Top-level `const` qualifiers are not useful, so avoid them. This is done
either by simply removing the top-level `const`, or making the function
return a reference to const where that is appropriate.
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Stefan Klug <stefan.klug@ideasonboard.com>
This reverts commit 10cdc914da.
The constructors introduced by that commit are not used anywhere,
and they do not match the existing practice for boolean controls.
Specifically, every single boolean control is described by calling
the `ControlInfo(ControlValue, ControlValue, ControlValue)`
constructor. Crucially, that constructor does not set `values_`,
while the two removed constructors do. And whether or not `values_`
has any elements is currently used as an implicit sign to decide
whether or not the control is "enum-like", and those are assumed
to have type `int32_t`.
For example, any boolean control described using any of the two
removed constructors would cause an assertion in failure in
`CameraSession::listControls()` when calling `value.get<int32_t>()`.
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Stefan Klug <stefan.klug@ideasonboard.com>
The abi-compliance checker reports 100% compatibility in this release.
As such the SONAME is maintained at 0.5.
Binary compatibility: 100%
Source compatibility: 100%
Total binary compatibility problems: 0, warnings: 0
Total source compatibility problems: 0, warnings: 0
This release brings 129 commits with some substantial development on gstreamer
for this release which is nice to see.
In core we can see development that cleans up ProcessManager in a drive to
reduce singleton patterns throughout libcamera. There's a new ClockRecovery
class from Raspberry Pi which supports the new Wallclock timestamps
metadata and provides the capablity for synchronising cameras across
devices which will very exciting to see in the future.
Pipelines are now able to limit the number of requests queued into the pipeline
handler which helps prevent issues when applications desire to use a larger
request queue. This is particularly beneficial to pipewire based applications
it seems with longer processing queues.
It's now possible to check if controls exist with new macros defined for
every control to help codebases compile against multiple versions:
- LIBCAMERA_HAS_$VENDOR_VENDOR_$MODE_$NAME
And for developers we now have a b4-config file to support easily using the
tool to send patche to the mailinglist!
On the apps side, there's lots of development on Gstreamer, including fixing
bayer formats for 10/12/14/16 bit depths. Both cam and gstreamer now report the
camera properties in full with text representations of all the enum/properties.
With the IPA components, Raspberry Pi includes development and updates to
Autofocus and makes the maximum digital gain limits configurable. RKISP1 now
has ExposureValue control implemented using exposure compensation in the
agc_mean_luminance implementation. The ST vd56g3 camera sensor support is added
for both RPi and libipa platforms.
For Pipeline Handlers notable updates include the imx8-isi being fixed for
newer generation chips from NXP adding multi-camera support, and Wallclock
timestamp support added to Rapsberry Pi.
Finally, on the documentation side - a notable change is the update to use
spinhx-doxylink. This should prevent stale links being used in the
documentation pages but adds an extra tool to the dependencies for building the
documentation.
Contributors:
27 Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
23 Laurent Pinchart <laurent.pinchart@ideasonboard.com>
17 Stefan Klug <stefan.klug@ideasonboard.com>
12 David Plowman <david.plowman@raspberrypi.com>
10 Naushir Patuck <naush@raspberrypi.com>
9 Nick Hollinghurst <nick.hollinghurst@raspberrypi.com>
7 Umang Jain <uajain@igalia.com>
5 Antoine Bouyer <antoine.bouyer@nxp.com>
5 Benjamin Mugnier <benjamin.mugnier@foss.st.com>
3 Jaslo Ziska <jaslo@ziska.de>
2 Daniel Scally <dan.scally@ideasonboard.com>
2 Kieran Bingham <kieran.bingham@ideasonboard.com>
1 Christian Rauch <Rauch.Christian@gmx.de>
1 Han-Lin Chen <hanlinchen@chromium.org>
1 Harvey Yang <chenghaoyang@chromium.org>
1 Hou Qi <qi.hou@nxp.com>
1 Milan Zamazal <mzamazal@redhat.com>
1 Nícolas F. R. A. Prado <nfraprado@collabora.com>
1 Paul Elder <paul.elder@ideasonboard.com>
143 files changed, 6004 insertions(+), 1620 deletions(-)
Integration overview:
The following commits in this release relate to either a bug fix or an
improvement to an existing commit.
