summaryrefslogtreecommitdiffstats
path: root/src/video_core
diff options
context:
space:
mode:
Diffstat (limited to 'src/video_core')
-rw-r--r--src/video_core/dma_pusher.cpp57
-rw-r--r--src/video_core/dma_pusher.h5
-rw-r--r--src/video_core/engines/fermi_2d.cpp8
-rw-r--r--src/video_core/engines/maxwell_3d.cpp66
-rw-r--r--src/video_core/engines/maxwell_3d.h17
-rw-r--r--src/video_core/engines/shader_bytecode.h6
-rw-r--r--src/video_core/engines/shader_header.h41
-rw-r--r--src/video_core/gpu.h2
-rw-r--r--src/video_core/rasterizer_cache.h18
-rw-r--r--src/video_core/rasterizer_interface.h4
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer.cpp32
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer.h4
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer_cache.cpp142
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer_cache.h91
-rw-r--r--src/video_core/renderer_opengl/gl_shader_decompiler.cpp44
-rw-r--r--src/video_core/renderer_opengl/gl_shader_disk_cache.cpp2
-rw-r--r--src/video_core/renderer_opengl/gl_shader_gen.cpp4
-rw-r--r--src/video_core/renderer_opengl/renderer_opengl.cpp1
-rw-r--r--src/video_core/renderer_opengl/renderer_opengl.h4
-rw-r--r--src/video_core/renderer_vulkan/vk_memory_manager.cpp2
-rw-r--r--src/video_core/shader/decode/memory.cpp2
-rw-r--r--src/video_core/shader/decode/other.cpp15
-rw-r--r--src/video_core/shader/track.cpp10
23 files changed, 409 insertions, 168 deletions
diff --git a/src/video_core/dma_pusher.cpp b/src/video_core/dma_pusher.cpp
index eb9bf1878..669541b4b 100644
--- a/src/video_core/dma_pusher.cpp
+++ b/src/video_core/dma_pusher.cpp
@@ -33,18 +33,36 @@ void DmaPusher::DispatchCalls() {
}
bool DmaPusher::Step() {
- if (dma_get != dma_put) {
- // Push buffer non-empty, read a word
- const auto address = gpu.MemoryManager().GpuToCpuAddress(dma_get);
- ASSERT_MSG(address, "Invalid GPU address");
+ if (!ib_enable || dma_pushbuffer.empty()) {
+ // pushbuffer empty and IB empty or nonexistent - nothing to do
+ return false;
+ }
- const CommandHeader command_header{Memory::Read32(*address)};
+ const CommandList& command_list{dma_pushbuffer.front()};
+ const CommandListHeader& command_list_header{command_list[dma_pushbuffer_subindex++]};
+ GPUVAddr dma_get = command_list_header.addr;
+ GPUVAddr dma_put = dma_get + command_list_header.size * sizeof(u32);
+ bool non_main = command_list_header.is_non_main;
- dma_get += sizeof(u32);
+ if (dma_pushbuffer_subindex >= command_list.size()) {
+ // We've gone through the current list, remove it from the queue
+ dma_pushbuffer.pop();
+ dma_pushbuffer_subindex = 0;
+ }
- if (!non_main) {
- dma_mget = dma_get;
- }
+ if (command_list_header.size == 0) {
+ return true;
+ }
+
+ // Push buffer non-empty, read a word
+ const auto address = gpu.MemoryManager().GpuToCpuAddress(dma_get);
+ ASSERT_MSG(address, "Invalid GPU address");
+
+ command_headers.resize(command_list_header.size);
+
+ Memory::ReadBlock(*address, command_headers.data(), command_list_header.size * sizeof(u32));
+
+ for (const CommandHeader& command_header : command_headers) {
// now, see if we're in the middle of a command
if (dma_state.length_pending) {
@@ -91,22 +109,11 @@ bool DmaPusher::Step() {
break;
}
}
- } else if (ib_enable && !dma_pushbuffer.empty()) {
- // Current pushbuffer empty, but we have more IB entries to read
- const CommandList& command_list{dma_pushbuffer.front()};
- const CommandListHeader& command_list_header{command_list[dma_pushbuffer_subindex++]};
- dma_get = command_list_header.addr;
- dma_put = dma_get + command_list_header.size * sizeof(u32);
- non_main = command_list_header.is_non_main;
-
- if (dma_pushbuffer_subindex >= command_list.size()) {
- // We've gone through the current list, remove it from the queue
- dma_pushbuffer.pop();
- dma_pushbuffer_subindex = 0;
- }
- } else {
- // Otherwise, pushbuffer empty and IB empty or nonexistent - nothing to do
- return {};
+ }
+
+ if (!non_main) {
+ // TODO (degasus): This is dead code, as dma_mget is never read.
+ dma_mget = dma_put;
}
return true;
diff --git a/src/video_core/dma_pusher.h b/src/video_core/dma_pusher.h
index 1097e5c49..27a36348c 100644
--- a/src/video_core/dma_pusher.h
+++ b/src/video_core/dma_pusher.h
@@ -75,6 +75,8 @@ private:
GPU& gpu;
+ std::vector<CommandHeader> command_headers; ///< Buffer for list of commands fetched at once
+
std::queue<CommandList> dma_pushbuffer; ///< Queue of command lists to be processed
std::size_t dma_pushbuffer_subindex{}; ///< Index within a command list within the pushbuffer
@@ -89,11 +91,8 @@ private:
DmaState dma_state{};
bool dma_increment_once{};
- GPUVAddr dma_put{}; ///< pushbuffer current end address
- GPUVAddr dma_get{}; ///< pushbuffer current read address
GPUVAddr dma_mget{}; ///< main pushbuffer last read address
bool ib_enable{true}; ///< IB mode enabled
- bool non_main{}; ///< non-main pushbuffer active
};
} // namespace Tegra
diff --git a/src/video_core/engines/fermi_2d.cpp b/src/video_core/engines/fermi_2d.cpp
index ec1a57226..540dcc52c 100644
--- a/src/video_core/engines/fermi_2d.cpp
+++ b/src/video_core/engines/fermi_2d.cpp
@@ -44,10 +44,10 @@ void Fermi2D::HandleSurfaceCopy() {
const u32 src_blit_y2{
static_cast<u32>((regs.blit_src_y + (regs.blit_dst_height * regs.blit_dv_dy)) >> 32)};
- const MathUtil::Rectangle<u32> src_rect{src_blit_x1, src_blit_y1, src_blit_x2, src_blit_y2};
- const MathUtil::Rectangle<u32> dst_rect{regs.blit_dst_x, regs.blit_dst_y,
- regs.blit_dst_x + regs.blit_dst_width,
- regs.blit_dst_y + regs.blit_dst_height};
+ const Common::Rectangle<u32> src_rect{src_blit_x1, src_blit_y1, src_blit_x2, src_blit_y2};
+ const Common::Rectangle<u32> dst_rect{regs.blit_dst_x, regs.blit_dst_y,
+ regs.blit_dst_x + regs.blit_dst_width,
+ regs.blit_dst_y + regs.blit_dst_height};
if (!rasterizer.AccelerateSurfaceCopy(regs.src, regs.dst, src_rect, dst_rect)) {
UNIMPLEMENTED();
diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp
index 2d2136067..144e7fa82 100644
--- a/src/video_core/engines/maxwell_3d.cpp
+++ b/src/video_core/engines/maxwell_3d.cpp
@@ -107,21 +107,23 @@ void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) {
auto debug_context = system.GetGPUDebugContext();
+ const u32 method = method_call.method;
+
// It is an error to write to a register other than the current macro's ARG register before it
// has finished execution.
