In order to initialise and deinitialise gpuisp we need to be able to setup
EGL in the same thread as Debayer::process() happens in.
This requires extending the Debayer object to provide start and stop
methods which are triggered through invokeMethod in the same way as
process() is.
Introduce start() and stop() methods to the Debayer class. Trigger those
methods as described above via invokeMethod. The debayer_egl class will
take care of initialising and de-initialising as necessary. Debayer CPU
sees no functional change.
Per feedback from Barnabas the stop method is using blocking
synchronisation and thus we drop ispWorkerThread_.removeMessages().
[bod: Made method blocking not queued per Robert's bugfixes]
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Reviewed-by: Milan Zamazal <mzamazal@redhat.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Move the initialisation of Bayer params and CCM to a new constructor in the
Debayer class.
Ensure we call the base class constructor from DebayerCpu's constructor in
the expected constructor order Debayer then DebayerCpu.
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Milan Zamazal <mzamazal@redhat.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The DebayerCpu class has a number of variables, embedded structures and
methods which are useful to DebayerGpu implementation.
Move relevant variables and methods to base class.
Since we want to call setParams() from the GPUISP and reuse the code in
the existing CPUISP as a first step, we need to move all of the
dependent variables in DebayerCPU to the Debayer base class including
LookupTable and redCcm_.
The DebayerEGL class will ultimately be able to consume both the CCM and
non-CCM data.
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Milan Zamazal <mzamazal@redhat.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Add a method to the SwstatsCpu class to process a whole Framebuffer in
one go, rather then line by line. This is useful for gathering stats
when debayering is not necessary or is not done on the CPU.
Reviewed-by: Milan Zamazal <mzamazal@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
[bod: various rebase splats fixed]
[bod: Added constructor Doxygen header]
[bod: Squashed a fix from Hans to calculate stats on every 4th frame]
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
patternSize_ is a private variable and its meaning is already documented
in the patternSize() getter documentation.
Move the list of valid sizes to the patternSize() getter documentation
and drop the patternSize_ documentation.
While at it also add 1x1 as valid size for use with future support
of single plane non Bayer input data.
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Milan Zamazal <mzamazal@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Update the documentation of the statsProcessFn() / processLine0() src[]
pointer argument to take into account that swstats_cpu may also be used
with planar input data or with non Bayer single plane input data.
The statsProcessFn typedef is private, so no documentation is generated
for it. Move the new updated src[] pointer argument documentation to
processLine0() so that it gets included in the generated docs.
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Milan Zamazal <mzamazal@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Encode the bayer shader files into a header as part of the build process.
Qcam already compiles the shader files down into a QT resource file which
it references internally.
In order to share the debayering shader programs outside of qcam create a
generic header which both qcam and libcamera can operate from.
Acked-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Milan Zamazal <mzamazal@redhat.com>
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Some kernel drivers give their entities names that will differ from
implementation to implementation; for example the drivers for the
Camera Receiver Unit and CSI-2 receiver in the RZ/V2H(P) SoC give their
entities names that include their memory address, in the format
"csi-16000400.csi2". Passing that entity name to
V4L2Subdevice::fromEntityName() is too inflexible given it would only
then work if that specific CSI-2 receiver were the one being used.
Add an overload for V4L2Subdevice::fromEntityName() to instead allow
users to pass a std::basic_regex, and use std::regex_search() instead
of a direct string comparison to find a matching entity. Ths allows
us to form regular expressions like "csi-[0-9a-f]{8}.csi2" to find
the entities.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Reviewed-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Some kernel drivers give their entities names that will differ from
implementation to implementation; for example the drivers for the
Camera Receiver Unit and CSI-2 receiver in the RZ/V2H(P) SoC give their
entities names that include their memory address, in the format
"csi-16000400.csi2". Passing that entity name to
MediaDevice::getEntityByName() is too inflexible given it would only
then work if that specific CSI-2 receiver were the one being used.
Add an overload for MediaDevice::getEntityByName() that accepts a
std::basic_regex instead of a string, and use std::regex_search()
instead of a direct string comparison to find a matching entity. This
allows us to search for entites using regex patterns like
"csi-[0-9a-f]{8}.csi2".
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Reviewed-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Some entities in a media graph have names that might differ from
implementation to implementation; for example the Camera Receiver
Unit and CSI-2 receiver on the RZ/V2H(P) SoC have entities with names
that include their address, in the form "csi-16000400.csi2". Passing
that entity name to DeviceMatch is too inflexible given it would only
work if that specific CSI-2 receiver were the one being used.
Add an overload for DeviceMatch::add() such that users can pass in a
std::regex instead of a string. Update DeviceMatch::match() to check
for entities that are matched by the regular expressions added with
the new overload after checking for any exact matches from the vector
of strings. This allows us to use regex to match on patterns like
"csi-[0-9a-f]{8}.csi2".
