sys/sysmacros.h was an incorrect choice, which doesn't work with musl.
POSIX mandates dev_t to be defined by sys/types.h, so utilise that
header instead.
Fixes: effe4d6ced ("libcamera: camera_manager, pipeline_handler: allow retrieving cameras by device numbers")
Signed-off-by: Madhavan Krishnan <madhavan.krishnan@linaro.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
When the log severity names were added, there was only 4 characters
reserved for their printing. When the FATAL level was added, this
increased to 5, and thus both DBG and ERR can be expanded to their full
spelling. This also brings the levels in line with the representation
that can be used when calling logSetLevel().
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Move all accesses to the state_ and disconnected_ members to functions
of the Private class. This will make it easier to implement
synchronization, and simplifies the Camera class implementation.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Use the d-pointer idiom ([1], [2]) to hide the private data members from
the Camera class interface. This will ease maintaining ABI
compatibility, and prepares for the implementation of the Camera class
threading model.
The FrameBufferAllocator class accesses the Camera private data members
directly. In order to hide them, this pattern is replaced with new
private member functions in the Camera class, and the
FrameBufferAllocator is updated accordingly.
[1] https://wiki.qt.io/D-Pointer
[2] https://en.cppreference.com/w/cpp/language/pimpl
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Making CameraManager::cameras() thread-safe requires returning a copy of
the cameras vector instead of a reference. This is also required for
hot-plugging support and is thus desirable.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The V4L2VideoDevice::exportBuffers(),
PipelineHandler::exportFrameBuffers() and
FrameBufferAllocator::allocate() functions all return the number of
allocated buffers on success, but are documented as returning 0 in that
case. Fix their documentation.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The slots list is touched from most of the Signal template functions. In
order to prepare for thread-safety, move handling of the list to a small
number of non-template functions in the SignalBase class.
This incidently fixes a bug in signal disconnection handling where the
signal wasn't removed from the object's signals list, as pointed out by
the signals unit test.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
We use static local variables to indicate errors in methods that return
a const reference. The local variables can thus be const, make them so.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
ConnectionTypeBlocking always invokes the method through inter-thread
message passing, which results in deadlocks if the sender and receiver
live in the same thread. The deadlock can easily be avoided by turning
the invocation into a direct call in this case. Do so to make
ConnectionTypeBlocking easier to use when some of the senders live in
the same thread as the receiver while the other senders don't.
Extend the object-invoke test to cover this usage.
While at it reformat the documentation to avoid long \brief lines.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
The current thread ID is useful when debugging concurrency issues. Print
it in log messages. The syslog target is left out as the thread ID would
have little use there, and partly duplicates the process ID.
The log messages now look as follows.
[19:10:33.206560546] [22096] INFO Camera camera_manager.cpp:274 libcamera v0.0.0+993-32696686
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
When inserting an element with emplace(), the element is constructed
in-place with the parameters to the emplace() method being forwarded to
the constructor of the element. For std::map containers, the element is
an std::pair<const Key, T>. The constructors of std::pair<T1, T2> fall
into three categories:
(1) Default, copy and move constructors (and related versions)
(2) Constructors that take lvalue or rvalue references to T1 and T2
(3) A forwarding constructor that forwards parameters to the
constructors of T1 and T2
The first category isn't useful in most cases for std::map::emplace(),
as the caller usually doesn't have an existing std::pair<const Key, T>
for the element to be inserted.
The constructor from the third category is useful to avoid constructing
intermediate Key or T instances when the caller doesn't have them
available. This constructor takes two std::tuple arguments that contain
the arguments for the Key and T constructors, respectively. Due to
template deduction rules, usage of such a constructor couldn't be
deduced by the compiler automatically in all cases, so the constructor
takes a first argument of type std::piecewise_construct_t that lets the
caller force the usage ot the forwarding constructor (also known for
this reason as the piecewise constructor). The caller uses a construct
such as
map.emplace(std::piecewise_construct,
std::forward_as_tuple(args_for_Key, ...),
std::forward_as_tuple(args_for_T, ...));
This syntax is a bit heavy, but is required to construct Key and T
in-place from arguments to their non-default constructor (it is also the
only std::pair non-default constructor that can be used for non-copyable
non-movable types).
When the caller of std::map::emplace() already has references to a Key
and a T, they can be passed to the std::pair piecewise constructor, and
this will create std::tuple instance to wrap the Key and T references
arguments to ultimately pass them to the Key and T copy constructors.
map.emplace(std::piecewise_construct,
std::forward_as_tuple(Key_value),
std::forward_as_tuple(T_value));
While this mechanism works, it's unnecessary complex. A constructor of
std::pair that takes references to Key and T can be used without any
performance penalty, as it will also call the copy constructor of Key
and T. In this case we can use a simpler constructor of std::pair, and
thus a simpler call of std::map::emplace.
map.emplace(Key_value, T_value);
We have a couple occurrences of this above misuse of piecewise
construction. Simplify them, which simplifies the code and reduces the
generated code size.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The gen-controls.py script hardcodes the path to the python interpreter
to /usr/bin/python3 in the first line of the script. This hardcodes
usage of the host python3, even when building in cross-compilation
environments that may ship their own version of python. Fix it by
setting the interpreter to '/usr/bin/env python3'.
