2 * Copyright © 2010 Intel Corporation
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
11 * The above copyright notice and this permission notice (including the next
12 * paragraph) shall be included in all copies or substantial portions of the
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
18 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
20 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
21 * DEALINGS IN THE SOFTWARE.
24 #include "main/core.h" /* for MAX2 */
26 #include "ir_visitor.h"
27 #include "glsl_types.h"
29 ir_rvalue::ir_rvalue()
31 this->type
= glsl_type::error_type
;
34 bool ir_rvalue::is_zero() const
39 bool ir_rvalue::is_one() const
44 bool ir_rvalue::is_negative_one() const
49 bool ir_rvalue::is_basis() const
55 * Modify the swizzle make to move one component to another
57 * \param m IR swizzle to be modified
58 * \param from Component in the RHS that is to be swizzled
59 * \param to Desired swizzle location of \c from
62 update_rhs_swizzle(ir_swizzle_mask
&m
, unsigned from
, unsigned to
)
65 case 0: m
.x
= from
; break;
66 case 1: m
.y
= from
; break;
67 case 2: m
.z
= from
; break;
68 case 3: m
.w
= from
; break;
69 default: assert(!"Should not get here.");
72 m
.num_components
= MAX2(m
.num_components
, (to
+ 1));
76 ir_assignment::set_lhs(ir_rvalue
*lhs
)
79 bool swizzled
= false;
82 ir_swizzle
*swiz
= lhs
->as_swizzle();
87 unsigned write_mask
= 0;
88 ir_swizzle_mask rhs_swiz
= { 0, 0, 0, 0, 0, 0 };
90 for (unsigned i
= 0; i
< swiz
->mask
.num_components
; i
++) {
94 case 0: c
= swiz
->mask
.x
; break;
95 case 1: c
= swiz
->mask
.y
; break;
96 case 2: c
= swiz
->mask
.z
; break;
97 case 3: c
= swiz
->mask
.w
; break;
98 default: assert(!"Should not get here.");
101 write_mask
|= (((this->write_mask
>> i
) & 1) << c
);
102 update_rhs_swizzle(rhs_swiz
, i
, c
);
105 this->write_mask
= write_mask
;
108 this->rhs
= new(mem_ctx
) ir_swizzle(this->rhs
, rhs_swiz
);
113 /* Now, RHS channels line up with the LHS writemask. Collapse it
114 * to just the channels that will be written.
116 ir_swizzle_mask rhs_swiz
= { 0, 0, 0, 0, 0, 0 };
118 for (int i
= 0; i
< 4; i
++) {
119 if (write_mask
& (1 << i
))
120 update_rhs_swizzle(rhs_swiz
, i
, rhs_chan
++);
122 this->rhs
= new(mem_ctx
) ir_swizzle(this->rhs
, rhs_swiz
);
125 assert((lhs
== NULL
) || lhs
->as_dereference());
127 this->lhs
= (ir_dereference
*) lhs
;
131 ir_assignment::whole_variable_written()
133 ir_variable
*v
= this->lhs
->whole_variable_referenced();
138 if (v
->type
->is_scalar())
141 if (v
->type
->is_vector()) {
142 const unsigned mask
= (1U << v
->type
->vector_elements
) - 1;
144 if (mask
!= this->write_mask
)
148 /* Either all the vector components are assigned or the variable is some
149 * composite type (and the whole thing is assigned.
154 ir_assignment::ir_assignment(ir_dereference
*lhs
, ir_rvalue
*rhs
,
155 ir_rvalue
*condition
, unsigned write_mask
)
157 this->ir_type
= ir_type_assignment
;
158 this->condition
= condition
;
161 this->write_mask
= write_mask
;
163 if (lhs
->type
->is_scalar() || lhs
->type
->is_vector()) {
164 int lhs_components
= 0;
165 for (int i
= 0; i
< 4; i
++) {
166 if (write_mask
& (1 << i
))
170 assert(lhs_components
== this->rhs
->type
->vector_elements
);
174 ir_assignment::ir_assignment(ir_rvalue
*lhs
, ir_rvalue
*rhs
,
175 ir_rvalue
*condition
)
177 this->ir_type
= ir_type_assignment
;
178 this->condition
= condition
;
181 /* If the RHS is a vector type, assume that all components of the vector
182 * type are being written to the LHS. The write mask comes from the RHS
183 * because we can have a case where the LHS is a vec4 and the RHS is a
184 * vec3. In that case, the assignment is:
186 * (assign (...) (xyz) (var_ref lhs) (var_ref rhs))
188 if (rhs
->type
->is_vector())
189 this->write_mask
= (1U << rhs
->type
->vector_elements
) - 1;
190 else if (rhs
->type
->is_scalar())
191 this->write_mask
= 1;
193 this->write_mask
= 0;
198 ir_expression::ir_expression(int op
, const struct glsl_type
*type
,
199 ir_rvalue
*op0
, ir_rvalue
*op1
,
200 ir_rvalue
*op2
, ir_rvalue
*op3
)
202 this->ir_type
= ir_type_expression
;
204 this->operation
= ir_expression_operation(op
);
205 this->operands
[0] = op0
;
206 this->operands
[1] = op1
;
207 this->operands
[2] = op2
;
208 this->operands
[3] = op3
;
210 int num_operands
= get_num_operands(this->operation
);
211 for (int i
= num_operands
; i
< 4; i
++) {
212 assert(this->operands
[i
] == NULL
);
217 ir_expression::ir_expression(int op
, ir_rvalue
*op0
)
219 this->ir_type
= ir_type_expression
;
221 this->operation
= ir_expression_operation(op
);
222 this->operands
[0] = op0
;
