Commit graph

18176 commits

Author SHA1 Message Date
Mathieu Duponchelle
280d86a841 rtpsession: Add disable-sr-timestamp property
The Onvif Streaming Spec, in section 6.11, mandates that when
Rate-Control is disabled potential RTCP packets shall have
their timestamps set to 0.

<https://www.onvif.org/specs/stream/ONVIF-Streaming-Spec.pdf>
2019-04-05 20:23:08 +02:00
Philipp Zabel
603c7a52fd v4l2: remove __user define from types-compat.h
Remove the now unused __user define.
2019-04-05 16:01:55 +00:00
Philipp Zabel
40db843efd v4l2object: use opRGB colorspace and xfer func defines
AdobeRGB defines have been renamed to opRGB in the kernel headers,
use the new names.
2019-04-05 16:01:55 +00:00
Philipp Zabel
e2ff87732d v4l2videodec: support orphaning
Recent kernels allow REQBUFS(0) on a queue that still has buffers in
use (mmapped or exported via dmabuf), orphaning all buffers on the queue.
If this is supported, the v4l2videodec element does not have to send a
drain request downstream.
2019-04-05 16:01:55 +00:00
Philipp Zabel
dfc20013cc v4l2bufferpool: support orphaning
Now that the v4l2allocator allows orphaning the V4L2 buffer queue, add
support for orphaning in the v4l2bufferpool. gst_v4l2_buffer_pool_orphan
can be used as a replacement for gst_v4l2_buffer_pool_stop, without
having to wait for buffers to be returned to the pool.
2019-04-05 16:01:55 +00:00
Philipp Zabel
0948ce0478 v4l2allocator: support orphaning
Recent kernels allow REQBUFS(0) on a queue that still has buffers in
use (mmapped or exported via dmabuf), orphaning all buffers on the queue.

Orphaning the allocator causes it to release all buffers with
REQBUFS(0), even if they are still in use. An orphaned allocator can
only be stopped. It can not be restarted or create new buffers.
2019-04-05 16:01:55 +00:00
Philipp Zabel
5d904530c1 v4l2: update kernel headers to latest from media tree
Update to the latest installed headers (output of make headers_install)
from the media tree, keeping the slight modifications to the includes.

This includes new HEVC controls, the AdobeRGB -> opRGB rename, a new
capabilities field for v4l2_requestbuffers and v4l2_create_buffers, new
32-bit YUV formats, and request_fd changes.
2019-04-05 16:01:55 +00:00
Nicolas Dufresne
bac8c00b85 shout2: Fix leak on error in start 2019-04-05 14:53:09 +00:00
Nicolas Dufresne
464ada3f29 test: rtpsession: Test FB Nack packing
We used to split the NACK if a smaller seqnum of a range of seqnum was
submited. This test also make sure that the three operations (append,
prepend, update) works properly.
2019-04-05 14:53:09 +00:00
Nicolas Dufresne
a31569713f test: rtpsession: Test handling of NACK surplus
This test verify that NACKs that didn't fit in one packet are properly
filtered and inserted into the following pipeline.
2019-04-05 14:53:09 +00:00
Nicolas Dufresne
6bb53e75fb rtpsession: Send as many nack seqnum as possible
In order to do that, we now split the nacks registration from the actual
FB nack packet construction. We then try and add as many FB Nacks as
possible into the active packets and leave the remaining seqnums in the
RTPSource. In order to avoid sending outdated NACK later on, we save the
seqnum calculated deadline and cleanup the outdated seqnums before the
next RTCP send.

Fixes #583
2019-04-05 14:53:09 +00:00
John Bassett
74a74bfc99 rtpsession: Fix race when sending PLI, FIR and NACK packets
Calling rtp_session_send_rtcp before marking the source as requiring a
pli/fir/nack meant the rtcp_thread could be scheduled and start running
before the source was updated. This meant the request would not be sent
early but instead was transmitted with the next regular RTCP packet.

Add test for nack generation.
2019-04-05 14:53:09 +00:00
Nicolas Dufresne
6b50d142f3 rtpsession: Fix early rtcp time comparision
If the current time is equal to the early rtcp time deadline, there is
no need to schedule a timer. This ensure that immediate feedback is
really immediate and simplify implementing unit tests with the test
clock, which stops perfectly on the timeout time.

