ref: 3b14f94924987fc66302bb8b803ad9065816f93e
dir: dav1d/src/lib.c
/*
* Copyright © 2018, VideoLAN and dav1d authors
* Copyright © 2018, Two Orioles, LLC
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice, this
* list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
* WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
* DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
* ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include "config.h"
#include "vcs_version.h"
#include <errno.h>
#include <string.h>
#if defined(__linux__) && defined(HAVE_DLSYM)
#include <dlfcn.h>
#endif
#include "dav1d/dav1d.h"
#include "dav1d/data.h"
#include "common/validate.h"
#include "src/cpu.h"
#include "src/fg_apply.h"
#include "src/internal.h"
#include "src/log.h"
#include "src/obu.h"
#include "src/qm.h"
#include "src/ref.h"
#include "src/thread_task.h"
#include "src/wedge.h"
static COLD void init_internal(void) {
dav1d_init_cpu();
dav1d_init_interintra_masks();
dav1d_init_qm_tables();
dav1d_init_thread();
dav1d_init_wedge_masks();
}
COLD const char *dav1d_version(void) {
return DAV1D_VERSION;
}
COLD void dav1d_default_settings(Dav1dSettings *s) {
s->n_frame_threads = 1;
s->n_tile_threads = 1;
s->apply_grain = 1;
s->allocator.cookie = NULL;
s->allocator.alloc_picture_callback = dav1d_default_picture_alloc;
s->allocator.release_picture_callback = dav1d_default_picture_release;
s->logger.cookie = NULL;
s->logger.callback = dav1d_log_default_callback;
s->operating_point = 0;
s->all_layers = 1; // just until the tests are adjusted
s->frame_size_limit = 0;
}
static void close_internal(Dav1dContext **const c_out, int flush);
NO_SANITIZE("cfi-icall") // CFI is broken with dlsym()
static COLD size_t get_stack_size_internal(const pthread_attr_t *const thread_attr) {
#if defined(__linux__) && defined(HAVE_DLSYM) && defined(__GLIBC__)
/* glibc has an issue where the size of the TLS is subtracted from the stack
* size instead of allocated separately. As a result the specified stack
* size may be insufficient when used in an application with large amounts
* of TLS data. The following is a workaround to compensate for that.
* See https://sourceware.org/bugzilla/show_bug.cgi?id=11787 */
size_t (*const get_minstack)(const pthread_attr_t*) =
dlsym(RTLD_DEFAULT, "__pthread_get_minstack");
if (get_minstack)
return get_minstack(thread_attr) - PTHREAD_STACK_MIN;
#endif
return 0;
}
COLD int dav1d_open(Dav1dContext **c_out, const Dav1dSettings *s) {
static pthread_once_t initted = PTHREAD_ONCE_INIT;
pthread_once(&initted, init_internal);
validate_input_or_ret(c_out != NULL, DAV1D_ERR(EINVAL));
validate_input_or_ret(s != NULL, DAV1D_ERR(EINVAL));
validate_input_or_ret(s->n_tile_threads >= 1 &&
s->n_tile_threads <= DAV1D_MAX_TILE_THREADS, DAV1D_ERR(EINVAL));
validate_input_or_ret(s->n_frame_threads >= 1 &&
s->n_frame_threads <= DAV1D_MAX_FRAME_THREADS, DAV1D_ERR(EINVAL));
validate_input_or_ret(s->allocator.alloc_picture_callback != NULL,
DAV1D_ERR(EINVAL));
validate_input_or_ret(s->allocator.release_picture_callback != NULL,
DAV1D_ERR(EINVAL));
validate_input_or_ret(s->operating_point >= 0 &&
s->operating_point <= 31, DAV1D_ERR(EINVAL));
pthread_attr_t thread_attr;
if (pthread_attr_init(&thread_attr)) return DAV1D_ERR(ENOMEM);
size_t stack_size = 1024 * 1024 + get_stack_size_internal(&thread_attr);
pthread_attr_setstacksize(&thread_attr, stack_size);
Dav1dContext *const c = *c_out = dav1d_alloc_aligned(sizeof(*c), 32);
