summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--src/common/thread.cpp35
-rw-r--r--src/common/thread.h20
-rw-r--r--src/core/hle/service/audio/audout_u.cpp28
-rw-r--r--src/core/hle/service/audio/audout_u.h3
-rw-r--r--src/core/hle/service/hid/controllers/debug_pad.cpp5
-rw-r--r--src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp13
-rw-r--r--src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h6
-rw-r--r--src/video_core/engines/maxwell_3d.h7
-rw-r--r--src/video_core/engines/shader_bytecode.h1
-rw-r--r--src/video_core/macro_interpreter.cpp29
-rw-r--r--src/video_core/macro_interpreter.h4
-rw-r--r--src/video_core/renderer_opengl/gl_shader_decompiler.cpp169
-rw-r--r--src/video_core/renderer_opengl/maxwell_to_gl.h5
13 files changed, 198 insertions, 127 deletions
diff --git a/src/common/thread.cpp b/src/common/thread.cpp
index 9e207118f..5144c0d9f 100644
--- a/src/common/thread.cpp
+++ b/src/common/thread.cpp
@@ -25,23 +25,6 @@
namespace Common {
-int CurrentThreadId() {
-#ifdef _MSC_VER
- return GetCurrentThreadId();
-#elif defined __APPLE__
- return mach_thread_self();
-#else
- return 0;
-#endif
-}
-
-#ifdef _WIN32
-// Supporting functions
-void SleepCurrentThread(int ms) {
- Sleep(ms);
-}
-#endif
-
#ifdef _MSC_VER
void SetThreadAffinity(std::thread::native_handle_type thread, u32 mask) {
@@ -62,7 +45,7 @@ void SwitchCurrentThread() {
// This is implemented much nicer in upcoming msvc++, see:
// http://msdn.microsoft.com/en-us/library/xcb2z8hs(VS.100).aspx
-void SetCurrentThreadName(const char* szThreadName) {
+void SetCurrentThreadName(const char* name) {
static const DWORD MS_VC_EXCEPTION = 0x406D1388;
#pragma pack(push, 8)
@@ -75,7 +58,7 @@ void SetCurrentThreadName(const char* szThreadName) {
#pragma pack(pop)
info.dwType = 0x1000;
- info.szName = szThreadName;
+ info.szName = name;
info.dwThreadID = -1; // dwThreadID;
info.dwFlags = 0;
@@ -107,10 +90,6 @@ void SetCurrentThreadAffinity(u32 mask) {
}
#ifndef _WIN32
-void SleepCurrentThread(int ms) {
- usleep(1000 * ms);
-}
-
void SwitchCurrentThread() {
usleep(1000 * 1);
}
@@ -118,15 +97,15 @@ void SwitchCurrentThread() {
// MinGW with the POSIX threading model does not support pthread_setname_np
#if !defined(_WIN32) || defined(_MSC_VER)
-void SetCurrentThreadName(const char* szThreadName) {
+void SetCurrentThreadName(const char* name) {
#ifdef __APPLE__
- pthread_setname_np(szThreadName);
+ pthread_setname_np(name);
#elif defined(__Bitrig__) || defined(__DragonFly__) || defined(__FreeBSD__) || defined(__OpenBSD__)
- pthread_set_name_np(pthread_self(), szThreadName);
+ pthread_set_name_np(pthread_self(), name);
#elif defined(__NetBSD__)
- pthread_setname_np(pthread_self(), "%s", (void*)szThreadName);
+ pthread_setname_np(pthread_self(), "%s", (void*)name);
#else
- pthread_setname_np(pthread_self(), szThreadName);
+ pthread_setname_np(pthread_self(), name);
#endif
}
#endif
diff --git a/src/common/thread.h b/src/common/thread.h
index 6cbdb96a3..2cf74452d 100644
--- a/src/common/thread.h
+++ b/src/common/thread.h
@@ -13,15 +13,8 @@
namespace Common {
-int CurrentThreadId();
-
-void SetThreadAffinity(std::thread::native_handle_type thread, u32 mask);
-void SetCurrentThreadAffinity(u32 mask);
-
class Event {
public:
- Event() : is_set(false) {}
-
void Set() {
std::lock_guard<std::mutex> lk(mutex);
if (!is_set) {
@@ -53,14 +46,14 @@ public:
}
private:
- bool is_set;
+ bool is_set = false;
std::condition_variable condvar;
std::mutex mutex;
};
class Barrier {
public:
- explicit Barrier(std::size_t count_) : count(count_), waiting(0), generation(0) {}
+ explicit Barrier(std::size_t count_) : count(count_) {}
/// Blocks until all "count" threads have called Sync()
void Sync() {
@@ -80,12 +73,13 @@ public:
private:
std::condition_variable condvar;
std::mutex mutex;
- const std::size_t count;
- std::size_t waiting;
- std::size_t generation; // Incremented once each time the barrier is used
+ std::size_t count;
+ std::size_t waiting = 0;
+ std::size_t generation = 0; // Incremented once each time the barrier is used
};
-void SleepCurrentThread(int ms);
+void SetThreadAffinity(std::thread::native_handle_type thread, u32 mask);
+void SetCurrentThreadAffinity(u32 mask);
void SwitchCurrentThread(); // On Linux, this is equal to sleep 1ms
void SetCurrentThreadName(const char* name);
diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp
index ff1edefbb..23e1f1165 100644
--- a/src/core/hle/service/audio/audout_u.cpp
+++ b/src/core/hle/service/audio/audout_u.cpp
@@ -44,8 +44,10 @@ enum class AudioState : u32 {
class IAudioOut final : public ServiceFramework<IAudioOut> {
public:
- IAudioOut(AudoutParams audio_params, AudioCore::AudioOut& audio_core)
- : ServiceFramework("IAudioOut"), audio_core(audio_core), audio_params(audio_params) {
+ IAudioOut(AudoutParams audio_params, AudioCore::AudioOut& audio_core, std::string&& device_name,
+ std::string&& unique_name)
+ : ServiceFramework("IAudioOut"), audio_core(audio_core), audio_params(audio_params),
+ device_name(std::move(device_name)) {
static const FunctionInfo functions[] = {
{0, &IAudioOut::GetAudioOutState, "GetAudioOutState"},
@@ -69,7 +71,7 @@ public:
Kernel::Event::Create(kernel, Kernel::ResetType::Sticky, "IAudioOutBufferReleased");
stream = audio_core.OpenStream(audio_params.sample_rate, audio_params.channel_count,
- "IAudioOut", [=]() { buffer_event->Signal(); });
+ std::move(unique_name), [=]() { buffer_event->Signal(); });
}
private:
@@ -177,6 +179,7 @@ private:
AudioCore::AudioOut& audio_core;
AudioCore::StreamPtr stream;
+ std::string device_name;
AudoutParams audio_params{};
@@ -199,7 +202,15 @@ void AudOutU::ListAudioOutsImpl(Kernel::HLERequestContext& ctx) {
void AudOutU::OpenAudioOutImpl(Kernel::HLERequestContext& ctx) {
LOG_DEBUG(Service_Audio, "called");
- ctx.WriteBuffer(DefaultDevice);
+ const auto device_name_data{ctx.ReadBuffer()};
+ std::string device_name;
+ if (device_name_data[0] != '\0') {
+ device_name.assign(device_name_data.begin(), device_name_data.end());
+ } else {
+ device_name.assign(DefaultDevice.begin(), DefaultDevice.end());
+ }
+ ctx.WriteBuffer(device_name);
+
IPC::RequestParser rp{ctx};
auto params{rp.PopRaw<AudoutParams>()};
if (params.channel_count <= 2) {
@@ -212,10 +223,9 @@ void AudOutU::OpenAudioOutImpl(Kernel::HLERequestContext& ctx) {
params.sample_rate = DefaultSampleRate;
}
- // TODO(bunnei): Support more than one IAudioOut interface. When we add this, ListAudioOutsImpl
- // will likely need to be updated as well.