223 this->operands
[1] = NULL
;
224 this->operands
[2] = NULL
;
225 this->operands
[3] = NULL
;
227 assert(op
<= ir_last_unop
);
229 switch (this->operation
) {
230 case ir_unop_bit_not
:
231 case ir_unop_logic_not
:
246 case ir_unop_round_even
:
249 case ir_unop_sin_reduced
:
250 case ir_unop_cos_reduced
:
253 this->type
= op0
->type
;
259 case ir_unop_bitcast_f2i
:
260 this->type
= glsl_type::get_instance(GLSL_TYPE_INT
,
261 op0
->type
->vector_elements
, 1);
267 case ir_unop_bitcast_i2f
:
268 case ir_unop_bitcast_u2f
:
269 this->type
= glsl_type::get_instance(GLSL_TYPE_FLOAT
,
270 op0
->type
->vector_elements
, 1);
275 this->type
= glsl_type::get_instance(GLSL_TYPE_BOOL
,
276 op0
->type
->vector_elements
, 1);
281 case ir_unop_bitcast_f2u
:
282 this->type
= glsl_type::get_instance(GLSL_TYPE_UINT
,
283 op0
->type
->vector_elements
, 1);
287 case ir_unop_unpack_half_2x16_split_x
:
288 case ir_unop_unpack_half_2x16_split_y
:
289 this->type
= glsl_type::float_type
;
293 this->type
= glsl_type::bool_type
;
296 case ir_unop_pack_snorm_2x16
:
297 case ir_unop_pack_snorm_4x8
:
298 case ir_unop_pack_unorm_2x16
:
299 case ir_unop_pack_unorm_4x8
:
300 case ir_unop_pack_half_2x16
:
301 this->type
= glsl_type::uint_type
;
304 case ir_unop_unpack_snorm_2x16
:
305 case ir_unop_unpack_unorm_2x16
:
306 case ir_unop_unpack_half_2x16
:
307 this->type
= glsl_type::vec2_type
;
310 case ir_unop_unpack_snorm_4x8
:
311 case ir_unop_unpack_unorm_4x8
:
312 this->type
= glsl_type::vec4_type
;
316 assert(!"not reached: missing automatic type setup for ir_expression");
317 this->type
= op0
->type
;
322 ir_expression::ir_expression(int op
, ir_rvalue
*op0
, ir_rvalue
*op1
)
324 this->ir_type
= ir_type_expression
;
326 this->operation
= ir_expression_operation(op
);
327 this->operands
[0] = op0
;
328 this->operands
[1] = op1
;
329 this->operands
[2] = NULL
;
330 this->operands
[3] = NULL
;
332 assert(op
> ir_last_unop
);
334 switch (this->operation
) {
335 case ir_binop_all_equal
:
336 case ir_binop_any_nequal
:
337 this->type
= glsl_type::bool_type
;
348 if (op0
->type
->is_scalar()) {
349 this->type
= op1
->type
;
350 } else if (op1
->type
->is_scalar()) {
351 this->type
= op0
->type
;
353 /* FINISHME: matrix types */
354 assert(!op0
->type
->is_matrix() && !op1
->type
->is_matrix());
355 assert(op0
->type
== op1
->type
);
356 this->type
= op0
->type
;
360 case ir_binop_logic_and
:
361 case ir_binop_logic_xor
:
362 case ir_binop_logic_or
:
363 case ir_binop_bit_and
:
364 case ir_binop_bit_xor
:
365 case ir_binop_bit_or
:
366 assert(!op0
->type
->is_matrix());
367 assert(!op1
->type
->is_matrix());
368 if (op0
->type
->is_scalar()) {
369 this->type
= op1
->type
;
370 } else if (op1
->type
->is_scalar()) {
371 this->type
= op0
->type
;
373 assert(op0
->type
->vector_elements
== op1
->type
->vector_elements
);
374 this->type
= op0
->type
;
379 case ir_binop_nequal
:
380 case ir_binop_lequal
:
381 case ir_binop_gequal
:
383 case ir_binop_greater
:
384 assert(op0
->type
== op1
->type
);
385 this->type
= glsl_type::get_instance(GLSL_TYPE_BOOL
,
386 op0
->type
->vector_elements
, 1);
390 this->type
= glsl_type::float_type
;
393 case ir_binop_pack_half_2x16_split
:
394 this->type
= glsl_type::uint_type
;
397 case ir_binop_lshift
:
398 case ir_binop_rshift
:
399 this->type
= op0
->type
;
402 case ir_binop_vector_extract
:
403 this->type
= op0
->type
->get_scalar_type();
407 assert(!"not reached: missing automatic type setup for ir_expression");
408 this->type
= glsl_type::float_type
;
413 ir_expression::get_num_operands(ir_expression_operation op
)
415 assert(op
<= ir_last_opcode
);
417 if (op
<= ir_last_unop
)
420 if (op
<= ir_last_binop
)
423 if (op
<= ir_last_triop
)
426 if (op
<= ir_last_quadop
)
433 static const char *const operator_strs
[] = {
482 "unpackHalf2x16_split_x",
483 "unpackHalf2x16_split_y",
514 "packHalf2x16_split",
527 const char *ir_expression::operator_string(ir_expression_operation op
)
529 assert((unsigned int) op
< Elements(operator_strs
));
530 assert(Elements(operator_strs
) == (ir_quadop_vector
+ 1));
531 return operator_strs
[op
];
534 const char *ir_expression::operator_string()
536 return operator_string(this->operation
);
540 depth_layout_string(ir_depth_layout layout
)
543 case ir_depth_layout_none
: return "";
544 case ir_depth_layout_any
: return "depth_any";
545 case ir_depth_layout_greater
: return "depth_greater";
546 case ir_depth_layout_less
: return "depth_less";
547 case ir_depth_layout_unchanged
: return "depth_unchanged";
555 ir_expression_operation
556 ir_expression::get_operator(const char *str
)
558 const int operator_count
= sizeof(operator_strs