if (!c) goto error;
memset(c, 0, sizeof(*c));
c->allocator = s->allocator;
c->logger = s->logger;
c->apply_grain = s->apply_grain;
c->operating_point = s->operating_point;
c->all_layers = s->all_layers;
c->frame_size_limit = s->frame_size_limit;
if (dav1d_mem_pool_init(&c->seq_hdr_pool) ||
dav1d_mem_pool_init(&c->frame_hdr_pool) ||
dav1d_mem_pool_init(&c->segmap_pool) ||
dav1d_mem_pool_init(&c->refmvs_pool) ||
dav1d_mem_pool_init(&c->cdf_pool))
{
goto error;
}
if (c->allocator.alloc_picture_callback == dav1d_default_picture_alloc) {
if (dav1d_mem_pool_init(&c->picture_pool)) goto error;
c->allocator.cookie = c->picture_pool;
}
/* On 32-bit systems extremely large frame sizes can cause overflows in
* dav1d_decode_frame() malloc size calculations. Prevent that from occuring
* by enforcing a maximum frame size limit, chosen to roughly correspond to
* the largest size possible to decode without exhausting virtual memory. */
if (sizeof(size_t) < 8 && s->frame_size_limit - 1 >= 8192 * 8192) {
c->frame_size_limit = 8192 * 8192;
if (s->frame_size_limit)
dav1d_log(c, "Frame size limit reduced from %u to %u.\n",
s->frame_size_limit, c->frame_size_limit);
}
c->frame_thread.flush = &c->frame_thread.flush_mem;
atomic_init(c->frame_thread.flush, 0);
c->n_fc = s->n_frame_threads;
c->fc = dav1d_alloc_aligned(sizeof(*c->fc) * s->n_frame_threads, 32);
if (!c->fc) goto error;
memset(c->fc, 0, sizeof(*c->fc) * s->n_frame_threads);
if (c->n_fc > 1) {
c->frame_thread.out_delayed =
calloc(c->n_fc, sizeof(*c->frame_thread.out_delayed));
if (!c->frame_thread.out_delayed) goto error;
}
for (int n = 0; n < s->n_frame_threads; n++) {
Dav1dFrameContext *const f = &c->fc[n];
f->c = c;
f->lf.last_sharpness = -1;
f->n_tc = s->n_tile_threads;
f->tc = dav1d_alloc_aligned(sizeof(*f->tc) * s->n_tile_threads, 64);
if (!f->tc) goto error;
memset(f->tc, 0, sizeof(*f->tc) * s->n_tile_threads);
if (f->n_tc > 1) {
if (pthread_mutex_init(&f->tile_thread.lock, NULL)) goto error;
if (pthread_cond_init(&f->tile_thread.cond, NULL)) {
pthread_mutex_destroy(&f->tile_thread.lock);
goto error;
}
if (pthread_cond_init(&f->tile_thread.icond, NULL)) {
pthread_mutex_destroy(&f->tile_thread.lock);
pthread_cond_destroy(&f->tile_thread.cond);
goto error;
}
f->tile_thread.inited = 1;
}
for (int m = 0; m < s->n_tile_threads; m++) {
Dav1dTileContext *const t = &f->tc[m];
t->f = f;
memset(t->cf_16bpc, 0, sizeof(t->cf_16bpc));
if (f->n_tc > 1) {
if (pthread_mutex_init(&t->tile_thread.td.lock, NULL)) goto error;
if (pthread_cond_init(&t->tile_thread.td.cond, NULL)) {
pthread_mutex_destroy(&t->tile_thread.td.lock);
goto error;
}
t->tile_thread.fttd = &f->tile_thread;
if (pthread_create(&t->tile_thread.td.thread, &thread_attr, dav1d_tile_task, t)) {
pthread_cond_destroy(&t->tile_thread.td.cond);
pthread_mutex_destroy(&t->tile_thread.td.lock);
goto error;
}
t->tile_thread.td.inited = 1;
}
}
dav1d_refmvs_init(&f->rf);
if (c->n_fc > 1) {
if (pthread_mutex_init(&f->frame_thread.td.lock, NULL)) goto error;
if (pthread_cond_init(&f->frame_thread.td.cond, NULL)) {
pthread_mutex_destroy(&f->frame_thread.td.lock);
goto error;
}
if (pthread_create(&f->frame_thread.td.thread, &thread_attr, dav1d_frame_task, f)) {
pthread_cond_destroy(&f->frame_thread.td.cond);
pthread_mutex_destroy(&f->frame_thread.td.lock);
goto error;
}
f->frame_thread.td.inited = 1;
}
}
// intra edge tree
c->intra_edge.root[BL_128X128] = &c->intra_edge.branch_sb128[0].node;
