The KMS sink currently displays the frame buffer on the top-left corner
of the screen, resulting in either a black area on the bottom and right
sides (if the frame buffer is smaller than the display resolution) of in
a restricted field of view (if the frame buffer is larger than the
display resolution). Improve this by scaling the frame buffer to full
screen if supported, and aligning the crop rectangle to the frame buffer
center if the field of view needs to be restricted.
The implementation test for possible composition options, from best to
worst. The tests are performed when the camera is started, as testing
atomic commits requires access to frame buffer objects, which are not
available at configure time. Changing this would require either a large
refactoring of the cam application to provide frame buffers earlier, or
extending the KMS API to support testing commits with dummy buffer
objects. Both are candidates for later development.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Eric Curtin <ecurtin@redhat.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Tested-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The drmRequest is KMSSink::processRequest() is created as a naked
pointer, passed to the constructor of the KMSSink::Request object that
stores it in a std::unique_ptr<>, and used later in the function. The
current implementation is safe, but could be prone to both memory leaks
and use-after-free bugs if modified. Improve it by replacing the naked
pointer with a std::unique_ptr<>.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Eric Curtin <ecurtin@redhat.com>
Tested-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The SDL_UpdateNVTexture() function, used for NV12 texture support, has
been introduced in SDL2 2.0.16. Compiling against older SDL versions
fails with
../src/cam/sdl_texture_yuv.cpp: In member function ‘virtual void SDLTextureNV12::update(const std::vector<libcamera::Span<const unsigned char> >&)’:
../src/cam/sdl_texture_yuv.cpp:19:2: error: ‘SDL_UpdateNVTexture’ was not declared in this scope; did you mean ‘SDL_UpdateYUVTexture’?
19 | SDL_UpdateNVTexture(ptr_, &rect_, data[0].data(), pitch_,
| ^~~~~~~~~~~~~~~~~~~
| SDL_UpdateYUVTexture
Fix it with conditional compilation based on the SDL version.
Fixes: 7b8df9fe6b ("cam: sdl_sink: Add NV12 texture support")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Eric Curtin <ecurtin@redhat.com>
In preparation for the addition of NV12 support in the SDL sink, rename
the sdl_texture_yuyv.{cpp,h} files to just "yuv". Separate
sdl_texture_nv12.{cpp,h} files could be added instead, but given how
short the implementation will be, grouping all YUV formats in a single
file is better.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Eric Curtin <ecurtin@redhat.com>
Tested-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The SDLTexture::update() function isn't meant to modify the data it
receives. Make the Span type const to ensure this at compile time. While
at it, pass the Span by value instead of reference, as a Span is only a
pointer and size, which will fit in registers and will avoid pointer
dereferences in the callee.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
We were using the libjpeg functionality of SDL2_image only, instead just
use libjpeg directly to reduce our dependancy count, it is a more
commonly available library.
Signed-off-by: Eric Curtin <ecurtin@redhat.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Now that the ControlList::get() function returns an instance of
std::optional<>, we can replace the ControlList::contains() calls with a
nullopt check on the return value of get(). This avoids double lookups
of controls through the code base.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Previously, ControlList::get<T>() would use default constructed objects to
indicate that a ControlList does not have the requested Control. This has
several disadvantages: 1) It requires types to be default constructible,
2) it does not differentiate between a default constructed object and an
object that happens to have the same state as a default constructed object.
std::optional<T> additionally stores the information if the object is valid
or not, and therefore is more expressive than a default constructed object.
Signed-off-by: Christian Rauch <Rauch.Christian@gmx.de>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The SDL_PixelFormatEnum type has been introduced in libsdl by
1a4c0d4e17e6 ("Fixed bug 4377 - SDL_PIXELFORMAT enum is anonymous, which
prevents its use in a templated function") which is only available after
release 2.0.10 of the library.
Debian 10 ships libsdl at version 2.0.9 and building cam with sdl
support there fails with error:
./src/cam/sdl_texture.h:27:8: error: ‘SDL_PixelFormatEnum’ does not name
a type; did you mean ‘SDL_PixelFormat’?
Fix that by using the base type uint32_t in place of
SDL_PixelFormatEnum.
Reported-by: https://buildbot.libcamera.org/#/builders/6/builds/355
Fixes: 11554a259f ("cam: sdl_sink: Add SDL sink with initial YUYV support")
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Reviewed-by: Eric Curtin <ecurtin@redhat.com>
There is a limitation that requires input and output to be pixel
for pixel identical in terms of height and width. Remove this
limitation to enable more hardware that doesn't match. Just start
drawing from top left 0, 0 corner.
Signed-off-by: Eric Curtin <ecurtin@redhat.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Eric Curtin <ecurtin@redhat.com>
Existing code is hardcoded to card0. Since recent fedora upgrades, we
have noticed on more than one machine that card1 is present as the
lowest numbered device, could theoretically be higher. This technique
tries every file starting with card and continue only when we have
successfully opened one. These devices with card1 as the lowest device
were simply failing when they do not see a /dev/dri/card0 file present.
