November 2009 - h323plus - lists.packetizer.com

PacPhone 2.3.0 Released
by Simon Horne 30 Nov '09

30 Nov '09

Guys www.pacphone.com PacPhone 2.3.0 has been released. It has been extensively tested with the latest GnuGk release. Special thanks to all the people who tested the pre-Release and provided feedback. Anyone has any questions let me know. Simon

1 0

missing RTP Frames
by Igor Pavlov 27 Nov '09

27 Nov '09

Hello. I missing RTP frames, when connect with my H264 plugin on connection speed greater then approximately 256 kBps. I inserted logging of rtp frames sequencenumber in function RTP_Session::SendReceiveStatus RTP_UDP::ReadDataPDU(RTP_DataFrame & frame); after call SendReceiveStatus status = ReadDataOrControlPDU(*dataSocket, frame, TRUE); and periodicaly I have holes in sequence of numbers. At the same time I have working sniffer, and I see that RTP packet missing in my logs is present in sniffers statistics. I tryed to find the place in h323plus sources where can I miss RTP packets, but couldn't find it :( The missed group of packets (even if length of group is 1 - single packet) always ends with rtp packet which have marker=0 Help me to find out the reason why I missing packets. Because of speed?

1 0

GNU Gatekeeper 2.3.1 released
by Jan Willamowius 26 Nov '09

26 Nov '09

Hi, a new stable version of the GNU Gatekeeper is now available. Source code and executables can be downloaded from http://www.gnugk.org/h323download.html Executables are available for - Linux (32 bit and 64 bit) - Windows (XP, Vista, Windows7, ...) - MacOSX (PPC and Intel) - Solaris 10 - FreeBSD - OpenBSD What's new in 2.3.1 ? - lots of interoperability fixes for H.460.18/.19 based on the feedback from the community - new routing policy "CatchAll" - new [ModeSelection] section to fine tune when to proxy calls - support for setting the proxy mode per call using Radius - support for ignoring case and alias type (calling "FOO" will match alias "foo") - H.460.23/.24 support updated to the now approved standard - many bug fixes, eg. for direct mode and URI dialing - a unified binary for MacOS X that works both on PPC and Intel - Firebird database driver updated - many small improvements in the manual See below for the full list of changes. Enjoy, Jan -- Jan Willamowius, jan(a)willamowius.de http://www.gnugk.org/ http://www.willamowius.de/gnugk-support.html Changes from 2.3.0 to 2.3.1 =========================== - Firebird database driver updated for Firebird 2.0.x and 2.1.x - BUGFIX(Toolkit.cxx) don't ignore ExternalIP settings when NetworkInterfaces= is set - BUGFIX(SoftPBX.cxx) make response messages more consistent - the old format (2.0) to specify neighbor gatekeepers is deprecated now and has been removed from the manual; it still works, but may be removed from future versions - BUGFIX(ProxyChannel.cxx) only close endpoints logical channel on CloseLogicalChannel, use [Proxy]SearchBothSidesOnCLC=1 to retain the old behavior - BUGFIX(Toolkit.cxx) Enable RewritePString to rewrite address correctly if unregistered full URI. - BUGFIX(Neighbor.cxx) srv policy now fails calls and doesn't pass them to the next policy if a SRV record exists for a domain, but doesn't have a destination for the call - allow to set proxy mode through 'proxy' variable in Radius h323-ivr-in attribute (thanks to Yuriy