summaryrefslogtreecommitdiffstats
path: root/src/video_core
diff options
context:
space:
mode:
Diffstat (limited to 'src/video_core')
-rw-r--r--src/video_core/CMakeLists.txt3
-rw-r--r--src/video_core/engines/maxwell_3d.h12
-rw-r--r--src/video_core/engines/shader_bytecode.h9
-rw-r--r--src/video_core/engines/shader_header.h11
-rw-r--r--src/video_core/memory_manager.cpp7
-rw-r--r--src/video_core/memory_manager.h3
-rw-r--r--src/video_core/morton.cpp353
-rw-r--r--src/video_core/morton.h21
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer.cpp49
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer.h31
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer_cache.cpp209
-rw-r--r--src/video_core/renderer_opengl/gl_shader_decompiler.cpp210
-rw-r--r--src/video_core/renderer_opengl/gl_state.cpp37
-rw-r--r--src/video_core/renderer_opengl/gl_state.h9
-rw-r--r--src/video_core/renderer_opengl/renderer_opengl.cpp4
-rw-r--r--src/video_core/utils.h164
16 files changed, 647 insertions, 485 deletions
diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt
index a780215c1..3f906a517 100644
--- a/src/video_core/CMakeLists.txt
+++ b/src/video_core/CMakeLists.txt
@@ -21,6 +21,8 @@ add_library(video_core STATIC
macro_interpreter.h
memory_manager.cpp
memory_manager.h
+ morton.cpp
+ morton.h
rasterizer_cache.cpp
rasterizer_cache.h
rasterizer_interface.h
@@ -62,7 +64,6 @@ add_library(video_core STATIC
textures/decoders.cpp
textures/decoders.h
textures/texture.h
- utils.h
video_core.cpp
video_core.h
)
diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h
index eff6abd55..4f137e693 100644
--- a/src/video_core/engines/maxwell_3d.h
+++ b/src/video_core/engines/maxwell_3d.h
@@ -631,7 +631,16 @@ public:
}
} zeta;
- INSERT_PADDING_WORDS(0x5B);
+ INSERT_PADDING_WORDS(0x41);
+
+ union {
+ BitField<0, 4, u32> stencil;
+ BitField<4, 4, u32> unknown;
+ BitField<8, 4, u32> scissor;
+ BitField<12, 4, u32> viewport;
+ } clear_flags;
+
+ INSERT_PADDING_WORDS(0x19);
std::array<VertexAttribute, NumVertexAttributes> vertex_attrib_format;
@@ -1134,6 +1143,7 @@ ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7);
ASSERT_REG_POSITION(color_mask_common, 0x3E4);
ASSERT_REG_POSITION(rt_separate_frag_data, 0x3EB);
ASSERT_REG_POSITION(zeta, 0x3F8);
+ASSERT_REG_POSITION(clear_flags, 0x43E);
ASSERT_REG_POSITION(vertex_attrib_format, 0x458);
ASSERT_REG_POSITION(rt_control, 0x487);
ASSERT_REG_POSITION(zeta_width, 0x48a);
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index c5f502ce1..639bedb80 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -82,6 +82,8 @@ union Attribute {
Position = 7,
Attribute_0 = 8,
Attribute_31 = 39,
+ ClipDistances0123 = 44,
+ ClipDistances4567 = 45,
PointCoord = 46,
// This attribute contains a tuple of (~, ~, InstanceId, VertexId) when inside a vertex
// shader, and a tuple of (TessCoord.x, TessCoord.y, TessCoord.z, ~) when inside a Tess Eval
@@ -262,7 +264,7 @@ enum class FlowCondition : u64 {
Fcsm_Tr = 0x1C, // TODO(bunnei): What is this used for?
};
-enum class ControlCode : u64 {
+enum class ConditionCode : u64 {
F = 0,
LT = 1,
EQ = 2,
@@ -570,7 +572,6 @@ union Instruction {
BitField<39, 2, u64> tab5cb8_2;
BitField<41, 3, u64> tab5c68_1;
BitField<44, 2, u64> tab5c68_0;
- BitField<47, 1, u64> cc;
BitField<48, 1, u64> negate_b;
} fmul;
@@ -832,7 +833,7 @@ union Instruction {
union {
BitField<0, 3, u64> pred0;
BitField<3, 3, u64> pred3;
- BitField<8, 5, ControlCode> cc; // flag in cc
+ BitField<8, 5, ConditionCode> cc; // flag in cc
BitField<39, 3, u64> pred39;
BitField<42, 1, u64> neg_pred39;
BitField<45, 4, PredOperation> op; // op with pred39
@@ -1236,7 +1237,7 @@ union Instruction {
BitField<60, 1, u64> is_b_gpr;
BitField<59, 1, u64> is_c_gpr;
BitField<20, 24, s64> smem_imm;
- BitField<0, 5, ControlCode> flow_control_code;
+ BitField<0, 5, ConditionCode> flow_condition_code;
Attribute attribute;
Sampler sampler;
diff --git a/src/video_core/engines/shader_header.h b/src/video_core/engines/shader_header.h
index a0e015c4b..99c34649f 100644
--- a/src/video_core/engines/shader_header.h
+++ b/src/video_core/engines/shader_header.h
@@ -62,7 +62,16 @@ struct Header {
INSERT_PADDING_BYTES(1); // ImapSystemValuesB
INSERT_PADDING_BYTES(16); // ImapGenericVector[32]
INSERT_PADDING_BYTES(2); // ImapColor
- INSERT_PADDING_BYTES(2); // ImapSystemValuesC
+ union {
+ BitField<0, 8, u16> clip_distances;
+ BitField<8, 1, u16> point_sprite_s;
+ BitField<9, 1, u16> point_sprite_t;
+ BitField<10, 1, u16> fog_coordinate;
+ BitField<12, 1, u16> tessellation_eval_point_u;
+ BitField<13, 1, u16> tessellation_eval_point_v;
+ BitField<14, 1, u16> instance_id;
+ BitField<15, 1, u16> vertex_id;
+ };
INSERT_PADDING_BYTES(5); // ImapFixedFncTexture[10]
INSERT_PADDING_BYTES(1); // ImapReserved
INSERT_PADDING_BYTES(3); // OmapSystemValuesA
diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp
index 77a20bb84..47247f097 100644
--- a/src/video_core/memory_manager.cpp
+++ b/src/video_core/memory_manager.cpp
@@ -9,6 +9,13 @@
namespace Tegra {
+MemoryManager::MemoryManager() {
+ // Mark the first page as reserved, so that 0 is not a valid GPUVAddr. Otherwise, games might
+ // try to use 0 as a valid address, which is also used to mean nullptr. This fixes a bug with
+ // Undertale using 0 for a render target.
+ PageSlot(0) = static_cast<u64>(PageStatus::Reserved);
+}
+
GPUVAddr MemoryManager::AllocateSpace(u64 size, u64 align) {
const std::optional<GPUVAddr> gpu_addr{FindFreeBlock(0, size, align, PageStatus::Unmapped)};
diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h
index 4eb338aa2..fb03497ca 100644
--- a/src/video_core/memory_manager.h
+++ b/src/video_core/memory_manager.h
@@ -18,7 +18,7 @@ using GPUVAddr = u64;
class MemoryManager final {
public:
- MemoryManager() = default;
+ MemoryManager();
GPUVAddr AllocateSpace(u64 size, u64 align);
GPUVAddr AllocateSpace(GPUVAddr gpu_addr, u64 size, u64 align);
@@ -37,6 +37,7 @@ private:
enum class PageStatus : u64 {
Unmapped = 0xFFFFFFFFFFFFFFFFULL,
Allocated = 0xFFFFFFFFFFFFFFFEULL,
+ Reserved = 0xFFFFFFFFFFFFFFFDULL,
};
std::optional<GPUVAddr> FindFreeBlock(GPUVAddr region_start, u64 size, u64 align,
diff --git a/src/video_core/morton.cpp b/src/video_core/morton.cpp
new file mode 100644
index 000000000..f14abba7d
--- /dev/null
+++ b/src/video_core/morton.cpp
@@ -0,0 +1,353 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <array>
+#include <cstring>
+#include "common/assert.h"
+#include "common/common_types.h"
+#include "core/memory.h"
+#include "video_core/morton.h"
+#include "video_core/surface.h"
+#include "video_core/textures/decoders.h"
+
+namespace VideoCore {
+
+using Surface::GetBytesPerPixel;
+using Surface::PixelFormat;
+
+using MortonCopyFn = void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr);
+using ConversionArray = std::array<MortonCopyFn, Surface::MaxPixelFormat>;
+
+template <bool morton_to_linear, PixelFormat format>
+static void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth,
+ u8* buffer, std::size_t buffer_size, VAddr addr) {
+ constexpr u32 bytes_per_pixel = GetBytesPerPixel(format);
+
+ // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
+ // pixel values.