- pipeline: imx8-isi: Cosmetic changes
- Fixes: 680cde6005 ("libcamera: imx8-isi: Split Bayer/YUV config generation")
- pipeline: imx8-isi: Fix match returned value in error case
- Fixes: 0ec982d210 ("libcamera: pipeline: Add IMX8 ISI pipeline")
- gstreamer: Replace NULL with nullptr
- Reported-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
- libcamera: base: Fix log level parsing when multiple categories are listed
- Fixes: 24c2caa1c1 ("libcamera: base: log: Use `std::string_view` to avoid some copies")
- libcamera: pipeline: uvcvideo: Silently ignore `AeEnable`
- Fixes: ffcecda4d5 ("libcamera: pipeline: uvcvideo: Report new AeEnable control as available")
- ipa: rpi: Defer initialising AF LensPosition ControlInfo and value
- Bug: https://bugs.libcamera.org/show_bug.cgi?id=258
- subprojects: libpisp: Update to 1.2.1
- Bug: https://github.com/raspberrypi/libpisp/pull/43
- ipa: rpi: Fix bug in AfState reporting
- Fixes: ea5f451c56 ("ipa: rpi: controller: AutoFocus bidirectional scanning")
- libcamera: base: bound_method: Move return value
- Bug: https://bugs.libcamera.org/show_bug.cgi?id=273#c1
- libcamera: camera: Fix up the AeEnable control during Camera::start()
- Fixes: 7abd413905 ("libcamera: camera: Pre-process AeEnable control")
- libcamera: Put buffer back to V4L2BufferCache when VIDIOC_QBUF fails
- Fixes: cadae67e45 ("libcamera: v4l2_videodevice: Add FrameBuffer interface")
- utils: codegen: ipc: Check `ipc_` instead of `isolate_`
- Bug: https://bugs.libcamera.org/show_bug.cgi?id=276
- libcamera: camera: Do not call `generateConfiguration()` synchronously
- Bug: https://bugs.libcamera.org/show_bug.cgi?id=273
- meson: Update subprojects .gitignore
- Fixes: a29c53f6a6 ("meson: Use libyaml wrap file from wrapdb")
And the following updates have been made in this release, grouped by category:
core:
- libcamera: base: log: Take `LogCategory` by reference
- libcamera: base: Fix log level parsing when multiple categories are listed
- libcamera: process: Use `pid_` member to decide if running
- libcamera: process: start(): Use span instead of vector
- libcamera: process: closeAllFdsExcept(): Take vector by value
- libcamera: process: Move `closeAllFdsExcept()`
- libcamera: process: Use `close_range()` when available
- libcamera: process: Remove `ProcessManager` singleton
- libcamera: process: Disable copy/move
- libcamera: process: Misc. cleanup around `execv()`
- libcamera: process: Return error if already running
- libcamera: process: Ensure that file descriptors are nonnegative
- libcamera: framebuffer: Add FrameMetadata::Status::FrameStartup
- libcamera: Add ClockRecovery class to generate wallclock timestamps
- libcamera: controls: Revert incorrect SPDX removal
- libcamera: controls: Remove hyphenation in control description text
- libcamera: controls: Use nanoseconds units for FrameWallClock
- libcamera: delayed_controls: Inherit from Object class
- libcamera: pipeline_handler: Move waitingRequests_ into camera class
- libcamera: pipeline_handler: Allow to limit the number of queued requests
- libcamera: base: bound_method: Move return value
- libcamera: controls: Generate macro for each control
- libcamera: camera: Fix up the AeEnable control during Camera::start()
- libcamera: Put buffer back to V4L2BufferCache when VIDIOC_QBUF fails
- libcamera: camera_manager: Log info message to report camera creation
- libcamera: v4l2_videodevice: Add color space to format string representation
- libcamera: Drop remaining file names from header comment blocks
- libcamera: control_serializer: Accept empty `ControlList`
- libcamera: camera: Do not call `generateConfiguration()` synchronously
- libcamera: utils: Add scope_exit class
