Scarfski/bgfx
1
0
Fork 0
mirror of https://github.com/bkaradzic/bgfx.git synced 2026-10-26 02:35:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

Implemented BGFX_SAMPLER_SAMPLE_STENCIL in all backends. (#3884)

Browse Source
This commit is contained in:
Branimir Karadžić
2026-08-13 15:38:58 -07:00
committed by GitHub
parent 3d3b08778e
commit 56ef37bcf8
12 changed files with 465 additions and 82 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/renderer_gl.cpp
+217 -28
View File
@@ -636,6 +636,7 @@ namespace bgfx { namespace gl
ARB_shader_storage_buffer_object,
ARB_shader_texture_lod,
ARB_shader_viewport_layer_array,
ARB_stencil_texturing,
ARB_texture_compression_bptc,
ARB_texture_compression_rgtc,
ARB_texture_cube_map_array,
@@ -857,6 +858,7 @@ namespace bgfx { namespace gl
{ "ARB_shader_storage_buffer_object", BGFX_CONFIG_RENDERER_OPENGL >= 43, true },
{ "ARB_shader_texture_lod", BGFX_CONFIG_RENDERER_OPENGL >= 30, true },
{ "ARB_shader_viewport_layer_array", false, true },
{ "ARB_stencil_texturing", BGFX_CONFIG_RENDERER_OPENGL >= 43, true },
{ "ARB_texture_compression_bptc", BGFX_CONFIG_RENDERER_OPENGL >= 44, true },
{ "ARB_texture_compression_rgtc", BGFX_CONFIG_RENDERER_OPENGL >= 30, true },
{ "ARB_texture_cube_map_array", BGFX_CONFIG_RENDERER_OPENGL >= 40, true },
@@ -2294,6 +2296,7 @@ namespace bgfx { namespace gl
, m_blitSupported(false)
, m_blitFboSupported(false)
, m_textureViewSupported(false)
, m_depthStencilTexturingSupported(false)
, m_readBackSupported(BX_ENABLED(BGFX_CONFIG_RENDERER_OPENGL) )
, m_readBackFboSupported(false)
, m_vaoSupport(false)
@@ -2959,6 +2962,12 @@ namespace bgfx { namespace gl
m_textureViewSupported = NULL != glTextureView;
}
m_depthStencilTexturingSupported = false
|| s_extension[Extension::ARB_stencil_texturing].m_supported
|| BX_ENABLED(BGFX_CONFIG_RENDERER_OPENGL >= 43)
|| BX_ENABLED(BGFX_CONFIG_RENDERER_OPENGLES >= 31)
;
m_blitFboSupported = !m_blitSupported && BX_ENABLED(BGFX_CONFIG_RENDERER_OPENGLES);
m_readBackFboSupported = !m_readBackSupported;
@@ -5022,6 +5031,7 @@ namespace bgfx { namespace gl
bool m_blitSupported;
bool m_blitFboSupported;
bool m_textureViewSupported;
bool m_depthStencilTexturingSupported;
bool m_readBackSupported;
bool m_readBackFboSupported;
bool m_vaoSupport;
@@ -5144,6 +5154,13 @@ namespace bgfx { namespace gl
GLSL_TYPE(GL_INT_SAMPLER_CUBE);
GLSL_TYPE(GL_UNSIGNED_INT_SAMPLER_CUBE);
GLSL_TYPE(GL_SAMPLER_CUBE_SHADOW);