+ const u32 tile_size_x{GetDefaultBlockWidth(format)};
+ const u32 tile_size_y{GetDefaultBlockHeight(format)};
+
+ if constexpr (morton_to_linear) {
+ Tegra::Texture::UnswizzleTexture(buffer, addr, tile_size_x, tile_size_y, bytes_per_pixel,
+ stride, height, depth, block_height, block_depth);
+ } else {
+ Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x,
+ (height + tile_size_y - 1) / tile_size_y, depth,
+ bytes_per_pixel, bytes_per_pixel, Memory::GetPointer(addr),
+ buffer, false, block_height, block_depth);
+ }
+}
+
+static constexpr ConversionArray morton_to_linear_fns = {
+ // clang-format off
+ MortonCopy<true, PixelFormat::ABGR8U>,
+ MortonCopy<true, PixelFormat::ABGR8S>,
+ MortonCopy<true, PixelFormat::ABGR8UI>,
+ MortonCopy<true, PixelFormat::B5G6R5U>,
+ MortonCopy<true, PixelFormat::A2B10G10R10U>,
+ MortonCopy<true, PixelFormat::A1B5G5R5U>,
+ MortonCopy<true, PixelFormat::R8U>,
+ MortonCopy<true, PixelFormat::R8UI>,
+ MortonCopy<true, PixelFormat::RGBA16F>,
+ MortonCopy<true, PixelFormat::RGBA16U>,
+ MortonCopy<true, PixelFormat::RGBA16UI>,
+ MortonCopy<true, PixelFormat::R11FG11FB10F>,
+ MortonCopy<true, PixelFormat::RGBA32UI>,
+ MortonCopy<true, PixelFormat::DXT1>,
+ MortonCopy<true, PixelFormat::DXT23>,
+ MortonCopy<true, PixelFormat::DXT45>,
+ MortonCopy<true, PixelFormat::DXN1>,
+ MortonCopy<true, PixelFormat::DXN2UNORM>,
+ MortonCopy<true, PixelFormat::DXN2SNORM>,
+ MortonCopy<true, PixelFormat::BC7U>,
+ MortonCopy<true, PixelFormat::BC6H_UF16>,
+ MortonCopy<true, PixelFormat::BC6H_SF16>,
+ MortonCopy<true, PixelFormat::ASTC_2D_4X4>,
+ MortonCopy<true, PixelFormat::G8R8U>,
+ MortonCopy<true, PixelFormat::G8R8S>,
+ MortonCopy<true, PixelFormat::BGRA8>,
+ MortonCopy<true, PixelFormat::RGBA32F>,
+ MortonCopy<true, PixelFormat::RG32F>,
+ MortonCopy<true, PixelFormat::R32F>,
+ MortonCopy<true, PixelFormat::R16F>,
+ MortonCopy<true, PixelFormat::R16U>,
+ MortonCopy<true, PixelFormat::R16S>,
+ MortonCopy<true, PixelFormat::R16UI>,
+ MortonCopy<true, PixelFormat::R16I>,
+ MortonCopy<true, PixelFormat::RG16>,
+ MortonCopy<true, PixelFormat::RG16F>,
+ MortonCopy<true, PixelFormat::RG16UI>,
+ MortonCopy<true, PixelFormat::RG16I>,
+ MortonCopy<true, PixelFormat::RG16S>,
+ MortonCopy<true, PixelFormat::RGB32F>,
+ MortonCopy<true, PixelFormat::RGBA8_SRGB>,
+ MortonCopy<true, PixelFormat::RG8U>,
+ MortonCopy<true, PixelFormat::RG8S>,
+ MortonCopy<true, PixelFormat::RG32UI>,
+ MortonCopy<true, PixelFormat::R32UI>,
+ MortonCopy<true, PixelFormat::ASTC_2D_8X8>,
+ MortonCopy<true, PixelFormat::ASTC_2D_8X5>,
+ MortonCopy<true, PixelFormat::ASTC_2D_5X4>,
+ MortonCopy<true, PixelFormat::BGRA8_SRGB>,
+ MortonCopy<true, PixelFormat::DXT1_SRGB>,
+ MortonCopy<true, PixelFormat::DXT23_SRGB>,
+ MortonCopy<true, PixelFormat::DXT45_SRGB>,
+ MortonCopy<true, PixelFormat::BC7U_SRGB>,
+ MortonCopy<true, PixelFormat::ASTC_2D_4X4_SRGB>,
+ MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>,
+ MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>,
+ MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>,
+ MortonCopy<true, PixelFormat::ASTC_2D_5X5>,
+ MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>,
+ MortonCopy<true, PixelFormat::ASTC_2D_10X8>,
+ MortonCopy<true, PixelFormat::ASTC_2D_10X8_SRGB>,
+ MortonCopy<true, PixelFormat::Z32F>,
+ MortonCopy<true, PixelFormat::Z16>,
+ MortonCopy<true, PixelFormat::Z24S8>,
+ MortonCopy<true, PixelFormat::S8Z24>,
+ MortonCopy<true, PixelFormat::Z32FS8>,
+ // clang-format on
+};
+
+static constexpr ConversionArray linear_to_morton_fns = {
+ // clang-format off
+ MortonCopy<false, PixelFormat::ABGR8U>,
+ MortonCopy<false, PixelFormat::ABGR8S>,
+ MortonCopy<false, PixelFormat::ABGR8UI>,
+ MortonCopy<false, PixelFormat::B5G6R5U>,
+ MortonCopy<false, PixelFormat::A2B10G10R10U>,
+ MortonCopy<false, PixelFormat::A1B5G5R5U>,
+ MortonCopy<false, PixelFormat::R8U>,
+ MortonCopy<false, PixelFormat::R8UI>,
+ MortonCopy<false, PixelFormat::RGBA16F>,
+ MortonCopy<false, PixelFormat::RGBA16U>,
+ MortonCopy<false, PixelFormat::RGBA16UI>,
+ MortonCopy<false, PixelFormat::R11FG11FB10F>,
+ MortonCopy<false, PixelFormat::RGBA32UI>,
+ MortonCopy<false, PixelFormat::DXT1>,
+ MortonCopy<false, PixelFormat::DXT23>,
+ MortonCopy<false, PixelFormat::DXT45>,
+ MortonCopy<false, PixelFormat::DXN1>,
+ MortonCopy<false, PixelFormat::DXN2UNORM>,
+ MortonCopy<false, PixelFormat::DXN2SNORM>,
+ MortonCopy<false, PixelFormat::BC7U>,
+ MortonCopy<false, PixelFormat::BC6H_UF16>,
+ MortonCopy<false, PixelFormat::BC6H_SF16>,
+ // TODO(Subv): Swizzling ASTC formats are not supported
+ nullptr,
+ MortonCopy<false, PixelFormat::G8R8U>,
+ MortonCopy<false, PixelFormat::G8R8S>,
+ MortonCopy<false, PixelFormat::BGRA8>,
+ MortonCopy<false, PixelFormat::RGBA32F>,
+ MortonCopy<false, PixelFormat::RG32F>,
+ MortonCopy<false, PixelFormat::R32F>,
+ MortonCopy<false, PixelFormat::R16F>,
+ MortonCopy<false, PixelFormat::R16U>,
+ MortonCopy<false, PixelFormat::R16S>,
+ MortonCopy<false, PixelFormat::R16UI>,
+ MortonCopy<false, PixelFormat::R16I>,
+ MortonCopy<false, PixelFormat::RG16>,
+ MortonCopy<false, PixelFormat::RG16F>,
+ MortonCopy<false, PixelFormat::RG16UI>,
+ MortonCopy<false, PixelFormat::RG16I>,
+ MortonCopy<false, PixelFormat::RG16S>,
+ MortonCopy<false, PixelFormat::RGB32F>,
+ MortonCopy<false, PixelFormat::RGBA8_SRGB>,
+ MortonCopy<false, PixelFormat::RG8U>,
+ MortonCopy<false, PixelFormat::RG8S>,
+ MortonCopy<false, PixelFormat::RG32UI>,
+ MortonCopy<false, PixelFormat::R32UI>,
+ nullptr,
+ nullptr,
+ nullptr,
+ MortonCopy<false, PixelFormat::BGRA8_SRGB>,
+ MortonCopy<false, PixelFormat::DXT1_SRGB>,
+ MortonCopy<false, PixelFormat::DXT23_SRGB>,
+ MortonCopy<false, PixelFormat::DXT45_SRGB>,
+ MortonCopy<false, PixelFormat::BC7U_SRGB>,
+ nullptr,
+ nullptr,
+ nullptr,
+ nullptr,
+ nullptr,
+ nullptr,
+ nullptr,
+ nullptr,
+ MortonCopy<false, PixelFormat::Z32F>,
+ MortonCopy<false, PixelFormat::Z16>,
+ MortonCopy<false, PixelFormat::Z24S8>,
+ MortonCopy<false, PixelFormat::S8Z24>,
+ MortonCopy<false, PixelFormat::Z32FS8>,
+ // clang-format on
+};
+
+constexpr MortonCopyFn GetSwizzleFunction(MortonSwizzleMode mode, Surface::PixelFormat format) {
+ switch (mode) {
+ case MortonSwizzleMode::MortonToLinear:
+ return morton_to_linear_fns[static_cast<std::size_t>(format)];
+ case MortonSwizzleMode::LinearToMorton:
+ return linear_to_morton_fns[static_cast<std::size_t>(format)];
+ }
+}
+
+/// 8x8 Z-Order coordinate from 2D coordinates
+static u32 MortonInterleave(u32 x, u32 y) {
+ static const u32 xlut[] = {0x00, 0x01, 0x04, 0x05, 0x10, 0x11, 0x14, 0x15};
+ static const u32 ylut[] = {0x00, 0x02, 0x08, 0x0a, 0x20, 0x22, 0x28, 0x2a};
+ return xlut[x % 8] + ylut[y % 8];
+}
+
+/// Calculates the offset of the position of the pixel in Morton order
+static u32 GetMortonOffset(u32 x, u32 y, u32 bytes_per_pixel) {
+ // Images are split into 8x8 tiles. Each tile is composed of four 4x4 subtiles each
+ // of which is composed of four 2x2 subtiles each of which is composed of four texels.
+ // Each structure is embedded into the next-bigger one in a diagonal pattern, e.g.
+ // texels are laid out in a 2x2 subtile like this:
+ // 2 3
+ // 0 1
+ //
+ // The full 8x8 tile has the texels arranged like this:
+ //
+ // 42 43 46 47 58 59 62 63
+ // 40 41 44 45 56 57 60 61
+ // 34 35 38 39 50 51 54 55
+ // 32 33 36 37 48 49 52 53
+ // 10 11 14 15 26 27 30 31
+ // 08 09 12 13 24 25 28 29
+ // 02 03 06 07 18 19 22 23
+ // 00 01 04 05 16 17 20 21
+ //
+ // This pattern is what's called Z-order curve, or Morton order.