) / sizeof(operator_strs
[0]);
559 for (int op
= 0; op
< operator_count
; op
++) {
560 if (strcmp(str
, operator_strs
[op
]) == 0)
561 return (ir_expression_operation
) op
;
563 return (ir_expression_operation
) -1;
566 ir_constant::ir_constant()
568 this->ir_type
= ir_type_constant
;
571 ir_constant::ir_constant(const struct glsl_type
*type
,
572 const ir_constant_data
*data
)
574 assert((type
->base_type
>= GLSL_TYPE_UINT
)
575 && (type
->base_type
<= GLSL_TYPE_BOOL
));
577 this->ir_type
= ir_type_constant
;
579 memcpy(& this->value
, data
, sizeof(this->value
));
582 ir_constant::ir_constant(float f
)
584 this->ir_type
= ir_type_constant
;
585 this->type
= glsl_type::float_type
;
586 this->value
.f
[0] = f
;
587 for (int i
= 1; i
< 16; i
++) {
588 this->value
.f
[i
] = 0;
592 ir_constant::ir_constant(unsigned int u
)
594 this->ir_type
= ir_type_constant
;
595 this->type
= glsl_type::uint_type
;
596 this->value
.u
[0] = u
;
597 for (int i
= 1; i
< 16; i
++) {
598 this->value
.u
[i
] = 0;
602 ir_constant::ir_constant(int i
)
604 this->ir_type
= ir_type_constant
;
605 this->type
= glsl_type::int_type
;
606 this->value
.i
[0] = i
;
607 for (int i
= 1; i
< 16; i
++) {
608 this->value
.i
[i
] = 0;
612 ir_constant::ir_constant(bool b
)
614 this->ir_type
= ir_type_constant
;
615 this->type
= glsl_type::bool_type
;
616 this->value
.b
[0] = b
;
617 for (int i
= 1; i
< 16; i
++) {
618 this->value
.b
[i
] = false;
622 ir_constant::ir_constant(const ir_constant
*c
, unsigned i
)
624 this->ir_type
= ir_type_constant
;
625 this->type
= c
->type
->get_base_type();
627 switch (this->type
->base_type
) {
628 case GLSL_TYPE_UINT
: this->value
.u
[0] = c
->value
.u
[i
]; break;
629 case GLSL_TYPE_INT
: this->value
.i
[0] = c
->value
.i
[i
]; break;
630 case GLSL_TYPE_FLOAT
: this->value
.f
[0] = c
->value
.f
[i
]; break;
631 case GLSL_TYPE_BOOL
: this->value
.b
[0] = c
->value
.b
[i
]; break;
632 default: assert(!"Should not get here."); break;
636 ir_constant::ir_constant(const struct glsl_type
*type
, exec_list
*value_list
)
638 this->ir_type
= ir_type_constant
;
641 assert(type
->is_scalar() || type
->is_vector() || type
->is_matrix()
642 || type
->is_record() || type
->is_array());
644 if (type
->is_array()) {
645 this->array_elements
= ralloc_array(this, ir_constant
*, type
->length
);
647 foreach_list(node
, value_list
) {
648 ir_constant
*value
= (ir_constant
*) node
;
649 assert(value
->as_constant() != NULL
);
651 this->array_elements
[i
++] = value
;
656 /* If the constant is a record, the types of each of the entries in
657 * value_list must be a 1-for-1 match with the structure components. Each
658 * entry must also be a constant. Just move the nodes from the value_list
659 * to the list in the ir_constant.
661 /* FINISHME: Should there be some type checking and / or assertions here? */
662 /* FINISHME: Should the new constant take ownership of the nodes from
663 * FINISHME: value_list, or should it make copies?
665 if (type
->is_record()) {
666 value_list
->move_nodes_to(& this->components
);
670 for (unsigned i
= 0; i
< 16; i
++) {
671 this->value
.u
[i
] = 0;
674 ir_constant
*value
= (ir_constant
*) (value_list
->head
);
676 /* Constructors with exactly one scalar argument are special for vectors
677 * and matrices. For vectors, the scalar value is replicated to fill all
678 * the components. For matrices, the scalar fills the components of the
679 * diagonal while the rest is filled with 0.
681 if (value
->type
->is_scalar() && value
->next
->is_tail_sentinel()) {
682 if (type
->is_matrix()) {
683 /* Matrix - fill diagonal (rest is already set to 0) */
684 assert(type
->base_type
== GLSL_TYPE_FLOAT
);
685 for (unsigned i
= 0; i
< type
->matrix_columns
; i
++)
686 this->value
.f
[i
* type
->vector_elements
+ i
] = value
->value
.f
[0];
688 /* Vector or scalar - fill all components */
689 switch (type
->base_type
) {
692 for (unsigned i
= 0; i
< type
->components(); i
++)
693 this->value
.u
[i
] = value
->value
.u
[0];
695 case GLSL_TYPE_FLOAT
:
696 for (unsigned i
= 0; i
< type
->components(); i
++)
697 this->value
.f
[i
] = value
->value
.f
[0];
700 for (unsigned i
= 0; i
< type
->components(); i
++)
701 this->value
.b
[i
] = value
->value
.b
[0];
704 assert(!"Should not get here.");
711 if (type
->is_matrix() && value
->type
->is_matrix()) {
712 assert(value
->next
->is_tail_sentinel());
714 /* From section 5.4.2 of the GLSL 1.20 spec:
715 * "If a matrix is constructed from a matrix, then each component
716 * (column i, row j) in the result that has a corresponding component
717 * (column i, row j) in the argument will be initialized from there."
