The STM32 contains a camera pipeline known as the DCMIPP (Digital
Camera-Memory Interface Pixel Processor) which receives data from a
parallel interface and dumps the post-processed data to memory. The
pipeline is capable of some processing in the form of downscaling
captured data through cropping or skipping the sensor's output.
The simple pipeline handler is quite capable of handling the DCMIPP
given its operation is handled entirely through configuring the pads
of a media graph, so add support for the driver to the pipeline's
supportedDevices array.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
At generateConfiguration() a YUV/RGB pixel format is preferred for the
StillCapture/VideoRecording/Viewfinder roles, but currently there are no
guarantees in place that the sensor provides a non-Bayer bus format from
which YUV/RGB can be generated.
This makes the default configuration generated for those roles not to
work if the sensor is a RAW-only one.
To improve the situation split the configuration generation in two,
one for YUV modes and one for Raw Bayer mode.
StreamRoles assigned to a YUV mode will try to first generate a YUV
configuration and then fallback to RAW if that's what the sensor can
provide.
As an additional requirement, for YUV streams, the generated mode has to
be validated with the sensor to confirm the desired sizes can be
generated as the ISI cannot up-scale. In order to test a format use the
newly introduced CameraSensor::tryFormat().
Reported-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com>
Tested-by: Daniel Scally <dan.scally@ideasonboard.com>
Add a function to the CameraSensor class that allows to test a format
without applying it to the subdevice and without modifying any control
value associated with the camera sensor.
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The ISICameraConfiguration::validate() function selects which media
bus format to configure the sensor with based on the pixel format
of the first configured stream using the media bus code associated to it
in the formatsMap_ map.
In order to remove the PixelFormamt-to-mbus-code association in
formatsMap_ provide a wrapper function for the newly introduced
getRawMediaBusFormat() and getYuvMediaBusFormat() that automatically
selects what media bus format to use based on the first stream pixel
format.
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com>
Tested-by: Daniel Scally <dan.scally@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
As per the RAW format selection, the media bus format selection
procedure relies on the direct association of PixelFormat and media
bus code in the formatsMap_ map.
As the ISI can generate YUV and RGB formats from any non-Bayer media
bus format, break out the YUV/RGB media bus format selection to a
separate function.
The newly introduced getYuvMediaBusFormat() tests a list of
known-supported media bus formats against the list of media bus
formats supported by the sensor and tries to prefer media bus
codes with the same encoding as the requested PixelFormat.
Use the newly introduced function in
ISICameraConfiguration::validateYuv() to make sure the sensor can
produce a YUV/RGB media bus format.
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com>
Tested-by: Daniel Scally <dan.scally@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The current implementation of the ISI pipeline handler handles
translation of PixelFormat to media bus formats from the sensor
through a centralized map.
As the criteria to select the correct media bus code depends on if the
output PixelFormat is a RAW Bayer format or not, start by splitting
the RAW media bus code procedure selection out by adding a function
for such purpose to the ISICameraData class.
Add the function to the ISICameraData and not to the
ISICameraConfiguration because:
- The sensor is a property of CameraData
- The same function will be re-used by the ISIPipelineHandler
during CameraConfiguration generation.
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com>
Tested-by: Daniel Scally <dan.scally@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
In `udevNotify()`, constructing an std::string from the device's
associated action is unnecessary as it is only compared against static
strings, and for that purpose an std::string_view works just as well,
while being cheaper to construct.
In the same vein, an std::string_view can be used to store the device's
devnode initially, and the string construction can be deferred until it
is needed.
Furthermore, previously `udev_device_get_devnode()` was called twice.
The extra call is now removed.
Signed-off-by: Barnabás Pőcze <pobrn@protonmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Previously, after `addV4L2Device()` had seen all dependecies, it would
remove the `MediaDeviceDeps` object from the `pending_` list, which
would result in it being destroyed. However, there would still be
(dangling) pointers to this object in `devMap_` that were added in
`addUdevDevice()` (line 103). So remove the entry with the given devnum
when it is removed from the corresponding `MediaDeviceDeps` object.
Signed-off-by: Barnabás Pőcze <pobrn@protonmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Files opened internally in libcamera without the O_CLOEXEC file will
remain open upon a call to one of the exec(3) functions. As exec()
doesn't destroy local or global objects, this can lead to various side
effects. Avoid this by opening file descriptors with O_CLOEXEC for all
internal files.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
When an executable using libcamera calls exec(3) while a camera is in
use, file descriptors corresponding to the V4L2 video devices are kept
open has they have been created without O_CLOEXEC. This results in the
video devices staying busy, preventing the new executable from using
them:
[91] ERROR V4L2 v4l2_videodevice.cpp:1047 /dev/video0[149:cap]: Unableto set format: Resource busy
Fix this by opening video devices with O_CLOEXEC, which is generally a
good idea in libraries.