Georgiewskiy) - BUGFIX(Routing.cxx) let dns policy handle LRQs that resolve locally, so URI dialing from neighbors works - the manual has been improved (thanks to Robert Kulagowski) - set Radius VSA attribute h323_rtp_proxy in stop event (thanks to Yuriy Georgiewskiy) - BUGFIX(ProxyChannel.cxx) fix memory leak when creation of a fastStart channel fails - status port command DisconnectIP now also works with unregistered endpoints using the std port - BUGFIX(ProxyChannel.cxx) fix handling of H.460.19 RTP keepalive packets - new switch [Proxy]DisableRTPQueueing=1 - BUGFIX(ProxyChannel.cxx) fully initialize UDP proxy IPs and ports - BUGFIX(Routing.cxx) don't apply dns policy to aliases of type dialedDigits - BUGFIX(Toolkit.h/.cxx) use default IP for networks specified in InternalNetwork=, not ExternalIP - enable data execution protection on Windows versions that support it (starting with XP SP3) - BUGFIX(gk.cxx) make sure Windows service writes same trace format as application (including file and line) - BUGFIX(Neighbor.cxx) fix memory leak in srv policy, only send the local part of the alias in the LRQ - new switch [RoutedMode]H46018NoNat to disable H.460.18 if no NAT detected (default enabled) - new routing policy 'catchall' to route all calls to one alias or IP - new Radius attributes on stop: h323pddtime, h323ringtime, h323routeattempts - BUGFIX(Neighbor.cxx) fix srv policy for direct mode, don't send LRQ if SRV points to us - new switch [RoutedMode]TranslateSorensonSourceInfo=1 to convert the Sorenson VP200 caller information - BUGFIX(GkStatus.cxx) don't flush status port socket when access is denied to avoid blocking - new switches in [Gatekeeper::Main]: CompareAliasType and CompareAliasCase to switch off alias type and case matching - BUGFIX(Routing.cxx) fix dns policy for direct mode - BUGFIX(Neighbor.cxx) try next policy if SRV lookup fails - neighbor type (GnuGk|CiscoGk|ClarentGk|GlonetGk) is now case insensitive - features documented: [AssignedAliases::SQL], [AssignedGatekeepers::SQL] and [GkQoSMonitor::SQL] - BUGFIX(RasTbl.cxx) use rewritten number when decrementing prefix capacity counters, so they match the increment - feature to set a ClientAuthId in SqlAuth that is provided as %{client-auth-id} on all accounting events - new status port command ResetCallCounters - changed ProxyForSameNAT default from 0 to 1 (to avoid media failure uncertainty) - BUGFIX(RasSrv.cxx) IsReachable looks at correct route table. (virtual was ignored) - changed [RoutedMode]EnableH46024= to be consistent with H.460.18 switch - BUGFIX(RasSrv.cxx) RCF messages H.460 features now supported Features not GenericData - BUGFIX(ProxyChannel.cxx) H.460.19 working with Fast Connect - BUGFIX(RasSrv.cxx,ProxyChannel.cxx) Tandberg MXP interoperability fixes - BUGFIX(ProxyChannel.cxx) check capacity on routes supplied by authenticators (Radius, SQL etc.) - BUGFIX(RasSrv.cxx) avoid cutting off status messages after 1000 chars (problem in PString::Printf) - BUGFIX(ProxyChannel.cxx) fix for fastStart with H.460.19 - new switch [RoutedMode]EnableH.460.24=, which defaults to OFF - send RTCP statistics to Radius server (thanks to Yuriy Georgiewskiy) - BUGFIX(Toolkit.cxx) fix a crash in RewriteE164() when casting aliases - new config section [ModeSelection] to specify the routing/proxy mode by IP network