if (executing_macro != 0) {
- ASSERT(method_call.method == executing_macro + 1);
+ ASSERT(method == executing_macro + 1);
}
// Methods after 0xE00 are special, they're actually triggers for some microcode that was
// uploaded to the GPU during initialization.
- if (method_call.method >= MacroRegistersStart) {
+ if (method >= MacroRegistersStart) {
// We're trying to execute a macro
if (executing_macro == 0) {
// A macro call must begin by writing the macro method's register, not its argument.
- ASSERT_MSG((method_call.method % 2) == 0,
+ ASSERT_MSG((method % 2) == 0,
"Can't start macro execution by writing to the ARGS register");
- executing_macro = method_call.method;
+ executing_macro = method;
}
macro_params.push_back(method_call.argument);
@@ -133,66 +135,62 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) {
return;
}
- ASSERT_MSG(method_call.method < Regs::NUM_REGS,
+ ASSERT_MSG(method < Regs::NUM_REGS,
"Invalid Maxwell3D register, increase the size of the Regs structure");
if (debug_context) {
debug_context->OnEvent(Tegra::DebugContext::Event::MaxwellCommandLoaded, nullptr);
}
- if (regs.reg_array[method_call.method] != method_call.argument) {
- regs.reg_array[method_call.method] = method_call.argument;
+ if (regs.reg_array[method] != method_call.argument) {
+ regs.reg_array[method] = method_call.argument;
// Color buffers
constexpr u32 first_rt_reg = MAXWELL3D_REG_INDEX(rt);
constexpr u32 registers_per_rt = sizeof(regs.rt[0]) / sizeof(u32);
- if (method_call.method >= first_rt_reg &&
- method_call.method < first_rt_reg + registers_per_rt * Regs::NumRenderTargets) {
- const std::size_t rt_index = (method_call.method - first_rt_reg) / registers_per_rt;
- dirty_flags.color_buffer |= 1u << static_cast<u32>(rt_index);
+ if (method >= first_rt_reg &&
+ method < first_rt_reg + registers_per_rt * Regs::NumRenderTargets) {
+ const std::size_t rt_index = (method - first_rt_reg) / registers_per_rt;
+ dirty_flags.color_buffer.set(rt_index);
}
// Zeta buffer
constexpr u32 registers_in_zeta = sizeof(regs.zeta) / sizeof(u32);
- if (method_call.method == MAXWELL3D_REG_INDEX(zeta_enable) ||
- method_call.method == MAXWELL3D_REG_INDEX(zeta_width) ||
- method_call.method == MAXWELL3D_REG_INDEX(zeta_height) ||
- (method_call.method >= MAXWELL3D_REG_INDEX(zeta) &&
- method_call.method < MAXWELL3D_REG_INDEX(zeta) + registers_in_zeta)) {
+ if (method == MAXWELL3D_REG_INDEX(zeta_enable) ||
+ method == MAXWELL3D_REG_INDEX(zeta_width) ||
+ method == MAXWELL3D_REG_INDEX(zeta_height) ||
+ (method >= MAXWELL3D_REG_INDEX(zeta) &&
+ method < MAXWELL3D_REG_INDEX(zeta) + registers_in_zeta)) {
dirty_flags.zeta_buffer = true;
}
// Shader
constexpr u32 shader_registers_count =
sizeof(regs.shader_config[0]) * Regs::MaxShaderProgram / sizeof(u32);
- if (method_call.method >= MAXWELL3D_REG_INDEX(shader_config[0]) &&
- method_call.method < MAXWELL3D_REG_INDEX(shader_config[0]) + shader_registers_count) {
+ if (method >= MAXWELL3D_REG_INDEX(shader_config[0]) &&
+ method < MAXWELL3D_REG_INDEX(shader_config[0]) + shader_registers_count) {
dirty_flags.shaders = true;
}
// Vertex format
- if (method_call.method >= MAXWELL3D_REG_INDEX(vertex_attrib_format) &&
- method_call.method <
- MAXWELL3D_REG_INDEX(vertex_attrib_format) + regs.vertex_attrib_format.size()) {
+ if (method >= MAXWELL3D_REG_INDEX(vertex_attrib_format) &&
+ method < MAXWELL3D_REG_INDEX(vertex_attrib_format) + regs.vertex_attrib_format.size()) {
dirty_flags.vertex_attrib_format = true;
}
// Vertex buffer
- if (method_call.method >= MAXWELL3D_REG_INDEX(vertex_array) &&
- method_call.method < MAXWELL3D_REG_INDEX(vertex_array) + 4 * 32) {
- dirty_flags.vertex_array |=
- 1u << ((method_call.method - MAXWELL3D_REG_INDEX(vertex_array)) >> 2);
- } else if (method_call.method >= MAXWELL3D_REG_INDEX(vertex_array_limit) &&
- method_call.method < MAXWELL3D_REG_INDEX(vertex_array_limit) + 2 * 32) {
- dirty_flags.vertex_array |=
- 1u << ((method_call.method - MAXWELL3D_REG_INDEX(vertex_array_limit)) >> 1);
- } else if (method_call.method >= MAXWELL3D_REG_INDEX(instanced_arrays) &&
- method_call.method < MAXWELL3D_REG_INDEX(instanced_arrays) + 32) {
- dirty_flags.vertex_array |=
- 1u << (method_call.method - MAXWELL3D_REG_INDEX(instanced_arrays));
+ if (method >= MAXWELL3D_REG_INDEX(vertex_array) &&
+ method < MAXWELL3D_REG_INDEX(vertex_array) + 4 * 32) {
+ dirty_flags.vertex_array.set((method - MAXWELL3D_REG_INDEX(vertex_array)) >> 2);
+ } else if (method >= MAXWELL3D_REG_INDEX(vertex_array_limit) &&
+ method < MAXWELL3D_REG_INDEX(vertex_array_limit) + 2 * 32) {
+ dirty_flags.vertex_array.set((method - MAXWELL3D_REG_INDEX(vertex_array_limit)) >> 1);
+ } else if (method >= MAXWELL3D_REG_INDEX(instanced_arrays) &&
+ method < MAXWELL3D_REG_INDEX(instanced_arrays) + 32) {
+ dirty_flags.vertex_array.set(method - MAXWELL3D_REG_INDEX(instanced_arrays));
}
}
- switch (method_call.method) {
+ switch (method) {
case MAXWELL3D_REG_INDEX(macros.data): {
ProcessMacroUpload(method_call.argument);
break;
diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h
index 0e3873ffd..7fbf1026e 100644
--- a/src/video_core/engines/maxwell_3d.h
+++ b/src/video_core/engines/maxwell_3d.h
@@ -5,8 +5,10 @@
#pragma once
#include <array>
+#include <bitset>
#include <unordered_map>
#include <vector>
+
#include "common/assert.h"
#include "common/bit_field.h"
#include "common/common_funcs.h"
@@ -503,7 +505,7 @@ public:
f32 translate_z;
INSERT_PADDING_WORDS(2);
- MathUtil::Rectangle<s32> GetRect() const {
+ Common::Rectangle<s32> GetRect() const {
return {
GetX(), // left
GetY() + GetHeight(), // top
