| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
| |
struct buff_data contains a pointer to struct video_data, so passing the
file descriptor again is redundant.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
|
|
|
|
|
|
|
|
|
| |
In particular:
1) save errno before it (possibly) gets overwritten by other calls
2) do not forget to enqueue the buffer again in case of error
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* commit '1afddbe59e96af75f1c07605afc95615569f388f':
avpacket: use AVBuffer to allow refcounting the packets.
Conflicts:
libavcodec/avpacket.c
libavcodec/utils.c
libavdevice/v4l2.c
libavformat/avidec.c
libavformat/flacdec.c
libavformat/id3v2.c
libavformat/matroskaenc.c
libavformat/mux.c
libavformat/utils.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
|
| |
| |
| |
| |
| |
| | |
This will allow us to avoid copying the packets in many cases.
This breaks ABI.
|
| |
| |
| |
| | |
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
just one buffer left available
This will avoid the possibility that we dequeue more buffers than we
have obtained from the v4l2 driver.
Fixes ticket #1570
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This ensures that fps and other values are estimated correctly,
the probesize also is intended to limit disk/protocol reads which
does not apply to lavfi inputs at all.
Fixes Ticket1051
Something similar could be usefull to other input devices
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
|
| |
| |
| |
| | |
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| | |
Also cleanup exported symbols
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
|
| |
| |
| |
| |
| |
| |
| | |
Set ret to 0 before enumerating standards in v4l2_set_parameters(), avoid
use of uninitialized variable.
Regression introduced in 60950adc18fe145a235211e75da68ab07123fcaa.
|
| |
| |
| |
| |
| | |
In particular, avoid use of non-thread-safe strerror(), and store errno
before calling av_log().
|
| |
| |
| |
| | |
Signed-off-by: rogerdpack <rogerpack2005@gmail.com>
|
| |
| |
| |
| | |
Signed-off-by: rogerdpack <rogerpack2005@gmail.com>
|
| |
| |
| |
| | |
Signed-off-by: rogerdpack <rogerpack2005@gmail.com>
|
|\ \ |
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Generally speaking, there are two types of v4l2 devices [1]:
1) devices that support a standard, like PAL or NTFS (tv cards, for example). For
this class of devices the framerate is fixed by the standard (for example PAL uses
25 fps) and the v4l2 driver cannot usually negotiate a different framerate (unless
it can skip frames on the driver side, to save I/O bandwidth).
2) devices for which the notion of standard does not make sense (webcams, for example).
For these devices it is usually possibile to request a desidered framerate.
In either case, the desidered frame rate can be requested when the VIDIOC_G_PARM
ioctl returns the V4L2_CAP_TIMEPERFRAME flag in the capability field.
Currently the code does not check for V4L2_CAP_TIMEPERFRAME and supports only the
second category of devices, returning a time per frame of 0/0 for devices in the
first group that do not permit to negotiate the framerate.
This patch adds support to read the correct framerate in all cases.
[1] http://linuxtv.org/downloads/v4l-dvb-apis/standard.html
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Since the user is expected to choose the standard by name (with -standard
option), add the possibility to list all the supported standards.
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>
|
| | |
| | |
| | |
| | | |
Signed-off-by: rogerdpack <rogerpack2005@gmail.com>
|
|/ /
| |
| |
| | |
Signed-off-by: rogerdpack <rogerpack2005@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Because libavdevice/v4l2 is a demuxer, it makes sense to look for
decoders instead of encoders when listing the formats supported by the
device.
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
After opening the device, the first thing we should do is selecting the
input. This is because the image formats (VIDIOC_ENUM_FMT ioctl) and the
standards (VIDIOC_ENUMSTD ioctl) supported may depend on the selected
input ([1] and [2]).
[1] http://linuxtv.org/downloads/v4l-dvb-apis/vidioc-enum-fmt.html
[2] http://linuxtv.org/downloads/v4l-dvb-apis/vidioc-enumstd.html
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
errors
Make device_try_init() return an error value, and allow to properly
report errors which are not due to a wrong pixel/codec configuration.
In particular, report intelligible feedback in case of busy device.
|
|\|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* commit '80ac87c13dc8c6c063e26a464c5c542357c0583f':
lavc: support ZenoXVID custom tag
libcdio: support recent cdio-paranoia
float_dsp: Add #ifdef HAVE_INLINE_ASM around vector_fmul_window
theora: Skip zero-sized headers
Conflicts:
configure
Merged-by: Michael Niedermayer <michaelni@gmx.at>
|
| |
| |
| |
| |
| | |
Upstream decided to split the paranoia interface and move the headers
accordingly.
|
| |
| |
| |
| | |
Fix a warning and random failures.
|
| |
| |
| |
| | |
Reviewed-by: Stefano Sabatini
|
| |
| |
| |
| | |
Reviewed-by: Stefano Sabatini
|
| | |
|
| |
| |
| |
| | |
In particular, propagate errno value in case of device init failure.
|
| | |
|
| | |
|
|\|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* commit '90cfc084e3e6d37ab88fc96a95f0401d8e8b4cd1':
avpacket: free side data in av_free_packet().
v4l2: do not assert on a value received from outside of Libav
v4l2: set the average framerate instead of codec timebase.
Conflicts:
libavcodec/avpacket.c
libavdevice/v4l2.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
|
| | |
|
| |
| |
| |
| | |
Codec timebase is supposed to be set by decoders only.
|
|\|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* commit '246da0b13551b1f80f067e4f258e5bd691f5ab33':
v4l2: avoid pointless indirection.
vdpau: Add VC-1 decoding via hwaccel infrastructure
vdpau: Add H.264 decoding via hwaccel infrastructure
Conflicts:
configure
libavcodec/Makefile
libavcodec/version.h
libavdevice/v4l2.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
|
| |
| |
| |
| |
| | |
v4l2_read_header() does no cleanup, so it can return directly, without
any need for goto.
|
|\|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* commit 'a0b7e289075dccf223b7f407790d8a86fc5d77e8':
aviobuf: Partial support for reading in read/write contexts
build: Avoid detecting bogus components named 'x'
Conflicts:
libavcodec/allcodecs.c
libavdevice/alldevices.c
libavformat/allformats.c
libavformat/aviobuf.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The function find_things() in configure is confused by component
registration calls as part of multiline macros defining combined
component registration. Coalesce those macros into one line to
work around the issue.
Signed-off-by: Diego Biurrun <diego@biurrun.de>
Signed-off-by: Martin Storsjö <martin@martin.st>
|
| |
| |
| |
| |
| |
| |
| | |
This breaks the sed in configure:find_things(). Fixes regression from
c73c87b. Fixes Ticket2079.
Found-by: jamal
|
| |
| |
| |
| | |
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
|
|\|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* qatar/master:
cosmetics: Prettyprint codec/format/filter registration files
Conflicts:
libavcodec/allcodecs.c
libavdevice/alldevices.c
libavfilter/allfilters.c
libavformat/allformats.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
|
| | |
|
| | |
|
| |
| |
| |
| | |
Tested-by: Val Malykh
|
| |
| |
| |
| |
| |
| | |
Every call to CoInitialize should have a matching CoUnititalize.
Based on patch by Don Moir <donmoir@comcast.net>.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Handling DirectShow events prevents infinite loops when there is an error in
the graph, such as a device being disconnected. This makes it possible for
dshow to return an error to the caller and run the cleanup code.
Based on patch by Don Moir <donmoir@comcast.net>.
|
| |
| |
| |
| |
| | |
Signed-off-by: rogerdpack <rogerpack2005@gmail.com>
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>
|