- ASSERT_MSG(!audio_out_interface, "Unimplemented");
- audio_out_interface = std::make_shared<IAudioOut>(params, *audio_core);
+ std::string unique_name{fmt::format("{}-{}", device_name, audio_out_interfaces.size())};
+ auto audio_out_interface = std::make_shared<IAudioOut>(
+ params, *audio_core, std::move(device_name), std::move(unique_name));
IPC::ResponseBuilder rb{ctx, 6, 0, 1};
rb.Push(RESULT_SUCCESS);
@@ -224,6 +234,8 @@ void AudOutU::OpenAudioOutImpl(Kernel::HLERequestContext& ctx) {
rb.Push<u32>(static_cast<u32>(AudioCore::Codec::PcmFormat::Int16));
rb.Push<u32>(static_cast<u32>(AudioState::Stopped));
rb.PushIpcInterface<Audio::IAudioOut>(audio_out_interface);
+
+ audio_out_interfaces.push_back(std::move(audio_out_interface));
}
AudOutU::AudOutU() : ServiceFramework("audout:u") {
diff --git a/src/core/hle/service/audio/audout_u.h b/src/core/hle/service/audio/audout_u.h
index dcaf64708..aed4c43b2 100644
--- a/src/core/hle/service/audio/audout_u.h
+++ b/src/core/hle/service/audio/audout_u.h
@@ -4,6 +4,7 @@
#pragma once
+#include <vector>
#include "core/hle/service/service.h"
namespace AudioCore {
@@ -24,7 +25,7 @@ public:
~AudOutU() override;
private:
- std::shared_ptr<IAudioOut> audio_out_interface;
+ std::vector<std::shared_ptr<IAudioOut>> audio_out_interfaces;
std::unique_ptr<AudioCore::AudioOut> audio_core;
void ListAudioOutsImpl(Kernel::HLERequestContext& ctx);
diff --git a/src/core/hle/service/hid/controllers/debug_pad.cpp b/src/core/hle/service/hid/controllers/debug_pad.cpp
index e76c83aee..c22357d8c 100644
--- a/src/core/hle/service/hid/controllers/debug_pad.cpp
+++ b/src/core/hle/service/hid/controllers/debug_pad.cpp
@@ -71,8 +71,9 @@ void Controller_DebugPad::OnUpdate(u8* data, std::size_t size) {
void Controller_DebugPad::OnLoadInputDevices() {
std::transform(Settings::values.debug_pad_buttons.begin(),
- Settings::values.debug_pad_buttons.end(), buttons.begin(),
- Input::CreateDevice<Input::ButtonDevice>);
+ Settings::values.debug_pad_buttons.begin() +
+ Settings::NativeButton::NUM_BUTTONS_HID,
+ buttons.begin(), Input::CreateDevice<Input::ButtonDevice>);
std::transform(Settings::values.debug_pad_analogs.begin(),
Settings::values.debug_pad_analogs.end(), analogs.begin(),
Input::CreateDevice<Input::AnalogDevice>);
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp
index 7a88ae029..792d26e52 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp
@@ -5,6 +5,8 @@
#include <cstring>
#include "common/assert.h"
#include "common/logging/log.h"
+#include "core/core_timing.h"
+#include "core/core_timing_util.h"
#include "core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h"
namespace Service::Nvidia::Devices {
@@ -33,6 +35,8 @@ u32 nvhost_ctrl_gpu::ioctl(Ioctl command, const std::vector<u8>& input, std::vec
return ZBCQueryTable(input, output);
case IoctlCommand::IocFlushL2:
return FlushL2(input, output);
+ case IoctlCommand::IocGetGpuTime:
+ return GetGpuTime(input, output);
}
UNIMPLEMENTED_MSG("Unimplemented ioctl");
return 0;
@@ -169,4 +173,13 @@ u32 nvhost_ctrl_gpu::FlushL2(const std::vector<u8>& input, std::vector<u8>& outp
return 0;
}
+u32 nvhost_ctrl_gpu::GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output) {
+ LOG_DEBUG(Service_NVDRV, "called");
+ IoctlGetGpuTime params{};
+ std::memcpy(&params, input.data(), input.size());
+ params.gpu_time = CoreTiming::cyclesToNs(CoreTiming::GetTicks());
+ std::memcpy(output.data(), &params, output.size());
+ return 0;
+}
+
} // namespace Service::Nvidia::Devices
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h
index 3bbf028ad..240435eea 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h
+++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h
@@ -156,6 +156,11 @@ private:
};
static_assert(sizeof(IoctlFlushL2) == 8, "IoctlFlushL2 is incorrect size");
+ struct IoctlGetGpuTime {
+ u64_le gpu_time;
+ };
+ static_assert(sizeof(IoctlGetGpuTime) == 8, "IoctlGetGpuTime is incorrect size");
+
u32 GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output);
u32 GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output);
u32 GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output);
@@ -164,6 +169,7 @@ private:
u32 ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output);
u32 ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output);
u32 FlushL2(const std::vector<u8>& input, std::vector<u8>& output);
+ u32 GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output);
};
} // namespace Service::Nvidia::Devices
diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h
index 9e480dc39..eff6abd55 100644
--- a/src/video_core/engines/maxwell_3d.h
+++ b/src/video_core/engines/maxwell_3d.h
@@ -389,6 +389,13 @@ public:
ReverseSubtract = 3,
Min = 4,
Max = 5,
+
+ // These values are used by Nouveau and some games.
