summaryrefslogtreecommitdiff
path: root/src
diff options
context:
space:
mode:
Diffstat (limited to 'src')
-rw-r--r--src/core/hle/service/hid/hid.cpp8
-rw-r--r--src/core/hle/service/service.cpp2
-rw-r--r--src/core/settings.h2
-rw-r--r--src/video_core/engines/maxwell_dma.cpp12
-rw-r--r--src/video_core/engines/shader_bytecode.h41
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer_cache.cpp23
-rw-r--r--src/video_core/renderer_opengl/gl_shader_cache.cpp8
-rw-r--r--src/video_core/renderer_opengl/gl_shader_decompiler.cpp223
-rw-r--r--src/video_core/renderer_opengl/gl_shader_util.cpp2
-rw-r--r--src/video_core/renderer_opengl/gl_stream_buffer.cpp2
10 files changed, 220 insertions, 103 deletions
diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp
index 0d31abe8b..a8e0c869f 100644
--- a/src/core/hle/service/hid/hid.cpp
+++ b/src/core/hle/service/hid/hid.cpp
@@ -2,7 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include <atomic>
#include "common/logging/log.h"
#include "core/core.h"
#include "core/core_timing.h"
@@ -78,7 +77,7 @@ private:
SharedMemory mem{};
std::memcpy(&mem, shared_mem->GetPointer(), sizeof(SharedMemory));
- if (is_device_reload_pending.exchange(false))
+ if (Settings::values.is_device_reload_pending.exchange(false))
LoadInputDevices();
// Set up controllers as neon red+blue Joy-Con attached to console
@@ -267,7 +266,6 @@ private:
CoreTiming::EventType* pad_update_event;
// Stored input state info
- std::atomic<bool> is_device_reload_pending{true};
std::array<std::unique_ptr<Input::ButtonDevice>, Settings::NativeButton::NUM_BUTTONS_HID>
buttons;
std::array<std::unique_ptr<Input::AnalogDevice>, Settings::NativeAnalog::NUM_STICKS_HID> sticks;
@@ -797,7 +795,9 @@ public:
}
};
-void ReloadInputDevices() {}
+void ReloadInputDevices() {
+ Settings::values.is_device_reload_pending.store(true);
+}
void InstallInterfaces(SM::ServiceManager& service_manager) {
std::make_shared<Hid>()->InstallAsService(service_manager);
diff --git a/src/core/hle/service/service.cpp b/src/core/hle/service/service.cpp
index 9d804652e..9bb7c7b26 100644
--- a/src/core/hle/service/service.cpp
+++ b/src/core/hle/service/service.cpp
@@ -74,8 +74,6 @@ using Kernel::SharedPtr;
namespace Service {
-std::unordered_map<std::string, SharedPtr<ClientPort>> g_kernel_named_ports;
-
/**
* Creates a function string for logging, complete with the name (or header code, depending
* on what's passed in) the port name, and all the cmd_buff arguments.
diff --git a/src/core/settings.h b/src/core/settings.h
index 5bf1863e6..08a16ef2c 100644
--- a/src/core/settings.h
+++ b/src/core/settings.h
@@ -5,6 +5,7 @@
#pragma once
#include <array>
+#include <atomic>
#include <string>
#include "common/common_types.h"
@@ -120,6 +121,7 @@ struct Values {
std::array<std::string, NativeAnalog::NumAnalogs> analogs;
std::string motion_device;
std::string touch_device;
+ std::atomic_bool is_device_reload_pending{true};
// Core
bool use_cpu_jit;
diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp
index 6e740713f..c24d33d5c 100644
--- a/src/video_core/engines/maxwell_dma.cpp
+++ b/src/video_core/engines/maxwell_dma.cpp
@@ -41,7 +41,6 @@ void MaxwellDMA::HandleCopy() {
// TODO(Subv): Perform more research and implement all features of this engine.
ASSERT(regs.exec.enable_swizzle == 0);
- ASSERT(regs.exec.enable_2d == 1);
ASSERT(regs.exec.query_mode == Regs::QueryMode::None);
ASSERT(regs.exec.query_intr == Regs::QueryIntr::None);
ASSERT(regs.exec.copy_mode == Regs::CopyMode::Unk2);
@@ -51,10 +50,19 @@ void MaxwellDMA::HandleCopy() {
ASSERT(regs.dst_params.pos_y == 0);
if (regs.exec.is_dst_linear == regs.exec.is_src_linear) {
- Memory::CopyBlock(dest_cpu, source_cpu, regs.x_count * regs.y_count);
+ size_t copy_size = regs.x_count;
+
+ // When the enable_2d bit is disabled, the copy is performed as if we were copying a 1D
+ // buffer of length `x_count`, otherwise we copy a 2D buffer of size (x_count, y_count).
