X-Git-Url: https://git.libre-soc.org/?a=blobdiff_plain;f=src%2Fmesa%2Fdrivers%2Fdri%2Fi965%2Fbrw_vec4_vs_visitor.cpp;h=058ee3a7dadf63e737ec9b1e0732671c16e505e3;hb=be095e11e41158f91bcb3f6fcbc2e2a91a5d9124;hp=f81ee4e5bc01bb02ebd4e6e67e61da88539dfef7;hpb=014b90221ad5cf833bfdd55b0336771d209f0f1d;p=mesa.git diff --git a/src/mesa/drivers/dri/i965/brw_vec4_vs_visitor.cpp b/src/mesa/drivers/dri/i965/brw_vec4_vs_visitor.cpp index f81ee4e5bc0..058ee3a7dad 100644 --- a/src/mesa/drivers/dri/i965/brw_vec4_vs_visitor.cpp +++ b/src/mesa/drivers/dri/i965/brw_vec4_vs_visitor.cpp @@ -30,121 +30,11 @@ namespace brw { void vec4_vs_visitor::emit_prolog() { - dst_reg sign_recovery_shift; - dst_reg normalize_factor; - dst_reg es3_normalize_factor; - - for (int i = 0; i < VERT_ATTRIB_MAX; i++) { - if (vs_prog_data->inputs_read & BITFIELD64_BIT(i)) { - uint8_t wa_flags = key->gl_attrib_wa_flags[i]; - dst_reg reg(ATTR, i); - dst_reg reg_d = reg; - reg_d.type = BRW_REGISTER_TYPE_D; - dst_reg reg_ud = reg; - reg_ud.type = BRW_REGISTER_TYPE_UD; - - /* Do GL_FIXED rescaling for GLES2.0. Our GL_FIXED attributes - * come in as floating point conversions of the integer values. - */ - if (wa_flags & BRW_ATTRIB_WA_COMPONENT_MASK) { - dst_reg dst = reg; - dst.type = brw_type_for_base_type(glsl_type::vec4_type); - dst.writemask = (1 << (wa_flags & BRW_ATTRIB_WA_COMPONENT_MASK)) - 1; - emit(MUL(dst, src_reg(dst), src_reg(1.0f / 65536.0f))); - } - - /* Do sign recovery for 2101010 formats if required. */ - if (wa_flags & BRW_ATTRIB_WA_SIGN) { - if (sign_recovery_shift.file == BAD_FILE) { - /* shift constant: <22,22,22,30> */ - sign_recovery_shift = dst_reg(this, glsl_type::uvec4_type); - emit(MOV(writemask(sign_recovery_shift, WRITEMASK_XYZ), src_reg(22u))); - emit(MOV(writemask(sign_recovery_shift, WRITEMASK_W), src_reg(30u))); - } - - emit(SHL(reg_ud, src_reg(reg_ud), src_reg(sign_recovery_shift))); - emit(ASR(reg_d, src_reg(reg_d), src_reg(sign_recovery_shift))); - } - - /* Apply BGRA swizzle if required. */ - if (wa_flags & BRW_ATTRIB_WA_BGRA) { - src_reg temp = src_reg(reg); - temp.swizzle = BRW_SWIZZLE4(2,1,0,3); - emit(MOV(reg, temp)); - } - - if (wa_flags & BRW_ATTRIB_WA_NORMALIZE) { - /* ES 3.0 has different rules for converting signed normalized - * fixed-point numbers than desktop GL. - */ - if ((wa_flags & BRW_ATTRIB_WA_SIGN) && !use_legacy_snorm_formula) { - /* According to equation 2.2 of the ES 3.0 specification, - * signed normalization conversion is done by: - * - * f = c / (2^(b-1)-1) - */ - if (es3_normalize_factor.file == BAD_FILE) { - /* mul constant: 1 / (2^(b-1) - 1) */ - es3_normalize_factor = dst_reg(this, glsl_type::vec4_type); - emit(MOV(writemask(es3_normalize_factor, WRITEMASK_XYZ), - src_reg(1.0f / ((1<<9) - 1)))); - emit(MOV(writemask(es3_normalize_factor, WRITEMASK_W), - src_reg(1.0f / ((1<<1) - 1)))); - } - - dst_reg dst = reg; - dst.type = brw_type_for_base_type(glsl_type::vec4_type); - emit(MOV(dst, src_reg(reg_d))); - emit(MUL(dst, src_reg(dst), src_reg(es3_normalize_factor))); - emit_minmax(BRW_CONDITIONAL_GE, dst, src_reg(dst), src_reg(-1.0f)); - } else { - /* The following equations are from the OpenGL 3.2 specification: - * - * 2.1 unsigned normalization - * f = c/(2^n-1) - * - * 2.2 signed normalization - * f = (2c+1)/(2^n-1) - * - * Both of these share a common divisor, which is represented by - * "normalize_factor" in the code below. - */ - if (normalize_factor.file == BAD_FILE) { - /* 1 / (2^b - 1) for b=<10,10,10,2> */ - normalize_factor = dst_reg(this, glsl_type::vec4_type); - emit(MOV(writemask(normalize_factor, WRITEMASK_XYZ), - src_reg(1.0f / ((1<<10) - 1)))); - emit(MOV(writemask(normalize_factor, WRITEMASK_W), - src_reg(1.0f / ((1<<2) - 1)))); - } - - dst_reg dst = reg; - dst.type = brw_type_for_base_type(glsl_type::vec4_type); - emit(MOV(dst, src_reg((wa_flags & BRW_ATTRIB_WA_SIGN) ? reg_d : reg_ud))); - - /* For signed normalization, we want the numerator to be 2c+1. */ - if (wa_flags & BRW_ATTRIB_WA_SIGN) { - emit(MUL(dst, src_reg(dst), src_reg(2.0f))); - emit(ADD(dst, src_reg(dst), src_reg(1.0f))); - } - - emit(MUL(dst, src_reg(dst), src_reg(normalize_factor))); - } - } - - if (wa_flags & BRW_ATTRIB_WA_SCALE) { - dst_reg dst = reg; - dst.type = brw_type_for_base_type(glsl_type::vec4_type); - emit(MOV(dst, src_reg((wa_flags & BRW_ATTRIB_WA_SIGN) ? reg_d : reg_ud))); - } - } - } } dst_reg * -vec4_vs_visitor::make_reg_for_system_value(int location, - const glsl_type *type) +vec4_vs_visitor::make_reg_for_system_value(int location) { /* VertexID is stored by the VF as the last vertex element, but * we don't represent it with a flag in inputs_read, so we call @@ -155,7 +45,11 @@ vec4_vs_visitor::make_reg_for_system_value(int location, switch (location) { case SYSTEM_VALUE_BASE_VERTEX: reg->writemask = WRITEMASK_X; - vs_prog_data->uses_vertexid = true; + vs_prog_data->uses_basevertex = true; + break; + case SYSTEM_VALUE_BASE_INSTANCE: + reg->writemask = WRITEMASK_Y; + vs_prog_data->uses_baseinstance = true; break; case SYSTEM_VALUE_VERTEX_ID: case SYSTEM_VALUE_VERTEX_ID_ZERO_BASE: @@ -166,6 +60,11 @@ vec4_vs_visitor::make_reg_for_system_value(int location, reg->writemask = WRITEMASK_W; vs_prog_data->uses_instanceid = true; break; + case SYSTEM_VALUE_DRAW_ID: + reg = new(mem_ctx) dst_reg(ATTR, VERT_ATTRIB_MAX + 1); + reg->writemask = WRITEMASK_X; + vs_prog_data->uses_drawid = true; + break; default: unreachable("not reached"); } @@ -201,6 +100,32 @@ vec4_vs_visitor::emit_urb_write_opcode(bool complete) } +void +vec4_vs_visitor::emit_urb_slot(dst_reg reg, int varying) +{ + reg.type = BRW_REGISTER_TYPE_F; + output_reg[varying].type = reg.type; + + switch (varying) { + case VARYING_SLOT_COL0: + case VARYING_SLOT_COL1: + case VARYING_SLOT_BFC0: + case VARYING_SLOT_BFC1: { + /* These built-in varyings are only supported in compatibility mode, + * and we only support GS in core profile. So, this must be a vertex + * shader. + */ + vec4_instruction *inst = emit_generic_urb_slot(reg, varying); + if (inst && key->clamp_vertex_color) + inst->saturate = true; + break; + } + default: + return vec4_visitor::emit_urb_slot(reg, varying); + } +} + + void vec4_vs_visitor::emit_clip_distances(dst_reg reg, int offset) { @@ -221,7 +146,7 @@ vec4_vs_visitor::emit_clip_distances(dst_reg reg, int offset) clip_vertex = VARYING_SLOT_POS; } - for (int i = 0; i + offset < key->base.nr_userclip_plane_consts && i < 4; + for (int i = 0; i + offset < key->nr_userclip_plane_consts && i < 4; ++i) { reg.writemask = 1 << i; emit(DP4(reg, @@ -234,9 +159,7 @@ vec4_vs_visitor::emit_clip_distances(dst_reg reg, int offset) void vec4_vs_visitor::setup_uniform_clipplane_values() { - for (int i = 0; i < key->base.nr_userclip_plane_consts; ++i) { - assert(this->uniforms < uniform_array_size); - this->uniform_vector_size[this->uniforms] = 4; + for (int i = 0; i < key->nr_userclip_plane_consts; ++i) { this->userplane[i] = dst_reg(UNIFORM, this->uniforms); this->userplane[i].type = BRW_REGISTER_TYPE_F; for (int j = 0; j < 4; ++j) { @@ -254,7 +177,7 @@ vec4_vs_visitor::emit_thread_end() setup_uniform_clipplane_values(); /* Lower legacy ff and ClipVertex clipping to clip distances */ - if (key->base.nr_userclip_plane_consts > 0) { + if (key->nr_userclip_plane_consts > 0) { current_annotation = "user clip distances"; output_reg[VARYING_SLOT_CLIP_DIST0] = dst_reg(this, glsl_type::vec4_type); @@ -276,20 +199,15 @@ vec4_vs_visitor::vec4_vs_visitor(const struct brw_compiler *compiler, void *log_data, const struct brw_vs_prog_key *key, struct brw_vs_prog_data *vs_prog_data, - struct gl_vertex_program *vp, - struct gl_shader_program *prog, + const nir_shader *shader, gl_clip_plane *clip_planes, void *mem_ctx, int shader_time_index, bool use_legacy_snorm_formula) - : vec4_visitor(compiler, log_data, - &vp->Base, &key->base, &vs_prog_data->base, prog, - MESA_SHADER_VERTEX, - mem_ctx, false /* no_spills */, - shader_time_index), + : vec4_visitor(compiler, log_data, &key->tex, &vs_prog_data->base, shader, + mem_ctx, false /* no_spills */, shader_time_index), key(key), vs_prog_data(vs_prog_data), - vp(vp), clip_planes(clip_planes), use_legacy_snorm_formula(use_legacy_snorm_formula) {