@@ -1094,19 +1096,18 @@ public:
MemoryManager& memory_manager;
struct DirtyFlags {
- u8 color_buffer = 0xFF;
- bool zeta_buffer = true;
-
- bool shaders = true;
+ std::bitset<8> color_buffer{0xFF};
+ std::bitset<32> vertex_array{0xFFFFFFFF};
bool vertex_attrib_format = true;
- u32 vertex_array = 0xFFFFFFFF;
+ bool zeta_buffer = true;
+ bool shaders = true;
void OnMemoryWrite() {
- color_buffer = 0xFF;
zeta_buffer = true;
shaders = true;
- vertex_array = 0xFFFFFFFF;
+ color_buffer.set();
+ vertex_array.set();
}
};
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index 1f425f90b..252592edd 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -376,9 +376,9 @@ enum class R2pMode : u64 {
};
enum class IpaInterpMode : u64 {
- Linear = 0,
- Perspective = 1,
- Flat = 2,
+ Pass = 0,
+ Multiply = 1,
+ Constant = 2,
Sc = 3,
};
diff --git a/src/video_core/engines/shader_header.h b/src/video_core/engines/shader_header.h
index cf2b76ff6..e86a7f04a 100644
--- a/src/video_core/engines/shader_header.h
+++ b/src/video_core/engines/shader_header.h
@@ -16,6 +16,13 @@ enum class OutputTopology : u32 {
TriangleStrip = 7,
};
+enum class AttributeUse : u8 {
+ Unused = 0,
+ Constant = 1,
+ Perspective = 2,
+ ScreenLinear = 3,
+};
+
// Documentation in:
// http://download.nvidia.com/open-gpu-doc/Shader-Program-Header/1/Shader-Program-Header.html#ImapTexture
struct Header {
@@ -84,9 +91,15 @@ struct Header {
} vtg;
struct {
- INSERT_PADDING_BYTES(3); // ImapSystemValuesA
- INSERT_PADDING_BYTES(1); // ImapSystemValuesB
- INSERT_PADDING_BYTES(32); // ImapGenericVector[32]
+ INSERT_PADDING_BYTES(3); // ImapSystemValuesA
+ INSERT_PADDING_BYTES(1); // ImapSystemValuesB
+ union {
+ BitField<0, 2, AttributeUse> x;
+ BitField<2, 2, AttributeUse> y;
+ BitField<4, 2, AttributeUse> w;
+ BitField<6, 2, AttributeUse> z;
+ u8 raw;
+ } imap_generic_vector[32];
INSERT_PADDING_BYTES(2); // ImapColor
INSERT_PADDING_BYTES(2); // ImapSystemValuesC
INSERT_PADDING_BYTES(10); // ImapFixedFncTexture[10]
@@ -103,6 +116,28 @@ struct Header {
const u32 bit = render_target * 4 + component;
return omap.target & (1 << bit);
}
+ AttributeUse GetAttributeIndexUse(u32 attribute, u32 index) const {
+ return static_cast<AttributeUse>(
+ (imap_generic_vector[attribute].raw >> (index * 2)) & 0x03);
+ }
+ AttributeUse GetAttributeUse(u32 attribute) const {
+ AttributeUse result = AttributeUse::Unused;
+ for (u32 i = 0; i < 4; i++) {
+ const auto index = GetAttributeIndexUse(attribute, i);
+ if (index == AttributeUse::Unused) {
+ continue;
+ }
+ if (result == AttributeUse::Unused || result == index) {
+ result = index;
+ continue;
+ }
+ LOG_CRITICAL(HW_GPU, "Generic Attribute Conflict in Interpolation Mode");
+ if (index == AttributeUse::Perspective) {
+ result = index;
+ }
+ }
+ return result;
+ }
} ps;
};
diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h
index 0f5bfdcbf..6313702f2 100644
--- a/src/video_core/gpu.h
+++ b/src/video_core/gpu.h
@@ -100,7 +100,7 @@ struct FramebufferConfig {
using TransformFlags = Service::NVFlinger::BufferQueue::BufferTransformFlags;
TransformFlags transform_flags;
- MathUtil::Rectangle<int> crop_rect;
+ Common::Rectangle<int> crop_rect;
};
namespace Engines {
diff --git a/src/video_core/rasterizer_cache.h b/src/video_core/rasterizer_cache.h
index bcf0c15a4..a7bcf26fb 100644
--- a/src/video_core/rasterizer_cache.h
+++ b/src/video_core/rasterizer_cache.h
@@ -129,6 +129,15 @@ protected:
return ++modified_ticks;
}
+ /// Flushes the specified object, updating appropriate cache state as needed
+ void FlushObject(const T& object) {
+ if (!object->IsDirty()) {
+ return;
+ }
+ object->Flush();
+ object->MarkAsModified(false, *this);
+ }
+
private:
/// Returns a list of cached objects from the specified memory region, ordered by access time
std::vector<T> GetSortedObjectsFromRegion(VAddr addr, u64 size) {
@@ -154,15 +163,6 @@ private:
return objects;
}
- /// Flushes the specified object, updating appropriate cache state as needed
- void FlushObject(const T& object) {
- if (!object->IsDirty()) {
- return;
- }
- object->Flush();
- object->MarkAsModified(false, *this);
- }
-
using ObjectSet = std::set<T>;
using ObjectCache = std::unordered_map<VAddr, T>;
using IntervalCache = boost::icl::interval_map<VAddr, ObjectSet>;
diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h
index b2a223705..6a1dc9cf6 100644
--- a/src/video_core/rasterizer_interface.h
+++ b/src/video_core/rasterizer_interface.h
@@ -47,8 +47,8 @@ public:
/// Attempt to use a faster method to perform a surface copy
virtual bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src,
const Tegra::Engines::Fermi2D::Regs::Surface& dst,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect) {
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect) {
return false;
}
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 12d876120..c8c1d6911 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -102,8 +102,8 @@ struct FramebufferCacheKey {
RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, Core::System& system,
ScreenInfo& info)
- : res_cache{*this}, shader_cache{*this, system}, emu_window{window}, screen_info{info},
- buffer_cache(*this, STREAM_BUFFER_SIZE), global_cache{*this} {
+ : res_cache{*this}, shader_cache{*this, system}, global_cache{*this}, emu_window{window},
+ screen_info{info}, buffer_cache(*this, STREAM_BUFFER_SIZE) {
// Create sampler objects
for (std::size_t i = 0; i < texture_samplers.size(); ++i) {
texture_samplers[i].Create();
@@ -200,7 +200,7 @@ GLuint RasterizerOpenGL::SetupVertexFormat() {
}
// Rebinding the VAO invalidates the vertex buffer bindings.