GLSL_TYPE(GL_SAMPLER_CUBE_MAP_ARRAY_SHADOW);
GLSL_TYPE(GL_SAMPLER_CUBE_MAP_ARRAY);
GLSL_TYPE(GL_INT_SAMPLER_CUBE_MAP_ARRAY);
GLSL_TYPE(GL_UNSIGNED_INT_SAMPLER_CUBE_MAP_ARRAY);
GLSL_TYPE(GL_IMAGE_1D);
GLSL_TYPE(GL_INT_IMAGE_1D);
GLSL_TYPE(GL_UNSIGNED_INT_IMAGE_1D);
@@ -5160,6 +5177,10 @@ namespace bgfx { namespace gl
GLSL_TYPE(GL_IMAGE_CUBE);
GLSL_TYPE(GL_INT_IMAGE_CUBE);
GLSL_TYPE(GL_UNSIGNED_INT_IMAGE_CUBE);
GLSL_TYPE(GL_IMAGE_CUBE_MAP_ARRAY);
GLSL_TYPE(GL_INT_IMAGE_CUBE_MAP_ARRAY);
GLSL_TYPE(GL_UNSIGNED_INT_IMAGE_CUBE_MAP_ARRAY);
}
#undef GLSL_TYPE
@@ -5242,6 +5263,13 @@ namespace bgfx { namespace gl
case GL_INT_SAMPLER_CUBE:
case GL_UNSIGNED_INT_SAMPLER_CUBE:
case GL_SAMPLER_CUBE_SHADOW:
case GL_SAMPLER_CUBE_MAP_ARRAY_SHADOW:
case GL_SAMPLER_CUBE_MAP_ARRAY:
case GL_INT_SAMPLER_CUBE_MAP_ARRAY:
case GL_UNSIGNED_INT_SAMPLER_CUBE_MAP_ARRAY:
case GL_IMAGE_1D:
case GL_INT_IMAGE_1D:
case GL_UNSIGNED_INT_IMAGE_1D:
@@ -5258,6 +5286,10 @@ namespace bgfx { namespace gl
case GL_IMAGE_CUBE:
case GL_INT_IMAGE_CUBE:
case GL_UNSIGNED_INT_IMAGE_CUBE:
case GL_IMAGE_CUBE_MAP_ARRAY:
case GL_INT_IMAGE_CUBE_MAP_ARRAY:
case GL_UNSIGNED_INT_IMAGE_CUBE_MAP_ARRAY:
return UniformType::Sampler;
};
@@ -5497,6 +5529,13 @@ namespace bgfx { namespace gl
case GL_INT_SAMPLER_CUBE:
case GL_UNSIGNED_INT_SAMPLER_CUBE:
case GL_SAMPLER_CUBE_SHADOW:
case GL_SAMPLER_CUBE_MAP_ARRAY_SHADOW:
case GL_SAMPLER_CUBE_MAP_ARRAY:
case GL_INT_SAMPLER_CUBE_MAP_ARRAY:
case GL_UNSIGNED_INT_SAMPLER_CUBE_MAP_ARRAY:
case GL_IMAGE_1D:
case GL_INT_IMAGE_1D:
case GL_UNSIGNED_INT_IMAGE_1D:
@@ -5512,6 +5551,10 @@ namespace bgfx { namespace gl
case GL_IMAGE_CUBE:
case GL_INT_IMAGE_CUBE:
case GL_UNSIGNED_INT_IMAGE_CUBE:
case GL_IMAGE_CUBE_MAP_ARRAY:
case GL_INT_IMAGE_CUBE_MAP_ARRAY:
case GL_UNSIGNED_INT_IMAGE_CUBE_MAP_ARRAY:
if (m_numSamplers < BX_COUNTOF(m_sampler) )
{
BX_TRACE("Sampler #%d at location %d.", m_numSamplers, loc);
@@ -5777,6 +5820,8 @@ namespace bgfx { namespace gl
m_height = _height;
m_depth = _depth;
m_currentSamplerHash = UINT32_MAX;
m_baseLevel = 0;
m_maxLevel = -1;
const bool writeOnly = 0 != (m_flags&BGFX_TEXTURE_RT_WRITE_ONLY);
const bool computeWrite = 0 != (m_flags&BGFX_TEXTURE_COMPUTE_WRITE );
@@ -5850,7 +5895,7 @@ namespace bgfx { namespace gl
, internalFmt
, m_width
, m_height
, _depth
, GL_TEXTURE_CUBE_MAP_ARRAY == _target ? _depth*6 : _depth
) );
m_immutableStorage = true;
}
@@ -6269,11 +6314,18 @@ namespace bgfx { namespace gl
