RepoMirrors
/
mpv
mirror of https://github.com/mpv-player/mpv
1
0
Fork 0
Code Issues Releases Wiki Activity

filter: don't randomly lose async wakeup notifications 

Another "what was I thinking" thing - destroying filters explicitly
skipped async wakeups for no reason. These were notifications for
filters that are not going to be destroyed too, and so their wakeup will
be lost, leading to stalled playback. This is completely unnecessary and
the special code can be removed.

Fixes #5488. (This case destroyed all audio filters due to AO init
failure, which could make clear out the f_demux_in.c wakeup for video,
and "freeze" playback.)
This commit is contained in:
wm4 2018-02-04 20:00:32 +01:00 committed by Kevin Mitchell
parent beb8d27912
commit 2f6dc93276
1 changed files with 4 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
filters/filter.c
View File

@ -165,13 +165,12 @@ void mp_filter_internal_mark_progress(struct mp_filter *f)
// Basically copy the async notifications to the sync ones. Done so that the
// sync notifications don't need any locking.
static void flush_async_notifications(struct filter_runner *r, bool queue)
static void flush_async_notifications(struct filter_runner *r)
{
pthread_mutex_lock(&r->async_lock);
for (int n = 0; n < r->num_async_pending; n++) {
struct mp_filter *f = r->async_pending[n];
if (queue)
add_pending(f);
add_pending(f);
f->in->async_pending = false;
}
r->num_async_pending = 0;
@ -185,7 +184,7 @@ bool mp_filter_run(struct mp_filter *filter)
r->filtering = true;
flush_async_notifications(r, true);
flush_async_notifications(r);
while (r->num_pending) {
struct mp_filter *next = r->pending[r->num_pending - 1];
@ -653,7 +652,7 @@ static void filter_destructor(void *p)
// Just make sure the filter is not still in the async notifications set.
// There will be no more new notifications at this point (due to destroy()).
flush_async_notifications(r, false);
flush_async_notifications(r);
for (int n = 0; n < r->num_pending; n++) {
if (r->pending[n] == f) {