- gpu.dirty_flags.vertex_array = 0xFFFFFFFF;
+ gpu.dirty_flags.vertex_array.set();
state.draw.vertex_array = vao_entry.handle;
return vao_entry.handle;
@@ -210,14 +210,14 @@ void RasterizerOpenGL::SetupVertexBuffer(GLuint vao) {
auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
const auto& regs = gpu.regs;
- if (!gpu.dirty_flags.vertex_array)
+ if (gpu.dirty_flags.vertex_array.none())
return;
MICROPROFILE_SCOPE(OpenGL_VB);
// Upload all guest vertex arrays sequentially to our buffer
for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) {
- if (~gpu.dirty_flags.vertex_array & (1u << index))
+ if (!gpu.dirty_flags.vertex_array[index])
continue;
const auto& vertex_array = regs.vertex_array[index];
@@ -244,7 +244,7 @@ void RasterizerOpenGL::SetupVertexBuffer(GLuint vao) {
}
}
- gpu.dirty_flags.vertex_array = 0;
+ gpu.dirty_flags.vertex_array.reset();
}
DrawParameters RasterizerOpenGL::SetupDraw() {
@@ -488,13 +488,13 @@ std::pair<bool, bool> RasterizerOpenGL::ConfigureFramebuffers(
OpenGLState& current_state, bool using_color_fb, bool using_depth_fb, bool preserve_contents,
std::optional<std::size_t> single_color_target) {
MICROPROFILE_SCOPE(OpenGL_Framebuffer);
- const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+ auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
const auto& regs = gpu.regs;
const FramebufferConfigState fb_config_state{using_color_fb, using_depth_fb, preserve_contents,
single_color_target};
- if (fb_config_state == current_framebuffer_config_state && gpu.dirty_flags.color_buffer == 0 &&
- !gpu.dirty_flags.zeta_buffer) {
+ if (fb_config_state == current_framebuffer_config_state &&
+ gpu.dirty_flags.color_buffer.none() && !gpu.dirty_flags.zeta_buffer) {
// Only skip if the previous ConfigureFramebuffers call was from the same kind (multiple or
// single color targets). This is done because the guest registers may not change but the
// host framebuffer may contain different attachments
@@ -721,10 +721,10 @@ void RasterizerOpenGL::DrawArrays() {
// Add space for at least 18 constant buffers
buffer_size += Maxwell::MaxConstBuffers * (MaxConstbufferSize + uniform_buffer_alignment);
- bool invalidate = buffer_cache.Map(buffer_size);
+ const bool invalidate = buffer_cache.Map(buffer_size);
if (invalidate) {
// As all cached buffers are invalidated, we need to recheck their state.
- gpu.dirty_flags.vertex_array = 0xFFFFFFFF;
+ gpu.dirty_flags.vertex_array.set();
}
const GLuint vao = SetupVertexFormat();
@@ -738,9 +738,13 @@ void RasterizerOpenGL::DrawArrays() {
shader_program_manager->ApplyTo(state);
state.Apply();
+ res_cache.SignalPreDrawCall();
+
// Execute draw call
params.DispatchDraw();
+ res_cache.SignalPostDrawCall();
+
// Disable scissor test
state.viewports[0].scissor.enabled = false;
@@ -779,8 +783,8 @@ void RasterizerOpenGL::FlushAndInvalidateRegion(VAddr addr, u64 size) {
bool RasterizerOpenGL::AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src,
const Tegra::Engines::Fermi2D::Regs::Surface& dst,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect) {
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect) {
MICROPROFILE_SCOPE(OpenGL_Blits);
res_cache.FermiCopySurface(src, dst, src_rect, dst_rect);
return true;
@@ -1034,7 +1038,7 @@ void RasterizerOpenGL::SyncViewport(OpenGLState& current_state) {
for (std::size_t i = 0; i < viewport_count; i++) {
auto& viewport = current_state.viewports[i];
const auto& src = regs.viewports[i];
- const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[i].GetRect()};
+ const Common::Rectangle<s32> viewport_rect{regs.viewport_transform[i].GetRect()};
viewport.x = viewport_rect.left;
viewport.y = viewport_rect.bottom;
viewport.width = viewport_rect.GetWidth();
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h
index 258d62259..2f0524f85 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer.h
@@ -62,8 +62,8 @@ public:
void FlushAndInvalidateRegion(VAddr addr, u64 size) override;
bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src,
const Tegra::Engines::Fermi2D::Regs::Surface& dst,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect) override;
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect) override;
bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
u32 pixel_stride) override;
bool AccelerateDrawBatch(bool is_indexed) override;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
index 74200914e..5fdf1164d 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
@@ -3,6 +3,7 @@
// Refer to the license.txt file included.
