From db309b2423a996cb792273080e73906b07f8b45b Mon Sep 17 00:00:00 2001 From: wwylele Date: Mon, 24 Jul 2017 14:13:33 +0300 Subject: [PATCH 01/31] pica/regs: layout geometry shader configuration regs All the register meanings are derived from ctrulib (3dbrew is outdated for most of them) --- src/video_core/regs_pipeline.h | 34 ++++++++++++++++++++++++++++++++-- src/video_core/regs_shader.h | 7 +++++++ 2 files changed, 39 insertions(+), 2 deletions(-) diff --git a/src/video_core/regs_pipeline.h b/src/video_core/regs_pipeline.h index 8b6369297..e78c3e331 100644 --- a/src/video_core/regs_pipeline.h +++ b/src/video_core/regs_pipeline.h @@ -147,7 +147,15 @@ struct PipelineRegs { // Number of vertices to render u32 num_vertices; - INSERT_PADDING_WORDS(0x1); + enum class UseGS : u32 { + No = 0, + Yes = 2, + }; + + union { + BitField<0, 2, UseGS> use_gs; + BitField<31, 1, u32> variable_primitive; + }; // The index of the first vertex to render u32 vertex_offset; @@ -218,7 +226,29 @@ struct PipelineRegs { GPUMode gpu_mode; - INSERT_PADDING_WORDS(0x18); + INSERT_PADDING_WORDS(0x4); + BitField<0, 4, u32> vs_outmap_total_minus_1_a; + INSERT_PADDING_WORDS(0x6); + BitField<0, 4, u32> vs_outmap_total_minus_1_b; + + enum class GSMode : u32 { + Point = 0, + VariablePrimitive = 1, + FixedPrimitive = 2, + }; + + union { + BitField<0, 8, GSMode> mode; + BitField<8, 4, u32> fixed_vertex_num_minus_1; + BitField<12, 4, u32> stride_minus_1; + BitField<16, 4, u32> start_index; + } gs_config; + + INSERT_PADDING_WORDS(0x1); + + u32 variable_vertex_main_num_minus_1; + + INSERT_PADDING_WORDS(0x9); enum class TriangleTopology : u32 { List = 0, diff --git a/src/video_core/regs_shader.h b/src/video_core/regs_shader.h index ddb1ee451..c15d4d162 100644 --- a/src/video_core/regs_shader.h +++ b/src/video_core/regs_shader.h @@ -24,9 +24,16 @@ struct ShaderRegs { INSERT_PADDING_WORDS(0x4); + enum ShaderMode { + GS = 0x08, + VS = 0xA0, + }; + union { // Number of input attributes to shader unit - 1 BitField<0, 4, u32> max_input_attribute_index; + BitField<8, 8, u32> input_to_uniform; + BitField<24, 8, ShaderMode> shader_mode; }; // Offset to shader program entry point (in words) From 46c6973d2bde25a2a8ae9ac434660798fd1dfaee Mon Sep 17 00:00:00 2001 From: wwylele Date: Tue, 25 Jul 2017 22:30:29 +0300 Subject: [PATCH 02/31] pica/shader: extend UnitState for GS Among four shader units in pica, a special unit can be configured to run both VS and GS program. GSUnitState represents this unit, which extends UnitState (which represents the other three normal units) with extra state for primitive emitting. It uses lots of raw pointers to represent internal structure in order to keep it standard layout type for JIT to access. This unit doesn't handle triangle winding (inverting) itself; instead, it calls a WindingSetter handler. This will be explained in the following commits --- src/video_core/shader/shader.cpp | 38 ++++++++++++++++++++++++++ src/video_core/shader/shader.h | 46 ++++++++++++++++++++++++++++++++ 2 files changed, 84 insertions(+) diff --git a/src/video_core/shader/shader.cpp b/src/video_core/shader/shader.cpp index 67ed19ba8..b12468d3a 100644 --- a/src/video_core/shader/shader.cpp +++ b/src/video_core/shader/shader.cpp @@ -82,6 +82,44 @@ void UnitState::WriteOutput(const ShaderRegs& config, AttributeBuffer& output) { } } +UnitState::UnitState(GSEmitter* emitter) : emitter_ptr(emitter) {} + +GSEmitter::GSEmitter() { + handlers = new Handlers; +} + +GSEmitter::~GSEmitter() { + delete handlers; +} + +void GSEmitter::Emit(Math::Vec4 (&vertex)[16]) { + ASSERT(vertex_id < 3); + std::copy(std::begin(vertex), std::end(vertex), buffer[vertex_id].begin()); + if (prim_emit) { + if (winding) + handlers->winding_setter(); + for (size_t i = 0; i < buffer.size(); ++i) { + AttributeBuffer output; + unsigned int output_i = 0; + for (unsigned int reg : Common::BitSet(output_mask)) { + output.attr[output_i++] = buffer[i][reg]; + } + handlers->vertex_handler(output); + } + } +} + +GSUnitState::GSUnitState() : UnitState(&emitter) {} + +void GSUnitState::SetVertexHandler(VertexHandler vertex_handler, WindingSetter winding_setter) { + emitter.handlers->vertex_handler = std::move(vertex_handler); + emitter.handlers->winding_setter = std::move(winding_setter); +} + +void GSUnitState::ConfigOutput(const ShaderRegs& config) { + emitter.output_mask = config.output_mask; +} + MICROPROFILE_DEFINE(GPU_Shader, "GPU", "Shader", MP_RGB(50, 50, 240)); #ifdef ARCHITECTURE_x86_64 diff --git a/src/video_core/shader/shader.h b/src/video_core/shader/shader.h index e156f6aef..caec96043 100644 --- a/src/video_core/shader/shader.h +++ b/src/video_core/shader/shader.h @@ -6,6 +6,7 @@ #include #include +#include #include #include #include "common/assert.h" @@ -31,6 +32,12 @@ struct AttributeBuffer { alignas(16) Math::Vec4 attr[16]; }; +/// Handler type for receiving vertex outputs from vertex shader or geometry shader +using VertexHandler = std::function; + +/// Handler type for signaling to invert the vertex order of the next triangle +using WindingSetter = std::function; + struct OutputVertex { Math::Vec4 pos; Math::Vec4 quat; @@ -60,6 +67,29 @@ ASSERT_POS(tc2, RasterizerRegs::VSOutputAttributes::TEXCOORD2_U); static_assert(std::is_pod::value, "Structure is not POD"); static_assert(sizeof(OutputVertex) == 24 * sizeof(float), "OutputVertex has invalid size"); +/** + * This structure contains state information for primitive emitting in geometry shader. + */ +struct GSEmitter { + std::array, 16>, 3> buffer; + u8 vertex_id; + bool prim_emit; + bool winding; + u32 output_mask; + + // Function objects are hidden behind a raw pointer to make the structure standard layout type, + // for JIT to use offsetof to access other members. + struct Handlers { + VertexHandler vertex_handler; + WindingSetter winding_setter; + } * handlers; + + GSEmitter(); + ~GSEmitter(); + void Emit(Math::Vec4 (&vertex)[16]); +}; +static_assert(std::is_standard_layout::value, "GSEmitter is not standard layout type"); + /** * This structure contains the state information that needs to be unique for a shader unit. The 3DS * has four shader units that process shaders in parallel. At the present, Citra only implements a @@ -67,6 +97,7 @@ static_assert(sizeof(OutputVertex) == 24 * sizeof(float), "OutputVertex has inva * here will make it easier for us to parallelize the shader processing later. */ struct UnitState { + explicit UnitState(GSEmitter* emitter = nullptr); struct Registers { // The registers are accessed by the shader JIT using SSE instructions, and are therefore // required to be 16-byte aligned. @@ -82,6 +113,8 @@ struct UnitState { // TODO: How many bits do these actually have? s32 address_registers[3]; + GSEmitter* emitter_ptr; + static size_t InputOffset(const SourceRegister& reg) { switch (reg.GetRegisterType()) { case RegisterType::Input: @@ -125,6 +158,19 @@ struct UnitState { void WriteOutput(const ShaderRegs& config, AttributeBuffer& output); }; +/** + * This is an extended shader unit state that represents the special unit that can run both vertex + * shader and geometry shader. It contains an additional primitive emitter and utilities for + * geometry shader. + */ +struct GSUnitState : public UnitState { + GSUnitState(); + void SetVertexHandler(VertexHandler vertex_handler, WindingSetter winding_setter); + void ConfigOutput(const ShaderRegs& config); + + GSEmitter emitter; +}; + struct ShaderSetup { struct { // The float uniforms are accessed by the shader JIT using SSE instructions, and are From 28128348f21d83c30979ef10399a8a764bb08a73 Mon Sep 17 00:00:00 2001 From: wwylele Date: Tue, 25 Jul 2017 22:43:25 +0300 Subject: [PATCH 03/31] pica/shader/interpreter: implement SETEMIT and EMIT --- src/video_core/shader/shader_interpreter.cpp | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/src/video_core/shader/shader_interpreter.cpp b/src/video_core/shader/shader_interpreter.cpp index 206c0978a..9d4da4904 100644 --- a/src/video_core/shader/shader_interpreter.cpp +++ b/src/video_core/shader/shader_interpreter.cpp @@ -636,6 +636,22 @@ static void RunInterpreter(const ShaderSetup& setup, UnitState& state, DebugData break; } + case OpCode::Id::EMIT: { + GSEmitter* emitter = state.emitter_ptr; + ASSERT_MSG(emitter, "Execute EMIT on VS"); + emitter->Emit(state.registers.output); + break; + } + + case OpCode::Id::SETEMIT: { + GSEmitter* emitter = state.emitter_ptr; + ASSERT_MSG(emitter, "Execute SETEMIT on VS"); + emitter->vertex_id = instr.setemit.vertex_id; + emitter->prim_emit = instr.setemit.prim_emit != 0; + emitter->winding = instr.setemit.winding != 0; + break; + } + default: LOG_ERROR(HW_GPU, "Unhandled instruction: 0x%02x (%s): 0x%08x", (int)instr.opcode.Value().EffectiveOpCode(), From bb63ae305279d9a73ea70133c89e92a36dc79f69 Mon Sep 17 00:00:00 2001 From: wwylele Date: Wed, 26 Jul 2017 00:39:43 +0300 Subject: [PATCH 04/31] correct constness --- src/video_core/shader/shader.cpp | 3 ++- src/video_core/shader/shader.h | 3 ++- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/video_core/shader/shader.cpp b/src/video_core/shader/shader.cpp index b12468d3a..e9063e616 100644 --- a/src/video_core/shader/shader.cpp +++ b/src/video_core/shader/shader.cpp @@ -21,7 +21,8 @@ namespace Pica { namespace Shader { -OutputVertex OutputVertex::FromAttributeBuffer(const RasterizerRegs& regs, AttributeBuffer& input) { +OutputVertex OutputVertex::FromAttributeBuffer(const RasterizerRegs& regs, + const AttributeBuffer& input) { // Setup output data union { OutputVertex ret{}; diff --git a/src/video_core/shader/shader.h b/src/video_core/shader/shader.h index caec96043..a3789da01 100644 --- a/src/video_core/shader/shader.h +++ b/src/video_core/shader/shader.h @@ -50,7 +50,8 @@ struct OutputVertex { INSERT_PADDING_WORDS(1); Math::Vec2 tc2; - static OutputVertex FromAttributeBuffer(const RasterizerRegs& regs, AttributeBuffer& output); + static OutputVertex FromAttributeBuffer(const RasterizerRegs& regs, + const AttributeBuffer& output); }; #define ASSERT_POS(var, pos) \ static_assert(offsetof(OutputVertex, var) == pos * sizeof(float24), "Semantic at wrong " \ From 36981a5aa6ffcc10417e533ab00de3b6f9bad067 Mon Sep 17 00:00:00 2001 From: wwylele Date: Wed, 26 Jul 2017 15:07:13 +0300 Subject: [PATCH 05/31] pica/primitive_assembly: Handle winding for GS primitive hwtest shows that, although GS always emit a group of three vertices as one primitive, it still respects to the topology type, as if the three vertices are input into the primitive assembler independently and sequentially. It is also shown that the winding flag in SETEMIT only takes effect for Shader topology type, which is believed to be the actual difference between List and Shader (hence removed the TODO). However, only Shader topology type is observed in official games when GS is in use, so the other mode seems to be just unintended usage. --- src/video_core/primitive_assembly.cpp | 15 ++++++++++++--- src/video_core/primitive_assembly.h | 7 +++++++ 2 files changed, 19 insertions(+), 3 deletions(-) diff --git a/src/video_core/primitive_assembly.cpp b/src/video_core/primitive_assembly.cpp index acd2ac5e2..9c3dd4cab 100644 --- a/src/video_core/primitive_assembly.cpp +++ b/src/video_core/primitive_assembly.cpp @@ -17,15 +17,18 @@ template void PrimitiveAssembler::SubmitVertex(const VertexType& vtx, TriangleHandler triangle_handler) { switch (topology) { - // TODO: Figure out what's different with TriangleTopology::Shader. case PipelineRegs::TriangleTopology::List: case PipelineRegs::TriangleTopology::Shader: if (buffer_index < 2) { buffer[buffer_index++] = vtx; } else { buffer_index = 0; - - triangle_handler(buffer[0], buffer[1], vtx); + if (topology == PipelineRegs::TriangleTopology::Shader && winding) { + triangle_handler(buffer[1], buffer[0], vtx); + winding = false; + } else { + triangle_handler(buffer[0], buffer[1], vtx); + } } break; @@ -50,10 +53,16 @@ void PrimitiveAssembler::SubmitVertex(const VertexType& vtx, } } +template +void PrimitiveAssembler::SetWinding() { + winding = true; +} + template void PrimitiveAssembler::Reset() { buffer_index = 0; strip_ready = false; + winding = false; } template diff --git a/src/video_core/primitive_assembly.h b/src/video_core/primitive_assembly.h index e8eccdf27..12de8e3b9 100644 --- a/src/video_core/primitive_assembly.h +++ b/src/video_core/primitive_assembly.h @@ -29,6 +29,12 @@ struct PrimitiveAssembler { */ void SubmitVertex(const VertexType& vtx, TriangleHandler triangle_handler); + /** + * Invert the vertex order of the next triangle. Called by geometry shader emitter. + * This only takes effect for TriangleTopology::Shader. + */ + void SetWinding(); + /** * Resets the internal state of the PrimitiveAssembler. */ @@ -45,6 +51,7 @@ private: int buffer_index; VertexType buffer[2]; bool strip_ready = false; + bool winding = false; }; } // namespace From 8285ca4ad8f9a5d07c9a2ba91367fcf3756f5153 Mon Sep 17 00:00:00 2001 From: wwylele Date: Wed, 26 Jul 2017 18:44:52 +0300 Subject: [PATCH 06/31] pica/shader/jit: implement SETEMIT and EMIT --- .../shader/shader_jit_x64_compiler.cpp | 49 ++++++++++++++++++- .../shader/shader_jit_x64_compiler.h | 2 + 2 files changed, 49 insertions(+), 2 deletions(-) diff --git a/src/video_core/shader/shader_jit_x64_compiler.cpp b/src/video_core/shader/shader_jit_x64_compiler.cpp index 42a57aab1..1b31623bd 100644 --- a/src/video_core/shader/shader_jit_x64_compiler.cpp +++ b/src/video_core/shader/shader_jit_x64_compiler.cpp @@ -75,8 +75,8 @@ const JitFunction instr_table[64] = { &JitShader::Compile_IF, // ifu &JitShader::Compile_IF, // ifc &JitShader::Compile_LOOP, // loop - nullptr, // emit - nullptr, // sete + &JitShader::Compile_EMIT, // emit + &JitShader::Compile_SETE, // sete &JitShader::Compile_JMP, // jmpc &JitShader::Compile_JMP, // jmpu &JitShader::Compile_CMP, // cmp @@ -772,6 +772,51 @@ void JitShader::Compile_JMP(Instruction instr) { } } +static void Emit(GSEmitter* emitter, Math::Vec4 (*output)[16]) { + emitter->Emit(*output); +} + +void JitShader::Compile_EMIT(Instruction instr) { + Label have_emitter, end; + mov(rax, qword[STATE + offsetof(UnitState, emitter_ptr)]); + test(rax, rax); + jnz(have_emitter); + + ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0); + mov(ABI_PARAM1, reinterpret_cast("Execute EMIT on VS")); + CallFarFunction(*this, LogCritical); + ABI_PopRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0); + jmp(end); + + L(have_emitter); + ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0); + mov(ABI_PARAM1, rax); + mov(ABI_PARAM2, STATE); + add(ABI_PARAM2, static_cast(offsetof(UnitState, registers.output))); + CallFarFunction(*this, Emit); + ABI_PopRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0); + L(end); +} + +void JitShader::Compile_SETE(Instruction instr) { + Label have_emitter, end; + mov(rax, qword[STATE + offsetof(UnitState, emitter_ptr)]); + test(rax, rax); + jnz(have_emitter); + + ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0); + mov(ABI_PARAM1, reinterpret_cast("Execute SETEMIT on VS")); + CallFarFunction(*this, LogCritical); + ABI_PopRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0); + jmp(end); + + L(have_emitter); + mov(byte[rax + offsetof(GSEmitter, vertex_id)], instr.setemit.vertex_id); + mov(byte[rax + offsetof(GSEmitter, prim_emit)], instr.setemit.prim_emit); + mov(byte[rax + offsetof(GSEmitter, winding)], instr.setemit.winding); + L(end); +} + void JitShader::Compile_Block(unsigned end) { while (program_counter < end) { Compile_NextInstr(); diff --git a/src/video_core/shader/shader_jit_x64_compiler.h b/src/video_core/shader/shader_jit_x64_compiler.h index 31af0ca48..4aee56b1d 100644 --- a/src/video_core/shader/shader_jit_x64_compiler.h +++ b/src/video_core/shader/shader_jit_x64_compiler.h @@ -66,6 +66,8 @@ public: void Compile_JMP(Instruction instr); void Compile_CMP(Instruction instr); void Compile_MAD(Instruction instr); + void Compile_EMIT(Instruction instr); + void Compile_SETE(Instruction instr); private: void Compile_Block(unsigned end); From 0f35755572fe63534813528de9a0710193f2e335 Mon Sep 17 00:00:00 2001 From: wwylele Date: Fri, 4 Aug 2017 17:03:17 +0300 Subject: [PATCH 07/31] pica/command_processor: build geometry pipeline and run geometry shader The geometry pipeline manages data transfer between VS, GS and primitive assembler. It has known