1 0

OpenMCU Codec issues
by Josh Barth 25 Nov '09

25 Nov '09

I have the ptlib, h323plus, plugins, and openmcu from h323plus.org I built each successfully, but I am not having any luck getting the additional codecs loaded when I launch OpenMCU, all I get is G.711. Did I not build something right? In order I built, ptlib, h323plus, openmcu, plugins Where should the codecs be located? (they are in /usr/local/lib/pwlib) These are modular correct, I don't need to re-compile h323plus or OpenMCU? Here is my trace: 0:00.081 OpenMCU Version 2.2.5 by H323plus on Unix Linux (2.6.26-2-sparc64-sparc64) at 2009/11/17 14:43:29.321 0:00.082 OpenMCU PWLib File handle high water mark set: 15 Thread unblock pipe 0:00.082 OpenMCU PWLib Created thread 0x162c78 H323 Listener:%0x 0:00.083 OpenMCU PWLib File handle high water mark set: 16 PTCPSocket 0:00.083 OpenMCU H323 Started listener Listener[ip$*:1720] 0:00.084 OpenMCU PWLib Thread high water mark set: 6 0:00.085 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.085 OpenMCU H323 Added capability: G.711-ALaw-64k <1> 0:00.085 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.086 OpenMCU H323 Added capability: G.711-uLaw-64k <2> 0:00.086 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.086 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.086 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.086 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.087 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.087 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.087 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.087 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.087 OpenMCU EXT No Extended Capabilities found to load 0:00.088 OpenMCU H323 FindCapability: "*-QCIF*" 0:00.088 OpenMCU H323 FindCapability: "*-SQCIF*" 0:00.088 OpenMCU H323 Added capability: UserInput/hookflash <3> 0:00.088 OpenMCU H323 Added capability: UserInput/basicString <4> 0:00.088 OpenMCU H323 Added capability: UserInput/dtmf <5> 0:00.089 OpenMCU H323 Added capability: UserInput/RFC2833 <6> 0:00.089 OpenMCU H323 Added capability: UserInput/Navigation <7> 0:00.089 OpenMCU H323 Added capability: UserInput/Softkey <8> 0:00.089 OpenMCU H323 Added capability: UserInput/PointDevice <9> 0:00.089 OpenMCU H323 Added capability: UserInput/Modal <10> 0:00.089 OpenMCU MCU Codecs (in preference order): Table: G.711-ALaw-64k <1> G.711-uLaw-64k <2> UserInput/hookflash <3> UserInput/basicString <4> UserInput/dtmf <5> UserInput/RFC2833 <6> UserInput/Navigation <7> UserInput/Softkey <8> UserInput/PointDevice <9> UserInput/Modal <10> Set: 0: 0: G.711-ALaw-64k <1> G.711-uLaw-64k <2> 1: 2: UserInput/hookflash <3> UserInput/basicString <4> UserInput/dtmf <5> UserInput/RFC2833 <6> UserInput/Navigation <7> UserInput/Softkey <8> UserInput/PointDevice <9> UserInput/Modal <10> 0:00.094 OpenMCU PWLib File handle high water mark set: 17 PTCPSocket 0:00.095 OpenMCU PWLib File handle high water mark set: 19 Thread unblock pipe 0:00.095 OpenMCU PWLib Created thread 0x168620 HTTP Service 0:00.095 H323 Listener:f6f8bb90 PWLib Started thread 0x162c78 H323 Listener:f6f8bb90 0:00.095 H323 Listener:f6f8bb90 H323 Awaiting TCP connections on port 1720 0:00.095 H323 Listener:f6f8bb90 TCP Waiting on socket accept on ip$*:1720

4 9

H264 Plugin in H323Plus
by ooi khai chin 24 Nov '09

24 Nov '09

Hi All, regarding the H.241 documentation, it mention that H.264 plugin in H.323 only support single NAL mode. Then, I refer the RFC 3984, it state that single NAL mode only support type 1-23 NAL unit, no FU-A, no FU-B, ... Currently, I am using H.264 hardware encoder which used baseline profile, not libx264. My H.264 encoder(hardware) will only return the whole encoded data and it's total length, not each NAL and it's length. So, how can I make it comply to H.323 if no fragmentation is allowed in H.323? Thanks.

2 1

H264 PLUGIN in h323plus
by ooi khai chin 24 Nov '09

24 Nov '09

Hi All, For the libx264, It directly fragment the encoded data if it larger than the maxpayloadsize (1.4KB by default). If I had encoded data 5.6K. Then, it will fragment and sent out, 4 times. Is it comply to H.323 standard? Because currently other voip (polycom, Mirial...) can't display my encoded data size greater than 1.4K. Thanks.