#include <algorithm>
+#include <optional>
#include <glad/glad.h>
#include "common/alignment.h"
@@ -399,7 +400,7 @@ static const FormatTuple& GetFormatTuple(PixelFormat pixel_format, ComponentType
return format;
}
-MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
+Common::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
u32 actual_height{std::max(1U, unaligned_height >> mip_level)};
if (IsPixelFormatASTC(pixel_format)) {
// ASTC formats must stop at the ATSC block size boundary
@@ -549,6 +550,8 @@ CachedSurface::CachedSurface(const SurfaceParams& params)
// alternatives. This signals a bug on those functions.
const auto width = static_cast<GLsizei>(params.MipWidth(0));
const auto height = static_cast<GLsizei>(params.MipHeight(0));
+ memory_size = params.MemorySize();
+ reinterpreted = false;
const auto& format_tuple = GetFormatTuple(params.pixel_format, params.component_type);
gl_internal_format = format_tuple.internal_format;
@@ -962,30 +965,31 @@ Surface RasterizerCacheOpenGL::GetColorBufferSurface(std::size_t index, bool pre
auto& gpu{Core::System::GetInstance().GPU().Maxwell3D()};
const auto& regs{gpu.regs};
- if ((gpu.dirty_flags.color_buffer & (1u << static_cast<u32>(index))) == 0) {
+ if (!gpu.dirty_flags.color_buffer[index]) {
return last_color_buffers[index];
}
- gpu.dirty_flags.color_buffer &= ~(1u << static_cast<u32>(index));
+ gpu.dirty_flags.color_buffer.reset(index);
ASSERT(index < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets);
if (index >= regs.rt_control.count) {
- return last_color_buffers[index] = {};
+ return current_color_buffers[index] = {};
}
if (regs.rt[index].Address() == 0 || regs.rt[index].format == Tegra::RenderTargetFormat::NONE) {
- return last_color_buffers[index] = {};
+ return current_color_buffers[index] = {};
}
const SurfaceParams color_params{SurfaceParams::CreateForFramebuffer(index)};
- return last_color_buffers[index] = GetSurface(color_params, preserve_contents);
+ return current_color_buffers[index] = GetSurface(color_params, preserve_contents);
}
void RasterizerCacheOpenGL::LoadSurface(const Surface& surface) {
surface->LoadGLBuffer();
surface->UploadGLTexture(read_framebuffer.handle, draw_framebuffer.handle);
surface->MarkAsModified(false, *this);
+ surface->MarkForReload(false);
}
Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool preserve_contents) {
@@ -997,18 +1001,23 @@ Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool pres
Surface surface{TryGet(params.addr)};
if (surface) {
if (surface->GetSurfaceParams().IsCompatibleSurface(params)) {
- // Use the cached surface as-is
+ // Use the cached surface as-is unless it's not synced with memory
+ if (surface->MustReload())
+ LoadSurface(surface);
return surface;
} else if (preserve_contents) {
// If surface parameters changed and we care about keeping the previous data, recreate
// the surface from the old one
Surface new_surface{RecreateSurface(surface, params)};
- Unregister(surface);
+ UnregisterSurface(surface);
Register(new_surface);
+ if (new_surface->IsUploaded()) {
+ RegisterReinterpretSurface(new_surface);
+ }
return new_surface;
} else {
// Delete the old surface before creating a new one to prevent collisions.
- Unregister(surface);
+ UnregisterSurface(surface);
}
}
@@ -1062,8 +1071,8 @@ void RasterizerCacheOpenGL::FastLayeredCopySurface(const Surface& src_surface,
}
static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect, GLuint read_fb_handle,
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect, GLuint read_fb_handle,
GLuint draw_fb_handle, GLenum src_attachment = 0, GLenum dst_attachment = 0,
std::size_t cubemap_face = 0) {
@@ -1193,7 +1202,7 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
void RasterizerCacheOpenGL::FermiCopySurface(
const Tegra::Engines::Fermi2D::Regs::Surface& src_config,
const Tegra::Engines::Fermi2D::Regs::Surface& dst_config,
- const MathUtil::Rectangle<u32>& src_rect, const MathUtil::Rectangle<u32>& dst_rect) {
+ const Common::Rectangle<u32>& src_rect, const Common::Rectangle<u32>& dst_rect) {
const auto& src_params = SurfaceParams::CreateForFermiCopySurface(src_config);
const auto& dst_params = SurfaceParams::CreateForFermiCopySurface(dst_config);
@@ -1257,7 +1266,11 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
case SurfaceTarget::TextureCubemap:
case SurfaceTarget::Texture2DArray:
case SurfaceTarget::TextureCubeArray:
- FastLayeredCopySurface(old_surface, new_surface);
+ if (old_params.pixel_format == new_params.pixel_format)
+ FastLayeredCopySurface(old_surface, new_surface);
+ else {
+ AccurateCopySurface(old_surface, new_surface);
+ }
break;
default:
LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
@@ -1286,4 +1299,107 @@ Surface RasterizerCacheOpenGL::TryGetReservedSurface(const SurfaceParams& params
return {};
}
+static std::optional<u32> TryFindBestMipMap(std::size_t memory, const SurfaceParams params,
+ u32 height) {
+ for (u32 i = 0; i < params.max_mip_level; i++) {
+ if (memory == params.GetMipmapSingleSize(i) && params.MipHeight(i) == height) {
+ return {i};
+ }
+ }
+ return {};
+}
+
+static std::optional<u32> TryFindBestLayer(VAddr addr, const SurfaceParams params, u32 mipmap) {
+ const std::size_t size = params.LayerMemorySize();
+ VAddr start = params.addr + params.GetMipmapLevelOffset(mipmap);
+ for (u32 i = 0; i < params.depth; i++) {
+ if (start == addr) {
+ return {i};
+ }
+ start += size;
+ }
+ return {};
+}
+
+static bool LayerFitReinterpretSurface(RasterizerCacheOpenGL& cache, const Surface render_surface,
+ const Surface blitted_surface) {
+ const auto& dst_params = blitted_surface->GetSurfaceParams();
+ const auto& src_params = render_surface->GetSurfaceParams();
+ const std::size_t src_memory_size = src_params.size_in_bytes;
+ const std::optional<u32> level =
+ TryFindBestMipMap(src_memory_size, dst_params, src_params.height);
+ if (level.has_value()) {
+ if (src_params.width == dst_params.MipWidthGobAligned(*level) &&