four modes: - no GS mode: sends VS output directly to the primitive assembler (what citra currently does) - GS mode 0: sends VS output to GS input registers, and sends GS output to primitive assembler - GS mode 1: sends VS output to GS uniform registers, and sends GS output to primitive assembler. It also takes an index from the index buffer at the beginning of each primitive for determine the primitive size. - GS mode 2: similar to mode 1, but doesn't take the index and uses a fixed primitive size. hwtest shows that immediate mode also supports GS (at least for mode 0), so the geometry pipeline gets refactored into its own class for supporting both drawing mode. In the immediate mode, some games don't set the pipeline registers to a valid value until the first attribute input, so a geometry pipeline reset flag is set in `pipeline.vs_default_attributes_setup.index` trigger, and the actual pipeline reconfigure is triggered in the first attribute input. In the normal drawing mode with index buffer, the vertex cache is a little bit modified to support the geometry pipeline. Instead of OutputVertex, it now holds AttributeBuffer, which is the input to the geometry pipeline. The AttributeBuffer->OutputVertex conversion is done inside the pipeline vertex handler. The actual hardware vertex cache is believed to be implemented in a similar way (because this is the only way that makes sense). Both geometry pipeline and GS unit rely on states preservation across drawing call, so they are put into the global state. In the future, the other three vertex shader units should be also placed in the global state, and a scheduler should be implemented on top of the four units. Note that the current gs_unit already allows running VS on it in the future. --- src/video_core/CMakeLists.txt | 2 + src/video_core/command_processor.cpp | 54 +++--- src/video_core/geometry_pipeline.cpp | 274 +++++++++++++++++++++++++++ src/video_core/geometry_pipeline.h | 49 +++++ src/video_core/pica.cpp | 21 +- src/video_core/pica_state.h | 11 ++ 6 files changed, 383 insertions(+), 28 deletions(-) create mode 100644 src/video_core/geometry_pipeline.cpp create mode 100644 src/video_core/geometry_pipeline.h diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index cffa4c952..82f47d8a9 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -1,6 +1,7 @@ set(SRCS command_processor.cpp debug_utils/debug_utils.cpp + geometry_pipeline.cpp pica.cpp primitive_assembly.cpp regs.cpp @@ -29,6 +30,7 @@ set(SRCS set(HEADERS command_processor.h debug_utils/debug_utils.h + geometry_pipeline.h gpu_debugger.h pica.h pica_state.h diff --git a/src/video_core/command_processor.cpp b/src/video_core/command_processor.cpp index f98ca3302..fb65a3a0a 100644 --- a/src/video_core/command_processor.cpp +++ b/src/video_core/command_processor.cpp @@ -161,6 +161,7 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { case PICA_REG_INDEX(pipeline.vs_default_attributes_setup.index): g_state.immediate.current_attribute = 0; + g_state.immediate.reset_geometry_pipeline = true; default_attr_counter = 0; break; @@ -234,16 +235,14 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { shader_engine->Run(g_state.vs, shader_unit); shader_unit.WriteOutput(regs.vs, output); - // Send to renderer - using Pica::Shader::OutputVertex; - auto AddTriangle = [](const OutputVertex& v0, const OutputVertex& v1, - const OutputVertex& v2) { - VideoCore::g_renderer->Rasterizer()->AddTriangle(v0, v1, v2); - }; - - g_state.primitive_assembler.SubmitVertex( - Shader::OutputVertex::FromAttributeBuffer(regs.rasterizer, output), - AddTriangle); + // Send to geometry pipeline + if (g_state.immediate.reset_geometry_pipeline) { + g_state.geometry_pipeline.Reconfigure(); + g_state.immediate.reset_geometry_pipeline = false; + } + ASSERT(!g_state.geometry_pipeline.NeedIndexInput()); + g_state.geometry_pipeline.Setup(shader_engine); + g_state.geometry_pipeline.SubmitVertex(output); } } } @@ -321,8 +320,8 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { // The size has been tuned for optimal balance between hit-rate and the cost of lookup const size_t VERTEX_CACHE_SIZE = 32; std::array vertex_cache_ids; - std::array vertex_cache; - Shader::OutputVertex output_vertex; + std::array vertex_cache; + Shader::AttributeBuffer vs_output; unsigned int vertex_cache_pos = 0; vertex_cache_ids.fill(-1); @@ -332,6 +331,11 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { shader_engine->SetupBatch(g_state.vs, regs.vs.main_offset); + g_state.geometry_pipeline.Reconfigure(); + g_state.geometry_pipeline.Setup(shader_engine); + if (g_state.geometry_pipeline.NeedIndexInput()) + ASSERT(is_indexed); + for (unsigned int index = 0; index < regs.pipeline.num_vertices; ++index) { // Indexed rendering doesn't use the start offset unsigned int vertex = @@ -345,6 +349,11 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { bool vertex_cache_hit = false; if (is_indexed) { + if (g_state.geometry_pipeline.NeedIndexInput()) { + g_state.geometry_pipeline.SubmitIndex(vertex); + continue; + } + if (g_debug_context && Pica::g_debug_context->recorder) { int size = index_u16 ? 2 : 1; memory_accesses.AddAccess(base_address + index_info.offset + size * index, @@ -353,7 +362,7 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { for (unsigned int i = 0; i < VERTEX_CACHE_SIZE; ++i) { if (vertex == vertex_cache_ids[i]) { - output_vertex = vertex_cache[i]; + vs_output = vertex_cache[i]; vertex_cache_hit = true; break; } @@ -362,7 +371,7 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { if (!vertex_cache_hit) { // Initialize data for the current vertex - Shader::AttributeBuffer input, output{}; + Shader::AttributeBuffer input; loader.LoadVertex(base_address, index, vertex, input, memory_accesses); // Send to vertex shader @@ -371,26 +380,17 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) { (void*)&input); shader_unit.LoadInput(regs.vs, input); shader_engine->Run(g_state.vs, shader_unit); - shader_unit.WriteOutput(regs.vs, output); - - // Retrieve vertex from register data - output_vertex = Shader::OutputVertex::FromAttributeBuffer(regs.rasterizer, output); + shader_unit.WriteOutput(regs.vs, vs_output); if (is_indexed) { - vertex_cache[vertex_cache_pos] = output_vertex; + vertex_cache[vertex_cache_pos] = vs_output; vertex_cache_ids[vertex_cache_pos] = vertex; vertex_cache_pos = (vertex_cache_pos + 1) % VERTEX_CACHE_SIZE; } } - // Send to renderer - using Pica::Shader::OutputVertex; - auto AddTriangle = [](const OutputVertex& v0, const OutputVertex& v1, - const OutputVertex& v2) { - VideoCore::g_renderer->Rasterizer()->AddTriangle(v0, v1, v2); - }; - - primitive_assembler.SubmitVertex(output_vertex, AddTriangle); + // Send to geometry pipeline + g_state.geometry_pipeline.SubmitVertex(vs_output); } for (auto& range : memory_accesses.ranges) { diff --git a/src/video_core/geometry_pipeline.cpp b/src/video_core/geometry_pipeline.cpp new file mode 100644 index 000000000..b146e2ecb --- /dev/null +++ b/src/video_core/geometry_pipeline.cpp @@ -0,0 +1,274 @@ +// Copyright 2017 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "video_core/geometry_pipeline.h" +#include "video_core/pica_state.h" +#include "video_core/regs.h" +#include "video_core/renderer_base.h" +#include "video_core/video_core.h" + +namespace Pica { + +/// An attribute buffering interface for different pipeline modes +class GeometryPipelineBackend { +public: + virtual ~GeometryPipelineBackend() = default; + + /// Checks if there is no incomplete data transfer + virtual bool IsEmpty() const = 0; + + /// Checks if the pipeline needs a direct input from index buffer + virtual bool NeedIndexInput() const = 0; + + /// Submits an index from index buffer + virtual void SubmitIndex(unsigned int val) = 0; + + /** + * Submits vertex attributes + * @param input attributes of a vertex output from vertex shader + * @return if the buffer is full and the geometry shader should be invoked + */ + virtual bool SubmitVertex(const Shader::AttributeBuffer& input) = 0; +}; + +// In the Point mode, vertex attributes are sent to the input registers in the geometry shader unit. +// The size of vertex shader outputs and geometry shader inputs are constants. Geometry shader is +// invoked upon inputs buffer filled up by vertex shader outputs. For example, if we have a geometry +// shader that takes 6 inputs, and the vertex shader outputs 2 attributes, it would take 3 vertices +// for one geometry shader invocation. +// TODO: what happens when the input size is not divisible by the output size? +class GeometryPipeline_Point : public GeometryPipelineBackend { +public: + GeometryPipeline_Point(const Regs& regs, Shader::GSUnitState& unit) : regs(regs), unit(unit) { + ASSERT(regs.pipeline.variable_primitive == 0); + ASSERT(regs.gs.input_to_uniform == 0); + vs_output_num = regs.pipeline.vs_outmap_total_minus_1_a + 1; + size_t gs_input_num = regs.gs.max_input_attribute_index + 1; + ASSERT(gs_input_num % vs_output_num == 0); + buffer_cur = attribute_buffer.attr; + buffer_end = attribute_buffer.attr + gs_input_num; + } + + bool IsEmpty() const override { + return buffer_cur == attribute_buffer.attr; + } + + bool NeedIndexInput() const override { + return false; + } + + void SubmitIndex(unsigned int val) override { + UNREACHABLE(); + } + + bool SubmitVertex(const Shader::AttributeBuffer& input) override { + buffer_cur = std::copy(input.attr, input.attr + vs_output_num, buffer_cur); + if (buffer_cur == buffer_end) { + buffer_cur = attribute_buffer.attr; + unit.LoadInput(regs.gs, attribute_buffer); + return true; + } + return false; + } + +private: + const Regs& regs; + Shader::GSUnitState& unit; + Shader::AttributeBuffer attribute_buffer; + Math::Vec4* buffer_cur; + Math::Vec4* buffer_end; + unsigned int vs_output_num; +}; + +// In VariablePrimitive mode, vertex attributes are buffered into the uniform registers in the +// geometry shader unit. The number of vertex is variable, which is specified by the first index +// value in the batch. This mode is usually used for subdivision. +class GeometryPipeline_VariablePrimitive : public GeometryPipelineBackend { +public: + GeometryPipeline_VariablePrimitive(const Regs& regs, Shader::ShaderSetup& setup) + : regs(regs), setup(setup) { + ASSERT(regs.pipeline.variable_primitive == 1); + ASSERT(regs.gs.input_to_uniform == 1); + vs_output_num = regs.pipeline.vs_outmap_total_minus_1_a + 1; + } + + bool IsEmpty() const override { + return need_index; + } + + bool NeedIndexInput() const override { + return need_index; + } + + void SubmitIndex(unsigned int val) override { + DEBUG_ASSERT(need_index); + + // The number of vertex input is put to the uniform register + float24 vertex_num = float24::FromFloat32(val); + setup.uniforms.f[0] = Math::MakeVec(vertex_num, vertex_num, vertex_num, vertex_num); + + // The second uniform register and so on are used for receiving input vertices + buffer_cur = setup.uniforms.f + 1; + + main_vertex_num = regs.pipeline.variable_vertex_main_num_minus_1 + 1; + total_vertex_num = val; + need_index = false; + } + + bool SubmitVertex(const Shader::AttributeBuffer& input) override { + DEBUG_ASSERT(!need_index); + if (main_vertex_num != 0) { + // For main vertices, receive all attributes + buffer_cur = std::copy(input.attr, input.attr + vs_output_num, buffer_cur); + --main_vertex_num; + } else { + // For other vertices, only receive the first attribute (usually the position) + *(buffer_cur++) = input.attr[0]; + } + --total_vertex_num; + + if (total_vertex_num == 0) { + need_index = true; + return true; + } + + return false; + } + +private: + bool need_index = true; + const Regs& regs; + Shader::ShaderSetup& setup; + unsigned int main_vertex_num; + unsigned int total_vertex_num; + Math::Vec4* buffer_cur; + unsigned int vs_output_num; +}; + +// In FixedPrimitive mode, vertex attributes are buffered into the uniform registers in the geometry +// shader unit. The number of vertex per shader invocation is constant. This is usually used for +// particle system. +class GeometryPipeline_FixedPrimitive : public GeometryPipelineBackend { +public: + GeometryPipeline_FixedPrimitive(const Regs& regs, Shader::ShaderSetup& setup) + : regs(regs), setup(setup) { + ASSERT(regs.pipeline.variable_primitive == 0); + ASSERT(regs.gs.input_to_uniform == 1); + vs_output_num = regs.pipeline.vs_outmap_total_minus_1_a + 1; + ASSERT(vs_output_num == regs.pipeline.gs_config.stride_minus_1 + 1); + size_t vertex_num = regs.pipeline.gs_config.fixed_vertex_num_minus_1 + 1; + buffer_cur = buffer_begin = setup.uniforms.f + regs.pipeline.gs_config.start_index; + buffer_end = buffer_begin + vs_output_num * vertex_num; + } + + bool IsEmpty() const override { + return buffer_cur == buffer_begin; + } + + bool NeedIndexInput() const override { + return false; + } + + void SubmitIndex(unsigned int val) override { + UNREACHABLE(); + } + + bool SubmitVertex(const Shader::AttributeBuffer& input) override { + buffer_cur = std::copy(input.attr, input.attr + vs_output_num, buffer_cur); + if (buffer_cur == buffer_end) { + buffer_cur = buffer_begin; + return true; + } + return false; + } + +private: + const Regs& regs; + Shader::ShaderSetup& setup; + Math::Vec4* buffer_begin; + Math::Vec4* buffer_cur; + Math::Vec4* buffer_end; + unsigned int vs_output_num; +}; + +GeometryPipeline::GeometryPipeline(State& state) : state(state) {} + +GeometryPipeline::~GeometryPipeline() = default; + +void GeometryPipeline::SetVertexHandler(Shader::VertexHandler vertex_handler) { + this->vertex_handler = vertex_handler; +} + +void GeometryPipeline::Setup(Shader::ShaderEngine* shader_engine) { + if (!backend) + return; + + this->shader_engine = shader_engine; + shader_engine->SetupBatch(state.gs, state.regs.gs.main_offset); +} + +void GeometryPipeline::Reconfigure() { + ASSERT(!backend || backend->IsEmpty()); + + if (state.regs.pipeline.use_gs == PipelineRegs::UseGS::No) { + backend = nullptr; + return; + } + + ASSERT(state.regs.pipeline.use_gs == PipelineRegs::UseGS::Yes); + + // The following assumes that when geometry shader is in use, the shader unit 3 is configured as + // a geometry shader unit. + // TODO: what happens if this is not true? + ASSERT(state.regs.pipeline.gs_unit_exclusive_configuration == 1); + ASSERT(state.regs.gs.shader_mode == ShaderRegs::ShaderMode::GS); + + state.gs_unit.ConfigOutput(state.regs.gs); + + ASSERT(state.regs.pipeline.vs_outmap_total_minus_1_a == + state.regs.pipeline.vs_outmap_total_minus_1_b); + + switch (state.regs.pipeline.gs_config.mode) { + case PipelineRegs::GSMode::Point: + backend = std::make_unique(state.regs, state.gs_unit); + break; + case PipelineRegs::GSMode::VariablePrimitive: + backend = std::make_unique(state.regs, state.gs); + break; + case PipelineRegs::GSMode::FixedPrimitive: + backend = std::make_unique(state.regs, state.gs); + break; + default: + UNREACHABLE(); + } +} + +bool GeometryPipeline::NeedIndexInput() const { + if (!backend) + return false; + return backend->NeedIndexInput(); +} + +void GeometryPipeline::SubmitIndex(unsigned int val) { + backend->SubmitIndex(val); +} + +void GeometryPipeline::SubmitVertex(const Shader::AttributeBuffer& input) { + if (!backend) { + // No backend means the geometry shader is disabled, so we send the vertex shader output + // directly to the primitive assembler. + vertex_handler(input); + } else { + if (backend->SubmitVertex(input)) { + shader_engine->Run(state.gs, state.gs_unit); + + // The uniform b15 is set to true after every geometry shader invocation. This is useful + // for the shader to know if this is the first invocation in a batch, if the program set + // b15 to false first. + state.gs.uniforms.b[15] = true; + } + } +} + +} // namespace Pica diff --git a/src/video_core/geometry_pipeline.h b/src/video_core/geometry_pipeline.h new file mode 100644 index 000000000..91fdd3192 --- /dev/null +++ b/src/video_core/geometry_pipeline.h @@ -0,0 +1,49 @@ +// Copyright 2017 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include +#include "video_core/shader/shader.h" + +namespace Pica { + +struct State; + +class GeometryPipelineBackend; + +/// A pipeline receiving from vertex shader and sending to geometry shader and primitive assembler +class GeometryPipeline { +public: + explicit GeometryPipeline(State& state); + ~GeometryPipeline(); + + /// Sets the handler for