Commit graph

119517 commits

Author SHA1 Message Date
Piotr Brzeziński
4020144a65 vtdec: Fix caps criticals during negotiation
Calling gst_pad_peer_query_caps() without a filter can give us EMPTY caps, whereas all the code below
assumes that's not the case. Replacing query+intersect with a filtered query ensures we always get a subset
of the template caps back.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6454>
2024-03-27 16:24:07 +00:00
Víctor Manuel Jáquez Leal
6eab0524ca jpegparse: avi1 tag can be progressive
AVI1 tag in APP0 is trivalue: 0 not interleaved, 1 odd, 2 even.

So if avi1 is zero then the frame is progressive.

Also, this patch adds a couple log messages.

Fixes: #3414
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6453>
2024-03-27 14:59:38 +00:00
Tim-Philipp Müller
d2f20d546d tests: add check to make sure -bad lib headers are C++ compiler clean
Only non-internal libs without external deps for now.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6448>
2024-03-26 16:15:12 +00:00
Tim-Philipp Müller
fe8b80704a ges: add check to make sure headers are C++ compiler clean
See https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/3421

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6448>
2024-03-26 16:15:12 +00:00
Thibault Saunier
256d990aea ges: frame-composition-meta: Stop using keyword 'operator' for field in C++
Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/3421

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6448>
2024-03-26 16:15:12 +00:00
Wojciech Kapsa
7a8663d051 libnice: bump subproject wrap to v0.1.22
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6431>
2024-03-22 21:37:41 +00:00
Hou Qi
c0d35575a8 v4l2: Also set max_width/max_height if enum framesize fail
Some driver doesn't implement enum_framesize. The maximum supported
size can be got by trying format with a very large size. Also need
to set max_width/max_height for this case, otherwise default encoded
buffer size 256kB is too small.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6430>
2024-03-22 17:27:58 +00:00
Tim-Philipp Müller
175d116f94 Back to development 2024-03-22 01:38:06 +01:00
Tim-Philipp Müller
0d0a1d9d16 Release 1.24.1 2024-03-21 21:47:53 +01:00
Thomas Goodwin
5a80a146f0 gst-inspect: fix --atleast-version to be implicitly applied to --exists
The --atleast-version implies --exists, but the implementation in
earlier commits had the version check applied any time the --exists was
checked, and the default value of the major and minor versions were set
to the GStreamer major and minor versions.  The resulting behavior would
have gst-inspect return '1' if the plugin's version didn't match
gstreamer's even when --atleast-version was not specified in the command
line args.  The change in this patch removes that behavior and adds
tests to verify that if --exists is specified WITHOUT --atleast-version
the version check will NOT be applied.  If both arguments are specified
and the version does not match the arg-supplied version number, a new
return code of '2' is used to uniquely identify the failure.

Fixes #3246

Signed-off-by: Thomas Goodwin <thomas.goodwin@laerdal.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6414>
2024-03-21 00:13:59 +01:00
Piotr Brzeziński
11c6432cca vtdec: Ignore output loop errors in drain() if we're flushing
In an early non-linked scenario, this was causing a ton of criticals about the queue array,
because the output callback would still fire for leftover frames that were still being processed by VT
at the time the output loop stopped. This makes sure they're flushed correctly as well.

Also renames gst_vtdec_loop to gst_vtdec_output_loop for consistency with related functions.

wip

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6411>
2024-03-20 14:21:24 +01:00
Piotr Brzeziński
bf88fb152d vtdec: Fix a deadlock during ProRes playback
Sometimes a call to negotiate (and thus drain) can happen from the output loop
(via finish_frame()), which will tell VT to output all internal frames, but that won't succeed
if we happen to decide to wait for the queue to empty (because the loop is waiting for draining to finish and
will not make space in the queue!). This commit adds an override for the queue size limit if we're draining/flushing.