+ AddGL = 0x8006,
+ SubtractGL = 0x8007,
+ ReverseSubtractGL = 0x8008,
+ MinGL = 0x800a,
+ MaxGL = 0x800b
};
enum class Factor : u32 {
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index 83a6fd875..c5f502ce1 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -153,6 +153,7 @@ enum class PredCondition : u64 {
NotEqual = 5,
GreaterEqual = 6,
LessThanWithNan = 9,
+ LessEqualWithNan = 11,
GreaterThanWithNan = 12,
NotEqualWithNan = 13,
GreaterEqualWithNan = 14,
diff --git a/src/video_core/macro_interpreter.cpp b/src/video_core/macro_interpreter.cpp
index 335a8d407..2b0dea5cd 100644
--- a/src/video_core/macro_interpreter.cpp
+++ b/src/video_core/macro_interpreter.cpp
@@ -35,6 +35,7 @@ void MacroInterpreter::Reset() {
// The next parameter index starts at 1, because $r1 already has the value of the first
// parameter.
next_parameter_index = 1;
+ carry_flag = false;
}
bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) {
@@ -135,14 +136,28 @@ MacroInterpreter::Opcode MacroInterpreter::GetOpcode(u32 offset) const {
return {macro_memory[offset + pc / sizeof(u32)]};
}
-u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const {
+u32 MacroInterpreter::GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) {
switch (operation) {
- case ALUOperation::Add:
- return src_a + src_b;
- // TODO(Subv): Implement AddWithCarry
- case ALUOperation::Subtract:
- return src_a - src_b;
- // TODO(Subv): Implement SubtractWithBorrow
+ case ALUOperation::Add: {
+ const u64 result{static_cast<u64>(src_a) + src_b};
+ carry_flag = result > 0xffffffff;
+ return static_cast<u32>(result);
+ }
+ case ALUOperation::AddWithCarry: {
+ const u64 result{static_cast<u64>(src_a) + src_b + (carry_flag ? 1ULL : 0ULL)};
+ carry_flag = result > 0xffffffff;
+ return static_cast<u32>(result);
+ }
+ case ALUOperation::Subtract: {
+ const u64 result{static_cast<u64>(src_a) - src_b};
+ carry_flag = result < 0x100000000;
+ return static_cast<u32>(result);
+ }
+ case ALUOperation::SubtractWithBorrow: {
+ const u64 result{static_cast<u64>(src_a) - src_b - (carry_flag ? 0ULL : 1ULL)};
+ carry_flag = result < 0x100000000;
+ return static_cast<u32>(result);
+ }
case ALUOperation::Xor:
return src_a ^ src_b;
case ALUOperation::Or:
diff --git a/src/video_core/macro_interpreter.h b/src/video_core/macro_interpreter.h
index 62d1ce289..cde360288 100644
--- a/src/video_core/macro_interpreter.h
+++ b/src/video_core/macro_interpreter.h
@@ -117,7 +117,7 @@ private:
bool Step(u32 offset, bool is_delay_slot);
/// Calculates the result of an ALU operation. src_a OP src_b;
- u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b) const;
+ u32 GetALUResult(ALUOperation operation, u32 src_a, u32 src_b);
/// Performs the result operation on the input result and stores it in the specified register
/// (if necessary).
