GStreamer-devel

rx-pipeline

Classic

List

Threaded

5 messages Options

Michael Trimarchi

rx-pipeline

Hi all

I have created an rx-pipeline. The socket is allocated before and is assigned to the udpsrc

0:04:45.048620955 13554 0xb3d04660 LOG udpsrc gstudpsrc.c:493:gst_udpsrc_create:<udpsrc0> ioctl says 172 bytes available
0:04:45.048647774 13554 0xb3d04660 LOG udpsrc gstudpsrc.c:564:gst_udpsrc_create:<udpsrc0> read 172 bytes
priv_cb_pipeline_bus: Unhandled bus message from element rtp-pipeline (async-done).
priv_cb_pipeline_bus: Unhandled bus message from element audiosinkringbuffer1 (stream-status).
0:04:45.119837968 13554 0xb3d04660 LOG udpsrc gstudpsrc.c:493:gst_udpsrc_create:<udpsrc0> ioctl says 172 bytes available
0:04:45.119887625 13554 0xb3d04660 LOG udpsrc gstudpsrc.c:564:gst_udpsrc_create:<udpsrc0> read 172 bytes
0:04:45.119938400 13554 0xb3d04660 LOG udpsrc gstudpsrc.c:493:gst_udpsrc_create:<udpsrc0> ioctl says 172 bytes available
0:04:45.119960469 13554 0xb3d04660 LOG udpsrc gstudpsrc.c:564:gst_udpsrc_create:<udpsrc0> read 172 bytes
0:04:45.120002584 13554 0xb3d04660 LOG udpsrc gstudpsrc.c:493:gst_udpsrc_create:<udpsrc0> ioctl says 172 bytes available
0:04:45.120024095 13554 0xb3d04660 LOG udpsrc gstudpsrc.c:564:gst_udpsrc_create:<udpsrc0> read 172 bytes
0:04:45.120065929 13554 0xb3d04660 LOG udpsrc gstudpsrc.c:450:gst_udpsrc_create:<udpsrc0> doing select, timeout 0

I starts to receive packet and then stop.

udpsrc = gst_element_factory_make ("udpsrc", "udpsrc0");
assert (udpsrc != NULL);

g_object_set(G_OBJECT(self->sm_udpsrc),
"caps", caps,
"sockfd", sock,
"port", port,
NULL);

Then it's connected to a depayloader element a decelement and audiosink element.

There are no ERROR message in the log and seems that it stucks. Any idea?

Michael

------------------------------------------------------------------------------
The Next 800 Companies to Lead America's Growth: New Video Whitepaper
David G. Thomson, author of the best-selling book "Blueprint to a
Billion" shares his insights and actions to help propel your
business during the next growth cycle. Listen Now!
http://p.sf.net/sfu/SAP-dev2dev
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel

Marco Ballesio

Re: rx-pipeline

Hi,

On Fri, Nov 5, 2010 at 2:01 PM, Michael Trimarchi <[hidden email]> wrote:

Hi all

I have created an rx-pipeline. The socket is allocated before and is assigned to the udpsrc

..snip..

I starts to receive packet and then stop.

udpsrc = gst_element_factory_make ("udpsrc", "udpsrc0");
assert (udpsrc != NULL);

g_object_set(G_OBJECT(self->sm_udpsrc),
"caps", caps,
"sockfd", sock,
"port", port,
NULL);

Then it's connected to a depayloader element a decelement and audiosink element.

There are no ERROR message in the log and seems that it stucks. Any idea?

run the pipeline with gdb and get a bt for where it gets stuck. Alternatively, enable core dumps, send a SIGABRT and check the backtrace with gdb (may you can attach it to pastebin and post a link here).

Regards

Michael

------------------------------------------------------------------------------
The Next 800 Companies to Lead America's Growth: New Video Whitepaper
David G. Thomson, author of the best-selling book "Blueprint to a
Billion" shares his insights and actions to help propel your
business during the next growth cycle. Listen Now!
http://p.sf.net/sfu/SAP-dev2dev
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel

------------------------------------------------------------------------------
The Next 800 Companies to Lead America's Growth: New Video Whitepaper
David G. Thomson, author of the best-selling book "Blueprint to a
Billion" shares his insights and actions to help propel your
business during the next growth cycle. Listen Now!
http://p.sf.net/sfu/SAP-dev2dev
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gstreamer-devel

Michael Trimarchi

Re: rx-pipeline

On 11/05/2010 02:30 PM, Marco Ballesio wrote:

> Hi,
>
> On Fri, Nov 5, 2010 at 2:01 PM, Michael Trimarchi<
> [hidden email]> wrote:
>
>> Hi all
>>
>> I have created an rx-pipeline. The socket is allocated before and is
>> assigned to the udpsrc
>>
>
> ..snip..
>
> I starts to receive packet and then stop.
>>
>> udpsrc = gst_element_factory_make ("udpsrc", "udpsrc0");
>> assert (udpsrc != NULL);
>>
>> g_object_set(G_OBJECT(self->sm_udpsrc),
>> "caps", caps,
>> "sockfd", sock,
>> "port", port,
>> NULL);
>>
>> Then it's connected to a depayloader element a decelement and audiosink
>> element.
>>
>> There are no ERROR message in the log and seems that it stucks. Any idea?
>>
>
> run the pipeline with gdb and get a bt for where it gets stuck.
> Alternatively, enable core dumps, send a SIGABRT and check the backtrace
> with gdb (may you can attach it to pastebin and post a link here).

