summaryrefslogtreecommitdiffstats
path: root/video
diff options
context:
space:
mode:
authorwm4 <wm4@nowhere>2018-01-09 09:18:18 +0100
committerMartin Herkt <652892+lachs0r@users.noreply.github.com>2018-01-09 09:19:56 +0100
commit0a406f97e09ae81ac376fb8438f92bf50785c4cf (patch)
treeab8cf5539957e5c594d546ea4585cb18d3508910 /video
parent6e974f77bdb9f194d4c2c0c1dc36703419dbd0df (diff)
downloadmpv-0a406f97e09ae81ac376fb8438f92bf50785c4cf.tar.bz2
mpv-0a406f97e09ae81ac376fb8438f92bf50785c4cf.tar.xz
video, audio: don't actively wait for demuxer input
If feed_packet() ended with DATA_WAIT, the player should have gone to sleep, until the demuxer wakes it up again when there is new data. But the call to read_frame() unconditionally overwrote this status code, so it never waited. The consequence was that the core burned CPU by effectively polling the demuxer status, which was noticeable especially when seeking in network streams (since seeking is async, decoders will start out with having to wait for network). Regression since commit 33e5755c.
Diffstat (limited to 'video')
-rw-r--r--video/decode/dec_video.c2
1 files changed, 2 insertions, 0 deletions
diff --git a/video/decode/dec_video.c b/video/decode/dec_video.c
index 243c38a0a8..6eaed8a85a 100644
--- a/video/decode/dec_video.c
+++ b/video/decode/dec_video.c
@@ -504,6 +504,8 @@ void video_work(struct dec_video *d_video)
read_frame(d_video);
if (!d_video->current_mpi) {
feed_packet(d_video);
+ if (d_video->current_state == DATA_WAIT)
+ return;
read_frame(d_video); // retry, to avoid redundant iterations
}
}