コード例 #1
0
static void
saturate_src(nir_builder *b, nir_tex_instr *tex, unsigned sat_mask)
{
   b->cursor = nir_before_instr(&tex->instr);

   /* Walk through the sources saturating the requested arguments. */
   for (unsigned i = 0; i < tex->num_srcs; i++) {
      if (tex->src[i].src_type != nir_tex_src_coord)
         continue;

      nir_ssa_def *src =
         nir_ssa_for_src(b, tex->src[i].src, tex->coord_components);

      /* split src into components: */
      nir_ssa_def *comp[4];

      for (unsigned j = 0; j < tex->coord_components; j++)
         comp[j] = nir_channel(b, src, j);

      /* clamp requested components, array index does not get clamped: */
      unsigned ncomp = tex->coord_components;
      if (tex->is_array)
         ncomp--;

      for (unsigned j = 0; j < ncomp; j++) {
         if ((1 << j) & sat_mask) {
            if (tex->sampler_dim == GLSL_SAMPLER_DIM_RECT) {
               /* non-normalized texture coords, so clamp to texture
                * size rather than [0.0, 1.0]
                */
               nir_ssa_def *txs = get_texture_size(b, tex);
               comp[j] = nir_fmax(b, comp[j], nir_imm_float(b, 0.0));
               comp[j] = nir_fmin(b, comp[j], nir_channel(b, txs, j));
            } else {
               comp[j] = nir_fsat(b, comp[j]);
            }
         }
      }

      /* and move the result back into a single vecN: */
      src = nir_vec(b, comp, tex->coord_components);

      nir_instr_rewrite_src(&tex->instr,
                            &tex->src[i].src,
                            nir_src_for_ssa(src));
   }
}
コード例 #2
0
static void
lower_load_sample_pos(lower_wpos_ytransform_state *state,
                      nir_intrinsic_instr *intr)
{
    nir_builder *b = &state->b;
    b->cursor = nir_after_instr(&intr->instr);

    nir_ssa_def *pos = &intr->dest.ssa;
    nir_ssa_def *scale = nir_channel(b, get_transform(state), 0);
    nir_ssa_def *neg_scale = nir_channel(b, get_transform(state), 2);
    /* Either y or 1-y for scale equal to 1 or -1 respectively. */
    nir_ssa_def *flipped_y =
        nir_fadd(b, nir_fmax(b, neg_scale, nir_imm_float(b, 0.0)),
                 nir_fmul(b, nir_channel(b, pos, 1), scale));
    nir_ssa_def *flipped_pos = nir_vec2(b, nir_channel(b, pos, 0), flipped_y);

    nir_ssa_def_rewrite_uses_after(&intr->dest.ssa, nir_src_for_ssa(flipped_pos),
                                   flipped_pos->parent_instr);
}
コード例 #3
0
static void
lower_load_pointcoord(lower_wpos_ytransform_state *state,
                      nir_intrinsic_instr *intr)
{
   nir_builder *b = &state->b;
   b->cursor = nir_after_instr(&intr->instr);

   nir_ssa_def *pntc = &intr->dest.ssa;
   nir_ssa_def *transform = get_transform(state);
   nir_ssa_def *y = nir_channel(b, pntc, 1);
   /* The offset is 1 if we're flipping, 0 otherwise. */
   nir_ssa_def *offset = nir_fmax(b, nir_channel(b, transform, 2),
                                  nir_imm_float(b, 0.0));
   /* Flip the sign of y if we're flipping. */
   nir_ssa_def *scaled = nir_fmul(b, y, nir_channel(b, transform, 0));

   /* Reassemble the vector. */
   nir_ssa_def *flipped_pntc = nir_vec2(b,
                                        nir_channel(b, pntc, 0),
                                        nir_fadd(b, offset, scaled));

   nir_ssa_def_rewrite_uses_after(&intr->dest.ssa, nir_src_for_ssa(flipped_pntc),
                                  flipped_pntc->parent_instr);
}
コード例 #4
0
static nir_shader *
build_resolve_compute_shader(struct radv_device *dev, bool is_integer, int samples)
{
	nir_builder b;
	char name[64];
	nir_if *outer_if = NULL;
	const struct glsl_type *sampler_type = glsl_sampler_type(GLSL_SAMPLER_DIM_MS,
								 false,
								 false,
								 GLSL_TYPE_FLOAT);
	const struct glsl_type *img_type = glsl_sampler_type(GLSL_SAMPLER_DIM_2D,
							     false,
							     false,
							     GLSL_TYPE_FLOAT);
	snprintf(name, 64, "meta_resolve_cs-%d-%s", samples, is_integer ? "int" : "float");
	nir_builder_init_simple_shader(&b, NULL, MESA_SHADER_COMPUTE, NULL);
	b.shader->info->name = ralloc_strdup(b.shader, name);
	b.shader->info->cs.local_size[0] = 16;
	b.shader->info->cs.local_size[1] = 16;
	b.shader->info->cs.local_size[2] = 1;