+
+ const unsigned int block_height = 8;
+ const unsigned int coarse_x = x & ~7;
+
+ u32 i = MortonInterleave(x, y);
+
+ const unsigned int offset = coarse_x * block_height;
+
+ return (i + offset) * bytes_per_pixel;
+}
+
+static u32 MortonInterleave128(u32 x, u32 y) {
+ // 128x128 Z-Order coordinate from 2D coordinates
+ static constexpr u32 xlut[] = {
+ 0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042,
+ 0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809,
+ 0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000,
+ 0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043,
+ 0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a,
+ 0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001,
+ 0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048,
+ 0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b,
+ 0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002,
+ 0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049,
+ 0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840,
+ 0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003,
+ 0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a,
+ 0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841,
+ 0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008,
+ 0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b,
+ 0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842,
+ 0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009,
+ 0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800,
+ 0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843,
+ 0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a,
+ 0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801,
+ 0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848,
+ 0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b,
+ 0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802,
+ 0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849,
+ 0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040,
+ 0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803,
+ 0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a,
+ 0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041,
+ 0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808,
+ 0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b,
+ 0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042,
+ 0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809,
+ 0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b,
+ };
+ static constexpr u32 ylut[] = {
+ 0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090,
+ 0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124,
+ 0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200,
+ 0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294,
+ 0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330,
+ 0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404,
+ 0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0,
+ 0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534,
+ 0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610,
+ 0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4,
+ 0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780,
+ 0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014,
+ 0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0,
+ 0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184,
+ 0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220,
+ 0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4,
+ 0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390,
+ 0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424,
+ 0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500,
+ 0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594,
+ 0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630,
+ 0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704,
+ 0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0,
+ 0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034,
+ 0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110,
+ 0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4,
+ 0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280,
+ 0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314,
+ 0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0,
+ 0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484,
+ 0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520,
+ 0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4,
+ 0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690,
+ 0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724,
+ 0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4,
+ };
+ return xlut[x % 128] + ylut[y % 128];
+}
+
+static u32 GetMortonOffset128(u32 x, u32 y, u32 bytes_per_pixel) {
+ // Calculates the offset of the position of the pixel in Morton order
+ // Framebuffer images are split into 128x128 tiles.
+
+ constexpr u32 block_height = 128;
+ const u32 coarse_x = x & ~127;
+
+ const u32 i = MortonInterleave128(x, y);
+
+ const u32 offset = coarse_x * block_height;
+
+ return (i + offset) * bytes_per_pixel;
+}
+
+void MortonSwizzle(MortonSwizzleMode mode, Surface::PixelFormat format, u32 stride,
+ u32 block_height, u32 height, u32 block_depth, u32 depth, u8* buffer,
+ std::size_t buffer_size, VAddr addr) {
+
+ GetSwizzleFunction(mode, format)(stride, block_height, height, block_depth, depth, buffer,
+ buffer_size, addr);
+}
+
+void MortonCopyPixels128(u32 width, u32 height, u32 bytes_per_pixel, u32 linear_bytes_per_pixel,
+ u8* morton_data, u8* linear_data, bool morton_to_linear) {
+ u8* data_ptrs[2];
+ for (u32 y = 0; y < height; ++y) {
+ for (u32 x = 0; x < width; ++x) {
+ const u32 coarse_y = y & ~127;
+ const u32 morton_offset =
+ GetMortonOffset128(x, y, bytes_per_pixel) + coarse_y * width * bytes_per_pixel;
+ const u32 linear_pixel_index = (x + y * width) * linear_bytes_per_pixel;
+
+ data_ptrs[morton_to_linear ? 1 : 0] = morton_data + morton_offset;
+ data_ptrs[morton_to_linear ? 0 : 1] = &linear_data[linear_pixel_index];
+
+ std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
+ }
+ }
+}
+
+} // namespace VideoCore \ No newline at end of file
diff --git a/src/video_core/morton.h b/src/video_core/morton.h
new file mode 100644
index 000000000..b9b9eca86
--- /dev/null
+++ b/src/video_core/morton.h
@@ -0,0 +1,21 @@
+// Copyright 2018 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+#include "video_core/surface.h"
+
+namespace VideoCore {
+
+enum class MortonSwizzleMode { MortonToLinear, LinearToMorton };
+
+void MortonSwizzle(MortonSwizzleMode mode, VideoCore::Surface::PixelFormat format, u32 stride,
+ u32 block_height, u32 height, u32 block_depth, u32 depth, u8* buffer,
+ std::size_t buffer_size, VAddr addr);
+
+void MortonCopyPixels128(u32 width, u32 height, u32 bytes_per_pixel, u32 linear_bytes_per_pixel,
+ u8* morton_data, u8* linear_data, bool morton_to_linear);
+
+} // namespace VideoCore \ No newline at end of file
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index ae6aaee4c..630a58e49 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -98,14 +98,9 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, ScreenInfo
has_ARB_direct_state_access = true;
} else if (extension == "GL_ARB_multi_bind") {
has_ARB_multi_bind = true;
- } else if (extension == "GL_ARB_separate_shader_objects") {
- has_ARB_separate_shader_objects = true;
- } else if (extension == "GL_ARB_vertex_attrib_binding") {
- has_ARB_vertex_attrib_binding = true;
}
}
- ASSERT_MSG(has_ARB_separate_shader_objects, "has_ARB_separate_shader_objects is unsupported");
OpenGLState::ApplyDefaultState();
// Create render framebuffer
@@ -542,6 +537,30 @@ void RasterizerOpenGL::Clear() {
ASSERT_MSG(regs.zeta_enable != 0, "Tried to clear stencil but buffer is not enabled!");
use_stencil = true;
clear_state.stencil.test_enabled = true;
+ if (regs.clear_flags.stencil) {
+ // Stencil affects the clear so fill it with the used masks
+ clear_state.stencil.front.test_func = GL_ALWAYS;
+ clear_state.stencil.front.test_mask = regs.stencil_front_func_mask;
+ clear_state.stencil.front.action_stencil_fail = GL_KEEP;
+ clear_state.stencil.front.action_depth_fail = GL_KEEP;
+ clear_state.stencil.front.action_depth_pass = GL_KEEP;
+ clear_state.stencil.front.write_mask = regs.stencil_front_mask;
+ if (regs.stencil_two_side_enable) {
+ clear_state.stencil.back.test_func = GL_ALWAYS;
+ clear_state.stencil.back.test_mask = regs.stencil_back_func_mask;
+ clear_state.stencil.back.action_stencil_fail = GL_KEEP;
+ clear_state.stencil.back.action_depth_fail = GL_KEEP;
+ clear_state.stencil.back.action_depth_pass = GL_KEEP;
+ clear_state.stencil.back.write_mask = regs.stencil_back_mask;
+ } else {
+ clear_state.stencil.back.test_func = GL_ALWAYS;
+ clear_state.stencil.back.test_mask = 0xFFFFFFFF;
+ clear_state.stencil.back.write_mask = 0xFFFFFFFF;
+ clear_state.stencil.back.action_stencil_fail = GL_KEEP;
+ clear_state.stencil.back.action_depth_fail = GL_KEEP;
+ clear_state.stencil.back.action_depth_pass = GL_KEEP;
+ }
+ }
}
if (!use_color && !use_depth && !use_stencil) {
@@ -553,6 +572,14 @@ void RasterizerOpenGL::Clear() {
ConfigureFramebuffers(clear_state, use_color, use_depth || use_stencil, false,
regs.clear_buffers.RT.Value());
+ if (regs.clear_flags.scissor) {
+ SyncScissorTest(clear_state);
+ }
+
+ if (regs.clear_flags.viewport) {
+ clear_state.EmulateViewportWithScissor();
+ }
+
clear_state.Apply();
if (use_color) {
@@ -588,7 +615,7 @@ void RasterizerOpenGL::DrawArrays() {
SyncLogicOpState();
SyncCullMode();
SyncPrimitiveRestart();
- SyncScissorTest();
+ SyncScissorTest(state);
// Alpha Testing is synced on shaders.
SyncTransformFeedback();
SyncPointState();
@@ -815,7 +842,7 @@ void RasterizerOpenGL::SamplerInfo::SyncWithConfig(const Tegra::Texture::TSCEntr
}
const u32 bias = config.mip_lod_bias.Value();
// Sign extend the 13-bit value.
