GCC5 does not provide prototypes for the math library functions defined
in the math.h header for the std:: namespace.
Include the C++ <cmath> header in place of <math.h> as it defines
overloads for the std::abs and std::fmod function.
This goes intentionally against the libcamera coding guidelines, and
is reported as warning by checkpatch.py.
Fixes: 968ab9bad0 ("libcamera: ipu3: imgu: Calculate ImgU pipe configuration")
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
The ImgUDevice::configureInput() function does not only configure the
input format but applies rectangles to the IF, BDS and GDC components.
Rename it to ImgUDevice::configure().
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Instrument the ImgUDevice::configureInput() function to use the provided
pipe configuration parameters to configure the IF, BDS and GDC
rectangles.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Collect the desired ImgU pipe configuration while assigning
streams in the pipeline handler validate() function and ask the
ImgUDevice class to calculate the pipe configuration parameters.
If the requested pipe configuration results in a non-valid
configuration, return an error from the validate() function.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Instrument the ImgU component to dynamically calculate the image
manipulation pipeline intermediate sizes.
To correctly configure the ImgU it is necessary to program the IF, BDS
and GDC sizes, which are currently fixed to the input frame size.
The procedure used to calculate the intermediate sizes has been ported
from the pipe_config.py python script, available at:
https://github.com/intel/intel-ipu3-pipecfg
at revision:
61e83f2f7606 ("Add more information into README")
Define two structures (ImgUDevice::Pipe and ImgUDevice::PipeConfig)
to allow the pipeline handler to supply and retrieve configuration
parameters from the ImgU.
Finally, add a new operation to the ImgUDevice that calculates
the pipe configuration parameters based on the requested input and
output sizes.
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
The IPU3CameraConfiguration::streams_ field was used to keep an
association between the StreamConfiguration and the assigned streams
before CameraConfiguration::setStream() was called at configure() time.
The stream assignment was based on the order in which elements were
inserted in the vector, implementing a fragile association between
streams and their intended configurations.
As it is now possible to assign streams at validation time, there is no
need to keep that association in place, and the streams_ vector is now
unused.
Remove it and the associated accessor method from the
IPU3CameraConfiguration class.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Remove the adjustStream() and assignStream() methods, and perform stream
adjustment and assignment while iterating the StreamConfiguration
items.
The adjustStream() implementation had some arbitrary assumption, like
the main output having to be as large as the sensor resolution, and did
not take into account the different alignment requirements between the
main output and the viewfinder output.
The assignStream() implementation also assumes only full-size streams
can be produced by the main output, and having it as a separate function
prevents adjusting streams according to which output they are assigned.
Blend the two implementation in a single loop and perform the required
stream adjustment and assignment in one go.
As streams are now assigned at validate() time, remove the same
operation from the configure() function.
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Add a const version of the CIO2Device::sensor() method, that retrieves
a const pointer to the sensor_ class member, to be called by users
which only own a const reference to a CIO2Device class instance.
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
The current implementation of the Camera::configure() method zeroes the
stream pointers assigned to the StreamConfiguration items before calling
the pipeline handler configure() operation, just after the
CameraConfiguration has been validated.
This discards the stream assignment performed at pipeline hander
validation time, requiring platforms that need to perform that early
assignment to maintain the association in place with custom data
structures.
To allow pipeline handlers to use StreamConfiguration::setStream() at
validate() time, zero the stream assignment before calling validate().
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
The IPU3 pipeline handler supports 2 processed RGB/YUV streams
and one RAW stream. Validate that the requested stream combination is
supported in the pipeline handler validate() implementation and return
an error in case it's not.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Report StreamFormats associated to each StreamConfiguration generated
by the IPU3 pipeline handler.
The StreamFormats are generated differently for RAW and processed
streams, with the former using the sensor enumerated resolutions and
the latter using a continuous range of sizes constructed by matching the
sensor capabilities with the platform constraints.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
The validate function overwrites the generated StreamConfiguration with
the one reported by the CIO2 unit when inspecting the RAW stream
configuration.
As we prepare to add StreamFormats to the IPU3 StreamConfiguration,
assigning to the CIO2 generated configuration would delete the
StreamFormats.
Fix this by updating relevant fields only in order to keep the
assigned StreamFormats.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Inspect the return status of validate() in the IPU3 pipeline handler
generateConfigurtion() implementation. If the generated configuration is
not valid, return a an empty configuration to the application.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Remove stream assignment from the IPU3 pipeline handler
generateConfiguration() implementation.
