While the M2M device backing the converter doesn't support multiple
streams natively, it can be run once per stream to produce multiple
outputs from the same input, with different output formats and sizes.
To support this, create a class to model a stream and move control of
the M2M device to the Stream class. The SimpleConverter class then
creates stream instances and iterates over them. Each stream needs its
own instance of the V4L2M2MDevice, to support different output
configurations. The SimpleConverter class retains a device instance to
support the query operations.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Phi-Bang Nguyen <pnguyen@baylibre.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
The SimpleConverter API signals completion of input and output buffer
pairs. This unnecessarily delays requeueing the input buffer to the
video capture queue until the output buffer completes, and also delays
signalling request completion until the input buffer completes. While
this shouldn't cause large delays in practice, it will also not scale
when multi-stream support will be added to the converter class.
To address the current issue and prepare for the future, decouple
signalling of input and output buffers completion.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Phi-Bang Nguyen <pnguyen@baylibre.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Simplify the SimpleConverter interface by opening the M2M device in the
constructor. The explicit call to open() is replaced by a check through
a new isValid() function, and the unused close() function is removed.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Phi-Bang Nguyen <pnguyen@baylibre.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
The number of buffers on the input and output of the converter don't
necessarily need to match. Use the buffer count from the input and
output configuration respectively. This removes the need to pass the
buffer count to the start() function, which brings it closer to the
pipeline handler API.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Phi-Bang Nguyen <pnguyen@baylibre.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Group the configuration parameters for the converter input in a
StreamConfiguration instance. This makes the configure() function
signature cleaner, and will allow passing additional parameters (such as
stride and buffer count).
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Phi-Bang Nguyen <pnguyen@baylibre.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The SimpleConverter class has different sets of functions, related to
static queries, device configuration and runtime operation. Group the
query functions together. While at it, swap the arguments to the
strideAndFrameSize() function to match the order in which pixel format
and size are usually specified.
No functional change is included.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Phi-Bang Nguyen <pnguyen@baylibre.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The stride (and frame size) calculation has been moved from configure
time to configuration validate time by commit 89fb1efac2 ("libcamera: simple:
Fill stride and frameSize at config validation"). This change has
however left one stray setting of the stride when configuring the
converter. Fix it.
While at it, turn the SimpleConverter::configure() output configuration
argument to a const reference to emphasize it can't be null and isn't
modified by the function, and rename it from cfg to outputCfg to make
its purpose clearer.
Fixes: 89fb1efac2 ("libcamera: simple: Fill stride and frameSize at config validation")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Phi-Bang Nguyen <pnguyen@baylibre.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Replace manual resource destruction with std::unique_ptr<> where
applicable. This removes the need for several destructors.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
When configuring the converter, the format is first set on the output
side based on the format of the camera pipeline output, and then the
format is set on the capture side to match the desired stream
configuration. The format parameter passed to
V4L2VideoDevice::setFormat() uses the same variable for both calls,
which has the unwanted side effect of carrying plane configuration from
the output side to the capture side of the converter. In particular, the
stride or plane size requested on the capture side can become
unnecessarily large when converting to a format with a lower number of
bits per pixel (for instance converting YUYV to NV12).
Fix this by resetting the format variable before using it to configure
the capture side.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The V4L2DeviceFormat class now has default initializers for all members,
explicit initialization isn't needed anymore.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Fill the stride and frameSize fields of the StreamConfiguration at
configuration validation time instead of at camera configuration time.
This allows applications to get the stride when trying a configuration
without modifying the active configuration of the camera.
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Extend the SimpleConverter to support scaling, with reporting of the
minimum and maximum output sizes supported for a given input size.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Report the stride when configuring the camera. The stride is retrieved
from the capture device first, and overridden by the converter if used.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The libcamera internal headers are located in src/libcamera/include/.
The directory is added to the compiler headers search path with a meson
include_directories() directive, and internal headers are included with
(e.g. for the internal semaphore.h header)
#include "semaphore.h"
All was well, until libcxx decided to implement the C++20
synchronization library. The __threading_support header gained a
#include <semaphore.h>
to include the pthread's semaphore support. As include_directories()
adds src/libcamera/include/ to the compiler search path with -I, the
internal semaphore.h is included instead of the pthread version.
Needless to say, the compiler isn't happy.
Three options have been considered to fix this issue:
- Use -iquote instead of -I. The -iquote option instructs gcc to only
consider the header search path for headers included with the ""
version. Meson unfortunately doesn't support this option.
- Rename the internal semaphore.h header. This was deemed to be the
beginning of a long whack-a-mole game, where namespace clashes with
system libraries would appear over time (possibly dependent on
particular system configurations) and would need to be constantly
fixed.
- Move the internal headers to another directory to create a unique
namespace through path components. This causes lots of churn in all
the existing source files through the all project.
The first option would be best, but isn't available to us due to missing
support in meson. Even if -iquote support was added, we would need to
fix the problem before a new version of meson containing the required
support would be released.
The third option is thus the only practical solution available. Bite the
bullet, and do it, moving headers to include/libcamera/internal/.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Jacopo Mondi <jacopo@jmondi.org>
Laurent Pinchart
Kieran Bingham
Hirokazu Honda
Paul Elder