GL_CHECK(glBindTexture(m_target, m_id) );
GL_CHECK(glPixelStorei(GL_UNPACK_ALIGNMENT, 1) );
GLenum target = isCubeMap()
? GL_TEXTURE_CUBE_MAP_POSITIVE_X
const bool cubeArray = GL_TEXTURE_CUBE_MAP_ARRAY == m_target;
GLenum target = isCubeMap() && !cubeArray
? GLenum(GL_TEXTURE_CUBE_MAP_POSITIVE_X + _side)
: m_target
;
const uint16_t zOffset = cubeArray
? uint16_t(_z*6 + _side)
: _z
;
const bool swizzle = true
&& TextureFormat::BGRA8 == m_requestedFormat
&& !s_textureFormat[m_requestedFormat].m_supported
@@ -6333,11 +6385,11 @@ namespace bgfx { namespace gl
? s_textureFormat[m_textureFormat].m_internalFmtSrgb
: s_textureFormat[m_textureFormat].m_internalFmt
;
GL_CHECK(compressedTexSubImage(target+_side
GL_CHECK(compressedTexSubImage(target
, _mip
, rect.m_x
, rect.m_y
, _z
, zOffset
, rect.m_width
, rect.m_height
, _depth
@@ -6365,11 +6417,11 @@ namespace bgfx { namespace gl
data = temp;
}
GL_CHECK(texSubImage(target+_side
GL_CHECK(texSubImage(target
, _mip
, rect.m_x
, rect.m_y
, _z
, zOffset
, rect.m_width
, rect.m_height
, _depth
@@ -6440,6 +6492,7 @@ namespace bgfx { namespace gl
&& (BX_ENABLED(BGFX_CONFIG_RENDERER_OPENGL) || s_renderGL->m_gles3 || s_extension[Extension::APPLE_texture_max_level].m_supported) )
{
GL_CHECK(glTexParameteri(targetMsaa, GL_TEXTURE_MAX_LEVEL, numMips-1) );
m_maxLevel = -1;
}
if (target == GL_TEXTURE_3D)
@@ -6489,13 +6542,48 @@ namespace bgfx { namespace gl
}
}
GLuint TextureGL::getViewId(uint8_t _firstMip, uint8_t _numMips, uint16_t _firstLayer, uint16_t _numLayers)
GLenum TextureGL::getViewTarget(uint16_t _numLayers, bool _layered) const
{
const uint8_t firstMip = bx::min<uint8_t>(_firstMip, uint8_t(m_numMips - 1) );
const uint8_t numMips = bx::min<uint8_t>(_numMips, uint8_t(m_numMips - firstMip) );
const uint16_t numLayers0 = uint16_t(bx::max<uint32_t>(m_numLayers, 1) );
const uint16_t firstLayer = bx::min<uint16_t>(_firstLayer, uint16_t(numLayers0 - 1) );
const uint16_t numLayers = bx::min<uint16_t>(_numLayers, uint16_t(numLayers0 - firstLayer) );
switch (m_target)
{
case GL_TEXTURE_2D_ARRAY:
case GL_TEXTURE_CUBE_MAP:
case GL_TEXTURE_CUBE_MAP_ARRAY:
if (_layered)
{
return GL_TEXTURE_2D_ARRAY;
}
if (1 == _numLayers)
{
return GL_TEXTURE_2D;
}
if (!isCubeMap() )
{
return GL_TEXTURE_2D_ARRAY;
}
return 6 == _numLayers
? GL_TEXTURE_CUBE_MAP
: GL_TEXTURE_CUBE_MAP_ARRAY
;
default:
break;
}
return m_target;
}
