Age | Commit message (Collapse) | Author | Files | Lines |
|
gcc 13 moved some includes around and as a result <cstdint> is no longer
transitively included [1]. Explicitly include it.
[1] https://gcc.gnu.org/gcc-13/porting_to.html#header-dep-changes
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
The use of --store-pin is meant to just set a filename for storing
the pin monitoring data, it is not meant to enable pin monitoring
itself, that's for --monitor-pin.
Using just --store-pin without --monitor-pin seemed to work, but it
was actually not logging some data, causing warnings to be issued
when analyzing the result.
With this patch --monitor-pin is required to actually enable pin
monitoring.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
The v4l2-tracer converts the errno integer to a string using
strerrorname_np(). Check if this function is available and,
if not, use strerror() instead.
Signed-off-by: Deborah Brouwer <deborah.brouwer@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
This format is similar to 420, except that it uses full length on
vertical.
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
Those formats are similar to NV12/NV21, except that it uses a full
length at the second plane. Add support for it.
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
Not all planar formats have decimation. Add support for full size
ones.
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
As we're planning to add more formats, let's first split the
switch to avoid if's for two and three plane formats.
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
Some formats use multiple planes with vertical and/or horizontal
decimation. The image size estimation for those planes are wrong,
producing false warnings.
Fix it.
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
If the returned image is too small, the conversion will fail.
So, before letting that happen, validate the image size.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
When raw mode is selected, it should not convert the image
to RGB24. This works fine on all modes but mmap (without threads).
Fix it.
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
use the resolution returned by the driver in order to allocate
the memory for storing the converted buffer.
That fixes a core dump when the driver returns a resolution
bigger than the requested image.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Drivers are allowed to generate buffers where stride != width. Where as
v4lconvert_grey_to_rgb24() assumed that stride == width is always true.
This resulted in wrong frames for monochrome sensors, with padding bytes
being visible diagonally and messing up the image alignment.
Tested with rasbperry pi unicam driver using strides of 32 paired with
the st-vgxy61 driver, which native resolution is 1944x1204, producing a
frame of 1952x1204.
Signed-off-by: Benjamin Mugnier <benjamin.mugnier@foss.st.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
The v4l2-tracer utility didn't support tracing simple controls without
a payload (i.e. the 'size' field is 0) and just a simple value.
This adds support for that, plus the four standard INTEGER64 controls
that use field value64.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Deborah Brouwer <deborah.brouwer@collabora.com>
Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
|
|
The v4l2-tracer traces, records and replays userspace applications that
implement the v4l2 memory-to-memory stateless video decoder interface.
It writes the traced system calls and video frame data to a file in JSON
format. It reads the JSON file to replay the same system calls and to send
the same video frame data to the kernel driver. The JSON file can be
replayed in different environments or on different machines.
Signed-off-by: Deborah Brouwer <deborah.brouwer@collabora.com>
Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
The channel format doesn't include the language of the audio PIDs, so how
is anyone supposed to select the right one?
Make it possible to request all of them. This also includes subtitles.
Link: https://lore.kernel.org/linux-media/16e9ea29-3373-f59f-772d-b7a9b2df21cb@0882a8b5-c6c3-11e9-b005-00805fc181fe
Signed-off-by: Simon Arlott <v4l@octiron.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
When there are multiple PID_* entries in the channels file, keep all of
them instead of only the last one.
Link: https://lore.kernel.org/linux-media/da8f5aa8-e413-1678-d93d-6e8d20fb97b6@0882a8b5-c6c3-11e9-b005-00805fc181fe
Signed-off-by: Simon Arlott <v4l@octiron.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
Logic attempted to move raw data from the generic DVB descriptor into a
linked list with a different structure. Calculation for the destination
length would always fail as a result. Commit moves specific data elements
Link: https://lore.kernel.org/linux-media/61b25413-d877-3a8f-344b-651e4b7885a2@rogers.com
[mchehab: patch was mangled by e-mailer]
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
When syncing from the mainline kernel code, the v4l2-tpg.h header
has to be patched, and the patch was outdated.
Add the missing get_random_u8() function.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Avoid entering the v4l-utils-po and libdvbv5-po directories if NLS
support is disabled, as the generated Makefiles in those directories are
empty then.
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
The main change was renaming Y/CbCr and Y/CrCb to Y/UV and Y/VU in
the pixelformat descriptors. Accept both variants.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
The atomisp driver can generate V4L2_PIX_FMT_NV16 buffers where
stride != width. Where as v4lconvert_nv16_to_yuyv() assumed that
stride == width is always true.
Add a stride argument to v4lconvert_nv16_to_yuyv() to fix this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
The atomisp driver can generate V4L2_PIX_FMT_NV12 buffers where
stride != width. Where as v4lconvert_nv12_*() assumed that
stride == width is always true.
