|
|
@ -40,16 +40,15 @@ MICROPROFILE_DEFINE(OpenGL_Blits, "OpenGL", "Blits", MP_RGB(100, 100, 255));
|
|
|
|
MICROPROFILE_DEFINE(OpenGL_CacheManagement, "OpenGL", "Cache Mgmt", MP_RGB(100, 255, 100));
|
|
|
|
MICROPROFILE_DEFINE(OpenGL_CacheManagement, "OpenGL", "Cache Mgmt", MP_RGB(100, 255, 100));
|
|
|
|
|
|
|
|
|
|
|
|
static bool IsVendorAmd() {
|
|
|
|
static bool IsVendorAmd() {
|
|
|
|
std::string gpu_vendor{reinterpret_cast<char const*>(glGetString(GL_VENDOR))};
|
|
|
|
const std::string_view gpu_vendor{reinterpret_cast<char const*>(glGetString(GL_VENDOR))};
|
|
|
|
return gpu_vendor == "ATI Technologies Inc." || gpu_vendor == "Advanced Micro Devices, Inc.";
|
|
|
|
return gpu_vendor == "ATI Technologies Inc." || gpu_vendor == "Advanced Micro Devices, Inc.";
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
RasterizerOpenGL::RasterizerOpenGL(Frontend::EmuWindow& window)
|
|
|
|
RasterizerOpenGL::RasterizerOpenGL()
|
|
|
|
: is_amd(IsVendorAmd()), shader_dirty(true),
|
|
|
|
: is_amd(IsVendorAmd()), vertex_buffer(GL_ARRAY_BUFFER, VERTEX_BUFFER_SIZE, is_amd),
|
|
|
|
vertex_buffer(GL_ARRAY_BUFFER, VERTEX_BUFFER_SIZE, is_amd),
|
|
|
|
|
|
|
|
uniform_buffer(GL_UNIFORM_BUFFER, UNIFORM_BUFFER_SIZE, false),
|
|
|
|
uniform_buffer(GL_UNIFORM_BUFFER, UNIFORM_BUFFER_SIZE, false),
|
|
|
|
index_buffer(GL_ELEMENT_ARRAY_BUFFER, INDEX_BUFFER_SIZE, false),
|
|
|
|
index_buffer(GL_ELEMENT_ARRAY_BUFFER, INDEX_BUFFER_SIZE, false),
|
|
|
|
texture_buffer(GL_TEXTURE_BUFFER, TEXTURE_BUFFER_SIZE, false), emu_window{window} {
|
|
|
|
texture_buffer(GL_TEXTURE_BUFFER, TEXTURE_BUFFER_SIZE, false) {
|
|
|
|
|
|
|
|
|
|
|
|
allow_shadow = GLAD_GL_ARB_shader_image_load_store && GLAD_GL_ARB_shader_image_size &&
|
|
|
|
allow_shadow = GLAD_GL_ARB_shader_image_load_store && GLAD_GL_ARB_shader_image_size &&
|
|
|
|
GLAD_GL_ARB_framebuffer_no_attachments;
|
|
|
|
GLAD_GL_ARB_framebuffer_no_attachments;
|
|
|
@ -169,7 +168,7 @@ RasterizerOpenGL::RasterizerOpenGL(Frontend::EmuWindow& window)
|
|
|
|
SyncEntireState();
|
|
|
|
SyncEntireState();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
RasterizerOpenGL::~RasterizerOpenGL() {}
|
|
|
|
RasterizerOpenGL::~RasterizerOpenGL() = default;
|
|
|
|
|
|
|
|
|
|
|
|
void RasterizerOpenGL::LoadDiskResources(const std::atomic_bool& stop_loading,
|
|
|
|
void RasterizerOpenGL::LoadDiskResources(const std::atomic_bool& stop_loading,
|
|
|
|
const VideoCore::DiskResourceLoadCallback& callback) {
|
|
|
|
const VideoCore::DiskResourceLoadCallback& callback) {
|
|
|
@ -272,17 +271,17 @@ RasterizerOpenGL::VertexArrayInfo RasterizerOpenGL::AnalyzeVertexArray(bool is_i
|
|
|
|
u32 vertex_max;
|
|
|
|
u32 vertex_max;
|
|
|
|
if (is_indexed) {
|
|
|
|
if (is_indexed) {
|
|
|
|
const auto& index_info = regs.pipeline.index_array;
|
|
|
|
const auto& index_info = regs.pipeline.index_array;
|
|
|
|
PAddr address = vertex_attributes.GetPhysicalBaseAddress() + index_info.offset;
|
|
|
|
const PAddr address = vertex_attributes.GetPhysicalBaseAddress() + index_info.offset;
|
|
|
|
const u8* index_address_8 = VideoCore::g_memory->GetPhysicalPointer(address);
|
|
|
|