dav1d_init_mode_tree(c->intra_edge.root[BL_128X128], c->intra_edge.tip_sb128, 1);
c->intra_edge.root[BL_64X64] = &c->intra_edge.branch_sb64[0].node;
dav1d_init_mode_tree(c->intra_edge.root[BL_64X64], c->intra_edge.tip_sb64, 0);
pthread_attr_destroy(&thread_attr);
return 0;
error:
if (c) close_internal(c_out, 0);
pthread_attr_destroy(&thread_attr);
return DAV1D_ERR(ENOMEM);
}
static void dummy_free(const uint8_t *const data, void *const user_data) {
assert(data && !user_data);
}
int dav1d_parse_sequence_header(Dav1dSequenceHeader *const out,
const uint8_t *const ptr, const size_t sz)
{
Dav1dData buf = { 0 };
int res;
validate_input_or_ret(out != NULL, DAV1D_ERR(EINVAL));
Dav1dSettings s;
dav1d_default_settings(&s);
s.logger.callback = NULL;
Dav1dContext *c;
res = dav1d_open(&c, &s);
if (res < 0) return res;
if (ptr) {
res = dav1d_data_wrap_internal(&buf, ptr, sz, dummy_free, NULL);
if (res < 0) goto error;
}
while (buf.sz > 0) {
res = dav1d_parse_obus(c, &buf, 1);
if (res < 0) goto error;
assert((size_t)res <= buf.sz);
buf.sz -= res;
buf.data += res;
}
if (!c->seq_hdr) {
res = DAV1D_ERR(EINVAL);
goto error;
}
memcpy(out, c->seq_hdr, sizeof(*out));
res = 0;
error:
dav1d_data_unref_internal(&buf);
dav1d_close(&c);
return res;
}
static int output_image(Dav1dContext *const c, Dav1dPicture *const out,
Dav1dPicture *const in)
{
const Dav1dFilmGrainData *fgdata = &in->frame_hdr->film_grain.data;
int has_grain = fgdata->num_y_points || fgdata->num_uv_points[0] ||
fgdata->num_uv_points[1];
// If there is nothing to be done, skip the allocation/copy
if (!c->apply_grain || !has_grain) {
dav1d_picture_move_ref(out, in);
return 0;
}
// Apply film grain to a new copy of the image to avoid corrupting refs
int res = dav1d_picture_alloc_copy(c, out, in->p.w, in);
if (res < 0) {
dav1d_picture_unref_internal(in);
dav1d_picture_unref_internal(out);
return res;
}
switch (out->p.bpc) {
#if CONFIG_8BPC
case 8:
dav1d_apply_grain_8bpc(&c->dsp[0].fg, out, in);
break;
#endif
#if CONFIG_16BPC
case 10:
case 12:
dav1d_apply_grain_16bpc(&c->dsp[(out->p.bpc >> 1) - 4].fg, out, in);
break;
#endif
default:
assert(0);
}
dav1d_picture_unref_internal(in);
return 0;
}
static int output_picture_ready(Dav1dContext *const c) {
if (!c->out.data[0]) return 0;
// skip lower spatial layers
if (c->operating_point_idc && !c->all_layers) {
const int max_spatial_id = ulog2(c->operating_point_idc >> 8);
if (max_spatial_id > c->out.frame_hdr->spatial_id) {
dav1d_picture_unref_internal(&c->out);
return 0;
}
}
return 1;
}
static int drain_picture(Dav1dContext *const c, Dav1dPicture *const out) {
unsigned drain_count = 0;
do {
const unsigned next = c->frame_thread.next;
Dav1dFrameContext *const f = &c->fc[next];
pthread_mutex_lock(&f->frame_thread.td.lock);
while (f->n_tile_data > 0)
pthread_cond_wait(&f->frame_thread.td.cond,
&f->frame_thread.td.lock);
pthread_mutex_unlock(&f->frame_thread.td.lock);
Dav1dThreadPicture *const out_delayed =
&c->frame_thread.out_delayed[next];
if (++c->frame_thread.next == c->n_fc)
c->frame_thread.next = 0;
if (out_delayed->p.data[0]) {
const unsigned progress =
atomic_load_explicit(&out_delayed->progress[1],
memory_order_relaxed);
if (out_delayed->visible && progress != FRAME_ERROR)
dav1d_picture_ref(&c->out, &out_delayed->p);
dav1d_thread_picture_unref(out_delayed);
if (output_picture_ready(c))
return output_image(c, out, &c->out);
}
} while (++drain_count < c->n_fc);
return DAV1D_ERR(EAGAIN);