This fix has been extracted from Pexip feature patch called
  "rtpsession: Allow instant transmission of RTCP packets"
2019-04-05 14:53:09 +00:00
Guillaume Desmottes
e3c4b80750 v4l2src: preserve features when fixating caps
The caps features were lost when sorting caps structures in
gst_v4l2src_fixate(). This was breaking alternate as
GST_CAPS_FEATURE_FORMAT_INTERLACED was removed from the caps.
2019-04-05 13:22:58 +00:00
Mathieu Duponchelle
74e3eb1f1d rtpgstpay: Set DELTA_UNIT flag when appropriate
When used in combination with a rtponviftimestamp element
downstream, forwarding this flag ensures it gets correctly
serialized in the ONVIF header extension.
2019-04-04 19:08:23 +02:00
Antonio Ospite
435f67debf docs: fix typo s/abonormally/abnormally/ 2019-04-03 16:42:26 +02:00
Antonio Ospite
d6939c4031 docs: fix typo s/incomming/incoming/ 2019-04-03 16:38:56 +02:00
Antonio Ospite
f7c8317668 rtp: fix indentation after G_DEFINE_TYPE
A missing colon after G_DEFINE_TYPE declaration was confusing gst-indent
and causing problem in the pre-commit hook.

Add the missing colon and fix the following function declaration to
follow the normal GStreamer style.
2019-04-03 16:37:34 +02:00
Antonio Ospite
114de8cc96 rtpsession: fix comment to refer to buffers instead of groups
One comments in gst_rtp_session_chain_send_rtp_common() is referring to
groups in a buffer list, however this concept of "group" comes from
GStreamer 0.10 and does not exist anymore in GStreamer 1.0, so update the
comment to refer to buffers instead.
2019-04-02 13:03:56 +02:00
Antonio Ospite
e98b0ca8da rtpsource: add comment to explain why probation queue is not always cleared 2019-04-02 13:03:56 +02:00
Antonio Ospite
6f12f1cecc test: rtpbin_buffer_list: add test to verify that stats are correct
Add a test to verify that stats about sent and received packets are
correct even when using buffer lists.

NOTE: the newly introduced get_session_source_stats() selects the
desired source (sender or receiver) by filtering them by type (using the
get_sender parameter) rather than by ssrc because this simplifies the
code and it's good enough for testing purposes as there is usually one
source per type in the test setup.

Filtering by ssrc would have required handling asynchronous signals like
"on-new-sender-ssrc", with the relative locking, just to retrieve the
actual ssrc of the sender.
2019-04-02 13:02:28 +02:00
Antonio Ospite
0fae88b5fd rtpsource: fix stats about received packets
The update_receiver_stats() function is called also when sending packets
in rtp_source_send_rtp(), and sending packets may happen using a buffer
list rather than individual buffers.

So update the stats using the actual number of packets sent.

NOTE: this is fine for the receive path too (rtp_process_send_rtp)
because the receive path does not support buffer lists and
pinfo->packets would always be equal to 1 in this case.
2019-04-02 09:26:03 +02:00
Antonio Ospite
a330012d6a test: rtpbin_buffer_list: move buffer list creation next to its validation
The tests create a buffer list and then use the chain_list callback to
verify that the correct packets have been pushed.

Move the creation and validation code next to each other so that the
reader can more easily understand what is going on.

While at it add some comments to introduce the two related functions.
2019-04-01 18:42:32 +00:00
Antonio Ospite
af8698e656 test: rtpbin_buffer_list: set the chain_list function directly in the test
The helper function set_chain_function does not really do anything useful, remove it.
2019-04-01 18:42:32 +00:00
Antonio Ospite
5567066bff test: rtpbin_buffer_list: make check_packet more flexible
Make it possible to differentiate between the position in the list and
the packet index in the global structures in check_packet, in some
future case the list may change, in case some element removes a buffer
from the list, and the two indices may not coincide.
2019-04-01 18:42:32 +00:00
Antonio Ospite
5807d79a9d test: rtpbin_buffer_list: factor out a function to create packets buffers 2019-04-01 18:42:32 +00:00
Antonio Ospite
6ab13c19b6 test: rtpbin_buffer_list: check if the chain_list function has been called
Make the test more useful to verify that the chain list function has
actually been called.
2019-04-01 18:42:32 +00:00
Antonio Ospite
623d3b930e test: rtpbin_buffer_list: port to GStreamer 1.0
Port the rtpbin_buffer_list test to GStreamer 1.0 and re-enable it.