+ src_params.height == dst_params.MipHeight(*level) &&
+ src_params.block_height >= dst_params.MipBlockHeight(*level)) {
+ const std::optional<u32> slot =
+ TryFindBestLayer(render_surface->GetAddr(), dst_params, *level);
+ if (slot.has_value()) {
+ glCopyImageSubData(render_surface->Texture().handle,
+ SurfaceTargetToGL(src_params.target), 0, 0, 0, 0,
+ blitted_surface->Texture().handle,
+ SurfaceTargetToGL(dst_params.target), *level, 0, 0, *slot,
+ dst_params.MipWidth(*level), dst_params.MipHeight(*level), 1);
+ blitted_surface->MarkAsModified(true, cache);
+ return true;
+ }
+ }
+ }
+ return false;
+}
+
+static bool IsReinterpretInvalid(const Surface render_surface, const Surface blitted_surface) {
+ const VAddr bound1 = blitted_surface->GetAddr() + blitted_surface->GetMemorySize();
+ const VAddr bound2 = render_surface->GetAddr() + render_surface->GetMemorySize();
+ if (bound2 > bound1)
+ return true;
+ const auto& dst_params = blitted_surface->GetSurfaceParams();
+ const auto& src_params = render_surface->GetSurfaceParams();
+ return (dst_params.component_type != src_params.component_type);
+}
+
+static bool IsReinterpretInvalidSecond(const Surface render_surface,
+ const Surface blitted_surface) {
+ const auto& dst_params = blitted_surface->GetSurfaceParams();
+ const auto& src_params = render_surface->GetSurfaceParams();
+ return (dst_params.height > src_params.height && dst_params.width > src_params.width);
+}
+
+bool RasterizerCacheOpenGL::PartialReinterpretSurface(Surface triggering_surface,
+ Surface intersect) {
+ if (IsReinterpretInvalid(triggering_surface, intersect)) {
+ UnregisterSurface(intersect);
+ return false;
+ }
+ if (!LayerFitReinterpretSurface(*this, triggering_surface, intersect)) {
+ if (IsReinterpretInvalidSecond(triggering_surface, intersect)) {
+ UnregisterSurface(intersect);
+ return false;
+ }
+ FlushObject(intersect);
+ FlushObject(triggering_surface);
+ intersect->MarkForReload(true);
+ }
+ return true;
+}
+
+void RasterizerCacheOpenGL::SignalPreDrawCall() {
+ if (texception && GLAD_GL_ARB_texture_barrier) {
+ glTextureBarrier();
+ }
+ texception = false;
+}
+
+void RasterizerCacheOpenGL::SignalPostDrawCall() {
+ for (u32 i = 0; i < Maxwell::NumRenderTargets; i++) {
+ if (current_color_buffers[i] != nullptr) {
+ Surface intersect = CollideOnReinterpretedSurface(current_color_buffers[i]->GetAddr());
+ if (intersect != nullptr) {
+ PartialReinterpretSurface(current_color_buffers[i], intersect);
+ texception = true;
+ }
+ }
+ }
+}
+
} // namespace OpenGL
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
index 89d733c50..797bbdc9c 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
@@ -28,12 +28,13 @@ namespace OpenGL {
class CachedSurface;
using Surface = std::shared_ptr<CachedSurface>;
-using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, MathUtil::Rectangle<u32>>;
+using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, Common::Rectangle<u32>>;
using SurfaceTarget = VideoCore::Surface::SurfaceTarget;
using SurfaceType = VideoCore::Surface::SurfaceType;
using PixelFormat = VideoCore::Surface::PixelFormat;
using ComponentType = VideoCore::Surface::ComponentType;
+using Maxwell = Tegra::Engines::Maxwell3D::Regs;
struct SurfaceParams {
enum class SurfaceClass {
@@ -71,7 +72,7 @@ struct SurfaceParams {
}
/// Returns the rectangle corresponding to this surface
- MathUtil::Rectangle<u32> GetRect(u32 mip_level = 0) const;
+ Common::Rectangle<u32> GetRect(u32 mip_level = 0) const;
/// Returns the total size of this surface in bytes, adjusted for compression
std::size_t SizeInBytesRaw(bool ignore_tiled = false) const {
@@ -140,10 +141,18 @@ struct SurfaceParams {
return offset;
}
+ std::size_t GetMipmapSingleSize(u32 mip_level) const {
+ return InnerMipmapMemorySize(mip_level, false, is_layered);
+ }
+
u32 MipWidth(u32 mip_level) const {
return std::max(1U, width >> mip_level);
}
+ u32 MipWidthGobAligned(u32 mip_level) const {
+ return Common::AlignUp(std::max(1U, width >> mip_level), 64U * 8U / GetFormatBpp());
+ }
+
u32 MipHeight(u32 mip_level) const {
return std::max(1U, height >> mip_level);
}
@@ -346,6 +355,10 @@ public:
return cached_size_in_bytes;
}
+ std::size_t GetMemorySize() const {
+ return memory_size;
+ }
+
void Flush() override {
FlushGLBuffer();
}
@@ -395,6 +408,26 @@ public:
Tegra::Texture::SwizzleSource swizzle_z,
Tegra::Texture::SwizzleSource swizzle_w);
+ void MarkReinterpreted() {
+ reinterpreted = true;
+ }
+
+ bool IsReinterpreted() const {
+ return reinterpreted;
+ }
+
+ void MarkForReload(bool reload) {
+ must_reload = reload;
+ }
+
+ bool MustReload() const {
+ return must_reload;
+ }
+
+ bool IsUploaded() const {
+ return params.identity == SurfaceParams::SurfaceClass::Uploaded;
+ }
+
private:
void UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, GLuint draw_fb_handle);
@@ -408,6 +441,9 @@ private:
GLenum gl_internal_format{};
std::size_t cached_size_in_bytes{};
std::array<GLenum, 4> swizzle{GL_RED, GL_GREEN, GL_BLUE, GL_ALPHA};
+ std::size_t memory_size;
+ bool reinterpreted = false;
+ bool must_reload = false;
};
class RasterizerCacheOpenGL final : public RasterizerCache<Surface> {
@@ -430,8 +466,11 @@ public:
/// Copies the contents of one surface to another
void FermiCopySurface(const Tegra::Engines::Fermi2D::Regs::Surface& src_config,
const Tegra::Engines::Fermi2D::Regs::Surface& dst_config,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect);
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect);
+
+ void SignalPreDrawCall();
+ void SignalPostDrawCall();
private:
void LoadSurface(const Surface& surface);
@@ -449,6 +488,10 @@ private:
/// Tries to get a reserved surface for the specified parameters
Surface TryGetReservedSurface(const SurfaceParams& params);
+ // Partialy reinterpret a surface based on a triggering_surface that collides with it.
+ // returns true if the reinterpret was successful, false in case it was not.
