You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 
cocos_lib/cocos/renderer/gfx-gles3/GLES3Commands.cpp

3267 lines
150 KiB

/****************************************************************************
Copyright (c) 2019-2023 Xiamen Yaji Software Co., Ltd.
http://www.cocos.com
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights to
use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
of the Software, and to permit persons to whom the Software is furnished to do so,
subject to the following conditions:
The above copyright notice and this permission notice shall be included in
all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.
****************************************************************************/
#include "GLES3Commands.h"
#include "GLES3Device.h"
#include "GLES3PipelineCache.h"
#include "GLES3QueryPool.h"
#include "GLES3Std.h"
#include "base/StringUtil.h"
#include "gfx-base/GFXDef-common.h"
#include "gfx-gles-common/GLESCommandPool.h"
#include "gfx-gles3/GLES3GPUObjects.h"
#include "gfx-gles3/states/GLES3Sampler.h"
#define BUFFER_OFFSET(idx) (static_cast<char *>(0) + (idx))
constexpr uint32_t USE_VAO = true;
namespace cc {
namespace gfx {
namespace {
GLenum mapGLInternalFormat(Format format) {
switch (format) {
case Format::A8: return GL_ALPHA;
case Format::L8: return GL_LUMINANCE;
case Format::LA8: return GL_LUMINANCE_ALPHA;
case Format::R8: return GL_R8;
case Format::R8SN: return GL_R8_SNORM;
case Format::R8UI: return GL_R8UI;
case Format::R8I: return GL_R8I;
case Format::RG8: return GL_RG8;
case Format::RG8SN: return GL_RG8_SNORM;
case Format::RG8UI: return GL_RG8UI;
case Format::RG8I: return GL_RG8I;
case Format::RGB8: return GL_RGB8;
case Format::RGB8SN: return GL_RGB8_SNORM;
case Format::RGB8UI: return GL_RGB8UI;
case Format::RGB8I: return GL_RGB8I;
case Format::SRGB8: return GL_SRGB_EXT;
case Format::RGBA8: return GL_RGBA8;
case Format::RGBA8SN: return GL_RGBA8_SNORM;
case Format::RGBA8UI: return GL_RGBA8UI;
case Format::RGBA8I: return GL_RGBA8I;
case Format::SRGB8_A8: return GL_SRGB_ALPHA_EXT;
case Format::R16I: return GL_R16I;
case Format::R16UI: return GL_R16UI;
case Format::R16F: return GL_R16F;
case Format::RG16I: return GL_RG16I;
case Format::RG16UI: return GL_RG16UI;
case Format::RG16F: return GL_RG16F;
case Format::RGB16I: return GL_RGB16I;
case Format::RGB16UI: return GL_RGB16UI;
case Format::RGB16F: return GL_RGB16F;
case Format::RGBA16I: return GL_RGBA16I;
case Format::RGBA16UI: return GL_RGBA16UI;
case Format::RGBA16F: return GL_RGBA16F;
case Format::R32I: return GL_R32I;
case Format::R32UI: return GL_R32UI;
case Format::R32F: return GL_R32F;
case Format::RG32I: return GL_RG32I;
case Format::RG32UI: return GL_RG32UI;
case Format::RG32F: return GL_RG32F;
case Format::RGB32I: return GL_RGB32I;
case Format::RGB32UI: return GL_RGB32UI;
case Format::RGB32F: return GL_RGB32F;
case Format::RGBA32I: return GL_RGBA32I;
case Format::RGBA32UI: return GL_RGBA32UI;
case Format::RGBA32F: return GL_RGBA32F;
case Format::R5G6B5: return GL_RGB565;
case Format::RGB5A1: return GL_RGB5_A1;
case Format::RGB9E5: return GL_RGB9_E5;
case Format::RGBA4: return GL_RGBA4;
case Format::RGB10A2: return GL_RGB10_A2;
case Format::RGB10A2UI: return GL_RGB10_A2UI;
case Format::R11G11B10F: return GL_R11F_G11F_B10F;
case Format::DEPTH: return GL_DEPTH_COMPONENT32F;
case Format::DEPTH_STENCIL: return GL_DEPTH24_STENCIL8;
case Format::BC1: return GL_COMPRESSED_RGB_S3TC_DXT1_EXT;
case Format::BC1_ALPHA: return GL_COMPRESSED_RGBA_S3TC_DXT1_EXT;
case Format::BC1_SRGB: return GL_COMPRESSED_SRGB_S3TC_DXT1_EXT;
case Format::BC1_SRGB_ALPHA: return GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT;
case Format::BC2: return GL_COMPRESSED_RGBA_S3TC_DXT3_EXT;
case Format::BC2_SRGB: return GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT;
case Format::BC3: return GL_COMPRESSED_RGBA_S3TC_DXT5_EXT;
case Format::BC3_SRGB: return GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT;
case Format::ETC_RGB8: return GL_ETC1_RGB8_OES;
case Format::ETC2_RGB8: return GL_COMPRESSED_RGB8_ETC2;
case Format::ETC2_SRGB8: return GL_COMPRESSED_SRGB8_ETC2;
case Format::ETC2_RGB8_A1: return GL_COMPRESSED_RGB8_PUNCHTHROUGH_ALPHA1_ETC2;
case Format::ETC2_SRGB8_A1: return GL_COMPRESSED_SRGB8_PUNCHTHROUGH_ALPHA1_ETC2;
case Format::ETC2_RGBA8: return GL_COMPRESSED_RGBA8_ETC2_EAC;
case Format::ETC2_SRGB8_A8: return GL_COMPRESSED_SRGB8_ALPHA8_ETC2_EAC;
case Format::EAC_R11: return GL_COMPRESSED_R11_EAC;
case Format::EAC_R11SN: return GL_COMPRESSED_SIGNED_R11_EAC;
case Format::EAC_RG11: return GL_COMPRESSED_RG11_EAC;
case Format::EAC_RG11SN: return GL_COMPRESSED_SIGNED_RG11_EAC;
case Format::PVRTC_RGB2: return GL_COMPRESSED_RGB_PVRTC_2BPPV1_IMG;
case Format::PVRTC_RGBA2: return GL_COMPRESSED_RGBA_PVRTC_2BPPV1_IMG;
case Format::PVRTC_RGB4: return GL_COMPRESSED_RGB_PVRTC_4BPPV1_IMG;
case Format::PVRTC_RGBA4: return GL_COMPRESSED_RGBA_PVRTC_4BPPV1_IMG;
case Format::ASTC_RGBA_4X4: return GL_COMPRESSED_RGBA_ASTC_4x4_KHR;
case Format::ASTC_RGBA_5X4: return GL_COMPRESSED_RGBA_ASTC_5x4_KHR;
case Format::ASTC_RGBA_5X5: return GL_COMPRESSED_RGBA_ASTC_5x5_KHR;
case Format::ASTC_RGBA_6X5: return GL_COMPRESSED_RGBA_ASTC_6x5_KHR;
case Format::ASTC_RGBA_6X6: return GL_COMPRESSED_RGBA_ASTC_6x6_KHR;
case Format::ASTC_RGBA_8X5: return GL_COMPRESSED_RGBA_ASTC_8x5_KHR;
case Format::ASTC_RGBA_8X6: return GL_COMPRESSED_RGBA_ASTC_8x6_KHR;
case Format::ASTC_RGBA_8X8: return GL_COMPRESSED_RGBA_ASTC_8x8_KHR;
case Format::ASTC_RGBA_10X5: return GL_COMPRESSED_RGBA_ASTC_10x5_KHR;
case Format::ASTC_RGBA_10X6: return GL_COMPRESSED_RGBA_ASTC_10x6_KHR;
case Format::ASTC_RGBA_10X8: return GL_COMPRESSED_RGBA_ASTC_10x8_KHR;
case Format::ASTC_RGBA_10X10: return GL_COMPRESSED_RGBA_ASTC_10x10_KHR;
case Format::ASTC_RGBA_12X10: return GL_COMPRESSED_RGBA_ASTC_12x10_KHR;
case Format::ASTC_RGBA_12X12: return GL_COMPRESSED_RGBA_ASTC_12x12_KHR;
case Format::ASTC_SRGBA_4X4: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_4x4_KHR;
case Format::ASTC_SRGBA_5X4: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x4_KHR;
case Format::ASTC_SRGBA_5X5: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x5_KHR;
case Format::ASTC_SRGBA_6X5: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x5_KHR;
case Format::ASTC_SRGBA_6X6: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x6_KHR;
case Format::ASTC_SRGBA_8X5: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x5_KHR;
case Format::ASTC_SRGBA_8X6: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x6_KHR;
case Format::ASTC_SRGBA_8X8: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x8_KHR;
case Format::ASTC_SRGBA_10X5: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x5_KHR;
case Format::ASTC_SRGBA_10X6: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x6_KHR;
case Format::ASTC_SRGBA_10X8: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x8_KHR;
case Format::ASTC_SRGBA_10X10: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x10_KHR;
case Format::ASTC_SRGBA_12X10: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x10_KHR;
case Format::ASTC_SRGBA_12X12: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x12_KHR;
default: {
CC_ABORT();
return GL_NONE;
}
}
}
GLenum mapGLFormat(Format format) {
switch (format) {
case Format::A8: return GL_ALPHA;
case Format::L8: return GL_LUMINANCE;
case Format::LA8: return GL_LUMINANCE_ALPHA;
case Format::R8:
case Format::R8SN:
case Format::R16F:
case Format::R32F: return GL_RED;
case Format::RG8:
case Format::RG8SN:
case Format::RG16F:
case Format::RG32F: return GL_RG;
case Format::RGB8:
case Format::RGB8SN:
case Format::RGB16F:
case Format::RGB32F:
case Format::R11G11B10F:
case Format::R5G6B5:
case Format::RGB9E5:
case Format::SRGB8: return GL_RGB;
case Format::RGBA8:
case Format::RGBA8SN:
case Format::RGBA16F:
case Format::RGBA32F:
case Format::RGBA4:
case Format::RGB5A1:
case Format::RGB10A2:
case Format::SRGB8_A8: return GL_RGBA;
case Format::R8UI:
case Format::R8I:
case Format::R16UI:
case Format::R16I:
case Format::R32UI:
case Format::R32I: return GL_RED_INTEGER;
case Format::RG8UI:
case Format::RG8I:
case Format::RG16UI:
case Format::RG16I:
case Format::RG32UI:
case Format::RG32I: return GL_RG_INTEGER;
case Format::RGB8UI:
case Format::RGB8I:
case Format::RGB16UI:
case Format::RGB16I:
case Format::RGB32UI:
case Format::RGB32I: return GL_RGB_INTEGER;
case Format::RGBA8UI:
case Format::RGBA8I:
case Format::RGBA16UI:
case Format::RGBA16I:
case Format::RGBA32UI:
case Format::RGBA32I:
case Format::RGB10A2UI: return GL_RGBA_INTEGER;
case Format::DEPTH: return GL_DEPTH_COMPONENT;
case Format::DEPTH_STENCIL: return GL_DEPTH_STENCIL;
case Format::BC1: return GL_COMPRESSED_RGB_S3TC_DXT1_EXT;
case Format::BC1_ALPHA: return GL_COMPRESSED_RGBA_S3TC_DXT1_EXT;
case Format::BC1_SRGB: return GL_COMPRESSED_SRGB_S3TC_DXT1_EXT;
case Format::BC1_SRGB_ALPHA: return GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT;
case Format::BC2: return GL_COMPRESSED_RGBA_S3TC_DXT3_EXT;
case Format::BC2_SRGB: return GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT;
case Format::BC3: return GL_COMPRESSED_RGBA_S3TC_DXT5_EXT;
case Format::BC3_SRGB: return GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT;
case Format::ETC_RGB8: return GL_ETC1_RGB8_OES;
case Format::ETC2_RGB8: return GL_COMPRESSED_RGB8_ETC2;
case Format::ETC2_SRGB8: return GL_COMPRESSED_SRGB8_ETC2;
case Format::ETC2_RGB8_A1: return GL_COMPRESSED_RGB8_PUNCHTHROUGH_ALPHA1_ETC2;
case Format::ETC2_SRGB8_A1: return GL_COMPRESSED_SRGB8_PUNCHTHROUGH_ALPHA1_ETC2;
case Format::ETC2_RGBA8: return GL_COMPRESSED_RGBA8_ETC2_EAC;
case Format::ETC2_SRGB8_A8: return GL_COMPRESSED_SRGB8_ALPHA8_ETC2_EAC;
case Format::EAC_R11: return GL_COMPRESSED_R11_EAC;
case Format::EAC_R11SN: return GL_COMPRESSED_SIGNED_R11_EAC;
case Format::EAC_RG11: return GL_COMPRESSED_RG11_EAC;
case Format::EAC_RG11SN: return GL_COMPRESSED_SIGNED_RG11_EAC;
case Format::PVRTC_RGB2: return GL_COMPRESSED_RGB_PVRTC_2BPPV1_IMG;
case Format::PVRTC_RGBA2: return GL_COMPRESSED_RGBA_PVRTC_2BPPV1_IMG;
case Format::PVRTC_RGB4: return GL_COMPRESSED_RGB_PVRTC_4BPPV1_IMG;
case Format::PVRTC_RGBA4: return GL_COMPRESSED_RGBA_PVRTC_4BPPV1_IMG;
case Format::ASTC_RGBA_4X4: return GL_COMPRESSED_RGBA_ASTC_4x4_KHR;
case Format::ASTC_RGBA_5X4: return GL_COMPRESSED_RGBA_ASTC_5x4_KHR;
case Format::ASTC_RGBA_5X5: return GL_COMPRESSED_RGBA_ASTC_5x5_KHR;
case Format::ASTC_RGBA_6X5: return GL_COMPRESSED_RGBA_ASTC_6x5_KHR;
case Format::ASTC_RGBA_6X6: return GL_COMPRESSED_RGBA_ASTC_6x6_KHR;
case Format::ASTC_RGBA_8X5: return GL_COMPRESSED_RGBA_ASTC_8x5_KHR;
case Format::ASTC_RGBA_8X6: return GL_COMPRESSED_RGBA_ASTC_8x6_KHR;
case Format::ASTC_RGBA_8X8: return GL_COMPRESSED_RGBA_ASTC_8x8_KHR;
case Format::ASTC_RGBA_10X5: return GL_COMPRESSED_RGBA_ASTC_10x5_KHR;
case Format::ASTC_RGBA_10X6: return GL_COMPRESSED_RGBA_ASTC_10x6_KHR;
case Format::ASTC_RGBA_10X8: return GL_COMPRESSED_RGBA_ASTC_10x8_KHR;
case Format::ASTC_RGBA_10X10: return GL_COMPRESSED_RGBA_ASTC_10x10_KHR;
case Format::ASTC_RGBA_12X10: return GL_COMPRESSED_RGBA_ASTC_12x10_KHR;
case Format::ASTC_RGBA_12X12: return GL_COMPRESSED_RGBA_ASTC_12x12_KHR;
case Format::ASTC_SRGBA_4X4: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_4x4_KHR;
case Format::ASTC_SRGBA_5X4: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x4_KHR;
case Format::ASTC_SRGBA_5X5: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x5_KHR;
case Format::ASTC_SRGBA_6X5: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x5_KHR;
case Format::ASTC_SRGBA_6X6: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x6_KHR;
case Format::ASTC_SRGBA_8X5: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x5_KHR;
case Format::ASTC_SRGBA_8X6: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x6_KHR;
case Format::ASTC_SRGBA_8X8: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x8_KHR;
case Format::ASTC_SRGBA_10X5: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x5_KHR;
case Format::ASTC_SRGBA_10X6: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x6_KHR;
case Format::ASTC_SRGBA_10X8: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x8_KHR;
case Format::ASTC_SRGBA_10X10: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x10_KHR;
case Format::ASTC_SRGBA_12X10: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x10_KHR;
case Format::ASTC_SRGBA_12X12: return GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x12_KHR;
default: {
CC_ABORT();
return GL_NONE;
}
}
}
GLenum mapGLType(Type type) {
switch (type) {
case Type::BOOL: return GL_BOOL;
case Type::BOOL2: return GL_BOOL_VEC2;
case Type::BOOL3: return GL_BOOL_VEC3;
case Type::BOOL4: return GL_BOOL_VEC4;
case Type::INT: return GL_INT;
case Type::INT2: return GL_INT_VEC2;
case Type::INT3: return GL_INT_VEC3;
case Type::INT4: return GL_INT_VEC4;
case Type::UINT: return GL_UNSIGNED_INT;
case Type::FLOAT: return GL_FLOAT;
case Type::FLOAT2: return GL_FLOAT_VEC2;
case Type::FLOAT3: return GL_FLOAT_VEC3;
case Type::FLOAT4: return GL_FLOAT_VEC4;
case Type::MAT2: return GL_FLOAT_MAT2;
case Type::MAT2X3: return GL_FLOAT_MAT2x3;
case Type::MAT2X4: return GL_FLOAT_MAT2x4;
case Type::MAT3X2: return GL_FLOAT_MAT3x2;
case Type::MAT3: return GL_FLOAT_MAT3;
case Type::MAT3X4: return GL_FLOAT_MAT3x4;
case Type::MAT4X2: return GL_FLOAT_MAT4x2;
case Type::MAT4X3: return GL_FLOAT_MAT4x3;
case Type::MAT4: return GL_FLOAT_MAT4;
case Type::SAMPLER2D: return GL_SAMPLER_2D;
case Type::SAMPLER2D_ARRAY: return GL_SAMPLER_2D_ARRAY;
case Type::SAMPLER3D: return GL_SAMPLER_3D;
case Type::SAMPLER_CUBE: return GL_SAMPLER_CUBE;
default: {
CC_ABORT();
return GL_NONE;
}
}
}
Type mapType(GLenum glType) {
switch (glType) {
case GL_BOOL: return Type::BOOL;
case GL_BOOL_VEC2: return Type::BOOL2;
case GL_BOOL_VEC3: return Type::BOOL3;
case GL_BOOL_VEC4: return Type::BOOL4;
case GL_INT: return Type::INT;
case GL_INT_VEC2: return Type::INT2;
case GL_INT_VEC3: return Type::INT3;
case GL_INT_VEC4: return Type::INT4;
case GL_UNSIGNED_INT: return Type::UINT;
case GL_UNSIGNED_INT_VEC2: return Type::UINT2;
case GL_UNSIGNED_INT_VEC3: return Type::UINT3;
case GL_UNSIGNED_INT_VEC4: return Type::UINT4;
case GL_FLOAT: return Type::FLOAT;
case GL_FLOAT_VEC2: return Type::FLOAT2;
case GL_FLOAT_VEC3: return Type::FLOAT3;
case GL_FLOAT_VEC4: return Type::FLOAT4;
case GL_FLOAT_MAT2: return Type::MAT2;
case GL_FLOAT_MAT2x3: return Type::MAT2X3;
case GL_FLOAT_MAT2x4: return Type::MAT2X4;
case GL_FLOAT_MAT3x2: return Type::MAT3X2;
case GL_FLOAT_MAT3: return Type::MAT3;
case GL_FLOAT_MAT3x4: return Type::MAT3X4;
case GL_FLOAT_MAT4x2: return Type::MAT4X2;
case GL_FLOAT_MAT4x3: return Type::MAT4X3;
case GL_FLOAT_MAT4: return Type::MAT4;
case GL_SAMPLER_2D: return Type::SAMPLER2D;
case GL_SAMPLER_2D_ARRAY: return Type::SAMPLER2D_ARRAY;
case GL_SAMPLER_3D: return Type::SAMPLER3D;
case GL_SAMPLER_CUBE: return Type::SAMPLER_CUBE;
default: {
CC_ABORT();
return Type::UNKNOWN;
}
}
}
GLenum formatToGLType(Format format) {
switch (format) {
case Format::R8: return GL_UNSIGNED_BYTE;
case Format::R8SN: return GL_BYTE;
case Format::R8UI: return GL_UNSIGNED_BYTE;
case Format::R8I: return GL_BYTE;
case Format::R16F: return GL_HALF_FLOAT;
case Format::R16UI: return GL_UNSIGNED_SHORT;
case Format::R16I: return GL_SHORT;
case Format::R32F: return GL_FLOAT;
case Format::R32UI: return GL_UNSIGNED_INT;
case Format::R32I: return GL_INT;
case Format::RG8: return GL_UNSIGNED_BYTE;
case Format::RG8SN: return GL_BYTE;
case Format::RG8UI: return GL_UNSIGNED_BYTE;
case Format::RG8I: return GL_BYTE;
case Format::RG16F: return GL_HALF_FLOAT;
case Format::RG16UI: return GL_UNSIGNED_SHORT;
case Format::RG16I: return GL_SHORT;
case Format::RG32F: return GL_FLOAT;
case Format::RG32UI: return GL_UNSIGNED_INT;
case Format::RG32I: return GL_INT;
case Format::RGB8:
case Format::SRGB8: return GL_UNSIGNED_BYTE;
case Format::RGB8SN: return GL_BYTE;
case Format::RGB8UI: return GL_UNSIGNED_BYTE;
case Format::RGB8I: return GL_BYTE;
