summaryrefslogtreecommitdiffstats
path: root/src/video_core
diff options
context:
space:
mode:
Diffstat (limited to 'src/video_core')
-rw-r--r--src/video_core/engines/engine_upload.cpp6
-rw-r--r--src/video_core/engines/engine_upload.h8
-rw-r--r--src/video_core/engines/maxwell_3d.cpp38
-rw-r--r--src/video_core/engines/maxwell_3d.h2
-rw-r--r--src/video_core/gpu_thread.cpp2
-rw-r--r--src/video_core/gpu_thread.h2
-rw-r--r--src/video_core/macro_interpreter.cpp6
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer.cpp8
-rw-r--r--src/video_core/renderer_opengl/gl_shader_gen.cpp14
-rw-r--r--src/video_core/renderer_opengl/maxwell_to_gl.h2
10 files changed, 48 insertions, 40 deletions
diff --git a/src/video_core/engines/engine_upload.cpp b/src/video_core/engines/engine_upload.cpp
index f8aa4ff55..082a40cd9 100644
--- a/src/video_core/engines/engine_upload.cpp
+++ b/src/video_core/engines/engine_upload.cpp
@@ -2,6 +2,8 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
+#include <cstring>
+
#include "common/assert.h"
#include "video_core/engines/engine_upload.h"
#include "video_core/memory_manager.h"
@@ -10,7 +12,9 @@
namespace Tegra::Engines::Upload {
State::State(MemoryManager& memory_manager, Registers& regs)
- : memory_manager(memory_manager), regs(regs) {}
+ : regs{regs}, memory_manager{memory_manager} {}
+
+State::~State() = default;
void State::ProcessExec(const bool is_linear) {
write_offset = 0;
diff --git a/src/video_core/engines/engine_upload.h b/src/video_core/engines/engine_upload.h
index 9c6e0d21c..ef4f5839a 100644
--- a/src/video_core/engines/engine_upload.h
+++ b/src/video_core/engines/engine_upload.h
@@ -4,10 +4,8 @@
#pragma once
-#include <cstddef>
#include <vector>
#include "common/bit_field.h"
-#include "common/common_funcs.h"
#include "common/common_types.h"
namespace Tegra {
@@ -57,10 +55,10 @@ struct Registers {
class State {
public:
State(MemoryManager& memory_manager, Registers& regs);
- ~State() = default;
+ ~State();
- void ProcessExec(const bool is_linear);
- void ProcessData(const u32 data, const bool is_last_call);
+ void ProcessExec(bool is_linear);
+ void ProcessData(u32 data, bool is_last_call);
private:
u32 write_offset = 0;
diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp
index d7b586db9..39968d403 100644
--- a/src/video_core/engines/maxwell_3d.cpp
+++ b/src/video_core/engines/maxwell_3d.cpp
@@ -34,9 +34,9 @@ void Maxwell3D::InitializeRegisterDefaults() {
// Depth range near/far is not always set, but is expected to be the default 0.0f, 1.0f. This is
// needed for ARMS.
- for (std::size_t viewport{}; viewport < Regs::NumViewports; ++viewport) {
- regs.viewports[viewport].depth_range_near = 0.0f;
- regs.viewports[viewport].depth_range_far = 1.0f;
+ for (auto& viewport : regs.viewports) {
+ viewport.depth_range_near = 0.0f;
+ viewport.depth_range_far = 1.0f;
}
// Doom and Bomberman seems to use the uninitialized registers and just enable blend
@@ -47,13 +47,13 @@ void Maxwell3D::InitializeRegisterDefaults() {
regs.blend.equation_a = Regs::Blend::Equation::Add;
regs.blend.factor_source_a = Regs::Blend::Factor::One;
regs.blend.factor_dest_a = Regs::Blend::Factor::Zero;
- for (std::size_t blend_index = 0; blend_index < Regs::NumRenderTargets; blend_index++) {
- regs.independent_blend[blend_index].equation_rgb = Regs::Blend::Equation::Add;
- regs.independent_blend[blend_index].factor_source_rgb = Regs::Blend::Factor::One;
- regs.independent_blend[blend_index].factor_dest_rgb = Regs::Blend::Factor::Zero;
- regs.independent_blend[blend_index].equation_a = Regs::Blend::Equation::Add;
- regs.independent_blend[blend_index].factor_source_a = Regs::Blend::Factor::One;
- regs.independent_blend[blend_index].factor_dest_a = Regs::Blend::Factor::Zero;
+ for (auto& blend : regs.independent_blend) {
+ blend.equation_rgb = Regs::Blend::Equation::Add;
+ blend.factor_source_rgb = Regs::Blend::Factor::One;
+ blend.factor_dest_rgb = Regs::Blend::Factor::Zero;
+ blend.equation_a = Regs::Blend::Equation::Add;
+ blend.factor_source_a = Regs::Blend::Factor::One;
+ blend.factor_dest_a = Regs::Blend::Factor::Zero;
}
regs.stencil_front_op_fail = Regs::StencilOp::Keep;
regs.stencil_front_op_zfail = Regs::StencilOp::Keep;
@@ -75,11 +75,11 @@ void Maxwell3D::InitializeRegisterDefaults() {
// TODO(bunnei): Some games do not initialize the color masks (e.g. Sonic Mania). Assuming a
// default of enabled fixes rendering here.
