24 +8,34 @@
namespace Service::APM {
+class Controller;
+class Module;
+
class APM final : public ServiceFramework<APM> {
public:
- explicit APM(std::shared_ptr<Module> apm, const char* name);
+ explicit APM(std::shared_ptr<Module> apm, Controller& controller, const char* name);
~APM() override;
private:
void OpenSession(Kernel::HLERequestContext& ctx);
+ void GetPerformanceMode(Kernel::HLERequestContext& ctx);
std::shared_ptr<Module> apm;
+ Controller& controller;
};
class APM_Sys final : public ServiceFramework<APM_Sys> {
public:
- explicit APM_Sys();
+ explicit APM_Sys(Controller& controller);
~APM_Sys() override;
+ void SetCpuBoostMode(Kernel::HLERequestContext& ctx);
+
private:
void GetPerformanceEvent(Kernel::HLERequestContext& ctx);
+ void GetCurrentPerformanceConfiguration(Kernel::HLERequestContext& ctx);
+
+ Controller& controller;
};
} // namespace Service::APM
diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp
index 3711e1ea1..679299f68 100644
--- a/src/core/hle/service/audio/audren_u.cpp
+++ b/src/core/hle/service/audio/audren_u.cpp
@@ -25,7 +25,8 @@ namespace Service::Audio {
class IAudioRenderer final : public ServiceFramework<IAudioRenderer> {
public:
- explicit IAudioRenderer(AudioCore::AudioRendererParameter audren_params)
+ explicit IAudioRenderer(AudioCore::AudioRendererParameter audren_params,
+ const std::size_t instance_number)
: ServiceFramework("IAudioRenderer") {
// clang-format off
static const FunctionInfo functions[] = {
@@ -48,8 +49,8 @@ public:
auto& system = Core::System::GetInstance();
system_event = Kernel::WritableEvent::CreateEventPair(
system.Kernel(), Kernel::ResetType::Manual, "IAudioRenderer:SystemEvent");
- renderer = std::make_unique<AudioCore::AudioRenderer>(system.CoreTiming(), audren_params,
- system_event.writable);
+ renderer = std::make_unique<AudioCore::AudioRenderer>(
+ system.CoreTiming(), audren_params, system_event.writable, instance_number);
}
private:
@@ -607,7 +608,7 @@ void AudRenU::OpenAudioRendererImpl(Kernel::HLERequestContext& ctx) {
IPC::ResponseBuilder rb{ctx, 2, 0, 1};
rb.Push(RESULT_SUCCESS);
- rb.PushIpcInterface<IAudioRenderer>(params);
+ rb.PushIpcInterface<IAudioRenderer>(params, audren_instance_count++);
}
bool AudRenU::IsFeatureSupported(AudioFeatures feature, u32_le revision) const {
diff --git a/src/core/hle/service/audio/audren_u.h b/src/core/hle/service/audio/audren_u.h
index 1d3c8df61..49f2733cf 100644
--- a/src/core/hle/service/audio/audren_u.h
+++ b/src/core/hle/service/audio/audren_u.h
@@ -33,6 +33,7 @@ private:
};
bool IsFeatureSupported(AudioFeatures feature, u32_le revision) const;
+ std::size_t audren_instance_count = 0;
};
} // namespace Service::Audio
diff --git a/src/core/hle/service/filesystem/filesystem.cpp b/src/core/hle/service/filesystem/filesystem.cpp
index 1ebfeb4bf..8ce110dd1 100644
--- a/src/core/hle/service/filesystem/filesystem.cpp
+++ b/src/core/hle/service/filesystem/filesystem.cpp
@@ -472,12 +472,12 @@ void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite) {
}
}
-void InstallInterfaces(SM::ServiceManager& service_manager, FileSys::VfsFilesystem& vfs) {
+void InstallInterfaces(Core::System& system) {
romfs_factory = nullptr;
- CreateFactories(vfs, false);
- std::make_shared<FSP_LDR>()->InstallAsService(service_manager);
- std::make_shared<FSP_PR>()->InstallAsService(service_manager);
- std::make_shared<FSP_SRV>()->InstallAsService(service_manager);
+ CreateFactories(*system.GetFilesystem(), false);
+ std::make_shared<FSP_LDR>()->InstallAsService(system.ServiceManager());
+ std::make_shared<FSP_PR>()->InstallAsService(system.ServiceManager());
+ std::make_shared<FSP_SRV>(system.GetReporter())->InstallAsService(system.ServiceManager());
}
} // namespace Service::FileSystem
diff --git a/src/core/hle/service/filesystem/filesystem.h b/src/core/hle/service/filesystem/filesystem.h
index 6481f237c..3849dd89e 100644
--- a/src/core/hle/service/filesystem/filesystem.h
+++ b/src/core/hle/service/filesystem/filesystem.h
@@ -65,7 +65,7 @@ FileSys::VirtualDir GetModificationDumpRoot(u64 title_id);
// above is called.
void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite = true);
-void InstallInterfaces(SM::ServiceManager& service_manager, FileSys::VfsFilesystem& vfs);
+void InstallInterfaces(Core::System& system);
// A class that wraps a VfsDirectory with methods that return ResultVal and ResultCode instead of
// pointers and booleans. This makes using a VfsDirectory with switch services much easier and
diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp
index e7df8fd98..d3cd46a9b 100644
--- a/src/core/hle/service/filesystem/fsp_srv.cpp
+++ b/src/core/hle/service/filesystem/fsp_srv.cpp
@@ -26,6 +26,7 @@
#include "core/hle/kernel/process.h"
#include "core/hle/service/filesystem/filesystem.h"
#include "core/hle/service/filesystem/fsp_srv.h"
+#include "core/reporter.h"
namespace Service::FileSystem {
@@ -613,7 +614,7 @@ private:
u64 next_entry_index = 0;
};
-FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") {
+FSP_SRV::FSP_SRV(const Core::Reporter& reporter) : ServiceFramework("fsp-srv"), reporter(reporter) {
// clang-format off
static const FunctionInfo functions[] = {
{0, nullptr, "OpenFileSystem"},
@@ -710,14 +711,14 @@ FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") {
{1001, nullptr, "SetSaveDataSize"},
{1002, nullptr, "SetSaveDataRootPath"},
{1003, nullptr, "DisableAutoSaveDataCreation"},
- {1004, nullptr, "SetGlobalAccessLogMode"},
+ {1004, &FSP_SRV::SetGlobalAccessLogMode, "SetGlobalAccessLogMode"},
{1005, &FSP_SRV::GetGlobalAccessLogMode, "GetGlobalAccessLogMode"},
- {1006, nullptr, "OutputAccessLogToSdCard"},
+ {1006, &FSP_SRV::OutputAccessLogToSdCard, "OutputAccessLogToSdCard"},
{1007, nullptr, "RegisterUpdatePartition"},
{1008, nullptr, "OpenRegisteredUpdatePartition"},
{1009, nullptr, "GetAndClearMemoryReportInfo"},
{1010, nullptr, "SetDataStorageRedirectTarget"},
- {1011, nullptr, "OutputAccessLogToSdCard2"},
+ {1011, &FSP_SRV::GetAccessLogVersionInfo, "GetAccessLogVersionInfo"},
{1100, nullptr, "OverrideSaveDataTransferTokenSignVerificationKey"},
{1110, nullptr, "CorruptSaveDataFileSystemBySaveDataSpaceId2"},
{1200, nullptr, "OpenMultiCommitManager"},
@@ -814,21 +815,22 @@ void FSP_SRV::OpenSaveDataInfoReaderBySaveDataSpaceId(Kernel::HLERequestContext&
rb.PushIpcInterface<ISaveDataInfoReader>(std::make_shared<ISaveDataInfoReader>(space));
}
-void FSP_SRV::GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) {
- LOG_WARNING(Service_FS, "(STUBBED) called");
+void FSP_SRV::SetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ log_mode = rp.PopEnum<LogMode>();
- enum class LogMode : u32 {
- Off,
- Log,
- RedirectToSdCard,
- LogToSdCard = Log | RedirectToSdCard,
- };
+ LOG_DEBUG(Service_FS, "called, log_mode={:08X}", static_cast<u32>(log_mode));
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(RESULT_SUCCESS);
+}
+
+void FSP_SRV::GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_FS, "called");
- // Given we always want to receive logging information,
- // we always specify logging as enabled.
IPC::ResponseBuilder rb{ctx, 3};
rb.Push(RESULT_SUCCESS);
- rb.PushEnum(LogMode::Log);
+ rb.PushEnum(log_mode);
}
void FSP_SRV::OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx) {
@@ -902,4 +904,26 @@ void FSP_SRV::OpenPatchDataStorageByCurrentProcess(Kernel::HLERequestContext& ct
rb.Push(FileSys::ERROR_ENTITY_NOT_FOUND);
}
+void FSP_SRV::OutputAccessLogToSdCard(Kernel::HLERequestContext& ctx) {
+ const auto raw = ctx.ReadBuffer();
+ auto log = Common::StringFromFixedZeroTerminatedBuffer(
+ reinterpret_cast<const char*>(raw.data()), raw.size());
+
+ LOG_DEBUG(Service_FS, "called, log='{}'", log);
+
+ reporter.SaveFilesystemAccessReport(log_mode, std::move(log));
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(RESULT_SUCCESS);
+}
+
+void FSP_SRV::GetAccessLogVersionInfo(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_FS, "called");
+
+ IPC::ResponseBuilder rb{ctx, 4};
+ rb.Push(RESULT_SUCCESS);
+ rb.PushEnum(AccessLogVersion::Latest);
+ rb.Push(access_log_program_index);
+}
+
} // namespace Service::FileSystem
diff --git a/src/core/hle/service/filesystem/fsp_srv.h b/src/core/hle/service/filesystem/fsp_srv.h
index d7572ba7a..b5486a193 100644
--- a/src/core/hle/service/filesystem/fsp_srv.h
+++ b/src/core/hle/service/filesystem/fsp_srv.h
@@ -7,15 +7,32 @@
#include <memory>
#include "core/hle/service/service.h"
+namespace Core {
+class Reporter;
+}
+
namespace FileSys {
class FileSystemBackend;
}
namespace Service::FileSystem {
+enum class AccessLogVersion : u32 {
+ V7_0_0 = 2,
+
+ Latest = V7_0_0,
+};
+
+enum class LogMode : u32 {
+ Off,
+ Log,
+ RedirectToSdCard,
+ LogToSdCard = Log | RedirectToSdCard,
+};
+
class FSP_SRV final : public ServiceFramework<FSP_SRV> {
public:
- explicit FSP_SRV();
+ explicit FSP_SRV(const Core::Reporter& reporter);
~FSP_SRV() override;
private:
@@ -26,13 +43,20 @@ private:
void OpenSaveDataFileSystem(Kernel::HLERequestContext& ctx);
void OpenReadOnlySaveDataFileSystem(Kernel::HLERequestContext& ctx);
void OpenSaveDataInfoReaderBySaveDataSpaceId(Kernel::HLERequestContext& ctx);
+ void SetGlobalAccessLogMode(Kernel::HLERequestContext& ctx);
void GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx);
void OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx);
void OpenDataStorageByDataId(Kernel::HLERequestContext& ctx);
void OpenPatchDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx);
+ void OutputAccessLogToSdCard(Kernel::HLERequestContext& ctx);
+ void GetAccessLogVersionInfo(Kernel::HLERequestContext& ctx);
FileSys::VirtualFile romfs;
u64 current_process_id = 0;
+ u32 access_log_program_index = 0;
+ LogMode log_mode = LogMode::LogToSdCard;
+
+ const Core::Reporter& reporter;
};
} // namespace Service::FileSystem
diff --git a/src/core/hle/service/friend/friend.cpp b/src/core/hle/service/friend/friend.cpp
index dec541f2e..d1ec12ef9 100644
--- a/src/core/hle/service/friend/friend.cpp
+++ b/src/core/hle/service/friend/friend.cpp
@@ -22,7 +22,7 @@ public:
{0, nullptr, "GetCompletionEvent"},
{1, nullptr, "Cancel"},
{10100, nullptr, "GetFriendListIds"},
- {10101, nullptr, "GetFriendList"},
+ {10101, &IFriendService::GetFriendList, "GetFriendList"},
{10102, nullptr, "UpdateFriendInfo"},
{10110, nullptr, "GetFriendProfileImage"},
{10200, nullptr, "SendFriendRequestForApplication"},
@@ -99,6 +99,23 @@ public:
}
private:
+ enum class PresenceFilter : u32 {
+ None = 0,
+ Online = 1,
+ OnlinePlay = 2,
+ OnlineOrOnlinePlay = 3,
+ };
+
+ struct SizedFriendFilter {
+ PresenceFilter presence;
+ u8 is_favorite;
+ u8 same_app;
+ u8 same_app_played;
+ u8 arbitary_app_played;
+ u64 group_id;
+ };
+ static_assert(sizeof(SizedFriendFilter) == 0x10, "SizedFriendFilter is an invalid size");