}
static int gen_picture(Dav1dContext *const c)
{
int res;
Dav1dData *const in = &c->in;
if (output_picture_ready(c))
return 0;
while (in->sz > 0) {
res = dav1d_parse_obus(c, in, 0);
if (res < 0) {
dav1d_data_unref_internal(in);
} else {
assert((size_t)res <= in->sz);
in->sz -= res;
in->data += res;
if (!in->sz) dav1d_data_unref_internal(in);
}
if (output_picture_ready(c))
break;
if (res < 0)
return res;
}
return 0;
}
int dav1d_send_data(Dav1dContext *const c, Dav1dData *const in)
{
validate_input_or_ret(c != NULL, DAV1D_ERR(EINVAL));
validate_input_or_ret(in != NULL, DAV1D_ERR(EINVAL));
validate_input_or_ret(in->data == NULL || in->sz, DAV1D_ERR(EINVAL));
if (in->data)
c->drain = 0;
if (c->in.data)
return DAV1D_ERR(EAGAIN);
dav1d_data_ref(&c->in, in);
int res = gen_picture(c);
if (!res)
dav1d_data_unref_internal(in);
return res;
}
int dav1d_get_picture(Dav1dContext *const c, Dav1dPicture *const out)
{
validate_input_or_ret(c != NULL, DAV1D_ERR(EINVAL));
validate_input_or_ret(out != NULL, DAV1D_ERR(EINVAL));
const int drain = c->drain;
c->drain = 1;
int res = gen_picture(c);
if (res < 0)
return res;
if (output_picture_ready(c))
return output_image(c, out, &c->out);
if (c->n_fc > 1 && drain)
return drain_picture(c, out);
return DAV1D_ERR(EAGAIN);
}
void dav1d_flush(Dav1dContext *const c) {
dav1d_data_unref_internal(&c->in);
c->drain = 0;
for (int i = 0; i < 8; i++) {
if (c->refs[i].p.p.data[0])
dav1d_thread_picture_unref(&c->refs[i].p);
dav1d_ref_dec(&c->refs[i].segmap);
dav1d_ref_dec(&c->refs[i].refmvs);
dav1d_cdf_thread_unref(&c->cdf[i]);
}
c->frame_hdr = NULL;
c->seq_hdr = NULL;
dav1d_ref_dec(&c->seq_hdr_ref);
c->mastering_display = NULL;
c->content_light = NULL;
c->itut_t35 = NULL;
dav1d_ref_dec(&c->mastering_display_ref);
dav1d_ref_dec(&c->content_light_ref);
dav1d_ref_dec(&c->itut_t35_ref);
if (c->n_fc == 1) return;
// mark each currently-running frame as flushing, so that we
// exit out as quickly as the running thread checks this flag
atomic_store(c->frame_thread.flush, 1);
for (unsigned n = 0, next = c->frame_thread.next; n < c->n_fc; n++, next++) {
if (next == c->n_fc) next = 0;
Dav1dFrameContext *const f = &c->fc[next];
pthread_mutex_lock(&f->frame_thread.td.lock);
if (f->n_tile_data > 0) {
while (f->n_tile_data > 0)
pthread_cond_wait(&f->frame_thread.td.cond,
&f->frame_thread.td.lock);
assert(!f->cur.data[0]);
}
pthread_mutex_unlock(&f->frame_thread.td.lock);
Dav1dThreadPicture *const out_delayed = &c->frame_thread.out_delayed[next];
if (out_delayed->p.data[0])
dav1d_thread_picture_unref(out_delayed);
}
atomic_store(c->frame_thread.flush, 0);
c->frame_thread.next = 0;
}
COLD void dav1d_close(Dav1dContext **const c_out) {
validate_input(c_out != NULL);
close_internal(c_out, 1);
}
static COLD void close_internal(Dav1dContext **const c_out, int flush) {
Dav1dContext *const c = *c_out;
if (!c) return;
if (flush) dav1d_flush(c);
for (unsigned n = 0; c->fc && n < c->n_fc; n++) {
Dav1dFrameContext *const f = &c->fc[n];
// clean-up threading stuff
if (c->n_fc > 1 && f->frame_thread.td.inited) {
pthread_mutex_lock(&f->frame_thread.td.lock);
f->frame_thread.die = 1;
pthread_cond_signal(&f->frame_thread.td.cond);
pthread_mutex_unlock(&f->frame_thread.td.lock);
pthread_join(f->frame_thread.td.thread, NULL);
freep(&f->frame_thread.b);
dav1d_freep_aligned(&f->frame_thread.pal_idx);
dav1d_freep_aligned(&f->frame_thread.cf);
freep(&f->frame_thread.tile_start_off);