GLuint TextureGL::getViewId(uint8_t _firstMip, uint8_t _numMips, uint16_t _firstLayer, uint16_t _numLayers, GLenum* _target, bool _layered)
{
const uint16_t facesPerLayer = isCubeMap() ? 6 : 1;
const uint8_t firstMip = bx::min<uint8_t>(_firstMip, uint8_t(m_numMips - 1) );
const uint8_t numMips = bx::min<uint8_t>(_numMips, uint8_t(m_numMips - firstMip) );
const uint16_t numLayers0 = uint16_t(bx::max<uint32_t>(m_numLayers, 1) * facesPerLayer);
const uint16_t firstLayer = bx::min<uint16_t>(_firstLayer, uint16_t(numLayers0 - 1) );
const uint16_t numLayers = bx::min<uint16_t>(_numLayers, uint16_t(numLayers0 - firstLayer) );
const bool fullRange = 0 == firstMip
&& numMips >= m_numMips
@@ -6503,19 +6591,38 @@ namespace bgfx { namespace gl
&& numLayers >= numLayers0
;
const GLenum target = getViewTarget(numLayers, _layered);
const bool cubeAligned = GL_TEXTURE_2D == target
|| GL_TEXTURE_2D_ARRAY == target
|| (0 == firstLayer%facesPerLayer && 0 == numLayers%facesPerLayer)
;
if (fullRange
|| !s_renderGL->m_textureViewSupported
|| !m_immutableStorage
|| GL_ZERO == m_internalFmt
|| !cubeAligned
|| NULL == glTextureView)
{
if (NULL != _target)
{
*_target = m_target;
}
return m_id;
}
if (NULL != _target)
{
*_target = target;
}
const uint16_t parent = uint16_t(this - s_renderGL->m_textures);
const uint64_t key = 0
| (uint64_t(parent) << 48)
| (uint64_t(_layered) << 47)
| (uint64_t(firstMip) << 40)
| (uint64_t(numMips) << 32)
| (uint64_t(firstLayer) << 16)
@@ -6531,7 +6638,7 @@ namespace bgfx { namespace gl
GL_CHECK(glGenTextures(1, &viewId) );
s_renderGL->m_textureViewStateCache.add(key, TextureViewGL{viewId}, parent);
GL_CHECK(glTextureView(viewId
, m_target
, target
, m_id
, m_internalFmt
, firstMip
@@ -6551,10 +6658,28 @@ namespace bgfx { namespace gl
;
const uint32_t index = (flags & BGFX_SAMPLER_BORDER_COLOR_MASK) >> BGFX_SAMPLER_BORDER_COLOR_SHIFT;
const GLuint id = getViewId(_firstMip, _numMips, _firstLayer, _numLayers);
GLenum target = m_target;
const GLuint id = getViewId(_firstMip, _numMips, _firstLayer, _numLayers, &target);
GL_CHECK(glActiveTexture(GL_TEXTURE0+_stage) );
GL_CHECK(glBindTexture(m_target, id) );
GL_CHECK(glBindTexture(target, id) );
m_depthStencilTexturing |= 0 != (flags & BGFX_SAMPLER_SAMPLE_STENCIL);
if (m_depthStencilTexturing)
{
BX_WARN(s_renderGL->m_depthStencilTexturingSupported
, "BGFX_SAMPLER_SAMPLE_STENCIL requires OpenGL 4.3, OpenGL ES 3.1 or GL_ARB_stencil_texturing; sampling depth instead."