Signed-off-by: Elias Naur <mail@eliasnaur.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Invalid, or not correctly reset requests can cause undefined behaviour
in the pipeline handlers due to unexpected request state.
If the status has not been reset to Request::RequestPending, it is
either not new, or has not been correctly procesed through
Request::reuse().
This can be caught early by validating the status of the request when it
is queued to a camera.
Reject invalid requests before processing them in the pipeline handlers.
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The ConverterFactoryBase::create() function returns a nullptr when no
converter is found. The only caller, SimpleCameraData::init(), checks if
the converter is valid with isValid(), but doesn't check if the pointer
is null, which can lead to a crash.
We could check both pointer validity and converter validity in the
caller, but to limit the complexity in callers, it is better to check
the converter validity in the create() function and return a null
pointer when no valid converter is found.
Signed-off-by: Suhrid Subramaniam <suhrid.subramaniam@mediatek.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
On Raspberry Pi Compute Module platforms, it is possible to attach a
single camera device only to the secondary Unicam port. The current
logic of PipelineHandlerRPi::match() will return a failure during
enumeration of the first Unicam media device (due to no sensor attached,
or sensor failure) and thus the second Unicam media device will never be
enumerated.
Fix this by looping over all Unicam instances in PipelineHandlerRPi::match()
until a camera is correctly registered, or return a failure otherwise.
Reported-on: https://github.com/raspberrypi/libcamera/issues/44
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Add a new parameter to the pipeline handler config file named
"unicam_timeout_value_ms" to allow users to override the automatically
computed Unicam timeout value.
This value is given in milliseconds, and setting a value of 0 (the
default value) disables the override.
An example use of this parameter would be if an application configured a
RAW stream, and provides buffers for the stream on every request. If the
application holds off on sending requests for a particular reason (e.g.
a timelapse use case), then we will possibly hit the watchdog timeout as
it is only a small multiple of the frame length. This override allows an
application to select a larger value with the knowledge that it may
space requests longer than the calculated timeout value.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Add an explicit helper function setCameraTimeout() in the pipeline
handler to set the Unicam timeout value. This function is signalled from
the IPA to set up an appropriate timeout. This replaces the
maxSensorFrameLengthMs value parameter returned back from
IPARPi::start().
Adjust the timeout to be 5x the maximum frame duration reported by the
IPA.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Currently the function `createPipelineHandlers` connects itself to the
`devicesAdded` signal at the end of each call. As the Signal object
supports multiple non-unique listeners connected to it, the former
function would be called exponentially often with each new emitted event
on `devicesAdded` (i.e. with udev plugging in a new camera)
Fix it by connecting the createPipelineHandlers() slot to `devicesAdded`
signal in CameraManager::Private::init() instead. This will prevent the
slot getting connected multiple times to the `devicesAdded` signal.
Signed-off-by: Sophie Friedrich <dev@flowerpot.me>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Currently it is not possible to display debug output from an isolated IPA
module. The standard descriptors are all closed and any specified log
file is explicitly deactivated for the IPA module. Since libcamera and the
isolated IPA module are separate processes, they cannot write to the same
file. However, if syslog is used, then this would be possible.
If syslog is specified as a log file, then this is left as it is for the
isolated IPA module.
Signed-off-by: Matthias Fend <matthias.fend@emfend.at>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Tested-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Commit 613d540267 ("pipeline: raspberrypi: Fix handling of colour
spaces") adjusts the colorspace to ColorSpace::Raw for raw streams.
However, if the colorspace is not requested for raw streams(nullopt),
we should still set the colorspace to ColorSpace::Raw, for raw streams.
Fixes: 613d540267 ("pipeline: raspberrypi: Fix handling of colour spaces")
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Umang Jain <umang.jain@ideasonboard.com>
Commit 6f6e1bf704 ("libcamera: camera_sensor: Apply flips at
setFormat()") extended the CameraSensor::setFormat() function
to apply vertical/horizontal flips on the sensor based on the
supplied Transform. To pass the Transform to the function the
V4L2SubdeviceFormat structure has been augmented with a Transform
member.
However as the newly added Transform is not used at all in the
V4L2Subdevice class, it should not be part of V4L2SubdeviceFormat.
Fix that by removing the transform field from V4L2SubdeviceFormat
and pass it as an explicit parameter to CameraSensor::setFormat().
Fixes: 6f6e1bf704 ("libcamera: camera_sensor: Apply flips at setFormat())
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Commit 1a614866a2 ("libcamera: camera_sensor: Validate Transform") has
removed usage of the IPU3CameraData::rotationTransform_ but hasn't
removed the field itself, nor its initialization. Drop those as they're
unused.
