Two independent instances of the ControlSerializer class are in use at
the IPC boundaries, one in the Proxy class that serializes data from the
pipeline handler to the IPA, and one in the ProxyWorker which serializes
data in the opposite direction.
Each instance operates autonomously, without any centralized point of
control, and each one assigns a numerical handle to each ControlInfoMap
it serializes. This creates a risk of potential collision on the handle
values, as both instances will use the same numerical space and
are not aware of what handles has been already used by the instance "on
the other side".
To fix that, partition the handles numerical space by initializing the
control serializer with a seed according to the role of the component
that creates the serializer and increment the handle number by 2, to
avoid any collision risk.
While this is temporary and rather hacky solution, it solves an issue
with isolated IPA modules without too much complexity added.
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
ControlInfoMap does not have a ControlId map associated, but rather
creates one with the generateIdMap() function at creation time.
As a consequence, when in the need to de-serialize a ControlInfoMap all
the ControlId it contains are created by the deserializer instance, not
being able to discern if the controls the ControlIdMap refers to are the
global libcamera controls (and properties) or instances local to the
V4L2 device that has first initialized the controls.
As a consequence the ControlId stored in a de-serialized map will always
be newly created entities, preventing lookup by ControlId reference on a
de-serialized ControlInfoMap.
In order to make it possible to use globally available ControlId
instances whenever possible, create ControlInfoMap with a reference to
an externally allocated ControlIdMap instead of generating one
internally.
As a consequence the class constructors take and additional argument,
which might be not pleasant to type in, but enforces the concepts that
ControlInfoMap should be created with controls part of the same id map.
As the ControlIdMap the ControlInfoMap refers to needs to be allocated
externally:
- Use the globally available controls::controls (or
properties::properties) id map when referring to libcamera controls
- The V4L2 device that creates ControlInfoMap by parsing the device's
controls has to allocate a ControlIdMap
- The ControlSerializer that de-serializes a ControlInfoMap has to
create and store the ControlIdMap the de-serialized info map refers to
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The ControlSerializer saves all ControlInfoMaps that it has already
(de)serialized, in order to (de)serialize ControlLists that contain the
ControlInfoMaps. Leverage this to cache ControlInfoMaps, such that the
ControlSerializer will not re-(de)serialize a ControlInfoMap that it has
previously (de)serialized.
Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
With the internal headers now in include/libcamera/internal/, we may
have identically named headers in include/libcamera/. Their header
guards would clash. Rename the header guards of internal headers to
prevent any issue.
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 libcamera internal headers are located in src/libcamera/include/.
The directory is added to the compiler headers search path with a meson
include_directories() directive, and internal headers are included with
(e.g. for the internal semaphore.h header)
#include "semaphore.h"
All was well, until libcxx decided to implement the C++20
synchronization library. The __threading_support header gained a
#include <semaphore.h>
to include the pthread's semaphore support. As include_directories()
adds src/libcamera/include/ to the compiler search path with -I, the
internal semaphore.h is included instead of the pthread version.
Needless to say, the compiler isn't happy.
Three options have been considered to fix this issue:
- Use -iquote instead of -I. The -iquote option instructs gcc to only
consider the header search path for headers included with the ""
version. Meson unfortunately doesn't support this option.
- Rename the internal semaphore.h header. This was deemed to be the
beginning of a long whack-a-mole game, where namespace clashes with
system libraries would appear over time (possibly dependent on
particular system configurations) and would need to be constantly
fixed.
- Move the internal headers to another directory to create a unique
namespace through path components. This causes lots of churn in all
the existing source files through the all project.
The first option would be best, but isn't available to us due to missing
support in meson. Even if -iquote support was added, we would need to
fix the problem before a new version of meson containing the required
support would be released.
The third option is thus the only practical solution available. Bite the
bullet, and do it, moving headers to include/libcamera/internal/.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Jacopo Mondi <jacopo@jmondi.org>
