1/*
2 * Copyright © 2016 Red Hat
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 FROM,
20 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21 * SOFTWARE.
22 */
23
24/* Lowers the additional tex_src_plane src, generated by nir_lower_tex
25 * for planar YUV textures, into separate samplers, matching the logic
26 * that mesa/st uses to insert additional sampler view/state (since both
27 * sides need to agree).
28 *
29 * This should run after nir_lower_samplers.
30 */
31
32#include "compiler/nir/nir.h"
33#include "st_nir.h"
34
35typedef struct {
36   unsigned lower_2plane;
37   unsigned lower_3plane;
38
39   /* Maps a primary sampler (used for Y) to the U or UV sampler.  In
40    * case of 3-plane YUV format, the V plane is next sampler after U.
41    */
42   unsigned char sampler_map[PIPE_MAX_SAMPLERS][2];
43} lower_tex_src_state;
44
45static void
46assign_extra_samplers(lower_tex_src_state *state, unsigned free_slots)
47{
48   unsigned mask = state->lower_2plane | state->lower_3plane;
49
50   while (mask) {
51      unsigned extra, y_samp = u_bit_scan(&mask);
52
53      extra = u_bit_scan(&free_slots);
54      state->sampler_map[y_samp][0] = extra;
55
56      if (state->lower_3plane & (1 << y_samp)) {
57         extra = u_bit_scan(&free_slots);
58         state->sampler_map[y_samp][1] = extra;
59      }
60   }
61}
62
63static void
64lower_tex_src_plane_block(lower_tex_src_state *state, nir_block *block)
65{
66   nir_foreach_instr(instr, block) {
67      if (instr->type != nir_instr_type_tex)
68         continue;
69
70      nir_tex_instr *tex = nir_instr_as_tex(instr);
71      int plane_index = nir_tex_instr_src_index(tex, nir_tex_src_plane);
72
73      if (plane_index < 0)
74         continue;
75
76      nir_const_value *plane = nir_src_as_const_value(tex->src[plane_index].src);
77      assume(plane);
78
79      if (plane->i32[0] > 0) {
80         unsigned y_samp = tex->texture_index;
81
82         assume(tex->texture_index == tex->sampler_index);
83         assume(((state->lower_3plane & (1 << y_samp)) && plane->i32[0] < 3) ||
84               (plane->i32[0] < 2));
85
86         tex->texture_index = tex->sampler_index =
87               state->sampler_map[y_samp][plane->i32[0] - 1];
88      }
89
90      nir_tex_instr_remove_src(tex, plane_index);
91   }
92}
93
94static void
95lower_tex_src_plane_impl(lower_tex_src_state *state, nir_function_impl *impl)
96{
97   nir_foreach_block(block, impl) {
98      lower_tex_src_plane_block(state, block);
99   }
100
101   nir_metadata_preserve(impl, nir_metadata_block_index |
102                               nir_metadata_dominance);
103}
104
105void
106st_nir_lower_tex_src_plane(struct nir_shader *shader, unsigned free_slots,
107                           unsigned lower_2plane, unsigned lower_3plane)
108{
109   lower_tex_src_state state = {0};
110
111   state.lower_2plane = lower_2plane;
112   state.lower_3plane = lower_3plane;
113
114   assign_extra_samplers(&state, free_slots);
115
116   nir_foreach_function(function, shader) {
117      if (function->impl)
118         lower_tex_src_plane_impl(&state, function->impl);
119   }
120}
121