1.. _screen: 2 3Screen 4====== 5 6A screen is an object representing the context-independent part of a device. 7 8Flags and enumerations 9---------------------- 10 11XXX some of these don't belong in this section. 12 13 14.. _pipe_cap: 15 16PIPE_CAP_* 17^^^^^^^^^^ 18 19Capability queries return information about the features and limits of the 20driver/GPU. For floating-point values, use :ref:`get_paramf`, and for boolean 21or integer values, use :ref:`get_param`. 22 23The integer capabilities: 24 25* ``PIPE_CAP_GRAPHICS``: Whether graphics is supported. If not, contexts can 26 only be created with PIPE_CONTEXT_COMPUTE_ONLY. 27* ``PIPE_CAP_NPOT_TEXTURES``: Whether :term:`NPOT` textures may have repeat modes, 28 normalized coordinates, and mipmaps. 29* ``PIPE_CAP_MAX_DUAL_SOURCE_RENDER_TARGETS``: How many dual-source blend RTs are support. 30 :ref:`Blend` for more information. 31* ``PIPE_CAP_ANISOTROPIC_FILTER``: Whether textures can be filtered anisotropically. 32* ``PIPE_CAP_POINT_SPRITE``: Whether point sprites are available. 33* ``PIPE_CAP_MAX_RENDER_TARGETS``: The maximum number of render targets that may be 34 bound. 35* ``PIPE_CAP_OCCLUSION_QUERY``: Whether occlusion queries are available. 36* ``PIPE_CAP_QUERY_TIME_ELAPSED``: Whether PIPE_QUERY_TIME_ELAPSED queries are available. 37* ``PIPE_CAP_TEXTURE_SHADOW_MAP``: indicates whether the fragment shader hardware 38 can do the depth texture / Z comparison operation in TEX instructions 39 for shadow testing. 40* ``PIPE_CAP_TEXTURE_SWIZZLE``: Whether swizzling through sampler views is 41 supported. 42* ``PIPE_CAP_MAX_TEXTURE_2D_SIZE``: The maximum size of 2D (and 1D) textures. 43* ``PIPE_CAP_MAX_TEXTURE_3D_LEVELS``: The maximum number of mipmap levels available 44 for a 3D texture. 45* ``PIPE_CAP_MAX_TEXTURE_CUBE_LEVELS``: The maximum number of mipmap levels available 46 for a cubemap. 47* ``PIPE_CAP_TEXTURE_MIRROR_CLAMP_TO_EDGE``: Whether mirrored texture coordinates are 48 supported with the clamp-to-edge wrap mode. 49* ``PIPE_CAP_TEXTURE_MIRROR_CLAMP``: Whether mirrored texture coordinates are supported 50 with clamp or clamp-to-border wrap modes. 51* ``PIPE_CAP_BLEND_EQUATION_SEPARATE``: Whether alpha blend equations may be different 52 from color blend equations, in :ref:`Blend` state. 53* ``PIPE_CAP_MAX_STREAM_OUTPUT_BUFFERS``: The maximum number of stream buffers. 54* ``PIPE_CAP_PRIMITIVE_RESTART``: Whether primitive restart is supported. 55* ``PIPE_CAP_PRIMITIVE_RESTART_FIXED_INDEX``: Subset of 56 PRIMITIVE_RESTART where the restart index is always the fixed maximum 57 value for the index type. 58* ``PIPE_CAP_INDEP_BLEND_ENABLE``: Whether per-rendertarget blend enabling and channel 59 masks are supported. If 0, then the first rendertarget's blend mask is 60 replicated across all MRTs. 61* ``PIPE_CAP_INDEP_BLEND_FUNC``: Whether per-rendertarget blend functions are 62 available. If 0, then the first rendertarget's blend functions affect all 63 MRTs. 64* ``PIPE_CAP_MAX_TEXTURE_ARRAY_LAYERS``: The maximum number of texture array 65 layers supported. If 0, the array textures are not supported at all and 66 the ARRAY texture targets are invalid. 67* ``PIPE_CAP_TGSI_FS_COORD_ORIGIN_UPPER_LEFT``: Whether the TGSI property 68 FS_COORD_ORIGIN with value UPPER_LEFT is supported. 69* ``PIPE_CAP_TGSI_FS_COORD_ORIGIN_LOWER_LEFT``: Whether the TGSI property 70 FS_COORD_ORIGIN with value LOWER_LEFT is supported. 71* ``PIPE_CAP_TGSI_FS_COORD_PIXEL_CENTER_HALF_INTEGER``: Whether the TGSI 72 property FS_COORD_PIXEL_CENTER with value HALF_INTEGER is supported. 73* ``PIPE_CAP_TGSI_FS_COORD_PIXEL_CENTER_INTEGER``: Whether the TGSI 74 property FS_COORD_PIXEL_CENTER with value INTEGER is supported. 75* ``PIPE_CAP_DEPTH_CLIP_DISABLE``: Whether the driver is capable of disabling 76 depth clipping (=1) (through pipe_rasterizer_state) or supports lowering 77 depth_clamp in the client shader code (=2), for this the driver must 78 currently use TGSI. 79* ``PIPE_CAP_DEPTH_CLIP_DISABLE_SEPARATE``: Whether the driver is capable of 80 disabling depth clipping (through pipe_rasterizer_state) separately for 81 the near and far plane. If not, depth_clip_near and depth_clip_far will be 82 equal. 83 ``PIPE_CAP_DEPTH_CLAMP_ENABLE``: Whether the driver is capable of 84 enabling depth clamping (through pipe_rasterizer_state) separately from depth 85 clipping. If not, depth_clamp will be the inverse of depth_clip_far. 86* ``PIPE_CAP_SHADER_STENCIL_EXPORT``: Whether a stencil reference value can be 87 written from a fragment shader. 88* ``PIPE_CAP_TGSI_INSTANCEID``: Whether TGSI_SEMANTIC_INSTANCEID is supported 89 in the vertex shader. 90* ``PIPE_CAP_VERTEX_ELEMENT_INSTANCE_DIVISOR``: Whether the driver supports 91 per-instance vertex attribs. 92* ``PIPE_CAP_FRAGMENT_COLOR_CLAMPED``: Whether fragment color clamping is 93 supported. That is, is the pipe_rasterizer_state::clamp_fragment_color 94 flag supported by the driver? If not, gallium frontends will insert 95 clamping code into the fragment shaders when needed. 96 97* ``PIPE_CAP_MIXED_COLORBUFFER_FORMATS``: Whether mixed colorbuffer formats are 98 supported, e.g. RGBA8 and RGBA32F as the first and second colorbuffer, resp. 99* ``PIPE_CAP_VERTEX_COLOR_UNCLAMPED``: Whether the driver is capable of 100 outputting unclamped vertex colors from a vertex shader. If unsupported, 101 the vertex colors are always clamped. This is the default for DX9 hardware. 102* ``PIPE_CAP_VERTEX_COLOR_CLAMPED``: Whether the driver is capable of 103 clamping vertex colors when they come out of a vertex shader, as specified 104 by the pipe_rasterizer_state::clamp_vertex_color flag. If unsupported, 105 the vertex colors are never clamped. This is the default for DX10 hardware. 106 If both clamped and unclamped CAPs are supported, the clamping can be 107 controlled through pipe_rasterizer_state. If the driver cannot do vertex 108 color clamping, gallium frontends may insert clamping code into the vertex 109 shader. 110* ``PIPE_CAP_GLSL_FEATURE_LEVEL``: Whether the driver supports features 111 equivalent to a specific GLSL version. E.g. for GLSL 1.3, report 130. 112* ``PIPE_CAP_GLSL_FEATURE_LEVEL_COMPATIBILITY``: Whether the driver supports 113 features equivalent to a specific GLSL version including all legacy OpenGL 114 features only present in the OpenGL compatibility profile. 115 The only legacy features that Gallium drivers must implement are 116 the legacy shader inputs and outputs (colors, texcoords, fog, clipvertex, 117 edgeflag). 118* ``PIPE_CAP_ESSL_FEATURE_LEVEL``: An optional cap to allow drivers to 119 report a higher GLSL version for GLES contexts. This is useful when a 120 driver does not support all the required features for a higher GL version, 121 but does support the required features for a higher GLES version. A driver 122 is allowed to return ``0`` in which case ``PIPE_CAP_GLSL_FEATURE_LEVEL`` is 123 used. 124 Note that simply returning the same value as the GLSL feature level cap is 125 incorrect. For example, GLSL version 3.30 does not require ``ARB_gpu_shader5``, 126 but ESSL version 3.20 es does require ``EXT_gpu_shader5`` 127* ``PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION``: Whether quads adhere to 128 the flatshade_first setting in ``pipe_rasterizer_state``. 129* ``PIPE_CAP_USER_VERTEX_BUFFERS``: Whether the driver supports user vertex 130 buffers. If not, gallium frontends must upload all data which is not in hw 131 resources. If user-space buffers are supported, the driver must also still 132 accept HW resource buffers. 133* ``PIPE_CAP_VERTEX_BUFFER_OFFSET_4BYTE_ALIGNED_ONLY``: This CAP describes a hw 134 limitation. If true, pipe_vertex_buffer::buffer_offset must always be aligned 135 to 4. If false, there are no restrictions on the offset. 136* ``PIPE_CAP_VERTEX_BUFFER_STRIDE_4BYTE_ALIGNED_ONLY``: This CAP describes a hw 137 limitation. If true, pipe_vertex_buffer::stride must always be aligned to 4. 138 If false, there are no restrictions on the stride. 