- const u32 mask = 1U << (13 - 1);
+ constexpr u32 mask = 1U << (13 - 1);
const float bias_lod = static_cast<s32>((bias ^ mask) - mask) / 256.f;
if (lod_bias != bias_lod) {
lod_bias = bias_lod;
@@ -947,8 +974,8 @@ void RasterizerOpenGL::SyncViewport(OpenGLState& current_state) {
auto& viewport = current_state.viewports[i];
viewport.x = viewport_rect.left;
viewport.y = viewport_rect.bottom;
- viewport.width = static_cast<GLfloat>(viewport_rect.GetWidth());
- viewport.height = static_cast<GLfloat>(viewport_rect.GetHeight());
+ viewport.width = viewport_rect.GetWidth();
+ viewport.height = viewport_rect.GetHeight();
viewport.depth_range_far = regs.viewports[i].depth_range_far;
viewport.depth_range_near = regs.viewports[i].depth_range_near;
}
@@ -1120,11 +1147,11 @@ void RasterizerOpenGL::SyncLogicOpState() {
state.logic_op.operation = MaxwellToGL::LogicOp(regs.logic_op.operation);
}
-void RasterizerOpenGL::SyncScissorTest() {
+void RasterizerOpenGL::SyncScissorTest(OpenGLState& current_state) {
const auto& regs = Core::System::GetInstance().GPU().Maxwell3D().regs;
for (std::size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumViewports; i++) {
const auto& src = regs.scissor_test[i];
- auto& dst = state.viewports[i].scissor;
+ auto& dst = current_state.viewports[i].scissor;
dst.enabled = (src.enable != 0);
if (dst.enabled == 0) {
return;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h
index 6e78ab4cd..f4354289c 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer.h
@@ -91,19 +91,20 @@ private:
void SyncWithConfig(const Tegra::Texture::TSCEntry& info);
private:
- Tegra::Texture::TextureFilter mag_filter;
- Tegra::Texture::TextureFilter min_filter;
- Tegra::Texture::TextureMipmapFilter mip_filter;
- Tegra::Texture::WrapMode wrap_u;
- Tegra::Texture::WrapMode wrap_v;
- Tegra::Texture::WrapMode wrap_p;
- bool uses_depth_compare;
- Tegra::Texture::DepthCompareFunc depth_compare_func;
- GLvec4 border_color;
- float min_lod;
- float max_lod;
- float lod_bias;
- float max_anisotropic;
+ Tegra::Texture::TextureFilter mag_filter = Tegra::Texture::TextureFilter::Nearest;
+ Tegra::Texture::TextureFilter min_filter = Tegra::Texture::TextureFilter::Nearest;
+ Tegra::Texture::TextureMipmapFilter mip_filter = Tegra::Texture::TextureMipmapFilter::None;
+ Tegra::Texture::WrapMode wrap_u = Tegra::Texture::WrapMode::ClampToEdge;
+ Tegra::Texture::WrapMode wrap_v = Tegra::Texture::WrapMode::ClampToEdge;
+ Tegra::Texture::WrapMode wrap_p = Tegra::Texture::WrapMode::ClampToEdge;
+ bool uses_depth_compare = false;
+ Tegra::Texture::DepthCompareFunc depth_compare_func =
+ Tegra::Texture::DepthCompareFunc::Always;
+ GLvec4 border_color = {};
+ float min_lod = 0.0f;
+ float max_lod = 16.0f;
+ float lod_bias = 0.0f;
+ float max_anisotropic = 1.0f;
};
/**
@@ -171,7 +172,7 @@ private:
void SyncMultiSampleState();
/// Syncs the scissor test state to match the guest state
- void SyncScissorTest();
+ void SyncScissorTest(OpenGLState& current_state);
/// Syncs the transform feedback state to match the guest state
void SyncTransformFeedback();
@@ -187,8 +188,6 @@ private:
bool has_ARB_direct_state_access = false;
bool has_ARB_multi_bind = false;
- bool has_ARB_separate_shader_objects = false;
- bool has_ARB_vertex_attrib_binding = false;
OpenGLState state;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
index b994e89dd..d458f77e4 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
@@ -15,6 +15,7 @@
#include "core/memory.h"
#include "core/settings.h"
#include "video_core/engines/maxwell_3d.h"
+#include "video_core/morton.h"
#include "video_core/renderer_opengl/gl_rasterizer.h"
#include "video_core/renderer_opengl/gl_rasterizer_cache.h"
#include "video_core/renderer_opengl/gl_state.h"
@@ -22,10 +23,11 @@
#include "video_core/surface.h"
#include "video_core/textures/astc.h"
#include "video_core/textures/decoders.h"
-#include "video_core/utils.h"
namespace OpenGL {
+using VideoCore::MortonSwizzle;
+using VideoCore::MortonSwizzleMode;
using VideoCore::Surface::ComponentTypeFromDepthFormat;
using VideoCore::Surface::ComponentTypeFromRenderTarget;
using VideoCore::Surface::ComponentTypeFromTexture;
@@ -265,11 +267,11 @@ static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex
{GL_COMPRESSED_RG_RGTC2, GL_RG, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
true}, // DXN2UNORM
{GL_COMPRESSED_SIGNED_RG_RGTC2, GL_RG, GL_INT, ComponentType::SNorm, true}, // DXN2SNORM
- {GL_COMPRESSED_RGBA_BPTC_UNORM_ARB, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
+ {GL_COMPRESSED_RGBA_BPTC_UNORM, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
true}, // BC7U
- {GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT_ARB, GL_RGB, GL_UNSIGNED_INT_8_8_8_8,
- ComponentType::Float, true}, // BC6H_UF16
- {GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT_ARB, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::Float,
+ {GL_COMPRESSED_RGB_BPTC_UNSIGNED_FLOAT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::Float,
+ true}, // BC6H_UF16
+ {GL_COMPRESSED_RGB_BPTC_SIGNED_FLOAT, GL_RGB, GL_UNSIGNED_INT_8_8_8_8, ComponentType::Float,
true}, // BC6H_SF16
{GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_4X4
{GL_RG8, GL_RG, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // G8R8U
@@ -306,8 +308,8 @@ static constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex
true}, // DXT23_SRGB
{GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
true}, // DXT45_SRGB
- {GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM_ARB, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8,
- ComponentType::UNorm, true}, // BC7U_SRGB
+ {GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, ComponentType::UNorm,
+ true}, // BC7U_SRGB
{GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_4X4_SRGB
{GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X8_SRGB
{GL_SRGB8_ALPHA8, GL_RGBA, GL_UNSIGNED_BYTE, ComponentType::UNorm, false}, // ASTC_2D_8X5_SRGB
@@ -346,7 +348,7 @@ static GLenum SurfaceTargetToGL(SurfaceTarget target) {
case SurfaceTarget::TextureCubemap:
return GL_TEXTURE_CUBE_MAP;
case SurfaceTarget::TextureCubeArray:
- return GL_TEXTURE_CUBE_MAP_ARRAY_ARB;
+ return GL_TEXTURE_CUBE_MAP_ARRAY;
}
LOG_CRITICAL(Render_OpenGL, "Unimplemented texture target={}", static_cast<u32>(target));
UNREACHABLE();
@@ -370,174 +372,7 @@ MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
return {0, actual_height, MipWidth(mip_level), 0};
}
-template <bool morton_to_gl, PixelFormat format>
-void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth, u8* gl_buffer,
- std::size_t gl_buffer_size, VAddr addr) {
- constexpr u32 bytes_per_pixel = GetBytesPerPixel(format);
-
- // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
- // pixel values.
- const u32 tile_size_x{GetDefaultBlockWidth(format)};
- const u32 tile_size_y{GetDefaultBlockHeight(format)};
-
- if (morton_to_gl) {
- Tegra::Texture::UnswizzleTexture(gl_buffer, addr, tile_size_x, tile_size_y, bytes_per_pixel,
- stride, height, depth, block_height, block_depth);
- } else {
- Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x,
- (height + tile_size_y - 1) / tile_size_y, depth,
- bytes_per_pixel, bytes_per_pixel, Memory::GetPointer(addr),
- gl_buffer, false, block_height, block_depth);
- }
-}
-
-using GLConversionArray = std::array<void (*)(u32, u32, u32, u32, u32, u8*, std::size_t, VAddr),
- VideoCore::Surface::MaxPixelFormat>;
-
-static constexpr GLConversionArray morton_to_gl_fns = {
- // clang-format off
- MortonCopy<true, PixelFormat::ABGR8U>,
- MortonCopy<true, PixelFormat::ABGR8S>,
- MortonCopy<true, PixelFormat::ABGR8UI>,
- MortonCopy<true, PixelFormat::B5G6R5U>,
- MortonCopy<true, PixelFormat::A2B10G10R10U>,
- MortonCopy<true, PixelFormat::A1B5G5R5U>,
- MortonCopy<true, PixelFormat::R8U>,
- MortonCopy<true, PixelFormat::R8UI>,
- MortonCopy<true, PixelFormat::RGBA16F>,
- MortonCopy<true, PixelFormat::RGBA16U>,
- MortonCopy<true, PixelFormat::RGBA16UI>,
- MortonCopy<true, PixelFormat::R11FG11FB10F>,
- MortonCopy<true, PixelFormat::RGBA32UI>,
- MortonCopy<true, PixelFormat::DXT1>,
- MortonCopy<true, PixelFormat::DXT23>,
- MortonCopy<true, PixelFormat::DXT45>,
- MortonCopy<true, PixelFormat::DXN1>,
- MortonCopy<true, PixelFormat::DXN2UNORM>,
- MortonCopy<true, PixelFormat::DXN2SNORM>,
- MortonCopy<true, PixelFormat::BC7U>,
- MortonCopy<true, PixelFormat::BC6H_UF16>,
- MortonCopy<true, PixelFormat::BC6H_SF16>,
- MortonCopy<true, PixelFormat::ASTC_2D_4X4>,
- MortonCopy<true, PixelFormat::G8R8U>,
- MortonCopy<true, PixelFormat::G8R8S>,
- MortonCopy<true, PixelFormat::BGRA8>,
- MortonCopy<true, PixelFormat::RGBA32F>,
- MortonCopy<true, PixelFormat::RG32F>,
- MortonCopy<true, PixelFormat::R32F>,
- MortonCopy<true, PixelFormat::R16F>,
- MortonCopy<true, PixelFormat::R16U>,
- MortonCopy<true, PixelFormat::R16S>,
- MortonCopy<true, PixelFormat::R16UI>,
- MortonCopy<true, PixelFormat::R16I>,
- MortonCopy<true, PixelFormat::RG16>,
- MortonCopy<true, PixelFormat::RG16F>,
- MortonCopy<true, PixelFormat::RG16UI>,
- MortonCopy<true, PixelFormat::RG16I>,
- MortonCopy<true, PixelFormat::RG16S>,
- MortonCopy<true, PixelFormat::RGB32F>,
- MortonCopy<true, PixelFormat::RGBA8_SRGB>,
- MortonCopy<true, PixelFormat::RG8U>,
- MortonCopy<true, PixelFormat::RG8S>,
- MortonCopy<true, PixelFormat::RG32UI>,
- MortonCopy<true, PixelFormat::R32UI>,
- MortonCopy<true, PixelFormat::ASTC_2D_8X8>,
- MortonCopy<true, PixelFormat::ASTC_2D_8X5>,
- MortonCopy<true, PixelFormat::ASTC_2D_5X4>,
- MortonCopy<true, PixelFormat::BGRA8_SRGB>,