const u8* index_address_8 = VideoCore::g_memory->GetPhysicalPointer(address);
|
|
|
|
const u16* index_address_16 = reinterpret_cast<const u16*>(index_address_8);
|
|
|
|
const u16* index_address_16 = reinterpret_cast<const u16*>(index_address_8);
|
|
|
|
bool index_u16 = index_info.format != 0;
|
|
|
|
const bool index_u16 = index_info.format != 0;
|
|
|
|
|
|
|
|
|
|
|
|
vertex_min = 0xFFFF;
|
|
|
|
vertex_min = 0xFFFF;
|
|
|
|
vertex_max = 0;
|
|
|
|
vertex_max = 0;
|
|
|
|
std::size_t size = regs.pipeline.num_vertices * (index_u16 ? 2 : 1);
|
|
|
|
const u32 size = regs.pipeline.num_vertices * (index_u16 ? 2 : 1);
|
|
|
|
res_cache.FlushRegion(address, size, nullptr);
|
|
|
|
res_cache.FlushRegion(address, size, nullptr);
|
|
|
|
for (u32 index = 0; index < regs.pipeline.num_vertices; ++index) {
|
|
|
|
for (u32 index = 0; index < regs.pipeline.num_vertices; ++index) {
|
|
|
|
u32 vertex = index_u16 ? index_address_16[index] : index_address_8[index];
|
|
|
|
const u32 vertex = index_u16 ? index_address_16[index] : index_address_8[index];
|
|
|
|
vertex_min = std::min(vertex_min, vertex);
|
|
|
|
vertex_min = std::min(vertex_min, vertex);
|
|
|
|
vertex_max = std::max(vertex_max, vertex);
|
|
|
|
vertex_max = std::max(vertex_max, vertex);
|
|
|
|
}
|
|
|
|
}
|
|
|
@ -291,9 +290,9 @@ RasterizerOpenGL::VertexArrayInfo RasterizerOpenGL::AnalyzeVertexArray(bool is_i
|
|
|
|
vertex_max = regs.pipeline.vertex_offset + regs.pipeline.num_vertices - 1;
|
|
|
|
vertex_max = regs.pipeline.vertex_offset + regs.pipeline.num_vertices - 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
u32 vertex_num = vertex_max - vertex_min + 1;
|
|
|
|
const u32 vertex_num = vertex_max - vertex_min + 1;
|
|
|
|
u32 vs_input_size = 0;
|
|
|
|
u32 vs_input_size = 0;
|
|
|
|
for (auto& loader : vertex_attributes.attribute_loaders) {
|
|
|
|
for (const auto& loader : vertex_attributes.attribute_loaders) {
|
|
|
|
if (loader.component_count != 0) {
|
|
|
|
if (loader.component_count != 0) {
|
|
|
|
vs_input_size += loader.byte_count * vertex_num;
|
|
|
|
vs_input_size += loader.byte_count * vertex_num;
|
|
|
|
}
|
|
|
|
}
|
|
|
@ -363,15 +362,15 @@ void RasterizerOpenGL::SetupVertexArray(u8* array_ptr, GLintptr buffer_offset,
|
|
|
|
for (std::size_t i = 0; i < enable_attributes.size(); ++i) {
|
|
|
|
for (std::size_t i = 0; i < enable_attributes.size(); ++i) {
|
|
|
|
if (enable_attributes[i] != hw_vao_enabled_attributes[i]) {
|
|
|
|
if (enable_attributes[i] != hw_vao_enabled_attributes[i]) {
|
|
|
|
if (enable_attributes[i]) {
|
|
|
|
if (enable_attributes[i]) {
|
|
|
|
glEnableVertexAttribArray(i);
|
|
|
|
glEnableVertexAttribArray(static_cast<GLuint>(i));
|
|
|
|
} else {
|
|
|
|
} else {
|
|
|
|
glDisableVertexAttribArray(i);
|
|
|
|
glDisableVertexAttribArray(static_cast<GLuint>(i));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
hw_vao_enabled_attributes[i] = enable_attributes[i];
|
|
|
|
hw_vao_enabled_attributes[i] = enable_attributes[i];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
if (vertex_attributes.IsDefaultAttribute(i)) {
|
|
|
|
if (vertex_attributes.IsDefaultAttribute(i)) {
|
|
|
|