+ if (regs.exec.enable_2d) {
+ copy_size = copy_size * regs.y_count;
+ }
+
+ Memory::CopyBlock(dest_cpu, source_cpu, copy_size);
return;
}
+ ASSERT(regs.exec.enable_2d == 1);
u8* src_buffer = Memory::GetPointer(source_cpu);
u8* dst_buffer = Memory::GetPointer(dest_cpu);
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index d2388673e..9176a8dbc 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -244,6 +244,16 @@ enum class TextureType : u64 {
TextureCube = 3,
};
+enum class TextureQueryType : u64 {
+ Dimension = 1,
+ TextureType = 2,
+ SamplePosition = 5,
+ Filter = 16,
+ LevelOfDetail = 18,
+ Wrap = 20,
+ BorderColor = 22,
+};
+
enum class IpaInterpMode : u64 { Linear = 0, Perspective = 1, Flat = 2, Sc = 3 };
enum class IpaSampleMode : u64 { Default = 0, Centroid = 1, Offset = 2 };
@@ -519,6 +529,21 @@ union Instruction {
} tex;
union {
+ BitField<22, 6, TextureQueryType> query_type;
+ BitField<31, 4, u64> component_mask;
+ } txq;
+
+ union {
+ BitField<28, 1, u64> array;
+ BitField<29, 2, TextureType> texture_type;
+ BitField<31, 4, u64> component_mask;
+
+ bool IsComponentEnabled(size_t component) const {
+ return ((1ull << component) & component_mask) != 0;
+ }
+ } tmml;
+
+ union {
BitField<28, 1, u64> array;
BitField<29, 2, TextureType> texture_type;
BitField<56, 2, u64> component;
@@ -670,11 +695,13 @@ public:
LDG, // Load from global memory
STG, // Store in global memory
TEX,
- TEXQ, // Texture Query
- TEXS, // Texture Fetch with scalar/non-vec4 source/destinations
- TLDS, // Texture Load with scalar/non-vec4 source/destinations
- TLD4, // Texture Load 4
- TLD4S, // Texture Load 4 with scalar / non - vec4 source / destinations
+ TXQ, // Texture Query
+ TEXS, // Texture Fetch with scalar/non-vec4 source/destinations
+ TLDS, // Texture Load with scalar/non-vec4 source/destinations
+ TLD4, // Texture Load 4
+ TLD4S, // Texture Load 4 with scalar / non - vec4 source / destinations
+ TMML_B, // Texture Mip Map Level
+ TMML, // Texture Mip Map Level
EXIT,
IPA,
FFMA_IMM, // Fused Multiply and Add
@@ -894,11 +921,13 @@ private:
INST("1110111011010---", Id::LDG, Type::Memory, "LDG"),
INST("1110111011011---", Id::STG, Type::Memory, "STG"),
INST("110000----111---", Id::TEX, Type::Memory, "TEX"),
- INST("1101111101001---", Id::TEXQ, Type::Memory, "TEXQ"),
+ INST("1101111101001---", Id::TXQ, Type::Memory, "TXQ"),
INST("1101100---------", Id::TEXS, Type::Memory, "TEXS"),
INST("1101101---------", Id::TLDS, Type::Memory, "TLDS"),
INST("110010----111---", Id::TLD4, Type::Memory, "TLD4"),
INST("1101111100------", Id::TLD4S, Type::Memory, "TLD4S"),
+ INST("110111110110----", Id::TMML_B, Type::Memory, "TMML_B"),
+ INST("1101111101011---", Id::TMML, Type::Memory, "TMML"),
INST("111000110000----", Id::EXIT, Type::Trivial, "EXIT"),
INST("11100000--------", Id::IPA, Type::Trivial, "IPA"),
INST("0011001-1-------", Id::FFMA_IMM, Type::Ffma, "FFMA_IMM"),
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
index 360fb0cd5..fa730b9e6 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
@@ -505,7 +505,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height) {
S8Z24 input_pixel{};
Z24S8 output_pixel{};
- const auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::S8Z24)};
+ constexpr auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::S8Z24)};
for (size_t y = 0; y < height; ++y) {
for (size_t x = 0; x < width; ++x) {
const size_t offset{bpp * (y * width + x)};
@@ -518,7 +518,7 @@ static void ConvertS8Z24ToZ24S8(std::vector<u8>& data, u32 width, u32 height) {
}
static void ConvertG8R8ToR8G8(std::vector<u8>& data, u32 width, u32 height) {
- const auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::G8R8U)};
+ constexpr auto bpp{CachedSurface::GetGLBytesPerPixel(PixelFormat::G8R8U)};
for (size_t y = 0; y < height; ++y) {
for (size_t x = 0; x < width; ++x) {
const size_t offset{bpp * (y * width + x)};
@@ -584,12 +584,13 @@ void CachedSurface::LoadGLBuffer() {
UNREACHABLE();
}
- gl_buffer.resize(params.depth * copy_size);
+ gl_buffer.resize(static_cast<size_t>(params.depth) * copy_size);
morton_to_gl_fns[static_cast<size_t>(params.pixel_format)](
params.width, params.block_height, params.height, gl_buffer.data(), copy_size,
params.addr);
} else {