+ bool PartialReinterpretSurface(Surface triggering_surface, Surface intersect);
+
/// Performs a slow but accurate surface copy, flushing to RAM and reinterpreting the data
void AccurateCopySurface(const Surface& src_surface, const Surface& dst_surface);
void FastLayeredCopySurface(const Surface& src_surface, const Surface& dst_surface);
@@ -465,12 +508,50 @@ private:
OGLFramebuffer read_framebuffer;
OGLFramebuffer draw_framebuffer;
+ bool texception = false;
+
/// Use a Pixel Buffer Object to download the previous texture and then upload it to the new one
/// using the new format.
OGLBuffer copy_pbo;
- std::array<Surface, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> last_color_buffers;
+ std::array<Surface, Maxwell::NumRenderTargets> last_color_buffers;
+ std::array<Surface, Maxwell::NumRenderTargets> current_color_buffers;
Surface last_depth_buffer;
+
+ using SurfaceIntervalCache = boost::icl::interval_map<VAddr, Surface>;
+ using SurfaceInterval = typename SurfaceIntervalCache::interval_type;
+
+ static auto GetReinterpretInterval(const Surface& object) {
+ return SurfaceInterval::right_open(object->GetAddr() + 1,
+ object->GetAddr() + object->GetMemorySize() - 1);
+ }
+
+ // Reinterpreted surfaces are very fragil as the game may keep rendering into them.
+ SurfaceIntervalCache reinterpreted_surfaces;
+
+ void RegisterReinterpretSurface(Surface reinterpret_surface) {
+ auto interval = GetReinterpretInterval(reinterpret_surface);
+ reinterpreted_surfaces.insert({interval, reinterpret_surface});
+ reinterpret_surface->MarkReinterpreted();
+ }
+
+ Surface CollideOnReinterpretedSurface(VAddr addr) const {
+ const SurfaceInterval interval{addr};
+ for (auto& pair :
+ boost::make_iterator_range(reinterpreted_surfaces.equal_range(interval))) {
+ return pair.second;
+ }
+ return nullptr;
+ }
+
+ /// Unregisters an object from the cache
+ void UnregisterSurface(const Surface& object) {
+ if (object->IsReinterpreted()) {
+ auto interval = GetReinterpretInterval(object);
+ reinterpreted_surfaces.erase(interval);
+ }
+ Unregister(object);
+ }
};
} // namespace OpenGL
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
index db18f4dbe..72ff6ac6a 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
@@ -20,6 +20,7 @@
namespace OpenGL::GLShader {
using Tegra::Shader::Attribute;
+using Tegra::Shader::AttributeUse;
using Tegra::Shader::Header;
using Tegra::Shader::IpaInterpMode;
using Tegra::Shader::IpaMode;
@@ -288,34 +289,22 @@ private:
code.AddNewLine();
}
- std::string GetInputFlags(const IpaMode& input_mode) {
- const IpaSampleMode sample_mode = input_mode.sampling_mode;
- const IpaInterpMode interp_mode = input_mode.interpolation_mode;
+ std::string GetInputFlags(AttributeUse attribute) {
std::string out;
- switch (interp_mode) {
- case IpaInterpMode::Flat:
+ switch (attribute) {
+ case AttributeUse::Constant:
out += "flat ";
break;
- case IpaInterpMode::Linear:
+ case AttributeUse::ScreenLinear:
out += "noperspective ";
break;
- case IpaInterpMode::Perspective:
+ case AttributeUse::Perspective:
// Default, Smooth
break;
default:
- UNIMPLEMENTED_MSG("Unhandled IPA interp mode: {}", static_cast<u32>(interp_mode));
- }
- switch (sample_mode) {
- case IpaSampleMode::Centroid:
- // It can be implemented with the "centroid " keyword in GLSL
- UNIMPLEMENTED_MSG("Unimplemented IPA sampler mode centroid");
- break;
- case IpaSampleMode::Default:
- // Default, n/a
- break;
- default:
- UNIMPLEMENTED_MSG("Unimplemented IPA sampler mode: {}", static_cast<u32>(sample_mode));
+ LOG_CRITICAL(HW_GPU, "Unused attribute being fetched");
+ UNREACHABLE();
}
return out;
}
@@ -324,16 +313,11 @@ private:
const auto& attributes = ir.GetInputAttributes();
for (const auto element : attributes) {
const Attribute::Index index = element.first;
- const IpaMode& input_mode = *element.second.begin();
if (index < Attribute::Index::Attribute_0 || index > Attribute::Index::Attribute_31) {
// Skip when it's not a generic attribute
continue;
}
- ASSERT(element.second.size() > 0);
- UNIMPLEMENTED_IF_MSG(element.second.size() > 1,
- "Multiple input flag modes are not supported in GLSL");
-
// TODO(bunnei): Use proper number of elements for these
u32 idx = static_cast<u32>(index) - static_cast<u32>(Attribute::Index::Attribute_0);
if (stage != ShaderStage::Vertex) {
@@ -345,8 +329,14 @@ private:
if (stage == ShaderStage::Geometry) {
attr = "gs_" + attr + "[]";
}
- code.AddLine("layout (location = " + std::to_string(idx) + ") " +
- GetInputFlags(input_mode) + "in vec4 " + attr + ';');
+ std::string suffix;
+ if (stage == ShaderStage::Fragment) {
+ const auto input_mode =
+ header.ps.GetAttributeUse(idx - GENERIC_VARYING_START_LOCATION);
+ suffix = GetInputFlags(input_mode);
+ }
+ code.AddLine("layout (location = " + std::to_string(idx) + ") " + suffix + "in vec4 " +
+ attr + ';');
}
if (!attributes.empty())
code.AddNewLine();
@@ -1584,4 +1574,4 @@ ProgramResult Decompile(const ShaderIR& ir, Maxwell::ShaderStage stage, const st
return {decompiler.GetResult(), decompiler.GetShaderEntries()};
}
-} // namespace OpenGL::GLShader \ No newline at end of file
+} // namespace OpenGL::GLShader
diff --git a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp
index 81882822b..82fc4d44b 100644
--- a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp
@@ -2,8 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#pragma once
-
#include <cstring>
#include <fmt/format.h>
#include <lz4.h>
diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp
index 04e1db911..7d96649af 100644
--- a/src/video_core/renderer_opengl/gl_shader_gen.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp
@@ -124,7 +124,7 @@ layout (location = 5) out vec4 FragColor5;
layout (location = 6) out vec4 FragColor6;
layout (location = 7) out vec4 FragColor7;