Reported-by: Madhavan Krishnan <madhavan.krishnan@linaro.org>
Suggested-by: Nicolas Dufresne <nicolas@ndufresne.ca>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Doxygen generates the following warning:
src/libcamera/ipa_interface.cpp:262: warning: explicit link request to 'dup()' could not be resolved
Fix it by disabling link generation by prefixing the function name with
a %.
Fixes: 4b9bd6c3ad ("libcamera: ipa_interface: Document the ownership of dmabufs passed to map_buffers()")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The ownership of the dmabuf file handles passed to map_buffers() is not
clear. Explicitly document that they are borrowed from the caller and
only guaranteed to be valid for the duration of the map_buffers() call.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
With the FrameBuffer rework completed there is no reason to keep the
camera prepared state around as buffer allocations are now decoupled
from the camera state. Remove the camera state simplifying the API.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
With the FrameBuffer interface in place there is no need for the Camera
to call into the specific pipelines allocation and freeing of buffers as
it no longer needs to be synchronized with buffer allocation by the
application.
Remove the function prototypes in the pipeline handler base class and
fold the functionality in the pipelines start() and stop() functions
where needed. A follow up patch will remove the now no-op
Camera::allocateBuffers() and Camera::freeBuffers().
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The Buffer interface is no longer in use and can be removed. While doing
so clean up the two odd names (dequeueFrameBuffer() and
queuedFrameBuffers_) that had to be used when adding the FrameBuffer
interface.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Switch to the FrameBuffer interface where all buffers are treated as
external buffers and are allocated outside the camera. Applications
allocating buffers using libcamera are switched to use the
FrameBufferAllocator helper.
Follow-up changes to this one will finalize the transition to the new
FrameBuffer interface by removing code that is left unused after this
change.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The FrameBuffer interface is based on the idea that all buffers are
allocated externally to libcamera and are only used by it. This is meant
to create a simpler API centered around usage of buffers, regardless of
where they come from.
Linux however lacks a centralized allocator at the moment, and not all
users of libcamera are expected to use another device that could provide
suitable buffers for the camera. This patch thus adds a helper class to
allocate buffers internally in libcamera, in a way that matches the
needs of the FrameBuffer-based API.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Extend the pipeline handlers to support the FrameBuffer API with three
new methods to handle allocation, importing and freeing of buffers. The
new methods will replace allocateBuffers() and freeBuffers().
The FrameBuffer API will use the methods on a stream level and either
allocate or import buffers for each active stream controlled from the
Camera class and an upcoming FrameBufferAllocator helper. With this new
API the implementation in pipeline handlers can be made simpler as all
streams don't need to be handled in allocateBuffers().
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The V4L2VideoDevice class can now operate using a FrameBuffer interface,
switch the IPU3 CIO2 and statistics buffer to use it. We can not convert
the application-facing buffers yet.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The V4L2VideoDevice class can now operate using a FrameBuffer interface,
switch the RkISP1 statistics and parameters buffer to use it. We can not
convert the application-facing buffers yet.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
It's common for applications to create and queue a new request in a
previous request completion handler. When the new request gets queued to
the RkISP1 pipeline handler it tries to find a parameters and statistic
buffer to be used with the request. The problem is if the pipeline depth
is already filled there are no internal buffers free to be used by the
new request.
This was solved by allocation one more parameters and statistic buffer
then the pipeline depth, this is waste full. Instead free the resources
of the request that has completed before it is signaled to the
application, this way if the pipeline depth is full it can reuse the
internal resources and the wasteful allocation can be removed.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Add a new interface in parallel with the existing Buffer implementation
to also support FrameBuffer. The reason it's added in parallel is to aid
in the migration from Buffer to FrameBuffer throughout libcamera. With
this change discrete parts of libcamera can be migrated and tested
independently.
As the new interface is added in parallel there are some oddities in
this change which will be undone in a follow up patch once libcamera
have migrated away from the Buffer interface.
- There is a nasty hack in V4L2VideoDevice::bufferAvailable(). It is
needed to allow both interfaces to exist and function at the same
time. The idea is if buffers are allocated using the FrameBuffer
interface V4L2VideoDevice::cache_ is set and we know to call the
FrameBuffer 'buffer ready' signal, and likewise if it's not to call
the Buffer variant.
- There is some code duplication between the two interfaces as they aim
to solve the same thing in slightly different ways. As all Buffer
related code is soon to be removed no effort to create code sharing
between them have been made.
- Some function and variables which can't be distinguished by their
argument types have been given a frameBuffer prefix instead of a
buffer prefix. They are clearly documented in the code and will be
renamed to the correct buffer prefix when the Buffer interface is
removed.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
In preparation for the FrameBuffer interface add a class that will deal
with keeping the cache between dmabuf file descriptors and V4L2 video
device buffer indexes.