- const u8* const texture_src_data_end{texture_src_data + (params.depth * copy_size)};
+ const u8* const texture_src_data_end{texture_src_data +
+ (static_cast<size_t>(params.depth) * copy_size)};
gl_buffer.assign(texture_src_data, texture_src_data_end);
}
@@ -608,18 +609,20 @@ void CachedSurface::UploadGLTexture(GLuint read_fb_handle, GLuint draw_fb_handle
MICROPROFILE_SCOPE(OpenGL_TextureUL);
- ASSERT(gl_buffer.size() ==
- params.width * params.height * GetGLBytesPerPixel(params.pixel_format) * params.depth);
+ ASSERT(gl_buffer.size() == static_cast<size_t>(params.width) * params.height *
+ GetGLBytesPerPixel(params.pixel_format) * params.depth);
const auto& rect{params.GetRect()};
// Load data from memory to the surface
- GLint x0 = static_cast<GLint>(rect.left);
- GLint y0 = static_cast<GLint>(rect.bottom);
- size_t buffer_offset = (y0 * params.width + x0) * GetGLBytesPerPixel(params.pixel_format);
+ const GLint x0 = static_cast<GLint>(rect.left);
+ const GLint y0 = static_cast<GLint>(rect.bottom);
+ const size_t buffer_offset =
+ static_cast<size_t>(static_cast<size_t>(y0) * params.width + static_cast<size_t>(x0)) *
+ GetGLBytesPerPixel(params.pixel_format);
const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type);
- GLuint target_tex = texture.handle;
+ const GLuint target_tex = texture.handle;
OpenGLState cur_state = OpenGLState::GetCurState();
const auto& old_tex = cur_state.texture_units[0];
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index 7e4b85ac3..61080f5cc 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -13,8 +13,8 @@ namespace OpenGL {
/// Gets the address for the specified shader stage program
static VAddr GetShaderAddress(Maxwell::ShaderProgram program) {
- auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
- auto& shader_config = gpu.regs.shader_config[static_cast<size_t>(program)];
+ const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+ const auto& shader_config = gpu.regs.shader_config[static_cast<size_t>(program)];
return *gpu.memory_manager.GpuToCpuAddress(gpu.regs.code_address.CodeAddress() +
shader_config.offset);
}
@@ -86,7 +86,7 @@ CachedShader::CachedShader(VAddr addr, Maxwell::ShaderProgram program_type)
}
GLuint CachedShader::GetProgramResourceIndex(const GLShader::ConstBufferEntry& buffer) {
- auto search{resource_cache.find(buffer.GetHash())};
+ const auto search{resource_cache.find(buffer.GetHash())};
if (search == resource_cache.end()) {
const GLuint index{
glGetProgramResourceIndex(program.handle, GL_UNIFORM_BLOCK, buffer.GetName().c_str())};
@@ -98,7 +98,7 @@ GLuint CachedShader::GetProgramResourceIndex(const GLShader::ConstBufferEntry& b
}
GLint CachedShader::GetUniformLocation(const GLShader::SamplerEntry& sampler) {
- auto search{uniform_cache.find(sampler.GetHash())};
+ const auto search{uniform_cache.find(sampler.GetHash())};
if (search == uniform_cache.end()) {
const GLint index{glGetUniformLocation(program.handle, sampler.GetName().c_str())};
uniform_cache[sampler.GetHash()] = index;
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
index 172ba8335..762e58aad 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
@@ -113,7 +113,7 @@ private:
/// Scans a range of code for labels and determines the exit method.
ExitMethod Scan(u32 begin, u32 end, std::set<u32>& labels) {
- auto [iter, inserted] =
+ const auto [iter, inserted] =
exit_method_map.emplace(std::make_pair(begin, end), ExitMethod::Undetermined);
ExitMethod& exit_method = iter->second;
if (!inserted)
@@ -131,22 +131,22 @@ private:
if (instr.pred.pred_index == static_cast<u64>(Pred::UnusedIndex)) {
return exit_method = ExitMethod::AlwaysEnd;
} else {
- ExitMethod not_met = Scan(offset + 1, end, labels);
+ const ExitMethod not_met = Scan(offset + 1, end, labels);
return exit_method = ParallelExit(ExitMethod::AlwaysEnd, not_met);
}
}
case OpCode::Id::BRA: {
- u32 target = offset + instr.bra.GetBranchTarget();
+ const u32 target = offset + instr.bra.GetBranchTarget();
labels.insert(target);
- ExitMethod no_jmp = Scan(offset + 1, end, labels);
- ExitMethod jmp = Scan(target, end, labels);
+ 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: {
// The SSY instruction uses a similar encoding as the BRA instruction.