receiving vertex outputs from vertex shader + void SetVertexHandler(Shader::VertexHandler vertex_handler); + + /** + * Setup the geometry shader unit if it is in use + * @param shader_engine the shader engine for the geometry shader to run + */ + void Setup(Shader::ShaderEngine* shader_engine); + + /// Reconfigures the pipeline according to current register settings + void Reconfigure(); + + /// Checks if the pipeline needs a direct input from index buffer + bool NeedIndexInput() const; + + /// Submits an index from index buffer. Call this only when NeedIndexInput returns true + void SubmitIndex(unsigned int val); + + /// Submits vertex attributes output from vertex shader + void SubmitVertex(const Shader::AttributeBuffer& input); + +private: + Shader::VertexHandler vertex_handler; + Shader::ShaderEngine* shader_engine; + std::unique_ptr backend; + State& state; +}; +} // namespace Pica diff --git a/src/video_core/pica.cpp b/src/video_core/pica.cpp index b95148a6a..218e06883 100644 --- a/src/video_core/pica.cpp +++ b/src/video_core/pica.cpp @@ -3,9 +3,11 @@ // Refer to the license.txt file included. #include +#include "video_core/geometry_pipeline.h" #include "video_core/pica.h" #include "video_core/pica_state.h" -#include "video_core/regs_pipeline.h" +#include "video_core/renderer_base.h" +#include "video_core/video_core.h" namespace Pica { @@ -24,6 +26,23 @@ void Zero(T& o) { memset(&o, 0, sizeof(o)); } +State::State() : geometry_pipeline(*this) { + auto SubmitVertex = [this](const Shader::AttributeBuffer& vertex) { + using Pica::Shader::OutputVertex; + auto AddTriangle = [this](const OutputVertex& v0, const OutputVertex& v1, + const OutputVertex& v2) { + VideoCore::g_renderer->Rasterizer()->AddTriangle(v0, v1, v2); + }; + primitive_assembler.SubmitVertex( + Shader::OutputVertex::FromAttributeBuffer(regs.rasterizer, vertex), AddTriangle); + }; + + auto SetWinding = [this]() { primitive_assembler.SetWinding(); }; + + g_state.gs_unit.SetVertexHandler(SubmitVertex, SetWinding); + g_state.geometry_pipeline.SetVertexHandler(SubmitVertex); +} + void State::Reset() { Zero(regs); Zero(vs); diff --git a/src/video_core/pica_state.h b/src/video_core/pica_state.h index 864a2c9e6..c6634a0bc 100644 --- a/src/video_core/pica_state.h +++ b/src/video_core/pica_state.h @@ -8,6 +8,7 @@ #include "common/bit_field.h" #include "common/common_types.h" #include "common/vector_math.h" +#include "video_core/geometry_pipeline.h" #include "video_core/primitive_assembly.h" #include "video_core/regs.h" #include "video_core/shader/shader.h" @@ -16,6 +17,7 @@ namespace Pica { /// Struct used to describe current Pica state struct State { + State(); void Reset(); /// Pica registers @@ -137,8 +139,17 @@ struct State { Shader::AttributeBuffer input_vertex; // Index of the next attribute to be loaded into `input_vertex`. u32 current_attribute = 0; + // Indicates the immediate mode just started and the geometry pipeline needs to reconfigure + bool reset_geometry_pipeline = true; } immediate; + // the geometry shader needs to be kept in the global state because some shaders relie on + // preserved register value across shader invocation. + // TODO: also bring the three vertex shader units here and implement the shader scheduler. + Shader::GSUnitState gs_unit; + + GeometryPipeline geometry_pipeline; + // This is constructed with a dummy triangle topology PrimitiveAssembler primitive_assembler; }; From 3e478ca13110639a67ad95880aae5d7d13e096b7 Mon Sep 17 00:00:00 2001 From: wwylele Date: Fri, 18 Aug 2017 15:04:56 +0300 Subject: [PATCH 08/31] SwRasterizer/Lighting: implement bump mapping --- src/video_core/swrasterizer/lighting.cpp | 28 ++++++++++++++++++---- src/video_core/swrasterizer/lighting.h | 3 ++- src/video_core/swrasterizer/rasterizer.cpp | 4 ++-- 3 files changed, 27 insertions(+), 8 deletions(-) diff --git a/src/video_core/swrasterizer/lighting.cpp b/src/video_core/swrasterizer/lighting.cpp index 39a3e396d..4f16bac07 100644 --- a/src/video_core/swrasterizer/lighting.cpp +++ b/src/video_core/swrasterizer/lighting.cpp @@ -22,14 +22,32 @@ static float LookupLightingLut(const Pica::State::Lighting& lighting, size_t lut std::tuple, Math::Vec4> ComputeFragmentsColors( const Pica::LightingRegs& lighting, const Pica::State::Lighting& lighting_state, - const Math::Quaternion& normquat, const Math::Vec3& view) { + const Math::Quaternion& normquat, const Math::Vec3& view, + const Math::Vec4 (&texture_color)[4]) { - // TODO(Subv): Bump mapping - Math::Vec3 surface_normal = {0.0f, 0.0f, 1.0f}; + Math::Vec3 surface_normal; + Math::Vec3 surface_tangent; if (lighting.config0.bump_mode != LightingRegs::LightingBumpMode::None) { - LOG_CRITICAL(HW_GPU, "unimplemented bump mapping"); - UNIMPLEMENTED(); + Math::Vec3 perturbation = + texture_color[lighting.config0.bump_selector].xyz().Cast() / 127.5f - + Math::MakeVec(1.0f, 1.0f, 1.0f); + if (lighting.config0.bump_mode == LightingRegs::LightingBumpMode::NormalMap) { + if (!lighting.config0.disable_bump_renorm) { + const float z_square = 1 - perturbation.xy().Length2(); + perturbation.z = std::sqrt(std::max(z_square, 0.0f)); + } + surface_normal = perturbation; + surface_tangent = Math::MakeVec(1.0f, 0.0f, 0.0f); + } else if (lighting.config0.bump_mode == LightingRegs::LightingBumpMode::TangentMap) { + surface_normal = Math::MakeVec(0.0f, 0.0f, 1.0f); + surface_tangent = perturbation; + } else { + LOG_ERROR(HW_GPU, "Unknown bump mode %u", lighting.config0.bump_mode.Value()); + } + } else { + surface_normal = Math::MakeVec(0.0f, 0.0f, 1.0f); + surface_tangent = Math::MakeVec(1.0f, 0.0f, 0.0f); } // Use the normalized the quaternion when performing the rotation diff --git a/src/video_core/swrasterizer/lighting.h b/src/video_core/swrasterizer/lighting.h index 438dca926..d807a3d94 100644 --- a/src/video_core/swrasterizer/lighting.h +++ b/src/video_core/swrasterizer/lighting.h @@ -13,6 +13,7 @@ namespace Pica { std::tuple, Math::Vec4> ComputeFragmentsColors( const Pica::LightingRegs& lighting, const Pica::State::Lighting& lighting_state, - const Math::Quaternion& normquat, const Math::Vec3& view); + const Math::Quaternion& normquat, const Math::Vec3& view, + const Math::Vec4 (&texture_color)[4]); } // namespace Pica diff --git a/src/video_core/swrasterizer/rasterizer.cpp b/src/video_core/swrasterizer/rasterizer.cpp index fdc1df199..862135614 100644 --- a/src/video_core/swrasterizer/rasterizer.cpp +++ b/src/video_core/swrasterizer/rasterizer.cpp @@ -437,8 +437,8 @@ static void ProcessTriangleInternal(const Vertex& v0, const Vertex& v1, const Ve GetInterpolatedAttribute(v0.view.y, v1.view.y, v2.view.y).ToFloat32(), GetInterpolatedAttribute(v0.view.z, v1.view.z, v2.view.z).ToFloat32(), }; - std::tie(primary_fragment_color, secondary_fragment_color) = - ComputeFragmentsColors(g_state.regs.lighting, g_state.lighting, normquat, view); + std::tie(primary_fragment_color, secondary_fragment_color) = ComputeFragmentsColors( + g_state.regs.lighting, g_state.lighting, normquat, view, texture_color); } for (unsigned tev_stage_index = 0; tev_stage_index < tev_stages.size(); From b5aa5703540adceb1fc867b577dad50388a47e15 Mon Sep 17 00:00:00 2001 From: wwylele Date: Fri, 18 Aug 2017 16:35:11 +0300 Subject: [PATCH 09/31] SwRasterizer/Lighting: implement LUT input CP --- src/video_core/swrasterizer/lighting.cpp | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/src/video_core/swrasterizer/lighting.cpp b/src/video_core/swrasterizer/lighting.cpp index 4f16bac07..b38964530 100644 --- a/src/video_core/swrasterizer/lighting.cpp +++ b/src/video_core/swrasterizer/lighting.cpp @@ -52,6 +52,7 @@ std::tuple, Math::Vec4> ComputeFragmentsColors( // Use the normalized the quaternion when performing the rotation auto normal = Math::QuaternionRotate(normquat, surface_normal); + auto tangent = Math::QuaternionRotate(normquat, surface_tangent); Math::Vec4 diffuse_sum = {0.0f, 0.0f, 0.0f, 1.0f}; Math::Vec4 specular_sum = {0.0f, 0.0f, 0.0f, 1.0f}; @@ -120,6 +121,16 @@ std::tuple, Math::Vec4> ComputeFragmentsColors( result = Math::Dot(light_vector, spot_dir.Cast() / 2047.0f); break; } + case LightingRegs::LightingLutInput::CP: + if (lighting.config0.config == LightingRegs::LightingConfig::Config7) { + const Math::Vec3 norm_half_vector = half_vector.Normalized(); + const Math::Vec3 half_vector_proj = + norm_half_vector - normal * Math::Dot(normal, norm_half_vector); + result = Math::Dot(half_vector_proj, tangent); + } else { + result = 0.0f; + } + break; default: LOG_CRITICAL(HW_GPU, "Unknown lighting LUT input %u\n", static_cast(input)); UNIMPLEMENTED(); From 17c6104d2afda7bf354c454f87561a3dbdf524e3 Mon Sep 17 00:00:00 2001 From: wwylele Date: Mon, 21 Aug 2017 12:03:38 +0300 Subject: [PATCH 10/31] gl_rasterizer/lighting: more accurate CP formula --- src/video_core/renderer_opengl/gl_shader_gen.cpp | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp index ae67aab05..d85f281e5 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.cpp +++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp @@ -594,8 +594,8 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) { // Note: even if the normal vector is modified by normal map, which is not the // normal of the tangent plane anymore, the half angle vector is still projected // using the modified normal vector. - std::string half_angle_proj = "normalize(half_vector) - normal / dot(normal, " - "normal) * dot(normal, normalize(half_vector))"; + std::string half_angle_proj = + "normalize(half_vector) - normal * dot(normal, normalize(half_vector))"; // Note: the half angle vector projection is confirmed not normalized before the dot // product. The result is in fact not cos(phi) as the name suggested. index = "dot(" + half_angle_proj + ", tangent)"; From c84e60b4700778db236d3177714a076c515f9ce7 Mon Sep 17 00:00:00 2001 From: wwylele Date: Tue, 8 Aug 2017 21:34:17 +0300 Subject: [PATCH 11/31] HID: use TouchDevice for touch pad --- src/core/frontend/input.h | 6 ++++++ src/core/hle/service/hid/hid.cpp | 12 ++++++++---- src/core/settings.h | 1 + 3 files changed, 15 insertions(+), 4 deletions(-) diff --git a/src/core/frontend/input.h b/src/core/frontend/input.h index 5916a901d..8c256beb5 100644 --- a/src/core/frontend/input.h +++ b/src/core/frontend/input.h @@ -126,4 +126,10 @@ using AnalogDevice = InputDevice>; */ using MotionDevice = InputDevice, Math::Vec3>>; +/** + * A touch device is an input device that returns a tuple of two floats and a bool. The floats are + * x and y coordinates in the range 0.0 - 1.0, and the bool indicates whether it is pressed. + */ +using TouchDevice = InputDevice>; + } // namespace Input diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp index 31f34a7ae..aa5d821f9 100644 --- a/src/core/hle/service/hid/hid.cpp +++ b/src/core/hle/service/hid/hid.cpp @@ -7,9 +7,9 @@ #include #include #include "common/logging/log.h" +#include "core/3ds.h" #include "core/core.h" #include "core/core_timing.h" -#include "core/frontend/emu_window.h" #include "core/frontend/input.h" #include "core/hle/ipc.h" #include "core/hle/kernel/event.h" @@ -19,7 +19,6 @@ #include "core/hle/service/hid/hid_spvr.h" #include "core/hle/service/hid/hid_user.h" #include "core/hle/service/service.h" -#include "video_core/video_core.h" namespace Service { namespace HID { @@ -59,6 +58,7 @@ static std::array, Settings::NativeButton:: buttons; static std::unique_ptr circle_pad; static std::unique_ptr motion_device; +static std::unique_ptr touch_device; DirectionState GetStickDirectionState(s16 circle_pad_x, s16 circle_pad_y) { // 30 degree and 60 degree are angular thresholds for directions @@ -96,6 +96,7 @@ static void LoadInputDevices() { circle_pad = Input::CreateDevice( Settings::values.analogs[Settings::NativeAnalog::CirclePad]); motion_device = Input::CreateDevice(Settings::values.motion_device); + touch_device = Input::CreateDevice(Settings::values.touch_device); } static void UnloadInputDevices() { @@ -104,6 +105,7 @@ static void UnloadInputDevices() { } circle_pad.reset(); motion_device.reset(); + touch_device.reset(); } static void UpdatePadCallback(u64 userdata, int cycles_late) { @@ -172,8 +174,10 @@ static void UpdatePadCallback(u64 userdata, int cycles_late) { // Get the current touch entry TouchDataEntry& touch_entry = mem->touch.entries[mem->touch.index]; bool pressed = false; - - std::tie(touch_entry.x, touch_entry.y, pressed) = VideoCore::g_emu_window->GetTouchState(); + float x, y; + std::tie(x, y, pressed) = touch_device->GetStatus(); + touch_entry.x = static_cast(x * Core::kScreenBottomWidth); + touch_entry.y = static_cast(y * Core::kScreenBottomHeight); touch_entry.valid.Assign(pressed ? 1 : 0); // TODO(bunnei): We're not doing anything with offset 0xA8 + 0x18 of HID SharedMemory, which diff --git a/src/core/settings.h b/src/core/settings.h index 7e15b119b..088d7651a 100644 --- a/src/core/settings.h +++ b/src/core/settings.h @@ -80,6 +80,7 @@ struct Values { std::array buttons; std::array analogs; std::string motion_device; + std::string touch_device; // Core bool use_cpu_jit; From 2617de1fe6f6f1fc846a8e038e1ea77a894554b2 Mon Sep 17 00:00:00 2001 From: wwylele Date: Wed, 9 Aug 2017 02:57:42 +0300 Subject: [PATCH 12/31] EmuWindow: refactor touch input into a TouchDevice --- src/citra/config.cpp | 2 + src/citra/default_ini.h | 4 ++ src/citra_qt/configuration/config.cpp | 3 ++ src/core/frontend/emu_window.cpp | 71 ++++++++++++++++++++++----- src/core/frontend/emu_window.h | 31 ++---------- 5 files changed, 72 insertions(+), 39 deletions(-) diff --git a/src/citra/config.cpp b/src/citra/config.cpp index 73846ed91..93569fc4c 100644 --- a/src/citra/config.cpp +++ b/src/citra/config.cpp @@ -78,6 +78,8 @@ void Config::ReadValues() { Settings::values.motion_device = sdl2_config->Get( "Controls", "motion_device", "engine:motion_emu,update_period:100,sensitivity:0.01"); + Settings::values.touch_device = + sdl2_config->Get("Controls", "touch_device", "engine:emu_window"); // Core Settings::values.use_cpu_jit = sdl2_config->GetBoolean("Core", "use_cpu_jit", true); diff --git a/src/citra/default_ini.h b/src/citra/default_ini.h index 9ea779dd8..fa9dda651 100644 --- a/src/citra/default_ini.h +++ b/src/citra/default_ini.h @@ -62,6 +62,10 @@ c_stick= # - "sensitivity": the coefficient converting mouse movement to tilting angle (default to 0.01) motion_device= +# for touch input, the following devices are available: +# - "emu_window" (default) for emulating touch input from mouse input to the emulation window. No parameters required +touch_device= + [Core] # Whether to use the Just-In-Time (JIT) compiler for CPU emulation # 0: Interpreter (slow), 1 (default): JIT (fast) diff --git a/src/citra_qt/configuration/config.cpp b/src/citra_qt/configuration/config.cpp index 6e42db007..e398c6f29 100644 --- a/src/citra_qt/configuration/config.cpp +++ b/src/citra_qt/configuration/config.cpp @@ -61,6 +61,8 @@ void Config::ReadValues() { qt_config->value("motion_device", "engine:motion_emu,update_period:100,sensitivity:0.01") .toString() .toStdString(); + Settings::values.touch_device = + qt_config->value("touch_device", "engine:emu_window").toString().toStdString(); qt_config->endGroup(); @@ -209,6 +211,7 @@ void Config::SaveValues() { QString::fromStdString(Settings::values.analogs[i])); } qt_config->setValue("motion_device", QString::fromStdString(Settings::values.motion_device)); + qt_config->setValue("touch_device", QString::fromStdString(Settings::values.touch_device)); qt_config->endGroup(); qt_config->beginGroup("Core"); diff --git a/src/core/frontend/emu_window.cpp b/src/core/frontend/emu_window.cpp index 60b20d4e2..787c517ff 100644 --- a/src/core/frontend/emu_window.cpp +++ b/src/core/frontend/emu_window.cpp @@ -2,14 +2,55 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include #include -#include "common/assert.h" -#include "core/3ds.h" -#include "core/core.h" +#include #include "core/frontend/emu_window.h" +#include "core/frontend/input.h" #include "core/settings.h" +class EmuWindow::TouchState : public Input::Factory, + public std::enable_shared_from_this { +public: + std::unique_ptr Create(const Common::ParamPackage&) override { + return std::make_unique(shared_from_this()); + } + + std::mutex mutex; + + bool touch_pressed = false; ///< True if touchpad area is currently pressed, otherwise false + + float touch_x = 0.0f; ///< Touchpad X-position + float touch_y = 0.0f; ///< Touchpad Y-position + +private: + class Device : public Input::TouchDevice { + public: + explicit Device(std::weak_ptr&& touch_state) : touch_state(touch_state) {} + std::tuple GetStatus() const override { + if (auto state = touch_state.lock()) { + std::lock_guard guard(state->mutex); + return std::make_tuple(state->touch_x, state->touch_y, state->touch_pressed); + } + return std::make_tuple(0.0f, 0.0f, false); + } + + private: + std::weak_ptr touch_state; + }; +}; + +EmuWindow::EmuWindow() { + // TODO: Find a better place to set this. + config.min_client_area_size = std::make_pair(400u, 480u); + active_config = config; + touch_state = std::make_shared(); + Input::RegisterFactory("emu_window", touch_state); +} + +EmuWindow::~EmuWindow() { + Input::UnregisterFactory("emu_window"); +} + /** * Check if the given x/y coordinates are within the touchpad specified by the framebuffer layout * @param layout FramebufferLayout object describing the framebuffer size and screen positions @@ -38,22 +79,26 @@ void EmuWindow::TouchPressed(unsigned framebuffer_x, unsigned framebuffer_y) { if (!IsWithinTouchscreen(framebuffer_layout, framebuffer_x, framebuffer_y)) return; - touch_x = Core::kScreenBottomWidth * (framebuffer_x - framebuffer_layout.bottom_screen.left) / - (framebuffer_layout.bottom_screen.right - framebuffer_layout.bottom_screen.left); - touch_y = Core::kScreenBottomHeight * (framebuffer_y - framebuffer_layout.bottom_screen.top) / - (framebuffer_layout.bottom_screen.bottom - framebuffer_layout.bottom_screen.top); + std::lock_guard guard(touch_state->mutex); + touch_state->touch_x = + static_cast(framebuffer_x - framebuffer_layout.bottom_screen.left) / + (framebuffer_layout.bottom_screen.right - framebuffer_layout.bottom_screen.left); + touch_state->touch_y = + static_cast(framebuffer_y - framebuffer_layout.bottom_screen.top) / + (framebuffer_layout.bottom_screen.bottom - framebuffer_layout.bottom_screen.top); - touch_pressed = true; + touch_state->touch_pressed = true; } void EmuWindow::TouchReleased() { - touch_pressed = false; - touch_x = 0; - touch_y = 0; + std::lock_guard guard(touch_state->mutex); + touch_state->touch_pressed = false; + touch_state->touch_x = 0; + touch_state->touch_y = 0; } void EmuWindow::TouchMoved(unsigned framebuffer_x, unsigned framebuffer_y) { - if (!touch_pressed) + if (!touch_state->touch_pressed) return; if (!IsWithinTouchscreen(framebuffer_layout, framebuffer_x, framebuffer_y)) diff --git a/src/core/frontend/emu_window.h b/src/core/frontend/emu_window.h index 7bdee251c..c10dee51b 100644 --- a/src/core/frontend/emu_window.h +++ b/src/core/frontend/emu_window.h @@ -4,11 +4,10 @@ #pragma once -#include +#include #include #include #include "common/common_types.h" -#include "common/math_util.h" #include "core/frontend/framebuffer_layout.h" /** @@ -68,17 +67,6 @@ public: */ void TouchMoved(unsigned framebuffer_x, unsigned framebuffer_y); - /** - * Gets the current touch screen state (touch X/Y coordinates and whether or not it is pressed). - * @note This should be called by the core emu thread to get a state set by the window thread. - * @todo Fix this function to be thread-safe. - * @return std::tuple of (x, y, pressed) where `x` and `y` are the touch coordinates and - * `pressed` is true if the touch screen is currently being pressed - */ - std::tuple GetTouchState() const { - return std::make_tuple(touch_x, touch_y, touch_pressed); - } - /** * Returns currently active configuration. * @note Accesses to the returned object need not be consistent because it may be modified in @@ -113,15 +101,8 @@ public: void UpdateCurrentFramebufferLayout(unsigned width, unsigned height); protected: - EmuWindow() { - // TODO: Find a better place to set this. - config.min_client_area_size = std::make_pair(400u, 480u); - active_config = config; - touch_x = 0; - touch_y = 0; - touch_pressed = false; - } - virtual ~EmuWindow() {} + EmuWindow(); + virtual ~EmuWindow(); /** * Processes any pending configuration changes from the last SetConfig call. @@ -177,10 +158,8 @@ private: /// ProcessConfigurationChanges) WindowConfig active_config; ///< Internal active configuration - bool touch_pressed; ///< True if touchpad area is currently pressed, otherwise false - - u16 touch_x; ///< Touchpad X-position in native 3DS pixel coordinates (0-320) - u16 touch_y; ///< Touchpad Y-position in native 3DS pixel coordinates (0-240) + class TouchState; + std::shared_ptr touch_state; /** * Clip the provided coordinates to be inside the touchscreen area. From 076b39510451b1a0601e8a4db0d690615c81200f Mon Sep 17 00:00:00 2001 From: stone3311 Date: Sat, 26 Aug 2017 18:35:45 +0200 Subject: [PATCH 13/31] Fix info about TODO list --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index e766918f7..31f5afe27 100644 --- a/README.md +++ b/README.md @@ -17,7 +17,7 @@ For development discussion, please join us @ #citra on freenode. Most of the development happens on GitHub. It's also where [our central repository](https://github.com/citra-emu/citra) is hosted. -If you want to contribute please take a look at the [Contributor's Guide](CONTRIBUTING.md), [TODO list](https://docs.google.com/document/d/1SWIop0uBI9IW8VGg97TAtoT_CHNoP42FzYmvG1F4QDA) and [Developer Information](https://github.com/citra-emu/citra/wiki/Developer-Information). You should as well contact any of the developers in the forum in order to know about the current state of the emulator. +If you want to contribute please take a look at the [Contributor's Guide](CONTRIBUTING.md) and [Developer Information](https://github.com/citra-emu/citra/wiki/Developer-Information). You should as well contact any of the developers in the forum in order to know about the current state of the emulator because the [TODO list](https://docs.google.com/document/d/1SWIop0uBI9IW8VGg97TAtoT_CHNoP42FzYmvG1F4QDA) isn't maintained anymore. ### Building From da88f3b8f0f9f1162b7ad41f70e2126195eee999 Mon Sep 17 00:00:00 2001 From: Subv Date: Mon, 21 Aug 2017 12:18:52 -0500 Subject: [PATCH 14/31] Warnings: Fixed a few missing-return warnings in video_core. --- src/video_core/regs_framebuffer.h | 10 ++++------ src/video_core/swrasterizer/framebuffer.cpp | 2 ++ src/video_core/swrasterizer/texturing.cpp | 4 ++++ 3 files changed, 10 insertions(+), 6 deletions(-) diff --git a/src/video_core/regs_framebuffer.h b/src/video_core/regs_framebuffer.h index a50bd4111..7b565f911 100644 --- a/src/video_core/regs_framebuffer.h +++ b/src/video_core/regs_framebuffer.h @@ -256,10 +256,9 @@ struct FramebufferRegs { return 3; case DepthFormat::D24S8: return 4; - default: - LOG_CRITICAL(HW_GPU, "Unknown depth format %u", format); - UNIMPLEMENTED(); } + + ASSERT_MSG(false, "Unknown depth format %u", format); } // Returns the number of bits per depth component of the specified depth format @@ -270,10 +269,9 @@ struct FramebufferRegs { case DepthFormat::D24: case DepthFormat::D24S8: return 24; - default: - LOG_CRITICAL(HW_GPU, "Unknown depth format %u", format); - UNIMPLEMENTED(); } + + ASSERT_MSG(false, "Unknown depth format %u", format); } INSERT_PADDING_WORDS(0x20); diff --git a/src/video_core/swrasterizer/framebuffer.cpp b/src/video_core/swrasterizer/framebuffer.cpp index 7de3aac75..f34eab6cf 100644 --- a/src/video_core/swrasterizer/framebuffer.cpp +++ b/src/video_core/swrasterizer/framebuffer.cpp @@ -352,6 +352,8 @@ u8 LogicOp(u8 src, u8 dest, FramebufferRegs::LogicOp op) { case FramebufferRegs::LogicOp::OrInverted: return ~src | dest; } + + UNREACHABLE(); }; } // namespace Rasterizer diff --git a/src/video_core/swrasterizer/texturing.cpp b/src/video_core/swrasterizer/texturing.cpp index 4f02b93f2..79b1ce841 100644 --- a/src/video_core/swrasterizer/texturing.cpp +++ b/src/video_core/swrasterizer/texturing.cpp @@ -89,6 +89,8 @@ Math::Vec3 GetColorModifier(TevStageConfig::ColorModifier factor, case ColorModifier::OneMinusSourceBlue: return (Math::Vec3(255, 255, 255) - values.bbb()).Cast(); } + + UNREACHABLE(); }; u8 GetAlphaModifier(TevStageConfig::AlphaModifier factor, const Math::Vec4& values) { @@ -119,6 +121,8 @@ u8 GetAlphaModifier(TevStageConfig::AlphaModifier factor, const Math::Vec4& case AlphaModifier::OneMinusSourceBlue: return 255 - values.b(); } + + UNREACHABLE(); }; Math::Vec3 ColorCombine(TevStageConfig::Operation op, const Math::Vec3 input[3]) { From 54411bef4eb16af0822820205a923690ea7e822a Mon Sep 17 00:00:00 2001 From: Subv Date: Mon, 17 Jul 2017 09:25:58 -0500 Subject: [PATCH 15/31] Services/UDS: Add functions to generate 802.11 auth and assoc response frames. --- src/core/CMakeLists.txt | 2 + src/core/hle/service/nwm/nwm_uds.h | 12 ++++ src/core/hle/service/nwm/uds_beacon.h | 11 --- src/core/hle/service/nwm/uds_connection.cpp | 79 +++++++++++++++++++++ src/core/hle/service/nwm/uds_connection.h | 51 +++++++++++++ 5 files changed, 144 insertions(+), 11 deletions(-) create mode 100644 src/core/hle/service/nwm/uds_connection.cpp create mode 100644 src/core/hle/service/nwm/uds_connection.h diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt index ea09819e5..0719138af 100644 --- a/src/core/CMakeLists.txt +++ b/src/core/CMakeLists.txt @@ -144,6 +144,7 @@ set(SRCS hle/service/nwm/nwm_tst.cpp hle/service/nwm/nwm_uds.cpp hle/service/nwm/uds_beacon.cpp + hle/service/nwm/uds_connection.cpp hle/service/nwm/uds_data.cpp hle/service/pm_app.cpp hle/service/ptm/ptm.cpp @@ -342,6 +343,7 @@ set(HEADERS hle/service/nwm/nwm_tst.h hle/service/nwm/nwm_uds.h hle/service/nwm/uds_beacon.h + hle/service/nwm/uds_connection.h hle/service/nwm/uds_data.h hle/service/pm_app.h hle/service/ptm/ptm.h diff --git a/src/core/hle/service/nwm/nwm_uds.h b/src/core/hle/service/nwm/nwm_uds.h index 141f49f9c..f1caaf974 100644 --- a/src/core/hle/service/nwm/nwm_uds.h +++ b/src/core/hle/service/nwm/nwm_uds.h @@ -42,6 +42,7 @@ using NodeList = std::vector; enum class NetworkStatus { NotConnected = 3, ConnectedAsHost = 6, + Connecting = 7, ConnectedAsClient = 9, ConnectedAsSpectator = 10, }; @@ -85,6 +86,17 @@ static_assert(offsetof(NetworkInfo, oui_value) == 0xC, "oui_value is at the wron static_assert(offsetof(NetworkInfo, wlan_comm_id) == 0x10, "wlancommid is at the wrong offset."); static_assert(sizeof(NetworkInfo) == 0x108, "NetworkInfo has incorrect size."); +/// Additional block tag ids in the Beacon and Association Response frames +enum class TagId : u8 { + SSID = 0, + SupportedRates = 1, + DSParameterSet = 2, + TrafficIndicationMap = 5, + CountryInformation = 7, + ERPInformation = 42, + VendorSpecific = 221 +}; + class NWM_UDS final : public Interface { public: NWM_UDS(); diff --git a/src/core/hle/service/nwm/uds_beacon.h b/src/core/hle/service/nwm/uds_beacon.h index caacf4c6f..c726b04d9 100644 --- a/src/core/hle/service/nwm/uds_beacon.h +++ b/src/core/hle/service/nwm/uds_beacon.h @@ -17,17 +17,6 @@ namespace NWM { using MacAddress = std::array; constexpr std::array NintendoOUI = {0x00, 0x1F, 0x32}; -/// Additional block tag ids in the Beacon frames -enum class TagId : u8 { - SSID = 0, - SupportedRates = 1, - DSParameterSet = 2, - TrafficIndicationMap = 5, - CountryInformation = 7, - ERPInformation = 42, - VendorSpecific = 221 -}; - /** * Internal vendor-specific tag ids as stored inside * VendorSpecific blocks in the Beacon frames. diff --git a/src/core/hle/service/nwm/uds_connection.cpp b/src/core/hle/service/nwm/uds_connection.cpp new file mode 100644 index 000000000..c8a76ec2a --- /dev/null +++ b/src/core/hle/service/nwm/uds_connection.cpp @@ -0,0 +1,79 @@ +// Copyright 2017 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "core/hle/service/nwm/nwm_uds.h" +#include "core/hle/service/nwm/uds_connection.h" +#include "fmt/format.h" + +namespace Service { +namespace NWM { + +// Note: These values were taken from a packet capture of an o3DS XL +// broadcasting a Super Smash Bros. 4 lobby. +constexpr u16 DefaultExtraCapabilities = 0x0431; + +std::vector GenerateAuthenticationFrame(AuthenticationSeq seq) { + AuthenticationFrame frame{}; + frame.auth_seq = static_cast(seq); + + std::vector data(sizeof(frame)); + std::memcpy(data.data(), &frame, sizeof(frame)); + + return data; +} + +AuthenticationSeq GetAuthenticationSeqNumber(const std::vector& body) { + AuthenticationFrame frame; + std::memcpy(&frame, body.data(), sizeof(frame)); + + return static_cast(frame.auth_seq); +} + +/** + * Generates an SSID tag of an 802.11 Beacon frame with an 8-byte character representation of the + * specified network id as the SSID value. + * @param network_id The network id to use. + * @returns A buffer with the SSID tag. + */ +static std::vector GenerateSSIDTag(u32 network_id) { + constexpr u8 SSIDSize = 8; + + struct { + u8 id = static_cast(TagId::SSID); + u8 size = SSIDSize; + } tag_header; + + std::vector buffer(sizeof(tag_header) + SSIDSize); + + std::memcpy(buffer.data(), &tag_header, sizeof(tag_header)); + + std::string network_name = fmt::format("{0:08X}", network_id); + + std::memcpy(buffer.data() + sizeof(tag_header), network_name.c_str(), SSIDSize); + + return buffer; +} + +std::vector GenerateAssocResponseFrame(AssocStatus status, u16 association_id, u32 network_id) { + AssociationResponseFrame frame{}; + frame.capabilities = DefaultExtraCapabilities; + frame.status_code = static_cast(status); + // The association id is ORed with this magic value (0xC000) + constexpr u16 AssociationIdMagic = 0xC000; + frame.assoc_id = association_id | AssociationIdMagic; + + std::vector data(sizeof(frame)); + std::memcpy(data.data(), &frame, sizeof(frame)); + + auto ssid_tag = GenerateSSIDTag(network_id); + data.insert(data.end(), ssid_tag.begin(), ssid_tag.end()); + + // TODO(Subv): Add the SupportedRates tag. + // TODO(Subv): Add the DSParameterSet tag. + // TODO(Subv): Add the ERPInformation tag. + return data; +} + +} // namespace NWM +} // namespace Service diff --git a/src/core/hle/service/nwm/uds_connection.h b/src/core/hle/service/nwm/uds_connection.h new file mode 100644 index 000000000..73f55a4fd --- /dev/null +++ b/src/core/hle/service/nwm/uds_connection.h @@ -0,0 +1,51 @@ +// Copyright 2017 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include +#include "common/common_types.h" +#include "common/swap.h" +#include "core/hle/service/service.h" + +namespace Service { +namespace NWM { + +/// Sequence number of the 802.11 authentication frames. +enum class AuthenticationSeq : u16 { SEQ1 = 1, SEQ2 = 2 }; + +enum class AuthAlgorithm : u16 { OpenSystem = 0 }; + +enum class AuthStatus : u16 { Successful = 0 }; + +enum class AssocStatus : u16 { Successful = 0 }; + +struct AuthenticationFrame { + u16_le auth_algorithm = static_cast(AuthAlgorithm::OpenSystem); + u16_le auth_seq; + u16_le status_code = static_cast(AuthStatus::Successful); +}; + +static_assert(sizeof(AuthenticationFrame) == 6, "AuthenticationFrame has wrong size"); + +struct AssociationResponseFrame { + u16_le capabilities; + u16_le status_code; + u16_le assoc_id; +}; + +static_assert(sizeof(AssociationResponseFrame) == 6, "AssociationResponseFrame has wrong size"); + +/// Generates an 802.11 authentication frame, starting at the frame body. +std::vector GenerateAuthenticationFrame(AuthenticationSeq seq); + +/// Returns the sequence number from the body of an Authentication frame. +AuthenticationSeq GetAuthenticationSeqNumber(const std::vector& body); + +/// Generates an 802.11 association response frame with the specified status, association id and +/// network id, starting at the frame body. +std::vector GenerateAssocResponseFrame(AssocStatus status, u16 association_id, u32 network_id); + +} // namespace NWM +} // namespace Service From 2e9f544ecc9a01ff59859b43d65c61a2838e7c34 Mon Sep 17 00:00:00 2001 From: Subv Date: Mon, 17 Jul 2017 09:39:12 -0500 Subject: [PATCH 16/31] Services/UDS: Store the received beacon frames until RecvBeaconBroadcastData is called, up to 15 beacons at the same time, removing any older beacon frames when the limit is exceeded. --- src/core/hle/service/nwm/nwm_uds.cpp | 65 ++++++++++++++++++++++++++-- 1 file changed, 62 insertions(+), 3 deletions(-) diff --git a/src/core/hle/service/nwm/nwm_uds.cpp b/src/core/hle/service/nwm/nwm_uds.cpp index 6dbdff044..8fdf160ff 100644 --- a/src/core/hle/service/nwm/nwm_uds.cpp +++ b/src/core/hle/service/nwm/nwm_uds.cpp @@ -4,6 +4,7 @@ #include #include +#include #include #include #include "common/common_types.h" @@ -15,8 +16,10 @@ #include "core/hle/result.h" #include "core/hle/service/nwm/nwm_uds.h" #include "core/hle/service/nwm/uds_beacon.h" +#include "core/hle/service/nwm/uds_connection.h" #include "core/hle/service/nwm/uds_data.h" #include "core/memory.h" +#include "network/network.h" namespace Service { namespace NWM { @@ -51,6 +54,52 @@ static NetworkInfo network_info; // Event that will generate and send the 802.11 beacon frames. static int beacon_broadcast_event; +// Mutex to synchronize access to the list of received beacons between the emulation thread and the +// network thread. +static std::mutex beacon_mutex; + +// Number of beacons to store before we start dropping the old ones. +// TODO(Subv): Find a more accurate value for this limit. +constexpr size_t MaxBeaconFrames = 15; + +// List of the last beacons received from the network. +static std::deque received_beacons; + +/** + * Returns a list of received 802.11 beacon frames from the specified sender since the last call. + */ +std::deque GetReceivedBeacons(const MacAddress& sender) { + std::lock_guard lock(beacon_mutex); + // TODO(Subv): Filter by sender. + return std::move(received_beacons); +} + +/// Sends a WifiPacket to the room we're currently connected to. +void SendPacket(Network::WifiPacket& packet) { + // TODO(Subv): Implement. +} + +// Inserts the received beacon frame in the beacon queue and removes any older beacons if the size +// limit is exceeded. +void HandleBeaconFrame(const Network::WifiPacket& packet) { + std::lock_guard lock(beacon_mutex); + + received_beacons.emplace_back(packet); + + // Discard old beacons if the buffer is full. + if (received_beacons.size() > MaxBeaconFrames) + received_beacons.pop_front(); +} + +/// Callback to parse and handle a received wifi packet. +void OnWifiPacketReceived(const Network::WifiPacket& packet) { + switch (packet.type) { + case Network::WifiPacket::PacketType::Beacon: + HandleBeaconFrame(packet); + break; + } +} + /** * NWM_UDS::Shutdown service function * Inputs: @@ -111,8 +160,7 @@ static void RecvBeaconBroadcastData(Interface* self) { u32 total_size = sizeof(BeaconDataReplyHeader); // Retrieve all beacon frames that were received from the desired mac address. - std::deque beacons = - GetReceivedPackets(WifiPacket::PacketType::Beacon, mac_address); + auto beacons = GetReceivedBeacons(mac_address); BeaconDataReplyHeader data_reply_header{}; data_reply_header.total_entries = beacons.size(); @@ -193,6 +241,9 @@ static void InitializeWithVersion(Interface* self) { rb.Push(RESULT_SUCCESS); rb.PushCopyHandles(Kernel::g_handle_table.Create(connection_status_event).Unwrap()); + // TODO(Subv): Connect the OnWifiPacketReceived function to the wifi packet received callback of + // the room we're currently in. + LOG_DEBUG(Service_NWM, "called sharedmem_size=0x%08X, version=0x%08X, sharedmem_handle=0x%08X", sharedmem_size, version, sharedmem_handle); } @@ -610,9 +661,17 @@ static void BeaconBroadcastCallback(u64 userdata, int cycles_late) { if (connection_status.status != static_cast(NetworkStatus::ConnectedAsHost)) return; - // TODO(Subv): Actually send the beacon. std::vector frame = GenerateBeaconFrame(network_info, node_info); + using Network::WifiPacket; + WifiPacket packet; + packet.type = WifiPacket::PacketType::Beacon; + packet.data = std::move(frame); + packet.destination_address = Network::BroadcastMac; + packet.channel = network_channel; + + SendPacket(packet); + // Start broadcasting the network, send a beacon frame every 102.4ms. CoreTiming::ScheduleEvent(msToCycles(DefaultBeaconInterval * MillisecondsPerTU) - cycles_late, beacon_broadcast_event, 0); From d088dbfbe1064bb5212e83c50e71e4b2ea5b00cd Mon Sep 17 00:00:00 2001 From: Subv Date: Mon, 17 Jul 2017 09:51:03 -0500 Subject: [PATCH 17/31] Services/UDS: Handle the connection sequence packets. There is currently no stage tracking, a client is considered "Connected" when it receives the EAPoL Logoff packet from the server, this is not yet implemented. --- src/core/hle/service/nwm/nwm_uds.cpp | 100 ++++++++++++++++++++++----- 1 file changed, 83 insertions(+), 17 deletions(-) diff --git a/src/core/hle/service/nwm/nwm_uds.cpp b/src/core/hle/service/nwm/nwm_uds.cpp index 8fdf160ff..893bbb1e7 100644 --- a/src/core/hle/service/nwm/nwm_uds.cpp +++ b/src/core/hle/service/nwm/nwm_uds.cpp @@ -91,12 +91,95 @@ void HandleBeaconFrame(const Network::WifiPacket& packet) { received_beacons.pop_front(); } +/* + * Returns an available index in the nodes array for the + * currently-hosted UDS network. + */ +static u16 GetNextAvailableNodeId() { + ASSERT_MSG(connection_status.status == static_cast(NetworkStatus::ConnectedAsHost), + "Can not accept clients if we're not hosting a network"); + + for (u16 index = 0; index < connection_status.max_nodes; ++index) { + if ((connection_status.node_bitmask & (1 << index)) == 0) + return index; + } + + // Any connection attempts to an already full network should have been refused. + ASSERT_MSG(false, "No available connection slots in the network"); +} + +/* + * Start a connection sequence with an UDS server. The sequence starts by sending an 802.11 + * authentication frame with SEQ1. + */ +void StartConnectionSequence(const MacAddress& server) { + ASSERT(connection_status.status == static_cast(NetworkStatus::NotConnected)); + + // TODO(Subv): Handle timeout. + + // Send an authentication frame with SEQ1 + using Network::WifiPacket; + WifiPacket auth_request; + auth_request.channel = network_channel; + auth_request.data = GenerateAuthenticationFrame(AuthenticationSeq::SEQ1); + auth_request.destination_address = server; + auth_request.type = WifiPacket::PacketType::Authentication; + + SendPacket(auth_request); +} + +/// Sends an Association Response frame to the specified mac address +void SendAssociationResponseFrame(const MacAddress& address) { + ASSERT_MSG(connection_status.status == static_cast(NetworkStatus::ConnectedAsHost)); + + using Network::WifiPacket; + WifiPacket assoc_response; + assoc_response.channel = network_channel; + // TODO(Subv): This will cause multiple clients to end up with the same association id, but + // we're not using that for anything. + u16 association_id = 1; + assoc_response.data = GenerateAssocResponseFrame(AssocStatus::Successful, association_id, + network_info.network_id); + assoc_response.destination_address = address; + assoc_response.type = WifiPacket::PacketType::AssociationResponse; + + SendPacket(assoc_response); +} + +/* + * Handles the authentication request frame and sends the authentication response and association + * response frames. Once an Authentication frame with SEQ1 is received by the server, it responds + * with an Authentication frame containing SEQ2, and immediately sends an Association response frame + * containing the details of the access point and the assigned association id for the new client. + */ +void HandleAuthenticationFrame(const Network::WifiPacket& packet) { + // Only the SEQ1 auth frame is handled here, the SEQ2 frame doesn't need any special behavior + if (GetAuthenticationSeqNumber(packet.data) == AuthenticationSeq::SEQ1) { + ASSERT_MSG(connection_status.status == static_cast(NetworkStatus::ConnectedAsHost)); + + // Respond with an authentication response frame with SEQ2 + using Network::WifiPacket; + WifiPacket auth_request; + auth_request.channel = network_channel; + auth_request.data = GenerateAuthenticationFrame(AuthenticationSeq::SEQ2); + auth_request.destination_address = packet.transmitter_address; + auth_request.type = WifiPacket::PacketType::Authentication; + + SendPacket(auth_request); + + SendAssociationResponseFrame(packet.transmitter_address); + } +} + /// Callback to parse and handle a received wifi packet. void OnWifiPacketReceived(const Network::WifiPacket& packet) { switch (packet.type) { case Network::WifiPacket::PacketType::Beacon: HandleBeaconFrame(packet); break; + case Network::WifiPacket::PacketType::Authentication: + HandleAuthenticationFrame(packet); + break; } } @@ -677,23 +760,6 @@ static void BeaconBroadcastCallback(u64 userdata, int cycles_late) { beacon_broadcast_event, 0); } -/* - * Returns an available index in the nodes array for the - * currently-hosted UDS network. - */ -static u32 GetNextAvailableNodeId() { - ASSERT_MSG(connection_status.status == static_cast(NetworkStatus::ConnectedAsHost), - "Can not accept clients if we're not hosting a network"); - - for (unsigned index = 0; index < connection_status.max_nodes; ++index) { - if ((connection_status.node_bitmask & (1 << index)) == 0) - return index; - } - - // Any connection attempts to an already full network should have been refused. - ASSERT_MSG(false, "No available connection slots in the network"); -} - /* * Called when a client connects to an UDS network we're hosting, * updates the connection status and signals the update event. From f64cd87604b7a760e2832c76938d83ec6a284b22 Mon Sep 17 00:00:00 2001 From: Subv Date: Mon, 17 Jul 2017 09:51:40 -0500 Subject: [PATCH 18/31] Services/UDS: Remove an old duplicated declaration of WifiPacket. --- src/core/hle/service/nwm/uds_beacon.cpp | 3 --- src/core/hle/service/nwm/uds_beacon.h | 19 ------------------- 2 files changed, 22 deletions(-) diff --git a/src/core/hle/service/nwm/uds_beacon.cpp b/src/core/hle/service/nwm/uds_beacon.cpp index 6332b404c..552eaf65e 100644 --- a/src/core/hle/service/nwm/uds_beacon.cpp +++ b/src/core/hle/service/nwm/uds_beacon.cpp @@ -325,8 +325,5 @@ std::vector GenerateBeaconFrame(const NetworkInfo& network_info, const NodeL return buffer; } -std::deque GetReceivedPackets(WifiPacket::PacketType type, const MacAddress& sender) { - return {}; -} } // namespace NWM } // namespace Service diff --git a/src/core/hle/service/nwm/uds_beacon.h b/src/core/hle/service/nwm/uds_beacon.h index c726b04d9..50cc76da2 100644 --- a/src/core/hle/service/nwm/uds_beacon.h +++ b/src/core/hle/service/nwm/uds_beacon.h @@ -124,20 +124,6 @@ struct BeaconData { static_assert(sizeof(BeaconData) == 0x12, "BeaconData has incorrect size."); -/// Information about a received WiFi packet. -/// Acts as our own 802.11 header. -struct WifiPacket { - enum class PacketType { Beacon, Data }; - - PacketType type; ///< The type of 802.11 frame, Beacon / Data. - - /// Raw 802.11 frame data, starting at the management frame header for management frames. - std::vector data; - MacAddress transmitter_address; ///< Mac address of the transmitter. - MacAddress destination_address; ///< Mac address of the receiver. - u8 channel; ///< WiFi channel where this frame was transmitted. -}; - /** * Decrypts the beacon data buffer for the network described by `network_info`. */ @@ -150,10 +136,5 @@ void DecryptBeaconData(const NetworkInfo& network_info, std::vector& buffer) */ std::vector GenerateBeaconFrame(const NetworkInfo& network_info, const NodeList& nodes); -/** - * Returns a list of received 802.11 frames from the specified sender - * matching the type since the last call. - */ -std::deque GetReceivedPackets(WifiPacket::PacketType type, const MacAddress& sender); } // namespace NWM } // namespace Service From 933508e2a2f7923cebc15d679b78933df8fb9ee5 Mon Sep 17 00:00:00 2001 From: MerryMage Date: Thu, 3 Aug 2017 12:22:51 +0100 Subject: [PATCH 19/31] interpolate: Interpolate on a frame-by-frame basis --- src/audio_core/hle/source.cpp | 49 ++++++++----------- src/audio_core/interpolate.cpp | 86 +++++++++++++++------------------- src/audio_core/interpolate.h | 27 ++++++----- 3 files changed, 74 insertions(+), 88 deletions(-) diff --git a/src/audio_core/hle/source.cpp b/src/audio_core/hle/source.cpp index 92484c526..de4e88cae 100644 --- a/src/audio_core/hle/source.cpp +++ b/src/audio_core/hle/source.cpp @@ -244,17 +244,27 @@ void Source::GenerateFrame() { break; } - const size_t size_to_copy = - std::min(state.current_buffer.size(), current_frame.size() - frame_position); - - std::copy(state.current_buffer.begin(), state.current_buffer.begin() + size_to_copy, - current_frame.begin() + frame_position); - state.current_buffer.erase(state.current_buffer.begin(), - state.current_buffer.begin() + size_to_copy); - - frame_position += size_to_copy; - state.next_sample_number += static_cast(size_to_copy); + switch (state.interpolation_mode) { + case InterpolationMode::None: + AudioInterp::None(state.interp_state, state.current_buffer, state.rate_multiplier, + current_frame, frame_position); + break; + case InterpolationMode::Linear: + AudioInterp::Linear(state.interp_state, state.current_buffer, state.rate_multiplier, + current_frame, frame_position); + break; + case InterpolationMode::Polyphase: + // TODO(merry): Implement polyphase interpolation + LOG_DEBUG(Audio_DSP, "Polyphase interpolation unimplemented; falling back to linear"); + AudioInterp::Linear(state.interp_state, state.current_buffer, state.rate_multiplier, + current_frame, frame_position); + break; + default: + UNIMPLEMENTED(); + break; + } } + state.next_sample_number += frame_position; state.filters.ProcessFrame(current_frame); } @@ -305,25 +315,6 @@ bool Source::DequeueBuffer() { return true; } - switch (state.interpolation_mode) { - case InterpolationMode::None: - state.current_buffer = - AudioInterp::None(state.interp_state, state.current_buffer, state.rate_multiplier); - break; - case InterpolationMode::Linear: - state.current_buffer = - AudioInterp::Linear(state.interp_state, state.current_buffer, state.rate_multiplier); - break; - case InterpolationMode::Polyphase: - // TODO(merry): Implement polyphase interpolation - state.current_buffer = - AudioInterp::Linear(state.interp_state, state.current_buffer, state.rate_multiplier); - break; - default: - UNIMPLEMENTED(); - break; - } - // the first playthrough starts at play_position, loops start at the beginning of the buffer state.current_sample_number = (!buf.has_played) ? buf.play_position : 0; state.next_sample_number = state.current_sample_number; diff --git a/src/audio_core/interpolate.cpp b/src/audio_core/interpolate.cpp index 8a5d4181a..16e68bc5c 100644 --- a/src/audio_core/interpolate.cpp +++ b/src/audio_core/interpolate.cpp @@ -13,74 +13,64 @@ namespace AudioInterp { constexpr u64 scale_factor = 1 << 24; constexpr u64 scale_mask = scale_factor - 1; -/// Here we step over the input in steps of rate_multiplier, until we consume all of the input. +/// Here we step over the input in steps of rate, until we consume all of the input. /// Three adjacent samples are passed to fn each step. template -static StereoBuffer16 StepOverSamples(State& state, const StereoBuffer16& input, - float rate_multiplier, Function fn) { - ASSERT(rate_multiplier > 0); +static void StepOverSamples(State& state, StereoBuffer16& input, float rate, + DSP::HLE::StereoFrame16& output, size_t& outputi, Function fn) { + ASSERT(rate > 0); - if (input.size() < 2) - return {}; + if (input.empty()) + return; - StereoBuffer16 output; - output.reserve(static_cast(input.size() / rate_multiplier)); + input.insert(input.begin(), {state.xn2, state.xn1}); - u64 step_size = static_cast(rate_multiplier * scale_factor); + const u64 step_size = static_cast(rate * scale_factor); + u64 fposition = state.fposition; + size_t inputi = 0; - u64 fposition = 0; - const u64 max_fposition = input.size() * scale_factor; + while (outputi < output.size()) { + inputi = static_cast(fposition / scale_factor); + + if (inputi + 2 >= input.size()) { + inputi = input.size() - 2; + break; + } - while (fposition < 1 * scale_factor) { u64 fraction = fposition & scale_mask; - - output.push_back(fn(fraction, state.xn2, state.xn1, input[0])); + output[outputi++] = fn(fraction, input[inputi], input[inputi + 1], input[inputi + 2]); fposition += step_size; } - while (fposition < 2 * scale_factor) { - u64 fraction = fposition & scale_mask; + state.xn2 = input[inputi]; + state.xn1 = input[inputi + 1]; + state.fposition = fposition - inputi * scale_factor; - output.push_back(fn(fraction, state.xn1, input[0], input[1])); - - fposition += step_size; - } - - while (fposition < max_fposition) { - u64 fraction = fposition & scale_mask; - - size_t index = static_cast(fposition / scale_factor); - output.push_back(fn(fraction, input[index - 2], input[index - 1], input[index])); - - fposition += step_size; - } - - state.xn2 = input[input.size() - 2]; - state.xn1 = input[input.size() - 1]; - - return output; + input.erase(input.begin(), input.begin() + inputi + 2); } -StereoBuffer16 