Commit graph

1118 commits

Author SHA1 Message Date
Sjoerd Simons
8e43c0ec9e v4l2: Optimize negotiation by removing the query filter
As cameras tend to have a quite specific set of capabilities (specific
framerates for each resolution), getting the peer caps filtered by our
probed caps can cause a big increase in the caps size which slows down
things quire a bit.

As for negotiation v4l2 iterates through the caps of the peer to find the
first intersection with the probed caps, getting the fully expanded
intersection of capabilities is not useful.

Using the same testcase as for bug #702632, adding this patch on top of
the patches suggested there speeds up getting the inital frame from
around ~14-15 seconds to around ~3-4 seconds.

https://bugzilla.gnome.org/show_bug.cgi?id=702638
2013-06-19 15:05:25 +02:00
Andoni Morales Alastruey
025f226e60 osxvideosink: fix support in VM's without hardware acceleration 2013-06-18 10:41:37 +02:00
Michael Olbrich
a1c34b5407 v4l2: make sure the element is not deleted before the pool
The pool accesses data from the v4l2object so it must exist at least
as long as the pool. Refcount the element which controls the object
live-time.

https://bugzilla.gnome.org/show_bug.cgi?id=701650
2013-06-07 20:07:42 +02:00
Michael Olbrich
0769a78dd1 v4l2: iterate controls with V4L2_CTRL_FLAG_NEXT_CTRL if possible
In v2.6.18 control classes where added to the v4l2 API.
Iterating over CIDs starting with V4L2_CID_BASE will only find controls for
the first control class.
By iterating with V4L2_CTRL_FLAG_NEXT_CTRL all controls are found.

This is necessary to make controls from other control classes available in
the extra-controls property.

If V4L2_CTRL_FLAG_NEXT_CTRL is not defined at compile time or not supported
at runtime then the old mechanism for iterating is used.

https://bugzilla.gnome.org/show_bug.cgi?id=701540
2013-06-05 13:00:21 +02:00
Michael Olbrich
74d217c44b v4l2: improve pixel aspect ratio handling
Instead of just assuming a aspect ratio of 1/1 use VIDIOC_CROPCAP to ask
the device.
This also add a pixel-aspect-ratio property to overwrite the value from the
driver and a force-aspect-ratio property to ignore it.

https://bugzilla.gnome.org/show_bug.cgi?id=700285
2013-06-04 17:37:15 +02:00
Stirling Westrup
86405d6ee7 v4l2: Fix compilation with older kernels
https://bugzilla.gnome.org/show_bug.cgi?id=701595
2013-06-04 17:04:11 +02:00
Michael Olbrich
3417a791ac v4l2: call VIDIOC_REQBUFS with count = 0 in pool_finalize
Without this the following sequence fails:

- set_caps()
  - object_stop() (does nothing)
  - set_format() -> VIDIOC_S_FMT
- set_config() -> VIDIOC_REQBUFS with count = N
- set_caps()
  - object_stop()
    - pool_finalize()
  - set_format() -> VIDIOC_S_FMT => EBUSY

Usually the pool is started after set_config(), in which case object_stop()
will result in a pool_stop and therefore VIDIOC_REQBUFS with count = 0 but
that is not guaranteed.
Also calling VIDIOC_REQBUFS with count = 0 in pool_finalize() if necessary
fixes this problem.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701543
2013-06-04 10:45:06 +02:00
Michael Olbrich
496995a7d5 v4l2: rework sink buffer refcounting
This is a followup patch for #700781, which is not quite correct.
The buffer handling is quite complicated here.
The original code intended to the the following:

- gst_v4l2_buffer_pool_process() calls QBUF and adds the buffer to the
  local list.
- The sink calls gst_buffer_unref() which returns the buffer to the pool
  but not the 'free list'.
- Some time later DQBUF returns the buffer and
  gst_v4l2_buffer_pool_release_buffer() puts in on the 'free list'.