ASSERT_MSG(instr.bra.constant_buffer == 0,
"Constant buffer SSY is not supported");
- u32 target = offset + instr.bra.GetBranchTarget();
+ const u32 target = offset + instr.bra.GetBranchTarget();
labels.insert(target);
// Continue scanning for an exit method.
break;
@@ -346,8 +346,8 @@ public:
*/
void SetRegisterToInputAttibute(const Register& reg, u64 elem, Attribute::Index attribute,
const Tegra::Shader::IpaMode& input_mode) {
- std::string dest = GetRegisterAsFloat(reg);
- std::string src = GetInputAttribute(attribute, input_mode) + GetSwizzle(elem);
+ const std::string dest = GetRegisterAsFloat(reg);
+ const std::string src = GetInputAttribute(attribute, input_mode) + GetSwizzle(elem);
shader.AddLine(dest + " = " + src + ';');
}
@@ -359,8 +359,8 @@ public:
* @param reg The register to use as the source value.
*/
void SetOutputAttributeToRegister(Attribute::Index attribute, u64 elem, const Register& reg) {
- std::string dest = GetOutputAttribute(attribute);
- std::string src = GetRegisterAsFloat(reg);
+ const std::string dest = GetOutputAttribute(attribute);
+ const std::string src = GetRegisterAsFloat(reg);
if (!dest.empty()) {
// Can happen with unknown/unimplemented output attributes, in which case we ignore the
@@ -393,9 +393,9 @@ public:
GLSLRegister::Type type) {
declr_const_buffers[cbuf_index].MarkAsUsedIndirect(cbuf_index, stage);
- std::string final_offset = fmt::format("({} + {})", index_str, offset / 4);
- std::string value = 'c' + std::to_string(cbuf_index) + '[' + final_offset + " / 4][" +
- final_offset + " % 4]";
+ const std::string final_offset = fmt::format("({} + {})", index_str, offset / 4);
+ const std::string value = 'c' + std::to_string(cbuf_index) + '[' + final_offset + " / 4][" +
+ final_offset + " % 4]";
if (type == GLSLRegister::Type::Float) {
return value;
@@ -468,10 +468,10 @@ public:
/// necessary.
std::string AccessSampler(const Sampler& sampler, Tegra::Shader::TextureType type,
bool is_array) {
- size_t offset = static_cast<size_t>(sampler.index.Value());
+ const size_t offset = static_cast<size_t>(sampler.index.Value());
// If this sampler has already been used, return the existing mapping.
- auto itr =
+ const auto itr =
std::find_if(used_samplers.begin(), used_samplers.end(),
[&](const SamplerEntry& entry) { return entry.GetOffset() == offset; });
@@ -481,8 +481,8 @@ public:
}
// Otherwise create a new mapping for this sampler
- size_t next_index = used_samplers.size();
- SamplerEntry entry{stage, offset, next_index, type, is_array};
+ const size_t next_index = used_samplers.size();
+ const SamplerEntry entry{stage, offset, next_index, type, is_array};
used_samplers.emplace_back(entry);
return entry.GetName();
}
@@ -699,7 +699,7 @@ private:
};
bool IsColorComponentOutputEnabled(u32 render_target, u32 component) const {
- u32 bit = render_target * 4 + component;
+ const u32 bit = render_target * 4 + component;
return enabled_color_outputs & (1 << bit);
}
};
@@ -707,7 +707,7 @@ private:
/// Gets the Subroutine object corresponding to the specified address.
const Subroutine& GetSubroutine(u32 begin, u32 end) const {
- auto iter = subroutines.find(Subroutine{begin, end, suffix});
+ const auto iter = subroutines.find(Subroutine{begin, end, suffix});
ASSERT(iter != subroutines.end());
return *iter;
}
@@ -752,7 +752,7 @@ private:
// Can't assign to the constant predicate.
ASSERT(pred != static_cast<u64>(Pred::UnusedIndex));
- std::string variable = 'p' + std::to_string(pred) + '_' + suffix;
+ const std::string variable = 'p' + std::to_string(pred) + '_' + suffix;
shader.AddLine(variable + " = " + value + ';');
declr_predicates.insert(std::move(variable));
}
@@ -1033,7 +1033,11 @@ private:
if (header.writes_depth) {
// The depth output is always 2 registers after the last color output, and current_reg
// already contains one past the last color register.
