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>
The V4L2Camera::getBufferFd() method doesn't need to run in the camera
thread. Call it directly.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
In the upcoming FrameBuffer API the memory mapping of buffers will be
left to the user of the FrameBuffer objects. Prepare the V4L2
compatibility layer to this upcoming change to ease conversion to the
new API.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
With the upcoming FrameBuffer API a new library wide FrameMetadata
object will be added which will replace the specific implementation in
the V4L2 compatibility layer.
Avoid name collisions while the new FrameBuffer API is added by renaming
the V4L2 compatibility layer specific implementation until it can be
replaced with the library wide implementation.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Add a test which exercises the whole FileDescriptor interface.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
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>
Refactoring of the camera_proxy have left the 'ret' variable undeclared,
declare it.
../../src/v4l2/v4l2_camera_proxy.cpp:273:2: error: use of undeclared identifier 'ret'
ret = vcam_->invokeMethod(&V4L2Camera::configure,
^
../../src/v4l2/v4l2_camera_proxy.cpp:278:6: error: use of undeclared identifier 'ret'
if (ret < 0)
Fixes: fce110c6d9 ("v4l2: camera_proxy: Break out try_fmt")
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Acked-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Most of the ioctl handlers in the V4L2CameraProxy class have an empty
line between the tracing printouts and the immediately following buffer
type validation. Align the two occasions where such an empty line is missing
with the others.
Cosmetic change only.
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Calling vidioc_s_fmt() calls vidioc_try_fmt() duplicating prinouts.
Breakout try format procedure and call it from both functions.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
The set pixelformat field of struct v4l2_pix_format structure was
wrongly converted to PixelFormat by calling v4l2ToDrm(), with an already
converted 'format' argument.
Fix this by calling the right drmToV4l2() conversion function.
Fixes: 0ce8f2390b ("v4l2: v4l2_compat: Add V4L2 compatibility layer")
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Commit 29c5508075 ("v4l2: camera_proxy: Create format info array")
introduced usage of the std::array template class, but didn't include
the corresponding header. This may cause a compilation breakage in the
future if the indirect include of <array> disappears due to unrelated
changes. Fix it.
Fixed: 29c5508075 ("v4l2: camera_proxy: Create format info array")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Create a PixelFormatInfo structure to store information about a format,
and add a global array of format info for all the formats currently
supported. Move the format helpers to use the information from the
array.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
To clarify code, adopt the following rules for format helpers:
- All variables representing V4L2 pixel formats shall use uint32_t
- All variables representing DRM pixel formats shall use PixelFormat
- Functions returning positive values only shall not have a signed
return type
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Create a new FileOperations structure to hold all the dynamically-loaded
C library file operations. The file operations are now exposed publicly,
to prepare for usage of mmap in the V4L2CameraProxy.
A new template helper function is added to retrieve a symbol with
dlsym() with proper casting to simplify the V4L2CompatManager
constructor.
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>
Now that Object::invokeMethod() supports returning a value, use it and
drop the return value method argument.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Test that a signal can be connected to non-void static and member slots.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Test that Object::invokeMethod() can be used to invoke a non-void
method. Verify that the return value is correctly propagated to the
caller.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Most of the bound method classes are named with a BoundMethod prefix,
except for BoundMemberMethod and BoundStaticMethod. Rename them to
BoundMethodMember and BoundMethodStatic respectively to make the code
more coherent.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
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>
Create a new BoundMethodPack class to replace the PackType type alias.
This will allow adding additional fields to the arguments pack, when
adding support for propagation of bound method return values.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The sequence and generator member types of BoundMethodArgs are not
dependent on the template arguments of BoundMethodArgs. To prepare for
template specialization of BoundMethodArgs and avoid code duplication,
move them to the BoundMethodBase class.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The BoundMethodBase::connectionType() method isn't used, drop it. While
it at make the connectionType_ member private as it is only used by the
class.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Mark the activate() and invoke() methods with the override keyword where
appropriate.
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>
Test the ConnectionTypeDirect type when the object lives in a different
thread. This test passes but generates a memory leak.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The checkstyle script expects hunks to be declared with a start line and
line count, however the unified diff format [0] declares that a single
line hunk will only have the start line:
> If a hunk contains just one line, only its start line number appears.
> Otherwise its line numbers look like ‘start,count’. An empty hunk is
> considered to start at the line that follows the hunk.
[0] https://www.gnu.org/software/diffutils/manual/html_node/Detailed-Unified.html#Detailed-Unified
Attempting to parse a single line hunk results in the following error:
File "./utils/checkstyle.py", line 110, in __init__
raise RuntimeError("Malformed diff hunk header '%s'" % line)
RuntimeError: Malformed diff hunk header '@@ -1 +1,2 @@
The DiffHunk class only makes use of the start line, and does not
utilise the line count, thus update the regex to make the unused
groups optional.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The checkstyle.py patch has a fault which it identified in itself when
updating the regex string.