case Format::RGB16F: return GL_HALF_FLOAT;
case Format::RGB16UI: return GL_UNSIGNED_SHORT;
case Format::RGB16I: return GL_SHORT;
case Format::RGB32F: return GL_FLOAT;
case Format::RGB32UI: return GL_UNSIGNED_INT;
case Format::RGB32I: return GL_INT;
case Format::RGBA8:
case Format::SRGB8_A8: return GL_UNSIGNED_BYTE;
case Format::RGBA8SN: return GL_BYTE;
case Format::RGBA8UI: return GL_UNSIGNED_BYTE;
case Format::RGBA8I: return GL_BYTE;
case Format::RGBA16F: return GL_HALF_FLOAT;
case Format::RGBA16UI: return GL_UNSIGNED_SHORT;
case Format::RGBA16I: return GL_SHORT;
case Format::RGBA32F: return GL_FLOAT;
case Format::RGBA32UI: return GL_UNSIGNED_INT;
case Format::RGBA32I: return GL_INT;
case Format::R5G6B5: return GL_UNSIGNED_SHORT_5_6_5;
case Format::R11G11B10F: return GL_UNSIGNED_INT_10F_11F_11F_REV;
case Format::RGB5A1: return GL_UNSIGNED_SHORT_5_5_5_1;
case Format::RGBA4: return GL_UNSIGNED_SHORT_4_4_4_4;
case Format::RGB10A2:
case Format::RGB10A2UI: return GL_UNSIGNED_INT_2_10_10_10_REV;
case Format::RGB9E5: return GL_UNSIGNED_INT_5_9_9_9_REV;
case Format::DEPTH: return GL_FLOAT;
case Format::DEPTH_STENCIL: return GL_UNSIGNED_INT_24_8;
case Format::BC1:
case Format::BC1_SRGB:
case Format::BC2:
case Format::BC2_SRGB:
case Format::BC3:
case Format::BC3_SRGB:
case Format::BC4: return GL_UNSIGNED_BYTE;
case Format::BC4_SNORM: return GL_BYTE;
case Format::BC5: return GL_UNSIGNED_BYTE;
case Format::BC5_SNORM: return GL_BYTE;
case Format::BC6H_SF16:
case Format::BC6H_UF16: return GL_FLOAT;
case Format::BC7:
case Format::BC7_SRGB:
case Format::ETC_RGB8:
case Format::ETC2_RGBA8:
case Format::ETC2_RGB8:
case Format::ETC2_SRGB8:
case Format::ETC2_RGB8_A1:
case Format::ETC2_SRGB8_A1:
case Format::EAC_R11: return GL_UNSIGNED_BYTE;
case Format::EAC_R11SN: return GL_BYTE;
case Format::EAC_RG11: return GL_UNSIGNED_BYTE;
case Format::EAC_RG11SN: return GL_BYTE;
case Format::PVRTC_RGB2:
case Format::PVRTC_RGBA2:
case Format::PVRTC_RGB4:
case Format::PVRTC_RGBA4:
case Format::PVRTC2_2BPP:
case Format::PVRTC2_4BPP:
case Format::ASTC_RGBA_4X4:
case Format::ASTC_RGBA_5X4:
case Format::ASTC_RGBA_5X5:
case Format::ASTC_RGBA_6X5:
case Format::ASTC_RGBA_6X6:
case Format::ASTC_RGBA_8X5:
case Format::ASTC_RGBA_8X6:
case Format::ASTC_RGBA_8X8:
case Format::ASTC_RGBA_10X5:
case Format::ASTC_RGBA_10X6:
case Format::ASTC_RGBA_10X8:
case Format::ASTC_RGBA_10X10:
case Format::ASTC_RGBA_12X10:
case Format::ASTC_RGBA_12X12:
case Format::ASTC_SRGBA_4X4:
case Format::ASTC_SRGBA_5X4:
case Format::ASTC_SRGBA_5X5:
case Format::ASTC_SRGBA_6X5:
case Format::ASTC_SRGBA_6X6:
case Format::ASTC_SRGBA_8X5:
case Format::ASTC_SRGBA_8X6:
case Format::ASTC_SRGBA_8X8:
case Format::ASTC_SRGBA_10X5:
case Format::ASTC_SRGBA_10X6:
case Format::ASTC_SRGBA_10X8:
case Format::ASTC_SRGBA_10X10:
case Format::ASTC_SRGBA_12X10:
case Format::ASTC_SRGBA_12X12:
return GL_UNSIGNED_BYTE;
default: {
CC_ABORT();
return GL_NONE;
}
}
}
GLenum getTextureViewTarget(GLES3GPUTextureView *view) {
switch (view->type) {
case TextureType::TEX2D_ARRAY: return GL_TEXTURE_2D_ARRAY;
case TextureType::TEX3D: return GL_TEXTURE_3D;
case TextureType::CUBE: return GL_TEXTURE_CUBE_MAP;
case TextureType::TEX2D: {
if (hasFlag(view->gpuTexture->flags, TextureFlagBit::EXTERNAL_OES)) {
return GL_TEXTURE_EXTERNAL_OES;
}
return view->gpuTexture->type == TextureType::CUBE ? GL_TEXTURE_CUBE_MAP_POSITIVE_X + view->baseLayer : GL_TEXTURE_2D;
}
default:
CC_ABORT();
return GL_NONE;
}
}
uint32_t glComponentCount(GLenum glType) {
switch (glType) {
case GL_FLOAT_MAT2:
case GL_FLOAT_MAT2x3:
case GL_FLOAT_MAT2x4: return 2;
case GL_FLOAT_MAT3x2:
case GL_FLOAT_MAT3:
case GL_FLOAT_MAT3x4: return 3;
case GL_FLOAT_MAT4x2:
case GL_FLOAT_MAT4x3:
case GL_FLOAT_MAT4: return 4;
default: {
return 1;
}
}
}
const GLenum GLES3_WRAPS[] = {
GL_REPEAT,
GL_MIRRORED_REPEAT,
GL_CLAMP_TO_EDGE,
GL_CLAMP_TO_EDGE,
};
const GLenum GLES3_CMP_FUNCS[] = {
GL_NEVER,
GL_LESS,
GL_EQUAL,
GL_LEQUAL,
GL_GREATER,
GL_NOTEQUAL,
GL_GEQUAL,
GL_ALWAYS,
};
const GLenum GLES3_STENCIL_OPS[] = {
GL_ZERO,
GL_KEEP,
GL_REPLACE,
GL_INCR,
GL_DECR,
GL_INVERT,
GL_INCR_WRAP,
GL_DECR_WRAP,
};
const GLenum GLES3_BLEND_OPS[] = {
GL_FUNC_ADD,
GL_FUNC_SUBTRACT,
GL_FUNC_REVERSE_SUBTRACT,
GL_MIN,
GL_MAX,
};
const GLenum GLES3_BLEND_FACTORS[] = {
GL_ZERO,
GL_ONE,
GL_SRC_ALPHA,
GL_DST_ALPHA,
GL_ONE_MINUS_SRC_ALPHA,
GL_ONE_MINUS_DST_ALPHA,
GL_SRC_COLOR,
GL_DST_COLOR,
GL_ONE_MINUS_SRC_COLOR,
GL_ONE_MINUS_DST_COLOR,
GL_SRC_ALPHA_SATURATE,
GL_CONSTANT_COLOR,
GL_ONE_MINUS_CONSTANT_COLOR,
GL_CONSTANT_ALPHA,
GL_ONE_MINUS_CONSTANT_ALPHA,
};
const GLenum GLES3_FILTERS[] = {
GL_NONE,
GL_NEAREST,
GL_LINEAR,
GL_NONE,
};
const GLenum GL_MEMORY_ACCESS[] = {
GL_READ_ONLY,
GL_WRITE_ONLY,
GL_READ_WRITE,
};
} // namespace
void cmdFuncGLES3CreateBuffer(GLES3Device *device, GLES3GPUBuffer *gpuBuffer) {
GLenum glUsage = hasFlag(gpuBuffer->memUsage, MemoryUsageBit::HOST) ? GL_DYNAMIC_DRAW : GL_STATIC_DRAW;
GLES3ObjectCache &gfxStateCache = device->stateCache()->gfxStateCache;
if (hasFlag(gpuBuffer->usage, BufferUsageBit::VERTEX)) {
gpuBuffer->glTarget = GL_ARRAY_BUFFER;
GL_CHECK(glGenBuffers(1, &gpuBuffer->glBuffer));
if (gpuBuffer->size) {
if (USE_VAO) {
if (device->stateCache()->glVAO) {
GL_CHECK(glBindVertexArray(0));
device->stateCache()->glVAO = 0;
}
}
gfxStateCache.gpuInputAssembler = nullptr;
if (device->stateCache()->glArrayBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, gpuBuffer->glBuffer));
}
GL_CHECK(glBufferData(GL_ARRAY_BUFFER, gpuBuffer->size, nullptr, glUsage));
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, 0));
device->stateCache()->glArrayBuffer = 0;
}
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::INDEX)) {
gpuBuffer->glTarget = GL_ELEMENT_ARRAY_BUFFER;
GL_CHECK(glGenBuffers(1, &gpuBuffer->glBuffer));
if (gpuBuffer->size) {
if (USE_VAO) {
if (device->stateCache()->glVAO) {
GL_CHECK(glBindVertexArray(0));
device->stateCache()->glVAO = 0;
}
}
gfxStateCache.gpuInputAssembler = nullptr;
if (device->stateCache()->glElementArrayBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, gpuBuffer->glBuffer));
}
GL_CHECK(glBufferData(GL_ELEMENT_ARRAY_BUFFER, gpuBuffer->size, nullptr, glUsage));
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0));
device->stateCache()->glElementArrayBuffer = 0;
}
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::UNIFORM)) {
gpuBuffer->glTarget = GL_UNIFORM_BUFFER;
GL_CHECK(glGenBuffers(1, &gpuBuffer->glBuffer));
if (gpuBuffer->size) {
if (device->stateCache()->glUniformBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_UNIFORM_BUFFER, gpuBuffer->glBuffer));
}
GL_CHECK(glBufferData(GL_UNIFORM_BUFFER, gpuBuffer->size, nullptr, glUsage));
GL_CHECK(glBindBuffer(GL_UNIFORM_BUFFER, 0));
device->stateCache()->glUniformBuffer = 0;
}
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::STORAGE)) {
gpuBuffer->glTarget = GL_SHADER_STORAGE_BUFFER;
GL_CHECK(glGenBuffers(1, &gpuBuffer->glBuffer));
if (gpuBuffer->size) {
if (device->stateCache()->glShaderStorageBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_SHADER_STORAGE_BUFFER, gpuBuffer->glBuffer));
}
GL_CHECK(glBufferData(GL_SHADER_STORAGE_BUFFER, gpuBuffer->size, nullptr, glUsage));
GL_CHECK(glBindBuffer(GL_SHADER_STORAGE_BUFFER, 0));
device->stateCache()->glShaderStorageBuffer = 0;
}
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::INDIRECT)) {
gpuBuffer->glTarget = GL_NONE;
} else if ((hasFlag(gpuBuffer->usage, BufferUsageBit::TRANSFER_DST)) ||
(hasFlag(gpuBuffer->usage, BufferUsageBit::TRANSFER_SRC))) {
gpuBuffer->buffer = static_cast<uint8_t *>(CC_MALLOC(gpuBuffer->size));
gpuBuffer->glTarget = GL_NONE;
} else {
CC_ABORT();
gpuBuffer->glTarget = GL_NONE;
}
}
void cmdFuncGLES3DestroyBuffer(GLES3Device *device, GLES3GPUBuffer *gpuBuffer) {
GLES3ObjectCache &gfxStateCache = device->stateCache()->gfxStateCache;
if (gpuBuffer->glBuffer) {
if (hasFlag(gpuBuffer->usage, BufferUsageBit::VERTEX)) {
if (USE_VAO) {
if (device->stateCache()->glVAO) {
GL_CHECK(glBindVertexArray(0));
device->stateCache()->glVAO = 0;
}
}
gfxStateCache.gpuInputAssembler = nullptr;
if (device->stateCache()->glArrayBuffer == gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, 0));
device->stateCache()->glArrayBuffer = 0;
}
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::INDEX)) {
if (USE_VAO) {
if (device->stateCache()->glVAO) {
GL_CHECK(glBindVertexArray(0));
device->stateCache()->glVAO = 0;
}
}
gfxStateCache.gpuInputAssembler = nullptr;
if (device->stateCache()->glElementArrayBuffer == gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0));
device->stateCache()->glElementArrayBuffer = 0;
}
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::UNIFORM)) {
ccstd::vector<GLuint> &ubo = device->stateCache()->glBindUBOs;
for (GLuint i = 0; i < ubo.size(); i++) {
if (ubo[i] == gpuBuffer->glBuffer) {
GL_CHECK(glBindBufferBase(GL_UNIFORM_BUFFER, i, 0));
device->stateCache()->glUniformBuffer = 0;
ubo[i] = 0;
}
}
if (device->stateCache()->glUniformBuffer == gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_UNIFORM_BUFFER, 0));
device->stateCache()->glUniformBuffer = 0;
}
}
if (hasFlag(gpuBuffer->usage, BufferUsageBit::STORAGE)) {
ccstd::vector<GLuint> &ssbo = device->stateCache()->glBindSSBOs;
for (GLuint i = 0; i < ssbo.size(); i++) {
if (ssbo[i] == gpuBuffer->glBuffer) {
GL_CHECK(glBindBufferBase(GL_SHADER_STORAGE_BUFFER, i, 0));
device->stateCache()->glShaderStorageBuffer = 0;
ssbo[i] = 0;
}
}
if (device->stateCache()->glShaderStorageBuffer == gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_SHADER_STORAGE_BUFFER, 0));
device->stateCache()->glShaderStorageBuffer = 0;
}
}
GL_CHECK(glDeleteBuffers(1, &gpuBuffer->glBuffer));
gpuBuffer->glBuffer = 0;
}
CC_SAFE_FREE(gpuBuffer->buffer)
}
void cmdFuncGLES3ResizeBuffer(GLES3Device *device, GLES3GPUBuffer *gpuBuffer) {
GLES3ObjectCache &gfxStateCache = device->stateCache()->gfxStateCache;
GLenum glUsage = (hasFlag(gpuBuffer->memUsage, MemoryUsageBit::HOST) ? GL_DYNAMIC_DRAW : GL_STATIC_DRAW);
if (hasFlag(gpuBuffer->usage, BufferUsageBit::VERTEX)) {
gpuBuffer->glTarget = GL_ARRAY_BUFFER;
if (gpuBuffer->size) {
if (USE_VAO) {
if (device->stateCache()->glVAO) {
GL_CHECK(glBindVertexArray(0));
device->stateCache()->glVAO = 0;
}
}
gfxStateCache.gpuInputAssembler = nullptr;
if (device->stateCache()->glArrayBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, gpuBuffer->glBuffer));
}
GL_CHECK(glBufferData(GL_ARRAY_BUFFER, gpuBuffer->size, nullptr, glUsage));
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, 0));
device->stateCache()->glArrayBuffer = 0;
}
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::INDEX)) {
gpuBuffer->glTarget = GL_ELEMENT_ARRAY_BUFFER;
if (gpuBuffer->size) {
if (USE_VAO) {
if (device->stateCache()->glVAO) {
GL_CHECK(glBindVertexArray(0));
device->stateCache()->glVAO = 0;
}
}
gfxStateCache.gpuInputAssembler = nullptr;
if (device->stateCache()->glElementArrayBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, gpuBuffer->glBuffer));
}
GL_CHECK(glBufferData(GL_ELEMENT_ARRAY_BUFFER, gpuBuffer->size, nullptr, glUsage));
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0));
device->stateCache()->glElementArrayBuffer = 0;
}
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::UNIFORM)) {
gpuBuffer->glTarget = GL_UNIFORM_BUFFER;
if (gpuBuffer->size) {
if (device->stateCache()->glUniformBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_UNIFORM_BUFFER, gpuBuffer->glBuffer));
}
GL_CHECK(glBufferData(GL_UNIFORM_BUFFER, gpuBuffer->size, nullptr, glUsage));
GL_CHECK(glBindBuffer(GL_UNIFORM_BUFFER, 0));
device->stateCache()->glUniformBuffer = 0;
}
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::INDIRECT)) {
gpuBuffer->indirects.resize(gpuBuffer->count);
gpuBuffer->glTarget = GL_NONE;
} else if ((hasFlag(gpuBuffer->usage, BufferUsageBit::TRANSFER_DST)) ||
(hasFlag(gpuBuffer->usage, BufferUsageBit::TRANSFER_SRC))) {
if (gpuBuffer->buffer) {
CC_FREE(gpuBuffer->buffer);
}
gpuBuffer->buffer = static_cast<uint8_t *>(CC_MALLOC(gpuBuffer->size));
gpuBuffer->glTarget = GL_NONE;
} else {
CC_ABORT();
gpuBuffer->glTarget = GL_NONE;
}
}
static void renderBufferStorage(GLES3Device *device, GLES3GPUTexture *gpuTexture) {
CC_ASSERT(gpuTexture->type == TextureType::TEX2D);
GLuint &glRenderbuffer = device->stateCache()->glRenderbuffer;
if (gpuTexture->glRenderbuffer != glRenderbuffer) {
GL_CHECK(glBindRenderbuffer(GL_RENDERBUFFER, gpuTexture->glRenderbuffer));
glRenderbuffer = gpuTexture->glRenderbuffer;
}
if (gpuTexture->glSamples > 1) {
GL_CHECK(glRenderbufferStorageMultisampleEXT(GL_RENDERBUFFER, gpuTexture->glSamples, gpuTexture->glInternalFmt, gpuTexture->width, gpuTexture->height));
} else {
GL_CHECK(glRenderbufferStorage(GL_RENDERBUFFER, gpuTexture->glInternalFmt, gpuTexture->width, gpuTexture->height));
}
}
static void textureStorage(GLES3Device *device, GLES3GPUTexture *gpuTexture) {
GLuint &glTexture = device->stateCache()->glTextures[device->stateCache()->texUint];
if (gpuTexture->glTexture != glTexture) {
glTexture = gpuTexture->glTexture;
}
uint32_t w = gpuTexture->width;
uint32_t h = gpuTexture->height;
uint32_t d = gpuTexture->type == TextureType::TEX2D_ARRAY ? gpuTexture->arrayLayer : gpuTexture->depth;
switch (gpuTexture->type) {
case TextureType::CUBE:
CC_ASSERT(gpuTexture->glSamples <= 1);
GL_CHECK(glBindTexture(GL_TEXTURE_CUBE_MAP, gpuTexture->glTexture));
GL_CHECK(glTexStorage2D(GL_TEXTURE_CUBE_MAP, gpuTexture->mipLevel, gpuTexture->glInternalFmt, w, h));
break;
case TextureType::TEX2D:
if (gpuTexture->glSamples > 1 && !gpuTexture->immutable && device->constantRegistry()->glMinorVersion >= 1) {
GL_CHECK(glBindTexture(GL_TEXTURE_2D_MULTISAMPLE, gpuTexture->glTexture));
GL_CHECK(glTexStorage2DMultisample(GL_TEXTURE_2D_MULTISAMPLE, gpuTexture->glSamples, gpuTexture->glInternalFmt, w, h, GL_FALSE));
} else {
auto target = hasFlag(gpuTexture->flags, TextureFlagBit::EXTERNAL_OES) ? GL_TEXTURE_EXTERNAL_OES : GL_TEXTURE_2D;
GL_CHECK(glBindTexture(target, gpuTexture->glTexture));
if (hasFlag(gpuTexture->flags, TextureFlagBit::MUTABLE_STORAGE)) {
GL_CHECK(glTexImage2D(GL_TEXTURE_2D, gpuTexture->mipLevel, gpuTexture->glInternalFmt, w, h, 0,
gpuTexture->glFormat, gpuTexture->glType, nullptr));
} else {
GL_CHECK(glTexStorage2D(GL_TEXTURE_2D, gpuTexture->mipLevel, gpuTexture->glInternalFmt, w, h));
}
}
break;
case TextureType::TEX3D:
CC_ASSERT(gpuTexture->glSamples <= 1);
GL_CHECK(glBindTexture(GL_TEXTURE_3D, gpuTexture->glTexture));
GL_CHECK(glTexStorage3D(GL_TEXTURE_3D, gpuTexture->mipLevel, gpuTexture->glInternalFmt, w, h, d));
break;
case TextureType::TEX2D_ARRAY:
if (gpuTexture->glSamples > 1 && !gpuTexture->immutable && device->constantRegistry()->glMinorVersion >= 1) {
GL_CHECK(glBindTexture(GL_TEXTURE_2D_MULTISAMPLE_ARRAY, gpuTexture->glTexture));
GL_CHECK(glTexStorage3DMultisample(GL_TEXTURE_2D_MULTISAMPLE_ARRAY, gpuTexture->glSamples, gpuTexture->glInternalFmt, w, h, d, GL_FALSE));
} else {
GL_CHECK(glBindTexture(GL_TEXTURE_2D_ARRAY, gpuTexture->glTexture));
GL_CHECK(glTexStorage3D(GL_TEXTURE_2D_ARRAY, gpuTexture->mipLevel, gpuTexture->glInternalFmt, w, h, d));
}
break;
default:
break;
}
}
static bool useRenderBuffer(const GLES3Device *device, Format format, TextureUsage usage) {
return !device->isTextureExclusive(format) &&
hasAllFlags(TextureUsage::COLOR_ATTACHMENT | TextureUsage::DEPTH_STENCIL_ATTACHMENT, usage);
}
void cmdFuncGLES3CreateTexture(GLES3Device *device, GLES3GPUTexture *gpuTexture) {
gpuTexture->glInternalFmt = mapGLInternalFormat(gpuTexture->format);
gpuTexture->glFormat = mapGLFormat(gpuTexture->format);
gpuTexture->glType = formatToGLType(gpuTexture->format);
bool supportRenderBufferMS = device->constantRegistry()->mMSRT > MSRTSupportLevel::NONE;
gpuTexture->useRenderBuffer = useRenderBuffer(device, gpuTexture->format, gpuTexture->usage) &&
(gpuTexture->glSamples <= 1 || supportRenderBufferMS);
if (gpuTexture->glSamples > 1) {
// Allocate render buffer when binding a framebuffer if the MSRT extension is not present.