139* ``PIPE_CAP_VERTEX_ELEMENT_SRC_OFFSET_4BYTE_ALIGNED_ONLY``: This CAP describes 140 a hw limitation. If true, pipe_vertex_element::src_offset must always be 141 aligned to 4. If false, there are no restrictions on src_offset. 142* ``PIPE_CAP_COMPUTE``: Whether the implementation supports the 143 compute entry points defined in pipe_context and pipe_screen. 144* ``PIPE_CAP_CONSTANT_BUFFER_OFFSET_ALIGNMENT``: Describes the required 145 alignment of pipe_constant_buffer::buffer_offset. 146* ``PIPE_CAP_START_INSTANCE``: Whether the driver supports 147 pipe_draw_info::start_instance. 148* ``PIPE_CAP_QUERY_TIMESTAMP``: Whether PIPE_QUERY_TIMESTAMP and 149 the pipe_screen::get_timestamp hook are implemented. 150* ``PIPE_CAP_TEXTURE_MULTISAMPLE``: Whether all MSAA resources supported 151 for rendering are also supported for texturing. 152* ``PIPE_CAP_MIN_MAP_BUFFER_ALIGNMENT``: The minimum alignment that should be 153 expected for a pointer returned by transfer_map if the resource is 154 PIPE_BUFFER. In other words, the pointer returned by transfer_map is 155 always aligned to this value. 156* ``PIPE_CAP_TEXTURE_BUFFER_OFFSET_ALIGNMENT``: Describes the required 157 alignment for pipe_sampler_view::u.buf.offset, in bytes. 158 If a driver does not support offset/size, it should return 0. 159* ``PIPE_CAP_BUFFER_SAMPLER_VIEW_RGBA_ONLY``: Whether the driver only 160 supports R, RG, RGB and RGBA formats for PIPE_BUFFER sampler views. 161 When this is the case it should be assumed that the swizzle parameters 162 in the sampler view have no effect. 163* ``PIPE_CAP_TGSI_TEXCOORD``: This CAP describes a hw limitation. 164 If true, the hardware cannot replace arbitrary shader inputs with sprite 165 coordinates and hence the inputs that are desired to be replaceable must 166 be declared with TGSI_SEMANTIC_TEXCOORD instead of TGSI_SEMANTIC_GENERIC. 167 The rasterizer's sprite_coord_enable state therefore also applies to the 168 TEXCOORD semantic. 169 Also, TGSI_SEMANTIC_PCOORD becomes available, which labels a fragment shader 170 input that will always be replaced with sprite coordinates. 171* ``PIPE_CAP_TEXTURE_BUFFER_SAMPLER``: Whether a sampler should still 172 be used for PIPE_BUFFER resources (normally a sampler is only used 173 if the texture target is PIPE_TEXTURE_*). 174* ``PIPE_CAP_PREFER_BLIT_BASED_TEXTURE_TRANSFER``: Whether it is preferable 175 to use a blit to implement a texture transfer which needs format conversions 176 and swizzling in gallium frontends. Generally, all hardware drivers with 177 dedicated memory should return 1 and all software rasterizers should return 0. 178* ``PIPE_CAP_QUERY_PIPELINE_STATISTICS``: Whether PIPE_QUERY_PIPELINE_STATISTICS 179 is supported. 180* ``PIPE_CAP_TEXTURE_BORDER_COLOR_QUIRK``: Bitmask indicating whether special 181 considerations have to be given to the interaction between the border color 182 in the sampler object and the sampler view used with it. 183 If PIPE_QUIRK_TEXTURE_BORDER_COLOR_SWIZZLE_R600 is set, the border color 184 may be affected in undefined ways for any kind of permutational swizzle 185 (any swizzle XYZW where X/Y/Z/W are not ZERO, ONE, or R/G/B/A respectively) 186 in the sampler view. 187 If PIPE_QUIRK_TEXTURE_BORDER_COLOR_SWIZZLE_NV50 is set, the border color 188 state should be swizzled manually according to the swizzle in the sampler 189 view it is intended to be used with, or herein undefined results may occur 190 for permutational swizzles. 191* ``PIPE_CAP_MAX_TEXTURE_BUFFER_SIZE``: The maximum accessible size with 192 a buffer sampler view, in texels. 193* ``PIPE_CAP_MAX_VIEWPORTS``: The maximum number of viewports (and scissors 194 since they are linked) a driver can support. Returning 0 is equivalent 195 to returning 1 because every driver has to support at least a single 196 viewport/scissor combination. 197* ``PIPE_CAP_ENDIANNESS``:: The endianness of the device. Either 198 PIPE_ENDIAN_BIG or PIPE_ENDIAN_LITTLE. 199* ``PIPE_CAP_MIXED_FRAMEBUFFER_SIZES``: Whether it is allowed to have 200 different sizes for fb color/zs attachments. This controls whether 201 ARB_framebuffer_object is provided. 202* ``PIPE_CAP_TGSI_VS_LAYER_VIEWPORT``: Whether ``TGSI_SEMANTIC_LAYER`` and 203 ``TGSI_SEMANTIC_VIEWPORT_INDEX`` are supported as vertex shader 204 outputs. Note that the viewport will only be used if multiple viewports are 205 exposed. 206* ``PIPE_CAP_MAX_GEOMETRY_OUTPUT_VERTICES``: The maximum number of vertices 207 output by a single invocation of a geometry shader. 208* ``PIPE_CAP_MAX_GEOMETRY_TOTAL_OUTPUT_COMPONENTS``: The maximum number of 209 vertex components output by a single invocation of a geometry shader. 210 This is the product of the number of attribute components per vertex and 211 the number of output vertices. 212* ``PIPE_CAP_MAX_TEXTURE_GATHER_COMPONENTS``: Max number of components 213 in format that texture gather can operate on. 1 == RED, ALPHA etc, 214 4 == All formats. 215* ``PIPE_CAP_TEXTURE_GATHER_SM5``: Whether the texture gather 216 hardware implements the SM5 features, component selection, 217 shadow comparison, and run-time offsets. 218* ``PIPE_CAP_BUFFER_MAP_PERSISTENT_COHERENT``: Whether 219 PIPE_MAP_PERSISTENT and PIPE_MAP_COHERENT are supported 220 for buffers. 221* ``PIPE_CAP_TEXTURE_QUERY_LOD``: Whether the ``LODQ`` instruction is 222 supported. 223* ``PIPE_CAP_MIN_TEXTURE_GATHER_OFFSET``: The minimum offset that can be used 224 in conjunction with a texture gather opcode. 225* ``PIPE_CAP_MAX_TEXTURE_GATHER_OFFSET``: The maximum offset that can be used 226 in conjunction with a texture gather opcode. 227* ``PIPE_CAP_SAMPLE_SHADING``: Whether there is support for per-sample 228 shading. The context->set_min_samples function will be expected to be 229 implemented. 230* ``PIPE_CAP_TEXTURE_GATHER_OFFSETS``: Whether the ``TG4`` instruction can 231 accept 4 offsets. 232* ``PIPE_CAP_TGSI_VS_WINDOW_SPACE_POSITION``: Whether 233 TGSI_PROPERTY_VS_WINDOW_SPACE_POSITION is supported, which disables clipping 234 and viewport transformation. 235* ``PIPE_CAP_MAX_VERTEX_STREAMS``: The maximum number of vertex streams 236 supported by the geometry shader. If stream-out is supported, this should be 237 at least 1. If stream-out is not supported, this should be 0. 238* ``PIPE_CAP_DRAW_INDIRECT``: Whether the driver supports taking draw arguments 239 { count, instance_count, start, index_bias } from a PIPE_BUFFER resource. 240 See pipe_draw_info. 241* ``PIPE_CAP_MULTI_DRAW_INDIRECT``: Whether the driver supports 242 pipe_draw_info::indirect_stride and ::indirect_count 243* ``PIPE_CAP_MULTI_DRAW_INDIRECT_PARAMS``: Whether the driver supports 244 taking the number of indirect draws from a separate parameter 245 buffer, see pipe_draw_indirect_info::indirect_draw_count. 246* ``PIPE_CAP_TGSI_FS_FINE_DERIVATIVE``: Whether the fragment shader supports 247 the FINE versions of DDX/DDY. 248* ``PIPE_CAP_VENDOR_ID``: The vendor ID of the underlying hardware. If it's 249 not available one should return 0xFFFFFFFF. 250* ``PIPE_CAP_DEVICE_ID``: The device ID (PCI ID) of the underlying hardware. 251 0xFFFFFFFF if not available. 252* ``PIPE_CAP_ACCELERATED``: Whether the renderer is hardware accelerated. 253* ``PIPE_CAP_VIDEO_MEMORY``: The amount of video memory in megabytes. 254* ``PIPE_CAP_UMA``: If the device has a unified memory architecture or on-card 255 memory and GART. 256* ``PIPE_CAP_CONDITIONAL_RENDER_INVERTED``: Whether the driver supports inverted 257 condition for conditional rendering. 258* ``PIPE_CAP_MAX_VERTEX_ATTRIB_STRIDE``: The maximum supported vertex stride. 259* ``PIPE_CAP_SAMPLER_VIEW_TARGET``: Whether the sampler view's target can be 260 different than the underlying resource's, as permitted by 261 ARB_texture_view. For example a 2d array texture may be reinterpreted as a 262 cube (array) texture and vice-versa. 263* ``PIPE_CAP_CLIP_HALFZ``: Whether the driver supports the 264 pipe_rasterizer_state::clip_halfz being set to true. This is required 265 for enabling ARB_clip_control. 