1 /*
2 * Copyright © 2010 Intel Corporation
3 *
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:
10 *
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
13 * Software.
14 *
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.
22 */
23
24 /**
25 * \file opt_swizzle.cpp
26 * Optimize swizzle operations.
27 *
28 * First, compact a sequence of swizzled swizzles into a single swizzle.
29 *
30 * If the final resulting swizzle doesn't change the order or count of
31 * components, then remove the swizzle so that other optimization passes see
32 * the value behind it.
33 */
34
35 #include "ir.h"
36 #include "ir_visitor.h"
37 #include "ir_rvalue_visitor.h"
38 #include "compiler/glsl_types.h"
39
40 namespace {
41
42 class ir_opt_swizzle_visitor : public ir_rvalue_visitor {
43 public:
ir_opt_swizzle_visitor()44 ir_opt_swizzle_visitor()
45 {
46 this->progress = false;
47 }
48
49 void handle_rvalue(ir_rvalue **rvalue);
50 bool progress;
51 };
52
53 } /* unnamed namespace */
54
55 void
handle_rvalue(ir_rvalue ** rvalue)56 ir_opt_swizzle_visitor::handle_rvalue(ir_rvalue **rvalue)
57 {
58 if (!*rvalue)
59 return;
60
61 ir_swizzle *swiz = (*rvalue)->as_swizzle();
62
63 if (!swiz)
64 return;
65
66 ir_swizzle *swiz2;
67
68 while ((swiz2 = swiz->val->as_swizzle()) != NULL) {
69 int mask2[4];
70
71 memset(&mask2, 0, sizeof(mask2));
72 if (swiz2->mask.num_components >= 1)
73 mask2[0] = swiz2->mask.x;
74 if (swiz2->mask.num_components >= 2)
75 mask2[1] = swiz2->mask.y;
76 if (swiz2->mask.num_components >= 3)
77 mask2[2] = swiz2->mask.z;
78 if (swiz2->mask.num_components >= 4)
79 mask2[3] = swiz2->mask.w;
80
81 if (swiz->mask.num_components >= 1)
82 swiz->mask.x = mask2[swiz->mask.x];
83 if (swiz->mask.num_components >= 2)
84 swiz->mask.y = mask2[swiz->mask.y];
85 if (swiz->mask.num_components >= 3)
86 swiz->mask.z = mask2[swiz->mask.z];
87 if (swiz->mask.num_components >= 4)
88 swiz->mask.w = mask2[swiz->mask.w];
89
90 swiz->val = swiz2->val;
91
92 this->progress = true;
93 }
94
95 if (swiz->type != swiz->val->type)
96 return;
97
98 int elems = swiz->val->type->vector_elements;
99 if (swiz->mask.x != 0)
100 return;
101 if (elems >= 2 && swiz->mask.y != 1)
102 return;
103 if (elems >= 3 && swiz->mask.z != 2)
104 return;
105 if (elems >= 4 && swiz->mask.w != 3)
106 return;
107
108 this->progress = true;
109 *rvalue = swiz->val;
110 }
111
112 bool
optimize_swizzles(exec_list * instructions)113 optimize_swizzles(exec_list *instructions)
114 {
115 ir_opt_swizzle_visitor v;
116 visit_list_elements(&v, instructions);
117
118 return v.progress;
119 }
120