719 unsigned cols
= MIN2(type
->matrix_columns
, value
->type
->matrix_columns
);
720 unsigned rows
= MIN2(type
->vector_elements
, value
->type
->vector_elements
);
721 for (unsigned i
= 0; i
< cols
; i
++) {
722 for (unsigned j
= 0; j
< rows
; j
++) {
723 const unsigned src
= i
* value
->type
->vector_elements
+ j
;
724 const unsigned dst
= i
* type
->vector_elements
+ j
;
725 this->value
.f
[dst
] = value
->value
.f
[src
];
729 /* "All other components will be initialized to the identity matrix." */
730 for (unsigned i
= cols
; i
< type
->matrix_columns
; i
++)
731 this->value
.f
[i
* type
->vector_elements
+ i
] = 1.0;
736 /* Use each component from each entry in the value_list to initialize one
737 * component of the constant being constructed.
739 for (unsigned i
= 0; i
< type
->components(); /* empty */) {
740 assert(value
->as_constant() != NULL
);
741 assert(!value
->is_tail_sentinel());
743 for (unsigned j
= 0; j
< value
->type
->components(); j
++) {
744 switch (type
->base_type
) {
746 this->value
.u
[i
] = value
->get_uint_component(j
);
749 this->value
.i
[i
] = value
->get_int_component(j
);
751 case GLSL_TYPE_FLOAT
:
752 this->value
.f
[i
] = value
->get_float_component(j
);
755 this->value
.b
[i
] = value
->get_bool_component(j
);
758 /* FINISHME: What to do? Exceptions are not the answer.
764 if (i
>= type
->components())
768 value
= (ir_constant
*) value
->next
;
773 ir_constant::zero(void *mem_ctx
, const glsl_type
*type
)
775 assert(type
->is_scalar() || type
->is_vector() || type
->is_matrix()
776 || type
->is_record() || type
->is_array());
778 ir_constant
*c
= new(mem_ctx
) ir_constant
;
780 memset(&c
->value
, 0, sizeof(c
->value
));
782 if (type
->is_array()) {
783 c
->array_elements
= ralloc_array(c
, ir_constant
*, type
->length
);
785 for (unsigned i
= 0; i
< type
->length
; i
++)
786 c
->array_elements
[i
] = ir_constant::zero(c
, type
->element_type());
789 if (type
->is_record()) {
790 for (unsigned i
= 0; i
< type
->length
; i
++) {
791 ir_constant
*comp
= ir_constant::zero(mem_ctx
, type
->fields
.structure
[i
].type
);
792 c
->components
.push_tail(comp
);
800 ir_constant::get_bool_component(unsigned i
) const
802 switch (this->type
->base_type
) {
803 case GLSL_TYPE_UINT
: return this->value
.u
[i
] != 0;
804 case GLSL_TYPE_INT
: return this->value
.i
[i
] != 0;
805 case GLSL_TYPE_FLOAT
: return ((int)this->value
.f
[i
]) != 0;
806 case GLSL_TYPE_BOOL
: return this->value
.b
[i
];
807 default: assert(!"Should not get here."); break;
810 /* Must return something to make the compiler happy. This is clearly an
817 ir_constant::get_float_component(unsigned i
) const
819 switch (this->type
->base_type
) {
820 case GLSL_TYPE_UINT
: return (float) this->value
.u
[i
];
821 case GLSL_TYPE_INT
: return (float) this->value
.i
[i
];
822 case GLSL_TYPE_FLOAT
: return this->value
.f
[i
];
823 case GLSL_TYPE_BOOL
: return this->value
.b
[i
] ? 1.0f
: 0.0f
;
824 default: assert(!"Should not get here."); break;
827 /* Must return something to make the compiler happy. This is clearly an
834 ir_constant::get_int_component(unsigned i
) const
836 switch (this->type
->base_type
) {
837 case GLSL_TYPE_UINT
: return this->value
.u
[i
];
838 case GLSL_TYPE_INT
: return this->value
.i
[i
];
839 case GLSL_TYPE_FLOAT
: return (int) this->value
.f
[i
];
840 case GLSL_TYPE_BOOL
: return this->value
.b
[i
] ? 1 : 0;
841 default: assert(!"Should not get here."); break;
844 /* Must return something to make the compiler happy. This is clearly an
851 ir_constant::get_uint_component(unsigned i
) const
853 switch (this->type
->base_type
) {
854 case GLSL_TYPE_UINT
: return this->value
.u
[i
];
855 case GLSL_TYPE_INT
: return this->value
.i
[i
];
856 case GLSL_TYPE_FLOAT
: return (unsigned) this->value
.f
[i
];
857 case GLSL_TYPE_BOOL
: return this->value
.b
[i
] ? 1 : 0;
858 default: assert(!"Should not get here."); break;
861 /* Must return something to make the compiler happy. This is clearly an
868 ir_constant::get_array_element(unsigned i
) const
870 assert(this->type
->is_array());
872 /* From page 35 (page 41 of the PDF) of the GLSL 1.20 spec:
874 * "Behavior is undefined if a shader subscripts an array with an index
875 * less than 0 or greater than or equal to the size the array was
878 * Most out-of-bounds accesses are removed before things could get this far.
879 * There are cases where non-constant array index values can get constant
884 else if (i
>= this->type
->length
)
885 i
= this->type
->length
- 1;
887 return array_elements
[i
];
891 ir_constant::get_record_field(const char *name
)
893 int idx
= this->type
->field_index(name
);
898 if (this->components
.is_empty())
901 exec_node
*node
= this->components
.head
;
902 for (int i
= 0; i
< idx
; i
++) {
905 /* If the end of the list is encountered before the element matching the
906 * requested field is found, return NULL.