- shader.AddLine("gl_FragDepth = " + regs.GetRegisterAsFloat(current_reg + 1) + ';');
+
+ shader.AddLine(
+ "gl_FragDepth = " +
+ regs.GetRegisterAsFloat(static_cast<Tegra::Shader::Register>(current_reg) + 1) +
+ ';');
}
}
@@ -1435,7 +1439,7 @@ private:
if (instr.alu_integer.negate_b)
op_b = "-(" + op_b + ')';
- std::string shift = std::to_string(instr.alu_integer.shift_amount.Value());
+ const std::string shift = std::to_string(instr.alu_integer.shift_amount.Value());
regs.SetRegisterToInteger(instr.gpr0, true, 0,
"((" + op_a + " << " + shift + ") + " + op_b + ')', 1, 1);
@@ -1453,7 +1457,7 @@ private:
case OpCode::Id::SEL_C:
case OpCode::Id::SEL_R:
case OpCode::Id::SEL_IMM: {
- std::string condition =
+ const std::string condition =
GetPredicateCondition(instr.sel.pred, instr.sel.neg_pred != 0);
regs.SetRegisterToInteger(instr.gpr0, true, 0,
'(' + condition + ") ? " + op_a + " : " + op_b, 1, 1);
@@ -1475,8 +1479,9 @@ private:
case OpCode::Id::LOP3_C:
case OpCode::Id::LOP3_R:
case OpCode::Id::LOP3_IMM: {
- std::string op_c = regs.GetRegisterAsInteger(instr.gpr39);
+ const std::string op_c = regs.GetRegisterAsInteger(instr.gpr39);
std::string lut;
+
if (opcode->GetId() == OpCode::Id::LOP3_R) {
lut = '(' + std::to_string(instr.alu.lop3.GetImmLut28()) + ')';
} else {
@@ -1491,9 +1496,9 @@ private:
case OpCode::Id::IMNMX_IMM: {
ASSERT_MSG(instr.imnmx.exchange == Tegra::Shader::IMinMaxExchange::None,
"Unimplemented");
- std::string condition =
+ const std::string condition =
GetPredicateCondition(instr.imnmx.pred, instr.imnmx.negate_pred != 0);
- std::string parameters = op_a + ',' + op_b;
+ const std::string parameters = op_a + ',' + op_b;
regs.SetRegisterToInteger(instr.gpr0, instr.imnmx.is_signed, 0,
'(' + condition + ") ? min(" + parameters + ") : max(" +
parameters + ')',
@@ -1510,7 +1515,7 @@ private:
break;
}
case OpCode::Type::Ffma: {
- std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
+ const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
std::string op_b = instr.ffma.negate_b ? "-" : "";
std::string op_c = instr.ffma.negate_c ? "-" : "";
@@ -1720,7 +1725,7 @@ private:
shader.AddLine("uint index = (" + regs.GetRegisterAsInteger(instr.gpr8, 0, false) +
" / 4) & (MAX_CONSTBUFFER_ELEMENTS - 1);");
- std::string op_a =
+ const std::string op_a =
regs.GetUniformIndirect(instr.cbuf36.index, instr.cbuf36.offset + 0, "index",
GLSLRegister::Type::Float);
@@ -1730,7 +1735,7 @@ private:
break;
case Tegra::Shader::UniformType::Double: {
- std::string op_b =
+ const std::string op_b =
regs.GetUniformIndirect(instr.cbuf36.index, instr.cbuf36.offset + 4,
"index", GLSLRegister::Type::Float);
regs.SetRegisterToFloat(instr.gpr0, 0, op_a, 1, 1);
@@ -1760,13 +1765,13 @@ private:
switch (texture_type) {
case Tegra::Shader::TextureType::Texture1D: {
- std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+ const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
coord = "float coords = " + x + ';';
break;
}
case Tegra::Shader::TextureType::Texture2D: {
- std::string x = regs.GetRegisterAsFloat(instr.gpr8);
- std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+ const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+ const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
coord = "vec2 coords = vec2(" + x + ", " + y + ");";
break;
}
@@ -1776,8 +1781,8 @@ private:
UNREACHABLE();
// Fallback to interpreting as a 2D texture for now
- std::string x = regs.GetRegisterAsFloat(instr.gpr8);
- std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+ const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+ const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
coord = "vec2 coords = vec2(" + x + ", " + y + ");";
texture_type = Tegra::Shader::TextureType::Texture2D;
}
@@ -1811,13 +1816,13 @@ private:
switch (texture_type) {
case Tegra::Shader::TextureType::Texture2D: {
if (is_array) {
- std::string index = regs.GetRegisterAsInteger(instr.gpr8);
- std::string x = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
- std::string y = regs.GetRegisterAsFloat(instr.gpr20);