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
This change integrates the MediaPipeline class into the imx8-isi
pipeline handler. Purpose is to allow a dynamic discovery and
configuration of the actual subdevices graph between the sensor and
the ISI crossbar. This brings support for more complex topologies and
simplifies the implementation.
Signed-off-by: Andrei Gansari <andrei.gansari@nxp.com>
Signed-off-by: Antoine Bouyer <antoine.bouyer@nxp.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Exposes internal MediaEntity::Entity list to help extracting more
information regarding linked entities.
For example, when the pad index of the last device in the list need to be
retrieved from the media pipeline user.
Exposes as const to with a dedicated access to prevent any corruption from
user. Then it is still protected so as when the list was private.
Since MediaPipeline::Entity needs also to be moved to public, then need to
add some documentation in cpp source. Existing documentation from header
file is applied when available.
Signed-off-by: Andrei Gansari <andrei.gansari@nxp.com>
Signed-off-by: Antoine Bouyer <antoine.bouyer@nxp.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Array controls (eg. ColourCorrectionMatrix, FrameDurationLimits,
ColourGains) are serialized properly by the ControlSerializer, but are
not deserialized properly. This is because their arrayness and size are
not considered during deserialization.
Fix this by adding arrayness and size to the serialized form of all
ControlValues. This is achieved by fully serializing the min/max/def
ControlValue's metadata associated with each ControlInfo entry in the
ControlInfoMap.
While at it, clean up the serialization format of ControlValues and
ControlLists:
- ControlValue's id is only used by ControlList, so add a new struct for
ControlList entries to contain it, and remove id from ControlValue
- Remove offset from ControlInfo's entry, as it is no longer needed,
since the serialized data of a ControlInfo has now been converted to
simply three serialized ControlValues
- Remove the type from the serialized data of ControlValue, as it is
already in the metadata entry
The issue regarding array controls was not noticed before because the
default value of the ControlInfo of other array controls had been set to
scalar values similar to how min/max are set, and ColourCorrectionMatrix
was the first control to properly define a non-scalar default value.
Bug: https://bugs.libcamera.org/show_bug.cgi?id=285
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Tested-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com> # rkisp1
Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
When the dewarper is present it can handle arbitrary orientations
specified in the requested camera configuration. In that case handle all
transformations inside the dewarper (even if the sensor supports some of
them) because that makes it easier to handle coordinates for lens
dewarping inside the dewarper.
This complicates the path selection a bit, as for transformations that
include a transpose, the format before the dewarper has swapped
width/height.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Add a transpose() function to size that applies the
Transformation::Transpose operation in the size. This is useful when
handling orientation adjustments.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
To do actual lens dewarping, the dewarper will be configured based on
the tuning file.
As a first step implement the basic loading of the
tuning file and enable/disable the dewarper for the given camera based
on the existence of the "Dewarp" entry under a new top level element
'modules' in the tuning file.
Note: This is an backwards incompatible change in that the dewarper is
currently included in the chain unconditionally. Some users may want to
not use the dewarper, so it is sensible to make that configurable.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
In raw mode the number of configurations is actively limited to 1. It is
therefore safe to move isRaw up one level to simplify the code and
prepare for later use.
During that rework it was noticed that the old code actually has a bug
in that it reduces the number of configurations to 1 in case a raw
config is found, but it doesn't reduce the config vector to that raw
config, but the first config.
Change that behavior to check the first config and either remove all
remaining configs if the first is raw or drop all raw configs if the
first is non-raw.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The dewarper integration into the rkisp1 pipeline is quite complicated.
Simplify that by switching to the now available ConverterDW100Module. As
there is no other known converter in combination with the rkisp1 ISP this
is a safe step to do.
This change also paves the way to implement dw100 specific features later.
The input crop implemented in the dw100 kernel driver is quite limited
in that it doesn't allow arbitrary crop rectangles but only scale
factors quantized to the underlying fixed point representation and only
aspect ratio preserving crops.
The vertex map based implementation allows for pixel perfect crops. The
only downside is that ScalerCrop can no longer be set dynamically on
older kernels. A corresponding warning is already implemented in the
converter module.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The DW100 Dewarp engine is present on i.MX8MP SoC and possibly others.
This patch provides a dedicated converter module that allows easy
integration of such a dewarper into a pipeline handler.
In this patch only the ScalerCrop control is implemented. Support for
additional functionality will be added in later patches.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Using a custom vertex map the dw100 dewarper is capable of doing
complex and useful transformations on the image data. This class
implements a pipeline featuring:
- Arbitrary ScalerCrop
- Full transform support (Flip, 90deg rotations)
- Arbitrary move, scale, rotate
ScalerCrop and Transform is implemented to provide a interface that is
standardized libcamera wide. The rest is implemented on top for more
flexible dw100 specific features.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
When the dewarper is used, an addition buffer loop with
kRkISP1MinBufferCount buffers is created between ISP and dewarper. When
the dewarper is configured, it stores the bufferCount value of the
requested stream configurations. This number of buffers is then imported
when the dewarper is started.