	nir_variable *input_img = nir_variable_create(b.shader, nir_var_uniform,
						      sampler_type, "s_tex");
	input_img->data.descriptor_set = 0;
	input_img->data.binding = 0;

	nir_variable *output_img = nir_variable_create(b.shader, nir_var_uniform,
						       img_type, "out_img");
	output_img->data.descriptor_set = 0;
	output_img->data.binding = 1;
	nir_ssa_def *invoc_id = nir_load_system_value(&b, nir_intrinsic_load_local_invocation_id, 0);
	nir_ssa_def *wg_id = nir_load_system_value(&b, nir_intrinsic_load_work_group_id, 0);
	nir_ssa_def *block_size = nir_imm_ivec4(&b,
						b.shader->info->cs.local_size[0],
						b.shader->info->cs.local_size[1],
						b.shader->info->cs.local_size[2], 0);

	nir_ssa_def *global_id = nir_iadd(&b, nir_imul(&b, wg_id, block_size), invoc_id);

	nir_intrinsic_instr *src_offset = nir_intrinsic_instr_create(b.shader, nir_intrinsic_load_push_constant);
	src_offset->src[0] = nir_src_for_ssa(nir_imm_int(&b, 0));
	src_offset->num_components = 2;
	nir_ssa_dest_init(&src_offset->instr, &src_offset->dest, 2, 32, "src_offset");
	nir_builder_instr_insert(&b, &src_offset->instr);

	nir_intrinsic_instr *dst_offset = nir_intrinsic_instr_create(b.shader, nir_intrinsic_load_push_constant);
	dst_offset->src[0] = nir_src_for_ssa(nir_imm_int(&b, 8));
	dst_offset->num_components = 2;
	nir_ssa_dest_init(&dst_offset->instr, &dst_offset->dest, 2, 32, "dst_offset");
	nir_builder_instr_insert(&b, &dst_offset->instr);

	nir_ssa_def *img_coord = nir_iadd(&b, global_id, &src_offset->dest.ssa);
	/* do a txf_ms on each sample */
	nir_ssa_def *tmp;

	nir_tex_instr *tex = nir_tex_instr_create(b.shader, 2);
	tex->sampler_dim = GLSL_SAMPLER_DIM_MS;
	tex->op = nir_texop_txf_ms;
	tex->src[0].src_type = nir_tex_src_coord;
	tex->src[0].src = nir_src_for_ssa(img_coord);
	tex->src[1].src_type = nir_tex_src_ms_index;
	tex->src[1].src = nir_src_for_ssa(nir_imm_int(&b, 0));
	tex->dest_type = nir_type_float;
	tex->is_array = false;
	tex->coord_components = 2;
	tex->texture = nir_deref_var_create(tex, input_img);
	tex->sampler = NULL;

	nir_ssa_dest_init(&tex->instr, &tex->dest, 4, 32, "tex");
	nir_builder_instr_insert(&b, &tex->instr);

	tmp = &tex->dest.ssa;
	nir_variable *color =
		nir_local_variable_create(b.impl, glsl_vec4_type(), "color");

	if (!is_integer && samples > 1) {
		nir_tex_instr *tex_all_same = nir_tex_instr_create(b.shader, 1);
		tex_all_same->sampler_dim = GLSL_SAMPLER_DIM_MS;
		tex_all_same->op = nir_texop_samples_identical;
		tex_all_same->src[0].src_type = nir_tex_src_coord;
		tex_all_same->src[0].src = nir_src_for_ssa(img_coord);
		tex_all_same->dest_type = nir_type_float;
		tex_all_same->is_array = false;
		tex_all_same->coord_components = 2;
		tex_all_same->texture = nir_deref_var_create(tex_all_same, input_img);
		tex_all_same->sampler = NULL;

		nir_ssa_dest_init(&tex_all_same->instr, &tex_all_same->dest, 1, 32, "tex");
		nir_builder_instr_insert(&b, &tex_all_same->instr);

