sr-users

sr-users@lists.kamailio.org

35 participants
37297 discussions

Failed to parse the R-URI with topoh enabled

by Ale

Hi All, I think I have a problem with topoh, but I can't identify it. Kamailio 5.6.6, used as a stateless proxy, receives a 487 from a phone and propagates it correctly. Next comes the ack that should be forwarded back to the phone, but instead it doesn't forward it and produces the following error "pv_get_ruri(): failed to parse the R-URI" The error is generated by xlog("...$ru from $fu...\n"") as first line of request_route. If I disable topoh everything works perfectly.Topoh is only configured with mask_key only. Furthermore, not all phones generate this problem, at the moment there are some snoms. ACK sip:user@phone_ip:2048;line=kkqgarrj SIP/2.0 Via: SIP/2.0/UDP proxy_sip_ip;branch=z9hG4bK2ec.e7126a3134bef9d974f57dd43ebd4ea2.0 Route: <sip:kamailio_ip:5060;lr;received=sip:phone_ip:2048> Max-Forwards: 66 From: <sip:111111111@sip.example.com>;tag=pDc1BQ0B57Ujj To: <sip:user@sip.example.com>;tag=vslvowy4y2 Call-ID: 40a6c905-c0c8-4c20-b3f7-397b3fce58b6 CSeq: 90285584 ACK Content Length: 0 There is no difference between the ACK of the snom and other phones that work. The only difference I noticed is that the snom 487 contains Contact field in the header. Any suggestions? Thank you

3 weeks, 4 days

How to handle e2e ACK with t_check_trans()

by Benoit Panizzon

Hi List To correctly route ACK belonging to a transaction (totag) the example code does: if ( is_method("ACK") ) { if ( t_check_trans() ) { route(RELAY); exit; } else { xlog("L_INFO", "$cfg(route): $rm NOT loose route NO transaction found DROP\n"); exit; } } Unfortunatels I fear that in a parallell forking scenario, the transaction is remove from memory as soon as the first 'negative' ACK is being processed so that the end 2 end ACK to the 200 OK of the branch that connected the call, never makes it through. I attempted to increase wt_timer to 10000 to have the transaction stay around 10 seconds but that does not solve the issue. I read that t_check_trans()'s result shall be considered more as a hint than as a confirmation of an existing transaction. Is there a clean / secure way to process e2e ACK and make sure they belong to a transaction? Or can I just safely remove the t_check_trans() call and route any ACK with a totag? Mit freundlichen Grüssen -Benoît Panizzon- -- I m p r o W a r e A G - Leiter Commerce Kunden ______________________________________________________ Zurlindenstrasse 29 Tel +41 61 826 93 00 CH-4133 Pratteln Fax +41 61 826 93 01 Schweiz Web http://www.imp.ch ______________________________________________________

3 weeks, 5 days

Dealing with flood of invalid SIP messages

by mayamatakeshi

Hi, I was going through some old company tickets that I am assigned to and found a case when possibly an attacker flooded our kamailio server with invalid sip messages like this: 2019-04-27T20:14:05.533554+09:00 IPX051 /usr/local/src/git/sip-router/kamailio[1732]: ERROR: <core> [parser/msg_parser.c:714]: ERROR: parse_msg: message=<[F#016sD#026Z<8D>97<F8><B5>;<A9><E7>-<D2>(<E2><F6> v;/#021k\<CC>8<B1>λ<F4>#004M<B6><BE><EC>#035#003<94><E1>=<A0><FF><E3><AF>Kwzr<8B>A#036B<D7>#027#023cu<82>Y<D4>#037<FB><AC>S_<C4>Qg<AB><DE>F<88>I#006<8C><FA><F4>~#y3G<C7>H<80>b<BC><AD>#035<89>#002<DB><C8>#001U<9E>#007<CB><F9>nT<E5><EE><8E><F1>#0144> At that time we manually banned the IP. But it would be helpful to have this done automatically by fail2ban. So I was thinking this log should include the src IP address. I looked at the latest kamailio commit and core/parser/msg_parser.c does this log the same way so I was thinking in opening an issue for this. But maybe this should be dealt with differently. Any ideas?