if (gpuTexture->useRenderBuffer &&
hasFlag(gpuTexture->flags, TextureFlagBit::LAZILY_ALLOCATED)) {
gpuTexture->memoryAllocated = false;
return;
}
}
if (gpuTexture->glTexture || gpuTexture->size == 0) {
return;
}
if (gpuTexture->useRenderBuffer) {
GL_CHECK(glGenRenderbuffers(1, &gpuTexture->glRenderbuffer));
renderBufferStorage(device, gpuTexture);
} else {
GL_CHECK(glGenTextures(1, &gpuTexture->glTexture));
textureStorage(device, gpuTexture);
}
}
void cmdFuncGLES3DestroyTexture(GLES3Device *device, GLES3GPUTexture *gpuTexture) {
device->framebufferCacheMap()->onTextureDestroy(gpuTexture);
if (gpuTexture->glTexture) {
for (GLuint &glTexture : device->stateCache()->glTextures) {
if (glTexture == gpuTexture->glTexture) {
glTexture = 0;
}
}
if (!hasFlag(gpuTexture->flags, TextureFlagBit::EXTERNAL_OES) && !hasFlag(gpuTexture->flags, TextureFlagBit::EXTERNAL_NORMAL)) {
GL_CHECK(glDeleteTextures(1, &gpuTexture->glTexture));
}
gpuTexture->glTexture = 0;
} else if (gpuTexture->glRenderbuffer) {
GLuint &glRenderbuffer = device->stateCache()->glRenderbuffer;
if (gpuTexture->glRenderbuffer == glRenderbuffer) {
GL_CHECK(glBindRenderbuffer(GL_RENDERBUFFER, 0));
glRenderbuffer = 0;
}
GL_CHECK(glDeleteRenderbuffers(1, &gpuTexture->glRenderbuffer));
gpuTexture->glRenderbuffer = 0;
}
}
void cmdFuncGLES3ResizeTexture(GLES3Device *device, GLES3GPUTexture *gpuTexture) {
if (!gpuTexture->memoryAllocated ||
hasFlag(gpuTexture->flags, TextureFlagBit::EXTERNAL_OES) ||
hasFlag(gpuTexture->flags, TextureFlagBit::EXTERNAL_NORMAL)) {
return;
}
if (!gpuTexture->useRenderBuffer) {
// immutable by default
cmdFuncGLES3DestroyTexture(device, gpuTexture);
cmdFuncGLES3CreateTexture(device, gpuTexture);
} else if (gpuTexture->size > 0){
renderBufferStorage(device, gpuTexture);
}
}
void cmdFuncGLES3CreateTextureView(GLES3Device */*device*/, GLES3GPUTextureView *gpuTextureView) {
gpuTextureView->glTarget = getTextureViewTarget(gpuTextureView);
}
void cmdFuncGLES3PrepareSamplerInfo(GLES3Device * /*device*/, GLES3GPUSampler *gpuSampler) {
if (gpuSampler->minFilter == Filter::LINEAR || gpuSampler->minFilter == Filter::ANISOTROPIC) {
if (gpuSampler->mipFilter == Filter::LINEAR || gpuSampler->mipFilter == Filter::ANISOTROPIC) {
gpuSampler->glMinFilter = GL_LINEAR_MIPMAP_LINEAR;
} else if (gpuSampler->mipFilter == Filter::POINT) {
gpuSampler->glMinFilter = GL_LINEAR_MIPMAP_NEAREST;
} else {
gpuSampler->glMinFilter = GL_LINEAR;
}
} else {
if (gpuSampler->mipFilter == Filter::LINEAR || gpuSampler->mipFilter == Filter::ANISOTROPIC) {
gpuSampler->glMinFilter = GL_NEAREST_MIPMAP_LINEAR;
} else if (gpuSampler->mipFilter == Filter::POINT) {
gpuSampler->glMinFilter = GL_NEAREST_MIPMAP_NEAREST;
} else {
gpuSampler->glMinFilter = GL_NEAREST;
}
}
if (gpuSampler->magFilter == Filter::LINEAR || gpuSampler->magFilter == Filter::ANISOTROPIC) {
gpuSampler->glMagFilter = GL_LINEAR;
} else {
gpuSampler->glMagFilter = GL_NEAREST;
}
gpuSampler->glWrapS = GLES3_WRAPS[toNumber(gpuSampler->addressU)];
gpuSampler->glWrapT = GLES3_WRAPS[toNumber(gpuSampler->addressV)];
gpuSampler->glWrapR = GLES3_WRAPS[toNumber(gpuSampler->addressW)];
}
GLuint GLES3GPUSampler::getGLSampler(uint16_t minLod, uint16_t maxLod) {
uint32_t hash = (minLod << 16) + maxLod;
if (!_cache.count(hash)) {
GLuint glSampler{0U};
GL_CHECK(glGenSamplers(1, &glSampler));
GL_CHECK(glSamplerParameteri(glSampler, GL_TEXTURE_MIN_FILTER, glMinFilter));
GL_CHECK(glSamplerParameteri(glSampler, GL_TEXTURE_MAG_FILTER, glMagFilter));
GL_CHECK(glSamplerParameteri(glSampler, GL_TEXTURE_WRAP_S, glWrapS));
GL_CHECK(glSamplerParameteri(glSampler, GL_TEXTURE_WRAP_T, glWrapT));
GL_CHECK(glSamplerParameteri(glSampler, GL_TEXTURE_WRAP_R, glWrapR));
GL_CHECK(glSamplerParameterf(glSampler, GL_TEXTURE_MIN_LOD, static_cast<GLfloat>(minLod)));
GL_CHECK(glSamplerParameterf(glSampler, GL_TEXTURE_MAX_LOD, static_cast<GLfloat>(maxLod)));
_cache[hash] = glSampler;
}
return _cache[hash];
}
bool cmdFuncGLES3CreateProgramByBinary(GLES3Device *device, GLES3GPUShader *gpuShader, GLES3GPUPipelineLayout *pipelineLayout) {
auto *pipelineCache = device->pipelineCache();
if (pipelineCache == nullptr || gpuShader->hash == INVALID_SHADER_HASH) {
return false;
}
ccstd::hash_t hash = gpuShader->hash;
ccstd::hash_combine(hash, pipelineLayout->hash);
auto *item = pipelineCache->fetchBinary(gpuShader->name, hash);
if (item != nullptr) {
GL_CHECK(gpuShader->glProgram = glCreateProgram());
GL_CHECK(glProgramBinary(gpuShader->glProgram, item->format, item->data.data(), item->data.size()));
return true;
}
return false;
}
bool cmdFuncGLES3CreateProgramBySource(GLES3Device *device, GLES3GPUShader *gpuShader, GLES3GPUPipelineLayout *pipelineLayout) {
GLenum glShaderStage = 0;
ccstd::string shaderStageStr;
GLint status;
for (size_t i = 0; i < gpuShader->gpuStages.size(); ++i) {
GLES3GPUShaderStage &gpuStage = gpuShader->gpuStages[i];
switch (gpuStage.type) {
case ShaderStageFlagBit::VERTEX: {
glShaderStage = GL_VERTEX_SHADER;
shaderStageStr = "Vertex Shader";
break;
}
case ShaderStageFlagBit::FRAGMENT: {
glShaderStage = GL_FRAGMENT_SHADER;
shaderStageStr = "Fragment Shader";
break;
}
case ShaderStageFlagBit::COMPUTE: {
glShaderStage = GL_COMPUTE_SHADER;
shaderStageStr = "Compute Shader";
break;
}
default: {
CC_ABORT();
return false;
}
}
GL_CHECK(gpuStage.glShader = glCreateShader(glShaderStage));
uint32_t version = device->constantRegistry()->glMinorVersion ? 310 : 300;
ccstd::string shaderSource = StringUtil::format("#version %u es\n", version) + gpuStage.source;
const char *source = shaderSource.c_str();
GL_CHECK(glShaderSource(gpuStage.glShader, 1, (const GLchar **)&source, nullptr));
GL_CHECK(glCompileShader(gpuStage.glShader));
GL_CHECK(glGetShaderiv(gpuStage.glShader, GL_COMPILE_STATUS, &status));
if (status != GL_TRUE) {
GLint logSize = 0;
GL_CHECK(glGetShaderiv(gpuStage.glShader, GL_INFO_LOG_LENGTH, &logSize));
++logSize;
auto *logs = static_cast<GLchar *>(CC_MALLOC(logSize));
GL_CHECK(glGetShaderInfoLog(gpuStage.glShader, logSize, nullptr, logs));
CC_LOG_ERROR("%s in %s compilation failed.", shaderStageStr.c_str(), gpuShader->name.c_str());
CC_LOG_ERROR(logs);
CC_FREE(logs);
GL_CHECK(glDeleteShader(gpuStage.glShader));
gpuStage.glShader = 0;
return false;
}
}
GL_CHECK(gpuShader->glProgram = glCreateProgram());
// link program
for (size_t i = 0; i < gpuShader->gpuStages.size(); ++i) {
GLES3GPUShaderStage &gpuStage = gpuShader->gpuStages[i];
GL_CHECK(glAttachShader(gpuShader->glProgram, gpuStage.glShader));
}
GL_CHECK(glLinkProgram(gpuShader->glProgram));
// detach & delete immediately
for (size_t i = 0; i < gpuShader->gpuStages.size(); ++i) {
GLES3GPUShaderStage &gpuStage = gpuShader->gpuStages[i];
if (gpuStage.glShader) {
GL_CHECK(glDetachShader(gpuShader->glProgram, gpuStage.glShader));
GL_CHECK(glDeleteShader(gpuStage.glShader));
gpuStage.glShader = 0;
}
}
GL_CHECK(glGetProgramiv(gpuShader->glProgram, GL_LINK_STATUS, &status));
if (status != 1) {
CC_LOG_ERROR("Failed to link Shader [%s].", gpuShader->name.c_str());
GLint logSize = 0;
GL_CHECK(glGetProgramiv(gpuShader->glProgram, GL_INFO_LOG_LENGTH, &logSize));
if (logSize) {
++logSize;
auto *logs = static_cast<GLchar *>(CC_MALLOC(logSize));
GL_CHECK(glGetProgramInfoLog(gpuShader->glProgram, logSize, nullptr, logs));
CC_LOG_ERROR(logs);
CC_FREE(logs);
return false;
}
}
auto *cache = device->pipelineCache();
if (cache != nullptr && gpuShader->hash != INVALID_SHADER_HASH) {
GLint binaryLength = 0;
GL_CHECK(glGetProgramiv(gpuShader->glProgram, GL_PROGRAM_BINARY_LENGTH, &binaryLength));
GLsizei length = 0;
IntrusivePtr<GLES3GPUProgramBinary> binary = ccnew GLES3GPUProgramBinary();
binary->name = gpuShader->name;
binary->hash = gpuShader->hash;
ccstd::hash_combine(binary->hash, pipelineLayout->hash);
binary->data.resize(binaryLength);
GL_CHECK(glGetProgramBinary(gpuShader->glProgram, binaryLength, &length, &binary->format, binary->data.data()));
cache->addBinary(binary);
}
return true;
}
// NOLINTNEXTLINE(google-readability-function-size, readability-function-size)
void cmdFuncGLES3CreateShader(GLES3Device *device, GLES3GPUShader *gpuShader, GLES3GPUPipelineLayout *pipelineLayout) {
if (!cmdFuncGLES3CreateProgramByBinary(device, gpuShader, pipelineLayout)) {
if (!cmdFuncGLES3CreateProgramBySource(device, gpuShader, pipelineLayout)) {
return;
}
}
CC_LOG_INFO("Shader '%s' compilation succeeded.", gpuShader->name.c_str());
GLint attrMaxLength = 0;
GLint attrCount = 0;
GL_CHECK(glGetProgramiv(gpuShader->glProgram, GL_ACTIVE_ATTRIBUTE_MAX_LENGTH, &attrMaxLength));
GL_CHECK(glGetProgramiv(gpuShader->glProgram, GL_ACTIVE_ATTRIBUTES, &attrCount));
GLchar glName[256];
GLsizei glLength;
GLsizei glSize;
GLenum glType;
gpuShader->glInputs.resize(attrCount);
for (GLint i = 0; i < attrCount; ++i) {
GLES3GPUInput &gpuInput = gpuShader->glInputs[i];
memset(glName, 0, sizeof(glName));
GL_CHECK(glGetActiveAttrib(gpuShader->glProgram, i, attrMaxLength, &glLength, &glSize, &glType, glName));
char *offset = strchr(glName, '[');
if (offset) {
glName[offset - glName] = '\0';
}
GL_CHECK(gpuInput.glLoc = glGetAttribLocation(gpuShader->glProgram, glName));
gpuInput.binding = gpuInput.glLoc;
gpuInput.name = glName;
gpuInput.type = mapType(glType);
gpuInput.count = glSize;
gpuInput.size = gpuInput.stride * gpuInput.count;
gpuInput.glType = glType;
}
// create buffers
GLint blockCount;
GL_CHECK(glGetProgramiv(gpuShader->glProgram, GL_ACTIVE_UNIFORM_BLOCKS, &blockCount));
GLint bufferCount = 0;
if (device->constantRegistry()->glMinorVersion > 0) {
GL_CHECK(glGetProgramInterfaceiv(gpuShader->glProgram, GL_SHADER_STORAGE_BLOCK, GL_ACTIVE_RESOURCES, &bufferCount));
}
gpuShader->glBuffers.resize(blockCount + bufferCount);
GLuint uboBinding = 0;
for (GLint i = 0; i < blockCount; ++i) {
GLES3GPUUniformBuffer &glBlock = gpuShader->glBuffers[i];
memset(glName, 0, sizeof(glName));
GL_CHECK(glGetActiveUniformBlockName(gpuShader->glProgram, i, 255, &glLength, glName));
char *offset = strchr(glName, '[');
if (offset) {
glName[offset - glName] = '\0';
}
glBlock.name = glName;
glBlock.set = INVALID_BINDING;
glBlock.binding = INVALID_BINDING;
for (size_t b = 0; b < gpuShader->blocks.size(); ++b) {
UniformBlock &block = gpuShader->blocks[b];
if (block.name == glBlock.name) {
glBlock.set = block.set;
glBlock.binding = block.binding;
glBlock.glBinding = uboBinding;
GL_CHECK(glUniformBlockBinding(gpuShader->glProgram, i, uboBinding));
uboBinding += block.count;
break;
}
}
}
for (GLint i = 0; i < bufferCount; ++i) {
GLES3GPUUniformBuffer &glBuffer = gpuShader->glBuffers[blockCount + i];
memset(glName, 0, sizeof(glName));
GL_CHECK(glGetProgramResourceName(gpuShader->glProgram, GL_SHADER_STORAGE_BLOCK, i, 255, &glLength, glName));
char *offset = strchr(glName, '[');
if (offset) {
glName[offset - glName] = '\0';
}
glBuffer.name = glName;
glBuffer.set = INVALID_BINDING;
glBuffer.binding = INVALID_BINDING;
glBuffer.isStorage = true;
for (size_t b = 0; b < gpuShader->buffers.size(); ++b) {
UniformStorageBuffer &buffer = gpuShader->buffers[b];
if (buffer.name == glBuffer.name) {
glBuffer.set = buffer.set;
glBuffer.binding = buffer.binding;
static GLenum prop = GL_BUFFER_BINDING;
// SSBO bindings should already be specified in shader layout qualifiers
GL_CHECK(glGetProgramResourceiv(gpuShader->glProgram, GL_SHADER_STORAGE_BLOCK, i, 1, &prop, 1, nullptr, (GLint *)&glBuffer.glBinding));
break;
}
}
}
// fallback subpassInputs into samplerTextures if not using FBF
if (device->constantRegistry()->mFBF == FBFSupportLevel::NONE) {
for (const auto &subpassInput : gpuShader->subpassInputs) {
gpuShader->samplerTextures.emplace_back();
auto &samplerTexture = gpuShader->samplerTextures.back();
samplerTexture.name = subpassInput.name;
samplerTexture.set = subpassInput.set;
samplerTexture.binding = subpassInput.binding;
samplerTexture.count = subpassInput.count;
samplerTexture.type = Type::SAMPLER2D;
}
}
// create uniform sampler textures
if (!gpuShader->samplerTextures.empty()) {
gpuShader->glSamplerTextures.resize(gpuShader->samplerTextures.size());
for (size_t i = 0; i < gpuShader->glSamplerTextures.size(); ++i) {
UniformSamplerTexture &samplerTexture = gpuShader->samplerTextures[i];
GLES3GPUUniformSamplerTexture &gpuSamplerTexture = gpuShader->glSamplerTextures[i];
gpuSamplerTexture.set = samplerTexture.set;
gpuSamplerTexture.binding = samplerTexture.binding;
gpuSamplerTexture.name = samplerTexture.name;
gpuSamplerTexture.type = samplerTexture.type;
gpuSamplerTexture.count = samplerTexture.count;
gpuSamplerTexture.glType = mapGLType(gpuSamplerTexture.type);
gpuSamplerTexture.glLoc = -1;
}
}
// texture unit index mapping optimization
ccstd::vector<GLES3GPUUniformSamplerTexture> glActiveSamplerTextures;
ccstd::vector<GLint> glActiveSamplerLocations;
const GLESBindingMapping &bindingMappings = device->bindingMappings();
ccstd::unordered_map<ccstd::string, uint32_t> &texUnitCacheMap = device->stateCache()->texUnitCacheMap;
// sampler bindings in the flexible set comes strictly after buffer bindings
// so we need to subtract the buffer count for these samplers
uint32_t flexibleSetBaseOffset = 0U;
for (const auto &buffer : gpuShader->blocks) {
if (buffer.set == bindingMappings.flexibleSet) {
flexibleSetBaseOffset++;
}
}
uint32_t arrayOffset = 0U;
for (uint32_t i = 0U; i < gpuShader->samplerTextures.size(); i++) {
const UniformSamplerTexture &samplerTexture = gpuShader->samplerTextures[i];