		nir_ssa_def *all_same = nir_ine(&b, &tex_all_same->dest.ssa, nir_imm_int(&b, 0));
		nir_if *if_stmt = nir_if_create(b.shader);
		if_stmt->condition = nir_src_for_ssa(all_same);
		nir_cf_node_insert(b.cursor, &if_stmt->cf_node);

		b.cursor = nir_after_cf_list(&if_stmt->then_list);
		for (int i = 1; i < samples; i++) {
			nir_tex_instr *tex_add = nir_tex_instr_create(b.shader, 2);
			tex_add->sampler_dim = GLSL_SAMPLER_DIM_MS;
			tex_add->op = nir_texop_txf_ms;
			tex_add->src[0].src_type = nir_tex_src_coord;
			tex_add->src[0].src = nir_src_for_ssa(img_coord);
			tex_add->src[1].src_type = nir_tex_src_ms_index;
			tex_add->src[1].src = nir_src_for_ssa(nir_imm_int(&b, i));
			tex_add->dest_type = nir_type_float;
			tex_add->is_array = false;
			tex_add->coord_components = 2;
			tex_add->texture = nir_deref_var_create(tex_add, input_img);
			tex_add->sampler = NULL;

			nir_ssa_dest_init(&tex_add->instr, &tex_add->dest, 4, 32, "tex");
			nir_builder_instr_insert(&b, &tex_add->instr);

			tmp = nir_fadd(&b, tmp, &tex_add->dest.ssa);
		}

		tmp = nir_fdiv(&b, tmp, nir_imm_float(&b, samples));
		nir_store_var(&b, color, tmp, 0xf);
		b.cursor = nir_after_cf_list(&if_stmt->else_list);
		outer_if = if_stmt;
	}
	nir_store_var(&b, color, &tex->dest.ssa, 0xf);

	if (outer_if)
		b.cursor = nir_after_cf_node(&outer_if->cf_node);

	nir_ssa_def *newv = nir_load_var(&b, color);
	nir_ssa_def *coord = nir_iadd(&b, global_id, &dst_offset->dest.ssa);
	nir_intrinsic_instr *store = nir_intrinsic_instr_create(b.shader, nir_intrinsic_image_store);
	store->src[0] = nir_src_for_ssa(coord);
	store->src[1] = nir_src_for_ssa(nir_ssa_undef(&b, 1, 32));
	store->src[2] = nir_src_for_ssa(newv);
	store->variables[0] = nir_deref_var_create(store, output_img);
	nir_builder_instr_insert(&b, &store->instr);
	return b.shader;
}
コード例 #5
0
}

static void
convert_yuv_to_rgb(nir_builder *b, nir_tex_instr *tex,
                   nir_ssa_def *y, nir_ssa_def *u, nir_ssa_def *v,
                   nir_ssa_def *a)
{
   nir_const_value m[3][4] = {
      { { .f32 = 1.16438356f }, { .f32 =  1.16438356f }, { .f32 = 1.16438356f }, { .f32 = 0.0f } },
      { { .f32 = 0.0f        }, { .f32 = -0.39176229f }, { .f32 = 2.01723214f }, { .f32 = 0.0f } },
      { { .f32 = 1.59602678f }, { .f32 = -0.81296764f }, { .f32 = 0.0f        }, { .f32 = 0.0f } },
   };

   nir_ssa_def *offset =
      nir_vec4(b,
               nir_imm_float(b, -0.874202214f),
               nir_imm_float(b,  0.531667820f),
               nir_imm_float(b, -1.085630787f),
               a);

   nir_ssa_def *result =
      nir_ffma(b, y, nir_build_imm(b, 4, 32, m[0]),
               nir_ffma(b, u, nir_build_imm(b, 4, 32, m[1]),
                        nir_ffma(b, v, nir_build_imm(b, 4, 32, m[2]),
                                 offset)));

   nir_ssa_def_rewrite_uses(&tex->dest.ssa, nir_src_for_ssa(result));
}

static void
lower_y_uv_external(nir_builder *b, nir_tex_instr *tex,
コード例 #6
0
/* NIR equiv of TGSI CMP instruction: */
static nir_ssa_def *
nir_cmp(nir_builder *b, nir_ssa_def *src0, nir_ssa_def *src1, nir_ssa_def *src2)
{
    return nir_bcsel(b, nir_flt(b, src0, nir_imm_float(b, 0.0)), src1, src2);
}
コード例 #7
0
ファイル: vtn_alu.c プロジェクト: elgambitero/Mesa-3D
void
vtn_handle_alu(struct vtn_builder *b, SpvOp opcode,
               const uint32_t *w, unsigned count)
{
   struct vtn_value *val = vtn_push_value(b, w[2], vtn_value_type_ssa);
   const struct glsl_type *type =
      vtn_value(b, w[1], vtn_value_type_type)->type->type;

   vtn_foreach_decoration(b, val, handle_no_contraction, NULL);

