falconia has uploaded this change for review.
twrtp: fix generation of loss info word in RTCP RR blocks
We generate RTCP SR and RR packets that include reception report blocks
describing what our twjit received - see RFC 3550 section 6.4.1. These
RR blocks include a 32-bit word that reports our receiver's detection
of packet loss in the incoming RTP stream; this word consists of an
8-bit "fraction lost" field and a 24-bit "cumulative number of packets
lost" field. Our previous code computed each of these separate fields
correctly, but the final step of combining them into the 32-bit word
applied the wrong shift, producing garbage as a result. Fix it.
Change-Id: Ib34239a4f4619fef097b4c1fb5fc3abed339850d
---
M src/twrtp.c
1 file changed, 1 insertion(+), 1 deletion(-)
git pull ssh://gerrit.osmocom.org:29418/libosmo-netif refs/changes/27/44027/1
diff --git a/src/twrtp.c b/src/twrtp.c
index 1a1b7be..4d81510 100644
--- a/src/twrtp.c
+++ b/src/twrtp.c
@@ -851,7 +851,7 @@
lost_fract = 0;
else
lost_fract = (newly_lost << 8) / delta_expect;
- lost_word = (lost_fract << 8) | (cumulative_lost & 0xFFFFFF);
+ lost_word = (lost_fract << 24) | (cumulative_lost & 0xFFFFFF);
rr->ssrc = htonl(rri->ssrc);
rr->lost_word = htonl(lost_word);
To view, visit change 44027. To unsubscribe, or for help writing mail filters, visit settings.