Reported-by: Ian Mullins <imullins@redhat.com>
Tested-by: Ian Mullins <imullins@redhat.com>
Signed-off-by: Eric Curtin <ecurtin@redhat.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
This adds more portability to existing cam sinks. You can pass a
YUYV camera buffer and SDL will handle the pixel buffer conversion
to the display. This allows cam reference implementation to display
images on VMs, Mac M1, Raspberry Pi, etc. This also enables cam
reference implementation, to run as a desktop application in Wayland or
X11. SDL also has support for Android and ChromeOS which has not been
tested. Also tested on simpledrm Raspberry Pi 4 framebuffer
successfully where existing kms sink did not work. Can also be used as
kmsdrm sink. Only supports one camera stream at present.
Signed-off-by: Eric Curtin <ecurtin@redhat.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Tested-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Add a "--script" option to the cam test application to allow specify a
capture script to be used to drive the capture session. Add to the
CameraSession class a script parser instance, created conditionally to
the OptCaptureScript option.
If the script parser has been created, use it at queueRequest time to
retrieve the list of controls that has to be associated with a Request,
and populate Request::controls() with it before queueing it to the
Camera.
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The CameraSession::processRequest() frame completion handler currently
re-queues completed requests by calling Camera::queueRequests()
explicitely.
In order to maintain the CameraSession::queuedRequest_ counter up to
date with the actual number of queued requests, call the
CameraSession::queueRequest() instead of operating the Camera directly.
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Add a parser class to the cam test application to control the capture
operations through a yaml script.
The capture script currently allow to specify a list of controls and
their associated values to be applied per-frame.
Also add a trivial capture script example to showcase the intended
script structure.
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
For consistency with UniqueFD, rename the fd() function to get().
Renaming UniqueFD::get() to fd() would have been another option, but was
rejected to keep as close as possible to the std::shared_ptr<> and
std::unique_ptr<> APIs.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Hirokazu Honda <hiroh@chromium.org>
Remove the verbose #ifndef/#define/#endif pattern for maintaining
header idempotency, and replace it with a simple #pragma once.
This simplifies the headers, and prevents redundant changes when
header files get moved.
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
"using namespace" in a header file propagates the namespace to
the files including the header file. So it should be avoided.
This removes "using namespace" in stream_options.h
Signed-off-by: Hirokazu Honda <hiroh@chromium.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The camera session keeps requeuing requests until the capture limit is
reached. This causes more request than the limit to complete, as there's
a queue of requests in flight. When capturing from multiple cameras
concurrently, this results in the captureDone signal being emitted for
every request completion after the limit is reached, instead of once per
camera session when reaching the limit.
Fix this by simply dropping any request that completes after the limit
is reached. We could instead avoid requeuing more requests than needed
to reach the limit, but that may cause request starvation in pipelines,
which are currently not handled consistently (or correctly).
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The FileSink class constructs stream names internally the same way that
the CameraSession does, except that it fails to add the camera name.
This results in files being written without the camera name.
This could be fixed in FileSink, but we would still duplicate code to
construct stream names. Pass the stream names map from CameraSession to
FileSink instead, and store it internally.
Fixes: 02001fecb0 ("cam: Turn BufferWriter into a FrameSink")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
When creating a DRM frame buffer, the dmabufs for the planes are
imported as GEM objects. For multi-planar formats, all planes may use
the same dmabuf, which results in multiple imports. This doesn't cause
any issue at import time, as DRM detects this situation and returns the
same GEM object. However, when destroying the frame buffer, the same GEM
object ends up being closed multiple times, which generates an error.
Fix this by avoiding multiple imports of the same dmabuf for the same
frame buffer. While the issue may theoretically occur with identical
dmabufs for different frame buffers, this is quite unlikely and is thus
not addressed.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The stride is not always identical for all planes for multi-planar
formats. Semi-planar YUV formats without horizontal subsampling often
have a chroma stride equal to twice the luma stride, and tri-planar YUV
formats with a 1/2 horizontal subsampling often have a chroma stride
equal to half the luma stride. This isn't correctly taken into account
when creating a DRM frame buffer, as the same stride is set for all
planes.
libcamera doesn't report per-plane stride values yet, but uses chroma
strides that match the above description for all currently supported
platforms. Calculation the chrome strides appropriately in the KMSSink
class, and pass them to DRM::createFrameBuffer().
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The new Image class represents a multi-planar image with direct access
to pixel data. It currently duplicates the function of the
MappedFrameBuffer class which is internal to libcamera, and will serve
as a design playground to improve the API until it is considered ready
to be made part of the libcamera public API.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Hirokazu Honda <hiroh@chromium.org>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The number of metadata planes should always match the number of frame
buffer planes. Enforce this by making the vector private and providing
accessor functions.
As this changes the public API, update all in-tree users accordingly.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Reviewed-by: Hirokazu Honda <hiroh@chromium.org>
Laurent Pinchart
Eric Curtin
Naushir Patuck
Christian Rauch
Jacopo Mondi
Nejc Galof
Kieran Bingham
Hirokazu Honda