- MortonCopy<true, PixelFormat::DXT1_SRGB>,
- MortonCopy<true, PixelFormat::DXT23_SRGB>,
- MortonCopy<true, PixelFormat::DXT45_SRGB>,
- MortonCopy<true, PixelFormat::BC7U_SRGB>,
- MortonCopy<true, PixelFormat::ASTC_2D_4X4_SRGB>,
- MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>,
- MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>,
- MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>,
- MortonCopy<true, PixelFormat::ASTC_2D_5X5>,
- MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>,
- MortonCopy<true, PixelFormat::ASTC_2D_10X8>,
- MortonCopy<true, PixelFormat::ASTC_2D_10X8_SRGB>,
- MortonCopy<true, PixelFormat::Z32F>,
- MortonCopy<true, PixelFormat::Z16>,
- MortonCopy<true, PixelFormat::Z24S8>,
- MortonCopy<true, PixelFormat::S8Z24>,
- MortonCopy<true, PixelFormat::Z32FS8>,
- // clang-format on
-};
-
-static constexpr GLConversionArray gl_to_morton_fns = {
- // clang-format off
- MortonCopy<false, PixelFormat::ABGR8U>,
- MortonCopy<false, PixelFormat::ABGR8S>,
- MortonCopy<false, PixelFormat::ABGR8UI>,
- MortonCopy<false, PixelFormat::B5G6R5U>,
- MortonCopy<false, PixelFormat::A2B10G10R10U>,
- MortonCopy<false, PixelFormat::A1B5G5R5U>,
- MortonCopy<false, PixelFormat::R8U>,
- MortonCopy<false, PixelFormat::R8UI>,
- MortonCopy<false, PixelFormat::RGBA16F>,
- MortonCopy<false, PixelFormat::RGBA16U>,
- MortonCopy<false, PixelFormat::RGBA16UI>,
- MortonCopy<false, PixelFormat::R11FG11FB10F>,
- MortonCopy<false, PixelFormat::RGBA32UI>,
- MortonCopy<false, PixelFormat::DXT1>,
- MortonCopy<false, PixelFormat::DXT23>,
- MortonCopy<false, PixelFormat::DXT45>,
- MortonCopy<false, PixelFormat::DXN1>,
- MortonCopy<false, PixelFormat::DXN2UNORM>,
- MortonCopy<false, PixelFormat::DXN2SNORM>,
- MortonCopy<false, PixelFormat::BC7U>,
- MortonCopy<false, PixelFormat::BC6H_UF16>,
- MortonCopy<false, PixelFormat::BC6H_SF16>,
- // TODO(Subv): Swizzling ASTC formats are not supported
- nullptr,
- MortonCopy<false, PixelFormat::G8R8U>,
- MortonCopy<false, PixelFormat::G8R8S>,
- MortonCopy<false, PixelFormat::BGRA8>,
- MortonCopy<false, PixelFormat::RGBA32F>,
- MortonCopy<false, PixelFormat::RG32F>,
- MortonCopy<false, PixelFormat::R32F>,
- MortonCopy<false, PixelFormat::R16F>,
- MortonCopy<false, PixelFormat::R16U>,
- MortonCopy<false, PixelFormat::R16S>,
- MortonCopy<false, PixelFormat::R16UI>,
- MortonCopy<false, PixelFormat::R16I>,
- MortonCopy<false, PixelFormat::RG16>,
- MortonCopy<false, PixelFormat::RG16F>,
- MortonCopy<false, PixelFormat::RG16UI>,
- MortonCopy<false, PixelFormat::RG16I>,
- MortonCopy<false, PixelFormat::RG16S>,
- MortonCopy<false, PixelFormat::RGB32F>,
- MortonCopy<false, PixelFormat::RGBA8_SRGB>,
- MortonCopy<false, PixelFormat::RG8U>,
- MortonCopy<false, PixelFormat::RG8S>,
- MortonCopy<false, PixelFormat::RG32UI>,
- MortonCopy<false, PixelFormat::R32UI>,
- nullptr,
- nullptr,
- nullptr,
- MortonCopy<false, PixelFormat::BGRA8_SRGB>,
- MortonCopy<false, PixelFormat::DXT1_SRGB>,
- MortonCopy<false, PixelFormat::DXT23_SRGB>,
- MortonCopy<false, PixelFormat::DXT45_SRGB>,
- MortonCopy<false, PixelFormat::BC7U_SRGB>,
- nullptr,
- nullptr,
- nullptr,
- nullptr,
- nullptr,
- nullptr,
- nullptr,
- nullptr,
- MortonCopy<false, PixelFormat::Z32F>,
- MortonCopy<false, PixelFormat::Z16>,
- MortonCopy<false, PixelFormat::Z24S8>,
- MortonCopy<false, PixelFormat::S8Z24>,
- MortonCopy<false, PixelFormat::Z32FS8>,
- // clang-format on
-};
-
-void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params,
+void SwizzleFunc(const MortonSwizzleMode& mode, const SurfaceParams& params,
std::vector<u8>& gl_buffer, u32 mip_level) {
u32 depth = params.MipDepth(mip_level);
if (params.target == SurfaceTarget::Texture2D) {
@@ -550,19 +385,19 @@ void SwizzleFunc(const GLConversionArray& functions, const SurfaceParams& params
const u64 layer_size = params.LayerMemorySize();
const u64 gl_size = params.LayerSizeGL(mip_level);
for (u32 i = 0; i < params.depth; i++) {
- functions[static_cast<std::size_t>(params.pixel_format)](
- params.MipWidth(mip_level), params.MipBlockHeight(mip_level),
- params.MipHeight(mip_level), params.MipBlockDepth(mip_level), 1,
- gl_buffer.data() + offset_gl, gl_size, params.addr + offset);
+ MortonSwizzle(mode, params.pixel_format, params.MipWidth(mip_level),
+ params.MipBlockHeight(mip_level), params.MipHeight(mip_level),
+ params.MipBlockDepth(mip_level), 1, gl_buffer.data() + offset_gl, gl_size,
+ params.addr + offset);
offset += layer_size;
offset_gl += gl_size;
}
} else {
const u64 offset = params.GetMipmapLevelOffset(mip_level);
- functions[static_cast<std::size_t>(params.pixel_format)](
- params.MipWidth(mip_level), params.MipBlockHeight(mip_level),
- params.MipHeight(mip_level), params.MipBlockDepth(mip_level), depth, gl_buffer.data(),
- gl_buffer.size(), params.addr + offset);
+ MortonSwizzle(mode, params.pixel_format, params.MipWidth(mip_level),
+ params.MipBlockHeight(mip_level), params.MipHeight(mip_level),
+ params.MipBlockDepth(mip_level), depth, gl_buffer.data(), gl_buffer.size(),
+ params.addr + offset);
}
}
@@ -726,7 +561,7 @@ static void CopySurface(const Surface& src_surface, const Surface& dst_surface,
const std::size_t buffer_size = std::max(src_params.size_in_bytes, dst_params.size_in_bytes);
glBindBuffer(GL_PIXEL_PACK_BUFFER, copy_pbo_handle);
- glBufferData(GL_PIXEL_PACK_BUFFER, buffer_size, nullptr, GL_STREAM_DRAW_ARB);
+ glBufferData(GL_PIXEL_PACK_BUFFER, buffer_size, nullptr, GL_STREAM_DRAW);
if (source_format.compressed) {
glGetCompressedTextureImage(src_surface->Texture().handle, src_attachment,
static_cast<GLsizei>(src_params.size_in_bytes), nullptr);
@@ -996,7 +831,7 @@ void CachedSurface::LoadGLBuffer() {
ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
params.block_width, static_cast<u32>(params.target));
for (u32 i = 0; i < params.max_mip_level; i++)
- SwizzleFunc(morton_to_gl_fns, params, gl_buffer[i], i);
+ SwizzleFunc(MortonSwizzleMode::MortonToLinear, params, gl_buffer[i], i);
} else {
const auto texture_src_data{Memory::GetPointer(params.addr)};
const auto texture_src_data_end{texture_src_data + params.size_in_bytes_gl};
@@ -1035,7 +870,7 @@ void CachedSurface::FlushGLBuffer() {
ASSERT_MSG(params.block_width == 1, "Block width is defined as {} on texture type {}",
params.block_width, static_cast<u32>(params.target));
- SwizzleFunc(gl_to_morton_fns, params, gl_buffer[0], 0);
+ SwizzleFunc(MortonSwizzleMode::LinearToMorton, params, gl_buffer[0], 0);
} else {
std::memcpy(Memory::GetPointer(GetAddr()), gl_buffer[0].data(), GetSizeInBytes());
}
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
index 8c43f0ead..9dbaf26d4 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
@@ -34,6 +34,17 @@ constexpr u32 PROGRAM_HEADER_SIZE = sizeof(Tegra::Shader::Header);
constexpr u32 MAX_GEOMETRY_BUFFERS = 6;
constexpr u32 MAX_ATTRIBUTES = 0x100; // Size in vec4s, this value is untested
+static const char* INTERNAL_FLAG_NAMES[] = {"zero_flag", "sign_flag", "carry_flag",
+ "overflow_flag"};
+
+enum class InternalFlag : u64 {
+ ZeroFlag = 0,
+ SignFlag = 1,
+ CarryFlag = 2,
+ OverflowFlag = 3,
+ Amount
+};
+
class DecompileFail : public std::runtime_error {
public:
using std::runtime_error::runtime_error;
@@ -267,14 +278,6 @@ private:
const std::string& suffix;
};
-enum class InternalFlag : u64 {
- ZeroFlag = 0,
- CarryFlag = 1,
- OverflowFlag = 2,
- NaNFlag = 3,
- Amount
-};
-
/**
* Used to manage shader registers that are emulated with GLSL. This class keeps track of the state
* of all registers (e.g. whether they are currently being used as Floats or Integers), and
@@ -381,7 +384,7 @@ public:
if (sets_cc) {
const std::string zero_condition = "( " + ConvertIntegerSize(value, size) + " == 0 )";
SetInternalFlag(InternalFlag::ZeroFlag, zero_condition);
- LOG_WARNING(HW_GPU, "Control Codes Imcomplete.");
+ LOG_WARNING(HW_GPU, "Condition codes implementation is incomplete.");
}
}
@@ -464,23 +467,25 @@ public:
shader.AddLine("lmem[" + index + "] = " + func + '(' + value + ");");
}
- std::string GetControlCode(const Tegra::Shader::ControlCode cc) const {
+ std::string GetConditionCode(const Tegra::Shader::ConditionCode cc) const {
switch (cc) {
- case Tegra::Shader::ControlCode::NEU:
+ case Tegra::Shader::ConditionCode::NEU:
return "!(" + GetInternalFlag(InternalFlag::ZeroFlag) + ')';
default:
- UNIMPLEMENTED_MSG("Unimplemented Control Code: {}", static_cast<u32>(cc));
+ UNIMPLEMENTED_MSG("Unimplemented condition code: {}", static_cast<u32>(cc));
return "false";
}
}
- std::string GetInternalFlag(const InternalFlag ii) const {
- const u32 code = static_cast<u32>(ii);
- return "internalFlag_" + std::to_string(code) + suffix;
+ std::string GetInternalFlag(const InternalFlag flag) const {
+ const auto index = static_cast<u32>(flag);
+ ASSERT(index < static_cast<u32>(InternalFlag::Amount));
+
+ return std::string(INTERNAL_FLAG_NAMES[index]) + '_' + suffix;
}
- void SetInternalFlag(const InternalFlag ii, const std::string& value) const {
- shader.AddLine(GetInternalFlag(ii) + " = " + value + ';');
+ void SetInternalFlag(const InternalFlag flag, const std::string& value) const {
+ shader.AddLine(GetInternalFlag(flag) + " = " + value + ';');
}
/**
@@ -495,27 +500,42 @@ public:
const Register& buf_reg) {
const std::string dest = GetOutputAttribute(attribute);
const std::string src = GetRegisterAsFloat(val_reg);
+ if (dest.empty())
+ return;
- if (!dest.empty()) {
- // Can happen with unknown/unimplemented output attributes, in which case we ignore the
- // instruction for now.