- libcamera: v4l2_videodevice: Use scope_exit
- libcamera: camera_sensor_properties: Add vd56g3 camera sensor
- libcamera: base: bound_method: Forward arguments when possible
- meson: Use libyaml wrap file from wrapdb
- meson: Update subprojects .gitignore
- pipeline: v4l2_subdevice: Add color space to format string representation
- utils: gen-debug-controls: Remove line number from control description
- utils: gen-debug-controls: Fix handling of controls that appear multiple times
- utils: gen-debug-controls: Improve log output
- utils: codegen: ipc: Check `ipc_` instead of `isolate_`
- controls: Add FrameWallClock control
- controls: Add camera synchronisation controls for Raspberry Pi
- camera_sensor: Expand on computeTransform() documentation
- subprojects: libpisp: Update to 1.2.1
- V4L2VideoDevice: Call FrameBuffer::Private::cancel() in streamOff()
- camera: Fix spell error
- b4: Add .b4-config
- package: Drop Gentoo ebuild
apps:
- android: camera_device: Do not pass `nullptr` to `Request::addBuffer()`
- apps: common: image: Fix assertion
- apps: cam: Print enum string for camera properties
- apps: lc-compliance: Replace manual include guard with pragma once
- qcam: Silence false positive warnings with Qt 6.9.0 and newer
- gstreamer: Document improvements when updating minimum GStreamer version
- gstreamer: Factor out video pool creation
- gstreamer: Reduce indentation in gst_libcamera_create_video_pool()
- gstreamer: Rename variable in gst_libcamera_create_video_pool()
- gstreamer: Fix leak of GstQuery and GstBufferPool in error path
- gstreamer: Drop incorrect unref on caps
- gstreamer: Replace NULL with nullptr
- gstreamer: Fix libcamerasrc responding latency before setting caps
- gstreamer: Use std::exchange() instead of g_steal_pointer()
- gstreamer: Make format_map[] const
- gstreamer: Fix reconfiguration condition check
- gstreamer: Log and check adjusted camera configuration
- gstreamer: Enable bayer formats with 10/12/14/16 bits
- gstreamer: Split value_set_rectangle() GValue helper
- gstreamer: Move existing GValue helpers to gstreamer-utils
- gstreamer: Report camera properties as device properties
ipa:
- ipa: rpi: Replace dropFrameCount in the IPA -> PH interface
- ipa: rpi: Rename dropFrameCount_ to invalidCount_
- ipa: rpi: Initialize enum controls with a list of values
- ipa: rpi: Defer initialising AF LensPosition ControlInfo and value
- ipa: rpi: controller: Improve findPeak() function in AF algorithm
- ipa: rpi: controller: AutoFocus weighting tweak
- ipa: rpi: controller: Autofocus CAF/PDAF stability tweak
- ipa: rpi: controller: AutoFocus tweak earlyTerminationByPhase()
- ipa: rpi: controller: Autofocus to use AWB statistics; re-trigger
- ipa: rpi: controller: AutoFocus bidirectional scanning
- ipa: rpi: Update IMX708 camera tuning files for AutoFocus changes
- ipa: rpi: Fix bug in AfState reporting
- ipa: rpi: agc: Change handling of colour gains less than 1
- ipa: rpi: agc: Make the maximum digital gain configurable
- ipa: rpi: agc: Rename "analogue gain" to "gain" where appropriate
- ipa: rpi: Advance the delay context counter even when IPAs don't run
- ipa: rpi: agc: Calculate digital gain in process()
- ipa: rpi: Update digital gain handling in IPA base and derived classes
- ipa: rpi: agc: Remove digital gain from AgcPrepareStatus
- ipa: rpi: Fix static initialisation order bug in the Controller
- ipa: rpi: Add vd56g3 support for rpi
- ipa: rpi: Add vd56g3 tuning files for rpi
- ipa: rkisp1: Move Sharpness control creation to Filter algorithm
- ipa: rkisp1: agc: Implement ExposureValue control
- libipa: pwl: Improve documentation
- libipa: pwl: Fix single point Pwl
- libipa: agc_mean_luminance: Add debug logging