   /* Collect the various SSA sources */
   const unsigned num_inputs = count - 3;
   struct vtn_ssa_value *vtn_src[4] = { NULL, };
   for (unsigned i = 0; i < num_inputs; i++)
      vtn_src[i] = vtn_ssa_value(b, w[i + 3]);

   if (glsl_type_is_matrix(vtn_src[0]->type) ||
       (num_inputs >= 2 && glsl_type_is_matrix(vtn_src[1]->type))) {
      vtn_handle_matrix_alu(b, opcode, val, vtn_src[0], vtn_src[1]);
      b->nb.exact = false;
      return;
   }

   val->ssa = vtn_create_ssa_value(b, type);
   nir_ssa_def *src[4] = { NULL, };
   for (unsigned i = 0; i < num_inputs; i++) {
      assert(glsl_type_is_vector_or_scalar(vtn_src[i]->type));
      src[i] = vtn_src[i]->def;
   }

   switch (opcode) {
   case SpvOpAny:
      if (src[0]->num_components == 1) {
         val->ssa->def = nir_imov(&b->nb, src[0]);
      } else {
         nir_op op;
         switch (src[0]->num_components) {
         case 2:  op = nir_op_bany_inequal2; break;
         case 3:  op = nir_op_bany_inequal3; break;
         case 4:  op = nir_op_bany_inequal4; break;
         default: unreachable("invalid number of components");
         }
         val->ssa->def = nir_build_alu(&b->nb, op, src[0],
                                       nir_imm_int(&b->nb, NIR_FALSE),
                                       NULL, NULL);
      }
      break;

   case SpvOpAll:
      if (src[0]->num_components == 1) {
         val->ssa->def = nir_imov(&b->nb, src[0]);
      } else {
         nir_op op;
         switch (src[0]->num_components) {
         case 2:  op = nir_op_ball_iequal2;  break;
         case 3:  op = nir_op_ball_iequal3;  break;
         case 4:  op = nir_op_ball_iequal4;  break;
         default: unreachable("invalid number of components");
         }
         val->ssa->def = nir_build_alu(&b->nb, op, src[0],
                                       nir_imm_int(&b->nb, NIR_TRUE),
                                       NULL, NULL);
      }
      break;

   case SpvOpOuterProduct: {
      for (unsigned i = 0; i < src[1]->num_components; i++) {
         val->ssa->elems[i]->def =
            nir_fmul(&b->nb, src[0], nir_channel(&b->nb, src[1], i));
      }
      break;
   }

   case SpvOpDot:
      val->ssa->def = nir_fdot(&b->nb, src[0], src[1]);
      break;

   case SpvOpIAddCarry:
      assert(glsl_type_is_struct(val->ssa->type));
      val->ssa->elems[0]->def = nir_iadd(&b->nb, src[0], src[1]);
      val->ssa->elems[1]->def = nir_uadd_carry(&b->nb, src[0], src[1]);
      break;

   case SpvOpISubBorrow:
      assert(glsl_type_is_struct(val->ssa->type));
      val->ssa->elems[0]->def = nir_isub(&b->nb, src[0], src[1]);
      val->ssa->elems[1]->def = nir_usub_borrow(&b->nb, src[0], src[1]);
      break;

   case SpvOpUMulExtended:
      assert(glsl_type_is_struct(val->ssa->type));
      val->ssa->elems[0]->def = nir_imul(&b->nb, src[0], src[1]);
      val->ssa->elems[1]->def = nir_umul_high(&b->nb, src[0], src[1]);
      break;

   case SpvOpSMulExtended:
      assert(glsl_type_is_struct(val->ssa->type));
      val->ssa->elems[0]->def = nir_imul(&b->nb, src[0], src[1]);
      val->ssa->elems[1]->def = nir_imul_high(&b->nb, src[0], src[1]);
      break;

   case SpvOpFwidth:
      val->ssa->def = nir_fadd(&b->nb,
                               nir_fabs(&b->nb, nir_fddx(&b->nb, src[0])),
                               nir_fabs(&b->nb, nir_fddy(&b->nb, src[0])));
      break;
   case SpvOpFwidthFine:
      val->ssa->def = nir_fadd(&b->nb,
                               nir_fabs(&b->nb, nir_fddx_fine(&b->nb, src[0])),
                               nir_fabs(&b->nb, nir_fddy_fine(&b->nb, src[0])));
      break;
   case SpvOpFwidthCoarse:
      val->ssa->def = nir_fadd(&b->nb,
                               nir_fabs(&b->nb, nir_fddx_coarse(&b->nb, src[0])),
                               nir_fabs(&b->nb, nir_fddy_coarse(&b->nb, src[0])));
      break;