3 weeks, 6 days

Issue with Retrieving Kamailio Statistics: Missing FIFO File and Errors

by satyaprakash ch

Hi, I'm encountering an issue while trying to retrieve Kamailio statistics on the server using the command kamctl stats. Instead of the expected output, I am receiving the following errors: kamctl stats: -e \E[37;31mERROR: Error opening Kamailio's FIFO /run/kamailio/kamailio_rpc.fifo -e \E[37;31mERROR: Make sure you have loaded the jsonrpcs module and set FIFO transport parameters Additionally, the file /run/kamailio/kamailio_rpc.fifo does not exist. Could you please provide suggestions ?

4 weeks, 1 day

Handeling error returned by rtpengine_manage(...)

by Wolter Eldering

Hello, I have a question about the rtpengine module and how to use it via KEMI/python. Will processing of the native script stop when rtpengine_manage returns with an error? Or will it continue and will the SDP be forwarded unchanged? #!ifdef WITH_RTPENGINE if(nat_uac_test("8")) { rtpengine_manage("SIP-source-address replace-origin replace-session-connection"); } else { rtpengine_manage("replace-origin replace-session-connection"); } #!else Regards, Wolter

1 month

Kamailio not deleteing rtp sessions

by Tihomir Culjaga

Hello, I'm using Kamailio + RTPEngine. Everything works well except one thing. I noticed rtp sessions are not being deleted after call termination. Instead, they remain lingering on RTPEngine waiting for a timeout. kamailio --version version: kamailio 5.5.2 (x86_64/linux) 55e232 flags: USE_TCP, USE_TLS, USE_SCTP, TLS_HOOKS, USE_RAW_SOCKS, DISABLE_NAGLE, USE_MCAST, DNS_IP_HACK, SHM_MMAP, PKG_MALLOC, Q_MALLOC, F_MALLOC, TLSF_MALLOC, DBG_SR_MEMORY, USE_FUTEX, FAST_LOCK-ADAPTIVE_WAIT, USE_DNS_CACHE, USE_DNS_FAILOVER, USE_NAPTR, USE_DST_BLOCKLIST, HAVE_RESOLV_RES, TLS_PTHREAD_MUTEX_SHARED ADAPTIVE_WAIT_LOOPS 1024, MAX_RECV_BUFFER_SIZE 262144, MAX_URI_SIZE 1024, BUF_SIZE 65535, DEFAULT PKG_SIZE 8MB poll method support: poll, epoll_lt, epoll_et, sigio_rt, select. id: 55e232 compiled on 14:13:04 Aug 25 2021 with gcc 8.4.1 rtpengine --version Version: 9.5.1.1-1.el8 rtpengine-ctl list totals tells me every session went to timeout :) --- snip --- Total statistics (does not include current running sessions): Uptime of rtpengine :914 seconds Total managed sessions :51 Total rejected sessions :0 Total timed-out sessions via TIMEOUT :51 Total timed-out sessions via SILENT_TIMEOUT :0 Total timed-out sessions via FINAL_TIMEOUT :0 Total timed-out sessions via OFFER_TIMEOUT :0 Total regular terminated sessions :0 Total forced terminated sessions :0 --- snip --- I'm using rtpengine_manage in two routes "route[RELAY]" and onreply_route[MANAGE_REPLY] . Do I need to add a specific rtpengine_delete somewhere ? route[RELAY] { ... $xavp(r=>$T_branch_idx) = "replace-origin replace-session-connection"; if (!nat_uac_test("8")) { $xavp(r=>$T_branch_idx) = $xavp(r=>$T_branch_idx) + " trust-address"; } if (!has_totag()) { xlog("L_INFO", "NEXIOS: proto = [$proto], ru = [$ru]"); if ($proto =~ "ws" && !($ru =~ "transport=ws")) { xlog("L_INFO", "####################### FROM WS to any ########################\n"); $xavp(r=>$T_branch_idx) = $xavp(r=>$T_branch_idx) + " rtcp-mux-demux DTLS=off SDES-off ICE=remove RTP/AVP direction=external direction=internal"; } else if (!