908 if (node
->is_tail_sentinel())
912 return (ir_constant
*) node
;
916 ir_constant::copy_offset(ir_constant
*src
, int offset
)
918 switch (this->type
->base_type
) {
921 case GLSL_TYPE_FLOAT
:
922 case GLSL_TYPE_BOOL
: {
923 unsigned int size
= src
->type
->components();
924 assert (size
<= this->type
->components() - offset
);
925 for (unsigned int i
=0; i
<size
; i
++) {
926 switch (this->type
->base_type
) {
928 value
.u
[i
+offset
] = src
->get_uint_component(i
);
931 value
.i
[i
+offset
] = src
->get_int_component(i
);
933 case GLSL_TYPE_FLOAT
:
934 value
.f
[i
+offset
] = src
->get_float_component(i
);
937 value
.b
[i
+offset
] = src
->get_bool_component(i
);
939 default: // Shut up the compiler
946 case GLSL_TYPE_STRUCT
: {
947 assert (src
->type
== this->type
);
948 this->components
.make_empty();
949 foreach_list(node
, &src
->components
) {
950 ir_constant
*const orig
= (ir_constant
*) node
;
952 this->components
.push_tail(orig
->clone(this, NULL
));
957 case GLSL_TYPE_ARRAY
: {
958 assert (src
->type
== this->type
);
959 for (unsigned i
= 0; i
< this->type
->length
; i
++) {
960 this->array_elements
[i
] = src
->array_elements
[i
]->clone(this, NULL
);
966 assert(!"Should not get here.");
972 ir_constant::copy_masked_offset(ir_constant
*src
, int offset
, unsigned int mask
)
974 assert (!type
->is_array() && !type
->is_record());
976 if (!type
->is_vector() && !type
->is_matrix()) {
982 for (int i
=0; i
<4; i
++) {
983 if (mask
& (1 << i
)) {
984 switch (this->type
->base_type
) {
986 value
.u
[i
+offset
] = src
->get_uint_component(id
++);
989 value
.i
[i
+offset
] = src
->get_int_component(id
++);
991 case GLSL_TYPE_FLOAT
:
992 value
.f
[i
+offset
] = src
->get_float_component(id
++);
995 value
.b
[i
+offset
] = src
->get_bool_component(id
++);
998 assert(!"Should not get here.");
1006 ir_constant::has_value(const ir_constant
*c
) const
1008 if (this->type
!= c
->type
)
1011 if (this->type
->is_array()) {
1012 for (unsigned i
= 0; i
< this->type
->length
; i
++) {
1013 if (!this->array_elements
[i
]->has_value(c
->array_elements
[i
]))
1019 if (this->type
->base_type
== GLSL_TYPE_STRUCT
) {
1020 const exec_node
*a_node
= this->components
.head
;
1021 const exec_node
*b_node
= c
->components
.head
;
1023 while (!a_node
->is_tail_sentinel()) {
1024 assert(!b_node
->is_tail_sentinel());
1026 const ir_constant
*const a_field
= (ir_constant
*) a_node
;
1027 const ir_constant
*const b_field
= (ir_constant
*) b_node
;
1029 if (!a_field
->has_value(b_field
))
1032 a_node
= a_node
->next
;
1033 b_node
= b_node
->next
;
1039 for (unsigned i
= 0; i
< this->type
->components(); i
++) {
1040 switch (this->type
->base_type
) {
1041 case GLSL_TYPE_UINT
:
1042 if (this->value
.u
[i
] != c
->value
.u
[i
])
1046 if (this->value
.i
[i
] != c
->value
.i
[i
])
1049 case GLSL_TYPE_FLOAT
:
1050 if (this->value
.f
[i
] != c
->value
.f
[i
])
1053 case GLSL_TYPE_BOOL
:
1054 if (this->value
.b
[i
] != c
->value
.b
[i
])
1058 assert(!"Should not get here.");
1067 ir_constant::is_zero() const
1069 if (!this->type
->is_scalar() && !this->type
->is_vector())
1072 for (unsigned c
= 0; c
< this->type
->vector_elements
; c
++) {
1073 switch (this->type
->base_type
) {
1074 case GLSL_TYPE_FLOAT
:
1075 if (this->value
.f
[c
] != 0.0)
1079 if (this->value
.i
[c
] != 0)
1082 case GLSL_TYPE_UINT
:
1083 if (this->value
.u
[c
] != 0)
1086 case GLSL_TYPE_BOOL
:
1087 if (this->value
.b
[c
] != false)
1091 /* The only other base types are structures, arrays, and samplers.
1092 * Samplers cannot be constants, and the others should have been
1093 * filtered out above.
1095 assert(!"Should not get here.");
1104 ir_constant::is_one() const
1106 if (!this->type
->is_scalar() && !this->type
->is_vector())
1109 for (unsigned c
= 0; c
< this->type
->vector_elements
; c
++) {
1110 switch (this->type
->base_type
) {
1111 case GLSL_TYPE_FLOAT
:
1112 if (this->value
.f
[c
] != 1.0)
1116 if (this->value
.i
[c
] != 1)
1119 case GLSL_TYPE_UINT
:
1120 if (this->value
.u
[c
] != 1)
1123 case GLSL_TYPE_BOOL
:
1124 if (this->value
.b
[c
] != true)
1128 /* The only other base types are structures, arrays, and samplers.
1129 * Samplers cannot be constants, and the others should have been
1130 * filtered out above.
1132 assert(!"Should not get here.");
1141 ir_constant::is_negative_one() const
1143 if (!this->type
->is_scalar() && !this->type
->is_vector())
1146 if (this->type
->is_boolean())
1149 for (unsigned c
= 0; c
< this->type
->vector_elements
; c
++) {
1150 switch (this->type
->base_type
) {
1151 case GLSL_TYPE_FLOAT
:
1152 if (this->value
.f
[c
] != -1.0)
1156 if (this->value
.i
[c
] != -1)
1159 case GLSL_TYPE_UINT
:
1160 if (int(this->value
.u
[c
]) != -1)
1164 /* The only other base types are structures, arrays, samplers, and
1165 * booleans. Samplers cannot be constants, and the others should
1166 * have been filtered out above.