266* ``PIPE_CAP_VERTEXID_NOBASE``: If true, the driver only supports 267 TGSI_SEMANTIC_VERTEXID_NOBASE (and not TGSI_SEMANTIC_VERTEXID). This means 268 gallium frontends for APIs whose vertexIDs are offset by basevertex (such as GL) 269 will need to lower TGSI_SEMANTIC_VERTEXID to TGSI_SEMANTIC_VERTEXID_NOBASE 270 and TGSI_SEMANTIC_BASEVERTEX, so drivers setting this must handle both these 271 semantics. Only relevant if geometry shaders are supported. 272 (BASEVERTEX could be exposed separately too via ``PIPE_CAP_DRAW_PARAMETERS``). 273* ``PIPE_CAP_POLYGON_OFFSET_CLAMP``: If true, the driver implements support 274 for ``pipe_rasterizer_state::offset_clamp``. 275* ``PIPE_CAP_MULTISAMPLE_Z_RESOLVE``: Whether the driver supports blitting 276 a multisampled depth buffer into a single-sampled texture (or depth buffer). 277 Only the first sampled should be copied. 278* ``PIPE_CAP_RESOURCE_FROM_USER_MEMORY``: Whether the driver can create 279 a pipe_resource where an already-existing piece of (malloc'd) user memory 280 is used as its backing storage. In other words, whether the driver can map 281 existing user memory into the device address space for direct device access. 282 The create function is pipe_screen::resource_from_user_memory. The address 283 and size must be page-aligned. 284* ``PIPE_CAP_RESOURCE_FROM_USER_MEMORY_COMPUTE_ONLY``: Same as 285 ``PIPE_CAP_RESOURCE_FROM_USER_MEMORY`` but indicates it is only supported from 286 the compute engines. 287* ``PIPE_CAP_DEVICE_RESET_STATUS_QUERY``: 288 Whether pipe_context::get_device_reset_status is implemented. 289* ``PIPE_CAP_MAX_SHADER_PATCH_VARYINGS``: 290 How many per-patch outputs and inputs are supported between tessellation 291 control and tessellation evaluation shaders, not counting in TESSINNER and 292 TESSOUTER. The minimum allowed value for OpenGL is 30. 293* ``PIPE_CAP_TEXTURE_FLOAT_LINEAR``: Whether the linear minification and 294 magnification filters are supported with single-precision floating-point 295 textures. 296* ``PIPE_CAP_TEXTURE_HALF_FLOAT_LINEAR``: Whether the linear minification and 297 magnification filters are supported with half-precision floating-point 298 textures. 299* ``PIPE_CAP_DEPTH_BOUNDS_TEST``: Whether bounds_test, bounds_min, and 300 bounds_max states of pipe_depth_stencil_alpha_state behave according 301 to the GL_EXT_depth_bounds_test specification. 302* ``PIPE_CAP_TGSI_TXQS``: Whether the `TXQS` opcode is supported 303* ``PIPE_CAP_FORCE_PERSAMPLE_INTERP``: If the driver can force per-sample 304 interpolation for all fragment shader inputs if 305 pipe_rasterizer_state::force_persample_interp is set. This is only used 306 by GL3-level sample shading (ARB_sample_shading). GL4-level sample shading 307 (ARB_gpu_shader5) doesn't use this. While GL3 hardware has a state for it, 308 GL4 hardware will likely need to emulate it with a shader variant, or by 309 selecting the interpolation weights with a conditional assignment 310 in the shader. 311* ``PIPE_CAP_SHAREABLE_SHADERS``: Whether shader CSOs can be used by any 312 pipe_context. Important for reducing jank at draw time by letting GL shaders 313 linked in one thread be used in another thread without recompiling. 314* ``PIPE_CAP_COPY_BETWEEN_COMPRESSED_AND_PLAIN_FORMATS``: 315 Whether copying between compressed and plain formats is supported where 316 a compressed block is copied to/from a plain pixel of the same size. 317* ``PIPE_CAP_CLEAR_TEXTURE``: Whether `clear_texture` will be 318 available in contexts. 319* ``PIPE_CAP_CLEAR_SCISSORED``: Whether `clear` can accept a scissored 320 bounding box. 321* ``PIPE_CAP_DRAW_PARAMETERS``: Whether ``TGSI_SEMANTIC_BASEVERTEX``, 322 ``TGSI_SEMANTIC_BASEINSTANCE``, and ``TGSI_SEMANTIC_DRAWID`` are 323 supported in vertex shaders. 324* ``PIPE_CAP_TGSI_PACK_HALF_FLOAT``: Whether the ``UP2H`` and ``PK2H`` 325 TGSI opcodes are supported. 326* ``PIPE_CAP_TGSI_FS_POSITION_IS_SYSVAL``: If gallium frontends should use 327 a system value for the POSITION fragment shader input. 328* ``PIPE_CAP_TGSI_FS_POINT_IS_SYSVAL``: If gallium frontends should use 329 a system value for the POINT fragment shader input. 330* ``PIPE_CAP_TGSI_FS_FACE_IS_INTEGER_SYSVAL``: If gallium frontends should use 331 a system value for the FACE fragment shader input. 332 Also, the FACE system value is integer, not float. 333* ``PIPE_CAP_SHADER_BUFFER_OFFSET_ALIGNMENT``: Describes the required 334 alignment for pipe_shader_buffer::buffer_offset, in bytes. Maximum 335 value allowed is 256 (for GL conformance). 0 is only allowed if 336 shader buffers are not supported. 337* ``PIPE_CAP_INVALIDATE_BUFFER``: Whether the use of ``invalidate_resource`` 338 for buffers is supported. 339* ``PIPE_CAP_GENERATE_MIPMAP``: Indicates whether pipe_context::generate_mipmap 340 is supported. 341* ``PIPE_CAP_STRING_MARKER``: Whether pipe->emit_string_marker() is supported. 342* ``PIPE_CAP_SURFACE_REINTERPRET_BLOCKS``: Indicates whether 343 pipe_context::create_surface supports reinterpreting a texture as a surface 344 of a format with different block width/height (but same block size in bits). 345 For example, a compressed texture image can be interpreted as a 346 non-compressed surface whose texels are the same number of bits as the 347 compressed blocks, and vice versa. The width and height of the surface is 348 adjusted appropriately. 349* ``PIPE_CAP_QUERY_BUFFER_OBJECT``: Driver supports 350 context::get_query_result_resource callback. 351* ``PIPE_CAP_PCI_GROUP``: Return the PCI segment group number. 352* ``PIPE_CAP_PCI_BUS``: Return the PCI bus number. 353* ``PIPE_CAP_PCI_DEVICE``: Return the PCI device number. 354* ``PIPE_CAP_PCI_FUNCTION``: Return the PCI function number. 355* ``PIPE_CAP_FRAMEBUFFER_NO_ATTACHMENT``: 356 If non-zero, rendering to framebuffers with no surface attachments 357 is supported. The context->is_format_supported function will be expected 358 to be implemented with PIPE_FORMAT_NONE yeilding the MSAA modes the hardware 359 supports. N.B., The maximum number of layers supported for rasterizing a 360 primitive on a layer is obtained from ``PIPE_CAP_MAX_TEXTURE_ARRAY_LAYERS`` 361 even though it can be larger than the number of layers supported by either 362 rendering or textures. 363* ``PIPE_CAP_ROBUST_BUFFER_ACCESS_BEHAVIOR``: Implementation uses bounds 364 checking on resource accesses by shader if the context is created with 365 PIPE_CONTEXT_ROBUST_BUFFER_ACCESS. See the ARB_robust_buffer_access_behavior 366 extension for information on the required behavior for out of bounds accesses 367 and accesses to unbound resources. 368* ``PIPE_CAP_CULL_DISTANCE``: Whether the driver supports the arb_cull_distance 369 extension and thus implements proper support for culling planes. 370* ``PIPE_CAP_PRIMITIVE_RESTART_FOR_PATCHES``: Whether primitive restart is 371 supported for patch primitives. 372* ``PIPE_CAP_TGSI_VOTE``: Whether the ``VOTE_*`` ops can be used in shaders. 373* ``PIPE_CAP_MAX_WINDOW_RECTANGLES``: The maxium number of window rectangles 374 supported in ``set_window_rectangles``. 375* ``PIPE_CAP_POLYGON_OFFSET_UNITS_UNSCALED``: If true, the driver implements support 376 for ``pipe_rasterizer_state::offset_units_unscaled``. 377* ``PIPE_CAP_VIEWPORT_SUBPIXEL_BITS``: Number of bits of subpixel precision for 378 floating point viewport bounds. 379* ``PIPE_CAP_RASTERIZER_SUBPIXEL_BITS``: Number of bits of subpixel precision used 380 by the rasterizer. 381* ``PIPE_CAP_MIXED_COLOR_DEPTH_BITS``: Whether there is non-fallback 382 support for color/depth format combinations that use a different 383 number of bits. For the purpose of this cap, Z24 is treated as 384 32-bit. If set to off, that means that a B5G6R5 + Z24 or RGBA8 + Z16 385 combination will require a driver fallback, and should not be 386 advertised in the GLX/EGL config list. 387* ``PIPE_CAP_TGSI_ARRAY_COMPONENTS``: If true, the driver interprets the 388 UsageMask of input and output declarations and allows declaring arrays 389 in overlapping ranges. The components must be a contiguous range, e.g. a 390 UsageMask of xy or yzw is allowed, but xz or yw isn't. Declarations with 391 overlapping locations must have matching semantic names and indices, and 392 equal interpolation qualifiers. 