Commit ab49dfd0 authored by Stian Selnes's avatar Stian Selnes Committed by Olivier Crête
Browse files

rtpjitterbuffer: Fix lost duration when gap after lost timer

This patch fixes an issue with the estimated gap duration when there is
a gap immediately after a lost timer has been processed. Previously
there was a discrepancy beteen the gap in seqnum and gap in dts which
would cause wrong calculated duration. The issue would only be seen with
retranmission enabled since when it's disabled lost timers are only
created when a packet is received and the actual gap length and last dts
is known.

https://bugzilla.gnome.org/show_bug.cgi?id=769768
parent dd020f5c
......@@ -3678,8 +3678,10 @@ do_lost_timeout (GstRtpJitterBuffer * jitterbuffer, TimerData * timer,
next_in_seqnum = (seqnum + lost_packets) & 0xffff;
/* we now only accept seqnum bigger than this */
if (gst_rtp_buffer_compare_seqnum (priv->next_in_seqnum, next_in_seqnum) > 0)
if (gst_rtp_buffer_compare_seqnum (priv->next_in_seqnum, next_in_seqnum) > 0) {
priv->next_in_seqnum = next_in_seqnum;
priv->last_in_dts = apply_offset (jitterbuffer, timer->timeout);
}
/* Avoid creating events if we don't need it. Note that we still need to create
* the lost *ITEM* since it will be used to notify the outgoing thread of
......
......@@ -1875,8 +1875,6 @@ GST_START_TEST (test_gap_exceeds_latency)
GstTestClock *testclock;
const gint jb_latency_ms = 200;
guint32 timestamp_ms = 0;
guint32 rtp_ts = 0;
gint i;
GstEvent *out_event;
GstBuffer *out_buf;
......@@ -1893,10 +1891,7 @@ GST_START_TEST (test_gap_exceeds_latency)
fail_unless_equals_int (GST_FLOW_OK,
gst_harness_push (h, generate_test_buffer (0)));
timestamp_ms += 20;
rtp_ts += PCMU_RTP_TS_DURATION;
gst_harness_set_time (h, timestamp_ms * GST_MSECOND);
gst_harness_set_time (h, 1 * PCMU_BUF_DURATION);
fail_unless_equals_int (GST_FLOW_OK,
gst_harness_push (h, generate_test_buffer (1)));
......@@ -1931,12 +1926,23 @@ GST_START_TEST (test_gap_exceeds_latency)
fail_unless_equals_int (GST_FLOW_OK,
gst_harness_push (h, generate_test_buffer (16)));
/* FIXME: something is up with the timestamp here!!?! */
/* Manually check the first rtx event */
out_event = gst_harness_pull_upstream_event (h);
verify_rtx_event (out_event, 6, 119999994, 10, PCMU_BUF_DURATION);
/* lost more rtx with weird timestamps... */
verify_rtx_event (out_event, 6, 6 * PCMU_BUF_DURATION, 10, PCMU_BUF_DURATION);
/* Go throught the rest of rtx events. A bit more relaxed since order is
* partly an implentation detail. */
for (i = 0; i < 12; i++) {
gst_event_unref (gst_harness_pull_upstream_event (h));
const GstStructure *s;
guint seqnum, retry;
fail_unless (out_event = gst_harness_pull_upstream_event (h));
fail_unless (s = gst_event_get_structure (out_event));
fail_unless (gst_structure_get_uint (s, "seqnum", &seqnum));
fail_unless (gst_structure_get_uint (s, "retry", &retry));
fail_unless (seqnum >= 6 && seqnum <= 12);
verify_rtx_event (out_event, seqnum, seqnum * PCMU_BUF_DURATION,
10 + retry * 40, PCMU_BUF_DURATION);
}
fail_unless_equals_int (0, gst_harness_upstream_events_in_queue (h));
......@@ -1948,19 +1954,20 @@ GST_START_TEST (test_gap_exceeds_latency)
gst_harness_push (h, generate_test_buffer (i)));
}
/* FIXME: wtf is going on with timestamps and durations here??!? */
gst_harness_crank_single_clock_wait (h);
out_event = gst_harness_pull_event (h);
verify_lost_event (out_event, 3, 41428571, 78571423);
verify_lost_event (out_event, 3, 3 * PCMU_BUF_DURATION,
3 * PCMU_BUF_DURATION);
/* FIXME: and these rtx... */
gst_harness_crank_single_clock_wait (h);
out_event = gst_harness_pull_event (h);
verify_lost_event (out_event, 6, 119999994, 21428571);
verify_lost_event (out_event, 6, 6 * PCMU_BUF_DURATION,
1 * PCMU_BUF_DURATION);
gst_harness_crank_single_clock_wait (h);
out_event = gst_harness_pull_event (h);
verify_lost_event (out_event, 7, 141428565, 21428571);
verify_lost_event (out_event, 7, 7 * PCMU_BUF_DURATION,
1 * PCMU_BUF_DURATION);
/* 8 */
for (i = 8; i <= 16; i++) {
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment