1 /*
2 * Copyright © 2019 Vasily Khoruzhick <[email protected]>
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 DEALINGS
21 * IN THE SOFTWARE.
22 */
23
24 #include "nir.h"
25 #include "nir_builder.h"
26
27 #include "lima_ir.h"
28
29 static bool
lima_nir_split_load_input_instr(nir_builder * b,nir_instr * instr,UNUSED void * cb_data)30 lima_nir_split_load_input_instr(nir_builder *b,
31 nir_instr *instr,
32 UNUSED void *cb_data)
33 {
34 if (instr->type != nir_instr_type_alu)
35 return false;
36
37 nir_alu_instr *alu = nir_instr_as_alu(instr);
38 if (alu->op != nir_op_mov)
39 return false;
40
41 nir_def *ssa = alu->src[0].src.ssa;
42 if (ssa->parent_instr->type != nir_instr_type_intrinsic)
43 return false;
44
45 nir_intrinsic_instr *intrin = nir_instr_as_intrinsic(ssa->parent_instr);
46 if (intrin->intrinsic != nir_intrinsic_load_input)
47 return false;
48
49 uint8_t swizzle = alu->src[0].swizzle[0];
50 int i;
51
52 for (i = 1; i < alu->def.num_components; i++)
53 if (alu->src[0].swizzle[i] != (swizzle + i))
54 break;
55
56 if (i != alu->def.num_components)
57 return false;
58
59 /* mali4xx can't access unaligned vec3, don't split load input */
60 if (alu->def.num_components == 3 && swizzle > 0)
61 return false;
62
63 /* mali4xx can't access unaligned vec2, don't split load input */
64 if (alu->def.num_components == 2 &&
65 swizzle != 0 && swizzle != 2)
66 return false;
67
68 b->cursor = nir_before_instr(&intrin->instr);
69 nir_intrinsic_instr *new_intrin = nir_intrinsic_instr_create(
70 b->shader,
71 intrin->intrinsic);
72 nir_def_init(&new_intrin->instr, &new_intrin->def,
73 alu->def.num_components, ssa->bit_size);
74 new_intrin->num_components = alu->def.num_components;
75 nir_intrinsic_set_base(new_intrin, nir_intrinsic_base(intrin));
76 nir_intrinsic_set_component(new_intrin, nir_intrinsic_component(intrin) + swizzle);
77 nir_intrinsic_set_dest_type(new_intrin, nir_intrinsic_dest_type(intrin));
78
79 /* offset */
80 new_intrin->src[0] = nir_src_for_ssa(intrin->src[0].ssa);
81
82 nir_builder_instr_insert(b, &new_intrin->instr);
83 nir_def_replace(&alu->def, &new_intrin->def);
84 return true;
85 }
86
87 /* Replaces a single load of several packed varyings and number of movs with
88 * a number of loads of smaller size
89 */
90 bool
lima_nir_split_load_input(nir_shader * shader)91 lima_nir_split_load_input(nir_shader *shader)
92 {
93 return nir_shader_instructions_pass(shader, lima_nir_split_load_input_instr,
94 nir_metadata_control_flow,
95 NULL);
96 }
97