This bug could happen for any formats, but was especially obvious for ProRes, which has dpb_size of 0.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6411>
2024-03-20 14:21:24 +01:00
Edward Hervey
db6803bd55 adaptivedemux2: Don't use g_str_equal on potentially NULL strings
It is only meant to be used as a callback. The fallback macro uses strcmp which
doesn't handle NULL strings gracefully. Instead use g_strcmp0

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6410>
2024-03-20 10:49:02 +01:00
Edward Hervey
77fa0ae0e7 hlsdemux2: Avoid NULL pointer usage
The pending/current variant are both NULL when the demuxer is resetted.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6410>
2024-03-20 10:49:02 +01:00
Edward Hervey
9ce063f5f6 adaptivedemux2: Handle context going away
This issue can happen when the scheduler loop was stopped (and context went
away). We no longer want to push/pop main context threads.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6410>
2024-03-20 10:49:02 +01:00
Edward Hervey
257de579b3 hlsdemux2: Improve detection of playlist updates
In the case we are not updating an existing playlist, we only want to reset the
download error count if the URI we are downloading was not the previous one we
were trying to load

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6410>
2024-03-20 10:49:02 +01:00
Edward Hervey
f8d8c6795d uridecodebin3: Don't hold lock when posting messages or signals
There's a very good chance that the receiver might react on those synchronously
and call back into uridecodebin3 (ex: for setting the next URI).

Make sure we release the lock if we need to do that.

Fixes #3400

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6403>
2024-03-19 12:02:27 +01:00
Ruijing Dong
b547c8eebb va: enc : checking surface alignment attribute
Apply surface alignment attribute when availalbe,
also fix frame cropping issue for va h265 encoder.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6399>
2024-03-18 20:55:38 +01:00
Edward Hervey
c8f42ab3af uridecodebin3: Handle potential double redirection errors
Some elements (like qtdemux) might post a redirection error message twice. We
only want to handle it once.

Fixes #3390

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6394>
2024-03-18 15:52:08 +01:00
Mark Nauwelaerts
f8ae970db2 dvdspu: avoid null dereference
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6391>
2024-03-18 11:43:39 +01:00
Philippe Normand
559278420b play: Fix a critical warning in error callback
`on_error()` can be called with a NULL details structure, so in that situation
the `gst_structure_copy()` would raise a critical warning. Create an empty
structure instead of attempting to copy a NULL one.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6387>
2024-03-17 21:43:44 +01:00
Tim-Philipp Müller
93255efece Revert "audiobasesink: Don't wait on gap events"
This reverts commit 8e923a8e2d.

This caused regressions, see #3303.

Without this commit, osxaudiosrc ! osxaudiosink won't work
right, but since that hasn't really been a huge problem
for years it's probably best to revert this until a proper
solution can be figured out.



(cherry picked from commit f04f86f3ee)

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6384>
2024-03-17 03:18:54 +00:00
Seungha Yang
9061e464a8 d3d12: Fix SDK debug layer activation
Debug layer must be enabled before creating device. Otherwise
already opened devices before the activation will be removed

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6383>
2024-03-16 15:58:13 +01:00
Edward Hervey
80cd85d03d decodebin3: Post error messages if there are no streams to output
This could happen because:
* No streams were selected
* Or we end up with no stream selected

Also post a warning message if we are missing plugins but there are other
streams to output

Fixes #3360

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6377>
2024-03-15 00:12:28 +00:00
Edward Hervey
2d875b5ed2 decodebin3: Remove failing stream from active selection also
It gets added in get_output_slot()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6377>
2024-03-15 00:12:28 +00:00
Robert Mader
aef872944e v4l2codecs: decoders: Add DMA_DRM caps support
In order to simplify caps negotiations for clients and, notably, be more
compatible with va* decoders.
Crucially this allows clients to know ahead of time whether buffers will
actually be DMABufs.

Similar to GstVaBaseDec we only announce system memory caps if the peer
has ANY caps. Further more, and again like va decoders, we fail in
`decide_allocation()` if DMA_DRM caps are used without VideoMeta.
Apart from buggy peers this can happen e.g. when a peer with ANY caps
is used in combination with caps filters.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6376>
2024-03-14 23:32:00 +00:00
Robert Mader
383545d856 v4l2codecs: decoders: Introduce and use set_output_state helper class
Allowing us to avoid some code duplication. This will become more
important with upcoming changes to caps generation.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6376>
2024-03-14 23:32:00 +00:00
Robert Mader
84d015cabb v4l2codecs: decoder: Clean up select_src_format()
Most importantly rely on video info helpers instead of manual parsing
of caps, which will allow us to use additional helpers in the future.