393 Components may overlap, notably when the gaps in an array of dvec3 are 394 filled in. 395* ``PIPE_CAP_STREAM_OUTPUT_PAUSE_RESUME``: Whether GL_ARB_transform_feeddback2 396 is supported, including pausing/resuming queries and having 397 `count_from_stream_output` set on indirect draws to implement 398 glDrawTransformFeedback. Required for OpenGL 4.0. 399* ``PIPE_CAP_STREAM_OUTPUT_INTERLEAVE_BUFFERS``: Whether interleaved stream 400 output mode is able to interleave across buffers. This is required for 401 ARB_transform_feedback3. 402* ``PIPE_CAP_TGSI_CAN_READ_OUTPUTS``: Whether every TGSI shader stage can read 403 from the output file. 404* ``PIPE_CAP_GLSL_OPTIMIZE_CONSERVATIVELY``: Tell the GLSL compiler to use 405 the minimum amount of optimizations just to be able to do all the linking 406 and lowering. 407* ``PIPE_CAP_FBFETCH``: The number of render targets whose value in the 408 current framebuffer can be read in the shader. 0 means framebuffer fetch 409 is not supported. 1 means that only the first render target can be read, 410 and a larger value would mean that multiple render targets are supported. 411* ``PIPE_CAP_FBFETCH_COHERENT``: Whether framebuffer fetches from the fragment 412 shader can be guaranteed to be coherent with framebuffer writes. 413* ``PIPE_CAP_TGSI_MUL_ZERO_WINS``: Whether TGSI shaders support the 414 ``TGSI_PROPERTY_MUL_ZERO_WINS`` shader property. 415* ``PIPE_CAP_DOUBLES``: Whether double precision floating-point operations 416 are supported. 417* ``PIPE_CAP_INT64``: Whether 64-bit integer operations are supported. 418* ``PIPE_CAP_INT64_DIVMOD``: Whether 64-bit integer division/modulo 419 operations are supported. 420* ``PIPE_CAP_TGSI_TEX_TXF_LZ``: Whether TEX_LZ and TXF_LZ opcodes are 421 supported. 422* ``PIPE_CAP_TGSI_CLOCK``: Whether the CLOCK opcode is supported. 423* ``PIPE_CAP_POLYGON_MODE_FILL_RECTANGLE``: Whether the 424 PIPE_POLYGON_MODE_FILL_RECTANGLE mode is supported for 425 ``pipe_rasterizer_state::fill_front`` and 426 ``pipe_rasterizer_state::fill_back``. 427* ``PIPE_CAP_SPARSE_BUFFER_PAGE_SIZE``: The page size of sparse buffers in 428 bytes, or 0 if sparse buffers are not supported. The page size must be at 429 most 64KB. 430* ``PIPE_CAP_TGSI_BALLOT``: Whether the BALLOT and READ_* opcodes as well as 431 the SUBGROUP_* semantics are supported. 432* ``PIPE_CAP_TGSI_TES_LAYER_VIEWPORT``: Whether ``TGSI_SEMANTIC_LAYER`` and 433 ``TGSI_SEMANTIC_VIEWPORT_INDEX`` are supported as tessellation evaluation 434 shader outputs. 435* ``PIPE_CAP_CAN_BIND_CONST_BUFFER_AS_VERTEX``: Whether a buffer with just 436 PIPE_BIND_CONSTANT_BUFFER can be legally passed to set_vertex_buffers. 437* ``PIPE_CAP_ALLOW_MAPPED_BUFFERS_DURING_EXECUTION``: As the name says. 438* ``PIPE_CAP_POST_DEPTH_COVERAGE``: whether 439 ``TGSI_PROPERTY_FS_POST_DEPTH_COVERAGE`` is supported. 440* ``PIPE_CAP_BINDLESS_TEXTURE``: Whether bindless texture operations are 441 supported. 442* ``PIPE_CAP_NIR_SAMPLERS_AS_DEREF``: Whether NIR tex instructions should 443 reference texture and sampler as NIR derefs instead of by indices. 444* ``PIPE_CAP_QUERY_SO_OVERFLOW``: Whether the 445 ``PIPE_QUERY_SO_OVERFLOW_PREDICATE`` and 446 ``PIPE_QUERY_SO_OVERFLOW_ANY_PREDICATE`` query types are supported. Note that 447 for a driver that does not support multiple output streams (i.e., 448 ``PIPE_CAP_MAX_VERTEX_STREAMS`` is 1), both query types are identical. 449* ``PIPE_CAP_MEMOBJ``: Whether operations on memory objects are supported. 450* ``PIPE_CAP_LOAD_CONSTBUF``: True if the driver supports ``TGSI_OPCODE_LOAD`` use 451 with constant buffers. 452* ``PIPE_CAP_TGSI_ANY_REG_AS_ADDRESS``: Any TGSI register can be used as 453 an address for indirect register indexing. 454* ``PIPE_CAP_TILE_RASTER_ORDER``: Whether the driver supports 455 GL_MESA_tile_raster_order, using the tile_raster_order_* fields in 456 pipe_rasterizer_state. 457* ``PIPE_CAP_MAX_COMBINED_SHADER_OUTPUT_RESOURCES``: Limit on combined shader 458 output resources (images + buffers + fragment outputs). If 0 the state 459 tracker works it out. 460* ``PIPE_CAP_FRAMEBUFFER_MSAA_CONSTRAINTS``: This determines limitations 461 on the number of samples that framebuffer attachments can have. 462 Possible values: 463 464 0. color.nr_samples == zs.nr_samples == color.nr_storage_samples 465 (standard MSAA quality) 466 1. color.nr_samples >= zs.nr_samples == color.nr_storage_samples 467 (enhanced MSAA quality) 468 2. color.nr_samples >= zs.nr_samples >= color.nr_storage_samples 469 (full flexibility in tuning MSAA quality and performance) 470 471 All color attachments must have the same number of samples and the same 472 number of storage samples. 473* ``PIPE_CAP_SIGNED_VERTEX_BUFFER_OFFSET``: 474 Whether pipe_vertex_buffer::buffer_offset is treated as signed. The u_vbuf 475 module needs this for optimal performance in workstation applications. 476* ``PIPE_CAP_CONTEXT_PRIORITY_MASK``: For drivers that support per-context 477 priorities, this returns a bitmask of ``PIPE_CONTEXT_PRIORITY_x`` for the 478 supported priority levels. A driver that does not support prioritized 479 contexts can return 0. 480* ``PIPE_CAP_FENCE_SIGNAL``: True if the driver supports signaling semaphores 481 using fence_server_signal(). 482* ``PIPE_CAP_CONSTBUF0_FLAGS``: The bits of pipe_resource::flags that must be 483 set when binding that buffer as constant buffer 0. If the buffer doesn't have 484 those bits set, pipe_context::set_constant_buffer(.., 0, ..) is ignored 485 by the driver, and the driver can throw assertion failures. 486* ``PIPE_CAP_PACKED_UNIFORMS``: True if the driver supports packed uniforms 487 as opposed to padding to vec4s. Requires ``PIPE_SHADER_CAP_INTEGERS`` if 488 ``lower_uniforms_to_ubo`` is set. 489* ``PIPE_CAP_CONSERVATIVE_RASTER_POST_SNAP_TRIANGLES``: Whether the 490 ``PIPE_CONSERVATIVE_RASTER_POST_SNAP`` mode is supported for triangles. 491 The post-snap mode means the conservative rasterization occurs after 492 the conversion from floating-point to fixed-point coordinates 493 on the subpixel grid. 494* ``PIPE_CAP_CONSERVATIVE_RASTER_POST_SNAP_POINTS_LINES``: Whether the 495 ``PIPE_CONSERVATIVE_RASTER_POST_SNAP`` mode is supported for points and lines. 496* ``PIPE_CAP_CONSERVATIVE_RASTER_PRE_SNAP_TRIANGLES``: Whether the 497 ``PIPE_CONSERVATIVE_RASTER_PRE_SNAP`` mode is supported for triangles. 498 The pre-snap mode means the conservative rasterization occurs before 499 the conversion from floating-point to fixed-point coordinates. 500* ``PIPE_CAP_CONSERVATIVE_RASTER_PRE_SNAP_POINTS_LINES``: Whether the 501 ``PIPE_CONSERVATIVE_RASTER_PRE_SNAP`` mode is supported for points and lines. 502* ``PIPE_CAP_CONSERVATIVE_RASTER_POST_DEPTH_COVERAGE``: Whether 503 ``PIPE_CAP_POST_DEPTH_COVERAGE`` works with conservative rasterization. 504* ``PIPE_CAP_CONSERVATIVE_RASTER_INNER_COVERAGE``: Whether 505 inner_coverage from GL_INTEL_conservative_rasterization is supported. 506* ``PIPE_CAP_MAX_CONSERVATIVE_RASTER_SUBPIXEL_PRECISION_BIAS``: The maximum 507 subpixel precision bias in bits during conservative rasterization. 508* ``PIPE_CAP_PROGRAMMABLE_SAMPLE_LOCATIONS``: True is the driver supports 509 programmable sample location through ```get_sample_pixel_grid``` and 510 ```set_sample_locations```. 511* ``PIPE_CAP_MAX_GS_INVOCATIONS``: Maximum supported value of 512 TGSI_PROPERTY_GS_INVOCATIONS. 513* ``PIPE_CAP_MAX_SHADER_BUFFER_SIZE``: Maximum supported size for binding 514 with set_shader_buffers. 515* ``PIPE_CAP_MAX_COMBINED_SHADER_BUFFERS``: Maximum total number of shader 516 buffers. A value of 0 means the sum of all per-shader stage maximums (see 517 ``PIPE_SHADER_CAP_MAX_SHADER_BUFFERS``). 518* ``PIPE_CAP_MAX_COMBINED_HW_ATOMIC_COUNTERS``: Maximum total number of atomic 519 counters. A value of 0 means the default value (MAX_ATOMIC_COUNTERS = 4096). 520* ``PIPE_CAP_MAX_COMBINED_HW_ATOMIC_COUNTER_BUFFERS``: Maximum total number of 521 atomic counter buffers. A value of 0 means the sum of all per-shader stage 522 maximums (see ``PIPE_SHADER_CAP_MAX_HW_ATOMIC_COUNTER_BUFFERS``). 