X-Git-Url: https://git.libre-soc.org/?a=blobdiff_plain;f=src%2Fmesa%2Fmain%2Fff_fragment_shader.cpp;h=05633d0d7d455b11d9b9db4eedbc2177c6f65c4f;hb=51c1c4d95a05b6eb6fce74e8d624615e4a1b38ab;hp=cfee334472109d870dd89b5863d978c0106c3c90;hpb=599aac95ff2149d881177ed75a48d97d3dcf47bd;p=mesa.git diff --git a/src/mesa/main/ff_fragment_shader.cpp b/src/mesa/main/ff_fragment_shader.cpp index cfee3344721..05633d0d7d4 100644 --- a/src/mesa/main/ff_fragment_shader.cpp +++ b/src/mesa/main/ff_fragment_shader.cpp @@ -1,10 +1,10 @@ /************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * + * Copyright 2007 VMware, Inc. * All Rights Reserved. * Copyright 2009 VMware, Inc. All Rights Reserved. * Copyright © 2010-2011 Intel Corporation - * + * * 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 @@ -12,45 +12,45 @@ * distribute, sub license, 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 (including the * next paragraph) 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 NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS 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. - * + * **************************************************************************/ -extern "C" { -#include "glheader.h" -#include "imports.h" -#include "mtypes.h" -#include "main/uniforms.h" +#include "main/glheader.h" +#include "main/context.h" + #include "main/macros.h" +#include "main/samplerobj.h" +#include "main/shaderobj.h" +#include "main/state.h" +#include "main/texenvprogram.h" +#include "main/texobj.h" +#include "main/uniforms.h" +#include "compiler/glsl/ir_builder.h" +#include "compiler/glsl/ir_optimization.h" +#include "compiler/glsl/glsl_parser_extras.h" +#include "compiler/glsl/glsl_symbol_table.h" +#include "compiler/glsl_types.h" +#include "program/ir_to_mesa.h" #include "program/program.h" -#include "program/prog_parameter.h" +#include "program/programopt.h" #include "program/prog_cache.h" #include "program/prog_instruction.h" +#include "program/prog_parameter.h" #include "program/prog_print.h" #include "program/prog_statevars.h" -#include "program/programopt.h" -#include "texenvprogram.h" -} -#include "main/uniforms.h" -#include "../glsl/glsl_types.h" -#include "../glsl/ir.h" -#include "../glsl/ir_builder.h" -#include "../glsl/glsl_symbol_table.h" -#include "../glsl/glsl_parser_extras.h" -#include "../glsl/ir_optimization.h" -#include "../glsl/ir_print_visitor.h" -#include "../program/ir_to_mesa.h" +#include "util/bitscan.h" using namespace ir_builder; @@ -69,14 +69,6 @@ using namespace ir_builder; */ -struct texenvprog_cache_item -{ - GLuint hash; - void *key; - struct gl_shader_program *data; - struct texenvprog_cache_item *next; -}; - static GLboolean texenv_doing_secondary_color(struct gl_context *ctx) { @@ -90,21 +82,9 @@ texenv_doing_secondary_color(struct gl_context *ctx) return GL_FALSE; } -struct mode_opt { -#ifdef __GNUC__ - __extension__ GLubyte Source:4; /**< SRC_x */ - __extension__ GLubyte Operand:3; /**< OPR_x */ -#else - GLubyte Source; /**< SRC_x */ - GLubyte Operand; /**< OPR_x */ -#endif -}; - struct state_key { - GLuint nr_enabled_units:8; - GLuint enabled_units:8; + GLuint nr_enabled_units:4; GLuint separate_specular:1; - GLuint fog_enabled:1; GLuint fog_mode:2; /**< FOG_x */ GLuint inputs_available:12; GLuint num_draw_buffers:4; @@ -114,148 +94,23 @@ struct state_key { GLuint enabled:1; GLuint source_index:4; /**< TEXTURE_x_INDEX */ GLuint shadow:1; + + /*** + * These are taken from struct gl_tex_env_combine_packed + * @{ + */ + GLuint ModeRGB:4; + GLuint ModeA:4; GLuint ScaleShiftRGB:2; GLuint ScaleShiftA:2; - - GLuint NumArgsRGB:3; /**< up to MAX_COMBINER_TERMS */ - GLuint ModeRGB:5; /**< MODE_x */ - - GLuint NumArgsA:3; /**< up to MAX_COMBINER_TERMS */ - GLuint ModeA:5; /**< MODE_x */ - - struct mode_opt OptRGB[MAX_COMBINER_TERMS]; - struct mode_opt OptA[MAX_COMBINER_TERMS]; - } unit[MAX_TEXTURE_UNITS]; + GLuint NumArgsRGB:3; + GLuint NumArgsA:3; + struct gl_tex_env_argument ArgsRGB[MAX_COMBINER_TERMS]; + struct gl_tex_env_argument ArgsA[MAX_COMBINER_TERMS]; + /** @} */ + } unit[MAX_TEXTURE_COORD_UNITS]; }; -#define FOG_LINEAR 0 -#define FOG_EXP 1 -#define FOG_EXP2 2 -#define FOG_UNKNOWN 3 - -static GLuint translate_fog_mode( GLenum mode ) -{ - switch (mode) { - case GL_LINEAR: return FOG_LINEAR; - case GL_EXP: return FOG_EXP; - case GL_EXP2: return FOG_EXP2; - default: return FOG_UNKNOWN; - } -} - -#define OPR_SRC_COLOR 0 -#define OPR_ONE_MINUS_SRC_COLOR 1 -#define OPR_SRC_ALPHA 2 -#define OPR_ONE_MINUS_SRC_ALPHA 3 -#define OPR_ZERO 4 -#define OPR_ONE 5 -#define OPR_UNKNOWN 7 - -static GLuint translate_operand( GLenum operand ) -{ - switch (operand) { - case GL_SRC_COLOR: return OPR_SRC_COLOR; - case GL_ONE_MINUS_SRC_COLOR: return OPR_ONE_MINUS_SRC_COLOR; - case GL_SRC_ALPHA: return OPR_SRC_ALPHA; - case GL_ONE_MINUS_SRC_ALPHA: return OPR_ONE_MINUS_SRC_ALPHA; - case GL_ZERO: return OPR_ZERO; - case GL_ONE: return OPR_ONE; - default: - assert(0); - return OPR_UNKNOWN; - } -} - -#define SRC_TEXTURE 0 -#define SRC_TEXTURE0 1 -#define SRC_TEXTURE1 2 -#define SRC_TEXTURE2 3 -#define SRC_TEXTURE3 4 -#define SRC_TEXTURE4 5 -#define SRC_TEXTURE5 6 -#define SRC_TEXTURE6 7 -#define SRC_TEXTURE7 8 -#define SRC_CONSTANT 9 -#define SRC_PRIMARY_COLOR 10 -#define SRC_PREVIOUS 11 -#define SRC_ZERO 12 -#define SRC_UNKNOWN 15 - -static GLuint translate_source( GLenum src ) -{ - switch (src) { - case GL_TEXTURE: return SRC_TEXTURE; - case GL_TEXTURE0: - case GL_TEXTURE1: - case GL_TEXTURE2: - case GL_TEXTURE3: - case GL_TEXTURE4: - case GL_TEXTURE5: - case GL_TEXTURE6: - case GL_TEXTURE7: return SRC_TEXTURE0 + (src - GL_TEXTURE0); - case GL_CONSTANT: return SRC_CONSTANT; - case GL_PRIMARY_COLOR: return SRC_PRIMARY_COLOR; - case GL_PREVIOUS: return SRC_PREVIOUS; - case GL_ZERO: - return SRC_ZERO; - default: - assert(0); - return SRC_UNKNOWN; - } -} - -#define MODE_REPLACE 0 /* r = a0 */ -#define MODE_MODULATE 1 /* r = a0 * a1 */ -#define MODE_ADD 2 /* r = a0 + a1 */ -#define MODE_ADD_SIGNED 3 /* r = a0 + a1 - 0.5 */ -#define MODE_INTERPOLATE 4 /* r = a0 * a2 + a1 * (1 - a2) */ -#define MODE_SUBTRACT 5 /* r = a0 - a1 */ -#define MODE_DOT3_RGB 6 /* r = a0 . a1 */ -#define MODE_DOT3_RGB_EXT 7 /* r = a0 . a1 */ -#define MODE_DOT3_RGBA 8 /* r = a0 . a1 */ -#define MODE_DOT3_RGBA_EXT 9 /* r = a0 . a1 */ -#define MODE_MODULATE_ADD_ATI 10 /* r = a0 * a2 + a1 */ -#define MODE_MODULATE_SIGNED_ADD_ATI 11 /* r = a0 * a2 + a1 - 0.5 */ -#define MODE_MODULATE_SUBTRACT_ATI 12 /* r = a0 * a2 - a1 */ -#define MODE_ADD_PRODUCTS 13 /* r = a0 * a1 + a2 * a3 */ -#define MODE_ADD_PRODUCTS_SIGNED 14 /* r = a0 * a1 + a2 * a3 - 0.5 */ -#define MODE_BUMP_ENVMAP_ATI 15 /* special */ -#define MODE_UNKNOWN 16 - -/** - * Translate GL combiner state into a MODE_x value - */ -static GLuint translate_mode( GLenum envMode, GLenum mode ) -{ - switch (mode) { - case GL_REPLACE: return MODE_REPLACE; - case GL_MODULATE: return MODE_MODULATE; - case GL_ADD: - if (envMode == GL_COMBINE4_NV) - return MODE_ADD_PRODUCTS; - else - return MODE_ADD; - case GL_ADD_SIGNED: - if (envMode == GL_COMBINE4_NV) - return MODE_ADD_PRODUCTS_SIGNED; - else - return MODE_ADD_SIGNED; - case GL_INTERPOLATE: return MODE_INTERPOLATE; - case GL_SUBTRACT: return MODE_SUBTRACT; - case GL_DOT3_RGB: return MODE_DOT3_RGB; - case GL_DOT3_RGB_EXT: return MODE_DOT3_RGB_EXT; - case GL_DOT3_RGBA: return MODE_DOT3_RGBA; - case GL_DOT3_RGBA_EXT: return MODE_DOT3_RGBA_EXT; - case GL_MODULATE_ADD_ATI: return MODE_MODULATE_ADD_ATI; - case GL_MODULATE_SIGNED_ADD_ATI: return MODE_MODULATE_SIGNED_ADD_ATI; - case GL_MODULATE_SUBTRACT_ATI: return MODE_MODULATE_SUBTRACT_ATI; - case GL_BUMP_ENVMAP_ATI: return MODE_BUMP_ENVMAP_ATI; - default: - assert(0); - return MODE_UNKNOWN; - } -} - /** * Do we need to clamp the results of the given texture env/combine mode? @@ -266,23 +121,22 @@ static GLboolean need_saturate( GLuint mode ) { switch (mode) { - case MODE_REPLACE: - case MODE_MODULATE: - case MODE_INTERPOLATE: + case TEXENV_MODE_REPLACE: + case TEXENV_MODE_MODULATE: + case TEXENV_MODE_INTERPOLATE: return GL_FALSE; - case MODE_ADD: - case MODE_ADD_SIGNED: - case MODE_SUBTRACT: - case MODE_DOT3_RGB: - case MODE_DOT3_RGB_EXT: - case MODE_DOT3_RGBA: - case MODE_DOT3_RGBA_EXT: - case MODE_MODULATE_ADD_ATI: - case MODE_MODULATE_SIGNED_ADD_ATI: - case MODE_MODULATE_SUBTRACT_ATI: - case MODE_ADD_PRODUCTS: - case MODE_ADD_PRODUCTS_SIGNED: - case MODE_BUMP_ENVMAP_ATI: + case TEXENV_MODE_ADD: + case TEXENV_MODE_ADD_SIGNED: + case TEXENV_MODE_SUBTRACT: + case TEXENV_MODE_DOT3_RGB: + case TEXENV_MODE_DOT3_RGB_EXT: + case TEXENV_MODE_DOT3_RGBA: + case TEXENV_MODE_DOT3_RGBA_EXT: + case TEXENV_MODE_MODULATE_ADD_ATI: + case TEXENV_MODE_MODULATE_SIGNED_ADD_ATI: + case TEXENV_MODE_MODULATE_SUBTRACT_ATI: + case TEXENV_MODE_ADD_PRODUCTS_NV: + case TEXENV_MODE_ADD_PRODUCTS_SIGNED_NV: return GL_TRUE; default: assert(0); @@ -290,20 +144,7 @@ need_saturate( GLuint mode ) } } - - -/** - * Translate TEXTURE_x_BIT to TEXTURE_x_INDEX. - */ -static GLuint translate_tex_src_bit( GLbitfield bit ) -{ - ASSERT(bit); - return ffs(bit) - 1; -} - - #define VERT_BIT_TEX_ANY (0xff << VERT_ATTRIB_TEX0) -#define VERT_RESULT_TEX_ANY (0xff << VERT_RESULT_TEX0) /** * Identify all possible varying inputs. The fragment program will @@ -311,102 +152,106 @@ static GLuint translate_tex_src_bit( GLbitfield bit ) * constants instead. * * This function figures out all the inputs that the fragment program - * has access to. The bitmask is later reduced to just those which - * are actually referenced. + * has access to and filters input bitmask. */ -static GLbitfield get_fp_input_mask( struct gl_context *ctx ) +static GLbitfield filter_fp_input_mask( GLbitfield fp_inputs, + struct gl_context *ctx ) { - /* _NEW_PROGRAM */ - const GLboolean vertexShader = - (ctx->Shader.CurrentVertexProgram && - ctx->Shader.CurrentVertexProgram->LinkStatus && - ctx->Shader.CurrentVertexProgram->_LinkedShaders[MESA_SHADER_VERTEX]); - const GLboolean vertexProgram = ctx->VertexProgram._Enabled; - GLbitfield fp_inputs = 0x0; - if (ctx->VertexProgram._Overriden) { /* Somebody's messing with the vertex program and we don't have * a clue what's happening. Assume that it could be producing * all possible outputs. */ - fp_inputs = ~0; + return fp_inputs; } - else if (ctx->RenderMode == GL_FEEDBACK) { + + if (ctx->RenderMode == GL_FEEDBACK) { /* _NEW_RENDERMODE */ - fp_inputs = (FRAG_BIT_COL0 | FRAG_BIT_TEX0); + return fp_inputs & (VARYING_BIT_COL0 | VARYING_BIT_TEX0); } - else if (!(vertexProgram || vertexShader)) { + + /* _NEW_PROGRAM */ + const GLboolean vertexShader = + ctx->_Shader->CurrentProgram[MESA_SHADER_VERTEX] != NULL; + const GLboolean vertexProgram = _mesa_arb_vertex_program_enabled(ctx); + + if (!(vertexProgram || vertexShader)) { /* Fixed function vertex logic */ - /* _NEW_ARRAY */ - GLbitfield64 varying_inputs = ctx->varying_vp_inputs; + GLbitfield possible_inputs = 0; + + /* _NEW_VARYING_VP_INPUTS */ + GLbitfield varying_inputs = ctx->varying_vp_inputs; + /* We only update ctx->varying_vp_inputs when in VP_MODE_FF _VPMode */ + assert(VP_MODE_FF == ctx->VertexProgram._VPMode); /* These get generated in the setup routine regardless of the * vertex program: */ /* _NEW_POINT */ - if (ctx->Point.PointSprite) - varying_inputs |= FRAG_BITS_TEX_ANY; + if (ctx->Point.PointSprite) { + /* All texture varyings are possible to use */ + possible_inputs = VARYING_BITS_TEX_ANY; + } + else { + /* _NEW_TEXTURE_STATE */ + const GLbitfield possible_tex_inputs = + ctx->Texture._TexGenEnabled | + ctx->Texture._TexMatEnabled | + ((varying_inputs & VERT_BIT_TEX_ANY) >> VERT_ATTRIB_TEX0); + + possible_inputs = (possible_tex_inputs << VARYING_SLOT_TEX0); + } /* First look at what values may be computed by the generated * vertex program: */ /* _NEW_LIGHT */ if (ctx->Light.Enabled) { - fp_inputs |= FRAG_BIT_COL0; + possible_inputs |= VARYING_BIT_COL0; if (texenv_doing_secondary_color(ctx)) - fp_inputs |= FRAG_BIT_COL1; + possible_inputs |= VARYING_BIT_COL1; } - /* _NEW_TEXTURE */ - fp_inputs |= (ctx->Texture._TexGenEnabled | - ctx->Texture._TexMatEnabled) << FRAG_ATTRIB_TEX0; - /* Then look at what might be varying as a result of enabled * arrays, etc: */ if (varying_inputs & VERT_BIT_COLOR0) - fp_inputs |= FRAG_BIT_COL0; + possible_inputs |= VARYING_BIT_COL0; if (varying_inputs & VERT_BIT_COLOR1) - fp_inputs |= FRAG_BIT_COL1; - - fp_inputs |= (((varying_inputs & VERT_BIT_TEX_ANY) >> VERT_ATTRIB_TEX0) - << FRAG_ATTRIB_TEX0); + possible_inputs |= VARYING_BIT_COL1; + return fp_inputs & possible_inputs; } - else { - /* calculate from vp->outputs */ - struct gl_program *vprog; - GLbitfield64 vp_outputs; - /* Choose GLSL vertex shader over ARB vertex program. Need this - * since vertex shader state validation comes after fragment state - * validation (see additional comments in state.c). - */ - if (vertexShader) - vprog = ctx->Shader.CurrentVertexProgram->_LinkedShaders[MESA_SHADER_VERTEX]->Program; - else - vprog = &ctx->VertexProgram.Current->Base; - - vp_outputs = vprog->OutputsWritten; + /* calculate from vp->outputs */ + struct gl_program *vprog; - /* These get generated in the setup routine regardless of the - * vertex program: - */ - /* _NEW_POINT */ - if (ctx->Point.PointSprite) - vp_outputs |= FRAG_BITS_TEX_ANY; + /* Choose GLSL vertex shader over ARB vertex program. Need this + * since vertex shader state validation comes after fragment state + * validation (see additional comments in state.c). + */ + if (ctx->_Shader->CurrentProgram[MESA_SHADER_GEOMETRY] != NULL) + vprog = ctx->_Shader->CurrentProgram[MESA_SHADER_GEOMETRY]; + else if (ctx->_Shader->CurrentProgram[MESA_SHADER_TESS_EVAL] != NULL) + vprog = ctx->_Shader->CurrentProgram[MESA_SHADER_TESS_EVAL]; + else if (vertexShader) + vprog = ctx->_Shader->CurrentProgram[MESA_SHADER_VERTEX]; + else + vprog = ctx->VertexProgram.Current; - if (vp_outputs & (1 << VERT_RESULT_COL0)) - fp_inputs |= FRAG_BIT_COL0; - if (vp_outputs & (1 << VERT_RESULT_COL1)) - fp_inputs |= FRAG_BIT_COL1; + GLbitfield possible_inputs = vprog->info.outputs_written; - fp_inputs |= (((vp_outputs & VERT_RESULT_TEX_ANY) >> VERT_RESULT_TEX0) - << FRAG_ATTRIB_TEX0); + /* These get generated in the setup routine regardless of the + * vertex program: + */ + /* _NEW_POINT */ + if (ctx->Point.PointSprite) { + /* All texture varyings are possible to use */ + possible_inputs |= VARYING_BITS_TEX_ANY; } - - return fp_inputs; + + return fp_inputs & possible_inputs; } @@ -416,79 +261,58 @@ static GLbitfield get_fp_input_mask( struct gl_context *ctx ) */ static GLuint make_state_key( struct gl_context *ctx, struct state_key *key ) { - GLuint i, j; - GLbitfield inputs_referenced = FRAG_BIT_COL0; - const GLbitfield inputs_available = get_fp_input_mask( ctx ); + GLbitfield inputs_referenced = VARYING_BIT_COL0; + GLbitfield mask; GLuint keySize; memset(key, 0, sizeof(*key)); - /* _NEW_TEXTURE */ - for (i = 0; i < ctx->Const.MaxTextureUnits; i++) { + /* _NEW_TEXTURE_OBJECT */ + mask = ctx->Texture._EnabledCoordUnits; + int i = -1; + while (mask) { + i = u_bit_scan(&mask); const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[i]; const struct gl_texture_object *texObj = texUnit->_Current; - const struct gl_tex_env_combine_state *comb = texUnit->_CurrentCombine; - GLenum format; + const struct gl_tex_env_combine_packed *comb = + &ctx->Texture.FixedFuncUnit[i]._CurrentCombinePacked; - if (!texUnit->_ReallyEnabled || !texUnit->Enabled) + if (!texObj) continue; - format = texObj->Image[0][texObj->BaseLevel]->_BaseFormat; - key->unit[i].enabled = 1; - key->enabled_units |= (1<nr_enabled_units = i + 1; - inputs_referenced |= FRAG_BIT_TEX(i); - - key->unit[i].source_index = - translate_tex_src_bit(texUnit->_ReallyEnabled); - - key->unit[i].shadow = - ((texObj->Sampler.CompareMode == GL_COMPARE_R_TO_TEXTURE) && - ((format == GL_DEPTH_COMPONENT) || - (format == GL_DEPTH_STENCIL_EXT))); + inputs_referenced |= VARYING_BIT_TEX(i); - key->unit[i].NumArgsRGB = comb->_NumArgsRGB; - key->unit[i].NumArgsA = comb->_NumArgsA; + key->unit[i].source_index = texObj->TargetIndex; - key->unit[i].ModeRGB = - translate_mode(texUnit->EnvMode, comb->ModeRGB); - key->unit[i].ModeA = - translate_mode(texUnit->EnvMode, comb->ModeA); + const struct gl_sampler_object *samp = _mesa_get_samplerobj(ctx, i); + if (samp->CompareMode == GL_COMPARE_R_TO_TEXTURE) { + const GLenum format = _mesa_texture_base_format(texObj); + key->unit[i].shadow = (format == GL_DEPTH_COMPONENT || + format == GL_DEPTH_STENCIL_EXT); + } + key->unit[i].ModeRGB = comb->ModeRGB; + key->unit[i].ModeA = comb->ModeA; key->unit[i].ScaleShiftRGB = comb->ScaleShiftRGB; key->unit[i].ScaleShiftA = comb->ScaleShiftA; + key->unit[i].NumArgsRGB = comb->NumArgsRGB; + key->unit[i].NumArgsA = comb->NumArgsA; - for (j = 0; j < MAX_COMBINER_TERMS; j++) { - key->unit[i].OptRGB[j].Operand = translate_operand(comb->OperandRGB[j]); - key->unit[i].OptA[j].Operand = translate_operand(comb->OperandA[j]); - key->unit[i].OptRGB[j].Source = translate_source(comb->SourceRGB[j]); - key->unit[i].OptA[j].Source = translate_source(comb->SourceA[j]); - } - - if (key->unit[i].ModeRGB == MODE_BUMP_ENVMAP_ATI) { - /* requires some special translation */ - key->unit[i].NumArgsRGB = 2; - key->unit[i].ScaleShiftRGB = 0; - key->unit[i].OptRGB[0].Operand = OPR_SRC_COLOR; - key->unit[i].OptRGB[0].Source = SRC_TEXTURE; - key->unit[i].OptRGB[1].Operand = OPR_SRC_COLOR; - key->unit[i].OptRGB[1].Source = texUnit->BumpTarget - GL_TEXTURE0 + SRC_TEXTURE0; - } + memcpy(key->unit[i].ArgsRGB, comb->ArgsRGB, sizeof comb->ArgsRGB); + memcpy(key->unit[i].ArgsA, comb->ArgsA, sizeof comb->ArgsA); } + key->nr_enabled_units = i + 1; + /* _NEW_LIGHT | _NEW_FOG */ if (texenv_doing_secondary_color(ctx)) { key->separate_specular = 1; - inputs_referenced |= FRAG_BIT_COL1; + inputs_referenced |= VARYING_BIT_COL1; } /* _NEW_FOG */ - if (ctx->Fog.Enabled) { - key->fog_enabled = 1; - key->fog_mode = translate_fog_mode(ctx->Fog.Mode); - inputs_referenced |= FRAG_BIT_FOGC; /* maybe */ - } + key->fog_mode = ctx->Fog._PackedEnabledMode; /* _NEW_BUFFERS */ key->num_draw_buffers = ctx->DrawBuffer->_NumColorDrawBuffers; @@ -499,7 +323,7 @@ static GLuint make_state_key( struct gl_context *ctx, struct state_key *key ) key->num_draw_buffers = 1; } - key->inputs_available = (inputs_available & inputs_referenced); + key->inputs_available = filter_fp_input_mask(inputs_referenced, ctx); /* compute size of state key, ignoring unused texture units */ keySize = sizeof(*key) - sizeof(key->unit) @@ -511,12 +335,11 @@ static GLuint make_state_key( struct gl_context *ctx, struct state_key *key ) /** State used to build the fragment program: */ -struct texenv_fragment_program { +class texenv_fragment_program : public ir_factory { +public: struct gl_shader_program *shader_program; struct gl_shader *shader; - exec_list *instructions; exec_list *top_instructions; - void *mem_ctx; struct state_key *state; ir_variable *src_texture[MAX_TEXTURE_COORD_UNITS]; @@ -537,22 +360,23 @@ struct texenv_fragment_program { }; static ir_rvalue * -get_current_attrib(struct texenv_fragment_program *p, GLuint attrib) +get_current_attrib(texenv_fragment_program *p, GLuint attrib) { ir_variable *current; ir_rvalue *val; current = p->shader->symbols->get_variable("gl_CurrentAttribFragMESA"); - current->max_array_access = MAX2(current->max_array_access, attrib); + assert(current); + current->data.max_array_access = MAX2(current->data.max_array_access, (int)attrib); val = new(p->mem_ctx) ir_dereference_variable(current); ir_rvalue *index = new(p->mem_ctx) ir_constant(attrib); return new(p->mem_ctx) ir_dereference_array(val, index); } static ir_rvalue * -get_gl_Color(struct texenv_fragment_program *p) +get_gl_Color(texenv_fragment_program *p) { - if (p->state->inputs_available & FRAG_BIT_COL0) { + if (p->state->inputs_available & VARYING_BIT_COL0) { ir_variable *var = p->shader->symbols->get_variable("gl_Color"); assert(var); return new(p->mem_ctx) ir_dereference_variable(var); @@ -562,44 +386,47 @@ get_gl_Color(struct texenv_fragment_program *p) } static ir_rvalue * -get_source(struct texenv_fragment_program *p, +get_source(texenv_fragment_program *p, GLuint src, GLuint unit) { ir_variable *var; ir_dereference *deref; switch (src) { - case SRC_TEXTURE: + case TEXENV_SRC_TEXTURE: return new(p->mem_ctx) ir_dereference_variable(p->src_texture[unit]); - case SRC_TEXTURE0: - case SRC_TEXTURE1: - case SRC_TEXTURE2: - case SRC_TEXTURE3: - case SRC_TEXTURE4: - case SRC_TEXTURE5: - case SRC_TEXTURE6: - case SRC_TEXTURE7: + case TEXENV_SRC_TEXTURE0: + case TEXENV_SRC_TEXTURE1: + case TEXENV_SRC_TEXTURE2: + case TEXENV_SRC_TEXTURE3: + case TEXENV_SRC_TEXTURE4: + case TEXENV_SRC_TEXTURE5: + case TEXENV_SRC_TEXTURE6: + case TEXENV_SRC_TEXTURE7: return new(p->mem_ctx) - ir_dereference_variable(p->src_texture[src - SRC_TEXTURE0]); + ir_dereference_variable(p->src_texture[src - TEXENV_SRC_TEXTURE0]); - case SRC_CONSTANT: + case TEXENV_SRC_CONSTANT: var = p->shader->symbols->get_variable("gl_TextureEnvColor"); assert(var); deref = new(p->mem_ctx) ir_dereference_variable(var); - var->max_array_access = MAX2(var->max_array_access, unit); + var->data.max_array_access = MAX2(var->data.max_array_access, (int)unit); return new(p->mem_ctx) ir_dereference_array(deref, new(p->mem_ctx) ir_constant(unit)); - case SRC_PRIMARY_COLOR: + case TEXENV_SRC_PRIMARY_COLOR: var = p->shader->symbols->get_variable("gl_Color"); assert(var); return new(p->mem_ctx) ir_dereference_variable(var); - case SRC_ZERO: + case TEXENV_SRC_ZERO: return new(p->mem_ctx) ir_constant(0.0f); - case SRC_PREVIOUS: + case TEXENV_SRC_ONE: + return new(p->mem_ctx) ir_constant(1.0f); + + case TEXENV_SRC_PREVIOUS: if (!p->src_previous) { return get_gl_Color(p); } else { @@ -613,7 +440,7 @@ get_source(struct texenv_fragment_program *p, } static ir_rvalue * -emit_combine_source(struct texenv_fragment_program *p, +emit_combine_source(texenv_fragment_program *p, GLuint unit, GLuint source, GLuint operand) @@ -623,26 +450,21 @@ emit_combine_source(struct texenv_fragment_program *p, src = get_source(p, source, unit); switch (operand) { - case OPR_ONE_MINUS_SRC_COLOR: + case TEXENV_OPR_ONE_MINUS_COLOR: return sub(new(p->mem_ctx) ir_constant(1.0f), src); - case OPR_SRC_ALPHA: - return src->type->is_scalar() - ? src : (ir_rvalue *) new(p->mem_ctx) ir_swizzle(src, 3, 3, 3, 3, 1); + case TEXENV_OPR_ALPHA: + return src->type->is_scalar() ? src : swizzle_w(src); - case OPR_ONE_MINUS_SRC_ALPHA: { - ir_rvalue *const scalar = (src->type->is_scalar()) - ? src : (ir_rvalue *) new(p->mem_ctx) ir_swizzle(src, 3, 3, 3, 3, 1); + case TEXENV_OPR_ONE_MINUS_ALPHA: { + ir_rvalue *const scalar = src->type->is_scalar() ? src : swizzle_w(src); return sub(new(p->mem_ctx) ir_constant(1.0f), scalar); } - case OPR_ZERO: - return new(p->mem_ctx) ir_constant(0.0f); - case OPR_ONE: - return new(p->mem_ctx) ir_constant(1.0f); - case OPR_SRC_COLOR: + case TEXENV_OPR_COLOR: return src; + default: assert(0); return src; @@ -659,29 +481,29 @@ static GLboolean args_match( const struct state_key *key, GLuint unit ) GLuint i, numArgs = key->unit[unit].NumArgsRGB; for (i = 0; i < numArgs; i++) { - if (key->unit[unit].OptA[i].Source != key->unit[unit].OptRGB[i].Source) + if (key->unit[unit].ArgsA[i].Source != key->unit[unit].ArgsRGB[i].Source) return GL_FALSE; - switch (key->unit[unit].OptA[i].Operand) { - case OPR_SRC_ALPHA: - switch (key->unit[unit].OptRGB[i].Operand) { - case OPR_SRC_COLOR: - case OPR_SRC_ALPHA: + switch (key->unit[unit].ArgsA[i].Operand) { + case TEXENV_OPR_ALPHA: + switch (key->unit[unit].ArgsRGB[i].Operand) { + case TEXENV_OPR_COLOR: + case TEXENV_OPR_ALPHA: break; default: return GL_FALSE; } break; - case OPR_ONE_MINUS_SRC_ALPHA: - switch (key->unit[unit].OptRGB[i].Operand) { - case OPR_ONE_MINUS_SRC_COLOR: - case OPR_ONE_MINUS_SRC_ALPHA: + case TEXENV_OPR_ONE_MINUS_ALPHA: + switch (key->unit[unit].ArgsRGB[i].Operand) { + case TEXENV_OPR_ONE_MINUS_COLOR: + case TEXENV_OPR_ONE_MINUS_ALPHA: break; default: return GL_FALSE; } break; - default: + default: return GL_FALSE; /* impossible */ } } @@ -690,20 +512,20 @@ static GLboolean args_match( const struct state_key *key, GLuint unit ) } static ir_rvalue * -smear(struct texenv_fragment_program *p, ir_rvalue *val) +smear(ir_rvalue *val) { if (!val->type->is_scalar()) return val; - return new(p->mem_ctx) ir_swizzle(val, 0, 0, 0, 0, 4); + return swizzle_xxxx(val); } static ir_rvalue * -emit_combine(struct texenv_fragment_program *p, +emit_combine(texenv_fragment_program *p, GLuint unit, GLuint nr, GLuint mode, - const struct mode_opt *opt) + const struct gl_tex_env_argument *opt) { ir_rvalue *src[MAX_COMBINER_TERMS]; ir_rvalue *tmp0, *tmp1; @@ -715,64 +537,57 @@ emit_combine(struct texenv_fragment_program *p, src[i] = emit_combine_source( p, unit, opt[i].Source, opt[i].Operand ); switch (mode) { - case MODE_REPLACE: + case TEXENV_MODE_REPLACE: return src[0]; - case MODE_MODULATE: + case TEXENV_MODE_MODULATE: return mul(src[0], src[1]); - case MODE_ADD: + case TEXENV_MODE_ADD: return add(src[0], src[1]); - case MODE_ADD_SIGNED: + case TEXENV_MODE_ADD_SIGNED: return add(add(src[0], src[1]), new(p->mem_ctx) ir_constant(-0.5f)); - case MODE_INTERPOLATE: + case TEXENV_MODE_INTERPOLATE: /* Arg0 * (Arg2) + Arg1 * (1-Arg2) */ tmp0 = mul(src[0], src[2]); tmp1 = mul(src[1], sub(new(p->mem_ctx) ir_constant(1.0f), src[2]->clone(p->mem_ctx, NULL))); return add(tmp0, tmp1); - case MODE_SUBTRACT: + case TEXENV_MODE_SUBTRACT: return sub(src[0], src[1]); - case MODE_DOT3_RGBA: - case MODE_DOT3_RGBA_EXT: - case MODE_DOT3_RGB_EXT: - case MODE_DOT3_RGB: { + case TEXENV_MODE_DOT3_RGBA: + case TEXENV_MODE_DOT3_RGBA_EXT: + case TEXENV_MODE_DOT3_RGB_EXT: + case TEXENV_MODE_DOT3_RGB: { tmp0 = mul(src[0], new(p->mem_ctx) ir_constant(2.0f)); tmp0 = add(tmp0, new(p->mem_ctx) ir_constant(-1.0f)); - tmp0 = new(p->mem_ctx) ir_swizzle(smear(p, tmp0), 0, 1, 2, 3, 3); tmp1 = mul(src[1], new(p->mem_ctx) ir_constant(2.0f)); tmp1 = add(tmp1, new(p->mem_ctx) ir_constant(-1.0f)); - tmp1 = new(p->mem_ctx) ir_swizzle(smear(p, tmp1), 0, 1, 2, 3, 3); - return dot(tmp0, tmp1); + return dot(swizzle_xyz(smear(tmp0)), swizzle_xyz(smear(tmp1))); } - case MODE_MODULATE_ADD_ATI: + case TEXENV_MODE_MODULATE_ADD_ATI: return add(mul(src[0], src[2]), src[1]); - case MODE_MODULATE_SIGNED_ADD_ATI: + case TEXENV_MODE_MODULATE_SIGNED_ADD_ATI: return add(add(mul(src[0], src[2]), src[1]), new(p->mem_ctx) ir_constant(-0.5f)); - case MODE_MODULATE_SUBTRACT_ATI: + case TEXENV_MODE_MODULATE_SUBTRACT_ATI: return sub(mul(src[0], src[2]), src[1]); - case MODE_ADD_PRODUCTS: + case TEXENV_MODE_ADD_PRODUCTS_NV: return add(mul(src[0], src[1]), mul(src[2], src[3])); - case MODE_ADD_PRODUCTS_SIGNED: + case TEXENV_MODE_ADD_PRODUCTS_SIGNED_NV: return add(add(mul(src[0], src[1]), mul(src[2], src[3])), new(p->mem_ctx) ir_constant(-0.5f)); - - case MODE_BUMP_ENVMAP_ATI: - /* special - not handled here */ - assert(0); - return src[0]; - default: + default: assert(0); return src[0]; } @@ -782,26 +597,22 @@ emit_combine(struct texenv_fragment_program *p, * Generate instructions for one texture unit's env/combiner mode. */ static ir_rvalue * -emit_texenv(struct texenv_fragment_program *p, GLuint unit) +emit_texenv(texenv_fragment_program *p, GLuint unit) { const struct