1168 assert(!"Should not get here.");
1177 ir_constant::is_basis() const
1179 if (!this->type
->is_scalar() && !this->type
->is_vector())
1182 if (this->type
->is_boolean())
1186 for (unsigned c
= 0; c
< this->type
->vector_elements
; c
++) {
1187 switch (this->type
->base_type
) {
1188 case GLSL_TYPE_FLOAT
:
1189 if (this->value
.f
[c
] == 1.0)
1191 else if (this->value
.f
[c
] != 0.0)
1195 if (this->value
.i
[c
] == 1)
1197 else if (this->value
.i
[c
] != 0)
1200 case GLSL_TYPE_UINT
:
1201 if (int(this->value
.u
[c
]) == 1)
1203 else if (int(this->value
.u
[c
]) != 0)
1207 /* The only other base types are structures, arrays, samplers, and
1208 * booleans. Samplers cannot be constants, and the others should
1209 * have been filtered out above.
1211 assert(!"Should not get here.");
1221 this->ir_type
= ir_type_loop
;
1222 this->cmp
= ir_unop_neg
;
1225 this->increment
= NULL
;
1226 this->counter
= NULL
;
1230 ir_dereference_variable::ir_dereference_variable(ir_variable
*var
)
1232 assert(var
!= NULL
);
1234 this->ir_type
= ir_type_dereference_variable
;
1236 this->type
= var
->type
;
1240 ir_dereference_array::ir_dereference_array(ir_rvalue
*value
,
1241 ir_rvalue
*array_index
)
1243 this->ir_type
= ir_type_dereference_array
;
1244 this->array_index
= array_index
;
1245 this->set_array(value
);
1249 ir_dereference_array::ir_dereference_array(ir_variable
*var
,
1250 ir_rvalue
*array_index
)
1252 void *ctx
= ralloc_parent(var
);
1254 this->ir_type
= ir_type_dereference_array
;
1255 this->array_index
= array_index
;
1256 this->set_array(new(ctx
) ir_dereference_variable(var
));
1261 ir_dereference_array::set_array(ir_rvalue
*value
)
1263 assert(value
!= NULL
);
1265 this->array
= value
;
1267 const glsl_type
*const vt
= this->array
->type
;
1269 if (vt
->is_array()) {
1270 type
= vt
->element_type();
1271 } else if (vt
->is_matrix()) {
1272 type
= vt
->column_type();
1273 } else if (vt
->is_vector()) {
1274 type
= vt
->get_base_type();
1279 ir_dereference_record::ir_dereference_record(ir_rvalue
*value
,
1282 assert(value
!= NULL
);
1284 this->ir_type
= ir_type_dereference_record
;
1285 this->record
= value
;
1286 this->field
= ralloc_strdup(this, field
);
1287 this->type
= this->record
->type
->field_type(field
);
1291 ir_dereference_record::ir_dereference_record(ir_variable
*var
,
1294 void *ctx
= ralloc_parent(var
);
1296 this->ir_type
= ir_type_dereference_record
;
1297 this->record
= new(ctx
) ir_dereference_variable(var
);
1298 this->field
= ralloc_strdup(this, field
);
1299 this->type
= this->record
->type
->field_type(field
);
1303 ir_dereference::is_lvalue() const
1305 ir_variable
*var
= this->variable_referenced();
1307 /* Every l-value derference chain eventually ends in a variable.
1309 if ((var
== NULL
) || var
->read_only
)
1312 /* From page 17 (page 23 of the PDF) of the GLSL 1.20 spec:
1314 * "Samplers cannot be treated as l-values; hence cannot be used
1315 * as out or inout function parameters, nor can they be
1318 if (this->type
->contains_sampler())
1325 static const char *tex_opcode_strs
[] = { "tex", "txb", "txl", "txd", "txf", "txf_ms", "txs", "lod" };
1327 const char *ir_texture::opcode_string()
1329 assert((unsigned int) op
<=
1330 sizeof(tex_opcode_strs
) / sizeof(tex_opcode_strs
[0]));
1331 return tex_opcode_strs
[op
];
1335 ir_texture::get_opcode(const char *str
)
1337 const int count
= sizeof(tex_opcode_strs
) / sizeof(tex_opcode_strs
[0]);
1338 for (int op
= 0; op
< count
; op
++) {
1339 if (strcmp(str
, tex_opcode_strs
[op
]) == 0)
1340 return (ir_texture_opcode
) op
;
1342 return (ir_texture_opcode
) -1;
1347 ir_texture::set_sampler(ir_dereference
*sampler
, const glsl_type
*type
)
1349 assert(sampler
!= NULL
);
1350 assert(type
!= NULL
);
1351 this->sampler
= sampler
;
1354 if (this->op
== ir_txs
) {
1355 assert(type
->base_type
== GLSL_TYPE_INT
);
1356 } else if (this->op
== ir_lod
) {
1357 assert(type
->vector_elements
== 2);
1358 assert(type
->base_type
== GLSL_TYPE_FLOAT
);
1360 assert(sampler
->type
->sampler_type
== (int) type
->base_type
);
1361 if (sampler
->type
->sampler_shadow
)
1362 assert(type
->vector_elements
== 4 || type
->vector_elements
== 1);
1364 assert(type
->vector_elements
== 4);
1370 ir_swizzle::init_mask(const unsigned *comp
, unsigned count
)
1372 assert((count
>= 1) && (count
<= 4));
1374 memset(&this->mask
, 0, sizeof(this->mask
));
1375 this->mask
.num_components
= count
;
1377 unsigned dup_mask
= 0;
1380 assert(comp
[3] <= 3);
1381 dup_mask
|= (1U << comp
[3])
1382 & ((1U << comp
[0]) | (1U << comp
[1]) | (1U << comp
[2]));
1383 this->mask
.w
= comp
[3];
1386 assert(comp
[2] <= 3);
1387 dup_mask
|= (1U << comp
[2])
1388 & ((1U << comp
[0]) | (1U << comp
[1]));
1389 this->mask
.z
= comp
[2];
1392 assert(comp
[1] <= 3);
1393 dup_mask
|= (1U << comp
[1])
1394 & ((1U << comp
[0]));
1395 this->mask
.y
= comp
[1];
1398 assert(comp
[0] <= 3);
1399 this->mask
.x
= comp
[0];
1402 this->mask
.has_duplicates
= dup_mask
!= 0;
1404 /* Based on the number of elements in the swizzle and the base type
1405 * (i.e., float, int, unsigned, or bool) of the vector being swizzled,
1406 * generate the type of the resulting value.
