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.
- Allocate and free the node name in uORBDeviceNode.
- Add protected build support by de-allocating the name with kmm_free, when
running in kernel side. strdup allocates from the kernel heap in NuttX kernel
space.
- Remove the CDev::unregister_driver_and_memory(), it is no longer used
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
In NuttX protected build there are separate work queues in kernel and user sides.
pthreads are only available in user side, so use tasks and kthreads for
memory protected builds.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
- use NuttX gdb script for nxthreads and thread backtrace
- update jlink_gdb_backtrace and jlink_debug_gdb helper targets to use
NuttX gdb script
- Debug/PX4 fix "perf" divide by zero
- Debug/PX4 add "dmesg"
dmesg was locking the console buffer, then writing to stdout (a pipe in
the case of the MAVLink shell).
This might block, waiting for mavlink to read from the pipe. If however
mavlink tries to write to the console at that time, the lock is already
taken.
This patch avoids nested locking by using a separate buffer.
This is more logical place for the "backend" implementation than
directly under platform.
This also allows making other implementations as "real" drivers, as well as proper configuration via Kconfigs
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Don't compare pointers to metadata, but the metadata contents.
In protected/kernel build there are two sets of metadata, on on kernel
side and another in user side. Thus the comparison of pointers would just
always fail. Compare orb_id instead
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Proxy all calls to the DeviceNode through Manager;
- This hides the DeviceNode from publishers and subscribers
- Manager can be made an interface class between user and kernel spaces in protected build
- This doesn't increase code size or harm the performance on flat build
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Make possible to define chip-select pin for internal SPI.
By defining chip-select pin, it's possible to start specific SPI device only.
This allows to have several same type of sensors on the same bus with different orientation.
- 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
In protected build, this needs to go to user-space initialization as it
calls apps (sercon) and launches mavlink.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
This is already removed from nuttx, and in posix the size of s->last_times
can be just checked with sizeof()
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Add interface functions for fetching latency buckets and counters and use
those in perf_counter.cpp. This cleans up the usage of perf counters, when variables defined in hrt_drv are not referenced directly from perf.
This also enables implementing kernel-userspace interface for those for
nuttx protected/kernel build.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Using cmake_host_system_information, grabs AVAILABLE_PHYSICAL_MEMORY and adds another job for every 1.5GB of available memory.
This is tested on a single system with 16 logical cores and 16GB RAM (~11.5GB available, reported correctly by cmake).
Beat Küng
David Sidrane
Daniel Agar
Julian Oes
PX4 BuildBot
Igor Mišić
Jukka Laitinen
Jani Paalijarvi
Guilherme Lawless