+
void DeclareCloseOnlinePlaySession(Kernel::HLERequestContext& ctx) {
// Stub used by Splatoon 2
LOG_WARNING(Service_ACC, "(STUBBED) called");
@@ -112,6 +129,22 @@ private:
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
}
+
+ void GetFriendList(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ const auto friend_offset = rp.Pop<u32>();
+ const auto uuid = rp.PopRaw<Common::UUID>();
+ [[maybe_unused]] const auto filter = rp.PopRaw<SizedFriendFilter>();
+ const auto pid = rp.Pop<u64>();
+ LOG_WARNING(Service_ACC, "(STUBBED) called, offset={}, uuid={}, pid={}", friend_offset,
+ uuid.Format(), pid);
+
+ IPC::ResponseBuilder rb{ctx, 3};
+ rb.Push(RESULT_SUCCESS);
+
+ rb.Push<u32>(0); // Friend count
+ // TODO(ogniK): Return a buffer of u64s which are the "NetworkServiceAccountId"
+ }
};
class INotificationService final : public ServiceFramework<INotificationService> {
diff --git a/src/core/hle/service/hid/controllers/npad.cpp b/src/core/hle/service/hid/controllers/npad.cpp
index fdd6d79a2..1e81f776f 100644
--- a/src/core/hle/service/hid/controllers/npad.cpp
+++ b/src/core/hle/service/hid/controllers/npad.cpp
@@ -548,6 +548,37 @@ void Controller_NPad::DisconnectNPad(u32 npad_id) {
connected_controllers[NPadIdToIndex(npad_id)].is_connected = false;
}
+void Controller_NPad::StartLRAssignmentMode() {
+ // Nothing internally is used for lr assignment mode. Since we have the ability to set the
+ // controller types from boot, it doesn't really matter about showing a selection screen
+ is_in_lr_assignment_mode = true;
+}
+
+void Controller_NPad::StopLRAssignmentMode() {
+ is_in_lr_assignment_mode = false;
+}
+
+bool Controller_NPad::SwapNpadAssignment(u32 npad_id_1, u32 npad_id_2) {
+ if (npad_id_1 == NPAD_HANDHELD || npad_id_2 == NPAD_HANDHELD || npad_id_1 == NPAD_UNKNOWN ||
+ npad_id_2 == NPAD_UNKNOWN) {
+ return true;
+ }
+ const auto npad_index_1 = NPadIdToIndex(npad_id_1);
+ const auto npad_index_2 = NPadIdToIndex(npad_id_2);
+
+ if (!IsControllerSupported(connected_controllers[npad_index_1].type) ||
+ !IsControllerSupported(connected_controllers[npad_index_2].type)) {
+ return false;
+ }
+
+ std::swap(connected_controllers[npad_index_1].type, connected_controllers[npad_index_2].type);
+
+ InitNewlyAddedControler(npad_index_1);
+ InitNewlyAddedControler(npad_index_2);
+
+ return true;
+}
+
bool Controller_NPad::IsControllerSupported(NPadControllerType controller) {
if (controller == NPadControllerType::Handheld) {
// Handheld is not even a supported type, lets stop here
diff --git a/src/core/hle/service/hid/controllers/npad.h b/src/core/hle/service/hid/controllers/npad.h
index 4ff50b3cd..4b6c1083f 100644
--- a/src/core/hle/service/hid/controllers/npad.h
+++ b/src/core/hle/service/hid/controllers/npad.h
@@ -124,6 +124,10 @@ public:
void ConnectAllDisconnectedControllers();
void ClearAllControllers();
+ void StartLRAssignmentMode();
+ void StopLRAssignmentMode();
+ bool SwapNpadAssignment(u32 npad_id_1, u32 npad_id_2);
+
// Logical OR for all buttons presses on all controllers
// Specifically for cheat engine and other features.
u32 GetAndResetPressState();
@@ -321,5 +325,6 @@ private:
void RequestPadStateUpdate(u32 npad_id);
std::array<ControllerPad, 10> npad_pad_states{};
bool IsControllerSupported(NPadControllerType controller);
+ bool is_in_lr_assignment_mode{false};
};
} // namespace Service::HID
diff --git a/src/core/hle/service/hid/errors.h b/src/core/hle/service/hid/errors.h
new file mode 100644
index 000000000..3583642e7
--- /dev/null
+++ b/src/core/hle/service/hid/errors.h
@@ -0,0 +1,13 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/result.h"
+
+namespace Service::HID {
+
+constexpr ResultCode ERR_NPAD_NOT_CONNECTED{ErrorModule::HID, 710};
+
+} // namespace Service::HID
diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp
index a4ad95d96..0bd24b8eb 100644
--- a/src/core/hle/service/hid/hid.cpp
+++ b/src/core/hle/service/hid/hid.cpp
@@ -16,6 +16,7 @@
#include "core/hle/kernel/readable_event.h"
#include "core/hle/kernel/shared_memory.h"
#include "core/hle/kernel/writable_event.h"
+#include "core/hle/service/hid/errors.h"
#include "core/hle/service/hid/hid.h"
#include "core/hle/service/hid/irs.h"
#include "core/hle/service/hid/xcd.h"
@@ -202,11 +203,11 @@ Hid::Hid() : ServiceFramework("hid") {
{123, nullptr, "SetNpadJoyAssignmentModeSingleByDefault"},
{124, &Hid::SetNpadJoyAssignmentModeDual, "SetNpadJoyAssignmentModeDual"},
{125, &Hid::MergeSingleJoyAsDualJoy, "MergeSingleJoyAsDualJoy"},
- {126, nullptr, "StartLrAssignmentMode"},
- {127, nullptr, "StopLrAssignmentMode"},
+ {126, &Hid::StartLrAssignmentMode, "StartLrAssignmentMode"},
+ {127, &Hid::StopLrAssignmentMode, "StopLrAssignmentMode"},
{128, &Hid::SetNpadHandheldActivationMode, "SetNpadHandheldActivationMode"},
{129, nullptr, "GetNpadHandheldActivationMode"},
- {130, nullptr, "SwapNpadAssignment"},
+ {130, &Hid::SwapNpadAssignment, "SwapNpadAssignment"},
{131, nullptr, "IsUnintendedHomeButtonInputProtectionEnabled"},
{132, nullptr, "EnableUnintendedHomeButtonInputProtection"},
{133, nullptr, "SetNpadJoyAssignmentModeSingleWithDestination"},
@@ -733,6 +734,49 @@ void Hid::SetPalmaBoostMode(Kernel::HLERequestContext& ctx) {
rb.Push(RESULT_SUCCESS);
}
+void Hid::StartLrAssignmentMode(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ const auto applet_resource_user_id{rp.Pop<u64>()};
+
+ LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id);
+ auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+ controller.StartLRAssignmentMode();
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::StopLrAssignmentMode(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ const auto applet_resource_user_id{rp.Pop<u64>()};
+
+ LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id);
+ auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+ controller.StopLRAssignmentMode();
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::SwapNpadAssignment(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ const auto npad_1{rp.Pop<u32>()};
+ const auto npad_2{rp.Pop<u32>()};
+ const auto applet_resource_user_id{rp.Pop<u64>()};
+
+ LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, npad_1={}, npad_2={}",
+ applet_resource_user_id, npad_1, npad_2);
+
+ auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+ IPC::ResponseBuilder rb{ctx, 2};
+ if (controller.SwapNpadAssignment(npad_1, npad_2)) {
+ rb.Push(RESULT_SUCCESS);
+ } else {
+ LOG_ERROR(Service_HID, "Npads are not connected!");
+ rb.Push(ERR_NPAD_NOT_CONNECTED);
+ }
+}
+
class HidDbg final : public ServiceFramework<HidDbg> {
public:
explicit HidDbg() : ServiceFramework{"hid:dbg"} {
diff --git a/src/core/hle/service/hid/hid.h b/src/core/hle/service/hid/hid.h
index d3660cad2..28260ef1b 100644
--- a/src/core/hle/service/hid/hid.h
+++ b/src/core/hle/service/hid/hid.h
@@ -119,6 +119,9 @@ private:
void StopSixAxisSensor(Kernel::HLERequestContext& ctx);
void SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx);
void SetPalmaBoostMode(Kernel::HLERequestContext& ctx);
+ void StartLrAssignmentMode(Kernel::HLERequestContext& ctx);
+ void StopLrAssignmentMode(Kernel::HLERequestContext& ctx);
+ void SwapNpadAssignment(Kernel::HLERequestContext& ctx);
std::shared_ptr<IAppletResource> applet_resource;
};
diff --git a/src/core/hle/service/ldr/ldr.cpp b/src/core/hle/service/ldr/ldr.cpp
index b839303ac..8ddad8682 100644
--- a/src/core/hle/service/ldr/ldr.cpp
+++ b/src/core/hle/service/ldr/ldr.cpp
@@ -345,14 +345,16 @@ public:
vm_manager
.MirrorMemory(*map_address, nro_address, nro_size, Kernel::MemoryState::ModuleCode)
.IsSuccess());
- ASSERT(vm_manager.UnmapRange(nro_address, nro_size).IsSuccess());
+ ASSERT(vm_manager.ReprotectRange(nro_address, nro_size, Kernel::VMAPermission::None)
+ .IsSuccess());
if (bss_size > 0) {
ASSERT(vm_manager
.MirrorMemory(*map_address + nro_size, bss_address, bss_size,
Kernel::MemoryState::ModuleCode)
.IsSuccess());
- ASSERT(vm_manager.UnmapRange(bss_address, bss_size).IsSuccess());
+ ASSERT(vm_manager.ReprotectRange(bss_address, bss_size, Kernel::VMAPermission::None)
+ .IsSuccess());
}
vm_manager.ReprotectRange(*map_address, header.text_size,
@@ -364,7 +366,8 @@ public:
Core::System::GetInstance().InvalidateCpuInstructionCaches();
- nro.insert_or_assign(*map_address, NROInfo{hash, nro_size + bss_size});
+ nro.insert_or_assign(*map_address,
+ NROInfo{hash, nro_address, nro_size, bss_address, bss_size});
IPC::ResponseBuilder rb{ctx, 4};
rb.Push(RESULT_SUCCESS);
@@ -409,9 +412,23 @@ public:
}
auto& vm_manager = Core::CurrentProcess()->VMManager();
- const auto& nro_size = iter->second.size;
+ const auto& nro_info = iter->second;
- ASSERT(vm_manager.UnmapRange(nro_address, nro_size).IsSuccess());
+ // Unmap the mirrored memory
+ ASSERT(
+ vm_manager.UnmapRange(nro_address, nro_info.nro_size + nro_info.bss_size).IsSuccess());
+
+ // Reprotect the source memory
+ ASSERT(vm_manager
+ .ReprotectRange(nro_info.nro_address, nro_info.nro_size,
+ Kernel::VMAPermission::ReadWrite)
+ .IsSuccess());
+ if (nro_info.bss_size > 0) {
+ ASSERT(vm_manager
+ .ReprotectRange(nro_info.bss_address, nro_info.bss_size,
+ Kernel::VMAPermission::ReadWrite)
+ .IsSuccess());
+ }
Core::System::GetInstance().InvalidateCpuInstructionCaches();
@@ -473,7 +490,10 @@ private:
struct NROInfo {
SHA256Hash hash;
- u64 size;
+ VAddr nro_address;
+ u64 nro_size;
+ VAddr bss_address;
+ u64 bss_size;
};
bool initialized = false;
diff --git a/src/core/hle/service/mii/mii.cpp b/src/core/hle/service/mii/mii.cpp
index ce84e25ed..0b3923ad9 100644
--- a/src/core/hle/service/mii/mii.cpp
+++ b/src/core/hle/service/mii/mii.cpp
@@ -48,7 +48,7 @@ public:
{19, nullptr, "Export"},
{20, nullptr, "IsBrokenDatabaseWithClearFlag"},
{21, &IDatabaseService::GetIndex, "GetIndex"},
- {22, nullptr, "SetInterfaceVersion"},
+ {22, &IDatabaseService::SetInterfaceVersion, "SetInterfaceVersion"},
{23, nullptr, "Convert"},
};
// clang-format on
@@ -350,8 +350,22 @@ private:
rb.Push(index);
}
+ void SetInterfaceVersion(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ current_interface_version = rp.PopRaw<u32>();
+
+ LOG_DEBUG(Service_Mii, "called, interface_version={:08X}", current_interface_version);
+
+ UNIMPLEMENTED_IF(current_interface_version != 1);
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(RESULT_SUCCESS);
+ }
+
MiiManager db;
+ u32 current_interface_version = 0;
+
// Last read offsets of Get functions
std::array<u32, 4> offsets{};
};
diff --git a/src/core/hle/service/pm/pm.cpp b/src/core/hle/service/pm/pm.cpp
index ebcc41a43..fe6b5f798 100644
--- a/src/core/hle/service/pm/pm.cpp
+++ b/src/core/hle/service/pm/pm.cpp
@@ -3,11 +3,44 @@
// Refer to the license.txt file included.