While on it, tighen the check for supported formats - failing that
indicates a bug in caps negotiation - and make some style changes.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6376>
2024-03-14 23:32:00 +00:00
Robert Mader
b5fe2319f3 v4l2codecs: decoder: Generalize size enumeration caps
By reducing the generated caps to the minimal number of fields and
using intersections instead of merges. This will allow us to reuse the
result in the future.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6376>
2024-03-14 23:32:00 +00:00
Robert Mader
7b6d6fe080 v4l2codecs: decoders: Use src template for negotiation filter
This ensures we don't create filter caps that are not supported by the
individual codec implementations, as well as that the resulting caps
have the required fields so they can be turned into a GstVideoFormat.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6376>
2024-03-14 23:32:00 +00:00
Piotr Brzeziński
4295e1dd30 avaudenc: Avoid double-freeing frame's extended data
This occured when attempting to encode 16 channel audio, would crash on the first buffer.
We only need to store ext_data, old ext_data_array (frame->extended_data) is already freed by `av_frame_unref`.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6375>
2024-03-14 23:39:35 +01:00
Piotr Brzeziński
9048bea3d3 avcodecmap: Increase max AAC channels to 16
This is the maximum amount supported by aacenc. 8-channel output fully works.
16-channel also encodes fine, but codec-utils isn't able to parse its channel config,
so output level will not be shown in caps. For that to work, GASpecificConfig parsing
needs to be implemented. It's not a critical issue and can be worked on at a later date.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6375>
2024-03-14 23:39:35 +01:00
Seungha Yang
615b7ca7ca asio: Fix {input,output}-channels property handling
Fixing regression introduced by the commit 06dc931b52

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6372>
2024-03-14 13:46:28 +00:00
Sebastian Dröge
0fe3ffe0c0 ptp: Initialize expected DELAY_REQ seqnum to an invalid value
This allows distinguishing pending syncs that didn't have a DELAY_REQ
sent from ones that did but used a seqnum of 0, like the very first one.

Specifically, if the first one or more syncs are still pending and we
send the first DELAY_REQ for a later pending sync, then the DELAY_RESP
would've been wrongly associated to the very first pending sync because
of the seqnum.

Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/3383

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6365>
2024-03-14 01:36:28 +00:00
Seungha Yang
470bcd4ec9 d3d11device: Fix adapter LUID comparison in wrapped device mode
Fix integer type mismatching

Fixes: https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/3382
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6364>
2024-03-14 00:35:19 +00:00
Alexander Slobodeniuk
6e57362f35 rtspsrc: remove 'deprecated' flag from the 'push-backchannel-sample' signal
It seems that it was added by accident when copying from push-backchannel-buffer

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6363>
2024-03-13 21:14:27 +00:00
Thomas Klausner
8ca53a1a67 shmallocator: fix build on Illumos
Closes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/3370

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6362>
2024-03-13 20:09:44 +00:00
Sebastian Dröge
5a2f99a255 mpg123audiodec: Correctly handle the case of clipping all decoded samples
Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/3365

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6357>
2024-03-13 14:23:56 +00:00
Piotr Brzeziński
07e1f13e2c audiovisualizer: Don't wrap temporary memory in buffers
Avoids potentially ending up with the buffermemory pointing to already-freed or reused addresses.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6339>
2024-03-13 12:36:28 +00:00
Piotr Brzeziński
6b369d8470 qtdemux: Fix wrapping temporary memory in buffers
That memory can disappear at any moment, doesn't cost much to just copy those few bytes.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6339>
2024-03-13 12:36:28 +00:00
Piotr Brzeziński
eff6167d2d audioencoder: Avoid wrapping temporarily mapped memory with a GstBuffer and passing that to subclass
Memory from gst_adapter_map() could live shorter than the GstMemory that the GstBuffer wraps around it, which in lucky
cases 'just' caused a re-use of the same memory for multiple (potentially still in use!) input buffers, but could easily
end up pointing to an already-freed memory.

