I detect a problem with the conference facility in pjmedia. In a slow/old machine, when the CPU is working in another process (something hard like opening a browser or another big program) the conference process stop its execution. How to test: First, enable the #define CONF_DEBUG in conference.c and build. Second, start in a terminal (I use gnome-terminal) the pjproject-1.14/pjsip-apps/bin/samples/i686-pc-linux-gnu/confsample program. Then, you will see a lot of debug lines, make a long click or try to select some text in the terminal to stop/delay execution. At least in my machine, the program still works with no hangs, but the conference stop and the get_frame() is never called again. Same effect opening something like Firefox, OpenOffice, etc. during conference execution. Thank you. Regards. -- ------------------------------------------------------------------------ Fabio E. Curi Gerencia de I+D Tecnologia En Seguridad SRL (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 http://www.tecnologiaenseguridad.com

Tested in Ubuntu 11, with an not so old machine, same behavior. Regards. Fabio Curi wrote: > I detect a problem with the conference facility in pjmedia. > > In a slow/old machine, when the CPU is working in another process > (something hard like opening a browser or another big program) the > conference process stop its execution. > > How to test: > > First, enable the #define CONF_DEBUG in conference.c and build. > Second, start in a terminal (I use gnome-terminal) the > pjproject-1.14/pjsip-apps/bin/samples/i686-pc-linux-gnu/confsample > program. > Then, you will see a lot of debug lines, make a long click or try to > select some text in the terminal to stop/delay execution. > > At least in my machine, the program still works with no hangs, but the > conference stop and the get_frame() is never called again. > > Same effect opening something like Firefox, OpenOffice, etc. during > conference execution. > > Thank you. > > Regards. > > > -- > ------------------------------------------------------------------------ > Fabio E. Curi > Gerencia de I+D > Tecnologia En Seguridad SRL > (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 > http://www.tecnologiaenseguridad.com > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org -- ------------------------------------------------------------------------ Fabio E. Curi Gerencia de I+D Tecnologia En Seguridad SRL (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 http://www.tecnologiaenseguridad.com

Hi Fabio, As the clock is from sound device (conference itself actually doesn't have clock), so the problem seems to be in the clock generator (in this case it's the sound device), it can be pjmedia sound dev wrapper bug or the backend bug or even perhaps expected behaviour/limitation for that specific backend. BR, nanang On Wed, May 9, 2012 at 8:02 PM, Fabio Curi <fcuri@tecnologiaenseguridad.com> wrote: > Tested in Ubuntu 11, with an not so old machine, same behavior. > > Regards. > > > Fabio Curi wrote: > > I detect a problem with the conference facility in pjmedia. > > In a slow/old machine, when the CPU is working in another process (something > hard like opening a browser or another big program) the conference process > stop its execution. > > How to test: > > First, enable the #define CONF_DEBUG in conference.c and build. > Second, start in a terminal (I use gnome-terminal) the > pjproject-1.14/pjsip-apps/bin/samples/i686-pc-linux-gnu/confsample program. > Then, you will see a lot of debug lines, make a long click or try to select > some text in the terminal to stop/delay execution. > > At least in my machine, the program still works with no hangs, but the > conference stop and the get_frame() is never called again. > > Same effect opening something like Firefox, OpenOffice, etc. during > conference execution. > > Thank you. > > Regards. > > > -- > ________________________________ > Fabio E. Curi > Gerencia de I+D > Tecnologia En Seguridad SRL > (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 > http://www.tecnologiaenseguridad.com > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > > > -- > ________________________________ > Fabio E. Curi > Gerencia de I+D > Tecnologia En Seguridad SRL > (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 > http://www.tecnologiaenseguridad.com > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >

Nanang, The problem is in portaudio, after the cpu delay, the PaAlsaStream_WaitForFrames() function in pa_linux_alsa.c never exit from a the while( pollPlayback || pollCapture ) {}. Then the capture/playback stops. I think this is a problem solved in the last portaudio version (may be), then I will try to change the portaudio version in pjmedia to test. It is possible? It is a trivial change or is too complex? Thank you. Regards. Nanang Izzuddin wrote: > Hi Fabio, > > As the clock is from sound device (conference itself actually doesn't > have clock), so the problem seems to be in the clock generator (in > this case it's the sound device), it can be pjmedia sound dev wrapper > bug or the backend bug or even perhaps expected behaviour/limitation > for that specific backend. > > BR, > nanang > > > On Wed, May 9, 2012 at 8:02 PM, Fabio Curi > <fcuri@tecnologiaenseguridad.com> wrote: >> Tested in Ubuntu 11, with an not so old machine, same behavior. >> >> Regards. >> >> >> Fabio Curi wrote: >> >> I detect a problem with the conference facility in pjmedia. >> >> In a slow/old machine, when the CPU is working in another process (something >> hard like opening a browser or another big program) the conference process >> stop its execution. >> >> How to test: >> >> First, enable the #define CONF_DEBUG in conference.c and build. >> Second, start in a terminal (I use gnome-terminal) the >> pjproject-1.14/pjsip-apps/bin/samples/i686-pc-linux-gnu/confsample program. >> Then, you will see a lot of debug lines, make a long click or try to select >> some text in the terminal to stop/delay execution. >> >> At least in my machine, the program still works with no hangs, but the >> conference stop and the get_frame() is never called again. >> >> Same effect opening something like Firefox, OpenOffice, etc. during >> conference execution. >> >> Thank you. >> >> Regards. >> >> >> -- >> ________________________________ >> Fabio E. Curi >> Gerencia de I+D >> Tecnologia En Seguridad SRL >> (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 >> http://www.tecnologiaenseguridad.com >> >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> >> >> -- >> ________________________________ >> Fabio E. Curi >> Gerencia de I+D >> Tecnologia En Seguridad SRL >> (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 >> http://www.tecnologiaenseguridad.com >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > -- ------------------------------------------------------------------------ Fabio E. Curi Gerencia de I+D Tecnologia En Seguridad SRL (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 http://www.tecnologiaenseguridad.com

We are slowly moving away from PortAudio, so I would suggest trying our native ALSA implementation instead. Best regards, Benny On Mon, May 14, 2012 at 5:20 AM, Fabio Curi <fcuri@tecnologiaenseguridad.com > wrote: > Nanang, > > The problem is in portaudio, after the cpu delay, the > PaAlsaStream_WaitForFrames() function in pa_linux_alsa.c never exit from a > the while( pollPlayback || pollCapture ) {}. Then the capture/playback > stops. > > I think this is a problem solved in the last portaudio version (may be), > then I will try to change the portaudio version in pjmedia to test. It is > possible? It is a trivial change or is too complex? > > Thank you. > > Regards. > > > > Nanang Izzuddin wrote: > > Hi Fabio, > > As the clock is from sound device (conference itself actually doesn't > have clock), so the problem seems to be in the clock generator (in > this case it's the sound device), it can be pjmedia sound dev wrapper > bug or the backend bug or even perhaps expected behaviour/limitation > for that specific backend. > > BR, > nanang > > > On Wed, May 9, 2012 at 8:02 PM, Fabio Curi<fcuri@tecnologiaenseguridad.com> <fcuri@tecnologiaenseguridad.com> wrote: > > Tested in Ubuntu 11, with an not so old machine, same behavior. > > Regards. > > > Fabio Curi wrote: > > I detect a problem with the conference facility in pjmedia. > > In a slow/old machine, when the CPU is working in another process (something > hard like opening a browser or another big program) the conference process > stop its execution. > > How to test: > > First, enable the #define CONF_DEBUG in conference.c and build. > Second, start in a terminal (I use gnome-terminal) the > pjproject-1.14/pjsip-apps/bin/samples/i686-pc-linux-gnu/confsample program. > Then, you will see a lot of debug lines, make a long click or try to select > some text in the terminal to stop/delay execution. > > At least in my machine, the program still works with no hangs, but the > conference stop and the get_frame() is never called again. > > Same effect opening something like Firefox, OpenOffice, etc. during > conference execution. > > Thank you. > > Regards. > > > -- > ________________________________ > Fabio E. Curi > Gerencia de I+D > Tecnologia En Seguridad SRL > (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992http://www.tecnologiaenseguridad.com > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing listpjsip@lists.pjsip.orghttp://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > > > -- > ________________________________ > Fabio E. Curi > Gerencia de I+D > Tecnologia En Seguridad SRL > (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992http://www.tecnologiaenseguridad.com > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing listpjsip@lists.pjsip.orghttp://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing listpjsip@lists.pjsip.orghttp://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > > > -- > ------------------------------ > Fabio E. Curi > Gerencia de I+D > Tecnologia En Seguridad SRL > (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 > http://www.tecnologiaenseguridad.com > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > >

I changed to the last portaudio. Look at my post "bad sound with pulseaudio" http://lists.pjsip.org/pipermail/pjsip_lists.pjsip.org/2012-May/014515.html With this version I solved my problem, but sometimes there is no sound at all - portaudio reports error. Regards, Mihail. 14.05.2012 2:20, Fabio Curi ?????: > Nanang, > > The problem is in portaudio, after the cpu delay, the > PaAlsaStream_WaitForFrames() function in pa_linux_alsa.c never exit > from a the while( pollPlayback || pollCapture ) {}. Then the > capture/playback stops. > > I think this is a problem solved in the last portaudio version (may > be), then I will try to change the portaudio version in pjmedia to > test. It is possible? It is a trivial change or is too complex? > > Thank you. > > Regards. > > > Nanang Izzuddin wrote: >> Hi Fabio, >> >> As the clock is from sound device (conference itself actually doesn't >> have clock), so the problem seems to be in the clock generator (in >> this case it's the sound device), it can be pjmedia sound dev wrapper >> bug or the backend bug or even perhaps expected behaviour/limitation >> for that specific backend. >> >> BR, >> nanang >> >> >> On Wed, May 9, 2012 at 8:02 PM, Fabio Curi >> <fcuri@tecnologiaenseguridad.com> wrote: >>> Tested in Ubuntu 11, with an not so old machine, same behavior. >>> >>> Regards. >>> >>> >>> Fabio Curi wrote: >>> >>> I detect a problem with the conference facility in pjmedia. >>> >>> In a slow/old machine, when the CPU is working in another process (something >>> hard like opening a browser or another big program) the conference process >>> stop its execution. >>> >>> How to test: >>> >>> First, enable the #define CONF_DEBUG in conference.c and build. >>> Second, start in a terminal (I use gnome-terminal) the >>> pjproject-1.14/pjsip-apps/bin/samples/i686-pc-linux-gnu/confsample program. >>> Then, you will see a lot of debug lines, make a long click or try to select >>> some text in the terminal to stop/delay execution. >>> >>> At least in my machine, the program still works with no hangs, but the >>> conference stop and the get_frame() is never called again. >>> >>> Same effect opening something like Firefox, OpenOffice, etc. during >>> conference execution. >>> >>> Thank you. >>> >>> Regards. >>> >>> >>> -- >>> ________________________________ >>> Fabio E. Curi >>> Gerencia de I+D >>> Tecnologia En Seguridad SRL >>> (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 >>> http://www.tecnologiaenseguridad.com >>> >>> >>> _______________________________________________ >>> Visit our blog:http://blog.pjsip.org >>> >>> pjsip mailing list >>> pjsip@lists.pjsip.org >>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>> >>> >>> >>> -- >>> ________________________________ >>> Fabio E. Curi >>> Gerencia de I+D >>> Tecnologia En Seguridad SRL >>> (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 >>> http://www.tecnologiaenseguridad.com >>> >>> _______________________________________________ >>> Visit our blog:http://blog.pjsip.org >>> >>> pjsip mailing list >>> pjsip@lists.pjsip.org >>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>> >> _______________________________________________ >> Visit our blog:http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> > > > -- > ------------------------------------------------------------------------ > Fabio E. Curi > Gerencia de I+D > Tecnologia En Seguridad SRL > (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 > http://www.tecnologiaenseguridad.com > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

 Nanang,

 The problem is in portaudio, after the cpu delay, the
 PaAlsaStream_WaitForFrames() function in pa_linux_alsa.c never
 exit from a the while( pollPlayback || pollCapture ) {}. Then the
 capture/playback stops.