- for (std::size_t color_mask = 0; color_mask < Regs::NumRenderTargets; color_mask++) {
- regs.color_mask[color_mask].R.Assign(1);
- regs.color_mask[color_mask].G.Assign(1);
- regs.color_mask[color_mask].B.Assign(1);
- regs.color_mask[color_mask].A.Assign(1);
+ for (auto& color_mask : regs.color_mask) {
+ color_mask.R.Assign(1);
+ color_mask.G.Assign(1);
+ color_mask.B.Assign(1);
+ color_mask.A.Assign(1);
}
// Commercial games seem to assume this value is enabled and nouveau sets this value manually.
@@ -178,13 +178,13 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) {
// Vertex buffer
if (method >= MAXWELL3D_REG_INDEX(vertex_array) &&
- method < MAXWELL3D_REG_INDEX(vertex_array) + 4 * 32) {
+ method < MAXWELL3D_REG_INDEX(vertex_array) + 4 * Regs::NumVertexArrays) {
dirty_flags.vertex_array.set((method - MAXWELL3D_REG_INDEX(vertex_array)) >> 2);
} else if (method >= MAXWELL3D_REG_INDEX(vertex_array_limit) &&
- method < MAXWELL3D_REG_INDEX(vertex_array_limit) + 2 * 32) {
+ method < MAXWELL3D_REG_INDEX(vertex_array_limit) + 2 * Regs::NumVertexArrays) {
dirty_flags.vertex_array.set((method - MAXWELL3D_REG_INDEX(vertex_array_limit)) >> 1);
} else if (method >= MAXWELL3D_REG_INDEX(instanced_arrays) &&
- method < MAXWELL3D_REG_INDEX(instanced_arrays) + 32) {
+ method < MAXWELL3D_REG_INDEX(instanced_arrays) + Regs::NumVertexArrays) {
dirty_flags.vertex_array.set(method - MAXWELL3D_REG_INDEX(instanced_arrays));
}
}
@@ -442,7 +442,7 @@ Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const {
const auto a_type = tic_entry.a_type.Value();
// TODO(Subv): Different data types for separate components are not supported
- ASSERT(r_type == g_type && r_type == b_type && r_type == a_type);
+ DEBUG_ASSERT(r_type == g_type && r_type == b_type && r_type == a_type);
return tic_entry;
}
diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h
index 4883b582a..48e4fec33 100644
--- a/src/video_core/engines/maxwell_3d.h
+++ b/src/video_core/engines/maxwell_3d.h
@@ -6,6 +6,7 @@
#include <array>
#include <bitset>
+#include <type_traits>
#include <unordered_map>
#include <vector>
@@ -1107,6 +1108,7 @@ public:
} regs{};
static_assert(sizeof(Regs) == Regs::NUM_REGS * sizeof(u32), "Maxwell3D Regs has wrong size");
+ static_assert(std::is_trivially_copyable_v<Regs>, "Maxwell3D Regs must be trivially copyable");
struct State {
struct ConstBufferInfo {
diff --git a/src/video_core/gpu_thread.cpp b/src/video_core/gpu_thread.cpp
index 03856013f..1e2ff46b0 100644
--- a/src/video_core/gpu_thread.cpp
+++ b/src/video_core/gpu_thread.cpp
@@ -118,7 +118,7 @@ void SynchState::WaitForSynchronization(u64 fence) {
// Wait for the GPU to be idle (all commands to be executed)
{
MICROPROFILE_SCOPE(GPU_wait);
- std::unique_lock<std::mutex> lock{synchronization_mutex};
+ std::unique_lock lock{synchronization_mutex};
synchronization_condition.wait(lock, [this, fence] { return signaled_fence >= fence; });
}
}
diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h
index 64a3335ba..05a168a72 100644
--- a/src/video_core/gpu_thread.h
+++ b/src/video_core/gpu_thread.h
@@ -103,7 +103,7 @@ struct SynchState final {
void TrySynchronize() {
if (IsSynchronized()) {
- std::lock_guard<std::mutex> lock{synchronization_mutex};
+ std::lock_guard lock{synchronization_mutex};
synchronization_condition.notify_one();
}
}
diff --git a/src/video_core/macro_interpreter.cpp b/src/video_core/macro_interpreter.cpp
index 524d9ea5a..fbea107ca 100644
--- a/src/video_core/macro_interpreter.cpp
+++ b/src/video_core/macro_interpreter.cpp
@@ -118,10 +118,10 @@ bool MacroInterpreter::Step(u32 offset, bool is_delay_slot) {
static_cast<u32>(opcode.operation.Value()));
}
- if (opcode.is_exit) {
+ // An instruction with the Exit flag will not actually
+ // cause an exit if it's executed inside a delay slot.