On the input stream of the dewarper the bufferCount is currently left
unchanged, meaning it carries the bufferCount as supplied by the user
instead of the bufferCount of the additional loop. Fix that by setting
the bufferCount to kRkISP1MinBufferCount.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
In order to allow digital zooming, scale down in the dewarper instead of
the resizer. That means forwarding the full sensor size data to the
dewarper. The ScalerCrop rectangle will also be applied at the dewarper.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The decision if the dewarper shall be used is not per pipeline but per
camera and per configuration (raw streams can't use it). Move the
corresponding flag into the camera data class. Rename the flag to
"usesDewarper" which is easier understand when we later add the ability
to enable/disable the dewarper on a per camera basis which will be
expressed using a "canUseDewarper" flag.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
In case the buffer returned from the ISP was canceled, the upstream
buffer was not correctly marked as canceled. Move the cancel
functionality into an own helper function and correctly cancel the
upstream buffers. Add missing cancellation in case queuing to the
dewarper fails.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Isaac Scott <isaac.scott@ideasonboard.com>
To actually use requests with the m2m device, requests need to be
allocated on the underlying media device. This can only be done if the
media device is opened which means acquiring it. Add a function to check
if the m2m device supports requests by acquiring the media device,
asking it and then releasing it again. Also add a function to allocate
requests that acquires the internal media device and releases it after
allocating the requests.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Add V4L2 request support to the V4L2M2MConverter class. Extend the
functions related to buffer queuing with an optional request parameter
that gets passed to the lower layers.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Isaac Scott <isaac.scott@ideasonboard.com>
The V4L2 requests API provides support to atomically tie controls to a
set of buffers. This is especially common for m2m devices. Such a
request is represented by an fd that is allocated via
MEDIA_IOC_REQUEST_ALLOC and then passed to the various V4L2 functions.
Implement a V4L2Request class to wrap such an fd and add the
corresponding utility functions.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
With the upcoming addition of V4L2 requests support, the converters need
to keep a handle to the corresponding media device.
Prepare for that by changing the constructor parameter from a raw
pointer to a shared pointer.
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Isaac Scott <isaac.scott@ideasonboard.com>
Adapt the PipelineHandler::acquireMediaDevice() support function to
return a shared pointer instead of the underlying raw pointer.
Propagate this update to all pipeline handlers that use the MediaDevice
and store a std::shared_ptr<MediaDevice> accordingly.
This is required to support media devices that are potentially shared
among multiple pipeline handlers, like a dewarper implemented as v4l2
m2m device.
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Stefan Klug <stefan.klug@ideasonboard.com>
Reviewed-by: Isaac Scott <isaac.scott@ideasonboard.com>
Virtual pipeline handler should provide colorSpace in
generateConfiguration() and validate the colorspace in validate().
It is mandatory for a pipeline handler to set the colorspace if it
is unset in the stream configuration, during validate().
For choosing the colorspace for the generated NV12 frames, following
points have been taken into account:
- The transfer function should be Rec.709 for NV12
- The YCbCr encoding has been chosen Rec.709 as it is the most common
than Rec.601/Rec.2020
- Range should be 'Limited' as with the NV12 pixel format.
Hence, the closest colorspace match is ColorSpace::Rec709 which is
set for the virtual pipeline handler.
Signed-off-by: Umang Jain <uajain@igalia.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Tested-by: Robert Mader <robert.mader@collabora.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Bayer pattern on the sensor can change while configuring it with the
intended capture format. This is due to the transform being applied on
the sensor which supports [v/h]flips.
During configure(), the simple pipeline handler does not detect any
bayer pattern changes that can arise due to the transformations being
applied via SimpleCameraData:setupFormats(). In such cases, the video
node will be configured in-correctly, without realising the bayer
pattern has changed on the sensor, for the given capture format.
This patch detects the bayer pattern change after the sensor has
been configured and retrieves the corresponding V4L2 pixel format
to correctly configure the video node and the input to converter or
Soft-ISP.
Signed-off-by: Umang Jain <uajain@igalia.com>
Reviewed-by: Milan Zamazal <mzamazal@redhat.com>
Tested-by: Milan Zamazal <mzamazal@redhat.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Bryan O'Donoghue
Hans de Goede
Milan Zamazal
Daniel Scally
Jacopo Mondi
Andrei Gansari
Paul Elder
Stefan Klug
Umang Jain
Kieran Bingham
Umang Jain