gallium/util: add util_vertex_state_cache for deduplicating the states

Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13050>
This commit is contained in:
Marek Olšák
2021-09-15 22:51:49 -04:00
committed by Marge Bot
parent e8cad57aa7
commit 7a2458f4bd
3 changed files with 203 additions and 0 deletions
+2
View File
@@ -321,6 +321,8 @@ files_libgallium = files(
'util/u_upload_mgr.h',
'util/u_vbuf.c',
'util/u_vbuf.h',
'util/u_vertex_state_cache.c',
'util/u_vertex_state_cache.h',
'util/u_video.h',
'util/u_viewport.h',
'nir/tgsi_to_nir.c',
@@ -0,0 +1,134 @@
/*
* Copyright 2021 Advanced Micro Devices, Inc.
* All Rights Reserved.
*
* 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
* on the rights to use, copy, modify, merge, publish, distribute, sub
* license, 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 "util/u_vertex_state_cache.h"
#include "util/u_inlines.h"
#include "util/hash_table.h"
#include "util/set.h"
static uint32_t key_hash(const void *key)
{
const struct pipe_vertex_state *state = key;
return _mesa_hash_data(&state->input, sizeof(state->input));
}
static bool key_equals(const void *a, const void *b)
{
const struct pipe_vertex_state *sa = a;
const struct pipe_vertex_state *sb = b;
return !memcmp(&sa->input, &sb->input, sizeof(sa->input));
}
void
util_vertex_state_cache_init(struct util_vertex_state_cache *cache,
pipe_create_vertex_state_func create,
pipe_vertex_state_destroy_func destroy)
{
simple_mtx_init(&cache->lock, mtx_plain);
cache->set = _mesa_set_create(NULL, key_hash, key_equals);
cache->create = create;
cache->destroy = destroy;
}
void
util_vertex_state_cache_deinit(struct util_vertex_state_cache *cache)
{
if (cache->set) {
set_foreach(cache->set, entry) {
fprintf(stderr, "mesa: vertex state cache should be empty\n");
assert(!"vertex state cache should be empty");
}
_mesa_set_destroy(cache->set, NULL);
simple_mtx_destroy(&cache->lock);
}
}
struct pipe_vertex_state *
util_vertex_state_cache_get(struct pipe_screen *screen,
struct pipe_vertex_buffer *buffer,
const struct pipe_vertex_element *elements,
unsigned num_elements,
struct pipe_resource *indexbuf,
uint32_t full_velem_mask,
struct util_vertex_state_cache *cache)
{
struct pipe_vertex_state key;
memset(&key, 0, sizeof(key));
key.input.indexbuf = indexbuf;
key.input.vbuffer.stride = buffer->stride;
assert(!buffer->is_user_buffer);
key.input.vbuffer.buffer_offset = buffer->buffer_offset;
key.input.vbuffer.buffer = buffer->buffer;
key.input.num_elements = num_elements;
for (unsigned i = 0; i < num_elements; i++)
key.input.elements[i] = elements[i];
key.input.full_velem_mask = full_velem_mask;
uint32_t hash = key_hash(&key);
/* Find the state in the live cache. */
simple_mtx_lock(&cache->lock);
struct set_entry *entry = _mesa_set_search_pre_hashed(cache->set, hash, &key);
struct pipe_vertex_state *state = entry ? (void*)entry->key : NULL;
/* Return if the state already exists. */
if (state) {
/* Increase the refcount. */
p_atomic_inc(&state->reference.count);
assert(state->reference.count >= 1);
simple_mtx_unlock(&cache->lock);
return state;
}
state = cache->create(screen, buffer, elements, num_elements, indexbuf,
full_velem_mask);
if (state) {
assert(key_hash(state) == hash);
_mesa_set_add_pre_hashed(cache->set, hash, state);
}
simple_mtx_unlock(&cache->lock);
return state;
}
void
util_vertex_state_destroy(struct pipe_screen *screen,
struct util_vertex_state_cache *cache,
struct pipe_vertex_state *state)
{
simple_mtx_lock(&cache->lock);
/* There could have been a thread race and the cache might have returned
* the vertex state being destroyed. Check the reference count and do
* nothing if it's positive.
*/
if (p_atomic_read(&state->reference.count) <= 0) {
_mesa_set_remove_key(cache->set, state);
cache->destroy(screen, state);
}
simple_mtx_unlock(&cache->lock);
}
@@ -0,0 +1,67 @@
/*
* Copyright 2021 Advanced Micro Devices, Inc.
* All Rights Reserved.
*
* 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
* on the rights to use, copy, modify, merge, publish, distribute, sub
* license, 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.
*/
/* This deduplicates pipe_vertex_state CSOs to enable draw merging in
* u_threaded_context because the draw merging is possible only if different
* display lists use the same pipe_vertex_state CSO.
*/
#ifndef U_VERTEX_STATE_CACHE_H
#define U_VERTEX_STATE_CACHE_H
#include "util/simple_mtx.h"
#include "pipe/p_screen.h"
#include "pipe/p_state.h"
struct util_vertex_state_cache {
simple_mtx_t lock;
struct set *set;
pipe_create_vertex_state_func create;
pipe_vertex_state_destroy_func destroy;
};
void
util_vertex_state_cache_init(struct util_vertex_state_cache *cache,
pipe_create_vertex_state_func create,
pipe_vertex_state_destroy_func destroy);
void
util_vertex_state_cache_deinit(struct util_vertex_state_cache *cache);
struct pipe_vertex_state *
util_vertex_state_cache_get(struct pipe_screen *screen,
struct pipe_vertex_buffer *buffer,
const struct pipe_vertex_element *elements,
unsigned num_elements,
struct pipe_resource *indexbuf,
uint32_t full_velem_mask,
struct util_vertex_state_cache *cache);
void
util_vertex_state_destroy(struct pipe_screen *screen,
struct util_vertex_state_cache *cache,
struct pipe_vertex_state *state);
#endif