($proto =~ "ws") && $ru =~ "transport=ws") { xlog("L_INFO", "####################### FROM any to WS ########################\n"); setbflag(FLB_TOWEBRTC); $xavp(r=>$T_branch_idx) = $xavp(r=>$T_branch_idx) + " rtcp-mux-offer generate-mid DTLS=active SDES-off ICE=force RTP/SAVPF direction=internal direction=external"; } else if (($proto =~ "udp") && $ru =~ "transport=udp" && $fs =~ "MY_IP2_ADDR") { xlog("L_INFO", "####################### FROM ?? to ?? $ru ########################\n"); } } # we engage RTPEngine if (has_body("application/sdp")) { xlog("L_INFO", "NEXIOS: # $xavp(r=>$T_branch_idx) #\n"); rtpengine_manage($xavp(r=>$T_branch_idx)); } if (!t_relay()) { sl_reply_error(); } exit; } onreply_route[MANAGE_REPLY] { ... $xavp(r=>$T_branch_idx) = "replace-origin replace-session-connection"; if (!nat_uac_test("8")) { $xavp(r=>$T_branch_idx) = $xavp(r=>$T_branch_idx) + " trust-address"; } if (has_totag()) { xlog("L_INFO", "onreply_route proto = [$proto], ru = [$ru]"); if ($proto =~ "ws" && !($ru =~ "transport=ws")) { xlog("L_INFO", "onreply_route ####################### FROM WS to any ########################\n"); $xavp(r=>$T_branch_idx) = $xavp(r=>$T_branch_idx) + " rtcp-mux-demux DTLS=off SDES-off ICE=remove RTP/AVP direction=external direction=internal"; } else if (!($proto =~ "ws") && $ru =~ "transport=ws") { xlog("L_INFO", "onreply_route ####################### FROM any to WS ########################\n"); setbflag(FLB_TOWEBRTC); $xavp(r=>$T_branch_idx) = $xavp(r=>$T_branch_idx) + " rtcp-mux-offer generate-mid DTLS=active SDES-off ICE=force RTP/SAVPF direction=internal direction=external"; } else if (($proto =~ "udp") && $ru =~ "transport=udp" && $fs =~ "MY_IP2_ADDR") { xlog("L_INFO", "onreply_route ####################### FROM ?? to ?? $ru ########################\n"); } } if (has_body("application/sdp")) { xlog("L_INFO", "onreply_route # $xavp(r=>$T_branch_idx) #\n"); rtpengine_manage($xavp(r=>$T_branch_idx)); } } Any ideas are appreciated. Regards, Tihomir.

1 month

parallel forking: how to fix 183 g711 then 200 ok g722?

by Benoit Panizzon

Hi Gang Parallel forking situation. A => B A => C A invites with g711,g722 which is forked to B and C B reports ringing by sending 183 + SDP g711 C report 180 ringing. C establishes connection 200 OK + SDP g722 https://www.rfc-editor.org/rfc/rfc3261.html#page-78 o If the initial offer is in an INVITE, the answer MUST be in a reliable non-failure message from UAS back to UAC which is correlated to that INVITE. For this specification, that is only the final 2xx response to that INVITE. That same exact answer MAY also be placed in any provisional responses sent prior to the answer. The UAC MUST treat the first session description it receives as the answer, and MUST ignore any session descriptions in subsequent responses to the initial INVITE. Now we run into the situation in which A received the 183 g711 and being RFC compliant, dropped the 2nd SDP from the 200 OK. => No audio! Has anyone encountered this situation and found a nice solution? Like generating an UPDATE or RE-INVITE towards A to change the codec in an RFC compliant way after the call is established? Or is the only solution to remove all codecs except the required g711a one when more than one contact are registered on an AOR? Mit freundlichen Grüssen -Benoît Panizzon- -- I m p r o W a r e A G - Leiter Commerce Kunden ______________________________________________________ Zurlindenstrasse 29 Tel +41 61 826 93 00 CH-4133 Pratteln Fax +41 61 826 93 01 Schweiz Web http://www.imp.ch ______________________________________________________