@@ -165,5 +165,7 @@ private:
std::vector<u32> parameters;
/// Index of the next parameter that will be fetched by the 'parm' instruction.
u32 next_parameter_index = 0;
+
+ bool carry_flag{};
};
} // namespace Tegra
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
index 90a88b91a..ba80e5832 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
@@ -1059,11 +1059,17 @@ private:
const std::string& op_a, const std::string& op_b) const {
using Tegra::Shader::PredCondition;
static const std::unordered_map<PredCondition, const char*> PredicateComparisonStrings = {
- {PredCondition::LessThan, "<"}, {PredCondition::Equal, "=="},
- {PredCondition::LessEqual, "<="}, {PredCondition::GreaterThan, ">"},
- {PredCondition::NotEqual, "!="}, {PredCondition::GreaterEqual, ">="},
- {PredCondition::LessThanWithNan, "<"}, {PredCondition::NotEqualWithNan, "!="},
- {PredCondition::GreaterThanWithNan, ">"}, {PredCondition::GreaterEqualWithNan, ">="}};
+ {PredCondition::LessThan, "<"},
+ {PredCondition::Equal, "=="},
+ {PredCondition::LessEqual, "<="},
+ {PredCondition::GreaterThan, ">"},
+ {PredCondition::NotEqual, "!="},
+ {PredCondition::GreaterEqual, ">="},
+ {PredCondition::LessThanWithNan, "<"},
+ {PredCondition::NotEqualWithNan, "!="},
+ {PredCondition::LessEqualWithNan, "<="},
+ {PredCondition::GreaterThanWithNan, ">"},
+ {PredCondition::GreaterEqualWithNan, ">="}};
const auto& comparison{PredicateComparisonStrings.find(condition)};
UNIMPLEMENTED_IF_MSG(comparison == PredicateComparisonStrings.end(),
@@ -1072,6 +1078,7 @@ private:
std::string predicate{'(' + op_a + ") " + comparison->second + " (" + op_b + ')'};
if (condition == PredCondition::LessThanWithNan ||
condition == PredCondition::NotEqualWithNan ||
+ condition == PredCondition::LessEqualWithNan ||
condition == PredCondition::GreaterThanWithNan ||
condition == PredCondition::GreaterEqualWithNan) {
predicate += " || isnan(" + op_a + ") || isnan(" + op_b + ')';
@@ -1243,14 +1250,7 @@ private:
regs.SetRegisterToInteger(dest, true, 0, result, 1, 1);
}
- void WriteTexsInstruction(const Instruction& instr, const std::string& coord,
- const std::string& texture) {
- // Add an extra scope and declare the texture coords inside to prevent
- // overwriting them in case they are used as outputs of the texs instruction.
- shader.AddLine('{');
- ++shader.scope;
- shader.AddLine(coord);
-
+ void WriteTexsInstruction(const Instruction& instr, const std::string& texture) {
// TEXS has two destination registers and a swizzle. The first two elements in the swizzle
// go into gpr0+0 and gpr0+1, and the rest goes into gpr28+0 and gpr28+1
@@ -1273,9 +1273,6 @@ private:
++written_components;
}
-
- --shader.scope;
- shader.AddLine('}');
}
static u32 TextureCoordinates(Tegra::Shader::TextureType texture_type) {
@@ -2563,7 +2560,6 @@ private:
}
// TODO: make sure coordinates are always indexed to gpr8 and gpr20 is always bias
// or lod.
- std::string op_c;
const std::string sampler =
GetSampler(instr.sampler, texture_type, is_array, depth_compare);
@@ -2586,27 +2582,35 @@ private:
}
case Tegra::Shader::TextureProcessMode::LB:
case Tegra::Shader::TextureProcessMode::LBA: {
- if (depth_compare) {
- if (is_array)
- op_c = regs.GetRegisterAsFloat(instr.gpr20.Value() + 2);
- else
- op_c = regs.GetRegisterAsFloat(instr.gpr20.Value() + 1);
- } else {
- op_c = regs.GetRegisterAsFloat(instr.gpr20);
- }
+ const std::string bias = [&]() {
+ if (depth_compare) {
+ if (is_array)
+ return regs.GetRegisterAsFloat(instr.gpr20.Value() + 2);
+ else
+ return regs.GetRegisterAsFloat(instr.gpr20.Value() + 1);
+ } else {
+ return regs.GetRegisterAsFloat(instr.gpr20);
+ }
+ }();
+ shader.AddLine("float bias = " + bias + ';');
+
// TODO: Figure if A suffix changes the equation at all.
