Events have a global, system-wide sequence number, which must be handled
atomically, (fetching and incrementing the sequence AND sending the event
to uORB must be atomic). Currently in FLAT mode, only one instance of this
sequence number exists, so it is OK to have it in px4_platform.
However, in PROTECTED mode px4_platform is instantiated both in kernel-
and user spaces, which makes two instances of this sequence number, which
causes problems in the mavlink event handling logic.
When mavlink receives and handles events, it expects that:
- The sequence numbers arrive in order (seq n is followed by n+1 etc)
- It increments by 1
- There is only one instance of the sequence number
In PROTECTED mode this is violated, as the kernel and user sequence
numbers run freely on their own. This patch fixes the issue by moving
the event backend to the kernel and by providing user access to it via
ioctl.
- new ROS2 platform in PX4 intended for creating configs that build and run entirely in ROS2
- PX4_CONFIG defaults to px4_ros2_default if no config specified and in a colcon workspace with ROS_VERSION=2
- currently doesn't do much other than allow you to build px4 msgs interface package
* Made voxl2 apps processor and slpi dsp processor builds into separate board builds so that they can
more easily be configured independently.
* Removed board specific link library command from px4_config.cmake and moved it to a more generic
board specific solution that can be used by any board that needs custom link libraries.
* Removed redundant cmake command for Qurt
* Removed unused definition from Qurt cmake file
* Removed unnecessary QURT_LIB cmake function
* Reorganized the voxl2 build structure to avoid 4 level board directories.
* Reverted cmake files to remove 4 level board naming code
* Updated documentation
the nuttx and posix specific options files since this option cannot be used with
the qurt platform. There are header files in the hexagon sdk that fail this check.
- new modules/simulation directory to collect all simulators and related modules
- new Tools/simulation directory to collect and organize scattered simulation submodules, scripts, etc
- simulation module renamed to simulator_mavlink
- sih renamed to simulator_sih (not a great name, but I wanted to be clear it was a simulator)
- ignition_simulator renamed to simulator_ignition_bridge
- large sitl_target.cmake split by simulation option and in some cases pushed to appropriate modules
- sitl targets broken down to what's actually available (eg jmavsim only has 1 model and 1 world)
- new Gazebo consistently referred to as Ignition for now (probably the least confusing thing until we fully drop Gazebo classic support someday)
- much simpler direct interface using Ignition Transport
- in tree models and worlds
- control allocation output configuration, no more magic actuator mapping to mavlink and back
- currently requires no custom Gazebo plugins (keeping things as lightweight and simple as possible)
Co-authored-by: Jaeyoung-Lim <jalim@ethz.ch>
Fixes the error
googletest-src/googletest/src/gtest-death-test.cc:1283:24: error: ‘dummy’ may be used uninitialized [-Werror=maybe-uninitialized]
with GCC 11
Implement an interface for protected build to access parameters.
The implementation only does IOCTL calls to the kernel, where the parameters
live.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Define __KERNEL__ macro during compilation and place the module in separate library
Remove default library linking to m or libc on NuttX. Add these in platform layer instead, since
they are different on kernel and user side
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
This adds the env option PX4_FUZZ which runs the LLVM libFuzzer which
throws random bytes at mavlink_receiver using MAVLink messages over UDP.
The MAVLink messages that are being sent are valid, so the CRC is
calculated but the payload and msgid, etc. are generally garbage, unless
the fuzzing gets a msgid right by chance.
As I understand it, libFuzzer watches the test coverage and will try to
execute as much of the code as possible.
There was an error that PX4_CRYPTO was cached accross variants, if it was
defined just for one variant.
This synchronizes the caching of BOARD_CRYPTO with other similar BOARD_ flags;
it is set as INTERNAL and "1" when enabled.
Also remove handling of BOARD_KEYSTORE; it is not used anywhere after
changing the crypto under src/drivers. If a separate keystore driver
is required, it is just selected as any other driver in px4board
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
- cmake NuttX build wrapper compile in place instead of copying source tree to build directory
- slightly faster skipping necessary copying (depending on system)
- allows debugging in place
- easier to work directly in NuttX following official documentation
- simplifies overall build which should make it easier to resolve any remaining NuttX dependency issues in the build system
- the downside is switching back and forth between different builds always require rebuilding NuttX, but I think this is worth the improved developer experience
- also no longer builds px4io and bootloader in every single build, for most users these rarely change and we're wasting a lot of build time
- mavlink/mavlink is now directly included as a submodule instead of the generated mavlink/c_library_v2
- this also switches to mavlink development.xml by default
Ville Juven
Jukka Laitinen
Matthias Grob
Daniel Agar
Beat Küng
tanja
Zachary Lowell
Eric Katzfey
Eric Katzfey
Hamish Willee
Peter van der Perk
Julian Oes
Jacob Crabill
mcsauder