523* ``PIPE_CAP_MAX_TEXTURE_UPLOAD_MEMORY_BUDGET``: Maximum recommend memory size 524 for all active texture uploads combined. This is a performance hint. 525 0 means no limit. 526* ``PIPE_CAP_MAX_VERTEX_ELEMENT_SRC_OFFSET``: The maximum supported value for 527 of pipe_vertex_element::src_offset. 528* ``PIPE_CAP_SURFACE_SAMPLE_COUNT``: Whether the driver 529 supports pipe_surface overrides of resource nr_samples. If set, will 530 enable EXT_multisampled_render_to_texture. 531* ``PIPE_CAP_TGSI_ATOMFADD``: Atomic floating point adds are supported on 532 images, buffers, and shared memory. 533* ``PIPE_CAP_RGB_OVERRIDE_DST_ALPHA_BLEND``: True if the driver needs blend state to use zero/one instead of destination alpha for RGB/XRGB formats. 534* ``PIPE_CAP_GLSL_TESS_LEVELS_AS_INPUTS``: True if the driver wants TESSINNER and TESSOUTER to be inputs (rather than system values) for tessellation evaluation shaders. 535* ``PIPE_CAP_DEST_SURFACE_SRGB_CONTROL``: Indicates whether the drivers 536 supports switching the format between sRGB and linear for a surface that is 537 used as destination in draw and blit calls. 538* ``PIPE_CAP_NIR_COMPACT_ARRAYS``: True if the compiler backend supports NIR's compact array feature, for all shader stages. 539* ``PIPE_CAP_MAX_VARYINGS``: The maximum number of fragment shader 540 varyings. This will generally correspond to 541 ``PIPE_SHADER_CAP_MAX_INPUTS`` for the fragment shader, but in some 542 cases may be a smaller number. 543* ``PIPE_CAP_COMPUTE_GRID_INFO_LAST_BLOCK``: Whether pipe_grid_info::last_block 544 is implemented by the driver. See struct pipe_grid_info for more details. 545* ``PIPE_CAP_COMPUTE_SHADER_DERIVATIVE``: True if the driver supports derivatives (and texture lookups with implicit derivatives) in compute shaders. 546* ``PIPE_CAP_TGSI_SKIP_SHRINK_IO_ARRAYS``: Whether the TGSI pass to shrink IO 547 arrays should be skipped and enforce keeping the declared array sizes instead. 548 A driver might rely on the input mapping that was defined with the original 549 GLSL code. 550* ``PIPE_CAP_IMAGE_LOAD_FORMATTED``: True if a format for image loads does not need to be specified in the shader IR 551* ``PIPE_CAP_THROTTLE``: Whether or not gallium frontends should throttle pipe_context 552 execution. 0 = throttling is disabled. 553* ``PIPE_CAP_DMABUF``: Whether Linux DMABUF handles are supported by 554 resource_from_handle and resource_get_handle. 555* ``PIPE_CAP_PREFER_COMPUTE_FOR_MULTIMEDIA``: Whether VDPAU, VAAPI, and 556 OpenMAX should use a compute-based blit instead of pipe_context::blit and compute pipeline for compositing images. 557* ``PIPE_CAP_FRAGMENT_SHADER_INTERLOCK``: True if fragment shader interlock 558 functionality is supported. 559* ``PIPE_CAP_CS_DERIVED_SYSTEM_VALUES_SUPPORTED``: True if driver handles 560 gl_LocalInvocationIndex and gl_GlobalInvocationID. Otherwise, gallium frontends will 561 lower those system values. 562* ``PIPE_CAP_ATOMIC_FLOAT_MINMAX``: Atomic float point minimum, 563 maximum, exchange and compare-and-swap support to buffer and shared variables. 564* ``PIPE_CAP_TGSI_DIV``: Whether opcode DIV is supported 565* ``PIPE_CAP_FRAGMENT_SHADER_TEXTURE_LOD``: Whether texture lookups with 566 explicit LOD is supported in the fragment shader. 567* ``PIPE_CAP_FRAGMENT_SHADER_DERIVATIVES``: True if the driver supports 568 derivatives in fragment shaders. 569* ``PIPE_CAP_VERTEX_SHADER_SATURATE``: True if the driver supports saturate 570 modifiers in the vertex shader. 571* ``PIPE_CAP_TEXTURE_SHADOW_LOD``: True if the driver supports shadow sampler 572 types with texture functions having interaction with LOD of texture lookup. 573* ``PIPE_CAP_SHADER_SAMPLES_IDENTICAL``: True if the driver supports a shader query to tell whether all samples of a multisampled surface are definitely identical. 574* ``PIPE_CAP_TGSI_ATOMINC_WRAP``: Atomic increment/decrement + wrap around are supported. 575* ``PIPE_CAP_PREFER_IMM_ARRAYS_AS_CONSTBUF``: True if gallium frontends should 576 turn arrays whose contents can be deduced at compile time into constant 577 buffer loads, or false if the driver can handle such arrays itself in a more 578 efficient manner. 579* ``PIPE_CAP_GL_SPIRV``: True if the driver supports ARB_gl_spirv extension. 580* ``PIPE_CAP_GL_SPIRV_VARIABLE_POINTERS``: True if the driver supports Variable Pointers in SPIR-V shaders. 581* ``PIPE_CAP_DEMOTE_TO_HELPER_INVOCATION``: True if driver supports demote keyword in GLSL programs. 582* ``PIPE_CAP_TGSI_TG4_COMPONENT_IN_SWIZZLE``: True if driver wants the TG4 component encoded in sampler swizzle rather than as a separate source. 583* ``PIPE_CAP_FLATSHADE``: Driver supports pipe_rasterizer_state::flatshade. Must be 1 584 for non-NIR drivers or gallium nine. 585* ``PIPE_CAP_ALPHA_TEST``: Driver supports alpha-testing. Must be 1 586 for non-NIR drivers or gallium nine. If set, frontend may set 587 ``pipe_depth_stencil_alpha_state->alpha_enabled`` and ``alpha_func``. 588 Otherwise, alpha test will be lowered to a comparison and discard_if in the 589 fragment shader. 590* ``PIPE_CAP_POINT_SIZE_FIXED``: Driver supports point-sizes that are fixed, 591 as opposed to writing gl_PointSize for every point. 592* ``PIPE_CAP_TWO_SIDED_COLOR``: Driver supports two-sided coloring. Must be 1 593 for non-NIR drivers. If set, pipe_rasterizer_state may be set to indicate 594 that backfacing primitives should use the back-side color as the FS input 595 color. If unset, mesa/st will lower it to gl_FrontFacing reads in the 596 fragment shader. 597* ``PIPE_CAP_CLIP_PLANES``: Driver supports user-defined clip-planes. 598* ``PIPE_CAP_MAX_VERTEX_BUFFERS``: Number of supported vertex buffers. 599* ``PIPE_CAP_OPENCL_INTEGER_FUNCTIONS``: Driver supports extended OpenCL-style integer functions. This includes averge, saturating additiong, saturating subtraction, absolute difference, count leading zeros, and count trailing zeros. 600* ``PIPE_CAP_INTEGER_MULTIPLY_32X16``: Driver supports integer multiplication between a 32-bit integer and a 16-bit integer. If the second operand is 32-bits, the upper 16-bits are ignored, and the low 16-bits are possibly sign extended as necessary. 601* ``PIPE_CAP_NIR_IMAGES_AS_DEREF``: Whether NIR image load/store intrinsics should be nir_intrinsic_image_deref_* instead of nir_intrinsic_image_*. Defaults to true. 602* ``PIPE_CAP_PACKED_STREAM_OUTPUT``: Driver supports packing optimization for stream output (e.g. GL transform feedback captured variables). Defaults to true. 603* ``PIPE_CAP_VIEWPORT_TRANSFORM_LOWERED``: Driver needs the nir_lower_viewport_transform pass to be enabled. This also means that the gl_Position value is modified and should be lowered for transform feedback, if needed. Defaults to false. 604* ``PIPE_CAP_PSIZ_CLAMPED``: Driver needs for the point size to be clamped. Additionally, the gl_PointSize has been modified and its value should be lowered for transform feedback, if needed. Defaults to false. 605* ``PIPE_CAP_GL_BEGIN_END_BUFFER_SIZE``: Buffer size used to upload vertices for glBegin/glEnd. 606* ``PIPE_CAP_VIEWPORT_SWIZZLE``: Whether pipe_viewport_state::swizzle can be used to specify pre-clipping swizzling of coordinates (see GL_NV_viewport_swizzle). 607* ``PIPE_CAP_SYSTEM_SVM``: True if all application memory can be shared with the GPU without explicit mapping. 608* ``PIPE_CAP_VIEWPORT_MASK``: Whether ``TGSI_SEMANTIC_VIEWPORT_MASK`` and ``TGSI_PROPERTY_LAYER_VIEWPORT_RELATIVE`` are supported (see GL_NV_viewport_array2). 609* ``PIPE_CAP_MAP_UNSYNCHRONIZED_THREAD_SAFE``: Whether mapping a buffer as unsynchronized from any thread is safe. 610* ``PIPE_CAP_GLSL_ZERO_INIT``: Choose a default zero initialization some glsl variables. If `1`, then all glsl shader variables and gl_FragColor are initialized to zero. If `2`, then shader out variables are not initialized but function out variables are. 611* ``PIPE_CAP_BLEND_EQUATION_ADVANCED``: Driver supports blend equation advanced without necessarily supporting FBFETCH. 612* ``PIPE_CAP_NIR_ATOMICS_AS_DEREF``: Whether NIR atomics instructions should reference atomics as NIR derefs instead of by indices. 