GLint glLoc = -1;
GL_CHECK(glLoc = glGetUniformLocation(gpuShader->glProgram, samplerTexture.name.c_str()));
if (glLoc >= 0) {
glActiveSamplerTextures.push_back(gpuShader->glSamplerTextures[i]);
glActiveSamplerLocations.push_back(glLoc);
if (texUnitCacheMap.count(samplerTexture.name) == 0U) {
uint32_t binding = samplerTexture.binding + bindingMappings.samplerTextureOffsets[samplerTexture.set] + arrayOffset;
if (samplerTexture.set == bindingMappings.flexibleSet) binding -= flexibleSetBaseOffset;
texUnitCacheMap[samplerTexture.name] = binding % device->getCapabilities().maxTextureUnits;
arrayOffset += samplerTexture.count - 1;
}
}
}
if (!glActiveSamplerTextures.empty()) {
ccstd::vector<bool> usedTexUnits(device->getCapabilities().maxTextureUnits, false);
// try to reuse existing mappings first
for (uint32_t i = 0U; i < glActiveSamplerTextures.size(); i++) {
GLES3GPUUniformSamplerTexture &glSamplerTexture = glActiveSamplerTextures[i];
if (texUnitCacheMap.count(glSamplerTexture.name)) {
uint32_t cachedUnit = texUnitCacheMap[glSamplerTexture.name];
glSamplerTexture.glLoc = glActiveSamplerLocations[i];
for (uint32_t t = 0U; t < glSamplerTexture.count; t++) {
while (usedTexUnits[cachedUnit]) { // the shader already compiles so we should be safe to do this here
cachedUnit = (cachedUnit + 1) % device->getCapabilities().maxTextureUnits;
}
glSamplerTexture.units.push_back(static_cast<GLint>(cachedUnit));
usedTexUnits[cachedUnit] = true;
}
}
}
// fill in the rest sequencially
uint32_t unitIdx = 0U;
for (uint32_t i = 0U; i < glActiveSamplerTextures.size(); i++) {
GLES3GPUUniformSamplerTexture &glSamplerTexture = glActiveSamplerTextures[i];
if (glSamplerTexture.glLoc < 0) {
glSamplerTexture.glLoc = glActiveSamplerLocations[i];
for (uint32_t t = 0U; t < glSamplerTexture.count; t++) {
while (usedTexUnits[unitIdx]) {
unitIdx = (unitIdx + 1) % device->getCapabilities().maxTextureUnits;
}
if (!texUnitCacheMap.count(glSamplerTexture.name)) {
texUnitCacheMap[glSamplerTexture.name] = unitIdx;
}
glSamplerTexture.units.push_back(static_cast<GLint>(unitIdx));
usedTexUnits[unitIdx] = true;
}
}
}
if (device->stateCache()->glProgram != gpuShader->glProgram) {
GL_CHECK(glUseProgram(gpuShader->glProgram));
}
for (GLES3GPUUniformSamplerTexture &gpuSamplerTexture : glActiveSamplerTextures) {
GL_CHECK(glUniform1iv(gpuSamplerTexture.glLoc, static_cast<GLsizei>(gpuSamplerTexture.units.size()), gpuSamplerTexture.units.data()));
}
if (device->stateCache()->glProgram != gpuShader->glProgram) {
GL_CHECK(glUseProgram(device->stateCache()->glProgram));
}
}
gpuShader->glSamplerTextures = std::move(glActiveSamplerTextures);
// create uniform storage images
if (!gpuShader->images.empty()) {
for (size_t i = 0; i < gpuShader->images.size(); ++i) {
UniformStorageImage &image = gpuShader->images[i];
GLint glLoc = -1;
GL_CHECK(glLoc = glGetUniformLocation(gpuShader->glProgram, image.name.c_str()));
if (glLoc >= 0) {
GLES3GPUUniformStorageImage gpuImage;
gpuImage.set = image.set;
gpuImage.binding = image.binding;
gpuImage.name = image.name;
gpuImage.type = image.type;
gpuImage.count = image.count;
gpuImage.glMemoryAccess = GL_MEMORY_ACCESS[toNumber(image.memoryAccess) - 1];
gpuImage.units.resize(gpuImage.count);
gpuImage.glLoc = glLoc;
// storage image units are immutable
GL_CHECK(glGetUniformiv(gpuShader->glProgram, gpuImage.glLoc, gpuImage.units.data()));
gpuShader->glImages.push_back(gpuImage);
}
}
}
}
void cmdFuncGLES3DestroyShader(GLES3Device *device, GLES3GPUShader *gpuShader) {
GLES3ObjectCache &gfxStateCache = device->stateCache()->gfxStateCache;
if (gpuShader->glProgram) {
if (device->stateCache()->glProgram == gpuShader->glProgram) {
GL_CHECK(glUseProgram(0));
device->stateCache()->glProgram = 0;
gfxStateCache.gpuPipelineState = nullptr;
}
GL_CHECK(glDeleteProgram(gpuShader->glProgram));
gpuShader->glProgram = 0;
}
}
void cmdFuncGLES3CreateRenderPass(GLES3Device * /*device*/, GLES3GPURenderPass *gpuRenderPass) {
auto &subPasses = gpuRenderPass->subpasses;
auto &attachments = gpuRenderPass->colorAttachments;
auto &drawBuffers = gpuRenderPass->drawBuffers;
bool hasDS = (gpuRenderPass->depthStencilAttachment.format != Format::UNKNOWN);
gpuRenderPass->drawBuffers.resize(subPasses.size());
gpuRenderPass->indices.resize(attachments.size() + hasDS, INVALID_BINDING);
for (uint32_t i = 0; i < subPasses.size(); ++i) {
auto &sub = subPasses[i];
auto &drawBuffer = drawBuffers[i];
std::vector<bool> visited(gpuRenderPass->colorAttachments.size() + hasDS);
for (auto &input : sub.inputs) {
visited[input] = true;
if(input == gpuRenderPass->colorAttachments.size()) {
// ds input
continue;
}
drawBuffer.emplace_back(gpuRenderPass->indices[input]);
}
for (auto &color : sub.colors) {
auto &index = gpuRenderPass->indices[color];
if (index == INVALID_BINDING) {
index = static_cast<uint32_t>(gpuRenderPass->colors.size());
gpuRenderPass->colors.emplace_back(color);
}
if (!visited[color]) {
drawBuffer.emplace_back(index);
}
}
for (auto &resolve : sub.resolves) {
if (resolve == INVALID_BINDING) {
gpuRenderPass->resolves.emplace_back(resolve);
continue;
}
auto &index = gpuRenderPass->indices[resolve];
if (index == INVALID_BINDING) {
index = static_cast<uint32_t>(gpuRenderPass->resolves.size());
gpuRenderPass->resolves.emplace_back(resolve);
}
}
if (sub.depthStencil != gfx::INVALID_BINDING) {
gpuRenderPass->depthStencil = sub.depthStencil;
gpuRenderPass->indices.back() = gpuRenderPass->depthStencil;
}
if (sub.depthStencilResolve != gfx::INVALID_BINDING) {
gpuRenderPass->depthStencilResolve = sub.depthStencilResolve;
}
}
}
void cmdFuncGLES3DestroyRenderPass(GLES3Device * /*device*/, GLES3GPURenderPass *gpuRenderPass) {
}
void cmdFuncGLES3CreateInputAssembler(GLES3Device *device, GLES3GPUInputAssembler *gpuInputAssembler) {
if (gpuInputAssembler->gpuIndexBuffer) {
switch (gpuInputAssembler->gpuIndexBuffer->stride) {
case 1: gpuInputAssembler->glIndexType = GL_UNSIGNED_BYTE; break;
case 2: gpuInputAssembler->glIndexType = GL_UNSIGNED_SHORT; break;
case 4: gpuInputAssembler->glIndexType = GL_UNSIGNED_INT; break;
default: {
CC_LOG_ERROR("Illegal index buffer stride.");
}
}
}
ccstd::vector<uint32_t> streamOffsets(device->getCapabilities().maxVertexAttributes, 0U);
gpuInputAssembler->glAttribs.resize(gpuInputAssembler->attributes.size());
for (size_t i = 0; i < gpuInputAssembler->glAttribs.size(); ++i) {
GLES3GPUAttribute &gpuAttribute = gpuInputAssembler->glAttribs[i];
const Attribute &attrib = gpuInputAssembler->attributes[i];
auto *gpuVB = static_cast<GLES3GPUBuffer *>(gpuInputAssembler->gpuVertexBuffers[attrib.stream]);
gpuAttribute.name = attrib.name;
gpuAttribute.glType = formatToGLType(attrib.format);
gpuAttribute.size = GFX_FORMAT_INFOS[static_cast<int>(attrib.format)].size;
gpuAttribute.count = GFX_FORMAT_INFOS[static_cast<int>(attrib.format)].count;
gpuAttribute.componentCount = glComponentCount(gpuAttribute.glType);
gpuAttribute.isNormalized = attrib.isNormalized;
gpuAttribute.isInstanced = attrib.isInstanced;
gpuAttribute.offset = streamOffsets[attrib.stream];
if (gpuVB) {
gpuAttribute.glBuffer = gpuVB->glBuffer;
gpuAttribute.stride = gpuVB->stride;
}
streamOffsets[attrib.stream] += gpuAttribute.size;
}
}
void cmdFuncGLES3DestroyInputAssembler(GLES3Device *device, GLES3GPUInputAssembler *gpuInputAssembler) {
GLES3ObjectCache &gfxStateCache = device->stateCache()->gfxStateCache;
for (auto it = gpuInputAssembler->glVAOs.begin(); it != gpuInputAssembler->glVAOs.end(); ++it) {
if (device->stateCache()->glVAO == it->second) {
GL_CHECK(glBindVertexArray(0));
device->stateCache()->glVAO = 0;
gfxStateCache.gpuInputAssembler = nullptr;
}
GL_CHECK(glDeleteVertexArrays(1, &it->second));
}
gpuInputAssembler->glVAOs.clear();
}
static GLES3GPUSwapchain *getSwapchainIfExists(const ccstd::vector<GLES3GPUTextureView *> &textureViews, const uint32_t *indices, size_t count) {
GLES3GPUSwapchain *swapchain{nullptr};
if (indices) {
for (size_t i = 0; i < count; ++i) {
if (indices[i] == INVALID_BINDING) {
continue;
}
auto *colorTexture = textureViews[indices[i]]->gpuTexture;
if (colorTexture->swapchain) {
swapchain = colorTexture->swapchain;
}
}
}
return swapchain;
}
static GLbitfield getColorBufferMask(Format format) {
GLbitfield mask = 0U;
const FormatInfo &info = GFX_FORMAT_INFOS[toNumber(format)];
if (info.hasDepth || info.hasStencil) {
if (info.hasDepth) mask |= GL_DEPTH_BUFFER_BIT;
if (info.hasStencil) mask |= GL_STENCIL_BUFFER_BIT;
} else {
mask = GL_COLOR_BUFFER_BIT;
}
return mask;
}
static void doResolve(GLES3Device *device, GLES3GPUFramebuffer *gpuFbo) {
device->context()->makeCurrent(gpuFbo->resolveFramebuffer.swapchain, gpuFbo->framebuffer.swapchain);
auto *cache = device->stateCache();
auto width = gpuFbo->width;
auto height = gpuFbo->height;
const auto fbHandle = gpuFbo->framebuffer.getHandle();
if (cache->glReadFramebuffer != fbHandle) {
GL_CHECK(glBindFramebuffer(GL_READ_FRAMEBUFFER, fbHandle));
cache->glReadFramebuffer = fbHandle;
}
const auto rsvHandle = gpuFbo->resolveFramebuffer.getHandle();
if (cache->glDrawFramebuffer != rsvHandle) {
GL_CHECK(glBindFramebuffer(GL_DRAW_FRAMEBUFFER, rsvHandle));
cache->glDrawFramebuffer = rsvHandle;
}
gpuFbo->resolveFramebuffer.processLoad(GL_DRAW_FRAMEBUFFER);
if (!gpuFbo->colorBlitPairs.empty()) {
auto resolveColorNum = gpuFbo->resolveFramebuffer.colors.size();
std::vector<GLenum> drawBuffers(resolveColorNum, GL_NONE);
for (auto &[src, dst] : gpuFbo->colorBlitPairs) {
drawBuffers[dst] = GL_COLOR_ATTACHMENT0 + dst;
GL_CHECK(glReadBuffer(GL_COLOR_ATTACHMENT0 + src));
if (rsvHandle != 0) {
GL_CHECK(glDrawBuffers(resolveColorNum, drawBuffers.data()));
}
GL_CHECK(glBlitFramebuffer(
0, 0, width, height,
0, 0, width, height,
GL_COLOR_BUFFER_BIT, GL_NEAREST));
drawBuffers[dst] = GL_NONE;
}
}
if (gpuFbo->dsResolveMask != 0 && rsvHandle != 0) {
GL_CHECK(glBlitFramebuffer(
0, 0, width, height,
0, 0, width, height,
gpuFbo->dsResolveMask, GL_NEAREST));
}
gpuFbo->framebuffer.processStore(GL_READ_FRAMEBUFFER);
gpuFbo->resolveFramebuffer.processStore(GL_DRAW_FRAMEBUFFER);
}
void cmdFuncGLES3CreateFramebuffer(GLES3Device *device, GLES3GPUFramebuffer *gpuFBO) {
const auto *renderPass = gpuFBO->gpuRenderPass;
const auto &colors = renderPass->colors;
const auto &resolves = renderPass->resolves;
const auto &indices = renderPass->indices;
const auto depthStencil = renderPass->depthStencil;
const auto depthStencilResolve = renderPass->depthStencilResolve;
gpuFBO->framebuffer.initialize(getSwapchainIfExists(gpuFBO->gpuColorViews, colors.data(), colors.size()));
gpuFBO->resolveFramebuffer.initialize(getSwapchainIfExists(gpuFBO->gpuColorViews, resolves.data(), resolves.size()));
auto supportLevel = device->constantRegistry()->mMSRT;
/*
* LEVEL0 does ont support on-chip resolve
* LEVEL1 only support COLOR_ATTACHMENT0
* LEVEL2 support COLOR_ATTACHMENT(i) + DEPTH_STENCIL
*/
uint32_t supportCount = supportLevel > MSRTSupportLevel::LEVEL1 ? 255 : static_cast<uint32_t>(supportLevel);
constexpr bool useDsResolve = true;
uint32_t resolveColorIndex = 0;
for (uint32_t i = 0; i < colors.size(); ++i) {
const auto &attachmentIndex = colors[i];
const auto &colorIndex = indices[attachmentIndex];
const auto &resolveIndex = resolves.empty() ? INVALID_BINDING : resolves[attachmentIndex];
const auto &desc = renderPass->colorAttachments[attachmentIndex];
const auto *view = gpuFBO->gpuColorViews[attachmentIndex];
CC_ASSERT(view != nullptr);
// need to resolve
if (view->gpuTexture->glSamples > 1 && resolveIndex != INVALID_BINDING) {
const auto &resolveDesc = renderPass->colorAttachments[resolveIndex];
const auto *resolveView = gpuFBO->gpuColorViews[resolveIndex];
CC_ASSERT(resolveView != nullptr);
bool lazilyAllocated = hasFlag(view->gpuTexture->flags, TextureFlagBit::LAZILY_ALLOCATED);
if (lazilyAllocated && // MS attachment should be memoryless
resolveView->gpuTexture->swapchain == nullptr && // not back buffer
i < supportCount) { // extension limit
auto validateDesc = resolveDesc;
// implicit MS take color slot, so color loadOP should be used.
// resolve attachment with Store::Discard is meaningless.
validateDesc.loadOp = desc.loadOp;
validateDesc.storeOp = StoreOp::STORE;
gpuFBO->framebuffer.bindColorMultiSample(resolveView, colorIndex, view->gpuTexture->glSamples, validateDesc);
} else {
// implicit MS not supported, fallback to MS Renderbuffer
gpuFBO->colorBlitPairs.emplace_back(colorIndex, resolveColorIndex);
gpuFBO->framebuffer.bindColor(view, colorIndex, desc);
gpuFBO->resolveFramebuffer.bindColor(resolveView, resolveColorIndex++, resolveDesc);
}
continue;
}
gpuFBO->framebuffer.bindColor(view, colorIndex, desc);
}
if (depthStencil != INVALID_BINDING) {
const auto &desc = renderPass->depthStencilAttachment;
const auto *view = gpuFBO->gpuDepthStencilView;
CC_ASSERT(view != nullptr);
if (view->gpuTexture->glSamples > 1 && depthStencilResolve != INVALID_BINDING) {
const auto &resolveDesc = renderPass->depthStencilResolveAttachment;
const auto *resolveView = gpuFBO->gpuDepthStencilResolveView;
bool lazilyAllocated = hasFlag(view->gpuTexture->flags, TextureFlagBit::LAZILY_ALLOCATED);
if (lazilyAllocated && // MS attachment should be memoryless
resolveView->gpuTexture->swapchain == nullptr && // not back buffer
supportCount > 1 && // extension limit
useDsResolve) { // enable ds resolve
auto validateDesc = resolveDesc;
// implicit MS take ds slot, so ds MS loadOP should be used.