If the buffer must be copied then (parent_class)->acquire_buffer() is
called directly to keep the buffer in the pool.

This has two problems:
1. If gst_v4l2_buffer_pool_release_buffer() is called before the buffer is
   returned to the pool, then the buffer is put on the 'free list' twice.
   This can happen if a reference to the buffer is kept outside the sink,
   of if DQBUF returns the buffer, that was just queued with QBUF.
2. If buffers are copied, then all buffers are in the pool at all times. As
   a result gst_v4l2_buffer_pool_stop() and gst_v4l2_buffer_pool_dqbuf()
   can access pool->buffers at the same time, which can lead to memory
   corruption.

The patch for #700781 fixes those problems, but with the side effect that
there are always buffers outside the pool (because they are queued) and
the pool is never stopped.
This patch fixes this by releasing the reference to the buffer after
handling it (to avoid problem 2.) so it can be returned to the pool.
gst_v4l2_buffer_pool_release_buffer() is only called if the buffer is
already in the pool (to avoid problem 1.).

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701375
2013-06-03 11:55:48 +02:00
Sebastian Dröge
82e7826af0 v4l2: Only conditionally use V4L2_CTRL_TYPE_INTEGER_MENU, it's not available in older versions 2013-05-29 20:35:41 +02:00
Michael Olbrich
0fb59275b0 v4l2: add a property for arbitrary v4l2 controls
This makes it possible to set any controls that can be set with
VIDIOC_S_CTRL.
The controls are set when the property is set (if the device is open)
and when the device is opened.

https://bugzilla.gnome.org/show_bug.cgi?id=698837
2013-05-29 20:18:11 +02:00
Michael Olbrich
f9cc655adf v4l2: keep a reference to all queued buffers
Without this, a queued buffer may be required, filled and queued before it
is dequeued.
Calling gst_buffer_pool_acquire_buffer() ensures that the buffer is set up
correctly and gst_buffer_unref() calls buffer_release().

https://bugzilla.gnome.org/show_bug.cgi?id=700781
2013-05-21 16:44:58 +02:00
Michael Olbrich
00ffe41e6a v4l2: Don't stop streaming when set_caps is called with unchanged caps
This can happen if other parts of the pipeline are reconfigured.
Stop streaming even for a short amount of time can be quite visible, so it
should be avoided if possible.

https://bugzilla.gnome.org/show_bug.cgi?id=700503
2013-05-20 11:13:18 +02:00
Michael Olbrich
57c0f4d6b5 v4l2: only add interlace-mode to the caps for raw formats
https://bugzilla.gnome.org/show_bug.cgi?id=700280
2013-05-14 13:01:27 +02:00
Michael Olbrich
16b0fd5ad9 v4l2: copy and set the actual size of the content
https://bugzilla.gnome.org/show_bug.cgi?id=700282
2013-05-14 12:57:14 +02:00
Michael Olbrich
113b60935a v4l2: request 0 buffers when stopping
Without this stopping the pool in *_set_caps() is useless.
S_FMT will still fail with EBUSY.

https://bugzilla.gnome.org/show_bug.cgi?id=699835
2013-05-07 16:39:41 +02:00
Philipp Zabel
59d7d5c6bb v4l2: fill out v4l2_buffer.bytesused field for v4l2sink
When queuing a buffer for a sink, bytesused must contain the actual
amount of data.

For a source, the driver must overwrite this, so it doesn't matter
what is set here.

https://bugzilla.gnome.org/show_bug.cgi?id=699598
2013-05-04 11:19:14 +01:00
Michael Olbrich
53da901cc9 v4l2: always generate video info from caps
In the past gst_video_info_from_caps() only video/x-raw. Now it also
supports other video/* and image/* formats.
With this patch the format won't be GST_VIDEO_FORMAT_UNKOWN and
gst_v4l2_buffer_pool_set_config() handles strides correctly.