613* ``PIPE_CAP_NO_CLIP_ON_COPY_TEX``: Driver doesn't want x/y/width/height clipped based on src size when doing a copy texture operation (eg: may want out-of-bounds reads that produce 0 instead of leaving the texture content undefined) 614* ``PIPE_CAP_MAX_TEXTURE_MB``: Maximum texture size in MB (default is 1024) 615* ``PIPE_CAP_DEVICE_PROTECTED_CONTENT``: Whether the device support protected / encrypted content. 616* ``PIPE_CAP_PREFER_REAL_BUFFER_IN_CONSTBUF0``: The state tracker is encouraged to upload constants into a real buffer and bind it into constant buffer 0 instead of binding a user pointer. This may enable a faster codepath in a gallium frontend for drivers that really prefer a real buffer. 617* ``PIPE_CAP_GL_CLAMP``: Driver natively supports GL_CLAMP. Required for non-NIR drivers with the GL frontend. NIR drivers with the cap unavailable will have GL_CLAMP lowered to txd/txl with a saturate on the coordinates. 618* ``PIPE_CAP_TEXRECT``: Driver supports rectangle textures. Required for OpenGL on `!prefers_nir` drivers. If this cap is not present, st/mesa will lower the NIR to use normal 2D texture sampling by using either `txs` or `nir_intrinsic_load_texture_scaling` to normalize the texture coordinates. 619* ``PIPE_CAP_SAMPLER_REDUCTION_MINMAX``: Driver supports EXT min/max sampler reduction. 620* ``PIPE_CAP_SAMPLER_REDUCTION_MINMAX_ARB``: Driver supports ARB min/max sampler reduction with format queries. 621* ``PIPE_CAP_EMULATE_NONFIXED_PRIMITIVE_RESTART``: Driver requests all draws using a non-fixed restart index to be rewritten to use a fixed restart index. 622* ``PIPE_CAP_SUPPORTED_PRIM_MODES``: A bitmask of the ``pipe_prim_type`` enum values that the driver can natively support. 623* ``PIPE_CAP_SUPPORTED_PRIM_MODES_WITH_RESTART``: A bitmask of the ``pipe_prim_type`` enum values that the driver can natively support for primitive restart. Only useful if ``PIPE_CAP_PRIMITIVE_RESTART`` is also exported. 624* ``PIPE_CAP_PREFER_BACK_BUFFER_REUSE``: Only applies to DRI_PRIME. If 1, the driver prefers that DRI3 tries to use the same back buffer each frame. If 0, this means DRI3 will at least use 2 back buffers and ping-pong between them to allow the tiled->linear copy to run in parallel. 625* ``PIPE_CAP_DRAW_VERTEX_STATE``: Driver supports `pipe_screen::create_vertex_state/vertex_state_destroy` and `pipe_context::draw_vertex_state`. Only used by display lists and designed to serve vbo_save. 626 627.. _pipe_capf: 628 629PIPE_CAPF_* 630^^^^^^^^^^^^^^^^ 631 632The floating-point capabilities are: 633 634* ``PIPE_CAPF_MAX_LINE_WIDTH``: The maximum width of a regular line. 635* ``PIPE_CAPF_MAX_LINE_WIDTH_AA``: The maximum width of a smoothed line. 636* ``PIPE_CAPF_MAX_POINT_WIDTH``: The maximum width and height of a point. 637* ``PIPE_CAPF_MAX_POINT_WIDTH_AA``: The maximum width and height of a smoothed point. 638* ``PIPE_CAPF_MAX_TEXTURE_ANISOTROPY``: The maximum level of anisotropy that can be 639 applied to anisotropically filtered textures. 640* ``PIPE_CAPF_MAX_TEXTURE_LOD_BIAS``: The maximum :term:`LOD` bias that may be applied 641 to filtered textures. 642* ``PIPE_CAPF_MIN_CONSERVATIVE_RASTER_DILATE``: The minimum conservative rasterization 643 dilation. 644* ``PIPE_CAPF_MAX_CONSERVATIVE_RASTER_DILATE``: The maximum conservative rasterization 645 dilation. 646* ``PIPE_CAPF_CONSERVATIVE_RASTER_DILATE_GRANULARITY``: The conservative rasterization 647 dilation granularity for values relative to the minimum dilation. 648 649 650.. _pipe_shader_cap: 651 652PIPE_SHADER_CAP_* 653^^^^^^^^^^^^^^^^^ 654 655These are per-shader-stage capabitity queries. Different shader stages may 656support different features. 657 658* ``PIPE_SHADER_CAP_MAX_INSTRUCTIONS``: The maximum number of instructions. 659* ``PIPE_SHADER_CAP_MAX_ALU_INSTRUCTIONS``: The maximum number of arithmetic instructions. 660* ``PIPE_SHADER_CAP_MAX_TEX_INSTRUCTIONS``: The maximum number of texture instructions. 661* ``PIPE_SHADER_CAP_MAX_TEX_INDIRECTIONS``: The maximum number of texture indirections. 662* ``PIPE_SHADER_CAP_MAX_CONTROL_FLOW_DEPTH``: The maximum nested control flow depth. 663* ``PIPE_SHADER_CAP_MAX_INPUTS``: The maximum number of input registers. 664* ``PIPE_SHADER_CAP_MAX_OUTPUTS``: The maximum number of output registers. 665 This is valid for all shaders except the fragment shader. 666* ``PIPE_SHADER_CAP_MAX_CONST_BUFFER_SIZE``: The maximum size per constant buffer in bytes. 667* ``PIPE_SHADER_CAP_MAX_CONST_BUFFERS``: Maximum number of constant buffers that can be bound 668 to any shader stage using ``set_constant_buffer``. If 0 or 1, the pipe will 669 only permit binding one constant buffer per shader. 670 671If a value greater than 0 is returned, the driver can have multiple 672constant buffers bound to shader stages. The CONST register file is 673accessed with two-dimensional indices, like in the example below. 674 675DCL CONST[0][0..7] # declare first 8 vectors of constbuf 0 676DCL CONST[3][0] # declare first vector of constbuf 3 677MOV OUT[0], CONST[0][3] # copy vector 3 of constbuf 0 678 679* ``PIPE_SHADER_CAP_MAX_TEMPS``: The maximum number of temporary registers. 680* ``PIPE_SHADER_CAP_TGSI_CONT_SUPPORTED``: Whether the continue opcode is supported. 681* ``PIPE_SHADER_CAP_INDIRECT_INPUT_ADDR``: Whether indirect addressing 682 of the input file is supported. 683* ``PIPE_SHADER_CAP_INDIRECT_OUTPUT_ADDR``: Whether indirect addressing 684 of the output file is supported. 685* ``PIPE_SHADER_CAP_INDIRECT_TEMP_ADDR``: Whether indirect addressing 686 of the temporary file is supported. 687* ``PIPE_SHADER_CAP_INDIRECT_CONST_ADDR``: Whether indirect addressing 688 of the constant file is supported. 689* ``PIPE_SHADER_CAP_SUBROUTINES``: Whether subroutines are supported, i.e. 690 BGNSUB, ENDSUB, CAL, and RET, including RET in the main block. 691* ``PIPE_SHADER_CAP_INTEGERS``: Whether integer opcodes are supported. 692 If unsupported, only float opcodes are supported. 693* ``PIPE_SHADER_CAP_INT64_ATOMICS``: Whether int64 atomic opcodes are supported. The device needs to support add, sub, swap, cmpswap, and, or, xor, min, and max. 694* ``PIPE_SHADER_CAP_FP16``: Whether half precision floating-point opcodes are supported. 695 If unsupported, half precision ops need to be lowered to full precision. 696* ``PIPE_SHADER_CAP_FP16_DERIVATIVES``: Whether half precision floating-point 697 DDX and DDY opcodes are supported. 698* ``PIPE_SHADER_CAP_FP16_CONST_BUFFERS``: Whether half precision floating-point 699 constant buffer loads are supported. Drivers are recommended to report 0 700 if x86 F16C is not supported by the CPU (or an equivalent instruction set 701 on other CPU architectures), otherwise they could be impacted by emulated 702 FP16 conversions in glUniform. 703* ``PIPE_SHADER_CAP_INT16``: Whether 16-bit signed and unsigned integer types 704 are supported. 705* ``PIPE_SHADER_CAP_GLSL_16BIT_CONSTS``: Lower mediump constants to 16-bit. 706 Note that 16-bit constants are not lowered to uniforms in GLSL. 707* ``PIPE_SHADER_CAP_MAX_TEXTURE_SAMPLERS``: The maximum number of texture 708 samplers. 709* ``PIPE_SHADER_CAP_PREFERRED_IR``: Preferred representation of the 710 program. It should be one of the ``pipe_shader_ir`` enum values. 711* ``PIPE_SHADER_CAP_MAX_SAMPLER_VIEWS``: The maximum number of texture 712 sampler views. Must not be lower than PIPE_SHADER_CAP_MAX_TEXTURE_SAMPLERS. 713* ``PIPE_SHADER_CAP_TGSI_DROUND_SUPPORTED``: Whether double precision rounding 714 is supported. If it is, DTRUNC/DCEIL/DFLR/DROUND opcodes may be used. 715* ``PIPE_SHADER_CAP_TGSI_DFRACEXP_DLDEXP_SUPPORTED``: Whether DFRACEXP and 716 DLDEXP are supported. 717* ``PIPE_SHADER_CAP_TGSI_LDEXP_SUPPORTED``: Whether LDEXP is supported. 718* ``PIPE_SHADER_CAP_TGSI_FMA_SUPPORTED``: Whether FMA and DFMA (doubles only) 719 are supported. 720* ``PIPE_SHADER_CAP_TGSI_ANY_INOUT_DECL_RANGE``: Whether the driver doesn't 721 ignore tgsi_declaration_range::Last for shader inputs and outputs. 