1 month

Statically Linking Multiple Modules Compile Error

by Steve Bezalel

Hello, I am trying to compile Kamailio with all its modules statically linked. How should I go about this? I followed through the installation documentation but could not find any information. I tried passing all the modules I want to statically link in the "static_modules" argument of `make` but it seems that definitions of modules clash as such. ``` make static_modules="siputils rr" ``` ``` /usr/bin/ld: modules/rr/rr_mod.o:/playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/rr/rr_mod.c:69: multiple definition of `module_version'; modules/siputils/siputils.o:/playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/siputils/siputils.c:78: first defined here /usr/bin/ld: modules/rr/rr_mod.o:/playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/rr/rr_mod.c:69: multiple definition of `module_flags'; modules/siputils/siputils.o:/playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/siputils/siputils.c:78: first defined here /usr/bin/ld: modules/rr/rr_mod.o:/playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/rr/rr_mod.c:69: multiple definition of `module_interface_ver'; modules/siputils/siputils.o:/playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/siputils/siputils.c:78: first defined here /usr/bin/ld: modules/rr/rr_mod.o:/playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/rr/rr_mod.c:166: multiple definition of `exports'; modules/siputils/siputils.o:/playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/siputils/siputils.c:241: first defined here /usr/bin/ld: modules/rr/rr_mod.o: in function `mod_register': /playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/rr/rr_mod.c:926: multiple definition of `mod_register'; modules/siputils/siputils.o:/playground/fuzzing-targets/sip/kamailio-dfsan/src/modules/siputils/siputils.c:817: first defined here ``` Any help would be appreciated. Thank you.