None(State& state, const StereoBuffer16& input, float rate_multiplier) { - return StepOverSamples( - state, input, rate_multiplier, +void None(State& state, StereoBuffer16& input, float rate, DSP::HLE::StereoFrame16& output, + size_t& outputi) { + StepOverSamples( + state, input, rate, output, outputi, [](u64 fraction, const auto& x0, const auto& x1, const auto& x2) { return x0; }); } -StereoBuffer16 Linear(State& state, const StereoBuffer16& input, float rate_multiplier) { +void Linear(State& state, StereoBuffer16& input, float rate, DSP::HLE::StereoFrame16& output, + size_t& outputi) { // Note on accuracy: Some values that this produces are +/- 1 from the actual firmware. - return StepOverSamples(state, input, rate_multiplier, - [](u64 fraction, const auto& x0, const auto& x1, const auto& x2) { - // This is a saturated subtraction. (Verified by black-box fuzzing.) - s64 delta0 = MathUtil::Clamp(x1[0] - x0[0], -32768, 32767); - s64 delta1 = MathUtil::Clamp(x1[1] - x0[1], -32768, 32767); + StepOverSamples(state, input, rate, output, outputi, + [](u64 fraction, const auto& x0, const auto& x1, const auto& x2) { + // This is a saturated subtraction. (Verified by black-box fuzzing.) + s64 delta0 = MathUtil::Clamp(x1[0] - x0[0], -32768, 32767); + s64 delta1 = MathUtil::Clamp(x1[1] - x0[1], -32768, 32767); - return std::array{ - static_cast(x0[0] + fraction * delta0 / scale_factor), - static_cast(x0[1] + fraction * delta1 / scale_factor), - }; - }); + return std::array{ + static_cast(x0[0] + fraction * delta0 / scale_factor), + static_cast(x0[1] + fraction * delta1 / scale_factor), + }; + }); } } // namespace AudioInterp diff --git a/src/audio_core/interpolate.h b/src/audio_core/interpolate.h index 19a7b66cb..59f59bc14 100644 --- a/src/audio_core/interpolate.h +++ b/src/audio_core/interpolate.h @@ -6,6 +6,7 @@ #include #include +#include "audio_core/hle/common.h" #include "common/common_types.h" namespace AudioInterp { @@ -14,31 +15,35 @@ namespace AudioInterp { using StereoBuffer16 = std::vector>; struct State { - // Two historical samples. + /// Two historical samples. std::array xn1 = {}; ///< x[n-1] std::array xn2 = {}; ///< x[n-2] + /// Current fractional position. + u64 fposition = 0; }; /** * No interpolation. This is equivalent to a zero-order hold. There is a two-sample predelay. * @param state Interpolation state. * @param input Input buffer. - * @param rate_multiplier Stretch factor. Must be a positive non-zero value. - * rate_multiplier > 1.0 performs decimation and rate_multipler < 1.0 - * performs upsampling. - * @return The resampled audio buffer. + * @param rate Stretch factor. Must be a positive non-zero value. + * rate > 1.0 performs decimation and rate < 1.0 performs upsampling. + * @param output The resampled audio buffer. + * @param outputi The index of output to start writing to. */ -StereoBuffer16 None(State& state, const StereoBuffer16& input, float rate_multiplier); +void None(State& state, StereoBuffer16& input, float rate, DSP::HLE::StereoFrame16& output, + size_t& outputi); /** * Linear interpolation. This is equivalent to a first-order hold. There is a two-sample predelay. * @param state Interpolation state. * @param input Input buffer. - * @param rate_multiplier Stretch factor. Must be a positive non-zero value. - * rate_multiplier > 1.0 performs decimation and rate_multipler < 1.0 - * performs upsampling. - * @return The resampled audio buffer. + * @param rate Stretch factor. Must be a positive non-zero value. + * rate > 1.0 performs decimation and rate < 1.0 performs upsampling. + * @param output The resampled audio buffer. + * @param outputi The index of output to start writing to. */ -StereoBuffer16 Linear(State& state, const StereoBuffer16& input, float rate_multiplier); +void Linear(State& state, StereoBuffer16& input, float rate, DSP::HLE::StereoFrame16& output, + size_t& outputi); } // namespace AudioInterp From 826606479682234c98e4dfa6e616e637a28d4fcc Mon Sep 17 00:00:00 2001 From: danzel Date: Tue, 29 Aug 2017 20:39:55 +1200 Subject: [PATCH 20/31] Use recursive_mutex instead of mutex to fix #2902 --- src/core/hle/lock.cpp | 2 +- src/core/hle/lock.h | 2 +- src/core/hle/svc.cpp | 2 +- src/core/memory.cpp | 4 ++-- 4 files changed, 5 insertions(+), 5 deletions(-) diff --git a/src/core/hle/lock.cpp b/src/core/hle/lock.cpp index 082f689c8..1c24c7ce9 100644 --- a/src/core/hle/lock.cpp +++ b/src/core/hle/lock.cpp @@ -7,5 +7,5 @@ #include namespace HLE { -std::mutex g_hle_lock; +std::recursive_mutex g_hle_lock; } diff --git a/src/core/hle/lock.h b/src/core/hle/lock.h index 8265621e1..5c99fe996 100644 --- a/src/core/hle/lock.h +++ b/src/core/hle/lock.h @@ -14,5 +14,5 @@ namespace HLE { * to the emulated memory is not protected by this mutex, and should be avoided in any threads other * than the CPU thread. */ -extern std::mutex g_hle_lock; +extern std::recursive_mutex g_hle_lock; } // namespace HLE diff --git a/src/core/hle/svc.cpp b/src/core/hle/svc.cpp index b98938cb4..dfc36748c 100644 --- a/src/core/hle/svc.cpp +++ b/src/core/hle/svc.cpp @@ -1334,7 +1334,7 @@ void CallSVC(u32 immediate) { MICROPROFILE_SCOPE(Kernel_SVC); // Lock the global kernel mutex when we enter the kernel HLE. - std::lock_guard lock(HLE::g_hle_lock); + std::lock_guard lock(HLE::g_hle_lock); const FunctionDef* info = GetSVCInfo(immediate); if (info) { diff --git a/src/core/memory.cpp b/src/core/memory.cpp index a3c5f4a9d..097bc5b47 100644 --- a/src/core/memory.cpp +++ b/src/core/memory.cpp @@ -183,7 +183,7 @@ T Read(const VAddr vaddr) { } // The memory access might do an MMIO or cached access, so we have to lock the HLE kernel state - std::lock_guard lock(HLE::g_hle_lock); + std::lock_guard lock(HLE::g_hle_lock); PageType type = current_page_table->attributes[vaddr >> PAGE_BITS]; switch (type) { @@ -224,7 +224,7 @@ void Write(const VAddr vaddr, const T data) { } // The memory access might do an MMIO or cached access, so we have to lock the HLE kernel state - std::lock_guard lock(HLE::g_hle_lock); + std::lock_guard lock(HLE::g_hle_lock); PageType type = current_page_table->attributes[vaddr >> PAGE_BITS]; switch (type) { From e2c41a589198ff3162da8047a4c33162b02b0f2b Mon Sep 17 00:00:00 2001 From: wwylele Date: Thu, 31 Aug 2017 12:24:00 +0300 Subject: [PATCH 21/31] video_core: report telemetry for gas mode --- src/video_core/renderer_opengl/gl_shader_gen.cpp | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp index 3f390491a..c8fc7a0ff 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.cpp +++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp @@ -8,6 +8,7 @@ #include "common/assert.h" #include "common/bit_field.h" #include "common/logging/log.h" +#include "core/core.h" #include "video_core/regs_framebuffer.h" #include "video_core/regs_lighting.h" #include "video_core/regs_rasterizer.h" @@ -1155,6 +1156,11 @@ vec4 secondary_fragment_color = vec4(0.0); // Blend the fog out += "last_tex_env_out.rgb = mix(fog_color.rgb, last_tex_env_out.rgb, fog_factor);\n"; + } else if (state.fog_mode == TexturingRegs::FogMode::Gas) { + Core::Telemetry().AddField(Telemetry::FieldType::Session, "VideoCore_Pica_UseGasMode", + true); + LOG_CRITICAL(Render_OpenGL, "Unimplemented gas mode"); + UNIMPLEMENTED(); } out += "gl_FragDepth = depth;\n"; From ab47bf6ad65b870f447fe2c1a39f3c288b906c3d Mon Sep 17 00:00:00 2001 From: James Rowe Date: Wed, 30 Aug 2017 19:39:24 -0600 Subject: [PATCH 22/31] Build: Add mingw64 compile support to appveyor Releases will be built with both mingw and msvc and the binaries of both builds will be uploaded to github releases --- appveyor.yml | 161 ++++++++++++++++++++++++++++++++++++++------------- 1 file changed, 122 insertions(+), 39 deletions(-) diff --git a/appveyor.yml b/appveyor.yml index 94e9969f5..1c390cdd5 100644 --- a/appveyor.yml +++ b/appveyor.yml @@ -7,6 +7,15 @@ cache: os: Visual Studio 2017 +environment: + # Tell msys2 to add mingw64 to the path + MSYSTEM: MINGW64 + # Tell msys2 to inherit the current directory when starting the shell + CHERE_INVOKING: 1 + matrix: + - BUILD_TYPE: mingw + - BUILD_TYPE: msvc + platform: - x64 @@ -15,72 +24,146 @@ configuration: install: - git submodule update --init --recursive + - ps: | + if ($env:BUILD_TYPE -eq 'mingw') { + $dependencies = "mingw64/mingw-w64-x86_64-cmake", + "mingw64/mingw-w64-x86_64-qt5", + "mingw64/mingw-w64-x86_64-curl", + "mingw64/mingw-w64-x86_64-SDL2" + # redirect err to null to prevent warnings from becoming errors + # workaround to prevent pacman from failing due to cyclical dependencies + C:\msys64\usr\bin\bash -lc "pacman --noconfirm -S mingw64/mingw-w64-x86_64-freetype mingw64/mingw-w64-x86_64-fontconfig" 2> $null + C:\msys64\usr\bin\bash -lc "pacman --noconfirm -S $dependencies" 2> $null + } before_build: - - mkdir build - - cd build - - cmake -G "Visual Studio 15 2017 Win64" -DCITRA_USE_BUNDLED_QT=1 -DCITRA_USE_BUNDLED_SDL2=1 -DCMAKE_USE_OPENSSL=0 .. + - mkdir %BUILD_TYPE%_build + - cd %BUILD_TYPE%_build + - ps: | + if ($env:BUILD_TYPE -eq 'msvc') { + # redirect stderr and change the exit code to prevent powershell from cancelling the build if cmake prints a warning + cmd /C 'cmake -G "Visual Studio 15 2017 Win64" -DCITRA_USE_BUNDLED_QT=1 -DCITRA_USE_BUNDLED_SDL2=1 -DCMAKE_USE_OPENSSL=0 .. 2>&1 && exit 0' + } else { + C:\msys64\usr\bin\bash.exe -lc "cmake -G 'MSYS Makefiles' -DUSE_SYSTEM_CURL=1 -DCMAKE_BUILD_TYPE=Release .. 2>&1" + } - cd .. -build: - project: build/citra.sln - parallel: true +build_script: + - ps: | + if ($env:BUILD_TYPE -eq 'msvc') { + # https://www.appveyor.com/docs/build-phase + msbuild msvc_build/citra.sln /maxcpucount /logger:"C:\Program Files\AppVeyor\BuildAgent\Appveyor.MSBuildLogger.dll" + } else { + C:\msys64\usr\bin\bash.exe -lc 'mingw32-make -C mingw_build/ 2>&1' + } after_build: - ps: | $GITDATE = $(git show -s --date=short --format='%ad') -replace "-","" $GITREV = $(git show -s --format='%h') - $GIT_LONG_HASH = $(git rev-parse HEAD) - # Where are these spaces coming from? Regardless, let's remove them - $MSVC_BUILD_NAME = "citra-windows-msvc-$GITDATE-$GITREV.zip" -replace " ", "" - $MSVC_BUILD_PDB = "citra-windows-msvc-$GITDATE-$GITREV-debugsymbols.zip" -replace " ", "" - $MSVC_SEVENZIP = "citra-windows-msvc-$GITDATE-$GITREV.7z" -replace " ", "" - $BINTRAY_VERSION = "nightly-$GIT_LONG_HASH" -replace " ", "" - - # set the build names as env vars so the artifacts can upload them - $env:MSVC_BUILD_NAME = $MSVC_BUILD_NAME - $env:MSVC_BUILD_PDB = $MSVC_BUILD_PDB - $env:MSVC_SEVENZIP = $MSVC_SEVENZIP - $env:GITREV = $GITREV - - 7z a -tzip $MSVC_BUILD_PDB .\build\bin\release\*.pdb - rm .\build\bin\release\*.pdb # Find out which kind of release we are producing by tag name if ($env:APPVEYOR_REPO_TAG_NAME) { - $RELEASE_DIST, $RELEASE_VERSION = $env:APPVEYOR_REPO_TAG_NAME.split('-') + $RELEASE_DIST, $RELEASE_VERSION = $env:APPVEYOR_REPO_TAG_NAME.split('-') } else { - # There is no repo tag - make assumptions - $RELEASE_DIST = "head" + # There is no repo tag - make assumptions + $RELEASE_DIST = "head" } - mkdir $RELEASE_DIST - Copy-Item .\build\bin\release\* -Destination $RELEASE_DIST -Recurse - Copy-Item .\license.txt -Destination $RELEASE_DIST - Copy-Item .\README.md -Destination $RELEASE_DIST + if ($env:BUILD_TYPE -eq 'msvc') { + # Where are these spaces coming from? Regardless, let's remove them + $MSVC_BUILD_ZIP = "citra-windows-msvc-$GITDATE-$GITREV.zip" -replace " ", "" + $MSVC_BUILD_PDB = "citra-windows-msvc-$GITDATE-$GITREV-debugsymbols.zip" -replace " ", "" + $MSVC_SEVENZIP = "citra-windows-msvc-$GITDATE-$GITREV.7z" -replace " ", "" - 7z a -tzip $MSVC_BUILD_NAME $RELEASE_DIST\* - 7z a $MSVC_SEVENZIP $RELEASE_DIST + # set the build names as env vars so the artifacts can upload them + $env:BUILD_ZIP = $MSVC_BUILD_ZIP + $env:BUILD_SYMBOLS = $MSVC_BUILD_PDB + $env:BUILD_UPDATE = $MSVC_SEVENZIP + + 7z a -tzip $MSVC_BUILD_PDB .\msvc_build\bin\release\*.pdb + rm .\msvc_build\bin\release\*.pdb + + mkdir $RELEASE_DIST + Copy-Item .\msvc_build\bin\release\* -Destination $RELEASE_DIST -Recurse + Copy-Item .\license.txt -Destination $RELEASE_DIST + Copy-Item .\README.md -Destination $RELEASE_DIST + 7z a -tzip $MSVC_BUILD_ZIP $RELEASE_DIST\* + 7z a $MSVC_SEVENZIP $RELEASE_DIST + } else { + $MINGW_BUILD_ZIP = "citra-windows-mingw-$GITDATE-$GITREV.zip" -replace " ", "" + $MINGW_SEVENZIP = "citra-windows-mingw-$GITDATE-$GITREV.7z" -replace " ", "" + # not going to bother adding separate debug symbols for mingw, so just upload a README for it + # if someone wants to add them, change mingw to compile with -g and use objdump and strip to separate the symbols from the binary + $MINGW_NO_DEBUG_SYMBOLS = "README_No_Debug_Symbols.txt" + Set-Content -Path $MINGW_NO_DEBUG_SYMBOLS -Value "This is a workaround for Appveyor since msvc has debug symbols but mingw doesnt" -Force + + # store the build information in env vars so we can use them as artifacts + $env:BUILD_ZIP = $MINGW_BUILD_ZIP + $env:BUILD_SYMBOLS = $MINGW_NO_DEBUG_SYMBOLS + $env:BUILD_UPDATE = $MINGW_SEVENZIP + + $CMAKE_SOURCE_DIR = "$env:APPVEYOR_BUILD_FOLDER" + $CMAKE_BINARY_DIR = "$CMAKE_SOURCE_DIR/mingw_build" + $RELEASE_DIST = $RELEASE_DIST + "-mingw" + + mkdir $RELEASE_DIST + mkdir $RELEASE_DIST/platforms + + # copy the compiled binaries and other release files to the release folder + Get-ChildItem "$CMAKE_BINARY_DIR" -Recurse -Filter "citra*.exe" | Copy-Item -destination $RELEASE_DIST + Copy-Item -path "$CMAKE_SOURCE_DIR/license.txt" -destination $RELEASE_DIST + Copy-Item -path "$CMAKE_SOURCE_DIR/README.md" -destination $RELEASE_DIST + + # copy all the dll dependencies to the release folder + # hardcoded list because we don't build static and determining the list of dlls from the binary is a pain. + $MingwDLLs = "Qt5Core.dll","Qt5Widgets.dll","Qt5Gui.dll","Qt5OpenGL.dll", + # QT dll dependencies + "libbz2-*.dll","libicudt*.dll","libicuin*.dll","libicuuc*.dll","libffi-*.dll", + "libfreetype-*.dll","libglib-*.dll","libgobject-*.dll","libgraphite2.dll","libiconv-*.dll", + "libharfbuzz-*.dll","libintl-*.dll","libpcre-*.dll","libpcre16-*.dll","libpng16-*.dll", + # Runtime/Other dependencies + "libgcc_s_seh-*.dll","libstdc++-*.dll","libwinpthread-*.dll","SDL2.dll","zlib1.dll", + # curl dependencies + "libcurl-*.dll","libnghttp2-*.dll","libeay32.dll","libgmp-*.dll","librtmp-*.dll", + "libgnutls-*.dll","libhogweed-*.dll","libnettle-*.dll","libssh2-*.dll", + "ssleay32.dll","libidn-*.dll","libp11-kit-*.dll","libtasn1-*.dll","libunistring-*.dll" + foreach ($file in $MingwDLLs) { + Copy-Item -path "C:/msys64/mingw64/bin/$file" -force -destination "$RELEASE_DIST" + } + + # copy the qt windows plugin dll to platforms + Copy-Item -path "C:/msys64/mingw64/share/qt5/plugins/platforms/qwindows.dll" -force -destination "$RELEASE_DIST/platforms" + + 7z a -tzip $MINGW_BUILD_ZIP $RELEASE_DIST\* + 7z a $MINGW_SEVENZIP $RELEASE_DIST + } test_script: - - cd build && ctest -VV -C Release && cd .. + - cd %BUILD_TYPE%_build + - ps: | + if ($env:BUILD_TYPE -eq 'msvc') { + ctest -VV -C Release + } else { + C:\msys64\usr\bin\bash.exe -lc "ctest -VV -C Release" + } + - cd .. artifacts: - - path: $(MSVC_BUILD_NAME) - name: msvcbuild + - path: $(BUILD_ZIP) + name: build type: zip - - path: $(MSVC_BUILD_PDB) - name: msvcdebug - type: zip - - path: $(MSVC_SEVENZIP) - name: msvcupdate + - path: $(BUILD_SYMBOLS) + name: debugsymbols + - path: $(BUILD_UPDATE) + name: update deploy: provider: GitHub release: $(appveyor_repo_tag_name) auth_token: secure: "dbpsMC/MgPKWFNJCXpQl4cR8FYhepkPLjgNp/pRMktZ8oLKTqPYErfreaIxb/4P1" - artifact: msvcupdate,msvcbuild + artifact: update,build draft: false prerelease: false on: From 40505bc4fcc5cb1043b90b0acdce9e0093422921 Mon Sep 17 00:00:00 2001 From: DaMan Date: Thu, 31 Aug 2017 18:37:11 -0400 Subject: [PATCH 23/31] Add manifest --- CMakeLists.txt | 4 ++-- dist/citra.manifest | 24 ++++++++++++++++++++++++ src/citra/citra.rc | 8 ++++++++ src/citra_qt/citra-qt.rc | 8 ++++++++ 4 files changed, 42 insertions(+), 2 deletions(-) create mode 100644 dist/citra.manifest diff --git a/CMakeLists.txt b/CMakeLists.txt index ddba04ef9..f8060270e 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -129,8 +129,8 @@ else() set(CMAKE_C_FLAGS_RELEASE "/O2 /GS- /MD" CACHE STRING "" FORCE) set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE}" CACHE STRING "" FORCE) - set(CMAKE_EXE_LINKER_FLAGS_DEBUG "/DEBUG" CACHE STRING "" FORCE) - set(CMAKE_EXE_LINKER_FLAGS_RELEASE "/DEBUG /INCREMENTAL:NO /OPT:REF,ICF" CACHE STRING "" FORCE) + set(CMAKE_EXE_LINKER_FLAGS_DEBUG "/DEBUG /MANIFEST:NO" CACHE STRING "" FORCE) + set(CMAKE_EXE_LINKER_FLAGS_RELEASE "/DEBUG /MANIFEST:NO /INCREMENTAL:NO /OPT:REF,ICF" CACHE STRING "" FORCE) endif() # Set file offset size to 64 bits. diff --git a/dist/citra.manifest b/dist/citra.manifest new file mode 100644 index 000000000..fd30b656f --- /dev/null +++ b/dist/citra.manifest @@ -0,0 +1,24 @@ + + + + + + + + + + + + True/PM + true + + + + + + + + + + + \ No newline at end of file diff --git a/src/citra/citra.rc b/src/citra/citra.rc index fea603004..c490ef302 100644 --- a/src/citra/citra.rc +++ b/src/citra/citra.rc @@ -1,3 +1,4 @@ +#include "winresrc.h" ///////////////////////////////////////////////////////////////////////////// // // Icon @@ -7,3 +8,10 @@ // remains consistent on all systems. CITRA_ICON ICON "../../dist/citra.ico" + +///////////////////////////////////////////////////////////////////////////// +// +// RT_MANIFEST +// + +1 RT_MANIFEST "../../dist/citra.manifest" diff --git a/src/citra_qt/citra-qt.rc b/src/citra_qt/citra-qt.rc index fea603004..c490ef302 100644 --- a/src/citra_qt/citra-qt.rc +++ b/src/citra_qt/citra-qt.rc @@ -1,3 +1,4 @@ +#include "winresrc.h" ///////////////////////////////////////////////////////////////////////////// // // Icon @@ -7,3 +8,10 @@ // remains consistent on all systems. CITRA_ICON ICON "../../dist/citra.ico" + +///////////////////////////////////////////////////////////////////////////// +// +// RT_MANIFEST +// + +1 RT_MANIFEST "../../dist/citra.manifest" From 12fbc8c8dff3265b03cffdd5bb5e6dd6537cd824 Mon Sep 17 00:00:00 2001 From: wwylele Date: Sun, 27 Aug 2017 07:33:27 +0300 Subject: [PATCH 24/31] pica/lighting: only apply Fresnel factor for the last light --- src/video_core/renderer_opengl/gl_shader_gen.cpp | 9 +++++---- src/video_core/swrasterizer/lighting.cpp | 7 ++++--- 2 files changed, 9 insertions(+), 7 deletions(-) diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp index 3f390491a..b5f359da6 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.cpp +++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp @@ -750,7 +750,8 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) { } // Fresnel - if (lighting.lut_fr.enable && + // Note: only the last entry in the light slots applies the Fresnel factor + if (light_index == lighting.src_num - 1 && lighting.lut_fr.enable && LightingRegs::IsLightingSamplerSupported(lighting.config, LightingRegs::LightingSampler::Fresnel)) { // Lookup fresnel LUT value @@ -759,17 +760,17 @@ static void WriteLighting(std::string& out, const PicaShaderConfig& config) { lighting.lut_fr.type, lighting.lut_fr.abs_input); value = "(" + std::to_string(lighting.lut_fr.scale) + " * " + value + ")"; - // Enabled for difffuse lighting alpha component + // Enabled for diffuse lighting alpha component if (lighting.fresnel_selector == LightingRegs::LightingFresnelSelector::PrimaryAlpha || lighting.fresnel_selector == LightingRegs::LightingFresnelSelector::Both) { - out += "diffuse_sum.a *= " + value + ";\n"; + out += "diffuse_sum.a = " + value + ";\n"; } // Enabled for the specular lighting alpha component if (lighting.fresnel_selector == LightingRegs::LightingFresnelSelector::SecondaryAlpha || lighting.fresnel_selector == LightingRegs::LightingFresnelSelector::Both) { - out += "specular_sum.a *= " + value + ";\n"; + out += "specular_sum.a = " + value + ";\n"; } } diff --git a/src/video_core/swrasterizer/lighting.cpp b/src/video_core/swrasterizer/lighting.cpp index b38964530..5fa748611 100644 --- a/src/video_core/swrasterizer/lighting.cpp +++ b/src/video_core/swrasterizer/lighting.cpp @@ -230,7 +230,8 @@ std::tuple, Math::Vec4> ComputeFragmentsColors( d1_lut_value * refl_value * light_config.specular_1.ToVec3f(); // Fresnel - if (lighting.config1.disable_lut_fr == 0 && + // Note: only the last entry in the light slots applies the Fresnel factor + if (light_index == lighting.max_light_index && lighting.config1.disable_lut_fr == 0 && LightingRegs::IsLightingSamplerSupported(lighting.config0.config, LightingRegs::LightingSampler::Fresnel)) { @@ -242,14 +243,14 @@ std::tuple, Math::Vec4> ComputeFragmentsColors( if (lighting.config0.fresnel_selector == LightingRegs::LightingFresnelSelector::PrimaryAlpha || lighting.config0.fresnel_selector == LightingRegs::LightingFresnelSelector::Both) { - diffuse_sum.a() *= lut_value; + diffuse_sum.a() = lut_value; } // Enabled for the specular lighting alpha component if (lighting.config0.fresnel_selector == LightingRegs::LightingFresnelSelector::SecondaryAlpha || lighting.config0.fresnel_selector == LightingRegs::LightingFresnelSelector::Both) { - specular_sum.a() *= lut_value; + specular_sum.a() = lut_value; } } From 59a9aaf388b71444116a42fe97a969947230908e Mon Sep 17 00:00:00 2001 From: wwylele Date: Wed, 2 Aug 2017 22:56:44 +0300 Subject: [PATCH 25/31] APT: load different shared font depending on the region --- src/core/hle/service/apt/apt.cpp | 286 ++++++++++++++++--------------- src/core/hle/service/cfg/cfg.cpp | 2 +- src/core/hle/service/cfg/cfg.h | 2 + 3 files changed, 155 insertions(+), 135 deletions(-) diff --git a/src/core/hle/service/apt/apt.cpp b/src/core/hle/service/apt/apt.cpp index 58d94768c..8c0ba73f2 100644 --- a/src/core/hle/service/apt/apt.cpp +++ b/src/core/hle/service/apt/apt.cpp @@ -19,6 +19,7 @@ #include "core/hle/service/apt/apt_s.h" #include "core/hle/service/apt/apt_u.h" #include "core/hle/service/apt/bcfnt/bcfnt.h" +#include "core/hle/service/cfg/cfg.h" #include "core/hle/service/fs/archive.h" #include "core/hle/service/ptm/ptm.h" #include "core/hle/service/service.h" @@ -198,6 +199,143 @@ void Initialize(Service::Interface* self) { Kernel::g_handle_table.Create(slot_data->parameter_event).Unwrap()); } +static u32 DecompressLZ11(const u8* in, u8* out) { + u32_le decompressed_size; + memcpy(&decompressed_size, in, sizeof(u32)); + in += 4; + + u8 type = decompressed_size & 0xFF; + ASSERT(type == 0x11); + decompressed_size >>= 8; + + u32 current_out_size = 0; + u8 flags = 0, mask = 1; + while (current_out_size < decompressed_size) { + if (mask == 1) { + flags = *(in++); + mask = 0x80; + } else { + mask >>= 1; + } + + if (flags & mask) { + u8 byte1 = *(in++); + u32 length = byte1 >> 4; + u32 offset; + if (length == 0) { + u8 byte2 = *(in++); + u8 byte3 = *(in++); + length = (((byte1 & 0x0F) << 4) | (byte2 >> 4)) + 0x11; + offset = (((byte2 & 0x0F) << 8) | byte3) + 0x1; + } else if (length == 1) { + u8 byte2 = *(in++); + u8 byte3 = *(in++); + u8 byte4 = *(in++); + length = (((byte1 & 0x0F) << 12) | (byte2 << 4) | (byte3 >> 4)) + 0x111; + offset = (((byte3 & 0x0F) << 8) | byte4) + 0x1; + } else { + u8 byte2 = *(in++); + length = (byte1 >> 4) + 0x1; + offset = (((byte1 & 0x0F) << 8) | byte2) + 0x1; + } + + for (u32 i = 0; i < length; i++) { + *out = *(out - offset); + ++out; + } + + current_out_size += length; + } else { + *(out++) = *(in++); + current_out_size++; + } + } + return decompressed_size; +} + +static bool LoadSharedFont() { + u8 font_region_code; + switch (CFG::GetRegionValue()) { + case 4: // CHN + font_region_code = 2; + break; + case 5: // KOR + font_region_code = 3; + break; + case 6: // TWN + font_region_code = 4; + break; + default: // JPN/EUR/USA + font_region_code = 1; + break; + } + + const u64_le shared_font_archive_id_low = 0x0004009b00014002 | ((font_region_code - 1) << 8); + const u64_le shared_font_archive_id_high = 0x00000001ffffff00; + std::vector shared_font_archive_id(16); + std::memcpy(&shared_font_archive_id[0], &shared_font_archive_id_low, sizeof(u64)); + std::memcpy(&shared_font_archive_id[8], &shared_font_archive_id_high, sizeof(u64)); + FileSys::Path archive_path(shared_font_archive_id); + auto archive_result = Service::FS::OpenArchive(Service::FS::ArchiveIdCode::NCCH, archive_path); + if (archive_result.Failed()) + return false; + + std::vector romfs_path(20, 0); // 20-byte all zero path for opening RomFS + FileSys::Path file_path(romfs_path); + FileSys::Mode open_mode = {}; + open_mode.read_flag.Assign(1); + auto file_result = Service::FS::OpenFileFromArchive(*archive_result, file_path, open_mode); + if (file_result.Failed()) + return false; + + auto romfs = std::move(file_result).Unwrap(); + std::vector romfs_buffer(romfs->backend->GetSize()); + romfs->backend->Read(0, romfs_buffer.size(), romfs_buffer.data()); + romfs->backend->Close(); + + const char16_t* file_name[4] = {u"cbf_std.bcfnt.lz", u"cbf_zh-Hans-CN.bcfnt.lz", + u"cbf_ko-Hang-KR.bcfnt.lz", u"cbf_zh-Hant-TW.bcfnt.lz"}; + const u8* font_file = + RomFS::GetFilePointer(romfs_buffer.data(), {file_name[font_region_code - 1]}); + if (font_file == nullptr) + return false; + + struct { + u32_le status; + u32_le region; + u32_le decompressed_size; + INSERT_PADDING_WORDS(0x1D); + } shared_font_header{}; + static_assert(sizeof(shared_font_header) == 0x80, "shared_font_header has incorrect size"); + + shared_font_header.status = 2; // successfully loaded + shared_font_header.region = font_region_code; + shared_font_header.decompressed_size = + DecompressLZ11(font_file, shared_font_mem->GetPointer(0x80)); + std::memcpy(shared_font_mem->GetPointer(), &shared_font_header, sizeof(shared_font_header)); + *shared_font_mem->GetPointer(0x83) = 'U'; // Change the magic from "CFNT" to "CFNU" + + return true; +} + +static bool LoadLegacySharedFont() { + // This is the legacy method to load shared font. + // The expected format is a decrypted, uncompressed BCFNT file with the 0x80 byte header + // generated by the APT:U service. The best way to get is by dumping it from RAM. We've provided + // a homebrew app to do this: https://github.com/citra-emu/3dsutils. Put the resulting file + // "shared_font.bin" in the Citra "sysdata" directory. + std::string filepath = FileUtil::GetUserPath(D_SYSDATA_IDX) + SHARED_FONT; + + FileUtil::CreateFullPath(filepath); // Create path if not already created + FileUtil::IOFile file(filepath, "rb"); + if (file.IsOpen()) { + file.ReadBytes(shared_font_mem->GetPointer(), file.GetSize()); + return true; + } + + return false; +} + void GetSharedFont(Service::Interface* self) { IPC::RequestParser rp(Kernel::GetCommandBuffer(), 0x44, 0, 0); // 0x00440000 IPC::RequestBuilder rb = rp.MakeBuilder(2, 2); @@ -206,11 +344,20 @@ void GetSharedFont(Service::Interface* self) { Core::Telemetry().AddField(Telemetry::FieldType::Session, "RequiresSharedFont", true); if (!shared_font_loaded) { - LOG_ERROR(Service_APT, "shared font file missing - go dump it from your 3ds"); - rb.Push(-1); // TODO: Find the right error code - rb.Skip(1 + 2, true); - Core::System::GetInstance().SetStatus(Core::System::ResultStatus::ErrorSharedFont); - return; + // On real 3DS, font loading happens on booting. However, we load it on demand to coordinate + // with CFG region auto configuration, which happens later than APT initialization. + if (LoadSharedFont()) { + shared_font_loaded = true; + } else if (LoadLegacySharedFont()) { + LOG_WARNING(Service_APT, "Loaded shared font by legacy method"); + shared_font_loaded = true; + } else { + LOG_ERROR(Service_APT, "shared font file missing - go dump it from your 3ds"); + rb.Push(-1); // TODO: Find the right error code + rb.Skip(1 + 2, true); + Core::System::GetInstance().SetStatus(Core::System::ResultStatus::ErrorSharedFont); + return; + } } // The shared font has to be relocated to the new address before being passed to the @@ -863,125 +1010,6 @@ void CheckNew3DS(Service::Interface* self) { LOG_WARNING(Service_APT, "(STUBBED) called"); } -static u32 DecompressLZ11(const u8* in, u8* out) { - u32_le decompressed_size; - memcpy(&decompressed_size, in, sizeof(u32)); - in += 4; - - u8 type = decompressed_size & 0xFF; - ASSERT(type == 0x11); - decompressed_size >>= 8; - - u32 current_out_size = 0; - u8 flags = 0, mask = 1; - while (current_out_size < decompressed_size) { - if (mask == 1) { - flags = *(in++); - mask = 0x80; - } else { - mask >>= 1; - } - - if (flags & mask) { - u8 byte1 = *(in++); - u32 length = byte1 >> 4; - u32 offset; - if (length == 0) { - u8 byte2 = *(in++); - u8 byte3 = *(in++); - length = (((byte1 & 0x0F) << 4) | (byte2 >> 4)) + 0x11; - offset = (((byte2 & 0x0F) << 8) | byte3) + 0x1; - } else if (length == 1) { - u8 byte2 = *(in++); - u8 byte3 = *(in++); - u8 byte4 = *(in++); - length = (((byte1 & 0x0F) << 12) | (byte2 << 4) | (byte3 >> 4)) + 0x111; - offset = (((byte3 & 0x0F) << 8) | byte4) + 0x1; - } else { - u8 byte2 = *(in++); - length = (byte1 >> 4) + 0x1; - offset = (((byte1 & 0x0F) << 8) | byte2) + 0x1; - } - - for (u32 i = 0; i < length; i++) { - *out = *(out - offset); - ++out; - } - - current_out_size += length; - } else { - *(out++) = *(in++); - current_out_size++; - } - } - return decompressed_size; -} - -static bool LoadSharedFont() { - // TODO (wwylele): load different font archive for region CHN/KOR/TWN - const u64_le shared_font_archive_id_low = 0x0004009b00014002; - const u64_le shared_font_archive_id_high = 0x00000001ffffff00; - std::vector shared_font_archive_id(16); - std::memcpy(&shared_font_archive_id[0], &shared_font_archive_id_low, sizeof(u64)); - std::memcpy(&shared_font_archive_id[8], &shared_font_archive_id_high, sizeof(u64)); - FileSys::Path archive_path(shared_font_archive_id); - auto archive_result = Service::FS::OpenArchive(Service::FS::ArchiveIdCode::NCCH, archive_path); - if (archive_result.Failed()) - return false; - - std::vector romfs_path(20, 0); // 20-byte all zero path for opening RomFS - FileSys::Path file_path(romfs_path); - FileSys::Mode open_mode = {}; - open_mode.read_flag.Assign(1); - auto file_result = Service::FS::OpenFileFromArchive(*archive_result, file_path, open_mode); - if (file_result.Failed()) - return false; - - auto romfs = std::move(file_result).Unwrap(); - std::vector romfs_buffer(romfs->backend->GetSize()); - romfs->backend->Read(0, romfs_buffer.size(), romfs_buffer.data()); - romfs->backend->Close(); - - const u8* font_file = RomFS::GetFilePointer(romfs_buffer.data(), {u"cbf_std.bcfnt.lz"}); - if (font_file == nullptr) - return false; - - struct { - u32_le status; - u32_le region; - u32_le decompressed_size; - INSERT_PADDING_WORDS(0x1D); - } shared_font_header{}; - static_assert(sizeof(shared_font_header) == 0x80, "shared_font_header has incorrect size"); - - shared_font_header.status = 2; // successfully loaded - shared_font_header.region = 1; // region JPN/EUR/USA - shared_font_header.decompressed_size = - DecompressLZ11(font_file, shared_font_mem->GetPointer(0x80)); - std::memcpy(shared_font_mem->GetPointer(), &shared_font_header, sizeof(shared_font_header)); - *shared_font_mem->GetPointer(0x83) = 'U'; // Change the magic from "CFNT" to "CFNU" - - return true; -} - -static bool LoadLegacySharedFont() { - // This is the legacy method to load shared font. - // The expected format is a decrypted, uncompressed BCFNT file with the 0x80 byte header - // generated by the APT:U service. The best way to get is by dumping it from RAM. We've provided - // a homebrew app to do this: https://github.com/citra-emu/3dsutils. Put the resulting file - // "shared_font.bin" in the Citra "sysdata" directory. - std::string filepath = FileUtil::GetUserPath(D_SYSDATA_IDX) + SHARED_FONT; - - FileUtil::CreateFullPath(filepath); // Create path if not already created - FileUtil::IOFile file(filepath, "rb"); - if (file.IsOpen()) { - file.ReadBytes(shared_font_mem->GetPointer(), file.GetSize()); - return true; - } - - return false; -} - void Init() { AddService(new APT_A_Interface); AddService(new APT_S_Interface); @@ -995,16 +1023,6 @@ void Init() { MemoryPermission::ReadWrite, MemoryPermission::Read, 0, Kernel::MemoryRegion::SYSTEM, "APT:SharedFont"); - if (LoadSharedFont()) { - shared_font_loaded = true; - } else if (LoadLegacySharedFont()) { - LOG_WARNING(Service_APT, "Loaded shared font by legacy method"); - shared_font_loaded = true; - } else { - LOG_WARNING(Service_APT, "Unable to load shared font"); - shared_font_loaded = false; - } - lock = Kernel::Mutex::Create(false, "APT_U:Lock"); cpu_percent = 0; diff --git a/src/core/hle/service/cfg/cfg.cpp b/src/core/hle/service/cfg/cfg.cpp index 3dbeb27cc..f26a1f65f 100644 --- a/src/core/hle/service/cfg/cfg.cpp +++ b/src/core/hle/service/cfg/cfg.cpp @@ -168,7 +168,7 @@ void