- texture = "texture(" + sampler + ", coords, " + op_c + ')';
+ texture = "texture(" + sampler + ", coords, bias)";
break;
}
case Tegra::Shader::TextureProcessMode::LL:
case Tegra::Shader::TextureProcessMode::LLA: {
- if (num_coordinates <= 2) {
- op_c = regs.GetRegisterAsFloat(instr.gpr20);
- } else {
- op_c = regs.GetRegisterAsFloat(instr.gpr20.Value() + 1);
- }
+ const std::string lod = [&]() {
+ if (num_coordinates <= 2) {
+ return regs.GetRegisterAsFloat(instr.gpr20);
+ } else {
+ return regs.GetRegisterAsFloat(instr.gpr20.Value() + 1);
+ }
+ }();
+ shader.AddLine("float lod = " + lod + ';');
+
// TODO: Figure if A suffix changes the equation at all.
- texture = "textureLod(" + sampler + ", coords, " + op_c + ')';
+ texture = "textureLod(" + sampler + ", coords, lod)";
break;
}
default: {
@@ -2633,7 +2637,6 @@ private:
break;
}
case OpCode::Id::TEXS: {
- std::string coord;
Tegra::Shader::TextureType texture_type{instr.texs.GetTextureType()};
bool is_array{instr.texs.IsArrayTexture()};
@@ -2646,17 +2649,21 @@ private:
if (depth_compare)
num_coordinates += 1;
+ // Scope to avoid variable name overlaps.
+ shader.AddLine('{');
+ ++shader.scope;
+
switch (num_coordinates) {
case 2: {
if (is_array) {
const std::string index = regs.GetRegisterAsInteger(instr.gpr8);
const std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
const std::string y = regs.GetRegisterAsFloat(instr.gpr20);
- coord = "vec3 coords = vec3(" + x + ", " + y + ", " + index + ");";
+ shader.AddLine("vec3 coords = vec3(" + x + ", " + y + ", " + index + ");");
} else {
const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
const std::string y = regs.GetRegisterAsFloat(instr.gpr20);
- coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+ shader.AddLine("vec2 coords = vec2(" + x + ", " + y + ");");
}
break;
}
@@ -2666,13 +2673,13 @@ private:
const std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2);
const std::string z = regs.GetRegisterAsFloat(instr.gpr20);
- coord =
- "vec4 coords = vec4(" + x + ", " + y + ", " + z + ", " + index + ");";
+ shader.AddLine("vec4 coords = vec4(" + x + ", " + y + ", " + z + ", " +
+ index + ");");
} else {
const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
const std::string z = regs.GetRegisterAsFloat(instr.gpr20);
- coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+ shader.AddLine("vec3 coords = vec3(" + x + ", " + y + ", " + z + ");");
}
break;
}
@@ -2683,7 +2690,7 @@ private:
// Fallback to interpreting as a 2D texture for now
const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
const std::string y = regs.GetRegisterAsFloat(instr.gpr20);
- coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+ shader.AddLine("vec2 coords = vec2(" + x + ", " + y + ");");
texture_type = Tegra::Shader::TextureType::Texture2D;
is_array = false;
}
@@ -2715,14 +2722,16 @@ private:
}
}
if (!depth_compare) {
- WriteTexsInstruction(instr, coord, texture);
+ WriteTexsInstruction(instr, texture);
} else {
- WriteTexsInstruction(instr, coord, "vec4(" + texture + ')');
+ WriteTexsInstruction(instr, "vec4(" + texture + ')');
}
+
+ shader.AddLine('}');
+ --shader.scope;
break;
}
case OpCode::Id::TLDS: {
- std::string coord;
const Tegra::Shader::TextureType texture_type{instr.tlds.GetTextureType()};
const bool is_array{instr.tlds.IsArrayTexture()};
@@ -2736,12 +2745,16 @@ private:
UNIMPLEMENTED_IF_MSG(instr.tlds.UsesMiscMode(Tegra::Shader::TextureMiscMode::MZ),
"MZ is not implemented");
- u32 op_c_offset = 0;
+ u32 extra_op_offset = 0;
+
+ // Scope to avoid variable name overlaps.