// resolve attachment with Store::Discard is meaningless.
validateDesc.depthLoadOp = desc.depthLoadOp;
validateDesc.depthStoreOp = StoreOp::STORE;
validateDesc.stencilLoadOp = desc.stencilLoadOp;
validateDesc.stencilStoreOp = StoreOp::STORE;
gpuFBO->framebuffer.bindDepthStencilMultiSample(resolveView, view->gpuTexture->glSamples, validateDesc);
} else {
// implicit MS not supported, fallback to MS Renderbuffer
gpuFBO->dsResolveMask = getColorBufferMask(desc.format);
gpuFBO->framebuffer.bindDepthStencil(view, desc);
gpuFBO->resolveFramebuffer.bindDepthStencil(resolveView, resolveDesc);
}
} else {
gpuFBO->framebuffer.bindDepthStencil(view, desc);
}
}
gpuFBO->framebuffer.finalize(device->stateCache());
gpuFBO->resolveFramebuffer.finalize(device->stateCache());
}
void cmdFuncGLES3DestroyFramebuffer(GLES3Device *device, GLES3GPUFramebuffer *gpuFBO) {
auto *cache = device->stateCache();
auto *framebufferCacheMap = device->framebufferCacheMap();
gpuFBO->framebuffer.destroy(cache, framebufferCacheMap);
gpuFBO->resolveFramebuffer.destroy(cache, framebufferCacheMap);
}
void completeBarrier(GLES3GPUGeneralBarrier *barrier) {
bool hasShaderWrites = false;
for (uint32_t mask = toNumber(barrier->prevAccesses); mask; mask = utils::clearLowestBit(mask)) {
switch (static_cast<AccessFlagBit>(utils::getLowestBit(mask))) {
case AccessFlagBit::COMPUTE_SHADER_WRITE:
case AccessFlagBit::VERTEX_SHADER_WRITE:
case AccessFlagBit::FRAGMENT_SHADER_WRITE:
case AccessFlagBit::COLOR_ATTACHMENT_WRITE:
case AccessFlagBit::DEPTH_STENCIL_ATTACHMENT_WRITE:
hasShaderWrites = true;
break;
default:
break;
}
}
if (hasShaderWrites) {
for (uint32_t mask = toNumber(barrier->nextAccesses); mask; mask = utils::clearLowestBit(mask)) {
switch (static_cast<AccessFlagBit>(utils::getLowestBit(mask))) {
case AccessFlagBit::INDIRECT_BUFFER:
barrier->glBarriers |= GL_COMMAND_BARRIER_BIT;
break;
case AccessFlagBit::INDEX_BUFFER:
barrier->glBarriers |= GL_ELEMENT_ARRAY_BARRIER_BIT;
break;
case AccessFlagBit::VERTEX_BUFFER:
barrier->glBarriers |= GL_VERTEX_ATTRIB_ARRAY_BARRIER_BIT;
break;
case AccessFlagBit::COMPUTE_SHADER_READ_UNIFORM_BUFFER:
case AccessFlagBit::VERTEX_SHADER_READ_UNIFORM_BUFFER:
case AccessFlagBit::FRAGMENT_SHADER_READ_UNIFORM_BUFFER:
barrier->glBarriersByRegion |= GL_UNIFORM_BARRIER_BIT;
break;
case AccessFlagBit::COMPUTE_SHADER_READ_TEXTURE:
case AccessFlagBit::VERTEX_SHADER_READ_TEXTURE:
case AccessFlagBit::FRAGMENT_SHADER_READ_TEXTURE:
case AccessFlagBit::FRAGMENT_SHADER_READ_COLOR_INPUT_ATTACHMENT:
case AccessFlagBit::FRAGMENT_SHADER_READ_DEPTH_STENCIL_INPUT_ATTACHMENT:
barrier->glBarriersByRegion |= GL_SHADER_IMAGE_ACCESS_BARRIER_BIT;
barrier->glBarriersByRegion |= GL_TEXTURE_FETCH_BARRIER_BIT;
break;
case AccessFlagBit::COMPUTE_SHADER_READ_OTHER:
case AccessFlagBit::VERTEX_SHADER_READ_OTHER:
case AccessFlagBit::FRAGMENT_SHADER_READ_OTHER:
barrier->glBarriersByRegion |= GL_SHADER_STORAGE_BARRIER_BIT;
break;
case AccessFlagBit::COLOR_ATTACHMENT_READ:
case AccessFlagBit::DEPTH_STENCIL_ATTACHMENT_READ:
barrier->glBarriersByRegion |= GL_FRAMEBUFFER_BARRIER_BIT;
break;
case AccessFlagBit::TRANSFER_READ:
barrier->glBarriersByRegion |= GL_FRAMEBUFFER_BARRIER_BIT;
barrier->glBarriers |= GL_TEXTURE_UPDATE_BARRIER_BIT;
barrier->glBarriers |= GL_BUFFER_UPDATE_BARRIER_BIT;
barrier->glBarriers |= GL_PIXEL_BUFFER_BARRIER_BIT;
break;
case AccessFlagBit::COMPUTE_SHADER_WRITE:
case AccessFlagBit::VERTEX_SHADER_WRITE:
case AccessFlagBit::FRAGMENT_SHADER_WRITE:
barrier->glBarriersByRegion |= GL_SHADER_IMAGE_ACCESS_BARRIER_BIT;
barrier->glBarriersByRegion |= GL_SHADER_STORAGE_BARRIER_BIT;
break;
case AccessFlagBit::COLOR_ATTACHMENT_WRITE:
case AccessFlagBit::DEPTH_STENCIL_ATTACHMENT_WRITE:
barrier->glBarriersByRegion |= GL_FRAMEBUFFER_BARRIER_BIT;
break;
case AccessFlagBit::TRANSFER_WRITE:
barrier->glBarriersByRegion |= GL_FRAMEBUFFER_BARRIER_BIT;
barrier->glBarriers |= GL_TEXTURE_UPDATE_BARRIER_BIT;
barrier->glBarriers |= GL_BUFFER_UPDATE_BARRIER_BIT;
barrier->glBarriers |= GL_PIXEL_BUFFER_BARRIER_BIT;
break;
case AccessFlagBit::HOST_PREINITIALIZED:
case AccessFlagBit::HOST_WRITE:
case AccessFlagBit::PRESENT:
default:
break;
}
}
}
}
void cmdFuncGLES3CreateGeneralBarrier(GLES3Device * /*device*/, GLES3GPUGeneralBarrier *barrier) {
completeBarrier(barrier);
}
void cmdFuncGLES3CreateQueryPool(GLES3Device * /*device*/, GLES3GPUQueryPool *gpuQueryPool) {
GL_CHECK(glGenQueries(gpuQueryPool->maxQueryObjects, gpuQueryPool->glQueryIds.data()));
}
void cmdFuncGLES3DestroyQueryPool(GLES3Device * /*device*/, GLES3GPUQueryPool *gpuQueryPool) {
GL_CHECK(glDeleteQueries(gpuQueryPool->maxQueryObjects, gpuQueryPool->glQueryIds.data()));
}
void cmdFuncGLES3Query(GLES3Device * /*device*/, GLES3QueryPool *queryPool, GLES3QueryType type, uint32_t id) {
GLES3GPUQueryPool *gpuQueryPool = queryPool->gpuQueryPool();
switch (type) {
case GLES3QueryType::BEGIN: {
auto queryId = queryPool->getIdCount();
GLuint glQueryId = gpuQueryPool->mapGLQueryId(queryId);
if (glQueryId != UINT_MAX) {
GL_CHECK(glBeginQuery(GL_ANY_SAMPLES_PASSED, glQueryId));
}
break;
}
case GLES3QueryType::END: {
auto queryId = queryPool->getIdCount();
GLuint glQueryId = gpuQueryPool->mapGLQueryId(queryId);
if (glQueryId != UINT_MAX) {
GL_CHECK(glEndQuery(GL_ANY_SAMPLES_PASSED));
queryPool->addId(id);
}
break;
}
case GLES3QueryType::RESET: {
queryPool->clearId();
break;
}
case GLES3QueryType::GET_RESULTS: {
auto queryCount = queryPool->getIdCount();
ccstd::vector<uint64_t> results(queryCount);
for (auto queryId = 0U; queryId < queryCount; queryId++) {
GLuint glQueryId = gpuQueryPool->mapGLQueryId(queryId);
if (glQueryId != UINT_MAX) {
GLuint result{0};
GL_CHECK(glGetQueryObjectuiv(glQueryId, GL_QUERY_RESULT, &result));
results[queryId] = static_cast<uint64_t>(result);
} else {
results[queryId] = 1ULL;
}
}
ccstd::unordered_map<uint32_t, uint64_t> mapResults;
for (auto queryId = 0U; queryId < queryCount; queryId++) {
uint32_t glID = queryPool->getId(queryId);
auto iter = mapResults.find(glID);
if (iter != mapResults.end()) {
iter->second += results[queryId];
} else {
mapResults[glID] = results[queryId];
}
}
{
std::lock_guard<std::mutex> lock(queryPool->getMutex());
queryPool->setResults(std::move(mapResults));
}
break;
}
default: {
break;
}
}
}
void cmdFuncGLES3BeginRenderPass(GLES3Device *device, GLES3GPURenderPass *gpuRenderPass, GLES3GPUFramebuffer *gpuFramebuffer,
const Rect *renderArea, const Color *clearColors, float clearDepth, uint32_t clearStencil) {
GLES3GPUStateCache *cache = device->stateCache();
GLES3ObjectCache &gfxStateCache = cache->gfxStateCache;
gfxStateCache.subpassIdx = 0;
gfxStateCache.gpuRenderPass = gpuRenderPass;
gfxStateCache.gpuFramebuffer = gpuFramebuffer;
gfxStateCache.renderArea = *renderArea;
gfxStateCache.clearColors.assign(clearColors, clearColors + gpuRenderPass->colorAttachments.size());
gfxStateCache.clearDepth = clearDepth;
gfxStateCache.clearStencil = clearStencil;
if (gpuFramebuffer && gpuRenderPass) {
auto &framebuffer = gpuFramebuffer->framebuffer;
auto &resolveFramebuffer = gpuFramebuffer->resolveFramebuffer;
if (gpuFramebuffer->resolveFramebuffer.isActive()) {
device->context()->makeCurrent(resolveFramebuffer.swapchain, framebuffer.swapchain);
} else {
device->context()->makeCurrent(framebuffer.swapchain);
}
const auto fbHandle = framebuffer.getHandle();
if (cache->glDrawFramebuffer != fbHandle) {
GL_CHECK(glBindFramebuffer(GL_DRAW_FRAMEBUFFER, fbHandle));
cache->glDrawFramebuffer = fbHandle;
}
if (cache->viewport.left != renderArea->x ||
cache->viewport.top != renderArea->y ||
cache->viewport.width != renderArea->width ||
cache->viewport.height != renderArea->height) {
GL_CHECK(glViewport(renderArea->x, renderArea->y, renderArea->width, renderArea->height));
cache->viewport.left = renderArea->x;
cache->viewport.top = renderArea->y;
cache->viewport.width = renderArea->width;
cache->viewport.height = renderArea->height;
}
if (cache->scissor != *renderArea) {
GL_CHECK(glScissor(renderArea->x, renderArea->y, renderArea->width, renderArea->height));
cache->scissor = *renderArea;
}
GLbitfield glClears = 0;
float fColors[4]{};
bool maskSet = false;
auto performLoadOp = [&](uint32_t attachmentIndex, uint32_t glAttachmentIndex) {
const ColorAttachment &colorAttachment = gpuRenderPass->colorAttachments[attachmentIndex];
if (colorAttachment.format != Format::UNKNOWN && colorAttachment.loadOp == LoadOp::CLEAR) {
if (!maskSet && cache->bs.targets[0].blendColorMask != ColorMask::ALL) {
GL_CHECK(glColorMask(true, true, true, true));
maskSet = true;
}
const Color &color = clearColors[attachmentIndex];
if (fbHandle) {
fColors[0] = color.x;
fColors[1] = color.y;
fColors[2] = color.z;
fColors[3] = color.w;
GL_CHECK(glClearBufferfv(GL_COLOR, glAttachmentIndex, fColors));
} else {
GL_CHECK(glClearColor(color.x, color.y, color.z, color.w));
glClears |= GL_COLOR_BUFFER_BIT;
}
}
};
auto performDepthStencilLoadOp = [&](uint32_t attachmentIndex) {
if (attachmentIndex != INVALID_BINDING) {
LoadOp depthLoadOp = gpuRenderPass->depthStencilAttachment.depthLoadOp;
LoadOp stencilLoadOp = gpuRenderPass->depthStencilAttachment.stencilLoadOp;
bool hasStencils = GFX_FORMAT_INFOS[toNumber(gpuRenderPass->depthStencilAttachment.format)].hasStencil;
if (attachmentIndex < gpuRenderPass->colorAttachments.size()) {
depthLoadOp = stencilLoadOp = gpuRenderPass->colorAttachments[attachmentIndex].loadOp;
hasStencils = GFX_FORMAT_INFOS[toNumber(gpuRenderPass->colorAttachments[attachmentIndex].format)].hasStencil;
}
if (gpuRenderPass->depthStencilAttachment.depthLoadOp == LoadOp::CLEAR) {
if (!cache->dss.depthWrite) {
GL_CHECK(glDepthMask(true));
}
GL_CHECK(glClearDepthf(clearDepth));
glClears |= GL_DEPTH_BUFFER_BIT;
}
if (hasStencils && gpuRenderPass->depthStencilAttachment.depthLoadOp == LoadOp::CLEAR) {
if (!cache->dss.stencilWriteMaskFront) {
GL_CHECK(glStencilMaskSeparate(GL_FRONT, 0xffffffff));
}
if (!cache->dss.stencilWriteMaskBack) {
GL_CHECK(glStencilMaskSeparate(GL_BACK, 0xffffffff));
}
GL_CHECK(glClearStencil(clearStencil));
glClears |= GL_STENCIL_BUFFER_BIT;
}
}
if (glClears) {
GL_CHECK(glClear(glClears));
}
// restore states
if (maskSet) {
ColorMask colorMask = cache->bs.targets[0].blendColorMask;
GL_CHECK(glColorMask((GLboolean)(colorMask & ColorMask::R),
(GLboolean)(colorMask & ColorMask::G),
(GLboolean)(colorMask & ColorMask::B),
(GLboolean)(colorMask & ColorMask::A)));
}
if ((glClears & GL_DEPTH_BUFFER_BIT) && !cache->dss.depthWrite) {
GL_CHECK(glDepthMask(false));
}
if (glClears & GL_STENCIL_BUFFER_BIT) {
if (!cache->dss.stencilWriteMaskFront) {
GL_CHECK(glStencilMaskSeparate(GL_FRONT, 0));
}
if (!cache->dss.stencilWriteMaskBack) {
GL_CHECK(glStencilMaskSeparate(GL_BACK, 0));
}
}
};
const auto &attachments = gpuRenderPass->colorAttachments;
const auto &indices = gpuRenderPass->indices;
for (uint32_t i = 0; i < attachments.size(); ++i) {
performLoadOp(i, indices[i]);
}
performDepthStencilLoadOp(gpuRenderPass->depthStencil);
framebuffer.processLoad(GL_DRAW_FRAMEBUFFER);
}
}
static void ensureScissorRect(GLES3GPUStateCache *cache, int32_t x, int32_t y, uint32_t width, uint32_t height) {
if (cache->scissor.x > x ||
cache->scissor.y > y ||
cache->scissor.width < width ||
cache->scissor.height < height) {
cache->scissor.x = std::min(cache->scissor.x, x);
cache->scissor.y = std::min(cache->scissor.y, y);
cache->scissor.width = std::max(cache->scissor.width, width);
cache->scissor.height = std::max(cache->scissor.height, height);
GL_CHECK(glScissor(cache->scissor.x, cache->scissor.y, cache->scissor.width, cache->scissor.height));
}
}
void cmdFuncGLES3EndRenderPass(GLES3Device *device) {
ccstd::vector<GLenum> invalidAttachments;
GLES3GPUStateCache *cache = device->stateCache();
GLES3ObjectCache &gfxStateCache = cache->gfxStateCache;
GLES3GPUFramebuffer *gpuFramebuffer = gfxStateCache.gpuFramebuffer;
if (gpuFramebuffer->resolveFramebuffer.isActive()) {
doResolve(device, gpuFramebuffer);
} else {
gpuFramebuffer->framebuffer.processStore(GL_DRAW_FRAMEBUFFER);
}
// if (device->constantRegistry()->mFBF == FBFSupportLevel::NON_COHERENT_EXT) {
// GL_CHECK(glFramebufferFetchBarrierEXT());
// } else if (device->constantRegistry()->mFBF == FBFSupportLevel::NON_COHERENT_QCOM) {
// GL_CHECK(glFramebufferFetchBarrierQCOM());
// }
}
// NOLINTNEXTLINE(google-readability-function-size, readability-function-size)
void cmdFuncGLES3BindState(GLES3Device *device, GLES3GPUPipelineState *gpuPipelineState, GLES3GPUInputAssembler *gpuInputAssembler,
const GLES3GPUDescriptorSet *const *gpuDescriptorSets, const uint32_t *dynamicOffsets, const DynamicStates *dynamicStates) {
GLES3ObjectCache &gfxStateCache = device->stateCache()->gfxStateCache;
GLES3GPUStateCache *cache = device->stateCache();
bool isShaderChanged = false;
if (gpuPipelineState && gpuPipelineState != gfxStateCache.gpuPipelineState) {
gfxStateCache.gpuPipelineState = gpuPipelineState;
gfxStateCache.glPrimitive = gpuPipelineState->glPrimitive;
if (gpuPipelineState->gpuShader) {
if (cache->glProgram != gpuPipelineState->gpuShader->glProgram) {
GL_CHECK(glUseProgram(gpuPipelineState->gpuShader->glProgram));
cache->glProgram = gpuPipelineState->gpuShader->glProgram;
isShaderChanged = true;
}
}
// bind rasterizer state
if (cache->rs.cullMode != gpuPipelineState->rs.cullMode) {
switch (gpuPipelineState->rs.cullMode) {
case CullMode::NONE: {
if (cache->isCullFaceEnabled) {
GL_CHECK(glDisable(GL_CULL_FACE));
cache->isCullFaceEnabled = false;
}
} break;
case CullMode::FRONT: {
if (!cache->isCullFaceEnabled) {
GL_CHECK(glEnable(GL_CULL_FACE));
cache->isCullFaceEnabled = true;
}
GL_CHECK(glCullFace(GL_FRONT));
} break;
case CullMode::BACK: {
if (!cache->isCullFaceEnabled) {
GL_CHECK(glEnable(GL_CULL_FACE));
cache->isCullFaceEnabled = true;
}
GL_CHECK(glCullFace(GL_BACK));
} break;
default:
break;
}
cache->rs.cullMode = gpuPipelineState->rs.cullMode;
}
if (cache->rs.isFrontFaceCCW != gpuPipelineState->rs.isFrontFaceCCW) {
GL_CHECK(glFrontFace(gpuPipelineState->rs.isFrontFaceCCW ? GL_CCW : GL_CW));
cache->rs.isFrontFaceCCW = gpuPipelineState->rs.isFrontFaceCCW;
}
if ((cache->rs.depthBias != gpuPipelineState->rs.depthBias) ||
(cache->rs.depthBiasSlop != gpuPipelineState->rs.depthBiasSlop)) {
GL_CHECK(glPolygonOffset(cache->rs.depthBias, cache->rs.depthBiasSlop));
cache->rs.depthBiasSlop = gpuPipelineState->rs.depthBiasSlop;
}
if (cache->rs.lineWidth != gpuPipelineState->rs.lineWidth) {
GL_CHECK(glLineWidth(gpuPipelineState->rs.lineWidth));
cache->rs.lineWidth = gpuPipelineState->rs.lineWidth;
}
// bind depth-stencil state
if (cache->dss.depthTest != gpuPipelineState->dss.depthTest) {
if (gpuPipelineState->dss.depthTest) {
GL_CHECK(glEnable(GL_DEPTH_TEST));
} else {
GL_CHECK(glDisable(GL_DEPTH_TEST));
}
cache->dss.depthTest = gpuPipelineState->dss.depthTest;
}
if (cache->dss.depthWrite != gpuPipelineState->dss.depthWrite) {
GL_CHECK(glDepthMask(static_cast<bool>(gpuPipelineState->dss.depthWrite)));
cache->dss.depthWrite = gpuPipelineState->dss.depthWrite;
}