Add a stride argument to v4lconvert_nv12_*() to fix this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
account
The atomisp driver can generate V4L2_PIX_FMT_RGB565 buffers where
stride != width. Where as v4lconvert_rgb565_to_rgb/bgr24() assumed that
stride == width is always true.
Add a stride argument to v4lconvert_rgb565_to_rgb/bgr24() to fix this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
account
The atomisp driver can generate V4L2_PIX_FMT_YUV420 buffers where
stride != width. Where as v4lconvert_yuv420_to_rgb/bgr24() assumed that
stride == width is always true.
Add a stride argument to v4lconvert_yuv420_to_rgb/bgr24() to fix this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
|
|
The old g++ 7.5.0 compiler doesn't support some of the newer struct
initialization methods.
Just assign the fields explicitly.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
|
|
The libbpf APIs bpf_load_program_xattr(), bpf_create_map_node() and
bpf_create_map_in_map_node() have been deprecated since v0.7. Convert
them to use bpf_prog_load() and bpf_map_create().
Also, modify config script to add a check for libbpf version.
Signed-off-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Signed-off-by: Sean Young <sean@mess.org>
|
|
The lirc documentation contains some drastic measures for disabling
in-kernel decoding. Provide a simpler method of disabling kernel
decoding: simply set /etc/rc_maps.cfg to:
* * empty.toml
Signed-off-by: Sean Young <sean@mess.org>
|
|
These options are useful to test what happens if Standby or
Image/Text View On messages are lost every so often.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
If DEBUG_KOBJECT_RELEASE=y then it takes a lot longer to clean up and
create /dev/char/* device nodes.
Give the system more time to settle after an rmmod or modprobe.
This fixes an issue where v4l2-ctl tries to use a /dev/char/81:xx device
that hasn't been created yet.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
The control tests didn't take support for such controls into account.
Fix the tests.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
DUe to an incorrect 'return' which should have been a 'continue'
only the first compound control would be shown instead of all
requested compound controls.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Add a check for nr_of_dims > 0, since such controls also have a
payload (i.e. are not simple integers).
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Make the flag2s function available to other utilities.
Signed-off-by: Deborah Brouwer <deborah.brouwer@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Just cec header changes, but they affect the cec utilities.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Add a string value for V4L2_PIX_FMT_FLAG_SET_CSC in pixflags2s().
Signed-off-by: Deborah Brouwer <deborah.brouwer@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
A number of Doxygen options have become obsolete, and Doxygen complains
about this when building the documentation:
warning: Tag 'TCL_SUBST' at line 229 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'COLS_IN_ALPHA_INDEX' at line 1031 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'LATEX_SOURCE_CODE' at line 1716 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'XML_SCHEMA' at line 1841 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'XML_DTD' at line 1847 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'PERL_PATH' at line 2057 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'CLASS_DIAGRAMS' at line 2070 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'MSCGEN_PATH' at line 2079 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
We don't use those options, so drop them.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
The list of INPUT files in doxygen_libdvbv5.cfg contains
desc_descriptor_id.h, which doesn't exist. The correct file name is
desc_registration_id.h. Fix it.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Main change: drop the references to V4L2_CAP_ASYNCIO.
Also fix some of the capability reporting functions (there were some
duplicates and in one place V4L2_CAP_IO_MC support was missing).
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Add support for the V4L2_PIX_FMT_YUVA32 and V4L2_PIX_FMT_YUVX32 pixel
formats in the v4l2-ctl, v4l2-compliance, qvidcap and qv4l2 utilities.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
While the vicodec tests did verify if the output of several decode
tests was what was expected by using 'cmp', it did not report the
result in a 'FAIL/OK' format, so failures would not be detected
during the daily build.
Fix this to prevent undetected regressions in the future.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Fixes:
dvb-dev-remote.c:1539:15: warning: 'send_buf' reading 80 bytes from a region of size 13 [-Wstringop-overread]
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Fixes:
firmware-tool.c:702:51: warning: pointer 'temp_data' may be used after 'free' [-Wuse-after-free]
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Verify that using an array control in a request works fine.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Changing the input will change the format, which will also change
the dimensions of the pixel array control, and reset the contents
of that array to the default value. Check that this is in fact
happening.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Don't show just the first changed dimension, show all dimensions.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
If this flag is set, get the new dimensions and show it.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Sync with the latest https://git.linuxtv.org/media_stage.git/
kernel.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Move the helper functions to decode tuner/modulator related fields
to v4l2-info.cpp to avoid code duplication in between v4l2-ctl and
rds-ctl.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|
|
Use the helper function v4l2_info_capability() from v4l2-info.cpp
to show the driver info instead of duplicating that code.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
|