dav1d_freep_aligned(&f->frame_thread.pal);
freep(&f->frame_thread.cbi);
pthread_mutex_destroy(&f->frame_thread.td.lock);
pthread_cond_destroy(&f->frame_thread.td.cond);
}
if (f->n_tc > 1 && f->tc && f->tile_thread.inited) {
pthread_mutex_lock(&f->tile_thread.lock);
for (int m = 0; m < f->n_tc; m++) {
Dav1dTileContext *const t = &f->tc[m];
t->tile_thread.die = 1;
// mark not created tile threads as available
if (!t->tile_thread.td.inited)
f->tile_thread.available |= 1ULL<<m;
}
pthread_cond_broadcast(&f->tile_thread.cond);
while (f->tile_thread.available != ~0ULL >> (64 - f->n_tc))
pthread_cond_wait(&f->tile_thread.icond,
&f->tile_thread.lock);
pthread_mutex_unlock(&f->tile_thread.lock);
for (int m = 0; m < f->n_tc; m++) {
Dav1dTileContext *const t = &f->tc[m];
if (f->n_tc > 1 && t->tile_thread.td.inited) {
pthread_join(t->tile_thread.td.thread, NULL);
pthread_mutex_destroy(&t->tile_thread.td.lock);
pthread_cond_destroy(&t->tile_thread.td.cond);
}
}
pthread_mutex_destroy(&f->tile_thread.lock);
pthread_cond_destroy(&f->tile_thread.cond);
pthread_cond_destroy(&f->tile_thread.icond);
freep(&f->tile_thread.task_idx_to_sby_and_tile_idx);
}
for (int m = 0; f->ts && m < f->n_ts; m++) {
Dav1dTileState *const ts = &f->ts[m];
pthread_cond_destroy(&ts->tile_thread.cond);
pthread_mutex_destroy(&ts->tile_thread.lock);
}
dav1d_free_aligned(f->ts);
dav1d_free_aligned(f->tc);
dav1d_free_aligned(f->ipred_edge[0]);
free(f->a);
free(f->tile);
free(f->lf.mask);
free(f->lf.lr_mask);
free(f->lf.level);
free(f->lf.tx_lpf_right_edge[0]);
dav1d_refmvs_clear(&f->rf);
dav1d_free_aligned(f->lf.cdef_line_buf);
dav1d_free_aligned(f->lf.lr_lpf_line[0]);
}
dav1d_free_aligned(c->fc);
dav1d_data_unref_internal(&c->in);
if (c->n_fc > 1 && c->frame_thread.out_delayed) {
for (unsigned n = 0; n < c->n_fc; n++)
if (c->frame_thread.out_delayed[n].p.data[0])
dav1d_thread_picture_unref(&c->frame_thread.out_delayed[n]);
free(c->frame_thread.out_delayed);
}
for (int n = 0; n < c->n_tile_data; n++)
dav1d_data_unref_internal(&c->tile[n].data);
free(c->tile);
for (int n = 0; n < 8; n++) {
dav1d_cdf_thread_unref(&c->cdf[n]);
if (c->refs[n].p.p.data[0])
dav1d_thread_picture_unref(&c->refs[n].p);
dav1d_ref_dec(&c->refs[n].refmvs);
dav1d_ref_dec(&c->refs[n].segmap);
}
dav1d_ref_dec(&c->seq_hdr_ref);
dav1d_ref_dec(&c->frame_hdr_ref);
dav1d_ref_dec(&c->mastering_display_ref);
dav1d_ref_dec(&c->content_light_ref);
dav1d_ref_dec(&c->itut_t35_ref);
dav1d_mem_pool_end(c->seq_hdr_pool);
dav1d_mem_pool_end(c->frame_hdr_pool);
dav1d_mem_pool_end(c->segmap_pool);
dav1d_mem_pool_end(c->refmvs_pool);
dav1d_mem_pool_end(c->cdf_pool);
dav1d_mem_pool_end(c->picture_pool);
dav1d_freep_aligned(c_out);
}
void dav1d_picture_unref(Dav1dPicture *const p) {
dav1d_picture_unref_internal(p);
}
uint8_t *dav1d_data_create(Dav1dData *const buf, const size_t sz) {
return dav1d_data_create_internal(buf, sz);
}
int dav1d_data_wrap(Dav1dData *const buf, const uint8_t *const ptr,
const size_t sz,
void (*const free_callback)(const uint8_t *data,
void *user_data),
void *const user_data)
{
return dav1d_data_wrap_internal(buf, ptr, sz, free_callback, user_data);
}
int dav1d_data_wrap_user_data(Dav1dData *buf,
const uint8_t *const user_data,
void (*const free_callback)(const uint8_t *user_data,
void *cookie),
void *const cookie)
{
return dav1d_data_wrap_user_data_internal(buf,
user_data,
free_callback,
cookie);
}
void dav1d_data_unref(Dav1dData *buf) {
dav1d_data_unref_internal(buf);
}