u32 reg = regs.vs.GetRegisterForAttribute(i);
|
|
|
|
const u32 reg = regs.vs.GetRegisterForAttribute(i);
|
|
|
|
if (!enable_attributes[reg]) {
|
|
|
|
if (!enable_attributes[reg]) {
|
|
|
|
const auto& attr = Pica::g_state.input_default_attributes.attr[i];
|
|
|
|
const auto& attr = Pica::g_state.input_default_attributes.attr[i];
|
|
|
|
glVertexAttrib4f(reg, attr.x.ToFloat32(), attr.y.ToFloat32(), attr.z.ToFloat32(),
|
|
|
|
glVertexAttrib4f(reg, attr.x.ToFloat32(), attr.y.ToFloat32(), attr.z.ToFloat32(),
|
|
|
@ -798,15 +797,16 @@ bool RasterizerOpenGL::Draw(bool accelerate, bool is_indexed) {
|
|
|
|
std::size_t max_vertices = 3 * (VERTEX_BUFFER_SIZE / (3 * sizeof(HardwareVertex)));
|
|
|
|
std::size_t max_vertices = 3 * (VERTEX_BUFFER_SIZE / (3 * sizeof(HardwareVertex)));
|
|
|
|
for (std::size_t base_vertex = 0; base_vertex < vertex_batch.size();
|
|
|
|
for (std::size_t base_vertex = 0; base_vertex < vertex_batch.size();
|
|
|
|
base_vertex += max_vertices) {
|
|
|
|
base_vertex += max_vertices) {
|
|
|
|
std::size_t vertices = std::min(max_vertices, vertex_batch.size() - base_vertex);
|
|
|
|
const std::size_t vertices = std::min(max_vertices, vertex_batch.size() - base_vertex);
|
|
|
|
std::size_t vertex_size = vertices * sizeof(HardwareVertex);
|
|
|
|
const std::size_t vertex_size = vertices * sizeof(HardwareVertex);
|
|
|
|
u8* vbo;
|
|
|
|
u8* vbo;
|
|
|
|
GLintptr offset;
|
|
|
|
GLintptr offset;
|
|
|
|
std::tie(vbo, offset, std::ignore) =
|
|
|
|
std::tie(vbo, offset, std::ignore) =
|
|
|
|
vertex_buffer.Map(vertex_size, sizeof(HardwareVertex));
|
|
|
|
vertex_buffer.Map(vertex_size, sizeof(HardwareVertex));
|
|
|
|
std::memcpy(vbo, vertex_batch.data() + base_vertex, vertex_size);
|
|
|
|
std::memcpy(vbo, vertex_batch.data() + base_vertex, vertex_size);
|
|
|
|
vertex_buffer.Unmap(vertex_size);
|
|
|
|
vertex_buffer.Unmap(vertex_size);
|
|
|
|
glDrawArrays(GL_TRIANGLES, offset / sizeof(HardwareVertex), (GLsizei)vertices);
|
|
|
|
glDrawArrays(GL_TRIANGLES, static_cast<GLint>(offset / sizeof(HardwareVertex)),
|
|
|
|
|
|
|
|
static_cast<GLsizei>(vertices));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
@ -1338,7 +1338,7 @@ void RasterizerOpenGL::NotifyPicaRegisterChanged(u32 id) {
|
|
|
|
case PICA_REG_INDEX(lighting.lut_data[5]):
|
|
|
|
case PICA_REG_INDEX(lighting.lut_data[5]):
|
|
|
|
case PICA_REG_INDEX(lighting.lut_data[6]):
|
|
|
|
case PICA_REG_INDEX(lighting.lut_data[6]):
|
|
|
|
case PICA_REG_INDEX(lighting.lut_data[7]): {
|
|
|
|
case PICA_REG_INDEX(lighting.lut_data[7]): {
|
|
|
|
auto& lut_config = regs.lighting.lut_config;
|
|
|
|
const auto& lut_config = regs.lighting.lut_config;
|
|
|
|
uniform_block_data.lighting_lut_dirty[lut_config.type] = true;
|
|
|
|
uniform_block_data.lighting_lut_dirty[lut_config.type] = true;
|
|
|
|
uniform_block_data.lighting_lut_dirty_any = true;
|
|
|
|
uniform_block_data.lighting_lut_dirty_any = true;
|
|
|
|
break;
|
|
|
|
break;
|
|
|
@ -1564,8 +1564,8 @@ void RasterizerOpenGL::SamplerInfo::Create() {
|
|
|
|
|
|
|
|
|
|
|
|
// default is 1000 and -1000
|
|
|
|
// default is 1000 and -1000
|
|
|
|
// Other attributes have correct defaults