);
if (s_renderGL->m_depthStencilTexturingSupported)
{
GL_CHECK(glTexParameteri(target, GL_DEPTH_STENCIL_TEXTURE_MODE, 0 != (flags & BGFX_SAMPLER_SAMPLE_STENCIL)
? GL_STENCIL_INDEX
: GL_DEPTH_COMPONENT
) );
}
}
if (s_renderGL->m_samplerObjectSupport)
{
@@ -6564,6 +6689,26 @@ namespace bgfx { namespace gl
{
setSamplerState(flags, _palette[index]);
}
if (id == m_id
&& 1 < m_numMips
&& GL_TEXTURE_2D_MULTISAMPLE != m_target
&& (BX_ENABLED(BGFX_CONFIG_RENDERER_OPENGL >= 30) || s_renderGL->m_gles3) )
{
const uint8_t firstMip = bx::min<uint8_t>(_firstMip, uint8_t(m_numMips - 1) );
const uint8_t numMips = bx::min<uint8_t>(_numMips, uint8_t(m_numMips - firstMip) );
const int32_t baseLevel = int32_t(firstMip);
const int32_t maxLevel = int32_t(firstMip + numMips - 1);
if (m_baseLevel != baseLevel
|| m_maxLevel != maxLevel)
{
GL_CHECK(glTexParameteri(m_target, GL_TEXTURE_BASE_LEVEL, baseLevel) );
GL_CHECK(glTexParameteri(m_target, GL_TEXTURE_MAX_LEVEL, maxLevel) );
m_baseLevel = baseLevel;
m_maxLevel = maxLevel;
}
}
}
void TextureGL::resolve(uint8_t _resolve) const
@@ -7557,8 +7702,10 @@ namespace bgfx { namespace gl
}
else
{
if (1 < texture.m_numLayers
&& !texture.isCubeMap() )
const bool cubeArray = GL_TEXTURE_CUBE_MAP_ARRAY == texture.m_target;
if ( (1 < texture.m_numLayers && !texture.isCubeMap() )
|| cubeArray)
{
if (1 < at.numLayers)
{
@@ -8392,12 +8539,31 @@ namespace bgfx { namespace gl
case Binding::Image:
{
const TextureGL& texture = m_textures[bind.m_idx];
TextureGL& texture = m_textures[bind.m_idx];
GLuint id = texture.m_id;
GLint level = bind.m_firstMip;
GLint layer = bind.m_firstLayer;
GLboolean layered = texture.isLayered() || UINT16_MAX != bind.m_numLayers ? GL_TRUE : GL_FALSE;
if (UINT16_MAX != bind.m_numLayers
&& texture.isLayered() )
{
const GLuint viewId = texture.getViewId(bind.m_firstMip, 1, bind.m_firstLayer, bind.m_numLayers, NULL, true);
if (viewId != texture.m_id)
{
id = viewId;
level = 0;
layer = 0;
}
}
GL_CHECK(glBindImageTexture(ii
, texture.m_id
, bind.m_firstMip
, texture.isLayered() || UINT16_MAX != bind.m_numLayers ? GL_TRUE : GL_FALSE
, bind.m_firstLayer
, id
, level
, layered
, layer
, s_access[bind.m_access]
, s_imageFormat[bind.m_format])
);
@@ -8920,6 +9086,10 @@ namespace bgfx { namespace gl
if (current.m_idx != bind.m_idx
|| current.m_type != bind.m_type
|| current.m_samplerFlags != bind.m_samplerFlags
|| current.m_firstLayer != bind.m_firstLayer
|| current.m_numLayers != bind.m_numLayers
|| current.m_firstMip != bind.m_firstMip
|| current.m_numMips != bind.m_numMips
|| programChanged)
{
if (kInvalidHandle != bind.m_idx)
@@ -8928,12 +9098,31 @@ namespace bgfx { namespace gl
{
case Binding::Image:
{
const TextureGL& texture = m_textures[bind.m_idx];
TextureGL& texture = m_textures[bind.m_idx];
GLuint id = texture.m_id;
GLint level = bind.m_firstMip;
GLint layer = bind.m_firstLayer;
GLboolean layered = texture.isLayered() || UINT16_MAX != bind.m_numLayers ? GL_TRUE : GL_FALSE;
if (UINT16_MAX != bind.m_numLayers
&& texture.isLayered() )
{
const GLuint viewId = texture.getViewId(bind.m_firstMip, 1, bind.m_firstLayer, bind.m_numLayers, NULL, true);
if (viewId != texture.m_id)
{
id = viewId;
level = 0;
layer = 0;
}
}
GL_CHECK(glBindImageTexture(stage
, texture.m_id
, bind.m_firstMip
, texture.isLayered() || UINT16_MAX != bind.m_numLayers ? GL_TRUE : GL_FALSE
, bind.m_firstLayer
, id
, level
, layered
, layer
, s_access[bind.m_access]
, s_imageFormat[bind.m_format])
);