+ 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 + ");";
} else {
- std::string x = regs.GetRegisterAsFloat(instr.gpr8);
- std::string y = regs.GetRegisterAsFloat(instr.gpr20);
+ const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+ const std::string y = regs.GetRegisterAsFloat(instr.gpr20);
coord = "vec2 coords = vec2(" + x + ", " + y + ");";
}
break;
@@ -1828,8 +1833,8 @@ private:
UNREACHABLE();
// Fallback to interpreting as a 2D texture for now
- std::string x = regs.GetRegisterAsFloat(instr.gpr8);
- std::string y = regs.GetRegisterAsFloat(instr.gpr20);
+ const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+ const std::string y = regs.GetRegisterAsFloat(instr.gpr20);
coord = "vec2 coords = vec2(" + x + ", " + y + ");";
texture_type = Tegra::Shader::TextureType::Texture2D;
is_array = false;
@@ -1850,8 +1855,8 @@ private:
LOG_CRITICAL(HW_GPU, "Unhandled 2d array texture");
UNREACHABLE();
} else {
- std::string x = regs.GetRegisterAsInteger(instr.gpr8);
- std::string y = regs.GetRegisterAsInteger(instr.gpr20);
+ const std::string x = regs.GetRegisterAsInteger(instr.gpr8);
+ const std::string y = regs.GetRegisterAsInteger(instr.gpr20);
coord = "ivec2 coords = ivec2(" + x + ", " + y + ");";
}
break;
@@ -1874,8 +1879,8 @@ private:
switch (instr.tld4.texture_type) {
case Tegra::Shader::TextureType::Texture2D: {
- std::string x = regs.GetRegisterAsFloat(instr.gpr8);
- std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+ const std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+ const std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
coord = "vec2 coords = vec2(" + x + ", " + y + ");";
break;
}
@@ -1920,6 +1925,74 @@ private:
WriteTexsInstruction(instr, coord, texture);
break;
}
+ case OpCode::Id::TXQ: {
+ // 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.
+ const std::string sampler =
+ GetSampler(instr.sampler, Tegra::Shader::TextureType::Texture2D, 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);
+ break;
+ }
+ default: {
+ LOG_CRITICAL(HW_GPU, "Unhandled texture query type: {}",
+ static_cast<u32>(instr.txq.query_type.Value()));
+ UNREACHABLE();
+ }
+ }
+ break;
+ }
+ case OpCode::Id::TMML: {
+ const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
+ const std::string op_b = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+ const bool is_array = instr.tmml.array != 0;
+ auto texture_type = instr.tmml.texture_type.Value();
+ const std::string sampler = GetSampler(instr.sampler, texture_type, is_array);
+
+ // TODO: add coordinates for different samplers once other texture types are
+ // implemented.
+ std::string coord;
+ switch (texture_type) {
+ case Tegra::Shader::TextureType::Texture1D: {
+ std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+ coord = "float coords = " + x + ';';
+ break;
+ }
+ case Tegra::Shader::TextureType::Texture2D: {
+ std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+ std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+ coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+ break;
+ }
+ default:
+ LOG_CRITICAL(HW_GPU, "Unhandled texture type {}",
+ static_cast<u32>(texture_type));
+ UNREACHABLE();
+
+ // Fallback to interpreting as a 2D texture for now
+ std::string x = regs.GetRegisterAsFloat(instr.gpr8);
+ std::string y = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1);
+ coord = "vec2 coords = vec2(" + x + ", " + y + ");";
+ texture_type = Tegra::Shader::TextureType::Texture2D;
+ }
+ // 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 = "textureQueryLod(" + sampler + ", coords)";
+ const std::string tmp = "vec2 tmp = " + texture + "*vec2(256.0, 256.0);";
+ shader.AddLine(tmp);
+
+ regs.SetRegisterToInteger(instr.gpr0, true, 0, "int(tmp.y)", 1, 1);
+ regs.SetRegisterToInteger(instr.gpr0.Value() + 1, false, 0, "uint(tmp.x)", 1, 1);
+ --shader.scope;
+ shader.AddLine('}');
+ break;
+ }
default: {
LOG_CRITICAL(HW_GPU, "Unhandled memory instruction: {}", opcode->GetName());
UNREACHABLE();
@@ -1959,12 +2032,12 @@ private:
// We can't use the constant predicate as destination.