- libipa: agc_mean_luminance: Add exposure compensation support
- libcamera: libipa: Add vd56g3 support for libipa
pipeline:
- pipeline: ipa: rpi: Split RPiCameraData::dropFrameCount_
- pipeline: rpi: Remove disable_startup_frame_drops config option
- pipeline: rpi: Remove ispOutputCount_ and ispOutputTotal_
- pipeline: rpi: Add wallclock timestamp support
- pipeline: rpi: Fix for enumerating the media graphs
- libcamera: pipeline: rpi: Do not set timestamps to 0 if unavailable
- libcamera: pipeline: rpi: Do not clear request metadata anymore
- pipeline: imx8-isi: Cosmetic changes
- pipeline: imx8-isi: Fix match returned value in error case
- pipeline: imx8-isi: Dynamically compute crossbar subdevice's first source.
- pipeline: imx8-isi: Add constexpr for maximum pipeline and resize if needed
- pipeline: imx8-isi: Add multicamera support
- libcamera: pipeline: uvcvideo: Silently ignore `AeEnable`
- libcamera: pipeline: uvcvideo: Handle controls during startup
- pipeline: rkisp1: Limit the maximum number of buffers queued in
- pipeline: rkisp1: Properly handle the bufferCount set in the stream configuration
- libcamera: pipeline: rkisp1: Don't rely on bufferCount
- libcamera: software_isp: Remove type casts in statistics computation
- pipeline: simple: Fix matching with empty media graphs
- libcamera: mali-c55: Set bytesused appropriately
- libcamera: mali-c55: Correct expected entity function
test:
- libcamera: test: Add a failing test for the log level parser
- test: libtest: CameraTest: Set env var before CameraManager
- test: ipa: Add basic Pwl test
- test: ipa: Add failing test for single point Pwl
documentation:
- Documentation: Make `doxygen-internal` depend on public inputs
- Documentation: application: Update mediactl URL
- Documentation: Remove unneeded options from Sphinx configuration
- Documentation: Drop documentation author names
- Documentation: Use Sphinx doxylink to generate links to doxygen
- Documentation: Replace links to Doxygen documentation with doxylink
Acked-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Use `std::{forward,move}` to forward the arguments, this enables the
use of move constructors, likely leading to less code and better runtime.
For example, move constructing a libstdc++ `std::shared_ptr` is noticeably
less code than copy constructing one.
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Iterate over all libcamera camera properties and report them as
device properties. Each libcamera ControlType is mapped to the
corresponding gstreamer GType. If the ControlValue is an array of
values (ControlValue::isArray()), GValue with type GST_TYPE_ARRAY
is used to set the value of that ControlValue with the corresponding
GType.
Signed-off-by: Umang Jain <uajain@igalia.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Move the existing GValue helpers from gstlibcamera-controls.cpp.in
to gstreamer-utils.cpp. The intention here is to make these helpers
available to other parts of gstreamer source element, for example,
reporting camera properties in GstDeviceProvider.
The function signature has been changed to match with the other
gstreamer-utils utility functions:
value_get_rectangle() => gst_libcamera_gvalue_get_rectangle()
value_set_rectangle() => gst_libcamera_gvalue_set_rectangle()
value_set_point() => gst_libcamera_gvalue_set_point()
value_set_size() => gst_libcamera_gvalue_set_size()
Signed-off-by: Umang Jain <uajain@igalia.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Split the value_set_rectangle() GValue helper into further
helpers pertaining to libcamera::Point and libcamera::Size.
This would help to cover additional cases where helpers
are needed for Point and Size individually (in subsequent commits).
