1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
|
/*
* Mesa 3-D graphics library
*
* Copyright 2003 VMware, Inc.
* Copyright 2009 VMware, Inc.
* All Rights Reserved.
* Copyright (C) 2016 Advanced Micro Devices, Inc.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
* THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
* DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
* OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
* USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
#include "main/glheader.h"
#include "main/context.h"
#include "main/varray.h"
#include "main/macros.h"
#include "main/sse_minmax.h"
#include "x86/common_x86_asm.h"
#include "util/hash_table.h"
struct minmax_cache_key {
GLintptr offset;
GLuint count;
GLenum type;
};
struct minmax_cache_entry {
struct minmax_cache_key key;
GLuint min;
GLuint max;
};
static uint32_t
vbo_minmax_cache_hash(const struct minmax_cache_key *key)
{
return _mesa_hash_data(key, sizeof(*key));
}
static bool
vbo_minmax_cache_key_equal(const struct minmax_cache_key *a,
const struct minmax_cache_key *b)
{
return (a->offset == b->offset) && (a->count == b->count) && (a->type == b->type);
}
static void
vbo_minmax_cache_delete_entry(struct hash_entry *entry)
{
free(entry->data);
}
static GLboolean
vbo_use_minmax_cache(struct gl_buffer_object *bufferObj)
{
if (bufferObj->UsageHistory & (USAGE_TEXTURE_BUFFER |
USAGE_ATOMIC_COUNTER_BUFFER |
USAGE_SHADER_STORAGE_BUFFER |
USAGE_TRANSFORM_FEEDBACK_BUFFER |
USAGE_PIXEL_PACK_BUFFER))
return GL_FALSE;
if ((bufferObj->Mappings[MAP_USER].AccessFlags &
(GL_MAP_PERSISTENT_BIT | GL_MAP_WRITE_BIT)) ==
(GL_MAP_PERSISTENT_BIT | GL_MAP_WRITE_BIT))
return GL_FALSE;
return GL_TRUE;
}
void
vbo_delete_minmax_cache(struct gl_buffer_object *bufferObj)
{
_mesa_hash_table_destroy(bufferObj->MinMaxCache, vbo_minmax_cache_delete_entry);
bufferObj->MinMaxCache = NULL;
}
static GLboolean
vbo_get_minmax_cached(struct gl_buffer_object *bufferObj,
GLenum type, GLintptr offset, GLuint count,
GLuint *min_index, GLuint *max_index)
{
GLboolean found = GL_FALSE;
struct minmax_cache_key key;
uint32_t hash;
struct hash_entry *result;
if (!bufferObj->MinMaxCache)
return GL_FALSE;
if (!vbo_use_minmax_cache(bufferObj))
return GL_FALSE;
mtx_lock(&bufferObj->Mutex);
if (bufferObj->MinMaxCacheDirty) {
_mesa_hash_table_clear(bufferObj->MinMaxCache, vbo_minmax_cache_delete_entry);
bufferObj->MinMaxCacheDirty = false;
goto out;
}
key.type = type;
key.offset = offset;
key.count = count;
hash = vbo_minmax_cache_hash(&key);
result = _mesa_hash_table_search_pre_hashed(bufferObj->MinMaxCache, hash, &key);
if (result) {
struct minmax_cache_entry *entry = result->data;
*min_index = entry->min;
*max_index = entry->max;
found = GL_TRUE;
}
out:
mtx_unlock(&bufferObj->Mutex);
return found;
}
static void
vbo_minmax_cache_store(struct gl_context *ctx,
struct gl_buffer_object *bufferObj,
GLenum type, GLintptr offset, GLuint count,
GLuint min, GLuint max)
{
struct minmax_cache_entry *entry;
struct hash_entry *table_entry;
uint32_t hash;
if (!vbo_use_minmax_cache(bufferObj))
return;
mtx_lock(&bufferObj->Mutex);
if (!bufferObj->MinMaxCache) {
bufferObj->MinMaxCache =
_mesa_hash_table_create(NULL,
(uint32_t (*)(const void *))vbo_minmax_cache_hash,
(bool (*)(const void *, const void *))vbo_minmax_cache_key_equal);
if (!bufferObj->MinMaxCache)
goto out;
}
entry = MALLOC_STRUCT(minmax_cache_entry);
if (!entry)
goto out;
entry->key.offset = offset;
entry->key.count = count;
entry->key.type = type;
entry->min = min;
entry->max = max;
hash = vbo_minmax_cache_hash(&entry->key);
table_entry = _mesa_hash_table_search_pre_hashed(bufferObj->MinMaxCache,
hash, &entry->key);
if (table_entry) {
/* It seems like this could happen when two contexts are rendering using
* the same buffer object from multiple threads.
*/
_mesa_debug(ctx, "duplicate entry in minmax cache\n");
free(entry);
goto out;
}
table_entry = _mesa_hash_table_insert_pre_hashed(bufferObj->MinMaxCache,
hash, &entry->key, entry);
if (!table_entry)
free(entry);
out:
mtx_unlock(&bufferObj->Mutex);
}
/**
* Compute min and max elements by scanning the index buffer for
* glDraw[Range]Elements() calls.