ASSERT(instr.fsetp.pred3 != static_cast<u64>(Pred::UnusedIndex));
- std::string second_pred =
+ const std::string second_pred =
GetPredicateCondition(instr.fsetp.pred39, instr.fsetp.neg_pred != 0);
- std::string combiner = GetPredicateCombiner(instr.fsetp.op);
+ const std::string combiner = GetPredicateCombiner(instr.fsetp.op);
- std::string predicate = GetPredicateComparison(instr.fsetp.cond, op_a, op_b);
+ const std::string predicate = GetPredicateComparison(instr.fsetp.cond, op_a, op_b);
// Set the primary predicate to the result of Predicate OP SecondPredicate
SetPredicate(instr.fsetp.pred3,
'(' + predicate + ") " + combiner + " (" + second_pred + ')');
@@ -1978,7 +2051,8 @@ private:
break;
}
case OpCode::Type::IntegerSetPredicate: {
- std::string op_a = regs.GetRegisterAsInteger(instr.gpr8, 0, instr.isetp.is_signed);
+ const std::string op_a =
+ regs.GetRegisterAsInteger(instr.gpr8, 0, instr.isetp.is_signed);
std::string op_b;
if (instr.is_b_imm) {
@@ -1995,12 +2069,12 @@ private:
// We can't use the constant predicate as destination.
ASSERT(instr.isetp.pred3 != static_cast<u64>(Pred::UnusedIndex));
- std::string second_pred =
+ const std::string second_pred =
GetPredicateCondition(instr.isetp.pred39, instr.isetp.neg_pred != 0);
- std::string combiner = GetPredicateCombiner(instr.isetp.op);
+ const std::string combiner = GetPredicateCombiner(instr.isetp.op);
- std::string predicate = GetPredicateComparison(instr.isetp.cond, op_a, op_b);
+ const std::string predicate = GetPredicateComparison(instr.isetp.cond, op_a, op_b);
// Set the primary predicate to the result of Predicate OP SecondPredicate
SetPredicate(instr.isetp.pred3,
'(' + predicate + ") " + combiner + " (" + second_pred + ')');
@@ -2014,20 +2088,20 @@ private:
break;
}
case OpCode::Type::PredicateSetPredicate: {
- std::string op_a =
+ const std::string op_a =
GetPredicateCondition(instr.psetp.pred12, instr.psetp.neg_pred12 != 0);
- std::string op_b =
+ const std::string op_b =
GetPredicateCondition(instr.psetp.pred29, instr.psetp.neg_pred29 != 0);
// We can't use the constant predicate as destination.
ASSERT(instr.psetp.pred3 != static_cast<u64>(Pred::UnusedIndex));
- std::string second_pred =
+ const std::string second_pred =
GetPredicateCondition(instr.psetp.pred39, instr.psetp.neg_pred39 != 0);
- std::string combiner = GetPredicateCombiner(instr.psetp.op);
+ const std::string combiner = GetPredicateCombiner(instr.psetp.op);
- std::string predicate =
+ const std::string predicate =
'(' + op_a + ") " + GetPredicateCombiner(instr.psetp.cond) + " (" + op_b + ')';
// Set the primary predicate to the result of Predicate OP SecondPredicate
@@ -2053,7 +2127,7 @@ private:
std::string op_b = instr.fset.neg_b ? "-" : "";
if (instr.is_b_imm) {
- std::string imm = GetImmediate19(instr);
+ const std::string imm = GetImmediate19(instr);
if (instr.fset.neg_imm)
op_b += "(-" + imm + ')';
else
@@ -2073,13 +2147,14 @@ private:
// The fset instruction sets a register to 1.0 or -1 (depending on the bf bit) if the
// condition is true, and to 0 otherwise.
- std::string second_pred =
+ const std::string second_pred =
GetPredicateCondition(instr.fset.pred39, instr.fset.neg_pred != 0);
- std::string combiner = GetPredicateCombiner(instr.fset.op);
+ const std::string combiner = GetPredicateCombiner(instr.fset.op);
- std::string predicate = "((" + GetPredicateComparison(instr.fset.cond, op_a, op_b) +
- ") " + combiner + " (" + second_pred + "))";
+ const std::string predicate = "((" +
+ GetPredicateComparison(instr.fset.cond, op_a, op_b) +
+ ") " + combiner + " (" + second_pred + "))";
if (instr.fset.bf) {
regs.SetRegisterToFloat(instr.gpr0, 0, predicate + " ? 1.0 : 0.0", 1, 1);
@@ -2090,7 +2165,7 @@ private:
break;
}
case OpCode::Type::IntegerSet: {
- std::string op_a = regs.GetRegisterAsInteger(instr.gpr8, 0, instr.iset.is_signed);
+ const std::string op_a = regs.GetRegisterAsInteger(instr.gpr8, 0, instr.iset.is_signed);
std::string op_b;
@@ -2107,13 +2182,14 @@ private:
// The iset instruction sets a register to 1.0 or -1 (depending on the bf bit) if the
// condition is true, and to 0 otherwise.