-layout (location = 0) in vec4 position;
+layout (location = 0) in noperspective vec4 position;
layout (std140, binding = EMULATION_UBO_BINDING) uniform fs_config {
vec4 viewport_flip;
@@ -172,4 +172,4 @@ void main() {
return {out, program.second};
}
-} // namespace OpenGL::GLShader \ No newline at end of file
+} // namespace OpenGL::GLShader
diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp
index 272fc2e8e..e60b2eb44 100644
--- a/src/video_core/renderer_opengl/renderer_opengl.cpp
+++ b/src/video_core/renderer_opengl/renderer_opengl.cpp
@@ -257,6 +257,7 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture,
const Tegra::FramebufferConfig& framebuffer) {
texture.width = framebuffer.width;
texture.height = framebuffer.height;
+ texture.pixel_format = framebuffer.pixel_format;
GLint internal_format;
switch (framebuffer.pixel_format) {
diff --git a/src/video_core/renderer_opengl/renderer_opengl.h b/src/video_core/renderer_opengl/renderer_opengl.h
index 7e13e566b..c168fa89e 100644
--- a/src/video_core/renderer_opengl/renderer_opengl.h
+++ b/src/video_core/renderer_opengl/renderer_opengl.h
@@ -39,7 +39,7 @@ struct TextureInfo {
/// Structure used for storing information about the display target for the Switch screen
struct ScreenInfo {
GLuint display_texture;
- const MathUtil::Rectangle<float> display_texcoords{0.0f, 0.0f, 1.0f, 1.0f};
+ const Common::Rectangle<float> display_texcoords{0.0f, 0.0f, 1.0f, 1.0f};
TextureInfo texture;
};
@@ -102,7 +102,7 @@ private:
/// Used for transforming the framebuffer orientation
Tegra::FramebufferConfig::TransformFlags framebuffer_transform_flags;
- MathUtil::Rectangle<int> framebuffer_crop_rect;
+ Common::Rectangle<int> framebuffer_crop_rect;
};
} // namespace OpenGL
diff --git a/src/video_core/renderer_vulkan/vk_memory_manager.cpp b/src/video_core/renderer_vulkan/vk_memory_manager.cpp
index 17ee93b91..0451babbf 100644
--- a/src/video_core/renderer_vulkan/vk_memory_manager.cpp
+++ b/src/video_core/renderer_vulkan/vk_memory_manager.cpp
@@ -238,7 +238,7 @@ bool VKMemoryManager::AllocMemory(vk::MemoryPropertyFlags wanted_properties, u32
VKMemoryCommitImpl::VKMemoryCommitImpl(VKMemoryAllocation* allocation, vk::DeviceMemory memory,
u8* data, u64 begin, u64 end)
- : allocation{allocation}, memory{memory}, data{data}, interval(std::make_pair(begin, end)) {}
+ : interval(std::make_pair(begin, end)), memory{memory}, allocation{allocation}, data{data} {}
VKMemoryCommitImpl::~VKMemoryCommitImpl() {
allocation->Free(this);
diff --git a/src/video_core/shader/decode/memory.cpp b/src/video_core/shader/decode/memory.cpp
index 55ec601ff..38f01ca50 100644
--- a/src/video_core/shader/decode/memory.cpp
+++ b/src/video_core/shader/decode/memory.cpp
@@ -48,7 +48,7 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) {
UNIMPLEMENTED_IF_MSG((instr.attribute.fmt20.immediate.Value() % sizeof(u32)) != 0,
"Unaligned attribute loads are not supported");
- Tegra::Shader::IpaMode input_mode{Tegra::Shader::IpaInterpMode::Perspective,
+ Tegra::Shader::IpaMode input_mode{Tegra::Shader::IpaInterpMode::Pass,
Tegra::Shader::IpaSampleMode::Default};
u64 next_element = instr.attribute.fmt20.element;
diff --git a/src/video_core/shader/decode/other.cpp b/src/video_core/shader/decode/other.cpp
index f9502e3d0..d750a2936 100644
--- a/src/video_core/shader/decode/other.cpp
+++ b/src/video_core/shader/decode/other.cpp
@@ -135,7 +135,18 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
instr.ipa.sample_mode.Value()};
const Node attr = GetInputAttribute(attribute.index, attribute.element, input_mode);
- const Node value = GetSaturatedFloat(attr, instr.ipa.saturate);
+ Node value = attr;
+ const Tegra::Shader::Attribute::Index index = attribute.index.Value();
+ if (index >= Tegra::Shader::Attribute::Index::Attribute_0 &&
+ index <= Tegra::Shader::Attribute::Index::Attribute_31) {
+ // TODO(Blinkhawk): There are cases where a perspective attribute use PASS.
+ // In theory by setting them as perspective, OpenGL does the perspective correction.
+ // A way must figured to reverse the last step of it.
+ if (input_mode.interpolation_mode == Tegra::Shader::IpaInterpMode::Multiply) {
+ value = Operation(OperationCode::FMul, PRECISE, value, GetRegister(instr.gpr20));
+ }
+ }
+ value = GetSaturatedFloat(value, instr.ipa.saturate);
SetRegister(bb, instr.gpr0, value);
break;
@@ -175,4 +186,4 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
return pc;
}
-} // namespace VideoCommon::Shader \ No newline at end of file
+} // namespace VideoCommon::Shader
diff --git a/src/video_core/shader/track.cpp b/src/video_core/shader/track.cpp
index be4635342..33b071747 100644
--- a/src/video_core/shader/track.cpp
+++ b/src/video_core/shader/track.cpp
@@ -20,9 +20,9 @@ std::pair<Node, s64> FindOperation(const NodeBlock& code, s64 cursor,
return {node, cursor};
}
if (const auto conditional = std::get_if<ConditionalNode>(node)) {
- const auto& code = conditional->GetCode();
- const auto [found, internal_cursor] =
- FindOperation(code, static_cast<s64>(code.size() - 1), operation_code);
+ const auto& conditional_code = conditional->GetCode();
+ const auto [found, internal_cursor] = FindOperation(
+ conditional_code, static_cast<s64>(conditional_code.size() - 1), operation_code);
if (found)
return {found, cursor};
}
@@ -58,8 +58,8 @@ Node ShaderIR::TrackCbuf(Node tracked, const NodeBlock& code, s64 cursor) {
return nullptr;
}
if (const auto conditional = std::get_if<ConditionalNode>(tracked)) {
- const auto& code = conditional->GetCode();
- return TrackCbuf(tracked, code, static_cast<s64>(code.size()));
+ const auto& conditional_code = conditional->GetCode();
+ return TrackCbuf(tracked, conditional_code, static_cast<s64>(conditional_code.size()));
}
return nullptr;
}