https://bugzilla.gnome.org/show_bug.cgi?id=699570
2013-05-03 13:14:58 +02:00
Michael Olbrich
1e777ede92 v4l2: try to allocate new buffers with VIDIOC_CREATE_BUFS if needed
If max_buffers is 0 then an arbitrary number of buffers (currently 4) is
allocated. If this is not enough v4l2src starts copying buffers.
With this patch VIDIOC_CREATE_BUFS is used to allocate a new buffer. If
this fails v4l2src falls back to copying buffers.

https://bugzilla.gnome.org/show_bug.cgi?id=699447
2013-05-03 11:18:04 +02:00
Andoni Morales Alastruey
bb980775f8 osxvideosink: fix setting window handle after transition
The destroyed flag was not reset properly and it's also not needed
as we can check osxwindow != NULL
2013-05-02 14:03:43 +02:00
Michael Olbrich
eebe5fc70d v4l2: fix and cleanup VIDIOC_EXPBUF handling
clear the struct, and provide a correct error message

https://bugzilla.gnome.org/show_bug.cgi?id=699337
2013-05-01 15:40:55 +02:00
Philipp Zabel
4d3f6850ca v4l2: handle return value -ENOTTY for unimplemented VIDIOC_G_PARM
Newer kernels return -ENOTTY, older kernels return -EINVAL if the ioctl
is not implemented. With this patch, GStreamer handles both cases.

https://bugzilla.gnome.org/show_bug.cgi?id=698825
2013-05-01 15:34:50 +02:00
Michael Olbrich
bf3a0b7e37 v4l2: fix broken boolean expression to detect non-frame buffers
https://bugzilla.gnome.org/show_bug.cgi?id=699294
2013-04-30 11:47:09 +02:00
Michael Olbrich
d4126ff986 v4l2: also poll for output devices
Note that the V4L2 API defines that for output devices POLLOUT
indicates that a buffer is ready to be dequeued.

https://bugzilla.gnome.org/show_bug.cgi?id=698992
2013-04-27 09:31:29 +01:00
Philipp Zabel
fd0123800c v4l2: fix copying of encoded buffers
The existence of a GstVideoFormatInfo does not guarantee, that
the buffer contains video frames, so the format must be checked.

Also, for encoded buffers the length is variable and must be set.

https://bugzilla.gnome.org/show_bug.cgi?id=698949
2013-04-26 15:36:49 +02:00
Michael Olbrich
3466796f10 v4l2: add support for mpeg4 and H.263
https://bugzilla.gnome.org/show_bug.cgi?id=698826
2013-04-26 14:26:15 +02:00
Philipp Zabel
305023fe9d v4l2: handle ENODATA return value for VIDIOC_ENUMSTD
In kernel v3.7-rc1, VIDIOC_ENUMSTD returns ENODATA if the current input
does not support the STD API.

https://bugzilla.gnome.org/show_bug.cgi?id=698827
2013-04-25 15:40:59 +02:00
Michael Olbrich
2e80c0d2c0 v4l2: bufferpool: reset buffer size in release_buffer
The buffer might still be in use elsewhere when dequeuing buffers for
outputs.

https://bugzilla.gnome.org/show_bug.cgi?id=698822
2013-04-25 12:34:24 +02:00
Michael Olbrich
8f9cc96a69 v4l2: bufferpool: remove unused includes
The hacks that needed these are long gone.

https://bugzilla.gnome.org/show_bug.cgi?id=698821
2013-04-25 11:30:48 +01:00
Todd Agulnick
a30634a248 osxaudio: Use gst_audio_channel_positions_to_mask() to create mask
https://bugzilla.gnome.org/show_bug.cgi?id=698807
2013-04-25 09:57:26 +02:00
Todd Agulnick
e7710c2655 osxaudio: Remove unused code 2013-04-25 09:56:57 +02:00
Diogo Carbonera Luvizon
0a7d4afc1b v4l2: save the format correctly
If TRY_FMT is not implemented,  gst_v4l2_object_get_nearest_size will
use S_FMT and will change the device's operation mode. To save the
old device mode we need to set the type field or else it will fail
to save the previous format.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=685209
2013-04-24 16:00:24 +02:00
Sebastian Dröge
3230182a61 osxaudio: Include gstaudioringbuffer.h to fix compilation in 1.0 2013-04-18 09:37:30 +02:00
Philippe Normand
d352201a34 osxaudiosink: channel-mask configuration fixes
Set channel-mask according to sink's layout in case of stereo layout.
Also initialize and reset the mask when an unrecognized channel is detected.

