X-Git-Url: https://git.libre-soc.org/?a=blobdiff_plain;f=src%2Fmesa%2Fswrast%2Fs_texcombine.c;h=c67c356c1e01f305dcc7c98f17f36073b76d467b;hb=26b8dfc8cadf0f1a8604fc77b226cc7de005f9ca;hp=95e2c082ba721055813881de7fa739a8a9922de4;hpb=653a83445f94620673f747a4ace6847a2c7fdb4d;p=mesa.git diff --git a/src/mesa/swrast/s_texcombine.c b/src/mesa/swrast/s_texcombine.c index 95e2c082ba7..c67c356c1e0 100644 --- a/src/mesa/swrast/s_texcombine.c +++ b/src/mesa/swrast/s_texcombine.c @@ -27,9 +27,9 @@ #include "main/glheader.h" #include "main/context.h" #include "main/colormac.h" -#include "main/image.h" #include "main/imports.h" -#include "shader/prog_instruction.h" +#include "main/pixeltransfer.h" +#include "program/prog_instruction.h" #include "s_context.h" #include "s_texcombine.h" @@ -45,10 +45,14 @@ typedef float (*float4_array)[4]; /** * Return array of texels for given unit. */ -static INLINE float4_array +static inline float4_array get_texel_array(SWcontext *swrast, GLuint unit) { +#ifdef _OPENMP + return (float4_array) (swrast->TexelBuffer + unit * MAX_WIDTH * 4 * omp_get_num_threads() + (MAX_WIDTH * 4 * omp_get_thread_num())); +#else return (float4_array) (swrast->TexelBuffer + unit * MAX_WIDTH * 4); +#endif } @@ -72,7 +76,7 @@ get_texel_array(SWcontext *swrast, GLuint unit) * \param rgba incoming/result fragment colors */ static void -texture_combine( GLcontext *ctx, GLuint unit, GLuint n, +texture_combine( struct gl_context *ctx, GLuint unit, GLuint n, const float4_array primary_rgba, const GLfloat *texelBuffer, GLchan (*rgbaChan)[4] ) @@ -86,10 +90,28 @@ texture_combine( GLcontext *ctx, GLuint unit, GLuint n, const GLfloat scaleA = (GLfloat) (1 << combine->ScaleShiftA); const GLuint numArgsRGB = combine->_NumArgsRGB; const GLuint numArgsA = combine->_NumArgsA; - GLfloat ccolor[MAX_COMBINER_TERMS][MAX_WIDTH][4]; /* temp color buffers */ - GLfloat rgba[MAX_WIDTH][4]; + float4_array ccolor[4], rgba; GLuint i, term; + /* alloc temp pixel buffers */ + rgba = (float4_array) malloc(4 * n * sizeof(GLfloat)); + if (!rgba) { + _mesa_error(ctx, GL_OUT_OF_MEMORY, "texture_combine"); + return; + } + + for (i = 0; i < numArgsRGB || i < numArgsA; i++) { + ccolor[i] = (float4_array) malloc(4 * n * sizeof(GLfloat)); + if (!ccolor[i]) { + while (i) { + free(ccolor[i]); + i--; + } + _mesa_error(ctx, GL_OUT_OF_MEMORY, "texture_combine"); + return; + } + } + for (i = 0; i < n; i++) { rgba[i][RCOMP] = CHAN_TO_FLOAT(rgbaChan[i][RCOMP]); rgba[i][GCOMP] = CHAN_TO_FLOAT(rgbaChan[i][GCOMP]); @@ -163,7 +185,7 @@ texture_combine( GLcontext *ctx, GLuint unit, GLuint n, const GLuint srcUnit = srcRGB - GL_TEXTURE0; ASSERT(srcUnit < ctx->Const.MaxTextureUnits); if (!ctx->Texture.Unit[srcUnit]._ReallyEnabled) - return; + goto end; argRGB[term] = get_texel_array(swrast, srcUnit); } } @@ -253,7 +275,7 @@ texture_combine( GLcontext *ctx, GLuint unit, GLuint n, const GLuint srcUnit = srcA - GL_TEXTURE0; ASSERT(srcUnit < ctx->Const.MaxTextureUnits); if (!ctx->Texture.Unit[srcUnit]._ReallyEnabled) - return; + goto end; argA[term] = get_texel_array(swrast, srcUnit); } } @@ -368,7 +390,7 @@ texture_combine( GLcontext *ctx, GLuint unit, GLuint n, (arg0[i][GCOMP] - 0.5F) * (arg1[i][GCOMP] - 0.5F) + (arg0[i][BCOMP] - 0.5F) * (arg1[i][BCOMP] - 0.5F)) * 4.0F * scaleRGB; - dot = CLAMP(dot, 0.0, 1.0F); + dot = CLAMP(dot, 0.0F, 1.0F); rgba[i][RCOMP] = rgba[i][GCOMP] = rgba[i][BCOMP] = dot; } break; @@ -411,7 +433,7 @@ texture_combine( GLcontext *ctx, GLuint unit, GLuint n, rgba[i][BCOMP] = 0.0; rgba[i][ACOMP] = 1.0; } - return; /* no alpha processing */ + goto end; /* no alpha processing */ default: _mesa_problem(ctx, "invalid combine mode"); } @@ -519,6 +541,12 @@ texture_combine( GLcontext *ctx, GLuint unit, GLuint n, UNCLAMPED_FLOAT_TO_CHAN(rgbaChan[i][BCOMP], rgba[i][BCOMP]); UNCLAMPED_FLOAT_TO_CHAN(rgbaChan[i][ACOMP], rgba[i][ACOMP]); } + +end: + for (i = 0; i < numArgsRGB || i < numArgsA; i++) { + free(ccolor[i]); + } + free(rgba); } @@ -556,12 +584,39 @@ swizzle_texels(GLuint swizzle, GLuint count, float4_array texels) * Apply texture mapping to a span of fragments. */ void -_swrast_texture_span( GLcontext *ctx, SWspan *span ) +_swrast_texture_span( struct gl_context *ctx, SWspan *span ) { SWcontext *swrast = SWRAST_CONTEXT(ctx); - GLfloat primary_rgba[MAX_WIDTH][4]; + float4_array primary_rgba; GLuint unit; + if (!swrast->TexelBuffer) { +#ifdef _OPENMP + const GLint maxThreads = omp_get_max_threads(); +#else + const GLint maxThreads = 1; +#endif + + /* TexelBuffer is also global and normally shared by all SWspan + * instances; when running with multiple threads, create one per + * thread. + */ + swrast->TexelBuffer = + (GLfloat *) MALLOC(ctx->Const.MaxTextureImageUnits * maxThreads * + MAX_WIDTH * 4 * sizeof(GLfloat)); + if (!swrast->TexelBuffer) { + _mesa_error(ctx, GL_OUT_OF_MEMORY, "texture_combine"); + return; + } + } + + primary_rgba = (float4_array) malloc(span->end * 4 * sizeof(GLfloat)); + + if (!primary_rgba) { + _mesa_error(ctx, GL_OUT_OF_MEMORY, "texture_span"); + return; + } + ASSERT(span->end <= MAX_WIDTH); /* @@ -600,9 +655,9 @@ _swrast_texture_span( GLcontext *ctx, SWspan *span ) /* adjust texture lod (lambda) */ if (span->arrayMask & SPAN_LAMBDA) { - if (texUnit->LodBias + curObj->LodBias != 0.0F) { + if (texUnit->LodBias + curObj->Sampler.LodBias != 0.0F) { /* apply LOD bias, but don't clamp yet */ - const GLfloat bias = CLAMP(texUnit->LodBias + curObj->LodBias, + const GLfloat bias = CLAMP(texUnit->LodBias + curObj->Sampler.LodBias, -ctx->Const.MaxTextureLodBias, ctx->Const.MaxTextureLodBias); GLuint i; @@ -611,10 +666,11 @@ _swrast_texture_span( GLcontext *ctx, SWspan *span ) } } - if (curObj->MinLod != -1000.0 || curObj->MaxLod != 1000.0) { + if (curObj->Sampler.MinLod != -1000.0 || + curObj->Sampler.MaxLod != 1000.0) { /* apply LOD clamping to lambda */ - const GLfloat min = curObj->MinLod; - const GLfloat max = curObj->MaxLod; + const GLfloat min = curObj->Sampler.MinLod; + const GLfloat max = curObj->Sampler.MaxLod; GLuint i; for (i = 0; i < span->end; i++) { GLfloat l = lambda[i]; @@ -655,9 +711,9 @@ _swrast_texture_span( GLcontext *ctx, SWspan *span ) /* adjust texture lod (lambda) */ if (span->arrayMask & SPAN_LAMBDA) { - if (texUnit->LodBias + curObj->LodBias != 0.0F) { + if (texUnit->LodBias + curObj->Sampler.LodBias != 0.0F) { /* apply LOD bias, but don't clamp yet */ - const GLfloat bias = CLAMP(texUnit->LodBias + curObj->LodBias, + const GLfloat bias = CLAMP(texUnit->LodBias + curObj->Sampler.LodBias, -ctx->Const.MaxTextureLodBias, ctx->Const.MaxTextureLodBias); GLuint i; @@ -666,10 +722,11 @@ _swrast_texture_span( GLcontext *ctx, SWspan *span ) } } - if (curObj->MinLod != -1000.0 || curObj->MaxLod != 1000.0) { + if (curObj->Sampler.MinLod != -1000.0 || + curObj->Sampler.MaxLod != 1000.0) { /* apply LOD clamping to lambda */ - const GLfloat min = curObj->MinLod; - const GLfloat max = curObj->MaxLod; + const GLfloat min = curObj->Sampler.MinLod; + const GLfloat max = curObj->Sampler.MaxLod; GLuint i; for (i = 0; i < span->end; i++) { GLfloat l = lambda[i]; @@ -677,16 +734,23 @@ _swrast_texture_span( GLcontext *ctx, SWspan *span ) } } } + else if (curObj->Sampler.MaxAnisotropy > 1.0 && + curObj->Sampler.MinFilter == GL_LINEAR_MIPMAP_LINEAR) { + /* sample_lambda_2d_aniso is beeing used as texture_sample_func, + * it requires the current SWspan *span as an additional parameter. + * In order to keep the same function signature, the unused lambda + * parameter will be modified to actually contain the SWspan pointer. + * This is a Hack. To make it right, the texture_sample_func + * signature and all implementing functions need to be modified. + */ + /* "hide" SWspan struct; cast to (GLfloat *) to suppress warning */ + lambda = (GLfloat *)span; + } /* Sample the texture (span->end = number of fragments) */ swrast->TextureSample[unit]( ctx, texUnit->_Current, span->end, texcoords, lambda, texels ); - /* GL_SGI_texture_color_table */ - if (texUnit->ColorTableEnabled) { - _mesa_lookup_rgba_float(&texUnit->ColorTable, span->end, texels); - } - /* GL_EXT_texture_swizzle */ if (curObj->_Swizzle != SWIZZLE_NOOP) { swizzle_texels(curObj->_Swizzle, span->end, texels); @@ -706,4 +770,6 @@ _swrast_texture_span( GLcontext *ctx, SWspan *span ) span->array->rgba ); } } + + free(primary_rgba); }