#include "core/hle/ipc_helpers.h"
+#include "core/hle/kernel/kernel.h"
+#include "core/hle/kernel/process.h"
#include "core/hle/service/pm/pm.h"
#include "core/hle/service/service.h"
namespace Service::PM {
+namespace {
+
+constexpr ResultCode ERROR_PROCESS_NOT_FOUND{ErrorModule::PM, 1};
+
+constexpr u64 NO_PROCESS_FOUND_PID{0};
+
+std::optional<Kernel::SharedPtr<Kernel::Process>> SearchProcessList(
+ const std::vector<Kernel::SharedPtr<Kernel::Process>>& process_list,
+ std::function<bool(const Kernel::SharedPtr<Kernel::Process>&)> predicate) {
+ const auto iter = std::find_if(process_list.begin(), process_list.end(), predicate);
+
+ if (iter == process_list.end()) {
+ return std::nullopt;
+ }
+
+ return *iter;
+}
+
+void GetApplicationPidGeneric(Kernel::HLERequestContext& ctx,
+ const std::vector<Kernel::SharedPtr<Kernel::Process>>& process_list) {
+ const auto process = SearchProcessList(process_list, [](const auto& process) {
+ return process->GetProcessID() == Kernel::Process::ProcessIDMin;
+ });
+
+ IPC::ResponseBuilder rb{ctx, 4};
+ rb.Push(RESULT_SUCCESS);
+ rb.Push(process.has_value() ? (*process)->GetProcessID() : NO_PROCESS_FOUND_PID);
+}
+
+} // Anonymous namespace
+
class BootMode final : public ServiceFramework<BootMode> {
public:
explicit BootMode() : ServiceFramework{"pm:bm"} {
@@ -41,14 +74,15 @@ private:
class DebugMonitor final : public ServiceFramework<DebugMonitor> {
public:
- explicit DebugMonitor() : ServiceFramework{"pm:dmnt"} {
+ explicit DebugMonitor(const Kernel::KernelCore& kernel)
+ : ServiceFramework{"pm:dmnt"}, kernel(kernel) {
// clang-format off
static const FunctionInfo functions[] = {
{0, nullptr, "GetDebugProcesses"},
{1, nullptr, "StartDebugProcess"},
- {2, nullptr, "GetTitlePid"},
+ {2, &DebugMonitor::GetTitlePid, "GetTitlePid"},
{3, nullptr, "EnableDebugForTitleId"},
- {4, nullptr, "GetApplicationPid"},
+ {4, &DebugMonitor::GetApplicationPid, "GetApplicationPid"},
{5, nullptr, "EnableDebugForApplication"},
{6, nullptr, "DisableDebug"},
};
@@ -56,21 +90,77 @@ public:
RegisterHandlers(functions);
}
+
+private:
+ void GetTitlePid(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ const auto title_id = rp.PopRaw<u64>();
+
+ LOG_DEBUG(Service_PM, "called, title_id={:016X}", title_id);
+
+ const auto process =
+ SearchProcessList(kernel.GetProcessList(), [title_id](const auto& process) {
+ return process->GetTitleID() == title_id;
+ });
+
+ if (!process.has_value()) {
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(ERROR_PROCESS_NOT_FOUND);
+ return;
+ }
+
+ IPC::ResponseBuilder rb{ctx, 4};
+ rb.Push(RESULT_SUCCESS);
+ rb.Push((*process)->GetProcessID());
+ }
+
+ void GetApplicationPid(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_PM, "called");
+ GetApplicationPidGeneric(ctx, kernel.GetProcessList());
+ }
+
+ const Kernel::KernelCore& kernel;
};
class Info final : public ServiceFramework<Info> {
public:
- explicit Info() : ServiceFramework{"pm:info"} {
+ explicit Info(const std::vector<Kernel::SharedPtr<Kernel::Process>>& process_list)
+ : ServiceFramework{"pm:info"}, process_list(process_list) {
static const FunctionInfo functions[] = {
- {0, nullptr, "GetTitleId"},
+ {0, &Info::GetTitleId, "GetTitleId"},
};
RegisterHandlers(functions);
}
+
+private:
+ void GetTitleId(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ const auto process_id = rp.PopRaw<u64>();
+
+ LOG_DEBUG(Service_PM, "called, process_id={:016X}", process_id);
+
+ const auto process = SearchProcessList(process_list, [process_id](const auto& process) {
+ return process->GetProcessID() == process_id;
+ });
+
+ if (!process.has_value()) {
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(ERROR_PROCESS_NOT_FOUND);
+ return;
+ }
+
+ IPC::ResponseBuilder rb{ctx, 4};
+ rb.Push(RESULT_SUCCESS);
+ rb.Push((*process)->GetTitleID());
+ }
+
+ const std::vector<Kernel::SharedPtr<Kernel::Process>>& process_list;
};
class Shell final : public ServiceFramework<Shell> {
public:
- explicit Shell() : ServiceFramework{"pm:shell"} {
+ explicit Shell(const Kernel::KernelCore& kernel)
+ : ServiceFramework{"pm:shell"}, kernel(kernel) {
// clang-format off
static const FunctionInfo functions[] = {
{0, nullptr, "LaunchProcess"},
@@ -79,21 +169,31 @@ public:
{3, nullptr, "GetProcessEventWaiter"},
{4, nullptr, "GetProcessEventType"},
{5, nullptr, "NotifyBootFinished"},
- {6, nullptr, "GetApplicationPid"},
+ {6, &Shell::GetApplicationPid, "GetApplicationPid"},
{7, nullptr, "BoostSystemMemoryResourceLimit"},
{8, nullptr, "EnableAdditionalSystemThreads"},
+ {9, nullptr, "GetUnimplementedEventHandle"},
};
// clang-format on
RegisterHandlers(functions);
}
+
+private:
+ void GetApplicationPid(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_PM, "called");
+ GetApplicationPidGeneric(ctx, kernel.GetProcessList());
+ }
+
+ const Kernel::KernelCore& kernel;
};
-void InstallInterfaces(SM::ServiceManager& sm) {
- std::make_shared<BootMode>()->InstallAsService(sm);
- std::make_shared<DebugMonitor>()->InstallAsService(sm);
- std::make_shared<Info>()->InstallAsService(sm);
- std::make_shared<Shell>()->InstallAsService(sm);
+void InstallInterfaces(Core::System& system) {
+ std::make_shared<BootMode>()->InstallAsService(system.ServiceManager());
+ std::make_shared<DebugMonitor>(system.Kernel())->InstallAsService(system.ServiceManager());
+ std::make_shared<Info>(system.Kernel().GetProcessList())
+ ->InstallAsService(system.ServiceManager());
+ std::make_shared<Shell>(system.Kernel())->InstallAsService(system.ServiceManager());
}
} // namespace Service::PM
diff --git a/src/core/hle/service/pm/pm.h b/src/core/hle/service/pm/pm.h
index cc8d3f215..852e7050c 100644
--- a/src/core/hle/service/pm/pm.h
+++ b/src/core/hle/service/pm/pm.h
@@ -4,8 +4,8 @@
#pragma once
-namespace Service::SM {
-class ServiceManager;
+namespace Core {
+class System;
}
namespace Service::PM {
@@ -16,6 +16,6 @@ enum class SystemBootMode {
};
/// Registers all PM services with the specified service manager.
-void InstallInterfaces(SM::ServiceManager& service_manager);
+void InstallInterfaces(Core::System& system);
} // namespace Service::PM
diff --git a/src/core/hle/service/service.cpp b/src/core/hle/service/service.cpp
index 5fc7d3cab..7eefd733f 100644
--- a/src/core/hle/service/service.cpp
+++ b/src/core/hle/service/service.cpp
@@ -195,8 +195,7 @@ ResultCode ServiceFrameworkBase::HandleSyncRequest(Kernel::HLERequestContext& co
// Module interface
/// Initialize ServiceManager
-void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system,
- FileSys::VfsFilesystem& vfs) {
+void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system) {
// NVFlinger needs to be accessed by several services like Vi and AppletOE so we instantiate it
// here and pass it into the respective InstallInterfaces functions.
auto nv_flinger = std::make_shared<NVFlinger::NVFlinger>(system.CoreTiming());
@@ -206,7 +205,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system,
Account::InstallInterfaces(system);
AM::InstallInterfaces(*sm, nv_flinger, system);
AOC::InstallInterfaces(*sm);
- APM::InstallInterfaces(*sm);
+ APM::InstallInterfaces(system);
Audio::InstallInterfaces(*sm);
BCAT::InstallInterfaces(*sm);
BPC::InstallInterfaces(*sm);
@@ -218,7 +217,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system,
EUPLD::InstallInterfaces(*sm);
Fatal::InstallInterfaces(*sm);
FGM::InstallInterfaces(*sm);
- FileSystem::InstallInterfaces(*sm, vfs);
+ FileSystem::InstallInterfaces(system);
Friend::InstallInterfaces(*sm);
Glue::InstallInterfaces(system);
GRC::InstallInterfaces(*sm);
@@ -242,7 +241,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system,
PCTL::InstallInterfaces(*sm);
PCV::InstallInterfaces(*sm);
PlayReport::InstallInterfaces(*sm);
- PM::InstallInterfaces(*sm);
+ PM::InstallInterfaces(system);
PSC::InstallInterfaces(*sm);
PSM::InstallInterfaces(*sm);
Set::InstallInterfaces(*sm);
diff --git a/src/core/hle/service/service.h b/src/core/hle/service/service.h
index abbfe5524..c6c4bdae5 100644
--- a/src/core/hle/service/service.h
+++ b/src/core/hle/service/service.h
@@ -182,8 +182,7 @@ private:
};
/// Initialize ServiceManager
-void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system,
- FileSys::VfsFilesystem& vfs);
+void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system);
/// Shutdown ServiceManager
void Shutdown();
diff --git a/src/core/reporter.cpp b/src/core/reporter.cpp
index 6ea26fda7..5d4c3e6ea 100644
--- a/src/core/reporter.cpp
+++ b/src/core/reporter.cpp
@@ -350,6 +350,24 @@ void Reporter::SaveErrorReport(u64 title_id, ResultCode result,
SaveToFile(std::move(out), GetPath("error_report", title_id, timestamp));
}
+void Reporter::SaveFilesystemAccessReport(Service::FileSystem::LogMode log_mode,
+ std::string log_message) const {
+ if (!IsReportingEnabled())
+ return;
+
+ const auto timestamp = GetTimestamp();
+ const auto title_id = system.CurrentProcess()->GetTitleID();
+ json out;
+
+ out["yuzu_version"] = GetYuzuVersionData();
+ out["report_common"] = GetReportCommonData(title_id, RESULT_SUCCESS, timestamp);
+
+ out["log_mode"] = fmt::format("{:08X}", static_cast<u32>(log_mode));
+ out["log_message"] = std::move(log_message);
+
+ SaveToFile(std::move(out), GetPath("filesystem_access_report", title_id, timestamp));
+}
+
void Reporter::SaveUserReport() const {
if (!IsReportingEnabled()) {
return;
diff --git a/src/core/reporter.h b/src/core/reporter.h
index 4266ca550..44256de50 100644
--- a/src/core/reporter.h
+++ b/src/core/reporter.h
@@ -16,6 +16,10 @@ namespace Kernel {
class HLERequestContext;
} // namespace Kernel
+namespace Service::FileSystem {
+enum class LogMode : u32;
+}
+
namespace Core {
class System;
@@ -49,6 +53,9 @@ public:
std::optional<std::string> custom_text_main = {},
std::optional<std::string> custom_text_detail = {}) const;
+ void SaveFilesystemAccessReport(Service::FileSystem::LogMode log_mode,
+ std::string log_message) const;
+
void SaveUserReport() const;
private:
diff --git a/src/core/settings.cpp b/src/core/settings.cpp
index 63aa59690..0dd1632ac 100644
--- a/src/core/settings.cpp
+++ b/src/core/settings.cpp
@@ -85,7 +85,6 @@ void LogSettings() {
LogSetting("System_RngSeed", Settings::values.rng_seed.value_or(0));
LogSetting("System_CurrentUser", Settings::values.current_user);
LogSetting("System_LanguageIndex", Settings::values.language_index);
- LogSetting("Core_CpuJitEnabled", Settings::values.cpu_jit_enabled);
LogSetting("Core_UseMultiCore", Settings::values.use_multi_core);
LogSetting("Renderer_UseResolutionFactor", Settings::values.resolution_factor);
LogSetting("Renderer_UseFrameLimit", Settings::values.use_frame_limit);
diff --git a/src/core/settings.h b/src/core/settings.h
index acf18d653..6638ce8f9 100644
--- a/src/core/settings.h
+++ b/src/core/settings.h
@@ -378,7 +378,6 @@ struct Values {
std::atomic_bool is_device_reload_pending{true};
// Core
- bool cpu_jit_enabled;
bool use_multi_core;
// Data Storage
diff --git a/src/core/telemetry_session.cpp b/src/core/telemetry_session.cpp
index 98f49042a..793d102d3 100644
--- a/src/core/telemetry_session.cpp
+++ b/src/core/telemetry_session.cpp
@@ -168,7 +168,6 @@ void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader) {
AddField(Telemetry::FieldType::UserConfig, "Audio_SinkId", Settings::values.sink_id);
AddField(Telemetry::FieldType::UserConfig, "Audio_EnableAudioStretching",
Settings::values.enable_audio_stretching);
- AddField(Telemetry::FieldType::UserConfig, "Core_UseCpuJit", Settings::values.cpu_jit_enabled);
AddField(Telemetry::FieldType::UserConfig, "Core_UseMultiCore",
Settings::values.use_multi_core);
AddField(Telemetry::FieldType::UserConfig, "Renderer_ResolutionFactor",
diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt
index 6839abe71..cd32c65d3 100644
--- a/src/video_core/CMakeLists.txt
+++ b/src/video_core/CMakeLists.txt
@@ -103,6 +103,8 @@ add_library(video_core STATIC
shader/decode/video.cpp
shader/decode/xmad.cpp
shader/decode/other.cpp
+ shader/control_flow.cpp
+ shader/control_flow.h
shader/decode.cpp
shader/node_helper.cpp
shader/node_helper.h
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index 404d4f5aa..4aaad2979 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -1278,6 +1278,7 @@ union Instruction {
union {
BitField<49, 1, u64> nodep_flag;
BitField<53, 4, u64> texture_info;
+ BitField<59, 1, u64> fp32_flag;
TextureType GetTextureType() const {
// The TLDS instruction has a weird encoding for the texture type.