The function aims to provide a suitable default for the requested use
cases. Defer stream assignment to validation and only initialize sizes
and formats.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
The mbusCodesToInfo map actually maps media bus codes to PixelFormat
instances. Rename the map according to its actual function and while at
it drop the 'static' keyword as the map is already defined in an
anonymous namespace.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
The monitor_ and notifier_ pointers are acted on in the destructor if
not set to nullptr, the pointers are however first initialized in
init() and enumerate(). Avoid acting on uninitialized pointers by
initializing them to nullptr in the constructor.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Umang Jain <email@uajain.com>
Object::deleteLater() ensures that the deletion of the Object
takes place in a thread it is bound to. Deleting the Object
in a different thread is a violation according to the libcamera
threading model.
On hot-unplug of a currently streaming camera, the last reference
of Camera when dropped from the application thread (for e.g. QCam's
thread), the destructor is then called from this thread. This is not
allowed by the libcamera threading model. Camera is meant to be deleted
in the thread it is bound to - in this case the CameraManager's thread.
Signed-off-by: Umang Jain <email@uajain.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
This commit adds support to schedule the deletion of an Object to the
thread it is bound to (similar to [1]). An Object getting destroyed
by a different thread is considered as a violation as per the
libcamera threading model.
This will be useful for an Object where its ownership is shared via
shared pointers in different threads. If the thread which drops the
last reference of the Object is a different thread, the destructors
get called in that particular thread, not the one Object is bound to.
Hence, in order to resolve this kind of situation, the creation of
shared pointer can be accompanied by a custom deleter which in turns
use deleteLater() to ensure the Object is destroyed in its own thread.
[1] https://doc.qt.io/qt-5/qobject.html#deleteLater
Signed-off-by: Umang Jain <email@uajain.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Extend the current dispatchMessages() to support dispatching of
selective messsages according to the Message::Type passed in
the function argument. dispatchMessages() can now be called
explicitly to force deliver selected type's message to the
thread for processing (typically when event loop is not
running).
Signed-off-by: Umang Jain <email@uajain.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Now SimpleCameraData::setupFormats() can fail if the camera sensor
supports media bus code which some entities down the pipeline don't.
When this happens continue with the next media bus code instead of
aborting the enumeration of the possible pipeline configurations.
Signed-off-by: Andrey Konovalov <andrey.konovalov@linaro.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The SimpleCameraData::init() function needs to setup links along the
pipeline, but doesn't need to repeat that operation for each media bus
code it tries. Move the link setup before the loop.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
This control is principally for returning, in the image metadata, the
CCM (Colour Correction Matrix) used by the imaging pipeline. By which
we mean the 3x3 matrix that is applied to the camera RGB pixels after
subtraction of black levels and white-balancing, but before any gamma
transformation.
Some implementations may also choose to let an application set
explicit colour matrices, using this as a control.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Remove use of vcsm allocations and replace with dma heap allocations.
The pipeline handler now passes the fd of the allocation over to the IPA
instead of the raw pointer.
Also use libcamera::FileDescriptor for fd lifetime management.
This commit must be built alongside the accompanying BCM2835 ISP kernel
driver changes at https://github.com/raspberrypi/linux/pull/3715.
Otherwise a mismatch will cause undefined behavior.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The IPA is meant to be started when starting the camera, and stopped
when stopping it. It was so far started early in order to handle the
IPAInterface::processEvent() call related to lens shading table
allocation before IPAInterface::configure() to pass the table to the
IPA. Now that the lens shading table is passed through configure(),
starting the IPA early isn't needed anymore.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Naushir Patuck <naush@raspberrypi.com>
The Raspberry Pi IPA uses the custom RPI_IPA_ACTION_SET_SENSOR_CONFIG
frame action to send the sensor staggered write configuration to the
pipeline handler when the IPA is configured. Replace this ad-hoc
mechanism by passing the corresponding data back from the IPA to the
pipeline handler through the configure() response. This allows
synchronous handling of the response on the pipeline handler side.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Naushir Patuck <naush@raspberrypi.com>
The IPAInterface::configure() function now accepts custom configuration
data. Use it to pass the lens shading table instead of using a custom
IPA event. This will allow starting the IPA when starting the camera,
instead of pre-starting it early in order to process the lens shading
table allocation event.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Naushir Patuck <naush@raspberrypi.com>
Instead of receiving sensor orientation configuration from the IPA,
retrieve it from the CameraSensor Rotation property, and configure the
HFLIP and VFLIP controls accordingly.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Naushir Patuck <naush@raspberrypi.com>
The PipelineHandlerRPi::configureIPA() function accesses plenty of
member data from the RPiCameraData class and no member from the
PipelineHandlerRPi class. Move it to RPiCameraData where it logically
belongs.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