Some other changes include:
  - the check on the caps has been moved from the buffer level to the
    pad level;
  - remove underscore prefix from static functions names, this is not
    idiomatic in C and rarely used in the other tests;
  - the unused header_buffer variable has been removed;
  - check_group() has been renamed to check_packet() because in
    GStreamer 1.0 there is no concept of "group" anymore, the comments
    have also been updated to reflect this.
2019-04-01 18:42:32 +00:00
Tim-Philipp Müller
abef4e4ea3 tests: jpegdec: bump discoverer timeout for valgrind
Tests might take a bit longer, esp. when run under valgrind
and/or they're running on the CI with other things going on,
so let's just bump the timeout to something higher and let
the test runner time us out if needed.
2019-04-01 18:20:53 +01:00
Nirbheek Chauhan
f23724b8a6 meson: Only ensure that moc is available on Linux
On other OSes, it's not possible to have qmake or the qt5 pkg-config
files and not have moc, and `moc` will not be in `PATH`, so this only
causes problems.
2019-04-01 18:20:28 +05:30
Olivier Crête
915a9c99bb rtpstorage: Limit the queue size
Limit to the queue size in case there is no arrival time or in case there is
a huge flood of packets.
2019-03-29 22:51:54 +00:00
Olivier Crête
0ecc52c2ee rtpbin: Request the FEC decoder even if ignore-pt is set 2019-03-28 16:24:17 -04:00
Olivier Crête
c2dd263562 rtpbin: Factor out the code that exposes the src pad 2019-03-28 16:24:12 -04:00
Olivier Crête
8bf074f21e rtpreddec: Add some more debug prints 2019-03-27 18:54:27 -04:00
Olivier Crête
c840328664 rtpstorage: Issue warning if request by size if 0
If the size is 0, then nothing will ever be in the storage, if a request is
received, it generally implies a misconfigured pipeline.
2019-03-26 19:41:06 -04:00
Olivier Crête
7a317ff732 rtpstorage: Add more debug messages 2019-03-26 19:41:06 -04:00
Olivier Crête
785219a317 rtpstorage: Make debug category available to sub objects 2019-03-26 19:41:06 -04:00
Olivier Crête
9b0a373eac rtpstorage: Add debug funcptr to chain function 2019-03-26 18:08:57 -04:00
Julian Bouzas
2ebdd70c21 flac: report latency in flacenc and flacdec
The FLAC specification states that the data is processed in blocks, regardless of the number of channels. Thus, The latency can be calculated using the blocksize and rate. For example a 1 second block sampled at 44.1KHz has a blocksize of 44100
2019-03-25 15:14:32 +01:00
Tim-Philipp Müller
d682c74c1e examples: rtsp: fix compiler warning
"control reaches end of non-void function"
2019-03-22 23:37:09 +00:00
Nicolas Dufresne
79fd0af152 gstrtpsession: Remove set but not use running-time 2019-03-22 20:01:52 +00:00
Olivier Crête
7ecbd7271d rtpmanager: Register chain functions to debug 2019-03-22 16:44:41 +00:00
Nicolas Dufresne
2ff7519d73 rtpbin: Allow reusing the sender AUX bin
This is needed for the case you don't know in advance all the sessions
you will be using, but would like to place all the related AUX element
in the same GstBin. As per current implementation, each time an sender
AUX bin is requested and returned, RTPBin will walk the src pads and
create sessions for these pads.

In the current implementation, if a src pad already have a sessions, it
returns an error and stops. As a side effect, if an AUX bin is reused in
a following AUX bin request, it can only work if the pads are created on
the last request.

This change simply relax the restriction in order to keep walking, and
just ensure that all newly created pads have a sessions.
2019-03-21 21:10:43 +00:00
Philipp Zabel
16e5b32bc1 v4l2videoenc: set GstVideoCodecFrame sync point flag
The V4L2 elements already set the delta unit buffer flag when dequeueing
the buffer, but gst_video_encoder_finish_frame overwrites it from the
passed codec frame's sync point flag. Set the flag correctly.
2019-03-21 18:05:51 +00:00
George Kiagiadakis
d5ce10240a gstrtpsession: improve stats about rtx requests 2019-03-21 13:40:31 -04:00
George Kiagiadakis
db647ee55b rtprtxsend: Improve looging of not found RTX packet
When an RTX packet is not found, display a message that say if the
packet have not arrived yet or if it was already removed from the RTX
packet queue.
2019-03-21 13:19:52 -04:00
Nicolas Dufresne
0aff8a7d30 rtpsession: Remove unused rtp_session_create_source 2019-03-21 13:19:52 -04:00
Tim-Philipp Müller
e8583cebe7 meson: add -Wno-unused also to C++ args when gst debug system is disabled
And check if argument is supported instead of just passing it blindly,
and make meson code slightly cleaner, centralising the argument setting
in one place.
2019-03-21 16:45:03 +00:00
Piotr Drąg
a0474578a0 Update LINGUAS 2019-03-21 15:59:31 +00:00
Seungha Yang
63bb1e3a4d qtdemux: Don't pass zero to denominator for framerate
Need to respect return of gst_video_guess_framerate() to ensure
non-zero denominator.

This patch is to fix below error with an abnormal (but has valid frame) file.
(gst-play-1.0:17940): GStreamer-CRITICAL **: passed '0' as denominator for `GstFraction'
2019-03-19 12:35:08 +09:00