The libcamera::Rectangle's GValue helper can be easily
constructed with the new Point and Size helpers. Hence,
this patch does not introduce any functional changes.
Signed-off-by: Umang Jain <uajain@igalia.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
For both vc4 and pisp, vd56g3.json has been generated using ctt with
rpi.dpc algorithm removed as this is already handled in the sensor's
ISP. vd56g3_mono.json has been adapted from vd56g3.json by removing
color correction related algorithms.
Adding Cyril Liotard and Jean Poire as co-developers for providing the
base vd56g3.json tuning files for vc4 and pisp respectively. Thank you.
Signed-off-by: Benjamin Mugnier <benjamin.mugnier@foss.st.com>
Co-developed-by: Cyril Liotard <cyril.liotard@st.com>
Signed-off-by: Cyril Liotard <cyril.liotard@st.com>
Co-developed-by: Jean Poire <jean.poire@st.com>
Signed-off-by: Jean Poire <jean.poire@st.com>
Reviewed-by: Naushir Patuck <naush@rasbperrypi.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Add unit cell size from the 'pixel size' element in the datasheet.
Delays are set to 2 in case a setting is entered at the very end of the
N frame, the N+1 frame will miss it and only the N+2 frame will use this
new setting.
Note that vd56g3 has a diagonal color test pattern, but does not match
any description in specified test patterns and therefore is ignored.
Signed-off-by: Benjamin Mugnier <benjamin.mugnier@foss.st.com>
Reviewed-by: Naushir Patuck <naush@raspberrypi.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The `ProcessManager` is a singleton class to handle `SIGCHLD` signals
and report the exit status to the particular `Process` instance.
However, having a singleton in a library is not favourable and it is
even less favourable if it installs a signal handler.
Using pidfd it is possible to avoid the need for the signal handler;
and the `Process` objects can watch their pidfd themselves, eliminating
the need for the `ProcessManager` class altogether.
`P_PIDFD` for `waitid()` was introduced in Linux 5.4, so this change
raises the minimum supported kernel version. `clone3()`, `CLONE_PIDFD`,
`pidfd_send_signal()` were all introduced earlier.
Furthermore, the call to the `unshare()` system call can be removed
as those options can be passed to `clone3()` directly.
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Instead of creating a new vector, take the vector by value to make it
possible for the caller to use move construction when calling the function.
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Instead of using a separate member variable, use `pid_ > 0` to determine
if the process is still running. Previously the value of `pid_` was not
reset to -1 when the process terminated, but since it is only meaningful
while the process is running, reset it to -1 in `Process::died()`.
Neither `pid_` nor `running_` are exposed, so this change has no effect
on the public interface or observable behaviour.
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
When no log category is specified, `nullptr` is passed, and
then the `_log()` function implementations replace that with
`LogCategory::defaultCategory()`. But since the call site always
knows the log category, this condition can be removed and the
`_LOG1()` macro can use `LogCategory::defaultCategory()`.
So remove the condition from the `_log()` implementations and
use references to refer to log categories.
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
As CameraConfiguration::validate() might alter the configuration
previously negotiated with downstream, log an info when this happens as
GStreamer source elements are not supposed to do that.
Also check if downstream can accept this new stream configuration and if
not, return a not-negotiated error.
Signed-off-by: Jaslo Ziska <jaslo@ziska.de>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Umang Jain <uajain@igalia.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
gst_pad_peer_query_accept_caps() might only check if the caps are
acceptable with the peer element, but not recursively with all
downstream elements. If the reconfigure flag was set because the
pipeline downstream changed, gst_pad_peer_query_accept_caps() might still
return true, even though downstream can't handle the current caps, which
causes a not-negotiated error.
This commit fixes this issue by emitting a query event which
recursively checks with all downstream elements. Because at this point
we are only interested in whether the current caps are still acceptable,
use the currently used caps as a filter and then check if the query
returned empty caps.