- std::string second_pred =
+ const std::string second_pred =
GetPredicateCondition(instr.iset.pred39, instr.iset.neg_pred != 0);
- std::string combiner = GetPredicateCombiner(instr.iset.op);
+ const std::string combiner = GetPredicateCombiner(instr.iset.op);
- std::string predicate = "((" + GetPredicateComparison(instr.iset.cond, op_a, op_b) +
- ") " + combiner + " (" + second_pred + "))";
+ const std::string predicate = "((" +
+ GetPredicateComparison(instr.iset.cond, op_a, op_b) +
+ ") " + combiner + " (" + second_pred + "))";
if (instr.iset.bf) {
regs.SetRegisterToFloat(instr.gpr0, 0, predicate + " ? 1.0 : 0.0", 1, 1);
@@ -2263,7 +2339,7 @@ private:
case OpCode::Id::BRA: {
ASSERT_MSG(instr.bra.constant_buffer == 0,
"BRA with constant buffers are not implemented");
- u32 target = offset + instr.bra.GetBranchTarget();
+ const u32 target = offset + instr.bra.GetBranchTarget();
shader.AddLine("{ jmp_to = " + std::to_string(target) + "u; break; }");
break;
}
@@ -2287,7 +2363,7 @@ private:
// has a similar structure to the BRA opcode.
ASSERT_MSG(instr.bra.constant_buffer == 0, "Constant buffer SSY is not supported");
- u32 target = offset + instr.bra.GetBranchTarget();
+ const u32 target = offset + instr.bra.GetBranchTarget();
EmitPushToSSYStack(target);
break;
}
@@ -2381,10 +2457,10 @@ private:
shader.AddLine("case " + std::to_string(label) + "u: {");
++shader.scope;
- auto next_it = labels.lower_bound(label + 1);
- u32 next_label = next_it == labels.end() ? subroutine.end : *next_it;
+ const auto next_it = labels.lower_bound(label + 1);
+ const u32 next_label = next_it == labels.end() ? subroutine.end : *next_it;
- u32 compile_end = CompileRange(label, next_label);
+ const u32 compile_end = CompileRange(label, next_label);
if (compile_end > next_label && compile_end != PROGRAM_END) {
// This happens only when there is a label inside a IF/LOOP block
shader.AddLine(" jmp_to = " + std::to_string(compile_end) + "u; break; }");
@@ -2447,7 +2523,8 @@ boost::optional<ProgramResult> DecompileProgram(const ProgramCode& program_code,
Maxwell3D::Regs::ShaderStage stage,
const std::string& suffix) {
try {
- auto subroutines = ControlFlowAnalyzer(program_code, main_offset, suffix).GetSubroutines();
+ const auto subroutines =
+ ControlFlowAnalyzer(program_code, main_offset, suffix).GetSubroutines();
GLSLGenerator generator(subroutines, program_code, main_offset, stage, suffix);
return ProgramResult{generator.GetShaderCode(), generator.GetEntries()};
} catch (const DecompileFail& exception) {
diff --git a/src/video_core/renderer_opengl/gl_shader_util.cpp b/src/video_core/renderer_opengl/gl_shader_util.cpp
index 5781d9d16..5f3fe067e 100644
--- a/src/video_core/renderer_opengl/gl_shader_util.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_util.cpp
@@ -25,7 +25,7 @@ GLuint LoadShader(const char* source, GLenum type) {
default:
UNREACHABLE();
}
- GLuint shader_id = glCreateShader(type);
+ const GLuint shader_id = glCreateShader(type);
glShaderSource(shader_id, 1, &source, nullptr);
LOG_DEBUG(Render_OpenGL, "Compiling {} shader...", debug_type);
glCompileShader(shader_id);
diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.cpp b/src/video_core/renderer_opengl/gl_stream_buffer.cpp
index e565afcee..aadf68f16 100644
--- a/src/video_core/renderer_opengl/gl_stream_buffer.cpp
+++ b/src/video_core/renderer_opengl/gl_stream_buffer.cpp
@@ -29,7 +29,7 @@ OGLStreamBuffer::OGLStreamBuffer(GLenum target, GLsizeiptr size, bool prefer_coh
if (GLAD_GL_ARB_buffer_storage) {
persistent = true;
coherent = prefer_coherent;
- GLbitfield flags =
+ const GLbitfield flags =
GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT | (coherent ? GL_MAP_COHERENT_BIT : 0);
glBufferStorage(gl_target, allocate_size, nullptr, flags);
mapped_ptr = static_cast<u8*>(glMapBufferRange(
generated by cgit v1.2.3 (git 2.25.1) at 2025-04-26 20:35:53 +0000