0:00:08.180251497 16201 0x805e128 DEBUG udpsrc gstudpsrc.c:837:gst_udpsrc_start:<udpsrc0> using provided socket 19
0:00:08.180343757 16201 0x805e128 DEBUG udpsrc gstudpsrc.c:873:gst_udpsrc_start:<udpsrc0> have udp buffer of 114688 bytes
0:00:08.180385033 16201 0x805e128 DEBUG udpsrc gstudpsrc.c:906:gst_udpsrc_start:<udpsrc0> bound, on port 16384
0:00:08.180558169 16201 0x805e128 LOG udpsrc gstudpsrc.c:992:gst_udpsrc_unlock_stop:<udpsrc0> No longer flushing
[New Thread 0xa3dd2b70 (LWP 16210)]
0:00:08.183153877 16201 0x805e128 LOG udpsrc gstudpsrc.c:992:gst_udpsrc_unlock_stop:<udpsrc0> No longer flushing

...

gst_udpsrc_create:<udpsrc0> ioctl says 172 bytes available
0:00:34.012355001 16201 0xb4400468 LOG udpsrc gstudpsrc.c:564:gst_udpsrc_create:<udpsrc0> read 172 bytes
0:00:34.036124526 16201 0xb4400468 LOG udpsrc gstudpsrc.c:493:gst_udpsrc_create:<udpsrc0> ioctl says 172 bytes available
0:00:34.036172576 16201 0xb4400468 LOG udpsrc gstudpsrc.c:564:gst_udpsrc_create:<udpsrc0> read 172 bytes
0:00:34.060245211 16201 0xb4400468 LOG udpsrc gstudpsrc.c:493:gst_udpsrc_create:<udpsrc0> ioctl says 172 bytes available
0:00:34.060299198 16201 0xb4400468 LOG udpsrc gstudpsrc.c:564:gst_udpsrc_create:<udpsrc0> read 172 bytes
0:00:34.084272869 16201 0xb4400468 LOG udpsrc gstudpsrc.c:450:gst_udpsrc_create:<udpsrc0> doing select, timeout 0
gst_udpsrc_create:<udpsrc0> doing select, timeout 0

452 ret = gst_poll_wait (udpsrc->fdset, timeout); <---------- seems to be block here
453 GST_LOG_OBJECT (udpsrc, "select returned %d", ret);
454 if (G_UNLIKELY (ret < 0)) {
455 if (errno == EBUSY)
456 goto stopped;
457 #ifdef G_OS_WIN32
458 if (WSAGetLastError () != WSAEINTR)
459 goto select_error;
460 #else
(gdb) l
461 if (errno != EAGAIN && errno != EINTR)
462 goto select_error;

Michael

>
> Regards
>
>
>>
>> Michael
>>
>>
>> ------------------------------------------------------------------------------
>> The Next 800 Companies to Lead America's Growth: New Video Whitepaper
>> David G. Thomson, author of the best-selling book "Blueprint to a
>> Billion" shares his insights and actions to help propel your
>> business during the next growth cycle. Listen Now!
>> http://p.sf.net/sfu/SAP-dev2dev
>> _______________________________________________
>> gstreamer-devel mailing list
>> [hidden email]
>> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel
>>
>
>
>
> ------------------------------------------------------------------------------
> The Next 800 Companies to Lead America's Growth: New Video Whitepaper
> David G. Thomson, author of the best-selling book "Blueprint to a
> Billion" shares his insights and actions to help propel your
> business during the next growth cycle. Listen Now!
> http://p.sf.net/sfu/SAP-dev2dev
>
>
>
> _______________________________________________
> gstreamer-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel

Marco Ballesio

Re: rx-pipeline

Hi,

so you had on the wire about 15.8 secs of bidirectional g711
PCMU data, undoubtedly more than what you got from your fd.

Maybe the fd is not connected to the proper socket (should be port 35364 or 7078, dunno what's the device you're running the pipeline onto).

Another guess: check whether you're setting the caps between udpsrc and depayloader to something like:

"application/x-rtp,media=audio,clock-rate=8000,encoding-name=PCMU,payload=0"

Regards

On Fri, Nov 5, 2010 at 6:43 PM, Michael Trimarchi <[hidden email]> wrote:

On 11/05/2010 05:21 PM, Marco Ballesio wrote:

tcpdump -s0 -i

Michael Trimarchi

Re: rx-pipeline

Hi
On 11/05/2010 07:59 PM, Marco Ballesio wrote:

> Hi,
>
> so you had on the wire about 15.8 secs of bidirectional g711
> PCMU data, undoubtedly more than what you got from your fd.
>
> Maybe the fd is not connected to the proper socket (should be port 35364 or
> 7078, dunno what's the device you're running the pipeline onto).
>
> Another guess: check whether you're setting the caps between udpsrc and
> depayloader to something like:
>
> "application/x-rtp,media=audio,clock-rate=8000,encoding-name=PCMU,payload=0"

You are right. My friend is fixed the problem during the weekend. The fd was not
connected to the right socked.

Michael

>
> Regards
>
> On Fri, Nov 5, 2010 at 6:43 PM, Michael Trimarchi<
> [hidden email]> wrote:
>
>> On 11/05/2010 05:21 PM, Marco Ballesio wrote:
>>
>>> tcpdump -s0 -i
>>>
>>
>>
>