- if (stage == Maxwell3D::Regs::ShaderStage::Geometry) {
- // TODO(Rodrigo): nouveau sets some attributes after setting emitting a geometry
- // shader. These instructions use a dirty register as buffer index, to avoid some
- // drivers from complaining about out of boundary writes, guard them.
- const std::string buf_index{"((" + GetRegisterAsInteger(buf_reg) + ") % " +
- std::to_string(MAX_GEOMETRY_BUFFERS) + ')'};
- shader.AddLine("amem[" + buf_index + "][" +
- std::to_string(static_cast<u32>(attribute)) + ']' +
- GetSwizzle(elem) + " = " + src + ';');
- } else {
- if (attribute == Attribute::Index::PointSize) {
- fixed_pipeline_output_attributes_used.insert(attribute);
- shader.AddLine(dest + " = " + src + ';');
- } else {
- shader.AddLine(dest + GetSwizzle(elem) + " = " + src + ';');
- }
- }
+ // Can happen with unknown/unimplemented output attributes, in which case we ignore the
+ // instruction for now.
+ if (stage == Maxwell3D::Regs::ShaderStage::Geometry) {
+ // TODO(Rodrigo): nouveau sets some attributes after setting emitting a geometry
+ // shader. These instructions use a dirty register as buffer index, to avoid some
+ // drivers from complaining about out of boundary writes, guard them.
+ const std::string buf_index{"((" + GetRegisterAsInteger(buf_reg) + ") % " +
+ std::to_string(MAX_GEOMETRY_BUFFERS) + ')'};
+ shader.AddLine("amem[" + buf_index + "][" +
+ std::to_string(static_cast<u32>(attribute)) + ']' + GetSwizzle(elem) +
+ " = " + src + ';');
+ return;
+ }
+
+ switch (attribute) {
+ case Attribute::Index::ClipDistances0123:
+ case Attribute::Index::ClipDistances4567: {
+ const u64 index = attribute == Attribute::Index::ClipDistances4567 ? 4 : 0 + elem;
+ UNIMPLEMENTED_IF_MSG(
+ ((header.vtg.clip_distances >> index) & 1) == 0,
+ "Shader is setting gl_ClipDistance{} without enabling it in the header", index);
+
+ fixed_pipeline_output_attributes_used.insert(attribute);
+ shader.AddLine(dest + '[' + std::to_string(index) + "] = " + src + ';');
+ break;
+ }
+ case Attribute::Index::PointSize:
+ fixed_pipeline_output_attributes_used.insert(attribute);
+ shader.AddLine(dest + " = " + src + ';');
+ break;
+ default:
+ shader.AddLine(dest + GetSwizzle(elem) + " = " + src + ';');
+ break;
}
}
@@ -631,8 +651,8 @@ private:
/// Generates declarations for internal flags.
void GenerateInternalFlags() {
- for (u32 ii = 0; ii < static_cast<u64>(InternalFlag::Amount); ii++) {
- const InternalFlag code = static_cast<InternalFlag>(ii);
+ for (u32 flag = 0; flag < static_cast<u32>(InternalFlag::Amount); flag++) {
+ const InternalFlag code = static_cast<InternalFlag>(flag);
declarations.AddLine("bool " + GetInternalFlag(code) + " = false;");
}
declarations.AddNewLine();
@@ -735,12 +755,19 @@ private:
void GenerateVertex() {
if (stage != Maxwell3D::Regs::ShaderStage::Vertex)
return;
+ bool clip_distances_declared = false;
+
declarations.AddLine("out gl_PerVertex {");
++declarations.scope;
declarations.AddLine("vec4 gl_Position;");
for (auto& o : fixed_pipeline_output_attributes_used) {
if (o == Attribute::Index::PointSize)
declarations.AddLine("float gl_PointSize;");
+ if (!clip_distances_declared && (o == Attribute::Index::ClipDistances0123 ||
+ o == Attribute::Index::ClipDistances4567)) {
+ declarations.AddLine("float gl_ClipDistance[];");
+ clip_distances_declared = true;
+ }
}
--declarations.scope;
declarations.AddLine("};");
@@ -911,6 +938,10 @@ private:
return "gl_PointSize";
case Attribute::Index::Position:
return "position";
+ case Attribute::Index::ClipDistances0123:
+ case Attribute::Index::ClipDistances4567: {
+ return "gl_ClipDistance";
+ }
default:
const u32 index{static_cast<u32>(attribute) -
static_cast<u32>(Attribute::Index::Attribute_0)};
@@ -1526,9 +1557,8 @@ private:
instr.fmul.tab5c68_0 != 1, "FMUL tab5cb8_0({}) is not implemented",
instr.fmul.tab5c68_0
.Value()); // SMO typical sends 1 here which seems to be the default
- UNIMPLEMENTED_IF_MSG(instr.fmul.cc != 0, "FMUL cc is not implemented");
UNIMPLEMENTED_IF_MSG(instr.generates_cc,
- "FMUL Generates an unhandled Control Code");
+ "Condition codes generation in FMUL is not implemented");
op_b = GetOperandAbsNeg(op_b, false, instr.fmul.negate_b);
@@ -1540,7 +1570,7 @@ private:
case OpCode::Id::FADD_R:
case OpCode::Id::FADD_IMM: {
UNIMPLEMENTED_IF_MSG(instr.generates_cc,
- "FADD Generates an unhandled Control Code");
+ "Condition codes generation in FADD is not implemented");
op_a = GetOperandAbsNeg(op_a, instr.alu.abs_a, instr.alu.negate_a);
op_b = GetOperandAbsNeg(op_b, instr.alu.abs_b, instr.alu.negate_b);
@@ -1590,7 +1620,7 @@ private:
case OpCode::Id::FMNMX_R:
case OpCode::Id::FMNMX_IMM: {
UNIMPLEMENTED_IF_MSG(instr.generates_cc,
- "FMNMX Generates an unhandled Control Code");
+ "Condition codes generation in FMNMX is not implemented");
op_a = GetOperandAbsNeg(op_a, instr.alu.abs_a, instr.alu.negate_a);
op_b = GetOperandAbsNeg(op_b, instr.alu.abs_b, instr.alu.negate_b);
@@ -1627,7 +1657,7 @@ private:
}
case OpCode::Id::FMUL32_IMM: {
UNIMPLEMENTED_IF_MSG(instr.op_32.generates_cc,
- "FMUL32 Generates an unhandled Control Code");
+ "Condition codes generation in FMUL32 is not implemented");
regs.SetRegisterToFloat(instr.gpr0, 0,
regs.GetRegisterAsFloat(instr.gpr8) + " * " +
@@ -1637,7 +1667,7 @@ private:
}
case OpCode::Id::FADD32I: {
UNIMPLEMENTED_IF_MSG(instr.op_32.generates_cc,
- "FADD32 Generates an unhandled Control Code");
+ "Condition codes generation in FADD32I is not implemented");
std::string op_a = regs.GetRegisterAsFloat(instr.gpr8);
std::string op_b = GetImmediate32(instr);
@@ -1672,7 +1702,8 @@ private:
switch (opcode->get().GetId()) {
case OpCode::Id::BFE_IMM: {
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "BFE Generates an unhandled Control Code");
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in BFE is not implemented");
std::string inner_shift =
'(' + op_a + " << " + std::to_string(instr.bfe.GetLeftShiftValue()) + ')';
@@ -1709,7 +1740,8 @@ private:
case OpCode::Id::SHR_C:
case OpCode::Id::SHR_R:
case OpCode::Id::SHR_IMM: {
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "SHR Generates an unhandled Control Code");
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in SHR is not implemented");
if (!instr.shift.is_signed) {
// Logical shift right
@@ -1724,8 +1756,8 @@ private:
case OpCode::Id::SHL_C:
case OpCode::Id::SHL_R:
case OpCode::Id::SHL_IMM:
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "SHL Generates an unhandled Control Code");
-
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in SHL is not implemented");
regs.SetRegisterToInteger(instr.gpr0, true, 0, op_a + " << " + op_b, 1, 1);
break;
default: {
@@ -1741,7 +1773,7 @@ private:
switch (opcode->get().GetId()) {
case OpCode::Id::IADD32I:
UNIMPLEMENTED_IF_MSG(instr.op_32.generates_cc,
- "IADD32 Generates an unhandled Control Code");
+ "Condition codes generation in IADD32I is not implemented");
if (instr.iadd32i.negate_a)
op_a = "-(" + op_a + ')';
@@ -1751,7 +1783,7 @@ private:
break;
case OpCode::Id::LOP32I: {
UNIMPLEMENTED_IF_MSG(instr.op_32.generates_cc,
- "LOP32I Generates an unhandled Control Code");
+ "Condition codes generation in LOP32I is not implemented");
if (instr.alu.lop32i.invert_a)
op_a = "~(" + op_a + ')';
@@ -1790,7 +1822,7 @@ private:
case OpCode::Id::IADD_R:
case OpCode::Id::IADD_IMM: {
UNIMPLEMENTED_IF_MSG(instr.generates_cc,
- "IADD Generates an unhandled Control Code");
+ "Condition codes generation in IADD is not implemented");
if (instr.alu_integer.negate_a)
op_a = "-(" + op_a + ')';
@@ -1806,7 +1838,7 @@ private:
case OpCode::Id::IADD3_R:
case OpCode::Id::IADD3_IMM: {
UNIMPLEMENTED_IF_MSG(instr.generates_cc,
- "IADD3 Generates an unhandled Control Code");
+ "Condition codes generation in IADD3 is not implemented");
std::string op_c = regs.GetRegisterAsInteger(instr.gpr39);
@@ -1869,7 +1901,7 @@ private:
case OpCode::Id::ISCADD_R:
case OpCode::Id::ISCADD_IMM: {
UNIMPLEMENTED_IF_MSG(instr.generates_cc,
- "ISCADD Generates an unhandled Control Code");
+ "Condition codes generation in ISCADD is not implemented");
if (instr.alu_integer.negate_a)
op_a = "-(" + op_a + ')';
@@ -1904,7 +1936,8 @@ private:
case OpCode::Id::LOP_C:
case OpCode::Id::LOP_R:
case OpCode::Id::LOP_IMM: {
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "LOP Generates an unhandled Control Code");
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in LOP is not implemented");
if (instr.alu.lop.invert_a)
op_a = "~(" + op_a + ')';
@@ -1920,7 +1953,7 @@ private:
case OpCode::Id::LOP3_R:
case OpCode::Id::LOP3_IMM: {
UNIMPLEMENTED_IF_MSG(instr.generates_cc,
- "LOP3 Generates an unhandled Control Code");
+ "Condition codes generation in LOP3 is not implemented");
const std::string op_c = regs.GetRegisterAsInteger(instr.gpr39);
std::string lut;
@@ -1939,7 +1972,7 @@ private:
case OpCode::Id::IMNMX_IMM: {
UNIMPLEMENTED_IF(instr.imnmx.exchange != Tegra::Shader::IMinMaxExchange::None);
UNIMPLEMENTED_IF_MSG(instr.generates_cc,
- "IMNMX Generates an unhandled Control Code");
+ "Condition codes generation in IMNMX is not implemented");
const std::string condition =
GetPredicateCondition(instr.imnmx.pred, instr.imnmx.negate_pred != 0);
@@ -2112,7 +2145,8 @@ private:
instr.ffma.tab5980_0.Value()); // Seems to be 1 by default based on SMO
UNIMPLEMENTED_IF_MSG(instr.ffma.tab5980_1 != 0, "FFMA tab5980_1({}) not implemented",
instr.ffma.tab5980_1.Value());
