<html><head><meta http-equiv="Content-Type" content="text/html; charset=us-ascii"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">If you understand the conditions that trigger this, a unit test that reproduces it would be very helpful as well. We look forward to see your pull requests on this issue. Thanks for all your hard work and diligence. It is a weird issue that I dont believe I have ever seen in the wild, I will be interested to better understand what is different about your environment that triggers its and what sort of device you are interloping with that causes this issue.<br class=""><div><br class=""><blockquote type="cite" class=""><div class="">On Apr 12, 2022, at 9:05 AM, Dragos Oancea <<a href="mailto:dragos@freeswitch.org" class="">dragos@freeswitch.org</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div dir="ltr" class="">You could open a github issue.</div><br class=""><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Apr 8, 2022 at 1:40 PM Andy Newlands <<a href="mailto:andynewlands@gmail.com" class="">andynewlands@gmail.com</a>> wrote:<br class=""></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr" class="">Further to my earlier quest, and after a lot of pain, I believe I have figured out the cause of the issue. It lies with what appears to be a bug do_flush() in switch_rtp.c - this would certainly manifest as seemingly random/intermittent behaviour in respect of DTMF handling/processing.<div class=""><br class=""></div><div class="">I have amended the code and I am currently testing my fix.<br class=""><div class=""><br class=""></div><div class=""><div class=""><div dir="ltr" class=""><div dir="ltr" class=""><div class="">Kind regards,<br class=""></div><div class=""><br class=""></div><div class="">Andy</div></div></div></div><br class=""></div></div></div><br class=""><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Mon, 28 Feb 2022 at 14:12, Andy Newlands <<a href="mailto:andynewlands@gmail.com" target="_blank" class="">andynewlands@gmail.com</a>> wrote:<br class=""></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr" class="">Hi,<div class=""><br class=""></div><div class="">We have a reproducible problem where spurious DTMF characters are introduced on calls from Freeswitch to a remote (3rd party) IVR, but ONLY when using SRTP (not with RTP)</div><div class=""><br class=""></div><div class="">When the spurious character is generated it is:</div><div class=""><br class=""></div><div class="">1. Only ever following a valid/legitimate character (one that was genuinely keyed)</div><div class="">2. Often has a very long duration (see below)</div><div class="">3. Appears to be a random character - often 0x00, but sometimes 0-9/A-F,#,* (and that's when we get problems with the IVR - FS automatically drops the 0x00 values).</div><div class=""><br class=""></div><div class="">We have been able to reproduce this from mobile/cell phones and traditional land-lines.</div><div class=""><br class=""></div><div class="">I added some code to: switch_rtp.c, in static handle_rfc2833_result_t handle_rfc2833(switch_rtp_t *rtp_session, switch_size_t bytes, int *do_cng), to try to address this.</div><div class=""><br class=""></div><div class="">It logs an error (with the character hex value and duration) for DTMF characters with "long" durations (over 10000) and then discards them. As you can see, below, this is a fairly frequent occurrence (although most of the time, the character value is null, so FS drps it, anyway - but sometimes, "real" characters are introduced).</div><div class=""><br class=""></div><div class="">2022-02-28 14:00:53.743665 [NOTICE] switch_vpx.c:599 VPX encoder reset (WxH/BW) from 0x0/0 to 352x288/1024<br class="">2022-02-28 14:00:56.103665 [INFO] switch_channel.c:522 RECV DTMF 4:2400<br class="">2022-02-28 14:00:56.303664 [ERR] switch_channel.c:557 Ignored invalid DTMF on Call-ID: 3b5d08d3-d898-40b6-ae00-5886713ca6dd, DTMF: 0x0 Duration: 5063<br class="">2022-02-28 14:00:56.383661 [INFO] switch_channel.c:522 RECV DTMF 4:2400<br class="">2022-02-28 14:00:56.583665 [INFO] switch_channel.c:522 RECV DTMF 5:2400<br class="">2022-02-28 14:00:57.063662 [INFO] switch_channel.c:522 RECV DTMF 5:2400<br class="">2022-02-28 14:00:57.123670 [INFO] switch_channel.c:522 RECV DTMF 8:2720<br class="">2022-02-28 14:00:57.323673 [ERR] switch_rtp.c:686 Discarded long-duration DTMF on Call-ID: 3b5d08d3-d898-40b6-ae00-5886713ca6dd, DTMF: 0x0 Duration: 54088<br class="">2022-02-28 14:00:57.603671 [INFO] switch_channel.c:522 RECV DTMF 9:2400<br class="">2022-02-28 14:00:57.824102 [INFO] switch_channel.c:522 RECV DTMF 8:2720<br class="">2022-02-28 14:00:57.903663 [ERR] switch_rtp.c:686 Discarded long-duration DTMF on Call-ID: 3b5d08d3-d898-40b6-ae00-5886713ca6dd, DTMF: 0x0 Duration: 57002<br class="">2022-02-28 14:00:58.183668 [INFO] switch_channel.c:522 RECV DTMF 6:2400<br class="">2022-02-28 14:00:58.363666 [ERR] switch_rtp.c:686 Discarded long-duration DTMF on Call-ID: 3b5d08d3-d898-40b6-ae00-5886713ca6dd, DTMF: 0x35 Duration: 57073<br class="">2022-02-28 14:00:58.503667 [INFO] switch_channel.c:522 RECV DTMF 9:2400<br class="">2022-02-28 14:00:58.643667 [INFO] switch_channel.c:522 RECV DTMF 1:2400<br class="">2022-02-28 14:00:58.963663 [INFO] switch_channel.c:522 RECV DTMF 6:2400<br class="">2022-02-28 14:00:59.283665 [INFO] switch_channel.c:522 RECV DTMF 1:2400<br class=""></div><div class=""><br class=""></div><div class="">The first spurious character is 0x00 (which FS will drop anyway) but this could easily be a valid character (and would escape detection because its duration is not too long - it's long but not unreasonable). So, this is an imperfect "fix" as we occasionally see a spurious character with a legitimate value and a "sensible" duration - sufficiently often for users to complain. The last [ERR] entry shows a '5' being introduced but this is dropped because the duration is "crazy" (57073)</div><div class=""><br class=""></div><div class="">As mentioned, this does not happen with secure media disabled - only with SRTP. So, I am wondering if the code correctly processes a digit then, sometimes (somehow) incorrectly attempts to decode part of an audio packet as though it were RFC2833.<div class=""><div dir="ltr" class=""><div dir="ltr" class=""><div class=""><br class=""></div><div class="">Has anyone else experienced this? Does anyone know what may be causing the problem and what the fix might be.</div><div class=""><br class=""></div><div class="">I can provide console output and tcp dumps if required, along with any other details that may help (I'm also comfortable making code changes - to for extra logging/diagnostics etc).</div></div></div></div></div></div></blockquote></div></blockquote></div></div></blockquote></div><br class=""></body></html>