@@ -1368,6 +1369,20 @@ union Instruction {
} bra;
union {
+ BitField<20, 24, u64> target;
+ BitField<5, 1, u64> constant_buffer;
+
+ s32 GetBranchExtend() const {
+ // Sign extend the branch target offset
+ u32 mask = 1U << (24 - 1);
+ u32 value = static_cast<u32>(target);
+ // The branch offset is relative to the next instruction and is stored in bytes, so
+ // divide it by the size of an instruction and add 1 to it.
+ return static_cast<s32>((value ^ mask) - mask) / sizeof(Instruction) + 1;
+ }
+ } brx;
+
+ union {
BitField<39, 1, u64> emit; // EmitVertex
BitField<40, 1, u64> cut; // EndPrimitive
} out;
@@ -1464,6 +1479,7 @@ public:
BFE_IMM,
BFI_IMM_R,
BRA,
+ BRX,
PBK,
LD_A,
LD_L,
@@ -1738,6 +1754,7 @@ private:
INST("111000101001----", Id::SSY, Type::Flow, "SSY"),
INST("111000101010----", Id::PBK, Type::Flow, "PBK"),
INST("111000100100----", Id::BRA, Type::Flow, "BRA"),
+ INST("111000100101----", Id::BRX, Type::Flow, "BRX"),
INST("1111000011111---", Id::SYNC, Type::Flow, "SYNC"),
INST("111000110100---", Id::BRK, Type::Flow, "BRK"),
INST("111000110000----", Id::EXIT, Type::Flow, "EXIT"),
@@ -1760,7 +1777,7 @@ private:
INST("1101111101010---", Id::TXQ_B, Type::Texture, "TXQ_B"),
INST("1101-00---------", Id::TEXS, Type::Texture, "TEXS"),
INST("11011100--11----", Id::TLD, Type::Texture, "TLD"),
- INST("1101101---------", Id::TLDS, Type::Texture, "TLDS"),
+ INST("1101-01---------", Id::TLDS, Type::Texture, "TLDS"),
INST("110010----111---", Id::TLD4, Type::Texture, "TLD4"),
INST("1101111100------", Id::TLD4S, Type::Texture, "TLD4S"),
INST("110111110110----", Id::TMML_B, Type::Texture, "TMML_B"),
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h
index bf67e3a70..d238c1257 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer.h
@@ -98,17 +98,19 @@ private:
/**
* Configures the color and depth framebuffer states.
- * @param must_reconfigure If true, tells the framebuffer to skip the cache and reconfigure
- * again. Used by the texture cache to solve texception conflicts
- * @param use_color_fb If true, configure color framebuffers.
- * @param using_depth_fb If true, configure the depth/stencil framebuffer.
- * @param preserve_contents If true, tries to preserve data from a previously used framebuffer.
+ *
+ * @param current_state The current OpenGL state.
+ * @param using_color_fb If true, configure color framebuffers.
+ * @param using_depth_fb If true, configure the depth/stencil framebuffer.
+ * @param preserve_contents If true, tries to preserve data from a previously used
+ * framebuffer.
* @param single_color_target Specifies if a single color buffer target should be used.
+ *
* @returns If depth (first) or stencil (second) are being stored in the bound zeta texture
- * (requires using_depth_fb to be true)
+ * (requires using_depth_fb to be true)
*/
std::pair<bool, bool> ConfigureFramebuffers(
- OpenGLState& current_state, bool use_color_fb = true, bool using_depth_fb = true,
+ OpenGLState& current_state, bool using_color_fb = true, bool using_depth_fb = true,
bool preserve_contents = true, std::optional<std::size_t> single_color_target = {});
/// Configures the current constbuffers to use for the draw command.
diff --git a/src/video_core/renderer_opengl/gl_sampler_cache.h b/src/video_core/renderer_opengl/gl_sampler_cache.h
index defbc2d81..34ee37f00 100644
--- a/src/video_core/renderer_opengl/gl_sampler_cache.h
+++ b/src/video_core/renderer_opengl/gl_sampler_cache.h
@@ -17,9 +17,9 @@ public:
~SamplerCacheOpenGL();
protected:
- OGLSampler CreateSampler(const Tegra::Texture::TSCEntry& tsc) const;
+ OGLSampler CreateSampler(const Tegra::Texture::TSCEntry& tsc) const override;
- GLuint ToSamplerType(const OGLSampler& sampler) const;
+ GLuint ToSamplerType(const OGLSampler& sampler) const override;
};
} // namespace OpenGL
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index f9b2b03a0..5d76ee12d 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -129,9 +129,11 @@ std::size_t CalculateProgramSize(const GLShader::ProgramCode& program) {
/// Hashes one (or two) program streams
u64 GetUniqueIdentifier(Maxwell::ShaderProgram program_type, const ProgramCode& code,
- const ProgramCode& code_b) {
- u64 unique_identifier =
- Common::CityHash64(reinterpret_cast<const char*>(code.data()), CalculateProgramSize(code));
+ const ProgramCode& code_b, std::size_t size_a = 0, std::size_t size_b = 0) {
+ if (size_a == 0) {
+ size_a = CalculateProgramSize(code);
+ }
+ u64 unique_identifier = Common::CityHash64(reinterpret_cast<const char*>(code.data()), size_a);
if (program_type != Maxwell::ShaderProgram::VertexA) {
return unique_identifier;
}
@@ -140,8 +142,11 @@ u64 GetUniqueIdentifier(Maxwell::ShaderProgram program_type, const ProgramCode&
std::size_t seed = 0;
boost::hash_combine(seed, unique_identifier);
- const u64 identifier_b = Common::CityHash64(reinterpret_cast<const char*>(code_b.data()),
- CalculateProgramSize(code_b));
+ if (size_b == 0) {
+ size_b = CalculateProgramSize(code_b);
+ }
+ const u64 identifier_b =
+ Common::CityHash64(reinterpret_cast<const char*>(code_b.data()), size_b);
boost::hash_combine(seed, identifier_b);
return static_cast<u64>(seed);
}
@@ -150,14 +155,17 @@ u64 GetUniqueIdentifier(Maxwell::ShaderProgram program_type, const ProgramCode&
GLShader::ProgramResult CreateProgram(const Device& device, Maxwell::ShaderProgram program_type,
ProgramCode program_code, ProgramCode program_code_b) {
GLShader::ShaderSetup setup(program_code);
+ setup.program.size_a = CalculateProgramSize(program_code);
+ setup.program.size_b = 0;
if (program_type == Maxwell::ShaderProgram::VertexA) {
// VertexB is always enabled, so when VertexA is enabled, we have two vertex shaders.
// Conventional HW does not support this, so we combine VertexA and VertexB into one
// stage here.
setup.SetProgramB(program_code_b);
+ setup.program.size_b = CalculateProgramSize(program_code_b);
}
- setup.program.unique_identifier =
- GetUniqueIdentifier(program_type, program_code, program_code_b);
+ setup.program.unique_identifier = GetUniqueIdentifier(
+ program_type, program_code, program_code_b, setup.program.size_a, setup.program.size_b);
switch (program_type) {
case Maxwell::ShaderProgram::VertexA:
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
index 5f2f1510c..bfc975a04 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
@@ -191,10 +191,12 @@ public:
// TODO(Subv): Figure out the actual depth of the flow stack, for now it seems
// unlikely that shaders will use 20 nested SSYs and PBKs.
- constexpr u32 FLOW_STACK_SIZE = 20;
- for (const auto stack : std::array{MetaStackClass::Ssy, MetaStackClass::Pbk}) {
- code.AddLine("uint {}[{}];", FlowStackName(stack), FLOW_STACK_SIZE);
- code.AddLine("uint {} = 0u;", FlowStackTopName(stack));
+ if (!ir.IsFlowStackDisabled()) {
+ constexpr u32 FLOW_STACK_SIZE = 20;
+ for (const auto stack : std::array{MetaStackClass::Ssy, MetaStackClass::Pbk}) {
+ code.AddLine("uint {}[{}];", FlowStackName(stack), FLOW_STACK_SIZE);
+ code.AddLine("uint {} = 0u;", FlowStackTopName(stack));
+ }
}
code.AddLine("while (true) {{");
@@ -1555,6 +1557,14 @@ private:
return {};
}
+ std::string BranchIndirect(Operation operation) {
+ const std::string op_a = VisitOperand(operation, 0, Type::Uint);
+
+ code.AddLine("jmp_to = {};", op_a);
+ code.AddLine("break;");
+ return {};
+ }
+
std::string PushFlowStack(Operation operation) {
const auto stack = std::get<MetaStackClass>(operation.GetMeta());
const auto target = std::get_if<ImmediateNode>(&*operation[0]);
@@ -1789,6 +1799,7 @@ private:
&GLSLDecompiler::ImageStore,
&GLSLDecompiler::Branch,
+ &GLSLDecompiler::BranchIndirect,
&GLSLDecompiler::PushFlowStack,
&GLSLDecompiler::PopFlowStack,
&GLSLDecompiler::Exit,
diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp
index 9148629ec..f9ee8429e 100644
--- a/src/video_core/renderer_opengl/gl_shader_gen.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp
@@ -29,14 +29,14 @@ layout (std140, binding = EMULATION_UBO_BINDING) uniform vs_config {
};
)";
- const ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET);
+ const ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET, setup.program.size_a);
ProgramResult program =
Decompile(device, program_ir, Maxwell3D::Regs::ShaderStage::Vertex, "vertex");
out += program.first;
if (setup.IsDualProgram()) {
- const ShaderIR program_ir_b(setup.program.code_b, PROGRAM_OFFSET);
+ const ShaderIR program_ir_b(setup.program.code_b, PROGRAM_OFFSET, setup.program.size_b);
ProgramResult program_b =
Decompile(device, program_ir_b, Maxwell3D::Regs::ShaderStage::Vertex, "vertex_b");
@@ -80,7 +80,7 @@ layout (std140, binding = EMULATION_UBO_BINDING) uniform gs_config {
};
)";
- const ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET);
+ const ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET, setup.program.size_a);
ProgramResult program =
Decompile(device, program_ir, Maxwell3D::Regs::ShaderStage::Geometry, "geometry");
out += program.first;
@@ -115,7 +115,7 @@ layout (std140, binding = EMULATION_UBO_BINDING) uniform fs_config {
};
)";
- const ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET);
+ const ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET, setup.program.size_a);
ProgramResult program =
Decompile(device, program_ir, Maxwell3D::Regs::ShaderStage::Fragment, "fragment");
diff --git a/src/video_core/renderer_opengl/gl_shader_gen.h b/src/video_core/renderer_opengl/gl_shader_gen.h
index 0536c8a03..7cbc590f8 100644
--- a/src/video_core/renderer_opengl/gl_shader_gen.h
+++ b/src/video_core/renderer_opengl/gl_shader_gen.h
@@ -27,6 +27,8 @@ struct ShaderSetup {
ProgramCode code;
ProgramCode code_b; // Used for dual vertex shaders
u64 unique_identifier;
+ std::size_t size_a;
+ std::size_t size_b;
} program;
/// Used in scenarios where we have a dual vertex shaders
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
index 02a9f5ecb..d2e9f4031 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
@@ -109,8 +109,8 @@ void VKBufferCache::Reserve(std::size_t max_size) {
}
}
-VKExecutionContext VKBufferCache::Send(VKExecutionContext exctx) {
- return stream_buffer->Send(exctx, buffer_offset - buffer_offset_base);
+void VKBufferCache::Send() {
+ stream_buffer->Send(buffer_offset - buffer_offset_base);
}
void VKBufferCache::AlignBuffer(std::size_t alignment) {
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h
index 3edf460df..49f13bcdc 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.h
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h
@@ -77,7 +77,7 @@ public:
void Reserve(std::size_t max_size);
/// Ensures that the set data is sent to the device.
- [[nodiscard]] VKExecutionContext Send(VKExecutionContext exctx);
+ void Send();
/// Returns the buffer cache handle.
vk::Buffer GetBuffer() const {
diff --git a/src/video_core/renderer_vulkan/vk_sampler_cache.h b/src/video_core/renderer_vulkan/vk_sampler_cache.h
index 771b05c73..1f73b716b 100644
--- a/src/video_core/renderer_vulkan/vk_sampler_cache.h
+++ b/src/video_core/renderer_vulkan/vk_sampler_cache.h
@@ -4,9 +4,6 @@
#pragma once
-#include <unordered_map>
-
-#include "common/common_types.h"
#include "video_core/renderer_vulkan/declarations.h"
#include "video_core/sampler_cache.h"
#include "video_core/textures/texture.h"
@@ -21,9 +18,9 @@ public:
~VKSamplerCache();
protected:
- UniqueSampler CreateSampler(const Tegra::Texture::TSCEntry& tsc) const;
+ UniqueSampler CreateSampler(const Tegra::Texture::TSCEntry& tsc) const override;
- vk::Sampler ToSamplerType(const UniqueSampler& sampler) const;
+ vk::Sampler ToSamplerType(const UniqueSampler& sampler) const override;
private:
const VKDevice& device;
diff --git a/src/video_core/renderer_vulkan/vk_scheduler.cpp b/src/video_core/renderer_vulkan/vk_scheduler.cpp
index f1fea1871..0f8116458 100644
--- a/src/video_core/renderer_vulkan/vk_scheduler.cpp
+++ b/src/video_core/renderer_vulkan/vk_scheduler.cpp
@@ -19,23 +19,19 @@ VKScheduler::VKScheduler(const VKDevice& device, VKResourceManager& resource_man
VKScheduler::~VKScheduler() = default;
-VKExecutionContext VKScheduler::GetExecutionContext() const {
- return VKExecutionContext(current_fence, current_cmdbuf);
-}
-
-VKExecutionContext VKScheduler::Flush(vk::Semaphore semaphore) {
+void VKScheduler::Flush(bool release_fence, vk::Semaphore semaphore) {
SubmitExecution(semaphore);
- current_fence->Release();
+ if (release_fence)
+ current_fence->Release();
AllocateNewContext();
- return GetExecutionContext();
}
-VKExecutionContext VKScheduler::Finish(vk::Semaphore semaphore) {
+void VKScheduler::Finish(bool release_fence, vk::Semaphore semaphore) {
SubmitExecution(semaphore);
current_fence->Wait();
- current_fence->Release();
+ if (release_fence)
+ current_fence->Release();
AllocateNewContext();
- return GetExecutionContext();
}
void VKScheduler::SubmitExecution(vk::Semaphore semaphore) {
diff --git a/src/video_core/renderer_vulkan/vk_scheduler.h b/src/video_core/renderer_vulkan/vk_scheduler.h
index cfaf5376f..0e5b49c7f 100644
--- a/src/video_core/renderer_vulkan/vk_scheduler.h
+++ b/src/video_core/renderer_vulkan/vk_scheduler.h
@@ -10,10 +10,43 @@
namespace Vulkan {
class VKDevice;
-class VKExecutionContext;
class VKFence;
class VKResourceManager;
+class VKFenceView {
+public:
+ VKFenceView() = default;
+ VKFenceView(VKFence* const& fence) : fence{fence} {}
+
+ VKFence* operator->() const noexcept {
+ return fence;
+ }
+
+ operator VKFence&() const noexcept {
+ return *fence;
+ }
+
+private:
+ VKFence* const& fence;
+};
+
+class VKCommandBufferView {
+public:
+ VKCommandBufferView() = default;
+ VKCommandBufferView(const vk::CommandBuffer& cmdbuf) : cmdbuf{cmdbuf} {}
+
+ const vk::CommandBuffer* operator->() const noexcept {
+ return &cmdbuf;
+ }
+
+ operator vk::CommandBuffer() const noexcept {
+ return cmdbuf;
+ }
+
+private:
+ const vk::CommandBuffer& cmdbuf;
+};
+
/// The scheduler abstracts command buffer and fence management with an interface that's able to do
/// OpenGL-like operations on Vulkan command buffers.
class VKScheduler {
@@ -21,16 +54,21 @@ public:
explicit VKScheduler(const VKDevice& device, VKResourceManager& resource_manager);
~VKScheduler();
- /// Gets the current execution context.
- [[nodiscard]] VKExecutionContext GetExecutionContext() const;
+ /// Gets a reference to the current fence.
+ VKFenceView GetFence() const {
+ return current_fence;
+ }
+
+ /// Gets a reference to the current command buffer.
+ VKCommandBufferView GetCommandBuffer() const {
+ return current_cmdbuf;
+ }
- /// Sends the current execution context to the GPU. It invalidates the current execution context
- /// and returns a new one.
- VKExecutionContext Flush(vk::Semaphore semaphore = nullptr);
+ /// Sends the current execution context to the GPU.
+ void Flush(bool release_fence = true, vk::Semaphore semaphore = nullptr);
- /// Sends the current execution context to the GPU and waits for it to complete. It invalidates
- /// the current execution context and returns a new one.
- VKExecutionContext Finish(vk::Semaphore semaphore = nullptr);
+ /// Sends the current execution context to the GPU and waits for it to complete.
+ void Finish(bool release_fence = true, vk::Semaphore semaphore = nullptr);
private:
void SubmitExecution(vk::Semaphore semaphore);
@@ -44,26 +82,4 @@ private:
VKFence* next_fence = nullptr;
};
-class VKExecutionContext {
- friend class VKScheduler;
-
-public:
- VKExecutionContext() = default;
-
- VKFence& GetFence() const {
- return *fence;
- }
-
- vk::CommandBuffer GetCommandBuffer() const {
- return cmdbuf;
- }
-
-private:
- explicit VKExecutionContext(VKFence* fence, vk::CommandBuffer cmdbuf)
- : fence{fence}, cmdbuf{cmdbuf} {}
-
- VKFence* fence{};
- vk::CommandBuffer cmdbuf;
-};
-
} // namespace Vulkan
diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
index 97ce214b1..1bb04607b 100644
--- a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
+++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
@@ -949,6 +949,14 @@ private:
return {};
}
+ Id BranchIndirect(Operation operation) {
+ const Id op_a = VisitOperand<Type::Uint>(operation, 0);
+
+ Emit(OpStore(jmp_to, op_a));
+ BranchingOp([&]() { Emit(OpBranch(continue_label)); });
+ return {};
+ }
+
Id PushFlowStack(Operation operation) {
const auto target = std::get_if<ImmediateNode>(&*operation[0]);
ASSERT(target);
@@ -1334,6 +1342,7 @@ private:
&SPIRVDecompiler::ImageStore,
&SPIRVDecompiler::Branch,
+ &SPIRVDecompiler::BranchIndirect,
&SPIRVDecompiler::PushFlowStack,
&SPIRVDecompiler::PopFlowStack,
&SPIRVDecompiler::Exit,
diff --git a/src/video_core/renderer_vulkan/vk_stream_buffer.cpp b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp
index 58ffa42f2..62f1427f5 100644
--- a/src/video_core/renderer_vulkan/vk_stream_buffer.cpp
+++ b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp
@@ -46,12 +46,12 @@ std::tuple<u8*, u64, bool> VKStreamBuffer::Reserve(u64 size) {
return {mapped_pointer + offset, offset, invalidation_mark.has_value()};
}
-VKExecutionContext VKStreamBuffer::Send(VKExecutionContext exctx, u64 size) {
+void VKStreamBuffer::Send(u64 size) {
ASSERT_MSG(size <= mapped_size, "Reserved size is too small");
if (invalidation_mark) {
// TODO(Rodrigo): Find a better way to invalidate than waiting for all watches to finish.
- exctx = scheduler.Flush();
+ scheduler.Flush();
std::for_each(watches.begin(), watches.begin() + *invalidation_mark,
[&](auto& resource) { resource->Wait(); });
invalidation_mark = std::nullopt;
@@ -62,11 +62,9 @@ VKExecutionContext VKStreamBuffer::Send(VKExecutionContext exctx, u64 size) {
ReserveWatches(WATCHES_RESERVE_CHUNK);
}
// Add a watch for this allocation.
- watches[used_watches++]->Watch(exctx.GetFence());
+ watches[used_watches++]->Watch(scheduler.GetFence());
offset += size;
-
- return exctx;
}
void VKStreamBuffer::CreateBuffers(VKMemoryManager& memory_manager, vk::BufferUsageFlags usage) {
diff --git a/src/video_core/renderer_vulkan/vk_stream_buffer.h b/src/video_core/renderer_vulkan/vk_stream_buffer.h
index 69d036ccd..842e54162 100644
--- a/src/video_core/renderer_vulkan/vk_stream_buffer.h
+++ b/src/video_core/renderer_vulkan/vk_stream_buffer.h
@@ -37,7 +37,7 @@ public:
std::tuple<u8*, u64, bool> Reserve(u64 size);
/// Ensures that "size" bytes of memory are available to the GPU, potentially recording a copy.
- [[nodiscard]] VKExecutionContext Send(VKExecutionContext exctx, u64 size);
+ void Send(u64 size);
vk::Buffer GetBuffer() const {
return *buffer;
diff --git a/src/video_core/shader/control_flow.cpp b/src/video_core/shader/control_flow.cpp
new file mode 100644
index 000000000..fdcc970ff
--- /dev/null
+++ b/src/video_core/shader/control_flow.cpp
@@ -0,0 +1,476 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <list>
+#include <map>
+#include <stack>
+#include <unordered_map>
+#include <unordered_set>
+#include <vector>
+
+#include "common/assert.h"
+#include "common/common_types.h"
+#include "video_core/shader/control_flow.h"
+#include "video_core/shader/shader_ir.h"
+
+namespace VideoCommon::Shader {
+
+using Tegra::Shader::Instruction;
+using Tegra::Shader::OpCode;
+
+constexpr s32 unassigned_branch = -2;
+
+struct Query {
+ u32 address{};
+ std::stack<u32> ssy_stack{};
+ std::stack<u32> pbk_stack{};
+};
+
+struct BlockStack {
+ BlockStack() = default;
+ BlockStack(const BlockStack& b) = default;
+ BlockStack(const Query& q) : ssy_stack{q.ssy_stack}, pbk_stack{q.pbk_stack} {}
+ std::stack<u32> ssy_stack{};
+ std::stack<u32> pbk_stack{};
+};
+
+struct BlockBranchInfo {
+ Condition condition{};
+ s32 address{exit_branch};
+ bool kill{};
+ bool is_sync{};
+ bool is_brk{};
+ bool ignore{};
+};
+
+struct BlockInfo {
+ u32 start{};
+ u32 end{};
+ bool visited{};
+ BlockBranchInfo branch{};
+
+ bool IsInside(const u32 address) const {
+ return start <= address && address <= end;
+ }
+};
+
+struct CFGRebuildState {
+ explicit CFGRebuildState(const ProgramCode& program_code, const std::size_t program_size,
+ const u32 start)
+ : program_code{program_code}, program_size{program_size}, start{start} {}
+
+ u32 start{};
+ std::vector<BlockInfo> block_info{};
+ std::list<u32> inspect_queries{};
+ std::list<Query> queries{};
+ std::unordered_map<u32, u32> registered{};
+ std::unordered_set<u32> labels{};
+ std::map<u32, u32> ssy_labels{};
+ std::map<u32, u32> pbk_labels{};
+ std::unordered_map<u32, BlockStack> stacks{};
+ const ProgramCode& program_code;
+ const std::size_t program_size;
+};
+
+enum class BlockCollision : u32 { None, Found, Inside };
+
+std::pair<BlockCollision, u32> TryGetBlock(CFGRebuildState& state, u32 address) {
+ const auto& blocks = state.block_info;
+ for (u32 index = 0; index < blocks.size(); index++) {
+ if (blocks[index].start == address) {
+ return {BlockCollision::Found, index};
+ }
+ if (blocks[index].IsInside(address)) {
+ return {BlockCollision::Inside, index};
+ }
+ }
+ return {BlockCollision::None, -1};
+}
+
+struct ParseInfo {
+ BlockBranchInfo branch_info{};
+ u32 end_address{};
+};
+
+BlockInfo& CreateBlockInfo(CFGRebuildState& state, u32 start, u32 end) {
+ auto& it = state.block_info.emplace_back();
+ it.start = start;
+ it.end = end;
+ const u32 index = static_cast<u32>(state.block_info.size() - 1);
+ state.registered.insert({start, index});
+ return it;
+}
+
+Pred GetPredicate(u32 index, bool negated) {
+ return static_cast<Pred>(index + (negated ? 8 : 0));
+}
+
+/**
+ * Returns whether the instruction at the specified offset is a 'sched' instruction.