|
|
|
|
// Other attributes have correct defaults
|
|
|
|
glSamplerParameterf(sampler.handle, GL_TEXTURE_MAX_LOD, lod_max);
|
|
|
|
glSamplerParameterf(sampler.handle, GL_TEXTURE_MAX_LOD, static_cast<float>(lod_max));
|
|
|
|
glSamplerParameterf(sampler.handle, GL_TEXTURE_MIN_LOD, lod_min);
|
|
|
|
glSamplerParameterf(sampler.handle, GL_TEXTURE_MIN_LOD, static_cast<float>(lod_min));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void RasterizerOpenGL::SamplerInfo::SyncWithConfig(
|
|
|
|
void RasterizerOpenGL::SamplerInfo::SyncWithConfig(
|
|
|
@ -1615,12 +1615,12 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(
|
|
|
|
|
|
|
|
|
|
|
|
if (lod_min != config.lod.min_level) {
|
|
|
|
if (lod_min != config.lod.min_level) {
|
|
|
|
lod_min = config.lod.min_level;
|
|
|
|
lod_min = config.lod.min_level;
|
|
|
|
glSamplerParameterf(s, GL_TEXTURE_MIN_LOD, lod_min);
|
|
|
|
glSamplerParameterf(s, GL_TEXTURE_MIN_LOD, static_cast<float>(lod_min));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
if (lod_max != config.lod.max_level) {
|
|
|
|
if (lod_max != config.lod.max_level) {
|
|
|
|
lod_max = config.lod.max_level;
|
|
|
|
lod_max = config.lod.max_level;
|
|
|
|
glSamplerParameterf(s, GL_TEXTURE_MAX_LOD, lod_max);
|
|
|
|
glSamplerParameterf(s, GL_TEXTURE_MAX_LOD, static_cast<float>(lod_max));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
if (!GLES && lod_bias != config.lod.bias) {
|
|
|
|
if (!GLES && lod_bias != config.lod.bias) {
|
|
|
@ -1837,13 +1837,16 @@ void RasterizerOpenGL::SyncCombinerColor() {
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void RasterizerOpenGL::SyncTevConstColor(int stage_index,
|
|
|
|
void RasterizerOpenGL::SyncTevConstColor(std::size_t stage_index,
|
|
|
|
const Pica::TexturingRegs::TevStageConfig& tev_stage) {
|
|
|
|
const Pica::TexturingRegs::TevStageConfig& tev_stage) {
|
|
|
|
auto const_color = PicaToGL::ColorRGBA8(tev_stage.const_color);
|
|
|
|
const auto const_color = PicaToGL::ColorRGBA8(tev_stage.const_color);
|
|
|
|
if (const_color != uniform_block_data.data.const_color[stage_index]) {
|
|
|
|
|
|
|
|
uniform_block_data.data.const_color[stage_index] = const_color;
|
|
|
|
if (const_color == uniform_block_data.data.const_color[stage_index]) {
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
uniform_block_data.data.const_color[stage_index] = const_color;
|
|
|
|
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void RasterizerOpenGL::SyncGlobalAmbient() {
|
|
|
|
void RasterizerOpenGL::SyncGlobalAmbient() {
|
|
|
@ -1990,7 +1993,7 @@ void RasterizerOpenGL::SyncAndUploadLUTs() {
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(),
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(),
|
|
|
|
new_data.size() * sizeof(GLvec2));
|
|
|
|
new_data.size() * sizeof(GLvec2));
|
|
|
|
uniform_block_data.data.lighting_lut_offset[index / 4][index % 4] =
|
|
|
|
uniform_block_data.data.lighting_lut_offset[index / 4][index % 4] =
|
|
|
|
(offset + bytes_used) / sizeof(GLvec2);
|
|
|
|
static_cast<GLint>((offset + bytes_used) / sizeof(GLvec2));
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec2);