Fixes: 1a614866a2 ("libcamera: camera_sensor: Validate Transform")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
The CameraSensor::validateSensorDriver() function prints a Warning
message when the camera sensor doesn't support flips. We don't mandate
flip support and can run without it without any problem, so a warning is
too harsh. Demote it to a Debug message.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Add the ability to read the platform configuration parameters from a
config file provided by the user through the LIBCAMERA_RPI_CONFIG_FILE
environment variable. Use the PipelineHandler::configurationFile()
helper to determine the full path of the file.
Provide an example configuration file named example.yaml. Currently two
parameters are available through the json file:
"min_unicam_buffers" The minimum number of internal Unicam buffers to
allocate.
"min_total_unicam_buffers" The minimum number of internal + external
Unicam buffers that must be allocated.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reorder the code such that the IPA requested startup drop frames count is
available before the pipeline handler allocates any stream buffers.
This will be used in a subsequent change to stop Unicam buffer allocations if
there are no startup drop frames required and the application has configured a
raw stream and always provides buffers for it.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Add a configuration structure to store platform specific parameters used by
the pipeline handler. Currently, these only store Unicam buffer counts,
replacing the hardcoded static values in the source code.
In subsequent commits, more parameters will be added to the configuration
structure, and parameters will be read in through a config file.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Add a new helper function PipelineHandler::configurationFile() that returns
the full path of a named configuration file. This configuration file may be read
by pipeline handlers for platform specific configuration parameters on
initialisation.
The mechanism for searching for the configuration file is similar to the IPA
configuration file:
- In the source tree if libcamera is not installed
- Otherwise in standard system locations (etc and share directories).
When stored in the source tree, configuration files shall be located in a 'data'
subdirectory of their respective pipeline handler directory.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
We implement a custom validateColorSpaces method that forces all
(non-raw) streams to same colour space, whilst distinguishing RGB
streams from YUV ones, as the former must have the YCbCr encoding and
range over-written.
When we apply the colour space, we always send the full YUV version as
that gets converted correctly to what our hardware drivers expect. It
is also careful to check what comes back as the YCbCr information gets
overwritten again on the way back.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Signed-off-by: Umang Jain <umang.jain@ideasonboard.com>
Pass the available lens controls to the IPA through the configure() function.
Validate that the V4L2_CID_FOCUS_ABSOLUTE does exist. If it doesn't, log a
warning message, and do not advertise focus related controls from the IPA.
Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: Nick Hollinghurst <nick.hollinghurst@raspberrypi.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Previously the x,y offsets in the min/max ScalerCrop control values
were zero. Here we make them the same as the sensor's analogue crop
offset which is I think less misleading.
With this change, it also seems reasonable to advertise the default
scaler crop value to be the true default that you will get. This makes
it possible for applications to see what that value will be without
having to start the camera and wait for frames.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Add support for Transform to the RkISP1 pipeline handler.
The pipeline rotates using the sensor's V/H flips, hence use the
CameraSensor helpers to handle transformation requests from
applications.
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Augment the CameraSensor::setFormat() function to configure horizontal
and vertical flips before applying the image format on the sensor.
Applying flips before format is crucial as they might change the Bayer
pattern ordering.
To allow users of the CameraSensor class to specify a Transform,
add to the V4L2SubdeviceFormat class a 'transform' member, by
default initialized to Transform::Identity.
Moving the handling of H/V flips to the CameraSensor class allows to
remove quite some boilerplate code from the IPU3 and RaspberryPi
pipeline handlers.
No functional changes intended.
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
The two pipeline handlers that currently support Transform (IPU3 and
RaspberryPi) implement it by operating H/V flips on the image sensor.
Centralize the code that validates a Transform request against the
sensor rotation capabilities in the CameraSensor class.
The implementation in the IPU3 pipeline handler was copied from the
RaspberryPi implementation, and is now centralized in CameraSensor to
make it easier for other platforms.
The CameraSensor::validateTransform() implementation comes directly from
the RaspberryPi pipeline handler, no functional changes intended.
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
During the camera sensor driver validation, verify if the sensor
supports horizontal and vertical flips and store a flag as
CameraSensor::supportFlips_ class member.
The flag will be later inspected when applying flips.
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: David Plowman <david.plowman@raspberrypi.com>
Support for the OmniVision OV8858 sensor is scheduled for inclusion in
the Linux kernel in version v6.3.
Add support for the sensor in libcamera by providing static properties
and a camera sensor helper in libipa.
The camera sensor helper expresses analogue gain increments in 1/128
step which differs from what is reported in the sensor documentation in
section "5.8 manual exposure compensation/ manual gain compensation" [0]
A more detailed analysis of the sensor gain model is reported at:
https://patchwork.linuxtv.org/project/linux-media/patch/20221106171129.166892-2-nicholas@rothemail.net/#142267
Record with a \todo note a reference to discussion on the gain model
implementation.
Signed-off-by: Nicholas Roth <nicholas@rothemail.net>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