This initial implementation ensures that no hot association is lost
while its eviction strategy could be improved in the future.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Move the metadata retrieved when dequeuing a V4L2 buffer into a
FrameMetadata object. This is done as a step to migrate to the
FrameBuffer interface as the functions added to Buffer around
FrameMetadata match the ones in FrameBuffer.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
In the FrameBuffer interface the stream will not be available from the
buffer object as the buffer might be allocated externally. The
application needs to explicitly state which stream the buffer is being
added for to the request.
Extend the addBuffer() function to get this information explicitly from
the caller.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The part in createPlane() that exports a dma buffer from a video device
will be used directly by the FrameBuffer interface. Break it out to a
separate function.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reading V4L2VideoDevice::queueBuffer() is confusing since buf.type is
first set to bufferType_ but then both variables are used in V4L2 macros
to operate based on which type of buffer is being processed. Align on
only using buf.type since it has the most existing users.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
There is no need to have a private helper function to access a private
data member when a friend statement is needed anyhow. Remove the helper
function to simplify the code and make it clear that a private member of
Buffer is accessed.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
It is not libcamera's responsibility to handle memory mappings. Switch
from the soon to be removed Plane class which deals with memory
mappings to FrameBuffer::Plane which just describes it. This makes the
transition to the full FrameBuffer easier.
As the full FrameBuffer interface has not yet spread to all parts of
libcamera core it is hard to create efficient caching of memory mappings
in the qcam application. This will be fixed in a later patch, for now
the dmabuf is mapped and unmapped each time it is seen by the
application.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Switch the IPA interfaces and implementations to use the Framebuffer
interface.
- The IPA interface is switched to use the simpler FrameBuffer::Plane
container when carrying dmabuf descriptions (fd and length) over the
pipeline/IPA boundary.
- The RkISP1 IPA implementation takes advantage of the new simpler and
safer (better control over file descriptors) FrameBuffer interface.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Add a new FrameBuffer class to describe memory used to store frames.
This change just adds the new interface, future patches will migrate all
parts of libcamera to use this and replace the old Buffer interface.
This change needs to specify the const keyword for Plane::length() as to
not get Doxygen confused with FrameBuffer::Plane::length added in this
change.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
With the introduction of FrameBuffer objects, the metadata information
related to captured frames will not be stored directly in the frame
buffer object. Add a new FrameMetadata class to hold this information.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Add a helper to make it easier to pass file descriptors around. The
helper class duplicates the fd which decouples it from the original fd
which could be closed by its owner while the new FileDescriptor remains
valid.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The association of buffers to a request can be done directly in
addBuffer() instead of when the request is queued to the camera. Keep
the check that a request contains buffers by moving it to
Camera::queueRequest() where prepare() was previously called.
As a bonus we can remove a friend statement in Request.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Propagate the return value of the bound method all the way to the caller
of activate(). The value is stored in the arguments pack for indirect
invocation.
As C++ doesn't allow instantiating a variable of type void, we need to
specialize the template class BoundMethodPack for methods returning
void. This in turn requires template specialization for the
BoundMethodArgs class in order to store the return value in the pack,
and for the BoundMemberMethod class to extract the return value from the
pack.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The bound method arguments pack will need to be accessed by the method
invoker in order to retrieve the method return value when using a
blocking connection type. We thus can't delete the pack unconditionally
in the bound method target thread. We also can't delete it
unconditionally in the invoker's thread, as for queued connections the
pack will be used in the target thread after the invoker completes.
This shows that ownership of the arguments pack is shared between two
contexts. As a result, manage it using std::shared_ptr<>.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The bound method implementation is restricted to binding to void methods
as return values are not supported. This complicates usage of bound
methods, as non-void methods used a slots or Object::invokeMethod()
targets need to be wrapped in a void method. Simplify this by supporting
arbitrary return types and ignoring the return value.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
When BoundMethodBase::activatePack() is called with the connection type
set to ConnectionTypeDirect, the method isn't deleted even if
deleteMethod is true, as is the case when called from
Object::invokeMethod(). This causes a memory leak. Fix it.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
IPCUnixSocket::send() sends a IPCUnixSocket::Header allocated on the
stack. All the fields of the header are initialized, but the padding
bytes are not. This results in random data being sent over the UNIX
socket, potentially leaking information.
Fix this by initializing the whole header to 0.
Fixes: 13dd7a01ec ("libcamera: ipc: unix: Add a IPC mechanism based on Unix sockets")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The V4L2 compatibility layer will need a way to map device numbers to
libcamera Camera instances. Expose a method in the camera manager to
retrieve Camera instances by devnum. The mapping from device numbers to
Camera instances is optionally declared by pipeline handlers when they
register cameras with the camera manager.
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
We are preparing to integrate the V4L2 adaptation layer, which will
intercept open() calls (among others) via LD_PRELOAD. To prevent
libcamera's own open() calls from being intercepted, replace them with a
direct syscall.
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