722* ``PIPE_SHADER_CAP_MAX_UNROLL_ITERATIONS_HINT``: This is the maximum number 723 of iterations that loops are allowed to have to be unrolled. It is only 724 a hint to gallium frontends. Whether any loops will be unrolled is not 725 guaranteed. 726* ``PIPE_SHADER_CAP_MAX_SHADER_BUFFERS``: Maximum number of memory buffers 727 (also used to implement atomic counters). Having this be non-0 also 728 implies support for the ``LOAD``, ``STORE``, and ``ATOM*`` TGSI 729 opcodes. 730* ``PIPE_SHADER_CAP_SUPPORTED_IRS``: Supported representations of the 731 program. It should be a mask of ``pipe_shader_ir`` bits. 732* ``PIPE_SHADER_CAP_MAX_SHADER_IMAGES``: Maximum number of image units. 733* ``PIPE_SHADER_CAP_LOWER_IF_THRESHOLD``: IF and ELSE branches with a lower 734 cost than this value should be lowered by gallium frontends for better 735 performance. This is a tunable for the GLSL compiler and the behavior is 736 specific to the compiler. 737* ``PIPE_SHADER_CAP_TGSI_SKIP_MERGE_REGISTERS``: Whether the merge registers 738 TGSI pass is skipped. This might reduce code size and register pressure if 739 the underlying driver has a real backend compiler. 740* ``PIPE_SHADER_CAP_MAX_HW_ATOMIC_COUNTERS``: If atomic counters are separate, 741 how many HW counters are available for this stage. (0 uses SSBO atomics). 742* ``PIPE_SHADER_CAP_MAX_HW_ATOMIC_COUNTER_BUFFERS``: If atomic counters are 743 separate, how many atomic counter buffers are available for this stage. 744 745.. _pipe_compute_cap: 746 747PIPE_COMPUTE_CAP_* 748^^^^^^^^^^^^^^^^^^ 749 750Compute-specific capabilities. They can be queried using 751pipe_screen::get_compute_param. 752 753* ``PIPE_COMPUTE_CAP_IR_TARGET``: A description of the target of the form 754 ``processor-arch-manufacturer-os`` that will be passed on to the compiler. 755 This CAP is only relevant for drivers that specify PIPE_SHADER_IR_NATIVE for 756 their preferred IR. 757 Value type: null-terminated string. Shader IR type dependent. 758* ``PIPE_COMPUTE_CAP_GRID_DIMENSION``: Number of supported dimensions 759 for grid and block coordinates. Value type: ``uint64_t``. Shader IR type dependent. 760* ``PIPE_COMPUTE_CAP_MAX_GRID_SIZE``: Maximum grid size in block 761 units. Value type: ``uint64_t []``. Shader IR type dependent. 762* ``PIPE_COMPUTE_CAP_MAX_BLOCK_SIZE``: Maximum block size in thread 763 units. Value type: ``uint64_t []``. Shader IR type dependent. 764* ``PIPE_COMPUTE_CAP_MAX_THREADS_PER_BLOCK``: Maximum number of threads that 765 a single block can contain. Value type: ``uint64_t``. Shader IR type dependent. 766 This may be less than the product of the components of MAX_BLOCK_SIZE and is 767 usually limited by the number of threads that can be resident simultaneously 768 on a compute unit. 769* ``PIPE_COMPUTE_CAP_MAX_GLOBAL_SIZE``: Maximum size of the GLOBAL 770 resource. Value type: ``uint64_t``. Shader IR type dependent. 771* ``PIPE_COMPUTE_CAP_MAX_LOCAL_SIZE``: Maximum size of the LOCAL 772 resource. Value type: ``uint64_t``. Shader IR type dependent. 773* ``PIPE_COMPUTE_CAP_MAX_PRIVATE_SIZE``: Maximum size of the PRIVATE 774 resource. Value type: ``uint64_t``. Shader IR type dependent. 775* ``PIPE_COMPUTE_CAP_MAX_INPUT_SIZE``: Maximum size of the INPUT 776 resource. Value type: ``uint64_t``. Shader IR type dependent. 777* ``PIPE_COMPUTE_CAP_MAX_MEM_ALLOC_SIZE``: Maximum size of a memory object 778 allocation in bytes. Value type: ``uint64_t``. 779* ``PIPE_COMPUTE_CAP_MAX_CLOCK_FREQUENCY``: Maximum frequency of the GPU 780 clock in MHz. Value type: ``uint32_t`` 781* ``PIPE_COMPUTE_CAP_MAX_COMPUTE_UNITS``: Maximum number of compute units 782 Value type: ``uint32_t`` 783* ``PIPE_COMPUTE_CAP_IMAGES_SUPPORTED``: Whether images are supported 784 non-zero means yes, zero means no. Value type: ``uint32_t`` 785* ``PIPE_COMPUTE_CAP_SUBGROUP_SIZE``: The size of a basic execution unit in 786 threads. Also known as wavefront size, warp size or SIMD width. 787* ``PIPE_COMPUTE_CAP_ADDRESS_BITS``: The default compute device address space 788 size specified as an unsigned integer value in bits. 789* ``PIPE_COMPUTE_CAP_MAX_VARIABLE_THREADS_PER_BLOCK``: Maximum variable number 790 of threads that a single block can contain. This is similar to 791 PIPE_COMPUTE_CAP_MAX_THREADS_PER_BLOCK, except that the variable size is not 792 known a compile-time but at dispatch-time. 793 794.. _pipe_bind: 795 796PIPE_BIND_* 797^^^^^^^^^^^ 798 799These flags indicate how a resource will be used and are specified at resource 800creation time. Resources may be used in different roles 801during their lifecycle. Bind flags are cumulative and may be combined to create 802a resource which can be used for multiple things. 803Depending on the pipe driver's memory management and these bind flags, 804resources might be created and handled quite differently. 805 806* ``PIPE_BIND_RENDER_TARGET``: A color buffer or pixel buffer which will be 807 rendered to. Any surface/resource attached to pipe_framebuffer_state::cbufs 808 must have this flag set. 809* ``PIPE_BIND_DEPTH_STENCIL``: A depth (Z) buffer and/or stencil buffer. Any 810 depth/stencil surface/resource attached to pipe_framebuffer_state::zsbuf must 811 have this flag set. 812* ``PIPE_BIND_BLENDABLE``: Used in conjunction with PIPE_BIND_RENDER_TARGET to 813 query whether a device supports blending for a given format. 814 If this flag is set, surface creation may fail if blending is not supported 815 for the specified format. If it is not set, a driver may choose to ignore 816 blending on surfaces with formats that would require emulation. 817* ``PIPE_BIND_DISPLAY_TARGET``: A surface that can be presented to screen. Arguments to 818 pipe_screen::flush_front_buffer must have this flag set. 819* ``PIPE_BIND_SAMPLER_VIEW``: A texture that may be sampled from in a fragment 820 or vertex shader. 821* ``PIPE_BIND_VERTEX_BUFFER``: A vertex buffer. 822* ``PIPE_BIND_INDEX_BUFFER``: An vertex index/element buffer. 823* ``PIPE_BIND_CONSTANT_BUFFER``: A buffer of shader constants. 824* ``PIPE_BIND_STREAM_OUTPUT``: A stream output buffer. 825* ``PIPE_BIND_CUSTOM``: 826* ``PIPE_BIND_SCANOUT``: A front color buffer or scanout buffer. 827* ``PIPE_BIND_SHARED``: A sharable buffer that can be given to another 828 process. 829* ``PIPE_BIND_GLOBAL``: A buffer that can be mapped into the global 830 address space of a compute program. 831* ``PIPE_BIND_SHADER_BUFFER``: A buffer without a format that can be bound 832 to a shader and can be used with load, store, and atomic instructions. 833* ``PIPE_BIND_SHADER_IMAGE``: A buffer or texture with a format that can be 834 bound to a shader and can be used with load, store, and atomic instructions. 835* ``PIPE_BIND_COMPUTE_RESOURCE``: A buffer or texture that can be 836 bound to the compute program as a shader resource. 837* ``PIPE_BIND_COMMAND_ARGS_BUFFER``: A buffer that may be sourced by the 838 GPU command processor. It can contain, for example, the arguments to 839 indirect draw calls. 840 841.. _pipe_usage: 842 843PIPE_USAGE_* 844^^^^^^^^^^^^ 845 846The PIPE_USAGE enums are hints about the expected usage pattern of a resource. 847Note that drivers must always support read and write CPU access at any time 848no matter which hint they got. 849 850* ``PIPE_USAGE_DEFAULT``: Optimized for fast GPU access. 851* ``PIPE_USAGE_IMMUTABLE``: Optimized for fast GPU access and the resource is 852 not expected to be mapped or changed (even by the GPU) after the first upload. 853* ``PIPE_USAGE_DYNAMIC``: Expect frequent write-only CPU access. What is 854 uploaded is expected to be used at least several times by the GPU. 855* ``PIPE_USAGE_STREAM``: Expect frequent write-only CPU access. What is 856 uploaded is expected to be used only once by the GPU. 857* ``PIPE_USAGE_STAGING``: Optimized for fast CPU access. 858 859 860Methods 861------- 862 863XXX to-do 864 865get_name 866^^^^^^^^ 867 868Returns an identifying name for the screen. 869 870The returned string should remain valid and immutable for the lifetime of 871pipe_screen. 872 873get_vendor 874^^^^^^^^^^ 875 876Returns the screen vendor. 877 878The returned string should remain valid and immutable for the lifetime of 879pipe_screen. 