+ * Sched instructions always appear before a sequence of 3 instructions.
+ */
+constexpr bool IsSchedInstruction(u32 offset, u32 main_offset) {
+ constexpr u32 SchedPeriod = 4;
+ u32 absolute_offset = offset - main_offset;
+
+ return (absolute_offset % SchedPeriod) == 0;
+}
+
+enum class ParseResult : u32 {
+ ControlCaught,
+ BlockEnd,
+ AbnormalFlow,
+};
+
+std::pair<ParseResult, ParseInfo> ParseCode(CFGRebuildState& state, u32 address) {
+ u32 offset = static_cast<u32>(address);
+ const u32 end_address = static_cast<u32>(state.program_size / sizeof(Instruction));
+ ParseInfo parse_info{};
+
+ const auto insert_label = [](CFGRebuildState& state, u32 address) {
+ const auto pair = state.labels.emplace(address);
+ if (pair.second) {
+ state.inspect_queries.push_back(address);
+ }
+ };
+
+ while (true) {
+ if (offset >= end_address) {
+ // ASSERT_OR_EXECUTE can't be used, as it ignores the break
+ ASSERT_MSG(false, "Shader passed the current limit!");
+ parse_info.branch_info.address = exit_branch;
+ parse_info.branch_info.ignore = false;
+ break;
+ }
+ if (state.registered.count(offset) != 0) {
+ parse_info.branch_info.address = offset;
+ parse_info.branch_info.ignore = true;
+ break;
+ }
+ if (IsSchedInstruction(offset, state.start)) {
+ offset++;
+ continue;
+ }
+ const Instruction instr = {state.program_code[offset]};
+ const auto opcode = OpCode::Decode(instr);
+ if (!opcode || opcode->get().GetType() != OpCode::Type::Flow) {
+ offset++;
+ continue;
+ }
+
+ switch (opcode->get().GetId()) {
+ case OpCode::Id::EXIT: {
+ const auto pred_index = static_cast<u32>(instr.pred.pred_index);
+ parse_info.branch_info.condition.predicate =
+ GetPredicate(pred_index, instr.negate_pred != 0);
+ if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
+ offset++;
+ continue;
+ }
+ const ConditionCode cc = instr.flow_condition_code;
+ parse_info.branch_info.condition.cc = cc;
+ if (cc == ConditionCode::F) {
+ offset++;
+ continue;
+ }
+ parse_info.branch_info.address = exit_branch;
+ parse_info.branch_info.kill = false;
+ parse_info.branch_info.is_sync = false;
+ parse_info.branch_info.is_brk = false;
+ parse_info.branch_info.ignore = false;
+ parse_info.end_address = offset;
+
+ return {ParseResult::ControlCaught, parse_info};
+ }
+ case OpCode::Id::BRA: {
+ if (instr.bra.constant_buffer != 0) {
+ return {ParseResult::AbnormalFlow, parse_info};
+ }
+ const auto pred_index = static_cast<u32>(instr.pred.pred_index);
+ parse_info.branch_info.condition.predicate =
+ GetPredicate(pred_index, instr.negate_pred != 0);
+ if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
+ offset++;
+ continue;
+ }
+ const ConditionCode cc = instr.flow_condition_code;
+ parse_info.branch_info.condition.cc = cc;
+ if (cc == ConditionCode::F) {
+ offset++;
+ continue;
+ }
+ const u32 branch_offset = offset + instr.bra.GetBranchTarget();
+ if (branch_offset == 0) {
+ parse_info.branch_info.address = exit_branch;
+ } else {
+ parse_info.branch_info.address = branch_offset;
+ }
+ insert_label(state, branch_offset);
+ parse_info.branch_info.kill = false;
+ parse_info.branch_info.is_sync = false;
+ parse_info.branch_info.is_brk = false;
+ parse_info.branch_info.ignore = false;
+ parse_info.end_address = offset;
+
+ return {ParseResult::ControlCaught, parse_info};
+ }
+ case OpCode::Id::SYNC: {
+ const auto pred_index = static_cast<u32>(instr.pred.pred_index);
+ parse_info.branch_info.condition.predicate =
+ GetPredicate(pred_index, instr.negate_pred != 0);
+ if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
+ offset++;
+ continue;
+ }
+ const ConditionCode cc = instr.flow_condition_code;
+ parse_info.branch_info.condition.cc = cc;
+ if (cc == ConditionCode::F) {
+ offset++;
+ continue;
+ }
+ parse_info.branch_info.address = unassigned_branch;
+ parse_info.branch_info.kill = false;
+ parse_info.branch_info.is_sync = true;
+ parse_info.branch_info.is_brk = false;
+ parse_info.branch_info.ignore = false;
+ parse_info.end_address = offset;
+
+ return {ParseResult::ControlCaught, parse_info};
+ }
+ case OpCode::Id::BRK: {
+ const auto pred_index = static_cast<u32>(instr.pred.pred_index);
+ parse_info.branch_info.condition.predicate =
+ GetPredicate(pred_index, instr.negate_pred != 0);
+ if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
+ offset++;
+ continue;
+ }
+ const ConditionCode cc = instr.flow_condition_code;
+ parse_info.branch_info.condition.cc = cc;
+ if (cc == ConditionCode::F) {
+ offset++;
+ continue;
+ }
+ parse_info.branch_info.address = unassigned_branch;
+ parse_info.branch_info.kill = false;
+ parse_info.branch_info.is_sync = false;
+ parse_info.branch_info.is_brk = true;
+ parse_info.branch_info.ignore = false;
+ parse_info.end_address = offset;
+
+ return {ParseResult::ControlCaught, parse_info};
+ }
+ case OpCode::Id::KIL: {
+ const auto pred_index = static_cast<u32>(instr.pred.pred_index);
+ parse_info.branch_info.condition.predicate =
+ GetPredicate(pred_index, instr.negate_pred != 0);
+ if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
+ offset++;
+ continue;
+ }
+ const ConditionCode cc = instr.flow_condition_code;
+ parse_info.branch_info.condition.cc = cc;
+ if (cc == ConditionCode::F) {
+ offset++;
+ continue;
+ }
+ parse_info.branch_info.address = exit_branch;
+ parse_info.branch_info.kill = true;
+ parse_info.branch_info.is_sync = false;
+ parse_info.branch_info.is_brk = false;
+ parse_info.branch_info.ignore = false;
+ parse_info.end_address = offset;
+
+ return {ParseResult::ControlCaught, parse_info};
+ }
+ case OpCode::Id::SSY: {
+ const u32 target = offset + instr.bra.GetBranchTarget();
+ insert_label(state, target);
+ state.ssy_labels.emplace(offset, target);
+ break;
+ }
+ case OpCode::Id::PBK: {
+ const u32 target = offset + instr.bra.GetBranchTarget();
+ insert_label(state, target);
+ state.pbk_labels.emplace(offset, target);
+ break;
+ }
+ case OpCode::Id::BRX: {
+ return {ParseResult::AbnormalFlow, parse_info};
+ }
+ default:
+ break;
+ }
+
+ offset++;
+ }
+ parse_info.branch_info.kill = false;
+ parse_info.branch_info.is_sync = false;
+ parse_info.branch_info.is_brk = false;
+ parse_info.end_address = offset - 1;
+ return {ParseResult::BlockEnd, parse_info};
+}
+
+bool TryInspectAddress(CFGRebuildState& state) {
+ if (state.inspect_queries.empty()) {
+ return false;
+ }
+
+ const u32 address = state.inspect_queries.front();
+ state.inspect_queries.pop_front();
+ const auto [result, block_index] = TryGetBlock(state, address);
+ switch (result) {
+ case BlockCollision::Found: {
+ return true;
+ }
+ case BlockCollision::Inside: {
+ // This case is the tricky one:
+ // We need to Split the block in 2 sepparate blocks
+ const u32 end = state.block_info[block_index].end;
+ BlockInfo& new_block = CreateBlockInfo(state, address, end);
+ BlockInfo& current_block = state.block_info[block_index];
+ current_block.end = address - 1;
+ new_block.branch = current_block.branch;
+ BlockBranchInfo forward_branch{};
+ forward_branch.address = address;
+ forward_branch.ignore = true;
+ current_block.branch = forward_branch;
+ return true;
+ }
+ default:
+ break;
+ }
+ const auto [parse_result, parse_info] = ParseCode(state, address);
+ if (parse_result == ParseResult::AbnormalFlow) {
+ // if it's AbnormalFlow, we end it as false, ending the CFG reconstruction
+ return false;
+ }
+
+ BlockInfo& block_info = CreateBlockInfo(state, address, parse_info.end_address);
+ block_info.branch = parse_info.branch_info;
+ if (parse_info.branch_info.condition.IsUnconditional()) {
+ return true;
+ }
+
+ const u32 fallthrough_address = parse_info.end_address + 1;
+ state.inspect_queries.push_front(fallthrough_address);
+ return true;
+}
+
+bool TryQuery(CFGRebuildState& state) {
+ const auto gather_labels = [](std::stack<u32>& cc, std::map<u32, u32>& labels,
+ BlockInfo& block) {
+ auto gather_start = labels.lower_bound(block.start);
+ const auto gather_end = labels.upper_bound(block.end);
+ while (gather_start != gather_end) {
+ cc.push(gather_start->second);
+ gather_start++;
+ }
+ };
+ if (state.queries.empty()) {
+ return false;
+ }
+ Query& q = state.queries.front();
+ const u32 block_index = state.registered[q.address];
+ BlockInfo& block = state.block_info[block_index];
+ // If the block is visted, check if the stacks match, else gather the ssy/pbk
+ // labels into the current stack and look if the branch at the end of the block
+ // consumes a label. Schedule new queries accordingly
+ if (block.visited) {
+ BlockStack& stack = state.stacks[q.address];
+ const bool all_okay = (stack.ssy_stack.size() == 0 || q.ssy_stack == stack.ssy_stack) &&
+ (stack.pbk_stack.size() == 0 || q.pbk_stack == stack.pbk_stack);
+ state.queries.pop_front();
+ return all_okay;
+ }
+ block.visited = true;
+ state.stacks[q.address] = BlockStack{q};
+ Query q2(q);
+ state.queries.pop_front();
+ gather_labels(q2.ssy_stack, state.ssy_labels, block);
+ gather_labels(q2.pbk_stack, state.pbk_labels, block);
+ if (!block.branch.condition.IsUnconditional()) {
+ q2.address = block.end + 1;
+ state.queries.push_back(q2);
+ }
+ Query conditional_query{q2};
+ if (block.branch.is_sync) {
+ if (block.branch.address == unassigned_branch) {
+ block.branch.address = conditional_query.ssy_stack.top();
+ }
+ conditional_query.ssy_stack.pop();
+ }
+ if (block.branch.is_brk) {
+ if (block.branch.address == unassigned_branch) {
+ block.branch.address = conditional_query.pbk_stack.top();
+ }
+ conditional_query.pbk_stack.pop();
+ }
+ conditional_query.address = block.branch.address;
+ state.queries.push_back(conditional_query);
+ return true;
+}
+