https://bugzilla.gnome.org/show_bug.cgi?id=698224
2013-04-18 09:16:19 +02:00
Olivier Crête
b5a9ed20e1 v4l2src: Disable renegotiation in the negotiate method
This way, we don't block the initial negotiation.

Thanks to Jeremy Whiting for doing all the testing.

https://bugzilla.gnome.org/show_bug.cgi?id=695981
2013-04-17 17:04:11 -04:00
Olivier Crête
3340baa374 Revert "v4l2: disable renegotiation"
This reverts commit d1b26e1d59.

This causes the initial negotiation to never happen if a reconfigure
event is received after gst_base_src_start_complete() but before the loop
starts.

https://bugzilla.gnome.org/show_bug.cgi?id=695981
2013-04-17 17:02:08 -04:00
Sebastian Dröge
27248393e4 osxaudio: Fix merge conflicts 2013-04-17 20:24:48 +02:00
Sebastian Dröge
2b1f967101 osxaudioringbuffer: First check the type, then cast 2013-04-17 09:52:14 +02:00
Takashi Nakajima
ce5246ed71 osxaudio: use GST_IS_OSX_AUDIO_SINK in ring buffer. 2013-04-17 09:52:14 +02:00
Takashi Nakajima
efda79b084 osxaudio: call set_channel_positions() in osxaudioringbuffer acquire() 2013-04-17 09:52:14 +02:00
Todd Agulnick
c2c85a094a osxaudio: use GST_AUDIO_INFO_* accessors
Changes include the following:

 * Update classname references
 * Replace GST_BOILERPLATE_FULL with G_DEFINE_TYPE
 * Use new GstAudioInfo struct and methods
 * Use new buffer memory allocation scheme

Conflicts:
	sys/osxaudio/gstosxaudioringbuffer.c
2013-04-17 09:52:13 +02:00
Todd Agulnick
ed94ef79f9 osxaudio: adjust for changes to glib mutex api. 2013-04-17 09:52:13 +02:00
Takashi Nakajima
09e980d2c9 osxaudio: try to fix up according to Sebastian's comments 2013-04-17 09:52:13 +02:00
Philippe Normand
ab64837bf2 osxaudio: build fixes
Enable the osxaudio plugin build in configure.ac and fix some
include directive order issues.
2013-04-17 09:52:13 +02:00
ted-n
d5d53ec611 osxaudio: fix layout for osxaudiosrc 2013-04-17 09:52:13 +02:00
ted-n
b217b6fdfb osxaudio: port to v.1.0 2013-04-17 09:52:13 +02:00
Andoni Morales Alastruey
baa23d314d osxvideosink: fix segfault accessing osxwindow when not set yet 2013-04-16 14:54:06 +02:00
Sebastian Dröge
b0b0557c48 gst: Add better support for static plugins 2013-04-15 15:54:11 +02:00
Josep Torra
0d6602eaaf osxaudio: Fixes error: "GST_LEVEL_DEFAULT" redefined 2013-04-15 08:41:34 +02:00
Josep Torra
220a9ff102 osxaudio: fixes implicit declaration of function 'getpid' 2013-04-15 08:41:28 +02:00
Todd Agulnick
a30476497f osxvideo: include pthread.h to fix compiler warning
https://bugzilla.gnome.org/show_bug.cgi?id=697303
2013-04-05 09:34:23 +01:00