2 1

H264 PLUGIN in H323Plus
by ooi khai chin 24 Nov '09

24 Nov '09

Hi All, Currently, my H264 plugin able to work with polycom. I found that the media statistic(polycom PVX ) only show the frame rate video format, and display the image if my sending encoded data is not more than 1.4KB. In my code, I do the fragmentation once it greater than 1.4KB(like libx264 do). However, I had read some documentations and all of them mentioned about H.241 only support the single NAL unit mode(mode 0). I wonder how this mode handle if encoded data greater than 1.4KB(usually happen)??? Any Idea? Thanks.

3 3

H264 PLUGIN in H323Plus
by ooi khai chin 23 Nov '09

23 Nov '09

Hi All, I captured the capability exchange from difference voip software like(polycom, vcon...), it show only "0.0.8.241.0.0.1", which is defined as OpalPluginCodec_Identifier_H264_Generic. 1. What is the difference with the others like (default, RFC3984 interleave, RFC3984 non interleave)? Another things is that When I use my H.264 Encoder API, it only return me the encoded data and the encoded length. 2. So, is it considered as single NAL unit? How it can be comply to H.323 since the default MTU value in H.323 is 1400bytes (like fragmentation or other)? I try to increate the MTU value, but it cause me error.

2 1

Re: [h323plus] [Openh323gk-users] PTLIB PASN_BMPString handling changes and endpoint not registered.
by Simon Horne 18 Nov '09

18 Nov '09

Guys I need to make you aware of a change that causes different registration endpoint identifier lengths to be generated based on the version of ptlib used. GnuGK 2.3.0 (ptlib 2.4.5) registrationConfirm { ... endpointIdentifier = 9 characters { 0034 0033 0039 0035 005f 0065 006e 0064 4395_end 0070 p } GnuGk 2.3.1 (ptlib cvs) registrationConfirm { ... endpointIdentifier = 10 characters { 0034 0033 0039 0035 005f 0065 006e 0064 4395_end 0070 0000 p } While this may not look too much of a problem however if you are registering a h323plus endpoint built off a later version of PTLIB (like PacPhone) on a GnuGk with an earlier version of PTLIB then your endpoint will register but will not be able to make any calls and you will get an error about endpoint not registered. This is because the gatekeeper generated a 9 character endpointIdentifer while the endpoint will send a 10 digit identifier to the gatekeeper. I have patched the latest PacPhone preRelease to resolve the problem www.pacphone.com/downloads/PacPhoneSetup230.exe I will be applying a fix to the H323plus CVS shortly. Simon