1 month

$hdr behaviour using jansson+http_connect

by Sergio Charrua

Hi all! My team and I are finalizing implementation of STIR/SHAKEN. The flow has been simplified and is something like: Call comes from SBC to Kamailio, Kamailio makes an HTTP request to a REST API, response is received by Kamailio (5.8.2) and depending on the response, it replies back to SBC with a SIP 300 Multiple Choice of a SIP 4xx error, denying the call. The INVITEs received by SBC may or may not contain the Identity header. On Kamailio side, the following code gets the required headers to build the JSON object sent to REST API: route[HANDLE_STIRSHAKEN] { if (!is_method("INVITE")) { return; } $var(requestTime) = $utimef(%s); $var(ruri) = "$ru"; #request URI $var(from) = "$hdr(From)"; #$fU From header $var(to) = "$hdr(To)"; #$tU; To header $var(contact) = "$hdr(Contact)"; #Contact header $var(callID) = "$ci"; # Call ID $var(identity) = "$hdr(Identity)"; # Identity header jansson_set("string","from" , $var(from) , "$var(post)"); jansson_set("string","to" , $var(to) , "$var(post)"); jansson_set("string","r-uri" , $var(ruri) , "$var(post)"); jansson_set("string","contact" , $var(contact) , "$var(post)"); jansson_set("string","call-id" , $var(callID) , "$var(post)"); jansson_set("string","identity" , $var(identity) , "$var(post)"); http_connect("api1", "/stsh", "application/json","$var(post)","$var(result)"); xlog("L_DEBUG", "HANDLE_STIRSHAKEN - Result = $var(result)\n"); [... do some other irrelevant stuff ...] } The usual INVITE is similar to this: INVITE sip:+33142XXXXXX@10.242.XXX.YYY:5060;user=phone SIP/2.0 Via: SIP/2.0/UDP 10.240.AAA.BBB:5064;branch=z9hG4bK+30f178b47ee25dd589f986ed7402b4eb3+sip+6+a64eb889 From: "+493XXXXXXX" <sip:+493XXXXXXX@10.240.AAA.BBB ;user=phone>;tag=10.240.AAA.BBB+6+ffe686f8+614d17a3 To: +33142XXXXXX <sip:+33142XXXXXX@10.242.XXX.YYY> CSeq: 1 INVITE Expires: 180 Content-Length: 234 Call-Info: <sip:10.240.AAA.BBB:5064>;method="NOTIFY;Event=telephone-event;Duration=2000" Supported: timer, 100rel Contact: <sip:+493XXXXXXX@10.240.AAA.BBB:5064;user=phone> Content-Type: application/sdp Call-ID: 0gQAAC8WAAACBAAALxYAAHRfFdrMLbnN/iMYu2NTT6kvhuhZAFl7MG1V2ozJLUcXg9dxbyQZ/PtoswtdHtcv9g--(a)10.240.AAA.BBB Min-SE: 90 Session-Expires: 90;refresher=uac X-dtmf: 5 P-Asserted-Identity: "+493XXXXXXX" <sip:+493XXXXXXX@10.240.AAA.BBB> Accept: application/sdp Accept: application/dtmf-relay Max-Forwards: 66 Allow: INVITE,CANCEL,BYE,ACK, INFO Identity: dGhpcyBpcyBzb21lIGR1bW15IGJhc2U2NCBlbmNvZGVkIElkZW50aXR5IGhlYWRlcg==.eW91IHdvdWxkIG5vdCBleHBlY3QgbWUgdG8gbGVhdmUgdGhlIHJlYWwgdmFsdWUgaGVyZSByaWdodD8=.dGhhbmtzIGZvciB5b3VyIGNvbXByZWhlbnNpb24=;info=< https://someURL.com/with/some/parameter;sc697295b.cer>;alg=ES256;ppt=shaken P-Origination-Id: 23A16F8B-F9AB-4536-9884-85C96EF3B964 P-Attestation-Indicator: C has-sdp: present [SDP stuff] (I have replaced some of the values for obvious reasons). - SBC is sending every 70 secs 2 automatically generated INVITEs to kamailio - 1st call contains the Identity header, the other call does not contain Identity header. - Both calls have different Call-IDs and different FROM and TO values, so they are different calls and can be perfectly found in logs as separated calls - Kamailio receives each call and extracts headers, as shown in code above, and sends the JSON request to HTTP REST API Here is the weird part.... every now and then, completely random, the JSON object request sent to the HTTP REST API contains an Identity value, supposedly extracted from the SIP header **but** the SIP message captured in SNGrep doesn't have **any* I*dentity header!! In fact, in the kamailio logs, where I sent the header values extracted from INVITE, the identity is empty! BUT the JSON object has the identity set with value. And the worst part is that, when this issue happens, the value of the Identity set in JSON belongs to a completely different call that was processed a few minutes before ! So I suspected about some Caching behaviour on the http_connect function and added the following lines before sending the JSON to REST API: if ( is_present_hf("Identity") ) { jansson_set("string","identity",* $var(identity),* "$var(post)"); } else jansson_set("string","identity", "", "$var(post)"); forcing the identity property of the JSON to an empty string, but this did not work either! I even added a new property which is the current timestamp making sure every request was unique, and that did not work either! The only way I could make it work was replacing the above lines using these lines: if ( is_present_hf("Identity") ) { jansson_set("string","identity", *(a)hf_value2.Identity*, "$var(post)"); } else jansson_set("string","identity", "", "$var(post)"); So, instead of using $hdr(Identity) , Kamailio extracts the Identity header using @hf_value2.Identity Question: why is this behaviour? Could this be a bug somewhere? Atenciosamente / Kind Regards / Cordialement / Un saludo, *Sérgio Charrua*

1 month

Re: Dispatcher ping: "binding to source address X.X.X.X:XXXX failed: Address already in use" for TCP and TLS destinations

by Mirko Brankovic

Hi, Has anyone found how to safely suppress the warning? We see it on 5.7.4 Thanks, Mirko

1 month, 1 week

← Newer
1
2
3
4
5
6
7
...
3730
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

sr-users