   case SpvOpVectorTimesScalar:
      /* The builder will take care of splatting for us. */
      val->ssa->def = nir_fmul(&b->nb, src[0], src[1]);
      break;

   case SpvOpIsNan:
      val->ssa->def = nir_fne(&b->nb, src[0], src[0]);
      break;

   case SpvOpIsInf:
      val->ssa->def = nir_feq(&b->nb, nir_fabs(&b->nb, src[0]),
                                      nir_imm_float(&b->nb, INFINITY));
      break;

   case SpvOpFUnordEqual:
   case SpvOpFUnordNotEqual:
   case SpvOpFUnordLessThan:
   case SpvOpFUnordGreaterThan:
   case SpvOpFUnordLessThanEqual:
   case SpvOpFUnordGreaterThanEqual: {
      bool swap;
      nir_alu_type src_alu_type = nir_get_nir_type_for_glsl_type(vtn_src[0]->type);
      nir_alu_type dst_alu_type = nir_get_nir_type_for_glsl_type(type);
      nir_op op = vtn_nir_alu_op_for_spirv_opcode(opcode, &swap, src_alu_type, dst_alu_type);

      if (swap) {
         nir_ssa_def *tmp = src[0];
         src[0] = src[1];
         src[1] = tmp;
      }

      val->ssa->def =
         nir_ior(&b->nb,
                 nir_build_alu(&b->nb, op, src[0], src[1], NULL, NULL),
                 nir_ior(&b->nb,
                         nir_fne(&b->nb, src[0], src[0]),
                         nir_fne(&b->nb, src[1], src[1])));
      break;
   }

   case SpvOpFOrdEqual:
   case SpvOpFOrdNotEqual:
   case SpvOpFOrdLessThan:
   case SpvOpFOrdGreaterThan:
   case SpvOpFOrdLessThanEqual:
   case SpvOpFOrdGreaterThanEqual: {
      bool swap;
      nir_alu_type src_alu_type = nir_get_nir_type_for_glsl_type(vtn_src[0]->type);
      nir_alu_type dst_alu_type = nir_get_nir_type_for_glsl_type(type);
      nir_op op = vtn_nir_alu_op_for_spirv_opcode(opcode, &swap, src_alu_type, dst_alu_type);

      if (swap) {
         nir_ssa_def *tmp = src[0];
         src[0] = src[1];
         src[1] = tmp;
      }

      val->ssa->def =
         nir_iand(&b->nb,
                  nir_build_alu(&b->nb, op, src[0], src[1], NULL, NULL),
                  nir_iand(&b->nb,
                          nir_feq(&b->nb, src[0], src[0]),
                          nir_feq(&b->nb, src[1], src[1])));
      break;
   }

   default: {
      bool swap;
      nir_alu_type src_alu_type = nir_get_nir_type_for_glsl_type(vtn_src[0]->type);
      nir_alu_type dst_alu_type = nir_get_nir_type_for_glsl_type(type);
      nir_op op = vtn_nir_alu_op_for_spirv_opcode(opcode, &swap, src_alu_type, dst_alu_type);

      if (swap) {
         nir_ssa_def *tmp = src[0];
         src[0] = src[1];
         src[1] = tmp;
      }

      val->ssa->def = nir_build_alu(&b->nb, op, src[0], src[1], src[2], src[3]);
      break;
   } /* default */
   }

   b->nb.exact = false;
}
コード例 #8
0
ファイル: nir_lower_tex.c プロジェクト: BNieuwenhuizen/mesa
   return &plane_tex->dest.ssa;
}

static void
convert_yuv_to_rgb(nir_builder *b, nir_tex_instr *tex,
                   nir_ssa_def *y, nir_ssa_def *u, nir_ssa_def *v)
{
   nir_const_value m[3] = {
      { .f32 = { 1.0f,  0.0f,         1.59602678f, 0.0f } },
      { .f32 = { 1.0f, -0.39176229f, -0.81296764f, 0.0f } },
      { .f32 = { 1.0f,  2.01723214f,  0.0f,        0.0f } }
   };