1408 type
= glsl_type::get_instance(val
->type
->base_type
, mask
.num_components
, 1);
1411 ir_swizzle::ir_swizzle(ir_rvalue
*val
, unsigned x
, unsigned y
, unsigned z
,
1412 unsigned w
, unsigned count
)
1415 const unsigned components
[4] = { x
, y
, z
, w
};
1416 this->ir_type
= ir_type_swizzle
;
1417 this->init_mask(components
, count
);
1420 ir_swizzle::ir_swizzle(ir_rvalue
*val
, const unsigned *comp
,
1424 this->ir_type
= ir_type_swizzle
;
1425 this->init_mask(comp
, count
);
1428 ir_swizzle::ir_swizzle(ir_rvalue
*val
, ir_swizzle_mask mask
)
1430 this->ir_type
= ir_type_swizzle
;
1433 this->type
= glsl_type::get_instance(val
->type
->base_type
,
1434 mask
.num_components
, 1);
1443 ir_swizzle::create(ir_rvalue
*val
, const char *str
, unsigned vector_length
)
1445 void *ctx
= ralloc_parent(val
);
1447 /* For each possible swizzle character, this table encodes the value in
1448 * \c idx_map that represents the 0th element of the vector. For invalid
1449 * swizzle characters (e.g., 'k'), a special value is used that will allow
1450 * detection of errors.
1452 static const unsigned char base_idx
[26] = {
1453 /* a b c d e f g h i j k l m */
1454 R
, R
, I
, I
, I
, I
, R
, I
, I
, I
, I
, I
, I
,
1455 /* n o p q r s t u v w x y z */
1456 I
, I
, S
, S
, R
, S
, S
, I
, I
, X
, X
, X
, X
1459 /* Each valid swizzle character has an entry in the previous table. This
1460 * table encodes the base index encoded in the previous table plus the actual
1461 * index of the swizzle character. When processing swizzles, the first
1462 * character in the string is indexed in the previous table. Each character
1463 * in the string is indexed in this table, and the value found there has the
1464 * value form the first table subtracted. The result must be on the range
1467 * For example, the string "wzyx" will get X from the first table. Each of
1468 * the charcaters will get X+3, X+2, X+1, and X+0 from this table. After
1469 * subtraction, the swizzle values are { 3, 2, 1, 0 }.
1471 * The string "wzrg" will get X from the first table. Each of the characters
1472 * will get X+3, X+2, R+0, and R+1 from this table. After subtraction, the
1473 * swizzle values are { 3, 2, 4, 5 }. Since 4 and 5 are outside the range
1474 * [0,3], the error is detected.
1476 static const unsigned char idx_map
[26] = {
1477 /* a b c d e f g h i j k l m */
1478 R
+3, R
+2, 0, 0, 0, 0, R
+1, 0, 0, 0, 0, 0, 0,
1479 /* n o p q r s t u v w x y z */
1480 0, 0, S
+2, S
+3, R
+0, S
+0, S
+1, 0, 0, X
+3, X
+0, X
+1, X
+2
1483 int swiz_idx
[4] = { 0, 0, 0, 0 };
1487 /* Validate the first character in the swizzle string and look up the base
1488 * index value as described above.
1490 if ((str
[0] < 'a') || (str
[0] > 'z'))
1493 const unsigned base
= base_idx
[str
[0] - 'a'];
1496 for (i
= 0; (i
< 4) && (str
[i
] != '\0'); i
++) {
1497 /* Validate the next character, and, as described above, convert it to a
1500 if ((str
[i
] < 'a') || (str
[i
] > 'z'))
1503 swiz_idx
[i
] = idx_map
[str
[i
] - 'a'] - base
;
1504 if ((swiz_idx
[i
] < 0) || (swiz_idx
[i
] >= (int) vector_length
))
1511 return new(ctx
) ir_swizzle(val
, swiz_idx
[0], swiz_idx
[1], swiz_idx
[2],
1521 ir_swizzle::variable_referenced() const
1523 return this->val
->variable_referenced();
1527 ir_variable::ir_variable(const struct glsl_type
*type
, const char *name
,
1528 ir_variable_mode mode
)
1529 : max_array_access(0), read_only(false), centroid(false), invariant(false),
1530 mode(mode
), interpolation(INTERP_QUALIFIER_NONE
)
1532 this->ir_type
= ir_type_variable
;
1534 this->name
= ralloc_strdup(this, name
);
1535 this->explicit_location
= false;
1536 this->has_initializer
= false;
1537 this->location
= -1;
1538 this->location_frac
= 0;
1539 this->warn_extension
= NULL
;
1540 this->constant_value
= NULL
;
1541 this->constant_initializer
= NULL
;
1542 this->origin_upper_left
= false;
1543 this->pixel_center_integer
= false;
1544 this->depth_layout
= ir_depth_layout_none
;
1547 if (type
&& type
->base_type
== GLSL_TYPE_SAMPLER
)
1548 this->read_only
= true;
1553 ir_variable::interpolation_string() const
1555 switch (this->interpolation
) {
1556 case INTERP_QUALIFIER_NONE
: return "no";
1557 case INTERP_QUALIFIER_SMOOTH
: return "smooth";
1558 case INTERP_QUALIFIER_FLAT
: return "flat";
1559 case INTERP_QUALIFIER_NOPERSPECTIVE
: return "noperspective";
1562 assert(!"Should not get here.");
1567 glsl_interp_qualifier
1568 ir_variable::determine_interpolation_mode(bool flat_shade
)