+std::optional<ShaderCharacteristics> ScanFlow(const ProgramCode& program_code, u32 program_size,
+ u32 start_address) {
+ CFGRebuildState state{program_code, program_size, start_address};
+ // Inspect Code and generate blocks
+ state.labels.clear();
+ state.labels.emplace(start_address);
+ state.inspect_queries.push_back(state.start);
+ while (!state.inspect_queries.empty()) {
+ if (!TryInspectAddress(state)) {
+ return {};
+ }
+ }
+ // Decompile Stacks
+ Query start_query{};
+ start_query.address = state.start;
+ state.queries.push_back(start_query);
+ bool decompiled = true;
+ while (!state.queries.empty()) {
+ if (!TryQuery(state)) {
+ decompiled = false;
+ break;
+ }
+ }
+ // Sort and organize results
+ std::sort(state.block_info.begin(), state.block_info.end(),
+ [](const BlockInfo& a, const BlockInfo& b) -> bool { return a.start < b.start; });
+ ShaderCharacteristics result_out{};
+ result_out.decompilable = decompiled;
+ result_out.start = start_address;
+ result_out.end = start_address;
+ for (auto& block : state.block_info) {
+ ShaderBlock new_block{};
+ new_block.start = block.start;
+ new_block.end = block.end;
+ new_block.ignore_branch = block.branch.ignore;
+ if (!new_block.ignore_branch) {
+ new_block.branch.cond = block.branch.condition;
+ new_block.branch.kills = block.branch.kill;
+ new_block.branch.address = block.branch.address;
+ }
+ result_out.end = std::max(result_out.end, block.end);
+ result_out.blocks.push_back(new_block);
+ }
+ if (result_out.decompilable) {
+ result_out.labels = std::move(state.labels);
+ return {result_out};
+ }
+ // If it's not decompilable, merge the unlabelled blocks together
+ auto back = result_out.blocks.begin();
+ auto next = std::next(back);
+ while (next != result_out.blocks.end()) {
+ if (state.labels.count(next->start) == 0 && next->start == back->end + 1) {
+ back->end = next->end;
+ next = result_out.blocks.erase(next);
+ continue;
+ }
+ back = next;
+ next++;
+ }
+ return {result_out};
+}
+} // namespace VideoCommon::Shader
diff --git a/src/video_core/shader/control_flow.h b/src/video_core/shader/control_flow.h
new file mode 100644
index 000000000..5e8ea3271
--- /dev/null
+++ b/src/video_core/shader/control_flow.h
@@ -0,0 +1,63 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <cstring>
+#include <list>
+#include <optional>
+#include <unordered_set>
+
+#include "video_core/engines/shader_bytecode.h"
+#include "video_core/shader/shader_ir.h"
+
+namespace VideoCommon::Shader {
+
+using Tegra::Shader::ConditionCode;
+using Tegra::Shader::Pred;
+
+constexpr s32 exit_branch = -1;
+
+struct Condition {
+ Pred predicate{Pred::UnusedIndex};
+ ConditionCode cc{ConditionCode::T};
+
+ bool IsUnconditional() const {
+ return predicate == Pred::UnusedIndex && cc == ConditionCode::T;
+ }
+ bool operator==(const Condition& other) const {
+ return std::tie(predicate, cc) == std::tie(other.predicate, other.cc);
+ }
+};
+
+struct ShaderBlock {
+ u32 start{};
+ u32 end{};
+ bool ignore_branch{};
+ struct Branch {
+ Condition cond{};
+ bool kills{};
+ s32 address{};
+ bool operator==(const Branch& b) const {
+ return std::tie(cond, kills, address) == std::tie(b.cond, b.kills, b.address);
+ }
+ } branch{};
+ bool operator==(const ShaderBlock& sb) const {
+ return std::tie(start, end, ignore_branch, branch) ==
+ std::tie(sb.start, sb.end, sb.ignore_branch, sb.branch);
+ }
+};
+
+struct ShaderCharacteristics {
+ std::list<ShaderBlock> blocks{};
+ bool decompilable{};
+ u32 start{};
+ u32 end{};
+ std::unordered_set<u32> labels{};
+};
+
+std::optional<ShaderCharacteristics> ScanFlow(const ProgramCode& program_code, u32 program_size,
+ u32 start_address);
+
+} // namespace VideoCommon::Shader
diff --git a/src/video_core/shader/decode.cpp b/src/video_core/shader/decode.cpp
index 2c9ff28f2..29c8895c5 100644
--- a/src/video_core/shader/decode.cpp
+++ b/src/video_core/shader/decode.cpp
@@ -11,6 +11,7 @@
#include "common/common_types.h"
#include "video_core/engines/shader_bytecode.h"
#include "video_core/engines/shader_header.h"
+#include "video_core/shader/control_flow.h"
#include "video_core/shader/node_helper.h"
#include "video_core/shader/shader_ir.h"
@@ -21,20 +22,6 @@ using Tegra::Shader::OpCode;
namespace {
-/// Merges exit method of two parallel branches.
-constexpr ExitMethod ParallelExit(ExitMethod a, ExitMethod b) {
- if (a == ExitMethod::Undetermined) {
- return b;
- }
- if (b == ExitMethod::Undetermined) {
- return a;
- }
- if (a == b) {
- return a;
- }
- return ExitMethod::Conditional;
-}
-
/**
* Returns whether the instruction at the specified offset is a 'sched' instruction.
* Sched instructions always appear before a sequence of 3 instructions.
@@ -51,85 +38,104 @@ constexpr bool IsSchedInstruction(u32 offset, u32 main_offset) {
void ShaderIR::Decode() {
std::memcpy(&header, program_code.data(), sizeof(Tegra::Shader::Header));
- std::set<u32> labels;
- const ExitMethod exit_method = Scan(main_offset, MAX_PROGRAM_LENGTH, labels);
- if (exit_method != ExitMethod::AlwaysEnd) {
- UNREACHABLE_MSG("Program does not always end");
- }
-
- if (labels.empty()) {
- basic_blocks.insert({main_offset, DecodeRange(main_offset, MAX_PROGRAM_LENGTH)});
+ disable_flow_stack = false;
+ const auto info = ScanFlow(program_code, program_size, main_offset);
+ if (info) {
+ const auto& shader_info = *info;
+ coverage_begin = shader_info.start;
+ coverage_end = shader_info.end;
+ if (shader_info.decompilable) {
+ disable_flow_stack = true;
+ const auto insert_block = ([this](NodeBlock& nodes, u32 label) {
+ if (label == exit_branch) {
+ return;
+ }
+ basic_blocks.insert({label, nodes});
+ });
+ const auto& blocks = shader_info.blocks;
+ NodeBlock current_block;
+ u32 current_label = exit_branch;
+ for (auto& block : blocks) {
+ if (shader_info.labels.count(block.start) != 0) {
+ insert_block(current_block, current_label);
+ current_block.clear();
+ current_label = block.start;
+ }
+ if (!block.ignore_branch) {
+ DecodeRangeInner(current_block, block.start, block.end);
+ InsertControlFlow(current_block, block);
+ } else {
+ DecodeRangeInner(current_block, block.start, block.end + 1);
+ }
+ }
+ insert_block(current_block, current_label);
+ return;
+ }
+ LOG_WARNING(HW_GPU, "Flow Stack Removing Failed! Falling back to old method");
+ // we can't decompile it, fallback to standard method
+ for (const auto& block : shader_info.blocks) {
+ basic_blocks.insert({block.start, DecodeRange(block.start, block.end + 1)});
+ }
return;
}
+ LOG_WARNING(HW_GPU, "Flow Analysis Failed! Falling back to brute force compiling");
+
+ // Now we need to deal with an undecompilable shader. We need to brute force
+ // a shader that captures every position.
+ coverage_begin = main_offset;
+ const u32 shader_end = static_cast<u32>(program_size / sizeof(u64));
+ coverage_end = shader_end;
+ for (u32 label = main_offset; label < shader_end; label++) {
+ basic_blocks.insert({label, DecodeRange(label, label + 1)});
+ }
+}
- labels.insert(main_offset);
-
- for (const u32 label : labels) {
- const auto next_it = labels.lower_bound(label + 1);
- const u32 next_label = next_it == labels.end() ? MAX_PROGRAM_LENGTH : *next_it;
+NodeBlock ShaderIR::DecodeRange(u32 begin, u32 end) {
+ NodeBlock basic_block;
+ DecodeRangeInner(basic_block, begin, end);
+ return basic_block;
+}
- basic_blocks.insert({label, DecodeRange(label, next_label)});
+void ShaderIR::DecodeRangeInner(NodeBlock& bb, u32 begin, u32 end) {
+ for (u32 pc = begin; pc < (begin > end ? MAX_PROGRAM_LENGTH : end);) {
+ pc = DecodeInstr(bb, pc);
}
}
-ExitMethod ShaderIR::Scan(u32 begin, u32 end, std::set<u32>& labels) {
- const auto [iter, inserted] =
- exit_method_map.emplace(std::make_pair(begin, end), ExitMethod::Undetermined);
- ExitMethod& exit_method = iter->second;
- if (!inserted)
- return exit_method;
-
- for (u32 offset = begin; offset != end && offset != MAX_PROGRAM_LENGTH; ++offset) {
- coverage_begin = std::min(coverage_begin, offset);
- coverage_end = std::max(coverage_end, offset + 1);
-
- const Instruction instr = {program_code[offset]};
- const auto opcode = OpCode::Decode(instr);
- if (!opcode)
- continue;
- switch (opcode->get().GetId()) {
- case OpCode::Id::EXIT: {
- // The EXIT instruction can be predicated, which means that the shader can conditionally
- // end on this instruction. We have to consider the case where the condition is not met
- // and check the exit method of that other basic block.
- using Tegra::Shader::Pred;
- if (instr.pred.pred_index == static_cast<u64>(Pred::UnusedIndex)) {
- return exit_method = ExitMethod::AlwaysEnd;
- } else {
- const ExitMethod not_met = Scan(offset + 1, end, labels);
- return exit_method = ParallelExit(ExitMethod::AlwaysEnd, not_met);
- }
+void ShaderIR::InsertControlFlow(NodeBlock& bb, const ShaderBlock& block) {
+ const auto apply_conditions = ([&](const Condition& cond, Node n) -> Node {
+ Node result = n;
+ if (cond.cc != ConditionCode::T) {
+ result = Conditional(GetConditionCode(cond.cc), {result});
}
- case OpCode::Id::BRA: {
- const u32 target = offset + instr.bra.GetBranchTarget();
- labels.insert(target);
- const ExitMethod no_jmp = Scan(offset + 1, end, labels);
- const ExitMethod jmp = Scan(target, end, labels);
- return exit_method = ParallelExit(no_jmp, jmp);
- }
- case OpCode::Id::SSY:
- case OpCode::Id::PBK: {
- // The SSY and PBK use a similar encoding as the BRA instruction.
- UNIMPLEMENTED_IF_MSG(instr.bra.constant_buffer != 0,
- "Constant buffer branching is not supported");
- const u32 target = offset + instr.bra.GetBranchTarget();
- labels.insert(target);
- // Continue scanning for an exit method.