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec2);
|
|
|
|
}
|
|
|
|
}
|
|
|
@ -2012,7 +2015,8 @@ void RasterizerOpenGL::SyncAndUploadLUTs() {
|
|
|
|
if (new_data != fog_lut_data || invalidate) {
|
|
|
|
if (new_data != fog_lut_data || invalidate) {
|
|
|
|
fog_lut_data = new_data;
|
|
|
|
fog_lut_data = new_data;
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(), new_data.size() * sizeof(GLvec2));
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(), new_data.size() * sizeof(GLvec2));
|
|
|
|
uniform_block_data.data.fog_lut_offset = (offset + bytes_used) / sizeof(GLvec2);
|
|
|
|
uniform_block_data.data.fog_lut_offset =
|
|
|
|
|
|
|
|
static_cast<GLint>((offset + bytes_used) / sizeof(GLvec2));
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec2);
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec2);
|
|
|
|
}
|
|
|
|
}
|
|
|
@ -2031,7 +2035,7 @@ void RasterizerOpenGL::SyncAndUploadLUTs() {
|
|
|
|
if (new_data != lut_data || invalidate) {
|
|
|
|
if (new_data != lut_data || invalidate) {
|
|
|
|
lut_data = new_data;
|
|
|
|
lut_data = new_data;
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(), new_data.size() * sizeof(GLvec2));
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(), new_data.size() * sizeof(GLvec2));
|
|
|
|
lut_offset = (offset + bytes_used) / sizeof(GLvec2);
|
|
|
|
lut_offset = static_cast<GLint>((offset + bytes_used) / sizeof(GLvec2));
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec2);
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec2);
|
|
|
|
}
|
|
|
|
}
|
|
|
@ -2072,7 +2076,8 @@ void RasterizerOpenGL::SyncAndUploadLUTs() {
|
|
|
|
if (new_data != proctex_lut_data || invalidate) {
|
|
|
|
if (new_data != proctex_lut_data || invalidate) {
|
|
|
|
proctex_lut_data = new_data;
|
|
|
|
proctex_lut_data = new_data;
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(), new_data.size() * sizeof(GLvec4));
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(), new_data.size() * sizeof(GLvec4));
|
|
|
|
uniform_block_data.data.proctex_lut_offset = (offset + bytes_used) / sizeof(GLvec4);
|
|
|
|
uniform_block_data.data.proctex_lut_offset =
|
|
|
|
|
|
|
|
static_cast<GLint>((offset + bytes_used) / sizeof(GLvec4));
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec4);
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec4);
|
|
|
|
}
|
|
|
|
}
|
|
|
@ -2094,7 +2099,7 @@ void RasterizerOpenGL::SyncAndUploadLUTs() {
|
|
|
|
proctex_diff_lut_data = new_data;
|
|
|
|
proctex_diff_lut_data = new_data;
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(), new_data.size() * sizeof(GLvec4));
|
|
|
|
std::memcpy(buffer + bytes_used, new_data.data(), new_data.size() * sizeof(GLvec4));
|
|
|
|
uniform_block_data.data.proctex_diff_lut_offset =
|
|
|
|
uniform_block_data.data.proctex_diff_lut_offset =
|
|
|
|
(offset + bytes_used) / sizeof(GLvec4);
|
|
|
|
static_cast<GLint>((offset + bytes_used) / sizeof(GLvec4));
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
uniform_block_data.dirty = true;
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec4);
|
|
|
|
bytes_used += new_data.size() * sizeof(GLvec4);
|
|
|
|
}
|
|
|
|
}
|
|
|
|