   nir_ssa_def *yuv =
      nir_vec4(b,
               nir_fmul(b, nir_imm_float(b, 1.16438356f),
                        nir_fadd(b, y, nir_imm_float(b, -0.0625f))),
               nir_channel(b, nir_fadd(b, u, nir_imm_float(b, -0.5f)), 0),
               nir_channel(b, nir_fadd(b, v, nir_imm_float(b, -0.5f)), 0),
               nir_imm_float(b, 0.0));

   nir_ssa_def *red = nir_fdot4(b, yuv, nir_build_imm(b, 4, 32, m[0]));
   nir_ssa_def *green = nir_fdot4(b, yuv, nir_build_imm(b, 4, 32, m[1]));
   nir_ssa_def *blue = nir_fdot4(b, yuv, nir_build_imm(b, 4, 32, m[2]));

   nir_ssa_def *result = nir_vec4(b, red, green, blue, nir_imm_float(b, 1.0f));

   nir_ssa_def_rewrite_uses(&tex->dest.ssa, nir_src_for_ssa(result));
}

static void
コード例 #9
0
ファイル: vtn_glsl450.c プロジェクト: etnaviv/mesa
static void
handle_glsl450_alu(struct vtn_builder *b, enum GLSLstd450 entrypoint,
                   const uint32_t *w, unsigned count)
{
   struct nir_builder *nb = &b->nb;
   const struct glsl_type *dest_type =
      vtn_value(b, w[1], vtn_value_type_type)->type->type;

   struct vtn_value *val = vtn_push_value(b, w[2], vtn_value_type_ssa);
   val->ssa = vtn_create_ssa_value(b, dest_type);

   /* Collect the various SSA sources */
   unsigned num_inputs = count - 5;
   nir_ssa_def *src[3] = { NULL, };
   for (unsigned i = 0; i < num_inputs; i++)
      src[i] = vtn_ssa_value(b, w[i + 5])->def;

   switch (entrypoint) {
   case GLSLstd450Radians:
      val->ssa->def = nir_fmul(nb, src[0], nir_imm_float(nb, 0.01745329251));
      return;
   case GLSLstd450Degrees:
      val->ssa->def = nir_fmul(nb, src[0], nir_imm_float(nb, 57.2957795131));
      return;
   case GLSLstd450Tan:
      val->ssa->def = nir_fdiv(nb, nir_fsin(nb, src[0]),
                               nir_fcos(nb, src[0]));
      return;

   case GLSLstd450Modf: {
      nir_ssa_def *sign = nir_fsign(nb, src[0]);
      nir_ssa_def *abs = nir_fabs(nb, src[0]);
      val->ssa->def = nir_fmul(nb, sign, nir_ffract(nb, abs));
      nir_store_deref_var(nb, vtn_nir_deref(b, w[6]),
                          nir_fmul(nb, sign, nir_ffloor(nb, abs)), 0xf);
      return;
   }

   case GLSLstd450ModfStruct: {
      nir_ssa_def *sign = nir_fsign(nb, src[0]);
      nir_ssa_def *abs = nir_fabs(nb, src[0]);
      assert(glsl_type_is_struct(val->ssa->type));
      val->ssa->elems[0]->def = nir_fmul(nb, sign, nir_ffract(nb, abs));
      val->ssa->elems[1]->def = nir_fmul(nb, sign, nir_ffloor(nb, abs));
      return;
   }

   case GLSLstd450Step:
      val->ssa->def = nir_sge(nb, src[1], src[0]);
      return;

   case GLSLstd450Length:
      val->ssa->def = build_length(nb, src[0]);
      return;
   case GLSLstd450Distance:
      val->ssa->def = build_length(nb, nir_fsub(nb, src[0], src[1]));
      return;
   case GLSLstd450Normalize:
      val->ssa->def = nir_fdiv(nb, src[0], build_length(nb, src[0]));
      return;

   case GLSLstd450Exp:
      val->ssa->def = build_exp(nb, src[0]);
      return;

   case GLSLstd450Log:
      val->ssa->def = build_log(nb, src[0]);
      return;

   case GLSLstd450FClamp:
      val->ssa->def = build_fclamp(nb, src[0], src[1], src[2]);
      return;
   case GLSLstd450UClamp:
      val->ssa->def = nir_umin(nb, nir_umax(nb, src[0], src[1]), src[2]);
      return;
   case GLSLstd450SClamp:
      val->ssa->def = nir_imin(nb, nir_imax(nb, src[0], src[1]), src[2]);
      return;