GetCountryCodeID(Service::Interface* self) { cmd_buff[2] = country_code_id; } -static u32 GetRegionValue() { +u32 GetRegionValue() { if (Settings::values.region_value == Settings::REGION_VALUE_AUTO_SELECT) return preferred_region_code; diff --git a/src/core/hle/service/cfg/cfg.h b/src/core/hle/service/cfg/cfg.h index 1659ebf32..282b6936b 100644 --- a/src/core/hle/service/cfg/cfg.h +++ b/src/core/hle/service/cfg/cfg.h @@ -101,6 +101,8 @@ void GetCountryCodeString(Service::Interface* self); */ void GetCountryCodeID(Service::Interface* self); +u32 GetRegionValue(); + /** * CFG::SecureInfoGetRegion service function * Inputs: From 79f177c6d28a2e7644e152a97639f66272d9f012 Mon Sep 17 00:00:00 2001 From: James Rowe Date: Sun, 3 Sep 2017 11:26:10 -0600 Subject: [PATCH 26/31] Fix icon for citra qt --- src/citra_qt/citra-qt.rc | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/citra_qt/citra-qt.rc b/src/citra_qt/citra-qt.rc index c490ef302..a48a9440d 100644 --- a/src/citra_qt/citra-qt.rc +++ b/src/citra_qt/citra-qt.rc @@ -6,7 +6,9 @@ // Icon with lowest ID value placed first to ensure application icon // remains consistent on all systems. -CITRA_ICON ICON "../../dist/citra.ico" +// QT requires that the default application icon is named IDI_ICON1 + +IDI_ICON1 ICON "../../dist/citra.ico" ///////////////////////////////////////////////////////////////////////////// From 589babbf7477423457dddbefbbb29623fa5c0624 Mon Sep 17 00:00:00 2001 From: mailwl Date: Sat, 12 Aug 2017 11:10:04 +0300 Subject: [PATCH 27/31] Mii Selector Applet: update Mii structures --- src/core/hle/applets/mii_selector.cpp | 6 +-- src/core/hle/applets/mii_selector.h | 57 ++++++++++++--------------- 2 files changed, 29 insertions(+), 34 deletions(-) diff --git a/src/core/hle/applets/mii_selector.cpp b/src/core/hle/applets/mii_selector.cpp index 705859f1e..f225c23a5 100644 --- a/src/core/hle/applets/mii_selector.cpp +++ b/src/core/hle/applets/mii_selector.cpp @@ -66,7 +66,7 @@ ResultCode MiiSelector::StartImpl(const Service::APT::AppletStartupParameter& pa // continue. MiiResult result; memset(&result, 0, sizeof(result)); - result.result_code = 0; + result.return_code = 0; // Let the application know that we're closing Service::APT::MessageParameter message; @@ -82,5 +82,5 @@ ResultCode MiiSelector::StartImpl(const Service::APT::AppletStartupParameter& pa } void MiiSelector::Update() {} -} -} // namespace +} // namespace Applets +} // namespace HLE diff --git a/src/core/hle/applets/mii_selector.h b/src/core/hle/applets/mii_selector.h index ec00e29d2..69db401d0 100644 --- a/src/core/hle/applets/mii_selector.h +++ b/src/core/hle/applets/mii_selector.h @@ -16,51 +16,46 @@ namespace HLE { namespace Applets { struct MiiConfig { - u8 unk_000; - u8 unk_001; - u8 unk_002; - u8 unk_003; - u8 unk_004; + u8 cancel_button_flag; + u8 enable_guest_mii_flag; + u8 show_on_top_screen_flag; + INSERT_PADDING_BYTES(5); + u16 title[0x40]; + INSERT_PADDING_BYTES(4); + u8 show_guest_miis_flag; INSERT_PADDING_BYTES(3); - u16 unk_008; - INSERT_PADDING_BYTES(0x82); - u8 unk_08C; - INSERT_PADDING_BYTES(3); - u16 unk_090; + u32 initially_selected_mii_index; + u8 guest_mii_whitelist[6]; + u8 user_mii_whitelist[0x64]; INSERT_PADDING_BYTES(2); - u32 unk_094; - u16 unk_098; - u8 unk_09A[0x64]; - u8 unk_0FE; - u8 unk_0FF; - u32 unk_100; + u32 magic_value; }; - static_assert(sizeof(MiiConfig) == 0x104, "MiiConfig structure has incorrect size"); #define ASSERT_REG_POSITION(field_name, position) \ static_assert(offsetof(MiiConfig, field_name) == position, \ "Field " #field_name " has invalid position") -ASSERT_REG_POSITION(unk_008, 0x08); -ASSERT_REG_POSITION(unk_08C, 0x8C); -ASSERT_REG_POSITION(unk_090, 0x90); -ASSERT_REG_POSITION(unk_094, 0x94); -ASSERT_REG_POSITION(unk_0FE, 0xFE); +ASSERT_REG_POSITION(title, 0x08); +ASSERT_REG_POSITION(show_guest_miis_flag, 0x8C); +ASSERT_REG_POSITION(initially_selected_mii_index, 0x90); +ASSERT_REG_POSITION(guest_mii_whitelist, 0x94); #undef ASSERT_REG_POSITION struct MiiResult { - u32 result_code; - u8 unk_04; - INSERT_PADDING_BYTES(7); - u8 unk_0C[0x60]; - u8 unk_6C[0x16]; + u32 return_code; + u32 guest_mii_selected_flag; + u32 selected_guest_mii_index; + // TODO(mailwl): expand to Mii Format structure: https://www.3dbrew.org/wiki/Mii + u8 selected_mii_data[0x5C]; INSERT_PADDING_BYTES(2); + u16 mii_data_checksum; + u16 guest_mii_name[0xC]; }; static_assert(sizeof(MiiResult) == 0x84, "MiiResult structure has incorrect size"); #define ASSERT_REG_POSITION(field_name, position) \ static_assert(offsetof(MiiResult, field_name) == position, \ "Field " #field_name " has invalid position") -ASSERT_REG_POSITION(unk_0C, 0x0C); -ASSERT_REG_POSITION(unk_6C, 0x6C); +ASSERT_REG_POSITION(selected_mii_data, 0x0C); +ASSERT_REG_POSITION(guest_mii_name, 0x6C); #undef ASSERT_REG_POSITION class MiiSelector final : public Applet { @@ -79,5 +74,5 @@ private: MiiConfig config; }; -} -} // namespace +} // namespace Applets +} // namespace HLE From 11f2eff17df600c57aba35384c9f82490368735d Mon Sep 17 00:00:00 2001 From: mailwl Date: Mon, 4 Sep 2017 12:15:15 +0300 Subject: [PATCH 28/31] Remove _flag in var names --- src/core/hle/applets/mii_selector.h | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/src/core/hle/applets/mii_selector.h b/src/core/hle/applets/mii_selector.h index 69db401d0..136ce8948 100644 --- a/src/core/hle/applets/mii_selector.h +++ b/src/core/hle/applets/mii_selector.h @@ -16,13 +16,13 @@ namespace HLE { namespace Applets { struct MiiConfig { - u8 cancel_button_flag; - u8 enable_guest_mii_flag; - u8 show_on_top_screen_flag; + u8 enable_cancel_button; + u8 enable_guest_mii; + u8 show_on_top_screen; INSERT_PADDING_BYTES(5); u16 title[0x40]; INSERT_PADDING_BYTES(4); - u8 show_guest_miis_flag; + u8 show_guest_miis; INSERT_PADDING_BYTES(3); u32 initially_selected_mii_index; u8 guest_mii_whitelist[6]; @@ -35,14 +35,14 @@ static_assert(sizeof(MiiConfig) == 0x104, "MiiConfig structure has incorrect siz static_assert(offsetof(MiiConfig, field_name) == position, \ "Field " #field_name " has invalid position") ASSERT_REG_POSITION(title, 0x08); -ASSERT_REG_POSITION(show_guest_miis_flag, 0x8C); +ASSERT_REG_POSITION(show_guest_miis, 0x8C); ASSERT_REG_POSITION(initially_selected_mii_index, 0x90); ASSERT_REG_POSITION(guest_mii_whitelist, 0x94); #undef ASSERT_REG_POSITION struct MiiResult { u32 return_code; - u32 guest_mii_selected_flag; + u32 is_guest_mii_selected; u32 selected_guest_mii_index; // TODO(mailwl): expand to Mii Format structure: https://www.3dbrew.org/wiki/Mii u8 selected_mii_data[0x5C]; From ef8925b7acfe03cb83963e3f9ec4592ae308f4ee Mon Sep 17 00:00:00 2001 From: James Rowe Date: Tue, 5 Sep 2017 18:57:47 -0600 Subject: [PATCH 29/31] Remove excess debug dlls for mingw build --- appveyor.yml | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/appveyor.yml b/appveyor.yml index 1c390cdd5..ec9ca3747 100644 --- a/appveyor.yml +++ b/appveyor.yml @@ -131,6 +131,12 @@ after_build: foreach ($file in $MingwDLLs) { Copy-Item -path "C:/msys64/mingw64/bin/$file" -force -destination "$RELEASE_DIST" } + # the above list copies a few extra debug dlls that aren't needed (thanks globbing patterns!) + # so we can remove them by hardcoding another list of extra dlls to remove + $DebugDLLs = "libicudtd*.dll","libicuind*.dll","libicuucd*.dll" + foreach ($file in $DebugDLLs) { + Remove-Item -path "$RELEASE_DIST/$file" + } # copy the qt windows plugin dll to platforms Copy-Item -path "C:/msys64/mingw64/share/qt5/plugins/platforms/qwindows.dll" -force -destination "$RELEASE_DIST/platforms" From 0c55bed0477fc42dd95b512cbce741723a8878a6 Mon Sep 17 00:00:00 2001 From: B3n30 Date: Sat, 9 Sep 2017 21:21:53 +0200 Subject: [PATCH 30/31] trvis_OSX: build with system curl --- .travis-build.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.travis-build.sh b/.travis-build.sh index 64f5aed94..bb4e6fc47 100755 --- a/.travis-build.sh +++ b/.travis-build.sh @@ -52,7 +52,7 @@ elif [ "$TRAVIS_OS_NAME" = "osx" ]; then export Qt5_DIR=$(brew --prefix)/opt/qt5 mkdir build && cd build - cmake .. -GXcode + cmake .. -DUSE_SYSTEM_CURL=ON -GXcode xcodebuild -configuration Release ctest -VV -C Release From 9e847b754933f09bdf19be0dc45fc32a65052965 Mon Sep 17 00:00:00 2001 From: James Rowe Date: Sat, 9 Sep 2017 11:24:13 -0600 Subject: [PATCH 31/31] Build: Enable SSL in mingw by linking against WinSSL The mingw builds aren't submitting telemetry because the curl library they are linked against is configured to use openSSL and openSSL looks for the certificates in the users home folder. This keeps it from contacting web services because it can't communicate over SSL. This commit adds a download in mingw builds that will download a precompiled curl for mingw linked against winssl and sspi. --- CMakeLists.txt | 28 ++++++++++------------------ CMakeModules/DownloadExternals.cmake | 18 ++++++++++++++++++ appveyor.yml | 11 ++++------- externals/CMakeLists.txt | 19 +++++++++++++++++-- 4 files changed, 49 insertions(+), 27 deletions(-) create mode 100644 CMakeModules/DownloadExternals.cmake diff --git a/CMakeLists.txt b/CMakeLists.txt index f8060270e..d9c2f78a2 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -2,6 +2,7 @@ cmake_minimum_required(VERSION 3.6) list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/CMakeModules") list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/externals/cmake-modules") +include(DownloadExternals) project(citra) @@ -12,6 +13,15 @@ option(ENABLE_QT "Enable the Qt frontend" ON) option(CITRA_USE_BUNDLED_QT "Download bundled Qt binaries" OFF) option(ENABLE_WEB_SERVICE "Enable web services (telemetry, etc.)" ON) +option(CITRA_USE_BUNDLED_CURL "FOR MINGW ONLY: Download curl configured against winssl instead of openssl" OFF) +if (ENABLE_WEB_SERVICE AND CITRA_USE_BUNDLED_CURL AND WINDOWS AND MSVC) + message("Turning off use bundled curl as msvc can compile curl on cpr") + SET(CITRA_USE_BUNDLED_CURL OFF CACHE BOOL "" FORCE) +endif() +if (ENABLE_WEB_SERVICE AND NOT CITRA_USE_BUNDLED_CURL AND MINGW) + message(AUTHOR_WARNING "Turning on CITRA_USE_BUNDLED_CURL. Override it only if you know what you are doing.") + SET(CITRA_USE_BUNDLED_CURL ON CACHE BOOL "" FORCE) +endif() if(NOT EXISTS ${CMAKE_SOURCE_DIR}/.git/hooks/pre-commit) message(STATUS "Copying pre-commit hook") @@ -151,24 +161,6 @@ set_property(DIRECTORY APPEND PROPERTY # System imported libraries # ====================== -# This function downloads a binary library package from our external repo. -# Params: -# remote_path: path to the file to download, relative to the remote repository root -# prefix_var: name of a variable which will be set with the path to the extracted contents -function(download_bundled_external remote_path lib_name prefix_var) - set(prefix "${CMAKE_BINARY_DIR}/externals/${lib_name}") - if (NOT EXISTS "${prefix}") - message(STATUS "Downloading binaries for ${lib_name}...") - file(DOWNLOAD - https://github.com/citra-emu/ext-windows-bin/raw/master/${remote_path}${lib_name}.7z - "${CMAKE_BINARY_DIR}/externals/${lib_name}.7z" SHOW_PROGRESS) - execute_process(COMMAND ${CMAKE_COMMAND} -E tar xf "${CMAKE_BINARY_DIR}/externals/${lib_name}.7z" - WORKING_DIRECTORY "${CMAKE_BINARY_DIR}/externals") - endif() - message(STATUS "Using bundled binaries at ${prefix}") - set(${prefix_var} "${prefix}" PARENT_SCOPE) -endfunction() - find_package(PNG QUIET) if (NOT PNG_FOUND) message(STATUS "libpng not found. Some debugging features have been disabled.") diff --git a/CMakeModules/DownloadExternals.cmake b/CMakeModules/DownloadExternals.cmake new file mode 100644 index 000000000..138a15d5a --- /dev/null +++ b/CMakeModules/DownloadExternals.cmake @@ -0,0 +1,18 @@ + +# This function downloads a binary library package from our external repo. +# Params: +# remote_path: path to the file to download, relative to the remote repository root +# prefix_var: name of a variable which will be set with the path to the extracted contents +function(download_bundled_external remote_path lib_name prefix_var) +set(prefix "${CMAKE_BINARY_DIR}/externals/${lib_name}") +if (NOT EXISTS "${prefix}") + message(STATUS "Downloading binaries for ${lib_name}...") + file(DOWNLOAD + https://github.com/citra-emu/ext-windows-bin/raw/master/${remote_path}${lib_name}.7z + "${CMAKE_BINARY_DIR}/externals/${lib_name}.7z" SHOW_PROGRESS) + execute_process(COMMAND ${CMAKE_COMMAND} -E tar xf "${CMAKE_BINARY_DIR}/externals/${lib_name}.7z" + WORKING_DIRECTORY "${CMAKE_BINARY_DIR}/externals") +endif() +message(STATUS "Using bundled binaries at ${prefix}") +set(${prefix_var} "${prefix}" PARENT_SCOPE) +endfunction() \ No newline at end of file diff --git a/appveyor.yml b/appveyor.yml index ec9ca3747..5524eb576 100644 --- a/appveyor.yml +++ b/appveyor.yml @@ -44,7 +44,7 @@ before_build: # redirect stderr and change the exit code to prevent powershell from cancelling the build if cmake prints a warning cmd /C 'cmake -G "Visual Studio 15 2017 Win64" -DCITRA_USE_BUNDLED_QT=1 -DCITRA_USE_BUNDLED_SDL2=1 -DCMAKE_USE_OPENSSL=0 .. 2>&1 && exit 0' } else { - C:\msys64\usr\bin\bash.exe -lc "cmake -G 'MSYS Makefiles' -DUSE_SYSTEM_CURL=1 -DCMAKE_BUILD_TYPE=Release .. 2>&1" + C:\msys64\usr\bin\bash.exe -lc "cmake -G 'MSYS Makefiles' -DUSE_SYSTEM_CURL=1 -DCITRA_USE_BUNDLED_CURL=1 -DCMAKE_BUILD_TYPE=Release .. 2>&1" } - cd .. @@ -112,9 +112,10 @@ after_build: # copy the compiled binaries and other release files to the release folder Get-ChildItem "$CMAKE_BINARY_DIR" -Recurse -Filter "citra*.exe" | Copy-Item -destination $RELEASE_DIST + # copy the libcurl dll + Get-ChildItem "$CMAKE_BINARY_DIR" -Recurse -Filter "libcurl.dll" | Copy-Item -destination $RELEASE_DIST Copy-Item -path "$CMAKE_SOURCE_DIR/license.txt" -destination $RELEASE_DIST Copy-Item -path "$CMAKE_SOURCE_DIR/README.md" -destination $RELEASE_DIST - # copy all the dll dependencies to the release folder # hardcoded list because we don't build static and determining the list of dlls from the binary is a pain. $MingwDLLs = "Qt5Core.dll","Qt5Widgets.dll","Qt5Gui.dll","Qt5OpenGL.dll", @@ -123,11 +124,7 @@ after_build: "libfreetype-*.dll","libglib-*.dll","libgobject-*.dll","libgraphite2.dll","libiconv-*.dll", "libharfbuzz-*.dll","libintl-*.dll","libpcre-*.dll","libpcre16-*.dll","libpng16-*.dll", # Runtime/Other dependencies - "libgcc_s_seh-*.dll","libstdc++-*.dll","libwinpthread-*.dll","SDL2.dll","zlib1.dll", - # curl dependencies - "libcurl-*.dll","libnghttp2-*.dll","libeay32.dll","libgmp-*.dll","librtmp-*.dll", - "libgnutls-*.dll","libhogweed-*.dll","libnettle-*.dll","libssh2-*.dll", - "ssleay32.dll","libidn-*.dll","libp11-kit-*.dll","libtasn1-*.dll","libunistring-*.dll" + "libgcc_s_seh-*.dll","libstdc++-*.dll","libwinpthread-*.dll","SDL2.dll","zlib1.dll" foreach ($file in $MingwDLLs) { Copy-Item -path "C:/msys64/mingw64/bin/$file" -force -destination "$RELEASE_DIST" } diff --git a/externals/CMakeLists.txt b/externals/CMakeLists.txt index 8e4bcf21f..4a4ba1101 100644 --- a/externals/CMakeLists.txt +++ b/externals/CMakeLists.txt @@ -1,5 +1,8 @@ # Definitions for all external bundled libraries +set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} ${PROJECT_SOURCE_DIR}/CMakeModules) +include(DownloadExternals) + # Catch add_library(catch-single-include INTERFACE) target_include_directories(catch-single-include INTERFACE catch/single_include) @@ -54,9 +57,21 @@ add_subdirectory(enet) target_include_directories(enet INTERFACE ./enet/include) if (ENABLE_WEB_SERVICE) + # msys installed curl is configured to use openssl, but that isn't portable + # since it relies on having the bundled certs install in the home folder for SSL + # by default on mingw, download the precompiled curl thats linked against windows native ssl + if (MINGW AND CITRA_USE_BUNDLED_CURL) + download_bundled_external("curl/" "curl-7_55_1" CURL_PREFIX) + set(CURL_PREFIX "${CMAKE_BINARY_DIR}/externals/curl-7_55_1") + set(CURL_FOUND YES) + set(CURL_INCLUDE_DIR "${CURL_PREFIX}/include" CACHE PATH "Path to curl headers") + set(CURL_LIBRARY "${CURL_PREFIX}/lib/libcurldll.a" CACHE PATH "Path to curl library") + set(CURL_DLL_DIR "${CURL_PREFIX}/lib/" CACHE PATH "Path to curl.dll") + set(USE_SYSTEM_CURL ON CACHE BOOL "") + endif() # CPR - option(BUILD_TESTING OFF) - option(BUILD_CPR_TESTS OFF) + set(BUILD_TESTING OFF CACHE BOOL "") + set(BUILD_CPR_TESTS OFF CACHE BOOL "") add_subdirectory(cpr) target_include_directories(cpr INTERFACE ./cpr/include)