| Age | Commit message (Collapse) | Author |
|---|
|
Request complete by themselves when all the buffers they contain have
completed, connecting the buffer's completed signal to be notified of
buffer completion. While this works for now, it prevents pipelines from
delaying request completion until all metadata is available, and makes
it impossible to ensure that requests complete in the order they are
queued.
To fix this, make request completion handling explicit in pipeline
handlers. The base PipelineHandler class is extended with
implementations of the queueRequest() and stop() methods and gets new
completeBuffer() and completeRequest() methods to help pipeline handlers
tracking requests and buffers.
The three existing pipeline handlers connect the bufferReady signal of
their capture video node to a slot of their respective camera data
instance, where they use the PipelineHandler helpers to notify buffer
and request completion. Request completion is handled synchronously with
buffer completion as the pipeline handlers don't need to support more
advanced use cases, but this paves the road for future work.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
Add a new field to the Request class to report its completion status,
and a new complete() method to update the status.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
Add a new field to the Buffer class to report its completion status,
with a new cancel() method to mark the buffer as cancelled.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
There is a need to better control the order of operations an application
performs on a camera for it to function correctly. Add a basic state
machine to ensure applications perform operations on the camera in good
order.
Internal to the Camera states are added; Available, Acquired,
Configured, Prepared and Running. Each state represents a higher state
of configuration of the camera ultimately leading to the highest state
where the camera is capturing frames. Each state supports a subset of
operations the application may perform.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
The arrays that store Stream pointers shall always contain unique
values. Storing them in vectors opens up for the same stream pointer
appearing twice. Remove this possibility by storing them in a set which
guarantees each element is unique.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
When a signal is connected to a member function slot, the slot is not
disconnected when the slot object is deleted. This can lead to calling a
member function of a deleted object if the signal isn't disconnected
manually by the slot object's destructor.
Make signal handling easier by implementing a base Object class that
tracks all connected signals and disconnects from them automatically
when the object is deleted, using template specialization resolution in
the Signal class.
As inheriting from the Object class may to a too harsh requirement for
Signal usage in applications, keep the existing behaviour working if the
slot doesn't inherit from the Object class. We may reconsider this later
and require all slot objects to inherit from the Object class.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
Fix issues reported by checkstyle.py in preparation for further changes.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
In order to support capture, the camera needs methods to allocate and
free buffers, to start and stop the capture and to queue requests.
Define those interfaces in the Camera class and implement them to call
the corresponding pipeline handler methods.
Once a camera is started the pipeline handler of the camera will begin
processing requests queued to the camera by the application until it
gets stopped.
Once a request is created it can be queued to the camera and the
application will be notified asynchronously once the request is
completed and be able to process all the buffers involved in the
request.
At this point the request objects don't support controls. This will be
extended in the future.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
The API towards the application and pipeline handler can be simplified
if the camera caches which streams have been selected and their
respective configuration.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Some operations on the camera requires the application to have exclusive
access to the camera. To help check for this in these operations add a
helper.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Implement a Request object used by applications to queue image capture
requests to a camera.
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
Add a cache of the active stream configuration to the stream object.
This cache is to be updated from the Camera object and can be accessed
read only from both the application and pipeline handlers.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Construct a stream object with a default internal pool.
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
Copy the information from the struct v4l2_buffer when dequeueing the
buffer as applications need this information to make sense of the
captured data.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Provide classes that represent frame buffers and pools of frame buffers.
An image within the system may use one or more Plane objects to track each
plane in the case of multi-planar image formats. The Buffer class manages all
of the data required to render or interpret the raw image data.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
The camera needs to be configured with the number of buffers required to
satisfy the applications use case. While the application can request any
number of buffers, the pipeline must take the constraints of the Linux
driver into consideration.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
|
Extend the camera to support reading and configuring formats for
groups of streams. The implementation in the Camera are minimalistic as
the heavy lifting are done by the pipeline handler implementations.
The most important functionality the camera provides in this context is
validation of data structures passed to it from the application and
access control to the pipeline handler.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
A camera consists of one or more video streams originating from the same
video source. The different streams could for example have access to
different hardware blocks in the video pipeline and therefore be able to
process the video source in different ways.
All static information describing each stream need to be recorded at
camera creation. After a camera is created an application can retrieve
the static information about its streams at any time.
Update all pipeline handlers to register one stream per camera, this
will be extended in the future for some of the pipelines.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Exclusive access must be obtained before performing operations that
change the device state. Define an internal flag to track ownership and
provide a means of protecting functions that change device
configuration.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
|
Add an initial StreamConfiguration implementation to hold configuration
data for a single stream of a Camera. In its current form not many
configuration parameters are supported but it's expected the number of
options will grow over time.
At this stage the pixel format is represented as an unsigned int to
allow for easy mapping to the V4L2 API. This might be subject to change
in the future as we finalize how libcamera shall represent pixel
formats.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Add an initial Stream implementation. The idea is that once capability
support is added to the library each stream will describe its
capabilities using this class. An application will then select one or
more streams based on these capabilities and use them to configure the
camera and capture.
At this stage the Stream class is empty as capabilities are yet to be
added. The class is still useful as it will be used to communicate how
many streams a Camera object provides.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
The class assignment operator is usually defined as returning a
reference to the object, to allow a = b = c statements. While the return
type makes no difference when deleting the operator in a class
definition, it's still a good practice to use the correct return type.