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "FFMA Generates an unhandled Control Code");
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in FFMA is not implemented");
switch (opcode->get().GetId()) {
case OpCode::Id::FFMA_CR: {
@@ -2222,7 +2256,8 @@ private:
case OpCode::Id::I2F_C: {
UNIMPLEMENTED_IF(instr.conversion.dest_size != Register::Size::Word);
UNIMPLEMENTED_IF(instr.conversion.selector);
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "I2F Generates an unhandled Control Code");
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in I2F is not implemented");
std::string op_a{};
@@ -2252,7 +2287,8 @@ private:
case OpCode::Id::F2F_R: {
UNIMPLEMENTED_IF(instr.conversion.dest_size != Register::Size::Word);
UNIMPLEMENTED_IF(instr.conversion.src_size != Register::Size::Word);
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "F2F Generates an unhandled Control Code");
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in F2F is not implemented");
std::string op_a = regs.GetRegisterAsFloat(instr.gpr20);
if (instr.conversion.abs_a) {
@@ -2290,7 +2326,8 @@ private:
case OpCode::Id::F2I_R:
case OpCode::Id::F2I_C: {
UNIMPLEMENTED_IF(instr.conversion.src_size != Register::Size::Word);
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "F2I Generates an unhandled Control Code");
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in F2I is not implemented");
std::string op_a{};
if (instr.is_b_gpr) {
@@ -3201,7 +3238,8 @@ private:
break;
}
case OpCode::Type::PredicateSetRegister: {
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "PSET Generates an unhandled Control Code");
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in PSET is not implemented");
const std::string op_a =
GetPredicateCondition(instr.pset.pred12, instr.pset.neg_pred12 != 0);
@@ -3260,14 +3298,14 @@ private:
const std::string pred =
GetPredicateCondition(instr.csetp.pred39, instr.csetp.neg_pred39 != 0);
const std::string combiner = GetPredicateCombiner(instr.csetp.op);
- const std::string control_code = regs.GetControlCode(instr.csetp.cc);
+ const std::string condition_code = regs.GetConditionCode(instr.csetp.cc);
if (instr.csetp.pred3 != static_cast<u64>(Pred::UnusedIndex)) {
SetPredicate(instr.csetp.pred3,
- '(' + control_code + ") " + combiner + " (" + pred + ')');
+ '(' + condition_code + ") " + combiner + " (" + pred + ')');
}
if (instr.csetp.pred0 != static_cast<u64>(Pred::UnusedIndex)) {
SetPredicate(instr.csetp.pred0,
- "!(" + control_code + ") " + combiner + " (" + pred + ')');
+ "!(" + condition_code + ") " + combiner + " (" + pred + ')');
}
break;
}
@@ -3398,7 +3436,8 @@ private:
case OpCode::Type::Xmad: {
UNIMPLEMENTED_IF(instr.xmad.sign_a);
UNIMPLEMENTED_IF(instr.xmad.sign_b);
- UNIMPLEMENTED_IF_MSG(instr.generates_cc, "XMAD Generates an unhandled Control Code");
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in XMAD is not implemented");
std::string op_a{regs.GetRegisterAsInteger(instr.gpr8, 0, instr.xmad.sign_a)};
std::string op_b;
@@ -3490,9 +3529,9 @@ private:
default: {
switch (opcode->get().GetId()) {
case OpCode::Id::EXIT: {
- const Tegra::Shader::ControlCode cc = instr.flow_control_code;
- UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ControlCode::T,
- "EXIT Control Code used: {}", static_cast<u32>(cc));
+ const Tegra::Shader::ConditionCode cc = instr.flow_condition_code;
+ UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T,
+ "EXIT condition code used: {}", static_cast<u32>(cc));
if (stage == Maxwell3D::Regs::ShaderStage::Fragment) {
EmitFragmentOutputsWrite();
@@ -3524,9 +3563,9 @@ private:
case OpCode::Id::KIL: {
UNIMPLEMENTED_IF(instr.flow.cond != Tegra::Shader::FlowCondition::Always);
- const Tegra::Shader::ControlCode cc = instr.flow_control_code;
- UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ControlCode::T,
- "KIL Control Code used: {}", static_cast<u32>(cc));
+ const Tegra::Shader::ConditionCode cc = instr.flow_condition_code;
+ UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T,
+ "KIL condition code used: {}", static_cast<u32>(cc));
// Enclose "discard" in a conditional, so that GLSL compilation does not complain
// about unexecuted instructions that may follow this.
@@ -3588,9 +3627,9 @@ private:
UNIMPLEMENTED_IF_MSG(instr.bra.constant_buffer != 0,
"BRA with constant buffers are not implemented");
- const Tegra::Shader::ControlCode cc = instr.flow_control_code;
- UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ControlCode::T,
- "BRA Control Code used: {}", static_cast<u32>(cc));
+ const Tegra::Shader::ConditionCode cc = instr.flow_condition_code;
+ UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T,
+ "BRA condition code used: {}", static_cast<u32>(cc));
const u32 target = offset + instr.bra.GetBranchTarget();
shader.AddLine("{ jmp_to = " + std::to_string(target) + "u; break; }");
@@ -3633,9 +3672,9 @@ private:
break;
}
case OpCode::Id::SYNC: {
- const Tegra::Shader::ControlCode cc = instr.flow_control_code;
- UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ControlCode::T,
- "SYNC Control Code used: {}", static_cast<u32>(cc));
+ const Tegra::Shader::ConditionCode cc = instr.flow_condition_code;
+ UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T,
+ "SYNC condition code used: {}", static_cast<u32>(cc));
// The SYNC opcode jumps to the address previously set by the SSY opcode
EmitPopFromFlowStack();
@@ -3643,10 +3682,10 @@ private:
}
case OpCode::Id::BRK: {
// The BRK opcode jumps to the address previously set by the PBK opcode
- const Tegra::Shader::ControlCode cc = instr.flow_control_code;
- if (cc != Tegra::Shader::ControlCode::T) {
- UNIMPLEMENTED_MSG("BRK Control Code used: {}", static_cast<u32>(cc));
- }
+ const Tegra::Shader::ConditionCode cc = instr.flow_condition_code;
+ UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T,
+ "BRK condition code used: {}", static_cast<u32>(cc));
+
EmitPopFromFlowStack();
break;
}
@@ -3657,6 +3696,9 @@ private:
break;
}
case OpCode::Id::VMAD: {
+ UNIMPLEMENTED_IF_MSG(instr.generates_cc,
+ "Condition codes generation in VMAD is not implemented");
+
const bool result_signed = instr.video.signed_a == 1 || instr.video.signed_b == 1;
const std::string op_a = GetVideoOperandA(instr);
const std::string op_b = GetVideoOperandB(instr);
@@ -3676,10 +3718,6 @@ private:
regs.SetRegisterToInteger(instr.gpr0, result_signed, 1, result, 1, 1,
instr.vmad.saturate == 1, 0, Register::Size::Word,
instr.vmad.cc);
- if (instr.generates_cc) {
- UNIMPLEMENTED_MSG("VMAD Generates an unhandled Control Code");
- }
-
break;
}
case OpCode::Id::VSETP: {
diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp
index d9910c6e8..934f4db78 100644
--- a/src/video_core/renderer_opengl/gl_state.cpp
+++ b/src/video_core/renderer_opengl/gl_state.cpp
@@ -233,6 +233,28 @@ void OpenGLState::ApplyStencilTest() const {
config_stencil(GL_BACK, stencil.back, cur_state.stencil.back);
}
}
+// Viewport does not affects glClearBuffer so emulate viewport using scissor test
+void OpenGLState::EmulateViewportWithScissor() {
+ auto& current = viewports[0];
+ if (current.scissor.enabled) {
+ const GLint left = std::max(current.x, current.scissor.x);
+ const GLint right =
+ std::max(current.x + current.width, current.scissor.x + current.scissor.width);
+ const GLint bottom = std::max(current.y, current.scissor.y);
+ const GLint top =
+ std::max(current.y + current.height, current.scissor.y + current.scissor.height);
+ current.scissor.x = std::max(left, 0);
+ current.scissor.y = std::max(bottom, 0);
+ current.scissor.width = std::max(right - left, 0);
+ current.scissor.height = std::max(top - bottom, 0);
+ } else {
+ current.scissor.enabled = true;
+ current.scissor.x = current.x;
+ current.scissor.y = current.y;
+ current.scissor.width = current.width;
+ current.scissor.height = current.height;
+ }
+}
void OpenGLState::ApplyViewport() const {
if (GLAD_GL_ARB_viewport_array && geometry_shaders.enabled) {
@@ -242,7 +264,9 @@ void OpenGLState::ApplyViewport() const {
const auto& updated = viewports[i];
if (updated.x != current.x || updated.y != current.y ||
updated.width != current.width || updated.height != current.height) {
- glViewportIndexedf(i, updated.x, updated.y, updated.width, updated.height);
+ glViewportIndexedf(
+ i, static_cast<GLfloat>(updated.x), static_cast<GLfloat>(updated.y),
+ static_cast<GLfloat>(updated.width), static_cast<GLfloat>(updated.height));
}
if (updated.depth_range_near != current.depth_range_near ||
updated.depth_range_far != current.depth_range_far) {
@@ -270,8 +294,7 @@ void OpenGLState::ApplyViewport() const {
const auto& updated = viewports[0];
if (updated.x != current.x || updated.y != current.y || updated.width != current.width ||
updated.height != current.height) {
- glViewport(static_cast<GLint>(updated.x), static_cast<GLint>(updated.y),
- static_cast<GLsizei>(updated.width), static_cast<GLsizei>(updated.height));
+ glViewport(updated.x, updated.y, updated.width, updated.height);
}
if (updated.depth_range_near != current.depth_range_near ||
updated.depth_range_far != current.depth_range_far) {
@@ -339,14 +362,14 @@ void OpenGLState::ApplyTargetBlending(std::size_t target, bool force) const {
if (blend_changed || updated.src_rgb_func != current.src_rgb_func ||
updated.dst_rgb_func != current.dst_rgb_func || updated.src_a_func != current.src_a_func ||
updated.dst_a_func != current.dst_a_func) {
- glBlendFuncSeparateiARB(static_cast<GLuint>(target), updated.src_rgb_func,