* If primitive restart is enabled, we need to ignore restart
* indexes when computing min/max.
*/
static void
vbo_get_minmax_index(struct gl_context *ctx,
const struct _mesa_prim *prim,
const struct _mesa_index_buffer *ib,
GLuint *min_index, GLuint *max_index,
const GLuint count)
{
const GLboolean restart = ctx->Array._PrimitiveRestart;
const GLuint restartIndex = _mesa_primitive_restart_index(ctx, ib->type);
const int index_size = vbo_sizeof_ib_type(ib->type);
const char *indices;
GLuint i;
indices = (char *) ib->ptr + prim->start * index_size;
if (_mesa_is_bufferobj(ib->obj)) {
GLsizeiptr size = MIN2(count * index_size, ib->obj->Size);
if (vbo_get_minmax_cached(ib->obj, ib->type, (GLintptr) indices, count,
min_index, max_index))
return;
indices = ctx->Driver.MapBufferRange(ctx, (GLintptr) indices, size,
GL_MAP_READ_BIT, ib->obj,
MAP_INTERNAL);
}
switch (ib->type) {
case GL_UNSIGNED_INT: {
const GLuint *ui_indices = (const GLuint *)indices;
GLuint max_ui = 0;
GLuint min_ui = ~0U;
if (restart) {
for (i = 0; i < count; i++) {
if (ui_indices[i] != restartIndex) {
if (ui_indices[i] > max_ui) max_ui = ui_indices[i];
if (ui_indices[i] < min_ui) min_ui = ui_indices[i];
}
}
}
else {
#if defined(USE_SSE41)
if (cpu_has_sse4_1) {
_mesa_uint_array_min_max(ui_indices, &min_ui, &max_ui, count);
}
else
#endif
for (i = 0; i < count; i++) {
if (ui_indices[i] > max_ui) max_ui = ui_indices[i];
if (ui_indices[i] < min_ui) min_ui = ui_indices[i];
}
}
*min_index = min_ui;
*max_index = max_ui;
break;
}
case GL_UNSIGNED_SHORT: {
const GLushort *us_indices = (const GLushort *)indices;
GLuint max_us = 0;
GLuint min_us = ~0U;
if (restart) {
for (i = 0; i < count; i++) {
if (us_indices[i] != restartIndex) {
if (us_indices[i] > max_us) max_us = us_indices[i];
if (us_indices[i] < min_us) min_us = us_indices[i];
}
}
}
else {
for (i = 0; i < count; i++) {
if (us_indices[i] > max_us) max_us = us_indices[i];
if (us_indices[i] < min_us) min_us = us_indices[i];
}
}
*min_index = min_us;
*max_index = max_us;
break;
}
case GL_UNSIGNED_BYTE: {
const GLubyte *ub_indices = (const GLubyte *)indices;
GLuint max_ub = 0;
GLuint min_ub = ~0U;
if (restart) {
for (i = 0; i < count; i++) {
if (ub_indices[i] != restartIndex) {
if (ub_indices[i] > max_ub) max_ub = ub_indices[i];
if (ub_indices[i] < min_ub) min_ub = ub_indices[i];
}
}
}
else {
for (i = 0; i < count; i++) {
if (ub_indices[i] > max_ub) max_ub = ub_indices[i];
if (ub_indices[i] < min_ub) min_ub = ub_indices[i];
}
}
*min_index = min_ub;
*max_index = max_ub;
break;
}
default:
unreachable("not reached");
}
if (_mesa_is_bufferobj(ib->obj)) {
vbo_minmax_cache_store(ctx, ib->obj, ib->type, prim->start, count,
*min_index, *max_index);
ctx->Driver.UnmapBuffer(ctx, ib->obj, MAP_INTERNAL);
}
}
/**
* Compute min and max elements for nr_prims
*/
void
vbo_get_minmax_indices(struct gl_context *ctx,
const struct _mesa_prim *prims,
const struct _mesa_index_buffer *ib,
GLuint *min_index,
GLuint *max_index,
GLuint nr_prims)
{
GLuint tmp_min, tmp_max;
GLuint i;
GLuint count;
*min_index = ~0;
*max_index = 0;
for (i = 0; i < nr_prims; i++) {
const struct _mesa_prim *start_prim;
start_prim = &prims[i];
count = start_prim->count;
/* Do combination if possible to reduce map/unmap count */
while ((i + 1 < nr_prims) &&
(prims[i].start + prims[i].count == prims[i+1].start)) {
count += prims[i+1].count;
i++;
}
vbo_get_minmax_index(ctx, start_prim, ib, &tmp_min, &tmp_max, count);
*min_index = MIN2(*min_index, tmp_min);
*max_index = MAX2(*max_index, tmp_max);
}
}
|