Fix it in the Camera and CameraManager classes.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Use static_cast<>() instead of reinterpret_cast<>() to perform
downcasts, as reinterpret_cast<>() isn't meant (and guaranteed to be
safe) for that purpose.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
The new removeCamera() method is meant to be used by pipeline handlers
to unregister a camera in case of device disconnection.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
As camera object have the potential to outlive the hardware they
represent, there is a need to inform the camera that the underlying
device has been disconnected, and in turn to notify applications.
Implement a disconnection notification mechanism that can be used by
pipeline handlers to notify the camera of disconnection. The camera then
block all new API calls and emit the disconnected signal.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
The PipelineHandler which creates a Camera is responsible for serving
any operation requested by the user. In order forward the public API
calls, the camera needs to store a reference to its pipeline handler.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
---
Changes since v1:
- Create pipeline handlers is shared pointers, make them inherit from
std::enable_shared_from_this<> and stored them in shared pointers.
|
|
The new interrupt() function allows interrupting in-progress blocking
processEvents() calls. This is useful to stop running event loops.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
The Camera class is explicitly reference-counted to manage the lifetime
of camera objects. Replace this open-coded implementation with usage of
the std::shared_ptr<> class.
This API change prevents pipeline handlers from subclassing the Camera
class. This isn't deemed to be an issue. Mark the class final to make
this explicit.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
Cameras are listed through a double indirection, first iterating over
all available pipeline handlers, and then listing the cameras they each
support. To simplify the API make the pipeline handlers register the
cameras with the camera manager directly, which lets the camera manager
easily expose the list of all available cameras.
The PipelineHandler API gets simplified as the handlers don't need to
expose the list of cameras they have created.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
|
|
The CameraManager takes ownership of the dispatcher passed to the
setEventDispatcher() function. Enforces this by using std::unique_ptr<>.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
|
|
Import the following files from the Linux UAPI to keep our V4L2
interface in sync:
media-bus-format.h
v4l2-common.h
v4l2-controls.h
v4l2-mediabus.h
v4l2-subdev.h
videodev2.h
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Convey the fact that the CameraManager class owns the DeviceEnumerator
instance it creates by using std::unique_ptr<> to store the pointer.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
libcamera.h hasn't been updated with the recently added public headers
event_notifier.h, signal.h and timer.h. Add them.
The event_dispatcher_poll.h header, which used to be public, is now
private, but hasn't been removed from libcamera.h. Remove it.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
Provide a poll-based event dispatcher implementation as convenience for
applications that don't need a custom event loop. The poll-based
dispatcher is automatically instantiated if the application doesn't
provide its own dispatcher.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Add three new classes, EventDispatcher, EventNotifier and Timer, that
define APIs for file descriptor event notification and timers. The
implementation of the EventDispatcher is meant to be provided to
libcamera by the application.
The event dispatcher is integrated twith the camera manager to implement
automatic registration of timers and events.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
|
|
Introduce a Signal class that allows connecting event sources (signals)
to event listeners (slots) without adding any boilerplate code usually
associated with the observer or listener design patterns.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
|
|
There can only be a single camera manager instance in the application.
Creating it as a singleton helps avoiding mistakes. It also allows the
camera manager to be used as a storage of global data, such as the
future event dispatcher.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
|
|
Alphabetically sorted includes help finding where to insert new include
statements, and avoiding duplicates.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
The method is declared but not implemented. Remove it.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
|
|
The class was just a placeholder, now that we have other objects
defined, remove it along with the associated test.
The libcamera/libcamera.h header is kept as a shortcut to include the
whole libcamera public API.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Provide a CameraManager class which will handle listing, instancing,
destruction and lifetime management of cameras.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Provide a Camera class which represents our main interface to handling
camera devices. This is a rework of Kieran's initial proposal and
Laurent's documentation of the file changed to fit the device
enumerators needs.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
|
|
In order to avoid depending on system headers that may be outdated compared
to the kernel version available at runtime, import the Linux kernel headers
related to the APIs that libcamera requires a recent version of. This allows
libcamera to use the latest kernel APIs even when compiled on older systems.
The library must of course test at runtime whether those APIs are available
and fallback to older APIs in order to support older kernels.
Import media.h for the recent media controller APIs. The file is extracted
verbatim from kernel v4.19.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
|
|
The build system defines two variables, public_api and sources, that
store the names of the public headers and the source files respectively.
These files will need to be referenced when generating documentation
from source code, so let's make the variable names more descriptive:
- Rename public_api to libcamera_api and use the files() function
- Rename sources to libcamera_sources
- Add a libcamera_headers variable to hold the internal headers
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
|
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
|
The initial main.cpp and libcamera.h files are missing boilerplate
header comments. libcamera.h is further missing include guards. Add
them.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
|
Tabs are disliked within the meson build system.
Replace indentation by spaces, in all existing locations.
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
|
The include directory was defined but not installed.
Add it to the meson build structure to incorporate it as part of the
library install.
To facilitate the same include paths in our internal includes, update
the structure for headers to match the install structure.
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
|
Define the starting points for the libcamera build using
meson and ninja build components.
An initial 'dummy' library class is created, and a test binary links
against the shared library calling it's init_lib() function.
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