--- utils/checkstyle.py
+++ utils/checkstyle.py
#105: : W605 invalid escape sequence '\+'
+ diff_header_regex = re.compile('@@ -([0-9]+)(,[0-9]+)? \+([0-9]+)(,?[0-9]+)? @@')
---
1 potential style issue detected, please review
This is documented further at:
https://www.flake8rules.com/rules/W605.html
Update the string literal prefix to declare a raw byte string for the
regex parser.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The ControlSerializer::serial_ member variable isn't initialized. Add a
constructor to the class to initialize 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 __open_2() and __openat_2() functions are defined by glibc as taking
2 and 3 arguments respectively, with variadic arguments for the file
mode as open() and openat() do. The V4L2 compatibility layer defines
them as aliases for open() and openat(), which results in compilation
failures with gcc:
../../src/v4l2/v4l2_compat.cpp: In function ‘int __openat_2(int, const char*, int)’:
../../src/v4l2/v4l2_compat.cpp:58:14: error: invalid conversion from ‘int’ to ‘const char*’ [-fpermissive]
58 | return open(dirfd, path, oflag);
| ^~~~~
| |
| int
../../src/v4l2/v4l2_compat.cpp:31:39: note: initializing argument 1 of ‘int open(const char*, int, ...)’
31 | LIBCAMERA_PUBLIC int open(const char *path, int oflag, ...)
| ~~~~~~~~~~~~^~~~
../../src/v4l2/v4l2_compat.cpp:58:21: error: invalid conversion from ‘const char*’ to ‘int’ [-fpermissive]
58 | return open(dirfd, path, oflag);
| ^~~~
| |
| const char*
../../src/v4l2/v4l2_compat.cpp:31:49: note: initializing argument 2 of ‘int open(const char*, int, ...)’
31 | LIBCAMERA_PUBLIC int open(const char *path, int oflag, ...)
|
Fix this by defining the two functions as wrappers around open() and
openat() without variadic arguments.
Fixes: 0ce8f2390b ("v4l2: v4l2_compat: Add V4L2 compatibility layer")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Add libcamera V4L2 compatibility layer.
This initial implementation supports the minimal set of V4L2 operations,
which allows getting, setting, and enumerating formats, and streaming
frames from a video device. Some data about the wrapped V4L2 video
device are hardcoded.
Add a build option named 'v4l2' and adjust the build system to
selectively compile the V4L2 compatibility layer.
For now we match the V4L2 device node to a libcamera camera based on a
devnum that a pipeline handler may optionally map to a libcamera
camera.
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Register all UVC Cameras along with their device numbers, to eventually
allow the V4L2 compatibility layer to match against it.
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
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>
strlcpy is available in libbsd, bionic, musl, and ulibc, but not in glibc.
Instead of checking for strlcpy availability and modifying dependencies,
implement it in utils, as a wrapper around strncpy.
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Invoking a method that takes a reference argument with
Object::invokeMethod() results in a compilation error:
../test/object-invoke.cpp:131:11: error: no matching member function for call to 'invokeMethod'
object_.invokeMethod(&InvokedObject::methodWithReference,
~~~~~~~~^~~~~~~~~~~~
../include/libcamera/object.h:33:7: note: candidate template ignored: deduced conflicting types for parameter 'Args' (<const int &> vs. <int>)
void invokeMethod(void (T::*func)(Args...), ConnectionType type, Args... args)
This is due to the fact that implicit type conversions (from value to
reference in this case) takes place after template argument type
deduction, during overload resolution. A similar issue would occur if
T::func took a long argument and invokeMethod() was called with an in
argument.
Fix this by specifying to sets of argument types in the invokeMethod()
template, one for the arguments to the invoked method, and one for the
arguments to invokeMethod() itself. The compiler can then first perform
type deduction separately, and implicit conversion in a second step.
Reported-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Object::invokeMethod() fails with a compilation error when the invoked
method takes a reference argument. Add a test case for this issue.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
The buffer index is a V4L2 concept that will be hidden from users with
the introduction of a new FrameBuffer class. In preparation for this,
remove the index from log messages.
Keep and move one debug log message where the index is available as the
V4L2 buffer is being dequeued for the video device and it's useful when
debugging.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Expecting pipeline handler implementations of queueRequest() to call
the base class queueRequest() at the correct point have led to different
behaviors between the pipelines.
Fix this by splitting queueRequest() into a base class implementation
which handles the bookkeeping and a new queueRequestDevice() that is
to be implemented by pipeline handlers and only deals with committing the
request to the device.
Signed-off-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Object instances receive messages dispatched from the event loop of the
thread they belong to. Deleting an object from a different thread is
thus dangerous, unless the caller ensures that no message delivery is in
progress. Document this in the Object class documentation.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
When the Linux headers were updated to v5.2, the version announced in
the README was not, so update it.
Fixes: 4984973679 ("include: linux: Update headers to Linux v5.2")
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Acked-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
The InvokeObject instance is created on the stack in the run() method,
and is thus destroyed before the thread the object is bound to
terminates. This creates a race condition as the object message handler
could be running when the object is deleted. Fix this by moving the
InvokeObject to a member field.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