 I think this is a problem solved in the last portaudio version
 (may be), then I will try to change the portaudio version in
 pjmedia to test. It is possible? It is a trivial change or is too
 complex?

 Thank you.

 Regards.



 Nanang Izzuddin wrote:

 Hi Fabio,

 As the clock is from sound device (conference itself actually doesn't
 have clock), so the problem seems to be in the clock generator (in
 this case it's the sound device), it can be pjmedia sound dev wrapper
 bug or the backend bug or even perhaps expected behaviour/limitation
 for that specific backend.

 BR,
 nanang


 On Wed, May 9, 2012 at 8:02 PM, Fabio Curi
 <fcuri@tecnologiaenseguridad.com>  <mailto:fcuri@tecnologiaenseguridad.com>  wrote:

 Tested in Ubuntu 11, with an not so old machine, same behavior.

 Regards.


 Fabio Curi wrote:

 I detect a problem with the conference facility in pjmedia.

 In a slow/old machine, when the CPU is working in another process (something
 hard like opening a browser or another big program) the conference process
 stop its execution.

 How to test:

 First, enable the #define CONF_DEBUG in conference.c and build.
 Second, start in a terminal (I use gnome-terminal) the
 pjproject-1.14/pjsip-apps/bin/samples/i686-pc-linux-gnu/confsample program.
 Then, you will see a lot of debug lines, make a long click or try to select
 some text in the terminal to stop/delay execution.

 At least in my machine, the program still works with no hangs, but the
 conference stop and the get_frame() is never called again.

 Same effect opening something like Firefox, OpenOffice, etc. during
 conference execution.

 Thank you.

 Regards.


 --
 ________________________________
 Fabio E. Curi
 Gerencia de I+D
 Tecnologia En Seguridad SRL
 (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992
 http://www.tecnologiaenseguridad.com


 _______________________________________________
 Visit our blog:http://blog.pjsip.org

 pjsip mailing list
 pjsip@lists.pjsip.org  <mailto:pjsip@lists.pjsip.org>
 http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org



 --
 ________________________________
 Fabio E. Curi
 Gerencia de I+D
 Tecnologia En Seguridad SRL
 (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992
 http://www.tecnologiaenseguridad.com

 _______________________________________________
 Visit our blog:http://blog.pjsip.org

 pjsip mailing list
 pjsip@lists.pjsip.org  <mailto:pjsip@lists.pjsip.org>
 http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

 _______________________________________________
 Visit our blog:http://blog.pjsip.org

 pjsip mailing list
 pjsip@lists.pjsip.org  <mailto:pjsip@lists.pjsip.org>
 http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

 -- 
 ------------------------------------------------------------------------
 Fabio E. Curi
 Gerencia de I+D
 Tecnologia En Seguridad SRL
 (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992
 http://www.tecnologiaenseguridad.com

 _______________________________________________
 Visit our blog: http://blog.pjsip.org

 pjsip mailing list
 pjsip@lists.pjsip.org <mailto:pjsip@lists.pjsip.org>
 http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