   case GLSLstd450Cross: {
      unsigned yzx[4] = { 1, 2, 0, 0 };
      unsigned zxy[4] = { 2, 0, 1, 0 };
      val->ssa->def =
         nir_fsub(nb, nir_fmul(nb, nir_swizzle(nb, src[0], yzx, 3, true),
                                   nir_swizzle(nb, src[1], zxy, 3, true)),
                      nir_fmul(nb, nir_swizzle(nb, src[0], zxy, 3, true),
                                   nir_swizzle(nb, src[1], yzx, 3, true)));
      return;
   }

   case GLSLstd450SmoothStep: {
      /* t = clamp((x - edge0) / (edge1 - edge0), 0, 1) */
      nir_ssa_def *t =
         build_fclamp(nb, nir_fdiv(nb, nir_fsub(nb, src[2], src[0]),
                                       nir_fsub(nb, src[1], src[0])),
                          nir_imm_float(nb, 0.0), nir_imm_float(nb, 1.0));
      /* result = t * t * (3 - 2 * t) */
      val->ssa->def =
         nir_fmul(nb, t, nir_fmul(nb, t,
            nir_fsub(nb, nir_imm_float(nb, 3.0),
                         nir_fmul(nb, nir_imm_float(nb, 2.0), t))));
      return;
   }

   case GLSLstd450FaceForward:
      val->ssa->def =
         nir_bcsel(nb, nir_flt(nb, nir_fdot(nb, src[2], src[1]),
                                   nir_imm_float(nb, 0.0)),
                       src[0], nir_fneg(nb, src[0]));
      return;

   case GLSLstd450Reflect:
      /* I - 2 * dot(N, I) * N */
      val->ssa->def =
         nir_fsub(nb, src[0], nir_fmul(nb, nir_imm_float(nb, 2.0),
                              nir_fmul(nb, nir_fdot(nb, src[0], src[1]),
                                           src[1])));
      return;

   case GLSLstd450Refract: {
      nir_ssa_def *I = src[0];
      nir_ssa_def *N = src[1];
      nir_ssa_def *eta = src[2];
      nir_ssa_def *n_dot_i = nir_fdot(nb, N, I);
      nir_ssa_def *one = nir_imm_float(nb, 1.0);
      nir_ssa_def *zero = nir_imm_float(nb, 0.0);
      /* k = 1.0 - eta * eta * (1.0 - dot(N, I) * dot(N, I)) */
      nir_ssa_def *k =
         nir_fsub(nb, one, nir_fmul(nb, eta, nir_fmul(nb, eta,
                      nir_fsub(nb, one, nir_fmul(nb, n_dot_i, n_dot_i)))));
      nir_ssa_def *result =
         nir_fsub(nb, nir_fmul(nb, eta, I),
                      nir_fmul(nb, nir_fadd(nb, nir_fmul(nb, eta, n_dot_i),
                                                nir_fsqrt(nb, k)), N));
      /* XXX: bcsel, or if statement? */
      val->ssa->def = nir_bcsel(nb, nir_flt(nb, k, zero), zero, result);
      return;
   }

   case GLSLstd450Sinh:
      /* 0.5 * (e^x - e^(-x)) */
      val->ssa->def =
         nir_fmul(nb, nir_imm_float(nb, 0.5f),
                      nir_fsub(nb, build_exp(nb, src[0]),
                                   build_exp(nb, nir_fneg(nb, src[0]))));
      return;

   case GLSLstd450Cosh:
      /* 0.5 * (e^x + e^(-x)) */
      val->ssa->def =
         nir_fmul(nb, nir_imm_float(nb, 0.5f),
                      nir_fadd(nb, build_exp(nb, src[0]),
                                   build_exp(nb, nir_fneg(nb, src[0]))));
      return;

   case GLSLstd450Tanh:
      /* (0.5 * (e^x - e^(-x))) / (0.5 * (e^x + e^(-x))) */
      val->ssa->def =
         nir_fdiv(nb, nir_fmul(nb, nir_imm_float(nb, 0.5f),
                                   nir_fsub(nb, build_exp(nb, src[0]),
                                                build_exp(nb, nir_fneg(nb, src[0])))),
                      nir_fmul(nb, nir_imm_float(nb, 0.5f),
                                   nir_fadd(nb, build_exp(nb, src[0]),
                                                build_exp(nb, nir_fneg(nb, src[0])))));
      return;