state_key *key = p->state; GLboolean rgb_saturate, alpha_saturate; GLuint rgb_shift, alpha_shift; if (!key->unit[unit].enabled) { - return get_source(p, SRC_PREVIOUS, 0); - } - if (key->unit[unit].ModeRGB == MODE_BUMP_ENVMAP_ATI) { - /* this isn't really a env stage delivering a color and handled elsewhere */ - return get_source(p, SRC_PREVIOUS, 0); + return get_source(p, TEXENV_SRC_PREVIOUS, 0); } - + switch (key->unit[unit].ModeRGB) { - case MODE_DOT3_RGB_EXT: + case TEXENV_MODE_DOT3_RGB_EXT: alpha_shift = key->unit[unit].ScaleShiftA; rgb_shift = 0; break; - case MODE_DOT3_RGBA_EXT: + case TEXENV_MODE_DOT3_RGBA_EXT: alpha_shift = 0; rgb_shift = 0; break; @@ -810,7 +621,7 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit) alpha_shift = key->unit[unit].ScaleShiftA; break; } - + /* If we'll do rgb/alpha shifting don't saturate in emit_combine(). * We don't want to clamp twice. */ @@ -828,13 +639,8 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit) else alpha_saturate = GL_FALSE; - ir_variable *temp_var = new(p->mem_ctx) ir_variable(glsl_type::vec4_type, - "texenv_combine", - ir_var_temporary); - p->instructions->push_tail(temp_var); - + ir_variable *temp_var = p->make_temp(glsl_type::vec4_type, "texenv_combine"); ir_dereference *deref; - ir_assignment *assign; ir_rvalue *val; /* Emit the RGB and A combine ops @@ -844,27 +650,23 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit) val = emit_combine(p, unit, key->unit[unit].NumArgsRGB, key->unit[unit].ModeRGB, - key->unit[unit].OptRGB); - val = smear(p, val); + key->unit[unit].ArgsRGB); + val = smear(val); if (rgb_saturate) val = saturate(val); - deref = new(p->mem_ctx) ir_dereference_variable(temp_var); - assign = new(p->mem_ctx) ir_assignment(deref, val); - p->instructions->push_tail(assign); + p->emit(assign(temp_var, val)); } - else if (key->unit[unit].ModeRGB == MODE_DOT3_RGBA_EXT || - key->unit[unit].ModeRGB == MODE_DOT3_RGBA) { + else if (key->unit[unit].ModeRGB == TEXENV_MODE_DOT3_RGBA_EXT || + key->unit[unit].ModeRGB == TEXENV_MODE_DOT3_RGBA) { ir_rvalue *val = emit_combine(p, unit, key->unit[unit].NumArgsRGB, key->unit[unit].ModeRGB, - key->unit[unit].OptRGB); - val = smear(p, val); + key->unit[unit].ArgsRGB); + val = smear(val); if (rgb_saturate) val = saturate(val); - deref = new(p->mem_ctx) ir_dereference_variable(temp_var); - assign = new(p->mem_ctx) ir_assignment(deref, val); - p->instructions->push_tail(assign); + p->emit(assign(temp_var, val)); } else { /* Need to do something to stop from re-emitting identical @@ -873,26 +675,20 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit) val = emit_combine(p, unit, key->unit[unit].NumArgsRGB, key->unit[unit].ModeRGB, - key->unit[unit].OptRGB); - val = smear(p, val); - val = new(p->mem_ctx) ir_swizzle(val, 0, 1, 2, 3, 3); + key->unit[unit].ArgsRGB); + val = swizzle_xyz(smear(val)); if (rgb_saturate) val = saturate(val); - deref = new(p->mem_ctx) ir_dereference_variable(temp_var); - assign = new(p->mem_ctx) ir_assignment(deref, val, NULL, WRITEMASK_XYZ); - p->instructions->push_tail(assign); + p->emit(assign(temp_var, val, WRITEMASK_XYZ)); val = emit_combine(p, unit, key->unit[unit].NumArgsA, key->unit[unit].ModeA, - key->unit[unit].OptA); - val = smear(p, val); - val = new(p->mem_ctx) ir_swizzle(val, 3, 3, 3, 3, 1); + key->unit[unit].ArgsA); + val = swizzle_w(smear(val)); if (alpha_saturate) val = saturate(val); - deref = new(p->mem_ctx) ir_dereference_variable(temp_var); - assign = new(p->mem_ctx) ir_assignment(deref, val, NULL, WRITEMASK_W); - p->instructions->push_tail(assign); + p->emit(assign(temp_var, val, WRITEMASK_W)); } deref = new(p->mem_ctx) ir_dereference_variable(temp_var); @@ -906,14 +702,15 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit) shift = new(p->mem_ctx) ir_constant((float)(1 << rgb_shift)); } else { - float const_data[4] = { - 1 << rgb_shift, - 1 << rgb_shift, - 1 << rgb_shift, - 1 << alpha_shift - }; - shift = new(p->mem_ctx) ir_constant(glsl_type::vec4_type, - (ir_constant_data *)const_data); + ir_constant_data const_data; + + const_data.f[0] = float(1 << rgb_shift); + const_data.f[1] = float(1 << rgb_shift); + const_data.f[2] = float(1 << rgb_shift); + const_data.f[3] = float(1 << alpha_shift); + + shift = new(p->mem_ctx) ir_constant(glsl_type::vec4_type, + &const_data); } return saturate(mul(deref, shift)); @@ -926,10 +723,9 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit) /** * Generate instruction for getting a texture source term. */ -static void load_texture( struct texenv_fragment_program *p, GLuint unit ) +static void load_texture( texenv_fragment_program *p, GLuint unit ) { ir_dereference *deref; - ir_assignment *assign; if (p->src_texture[unit]) return; @@ -937,7 +733,7 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit ) const GLuint texTarget = p->state->unit[unit].source_index; ir_rvalue *texcoord; - if (!(p->state->inputs_available & (FRAG_BIT_TEX0 << unit))) { + if (!(p->state->inputs_available & (VARYING_BIT_TEX0 << unit))) { texcoord = get_current_attrib(p, VERT_ATTRIB_TEX0 + unit); } else if (p->texcoord_tex[unit]) { texcoord = new(p->mem_ctx) ir_dereference_variable(p->texcoord_tex[unit]); @@ -947,19 +743,15 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit ) texcoord = new(p->mem_ctx) ir_dereference_variable(tc_array); ir_rvalue *index = new(p->mem_ctx) ir_constant(unit); texcoord = new(p->mem_ctx) ir_dereference_array(texcoord, index); - tc_array->max_array_access = MAX2(tc_array->max_array_access, unit); + tc_array->data.max_array_access = MAX2(tc_array->data.max_array_access, (int)unit); } if (!p->state->unit[unit].enabled) { - p->src_texture[unit] = new(p->mem_ctx) ir_variable(glsl_type::vec4_type, - "dummy_tex", - ir_var_temporary); - p->instructions->push_tail(p->src_texture[unit]); - - deref = new(p->mem_ctx) ir_dereference_variable(p->src_texture[unit]); - assign = new(p->mem_ctx) ir_assignment(deref, - new(p->mem_ctx) ir_constant(0.0f)); - p->instructions->push_tail(assign); + p->src_texture[unit] = p->make_temp(glsl_type::vec4_type, + "dummy_tex"); + p->emit(p->src_texture[unit]); + + p->emit(assign(p->src_texture[unit], new(p->mem_ctx) ir_constant(0.0f))); return ; } @@ -969,62 +761,60 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit ) switch (texTarget) { case TEXTURE_1D_INDEX: if (p->state->unit[unit].shadow) - sampler_type = p->shader->symbols->get_type("sampler1DShadow"); + sampler_type = glsl_type::sampler1DShadow_type; else - sampler_type = p->shader->symbols->get_type("sampler1D"); + sampler_type = glsl_type::sampler1D_type; coords = 1; break; case TEXTURE_1D_ARRAY_INDEX: if (p->state->unit[unit].shadow) - sampler_type = p->shader->symbols->get_type("sampler1DArrayShadow"); + sampler_type = glsl_type::sampler1DArrayShadow_type; else - sampler_type = p->shader->symbols->get_type("sampler1DArray"); + sampler_type = glsl_type::sampler1DArray_type; coords = 2; break; case TEXTURE_2D_INDEX: if (p->state->unit[unit].shadow) - sampler_type = p->shader->symbols->get_type("sampler2DShadow"); + sampler_type = glsl_type::sampler2DShadow_type; else - sampler_type = p->shader->symbols->get_type("sampler2D"); + sampler_type = glsl_type::sampler2D_type; coords = 2; break; case TEXTURE_2D_ARRAY_INDEX: if (p->state->unit[unit].shadow) - sampler_type = p->shader->symbols->get_type("sampler2DArrayShadow"); + sampler_type = glsl_type::sampler2DArrayShadow_type; else - sampler_type = p->shader->symbols->get_type("sampler2DArray"); + sampler_type = glsl_type::sampler2DArray_type; coords = 3; break; case TEXTURE_RECT_INDEX: if (p->state->unit[unit].shadow) - sampler_type = p->shader->symbols->get_type("sampler2DRectShadow"); + sampler_type = glsl_type::sampler2DRectShadow_type; else - sampler_type = p->shader->symbols->get_type("sampler2DRect"); + sampler_type = glsl_type::sampler2DRect_type; coords = 2; break; case TEXTURE_3D_INDEX: assert(!p->state->unit[unit].shadow); - sampler_type = p->shader->symbols->get_type("sampler3D"); + sampler_type = glsl_type::sampler3D_type; coords = 3; break; case TEXTURE_CUBE_INDEX: if (p->state->unit[unit].shadow) - sampler_type = p->shader->symbols->get_type("samplerCubeShadow"); + sampler_type = glsl_type::samplerCubeShadow_type; else - sampler_type = p->shader->symbols->get_type("samplerCube"); + sampler_type = glsl_type::samplerCube_type; coords = 3; break; case TEXTURE_EXTERNAL_INDEX: assert(!p->state->unit[unit].shadow); - sampler_type = p->shader->symbols->get_type("samplerExternalOES"); + sampler_type = glsl_type::samplerExternalOES_type; coords = 2; break; } - p->src_texture[unit] = new(p->mem_ctx) ir_variable(glsl_type::vec4_type, - "tex", - ir_var_temporary); - p->instructions->push_tail(p->src_texture[unit]); + p->src_texture[unit] = p->make_temp(glsl_type::vec4_type, + "tex"); ir_texture *tex = new(p->mem_ctx) ir_texture(ir_tex); @@ -1034,6 +824,13 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit ) sampler_name, ir_var_uniform); p->top_instructions->push_head(sampler); + + /* Set the texture unit for this sampler in the same way that + * layout(binding=X) would. + */ + sampler->data.explicit_binding = true; + sampler->data.binding = unit; + deref = new(p->mem_ctx) ir_dereference_variable(sampler); tex->set_sampler(deref, glsl_type::vec4_type); @@ -1041,40 +838,38 @@ static void load_texture( struct texenv_fragment_program *p, GLuint unit ) if (p->state->unit[unit].shadow) { texcoord = texcoord->clone(p->mem_ctx, NULL); - tex->shadow_comparitor = new(p->mem_ctx) ir_swizzle(texcoord, + tex->shadow_comparator = new(p->mem_ctx) ir_swizzle(texcoord, coords, 0, 0, 0, 1); coords++; } texcoord = texcoord->clone(p->mem_ctx, NULL); - tex->projector = new(p->mem_ctx) ir_swizzle(texcoord, 3, 0, 0, 0, 1); + tex->projector = swizzle_w(texcoord); - deref = new(p->mem_ctx) ir_dereference_variable(p->src_texture[unit]); - assign = new(p->mem_ctx) ir_assignment(deref, tex); - p->instructions->push_tail(assign); + p->emit(assign(p->src_texture[unit], tex)); } static void -load_texenv_source(struct texenv_fragment_program *p, +load_texenv_source(texenv_fragment_program *p, GLuint src, GLuint unit) { switch (src) { - case SRC_TEXTURE: + case TEXENV_SRC_TEXTURE: load_texture(p, unit); break; - case SRC_TEXTURE0: - case SRC_TEXTURE1: - case SRC_TEXTURE2: - case SRC_TEXTURE3: - case SRC_TEXTURE4: - case SRC_TEXTURE5: - case SRC_TEXTURE6: - case SRC_TEXTURE7: - load_texture(p, src - SRC_TEXTURE0); + case TEXENV_SRC_TEXTURE0: + case TEXENV_SRC_TEXTURE1: + case TEXENV_SRC_TEXTURE2: + case TEXENV_SRC_TEXTURE3: + case TEXENV_SRC_TEXTURE4: + case TEXENV_SRC_TEXTURE5: + case TEXENV_SRC_TEXTURE6: + case TEXENV_SRC_TEXTURE7: + load_texture(p, src - TEXENV_SRC_TEXTURE0); break; - + default: /* not a texture src - do nothing */ break; @@ -1086,93 +881,22 @@ load_texenv_source(struct texenv_fragment_program *p, * Generate instructions for loading all texture source terms. */ static GLboolean -load_texunit_sources( struct texenv_fragment_program *p, GLuint unit ) +load_texunit_sources( texenv_fragment_program *p, GLuint unit ) { const struct state_key *key = p->state; GLuint i; for (i = 0; i < key->unit[unit].NumArgsRGB; i++) { - load_texenv_source( p, key->unit[unit].OptRGB[i].Source, unit ); + load_texenv_source( p, key->unit[unit].ArgsRGB[i].Source, unit ); } for (i = 0; i < key->unit[unit].NumArgsA; i++) { - load_texenv_source( p, key->unit[unit].OptA[i].Source, unit ); + load_texenv_source( p, key->unit[unit].ArgsA[i].Source, unit ); } return GL_TRUE; } -/** - * Generate instructions for loading bump map textures. - */ -static void -load_texunit_bumpmap( struct texenv_fragment_program *p, GLuint unit ) -{ - const struct state_key *key = p->state; - GLuint bumpedUnitNr = key->unit[unit].OptRGB[1].Source - SRC_TEXTURE0; - ir_rvalue *bump; - ir_rvalue *texcoord; - ir_variable *rot_mat_0_var, *rot_mat_1_var; - ir_dereference_variable *rot_mat_0, *rot_mat_1; - - rot_mat_0_var = p->shader->symbols->get_variable("gl_BumpRotMatrix0MESA"); - rot_mat_1_var = p->shader->symbols->get_variable("gl_BumpRotMatrix1MESA"); - rot_mat_0 = new(p->mem_ctx) ir_dereference_variable(rot_mat_0_var); - rot_mat_1 = new(p->mem_ctx) ir_dereference_variable(rot_mat_1_var); - - ir_variable *tc_array = p->shader->symbols->get_variable("gl_TexCoord"); - assert(tc_array); - texcoord = new(p->mem_ctx) ir_dereference_variable(tc_array); - ir_rvalue *index = new(p->mem_ctx) ir_constant(bumpedUnitNr); - texcoord = new(p->mem_ctx) ir_dereference_array(texcoord, index); - tc_array->max_array_access = MAX2(tc_array->max_array_access, unit); - - load_texenv_source( p, unit + SRC_TEXTURE0, unit ); - - /* Apply rot matrix and add coords to be available in next phase. - * dest = Arg1 + (Arg0.xx * rotMat0) + (Arg0.yy * rotMat1) - * note only 2 coords are affected the rest are left unchanged (mul by 0) - */ - ir_dereference *deref; - ir_assignment *assign; - ir_rvalue *bump_x, *bump_y; - - texcoord = smear(p, texcoord); - - /* bump_texcoord = texcoord */ - ir_variable *bumped = new(p->mem_ctx) ir_variable(texcoord->type, - "bump_texcoord", - ir_var_temporary); - p->instructions->push_tail(bumped); - - deref = new(p->mem_ctx) ir_dereference_variable(bumped); - assign = new(p->mem_ctx) ir_assignment(deref, texcoord); - p->instructions->push_tail(assign); - - /* bump_texcoord.xy += arg0.x * rotmat0 + arg0.y * rotmat1 */ - bump = get_source(p, key->unit[unit].OptRGB[0].Source, unit); - bump_x = new(p->mem_ctx) ir_swizzle(bump, 0, 0, 0, 0, 1); - bump = bump->clone(p->mem_ctx, NULL); - bump_y = new(p->mem_ctx) ir_swizzle(bump, 1, 0, 0, 0, 1); - - bump_x = mul(bump_x, rot_mat_0); - bump_y = mul(bump_y, rot_mat_1); - - ir_expression *expr; - - deref = new(p->mem_ctx) ir_dereference_variable(bumped); - expr = add(new(p->mem_ctx) ir_swizzle(deref, - 0, 1, 1, 1, - 2), - add(bump_x, bump_y)); - - deref = new(p->mem_ctx) ir_dereference_variable(bumped); - assign = new(p->mem_ctx) ir_assignment(deref, expr, NULL, WRITEMASK_XY); - p->instructions->push_tail(assign); - - p->texcoord_tex[bumpedUnitNr] = bumped; -} - /** * Applies the fog calculations. * @@ -1181,38 +905,32 @@ load_texunit_bumpmap( struct texenv_fragment_program *p, GLuint unit ) * GL_FOG_COORDINATE_EXT is set to GL_FRAGMENT_DEPTH_EXT. */ static ir_rvalue * -emit_fog_instructions(struct texenv_fragment_program *p, +emit_fog_instructions(texenv_fragment_program *p, ir_rvalue *fragcolor) { struct state_key *key = p->state; ir_rvalue *f, *temp; ir_variable *params, *oparams; ir_variable *fogcoord; - ir_assignment *assign; /* Temporary storage for the whole fog result. Fog calculations * only affect rgb so we're hanging on to the .a value of fragcolor * this way. */ - ir_variable *fog_result = new(p->mem_ctx) ir_variable(glsl_type::vec4_type, - "fog_result", - ir_var_auto); - p->instructions->push_tail(fog_result); - temp = new(p->mem_ctx) ir_dereference_variable(fog_result); - assign = new(p->mem_ctx) ir_assignment(temp, fragcolor); - p->instructions->push_tail(assign); + ir_variable *fog_result = p->make_temp(glsl_type::vec4_type, "fog_result"); + p->emit(assign(fog_result, fragcolor)); - temp = new(p->mem_ctx) ir_dereference_variable(fog_result); - fragcolor = new(p->mem_ctx) ir_swizzle(temp, 0, 1, 2, 3, 3); + fragcolor = swizzle_xyz(fog_result); oparams = p->shader->symbols->get_variable("gl_FogParamsOptimizedMESA"); + assert(oparams); fogcoord = p->shader->symbols->get_variable("gl_FogFragCoord"); + assert(fogcoord); params = p->shader->symbols->get_variable("gl_Fog"); + assert(params); f = new(p->mem_ctx) ir_dereference_variable(fogcoord); - ir_variable *f_var = new(p->mem_ctx) ir_variable(glsl_type::float_type, - "fog_factor", ir_var_auto); - p->instructions->push_tail(f_var); + ir_variable *f_var = p->make_temp(glsl_type::float_type, "fog_factor"); switch (key->fog_mode) { case FOG_LINEAR: @@ -1221,13 +939,7 @@ emit_fog_instructions(struct texenv_fragment_program *p, * gl_MesaFogParamsOptimized gives us (-1 / (end - start)) and * (end / (end - start)) so we can generate a single MAD. */ - temp = new(p->mem_ctx) ir_dereference_variable(oparams); - temp = new(p->mem_ctx) ir_swizzle(temp, 0, 0, 0, 0, 1); - f = mul(f, temp); - - temp = new(p->mem_ctx) ir_dereference_variable(oparams); - temp = new(p->mem_ctx) ir_swizzle(temp, 1, 0, 0, 0, 1); - f = add(f, temp); + f = add(mul(f, swizzle_x(oparams)), swizzle_y(oparams)); break; case FOG_EXP: /* f = e^(-(density * fogcoord)) @@ -1236,9 +948,7 @@ emit_fog_instructions(struct texenv_fragment_program *p, * use EXP2 which is generally the native instruction without * having to do any further math on the fog density uniform. */ - temp = new(p->mem_ctx) ir_dereference_variable(oparams); - temp = new(p->mem_ctx) ir_swizzle(temp, 2, 0, 0, 0, 1); - f = mul(f, temp); + f = mul(f, swizzle_z(oparams)); f = new(p->mem_ctx) ir_expression(ir_unop_neg, f); f = new(p->mem_ctx) ir_expression(ir_unop_exp2, f); break; @@ -1249,65 +959,34 @@ emit_fog_instructions(struct texenv_fragment_program *p, * can do this like FOG_EXP but with a squaring after the * multiply by density. */ - ir_variable *temp_var = new(p->mem_ctx) ir_variable(glsl_type::float_type, - "fog_temp", - ir_var_auto); - p->instructions->push_tail(temp_var); - - temp = new(p->mem_ctx) ir_dereference_variable(oparams); - temp = new(p->mem_ctx) ir_swizzle(temp, 3, 0, 0, 0, 1); - f = mul(f, temp); - - temp = new(p->mem_ctx) ir_dereference_variable(temp_var); - ir_assignment *assign = new(p->mem_ctx) ir_assignment(temp, f); - p->instructions->push_tail(assign); - - f = new(p->mem_ctx) ir_dereference_variable(temp_var); - temp = new(p->mem_ctx) ir_dereference_variable(temp_var); - f = mul(f, temp); + ir_variable *temp_var = p->make_temp(glsl_type::float_type, "fog_temp"); + p->emit(assign(temp_var, mul(f, swizzle_w(oparams)))); + + f = mul(temp_var, temp_var); f = new(p->mem_ctx) ir_expression(ir_unop_neg, f); f = new(p->mem_ctx) ir_expression(ir_unop_exp2, f); break; } - f = saturate(f); - - temp = new(p->mem_ctx) ir_dereference_variable(f_var); - assign = new(p->mem_ctx) ir_assignment(temp, f); - p->instructions->push_tail(assign); + p->emit(assign(f_var, saturate(f))); - f = new(p->mem_ctx) ir_dereference_variable(f_var); - f = sub(new(p->mem_ctx) ir_constant(1.0f), f); + f = sub(new(p->mem_ctx) ir_constant(1.0f), f_var); temp = new(p->mem_ctx) ir_dereference_variable(params); temp = new(p->mem_ctx) ir_dereference_record(temp, "color"); - temp = new(p->mem_ctx) ir_swizzle(temp, 0, 1, 2, 3, 3); - temp = mul(temp, f); + temp = mul(swizzle_xyz(temp), f); - f = new(p->mem_ctx) ir_dereference_variable(f_var); - f = add(temp, mul(fragcolor, f)); - - ir_dereference *deref = new(p->mem_ctx) ir_dereference_variable(fog_result); - assign = new(p->mem_ctx) ir_assignment(deref, f, NULL, WRITEMASK_XYZ); - p->instructions->push_tail(assign); + p->emit(assign(fog_result, add(temp, mul(fragcolor, f_var)), WRITEMASK_XYZ)); return new(p->mem_ctx) ir_dereference_variable(fog_result); } static void -emit_instructions(struct texenv_fragment_program *p) +emit_instructions(texenv_fragment_program *p) { struct state_key *key = p->state; GLuint unit; - if (key->enabled_units) { - /* Zeroth pass - bump map textures first */ - for (unit = 0; unit < key->nr_enabled_units; unit++) { - if (key->unit[unit].enabled && - key->unit[unit].ModeRGB == MODE_BUMP_ENVMAP_ATI) { - load_texunit_bumpmap(p, unit); - } - } - + if (key->nr_enabled_units) { /* First pass - to support texture_env_crossbar, first identify * all referenced texture sources and emit texld instructions * for each: @@ -1326,54 +1005,36 @@ emit_instructions(struct texenv_fragment_program *p) } } - ir_rvalue *cf = get_source(p, SRC_PREVIOUS, 0); - ir_dereference_variable *deref; - ir_assignment *assign; + ir_rvalue *cf = get_source(p, TEXENV_SRC_PREVIOUS, 0); if (key->separate_specular) { - ir_rvalue *tmp0; - ir_variable *spec_result = new(p->mem_ctx) ir_variable(glsl_type::vec4_type, - "specular_add", - ir_var_temporary); - - p->instructions->push_tail(spec_result); - - deref = new(p->mem_ctx) ir_dereference_variable(spec_result); - assign = new(p->mem_ctx) ir_assignment(deref, cf); - p->instructions->push_tail(assign); - - deref = new(p->mem_ctx) ir_dereference_variable(spec_result); - tmp0 = new(p->mem_ctx) ir_swizzle(deref, 0, 1, 2, 3, 3); + ir_variable *spec_result = p->make_temp(glsl_type::vec4_type, + "specular_add"); + p->emit(assign(spec_result, cf)); ir_rvalue *secondary; - if (p->state->inputs_available & FRAG_BIT_COL1) { + if (p->state->inputs_available & VARYING_BIT_COL1) { ir_variable *var = p->shader->symbols->get_variable("gl_SecondaryColor"); assert(var); - secondary = new(p->mem_ctx) ir_dereference_variable(var); + secondary = swizzle_xyz(var); } else { - secondary = get_current_attrib(p, VERT_ATTRIB_COLOR1); + secondary = swizzle_xyz(get_current_attrib(p, VERT_ATTRIB_COLOR1)); } - secondary = new(p->mem_ctx) ir_swizzle(secondary, 0, 1, 2, 3, 3); - - tmp0 = add(tmp0, secondary); - deref = new(p->mem_ctx) ir_dereference_variable(spec_result); - assign = new(p->mem_ctx) ir_assignment(deref, tmp0, NULL, WRITEMASK_XYZ); - p->instructions->push_tail(assign); + p->emit(assign(spec_result, add(swizzle_xyz(spec_result), secondary), + WRITEMASK_XYZ)); cf = new(p->mem_ctx) ir_dereference_variable(spec_result); } - if (key->fog_enabled) { + if (key->fog_mode) { cf = emit_fog_instructions(p, cf); } ir_variable *frag_color = p->shader->symbols->get_variable("gl_FragColor"); assert(frag_color); - deref = new(p->mem_ctx) ir_dereference_variable(frag_color); - assign = new(p->mem_ctx) ir_assignment(deref, cf); - p->instructions->push_tail(assign); + p->emit(assign(frag_color, cf)); } /** @@ -1383,21 +1044,23 @@ emit_instructions(struct texenv_fragment_program *p) static struct gl_shader_program * create_new_program(struct gl_context *ctx, struct state_key *key) { - struct texenv_fragment_program p; + texenv_fragment_program p; unsigned int unit; _mesa_glsl_parse_state *state; - memset(&p, 0, sizeof(p)); p.mem_ctx = ralloc_context(NULL); - p.shader = ctx->Driver.NewShader(ctx, 0, GL_FRAGMENT_SHADER); + p.shader = _mesa_new_shader(0, MESA_SHADER_FRAGMENT); +#ifdef DEBUG + p.shader->SourceChecksum = 0xf18ed; /* fixed */ +#endif p.shader->ir = new(p.shader) exec_list; - state = new(p.shader) _mesa_glsl_parse_state(ctx, GL_FRAGMENT_SHADER, + state = new(p.shader) _mesa_glsl_parse_state(ctx, MESA_SHADER_FRAGMENT, p.shader); p.shader->symbols = state->symbols; p.top_instructions = p.shader->ir; p.instructions = p.shader->ir; p.state = key; - p.shader_program = ctx->Driver.NewShaderProgram(ctx, 0); + p.shader_program = _mesa_new_shader_program(0); /* Tell the linker to ignore the fact that we're building a * separate shader, in case we're in a GLES2 context that would @@ -1405,10 +1068,23 @@ create_new_program(struct gl_context *ctx, struct state_key *key) * fixed function program in a GLES2 context at all, but that's a * big mess to clean up. */ - p.shader_program->InternalSeparateShader = GL_TRUE; + p.shader_program->SeparateShader = GL_TRUE; + + /* The legacy GLSL shadow functions follow the depth texture + * mode and return vec4. The GLSL 1.30 shadow functions return float and + * ignore the depth texture mode. That's a shader and state dependency + * that's difficult to deal with. st/mesa uses a simple but not + * completely correct solution: if the shader declares GLSL >= 1.30 and + * the depth texture mode is GL_ALPHA (000X), it sets the XXXX swizzle + * instead. Thus, the GLSL 1.30 shadow function will get the result in .x + * and legacy shadow functions will get it in .w as expected. + * For the fixed-function fragment shader, use 120 to get correct behavior + * for GL_ALPHA. + */ + state->language_version = 120; - state->language_version = 130; - if (ctx->Extensions.OES_EGL_image_external) + state->es_shader = false; + if (_mesa_is_gles(ctx) && ctx->Extensions.OES_EGL_image_external) state->OES_EGL_image_external_enable = true; _mesa_glsl_initialize_types(state); _mesa_glsl_initialize_variables(p.instructions, state); @@ -1421,11 +1097,11 @@ create_new_program(struct gl_context *ctx, struct state_key *key) p.src_previous = NULL; ir_function *main_f = new(p.mem_ctx) ir_function("main"); - p.instructions->push_tail(main_f); + p.emit(main_f); state->symbols->add_function(main_f); ir_function_signature *main_sig = - new(p.mem_ctx) ir_function_signature(p.shader->symbols->get_type("void")); + new(p.mem_ctx) ir_function_signature(glsl_type::void_type); main_sig->is_defined = true; main_f->add_signature(main_sig); @@ -1435,13 +1111,20 @@ create_new_program(struct gl_context *ctx, struct state_key *key) validate_ir_tree(p.shader->ir); - while (do_common_optimization(p.shader->ir, false, false, 32)) - ; + const struct gl_shader_compiler_options *options = + &ctx->Const.ShaderCompilerOptions[MESA_SHADER_FRAGMENT]; + + /* Conservative approach: Don't optimize here, the linker does it too. */ + if (!ctx->Const.GLSLOptimizeConservatively) { + while (do_common_optimization(p.shader->ir, false, false, options, + ctx->Const.NativeIntegers)) + ; + } + reparent_ir(p.shader->ir, p.shader->ir); - p.shader->CompileStatus = true; + p.shader->CompileStatus = COMPILE_SUCCESS; p.shader->Version = state->language_version; - p.shader->num_builtins_to_link = state->num_builtins_to_link; p.shader_program->Shaders = (gl_shader **)malloc(sizeof(*p.shader_program->Shaders)); p.shader_program->Shaders[0] = p.shader; @@ -1449,56 +1132,9 @@ create_new_program(struct gl_context *ctx, struct state_key *key) _mesa_glsl_link_shader(ctx, p.shader_program); - /* Set the sampler uniforms, and relink to get them into the linked - * program. - */ - struct gl_shader *const fs = - p.shader_program->_LinkedShaders[MESA_SHADER_FRAGMENT]; - struct gl_program *const fp = fs->Program; - - _mesa_generate_parameters_list_for_uniforms(p.shader_program, fs, - fp->Parameters); - - _mesa_associate_uniform_storage(ctx, p.shader_program, fp->Parameters); - - for (unsigned int i = 0; i < MAX_TEXTURE_UNITS; i++) { - /* Enough space for 'sampler_999\0'. - */ - char name[12]; - - snprintf(name, sizeof(name), "sampler_%d", i); - - int loc = _mesa_get_uniform_location(ctx, p.shader_program, name); - if (loc != -1) { - unsigned base; - unsigned idx; - - /* Avoid using _mesa_uniform() because it flags state - * updates, so if we're generating this shader_program in a - * state update, we end up recursing. Instead, just set the - * value, which is picked up at re-link. - */ - _mesa_uniform_split_location_offset(loc, &base, &idx); - assert(idx == 0); - - struct gl_uniform_storage *const storage = - &p.shader_program->UniformStorage[base]; - - /* Update the storage, the SamplerUnits in the shader program, and - * the SamplerUnits in the assembly shader. - */ - storage->storage[idx].i = i; - fp->SamplerUnits[storage->sampler] = i; - p.shader_program->SamplerUnits[storage->sampler] = i; - _mesa_propagate_uniforms_to_driver_storage(storage, 0, 1); - } - } - _mesa_update_shader_textures_used(p.shader_program, fp); - (void) ctx->Driver.ProgramStringNotify(ctx, fp->Target, fp); - - if (!p.shader_program->LinkStatus) + if (!p.shader_program->data->LinkStatus) _mesa_problem(ctx, "Failed to link fixed function fragment shader: %s\n", - p.shader_program->InfoLog); + p.shader_program->data->InfoLog); ralloc_free(p.mem_ctx); return p.shader_program;