- break;
+ if (cond.predicate != Pred::UnusedIndex) {
+ u32 pred = static_cast<u32>(cond.predicate);
+ const bool is_neg = pred > 7;
+ if (is_neg) {
+ pred -= 8;
+ }
+ result = Conditional(GetPredicate(pred, is_neg), {result});
}
- default:
- break;
+ return result;
+ });
+ if (block.branch.address < 0) {
+ if (block.branch.kills) {
+ Node n = Operation(OperationCode::Discard);
+ n = apply_conditions(block.branch.cond, n);
+ bb.push_back(n);
+ global_code.push_back(n);
+ return;
}
+ Node n = Operation(OperationCode::Exit);
+ n = apply_conditions(block.branch.cond, n);
+ bb.push_back(n);
+ global_code.push_back(n);
+ return;
}
- return exit_method = ExitMethod::AlwaysReturn;
-}
-
-NodeBlock ShaderIR::DecodeRange(u32 begin, u32 end) {
- NodeBlock basic_block;
- for (u32 pc = begin; pc < (begin > end ? MAX_PROGRAM_LENGTH : end);) {
- pc = DecodeInstr(basic_block, pc);
- }
- return basic_block;
+ Node n = Operation(OperationCode::Branch, Immediate(block.branch.address));
+ n = apply_conditions(block.branch.cond, n);
+ bb.push_back(n);
+ global_code.push_back(n);
}
u32 ShaderIR::DecodeInstr(NodeBlock& bb, u32 pc) {
@@ -140,15 +146,18 @@ u32 ShaderIR::DecodeInstr(NodeBlock& bb, u32 pc) {
const Instruction instr = {program_code[pc]};
const auto opcode = OpCode::Decode(instr);
+ const u32 nv_address = ConvertAddressToNvidiaSpace(pc);
// Decoding failure
if (!opcode) {
UNIMPLEMENTED_MSG("Unhandled instruction: {0:x}", instr.value);
+ bb.push_back(Comment(fmt::format("{:05x} Unimplemented Shader instruction (0x{:016x})",
+ nv_address, instr.value)));
return pc + 1;
}
- bb.push_back(
- Comment(fmt::format("{}: {} (0x{:016x})", pc, opcode->get().GetName(), instr.value)));
+ bb.push_back(Comment(
+ fmt::format("{:05x} {} (0x{:016x})", nv_address, opcode->get().GetName(), instr.value)));
using Tegra::Shader::Pred;
UNIMPLEMENTED_IF_MSG(instr.pred.full_pred == Pred::NeverExecute,
diff --git a/src/video_core/shader/decode/decode_integer_set.cpp b/src/video_core/shader/decode/decode_integer_set.cpp
deleted file mode 100644
index e69de29bb..000000000
--- a/src/video_core/shader/decode/decode_integer_set.cpp
+++ /dev/null
diff --git a/src/video_core/shader/decode/other.cpp b/src/video_core/shader/decode/other.cpp
index d46a8ab82..42e3de02f 100644
--- a/src/video_core/shader/decode/other.cpp
+++ b/src/video_core/shader/decode/other.cpp
@@ -91,11 +91,46 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
break;
}
case OpCode::Id::BRA: {
- UNIMPLEMENTED_IF_MSG(instr.bra.constant_buffer != 0,
- "BRA with constant buffers are not implemented");
+ Node branch;
+ if (instr.bra.constant_buffer == 0) {
+ const u32 target = pc + instr.bra.GetBranchTarget();
+ branch = Operation(OperationCode::Branch, Immediate(target));
+ } else {
+ const u32 target = pc + 1;
+ const Node op_a = GetConstBuffer(instr.cbuf36.index, instr.cbuf36.GetOffset());
+ const Node convert = SignedOperation(OperationCode::IArithmeticShiftRight, true,
+ PRECISE, op_a, Immediate(3));
+ const Node operand =
+ Operation(OperationCode::IAdd, PRECISE, convert, Immediate(target));
+ branch = Operation(OperationCode::BranchIndirect, convert);
+ }
- const u32 target = pc + instr.bra.GetBranchTarget();
- const Node branch = Operation(OperationCode::Branch, Immediate(target));
+ const Tegra::Shader::ConditionCode cc = instr.flow_condition_code;
+ if (cc != Tegra::Shader::ConditionCode::T) {
+ bb.push_back(Conditional(GetConditionCode(cc), {branch}));
+ } else {
+ bb.push_back(branch);
+ }
+ break;
+ }
+ case OpCode::Id::BRX: {
+ Node operand;
+ if (instr.brx.constant_buffer != 0) {
+ const s32 target = pc + 1;
+ const Node index = GetRegister(instr.gpr8);
+ const Node op_a =
+ GetConstBufferIndirect(instr.cbuf36.index, instr.cbuf36.GetOffset() + 0, index);
+ const Node convert = SignedOperation(OperationCode::IArithmeticShiftRight, true,
+ PRECISE, op_a, Immediate(3));
+ operand = Operation(OperationCode::IAdd, PRECISE, convert, Immediate(target));
+ } else {
+ const s32 target = pc + instr.brx.GetBranchExtend();
+ const Node op_a = GetRegister(instr.gpr8);
+ const Node convert = SignedOperation(OperationCode::IArithmeticShiftRight, true,
+ PRECISE, op_a, Immediate(3));
+ operand = Operation(OperationCode::IAdd, PRECISE, convert, Immediate(target));
+ }
+ const Node branch = Operation(OperationCode::BranchIndirect, operand);
const Tegra::Shader::ConditionCode cc = instr.flow_condition_code;
if (cc != Tegra::Shader::ConditionCode::T) {
@@ -109,6 +144,10 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
UNIMPLEMENTED_IF_MSG(instr.bra.constant_buffer != 0,
"Constant buffer flow is not supported");
+ if (disable_flow_stack) {
+ break;
+ }
+
// The SSY opcode tells the GPU where to re-converge divergent execution paths with SYNC.
const u32 target = pc + instr.bra.GetBranchTarget();
bb.push_back(
@@ -119,6 +158,10 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
UNIMPLEMENTED_IF_MSG(instr.bra.constant_buffer != 0,
"Constant buffer PBK is not supported");
+ if (disable_flow_stack) {
+ break;
+ }
+
// PBK pushes to a stack the address where BRK will jump to.
const u32 target = pc + instr.bra.GetBranchTarget();
bb.push_back(
@@ -130,6 +173,10 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T, "SYNC condition code used: {}",
static_cast<u32>(cc));
+ if (disable_flow_stack) {
+ break;
+ }
+
// The SYNC opcode jumps to the address previously set by the SSY opcode
bb.push_back(Operation(OperationCode::PopFlowStack, MetaStackClass::Ssy));
break;
@@ -138,6 +185,9 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
const Tegra::Shader::ConditionCode cc = instr.flow_condition_code;
UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T, "BRK condition code used: {}",
static_cast<u32>(cc));
+ if (disable_flow_stack) {
+ break;
+ }
// The BRK opcode jumps to the address previously set by the PBK opcode
bb.push_back(Operation(OperationCode::PopFlowStack, MetaStackClass::Pbk));
diff --git a/src/video_core/shader/decode/texture.cpp b/src/video_core/shader/decode/texture.cpp
index cb480be9b..323be3f14 100644
--- a/src/video_core/shader/decode/texture.cpp
+++ b/src/video_core/shader/decode/texture.cpp
@@ -269,7 +269,13 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) {
LOG_WARNING(HW_GPU, "TLDS.NODEP implementation is incomplete");
}
- WriteTexsInstructionFloat(bb, instr, GetTldsCode(instr, texture_type, is_array));
+ const Node4 components = GetTldsCode(instr, texture_type, is_array);
+
+ if (instr.tlds.fp32_flag) {
+ WriteTexsInstructionFloat(bb, instr, components);
+ } else {
+ WriteTexsInstructionHalfFloat(bb, instr, components);
+ }
break;
}
default:
diff --git a/src/video_core/shader/node.h b/src/video_core/shader/node.h
index 0ac83fcf0..7427ed896 100644
--- a/src/video_core/shader/node.h
+++ b/src/video_core/shader/node.h
@@ -148,11 +148,12 @@ enum class OperationCode {
ImageStore, /// (MetaImage, float[N] coords) -> void
- Branch, /// (uint branch_target) -> void
- PushFlowStack, /// (uint branch_target) -> void
- PopFlowStack, /// () -> void
- Exit, /// () -> void
- Discard, /// () -> void
+ Branch, /// (uint branch_target) -> void
+ BranchIndirect, /// (uint branch_target) -> void
+ PushFlowStack, /// (uint branch_target) -> void
+ PopFlowStack, /// () -> void
+ Exit, /// () -> void
+ Discard, /// () -> void
EmitVertex, /// () -> void
EndPrimitive, /// () -> void
diff --git a/src/video_core/shader/shader_ir.cpp b/src/video_core/shader/shader_ir.cpp
index 11b545cca..5994bfc4e 100644
--- a/src/video_core/shader/shader_ir.cpp
+++ b/src/video_core/shader/shader_ir.cpp
@@ -22,8 +22,8 @@ using Tegra::Shader::PredCondition;
using Tegra::Shader::PredOperation;
using Tegra::Shader::Register;
-ShaderIR::ShaderIR(const ProgramCode& program_code, u32 main_offset)
- : program_code{program_code}, main_offset{main_offset} {
+ShaderIR::ShaderIR(const ProgramCode& program_code, u32 main_offset, const std::size_t size)
+ : program_code{program_code}, main_offset{main_offset}, program_size{size} {
Decode();
}
diff --git a/src/video_core/shader/shader_ir.h b/src/video_core/shader/shader_ir.h
index e22548208..6145f0a70 100644
--- a/src/video_core/shader/shader_ir.h
+++ b/src/video_core/shader/shader_ir.h
@@ -22,18 +22,12 @@
namespace VideoCommon::Shader {
+struct ShaderBlock;
+
using ProgramCode = std::vector<u64>;
constexpr u32 MAX_PROGRAM_LENGTH = 0x1000;
-/// Describes the behaviour of code path of a given entry point and a return point.
-enum class ExitMethod {
- Undetermined, ///< Internal value. Only occur when analyzing JMP loop.
- AlwaysReturn, ///< All code paths reach the return point.
- Conditional, ///< Code path reaches the return point or an END instruction conditionally.
- AlwaysEnd, ///< All code paths reach a END instruction.
-};
-
class ConstBuffer {
public:
explicit ConstBuffer(u32 max_offset, bool is_indirect)
@@ -73,7 +67,7 @@ struct GlobalMemoryUsage {
class ShaderIR final {
public:
- explicit ShaderIR(const ProgramCode& program_code, u32 main_offset);
+ explicit ShaderIR(const ProgramCode& program_code, u32 main_offset, std::size_t size);
~ShaderIR();
const std::map<u32, NodeBlock>& GetBasicBlocks() const {
@@ -129,12 +123,20 @@ public:
return header;
}
+ bool IsFlowStackDisabled() const {
+ return disable_flow_stack;
+ }
+
+ u32 ConvertAddressToNvidiaSpace(const u32 address) const {
+ return (address - main_offset) * sizeof(Tegra::Shader::Instruction);
+ }
+
private:
void Decode();
- ExitMethod Scan(u32 begin, u32 end, std::set<u32>& labels);
-
NodeBlock DecodeRange(u32 begin, u32 end);
+ void DecodeRangeInner(NodeBlock& bb, u32 begin, u32 end);
+ void InsertControlFlow(NodeBlock& bb, const ShaderBlock& block);
/**
* Decodes a single instruction from Tegra to IR.
@@ -326,10 +328,11 @@ private:
const ProgramCode& program_code;
const u32 main_offset;
+ const std::size_t program_size;
+ bool disable_flow_stack{};
u32 coverage_begin{};
u32 coverage_end{};
- std::map<std::pair<u32, u32>, ExitMethod> exit_method_map;
std::map<u32, NodeBlock> basic_blocks;
NodeBlock global_code;
diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp
index 73978ff5b..b7f3fdf75 100644
--- a/src/yuzu/configuration/config.cpp
+++ b/src/yuzu/configuration/config.cpp
@@ -436,8 +436,6 @@ void Config::ReadControlValues() {
void Config::ReadCoreValues() {
qt_config->beginGroup(QStringLiteral("Core"));
- Settings::values.cpu_jit_enabled =
- ReadSetting(QStringLiteral("cpu_jit_enabled"), true).toBool();
Settings::values.use_multi_core = ReadSetting(QStringLiteral("use_multi_core"), false).toBool();
qt_config->endGroup();
@@ -831,7 +829,6 @@ void Config::SaveControlValues() {
void Config::SaveCoreValues() {
qt_config->beginGroup(QStringLiteral("Core"));
- WriteSetting(QStringLiteral("cpu_jit_enabled"), Settings::values.cpu_jit_enabled, true);
WriteSetting(QStringLiteral("use_multi_core"), Settings::values.use_multi_core, false);
qt_config->endGroup();
diff --git a/src/yuzu_cmd/config.cpp b/src/yuzu_cmd/config.cpp
index 30b22341b..067d58d80 100644
--- a/src/yuzu_cmd/config.cpp
+++ b/src/yuzu_cmd/config.cpp
@@ -340,7 +340,6 @@ void Config::ReadValues() {
}
// Core
- Settings::values.cpu_jit_enabled = sdl2_config->GetBoolean("Core", "cpu_jit_enabled", true);
Settings::values.use_multi_core = sdl2_config->GetBoolean("Core", "use_multi_core", false);
// Renderer
diff --git a/src/yuzu_cmd/default_ini.h b/src/yuzu_cmd/default_ini.h
index 4f1add434..0cfc111a6 100644
--- a/src/yuzu_cmd/default_ini.h
+++ b/src/yuzu_cmd/default_ini.h
@@ -76,10 +76,6 @@ motion_device=
touch_device=
[Core]
-# Whether to use the Just-In-Time (JIT) compiler for CPU emulation
-# 0: Interpreter (slow), 1 (default): JIT (fast)
-cpu_jit_enabled =
-
# Whether to use multi-core for CPU emulation
# 0 (default): Disabled, 1: Enabled
use_multi_core=
diff --git a/src/yuzu_tester/config.cpp b/src/yuzu_tester/config.cpp
index b96b7d279..9a11dc6c3 100644
--- a/src/yuzu_tester/config.cpp
+++ b/src/yuzu_tester/config.cpp
@@ -114,7 +114,6 @@ void Config::ReadValues() {
}
// Core
- Settings::values.cpu_jit_enabled = sdl2_config->GetBoolean("Core", "cpu_jit_enabled", true);
Settings::values.use_multi_core = sdl2_config->GetBoolean("Core", "use_multi_core", false);
// Renderer
diff --git a/src/yuzu_tester/default_ini.h b/src/yuzu_tester/default_ini.h
index 0f880d8c7..9a3e86d68 100644
--- a/src/yuzu_tester/default_ini.h
+++ b/src/yuzu_tester/default_ini.h
@@ -8,10 +8,6 @@ namespace DefaultINI {
const char* sdl2_config_file = R"(
[Core]
-# Whether to use the Just-In-Time (JIT) compiler for CPU emulation
-# 0: Interpreter (slow), 1 (default): JIT (fast)
-cpu_jit_enabled =
-
# Whether to use multi-core for CPU emulation
# 0 (default): Disabled, 1: Enabled
use_multi_core=