1570 if (this->interpolation
!= INTERP_QUALIFIER_NONE
)
1571 return (glsl_interp_qualifier
) this->interpolation
;
1572 int location
= this->location
;
1574 location
== VARYING_SLOT_COL0
|| location
== VARYING_SLOT_COL1
;
1575 if (flat_shade
&& is_gl_Color
)
1576 return INTERP_QUALIFIER_FLAT
;
1578 return INTERP_QUALIFIER_SMOOTH
;
1582 ir_function_signature::ir_function_signature(const glsl_type
*return_type
)
1583 : return_type(return_type
), is_defined(false), _function(NULL
)
1585 this->ir_type
= ir_type_function_signature
;
1586 this->is_builtin
= false;
1587 this->origin
= NULL
;
1592 modes_match(unsigned a
, unsigned b
)
1597 /* Accept "in" vs. "const in" */
1598 if ((a
== ir_var_const_in
&& b
== ir_var_function_in
) ||
1599 (b
== ir_var_const_in
&& a
== ir_var_function_in
))
1607 ir_function_signature::qualifiers_match(exec_list
*params
)
1609 exec_list_iterator iter_a
= parameters
.iterator();
1610 exec_list_iterator iter_b
= params
->iterator();
1612 /* check that the qualifiers match. */
1613 while (iter_a
.has_next()) {
1614 ir_variable
*a
= (ir_variable
*)iter_a
.get();
1615 ir_variable
*b
= (ir_variable
*)iter_b
.get();
1617 if (a
->read_only
!= b
->read_only
||
1618 !modes_match(a
->mode
, b
->mode
) ||
1619 a
->interpolation
!= b
->interpolation
||
1620 a
->centroid
!= b
->centroid
) {
1622 /* parameter a's qualifiers don't match */
1634 ir_function_signature::replace_parameters(exec_list
*new_params
)
1636 /* Destroy all of the previous parameter information. If the previous
1637 * parameter information comes from the function prototype, it may either
1638 * specify incorrect parameter names or not have names at all.
1640 foreach_iter(exec_list_iterator
, iter
, parameters
) {
1641 assert(((ir_instruction
*) iter
.get())->as_variable() != NULL
);
1646 new_params
->move_nodes_to(¶meters
);
1650 ir_function::ir_function(const char *name
)
1652 this->ir_type
= ir_type_function
;
1653 this->name
= ralloc_strdup(this, name
);
1658 ir_function::has_user_signature()
1660 foreach_list(n
, &this->signatures
) {
1661 ir_function_signature
*const sig
= (ir_function_signature
*) n
;
1662 if (!sig
->is_builtin
)
1670 ir_rvalue::error_value(void *mem_ctx
)
1672 ir_rvalue
*v
= new(mem_ctx
) ir_rvalue
;
1674 v
->type
= glsl_type::error_type
;
1680 visit_exec_list(exec_list
*list
, ir_visitor
*visitor
)
1682 foreach_iter(exec_list_iterator
, iter
, *list
) {
1683 ((ir_instruction
*)iter
.get())->accept(visitor
);
1689 steal_memory(ir_instruction
*ir
, void *new_ctx
)
1691 ir_variable
*var
= ir
->as_variable();
1692 ir_constant
*constant
= ir
->as_constant();
1693 if (var
!= NULL
&& var
->constant_value
!= NULL
)
1694 steal_memory(var
->constant_value
, ir
);
1696 if (var
!= NULL
&& var
->constant_initializer
!= NULL
)
1697 steal_memory(var
->constant_initializer
, ir
);
1699 /* The components of aggregate constants are not visited by the normal
1700 * visitor, so steal their values by hand.
1702 if (constant
!= NULL
) {
1703 if (constant
->type
->is_record()) {
1704 foreach_iter(exec_list_iterator
, iter
, constant
->components
) {
1705 ir_constant
*field
= (ir_constant
*)iter
.get();
1706 steal_memory(field
, ir
);
1708 } else if (constant
->type
->is_array()) {
1709 for (unsigned int i
= 0; i
< constant
->type
->length
; i
++) {
1710 steal_memory(constant
->array_elements
[i
], ir
);
1715 ralloc_steal(new_ctx
, ir
);
1720 reparent_ir(exec_list
*list
, void *mem_ctx
)
1722 foreach_list(node
, list
) {
1723 visit_tree((ir_instruction
*) node
, steal_memory
, mem_ctx
);
1729 try_min_one(ir_rvalue
*ir
)
1731 ir_expression
*expr
= ir
->as_expression();
1733 if (!expr
|| expr
->operation
!= ir_binop_min
)
1736 if (expr
->operands
[0]->is_one())
1737 return expr
->operands
[1];
1739 if (expr
->operands
[1]->is_one())
1740 return expr
->operands
[0];
1746 try_max_zero(ir_rvalue
*ir
)
1748 ir_expression
*expr
= ir
->as_expression();
1750 if (!expr
|| expr
->operation
!= ir_binop_max
)
1753 if (expr
->operands
[0]->is_zero())
1754 return expr
->operands
[1];
1756 if (expr
->operands
[1]->is_zero())
1757 return expr
->operands
[0];
1763 ir_rvalue::as_rvalue_to_saturate()
1765 ir_expression
*expr
= this->as_expression();
1770 ir_rvalue
*max_zero
= try_max_zero(expr
);
1772 return try_min_one(max_zero
);
1774 ir_rvalue
*min_one
= try_min_one(expr
);
1776 return try_max_zero(min_one
);
1785 vertices_per_prim(GLenum prim
)
1794 case GL_LINES_ADJACENCY
:
1796 case GL_TRIANGLES_ADJACENCY
:
1799 assert(!"Bad primitive");