SOLVED. I try with the last portaudio version, but the sound is bad and needs debugging. After read this post, I apply the patch from the mail "sharing soundcard with ALSA problem still remains" (Dan Arrhenius) and compile with CFLAGS="-DPJMEDIA_AUDIO_DEV_HAS_ALSA" make The conference is working now, without stop or delays, even after various hours of test in my notebook. Tomorrow I will test in others PCs . Alsa did not work without the patch. May be is a good idea apply this patch in the main repository. Thank you. Regards. Benny Prijono wrote: > We are slowly moving away from PortAudio, so I would suggest trying > our native ALSA implementation instead. > > Best regards, > Benny > > On Mon, May 14, 2012 at 5:20 AM, Fabio Curi > <fcuri@tecnologiaenseguridad.com > <mailto:fcuri@tecnologiaenseguridad.com>> wrote: > > Nanang, > > The problem is in portaudio, after the cpu delay, the > PaAlsaStream_WaitForFrames() function in pa_linux_alsa.c never > exit from a the while( pollPlayback || pollCapture ) {}. Then the > capture/playback stops. > > I think this is a problem solved in the last portaudio version > (may be), then I will try to change the portaudio version in > pjmedia to test. It is possible? It is a trivial change or is too > complex? > > Thank you. > > Regards. > > > > Nanang Izzuddin wrote: >> Hi Fabio, >> >> As the clock is from sound device (conference itself actually doesn't >> have clock), so the problem seems to be in the clock generator (in >> this case it's the sound device), it can be pjmedia sound dev wrapper >> bug or the backend bug or even perhaps expected behaviour/limitation >> for that specific backend. >> >> BR, >> nanang >> >> >> On Wed, May 9, 2012 at 8:02 PM, Fabio Curi >> <fcuri@tecnologiaenseguridad.com> <mailto:fcuri@tecnologiaenseguridad.com> wrote: >>> Tested in Ubuntu 11, with an not so old machine, same behavior. >>> >>> Regards. >>> >>> >>> Fabio Curi wrote: >>> >>> I detect a problem with the conference facility in pjmedia. >>> >>> In a slow/old machine, when the CPU is working in another process (something >>> hard like opening a browser or another big program) the conference process >>> stop its execution. >>> >>> How to test: >>> >>> First, enable the #define CONF_DEBUG in conference.c and build. >>> Second, start in a terminal (I use gnome-terminal) the >>> pjproject-1.14/pjsip-apps/bin/samples/i686-pc-linux-gnu/confsample program. >>> Then, you will see a lot of debug lines, make a long click or try to select >>> some text in the terminal to stop/delay execution. >>> >>> At least in my machine, the program still works with no hangs, but the >>> conference stop and the get_frame() is never called again. >>> >>> Same effect opening something like Firefox, OpenOffice, etc. during >>> conference execution. >>> >>> Thank you. >>> >>> Regards. >>> >>> >>> -- >>> ________________________________ >>> Fabio E. Curi >>> Gerencia de I+D >>> Tecnologia En Seguridad SRL >>> (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 >>> http://www.tecnologiaenseguridad.com >>> >>> >>> _______________________________________________ >>> Visit our blog:http://blog.pjsip.org >>> >>> pjsip mailing list >>> pjsip@lists.pjsip.org <mailto:pjsip@lists.pjsip.org> >>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>> >>> >>> >>> -- >>> ________________________________ >>> Fabio E. Curi >>> Gerencia de I+D >>> Tecnologia En Seguridad SRL >>> (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 >>> http://www.tecnologiaenseguridad.com >>> >>> _______________________________________________ >>> Visit our blog:http://blog.pjsip.org >>> >>> pjsip mailing list >>> pjsip@lists.pjsip.org <mailto:pjsip@lists.pjsip.org> >>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>> >> _______________________________________________ >> Visit our blog:http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org <mailto:pjsip@lists.pjsip.org> >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> > > > -- > ------------------------------------------------------------------------ > Fabio E. Curi > Gerencia de I+D > Tecnologia En Seguridad SRL > (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 > http://www.tecnologiaenseguridad.com > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org <mailto:pjsip@lists.pjsip.org> > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org -- ------------------------------------------------------------------------ Fabio E. Curi Gerencia de I+D Tecnologia En Seguridad SRL (054 11) 5238-5553/54/58/59 int 26 - ID: 184*3992 http://www.tecnologiaenseguridad.com