Signed-off-by: Jaslo Ziska <jaslo@ziska.de>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Umang Jain <uajain@igalia.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Use the Sphinx doxylink extension to generate links to the
Doxygen-generated documentation automatically. Not only does this fix
currently broken links, but it also ensures that any removal or rename
of a class or function referenced to from the Sphinx documentation
without a corresponding documentation update will be caught by a
documentation build error.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Stefan Klug <stefan.klug@ideasonboard.com>
The libcamera Sphinx documentation needs to link to the API
documentation generated by Doxygen. The links currently point to the
documentation hosted on the official https://libcamera.org/ website.
This causes multiple issues:
- Doxygen generates URLs with MD5 hashes of function signatures, making
the link targets unstable.
- When testing documentation builds that include API changes, links to
new API elements will be broken.
- The generated documentation can't be browsed offline.
Fix this by using the Sphinx doxylink extension. This allows specifying
link targets as class and function names, with the link being
automatically generated using the same MD5 hashing as Doxygen. The root
of the link target is configured in a central location, which defaults
to the build directory and can be overridden to point to the libcamera
website when pushing the documentation.
This commit only introduces the infrastructure to use doxylink. Manual
links will be replaced separately.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Stefan Klug <stefan.klug@ideasonboard.com>
The name of the documentation authors listed in conf.py is seriously out
of date. As the information is bound to bitrot, replace the names by a
generic mention of the libcamera documentation authors.
While at it, update the copyright dates.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Stefan Klug <stefan.klug@ideasonboard.com>
libcamera only generates HTML documentation. Drop the unused
configuration options for other output formats, as well as unneeded
entries in the exclude patterns.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Stefan Klug <stefan.klug@ideasonboard.com>
The linux kernel documentation for the Media Controller has moved.
Update the URL accordingly to the new location.
The existing link pointed to the 'introduction' page - but this isn't
easy to identify or get an overview of the full documentation available
for media controller. Instead point the link to the top level of the
media controller userspace API pages.
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Commit a29c53f6a6 ("meson: Use libyaml wrap file from wrapdb")
modified the libyaml wrap in a way that resulted in the subproject
source code being stored in a different subdirectory, but didn't update
.gitignore accordingly. Fix it.
Fixes: a29c53f6a6 ("meson: Use libyaml wrap file from wrapdb")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Most pipeline handler methods are dispatched in the internal `CameraManager`
thread. `PipelineHandler::generateConfiguration()` is called directly,
however, which likely goes against the expectations, so call it on the
internal thread.
Before 14618cdd0c ("libcamera: base: bound_method: Move return value")
this could not be done because the function returns `std::unique_ptr`, and
that type cannot be copied.
Bug: https://bugs.libcamera.org/show_bug.cgi?id=273
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Only try to send the "Exit" message if the `IPCPipeUnixSocket` object
exists. If the constructor fails, then `ipc_` might remain nullptr,
which would lead to a nullptr dereference in the destructor.
This change also modifies the constructor so that only a valid
`IPCPipeUnixSocket` object will be saved into the `ipc_` member,
which avoids error messages when `IPCPipeUnixSocket::sendAsync()`
is callded in the inappropriate state.
Bug: https://bugs.libcamera.org/show_bug.cgi?id=276
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Accept empty `ControlList`s without an info map. Serialization
supports `ControlList`s without an associated `ControlInfoMap`
object. However, for deserialization, a "v4l2" control list
without an info map resulted in a fatal error.
After 30114cadd8 ("ipa: rpi: Defer initialising AF LensPosition ControlInfo and value")
the `lensControls` member of `ipa::RPi::ConfigResult` is left empty
and without an info map in every case, not just when a lens is not present.
This causes the above assertion to trigger every single time.
Also fix the indefinite article of "a ControlInfoMap" in the
error message.
