- increase stack sizes to run cleanly under stackcheck
- this is likely overkill for most boards, but using stackcheck to set our minimum ensures we have a very safe margin on regular builds and it's something we can currently afford
- remove holybro_durandal-v1_stackcheck from test rack (there's only one unit)
- update to NuttX with stm32f4 and stm32f7 SPI DMA internal buffers
- remove explicit DMA buffer allocations from new IMU drivers
- restore original BOARD_DMA_ALLOC_POOL_SIZE
- decrease SPI DMA thresholds based on fmu-v2/v3/v4/v5 bench testing
instead of a static per-driver array.
Reduces BSS RAM usage by a couple of 100 Bytes (linear increase with num
drivers).
Downsides:
- a bit more runtime overhead
- less isolation, locking required
- a bit more complex
Running the sensors module out of the same WQ thread as the estimator, position, and attitude controllers is a bit safer and prevents potential priority and starvation issues. There is a very small increase in latency (~50 us) between sensors and ekf2 execution (on average). This also saves a little bit of memory (~ 3 kB) and cpu (~1-1.5% depending on the board).
Chip-select and SPI initialization uses the new config, whereas the drivers
still use the existing defines.
The configuration in board_config.h can be removed after all drivers are
updated.
The existing behavior is unexpected: if the work item is already on the
runnable queue, it will still be triggered after a call to ScheduleClear().
This can lead to race conditions.
* Introduce "px4_arch_adc_reference_v"
* Revert "Introduce "px4_arch_adc_reference_v""
This reverts commit 93691fbbd55a1b8da8c190e225b318067d90399b.
* use structure to return sample count and vref at the same time
* Revert "use structure to return sample count and vref at the same time"
This reverts commit 9cfd1c173cda51495f766a3f678c2202d67725fd.
* Revert "Revert "Introduce "px4_arch_adc_reference_v"""
This reverts commit edb7f7603e4471163ffb0fc6fc62ad2e30336e91.
* fix missed reference
* remove unecessary channel specific vref
* Update src/drivers/drv_adc.h
Co-Authored-By: David Sidrane <David.Sidrane@Nscdg.com>
* Update src/drivers/drv_adc.h
Co-Authored-By: David Sidrane <David.Sidrane@Nscdg.com>
* Introduce BOARD_ADC_POS_REF_V
Co-authored-by: David Sidrane <David.Sidrane@Nscdg.com>
- checked register mechanism and simple watchdog
- driver checks for errors gradually and can reconfigure itself
- respect IMU_GYRO_RATEMAX at the driver level
- fixed sensor INT16_MIN and INT16_MAX handling (y & z axis are flipped before publishing)
- increased sensor_gyro_fifo max size (enables running the driver much slower, but still transferring all raw data)
- PX4Accelerometer/PX4Gyroscope remove unnecessary memsets
It turns out that ekf2 needs more stack when sideslip fusion fails.
Sideslip fusion is currently only enabled for fixedwing by default and
not executed in testing.
- gyro filtering (low-pass and notch) only performed on primary gyro in `sensors/vehicle_angular_velocity` instead of every gyro in `PX4Gyroscope`
- sample rate is calculated from actual updates (the fixed value was slightly wrong in many cases, and very wrong in a few)
- In the FIFO case the array is now averaged and published in `sensor_gyro` for filtering downstream. I'll update this in the future to use the full FIFO array (if available), but right now it should be fine.
- this is a new module for temperature compensation that consolidates the functionality previously handled in the sensors module (calculating runtime thermal corrections) and the events module (online thermal calibration)
- by collecting this functionality into a single module we can optionally disable it on systems where it's not used and save some flash (if disabled at build time) or memory (disabled at run time)
The bulk of this change was tightly coupled and needed to be deleted in one pass. Some of the smaller changes were things that broke as a result of the initial purge and subsequently fixed by further eradicating unnecessary platform differences. Finally, I deleted any dead code I came across in the related files I touched while going through everything.
- DriverFramework (src/lib/DriverFramework submodule) completely removed
- added dspal submodule in qurt platform (was brought in via DriverFramework)
- all df wrapper drivers removed
- all boards using df wrapper drivers updated to use in tree equivalents
- unused empty arch/board.h on posix and qurt removed
- unused IOCTLs removed (pub block, priv, etc)
- Integrator delete methods only used from df wrapper drivers
- commander: sensor calibration use "NuttX version" everywhere for now
- sensors: update to px4_{open, read, close} instead of DevMgr wrapper (adc open for analog differential pressure)
- battery_status: update to px4_{open, read, close} instead of DevMgr wrapper (adc open for analog differential pressure)
- cdev cleanup conflicting typedefs and names with actual OS (pollevent_t, etc)
- load_mon and top remove from linux boards (unused)
- delete unused PX4_MAIN_FUNCTION
- delete unused getreg32 macro
- delete unused SIOCDEVPRIVATE define
- named each platform tasks consistently
- posix list_devices and list_topics removed (list_files now shows all virtual files)
Setting PTHREAD_EXPLICIT_SCHED was necessary on linux to create WQ threads with priorities relative to max, but unfortunately we can't rely pthread_attr_setinheritsched as it's dependent on system ulimit configuration or running privileged. Instead we can create the wq:manager at the maximum desired priority and allow each WQ thread to have a relative priority.
Beat Küng
Daniel Agar
mcsauder
Jacob Dahl
AlexKlimaj
SalimTerryLi
Matthias Grob
David Sidrane
TSC21
Julian Oes