Manifested when an AudioToolbox encoder kept getting silence inserted in seemingly random circumstances, turned out
to be the memory being re-used by GStreamer at the same time that the AT API was processing it...

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6339>
2024-03-13 12:36:28 +00:00
Seungha Yang
1a7d34a0e4 d3d12device: Fix IDXGIFactory2 leak
factory passed to gst_d3d12_device_find_adapter() method is valid
handle already

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6353>
2024-03-13 11:32:01 +00:00
Antonio Larrosa
a7b128bb27 gitlint: Allow curly brackets in commit prefix
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6344>
2024-03-13 09:58:25 +00:00
Antonio Larrosa
c04cdddd37 va{h264,h265,av1}enc: fix potential crash on devices without rate control
This fixes a crash in `gst_va_h264_enc_class_init` and `gst_va_h265_enc_class_init`
(and probably also in gst_va_av1_enc_class_init) when calling
`g_object_class_install_properties (object_class, n_props, properties);`

When rate_control_type is 0, the following code is executed in :

```
  } else {
    n_props--;
    properties[PROP_RATE_CONTROL] = NULL;
  }
```

n_props has initially a value of N_PROPERTIES but PROP_RATE_CONTROL
is not the last element in the array, so it's making
g_object_class_install_properties fail to iterate over the
properties array.

This applies the same fix to gstvah264enc.c, gstvah265enc.c and
gstvaav1enc.c.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6344>
2024-03-13 09:58:24 +00:00
Sebastian Dröge
cee1a14bb2 videoparsers: Don't verbosely warn about CEA_708_PROCESS_EM_DATA_FLAG not being set
And the same for CEA_708_PROCESS_CC_DATA_FLAG. This is not really a
problem and was polluting logs with warnings for every single frame.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6352>
2024-03-13 08:06:06 +00:00
Nicolas Dufresne
8c850dba9e glupload: Do not propose allocators with sysmem
None of the GL allocators actually offer a generic alloc() implementation. As a
side effect, they cannot be offered as they don't work with generic video
buffer pool.

Our specialized buffer pool can be dropped by tee or alphacombine as sharing the
same buffer pool over two branch is not supported by the pool API.

Fixes #3372

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6350>
2024-03-13 03:10:20 +00:00
Seungha Yang
44d4eb096d cuda,d3d11,d3d12bufferpool: Disable preallocation
Do not chain up to parent's GstBufferPool::start() which will do
preallocation. We don't want it to be preallocated
since there are various cases where negotiated downstream buffer pool is
not used at all (e.g., zero-copy decoding, IPC elements).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6345>
2024-03-13 02:29:42 +00:00
Antonio Larrosa
89041ad29d registry, ptp: Canonicalize the library path returned by dladdr
On systems using UsrMerge (like openSUSE or Fedora), /lib64 is
a symlink to /usr/lib64. So dladdr is returning the path to
the gstreamer library in /lib64 in priv_gst_get_relocated_libgstreamer.
Later gst_plugin_loader_spawn tries to build the path to the
gst-plugin-scanner helper from /lib64 and ends up trying to use
/lib64/../libexec/gstreamer-1.0/gst-plugin-scanner which doesn't exist.

By canonicalizing the path with a call to realpath, gst-plugin-scanner
is found correctly under
/usr/lib64/../libexec/gstreamer-1.0/gst-plugin-scanner

Similar change applied to gstreamer/libs/gst/net/gstptpclock.c

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6343>
2024-03-12 23:54:59 +00:00
Nirbheek Chauhan
45ce449854 gsturi: Sort by feature name to break a feature rank tie
This matches autoplug in other places such as decodebin, otherwise we
will pick "randomly" based on the order in which plugins are
registered, which is mostly dependent on the order in which readdir()
returns items.

So let's make it predictable.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6342>
2024-03-12 19:02:40 +00:00
Edward Hervey
48e0c6218d playbin3: Remove un-needed URI NULL check
This will mimic the playbin2 behaviour, which sets the "next" entry to be
NULL.

The biggest impact this has is that when going back to READY the current play
entry will be discarded (instead of being kept around for when you go back to
PAUSED/PLAYING).

Fixes #3371

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6338>
2024-03-12 14:16:34 +00:00