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Isaac Scott <isaac.scott@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
libcamera includes packaging information for Gentoo Linux. This is due
to historical reasons, the ebuild is not testing or maintained, and we
don't plan to add packaging information for other major distributions
such as Debian or Fedora. Drop the Gentoo ebuild.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Header comment blocks used to contain the file name. Considered as
useless information, the names have been removed, with the last ones
supposed to be dropped in commit d3bf27180e ("libcamera: Drop
remaining file names from header comment blocks"). A few have however
been forgotten, and more crept back since. Remove them.
While at it, fix one typo in a header comment block by replacing
'MaliC55 with Mali-C55', and add a missing blank line in
src/ipa/rpi/pisp/pisp.cpp.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com>
To debug color issues in a pipeline it is necessary to know the color
space used. Therefore add the color space to the string representation
of V4L2DeviceFormat that is returned by toString() and operator<<(). The
format gets for example printed as
[0:05:52.981684000] [230] DEBUG RkISP1 rkisp1_path.cpp:468 Configured main resizer video device with 1920x1080-NV12/sYCC
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Umang Jain <uajain@igalia.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
In Qt 6.9.0 the qtmochelpers.h header introduced a construct that gcc 12
and gcc 13 incorrectly flag as variable shadowing:
In file included from src/apps/qcam/qcam.p/moc_viewfinder_gl.cpp:12:
/usr/include/qt6/QtCore/qtmochelpers.h: In instantiation of ‘constexpr void QtMocHelpers::detail::UintDataStorage<std::integer_sequence<int, Idx ...>, T ...>::forEach(F&&) const [with F = QtMocHelpers::UintData<QtMocHelpers::SignalData<void(libcamera::FrameBuffer*)> >::copyTo<{anonymous}::qt_meta_tag_ZN12ViewFinderGLE_t, QtMocHelpers::MetaObjectContents<24, 10, 60, 3> >(QtMocHelpers::MetaObjectContents<24, 10, 60, 3>&, size_t, uint&) const::<lambda(const auto:39&)>; int ...Idx = {0}; T = {QtMocHelpers::SignalData<void(libcamera::FrameBuffer*)>}]’:
/usr/include/qt6/QtCore/qtmochelpers.h:255:21: required from ‘constexpr const QtMocHelpers::MetaObjectContents<24, 10, 60, 3> ViewFinderGL::qt_staticMetaObjectContent<{anonymous}::qt_meta_tag_ZN12ViewFinderGLE_t>’
src/apps/qcam/qcam.p/../../../../../../src/apps/qcam/viewfinder_gl.h:32:2: required from ‘constexpr const auto ViewFinderGL::qt_staticMetaObjectStaticContent<{anonymous}::qt_meta_tag_ZN12ViewFinderGLE_t>’
src/apps/qcam/qcam.p/moc_viewfinder_gl.cpp:63:5: required from here
src/apps/qcam/qcam.p/../../../../../../src/apps/qcam/viewfinder_gl.h:32:2: in ‘constexpr’ expansion of ‘ViewFinderGL::qt_create_metaobjectdata<{anonymous}::qt_meta_tag_ZN12ViewFinderGLE_t>()’
src/apps/qcam/qcam.p/moc_viewfinder_gl.cpp:58:87: in ‘constexpr’ expansion of ‘QtMocHelpers::metaObjectData<ViewFinderGL, {anonymous}::qt_meta_tag_ZN12ViewFinderGLE_t, StringRefStorage<char [13], char [15], char [1], char [24], char [7]>, UintData<SignalData<void(libcamera::FrameBuffer*)> >, UintData<>, UintData<> >(0, qt_stringData, qt_methods, qt_properties, qt_enums, QtMocHelpers::UintData<>(), const QtMocHelpers::detail::UintDataBlock<0, 0>{uint [1](), uint [1]()})’
/usr/include/qt6/QtCore/qtmochelpers.h:563:36: in ‘constexpr’ expansion of ‘(& methods)->QtMocHelpers::UintData<QtMocHelpers::SignalData<void(libcamera::FrameBuffer*)> >::copyTo<{anonymous}::qt_meta_tag_ZN12ViewFinderGLE_t, QtMocHelpers::MetaObjectContents<24, 10, 60, 3> >(result, ((size_t)dataoffset), metatypeoffset)’
/usr/include/qt6/QtCore/qtmochelpers.h:201:57: error: declaration of ‘entry’ shadows a member of ‘QtMocHelpers::detail::UintDataStorage<std::integer_sequence<int, 0>, QtMocHelpers::SignalData<void(libcamera::FrameBuffer*)> >’ [-Werror=shadow]
201 | [[maybe_unused]] auto invoke = [&f](const auto &entry) { f(entry.entry); return 0; };
| ~~~~~~~~~~~~^~~~~
/usr/include/qt6/QtCore/qtmochelpers.h:180:7: note: shadowed declaration is here
180 | T entry;
| ^~~~~
There is little we can do but silence the warning. Do so selectively
based on the Qt and gcc versions, to still detect variable shadowing
with compilers unaffected by the issue.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Umang Jain <uajain@igalia.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Camera creation is one of the most important events generated by
libcamera, but we are completely silent about it. The lack of a log
message makes it more difficult to identify problems and provide
support. Fix it by adding an Info message that reports the camera id and
its pipeline handler when the camera is added.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Umang Jain <uajain@igalia.com>
Reviewed-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The match() function currently reports that it is not possible to create
any cameras if it encounters an empty media graph.
Fix this by looping over all media graphs and only returning false when
all of them fail to create a camera.
It is worth noting that an issue does exist when on a partial match that
ends in an invalid match, any media devices that were acquired will stay
acquired. This is not a new issue though, as any acquired media devices
in general are not released until pipeline handler deconstruction. This
requires a rework of how we do matching and pipeline handler
construction, so it is captured in a comment.
In the meantime, this fix fixes a problem without increasing the net
number of problems.
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Tested-by: Hui Fang <hui.fang@nxp.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
In Camera::queueRequest() the control list is updated transparently by
converting AeEnable into ExposureTimeMode and AnalogueGainMode
controls.
However, this was not happening during Camera::start(), meaning that
setting AeEnable there was having no effect. It should behave the same
here too.
Fixes: 7abd413905 ("libcamera: camera: Pre-process AeEnable control")
Signed-off-by: David Plowman <david.plowman@raspberrypi.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
ISI can be used to support multiple cameras at the same time in the media
device. A dedicated pad is assigned to each camera by HW (dts). We then
assign one (or more) source pad(s) depending on ISI constraints in the SoC.
Since ISI may have different number of pipes depending on SoC, the number
of cameras is computed _before_ going through all pipeline's components,
to limit number of pipes that could be assigned to a camera.
For instance, 3 is targeted as maximum amount of streams per camera as
defined by 'kNumStreams' constant. If 2 cameras are connected, with
only 5 ISI pipes, this target cannot be achieved. In such case, 2 streams
are assigned to each camera.
On the other hand, if ISI has 8 source pads (from index 6 to 13) and 2
cameras, first three source pads [6:8] are assigned to first camera, and
the three next source pads [9:11] are assigned to the second camera. All
these pads (sink and sources) must have same format to prevent configuration
mismatch during start.
However, since ISI routing must be performed _before_ any stream becomes
active, 'setRouting' is now executed during the 'match' step, instead of
'configure'. Indeed, it is up to the application to decide when a camera
'start' is executed: this could happen before or after the other camera
configuration is executed. Moving routing into the 'match' step makes sure
the routing is correctly applied before any 'start' is executed.
Signed-off-by: Antoine Bouyer <antoine.bouyer@nxp.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
This patch adds number of streams per camera as constructor parameter, and
limits stream count to 3. Some applications may need up to 3 streams for
preview + capture + video record. Currently, imx8-isi pipeline only supports
up to 2. Increase constant parameter to 3 to match these applications'
requirements.
Minimum value between default value 3, and total amount of ISI's pipes is
now applied. For SOCs which only have 1 ISI pipe (ie i.MX93), available
stream count becomes 1.
Signed-off-by: Antoine Bouyer <antoine.bouyer@nxp.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