- updated.dst_rgb_func, updated.src_a_func, updated.dst_a_func);
+ glBlendFuncSeparatei(static_cast<GLuint>(target), updated.src_rgb_func,
+ updated.dst_rgb_func, updated.src_a_func, updated.dst_a_func);
}
if (blend_changed || updated.rgb_equation != current.rgb_equation ||
updated.a_equation != current.a_equation) {
- glBlendEquationSeparateiARB(static_cast<GLuint>(target), updated.rgb_equation,
- updated.a_equation);
+ glBlendEquationSeparatei(static_cast<GLuint>(target), updated.rgb_equation,
+ updated.a_equation);
}
}
diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h
index bdc743b0f..032fc43f0 100644
--- a/src/video_core/renderer_opengl/gl_state.h
+++ b/src/video_core/renderer_opengl/gl_state.h
@@ -156,10 +156,10 @@ public:
} draw;
struct viewport {
- GLfloat x;
- GLfloat y;
- GLfloat width;
- GLfloat height;
+ GLint x;
+ GLint y;
+ GLint width;
+ GLint height;
GLfloat depth_range_near; // GL_DEPTH_RANGE
GLfloat depth_range_far; // GL_DEPTH_RANGE
struct {
@@ -206,6 +206,7 @@ public:
OpenGLState& ResetBuffer(GLuint handle);
OpenGLState& ResetVertexArray(GLuint handle);
OpenGLState& ResetFramebuffer(GLuint handle);
+ void EmulateViewportWithScissor();
private:
static OpenGLState cur_state;
diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp
index 27b5b8960..4fd0d66c5 100644
--- a/src/video_core/renderer_opengl/renderer_opengl.cpp
+++ b/src/video_core/renderer_opengl/renderer_opengl.cpp
@@ -19,9 +19,9 @@
#include "core/settings.h"
#include "core/telemetry_session.h"
#include "core/tracer/recorder.h"
+#include "video_core/morton.h"
#include "video_core/renderer_opengl/gl_rasterizer.h"
#include "video_core/renderer_opengl/renderer_opengl.h"
-#include "video_core/utils.h"
namespace OpenGL {
@@ -490,7 +490,7 @@ bool RendererOpenGL::Init() {
Core::Telemetry().AddField(Telemetry::FieldType::UserSystem, "GPU_Model", gpu_model);
Core::Telemetry().AddField(Telemetry::FieldType::UserSystem, "GPU_OpenGL_Version", gl_version);
- if (!GLAD_GL_VERSION_3_3) {
+ if (!GLAD_GL_VERSION_4_3) {
return false;
}
diff --git a/src/video_core/utils.h b/src/video_core/utils.h
deleted file mode 100644
index e0a14d48f..000000000
--- a/src/video_core/utils.h
+++ /dev/null
@@ -1,164 +0,0 @@
-// Copyright 2014 Citra Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include "common/common_types.h"
-
-namespace VideoCore {
-
-// 8x8 Z-Order coordinate from 2D coordinates
-static inline u32 MortonInterleave(u32 x, u32 y) {
- static const u32 xlut[] = {0x00, 0x01, 0x04, 0x05, 0x10, 0x11, 0x14, 0x15};
- static const u32 ylut[] = {0x00, 0x02, 0x08, 0x0a, 0x20, 0x22, 0x28, 0x2a};
- return xlut[x % 8] + ylut[y % 8];
-}
-
-/**
- * Calculates the offset of the position of the pixel in Morton order
- */
-static inline u32 GetMortonOffset(u32 x, u32 y, u32 bytes_per_pixel) {
- // Images are split into 8x8 tiles. Each tile is composed of four 4x4 subtiles each
- // of which is composed of four 2x2 subtiles each of which is composed of four texels.
- // Each structure is embedded into the next-bigger one in a diagonal pattern, e.g.
- // texels are laid out in a 2x2 subtile like this:
- // 2 3
- // 0 1
- //
- // The full 8x8 tile has the texels arranged like this:
- //
- // 42 43 46 47 58 59 62 63
- // 40 41 44 45 56 57 60 61
- // 34 35 38 39 50 51 54 55
- // 32 33 36 37 48 49 52 53
- // 10 11 14 15 26 27 30 31
- // 08 09 12 13 24 25 28 29
- // 02 03 06 07 18 19 22 23
- // 00 01 04 05 16 17 20 21
- //
- // This pattern is what's called Z-order curve, or Morton order.
-
- const unsigned int block_height = 8;
- const unsigned int coarse_x = x & ~7;
-
- u32 i = VideoCore::MortonInterleave(x, y);
-
- const unsigned int offset = coarse_x * block_height;
-
- return (i + offset) * bytes_per_pixel;
-}
-
-static inline u32 MortonInterleave128(u32 x, u32 y) {
- // 128x128 Z-Order coordinate from 2D coordinates
- static constexpr u32 xlut[] = {
- 0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042,
- 0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809,
- 0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000,
- 0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043,
- 0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a,
- 0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001,
- 0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048,
- 0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b,
- 0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002,
- 0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049,
- 0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840,
- 0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003,
- 0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a,
- 0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841,
- 0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008,
- 0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b,
- 0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842,
- 0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009,
- 0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800,
- 0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843,
- 0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a,
- 0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801,
- 0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848,
- 0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b,
- 0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802,
- 0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849,
- 0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040,
- 0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803,
- 0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a,
- 0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041,
- 0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808,
- 0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b,
- 0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042,
- 0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809,
- 0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b,
- };
- static constexpr u32 ylut[] = {
- 0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090,
- 0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124,
- 0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200,
- 0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294,
- 0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330,
- 0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404,
- 0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0,
- 0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534,
- 0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610,
- 0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4,
- 0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780,
- 0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014,
- 0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0,
- 0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184,
- 0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220,
- 0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4,
- 0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390,
- 0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424,
- 0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500,
- 0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594,
- 0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630,
- 0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704,
- 0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0,
- 0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034,
- 0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110,
- 0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4,
- 0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280,
- 0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314,
- 0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0,
- 0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484,
- 0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520,
- 0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4,
- 0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690,
- 0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724,
- 0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4,
- };
- return xlut[x % 128] + ylut[y % 128];
-}
-
-static inline u32 GetMortonOffset128(u32 x, u32 y, u32 bytes_per_pixel) {
- // Calculates the offset of the position of the pixel in Morton order
- // Framebuffer images are split into 128x128 tiles.
-
- const unsigned int block_height = 128;
- const unsigned int coarse_x = x & ~127;
-
- u32 i = MortonInterleave128(x, y);
-
- const unsigned int offset = coarse_x * block_height;
-
- return (i + offset) * bytes_per_pixel;
-}
-
-static inline void MortonCopyPixels128(u32 width, u32 height, u32 bytes_per_pixel,
- u32 gl_bytes_per_pixel, u8* morton_data, u8* gl_data,
- bool morton_to_gl) {
- u8* data_ptrs[2];
- for (unsigned y = 0; y < height; ++y) {
- for (unsigned x = 0; x < width; ++x) {
- const u32 coarse_y = y & ~127;
- u32 morton_offset =
- GetMortonOffset128(x, y, bytes_per_pixel) + coarse_y * width * bytes_per_pixel;
- u32 gl_pixel_index = (x + y * width) * gl_bytes_per_pixel;
-
- data_ptrs[morton_to_gl] = morton_data + morton_offset;
- data_ptrs[!morton_to_gl] = &gl_data[gl_pixel_index];
-
- memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
- }
- }
-}
-
-} // namespace VideoCore
generated by cgit v1.2.3 (git 2.25.1) at 2024-07-15 10:42:00 +0200