+ if (opcode.is_exit && !is_delay_slot) {
// Exit has a delay slot, execute the next instruction
- // Note: Executing an exit during a branch delay slot will cause the instruction at the
- // branch target to be executed before exiting.
Step(offset, true);
return false;
}
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 3cc945235..dbd8049f5 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -261,8 +261,8 @@ DrawParameters RasterizerOpenGL::SetupDraw() {
// MakeQuadArray always generates u32 indexes
params.index_format = GL_UNSIGNED_INT;
params.count = (regs.vertex_buffer.count / 4) * 6;
- params.index_buffer_offset =
- primitive_assembler.MakeQuadArray(regs.vertex_buffer.first, params.count);
+ params.index_buffer_offset = primitive_assembler.MakeQuadArray(
+ regs.vertex_buffer.first, regs.vertex_buffer.count);
}
return params;
}
@@ -1135,7 +1135,9 @@ void RasterizerOpenGL::SyncTransformFeedback() {
void RasterizerOpenGL::SyncPointState() {
const auto& regs = system.GPU().Maxwell3D().regs;
- state.point.size = regs.point_size;
+ // Limit the point size to 1 since nouveau sometimes sets a point size of 0 (and that's invalid
+ // in OpenGL).
+ state.point.size = std::max(1.0f, regs.point_size);
}
void RasterizerOpenGL::SyncPolygonOffset() {
diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp
index 6abf948f8..7ab0b4553 100644
--- a/src/video_core/renderer_opengl/gl_shader_gen.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp
@@ -33,14 +33,14 @@ layout (std140, binding = EMULATION_UBO_BINDING) uniform vs_config {
};
)";
- ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET);
+ const ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET);
ProgramResult program =
Decompile(device, program_ir, Maxwell3D::Regs::ShaderStage::Vertex, "vertex");
out += program.first;
if (setup.IsDualProgram()) {
- ShaderIR program_ir_b(setup.program.code_b, PROGRAM_OFFSET);
+ const ShaderIR program_ir_b(setup.program.code_b, PROGRAM_OFFSET);
ProgramResult program_b =
Decompile(device, program_ir_b, Maxwell3D::Regs::ShaderStage::Vertex, "vertex_b");
@@ -76,7 +76,7 @@ void main() {
}
})";
- return {out, program.second};
+ return {std::move(out), std::move(program.second)};
}
ProgramResult GenerateGeometryShader(const Device& device, const ShaderSetup& setup) {
@@ -97,7 +97,7 @@ layout (std140, binding = EMULATION_UBO_BINDING) uniform gs_config {
};
)";
- ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET);
+ const ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET);
ProgramResult program =
Decompile(device, program_ir, Maxwell3D::Regs::ShaderStage::Geometry, "geometry");
out += program.first;
@@ -107,7 +107,7 @@ void main() {
execute_geometry();
};)";
- return {out, program.second};
+ return {std::move(out), std::move(program.second)};
}
ProgramResult GenerateFragmentShader(const Device& device, const ShaderSetup& setup) {
@@ -160,7 +160,7 @@ bool AlphaFunc(in float value) {
}
)";
- ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET);
+ const ShaderIR program_ir(setup.program.code, PROGRAM_OFFSET);
ProgramResult program =
Decompile(device, program_ir, Maxwell3D::Regs::ShaderStage::Fragment, "fragment");
@@ -172,7 +172,7 @@ void main() {
}
)";
- return {out, program.second};
+ return {std::move(out), std::move(program.second)};
}
} // namespace OpenGL::GLShader
diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h
index 95b773135..ed7b5cff0 100644
--- a/src/video_core/renderer_opengl/maxwell_to_gl.h
+++ b/src/video_core/renderer_opengl/maxwell_to_gl.h
@@ -126,6 +126,8 @@ inline GLenum PrimitiveTopology(Maxwell::PrimitiveTopology topology) {
return GL_TRIANGLES;
case Maxwell::PrimitiveTopology::TriangleStrip:
return GL_TRIANGLE_STRIP;
+ case Maxwell::PrimitiveTopology::TriangleFan:
+ return GL_TRIANGLE_FAN;
default:
LOG_CRITICAL(Render_OpenGL, "Unimplemented topology={}", static_cast<u32>(topology));
UNREACHABLE();
generated by cgit v1.2.3 (git 2.25.1) at 2024-08-30 03:15:13 +0200