if (cache->dss.depthFunc != gpuPipelineState->dss.depthFunc) {
GL_CHECK(glDepthFunc(GLES3_CMP_FUNCS[(int)gpuPipelineState->dss.depthFunc]));
cache->dss.depthFunc = gpuPipelineState->dss.depthFunc;
}
// bind depth-stencil state - front
if (gpuPipelineState->dss.stencilTestFront || gpuPipelineState->dss.stencilTestBack) {
if (!cache->isStencilTestEnabled) {
GL_CHECK(glEnable(GL_STENCIL_TEST));
cache->isStencilTestEnabled = true;
}
} else {
if (cache->isStencilTestEnabled) {
GL_CHECK(glDisable(GL_STENCIL_TEST));
cache->isStencilTestEnabled = false;
}
}
if (cache->dss.stencilFuncFront != gpuPipelineState->dss.stencilFuncFront ||
cache->dss.stencilRefFront != gpuPipelineState->dss.stencilRefFront ||
cache->dss.stencilReadMaskFront != gpuPipelineState->dss.stencilReadMaskFront) {
GL_CHECK(glStencilFuncSeparate(GL_FRONT,
GLES3_CMP_FUNCS[(int)gpuPipelineState->dss.stencilFuncFront],
gpuPipelineState->dss.stencilRefFront,
gpuPipelineState->dss.stencilReadMaskFront));
cache->dss.stencilFuncFront = gpuPipelineState->dss.stencilFuncFront;
cache->dss.stencilRefFront = gpuPipelineState->dss.stencilRefFront;
cache->dss.stencilReadMaskFront = gpuPipelineState->dss.stencilReadMaskFront;
}
if (cache->dss.stencilFailOpFront != gpuPipelineState->dss.stencilFailOpFront ||
cache->dss.stencilZFailOpFront != gpuPipelineState->dss.stencilZFailOpFront ||
cache->dss.stencilPassOpFront != gpuPipelineState->dss.stencilPassOpFront) {
GL_CHECK(glStencilOpSeparate(GL_FRONT,
GLES3_STENCIL_OPS[(int)gpuPipelineState->dss.stencilFailOpFront],
GLES3_STENCIL_OPS[(int)gpuPipelineState->dss.stencilZFailOpFront],
GLES3_STENCIL_OPS[(int)gpuPipelineState->dss.stencilPassOpFront]));
cache->dss.stencilFailOpFront = gpuPipelineState->dss.stencilFailOpFront;
cache->dss.stencilZFailOpFront = gpuPipelineState->dss.stencilZFailOpFront;
cache->dss.stencilPassOpFront = gpuPipelineState->dss.stencilPassOpFront;
}
if (cache->dss.stencilWriteMaskFront != gpuPipelineState->dss.stencilWriteMaskFront) {
GL_CHECK(glStencilMaskSeparate(GL_FRONT, gpuPipelineState->dss.stencilWriteMaskFront));
cache->dss.stencilWriteMaskFront = gpuPipelineState->dss.stencilWriteMaskFront;
}
// bind depth-stencil state - back
if (cache->dss.stencilFuncBack != gpuPipelineState->dss.stencilFuncBack ||
cache->dss.stencilRefBack != gpuPipelineState->dss.stencilRefBack ||
cache->dss.stencilReadMaskBack != gpuPipelineState->dss.stencilReadMaskBack) {
GL_CHECK(glStencilFuncSeparate(GL_BACK,
GLES3_CMP_FUNCS[(int)gpuPipelineState->dss.stencilFuncBack],
gpuPipelineState->dss.stencilRefBack,
gpuPipelineState->dss.stencilReadMaskBack));
cache->dss.stencilFuncBack = gpuPipelineState->dss.stencilFuncBack;
cache->dss.stencilRefBack = gpuPipelineState->dss.stencilRefBack;
cache->dss.stencilReadMaskBack = gpuPipelineState->dss.stencilReadMaskBack;
}
if (cache->dss.stencilFailOpBack != gpuPipelineState->dss.stencilFailOpBack ||
cache->dss.stencilZFailOpBack != gpuPipelineState->dss.stencilZFailOpBack ||
cache->dss.stencilPassOpBack != gpuPipelineState->dss.stencilPassOpBack) {
GL_CHECK(glStencilOpSeparate(GL_BACK,
GLES3_STENCIL_OPS[(int)gpuPipelineState->dss.stencilFailOpBack],
GLES3_STENCIL_OPS[(int)gpuPipelineState->dss.stencilZFailOpBack],
GLES3_STENCIL_OPS[(int)gpuPipelineState->dss.stencilPassOpBack]));
cache->dss.stencilFailOpBack = gpuPipelineState->dss.stencilFailOpBack;
cache->dss.stencilZFailOpBack = gpuPipelineState->dss.stencilZFailOpBack;
cache->dss.stencilPassOpBack = gpuPipelineState->dss.stencilPassOpBack;
}
if (cache->dss.stencilWriteMaskBack != gpuPipelineState->dss.stencilWriteMaskBack) {
GL_CHECK(glStencilMaskSeparate(GL_BACK, gpuPipelineState->dss.stencilWriteMaskBack));
cache->dss.stencilWriteMaskBack = gpuPipelineState->dss.stencilWriteMaskBack;
}
// bind blend state
if (cache->bs.isA2C != gpuPipelineState->bs.isA2C) {
if (cache->bs.isA2C) {
GL_CHECK(glEnable(GL_SAMPLE_ALPHA_TO_COVERAGE));
} else {
GL_CHECK(glDisable(GL_SAMPLE_ALPHA_TO_COVERAGE));
}
cache->bs.isA2C = gpuPipelineState->bs.isA2C;
}
if (cache->bs.blendColor.x != gpuPipelineState->bs.blendColor.x ||
cache->bs.blendColor.y != gpuPipelineState->bs.blendColor.y ||
cache->bs.blendColor.z != gpuPipelineState->bs.blendColor.z ||
cache->bs.blendColor.w != gpuPipelineState->bs.blendColor.w) {
GL_CHECK(glBlendColor(gpuPipelineState->bs.blendColor.x,
gpuPipelineState->bs.blendColor.y,
gpuPipelineState->bs.blendColor.z,
gpuPipelineState->bs.blendColor.w));
cache->bs.blendColor = gpuPipelineState->bs.blendColor;
}
if (!gpuPipelineState->bs.targets.empty()) {
BlendTarget &cacheTarget = cache->bs.targets[0];
const BlendTarget &target = gpuPipelineState->bs.targets[0];
if (cacheTarget.blend != target.blend) {
if (!cacheTarget.blend) {
GL_CHECK(glEnable(GL_BLEND));
} else {
GL_CHECK(glDisable(GL_BLEND));
}
cacheTarget.blend = target.blend;
}
if (cacheTarget.blendEq != target.blendEq ||
cacheTarget.blendAlphaEq != target.blendAlphaEq) {
GL_CHECK(glBlendEquationSeparate(GLES3_BLEND_OPS[(int)target.blendEq],
GLES3_BLEND_OPS[(int)target.blendAlphaEq]));
cacheTarget.blendEq = target.blendEq;
cacheTarget.blendAlphaEq = target.blendAlphaEq;
}
if (cacheTarget.blendSrc != target.blendSrc ||
cacheTarget.blendDst != target.blendDst ||
cacheTarget.blendSrcAlpha != target.blendSrcAlpha ||
cacheTarget.blendDstAlpha != target.blendDstAlpha) {
GL_CHECK(glBlendFuncSeparate(GLES3_BLEND_FACTORS[(int)target.blendSrc],
GLES3_BLEND_FACTORS[(int)target.blendDst],
GLES3_BLEND_FACTORS[(int)target.blendSrcAlpha],
GLES3_BLEND_FACTORS[(int)target.blendDstAlpha]));
cacheTarget.blendSrc = target.blendSrc;
cacheTarget.blendDst = target.blendDst;
cacheTarget.blendSrcAlpha = target.blendSrcAlpha;
cacheTarget.blendDstAlpha = target.blendDstAlpha;
}
if (cacheTarget.blendColorMask != target.blendColorMask) {
GL_CHECK(glColorMask((GLboolean)(target.blendColorMask & ColorMask::R),
(GLboolean)(target.blendColorMask & ColorMask::G),
(GLboolean)(target.blendColorMask & ColorMask::B),
(GLboolean)(target.blendColorMask & ColorMask::A)));
cacheTarget.blendColorMask = target.blendColorMask;
}
}
} // if
// bind descriptor sets
if (gpuPipelineState && gpuPipelineState->gpuShader && gpuPipelineState->gpuPipelineLayout) {
const ccstd::vector<ccstd::vector<int>> &dynamicOffsetIndices = gpuPipelineState->gpuPipelineLayout->dynamicOffsetIndices;
size_t bufferLen = gpuPipelineState->gpuShader->glBuffers.size();
for (size_t j = 0; j < bufferLen; j++) {
const GLES3GPUUniformBuffer &glBuffer = gpuPipelineState->gpuShader->glBuffers[j];
const GLES3GPUDescriptorSet *gpuDescriptorSet = gpuDescriptorSets[glBuffer.set];
if (gpuDescriptorSet == nullptr) {
continue;
}
const uint32_t descriptorIndex = gpuDescriptorSet->descriptorIndices->at(glBuffer.binding);
const GLES3GPUDescriptor &gpuDescriptor = gpuDescriptorSet->gpuDescriptors[descriptorIndex];
if (!gpuDescriptor.gpuBuffer) {
// CC_LOG_ERROR("Buffer binding '%s' at set %d binding %d is not bounded",
// glBuffer.name.c_str(), glBuffer.set, glBuffer.binding);
continue;
}
uint32_t offset = gpuDescriptor.gpuBuffer->glOffset;
const ccstd::vector<int> &dynamicOffsetSetIndices = dynamicOffsetIndices[glBuffer.set];
int dynamicOffsetIndex = glBuffer.binding < dynamicOffsetSetIndices.size() ? dynamicOffsetSetIndices[glBuffer.binding] : -1;
if (dynamicOffsetIndex >= 0) offset += dynamicOffsets[dynamicOffsetIndex];
if (glBuffer.isStorage) {
if (cache->glBindSSBOs[glBuffer.glBinding] != gpuDescriptor.gpuBuffer->glBuffer ||
cache->glBindSSBOOffsets[glBuffer.glBinding] != offset) {
if (offset) {
GL_CHECK(glBindBufferRange(GL_SHADER_STORAGE_BUFFER, glBuffer.glBinding, gpuDescriptor.gpuBuffer->glBuffer,
offset, gpuDescriptor.gpuBuffer->size));
} else {
GL_CHECK(glBindBufferBase(GL_SHADER_STORAGE_BUFFER, glBuffer.glBinding, gpuDescriptor.gpuBuffer->glBuffer));
}
cache->glShaderStorageBuffer = cache->glBindSSBOs[glBuffer.glBinding] = gpuDescriptor.gpuBuffer->glBuffer;
cache->glBindSSBOOffsets[glBuffer.glBinding] = offset;
}
} else {
if (cache->glBindUBOs[glBuffer.glBinding] != gpuDescriptor.gpuBuffer->glBuffer ||
cache->glBindUBOOffsets[glBuffer.glBinding] != offset) {
if (offset) {
GL_CHECK(glBindBufferRange(GL_UNIFORM_BUFFER, glBuffer.glBinding, gpuDescriptor.gpuBuffer->glBuffer,
offset, gpuDescriptor.gpuBuffer->size));
} else {
GL_CHECK(glBindBufferBase(GL_UNIFORM_BUFFER, glBuffer.glBinding, gpuDescriptor.gpuBuffer->glBuffer));
}
cache->glUniformBuffer = cache->glBindUBOs[glBuffer.glBinding] = gpuDescriptor.gpuBuffer->glBuffer;
cache->glBindUBOOffsets[glBuffer.glBinding] = offset;
}
}
}
size_t samplerTextureLen = gpuPipelineState->gpuShader->glSamplerTextures.size();
for (size_t j = 0; j < samplerTextureLen; j++) {
const GLES3GPUUniformSamplerTexture &glSamplerTexture = gpuPipelineState->gpuShader->glSamplerTextures[j];
const GLES3GPUDescriptorSet *gpuDescriptorSet = gpuDescriptorSets[glSamplerTexture.set];
const uint32_t descriptorIndex = gpuDescriptorSet->descriptorIndices->at(glSamplerTexture.binding);
const GLES3GPUDescriptor *gpuDescriptor = &gpuDescriptorSet->gpuDescriptors[descriptorIndex];
GLES3GPUTextureView *gpuTextureView = nullptr;
GLES3GPUTexture *gpuTexture = nullptr;
uint32_t minLod = 0;
uint32_t maxLod = 0;
for (size_t u = 0; u < glSamplerTexture.units.size(); u++, gpuDescriptor++) {
auto unit = static_cast<uint32_t>(glSamplerTexture.units[u]);
auto *sampler = gpuDescriptor->gpuSampler ? gpuDescriptor->gpuSampler : static_cast<GLES3Sampler *>(device->getSampler({}))->gpuSampler();
if (!gpuDescriptor->gpuTextureView || !gpuDescriptor->gpuTextureView->gpuTexture || !sampler) {
// CC_LOG_ERROR("Sampler texture '%s' at binding %d set %d index %d is not bounded",
// glSamplerTexture.name.c_str(), glSamplerTexture.set, glSamplerTexture.binding, u);
continue;
}
gpuTextureView = gpuDescriptor->gpuTextureView;
gpuTexture = gpuTextureView->gpuTexture;
minLod = gpuTextureView->baseLevel;
maxLod = minLod + gpuTextureView->levelCount;
if (gpuTexture->size > 0) {
GLuint glTexture = gpuTexture->glTexture;
if (cache->glTextures[unit] != glTexture) {
if (cache->texUint != unit) {
GL_CHECK(glActiveTexture(GL_TEXTURE0 + unit));
cache->texUint = unit;
}
GL_CHECK(glBindTexture(gpuTextureView->glTarget, glTexture));
cache->glTextures[unit] = glTexture;
}
GLuint glSampler = sampler->getGLSampler(minLod, maxLod);
if (cache->glSamplers[unit] != glSampler) {
GL_CHECK(glBindSampler(unit, glSampler));
cache->glSamplers[unit] = glSampler;
}
}
}
}
size_t imageLen = gpuPipelineState->gpuShader->glImages.size();
for (size_t j = 0; j < imageLen; j++) {
const GLES3GPUUniformStorageImage &glImage = gpuPipelineState->gpuShader->glImages[j];
const GLES3GPUDescriptorSet *gpuDescriptorSet = gpuDescriptorSets[glImage.set];
const uint32_t descriptorIndex = gpuDescriptorSet->descriptorIndices->at(glImage.binding);
const GLES3GPUDescriptor *gpuDescriptor = &gpuDescriptorSet->gpuDescriptors[descriptorIndex];
GLES3GPUTexture *gpuTexture = nullptr;
for (size_t u = 0; u < glImage.units.size(); u++, gpuDescriptor++) {
auto unit = static_cast<uint32_t>(glImage.units[u]);
if (!gpuDescriptor->gpuTextureView || !gpuDescriptor->gpuTextureView->gpuTexture) {
// CC_LOG_ERROR("Storage image '%s' at binding %d set %d index %d is not bounded",
// glImage.name.c_str(), glImage.set, glImage.binding, u);
continue;
}
gpuTexture = gpuDescriptor->gpuTextureView->gpuTexture;
if (gpuTexture->size > 0) {
GLuint glTexture = gpuTexture->glTexture;
if (cache->glImages[unit] != glTexture) {
GL_CHECK(glBindImageTexture(unit, glTexture, 0, GL_TRUE, 0, glImage.glMemoryAccess, gpuTexture->glInternalFmt));
}
}
}
}
} // if
// bind vao
if (gpuInputAssembler && gpuPipelineState && gpuPipelineState->gpuShader &&
(isShaderChanged || gpuInputAssembler != gfxStateCache.gpuInputAssembler)) {
gfxStateCache.gpuInputAssembler = gpuInputAssembler;
if (USE_VAO) {
size_t hash = gpuPipelineState->gpuShader->glProgram ^ device->constantRegistry()->currentBoundThreadID;
GLuint glVAO = gpuInputAssembler->glVAOs[hash];
if (!glVAO) {
GL_CHECK(glGenVertexArrays(1, &glVAO));
gpuInputAssembler->glVAOs[hash] = glVAO;
GL_CHECK(glBindVertexArray(glVAO));
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, 0));
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0));
for (auto &gpuInput : gpuPipelineState->gpuShader->glInputs) {
for (size_t a = 0; a < gpuInputAssembler->attributes.size(); ++a) {
const GLES3GPUAttribute &gpuAttribute = gpuInputAssembler->glAttribs[a];
if (gpuAttribute.name == gpuInput.name) {
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, gpuAttribute.glBuffer));
for (uint32_t c = 0; c < gpuAttribute.componentCount; ++c) {
GLuint glLoc = gpuInput.glLoc + c;
uint32_t attribOffset = gpuAttribute.offset + gpuAttribute.size * c;
GL_CHECK(glEnableVertexAttribArray(glLoc));
cache->glEnabledAttribLocs[glLoc] = true;
GL_CHECK(glVertexAttribPointer(glLoc, gpuAttribute.count, gpuAttribute.glType, gpuAttribute.isNormalized, gpuAttribute.stride, BUFFER_OFFSET(attribOffset)));
GL_CHECK(glVertexAttribDivisor(glLoc, gpuAttribute.isInstanced ? 1 : 0));
}
break;
}
}
}
if (gpuInputAssembler->gpuIndexBuffer) {
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, gpuInputAssembler->gpuIndexBuffer->glBuffer));
}
GL_CHECK(glBindVertexArray(0));
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, 0));
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0));
cache->glVAO = 0;
cache->glArrayBuffer = 0;
cache->glElementArrayBuffer = 0;
}
if (cache->glVAO != glVAO) {
GL_CHECK(glBindVertexArray(glVAO));
cache->glVAO = glVAO;
}
} else {
for (auto &&glCurrentAttribLoc : cache->glCurrentAttribLocs) {
glCurrentAttribLoc = false;
}
for (auto &gpuInput : gpuPipelineState->gpuShader->glInputs) {
for (size_t a = 0; a < gpuInputAssembler->attributes.size(); ++a) {
const GLES3GPUAttribute &gpuAttribute = gpuInputAssembler->glAttribs[a];
if (gpuAttribute.name == gpuInput.name) {
if (cache->glArrayBuffer != gpuAttribute.glBuffer) {
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, gpuAttribute.glBuffer));
cache->glArrayBuffer = gpuAttribute.glBuffer;
}
for (uint32_t c = 0; c < gpuAttribute.componentCount; ++c) {
GLuint glLoc = gpuInput.glLoc + c;
uint32_t attribOffset = gpuAttribute.offset + gpuAttribute.size * c;
GL_CHECK(glEnableVertexAttribArray(glLoc));
cache->glCurrentAttribLocs[glLoc] = true;
cache->glEnabledAttribLocs[glLoc] = true;
GL_CHECK(glVertexAttribPointer(glLoc, gpuAttribute.count, gpuAttribute.glType, gpuAttribute.isNormalized, gpuAttribute.stride, BUFFER_OFFSET(attribOffset)));
GL_CHECK(glVertexAttribDivisor(glLoc, gpuAttribute.isInstanced ? 1 : 0));
}
break;
}
}
}
if (gpuInputAssembler->gpuIndexBuffer) {
if (cache->glElementArrayBuffer != gpuInputAssembler->gpuIndexBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, gpuInputAssembler->gpuIndexBuffer->glBuffer));
cache->glElementArrayBuffer = gpuInputAssembler->gpuIndexBuffer->glBuffer;
}
}
for (uint32_t a = 0; a < cache->glCurrentAttribLocs.size(); ++a) {
if (cache->glEnabledAttribLocs[a] != cache->glCurrentAttribLocs[a]) {
GL_CHECK(glDisableVertexAttribArray(a));