880 881get_device_vendor 882^^^^^^^^^^^^^^^^^ 883 884Returns the actual vendor of the device driving the screen 885(as opposed to the driver vendor). 886 887The returned string should remain valid and immutable for the lifetime of 888pipe_screen. 889 890.. _get_param: 891 892get_param 893^^^^^^^^^ 894 895Get an integer/boolean screen parameter. 896 897**param** is one of the :ref:`PIPE_CAP` names. 898 899.. _get_paramf: 900 901get_paramf 902^^^^^^^^^^ 903 904Get a floating-point screen parameter. 905 906**param** is one of the :ref:`PIPE_CAPF` names. 907 908context_create 909^^^^^^^^^^^^^^ 910 911Create a pipe_context. 912 913**priv** is private data of the caller, which may be put to various 914unspecified uses, typically to do with implementing swapbuffers 915and/or front-buffer rendering. 916 917is_format_supported 918^^^^^^^^^^^^^^^^^^^ 919 920Determine if a resource in the given format can be used in a specific manner. 921 922**format** the resource format 923 924**target** one of the PIPE_TEXTURE_x flags 925 926**sample_count** the number of samples. 0 and 1 mean no multisampling, 927the maximum allowed legal value is 32. 928 929**storage_sample_count** the number of storage samples. This must be <= 930sample_count. See the documentation of ``pipe_resource::nr_storage_samples``. 931 932**bindings** is a bitmask of :ref:`PIPE_BIND` flags. 933 934Returns TRUE if all usages can be satisfied. 935 936 937can_create_resource 938^^^^^^^^^^^^^^^^^^^ 939 940Check if a resource can actually be created (but don't actually allocate any 941memory). This is used to implement OpenGL's proxy textures. Typically, a 942driver will simply check if the total size of the given resource is less than 943some limit. 944 945For PIPE_TEXTURE_CUBE, the pipe_resource::array_size field should be 6. 946 947 948.. _resource_create: 949 950resource_create 951^^^^^^^^^^^^^^^ 952 953Create a new resource from a template. 954The following fields of the pipe_resource must be specified in the template: 955 956**target** one of the pipe_texture_target enums. 957Note that PIPE_BUFFER and PIPE_TEXTURE_X are not really fundamentally different. 958Modern APIs allow using buffers as shader resources. 959 960**format** one of the pipe_format enums. 961 962**width0** the width of the base mip level of the texture or size of the buffer. 963 964**height0** the height of the base mip level of the texture 965(1 for 1D or 1D array textures). 966 967**depth0** the depth of the base mip level of the texture 968(1 for everything else). 969 970**array_size** the array size for 1D and 2D array textures. 971For cube maps this must be 6, for other textures 1. 972 973**last_level** the last mip map level present. 974 975**nr_samples**: Number of samples determining quality, driving the rasterizer, 976shading, and framebuffer. It is the number of samples seen by the whole 977graphics pipeline. 0 and 1 specify a resource which isn't multisampled. 978 979**nr_storage_samples**: Only color buffers can set this lower than nr_samples. 980Multiple samples within a pixel can have the same color. ``nr_storage_samples`` 981determines how many slots for different colors there are per pixel. 982If there are not enough slots to store all sample colors, some samples will 983have an undefined color (called "undefined samples"). 984 985The resolve blit behavior is driver-specific, but can be one of these two: 986 9871. Only defined samples will be averaged. Undefined samples will be ignored. 9882. Undefined samples will be approximated by looking at surrounding defined 989 samples (even in different pixels). 990 991Blits and MSAA texturing: If the sample being fetched is undefined, one of 992the defined samples is returned instead. 993 994Sample shading (``set_min_samples``) will operate at a sample frequency that 995is at most ``nr_storage_samples``. Greater ``min_samples`` values will be 996replaced by ``nr_storage_samples``. 997 998**usage** one of the :ref:`PIPE_USAGE` flags. 999 1000**bind** bitmask of the :ref:`PIPE_BIND` flags. 1001 1002**flags** bitmask of PIPE_RESOURCE_FLAG flags. 1003 1004**next**: Pointer to the next plane for resources that consist of multiple 1005memory planes. 1006 1007As a corollary, this mean resources for an image with multiple planes have 1008to be created starting from the highest plane. 1009 1010resource_changed 1011^^^^^^^^^^^^^^^^ 1012 1013Mark a resource as changed so derived internal resources will be recreated 1014on next use. 1015 1016When importing external images that can't be directly used as texture sampler 1017source, internal copies may have to be created that the hardware can sample 1018from. When those resources are reimported, the image data may have changed, and 1019the previously derived internal resources must be invalidated to avoid sampling 1020from old copies. 1021 1022 1023 1024resource_destroy 1025^^^^^^^^^^^^^^^^ 1026 1027Destroy a resource. A resource is destroyed if it has no more references. 1028 1029 1030 1031get_timestamp 1032^^^^^^^^^^^^^ 1033 1034Query a timestamp in nanoseconds. The returned value should match 1035PIPE_QUERY_TIMESTAMP. This function returns immediately and doesn't 1036wait for rendering to complete (which cannot be achieved with queries). 1037 1038 1039 1040get_driver_query_info 1041^^^^^^^^^^^^^^^^^^^^^ 1042 1043Return a driver-specific query. If the **info** parameter is NULL, 1044the number of available queries is returned. Otherwise, the driver 1045query at the specified **index** is returned in **info**. 1046The function returns non-zero on success. 1047The driver-specific query is described with the pipe_driver_query_info 1048structure. 1049 1050get_driver_query_group_info 1051^^^^^^^^^^^^^^^^^^^^^^^^^^^ 1052 1053Return a driver-specific query group. If the **info** parameter is NULL, 1054the number of available groups is returned. Otherwise, the driver 1055query group at the specified **index** is returned in **info**. 1056The function returns non-zero on success. 1057The driver-specific query group is described with the 1058pipe_driver_query_group_info structure. 1059 1060 1061 1062get_disk_shader_cache 1063^^^^^^^^^^^^^^^^^^^^^ 1064 1065Returns a pointer to a driver-specific on-disk shader cache. If the driver 1066failed to create the cache or does not support an on-disk shader cache NULL is 1067returned. The callback itself may also be NULL if the driver doesn't support 1068an on-disk shader cache. 1069 1070 1071is_dmabuf_modifier_supported 1072^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 1073 1074Query whether the driver supports a **modifier** in combination with a 1075**format**, and whether it is only supported with "external" texture targets. 1076If the combination is supported in any fashion, true is returned. If the 1077**external_only** parameter is not NULL, the bool it points to is set to 1078false if non-external texture targets are supported with the specified modifier+ 1079format, or true if only external texture targets are supported. 1080 1081 1082get_dmabuf_modifier_planes 1083^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 1084 1085Query the number of planes required by the image layout specified by the 1086**modifier** and **format** parameters. The value returned includes both planes 1087dictated by **format** and any additional planes required for driver-specific 1088auxiliary data necessary for the layout defined by **modifier**. 1089If the proc is NULL, no auxiliary planes are required for any layout supported by 1090**screen** and the number of planes can be derived directly from **format**. 1091 1092 1093Thread safety 1094------------- 1095 1096Screen methods are required to be thread safe. While gallium rendering 1097contexts are not required to be thread safe, it is required to be safe to use 1098different contexts created with the same screen in different threads without 1099locks. It is also required to be safe using screen methods in a thread, while 1100using one of its contexts in another (without locks). 1101