1 0

Re: [h323plus] OpenMCU Codec issues
by Simon Horne 17 Nov '09

17 Nov '09

Josh Put the plugins in /usr/lib/ptlib they will automatically be read in at application startup. There is no need to recompile OpenMCU. Simon _____ From: Josh Barth [mailto:jbarth@buttechc.com] Sent: Wednesday, 18 November 2009 10:38 AM To: 'Simon Horne' Subject: RE: [h323plus] OpenMCU Codec issues So I still have the same issue, I have set the PTLIBPLUGINDIR env var. Should the structure then be /usr/local/lib/ptlib and then each of the libraries? Or should they exist in codecs/ or codecs/audio/ and codecs/video/ I am right in thinking once these are placed properly OpenMCU should pick them up? I don't have to place them, then recompile OpenMCU do I? Thanks very much, Josh From: Simon Horne [mailto:s.horne@packetizer.com] Sent: Tuesday, November 17, 2009 5:17 PM To: Josh Barth Subject: RE: [h323plus] OpenMCU Codec issues Josh The default directory for plugins should be /usr/lib/ptlib unless you set the environment variable PTLIBPLUGINDIR to point to your plugin directory. Simon _____ From: h323plus-bounces(a)lists.packetizer.com [mailto:h323plus-bounces@lists.packetizer.com] On Behalf Of Josh Barth Sent: Wednesday, 18 November 2009 9:45 AM To: 'h323plus(a)lists.packetizer.com' Subject: [h323plus] OpenMCU Codec issues I have the ptlib, h323plus, plugins, and openmcu from h323plus.org I built each successfully, but I am not having any luck getting the additional codecs loaded when I launch OpenMCU, all I get is G.711. Did I not build something right? In order I built, ptlib, h323plus, openmcu, plugins Where should the codecs be located? (they are in /usr/local/lib/pwlib) These are modular correct, I don't need to re-compile h323plus or OpenMCU? Here is my trace: 0:00.081 OpenMCU Version 2.2.5 by H323plus on Unix Linux (2.6.26-2-sparc64-sparc64) at 2009/11/17 14:43:29.321 0:00.082 OpenMCU PWLib File handle high water mark set: 15 Thread unblock pipe 0:00.082 OpenMCU PWLib Created thread 0x162c78 H323 Listener:%0x 0:00.083 OpenMCU PWLib File handle high water mark set: 16 PTCPSocket 0:00.083 OpenMCU H323 Started listener Listener[ip$*:1720] 0:00.084 OpenMCU PWLib Thread high water mark set: 6 0:00.085 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.085 OpenMCU H323 Added capability: G.711-ALaw-64k <1> 0:00.085 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.086 OpenMCU H323 Added capability: G.711-uLaw-64k <2> 0:00.086 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.086 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.086 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.086 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.087 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.087 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.087 OpenMCU H323 FindCapability: "G.711-ALaw-64k{sw}" 0:00.087 OpenMCU H323 FindCapability: "G.711-uLaw-64k{sw}" 0:00.087 OpenMCU EXT No Extended Capabilities found to load 0:00.088 OpenMCU H323 FindCapability: "*-QCIF*" 0:00.088 OpenMCU H323 FindCapability: "*-SQCIF*" 0:00.088 OpenMCU H323 Added capability: UserInput/hookflash <3> 0:00.088 OpenMCU H323 Added capability: UserInput/basicString <4> 0:00.088 OpenMCU H323 Added capability: UserInput/dtmf <5> 0:00.089 OpenMCU H323 Added capability: UserInput/RFC2833 <6> 0:00.089 OpenMCU H323 Added capability: UserInput/Navigation <7> 0:00.089 OpenMCU H323 Added capability: UserInput/Softkey <8> 0:00.089 OpenMCU H323 Added capability: UserInput/PointDevice <9> 0:00.089 OpenMCU H323 Added capability: UserInput/Modal <10> 0:00.089 OpenMCU MCU Codecs (in preference order): Table: G.711-ALaw-64k <1> G.711-uLaw-64k <2> UserInput/hookflash <3> UserInput/basicString <4> UserInput/dtmf <5> UserInput/RFC2833 <6> UserInput/Navigation <7> UserInput/Softkey <8> UserInput/PointDevice <9> UserInput/Modal <10> Set: 0: 0: G.711-ALaw-64k <1> G.711-uLaw-64k <2> 1: 2: UserInput/hookflash <3> UserInput/basicString <4> UserInput/dtmf <5> UserInput/RFC2833 <6> UserInput/Navigation <7> UserInput/Softkey <8> UserInput/PointDevice <9> UserInput/Modal <10> 0:00.094 OpenMCU PWLib File handle high water mark set: 17 PTCPSocket 0:00.095 OpenMCU PWLib File handle high water mark set: 19 Thread unblock pipe 0:00.095 OpenMCU PWLib Created thread 0x168620 HTTP Service 0:00.095 H323 Listener:f6f8bb90 PWLib Started thread 0x162c78 H323 Listener:f6f8bb90 0:00.095 H323 Listener:f6f8bb90 H323 Awaiting TCP connections on port 1720 0:00.095 H323 Listener:f6f8bb90 TCP Waiting on socket accept on ip$*:1720

1 0