cache->glEnabledAttribLocs[a] = false;
}
}
}
} // if
if (gpuPipelineState && !gpuPipelineState->dynamicStates.empty()) {
for (DynamicStateFlagBit dynamicState : gpuPipelineState->dynamicStates) {
switch (dynamicState) {
case DynamicStateFlagBit::LINE_WIDTH:
if (cache->rs.lineWidth != dynamicStates->lineWidth) {
cache->rs.lineWidth = dynamicStates->lineWidth;
GL_CHECK(glLineWidth(dynamicStates->lineWidth));
}
break;
case DynamicStateFlagBit::DEPTH_BIAS:
if ((cache->rs.depthBias != dynamicStates->depthBiasConstant) ||
(cache->rs.depthBiasSlop != dynamicStates->depthBiasSlope)) {
GL_CHECK(glPolygonOffset(dynamicStates->depthBiasConstant, dynamicStates->depthBiasSlope));
cache->rs.depthBias = dynamicStates->depthBiasConstant;
cache->rs.depthBiasSlop = dynamicStates->depthBiasSlope;
}
break;
case DynamicStateFlagBit::BLEND_CONSTANTS:
if ((cache->bs.blendColor.x != dynamicStates->blendConstant.x) ||
(cache->bs.blendColor.y != dynamicStates->blendConstant.y) ||
(cache->bs.blendColor.z != dynamicStates->blendConstant.z) ||
(cache->bs.blendColor.w != dynamicStates->blendConstant.w)) {
GL_CHECK(glBlendColor(dynamicStates->blendConstant.x,
dynamicStates->blendConstant.y,
dynamicStates->blendConstant.z,
dynamicStates->blendConstant.w));
cache->bs.blendColor = dynamicStates->blendConstant;
}
break;
case DynamicStateFlagBit::STENCIL_WRITE_MASK: {
const auto &front = dynamicStates->stencilStatesFront;
const auto &back = dynamicStates->stencilStatesBack;
if (cache->dss.stencilWriteMaskFront != front.writeMask) {
GL_CHECK(glStencilMaskSeparate(GL_FRONT, front.writeMask));
cache->dss.stencilWriteMaskFront = front.writeMask;
}
if (cache->dss.stencilWriteMaskBack != back.writeMask) {
GL_CHECK(glStencilMaskSeparate(GL_BACK, back.writeMask));
cache->dss.stencilWriteMaskBack = back.writeMask;
}
} break;
case DynamicStateFlagBit::STENCIL_COMPARE_MASK: {
const auto &front = dynamicStates->stencilStatesFront;
const auto &back = dynamicStates->stencilStatesBack;
if ((cache->dss.stencilRefFront != front.reference) ||
(cache->dss.stencilReadMaskFront != front.compareMask)) {
GL_CHECK(glStencilFuncSeparate(GL_FRONT,
GLES3_CMP_FUNCS[toNumber(cache->dss.stencilFuncFront)],
front.reference,
front.compareMask));
cache->dss.stencilRefFront = front.reference;
cache->dss.stencilReadMaskFront = front.compareMask;
}
if ((cache->dss.stencilRefBack != back.reference) ||
(cache->dss.stencilReadMaskBack != back.compareMask)) {
GL_CHECK(glStencilFuncSeparate(GL_BACK,
GLES3_CMP_FUNCS[toNumber(cache->dss.stencilFuncBack)],
back.reference,
back.compareMask));
cache->dss.stencilRefBack = back.reference;
cache->dss.stencilReadMaskBack = back.compareMask;
}
} break;
default:
CC_LOG_ERROR("Invalid dynamic states.");
break;
} // switch
}
}
}
void cmdFuncGLES3Draw(GLES3Device *device, const DrawInfo &drawInfo) {
GLES3ObjectCache &gfxStateCache = device->stateCache()->gfxStateCache;
GLES3GPUPipelineState *gpuPipelineState = gfxStateCache.gpuPipelineState;
GLES3GPUInputAssembler *gpuInputAssembler = gfxStateCache.gpuInputAssembler;
GLenum glPrimitive = gfxStateCache.glPrimitive;
if (gpuInputAssembler && gpuPipelineState) {
if (!gpuInputAssembler->gpuIndirectBuffer) {
if (gpuInputAssembler->gpuIndexBuffer) {
if (drawInfo.indexCount > 0) {
uint8_t *offset = nullptr;
offset += drawInfo.firstIndex * gpuInputAssembler->gpuIndexBuffer->stride;
if (drawInfo.instanceCount == 0) {
GL_CHECK(glDrawElements(glPrimitive, drawInfo.indexCount, gpuInputAssembler->glIndexType, offset));
} else {
GL_CHECK(glDrawElementsInstanced(glPrimitive, drawInfo.indexCount, gpuInputAssembler->glIndexType, offset, drawInfo.instanceCount));
}
}
} else if (drawInfo.vertexCount > 0) {
if (drawInfo.instanceCount == 0) {
GL_CHECK(glDrawArrays(glPrimitive, drawInfo.firstVertex, drawInfo.vertexCount));
} else {
GL_CHECK(glDrawArraysInstanced(glPrimitive, drawInfo.firstVertex, drawInfo.vertexCount, drawInfo.instanceCount));
}
}
} else {
for (size_t j = 0; j < gpuInputAssembler->gpuIndirectBuffer->indirects.size(); ++j) {
const DrawInfo &draw = gpuInputAssembler->gpuIndirectBuffer->indirects[j];
if (gpuInputAssembler->gpuIndexBuffer) {
if (draw.indexCount > 0) {
uint8_t *offset = nullptr;
offset += draw.firstIndex * gpuInputAssembler->gpuIndexBuffer->stride;
if (draw.instanceCount == 0) {
GL_CHECK(glDrawElements(glPrimitive, draw.indexCount, gpuInputAssembler->glIndexType, offset));
} else {
GL_CHECK(glDrawElementsInstanced(glPrimitive, draw.indexCount, gpuInputAssembler->glIndexType, offset, draw.instanceCount));
}
}
} else if (draw.vertexCount > 0) {
if (draw.instanceCount == 0) {
GL_CHECK(glDrawArrays(glPrimitive, draw.firstVertex, draw.vertexCount));
} else {
GL_CHECK(glDrawArraysInstanced(glPrimitive, draw.firstVertex, draw.vertexCount, draw.instanceCount));
}
}
}
}
}
}
void cmdFuncGLES3Dispatch(GLES3Device *device, const GLES3GPUDispatchInfo &info) {
GLES3GPUStateCache *cache = device->stateCache();
if (info.indirectBuffer) {
if (cache->glDispatchIndirectBuffer != info.indirectBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_DISPATCH_INDIRECT_BUFFER, info.indirectBuffer->glBuffer));
cache->glDispatchIndirectBuffer = info.indirectBuffer->glBuffer;
}
GL_CHECK(glDispatchComputeIndirect(info.indirectOffset));
} else {
GL_CHECK(glDispatchCompute(info.groupCountX, info.groupCountY, info.groupCountZ));
}
}
void cmdFuncGLES3MemoryBarrier(GLES3Device * /*device*/, GLbitfield barriers, GLbitfield barriersByRegion) {
if (barriers) GL_CHECK(glMemoryBarrier(barriers));
if (barriersByRegion) GL_CHECK(glMemoryBarrierByRegion(barriersByRegion));
}
void cmdFuncGLES3InsertMarker(GLES3Device *device, GLsizei length, const char *marker) {
if (device->constantRegistry()->debugMarker) {
glInsertEventMarkerEXT(length, marker);
}
}
void cmdFuncGLES3PushGroupMarker(GLES3Device *device, GLsizei length, const char *marker) {
if (device->constantRegistry()->debugMarker) {
glPushGroupMarkerEXT(length, marker);
}
}
void cmdFuncGLES3PopGroupMarker(GLES3Device *device) {
if (device->constantRegistry()->debugMarker) {
glPopGroupMarkerEXT();
}
}
static void uploadBufferData(GLenum target, GLintptr offset, GLsizeiptr length, const void *buffer) {
#if 0
GL_CHECK(glBufferSubData(target, offset, length, buffer));
#else
void *dst{nullptr};
GL_CHECK(dst = glMapBufferRange(target, offset, length, GL_MAP_WRITE_BIT | GL_MAP_INVALIDATE_BUFFER_BIT));
if (!dst) {
GL_CHECK(glBufferSubData(target, offset, length, buffer));
return;
}
memcpy(dst, buffer, length);
GL_CHECK(glUnmapBuffer(target));
#endif
}
void cmdFuncGLES3UpdateBuffer(GLES3Device *device, GLES3GPUBuffer *gpuBuffer, const void *buffer, uint32_t offset, uint32_t size) {
GLES3ObjectCache &gfxStateCache = device->stateCache()->gfxStateCache;
if (hasFlag(gpuBuffer->usage, BufferUsageBit::INDIRECT)) {
memcpy(reinterpret_cast<uint8_t *>(gpuBuffer->indirects.data()) + offset, buffer, size);
} else if (hasFlag(gpuBuffer->usage, BufferUsageBit::TRANSFER_SRC) && gpuBuffer->buffer != nullptr) {
memcpy(gpuBuffer->buffer + offset, buffer, size);
} else {
switch (gpuBuffer->glTarget) {
case GL_ARRAY_BUFFER: {
if (device->stateCache()->glVAO) {
GL_CHECK(glBindVertexArray(0));
device->stateCache()->glVAO = 0;
}
gfxStateCache.gpuInputAssembler = nullptr;
if (device->stateCache()->glArrayBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_ARRAY_BUFFER, gpuBuffer->glBuffer));
device->stateCache()->glArrayBuffer = gpuBuffer->glBuffer;
}
uploadBufferData(GL_ARRAY_BUFFER, offset, size, buffer);
break;
}
case GL_ELEMENT_ARRAY_BUFFER: {
if (device->stateCache()->glVAO) {
GL_CHECK(glBindVertexArray(0));
device->stateCache()->glVAO = 0;
}
gfxStateCache.gpuInputAssembler = nullptr;
if (device->stateCache()->glElementArrayBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, gpuBuffer->glBuffer));
device->stateCache()->glElementArrayBuffer = gpuBuffer->glBuffer;
}
uploadBufferData(GL_ELEMENT_ARRAY_BUFFER, offset, size, buffer);
break;
}
case GL_UNIFORM_BUFFER: {
if (device->stateCache()->glUniformBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_UNIFORM_BUFFER, gpuBuffer->glBuffer));
device->stateCache()->glUniformBuffer = gpuBuffer->glBuffer;
}
uploadBufferData(GL_UNIFORM_BUFFER, offset, size, buffer);
break;
}
case GL_SHADER_STORAGE_BUFFER: {
if (device->stateCache()->glShaderStorageBuffer != gpuBuffer->glBuffer) {
GL_CHECK(glBindBuffer(GL_SHADER_STORAGE_BUFFER, gpuBuffer->glBuffer));
device->stateCache()->glShaderStorageBuffer = gpuBuffer->glBuffer;
}
uploadBufferData(GL_SHADER_STORAGE_BUFFER, offset, size, buffer);
break;
}
default:
CC_ABORT();
break;
}
}
}
uint8_t *funcGLES3PixelBufferPick(const uint8_t *buffer, Format format, uint32_t offset, Extent stride, Extent extent) {
const auto blockHeight = formatAlignment(format).second;
const uint32_t bufferSize = formatSize(format, extent.width, extent.height, extent.depth);
const uint32_t rowStride = formatSize(format, stride.width, 1, 1);
const uint32_t sliceStride = formatSize(format, stride.width, stride.height, 1);
const uint32_t chunkSize = formatSize(format, extent.width, 1, 1);
uint8_t *stagingBuffer = GLES3Device::getInstance()->getStagingBuffer(bufferSize);
uint32_t bufferOffset = 0;
uint32_t destOffset = 0;
for (uint32_t i = 0; i < extent.depth; i++) {
bufferOffset = offset + sliceStride * i;
for (uint32_t j = 0; j < extent.height; j += blockHeight) {
memcpy(stagingBuffer + destOffset, buffer + bufferOffset, chunkSize);
destOffset += chunkSize;
bufferOffset += rowStride;
}
}
return stagingBuffer;
}
void cmdFuncGLES3CopyBuffersToTexture(GLES3Device *device, const uint8_t *const *buffers, GLES3GPUTexture *gpuTexture, const BufferTextureCopy *regions, uint32_t count) {
if (!gpuTexture->memoryAllocated) return;
GLuint &glTexture = device->stateCache()->glTextures[device->stateCache()->texUint];
if (glTexture != gpuTexture->glTexture) {
glTexture = gpuTexture->glTexture;
}
bool isCompressed = GFX_FORMAT_INFOS[static_cast<int>(gpuTexture->format)].isCompressed;
uint32_t n = 0;
const auto blockSize = formatAlignment(gpuTexture->format);
uint32_t destWidth = 0;
uint32_t destHeight = 0;
auto target = GL_NONE;
switch (gpuTexture->type) {
case TextureType::TEX2D: {
Extent extent{};
Offset offset{};
Extent stride{};
target = GL_TEXTURE_2D;
GL_CHECK(glBindTexture(GL_TEXTURE_2D, gpuTexture->glTexture));
for (size_t i = 0; i < count; ++i) {
const BufferTextureCopy &region = regions[i];
uint32_t mipLevel = region.texSubres.mipLevel;
offset.x = region.texOffset.x == 0 ? 0 : utils::alignTo(region.texOffset.x, static_cast<int32_t>(blockSize.first));
offset.y = region.texOffset.y == 0 ? 0 : utils::alignTo(region.texOffset.y, static_cast<int32_t>(blockSize.second));
extent.width = utils::alignTo(region.texExtent.width, static_cast<uint32_t>(blockSize.first));
extent.height = utils::alignTo(region.texExtent.height, static_cast<uint32_t>(blockSize.second));
stride.width = region.buffStride > 0 ? region.buffStride : extent.width;
stride.height = region.buffTexHeight > 0 ? region.buffTexHeight : extent.height;
destWidth = (region.texExtent.width + offset.x == (gpuTexture->width >> mipLevel)) ? region.texExtent.width : extent.width;
destHeight = (region.texExtent.height + offset.y == (gpuTexture->height >> mipLevel)) ? region.texExtent.height : extent.height;
const uint8_t *buff;
if (stride.width == extent.width && stride.height == extent.height) {
buff = buffers[n++] + region.buffOffset;
} else {
buff = funcGLES3PixelBufferPick(buffers[n++], gpuTexture->format, region.buffOffset, stride, extent);
}
if (isCompressed) {
auto memSize = static_cast<GLsizei>(formatSize(gpuTexture->format, extent.width, extent.height, 1));
GL_CHECK(glCompressedTexSubImage2D(GL_TEXTURE_2D,
mipLevel,
offset.x, offset.y,
destWidth, destHeight,
gpuTexture->glFormat,
memSize,
(GLvoid *)buff));
} else if (hasFlag(gpuTexture->flags, TextureFlagBit::MUTABLE_STORAGE)) {
GL_CHECK(glTexImage2D(GL_TEXTURE_2D,
gpuTexture->mipLevel,
gpuTexture->glInternalFmt,
destWidth,
destHeight,
0,
gpuTexture->glFormat,
gpuTexture->glType,
(GLvoid *)buff));
} else {
GL_CHECK(glTexSubImage2D(GL_TEXTURE_2D,
mipLevel,
offset.x, offset.y,
destWidth, destHeight,
gpuTexture->glFormat,
gpuTexture->glType,
(GLvoid *)buff));
}
}
break;
}
case TextureType::TEX2D_ARRAY: {
Extent extent{};
Offset offset{};
Extent stride{};
target = GL_TEXTURE_2D_ARRAY;
GL_CHECK(glBindTexture(GL_TEXTURE_2D_ARRAY, gpuTexture->glTexture));
for (size_t i = 0; i < count; ++i) {
const BufferTextureCopy &region = regions[i];
uint32_t d = region.texSubres.layerCount;
uint32_t layerCount = d + region.texSubres.baseArrayLayer;
uint32_t mipLevel = region.texSubres.mipLevel;
offset.x = region.texOffset.x == 0 ? 0 : utils::alignTo(region.texOffset.x, static_cast<int32_t>(blockSize.first));
offset.y = region.texOffset.y == 0 ? 0 : utils::alignTo(region.texOffset.y, static_cast<int32_t>(blockSize.second));
extent.width = utils::alignTo(region.texExtent.width, static_cast<uint32_t>(blockSize.first));
extent.height = utils::alignTo(region.texExtent.height, static_cast<uint32_t>(blockSize.second));
extent.depth = 1; // 1 slice/loop
stride.width = region.buffStride > 0 ? region.buffStride : extent.width;
stride.height = region.buffTexHeight > 0 ? region.buffTexHeight : extent.height;
destWidth = (region.texExtent.width + offset.x == (gpuTexture->width >> mipLevel)) ? region.texExtent.width : extent.width;
destHeight = (region.texExtent.height + offset.y == (gpuTexture->height >> mipLevel)) ? region.texExtent.height : extent.height;
for (uint32_t z = region.texSubres.baseArrayLayer; z < layerCount; ++z) {
offset.z = static_cast<int32_t>(z);
const uint8_t *buff;
if (stride.width == extent.width && stride.height == extent.height) {
buff = buffers[n++] + region.buffOffset;
} else {
buff = funcGLES3PixelBufferPick(buffers[n++], gpuTexture->format, region.buffOffset, stride, extent);
}
if (isCompressed) {
auto memSize = static_cast<GLsizei>(formatSize(gpuTexture->format, extent.width, extent.height, 1));
GL_CHECK(glCompressedTexSubImage3D(GL_TEXTURE_2D_ARRAY,
mipLevel,
offset.x, offset.y, offset.z,
destWidth, destHeight, extent.depth,
gpuTexture->glFormat,
memSize,
(GLvoid *)buff));
} else {
GL_CHECK(glTexSubImage3D(GL_TEXTURE_2D_ARRAY,
mipLevel,
offset.x, offset.y, offset.z,
destWidth, destHeight, extent.depth,
gpuTexture->glFormat,
gpuTexture->glType,
(GLvoid *)buff));
}
}
}
break;
}
case TextureType::TEX3D: {
Extent extent{};
Offset offset{};
Extent stride{};
target = GL_TEXTURE_3D;
GL_CHECK(glBindTexture(GL_TEXTURE_3D, gpuTexture->glTexture));
for (size_t i = 0; i < count; ++i) {
const BufferTextureCopy &region = regions[i];
uint32_t mipLevel = region.texSubres.mipLevel;
offset.x = region.texOffset.x == 0 ? 0 : utils::alignTo(region.texOffset.x, static_cast<int32_t>(blockSize.first));
offset.y = region.texOffset.y == 0 ? 0 : utils::alignTo(region.texOffset.y, static_cast<int32_t>(blockSize.second));
offset.z = region.texOffset.z;
extent.width = utils::alignTo(region.texExtent.width, static_cast<uint32_t>(blockSize.first));