   case GLSLstd450Asinh:
      val->ssa->def = nir_fmul(nb, nir_fsign(nb, src[0]),
         build_log(nb, nir_fadd(nb, nir_fabs(nb, src[0]),
                       nir_fsqrt(nb, nir_fadd(nb, nir_fmul(nb, src[0], src[0]),
                                                  nir_imm_float(nb, 1.0f))))));
      return;
   case GLSLstd450Acosh:
      val->ssa->def = build_log(nb, nir_fadd(nb, src[0],
         nir_fsqrt(nb, nir_fsub(nb, nir_fmul(nb, src[0], src[0]),
                                    nir_imm_float(nb, 1.0f)))));
      return;
   case GLSLstd450Atanh: {
      nir_ssa_def *one = nir_imm_float(nb, 1.0);
      val->ssa->def = nir_fmul(nb, nir_imm_float(nb, 0.5f),
         build_log(nb, nir_fdiv(nb, nir_fadd(nb, one, src[0]),
                                    nir_fsub(nb, one, src[0]))));
      return;
   }

   case GLSLstd450Asin:
      val->ssa->def = build_asin(nb, src[0], 0.086566724, -0.03102955);
      return;

   case GLSLstd450Acos:
      val->ssa->def = nir_fsub(nb, nir_imm_float(nb, M_PI_2f),
                               build_asin(nb, src[0], 0.08132463, -0.02363318));
      return;

   case GLSLstd450Atan:
      val->ssa->def = build_atan(nb, src[0]);
      return;

   case GLSLstd450Atan2:
      val->ssa->def = build_atan2(nb, src[0], src[1]);
      return;

   case GLSLstd450Frexp: {
      nir_ssa_def *exponent;
      val->ssa->def = build_frexp(nb, src[0], &exponent);
      nir_store_deref_var(nb, vtn_nir_deref(b, w[6]), exponent, 0xf);
      return;
   }

   case GLSLstd450FrexpStruct: {
      assert(glsl_type_is_struct(val->ssa->type));
      val->ssa->elems[0]->def = build_frexp(nb, src[0],
                                            &val->ssa->elems[1]->def);
      return;
   }

   default:
      val->ssa->def =
         nir_build_alu(&b->nb, vtn_nir_alu_op_for_spirv_glsl_opcode(entrypoint),
                       src[0], src[1], src[2], NULL);
      return;
   }
}
コード例 #10
0
ファイル: vtn_glsl450.c プロジェクト: etnaviv/mesa
/**
 * Return ln(x) - the natural logarithm of x.
 */
static nir_ssa_def *
build_log(nir_builder *b, nir_ssa_def *x)
{
   return nir_fmul(b, nir_flog2(b, x), nir_imm_float(b, 1.0 / M_LOG2E));
}
コード例 #11
0
ファイル: vtn_glsl450.c プロジェクト: etnaviv/mesa
/**
 * Return e^x.
 */
static nir_ssa_def *
build_exp(nir_builder *b, nir_ssa_def *x)
{
   return nir_fexp2(b, nir_fmul(b, x, nir_imm_float(b, M_LOG2E)));
}
コード例 #12
0
ファイル: nir_lower_tex.c プロジェクト: jon-turney/mesa
}

static void
convert_yuv_to_rgb(nir_builder *b, nir_tex_instr *tex,
                   nir_ssa_def *y, nir_ssa_def *u, nir_ssa_def *v,
                   nir_ssa_def *a)
{
   nir_const_value m[3] = {
      { .f32 = { 1.0f,  0.0f,         1.59602678f, 0.0f } },
      { .f32 = { 1.0f, -0.39176229f, -0.81296764f, 0.0f } },
      { .f32 = { 1.0f,  2.01723214f,  0.0f,        0.0f } }
   };

   nir_ssa_def *yuv =
      nir_vec4(b,
               nir_fmul(b, nir_imm_float(b, 1.16438356f),
                        nir_fadd(b, y, nir_imm_float(b, -16.0f / 255.0f))),
               nir_channel(b, nir_fadd(b, u, nir_imm_float(b, -128.0f / 255.0f)), 0),
               nir_channel(b, nir_fadd(b, v, nir_imm_float(b, -128.0f / 255.0f)), 0),
               nir_imm_float(b, 0.0));

   nir_ssa_def *red = nir_fdot4(b, yuv, nir_build_imm(b, 4, 32, m[0]));
   nir_ssa_def *green = nir_fdot4(b, yuv, nir_build_imm(b, 4, 32, m[1]));
   nir_ssa_def *blue = nir_fdot4(b, yuv, nir_build_imm(b, 4, 32, m[2]));

   nir_ssa_def *result = nir_vec4(b, red, green, blue, a);

   nir_ssa_def_rewrite_uses(&tex->dest.ssa, nir_src_for_ssa(result));
}

static void