*/
-#include "brw_vs.h"
-#include "main/context.h"
-
+#include "brw_vec4_vs.h"
+#include "common/gen_debug.h"
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 = vs_compile->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(with_writemask(sign_recovery_shift, WRITEMASK_XYZ), src_reg(22u)));
- emit(MOV(with_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 (_mesa_is_gles3(ctx) && (wa_flags & BRW_ATTRIB_WA_SIGN)) {
- /* 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(with_writemask(es3_normalize_factor, WRITEMASK_XYZ),
- src_reg(1.0f / ((1<<9) - 1))));
- emit(MOV(with_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_G, 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(with_writemask(normalize_factor, WRITEMASK_XYZ),
- src_reg(1.0f / ((1<<10) - 1))));
- emit(MOV(with_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(ir_variable *ir)
+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
* it VERT_ATTRIB_MAX, which setup_attributes() picks up on.
*/
dst_reg *reg = new(mem_ctx) dst_reg(ATTR, VERT_ATTRIB_MAX);
- vs_prog_data->uses_vertexid = true;
- switch (ir->location) {
- case SYSTEM_VALUE_VERTEX_ID:
+ switch (location) {
+ case SYSTEM_VALUE_BASE_VERTEX:
reg->writemask = WRITEMASK_X;
+ vs_prog_data->uses_basevertex = true;
break;
- case SYSTEM_VALUE_INSTANCE_ID:
+ case SYSTEM_VALUE_BASE_INSTANCE:
reg->writemask = WRITEMASK_Y;
+ vs_prog_data->uses_baseinstance = true;
break;
- default:
- assert(!"not reached");
+ case SYSTEM_VALUE_VERTEX_ID:
+ case SYSTEM_VALUE_VERTEX_ID_ZERO_BASE:
+ reg->writemask = WRITEMASK_Z;
+ vs_prog_data->uses_vertexid = true;
+ break;
+ case SYSTEM_VALUE_INSTANCE_ID:
+ 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");
}
return reg;
}
+void
+vec4_vs_visitor::emit_urb_slot(dst_reg reg, int varying)
+{
+ reg.type = BRW_REGISTER_TYPE_F;
+ output_reg[varying][0].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, 0);
+ 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)
+{
+ /* From the GLSL 1.30 spec, section 7.1 (Vertex Shader Special Variables):
+ *
+ * "If a linked set of shaders forming the vertex stage contains no
+ * static write to gl_ClipVertex or gl_ClipDistance, but the
+ * application has requested clipping against user clip planes through
+ * the API, then the coordinate written to gl_Position is used for
+ * comparison against the user clip planes."
+ *
+ * This function is only called if the shader didn't write to
+ * gl_ClipDistance. Accordingly, we use gl_ClipVertex to perform clipping
+ * if the user wrote to it; otherwise we use gl_Position.
+ */
+ gl_varying_slot clip_vertex = VARYING_SLOT_CLIP_VERTEX;
+ if (!(prog_data->vue_map.slots_valid & VARYING_BIT_CLIP_VERTEX)) {
+ clip_vertex = VARYING_SLOT_POS;
+ }
+
+ for (int i = 0; i + offset < key->nr_userclip_plane_consts && i < 4;
+ ++i) {
+ reg.writemask = 1 << i;
+ emit(DP4(reg,
+ src_reg(output_reg[clip_vertex][0]),
+ src_reg(this->userplane[i + offset])));
+ }
+}
+
+
+void
+vec4_vs_visitor::setup_uniform_clipplane_values()
+{
+ 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) {
+ stage_prog_data->param[this->uniforms * 4 + j] =
+ (gl_constant_value *) &clip_planes[i][j];
+ }
+ ++this->uniforms;
+ }
+}
+
+
void
vec4_vs_visitor::emit_thread_end()
{
+ setup_uniform_clipplane_values();
+
+ /* Lower legacy ff and ClipVertex clipping to clip distances */
+ if (key->nr_userclip_plane_consts > 0) {
+ current_annotation = "user clip distances";
+
+ output_reg[VARYING_SLOT_CLIP_DIST0][0] =
+ dst_reg(this, glsl_type::vec4_type);
+ output_reg[VARYING_SLOT_CLIP_DIST1][0] =
+ dst_reg(this, glsl_type::vec4_type);
+ output_num_components[VARYING_SLOT_CLIP_DIST0][0] = 4;
+ output_num_components[VARYING_SLOT_CLIP_DIST1][0] = 4;
+
+ emit_clip_distances(output_reg[VARYING_SLOT_CLIP_DIST0][0], 0);
+ emit_clip_distances(output_reg[VARYING_SLOT_CLIP_DIST1][0], 4);
+ }
+
/* For VS, we always end the thread by emitting a single vertex.
* emit_urb_write_opcode() will take care of setting the eot flag on the
* SEND instruction.
}
-vec4_vs_visitor::vec4_vs_visitor(struct brw_context *brw,
- struct brw_vs_compile *vs_compile,
+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_shader_program *prog,
- struct brw_shader *shader,
- void *mem_ctx)
- : vec4_visitor(brw, &vs_compile->base, &vs_compile->vp->program.Base,
- &vs_compile->key.base, &vs_prog_data->base, prog, shader,
- mem_ctx, INTEL_DEBUG & DEBUG_VS, false /* no_spills */),
- vs_compile(vs_compile),
- vs_prog_data(vs_prog_data)
+ 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, &key->tex, &vs_prog_data->base, shader,
+ mem_ctx, false /* no_spills */, shader_time_index),
+ key(key),
+ vs_prog_data(vs_prog_data),
+ clip_planes(clip_planes),
+ use_legacy_snorm_formula(use_legacy_snorm_formula)
{
}