+ shader.AddLine('{');
+ ++shader.scope;
switch (texture_type) {
case Tegra::Shader::TextureType::Texture1D: {
const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
- coord = "int coords = " + x + ';';
+ shader.AddLine("int coords = " + x + ';');
break;
}
case Tegra::Shader::TextureType::Texture2D: {
@@ -2749,8 +2762,8 @@ private:
const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
const std::string y = regs.GetRegisterAsInteger(instr.gpr20);
- coord = "ivec2 coords = ivec2(" + x + ", " + y + ");";
- op_c_offset = 1;
+ shader.AddLine("ivec2 coords = ivec2(" + x + ", " + y + ");");
+ extra_op_offset = 1;
break;
}
default:
@@ -2765,9 +2778,10 @@ private:
break;
}
case Tegra::Shader::TextureProcessMode::LL: {
- const std::string op_c =
- regs.GetRegisterAsInteger(instr.gpr20.Value() + op_c_offset);
- texture = "texelFetch(" + sampler + ", coords, " + op_c + ')';
+ shader.AddLine(
+ "float lod = " +
+ regs.GetRegisterAsInteger(instr.gpr20.Value() + extra_op_offset) + ';');
+ texture = "texelFetch(" + sampler + ", coords, lod)";
break;
}
default: {
@@ -2776,7 +2790,10 @@ private:
static_cast<u32>(instr.tlds.GetTextureProcessMode()));
}
}
- WriteTexsInstruction(instr, coord, texture);
+ WriteTexsInstruction(instr, texture);
+
+ --shader.scope;
+ shader.AddLine('}');
break;
}
case OpCode::Id::TLD4: {
@@ -2799,18 +2816,23 @@ private:
if (depth_compare)
num_coordinates += 1;
+ // Add an extra scope and declare the texture coords inside to prevent
+ // overwriting them in case they are used as outputs of the texs instruction.
+ shader.AddLine('{');
+ ++shader.scope;
+
switch (num_coordinates) {
case 2: {
const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
- coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+ shader.AddLine("vec2 coords = vec2(" + x + ", " + y + ");");
break;
}
case 3: {
const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
const std::string z = regs.GetRegisterAsFloat(instr.gpr8.Value() + 2);
- coord = "vec3 coords = vec3(" + x + ", " + y + ", " + z + ");";
+ shader.AddLine("vec3 coords = vec3(" + x + ", " + y + ", " + z + ");");
break;
}
default:
@@ -2818,17 +2840,13 @@ private:
static_cast<u32>(num_coordinates));
const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
- coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+ shader.AddLine("vec2 coords = vec2(" + x + ", " + y + ");");
texture_type = Tegra::Shader::TextureType::Texture2D;
}
const std::string sampler =
GetSampler(instr.sampler, texture_type, false, depth_compare);
- // Add an extra scope and declare the texture coords inside to prevent
- // overwriting them in case they are used as outputs of the texs instruction.
- shader.AddLine("{");
- ++shader.scope;
- shader.AddLine(coord);
+
const std::string texture = "textureGather(" + sampler + ", coords, " +
std::to_string(instr.tld4.component) + ')';
if (!depth_compare) {
@@ -2845,7 +2863,7 @@ private:
regs.SetRegisterToFloat(instr.gpr0, 0, texture, 1, 1, false);
}
--shader.scope;
- shader.AddLine("}");
+ shader.AddLine('}');
break;
}
case OpCode::Id::TLD4S: {
@@ -2856,6 +2874,10 @@ private:
instr.tld4s.UsesMiscMode(Tegra::Shader::TextureMiscMode::AOFFI),
"AOFFI is not implemented");
+ // Scope to avoid variable name overlaps.