extent.height = utils::alignTo(region.texExtent.height, static_cast<uint32_t>(blockSize.second));
extent.depth = region.texExtent.depth;
stride.width = region.buffStride > 0 ? region.buffStride : extent.width;
stride.height = region.buffTexHeight > 0 ? region.buffTexHeight : extent.height;
destWidth = (region.texExtent.width + offset.x == (gpuTexture->width >> mipLevel)) ? region.texExtent.width : extent.width;
destHeight = (region.texExtent.height + offset.y == (gpuTexture->height >> mipLevel)) ? region.texExtent.height : extent.height;
const uint8_t *buff;
if (stride.width == extent.width && stride.height == extent.height) {
buff = buffers[n++] + region.buffOffset;
} else {
buff = funcGLES3PixelBufferPick(buffers[n++], gpuTexture->format, region.buffOffset, stride, extent);
}
if (isCompressed) {
auto memSize = static_cast<GLsizei>(formatSize(gpuTexture->format, extent.width, extent.height, extent.depth));
GL_CHECK(glCompressedTexSubImage3D(GL_TEXTURE_3D,
mipLevel,
offset.x, offset.y, offset.z,
destWidth, destHeight, extent.depth,
gpuTexture->glFormat,
memSize,
(GLvoid *)buff));
} else {
GL_CHECK(glTexSubImage3D(GL_TEXTURE_3D,
mipLevel,
offset.x, offset.y, offset.z,
destWidth, destHeight, extent.depth,
gpuTexture->glFormat,
gpuTexture->glType,
(GLvoid *)buff));
}
}
break;
}
case TextureType::CUBE: {
Extent extent{};
Offset offset{};
Extent stride{};
target = GL_TEXTURE_CUBE_MAP;
GL_CHECK(glBindTexture(GL_TEXTURE_CUBE_MAP, gpuTexture->glTexture));
for (size_t i = 0; i < count; ++i) {
const BufferTextureCopy &region = regions[i];
uint32_t mipLevel = region.texSubres.mipLevel;
offset.x = region.texOffset.x == 0 ? 0 : utils::alignTo(region.texOffset.x, static_cast<int32_t>(blockSize.first));
offset.y = region.texOffset.y == 0 ? 0 : utils::alignTo(region.texOffset.y, static_cast<int32_t>(blockSize.second));
offset.z = 1;
extent.width = utils::alignTo(region.texExtent.width, static_cast<uint32_t>(blockSize.first));
extent.height = utils::alignTo(region.texExtent.height, static_cast<uint32_t>(blockSize.second));
extent.depth = 1;
stride.width = region.buffStride > 0 ? region.buffStride : extent.width;
stride.height = region.buffTexHeight > 0 ? region.buffTexHeight : extent.height;
destWidth = (region.texExtent.width + offset.x == (gpuTexture->width >> mipLevel)) ? region.texExtent.width : extent.width;
destHeight = (region.texExtent.height + offset.y == (gpuTexture->height >> mipLevel)) ? region.texExtent.height : extent.height;
uint32_t faceCount = region.texSubres.baseArrayLayer + region.texSubres.layerCount;
for (uint32_t f = region.texSubres.baseArrayLayer; f < faceCount; ++f) {
const uint8_t *buff;
if (stride.width == extent.width && stride.height == extent.height) {
buff = buffers[n++] + region.buffOffset;
} else {
buff = funcGLES3PixelBufferPick(buffers[n++], gpuTexture->format, region.buffOffset, stride, extent);
}
if (isCompressed) {
auto memSize = static_cast<GLsizei>(formatSize(gpuTexture->format, extent.width, extent.height, 1));
GL_CHECK(glCompressedTexSubImage2D(GL_TEXTURE_CUBE_MAP_POSITIVE_X + f,
mipLevel,
offset.x, offset.y,
destWidth, destHeight,
gpuTexture->glFormat,
memSize,
(GLvoid *)buff));
} else {
GL_CHECK(glTexSubImage2D(GL_TEXTURE_CUBE_MAP_POSITIVE_X + f,
mipLevel,
offset.x, offset.y,
destWidth, destHeight,
gpuTexture->glFormat,
gpuTexture->glType,
(GLvoid *)buff));
}
}
}
break;
}
default:
CC_ABORT();
break;
}
if (!isCompressed && hasFlag(gpuTexture->flags, TextureFlagBit::GEN_MIPMAP) && target != GL_NONE) {
GL_CHECK(glBindTexture(target, gpuTexture->glTexture));
GL_CHECK(glGenerateMipmap(target));
}
}
void cmdFuncGLES3CopyTextureToBuffers(GLES3Device *device, GLES3GPUTextureView *gpuTextureView, uint8_t *const *buffers, const BufferTextureCopy *regions, uint32_t count) {
const auto *gpuTexture = gpuTextureView->gpuTexture;
auto glFormat = gpuTexture->glFormat;
auto glType = gpuTexture->glType;
for (uint32_t i = 0; i < count; ++i) {
auto region = regions[i];
uint8_t *copyDst = buffers[i];
auto framebuffer = device->framebufferCacheMap()->getFramebufferFromTexture(gpuTextureView, region.texSubres);
if (device->stateCache()->glReadFramebuffer != framebuffer) {
GL_CHECK(glBindFramebuffer(GL_READ_FRAMEBUFFER, framebuffer));
device->stateCache()->glReadFramebuffer = framebuffer;
}
GL_CHECK(glReadPixels(region.texOffset.x, region.texOffset.y, region.texExtent.width, region.texExtent.height, glFormat, glType, copyDst));
}
}
void cmdFuncGLES3BlitTexture(GLES3Device *device, GLES3GPUTextureView *gpuTextureSrcView, GLES3GPUTextureView *gpuTextureDstView,
const TextureBlit *regions, uint32_t count, Filter filter) {
GLES3GPUStateCache *cache = device->stateCache();
const auto *gpuTextureSrc = gpuTextureSrcView->gpuTexture;
const auto *gpuTextureDst = gpuTextureDstView->gpuTexture;
GLbitfield mask = getColorBufferMask(gpuTextureSrc->format);
for (uint32_t i = 0U; i < count; ++i) {
const TextureBlit &region = regions[i];
device->context()->makeCurrent(gpuTextureDst->swapchain, gpuTextureSrc->swapchain);
GLuint srcFramebuffer = 0;
if (gpuTextureSrc->swapchain) {
srcFramebuffer = gpuTextureSrc->swapchain->glFramebuffer;
} else {
srcFramebuffer = device->framebufferCacheMap()->getFramebufferFromTexture(gpuTextureSrcView, region.srcSubres);
}
if (cache->glReadFramebuffer != srcFramebuffer) {
GL_CHECK(glBindFramebuffer(GL_READ_FRAMEBUFFER, srcFramebuffer));
cache->glReadFramebuffer = srcFramebuffer;
}
GLuint dstFramebuffer = 0;
if (gpuTextureDst->swapchain) {
dstFramebuffer = gpuTextureDst->swapchain->glFramebuffer;
} else {
dstFramebuffer = device->framebufferCacheMap()->getFramebufferFromTexture(gpuTextureDstView, region.dstSubres);
}
if (cache->glDrawFramebuffer != dstFramebuffer) {
GL_CHECK(glBindFramebuffer(GL_DRAW_FRAMEBUFFER, dstFramebuffer));
cache->glDrawFramebuffer = dstFramebuffer;
}
ensureScissorRect(cache, region.dstOffset.x, region.dstOffset.y, region.dstExtent.width, region.dstExtent.height);
GL_CHECK(glBlitFramebuffer(
region.srcOffset.x,
region.srcOffset.y,
region.srcOffset.x + region.srcExtent.width,
region.srcOffset.y + region.srcExtent.height,
region.dstOffset.x,
region.dstOffset.y,
region.dstOffset.x + region.dstExtent.width,
region.dstOffset.y + region.dstExtent.height,
mask, GLES3_FILTERS[(uint32_t)filter]));
}
}
void cmdFuncGLES3ExecuteCmds(GLES3Device *device, GLES3CmdPackage *cmdPackage) {
if (!cmdPackage->cmds.size()) return;
static uint32_t cmdIndices[static_cast<int>(GLESCmdType::COUNT)] = {0};
memset(cmdIndices, 0, sizeof(cmdIndices));
for (uint32_t i = 0; i < cmdPackage->cmds.size(); ++i) {
GLESCmdType cmdType = cmdPackage->cmds[i];
uint32_t &cmdIdx = cmdIndices[static_cast<int>(cmdType)];
switch (cmdType) {
case GLESCmdType::BEGIN_RENDER_PASS: {
GLES3CmdBeginRenderPass *cmd = cmdPackage->beginRenderPassCmds[cmdIdx];
cmdFuncGLES3BeginRenderPass(device, cmd->gpuRenderPass, cmd->gpuFBO,
&cmd->renderArea, cmd->clearColors, cmd->clearDepth, cmd->clearStencil);
break;
}
case GLESCmdType::END_RENDER_PASS: {
cmdFuncGLES3EndRenderPass(device);
break;
}
case GLESCmdType::BIND_STATES: {
GLES3CmdBindStates *cmd = cmdPackage->bindStatesCmds[cmdIdx];
cmdFuncGLES3BindState(device, cmd->gpuPipelineState, cmd->gpuInputAssembler,
cmd->gpuDescriptorSets.data(), cmd->dynamicOffsets.data(), &cmd->dynamicStates);
break;
}
case GLESCmdType::DRAW: {
GLES3CmdDraw *cmd = cmdPackage->drawCmds[cmdIdx];
cmdFuncGLES3Draw(device, cmd->drawInfo);
break;
}
case GLESCmdType::DISPATCH: {
GLES3CmdDispatch *cmd = cmdPackage->dispatchCmds[cmdIdx];
cmdFuncGLES3Dispatch(device, cmd->dispatchInfo);
break;
}
case GLESCmdType::BARRIER: {
GLES3CmdBarrier *cmd = cmdPackage->barrierCmds[cmdIdx];
cmdFuncGLES3MemoryBarrier(device, cmd->barriers, cmd->barriersByRegion);
break;
}
case GLESCmdType::UPDATE_BUFFER: {
GLES3CmdUpdateBuffer *cmd = cmdPackage->updateBufferCmds[cmdIdx];
cmdFuncGLES3UpdateBuffer(device, cmd->gpuBuffer, cmd->buffer, cmd->offset, cmd->size);
break;
}
case GLESCmdType::COPY_BUFFER_TO_TEXTURE: {
GLES3CmdCopyBufferToTexture *cmd = cmdPackage->copyBufferToTextureCmds[cmdIdx];
cmdFuncGLES3CopyBuffersToTexture(device, cmd->buffers, cmd->gpuTexture, cmd->regions, cmd->count);
break;
}
case GLESCmdType::BLIT_TEXTURE: {
GLES3CmdBlitTexture *cmd = cmdPackage->blitTextureCmds[cmdIdx];
cmdFuncGLES3BlitTexture(device, cmd->gpuTextureSrcView, cmd->gpuTextureDstView, cmd->regions, cmd->count, cmd->filter);
break;
}
case GLESCmdType::QUERY: {
GLES3CmdQuery *cmd = cmdPackage->queryCmds[cmdIdx];
cmdFuncGLES3Query(device, cmd->queryPool, cmd->type, cmd->id);
break;
}
default:
break;
}
cmdIdx++;
}
}
void GLES3GPUFramebufferObject::initialize(GLES3GPUSwapchain *swc) {
swapchain = swc;
}
void GLES3GPUFramebufferObject::bindColor(const GLES3GPUTextureView *texture, uint32_t colorIndex, const ColorAttachment &attachment) {
bindColorMultiSample(texture, colorIndex, 1, attachment);
}
void GLES3GPUFramebufferObject::bindColorMultiSample(const GLES3GPUTextureView *texture, uint32_t colorIndex, GLint samples, const ColorAttachment &attachment) {
if (colorIndex >= colors.size()) {
colors.resize(colorIndex + 1);
}
bool isDefaultFb = swapchain != nullptr && !swapchain->isXR;
if (attachment.loadOp == LoadOp::DISCARD) {
loadInvalidates.emplace_back(isDefaultFb ? GL_COLOR : GL_COLOR_ATTACHMENT0 + colorIndex);
}
if (attachment.storeOp == StoreOp::DISCARD) {
storeInvalidates.emplace_back(isDefaultFb ? GL_COLOR : GL_COLOR_ATTACHMENT0 + colorIndex);
}
colors[colorIndex] = {texture, samples};
}
void GLES3GPUFramebufferObject::bindDepthStencil(const GLES3GPUTextureView *texture, const DepthStencilAttachment &attachment) {
bindDepthStencilMultiSample(texture, 1, attachment);
}
void GLES3GPUFramebufferObject::bindDepthStencilMultiSample(const GLES3GPUTextureView *texture, GLint samples, const DepthStencilAttachment &attachment) {
const FormatInfo &info = GFX_FORMAT_INFOS[toNumber(texture->gpuTexture->format)];
bool isDefaultFb = swapchain != nullptr && !swapchain->isXR;
bool hasDepth = info.hasDepth;
bool hasStencil = info.hasStencil;
dsAttachment = hasDepth && hasStencil ? GL_DEPTH_STENCIL_ATTACHMENT :
hasDepth ? GL_DEPTH_ATTACHMENT : GL_STENCIL_ATTACHMENT;
if (hasDepth) {
auto att = isDefaultFb ? GL_DEPTH : GL_DEPTH_ATTACHMENT;
if (attachment.depthLoadOp == LoadOp::DISCARD) {
loadInvalidates.emplace_back(att);
}
if (attachment.depthStoreOp == StoreOp::DISCARD) {
storeInvalidates.emplace_back(att);
}
}
if (hasStencil) {
auto att = isDefaultFb ? GL_STENCIL : GL_STENCIL_ATTACHMENT;
if (attachment.stencilLoadOp == LoadOp::DISCARD) {
loadInvalidates.emplace_back(att);
}
if (attachment.stencilStoreOp == StoreOp::DISCARD) {
storeInvalidates.emplace_back(att);
}
}
depthStencil.first = texture;
depthStencil.second = samples;
}
bool GLES3GPUFramebufferObject::isActive() const {
return swapchain != nullptr || (handle != 0);
}
void GLES3GPUFramebufferObject::processLoad(GLenum target) {
if (!loadInvalidates.empty()) {
GL_CHECK(glInvalidateFramebuffer(target, utils::toUint(loadInvalidates.size()), loadInvalidates.data()));
}
}
void GLES3GPUFramebufferObject::processStore(GLenum target) {
if (!storeInvalidates.empty()) {
GL_CHECK(glInvalidateFramebuffer(target, utils::toUint(storeInvalidates.size()), storeInvalidates.data()));
}
}
void GLES3GPUFramebufferObject::finalize(GLES3GPUStateCache *cache) {
if (swapchain != nullptr) {
return;
}
if (colors.empty() && dsAttachment == GL_NONE) {
return;
}
GL_CHECK(glGenFramebuffers(1, &handle));
GL_CHECK(glBindFramebuffer(GL_DRAW_FRAMEBUFFER, handle));
cache->glDrawFramebuffer = handle;
auto bindAttachment = [](GLenum attachment, GLint samples, const GLES3GPUTextureView *view) {
auto *texture = view->gpuTexture;
if (samples > 1) {
CC_ASSERT(view->gpuTexture->glTexture != 0);
GL_CHECK(glFramebufferTexture2DMultisampleEXT(GL_FRAMEBUFFER,
attachment,
GL_TEXTURE_2D,
texture->glTexture,
view->baseLevel,
static_cast<GLsizei>(samples)));
return;
}
if (texture->useRenderBuffer) {
/*
* Renderbuffer is not allocated if using lazily allocated flag.
* If the attachment does not meet the implicit MS condition, renderbuffer should be allocated here.
*/
if (view->gpuTexture->glRenderbuffer == 0) {
GL_CHECK(glGenRenderbuffers(1, &view->gpuTexture->glRenderbuffer));
renderBufferStorage(GLES3Device::getInstance(), texture);
}
GL_CHECK(glFramebufferRenderbuffer(GL_DRAW_FRAMEBUFFER, attachment, GL_RENDERBUFFER, texture->glRenderbuffer));
} else {
GL_CHECK(glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, attachment, view->glTarget, texture->glTexture, view->baseLevel));
}
};
ccstd::vector<GLenum> drawBuffers(colors.size(), GL_NONE);
for (uint32_t i = 0; i < colors.size(); ++i) {
const auto &[view, samples] = colors[i];
auto att = static_cast<GLenum>(GL_COLOR_ATTACHMENT0 + i);
drawBuffers[i] = att;
bindAttachment(att, samples, view);
}
if (depthStencil.first != nullptr) {
const auto &[view, samples] = depthStencil;
bindAttachment(dsAttachment, samples, view);
}
if (!drawBuffers.empty()) {
GL_CHECK(glDrawBuffers(drawBuffers.size(), drawBuffers.data()));
}
GLenum status;
GL_CHECK(status = glCheckFramebufferStatus(GL_DRAW_FRAMEBUFFER));
if (status != GL_FRAMEBUFFER_COMPLETE) {
switch (status) {
case GL_FRAMEBUFFER_INCOMPLETE_ATTACHMENT:
CC_LOG_ERROR("checkFramebufferStatus() - FRAMEBUFFER_INCOMPLETE_ATTACHMENT");
break;
case GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT:
CC_LOG_ERROR("checkFramebufferStatus() - FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT");
break;
case GL_FRAMEBUFFER_INCOMPLETE_DIMENSIONS:
CC_LOG_ERROR("checkFramebufferStatus() - FRAMEBUFFER_INCOMPLETE_DIMENSIONS");
break;
case GL_FRAMEBUFFER_INCOMPLETE_MULTISAMPLE:
CC_LOG_ERROR("checkFramebufferStatus() - FRAMEBUFFER_INCOMPLETE_MULTISAMPLE");
break;
case GL_FRAMEBUFFER_UNSUPPORTED:
CC_LOG_ERROR("checkFramebufferStatus() - FRAMEBUFFER_UNSUPPORTED");
break;
default:
CC_LOG_ERROR("checkFramebufferStatus() - %x", status);
break;
}
}
}
void GLES3GPUFramebufferObject::destroy(GLES3GPUStateCache *cache, GLES3GPUFramebufferCacheMap *framebufferCacheMap) {
if (swapchain) {
swapchain = nullptr;
} else {
if (cache->glDrawFramebuffer == handle) {
GL_CHECK(glBindFramebuffer(GL_DRAW_FRAMEBUFFER, 0));
cache->glDrawFramebuffer = 0;
}
GL_CHECK(glDeleteFramebuffers(1, &handle));
framebufferCacheMap->unregisterExternal(handle);
handle = 0U;
}
}
void GLES3GPUFramebufferHub::update(GLES3GPUTexture *texture) {
auto &pool = _framebuffers[texture];
for (auto *framebuffer : pool) {
cmdFuncGLES3DestroyFramebuffer(GLES3Device::getInstance(), framebuffer);
cmdFuncGLES3CreateFramebuffer(GLES3Device::getInstance(), framebuffer);
}
}
GLint cmdFuncGLES3GetMaxSampleCount(const GLES3Device *device, Format format, TextureUsage usage, TextureFlags flags) {
std::ignore = flags;
auto internalFormat = mapGLInternalFormat(format);
auto target = useRenderBuffer(device, format, usage) ? GL_RENDERBUFFER : GL_TEXTURE_2D_MULTISAMPLE;
GLint maxSamples = 1;
GL_CHECK(glGetInternalformativ(target, internalFormat, GL_SAMPLES, 1, &maxSamples));
return maxSamples;
}
} // namespace gfx
} // namespace cc