+ shader.AddLine('{');
+ ++shader.scope;
+
const bool depth_compare =
instr.tld4s.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC);
const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
@@ -2863,28 +2885,33 @@ private:
// TODO(Subv): Figure out how the sampler type is encoded in the TLD4S instruction.
const std::string sampler = GetSampler(
instr.sampler, Tegra::Shader::TextureType::Texture2D, false, depth_compare);
- std::string coord;
if (!depth_compare) {
- coord = "vec2 coords = vec2(" + op_a + ", " + op_b + ");";
+ shader.AddLine("vec2 coords = vec2(" + op_a + ", " + op_b + ");");
} else {
// Note: TLD4S coordinate encoding works just like TEXS's
- const std::string op_c = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
- coord = "vec3 coords = vec3(" + op_a + ", " + op_c + ", " + op_b + ");";
+ shader.AddLine(
+ "float op_y = " + regs.GetRegisterAsFloat(instr.gpr8.Value() + 1) + ';');
+ shader.AddLine("vec3 coords = vec3(" + op_a + ", op_y, " + op_b + ");");
}
const std::string texture = "textureGather(" + sampler + ", coords, " +
std::to_string(instr.tld4s.component) + ')';
if (!depth_compare) {
- WriteTexsInstruction(instr, coord, texture);
+ WriteTexsInstruction(instr, texture);
} else {
- WriteTexsInstruction(instr, coord, "vec4(" + texture + ')');
+ WriteTexsInstruction(instr, "vec4(" + texture + ')');
}
+
+ --shader.scope;
+ shader.AddLine('}');
break;
}
case OpCode::Id::TXQ: {
UNIMPLEMENTED_IF_MSG(instr.txq.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP),
"NODEP is not implemented");
+ ++shader.scope;
+ shader.AddLine('{');
// TODO: the new commits on the texture refactor, change the way samplers work.
// Sadly, not all texture instructions specify the type of texture their sampler
// uses. This must be fixed at a later instance.
@@ -2892,8 +2919,14 @@ private:
GetSampler(instr.sampler, Tegra::Shader::TextureType::Texture2D, false, false);
switch (instr.txq.query_type) {
case Tegra::Shader::TextureQueryType::Dimension: {
- const std::string texture = "textureQueryLevels(" + sampler + ')';
- regs.SetRegisterToInteger(instr.gpr0, true, 0, texture, 1, 1);
+ const std::string texture = "textureSize(" + sampler + ", " +
+ regs.GetRegisterAsInteger(instr.gpr8) + ')';
+ const std::string mip_level = "textureQueryLevels(" + sampler + ')';
+ shader.AddLine("ivec2 sizes = " + texture + ';');
+ regs.SetRegisterToInteger(instr.gpr0, true, 0, "sizes.x", 1, 1);
+ regs.SetRegisterToInteger(instr.gpr0.Value() + 1, true, 0, "sizes.y", 1, 1);
+ regs.SetRegisterToInteger(instr.gpr0.Value() + 2, true, 0, "0", 1, 1);
+ regs.SetRegisterToInteger(instr.gpr0.Value() + 3, true, 0, mip_level, 1, 1);
break;
}
default: {
@@ -2901,6 +2934,8 @@ private:
static_cast<u32>(instr.txq.query_type.Value()));
}
}
+ --shader.scope;
+ shader.AddLine('}');
break;
}
case OpCode::Id::TMML: {
diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h
index 065b3929c..a8833c06e 100644
--- a/src/video_core/renderer_opengl/maxwell_to_gl.h
+++ b/src/video_core/renderer_opengl/maxwell_to_gl.h
@@ -218,14 +218,19 @@ inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) {
inline GLenum BlendEquation(Maxwell::Blend::Equation equation) {
switch (equation) {
case Maxwell::Blend::Equation::Add:
+ case Maxwell::Blend::Equation::AddGL:
return GL_FUNC_ADD;
case Maxwell::Blend::Equation::Subtract:
+ case Maxwell::Blend::Equation::SubtractGL:
return GL_FUNC_SUBTRACT;
case Maxwell::Blend::Equation::ReverseSubtract:
+ case Maxwell::Blend::Equation::ReverseSubtractGL:
return GL_FUNC_REVERSE_SUBTRACT;
case Maxwell::Blend::Equation::Min:
+ case Maxwell::Blend::Equation::MinGL:
return GL_MIN;
case Maxwell::Blend::Equation::Max:
+ case Maxwell::Blend::Equation::MaxGL:
return GL_MAX;
}
LOG_ERROR(Render_OpenGL, "Unimplemented blend equation={}", static_cast<u32>(equation));
generated by cgit v1.2.3 (git 2.25.1) at 2024-10-30 22:55:16 +0100